From kde-i18n-doc Mon Sep 15 13:52:10 2014 From: Chusslove Illich Date: Mon, 15 Sep 2014 13:52:10 +0000 To: kde-i18n-doc Subject: Centralized summit operations (was: Akademy 2014, notes of the Localization BoF) Message-Id: <201409151552.10642.caslav.ilic () gmx ! net> X-MARC-Message: https://marc.info/?l=kde-i18n-doc&m=141078919302387 MIME-Version: 1 Content-Type: multipart/mixed; boundary="--nextPart18327045.upXBMWt5op" --nextPart18327045.upXBMWt5op Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable > # Use summit workflow officially > Some questions / issues > - perf / infrastructure issues Performance is low enough that I don't think centralization is feasible without processing multiple languages in parallel. Here are the numbers for the fr summit (most translated), on a 2.8 GHz Core 2 machine (rounded upwards): messages summit merge: 5 min messages summit scatter: 10 min docmessages summit merge: 3 min docmessages summit scatter: 2 min Performance also depends on what "special" things are done on scatter, as set in per-language summit configs (LANG/summit/messages.extras.summit). E.g. sr summit takes 40 min to scatter, due to post-processing and review state checks. But this could be handled by dropping a language from automatic scatter if it becomes too demanding. > - would it be possible to avoid per team merging by making scripty > directly work on summit branch ? It would, but then nothing "special" could happen on merge. For example, as it is now, one special thing that set for all languages is "fine" wrapping of strings (e.g. line breaking tags, etc). And in per-language configs other such special things happen, e.g. automatic fill-out of some header fields. Other than scatter and merge operations, that are performed per language, there is the gather operation, that is performed on templates (I do this). On the performance side, gather currently takes one hour, due to complicated branch situation (before kf5 branches it took about 10 minutes). Also, gather requires manual intervention whenever a new catalog appears or dissapears (something like 1-3 times per week). This usually takes few minutes at most, but human judgment is necessary. > - people in BoF mostly agree. Need to be discussed. As I observed, a few teams looked into summit in the past and decided not to use it. I don't think it was only because it was not official. =2D-=20 Chusslove Illich (=D0=A7=D0=B0=D1=81=D0=BB=D0=B0=D0=B2 =D0=98=D0=BB=D0=B8= =D1=9B) --nextPart18327045.upXBMWt5op Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) iEYEABECAAYFAlQW7woACgkQMSGXgigGr3HZBQCfUc5YFRBec+zuyZdRogmeydu1 4oYAoIPH9XJDSqc6hoO9gvug/78hthr4 =wvJD -----END PGP SIGNATURE----- --nextPart18327045.upXBMWt5op--