#1034 dbtoepub doesn't scan sub-files and xincludes

[Raphael Hertzog]

dbtoepub doesn't cope well with docbook documents split over multiple files (either via DTD system entities or XInclude). The conversion to XHTML (by xsltproc) is fine but the images are not integrated in the zip file.

This is because REXML::Parsers::PullParser doesn't expand the sub-files and thus never scans all the graphics elements. Instead it returns the Xinclude tag or the entity and those are skipped by get_image_refs().

[Raphael Hertzog]

Test case with "dbtoepub test.xml"

[Keith Fahlgren]

A fix for this issue has been added to the current codebase.
Please test the fix with the latest snapshot from:

http://docbook.sourceforge.net/snapshots/

[Keith Fahlgren]

This should be resolved in 1.75.0. Please test and let me know if you're not happy with the resolution.

r8424 | abdelazer | 2009-05-06 18:03:07 -0700 (Wed, 06 May 2009) | 1 line
Changed paths:
M /trunk/xsl/epub/bin/lib/docbook.rb
M /trunk/xsl/epub/bin/spec/epub_regressions_spec.rb
A /trunk/xsl/epub/bin/spec/files/onegraphic.entity.xml
A /trunk/xsl/epub/bin/spec/files/onegraphic.xinclude.xml
M /trunk/xsl/epub/bin/spec/files/orm.book.001.xml
A /trunk/xsl/epub/bin/spec/files/xincludeents.xml
M /trunk/xsl/epub/docbook.xsl

Slightly more nuanced handling of imageobject alternatives and better support in dbtoepub for XIncludes and ENTITYs to resolve Issue #2750442

[Keith Fahlgren]

Had to refine the order in which the collapsing was done. Will ask for more testing.

[Joachim Breitner]

I believe that this bug is not fully fixed yet. If the main file ./foo.xml includes ./parts/part1.xml, which in turn tries to include an image img.png, then dbtoepub wrongly looks in ./img.png instead of ./parts/img.png. It works correctly when the xinclude-resolving is left to xsltproc when generating HTML, as it considers the basedir of included xml files.