IRCloggy #git 2011-04-28

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.

2011-04-28

← Destos left00:00
patrikf I don't know, I've never used submodules, and I'm not sure it will solve your problem.00:00
we do have a channel log though.00:00
← yhager left00:00
m3lvin patrikf: Thanks for the help, I followed these instructions http://stackoverflow.com/questions/1260748/how-do-i-remove-a-git-submodule and now the repository is working as it should.00:02
← erichammond left00:02
→ kingfishr joined00:02
patrikf interesting00:02
← smolyn left00:02
← pdelgallego_ left00:03
→ SjB joined00:04
→ rondon joined00:04
← pdelgallego left00:06
← dreiss left00:06
← dropdrive left00:07
→ erichammond joined00:08
← SjB left00:09
← kingfishr left00:10
← Vampire0 left00:11
→ yhager joined00:11
← cybernd left00:12
→ alberto56 joined00:13
← Brklyn left00:13
→ quadz|2 joined00:14
← bonhoffer left00:17
← dr_win left00:17
→ dropdrive joined00:17
→ bonhoffer joined00:17
← quadz left00:17
→ Vampire0 joined00:18
→ iftekhar joined00:19
→ rchavik joined00:19
← rchavik left00:19
→ rchavik joined00:19
← Anubisss left00:19
← siprbaum left00:20
→ siprbaum joined00:20
← erez left00:20
← jfkw left00:20
→ erez joined00:21
→ Anubisss joined00:21
→ justin_george joined00:21
← aidalgol left00:22
→ [ESphynx] joined00:22
← cjz left00:22
← justin-george left00:22
← booginga left00:23
→ aidalgol joined00:23
← ESphynx left00:24
← robinbowes left00:25
← bwidmer left00:26
→ kingfishr joined00:26
→ bwidmer joined00:26
→ robinbowes joined00:27
→ jfelchner joined00:27
← m3lvin left00:27
← MikhX left00:28
← engla left00:28
→ Ralf joined00:29
← aroop left00:29
← dtrott left00:30
← caseymcg left00:31
→ mtx_init joined00:31
← pavelz left00:32
← dim_ left00:33
→ n8o-mba joined00:34
→ dim_ joined00:34
← mtx_init left00:34
← nimrod10` left00:35
← hyperair left00:35
→ slackstation_ joined00:35
← tjholowaychuk left00:37
← slackstation left00:37
slackstation_ → slackstation00:37
← mikepack left00:37
← kingfishr left00:38
← linusoleander left00:38
← kenichi_ left00:39
→ nimrod10 joined00:40
acidjazz does git merge remove the specified branch before merging into the checked out one00:41
or does that branch still exist00:41
chrisf|work acidjazz: git merge doesnt remove *anything*.00:42
it creates a commit with multiple parents, which describes the merge.00:42
← dot_prime left00:42
← ketas left00:43
← test93939 left00:45
← c00kiemon5ter left00:45
→ joel_falcou joined00:46
→ mtx_init joined00:46
← SimonP86 left00:46
← steveoliver left00:47
→ c00kiemon5ter joined00:47
← nvardar left00:47
acidjazz chrisf|work: how do i merge a branch w/ the one i currently hvae checked out?00:49
← alip left00:50
patrikf acidjazz: i think you didn't fully understand what "merge" means in the git sense00:51
chrisf|work acidjazz: `git merge <some other branch>`00:51
patrikf you can safely remove a branch after merging, as all history has been merged, if that answers your question00:52
← metcalfc left00:52
patrikf git doesn't do that automatically because often you'll want to work on that branch again (apart from the technical difficulties that would arise with such behaviour)00:52
→ alip joined00:53
→ montylounge joined00:54
→ cooloney joined00:55
← GreYFoXGTi left00:56
← jbrokc left00:56
→ test93939 joined00:56
← Ins|de left00:56
→ GreYFoXGTi joined00:56
← test93939 left00:56
← atannus left00:57
→ caseymcg joined00:58
← foocraft left00:58
← cooloney left00:59
→ ISF joined00:59
← ISF left00:59
→ atannus joined00:59
→ cooloney joined00:59
→ kingfishr joined00:59
acidjazz i gotcha01:01
ok so heres what annoys me01:02
i have dev and beta on the same server01:02
diff directories01:02
my git server is remote01:02
i added it as origin01:02
i wanna merge dev to beta right01:02
in dev i git commit -a01:02
then git push01:02
then move to beta01:02
git fetch -all; git checkout dev; git rebase origin/dev; git checkout beta; git rebase origin/beta; git merge dev01:03
that is _retarded_01:03
http://stackoverflow.com/questions/4318161/can-git-pull-all-update-all-my-local-branches01:04
_retarded_01:04
← Chaser left01:05
← pyrony left01:05
← squiggy left01:06
← mikl left01:07
→ mikl joined01:07
← mikl left01:07
← rondon left01:11
→ Vampire0_ joined01:11
← quadz|2 left01:12
← ISF_ left01:13
← rettub left01:14
← Vampire0 left01:15
← dim_ left01:15
→ dim_ joined01:16
← caseymcg left01:16
→ rettub joined01:16
← erichammond left01:17
← justin_george left01:17
← GiT left01:17
→ spatulasnout joined01:18
→ ISF joined01:19
← ki_ left01:19
← fubada left01:20
← bonhoffer left01:20
→ bonhoffer joined01:20
→ foocraft joined01:21
→ pavelz joined01:23
← mtx_init left01:24
→ mtx_init joined01:24
→ dajhorn joined01:25
← owst left01:28
→ owst joined01:29
← spearce left01:31
→ DaveIngram joined01:32
→ shiba_yu36 joined01:33
← iftekhar left01:35
← errordeveloper left01:36
→ sterlo joined01:38
← Ramblurr left01:38
→ errordeveloper joined01:39
→ jfkw joined01:39
→ xiangfu joined01:39
→ spearce joined01:40
← spearce left01:40
→ lightcap joined01:41
← cakehero left01:41
← kingfishr left01:42
← bonhoffer left01:43
→ cschneid_ joined01:43
→ ShadowBelmolve_ joined01:43
→ bonhoffer joined01:43
← cschneid left01:43
→ seivan joined01:43
← ShadowBelmolve left01:44
← towski left01:44
← xiangfu left01:44
← n8o-mba left01:46
→ kingfishr joined01:46
→ tjholowaychuk joined01:47
← bitkiller left01:47
← whyking_ left01:48
← davidrussell left01:49
→ Ramblurr joined01:50
→ Avi0 joined01:50
← rgr left01:50
← theoretical left01:51
→ davidrussell joined01:51
← gusnan left01:52
← jfelchner left01:53
→ mosno joined01:54
mosno http://tinyurl.com/GitSurvey2010Results -> "You don't have permission to access /index.php/GitSurvey2010 on this server."01:55
→ t0rc joined01:56
mosno also, is it a bad idea to rsync -aH a .git dir from one system to another?01:57
i am the only user of both systems01:57
PerlJam mosno: why would it?01:59
mosno PerlJam: i can't think of a reason, hence my question01:59
← vsayer left01:59
PerlJam mosno: then there you go! :-)01:59
→ xiangfu joined01:59
→ curtana joined02:02
← DaveIngram left02:02
← milk left02:05
mosno PerlJam: just because i can't think of a reason does not imply that #git can't :)02:05
→ erichammond joined02:07
→ boombatower joined02:07
← boombatower left02:07
→ boombatower joined02:07
→ smolyn joined02:07
→ vsayer joined02:09
← bonhoffer left02:09
→ metcalfc joined02:10
→ MikhX joined02:10
← helgikrs left02:11
frogonwheels mosno: why don't you use git's updates?02:11
→ stliu joined02:11
→ jfelchner joined02:12
← diegoviola left02:13
→ helgikrs joined02:13
frogonwheels mosno: using a push/pull would be better in that it would be able to handle objects that got packed much more efficiently.02:13
mosno: look at this https://gist.github.com/767418 (sorry it's a dos batch file, but the idea is still there using git clone --mirror02:15
← Caelum left02:16
← _bitbucket left02:16
← sechrist left02:17
← errordeveloper left02:18
← montylounge left02:18
← Risar left02:19
← cwillu_at_work left02:19
← jones- left02:20
← doki_pen left02:20
← kpreid left02:20
→ Risar joined02:21
← xiangfu left02:21
→ cwillu_at_work joined02:22
→ rgr joined02:23
→ hyperair joined02:24
← Anubisss left02:24
→ Caelum joined02:25
→ flaguy48 joined02:26
→ dyer joined02:26
← dyer left02:26
→ dyer joined02:26
→ xerora joined02:28
→ DaveIngram joined02:28
← DaveIngram left02:28
→ _bitbucket joined02:28
mosno frogonwheels: because i'm a huge noob? :) i have been using clone mirror in another scenario, i guess it would apply here, too. i think i initially didn't use it because i wanted more fine-grained control but i didn't know enough about git. thanks :)02:30
frogonwheels: i just have to take the time and read the doco02:31
→ intripoon_ joined02:31
frogonwheels mosno: we're always noobs at something. if we never garbage-collected into packfiles, the rsync idea would work perfectly.02:31
→ jones- joined02:31
← Ramblurr left02:32
→ Ramblurr joined02:32
← intripoon left02:34
← Ralf left02:35
→ _Ralf_ joined02:35
← seivan left02:35
← ptl left02:35
→ tao_ joined02:35
← dyer left02:35
tao_ I'm in a branch via "git checkout -b mybranch". I want to revert all changes. How do I do that?02:36
or abandon branch entirely, if possible02:37
frogonwheels tao_: git checkout <the branch you were in>02:38
→ bonhoffer joined02:39
frogonwheels tao_: git branch -D mybranch to discard the branch then.02:39
chrisf|work although you may want to let it sit, and think on it for a while before doing that, in case you decide those commits are interesting after all..02:39
frogonwheels tao_: yep -there's not harm in letting it sit round for a bit - there might be something useful in it after all02:40
tao_: points for doing the experiment in a new branch to start with.02:40
← owst left02:41
tao_ I've made no commits in it yet02:42
→ ptl joined02:43
→ DrDogwelder joined02:43
frogonwheels tao_: git reset --hard will THROW AWAY any changes you've made.02:43
tao_ and leave me with a clean branch?02:43
chrisf|work yes02:44
tao_ cool, that's what I want02:44
tao_ still finds it weird that git requires me to 'git add .' tracked and modified files before commit02:45
tao_ it's tracked, why do I need to add it again?02:45
ChanServ set mode: +v02:45
zivester how can i see the diff between my local copy and the remote origin/master, without updating my local copy? `git diff origin/master` doesn't work, and i dont want to fetch every branch from 'origin' is it possible just to fetch one branch, so the diff actually works ?02:45
offby1 you can't.02:46
well.02:46
← rgr left02:47
offby1 I _think_ you can fetch one branch. Of course it'll pull down everything needed for that branch, which may well include everything needed for other branches (which happen to be ancestors of the one you want).02:47
← Textmode left02:47
offby1 But when I fetch, I just say "git remote update" and let it get everything from everywhere.02:47
zivester is git remove update the same as `git fetch remotename` ?02:48
remote*02:48
offby1 I think it's the same same as "git fetch --all"02:49
in other words, "git fetch remotename" fetches just from that one remote, but "git remote update" fetches from _all_ your remotes. (You might have more than one.)02:49
← foocraft left02:49
offby1 I typically have two: my company's main repo, and my "fork" of it, where I keep semi-private work.02:49
← frogonwheels left02:50
→ schnatta_ joined02:50
← schnatta left02:50
→ rapeman joined02:51
← chrisdev left02:51
→ frogonwheels joined02:51
zivester ok, thanks for the explanation.. i guess ill just `git fetch remotename` and do a `git diff origin/master` there's really no reason why I can't have all my reps fetched, i just felt it unnecessary.. but if i understand you correctly, if my branch is a descendent of another branch, it's going to fetch all those files anyway02:51
← rapeman left02:52
← oc80z left02:52
offby1 yep02:52
general-purpose git advice: 1) Read "Git For Computer Scientists" (http://eagain.net/articles/git-for-computer-scientists/); 2) Run "gitk --all"; 3) Learn about "git reflog"02:52
← eletuchy left02:53
← markizano left02:53
→ oc80z joined02:53
zivester omg flashbacks to college 5 years ago02:53
→ amerine joined02:54
← Bass10 left02:55
← kermit left02:56
offby1 ?02:57
← welly_ left02:58
→ pyrony joined02:58
→ dreiss joined02:59
← ISF left02:59
kingfishr On one computer, when I do a fast-forward, it prints out some useful info (each file that changed, total changed lines), but on my other machine it just prints "Fast-forwarded master to 0cd8f...".03:02
WHat is the difference? The machines share a .gitconfig.03:03
→ Tac_Home joined03:05
kingfishr Also, the second machine was on an older git version, but I updated to a newer git version and it's the same, so it's not some brand-new feature.03:05
← alberto56 left03:08
→ seivan joined03:08
→ Lncn joined03:09
Lncn okay, so I want to replace FTP'ing my website with git pushing... This is harder to figure out than I could have imagined. If I "git init" on my server, what do I "get remote add" on my local box?03:10
frogonwheels Lncn: here's one idea: http://toroid.org/ams/git-website-howto03:11
offby1 kingfishr: no idea, but I'd be curious to know if you ever found out :-|03:11
Lncn frogonwheels: thanks man03:12
kingfishr offby1: OK, i think the option may be merge.stat. However, I haven't explicitly set it to be false...maybe the systems have different defaults.03:12
frogonwheels kingfishr: it's a 'diffstat' that's being printed out03:12
kingfishr frogonwheels: yeah, i found that after a tangled web of googling :D03:12
← MikhX left03:13
→ gnufied joined03:14
offby1 kingfishr: maybe one has an /etc/git/config or something03:14
→ kermit joined03:15
← jsick3 left03:16
→ ketas joined03:17
← pyrony left03:17
← ShadowBelmolve_ left03:19
→ pyrony joined03:21
← dajhorn left03:21
← deekayen left03:21
→ sechrist joined03:28
← sterlo left03:28
← dim_ left03:28
← tjholowaychuk left03:29
→ dim_ joined03:29
→ ymasory joined03:29
→ ISF joined03:29
→ dr_win joined03:29
ISF → Guest8129903:30
→ eletuchy joined03:30
kingfishr offby1: I can't find any global config on either system, but that seems like a likely scenario03:31
in any case, adding merge.stat = 1 to my ~/.gitconfig made it work03:31
← pyrony left03:33
→ gigamonkey joined03:33
← eletuchy left03:33
← aef_ left03:34
→ sterlo joined03:35
yaotti_aw → yaotti03:35
← zivester left03:36
← seivan left03:36
← [intra]lanman left03:37
→ aef joined03:37
← bonhoffer left03:37
→ zivester joined03:37
← lightcap left03:38
→ lightcap joined03:39
← lightcap left03:40
→ lightcap joined03:41
← DrDogwelder left03:41
→ victori joined03:42
→ quiliro joined03:42
quiliro hi03:43
i am starting to find the beauty in git03:43
i acn now push my project to github03:43
← kukks left03:44
quiliro but i cannot get anyone to pull it03:44
maybe i am giving them the wrong instructions03:44
git clone [email@hidden.address]03:45
← comawhite left03:45
→ comawhite joined03:45
quiliro is that the correct command?03:45
it is supposed to be public03:45
← kingfishr_ left03:45
quiliro ...and gplv303:45
→ alecthegeek joined03:45
→ MikhX joined03:46
← Lncn left03:46
← Guest81299 left03:48
offby1 looks reasonable; I'll try it03:49
← cammm left03:49
offby1 Permission denied (publickey). fatal: The remote end hung up unexpectedly03:49
quiliro: you may need to tell github that it's a "public" project.03:49
or else you'll need to collect SSH public keys from those whom you want to give access.03:49
quiliro oh...ok... i'll do it now03:50
offby1 oh hold on03:50
that URL is just for you (and other committers); for the public, you should give out this URL instead https://github.com/quiliro/sisma.git03:50
works fine03:51
quiliro thanks offby103:51
← jfkw left03:51
quiliro offby1: can i do something for you?03:51
→ segher joined03:52
offby1 I dunno, like what? Make me a sandwich.03:52
→ kingfishr_ joined03:52
quiliro lol... if you live near the equator it might be possible03:53
depending on the longitude offby103:53
← lightcap left03:53
quiliro maybe i con test something for you03:54
← yhager left03:54
← ptl left03:55
→ drubies joined03:55
← pyroscope left03:56
→ foocraft joined03:57
→ roop joined03:58
→ ExtraSpice joined04:00
→ ptl joined04:02
← OOPMan left04:02
→ fivetwentysix joined04:03
→ Textmode joined04:05
→ rchavik_ joined04:05
← kingfishr_ left04:06
← amerine left04:07
← rchavik_ left04:07
→ eletuchy joined04:07
→ rchavik_ joined04:08
← mtx_init left04:09
← blop left04:12
mosno frogonwheels: that batch file example is a bit hard to follow04:12
frogonwheels: say i have 2 git repos but no remote branches, can i still push/pull?04:13
→ bindaas joined04:13
→ blop joined04:13
mosno frogonwheels: ie. i did rsync -aH to create the second repo, then i did some commits on branch FOO of the second repo, now i want to push those commits to branch FOO on the first repo... how?04:14
(the first repo is on my production machine, the second repo is on my dev/testing machine)04:14
→ ki_ joined04:16
← blop left04:17
→ blop joined04:17
quiliro offby1: git clone git://github.com/quiliro/sisma.git04:17
→ drake01 joined04:19
← MikhX left04:19
→ avinash_hm joined04:19
← sterlo left04:19
offby1 quiliro: works fine04:19
→ rondon joined04:21
← orafu left04:21
quiliro thks offby104:22
→ orafu joined04:23
→ OOPMan joined04:23
quiliro offby1: i had been reading four hours to find clone04:23
and make it work04:23
→ Samual joined04:24
→ tgunr_ joined04:27
→ sterlo joined04:27
→ tgunr__ joined04:28
← davidrussell left04:28
← rchavik_ left04:29
← tgunr_ left04:30
→ MikhX joined04:31
← avdg left04:31
→ harshu_ joined04:31
← tao_ left04:32
← tgunr__ left04:37
← Bauer left04:38
← dim_ left04:39
→ dim_ joined04:40
→ MissionCritical joined04:41
← roop left04:42
← gedevan left04:44
→ Remixman joined04:44
← erichammond left04:45
→ codetroll joined04:45
