Jeroen Ruigrok van der Werven wrote:
> [Finally got Internet again.]
>
> -On [20080513 19:55], Christian Boos ([EMAIL PROTECTED]) wrote:
>   
>> I just had the impression that we left the discussion on #IRC with no 
>> real conclusion, so that's why I wanted to raise the topic again on 
>> trac-dev. It's nothing /that/ serious,
>>     
>
> I take translations quite seriously. As does anyone I know who works on this
> kind of subject. There's more involved than 'simple translation'.
>   

I was saying that /whether to keep the messages.pot in the repository or 
not/ was nothing that serious.
The same for storing periodically the results of automatic runs of 
extract/updates. I still think that's useless and annoying (especially 
if done after every 5 or 10 normal commit) as even no actual changes to 
strings will propagate as hundreds of line offset changes (times the 
number of .po files), but hey, if you think that serves a purpose 
(allowing the translators to start working on up-to-date files), well, 
I'll shut up.
I'm not saying that the translation topic is not important, be it the 
contributions from translators or the i18n work.

>> We're only at the beginning, so experimenting should be OK - you had your
>> time for experimenting in the sandbox and I didn't join the fun by then,
>> but now that i18n is in trunk (and the more urgent stuff for 0.11 is behind
>> me) I felt inclined to join the work on i18n, experiment and try to improve
>> things there as well.
>>     
>
> I am not sure how you intended this, but the way you formulated this is
> sounds as if I only had a certain right while this was in the branch. And
> the way you proceeded did not give me the feeling of joining, but rather
> putting down your view as the only way forward.
>   

No, I only spent a day to /experiment/ with "my" way for updating the 
catalogs, and this mainly as a side-effect of doing others, more 
important, contributions to the i18n framework (and /those/ should have 
deserved a discussion). This doesn't mean that this automatically 
becomes "the" official way for doing things ever after. At the end of 
the day, I could even have realized that I was wrong and had missed 
something which prevented me to understand your arguments. And indeed, I 
understood that :
 - you need to have the exact same version of Babel in order to get the 
same kind of extraction
 - installing Babel trunk from scratch is not straightforward (but well 
explained in http://babel.edgewall.org/wiki/SubversionCheckout)
 - getting the extract/update/compile right is not exactly trivial 
either, notably the fact that it helps a lot to do an extra update after 
the manual edits, in order to validate them
And therefore I collected all these informations as I was progressing. 
My take was that with those detailed explanations, translators could do 
it as well, but I'm quite willing to admit I'm wrong on that point and 
the purpose of my mail was to discuss *this*, get feedback from 
translators who would say either "sure, we don't mind, it's better that 
way" or "are you kidding? are you seriously asking us to do all this?", 
or anything in between. As feedback has actually been "", feel free to 
move those detailed explanations about the extract/update commands steps 
in a === For Developers === section, and keep the simpler workflow for 
translators (or I'll do it if you want).

But after doing this experimentation, I verified at least that a two 
step commit as I suggested was actually more useful for tracking the 
changes (a la r7053:7054).
And there's nothing revolutionary here, see for example the translation 
guidelines for the Subversion project (section "Updating existing po 
files"):
...

It is recommended that the .po update is done by using two commits; one
after the "make locale-gnu-po-update", and one after the translation is
done. This has two advantages:
...
(http://svn.collab.net/repos/svn/trunk/TRANSLATING)


Btw, those guidelines will be worth looking at again when we'll have to deal 
with merge of translations.



>> Not a big deal really: 0.12/TracInstall presents the minimal things an 
>> user has to do in order to get the translations and TracL10N is aimed at 
>> translators and developers with much more detailed instructions.
>>     
>
> And TracL10N reflects your opinion on the workflow despite my objections.
> I totally do not agree with your stance that people should have to run Trac
> to contribute translations. And you have shown that you are not aware how
> translations in open source projects get handled, which is quite ok, but
> then I am baffled why you put forth your own idea.
>   

Well, you're right, I'm not that aware how things are done elsewhere. 
However I did look at the french translation, saw the same problems as 
mentioned by Richard Liao at the beginning of this thread, and fixed them.
Besides, I think that translating the Trac strings without at the same 
time seeing how they are used in context is not optimal because you 
won't see how the individual translations play together, in a given 
page. But again, that's only my (not so informed) opinion.

> ... I am uncomfortable with especially when objections
> seemingly are being overturned. And when that happens I'll just divert my
> energy and attention elsewhere since my involvement will not matter anyway.
>   

It does matter. Sorry for all this mess (well, we didn't have a decent 
flame on Trac-dev since ages, so this is now fixed :-) ).

-- Christian



--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Trac 
Development" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/trac-dev?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to