From 1 to 227 errors in just one fix (for MT)
Ricardo Palomares Martínez
Bug 436839 (https://bugzilla.mozilla.org/show_bug.cgi?id=436839)
started as an error on es-ES translation, and it has turned to affect
a nice bunch of locales. Since we use MozillaTranslator, I felt guilty
for not having a check in it that compare ending char in original vs.
translation strings, and I'm implementing it.
Since it already works in the synthetic test product I've used, I'm
testing it for the parts of es-ES I'm in charge of. The good news are
that the check seems to work very well, so I'll release a new MT
version very soon (it depends on another fixed bug).
The bad news are that I've found 43 errors due to this new check in
es-ES shared components, 91 in Thunderbird and 93 in SeaMonkey (not
tried yet in Calendar).
I can and will fix TB and SM errors, but is there any chance to get
the 43 errors in shared components fixed for some 3.0.x Firefox version?
BTW, this could happen to other teams using MozillaTranslator. Sorry
for not having solved this before.
Ricardo.
Ricardo Palomares Martínez
> Bug 436839 (https://bugzilla.mozilla.org/show_bug.cgi?id=436839)
> started as an error on es-ES translation, and it has turned to affect
> a nice bunch of locales. Since we use MozillaTranslator, I felt guilty
> for not having a check in it that compare ending char in original vs.
> translation strings, and I'm implementing it.
>
> Since it already works in the synthetic test product I've used, I'm
> testing it for the parts of es-ES I'm in charge of. The good news are
> that the check seems to work very well, so I'll release a new MT
> version very soon (it depends on another fixed bug).
I still have this pending, sorry.
> The bad news are that I've found 43 errors due to this new check in
> es-ES shared components, 91 in Thunderbird and 93 in SeaMonkey (not
> tried yet in Calendar).
>
> I can and will fix TB and SM errors, but is there any chance to get
> the 43 errors in shared components fixed for some 3.0.x Firefox version?
Any opinion on this, Axel? Should I file a bug for those 43 errors in
shared components? Please, don't tell me you need 43 bugs. :-)
Ricardo
Axel Hecht
Bugs should be filed per issue, and this sounds like you have one issue
in 43 places, so one bug sounds like the way to go.
Axel
Ricardo Palomares Martínez
> Ricardo Palomares Martínez wrote:
>> Ricardo Palomares Martínez escribió:
>>> es-ES shared components, 91 in Thunderbird and 93 in SeaMonkey (not
>>> tried yet in Calendar).
>>>
>>> I can and will fix TB and SM errors, but is there any chance to get
>>> the 43 errors in shared components fixed for some 3.0.x Firefox version?
>>
>>
>> Any opinion on this, Axel? Should I file a bug for those 43 errors in
>> shared components? Please, don't tell me you need 43 bugs. :-)
>
> Bugs should be filed per issue, and this sounds like you have one issue
> in 43 places, so one bug sounds like the way to go.
Hi, while reviewing and fixing I've found three kind of errors:
- Differences in newline escape chars (4 files affected)
- Typos and some other errors (5 files affected)
- Differences in ending chars between original and translation (can't
confirm now the exact number of files, but I'd say between 20 and 40)
So I think I should file three bugs. The problem I have is that I
can't really produce "standard" diffs for each of the three bugs while
isolating the changes for each one. In fact, I can't confirm the files
affected by the third point because I can't export files and,
therefore, run a diff nor attaching it to the bug without modifying
current files and making impossible to commit on a per-bug basis.
Since the time window to commit patches between dot releases is so
small, do you think I could file the three bugs and attach patches for
reviewing for the first two, get the approval, and try to produce and
attach the patch for the third one? If it turns to be too tight, this
third bug would remain open until the next dot release. Does this fit
you, Axel?
TIA
Axel Hecht
Sounds ok. Mind using http://l10n.mozilla.org/dashboard/bugzilla.html to
CC the relevant locale owners on the affected localizations?
Thanks
Axel, who had to redo that page for bugzilla reorg.
Ricardo Palomares Martínez
>>>> Ricardo Palomares Martínez escribió:
>>>>> The bad news are that I've found 43 errors due to this new check in
>>>>> es-ES shared components, 91 in Thunderbird and 93 in SeaMonkey (not
>>>>> tried yet in Calendar).
>
> CC the relevant locale owners on the affected localizations?
Sorry, I think there was a misunderstanding. The errors are *just* for
es-ES, not all localizations.
Actually, there are a number of en-US strings that are likely to have
non-needed extra ending whitespaces, so the fix would be on en-US side
instead of the rest of locales, but it is not easy to decide on them
without going case-by-case, I guess.
For es-ES, some changes are to fit those extra whitespaces, and others
are due to really different ending chars (whitespace instead dot, dot
instead comma, etc.)
Ricardo
Axel Hecht
Sounds OK, get those bugs filed.
Axel
Ricardo Palomares Martínez
> Ricardo Palomares Martínez wrote:
>> Sorry, I think there was a misunderstanding. The errors are *just* for
>> es-ES, not all localizations.
>>
OK, so here they are:
- bug 449242 for newline escape chars fixes
- bug 449244 for typos and other minor errors
- bug 449245 for differences in ending chars
I haven't provided a patch yet for 449245, since some files in patches
for 449242 and 449244 may be also affected by 449245. As this is not a
big problem for L10N, if 449245 doesn't make it for Fx 3.0.2, it won't
be a big problem.
TIA