Lady’s Gitweb - Gitweb/blob - gitweb.perl

   1 #!/usr/bin/perl
   2
   3 # gitweb - simple web interface to track changes in git repositories
   4 #
   5 # (C) 2005-2006, Kay Sievers <kay.sievers@vrfy.org>
   6 # (C) 2005, Christian Gierke
   7 #
   8 # This program is licensed under the GPLv2
   9
  10 use 5.008;
  11 use strict;
  12 use warnings;
  13 # handle ACL in file access tests
  14 use filetest 'access';
  15 use CGI qw(:standard :escapeHTML -nosticky);
  16 use CGI::Util qw(unescape);
  17 use CGI::Carp qw(fatalsToBrowser set_message);
  18 use Encode;
  19 use Fcntl ':mode';
  20 use File::Find qw();
  21 use File::Basename qw(basename);
  22 use Time::HiRes qw(gettimeofday tv_interval);
  23 use Digest::MD5 qw(md5_hex);
  24
  25 binmode STDOUT, ':utf8';
  26
  27 if (!defined($CGI::VERSION) || $CGI::VERSION < 4.08) {
  28         eval 'sub CGI::multi_param { CGI::param(@_) }'
  29 }
  30
  31 our $t0 = [ gettimeofday() ];
  32 our $number_of_git_cmds = 0;
  33
  34 BEGIN {
  35         CGI->compile() if $ENV{'MOD_PERL'};
  36 }
  37
  38 our $version = "++GIT_VERSION++";
  39
  40 our ($my_url, $my_uri, $base_url, $path_info, $home_link);
  41 sub evaluate_uri {
  42         our $cgi;
  43
  44         our $my_url = $cgi->url();
  45         our $my_uri = $cgi->url(-absolute => 1);
  46
  47         # Base URL for relative URLs in gitweb ($logo, $favicon, ...),
  48         # needed and used only for URLs with nonempty PATH_INFO
  49         our $base_url = $my_url;
  50
  51         # When the script is used as DirectoryIndex, the URL does not contain the name
  52         # of the script file itself, and $cgi->url() fails to strip PATH_INFO, so we
  53         # have to do it ourselves. We make $path_info global because it's also used
  54         # later on.
  55         #
  56         # Another issue with the script being the DirectoryIndex is that the resulting
  57         # $my_url data is not the full script URL: this is good, because we want
  58         # generated links to keep implying the script name if it wasn't explicitly
  59         # indicated in the URL we're handling, but it means that $my_url cannot be used
  60         # as base URL.
  61         # Therefore, if we needed to strip PATH_INFO, then we know that we have
  62         # to build the base URL ourselves:
  63         our $path_info = decode_utf8($ENV{"PATH_INFO"});
  64         if ($path_info) {
  65                 # $path_info has already been URL-decoded by the web server, but
  66                 # $my_url and $my_uri have not. URL-decode them so we can properly
  67                 # strip $path_info.
  68                 $my_url = unescape($my_url);
  69                 $my_uri = unescape($my_uri);
  70                 if ($my_url =~ s,\Q$path_info\E$,, &&
  71                     $my_uri =~ s,\Q$path_info\E$,, &&
  72                     defined $ENV{'SCRIPT_NAME'}) {
  73                         $base_url = $cgi->url(-base => 1) . $ENV{'SCRIPT_NAME'};
  74                 }
  75         }
  76
  77         # target of the home link on top of all pages
  78         our $home_link = $my_uri || "/";
  79 }
  80
  81 # core git executable to use
  82 # this can just be "git" if your webserver has a sensible PATH
  83 our $GIT = "++GIT_BINDIR++/git";
  84
  85 # absolute fs-path which will be prepended to the project path
  86 #our $projectroot = "/pub/scm";
  87 our $projectroot = "++GITWEB_PROJECTROOT++";
  88
  89 # fs traversing limit for getting project list
  90 # the number is relative to the projectroot
  91 our $project_maxdepth = "++GITWEB_PROJECT_MAXDEPTH++";
  92
  93 # string of the home link on top of all pages
  94 our $home_link_str = "++GITWEB_HOME_LINK_STR++";
  95
  96 # extra breadcrumbs preceding the home link
  97 our @extra_breadcrumbs = ();
  98
  99 # name of your site or organization to appear in page titles
 100 # replace this with something more descriptive for clearer bookmarks
 101 our $site_name = "++GITWEB_SITENAME++"
 102                  || ($ENV{'SERVER_NAME'} || "Untitled") . " Git";
 103
 104 # html snippet to include in the <head> section of each page
 105 our $site_html_head_string = "++GITWEB_SITE_HTML_HEAD_STRING++";
 106 # filename of html text to include at top of each page
 107 our $site_header = "++GITWEB_SITE_HEADER++";
 108 # html text to include at home page
 109 our $home_text = "++GITWEB_HOMETEXT++";
 110 # filename of html text to include at bottom of each page
 111 our $site_footer = "++GITWEB_SITE_FOOTER++";
 112
 113 # URI of stylesheets
 114 our @stylesheets = ("++GITWEB_CSS++");
 115 # URI of a single stylesheet, which can be overridden in GITWEB_CONFIG.
 116 our $stylesheet = undef;
 117
 118 # URI of GIT logo (72x27 size)
 119 our $logo = "++GITWEB_LOGO++";
 120 # URI of GIT favicon, assumed to be image/png type
 121 our $favicon = "++GITWEB_FAVICON++";
 122 # URI of gitweb.js (JavaScript code for gitweb)
 123 our $javascript = "++GITWEB_JS++";
 124
 125 # URI and label (title) of GIT logo link
 126 #our $logo_url = "http://www.kernel.org/pub/software/scm/git/docs/";
 127 #our $logo_label = "git documentation";
 128 our $logo_url = "http://git-scm.com/";
 129 our $logo_label = "git homepage";
 130
 131 # source of projects list
 132 our $projects_list = "++GITWEB_LIST++";
 133
 134 # the width (in characters) of the projects list "Description" column
 135 our $projects_list_description_width = 25;
 136
 137 # group projects by category on the projects list
 138 # (enabled if this variable evaluates to true)
 139 our $projects_list_group_categories = 0;
 140
 141 # default category if none specified
 142 # (leave the empty string for no category)
 143 our $project_list_default_category = "";
 144
 145 # default order of projects list
 146 # valid values are none, project, descr, owner, and age
 147 our $default_projects_order = "project";
 148
 149 # show repository only if this file exists
 150 # (only effective if this variable evaluates to true)
 151 our $export_ok = "++GITWEB_EXPORT_OK++";
 152
 153 # don't generate age column on the projects list page
 154 our $omit_age_column = 0;
 155
 156 # don't generate information about owners of repositories
 157 our $omit_owner=0;
 158
 159 # show repository only if this subroutine returns true
 160 # when given the path to the project, for example:
 161 #    sub { return -e "$_[0]/git-daemon-export-ok"; }
 162 our $export_auth_hook = undef;
 163
 164 # only allow viewing of repositories also shown on the overview page
 165 our $strict_export = "++GITWEB_STRICT_EXPORT++";
 166
 167 # list of git base URLs used for URL to where fetch project from,
 168 # i.e. full URL is "$git_base_url/$project"
 169 our @git_base_url_list = grep { $_ ne '' } ("++GITWEB_BASE_URL++");
 170
 171 # default blob_plain mimetype and default charset for text/plain blob
 172 our $default_blob_plain_mimetype = 'text/plain';
 173 our $default_text_plain_charset  = undef;
 174
 175 # file to use for guessing MIME types before trying /etc/mime.types
 176 # (relative to the current git repository)
 177 our $mimetypes_file = undef;
 178
 179 # assume this charset if line contains non-UTF-8 characters;
 180 # it should be valid encoding (see Encoding::Supported(3pm) for list),
 181 # for which encoding all byte sequences are valid, for example
 182 # 'iso-8859-1' aka 'latin1' (it is decoded without checking, so it
 183 # could be even 'utf-8' for the old behavior)
 184 our $fallback_encoding = 'latin1';
 185
 186 # rename detection options for git-diff and git-diff-tree
 187 # - default is '-M', with the cost proportional to
 188 #   (number of removed files) * (number of new files).
 189 # - more costly is '-C' (which implies '-M'), with the cost proportional to
 190 #   (number of changed files + number of removed files) * (number of new files)
 191 # - even more costly is '-C', '--find-copies-harder' with cost
 192 #   (number of files in the original tree) * (number of new files)
 193 # - one might want to include '-B' option, e.g. '-B', '-M'
 194 our @diff_opts = ('-M'); # taken from git_commit
 195
 196 # Disables features that would allow repository owners to inject script into
 197 # the gitweb domain.
 198 our $prevent_xss = 0;
 199
 200 # Path to the highlight executable to use (must be the one from
 201 # http://www.andre-simon.de due to assumptions about parameters and output).
 202 # Useful if highlight is not installed on your webserver's PATH.
 203 # [Default: highlight]
 204 our $highlight_bin = "++HIGHLIGHT_BIN++";
 205
 206 # information about snapshot formats that gitweb is capable of serving
 207 our %known_snapshot_formats = (
 208         # name => {
 209         #       'display' => display name,
 210         #       'type' => mime type,
 211         #       'suffix' => filename suffix,
 212         #       'format' => --format for git-archive,
 213         #       'compressor' => [compressor command and arguments]
 214         #                       (array reference, optional)
 215         #       'disabled' => boolean (optional)}
 216         #
 217         'tgz' => {
 218                 'display' => 'tar.gz',
 219                 'type' => 'application/x-gzip',
 220                 'suffix' => '.tar.gz',
 221                 'format' => 'tar',
 222                 'compressor' => ['gzip', '-n']},
 223
 224         'tbz2' => {
 225                 'display' => 'tar.bz2',
 226                 'type' => 'application/x-bzip2',
 227                 'suffix' => '.tar.bz2',
 228                 'format' => 'tar',
 229                 'compressor' => ['bzip2']},
 230
 231         'txz' => {
 232                 'display' => 'tar.xz',
 233                 'type' => 'application/x-xz',
 234                 'suffix' => '.tar.xz',
 235                 'format' => 'tar',
 236                 'compressor' => ['xz'],
 237                 'disabled' => 1},
 238
 239         'zip' => {
 240                 'display' => 'zip',
 241                 'type' => 'application/x-zip',
 242                 'suffix' => '.zip',
 243                 'format' => 'zip'},
 244 );
 245
 246 # Aliases so we understand old gitweb.snapshot values in repository
 247 # configuration.
 248 our %known_snapshot_format_aliases = (
 249         'gzip'  => 'tgz',
 250         'bzip2' => 'tbz2',
 251         'xz'    => 'txz',
 252
 253         # backward compatibility: legacy gitweb config support
 254         'x-gzip' => undef, 'gz' => undef,
 255         'x-bzip2' => undef, 'bz2' => undef,
 256         'x-zip' => undef, '' => undef,
 257 );
 258
 259 # Pixel sizes for icons and avatars. If the default font sizes or lineheights
 260 # are changed, it may be appropriate to change these values too via
 261 # $GITWEB_CONFIG.
 262 our %avatar_size = (
 263         'default' => 16,
 264         'double'  => 32
 265 );
 266
 267 # Used to set the maximum load that we will still respond to gitweb queries.
 268 # If server load exceed this value then return "503 server busy" error.
 269 # If gitweb cannot determined server load, it is taken to be 0.
 270 # Leave it undefined (or set to 'undef') to turn off load checking.
 271 our $maxload = 300;
 272
 273 # configuration for 'highlight' (http://www.andre-simon.de/)
 274 # match by basename
 275 our %highlight_basename = (
 276         #'Program' => 'py',
 277         #'Library' => 'py',
 278         'SConstruct' => 'py', # SCons equivalent of Makefile
 279         'Makefile' => 'make',
 280 );
 281 # match by extension
 282 our %highlight_ext = (
 283         # main extensions, defining name of syntax;
 284         # see files in /usr/share/highlight/langDefs/ directory
 285         (map { $_ => $_ } qw(py rb java css js tex bib xml awk bat ini spec tcl sql)),
 286         # alternate extensions, see /etc/highlight/filetypes.conf
 287         (map { $_ => 'c'   } qw(c h)),
 288         (map { $_ => 'sh'  } qw(sh bash zsh ksh)),
 289         (map { $_ => 'cpp' } qw(cpp cxx c++ cc)),
 290         (map { $_ => 'php' } qw(php php3 php4 php5 phps)),
 291         (map { $_ => 'pl'  } qw(pl perl pm)), # perhaps also 'cgi'
 292         (map { $_ => 'make'} qw(make mak mk)),
 293         (map { $_ => 'xml' } qw(xml xhtml html htm)),
 294 );
 295
 296 # You define site-wide feature defaults here; override them with
 297 # $GITWEB_CONFIG as necessary.
 298 our %feature = (
 299         # feature => {
 300         #       'sub' => feature-sub (subroutine),
 301         #       'override' => allow-override (boolean),
 302         #       'default' => [ default options...] (array reference)}
 303         #
 304         # if feature is overridable (it means that allow-override has true value),
 305         # then feature-sub will be called with default options as parameters;
 306         # return value of feature-sub indicates if to enable specified feature
 307         #
 308         # if there is no 'sub' key (no feature-sub), then feature cannot be
 309         # overridden
 310         #
 311         # use gitweb_get_feature(<feature>) to retrieve the <feature> value
 312         # (an array) or gitweb_check_feature(<feature>) to check if <feature>
 313         # is enabled
 314
 315         # Enable the 'blame' blob view, showing the last commit that modified
 316         # each line in the file. This can be very CPU-intensive.
 317
 318         # To enable system wide have in $GITWEB_CONFIG
 319         # $feature{'blame'}{'default'} = [1];
 320         # To have project specific config enable override in $GITWEB_CONFIG
 321         # $feature{'blame'}{'override'} = 1;
 322         # and in project config gitweb.blame = 0|1;
 323         'blame' => {
 324                 'sub' => sub { feature_bool('blame', @_) },
 325                 'override' => 0,
 326                 'default' => [0]},
 327
 328         # Enable the 'snapshot' link, providing a compressed archive of any
 329         # tree. This can potentially generate high traffic if you have large
 330         # project.
 331
 332         # Value is a list of formats defined in %known_snapshot_formats that
 333         # you wish to offer.
 334         # To disable system wide have in $GITWEB_CONFIG
 335         # $feature{'snapshot'}{'default'} = [];
 336         # To have project specific config enable override in $GITWEB_CONFIG
 337         # $feature{'snapshot'}{'override'} = 1;
 338         # and in project config, a comma-separated list of formats or "none"
 339         # to disable.  Example: gitweb.snapshot = tbz2,zip;
 340         'snapshot' => {
 341                 'sub' => \&feature_snapshot,
 342                 'override' => 0,
 343                 'default' => ['tgz']},
 344
 345         # Enable text search, which will list the commits which match author,
 346         # committer or commit text to a given string.  Enabled by default.
 347         # Project specific override is not supported.
 348         #
 349         # Note that this controls all search features, which means that if
 350         # it is disabled, then 'grep' and 'pickaxe' search would also be
 351         # disabled.
 352         'search' => {
 353                 'override' => 0,
 354                 'default' => [1]},
 355
 356         # Enable grep search, which will list the files in currently selected
 357         # tree containing the given string. Enabled by default. This can be
 358         # potentially CPU-intensive, of course.
 359         # Note that you need to have 'search' feature enabled too.
 360
 361         # To enable system wide have in $GITWEB_CONFIG
 362         # $feature{'grep'}{'default'} = [1];
 363         # To have project specific config enable override in $GITWEB_CONFIG
 364         # $feature{'grep'}{'override'} = 1;
 365         # and in project config gitweb.grep = 0|1;
 366         'grep' => {
 367                 'sub' => sub { feature_bool('grep', @_) },
 368                 'override' => 0,
 369                 'default' => [1]},
 370
 371         # Enable the pickaxe search, which will list the commits that modified
 372         # a given string in a file. This can be practical and quite faster
 373         # alternative to 'blame', but still potentially CPU-intensive.
 374         # Note that you need to have 'search' feature enabled too.
 375
 376         # To enable system wide have in $GITWEB_CONFIG
 377         # $feature{'pickaxe'}{'default'} = [1];
 378         # To have project specific config enable override in $GITWEB_CONFIG
 379         # $feature{'pickaxe'}{'override'} = 1;
 380         # and in project config gitweb.pickaxe = 0|1;
 381         'pickaxe' => {
 382                 'sub' => sub { feature_bool('pickaxe', @_) },
 383                 'override' => 0,
 384                 'default' => [1]},
 385
 386         # Enable showing size of blobs in a 'tree' view, in a separate
 387         # column, similar to what 'ls -l' does.  This cost a bit of IO.
 388
 389         # To disable system wide have in $GITWEB_CONFIG
 390         # $feature{'show-sizes'}{'default'} = [0];
 391         # To have project specific config enable override in $GITWEB_CONFIG
 392         # $feature{'show-sizes'}{'override'} = 1;
 393         # and in project config gitweb.showsizes = 0|1;
 394         'show-sizes' => {
 395                 'sub' => sub { feature_bool('showsizes', @_) },
 396                 'override' => 0,
 397                 'default' => [1]},
 398
 399         # Make gitweb use an alternative format of the URLs which can be
 400         # more readable and natural-looking: project name is embedded
 401         # directly in the path and the query string contains other
 402         # auxiliary information. All gitweb installations recognize
 403         # URL in either format; this configures in which formats gitweb
 404         # generates links.
 405
 406         # To enable system wide have in $GITWEB_CONFIG
 407         # $feature{'pathinfo'}{'default'} = [1];
 408         # Project specific override is not supported.
 409
 410         # Note that you will need to change the default location of CSS,
 411         # favicon, logo and possibly other files to an absolute URL. Also,
 412         # if gitweb.cgi serves as your indexfile, you will need to force
 413         # $my_uri to contain the script name in your $GITWEB_CONFIG.
 414         'pathinfo' => {
 415                 'override' => 0,
 416                 'default' => [0]},
 417
 418         # Make gitweb consider projects in project root subdirectories
 419         # to be forks of existing projects. Given project $projname.git,
 420         # projects matching $projname/*.git will not be shown in the main
 421         # projects list, instead a '+' mark will be added to $projname
 422         # there and a 'forks' view will be enabled for the project, listing
 423         # all the forks. If project list is taken from a file, forks have
 424         # to be listed after the main project.
 425
 426         # To enable system wide have in $GITWEB_CONFIG
 427         # $feature{'forks'}{'default'} = [1];
 428         # Project specific override is not supported.
 429         'forks' => {
 430                 'override' => 0,
 431                 'default' => [0]},
 432
 433         # Insert custom links to the action bar of all project pages.
 434         # This enables you mainly to link to third-party scripts integrating
 435         # into gitweb; e.g. git-browser for graphical history representation
 436         # or custom web-based repository administration interface.
 437
 438         # The 'default' value consists of a list of triplets in the form
 439         # (label, link, position) where position is the label after which
 440         # to insert the link and link is a format string where %n expands
 441         # to the project name, %f to the project path within the filesystem,
 442         # %h to the current hash (h gitweb parameter) and %b to the current
 443         # hash base (hb gitweb parameter); %% expands to %.
 444
 445         # To enable system wide have in $GITWEB_CONFIG e.g.
 446         # $feature{'actions'}{'default'} = [('graphiclog',
 447         #       '/git-browser/by-commit.html?r=%n', 'summary')];
 448         # Project specific override is not supported.
 449         'actions' => {
 450                 'override' => 0,
 451                 'default' => []},
 452
 453         # Allow gitweb scan project content tags of project repository,
 454         # and display the popular Web 2.0-ish "tag cloud" near the projects
 455         # list.  Note that this is something COMPLETELY different from the
 456         # normal Git tags.
 457
 458         # gitweb by itself can show existing tags, but it does not handle
 459         # tagging itself; you need to do it externally, outside gitweb.
 460         # The format is described in git_get_project_ctags() subroutine.
 461         # You may want to install the HTML::TagCloud Perl module to get
 462         # a pretty tag cloud instead of just a list of tags.
 463
 464         # To enable system wide have in $GITWEB_CONFIG
 465         # $feature{'ctags'}{'default'} = [1];
 466         # Project specific override is not supported.
 467
 468         # In the future whether ctags editing is enabled might depend
 469         # on the value, but using 1 should always mean no editing of ctags.
 470         'ctags' => {
 471                 'override' => 0,
 472                 'default' => [0]},
 473
 474         # The maximum number of patches in a patchset generated in patch
 475         # view. Set this to 0 or undef to disable patch view, or to a
 476         # negative number to remove any limit.
 477
 478         # To disable system wide have in $GITWEB_CONFIG
 479         # $feature{'patches'}{'default'} = [0];
 480         # To have project specific config enable override in $GITWEB_CONFIG
 481         # $feature{'patches'}{'override'} = 1;
 482         # and in project config gitweb.patches = 0|n;
 483         # where n is the maximum number of patches allowed in a patchset.
 484         'patches' => {
 485                 'sub' => \&feature_patches,
 486                 'override' => 0,
 487                 'default' => [16]},
 488
 489         # Avatar support. When this feature is enabled, views such as
 490         # shortlog or commit will display an avatar associated with
 491         # the email of the committer(s) and/or author(s).
 492
 493         # Currently available providers are gravatar and picon.
 494         # If an unknown provider is specified, the feature is disabled.
 495
 496         # Picon currently relies on the indiana.edu database.
 497
 498         # To enable system wide have in $GITWEB_CONFIG
 499         # $feature{'avatar'}{'default'} = ['<provider>'];
 500         # where <provider> is either gravatar or picon.
 501         # To have project specific config enable override in $GITWEB_CONFIG
 502         # $feature{'avatar'}{'override'} = 1;
 503         # and in project config gitweb.avatar = <provider>;
 504         'avatar' => {
 505                 'sub' => \&feature_avatar,
 506                 'override' => 0,
 507                 'default' => ['']},
 508
 509         # Enable displaying how much time and how many git commands
 510         # it took to generate and display page.  Disabled by default.
 511         # Project specific override is not supported.
 512         'timed' => {
 513                 'override' => 0,
 514                 'default' => [0]},
 515
 516         # Enable turning some links into links to actions which require
 517         # JavaScript to run (like 'blame_incremental').  Not enabled by
 518         # default.  Project specific override is currently not supported.
 519         'javascript-actions' => {
 520                 'override' => 0,
 521                 'default' => [0]},
 522
 523         # Enable and configure ability to change common timezone for dates
 524         # in gitweb output via JavaScript.  Enabled by default.
 525         # Project specific override is not supported.
 526         'javascript-timezone' => {
 527                 'override' => 0,
 528                 'default' => [
 529                         'local',     # default timezone: 'utc', 'local', or '(-|+)HHMM' format,
 530                                      # or undef to turn off this feature
 531                         'gitweb_tz', # name of cookie where to store selected timezone
 532                         'datetime',  # CSS class used to mark up dates for manipulation
 533                 ]},
 534
 535         # Syntax highlighting support. This is based on Daniel Svensson's
 536         # and Sham Chukoury's work in gitweb-xmms2.git.
 537         # It requires the 'highlight' program present in $PATH,
 538         # and therefore is disabled by default.
 539
 540         # To enable system wide have in $GITWEB_CONFIG
 541         # $feature{'highlight'}{'default'} = [1];
 542
 543         'highlight' => {
 544                 'sub' => sub { feature_bool('highlight', @_) },
 545                 'override' => 0,
 546                 'default' => [0]},
 547
 548         # Enable displaying of remote heads in the heads list
 549
 550         # To enable system wide have in $GITWEB_CONFIG
 551         # $feature{'remote_heads'}{'default'} = [1];
 552         # To have project specific config enable override in $GITWEB_CONFIG
 553         # $feature{'remote_heads'}{'override'} = 1;
 554         # and in project config gitweb.remoteheads = 0|1;
 555         'remote_heads' => {
 556                 'sub' => sub { feature_bool('remote_heads', @_) },
 557                 'override' => 0,
 558                 'default' => [0]},
 559
 560         # Enable showing branches under other refs in addition to heads
 561
 562         # To set system wide extra branch refs have in $GITWEB_CONFIG
 563         # $feature{'extra-branch-refs'}{'default'} = ['dirs', 'of', 'choice'];
 564         # To have project specific config enable override in $GITWEB_CONFIG
 565         # $feature{'extra-branch-refs'}{'override'} = 1;
 566         # and in project config gitweb.extrabranchrefs = dirs of choice
 567         # Every directory is separated with whitespace.
 568
 569         'extra-branch-refs' => {
 570                 'sub' => \&feature_extra_branch_refs,
 571                 'override' => 0,
 572                 'default' => []},
 573
 574         # Redact e-mail addresses.
 575
 576         # To enable system wide have in $GITWEB_CONFIG
 577         # $feature{'email-privacy'}{'default'} = [1];
 578         'email-privacy' => {
 579                 'sub' => sub { feature_bool('email-privacy', @_) },
 580                 'override' => 1,
 581                 'default' => [0]},
 582 );
 583
 584 sub gitweb_get_feature {
 585         my ($name) = @_;
 586         return unless exists $feature{$name};
 587         my ($sub, $override, @defaults) = (
 588                 $feature{$name}{'sub'},
 589                 $feature{$name}{'override'},
 590                 @{$feature{$name}{'default'}});
 591         # project specific override is possible only if we have project
 592         our $git_dir; # global variable, declared later
 593         if (!$override || !defined $git_dir) {
 594                 return @defaults;
 595         }
 596         if (!defined $sub) {
 597                 warn "feature $name is not overridable";
 598                 return @defaults;
 599         }
 600         return $sub->(@defaults);
 601 }
 602
 603 # A wrapper to check if a given feature is enabled.
 604 # With this, you can say
 605 #
 606 #   my $bool_feat = gitweb_check_feature('bool_feat');
 607 #   gitweb_check_feature('bool_feat') or somecode;
 608 #
 609 # instead of
 610 #
 611 #   my ($bool_feat) = gitweb_get_feature('bool_feat');
 612 #   (gitweb_get_feature('bool_feat'))[0] or somecode;
 613 #
 614 sub gitweb_check_feature {
 615         return (gitweb_get_feature(@_))[0];
 616 }
 617
 618
 619 sub feature_bool {
 620         my $key = shift;
 621         my ($val) = git_get_project_config($key, '--bool');
 622
 623         if (!defined $val) {
 624                 return ($_[0]);
 625         } elsif ($val eq 'true') {
 626                 return (1);
 627         } elsif ($val eq 'false') {
 628                 return (0);
 629         }
 630 }
 631
 632 sub feature_snapshot {
 633         my (@fmts) = @_;
 634
 635         my ($val) = git_get_project_config('snapshot');
 636
 637         if ($val) {
 638                 @fmts = ($val eq 'none' ? () : split /\s*[,\s]\s*/, $val);
 639         }
 640
 641         return @fmts;
 642 }
 643
 644 sub feature_patches {
 645         my @val = (git_get_project_config('patches', '--int'));
 646
 647         if (@val) {
 648                 return @val;
 649         }
 650
 651         return ($_[0]);
 652 }
 653
 654 sub feature_avatar {
 655         my @val = (git_get_project_config('avatar'));
 656
 657         return @val ? @val : @_;
 658 }
 659
 660 sub feature_extra_branch_refs {
 661         my (@branch_refs) = @_;
 662         my $values = git_get_project_config('extrabranchrefs');
 663
 664         if ($values) {
 665                 $values = config_to_multi ($values);
 666                 @branch_refs = ();
 667                 foreach my $value (@{$values}) {
 668                         push @branch_refs, split /\s+/, $value;
 669                 }
 670         }
 671
 672         return @branch_refs;
 673 }
 674
 675 # checking HEAD file with -e is fragile if the repository was
 676 # initialized long time ago (i.e. symlink HEAD) and was pack-ref'ed
 677 # and then pruned.
 678 sub check_head_link {
 679         my ($dir) = @_;
 680         my $headfile = "$dir/HEAD";
 681         return ((-e $headfile) ||
 682                 (-l $headfile && readlink($headfile) =~ /^refs\/heads\//));
 683 }
 684
 685 sub check_export_ok {
 686         my ($dir) = @_;
 687         return (check_head_link($dir) &&
 688                 (!$export_ok || -e "$dir/$export_ok") &&
 689                 (!$export_auth_hook || $export_auth_hook->($dir)));
 690 }
 691
 692 # process alternate names for backward compatibility
 693 # filter out unsupported (unknown) snapshot formats
 694 sub filter_snapshot_fmts {
 695         my @fmts = @_;
 696
 697         @fmts = map {
 698                 exists $known_snapshot_format_aliases{$_} ?
 699                        $known_snapshot_format_aliases{$_} : $_} @fmts;
 700         @fmts = grep {
 701                 exists $known_snapshot_formats{$_} &&
 702                 !$known_snapshot_formats{$_}{'disabled'}} @fmts;
 703 }
 704
 705 sub filter_and_validate_refs {
 706         my @refs = @_;
 707         my %unique_refs = ();
 708
 709         foreach my $ref (@refs) {
 710                 die_error(500, "Invalid ref '$ref' in 'extra-branch-refs' feature") unless (is_valid_ref_format($ref));
 711                 # 'heads' are added implicitly in get_branch_refs().
 712                 $unique_refs{$ref} = 1 if ($ref ne 'heads');
 713         }
 714         return sort keys %unique_refs;
 715 }
 716
 717 # If it is set to code reference, it is code that it is to be run once per
 718 # request, allowing updating configurations that change with each request,
 719 # while running other code in config file only once.
 720 #
 721 # Otherwise, if it is false then gitweb would process config file only once;
 722 # if it is true then gitweb config would be run for each request.
 723 our $per_request_config = 1;
 724
 725 # read and parse gitweb config file given by its parameter.
 726 # returns true on success, false on recoverable error, allowing
 727 # to chain this subroutine, using first file that exists.
 728 # dies on errors during parsing config file, as it is unrecoverable.
 729 sub read_config_file {
 730         my $filename = shift;
 731         return unless defined $filename;
 732         # die if there are errors parsing config file
 733         if (-e $filename) {
 734                 do $filename;
 735                 die $@ if $@;
 736                 return 1;
 737         }
 738         return;
 739 }
 740
 741 our ($GITWEB_CONFIG, $GITWEB_CONFIG_SYSTEM, $GITWEB_CONFIG_COMMON);
 742 sub evaluate_gitweb_config {
 743         our $GITWEB_CONFIG = $ENV{'GITWEB_CONFIG'} || "++GITWEB_CONFIG++";
 744         our $GITWEB_CONFIG_SYSTEM = $ENV{'GITWEB_CONFIG_SYSTEM'} || "++GITWEB_CONFIG_SYSTEM++";
 745         our $GITWEB_CONFIG_COMMON = $ENV{'GITWEB_CONFIG_COMMON'} || "++GITWEB_CONFIG_COMMON++";
 746
 747         # Protect against duplications of file names, to not read config twice.
 748         # Only one of $GITWEB_CONFIG and $GITWEB_CONFIG_SYSTEM is used, so
 749         # there possibility of duplication of filename there doesn't matter.
 750         $GITWEB_CONFIG = ""        if ($GITWEB_CONFIG eq $GITWEB_CONFIG_COMMON);
 751         $GITWEB_CONFIG_SYSTEM = "" if ($GITWEB_CONFIG_SYSTEM eq $GITWEB_CONFIG_COMMON);
 752
 753         # Common system-wide settings for convenience.
 754         # Those settings can be overridden by GITWEB_CONFIG or GITWEB_CONFIG_SYSTEM.
 755         read_config_file($GITWEB_CONFIG_COMMON);
 756
 757         # Use first config file that exists.  This means use the per-instance
 758         # GITWEB_CONFIG if exists, otherwise use GITWEB_SYSTEM_CONFIG.
 759         read_config_file($GITWEB_CONFIG) and return;
 760         read_config_file($GITWEB_CONFIG_SYSTEM);
 761 }
 762
 763 # Get loadavg of system, to compare against $maxload.
 764 # Currently it requires '/proc/loadavg' present to get loadavg;
 765 # if it is not present it returns 0, which means no load checking.
 766 sub get_loadavg {
 767         if( -e '/proc/loadavg' ){
 768                 open my $fd, '<', '/proc/loadavg'
 769                         or return 0;
 770                 my @load = split(/\s+/, scalar <$fd>);
 771                 close $fd;
 772
 773                 # The first three columns measure CPU and IO utilization of the last one,
 774                 # five, and 10 minute periods.  The fourth column shows the number of
 775                 # currently running processes and the total number of processes in the m/n
 776                 # format.  The last column displays the last process ID used.
 777                 return $load[0] || 0;
 778         }
 779         # additional checks for load average should go here for things that don't export
 780         # /proc/loadavg
 781
 782         return 0;
 783 }
 784
 785 # version of the core git binary
 786 our $git_version;
 787 sub evaluate_git_version {
 788         our $git_version = qx("$GIT" --version) =~ m/git version (.*)$/ ? $1 : "unknown";
 789         $number_of_git_cmds++;
 790 }
 791
 792 sub check_loadavg {
 793         if (defined $maxload && get_loadavg() > $maxload) {
 794                 die_error(503, "The load average on the server is too high");
 795         }
 796 }
 797
 798 # ======================================================================
 799 # input validation and dispatch
 800
 801 # Various hash size-related values.
 802 my $sha1_len = 40;
 803 my $sha256_extra_len = 24;
 804 my $sha256_len = $sha1_len + $sha256_extra_len;
 805
 806 # A regex matching $len hex characters. $len may be a range (e.g. 7,64).
 807 sub oid_nlen_regex {
 808         my $len = shift;
 809         my $hchr = qr/[0-9a-fA-F]/;
 810         return qr/(?:(?:$hchr){$len})/;
 811 }
 812
 813 # A regex matching two sets of $nlen hex characters, prefixed by the literal
 814 # string $prefix and with the literal string $infix between them.
 815 sub oid_nlen_prefix_infix_regex {
 816         my $nlen = shift;
 817         my $prefix = shift;
 818         my $infix = shift;
 819
 820         my $rx = oid_nlen_regex($nlen);
 821
 822         return qr/^\Q$prefix\E$rx\Q$infix\E$rx$/;
 823 }
 824
 825 # A regex matching a valid object ID.
 826 our $oid_regex;
 827 {
 828         my $x = oid_nlen_regex($sha1_len);
 829         my $y = oid_nlen_regex($sha256_extra_len);
 830         $oid_regex = qr/(?:$x(?:$y)?)/;
 831 }
 832
 833 # input parameters can be collected from a variety of sources (presently, CGI
 834 # and PATH_INFO), so we define an %input_params hash that collects them all
 835 # together during validation: this allows subsequent uses (e.g. href()) to be
 836 # agnostic of the parameter origin
 837
 838 our %input_params = ();
 839
 840 # input parameters are stored with the long parameter name as key. This will
 841 # also be used in the href subroutine to convert parameters to their CGI
 842 # equivalent, and since the href() usage is the most frequent one, we store
 843 # the name -> CGI key mapping here, instead of the reverse.
 844 #
 845 # XXX: Warning: If you touch this, check the search form for updating,
 846 # too.
 847
 848 our @cgi_param_mapping = (
 849         project => "p",
 850         action => "a",
 851         file_name => "f",
 852         file_parent => "fp",
 853         hash => "h",
 854         hash_parent => "hp",
 855         hash_base => "hb",
 856         hash_parent_base => "hpb",
 857         page => "pg",
 858         order => "o",
 859         searchtext => "s",
 860         searchtype => "st",
 861         snapshot_format => "sf",
 862         extra_options => "opt",
 863         search_use_regexp => "sr",
 864         ctag => "by_tag",
 865         diff_style => "ds",
 866         project_filter => "pf",
 867         # this must be last entry (for manipulation from JavaScript)
 868         javascript => "js"
 869 );
 870 our %cgi_param_mapping = @cgi_param_mapping;
 871
 872 # we will also need to know the possible actions, for validation
 873 our %actions = (
 874         "blame" => \&git_blame,
 875         "blame_incremental" => \&git_blame_incremental,
 876         "blame_data" => \&git_blame_data,
 877         "blobdiff" => \&git_blobdiff,
 878         "blobdiff_plain" => \&git_blobdiff_plain,
 879         "blob" => \&git_blob,
 880         "blob_plain" => \&git_blob_plain,
 881         "commitdiff" => \&git_commitdiff,
 882         "commitdiff_plain" => \&git_commitdiff_plain,
 883         "commit" => \&git_commit,
 884         "forks" => \&git_forks,
 885         "heads" => \&git_heads,
 886         "history" => \&git_history,
 887         "log" => \&git_log,
 888         "patch" => \&git_patch,
 889         "patches" => \&git_patches,
 890         "remotes" => \&git_remotes,
 891         "rss" => \&git_rss,
 892         "atom" => \&git_atom,
 893         "search" => \&git_search,
 894         "search_help" => \&git_search_help,
 895         "shortlog" => \&git_shortlog,
 896         "summary" => \&git_summary,
 897         "tag" => \&git_tag,
 898         "tags" => \&git_tags,
 899         "tree" => \&git_tree,
 900         "snapshot" => \&git_snapshot,
 901         "object" => \&git_object,
 902         # those below don't need $project
 903         "opml" => \&git_opml,
 904         "project_list" => \&git_project_list,
 905         "project_index" => \&git_project_index,
 906 );
 907
 908 # finally, we have the hash of allowed extra_options for the commands that
 909 # allow them
 910 our %allowed_options = (
 911         "--no-merges" => [ qw(rss atom log shortlog history) ],
 912 );
 913
 914 # fill %input_params with the CGI parameters. All values except for 'opt'
 915 # should be single values, but opt can be an array. We should probably
 916 # build an array of parameters that can be multi-valued, but since for the time
 917 # being it's only this one, we just single it out
 918 sub evaluate_query_params {
 919         our $cgi;
 920
 921         while (my ($name, $symbol) = each %cgi_param_mapping) {
 922                 if ($symbol eq 'opt') {
 923                         $input_params{$name} = [ map { decode_utf8($_) } $cgi->multi_param($symbol) ];
 924                 } else {
 925                         $input_params{$name} = decode_utf8($cgi->param($symbol));
 926                 }
 927         }
 928 }
 929
 930 # now read PATH_INFO and update the parameter list for missing parameters
 931 sub evaluate_path_info {
 932         return if defined $input_params{'project'};
 933         return if !$path_info;
 934         $path_info =~ s,^/+,,;
 935         return if !$path_info;
 936
 937         # find which part of PATH_INFO is project
 938         my $project = $path_info;
 939         $project =~ s,/+$,,;
 940         while ($project && !check_head_link("$projectroot/$project")) {
 941                 $project =~ s,/*[^/]*$,,;
 942         }
 943         return unless $project;
 944         $input_params{'project'} = $project;
 945
 946         # do not change any parameters if an action is given using the query string
 947         return if $input_params{'action'};
 948         $path_info =~ s,^\Q$project\E/*,,;
 949
 950         # next, check if we have an action
 951         my $action = $path_info;
 952         $action =~ s,/.*$,,;
 953         if (exists $actions{$action}) {
 954                 $path_info =~ s,^$action/*,,;
 955                 $input_params{'action'} = $action;
 956         }
 957
 958         # list of actions that want hash_base instead of hash, but can have no
 959         # pathname (f) parameter
 960         my @wants_base = (
 961                 'tree',
 962                 'history',
 963         );
 964
 965         # we want to catch, among others
 966         # [$hash_parent_base[:$file_parent]..]$hash_parent[:$file_name]
 967         my ($parentrefname, $parentpathname, $refname, $pathname) =
 968                 ($path_info =~ /^(?:(.+?)(?::(.+))?\.\.)?([^:]+?)?(?::(.+))?$/);
 969
 970         # first, analyze the 'current' part
 971         if (defined $pathname) {
 972                 # we got "branch:filename" or "branch:dir/"
 973                 # we could use git_get_type(branch:pathname), but:
 974                 # - it needs $git_dir
 975                 # - it does a git() call
 976                 # - the convention of terminating directories with a slash
 977                 #   makes it superfluous
 978                 # - embedding the action in the PATH_INFO would make it even
 979                 #   more superfluous
 980                 $pathname =~ s,^/+,,;
 981                 if (!$pathname || substr($pathname, -1) eq "/") {
 982                         $input_params{'action'} ||= "tree";
 983                         $pathname =~ s,/$,,;
 984                 } else {
 985                         # the default action depends on whether we had parent info
 986                         # or not
 987                         if ($parentrefname) {
 988                                 $input_params{'action'} ||= "blobdiff_plain";
 989                         } else {
 990                                 $input_params{'action'} ||= "blob_plain";
 991                         }
 992                 }
 993                 $input_params{'hash_base'} ||= $refname;
 994                 $input_params{'file_name'} ||= $pathname;
 995         } elsif (defined $refname) {
 996                 # we got "branch". In this case we have to choose if we have to
 997                 # set hash or hash_base.
 998                 #
 999                 # Most of the actions without a pathname only want hash to be
1000                 # set, except for the ones specified in @wants_base that want
1001                 # hash_base instead. It should also be noted that hand-crafted
1002                 # links having 'history' as an action and no pathname or hash
1003                 # set will fail, but that happens regardless of PATH_INFO.
1004                 if (defined $parentrefname) {
1005                         # if there is parent let the default be 'shortlog' action
1006                         # (for http://git.example.com/repo.git/A..B links); if there
1007                         # is no parent, dispatch will detect type of object and set
1008                         # action appropriately if required (if action is not set)
1009                         $input_params{'action'} ||= "shortlog";
1010                 }
1011                 if ($input_params{'action'} &&
1012                     grep { $_ eq $input_params{'action'} } @wants_base) {
1013                         $input_params{'hash_base'} ||= $refname;
1014                 } else {
1015                         $input_params{'hash'} ||= $refname;
1016                 }
1017         }
1018
1019         # next, handle the 'parent' part, if present
1020         if (defined $parentrefname) {
1021                 # a missing pathspec defaults to the 'current' filename, allowing e.g.
