On Sunday 16 of September 2007 20:44:34 Eddy Petrișor wrote: > Jaka Kranjc wrote: > > On Monday 10 of September 2007 22:21:59 Eddy Petrișor wrote: > >> Is there a reason for those fixes not to end up in both branches? > >> Anyway, when the trunk-to-beta copying was done, if your translations > >> were there, they couldn't have gotten lost. > >> > >> I understand your position, I am a translator myself, too, but currently > >> the two branches should contain (almost?) the same strings, so there > >> shouldn't have been any regressions. > >> > >> OTOH I made sure that: > >> - if after the trunk->beta copying there was a regression in the stats I > >> reverted the copy and merged the translation the other way around > >> (prior-copy-beta -> trunk) > >> - I got the higher of the stats (if one was lower) and updated it in the > >> other branch; if they were at the same level, > >> - I did no copying, although the trunk to beta copy might have changed a > >> few things within the translation, but not the stats themselves > >> > >> So now both trunk and beta should have the highest stats possible, and > >> with changes that were made in trunk, and did not affect stats, copied > >> in beta. > > > > Wormux is a small package and it so is less likely to have problems. > > But this change wasn't bulletproof. For example trunk could have had a > > worse statistic - lets say I did a few changes and marked a few > > translations as fuzzy, so they can be found more easily and improved. But > > I wouldn't want to do that in the beta branch, because a non-optimal > > translation is better than none. > > So in this case your algo would sync from beta to trunk and discard at > > best just the fuzzy marks. You say you did no copying - I wonder what > > msgmerge > No I did do copies. But if you had had made changes in trunk, not all is > >> - I did no copying, although the trunk to beta copy might have changed a ?
> lost since the important thing at the moment was the beta, while for trunk > you have been announced about the possible regression and you have all the > time in the world (until the next beta or the final, should that be the > case) to fix/revert the change I made. Yes, the beta - in this case it would be damaged from the copy. But it is not just about that. If I hadn't been around and attentive, this could have slipped by, like I'm sure it did with many translators. Again, wormux is small, so any damage would be small, but hacks should be avoided nonetheless. And I still don't know how to check a diff of a change to another branch. Is svn really that crappy that you have to pass URLS for such things? Mnja well, it is so crappy that it has to talk to the server if you want to see the log, so ... Not to question if it can do partial reverts. > > does when it finds two msgstr for one msgid ... > > > > Just let the translators handle it. ;) > > I am one of them and I have quite a good understanding of the process > surrounding l10n/i18n. > > As I said, desperate times call for desperate measures :-) I didn't mean the managers, that's the whole point. Like I said in my first mail, just state that it is obligatory for submitters of updates to relay in which branches it should go if there is more than one. There should be no copying, only msgmerge in the correct order. -- We cannot command nature except by obeying her. --Sir Francis Bacon Have a sourcerous day! www.sourcemage.org
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Wormux-update mailing list [email protected] https://mail.gna.org/listinfo/wormux-update
