Re: LSR updates and translations

[Federico Bruni]

2012/6/20 Phil Holmes <address@hidden>:
> ----- Original Message ----- From: "Francisco Vila" <address@hidden>
> To: "Phil Holmes" <address@hidden>
>
>
>> Good, I like it in principle, but how will be changes being tracked
>> using SHA IDs?
>>
>> Currently, translations have a SHA ID which is that of the snippet it
>> is a translation of. Any change to the original snippet is done in a
>> commit whose SHA ID is used to generate a diff against the ID of the
>> translation. This determines what the translators have to work on.
>
>
>
> That's a bit I'd deliberately glossed over since I didn't/don't fully
> understand it as yet.
>
> My idea would be that, using the new system, changed snippets would look
> exactly like changed documentation files.  If I modify, say,
> /Documentation/notation/ancient.itely, how do the translators pick up that
> this has been changed?
>

We copy in the translated file a committish which identifies our
latest revision.
We may use the committish of the latest commit (of the repository), as
recommended in the CG.
Or the committish of the latest commit of the original file, as me
(and Francisco) do in order to easily see in the log of the file which
version we translated/updated.

As long as Documentation/snippets is tracked by Git, translators
should be able to keep track of the changes in the snippets.
Check check-translation.py to see if your changes will break something.

When you have a patch, I can test a translation workflow and see if it
works as expected.

Anyway, your plan seems great.
Moving away the translations from the Git-tracked snippets is the
easiest solution, as we discussed months ago.

Thanks!
Federico