1022                 # someproject/blobdiff/oldrev..newrev:/filename
1023                 if ($parentpathname) {
1024                         $parentpathname =~ s,^/+,,;
1025                         $parentpathname =~ s,/$,,;
1026                         $input_params{'file_parent'} ||= $parentpathname;
1027                 } else {
1028                         $input_params{'file_parent'} ||= $input_params{'file_name'};
1029                 }
1030                 # we assume that hash_parent_base is wanted if a path was specified,
1031                 # or if the action wants hash_base instead of hash
1032                 if (defined $input_params{'file_parent'} ||
1033                         grep { $_ eq $input_params{'action'} } @wants_base) {
1034                         $input_params{'hash_parent_base'} ||= $parentrefname;
1035                 } else {
1036                         $input_params{'hash_parent'} ||= $parentrefname;
1037                 }
1038         }
1039
1040         # for the snapshot action, we allow URLs in the form
1041         # $project/snapshot/$hash.ext
1042         # where .ext determines the snapshot and gets removed from the
1043         # passed $refname to provide the $hash.
1044         #
1045         # To be able to tell that $refname includes the format extension, we
1046         # require the following two conditions to be satisfied:
1047         # - the hash input parameter MUST have been set from the $refname part
1048         #   of the URL (i.e. they must be equal)
1049         # - the snapshot format MUST NOT have been defined already (e.g. from
1050         #   CGI parameter sf)
1051         # It's also useless to try any matching unless $refname has a dot,
1052         # so we check for that too
1053         if (defined $input_params{'action'} &&
1054                 $input_params{'action'} eq 'snapshot' &&
1055                 defined $refname && index($refname, '.') != -1 &&
1056                 $refname eq $input_params{'hash'} &&
1057                 !defined $input_params{'snapshot_format'}) {
1058                 # We loop over the known snapshot formats, checking for
1059                 # extensions. Allowed extensions are both the defined suffix
1060                 # (which includes the initial dot already) and the snapshot
1061                 # format key itself, with a prepended dot
1062                 while (my ($fmt, $opt) = each %known_snapshot_formats) {
1063                         my $hash = $refname;
1064                         unless ($hash =~ s/(\Q$opt->{'suffix'}\E|\Q.$fmt\E)$//) {
1065                                 next;
1066                         }
1067                         my $sfx = $1;
1068                         # a valid suffix was found, so set the snapshot format
1069                         # and reset the hash parameter
1070                         $input_params{'snapshot_format'} = $fmt;
1071                         $input_params{'hash'} = $hash;
1072                         # we also set the format suffix to the one requested
1073                         # in the URL: this way a request for e.g. .tgz returns
1074                         # a .tgz instead of a .tar.gz
1075                         $known_snapshot_formats{$fmt}{'suffix'} = $sfx;
1076                         last;
1077                 }
1078         }
1079 }
1080
1081 our ($action, $project, $file_name, $file_parent, $hash, $hash_parent, $hash_base,
1082      $hash_parent_base, @extra_options, $page, $searchtype, $search_use_regexp,
1083      $searchtext, $search_regexp, $project_filter);
1084 sub evaluate_and_validate_params {
1085         our $action = $input_params{'action'};
1086         if (defined $action) {
1087                 if (!is_valid_action($action)) {
1088                         die_error(400, "Invalid action parameter");
1089                 }
1090         }
1091
1092         # parameters which are pathnames
1093         our $project = $input_params{'project'};
1094         if (defined $project) {
1095                 if (!is_valid_project($project)) {
1096                         undef $project;
1097                         die_error(404, "No such project");
1098                 }
1099         }
1100
1101         our $project_filter = $input_params{'project_filter'};
1102         if (defined $project_filter) {
1103                 if (!is_valid_pathname($project_filter)) {
1104                         die_error(404, "Invalid project_filter parameter");
1105                 }
1106         }
1107
1108         our $file_name = $input_params{'file_name'};
1109         if (defined $file_name) {
1110                 if (!is_valid_pathname($file_name)) {
1111                         die_error(400, "Invalid file parameter");
1112                 }
1113         }
1114
1115         our $file_parent = $input_params{'file_parent'};
1116         if (defined $file_parent) {
1117                 if (!is_valid_pathname($file_parent)) {
1118                         die_error(400, "Invalid file parent parameter");
1119                 }
1120         }
1121
1122         # parameters which are refnames
1123         our $hash = $input_params{'hash'};
1124         if (defined $hash) {
1125                 if (!is_valid_refname($hash)) {
1126                         die_error(400, "Invalid hash parameter");
1127                 }
1128         }
1129
1130         our $hash_parent = $input_params{'hash_parent'};
1131         if (defined $hash_parent) {
1132                 if (!is_valid_refname($hash_parent)) {
1133                         die_error(400, "Invalid hash parent parameter");
1134                 }
1135         }
1136
1137         our $hash_base = $input_params{'hash_base'};
1138         if (defined $hash_base) {
1139                 if (!is_valid_refname($hash_base)) {
1140                         die_error(400, "Invalid hash base parameter");
1141                 }
1142         }
1143
1144         our @extra_options = @{$input_params{'extra_options'}};
1145         # @extra_options is always defined, since it can only be (currently) set from
1146         # CGI, and $cgi->param() returns the empty array in array context if the param
1147         # is not set
1148         foreach my $opt (@extra_options) {
1149                 if (not exists $allowed_options{$opt}) {
1150                         die_error(400, "Invalid option parameter");
1151                 }
1152                 if (not grep(/^$action$/, @{$allowed_options{$opt}})) {
1153                         die_error(400, "Invalid option parameter for this action");
1154                 }
1155         }
1156
1157         our $hash_parent_base = $input_params{'hash_parent_base'};
1158         if (defined $hash_parent_base) {
1159                 if (!is_valid_refname($hash_parent_base)) {
1160                         die_error(400, "Invalid hash parent base parameter");
1161                 }
1162         }
1163
1164         # other parameters
1165         our $page = $input_params{'page'};
1166         if (defined $page) {
1167                 if ($page =~ m/[^0-9]/) {
1168                         die_error(400, "Invalid page parameter");
1169                 }
1170         }
1171
1172         our $searchtype = $input_params{'searchtype'};
1173         if (defined $searchtype) {
1174                 if ($searchtype =~ m/[^a-z]/) {
1175                         die_error(400, "Invalid searchtype parameter");
1176                 }
1177         }
1178
1179         our $search_use_regexp = $input_params{'search_use_regexp'};
1180
1181         our $searchtext = $input_params{'searchtext'};
1182         our $search_regexp = undef;
1183         if (defined $searchtext) {
1184                 if (length($searchtext) < 2) {
1185                         die_error(403, "At least two characters are required for search parameter");
1186                 }
1187                 if ($search_use_regexp) {
1188                         $search_regexp = $searchtext;
1189                         if (!eval { qr/$search_regexp/; 1; }) {
1190                                 (my $error = $@) =~ s/ at \S+ line \d+.*\n?//;
1191                                 die_error(400, "Invalid search regexp '$search_regexp'",
1192                                           esc_html($error));
1193                         }
1194                 } else {
1195                         $search_regexp = quotemeta $searchtext;
1196                 }
1197         }
1198 }
1199
1200 # path to the current git repository
1201 our $git_dir;
1202 sub evaluate_git_dir {
1203         our $git_dir = "$projectroot/$project" if $project;
1204 }
1205
1206 our (@snapshot_fmts, $git_avatar, @extra_branch_refs);
1207 sub configure_gitweb_features {
1208         # list of supported snapshot formats
1209         our @snapshot_fmts = gitweb_get_feature('snapshot');
1210         @snapshot_fmts = filter_snapshot_fmts(@snapshot_fmts);
1211
1212         our ($git_avatar) = gitweb_get_feature('avatar');
1213         $git_avatar = '' unless $git_avatar =~ /^(?:gravatar|picon)$/s;
1214
1215         our @extra_branch_refs = gitweb_get_feature('extra-branch-refs');
1216         @extra_branch_refs = filter_and_validate_refs (@extra_branch_refs);
1217 }
1218
1219 sub get_branch_refs {
1220         return ('heads', @extra_branch_refs);
1221 }
1222
1223 # custom error handler: 'die <message>' is Internal Server Error
1224 sub handle_errors_html {
1225         my $msg = shift; # it is already HTML escaped
1226
1227         # to avoid infinite loop where error occurs in die_error,
1228         # change handler to default handler, disabling handle_errors_html
1229         set_message("Error occurred when inside die_error:\n$msg");
1230
1231         # you cannot jump out of die_error when called as error handler;
1232         # the subroutine set via CGI::Carp::set_message is called _after_
1233         # HTTP headers are already written, so it cannot write them itself
1234         die_error(undef, undef, $msg, -error_handler => 1, -no_http_header => 1);
1235 }
1236 set_message(\&handle_errors_html);
1237
1238 # dispatch
1239 sub dispatch {
1240         if (!defined $action) {
1241                 if (defined $hash) {
1242                         $action = git_get_type($hash);
1243                         $action or die_error(404, "Object does not exist");
1244                 } elsif (defined $hash_base && defined $file_name) {
1245                         $action = git_get_type("$hash_base:$file_name");
1246                         $action or die_error(404, "File or directory does not exist");
1247                 } elsif (defined $project) {
1248                         $action = 'summary';
1249                 } else {
1250                         $action = 'project_list';
1251                 }
1252         }
1253         if (!defined($actions{$action})) {
1254                 die_error(400, "Unknown action");
1255         }
1256         if ($action !~ m/^(?:opml|project_list|project_index)$/ &&
1257             !$project) {
1258                 die_error(400, "Project needed");
1259         }
1260         $actions{$action}->();
1261 }
1262
1263 sub reset_timer {
1264         our $t0 = [ gettimeofday() ]
1265                 if defined $t0;
1266         our $number_of_git_cmds = 0;
1267 }
1268
1269 our $first_request = 1;
1270 sub run_request {
1271         reset_timer();
1272
1273         evaluate_uri();
1274         if ($first_request) {
1275                 evaluate_gitweb_config();
1276                 evaluate_git_version();
1277         }
1278         if ($per_request_config) {
1279                 if (ref($per_request_config) eq 'CODE') {
1280                         $per_request_config->();
1281                 } elsif (!$first_request) {
1282                         evaluate_gitweb_config();
1283                 }
1284         }
1285         check_loadavg();
1286
1287         # $projectroot and $projects_list might be set in gitweb config file
1288         $projects_list ||= $projectroot;
1289
1290         evaluate_query_params();
1291         evaluate_path_info();
1292         evaluate_and_validate_params();
1293         evaluate_git_dir();
1294
1295         configure_gitweb_features();
1296
1297         dispatch();
1298 }
1299
1300 our $is_last_request = sub { 1 };
1301 our ($pre_dispatch_hook, $post_dispatch_hook, $pre_listen_hook);
1302 our $CGI = 'CGI';
1303 our $cgi;
1304 our $FCGI_Stream_PRINT_raw = \&FCGI::Stream::PRINT;
1305 sub configure_as_fcgi {
1306         require CGI::Fast;
1307         our $CGI = 'CGI::Fast';
1308         # FCGI is not Unicode aware hence the UTF-8 encoding must be done manually.
1309         # However no encoding must be done within git_blob_plain() and git_snapshot()
1310         # which must still output in raw binary mode.
1311         no warnings 'redefine';
1312         my $enc = Encode::find_encoding('UTF-8');
1313         *FCGI::Stream::PRINT = sub {
1314                 my @OUTPUT = @_;
1315                 for (my $i = 1; $i < @_; $i++) {
1316                         $OUTPUT[$i] = $enc->encode($_[$i], Encode::FB_CROAK|Encode::LEAVE_SRC);
1317                 }
1318                 @_ = @OUTPUT;
1319                 goto $FCGI_Stream_PRINT_raw;
1320         };
1321
1322         my $request_number = 0;
1323         # let each child service 100 requests
1324         our $is_last_request = sub { ++$request_number > 100 };
1325 }
1326 sub evaluate_argv {
1327         my $script_name = $ENV{'SCRIPT_NAME'} || $ENV{'SCRIPT_FILENAME'} || __FILE__;
1328         configure_as_fcgi()
1329                 if $script_name =~ /\.fcgi$/;
1330
1331         return unless (@ARGV);
1332
1333         require Getopt::Long;
1334         Getopt::Long::GetOptions(
1335                 'fastcgi|fcgi|f' => \&configure_as_fcgi,
1336                 'nproc|n=i' => sub {
1337                         my ($arg, $val) = @_;
1338                         return unless eval { require FCGI::ProcManager; 1; };
1339                         my $proc_manager = FCGI::ProcManager->new({
1340                                 n_processes => $val,
1341                         });
1342                         our $pre_listen_hook    = sub { $proc_manager->pm_manage()        };
1343                         our $pre_dispatch_hook  = sub { $proc_manager->pm_pre_dispatch()  };
1344                         our $post_dispatch_hook = sub { $proc_manager->pm_post_dispatch() };
1345                 },
1346         );
1347 }
1348
1349 sub run {
1350         evaluate_argv();
1351
1352         $first_request = 1;
1353         $pre_listen_hook->()
1354                 if $pre_listen_hook;
1355
1356  REQUEST:
1357         while ($cgi = $CGI->new()) {
1358                 $pre_dispatch_hook->()
1359                         if $pre_dispatch_hook;
1360
1361                 run_request();
1362
1363                 $post_dispatch_hook->()
1364                         if $post_dispatch_hook;
1365                 $first_request = 0;
1366
1367                 last REQUEST if ($is_last_request->());
1368         }
1369
1370  DONE_GITWEB:
1371         1;
1372 }
1373
1374 run();
1375
1376 if (defined caller) {
1377         # wrapped in a subroutine processing requests,
1378         # e.g. mod_perl with ModPerl::Registry, or PSGI with Plack::App::WrapCGI
1379         return;
1380 } else {
1381         # pure CGI script, serving single request
1382         exit;
1383 }
1384
1385 ## ======================================================================
1386 ## action links
1387
1388 # possible values of extra options
1389 # -full => 0|1      - use absolute/full URL ($my_uri/$my_url as base)
1390 # -replay => 1      - start from a current view (replay with modifications)
1391 # -path_info => 0|1 - don't use/use path_info URL (if possible)
1392 # -anchor => ANCHOR - add #ANCHOR to end of URL, implies -replay if used alone
1393 sub href {
1394         my %params = @_;
1395         # default is to use -absolute url() i.e. $my_uri
1396         my $href = $params{-full} ? $my_url : $my_uri;
1397
1398         # implicit -replay, must be first of implicit params
1399         $params{-replay} = 1 if (keys %params == 1 && $params{-anchor});
1400
1401         $params{'project'} = $project unless exists $params{'project'};
1402
1403         if ($params{-replay}) {
1404                 while (my ($name, $symbol) = each %cgi_param_mapping) {
1405                         if (!exists $params{$name}) {
1406                                 $params{$name} = $input_params{$name};
1407                         }
1408                 }
1409         }
1410
1411         my $use_pathinfo = gitweb_check_feature('pathinfo');
1412         if (defined $params{'project'} &&
1413             (exists $params{-path_info} ? $params{-path_info} : $use_pathinfo)) {
1414                 # try to put as many parameters as possible in PATH_INFO:
1415                 #   - project name
1416                 #   - action
1417                 #   - hash_parent or hash_parent_base:/file_parent
1418                 #   - hash or hash_base:/filename
1419                 #   - the snapshot_format as an appropriate suffix
1420
1421                 # When the script is the root DirectoryIndex for the domain,
1422                 # $href here would be something like http://gitweb.example.com/
1423                 # Thus, we strip any trailing / from $href, to spare us double
1424                 # slashes in the final URL
1425                 $href =~ s,/$,,;
1426
1427                 # Then add the project name, if present
1428                 $href .= "/".esc_path_info($params{'project'});
1429                 delete $params{'project'};
1430
1431                 # since we destructively absorb parameters, we keep this
1432                 # boolean that remembers if we're handling a snapshot
1433                 my $is_snapshot = $params{'action'} eq 'snapshot';
1434
1435                 # Summary just uses the project path URL, any other action is
1436                 # added to the URL
1437                 if (defined $params{'action'}) {
1438                         $href .= "/".esc_path_info($params{'action'})
1439                                 unless $params{'action'} eq 'summary';
1440                         delete $params{'action'};
1441                 }
1442
1443                 # Next, we put hash_parent_base:/file_parent..hash_base:/file_name,
1444                 # stripping nonexistent or useless pieces
1445                 $href .= "/" if ($params{'hash_base'} || $params{'hash_parent_base'}
1446                         || $params{'hash_parent'} || $params{'hash'});
1447                 if (defined $params{'hash_base'}) {
1448                         if (defined $params{'hash_parent_base'}) {
1449                                 $href .= esc_path_info($params{'hash_parent_base'});
1450                                 # skip the file_parent if it's the same as the file_name
1451                                 if (defined $params{'file_parent'}) {
1452                                         if (defined $params{'file_name'} && $params{'file_parent'} eq $params{'file_name'}) {
1453                                                 delete $params{'file_parent'};
1454                                         } elsif ($params{'file_parent'} !~ /\.\./) {
1455                                                 $href .= ":/".esc_path_info($params{'file_parent'});
1456                                                 delete $params{'file_parent'};
1457                                         }
1458                                 }
1459                                 $href .= "..";
1460                                 delete $params{'hash_parent'};
1461                                 delete $params{'hash_parent_base'};
1462                         } elsif (defined $params{'hash_parent'}) {
1463                                 $href .= esc_path_info($params{'hash_parent'}). "..";
1464                                 delete $params{'hash_parent'};
1465                         }
1466
1467                         $href .= esc_path_info($params{'hash_base'});
1468                         if (defined $params{'file_name'} && $params{'file_name'} !~ /\.\./) {
1469                                 $href .= ":/".esc_path_info($params{'file_name'});
1470                                 delete $params{'file_name'};
1471                         }
1472                         delete $params{'hash'};
1473                         delete $params{'hash_base'};
1474                 } elsif (defined $params{'hash'}) {
1475                         $href .= esc_path_info($params{'hash'});
1476                         delete $params{'hash'};
1477                 }
1478
1479                 # If the action was a snapshot, we can absorb the
1480                 # snapshot_format parameter too
1481                 if ($is_snapshot) {
1482                         my $fmt = $params{'snapshot_format'};
1483                         # snapshot_format should always be defined when href()
1484                         # is called, but just in case some code forgets, we
1485                         # fall back to the default
1486                         $fmt ||= $snapshot_fmts[0];
1487                         $href .= $known_snapshot_formats{$fmt}{'suffix'};
1488                         delete $params{'snapshot_format'};
1489                 }
1490         }
1491
1492         # now encode the parameters explicitly
1493         my @result = ();
1494         for (my $i = 0; $i < @cgi_param_mapping; $i += 2) {
1495                 my ($name, $symbol) = ($cgi_param_mapping[$i], $cgi_param_mapping[$i+1]);
1496                 if (defined $params{$name}) {
1497                         if (ref($params{$name}) eq "ARRAY") {
1498                                 foreach my $par (@{$params{$name}}) {
1499                                         push @result, $symbol . "=" . esc_param($par);
1500                                 }
1501                         } else {
1502                                 push @result, $symbol . "=" . esc_param($params{$name});
1503                         }
1504                 }
1505         }
1506         $href .= "?" . join(';', @result) if scalar @result;
1507
1508         # final transformation: trailing spaces must be escaped (URI-encoded)
1509         $href =~ s/(\s+)$/CGI::escape($1)/e;
1510
1511         if ($params{-anchor}) {
1512                 $href .= "#".esc_param($params{-anchor});
1513         }
1514
1515         return $href;
1516 }
1517
1518
1519 ## ======================================================================
1520 ## validation, quoting/unquoting and escaping
1521
1522 sub is_valid_action {
1523         my $input = shift;
1524         return undef unless exists $actions{$input};
1525         return 1;
1526 }
1527
1528 sub is_valid_project {
1529         my $input = shift;
1530
1531         return unless defined $input;
1532         if (!is_valid_pathname($input) ||
1533                 !(-d "$projectroot/$input") ||
1534                 !check_export_ok("$projectroot/$input") ||
1535                 ($strict_export && !project_in_list($input))) {
1536                 return undef;
1537         } else {
1538                 return 1;
1539         }
1540 }
1541
1542 sub is_valid_pathname {
1543         my $input = shift;
1544
1545         return undef unless defined $input;
1546         # no '.' or '..' as elements of path, i.e. no '.' or '..'
1547         # at the beginning, at the end, and between slashes.
1548         # also this catches doubled slashes
1549         if ($input =~ m!(^|/)(|\.|\.\.)(/|$)!) {
1550                 return undef;
1551         }
1552         # no null characters
1553         if ($input =~ m!\0!) {
1554                 return undef;
1555         }
1556         return 1;
1557 }
1558
1559 sub is_valid_ref_format {
1560         my $input = shift;
1561
1562         return undef unless defined $input;
1563         # restrictions on ref name according to git-check-ref-format
1564         if ($input =~ m!(/\.|\.\.|[\000-\040\177 ~^:?*\[]|/$)!) {
1565                 return undef;
1566         }
1567         return 1;
1568 }
1569
1570 sub is_valid_refname {
1571         my $input = shift;
1572
1573         return undef unless defined $input;
1574         # textual hashes are O.K.
1575         if ($input =~ m/^$oid_regex$/) {
1576                 return 1;
1577         }
1578         # it must be correct pathname
1579         is_valid_pathname($input) or return undef;
1580         # check git-check-ref-format restrictions
1581         is_valid_ref_format($input) or return undef;
1582         return 1;
1583 }
1584
1585 # decode sequences of octets in utf8 into Perl's internal form,
1586 # which is utf-8 with utf8 flag set if needed.  gitweb writes out
1587 # in utf-8 thanks to "binmode STDOUT, ':utf8'" at beginning
1588 sub to_utf8 {
1589         my $str = shift;
1590         return undef unless defined $str;
1591
1592         if (utf8::is_utf8($str) || utf8::decode($str)) {
1593                 return $str;
1594         } else {
1595                 return decode($fallback_encoding, $str, Encode::FB_DEFAULT);
1596         }
1597 }
1598
1599 # quote unsafe chars, but keep the slash, even when it's not
1600 # correct, but quoted slashes look too horrible in bookmarks
1601 sub esc_param {
1602         my $str = shift;
1603         return undef unless defined $str;
1604         $str =~ s/([^A-Za-z0-9\-_.~()\/:@ ]+)/CGI::escape($1)/eg;
1605         $str =~ s/ /\+/g;
1606         return $str;
1607 }
1608
1609 # the quoting rules for path_info fragment are slightly different
1610 sub esc_path_info {
1611         my $str = shift;
1612         return undef unless defined $str;
1613
1614         # path_info doesn't treat '+' as space (specially), but '?' must be escaped
1615         $str =~ s/([^A-Za-z0-9\-_.~();\/;:@&= +]+)/CGI::escape($1)/eg;
1616
1617         return $str;
1618 }
1619
1620 # quote unsafe chars in whole URL, so some characters cannot be quoted
1621 sub esc_url {
1622         my $str = shift;
1623         return undef unless defined $str;
1624         $str =~ s/([^A-Za-z0-9\-_.~();\/;?:@&= ]+)/CGI::escape($1)/eg;
1625         $str =~ s/ /\+/g;
1626         return $str;
1627 }
1628
1629 # quote unsafe characters in HTML attributes
1630 sub esc_attr {
1631
1632         # for XHTML conformance escaping '"' to '&quot;' is not enough
1633         return esc_html(@_);
1634 }
1635
1636 # replace invalid utf8 character with SUBSTITUTION sequence
1637 sub esc_html {
1638         my $str = shift;
1639         my %opts = @_;
1640
1641         return undef unless defined $str;
1642
1643         $str = to_utf8($str);
1644         $str = $cgi->escapeHTML($str);
1645         if ($opts{'-nbsp'}) {
1646                 $str =~ s/ /&nbsp;/g;
1647         }
1648         $str =~ s|([[:cntrl:]])|(($1 ne "\t") ? quot_cec($1) : $1)|eg;
1649         return $str;
1650 }
1651
1652 # quote control characters and escape filename to HTML
1653 sub esc_path {
1654         my $str = shift;
1655         my %opts = @_;
1656
1657         return undef unless defined $str;
1658
1659         $str = to_utf8($str);
1660         $str = $cgi->escapeHTML($str);
1661         if ($opts{'-nbsp'}) {
1662                 $str =~ s/ /&nbsp;/g;
1663         }
1664         $str =~ s|([[:cntrl:]])|quot_cec($1)|eg;
1665         return $str;
1666 }
1667
1668 # Sanitize for use in XHTML + application/xml+xhtml (valid XML 1.0)
1669 sub sanitize {
1670         my $str = shift;
1671
1672         return undef unless defined $str;
1673
1674         $str = to_utf8($str);
1675         $str =~ s|([[:cntrl:]])|(index("\t\n\r", $1) != -1 ? $1 : quot_cec($1))|eg;
1676         return $str;
1677 }
1678
1679 # Make control characters "printable", using character escape codes (CEC)
1680 sub quot_cec {
1681         my $cntrl = shift;
1682         my %opts = @_;
1683         my %es = ( # character escape codes, aka escape sequences
1684                 "\t" => '\t',   # tab             (HT)
1685                 "\n" => '\n',   # line feed       (LF)
1686                 "\r" => '\r',   # carriage return (CR)
1687                 "\f" => '\f',   # form feed       (FF)
1688                 "\b" => '\b',   # backspace       (BS)
1689                 "\a" => '\a',   # alarm (bell)    (BEL)
1690                 "\e" => '\e',   # escape          (ESC)
1691                 "\013" => '\v', # vertical tab    (VT)
1692                 "\000" => '\0', # nul character   (NUL)
1693         );
1694         my $chr = ( (exists $es{$cntrl})
1695                     ? $es{$cntrl}
1696                     : sprintf('\%2x', ord($cntrl)) );
1697         if ($opts{-nohtml}) {
1698                 return $chr;
1699         } else {
1700                 return "<span class=\"cntrl\">$chr</span>";
1701         }
1702 }
1703
1704 # Alternatively use unicode control pictures codepoints,
1705 # Unicode "printable representation" (PR)
1706 sub quot_upr {
1707         my $cntrl = shift;
1708         my %opts = @_;
1709
1710         my $chr = sprintf('&#%04d;', 0x2400+ord($cntrl));
1711         if ($opts{-nohtml}) {
1712                 return $chr;
1713         } else {
1714                 return "<span class=\"cntrl\">$chr</span>";
1715         }
1716 }
1717
1718 # git may return quoted and escaped filenames
1719 sub unquote {
1720         my $str = shift;
1721
1722         sub unq {
1723                 my $seq = shift;
1724                 my %es = ( # character escape codes, aka escape sequences
1725                         't' => "\t",   # tab            (HT, TAB)
1726                         'n' => "\n",   # newline        (NL)
1727                         'r' => "\r",   # return         (CR)
1728                         'f' => "\f",   # form feed      (FF)
1729                         'b' => "\b",   # backspace      (BS)
1730                         'a' => "\a",   # alarm (bell)   (BEL)
1731                         'e' => "\e",   # escape         (ESC)
1732                         'v' => "\013", # vertical tab   (VT)
1733                 );
1734
1735                 if ($seq =~ m/^[0-7]{1,3}$/) {
1736                         # octal char sequence
1737                         return chr(oct($seq));
1738                 } elsif (exists $es{$seq}) {
1739                         # C escape sequence, aka character escape code
1740                         return $es{$seq};
1741                 }
1742                 # quoted ordinary character
1743                 return $seq;
1744         }
1745
1746         if ($str =~ m/^"(.*)"$/) {
1747                 # needs unquoting
1748                 $str = $1;
1749                 $str =~ s/\\([^0-7]|[0-7]{1,3})/unq($1)/eg;
1750         }
1751         return $str;
1752 }
1753
1754 # escape tabs (convert tabs to spaces)
1755 sub untabify {
1756         my $line = shift;
1757
1758         while ((my $pos = index($line, "\t")) != -1) {
1759                 if (my $count = (8 - ($pos % 8))) {
1760                         my $spaces = ' ' x $count;
1761                         $line =~ s/\t/$spaces/;
1762                 }
1763         }
1764
1765         return $line;
1766 }
1767
1768 sub project_in_list {
1769         my $project = shift;
1770         my @list = git_get_projects_list();
1771         return @list && scalar(grep { $_->{'path'} eq $project } @list);
1772 }
1773
1774 ## ----------------------------------------------------------------------
1775 ## HTML aware string manipulation
1776
1777 # Try to chop given string on a word boundary between position
1778 # $len and $len+$add_len. If there is no word boundary there,
1779 # chop at $len+$add_len. Do not chop if chopped part plus ellipsis
1780 # (marking chopped part) would be longer than given string.
1781 sub chop_str {
1782         my $str = shift;
1783         my $len = shift;
1784         my $add_len = shift || 10;
1785         my $where = shift || 'right'; # 'left' | 'center' | 'right'
1786
1787         # Make sure perl knows it is utf8 encoded so we don't
1788         # cut in the middle of a utf8 multibyte char.
1789         $str = to_utf8($str);
1790
1791         # allow only $len chars, but don't cut a word if it would fit in $add_len
1792         # if it doesn't fit, cut it if it's still longer than the dots we would add
1793         # remove chopped character entities entirely
1794
1795         # when chopping in the middle, distribute $len into left and right part
1796         # return early if chopping wouldn't make string shorter
1797         if ($where eq 'center') {
1798                 return $str if ($len + 5 >= length($str)); # filler is length 5
1799                 $len = int($len/2);
1800         } else {
1801                 return $str if ($len + 4 >= length($str)); # filler is length 4
1802         }
1803
1804         # regexps: ending and beginning with word part up to $add_len
1805         my $endre = qr/.{$len}\w{0,$add_len}/;
1806         my $begre = qr/\w{0,$add_len}.{$len}/;
1807
1808         if ($where eq 'left') {
1809                 $str =~ m/^(.*?)($begre)$/;
1810                 my ($lead, $body) = ($1, $2);
1811                 if (length($lead) > 4) {
1812                         $lead = " ...";
1813                 }
1814                 return "$lead$body";
1815
1816         } elsif ($where eq 'center') {
1817                 $str =~ m/^($endre)(.*)$/;
1818                 my ($left, $str)  = ($1, $2);
1819                 $str =~ m/^(.*?)($begre)$/;
1820                 my ($mid, $right) = ($1, $2);
1821                 if (length($mid) > 5) {
1822                         $mid = " ... ";
1823                 }
1824                 return "$left$mid$right";
1825
1826         } else {
1827                 $str =~ m/^($endre)(.*)$/;
1828                 my $body = $1;
1829                 my $tail = $2;
1830                 if (length($tail) > 4) {
1831                         $tail = "... ";
1832                 }
1833                 return "$body$tail";
1834         }
1835 }
1836
1837 # takes the same arguments as chop_str, but also wraps a <span> around the
1838 # result with a title attribute if it does get chopped. Additionally, the
1839 # string is HTML-escaped.
1840 sub chop_and_escape_str {
1841         my ($str) = @_;
1842
1843         my $chopped = chop_str(@_);
1844         $str = to_utf8($str);
1845         if ($chopped eq $str) {
1846                 return esc_html($chopped);
1847         } else {
1848                 $str =~ s/[[:cntrl:]]/?/g;
1849                 return $cgi->span({-title=>$str}, esc_html($chopped));
1850         }
1851 }
1852
1853 # Highlight selected fragments of string, using given CSS class,
1854 # and escape HTML.  It is assumed that fragments do not overlap.
1855 # Regions are passed as list of pairs (array references).
1856 #
1857 # Example: esc_html_hl_regions("foobar", "mark", [ 0, 3 ]) returns
1858 # '<span class="mark">foo</span>bar'
1859 sub esc_html_hl_regions {
1860         my ($str, $css_class, @sel) = @_;
1861         my %opts = grep { ref($_) ne 'ARRAY' } @sel;
1862         @sel     = grep { ref($_) eq 'ARRAY' } @sel;
1863         return esc_html($str, %opts) unless @sel;
1864
1865         my $out = '';
1866         my $pos = 0;
1867
1868         for my $s (@sel) {
1869                 my ($begin, $end) = @$s;
1870
1871                 # Don't create empty <span> elements.
1872                 next if $end <= $begin;
1873
1874                 my $escaped = esc_html(substr($str, $begin, $end - $begin),
1875                                        %opts);
1876
1877                 $out .= esc_html(substr($str, $pos, $begin - $pos), %opts)
1878                         if ($begin - $pos > 0);
1879                 $out .= $cgi->span({-class => $css_class}, $escaped);
1880
1881                 $pos = $end;
1882         }
1883         $out .= esc_html(substr($str, $pos), %opts)
1884                 if ($pos < length($str));
1885
1886         return $out;
1887 }
1888
1889 # return positions of beginning and end of each match
1890 sub matchpos_list {
1891         my ($str, $regexp) = @_;
1892         return unless (defined $str && defined $regexp);
1893
1894         my @matches;
1895         while ($str =~ /$regexp/g) {
1896                 push @matches, [$-[0], $+[0]];
1897         }
1898         return @matches;
1899 }
1900
1901 # highlight match (if any), and escape HTML
1902 sub esc_html_match_hl {
1903         my ($str, $regexp) = @_;
1904         return esc_html($str) unless defined $regexp;
1905
1906         my @matches = matchpos_list($str, $regexp);
1907         return esc_html($str) unless @matches;
1908
1909         return esc_html_hl_regions($str, 'match', @matches);
1910 }
1911
1912
1913 # highlight match (if any) of shortened string, and escape HTML
1914 sub esc_html_match_hl_chopped {
1915         my ($str, $chopped, $regexp) = @_;
1916         return esc_html_match_hl($str, $regexp) unless defined $chopped;
1917
1918         my @matches = matchpos_list($str, $regexp);
1919         return esc_html($chopped) unless @matches;
1920
1921         # filter matches so that we mark chopped string
1922         my $tail = "... "; # see chop_str
1923         unless ($chopped =~ s/\Q$tail\E$//) {
1924                 $tail = '';
1925         }
1926         my $chop_len = length($chopped);
1927         my $tail_len = length($tail);
1928         my @filtered;
1929
1930         for my $m (@matches) {
1931                 if ($m->[0] > $chop_len) {
1932                         push @filtered, [ $chop_len, $chop_len + $tail_len ] if ($tail_len > 0);
1933                         last;
1934                 } elsif ($m->[1] > $chop_len) {
1935                         push @filtered, [ $m->[0], $chop_len + $tail_len ];
1936                         last;
1937                 }
1938                 push @filtered, $m;
1939         }
1940
1941         return esc_html_hl_regions($chopped . $tail, 'match', @filtered);
1942 }
1943
1944 ## ----------------------------------------------------------------------
1945 ## functions returning short strings
1946
1947 # CSS class for given age value (in seconds)
1948 sub age_class {
1949         my $age = shift;
1950
1951         if (!defined $age) {
1952                 return "noage";
1953         } elsif ($age < 60*60*2) {
1954                 return "age0";
1955         } elsif ($age < 60*60*24*2) {
1956                 return "age1";
1957         } else {
1958                 return "age2";
1959         }
1960 }
1961
1962 # convert age in seconds to "nn units ago" string
1963 sub age_string {
1964         my $age = shift;
1965         my $age_str;
1966
1967         if ($age > 60*60*24*365*2) {
1968                 $age_str = (int $age/60/60/24/365);
1969                 $age_str .= " years ago";
1970         } elsif ($age > 60*60*24*(365/12)*2) {
1971                 $age_str = int $age/60/60/24/(365/12);
1972                 $age_str .= " months ago";
1973         } elsif ($age > 60*60*24*7*2) {
1974                 $age_str = int $age/60/60/24/7;
1975                 $age_str .= " weeks ago";
1976         } elsif ($age > 60*60*24*2) {
1977                 $age_str = int $age/60/60/24;
1978                 $age_str .= " days ago";
1979         } elsif ($age > 60*60*2) {
1980                 $age_str = int $age/60/60;
1981                 $age_str .= " hours ago";
1982         } elsif ($age > 60*2) {
1983                 $age_str = int $age/60;
1984                 $age_str .= " min ago";
1985         } elsif ($age > 2) {
1986                 $age_str = int $age;
1987                 $age_str .= " sec ago";
1988         } else {
1989                 $age_str .= " right now";
1990         }
1991         return $age_str;
1992 }
1993
1994 use constant {
1995         S_IFINVALID => 0030000,
1996         S_IFGITLINK => 0160000,
1997 };
1998
1999 # submodule/subproject, a commit object reference
2000 sub S_ISGITLINK {
2001         my $mode = shift;
2002
2003         return (($mode & S_IFMT) == S_IFGITLINK)
2004 }
2005
2006 # convert file mode in octal to symbolic file mode string
2007 sub mode_str {
2008         my $mode = oct shift;
2009
2010         if (S_ISGITLINK($mode)) {
2011                 return 'm---------';
2012         } elsif (S_ISDIR($mode & S_IFMT)) {
2013                 return 'drwxr-xr-x';
2014         } elsif (S_ISLNK($mode)) {
2015                 return 'lrwxrwxrwx';
2016         } elsif (S_ISREG($mode)) {
2017                 # git cares only about the executable bit
2018                 if ($mode & S_IXUSR) {
2019                         return '-rwxr-xr-x';
2020                 } else {
2021                         return '-rw-r--r--';
2022                 };
2023         } else {
2024                 return '----------';
2025         }
2026 }
2027
2028 # convert file mode in octal to file type string
2029 sub file_type {
2030         my $mode = shift;
2031
2032         if ($mode !~ m/^[0-7]+$/) {
2033                 return $mode;
2034         } else {
2035                 $mode = oct $mode;
2036         }
2037
2038         if (S_ISGITLINK($mode)) {
2039                 return "submodule";
2040         } elsif (S_ISDIR($mode & S_IFMT)) {
2041                 return "directory";
2042         } elsif (S_ISLNK($mode)) {
2043                 return "symlink";
2044         } elsif (S_ISREG($mode)) {
2045                 return "file";
2046         } else {
2047                 return "unknown";
2048         }
2049 }
2050
2051 # convert file mode in octal to file type description string
2052 sub file_type_long {
2053         my $mode = shift;
2054
2055         if ($mode !~ m/^[0-7]+$/) {
2056                 return $mode;
2057         } else {
2058                 $mode = oct $mode;
2059         }
2060
2061         if (S_ISGITLINK($mode)) {
2062                 return "submodule";
2063         } elsif (S_ISDIR($mode & S_IFMT)) {
2064                 return "directory";
2065         } elsif (S_ISLNK($mode)) {
2066                 return "symlink";
2067         } elsif (S_ISREG($mode)) {
2068                 if ($mode & S_IXUSR) {
2069                         return "executable";
2070                 } else {
2071                         return "file";
2072                 };
2073         } else {
2074                 return "unknown";
2075         }
2076 }
2077
2078
2079 ## ----------------------------------------------------------------------
2080 ## functions returning short HTML fragments, or transforming HTML fragments
2081 ## which don't belong to other sections
2082
2083 # format line of commit message.
2084 sub format_log_line_html {
2085         my $line = shift;
2086
2087         # Potentially abbreviated OID.
2088         my $regex = oid_nlen_regex("7,64");
2089
2090         $line = esc_html($line, -nbsp=>1);
2091         $line =~ s{
2092         \b
2093         (
2094             # The output of "git describe", e.g. v2.10.0-297-gf6727b0
2095             # or hadoop-20160921-113441-20-g094fb7d
2096             (?<!-) # see strbuf_check_tag_ref(). Tags can't start with -
2097             [A-Za-z0-9.-]+
2098             (?!\.) # refs can't end with ".", see check_refname_format()
2099             -g$regex
2100             |
2101             # Just a normal looking Git SHA1
2102             $regex
2103         )
2104         \b
2105     }{
2106                 $cgi->a({-href => href(action=>"object", hash=>$1),
2107                                         -class => "text"}, $1);
2108         }egx;
2109
2110         return $line;
2111 }
2112
2113 # format marker of refs pointing to given object
2114
2115 # the destination action is chosen based on object type and current context:
2116 # - for annotated tags, we choose the tag view unless it's the current view
2117 #   already, in which case we go to shortlog view
2118 # - for other refs, we keep the current view if we're in history, shortlog or
2119 #   log view, and select shortlog otherwise
2120 sub format_ref_marker {
2121         my ($refs, $id) = @_;
2122         my $markers = '';
2123
2124         if (defined $refs->{$id}) {
2125                 foreach my $ref (@{$refs->{$id}}) {
2126                         # this code exploits the fact that non-lightweight tags are the
2127                         # only indirect objects, and that they are the only objects for which
2128                         # we want to use tag instead of shortlog as action
2129                         my ($type, $name) = qw();
2130                         my $indirect = ($ref =~ s/\^\{\}$//);
2131                         # e.g. tags/v2.6.11 or heads/next
2132                         if ($ref =~ m!^(.*?)s?/(.*)$!) {
2133                                 $type = $1;
2134                                 $name = $2;
2135                         } else {
2136                                 $type = "ref";
2137                                 $name = $ref;
2138                         }
2139
2140                         my $class = $type;
2141                         $class .= " indirect" if $indirect;
2142
2143                         my $dest_action = "shortlog";
2144
2145                         if ($indirect) {
2146                                 $dest_action = "tag" unless $action eq "tag";
2147                         } elsif ($action =~ /^(history|(short)?log)$/) {
2148                                 $dest_action = $action;
2149                         }
2150
2151                         my $dest = "";
2152                         $dest .= "refs/" unless $ref =~ m!^refs/!;
2153                         $dest .= $ref;
2154
2155                         my $link = $cgi->a({
2156                                 -href => href(
2157                                         action=>$dest_action,
2158                                         hash=>$dest
2159                                 )}, esc_html($name));
2160
2161                         $markers .= " <span class=\"".esc_attr($class)."\" title=\"".esc_attr($ref)."\">" .
