2 SPDX-FileCopyrightText: 2024 Lady <https://www.ladys.computer/about/#lady>
3 SPDX-License-Identifier: CC0-1.0
7 <b>A make·file for X·M·L.</b>
9 <dfn>⛩️📰 书社</dfn> aims to make it easy to generate websites with
10 X·S·L·T and G·N·U Make.
11 It is consequently only a good choice for people who like X·S·L·T and
12 G·N·U Make and wish it were easier to make websites with them.
14 It makes things easier by :—
16 - Automatically identifying source files and characterizing them by
17 type (X·M·L, text, or asset).
19 - Parsing supported text types into X·M·L trees.
21 - Enabling easy inclusion of source files within each other.
23 It aims to do this with zero dependencies beyond the programs already
24 installed on your computer†.
26 † Assuming an operating system with a fairly featureful, and
27 Posix‐compliant, development setup (e·g, macOS).
28 In fact, on Linux you will probably need to install a few programs:
29 `libxml2-utils`, `xsltproc`, `sharutils`, and `pax`.
33 <i lang="cmn-Hans">书社</i> is a Chinese word meaning “publishing
36 The first character, <i lang="cmn-Hans">书</i>, is the simplified form
39 The second character, <i lang="cmn-Hans">社</i>, contemporarily means
40 “association”, but historically referred to the god of the soil and
41 related altars or festivities.
42 In Japanese, it is an alternate spelling for <i lang="ja">やしろ</i>,
43 the word for “Shinto shrine”.
45 The name <i lang="cmn-Hans">书社</i> was chosen to play on this pun, as
46 it is intended as a publishing program for webshrines.
48 In Ascii environments, ⛩️📰 书社 should be written `Shushe`, following
49 the pinyin transliteration.
53 In most cases, ⛩️📰 书社 aims to require only functionality which is
54 present in all Posix‐compliant operating systems.
55 There are a few exceptions.
56 Details on particular programs are given below; if a program is not
57 listed, it is assumed that any Posix‐compliant implementation will
62 This is a Posix utility, but ⛩️📰 书社 currently depends on
63 unspecified behaviour.
64 When the G·N·U version of `stat` is being used, then the G·N·U version
65 of `date` is also expected.
69 This is a Posix utility, but ⛩️📰 书社 currently depends on
70 unspecified behaviour.
71 It requires support for the following additional options :—
73 - **`-C`**, when supplied with `-m`, must be useable to compile a
74 `.mgc` magicfile for use with future invocations of `file`.
76 - **`--files-from`** must be useable to provide a file that `file`
77 should read file·names from, and `-` must be useable in this
78 context to specify the standard input.
80 - **`--mime-type`** must cause `file` to print the internet media type
81 of the file with no charset parameter.
83 - **`--separator`** must be useable to set the separator that `file`
84 uses to separate file names from types.
86 These options are implemented by the
87 [Fine Free File Command](https://darwinsys.com/file/), which is used
88 by most operating systems.
92 This is not a Posix utility.
93 Usage of `git` is optional, but recommended (and activated by default).
94 To disable it, set `GIT=`.
98 This is a Posix utility, but ⛩️📰 书社 currently depends on
99 unspecified behaviour.
100 ⛩️📰 书社 requires specifically the G·N·U version of `make`, and
101 depends on functionality present in version 3.81 or later.
102 It is not expected to work in previous versions, or with other
103 implementations of Make.
107 This is a Posix utility, but not included in the Linux Standard Base or
108 installed by default in many distributions.
109 Only `ustar` format support is required.
113 This is not a Posix utility, and nor is it particularly portable.
114 To get around incompatibilities, ⛩️📰 书社 attempts to recognize G·N·U
115 `stat` by searching for the string `GNU` when invoked with the
116 `--version` option, and falls back to B·S·D behaviour otherwise.
118 ### `uudecode` and `uuencode`
120 These are Posix utilities, but not included in the Linux Standard Base
121 or installed by default in many distributions.
122 The G·N·U [Sharutils](https://www.gnu.org/software/sharutils/) package
123 can be installed to access them.
127 This is a Posix utility, but ⛩️📰 书社 currently depends on
128 unspecified behaviour.
