IRCloggy #git 2011-05-22

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.

2011-05-22

→ __name__ joined00:02
← pastie left00:02
→ pastie joined00:03
→ candybar joined00:03
← gozoner left00:04
← bpierre left00:07
→ ymasory joined00:08
← psoo left00:09
→ joel_falcou joined00:09
← joel_falcou left00:09
← brian_g left00:10
← DoctorDarren left00:13
→ DoctorDarren joined00:13
→ gozoner joined00:15
← timemachine3030 left00:16
← kassah left00:17
← gozoner left00:17
← Dave^| left00:17
→ kassah joined00:17
← Corren left00:18
← warthog9 left00:20
← lanthan__ left00:21
AAA_awright Like if you press00:21
→ smw joined00:23
← towski left00:26
→ jonatasnona joined00:26
→ gbacon_ joined00:41
→ dankest joined00:44
→ mysphyt joined00:45
dankest Is there a way to completely overwrite a github repository with a push? Like a push force?00:45
My peer made a commit to the README before I did my initla project push00:46
and I can't seem to pull either00:46
→ Buglouse joined00:47
← theoretical left00:48
← _Qman left00:48
→ _Qman joined00:48
→ timemachine3030 joined00:49
ayust dankest: "git push --force" ?00:50
← socketwiz left00:53
← nicoulaj left00:53
→ linusoleander joined00:53
→ Tac_Home joined00:53
→ marienz_ joined00:54
← SimonNa left00:57
← timemachine3030 left00:58
← [intra]lanman left00:59
← marienz left01:01
→ timemachine3030 joined01:02
← kukks left01:03
← simosx left01:03
← timemachine3030 left01:03
→ timemachine3030 joined01:04
← kpreid left01:05
← Ivoz_ left01:08
← ajt_ left01:08
← __name__ left01:10
rhalff dankest: i would say if you can't push and commit your peer even when using force there ain't no use in pull'n it either.01:10
← vdv left01:11
← Jippi left01:11
→ Vampire0_ joined01:14
→ ISF joined01:15
← khmarbaise left01:16
← Vampire0 left01:18
tkil hm. is the timestamp after author in gitk the "author timestamp"? because if so, it looks like git svn dcommit clobbers that timestamp as well.01:18
→ goldins73 joined01:19
← harrisonk left01:20
← rhalff left01:23
← electrogeek left01:24
← Tianon left01:27
← goldins73 left01:28
← lamont left01:28
← Textmode left01:32
→ lamont joined01:34
← kst left01:35
→ electrogeek joined01:35
← electrogeek left01:35
→ electrogeek joined01:35
dankest thanks01:36
→ Tianon joined01:36
← Tianon left01:36
→ Tianon joined01:36
← mysphyt left01:42
← DoctorDarren left01:42
← mrtazz left01:43
← fishey left01:43
→ mrtazz joined01:43
← hackez_work left01:43
→ hackez_work joined01:43
→ fishey joined01:44
→ rcu joined01:46
← rcu left01:47
→ shiba_yu36 joined01:48
→ psoo joined01:58
← afief left02:00
← Yuuhi` left02:01
← nwp_ left02:02
← tkil left02:03
→ Yuuhi` joined02:03
← linusoleander left02:07
← demi` left02:09
← khiroc left02:09
← berserkr left02:09
→ mikepack joined02:09
→ dys` joined02:10
← dys left02:12
← dankest left02:13
→ recharge joined02:14
→ joel_falcou joined02:15
→ Murr joined02:17
→ nils_r joined02:17
← Leemp2 left02:17
→ angasule joined02:18
ralc sometimes i wanna make sweet sweet love to the inventors of git02:18
ChanServ set mode: +v02:19
← jonatasnona left02:20
→ Murr__ joined02:21
DrNick ralc:02:21
he's married02:21
and his wife is a world karate champion02:22
ralc sounds like the setup of an awesome evening02:22
← Murr left02:23
← Anubisss left02:23
← Murr__ left02:23
→ jonatasnona joined02:25
→ linusoleander joined02:25
← recharge left02:25
→ Murr joined02:27
← TDT left02:30
→ demi` joined02:31
← nils_r left02:32
→ cylence joined02:33
← Murr left02:35
→ tansell-laptop joined02:35
→ Murr joined02:35
← bitkiller left02:39
← Murr left02:41
← mikepack left02:44
→ recharge joined02:45
← jjuran left02:45
← Bass10 left02:46
→ mikepack joined02:46
→ hakunin joined02:47
→ intripoon_ joined02:48
← recharge left02:49
pgib mmm karate champion02:50
← mikepack left02:50
→ Adaptee joined02:51
→ nealsmith joined02:51
← nealsmith left02:51
← intripoon left02:51
→ ira_gershwin joined02:52
→ nardev joined02:55
← peterhil left02:56
milki >.>02:58
← linusoleander left02:59
→ Murr joined02:59
← tansell-laptop left03:01
← diegoviola left03:02
← devth left03:05
→ devth joined03:05
← alip left03:05
ChanServ set mode: +v03:05
← jonatasnona left03:05
→ jonatasnona joined03:06
→ alip joined03:09
→ bdiu joined03:09
← charon left03:11
→ warthog9 joined03:12
← BeerSerc left03:12
← hakunin left03:14
→ BeerSerc joined03:15
→ charon joined03:15
→ dl9pf_ joined03:15
→ alester joined03:15
← dl9pf left03:16
← ira_gershwin left03:17
← pastie left03:20
← gbacon_ left03:22
→ pastie joined03:24
← pyroscope left03:28
← german__ left03:31
← pizza_ left03:33
→ eydaimon joined03:33
eydaimon how can I get the current commit rev? (aside from git log)03:34
← devth left03:34
→ pizza_ joined03:34
← aef left03:34
milki git show-ref?03:34
git reflog?03:35
eydaimon thanks03:35
→ aef joined03:37
eydaimon how can I check what the ref is for the remote?03:40
nm03:40
duh03:40
← angasule left03:41
eydaimon hm03:41
← ki__ left03:42
→ german__ joined03:43
← bdiu left03:43
← ralc left03:45
→ ralc joined03:45
→ symbole joined03:45
← ralc left03:45
→ zhengyang joined03:45
→ seivan joined03:45
→ thiago_muc joined03:48
eydaimon guess a git fetch is required first03:48
is there a way to see the remote ref without first doing that?03:48
← warthog9 left03:49
→ bdiu joined03:51
→ pyroscope joined03:51
← EnTeQuAk left03:54
← aef left03:55
← psoo left03:55
→ aef joined03:56
→ bindaas joined03:57
→ EnTeQuAk joined03:57
→ ralc joined03:58
← int3__ left03:59
← OOPMan left04:03
→ leo2007 joined04:04
dys` → dys04:07
← JimmyRcom left04:08
→ JimmyRcom joined04:08
→ int3__ joined04:10
← ralc left04:11
← g0bl1n left04:13
← Murr left04:13
→ xiangfu joined04:15
← Opettaja left04:20
→ towski joined04:20
→ warthog9 joined04:21
← int3__ left04:22
→ [Pi] joined04:23
→ OOPMan joined04:23
[Pi] I uploaded a project onto GH, first version working. then my next three commits were horribly broken -- I was just loading problem code up so that other people could download it and help fix. anyway, now I have a working version #2. is there some way to tidy up ie remove these three commits, and then commit my second version?04:25
→ int3__ joined04:26
← madsy left04:26
← bdiu left04:28
joel_falcou [Pi]: you shold have worked in a branch04:29
jpalmer joel_falcou: even in a branch, you can have bad commits. (including when you merge the branch back into another)04:31
→ kpreid joined04:31
joel_falcou jpalmer: sure04:31
[Pi] joel_falcou: ok I will read up on that before I do it again.... in the meanwhile, how do I get rid of these last 3 commits?04:31
joel_falcou but havign bad commit in a dev branch that get merged when everythign is fixed is prolly better04:31
rewriting history make me cringe04:31
but YMMV04:31
[Pi]: why ?04:31
the stuff is fixed04:31
why bother04:31
[Pi] joel_falcou: tidiness04:32
jpalmer [Pi]: my peronal thought is, leave the bad commits too. the whole point of revision control is so you have... revisions. good or bad.04:32
joel_falcou yup exactly04:32
[Pi] meh, there is absolutely no benefit in this case04:32
jpalmer and there is absolutely no harm in it either.04:32
joel_falcou [Pi]: you really dont care04:32
[Pi] it is just extra mass. I would rather have each commit actually mean something04:32
joel_falcou it will04:33
[Pi] ^mess04:33
joel_falcou broken commit ahppens to all of us04:33
← fishey left04:33
[Pi] ok so if I am going to commit broken stuff then I create a branch, say someone fixes it on the branch, then I do some more fiddling my end and get to commit a new commit for the main tree04:34
that's what I should have done yes?04:34
and then I can delete the branch?04:34
joel_falcou http://nvie.com/posts/a-successful-git-branching-model/04:35
yup04:35
or let it live for reference04:35
[Pi] I am not in favour of leaving behind a trace of all the mess of my development process04:35
unless there is good reason04:35
jpalmer [Pi]: generally speaking, I use branches for any work that is new (features, bug fixes, etc) and I commit the code in that branch if it's going to live more than 1 day. when I've got the bug fixed, or feature added (and working) I merge it back into master, and delete the branch.04:35
[Pi] even if I do this, I would like to do it selectively, otherwise I will be inundated -- the essential will get lost amongst the mundane04:36
jpalmer: oh good, so deleting branches is at least standard practice04:36
so is there anyway to tidy up my main tree? I have 3 messed up commits... how do I get rid of them? I wouldn't mind even erasing everything and just committing what I have now04:37
jpalmer I don't know about 'standard' practice, but it's what my teams do. when the work in a branch is done and merged back in.. why keep the branch?04:37
joel_falcou i agree04:38
[Pi] yeah keep things tidy04:38
jpalmer [Pi]: deleting everything, and commiting now.. *again* defeats the purpose of revision control.04:38
[Pi] that's what I'm trying to do early on with this project04:38
jpalmer [Pi]: if you don't want to have revisions, then why use revision control at all?04:38
[Pi] jpalmer: at the moment I don't need it... it is a young project... but as it grows in size I will need revisions for sure04:39
so I'm just trying to get everything started on the right foot04:39
jpalmer even a "bad commit" can be useful. 2 years from now, you may need that commit to figure out "what was I thinking when I wrote that?!?" the code can help, the commit message can help.. leave it alone.04:39
[Pi] no. There is no benefit. It is a mess. I want my commit tree to be clean04:40
I don't know why you're trying to persuade me to keep junk04:40
is it because it is difficult / voodoo to remove recent commits?04:40
jpalmer ehh, you'll learn eventually I guess. it's your repo. do what you want with it.04:40
We were just trying to save you the headache and heartache of learning this on your own. but if you're going to ignore us.. you'll get smarter at some point. good luck.04:41
joel_falcou [Pi]: history rewriting is really more meesy than a few botched commit really04:41
[Pi] joel_falcou: thx04:43
jpalmer [Pi]: as for branching, generally speaking, it's not to "hide" or "get rid of" commits with non-functional code. it's more about keeping 'master' working a majority of the time. In fact, when merging your branch back into master, it keeps the commit history.04:43
→ ilteris joined04:44
← 13WAARIOC left04:44
[Pi] jpalmer: right, I didn't know about branching when I did this. My bad for not rtfm.04:45
→ mikepack joined04:46
jpalmer I'm not trying to flame you for not knowing something. I'm trying to help you realize that getting rid of history is generally not a good thing to do. Especially if you don't have a *really really really* good reason to do so.04:46
← mcdebugger left04:46