2162                                 $link . "</span>";
2163                 }
2164         }
2165
2166         if ($markers) {
2167                 return ' <span class="refs">'. $markers . '</span>';
2168         } else {
2169                 return "";
2170         }
2171 }
2172
2173 # format, perhaps shortened and with markers, title line
2174 sub format_subject_html {
2175         my ($long, $short, $href, $extra) = @_;
2176         $extra = '' unless defined($extra);
2177
2178         if (length($short) < length($long)) {
2179                 $long =~ s/[[:cntrl:]]/?/g;
2180                 return $cgi->a({-href => $href, -class => "list subject",
2181                                 -title => to_utf8($long)},
2182                        esc_html($short)) . $extra;
2183         } else {
2184                 return $cgi->a({-href => $href, -class => "list subject"},
2185                        esc_html($long)) . $extra;
2186         }
2187 }
2188
2189 # Rather than recomputing the url for an email multiple times, we cache it
2190 # after the first hit. This gives a visible benefit in views where the avatar
2191 # for the same email is used repeatedly (e.g. shortlog).
2192 # The cache is shared by all avatar engines (currently gravatar only), which
2193 # are free to use it as preferred. Since only one avatar engine is used for any
2194 # given page, there's no risk for cache conflicts.
2195 our %avatar_cache = ();
2196
2197 # Compute the picon url for a given email, by using the picon search service over at
2198 # http://www.cs.indiana.edu/picons/search.html
2199 sub picon_url {
2200         my $email = lc shift;
2201         if (!$avatar_cache{$email}) {
2202                 my ($user, $domain) = split('@', $email);
2203                 $avatar_cache{$email} =
2204                         "//www.cs.indiana.edu/cgi-pub/kinzler/piconsearch.cgi/" .
2205                         "$domain/$user/" .
2206                         "users+domains+unknown/up/single";
2207         }
2208         return $avatar_cache{$email};
2209 }
2210
2211 # Compute the gravatar url for a given email, if it's not in the cache already.
2212 # Gravatar stores only the part of the URL before the size, since that's the
2213 # one computationally more expensive. This also allows reuse of the cache for
2214 # different sizes (for this particular engine).
2215 sub gravatar_url {
2216         my $email = lc shift;
2217         my $size = shift;
2218         $avatar_cache{$email} ||=
2219                 "//www.gravatar.com/avatar/" .
2220                         md5_hex($email) . "?s=";
2221         return $avatar_cache{$email} . $size;
2222 }
2223
2224 # Insert an avatar for the given $email at the given $size if the feature
2225 # is enabled.
2226 sub git_get_avatar {
2227         my ($email, %opts) = @_;
2228         my $pre_white  = ($opts{-pad_before} ? "&nbsp;" : "");
2229         my $post_white = ($opts{-pad_after}  ? "&nbsp;" : "");
2230         $opts{-size} ||= 'default';
2231         my $size = $avatar_size{$opts{-size}} || $avatar_size{'default'};
2232         my $url = "";
2233         if ($git_avatar eq 'gravatar') {
2234                 $url = gravatar_url($email, $size);
2235         } elsif ($git_avatar eq 'picon') {
2236                 $url = picon_url($email);
2237         }
2238         # Other providers can be added by extending the if chain, defining $url
2239         # as needed. If no variant puts something in $url, we assume avatars
2240         # are completely disabled/unavailable.
2241         if ($url) {
2242                 return $pre_white .
2243                        "<img width=\"$size\" " .
2244                             "class=\"avatar\" " .
2245                             "src=\"".esc_url($url)."\" " .
2246                             "alt=\"\" " .
2247                        "/>" . $post_white;
2248         } else {
2249                 return "";
2250         }
2251 }
2252
2253 sub format_search_author {
2254         my ($author, $searchtype, $displaytext) = @_;
2255         my $have_search = gitweb_check_feature('search');
2256
2257         if ($have_search) {
2258                 my $performed = "";
2259                 if ($searchtype eq 'author') {
2260                         $performed = "authored";
2261                 } elsif ($searchtype eq 'committer') {
2262                         $performed = "committed";
2263                 }
2264
2265                 return $cgi->a({-href => href(action=>"search", hash=>$hash,
2266                                 searchtext=>$author,
2267                                 searchtype=>$searchtype), class=>"list",
2268                                 title=>"Search for commits $performed by $author"},
2269                                 $displaytext);
2270
2271         } else {
2272                 return $displaytext;
2273         }
2274 }
2275
2276 # format the author name of the given commit with the given tag
2277 # the author name is chopped and escaped according to the other
2278 # optional parameters (see chop_str).
2279 sub format_author_html {
2280         my $tag = shift;
2281         my $co = shift;
2282         my $author = chop_and_escape_str($co->{'author_name'}, @_);
2283         return "<$tag class=\"author\">" .
2284                format_search_author($co->{'author_name'}, "author",
2285                        git_get_avatar($co->{'author_email'}, -pad_after => 1) .
2286                        $author) .
2287                "</$tag>";
2288 }
2289
2290 # format git diff header line, i.e. "diff --(git|combined|cc) ..."
2291 sub format_git_diff_header_line {
2292         my $line = shift;
2293         my $diffinfo = shift;
2294         my ($from, $to) = @_;
2295
2296         if ($diffinfo->{'nparents'}) {
2297                 # combined diff
2298                 $line =~ s!^(diff (.*?) )"?.*$!$1!;
2299                 if ($to->{'href'}) {
2300                         $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2301                                          esc_path($to->{'file'}));
2302                 } else { # file was deleted (no href)
2303                         $line .= esc_path($to->{'file'});
2304                 }
2305         } else {
2306                 # "ordinary" diff
2307                 $line =~ s!^(diff (.*?) )"?a/.*$!$1!;
2308                 if ($from->{'href'}) {
2309                         $line .= $cgi->a({-href => $from->{'href'}, -class => "path"},
2310                                          'a/' . esc_path($from->{'file'}));
2311                 } else { # file was added (no href)
2312                         $line .= 'a/' . esc_path($from->{'file'});
2313                 }
2314                 $line .= ' ';
2315                 if ($to->{'href'}) {
2316                         $line .= $cgi->a({-href => $to->{'href'}, -class => "path"},
2317                                          'b/' . esc_path($to->{'file'}));
2318                 } else { # file was deleted
2319                         $line .= 'b/' . esc_path($to->{'file'});
2320                 }
2321         }
2322
2323         return "<div class=\"diff header\">$line</div>\n";
2324 }
2325
2326 # format extended diff header line, before patch itself
2327 sub format_extended_diff_header_line {
2328         my $line = shift;
2329         my $diffinfo = shift;
2330         my ($from, $to) = @_;
2331
2332         # match <path>
2333         if ($line =~ s!^((copy|rename) from ).*$!$1! && $from->{'href'}) {
2334                 $line .= $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2335                                        esc_path($from->{'file'}));
2336         }
2337         if ($line =~ s!^((copy|rename) to ).*$!$1! && $to->{'href'}) {
2338                 $line .= $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2339                                  esc_path($to->{'file'}));
2340         }
2341         # match single <mode>
2342         if ($line =~ m/\s(\d{6})$/) {
2343                 $line .= '<span class="info"> (' .
2344                          file_type_long($1) .
2345                          ')</span>';
2346         }
2347         # match <hash>
2348         if ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", ",") |
2349             $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", ",")) {
2350                 # can match only for combined diff
2351                 $line = 'index ';
2352                 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2353                         if ($from->{'href'}[$i]) {
2354                                 $line .= $cgi->a({-href=>$from->{'href'}[$i],
2355                                                   -class=>"hash"},
2356                                                  substr($diffinfo->{'from_id'}[$i],0,7));
2357                         } else {
2358                                 $line .= '0' x 7;
2359                         }
2360                         # separator
2361                         $line .= ',' if ($i < $diffinfo->{'nparents'} - 1);
2362                 }
2363                 $line .= '..';
2364                 if ($to->{'href'}) {
2365                         $line .= $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2366                                          substr($diffinfo->{'to_id'},0,7));
2367                 } else {
2368                         $line .= '0' x 7;
2369                 }
2370
2371         } elsif ($line =~ oid_nlen_prefix_infix_regex($sha1_len, "index ", "..") |
2372                  $line =~ oid_nlen_prefix_infix_regex($sha256_len, "index ", "..")) {
2373                 # can match only for ordinary diff
2374                 my ($from_link, $to_link);
2375                 if ($from->{'href'}) {
2376                         $from_link = $cgi->a({-href=>$from->{'href'}, -class=>"hash"},
2377                                              substr($diffinfo->{'from_id'},0,7));
2378                 } else {
2379                         $from_link = '0' x 7;
2380                 }
2381                 if ($to->{'href'}) {
2382                         $to_link = $cgi->a({-href=>$to->{'href'}, -class=>"hash"},
2383                                            substr($diffinfo->{'to_id'},0,7));
2384                 } else {
2385                         $to_link = '0' x 7;
2386                 }
2387                 my ($from_id, $to_id) = ($diffinfo->{'from_id'}, $diffinfo->{'to_id'});
2388                 $line =~ s!$from_id\.\.$to_id!$from_link..$to_link!;
2389         }
2390
2391         return $line . "<br/>\n";
2392 }
2393
2394 # format from-file/to-file diff header
2395 sub format_diff_from_to_header {
2396         my ($from_line, $to_line, $diffinfo, $from, $to, @parents) = @_;
2397         my $line;
2398         my $result = '';
2399
2400         $line = $from_line;
2401         #assert($line =~ m/^---/) if DEBUG;
2402         # no extra formatting for "^--- /dev/null"
2403         if (! $diffinfo->{'nparents'}) {
2404                 # ordinary (single parent) diff
2405                 if ($line =~ m!^--- "?a/!) {
2406                         if ($from->{'href'}) {
2407                                 $line = '--- a/' .
2408                                         $cgi->a({-href=>$from->{'href'}, -class=>"path"},
2409                                                 esc_path($from->{'file'}));
2410                         } else {
2411                                 $line = '--- a/' .
2412                                         esc_path($from->{'file'});
2413                         }
2414                 }
2415                 $result .= qq!<div class="diff from_file">$line</div>\n!;
2416
2417         } else {
2418                 # combined diff (merge commit)
2419                 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
2420                         if ($from->{'href'}[$i]) {
2421                                 $line = '--- ' .
2422                                         $cgi->a({-href=>href(action=>"blobdiff",
2423                                                              hash_parent=>$diffinfo->{'from_id'}[$i],
2424                                                              hash_parent_base=>$parents[$i],
2425                                                              file_parent=>$from->{'file'}[$i],
2426                                                              hash=>$diffinfo->{'to_id'},
2427                                                              hash_base=>$hash,
2428                                                              file_name=>$to->{'file'}),
2429                                                  -class=>"path",
2430                                                  -title=>"diff" . ($i+1)},
2431                                                 $i+1) .
2432                                         '/' .
2433                                         $cgi->a({-href=>$from->{'href'}[$i], -class=>"path"},
2434                                                 esc_path($from->{'file'}[$i]));
2435                         } else {
2436                                 $line = '--- /dev/null';
2437                         }
2438                         $result .= qq!<div class="diff from_file">$line</div>\n!;
2439                 }
2440         }
2441
2442         $line = $to_line;
2443         #assert($line =~ m/^\+\+\+/) if DEBUG;
2444         # no extra formatting for "^+++ /dev/null"
2445         if ($line =~ m!^\+\+\+ "?b/!) {
2446                 if ($to->{'href'}) {
2447                         $line = '+++ b/' .
2448                                 $cgi->a({-href=>$to->{'href'}, -class=>"path"},
2449                                         esc_path($to->{'file'}));
2450                 } else {
2451                         $line = '+++ b/' .
2452                                 esc_path($to->{'file'});
2453                 }
2454         }
2455         $result .= qq!<div class="diff to_file">$line</div>\n!;
2456
2457         return $result;
2458 }
2459
2460 # create note for patch simplified by combined diff
2461 sub format_diff_cc_simplified {
2462         my ($diffinfo, @parents) = @_;
2463         my $result = '';
2464
2465         $result .= "<div class=\"diff header\">" .
2466                    "diff --cc ";
2467         if (!is_deleted($diffinfo)) {
2468                 $result .= $cgi->a({-href => href(action=>"blob",
2469                                                   hash_base=>$hash,
2470                                                   hash=>$diffinfo->{'to_id'},
2471                                                   file_name=>$diffinfo->{'to_file'}),
2472                                     -class => "path"},
2473                                    esc_path($diffinfo->{'to_file'}));
2474         } else {
2475                 $result .= esc_path($diffinfo->{'to_file'});
2476         }
2477         $result .= "</div>\n" . # class="diff header"
2478                    "<div class=\"diff nodifferences\">" .
2479                    "Simple merge" .
2480                    "</div>\n"; # class="diff nodifferences"
2481
2482         return $result;
2483 }
2484
2485 sub diff_line_class {
2486         my ($line, $from, $to) = @_;
2487
2488         # ordinary diff
2489         my $num_sign = 1;
2490         # combined diff
2491         if ($from && $to && ref($from->{'href'}) eq "ARRAY") {
2492                 $num_sign = scalar @{$from->{'href'}};
2493         }
2494
2495         my @diff_line_classifier = (
2496                 { regexp => qr/^\@\@{$num_sign} /, class => "chunk_header"},
2497                 { regexp => qr/^\\/,               class => "incomplete"  },
2498                 { regexp => qr/^ {$num_sign}/,     class => "ctx" },
2499                 # classifier for context must come before classifier add/rem,
2500                 # or we would have to use more complicated regexp, for example
2501                 # qr/(?= {0,$m}\+)[+ ]{$num_sign}/, where $m = $num_sign - 1;
2502                 { regexp => qr/^[+ ]{$num_sign}/,   class => "add" },
2503                 { regexp => qr/^[- ]{$num_sign}/,   class => "rem" },
2504         );
2505         for my $clsfy (@diff_line_classifier) {
2506                 return $clsfy->{'class'}
2507                         if ($line =~ $clsfy->{'regexp'});
2508         }
2509
2510         # fallback
2511         return "";
2512 }
2513
2514 # assumes that $from and $to are defined and correctly filled,
2515 # and that $line holds a line of chunk header for unified diff
2516 sub format_unidiff_chunk_header {
2517         my ($line, $from, $to) = @_;
2518
2519         my ($from_text, $from_start, $from_lines, $to_text, $to_start, $to_lines, $section) =
2520                 $line =~ m/^\@{2} (-(\d+)(?:,(\d+))?) (\+(\d+)(?:,(\d+))?) \@{2}(.*)$/;
2521
2522         $from_lines = 0 unless defined $from_lines;
2523         $to_lines   = 0 unless defined $to_lines;
2524
2525         if ($from->{'href'}) {
2526                 $from_text = $cgi->a({-href=>"$from->{'href'}#l$from_start",
2527                                      -class=>"list"}, $from_text);
2528         }
2529         if ($to->{'href'}) {
2530                 $to_text   = $cgi->a({-href=>"$to->{'href'}#l$to_start",
2531                                      -class=>"list"}, $to_text);
2532         }
2533         $line = "<span class=\"chunk_info\">@@ $from_text $to_text @@</span>" .
2534                 "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2535         return $line;
2536 }
2537
2538 # assumes that $from and $to are defined and correctly filled,
2539 # and that $line holds a line of chunk header for combined diff
2540 sub format_cc_diff_chunk_header {
2541         my ($line, $from, $to) = @_;
2542
2543         my ($prefix, $ranges, $section) = $line =~ m/^(\@+) (.*?) \@+(.*)$/;
2544         my (@from_text, @from_start, @from_nlines, $to_text, $to_start, $to_nlines);
2545
2546         @from_text = split(' ', $ranges);
2547         for (my $i = 0; $i < @from_text; ++$i) {
2548                 ($from_start[$i], $from_nlines[$i]) =
2549                         (split(',', substr($from_text[$i], 1)), 0);
2550         }
2551
2552         $to_text   = pop @from_text;
2553         $to_start  = pop @from_start;
2554         $to_nlines = pop @from_nlines;
2555
2556         $line = "<span class=\"chunk_info\">$prefix ";
2557         for (my $i = 0; $i < @from_text; ++$i) {
2558                 if ($from->{'href'}[$i]) {
2559                         $line .= $cgi->a({-href=>"$from->{'href'}[$i]#l$from_start[$i]",
2560                                           -class=>"list"}, $from_text[$i]);
2561                 } else {
2562                         $line .= $from_text[$i];
2563                 }
2564                 $line .= " ";
2565         }
2566         if ($to->{'href'}) {
2567                 $line .= $cgi->a({-href=>"$to->{'href'}#l$to_start",
2568                                   -class=>"list"}, $to_text);
2569         } else {
2570                 $line .= $to_text;
2571         }
2572         $line .= " $prefix</span>" .
2573                  "<span class=\"section\">" . esc_html($section, -nbsp=>1) . "</span>";
2574         return $line;
2575 }
2576
2577 # process patch (diff) line (not to be used for diff headers),
2578 # returning HTML-formatted (but not wrapped) line.
2579 # If the line is passed as a reference, it is treated as HTML and not
2580 # esc_html()'ed.
2581 sub format_diff_line {
2582         my ($line, $diff_class, $from, $to) = @_;
2583
2584         if (ref($line)) {
2585                 $line = $$line;
2586         } else {
2587                 chomp $line;
2588                 $line = untabify($line);
2589
2590                 if ($from && $to && $line =~ m/^\@{2} /) {
2591                         $line = format_unidiff_chunk_header($line, $from, $to);
2592                 } elsif ($from && $to && $line =~ m/^\@{3}/) {
2593                         $line = format_cc_diff_chunk_header($line, $from, $to);
2594                 } else {
2595                         $line = esc_html($line, -nbsp=>1);
2596                 }
2597         }
2598
2599         my $diff_classes = "diff";
2600         $diff_classes .= " $diff_class" if ($diff_class);
2601         $line = "<div class=\"$diff_classes\">$line</div>\n";
2602
2603         return $line;
2604 }
2605
2606 # Generates undef or something like "_snapshot_" or "snapshot (_tbz2_ _zip_)",
2607 # linked.  Pass the hash of the tree/commit to snapshot.
2608 sub format_snapshot_links {
2609         my ($hash) = @_;
2610         my $num_fmts = @snapshot_fmts;
2611         if ($num_fmts > 1) {
2612                 # A parenthesized list of links bearing format names.
2613                 # e.g. "snapshot (_tar.gz_ _zip_)"
2614                 return "snapshot (" . join(' ', map
2615                         $cgi->a({
2616                                 -href => href(
2617                                         action=>"snapshot",
2618                                         hash=>$hash,
2619                                         snapshot_format=>$_
2620                                 )
2621                         }, $known_snapshot_formats{$_}{'display'})
2622                 , @snapshot_fmts) . ")";
2623         } elsif ($num_fmts == 1) {
2624                 # A single "snapshot" link whose tooltip bears the format name.
2625                 # i.e. "_snapshot_"
2626                 my ($fmt) = @snapshot_fmts;
2627                 return
2628                         $cgi->a({
2629                                 -href => href(
2630                                         action=>"snapshot",
2631                                         hash=>$hash,
2632                                         snapshot_format=>$fmt
2633                                 ),
2634                                 -title => "in format: $known_snapshot_formats{$fmt}{'display'}"
2635                         }, "snapshot");
2636         } else { # $num_fmts == 0
2637                 return undef;
2638         }
2639 }
2640
2641 ## ......................................................................
2642 ## functions returning values to be passed, perhaps after some
2643 ## transformation, to other functions; e.g. returning arguments to href()
2644
2645 # returns hash to be passed to href to generate gitweb URL
2646 # in -title key it returns description of link
2647 sub get_feed_info {
2648         my $format = shift || 'Atom';
2649         my %res = (action => lc($format));
2650         my $matched_ref = 0;
2651
2652         # feed links are possible only for project views
2653         return unless (defined $project);
2654         # some views should link to OPML, or to generic project feed,
2655         # or don't have specific feed yet (so they should use generic)
2656         return if (!$action || $action =~ /^(?:tags|heads|forks|tag|search)$/x);
2657
2658         my $branch = undef;
2659         # branches refs uses 'refs/' + $get_branch_refs()[x] + '/' prefix
2660         # (fullname) to differentiate from tag links; this also makes
2661         # possible to detect branch links
2662         for my $ref (get_branch_refs()) {
2663                 if ((defined $hash_base && $hash_base =~ m!^refs/\Q$ref\E/(.*)$!) ||
2664                     (defined $hash      && $hash      =~ m!^refs/\Q$ref\E/(.*)$!)) {
2665                         $branch = $1;
2666                         $matched_ref = $ref;
2667                         last;
2668                 }
2669         }
2670         # find log type for feed description (title)
2671         my $type = 'log';
2672         if (defined $file_name) {
2673                 $type  = "history of $file_name";
2674                 $type .= "/" if ($action eq 'tree');
2675                 $type .= " on '$branch'" if (defined $branch);
2676         } else {
2677                 $type = "log of $branch" if (defined $branch);
2678         }
2679
2680         $res{-title} = $type;
2681         $res{'hash'} = (defined $branch ? "refs/$matched_ref/$branch" : undef);
2682         $res{'file_name'} = $file_name;
2683
2684         return %res;
2685 }
2686
2687 ## ----------------------------------------------------------------------
2688 ## git utility subroutines, invoking git commands
2689
2690 # returns path to the core git executable and the --git-dir parameter as list
2691 sub git_cmd {
2692         $number_of_git_cmds++;
2693         return $GIT, '--git-dir='.$git_dir;
2694 }
2695
2696 # quote the given arguments for passing them to the shell
2697 # quote_command("command", "arg 1", "arg with ' and ! characters")
2698 # => "'command' 'arg 1' 'arg with '\'' and '\!' characters'"
2699 # Try to avoid using this function wherever possible.
2700 sub quote_command {
2701         return join(' ',
2702                 map { my $a = $_; $a =~ s/(['!])/'\\$1'/g; "'$a'" } @_ );
2703 }
2704
2705 # get HEAD ref of given project as hash
2706 sub git_get_head_hash {
2707         return git_get_full_hash(shift, 'HEAD');
2708 }
2709
2710 sub git_get_full_hash {
2711         return git_get_hash(@_);
2712 }
2713
2714 sub git_get_short_hash {
2715         return git_get_hash(@_, '--short=7');
2716 }
2717
2718 sub git_get_hash {
2719         my ($project, $hash, @options) = @_;
2720         my $o_git_dir = $git_dir;
2721         my $retval = undef;
2722         $git_dir = "$projectroot/$project";
2723         if (open my $fd, '-|', git_cmd(), 'rev-parse',
2724             '--verify', '-q', @options, $hash) {
2725                 $retval = <$fd>;
2726                 chomp $retval if defined $retval;
2727                 close $fd;
2728         }
2729         if (defined $o_git_dir) {
2730                 $git_dir = $o_git_dir;
2731         }
2732         return $retval;
2733 }
2734
2735 # get type of given object
2736 sub git_get_type {
2737         my $hash = shift;
2738
2739         open my $fd, "-|", git_cmd(), "cat-file", '-t', $hash or return;
2740         my $type = <$fd>;
2741         close $fd or return;
2742         chomp $type;
2743         return $type;
2744 }
2745
2746 # repository configuration
2747 our $config_file = '';
2748 our %config;
2749
2750 # store multiple values for single key as anonymous array reference
2751 # single values stored directly in the hash, not as [ <value> ]
2752 sub hash_set_multi {
2753         my ($hash, $key, $value) = @_;
2754
2755         if (!exists $hash->{$key}) {
2756                 $hash->{$key} = $value;
2757         } elsif (!ref $hash->{$key}) {
2758                 $hash->{$key} = [ $hash->{$key}, $value ];
2759         } else {
2760                 push @{$hash->{$key}}, $value;
2761         }
2762 }
2763
2764 # return hash of git project configuration
2765 # optionally limited to some section, e.g. 'gitweb'
2766 sub git_parse_project_config {
2767         my $section_regexp = shift;
2768         my %config;
2769
2770         local $/ = "\0";
2771
2772         open my $fh, "-|", git_cmd(), "config", '-z', '-l',
2773                 or return;
2774
2775         while (my $keyval = <$fh>) {
2776                 chomp $keyval;
2777                 my ($key, $value) = split(/\n/, $keyval, 2);
2778
2779                 hash_set_multi(\%config, $key, $value)
2780                         if (!defined $section_regexp || $key =~ /^(?:$section_regexp)\./o);
2781         }
2782         close $fh;
2783
2784         return %config;
2785 }
2786
2787 # convert config value to boolean: 'true' or 'false'
2788 # no value, number > 0, 'true' and 'yes' values are true
2789 # rest of values are treated as false (never as error)
2790 sub config_to_bool {
2791         my $val = shift;
2792
2793         return 1 if !defined $val;             # section.key
2794
2795         # strip leading and trailing whitespace
2796         $val =~ s/^\s+//;
2797         $val =~ s/\s+$//;
2798
2799         return (($val =~ /^\d+$/ && $val) ||   # section.key = 1
2800                 ($val =~ /^(?:true|yes)$/i));  # section.key = true
2801 }
2802
2803 # convert config value to simple decimal number
2804 # an optional value suffix of 'k', 'm', or 'g' will cause the value
2805 # to be multiplied by 1024, 1048576, or 1073741824
2806 sub config_to_int {
2807         my $val = shift;
2808
2809         # strip leading and trailing whitespace
2810         $val =~ s/^\s+//;
2811         $val =~ s/\s+$//;
2812
2813         if (my ($num, $unit) = ($val =~ /^([0-9]*)([kmg])$/i)) {
2814                 $unit = lc($unit);
2815                 # unknown unit is treated as 1
2816                 return $num * ($unit eq 'g' ? 1073741824 :
2817                                $unit eq 'm' ?    1048576 :
2818                                $unit eq 'k' ?       1024 : 1);
2819         }
2820         return $val;
2821 }
2822
2823 # convert config value to array reference, if needed
2824 sub config_to_multi {
2825         my $val = shift;
2826
2827         return ref($val) ? $val : (defined($val) ? [ $val ] : []);
2828 }
2829
2830 sub git_get_project_config {
2831         my ($key, $type) = @_;
2832
2833         return unless defined $git_dir;
2834
2835         # key sanity check
2836         return unless ($key);
2837         # only subsection, if exists, is case sensitive,
2838         # and not lowercased by 'git config -z -l'
2839         if (my ($hi, $mi, $lo) = ($key =~ /^([^.]*)\.(.*)\.([^.]*)$/)) {
2840                 $lo =~ s/_//g;
2841                 $key = join(".", lc($hi), $mi, lc($lo));
2842                 return if ($lo =~ /\W/ || $hi =~ /\W/);
2843         } else {
2844                 $key = lc($key);
2845                 $key =~ s/_//g;
2846                 return if ($key =~ /\W/);
2847         }
2848         $key =~ s/^gitweb\.//;
2849
2850         # type sanity check
2851         if (defined $type) {
2852                 $type =~ s/^--//;
2853                 $type = undef
2854                         unless ($type eq 'bool' || $type eq 'int');
2855         }
2856
2857         # get config
2858         if (!defined $config_file ||
2859             $config_file ne "$git_dir/config") {
2860                 %config = git_parse_project_config('gitweb');
2861                 $config_file = "$git_dir/config";
2862         }
2863
2864         # check if config variable (key) exists
2865         return unless exists $config{"gitweb.$key"};
2866
2867         # ensure given type
2868         if (!defined $type) {
2869                 return $config{"gitweb.$key"};
2870         } elsif ($type eq 'bool') {
2871                 # backward compatibility: 'git config --bool' returns true/false
2872                 return config_to_bool($config{"gitweb.$key"}) ? 'true' : 'false';
2873         } elsif ($type eq 'int') {
2874                 return config_to_int($config{"gitweb.$key"});
2875         }
2876         return $config{"gitweb.$key"};
2877 }
2878
2879 # get hash of given path at given ref
2880 sub git_get_hash_by_path {
2881         my $base = shift;
2882         my $path = shift || return undef;
2883         my $type = shift;
2884
2885         $path =~ s,/+$,,;
2886
2887         open my $fd, "-|", git_cmd(), "ls-tree", $base, "--", $path
2888                 or die_error(500, "Open git-ls-tree failed");
2889         my $line = <$fd>;
2890         close $fd or return undef;
2891
2892         if (!defined $line) {
2893                 # there is no tree or hash given by $path at $base
2894                 return undef;
2895         }
2896
2897         #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa  panic.c'
2898         $line =~ m/^([0-9]+) (.+) ($oid_regex)\t/;
2899         if (defined $type && $type ne $2) {
2900                 # type doesn't match
2901                 return undef;
2902         }
2903         return $3;
2904 }
2905
2906 # get path of entry with given hash at given tree-ish (ref)
2907 # used to get 'from' filename for combined diff (merge commit) for renames
2908 sub git_get_path_by_hash {
2909         my $base = shift || return;
2910         my $hash = shift || return;
2911
2912         local $/ = "\0";
2913
2914         open my $fd, "-|", git_cmd(), "ls-tree", '-r', '-t', '-z', $base
2915                 or return undef;
2916         while (my $line = <$fd>) {
2917                 chomp $line;
2918
2919                 #'040000 tree 595596a6a9117ddba9fe379b6b012b558bac8423  gitweb'
2920                 #'100644 blob e02e90f0429be0d2a69b76571101f20b8f75530f  gitweb/README'
2921                 if ($line =~ m/(?:[0-9]+) (?:.+) $hash\t(.+)$/) {
2922                         close $fd;
2923                         return $1;
2924                 }
2925         }
2926         close $fd;
2927         return undef;
2928 }
2929
2930 ## ......................................................................
2931 ## git utility functions, directly accessing git repository
2932
2933 # get the value of config variable either from file named as the variable
2934 # itself in the repository ($GIT_DIR/$name file), or from gitweb.$name
2935 # configuration variable in the repository config file.
2936 sub git_get_file_or_project_config {
2937         my ($path, $name) = @_;
2938
2939         $git_dir = "$projectroot/$path";
2940         open my $fd, '<', "$git_dir/$name"
2941                 or return git_get_project_config($name);
2942         my $conf = <$fd>;
2943         close $fd;
2944         if (defined $conf) {
2945                 chomp $conf;
2946         }
2947         return $conf;
2948 }
2949
2950 sub git_get_project_description {
2951         my $path = shift;
2952         return git_get_file_or_project_config($path, 'description');
2953 }
2954
2955 sub git_get_project_category {
2956         my $path = shift;
2957         return git_get_file_or_project_config($path, 'category');
2958 }
2959
2960
2961 # supported formats:
2962 # * $GIT_DIR/ctags/<tagname> file (in 'ctags' subdirectory)
2963 #   - if its contents is a number, use it as tag weight,
2964 #   - otherwise add a tag with weight 1
2965 # * $GIT_DIR/ctags file, each line is a tag (with weight 1)
2966 #   the same value multiple times increases tag weight
2967 # * `gitweb.ctag' multi-valued repo config variable
2968 sub git_get_project_ctags {
2969         my $project = shift;
2970         my $ctags = {};
2971
2972         $git_dir = "$projectroot/$project";
2973         if (opendir my $dh, "$git_dir/ctags") {
2974                 my @files = grep { -f $_ } map { "$git_dir/ctags/$_" } readdir($dh);
2975                 foreach my $tagfile (@files) {
2976                         open my $ct, '<', $tagfile
2977                                 or next;
2978                         my $val = <$ct>;
2979                         chomp $val if $val;
2980                         close $ct;
2981
2982                         (my $ctag = $tagfile) =~ s#.*/##;
2983                         if ($val =~ /^\d+$/) {
2984                                 $ctags->{$ctag} = $val;
2985                         } else {
2986                                 $ctags->{$ctag} = 1;
2987                         }
2988                 }
2989                 closedir $dh;
2990
2991         } elsif (open my $fh, '<', "$git_dir/ctags") {
2992                 while (my $line = <$fh>) {
2993                         chomp $line;
2994                         $ctags->{$line}++ if $line;
2995                 }
2996                 close $fh;
2997
2998         } else {
2999                 my $taglist = config_to_multi(git_get_project_config('ctag'));
3000                 foreach my $tag (@$taglist) {
3001                         $ctags->{$tag}++;
3002                 }
3003         }
3004
3005         return $ctags;
3006 }
3007
3008 # return hash, where keys are content tags ('ctags'),
3009 # and values are sum of weights of given tag in every project
3010 sub git_gather_all_ctags {
3011         my $projects = shift;
3012         my $ctags = {};
3013
3014         foreach my $p (@$projects) {
3015                 foreach my $ct (keys %{$p->{'ctags'}}) {
3016                         $ctags->{$ct} += $p->{'ctags'}->{$ct};
3017                 }
3018         }
3019
3020         return $ctags;
3021 }
3022
3023 sub git_populate_project_tagcloud {
3024         my $ctags = shift;
3025
3026         # First, merge different-cased tags; tags vote on casing
3027         my %ctags_lc;
3028         foreach (keys %$ctags) {
3029                 $ctags_lc{lc $_}->{count} += $ctags->{$_};
3030                 if (not $ctags_lc{lc $_}->{topcount}
3031                     or $ctags_lc{lc $_}->{topcount} < $ctags->{$_}) {
3032                         $ctags_lc{lc $_}->{topcount} = $ctags->{$_};
3033                         $ctags_lc{lc $_}->{topname} = $_;
3034                 }
3035         }
3036
3037         my $cloud;
3038         my $matched = $input_params{'ctag'};
3039         if (eval { require HTML::TagCloud; 1; }) {
3040                 $cloud = HTML::TagCloud->new;
3041                 foreach my $ctag (sort keys %ctags_lc) {
3042                         # Pad the title with spaces so that the cloud looks
3043                         # less crammed.
3044                         my $title = esc_html($ctags_lc{$ctag}->{topname});
3045                         $title =~ s/ /&nbsp;/g;
3046                         $title =~ s/^/&nbsp;/g;
3047                         $title =~ s/$/&nbsp;/g;
3048                         if (defined $matched && $matched eq $ctag) {
3049                                 $title = qq(<span class="match">$title</span>);
3050                         }
3051                         $cloud->add($title, href(project=>undef, ctag=>$ctag),
3052                                     $ctags_lc{$ctag}->{count});
3053                 }
3054         } else {
3055                 $cloud = {};
3056                 foreach my $ctag (keys %ctags_lc) {
3057                         my $title = esc_html($ctags_lc{$ctag}->{topname}, -nbsp=>1);
3058                         if (defined $matched && $matched eq $ctag) {
3059                                 $title = qq(<span class="match">$title</span>);
3060                         }
3061                         $cloud->{$ctag}{count} = $ctags_lc{$ctag}->{count};
3062                         $cloud->{$ctag}{ctag} =
3063                                 $cgi->a({-href=>href(project=>undef, ctag=>$ctag)}, $title);
3064                 }
3065         }
3066         return $cloud;
3067 }
3068
3069 sub git_show_project_tagcloud {
3070         my ($cloud, $count) = @_;
3071         if (ref $cloud eq 'HTML::TagCloud') {
3072                 return $cloud->html_and_css($count);
3073         } else {
3074                 my @tags = sort { $cloud->{$a}->{'count'} <=> $cloud->{$b}->{'count'} } keys %$cloud;
3075                 return
3076                         '<div id="htmltagcloud"'.($project ? '' : ' align="center"').'>' .
3077                         join (', ', map {
3078                                 $cloud->{$_}->{'ctag'}
3079                         } splice(@tags, 0, $count)) .
3080                         '</div>';
3081         }
3082 }
3083
3084 sub git_get_project_url_list {
3085         my $path = shift;
3086
3087         $git_dir = "$projectroot/$path";
3088         open my $fd, '<', "$git_dir/cloneurl"
3089                 or return wantarray ?
3090                 @{ config_to_multi(git_get_project_config('url')) } :
3091                    config_to_multi(git_get_project_config('url'));
3092         my @git_project_url_list = map { chomp; $_ } <$fd>;
3093         close $fd;
3094
3095         return wantarray ? @git_project_url_list : \@git_project_url_list;
3096 }
3097
3098 sub git_get_projects_list {
3099         my $filter = shift || '';
3100         my $paranoid = shift;
3101         my @list;
3102
3103         if (-d $projects_list) {
3104                 # search in directory
3105                 my $dir = $projects_list;
3106                 # remove the trailing "/"
3107                 $dir =~ s!/+$!!;
3108                 my $pfxlen = length("$dir");
3109                 my $pfxdepth = ($dir =~ tr!/!!);
3110                 # when filtering, search only given subdirectory
3111                 if ($filter && !$paranoid) {
3112                         $dir .= "/$filter";
3113                         $dir =~ s!/+$!!;
3114                 }
3115
3116                 File::Find::find({
3117                         follow_fast => 1, # follow symbolic links
3118                         follow_skip => 2, # ignore duplicates
3119                         dangling_symlinks => 0, # ignore dangling symlinks, silently
3120                         wanted => sub {
3121                                 # global variables
3122                                 our $project_maxdepth;
3123                                 our $projectroot;
3124                                 # skip project-list toplevel, if we get it.
3125                                 return if (m!^[/.]$!);
3126                                 # only directories can be git repositories
3127                                 return unless (-d $_);
3128                                 # need search permission
3129                                 return unless (-x $_);
3130                                 # don't traverse too deep (Find is super slow on os x)
3131                                 # $project_maxdepth excludes depth of $projectroot
3132                                 if (($File::Find::name =~ tr!/!!) - $pfxdepth > $project_maxdepth) {
3133                                         $File::Find::prune = 1;
3134                                         return;
3135                                 }
3136
3137                                 my $path = substr($File::Find::name, $pfxlen + 1);
3138                                 # paranoidly only filter here
3139                                 if ($paranoid && $filter && $path !~ m!^\Q$filter\E/!) {
3140                                         next;
3141                                 }
3142                                 # we check related file in $projectroot
3143                                 if (check_export_ok("$projectroot/$path")) {
3144                                         push @list, { path => $path };
3145                                         $File::Find::prune = 1;
3146                                 }
3147                         },
3148                 }, "$dir");
3149
3150         } elsif (-f $projects_list) {
3151                 # read from file(url-encoded):
3152                 # 'git%2Fgit.git Linus+Torvalds'
3153                 # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3154                 # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3155                 open my $fd, '<', $projects_list or return;
3156         PROJECT:
3157                 while (my $line = <$fd>) {
3158                         chomp $line;
3159                         my ($path, $owner) = split ' ', $line;
3160                         $path = unescape($path);
3161                         $owner = unescape($owner);
3162                         if (!defined $path) {
3163                                 next;
3164                         }
3165                         # if $filter is rpovided, check if $path begins with $filter
3166                         if ($filter && $path !~ m!^\Q$filter\E/!) {
3167                                 next;
3168                         }
3169                         if (check_export_ok("$projectroot/$path")) {
3170                                 my $pr = {
3171                                         path => $path
3172                                 };
3173                                 if ($owner) {
3174                                         $pr->{'owner'} = to_utf8($owner);
3175                                 }
3176                                 push @list, $pr;
3177                         }
3178                 }
3179                 close $fd;
3180         }
3181         return @list;
3182 }
3183
3184 # written with help of Tree::Trie module (Perl Artistic License, GPL compatible)
3185 # as side effects it sets 'forks' field to list of forks for forked projects
3186 sub filter_forks_from_projects_list {
3187         my $projects = shift;
3188
3189         my %trie; # prefix tree of directories (path components)
3190         # generate trie out of those directories that might contain forks
3191         foreach my $pr (@$projects) {
3192                 my $path = $pr->{'path'};
3193                 $path =~ s/\.git$//;      # forks of 'repo.git' are in 'repo/' directory
3194                 next if ($path =~ m!/$!); # skip non-bare repositories, e.g. 'repo/.git'
3195                 next unless ($path);      # skip '.git' repository: tests, git-instaweb
3196                 next unless (-d "$projectroot/$path"); # containing directory exists
3197                 $pr->{'forks'} = [];      # there can be 0 or more forks of project
3198
3199                 # add to trie
3200                 my @dirs = split('/', $path);
3201                 # walk the trie, until either runs out of components or out of trie
3202                 my $ref = \%trie;
3203                 while (scalar @dirs &&
3204                        exists($ref->{$dirs[0]})) {
3205                         $ref = $ref->{shift @dirs};
3206                 }
3207                 # create rest of trie structure from rest of components
3208                 foreach my $dir (@dirs) {
3209                         $ref = $ref->{$dir} = {};
3210                 }
3211                 # create end marker, store $pr as a data
3212                 $ref->{''} = $pr if (!exists $ref->{''});
3213         }
3214
3215         # filter out forks, by finding shortest prefix match for paths
3216         my @filtered;
3217  PROJECT:
3218         foreach my $pr (@$projects) {
3219                 # trie lookup
3220                 my $ref = \%trie;
3221         DIR:
3222                 foreach my $dir (split('/', $pr->{'path'})) {
3223                         if (exists $ref->{''}) {
3224                                 # found [shortest] prefix, is a fork - skip it
3225                                 push @{$ref->{''}{'forks'}}, $pr;
3226                                 next PROJECT;
3227                         }
3228                         if (!exists $ref->{$dir}) {
3229                                 # not in trie, cannot have prefix, not a fork
3230                                 push @filtered, $pr;
3231                                 next PROJECT;
3232                         }
3233                         # If the dir is there, we just walk one step down the trie.