129 It requires support for the `-0` flag, which must disable the special
130 quote and whitespace handling of `xargs` in favour of null‐terminated
133 ### `xmlcatalog` and `xmllint`
135 These are not a Posix utilities.
136 They is a part of `libxml2`, but may need to be installed separately
137 (e·g by the name `libxml2-utils`).
141 This is not a Posix utility.
142 It is a part of `libxslt`, but may need to be installed separately.
146 Place source files in `sources/` and run `make install` to compile
147 the result to `public/`.
148 Compilation involves the following steps :—
150 1. ⛩️📰 书社 compiles all of the magic files in `magic/` into a single
151 file, `build/magic.mgc`.
153 2. ⛩️📰 书社 processes all of the parsers in `parsers/` and determines
154 the list of supported plaintext types.
156 3. ⛩️📰 书社 identifies all of the source files and includes and uses
157 `build/magic.mgc` to classify them by media type.
159 4. ⛩️📰 书社 parses all plaintext and X·M·L source files and includes
160 and then builds a dependency tree between them.
162 5. ⛩️📰 书社 uses the dependency tree to establish prerequisites for
165 6. ⛩️📰 书社 compiles each output file to `build/result`.
167 7. ⛩️📰 书社 copies most output files from `build/result` to
168 `build/public`, but it does some additional processing instead on
169 those which indicate a non‐X·M·L desired final output form.
171 8. ⛩️📰 书社 copies the final resulting files to `public`.
173 You can use `make list` to list each identified source file or include
174 alongside its computed type and dependencies.
175 As this is a Make‐based program, steps will only be run if the
176 corresponding buildfile or output file is older than its
181 The ⛩️📰 书社 name·space is `urn:fdc:ladys.computer:20231231:Shu1She4`.
183 This document uses a few name·space prefixes, with the following
186 | Prefix | Expansion |
187 | ---------: | :-------------------------------------------- |
188 | `catalog:` | `urn:oasis:names:tc:entity:xmlns:xml:catalog` |
189 | `exsl:` | `http://exslt.org/common` |
190 | `exslstr:` | `http://exslt.org/strings` |
191 | `html:` | `http://www.w3.org/1999/xhtml` |
192 | `svg:` | `http://www.w3.org/2000/svg` |
193 | `xlink:` | `http://www.w3.org/1999/xlink` |
194 | `xslt:` | `http://www.w3.org/1999/XSL/Transform` |
195 | `书社:` | `urn:fdc:ladys.computer:20231231:Shu1She4` |
197 ## Setup and Configuration
199 ⛩️📰 书社 depends on the following programs to run.
200 In every case, you may supply your own implementation by overriding the
201 corresponding (allcaps) variable (e·g, set `MKDIR` to supply your own
202 `mkdir` implementation).
212 - `git` (optional; set `GIT=` to disable)
219 - `pax` (only when generating archives)
224 - `stat` (BSD *or* GNU)
230 - `xargs` (requires support for `-0`)
231 - `xmlcatalog` (provided by `libxml2`)
232 - `xmllint` (provided by `libxml2`)
233 - `xsltproc` (provided by `libxslt`)
235 The following additional variables can be used to control the behaviour
239 The location of the source files (default: `sources`).
240 Multiple source directories can be provided, so long as the same
241 file subpath doesn’t exist in more than one of them.
244 The location of source includes (default: `sources/includes`).
245 This can be inside of `SRCDIR`, but needn’t be.
246 Multiple include directories can be provided, so long as the same
247 file subpath doesn’t exist in more than one of them.
250 The location of the (temporary) build directory (default: `build`).
251 `make clean` will delete this, and it is recommended that it not be
252 used for programs aside from ⛩️📰 书社.
255 The location of directory to output files to (default: `public`).
256 `make install` will overwrite files in this directory which
257 correspond to those in `SRCDIR`.
258 It *will not* touch other files, including those generated from files
259 in `SRCDIR` which have since been deleted.
261 Files are first compiled to `$(BUILDDIR)/public` before they are
262 copied to `DESTDIR`, so this folder is relatively quick and
263 inexpensive to re·create.
