Hey,
The bug list has some encoding problems, at least in French, which makes it hard to promote to French speakers I'm quite lost between all the pdfReaders mailing lists / trac / web@ , so sorry if email is a lousy way to raise the issue.
We fixed many of those problems for the DFD website last year, but it's really not my era of expertise.
Thanks! Lucile
On 31/10/13 11:39, Lucile Falgueyrac wrote:
The bug list has some encoding problems, at least in French, which makes it hard to promote to French speakers I'm quite lost between all the pdfReaders mailing lists / trac / web@ , so sorry if email is a lousy way to raise the issue.
web@ is responsible for the issue described. We have also been long aware of this, but fixing those problems has proven problematic. As far as I know, this should be tracked on Trac too.
Lucile Falgueyrac lucile.falg@fsfe.org, Thu, 31 Oct 2013 12:39:15 +0100:
The bug list has some encoding problems, at least in French, which makes it hard to promote to French speakers
That goes for the French section in all language versions of the page: http://fsfe.org/campaigns/pdfreaders/buglist
In the repository the bugs reside in fsfe-web/trunk/campaigns/pdfreaders/bugs-fr.en.xml
The file uses UTF-8 encoding, which would be able to hold french accents generically. Most of the names are stored with XML/Unicode circumscriptions however and for some reason those are wrong - the numeric representation is probably not actually Unicode.
This is not just a wrongly encoded file, all characters actually stored in the file are from the ASCII name space, so there cannot even be encoding errors. The errors are on the XML/HTML-Representation-Level and this makes it tricky to convert the file using automatic tools. To make it even more difficult, the encoding is mixed - some of the circumscriptions are correct, some are not.
Fixing this automatically will be tricky, since the numeric representation has to be resolved (the GNU implementation of 'echo' can do this) before automatic conversion tools like 'iconv' can handle the input.
I do not know if I can get to this soon. Maybe someone can fix the bug manually for the time being, though at some point we should have a tool to handle this anyway.