← MikhX left04:47
→ MikhX joined04:49
→ tty1 joined04:49
← ymasory left04:51
← segher left04:52
→ lightcap joined04:52
← WebDragon left04:52
→ segher joined04:54
→ steveoliver joined04:55
← steveoliver left04:56
→ yhager joined04:56
→ fr0sty joined04:58
→ dr0id joined04:59
← dr0id left04:59
→ dr0id joined04:59
→ freimatz joined05:00
mosno frogonwheels: nevermind, i just removed the .git dir on the production server, rsynced .git from the testing server, and did a git checkout -f FOO05:01
frogonwheels: from now on, i'll use clone instead of rsync05:01
→ ThomasLocke joined05:02
← yhager left05:03
← RaceCondition left05:05
→ yhager joined05:07
← MikhX left05:07
← jfelchner left05:08
← alecthegeek left05:08
→ MikhX joined05:09
→ LongBeach joined05:09
→ roop joined05:10
→ paolodedios joined05:11
← lightcap left05:12
← paolodedios left05:13
← segher left05:14
→ seivan joined05:15
→ segher joined05:16
→ pyroscope joined05:17
← drake01 left05:18
← rondon left05:18
← metcalfc left05:22
→ mishok13 joined05:23
→ n8o-mba joined05:24
← dim_ left05:24
→ alecthegeek joined05:24
→ n1ck3 joined05:25
← Niedar left05:25
← namemon left05:26
→ dim_ joined05:26
← friskd left05:26
← MikhX left05:27
→ mtx_init joined05:28
← n8o-mba left05:28
→ MikhX joined05:29
→ aminpy joined05:29
← smolyn left05:29
← mosno left05:31
← Dvyjones left05:32
yaotti → yaotti_aw05:33
yaotti_aw → yaotti05:33
→ mosno joined05:35
ivey → ivey_off05:35
→ airborn joined05:38
← mosno left05:38
→ amerine joined05:41
← sterlo left05:42
→ sandstrom joined05:44
← sandstrom left05:44
→ sandstrom joined05:44
← gnufied left05:45
← MikhX left05:47
→ MikhX joined05:49
→ namemon joined05:50
→ nag joined05:50
→ cesario joined05:50
nag is gitosis still the preferred solution for setting up an authoritative remote branch?05:51
yaotti → yaotti_aw05:51
selckin !gitosis05:52
jast [automatic message] gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite05:52
nag thank you05:52
yaotti_aw → yaotti05:53
← mdel_ left05:54
→ Joel- joined05:56
← t0rc left05:57
→ lightcap joined05:57
← wsuff left05:58
→ PDXNative joined06:02
← PDXNative left06:03
→ mosno joined06:03
→ doki_pen joined06:04
← namemon left06:05
→ TrineLocke joined06:05
→ dc5ala joined06:05
← MikhX left06:05
← infid left06:06
→ thews joined06:07
← boombatower left06:07
→ workmad3 joined06:08
→ infid joined06:08
→ richardjoo joined06:08
← jettero left06:08
← joel_falcou left06:08
ChanServ set mode: +v06:09
→ cbreak_work joined06:09
← cesario left06:09
← eregon left06:09
→ eregon joined06:10
← alip left06:10
ChanServ set mode: +v06:10
→ jettero joined06:10
← jettero left06:10
→ jettero joined06:10
→ alip joined06:12
← richardjoo left06:12
→ richardjoo joined06:12
← dim_ left06:13
← angelsl left06:14
→ larsl joined06:15
→ dim_ joined06:15
← iivvoo left06:16
→ iivvoo joined06:16
→ durre joined06:17
← alecthegeek left06:18
→ Dvyjones joined06:19
← jonanin left06:21
→ KindOne joined06:21
→ jonanin joined06:21
KindOne I'm using TortoiseGit, any reason as to why it's taking 100% of one of my core06:22
I'm using TortoiseGit, any reason as to why it's taking 100% of one of my core's, I'm not downloading anything via GIT *06:22
→ petrux joined06:23
→ guardian joined06:24
→ MikhX joined06:26
→ wo9x joined06:27
cbreak_work have you considered stopping to use that abysmal svn imitating GUI and switching to command line?06:27
durre is this the right way to "checkout" a remote repository to start working on it's branches? "git clone [email@hidden.address] test1"06:29
← phantomcircuit left06:30
← roop left06:30
← Ruudjah left06:31
→ thiago_home joined06:33
fairuz_ → fairuz06:35
cbreak_work durre: that will give you a clone of the remote repository06:35
durre: that's the normal way to use git on an existing project06:35
you don't need to specify the destination though, it's optional06:35
← Textmode left06:37
← ptl left06:37
→ psoo joined06:37
→ anathematic joined06:39
→ gedevan joined06:42
→ sgh joined06:42
→ skoop joined06:42
← KindOne left06:43
← anathematic left06:43
→ anathematic joined06:43
→ ptl joined06:43
← guardian left06:44
← sandstrom left06:45
→ angelsl joined06:45
→ ph^ joined06:45
→ dr_ick joined06:46
→ jkp joined06:48
→ giallu joined06:48
→ gnufied joined06:50
← airborn left06:50
→ Swaagie joined06:55
← anathematic left06:57
→ eijk joined06:58
→ TTilus joined06:59
← kingfishr left06:59
TTilus when pushing changes to remote git suddenly ate awfull lot of memory (resident size ~1.3GiB) on "Compressing objects" stage and slowed down to crawl, whats up?07:01
nothing in the repo im pushing from strikes me as exceptional, same amount of files of this size (3 x ~30MiB) have existed in the repo from the beginning07:04
what should i be looking for? how to diagnose?07:04
repo tracks an svn repo using git svn07:04
→ pyrony joined07:05
cbreak_work TTilus: I'd check the sizes of things on .git/objects07:06
TTilus a friend of mine was unable to git svn clone the repo, memory consumption is thru the roof on delta compression stage and then comes "fatal, out of memory, malloc failed" and git svn crashes07:07
← wo9x left07:07
← thiago_home left07:08
→ Titosemi joined07:08
→ mklappstuhl joined07:09
TTilus goes digging .git/objects/07:09
cbreak_work TTilus: maybe someone committed a big file in subversion07:10
wereHamster try to push in smaller batches07:10
← pyrony left07:10
cbreak_work obviously, you can only git svn clone svn repositories07:10
→ glcrazy joined07:10
TTilus wereHamster: oh, how do i do that?07:10
wereHamster git push <wherever> HEAD~100:master or so07:10
TTilus cbreak_work: ah, thanks07:11
→ dharrigan joined07:12
← dharrigan left07:12
→ dharrigan joined07:12
← frerich left07:12
→ mastroDani joined07:13
TTilus cbreak_work: and yes, its the very same svn repo&branch my fried failed to git svn clone that i pull from and had difficult time pushing changes07:13
← Remixman left07:13
→ frerich joined07:14
← frerich left07:14
→ frerich joined07:14
→ arunce joined07:15
→ etnt joined07:16
yaotti → yaotti_aw07:16
→ teweWork joined07:16
← Tac_Home left07:17
← xerora left07:17
→ lanthan_ joined07:18
rob_w_away → rob_w07:18
→ drev1 joined07:18
cbreak_work TTilus: did you run du -h .git/objects?07:18
TTilus there is a *.pack file of size 2.6 in .git/objects/pack, would it cause this?07:18
cbreak_work 2.6 what?07:19
TTilus GiB07:19
cbreak_work yes.07:19
that would cause this.07:19
pack files are files that contain many objects07:20
TTilus next biggest is ./53, its 85 MiB07:20
→ sandstrom joined07:20
← sandstrom left07:20
→ sandstrom joined07:20
← lanthan left07:21
cbreak_work don't know how you can find out which is the big object though07:21
maybe the idx file contains that info07:22
← psoo left07:23
← shiba_yu36 left07:23
← dr_ick left07:23
← tonyj left07:23
← seivan left07:23
→ friskd joined07:24
→ robertdavid joined07:24
→ guardian joined07:25
→ psoo joined07:25
FauxFaux There's a script that identifies large objects; trivial to google as it's everywhere.07:25
→ tonyj joined07:25
← OOPMan left07:26
→ seivan joined07:26
→ opalka joined07:28
← opalka left07:28
→ opalka joined07:28
← Mannequin left07:29
← robertdavid left07:29
→ cesario joined07:29
→ pantsman joined07:30
← pantsman left07:30
→ pantsman joined07:30
→ Mannequin joined07:30
→ pdelgallego joined07:30
→ pdelgallego_ joined07:30
stliu → stliu_brb07:30
← mtx_init left07:32
← jlewis left07:32
→ saschagehlich joined07:32
yaotti_aw → yaotti07:32
→ icwiener joined07:33
← foocraft left07:35
→ davzie joined07:36
→ jlewis joined07:37
← drev1 left07:37
→ d0k joined07:37
← dr0id left07:40
→ whitman joined07:40
TTilus found one http://stubbisms.wordpress.com/2009/07/10/git-script-to-show-largest-pack-objects-and-trim-your-waist-line/07:42
it says largest object is 98 MiB07:42
does not really strike to me as big07:42
FauxFaux You have a single 100mb binary file in your repo..?07:43
→ drake01 joined07:43
← cyphase left07:44
← drake01 left07:45
cbreak_work TTilus: maybe you have several of them07:46
you only need 26 100mb files for 2.6gb07:46
← jkp left07:48
← richardjoo left07:48
TTilus FauxFaux: yes i do have, one07:49
cbreak_work: i don't, but the biggest one seems to change a lot07:50
cbreak_work change?07:50
you know, with 26 files I mean07:50
→ soc42 joined07:50
cbreak_work 26 files, ever.07:50
every little change is a new file07:50
if you change the last byte, new file.07:50
if you change anything in the contents, new file.07:50
TTilus exactly07:51
cbreak_work so, 26 files just means 26 different blobs with big size07:51
committing big binary files that change a lot is stupid with git07:51
TTilus and thats what i seem to have here07:51
← pantsman left07:52
→ troessner joined07:52
TTilus should batching push/pull/fetch whatever help, as suggested by wereHamster07:52
FauxFaux Are there actually any proposals for any way to make large files work in git? I'm not aware of any serious attempts.07:52
TTilus FauxFaux: what could you do if delta doesnt help?07:53
FauxFaux Someone should write a tool that can actually cope so I can stop recommending svn for that use case. /o\ /o\ /o\07:53
FauxFaux would use svn. /o /o /o\07:53
← mosno left07:53
TTilus FauxFaux: in this case the upstream repo actually _is_ svn repo =D07:53
← friskd left07:54
cbreak_work I would just store the big files outside revision control07:54
FauxFaux cbreak_work: So strongly disagree.07:54
TTilus cbreak_work: in this setup we need revision control for them07:54
← jophish left07:55
→ Dave^| joined07:55
TTilus cbreak_work: giving up revision control would be like jumping back to stone age07:55
← segher left07:55
cbreak_work the changing big binary files I work with are usually just derivatives, so why bother storing them?07:56
→ erichammond joined07:56
cbreak_work if you want them, just recompile, rerender07:56
FauxFaux You have a system that has 100% repeatable builds of anything?07:56
Please start selling that for millions.07:57
← cm2101 left07:57
← MikhX left07:57
→ Juo joined07:58
→ segher joined07:58
→ cyphase joined07:59
wereHamster FauxFaux: are you using a compiler which produces different results depending on the moon phase?07:59
FauxFaux I am using multiple compilers that fail to produce bit-identical outputs, and depend on resources available on the system and the version of the compiler and .. yes.08:00
TTilus cbreak_work: if they only were (in my case) i wouldnt be here =D08:00
wereHamster resources available? WTF?08:00
→ nevyn^_ joined08:00
TTilus even if i had system that has 100% repeatable builds it wouldnt help me here08:01
FauxFaux Uh, resources is a FauxFaux-only term that means "libraries/headers/jars/shell-scripts/..".08:01
→ rgb joined08:01
FauxFaux (I'd normally use the term "artifacts" but that means nothing to people who don't use maven.)08:01
← dc5ala left08:02
→ sigh joined08:03
→ engla joined08:03
← engla left08:03
→ engla joined08:03
← rgb247 left08:04
ichilton If you are doing different versions of the same product for different customers - with their own branding, settings etc - would you store those as branches in git?08:05
← mklappstuhl left08:05
ichilton that's what I did in subversion but git doesn't seem as suited to this08:05
selckin i'd create a proper architecture08:06
ichilton if I want to merge a change from the master into one of the client sites, i've got to git cherry-pick it as doing a normal merge tries to overwrite all the branch specific customisations08:06
FauxFaux While that strikes me as a strange idea in svn, I don't see why it's any less of a strange idea in git.08:06
ichilton selckin: what do you mean by that?08:06
FauxFaux: why strange?08:06
→ alecthegeek joined08:06
FauxFaux 'cos you end up with O(n) cherry-picks.08:07
(In either.)08:07
ichilton you lost me :)08:07
FauxFaux Every change you make in master has to be applied once for every customer; that's horrible.08:07
ichilton ah08:07
yes08:08
← eregon left08:08
ichilton but what's the alternative?08:08
→ eregon joined08:08
selckin if they are settings, they should plugin in to the main app somehow, so you'd have 1 release of your app, plus 1 repo per customer08:08
FauxFaux Move all the config to a subdirectory, or have the build-script overlay changed files, or something like that.08:08
selckin the customer repo should not by any means have your entire app, it should depend on it08:08
FauxFaux Not aware of any (non-maven <3 <3) tools to do this, though.08:08
ichilton humm08:09
that sounds equally difficult to manage08:09
as each site can change anything about the original08:09
so you'd have files everywhere to have to overlay08:09
FauxFaux Very scalable business model you have there. ¬_¬08:09
→ PigeonFriend joined08:11
← yhager left08:12
cbreak_work why does merging overwrite customisation?08:12
it should only change what changed since your last merge-base08:13
ichilton yeah but I only want to merge certain commits08:14
in svn you merge essentially a diff of those commits08:14
in git it merges a diff between the branches08:14
cbreak_work in git, a merge is an actual merge08:14
not a diff08:14
→ jatt joined08:14
ichilton nod08:14
cbreak_work it merges branches08:15
ichilton yep08:15
← gnufied left08:15
← wilmoore left08:15
← mattly left08:15
FauxFaux cherry-pick does exactly what you want, what's the problem? :)08:15
ichilton which is why branches in git a more suited to streams of development rather than separate repos with the ability to merge changes08:15
→ Remixman joined08:16
→ tatsuya joined08:16
→ mikl joined08:16
FauxFaux That's a strange comment, given a branch in git /is/ a seperate repo.08:16
→ LouisJB joined08:16
← preflex left08:17
← lanthan_ left08:18
← Juo left08:19
cbreak_work not neccessarily08:19
you can have multiple branches in the same repositories08:19
even multiple completely independent branches08:19
← pavelz left08:19
cbreak_work like Linux' development branch and Qt's master branch08:19
just doesn't make sense to do...08:19
FauxFaux I meant, in terms of what git clone -b / git fetch do.08:20
← saschagehlich left08:21
→ saschagehlich_ joined08:21
→ preflex joined08:21
→ Juo joined08:22
→ tatsuya_ joined08:23
← dreiss left08:23
← tatsuya left08:23
← Juo left08:23
← saschagehlich_ left08:25
→ saschagehlich joined08:25
→ noonoooo joined08:25
→ dr0id joined08:27
← LouisJB left08:28
→ LouisJB joined08:28
← noonoooo left08:30
→ Juo joined08:31
→ giuserpe_ joined08:31
← alip left08:31
→ xiangfu joined08:31
→ SRabbelier|Lappy joined08:32
→ alip joined08:33
← iAmerikan left08:35
→ achadwick joined08:35
← ptl left08:36
→ noonoooo joined08:37
→ RaceCondition joined08:37
← andyl left08:40
→ ptl joined08:41
→ Schmallon joined08:42
← lightcap left08:42
→ andyl joined08:44
→ mklappstuhl joined08:44
→ hynek joined08:45
← alecthegeek left08:46
→ jhoepken joined08:48
→ mattions joined08:49
← Avi0 left08:51
← gregglind left08:57
→ gregglind joined08:57
← da5id left08:59
→ n8o-mba joined09:00
→ sebbi joined09:01
← helgikrs left09:01
→ dpy joined09:01
← thoolihan left09:04
← n8o-mba left09:05
← dim_ left09:08
→ dim_ joined09:09
← robinbowes left09:10
← LekeFly left09:16
→ Alien_Freak joined09:16
← csgeek left09:17
→ eletuchy_ joined09:17
→ helgikrs joined09:17
← bremner_ left09:17
→ bremner_ joined09:19
→ errordeveloper joined09:19
← eletuchy left09:19
← harshu_ left09:21
→ arunce_ joined09:22
← Garen left09:22
→ Garen joined09:22
→ robinbowes joined09:23
→ lanthan joined09:23
← arunce left09:24
← lanthan left09:24
→ harshu__ joined09:24
→ robertdavid joined09:25
→ arunce joined09:26
→ lanthan joined09:26
← blast_hardcheese left09:28
← teweWork left09:28
← arunce_ left09:29
← mklappstuhl left09:30
→ rgr joined09:31
→ blast_hardcheese joined09:31
← saschagehlich left09:31
→ kingfishr joined09:31
→ foocraft joined09:33
← Spockz left09:34
→ kms joined09:34
→ mklappstuhl joined09:36
← d0k left09:41
→ Behi joined09:44
→ teweWork joined09:45
→ snearch joined09:45
→ whyking_ joined09:46
→ iAmerikan joined09:47
→ namemon joined09:48
→ tao_ joined09:48
→ icwiener_ joined09:48
← icwiener left09:49
→ tgunr_ joined09:50
→ da5id joined09:50
← dim_ left09:50
→ deuterium joined09:51
→ dim_ joined09:51
← tgunr left09:51
→ saschagehlich joined09:52
← saschagehlich left09:56
→ saschagehlich joined09:57
← mklappstuhl left09:57
← namemon left09:58
← mikl left10:00
→ rondon joined10:00
→ henrikbjorn joined10:01
← nevyn^_ left10:03
→ linusoleander joined10:03
← aidalgol left10:06
← cooloney left10:09
← ph^ left10:10
ec|detached → elliottcable10:12
→ JacobSingh joined10:14
← erez left10:14
JacobSingh I need to rebase 100 or so commits on a branch into another branch. I want to make one commit giving credit to author X. Author X did about 95 of the 100 commits, however, the first one is by author Y.10:15
→ nils_r joined10:15
→ bonhoffer joined10:15
JacobSingh If I s(quash) all the commits except the first one, I can edit the commit message, but the wrong dude is going to get credit for it10:15
