[[TF-dev]] Re: [[TF-dev]] Re: ru translation for installer/nbi and "\n" in messages

Hi Dmitri,

just replied to you in another email.

Explanation:
- When converting OmegaT TMX files into our CTI (in fact a tool called 
"WorldServer") TM, there is about 10% loss exactly due to the reasons 
you describe below (unwanted "\n" splitting of segments)
 > let me look into this again with our internal support team and see if 
a fix can be found (script or other conversion tool)
- To the commits made today: we take all the translations from CTI and 
for a string which is not 100% translated in CTI, we replace it with 
English original sentence. So committed are only sentences which are 
100% translated. Because RU projects are still not 100% translated (only 
Platform), there might be commits that contain also English sentences.
- Once it is all 100% translated, nothing will be reverted.

Let me work on the TMX conversion issue and find out if it can be fixed.

Thanks for patience,
Petra

Dmitry Lipin wrote:
> Hi Petra,
>
> Could you please respond?
>
> Due to that issue one of the latest Masaki`s commit (a few MB in size) 
> [1] reverted a lot of "ru" translated sentences into original 
> (English) ones. At least:
> http://hg.netbeans.org/main/l10n/diff/68717eec51e4/src/ru/other/installer/components/products/mysql/src/org/netbeans/installer/products/mysql/wizard/panels/Bundle_ru.properties 
>
> http://hg.netbeans.org/main/l10n/diff/68717eec51e4/src/ru/other/installer/engine/branding/jdk/src/native/launcher/windows/i18n/launcher_ru.properties 
>
> http://hg.netbeans.org/main/l10n/diff/68717eec51e4/src/ru/other/installer/engine/branding/jdk/src/org/netbeans/installer/wizard/components/Bundle_ru.properties 
>
> http://hg.netbeans.org/main/l10n/diff/68717eec51e4/src/ru/other/installer/engine/branding/jtb/src/native/launcher/windows/i18n/launcher_ru.properties 
>
> http://hg.netbeans.org/main/l10n/diff/68717eec51e4/src/ru/other/installer/engine/branding/mysql/src/native/launcher/windows/i18n/launcher_ru.properties 
>
> http://hg.netbeans.org/main/l10n/diff/68717eec51e4/src/ru/other/installer/engine/branding/mysql/src/org/netbeans/installer/wizard/components/actions/netbeans/Bundle_ru.properties 
>
> http://hg.netbeans.org/main/l10n/diff/68717eec51e4/src/ru/other/installer/engine/src/org/netbeans/installer/product/components/Bundle_ru.properties 
>
> http://hg.netbeans.org/main/l10n/diff/68717eec51e4/src/ru/other/installer/engine/src/org/netbeans/installer/utils/applications/Bundle_ru.properties 
>
> http://hg.netbeans.org/main/l10n/diff/68717eec51e4/src/ru/other/installer/engine/src/org/netbeans/installer/wizard/components/panels/netbeans/Bundle_ru.properties 
>
>
>
> [1] http://hg.netbeans.org/main/l10n/rev/68717eec51e4
>
> Thanks,
> Dmitry
>
> Dmitry Lipin wrote:
>> Hi Petra,
>>
>> It looks like OmegaT and OpenCTI (and old Sun`s CTI) uses different 
>> splitters of the messages been translated. In particular, "\n" is 
>> been handled inappropriately.
>> As the result, there are a lot of messages shown as not translated in 
>> OpenCTI but which are, in fact, translated (included in TMX) and 
>> corresponding translation is already pushed to main/l10 repository. 
>> So, what can we do with that? I am not eager to do manual "monkey" 
>> work of finding the translation using OmegaT and putting it in CTI 
>> for each not-translated message :)
>>
>> BTW, I can`t say for sure - nor check - whether the same thing 
>> happens for other translation of Java SE / platform messages.
>> Can we at, least, compare percentage of translation that gives OmegaT 
>> and CTI with the same TMX that Maxim provided? If the values differs 
>> significantly then we need to find out a solution to that.
>>
>> Thanks,
>> Dmitry
>>
>