I'm adding to that: Tx can't manage an SVN repository, the best way would be to run a sync script from an external machine upon demand or installing a script on the SVN machine itself thus saving the hassle of maintaining an additional server.
For some reason I remembered that it was possible in the past, I can't find this feature any longer. Great insights Atilla, lets keep this momentum going until we feel safe enough to work on our resources solely on Tx. Kind regards, Yaron Shahrabani <Hebrew translator> On Thu, Apr 4, 2013 at 9:43 PM, atilla ontas <[email protected]> wrote: > 2013/4/4 Oliver Burger <[email protected]> > >> Well, it doesn't matter if the translations are in a separate >> repository or not and I'm not sure, what it woudl mean for the devs, >> to separate them. >> The problem with tx is - or at least was - the manual work involved >> upon string changes. >> >> As an example: >> When uploading a file into tx, it parses that file and adds the >> strings into the database. When the string changes in the original pot >> (even it's only a typo), and you reupload the pot file, tx will parse >> it and see a new string and one string missing and it will act >> accordingly: >> It will remove the translations for the old string from its database >> and it will add the new untranslated string, since it can not know, >> that this was only a string change. >> So the task for us (mostly me) was: >> - Tell you not to touch certain translation resources in a certain time >> span >> - Download all files from this resource from tx >> - merge them manually with the changed po files (it couldn't be >> automated, since TX also had problems fuzzying strings correctly) >> - reupload all files to tx and tell you to check >> >> I would rather have a solution without the need to do so.... >> >> Cheers, >> >> Oliver >> > Hi again. I think most of the problems you mentioned are past day problems. > -Tx can be locked for certain resources to not accept new translations. > See: https://www.transifex.com/projects/p/chakra-project/language/tr/ > Resources wtih red dot mark locked. > > - As i mentioned earlier even a bash script can automate pulling from tx > process on server side. İs it so hard to do? I really do not know... > > - Merging should be done in tx. Say, use hourly cron job to pull from and > push to tx. Than tx should care of translations and resources merging? > > - As i mentioned earlier again, i can tell tx can automatically pull > resources and transations from svn from what i read. > > - Tx do not mess up translator information anymore. It has been fixed as i > know. > > - For the fuzzy strings, tx still adds them into database as suggestions > and every tx translator can see and if possible use them in online > translation. > My 2 cents. >