jpalmer (and to be blunt, I can honeslty say I've never heard of a reason "good enough" to change history)04:46
nevyn jpalmer: a linear history04:47
jpalmer: being tidy.04:47
jpalmer heck, look at what happened when history got changed in all of the 'back to the future' movies. *nothing* good comes of it :P04:47
joel_falcou beign tidy is non sense w/r to RevControl04:47
really04:47
jpalmer nevyn: neither of which are *good* reasons, IMO.04:47
joel_falcou and error are part of the dev process and contains informations04:48
nevyn joel_falcou: explain that to the change managment board.04:48
joel_falcou ?04:48
nevyn: i dont get you04:48
nevyn change control want to know exactly what's going to be done in production. they don't want the dev missteps04:48
it just confuses them.04:48
jpalmer change management boards are reviewing your source code?04:49
joel_falcou hence branches nevyn04:49
and what the f*** is CMB >04:49
nevyn jpalmer: infrastructure code.04:49
joel_falcou is a poor academia guy so bear with him :p04:49
joel_falcou again04:49
fuck up in branch04:49
merge to single point history04:49
delete branch04:49
this is better than rewrite of history04:49
Arrowmaster except it removes the flow04:50
[Pi] sorry, I can't see it. I'm not deliberately trying to be Obtuse. but I have a good initial commit, a decent 1.0, followed by three 'wtf why is this not working' commits, which are completely worthless because I didn't know how to set up GL frame buffers, they are just pure floundering. now I have fixed everything and I have good 2.0. there is absolutely no benefit of keeping the flounderings, other than to remind myself that I can occasionally be an idio04:50
I'm well aware of the fact, so I don't need a reminder.04:50
joel_falcou Arrowmaster: depends of your workflow04:50
jpalmer joel_falcou: in the enterprise world, changes to infrastructure have to go through a review process. people who make sure the change isn't going to negatively impact other things.04:50
joel_falcou jpalmer: ok04:50
Arrowmaster: the branch is still there in the history04:50
just no pushed anymore04:50
the flow is still visible04:50
symbole Sometimes QA folk read commit messages to get a sense of what's going on.04:51
Arrowmaster if its all in one commit then the flow of what happened to it is gone04:51
joel_falcou Arrowmaster: no04:51
you have the branch with all the change just behind04:51
jpalmer [Pi]: the code is fixed in the recent commit, right? 2 year from now when you need to fix a bug in framebuffers, and haven't DONE it for 2 years.. those commits can and will be useful.04:51
joel_falcou the merge point is just there and have two parent04:51
the previous linear commit04:51
the branch that lead there04:51
Arrowmaster are you saying to git merge --no-ff or git merge --squash?04:51
joel_falcou git merge --no-ff04:52
i thought it was obvious04:52
Arrowmaster that can still leave cruft that shouldnt be there04:52
← mdel left04:52
joel_falcou sorry for the context lose04:52
Arrowmaster: like what kind of cruft ?04:52
you merge whwent the stuff works04:52
period04:52
jpalmer [Pi]: if you're that concerned about "tidy" and don't really need the revision history.. it's your repo. wipe and re-up.04:52
Arrowmaster like commits that fix previous commits in the branch04:53
← f0i left04:53
joel_falcou Arrowmaster: again i prefer this cruft to rewriting public history04:53
and again YMMV04:53
→ f0i joined04:53
jpalmer the bottom line for *everyone* is, these are your repos. o what you feel is best for you. I personally would *never* change history.04:54
[Pi] jpalmer: Wrong code doesn't help anyone. without seeing my code, I don't see why you are going out on a limb to suggest it may be useful in the future. It won't. I can show it to you if you like, and you will have to agree. There is no benefit of it. nothing to be learned.04:54
→ jjuran joined04:54
joel_falcou [Pi]: i guess we all understood, it's your repo, do w/e you want04:54
no need to go in logohrea on this04:55
jpalmer [Pi]: again, you'll learn in due time. a simple commit (even a bad one) can help jog your memory 2 years from now when you have to go back and work on that code. you'll have a better understanding of why you did something the way you did. but to each his own. wipe and re-up.04:55
just hope you don't regret your decision later. because when you change history, it's.. gone.04:56
→ mdel joined04:56
jpalmer Leaving it there really doesn't hurt anything. and given the possibilty that it may HELP you later.. I dunno.. I just don't see it being a good thing.04:56
but I'll stop my rant, it's past my bedtime anyway :P night all.04:57
← JasonWoof left04:58
[Pi] joel_falcou: 'do w/e you want' <-- this is why I'm here asking; I'm trying to find what are the alternatives, what can I do, what are the consequences / reasons04:59
joel_falcou we told you everythign you had to know04:59
rewriting history is a bad practice04:59
← cylence left05:00
← deekayen left05:00
← mdel left05:00
→ Opettaja joined05:01
→ gremset joined05:01
[Pi] ok seeing as I only have one good commit, I can download it to my desktop, then erase everything in the repo, then commit the good original, then commit my good v2. then read up on branches, so I don't mess up like this again...05:01
how do I errase everything in my repo?05:01
jpalmer delete the repo, and recreate.05:02
(I find it interesting that.. even after all this "don't do it" stuff, you still wan tto do it. but I digress.05:02
→ rchavik joined05:03
← rchavik left05:03
→ rchavik joined05:03
→ deekayen joined05:04
→ tansell-laptop joined05:05
[Pi] is there any way I can delete all of the repo's contents from the command line without actually going onto the site and deleting the repo itself ( which would delete the actual container rather than just the contents )?05:06
symbole I have branch A with file X that consists of "foo", and I also have branch B with file X that consists of "bar". When I merge B into A, I don't get a conflict. I don't quite understand it why.05:06
← 16SAAGMKN left05:07
→ paja joined05:09
← alip left05:10
← Adaptee left05:10
← banghouse left05:11
→ alip joined05:13
→ cylence joined05:16
→ dr0id joined05:17
← Tac_Home left05:17
← pizza_ left05:19
← alip left05:19
→ Error404NotFound joined05:19
symbole It's a fast-forward merge too.05:20
I guess I need to look at how that works more closely.05:20
→ pizza_ joined05:21
dr0id → Andr0id05:22
→ alip joined05:24
← alester left05:25
→ tewecske joined05:26
→ LongBeach joined05:26
→ KSHawkEye joined05:26
KSHawkEye Hey, wondering what the --depth <depth> actually did, and since it was removed what the alternates are?05:27
→ boombatower joined05:27
← boombatower left05:27
→ boombatower joined05:27
→ _Ryth joined05:30
ChanServ set mode: +v05:30
KSHawkEye (I'm guessing that it only clones the main directory and no subdirectories, but since it was removed how can I do this?)05:30
Arrowmaster thats not what it did05:30
KSHawkEye Oh.. what did it do?05:31
Arrowmaster it limited the amount of history that it cloned05:31
KSHawkEye Okay, so is there any way to do that then? just clone a single file, or a single dir?05:31
Arrowmaster no05:31
KSHawkEye Arrowmaster: You positive?05:31
Arrowmaster 100%05:32
KSHawkEye Thanks05:32
← joel_falcou left05:33
→ envi joined05:33
→ mdel joined05:34
← mikepack left05:37
_Ryth I just asked another dev on this project about git "hey, I've made a rather large change to hamaxe.cs How do i Keep my changes and merge your newest commit into mine?" Is there a way05:38
Sorry for poor formatting there ^05:38
→ Anatoliy joined05:39
→ mikepack joined05:40
← Anatoliy left05:41
← eydaimon left05:42
Arrowmaster KSHawkEye: were did you get the impression that --depth was removed from git clone?05:44
KSHawkEye Someone recommended it to me05:45
oh05:45
sorry05:45
← ISF left05:45
KSHawkEye read that wrong, and because I tried --depth and it failed, I also couldn't find it in the help or anything05:46
Arrowmaster then your git version must be old enough to not have it05:46
KSHawkEye It's ubuntu 11?05:46
git --help | grep "depth" doesn't reveal anything, and man git doesn't list it05:49
jast *automatic message* the 'git' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git.html05:49
Arrowmaster KSHawkEye: should be in man git-clone05:49
jast KSHawkEye: *automatic message* the 'git-clone' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-clone.html05:49
KSHawkEye Arrowmaster: Ah, thanks, didn't know there was another man for git-clone, thanks05:50
Arrowmaster yeah theres a different man page for all subcommands05:50
← jonatasnona left05:51
← Ruudjah left05:53
← seivan left05:57
← jfkw left06:03
← mikepack left06:03
← chrisf left06:05
← f00li5h left06:05
→ chrisf joined06:06
← mdel left06:08
→ adaptee joined06:09
← KSHawkEye left06:12
← halfline left06:14
→ cesario joined06:15
← RaceCondition left06:17
→ gozoner joined06:18
← gozoner left06:22
→ Tac_Home joined06:23
← adamm left06:25
← eli left06:25
→ seivan joined06:26
← symbole left06:27
→ jackdempsey joined06:28
→ phantomcircuit joined06:28
← jackdempsey left06:28
phantomcircuit so i have a shared bare git repo, how can i keep the permissions from getting clobbered by pushes from different people?06:29
→ eli joined06:29
← hyperair left06:29
→ namemon joined06:30
tango_ phantomcircuit: did you init it shared?06:31
phantomcircuit nope is there some way to fix that?06:32
sitaram phantomcircuit: http://sitaramc.github.com/gitolite/doc/overkill.html has some information that should be useful. It's not hard to do but you need to understand how unix permissions work06:35
→ aspotashev joined06:39
← madewokherd left06:44
→ legumbre_ joined06:46
← warthog9 left06:47
← legumbre left06:47
← Rickardo_ left06:48
→ rgr joined06:49
→ akosikeno joined06:49
← Tac_Home left06:49
→ Anatoliy joined06:54
← Anatoliy left06:55
← Wooga left06:56
→ Tac_Home joined06:57
→ ExtraSpice joined06:58
← gremset left06:59
→ Wooga joined06:59
← alip left07:00
← Wooga left07:02
→ Wooga joined07:03
→ alip joined07:04
← erez_ left07:05
→ Titosemi joined07:06
← rgr left07:07
→ rgr joined07:08
← Anxi80 left07:08
← Andr0id left07:08
→ Anxi80 joined07:08
→ vdv joined07:09
→ airborn joined07:09
← jfelchner left07:10
→ jfelchner joined07:10
→ nils_r joined07:11
← dhruvasagar left07:11
← Tac_Home left07:12
→ halfline joined07:13
→ dhruvasagar joined07:13
→ Tac_Home joined07:13
→ lxsameer joined07:16
lxsameer how can i clone a specific branch of a repository07:17
→ tore__ joined07:18
selckin can't07:20
← towski left07:21
lxsameer selckin: thsnkd07:21
bremner_ you can set the remote HEAD so that the default branch checked out in a clone is whatever you want. But you still transfer everything. Maybe fancy stuff with bundles can avoid some transferring.07:27
