IRCloggy #git 2010-07-06

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.

2010-07-06

→ katol joined00:00
→ anjen joined00:01
→ ryanakca joined00:01
katol Hello, when I try to rebase a branch A with another branch B I get a "Patch Failed at 0001 blah blah blah" where blah blah blah is my commit message but I don't get any conflicted filename. what should the problem be?00:03
← p3rg2z left00:03
← picard left00:03
→ Mage_Dude_ joined00:05
→ Basstard` joined00:06
← adymitruk left00:07
← d0k left00:08
← Wraithan left00:16
→ whaley joined00:16
→ stringo0 joined00:19
← panpainter left00:19
← marienz left00:20
→ marienz joined00:20
← whaley left00:20
→ bcardarella joined00:21
→ jcromartie joined00:22
→ tstellar joined00:22
→ panpainter joined00:22
tstellar Is there a way to have git diff color only trailing whitespace?00:22
jhelwig tstellar: Take a look at the core.whitespace setting. man git-config00:26
Gitbot tstellar: the git-config manpage can be found at http://git.or.cz/man/git-config00:26
→ eletuchy_ joined00:27
→ MattDiPasquale joined00:28
tstellar ok, thanks00:28
← Tommy[D] left00:28
← eletuchy left00:29
← mithro left00:29
← MattDiPasquale left00:30
→ plediii joined00:31
← iravid left00:33
→ rorist_ joined00:35
← rorist left00:37
→ rorist`love joined00:39
← rorist_ left00:41
→ alin joined00:41
→ Tommy[D] joined00:41
→ MattDiPasquale joined00:42
← marienz left00:42
← tstellar left00:42
→ rorist joined00:42
← khelll left00:44
← rorist left00:44
← rorist`love left00:46
→ marienz joined00:46
← kukks left00:50
← MattDiPasquale left00:50
→ iravid joined00:50
→ navetz joined00:50
← mikeric left00:53
→ scarabx joined00:56
← alezandro left00:58
→ patmaddox joined01:00
← ddunbar left01:03
← katol left01:03
← ExtraSpice left01:03
← Tuller left01:04
→ chrislerum joined01:04
→ ddunbar joined01:07
← marienz left01:09
← parasti left01:09
→ _numbers joined01:09
← hakunin left01:10
→ warlock_mza joined01:14
← panpainter left01:14
→ alezandro joined01:17
← Vampire0 left01:19
← warlock_mza left01:19
→ bc0906524 joined01:20
bc0906524 → sleepy_tiger01:20
→ jrmuizel joined01:21
← Bombe left01:29
→ Bombe joined01:29
_numbers how can i delete all history in a git repo past a specific ref?01:31
i accidentally commited a submodule the wrong way and it included all this BS history in the main branch01:31
← chrislerum left01:31
_numbers and i want to get rid of it01:31
but git rebase -i is not doing very well01:31
→ Vampire0 joined01:31
_numbers keeps stopping at like every commit asking to resolve merge conflicts for files that are 'both modified' and 'deleted by us'01:31
davr everything past a certain ref? just use git reset --hard <ref>01:33
_numbers everything that happened before the ref, i mean01:33
not after01:33
davr so you want to remove history before point X? that will only work if the files in X-1 no longer exist at X01:34
_numbers well im okay with squashing them then01:34
but i tried rebase -i and selected squash for all 374 commits01:34
but it has similar problems with merge conflicts01:35
offby1 _numbers: you might get some relief with "git rerere"01:37
if the rebase keeps asking you to resolve the same conflicts over and over.01:37
parndt you can also do a -s recursive -X ours or -s recursive -X theirs depending on which side you want to override the other01:38
_numbers i want ours01:39
engla|ger fixing merging is not the solution. all you want is to make a new root commit, use git reset or so01:39
_numbers ill try that01:39
→ octavio-rdz joined01:39
engla|ger just do git reset --soft <root commit> then you go git commit --amend01:40
If I understand this correctly01:40
no use in doing endless squashing and merging if you know exactly what the end result is going to look like. just use reset01:42
_numbers parndt: error: unknown switch `X'01:42
← ddunbar left01:43
→ codeshepherd joined01:44
parndt _numbers: git --version ?01:44
_numbers: but I'd try engla|ger's suggestion first01:44
_numbers engla|ger: i did it but now Your branch and 'origin/development' have diverged, and have 1 and 750 different commit(s) each, respectively.01:44
parndt: git version 1.7.0.401:44
git log shows no history behind this point tho, which is promising01:45
parndt hm ok i'm on 1.7.1 and have it01:45
_numbers i think its supposed to be -Xown01:45
no space01:45
thats what it says in docs01:45
but i see no example01:45
how do i move back to the first commit when i am at the back?01:45
parndt _numbers: yeah seems like there's more than one way of doing it depending on pull/merge or what you're doing01:46
→ jtdowney joined01:46
← jcromartie left01:46
engla|ger _numbers: yes, your rewritten branch will be very different from its original upstream branch, since you change the history01:47
← patmaddox left01:47
_numbers but im stuck on the first commit and dont see anything afterward01:47
i mean i dont know the git command to move to the end of the 750 diff commits01:47
it basically chopped me down to 1 commit01:48
engla|ger _numbers: the commands I gave you would flatten all the history from the point you were checked out until the <root commit> into only 1 commit, yes01:48
_numbers oh i dont want that unfortunately01:48
→ ddunbar joined01:48
_numbers there are 750 commits01:49
the commit i want is in the middle like #37501:49
every commit between 1 - 374 i want deleted01:49
← aziz left01:49
_numbers everything from 375 - 750 i want to keep01:49
parndt do you have a copy of the repo before you did this01:49
_numbers yep01:49
engla|ger by deleted you mean squashed; the changes kept01:49
well no worries, the reflog will have it all01:49
parndt then why don't you use that copy and do a forced push to origin01:49
_numbers parndt: oh you mean before i got the 375 commits i didnt want? no that copy is very old now. i just noticed it much later.01:50
parndt ah.01:50
engla|ger _numbers: if you check out commmit #347 and repeat these steps you'll have your squashed base01:50
→ simplechat joined01:50
engla|ger _numbers: then you rebase the old history #348->end on top of that01:50
parndt using git you can checkout commits like they're tags can't you?01:50
_numbers engla|ger: what cmds would allow me to reboase old history 348->end onto current?01:51
engla|ger git rebase --onto <newbase> #375 <oldbranch#750> If I recall correctly, check man git-rebase; replace #375 etc with the correct commit identifiers of course01:53
Gitbot engla|ger: the git-rebase manpage can be found at http://git.or.cz/man/git-rebase01:53
→ Transformer joined01:53
_numbers refs i want: 66fc7ed6...1a89ae0201:53
← Transformer left01:53
engla|ger _numbers: you'll have to think for yourself to pull this off I guess. rebase has an option to preserve merges. I'm going to bed, gl hf!01:55
_numbers thx01:55
← engla|ger left02:00
→ hakunin joined02:05
← simplechat left02:07
_numbers git rebase --onto development_clean 2f366e3 1a89ae # keeps getting the crap i dont want from 1 - 37502:08
→ lazylisper joined02:10
← hauserserser_ left02:11
← sagsousuke left02:12
← kpreid left02:13
→ chrislerum joined02:15
← coppro left02:16
→ coppro joined02:17
jmspeex With a project containing submodules, how can I tell git to pull everything -- including the submodules?02:18
← orafu left02:18
jmspeex doing "git submodule update" does nothing02:19
→ consolers joined02:19
coppro submodules are fixed at a specific version02:19
jmspeex Well, that specific version got advanced02:20
→ cytrinox_ joined02:20
← mw left02:21
jmspeex i.e. from one machine, I modified the submodules, I checked in and pushed the changes, then I checked in the toplevel project and pushed it02:23
← cytrinox left02:23
cytrinox_ → cytrinox02:23
jmspeex Then on my other machine, I did a git pull followed by a "git submodule update", but I still don't have the latest version of the submodules02:24
← _numbers left02:26
jmspeex Also, gitk tells me "Submodule celt c7bcf38...b5077ef (commits not present)"02:26
(same for the other submodule)02:26
→ alester joined02:29
← Elfe left02:31
→ Elfe_ joined02:31
→ whaley joined02:34
→ panpainter joined02:34
← scarabx left02:40
← shuaib left02:43
jmspeex Just tried "git submodule foreach git pull" and it worked. Is that the correct way to do this?02:45
← aresnick left02:45
→ dtangren joined02:45
← codeshepherd left02:48
← timj_ left02:52
→ timj joined02:52
→ phantomcircuit joined02:54
← n3kl left02:58
← jtdowney left03:00
→ n3kl joined03:00
← n3kl left03:00
→ n3kl joined03:00
← pabloh left03:00
→ pabloh joined03:04
→ \mjd joined03:05
← hearit left03:08
→ aluink joined03:10
→ codeshepherd joined03:11
← dtangren left03:12
→ Wraithan joined03:21
Wraithan is there a command to no matter what branch you are on, only push the current branch?03:22
I am sick of getting stuff like ! [rejected] live -> live (non-fast-forward) in cases when I just need to go to live branch and pull03:23
I've solved it for now by having a push alias that is 'git push origin' then doing push master or push live03:23
but was hoping there was a more generic way, so I didn't need to specify03:24
→ a9b joined03:26
→ multigrain joined03:26
→ hyperair joined03:29
← hakunin left03:29
← warthog9 left03:31
← bcardarella left03:34
\mjd Wraithan: git push remote something03:34
Oh, I see.03:35
Wraithan yes, I know that one03:35
\mjd Yes, there is a configuration option.03:35
Just a second.03:35
Wraithan That'd be awesome03:36
→ jony joined03:37
← bamboo434 left03:38
\mjd "... to default to pushing only the current branch to origin use git config remote.origin.push HEAD. "03:39
(That's from the git-push manual.)03:39
remote.<name>.push03:40
The default set of "refspec" for git-push(1). See git-push(1).03:40
03:40
→ urkud__ joined03:43
→ benbinary joined03:45
← vol7ron left03:45
← benbinary left03:46
← hyperair left03:47
→ t0rc joined03:50
← jrmuizel left03:52
→ patmaddox joined03:53
← urkud__ left03:54
← batok left03:57
→ hyperair joined03:58
← patmaddox left03:59
→ doener joined03:59
→ coderdad joined03:59
← whaley left04:03
→ yashi joined04:04
← hyperair left04:04
→ zehrique joined04:08
→ whaley joined04:12
→ hyperair joined04:14
→ rado1 joined04:15
← pabloh left04:15
← apsysoev left04:18
← alezandro left04:20
← yashi left04:20
← octavio-rdz left04:21
→ yashi joined04:22
← \mjd left04:22
→ patmaddox joined04:27
← ignatp left04:31
← molokai left04:35
→ molokai joined04:36
→ OFICINA7 joined04:40
ben_h → ben_h_04:40
ben_h_ → ben_h04:41
← slonopotamus_ left04:43
← patmaddox left04:47
→ brothers joined04:47
← Joeconyers left04:47
← zehrique left04:48
← codeshepherd left04:51
→ slonopotamus_ joined04:51
slonopotamus_ → slonopotamus04:52
→ codeshepherd joined04:53
← Kage left04:53
→ g3d joined04:55
← marc_kdab left04:56
← Bass10 left04:56
→ Heimidal joined04:56
→ mbroeker joined04:57
→ sftp_ joined04:58
← sftp left04:59
→ freimatz joined05:00
← Yuuhi left05:01
→ Yuuhi joined05:02
→ LongBeach joined05:03
← t0rc left05:05
← plediii left05:08
← coderdad left05:09
→ joachimbean joined05:10
← karptonite left05:11
← joachimbean left05:11
← molokai left05:12
→ molokai joined05:14
→ jonmasters joined05:14
← pallaert left05:15
→ karptonite joined05:16
← karptonite left05:16
→ vu3rdd joined05:19
→ psankar joined05:24
← cjs left05:24
→ girishr joined05:24
→ fmardini joined05:25
→ ezyang joined05:25
← hyperair left05:26
ezyang Hey all, did the semantics of `git diff-files --name-only` change between 1.6 and 1.7?05:26
I was under the impression that it was only supposed to show file differences between the index and the working copy, but my 1.6 copy of Git is showing all files.05:26
← stringo0 left05:28
← jsquared left05:29
ezyang huh, and now it's working. Must have been a filesystem hiccup.05:30
→ maleknet joined05:30
→ dnquark joined05:32
blueshift maybe PEBKAC05:32
dnquark so I think I lost a day's worth of work thanks to git05:32
(I helped)05:32
doener ezyang: IIRC diff-files doesn't update the index, so maybe it was just out of date05:33
← OFICINA7 left05:33
dnquark can someone explain to me why git might randomly delete a file?05:33
← coppro left05:33
slonopotamus dnquark: it doesn't05:33
NfNitLoop dnquark: Yeah, best it can do is pseudorandomly delete a file.05:34
:p05:34
dnquark slonopotamus: a fan of vinni-puh, i assume?05:34
doener dnquark: if you committed your stuff, it's pretty hard to actually lose it with git05:34
slonopotamus dnquark: just was in good mood when was choosing nickname05:34
dnquark so here's how it went down05:34
← programble left05:34
dnquark one of my machines runs windows05:35
← lazylisper left05:35
→ coppro joined05:35
dnquark I did a git pull, but that file was open05:35
I got an error: git check-out index unable to create file foo (file exists)05:35
← Wraithan left05:35
dnquark I closed the file, did a git commit and pull05:35
at which point it deleted the file from HEAD05:36
NfNitLoop dnquark: aaah, this is likely a problem with upper/lower case names on Windows.05:36
dnquark this I didn't notice until much later05:36
NfNitLoop You probably have 2 files in a directory that have the same name (but differ in case).05:36
dnquark so the question is -- why did git delete this file... I'm sure it was checked in05:36
→ lazylisper joined05:36
NfNitLoop git doesn't detect this. It'll happily write one, then write the other (over the first)05:36
doener NfNitLoop: that doesn't delete the file though, does it?05:37
dnquark no, no case problems here...05:37
← mbroeker left05:37
NfNitLoop Oh, hrmm, ok.05:37
ezyang doener: But git status was giving the right... oh?05:37
→ mtkd joined05:37
doener ezyang: status updates the index05:37
NfNitLoop doener: actually, you're right, it doesn't delete it. Just shows some *weird* behavior.05:37
ezyang interesting.05:37
dnquark after that, when I did a git pull, it had a conflict w/ file deleted in HEAD and modified in ....05:37
ezyang So what do I have to do to manually update the index?05:37
doener ezyang: git update-index --refresh IIRC05:37
ezyang cool, thanks!05:37
→ mbroeker joined05:38
blueshift dnquark: does a git reset --hard help?05:38
dnquark not realizing what was going on I stupidly did a git checkout --theirs and a commit, effectively blowing away my stuff05:38
doener ezyang: don't trust me on that one though05:38
slonopotamus dnquark: 'deleted file from HEAD', 'why did git delete this file' ???05:38
dnquark so i'm still trying to understand why the file got deleted from HEAD when it did05:38
slonopotamus blueshift: git reset --hard causes dataloss :)05:39
dnquark and what the correct course of action should have been05:39
(probably git reset --hard)05:39
blueshift dnquark: if it was in git somewhere some time you should be able to get it back05:39
dnquark hm05:40
→ pedrobelo joined05:40
← jfkw left05:41
dnquark ok, so this whole process started with a commit05:41
thank god for screen scrollback buffers05:41
blueshift and the reflog!05:42
dnquark yeah, well, that too i guess05:42
→ apsysoev joined05:42
→ urkud__ joined05:43
→ jsquared joined05:44
→ cilly joined05:44
← consolers left05:44
← panpainter left05:44
← codeshepherd left05:45
→ codeshepherd joined05:47
ezyang doener: Ok, I guess I'll check the git status source code.05:48
→ marc_kdab joined05:49
dnquark should I use git reset <hash> <filename> or git checkout <hash> <filename>?05:49