I can't seem to squash the first one and bring in one of the later ones...10:15
FauxFaux You can adjust the author at will; I think e(dit) will let you re-do the commit with --author.10:16
nils_r Is it possible to commit, pull, push etc to submodules?10:16
JacobSingh FauxFaux: So I'm correct that it is impossible to grab one from the middle, huh?10:16
I want to "s" them all, and then mark the first one as "e"10:17
and then commit --ammend --author = "The Right guy"10:17
← AAA_awright left10:17
yaotti → yaotti_aw10:17
FauxFaux It sounds like you want to cherry-pick --no-commit one commit, then merge --squash the remaining.10:17
Or do that a few times, then rebase -i master and reorder the commits.10:17
← captainramen left10:17
yaotti_aw → yaotti10:17
JacobSingh FauxFaux: ah, that's an interesting idea...10:18
So I go to master, and then cherry the ones I don't want as part of the big commit message10:18
← dim_ left10:18
← IcePic left10:18
← Joel- left10:18
JacobSingh but then do I commit that? That part confuses me10:19
← workmad3 left10:19
→ dim_ joined10:19
← rondon left10:20
→ workmad3 joined10:20
FauxFaux Commit what? cherry-pick will commit..10:20
← bonhoffer left10:21
→ mikl joined10:21
→ mklappstuhl joined10:21
mikl → Guest3741210:21
← mastroDani left10:21
← candybar left10:22
→ IcePic joined10:23
ichilton i'm doing this, but it doesn't seem to do anything:10:24
git cherry-pick -x e13cc94fd9f015b263ee0afe80c40f5d065ab1f410:24
am I doing something wrong?10:24
bremner_ are you on the branch you are cherry-picking from?10:24
→ OOPMan joined10:25
← teweWork left10:25
← drubies left10:26
ichilton bremner_: no, the one I want to change10:26
→ muneeb joined10:26
← angelsl left10:27
→ gnufied joined10:28
← soc42 left10:29
ichilton n/m - I forgot to push up to the server and pull down so it didn't know about that commit :)10:30
shame it didn't say invalid sha1!10:31
→ erez joined10:31
yaotti → yaotti_aw10:34
bremner_ strange10:36
← mklappstuhl left10:36
→ mklappstuhl joined10:38
→ md-2 joined10:38
→ teweWork joined10:39
→ mattly joined10:40
← adreghiciu left10:41
→ Psy-Q joined10:41
ChanServ set mode: +v10:42
Psy-Q hello10:42
we are seeing a very weird error that we can't resolve. when we merge or pull, we get: error: git checkout-index: unable to create symlink foo (File name too long)10:42
← mklappstuhl left10:42
stliu_brb → stliu10:43
← Swaagie left10:43
Psy-Q i can't find anything about this error and it seems impossible to resolve it. the file name is certainly not too long, it's 90 characters including the entire path and 13 characters for the filename itself10:44
yaotti_aw → yaotti10:44
albel727 Psy-Q: someone from your dev-team uses windows, I presume?10:46
→ Error404NotFound joined10:46
Psy-Q albel727: no, we're on linux and os x10:46
it doesn't seem to have anything to do with that windows symlink bug that git had on windows10:46
but who knows10:46
→ adreghiciu joined10:46
albel727 Psy-Q: hmm, strange. well, what can I say. your only chance is to fix that symlink manually.10:47
ichilton Anyone any ideas why i'm getting this when doing a git push? -10:47
! [rejected] master -> master (non-fast-forward)10:47
error: failed to push some refs to 'git@1.2.3.4:myrepo.git'10:47
To prevent you from losing history, non-fast-forward updates were rejected10:47
Psy-Q albel727: when i symlink to that file manually, it tries to overwrite on the next pull and then the same error happens10:47
albel727 ichilton: !faq non-ff10:47
jast ichilton: [automatic message] Your push would lose changes on the remote: please see https://git.wiki.kernel.org/index.php/GitFaq#non-ff10:47
ichilton All I did was make some changes to a branch, commit and push - it's done that but keeps giving that error10:47
Psy-Q maybe it has something to do with someone committing deletions on the physical file and then creating a symlink with the same name?10:48
→ derek_ joined10:48
albel727 Psy-Q: ofc. you have to fix the incoming commit. or rather, make a fixing commit and checkout only that.10:48
← derek_ left10:48
→ kemo joined10:48
→ vanmik joined10:48
albel727 is not aware of how this bug happens. He only knows how to fix it.10:48
Psy-Q albel727: so i reset --hard to the stage before that commit, remove the file, commit that and then merge in the broken commit? :\10:49
← ptr_uzl left10:49
albel727 Psy-Q: no, it won't work.10:49
← hyperair left10:49
albel727 it goes as follows.10:49
first git fetch10:49
then git reset origin/branch10:49
then git rm --cached <symlink_file_name>10:50
then recreate symlink10:50
← henrikbjorn left10:50
albel727 git add it10:50
git commit10:50
that should fix it.10:50
← alip left10:50
Psy-Q oh, alright, thanks. i'll try that10:50
removing the file and committing that produced the same error, but in a different text format :)10:51
albel727 ?10:51
← harshu__ left10:51
Psy-Q after that the error was printed as "fatal: failed to symlink 'foo': File name too long"10:52
albel727 after which "that"?10:52
Psy-Q before, it was "error: git checkout-index: unable to create symlink 'foo' (File name too long)"10:53
the "that" that i tried first10:53
albel727 which command are you executing?10:53
Psy-Q trying your advice now10:53
← tatsuya_ left10:53
→ angelsl joined10:54
→ alip joined10:55
→ SimonP86 joined10:56
← vanmik left10:56
Psy-Q i was able to remove the symlink and add a new one, as well as commit those steps. but on pull i still get "File name too long"10:56
albel727 that is really strange. did you reset to upstream?10:57
← fivetwentysix left10:57
Psy-Q i reset --hard to the last commit that seemed safe. is that wrong?10:57
i can go back some more commits10:58
albel727 and the "last commit that seemed safe" was on your branch or on upstream?10:58
Psy-Q ah. i'm going back two more10:58
→ paja joined10:58
Psy-Q that was local. upstream broke it10:58
albel727 it's wrong anyway.10:58
don't go anywhere.10:58
I have to ask, btw. do you have any diverging commits?10:59
(completely forgot to ask before)10:59
(and no, you're not supposed to reset to any state, besides upstream current state)11:00
(or the whole point of your manipulation is lost)11:00
Psy-Q ah, okay. i think i see what the point of it is11:00
i'll work on a separate clone11:00
albel727 and your reset should not be --hard11:00
or you'll be unable to reset to upstream state.11:01
Psy-Q no diverging commits by the way11:01
→ tatsuya joined11:01
Psy-Q i can work on a more or less happy state11:01
albel727 good to know11:01
albel727 suspects, that you'll be unable to do another clone at the moment.11:03
← teweWork left11:03
← kemo left11:03
→ teweWork joined11:04
Psy-Q whew. that seems to fix it11:06
where do i send the beer? :)11:06
albel727 as a background info, the reason for that error, is that due to some obscure bug, someone ended up committing not symlink contents (i.e which file it points to), but the pointed file content itself, while still preserving symlink flag on resulting blob. on pull, git tries to make a symlink pointing to a big unparseable gibberish, and linux objects. )11:06
glad to know )11:07
Psy-Q oooh, okay! so it's similar to the same issue on windows then11:07
albel727 but I doubt your beer will make it here, so a thanks is enough =)11:07
yeah11:07
the main point is a commit outside of linux.11:07
(on fs not supporting symlinks natively/good enough)11:07
Psy-Q thanks, then :) i can have luxurious alcoholic beverages shipped to places though?11:08
← arunce left11:08
Psy-Q ah, that's OS X's fault then11:08
we only have OS X and linux anyway11:08
albel727 yeah, seems so.11:08
→ rendar joined11:08
rendar can i push from a bare repo to another bare repo?11:08
→ GhettoCode joined11:08
albel727 why not?11:08
← ketas left11:08
rendar albel727, it gives me some errors, No refs in common and none specified; doing nothing.11:09
albel727 rendar: you just have to specify what and where you want to push. you may also want to check out --all, or --mirror flag,11:10
rendar albel727, ok11:10
albel727 you have to understand, what a ref is. the basic syntax is git push <repo> sourceref:targetref11:11
← aminpy left11:11
← mattions left11:12
rendar albel727, i've done git push <repo> master and it worked11:12
albel727 you don't usually have to specify sourceref and targetref, because for bare-nonbare interaction (or rather clone-cloned repo relationship) git has some configuration in ,git/config11:12
yeah, it should have11:12
worked, I mean.11:13
← robertdavid left11:13
rendar yep11:13
← rgr left11:13
← linusoleander left11:13
→ linusoleander joined11:15
→ rgr joined11:16
→ thoolihan joined11:16
← rson left11:17
← pdelgallego left11:18
← pdelgallego_ left11:19
→ nimrod10` joined11:19
→ spaceonline joined11:19
← seivan left11:19
← LouisJB left11:19
albel727 Psy-Q: btw, don't forget to push your commit. or others will have the same problems all over again =) and if someone fixes it separately from you, it might end up as a horrible merge conflict 0_011:21
← nimrod10 left11:22
Psy-Q albel727: i hope i was faster :) i pushed everything now, seems to work at least for the two people using OS X here11:22
albel727 k, good to know )11:23
and also tell people on OS X to try and avoid editing the file via symlink.11:23
Psy-Q yeah, i told everyone not try to use symlinks at all anymore on OS X11:24
albel727 would have avoided symlinks altogether, or made them separate noncommitable.11:24
albel727 ok11:24
s/separate noncommitable/not tracked by git/11:25
selckin why would i get conflict when trying to just squash a whole bunch of stuff with rebase -i ?11:26
← erez left11:26
→ ketas joined11:26
albel727 hmm, a good question, indeed.11:26
← mishok13 left11:26
albel727 can't recall seeing anything like that.11:26
→ dermoth_ joined11:27
→ Yuuhi joined11:27
albel727 tries to remember, whether git-rebase alerts about uncommitted local changes. or is it like git-checkout, which doesn't care as long as there are no conflicts.11:28
→ sou_ joined11:29
← dim_ left11:29
← gnufied left11:29
← Remixman left11:30
→ Remixman joined11:30
albel727 are you sure, you haven't reordered anything?11:30
← dermoth|home left11:31
selckin workdir is clean, just doing git rebase -i 1500commits in the past, %s/pick/fixup/g type thing11:31
conflicts all over11:31
← schnatta_ left11:31
albel727 hmm. maybe try git merge --squash first?11:31
→ dim_ joined11:31
PerlJam 1500 commits in the past? You've probably merged a few times since then?11:32
selckin sure11:32
← jophish_ left11:32
→ erez joined11:32
selckin should it not be able to handle that?11:33
PerlJam sure, but depending on what you changed, git may have to "replay" a commit that doesn't apply cleanly.11:33
albel727 hm. it should ignore that completely, afaik. interactive rebase just chooses one history.11:33
one history path.11:33
→ fr0sty|mobile joined11:34
PerlJam (i.e. you changed something that makes the commit "unclean" in the new rebased universe)11:34
← fr0sty left11:34
PerlJam merges just seem to be the synchronization points where such issues show themselves (in my limited experience)11:35
albel727 shouldn't that happen only if he uses --onto, or commit reordering?11:35
→ deuterium_ joined11:35
→ Remixman_ joined11:35
→ linusoleander_ joined11:35
← Remixman left11:36
← deuterium_ left11:36
→ hyperair joined11:37
← nimrod10` left11:37
→ nimrod10 joined11:38
← nimrod10 left11:38
→ nimrod10 joined11:38
→ ereslibre joined11:38
← ereslibre left11:38
→ ereslibre joined11:38
← flaguy48 left11:38
→ pdelgallego joined11:38
→ pdelgallego_ joined11:38
albel727 strange. anyway, I would have used git merge --squash or git-commit-tree <top_tree> -p <1500th parent>11:39
← linusoleander left11:39
linusoleander_ → linusoleander11:39
→ Remixman__ joined11:41
→ mishok13 joined11:42
← Remixman_ left11:42
← _Ralf_ left11:42
→ Ralf joined11:42
← adreghiciu left11:43
→ nevyn^_ joined11:43
→ Titosemi1 joined11:43
← gretch left11:43
→ adreghiciu joined11:44
← RaceCondition left11:44
← Titosemi left11:45
→ flaguy48 joined11:45
← Titosemi1 left11:45
← whyking_ left11:46
← adreghiciu left11:46
→ whyking_ joined11:46
yaotti → yaotti_aw11:47
→ level09 joined11:51
level09 I want to go back a few commits, is there a way to do so without reset --hard ?11:52
git checkout puts me in a detached head state11:52
→ bgy joined11:53
cehteh git checkout -b newbranch11:53
use newbranch as temporary work branch11:53
later merge or dispose it11:53
level09 but will it allow me to merge it ?11:53
cehteh why not?11:54
level09 since it is not fast forward11:54
or something11:54
no?11:54
I'll try that and see11:54
thanks :)11:54
cehteh you may also rebase it later to get a more linear history .. but rtfm to understand exactly what you are doing11:54
→ dyer joined11:55
← dyer left11:55
→ dyer joined11:55
cehteh merging doesnt require fast forwards, thats the purpose of merging11:55
← Yuuhi left11:55
← verto left11:55
→ verto joined11:55
← hyperair left11:55
← xiangfu left11:55
← dyer left11:56
→ owst joined11:57
→ robertdavid joined11:59
→ Yuuhi joined12:00
→ Remixman joined12:00
← Remixman__ left12:00
← nevyn^_ left12:01
→ babilen joined12:02
← glcrazy left12:03
← colazero left12:03
← juangiordana left12:05
→ hyperair joined12:05
← Remixman left12:06
→ Remixman_ joined12:06
→ joel_falcou joined12:06
← angelsl left12:07
→ SimonP86_ joined12:07
← jhoepken left12:07
→ SimonP86__ joined12:08
← verto left12:09
→ verto_ joined12:09
← dim_ left12:09
→ saschagehlich_ joined12:09
← saschagehlich_ left12:10
→ saschagehlich_ joined12:10
← saschagehlich left12:10
→ dim_ joined12:11
← SimonP86 left12:11
→ bitkiller joined12:12
← SimonP86_ left12:12
← gigamonkey left12:12
→ tvw joined12:13
→ angelsl joined12:13
← JacobSingh left12:13
→ [intra]lanman joined12:14
← [intra]lanman left12:14
→ [intra]lanman joined12:14
← gebi left12:14
← Robin_Watts left12:14
→ JacobSingh joined12:15
→ brockm joined12:15
← angelsl left12:17
← gedevan left12:17
→ angelsl joined12:17
→ sigh_ joined12:17
→ jfkw joined12:17
← sigh left12:18
← angelsl left12:19
→ angelsl joined12:19
← etnt left12:19
← level09 left12:19
→ paolodedios joined12:19
← troessner left12:22
→ troessner joined12:23
→ glcrazy joined12:23
yaotti_aw → yaotti12:23
← angelsl left12:23
→ angelsl joined12:23
← sou_ left12:24
← sandstrom left12:25
← angelsl left12:27
→ afeijo joined12:27
→ angelsl joined12:27
afeijo hi guys12:27
I removed a file, I want to get it from the repository, I ran 'git pull' but it didnt work. What would be the right way?12:28
← dim_ left12:28
brockm git checkout --filename12:28
frag4 git checkut12:28
← SimonP86__ left12:28
→ dim_ joined12:28
→ sandstrom joined12:28
← sandstrom left12:28
→ sandstrom joined12:28
afeijo hmmm it show "D FILENAME" but the file isnt here12:29
FauxFaux git checkout -- filename12:29
← opalka left12:29
brockm Did you use the real file's name or "filename" ? :P12:29
FauxFaux He probably did what frag4 said.12:29
afeijo now I got it! with git checkout -f12:30
thanks12:30
FauxFaux, lol I'm not that noob12:30
brockm oh...kay12:30
→ gusnan joined12:30
← Schmallon left12:30
→ jsick3 joined12:31
→ gretch joined12:34
← Modius left12:34
← KKA left12:34
→ Modius joined12:35
← Lgb left12:35
← flamingspinach left12:35
← darkskiez left12:35
← eregon left12:35
← bartman` left12:35
← ngng left12:35
← IcePic left12:35
→ KKA joined12:35
→ eregon joined12:35
← dwmw2 left12:36
← kinesis left12:36
→ flamingspinach joined12:36
→ darkskiez joined12:36
← kanru left12:36
→ n8o-mba joined12:36
→ kanru joined12:36
→ ketas- joined12:36
← saschagehlich_ left12:36
→ dwmw2 joined12:37
← Modius left12:37
→ Modius_ joined12:37
← jspiros left12:38
→ jspiros joined12:38
← paolodedios left12:38
→ bartman` joined12:38
← [intra]lanman left12:39
← ketas left12:39
← Error404NotFound left12:39
→ Remixman__ joined12:39
→ Lgb joined12:39
→ ngng joined12:39
→ kinesis joined12:39
← Remixman_ left12:39
← necrodearia left12:40
ketas- → ketas12:41
← n8o-mba left12:41
→ fivetwentysix joined12:41
→ ecuderpam joined12:41
ecuderpam What gives?12:42
Hi. I did a git svn fetch from a project with a few commits. git svn log shows me 7 revisions, but git log only shows me 1 commit, made by root. None of the svn commits were made by root.12:42
→ [intra]lanman joined12:42
ecuderpam Also, I have a directory in my working copy that is not versioned, there's no .gitignore but git status doesn't mention it.12:43
← Remixman__ left12:43
← cesario left12:44
→ cesario joined12:44
→ necrodearia joined12:45
← foocraft left12:45
← dcarter left12:45
← SRabbelier|Lappy left12:46
cbreak_work ecuderpam: git does not care about directories, just files12:46
→ amirite joined12:46
ecuderpam Hmm, I just did a git reset svn/trunk and the first part looks ok now. I thought git svn fetch updated me.12:46
→ kurko_ joined12:46
ecuderpam cbreak_work: Ah, yes, it's empty, I forgot.12:46
