Book logo xindy

A Flexible Indexing System

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: isolatin{1,2}{m,s}.xdy

Reiner Schlotte writes:
 > Hello,
 > I have two remarks and one question concerning the files in
 > xindy-2.0/modules/tex:
 > - The fact that the files isolatin1m.xdy and isolatin1s.xdy
 >   are identical except for the words "merge" and "sort" is
 >   a clear indication to me that we need something like
 >   "merge-and-sort-rule" to do both in one step. 
 >   Furthermore I think
 >   this should read "treat-identically-in-every-respect" (at least for
 >   the (merge-rule "^^e4" "" :string :again)-kind), because
 >   I cannot think of any situation where xindy should treat them
 >   differently (one exception: if xindy were to change the four bytes
 >   of input "^^e4" to one byte of output: "").

Hm. If merge-rules are applied, the result of the mapping is used as
the starting key for the sort rules. All sort rules are applied to the
result of the merge mapping, not the initial. Hence, the same sort
mappings do not make sense anymore. E.g., if ^^e4 is rewritten to ""
in a merge-rule then a the same sort rule again will usually not match
any further.

Keep in mind that merge rules and sort rules implement the concepts of
identity and order, therefore if two keys are said to be identical via
merge rules, they cannot be sorted differently anymore.

Therefore putting both specifications into one file is probably not
what you want.

 > - It should be noted that the first section of the isolatin??.xdy-
 >   files isn't special to the ISO latin?-encodings, but applies to
 >   _all_ 8bit-encodings.

Yes, that's right. Nontheless I thought that it might be easier from
the users point of view to include just one file, not two different
ones, depending whether he/she uses \usepackage[latin1]{inputenc}, or
not. But maybe I should separate them and offer an additional styule
that glues both together again. What's your opinion?

Another problem ist that I have written Perl-scripts that
automatically generate these files from the definition files from the
inputenc package. But many of the codes listed in there are
transformed by TeX, and the result in an .idx file is different from
what is defined there. I currently do not know a good way how to
generate these files automatically. A snapshot of these scripts is
available at

 > - Why aren't there any rules for ^^80 through ^^9f in the
 >   isolatin??.xdy-files? From the TeXbook:
 >     Text output is produced with this convention only when representing
 >     characters of code $\ge128$ that a \TeX\ installer has chosen not to
 >     output directly.
 >                            ^^^
 >   Is there something special about these characters that I'm not aware of?

Hm. Good point. Actually, I wasn't sure if these characters are ever
output this way. But since Don has written it explicitly, this needs
to be changed.

Thanks for your comments.


Roger Kehr
Computer Science Department         Darmstadt University of Technology