Re: dia 0.90RC3 encoding property

[Vitaly Lipatov <LAV VL3143 spb edu>]

On 3 Июнь 2002 23:33, Hans Breuer wrote:
> At 10:13 03.06.02 +0400, Vitaly Lipatov wrote:
>
> There is some code in lib/dia_xml.c (around line 134) which tries
> to be smart about default encoding and valid UTF-8. If
> there are no bytes found where the MSB is set it assumes
> well formed utf8. In you case this is plain wrong.
> Placing the correct encoding like:
>
> <?xml version="1.0" encoding="CP1252"?>
>
> in your dia file should help. Beware: the encoding string works
> for german on win32. I don't know the correct encoding
> string for russian on Linux ...
koi8-r

>
> But if you prepare a file which has the offending bits set
> (without an encoding definition) Dia will complain about
> the missing encoding and will show what it assumes to be the
> default.

Thank you for useful information.
>
> You could also apply the attached patch, which does not
> only seek for the MSB but the '&' char encode character
> too. Finally it fixes the re-writing of the temporay
> file including you default encoding.
You patch works fine.
But I saw  problem with SIGSEGV on opening file with
encoding=koi8-r or encoding=cp1251
See my attachment and comments in my previous letter (answer to Cyrille)

>
> Hope this helps,
> 	Hans

-- 
Lav
Виталий Липатов
ФГУП "ЦНИИ Судовой Электротехники и Технологии", Санкт-Петербург
GNU! Linux! LaTeX! LyX!