FauxFaux git svn rebase is what you're thinking of.12:46
→ avdg1 joined12:46
amirite how do i delete a file from git and from the entire history12:46
cbreak_work git svn fetch only fetches, you should use rebase12:46
FauxFaux amirite: man git-filter-branch12:46
jast amirite: [automatic message] the 'git-filter-branch' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-filter-branch.html12:46
cbreak_work amirite: man git-filter-branch12:46
jast amirite: [automatic message] the 'git-filter-branch' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-filter-branch.html12:46
FauxFaux Winning.12:46
ecuderpam git svn fetch && git svn rebase, or just git svn rebase?12:46
← avdg1 left12:46
→ etnt joined12:46
cbreak_work I was first here. Ha.12:47
→ avdg joined12:47
cbreak_work rebase fetches implicitly12:47
FauxFaux Can someone make jast honour "man git filter-branch"? Using the old names pains me.12:47
jast [automatic message] the 'git' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git.html12:47
FauxFaux YES THAT EXACTLY THAT12:47
ecuderpam cbreak_work: Excellent, thank you.12:47
bremner_ FauxFaux: I dunno how to make jast do things.12:48
kurko_ I have commited my code (let's call commit 1). Then I pulled from origin/master (let's call commit 2). There was a conflict. After having fixed it, what do you write for your new commit message (commit 3)? Do you describe the conflict, like 'fixed conflicts in the x file'? Or just 'merged from origin/master'?12:48
FauxFaux I assume jast is owned by someone, or is an irc client script.12:48
← clever left12:49
→ jonatasnona joined12:49
bremner_ jast is owned by jast12:49
← fivetwentysix left12:49
amirite what if i want to git-filter-branch on all files that used to exist in a given directory12:49
bremner_ amirite: that's one of the examples in the man page, I believe12:50
← f00li5h left12:50
brockm I REALLY hope jast isn't an irc client script...12:50
FauxFaux amirite: The man page explains that.12:50
brockm: Why?12:50
amirite thanks12:50
← EricInBNE left12:50
← ptl left12:51
selckin there isn't much diffrence betwean a bot and client script12:51
brockm I suppose if it's used correctly, no.12:52
Most of the time client script means "I host this at home and it disconnects frequently/is not backed up and has more limitations than it should"12:52
FauxFaux brockm: Most of us run irc clients in screen in machines in datacentres. :)12:53
selckin more people then you know run text-clients on some server12:53
brockm No, I realize that a lot of people do that12:53
ssh in and screen.12:53
→ JP-Dery joined12:54
rob_w → rob_w_away12:54
→ JamesA joined12:55
→ RichiH_ joined12:55
→ ptl joined12:55
→ fivetwentysix joined12:56
← RichiH left12:56
→ joschi_ joined12:56
→ masterhumper joined12:56
→ cxreg2 joined12:56
→ siyb_ joined12:56
→ vals_ joined12:56
→ homa_ran1 joined12:56
→ _bitbuck1t joined12:56
→ PerlPilot joined12:56
→ hyperbor1ean joined12:56
→ dotCOMmie joined12:56
→ Anubisss joined12:56
→ huehnts joined12:56
→ tonyj_ joined12:57
→ hjertnes_ joined12:57
→ batrick_ joined12:57
→ barnex1 joined12:57
→ Edulix_ joined12:57
→ frogonwheels_ joined12:57
→ sraue_ joined12:58
→ wagle_ joined12:58
→ SNEI joined12:58
← errordeveloper left12:58
→ preflex_ joined12:58
→ n1ck31 joined12:58
→ thiago_ joined12:58
→ milki_ joined12:58
→ kingfishr_ joined12:58
→ walf444 joined12:59
→ SRabbelier|Lappy joined12:59
← kingfishr left12:59
← preflex left12:59
← tonyj left12:59
← n1ck3 left12:59
← frogonwheels left12:59
← _bitbucket left12:59
← rchavik left12:59
← wagle left12:59
← tango_ left12:59
← sraue left12:59
← PerlJam left12:59
← hegge left12:59
← milki left12:59
← SJS left12:59
← walf443 left12:59
← thiago left12:59
← malfy left12:59
← barnex left12:59
← hyperboreean left12:59
← joschi left12:59
← CvJWoNSPa left12:59
← SunilThaha left12:59
← cxreg left12:59
← chitragupt left12:59
← batrick left12:59
← hjertnes left12:59
← sejo left12:59
← siyb left12:59
← Edulix left12:59
barnex1 → barnex12:59
→ xiangfu joined12:59
→ rchavik_ joined12:59
→ jeffhung joined12:59
→ malfy_ joined12:59
joschi_ → joschi12:59
→ jonmasters_ joined12:59
→ sunil joined13:00
→ chitragupt joined13:00
→ drubies joined13:00
→ dv310p3r joined13:00
← verto_ left13:00
→ jophish joined13:01
← sraue_ left13:01
→ f0i joined13:02
→ sraue joined13:02
→ foocraft joined13:02
← Davey left13:03
← snearch left13:04
→ bartek joined13:04
← xiangfu left13:04
→ hegge joined13:06
→ Ratler joined13:06
→ sebersole joined13:07
← adaptee left13:07
← gregglind left13:08
← dim_ left13:08
→ dim_ joined13:09
shamess git-diff doesn't seem to have coloured output, unless I use --color. Can I set "always" as default somewhere?13:10
FauxFaux Yes, man git-config.13:11
jast Yes: [automatic message] the 'git-config' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-config.html13:11
→ rayners joined13:11
← GhettoCode left13:11
shamess ah, epic. thanks again, FauxFaux.13:11
→ errordeveloper joined13:11
← codetroll left13:14
→ verto_ joined13:15
→ deekayen joined13:15
→ schnatta joined13:16
← nils_r left13:16
Alien_Freak → csgeek13:17
→ xiangfu joined13:17
← ianmcorvidae left13:18
→ dajhorn joined13:19
← Ramblurr left13:20
→ juangiordana joined13:21
RichiH_ → RichiH13:21
ivey_off → ivey13:21
→ Ramblurr joined13:22
← engla left13:22
→ IcePic joined13:23
→ dyer joined13:24
← dyer left13:24
→ dyer joined13:24
→ q0rban_ joined13:25
→ nicoulaj joined13:26
← q0rban_ left13:27
→ q0rban joined13:27
→ chrisdev joined13:27
← foocraft left13:28
← SethRobertson left13:28
→ opalka joined13:28
← dyer left13:29
→ foocraft joined13:29
→ SethRobertson joined13:29
→ ianmcorvidae joined13:29
→ v0n joined13:29
← jfkw left13:30
← dim_ left13:30
→ dim_ joined13:31
→ Emtec joined13:34
→ Ryback_ joined13:36
← Ryback_ left13:36
→ Ryback_ joined13:36
→ winterweaver joined13:36
← petrux left13:37
winterweaver Im in the process of creating an app that will be hosted on github. One of the components of this app, is someone else's git project, also on github. Is it possible to have a github repo within another repo? or does anyone have any recommendations on how to proceed?13:38
bremner_ winterweaver: you want submodules or gitslave13:39
winterweaver bremner_ thanks, I'll google that :)13:39
→ whyking__ joined13:39
← whyking__ left13:40
← jmcantrell left13:40
← whyking_ left13:40
← alip left13:40
→ Fullmoon joined13:41
← Fullmoon left13:41
→ jmcantrell joined13:41
← jmcantrell left13:41
← rgr left13:42
→ candybar joined13:43
→ boriskr joined13:43
← erez left13:43
→ rgr joined13:43
→ jmcantrell joined13:44
→ nixness joined13:44
← foocraft left13:44
→ alip joined13:45
← xiangfu left13:45
← rchavik_ left13:45
→ dyer joined13:45
← dyer left13:45
→ dyer joined13:45
← jmcantrell left13:45
← tao_ left13:46
ChanServ set mode: +v13:46
→ PabloM joined13:47
→ sanguisleft joined13:47
→ kenneth_reitz joined13:47
sanguisleft is there a way to have ignored files not show up in git-status messages13:48
?13:48
FauxFaux They won't, once the .gitignore is added.13:48
Emtec Hey. I have a problem. I search command to show all files with selected branch (with last_edit file, and last_commit_hash for files). I try used git ls-tree etc. but this command haven't option which i need. Git-core have command with option wich one i give ? (sorry for english).13:48
→ rchavik_ joined13:48
← sgh left13:48
→ nevyn^_ joined13:49
→ jmcantrell joined13:49
← kenneth_reitz left13:49
FauxFaux What's missing from ls-tree --recursive, Emtec?13:49
(I don't know what you mean by "last_edit file".)13:50
Emtec ls-tree show me <mode> blob hash file_name13:50
hmm sec, i'll show you13:50
← achadwick left13:50
Emtec https://github.com/andrewsmedina/django13:51
name age13:51
message13:51
← dharrigan left13:51
Emtec name file, age = last_edit, message = commit13:51
ls-tree not show last_edit file13:51
→ sgh joined13:52
← sgh left13:52
Emtec Git have command for this option, or I must make this same ?13:52
→ saschagehlich joined13:52
← nevyn^_ left13:52
sanguisleft FauxFaux: my git config look like https://gist.github.com/903720 but still I have 100's of strings like "sites/default/files/imagefield_thumbs/specialoffer_image.jpg" in my status messages. what is wrong with my line 11?13:52
→ cizra2 joined13:53
→ tgunr joined13:53
FauxFaux sanguisleft: excludesfile specifies the path to a file containing excludes, not the excludes themselves.13:53
Emtec: Working out the age of a file is hard work, unfortunately; git doesn't store it directly; github probably cache it heavily.13:53
cizra2 Hi. How do I see the names of the commits that make FETCH_HEAD different from master?13:53
sanguisleft merph13:53
← dim_ left13:53
sanguisleft ok13:53
FauxFaux git log FETCH_HEAD..master or the other way around.13:54
← ereslibre left13:54
→ ereslibre_laptop joined13:54
← ereslibre_laptop left13:54
→ ereslibre_laptop joined13:54
cizra2 FauxFaux: thanks!13:54
Emtec Yhym, thank FauxFaux for help ;)13:54
→ dharrigan joined13:54
← dharrigan left13:54
→ dharrigan joined13:54
Emtec So, i must make this ;):)13:54
← atannus left13:54
→ dim_ joined13:55
FauxFaux It's basically the output of $(log -1 --format="something" file), fwiw.13:55
→ mdel joined13:56
← jonatasnona left13:56
Emtec Hm, and ls-tree show me hash. This hash is last_commit_hash on file ?13:56
← tatsuya left13:56
→ jonatasnona joined13:57
← cizra2 left13:57
→ ericbarnes joined13:57
→ grahamc joined13:58
← opalka left14:01
→ replore joined14:01
← psoo left14:01
→ psoo joined14:02
elliottcable → ec|detached14:02
→ opalka joined14:03
→ RaceCondition joined14:04
→ logik-bomb joined14:04
← logik-bomb left14:04
→ logik-bomb joined14:04
ec|detached → elliottcable14:05
→ jfkw joined14:07
← kermit left14:07
← owst left14:08
→ owst joined14:08
→ Jippi joined14:09
← troessner left14:09
→ jbrokc joined14:10
→ troessner joined14:10
← opalka left14:10
← jmcantrell left14:10
FauxFaux Emtec: Yes.14:10
→ jmcantrell joined14:10
→ cannonball joined14:10
← mojavy left14:10
← thiago_ left14:10
→ thiago_ joined14:10
thiago_ → thiago14:10
← aef left14:12
← kurko_ left14:12
→ aef joined14:12
→ kurko_ joined14:13
← halfline left14:14
→ Niedar joined14:14
→ airborn joined14:15
← jbrokc left14:15
→ nevyn^_ joined14:16
← kurko_ left14:16
→ tearsong joined14:16
← owst left14:17
← sigh_ left14:17
→ owst joined14:17
→ LekeFly joined14:18
← airborn left14:18
← dharrigan left14:18
← candybar left14:19
← SRabbelier|Lappy left14:21
← linusoleander left14:22
→ jedahan joined14:22
→ panike joined14:22
← owst left14:23
→ owst joined14:24
→ nowhere_man joined14:24
← Jippi left14:24
dr0id hey I am back withg it issues , dang :P14:24
did a git push lhost testing14:24
now it pushed, but I can only see .git folder on my server14:24
I need to see the files and mess with them, but I don't, please guide me14:25
dr0id has faced such issues in the past, and knows it's not an issue :P14:25
→ erez joined14:25
→ spearce joined14:25
← rendar left14:26
nowhere_man dr0id: I suppose it's as in mercurial, you have to update the working copy14:26
pushing only just updates the object store14:26
dr0id I am new to VCS really. it has been tough for me :(14:26
jedahan git submodule split does not exist? i see references to it and its exactly what I need.14:26
← owst left14:26
dr0id there was nothing on production, just did a push from development14:26
I don't see the files, but just .git14:26
any tutorial that can explain me what happened ? or someone who could guide me rly :(14:27
nowhere_man dr0id: it's normal14:27
when you push, git updates the object store, the repository of versions14:27
that way, you can push to a friend's repo without messing his working copy14:27
dr0id I see.14:28
so, umm, I have no working copy right now on the production. how do I make a working copy out of the pushed objects ?14:28
bremner_ !faq non-bare14:29
jast [automatic message] Pushing to non-bare repositories is discouraged: please see https://git.wiki.kernel.org/index.php/GitFaq#non-bare14:29
dr0id core.bare=false14:29
→ Jippi joined14:29
bremner_ of course you can use git checkout on the server14:29
dr0id so it's a non-bare repo ?14:29
→ bkosborne joined14:29
dr0id git checkout is for checking out branches I guess14:29
bremner_ if it has a .git, then yes14:29
dr0id I have only 1 branch called testing14:30
so why would I checkout branches ? :P I need the files to work on heh14:30
sorry, but I am a newb.14:30
→ owst joined14:31
bremner_ well. What might be a combination of "use git checkout" and "I have one branch, named testing"?14:31
but the underlying problem is in the faq question linked to above14:31
dr0id ok, I will read that. thank you14:31
→ eric-atl joined14:34
← PabloM left14:35
← Jippi left14:35
→ Myth17 joined14:36
→ Kriss-|- joined14:36
Kriss-|- how do you add newline to the git commit -m ?14:36
sanguisleft why would the line "files/imagecache/*" in a .gitignore file still allow files like "sites/default/files/imagecache/product/jack-in-the-box.jpg" to show in the untracked files list14:37
FauxFaux Kriss-|-: Commit without -m? :)14:37
→ saschagehlich_ joined14:37
→ ThunderChicken joined14:37
← saschagehlich left14:37
→ _psychic_ joined14:37
FauxFaux Try removing the star? Shrug.14:37
Kriss-|- FauxFaux how ?14:38
cmn just delete it from that line in the file14:38
Kriss-|- your speaking to me ?14:39
→ nitish joined14:39
← nitish left14:39
→ nitish joined14:39
dr0id committing without -m should open your favourite text editor , for example, vim14:39
and then you can have fun I guess14:39
Kriss-|- how do i save it, so it comes with the push ?14:40
do i have to add the file with somehow when pushing ?14:40
→ jbrokc joined14:40
← gusnan left14:41
← Wooga left14:41
→ LetoThe2nd joined14:41
dr0id $ git checkout14:41
fatal: You are on a branch yet to be born14:41
heh14:42
bremner_: bugger, I got ya now :P did git checkout testing :P14:42
and can see the files, heh14:43
← Myth17 left14:43
bremner_ you'll have to do that every time, unless you work out something fancier, or pull on the server14:43
← owst left14:43
→ Wooga joined14:43
→ owst joined14:44
dr0id ok sir :)14:44
← mastro left14:45
LetoThe2nd i have a local working copy pulled from a svn server with git svn clone. after some changes, i cant svn dcommit, because the path on the server is not found probably this is because it contains whitespaces. but, it worked for a long time, this phenomenon came to light only today. have there been some changes in this area? might reverting to an older git version fix it?14:45
→ mkramer joined14:45
→ airborn joined14:46
→ eletuchy joined14:47
← logik-bomb left14:47
← mkramer left14:47
← eletuchy_ left14:49
→ theoretical joined14:50
→ Remixman joined14:50
→ Jippi joined14:50
← joel_falcou left14:51
→ n8o-mba joined14:51
→ beatak joined14:52
← LekeFly left14:52
→ gnufied joined14:53
← ebi left14:54
LetoThe2nd any ideas? or do i have to manually pull the changes into a fresh clone?14:54
→ DSpair joined14:54
DSpair Hello all.14:54
→ wilmoore joined14:55
→ nitish_mythology joined14:55
← Tac left14:55
← Modius_ left14:55
DSpair Question on tags in git. I notice that once you tag and then make new commits, "git describe" shows the original tag with an appended ID. What I would like to know is if it is possible to manipulate how those extended IDs are formatted?14:55
→ henrikbjorn joined14:56
→ Modius joined14:56
← Jippi left14:56
→ alex__c2022 joined14:56
← etnt left14:57
→ mastro joined14:57
→ resixian joined14:57
→ fivetwentysix_ joined14:58
← giallu left14:58
← LetoThe2nd left14:58
← fivetwentysix left14:58
fivetwentysix_ → fivetwentysix14:58
← henrikbjorn left14:59
← nitish left14:59
→ henrikbjorn joined14:59
← angelsl left15:00
adamholt → adamholt_away15:00
adamholt_away → adamholt15:00
← fivetwentysix left15:01
→ morganick joined15:01
← eletuchy left15:02
← gretch left15:02
nitish_mythology → Myth1715:02
→ eletuchy joined15:03
fr0sty-away DSpair: have you consulted man git-describe?15:03
jast DSpair: [automatic message] the 'git-describe' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-describe.html15:03
→ smolyn joined15:03
← dr0id left15:04
→ alester joined15:05
→ babusri joined15:05
→ kukks joined15:06
jedahan git filter-branch --index-filter "git rm -rf --cached --ignore-unmatch some_huge_file.tar.gz"; rm -rf .git/refs/original .git/logs; git gc --aggressive; git prune; still leaves me with a huge directory and those files still exist in the pack. Suggestions?15:07
← ereslibre_laptop left15:10
→ ereslibre_laptop joined15:10
← ereslibre_laptop left15:10
→ ereslibre_laptop joined15:10
→ alberto56 joined15:10
← sandstrom left15:11
← ki_ left15:12
← alester left15:13
→ tjholowaychuk joined15:13
→ ph^ joined15:13
→ jeremia joined15:13
→ ZeroCoder joined15:13
← jeremia left15:14
← da5id left15:14
← hyperair left15:14
→ alester joined15:15
