[tex-live] xindy vs utf8 latex

Joachim Schrod jschrod at acm.org
Wed May 7 02:03:52 CEST 2014

On 05/06/14 11:05, Lars Madsen wrote:
> Hi
>
> I was wondering, is any work being done on making xindy/texindy work with latex and utf8?
>
> This question highlights the problem and in an answer there is a good workaround.
>
> http://tex.stackexchange.com/q/153858/3929
>
> If a xindy fix is too far in the future, would it make sense to provide the ie2utf script from https://github.com/michal-h21/iec2utf as a part of TL, and perhaps even make some perl wrapper for it, making it easier to use for the casual user?
>
> (is piping available on all platforms?)

xindy for TL 2014 will support "-L utf8" texindy option for Latin
scripts. It will not support it for non-Latin scripts. I hope that
most users that used \inputenc{utf8} and who complained about
problems on tex.stackexchange.com will be served by that change.

Support for non-Latin scripts in LICR is hard to do. I recommend
everybody who uses non-Latin scripts to use LuaLaTeX or XeLaTeX.
Then one *must not* use texindy, but xindy. I have enhanced the
texindy man page to tell about that more prominently.

I'm thinking about adding an -I option to texindy that skips all
the heuristics that texindy does and just forwards the call to
xindy. To make that clear: The real difference between the commands
texindy and xindy is that texindy tries to find a markup
normalization for LaTeX Internal Character Representations (LICR).
If one doesn't have LICR in one's raw index, one shouldn't use
texindy. Most advices on tex.stackexchange.com concerning that
problem are simply wrong, do not work, and will produce subtle
problems.

FWIW: IMHO, ie2utf is *not* a good workaround. AFAICS, the author
did not understand how xindy works. If this mailing lists wants
more explanations, I'm willing to post them.

Joachim

--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Joachim Schrod, xindy maintainer
Email: jschrod at acm.org