Added bunch of new parameters (attribute sets) that affect list presentation: list.block.properties, itemizedlist.properties, orderedlist.properties, itemizedlist.label.properties and orderedlist.label.properties. Default behaviour of stylesheets has not been changed but further customizations will be much more easier.
Michael Smith [Tue, 20 Sep 2005 11:30:45 +0000 (11:30 +0000)]
Fixed handling of groff font requests output for Refsect1/Title
instances. Closes #1296252. Thanks to Paul DuBois for reporting.
Problem:
Instances of \\FB (instead of \\fB) in output for
Refsect1/Title.
Cause:
All content of Refsect1/Title gets uppercased (including any groff
requests that end up in the output.
Fix:
Added s/\\FB/\fB/, etc., matches to value of man.string.subst.map
param. (man.string.subst.map is basically used for "cleaning up"
borked groff markup output from initial formatting pass).
Affects:
This affects output of titles for all Refsect1-level headings in
manpages output.
Michael Smith [Tue, 20 Sep 2005 05:42:26 +0000 (05:42 +0000)]
Made some minor changes to get deal with a couple corner cases.
If somebody (say, Mike) checks in the VERSION file twice with the
same change description, it's possible for the merged changelog to
end with up something like:
* VERSION, VERSION: Version 1.69.1 released.
The existing version of mergechangelogs got confused by that, so I
added a regex match to catch instances like that.
Also, in running the snapshot builds, with the way I have the
snapshot build currently set up, it's useful for the build to just
automatically assume we are at version X.XX.1 if the previous
version was X.XX.0.
----------------------------------------------------------------------
You can enable crop marks in your document by setting crop.marks=1 and xep.extensions=1.
Appearance of crop marks can be controlled by parameters crop.mark.bleed (6pt), crop.mark.offset (24pt) and crop.mark.width (0.5pt).
Also there is new named template called user-xep-pis. You can overwrite it in order to produce some PIs that can control XEP as described in http://www.renderx.com/reference.html#Output_Formats
Michael Smith [Fri, 9 Sep 2005 03:58:58 +0000 (03:58 +0000)]
Added colon to list of delimiters for parsing datetime strings
into tokens. Closes #1285053. Thanks to Jeroen Ruigrok for the bug
report, and to Mauritz Jeanson for spotting the cause.
Michael Smith [Fri, 2 Sep 2005 06:50:52 +0000 (06:50 +0000)]
Removed catalog.xml from version control and updated build to
automatically generate it. (To ensure it always contains the
correct version number for the release.) Thanks to John L. Clark
for the suggestion.
Bob Stayton [Tue, 16 Aug 2005 19:26:31 +0000 (19:26 +0000)]
Fixed bug in char attribute, where text-align was being set to
any char value, even if the align attribute was not 'char'.
The CALS spec says text-align set to the value of char only
if the align attribute is 'char'.
Michael Smith [Thu, 11 Aug 2005 05:29:15 +0000 (05:29 +0000)]
Set default value of annotation.support param to 0 (off). Also,
wrapped contents of the apply-annotations template in a "if"
statement, so that those contents get evaluated only if
annotation.support is non-zero.
The reason for these changes was a report from Doug du Boulay (on
the docbook-apps list) that processing of the apply-annotations
template was consuming 75% of the total processing time for a
2.5Mb file. Anatoly Techtonik followed with a comment that with
1.69.0, "Compiling took about full working day on 12Mb file while
earlier it was about 2-4 hours".
Michael Smith [Tue, 9 Aug 2005 09:11:02 +0000 (09:11 +0000)]
Output .sp macro in mixed blocks where we were outputting
completely blank line before. Thanks (again) to Costin Stroie for
pointing out the problems and providing a test case.
Also, corrected the test for checking first preceding siblings of
text nodes in mixed blocks (the test is used for determining
whether or not we need to insert a line break before those nodes.)
Michael Smith [Fri, 5 Aug 2005 08:09:50 +0000 (08:09 +0000)]
Fixed bug that caused literal .sp to show up in output. (Thanks to
Costin Stroie for reporting).
WARNING: One of the consequences of this change is that the
stylesheet no longer generates spaces at the end of certain lines.
