Elie Roux wrote:
Elie Roux a écrit :
Sorry for this mail, I just haven't tested enough...
So here is something more consistent:
With my package, one can do the same as in XeTeX: reading files written
in another input encoding, but with the same limitation as XeTeX (tell
me if I'm wrong): reading all files as if they were written in the same
encoding... The problem is that output files (like aux files) are all
written in UTF-8. So for example if we read files in latin1, we will
read the aux file as if it were in latin1, when it is written in UTF-8,
so it will make some errors.
The only solution I see is to register a callback in open_read_file,
that will convert from utf8 to latin1, before input is converted again
to utf8. This solution seems heavy and not really efficient. So I was
wondering if there was another way to read the aux file in UTF-8, or
(even better) a way to write the files in latin1? I think the second
solution would be better, but there seems to be no corresponding
callback... Do you think there would be a way to add such a callback?
i don't know how latex handles that, but why not just overloading the
write-to-aux (or whatever) macros, intercept the to-be-written string
and convert it; it's the only controlled way since changing 'all' writes
will mess up other things (like files that users write)
Hans
-----------------------------------------------------------------
Hans Hagen | PRAGMA ADE
Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com
| www.pragma-pod.nl
-----------------------------------------------------------------