resixian on a branch i added a submodule. but now if i try to co master is says i must remove some files (from within the modules's directory) then when i do checkout master i get an error about not being able to rm <moduel directory> (because its not empty)15:15
← cbreak_work left15:16
→ bryanl joined15:16
← alberto56 left15:16
→ sandstrom joined15:18
← sandstrom left15:18
→ sandstrom joined15:18
→ rlb3 joined15:18
← mastro left15:19
DSpair fr0sty-away: I looked at the man page and I see that I can disregard the commit count, but I would like to replace the commit ID portion or just get the commit count.15:19
→ PabloM joined15:19
← khaije|m` left15:22
fr0sty-away what use is the commit count without the hash?15:22
→ khaije|m` joined15:22
→ Tac joined15:22
← stliu left15:22
fr0sty-away if you really want that, just chop the end off yourself with sed (or similar)15:22
DSpair I want to extract the tag (version) and the commit count for automatically updating my package builder.15:22
fr0sty-away you know that tag+commit_count is ambiguous, right?15:23
DSpair For example, I want to have my build system generate a new debian/RPM package and use the tag + commit count as the new package version.15:23
← dim_ left15:23
→ petrux joined15:23
→ mkramer joined15:23
← nevyn^_ left15:24
→ dim_ joined15:24
→ ki_ joined15:24
← babusri left15:24
masterhumper → sejo15:25
DSpair fr0sty-away: Yeah, I guess a quick awk of the output is good enough.15:26
← rchavik_ left15:26
→ nitish joined15:26
← nitish left15:26
→ nitish joined15:26
→ Prism joined15:27
← teweWork left15:27
→ alvaro_o joined15:27
← alvaro_o left15:27
← mkramer left15:30
→ joel_falcou joined15:30
← Myth17 left15:30
← khaije|m` left15:31
← wilmoore left15:31
→ khaije|m` joined15:31
→ necromancer joined15:31
→ mastro joined15:32
← sandstrom left15:32
← jatt left15:33
← bkosborne left15:33
← lanthan left15:35
← Juo left15:35
→ CoinOpeBoy joined15:36
→ gedevan joined15:36
→ devth joined15:37
→ nevyn^_ joined15:38
→ Juo joined15:40
→ dreiss joined15:40
← henrikbjorn left15:40
← ericbarnes left15:41
→ pyrony joined15:41
← khaije|m` left15:41
→ khaije|m` joined15:42
→ dharrigan joined15:43
← dharrigan left15:43
→ dharrigan joined15:43
→ mikewintermute joined15:44
→ aroop joined15:44
← skoop left15:44
← paja left15:44
← freimatz left15:46
← pyrony left15:46
← nitish left15:47
→ ymasory joined15:47
→ justin-george joined15:47
→ tychoish_ joined15:47
← tychoish_ left15:47
← md-2 left15:47
→ LiamH joined15:48
→ d0k joined15:48
← verto_ left15:48
jast small test here15:48
← Guest37412 left15:48
jast just ignore15:48
→ steveoliver joined15:48
thiago hey jast15:48
← amerine left15:48
jast man git filter-branch15:49
hmm.15:49
→ Error404NotFound joined15:49
thiago dash?15:49
← Error404NotFound left15:49
→ Error404NotFound joined15:49
→ nitish joined15:49
← tgunr left15:49
→ tgunr joined15:49
← nevyn^_ left15:50
jast trying to allow version without dash because people keep moaning at me :)15:50
← NickASloan left15:50
thiago but they can't type that in the command-line15:50
bremner_ they can on debian15:50
thiago well, they can, but they get the main git manpage, then an error15:51
bremner_ no, it works15:51
thiago No manual entry for filter-branch15:51
← Juo left15:51
← khaije|m` left15:52
→ adymitruk joined15:52
jast it has been patched in some distributions15:52
← aroop left15:52
→ NickASloan joined15:52
thiago man was patched?15:52
jast yes15:52
→ clever joined15:52
→ khaije|m` joined15:52
jast might even be an upstream patch15:52
I don't quite remember the discussion15:52
← TrineLocke left15:52
← rlb3 left15:53
← smolyn left15:53
ChanServ set mode: +v15:53
→ da5id joined15:53
→ urkud_ joined15:53
adymitruk Svn rocks!15:53
jast let's try this again15:53
→ rlb3 joined15:53
jast man git filter-branch15:53
bleh15:53
adymitruk Yet another guy being miss guided on SO15:53
http://stackoverflow.com/questions/5793734/svn-git-or-mercurial-for-local-servers/5794766#579476615:54
← siyb_ left15:54
→ AdamVancouver joined15:55
← adymitruk left15:55
→ AkumaStreak joined15:56
Silex where's the misguiding?15:56
bremner_ he trolled us and left15:57
Silex ah, right15:57
jast trolled us?15:57
→ Agos joined15:58
← Agos left15:58
bremner_ well, attempted to get us all excited?15:58
jast he strongly advocated git on that SO page, and we seemed to be bitter that he got downvoted for that15:58
s/w/h/15:58
Silex I think most of the answers are good guidance on that page15:58
"use git/mercurial, avoid svn" looks sane to me15:59
← rettub left15:59
jast yeah... the topmost answer recommends SVN based on false premises, though15:59
← troessner left16:00
→ troessner joined16:01
→ verto_ joined16:01
→ rettub joined16:01
Silex hum, true16:02
bremner_ dunno, the whole thing just reminded me of http://xkcd.com/386/16:02
← AdamVancouver left16:02
Silex hehe16:03
← jsjgruber left16:04
← dyer left16:05
← gnufied left16:05
→ dyer joined16:05
← dyer left16:05
→ dyer joined16:05
→ ShadowBelmolve joined16:05
→ sandstrom joined16:06
← sandstrom left16:06
→ sandstrom joined16:06
← winterweaver left16:07
→ wilmoore joined16:07
← dreiss left16:07
→ _pingu joined16:08
→ adymitruk joined16:09
← f0i left16:10
← _pingu left16:10
→ phantomcircuit joined16:11
→ inc0 joined16:13
inc0 hi, how can I ignore commiting changes to specified file? I cloned repo and want to add settings file to ignore list, how do I do that?16:14
panike inc0: man gitignore16:15
jast inc0: [automatic message] the 'gitignore' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/gitignore.html16:15
← urkud_ left16:15
← sebbi left16:15
→ Christie joined16:16
→ d3vic3 joined16:16
→ ddkilzer joined16:17
← dyer left16:18
← SNEI left16:19
fr0sty-away inc0: but keep in mind that gitignore doesn't work on tracked files.16:19
inc0 I still cant get it work16:19
→ SJS joined16:19
inc0 how to make file untracked?16:19
fr0sty-away git rm --cached <filename> && git commit.16:20
→ f0i joined16:20
inc0 and hold on, I'll paste my /.git/info/exclude16:20
→ amerine joined16:20
← petrux left16:21
→ mattions joined16:21
inc0 /settings/local* <-only one line, I want to remove file /settings/local.py from being tracked16:22
→ slonopotamus joined16:22
← ereslibre_laptop left16:23
→ atannus joined16:23
→ drake01 joined16:23
← justin-george left16:24
→ towski joined16:24
fr0sty-away inc0: once a file is tracked by git gitignore and .git/info/exclude (and core.excludesfile, for that matter) have _no_ effect.16:24
→ SeanMcMillan joined16:25
inc0 well now after I rmed this file git status marks it as deleted16:26
→ DelphiWorld joined16:26
DelphiWorld hi16:26
→ pyrony joined16:26
DelphiWorld how do i import a svn repo into a local git repo?16:27
bryanl if i have a branch that i've pushed to origin, is there any way i can rebase those commits at a later date? Or am I stuck?16:27
← clever left16:27
→ rendar joined16:28
→ mheusser joined16:28
→ teweWork joined16:28
albel727 you're stuck. or not, if you're the only one working with origin repo.16:29
← slonopotamus left16:29
← owst left16:29
→ smolyn joined16:29
← psoo left16:30
← DelphiWorld left16:30
← avinash_hm left16:30
→ troyt joined16:31
jast another test, please ignore16:31
man git filter-branch16:31
third failure... and that's why I didn't bother trying to implement it before16:32
→ aroop joined16:32
← SeanMcMillan left16:32
→ kenichi_ joined16:33
→ hyperair joined16:33
bremner_ dmesg16:33
oops16:33
→ SeanMcMillan joined16:33
vals_ → tango_16:33
→ owst joined16:33
→ clever joined16:33
← verto_ left16:33
← kenichi_ left16:33
→ nitish_mythology joined16:33
→ dhoss_ joined16:34
ChanServ set mode: +v16:34
← chrisf|work left16:34
→ kenichi_ joined16:34
yaotti → yaotti_aw16:34
SethRobertson I would be nice, jast, if there was a web page of what commands jastbot accepts... I know a nebulous subset of them.16:34
SeanMcMillan If you rebase a branch you've already pushed, it will just hurt people who have that branch checked out.16:34
→ chrisf|work joined16:35
→ verto_ joined16:35
→ winterweaver joined16:35
yaotti_aw → yaotti16:35
jast SeanMcMillan: UTSL: http://j.mp/gitircsrc16:35
wrong nick, sorry16:35
← inc0 left16:35
SeanMcMillan np16:35
SethRobertson thanks16:35
← dhoss_ left16:35
jast if you have an idea how to fix the new syntax, feel free to let me know :}16:35
→ Myth17 joined16:36
← Myth17 left16:36
→ Myth17 joined16:36
→ Vile joined16:36
← rgr left16:36
← nitish left16:36
winterweaver so I can add a submodule only in the toplevel of my working tree. But I would like the submodule to be checked out into a sub folder within my tree. Is there a way to do that? I was thinking I could create a symbolic link in the destination dir?16:37
← nicoulaj left16:37
jast winterweaver: no such restriction exists16:38
← rgb left16:38
→ gebi joined16:39
← nitish_mythology left16:39
winterweaver jast: I got this error --> You need to run this command from the toplevel of the working tree.16:39
→ friskd joined16:39
SethRobertson Try running the command from the top level, with a directory path.16:39
→ dhoss_ joined16:40
← dhoss_ left16:40
jast git submodule add /path/to/other/repository subdir/newlocationofsubmodule16:40
← sandstrom left16:41
← dpy left16:41
winterweaver oh i see... didn't see the second line on the man page there /doh16:41
ChanServ set mode: +v16:41
Vile Good morning, All! Can you suggest any ideas about getting git cvsserver to do two things: 1. work with submodules correctly, 2. allow login through pserver ?16:41
← jbrokc left16:42
jast well, git doesn't deal in authentication, so so much for 216:42
← workmad3 left16:42
→ dhoss_ joined16:43
← dhoss_ left16:43
← SeanMcMillan left16:43
→ pdhborges joined16:43
Vile As far as I know it does authenticate through :ext:, but not through :pserver: for some reason16:43
SethRobertson And CVS doesn't understand submodules at all. Gitslave would be more CVS-like (though of course CVS would not do anything useful with the gitslave metadata16:43
→ veganadian joined16:44
→ avinash_hm joined16:44
jast iirc, :ext: means SSH takes care of authentication16:44
← amerine left16:44
→ thiago_home joined16:45
Vile SethRobertson: my main problem is that I need to provide access to git repository on some platforms that are not supported by git16:45
And, what's worse, some of those repositories contain submodules16:46
← dharrigan left16:46
Vile And commit access required :(16:47
SethRobertson Do you need to provide versioned access or can you just supply tarballs?16:47
Vile Versioned access16:47
SethRobertson What do you expect to happen with a submodule and CVS?16:47
← unreal left16:47
← PigeonFriend left16:47
← drake01 left16:48
Vile Don't know. For example something like this, I setup intermediate repository, pull submodules into it and give CVS access to those16:48
→ drake01 joined16:48
← mishok13 left16:49
→ drev1 joined16:50
Vile I can even create a script that will pull submodule separately16:50
SethRobertson That would require git subtree merge. Getting changes back out would be challenging.16:50
Vile but, git cvsserver breaks half way through when I'm trying to get repository with submodules inside16:51
→ kipras joined16:51
Vile Good thing that main repository only will be changed through CVS, and not submodules16:51
yaotti → yaotti_aw16:52
→ ch077179 joined16:53
Vile or, alternatively, I can try to build git on that platform :(16:53
→ candybar joined16:53
← candybar left16:53
→ candybar joined16:53
SethRobertson Your best bet might be to programatically convert the repos to non-submodules (retroactively), do a git-subtree-merge, then serve up the data.16:53
← hackez_work left16:53
→ helvete joined16:53
→ qdw joined16:53
SethRobertson Or switch from submodules to gitslave which will be much more friendly to one and all16:54
What platform?16:54
Vile QNX 4/QNX 6.316:54
SethRobertson Um, done? Google seems to suggest that qnx.com provides it16:54
Vile for 6.4/6.516:55
→ sandstrom joined16:55
← sandstrom left16:55
→ sandstrom joined16:55
→ cilly joined16:56
SethRobertson Not being familiar with QNX versioning systems, would it be difficult to take their QNX-ported source code and update for your version?16:56
← pdhborges left16:57
→ halfline joined16:57
SethRobertson Oh good grief. Their "git" isn't git, it is "GNU Interactive Tools"16:57
→ RichardBronosky joined16:57
→ gigamonkey joined16:58
→ nvardar joined16:58
Vile SethRobertson: 6.3 looks doable. But QNX 4 is a different story, starting from the fact that it uses Watcom C++16:59
cmn SethRobertson: it's an unfortunate name clash, try git-core16:59
← morganick left16:59
→ AdamVancouver joined16:59
→ namemon joined16:59
← owst left17:00
→ saschpe joined17:01
→ jhoepken joined17:01
→ Chillance joined17:01
gebi Vile: for git on qnx, http://lists-archives.org/git/707734-git-on-qnx.html has the required configure patches, but some tests fail17:01
→ ttelford joined17:01
← n1ck31 left17:01
→ etnt joined17:01
← adymitruk left17:02
Vile gebi: Thanks, I will check it out17:02
← davzie left17:02
→ nitish joined17:03
← dim_ left17:04
→ kumbayo joined17:05
← troyt left17:05
→ nitish_mythology joined17:05
← nitish_mythology left17:05
→ nitish_mythology joined17:05
Vile gebi: looks like this can sort out QNX 6.3 compatibility17:05
→ dim_ joined17:06
gebi seems most things are fixed in qnx 6.4.117:07
← Myth17 left17:07
→ simosx joined17:08
← simosx left17:08
→ simosx joined17:08
gebi maybe write a mail to [email@hidden.address] for a request to add git, git is a wonderfull testsuite for a plattform too ;)17:08
Vile gebi: yes, QSSL/RIM brought libraries closer to standard in recent versions. my problem is that we still have to support old versions17:09
← nitish left17:09
→ ebi joined17:10
→ justin-george joined17:10
→ yopp joined17:11
← nitish_mythology left17:11
→ phantomcircuit_ joined17:12
← teweWork left17:12
→ paolodedios joined17:12
← phantomcircuit_ left17:13
← deuterium left17:13
← pyrony left17:13
→ MikhX joined17:13
gebi ah nice to hear, i've not used qnx since ages17:13
← whitman left17:14
→ ericbarnes joined17:14
→ justin_george joined17:14
bremner_ Vile: where is the qnx stuff based, waterloo?17:15
→ metcalfc joined17:15
← avinash_hm left17:15
bremner_ oh, maybe I misread. I was just curious about RIM's qnx activities.17:16
← mdel left17:16
← justin-george left17:16
← jhoepken left17:17
→ dreiss joined17:18
→ cjz joined17:18
← drake01 left17:19
→ drake01 joined17:20
→ paja joined17:20
← orafu left17:21
← MikhX left17:22
gigamonkey How does show-branch decide what branch name to use for each commit? I.e. if a commit is present in multiple branches, say foo and bar it could just as well be bar~13 as foo~28, right?17:22
→ Vampire0 joined17:22
thiago_home right17:23
→ mikepack joined17:23
thiago_home I think it chooses the one that is closest17:23
← etnt left17:23
dropdrive Which of the git/emacs packages does #git recommend?17:23
bremner_ magit17:24
← Vampire0_ left17:24
← frogonwheels_ left17:26
bitkiller i have a branch that appears in gitk --all, though i've got no reference to it. how can i clean it up?17:26
gigamonkey What does a revision like "master^2~10" mean? I can't find an explanation in the gitrevisions man page.17:27
→ cybernd joined17:27
patrikf gigamonkey: try again17:27
gigamonkey looking at the man page?17:28
→ kermit joined17:28
patrikf gigamonkey: yeah. note that the different modifiers can be nested as in this example17:28
otherwise, it's very straightforward to find it in the manpage17:28
gigamonkey I'm not sure what you mean by nested.17:28
← cesario left17:28
gigamonkey I'd parse ^2~10 as the same as ~1217:28
But I don't think that's right because there's another commit labeled master~1217:29
patrikf ^2 means second parent17:29
panike bitkiller: what is the name of the branch (full name including refs/... stuff)17:29
→ pyrony joined17:29
→ drake01__ joined17:29
patrikf gigamonkey: i.e. when you have a merge commit, ^ = ^1 is one side and ^2 is the other side17:29
gigamonkey Ah. I read "first parent" as "one generation ago" not "first of several co-parents"17:30
← paolodedios left17:30
bitkiller panike, i don't have any ref to that branch17:31
gigamonkey Thanks.17:32
← dv310p3r left17:33
panike Can you do 'git branch --contains <sha1>'?17:33
Where sha1 is a commit on the branch?17:33
yrlnry what would that do?17:33
Oh, it seems you can.17:34
aroop → aroop|away17:34
yrlnry (Sorry, didn;'t know about--contains before)17:34
worked for me, anyway.17:35
hyperbor1ean → hyperboreean17:35
← giuserpe_ left17:35
← drake01 left17:35
← pdelgallego left17:36
← drake01__ left17:36
aroop|away → aroop17:36
→ drake01 joined17:36
← pdelgallego_ left17:36
← guardian left17:37
→ alberto56 joined17:38
bitkiller panike, there was a stash mark on that branch17:38
→ Sebastian_ joined17:38
bitkiller panike, clering the stash solved17:38
← cbreak left17:38
bitkiller panike, clearing the stash solved my problem17:39
→ cbreak joined17:39
panike OK, excellent; I had not considered stashes17:39
