On 23/04/13 01:56, Alan wrote:
> Hi,
>
> On Mon, 22 Apr 2013 15:54:07 +0200 sajolida@??? wrote:
>
>> On 21/04/13 01:51, Alan wrote:
>>> Hi,
>>>
>>> When finalizing feature/about_tails, I encountered a wired issue
>>> with PO files. Python's gettext crashes with a LookupError while
>>> reading encoding of the PO file header, which is... "CHARSET". This
>>> is obviously a wrong charset, but it's the one created by
>>> refresh-translations for newly created PO files:
>>>
>>> "Content-Type: text/plain; charset=CHARSET\n"
>>>
>>> I currently workaroud that in my python scripts, but I'm afraid that
>>> the actual problem is in refresh-tranlsation. What do you think?
>>
>> I agree. The script refresh-translations runs xgettext to generate new
>> POT files but it doesn't take into account the previous POT files, and
>> for example the charset that could have been defined in it.
>>
>> Can you review and test the branch bugfix/preserve_pot_charset?
>>
> I tested the branch and it works. However, its 1st commit is not useful
> and its name is misleading: as far as I understand, it actually doesn't
> preserve anything, it just sets UTF-8 as a default encoding. I'm in
> favor of merging it anyway, *but* I would strongly prefer that we
> rewrite its history and rename it before. Would you please do that?
Which commit? d1a46fb? 1868290? None is using the verb "preserve" so I'm
not sure which one you are referring to, and what part of the message is
problematic. I've read them both again and they are fine with me, so I'd
rather you change them directly.
The name of the branch mentions `preserve` but it's sometimes hard to
know where the dev will lead you when you choose the name of a branch.