264 It’s reasonable to simply delete it before every `make install` to
265 ensure stale content is removed.
268 The location of the ⛩️📰 书社 `GNUmakefile`.
269 This should be set automatically when calling Make and shouldn’t ever
270 need to be set manually.
271 This variable is used to find the ⛩️📰 书社 `lib/` folder, which is
272 expected to be in the same location.
275 A white·space‐separated list of magic files to use (default:
276 `$(THISDIR)/magic/*`).
279 The value of this variable is appended to `MAGIC` by default, to
280 enable additional magic files without overriding the existing ones.
283 Rules to use with `find` when searching for source files.
284 The default ignores files that start with a period or hyphen‐minus,
285 those which end with a cloparen, and those which contain a hash,
286 buck, percent, asterisk, colon, semi, eroteme, bracket, backslash,
289 - **`EXTRAFINDRULES`:**
290 The value of this variable is appended to `FINDRULES` by default, to
291 enable additional rules without overriding the existing ones.
293 - **`FINDINCLUDERULES`:**
294 Rules to use with `find` when searching for includes (default:
297 - **`EXTRAFINDINCLUDERULES`:**
298 The value of this variable is appended to `FINDINCLUDERULES` by
299 default, to enable additional rules without overriding the existing
303 A white·space‐separated list of parsers to use (default:
304 `$(THISDIR)/parsers/*.xslt`).
306 - **`EXTRAPARSERS`:**
307 The value of this variable is appended to `PARSERS` by default, to
308 enable additional parsers without overriding the existing ones.
311 A white·space‐separated list of transforms to use (default:
312 `$(THISDIR)/transforms/*.xslt`).
314 - **`EXTRATRANSFORMS`:**
315 The value of this variable is appended to `TRANSFORMS` by default, to
316 enable additional transforms without overriding the existing ones.
319 A white·space‐separated list of media types to consider X·M·L
320 (default: `application/xml text/xml`).
323 The current version of ⛩️📰 书社 (default: derived from the current
324 git tag/branch/commit).
327 The current version of the source files (default: derived from the
328 current git tag/branch/commit).
331 If this variable has a value, every recipe instruction will be
332 printed when it runs (default: empty).
333 This is helpful for debugging, but typically too noisy for general
338 Source files may be placed in `SRCDIR` in any manner; the file
339 structure used there will match the output.
340 The type of source files is *not* determined by file extension, but
341 rather by magic number; this means that files **must** begin with
342 something recognizable.
343 Supported magic numbers include :—
345 - `<?xml` for `application/xml` files
346 - `#!js` for `text/javascript` files
347 - `@charset "` for `text/css` files
348 - `#!tsv` for `text/tab-separated-values` files
349 - `%%` for `text/record-jar` files (unregistered; see
350 [[draft-phillips-record-jar-01][]])
352 Text formats with associated X·S·L·T parsers are wrapped in a H·T·M·L
353 `<script>` element whose `@type` gives its media type, and then
354 passed to the parser to process.
355 Source files whose media type does not have an associated X·S·L·T
356 parser are considered “assets” and will not be transformed.
358 **☡ For compatibility with this program, source file·names must not
359 contain Ascii white·space, colons (`:`), semis (`;`), pipes (`|`),
360 bucks (`$`), percents (`%`), hashes (`#`), asterisks (`*`), brackets
361 (`[` or `]`), erotemes (`?`), backslashes (`\`), or control
362 characters, must not begin with a hyphen‐minus (`-`), must not end
363 with a cloparen (`)`), and must not contain quoted braces (`"{` or
365 The former characters have the potential to conflict with make syntax,
366 a leading hyphen‐minus is confusable for a command‐line argument, a
367 trailing cloparen [activates a bug in G·N·U Make
368 3.81](https://stackoverflow.com/questions/17148468/capturing-filenames-including-parentheses-with-gnu-makes-wildcard-function#comment24825307_17148894),
369 and quoted braces are used internally by the program.
373 Parsers are used to convert plaintext files into X·M·L trees, as well
374 as convert plaintext formats which are already included inline in
375 existing source X·M·L documents.
