Bug 31525
Summary: | openjade-1.3 has known (SourceForge) bug with a fix | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | [Retired] Red Hat Raw Hide | Reporter: | matti aarnio <matti.aarnio> | ||||||||
Component: | openjade | Assignee: | Tim Waugh <twaugh> | ||||||||
Status: | CLOSED RAWHIDE | QA Contact: | Brock Organ <borgan> | ||||||||
Severity: | high | Docs Contact: | |||||||||
Priority: | medium | ||||||||||
Version: | 1.0 | ||||||||||
Target Milestone: | --- | ||||||||||
Target Release: | --- | ||||||||||
Hardware: | i386 | ||||||||||
OS: | Linux | ||||||||||
URL: | http://sourceforge.net/tracker/index.php?func=detail&aid=216660&group_id=2115&atid=102115 | ||||||||||
Whiteboard: | |||||||||||
Fixed In Version: | Doc Type: | Bug Fix | |||||||||
Doc Text: | Story Points: | --- | |||||||||
Clone Of: | Environment: | ||||||||||
Last Closed: | 2001-06-14 13:49:23 UTC | Type: | --- | ||||||||
Regression: | --- | Mount Type: | --- | ||||||||
Documentation: | --- | CRM: | |||||||||
Verified Versions: | Category: | --- | |||||||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||
Cloudforms Team: | --- | Target Upstream Version: | |||||||||
Embargoed: | |||||||||||
Attachments: |
|
Description
matti aarnio
2001-03-12 17:37:00 UTC
Could you please verify that the patch actually does work? Thanks. we shall see, I constructed patch for it, and am now running 'rpm -bb' .. My initial goal was to have OpenJade CVS version compiling at current rawhide, but alas, that doesn't seem to be possible (autoconf/automake feature problems.) Created attachment 12482 [details]
patch -p3
ok, it works. $MAILVAR/ appears successfully at TOC of http://zmailer.org/zman/zmanual.pdf Your patch is different to the one on sourceforge. The sourceforge version makes more sense to me. Do you know why yours works? Did you try the sourceforge (add a line) version before the replace-a-line version you've attached? Created attachment 12567 [details]
Here's the testcase, for posterity.
I went thru both the C++ code, and the resulting TeX output. The problem lurks at place where code tries to store a character in to string beyond its end. That is why adding the extra char (dummy) is needed. Changeing the "protection string" into one with non-special starter ('?') caused header title to be converted from: $XXX to ?XXX which naturally the TeX ate just fine. The wrong result is not apparent at my output, as only top-level chapter titles go to page headers, but it shows up at PDF bookmarks. Oh yes, sorry for the difference - my bug at reading that HTMLized patch.. Reading more of the relevant C++ code, I think the bug is considerably bigger.. Input: <!DOCTYPE article PUBLIC "-//OASIS//DTD DocBook V4.1//EN" []> <article> <section> <title> \ ^ _ ~ { } & # % $ VARNAME</title> <para>Words.</para> </section> </article> Produces (jade)TeX: .... {title-sosofo-mode}}\char92{} \char94{} \char95{} \char126{} \{ \} \Entity{amp} \# \% \$ VARNAME\endNode{}\def\HeadingText{% \ ^ _ ~ \{ \} \# \% \$ VARNAME}% \endHeadPar{}\endSeq{} .... Which contains parts of protected chars, and parts of NON-PROTECTED chars in the HeadingText block. The fix in question did help for chars { } # % % $ but not for \ ^ _ ~ (and & - propably not any of entities, e.g. < included..) More fixes are needed. Created attachment 12626 [details]
New version of TeXFOT-fix -- proper one, but incomplete (see comments)
Proper Solution will need some way to insert arbitrary characters (at least of codepoints <= 255) into TeX without it being a macro, but try as I might, I don't seem to be able to find a way to do that. Alternate would be to fix the HYPERREF package in tetex to allow: Package hyperref Warning: Token not allowed in a PDFDocEncoded string, (hyperref) removing `\char' on input line 142. No doubt it is something trivial and mentioned somewhere, but my TeXnical skills are inadequate here... Mine too. openjade-1.3-13 has the one-line patch applied (the sourceforge version), so it's partially fixed there. I don't think that Adam is planning any maintenance release for openjade 1.3. It might be worth checking if this is fixed in the current CVS. It isn't, and that is second bug I spotted with my quick overview... - The new inMath_ counter is mishandled at endMathSeq (propably a feature that is not yet finalized..) - protectedChar_ is initialized too short. Otherwise the CVS version has fancy table changes, and ligature breaks for -, <, and > chars. Care to whip up a quick patch? I have CVS commit writes on openjade now. :-)) Closing; the original bug is now fixed. Please open a separate bug report with more details about the others that you found, if not already fixed. And patches. ;-) |