Bob Stayton [Thu, 5 Mar 2009 23:34:58 +0000 (23:34 +0000)]
Fix bug 1546008: in qandaentry in a TOC, use its blockinfo/titleabbrev or blockinfo/title
instead of question, if available. For DocBook 5, use the info versions.
Bob Stayton [Thu, 5 Mar 2009 23:33:56 +0000 (23:33 +0000)]
Fix bug 1546008: in qandaentry in a TOC, use its blockinfo/titleabbrev or blockinfo/title
instead of question, if available. For DocBook 5, use the info versions.
Keith Fahlgren [Fri, 20 Feb 2009 13:57:14 +0000 (13:57 +0000)]
The DocBook 1.74.* had problems with different XSLT processors.
Add instructions to release instructions and make Makefiles more
flexible to allow 'make check' to *actually* be run with something
other than xsltproc. Thanks to Sorin Ristache for reporting this
issue.
Because of problems with the 1.74.* releases, it is now
recommended that "make check" be run multiple times with
different XSLT processors.
One recommended way to invoke "make check" is the following:
After running with "xsltproc", you should run it again with
"saxon":
rm -f DOCBOOK-BUILD.LOG && \
. ~/docbk.sh && \
make check 2>&1 \
XSLTENGINE=saxon \
| tee DOCBOOK-BUILD.LOG && \
$DOCBOOK_SVN/buildtools/build-check DOCBOOK-BUILD.LOG
Keith Fahlgren [Tue, 17 Feb 2009 04:03:20 +0000 (04:03 +0000)]
Liza Daly reported that the dc:identifer-generation code was garbage (she was right).
Added new tests:
- should include at least one dc:identifier
- should include an ISBN as URN for dc:identifier if an ISBN was in the metadata
- should include an ISSN as URN for dc:identifier if an ISSN was in the metadata
- should include an biblioid as a dc:identifier if an biblioid was in the metadata
- should include a URN for a biblioid with @class attribute as a dc:identifier if an biblioid was in the metadata
Keith Fahlgren [Wed, 11 Feb 2009 20:58:08 +0000 (20:58 +0000)]
An XSLT2 stylehseet to obfuscate the text of a DocBook document so that you can submit it as a failing test case (probably because of unhandled markup)
Michael Smith [Mon, 9 Feb 2009 19:56:29 +0000 (19:56 +0000)]
Put the following at the top of generated roff for each page:
\" t
purpose is to explicitly tell AT&T troff that the page needs to be
pre-processed through tbl(1); groff can figure it out
automatically, but apparently AT&T troff needs to be explicitly told
Michael Smith [Mon, 9 Feb 2009 16:22:14 +0000 (16:22 +0000)]
This change closes bug 2013441 and makes man-page output generally
usable again on Solaris and other platforms that use AT&T troff
instead of groff.
The change was to add man.output.better.ps.enabled parameter, with
its default value set to zero.
If the value of the man.output.better.ps.enabled parameter is
non-zero, certain markup is embedded in each generated man page
such that PostScript output from the man -Tps command for that
page will include a number of enhancements designed to improve the
quality of that output.
If man.output.better.ps.enabled is zero (the default), no such
markup is embedded in generated man pages, and no enhancements are
included in the PostScript output generated from those man pages
by the man -Tps command.
WARNING: The enhancements provided by this parameter rely on
features that are specific to groff (GNU troff) and that are not
part of "classic" AT&T troff or any of its derivatives. Therefore,
any man pages you generate with this parameter enabled will be
readable only on systems on which the groff (GNU troff) program is
installed, such as GNU/Linux systems. The pages will not not be
readable on systems on with the classic troff (AT&T troff) command
is installed.
NOTE: The value of this parameter only affects PostScript output
generated from the man command. It has no effect on output
generated using the FO backend.
TIP: You can generate PostScript output for any man page by
running the following command:
man FOO -Tps > FOO.ps
You can then generate PDF output by running the following command:
Keith Fahlgren [Fri, 6 Feb 2009 04:57:22 +0000 (04:57 +0000)]
Improve both SEO and semantic structure of chunked ePub output by ensuring that
we always send out one and only one h1 in each XHTML chunk.
DocBook::Epub
- should include one and only one <h1> in each HTML file in rendered ePub files
for <book>s
- should include one and only one <h1> in each HTML file in rendered ePub files
for <book>s even if they do not have section markup
Dongsheng Song [Thu, 25 Dec 2008 02:13:45 +0000 (02:13 +0000)]
Simplified Chinese translation.
* defguide5/zh/stylesheets/fo.xsl,defguide5/zh/stylesheets/zh/fo.xsl,
defguide/zh/stylesheets/fo.xsl,defguide/zh/stylesheets/zh/fo.xsl: change symbol font.
* gentext/locale/zh_cn.xml: Use publishing term instead.
Bob Stayton [Wed, 1 Oct 2008 00:11:49 +0000 (00:11 +0000)]
Fix two bugs in info.reference template:
- mediaobject omitted from list of info containers.
- replace [1] with [position() = last()] in the target predicate because
when parens are used, the ancestor node list is in document order
instead of closest ancestor order.
Added a new top-level stylesheet module, chunk-changebars.xsl, to be
used for generating chunked output with highlighting based on change
(@revisionflag) markup. The module imports/includes the standard chunking
and changebars templates and contains additional logic for chunked output.
See FRs #1015180 and #1819915.
Modified declaration of the "col" parameter in the template with
match="entry|entrytbl". This is a workaround for bug #1816373.
The parameter was reset to 1 for cells with @revisionflag when
using changebars.xsl, which resulted in malformed output.
projects / docbook-dsssl / log