Official switch of translations to gettext

[yann]

The messages on "A Choice Must Be Made" have one English string missing (the first), one Norwegian string missing (the last), and are therefore all wrong by one.
The rest seem fine, although I haven't checked all 2000 strings yet.

Ah, I see. For some reason I missed some scenarios where id's are located after the messages. Maybe they were updated after my 1st pass. I'll update them.

[yann]

For some reason I missed some scenarios where id's are located after the messages. Maybe they were updated after my 1st pass. I'll update them.

The only impacted file aside from A Choice must be made appear to be Gryphon Mountain. People having already imported their language and not willing to fix by hand may just turn the translations for those files to the empty string (eg. by block remove of the lines for those files, and running "make -C po update-po"), and rerun the importer.

[sanna]

No problem, all those strings end up concatenated, just cut them as you like. But if you're editing the po file by hand, I strongly suggest you have a look at specialized tools (emacs po-mode, kbabel, poedit, etc.)

Well, I'll look into if I can get po-mode to work for emacs. Unfortunately poedit aren't available for macosx, and I'm not going to install either gnome or kde to get kbabel or gtranslator.

[yann]

Well, the crash location was somewhere inside the libintl-2.dll. I will look into it further..

libintl-2.dll ? Is there a native libintl on windows ?

[Dave]

Well, the crash location was somewhere inside the libintl-2.dll. I will look into it further..

libintl-2.dll ? Is there a native libintl on windows ?

libintl-2.dll appears to be the libintl implementation on Windows...

David

[ettin]

Provisional statistics.

[Viliam]

Provisional statistics.

sk.po Not translated 3066 (100%)

You made me cry...

[sanna]

I was wondering, what happens to the defines in the various translation/...cfg:s? They're not there in the po-file, so where are they? Lost?

[ettin]

Provisional statistics.

sk.po Not translated 3066 (100%)
You made me cry...

I was wondering, what happens to the defines in the various translation/...cfg:s? They're not there in the po-file, so where are they? Lost?

Have you read the first post of this thread?

This means that the translations in the old data/translations/*.cfg file will need to be imported into po/*.po (or users of 0.8.3 will likely don't see your translation).
To avoid getting in the way of current translation work, I won't run the migration script myself on all langauges. The script will have to be run on a langauge once the coordinator for this language gives the green light, to avoid problems with not-yet-committed translation updates.

[sanna]

Have you read the first post of this thread?

Yes, I've read it, (as probably could be surmised by the fact that I made the second post in this thread. ) and as far as I can see it does not mention what happens to defines within translation files. Nor can I find any information specifically pertaining to defines on the WIKI.

I've run the testmigration on my translation, and the definitions are still used in the actual translations, but the #defines---#enddef are not included. Therefore I ask whether or not I have to manually revert all uses of defines to literal text, or whether it will still be possible to use defines under gettext.

I do not wish to offend anybody, nor give the hardworking gettext-developers any additional work load, but this is quite important if I'm going to get the swedish translation working. I've been using defines quite extensively!

[yeti]

I'm slowly wading through the unmigrated yet translation needing messages...

This needs different translations in reports.cpp (adjective or substantive) and the others (verb):

Code: Select all

#: data/units/Cockatrice.cfg:26 src/actions.cpp:705 src/actions.cpp:842
#: src/reports.cpp:122
#, fuzzy
msgid "stone"
msgstr ""

[yann]

I'm slowly wading through the unmigrated yet translation needing messages...

This needs different translations in reports.cpp (adjective or substantive) and the others (verb):

Code: Select all

#: data/units/Cockatrice.cfg:26 src/actions.cpp:705 src/actions.cpp:842
#: src/reports.cpp:122
#, fuzzy
msgid "stone"
msgstr ""

Hm. All 3 .cpp files use it similarly to "poisonned", so IMHO at least those 3 strings should be in sync. And I'm not sure the special-attack string should b treated as a verb either, when it is compared to other strings availabel there (magical, charge...), although I understand they may need to be different in some languages.

Since it's not currently possible to add a prefix to special=stone, we could add a single prefix to the 3 other occurences, say "special effect label^stone". Would that be OK ?

[yann]

Provisional statistics.

sk.po Not translated 3066 (100%)

You made me cry...

We're waiting for your po-file

[yeti]

Hm. All 3 .cpp files use it similarly to "poisonned", so IMHO at least those 3 strings should be in sync.

I'm sorry, the special attack (.cfg) is the different one, of course.

And I'm not sure the special-attack string should b treated as a verb either, when it is compared to other strings availabel there (magical, charge...), although I understand they may need to be different in some languages.

Actually, it needs to be a kind verbal noun in Czech -- anyway, there can be generally a difference between an action name (attack special), and its outcome (unit state, usually am adjective).

Since it's not currently possible to add a prefix to special=stone, we could add a single prefix to the 3 other occurences, say "special effect label^stone". Would that be OK ?

That would be perfectly OK, thanks.

[yeti]

I suggest to add --no-fuzzy-matching option to msgmerge.

I've already done some po updates, and all the inexact translations it invented was total rubbish that would confuse players more than an untralsated message. I don't like the idea going through all fuzzy messages and changing them to untranslated after each po update just to make the translation usable.