← jsquared left05:50
→ gebi joined05:51
← girishr left05:51
dnquark ok, crisis averted05:51
even though I *still* have no clue why git decided that that file needed to be deleted in HEAD05:52
→ shugart joined05:53
shugart okay, so I'm stumped, what am I doing wrong. I commit a ruby on rails framework, but it's only adding the top level directory and no subdirectories05:54
ezyang It looks like the precise semantics of git status is `git update-index --refresh --unmerged -q05:54
Does this sound plausible?05:54
possibly needing the refresh after unmerged and q05:55
← seanius left05:55
→ seanius joined05:57
← lazylisper left06:00
→ theprojectabot joined06:02
→ HG` joined06:03
→ jsquared joined06:03
→ girishr joined06:04
← theprojectabot left06:06
→ theprojectabot joined06:07
← theprojectabot left06:08
← girishr left06:11
shugart nevermind got it working, silly me forgot that I was pulling from a git repo in a subdirectory, and thus messed me up06:11
← codeshepherd left06:12
← didlix left06:12
→ roop joined06:13
← Mage left06:13
← a9b left06:13
→ a9b joined06:15
← shugart left06:15
→ drizzd joined06:16
→ adripop joined06:17
→ warthog9 joined06:17
← adripop left06:18
→ cbreak_work joined06:19
→ eighty4 joined06:20
NfNitLoop → estas06:20
estas → NfNitLoop06:20
→ codeshepherd joined06:21
KB1JWQ → Corey06:22
← slonopotamus left06:23
Corey → KB1JWQ06:26
← madewokherd left06:27
← warthog9 left06:28
← rado1 left06:30
→ girishr joined06:31
→ [Po]lentino joined06:32
→ psoo joined06:32
← _Vi left06:37
← girishr left06:38
→ bababfds joined06:39
← Yuffster left06:40
← enherit left06:41
→ warthog9 joined06:41
← gebi left06:42
← hagabaka left06:42
← molokai left06:48
→ thierryp joined06:49
→ codeshepherd_ joined06:49
→ girishr joined06:49
→ slonopotamus joined06:50
→ luist__ joined06:50
→ dreiss joined06:51
→ lhz joined06:51
← codeshepherd left06:53
codeshepherd_ → codeshepherd06:53
→ ExtraSpice joined06:53
← luist_ left06:54
← ereslibre_laptop left06:56
→ poseid joined06:59
← locklace left07:01
← thierryp left07:02
→ thierryp joined07:02
AAA_awright Is there any method or plan of signing commits instead of using the commit name/email or author name/email?07:03
Why would you want to only be able to sign tags and not commits?07:03
← multigrain left07:04
← timyhac left07:04
→ NET||abuse joined07:05
→ dr_win joined07:06
← thierryp left07:07
→ thierryp joined07:07
→ trivol joined07:07
→ locklace joined07:07
→ pantsman joined07:09
← thierryp left07:12
→ thierryp joined07:13
← spearce left07:13
johnw then what happens if you try to rebase, and one of the affected commits is signed?07:13
← benjo2 left07:14
→ benjo2 joined07:14
→ girishr_ joined07:15
→ shruggar joined07:15
→ rchavik joined07:15
shruggar can anyone explain the --cached option of diff-index? It says "do not consider the on-disk file at all", but I don't understand in what situations it otherwise /would/ consider the on-disk files07:16
← benjo2 left07:17
← girishr left07:17
← roop left07:17
→ roop joined07:18
→ giskard joined07:19
blueshift johnw: signed? you mean signed-off?07:19
johnw i think he meant gpg-signed07:20
→ Pupeno joined07:20
blueshift oh, I didn't see that you were answering someone else07:20
Pupeno Is git's "database" ACID?07:20
← centr0 left07:20
← girishr_ left07:21
← thierryp left07:21
→ thierryp joined07:21
→ girishr joined07:21
parndt → parndt_away07:23
shruggar Pupeno, should be, so long as you use standard git tools to access it07:23
Pupeno Is there any docs about it? To show to someone else.07:24
→ rox` joined07:25
→ timyhac joined07:25
→ steffkes joined07:26
→ ProperN[out] joined07:27
← ProperN[out] left07:28
→ ProperN[out] joined07:28
← thierryp left07:28
→ dpino joined07:28
→ Nickeeh joined07:30
→ mastro joined07:30
→ lacostej joined07:30
← ProperNoun left07:31
ProperN[out] → ProperNoun07:31
shruggar I don't seem to be able to find anything from quick googling and grepping07:32
→ priidu_ joined07:32
→ thierryp joined07:32
← tildeequals left07:33
→ masterkorp joined07:33
→ tg joined07:34
shruggar there are some things which imply it, such as docks about lockfiles07:34
← carutsu left07:34
← dahankzter left07:35
shruggar and for the most part the git object model ensures safety07:35
lacostej Cloning the ubuntu lucid kernel repository hangs on my machine. I've tried 3 times and it hanged each time between 3 and 15% of the receive process. Nothing seems to happen. The screen isn't refreshed. No network activity. Maybe I should let it hang more (I waited up to 30 min I guess). Using 1.7.0.4. I am trying again. If it happens again should I do something special ?07:37
← thierryp left07:37
→ hyperair joined07:39
→ galderz joined07:42
→ xiong joined07:45
← urkud__ left07:48
doener shruggar: see the "RAW OUTPUT FORMAT" section in the man page. diff-index normally does a diff between a tree and the working tree07:50
→ gemmes_ joined07:50
→ ereslibre joined07:51
doener shruggar: ... and at least IIRC that section is actually right. The diff-* manpages tend to confuse me07:51
← gemmes_ left07:51
← johnw left07:51
shruggar ah, I thought that's what diff-files was for,,, so diff-index, by default, does not use the index? o.O07:51
lacostej hang again at 34% this time.07:53
← shruggar left07:54
wereHamster lacostej: what's the url?07:54
lacostej git://kernel.ubuntu.com/ubuntu/ubuntu-lucid.git07:55
→ project2501b joined07:56
← timyhac left07:57
← codeshepherd left07:57
← dreiss left07:57
→ dahankzter joined07:58
→ urkud__ joined08:00
← pedrobelo left08:01
→ pedrobelo joined08:01
→ tildeequals joined08:01
← pedrobelo left08:02
lacostej wereHamster: I'll let it hang for a while. Going for a tour. Back in about 1h. I'll read the messages here if there are any.08:04
→ shruggar joined08:05
← Mage_Dude_ left08:05
→ tar- joined08:06
→ barismetin joined08:09
→ freshtonic_ joined08:09
→ Mage_Dude_ joined08:10
← freshtonic left08:12
freshtonic_ → freshtonic08:12
→ ph^ joined08:13
→ gergap joined08:14
← Bombe left08:16
→ Bombe joined08:17
→ lazylisper joined08:17
← Nickeeh left08:20
← freshtonic left08:20
→ sftp joined08:21
← sftp_ left08:21
→ thierryp joined08:22
← lanthan left08:22
→ mrtazz joined08:23
← chrislerum left08:24
← pantsman left08:24
→ ph^_ joined08:25
→ lanthan joined08:26
← priidu_ left08:26
→ Ilari_antrcomp joined08:26
← poseid left08:27
→ tatsuya joined08:28
→ priidu joined08:28
← seanius left08:31
→ ntoll joined08:34
doener shruggar: diff-files is worktree against index, not against some tree08:34
shruggar 's head explodes08:35
shruggar so: "diff-index -q HEAD && diff-index -q --cached HEAD" would be "is the work-tree or the index different from HEAD", right?08:36
→ seanius joined08:37
← jds left08:38
→ cmyk joined08:39
cmyk hi08:39
I am following this guide: http://scie.nti.st/2007/11/14/hosting-git-repositories-the-easy-and-secure-way08:39
→ ProperN[out] joined08:40
Ilari_antrcomp cmyk: Gitosis is obsolete.08:40
cmyk oh08:40
← ProperN[out] left08:40
→ ProperN[out] joined08:40
← Zouppen left08:42
→ khelll joined08:43
← ProperNoun left08:44
ProperN[out] → ProperNoun08:44
Ilari_antrcomp cmyk: Use gitolite instead.08:45
shruggar should we go around all the google results and tell people "Note: gitosis is obsolete. Users are following your guide without this knowledge, it may be helpful if you were to note it at the top of <url>"08:46
adrien I used git-reset something but went one step too far: I pushed a commit to a remote server and reset to the commit before that one, is there a way I can get the commit back? (otherwise I'm not going to be able to push anymore I guess)08:46
I have the commit id but I'm not sure what to do08:46
shruggar adrien, use "git reflog" to find where you previously were08:46
cmyk Ilari_antrcomp: All I want to do is run a git server on my os x box, it's only going to be intranet08:47
adrien can I simply "checkout" the commit?08:47
→ tommie joined08:48
tommie hm, why did my etckeeper start acting strangely?08:48
| warning: LF will be replaced by CRLF in .gitignore08:48
| warning: LF will be replaced by CRLF in etckeeper/commit.d/50vcs-commit08:48
why would git do such a thing to my /etc files?08:48
→ curtana joined08:49
Ilari_antrcomp cmyk: Gitolite is quite good for just that.08:49
shruggar adrien, try: git reset --keep <the-commit>08:49
cmyk Ilari_antrcomp: okay, thanks08:49
Ilari_antrcomp (another computer because the other one is running 8 POV-Rays in parallel)...08:50
sitaram up_the_irons: ping08:50
← tommie left08:50
→ codeshepherd joined08:51
adrien shruggar: great, thanks a lot =)08:52
← seanius left08:53
→ seanius joined08:53
← Heimidal left08:55
→ dajero joined08:57
Ilari_antrcomp cmyk: AFAIK, the feature set of gitolite is strict superset of that of gitosis. Well, Gitolite doesn't support multi-keys, but has another way to get the same effect.08:58
cmyk I am a complete newb to git08:58
wereHamster cmyk: http://progit.org08:59
cmyk awesome, thanks wereHamster09:00
→ Nickeeh joined09:02
→ ustunozgur joined09:03
→ jtdowney joined09:04
← thierryp left09:06
The_Ball Is there any UI tool where I can look at a git tracked file and see where every line has "come from"?09:08
→ thierryp joined09:09
→ bluenovember joined09:10
→ thierryp_ joined09:10
_rane git blame09:11
← thierryp left09:13
→ hohum joined09:13
→ LucasWilcox joined09:14
← urkud__ left09:15
← Stummi left09:15
→ urkud__ joined09:15
→ Stummi joined09:15
← ramok left09:16
→ ramok joined09:16
← seanius left09:19
→ ThiefMaster joined09:19
← ThiefMaster left09:19
→ ThiefMaster joined09:19
→ seanius joined09:19
→ aiju joined09:19
aiju something really weird happened.. somehow git seems to have mixed several commits together, now some files disappeared, some are full of conflicts.. just plain WTF09:20
shruggar did you try to merge several commits together? :)09:20
aiju somehow i was in "no branch"09:21
shruggar what were you doing when it happened?09:21
aiju i don't really know.. just noticed it09:21
i did some commit reverting yesterday, which seems to be the cause of it09:22
git log is perfectly fine though09:22
shruggar yes, a revert is a type of merge09:22
wereHamster more like cherry-pick09:23
aiju changes i did decades ago seem to be reverted and it's all plain wtf09:23
wereHamster aiju: gitk --all09:23
curtana or at worst, reflog09:24
aiju looks kinda funny09:24
sadly, it lacks recent changes09:25
which do appear in git log09:25
shruggar gitk --all shows the notes ref... perhaps it shouldn't?09:25
← Textmode left09:26
aiju i just cloned the local repo, now it looks fine09:26
→ thierryp joined09:26
aiju still i'd like to repair the original one09:26
← thierryp_ left09:27
lacostej still stuck...09:27
aiju git stash did the job09:28
it's back to the way it was this morning09:28
(i.e. i have my files back)09:29
how to merge those two branches now?09:30
the "no branch" and master09:30
→ ijuma_ joined09:30
lacostej what's the proper way to debug a git clone ? Can I gdb into it to get a stack trace ? Can I use this: https://git.wiki.kernel.org/index.php/Aliases#A_.27debug.27_alias_to_help_debugging_builtins ?09:30
← dnquark left09:31
Ilari_antrcomp lacostej: git clone segfaults?09:31
lacostej no it hangs09:31
Ilari_antrcomp lacostej: What protocol?09:31
← MacGyverNL_ left09:31
lacostej Ilari_antrcomp: git clone git://kernel.ubuntu.com/ubuntu/ubuntu-lucid.git09:32
→ rweng joined09:32
Ilari_antrcomp lacostej: So git://... Try 'telnet kernel.ubuntu.com 9418' and see if it connects...09:32
Nugget_ telnet is eeeeeeevil!09:32
aiju Nugget_: so is ftp.. still people use it09:32
lacostej works09:32
aiju (my CS teacher called it "state of art")09:33
rweng hi, can anyone give me a hint what i am doing wrong? i got a file latex/.gitignore with content "*.out" - however, if i do "git status" the file latex/test.out appears09:33
Ilari_antrcomp lacostej: Git clone might have verbosity options...09:33
lacostej thing is that I need to download 900M, so it takes quite a bit of time at 250kbits09:33
← thierryp left09:33
shruggar rweng, appears under what section?09:33
rweng Changes to be committed: new file: latex/test.out09:34
uh, could it be because i added it before?09:34
shruggar rweng, sounds like it was added before, yes09:34
Ilari_antrcomp lacostej: It hangs mid-transfer?09:34
rweng arg, how dumb of me09:34
aiju how to switch branch?09:34
shruggar rweng, git rm --cached to remove it from the index09:34
rweng shruggar: thanks!09:34
lacostej yes. Now at "Receiving objects: 34% (594474/1737004), 293.72 MiB | 72 KiB/s " Last time was 13%. Tried 4-5 times09:34
wereHamster lacostej: I could give you a bundle which you can download through http..09:34
lacostej hehe. I am sure I can get it work. I just want to find out why :)09:35
→ tvw joined09:35
Ilari_antrcomp And git:// does not have debug mode that prints all sorts of connection parameters... :-/09:35
lacostej gdb bt doesn't give me much...09:37
→ romgit joined09:37
← Sjors left09:37
Ilari_antrcomp lacostej: Strace would probably show it blocking on read of network connection...09:37
aiju could i just delete the master branch and rename the second one "master"? :D09:38
lacostej probably but then the git executable should be able to recover from transient connection errors, right ?09:38
I wonder about the context of the blocking read09:38
Ilari_antrcomp lacostej: Git clone can't recover from connection going down.09:38
lacostej ah...09:39
then I get it09:39
that's bad. When you're downloading 1G...09:39
Ilari_antrcomp lacostej: TCP can recover from brief interruptions, but not systematic packet lossage.09:39
lacostej: Resuming download is HARD.09:39
Otherwise there would already be resumable fetch/clone.09:40
lacostej: And network sniffing would probably reveal your computer trying to repeatedly ACK the partial transmission with nothing coming back.09:40
aiju how to delete commits?09:41
Ilari_antrcomp aiju: reset.09:41
aiju my local repository is completely fine now.. i can't push it though09:41
Ilari_antrcomp aiju: Not fast-forward?09:41
aiju pardon?09:41
Ilari_antrcomp aiju: What's the rejection message for branch update? Non-fast forward?09:42
aiju yeah09:42
Ilari_antrcomp aiju: Either merge the changes from upstream or force the push (loses data).09:42
→ a_cezam joined09:42
→ oriba joined09:43
aiju -f did the job, ty09:43
← oriba left09:43
Ilari_antrcomp aiju: Wonder what the heck did you nuke from remote...09:43
aiju broken stuff09:43
"git revert" accidents09:43
Ilari_antrcomp aiju: If others have fetched that, they are now annoyed.09:43
aiju assuming, that others exist..09:44
→ RORUBUN joined09:44
a_cezam hi09:46
→ MacGyverNL joined09:47
→ felipec joined09:47
← slonopotamus left09:47
→ Dave^| joined09:47
a_cezam is there anyone?09:48
shruggar a_cezam, just ask your question09:48
a_cezam ok09:49
codejunky hides09:49
aiju a_cezam: no, we're all just a bunch of bots09:49
codejunky ;-)09:49
aiju beeps09:49
→ intripoon joined09:51
→ Paraselene_ joined09:53
← a_cezam left09:53
lacostej Ilari_antrcomp: to close this thread, reading past threads regarding resumable clones. Found this: https://git.wiki.kernel.org/index.php/SoC2009Ideas#Restartable_Clone . I will try again fetching my repo. Thanks for your time09:54
shruggar yeah, I thought it was a GSoC idea09:56
