Red Hat Bugzilla – Bug 129719
Upstream indic translations need to be incorporated into OOo packages
Last modified: 2007-11-30 17:10:47 EST
Description of problem: The upstream translations for the 5 indic languages (hi, pa, ta, gu, bn) need to be incorporated into our packages. Version-Release number of selected component (if applicable): 1.1.1-6 How reproducible: Every time Steps to Reproduce: 1. Start openoffice.org in any indic locale. 2. 3. Actual results: openoffice.org UI is in English Expected results: For indic locales, the UI should be in the corresponding language. Additional info: The translations for hi, ta, bn & gu are available, pa translations have not yet started. More information on each language is available on http://<lang_code>.openoffice.org for the respective language. The translations are often on other sites, links available from the URL above.
If they are available for OOo 1.1.x then they should get into FC3. If they are only for 2.0, we are out of luck for the 1.1.x series (can't backport languages without a bunch of work).
Hi Jatin, Since it's impossible for us to tell from http://<lang_code>.openoffice.org weather OOo 1.1.x translations for hi, ta, bn & gu are available or not, would you pls investigate it furthur? Thanks, Sarah
Which indic-type languages do we need? OOo has the "hindi" language tag used for translations at this point, but I'm not sure how good the coverage is. All files that contain translated strings are *.src files. Two good tools to use to examine the sources are: http://ooo.ximian.com (and use the LXR on the right side) http://sw.openoffice.org/source/browse/sw/sw/ (direct link to Writer source code)
Hi all, I am sorry i took so long to respond. The URL below contains the translated glossaries for hindi and tamil. http://www.ncb.ernet.in/bharateeyaoo/index1.html Scroll down till you reach the Glossaries section. There are several other related links on same the page. The bengali and gujarati glossaries have been translated, but haven't yet been tested. They are also not publicly available (no JCA), tho they are available to us. Should we use the glossaries then? Punjabi glossary has not been translated. Thus should the pa team translate OOo 1.1 or 2? Which version would be better considering the schedules for OOo 2 and RHEL?
The gujarati translations are available here: http://gu.openoffice.org/servlets/ProjectDocumentList?folderID=361&expandFolder=361&folderID=374
Status update: Hindi translations are in; however not for other four indic languages.
I can get the Hindi translation, but the other two aren't in .gsi file format... do you have any idea how to convert the others? gsi file format looks like this and includes the location where the string is supposed to actually go: basctl source\basicide\basidesh.src 0 string RID_STR_FILTER_ALLFILES 0 1 <All> 20040719 07:34:35
Hmm, I mean, I can get the Gujarati translation, but teh Hindi translation is one of the ones that's not in .gsi format.
umm... since ooo in RHEL4B2 is already available in hindi, i guess you mean that: 1. hindi is available in .gsi format and 2. you need the gujarati and tamil translations in .gsi format?
No, the gujarati translation from: http://gu.openoffice.org/servlets/ProjectDocumentList?folderID=361&expandFolder=361&folderID=374 IS in .gsi format, which is what I need. The other two (Hindi and Tamil) aren't in a format I can use... --------------------------- I am sorry i took so long to respond. The URL below contains the translated glossaries for hindi and tamil. http://www.ncb.ernet.in/bharateeyaoo/index1.html Scroll down till you reach the Glossaries section. There are several other related links on same the page. --------------------------- On that page, there are glossary files for Hindi and Tamil. If you say that the current OOo in RHEL4 _has_ a good Hindi translation, the thats great. I still can't use the Tamil translation though until its in a .gsi file format... Dan
Jatin: I did get a Tamil .gsi file a while back in email from you, on Sept. 20, titled "Tamil openoffice.org translation issues". I will incorporate that file into the next build (-12). So we still need hi + bn translations.
As of now, only hi, tn and gu are translated for OOo. With hindi already in the package, and .gsi files for tn and gu found, you now have all available translations.
of course, i meant ta (for tamil) and not tn in the comment above.
The languages were not getting merged correctly. I have a fix for this which I will build into the next package.
Tested with openoffice.org-1.1.2-14.6.EL4. bn_IN: not translated gu_IN: majority of the strings in the menu are translated. hi_IN: good pa_IN: not translated ta_IN: good
We still see a UI bug on gu_IN and ta_IN at bug 137954
Created attachment 107607 [details] gu_IN translation with euro translation Dan, please have a look into this problem. Currently if gu_IN translation is not available, euro translations will be used (instead of falling back to english).
Created attachment 107608 [details] ta_IN translations mixed with euro translations
A fix for this problem (replacing untranslated strings with English instead of German) will be available in 1.1.2-17 and higher.
Tested with 1.1.2-17.6.EL4. bn_IN: not translated, but no empty square. gu_IN: majority of strings are translated and no empty square. hi_IN: few strings in Tools and Help isn't translated, but no empty square. pa_IN: not translated, but no empty square. ta_IN: few strings in File isn't translated, but no empty square.
Ok, I think the program bugs has been gone. it's time to change the severity to translation then.
Ok, the problem with the gu_IN translation is that the .gsi file is not the correct format, most lines are missing a required tab somewhere in them (there must be 15 tabs on each line, many have only 14). There's actually a defined format for the .gsi files, and they technically have 16 fields or something, but not all are used, so it would be easy to forget a tab in there somewhere. I'll have to hack up the translation merge program to accept this odd .gsi file format once I figure out where the missing tab is and if its consistently missing in the same place. I'll be able to get this in for 1.1.2-18 which should be our RHEL4 GA final build.
A patch which accepts lines with between 13 and 16 fields inclusive will get added to 1.1.2-18. That should take care of the majority of cases, though I see that the gu_IN translations are majorly broken in this area (the spread of fields is from 12 - 20, which is utter crack. Somebody probably removed/added tabs to make it look pretty in a text editor, but of course that breaks transex3). It should make the translation much more complete at the cost of a few string mixups.
1.1.2-18.6.EL4 is in dist-4E
Confirmed fixed in 1.1.2-18.6.EL4. bn_IN: Untranslated gu_IN: Mostly translated and the european strings are gone hi_IN: Mostly translated pa_IN: Untranslated ta_IN: Mostly translated and the european strings are gone.