This is useful when using ⛩️📰 书社 directly as a static site
generator to provide archive downloads (archives are not compressed; it
is assumed that they will be gzipped over the wire). This requires
a recursive call to Make for each archive file, which performs the
following steps :—
- Extracts all of the elements slated for archiving into separate
files.
- Restarts.
- Processes the resulting extracted files and then archives them.
The extraction step in particular is somewhat convoluted; it requires
dynamically generating a transform which has the appropriate
`<exsl:document>` elements for a given source file, and then applying
that transform in a second call to `xsltproc`.
X·M·L outputs are now passed through an extra call to `xmllint` to
remove any unnecessary namespace attributes instead of just symlinked;
the symlinks weren’t compatible with archiving anyway.