but there are other things (like GitTorrent) which I'd rather see, and which would implicitly solve the restartable clone problem09:56
rajeshsr_ when git doc uses working tree does it mean the whole git tree or the current working directory?09:57
→ Titosemi joined09:57
shruggar it generally means the entire tree09:58
rajeshsr_ shruggar, ok, thanks09:58
→ sagsousuke joined09:58
← henkboom left09:59
rajeshsr_ Is it ok to switch branches without commiting changes? i sometimes get error, sometimes don't. Can anyone clarify about this behavior?09:59
cmyk would this work?: git config --global core.editor mate09:59
shruggar rajeshsr_, you'll get errors if files you have modified differ between the current HEAD and the branch you're switching to10:00
cmyk got this after adding the message and saving: Aborting commit due to empty commit message.10:00
→ gebi joined10:00
→ henkboom joined10:02
shruggar rajeshsr_, it's usually a case of "it's okay to do that, but why are you doing it?" because it may be that you're going about things the wrong way. For example, relevant questions might be: why are you leaving them un-committed? Why did you start editing them when in the wrong branch? Why are you checking out another branch if that's not the one you're working on?10:02
Ilari_antrcomp cmyk: The editor must wait until it is closed (you presumably need some command-line option for the editor command).10:02
cmyk i'll look at the tm bundle10:02
rajeshsr_ shruggar, yeah, those are good questions! I need to think about them, to understand my usage patterns! :)10:04
← felipec left10:04
→ icwiener joined10:04
→ thiago_akademy joined10:05
→ felipec joined10:05
rajeshsr_ So, simple said it causes error when their HEAD of one differs from the modification you made? So untracked files don't have a problem?10:05
shruggar rajeshsr_, untracked files have a problem if they exist in the branch you're switching to10:05
← felipec left10:06
← lacostej left10:06
→ felipec joined10:07
rajeshsr_ shruggar, well, with untracked files i never see a problem. It is untracked in all branches..10:07
← a9b left10:07
shruggar rajeshsr_, if it's untracked in both the branch you're on, and the branch you're switching to, then it's not a problem10:07
← aiju left10:08
← hobodave left10:08
rajeshsr_ ha, you mean if it is commited in one while it is untracked in another. Is that even possible?!10:09
shruggar rajeshsr_, of course it's possible, why wouldn't it be?10:09
rajeshsr_, if you add a file in the branch you're currently on, and commit, what would you expect to happen to all the other branches?10:10
← dpino left10:10
_Unode → Unode10:10
rajeshsr_ they won't be visible to other branches as their HEAD points to something else..10:10
shruggar the other branches won't be effected in any way10:11
→ tizbac joined10:11
shruggar (affected?)10:11
rajeshsr_ i think so. they won't get affected..10:11
← anjen left10:12
→ thierryp joined10:12
→ mital joined10:13
shruggar right, so if you were on one of those other branches, that file wouldn't be tracked as far as you're concerned, so if you created it, it would be "untracked files:". But switching to the branch where you did add+commit would complain "untracked files would be overwritten"10:13
→ trash joined10:14
← codeshepherd left10:14
wereHamster cmyk: mate -w10:14
→ codeshepherd joined10:14
← thierryp left10:14
→ thierryp_ joined10:14
← thierryp_ left10:15
← romgit left10:15
→ rolfb joined10:16
← urkud__ left10:16
← rchavik left10:17
→ rchavik joined10:17
→ elyob joined10:19
mital hey I am trying to setup git daemon.. I found this blog post : http://rockfloat.com/blog/?id=40 ... when I run "sudo runsvdir /etc/service &" it gives me error "runsv git-daemon: fatal: unable to lock supervise/lock: temporary failure"10:19
→ lucsky joined10:19
mital can anybody help me what should I do ?10:19
I am also not able to find script /etc/init.d/git-daemon-run10:20
→ didlix joined10:21
→ urkud__ joined10:22
← codeshepherd left10:25
→ d0k joined10:25
→ dpino joined10:25
→ codeshepherd joined10:25
→ Belna joined10:28
→ kamilo00111 joined10:28
← kamilo00111 left10:29
← rweng left10:31
← Nickeeh left10:32
→ level09 joined10:32
level09 i want to ignore all hidden dirs/files except .htaccess , will this work :10:32
*.10:32
← dajero left10:32
level09 !*.htaccess10:32
Ilari_antrcomp level09: '.*' and '!.htaccess'?10:33
level09 Ilari_antrcomp: my concern is .. will it look for it recursively ?10:33
→ dajero joined10:33
Ilari_antrcomp level09: If you don't want .htaccess to be recursive rule: '!/.htaccess'.10:34
level09 okay thanks ! actually I want it to be recursive10:35
mital am I missing any package ?? : aptitude shows me : git-common and git-daemon-run installed10:35
I am not able to find script /etc/init.d/git-daemon-run10:36
Ilari_antrcomp mital: It uses some weird init system.10:36
mital: (blame git-daemon-run for that).10:36
wereHamster runit10:36
mital hm10:36
what should I do now ?10:36
Ilari_antrcomp mital: Edit the startup script... Its there somewhere...10:37
→ oriba joined10:38
← sleepy_tiger left10:38
mital when I run "sudo runsvdir /etc/service &" it gives me error "runsv git-daemon: fatal: unable to lock supervise/lock: temporary failure"10:38
level09 when using git rm .* why I'm getting error10:38
'..' is outside repository10:38
← dajero left10:38
mital Ilari_antrcomp: do u know what reason this could be ?10:38
level09 i want to remove hidden dirs files from repo10:39
*from tracking rather10:39
Ilari_antrcomp level09: Shell expands .* to ..10:39
wereHamster level09: --ignore-unmatch10:40
Ilari_antrcomp mital: Something's locking that lock.10:40
level09 wereHamster: thanks ! trying ..10:40
→ thierryp joined10:40
← giskard left10:41
→ lacostej joined10:43
← codeshepherd left10:44
← oriba left10:44
→ Nickeeh joined10:45
→ vegeste joined10:46
← peper left10:46
← lazylisper left10:49
→ consolers joined10:53
→ Joeconyers joined10:53
← thierryp left10:59
← Titosemi left10:59
← seanius left10:59
← jtdowney left10:59
← lhz left10:59
← psoo left10:59
← eighty4 left10:59
← cilly left10:59
← slakmagik left10:59
← disappearedng left10:59
← adrien left10:59
← cwillu_at_work left10:59
← xiong left10:59
→ hukl joined11:00
→ aziz joined11:00
→ codeshepherd joined11:01
← keyvan left11:02
→ thierryp joined11:02
→ Titosemi joined11:02
→ seanius joined11:02
→ jtdowney joined11:02
→ lhz joined11:02
→ psoo joined11:02
→ eighty4 joined11:02
→ cilly joined11:02
→ slakmagik joined11:02
→ disappearedng joined11:02
→ adrien joined11:02
→ cwillu_at_work joined11:02
← mbroeker left11:03
→ darkskiez joined11:03
darkskiez is there something equivalent to "git branch --contains " in gitweb ?11:04
← g3d left11:04
← thierryp left11:05
← Titosemi left11:05
← seanius left11:05
← jtdowney left11:05
← lhz left11:05
← psoo left11:05
← eighty4 left11:05
← cilly left11:05
← slakmagik left11:05
← disappearedng left11:05
← adrien left11:05
← cwillu_at_work left11:05
→ hearit joined11:06
← shadowmaster left11:06
→ keyvan joined11:08
→ dajero joined11:09
← lacostej left11:09
→ cilly joined11:09
→ slakmagik joined11:09
→ adrien joined11:09
→ seanius joined11:10
mital where would be git configuration file... I want to enable receive-pack service11:10
can anybody help me11:10
→ jcromartie joined11:11
shruggar mital: http://progit.org/book/ch4-9.html11:11
Ilari_antrcomp mital: Do you REALLY want to do that?11:12
shruggar though I'm not sure if "service" is what you really want11:12
→ psil joined11:12
Ilari_antrcomp mital: git-daemon has absolutely NO AUTHENTICATION.11:12
← urkud__ left11:13
→ psoo joined11:13
→ Titosemi joined11:13
mital hmm I have no idea how to setup the authentication bit... so just wanted to check with no authentication it works... then I can start the authentication bit11:13
→ thierryp joined11:14
shruggar mital, what do you actually want to do?11:14
→ eighty4 joined11:14
mital I am setting up a local git server...11:14
shruggar mital, for what purpose?11:14
→ lhz joined11:14
mital I want to host some of my projects locally ... and want to access them from different machines/ os etc..11:15
may be more people will be accessing it11:15
Ilari_antrcomp mital: Install Gitolite.11:15
→ disappearedng joined11:15
→ cwillu_at_work joined11:16
→ icwiener_ joined11:20
← icwiener left11:22
← Nickeeh left11:26
← keyvan left11:26
← Joeconyers left11:28
Aaaarg what do I do if want master to point to a previous commit?11:29
Ilari Aaaarg: reset?11:29
avar create a new commit with that as the parent, watch git implode ! >:)11:29
→ kukks joined11:30
→ Kage joined11:30
shruggar smacks avar11:31
shruggar otherwise known as a "branch"11:32
→ lacostej joined11:34
← codeshepherd left11:34
→ hauserserser_ joined11:34
→ codeshepherd joined11:35
Aaaarg yep git reset, thanks Ilari11:35
→ elijahbal joined11:37
elijahbal Hello.11:38
→ another_newbie joined11:38
shruggar that's my name, don't wear it out11:38
← consolers left11:38
elijahbal I would like to maintain a C project, but I want to do by splitting the files in separate files.11:38
How to 1) create a branch 2) merge in this case ?11:39
I don't see how to merge with the master branch if I split the initial project in many files.11:39
shruggar elijahbal, can you rephrase the first line? You seem to be using the word "files" in a way I don't understand11:40
elijahbal I have a project with big files (typically one thousand lines), and I think it's not maintanable.11:40
I would like to create one file for each function in the project.11:40
That means split the project in many small files.11:41
What do you think of it ?11:41
shruggar Okay, have you tried a small test to see what happens?11:41
elijahbal I didn't try a test.11:41
No11:41
wereHamster git doesn't care whether a file has 100 or 1000 lines11:41
not at that small scale anyway11:42
elijahbal So the merge can be done anyway ?11:42
shruggar elijahbal, in what direction?11:42
elijahbal Listen this is a test case11:42
I have a program foo, with functions main(), a(), b(), [...], z() into it.11:43
shruggar Git should support this kind of merging, in theory, but I don't know how theoretical that support is11:43
elijahbal All this functions are stored in file called "foo.c"11:43
Now what I want to do is to let main() in "foo.c", and create as many files as functions : a() in "a.c", b() in "b.c", [...] , z() in "z.c".11:44
→ Nickeeh joined11:44
elijahbal But mainstream make a change in foo.c11:44
But mainstream make a change in foo.c in the functions x().11:45
How can I propagate this modification to the file "x.c" ? Do you think it's possible ?11:45
wereHamster I don't tink git is that smart yet, but create a test case and try it11:45
← bmalee left11:47
avar Why would you have a C project with 1 function per file anyway? That's just crazy.11:47
elijahbal avar: that is a *good* practice IMO.11:48
that is called "modular programming.11:48
"11:48
→ t0rc joined11:48
codejunky that sound like dan bernstein11:48
shruggar I think going from foo.c to split-a.c should work (or at least sometimes work). It's not exactly a "git question" so much as a "merge algorithms question". That is: if git can't do it for you, maybe some other merge tool may be able to. Git can certainly track the history of the change, though, however you propagate that change11:49
← t0rc left11:49
avar elijahbal: Sounds like a crazy workaround by someone who hasn't coded much in C or has a bad text editor, just my 2 eurocents.11:50
← psankar left11:50
→ t0rc joined11:50
avar It's much easier to grok codebases that have lots of C functions per file grouped by functionality.11:50
shruggar unless the functionality bleeds over11:51
then it's just "shouldn't this be in x.c?"11:51
← roop left11:51
elijahbal You avoid this problem if you code a function in a file called with the name of the function11:52
→ ehabkost joined11:52
avar Are you going to have a corresponding header file for each function too?11:52
elijahbal also, when you open a file and you see 8-9 functions into it, your brain has to adapt to the situation. Which pieces of code are you going to modify ?11:53
shruggar I'm talking out my ass anyway. I use ctags, I don't care what file a function is in :)11:53
← rolfb left11:53
avar so instead of #include <str.h> you have #include <strcat.h>, #include <strlen.h> ....11:53
elijahbal Setting the header by functionality is good enough.11:53
I use a global header for my project currently.11:53
But I could split it up by functionnality, although I don't think it is of a real interest.11:54
I just make one include per file, it's enough.11:54
#include "my_project.h" , and "my_project.h" contains all the stuff I need (include foo,bar,etc)11:54
blueshift why would you make a rule to split all functions into separate files?11:55
→ parasti joined11:55
blueshift sounds like someone just discovered header files and linking and got overexcited11:55
elijahbal ha... It's just because it easier to make modifications.11:56
← SvenDowideit left11:56
→ JdGordon joined11:56
elijahbal You don't work on one project, you work on one function at time.11:56
shruggar I don't work on one function at a time..11:57
elijahbal You can trust me, it's good for the sake of mind.11:57
shruggar: that's a problem for me typically, I can not overload my brain.11:57
JdGordon I'm trying to psuh to a remote branch and having no luck.. anyideas why http://pastebin.com/CamrTYSZ isnt working?11:57
blueshift elijahbal: you can work on one function at a time when they're all in the same file. why does multiple files help?11:57
maybe you need a better editor with code folding / multiple views11:58
shruggar I work on one task at a time. This usually involves modifications to several functions. I tend to commit a modification to one function at a time, but I use "git add --patch" and never give a thought to what file those modifications are in11:58
← dajero left11:58
blueshift JdGordon: you rebased, or someone else pushed11:59
pull and merge11:59
→ thiago_a1ademy joined11:59
thiago_a1ademy → thiago12:00
JdGordon ah ok, thanks12:00
← thiago_akademy left12:00
thiago → Guest6607812:00
Guest66078 → thiago_akademyh12:00
thiago_akademyh → thiago_akademy12:00
→ dajero joined12:00
elijahbal shruggar: and basically, when you do this, it happens that a function depends on many things in the same file.12:01
You can avoid this if you set up a function per file.12:01
Thus, the day you need only *one* thing, you cut/paste your functions elsewhere.12:01
JdGordon umm.. ok how to I undo that pull?12:02
blueshift git help reset12:02
JdGordon hang on, if you cant rebase how are you supposed to work with big patches and leep it in sync with upstream?12:03
JdGordon is obviously doing it wrong12:03
blueshift you can rebase stuff you haven't published12:03
but if you rebase it to the tip of origin, then try to push.. someone else may have done so in the meantime12:04
in which case fetch, merge/rebase, push again12:04
← dajero left12:05
→ dajero joined12:05
JdGordon my public repo has master which is a clone of upstream and some branches which I want to keep seperate but in sync with master... is that wrong?12:05
→ oriba joined12:06
→ keyvan joined12:07
blueshift possible workflows is a big topic - and time for me to go. sorry12:07