376 ⛩️📰 书社 comes with some parsers; namely :—
378 - **`parsers/plain.xslt`:**
379 Wraps `text/plain` contents in a `<html:pre>` element.
381 - **`parsers/record-jar.xslt`:**
382 Converts `text/record-jar` contents into a `<html:div>` of
383 `<html:dl>` elements (one for each record).
385 - **`parsers/tsv.xslt`:**
386 Converts `text/tab-separated-values` contents into an `<html:table>`
389 New ⛩️📰 书社 parsers which target plaintext formats should have an
390 `<xslt:template>` element with no `@name` or `@mode` and whose
393 - Starts with an appropriately‐name·spaced qualified name for a
394 `<html:script>` element.
396 - Follows this with the string `[@type=`.
398 - Follows this with a quoted string giving a media type supported by
400 Media type parameters are *not* supported.
402 - Follows this with the string `]`.
404 For example, the trivial `text/plain` parser is defined as follows :—
407 <?xml version="1.0"?>
409 xmlns="http://www.w3.org/1999/XSL/Transform"
410 xmlns:html="http://www.w3.org/1999/xhtml"
411 xmlns:书社="urn:fdc:ladys.computer:20231231:Shu1She4"
414 <书社:id>example:text/plain</书社:id>
415 <template match="html:script[@type='text/plain']">
416 <html:pre><value-of select="."/></html:pre>
421 ⛩️📰 书社 will scan the provided parsers for this pattern to determine
422 the set of allowed plaintext file types.
423 Multiple such `<xslt:template>` elements may be provided in a single
424 parser, for example if the parser supports multiple media types.
425 Alternatively, you can set the `@书社:supported-media-types` attribute
426 on the root element of the parser to override media type support
429 Even when `@书社:supported-media-types` is set, it is a requirement
430 that each parser transform any `<html:script>` elements with a
431 `@type` which matches their registered types into something else.
432 Otherwise the parser will be stuck in an endless loop.
433 The result tree of applying the transform to the `<html:script>`
434 element will be reparsed (in case any new `<html:script>` elements
435 were added in its subtree), and a `@书社:parsed-by` attribute will be
436 added to each toplevel element in the result.
437 The value of this attribute will be the value of the `<书社:id>`
438 toplevel element in the parser.
440 It is possible for parsers to support zero plaintext types.
441 This is useful when targeting specific dialects of X·M·L; parsers in
442 this sense operate on the same basic principles as transforms
444 The major distinction between X·M·L parsers and transforms is where in
445 the process the transformation happens:
446 Parsers are applied *prior* to embedding (and can be used to generate
447 embeds); transforms are applied *after*.
449 It is **strongly recommended** that auxillary templates in parsers be
450 name·spaced (by `@name` or `@mode`) whenever possible, to avoid
451 conflicts between parsers.
453 ### Attributes added during parsing
455 ⛩️📰 书社 will add a few attributes to the output of the parsing step,
458 - A `@书社:cksum` attribute on toplevel result elements, giving the
459 `cksum` checksum of the corresponding source file.
461 - For the elements which result from parsing plaintext `<html:script>`
464 - A `@书社:parsed-by` attribute, giving a space‐separated list of
465 parsers which parsed the node.
466 (Generally, this will be a list of one, but it is possible for the
467 result of a parse to be another plaintext node, which may be
468 parsed by a different parser.)
470 - A `@书社:media-type` attribute, giving the identified media type of
475 Documents can be embedded in other documents using a `<书社:link>`
476 element with `@xlink:show="embed"`.
477 The `@xlink:href`s of these elements should have the format
478 `about:shushe?source=<path>`, where `<path>` provides the path to the
479 file within `SRCDIR`.
480 Includes, which do not generate outputs of their own but may still be
481 freely embedded, instead use the format
482 `about:shushe?include=<path>`, where `<path>` provides the path
485 Embeds are replaced with the parsed contents of a file, unless the file
486 is an asset, in which case an `<html:object>` element is produced
487 instead (with the contents of the asset file provided as a base64
489 Embed replacements will be given a `@书社:identifier` attribute whose
490 value will match the `@xlink:href` of the embed.
492 Embedding takes place after parsing but before transformation, so
493 parsers are able to generate their own embeds.
