IRCloggy #git 2007-01-03

Logs Search ←Prev date Next date→ Channels Documentation

Provider of IRC logs since 2005.
WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

2007-01-03

→ pnatal joined00:06
→ GeertB joined00:26
rambokid heck, how do i extract the commit message of an existing commit?00:32
gitster "git show -s $commit"?00:33
rambokid (without the commit...Author:... blurbs of git-log)00:33
gitster sed them out of "git cat-file commit $commit" output then.00:33
rambokid gitster: not good enough, since depending on whether it was a merge or not, the output has multiple header lines00:33
gitster like: git-cat-file commit $commit | sed -e '1,/^$/d'00:34
At least that is "good enough" for git Porcelain-ish.00:35
rambokid gitster: ok thanks, just to make sure, 1,/^$/ will match all lines up to the first empty line and no more, right?00:38
that should work fine then.00:38
gitster most of the thing you would want to script are found in git Porcelain-ish ;-).00:38
rambokid gitster: ok, since you're here... i'm trying to merge commits to two branches. both branches operate in their own dir so there will never be a conflict. the order in which the commits have to be merged are like what gitk --all -d produces. i'm currently using a combination of git-rev-list + reverse commit id order + shell scripted merges for each individual commit. is there anything easier i'm overlooking?00:41
gitster what do you mean by "order in which the commits have to be merged"?00:42
for that matter, "merge commits to two branches" does not make sense to me...00:43
rambokid well, i have two branches (directories of a project) that were developed concurrently and tracked by different revision control systems. i can import both into git with their own history and the same root. now i want to merge them into a single line of commits ordered by dates of the commits.00:44
gitster ah, ok.00:45
rambokid e.g. if one dir is a/ and has commits a1 a2 a3 and the other is b with commits b1 b2 b3 and those commits were executed exactly interleaving in the past, the result should be a1 b1 a2 b2 a3 b3.00:45
gitster is each of the (currently separate, but you are trying to make them into one) development history linear, or do they have their own merges?00:46
rambokid linear and simple.00:46
gitster then I would sort "rev-list --timestamp" output numerically.00:46
rambokid gitster: ah, ok. but still script the mergin, like for r in `cat merged-commit-ids` ; do git-resolve head $r msg; done right?00:47
s/mergin/merge/00:47
gitster something like that, I guess.00:48
rambokid ok, thanks. just checking whether i'm on track here ;)00:48
actually, it looks like cherry-pick is more what i want here and it already does pick up the correct commit message00:52
→ benoitc joined01:04
→ spuk joined01:12
→ ShadeHawk joined01:13
→ kmacleod joined01:15
→ ag joined01:38
→ mchehab joined01:55
→ hon joined02:38
hon git shines. is there an equally great build system being developed?02:40
→ benoitc joined03:04
→ kanru joined04:11
Romster hon, package manager?04:16
or chrooted enviorment to build sources in?04:16
→ xjjk joined04:29
→ xjjk joined04:37
→ jeffpc joined04:37
→ rkaway2 joined05:15
→ russell_ joined05:45
→ devogon joined07:38
→ Romster joined07:59
→ ferdy joined08:37
→ gitster joined08:45
→ ehird joined09:27
→ nud_ joined09:29
← ehird left09:33
→ FrankBA_ joined09:40
→ pdmef joined10:03
→ normalperson joined10:10
→ chris2 joined10:14
→ kanru joined10:20
→ pnatal joined10:34
→ pnatal joined10:36
→ benlau joined10:47
← hon left12:04
→ GeertB joined12:15
→ spuk-_ joined12:32
→ GyrosGeier joined12:50
→ ferdy joined13:00
→ Romster joined13:25
→ Glanzmann joined13:32
Glanzmann Where do I clone Linus Kernel Tree from?13:33
→ timlarson_ joined13:38
moh git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git13:38
Glanzmann Thanks.13:41
→ GyrosGeier joined13:52
→ kanru_ joined14:19
→ FrankBA_ joined14:20