JdGordon "in which case fetch, merge/rebase, push again" ?12:08
→ sandsmark joined12:08
sandsmark is it possible to fetch a single file from a remote git repository?12:09
_rane no12:09
← beeradb left12:09
Pieter you can if it's tagged first12:10
→ peper joined12:10
sandsmark hmm, ok12:10
elliottcable → Paws12:11
Paws → elliottcable12:11
← thierryp left12:11
← jony left12:12
→ thierryp joined12:12
← thierryp left12:13
← thiago_akademy left12:14
→ SvenDowideit joined12:15
→ beeradb joined12:20
→ eli joined12:21
→ sylr joined12:23
← eighty4 left12:24
→ froschi joined12:24
luist__ → luist12:25
← darkskiez left12:31
→ mw joined12:31
← froschi left12:31
← oriba left12:32
→ rolfb joined12:33
← rolfb left12:33
rajeshsr_ how is cherrypicking different from merge?12:33
isn't merge more useful? cherrypicking makes you lose current state into past commit, right?12:34
← gergap left12:36
← lucsky left12:36
wereHamster cherry-pick is this: apply the diff from the given commit and reues its commit message when commiting it12:36
→ xperience joined12:36
→ gergap joined12:37
← gergap left12:37
← JdGordon left12:38
→ giallu joined12:38
rajeshsr_ wereHamster, isn't merge also doing something like that?12:39
→ gergap joined12:39
_rane cherry-pick is for a single commit12:41
→ malumalu joined12:41
_rane merge is for many12:41
roughly speaking12:42
→ bentob0x joined12:43
sitaram rajeshsr_: merge is doing something like that but it is preserving history; the new commit is built on top of the current HEAD12:44
rajeshsr_: and the new commit has 2 parents, so any future merges know about the ancestry. With cherry-pick that's not going to happen12:44
→ lucsky joined12:44
→ ProperN[out] joined12:45
← ProperN[out] left12:46
→ ProperN[out] joined12:46
→ awx joined12:46
← girishr left12:46
sitaram rajeshsr_: some detail (all of it was already said above but maybe in a different format) in "cherry-pick versus merge" section of http://sitaramc.github.com/the-list-and-irc/special-branches.html12:48
← ProperNoun left12:48
ProperN[out] → ProperNoun12:48
The_Ball _rane, thanks for that (git blame)12:48
→ radioman-lt joined12:50
→ molokai joined12:51
← codeshepherd left12:51
rajeshsr_ sitaram, so cherrypicking is used when the 2 commits are siblings-like?12:51
← dajero left12:52
→ codeshepherd joined12:52
→ poseid joined12:52
sitaram rajeshsr_: which 2 commits?12:52
wereHamster it's used when you don't want to preserve ancestry12:53
→ ricardo joined12:53
rajeshsr_ the present HEAD and the commit you cherrypicked...12:53
_rane aaaaaaa12:53
sitaram rajeshsr_: sort of...12:53
← ricardo left12:53
sitaram rajeshsr_: I have to leave now; will rejoin when I get home from work :)12:53
rajeshsr_ sitaram, ok, thank you very much!12:54
→ girishr joined12:57
← girishr left12:59
→ Joeconyers joined12:59
→ ndim joined12:59
→ ricardo joined13:00
← ricardo left13:00
→ rccrv joined13:01
← mtkd left13:03
← codeshepherd left13:03
→ mtkd joined13:03
← RobertLaptop left13:04
→ codeshepherd joined13:04
→ IslandUsurper joined13:06
← tizbac left13:06
← ajpiano left13:07
→ coderdad joined13:09
→ gbacon joined13:10
→ simplechat joined13:10
← coderdad left13:10
→ thiago_akademy joined13:10
→ coderdad joined13:10
← flaguy48 left13:10
← coderdad left13:11
→ coderdad joined13:11
← gebi left13:12
→ tstclair joined13:12
→ coderdad_ joined13:13
← coderdad left13:13
bdiego_away → bdiego13:14
← gbacon left13:15
→ scarabx joined13:15
→ gbacon joined13:15
→ dajero joined13:17
→ Enselic joined13:20
← lacostej left13:21
→ girishr joined13:22
← girishr left13:22
← rccrv left13:25
→ MattDiPasquale joined13:28
→ jony joined13:29
→ LiamH joined13:30
← EricInBNE left13:30
← gbacon left13:31
→ gebi joined13:34
← thiago_akademy left13:34
← frakturfreak left13:35
→ frakturfreak joined13:35
→ j0shua joined13:35
← dbatbold left13:37
→ dbatbold joined13:37
← scarabx left13:39
← _Zenith_ left13:40
→ RobertLaptop joined13:41
→ beatak joined13:43
→ tcurdt joined13:43
→ binjured joined13:46
sitaram rajeshsr_: (I'm back) if by "siblings" you mean the commit being cherry-picked should be a valid successor of the current HEAD, yes, otherwise you may get horrible conflicts13:49
level09 I'm behind the origin/master by 20 commit, i want to overwrite the remote with my branch , how do i do that?13:49
sitaram but this is subjective13:49
← ph^_ left13:50
sitaram level09: you want kill those 20 commits forever on the server?13:50
level09 yes13:50
git reset --hard Head~5 gave me an error :S13:50
on the server13:50
sorry it is git reset --hard Head~2013:50
any other way to do it13:51
sitaram directly on the server? apart from the HEAD being uppercase, I'm not sure that will work anyway13:51
→ multigrain joined13:51
sitaram level09: the correct way is, on your client, "git push remotename +branchname"13:51
level09: that "+" makes it a force push13:51
level09 great thanks !13:51
in my case it is git push origin+master13:52
sitaram level09: you need a space between origin and +master13:52
← Enselic left13:52
level09 that explains the error i got :DD13:54
! [remote rejected] master -> master (non-fast-forward)13:54
what do I do about that :D13:54
→ TeckniX joined13:55
→ Ryback_ joined13:57
curtana level09, you do a force push: -f ... but only if you understand the consequences13:57
basically if anyone else has pulled since you pushed, they'll have to fix up their history just as you just did13:57
if no one has then use -f13:57
level09 ok thanks13:58
i'll try that13:58
error: failed to push some refs to ..13:58
mapreduce I am on branch origin/somebranch. git push origin somebranch says that everything is up to date. git log somebranch on the remote repo shows an old commit as HEAD.14:00
What am I missing?14:00
← RORUBUN left14:00
mapreduce It seems strange to me to be on branch origin/somebranch instead of on somebranch that tracks origin/somebranch.14:00
level09 hmm whay do i get this on the remote : fatal: unable to read tree14:00
mapreduce I think I got there by doing git checkout origin/somebranch14:01
→ asedeno_work joined14:02
cbreak_work mapreduce: could it be that you have a local branch "origin/somebranch"?14:03
do git branch -a to show all branches14:03
level09 i think git rm --cached -r . has messed up everything14:04
mapreduce cbreak_work: That appears to be the case.14:04
→ tizbac joined14:04
mapreduce * origin/somebranch (no -> thing)14:04
cbreak_work what I'd do is rename that one to something less stupid :)14:05
← [Po]lentino left14:05
← maleknet left14:05
← luist left14:05
hendry i pushed to a my/.git but the contents of my/ aren't being updated. what have i missed?14:05
→ montylounge_ joined14:06
mapreduce cbreak_work: Sounds good, thanks.14:06
← marc_kdab left14:06
shruggar hendry: sounds like faq nonbare14:06
hendry, in short: push does not have anything to do with work-trees14:06
cbreak_work yeah. because merging requires user interaction14:07
pushing by definition does not14:07
shruggar doesn't know how to operate the bot for faq entries14:07
cbreak_work I think you have to write faq non-bare14:08
Gitbot cbreak_work: Pushing to non-bare repositories is discouraged. See https://git.wiki.kernel.org/index.php/GitFaq#non-bare14:08
shruggar ah, hyphen, of course14:08
→ flazz joined14:08
level09 can I delete my master branch and rename another to master ?14:08
cbreak_work yes14:08
or you can reset --hard your master to the other branches' head14:08
level09 let me try that :)14:09
cbreak_work be warned14:09
reset can kill14:09
← rox` left14:09
cbreak_work if you lose a finger or an eye it's not my fault.14:10
shruggar I recommend that we start saying "reset --keep" when we mean "reset --hard", even though it's not what we mean, since it's safer14:11
cbreak_work nice option14:12
another_newbie when merging dev branch into master, how to ignore the dbsetting file14:12
shruggar still kicking around reset --crunchy at home14:12
cbreak_work but no, part of the fun of #git is handing out weapons of mass destruction to unsuspecting users14:12
→ jds joined14:12
cbreak_work another_newbie: in general, merging merges whole branches14:13
you can hack something together I guess if you really really want to14:13
→ maleknet joined14:13
→ jfkw joined14:14
another_newbie can we ignore one file14:14
→ xiong joined14:14
← rchavik left14:15
← masterkorp left14:15
← mtkd left14:15
another_newbie when we develop locally we use local db settings in db connection file14:15
but that shouldn't be added to master14:15
→ rox` joined14:15
another_newbie how can we handle that14:15
→ thunk joined14:15
level09 i checkout on a remote to an earlier branch14:16
but on local , it still says the remote is ahead by xx commits14:16
sorry earlier commit , not branch14:16
cbreak_work another_newbie: don't track the file if you don't want it tracke14:16
settings are not part of the source code, they should not be tracked anyway, since they get regenerated the next time you start your app anyway14:16
another_newbie what if its already added when initiaing the reposiotry14:17
cbreak_work you can remove it again14:17
another_newbie how14:17
cbreak_work git rm --cached14:17
it'll be deleted everywhere, so no more trouble with it :)14:17
→ thierryp joined14:17
→ aresnick joined14:17
→ luist joined14:17
cbreak_work (everywhere besides on your local machine, you get to keep your own copy with --cached)14:17
← khelll left14:17
another_newbie if its already in the initial commit14:18
← jcromartie left14:19
level09 cbreak_work: everytime I add something to .gitignore I can excute git rm --cached ?14:19
→ hagabaka joined14:19
→ dec joined14:19
→ alezandro joined14:20
→ lanthan_ joined14:20
← maleknet left14:21
another_newbie should add the config-setting files in .gitignore14:21
← ereslibre left14:22
cbreak_work level09: unrelated.14:23
another_newbie: yes14:23
you don't want to track configuration after all14:23
← lanthan left14:23
level09 cbreak_work: why? if I dont remove the cached files they will still be tracked14:23
no ?14:23
cbreak_work level09: you can also ignore files that aren't tracked in the first place14:24
that's what I do... :)14:24
→ disappearedng_ joined14:24
level09 cbreak_work: agreed, but sometimes you discover new files while working14:24
← bentob0x left14:24
← d0k left14:24
cbreak_work I am very careful with what I add14:25
← asedeno_work left14:26
← disappearedng left14:26
← mital left14:29
→ mann3qu1n joined14:29
← mann3qu1n left14:29
→ mann3qu1n joined14:29
← mann3qu1n left14:31
→ ereslibre joined14:32
← ereslibre left14:32
→ ereslibre joined14:32
→ smorg joined14:32
→ d0k joined14:33
← comawhite left14:34
montylounge_ → montylounge14:34
→ comawhite joined14:35
→ mann3qu1n joined14:35
→ tedoc2000 joined14:36
← mann3qu1n left14:36
← HG` left14:37
→ killasmurf86 joined14:37
killasmurf86 hi, i'm wondering how to split one project into two project, without keeping history of irrelevant files, for each project14:39
← another_newbie left14:40
shruggar killasmurf86, use git-filter-branch to remove all irrelevant history from one branch, then do it again to remove the opposite from another14:40
killasmurf86, it's likely that there is some overlap14:40
level09 git checkout -f is taking me out of the master branch, once i checkout to it , the checkout is undone14:40
killasmurf86 shruggar: thanks, so I was looking at right manual :D14:40
→ OneTrueDave joined14:44
← Joeconyers left14:45
→ AAABeef joined14:45
OneTrueDave Given an old revision, how do you get a diff between that and the previous revision, i.e. see exactly what that commit changed?14:45
→ Joeconyers joined14:45
shruggar OneTrueDave, "git show <the-commit>" is probably what you want14:46
← dajero left14:46
OneTrueDave yep, thanks!14:46
shruggar OneTrueDave, more-literally, however: git diff <the-commit>^..<the-commit>14:46
→ altrux joined14:47
OneTrueDave what does the caret signify14:47
→ Yuffster joined14:47
shruggar OneTrueDave, first-parent-of. see man git-rev-parse14:47
Gitbot OneTrueDave: the git-rev-parse manpage can be found at http://git.or.cz/man/git-rev-parse14:47
Arjen A shortcut for that is $commit^!14:47
OneTrueDave okay, so show the diff between the first parent and the actual commit. makes sense.14:48
thanks14:49
← OneTrueDave left14:49
← warthog9 left14:49
← apsysoev left14:50
→ nanotube_ joined14:50
← killasmurf86 left14:50
→ ereslibre_laptop joined14:50
← ereslibre_laptop left14:50
→ ereslibre_laptop joined14:50
← coderdad_ left14:50
← nanotube_ left14:51
→ coderdad joined14:51
← ereslibre left14:51
level09 why am I getting this error when trying hard reset fatal: unable to read tree14:52
→ marc_kdab joined14:52
→ hipe__ joined14:53
→ ereslibre joined14:55
→ chrislerum joined14:55
← ereslibre_laptop left14:56
← coderdad left14:56
→ RobertLaptop_ joined14:57
→ ajpiano joined14:57
→ nsussman` joined14:58
← christophsturm left14:59
← RobertLaptop left14:59
→ alus joined14:59
fr0sty level09: what does 'git status' tell you?15:00
alus I have an svn repo which is receiving upstream changes, and I'd like to commit revisions from it to my git repo fork15:00
→ h32Lg joined15:00
← h32Lg left15:00
→ h32Lg joined15:00
alus is there a function for that, or do I have to monkey patch by hand?15:00
← radioman-lt left15:01
→ radioman-lt joined15:01
→ mbroeker joined15:02
→ rayners joined15:03
level09 fr0sty: nothing to commit - clean directory15:03
fr0sty where are you trying to reset ?15:04
'to what', really15:04
curtana alus, are you using git-svn?15:04
→ Nemurenai joined15:04
curtana wishes there was a git-svn --mirror that made a bare git repo15:05
← smorg left15:05
→ nevyn^ joined15:05
shruggar curtana, there's not because it's can't be made very sanely, due to mostly-necessary commit-rewriting15:06
→ warthog9 joined15:06
curtana double negative whoa head asplode15:07
← iravid left15:07
shruggar no double-negative there, though a misplaced "'s"15:07
→ iravid joined15:08
← hipe__ left15:08
level09 fr0sty: i did a push from local, I want to reset everything to a point before that merge15:08
alus curtana: I don't know what git-svn is, so probably not15:08
shruggar alus: if you are tracking an svn repository with git, and don't know what git-svn is, then you should probably look into man git-svn15:09
Gitbot alus: the git-svn manpage can be found at http://git.or.cz/man/git-svn15:09
alus "tracking" is maybe to specific a definition15:09
s/to/too15:10
fr0sty level09: you pushed from local, and are doing 'git reset <sha1>' on the local machine?15:10
if, so, what does 'git show <sha1>' tell you?15:10
← navetz left15:10
alus sometimes there are changes in the svn repo which I want. sometimes they want changes I make, and obviously I skip importing those15:10
level09 fr0sty: no , I sshed to the remote to try to reset it15:10
I was able to reset local with no issues15:11
→ Sr_X joined15:11
level09 tried a forceful push , but it was rejected15:11
fr0sty try adding a plus to the push command.15:11
alus git-svn looks too complicated15:12
fr0sty 'git push <remote> +<localbranch:remotebranch> ?15:12
ToxicFrog alus: what's complicated about it?15:12
fr0sty also, is the remote a bare repository?15:12
→ RORgasm joined15:12
alus ToxicFrog: uh, it does tons of things and I only want a tiny thing15:13