→ Matt_R joined17:40
→ MikhX joined17:40
bitkiller well, i tried popping it, but it wasnt popping at all....it was applying always the same stashed changes17:40
is that normal?17:40
ChanServ set mode: +v17:41
← drake01 left17:41
Sebastian_ hey there, we often use fixup commits to make changes to a branch before it is merged, so that reviewers can see how the found issues are addressed. Is the same possible for a tricky merge? git commit lets me create fixup commits for a merge without any issue, but I don't know how I would get them auto-squashed later.17:41
panike bitkiller: according to man git-stash, that is what the git stash pop command means17:42
jast bitkiller: [automatic message] the 'git-stash' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-stash.html17:42
→ drake01 joined17:42
Sebastian_ I don't need to change any commits before the merge, just the merge commit itself.17:42
panike Looks like you meant the 'git stash drop' command?17:42
← bgerber left17:42
← Adys left17:43
← cilly left17:43
→ mdel joined17:43
→ binjured_ joined17:43
← erichammond left17:44
bitkiller panike, i meant pop --> apply and remove17:45
→ Adys joined17:45
bitkiller though it was applying and not removing17:45
patrikf maybe there were conflicts17:45
← dim_ left17:45
bitkiller hmmm17:45
← RaceCondition left17:46
← kirkwood left17:46
bitkiller but wouldnt it remove anyway? (yes there were conflict)17:46
→ dv310p3r joined17:46
→ dim_ joined17:46
patrikf the manpage will answer that17:46
bitkiller patrikf, yes, and it answered17:47
binjured_ i ran a merge and got a conflict. i have branch HEAD and branch "local". what command do i run if i just want to resolve the conflict by taking "local?"17:47
bitkiller in case of conflicts, it's not removed. must remove manually with drop17:48
→ frogonwheels_ joined17:48
→ kenneth_reitz joined17:48
← drake01 left17:48
→ drake01 joined17:49
panike binjured: from man git-merge, seems like 'git merge -s theirs local' is what you are looking for17:49
jast binjured: [automatic message] the 'git-merge' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-merge.html17:49
→ vals_ joined17:50
← vals_ left17:50
→ vals__ joined17:50
← vals__ left17:50
→ vals_ joined17:50
→ captnemofulcrum joined17:50
panike Oh sorry, you will wnat to reset HEAD first and then run the merge command above17:51
→ jrots joined17:51
← tango_ left17:51
← spaceonline left17:51
→ dannas joined17:51
panike Or just manually edit the conflics present after the merge, to only use changes from 'local'17:52
→ newbie joined17:52
→ _pingu joined17:52
newbie → Guest8230417:52
← panike left17:52
vals_ → tango_17:52
→ ddilinger joined17:53
← brockm left17:53
ddilinger with gitolite, what might i need to setup to get it to send out emails whenever someone push's ?17:53
→ spaceonline joined17:53
← empity left17:54
→ jbrokc joined17:54
SethRobertson http://stackoverflow.com/questions/5055190/how-to-setup-post-receive-email-git-hook-with-gitolite17:55
→ kingfishr__ joined17:55
← f0i left17:55
→ skoop joined17:55
SethRobertson Found by googling "gitolite hooks email"17:55
← quiliro left17:55
← verto_ left17:56
← JacobSingh left17:56
← Guest82304 left17:57
← kumbayo left17:57
→ brockm joined17:57
→ RaceCondition joined17:57
← ecuderpam left17:58
← mattions left17:58
→ pdelgallego_ joined17:59
→ pdelgallego joined17:59
← babilen left18:01
← ki_ left18:01
→ pdhborges joined18:02
← pdhborges left18:02
→ deuterium joined18:03
← alberto56 left18:04
← helvete left18:04
→ hacim joined18:04
→ diegoviola joined18:04
hacim i'm trying to do a 'git fetch' from a remote repository, I authenticate, and then it just sits there, I've straced the processes on both side and they both seem to be waiting on a read()18:04
the remote side has the process 'git-upload-pack repositories/puppet.git'18:05
yrlnry Maybe something in your login profile on the remote machine that is waiting for input?18:05
← _pingu left18:06
hacim yrlnry: i cant imagine what, it worked yesterday and I haven't been online to change anything18:06
yrlnry I have nothing useful to add, sorry.18:07
jast do other remote operations work (i.e. non-git stuff)?18:07
← drake01 left18:07
hacim jast: they do18:08
i found this in the log: Apr 28 11:00:32 puppetmaster git-daemon[16715]: The remote end hung up unexpectedly18:08
jast hm.\18:08
→ drake01 joined18:08
← drake01 left18:08
jast obscure firewall misconfiguration? I don't know :)18:08
← [ESphynx] left18:08
← tvw left18:09
← AdamVancouver left18:10
→ sylr joined18:10
← ttelford left18:11
→ troyt joined18:11
→ mojavy joined18:12
→ kumbayo joined18:12
hacim interesting, i tried to clone on the server18:13
using the file path18:13
→ gnufied joined18:13
hacim error: copy-fd: write returned No space left on device18:13
fatal: failed to copy file to 'puppet/.git/objects/pack/pack-836873fbc927e0ae3e699a11ba3630a1b2b84c9f.pack': No space left on device18:13
← namemon left18:13
hacim thats a useful error message, but I have plenty of space and inodes18:13
← pdelgallego left18:13
← Emtec left18:13
← pdelgallego_ left18:13
→ ki_ joined18:14
hacim well, i tried a different filesystem and no problem18:14
→ HG` joined18:14
→ owst joined18:14
SethRobertson Sounds like either you did run out of space, or the filesystem file size limit might have been smaller than the pack file size.18:15
cbreak hacim: maybe it's some kind of server enforced limit18:15
or the error is a misdiagnosis18:15
SethRobertson Quotas could do it18:15
hacim cbreak: I think my /tmp is too small18:15
cbreak and writing failed for something else18:15
you made /tmp on a different volume?18:15
hacim /tmp is a memory filesystem, only 16Mb18:16
SethRobertson I don't think in general /tmp is where git writes incoming pack files (unless you checkout in /tmp)18:16
cbreak doesn't sound very efficient :/18:16
hacim well, I am using gitosis and git-daemon18:16
so maybe they do?18:16
cbreak I'd put /tmp on the normal root fs18:16
gitosis is just a bunch of scripts18:16
git daemon is usually just for read only access18:16
SethRobertson hacim: Oh, and !gitosis while we are at it18:17
jast hacim: [automatic message] gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite18:17
hacim yeah yeah18:17
I know18:17
migrating to gitolite is on my list18:17
← Vile left18:17
cbreak I know git creates temporary pack files when you clone from a smart protocol18:17
← sylr left18:17
← vdv left18:18
→ cilly joined18:18
← jrots left18:18
SethRobertson On the server, right?18:19
→ McManiaC joined18:20
← cilly left18:21
→ dyer joined18:21
← dyer left18:21
→ dyer joined18:21
→ babilen joined18:22
→ verto joined18:22
← owst left18:23
← yopp left18:24
→ tanuki_ joined18:25
← troessner left18:27
→ troessner joined18:27
→ carve joined18:28
carve I'm having issues setting up gitolite with a win7 box, could anyone help me?18:28
thiago_home using win7 as a server? why?18:28
carve server = debian, home box = win7 x64 (having issues on this box, the server is setup)18:28
thiago_home but gitolite is on the debian machine?18:29
→ mando joined18:29
carve yes18:29
hacim ok, i got rid of /tmp as a memory based filesystem18:29
carve my problem is with keys18:29
hacim but I'm still not able to pull18:29
← diegoviola left18:29
thiago_home hacim: ssh into the server. Do you get in or do you get authentication failure?18:29
→ linusoleander joined18:30
→ RC joined18:30
hacim thiago_home: it works fine18:30
carve i've added my pub ssh key for the root user, do I need to create one on THIS win7 box and xfer it over and add to gitolite?18:31
→ henrikbjorn joined18:31
hacim thiago_home: i can authenticate fine, and I see git packs being talked about in strace18:31
→ HG`` joined18:31
hacim for example, I see:18:31
write(5, "0073want b6558e8d2313bea8049cbf5"..., 119) = 11918:31
write(5, "0032have a25e946e2db11fc6bdc9824"..., 1604) = 160418:31
thiago_home so what's the problem with pulling?18:31
SethRobertson hacim: /tmp is not used on the client side.18:31
hacim thiago_home: i have no idea. it just does the above, and then does read(6,18:32
thiago_home: and then doesn't go any further18:32
SethRobertson hacim: It has to be something with the filesystem. Too small, quotas, file too big, etc18:32
hacim SethRobertson: I know, i was talking about the remote side18:32
→ milx joined18:32
→ lanthan joined18:32
SethRobertson hacim: try `git repack`18:32
or git gc --aggressive18:32
hacim SethRobertson: locally?18:32
SethRobertson Which side is generating the disk full error message?18:33
← lanthan left18:33
← HG` left18:33
hacim SethRobertson: well that was the server, when i tried to clone to /tmp, but I resolved that18:33
← joel_falcou left18:33
thiago_home realises that he switched person talking to18:34
SethRobertson hacim: I guess I have lost track of what the problem is, then.18:34
thiago_home carve: ssh into the server. Do you get in?18:34
milx Hello..... question - I know rewriting history is bad, but so long as my collaborators are aware, can I destroy/squash commits and push back to the central repository? What will happen when they next pull?18:34
→ opalka joined18:34
← opalka left18:34
→ opalka joined18:34
→ lanthan joined18:34
jast milx: they'll get the old history merged with the new history18:35
← ki_ left18:35
yaotti_aw → yaotti18:35
milx jast: hmm - that sounds bad. is it? Will the old commits just be sitting waiting for garbage collection or will they actually be visible in the branch?18:36
hacim SethRobertson: the problem is: I do a 'git fetch' on my client, it authenticates, but then never gets anything18:36
→ gionnico joined18:36
gionnico hello18:36
McManiaC is it possible to restrict git to go to subdirectories? I want to manage my dotfiles in $HOME with git, but I don't want to see I'm in a git project if I go to my document-folder18:36
jast milx: they'll be visible18:36
gionnico how can one get a good repository if he did git pull then corrupted something inside?18:36
(rm some file)18:36
→ pantsman joined18:36
← pantsman left18:36
→ pantsman joined18:36
jast McManiaC: no.18:37
→ verto_ joined18:37
McManiaC :(18:37
milx jast: sounds like if I split or squash some commits and push back - the changes will be applied twice in their branch when they pull. Practical example - one commit I received today really needs to be split into two so that only half of it can be merged to another branch. How should I handle it?18:37
← fr0sty|mobile left18:38
SethRobertson hacim: If you avoid using gitosis and just ssh into the server as yourself, as a test, does it work? If the answer is yes, you know where the problem is and our opinion on the chances of getting it working.18:38
→ fr0sty joined18:40
carve Could anyone tell me where I need my ssh keys for my win7 box? (currently they're in C:\Users\Carve\.ssh - but I still get a PW prompt with TortoiseGIT and Git bash says "fatal: The remote host blah blah blah")18:41
→ gebi_ joined18:41
→ erichammond joined18:41
hacim SethRobertson: 'and our opinion on the chances of getting it working'?18:41
→ p3rg2z joined18:41
SethRobertson hacim: !gitosis18:41
jast hacim: [automatic message] gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite18:41
← mikewintermute left18:42
jast carve: git on bash should ask for a password if the SSH key isn't in the right place18:42
← sanguisleft left18:42
carve jast - doesn't work18:43
it just hangs up the connection18:43
jast if it doesn't ask for a password, the problem is something else18:43
SethRobertson carve: More specifically, tortoisegit appears to use putty. Can you use putty individually to ssh around?18:43
carve I had gitosis installed previously18:43
jast it means the server authenticated you, but e.g. there might be a permissions problem on the server18:43
hacim christ18:43
carve I used plink w/ tortoiseGIT18:43
but I have gitolite now18:43
I can ssh in myself (of course I don't use SSH keys yet as I haven't set them up)18:44
← gebi left18:44
jast presumably gitolite (or your permissions within it) isn't set up correctly18:44
← RichardBronosky left18:44
carve I can however import the id_rsa key into plink correct?18:44
← homa_ran1 left18:44
→ homa_rano joined18:45
SethRobertson Yes, but does plink store it in the same place as git expects it? There is a web page that claims that storing it in Pagaent will let git's ssh find it.18:45
← ShadowBelmolve left18:45
carve I tried with both18:45
pagent and plink18:45
SethRobertson Are you using the msysgit ssh or PuTTY with TortuiseGit?18:48
Apparently it is some kind of configuration option.18:48
← DSpair left18:48
carve msysgit and tortoiseGIT18:48
→ trobrock joined18:48
SethRobertson Yes, but apparently you can either use msysgit's native ssh, or you can use an external ssh client, like PuTTY, while still using msysgit and TortuiseGit.18:49
carve msysgit is using plink18:50
thats the only option I setup18:50
besides the defaults for both apps18:50
← Kriss-|- left18:51
→ joel_falcou joined18:51
SethRobertson carve: More generally, these instructions purport to work. You should really test the ssh side without using git first, then without using gitosis, before doing the whole thing. http://petermorlion.blogspot.com/2010/03/okay-i-finally-got-git-to-work.html18:51
← trobrock left18:51
carve yeah I'm setting up ssh first18:52
just to see if thats the issue18:52
well.. setting it up to use keys instead of logins18:52
→ __sorin__ joined18:52
→ birbilakos joined18:53
__sorin__ Hi. Is there a list of commands that open gitk, and is there a way to tell Git to open something else? Thanks.18:53
← dim_ left18:53
← ymasory left18:54
← verto_ left18:54
← verto left18:54
→ ymasory joined18:54
← birbilakos left18:54
yaotti → yaotti_aw18:54
SethRobertson __sorin__: AFAIK only git-gui opens gitk18:55
fr0sty-away 'gitk' is the only one I'm aware of...18:55
though I've never used git-gui...18:55
→ connerk joined18:55
← connerk left18:55
SethRobertson You have not missed anything18:55
→ pdelgallego joined18:56
__sorin__ SethRobertson, There are some switches to some commands that open gitk. I just can't think of them. I always open it by mistake.18:56
SethRobertson __sorin__: Survey say: git-bisect, git-citool, and git-gui18:57
→ tvw joined18:57
milx jast: If I already pushed a branch to a remote server (github) and then want to rebase/squash stuff and re-push - that's OK right? Is the only problem if someone else has already cloned it?18:58
jast milx: yep18:58
→ berserkr joined18:58
milx jast: ok. and if someone had already cloned it what would be the best approach then - start a fresh branch and abandon the old one?18:58
SethRobertson __sorin__: You can try making a gitk supporting the same options in your path before gitk which runs a different command and see if that helps18:58
jast good question18:59
that would work, I guess18:59
← p3rg2z left18:59
jast I don't know if there's a better method18:59
I never need to do stuff like that :)18:59
← gnufied left18:59
__sorin__ SethRobertson, I would have liked the ability to override gitk via gitconfig or a hook.18:59
SethRobertson milx: It is not fatal if you rewrite history. It is just annoying to the other people. It probably is more annoying to have to change branches.19:00
→ sanguisleft joined19:00
milx SethRobertson: OK - so assuming everyone sharing the branch is aware of the history rewrite - how can they update their branch safely to get the new history? just pull?19:01
← HG`` left19:01
← dyer left19:01
SethRobertson `git pull --rebase` is normally all you need. If the history change involves tags being rewritten, you probably have to do something more complicated.19:01
hacim SethRobertson: turns out it was something wrong with my local repository because I just recloned and now everything is fine with the new one19:02
milx SethRobertson: OK thanks will try19:02
→ skyl joined19:03
skyl I have some commits I found from ``git log --grep 852``, say. Can I run a show of some kind that will show the diffs one after the other so I can page through them .. or sth19:04
← RaceCondition left19:04
← robertdavid left19:04
← Matt_R left19:05
SethRobertson skyl: git log --grep a --format=%h | while read f; do git show $f | less; done19:07
McManiaC is there a git bug tracker or sth?19:09
← milx left19:10
→ bgerber joined19:10
McManiaC or a wiki/irc channel for git developers?19:10
→ unreal joined19:10
iveqy McManiaC: there's a mailing list19:11
→ ShadowBelmolve joined19:12
→ pdelgallego_ joined19:12
adamholt → adamholt_away19:12
adamholt_away → adamholt19:12
← juangiordana left19:12
→ p3rg2z joined19:12
← jonatasnona left19:13
→ dajhorn_ joined19:14
← RC left19:14
← dajhorn left19:15
← rlb3 left19:15
→ Vile joined19:16
→ dyer joined19:16
← dyer left19:16
→ dyer joined19:16
preflex_ → preflex19:16
SethRobertson McManiaC: !bugtracker19:19