494 ⛩️📰 书社 is able to detect the transitive embed dependencies of files
495 and update them accordingly; it will signal an error if the
496 dependencies are recursive.
498 ## Output Redirection
500 By default, ⛩️📰 书社 installs files to the same location in `DESTDIR`
501 as they were placed in their `SRCDIR`.
502 This behaviour can be customized by setting the `@书社:destination`
503 attribute on the root element, whose value can give a different path.
504 This attribute is read after parsing, but before transformation (where
505 it is silently dropped).
509 Transforms are used to convert X·M·L files into their final output,
510 after all necessary parsing and embedding has taken place.
511 ⛩️📰 书社 comes with some transforms; namely :—
513 - **`transforms/asset.xslt`:**
514 Converts `<html:object>` elements which correspond to recognized
515 media types into the appropriate H·T·M·L elements, and deletes
516 `<html:style>` elements from the body of the document and moves
519 - **`transforms/metadata.xslt`:**
520 Provides basic `<html:head>` metadata.
521 This metadata is generated from `<html:meta>` elements with one of
522 the following `@itemprop` attributes :—
524 - **`urn:fdc:ladys.computer:20231231:Shu1She4:title`:**
525 Provides the title of the page.
527 ⛩️📰 书社 automatically encapsulates H·T·M·L embeds so that their
528 metadata does not propogate up to the embedding document.
529 To undo this behaviour, remove the `@itemscope` and `@itemtype`
530 attributes from the embed during the transformation phase.
532 - **`transforms/serialization.xslt`:**
533 Replaces `<书社:serialize-xml>` elements with the (escaped)
534 serialized X·M·L of their contents.
535 This replacement happens during the application phase, after most
536 other transformations have taken place.
538 If a `@with-namespaces` attribute is provided, any name·space nodes
539 on the toplevel serialized elements whose U·R·I’s correspond to the
540 definitions of the provided prefixes, as defined for the
541 `<书社:serialize-xml>` element, will be declared using name·space
542 attributes on the serialized elements.
543 Otherwise, only name·space nodes which _differ_ from the definitions
544 on the `<书社:serialize-xml>` element will be declared.
545 The string `#default` may be used to represent the default
547 Multiple prefixes may be provided, separated by white·space.
549 When it comes to name·spaces used internally by ⛩️📰 书社, the
550 prefix used by ⛩️📰 书社 may be declared _in addition to_ the
551 prefix(es) used in the source document(s).
552 It is not possible to selectively only declare one prefix for a
553 name·space to the exclusion of others.
555 `<书社:raw-output>` elements may be used inside of
556 `<书社:serialize-xml>` elements to inject raw output into the
559 The following are recommendations on effective creation of
562 - Make template matchers as specific as possible.
563 It is likely an error if two transforms have templates which match
564 the same element (unless the templates have different priority).
566 - Name·space templates (with `@name` or `@mode`) whenever possible.
568 - Set `@exclude-result-prefixes` on the root `xslt:transform` element
569 to reduce the number of declared name·spaces in the final result.
573 The following params are made available globally in parsers and
580 The checksum of the source file (⅌ `cksum`).
583 The ⛩️📰 书社 identifier of the source file (a u·r·i beginning with
587 The value of the `SRCREV` variable (if present).
590 The time at which the source file was last modified.
593 The value of the `THISREV` variable (if present).
595 The following params are only available in transforms :—
598 The path of the catalog file (within `BUILDDIR`).
601 The path of the output file (within `DESTDIR`).
605 ⛩️📰 书社 will wrap the final output of the transforms in appropriate
606 `<html:html>` and `<html:body>` elements, so it is not necessary for
607 transforms to do this explicitly.
608 After performing the initial transform, ⛩️📰 书社 will match the root
609 node of the result in the following modes to fill in areas of the
613 The result of matching in this mode is prepended into the
614 `<html:body>` of the output (before the transformation result).
617 The result of matching in this mode is appended into the
618 `<html:body>` of the output (after the transformation result).
621 The result of matching in this mode is inserted into the
622 `<html:head>` of the output.
624 In addition to being called with the transform result, each of these
625 modes will additionally be called with a `<xslt:include>` element
626 corresponding to each transform.
