Ralf Hoeller writes:
> hi,
>
> i'm missing something in the docs.
> (i'm probably too blind to see, but ...)
>
> i haven't found something about how to produce the *.raw
> file. there is nothing about how to work with multiple
> indexes using the style files.
The program tex2xindy is what you are looking for. In section
1.6 What is tex2xindy?
of the xindy by topic it says:
tex2xindy is a filter that parses `.idx' or similar files and
converts the \indexentry macros into a form readable by xindy.
...
The program makeindex4 uses tex2xindy to give xindy a makeindex-like
flavour.
Nontheless the documentation can be improved.
> i think there should be a "real-world" example, not only
> these empty *.tex and well prepared *.raw
> files in the Doc/tutorial directory.
If you look closely to the .tex files in the tutorial directory you
see that are used to typeset the index, not to compile it. The
tutorial is not intended to present real-world examples. As I
mention in the topic-guide, one should also consult the examples in
the testsuite. There you can find one real-world example of an index
processed for the scriptum of a lecture here at the insitute processed
with makeindex4.
> also the tutorial or manual or xindy-by-topics they talk
> a lot about the lisp-programming,
No. Show me one part where I have written something about Lisp
programming. If you look closely there are is one phrase in the
tutorial mentioning `...Lisp-like form with lots of braces...' Period.
> of xindy (compared to makeindex), but no hints how to
> start with a "real-world" tex-file.
See above.
> my situation:
> i used index.sty to produce multiple indexes in combination
> with makeindex. it works fine, but i miss 8-bit characters
> like umlauts.
> i know the real TeX people love to type \"{a} all the time
> and would never go beyond "a, but i prefer ä (yes, i'm a very
> lazy person and i use my computer just for work).
> so i was looking for a maketex-version that could handle
> umlauts. xindy seems to do the job.
xindy is not a replacement for makeindex. It has a well-defined
indexing model different from makeindex. Hence, one should not expect
xindy to behave exactly as makeindex. makeindex has some built-in
assumptions how the index looks like and this works for many people
but is not usable in more complicated indexes. For these one needs a
well defined model on which xindy is built. The result is that the
do-what-I-mean semantics of makeindex is not very compatible with
xindy's underlying indexing model. Thats the reason why we have not
invested too much effort in giving xindy a makeindex-like user
interface.
> i installed xindy,
> nearly everything worked fine (the directories should be
> created first before they can be a target:
> cp xindy.run xindy.mem /usr/local/lib/xindy/binaries/`cat PLATFORM`
> cp: when copying multiple files, last argument must be a directory)
This bug has already been reported and will be fixed in the next
release.
> after running latex i have two files concerning index-preparation:
> *.ndx *.idx
> running xindy in the way:
>
> wald1:[diss] >xindy -l xindy.xlg ralf.xdy diss.ndx
> This is `xindy' version 1.2 (i386-linux-elf binary version 1.2).
>
> Opening logfile "xindy.xlg" (done)
> Reading indexstyle...
> Loading module "ralf.xdy"...
> Finished loading module "ralf.xdy".
> Finished reading indexstyle.
> Finalizing indexstyle... (done)
>
> Reading raw-index "diss.ndx"...
> ERROR: EVAL: Die Variable |iNDEXENTRY| hat keinen Wert.
> wald1:[diss] >
>
> is not very satisfying.
> if i use an empty file called diss.raw verything works fine,
Since you ran xindy on the .ndx file which is probably not in the .raw
format the result is absolutely correct.
> .. on the default index-file *.idx, but there is no possibility
> to do both indexes (and the umlauts are in the other one ;-)).
I haven't understood the last sentence. What files do you mean?
I have no personal experience with the index.sty package and I also
don't know what kind of multiple indexes it produces.
Best regards.
Roger
--
======================================================================
Roger Kehr kehr@iti.informatik.th-darmstadt.de
Computer Science Department Technical University of Darmstadt