spuk-_ → spuk-14:25
→ Romster joined14:44
kanru_ → kanru14:54
→ cworth joined15:10
→ ArjenL joined15:47
→ krh joined16:09
→ krh joined16:18
krh ugh16:18
stupid wireless16:18
oops16:18
→ GeertB joined16:23
→ kanru joined17:02
→ linuxmigration joined17:10
→ tchan joined17:12
→ lyakh joined17:18
→ GeertB joined17:18
→ chris2_ joined17:23
→ kanru joined17:25
→ Oejet joined17:53
→ mkrufky joined18:20
mkrufky I found quite a nasty bug in gitweb... is git@vger the proper place to report it?18:21
i will give a short description:18:21
ever since the recent update to gitweb on kernel.org, you can no longer view a shortlog on any branch / head other than master18:21
for example:18:21
http://master.kernel.org/git/?p=linux/kernel/git/mkrufky/v4l-dvb-2.6.x.y.git;a=summary18:22
the master branch is current as of immediately before 2.6.20-rc3 ... but i've branched off for each kernel release, to include the stable kernel history for each 2.6.x.y kernel18:22
if you click on one of the branch links, you only get the shortlog for master18:23
it seems to be impossible to view the shortlog for any head / branch other than master18:23
chris2_ → chris219:13
→ robinr joined19:18
→ GyrosGeier joined19:19
→ pedrocr joined19:33
pedrocr is it possible to have utf8 in git filenames? I tried to commit one and I get "error: Compara303247303243oCALM/ComparacaoCALM.tex: does not exist and --remove not passed"19:34
"cg add" works but it then fails on "cg commit"19:34
→ xjjk joined19:43
→ devogon joined20:14
ArjenL → Arjen20:33
← Arjen left20:33
→ cworth joined20:39
→ spuk- joined20:48
jeffpc pasky: you may have noticed by now, I renamed gq to guilt, and set up a new mirror on repo.or.cz21:00
If you want, you can remove the gq mirror21:01
→ mndrix joined21:16
→ Romster joined21:30
→ clee joined21:48
clee hm.21:48
any git gurus around?21:48
→ robfitz joined21:49
clee I'm getting this error from git-prune:21:49
error: Packfile .git/objects/pack/pack-540263fe66ab9398cc796f000d52531a5c6f3df3.pack SHA1 mismatch with itself21:49
that packfile is about 2.3GB21:49
→ nud_ joined21:50
nud_ → nud21:52
nud clee: ask, and someone will answer if he knows.21:53
clee well, I'm mostly wondering how the packfile got that way21:54
I generated it with git-repack21:54
it took over nine hours.21:54
(the .git directory I'm working with here is over 17GB before packing.)21:54
→ lyakh joined22:07
→ dancor joined22:35
dancor i keep getting stuck in a phantom 'Transaction is out of date' state with git svn22:35
robinr clee: I sugggest you ask on the mailing list. Sounds like a serious error somewhere.22:39
← mkrufky left22:47
clee robinr: okay, t hanks23:02
er, thanks23:02
stupid keyboard.23:02
also, I think git-svnimport fucked up.23:02
I generated the git repo with git-svnimport and the .git/svn2git log has over 2600 duplicate revision IDs23:03
not SHA1 commit IDs, but Subversion revisions23:03
dancor how can i debug being in 'Transaction out of date' on 'git svn dcommit' when i'm not out of date23:10
in fact, i have just done 'git svn fetch'23:11
the only solution is to save the diff to a file, reset --hard to long ago, fetch, patch, commit23:12
it's quite ridiculous23:12
normalperson dancor: you fetched, but you didn't update your working HEAD23:12
after fetch, run 'git rebase remotes/git-svn'23:12
git-svn fetch won't ever update your working HEAD23:13
dancor Current branch master is up to date.23:13
but problem persists23:13
i guess the problem is that in the past i have not rebased every time i fetched?23:13
normalperson dancor: probably23:14
well, you should always rebase before you dcommit23:14
dancor oh so that is the right time23:14
normalperson (well, 'always' unless you really know what you're doing :)23:14
dancor i do not23:15
so always23:15
normalperson yes :)23:15
dancor so after commiting to master23:17
rebase23:17
then dcommit23:17
normalperson yes23:17

Logs Search ←Prev date Next date→ Channels Documentation