Hi.
Why is it that Narro requires me to re-approve translations that have been approved long ago?
Because the context changed. Even if the text is the same, it might have been moved in a different file or the corresponding key or comment has changed.
I understand that some words are translated always as is, but that would require the user to mark them.
For example, every once in a while, I'm required to re-review and re-approve a translation for the word "Bookmarks", that had been translated long ago. I distinctly remember having to approve it at least six times over. That doesn't make much sense.
The word Bookmarks is moved pretty often from one file to another from one version to another.
At this point, I have to re-approve some 6000 strings from Firefox Aurora and Beta for Serbian. This happens every once in a while and is very wasteful, because I have to redo the work I have already completed. With the new Mozilla release cycle of every few weeks, I simply can not afford the time to go over thousands of bogus unapproved messages; and am not quite thrilled to do so, because it's unnecessary work.
Six thousand is huge, almost everything, are you sure that's accurate? if so, something is not right; less than 20% of the whole number of strings are changed from one version to another.
Now, theoretically, I should be able to export the translations and reimport them marking all as approved. The problem with that is that Narro has had a history of importing wrong things (so I don't quite trust it with auto-approval), and there is no way to roll back if you screw things up.
Maybe the text is wrong if it says automatically approve. There is no automation in that. If the translation exists, it's approved, that's it.
Whenever you import, you have to provide mirror like file structures for en-US and your locale. I'm already providing them from en-US on the server, you only have to provide translations for your locale in the structure that you find on Mercurial.
You're in complete control, you can do how many imports you like, your translations are versioned on Mercurial, so you can go back to whatever versions of the translations you want.
There is almost no check when you import things, so you *can* import the wrong thing.
I have a few questions:
- Why do the approved strings suddenly become unapproved?
Because the context changes. Either the text was move in a different file, added as new to a file or the corresponding key or comment changed.
- Is it possible in Narro to just translate the difference between the last release and the new one, without having to re-approve everything?
That's how it works. If it's not working like that for you it's either a bug, either the import options are not right.
- Am I doing something wrong that makes this happen?
Most probably, as I'm translating for years now and only change about 200 texts per release.
Give me an example, project, links, how you do it and I'll see what the problem is.
--
Alexandru Szasz