3234                         $ref = $ref->{$dir};
3235                 }
3236                 # we ran out of trie
3237                 # (shouldn't happen: it's either no match, or end marker)
3238                 push @filtered, $pr;
3239         }
3240
3241         return @filtered;
3242 }
3243
3244 # note: fill_project_list_info must be run first,
3245 # for 'descr_long' and 'ctags' to be filled
3246 sub search_projects_list {
3247         my ($projlist, %opts) = @_;
3248         my $tagfilter  = $opts{'tagfilter'};
3249         my $search_re = $opts{'search_regexp'};
3250
3251         return @$projlist
3252                 unless ($tagfilter || $search_re);
3253
3254         # searching projects require filling to be run before it;
3255         fill_project_list_info($projlist,
3256                                $tagfilter  ? 'ctags' : (),
3257                                $search_re ? ('path', 'descr') : ());
3258         my @projects;
3259  PROJECT:
3260         foreach my $pr (@$projlist) {
3261
3262                 if ($tagfilter) {
3263                         next unless ref($pr->{'ctags'}) eq 'HASH';
3264                         next unless
3265                                 grep { lc($_) eq lc($tagfilter) } keys %{$pr->{'ctags'}};
3266                 }
3267
3268                 if ($search_re) {
3269                         next unless
3270                                 $pr->{'path'} =~ /$search_re/ ||
3271                                 $pr->{'descr_long'} =~ /$search_re/;
3272                 }
3273
3274                 push @projects, $pr;
3275         }
3276
3277         return @projects;
3278 }
3279
3280 our $gitweb_project_owner = undef;
3281 sub git_get_project_list_from_file {
3282
3283         return if (defined $gitweb_project_owner);
3284
3285         $gitweb_project_owner = {};
3286         # read from file (url-encoded):
3287         # 'git%2Fgit.git Linus+Torvalds'
3288         # 'libs%2Fklibc%2Fklibc.git H.+Peter+Anvin'
3289         # 'linux%2Fhotplug%2Fudev.git Greg+Kroah-Hartman'
3290         if (-f $projects_list) {
3291                 open(my $fd, '<', $projects_list);
3292                 while (my $line = <$fd>) {
3293                         chomp $line;
3294                         my ($pr, $ow) = split ' ', $line;
3295                         $pr = unescape($pr);
3296                         $ow = unescape($ow);
3297                         $gitweb_project_owner->{$pr} = to_utf8($ow);
3298                 }
3299                 close $fd;
3300         }
3301 }
3302
3303 sub git_get_project_owner {
3304         my $project = shift;
3305         my $owner;
3306
3307         return undef unless $project;
3308         $git_dir = "$projectroot/$project";
3309
3310         if (!defined $gitweb_project_owner) {
3311                 git_get_project_list_from_file();
3312         }
3313
3314         if (exists $gitweb_project_owner->{$project}) {
3315                 $owner = $gitweb_project_owner->{$project};
3316         }
3317         if (!defined $owner){
3318                 $owner = git_get_project_config('owner');
3319         }
3320         if (!defined $owner) {
3321                 $owner = get_file_owner("$git_dir");
3322         }
3323
3324         return $owner;
3325 }
3326
3327 sub git_get_last_activity {
3328         my ($path) = @_;
3329         my $fd;
3330
3331         $git_dir = "$projectroot/$path";
3332         open($fd, "-|", git_cmd(), 'for-each-ref',
3333              '--format=%(committer)',
3334              '--sort=-committerdate',
3335              '--count=1',
3336              map { "refs/$_" } get_branch_refs ()) or return;
3337         my $most_recent = <$fd>;
3338         close $fd or return;
3339         if (defined $most_recent &&
3340             $most_recent =~ / (\d+) [-+][01]\d\d\d$/) {
3341                 my $timestamp = $1;
3342                 my $age = time - $timestamp;
3343                 return ($age, age_string($age));
3344         }
3345         return (undef, undef);
3346 }
3347
3348 # Implementation note: when a single remote is wanted, we cannot use 'git
3349 # remote show -n' because that command always work (assuming it's a remote URL
3350 # if it's not defined), and we cannot use 'git remote show' because that would
3351 # try to make a network roundtrip. So the only way to find if that particular
3352 # remote is defined is to walk the list provided by 'git remote -v' and stop if
3353 # and when we find what we want.
3354 sub git_get_remotes_list {
3355         my $wanted = shift;
3356         my %remotes = ();
3357
3358         open my $fd, '-|' , git_cmd(), 'remote', '-v';
3359         return unless $fd;
3360         while (my $remote = <$fd>) {
3361                 chomp $remote;
3362                 $remote =~ s!\t(.*?)\s+\((\w+)\)$!!;
3363                 next if $wanted and not $remote eq $wanted;
3364                 my ($url, $key) = ($1, $2);
3365
3366                 $remotes{$remote} ||= { 'heads' => () };
3367                 $remotes{$remote}{$key} = $url;
3368         }
3369         close $fd or return;
3370         return wantarray ? %remotes : \%remotes;
3371 }
3372
3373 # Takes a hash of remotes as first parameter and fills it by adding the
3374 # available remote heads for each of the indicated remotes.
3375 sub fill_remote_heads {
3376         my $remotes = shift;
3377         my @heads = map { "remotes/$_" } keys %$remotes;
3378         my @remoteheads = git_get_heads_list(undef, @heads);
3379         foreach my $remote (keys %$remotes) {
3380                 $remotes->{$remote}{'heads'} = [ grep {
3381                         $_->{'name'} =~ s!^$remote/!!
3382                         } @remoteheads ];
3383         }
3384 }
3385
3386 sub git_get_references {
3387         my $type = shift || "";
3388         my %refs;
3389         # 5dc01c595e6c6ec9ccda4f6f69c131c0dd945f8c refs/tags/v2.6.11
3390         # c39ae07f393806ccf406ef966e9a15afc43cc36a refs/tags/v2.6.11^{}
3391         open my $fd, "-|", git_cmd(), "show-ref", "--dereference",
3392                 ($type ? ("--", "refs/$type") : ()) # use -- <pattern> if $type
3393                 or return;
3394
3395         while (my $line = <$fd>) {
3396                 chomp $line;
3397                 if ($line =~ m!^($oid_regex)\srefs/($type.*)$!) {
3398                         if (defined $refs{$1}) {
3399                                 push @{$refs{$1}}, $2;
3400                         } else {
3401                                 $refs{$1} = [ $2 ];
3402                         }
3403                 }
3404         }
3405         close $fd or return;
3406         return \%refs;
3407 }
3408
3409 sub git_get_rev_name_tags {
3410         my $hash = shift || return undef;
3411
3412         open my $fd, "-|", git_cmd(), "name-rev", "--tags", $hash
3413                 or return;
3414         my $name_rev = <$fd>;
3415         close $fd;
3416
3417         if ($name_rev =~ m|^$hash tags/(.*)$|) {
3418                 return $1;
3419         } else {
3420                 # catches also '$hash undefined' output
3421                 return undef;
3422         }
3423 }
3424
3425 ## ----------------------------------------------------------------------
3426 ## parse to hash functions
3427
3428 sub parse_date {
3429         my $epoch = shift;
3430         my $tz = shift || "-0000";
3431
3432         my %date;
3433         my @months = ("Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec");
3434         my @days = ("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat");
3435         my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($epoch);
3436         $date{'hour'} = $hour;
3437         $date{'minute'} = $min;
3438         $date{'mday'} = $mday;
3439         $date{'day'} = $days[$wday];
3440         $date{'month'} = $months[$mon];
3441         $date{'rfc2822'}   = sprintf "%s, %d %s %4d %02d:%02d:%02d +0000",
3442                              $days[$wday], $mday, $months[$mon], 1900+$year, $hour ,$min, $sec;
3443         $date{'mday-time'} = sprintf "%d %s %02d:%02d",
3444                              $mday, $months[$mon], $hour ,$min;
3445         $date{'iso-8601'}  = sprintf "%04d-%02d-%02dT%02d:%02d:%02dZ",
3446                              1900+$year, 1+$mon, $mday, $hour ,$min, $sec;
3447
3448         my ($tz_sign, $tz_hour, $tz_min) =
3449                 ($tz =~ m/^([-+])(\d\d)(\d\d)$/);
3450         $tz_sign = ($tz_sign eq '-' ? -1 : +1);
3451         my $local = $epoch + $tz_sign*((($tz_hour*60) + $tz_min)*60);
3452         ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($local);
3453         $date{'hour_local'} = $hour;
3454         $date{'minute_local'} = $min;
3455         $date{'tz_local'} = $tz;
3456         $date{'iso-tz'} = sprintf("%04d-%02d-%02d %02d:%02d:%02d %s",
3457                                   1900+$year, $mon+1, $mday,
3458                                   $hour, $min, $sec, $tz);
3459         return %date;
3460 }
3461
3462 sub hide_mailaddrs_if_private {
3463         my $line = shift;
3464         return $line unless gitweb_check_feature('email-privacy');
3465         $line =~ s/<[^@>]+@[^>]+>/<redacted>/g;
3466         return $line;
3467 }
3468
3469 sub parse_tag {
3470         my $tag_id = shift;
3471         my %tag;
3472         my @comment;
3473
3474         open my $fd, "-|", git_cmd(), "cat-file", "tag", $tag_id or return;
3475         $tag{'id'} = $tag_id;
3476         while (my $line = <$fd>) {
3477                 chomp $line;
3478                 if ($line =~ m/^object ($oid_regex)$/) {
3479                         $tag{'object'} = $1;
3480                 } elsif ($line =~ m/^type (.+)$/) {
3481                         $tag{'type'} = $1;
3482                 } elsif ($line =~ m/^tag (.+)$/) {
3483                         $tag{'name'} = $1;
3484                 } elsif ($line =~ m/^tagger (.*) ([0-9]+) (.*)$/) {
3485                         $tag{'author'} = hide_mailaddrs_if_private($1);
3486                         $tag{'author_epoch'} = $2;
3487                         $tag{'author_tz'} = $3;
3488                         if ($tag{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3489                                 $tag{'author_name'}  = $1;
3490                                 $tag{'author_email'} = $2;
3491                         } else {
3492                                 $tag{'author_name'} = $tag{'author'};
3493                         }
3494                 } elsif ($line =~ m/--BEGIN/) {
3495                         push @comment, $line;
3496                         last;
3497                 } elsif ($line eq "") {
3498                         last;
3499                 }
3500         }
3501         push @comment, <$fd>;
3502         $tag{'comment'} = \@comment;
3503         close $fd or return;
3504         if (!defined $tag{'name'}) {
3505                 return
3506         };
3507         return %tag
3508 }
3509
3510 sub parse_commit_text {
3511         my ($commit_text, $withparents) = @_;
3512         my @commit_lines = split '\n', $commit_text;
3513         my %co;
3514
3515         pop @commit_lines; # Remove '\0'
3516
3517         if (! @commit_lines) {
3518                 return;
3519         }
3520
3521         my $header = shift @commit_lines;
3522         if ($header !~ m/^$oid_regex/) {
3523                 return;
3524         }
3525         ($co{'id'}, my @parents) = split ' ', $header;
3526         while (my $line = shift @commit_lines) {
3527                 last if $line eq "\n";
3528                 if ($line =~ m/^tree ($oid_regex)$/) {
3529                         $co{'tree'} = $1;
3530                 } elsif ((!defined $withparents) && ($line =~ m/^parent ($oid_regex)$/)) {
3531                         push @parents, $1;
3532                 } elsif ($line =~ m/^author (.*) ([0-9]+) (.*)$/) {
3533                         $co{'author'} = hide_mailaddrs_if_private(to_utf8($1));
3534                         $co{'author_epoch'} = $2;
3535                         $co{'author_tz'} = $3;
3536                         if ($co{'author'} =~ m/^([^<]+) <([^>]*)>/) {
3537                                 $co{'author_name'}  = $1;
3538                                 $co{'author_email'} = $2;
3539                         } else {
3540                                 $co{'author_name'} = $co{'author'};
3541                         }
3542                 } elsif ($line =~ m/^committer (.*) ([0-9]+) (.*)$/) {
3543                         $co{'committer'} = hide_mailaddrs_if_private(to_utf8($1));
3544                         $co{'committer_epoch'} = $2;
3545                         $co{'committer_tz'} = $3;
3546                         if ($co{'committer'} =~ m/^([^<]+) <([^>]*)>/) {
3547                                 $co{'committer_name'}  = $1;
3548                                 $co{'committer_email'} = $2;
3549                         } else {
3550                                 $co{'committer_name'} = $co{'committer'};
3551                         }
3552                 }
3553         }
3554         if (!defined $co{'tree'}) {
3555                 return;
3556         };
3557         $co{'parents'} = \@parents;
3558         $co{'parent'} = $parents[0];
3559
3560         foreach my $title (@commit_lines) {
3561                 $title =~ s/^    //;
3562                 if ($title ne "") {
3563                         $co{'title'} = chop_str($title, 80, 5);
3564                         $co{'title_short'} = chop_str($title, 50, 5);
3565                         last;
3566                 }
3567         }
3568         if (! defined $co{'title'} || $co{'title'} eq "") {
3569                 $co{'title'} = $co{'title_short'} = '(no commit message)';
3570         }
3571         # remove added spaces, redact e-mail addresses if applicable.
3572         foreach my $line (@commit_lines) {
3573                 $line =~ s/^    //;
3574                 $line = hide_mailaddrs_if_private($line);
3575         }
3576         $co{'comment'} = \@commit_lines;
3577
3578         my $age = time - $co{'committer_epoch'};
3579         $co{'age'} = $age;
3580         $co{'age_string'} = age_string($age);
3581         my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime($co{'committer_epoch'});
3582         if ($age > 60*60*24*7*2) {
3583                 $co{'age_string_date'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3584                 $co{'age_string_age'} = $co{'age_string'};
3585         } else {
3586                 $co{'age_string_date'} = $co{'age_string'};
3587                 $co{'age_string_age'} = sprintf "%4i-%02u-%02i", 1900 + $year, $mon+1, $mday;
3588         }
3589         return %co;
3590 }
3591
3592 sub parse_commit {
3593         my ($commit_id) = @_;
3594         my %co;
3595
3596         local $/ = "\0";
3597
3598         open my $fd, "-|", git_cmd(), "rev-list",
3599                 "--parents",
3600                 "--header",
3601                 "--max-count=1",
3602                 $commit_id,
3603                 "--",
3604                 or die_error(500, "Open git-rev-list failed");
3605         %co = parse_commit_text(<$fd>, 1);
3606         close $fd;
3607
3608         return %co;
3609 }
3610
3611 sub parse_commits {
3612         my ($commit_id, $maxcount, $skip, $filename, @args) = @_;
3613         my @cos;
3614
3615         $maxcount ||= 1;
3616         $skip ||= 0;
3617
3618         local $/ = "\0";
3619
3620         open my $fd, "-|", git_cmd(), "rev-list",
3621                 "--header",
3622                 @args,
3623                 ("--max-count=" . $maxcount),
3624                 ("--skip=" . $skip),
3625                 @extra_options,
3626                 $commit_id,
3627                 "--",
3628                 ($filename ? ($filename) : ())
3629                 or die_error(500, "Open git-rev-list failed");
3630         while (my $line = <$fd>) {
3631                 my %co = parse_commit_text($line);
3632                 push @cos, \%co;
3633         }
3634         close $fd;
3635
3636         return wantarray ? @cos : \@cos;
3637 }
3638
3639 # parse line of git-diff-tree "raw" output
3640 sub parse_difftree_raw_line {
3641         my $line = shift;
3642         my %res;
3643
3644         # ':100644 100644 03b218260e99b78c6df0ed378e59ed9205ccc96d 3b93d5e7cc7f7dd4ebed13a5cc1a4ad976fc94d8 M   ls-files.c'
3645         # ':100644 100644 7f9281985086971d3877aca27704f2aaf9c448ce bc190ebc71bbd923f2b728e505408f5e54bd073a M   rev-tree.c'
3646         if ($line =~ m/^:([0-7]{6}) ([0-7]{6}) ($oid_regex) ($oid_regex) (.)([0-9]{0,3})\t(.*)$/) {
3647                 $res{'from_mode'} = $1;
3648                 $res{'to_mode'} = $2;
3649                 $res{'from_id'} = $3;
3650                 $res{'to_id'} = $4;
3651                 $res{'status'} = $5;
3652                 $res{'similarity'} = $6;
3653                 if ($res{'status'} eq 'R' || $res{'status'} eq 'C') { # renamed or copied
3654                         ($res{'from_file'}, $res{'to_file'}) = map { unquote($_) } split("\t", $7);
3655                 } else {
3656                         $res{'from_file'} = $res{'to_file'} = $res{'file'} = unquote($7);
3657                 }
3658         }
3659         # '::100755 100755 100755 60e79ca1b01bc8b057abe17ddab484699a7f5fdb 94067cc5f73388f33722d52ae02f44692bc07490 94067cc5f73388f33722d52ae02f44692bc07490 MR git-gui/git-gui.sh'
3660         # combined diff (for merge commit)
3661         elsif ($line =~ s/^(::+)((?:[0-7]{6} )+)((?:$oid_regex )+)([a-zA-Z]+)\t(.*)$//) {
3662                 $res{'nparents'}  = length($1);
3663                 $res{'from_mode'} = [ split(' ', $2) ];
3664                 $res{'to_mode'} = pop @{$res{'from_mode'}};
3665                 $res{'from_id'} = [ split(' ', $3) ];
3666                 $res{'to_id'} = pop @{$res{'from_id'}};
3667                 $res{'status'} = [ split('', $4) ];
3668                 $res{'to_file'} = unquote($5);
3669         }
3670         # 'c512b523472485aef4fff9e57b229d9d243c967f'
3671         elsif ($line =~ m/^($oid_regex)$/) {
3672                 $res{'commit'} = $1;
3673         }
3674
3675         return wantarray ? %res : \%res;
3676 }
3677
3678 # wrapper: return parsed line of git-diff-tree "raw" output
3679 # (the argument might be raw line, or parsed info)
3680 sub parsed_difftree_line {
3681         my $line_or_ref = shift;
3682
3683         if (ref($line_or_ref) eq "HASH") {
3684                 # pre-parsed (or generated by hand)
3685                 return $line_or_ref;
3686         } else {
3687                 return parse_difftree_raw_line($line_or_ref);
3688         }
3689 }
3690
3691 # parse line of git-ls-tree output
3692 sub parse_ls_tree_line {
3693         my $line = shift;
3694         my %opts = @_;
3695         my %res;
3696
3697         if ($opts{'-l'}) {
3698                 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa   16717  panic.c'
3699                 $line =~ m/^([0-9]+) (.+) ($oid_regex) +(-|[0-9]+)\t(.+)$/s;
3700
3701                 $res{'mode'} = $1;
3702                 $res{'type'} = $2;
3703                 $res{'hash'} = $3;
3704                 $res{'size'} = $4;
3705                 if ($opts{'-z'}) {
3706                         $res{'name'} = $5;
3707                 } else {
3708                         $res{'name'} = unquote($5);
3709                 }
3710         } else {
3711                 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa  panic.c'
3712                 $line =~ m/^([0-9]+) (.+) ($oid_regex)\t(.+)$/s;
3713
3714                 $res{'mode'} = $1;
3715                 $res{'type'} = $2;
3716                 $res{'hash'} = $3;
3717                 if ($opts{'-z'}) {
3718                         $res{'name'} = $4;
3719                 } else {
3720                         $res{'name'} = unquote($4);
3721                 }
3722         }
3723
3724         return wantarray ? %res : \%res;
3725 }
3726
3727 # generates _two_ hashes, references to which are passed as 2 and 3 argument
3728 sub parse_from_to_diffinfo {
3729         my ($diffinfo, $from, $to, @parents) = @_;
3730
3731         if ($diffinfo->{'nparents'}) {
3732                 # combined diff
3733                 $from->{'file'} = [];
3734                 $from->{'href'} = [];
3735                 fill_from_file_info($diffinfo, @parents)
3736                         unless exists $diffinfo->{'from_file'};
3737                 for (my $i = 0; $i < $diffinfo->{'nparents'}; $i++) {
3738                         $from->{'file'}[$i] =
3739                                 defined $diffinfo->{'from_file'}[$i] ?
3740                                         $diffinfo->{'from_file'}[$i] :
3741                                         $diffinfo->{'to_file'};
3742                         if ($diffinfo->{'status'}[$i] ne "A") { # not new (added) file
3743                                 $from->{'href'}[$i] = href(action=>"blob",
3744                                                            hash_base=>$parents[$i],
3745                                                            hash=>$diffinfo->{'from_id'}[$i],
3746                                                            file_name=>$from->{'file'}[$i]);
3747                         } else {
3748                                 $from->{'href'}[$i] = undef;
3749                         }
3750                 }
3751         } else {
3752                 # ordinary (not combined) diff
3753                 $from->{'file'} = $diffinfo->{'from_file'};
3754                 if ($diffinfo->{'status'} ne "A") { # not new (added) file
3755                         $from->{'href'} = href(action=>"blob", hash_base=>$hash_parent,
3756                                                hash=>$diffinfo->{'from_id'},
3757                                                file_name=>$from->{'file'});
3758                 } else {
3759                         delete $from->{'href'};
3760                 }
3761         }
3762
3763         $to->{'file'} = $diffinfo->{'to_file'};
3764         if (!is_deleted($diffinfo)) { # file exists in result
3765                 $to->{'href'} = href(action=>"blob", hash_base=>$hash,
3766                                      hash=>$diffinfo->{'to_id'},
3767                                      file_name=>$to->{'file'});
3768         } else {
3769                 delete $to->{'href'};
3770         }
3771 }
3772
3773 ## ......................................................................
3774 ## parse to array of hashes functions
3775
3776 sub git_get_heads_list {
3777         my ($limit, @classes) = @_;
3778         @classes = get_branch_refs() unless @classes;
3779         my @patterns = map { "refs/$_" } @classes;
3780         my @headslist;
3781
3782         open my $fd, '-|', git_cmd(), 'for-each-ref',
3783                 ($limit ? '--count='.($limit+1) : ()),
3784                 '--sort=-HEAD', '--sort=-committerdate',
3785                 '--format=%(objectname) %(refname) %(subject)%00%(committer)',
3786                 @patterns
3787                 or return;
3788         while (my $line = <$fd>) {
3789                 my %ref_item;
3790
3791                 chomp $line;
3792                 my ($refinfo, $committerinfo) = split(/\0/, $line);
3793                 my ($hash, $name, $title) = split(' ', $refinfo, 3);
3794                 my ($committer, $epoch, $tz) =
3795                         ($committerinfo =~ /^(.*) ([0-9]+) (.*)$/);
3796                 $ref_item{'fullname'}  = $name;
3797                 my $strip_refs = join '|', map { quotemeta } get_branch_refs();
3798                 $name =~ s!^refs/($strip_refs|remotes)/!!;
3799                 $ref_item{'name'} = $name;
3800                 # for refs neither in 'heads' nor 'remotes' we want to
3801                 # show their ref dir
3802                 my $ref_dir = (defined $1) ? $1 : '';
3803                 if ($ref_dir ne '' and $ref_dir ne 'heads' and $ref_dir ne 'remotes') {
3804                     $ref_item{'name'} .= ' (' . $ref_dir . ')';
3805                 }
3806
3807                 $ref_item{'id'}    = $hash;
3808                 $ref_item{'title'} = $title || '(no commit message)';
3809                 $ref_item{'epoch'} = $epoch;
3810                 if ($epoch) {
3811                         $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3812                 } else {
3813                         $ref_item{'age'} = "unknown";
3814                 }
3815
3816                 push @headslist, \%ref_item;
3817         }
3818         close $fd;
3819
3820         return wantarray ? @headslist : \@headslist;
3821 }
3822
3823 sub git_get_tags_list {
3824         my $limit = shift;
3825         my @tagslist;
3826
3827         open my $fd, '-|', git_cmd(), 'for-each-ref',
3828                 ($limit ? '--count='.($limit+1) : ()), '--sort=-creatordate',
3829                 '--format=%(objectname) %(objecttype) %(refname) '.
3830                 '%(*objectname) %(*objecttype) %(subject)%00%(creator)',
3831                 'refs/tags'
3832                 or return;
3833         while (my $line = <$fd>) {
3834                 my %ref_item;
3835
3836                 chomp $line;
3837                 my ($refinfo, $creatorinfo) = split(/\0/, $line);
3838                 my ($id, $type, $name, $refid, $reftype, $title) = split(' ', $refinfo, 6);
3839                 my ($creator, $epoch, $tz) =
3840                         ($creatorinfo =~ /^(.*) ([0-9]+) (.*)$/);
3841                 $ref_item{'fullname'} = $name;
3842                 $name =~ s!^refs/tags/!!;
3843
3844                 $ref_item{'type'} = $type;
3845                 $ref_item{'id'} = $id;
3846                 $ref_item{'name'} = $name;
3847                 if ($type eq "tag") {
3848                         $ref_item{'subject'} = $title;
3849                         $ref_item{'reftype'} = $reftype;
3850                         $ref_item{'refid'}   = $refid;
3851                 } else {
3852                         $ref_item{'reftype'} = $type;
3853                         $ref_item{'refid'}   = $id;
3854                 }
3855
3856                 if ($type eq "tag" || $type eq "commit") {
3857                         $ref_item{'epoch'} = $epoch;
3858                         if ($epoch) {
3859                                 $ref_item{'age'} = age_string(time - $ref_item{'epoch'});
3860                         } else {
3861                                 $ref_item{'age'} = "unknown";
3862                         }
3863                 }
3864
3865                 push @tagslist, \%ref_item;
3866         }
3867         close $fd;
3868
3869         return wantarray ? @tagslist : \@tagslist;
3870 }
3871
3872 ## ----------------------------------------------------------------------
3873 ## filesystem-related functions
3874
3875 sub get_file_owner {
3876         my $path = shift;
3877
3878         my ($dev, $ino, $mode, $nlink, $st_uid, $st_gid, $rdev, $size) = stat($path);
3879         my ($name, $passwd, $uid, $gid, $quota, $comment, $gcos, $dir, $shell) = getpwuid($st_uid);
3880         if (!defined $gcos) {
3881                 return undef;
3882         }
3883         my $owner = $gcos;
3884         $owner =~ s/[,;].*$//;
3885         return to_utf8($owner);
3886 }
3887
3888 # assume that file exists
3889 sub insert_file {
3890         my $filename = shift;
3891
3892         open my $fd, '<', $filename;
3893         print map { to_utf8($_) } <$fd>;
3894         close $fd;
3895 }
3896
3897 ## ......................................................................
3898 ## mimetype related functions
3899
3900 sub mimetype_guess_file {
3901         my $filename = shift;
3902         my $mimemap = shift;
3903         -r $mimemap or return undef;
3904
3905         my %mimemap;
3906         open(my $mh, '<', $mimemap) or return undef;
3907         while (<$mh>) {
3908                 next if m/^#/; # skip comments
3909                 my ($mimetype, @exts) = split(/\s+/);
3910                 foreach my $ext (@exts) {
3911                         $mimemap{$ext} = $mimetype;
3912                 }
3913         }
3914         close($mh);
3915
3916         $filename =~ /\.([^.]*)$/;
3917         return $mimemap{$1};
3918 }
3919
3920 sub mimetype_guess {
3921         my $filename = shift;
3922         my $mime;
3923         $filename =~ /\./ or return undef;
3924
3925         if ($mimetypes_file) {
3926                 my $file = $mimetypes_file;
3927                 if ($file !~ m!^/!) { # if it is relative path
3928                         # it is relative to project
3929                         $file = "$projectroot/$project/$file";
3930                 }
3931                 $mime = mimetype_guess_file($filename, $file);
3932         }
3933         $mime ||= mimetype_guess_file($filename, '/etc/mime.types');
3934         return $mime;
3935 }
3936
3937 sub blob_mimetype {
3938         my $fd = shift;
3939         my $filename = shift;
3940
3941         if ($filename) {
3942                 my $mime = mimetype_guess($filename);
3943                 $mime and return $mime;
3944         }
3945
3946         # just in case
3947         return $default_blob_plain_mimetype unless $fd;
3948
3949         if (-T $fd) {
3950                 return 'text/plain';
3951         } elsif (! $filename) {
3952                 return 'application/octet-stream';
3953         } elsif ($filename =~ m/\.png$/i) {
3954                 return 'image/png';
3955         } elsif ($filename =~ m/\.gif$/i) {
3956                 return 'image/gif';
3957         } elsif ($filename =~ m/\.jpe?g$/i) {
3958                 return 'image/jpeg';
3959         } else {
3960                 return 'application/octet-stream';
3961         }
3962 }
3963
3964 sub blob_contenttype {
3965         my ($fd, $file_name, $type) = @_;
3966
3967         $type ||= blob_mimetype($fd, $file_name);
3968         if ($type eq 'text/plain' && defined $default_text_plain_charset) {
3969                 $type .= "; charset=$default_text_plain_charset";
3970         }
3971
3972         return $type;
3973 }
3974
3975 # guess file syntax for syntax highlighting; return undef if no highlighting
3976 # the name of syntax can (in the future) depend on syntax highlighter used
3977 sub guess_file_syntax {
3978         my ($highlight, $file_name) = @_;
3979         return undef unless ($highlight && defined $file_name);
3980         my $basename = basename($file_name, '.in');
3981         return $highlight_basename{$basename}
3982                 if exists $highlight_basename{$basename};
3983
3984         $basename =~ /\.([^.]*)$/;
3985         my $ext = $1 or return undef;
3986         return $highlight_ext{$ext}
3987                 if exists $highlight_ext{$ext};
3988
3989         return undef;
3990 }
3991
3992 # run highlighter and return FD of its output,
3993 # or return original FD if no highlighting
3994 sub run_highlighter {
3995         my ($fd, $highlight, $syntax) = @_;
3996         return $fd unless ($highlight);
3997
3998         close $fd;
3999         my $syntax_arg = (defined $syntax) ? "--syntax $syntax" : "--force";
4000         open $fd, quote_command(git_cmd(), "cat-file", "blob", $hash)." | ".
4001                   quote_command($^X, '-CO', '-MEncode=decode,FB_DEFAULT', '-pse',
4002                     '$_ = decode($fe, $_, FB_DEFAULT) if !utf8::decode($_);',
4003                     '--', "-fe=$fallback_encoding")." | ".
4004                   quote_command($highlight_bin).
4005                   " --replace-tabs=8 --fragment $syntax_arg |"
4006                 or die_error(500, "Couldn't open file or run syntax highlighter");
4007         return $fd;
4008 }
4009
4010 ## ======================================================================
4011 ## functions printing HTML: header, footer, error page
4012
4013 sub get_page_title {
4014         my $title = to_utf8($site_name);
4015
4016         unless (defined $project) {
4017                 if (defined $project_filter) {
4018                         $title .= " - projects in '" . esc_path($project_filter) . "'";
4019                 }
4020                 return $title;
4021         }
4022         $title .= " - " . to_utf8($project);
4023
4024         return $title unless (defined $action);
4025         $title .= "/$action"; # $action is US-ASCII (7bit ASCII)
4026
4027         return $title unless (defined $file_name);
4028         $title .= " - " . esc_path($file_name);
4029         if ($action eq "tree" && $file_name !~ m|/$|) {
4030                 $title .= "/";
4031         }
4032
4033         return $title;
4034 }
4035
4036 sub get_content_type_html {
4037         # require explicit support from the UA if we are to send the page as
4038         # 'application/xhtml+xml', otherwise send it as plain old 'text/html'.
4039         # we have to do this because MSIE sometimes globs '*/*', pretending to
4040         # support xhtml+xml but choking when it gets what it asked for.
4041         if (defined $cgi->http('HTTP_ACCEPT') &&
4042             $cgi->http('HTTP_ACCEPT') =~ m/(,|;|\s|^)application\/xhtml\+xml(,|;|\s|$)/ &&
4043             $cgi->Accept('application/xhtml+xml') != 0) {
4044                 return 'application/xhtml+xml';
4045         } else {
4046                 return 'text/html';
4047         }
4048 }
4049
4050 sub print_feed_meta {
4051         if (defined $project) {
4052                 my %href_params = get_feed_info();
4053                 if (!exists $href_params{'-title'}) {
4054                         $href_params{'-title'} = 'log';
4055                 }
4056
4057                 foreach my $format (qw(RSS Atom)) {
4058                         my $type = lc($format);
4059                         my %link_attr = (
4060                                 '-rel' => 'alternate',
4061                                 '-title' => esc_attr("$project - $href_params{'-title'} - $format feed"),
4062                                 '-type' => "application/$type+xml"
4063                         );
4064
4065                         $href_params{'extra_options'} = undef;
4066                         $href_params{'action'} = $type;
4067                         $link_attr{'-href'} = esc_attr(href(%href_params));
4068                         print "<link ".
4069                               "rel=\"$link_attr{'-rel'}\" ".
4070                               "title=\"$link_attr{'-title'}\" ".
4071                               "href=\"$link_attr{'-href'}\" ".
4072                               "type=\"$link_attr{'-type'}\" ".
4073                               "/>\n";
4074
4075                         $href_params{'extra_options'} = '--no-merges';
4076                         $link_attr{'-href'} = esc_attr(href(%href_params));
4077                         $link_attr{'-title'} .= ' (no merges)';
4078                         print "<link ".
4079                               "rel=\"$link_attr{'-rel'}\" ".
4080                               "title=\"$link_attr{'-title'}\" ".
4081                               "href=\"$link_attr{'-href'}\" ".
4082                               "type=\"$link_attr{'-type'}\" ".
4083                               "/>\n";
4084                 }
4085
4086         } else {
4087                 printf('<link rel="alternate" title="%s projects list" '.
4088                        'href="%s" type="text/plain; charset=utf-8" />'."\n",
4089                        esc_attr($site_name),
4090                        esc_attr(href(project=>undef, action=>"project_index")));
4091                 printf('<link rel="alternate" title="%s projects feeds" '.
4092                        'href="%s" type="text/x-opml" />'."\n",
4093                        esc_attr($site_name),
4094                        esc_attr(href(project=>undef, action=>"opml")));
4095         }
4096 }
4097
4098 sub print_header_links {
4099         my $status = shift;
4100
4101         # print out each stylesheet that exist, providing backwards capability
4102         # for those people who defined $stylesheet in a config file
4103         if (defined $stylesheet) {
4104                 print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4105         } else {
4106                 foreach my $stylesheet (@stylesheets) {
4107                         next unless $stylesheet;
4108                         print '<link rel="stylesheet" type="text/css" href="'.esc_url($stylesheet).'"/>'."\n";
4109                 }
4110         }
4111         print_feed_meta()
4112                 if ($status eq '200 OK');
4113         if (defined $favicon) {
4114                 print qq(<link rel="shortcut icon" href=").esc_url($favicon).qq(" type="image/png" />\n);
4115         }
4116 }
4117
4118 sub print_nav_breadcrumbs_path {
4119         my $dirprefix = undef;
4120         while (my $part = shift) {
4121                 $dirprefix .= "/" if defined $dirprefix;
4122                 $dirprefix .= $part;
4123                 print $cgi->a({-href => href(project => undef,
4124                                              project_filter => $dirprefix,
4125                                              action => "project_list")},
4126                               esc_html($part)) . " / ";
4127         }
4128 }
4129
4130 sub print_nav_breadcrumbs {
4131         my %opts = @_;
4132
4133         for my $crumb (@extra_breadcrumbs, [ $home_link_str => $home_link ]) {
4134                 print $cgi->a({-href => esc_url($crumb->[1])}, $crumb->[0]) . " / ";
4135         }
4136         if (defined $project) {
4137                 my @dirname = split '/', $project;
4138                 my $projectbasename = pop @dirname;
4139                 print_nav_breadcrumbs_path(@dirname);
4140                 print $cgi->a({-href => href(action=>"summary")}, esc_html($projectbasename));
4141                 if (defined $action) {
4142                         my $action_print = $action ;
4143                         if (defined $opts{-action_extra}) {
4144                                 $action_print = $cgi->a({-href => href(action=>$action)},
4145                                         $action);
4146                         }
4147                         print " / $action_print";
4148                 }
4149                 if (defined $opts{-action_extra}) {
4150                         print " / $opts{-action_extra}";
4151                 }
4152                 print "\n";
4153         } elsif (defined $project_filter) {
4154                 print_nav_breadcrumbs_path(split '/', $project_filter);
4155         }
4156 }
4157
4158 sub print_search_form {
4159         if (!defined $searchtext) {
4160                 $searchtext = "";
4161         }
4162         my $search_hash;
4163         if (defined $hash_base) {
4164                 $search_hash = $hash_base;
4165         } elsif (defined $hash) {
4166                 $search_hash = $hash;
4167         } else {
4168                 $search_hash = "HEAD";
4169         }
4170         my $action = $my_uri;
4171         my $use_pathinfo = gitweb_check_feature('pathinfo');
4172         if ($use_pathinfo) {
4173                 $action .= "/".esc_url($project);
4174         }
4175         print $cgi->start_form(-method => "get", -action => $action) .
4176               "<div class=\"search\">\n" .
4177               (!$use_pathinfo &&
4178               $cgi->input({-name=>"p", -value=>$project, -type=>"hidden"}) . "\n") .
4179               $cgi->input({-name=>"a", -value=>"search", -type=>"hidden"}) . "\n" .
4180               $cgi->input({-name=>"h", -value=>$search_hash, -type=>"hidden"}) . "\n" .
4181               $cgi->popup_menu(-name => 'st', -default => 'commit',
4182                                -values => ['commit', 'grep', 'author', 'committer', 'pickaxe']) .
4183               " " . $cgi->a({-href => href(action=>"search_help"),
4184                              -title => "search help" }, "?") . " search:\n",
4185               $cgi->textfield(-name => "s", -value => $searchtext, -override => 1) . "\n" .
4186               "<span title=\"Extended regular expression\">" .
4187               $cgi->checkbox(-name => 'sr', -value => 1, -label => 're',
4188                              -checked => $search_use_regexp) .
4189               "</span>" .
4190               "</div>" .
4191               $cgi->end_form() . "\n";
4192 }
4193
4194 sub git_header_html {
4195         my $status = shift || "200 OK";
4196         my $expires = shift;
4197         my %opts = @_;
4198
4199         my $title = get_page_title();
4200         print $cgi->header(-type=>get_content_type_html(), -charset => 'utf-8',
4201                            -status=> $status, -expires => $expires)
4202                 unless ($opts{'-no_http_header'});
4203         my $mod_perl_version = $ENV{'MOD_PERL'} ? " $ENV{'MOD_PERL'}" : '';
4204         print <<EOF;
4205 <?xml version="1.0" encoding="utf-8"?>
4206 <!DOCTYPE html [
4207         <!ENTITY nbsp "&#xA0;">
4208         <!ENTITY sdot "&#x22C5;">
4209 ]>
4210 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en-US" lang="en-US">
4211 <!-- git web interface version $version, (C) 2005-2006, Kay Sievers <kay.sievers\@vrfy.org>, Christian Gierke -->
4212 <!-- git core binaries version $git_version -->
4213 <head>
4214 <meta name="generator" content="gitweb/$version git/$git_version$mod_perl_version"/>
4215 <meta name="robots" content="index, nofollow"/>
4216 <title>$title</title>
4217 EOF
4218         # the stylesheet, favicon etc urls won't work correctly with path_info
4219         # unless we set the appropriate base URL
4220         if ($ENV{'PATH_INFO'}) {
4221                 print "<base href=\"".esc_url($base_url)."\" />\n";
4222         }
4223         print_header_links($status);
4224
4225         if (defined $site_html_head_string) {
4226                 print to_utf8($site_html_head_string);
4227         }
4228
4229         print "</head>\n" .
4230               "<body>\n";
4231
4232         if (defined $site_header && -f $site_header) {
4233                 insert_file($site_header);
4234         }
4235
4236         print "<div class=\"page_header\">\n";
4237         if (defined $logo) {
4238                 print $cgi->a({-href => esc_url($logo_url),
4239                                -title => $logo_label},
4240                               $cgi->img({-src => esc_url($logo),
4241                                          -width => 72, -height => 27,
4242                                          -alt => "git",
4243                                          -class => "logo"}));
4244         }
4245         print_nav_breadcrumbs(%opts);
4246         print "</div>\n";
4247
4248         my $have_search = gitweb_check_feature('search');
4249         if (defined $project && $have_search) {
4250                 print_search_form();
4251         }
4252 }
4253
4254 sub git_footer_html {
4255         my $feed_class = 'rss_logo';
4256
4257         print "<div class=\"page_footer\">\n";
4258         if (defined $project) {
4259                 my $descr = git_get_project_description($project);
4260                 if (defined $descr) {
4261                         print "<div class=\"page_footer_text\">" . esc_html($descr) . "</div>\n";
4262                 }
4263
4264                 my %href_params = get_feed_info();
4265                 if (!%href_params) {
4266                         $feed_class .= ' generic';
4267                 }
4268                 $href_params{'-title'} ||= 'log';
4269
4270                 foreach my $format (qw(RSS Atom)) {
4271                         $href_params{'action'} = lc($format);
4272                         print $cgi->a({-href => href(%href_params),
4273                                       -title => "$href_params{'-title'} $format feed",
4274                                       -class => $feed_class}, $format)."\n";
4275                 }
4276
4277         } else {
4278                 print $cgi->a({-href => href(project=>undef, action=>"opml",
4279                                              project_filter => $project_filter),
4280                               -class => $feed_class}, "OPML") . " ";
4281                 print $cgi->a({-href => href(project=>undef, action=>"project_index",
4282                                              project_filter => $project_filter),
4283                               -class => $feed_class}, "TXT") . "\n";
4284         }
4285         print "</div>\n"; # class="page_footer"
4286
4287         if (defined $t0 && gitweb_check_feature('timed')) {
4288                 print "<div id=\"generating_info\">\n";
4289                 print 'This page took '.
4290                       '<span id="generating_time" class="time_span">'.
4291                       tv_interval($t0, [ gettimeofday() ]).
4292                       ' seconds </span>'.
4293                       ' and '.
4294                       '<span id="generating_cmd">'.
4295                       $number_of_git_cmds.
4296                       '</span> git commands '.
4297                       " to generate.\n";
4298                 print "</div>\n"; # class="page_footer"
4299         }
4300
4301         if (defined $site_footer && -f $site_footer) {
4302                 insert_file($site_footer);
4303         }
4304
4305         print qq!<script type="text/javascript" src="!.esc_url($javascript).qq!"></script>\n!;
4306         if (defined $action &&
4307             $action eq 'blame_incremental') {
4308                 print qq!<script type="text/javascript">\n!.
4309                       qq!startBlame("!. esc_attr(href(action=>"blame_data", -replay=>1)) .qq!",\n!.
4310                       qq!           "!. esc_attr(href()) .qq!");\n!.
4311                       qq!</script>\n!;
4312         } else {
4313                 my ($jstimezone, $tz_cookie, $datetime_class) =
4314                         gitweb_get_feature('javascript-timezone');
4315
4316                 print qq!<script type="text/javascript">\n!.
4317                       qq!window.onload = function () {\n!;
4318                 if (gitweb_check_feature('javascript-actions')) {
4319                         print qq!       fixLinks();\n!;
4320                 }
4321                 if ($jstimezone && $tz_cookie && $datetime_class) {
4322                         print qq!       var tz_cookie = { name: '$tz_cookie', expires: 14, path: '/' };\n!. # in days
4323                               qq!       onloadTZSetup('$jstimezone', tz_cookie, '$datetime_class');\n!;
4324                 }
4325                 print qq!};\n!.