jast McManiaC: [automatic message] the developers of git don't use a bug tracker. If you want to report a bug, send an e-mail to the mailing list at [email@hidden.address] you can review past discussions at http://j.mp/gitlist19:19
→ lolfrenz joined19:20
lolfrenz is it common task to want to apply a commit from one branch to another?19:20
← veganadian left19:20
SethRobertson Yes. You can "merge" "rebase" or "cherry-pick" them19:21
lolfrenz I assume cherry-pick is just to "steal" one commit, let me rtfm19:21
SethRobertson merge/rebase is preferred if possible. If not, then cherry-pick19:21
← steveoliver left19:21
→ gnufied joined19:22
→ TrineLocke joined19:25
← p3rg2z left19:26
→ Bass10 joined19:26
← NetHawk left19:27
→ urkud_ joined19:27
← jedahan left19:28
→ dekroning joined19:28
← Bass10 left19:28
→ codebeaker joined19:28
→ Bass10 joined19:29
skyl thanks SethRobertson19:29
← shennyg_ left19:31
→ shennyg joined19:31
← d3vic3 left19:34
→ cesario joined19:34
← jophish left19:35
→ foo_ joined19:36
← lolfrenz left19:37
→ LekeFly joined19:38
→ Tay101 joined19:38
→ drizzd joined19:38
Tay101 hi19:38
← gedevan left19:39
← McManiaC left19:40
← mando left19:41
→ giallu joined19:41
← LekeFly left19:41
← foo_ left19:41
→ NetHawk joined19:42
→ boombatower joined19:42
← boombatower left19:42
→ boombatower joined19:42
→ psoo joined19:42
→ jophish joined19:43
← Prism left19:44
→ hvoigt joined19:44
← ymasory left19:44
→ ymasory joined19:45
← winterweaver left19:46
aroop → aroop|away19:47
aroop|away → aroop19:47
→ quiliro joined19:48
→ Juo joined19:49
← Juo left19:50
← psoo left19:50
→ Juo joined19:51
← OOPMan left19:52
→ juangiordana joined19:53
← candybar left19:55
← troessner left19:56
← sandstrom left19:58
← curtana left19:58
← replore left19:58
speakman How do I search for in which commit a particular line in a particular file last existed?19:59
← ph^ left19:59
→ psoo joined19:59
→ alberto56 joined20:00
→ dl9pf joined20:00
→ gedevan joined20:00
← dv310p3r left20:01
← dl9pf_ left20:01
← eric-atl left20:01
batrick_ → batrick20:01
← noonoooo left20:01
← Vile left20:02
← Tay101 left20:02
albel727 speakman: man git-grep?20:02
jast speakman: [automatic message] the 'git-grep' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-grep.html20:02
← hvoigt left20:02
← ericbarnes left20:03
→ mando joined20:03
albel727 speakman: oh wait. that's not it.20:03
speakman albel727: :)20:03
→ hvoigt joined20:03
speakman Although git grep does take an <tree> object. But then I have to know which tree object :)20:04
← eletuchy left20:05
albel727 yeah, that's the main problem. )20:05
← khaije|m` left20:05
→ khaije|m` joined20:05
albel727 well, ofc you can go with git rev-list | xargs -n 1 git grep...20:06
gionnico how do you merge a file if there are conflicting versions?20:06
pantsman git log -S"some string from the line" -- path/to/file20:06
speakman To clearify my question: "When did this file last contain a certain line?"20:06
albel727 damn. how could I forget that.20:06
speakman :D20:06
→ flavius joined20:07
→ OOPMan joined20:08
flavius Hi. There was a ressource on the web which taught git very well, "under the hood", by "forcing" you not to use specific commands. Unfortunately I haven't bookmarked it and I don't remember any keyword about it. Do you know which one I mean?20:08
albel727 gionnico: you edit the file to be the way you want it, and then just git add it.20:08
gionnico albel727: doesn't seem so easy.. git add -A is the same?20:08
→ mm_202 joined20:09
gionnico Bob: change one file, git commit, git push. Alice: change the same file, and find out that there was a change..20:09
flavius oh I think it was http://progit.org/book/20:09
albel727 gionnico: yeah, that's outta work. anyway man git-merge "how to resove conflicts"20:09
jast gionnico: [automatic message] the 'git-merge' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-merge.html20:09
speakman progit <320:09
→ tewecske joined20:10
← kumbayo left20:11
← dannas left20:11
→ rendez joined20:11
SethRobertson speakman: git log -S<data>20:11
← Dave^| left20:11
← gnufied left20:11
→ workmad3 joined20:12
← TrineLocke left20:12
← stamina left20:13
← sebersole left20:13
→ stamina joined20:13
← helgikrs left20:13
→ guardian joined20:13
← khaije|m` left20:15
→ excess^ joined20:15
← henrikbjorn left20:15
→ s0ber_ joined20:15
← s0ber left20:15
→ khaije|m` joined20:15
s0ber_ → s0ber20:15
→ Heimidal joined20:15
← Heimidal left20:16
→ Heimidal joined20:16
→ TrineLocke joined20:16
→ owst joined20:17
→ steveoliver joined20:21
→ fr0sty|mobile joined20:21
brandonw progit + git man pages20:21
so good :D20:21
avdg + hack internals :p20:22
→ gnufied joined20:22
→ gfto joined20:22
← _cballou left20:23
→ bodom joined20:24
← fr0sty left20:24
→ replore_ joined20:24
← kenneth_reitz left20:25
← alberto56 left20:25
← airborn left20:25
← khaije|m` left20:26
→ khaije|m` joined20:27
← owst left20:28
← dajhorn_ left20:29
→ Nate75Sanders joined20:30
→ ShadowBelmolve_ joined20:31
→ sandstrom joined20:32
← errordeveloper left20:33
→ kumbayo joined20:33
← ShadowBelmolve left20:34
→ hackez_work joined20:35
← lanthan left20:35
brockm If I have a project that is based on another project, but I want to be able to commit changes to both, ie keep the original branch and my personal changes seperate... would I set those up in the same repository or make one for each?20:35
carve can you utilize dsa keys with gitolite (debian server) and a win7 box as the client, or do I need to create a rsa keypair too? (SSH is working with the DSA as-is for just plain SSH)20:35
→ derek__ joined20:36
← derek__ left20:36
→ vanmik joined20:36
carve brock you could fork the original and just merge the forks changes20:36
brockm okay but specifically how would I have a master for that fork AND a master for my personal changes in the same repository?20:37
← erichammond left20:37
carve wouldn't the fork have it's own 'master' ?20:37
→ diegoviola joined20:37
→ errordeveloper joined20:37
brockm I'm new to git and I've been reading on git branching models and git flow...so20:38
← khaije|m` left20:38
→ khaije|m` joined20:38
carve a fork is pretty much a clone of whatever you 'fork'20:38
brockm I guess what I'm saying is, how would I support that kind of a repo structure with two projects?20:38
carve so in theory i would say it has its own master20:39
brockm yeah...20:39
← hvoigt left20:39
brockm I know that, it does have it's own master.20:39
My question is20:39
carve R access for gitolite-admin DENIED to id_dsa <- Anyone know how to resolve this?20:40
→ hvoigt joined20:40
brockm if I remote add origin [email@hidden.address]20:40
← hvoigt left20:40
brockm then I git pull origin master20:40
that means my local repo now is a copy of that repo's master branch right?20:41
carve yes20:41
brockm okay well let's say I wanted to branch to develop for MY site20:41
and commit to my project without commiting back to origin20:42
and then later, branch origin and include something from my project20:42
← codebeaker left20:42
← urkud_ left20:42
→ guns joined20:43
← bindaas left20:48
→ mcbluebeard joined20:48
← flavius left20:48
← khaije|m` left20:49
← chrisdev left20:49
mcbluebeard I did a filter-branch through sed to remove a committed password and would now like to push to a remote repository but am (obviously) getting rejected due to non-ffwd changes. How can I force the push?20:49
mmattice -f maybe20:50
git help push is probably useful20:50
but if the remote repo won't let you, you're hosed a bit20:50
→ khaije|m` joined20:50
gionnico how do you go back to some commit ?20:51
← replore_ left20:51
→ replore joined20:52
→ namemon joined20:52
← phantomcircuit left20:53
→ phantomcircuit joined20:53
← tewecske left20:53
dotCOMmie If I have a downstream and an upstream long lived branch how can I find out which branch a short lived story branch was started from?20:53
← jsick3 left20:54
→ emptyflask joined20:55
dotCOMmie merge-base takes me 1/2 way there but I cant figure out how to differentiate a story branch that was started of downstream and a story branch that was started from upstream and got merged down stream20:55
→ ShadowBelmolve joined20:55
→ tewecske joined20:55
← bartek left20:55
→ aidalgol joined20:56
← ShadowBelmolve_ left20:56
ayust dotCOMmie: compare mergebase(story,upstream) to mergebase(upstream,downstream)?20:57
← guardian left20:58
dotCOMmie ayust: that does not handle the latter case of branch which was started from upstream and got merged downstream20:58
← JEEB left20:58
← kermit left21:00
← khaije|m` left21:01
← rubydiamond left21:01
carve How do you resolve this issue: R access for gitolite-admin DENIED21:01
Using a Windows box trying to clone the gitolite initial directory for admin21:01
→ khaije|m` joined21:01
ayust dotCOMmie: if it was started upstream and merged downstream, however, that's effectively merging the portion of upstream that wasn't mergef when story was started, into downstream, and then merging the story21:02
← emptyflask left21:02
← LongBeach left21:02
→ emptyflask joined21:02
ayust so as far as git commit history goes, the two are effectively the same, no?21:02
merged*21:03
→ guardian joined21:03
← guardian left21:03
→ Schmallon joined21:04
dotCOMmie ayust: yeah. Perhaps some context would help. I want to find all branches which would be ok to merge upstream as they would not pull in any dowstream commits along with them21:04
→ sylr joined21:05
dotCOMmie and so if you branch of story from upstream then merge it to downstream its still safe to merge the story branch upstream21:05
ayust dotCOMmie: it should be okay to compare merge-base (upstream, downstream) with merge-base( upstream, story) then... as long as the latter is a parent of the former21:06
it should be safe to merge ?21:06
cbreak merging is always safe as long as you have no local uncommitted changes21:06
ayust cbreak: different definition of safe21:07
← boutcher left21:07
→ chrisdev joined21:08
wereHamster ayust: it's safe in every sense of the word21:08
← sanguisleft left21:08
← drubies left21:08
→ boutcher joined21:09
→ candybar joined21:09
← saschpe left21:09
dotCOMmie for the purposes of this issue safe is: not pull changes against the stream21:09
adamholt → adamholt_away21:09
adamholt_away → adamholt21:09
ayust ^ wereHamster21:10
← q0rban left21:10
← replore left21:10
← setmeaway left21:10
wereHamster eh? 'aginst the stream'?21:11
← boutcher left21:11
→ setmeaway joined21:11
← setmeaway left21:11
← kipras left21:12
→ setmeaway joined21:12
← sylr left21:12
dotCOMmie ayust: I think that might make sense21:12
← khaije|m` left21:12
dotCOMmie is there a nice way to see if one commit is parent of another?21:12
ayust git log sha1..sha221:13
dotCOMmie yeah thats what I was thinking d:21:13
ayust (or just git rev-list)21:13
→ ereslibre joined21:14
← ereslibre left21:14
→ ereslibre joined21:14
→ khaije|m` joined21:14
← rendez left21:14
dotCOMmie wereHamster: if you have a development branch and a maintenance branch. The maintenance branch is upstream, it is safe to merge from it to a development branch. You do not however want to pull commits from the development branch to a maintenance branch.21:15
← deekayen left21:15
← Behi left21:16
← kumbayo left21:16
wereHamster is that a question?21:16
← boriskr left21:17
dotCOMmie wereHamster: I'm explaining "against the stream"21:17
wereHamster is still not getting it21:17
← captnemofulcrum left21:18
→ bsajsb joined21:18
← bsajsb left21:19
dotCOMmie If you are developing a software and you release it. And a customer says there is this small bug, please fix it. You want to be able to provide them with a minimal patch to just address that issue vs having to provide them with a patch for that bug plus many other perhaps untested changes.21:19
→ deekayen joined21:19
← deekayen left21:19
→ deekayen joined21:19
wereHamster is still not getting where the 'stream' comes into play21:20
← deekayen left21:20
← saschagehlich_ left21:20
→ deekayen joined21:21
dotCOMmie The stream is that you can have constant merges from your maintenance branch to your development branch21:21
and you don't want the changes going from the development branch to a maintenance branch21:21
wereHamster that's common workflow. NOthing special there21:22
dotCOMmie right21:22
← afeijo left21:22
← gedevan left21:23
← spaceonline left21:23
→ sigh joined21:23
→ lolfrenz joined21:24
← khaije|m` left21:25
← ExtraSpice left21:27
← bodom left21:27
→ eletuchy joined21:28
→ khaije|m` joined21:28
← bran left21:28
→ jnylen joined21:29
jnylen Hi all... git stash pop had a merge conflict which I have resolved... now how do I finish the operation?21:29
← PabloM left21:30
→ fr0sty joined21:30
← ericdc left21:30
← grahamc left21:30
← fr0sty|mobile left21:30
thiago_home jnylen: git add21:31
← candybar left21:32
→ jhaddad joined21:33
nixness → foocraft21:33
jnylen thiago_home: normally 'git stash pop' doesn't add anything to the index21:33
i think that's why i was confused21:33
thiago_home right21:33
jnylen because now the conflicting files are added to the index, which I didn't want21:33
so 'git reset' got things back to what i expected21:33
thiago_home yeah21:34
→ candybar joined21:34
jhaddad i've got a big project at work where i've done a whole bunch of commits in a branch, and pushed up to a remote branch of the same name. is there any harm of interactively rebasing it locally before I merge back into master?21:34
mmattice if someone else is playing with that branch, it may annoy them21:34
jhaddad mmattice: no one else is using it21:34
i just didn't know if it would do anything weird with the merge history.. never tried it before.21:35
→ alecthegeek joined21:35
pantsman it won't do anything "weird"21:36
jhaddad pantsman: k, thanks21:37
mmattice you can always test it and push it somewhere safe to try it out21:37
← schweize1 left21:38
→ kwebble joined21:38
← khaije|m` left21:38
← Juo left21:39
→ pdhborges joined21:39
← joel_falcou left21:39
→ joel_falcou joined21:39
← pantsman left21:39
← larsl left21:40
→ khaije|m` joined21:43
← dyer left21:44
← rayners left21:44
→ juangiordana_ joined21:46
→ LekeFly joined21:47
← joel_falcou left21:47
← juangiordana left21:48
← binjured_ left21:49
← AkumaStreak left21:49
← aidalgol left21:50
← vsayer left21:51
← alecthegeek left21:52
← kwebble left21:53
devth curious about when appropriate to use push -f: let's say i'm working on repo w/ 1 other dev, who is currently asleep under their desk. i commit and push. then i realize i want to fix up that last commit with --amend. save to push -f here?21:53
s/save/safe21:53
albel727 yeah. as long as you're sure nobody has pulled during that time, it's ok.21:55
← khaije|m` left21:55
devth right. so push -f doesn't blow anything up except that last commmit21:55
albel727 in described situation - yes.21:55
devth ok, ty21:56
RichiH devth: in theory, if you are only two devs, you can even push -f after he pulled21:56
if you communicate clearly21:56
avdg or both devs have enough git knowledge ;-)21:56
jhaddad RichiH: how would that work21:56
devth RichiH: if he made commits, how would he cleanly pull? some rebase?21:56
→ lightcap joined21:56
devth cherry-pick?21:56
avdg rebase, merge, whatever to sync21:57
RichiH though if he pulls and pushes again without doing anything else, you will end up with both patch paths in your upstream if both series end in the same status. if not, things get conflicted21:57
← tanuki_ left21:57
RichiH devth: depends on the situation, really21:57
→ gholms|work joined21:57
→ khaije|m` joined21:58
devth well, he wouldn't just be able to do a "git pull" right? he'd have to git fetch then figure out how to merge origin/<branch> into his stuff21:58
← ereslibre left21:58
avdg the git tools are not really interesting actually, whats more interesting are whats happening with the commits and the branch pointers21:58
but that requires some git core knowledge (how git objects works)21:59
devth avdg: yeah, i've read the progit interals a few times, but it's discussions like this that really help me understand :)21:59
→ replore joined21:59
← phantomcircuit left21:59
← _psychic_ left22:00
avdg the tools are just to help you with what you want to do22:00
adamholt → adamholt_away22:00
RichiH devth: if you rebase your local repo but keep HEAD the same and then push --force, B pulls into his repo and pushes again, and you pull again, you will both have two parallel commit paths which both end in the same HEAD22:00
adamholt_away → adamholt22:00
← excess^ left22:00
RichiH and by rebase, i meant interactive rebase with squashing, reordering, picking, etc22:01
← drev1 left22:01
→ vsayer joined22:01
← opalka left22:01
avdg :p interactive is really cool in git btw ;-)22:02
gholms|work Bleh, rebasing stuff that happened prior to a subtree merge is really hard. :-\22:02
devth RichiH: ok right. but assume he pulled, i rebased -i & pushed -f, he made commits, now he wants to push his commits. how can he?22:02