Arrowmaster lxsameer: git clone -b BRANCHNAME REPOURL07:31
← tore__ left07:31
← aspotashev left07:31
→ impulze joined07:35
→ aspotashev joined07:38
← madduck left07:39
← impulze left07:39
→ ggregori joined07:39
← airborn left07:41
← sechrist_ left07:41
← boombatower left07:44
← shiba_yu36 left07:46
← Opettaja left07:49
← Buglouse left07:50
→ warthog9 joined07:54
← [Pi] left07:55
← leo2007 left07:55
← nils_r left07:59
→ shiba_yu36 joined08:00
→ dirk2 joined08:02
← dirk2 left08:02
→ dirk2 joined08:03
→ madduck joined08:03
← dirk2 left08:04
→ jdhore1 joined08:04
← Tac_Home left08:06
jdhore1 When I ran hg, I did basically the following to backport stuff from the development repo/tree to stable: hg export commit# > ../other-repo/commit.patch && cd ../other-repo/ && hg import commit.patch ...How would I go about doing that with git?08:06
→ Opettaja joined08:07
← Opettaja left08:07
→ Opettaja joined08:07
→ Tac_Home joined08:08
selckin git format-patch/ git am08:08
why isn't it a branch in the same repo?08:08
→ dr0id joined08:09
jdhore1 selckin, Because hg's branching is hateful08:09
selckin well git isn't08:09
jdhore1 I know08:09
dr0id heh08:10
jdhore1 For the next stable tree, it'll be a branch08:10
→ Murr joined08:10
→ Juo joined08:11
→ Dave^| joined08:11
← giallu left08:11
← lxsameer left08:11
→ syphar joined08:13
← syphar left08:13
← OOPMan left08:14
jdhore1 hmm...git format-patch doesn't seem to do what I want...It's giving me every commit from the commit i specified to head, i just want the commit i specified08:15
teuf jdhore1: git format-patch -1 SHA108:16
→ deuterium_ joined08:18
← Tac_Home left08:18
jdhore1 teuf, thanks08:18
← deuterium_ left08:21
← Murr left08:21
→ aminpy joined08:21
← xerora left08:23
→ [Pi] joined08:23
→ gebi joined08:23
← akosikeno left08:24
→ workmad3 joined08:25
jdhore1 This is why I love git so damn much right now08:25
← german__ left08:26
← HedgeMage left08:27
teuf jdhore1: reading the backlog, if what you want is applying a commit from one branch to another, git cherry-pick does exactly thtat08:27
→ Textmode joined08:27
jdhore1 Yeah, I know about cherry-pick between branches...If only it worked between trees/repos though... :(08:28
teuf you can add the repo you want to cherry-pick from as a remote, and then cherry-pick from that08:28
it's a bit weird to do that if the trees are unrelated imo, but that's also helpful08:29
← rgr left08:30
jdhore1 That's not a bad idea, but i think i'm going to stick with my current way and ease myself into the git way of doing things kind of slowly08:31
← bindaas left08:33
→ d0k joined08:37
→ satanas joined08:41
→ citrusme joined08:42
← tewecske left08:42
ChanServ set mode: +v08:44
satanas hi guys. just a stupid question. a staged file is a file which is included in the index, right?08:44
→ tewecske joined08:44
→ sechrist joined08:44
selckin sure08:45
← thiago_muc left08:45
→ thiago_muc joined08:45
thiago_muc → thiago08:45
← thiago left08:49
→ bindaas joined08:50
satanas thanks for answering. and a tracked file what exactly is? Im not an english native speaker so its not as easy. Im confused08:50
← bindaas left08:50
→ bindaas joined08:51
selckin a file git knows about and is being tracked for changes08:51
→ nevyn^ joined08:51
satanas thanks selckin. but i translated word 'tracked' and it gives me a stupid translation. tracked stands for being in a commit?08:52
→ rcu joined08:53
← rcu left08:54
← tansell-laptop left08:54
selckin watched08:54
monitored08:54
but yes, a file in the repo08:54
satanas thanks a lot selckin. regards!08:55
→ boombatower joined08:56
← cesario left08:59
jdhore1 I think i'm falling in love with another piece of software...This is good :D08:59
← jjuran left09:00
→ radarek joined09:01
← satanas left09:02
← Opettaja left09:03
→ juvenal_ joined09:03
← juvenal left09:03
juvenal_ → juvenal09:03
→ Opettaja joined09:05
→ jatt joined09:07
→ BiggFREE joined09:09
← envi left09:09
→ noonoooo joined09:15
← citrusme left09:15
→ brian_g joined09:20
→ rgr joined09:22
→ eletuchy joined09:24
← namemon left09:29
→ namemon joined09:30
← dr0id left09:30
→ dr0id joined09:31
→ Raging_Hog joined09:31
← DrNick left09:31
← sechrist left09:31
→ sechrist joined09:32
→ leo2007 joined09:32
→ DrNick joined09:33
← BiggFREE left09:35
← _Ryth left09:37
← Samual left09:40
← radarek left09:46
← cylence left09:51
→ ch077179 joined09:53
← intripoon_ left09:53
← ch077179 left09:53
→ ch077179 joined09:54
→ royco-cup joined09:57
royco-cup Hey guys!09:57
can someone give me a hand on a git problem?09:57
nevyn just ask your question09:58
royco-cup ok...09:58
→ bauruine joined09:58
royco-cup I have changed modes and chown the .git and now it doesn't recogize the repo09:58
says its not a repo09:58
← jdhore1 left09:59
nevyn why did you do that?09:59
royco-cup they were associated to root (some of the files....) and I needed tem to be apache09:59
← nevyn^ left09:59
royco-cup nevyn: I reckon some file needs to be rw-rw-r ?10:00
guys is there any specific permission that the .git files need to be in order to be a repo?10:01
← ringlej left10:01
nevyn the permissions of the index files need to be whatever they need to be the directories within probably should be executable. but..nothing too special10:01
→ tore__ joined10:01
→ ezyang joined10:01
→ fcuk112 joined10:01
ezyang What's the easiest way to stage all files that I removed?10:01
→ gremset joined10:02
ezyang I.e. I did an "rm file" and now I want all of these missing files to be "git rm'd"10:02
royco-cup nevyn: thats weird... I dind't touch the files... just .git10:02
nevyn royco-cup: .git is the index files10:03
royco-cup nevyn: i see10:04
nevyn royco-cup: the working tree isn't that important in git terms.10:04
royco-cup nevyn: well I just tried having all u=rwx inside .git and still no luck10:05
← [Pi] left10:05
→ intripoon joined10:09
→ khmarbaise joined10:09
→ Jippi joined10:10
→ __name__ joined10:11
← __name__ left10:11
→ __name__ joined10:11
← royco-cup left10:11
→ [Pi] joined10:12
→ saschpe joined10:13
← blop left10:13
→ blop joined10:15
← khmarbaise left10:18
← blop left10:19
→ blop joined10:19
→ dirk2 joined10:19
← dys left10:20
← dirk2 left10:20
→ Anubisss joined10:22
→ RaceCondition joined10:22
← caseymcg left10:22
← chrisf left10:24
← gebi left10:26
→ gebi joined10:26
→ tvw joined10:27
→ rendar joined10:28
→ jstemmer joined10:29
→ khmarbaise joined10:33
→ SimonNa joined10:35
→ Agos joined10:37
→ jjuran joined10:37
→ Dave^|| joined10:39
→ saschagehlich joined10:40
← Dave^| left10:41
→ verto joined10:41
← verto_ left10:42
→ cytrinox joined10:44
→ evaluate joined10:45
→ crhan joined10:45
evaluate Hello.10:45
← ezyang left10:45
FauxFaux Patientest visitor ever.10:46
zomg Maybe he just wanted to say hello, 'cause he's a nice guy10:47
evaluate If I create a branch with git, then make some commits to that branch and finally merge the branch into master, the history of master will only contain a single entry where the whole history of the branch will be 'compressed'. Is there a way to merge a branch into master but still retain the history of that branch as separate commits?10:47
Sorry, I just like to write everything as a single 'line', I hate people that use the Enter key as punctuation...10:48
zomg Don't worry =)10:48
What did you use to see your history after the merge?10:48
→ ramusara joined10:48
FauxFaux evaluate: I'm guessing you mean man git merge --no-ff, but that's not what you're asking.10:48
jast evaluate: *automatic message* the 'git-merge' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-merge.html10:48
zomg If you just say "git merge x" then it should actually have a merge commit and the history should show up10:49
→ radarek joined10:49
zomg but if you use "git log", it won't show the merged branch side of the commits, so it looks like there's just the merge commit10:49
→ drizzd joined10:49
FauxFaux (It sounds like you're using git merge --squash and you're expecting the default behaviour.)10:49
zomg iirc anyway10:49
evaluate zomg, hmm, you're right. I'm not sure if I was too sleepy but I saw the merge commit and saw everything as a single commit, but the commits that I've done in the branch also appear.10:51
zomg gitk or such will display a nicer graphical view of them10:52
FauxFaux git config --global alias.lg "log --graph --pretty=format:'%Cred%h%Creset -%C(yellow)%d%Creset %s %Cgreen(%cr) %C(bold blue)<%an>%Creset' --abbrev-commit --date=relative" && git lg10:52
← adaptee left10:54
→ tty1 joined10:55
evaluate ok, now that I'm here, I'd also have one more curiosity. Is there a way to avoid conflicts when working with branches? For example if I change the version number in the configure file, git will usually see that as a conflict when doing a merge, is there a commonly accepted way to avoid that or is everyone just manually fixing these conflicts?10:55
← shiba_yu36 left10:56
← sechrist left10:56
→ chisler joined10:57
zomg evaluate: with config files you'd typically commit a "template" configuration file, like config.ini.base10:57
FauxFaux The best way to fix that kind of problem is to not have the version number in git and generate it at build time. :)10:57
← jfelchner left10:58
zomg then you would just create a copy of it, config.ini, and do your own configurations there10:58
→ eletuchy_ joined10:58
← ch077179 left10:59
→ ch077179 joined10:59
evaluate But that would mean that people who want to build the program from the git repository would have to go through some extra steps before they can build...10:59
zomg You could include the step in your build script11:00
Or you could provide a default config, which the user could extend in a custom config file (config.ini, which is committed, and a user.ini, which isn't, or something like that)11:00
← eletuchy left11:01
→ sechrist joined11:01
← Leonidas left11:03
← flowbee left11:04
→ Schmallon joined11:05
→ berserkr joined11:06
evaluate I think that requiring the user to execute a script when building from git might be best in this case then. Thank you for your help!11:06
zomg np11:07
→ moonflux joined11:08
← drl left11:09
← alip left11:10
→ psoo joined11:11
→ alip joined11:11
← jjuran left11:13
→ g0bl1n joined11:14
→ drl joined11:15
adamholt → adamholt_away11:15
adamholt_away → adamholt11:15
→ ringlej joined11:16
→ flowbee joined11:18
→ theoretical joined11:18
← babilen left11:20
→ fcuk112_ joined11:20
← fcuk112 left11:21
← psoo left11:21
→ LouisJB joined11:23
dr0id → MikeSee11:25
MikeSee → dr0id11:26
← aspotashev left11:26
→ Leonidas_ joined11:28
Leonidas_ → Leonidas11:29
← tvw left11:29
→ linusoleander joined11:35
← gebi left11:36
← Juo left11:36
→ ralc joined11:37
← linusoleander left11:37
→ jjuran joined11:37
→ psoo joined11:37
→ angasule joined11:40
← drl left11:41
