On Tuesday 25 January 2005 6:20, Johannes Wiedersich wrote:
> I opened tellico, created a new bibliography, entered one entry, type
> book containing the title: "Umlautetest: Äb Öb Üb bä bö bü".
Tellico's Unicode<->TeX character translation is handled by the
bibtex-translation.xml file, which maps TeX strings to UTF-8 characters.
One advantage of this is that the user is able to (albeit not conveniently)
modify Tellico's behaviour.
> Exporting this bibliography to bibtex (File --> Export --> Export to
> bibtex) and saved it as test.bib.
>
> $test.bib===========================================================
> @comment{Generated by Tellico 0.13.1}
> @book{test:bib,
> title = {Umlautetest: {\"A}b Öb \"{U}b B{\"a} B{\"o} B\"{u}},
> author = {Wieders{\"o}ch, J.}
> }
> ====================================================================
>
> Here Ä, ä and ö are handled correctly to {\"A} etc.,
>
> Ö (capitals) is NOT changed to Bibtex/Latex conventions and
I fully admit that the character map is not complete. I haven't come across
a good list of unicode characters with equivalent TeX strings yet. I added
those I found to be most common, mostly taken from the BibDesk application,
with permission from the author. I'll add the Ö character now. I'd welcome
any other additions. The XML format in bibtex-translation.xml should be
self-explanatory.
> both Ü and ü are leading to bibtex/latex errors:
> Ü should translate to {\"U}
> ü should translate to {\"u}
That was a bug in the mapping. Since I've found many cases of bibtex files
with the incorrect escaping, I wanted a one-to-many mapping. Unfortunately,
there was bug which reversed the string list. Attached is a patch which
fixes that error. It will be in the next release of Tellico.
> Importing a large bibtex database, editing it in tellico and exporting
> it again could potentially destroy the whole database. Ie, rendering it
> unreadable by other bibtex applications.
Yes and no. Yes, with the above bug with the incorrect mapping, With this
patch, we'll be ok. No, concerning the Ö character, since that's not a
valid bibtex character in the first place, and Tellico only replaces the
strings it recognizes.
> @comment-line at the beginning of the generated output file potentially
> also leads to error messages in other bibtex-applications.
No, the comment line is fully bibtex-compliant. All valid parsers should
have no problem.
Thanks for the bug report. The bibtex capabilities of Tellico don't get much
exercise.
Robby
Index: src/translators/bibtexhandler.cpp
===================================================================
--- src/translators/bibtexhandler.cpp (revision 1049)
+++ src/translators/bibtexhandler.cpp (working copy)
@@ -119,7 +119,7 @@
// to represent a character in LaTex. Iterate backwards so that the first one listed
// becomes the 'primary' key
QString s = keyList.item(i).toElement().attribute(QString::fromLatin1("char"));
- for(int j = strList.count()-1; j > -1; --j) {
+ for(int j = 0; j < strList.count(); ++j) {
s_utf8LatexMap[s].append(strList.item(j).toElement().text());
// kdDebug() << "BibtexHandler::loadTranslationMaps - "
// << s << " = " << strList.item(j).toElement().text() << endl;
