11:03 < Luke-Jr> wumpus: going over translations, I notice a *lot* of mismatching %1 with 1%, %n with %1, and & with %

11:03 < Luke-Jr> wumpus: I wonder if there's a good way to communicate to translators that they *need* to get these correct..

11:24 < Luke-Jr> hmm

11:25 < Luke-Jr> Italian omits %n for the singular form translation of "%n GB of free space available"

11:25 < wumpus> Luke-Jr: yes there's quite a few invalid translation messages; in practice they'll just be ignored

11:25 < Luke-Jr> this appears to be undocumented-but-works in Qt

11:25 < wumpus> it's possible to make announcements on transifex,could mention which languages are affected

11:25 < Luke-Jr> but our script complains about it..

11:25 < wumpus> could make an exception for that specific case

11:30 < GitHub80> [bitcoin] MarcoFalke opened pull request #7251: [0.12] gitian: Set reference date to actual release date (0.12...MarcoFalke-2015-gitianTime-0.12) https://github.com/bitcoin/bitcoin/pull/7251

11:32 < Luke-Jr> why do we get an Esperanto translation but no Tonal? :P

11:36 < Luke-Jr> (excluding list of langs in announcement since it's virtually all)

11:36 < GitHub54> [bitcoin] MarcoFalke opened pull request #7252: [gitian] Set reference date to actual release date (master...MarcoFalke-2015-gitianTime-0.13) https://github.com/bitcoin/bitcoin/pull/7252

11:38 < MarcoFalke> wumpus, do you still rebase backport commits even if they merge cleanly?

11:38 < MarcoFalke> I clean merge should be preferred, but then the commits don't include the cherry-pick information you usually include...

11:40 < wumpus> I don't care much how the commit ends up on the 0.12 branch, if a merge works that's fine with me too...

11:41 < MarcoFalke> E.g. #7245 merges fine whereas mempool_limit.py needs rebase.

11:43 < Luke-Jr> MarcoFalke: the cherry-pick info is kinda useless if you merge anyway

11:43 < Luke-Jr> since you get the same commit hash

11:43 < MarcoFalke> That's why I'd prefer a merge

11:44 < Luke-Jr> merge when possible IMO

11:44 < Luke-Jr> I intentionally try to build my bugfixes on the commit introducing the bug, so they merge cleanly anywhere.

11:53 < wumpus> right, as long as the merge commit has the pull request id in the right format, the changelog script will pick it up, no need to add extra info

13:16 < GitHub47> [bitcoin] luke-jr opened pull request #7253: Bugfix: update-translations: Allow numerus translations to omit %n specifier (usually when it only has one possible value) (master...numerus_omit_n) https://github.com/bitcoin/bitcoin/pull/7253