i18n bug?

[Nicolas Goutte]

On Wednesday 06 July 2005 16:12, David Faure wrote:
> On Wednesday 06 July 2005 15:11, Albert Astals Cid wrote:
> > No, it is not, that string is valid (ú is ú),
>
> Yes but encoded twice, since I can see accentuated letters just fine at the
> top of the file. But you're right, this isn't the reason.
>
> > but even a simple file like test3.po (attached) makes the test program
> > fail too.
>
> I see.
>
> OK I debugged it and it's a bug in msgfmt.
> msgfmt (GNU gettext-tools) 0.14.1 produces a wrong .mo file
> msgfmt (GNU gettext) 0.10.35 (from gettext-kde on ftp.kde.org) produces a
> working .mo file
>
> More precisely, msgfmt-0.14.x produces a wrong hashtable (1,0,2,0,3,4,5)
> whereas the code, and the working version of msgfmt, work with a hashtable
> saying (1,0,5,3,4,0,2).
>
> BTW a workaround is "msgfmt --no-hash test3.po" ;)
>
> I'll try 0.14.5 now from sources, to see why it produces something
> different.

Have you the result of the test?

Also somebody having time should test if this problem is visible with the 
gettext runtime or not. May be the bug is the old runtime code that KDE uses. 
(The hash function has a sublte change, but I do not know if it is the 
problem, as gettext 0.13.1 has laready that change.)

Have a nice day!