On Wed, Jul 21, 2004 at 07:12:27AM +0200, Christian Perrier wrote: > Quoting Steve Langasek (vorlon@debian.org): > > It is not safe to assume that s/palo/efi/g will give grammatically > > correct results in all languages. If you do commit a patch including > > such automated translations, please mark the translations as "fuzzy" for > > any languages you don't personally know. > And I have to add that if I see some new package with fuzzy strings > appear *now*, the same day that a lot of partman strings have been > inadvertently fuzzied and added in the middle of a freeze, I do not > guarantee to stay calm and not throw my l10n coordination hat away for > a few weeks. > At the minimum, the huge string mentioned by Joey requires careful > attention by the translation teams. Those who work with a proofreading > process will not have time to complete it seriously in the string > freeze schedule. This string is a perfect example of a string which > may need English changes (for consistency), then careful translation > (handling a lot of jargon), then review. > So, about partman-efi, we have the following choices: > -let it come into the archive but forget about it for this release (we > will mark it to be ignored in l10n stats) > -delay its introduction to the archive after the end of the string > freeze > -do nothing, let the mess continue and have a very angry bubulle leave > to Spain crossing fingers...:-) > /me begins to very quite tired of last minute changes Having gone through the same situation on alpha, I can say that the kind of changes we're talking about can mean the difference between having a TC with 25% of users making it to stage2 of the install, and having one with 80% of users making it to stage2 of the install. It's very frustrating on all sides to get no feedback from the partitioner that anything is wrong, and then be told at the end when the bootloader should install that there's a problem that requires reformatting, and even though I made a point to get all of the aboot-installer changes in before the string freeze, I think this is the kind of bug that it's important to try to address in *some* manner before TC2 is released. Perhaps this means covering it in big print in the errata, or perhaps it means ignoring this package for the translation stats until after TC2. -- Steve Langasek postmodern programmer
Attachment:
signature.asc
Description: Digital signature