← paja left11:41
→ maguey_ joined11:41
← rgr left11:46
← nellyb67 left11:47
← sechrist left11:48
→ paja joined11:50
→ TDT joined11:50
← aminpy left11:51
→ sechrist joined11:51
→ linusoleander joined11:52
→ aspotashev joined11:52
→ drl joined11:54
← jophish left11:55
← dcsobralWork left11:56
→ peterhil joined11:56
← Agos left11:57
← Titosemi left11:58
← maguey_ left11:58
← fcuk112_ left12:01
→ kumbayo joined12:02
← psoo left12:03
→ Titosemi joined12:05
← nardev left12:07
→ nardev joined12:07
← nardev left12:08
→ nardev joined12:08
← g0bl1n left12:09
← linusoleander left12:14
adamholt → adamholt_away12:15
adamholt_away → adamholt12:15
← sechrist left12:15
← kassah left12:17
→ kassah joined12:17
→ g0bl1n joined12:18
→ jophish joined12:19
← Opettaja left12:19
← Schmallon left12:21
→ snearch joined12:22
← ebi left12:26
→ Spockz joined12:28
→ ebi joined12:28
→ chessguy joined12:28
→ adaptee joined12:29
chessguy i'm not sure if this is the best place to ask this, but i had my machine all set up to use github, and now suddenly whenever i try to do anything with remote on one of my github repositories, i get a "Permission denied (publickey)."12:29
← Titosemi left12:29
→ Titosemi joined12:29
cbreak do you use a public key?12:30
chessguy yes12:31
→ tyteen4a03 joined12:32
← RobertLaptop left12:32
← tore__ left12:32
→ gusnan joined12:34
← crhan left12:34
← rushed left12:34
← Titosemi left12:35
→ Titosemi joined12:35
cbreak chessguy: does github know about it?12:37
chessguy cbreak, it did...12:37
let me try adding it again12:38
cbreak chessguy: are you on a unix?12:39
if so, try ssh [email@hidden.address]12:39
→ babilen joined12:39
chessguy cbreak, on windows, but using git bash12:41
cbreak, ultimately, it gives the same error12:42
cbreak http://paste.the-color-black.net/130590is what I get.12:43
so it seems you don't use the correct key12:43
is it in ~/.ssh/id_rsa?12:43
→ mcdebugger joined12:43
→ whaley joined12:43
→ lasfjie3w joined12:44
← khmarbaise left12:44
moonflux I'm trying to do the following: upstream cherry-picked some of the commits in my branch. I already found out that I can use gitcherry12:44
ups12:45
... to find out which were picked12:45
not I'd like to drop the picked commits from my branch12:45
not==now12:45
fat finger day12:45
chessguy cbreak, no, if i do "ssh -v [email@hidden.address] one of the things it says is "debug1: Trying private key: /c/Program Files/emacs-23.2/.ssh/identity"12:45
cbreak moonflux: why not git rebase upstreambranch?12:45
chessguy: that sounds weird12:46
moonflux cbreak: because upstream also merged a ton of other stuff and I have to amend my stuff to make rebase succeed12:46
cbreak try to find out where your private key is, then use it12:46
maybe you remember where you put it12:46
moonflux for now I just want to see what I still have to work on12:46
mjt that's an.. interesting home directory :)12:46
cbreak moonflux: but if you don't do that, you will not have the changes they picked12:47
moonflux: so if your other changes depend on those, it will break12:47
← theoretical left12:47
← pgib left12:47
moonflux cbreak: well, I have two cases (a) independent changes (like changing .gitignore) -- I can just drop these for now (b) stuff where other stuff depends on. I#d like to have these as merge points12:47
but for now I'm looking for a solution for (a) first12:48
cbreak merge points?12:48
you can drop things with git rebase -i12:48
← g0bl1n left12:48
cbreak and then deleting the line that represents that commit12:48
moonflux well, I haven't thought much about the merging yet ;)12:48
cbreak I don't think your approach is reasonable12:48
chessguy cbreak, well there is a key at "/c/Program Files/emacs-23.2/.ssh/known-hosts"12:48
cbreak a rebase seems much better12:48
chessguy: thta's not a key12:49
chessguy: that's a list of remote hosts your ssh knows12:49
moonflux I'll start with the rebase and see what I have to do then. thanks12:49
(the rebase -i I mean)12:49
→ gozoner joined12:50
chessguy cbreak, well it starts off "github.com,207.97.227.239 ssh-rsa AAAAB..."12:51
isn't the part at the end a key?12:51
moonflux cbreak: meh. you're right, I'm going to do the full rebase now everything else gives me headaches ;)12:52
mjt it's the host key12:52
cbreak chessguy: it's the public key of the remote12:52
chessguy ah12:52
cbreak chessguy: your problem is that you need YOUR OWN KEY12:52
chessguy gotcha12:52
cbreak the private key that you use to prove that you are indeed yourself12:52
chessguy i have a gem-private_key.pem12:54
cbreak don't know what that is12:55
→ maguey_ joined12:55
jast probably an SSL key and thus not relevant12:55
← CannedCorn left12:55
← maguey_ left12:55
cbreak a valid private key should start with this: -----BEGIN RSA PRIVATE KEY-----\nProc-Type: 4,ENCRYPTED12:55
jast well, we already found a private key, right? in that emacs dir12:55
cbreak well, a valid RSA one :)12:55
mjt jast: that was ssh-known-hosts file12:56
cbreak no, he said he gets a warning that that key is not found12:56
jast question is whether it's a personal key or something that was bundled with something12:56
mjt: "debug1: Trying private key: /c/Program Files/emacs-23.2/.ssh/identity"12:56
cbreak chessguy: your best option is to remember where you put the private key :)12:57
mjt ah12:57
cbreak or you can hope that the windows file search does not suck terribly, and just search for the string -----BEGIN RSA PRIVATE KEY-----12:57
mjt btw, file named "identify" usually stores ssh1 key12:57
i doubt github works over ssh112:57
ssh2 keys are named id_rsa and id_dsa12:58
jast that "Proc-Type" header isn't in any of my SSH private keys, btw12:58
mjt neither is in mine12:58
cbreak jast: I use encrypted private keys12:58
with a passphrase12:58
you might use unencrypted ones12:58
mjt sure, as everyone else?12:59
← Jippi left12:59
→ german__ joined12:59
mjt hm. one of my private keys starts with that "4,encrypted" indeed12:59
jast oh, yeah, right12:59
mjt heh. i looked at my old test key ;)13:00
cbreak I use OpenSSH_5.2p1, OpenSSL 0.9.8l 5 Nov 200913:00
(which is what is bundled with os x 10.6.7)13:00
→ chessguy_ joined13:00
← babilen left13:00
← chessguy left13:01
← chessguy_ left13:01
→ chessguy joined13:02
chessguy aaaaaand...cue the Blue Screen of Love13:02
→ shiba_yu36 joined13:02
→ fcuk112 joined13:03
→ babilen joined13:03
chessguy ah13:05
i see what happened13:05
← shiba_yu36 left13:05
chessguy for some reason when i installed "git bash", it set "/c/Program Files/emacs-23.2/" as my home directory13:05
← sraue left13:08
→ akerl joined13:09
→ sraue joined13:09
→ madsy joined13:10
← xxoxx left13:12
← babilen left13:13
→ babilen joined13:14
← babilen left13:14
→ rgr joined13:15
← saschpe left13:17
chessguy ok, so i've generated a new public and private key. but when i "ssh -v [email@hidden.address] it's not offering that key13:17
→ jfkw joined13:17
akerl chessguy: where is your key saved?13:18
chessguy in /c/projects/id_rsa13:18
→ babilen joined13:19
akerl Generally, unless you put it in your home .ssh folder, it won't be automatically checked to send. If you don't want to put it there, I suggest checking out .ssh/config13:19
← babilen left13:19
chessguy when i try to "ssh-add /c/projects/id_rsa", i get "Could not open a connection to your authentication agent."13:20
→ xxoxx joined13:20
← chrizz left13:20
→ babilen joined13:20
chessguy akerl, i like that idea, but for some reason git thinks that my emacs directory is my home directory13:20
moonflux can't I tell git on rebase 'see, I know this commit was applied, it just doesn't look the same because other stuff was changed before, I know it has that sha, merge so you get the upstream result up to there and show me the next"? ie. map shas before merge/rebase. right now in my first failing commit I have to decide which parts of upstream I want to pick now and which I'll keep for later since the changes will be introduced in a later change as I'm afraid13:20
I'll completely break my commit history. (or shouldn't I be afraid?)13:20
← jjuran left13:21
akerl chessguy: Thats... odd. Git shouldn't be choosing your home directory itself, AFAIK it lets your ssh_client config handle that13:21
← dr0id left13:22
moonflux dammit, gotta leave, will try again later13:23
← moonflux left13:23
chessguy ohhhh, i think i know what's going on here13:23
← Titosemi left13:23
chessguy as usual, i am my own worst enemy :)13:23
→ chrizz joined13:25
← verto left13:29
→ Myth17 joined13:29
← EricInBNE left13:29
chessguy bwahaha13:30
* branch master -> FETCH_HEAD13:30
Already up-to-date.13:30
all better13:30
thanks guys!13:32
← orafu left13:32
← babilen left13:32
→ shiba_yu36 joined13:33
cbreak moonflux: easy, just git rebase --skip13:34
just read the output that git rebase gives you on a conflict :)13:34
← rgr left13:35
→ babilen joined13:35
→ SimonP86 joined13:36
→ ColloquyUser joined13:38
← primalmotion left13:38
→ rgr joined13:38
← babilen left13:40
→ babilen joined13:41
→ thiago joined13:42
← gusnan left13:42
→ rcu joined13:45
← babilen left13:46
← namemon left13:46
← mcdebugger left13:46
← rcu left13:46
← drl left13:46
→ Modius_ joined13:48
← gbacon left13:49
→ Anatoliy joined13:50
← Modius left13:51
← gremset left13:52
← fcuk112 left13:54
→ theoretical joined13:54
← Anatoliy left13:55
→ namemon joined13:55
→ bid joined13:56
← german__ left13:57
→ CannedCorn joined13:57
← aspotashev left13:57
← karstensrage left13:58
← SimonP86 left13:58
bid hello, whats the command for force checkout? i want to checkout older revision and not marge it with my current work. (the client is running on windows os) thanks13:58
cbreak just use a normal checkout13:58
→ primalmotion joined13:59
→ SimonP86 joined13:59
→ karstensrage joined14:00
→ Anatoliy joined14:00
bid it will run over my changes? what about aditional files? it will delete the ones that are not related to the revison i checked out?14:00
karstensrage → Guest1843914:00
cbreak run over?14:00
bid sorry for my english,14:00
cbreak it will delete tracked files that don't exist in the branch /commit you switch to14:01
bid override maybe its a better word14:01
cbreak if you want to kill local uncommitted changes, use git reset --hard HEAD14:01
bid yes thats what i want14:01
cbreak this is ireversible.14:02
bid thats good, i am doing it for my QA system14:02
cbreak while you are in a detached head state, DO NOT COMMIT14:02
switch back to a branch with git checkout branchname before committing14:02
bid no commits will be done14:03
→ ronr_ joined14:03
← Mizar-jp left14:04
bid cbreak: general question, is it the same in SVN? i know its not the chan but #svn is slow today :(14:06
cbreak no14:06
← Anatoliy left14:06
→ hyperair joined14:06
← namemon left14:06
← sbell left14:06
cbreak svn sucks. :)14:06