4326                       qq!</script>\n!;
4327         }
4328
4329         print "</body>\n" .
4330               "</html>";
4331 }
4332
4333 # die_error(<http_status_code>, <error_message>[, <detailed_html_description>])
4334 # Example: die_error(404, 'Hash not found')
4335 # By convention, use the following status codes (as defined in RFC 2616):
4336 # 400: Invalid or missing CGI parameters, or
4337 #      requested object exists but has wrong type.
4338 # 403: Requested feature (like "pickaxe" or "snapshot") not enabled on
4339 #      this server or project.
4340 # 404: Requested object/revision/project doesn't exist.
4341 # 500: The server isn't configured properly, or
4342 #      an internal error occurred (e.g. failed assertions caused by bugs), or
4343 #      an unknown error occurred (e.g. the git binary died unexpectedly).
4344 # 503: The server is currently unavailable (because it is overloaded,
4345 #      or down for maintenance).  Generally, this is a temporary state.
4346 sub die_error {
4347         my $status = shift || 500;
4348         my $error = esc_html(shift) || "Internal Server Error";
4349         my $extra = shift;
4350         my %opts = @_;
4351
4352         my %http_responses = (
4353                 400 => '400 Bad Request',
4354                 403 => '403 Forbidden',
4355                 404 => '404 Not Found',
4356                 500 => '500 Internal Server Error',
4357                 503 => '503 Service Unavailable',
4358         );
4359         git_header_html($http_responses{$status}, undef, %opts);
4360         print <<EOF;
4361 <div class="page_body">
4362 <br /><br />
4363 $status - $error
4364 <br />
4365 EOF
4366         if (defined $extra) {
4367                 print "<hr />\n" .
4368                       "$extra\n";
4369         }
4370         print "</div>\n";
4371
4372         git_footer_html();
4373         goto DONE_GITWEB
4374                 unless ($opts{'-error_handler'});
4375 }
4376
4377 ## ----------------------------------------------------------------------
4378 ## functions printing or outputting HTML: navigation
4379
4380 sub git_print_page_nav {
4381         my ($current, $suppress, $head, $treehead, $treebase, $extra) = @_;
4382         $extra = '' if !defined $extra; # pager or formats
4383
4384         my @navs = qw(summary shortlog log commit commitdiff tree);
4385         if ($suppress) {
4386                 @navs = grep { $_ ne $suppress } @navs;
4387         }
4388
4389         my %arg = map { $_ => {action=>$_} } @navs;
4390         if (defined $head) {
4391                 for (qw(commit commitdiff)) {
4392                         $arg{$_}{'hash'} = $head;
4393                 }
4394                 if ($current =~ m/^(tree | log | shortlog | commit | commitdiff | search)$/x) {
4395                         for (qw(shortlog log)) {
4396                                 $arg{$_}{'hash'} = $head;
4397                         }
4398                 }
4399         }
4400
4401         $arg{'tree'}{'hash'} = $treehead if defined $treehead;
4402         $arg{'tree'}{'hash_base'} = $treebase if defined $treebase;
4403
4404         my @actions = gitweb_get_feature('actions');
4405         my %repl = (
4406                 '%' => '%',
4407                 'n' => $project,         # project name
4408                 'f' => $git_dir,         # project path within filesystem
4409                 'h' => $treehead || '',  # current hash ('h' parameter)
4410                 'b' => $treebase || '',  # hash base ('hb' parameter)
4411         );
4412         while (@actions) {
4413                 my ($label, $link, $pos) = splice(@actions,0,3);
4414                 # insert
4415                 @navs = map { $_ eq $pos ? ($_, $label) : $_ } @navs;
4416                 # munch munch
4417                 $link =~ s/%([%nfhb])/$repl{$1}/g;
4418                 $arg{$label}{'_href'} = $link;
4419         }
4420
4421         print "<div class=\"page_nav\">\n" .
4422                 (join " | ",
4423                  map { $_ eq $current ?
4424                        $_ : $cgi->a({-href => ($arg{$_}{_href} ? $arg{$_}{_href} : href(%{$arg{$_}}))}, "$_")
4425                  } @navs);
4426         print "<br/>\n$extra<br/>\n" .
4427               "</div>\n";
4428 }
4429
4430 # returns a submenu for the navigation of the refs views (tags, heads,
4431 # remotes) with the current view disabled and the remotes view only
4432 # available if the feature is enabled
4433 sub format_ref_views {
4434         my ($current) = @_;
4435         my @ref_views = qw{tags heads};
4436         push @ref_views, 'remotes' if gitweb_check_feature('remote_heads');
4437         return join " | ", map {
4438                 $_ eq $current ? $_ :
4439                 $cgi->a({-href => href(action=>$_)}, $_)
4440         } @ref_views
4441 }
4442
4443 sub format_paging_nav {
4444         my ($action, $page, $has_next_link) = @_;
4445         my $paging_nav;
4446
4447
4448         if ($page > 0) {
4449                 $paging_nav .=
4450                         $cgi->a({-href => href(-replay=>1, page=>undef)}, "first") .
4451                         " &sdot; " .
4452                         $cgi->a({-href => href(-replay=>1, page=>$page-1),
4453                                  -accesskey => "p", -title => "Alt-p"}, "prev");
4454         } else {
4455                 $paging_nav .= "first &sdot; prev";
4456         }
4457
4458         if ($has_next_link) {
4459                 $paging_nav .= " &sdot; " .
4460                         $cgi->a({-href => href(-replay=>1, page=>$page+1),
4461                                  -accesskey => "n", -title => "Alt-n"}, "next");
4462         } else {
4463                 $paging_nav .= " &sdot; next";
4464         }
4465
4466         return $paging_nav;
4467 }
4468
4469 ## ......................................................................
4470 ## functions printing or outputting HTML: div
4471
4472 sub git_print_header_div {
4473         my ($action, $title, $hash, $hash_base) = @_;
4474         my %args = ();
4475
4476         $args{'action'} = $action;
4477         $args{'hash'} = $hash if $hash;
4478         $args{'hash_base'} = $hash_base if $hash_base;
4479
4480         print "<div class=\"header\">\n" .
4481               $cgi->a({-href => href(%args), -class => "title"},
4482               $title ? $title : $action) .
4483               "\n</div>\n";
4484 }
4485
4486 sub format_repo_url {
4487         my ($name, $url) = @_;
4488         return "<tr class=\"metadata_url\"><td>$name</td><td>$url</td></tr>\n";
4489 }
4490
4491 # Group output by placing it in a DIV element and adding a header.
4492 # Options for start_div() can be provided by passing a hash reference as the
4493 # first parameter to the function.
4494 # Options to git_print_header_div() can be provided by passing an array
4495 # reference. This must follow the options to start_div if they are present.
4496 # The content can be a scalar, which is output as-is, a scalar reference, which
4497 # is output after html escaping, an IO handle passed either as *handle or
4498 # *handle{IO}, or a function reference. In the latter case all following
4499 # parameters will be taken as argument to the content function call.
4500 sub git_print_section {
4501         my ($div_args, $header_args, $content);
4502         my $arg = shift;
4503         if (ref($arg) eq 'HASH') {
4504                 $div_args = $arg;
4505                 $arg = shift;
4506         }
4507         if (ref($arg) eq 'ARRAY') {
4508                 $header_args = $arg;
4509                 $arg = shift;
4510         }
4511         $content = $arg;
4512
4513         print $cgi->start_div($div_args);
4514         git_print_header_div(@$header_args);
4515
4516         if (ref($content) eq 'CODE') {
4517                 $content->(@_);
4518         } elsif (ref($content) eq 'SCALAR') {
4519                 print esc_html($$content);
4520         } elsif (ref($content) eq 'GLOB' or ref($content) eq 'IO::Handle') {
4521                 print <$content>;
4522         } elsif (!ref($content) && defined($content)) {
4523                 print $content;
4524         }
4525
4526         print $cgi->end_div;
4527 }
4528
4529 sub format_timestamp_html {
4530         my $date = shift;
4531         my $strtime = $date->{'rfc2822'};
4532
4533         my (undef, undef, $datetime_class) =
4534                 gitweb_get_feature('javascript-timezone');
4535         if ($datetime_class) {
4536                 $strtime = qq!<span class="$datetime_class">$strtime</span>!;
4537         }
4538
4539         my $localtime_format = '(%02d:%02d %s)';
4540         if ($date->{'hour_local'} < 6) {
4541                 $localtime_format = '(<span class="atnight">%02d:%02d</span> %s)';
4542         }
4543         $strtime .= ' ' .
4544                     sprintf($localtime_format,
4545                             $date->{'hour_local'}, $date->{'minute_local'}, $date->{'tz_local'});
4546
4547         return $strtime;
4548 }
4549
4550 # Outputs the author name and date in long form
4551 sub git_print_authorship {
4552         my $co = shift;
4553         my %opts = @_;
4554         my $tag = $opts{-tag} || 'div';
4555         my $author = $co->{'author_name'};
4556
4557         my %ad = parse_date($co->{'author_epoch'}, $co->{'author_tz'});
4558         print "<$tag class=\"author_date\">" .
4559               format_search_author($author, "author", esc_html($author)) .
4560               " [".format_timestamp_html(\%ad)."]".
4561               git_get_avatar($co->{'author_email'}, -pad_before => 1) .
4562               "</$tag>\n";
4563 }
4564
4565 # Outputs table rows containing the full author or committer information,
4566 # in the format expected for 'commit' view (& similar).
4567 # Parameters are a commit hash reference, followed by the list of people
4568 # to output information for. If the list is empty it defaults to both
4569 # author and committer.
4570 sub git_print_authorship_rows {
4571         my $co = shift;
4572         # too bad we can't use @people = @_ || ('author', 'committer')
4573         my @people = @_;
4574         @people = ('author', 'committer') unless @people;
4575         foreach my $who (@people) {
4576                 my %wd = parse_date($co->{"${who}_epoch"}, $co->{"${who}_tz"});
4577                 print "<tr><td>$who</td><td>" .
4578                       format_search_author($co->{"${who}_name"}, $who,
4579                                            esc_html($co->{"${who}_name"})) . " " .
4580                       format_search_author($co->{"${who}_email"}, $who,
4581                                            esc_html("<" . $co->{"${who}_email"} . ">")) .
4582                       "</td><td rowspan=\"2\">" .
4583                       git_get_avatar($co->{"${who}_email"}, -size => 'double') .
4584                       "</td></tr>\n" .
4585                       "<tr>" .
4586                       "<td></td><td>" .
4587                       format_timestamp_html(\%wd) .
4588                       "</td>" .
4589                       "</tr>\n";
4590         }
4591 }
4592
4593 sub git_print_page_path {
4594         my $name = shift;
4595         my $type = shift;
4596         my $hb = shift;
4597
4598
4599         print "<div class=\"page_path\">";
4600         print $cgi->a({-href => href(action=>"tree", hash_base=>$hb),
4601                       -title => 'tree root'}, to_utf8("[$project]"));
4602         print " / ";
4603         if (defined $name) {
4604                 my @dirname = split '/', $name;
4605                 my $basename = pop @dirname;
4606                 my $fullname = '';
4607
4608                 foreach my $dir (@dirname) {
4609                         $fullname .= ($fullname ? '/' : '') . $dir;
4610                         print $cgi->a({-href => href(action=>"tree", file_name=>$fullname,
4611                                                      hash_base=>$hb),
4612                                       -title => $fullname}, esc_path($dir));
4613                         print " / ";
4614                 }
4615                 if (defined $type && $type eq 'blob') {
4616                         print $cgi->a({-href => href(action=>"blob_plain", file_name=>$file_name,
4617                                                      hash_base=>$hb),
4618                                       -title => $name}, esc_path($basename));
4619                 } elsif (defined $type && $type eq 'tree') {
4620                         print $cgi->a({-href => href(action=>"tree", file_name=>$file_name,
4621                                                      hash_base=>$hb),
4622                                       -title => $name}, esc_path($basename));
4623                         print " / ";
4624                 } else {
4625                         print esc_path($basename);
4626                 }
4627         }
4628         print "<br/></div>\n";
4629 }
4630
4631 sub git_print_log {
4632         my $log = shift;
4633         my %opts = @_;
4634
4635         if ($opts{'-remove_title'}) {
4636                 # remove title, i.e. first line of log
4637                 shift @$log;
4638         }
4639         # remove leading empty lines
4640         while (defined $log->[0] && $log->[0] eq "") {
4641                 shift @$log;
4642         }
4643
4644         # print log
4645         my $skip_blank_line = 0;
4646         foreach my $line (@$log) {
4647                 if ($line =~ m/^\s*([A-Z][-A-Za-z]*-([Bb]y|[Tt]o)|C[Cc]|(Clos|Fix)es): /) {
4648                         if (! $opts{'-remove_signoff'}) {
4649                                 print "<span class=\"signoff\">" . esc_html($line) . "</span><br/>\n";
4650                                 $skip_blank_line = 1;
4651                         }
4652                         next;
4653                 }
4654
4655                 if ($line =~ m,\s*([a-z]*link): (https?://\S+),i) {
4656                         if (! $opts{'-remove_signoff'}) {
4657                                 print "<span class=\"signoff\">" . esc_html($1) . ": " .
4658                                         "<a href=\"" . esc_html($2) . "\">" . esc_html($2) . "</a>" .
4659                                         "</span><br/>\n";
4660                                 $skip_blank_line = 1;
4661                         }
4662                         next;
4663                 }
4664
4665                 # print only one empty line
4666                 # do not print empty line after signoff
4667                 if ($line eq "") {
4668                         next if ($skip_blank_line);
4669                         $skip_blank_line = 1;
4670                 } else {
4671                         $skip_blank_line = 0;
4672                 }
4673
4674                 print format_log_line_html($line) . "<br/>\n";
4675         }
4676
4677         if ($opts{'-final_empty_line'}) {
4678                 # end with single empty line
4679                 print "<br/>\n" unless $skip_blank_line;
4680         }
4681 }
4682
4683 # return link target (what link points to)
4684 sub git_get_link_target {
4685         my $hash = shift;
4686         my $link_target;
4687
4688         # read link
4689         open my $fd, "-|", git_cmd(), "cat-file", "blob", $hash
4690                 or return;
4691         {
4692                 local $/ = undef;
4693                 $link_target = <$fd>;
4694         }
4695         close $fd
4696                 or return;
4697
4698         return $link_target;
4699 }
4700
4701 # given link target, and the directory (basedir) the link is in,
4702 # return target of link relative to top directory (top tree);
4703 # return undef if it is not possible (including absolute links).
4704 sub normalize_link_target {
4705         my ($link_target, $basedir) = @_;
4706
4707         # absolute symlinks (beginning with '/') cannot be normalized
4708         return if (substr($link_target, 0, 1) eq '/');
4709
4710         # normalize link target to path from top (root) tree (dir)
4711         my $path;
4712         if ($basedir) {
4713                 $path = $basedir . '/' . $link_target;
4714         } else {
4715                 # we are in top (root) tree (dir)
4716                 $path = $link_target;
4717         }
4718
4719         # remove //, /./, and /../
4720         my @path_parts;
4721         foreach my $part (split('/', $path)) {
4722                 # discard '.' and ''
4723                 next if (!$part || $part eq '.');
4724                 # handle '..'
4725                 if ($part eq '..') {
4726                         if (@path_parts) {
4727                                 pop @path_parts;
4728                         } else {
4729                                 # link leads outside repository (outside top dir)
4730                                 return;
4731                         }
4732                 } else {
4733                         push @path_parts, $part;
4734                 }
4735         }
4736         $path = join('/', @path_parts);
4737
4738         return $path;
4739 }
4740
4741 # print tree entry (row of git_tree), but without encompassing <tr> element
4742 sub git_print_tree_entry {
4743         my ($t, $basedir, $hash_base, $have_blame) = @_;
4744
4745         my %base_key = ();
4746         $base_key{'hash_base'} = $hash_base if defined $hash_base;
4747
4748         # The format of a table row is: mode list link.  Where mode is
4749         # the mode of the entry, list is the name of the entry, an href,
4750         # and link is the action links of the entry.
4751
4752         print "<td class=\"mode\">" . mode_str($t->{'mode'}) . "</td>\n";
4753         if (exists $t->{'size'}) {
4754                 print "<td class=\"size\">$t->{'size'}</td>\n";
4755         }
4756         if ($t->{'type'} eq "blob") {
4757                 print "<td class=\"list\">" .
4758                         $cgi->a({-href => href(action=>"blob", hash=>$t->{'hash'},
4759                                                file_name=>"$basedir$t->{'name'}", %base_key),
4760                                 -class => "list"}, esc_path($t->{'name'}));
4761                 if (S_ISLNK(oct $t->{'mode'})) {
4762                         my $link_target = git_get_link_target($t->{'hash'});
4763                         if ($link_target) {
4764                                 my $norm_target = normalize_link_target($link_target, $basedir);
4765                                 if (defined $norm_target) {
4766                                         print " -> " .
4767                                               $cgi->a({-href => href(action=>"object", hash_base=>$hash_base,
4768                                                                      file_name=>$norm_target),
4769                                                        -title => $norm_target}, esc_path($link_target));
4770                                 } else {
4771                                         print " -> " . esc_path($link_target);
4772                                 }
4773                         }
4774                 }
4775                 print "</td>\n";
4776                 print "<td class=\"link\">";
4777                 print $cgi->a({-href => href(action=>"blob", hash=>$t->{'hash'},
4778                                              file_name=>"$basedir$t->{'name'}", %base_key)},
4779                               "blob");
4780                 if ($have_blame) {
4781                         print " | " .
4782                               $cgi->a({-href => href(action=>"blame", hash=>$t->{'hash'},
4783                                                      file_name=>"$basedir$t->{'name'}", %base_key)},
4784                                       "blame");
4785                 }
4786                 if (defined $hash_base) {
4787                         print " | " .
4788                               $cgi->a({-href => href(action=>"history", hash_base=>$hash_base,
4789                                                      hash=>$t->{'hash'}, file_name=>"$basedir$t->{'name'}")},
4790                                       "history");
4791                 }
4792                 print " | " .
4793                         $cgi->a({-href => href(action=>"blob_plain", hash_base=>$hash_base,
4794                                                file_name=>"$basedir$t->{'name'}")},
4795                                 "raw");
4796                 print "</td>\n";
4797
4798         } elsif ($t->{'type'} eq "tree") {
4799                 print "<td class=\"list\">";
4800                 print $cgi->a({-href => href(action=>"tree", hash=>$t->{'hash'},
4801                                              file_name=>"$basedir$t->{'name'}",
4802                                              %base_key)},
4803                               esc_path($t->{'name'}));
4804                 print "</td>\n";
4805                 print "<td class=\"link\">";
4806                 print $cgi->a({-href => href(action=>"tree", hash=>$t->{'hash'},
4807                                              file_name=>"$basedir$t->{'name'}",
4808                                              %base_key)},
4809                               "tree");
4810                 if (defined $hash_base) {
4811                         print " | " .
4812                               $cgi->a({-href => href(action=>"history", hash_base=>$hash_base,
4813                                                      file_name=>"$basedir$t->{'name'}")},
4814                                       "history");
4815                 }
4816                 print "</td>\n";
4817         } else {
4818                 # unknown object: we can only present history for it
4819                 # (this includes 'commit' object, i.e. submodule support)
4820                 print "<td class=\"list\">" .
4821                       esc_path($t->{'name'}) .
4822                       "</td>\n";
4823                 print "<td class=\"link\">";
4824                 if (defined $hash_base) {
4825                         print $cgi->a({-href => href(action=>"history",
4826                                                      hash_base=>$hash_base,
4827                                                      file_name=>"$basedir$t->{'name'}")},
4828                                       "history");
4829                 }
4830                 print "</td>\n";
4831         }
4832 }
4833
4834 ## ......................................................................
4835 ## functions printing large fragments of HTML
4836
4837 # get pre-image filenames for merge (combined) diff
4838 sub fill_from_file_info {
4839         my ($diff, @parents) = @_;
4840
4841         $diff->{'from_file'} = [ ];
4842         $diff->{'from_file'}[$diff->{'nparents'} - 1] = undef;
4843         for (my $i = 0; $i < $diff->{'nparents'}; $i++) {
4844                 if ($diff->{'status'}[$i] eq 'R' ||
4845                     $diff->{'status'}[$i] eq 'C') {
4846                         $diff->{'from_file'}[$i] =
4847                                 git_get_path_by_hash($parents[$i], $diff->{'from_id'}[$i]);
4848                 }
4849         }
4850
4851         return $diff;
4852 }
4853
4854 # is current raw difftree line of file deletion
4855 sub is_deleted {
4856         my $diffinfo = shift;
4857
4858         return $diffinfo->{'to_id'} eq ('0' x 40) || $diffinfo->{'to_id'} eq ('0' x 64);
4859 }
4860
4861 # does patch correspond to [previous] difftree raw line
4862 # $diffinfo  - hashref of parsed raw diff format
4863 # $patchinfo - hashref of parsed patch diff format
4864 #              (the same keys as in $diffinfo)
4865 sub is_patch_split {
4866         my ($diffinfo, $patchinfo) = @_;
4867
4868         return defined $diffinfo && defined $patchinfo
4869                 && $diffinfo->{'to_file'} eq $patchinfo->{'to_file'};
4870 }
4871
4872
4873 sub git_difftree_body {
4874         my ($difftree, $hash, @parents) = @_;
4875         my ($parent) = $parents[0];
4876         my $have_blame = gitweb_check_feature('blame');
4877         print "<div class=\"list_head\">\n";
4878         if ($#{$difftree} > 10) {
4879                 print(($#{$difftree} + 1) . " files changed:\n");
4880         }
4881         print "</div>\n";
4882
4883         print "<table class=\"" .
4884               (@parents > 1 ? "combined " : "") .
4885               "diff_tree\">\n";
4886
4887         # header only for combined diff in 'commitdiff' view
4888         my $has_header = @$difftree && @parents > 1 && $action eq 'commitdiff';
4889         if ($has_header) {
4890                 # table header
4891                 print "<thead><tr>\n" .
4892                        "<th></th><th></th>\n"; # filename, patchN link
4893                 for (my $i = 0; $i < @parents; $i++) {
4894                         my $par = $parents[$i];
4895                         print "<th>" .
4896                               $cgi->a({-href => href(action=>"commitdiff",
4897                                                      hash=>$hash, hash_parent=>$par),
4898                                        -title => 'commitdiff to parent number ' .
4899                                                   ($i+1) . ': ' . substr($par,0,7)},
4900                                       $i+1) .
4901                               "&nbsp;</th>\n";
4902                 }
4903                 print "</tr></thead>\n<tbody>\n";
4904         }
4905
4906         my $alternate = 1;
4907         my $patchno = 0;
4908         foreach my $line (@{$difftree}) {
4909                 my $diff = parsed_difftree_line($line);
4910
4911                 if ($alternate) {
4912                         print "<tr class=\"dark\">\n";
4913                 } else {
4914                         print "<tr class=\"light\">\n";
4915                 }
4916                 $alternate ^= 1;
4917
4918                 if (exists $diff->{'nparents'}) { # combined diff
4919
4920                         fill_from_file_info($diff, @parents)
4921                                 unless exists $diff->{'from_file'};
4922
4923                         if (!is_deleted($diff)) {
4924                                 # file exists in the result (child) commit
4925                                 print "<td>" .
4926                                       $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
4927                                                              file_name=>$diff->{'to_file'},
4928                                                              hash_base=>$hash),
4929                                               -class => "list"}, esc_path($diff->{'to_file'})) .
4930                                       "</td>\n";
4931                         } else {
4932                                 print "<td>" .
4933                                       esc_path($diff->{'to_file'}) .
4934                                       "</td>\n";
4935                         }
4936
4937                         if ($action eq 'commitdiff') {
4938                                 # link to patch
4939                                 $patchno++;
4940                                 print "<td class=\"link\">" .
4941                                       $cgi->a({-href => href(-anchor=>"patch$patchno")},
4942                                               "patch") .
4943                                       " | " .
4944                                       "</td>\n";
4945                         }
4946
4947                         my $has_history = 0;
4948                         my $not_deleted = 0;
4949                         for (my $i = 0; $i < $diff->{'nparents'}; $i++) {
4950                                 my $hash_parent = $parents[$i];
4951                                 my $from_hash = $diff->{'from_id'}[$i];
4952                                 my $from_path = $diff->{'from_file'}[$i];
4953                                 my $status = $diff->{'status'}[$i];
4954
4955                                 $has_history ||= ($status ne 'A');
4956                                 $not_deleted ||= ($status ne 'D');
4957
4958                                 if ($status eq 'A') {
4959                                         print "<td  class=\"link\" align=\"right\"> | </td>\n";
4960                                 } elsif ($status eq 'D') {
4961                                         print "<td class=\"link\">" .
4962                                               $cgi->a({-href => href(action=>"blob",
4963                                                                      hash_base=>$hash,
4964                                                                      hash=>$from_hash,
4965                                                                      file_name=>$from_path)},
4966                                                       "blob" . ($i+1)) .
4967                                               " | </td>\n";
4968                                 } else {
4969                                         if ($diff->{'to_id'} eq $from_hash) {
4970                                                 print "<td class=\"link nochange\">";
4971                                         } else {
4972                                                 print "<td class=\"link\">";
4973                                         }
4974                                         print $cgi->a({-href => href(action=>"blobdiff",
4975                                                                      hash=>$diff->{'to_id'},
4976                                                                      hash_parent=>$from_hash,
4977                                                                      hash_base=>$hash,
4978                                                                      hash_parent_base=>$hash_parent,
4979                                                                      file_name=>$diff->{'to_file'},
4980                                                                      file_parent=>$from_path)},
4981                                                       "diff" . ($i+1)) .
4982                                               " | </td>\n";
4983                                 }
4984                         }
4985
4986                         print "<td class=\"link\">";
4987                         if ($not_deleted) {
4988                                 print $cgi->a({-href => href(action=>"blob",
4989                                                              hash=>$diff->{'to_id'},
4990                                                              file_name=>$diff->{'to_file'},
4991                                                              hash_base=>$hash)},
4992                                               "blob");
4993                                 print " | " if ($has_history);
4994                         }
4995                         if ($has_history) {
4996                                 print $cgi->a({-href => href(action=>"history",
4997                                                              file_name=>$diff->{'to_file'},
4998                                                              hash_base=>$hash)},
4999                                               "history");
5000                         }
5001                         print "</td>\n";
5002
5003                         print "</tr>\n";
5004                         next; # instead of 'else' clause, to avoid extra indent
5005                 }
5006                 # else ordinary diff
5007
5008                 my ($to_mode_oct, $to_mode_str, $to_file_type);
5009                 my ($from_mode_oct, $from_mode_str, $from_file_type);
5010                 if ($diff->{'to_mode'} ne ('0' x 6)) {
5011                         $to_mode_oct = oct $diff->{'to_mode'};
5012                         if (S_ISREG($to_mode_oct)) { # only for regular file
5013                                 $to_mode_str = sprintf("%04o", $to_mode_oct & 0777); # permission bits
5014                         }
5015                         $to_file_type = file_type($diff->{'to_mode'});
5016                 }
5017                 if ($diff->{'from_mode'} ne ('0' x 6)) {
5018                         $from_mode_oct = oct $diff->{'from_mode'};
5019                         if (S_ISREG($from_mode_oct)) { # only for regular file
5020                                 $from_mode_str = sprintf("%04o", $from_mode_oct & 0777); # permission bits
5021                         }
5022                         $from_file_type = file_type($diff->{'from_mode'});
5023                 }
5024
5025                 if ($diff->{'status'} eq "A") { # created
5026                         my $mode_chng = "<span class=\"file_status new\">[new $to_file_type";
5027                         $mode_chng   .= " with mode: $to_mode_str" if $to_mode_str;
5028                         $mode_chng   .= "]</span>";
5029                         print "<td>";
5030                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5031                                                      hash_base=>$hash, file_name=>$diff->{'file'}),
5032                                       -class => "list"}, esc_path($diff->{'file'}));
5033                         print "</td>\n";
5034                         print "<td>$mode_chng</td>\n";
5035                         print "<td class=\"link\">";
5036                         if ($action eq 'commitdiff') {
5037                                 # link to patch
5038                                 $patchno++;
5039                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5040                                               "patch") .
5041                                       " | ";
5042                         }
5043                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5044                                                      hash_base=>$hash, file_name=>$diff->{'file'})},
5045                                       "blob");
5046                         print "</td>\n";
5047
5048                 } elsif ($diff->{'status'} eq "D") { # deleted
5049                         my $mode_chng = "<span class=\"file_status deleted\">[deleted $from_file_type]</span>";
5050                         print "<td>";
5051                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'from_id'},
5052                                                      hash_base=>$parent, file_name=>$diff->{'file'}),
5053                                        -class => "list"}, esc_path($diff->{'file'}));
5054                         print "</td>\n";
5055                         print "<td>$mode_chng</td>\n";
5056                         print "<td class=\"link\">";
5057                         if ($action eq 'commitdiff') {
5058                                 # link to patch
5059                                 $patchno++;
5060                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5061                                               "patch") .
5062                                       " | ";
5063                         }
5064                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'from_id'},
5065                                                      hash_base=>$parent, file_name=>$diff->{'file'})},
5066                                       "blob") . " | ";
5067                         if ($have_blame) {
5068                                 print $cgi->a({-href => href(action=>"blame", hash_base=>$parent,
5069                                                              file_name=>$diff->{'file'})},
5070                                               "blame") . " | ";
5071                         }
5072                         print $cgi->a({-href => href(action=>"history", hash_base=>$parent,
5073                                                      file_name=>$diff->{'file'})},
5074                                       "history");
5075                         print "</td>\n";
5076
5077                 } elsif ($diff->{'status'} eq "M" || $diff->{'status'} eq "T") { # modified, or type changed
5078                         my $mode_chnge = "";
5079                         if ($diff->{'from_mode'} != $diff->{'to_mode'}) {
5080                                 $mode_chnge = "<span class=\"file_status mode_chnge\">[changed";
5081                                 if ($from_file_type ne $to_file_type) {
5082                                         $mode_chnge .= " from $from_file_type to $to_file_type";
5083                                 }
5084                                 if (($from_mode_oct & 0777) != ($to_mode_oct & 0777)) {
5085                                         if ($from_mode_str && $to_mode_str) {
5086                                                 $mode_chnge .= " mode: $from_mode_str->$to_mode_str";
5087                                         } elsif ($to_mode_str) {
5088                                                 $mode_chnge .= " mode: $to_mode_str";
5089                                         }
5090                                 }
5091                                 $mode_chnge .= "]</span>\n";
5092                         }
5093                         print "<td>";
5094                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5095                                                      hash_base=>$hash, file_name=>$diff->{'file'}),
5096                                       -class => "list"}, esc_path($diff->{'file'}));
5097                         print "</td>\n";
5098                         print "<td>$mode_chnge</td>\n";
5099                         print "<td class=\"link\">";
5100                         if ($action eq 'commitdiff') {
5101                                 # link to patch
5102                                 $patchno++;
5103                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5104                                               "patch") .
5105                                       " | ";
5106                         } elsif ($diff->{'to_id'} ne $diff->{'from_id'}) {
5107                                 # "commit" view and modified file (not onlu mode changed)
5108                                 print $cgi->a({-href => href(action=>"blobdiff",
5109                                                              hash=>$diff->{'to_id'}, hash_parent=>$diff->{'from_id'},
5110                                                              hash_base=>$hash, hash_parent_base=>$parent,
5111                                                              file_name=>$diff->{'file'})},
5112                                               "diff") .
5113                                       " | ";
5114                         }
5115                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5116                                                      hash_base=>$hash, file_name=>$diff->{'file'})},
5117                                        "blob") . " | ";
5118                         if ($have_blame) {
5119                                 print $cgi->a({-href => href(action=>"blame", hash_base=>$hash,
5120                                                              file_name=>$diff->{'file'})},
5121                                               "blame") . " | ";
5122                         }
5123                         print $cgi->a({-href => href(action=>"history", hash_base=>$hash,
5124                                                      file_name=>$diff->{'file'})},
5125                                       "history");
5126                         print "</td>\n";
5127
5128                 } elsif ($diff->{'status'} eq "R" || $diff->{'status'} eq "C") { # renamed or copied
5129                         my %status_name = ('R' => 'moved', 'C' => 'copied');
5130                         my $nstatus = $status_name{$diff->{'status'}};
5131                         my $mode_chng = "";
5132                         if ($diff->{'from_mode'} != $diff->{'to_mode'}) {
5133                                 # mode also for directories, so we cannot use $to_mode_str
5134                                 $mode_chng = sprintf(", mode: %04o", $to_mode_oct & 0777);
5135                         }
5136                         print "<td>" .
5137                               $cgi->a({-href => href(action=>"blob", hash_base=>$hash,
5138                                                      hash=>$diff->{'to_id'}, file_name=>$diff->{'to_file'}),
5139                                       -class => "list"}, esc_path($diff->{'to_file'})) . "</td>\n" .
5140                               "<td><span class=\"file_status $nstatus\">[$nstatus from " .
5141                               $cgi->a({-href => href(action=>"blob", hash_base=>$parent,
5142                                                      hash=>$diff->{'from_id'}, file_name=>$diff->{'from_file'}),
5143                                       -class => "list"}, esc_path($diff->{'from_file'})) .
5144                               " with " . (int $diff->{'similarity'}) . "% similarity$mode_chng]</span></td>\n" .
5145                               "<td class=\"link\">";
5146                         if ($action eq 'commitdiff') {
5147                                 # link to patch
5148                                 $patchno++;
5149                                 print $cgi->a({-href => href(-anchor=>"patch$patchno")},
5150                                               "patch") .
5151                                       " | ";
5152                         } elsif ($diff->{'to_id'} ne $diff->{'from_id'}) {
5153                                 # "commit" view and modified file (not only pure rename or copy)
5154                                 print $cgi->a({-href => href(action=>"blobdiff",
5155                                                              hash=>$diff->{'to_id'}, hash_parent=>$diff->{'from_id'},
5156                                                              hash_base=>$hash, hash_parent_base=>$parent,
5157                                                              file_name=>$diff->{'to_file'}, file_parent=>$diff->{'from_file'})},
5158                                               "diff") .
5159                                       " | ";
5160                         }
5161                         print $cgi->a({-href => href(action=>"blob", hash=>$diff->{'to_id'},
5162                                                      hash_base=>$parent, file_name=>$diff->{'to_file'})},
5163                                       "blob") . " | ";
5164                         if ($have_blame) {
5165                                 print $cgi->a({-href => href(action=>"blame", hash_base=>$hash,
5166                                                              file_name=>$diff->{'to_file'})},
5167                                               "blame") . " | ";
5168                         }
5169                         print $cgi->a({-href => href(action=>"history", hash_base=>$hash,
5170                                                     file_name=>$diff->{'to_file'})},
5171                                       "history");
5172                         print "</td>\n";
5173
5174                 } # we should not encounter Unmerged (U) or Unknown (X) status
5175                 print "</tr>\n";
5176         }
5177         print "</tbody>" if $has_header;
5178         print "</table>\n";
5179 }
5180
5181 # Print context lines and then rem/add lines in a side-by-side manner.
5182 sub print_sidebyside_diff_lines {
5183         my ($ctx, $rem, $add) = @_;
5184
5185         # print context block before add/rem block
5186         if (@$ctx) {
5187                 print join '',
5188                         '<div class="chunk_block ctx">',
5189                                 '<div class="old">',
5190                                 @$ctx,
5191                                 '</div>',
5192                                 '<div class="new">',
5193                                 @$ctx,
5194                                 '</div>',
5195                         '</div>';
5196         }
5197
5198         if (!@$add) {
5199                 # pure removal
5200                 print join '',
5201                         '<div class="chunk_block rem">',
5202                                 '<div class="old">',
5203                                 @$rem,
5204                                 '</div>',
5205                         '</div>';
5206         } elsif (!@$rem) {
5207                 # pure addition
5208                 print join '',
5209                         '<div class="chunk_block add">',
5210                                 '<div class="new">',
5211                                 @$add,
5212                                 '</div>',
5213                         '</div>';
5214         } else {
5215                 print join '',
5216                         '<div class="chunk_block chg">',
5217                                 '<div class="old">',
5218                                 @$rem,
5219                                 '</div>',
5220                                 '<div class="new">',
5221                                 @$add,
5222                                 '</div>',
5223                         '</div>';
5224         }
5225 }
5226
5227 # Print context lines and then rem/add lines in inline manner.
5228 sub print_inline_diff_lines {
5229         my ($ctx, $rem, $add) = @_;
5230
5231         print @$ctx, @$rem, @$add;
5232 }
5233
5234 # Format removed and added line, mark changed part and HTML-format them.
5235 # Implementation is based on contrib/diff-highlight
5236 sub format_rem_add_lines_pair {
5237         my ($rem, $add, $num_parents) = @_;
5238
5239         # We need to untabify lines before split()'ing them;
5240         # otherwise offsets would be invalid.
5241         chomp $rem;
5242         chomp $add;
5243         $rem = untabify($rem);
5244         $add = untabify($add);
5245
5246         my @rem = split(//, $rem);
5247         my @add = split(//, $add);
5248         my ($esc_rem, $esc_add);
5249         # Ignore leading +/- characters for each parent.
5250         my ($prefix_len, $suffix_len) = ($num_parents, 0);
5251         my ($prefix_has_nonspace, $suffix_has_nonspace);
5252
5253         my $shorter = (@rem < @add) ? @rem : @add;
5254         while ($prefix_len < $shorter) {
5255                 last if ($rem[$prefix_len] ne $add[$prefix_len]);
5256
5257                 $prefix_has_nonspace = 1 if ($rem[$prefix_len] !~ /\s/);
5258                 $prefix_len++;
5259         }
5260
5261         while ($prefix_len + $suffix_len < $shorter) {
5262                 last if ($rem[-1 - $suffix_len] ne $add[-1 - $suffix_len]);
5263
5264                 $suffix_has_nonspace = 1 if ($rem[-1 - $suffix_len] !~ /\s/);
5265                 $suffix_len++;
5266         }
5267
5268         # Mark lines that are different from each other, but have some common
5269         # part that isn't whitespace.  If lines are completely different, don't
5270         # mark them because that would make output unreadable, especially if
5271         # diff consists of multiple lines.
5272         if ($prefix_has_nonspace || $suffix_has_nonspace) {
5273                 $esc_rem = esc_html_hl_regions($rem, 'marked',
5274                         [$prefix_len, @rem - $suffix_len], -nbsp=>1);
5275                 $esc_add = esc_html_hl_regions($add, 'marked',
5276                         [$prefix_len, @add - $suffix_len], -nbsp=>1);
5277         } else {
5278                 $esc_rem = esc_html($rem, -nbsp=>1);
5279                 $esc_add = esc_html($add, -nbsp=>1);
5280         }
5281
5282         return format_diff_line(\$esc_rem, 'rem'),
5283                format_diff_line(\$esc_add, 'add');
5284 }
5285
5286 # HTML-format diff context, removed and added lines.
5287 sub format_ctx_rem_add_lines {
5288         my ($ctx, $rem, $add, $num_parents) = @_;
5289         my (@new_ctx, @new_rem, @new_add);
5290         my $can_highlight = 0;
5291         my $is_combined = ($num_parents > 1);
5292
5293         # Highlight if every removed line has a corresponding added line.
5294         if (@$add > 0 && @$add == @$rem) {
5295                 $can_highlight = 1;
5296
5297                 # Highlight lines in combined diff only if the chunk contains
5298                 # diff between the same version, e.g.
5299                 #
5300                 #    - a
5301                 #   -  b
5302                 #    + c
5303                 #   +  d
5304                 #
5305                 # Otherwise the highlighting would be confusing.
5306                 if ($is_combined) {
5307                         for (my $i = 0; $i < @$add; $i++) {
5308                                 my $prefix_rem = substr($rem->[$i], 0, $num_parents);
5309                                 my $prefix_add = substr($add->[$i], 0, $num_parents);
5310
5311                                 $prefix_rem =~ s/-/+/g;
5312
5313                                 if ($prefix_rem ne $prefix_add) {
5314                                         $can_highlight = 0;
5315                                         last;
5316                                 }
5317                         }
5318                 }
5319         }
5320
5321         if ($can_highlight) {
5322                 for (my $i = 0; $i < @$add; $i++) {
5323                         my ($line_rem, $line_add) = format_rem_add_lines_pair(
5324                                 $rem->[$i], $add->[$i], $num_parents);
5325                         push @new_rem, $line_rem;
5326                         push @new_add, $line_add;
5327                 }
5328         } else {
5329                 @new_rem = map { format_diff_line($_, 'rem') } @$rem;
5330                 @new_add = map { format_diff_line($_, 'add') } @$add;
5331         }
5332
5333         @new_ctx = map { format_diff_line($_, 'ctx') } @$ctx;
5334
5335         return (\@new_ctx, \@new_rem, \@new_add);
5336 }
5337
5338 # Print context lines and then rem/add lines.
5339 sub print_diff_lines {
5340         my ($ctx, $rem, $add, $diff_style, $num_parents) = @_;
5341         my $is_combined = $num_parents > 1;
5342
5343         ($ctx, $rem, $add) = format_ctx_rem_add_lines($ctx, $rem, $add,
5344                 $num_parents);
5345
5346         if ($diff_style eq 'sidebyside' && !$is_combined) {
5347                 print_sidebyside_diff_lines($ctx, $rem, $add);
5348         } else {
5349                 # default 'inline' style and unknown styles
5350                 print_inline_diff_lines($ctx, $rem, $add);
5351         }
5352 }
5353
5354 sub print_diff_chunk {
5355         my ($diff_style, $num_parents, $from, $to, @chunk) = @_;
5356         my (@ctx, @rem, @add);
5357
5358         # The class of the previous line.
