Re: Problems building snap-20070419
Hello, El Thu, 19 Apr 2007 03:01:46 -0500 Jason L Tibbitts III [EMAIL PROTECTED] escribió: rm -f fvwm.ar.gmo /usr/bin/msgfmt -c --statistics -o fvwm.ar.gmo fvwm.ar.po fvwm.ar.po:57:10: invalid multibyte sequence fvwm.ar.po:57:12: invalid multibyte sequence /usr/bin/msgfmt: found 2 fatal errors make[1]: *** [fvwm.ar.gmo] Error 1 make[1]: Leaving directory `/home/tibbs/fvwm/po' make: *** [distdir] Error 1 Mmm, strange. Someone changed the string for the arabic translation for More I will not discuss that, cause I don't know any word of Arabic, but looking at it, I see this: msgstr أ٣^ãثر, which seems to be weird. I am not an Arab, but, I bet that that is not correct, uh? The original translation I did was this: أكثر Taken from google, and maybe, not too accurate. Maybe seventhguardian or someone can give a better insight on what happened. Anyhow, if that is the problem, I wonder why is it working for me and some other people. I *think* that the format of the file was messed in the middle, it might have been my MTA or any other thing. All the files I provide *should* be encoded at utf8, if not, there's a problem in the middle. Maybe that was what happened to this patch... And now that I say that... Does someone know why some po files are not using utf8, but iso-8859-whatever? I think they all should be converted. I might make a patch for that, this time, gzipped to avoid any chance that my mail servers mess up my encoding. But I doubt that is the cause... -- Jesús Guerrero
Re: Problems building snap-20070419
On 4/19/07, Jesús Guerrero [EMAIL PROTECTED] wrote: Hello, El Thu, 19 Apr 2007 03:01:46 -0500 Jason L Tibbitts III [EMAIL PROTECTED] escribió: rm -f fvwm.ar.gmo /usr/bin/msgfmt -c --statistics -o fvwm.ar.gmo fvwm.ar.po fvwm.ar.po:57:10: invalid multibyte sequence fvwm.ar.po:57:12: invalid multibyte sequence /usr/bin/msgfmt: found 2 fatal errors make[1]: *** [fvwm.ar.gmo] Error 1 make[1]: Leaving directory `/home/tibbs/fvwm/po' make: *** [distdir] Error 1 Mmm, strange. Someone changed the string for the arabic translation for More I will not discuss that, cause I don't know any word of Arabic, but looking at it, I see this: msgstr أ٣^ãثر, which seems to be weird. I am not an Arab, but, I bet that that is not correct, uh? The original translation I did was this: أكثر Taken from google, and maybe, not too accurate. Maybe seventhguardian or someone can give a better insight on what happened. Anyhow, if that is the problem, I wonder why is it working for me and some other people. What I did was a copy-paste from your patch, but I have no idea where things got messed up. I used nano as an editor, but the text appears like this: #. ./fvwm/ConfigFvwmSetup: line 144 msgid Move msgstr Øر٠It doesn't look good either, but I have no idea on how this should be done.. I will remove the arabic translation for now, hoping someone will do it right eventually.. Cheers, Renato
Re: Problems building snap-20070419
El Thu, 19 Apr 2007 19:14:01 +0100 seventh guardian [EMAIL PROTECTED] escribió: It doesn't look good either, but I have no idea on how this should be done.. I will remove the arabic translation for now, hoping someone will do it right eventually.. I think the only right solution, is the total migration to utf8, in which regards the po/ directory. Of course, that implies that to manipulate the files, you will need an unicode capable environment. But it is the only way that you can edit files with different character sets without messing it all up. Converting the files to utf8 is trivial for me, I can easily do that. I can also patch the .po files which have a different charset, to use utf8. Since it is a technical patch, and not a proper locale edition, I don't see a reason to regenerate them, so, original authors and such will still be kept. I can do that and submit the files into a tarball, so we make sure they are not messed in the way. But that would only work if the files are used to substitute the ones in the cvs repository. Any copy paste done in a non utf8 aware environment or with a non aware editor can mess them up again. Anything that's non-utf8 is just not ok when you have to deal with multiple encodings (for me). Let me know what do you think about the idea. As you can see, at the current state, this is just a mess: [snip] $ for i in *.po; do echo -n $i - ; enca -L none $i; done fvwm.ar.po -Universal transformation format 8 bits; UTF-8 Surrounded by/intermixed with non-text data fvwm.de.po -Unrecognized encoding fvwm.fr.po -Unrecognized encoding FvwmScript.ar.po -7bit ASCII characters FvwmScript.de.po -Unrecognized encoding FvwmScript.fr.po -Unrecognized encoding FvwmScript.sv_SE.po -Unrecognized encoding FvwmScript.zh_CN.po -Universal transformation format 8 bits; UTF-8 fvwm.sv_SE.po -Unrecognized encoding FvwmTaskBar.ar.po -Universal transformation format 8 bits; UTF-8 FvwmTaskBar.de.po -7bit ASCII characters FvwmTaskBar.fr.po -7bit ASCII characters FvwmTaskBar.sv_SE.po -Unrecognized encoding FvwmTaskBar.zh_CN.po -Universal transformation format 8 bits; UTF-8 fvwm.zh_CN.po -Universal transformation format 8 bits; UTF-8 [/snip] That is, just like the files are in the repo. A complete mess. All the files should be enconded the same, and the charset line in the po files, changed to reflect utf8. Let me know what do you think, and if there is nothing wrong with this, I will do the job and briefly submit a gzipped tarball to avoid problems with the encoding getting messed in the way. -- Jesús Guerrero