it doesn't have real branches for example, and you can't jump around in the history14:06
bid its legacy here :\14:06
cbreak http://eagain.net/articles/git-for-computer-scientists/14:07
that's how git is14:07
→ Mizar-jp joined14:07
bid its great i know, but for now we are using SVN :\ i will move to git thats for sure14:08
← radarek left14:08
→ mklappstuhl joined14:09
→ psoo joined14:10
← deuterium left14:10
→ Titosemi joined14:11
→ ajt_ joined14:11
→ huntca joined14:13
← ColloquyUser left14:15
→ babilen joined14:16
andres cbreak: while I aggree with you in that svn is crap you can jump at any point in the history14:16
(takes a while though ;))14:17
← seivan left14:17
← jophish left14:19
← shiba_yu36 left14:19
→ shiba_yu36 joined14:19
→ oriba joined14:21
ronr_ I'm not sure whether this is a git issue or a 'linux' issue. I'm trying to connect to a git repository with gitosis using 'ssh gitosis@myhost'. I've created a key, copied it to the git server, added it to the gitosis.conf. it still seems that if I try to ssh to the server, it requests a password until it fails on Permission denied (publickey,password).14:22
Any idea how to solve it?14:22
→ beatak joined14:22
→ jophish joined14:22
→ mastroDani joined14:22
← shiba_yu36 left14:22
bremner_ !gitosis14:22
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/gitolite14:22
→ shiba_yu36 joined14:23
selckin error still at the ssh level, key is wrong & password is14:23
← oriba left14:23
cbreak ronr_: do you really use the key?14:23
ronr_: where's your private key?14:23
bremner_ even if you don't switch, read the gitolite ssh troubleshooting docs14:23
selckin also ssh will not allow login if permissions on ~/.ssh are wrong14:23
cbreak can you upload your private key to a pastebin? :D14:23
← beatak left14:23
→ andreim joined14:24
andreim hello, I deleted some files using rm instead of git rm and now the changes are not staged for commit. how can I stage them?14:24
cbreak andreim: man git-add14:25
jast andreim: *automatic message* the 'git-add' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-add.html14:25
cbreak look for -u and -A14:25
note that both can take path arguments too14:25
→ jfelchner joined14:25
← jfelchner left14:25
→ SimplCreative joined14:25
andreim cool ! that worked14:25
thank you14:25
← andreim left14:26
← ajt_ left14:26
ronr_ selckin: ah, you're here too :)14:27
cbreak: dunno if I really use the key, first time I'm doing it from linux. the private key is at ~/.ssh14:27
sorry for the delayed responses. I lag real bad.14:28
cbreak in ~/.ssh/id_rsa?14:28
try to ssh-add it14:28
ronr_ cbreak: considering it'll be absolutely useless for you, sure ;)14:28
← psoo left14:28
ronr_ ^regarding the private key @ pastebin :)14:28
the permissions on the key seem fine.14:29
"Could not open a connection to your authentication agent." <-- could that be an issue? :p14:30
cbreak it's optional14:30
you can ssh -u keyfilepath too I think14:30
selckin -i i think14:30
cbreak right14:31
selckin check the servers ssh log too, see if it's complaining about file permissions14:31
cbreak -i is the one :)14:31
← blueyed left14:31
ronr_ selckin: I wouldn't know where to look.14:32
selckin /var/log/sshd* probably14:32
← eletuchy_ left14:32
ronr_ tried with -i, gives me a 'invalid ELF header' error.14:32
cbreak uhm...14:33
that makes zero sense14:33
like, none at all14:33
maybe your remote ssh installation is broken14:33
← [Pi] left14:33
→ blueyed joined14:33
ronr_ I have no problem ssh'ing to it from the windows machine.14:34
selckin do you use a key there?14:34
cbreak weir14:34
weird14:34
← jophish left14:34
ronr_ yup14:34
selckin putty & openssh's keys are interchangeable without converting14:34
→ dv310p3r joined14:34
selckin *aren't14:34
ronr_ no, no. what I said was misleading, sorry.14:35
→ bitkiller joined14:35
ronr_ using the git-bash I created a key on windows, and I can git clone using ssh without any problem. haven't tried using the key with PuTTy.14:35
selckin then it should just work if copy the same key to linux ~.14:36
← rchavik left14:37
→ [Pi] joined14:37
ronr_ hmm, I could try using it just to test it out, but I wouldn't want to use the same key.14:38
→ shiba_yu_ joined14:38
← shiba_yu36 left14:39
→ lhunath joined14:39
lhunath How do I get rid of a submodule?14:39
→ akosikeno joined14:39
lhunath I deleted it from .gitmodules, but any files in that directory are still ignored.14:39
and git submodule still complains about it.14:39
ronr_ I don't see anything in the logs, either.14:39
lhunath grep sees a reference to it in .git/index, but I can't edit that.14:40
→ tore__ joined14:40
cbreak lhunath: git rm -r submodule14:40
maybe that works :)14:40
lhunath heh. cool. it did.14:41
→ banghouse joined14:43
→ madewokherd joined14:43
→ norican joined14:43
→ shiba_yu36 joined14:44
← shiba_yu_ left14:44
→ tvw joined14:44
← chisler left14:45
← rubydiamond left14:50
← angelsl left14:51
→ chisler joined14:52
← rgr left14:52
→ Garda joined14:52
→ chisler_ joined14:53
← lhunath left14:53
Garda hello14:54
I just typed "git remote add origin [email@hidden.address]14:54
notice the "giit"14:54
how do I fix it?14:54
cbreak edit .git/config or just git remote rm it and add the correct one14:55
or man git-remote set-url14:55
jast *automatic message* the 'git-remote' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-remote.html14:55
← SimplCreative left14:55
→ SimplCreative joined14:55
← ymasory left14:55
→ noviktk joined14:55
← chisler left14:56
chisler_ → chisler14:56
→ diegoviola joined14:56
Garda done14:57
thanks cbreak14:57
← Garda left14:57
→ ajt_ joined14:59
→ t0rc joined15:00
← norican left15:00
← evaluate left15:02
→ seivan joined15:02
→ OSInet joined15:02
← boombatower left15:03
→ jophish joined15:03
← jailbox left15:03
← diegoviola left15:04
→ OOPMan joined15:04
→ diegoviola joined15:04
← noviktk left15:04
→ RobertLaptop joined15:06
→ ymasory joined15:07
→ LiohAu joined15:07
→ radarek joined15:07
LiohAu i would like to install gitolite, do you know any up to date guide for debian ?15:07
← berserkr left15:10
← RobertLaptop left15:12
FauxFaux LiohAu: It's pretty trivial.15:13
necromancer`zzz → necromancer15:13
cbreak LiohAu: https://github.com/sitaramc/gitolite15:13
FauxFaux I can't even remember reading any docs after installing the package.15:13
cbreak always up-to-date :)15:13
and no need to do debian specific stuff15:14
just do the same as on any unix15:14
FauxFaux Or run "sudo apt-get install gitolite" and then go home and screw your wife? :)15:14
cbreak my wife? have fun with that...15:15
← SimplCreative left15:16
→ SimplCreative joined15:16
→ f00li5h joined15:16
LiohAu FauxFaux: i tried that already, but it fail15:17
→ tkil joined15:17
tkil not sure if it's worth reporting as a bug, but i confused myself when i ran "git diff no-such-file" and it gave me a usage string, instead of telling me that "no-such-file" doesn't exist...15:18
→ gusnan joined15:18
tkil verseion 1.7.4.4 on fedora 14 via the rpm.15:18
← bitkiller left15:19
→ bitkiller joined15:19
→ kukks joined15:19
cbreak tkil: that's because the first argument to git diff does not have to be a file15:20
LiohAu well i reinstalled it, and it seems that it worked15:20
tkil right, it can also be a commitish ... but if it doesn't match, shouldn't i still get an error?15:20
and it gave me the same usage error even if i tried: git diff -- no-such-file15:21
and anything past the "--" *should* be a path, if i'm reading the man page correctly.15:21
cbreak yes15:22
tkil huh. and i get a sane error if i'm in a .git repo... wonder if i'm hitting a corner case where git diff requires exactly two paths if it's not in a git dir or subdir...15:22
ronr_ okay, I'm afraid to ask this, but could the problem be related to running as root?15:22
I've configured both root and another user. the non-root works, the root doesn't.15:22
tkil ok, that makes a bit more sense, but i'd still like a clearer error message. let me see if i can come up with the tiny patch that woudl do that.15:23
or maybe i go drink more caffeine until i can type properly. :(15:23
cbreak git diff isn't supposed to work if it's not inside a repo15:23
ronr_: working as root isn't supposed to be done either15:24
selckin ronr_: check /etc/ssh/sshd_confnig for PermitRootLogin15:24
but kinda bad idea15:24
cbreak ronr_: note that root has a different home directory15:25
tkil looks like they set it up so it works as "plain" diff, but with all the git bonuses (modes, creation, etc). or so i'm guessing; looking at the code now.15:25
cbreak so it needs its own key15:25
→ ThomasLocke joined15:25
→ berserkr joined15:27
→ nlisgo joined15:27
→ RobertLaptop joined15:29
← SimplCreative left15:30
→ Buglouse joined15:31
→ rgr joined15:31
→ aminpy joined15:31
→ logik-bomb joined15:31
adamholt → adamholt_away15:32
adamholt_away → adamholt15:32
elliottcable → ryyvbggpnoyr15:33
→ devth joined15:34
← RobertLaptop left15:34
ryyvbggpnoyr → nyrktbeqba15:35
← ronr_ left15:35
→ rubydiamond joined15:35
→ ronr_ joined15:35
→ Swaagie joined15:35
← dwmw2 left15:38
necromancer → necromancer`zzz15:38
→ Corren joined15:38
nyrktbeqba → elliottcable15:40
← Pathin_ left15:42
← juvenal left15:42
→ _1marc joined15:43
→ Pathin joined15:44
→ nowhereman joined15:45
→ RobertLaptop joined15:46
LiohAu can i use a dsa key instead of rsa while installing gitolite?15:46
bremner_ should be fine15:47
← nowhere_man left15:48
← mastroDani left15:48
← Corren left15:48
← tvw left15:49
→ joel_falcou joined15:49
← RobertLaptop left15:51
← timemachine3030 left15:51
necromancer`zzz → necromancer15:57
→ tty3 joined15:57
tty3 → Guest4650215:57
← Pathin left15:58
← xiangfu left15:58
→ orafu joined15:59
← ronr_ left15:59
→ ronr_ joined16:00
← fr0stbyte left16:00
← nlisgo left16:00
← tty1 left16:00
← cytrinox left16:00
← akosikeno left16:01
← akerl left16:01
→ cytrinox joined16:02
necromancer → necromancer`zzz16:02
→ mikewintermute joined16:06
→ namemon joined16:09
← Myth17 left16:09
← namemon left16:14
→ icwiener joined16:15
→ jophish_ joined16:15
← saschagehlich left16:16
← jophish left16:16
→ Corren joined16:17
← gnufied left16:17
← paja left16:18
← radarek left16:18
angasule I'm giving a presentation about git soon, and I know I will get lots of questions related to the 'game developer' workflow (lots of large images, some large videos, and their 'sources'), what is the recommended solution for that these days? Partial solutions like git-p4 or whatever are acceptable if that is the right answer, of course16:18
→ paja joined16:18
→ timemachine3030_ joined16:19
bremner_ git-annex might be worth checking out16:19
cbreak git on its own deals rather poorly with binary files that change a lot and/or are big16:20
← gusnan left16:20
→ RobertLaptop joined16:21