ToxicFrog git svn clone foo; git svn rebase to get changes from svn, git svn dcommit to push changes to svn.15:13
← cbreak_work left15:13
GarethAdams|Work alus: my watch is water resistant to 50m, but I got it even though I didn't need to15:14
alus GarethAdams|Work: but you don't have to dive to 50m to read the time15:14
← starcoder|Svr left15:15
GarethAdams|Work alus: right, but you don't need to *use* all of the other features just because they're there15:15
s/but //15:15
alus GarethAdams|Work: well, the feature I want is burried in tons of other functionality15:15
← Nickeeh left15:15
GarethAdams|Work alus: ToxicFrog just described what you need to know15:15
level09 fr0sty: yes it is bare remote15:16
alus not really, no. how do I pull specific changesets, not all of them?15:16
er, svn calls them "revisions"15:16
→ smorg joined15:16
→ slonopotamus joined15:16
ToxicFrog alus: from what you said, the only ones you don't want to pull are the ones that are upstream incorporations of your own changes15:16
and git svn rebase skips those automatically15:17
alus hm15:17
level09 still rejected with +sign15:17
git push origin +master15:17
ToxicFrog If you want something more like cherry-pick, I have no idea, sorry15:17
alus well I'll try that some day.15:17
ToxicFrog The idea with git svn is that you should be able to commit from git to svn as well as pull from svn to git, so stuff like that may not be supported at all.15:17
← leavittx_ left15:18
alus doees it actually affect the giv repo other than the changes it pulls in from svn?15:18
← leavittx left15:18
alus I mean, it has to store the svn repo url somewhere. does that get committed to the git repo?15:19
fr0sty level09: you can't reset a bare repository15:19
→ Textmode joined15:19
sitaram fr0sty: well you can, with plumbing commands, but it isn't pretty :)15:20
level09 what do I do then ?15:20
← frakturfreak left15:22
fr0sty level09: git push -f origin +master15:22
→ frakturfreak joined15:22
→ hipe__ joined15:23
level09 ! [remote rejected] master -> master (non-fast-forward)15:23
error: failed to push some refs to ..15:23
→ dreiss joined15:23
level09 :(15:23
→ spaceonline joined15:25
← vu3rdd left15:25
→ girishr joined15:27
→ starcoder|Svr joined15:27
→ raichoo joined15:27
← drubies left15:29
← freimatz left15:30
← priidu left15:30
→ drubies joined15:30
← codeshepherd left15:31
fr0sty if you have access to the remote machine, try unsetting receive.denyNonFastForwards temporarily15:32
← slonopotamus left15:32
level09 then push again ?:15:33
← aresnick left15:34
level09 where is that setting ? in .git/config ?15:34
← keyvan left15:34
← Joeconyers left15:35
→ keyvan joined15:35
fr0sty yes, push again, in reponame.git/config in the remote repostitory.15:35
level09 thanks i'm gonna try that15:35
← intripoon left15:35
→ WebDragon joined15:36
→ kenichi joined15:36
Unode → _Unode15:36
dennda win215:36
level09 that worked !! :DD15:36
← dreiss left15:36
level09 fr0sty: thanks, is there any downsides for what I did ?15:37
← t0rc left15:38
← curtana left15:38
fr0sty anyone else who has previously pulled from the remote will have to jump through some hoops to recover this.15:39
and leaving denyNonFastForward = false will allow anyone else to do the same thing in the future.15:39
level09: man git-rebase and looke at 'RECOVERING FROM UPSTREAM REBASE' for an explaination of recovery steps.15:39
Gitbot level09: the git-rebase manpage can be found at http://git.or.cz/man/git-rebase15:39
→ curtana joined15:40
level09 thanks ! really helpful15:40
← kenyon left15:40
→ ereslibre_laptop joined15:40
→ ponzao____ joined15:40
level09 fr0sty: when adding new files to .gitignore , is it right to run git rm --cached . to remove all tracked files from cache ?15:40
← ereslibre left15:40
curtana level09, check out git-clean15:42
← warthog9 left15:42
level09 curtana: git clean will physically delete the files15:42
I want to keep these files , while untracked15:42
← psoo left15:42
← ToxicFrog left15:42
← sh1mmer left15:43
curtana oh15:43
→ hohoho joined15:43
→ ToxicFrog joined15:45
_Unode → Unode15:45
→ kenyon joined15:46
← keyvan left15:46
← spaceonline left15:47
→ steveire_ joined15:49
← steveire_ left15:49
→ steveire_ joined15:49
← jony left15:50
← marc_kdab left15:51
→ marc_kdab joined15:51
→ Joeconyers joined15:52
→ keyvan joined15:55
← andrewl left15:55
→ Nickeeh joined15:56
← Nemurenai left15:56
→ sh1mmer joined15:57
← steffkes left15:57
→ kpreid joined15:59
→ rubydiamond joined15:59
level09 anyway , is it safe to run git rm --cached .15:59
→ dsdeiz joined16:00
← dsdeiz left16:00
→ coderdad joined16:00
→ hobodave joined16:01
shruggar "safe" is a loaded term. There are situations where it can lose data, but you are most likely not in that situation16:01
← barismetin left16:01
shruggar for example, if you git-add'ed a file, didn't commit, then removed the physical copy of it, git rm --cached would remove the only remaining trace of the file. Other than that, you're safe16:02
→ stringo0 joined16:02
shruggar though even then, git-fsck could find it, just not in a friendly manner16:02
→ bewst1 joined16:02
shruggar most git commands are safe16:03
← coderdad left16:03
← xiong left16:03
→ coderdad joined16:03
bewst1 Is there a tool out there that will represent file moves in a git repo as moves? I'm interested in reconstructing the actual logical changes made in a commit.16:03
← hukl left16:04
Ilari_antrcomp bewst1: diff with request to track moves&copies?16:04
bewst1 yep16:04
do you mean "git diff" already does this?16:04
← level09 left16:05
Ilari_antrcomp bewst1: -C -C -M ...16:05
bewst1 two -C's?16:05
Ilari_antrcomp bewst1: Yes...16:06
← fmardini left16:06
bewst1 thank you!16:07
→ patmaddox joined16:08
← coderdad left16:08
→ kenneth_reitz joined16:09
→ azol joined16:09
← chrislerum left16:09
← nevyn^ left16:09
hobodave hey guys16:14
how can I "inspect" a blob given it's hash?16:15
→ Khisanth joined16:15
→ valberg joined16:15
valberg hi16:15
hobodave hi valberg16:15
→ warthog9 joined16:15
← keyvan left16:16
← GarethAdams|Work left16:16
← Pupeno left16:16
valberg is it possible to point git.domain.com to a folder containing git repos, so it's possible to do like this: git clone ssh://user@git.domain.com/project ?16:16
→ keyvan joined16:16
← trivol left16:16
hobodave yes16:17
that's how most git repos are setup16:17
I suggest gitosis16:17
cehteh mhm16:17
valberg k16:18
Nugget_ I don't think that's accurate.16:18
if you're using ssh as the transport (with unique logins for users) then you can't remap paths16:19
→ ProperN[out] joined16:19
← ProperN[out] left16:19
→ ProperN[out] joined16:19
Nugget_ gitosis allows that because you can remap the paths on account of it using a single user account on the repo for all user access16:19
but with just native git, it's just using the user's individual environment and paths are relative/absolute on that basis16:20
valberg Nugget_: okay - thnx16:20
← binjured left16:20
hobodave Nugget_: it's still just a directory16:20
Nugget_ git-daemon supports remapping, but I don't think ssh:// git paths invoke it in that manner16:20
hobodave: define "it's"16:21
valberg my thing is, that i have my git/ folder in /var/git/16:21
hobodave ssh://user@git.domain.com/project doesn't need remapping, it references HOME/project for that user16:21
Nugget_ hobodave: right, that's precisely the problem16:21
hobodave how so?16:21
Nugget_ are you suggesting creating symlinks for all users in their home directories for each git repo that's on the host?16:21
hobodave no, why would I do that?16:21
Nugget_ so multiple users can share a common repo16:22
← ProperNoun left16:22
ProperN[out] → ProperNoun16:22
Nugget_ which is, I presume, what's being discussed16:22
hobodave that was never mentioned16:22
valberg and i want git.domain.com to point to /var/git/ so i can have all my projects in /var/git/ and clone them with git clone git.domain.com/projectname16:22
hobodave I didn't make any assumptions based on his question16:22
Nugget_ then why did you suggest gitosis?16:22
hobodave for ease of maintenance16:22
→ coderdad joined16:22
Nugget_ what maintenance do you envision for a single user?16:23
→ chrislerum joined16:23
hobodave I'm not arguing with you about this. I answered his question as asked.16:23
valberg just to make it clear, i want many users to be able to access many repos easily16:24
hobodave then I suggest using gitosis valberg16:24
IslandUsurper → IslandUsurperAFK16:24
valberg thnx16:24
→ rado1 joined16:24
→ flazz_ joined16:25
← flazz left16:27
Nugget_ I don't use either, but my understanding is that gitosis is quasi-abandonware and that gitolite is the preferrred solution these days16:27
hobodave I wasn't aware of that16:27
gitosis "just works" for me16:27
← multigrain left16:27
← mbroeker left16:27
Nugget_ it's been discussed in here, but that's the extent of my understanding16:27
sitaram hobodave: gitosis works fine, and if you're already using it and the feature set is enough there's no need to change16:28
hobodave cool sitaram, that's what I assumed16:28
← Khisanth left16:29
hobodave considering gitosis is 99% SSH, it'd be hard for it to break :)16:29
sitaram valberg: if you haven't alread installed gitosis, I suggest you look at gitolite. Gitosis hasn't been updated in about 2 years... (which as I said, doesn't matter if you're already using it and are happy)16:29
→ multigrain joined16:29
sitaram hobodave: true. However, I wrote gitolite mainly because I wanted per-branch controls16:29
valberg sitaram: okay cool - then i'll check on gitolite16:30
:)16:30
hobodave yea I just saw your name in the repo16:30
interesting16:30
sitaram oh yeah; my usual disclaimer: I'm the author so I'm biased ;-)16:30
hobodave well at least you didn't blindly bash gitosis :)16:30
sitaram hobodave: nothing to bash; it works as advertised. Just that it was not enough for me16:31
hobodave: I used to support gitosis on this channel for months before I wrote gitolite. I still have a gitosis-troubleshooting page somewhere16:31
http://sitaramc.github.com/the-list-and-irc/gitosis-troubleshooting.html16:32
hobodave bookmarked, thanks :)16:32
If I ever need to setup a new git server I'll probably end up looking at gitolite16:32
is the setup/configuration significantly different from gitosis?16:32
did you base it off gitosis?16:33
← shruggar left16:33
→ rorist joined16:34
Tass_ how to commit something to a new branch?16:35
→ slonopotamus joined16:35
← kenneth_reitz left16:35
hobodave Tass_: git co -b new_branch_name , then git commit like usual16:35
sorry git checkout16:36
co is an alias I use16:36
Ilari_antrcomp hobodave: Syntax is different. But the basic principles are mostly the same.16:36
→ masak joined16:36
← simplechat left16:37
← whaley left16:37
masak when I do `git tag -a -m"some commit message" $TAGNAME`, where does the commit message end up, and how can I query for it later?16:37
← rorist left16:37
← gergap left16:38
Ilari_antrcomp masak: It winds up as tag message.16:38
slonopotamus masak: it isn't a commit message because no commit is created. tag object is, however16:38
← rtl left16:38
← olivier__ left16:38
hobodave masak: git show $TAGNAME16:39
masak ah, so somewhere there's a tag object with that message.16:39
sitaram hobodave: the config language is *quite* different -- it has to be, I have a lot more features16:39
masak hobodave: thanks.16:39
valberg sitaram: just to clarify, will i be able to poing git.domain.com to /var/git/ directory?16:39
← tvw left16:39
sitaram valberg: (assuming "point") do you mean you want a URL like [email@hidden.address] to go to /var/git/reponame.git?16:40
valberg sitaram: yeah16:40
→ rorist joined16:41
sitaram valberg: sure. By default it points to ~/repositories (for whatever user you installed it to), but you can -- at a certain point in the install -- edit a file called ~/.gitolite.rc and point a variable called "$REPO_BASE" wherever you want16:41
valberg sitaram: awesome!16:41
sitaram: then i'll definately check it out :D thnx for the heads up!16:42
→ dreiss joined16:42
→ binjured joined16:42
sitaram valberg: you can even change it later, as long as you change it *and* move the repositories from ~/repositories to /var/git in one shot, with no one accessing it in between16:42
← thedeeno left16:42
sitaram valberg: meaning as long as they are consistent it'll all work fine16:43
valberg cool cool16:43
← eno left16:43
sitaram they ==> the value of $REPO_BASE in ~/.gitolite.rc and the actual location of the repos16:43
→ WebDragon|laptop joined16:43
← sh1mmer left16:44
→ eno joined16:44
← slonopotamus left16:45
← ezyang left16:46
← WebDragon left16:47
→ jony joined16:48
← masak left16:49
← bewst1 left16:51
→ plediii joined16:51
← aziz left16:52
→ aziz joined16:52
→ pantsman- joined16:52
→ carutsu joined16:53
← KB1JWQ left16:53
→ shruggar joined16:53
→ EvanR-work joined16:54
EvanR-work were having a weird problem, the three of us have synchronized repos, but one of us has an anomaly in his repo16:54
← pantsman- left16:54
EvanR-work git fsck reveals, for him, a missing blob with hash equal to a submodule16:54
→ pantsman- joined16:55
← tstclair left16:55
← luist left16:55
EvanR-work we do not have this reference in our working repo16:55
when he commits, we usually cant fetch from him16:55
← pantsman- left16:55
→ pantsman joined16:56
EvanR-work git gc fails for him in any case16:56
i dont see how this could have occured in his repo and not ours16:57
→ luist joined16:58
← steveire_ left16:59
← stringo0 left17:00
← dpino left17:01
→ sh1mmer joined17:01
→ RBF joined17:01
→ aresnick joined17:01
← galderz left17:02
→ stringoO joined17:02
← girishr left17:03
← RBF left17:03
← multigrain left17:04
→ RBF joined17:04
fr0sty EvanR-work: define "syncrhonized17:04
← RBF left17:04
fr0sty also, is this submodule an abberation in your set of repos ?17:05
→ RBF joined17:05
→ gmcinnes joined17:05
fr0sty i.e. do you all have the offencint submodule?17:05
← RBF left17:06
→ RBF joined17:06
← RBF left17:06
gmcinnes oh noes! I deleted a remote branch using "push origin :foo" How can I resurrect it?17:06
→ ph^_ joined17:07
← chrislerum left17:07
→ thedeeno joined17:07
fr0sty just re-push it from your repository (if you have it).17:08
→ eighty4 joined17:09
← malumalu left17:09
fr0sty gmcinnes: what were you trying to do?17:09
gmcinnes fr0sty: I'm cleaning up some legacy stuff from the repo, and I deleted a branch I want to keep.17:10
→ malumalu joined17:10
← vegeste left17:10
valberg sitaram: is there a .deb package of gitolite for ubuntu 9.10 ?17:11
EvanR-work fr0sty: yes we do17:11
fr0sty gmcinnes: do you still have the branch/commit locally?17:11
gmcinnes fr0sty: Unfotunately, its legacy enough I may not have the branch locally, and no-one else might have either.17:11
sitaram valberg: ask Rhonda, but there is one for debian something and since the requirements are pretty simple it ought to work on any similar distro17:11
Rhonda: ping and ^^17:12
gmcinnes fr0sty: I mean I must have the objects, I guess, but I don't know the commit it pointed to.17:12
fr0sty gmcinnes: does anyone else have a clone of the remote repository? They will still have branch 'foo'17:12
→ enherit joined17:12
gmcinnes fr0sty: My repo is freshly cloned. I should have the objects locally, I guess.17:12