I had it doing that as a sort of hack to prevent certain kinds of
problems in output. It seems like the hack my no longer be
necessary. But if I am wrong, this change may cause spaces to show
up at the beginning of some lines in "mixed content". If you see
those, then this change has introduced a regression, and I will
need to go back and figure out the right way to fix it (instead of
the kludgy way I was doing it before.
Also, instances like the following were showing up in some cases.
.PP
.sp
That is, a .PP macro immediately followed by a .sp macro. I can
think of no good read to ever generate that, and I am not sure why
it is getting generated. So I have taked the lazy way out and
caused the stylesheet to now strip out the .sp from all such
instances, leaving just the .PP
Michael Smith [Sun, 24 Jul 2005 14:10:31 +0000 (14:10 +0000)]
Fixed handling of lists with titles. Also reworked handling of
space around lists. (Closes #1243003 and 1241371).
Among other things, this change "normalizes" line space before and
after verbatims. For most cases, the stylesheet now attempts to
ensure each verbatim is preceded and followed by exactly one line
of space.
Michael Smith [Wed, 13 Jul 2005 17:03:08 +0000 (17:03 +0000)]
Removed unnecessary .LP requests that were being generated after
lists. I don't know why those were there, but they have apparently
been in since the code was first added to the repository.
Michael Smith [Wed, 13 Jul 2005 03:57:50 +0000 (03:57 +0000)]
Prevent hyphenation of computer inlines, filenames, and URLs.
(Closes #1124391; thanks to Ryan Hofschneider for reporting).
By default, even when hyphenation is enabled (globally),
hyphenation is now suppressed for "computer inlines" (currently,
just classname, constant, envar, errorcode, option, replaceable,
userinput, type, varname -- but probably need to add more) and for
filenames, and for URLs from Ulink.
It can be (re)enabled using the man.hyphenate.computer.inlines,
man.hyphenate.filenames, and man.hyphenate.urls params.
Michael Smith [Mon, 11 Jul 2005 10:29:58 +0000 (10:29 +0000)]
Implemented out-of-line handling of URLs for links (Ulinks).
(Closes #1234711 and #1233909)
For display of links, this implementation gives users three choices:
1. Number & list links. Each link is numbered and a numbered
list of all links is added to the end of the document.
or
2. Only list links. Links are not numbered, but an (unnumbered)
list of links is added to the end of the document.
or
3. Suppress links. Don't number links and don't add any list of
links to the end of the document
Users can also choose whether links should be underlined.
Default is "the works" -- list, number, and underline links. The
man.links.* parameters can be used to change the defaults.
The default heading for the link list is REFERENCES. That can be
changed using the man.links.list.heading param.
There is a performance hit for listing and numbering links. But it
is not an unreasonable one.
The code currently only checks for Ulinks. It can be updated later
(possibly) to deal with DocBook NG linking. But come to think of
it, DocBook NG links get converted by the strip-ns template. So
DocBook NG users will just need to run that first. Or, if I can
ever manage to get the single-pass conversion process to work with
manpages, it will get done automatically.
Michael Smith [Sun, 10 Jul 2005 16:39:04 +0000 (16:39 +0000)]
Ulink handling -- fixed performance issue (per-Refentry sets of
Ulinks are now assembled once per document and stored in a
global variable).
It now takes roughly on the order of 15 to 20 percent longer to
process a doc with links numbered than it does to process it with
links unnumbered. Which may sound like a lot, but with a doc that
has 100+ links, on my machine, that still amounts to less than a
one-second difference (only about 0.7 seconds).
Bob Stayton [Sun, 10 Jul 2005 08:14:08 +0000 (08:14 +0000)]
Added support for specifying number format for auto labels for
chapter, appendix, part, preface from patch #1234737
to satsify feature request #1178038.
Michael Smith [Sat, 9 Jul 2005 10:07:58 +0000 (10:07 +0000)]
Checkpointing further ulink changes. All visible links are now
displayed in the link list, and numbered. But it performance
leaves something to be desired. And not working properly for case
where document contains multiple Refentry.
Michael Smith [Sat, 9 Jul 2005 09:00:14 +0000 (09:00 +0000)]
Checkpointing. Now link list contains unique URLs (removes dupes),
but needs more work -- some URLs are getting suppressed that
shouldn't be; one symptom of that is that numbering is thrown off.