Please help with unicode problem on tasks page generation
I commited on
code that generates i18n tasks pages using DDTP descriptions - which
means it WOULD do so if I would not have to struggle with a really stupid
UTF-8 problem. If you call
tasks.py 2> tasks.err
you get all the problematic descriptions (140kByte) which seem to be
all non-ASCII character containing texts. Even in untranslated descriptions
does this problem happen - just grep for "(lang='en')" in the output above.
The strange thing is that genshi.Markup tries unicode(string) which fails -
but it should not, because all the descriptions are properly formated
I verified that the code would work otherwise because some short
descriptions in the German translation came through. So if you could
lift me above this hurdle we would come quite close to translated
What else needs to be done:
1. Translation of some fixed texts.
I'm not convinced that I found the most clever way to inject translated
data['stringname'] = _('String to translate')
in the template. There must be a more clever way but for the moment
it works and if you have no better idea to put the string into the
template directly which leaves the problem to tell genshi for which
language to translate, we can finish all strings this way and get rid
of all the fixed strings in the template. In principle it is probably
not worth the effort to think about whether this is elegant or not.
2. Joining the scripts tasks_idx.py and tasks.py to have only a single
script generating index and single tasks pages.
3. Obtain also DDTP translations of meta package descriptions to
render this in the left column of the tasks page and at the index page.
4. Take over all the surrounding stuff for other CDDs as it is in the
current update_tasks script. I think about putting the CDD specific
stuff into a config file to enable easy adding of further CDDs but
I have not yet decided about the format of the config file.
Kind regards and I would be very happy if someone could tackle this unicode