fr0sty: but how do I find out what it used to point to?17:13
fr0sty EvanR-work: is it possible he committed locally to the submodule and forgot to push it?17:13
refs/remotes/<name>/foo should still exist locally.17:14
if so, 'git push origin foo' will put it back.17:14
Rhonda peeks at valberg and sitaram17:14
valberg :)17:15
Rhonda valberg: http://packages.debian.org/unstable/gitolite17:15
valberg cool17:15
Rhonda And 9.10 is old, why don't you use 10.04 which is LTS again anyway? ;)17:15
… and neither has gitolite directly, sorry to had been too late for that.17:15
→ rtl joined17:15
doener fr0sty: the remote tracking branch should have been delete when he did the push, and "foo" is only a shortname for refs/heads/foo not for refs/remotes/<name>/foo17:16
Rhonda I am thinking about doing lucid-backports for gitolite at least, but karmic …17:16
← shruggar left17:16
gmcinnes doener: yes. It's true. The tracking branch is gone.17:16
valberg Rhonda: it's a longer story - but the short version is: it wasn't possible to install 10.04 as a virtual server in xen when i set it up back in the days17:17
doener gmcinnes: did you have that branch checked out locally?17:17
→ hagebake joined17:17
gmcinnes doener: unfortunately not. And its old enough I doubt anyone else did.17:17
→ ProperN[out] joined17:18
→ doki_pen joined17:18
← hagabaka left17:18
← doki_pen left17:18
← ProperNoun left17:18
→ doki_pen joined17:19
← ProperN[out] left17:19
→ ProperN[out] joined17:19
← jony left17:19
doener gmcinnes: gitk $(git fsck --unreachable --full --no-reflogs | grep 'unreachable commit' | cut -d' ' -f3)17:19
fr0sty doener: anyone with a clone of the old repo can still 'git checkout -b saveme origin/foo'17:19
ProperN[out] → ProperNoun17:19
→ frakturfreak_ joined17:20
fr0sty assuming they haven't run 'git fetch' since the recent unpleasantness.17:20
→ Khisanth joined17:20
doener fetch is fine, fetch --prune is not17:20
fr0sty even better...17:21
→ gbacon joined17:21
Rhonda valberg: But that shouldn't hinder you to upgrade, shoudl it? :)17:21
valberg: Anyway, have fun and bug me if something bugs you in the package.17:21
→ trochala joined17:22
← trochala left17:23
→ trochala joined17:23
doener fr0sty: and of course, if someone has an up-to-date remote tracking branch for "foo", that someone can do "git push origin refs/remotes/origin/foo:refs/heads/foo" to restore the branch head17:23
← frakturfreak left17:24
The_Ball when viewing individual commits in gitk the patches are very small, however when I do "git diff sha1 sha2" then the diff is huge. I suppose it's comparing to the current tree, how can I see the changed introduced by a single commit?17:24
doener I just said that "git push origin foo" won't do if there's only a remote tracking branch refs/remotes/origin/foo, but no refs/heads/foo17:24
→ pavlix joined17:25
fr0sty The_Ball: 'git show <sha1>'17:26
The_Ball fr0sty, doh! thanks17:26
← tatsuya left17:26
fr0sty you can also: 'git log -p <sha1>..<sha2>' for extra fun (show every commit in between 1 and 2)17:27
→ khmarbaise joined17:28
→ apsysoev joined17:30
← bluenovember left17:31
→ peritus- joined17:33
peritus- hey. can someone tell me what's undeterministic about the following script: http://dpaste.de/GCrV/ ?17:35
gmcinnes doener: I found someone else who could check out the branch locally. I got them to give me the sha for the last commit on that branch, but when I tried to create a branch based on that locally, there's nothing in it. What am I doing wrong?17:35
→ ProperN[out] joined17:36
peritus- .. that is, i want to create the same repository twice from scratch (w/ all the same sha1's, etc.)17:36
← ProperN[out] left17:36
→ ProperN[out] joined17:36
← ProperNoun left17:36
← curtana left17:37
ProperN[out] → ProperNoun17:37
IslandUsurperAFK → IslandUsurper17:38
→ brett_h joined17:38
→ dizzo joined17:38
→ gnufied joined17:39
← trochala left17:39
Ilari_antrcomp peritus-: Didn't read it, but I guess committer/author timestamps?17:40
→ jmcantrell joined17:40
→ trochala joined17:40
← trochala left17:40
→ trochala joined17:40
peritus- Ilari_antrcomp: I do git commit -date=2005-04-07T22:13:13 .. that should do ?17:41
← coppro left17:41
→ tvw joined17:41
← rubydiamond left17:41
← plediii left17:41
peritus- Ilari_antrcomp: make that two dashes17:41
rkreis did you compare files and maybe some kind of dump?17:42
→ coppro joined17:42
valberg sitaram: does the git user need to have a home directory?17:42
rkreis like git whatchanged -p or so?17:42
Ilari_antrcomp valberg: Yes.17:42
sitaram valberg: yes; I suppose so... try without and see; I haven't tested that though17:42
valberg i'll try :)17:42
← kpreid left17:43
→ kpreid_ joined17:43
← harv left17:44
valberg sitaram: i was able to run the gl-setup command - but there is no autocompletion , so it's quite annoying :)17:44
sitaram no auto-completion in your $SHELL?17:45
← kokx left17:47
→ kokx joined17:47
fr0sty gmcinnes: have them do the 'git push...' for you if they have the commit.17:48
valberg sitaram: well, i didn't log into the git user with the -s parameter17:48
but i just realized that i don't have to have my repo i /var/git/ , so it's just fine to have the default17:49
→ k3yvn joined17:49
sitaram valberg: and as I said you can always move it later, as long as you fix that variable at the same time17:50
← keyvan left17:50
valberg sitaram: true true17:51
→ rubydiamond joined17:53
valberg sitaram: btw, i've changed the port of my sshd, how do i clone from my repo then ?17:54
← jds left17:54
→ mabrand joined17:54
→ kar8nga joined17:55
sitaram valberg: did you install using a .DEB?17:55
valberg sitaram: yep17:55
← d0k left17:55
sitaram valberg: ok, man git-clone has a bunch of syntaxes in the "GIT URLS" section; you'd probably use ssh://git@host:port/reponame17:56
Gitbot valberg: the git-clone manpage can be found at http://git.or.cz/man/git-clone17:56
→ HG` joined17:57
valberg sitaram: yeah that's totally right17:57
→ ChUrP joined17:57
← hohoho left17:58
→ Jaxan joined17:59
valberg sitaram: why is it that i have to clone the gitolite-admin to my machine ?17:59
→ zehrique joined17:59
← h32Lg left18:00
→ sixteneighty joined18:00
→ plediii joined18:00
← alus left18:01
← plediii left18:01
sitaram valberg: because all administration is done by making changes to the clone and pushing those changes :)18:01
valberg awesome18:01
→ tstclair joined18:01
sitaram valberg: might sound weird but it's very cool because (1) you get auditability for all config changes (2) you can do it from your client without having to log on to the server18:01
valberg: this is also what gitosis did, so credit goes to Tv for that idea18:02
valberg sitaram: so if i want to allow another user on the server to configure it, i'd just edit gitolite-admin/conf/gitolite.conf and the push ?18:02
s/push/commit and push18:02
Ilari_antrcomp valberg: Yes.18:03
sitaram s/commit/add, commit/ :)18:03
valberg heh heh fair enough :) got me there18:03
sitaram you can even allow someone else to share the admin workload by adding them into the RW+ list for the special "gitolite-admin" repo; no need to give him shell access to the server this way18:04
← jmcantrell left18:04
→ jmcantrell joined18:04
→ thiago_akademy joined18:05
← RORgasm left18:05
→ RORgasm joined18:06
→ girasquid joined18:06
Ilari_antrcomp valberg: And gitolite even tries to prevent users with full push access to gitolite-admin from gaining shell access (assuming they don't have it already).18:06
girasquid I have a git repo that's about 90MB, with most of that being in the .git/objects/pack directory. How can I trim that down?18:07
← RORgasm left18:08
Ilari_antrcomp girasquid: 'git gc' (but that might not do much)?18:08
← beeradb left18:09
→ RORgasm joined18:09
girasquid Ilari_antrcomp: holy crap, that bumped the size down to like 25MB. Thanks!18:09
→ WebDragon|away joined18:10
WebDragon|away → WebDragon18:11
→ bremner joined18:12
← poseid left18:13
→ programble joined18:14
← WebDragon|laptop left18:14
flazz_ → flazz18:15
elliottcable → coolSTER18:15
← tcurdt left18:17
Circuitsoft Hello. git-svn fetch is giving me a checksum mismatch message.18:18
← jailbox left18:19
Circuitsoft Do I have a broken object I need to remove?18:19
$ nice git svn fetch18:19
Checksum mismatch: some/file/from/svn f83cfec96f06b3f3a8c10617fd449d009efbe1e818:19
expected: cee883ec7e1ac0a5982fe7e5bba46b0118:19
got: 3577ded5f7e20da58c97f41d1ac5cc8118:19
← girasquid left18:19
→ zenazn joined18:19
← zehrique left18:20
rkreis Circuitsoft, what about git fsck?18:21
coolSTER → elliottcable18:22
Circuitsoft Lists a bunch of dangling blobs. Nothing else.18:22
→ poseid joined18:23
→ kipras joined18:23
Circuitsoft 101 to be exact.18:23
rkreis "It tests SHA1 and general object sanity" weird18:23
← mw left18:23
Circuitsoft Trying git fsck --full18:25
→ Titosemi_ joined18:27
← tvw left18:28
← jmcantrell left18:28
Circuitsoft Now /that/ is taking a while.18:28
→ andrewl joined18:29
Circuitsoft Hmm... A line from top:18:29
1737 aaustin 35 10 3054m 2.9g 2.9g R 100.4 38.4 2:58.57 git18:29
← Titosemi left18:29
Titosemi_ → Titosemi18:29
Circuitsoft 2.9g resident ram18:29
← azol left18:30
peritus- can someone tell me what's undeterministic about the following script: http://dpaste.de/GCrV/ ?18:31
I want to create the same repository twice from scratch (w/ all the same sha1's, etc.)18:32
← ereslibre_laptop left18:34
FauxFaux The commit or author time.18:35
→ ods15 joined18:35
→ ereslibre joined18:37
ods15 Hi. I am still in a phase of setting up my git repository correctly, and i want to move some work i've done on one git repo to a completely different git repo. how can i do this most sanely? pretty much what i had in mind was format-patch for the ~15 commits and then git-apply. nothing compilcated in these commits, only new files18:37
the "completely different" repo has different history and log, but the HEAD looks the same code-wise18:37
→ jailbox joined18:38
ods15 (the first git repo was a result of a failed experiment)18:38
drizzd ods15: git-format-patch applies with git-am, not git-apply18:38
→ masterkorp joined18:38
Circuitsoft My fsck --full has finished, and now I only got 55 dangling blobs.18:38
→ shruggar joined18:38
ods15 drizzd, ok, what is the local operation i should? git format-patch, take all the files, put in a dir, then git am on that dir?18:39
← trochala left18:39
→ ive joined18:39
FauxFaux Circuitsoft: My wild guess would be that your repo is fine but git svn is unhappy as you rebased or merged or ... it's tag.18:39
drizzd ods15: it would be even easier to add your repo as a remote and use rebase --onto18:39
ods15: let's say you have two repos A and B with completely different histories18:40
→ ProperN[out] joined18:40
← ProperN[out] left18:40
→ ProperN[out] joined18:40
drizzd but they are the same at commits A-master and B-master, respectively18:40
← ProperNoun left18:40
Circuitsoft FauxFaux: The odd thing is, that It's been svn-fetching for 35000 or so revs, and it just stopped at this one.18:40
← mastro left18:41
drizzd then in repo B you have some changes B-topic, which you want to apply onto A-master18:41
does that correctly represent your situation?18:41
← alezandro left18:41
ProperN[out] → ProperNoun18:41
← sandsmark left18:41
ods15 drizzd, yes, exact18:41
actually i have some branches in B, and i want to do this in several steps, but each step looks like this18:42
→ Fullmoon joined18:42
drizzd ods15: so all you have to do is cd to repo A, then git remote add -f B path/to/b; git checkout -t B/topic; git rebase --onto master B/master18:42
that will take the range of commits B/master..B/topic and apply them on top of master18:43
ods15 ok, i'll try that.. be back in a bit18:43
drizzd make sure master and B/master are in fact the same or at least similar, otherwise you will get lots of conflicts18:43
ods15 i was about to same, they are similar, not identical18:44
i know there are no conflicts18:44
(umm, i hope :)18:44
← ive left18:44
drizzd well, if you do get conflicts, rebase will let you resolve them, but rebasing generally involves more work than merging, if there are conflicts18:45
ods15 ok18:45
← poseid left18:46
→ Chillance joined18:46
ods15 can you explain -t? i don't understand it...18:47
fr0sty peritus-: --date only changes the author date, not the commiter date. look into GIT_COMMITER_DATE and GIT_AUTHOR_DATE18:48
← marc_kdab left18:49
valberg sitaram: thnx a lot for the help :) it works as a charm! :D18:49
→ marc_kdab joined18:49
→ Jochen_Hayek joined18:49
→ mtkd joined18:49
→ heptadecagram joined18:50
← j0shua left18:54
→ bcardarella joined18:54
→ justin-george joined18:56
→ oriba joined18:57
← oriba left18:57
→ Nemurenai joined18:58
→ rindolf joined18:59
rindolf Hi all. I have a "remotes/origin/convert-user-auth-to-catalyst" and I'd like to map it to a local "convert-user-auth-to-catalyst" branch. How can I do it?18:59
← ereslibre left19:00
→ bx2 joined19:00
rindolf I still don't have the local branch.19:00
← HG` left19:00
← bx2 left19:02
← project2501b left19:02
→ chrislerum joined19:02
→ p3rg2z joined19:02
drizzd ods15: oh, that's short for git checkout -b topic B/topic19:04
→ urkud__ joined19:05
hagebake → hagabaka19:05
← rox` left19:06
← ntoll left19:06
ods15 drizzd, actually, it didn't work forme for some reason, i had to write what you just said explicity19:07
git version maybe? 1.6.0.619:07
→ rox` joined19:07
ods15 anyway, the rebase worked, and i learned a little bit more about rebase! (i thought i already had it figured out yesterday :) so, great, thanks!19:07
drizzd ods15: possibly, but the feature is not new19:08
→ macek joined19:08
← iatrou left19:09
→ centr0 joined19:09
ods15 ok, now that you showed me this "remote", i think i realize i did something wrong here... i had a "pristine" git repo, and i decided to leave that to the side, and my intention later on is that it will become a central repo.. for what i did just now, i used "cp -r" to get my "own" repo to work on, where i did this merge19:10
i should have created a new repo, and used git remote? or git pull? (never used "git pull" before)19:10
← hipe__ left19:12
→ hipe joined19:12
← comawhite left19:13
→ poseid joined19:15
← hipe left19:15
LiamH Is there an accepted way of using git as a client for darcs repositories, analogous to git svn? Everything I can find in Google looks at least a couple years old.19:15
→ _numbers joined19:16
_numbers i am so frustrated with git19:16
← ChUrP left19:17
→ project2501b joined19:17
_numbers two major problems: 1. large git repos become slow and cumbersome to share, 2. history cannot be manipulated to split repo into volumes or cleanup old commits which really shouldn't be in there.19:18
very difficult to work with in those two respects19:18
→ jf_ joined19:18
→ comawhite joined19:18
jf_ hi, i've been trying for sometime now to mergeback changes to a svn repository and i would like to keep the right author within svn when i'll do the "git svn dcommit" .. is there a way to tell git that?19:19
peritus- ok, i updated http://dpaste.de/uz0m/ .. still doesn't produce the same repo deterministically19:19
_numbers: can be adressed with git-split and/or http://progit.org/2010/03/17/replace.html19:20
→ spearce joined19:20
