On Mon, Oct 31, 2016 at 08:03:32PM +0100, Marcus wrote: > Am 10/31/2016 02:59 AM, schrieb Ariel Constenla-Haile: > >On Sun, Oct 30, 2016 at 09:48:18PM +0100, Andrea Pescetti wrote: > >>On 29/10/2016 bugzilla wrote: > >>>https://bz.apache.org/ooo/show_bug.cgi?id=127201 > >>>--- Description --- > >>>The current translations in the source code repository are from more than > >>>two > >>>years ago. The latest translations from Pootle should be integrated back > >>>in the > >>>source code. > >> > >>I would advise against this. I'll explain the reason: Pootle does not have > >>the 4.1.x strings, so translators in Pootle are translating strings that > >>came from trunk (4.2.0, so to say). > > > >Things are rather more complicated than this, due to more than two years > >of inactivity merging translations; take this example: > > > >https://translate.apache.org/es/aoo40/translate/cui/source/tabpages.po#unit=13107269 > > > >In the current SDF files, both on trunk and AOO41* branch, this string > >isn't inside RID_OFAPAGE_AUTOCORR_EXCEPT, all these SDFs have the string > >inside RID_OFAPAGE_AUTOCORR_QUOTE, this is a change (may be a bug, we > >should check this) introduced with the integration of ia2: > > > >https://svn.apache.org/viewvc/openoffice/branches/AOO414/main/cui/source/tabpages/autocdlg.src?r1=1465678&r2=1541847&diff_format=h > > > >several string were move from one tabtage to the other. > > > >In trunk and AOO414 source, STR_PB_DBL_START and others are part of > >RID_OFAPAGE_AUTOCORR_EXCEPT, the same on pootle. So pootle does have > >4.1.* strings. But 4.1.* are not on the SDFs from trunk and AOO414. > > > > > >>This means we don't have a clean way for > >>reintegrating strings into the AOO41x branches. > >> > >>Of course, integrating translations should be a priority for 4.2.0; and of > >>course, it is still possible to integrate translation fixes into 4.1.4 (I > >>did this for fixing a few Italian strings in 4.1.2, but I edited the SDF > >>files directly, after doing the same fixes in Pootle for trunk). > >> > >>So if this is about integrating some important fixes, I'm all for it. If it > >>is about completely reimporting translations, I would prefer to do it in > >>4.2.0 since it is more correct in general: Pootle has 4.2.0, the time for > >>updating translations is 4.2.0 and translators expect a translation schedule > >>for 4.2.0. > > > >Pootle does not have trunk, see the example above. If we get the > >original en-US SDF used to generate the pootle templates, the date in it > >may give us an idea of what svn revision is reflecting pootle. > > > >So far, what we have in the SDFs on trunk and the branches is a merge > >that has as base an en-US SDF generated on 20130618 17:22:18, that is > >the date on al the SDFs on trunk, branch AOO41* and > >http://home.apache.org/~jsc/translation/aoo40new_en-US.sdf > > > >I assume the current pootle strings are closer to branch AOO412-4 than > >trunk, this is my conclusion after doing some test with jsc SDF on > >home.apache.org and one generated from AOO414 source. So I will > >integrate translations on demand, starting with the Spanish translation > >as a proof of concept; integration won't be automatic, a lot of manual > >work is needed, and the pootle server won't be modified in any way, all > >work is done offline by downloading the po files. > > sorry for asking as it's not clean for me. Do you plan to update the > translations already for 4.1.4 or not before 4.2.0? If 4.1.4, then the final > decision depends on the tests you will do?
I have already been studying and testing for two weeks, I will integrate translations on AOO414 branch but only on demand, because (a) it is a lot of work to merge translations from pootle to svn and (b) we need people to start doing QA as soon as possible not on the RC1. For (a) I'll try to do my best, for (b) we have working buildbots (Windows buildbot is waiting for some Infra fixes). Regards -- Ariel Constenla-Haile
signature.asc
Description: Digital signature
