10:55 a.m.
Hello,
I'm experiencing some encoding issues with po4a:
I'm using a latin1 man page (this is just in order to reproduce the issue,
in fact the issue was encountered with LaTeX).
I need to specify the encoding of the master document (-M ISO-8859-1).
The generated PO file (with po4a-normalize, or po4a-gettextize) indicate
an utf-8 charset.
File indicates that the po file is an UTF-8 Unicode PO
But the encoding of strings is not good:
instead of the Unicode 00E9 (é, encoded as C3A9 in UTF-8),
I've got an Unicode 00C3 followed by and Unicode 00A9 (respectively
encoded as C383 and C2A9)
Do you have an idea where it could come from?
--
Nekral
11:23 a.m.
Sorry, I was wrong.
The latin1 man page was in fact UTF-8 encoded.
I still need to investigate my issue with LaTex, but you can ignore the
previous mail.
--
Nekral
11:55 a.m.
OK, this issue is solved.
I extracted the comments of the TeX file and added them to the PO (as PO
comments). These comments were no re-encoded.
I now encode them in the TeX module. I think it is better not to encode
the comments in Transtractor because the comment encoding is not known
there (it is not necessary the input charset, as in my case).
Sorry for the noise,
--
Nekral
2:32 p.m.
On Sat, Feb 26, 2005 at 06:55:39PM +0100, Nicolas François wrote:
OK, this issue is solved.
I extracted the comments of the TeX file and added them to the PO (as PO
comments). These comments were no re-encoded.
I now encode them in the TeX module. I think it is better not to encode
the comments in Transtractor because the comment encoding is not known
there (it is not necessary the input charset, as in my case).
I disagree, it does not make sense to have comments in a different encoding
than the rest of input files.
Denis
3:07 p.m.
Sorry Denis, I can't answer your mail directly (I'm currently not
receiving mail, I'm just looking on the Alioth mailing list).
But here is the answer anyway.
I think I wasn't clear.
A module may want to add comments with :
options{'comment'} = dgettext("NOTE for translator: ...");
In this case, the Transtractor don't know the encoding of the comments.
This encoding can be different from the input file document.
So, I thought comments should always be encoded in the po_out encoding by
every modules, not by the Transtractor.
Another way could be to specify the the encoding with the comment string
(or force comments to have the same encoding as the input document).
I don't know what is the preferred way of fixing this. (There is not a
long history of using comments: only the TeX module does use this
functionality).
Kind Regards,
--
Nekral
4:59 p.m.
Hi,
On Sat, 26 Feb 2005, Nicolas François wrote:
A module may want to add comments with :
options{'comment'} = dgettext("NOTE for translator: ...");
This would be a comment in the po file? I think it shouldn't be
translated. The generated po file should be identical idependent on the
configured locale in the machine.
Another way could be to specify the the encoding with the comment
string
(or force comments to have the same encoding as the input document).
If you use comments like "English sentence: ".$part_of_input_doc, it will
be in the input document charset (I think it's the right way). When
pushing a comment, it should be automatically converted from the input
charset to the po file's one (I don't remember if it's already done).
Regards,
Jordi Vilalta
5:13 p.m.
On Sat, Feb 26, 2005 at 11:59:01PM +0100, Jordi Vilalta wrote:
On Sat, 26 Feb 2005, Nicolas François wrote:
>A module may want to add comments with :
>options{'comment'} = dgettext("NOTE for translator: ...");
This would be a comment in the po file? I think it shouldn't be
translated. The generated po file should be identical idependent on the
configured locale in the machine.
OK. You convinced me.
I will move the conversion from TeX to Transtractor.
Kind Regards,
--
Nekral
7006
days inactive
7006
days old
6 comments
3 participants
participants (3)
-
Denis Barbier -
Jordi Vilalta -
Nicolas François