fr0sty try spelling AUTHOR_DATE correctly?19:23
(GIT_AUTHOR_DATE, that is)19:23
→ ChUrP joined19:23
← urkud__ left19:24
→ gmarselis__ joined19:26
← doener left19:27
← cmyk left19:28
← project2501b left19:29
jf_ would it be possible to fake the username provided to svn with the author of the patch on git's branch?19:29
peritus- fr0sty: works. thanks for you patience19:29
← RORgasm left19:29
→ RORgasm joined19:30
→ urkud__ joined19:31
← f00li5h left19:31
→ tobiassjosten joined19:31
→ hipe joined19:31
← poseid left19:31
← eletuchy_ left19:32
← eighty4 left19:32
← MattDiPasquale left19:34
← hauserserser_ left19:34
→ Bass10 joined19:36
→ rubyruy joined19:37
→ poseid joined19:38
→ picardo joined19:38
← Bass10 left19:38
picardo hi19:38
how would i overwrite a version in the remote server?19:38
→ hauserserser_ joined19:38
cbreak by force pushing19:38
picardo i made quite a few changes and now i am getting the non-fast forward error19:38
cbreak prepare to annoy everyone else who pulled though19:39
wereHamster picardo: faq non-ff19:39
Gitbot picardo: Your push would lose changes on the remote. See https://git.wiki.kernel.org/index.php/GitFaq#non-ff19:39
picardo it's just me who uses it19:39
→ Bass10 joined19:39
picardo thanks19:39
← ods15 left19:39
← felipec left19:40
← rtl left19:41
← redondos left19:42
→ johnw joined19:42
→ redondos joined19:42
← rubydiamond left19:43
← tstclair left19:43
← jzacsh left19:44
→ tstclair joined19:44
→ webus joined19:44
← peritus- left19:44
← poseid left19:45
→ jzacsh joined19:46
→ jrmuizel joined19:47
← jrmuizel left19:47
→ poseid joined19:49
→ alezandro joined19:49
← disappearedng_ left19:50
→ peritus- joined19:50
→ krawek joined19:50
→ d0k joined19:50
← ruskie left19:51
→ priidu joined19:51
→ Tureba joined19:51
← k3yvn left19:53
→ ruskie joined19:53
← poseid left19:53
→ disappearedng_ joined19:54
← marc_kdab left19:55
→ marc_kdab joined19:57
→ khelll joined19:57
→ j0shua joined19:58
← _numbers left19:58
← redondos left19:59
→ redondos joined19:59
→ djustice joined20:02
→ anjen joined20:02
djustice kk. accidentally commitd a wrong .png file, eternal blob. -.- how do i erase it from the git history?20:02
cbreak for total history rewrite, consider git filter-branch or git rebase -i20:04
but there's a price you have to pay20:04
and the consequences for missuse can be severe20:05
← chrislerum left20:05
shruggar aka "but if other work is based on a commit which references the mistake, you'll need to inform whoever made that work"20:06
people seem to be in a "git is dangerous" mood today20:06
cbreak I always am20:06
→ sgh joined20:06
cbreak git's dangerous because it's powerful20:06
→ ive joined20:07
→ froschi joined20:07
djustice hmm. i see. thx.20:07
← froschi left20:07
← cilly left20:07
rudi_s djustice: If you haven't pushed the commits your're editing with git rebase -i then it's fine, otherwise you have the problem mentioned above.20:08
shruggar git has a reflog and a fsck command, the only danger comes from not noticing a mistake within a months. For almost anything else, you're only in danger of annoyance20:08
→ shadowmaster joined20:08
shruggar however, giving the people the impression that git is dangerous, or easy to break, leads them to say "oh no I screwed up!" kill there repos, and re-clone. That cannot be recovered from.20:09
*their repos20:09
shruggar whips himself20:09
→ WebDragon|laptop joined20:09
← lhz left20:09
cbreak all hail the power of the reflog!20:11
johnw you can flog yourself with it, and then re-flog as necessary20:11
← gmcinnes left20:11
← posciak left20:11
← peritus- left20:11
shruggar thanks, johnw, now my stomach is inside-out from a groan. Well played.20:12
→ posciak joined20:12
vinse how do i reference the next commit?20:12
sgh shruggar: true tue ..... I accidential lost a branch a couple of weeks ago. I found out 5 days after the fact "git fsck --lost-found > somefile" at least gave me the data back.20:12
vinse im like 6 commits below head, and i want to checkout each one incrementally to find where a bug was introduced20:12
johnw vinse: use git bisect?20:13
shruggar vinse, sounds like you want bisect20:13
← WebDragon left20:13
← webus left20:13
vinse ok then, will check it out20:13
← rindolf left20:13
vinse " Find by binary search the change that introduced a bug" golly, that kinda does sound like what i was trying to do!20:14
← Yuffster left20:14
sgh Is it possible to setup git to always represent time in my local timezone ?20:14
← masterkorp left20:14
→ chrislerum joined20:15
← sixteneighty left20:15
shruggar sgh: git config --global log.date local20:16
sgh, see man git-config for more20:16
Gitbot sgh: the git-config manpage can be found at http://git.or.cz/man/git-config20:16
→ MattDiPasquale joined20:17
→ Aaaarg_ joined20:19
str1ngs I'm having a problem with sparse checkouts if I echo foo/ > .git/info/sparse-checkout . I still end up seeing foo* in the tree. is there away to explicty add a path?20:19
sgh shruggar: thanks ... I did look, nut I missed log.date ..20:19
→ steveire_ joined20:20
← Aaaarg left20:20
Aaaarg_ → Aaaarg20:20
→ whaley joined20:21
→ endpoint_david joined20:22
→ rtl joined20:23
endpoint_david is there currently a way to see the commit-id from a git-grep line; basically showing the output of `git blame` for lines matched by `git grep`?20:23
← centr0 left20:23
→ eighty4 joined20:24
endpoint_david not seeing it on the man page for `git grep`, but I've thought many times that would be a useful feature20:24
→ mw joined20:26
← mw left20:26
→ mw joined20:26
← rorist left20:27
→ tvw joined20:27
→ eletuchy joined20:28
← alester left20:29
→ alester joined20:30
Circuitsoft git grep -n ab | sed 's/^\([^:]*\):\([0-9]*\):.*/git blame \1 -L \2,\2/' | sh20:31
Replace "ab" with what you're searching for20:31
← ph^ left20:32
endpoint_david Circuitsoft: nice, thanks20:32
→ project2501b joined20:33
← pragma_ left20:35
← gmarselis__ left20:37
vinse yes bisect was exactly what i needed, thanks20:38
but out of curiosity, how can i check out the next (or previous for that matter) commit when i'm not at head?20:39
or refer to them at all, not necessarily checkout20:39
like HEAD^ , but instead of HEAD, somethign that means "where i am now"20:40
fr0sty vinse: git checkout <SHA1>^20:40
cbreak you're always at HEAD20:40
endpoint_david vinse: HEAD does mean where you are now20:40
cbreak because HEAD is where you are20:40
fr0sty (but beware of merges, you might need <SHA1>^2 (or three...)20:41
vinse i thought HEAD meant "the most recent commit in the current branch" ?20:41
→ harv joined20:41
cbreak no20:41
vinse no? where the hell did i get. that's pretty specific to have made up :/20:41
cbreak branch head != HEAD20:42
endpoint_david vinse: that's often the case (i.e., when you check out a branch explicitly)20:42
→ Nebukadneza joined20:42
endpoint_david but not always20:42
Nebukadneza heho20:42
jf_ i'm still searching for a way to dcommit from git to svn to another username than mine, is this possible?20:42
endpoint_david you can have a detached HEAD, which just means that it points to a ref that's not named20:42
vinse so HEAD just means "currently checked out"? ok, good info20:42
← khelll left20:43
vinse ok so HEAD^x is x commits back, is there an analogous way to go x commits forward?20:43
← khmarbaise left20:43
endpoint_david vinse: no, as there could be any number of commits based on that commit20:43
(i.e., forward to what)20:43
→ pragma_ joined20:44
Nebukadneza i was reading the git blame and log manpages, but i somehow can't come up with anything that does this - i was wondering what the easiest way to count lines of code per author via git...20:44
cbreak jf_: can you svn commit with other names?20:44
pragma_ → Guest760620:44
jf_ cbreak: yes, i own the repository20:44
cbreak if so, maybe give those credentials to git20:44
vinse Nebukadneza: git blame | grep "author name" | wc -l ?20:45
cbreak vinse: git's history is a DAG, the D stands for Directed. And the direction is into the past20:45
jf_ cbreak: sure, i don't know how to do that20:45
Nebukadneza vinse: well, almost - if there were a way to get the "history" for every line20:45
because - if someone ran e.g. a codeformatter over some code all of that would be credited to him20:45
cbreak jf_: don't know either. but somehow you will have to have done it or it couldn't authenticate to svn20:45
endpoint_david Nebukadneza: you can't distinguish intent from the tool-level20:46
← RORgasm left20:46
vinse Nebukadneza: i'm not sure how you're going to ...20:46
nm, what endpoint_david said20:46
→ RORgasm joined20:46
Nebukadneza endpoint_david: sorry, didn'tget that one20:46
← djustice left20:46
cbreak you touch it, you own it.20:46
→ djustice joined20:47
→ masterkorp joined20:47
cbreak git doesn't care what you changed. if you changed a line, you changed it.20:47
vinse Nebukadneza: blame tells you the last person who changed it, how's it going to know whether it was a "significant change" or not?20:47
endpoint_david Nebukadneza: you can't tell from your script whether it was a change that should be counted or not20:47
← ph^_ left20:47
→ ph^ joined20:47
vinse Nebukadneza: did any of those 3 simultaneous explanations do it for you? ;p20:47
Nebukadneza sure, but if there would be a way to get the history of a line one could use the first commit e.g.20:48
cbreak Nebukadneza: lines do never get changed20:48
they get added or removed20:48
endpoint_david Nebukadneza: blame already reports the earliest unchanged version of the line20:48
in its current form20:49
cbreak if you "change" a line, a new one gets created to replace an older one20:49
Nebukadneza cbreak: in diffs yeah - i wasn't sure whether git had additional magic20:49
git has so much magic for so many things you'd never believe so i thought maybe it had something for this ...20:49
cbreak git stores files in full each time you commit a change :)20:49
it doesn't even care if you just changed a spelling mistake or exchanged the whole contents of the file20:49
Nebukadneza :)20:50
kay20:50
→ khmarbaise joined20:51
Nebukadneza i'll go for blame, it'll be good enough20:51
cbreak also check out git annotate20:52
Nebukadneza okay, thanks for the pointer :)20:52
→ JPrince88 joined20:52
JPrince88 I'm trying to do a git pull but I get error: Untracked working tree file 'test/fixtures/users.yml' would be overwritten by merge. Aborting20:53
I don't care if users.yml is written over, I don't use fixtures. How can I make it do this and overwrite this anyway?20:53
→ ereslibre joined20:53
mmattice rm it20:53
or follow what git status says20:54
Circuitsoft Easiest way I can think of is to just delete it before pulling.20:54
← Nemurenai left20:54
JPrince88 ah okay20:54
thanks guys it worked20:54
now I just have to deal with some merge issues20:54
← giallu left20:54
JPrince88 When I do git commit after finishing the pull20:56
does it merge with current files or the last git commit statement i made?20:56
→ engla joined20:57
Circuitsoft I think the pull operation will merge with and into the working copy.20:58
← steveire_ left20:59
→ AkumaStreak joined20:59
→ khelll joined20:59
cbreak merge can only merge branches21:04
nothing else21:04
it will merge the current branch with what ever you pulled21:05
→ flaguy48 joined21:05
kokx cbreak: and tags ?21:05
cbreak tags are immutable labels21:05
← endpoint_david left21:06
cbreak if they are a label for a commit, then the commit and it's ancestors can be considered a branch21:06
← macek left21:06
→ hearit_ joined21:06
kokx cbreak: so, they can be merged ;)21:06
← tobiassjosten left21:06
← ereslibre left21:06
cbreak depends21:06
kokx you cannot merge to them, but you can merge them into a branch21:06
← hearit_ left21:07
← murr4y left21:08
← jf_ left21:08
← rubyruy left21:08
← ive left21:08
← mtkd left21:08
← kar8nga left21:09
jmspeex Can someone help me understand why "git submodule update" doesn't behave like I expect?21:10
Circuitsoft jmspeex: What do you expect?21:10
jmspeex I expect it to actually do something when there are changes in the remote repository21:10
kokx jmspeex: well, thats actually unexpected behaviour from my side21:10
jmspeex as far as I can tell, it does not do anything at all21:11
Circuitsoft cd <submodule>; git fetch; cd <root_repo>; git submodule update21:11
kokx Circuitsoft: no, doesn't work21:11
→ spearce` joined21:11
kokx because you need to checkout the branch ;)21:11
← lucsky left21:11
kokx if you add git checkout origin/master after git fetch, it works21:11
→ stringo0 joined21:12
jmspeex Circuitsoft: did exactly that21:12
kokx jmspeex: git submodule update makes sure that the submodule is up2date with the commit ID in the repository21:12
jmspeex kokx: well it doesn't for me21:13
← stringoO left21:13
← spearce left21:13
spearce` → spearce21:13
kokx jmspeex: it does21:13
← ajpiano left21:13
kokx and i mean, the commit ID in the root repository21:13
jmspeex if I do a "git diff", this is what I see:21:13
-Subproject commit 130e02e2c494769231337746745849dddbd48f7b21:13
+Subproject commit 6f9a028cad836b8f41f479d244be849fc16e48b821:13
kokx yes21:13
thats correct21:13
jmspeex the newer version is 130e02 and what I used to have was 6f9a021:13
kokx hmmz, thats weird21:14
jmspeex My expectation was that "git submodule update" would advance that submodule to point to 130e02, which is the newer commit21:14
← henkboom left21:14
kokx no, thats not true jmspeex21:14
jmspeex What is not true?21:15
kokx what you expect21:15
bdiego → bdiego_work21:15
kokx ok, inside your root repository, there is a subproject commit ID stored21:15
and git submodule update, makes sure your submodule is updated to that specific commit ID21:15
jmspeex Let me make sure we're in sync. On machine A, I committed 130e02 in the submodule. Then I git a commit in the "superproject".21:15
Of course, I pushed all these changes21:15
Now on machine B, I'd like to "sync everything" to have the same as I have on machine A21:16
so I did a "git pull" in the superproject, along with a "git submodule foreach git fetch"21:16
After that, I did a "git submodule update"21:16
I would have expected this to be sufficient -- it's not21:17
kokx ah, you should have committed after the git submodule foreach git fetch i think21:17
kpreid_ → kpreid21:17
kokx though i'm not 100% sure21:17
jmspeex committed what?? Machine A already committed the fact that the superproject is using a newer version21:17
→ plediii joined21:17
→ cybergirl joined21:17
kokx hmmz ok jmspeex, that is weird indeed21:17
jmspeex To be honest, I have never seen "git submodule update" do *anything* at all since the first time I used it to clone the submodules21:18
→ henkboom joined21:18
→ ebneter joined21:18
← cybergirl left21:18
Circuitsoft On Machine A, in the superproject, git add <submodule_dir>, git commit, git push.21:19
On machine B, in the superproject, git pull, git submodule update.21:19
These confusions are why I try to avoid submodule now, and use git subtree when I can.21:19
← Fullmoon left21:19
jmspeex That's pretty much what I did (except I used "git commit -a")21:19
← bauruine left21:20
Circuitsoft I'm not sure I would trust git commit -a for this type of thing.21:20
plediii jmspeex: I think it doesn't do anything until you clone the super project.21:20