→ canton7 joined16:22
angasule bremner_: ah, I was wondering about it, but I haven't yet read a good description of it, thanks16:22
→ Samual joined16:22
← rgr left16:22
angasule cbreak: yes, of course, the solution would pretty much involve having something else manage assets16:22
→ airborn joined16:22
← kassah left16:23
angasule the ideal thing would be svn/p4 as a submodule, if it were possible16:23
← canton7 left16:23
→ HG` joined16:23
← ronr_ left16:24
← jpieper left16:25
← RobertLaptop left16:26
→ rgr joined16:27
← tore__ left16:28
→ Schmallon joined16:28
→ steph021_mac joined16:32
Thor` hey16:33
I would like to customize the error message ssh gets when trying to login as git@domain with gitolite16:34
← sunn left16:34
← kumbayo left16:35
← ramusara left16:35
← mtx_init left16:36
← Corren left16:37
bremner_ Thor`: customize it how?16:38
→ mtx_init joined16:38
→ RobertLaptop joined16:38
Thor` bremner_: when you ssh directly on [email@hidden.address]16:39
→ pen joined16:39
Thor` bremner_: you can read a message, in addition to the "PTY allocation request failed on channel 016:39
message.16:39
sitaram Thor`: I think you can do that; at least you can *add* to the existing stuff16:40
← devth left16:40
sitaram Thor`: what exactly does it give you now?16:40
→ empity joined16:40
Thor` sitaram: the thing I'd like to do is to print to an user the repositories (s)he can access, with the associated rights.16:41
→ sunn joined16:41
← sunn left16:41
→ sunn joined16:41
selckin think it does that if you ssh to it16:41
sitaram Thor`: huh? that already happens doesn't it?16:41
Thor` sitaram: I remembered that it does that, yes16:42
sitaram Thor`: perhaps you need to read this: http://sitaramc.github.com/gitolite/doc/report-output.html#_side_note_openssh_5_616:42
Thor` sitaram: but all I get is a "PTY allocation request failed on channel 0" message, and nothing else.16:42
← RobertLaptop left16:42
→ kassah joined16:43
Thor` sitaram: hmmm that would explain the problem16:43
→ temoto joined16:43
→ hendeby joined16:43
Thor` Version: 1:5.5p1-616:44
hmmm16:44
temoto Hello. Do they say something in lines of 'git represents with snapshots of tree (as opposed to changes)' ?16:44
(i know that commits store diffs, Q is not about internal representation)16:45
Thor` sitaram: ok it works fine with the -T flag16:45
sitaram: thank you very much for this great piece of code ;)16:46
tkil temoto -- could you ask that differently? i'm not sure what you're asking.16:46
→ tvw joined16:46
← selckin left16:46
temoto To be honest, i don't actually understand the question myself.16:47
tkil ha. :)16:47
sitaram temoto: depending on who "they" refers to, almost everything has been said by someone somewhere ;-)16:47
← zhengyang left16:47
← RaceCondition left16:47
sitaram Thor`: is that version on client or server?16:47
tkil i think that the short answer is "yes": git (from the outside) stores snapshots of entire trees, and generates diffs "on demand".16:47
→ DNomizo joined16:48
Thor` sitaram: server.16:48
tkil as contrast to sccs and probably cvs and svn, which store current state and diffs to go back in time. (er, maybe. it's been a *long* time since i've thought about that.)16:48
→ RaceCondition joined16:48
Thor` sitaram: I did an apt-cache show openssh-server16:48
sitaram Thor`: check client; it's the *client* that is killing connection16:48
→ Tac_Home joined16:48
← DNomizo left16:48
← paja left16:48
Thor` $ apt-cache show openssh-server | grep Version16:48
Version: 1:5.5p1-616:48
→ DoctorDarren joined16:48
Thor` sitaram: oooh16:48
temoto sitaram, they as in 'you know, they say sky is high'16:48
Thor` sitaram: okay, I didn't get that16:49
← nardev left16:49
tkil temoto -- i'm assuming you're not a native english speaker (and that's fine!)... it's just that your first question was worded oddly.16:49
→ nardev joined16:49
Thor` sitaram: $ apt-cache show openssh-client | grep Version16:49
Version: 1:5.8p1-1ubuntu316:49
tkil temoto -- so i'm trying to clarify it, is all.16:49
sitaram just do "ssh -V" on the same terminal you tried "ssh [email@hidden.address]16:49
Thor`: yeah 5.8 > 5.6 so of course you see the behaviour16:50
Thor` sitaram: same thing16:50
→ fcuk112 joined16:50
Thor` $ ssh -v16:50
OpenSSH_5.8p1 Debian-1ubuntu3, OpenSSL 0.9.8o 01 Jun 201016:50
tkil temoto -- there are many documents on the design and motivations on the git web site; have you had a chance to look through them?16:50
Thor` sitaram: yes.16:50
sitaram: you're right. :)16:50
sitaram preens16:50
Thor` sitaram: again, thank you very much for your help and gitolite.16:51
eagles0513875 hi guys :D16:51
temoto tkil, I heard it in argument about the need of 'merge' commits. Like merge commits pollute history, but they're inevitable because ... something regarded to snapshots vs. changesets.16:51
→ saleh joined16:51
sitaram Thor`: you're welcome!16:52
eagles0513875 hey guys16:52
sitaram temoto: merges have nothing to do with snapshots versus changesets16:52
eagles0513875 just out of curiosity16:52
im looking to create a repo16:52
tkil temoto -- merge commits are needed in any non-linear development model; git might make them more obvious than earlier (especially non-distributed) SCMs did.16:52
eagles0513875 and i follow the steps on the site16:52
how then do i determine what the git url will be16:52
temoto tkil, yeah, i know and like how git works.16:52
← Tac_Home left16:53
← int3__ left16:53
tkil temoto -- snapshots vs changesets seems like a false dichotomy, at least at the user level.16:53
sitaram temoto: merges have to do only with parallel development of different features and/or by different people. "non-linear", as tkil said16:53
tkil temoto -- storage in terms of snapshots vs. changesets ... hm. should only effect efficiency of operations, not the semantics: if you're going to merge, it's going to be a merge, and how it's stored by the SCM shouldn't matter to the user.16:55
← Guest18439 left16:55
→ xixor joined16:55
temoto I wonder would it be easier for those people if merge commits that just join two commits w/o diff would be hidden from history.16:55
→ Tac_Home joined16:55
→ Jippi joined16:55
xixor Hi everyone, sorry to be a n00b, but is there way I can get a list of which files have not been committed?16:55
tkil temoto -- if they don't interact, you can do that, i'm pretty sure.16:55
xixor -- git status16:55
← bindaas left16:56
→ Corren joined16:56
→ gnufied joined16:56
→ karstensrage joined16:56
tkil temoto -- that's basically the difference between a "proper merge" and a rebase, with all the pros and cons of the latter.16:57
← hendeby left16:57
→ adamm joined16:57
xixor tkil: ah, great, thanks16:57
temoto tkil, they don't interact with merge commit id?16:57
→ bindaas joined16:58
cbreak rebase does not create a merge commit16:59
tkil temoto -- if you rebase, then there's no merge commit at all. (but see my comment above about "if they don't interact", which is vague but hopefully close to truth)16:59
temoto -- the git-rebase man page has some good examples (and pretty pictures!)16:59
← _1marc left17:00
temoto tkil, that's exactly what i didn't understand: you said 'if they don't interact'. With what?17:00
tkil temoto -- the question of merge or rebase implies that you have two heads of development, yes?17:00
← tyteen4a03 left17:00
← thiago left17:01
← nardev left17:01
temoto I know how things work in git, just bad at language :)17:01
tkil, sure17:01
tkil temoto -- by "interact", i meant whether those two heads touched the same code or interfaces. if they're completely independent, then you should be able to rebase each head one at a time without any problems.17:01
temoto -- :) not trying to be condescending or obtuse, just trying to be clear, sorry.17:01
if, on the other hand, those two heads deal with the same code, then you start needing to do messier rebase: rebase to head 1, patch up conflicts with head 2, rebase to head 2.17:02
at some point it's just easier to have a merge commit. :-/17:02
this might be relevant to this discussion: http://code-redefined.blogspot.com/2010/11/git-rebase-tricks-cleaning-up-twisted.html17:03
temoto I usually try rebase and merge only in face of conflicts.17:03
→ _iron joined17:03
tkil temoto -- if you are a consumer or local-only repo, that's fine.17:03
FauxFaux 22/17:28:46 < ajmiles> dirtus 3 Faux ? :p17:04
Mispaste.17:04
temoto tkil, how is it not fine if i'm one of 4 commiters?17:04
→ ISF joined17:05
→ moonflux joined17:05
← RaceCondition left17:05
← ralc left17:05
→ RaceCondition joined17:08
lasfjie3w → pgib17:08
tkil temoto -- if other people pull from your repo ("producer repo"), then a rebase can confuse the remote "consumer" repos quite badly. see, among others: http://www.kernel.org/pub/software/scm/git/docs/git-rebase.html#_recovering_from_upstream_rebase17:10
i guess "upstream" and "downstream" are more common terms, sorry about that. :(17:11
← codeshah left17:11
← _iron left17:11
→ Swimming_bird joined17:12
← cthuluh left17:12
temoto tkil, i didn't mean interactive rebase changing history. git pull --rebase doesn't confuse downstreams.17:12
← ThomasLocke left17:12
temoto and it must be said that we're using a centralized repo workflow17:13
adamholt → adamholt_away17:13
adamholt_away → adamholt17:14
tkil temoto -- if you only ever pull into that repo/head, then you're fine.17:14
but if the centralized repo is ever rebased, you'll have issues.17:14
→ ronr_ joined17:14
tkil interactive doesn't have anything to do with it.17:14
← OSInet left17:14
temoto Nah, we're only pushing into centralized repo, but then only fast-forward.17:15
→ mmw joined17:15
tkil if centralized repo has A -> B -> C -> D, and someone pulls it and works on D, but then you rebase and the central repo becomes A -> B' -> C' -> D', then the person working against D is in a bind.17:16
ok. so long as it works for you. :)17:16
← steph021_mac left17:16
temoto :)17:17
→ ThomasLocke joined17:18
→ farresin joined17:18
← Error404NotFound left17:19
ChanServ set mode: +v17:19
→ kephas joined17:20
← kipras left17:20
farresin hi guys. which is the difference between a staged file and a tracked file? im not a native english speaker so i dont understand this very well17:20
→ kipras joined17:20
→ thiago joined17:21
← babilen left17:22
temoto farresin, tracked file is a file that git knows about, tracks changes, but currently the file is not yet changed.17:23
← nowhereman left17:23
temoto farresin, changed file becomes staged after git add17:23
← fcuk112 left17:23
temoto farresin, stage is what going to be commited.17:23
→ pantsman joined17:23
← pantsman left17:23
→ pantsman joined17:23
← saleh left17:24
temoto farresin, file state path: untracked -> staged -> (commited (tracked) -> unchanged -> changed -> staged -> )+17:25