5359         my $prev_class = '';
5360
5361         return unless @chunk;
5362
5363         # incomplete last line might be among removed or added lines,
5364         # or both, or among context lines: find which
5365         for (my $i = 1; $i < @chunk; $i++) {
5366                 if ($chunk[$i][0] eq 'incomplete') {
5367                         $chunk[$i][0] = $chunk[$i-1][0];
5368                 }
5369         }
5370
5371         # guardian
5372         push @chunk, ["", ""];
5373
5374         foreach my $line_info (@chunk) {
5375                 my ($class, $line) = @$line_info;
5376
5377                 # print chunk headers
5378                 if ($class && $class eq 'chunk_header') {
5379                         print format_diff_line($line, $class, $from, $to);
5380                         next;
5381                 }
5382
5383                 ## print from accumulator when have some add/rem lines or end
5384                 # of chunk (flush context lines), or when have add and rem
5385                 # lines and new block is reached (otherwise add/rem lines could
5386                 # be reordered)
5387                 if (!$class || ((@rem || @add) && $class eq 'ctx') ||
5388                     (@rem && @add && $class ne $prev_class)) {
5389                         print_diff_lines(\@ctx, \@rem, \@add,
5390                                          $diff_style, $num_parents);
5391                         @ctx = @rem = @add = ();
5392                 }
5393
5394                 ## adding lines to accumulator
5395                 # guardian value
5396                 last unless $line;
5397                 # rem, add or change
5398                 if ($class eq 'rem') {
5399                         push @rem, $line;
5400                 } elsif ($class eq 'add') {
5401                         push @add, $line;
5402                 }
5403                 # context line
5404                 if ($class eq 'ctx') {
5405                         push @ctx, $line;
5406                 }
5407
5408                 $prev_class = $class;
5409         }
5410 }
5411
5412 sub git_patchset_body {
5413         my ($fd, $diff_style, $difftree, $hash, @hash_parents) = @_;
5414         my ($hash_parent) = $hash_parents[0];
5415
5416         my $is_combined = (@hash_parents > 1);
5417         my $patch_idx = 0;
5418         my $patch_number = 0;
5419         my $patch_line;
5420         my $diffinfo;
5421         my $to_name;
5422         my (%from, %to);
5423         my @chunk; # for side-by-side diff
5424
5425         print "<div class=\"patchset\">\n";
5426
5427         # skip to first patch
5428         while ($patch_line = <$fd>) {
5429                 chomp $patch_line;
5430
5431                 last if ($patch_line =~ m/^diff /);
5432         }
5433
5434  PATCH:
5435         while ($patch_line) {
5436
5437                 # parse "git diff" header line
5438                 if ($patch_line =~ m/^diff --git (\"(?:[^\\\"]*(?:\\.[^\\\"]*)*)\"|[^ "]*) (.*)$/) {
5439                         # $1 is from_name, which we do not use
5440                         $to_name = unquote($2);
5441                         $to_name =~ s!^b/!!;
5442                 } elsif ($patch_line =~ m/^diff --(cc|combined) ("?.*"?)$/) {
5443                         # $1 is 'cc' or 'combined', which we do not use
5444                         $to_name = unquote($2);
5445                 } else {
5446                         $to_name = undef;
5447                 }
5448
5449                 # check if current patch belong to current raw line
5450                 # and parse raw git-diff line if needed
5451                 if (is_patch_split($diffinfo, { 'to_file' => $to_name })) {
5452                         # this is continuation of a split patch
5453                         print "<div class=\"patch cont\">\n";
5454                 } else {
5455                         # advance raw git-diff output if needed
5456                         $patch_idx++ if defined $diffinfo;
5457
5458                         # read and prepare patch information
5459                         $diffinfo = parsed_difftree_line($difftree->[$patch_idx]);
5460
5461                         # compact combined diff output can have some patches skipped
5462                         # find which patch (using pathname of result) we are at now;
5463                         if ($is_combined) {
5464                                 while ($to_name ne $diffinfo->{'to_file'}) {
5465                                         print "<div class=\"patch\" id=\"patch". ($patch_idx+1) ."\">\n" .
5466                                               format_diff_cc_simplified($diffinfo, @hash_parents) .
5467                                               "</div>\n";  # class="patch"
5468
5469                                         $patch_idx++;
5470                                         $patch_number++;
5471
5472                                         last if $patch_idx > $#$difftree;
5473                                         $diffinfo = parsed_difftree_line($difftree->[$patch_idx]);
5474                                 }
5475                         }
5476
5477                         # modifies %from, %to hashes
5478                         parse_from_to_diffinfo($diffinfo, \%from, \%to, @hash_parents);
5479
5480                         # this is first patch for raw difftree line with $patch_idx index
5481                         # we index @$difftree array from 0, but number patches from 1
5482                         print "<div class=\"patch\" id=\"patch". ($patch_idx+1) ."\">\n";
5483                 }
5484
5485                 # git diff header
5486                 #assert($patch_line =~ m/^diff /) if DEBUG;
5487                 #assert($patch_line !~ m!$/$!) if DEBUG; # is chomp-ed
5488                 $patch_number++;
5489                 # print "git diff" header
5490                 print format_git_diff_header_line($patch_line, $diffinfo,
5491                                                   \%from, \%to);
5492
5493                 # print extended diff header
5494                 print "<div class=\"diff extended_header\">\n";
5495         EXTENDED_HEADER:
5496                 while ($patch_line = <$fd>) {
5497                         chomp $patch_line;
5498
5499                         last EXTENDED_HEADER if ($patch_line =~ m/^--- |^diff /);
5500
5501                         print format_extended_diff_header_line($patch_line, $diffinfo,
5502                                                                \%from, \%to);
5503                 }
5504                 print "</div>\n"; # class="diff extended_header"
5505
5506                 # from-file/to-file diff header
5507                 if (! $patch_line) {
5508                         print "</div>\n"; # class="patch"
5509                         last PATCH;
5510                 }
5511                 next PATCH if ($patch_line =~ m/^diff /);
5512                 #assert($patch_line =~ m/^---/) if DEBUG;
5513
5514                 my $last_patch_line = $patch_line;
5515                 $patch_line = <$fd>;
5516                 chomp $patch_line;
5517                 #assert($patch_line =~ m/^\+\+\+/) if DEBUG;
5518
5519                 print format_diff_from_to_header($last_patch_line, $patch_line,
5520                                                  $diffinfo, \%from, \%to,
5521                                                  @hash_parents);
5522
5523                 # the patch itself
5524         LINE:
5525                 while ($patch_line = <$fd>) {
5526                         chomp $patch_line;
5527
5528                         next PATCH if ($patch_line =~ m/^diff /);
5529
5530                         my $class = diff_line_class($patch_line, \%from, \%to);
5531
5532                         if ($class eq 'chunk_header') {
5533                                 print_diff_chunk($diff_style, scalar @hash_parents, \%from, \%to, @chunk);
5534                                 @chunk = ();
5535                         }
5536
5537                         push @chunk, [ $class, $patch_line ];
5538                 }
5539
5540         } continue {
5541                 if (@chunk) {
5542                         print_diff_chunk($diff_style, scalar @hash_parents, \%from, \%to, @chunk);
5543                         @chunk = ();
5544                 }
5545                 print "</div>\n"; # class="patch"
5546         }
5547
5548         # for compact combined (--cc) format, with chunk and patch simplification
5549         # the patchset might be empty, but there might be unprocessed raw lines
5550         for (++$patch_idx if $patch_number > 0;
5551              $patch_idx < @$difftree;
5552              ++$patch_idx) {
5553                 # read and prepare patch information
5554                 $diffinfo = parsed_difftree_line($difftree->[$patch_idx]);
5555
5556                 # generate anchor for "patch" links in difftree / whatchanged part
5557                 print "<div class=\"patch\" id=\"patch". ($patch_idx+1) ."\">\n" .
5558                       format_diff_cc_simplified($diffinfo, @hash_parents) .
5559                       "</div>\n";  # class="patch"
5560
5561                 $patch_number++;
5562         }
5563
5564         if ($patch_number == 0) {
5565                 if (@hash_parents > 1) {
5566                         print "<div class=\"diff nodifferences\">Trivial merge</div>\n";
5567                 } else {
5568                         print "<div class=\"diff nodifferences\">No differences found</div>\n";
5569                 }
5570         }
5571
5572         print "</div>\n"; # class="patchset"
5573 }
5574
5575 # . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5576
5577 sub git_project_search_form {
5578         my ($searchtext, $search_use_regexp) = @_;
5579
5580         my $limit = '';
5581         if ($project_filter) {
5582                 $limit = " in '$project_filter/'";
5583         }
5584
5585         print "<div class=\"projsearch\">\n";
5586         print $cgi->start_form(-method => 'get', -action => $my_uri) .
5587               $cgi->hidden(-name => 'a', -value => 'project_list')  . "\n";
5588         print $cgi->hidden(-name => 'pf', -value => $project_filter). "\n"
5589                 if (defined $project_filter);
5590         print $cgi->textfield(-name => 's', -value => $searchtext,
5591                               -title => "Search project by name and description$limit",
5592                               -size => 60) . "\n" .
5593               "<span title=\"Extended regular expression\">" .
5594               $cgi->checkbox(-name => 'sr', -value => 1, -label => 're',
5595                              -checked => $search_use_regexp) .
5596               "</span>\n" .
5597               $cgi->submit(-name => 'btnS', -value => 'Search') .
5598               $cgi->end_form() . "\n" .
5599               $cgi->a({-href => href(project => undef, searchtext => undef,
5600                                      project_filter => $project_filter)},
5601                       esc_html("List all projects$limit")) . "<br />\n";
5602         print "</div>\n";
5603 }
5604
5605 # entry for given @keys needs filling if at least one of keys in list
5606 # is not present in %$project_info
5607 sub project_info_needs_filling {
5608         my ($project_info, @keys) = @_;
5609
5610         # return List::MoreUtils::any { !exists $project_info->{$_} } @keys;
5611         foreach my $key (@keys) {
5612                 if (!exists $project_info->{$key}) {
5613                         return 1;
5614                 }
5615         }
5616         return;
5617 }
5618
5619 # fills project list info (age, description, owner, category, forks, etc.)
5620 # for each project in the list, removing invalid projects from
5621 # returned list, or fill only specified info.
5622 #
5623 # Invalid projects are removed from the returned list if and only if you
5624 # ask 'age' or 'age_string' to be filled, because they are the only fields
5625 # that run unconditionally git command that requires repository, and
5626 # therefore do always check if project repository is invalid.
5627 #
5628 # USAGE:
5629 # * fill_project_list_info(\@project_list, 'descr_long', 'ctags')
5630 #   ensures that 'descr_long' and 'ctags' fields are filled
5631 # * @project_list = fill_project_list_info(\@project_list)
5632 #   ensures that all fields are filled (and invalid projects removed)
5633 #
5634 # NOTE: modifies $projlist, but does not remove entries from it
5635 sub fill_project_list_info {
5636         my ($projlist, @wanted_keys) = @_;
5637         my @projects;
5638         my $filter_set = sub { return @_; };
5639         if (@wanted_keys) {
5640                 my %wanted_keys = map { $_ => 1 } @wanted_keys;
5641                 $filter_set = sub { return grep { $wanted_keys{$_} } @_; };
5642         }
5643
5644         my $show_ctags = gitweb_check_feature('ctags');
5645  PROJECT:
5646         foreach my $pr (@$projlist) {
5647                 if (project_info_needs_filling($pr, $filter_set->('age', 'age_string'))) {
5648                         my (@activity) = git_get_last_activity($pr->{'path'});
5649                         unless (@activity) {
5650                                 next PROJECT;
5651                         }
5652                         ($pr->{'age'}, $pr->{'age_string'}) = @activity;
5653                 }
5654                 if (project_info_needs_filling($pr, $filter_set->('descr', 'descr_long'))) {
5655                         my $descr = git_get_project_description($pr->{'path'}) || "";
5656                         $descr = to_utf8($descr);
5657                         $pr->{'descr_long'} = $descr;
5658                         $pr->{'descr'} = chop_str($descr, $projects_list_description_width, 5);
5659                 }
5660                 if (project_info_needs_filling($pr, $filter_set->('owner'))) {
5661                         $pr->{'owner'} = git_get_project_owner("$pr->{'path'}") || "";
5662                 }
5663                 if ($show_ctags &&
5664                     project_info_needs_filling($pr, $filter_set->('ctags'))) {
5665                         $pr->{'ctags'} = git_get_project_ctags($pr->{'path'});
5666                 }
5667                 if ($projects_list_group_categories &&
5668                     project_info_needs_filling($pr, $filter_set->('category'))) {
5669                         my $cat = git_get_project_category($pr->{'path'}) ||
5670                                                            $project_list_default_category;
5671                         $pr->{'category'} = to_utf8($cat);
5672                 }
5673
5674                 push @projects, $pr;
5675         }
5676
5677         return @projects;
5678 }
5679
5680 sub sort_projects_list {
5681         my ($projlist, $order) = @_;
5682
5683         sub order_str {
5684                 my $key = shift;
5685                 return sub { $a->{$key} cmp $b->{$key} };
5686         }
5687
5688         sub order_num_then_undef {
5689                 my $key = shift;
5690                 return sub {
5691                         defined $a->{$key} ?
5692                                 (defined $b->{$key} ? $a->{$key} <=> $b->{$key} : -1) :
5693                                 (defined $b->{$key} ? 1 : 0)
5694                 };
5695         }
5696
5697         my %orderings = (
5698                 project => order_str('path'),
5699                 descr => order_str('descr_long'),
5700                 owner => order_str('owner'),
5701                 age => order_num_then_undef('age'),
5702         );
5703
5704         my $ordering = $orderings{$order};
5705         return defined $ordering ? sort $ordering @$projlist : @$projlist;
5706 }
5707
5708 # returns a hash of categories, containing the list of project
5709 # belonging to each category
5710 sub build_projlist_by_category {
5711         my ($projlist, $from, $to) = @_;
5712         my %categories;
5713
5714         $from = 0 unless defined $from;
5715         $to = $#$projlist if (!defined $to || $#$projlist < $to);
5716
5717         for (my $i = $from; $i <= $to; $i++) {
5718                 my $pr = $projlist->[$i];
5719                 push @{$categories{ $pr->{'category'} }}, $pr;
5720         }
5721
5722         return wantarray ? %categories : \%categories;
5723 }
5724
5725 # print 'sort by' <th> element, generating 'sort by $name' replay link
5726 # if that order is not selected
5727 sub print_sort_th {
5728         print format_sort_th(@_);
5729 }
5730
5731 sub format_sort_th {
5732         my ($name, $order, $header) = @_;
5733         my $sort_th = "";
5734         $header ||= ucfirst($name);
5735
5736         if ($order eq $name) {
5737                 $sort_th .= "<th>$header</th>\n";
5738         } else {
5739                 $sort_th .= "<th>" .
5740                             $cgi->a({-href => href(-replay=>1, order=>$name),
5741                                      -class => "header"}, $header) .
5742                             "</th>\n";
5743         }
5744
5745         return $sort_th;
5746 }
5747
5748 sub git_project_list_rows {
5749         my ($projlist, $from, $to, $check_forks) = @_;
5750
5751         $from = 0 unless defined $from;
5752         $to = $#$projlist if (!defined $to || $#$projlist < $to);
5753
5754         my $alternate = 1;
5755         for (my $i = $from; $i <= $to; $i++) {
5756                 my $pr = $projlist->[$i];
5757
5758                 if ($alternate) {
5759                         print "<tr class=\"dark\">\n";
5760                 } else {
5761                         print "<tr class=\"light\">\n";
5762                 }
5763                 $alternate ^= 1;
5764
5765                 if ($check_forks) {
5766                         print "<td>";
5767                         if ($pr->{'forks'}) {
5768                                 my $nforks = scalar @{$pr->{'forks'}};
5769                                 if ($nforks > 0) {
5770                                         print $cgi->a({-href => href(project=>$pr->{'path'}, action=>"forks"),
5771                                                        -title => "$nforks forks"}, "+");
5772                                 } else {
5773                                         print $cgi->span({-title => "$nforks forks"}, "+");
5774                                 }
5775                         }
5776                         print "</td>\n";
5777                 }
5778                 print "<td>" . $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary"),
5779                                         -class => "list"},
5780                                        esc_html_match_hl($pr->{'path'}, $search_regexp)) .
5781                       "</td>\n" .
5782                       "<td>" . $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary"),
5783                                         -class => "list",
5784                                         -title => $pr->{'descr_long'}},
5785                                         $search_regexp
5786                                         ? esc_html_match_hl_chopped($pr->{'descr_long'},
5787                                                                     $pr->{'descr'}, $search_regexp)
5788                                         : esc_html($pr->{'descr'})) .
5789                       "</td>\n";
5790                 unless ($omit_owner) {
5791                         print "<td><i>" . chop_and_escape_str($pr->{'owner'}, 15) . "</i></td>\n";
5792                 }
5793                 unless ($omit_age_column) {
5794                         print "<td class=\"". age_class($pr->{'age'}) . "\">" .
5795                             (defined $pr->{'age_string'} ? $pr->{'age_string'} : "No commits") . "</td>\n";
5796                 }
5797                 print"<td class=\"link\">" .
5798                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"summary")}, "summary")   . " | " .
5799                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"shortlog")}, "shortlog") . " | " .
5800                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"log")}, "log") . " | " .
5801                       $cgi->a({-href => href(project=>$pr->{'path'}, action=>"tree")}, "tree") .
5802                       ($pr->{'forks'} ? " | " . $cgi->a({-href => href(project=>$pr->{'path'}, action=>"forks")}, "forks") : '') .
5803                       "</td>\n" .
5804                       "</tr>\n";
5805         }
5806 }
5807
5808 sub git_project_list_body {
5809         # actually uses global variable $project
5810         my ($projlist, $order, $from, $to, $extra, $no_header) = @_;
5811         my @projects = @$projlist;
5812
5813         my $check_forks = gitweb_check_feature('forks');
5814         my $show_ctags  = gitweb_check_feature('ctags');
5815         my $tagfilter = $show_ctags ? $input_params{'ctag'} : undef;
5816         $check_forks = undef
5817                 if ($tagfilter || $search_regexp);
5818
5819         # filtering out forks before filling info allows to do less work
5820         @projects = filter_forks_from_projects_list(\@projects)
5821                 if ($check_forks);
5822         # search_projects_list pre-fills required info
5823         @projects = search_projects_list(\@projects,
5824                                          'search_regexp' => $search_regexp,
5825                                          'tagfilter'  => $tagfilter)
5826                 if ($tagfilter || $search_regexp);
5827         # fill the rest
5828         my @all_fields = ('descr', 'descr_long', 'ctags', 'category');
5829         push @all_fields, ('age', 'age_string') unless($omit_age_column);
5830         push @all_fields, 'owner' unless($omit_owner);
5831         @projects = fill_project_list_info(\@projects, @all_fields);
5832
5833         $order ||= $default_projects_order;
5834         $from = 0 unless defined $from;
5835         $to = $#projects if (!defined $to || $#projects < $to);
5836
5837         # short circuit
5838         if ($from > $to) {
5839                 print "<center>\n".
5840                       "<b>No such projects found</b><br />\n".
5841                       "Click ".$cgi->a({-href=>href(project=>undef)},"here")." to view all projects<br />\n".
5842                       "</center>\n<br />\n";
5843                 return;
5844         }
5845
5846         @projects = sort_projects_list(\@projects, $order);
5847
5848         if ($show_ctags) {
5849                 my $ctags = git_gather_all_ctags(\@projects);
5850                 my $cloud = git_populate_project_tagcloud($ctags);
5851                 print git_show_project_tagcloud($cloud, 64);
5852         }
5853
5854         print "<table class=\"project_list\">\n";
5855         unless ($no_header) {
5856                 print "<tr>\n";
5857                 if ($check_forks) {
5858                         print "<th></th>\n";
5859                 }
5860                 print_sort_th('project', $order, 'Project');
5861                 print_sort_th('descr', $order, 'Description');
5862                 print_sort_th('owner', $order, 'Owner') unless $omit_owner;
5863                 print_sort_th('age', $order, 'Last Change') unless $omit_age_column;
5864                 print "<th></th>\n" . # for links
5865                       "</tr>\n";
5866         }
5867
5868         if ($projects_list_group_categories) {
5869                 # only display categories with projects in the $from-$to window
5870                 @projects = sort {$a->{'category'} cmp $b->{'category'}} @projects[$from..$to];
5871                 my %categories = build_projlist_by_category(\@projects, $from, $to);
5872                 foreach my $cat (sort keys %categories) {
5873                         unless ($cat eq "") {
5874                                 print "<tr>\n";
5875                                 if ($check_forks) {
5876                                         print "<td></td>\n";
5877                                 }
5878                                 print "<td class=\"category\" colspan=\"5\">".esc_html($cat)."</td>\n";
5879                                 print "</tr>\n";
5880                         }
5881
5882                         git_project_list_rows($categories{$cat}, undef, undef, $check_forks);
5883                 }
5884         } else {
5885                 git_project_list_rows(\@projects, $from, $to, $check_forks);
5886         }
5887
5888         if (defined $extra) {
5889                 print "<tr>\n";
5890                 if ($check_forks) {
5891                         print "<td></td>\n";
5892                 }
5893                 print "<td colspan=\"5\">$extra</td>\n" .
5894                       "</tr>\n";
5895         }
5896         print "</table>\n";
5897 }
5898
5899 sub git_log_body {
5900         # uses global variable $project
5901         my ($commitlist, $from, $to, $refs, $extra) = @_;
5902
5903         $from = 0 unless defined $from;
5904         $to = $#{$commitlist} if (!defined $to || $#{$commitlist} < $to);
5905
5906         for (my $i = 0; $i <= $to; $i++) {
5907                 my %co = %{$commitlist->[$i]};
5908                 next if !%co;
5909                 my $commit = $co{'id'};
5910                 my $ref = format_ref_marker($refs, $commit);
5911                 git_print_header_div('commit',
5912                                "<span class=\"age\">$co{'age_string'}</span>" .
5913                                esc_html($co{'title'}) . $ref,
5914                                $commit);
5915                 print "<div class=\"title_text\">\n" .
5916                       "<div class=\"log_link\">\n" .
5917                       $cgi->a({-href => href(action=>"commit", hash=>$commit)}, "commit") .
5918                       " | " .
5919                       $cgi->a({-href => href(action=>"commitdiff", hash=>$commit)}, "commitdiff") .
5920                       " | " .
5921                       $cgi->a({-href => href(action=>"tree", hash=>$commit, hash_base=>$commit)}, "tree") .
5922                       "<br/>\n" .
5923                       "</div>\n";
5924                       git_print_authorship(\%co, -tag => 'span');
5925                       print "<br/>\n</div>\n";
5926
5927                 print "<div class=\"log_body\">\n";
5928                 git_print_log($co{'comment'}, -final_empty_line=> 1);
5929                 print "</div>\n";
5930         }
5931         if ($extra) {
5932                 print "<div class=\"page_nav\">\n";
5933                 print "$extra\n";
5934                 print "</div>\n";
5935         }
5936 }
5937
5938 sub git_shortlog_body {
5939         # uses global variable $project
5940         my ($commitlist, $from, $to, $refs, $extra) = @_;
5941
5942         $from = 0 unless defined $from;
5943         $to = $#{$commitlist} if (!defined $to || $#{$commitlist} < $to);
5944
5945         print "<table class=\"shortlog\">\n";
5946         my $alternate = 1;
5947         for (my $i = $from; $i <= $to; $i++) {
5948                 my %co = %{$commitlist->[$i]};
5949                 my $commit = $co{'id'};
5950                 my $ref = format_ref_marker($refs, $commit);
5951                 if ($alternate) {
5952                         print "<tr class=\"dark\">\n";
5953                 } else {
5954                         print "<tr class=\"light\">\n";
5955                 }
5956                 $alternate ^= 1;
5957                 # git_summary() used print "<td><i>$co{'age_string'}</i></td>\n" .
5958                 print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
5959                       format_author_html('td', \%co, 10) . "<td>";
5960                 print format_subject_html($co{'title'}, $co{'title_short'},
5961                                           href(action=>"commit", hash=>$commit), $ref);
5962                 print "</td>\n" .
5963                       "<td class=\"link\">" .
5964                       $cgi->a({-href => href(action=>"commit", hash=>$commit)}, "commit") . " | " .
5965                       $cgi->a({-href => href(action=>"commitdiff", hash=>$commit)}, "commitdiff") . " | " .
5966                       $cgi->a({-href => href(action=>"tree", hash=>$commit, hash_base=>$commit)}, "tree");
5967                 my $snapshot_links = format_snapshot_links($commit);
5968                 if (defined $snapshot_links) {
5969                         print " | " . $snapshot_links;
5970                 }
5971                 print "</td>\n" .
5972                       "</tr>\n";
5973         }
5974         if (defined $extra) {
5975                 print "<tr>\n" .
5976                       "<td colspan=\"4\">$extra</td>\n" .
5977                       "</tr>\n";
5978         }
5979         print "</table>\n";
5980 }
5981
5982 sub git_history_body {
5983         # Warning: assumes constant type (blob or tree) during history
5984         my ($commitlist, $from, $to, $refs, $extra,
5985             $file_name, $file_hash, $ftype) = @_;
5986
5987         $from = 0 unless defined $from;
5988         $to = $#{$commitlist} unless (defined $to && $to <= $#{$commitlist});
5989
5990         print "<table class=\"history\">\n";
5991         my $alternate = 1;
5992         for (my $i = $from; $i <= $to; $i++) {
5993                 my %co = %{$commitlist->[$i]};
5994                 if (!%co) {
5995                         next;
5996                 }
5997                 my $commit = $co{'id'};
5998
5999                 my $ref = format_ref_marker($refs, $commit);
6000
6001                 if ($alternate) {
6002                         print "<tr class=\"dark\">\n";
6003                 } else {
6004                         print "<tr class=\"light\">\n";
6005                 }
6006                 $alternate ^= 1;
6007                 print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
6008         # shortlog:   format_author_html('td', \%co, 10)
6009                       format_author_html('td', \%co, 15, 3) . "<td>";
6010                 # originally git_history used chop_str($co{'title'}, 50)
6011                 print format_subject_html($co{'title'}, $co{'title_short'},
6012                                           href(action=>"commit", hash=>$commit), $ref);
6013                 print "</td>\n" .
6014                       "<td class=\"link\">" .
6015                       $cgi->a({-href => href(action=>$ftype, hash_base=>$commit, file_name=>$file_name)}, $ftype) . " | " .
6016                       $cgi->a({-href => href(action=>"commitdiff", hash=>$commit)}, "commitdiff");
6017
6018                 if ($ftype eq 'blob') {
6019                         print " | " .
6020                               $cgi->a({-href => href(action=>"blob_plain", hash_base=>$commit, file_name=>$file_name)}, "raw");
6021
6022                         my $blob_current = $file_hash;
6023                         my $blob_parent  = git_get_hash_by_path($commit, $file_name);
6024                         if (defined $blob_current && defined $blob_parent &&
6025                                         $blob_current ne $blob_parent) {
6026                                 print " | " .
6027                                         $cgi->a({-href => href(action=>"blobdiff",
6028                                                                hash=>$blob_current, hash_parent=>$blob_parent,
6029                                                                hash_base=>$hash_base, hash_parent_base=>$commit,
6030                                                                file_name=>$file_name)},
6031                                                 "diff to current");
6032                         }
6033                 }
6034                 print "</td>\n" .
6035                       "</tr>\n";
6036         }
6037         if (defined $extra) {
6038                 print "<tr>\n" .
6039                       "<td colspan=\"4\">$extra</td>\n" .
6040                       "</tr>\n";
6041         }
6042         print "</table>\n";
6043 }
6044
6045 sub git_tags_body {
6046         # uses global variable $project
6047         my ($taglist, $from, $to, $extra) = @_;
6048         $from = 0 unless defined $from;
6049         $to = $#{$taglist} if (!defined $to || $#{$taglist} < $to);
6050
6051         print "<table class=\"tags\">\n";
6052         my $alternate = 1;
6053         for (my $i = $from; $i <= $to; $i++) {
6054                 my $entry = $taglist->[$i];
6055                 my %tag = %$entry;
6056                 my $comment = $tag{'subject'};
6057                 my $comment_short;
6058                 if (defined $comment) {
6059                         $comment_short = chop_str($comment, 30, 5);
6060                 }
6061                 if ($alternate) {
6062                         print "<tr class=\"dark\">\n";
6063                 } else {
6064                         print "<tr class=\"light\">\n";
6065                 }
6066                 $alternate ^= 1;
6067                 if (defined $tag{'age'}) {
6068                         print "<td><i>$tag{'age'}</i></td>\n";
6069                 } else {
6070                         print "<td></td>\n";
6071                 }
6072                 print "<td>" .
6073                       $cgi->a({-href => href(action=>$tag{'reftype'}, hash=>$tag{'refid'}),
6074                                -class => "list name"}, esc_html($tag{'name'})) .
6075                       "</td>\n" .
6076                       "<td>";
6077                 if (defined $comment) {
6078                         print format_subject_html($comment, $comment_short,
6079                                                   href(action=>"tag", hash=>$tag{'id'}));
6080                 }
6081                 print "</td>\n" .
6082                       "<td class=\"selflink\">";
6083                 if ($tag{'type'} eq "tag") {
6084                         print $cgi->a({-href => href(action=>"tag", hash=>$tag{'id'})}, "tag");
6085                 } else {
6086                         print "&nbsp;";
6087                 }
6088                 print "</td>\n" .
6089                       "<td class=\"link\">" . " | " .
6090                       $cgi->a({-href => href(action=>$tag{'reftype'}, hash=>$tag{'refid'})}, $tag{'reftype'});
6091                 if ($tag{'reftype'} eq "commit") {
6092                         print " | " . $cgi->a({-href => href(action=>"shortlog", hash=>$tag{'fullname'})}, "shortlog") .
6093                               " | " . $cgi->a({-href => href(action=>"log", hash=>$tag{'fullname'})}, "log");
6094                 } elsif ($tag{'reftype'} eq "blob") {
6095                         print " | " . $cgi->a({-href => href(action=>"blob_plain", hash=>$tag{'refid'})}, "raw");
6096                 }
6097                 print "</td>\n" .
6098                       "</tr>";
6099         }
6100         if (defined $extra) {
6101                 print "<tr>\n" .
6102                       "<td colspan=\"5\">$extra</td>\n" .
6103                       "</tr>\n";
6104         }
6105         print "</table>\n";
6106 }
6107
6108 sub git_heads_body {
6109         # uses global variable $project
6110         my ($headlist, $head_at, $from, $to, $extra) = @_;
6111         $from = 0 unless defined $from;
6112         $to = $#{$headlist} if (!defined $to || $#{$headlist} < $to);
6113
6114         print "<table class=\"heads\">\n";
6115         my $alternate = 1;
6116         for (my $i = $from; $i <= $to; $i++) {
6117                 my $entry = $headlist->[$i];
6118                 my %ref = %$entry;
6119                 my $curr = defined $head_at && $ref{'id'} eq $head_at;
6120                 if ($alternate) {
6121                         print "<tr class=\"dark\">\n";
6122                 } else {
6123                         print "<tr class=\"light\">\n";
6124                 }
6125                 $alternate ^= 1;
6126                 print "<td><i>$ref{'age'}</i></td>\n" .
6127                       ($curr ? "<td class=\"current_head\">" : "<td>") .
6128                       $cgi->a({-href => href(action=>"shortlog", hash=>$ref{'fullname'}),
6129                                -class => "list name"},esc_html($ref{'name'})) .
6130                       "</td>\n" .
6131                       "<td class=\"link\">" .
6132                       $cgi->a({-href => href(action=>"shortlog", hash=>$ref{'fullname'})}, "shortlog") . " | " .
6133                       $cgi->a({-href => href(action=>"log", hash=>$ref{'fullname'})}, "log") . " | " .
6134                       $cgi->a({-href => href(action=>"tree", hash=>$ref{'fullname'}, hash_base=>$ref{'fullname'})}, "tree") .
6135                       "</td>\n" .
6136                       "</tr>";
6137         }
6138         if (defined $extra) {
6139                 print "<tr>\n" .
6140                       "<td colspan=\"3\">$extra</td>\n" .
6141                       "</tr>\n";
6142         }
6143         print "</table>\n";
6144 }
6145
6146 # Display a single remote block
6147 sub git_remote_block {
6148         my ($remote, $rdata, $limit, $head) = @_;
6149
6150         my $heads = $rdata->{'heads'};
6151         my $fetch = $rdata->{'fetch'};
6152         my $push = $rdata->{'push'};
6153
6154         my $urls_table = "<table class=\"projects_list\">\n" ;
6155
6156         if (defined $fetch) {
6157                 if ($fetch eq $push) {
6158                         $urls_table .= format_repo_url("URL", $fetch);
6159                 } else {
6160                         $urls_table .= format_repo_url("Fetch URL", $fetch);
6161                         $urls_table .= format_repo_url("Push URL", $push) if defined $push;
6162                 }
6163         } elsif (defined $push) {
6164                 $urls_table .= format_repo_url("Push URL", $push);
6165         } else {
6166                 $urls_table .= format_repo_url("", "No remote URL");
6167         }
6168
6169         $urls_table .= "</table>\n";
6170
6171         my $dots;
6172         if (defined $limit && $limit < @$heads) {
6173                 $dots = $cgi->a({-href => href(action=>"remotes", hash=>$remote)}, "...");
6174         }
6175
6176         print $urls_table;
6177         git_heads_body($heads, $head, 0, $limit, $dots);
6178 }
6179
6180 # Display a list of remote names with the respective fetch and push URLs
6181 sub git_remotes_list {
6182         my ($remotedata, $limit) = @_;
6183         print "<table class=\"heads\">\n";
6184         my $alternate = 1;
6185         my @remotes = sort keys %$remotedata;
6186
6187         my $limited = $limit && $limit < @remotes;
6188
6189         $#remotes = $limit - 1 if $limited;
6190
6191         while (my $remote = shift @remotes) {
6192                 my $rdata = $remotedata->{$remote};
6193                 my $fetch = $rdata->{'fetch'};
6194                 my $push = $rdata->{'push'};
6195                 if ($alternate) {
6196                         print "<tr class=\"dark\">\n";
6197                 } else {
6198                         print "<tr class=\"light\">\n";
6199                 }
6200                 $alternate ^= 1;
6201                 print "<td>" .
6202                       $cgi->a({-href=> href(action=>'remotes', hash=>$remote),
6203                                -class=> "list name"},esc_html($remote)) .
6204                       "</td>";
6205                 print "<td class=\"link\">" .
6206                       (defined $fetch ? $cgi->a({-href=> $fetch}, "fetch") : "fetch") .
6207                       " | " .
6208                       (defined $push ? $cgi->a({-href=> $push}, "push") : "push") .
6209                       "</td>";
6210
6211                 print "</tr>\n";
6212         }
6213
6214         if ($limited) {
6215                 print "<tr>\n" .
6216                       "<td colspan=\"3\">" .
6217                       $cgi->a({-href => href(action=>"remotes")}, "...") .
6218                       "</td>\n" . "</tr>\n";
6219         }
6220
6221         print "</table>";
6222 }
6223
6224 # Display remote heads grouped by remote, unless there are too many
6225 # remotes, in which case we only display the remote names
6226 sub git_remotes_body {
6227         my ($remotedata, $limit, $head) = @_;
6228         if ($limit and $limit < keys %$remotedata) {
6229                 git_remotes_list($remotedata, $limit);
6230         } else {
6231                 fill_remote_heads($remotedata);
6232                 while (my ($remote, $rdata) = each %$remotedata) {
6233                         git_print_section({-class=>"remote", -id=>$remote},
6234                                 ["remotes", $remote, $remote], sub {
6235                                         git_remote_block($remote, $rdata, $limit, $head);
6236                                 });
6237                 }
6238         }
6239 }
6240
6241 sub git_search_message {
6242         my %co = @_;
6243
6244         my $greptype;
6245         if ($searchtype eq 'commit') {
6246                 $greptype = "--grep=";
6247         } elsif ($searchtype eq 'author') {
6248                 $greptype = "--author=";
6249         } elsif ($searchtype eq 'committer') {
6250                 $greptype = "--committer=";
6251         }
6252         $greptype .= $searchtext;
6253         my @commitlist = parse_commits($hash, 101, (100 * $page), undef,
6254                                        $greptype, '--regexp-ignore-case',
6255                                        $search_use_regexp ? '--extended-regexp' : '--fixed-strings');
6256
6257         my $paging_nav = '';
6258         if ($page > 0) {
6259                 $paging_nav .=
6260                         $cgi->a({-href => href(-replay=>1, page=>undef)},
6261                                 "first") .
6262                         " &sdot; " .
6263                         $cgi->a({-href => href(-replay=>1, page=>$page-1),
6264                                  -accesskey => "p", -title => "Alt-p"}, "prev");
6265         } else {
6266                 $paging_nav .= "first &sdot; prev";
6267         }
6268         my $next_link = '';
6269         if ($#commitlist >= 100) {
6270                 $next_link =
6271                         $cgi->a({-href => href(-replay=>1, page=>$page+1),
6272                                  -accesskey => "n", -title => "Alt-n"}, "next");
6273                 $paging_nav .= " &sdot; $next_link";
6274         } else {
6275                 $paging_nav .= " &sdot; next";
6276         }
6277
6278         git_header_html();
6279
6280         git_print_page_nav('','', $hash,$co{'tree'},$hash, $paging_nav);
6281         git_print_header_div('commit', esc_html($co{'title'}), $hash);
6282         if ($page == 0 && !@commitlist) {
6283                 print "<p>No match.</p>\n";
6284         } else {
6285                 git_search_grep_body(\@commitlist, 0, 99, $next_link);
6286         }
6287
6288         git_footer_html();
6289 }
6290
6291 sub git_search_changes {
6292         my %co = @_;
6293
6294         local $/ = "\n";
6295         open my $fd, '-|', git_cmd(), '--no-pager', 'log', @diff_opts,
6296                 '--pretty=format:%H', '--no-abbrev', '--raw', "-S$searchtext",
6297                 ($search_use_regexp ? '--pickaxe-regex' : ())
6298                         or die_error(500, "Open git-log failed");
6299
6300         git_header_html();
6301
6302         git_print_page_nav('','', $hash,$co{'tree'},$hash);
6303         git_print_header_div('commit', esc_html($co{'title'}), $hash);
6304
6305         print "<table class=\"pickaxe search\">\n";
6306         my $alternate = 1;
6307         undef %co;
6308         my @files;
6309         while (my $line = <$fd>) {
6310                 chomp $line;
6311                 next unless $line;
6312
6313                 my %set = parse_difftree_raw_line($line);
6314                 if (defined $set{'commit'}) {
6315                         # finish previous commit
6316                         if (%co) {
6317                                 print "</td>\n" .
6318                                       "<td class=\"link\">" .
6319                                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'})},
6320                                               "commit") .
6321                                       " | " .
6322                                       $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'},
6323                                                              hash_base=>$co{'id'})},
6324                                               "tree") .
6325                                       "</td>\n" .
6326                                       "</tr>\n";
6327                         }
6328
6329                         if ($alternate) {
6330                                 print "<tr class=\"dark\">\n";
6331                         } else {
6332                                 print "<tr class=\"light\">\n";
6333                         }
6334                         $alternate ^= 1;
6335                         %co = parse_commit($set{'commit'});
6336                         my $author = chop_and_escape_str($co{'author_name'}, 15, 5);
6337                         print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
6338                               "<td><i>$author</i></td>\n" .
6339                               "<td>" .
6340                               $cgi->a({-href => href(action=>"commit", hash=>$co{'id'}),
6341                                       -class => "list subject"},
6342                                       chop_and_escape_str($co{'title'}, 50) . "<br/>");
6343                 } elsif (defined $set{'to_id'}) {
6344                         next if is_deleted(\%set);
6345
6346                         print $cgi->a({-href => href(action=>"blob", hash_base=>$co{'id'},
6347                                                      hash=>$set{'to_id'}, file_name=>$set{'to_file'}),
6348                                       -class => "list"},
6349                                       "<span class=\"match\">" . esc_path($set{'file'}) . "</span>") .
6350                               "<br/>\n";
6351                 }
6352         }
6353         close $fd;
6354
6355         # finish last commit (warning: repetition!)
6356         if (%co) {
6357                 print "</td>\n" .
6358                       "<td class=\"link\">" .
6359                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'})},
6360                               "commit") .
6361                       " | " .
6362                       $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'},
6363                                              hash_base=>$co{'id'})},
6364                               "tree") .
6365                       "</td>\n" .
6366                       "</tr>\n";
6367         }
6368
6369         print "</table>\n";
6370
6371         git_footer_html();
6372 }
6373
6374 sub git_search_files {
6375         my %co = @_;
6376
6377         local $/ = "\n";
6378         open my $fd, "-|", git_cmd(), 'grep', '-n', '-z',
6379                 $search_use_regexp ? ('-E', '-i') : '-F',
6380                 $searchtext, $co{'tree'}
6381                         or die_error(500, "Open git-grep failed");
6382
6383         git_header_html();
6384
6385         git_print_page_nav('','', $hash,$co{'tree'},$hash);
6386         git_print_header_div('commit', esc_html($co{'title'}), $hash);
6387
6388         print "<table class=\"grep_search\">\n";
6389         my $alternate = 1;
6390         my $matches = 0;
6391         my $lastfile = '';
6392         my $file_href;
6393         while (my $line = <$fd>) {
6394                 chomp $line;
6395                 my ($file, $lno, $ltext, $binary);
6396                 last if ($matches++ > 1000);
6397                 if ($line =~ /^Binary file (.+) matches$/) {
6398                         $file = $1;
6399                         $binary = 1;
6400                 } else {
6401                         ($file, $lno, $ltext) = split(/\0/, $line, 3);
6402                         $file =~ s/^$co{'tree'}://;
6403                 }
6404                 if ($file ne $lastfile) {
6405                         $lastfile and print "</td></tr>\n";
6406                         if ($alternate++) {
6407                                 print "<tr class=\"dark\">\n";
6408                         } else {
6409                                 print "<tr class=\"light\">\n";
6410                         }
6411                         $file_href = href(action=>"blob", hash_base=>$co{'id'},
6412                                           file_name=>$file);
6413                         print "<td class=\"list\">".
