[tex-live] Re: UTF-8 support

Vladimir Volovich vvv@vsu.ru
22 Jan 2003 11:10:25 +0300

"PO" == Petr Olsak writes:

 >> A support for UTF-8 is already provided by ucs.sty available from
 >> CTAN. I think it should be included on the next CD.

 PO> NO!!

 PO> The UTF-8 support cannot be 100% perfect, if you realise it only
 PO> at macro level. I have made my own macros for UTF-8 support
 PO> before my second version of encTeX and I know very deep all
 PO> problems connected with thist task.

 PO> Please, don't assume that all UTF-8 codes on input side and on
 PO> the output to \write files can be implemented by macros.

if we are talking about LaTeX, then properly done writes to files
(AUX, etc) - those using standard LaTeX commands to do writes - will
save the characters properly in LaTeX internal character
representation (LICR) form, even when you are using UTF-8 input
encoding, e.g. if  (in UTF-8 encoded document) will be written to a
file, it will be represented in the encoding-independent form: \'A

the characters in AUX and other files written by TeX should be
encoding-independent, as the document can use several input encodings,
so writing UTF-8 to files is not a good idea.

 PO> People are using his own macros with \catcodes, \uccodes etc. set
 PO> to idividual letters and they do the \futurelet of the next
 PO> character.

these "dirty tricks" are not permitted in LaTeX (even for
macro-package writers) - catcodes of letters (when using inputenc) and
uc/lccodes are fixed.

 PO> If you don't use encTeX, you are not able to do these tasks in
 PO> normal sense.

if one uses proper interfaces (e.g. protected writes in LaTeX) then
there will be no problems.