→ avinashhm joined17:26
→ Juo joined17:26
← mmw left17:26
← bid left17:28
← ronr_ left17:31
→ gremset joined17:32
xixor hi, is it possible to put .gitignore files in subdirectories to ignore certain extensions only in those certain subdirectoreis?17:32
→ ningia joined17:32
← gozoner left17:39
legumbre_ → legumbre17:39
→ gozoner joined17:39
→ slonopotamus joined17:39
→ sbell joined17:40
temoto xixor, yes, it just works like you want by default.17:41
→ jfelchner joined17:42
xixor ah, ok17:42
temoto xixor, but somehow it's better to have one .gitignore - you know where to look why files are invisible.17:42
xixor so you would add the specific path? /path/to/ignore/*.log ?17:42
→ babilen joined17:42
← angasule left17:42
temoto Paths are relative to working directory.17:42
Yes.17:42
→ gusnan joined17:43
cbreak depending on what you write they will apply to all, or only relative to your current location17:43
you can have .gitignore in subdirectories as well17:44
you can have .git/info/exclude17:44
temoto In this specific case, i have *.log in my global ~/.gitignore17:44
cbreak that should apply to all files in the whole repository17:44
or rather, that should apply to all untracked files in the whole working directory17:45
adamholt → adamholt_away17:45
adamholt_away → adamholt17:45
xixor ok17:45
tkil man. it is really hard to train my fingers to linux coding style, instead of my own...17:45
← charon left17:46
→ charon joined17:46
wwalker tkil: which style is bothering you?17:46
tkil almost all of it. :) half-cuddled 'else', tabs instead of spaces, no spaces after opening parens...17:47
i'm fully aware that i can't change it, i'm just whining.17:47
→ saleh joined17:47
xixor what style? linux kernel C style ?17:47
tkil i use *lots* of whitespace, hard to remind my fingers to not add it.17:48
xixor -- yes. am working on a patch for git, it's in the kernel style.17:48
temoto linux style has one thing right: indentation is tabs. :)17:49
tkil uhg.17:50
wwalker tabs is the right thing :)17:50
temoto lolzy17:50
tkil i'm the type of person who puts a tabs-to-spaces hook on his repo. :)17:50
→ RaceCondition_ joined17:51
slonopotamus tkil: try astyle17:51
offby1 launches into a tab tirade17:51
xixor tkil: ah, interesting17:51
wwalker I put 'leading spaces reject commit' hooks17:51
slonopotamus tabs are broken17:51
xixor tkil: do they have a format requirements document you are working from? Or just matching some example code?17:51
tkil wwalker -- must be awesome for people who use /*\n * comment blocks... :)17:51
cbreak trailing spaces reject hook ftw17:51
tkil xixor -- SFAIK, it's polite to match style when patching, so i'm going along with it.17:51
like i said, i'm just whining.17:51
cbreak -- totally agree!17:52
wwalker tkil: it caused problems, I pulled it, but got my point acoss first17:52
xixor tkil: ah, yeah, always following existing code style is good practice17:52
wwalker :-( most of my devs didn't know what a comment was....17:52
tkil slonopotamus -- "astyle"?17:52
slonopotamus tkil: try google17:52
wwalker cbreak: Yes!!!17:52
tkil slonopotamus -- i use emacs c-mode, and it has linux style built in, so not too much pain.17:52
temoto We have 'any tabs reject' in pre-commit. Following many styles trains tolerance and calmness.17:53
slonopotamus tkil: that should work too :)17:53
tkil slonopotamus -- mostly on-the-line issues here. e.g., linux style is: if (foo || bar) {\n while mine is: if ( foo || bar )\n{17:53
etc17:53
slonopotamus tkil: astyle would fix that17:53
tkil wwalker -- also fun, do you do any work with windows devs? 4-char tabs... shudder.17:53
→ gebi joined17:54
tkil slonopotamus -- sure. i'll keep it in mind.17:54
← RaceCondition left17:54
RaceCondition_ → RaceCondition17:54
slonopotamus tkil: you write your code whatever you like, run astyle over it and voila - it is in upstream expected style17:54
tkil ... and try to resist the urge to whine around a bunch of opinionated nerds. ;->17:54
slonopotamus tkil: well, except small quirks, but who doesn't have them?17:54
← cube_ left17:54
tkil slonopotamus -- no, my style is right ... you all are the ones with quirks. :)17:54
→ g0bl1n joined17:54
wwalker tkil: yes, but I simply ignore their(win devs) existence, and point to the Snow Leopard and Fedora DVDs on my desk17:55
→ jpieper joined17:55
← saleh left17:55
slonopotamus tkil: you can convert files into your style before editing and back to upstream after editing :P17:55
tkil wwalker -- i don't have that luxury, the guy i'm subcontracting for is a windows dev. i'm writing linux, but he wrote some compat code for other instruments, i had to deal with the tabs...17:55
slonopotamus -- nah, it doesn't bug me that much. and i need to learn to read it anyway, don't want to have to convert stuff just to comprehend it.17:55
tkil goes to paint the braces^H^H^H^H^Hshed a different color.17:56
wwalker tabs are still the right thing, they mean "an indention level" set your editor to indent as many spaces as you want per indention level, 2, 3, 4, 5, 8(eek)17:57
→ _1marc joined17:57
wwalker then all devs looking at the code can see it an the indention depth they are comfortable with17:57
tkil wwalker -- "your editor". git-web? your other devs' editors? http://www.jwz.org/doc/tabs-vs-spaces.html17:58
wwalker I implemented the "tabs are mandatory" when I inherited a code base that was indented with 1 space per level.17:58
tkil wwalker -- tabs used to indent function args?17:58
slonopotamus using more than 2 spaces for intendation is just a waste of space (no pun intended)17:58
tkil (i had a coworker that used tabs-to-indent-level, then spaces-to-align-args)17:58
→ radarek joined17:59
tkil slonopotamus -- Linus has an excellent point that 8-char indents (tabs or spaces) does tend to discourage deeply-nested functions...17:59
slonopotamus tkil: that is a very good example that shows that tabs are broken :)17:59
ChanServ set mode: +v17:59
tkil slonopotamus -- i've got my opinions, yes. :)17:59
wwalker well, there are 3 editors, vim, emacs, eclipse, all 3 allow you to set the width of a tab.17:59
slonopotamus tkil: 640x480 screen tends to discourage big functions too.17:59
temoto tkil, from that perspective, google's style of 2-space indentation makes me scary as hell.18:00
← p4tux left18:00
slonopotamus tkil: also does 24pt font18:00
tkil wwalker -- when given the luxury, i tend to C-x h M-x untabify C-x h M-x indent-region18:00
but i don't always have that luxury. :)18:00
temoto -- indeed.18:00
ANYWAY. right. back to coding. honest.18:00
farresin temoto just one thing. a tracked file is a file that has been included in a commit at least 1 time right? if not, then the file is not tracked yet18:01
wwalker shiftwidth=2 tabstop=218:01
:)18:01
temoto farresin, yes.18:01
farresin ok. thanks ;) regards!18:01
→ dwmw2_gone_ joined18:01
temoto wwalker, also sts=2 for proper backspace behaviour18:01
← tvw left18:02
tkil best(?) indent horror i ever heard of: someone used fibonacci sequence indents. (i think it might be referenced in that JWZ article, it was in the early netscape codebase)18:04
offby1 farresin: I might amend that and say that a tracked file could also be one that isn't present in any commit, but is currently in the index18:04
→ cthuluh joined18:04
farresin ok. offby1. i will remember this. thanks for the info. regards!18:05
← temoto left18:07
← erichammond left18:07
← farresin left18:07
→ towski joined18:07
→ farresin joined18:08
← logik-bomb left18:08
← gremset left18:09
→ mishok13 joined18:09
→ gremset joined18:10
← farresin left18:10
← cryptide left18:11
← DoctorDarren left18:14
→ sechrist joined18:14
← HG` left18:16
← mikewintermute left18:18
← Titosemi left18:19
→ poga joined18:19
← avinashhm left18:19
→ mcdebugger joined18:20
→ nevyn^ joined18:20
← noonoooo left18:21
← Sonderblade left18:22
→ nils_r joined18:22
← topi left18:22
→ serverhorror_ joined18:22
→ jailbox joined18:23
→ joshsmith joined18:24
← joshsmith left18:24
→ joshsmith joined18:24
→ boue joined18:25
← Tac_Home left18:27
← cyphase left18:28
← slonopotamus left18:29
necromancer`zzz → necromancer18:33
→ Tac_Home joined18:34
→ rkhessel joined18:35
← rubydiamond left18:36
→ bid joined18:38
→ brendon joined18:40
← bid left18:41
→ bid joined18:42
← xybre left18:42
→ cyphase joined18:43
← chisler left18:43
→ xybre joined18:43
← asedeno_work left18:43
→ caseymcg joined18:45
necromancer → necromancer`zzz18:47
← bid left18:47
→ bid joined18:47
→ legumbre_ joined18:48
← leo2007 left18:50
← airborn left18:50
← legumbre left18:50
→ hendeby joined18:50
legumbre_ → legumbre18:51
hendeby → Guest9751418:51
→ airborn joined18:51
← bid left18:52
→ bid joined18:52
← Guest97514 left18:52
necromancer`zzz → necromancer18:52
→ hakunin joined18:52
→ asedeno_work joined18:53
→ dankest joined18:54
← brendon left18:54
→ Guest97514 joined18:57
← sechrist left18:57
← mklappstuhl left18:58
← hakunin left18:59
→ chrisf joined18:59
→ foobar_ joined19:00
→ logik-bomb joined19:01
← logik-bomb left19:01
→ logik-bomb joined19:01
← logik-bomb left19:01
→ fcuk112 joined19:02
← [Pi] left19:02
← Schmallon left19:03
← Guest97514 left19:04
→ mdel joined19:04
→ foobar__ joined19:04
← foobar__ left19:05
→ Gavrila joined19:06
← chrisf left19:06
→ hendeby_ joined19:06
→ Amzul joined19:07
← hendeby_ left19:07
← _1marc left19:10
← bid left19:10
← asedeno_work left19:12
← LiohAu left19:12
→ _1marc joined19:12
← ymasory left19:12
→ hendeby_ joined19:14
← caseymcg left19:14
← hendeby_ left19:14
← Amzul left19:14
→ bid joined19:15
→ chisler joined19:16
← binjured left19:19
← bid left19:20
→ binjured joined19:20
→ bid joined19:20
→ rhalff joined19:21
→ jophish joined19:24
← jophish_ left19:25
← _1marc left19:26
→ Amzul joined19:26
→ harrisonk joined19:27
→ Sonderblade joined19:27
→ asedeno_work joined19:27
← berserkr left19:28
← Sonderblade left19:28
→ ymasory joined19:28
← bid left19:29
← Amzul left19:30
→ bid joined19:30
→ _1marc joined19:31
→ pidgas joined19:32
harrisonk I want to clone a git repo here: https://gitorious.org/fg/fgdata but I know it's massive and I only want 1 folder out of the dozen or so that are in it, thus what command options should I use? git clone git_url_here /textures ?19:32
→ pl0sh joined19:33
← binjured left19:33
→ binjured joined19:33
offby1 harrisonk: I am not sure there are any options to do what you want.19:33
pl0sh hey guys, can I use git as a remote repository on my own server?19:34
offby1 I vaguely recall there's some option like "--depth" which only gets the top-level items, but I don't know if that'd help you19:34
jast pl0sh: sure19:34
offby1 harrisonk: nah, --depth has nothing to do with it; I misremembered it19:34
→ dinsdale joined19:34