gholms|work → gholms22:02
→ _psychic_ joined22:02
avdg he can always sync22:02
he just have the old commits, thats really everything22:02
→ Guest1729 joined22:02
avdg he is on an other track22:02
devth so a merge commit gets him into the same place as me22:03
but his history is different?22:03
avdg well, rebase can still get ride of that if you really wants that22:03
especially with interactive mode22:03
RichiH devth: in that case, you owe him beer22:04
→ EricInBNE joined22:04
RichiH and he will need to clone into a new repo and merge his stuff more or less manually22:04
devth before this discussion, i would have done this: B saves his stuff on another branch, reset --hard origin/master, then cherry-pick the work from that branch22:04
avdg or let him do the job :p22:04
← TrineLocke left22:05
devth lol, true, but this is all hypothetical22:05
→ Swimmin__ joined22:05
RichiH devth: there are several ways to do this22:05
← Swimming_bird left22:05
RichiH the basic guideline is: if someone else pulled changes already, you must only change history if there is a _really_ good reason to do so and only after talking to the others and getting their OK22:05
also, you will owe them beer22:06
devth hah. ok.22:06
so assume you bought the beers, you got their OK22:06
→ TrineLocke joined22:06
devth i what i set before about reset --hard the way for them to get their local histories inline with mine after i push -f?22:06
RichiH in that case, it depends22:06
devth *is22:06
← candybar left22:07
RichiH personally, i would most likely just clone into a new repo, merge all changes i want, and get rid of the old repo after making sure everything is OK. it's a bit of a ham-fisted approach, but it works22:08
← khaije|m` left22:08
devth ah, ok22:08
← TrineLocke left22:08
← Schmallon left22:08
→ candybar joined22:09
→ mtx_init joined22:09
→ alecthegeek joined22:10
← ThomasLocke left22:10
← pdelgallego left22:10
← pdelgallego_ left22:10
← PerlPilot left22:10
→ PerlJam joined22:11
← tty1 left22:11
hacim gah, what is wrong with this push22:11
avdg ?22:11
hacim i do a git push, and it does this: http://paste.debian.net/115416/22:11
and then it never finishes22:12
elliottcable → ec|detached22:12
avdg patient?22:12
hacim i waited 30 minutes22:12
avdg hmm22:12
hacim and I'm sitting next to the machine and it is underloaded22:12
avdg looks like a very slow server or slow connection22:13
→ sylr joined22:13
← gionnico left22:13
hacim i tried git gc --aggressive and git fsck on both server and local repo22:13
→ tty1 joined22:13
hacim the load is 0.1 on the server, and i'm on the local lan where ssh interactivity in the shell is mighty fast22:13
← tgunr left22:13
avdg maybe permission22:13
→ gionnico joined22:14
avdg but nothing to be sure22:14
what protocol are you using?22:14
→ khaije|m` joined22:15
← jbrokc left22:15
← berserkr left22:15
← resmo left22:15
← tgunr_ left22:15
→ jbrokc joined22:16
← fr0sty left22:16
hacim avdg: git://22:17
← sylr left22:17
← drizzd left22:17
gholms Pushing over git://. Interesting.22:17
speakman hm! adding a git pre-commit script makes "reword" when "git rebase -i" not work!22:17
← tty1 left22:18
avdg hacim: did you tried over ssh?22:18
→ tty2 joined22:18
avdg knows he's too inexperienced about the git protocol22:19
← gionnico left22:20
← gfto left22:22
← tvw left22:22
wagle_ → wagle22:22
→ syphar joined22:23
← alex__c2022 left22:23
→ mklappstuhl joined22:24
← Heimidal left22:25
← eijk left22:25
← mheusser left22:25
→ rcreasey joined22:26
rcreasey I need some help removing some commits an over-zealous developer made and pushed to our repo22:26
→ jophish_ joined22:26
rcreasey I basically want to remove all changes he's made22:26
← khaije|m` left22:26
avdg git rebase -i ?22:26
gholms git revert?22:27
thiago_home rcreasey: git revert22:27
← syphar left22:27
thiago_home it will apply the reverse of the patches22:27
← psoo left22:27
rcreasey revert just unwinds the changes22:27
there are a couple i'm going to need to keep that other people have made22:27
so i'm wanting to rebase?22:27
thiago_home no22:27
you want to revert22:27
gholms git revert works on a commit-by-commit basis.22:27
rcreasey oh, cool22:28
thiago_home never rebase a branch that people work on22:28
← skoop left22:28
← ZeroCoder left22:28
rcreasey so git revert, how do i tell it to go back to the commit I want?22:29
say HEAD~1422:29
thiago_home revert all commits in-between22:29
alternatively: git checkout HEAD~14 .; git commit22:29
← sigh left22:29
rcreasey ah, so it's just clipping those commits22:29
thiago_home clipping?22:30
rcreasey so if i did git revert HEAD~1422:30
→ pdelgallego joined22:30
gholms That would undo one commit.22:30
→ pdelgallego_ joined22:30
rcreasey just HEAD~1422:30
gotcha.22:30
→ alberto56 joined22:30
rcreasey and uh, dumb question22:30
→ whyking_ joined22:30
rcreasey revision hashes22:30
what's the easiest way to get the shortform?22:30
→ resmo joined22:31
thiago_home for what? For an arbitrary commit?22:31
gholms Keep chopping off characters until git says it's too ambiguous.22:31
rcreasey ah, ok22:31
← n8o-mba left22:32
→ simplexity joined22:32
← simplexity left22:32
→ simplexity joined22:32
← simplexity left22:33
← tearsong left22:34
← kingfishr__ left22:34
→ welly joined22:34
rcreasey so there's like 10 commits this guy did that I need to remove22:35
i should still be git reverting?22:36
yaotti_aw → yaotti22:36
rcreasey but should i start with the most recent commit and work backward in time?22:36
gholms Yes22:36
← v0n left22:36
→ khaije|m` joined22:36
gholms gives up on his rebase22:37
rcreasey does it work on merge commits?22:37
→ bonhoffer joined22:37
→ n8o-mba joined22:37
→ f00li5h joined22:37
gholms Reverting a merge is... umm...22:38
gholms lets someone more knowledgeable take this one22:38
rcreasey i already want to stab this guy in the eyes22:38
← ymasory left22:38
gholms rcreasey: Some light reading on that: http://www.kernel.org/pub/software/scm/git/docs/howto/revert-a-faulty-merge.txt22:38
→ sylr joined22:39
← bonhoffer left22:39
→ joel_falcou joined22:40
→ dpy joined22:40
← __sorin__ left22:40
← necromancer left22:41
← vanmik left22:41
→ phantomcircuit joined22:41
← glcrazy left22:42
← pdhborges left22:43
← ngng left22:46
← sylr left22:46
→ ngng joined22:46
← khaije|m` left22:47
← giallu left22:48
← cyphase left22:48
→ angelsl joined22:49
→ gadams joined22:50
→ mkanat joined22:52
→ jds joined22:52
→ khaije|m` joined22:53
← jophish_ left22:53
yaotti → yaotti_aw22:53
→ Octalot joined22:54
← LekeFly left22:54
→ ereslibre joined22:54
lolfrenz can I merge two commits into one?22:55
thiago_home that's not what merge means22:55
you want to squash two commits together22:55
have you pushed them somwhere where other people use them?22:55
lolfrenz nope22:55
thiago_home then git rebase -i $SUITABLE_COMMIT22:56
the suitable commit may be one commit before the oldest of the two22:56
← Samual left22:57
← mdel left22:57
← cesario left22:57
← mkanat left22:58
→ JEEB joined22:58
← n8o-mba left22:58
← cannonball left22:58
← _psychic_ left22:59
rhl hi, I have an svn repository in which I have a directory with read/write access (otherwise I just have read access), and I am trying to use git svn to pull just the folder, how do I do that?22:59
← mklappstuhl left22:59
← DrNick left22:59
→ cyphase joined23:00
← cyphase left23:00
→ cyphase joined23:00
→ DrNick joined23:00
← replore left23:00
bla have you tried passing it just the necessary path to git svn clone?23:01
If it won't work you can give it a relative path to trunk directory23:01
gholms I tried that once. It backed out to the top level of the repo and began parsing commits.23:01
All 1.1 million of them. >_<23:01
← ch077179 left23:01
rhl I get something like: svn-remote.svn.url already set: <url> wanted to set to <url>23:02
← linusoleander left23:02
→ dyer joined23:03
← dyer left23:03
→ dyer joined23:03
→ dnorman joined23:04
rhl bla: specifically: http://fpaste.org/rwPm/23:04
← dnorman left23:04
→ dnorman joined23:04
← namemon left23:04
← dnorman left23:04
← khaije|m` left23:05
rhl I have been able to check out it's parent directory, but i'm not sure if I commit back with changes in the subdirectory what will happen23:05
→ replore_ joined23:05
← Amorphous left23:05
rhl bla, ah, I see, nm, it creates the directory on failure, if I delete it, it works the second time through23:06
← chitragupt left23:06
→ linusoleander joined23:06
← ereslibre left23:07
← alecthegeek left23:07
← workmad3 left23:07
← deekayen left23:07
← ngng left23:07
← linusoleander left23:07
→ ngng joined23:08
→ linusoleander joined23:08
← milki_ left23:09
→ milki joined23:09
milki how can i delete branches that were made with 'git pull origin :refs/blah' where blah is not one of heads/tags/remotes via git23:09
its says new branch, but i cant reference the branch23:10
← phantomcircuit left23:11
→ ymasory joined23:11
gadams In gitolite how do I modify where the default repositories go?23:11
← alip left23:12
→ khaije|m` joined23:12
dropdrive Is magit being maintained? I see very little (no?) activity on it in the past year. I'm hesistant to start using something that has stagnated.23:12
← kenichi_ left23:13
milki gadams: its in the conf file for gitolite23:13
bremner_ dropdrive: what repo?23:13
milki gadams: when you install it23:14
→ smolyn_ joined23:14
→ JustinKSU joined23:14
gadams milki, the ENV{HOME} . "/projects.list", that variable?23:15
← mm_202 left23:15
← deuterium left23:15
milki shouldnt there be coments for it23:15
JustinKSU Help: I have used Eclipse to commit my code change in my GitHub Fork, but I don't see the change on github.org. Am I doing something wrong?23:15
← dr_win left23:15
milki i dont think its that one23:15
← smolyn left23:15
milki the default is /repositories23:15
→ drubies joined23:16
→ alip joined23:16
→ phantomcircuit joined23:16
milki $REPO_BASE23:16
gadams: ^23:16
though it says not to change it23:17
:P23:17
gadams milki, I just found that too. Thank you sir. I'm going to go facepalm myself23:17
milki make sure to migrate the admin repo properly23:17
i guess23:17
gadams I'm doing a clean install23:17
→ nakkor joined23:17
milki o ok23:18
→ Mark_Morcom joined23:18
→ bonhoffer joined23:18
milki is there a way to pull/clone refs from /refs that arent heads/tags/remotes!?23:18
← simosx left23:19
milki cant seem to figure out how to reference them -.-23:19
thiago_home yes23:19
git fetch reponame $REFNAME23:19
milki cant find remote ref23:20
thiago_home then you typed it wrong23:20
git ls-remote reponame to list all available refs23:20
all that exist23:20
milki well, i guess its already fetched because i can do git show SHA23:20
→ _psychic_ joined23:20
milki but its not under .git/refs23:20
← _psychic_ left23:20
thiago_home do you see the ref in the output of ls-remote?23:20
milki yes23:21
thiago_home what is it called?23:21
milki oo23:21
i need pull name23:21
s/pull/full23:21
from /23:21
ok23:21
thiago_home yes23:21
← nimrod10 left23:21
rcreasey bleh23:22
okay, this repo is totally fucked up23:22
milki how can i delete branches that were made with 'git pull origin :refs/blah' where blah is not one of heads/tags/remotes via git23:22
→ Amorphous joined23:22
milki it didnt really create a branch but it shows up in show-ref23:22
rcreasey is there anyway for me to just delete the last 15 commits and reset the repo to that point in time?23:22
← khaije|m` left23:22
milki rcreasey: git rebase --hard23:22
and uh23:23
→ kermit joined23:23
milki HEAD^15?23:23
but its permanent23:23
so...be careful...23:23
rcreasey good riddance23:23
milki or find the sha-1 hash of the commit you wnat to roll back to23:23
rcreasey HEAD~15?23:23
milki just to be safe23:23
rcreasey yeah23:23
i have that.23:23
ec|detached → elliottcable23:23
→ Samual joined23:24
gadams okay when I change $REPO_BASE it still attempts to /home/git/srv/git/repositories instead of /srv/git/repositories23:24
← javaanse_jongens left23:24
← Ryback_ left23:24
← bonhoffer left23:27
← gnufied left23:27
→ Textmode joined23:27
← JustinKSU left23:27
← ShadowBelmolve left23:28
→ khaije|m` joined23:29
← Swimmin__ left23:29
→ tao_ joined23:30
← jds left23:30
← replore_ left23:32
dropdrive bremner_: I couldn't find any that were recently updated.23:33
bremner_: (magit, that is.)23:33
bremner_ dropdrive: there's plenty of activity on the mailing list. Did you look on github?23:33
aroop → aroop|away23:33
bremner_ oh, it's just being reorganized on github. hmm.23:34
→ binjured_ joined23:34
dropdrive bremner_: Oh, which repo should I be using?23:34
← binjured_ left23:34
→ paolodedios joined23:35
dropdrive bremner_: (I couldn't find philjackson's.)23:35
→ javaanse_jongens joined23:35
bremner_ https://github.com/magit/magit23:35
← dpy left23:36
→ jophish_ joined23:36
→ tgunr joined23:36
← icwiener_ left23:37
→ jrots joined23:37
dropdrive bremner_: Thanks.23:37
← Chillance left23:38
bremner_ dropdrive: welcome!23:38
→ bonhoffer joined23:38
shadowmaster → Shadowmbit23:38
bremner_ I think google hasn't caught up with the re-org yet, so I can't even yell at you ;)23:38
gadams Has anyone ever changed $REPO_BASE when installtion gitolite?23:39
← beatak left23:39
← guns left23:39
← LiamH left23:40
← khaije|m` left23:41
← Christie left23:41
→ Christie joined23:41
← jophish_ left23:41
rcreasey okay, here's another problem i have. the same overzealous developer went and pushed his changes on to the master branch on a repo using the gitflow model (master/develop/features branches)23:42
I've cleaned up everyting but the master branch23:42
and master is currently sitting at 29 behind / 12 ahead of develop23:43
those 12 are obviously the ones that this guy committed23:43
how do I remove just those commits?23:43
→ drev1 joined23:43
devth rcreasey: sounds like you should merge master into develop to get those 12 commits in the right spot23:45
rcreasey: then you could reset master back to where it should be (HEAD~12)23:45
→ ShadowBelmolve joined23:46
← ShadowBelmolve left23:46
← thiago_home left23:46
← Christie left23:47
→ drev1_ joined23:47
← drev1 left23:48
gholms ^ That is, of course, if the commits merely went into the wrong branch.23:48
→ khaije|m` joined23:50
rcreasey no, i want them totally removed.23:50
← tao_ left23:50
← cjz left23:50
rcreasey also, when doing git reverts23:50
can i just do23:50
→ drev1 joined23:50
rcreasey git revert A; git revert B; git revert C ?23:50
or do i need to do something inbetween each?23:50
aroop|away → aroop23:51
← ormaaj left23:51
devth revert undoes commits by creating new commits23:51
rcreasey ah23:51
devth do you want to maintain that history, or do you want to completely abandon those commits?23:51
rcreasey well, this is being done on master23:51
gholms recommends maintaining that history23:52
rcreasey so i'd rather them just disappear entirely23:52
← drev1_ left23:52
rcreasey because they need to come as feature branch merges onto develop23:52
gholms Unless no one has fetched, of course. But if anyone has fetched those commits they'll appear again in the future.23:52
rcreasey (and then merges from develop onto master)23:52
← Yuuhi left23:53
rcreasey well, if i maintain the history of those commits (and thusly the revert commits to undo them), doesn't that break the fast forward that master would have in relation to develop?23:53
gholms Move develop forward after you revert it in master.23:53
devth if he hasn't pushed yet, i'd go with the cleaner history and use reset23:54
but if he pushed, use gholms's approach23:54
gholms Sure, if he hasn't pushed.23:54
rcreasey yeah, this is already up on github23:54
which is why i'm having to deal with this23:54
gholms Since you'll have canceled out the content of the extra commits everyone else should be able to merge back into develop and then master.23:54
Shadowmbit → shadowmaster23:54
← ymasory left23:55
rcreasey okay, so just so i'm clear23:55
← tjholowaychuk left23:55
rcreasey do a bunch of git revert -n <each of his revisions> on master23:55
then merge in develop to master23:55
gholms recommends git revert -n master~N..master23:55
gholms (where N is a number)23:55
master will be ahead of develop.23:56
rcreasey but master is currently sitting 29 behind / 12 ahead of develop23:56
gholms So after you revert the changes in master you will need to merge master into develop.23:56
devth develop will pull in those 12+1 commits though23:57
when you merge23:57
gholms Yeah, and when you're done master will be 20 behind / 0 ahead of develop.23:57
29, rather23:57
rcreasey you mean 30?23:57
29+123:57
?23:57
gholms Yep. ;)23:58
← nvardar left23:58
→ jophish_ joined23:58
gholms Because you'll be merging master into develop the things that are only in develop won't go back into master just yet.23:58
devth gl w/ that, i've gotta run :)23:58
← devth left23:59

Logs Search ←Prev date Next date→ Channels Documentation