Re: xindy support for Welsh language -- accents lost with -C utf8?

2009/4/7 David Stone <dfstone <at> bcs.org.uk>:
> Joachim Schrod <jschrod <at> acm.org> writes:
>
>> David Stone writes:
>>> The only outstanding query I had relates to accented letters.  I find
>>> that if I have accented letters, then the accent may be lost in the
>>> index.
>>
>> I haven't tried it, but that's most probably the bug in texindy that
>> was uncovered a few weeks ago. If you don't use xindy from TeX-Live,
>
> I use xindy from texlive 2008, the dvd.
>
-C in texindy serves two purposes:
1. find the correct sort rules in the given codepage
2. load the tex/inputenc/...xdy module for conversion from LICR

If your idx file is in UTF-8, you need no tex/inputenc/*.xdy modules,
you just need to treat the input as if it came from Omega (and XeTeX
writes the same file). If you say that the files comes from Omega,
texindy ignores -C, assumes UTF-8 and does not read anything from
tex/inputenc. Try:

texindy -I omega -L your_language somefile.idx

I use this with Czech, it also contains accented characters, and it
works. And I use the same syntax with Hindi which is just under
development, and it works too.

>> substitute your texindy command in bin/ with the attached one to
>> solve it.
>
> It seems very similar to it.  But I do not think it is quite the same.
> I tried the version you just sent, with the simple French test given
> in Denis Bitouzé's e-mail of 2009-02-19 18:11:15 GMT, subject `Accents
> lost, even with LC2 companion and "xindy revisited" conference
> example':
>
> \documentclass{article}
> \usepackage[utf8]{inputenc}
> \usepackage[T1]{fontenc}
> \usepackage{makeidx}
> \usepackage[frenchb]{babel}
> \makeindex
> \begin{document}
> cote\index{cote}
> \newpage
> côte\index{côte}
> \newpage
> coté\index{coté}
> \newpage
> côté\index{côté}
> \printindex
> \end{document}
>
> The new version you just sent me does indeed fix that case, as I have
> just verified for myself.
>
> But if I add -C utf8 to the texindy command line, then the bug (or
> another one with similar effect) is still apparent, for this French
> example, just as in my Welsh testing.
>
> The command line I used was:
>
> /usr/local/texlive/2008/bin/i386-linux/texindy_20090407 -C utf8 -L french test_french_index.idx
>
> (where texindy_20090407 was the name I gave to the file you sent, and
> test_french_index.tex was the file the contents of which are above.)
>
> It failed, giving `cote, 1--4', with the above command line.  But if I
> use:
>
> /usr/local/texlive/2008/bin/i386-linux/texindy_20090407 -L french test_french_index.idx
>
> it gave the correct index, with 4 entries.
>
> Let me know what diagnostics and files would be helpful, if any.
>
> --
> David Stone <dfstone <at> bcs.org.uk>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> High Quality Requirements in a Collaborative Environment.
> Download a free trial of Rational Requirements Composer Now!
> http://p.sf.net/sfu/www-ibm-com
> ________________________________________________________________________
> xindy-discuss mailing list -- general xindy announcements and discussions
> Unsubscribe via https://lists.sourceforge.net/lists/listinfo/xindy-discuss

--

-- 
Zdeněk Wagner
http://hroch486.icpf.cas.cz/wagner/
http://icebearsoft.euweb.cz

------------------------------------------------------------------------------
This SF.net email is sponsored by:
High Quality Requirements in a Collaborative Environment.
Download a free trial of Rational Requirements Composer Now!
http://p.sf.net/sfu/www-ibm-com