6414                                 $cgi->a({-href => $file_href, -class => "list"}, esc_path($file));
6415                         print "</td><td>\n";
6416                         $lastfile = $file;
6417                 }
6418                 if ($binary) {
6419                         print "<div class=\"binary\">Binary file</div>\n";
6420                 } else {
6421                         $ltext = untabify($ltext);
6422                         if ($ltext =~ m/^(.*)($search_regexp)(.*)$/i) {
6423                                 $ltext = esc_html($1, -nbsp=>1);
6424                                 $ltext .= '<span class="match">';
6425                                 $ltext .= esc_html($2, -nbsp=>1);
6426                                 $ltext .= '</span>';
6427                                 $ltext .= esc_html($3, -nbsp=>1);
6428                         } else {
6429                                 $ltext = esc_html($ltext, -nbsp=>1);
6430                         }
6431                         print "<div class=\"pre\">" .
6432                                 $cgi->a({-href => $file_href.'#l'.$lno,
6433                                         -class => "linenr"}, sprintf('%4i', $lno)) .
6434                                 ' ' .  $ltext . "</div>\n";
6435                 }
6436         }
6437         if ($lastfile) {
6438                 print "</td></tr>\n";
6439                 if ($matches > 1000) {
6440                         print "<div class=\"diff nodifferences\">Too many matches, listing trimmed</div>\n";
6441                 }
6442         } else {
6443                 print "<div class=\"diff nodifferences\">No matches found</div>\n";
6444         }
6445         close $fd;
6446
6447         print "</table>\n";
6448
6449         git_footer_html();
6450 }
6451
6452 sub git_search_grep_body {
6453         my ($commitlist, $from, $to, $extra) = @_;
6454         $from = 0 unless defined $from;
6455         $to = $#{$commitlist} if (!defined $to || $#{$commitlist} < $to);
6456
6457         print "<table class=\"commit_search\">\n";
6458         my $alternate = 1;
6459         for (my $i = $from; $i <= $to; $i++) {
6460                 my %co = %{$commitlist->[$i]};
6461                 if (!%co) {
6462                         next;
6463                 }
6464                 my $commit = $co{'id'};
6465                 if ($alternate) {
6466                         print "<tr class=\"dark\">\n";
6467                 } else {
6468                         print "<tr class=\"light\">\n";
6469                 }
6470                 $alternate ^= 1;
6471                 print "<td title=\"$co{'age_string_age'}\"><i>$co{'age_string_date'}</i></td>\n" .
6472                       format_author_html('td', \%co, 15, 5) .
6473                       "<td>" .
6474                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'}),
6475                                -class => "list subject"},
6476                               chop_and_escape_str($co{'title'}, 50) . "<br/>");
6477                 my $comment = $co{'comment'};
6478                 foreach my $line (@$comment) {
6479                         if ($line =~ m/^(.*?)($search_regexp)(.*)$/i) {
6480                                 my ($lead, $match, $trail) = ($1, $2, $3);
6481                                 $match = chop_str($match, 70, 5, 'center');
6482                                 my $contextlen = int((80 - length($match))/2);
6483                                 $contextlen = 30 if ($contextlen > 30);
6484                                 $lead  = chop_str($lead,  $contextlen, 10, 'left');
6485                                 $trail = chop_str($trail, $contextlen, 10, 'right');
6486
6487                                 $lead  = esc_html($lead);
6488                                 $match = esc_html($match);
6489                                 $trail = esc_html($trail);
6490
6491                                 print "$lead<span class=\"match\">$match</span>$trail<br />";
6492                         }
6493                 }
6494                 print "</td>\n" .
6495                       "<td class=\"link\">" .
6496                       $cgi->a({-href => href(action=>"commit", hash=>$co{'id'})}, "commit") .
6497                       " | " .
6498                       $cgi->a({-href => href(action=>"commitdiff", hash=>$co{'id'})}, "commitdiff") .
6499                       " | " .
6500                       $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'}, hash_base=>$co{'id'})}, "tree");
6501                 print "</td>\n" .
6502                       "</tr>\n";
6503         }
6504         if (defined $extra) {
6505                 print "<tr>\n" .
6506                       "<td colspan=\"3\">$extra</td>\n" .
6507                       "</tr>\n";
6508         }
6509         print "</table>\n";
6510 }
6511
6512 ## ======================================================================
6513 ## ======================================================================
6514 ## actions
6515
6516 sub git_project_list {
6517         my $order = $input_params{'order'};
6518         if (defined $order && $order !~ m/none|project|descr|owner|age/) {
6519                 die_error(400, "Unknown order parameter");
6520         }
6521
6522         my @list = git_get_projects_list($project_filter, $strict_export);
6523         if (!@list) {
6524                 die_error(404, "No projects found");
6525         }
6526
6527         git_header_html();
6528         if (defined $home_text && -f $home_text) {
6529                 print "<div class=\"index_include\">\n";
6530                 insert_file($home_text);
6531                 print "</div>\n";
6532         }
6533
6534         git_project_search_form($searchtext, $search_use_regexp);
6535         git_project_list_body(\@list, $order);
6536         git_footer_html();
6537 }
6538
6539 sub git_forks {
6540         my $order = $input_params{'order'};
6541         if (defined $order && $order !~ m/none|project|descr|owner|age/) {
6542                 die_error(400, "Unknown order parameter");
6543         }
6544
6545         my $filter = $project;
6546         $filter =~ s/\.git$//;
6547         my @list = git_get_projects_list($filter);
6548         if (!@list) {
6549                 die_error(404, "No forks found");
6550         }
6551
6552         git_header_html();
6553         git_print_page_nav('','');
6554         git_print_header_div('summary', "$project forks");
6555         git_project_list_body(\@list, $order);
6556         git_footer_html();
6557 }
6558
6559 sub git_project_index {
6560         my @projects = git_get_projects_list($project_filter, $strict_export);
6561         if (!@projects) {
6562                 die_error(404, "No projects found");
6563         }
6564
6565         print $cgi->header(
6566                 -type => 'text/plain',
6567                 -charset => 'utf-8',
6568                 -content_disposition => 'inline; filename="index.aux"');
6569
6570         foreach my $pr (@projects) {
6571                 if (!exists $pr->{'owner'}) {
6572                         $pr->{'owner'} = git_get_project_owner("$pr->{'path'}");
6573                 }
6574
6575                 my ($path, $owner) = ($pr->{'path'}, $pr->{'owner'});
6576                 # quote as in CGI::Util::encode, but keep the slash, and use '+' for ' '
6577                 $path  =~ s/([^a-zA-Z0-9_.\-\/ ])/sprintf("%%%02X", ord($1))/eg;
6578                 $owner =~ s/([^a-zA-Z0-9_.\-\/ ])/sprintf("%%%02X", ord($1))/eg;
6579                 $path  =~ s/ /\+/g;
6580                 $owner =~ s/ /\+/g;
6581
6582                 print "$path $owner\n";
6583         }
6584 }
6585
6586 sub git_summary {
6587         my $descr = git_get_project_description($project) || "none";
6588         my %co = parse_commit("HEAD");
6589         my %cd = %co ? parse_date($co{'committer_epoch'}, $co{'committer_tz'}) : ();
6590         my $head = $co{'id'};
6591         my $remote_heads = gitweb_check_feature('remote_heads');
6592
6593         my $owner = git_get_project_owner($project);
6594
6595         my $refs = git_get_references();
6596         # These get_*_list functions return one more to allow us to see if
6597         # there are more ...
6598         my @taglist  = git_get_tags_list(16);
6599         my @headlist = git_get_heads_list(16);
6600         my %remotedata = $remote_heads ? git_get_remotes_list() : ();
6601         my @forklist;
6602         my $check_forks = gitweb_check_feature('forks');
6603
6604         if ($check_forks) {
6605                 # find forks of a project
6606                 my $filter = $project;
6607                 $filter =~ s/\.git$//;
6608                 @forklist = git_get_projects_list($filter);
6609                 # filter out forks of forks
6610                 @forklist = filter_forks_from_projects_list(\@forklist)
6611                         if (@forklist);
6612         }
6613
6614         git_header_html();
6615         git_print_page_nav('summary','', $head);
6616
6617         print "<div class=\"title\">&nbsp;</div>\n";
6618         print "<table class=\"projects_list\">\n" .
6619               "<tr id=\"metadata_desc\"><td>description</td><td>" . esc_html($descr) . "</td></tr>\n";
6620         if ($owner and not $omit_owner) {
6621                 print  "<tr id=\"metadata_owner\"><td>owner</td><td>" . esc_html($owner) . "</td></tr>\n";
6622         }
6623         if (defined $cd{'rfc2822'}) {
6624                 print "<tr id=\"metadata_lchange\"><td>last change</td>" .
6625                       "<td>".format_timestamp_html(\%cd)."</td></tr>\n";
6626         }
6627
6628         # use per project git URL list in $projectroot/$project/cloneurl
6629         # or make project git URL from git base URL and project name
6630         my $url_tag = "URL";
6631         my @url_list = git_get_project_url_list($project);
6632         @url_list = map { "$_/$project" } @git_base_url_list unless @url_list;
6633         foreach my $git_url (@url_list) {
6634                 next unless $git_url;
6635                 print format_repo_url($url_tag, $git_url);
6636                 $url_tag = "";
6637         }
6638
6639         # Tag cloud
6640         my $show_ctags = gitweb_check_feature('ctags');
6641         if ($show_ctags) {
6642                 my $ctags = git_get_project_ctags($project);
6643                 if (%$ctags) {
6644                         # without ability to add tags, don't show if there are none
6645                         my $cloud = git_populate_project_tagcloud($ctags);
6646                         print "<tr id=\"metadata_ctags\">" .
6647                               "<td>content tags</td>" .
6648                               "<td>".git_show_project_tagcloud($cloud, 48)."</td>" .
6649                               "</tr>\n";
6650                 }
6651         }
6652
6653         print "</table>\n";
6654
6655         # If XSS prevention is on, we don't include README.html.
6656         # TODO: Allow a readme in some safe format.
6657         if (!$prevent_xss && -s "$projectroot/$project/README.html") {
6658                 print "<div class=\"title\">readme</div>\n" .
6659                       "<div class=\"readme\">\n";
6660                 insert_file("$projectroot/$project/README.html");
6661                 print "\n</div>\n"; # class="readme"
6662         }
6663
6664         # we need to request one more than 16 (0..15) to check if
6665         # those 16 are all
6666         my @commitlist = $head ? parse_commits($head, 17) : ();
6667         if (@commitlist) {
6668                 git_print_header_div('shortlog');
6669                 git_shortlog_body(\@commitlist, 0, 15, $refs,
6670                                   $#commitlist <=  15 ? undef :
6671                                   $cgi->a({-href => href(action=>"shortlog")}, "..."));
6672         }
6673
6674         if (@taglist) {
6675                 git_print_header_div('tags');
6676                 git_tags_body(\@taglist, 0, 15,
6677                               $#taglist <=  15 ? undef :
6678                               $cgi->a({-href => href(action=>"tags")}, "..."));
6679         }
6680
6681         if (@headlist) {
6682                 git_print_header_div('heads');
6683                 git_heads_body(\@headlist, $head, 0, 15,
6684                                $#headlist <= 15 ? undef :
6685                                $cgi->a({-href => href(action=>"heads")}, "..."));
6686         }
6687
6688         if (%remotedata) {
6689                 git_print_header_div('remotes');
6690                 git_remotes_body(\%remotedata, 15, $head);
6691         }
6692
6693         if (@forklist) {
6694                 git_print_header_div('forks');
6695                 git_project_list_body(\@forklist, 'age', 0, 15,
6696                                       $#forklist <= 15 ? undef :
6697                                       $cgi->a({-href => href(action=>"forks")}, "..."),
6698                                       'no_header');
6699         }
6700
6701         git_footer_html();
6702 }
6703
6704 sub git_tag {
6705         my %tag = parse_tag($hash);
6706
6707         if (! %tag) {
6708                 die_error(404, "Unknown tag object");
6709         }
6710
6711         my $head = git_get_head_hash($project);
6712         git_header_html();
6713         git_print_page_nav('','', $head,undef,$head);
6714         git_print_header_div('commit', esc_html($tag{'name'}), $hash);
6715         print "<div class=\"title_text\">\n" .
6716               "<table class=\"object_header\">\n" .
6717               "<tr>\n" .
6718               "<td>object</td>\n" .
6719               "<td>" . $cgi->a({-class => "list", -href => href(action=>$tag{'type'}, hash=>$tag{'object'})},
6720                                $tag{'object'}) . "</td>\n" .
6721               "<td class=\"link\">" . $cgi->a({-href => href(action=>$tag{'type'}, hash=>$tag{'object'})},
6722                                               $tag{'type'}) . "</td>\n" .
6723               "</tr>\n";
6724         if (defined($tag{'author'})) {
6725                 git_print_authorship_rows(\%tag, 'author');
6726         }
6727         print "</table>\n\n" .
6728               "</div>\n";
6729         print "<div class=\"page_body\">";
6730         my $comment = $tag{'comment'};
6731         foreach my $line (@$comment) {
6732                 chomp $line;
6733                 print esc_html($line, -nbsp=>1) . "<br/>\n";
6734         }
6735         print "</div>\n";
6736         git_footer_html();
6737 }
6738
6739 sub git_blame_common {
6740         my $format = shift || 'porcelain';
6741         if ($format eq 'porcelain' && $input_params{'javascript'}) {
6742                 $format = 'incremental';
6743                 $action = 'blame_incremental'; # for page title etc
6744         }
6745
6746         # permissions
6747         gitweb_check_feature('blame')
6748                 or die_error(403, "Blame view not allowed");
6749
6750         # error checking
6751         die_error(400, "No file name given") unless $file_name;
6752         $hash_base ||= git_get_head_hash($project);
6753         die_error(404, "Couldn't find base commit") unless $hash_base;
6754         my %co = parse_commit($hash_base)
6755                 or die_error(404, "Commit not found");
6756         my $ftype = "blob";
6757         if (!defined $hash) {
6758                 $hash = git_get_hash_by_path($hash_base, $file_name, "blob")
6759                         or die_error(404, "Error looking up file");
6760         } else {
6761                 $ftype = git_get_type($hash);
6762                 if ($ftype !~ "blob") {
6763                         die_error(400, "Object is not a blob");
6764                 }
6765         }
6766
6767         my $fd;
6768         if ($format eq 'incremental') {
6769                 # get file contents (as base)
6770                 open $fd, "-|", git_cmd(), 'cat-file', 'blob', $hash
6771                         or die_error(500, "Open git-cat-file failed");
6772         } elsif ($format eq 'data') {
6773                 # run git-blame --incremental
6774                 open $fd, "-|", git_cmd(), "blame", "--incremental",
6775                         $hash_base, "--", $file_name
6776                         or die_error(500, "Open git-blame --incremental failed");
6777         } else {
6778                 # run git-blame --porcelain
6779                 open $fd, "-|", git_cmd(), "blame", '-p',
6780                         $hash_base, '--', $file_name
6781                         or die_error(500, "Open git-blame --porcelain failed");
6782         }
6783         binmode $fd, ':utf8';
6784
6785         # incremental blame data returns early
6786         if ($format eq 'data') {
6787                 print $cgi->header(
6788                         -type=>"text/plain", -charset => "utf-8",
6789                         -status=> "200 OK");
6790                 local $| = 1; # output autoflush
6791                 while (my $line = <$fd>) {
6792                         print to_utf8($line);
6793                 }
6794                 close $fd
6795                         or print "ERROR $!\n";
6796
6797                 print 'END';
6798                 if (defined $t0 && gitweb_check_feature('timed')) {
6799                         print ' '.
6800                               tv_interval($t0, [ gettimeofday() ]).
6801                               ' '.$number_of_git_cmds;
6802                 }
6803                 print "\n";
6804
6805                 return;
6806         }
6807
6808         # page header
6809         git_header_html();
6810         my $formats_nav =
6811                 $cgi->a({-href => href(action=>"blob", -replay=>1)},
6812                         "blob") .
6813                 " | ";
6814         if ($format eq 'incremental') {
6815                 $formats_nav .=
6816                         $cgi->a({-href => href(action=>"blame", javascript=>0, -replay=>1)},
6817                                 "blame") . " (non-incremental)";
6818         } else {
6819                 $formats_nav .=
6820                         $cgi->a({-href => href(action=>"blame_incremental", -replay=>1)},
6821                                 "blame") . " (incremental)";
6822         }
6823         $formats_nav .=
6824                 " | " .
6825                 $cgi->a({-href => href(action=>"history", -replay=>1)},
6826                         "history") .
6827                 " | " .
6828                 $cgi->a({-href => href(action=>$action, file_name=>$file_name)},
6829                         "HEAD");
6830         git_print_page_nav('','', $hash_base,$co{'tree'},$hash_base, $formats_nav);
6831         git_print_header_div('commit', esc_html($co{'title'}), $hash_base);
6832         git_print_page_path($file_name, $ftype, $hash_base);
6833
6834         # page body
6835         if ($format eq 'incremental') {
6836                 print "<noscript>\n<div class=\"error\"><center><b>\n".
6837                       "This page requires JavaScript to run.\n Use ".
6838                       $cgi->a({-href => href(action=>'blame',javascript=>0,-replay=>1)},
6839                               'this page').
6840                       " instead.\n".
6841                       "</b></center></div>\n</noscript>\n";
6842
6843                 print qq!<div id="progress_bar" style="width: 100%; background-color: yellow"></div>\n!;
6844         }
6845
6846         print qq!<div class="page_body">\n!;
6847         print qq!<div id="progress_info">... / ...</div>\n!
6848                 if ($format eq 'incremental');
6849         print qq!<table id="blame_table" class="blame" width="100%">\n!.
6850               #qq!<col width="5.5em" /><col width="2.5em" /><col width="*" />\n!.
6851               qq!<thead>\n!.
6852               qq!<tr><th>Commit</th><th>Line</th><th>Data</th></tr>\n!.
6853               qq!</thead>\n!.
6854               qq!<tbody>\n!;
6855
6856         my @rev_color = qw(light dark);
6857         my $num_colors = scalar(@rev_color);
6858         my $current_color = 0;
6859
6860         if ($format eq 'incremental') {
6861                 my $color_class = $rev_color[$current_color];
6862
6863                 #contents of a file
6864                 my $linenr = 0;
6865         LINE:
6866                 while (my $line = <$fd>) {
6867                         chomp $line;
6868                         $linenr++;
6869
6870                         print qq!<tr id="l$linenr" class="$color_class">!.
6871                               qq!<td class="sha1"><a href=""> </a></td>!.
6872                               qq!<td class="linenr">!.
6873                               qq!<a class="linenr" href="">$linenr</a></td>!;
6874                         print qq!<td class="pre">! . esc_html($line) . "</td>\n";
6875                         print qq!</tr>\n!;
6876                 }
6877
6878         } else { # porcelain, i.e. ordinary blame
6879                 my %metainfo = (); # saves information about commits
6880
6881                 # blame data
6882         LINE:
6883                 while (my $line = <$fd>) {
6884                         chomp $line;
6885                         # the header: <SHA-1> <src lineno> <dst lineno> [<lines in group>]
6886                         # no <lines in group> for subsequent lines in group of lines
6887                         my ($full_rev, $orig_lineno, $lineno, $group_size) =
6888                            ($line =~ /^($oid_regex) (\d+) (\d+)(?: (\d+))?$/);
6889                         if (!exists $metainfo{$full_rev}) {
6890                                 $metainfo{$full_rev} = { 'nprevious' => 0 };
6891                         }
6892                         my $meta = $metainfo{$full_rev};
6893                         my $data;
6894                         while ($data = <$fd>) {
6895                                 chomp $data;
6896                                 last if ($data =~ s/^\t//); # contents of line
6897                                 if ($data =~ /^(\S+)(?: (.*))?$/) {
6898                                         $meta->{$1} = $2 unless exists $meta->{$1};
6899                                 }
6900                                 if ($data =~ /^previous /) {
6901                                         $meta->{'nprevious'}++;
6902                                 }
6903                         }
6904                         my $short_rev = substr($full_rev, 0, 8);
6905                         my $author = $meta->{'author'};
6906                         my %date =
6907                                 parse_date($meta->{'author-time'}, $meta->{'author-tz'});
6908                         my $date = $date{'iso-tz'};
6909                         if ($group_size) {
6910                                 $current_color = ($current_color + 1) % $num_colors;
6911                         }
6912                         my $tr_class = $rev_color[$current_color];
6913                         $tr_class .= ' boundary' if (exists $meta->{'boundary'});
6914                         $tr_class .= ' no-previous' if ($meta->{'nprevious'} == 0);
6915                         $tr_class .= ' multiple-previous' if ($meta->{'nprevious'} > 1);
6916                         print "<tr id=\"l$lineno\" class=\"$tr_class\">\n";
6917                         if ($group_size) {
6918                                 print "<td class=\"sha1\"";
6919                                 print " title=\"". esc_html($author) . ", $date\"";
6920                                 print " rowspan=\"$group_size\"" if ($group_size > 1);
6921                                 print ">";
6922                                 print $cgi->a({-href => href(action=>"commit",
6923                                                              hash=>$full_rev,
6924                                                              file_name=>$file_name)},
6925                                               esc_html($short_rev));
6926                                 if ($group_size >= 2) {
6927                                         my @author_initials = ($author =~ /\b([[:upper:]])\B/g);
6928                                         if (@author_initials) {
6929                                                 print "<br />" .
6930                                                       esc_html(join('', @author_initials));
6931                                                 #           or join('.', ...)
6932                                         }
6933                                 }
6934                                 print "</td>\n";
6935                         }
6936                         # 'previous' <sha1 of parent commit> <filename at commit>
6937                         if (exists $meta->{'previous'} &&
6938                             $meta->{'previous'} =~ /^($oid_regex) (.*)$/) {
6939                                 $meta->{'parent'} = $1;
6940                                 $meta->{'file_parent'} = unquote($2);
6941                         }
6942                         my $linenr_commit =
6943                                 exists($meta->{'parent'}) ?
6944                                 $meta->{'parent'} : $full_rev;
6945                         my $linenr_filename =
6946                                 exists($meta->{'file_parent'}) ?
6947                                 $meta->{'file_parent'} : unquote($meta->{'filename'});
6948                         my $blamed = href(action => 'blame',
6949                                           file_name => $linenr_filename,
6950                                           hash_base => $linenr_commit);
6951                         print "<td class=\"linenr\">";
6952                         print $cgi->a({ -href => "$blamed#l$orig_lineno",
6953                                         -class => "linenr" },
6954                                       esc_html($lineno));
6955                         print "</td>";
6956                         print "<td class=\"pre\">" . esc_html($data) . "</td>\n";
6957                         print "</tr>\n";
6958                 } # end while
6959
6960         }
6961
6962         # footer
6963         print "</tbody>\n".
6964               "</table>\n"; # class="blame"
6965         print "</div>\n";   # class="blame_body"
6966         close $fd
6967                 or print "Reading blob failed\n";
6968
6969         git_footer_html();
6970 }
6971
6972 sub git_blame {
6973         git_blame_common();
6974 }
6975
6976 sub git_blame_incremental {
6977         git_blame_common('incremental');
6978 }
6979
6980 sub git_blame_data {
6981         git_blame_common('data');
6982 }
6983
6984 sub git_tags {
6985         my $head = git_get_head_hash($project);
6986         git_header_html();
6987         git_print_page_nav('','', $head,undef,$head,format_ref_views('tags'));
6988         git_print_header_div('summary', $project);
6989
6990         my @tagslist = git_get_tags_list();
6991         if (@tagslist) {
6992                 git_tags_body(\@tagslist);
6993         }
6994         git_footer_html();
6995 }
6996
6997 sub git_heads {
6998         my $head = git_get_head_hash($project);
6999         git_header_html();
7000         git_print_page_nav('','', $head,undef,$head,format_ref_views('heads'));
7001         git_print_header_div('summary', $project);
7002
7003         my @headslist = git_get_heads_list();
7004         if (@headslist) {
7005                 git_heads_body(\@headslist, $head);
7006         }
7007         git_footer_html();
7008 }
7009
7010 # used both for single remote view and for list of all the remotes
7011 sub git_remotes {
7012         gitweb_check_feature('remote_heads')
7013                 or die_error(403, "Remote heads view is disabled");
7014
7015         my $head = git_get_head_hash($project);
7016         my $remote = $input_params{'hash'};
7017
7018         my $remotedata = git_get_remotes_list($remote);
7019         die_error(500, "Unable to get remote information") unless defined $remotedata;
7020
7021         unless (%$remotedata) {
7022                 die_error(404, defined $remote ?
7023                         "Remote $remote not found" :
7024                         "No remotes found");
7025         }
7026
7027         git_header_html(undef, undef, -action_extra => $remote);
7028         git_print_page_nav('', '',  $head, undef, $head,
7029                 format_ref_views($remote ? '' : 'remotes'));
7030
7031         fill_remote_heads($remotedata);
7032         if (defined $remote) {
7033                 git_print_header_div('remotes', "$remote remote for $project");
7034                 git_remote_block($remote, $remotedata->{$remote}, undef, $head);
7035         } else {
7036                 git_print_header_div('summary', "$project remotes");
7037                 git_remotes_body($remotedata, undef, $head);
7038         }
7039
7040         git_footer_html();
7041 }
7042
7043 sub git_blob_plain {
7044         my $type = shift;
7045         my $expires;
7046
7047         if (!defined $hash) {
7048                 if (defined $file_name) {
7049                         my $base = $hash_base || git_get_head_hash($project);
7050                         $hash = git_get_hash_by_path($base, $file_name, "blob")
7051                                 or die_error(404, "Cannot find file");
7052                 } else {
7053                         die_error(400, "No file name defined");
7054                 }
7055         } elsif ($hash =~ m/^$oid_regex$/) {
7056                 # blobs defined by non-textual hash id's can be cached
7057                 $expires = "+1d";
7058         }
7059
7060         open my $fd, "-|", git_cmd(), "cat-file", "blob", $hash
7061                 or die_error(500, "Open git-cat-file blob '$hash' failed");
7062
7063         # content-type (can include charset)
7064         $type = blob_contenttype($fd, $file_name, $type);
7065
7066         # "save as" filename, even when no $file_name is given
7067         my $save_as = "$hash";
7068         if (defined $file_name) {
7069                 $save_as = $file_name;
7070         } elsif ($type =~ m/^text\//) {
7071                 $save_as .= '.txt';
7072         }
7073
7074         # With XSS prevention on, blobs of all types except a few known safe
7075         # ones are served with "Content-Disposition: attachment" to make sure
7076         # they don't run in our security domain.  For certain image types,
7077         # blob view writes an <img> tag referring to blob_plain view, and we
7078         # want to be sure not to break that by serving the image as an
7079         # attachment (though Firefox 3 doesn't seem to care).
7080         my $sandbox = $prevent_xss &&
7081                 $type !~ m!^(?:text/[a-z]+|image/(?:gif|png|jpeg))(?:[ ;]|$)!;
7082
7083         # serve text/* as text/plain
7084         if ($prevent_xss &&
7085             ($type =~ m!^text/[a-z]+\b(.*)$! ||
7086              ($type =~ m!^[a-z]+/[a-z]\+xml\b(.*)$! && -T $fd))) {
7087                 my $rest = $1;
7088                 $rest = defined $rest ? $rest : '';
7089                 $type = "text/plain$rest";
7090         }
7091
7092         print $cgi->header(
7093                 -type => $type,
7094                 -expires => $expires,
7095                 -content_disposition =>
7096                         ($sandbox ? 'attachment' : 'inline')
7097                         . '; filename="' . $save_as . '"');
7098         local $/ = undef;
7099         local *FCGI::Stream::PRINT = $FCGI_Stream_PRINT_raw;
7100         binmode STDOUT, ':raw';
7101         print <$fd>;
7102         binmode STDOUT, ':utf8'; # as set at the beginning of gitweb.cgi
7103         close $fd;
7104 }
7105
7106 sub git_blob {
7107         my $expires;
7108
7109         if (!defined $hash) {
7110                 if (defined $file_name) {
7111                         my $base = $hash_base || git_get_head_hash($project);
7112                         $hash = git_get_hash_by_path($base, $file_name, "blob")
7113                                 or die_error(404, "Cannot find file");
7114                 } else {
7115                         die_error(400, "No file name defined");
7116                 }
7117         } elsif ($hash =~ m/^$oid_regex$/) {
7118                 # blobs defined by non-textual hash id's can be cached
7119                 $expires = "+1d";
7120         }
7121
7122         my $have_blame = gitweb_check_feature('blame');
7123         open my $fd, "-|", git_cmd(), "cat-file", "blob", $hash
7124                 or die_error(500, "Couldn't cat $file_name, $hash");
7125         my $mimetype = blob_mimetype($fd, $file_name);
7126         # use 'blob_plain' (aka 'raw') view for files that cannot be displayed
7127         if ($mimetype !~ m!^(?:text/|image/(?:gif|png|jpeg)$)! && -B $fd) {
7128                 close $fd;
7129                 return git_blob_plain($mimetype);
7130         }
7131         # we can have blame only for text/* mimetype
7132         $have_blame &&= ($mimetype =~ m!^text/!);
7133
7134         my $highlight = gitweb_check_feature('highlight');
7135         my $syntax = guess_file_syntax($highlight, $file_name);
7136         $fd = run_highlighter($fd, $highlight, $syntax);
7137
7138         git_header_html(undef, $expires);
7139         my $formats_nav = '';
7140         if (defined $hash_base && (my %co = parse_commit($hash_base))) {
7141                 if (defined $file_name) {
7142                         if ($have_blame) {
7143                                 $formats_nav .=
7144                                         $cgi->a({-href => href(action=>"blame", -replay=>1)},
7145                                                 "blame") .
7146                                         " | ";
7147                         }
7148                         $formats_nav .=
7149                                 $cgi->a({-href => href(action=>"history", -replay=>1)},
7150                                         "history") .
7151                                 " | " .
7152                                 $cgi->a({-href => href(action=>"blob_plain", -replay=>1)},
7153                                         "raw") .
7154                                 " | " .
7155                                 $cgi->a({-href => href(action=>"blob",
7156                                                        hash_base=>"HEAD", file_name=>$file_name)},
7157                                         "HEAD");
7158                 } else {
7159                         $formats_nav .=
7160                                 $cgi->a({-href => href(action=>"blob_plain", -replay=>1)},
7161                                         "raw");
7162                 }
7163                 git_print_page_nav('','', $hash_base,$co{'tree'},$hash_base, $formats_nav);
7164                 git_print_header_div('commit', esc_html($co{'title'}), $hash_base);
7165         } else {
7166                 print "<div class=\"page_nav\">\n" .
7167                       "<br/><br/></div>\n" .
7168                       "<div class=\"title\">".esc_html($hash)."</div>\n";
7169         }
7170         git_print_page_path($file_name, "blob", $hash_base);
7171         print "<div class=\"page_body\">\n";
7172         if ($mimetype =~ m!^image/!) {
7173                 print qq!<img class="blob" type="!.esc_attr($mimetype).qq!"!;
7174                 if ($file_name) {
7175                         print qq! alt="!.esc_attr($file_name).qq!" title="!.esc_attr($file_name).qq!"!;
7176                 }
7177                 print qq! src="! .
7178                       esc_attr(href(action=>"blob_plain", hash=>$hash,
7179                            hash_base=>$hash_base, file_name=>$file_name)) .
7180                       qq!" />\n!;
7181         } else {
7182                 my $nr;
7183                 while (my $line = <$fd>) {
7184                         chomp $line;
7185                         $nr++;
7186                         $line = untabify($line);
7187                         printf qq!<div class="pre"><a id="l%i" href="%s#l%i" class="linenr">%4i</a> %s</div>\n!,
7188                                $nr, esc_attr(href(-replay => 1)), $nr, $nr,
7189                                $highlight ? sanitize($line) : esc_html($line, -nbsp=>1);
7190                 }
7191         }
7192         close $fd
7193                 or print "Reading blob failed.\n";
7194         print "</div>";
7195         git_footer_html();
7196 }
7197
7198 sub git_tree {
7199         if (!defined $hash_base) {
7200                 $hash_base = "HEAD";
7201         }
7202         if (!defined $hash) {
7203                 if (defined $file_name) {
7204                         $hash = git_get_hash_by_path($hash_base, $file_name, "tree");
7205                 } else {
7206                         $hash = $hash_base;
7207                 }
7208         }
7209         die_error(404, "No such tree") unless defined($hash);
7210
7211         my $show_sizes = gitweb_check_feature('show-sizes');
7212         my $have_blame = gitweb_check_feature('blame');
7213
7214         my @entries = ();
7215         {
7216                 local $/ = "\0";
7217                 open my $fd, "-|", git_cmd(), "ls-tree", '-z',
7218                         ($show_sizes ? '-l' : ()), @extra_options, $hash
7219                         or die_error(500, "Open git-ls-tree failed");
7220                 @entries = map { chomp; $_ } <$fd>;
7221                 close $fd
7222                         or die_error(404, "Reading tree failed");
7223         }
7224
7225         my $refs = git_get_references();
7226         my $ref = format_ref_marker($refs, $hash_base);
7227         git_header_html();
7228         my $basedir = '';
7229         if (defined $hash_base && (my %co = parse_commit($hash_base))) {
7230                 my @views_nav = ();
7231                 if (defined $file_name) {
7232                         push @views_nav,
7233                                 $cgi->a({-href => href(action=>"history", -replay=>1)},
7234                                         "history"),
7235                                 $cgi->a({-href => href(action=>"tree",
7236                                                        hash_base=>"HEAD", file_name=>$file_name)},
7237                                         "HEAD"),
7238                 }
7239                 my $snapshot_links = format_snapshot_links($hash);
7240                 if (defined $snapshot_links) {
7241                         # FIXME: Should be available when we have no hash base as well.
7242                         push @views_nav, $snapshot_links;
7243                 }
7244                 git_print_page_nav('tree','', $hash_base, undef, undef,
7245                                    join(' | ', @views_nav));
7246                 git_print_header_div('commit', esc_html($co{'title'}) . $ref, $hash_base);
7247         } else {
7248                 undef $hash_base;
7249                 print "<div class=\"page_nav\">\n";
7250                 print "<br/><br/></div>\n";
7251                 print "<div class=\"title\">".esc_html($hash)."</div>\n";
7252         }
7253         if (defined $file_name) {
7254                 $basedir = $file_name;
7255                 if ($basedir ne '' && substr($basedir, -1) ne '/') {
7256                         $basedir .= '/';
7257                 }
7258                 git_print_page_path($file_name, 'tree', $hash_base);
7259         }
7260         print "<div class=\"page_body\">\n";
7261         print "<table class=\"tree\">\n";
7262         my $alternate = 1;
7263         # '..' (top directory) link if possible
7264         if (defined $hash_base &&
7265             defined $file_name && $file_name =~ m![^/]+$!) {
7266                 if ($alternate) {
7267                         print "<tr class=\"dark\">\n";
7268                 } else {
7269                         print "<tr class=\"light\">\n";
7270                 }
7271                 $alternate ^= 1;
7272
7273                 my $up = $file_name;
7274                 $up =~ s!/?[^/]+$!!;
7275                 undef $up unless $up;
7276                 # based on git_print_tree_entry
7277                 print '<td class="mode">' . mode_str('040000') . "</td>\n";
7278                 print '<td class="size">&nbsp;</td>'."\n" if $show_sizes;
7279                 print '<td class="list">';
7280                 print $cgi->a({-href => href(action=>"tree",
7281                                              hash_base=>$hash_base,
7282                                              file_name=>$up)},
7283                               "..");
7284                 print "</td>\n";
7285                 print "<td class=\"link\"></td>\n";
7286
7287                 print "</tr>\n";
7288         }
7289         foreach my $line (@entries) {
7290                 my %t = parse_ls_tree_line($line, -z => 1, -l => $show_sizes);
7291
7292                 if ($alternate) {
7293                         print "<tr class=\"dark\">\n";
7294                 } else {
7295                         print "<tr class=\"light\">\n";
7296                 }
7297                 $alternate ^= 1;
7298
7299                 git_print_tree_entry(\%t, $basedir, $hash_base, $have_blame);
7300
7301                 print "</tr>\n";
7302         }
7303         print "</table>\n" .
7304               "</div>";
7305         git_footer_html();
7306 }
7307
7308 sub sanitize_for_filename {
7309     my $name = shift;
7310
7311     $name =~ s!/!-!g;
7312     $name =~ s/[^[:alnum:]_.-]//g;
7313
7314     return $name;
7315 }
7316
7317 sub snapshot_name {
7318         my ($project, $hash) = @_;
7319
7320         # path/to/project.git  -> project
7321         # path/to/project/.git -> project
7322         my $name = to_utf8($project);
7323         $name =~ s,([^/])/*\.git$,$1,;
7324         $name = sanitize_for_filename(basename($name));
7325
7326         my $ver = $hash;
7327         if ($hash =~ /^[0-9a-fA-F]+$/) {
7328                 # shorten SHA-1 hash
7329                 my $full_hash = git_get_full_hash($project, $hash);
7330                 if ($full_hash =~ /^$hash/ && length($hash) > 7) {
7331                         $ver = git_get_short_hash($project, $hash);
7332                 }
7333         } elsif ($hash =~ m!^refs/tags/(.*)$!) {
7334                 # tags don't need shortened SHA-1 hash
7335                 $ver = $1;
7336         } else {
7337                 # branches and other need shortened SHA-1 hash
7338                 my $strip_refs = join '|', map { quotemeta } get_branch_refs();
7339                 if ($hash =~ m!^refs/($strip_refs|remotes)/(.*)$!) {
7340                         my $ref_dir = (defined $1) ? $1 : '';
7341                         $ver = $2;
7342
7343                         $ref_dir = sanitize_for_filename($ref_dir);
7344                         # for refs neither in heads nor remotes we want to
7345                         # add a ref dir to archive name
7346                         if ($ref_dir ne '' and $ref_dir ne 'heads' and $ref_dir ne 'remotes') {
7347                                 $ver = $ref_dir . '-' . $ver;
7348                         }
7349                 }
7350                 $ver .= '-' . git_get_short_hash($project, $hash);
7351         }
7352         # special case of sanitization for filename - we change
7353         # slashes to dots instead of dashes
7354         # in case of hierarchical branch names
7355         $ver =~ s!/!.!g;
7356         $ver =~ s/[^[:alnum:]_.-]//g;
7357
7358         # name = project-version_string
7359         $name = "$name-$ver";
7360
7361         return wantarray ? ($name, $name) : $name;
7362 }
7363
7364 sub exit_if_unmodified_since {
7365         my ($latest_epoch) = @_;
7366         our $cgi;
7367
7368         my $if_modified = $cgi->http('IF_MODIFIED_SINCE');
7369         if (defined $if_modified) {
7370                 my $since;
7371                 if (eval { require HTTP::Date; 1; }) {
7372                         $since = HTTP::Date::str2time($if_modified);
7373                 } elsif (eval { require Time::ParseDate; 1; }) {
7374                         $since = Time::ParseDate::parsedate($if_modified, GMT => 1);
7375                 }
7376                 if (defined $since && $latest_epoch <= $since) {
7377                         my %latest_date = parse_date($latest_epoch);
7378                         print $cgi->header(
7379                                 -last_modified => $latest_date{'rfc2822'},
7380                                 -status => '304 Not Modified');
7381                         goto DONE_GITWEB;
7382                 }
7383         }
7384 }
7385
7386 sub git_snapshot {
7387         my $format = $input_params{'snapshot_format'};
7388         if (!@snapshot_fmts) {
7389                 die_error(403, "Snapshots not allowed");
7390         }
7391         # default to first supported snapshot format
7392         $format ||= $snapshot_fmts[0];
7393         if ($format !~ m/^[a-z0-9]+$/) {
7394                 die_error(400, "Invalid snapshot format parameter");
7395         } elsif (!exists($known_snapshot_formats{$format})) {
7396                 die_error(400, "Unknown snapshot format");
7397         } elsif ($known_snapshot_formats{$format}{'disabled'}) {
7398                 die_error(403, "Snapshot format not allowed");
7399         } elsif (!grep($_ eq $format, @snapshot_fmts)) {
7400                 die_error(403, "Unsupported snapshot format");
7401         }
7402
7403         my $type = git_get_type("$hash^{}");
7404         if (!$type) {
7405                 die_error(404, 'Object does not exist');
7406         }  elsif ($type eq 'blob') {
7407                 die_error(400, 'Object is not a tree-ish');
7408         }
7409
7410         my ($name, $prefix) = snapshot_name($project, $hash);
7411         my $filename = "$name$known_snapshot_formats{$format}{'suffix'}";
7412
7413         my %co = parse_commit($hash);
7414         exit_if_unmodified_since($co{'committer_epoch'}) if %co;
7415
7416         my $cmd = quote_command(
7417                 git_cmd(), 'archive',
7418                 "--format=$known_snapshot_formats{$format}{'format'}",
7419                 "--prefix=$prefix/", $hash);
7420         if (exists $known_snapshot_formats{$format}{'compressor'}) {
7421                 $cmd .= ' | ' . quote_command(@{$known_snapshot_formats{$format}{'compressor'}});
7422         }
7423
7424         $filename =~ s/(["\\])/\\$1/g;
7425         my %latest_date;
7426         if (%co) {
7427                 %latest_date = parse_date($co{'committer_epoch'}, $co{'committer_tz'});
7428         }
7429
7430         print $cgi->header(
7431                 -type => $known_snapshot_formats{$format}{'type'},
7432                 -content_disposition => 'inline; filename="' . $filename . '"',
7433                 %co ? (-last_modified => $latest_date{'rfc2822'}) : (),
7434                 -status => '200 OK');
7435
7436         open my $fd, "-|", $cmd
7437                 or die_error(500, "Execute git-archive failed");
7438         local *FCGI::Stream::PRINT = $FCGI_Stream_PRINT_raw;
7439         binmode STDOUT, ':raw';
7440         print <$fd>;
7441         binmode STDOUT, ':utf8'; # as set at the beginning of gitweb.cgi
7442         close $fd;
7443 }
7444
7445 sub git_log_generic {
7446         my ($fmt_name, $body_subr, $base, $parent, $file_name, $file_hash) = @_;
7447
7448         my $head = git_get_head_hash($project);
7449         if (!defined $base) {
7450                 $base = $head;
7451         }
7452         if (!defined $page) {
7453                 $page = 0;
7454         }
7455         my $refs = git_get_references();
7456
7457         my $commit_hash = $base;
7458         if (defined $parent) {
7459                 $commit_hash = "$parent..$base";
7460         }
7461         my @commitlist =
7462                 parse_commits($commit_hash, 101, (100 * $page),
7463                               defined $file_name ? ($file_name, "--full-history") : ());
7464
7465         my $ftype;
7466         if (!defined $file_hash && defined $file_name) {
7467                 # some commits could have deleted file in question,
7468                 # and not have it in tree, but one of them has to have it
7469                 for (my $i = 0; $i < @commitlist; $i++) {
7470                         $file_hash = git_get_hash_by_path($commitlist[$i]{'id'}, $file_name);
7471                         last if defined $file_hash;
7472                 }
7473         }
7474         if (defined $file_hash) {
7475                 $ftype = git_get_type($file_hash);
7476         }
7477         if (defined $file_name && !defined $ftype) {
7478                 die_error(500, "Unknown type of object");
7479         }
7480         my %co;
7481         if (defined $file_name) {
7482                 %co = parse_commit($base)
7483                         or die_error(404, "Unknown commit object");
7484         }
7485
7486
7487         my $paging_nav = format_paging_nav($fmt_name, $page, $#commitlist >= 100);
7488         my $next_link = '';
7489         if ($#commitlist >= 100) {
7490                 $next_link =
7491                         $cgi->a({-href => href(-replay=>1, page=>$page+1),
7492                                  -accesskey => "n", -title => "Alt-n"}, "next");
7493         }
7494         my $patch_max = gitweb_get_feature('patches');
7495         if ($patch_max && !defined $file_name &&
7496                 !gitweb_check_feature('email-privacy')) {
7497                 if ($patch_max < 0 || @commitlist <= $patch_max) {
7498                         $paging_nav .= " &sdot; " .