627 If a transform has a `<书社:id>` top‐level element whose value is an
628 i·r·i, its `<xslt:include>` element will have a corresponding
630 This mechanism can be used to allow transforms to insert content
631 without matching any elements in the result; for example, the
632 following transform adds a link to a stylesheet to the `<html:head>`
636 <?xml version="1.0"?>
638 xmlns="http://www.w3.org/1999/XSL/Transform"
639 xmlns:html="http://www.w3.org/1999/xhtml"
640 xmlns:xslt="http://www.w3.org/1999/XSL/Transform"
641 xmlns:书社="urn:fdc:ladys.computer:20231231:Shu1She4"
642 exclude-result-prefixes="书社"
645 <书社:id>example:add-stylesheet-links.xslt</书社:id>
646 <template match="xslt:include[@书社:id='example:add-stylesheet-links.xslt']" mode="书社:metadata">
647 <html:link rel="stylesheet" type="text/css" href="/style.css"/>
652 Output wrapping can be entirely disabled by adding a
653 `@书社:disable-output-wrapping` attribute to the top‐level element in
655 It will not be performed on outputs whose root elements are
656 `<书社:archive>`, `<书社:base64-binary>`, or `<书社:raw-text>`
659 ## Applying Attributes
661 The `<书社:apply-attributes>` element will apply any attributes on the
662 element to the element(s) it wraps.
663 It is especially useful in combination with embeds.
665 The `<书社:apply-attributes-to-root>` element will apply any attributes
666 on the element to the root node of the final transformation result.
667 It is especially useful in combination with output wrapping.
669 In both cases, attributes from various sources are combined with
670 white·space between them.
671 Attribute application takes place after all ordinary transforms have
674 Both elements ignore attributes in the `xml:` name·space, except for
675 `@xml:lang`, which ignores all but the first definition (including
676 any already present on the root element).
677 On H·T·M·L and S·V·G elements, `@lang` has the same behaviour as
680 ## Other Kinds of Output
682 There are a few special elements in the `书社:` name·space which, if
683 they appear as the toplevel element in a transformation result, cause
684 ⛩️📰 书社 to produce something other than an X·M·L file.
687 - **`<书社:archive>`:**
688 Each child element with a `@书社:archived-as` attribute will be
689 archived as a separate file in a resulting tarball (this attribute
690 gives the file name).
691 These elements will be processed the same as the root elements of any
692 other file (e·g, they will be wrapped; they can themselves specify
693 non X·M·L output types, ⁊·c).
694 Other child elements will be ignored.
696 - **`<书社:base64-binary>`:**
697 The text nodes in the transformation result will, after removing all
698 Ascii whitespace, be treated as a Base·64 string, which is then
701 - **`<书社:raw-text>`:**
702 A plaintext (U·T·F‐8) file will be produced from the text nodes in
703 the transformation result.
707 It is possible to have a single source file produce multiple output
708 files via `<书社:page>` elements, whose `@name` gives the name of the
710 If a parsed document has a `@书社:destination` which contains `%s`,
711 the `%s` will be replaced with the `@name` for each `<书社:page>` (and
712 removed for the main output).
713 Otherwise, the `@name` is inserted before the first period of the
714 filename (or at the end of the filename for those with no period).
715 If `<书社:page>`s do not have a `@name`, they are numbered
717 The destination of pages must be in the same directory as their parent.
719 Pagination essentially forms a limited convenience for the more
720 sophisticated technique of creating an archive with ⛩️📰 书社 and
722 Pages are, from Make’s point of view, untracked side·effects of
723 installing the main output, meaning they cannot be targeted directly
724 and will not appear in `make list` or `make listout`.
725 They are intended solely for the like of indices and feeds, for which
726 convenience and necessity outweigh their flaws.
730 This repository conforms to [REUSE][].
732 Most source files are licensed under the terms of the <cite>Mozilla
733 Public License, version 2.0</cite>.
735 [REUSE]: <https://reuse.software/spec/>
736 [draft-phillips-record-jar-01]: <https://datatracker.ietf.org/doc/html/draft-phillips-record-jar-01>