jmspeex Circuitsoft: git commit -a clearly showed me it was updating the submodule "pointer"21:20
plediii: what doesn't do anything?21:21
plediii git submodule update21:21
there has to be something for it to do before it will do anything21:21
if you have a machine a and a machine b you should be seeing it do stuff21:21
→ murr4y joined21:22
← hyperair left21:22
jmspeex OK, I think I got it. It seems like my .git/config file got corrupted and the submodule entries had disappeared21:22
Of course, "git submodule update" thought the best way of failing was silently21:23
← pantsman left21:23
← shruggar left21:26
→ ChUrP1 joined21:26
→ vdboor_ joined21:28
← ChUrP left21:28
← WebDragon|laptop left21:30
← IslandUsurper left21:30
← Kage left21:32
← johnw left21:33
← chrislerum left21:33
→ nobody joined21:34
← Titosemi left21:34
← gebi left21:35
← icwiener_ left21:36
parndt_away → parndt21:36
→ chrislerum joined21:37
← sgh left21:38
← RORgasm left21:40
← coppro left21:42
→ coppro joined21:44
← lanthan_ left21:44
→ Fullmoon joined21:45
→ sergio_ joined21:45
→ freshtonic joined21:46
← bcardarella left21:47
przemoc hi, shouldn't following command give full and correct message text? git log -1 --format=format:'%s%n%n%-b' revision21:48
the problem is with imported repos (svn), where commit logs usually don't have blank line and everything goes as the subject21:49
→ mizerydearia joined21:49
przemoc %s produces here one-line string instead of multiple-line one21:49
is there any better way to get commit's exact message?21:50
← bartman left21:51
← Dave^| left21:51
← coppro left21:52
→ hilem joined21:52
→ Kage joined21:53
hilem i have a local branch, which doesn't exist in the remote repository, how do i push the local branch to the remote repository?21:53
→ Blackguard joined21:53
← Jaxan left21:53
sergio_ is there a simple way to do git2svn ?21:53
thiago_akademy try one of the existing config files21:54
→ bartman joined21:55
sergio_ hilem: http://progit.org/book/ch3-5.html21:55
Circuitsoft When I do a git-svn fetch, I starts at rev 1, even though HEAD is an svn clone from rev 1895521:55
How do I get it to start where it left off?21:55
rkreis maybe you can merge HEAD into remotes/svn?21:55
sergio_ hilem: the "Pushing" part of that chapter21:55
← tstclair left21:56
nsussman` → nsuss21:56
nsuss → nsussman`21:56
hilem sergio_: thanks! exactly what i needed21:58
Circuitsoft remotes/git-svn is already pointed to the same place.21:59
← j0shua left21:59
alester OK, here's what I have never understood.22:01
← rayners left22:01
alester I do a "git add" and add some files, but no commit.22:01
and then I say "Oh crap, what did I just add"22:02
and I do a "git diff" and there are no diffs.22:02
← kipras left22:02
alester How do I know what is in that commit-to-be?22:02
cbreak git diff shows the diff between working dir and index22:02
Circuitsoft git status22:02
cbreak if you want to show between index and HEAD, use git diff --cached22:02
alester right, I see they're green22:02
Circuitsoft or "git diff --cached"22:02
cbreak alester: also check out man git-diff22:02
Gitbot alester: the git-diff manpage can be found at http://git.or.cz/man/git-diff22:02
alester AHA22:02
OK, so it's still diff,22:02
cbreak if you do git diff HEAD, it shows between working dir and HEAD22:03
przemoc and w.r.t. commit message I am asking about sth better (i.e. it should be one command) than: git log -1 rev | sed '1,4d;s,^ ,,'22:03
← khmarbaise left22:03
→ spearce` joined22:04
← RobertLaptop_ left22:06
← rtl left22:07
← spearce left22:07
spearce` → spearce22:07
→ Buglouse joined22:07
→ julius__ joined22:09
julius__ hi22:09
← andrewl left22:10
← thierryp left22:10
julius__ ive accidently removed some lines from a file that needs manual merging, how can i get the changes back?22:10
→ ignatp joined22:12
← priidu left22:12
cbreak julius__: taken a look at git checkout?22:12
julius__: man git-checkout22:12
Gitbot julius__: the git-checkout manpage can be found at http://git.or.cz/man/git-checkout22:12
cbreak look for -p22:12
← LiamH left22:13
← tizbac left22:13
julius__ thats not a valid option in git 1.6.3.3 (ubuntu 9.04)22:14
thiago_akademy upgrade22:14
→ Cstep4 joined22:15
julius__ is planned..but the 700mb download needs to wait until i sleep22:15
the bandwith is very limited here22:15
→ priidu joined22:15
engla that's understandable22:16
I'm sure people here intended to suggest upgrading only git22:16
cbreak you're doing something wrong22:16
thiago_akademy upgrade Git only then22:16
← gbacon left22:16
engla julius__: use git help checkout to get the documentation for your installed version of git22:16
cbreak 700mb for a new git? what kind of garbage system do you use?22:17
julius__ no, thats a whole upgrade22:17
engla cbreak: stop being so aggressive, obviously this user didn't mean that for just git22:17
→ scarabx joined22:17
julius__ just upgrading git is above my knowlegde when it comes to a dist upgrade22:17
thiago_akademy get the source and build22:18
it's very easy22:18
cbreak git's fairly low on dependencies afaik22:18
I don't remember having to get more packages for compiling it22:18
← tvw left22:19
← seanius left22:20
wereHamster git pretty much only depends on libc and libz22:20
and a C compiler22:20
→ epsas joined22:20
epsas Hello22:20
julius__ ah crap. i just checked the upgrade wont upgrade git. it will stay at 1.6.3.322:20
epsas I want to write a post-update script that will execute if a certain branch has been updated22:21
cbreak 1.6.3 is fairly old22:21
epsas however, I am not sure how that information would be exposed to the post-update hook22:21
wereHamster epsas: man githooks22:21
Gitbot epsas: the githooks manpage can be found at http://git.or.cz/man/githooks22:21
→ seanius joined22:21
engla julius__: I don't use Ubuntu, but what about this? It might be a place to grab the latest git for you https://launchpad.net/~git-core/+archive/ppa22:22
→ linkdd joined22:24
linkdd hi22:24
julius__ engla, thx22:24
linkdd I have a question, I see this page : http://www.kernel.org/pub/software/scm/git/docs/v1.0.13/git-clone-pack.html talking about "git clone-pack"22:24
but when I try to run this command, git told me : "git: 'clone-pack' is not a git command. See 'git --help'."22:25
← Aaaarg left22:25
wereHamster linkdd: that's for version 1.0.13. We're now at 1.722:25
linkdd ah ok22:25
→ Aaaarg joined22:25
→ centr0 joined22:25
linkdd is there another way to do the same ?22:25
wereHamster linkdd: why don't you use one of the many tutorials or books?22:26
linkdd: also, man git-clone22:26
Gitbot linkdd: the git-clone manpage can be found at http://git.or.cz/man/git-clone22:26
cbreak what did it do?22:26
← altrux left22:26
← warthog9 left22:26
wereHamster and progit.org22:26
epsas fuck this channel22:26
← epsas left22:26
← iravid left22:26
linkdd wereHamster: I search but without find anything22:26
and I don't know what option to pass at git clone -u *** <repository>22:27
(where the option is ***)22:27
wereHamster what does -u do?22:27
ah, <upload-pack>. You only need that if git-upload-pack is not in the PATH on the remote host22:28
→ thierryp joined22:28
linkdd ok22:28
i search for it because I need to clone a repository into an existing repository (without creating a new folder containing the cloned repository)22:29
wereHamster linkdd: man git-fetch22:29
Gitbot linkdd: the git-fetch manpage can be found at http://git.or.cz/man/git-fetch22:29
linkdd maybe i can do that with git-merge ?22:29
engla linkdd: what does clone into an existing repository mean?22:29
← julius__ left22:29
← ignatp left22:30
← priidu left22:30
← khelll left22:30
→ andrewl joined22:30
← Cstep4 left22:30
linkdd git clone <first-repository> ; cd <first-repository> ; clone <second-repository> into <first-repository> and not into <first repository>/<second repository>22:31
← Sr_X left22:31
← TeckniX left22:31
cbreak that makes no sense to me22:31
engla linkdd: still missing definition22:31
cbreak you want to use fetch?22:31
dennda How can I only commit some lines of a file that I changed?22:32
← flazz left22:32
dennda aka line-based staging22:32
engla dennda: use git add -p22:32
cbreak dennda: man git-add22:32
Gitbot dennda: the git-add manpage can be found at http://git.or.cz/man/git-add22:32
linkdd engla: I don't know how to explain that :/22:32
cloning two repository in the same folder22:33
→ tschaub joined22:33
engla linkdd: ok, you can't do that. you can only have one repository per folder22:33
cbreak linkdd: man git-fetch22:33
Gitbot linkdd: the git-fetch manpage can be found at http://git.or.cz/man/git-fetch22:33
→ Mage joined22:34
linkdd cbreak: git fetch <repository> don't create the files in the directory22:34
cbreak of course not22:34
why should it?22:34
engla linkdd: fetch the secondary repository, *then* you can start to merge the branches22:34
cbreak it fetches completely different branches22:34
→ sgharms joined22:34
linkdd okay22:34
cbreak you have to check them out if you want them22:34
→ npmap joined22:34
sgharms If I do git merge --no-commit test larry, in the event of conflict in say "fileA" is test's HEAD listed first, or is larry's HEAD listed first?22:35
dennda Thanks, git is just awesome22:35
wereHamster sgharms: there is onl one HEAD in a git repo. It's what you currently have checked out22:36
← raichoo left22:36
engla sgharms: an octupus merge won't allow you to proceed to conflict resolution (?)22:36
← radioman-lt left22:36
sgharms Ehh.. let me rephrase22:37
← eighty4 left22:37
sgharms if i do git diff fileA22:37
I see at the top: <<<<<<< HEAD:path/to/filea22:37
linkdd cbreak: how to check them out ?22:37
sgharms in the middle the === bar and then at the foot >>> Another commit comment22:37
which branch does the "above bar" refer to and which branch does the "below bar" refer to?22:38
cbreak linkdd: man git-checkout22:38
Gitbot linkdd: the git-checkout manpage can be found at http://git.or.cz/man/git-checkout22:38
wereHamster sgharms: the first section is what was in your current branch, the second section what was in the branch you are trying to merge22:38
engla sgharms: did you really issue 'git merge --no-commit test larry' ?22:38
sgharms yes.22:38
cbreak linkdd: take a look at -t22:38
wereHamster sgharms: man git-merge has a description of the conflict markers22:38
Gitbot sgharms: the git-merge manpage can be found at http://git.or.cz/man/git-merge22:39
sgharms er, engla yes I did.22:39
engla sgharms: I'll leave you on your own for octopus merges with conflict resolution. It shouldn't happen.22:39
→ t0rc joined22:41
→ ajpiano joined22:42
linkdd cbreak: so, i need to use : git fetch /path/to/repo/.git ; then use git-checkout ?22:42
cbreak tip: add it as remote22:43
← thierryp left22:43
cbreak do a git branch -r to look at which branches you got22:43
← hipe left22:46
→ hipe joined22:46
← montylounge left22:46
npmap Hello.22:47
→ ignatp joined22:47
npmap What is the recommended way of serving git repos a la github via ssh?22:47
← masterkorp left22:48
← Q| left22:49
← malumalu left22:49
wereHamster npmap: gitolite22:49
Nebukadneza mhhh22:50
one other question - is there a way to rewrite history and rename a committer (e.g. "John Doe" to "JDoe")?22:50
linkdd cbreak: thanks wvery thing is working22:51
wereHamster Nebukadneza: git filer-branch22:51
← hipe left22:51
← JPrince88 left22:51
Nebukadneza wereHamster: thanks22:52
npmap wereHamster: thanks. I will set it up.22:53
← drizzd left22:54
Nebukadneza wereHamster: ah, so this: http://coffee.geek.nz/how-change-author-git.html is about right?22:54
tschaub I'm not sure I'm using git-svn correctly. I'm working on a project where the v2 line of development is going on in a subversion repo. The v3 line of development is using git.22:55
I've initialized my git repo with the svn subdir location that has 2.x development22:55
and created a 2.x branch22:55
→ madewokherd joined22:55
linkdd cbreak: it will help me with my project (if you want more informations go to #gitli, it is a linux distribution entirely managed with git)22:56
wereHamster Nebukadneza: yes, that's about right22:56
tschaub then I regularly `git svn fetch` and `git merge git-svn`22:56
Nebukadneza wereHamster: okay :)22:56
cbreak git's not a deployment system22:56
Nebukadneza theoretically i can apply this to my repository, push it to some remote and everything is "like locally" there too, right?22:56
tschaub merging changes from the 2.x subversion repo into the 2.x git branch22:56
then `git merge` 2.x into master22:57
this always results in commits with the message "Merge remote branch 'git-svn' into 2.x"22:57
I didn't know if there was a way to merge fetched changes from svn without those merge commits22:58
cbreak you can't merge without merge commits22:58
unless it's fast forward22:58
tschaub right, fast forward was what I was looking for22:58
guess I need to better understand what that means22:59
← thiago_akademy left22:59
cbreak you can't make fast forward merges on demand22:59
tschaub right, it has to do with the nature of the changes?22:59
cbreak you can only make them if the thing you want to merge in is a direct descendent22:59
→ ChUrP joined23:00
tschaub ok, since my repo is a clone from svn, it isn't obvious to me whether or not to expect that changes have a direct descendent23:00
← p3rg2z left23:00
← pavlix left23:00
→ Q| joined23:01
← ChUrP1 left23:01
tschaub I also had to `git show origin/2.x | head -n 1 | sed 's/commit //' > .git/refs/remotes/git-svn` to get my working copy correctly configured to be able to use git-svn23:02
← ph^ left23:03
tschaub this seemed more complicated than necessary - so I was curious if I've misconfigured something23:03
wereHamster a long way to say: git update-ref refs/remotes/git-svn origin/2.x23:03
tschaub ah, thank you23:03
tschaub wonders what else I've mis-documented in http://github.com/openlayers/openlayers/blob/master/readme-dev.md23:04
→ stringoO joined23:06
→ jtdowney joined23:07
← coderdad left23:07
→ coderdad joined23:07
← coderdad left23:12
← thomas_b left23:13
← didlix left23:13
tschaub thanks wereHamster and cbreak for the tips23:14
← linkdd left23:14
→ ChUrP1 joined23:15
→ jmcantrell joined23:17
← jtdowney left23:17
← brizly1 left23:18
→ brizly joined23:18
→ jtdowney joined23:18
← ChUrP left23:19
← sagsousuke left23:21
tschaub hmm, since changes to my 2.x branch only come from git-svn, it seems to me like those should be fast-forward merges23:21
→ JPrince88_ joined23:21
← JPrince88_ left23:22
→ gtklocker joined23:22
→ Tuller joined23:23
← sftp left23:24
← Yuuhi left23:25
→ ChUrP joined23:25
→ sftp joined23:25
← alester left23:25
→ johnw joined23:26
← Ryback_ left23:26
← eletuchy left23:26
← ChUrP1 left23:26
→ eletuchy joined23:27
→ alester joined23:27
← alester left23:27
← beatak left23:31
→ flazz joined23:33
← tildeequals left23:34
← chrislerum left23:36
← Paraselene_ left23:38
parndt → parndt_away23:38
← ChUrP left23:39
→ ChUrP joined23:40
→ friskd joined23:45
← Nickeeh left23:47
→ jksz joined23:52
← vdboor_ left23:53
← nsussman` left23:54
← ChUrP left23:56
← n3kl left23:57
← hearit left23:58
→ n3kl joined23:58
← n3kl left23:58
→ n3kl joined23:58
← aziz left23:59

Logs Search ←Prev date Next date→ Channels Documentation