7499                                 $cgi->a({-href => href(action=>"patches", -replay=>1)},
7500                                         "patches");
7501                 }
7502         }
7503
7504         git_header_html();
7505         git_print_page_nav($fmt_name,'', $hash,$hash,$hash, $paging_nav);
7506         if (defined $file_name) {
7507                 git_print_header_div('commit', esc_html($co{'title'}), $base);
7508         } else {
7509                 git_print_header_div('summary', $project)
7510         }
7511         git_print_page_path($file_name, $ftype, $hash_base)
7512                 if (defined $file_name);
7513
7514         $body_subr->(\@commitlist, 0, 99, $refs, $next_link,
7515                      $file_name, $file_hash, $ftype);
7516
7517         git_footer_html();
7518 }
7519
7520 sub git_log {
7521         git_log_generic('log', \&git_log_body,
7522                         $hash, $hash_parent);
7523 }
7524
7525 sub git_commit {
7526         $hash ||= $hash_base || "HEAD";
7527         my %co = parse_commit($hash)
7528             or die_error(404, "Unknown commit object");
7529
7530         my $parent  = $co{'parent'};
7531         my $parents = $co{'parents'}; # listref
7532
7533         # we need to prepare $formats_nav before any parameter munging
7534         my $formats_nav;
7535         if (!defined $parent) {
7536                 # --root commitdiff
7537                 $formats_nav .= '(initial)';
7538         } elsif (@$parents == 1) {
7539                 # single parent commit
7540                 $formats_nav .=
7541                         '(parent: ' .
7542                         $cgi->a({-href => href(action=>"commit",
7543                                                hash=>$parent)},
7544                                 esc_html(substr($parent, 0, 7))) .
7545                         ')';
7546         } else {
7547                 # merge commit
7548                 $formats_nav .=
7549                         '(merge: ' .
7550                         join(' ', map {
7551                                 $cgi->a({-href => href(action=>"commit",
7552                                                        hash=>$_)},
7553                                         esc_html(substr($_, 0, 7)));
7554                         } @$parents ) .
7555                         ')';
7556         }
7557         if (gitweb_check_feature('patches') && @$parents <= 1 &&
7558                 !gitweb_check_feature('email-privacy')) {
7559                 $formats_nav .= " | " .
7560                         $cgi->a({-href => href(action=>"patch", -replay=>1)},
7561                                 "patch");
7562         }
7563
7564         if (!defined $parent) {
7565                 $parent = "--root";
7566         }
7567         my @difftree;
7568         open my $fd, "-|", git_cmd(), "diff-tree", '-r', "--no-commit-id",
7569                 @diff_opts,
7570                 (@$parents <= 1 ? $parent : '-c'),
7571                 $hash, "--"
7572                 or die_error(500, "Open git-diff-tree failed");
7573         @difftree = map { chomp; $_ } <$fd>;
7574         close $fd or die_error(404, "Reading git-diff-tree failed");
7575
7576         # non-textual hash id's can be cached
7577         my $expires;
7578         if ($hash =~ m/^$oid_regex$/) {
7579                 $expires = "+1d";
7580         }
7581         my $refs = git_get_references();
7582         my $ref = format_ref_marker($refs, $co{'id'});
7583
7584         git_header_html(undef, $expires);
7585         git_print_page_nav('commit', '',
7586                            $hash, $co{'tree'}, $hash,
7587                            $formats_nav);
7588
7589         if (defined $co{'parent'}) {
7590                 git_print_header_div('commitdiff', esc_html($co{'title'}) . $ref, $hash);
7591         } else {
7592                 git_print_header_div('tree', esc_html($co{'title'}) . $ref, $co{'tree'}, $hash);
7593         }
7594         print "<div class=\"title_text\">\n" .
7595               "<table class=\"object_header\">\n";
7596         git_print_authorship_rows(\%co);
7597         print "<tr><td>commit</td><td class=\"sha1\">$co{'id'}</td></tr>\n";
7598         print "<tr>" .
7599               "<td>tree</td>" .
7600               "<td class=\"sha1\">" .
7601               $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'}, hash_base=>$hash),
7602                        class => "list"}, $co{'tree'}) .
7603               "</td>" .
7604               "<td class=\"link\">" .
7605               $cgi->a({-href => href(action=>"tree", hash=>$co{'tree'}, hash_base=>$hash)},
7606                       "tree");
7607         my $snapshot_links = format_snapshot_links($hash);
7608         if (defined $snapshot_links) {
7609                 print " | " . $snapshot_links;
7610         }
7611         print "</td>" .
7612               "</tr>\n";
7613
7614         foreach my $par (@$parents) {
7615                 print "<tr>" .
7616                       "<td>parent</td>" .
7617                       "<td class=\"sha1\">" .
7618                       $cgi->a({-href => href(action=>"commit", hash=>$par),
7619                                class => "list"}, $par) .
7620                       "</td>" .
7621                       "<td class=\"link\">" .
7622                       $cgi->a({-href => href(action=>"commit", hash=>$par)}, "commit") .
7623                       " | " .
7624                       $cgi->a({-href => href(action=>"commitdiff", hash=>$hash, hash_parent=>$par)}, "diff") .
7625                       "</td>" .
7626                       "</tr>\n";
7627         }
7628         print "</table>".
7629               "</div>\n";
7630
7631         print "<div class=\"page_body\">\n";
7632         git_print_log($co{'comment'});
7633         print "</div>\n";
7634
7635         git_difftree_body(\@difftree, $hash, @$parents);
7636
7637         git_footer_html();
7638 }
7639
7640 sub git_object {
7641         # object is defined by:
7642         # - hash or hash_base alone
7643         # - hash_base and file_name
7644         my $type;
7645
7646         # - hash or hash_base alone
7647         if ($hash || ($hash_base && !defined $file_name)) {
7648                 my $object_id = $hash || $hash_base;
7649
7650                 open my $fd, "-|", quote_command(
7651                         git_cmd(), 'cat-file', '-t', $object_id) . ' 2> /dev/null'
7652                         or die_error(404, "Object does not exist");
7653                 $type = <$fd>;
7654                 defined $type && chomp $type;
7655                 close $fd
7656                         or die_error(404, "Object does not exist");
7657
7658         # - hash_base and file_name
7659         } elsif ($hash_base && defined $file_name) {
7660                 $file_name =~ s,/+$,,;
7661
7662                 system(git_cmd(), "cat-file", '-e', $hash_base) == 0
7663                         or die_error(404, "Base object does not exist");
7664
7665                 # here errors should not happen
7666                 open my $fd, "-|", git_cmd(), "ls-tree", $hash_base, "--", $file_name
7667                         or die_error(500, "Open git-ls-tree failed");
7668                 my $line = <$fd>;
7669                 close $fd;
7670
7671                 #'100644 blob 0fa3f3a66fb6a137f6ec2c19351ed4d807070ffa  panic.c'
7672                 unless ($line && $line =~ m/^([0-9]+) (.+) ($oid_regex)\t/) {
7673                         die_error(404, "File or directory for given base does not exist");
7674                 }
7675                 $type = $2;
7676                 $hash = $3;
7677         } else {
7678                 die_error(400, "Not enough information to find object");
7679         }
7680
7681         print $cgi->redirect(-uri => href(action=>$type, -full=>1,
7682                                           hash=>$hash, hash_base=>$hash_base,
7683                                           file_name=>$file_name),
7684                              -status => '302 Found');
7685 }
7686
7687 sub git_blobdiff {
7688         my $format = shift || 'html';
7689         my $diff_style = $input_params{'diff_style'} || 'inline';
7690
7691         my $fd;
7692         my @difftree;
7693         my %diffinfo;
7694         my $expires;
7695
7696         # preparing $fd and %diffinfo for git_patchset_body
7697         # new style URI
7698         if (defined $hash_base && defined $hash_parent_base) {
7699                 if (defined $file_name) {
7700                         # read raw output
7701                         open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7702                                 $hash_parent_base, $hash_base,
7703                                 "--", (defined $file_parent ? $file_parent : ()), $file_name
7704                                 or die_error(500, "Open git-diff-tree failed");
7705                         @difftree = map { chomp; $_ } <$fd>;
7706                         close $fd
7707                                 or die_error(404, "Reading git-diff-tree failed");
7708                         @difftree
7709                                 or die_error(404, "Blob diff not found");
7710
7711                 } elsif (defined $hash &&
7712                          $hash =~ $oid_regex) {
7713                         # try to find filename from $hash
7714
7715                         # read filtered raw output
7716                         open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7717                                 $hash_parent_base, $hash_base, "--"
7718                                 or die_error(500, "Open git-diff-tree failed");
7719                         @difftree =
7720                                 # ':100644 100644 03b21826... 3b93d5e7... M     ls-files.c'
7721                                 # $hash == to_id
7722                                 grep { /^:[0-7]{6} [0-7]{6} $oid_regex $hash/ }
7723                                 map { chomp; $_ } <$fd>;
7724                         close $fd
7725                                 or die_error(404, "Reading git-diff-tree failed");
7726                         @difftree
7727                                 or die_error(404, "Blob diff not found");
7728
7729                 } else {
7730                         die_error(400, "Missing one of the blob diff parameters");
7731                 }
7732
7733                 if (@difftree > 1) {
7734                         die_error(400, "Ambiguous blob diff specification");
7735                 }
7736
7737                 %diffinfo = parse_difftree_raw_line($difftree[0]);
7738                 $file_parent ||= $diffinfo{'from_file'} || $file_name;
7739                 $file_name   ||= $diffinfo{'to_file'};
7740
7741                 $hash_parent ||= $diffinfo{'from_id'};
7742                 $hash        ||= $diffinfo{'to_id'};
7743
7744                 # non-textual hash id's can be cached
7745                 if ($hash_base =~ m/^$oid_regex$/ &&
7746                     $hash_parent_base =~ m/^$oid_regex$/) {
7747                         $expires = '+1d';
7748                 }
7749
7750                 # open patch output
7751                 open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7752                         '-p', ($format eq 'html' ? "--full-index" : ()),
7753                         $hash_parent_base, $hash_base,
7754                         "--", (defined $file_parent ? $file_parent : ()), $file_name
7755                         or die_error(500, "Open git-diff-tree failed");
7756         }
7757
7758         # old/legacy style URI -- not generated anymore since 1.4.3.
7759         if (!%diffinfo) {
7760                 die_error('404 Not Found', "Missing one of the blob diff parameters")
7761         }
7762
7763         # header
7764         if ($format eq 'html') {
7765                 my $formats_nav =
7766                         $cgi->a({-href => href(action=>"blobdiff_plain", -replay=>1)},
7767                                 "raw");
7768                 $formats_nav .= diff_style_nav($diff_style);
7769                 git_header_html(undef, $expires);
7770                 if (defined $hash_base && (my %co = parse_commit($hash_base))) {
7771                         git_print_page_nav('','', $hash_base,$co{'tree'},$hash_base, $formats_nav);
7772                         git_print_header_div('commit', esc_html($co{'title'}), $hash_base);
7773                 } else {
7774                         print "<div class=\"page_nav\"><br/>$formats_nav<br/></div>\n";
7775                         print "<div class=\"title\">".esc_html("$hash vs $hash_parent")."</div>\n";
7776                 }
7777                 if (defined $file_name) {
7778                         git_print_page_path($file_name, "blob", $hash_base);
7779                 } else {
7780                         print "<div class=\"page_path\"></div>\n";
7781                 }
7782
7783         } elsif ($format eq 'plain') {
7784                 print $cgi->header(
7785                         -type => 'text/plain',
7786                         -charset => 'utf-8',
7787                         -expires => $expires,
7788                         -content_disposition => 'inline; filename="' . "$file_name" . '.patch"');
7789
7790                 print "X-Git-Url: " . $cgi->self_url() . "\n\n";
7791
7792         } else {
7793                 die_error(400, "Unknown blobdiff format");
7794         }
7795
7796         # patch
7797         if ($format eq 'html') {
7798                 print "<div class=\"page_body\">\n";
7799
7800                 git_patchset_body($fd, $diff_style,
7801                                   [ \%diffinfo ], $hash_base, $hash_parent_base);
7802                 close $fd;
7803
7804                 print "</div>\n"; # class="page_body"
7805                 git_footer_html();
7806
7807         } else {
7808                 while (my $line = <$fd>) {
7809                         $line =~ s!a/($hash|$hash_parent)!'a/'.esc_path($diffinfo{'from_file'})!eg;
7810                         $line =~ s!b/($hash|$hash_parent)!'b/'.esc_path($diffinfo{'to_file'})!eg;
7811
7812                         print $line;
7813
7814                         last if $line =~ m!^\+\+\+!;
7815                 }
7816                 local $/ = undef;
7817                 print <$fd>;
7818                 close $fd;
7819         }
7820 }
7821
7822 sub git_blobdiff_plain {
7823         git_blobdiff('plain');
7824 }
7825
7826 # assumes that it is added as later part of already existing navigation,
7827 # so it returns "| foo | bar" rather than just "foo | bar"
7828 sub diff_style_nav {
7829         my ($diff_style, $is_combined) = @_;
7830         $diff_style ||= 'inline';
7831
7832         return "" if ($is_combined);
7833
7834         my @styles = (inline => 'inline', 'sidebyside' => 'side by side');
7835         my %styles = @styles;
7836         @styles =
7837                 @styles[ map { $_ * 2 } 0..$#styles/2 ];
7838
7839         return join '',
7840                 map { " | ".$_ }
7841                 map {
7842                         $_ eq $diff_style ? $styles{$_} :
7843                         $cgi->a({-href => href(-replay=>1, diff_style => $_)}, $styles{$_})
7844                 } @styles;
7845 }
7846
7847 sub git_commitdiff {
7848         my %params = @_;
7849         my $format = $params{-format} || 'html';
7850         my $diff_style = $input_params{'diff_style'} || 'inline';
7851
7852         my ($patch_max) = gitweb_get_feature('patches');
7853         if ($format eq 'patch') {
7854                 die_error(403, "Patch view not allowed") unless $patch_max;
7855         }
7856
7857         $hash ||= $hash_base || "HEAD";
7858         my %co = parse_commit($hash)
7859             or die_error(404, "Unknown commit object");
7860
7861         # choose format for commitdiff for merge
7862         if (! defined $hash_parent && @{$co{'parents'}} > 1) {
7863                 $hash_parent = '--cc';
7864         }
7865         # we need to prepare $formats_nav before almost any parameter munging
7866         my $formats_nav;
7867         if ($format eq 'html') {
7868                 $formats_nav =
7869                         $cgi->a({-href => href(action=>"commitdiff_plain", -replay=>1)},
7870                                 "raw");
7871                 if ($patch_max && @{$co{'parents'}} <= 1 &&
7872                         !gitweb_check_feature('email-privacy')) {
7873                         $formats_nav .= " | " .
7874                                 $cgi->a({-href => href(action=>"patch", -replay=>1)},
7875                                         "patch");
7876                 }
7877                 $formats_nav .= diff_style_nav($diff_style, @{$co{'parents'}} > 1);
7878
7879                 if (defined $hash_parent &&
7880                     $hash_parent ne '-c' && $hash_parent ne '--cc') {
7881                         # commitdiff with two commits given
7882                         my $hash_parent_short = $hash_parent;
7883                         if ($hash_parent =~ m/^$oid_regex$/) {
7884                                 $hash_parent_short = substr($hash_parent, 0, 7);
7885                         }
7886                         $formats_nav .=
7887                                 ' (from';
7888                         for (my $i = 0; $i < @{$co{'parents'}}; $i++) {
7889                                 if ($co{'parents'}[$i] eq $hash_parent) {
7890                                         $formats_nav .= ' parent ' . ($i+1);
7891                                         last;
7892                                 }
7893                         }
7894                         $formats_nav .= ': ' .
7895                                 $cgi->a({-href => href(-replay=>1,
7896                                                        hash=>$hash_parent, hash_base=>undef)},
7897                                         esc_html($hash_parent_short)) .
7898                                 ')';
7899                 } elsif (!$co{'parent'}) {
7900                         # --root commitdiff
7901                         $formats_nav .= ' (initial)';
7902                 } elsif (scalar @{$co{'parents'}} == 1) {
7903                         # single parent commit
7904                         $formats_nav .=
7905                                 ' (parent: ' .
7906                                 $cgi->a({-href => href(-replay=>1,
7907                                                        hash=>$co{'parent'}, hash_base=>undef)},
7908                                         esc_html(substr($co{'parent'}, 0, 7))) .
7909                                 ')';
7910                 } else {
7911                         # merge commit
7912                         if ($hash_parent eq '--cc') {
7913                                 $formats_nav .= ' | ' .
7914                                         $cgi->a({-href => href(-replay=>1,
7915                                                                hash=>$hash, hash_parent=>'-c')},
7916                                                 'combined');
7917                         } else { # $hash_parent eq '-c'
7918                                 $formats_nav .= ' | ' .
7919                                         $cgi->a({-href => href(-replay=>1,
7920                                                                hash=>$hash, hash_parent=>'--cc')},
7921                                                 'compact');
7922                         }
7923                         $formats_nav .=
7924                                 ' (merge: ' .
7925                                 join(' ', map {
7926                                         $cgi->a({-href => href(-replay=>1,
7927                                                                hash=>$_, hash_base=>undef)},
7928                                                 esc_html(substr($_, 0, 7)));
7929                                 } @{$co{'parents'}} ) .
7930                                 ')';
7931                 }
7932         }
7933
7934         my $hash_parent_param = $hash_parent;
7935         if (!defined $hash_parent_param) {
7936                 # --cc for multiple parents, --root for parentless
7937                 $hash_parent_param =
7938                         @{$co{'parents'}} > 1 ? '--cc' : $co{'parent'} || '--root';
7939         }
7940
7941         # read commitdiff
7942         my $fd;
7943         my @difftree;
7944         if ($format eq 'html') {
7945                 open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7946                         "--no-commit-id", "--patch-with-raw", "--full-index",
7947                         $hash_parent_param, $hash, "--"
7948                         or die_error(500, "Open git-diff-tree failed");
7949
7950                 while (my $line = <$fd>) {
7951                         chomp $line;
7952                         # empty line ends raw part of diff-tree output
7953                         last unless $line;
7954                         push @difftree, scalar parse_difftree_raw_line($line);
7955                 }
7956
7957         } elsif ($format eq 'plain') {
7958                 open $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
7959                         '-p', $hash_parent_param, $hash, "--"
7960                         or die_error(500, "Open git-diff-tree failed");
7961         } elsif ($format eq 'patch') {
7962                 # For commit ranges, we limit the output to the number of
7963                 # patches specified in the 'patches' feature.
7964                 # For single commits, we limit the output to a single patch,
7965                 # diverging from the git-format-patch default.
7966                 my @commit_spec = ();
7967                 if ($hash_parent) {
7968                         if ($patch_max > 0) {
7969                                 push @commit_spec, "-$patch_max";
7970                         }
7971                         push @commit_spec, '-n', "$hash_parent..$hash";
7972                 } else {
7973                         if ($params{-single}) {
7974                                 push @commit_spec, '-1';
7975                         } else {
7976                                 if ($patch_max > 0) {
7977                                         push @commit_spec, "-$patch_max";
7978                                 }
7979                                 push @commit_spec, "-n";
7980                         }
7981                         push @commit_spec, '--root', $hash;
7982                 }
7983                 open $fd, "-|", git_cmd(), "format-patch", @diff_opts,
7984                         '--encoding=utf8', '--stdout', @commit_spec
7985                         or die_error(500, "Open git-format-patch failed");
7986         } else {
7987                 die_error(400, "Unknown commitdiff format");
7988         }
7989
7990         # non-textual hash id's can be cached
7991         my $expires;
7992         if ($hash =~ m/^$oid_regex$/) {
7993                 $expires = "+1d";
7994         }
7995
7996         # write commit message
7997         if ($format eq 'html') {
7998                 my $refs = git_get_references();
7999                 my $ref = format_ref_marker($refs, $co{'id'});
8000
8001                 git_header_html(undef, $expires);
8002                 git_print_page_nav('commitdiff','', $hash,$co{'tree'},$hash, $formats_nav);
8003                 git_print_header_div('commit', esc_html($co{'title'}) . $ref, $hash);
8004                 print "<div class=\"title_text\">\n" .
8005                       "<table class=\"object_header\">\n";
8006                 git_print_authorship_rows(\%co);
8007                 print "</table>".
8008                       "</div>\n";
8009                 print "<div class=\"page_body\">\n";
8010                 if (@{$co{'comment'}} > 1) {
8011                         print "<div class=\"log\">\n";
8012                         git_print_log($co{'comment'}, -final_empty_line=> 1, -remove_title => 1);
8013                         print "</div>\n"; # class="log"
8014                 }
8015
8016         } elsif ($format eq 'plain') {
8017                 my $refs = git_get_references("tags");
8018                 my $tagname = git_get_rev_name_tags($hash);
8019                 my $filename = basename($project) . "-$hash.patch";
8020
8021                 print $cgi->header(
8022                         -type => 'text/plain',
8023                         -charset => 'utf-8',
8024                         -expires => $expires,
8025                         -content_disposition => 'inline; filename="' . "$filename" . '"');
8026                 my %ad = parse_date($co{'author_epoch'}, $co{'author_tz'});
8027                 print "From: " . to_utf8($co{'author'}) . "\n";
8028                 print "Date: $ad{'rfc2822'} ($ad{'tz_local'})\n";
8029                 print "Subject: " . to_utf8($co{'title'}) . "\n";
8030
8031                 print "X-Git-Tag: $tagname\n" if $tagname;
8032                 print "X-Git-Url: " . $cgi->self_url() . "\n\n";
8033
8034                 foreach my $line (@{$co{'comment'}}) {
8035                         print to_utf8($line) . "\n";
8036                 }
8037                 print "---\n\n";
8038         } elsif ($format eq 'patch') {
8039                 my $filename = basename($project) . "-$hash.patch";
8040
8041                 print $cgi->header(
8042                         -type => 'text/plain',
8043                         -charset => 'utf-8',
8044                         -expires => $expires,
8045                         -content_disposition => 'inline; filename="' . "$filename" . '"');
8046         }
8047
8048         # write patch
8049         if ($format eq 'html') {
8050                 my $use_parents = !defined $hash_parent ||
8051                         $hash_parent eq '-c' || $hash_parent eq '--cc';
8052                 git_difftree_body(\@difftree, $hash,
8053                                   $use_parents ? @{$co{'parents'}} : $hash_parent);
8054                 print "<br/>\n";
8055
8056                 git_patchset_body($fd, $diff_style,
8057                                   \@difftree, $hash,
8058                                   $use_parents ? @{$co{'parents'}} : $hash_parent);
8059                 close $fd;
8060                 print "</div>\n"; # class="page_body"
8061                 git_footer_html();
8062
8063         } elsif ($format eq 'plain') {
8064                 local $/ = undef;
8065                 print <$fd>;
8066                 close $fd
8067                         or print "Reading git-diff-tree failed\n";
8068         } elsif ($format eq 'patch') {
8069                 local $/ = undef;
8070                 print <$fd>;
8071                 close $fd
8072                         or print "Reading git-format-patch failed\n";
8073         }
8074 }
8075
8076 sub git_commitdiff_plain {
8077         git_commitdiff(-format => 'plain');
8078 }
8079
8080 # format-patch-style patches
8081 sub git_patch {
8082         git_commitdiff(-format => 'patch', -single => 1);
8083 }
8084
8085 sub git_patches {
8086         git_commitdiff(-format => 'patch');
8087 }
8088
8089 sub git_history {
8090         git_log_generic('history', \&git_history_body,
8091                         $hash_base, $hash_parent_base,
8092                         $file_name, $hash);
8093 }
8094
8095 sub git_search {
8096         $searchtype ||= 'commit';
8097
8098         # check if appropriate features are enabled
8099         gitweb_check_feature('search')
8100                 or die_error(403, "Search is disabled");
8101         if ($searchtype eq 'pickaxe') {
8102                 # pickaxe may take all resources of your box and run for several minutes
8103                 # with every query - so decide by yourself how public you make this feature
8104                 gitweb_check_feature('pickaxe')
8105                         or die_error(403, "Pickaxe search is disabled");
8106         }
8107         if ($searchtype eq 'grep') {
8108                 # grep search might be potentially CPU-intensive, too
8109                 gitweb_check_feature('grep')
8110                         or die_error(403, "Grep search is disabled");
8111         }
8112
8113         if (!defined $searchtext) {
8114                 die_error(400, "Text field is empty");
8115         }
8116         if (!defined $hash) {
8117                 $hash = git_get_head_hash($project);
8118         }
8119         my %co = parse_commit($hash);
8120         if (!%co) {
8121                 die_error(404, "Unknown commit object");
8122         }
8123         if (!defined $page) {
8124                 $page = 0;
8125         }
8126
8127         if ($searchtype eq 'commit' ||
8128             $searchtype eq 'author' ||
8129             $searchtype eq 'committer') {
8130                 git_search_message(%co);
8131         } elsif ($searchtype eq 'pickaxe') {
8132                 git_search_changes(%co);
8133         } elsif ($searchtype eq 'grep') {
8134                 git_search_files(%co);
8135         } else {
8136                 die_error(400, "Unknown search type");
8137         }
8138 }
8139
8140 sub git_search_help {
8141         git_header_html();
8142         git_print_page_nav('','', $hash,$hash,$hash);
8143         print <<EOT;
8144 <p><strong>Pattern</strong> is by default a normal string that is matched precisely (but without
8145 regard to case, except in the case of pickaxe). However, when you check the <em>re</em> checkbox,
8146 the pattern entered is recognized as the POSIX extended
8147 <a href="https://en.wikipedia.org/wiki/Regular_expression">regular expression</a> (also case
8148 insensitive).</p>
8149 <dl>
8150 <dt><b>commit</b></dt>
8151 <dd>The commit messages and authorship information will be scanned for the given pattern.</dd>
8152 EOT
8153         my $have_grep = gitweb_check_feature('grep');
8154         if ($have_grep) {
8155                 print <<EOT;
8156 <dt><b>grep</b></dt>
8157 <dd>All files in the currently selected tree (HEAD unless you are explicitly browsing
8158     a different one) are searched for the given pattern. On large trees, this search can take
8159 a while and put some strain on the server, so please use it with some consideration. Note that
8160 due to git-grep peculiarity, currently if regexp mode is turned off, the matches are
8161 case-sensitive.</dd>
8162 EOT
8163         }
8164         print <<EOT;
8165 <dt><b>author</b></dt>
8166 <dd>Name and e-mail of the change author and date of birth of the patch will be scanned for the given pattern.</dd>
8167 <dt><b>committer</b></dt>
8168 <dd>Name and e-mail of the committer and date of commit will be scanned for the given pattern.</dd>
8169 EOT
8170         my $have_pickaxe = gitweb_check_feature('pickaxe');
8171         if ($have_pickaxe) {
8172                 print <<EOT;
8173 <dt><b>pickaxe</b></dt>
8174 <dd>All commits that caused the string to appear or disappear from any file (changes that
8175 added, removed or "modified" the string) will be listed. This search can take a while and
8176 takes a lot of strain on the server, so please use it wisely. Note that since you may be
8177 interested even in changes just changing the case as well, this search is case sensitive.</dd>
8178 EOT
8179         }
8180         print "</dl>\n";
8181         git_footer_html();
8182 }
8183
8184 sub git_shortlog {
8185         git_log_generic('shortlog', \&git_shortlog_body,
8186                         $hash, $hash_parent);
8187 }
8188
8189 ## ......................................................................
8190 ## feeds (RSS, Atom; OPML)
8191
8192 sub git_feed {
8193         my $format = shift || 'atom';
8194         my $have_blame = gitweb_check_feature('blame');
8195
8196         # Atom: http://www.atomenabled.org/developers/syndication/
8197         # RSS:  http://www.notestips.com/80256B3A007F2692/1/NAMO5P9UPQ
8198         if ($format ne 'rss' && $format ne 'atom') {
8199                 die_error(400, "Unknown web feed format");
8200         }
8201
8202         # log/feed of current (HEAD) branch, log of given branch, history of file/directory
8203         my $head = $hash || 'HEAD';
8204         my @commitlist = parse_commits($head, 150, 0, $file_name);
8205
8206         my %latest_commit;
8207         my %latest_date;
8208         my $content_type = "application/$format+xml";
8209         if (defined $cgi->http('HTTP_ACCEPT') &&
8210                  $cgi->Accept('text/xml') > $cgi->Accept($content_type)) {
8211                 # browser (feed reader) prefers text/xml
8212                 $content_type = 'text/xml';
8213         }
8214         if (defined($commitlist[0])) {
8215                 %latest_commit = %{$commitlist[0]};
8216                 my $latest_epoch = $latest_commit{'committer_epoch'};
8217                 exit_if_unmodified_since($latest_epoch);
8218                 %latest_date = parse_date($latest_epoch, $latest_commit{'committer_tz'});
8219         }
8220         print $cgi->header(
8221                 -type => $content_type,
8222                 -charset => 'utf-8',
8223                 %latest_date ? (-last_modified => $latest_date{'rfc2822'}) : (),
8224                 -status => '200 OK');
8225
8226         # Optimization: skip generating the body if client asks only
8227         # for Last-Modified date.
8228         return if ($cgi->request_method() eq 'HEAD');
8229
8230         # header variables
8231         my $title = "$site_name - $project/$action";
8232         my $feed_type = 'log';
8233         if (defined $hash) {
8234                 $title .= " - '$hash'";
8235                 $feed_type = 'branch log';
8236                 if (defined $file_name) {
8237                         $title .= " :: $file_name";
8238                         $feed_type = 'history';
8239                 }
8240         } elsif (defined $file_name) {
8241                 $title .= " - $file_name";
8242                 $feed_type = 'history';
8243         }
8244         $title .= " $feed_type";
8245         $title = esc_html($title);
8246         my $descr = git_get_project_description($project);
8247         if (defined $descr) {
8248                 $descr = esc_html($descr);
8249         } else {
8250                 $descr = "$project " .
8251                          ($format eq 'rss' ? 'RSS' : 'Atom') .
8252                          " feed";
8253         }
8254         my $owner = git_get_project_owner($project);
8255         $owner = esc_html($owner);
8256
8257         #header
8258         my $alt_url;
8259         if (defined $file_name) {
8260                 $alt_url = href(-full=>1, action=>"history", hash=>$hash, file_name=>$file_name);
8261         } elsif (defined $hash) {
8262                 $alt_url = href(-full=>1, action=>"log", hash=>$hash);
8263         } else {
8264                 $alt_url = href(-full=>1, action=>"summary");
8265         }
8266         $alt_url = esc_attr($alt_url);
8267         print qq!<?xml version="1.0" encoding="utf-8"?>\n!;
8268         if ($format eq 'rss') {
8269                 print <<XML;
8270 <rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/">
8271 <channel>
8272 XML
8273                 print "<title>$title</title>\n" .
8274                       "<link>$alt_url</link>\n" .
8275                       "<description>$descr</description>\n" .
8276                       "<language>en</language>\n" .
8277                       # project owner is responsible for 'editorial' content
8278                       "<managingEditor>$owner</managingEditor>\n";
8279                 if (defined $logo || defined $favicon) {
8280                         # prefer the logo to the favicon, since RSS
8281                         # doesn't allow both
8282                         my $img = esc_url($logo || $favicon);
8283                         print "<image>\n" .
8284                               "<url>$img</url>\n" .
8285                               "<title>$title</title>\n" .
8286                               "<link>$alt_url</link>\n" .
8287                               "</image>\n";
8288                 }
8289                 if (%latest_date) {
8290                         print "<pubDate>$latest_date{'rfc2822'}</pubDate>\n";
8291                         print "<lastBuildDate>$latest_date{'rfc2822'}</lastBuildDate>\n";
8292                 }
8293                 print "<generator>gitweb v.$version/$git_version</generator>\n";
8294         } elsif ($format eq 'atom') {
8295                 print <<XML;
8296 <feed xmlns="http://www.w3.org/2005/Atom">
8297 XML
8298                 print "<title>$title</title>\n" .
8299                       "<subtitle>$descr</subtitle>\n" .
8300                       '<link rel="alternate" type="text/html" href="' .
8301                       $alt_url . '" />' . "\n" .
8302                       '<link rel="self" type="' . $content_type . '" href="' .
8303                       $cgi->self_url() . '" />' . "\n" .
8304                       "<id>" . esc_url(href(-full=>1)) . "</id>\n" .
8305                       # use project owner for feed author
8306                       "<author><name>$owner</name></author>\n";
8307                 if (defined $favicon) {
8308                         print "<icon>" . esc_url($favicon) . "</icon>\n";
8309                 }
8310                 if (defined $logo) {
8311                         # not twice as wide as tall: 72 x 27 pixels
8312                         print "<logo>" . esc_url($logo) . "</logo>\n";
8313                 }
8314                 if (! %latest_date) {
8315                         # dummy date to keep the feed valid until commits trickle in:
8316                         print "<updated>1970-01-01T00:00:00Z</updated>\n";
8317                 } else {
8318                         print "<updated>$latest_date{'iso-8601'}</updated>\n";
8319                 }
8320                 print "<generator version='$version/$git_version'>gitweb</generator>\n";
8321         }
8322
8323         # contents
8324         for (my $i = 0; $i <= $#commitlist; $i++) {
8325                 my %co = %{$commitlist[$i]};
8326                 my $commit = $co{'id'};
8327                 # we read 150, we always show 30 and the ones more recent than 48 hours
8328                 if (($i >= 20) && ((time - $co{'author_epoch'}) > 48*60*60)) {
8329                         last;
8330                 }
8331                 my %cd = parse_date($co{'author_epoch'}, $co{'author_tz'});
8332
8333                 # get list of changed files
8334                 open my $fd, "-|", git_cmd(), "diff-tree", '-r', @diff_opts,
8335                         $co{'parent'} || "--root",
8336                         $co{'id'}, "--", (defined $file_name ? $file_name : ())
8337                         or next;
8338                 my @difftree = map { chomp; $_ } <$fd>;
8339                 close $fd
8340                         or next;
8341
8342                 # print element (entry, item)
8343                 my $co_url = href(-full=>1, action=>"commitdiff", hash=>$commit);
8344                 if ($format eq 'rss') {
8345                         print "<item>\n" .
8346                               "<title>" . esc_html($co{'title'}) . "</title>\n" .
8347                               "<author>" . esc_html($co{'author'}) . "</author>\n" .
8348                               "<pubDate>$cd{'rfc2822'}</pubDate>\n" .
8349                               "<guid isPermaLink=\"true\">$co_url</guid>\n" .
8350                               "<link>" . esc_html($co_url) . "</link>\n" .
8351                               "<description>" . esc_html($co{'title'}) . "</description>\n" .
8352                               "<content:encoded>" .
8353                               "<![CDATA[\n";
8354                 } elsif ($format eq 'atom') {
8355                         print "<entry>\n" .
8356                               "<title type=\"html\">" . esc_html($co{'title'}) . "</title>\n" .
8357                               "<updated>$cd{'iso-8601'}</updated>\n" .
8358                               "<author>\n" .
8359                               "  <name>" . esc_html($co{'author_name'}) . "</name>\n";
8360                         if ($co{'author_email'}) {
8361                                 print "  <email>" . esc_html($co{'author_email'}) . "</email>\n";
8362                         }
8363                         print "</author>\n" .
8364                               # use committer for contributor
8365                               "<contributor>\n" .
8366                               "  <name>" . esc_html($co{'committer_name'}) . "</name>\n";
8367                         if ($co{'committer_email'}) {
8368                                 print "  <email>" . esc_html($co{'committer_email'}) . "</email>\n";
8369                         }
8370                         print "</contributor>\n" .
8371                               "<published>$cd{'iso-8601'}</published>\n" .
8372                               "<link rel=\"alternate\" type=\"text/html\" href=\"" . esc_attr($co_url) . "\" />\n" .
8373                               "<id>" . esc_html($co_url) . "</id>\n" .
8374                               "<content type=\"xhtml\" xml:base=\"" . esc_url($my_url) . "\">\n" .
8375                               "<div xmlns=\"http://www.w3.org/1999/xhtml\">\n";
8376                 }
8377                 my $comment = $co{'comment'};
8378                 print "<pre>\n";
8379                 foreach my $line (@$comment) {
8380                         $line = esc_html($line);
8381                         print "$line\n";
8382                 }
8383                 print "</pre><ul>\n";
8384                 foreach my $difftree_line (@difftree) {
8385                         my %difftree = parse_difftree_raw_line($difftree_line);
8386                         next if !$difftree{'from_id'};
8387
8388                         my $file = $difftree{'file'} || $difftree{'to_file'};
8389
8390                         print "<li>" .
8391                               "[" .
8392                               $cgi->a({-href => href(-full=>1, action=>"blobdiff",
8393                                                      hash=>$difftree{'to_id'}, hash_parent=>$difftree{'from_id'},
8394                                                      hash_base=>$co{'id'}, hash_parent_base=>$co{'parent'},
8395                                                      file_name=>$file, file_parent=>$difftree{'from_file'}),
8396                                       -title => "diff"}, 'D');
8397                         if ($have_blame) {
8398                                 print $cgi->a({-href => href(-full=>1, action=>"blame",
8399                                                              file_name=>$file, hash_base=>$commit),
8400                                               -title => "blame"}, 'B');
8401                         }
8402                         # if this is not a feed of a file history
8403                         if (!defined $file_name || $file_name ne $file) {
8404                                 print $cgi->a({-href => href(-full=>1, action=>"history",
8405                                                              file_name=>$file, hash=>$commit),
8406                                               -title => "history"}, 'H');
8407                         }
8408                         $file = esc_path($file);
8409                         print "] ".
8410                               "$file</li>\n";
8411                 }
8412                 if ($format eq 'rss') {
8413                         print "</ul>]]>\n" .
8414                               "</content:encoded>\n" .
8415                               "</item>\n";
8416                 } elsif ($format eq 'atom') {
8417                         print "</ul>\n</div>\n" .
8418                               "</content>\n" .
8419                               "</entry>\n";
8420                 }
8421         }
8422
8423         # end of feed
8424         if ($format eq 'rss') {
8425                 print "</channel>\n</rss>\n";
8426         } elsif ($format eq 'atom') {
8427                 print "</feed>\n";
8428         }
8429 }
8430
8431 sub git_rss {
8432         git_feed('rss');
8433 }
8434
8435 sub git_atom {
8436         git_feed('atom');
8437 }
8438
8439 sub git_opml {
8440         my @list = git_get_projects_list($project_filter, $strict_export);
8441         if (!@list) {
8442                 die_error(404, "No projects found");
8443         }
8444
8445         print $cgi->header(
8446                 -type => 'text/xml',
8447                 -charset => 'utf-8',
8448                 -content_disposition => 'inline; filename="opml.xml"');
8449
8450         my $title = esc_html($site_name);
8451         my $filter = " within subdirectory ";
8452         if (defined $project_filter) {
8453                 $filter .= esc_html($project_filter);
8454         } else {
8455                 $filter = "";
8456         }
8457         print <<XML;
8458 <?xml version="1.0" encoding="utf-8"?>
8459 <opml version="1.0">
8460 <head>
8461   <title>$title OPML Export$filter</title>
8462 </head>
8463 <body>
8464 <outline text="git RSS feeds">
8465 XML
8466
8467         foreach my $pr (@list) {
8468                 my %proj = %$pr;
8469                 my $head = git_get_head_hash($proj{'path'});
8470                 if (!defined $head) {
8471                         next;
8472                 }
8473                 $git_dir = "$projectroot/$proj{'path'}";
8474                 my %co = parse_commit($head);
8475                 if (!%co) {
8476                         next;
8477                 }
8478
8479                 my $path = esc_html(chop_str($proj{'path'}, 25, 5));
8480                 my $rss  = esc_attr(href('project' => $proj{'path'}, 'action' => 'rss', -full => 1));
8481                 my $html = esc_attr(href('project' => $proj{'path'}, 'action' => 'summary', -full => 1));
8482                 print "<outline type=\"rss\" text=\"$path\" title=\"$path\" xmlUrl=\"$rss\" htmlUrl=\"$html\"/>\n";
8483         }
8484         print <<XML;
8485 </outline>
8486 </body>
8487 </opml>
8488 XML
8489 }