Menu
▾
▴
#525 TMX: match other languages if no project target language
closed-fixed
5
2011-11-21
2009-03-21
No
When a multilingual TMX file is parsed, even languages that do not match the project target language are considered for matching purposes. With a lot of languages for a source string, the result can be that 5 100% matches with 5 different target languages are proposed when the project target language, if it had been considered, coud have produced matches even inferior to 100% but much more useful for the translator.
Your title - Multilingual TMX files and target language - does not specify what you want.
The fact that TUs with a different target language than the project's are used was done on purpose (it can be very useful to have DE-EN translations as reference in a DE-FR project), but not with multilingual TMXs in mind.
Didier
Moving to RFE.
Didier
As already mentioned, it can be sometimes beneficial to have a different target language displayed.
On the other hand, as your rightly mention, it can sometimes be an issue.
It would be best to have an option to trigger that behaviour, and possibly to specify which alternative languages are useful.
Didier
Implemented in SVN (/trunk).
What has finally been done is a compromise.
For a given TU, if there is a TUV for the target language, only that TUV is loaded, the other ones are ignored.
If there is no TUV for the target language, all other TUVs are loaded, regardless of the language.
This solves the issue of loading unwanted non-project languages from multilingual TMXs, while still allowing to load monolingual TMXs in other target languages than the project target languages.
Didier
Closing...
This feature was implemented in the released version 2.5.0 update 3 of OmegaT.