Re: [Gnash-dev] release ?

[Chris Leonard]

On Sat, Dec 31, 2011 at 10:13 AM, Rob Savoye <address@hidden> wrote:
> On 12/31/11 02:42, Sandro Santilli wrote:
>> On Sat, Dec 31, 2011 at 09:24:34AM +0100, Benjamin Wolsey wrote:
>>
>>> It would help if log_debug messages weren't marked as translatable,
>>> since they are of interest only to developers and it's not helpful if we
>>> get debugging reports in Spanish or Japanese anyway. I believe that was
>>> the policy up to now (at least it's what I did).
>
>  We've never had a policy on this at all, different developers did
> different things. I do believe other non-english speaking developers
> would benefit by some of the debug output.
>
>> Agreed, log_debug shouldn't be translated.
>
>  Actually, I disagree that all log_debug messages shouldn't be
> translated. Some are purely for developers, but many aren't. I also
> found when going through the files that there were many log_debug
> messages that should be log_error as well. To really do this right,
> somebody should go through all the log_* messages and make some things
> log_error, some log_debug translated, and some not. I will go back
> through *some* of the files this weekend and do this, but it'd be better
> if that wasn't a solo task.
>
>  As far as priorities, I'm not super happy about ignoring Cygnal, but
> as nobody but me probably runs Cygnal, I guess those could be put off
> for the future. Gnash is overly verbose at times too, many messages, but
> log_error and log_debug were added during development, but not all
> probably still need to be there.


Given the current (limited) state of L10n and the short timeline, we
need to start somewhere.  I created a draft reduced-gnash.pot.  I have
included all UI strings derived from the code in the conveniently
named GUI directory in the Gnash source repository.

This has  1542 words in 321 strings.  This is about 3-5 "mythical"
localizer days of work at common rates of L10n.  Can we agree to start
here?  The full gnash.po is also available in parallel, so nothing
stops a localizer or team from getting ambitious.  I will handle the
merging of the reduced into the full PO for the build process, as well
as reviewing for common build-breaking L10n errors (printf, terminal
newline mismatch, etc.).

I sincerely think that is a reasonable place to start for the current
sprint to the release.  We can revisit the reduced.po contrstruction
algorithm at any time and all strings are easily transferable into the
full PO, so nothing is really lost by starting here and there is much
to be gained by offering an achievable chunk of work to the
localizers.

cjl

reduced-gnash.pot
Description: Text Data