jast harrisonk: git cannot clone individual folders. you *can* tell it to skip old history (which is what --depth is for) but that's about it19:35
pl0sh jast: is there any tutorial or book I can read to set up my own repository on freebsd?19:35
harrisonk drat19:35
that repo is about 8 gigs in size if I remember corectly19:35
jast pl0sh: normally you just need sshd and proper permissions. if you want to manage a non-tiny number of repositories and users, check out gitolite; otherwise it's probably fairly easy to do manually19:35
harrisonk thus why I want only 1 folder19:35
offby1 yiukes19:36
yikes19:36
jast harrisonk: the remote might support using 'git archive' to download arbitrary trees19:36
but that won't enable you to commit anything19:36
← bid left19:36
→ bid joined19:36
← aminpy left19:37
harrisonk that leads me to another question: can I "pause" a git clone?19:37
pl0sh ok jast thanks!19:37
wereHamster harrisonk: no19:37
harrisonk well not pause it but rather download part today (for example) and tomorrow download the rest19:38
wereHamster no19:38
pl0sh jast: I think I found an article on the net19:39
jast resumable clone has been an outstanding feature for ages, but sadly it's a tricky challenge and no one has felt up+motivated yet19:39
← Gavrila left19:39
jast *up-to-it19:39
← pen left19:40
← gozoner left19:40
→ gozoner joined19:40
← pidgas left19:40
Vampire0_ → Vampire019:41
← Modius_ left19:41
→ Modius joined19:41
→ berserkr joined19:42
→ pen joined19:43
→ Gavrila joined19:45
← poga left19:47
← ymasory left19:47
→ jcon joined19:48
← workmad3 left19:51
← _1marc left19:52
→ pidgas joined19:53
← OOPMan left19:53
← boue left19:53
← snearch left19:58
tkil if you have shell access to the remote machine, there are ways you could do incremental download "by hand".19:59
hm. for that matter, you might be able to get a list of commits, fetch them individually, then repack locally.19:59
← Raging_Hog left19:59
tkil all a big PITA, but maybe worth it19:59
← CannedCorn left20:01
→ erichammond joined20:05
← Corren left20:07
tkil might also be the solution to whomever wanted to grab the textures ... if you can get just the remote commit, then tree, then filename, then blob....20:09
← d0k left20:09
tkil doesn't know enough git to know if that is even possible, though.20:09
→ LiohAu joined20:09
← dinsdale left20:09
← airborn left20:09
→ OOPMan joined20:09
← chisler left20:09
← alip left20:10
→ akkad` joined20:11
→ alip joined20:13
→ fr0stbyte joined20:16
→ CannedCorn joined20:17
← drizzd left20:18
→ paja joined20:19
← Swaagie left20:19
← joeytwiddle left20:21
mjt is there a way to "stash" a signed tag temporarily and restore it later? The tag is generated by an app and I'm trying to understand what it does but it complains that the tag (whcih I want to keep) already exist. Short of copying whole repository to a temp location..20:24
← rkhessel left20:24
← nevyn^ left20:26
offby1 mjt: geez, dunno20:28
maybe you can rename it20:28
mjt it should be much safer to just copy it for experiments :)20:29
tkil http://stackoverflow.com/questions/1028649/rename-a-tag-in-git maybe?20:29
elenril if i've accidentally 'git stash pop' while on a wrong branch, is there a way to get my stash back?20:30
teuf elenril: git reflog --all show stash entries, I don't know if you can get it back from there20:32
cbreak mjt: just rename it20:33
→ cesario joined20:34
elenril teuf: thanks, that seems to work20:34
→ EricInBNE joined20:34
→ joeytwiddle joined20:36
← gusnan left20:37
← adamm left20:37
← __name__ left20:37
mjt hm. now i'm confused. All docs says it's impossible to rename a tag :)20:37
→ dinsdale joined20:37
→ Murr_ joined20:38
← Murr_ left20:38
← MikhX left20:40
→ MikhX joined20:40
→ angasule joined20:40
→ Murr_ joined20:40
→ ymasory joined20:41
cbreak mjt: tags are refered to by files in .git/refs/tags20:41
← pl0sh left20:41
cbreak just rename the file20:41
mjt ah!20:41
cbreak (I don't know if this works, but my git experience says it should)20:42
there might be something special with signed tags that makes that not work though20:42
mjt that makes alot of sense. i knew but forgot about that fact :)20:42
cbreak make sure you remember the original name20:42
mjt thanks!20:42
cbreak so you can rename it back20:42
np20:42
← Juo left20:42
← fr0stbyte left20:44
→ Juo joined20:44
→ Opettaja joined20:44
→ fr0stbyte joined20:44
← radarek left20:44
← MikhX left20:45
→ MikhX joined20:45
← setmeaway left20:46
→ setmeaway joined20:47
← jstemmer left20:48
→ atrus joined20:49
→ adamm joined20:50
→ Hypnos joined20:50
→ airborn joined20:51
← jophish left20:52
← gnufied left20:54
← ricky left20:54
→ ricky joined20:55
→ jjuran joined20:55
← ExtraSpice left21:02
→ skoop joined21:03
← brian_g left21:04
→ brian_g joined21:05
← brian_g left21:06
← Murr_ left21:06
→ brian_g joined21:06
← LongBeach left21:08
← orospakr left21:11
→ radarek joined21:11
→ orospakr joined21:11
← chrisf|work left21:11
→ mastroDani joined21:13
← rhalff left21:14
← mastroDani left21:15
← cesario left21:16
← Gavrila left21:16
→ chrisf joined21:17
← Hypnos left21:19
← dinsdale left21:21
← jjuran left21:21
→ ralc joined21:23
← mcdebugger left21:28
→ npmccallum joined21:31
ChanServ set mode: +v21:34
npmccallum I have multiple changes in my working tree that I'd like to commit piece by piece, but they are the same file... is there anything like git stash that will go section by section through a diff and allow you to stash *part* of the file?21:34
Fissure add -p21:34
← ymasory left21:34
npmccallum Fissure: thanks! I knew I saw something like that, but I couldn't remember what it was...21:35
Fissure well, that does the opposite, but you can commit all of those and then do a rebase -i to test after each commit21:35
→ workmad3 joined21:35
← g0bl1n left21:35
→ dr_win_ joined21:36
← dr_win left21:37
tkil npmccallum -- if i find that i have a file with multiple logically independent changes in it, i tend to move it to the side, revert the original, and then use a merge or diff tool (i use ediff) to apply partial patches back and commit that way. one more option...21:37
← Adys left21:38
Fissure git-gui can do partial adds as well21:38
← Guest46502 left21:38
→ canton7 joined21:38
Fissure kinda silly to bring in external things when it's handled natively so well21:38
tkil Fissure -- hm. how does that handle the situation where two logical changes end up in the same patch hunk?21:39
bremner_ magit (emacs) is good at partial adds as well. Most of these things can add a single line21:39
tkil and i use emacs "natively", so ediff is actually more native than yet another gui tool. :)21:39
→ devth joined21:39
bremner_ tkil: then you should use magit21:39
Fissure git-gui can do line-by-line, and add -p has an option to let you edit the diff you want to apply manually21:39
tkil bremner -- huh. i thought i tried that once, but maybe not. i had been using git-status, am experimenting with dvc now.21:40
→ _1marc joined21:40
→ Adys joined21:40
tkil dvc might be the eventual winner, since i flip back and forth between git, hg, svn, and other crap pretty often (weekly if not daily)21:41
still has some rough edges tho.21:41
→ tty1 joined21:41
← RaceCondition left21:42
← radarek left21:42
← Tac_Home left21:42
→ codeshah joined21:43
Fissure i can't use raw svn anymore... too painful21:43
← vdv left21:43
tkil i'm using it through git svn now, but central repo is still svn.21:44
← ggregori left21:44
← LiohAu left21:46
Thor` sitaram: are you still here ?21:48
I'd like to know if there is a way to "mark" a repository web-accessible (from gitweb) with a flag in the gitolite.conf file21:50
and if not, to suggest this feature.21:50
→ ymasory joined21:51
Thor` it could be done through ln -s21:51
wereHamster Thor`: he'll probably wake up soon21:51
Thor` but both directories (the repository directory and the web directory) should belong to the same user.21:51
wereHamster: ok :)21:52
← dankest left21:52
Thor` and also, gitolite should be able to know the gitweb configuration location21:52
but that could be very great.21:52
← ISF left21:52
← bindaas left21:53
→ durarara joined21:56
← mishok13 left21:57
→ dcsobralWork joined21:57
→ RevoOf joined21:59
→ RaceCondition joined22:00
← skoop left22:00
← CannedCorn left22:00
← codeshah left22:03
← ymasory left22:06
← airborn left22:08
← serverhorror_ left22:13
← timemachine3030_ left22:14
← diegoviola left22:17
→ vvsh joined22:19
→ LiohAu joined22:23
← Dave^|| left22:26
← nils_r left22:26
tkil and now to see how hard the git list laughs at my patch...22:27
← pantsman left22:27
tkil http://fpaste.org/qfi8/ if anyone wants to get a head start on the giggling.22:27
← ThomasLocke left22:28
← EricInBNE left22:29
adamholt → adamholt_away22:29
adamholt_away → adamholt22:29
→ Tac_Home joined22:31
← Tac_Home left22:33
→ gremset_ joined22:34
← flowbee left22:36
← gremset left22:37
← vvsh left22:42
→ caseymcg joined22:43
← fcuk112 left22:45
← canton7 left22:46
← moonflux left22:47
→ BiggFREE joined22:47
→ angelsl joined22:48
← icwiener left22:50
→ g0bl1n joined22:56
→ juvenal joined22:57
→ nazgul102 joined23:00
← nazgul101 left23:03
← tewecske left23:03
→ ISF joined23:04
→ javaanse_jongens joined23:04
← bid left23:04
← AAA_awright left23:10
← pgib left23:12
→ linusoleander joined23:13
← javaanse_jongens left23:14
← foobar_ left23:14
← Jippi left23:15
→ RobertLaptop joined23:16
→ AAA_awright joined23:20
← Amorphous left23:20
→ bid joined23:20
→ ilteris___ joined23:21
← jatt left23:23
→ jonatasnona joined23:25
→ MikhX_ joined23:27
→ gigamonkey joined23:27
← MikhX left23:29
MikhX_ → MikhX23:29
f00li5h is there an FAQ somewhere? i'd like to know the least pesky way to run a git server23:29
jast as in several repositories with different users?23:32
Kobaz gitolit23:33
gitolite23:33
← pen left23:35
→ Amorphous joined23:37
← Zenopus left23:37
f00li5h paws at gitolite23:39
f00li5h jast: sure.23:39
it's just me comitting to a wc on my dev box, but the disk is starting to look shakey, and i'd liek to put the repo somewhere other than my exceptionally stealable laptop23:39
→ rchavik joined23:42
frogonwheels f00li5h: gitolite is great - (we use it), but I also have a script that backs up to a git clone --mirror clone of my repos.23:46
f00li5h paws at that locomotive anphibian23:47
Kobaz is there an asyncronous way to stream a file to a channel?23:48
er23:48
f00li5h hmm,23:51
if i move all this gitty stuff to an internet ... i suppose i should have some kind of web UI...23:51
is there one (that's not github)?23:51
peterhil f00li5h: There is gitourious and git-web, that you can install yourself23:52
f00li5h i can!23:52
peterhil At least gitorious offers private repos too.23:52
git-web is part of git23:52
← helgikrs left23:52
peterhil Just make sure you include it when doing config23:52
f00li5h hrm23:53
← kipras left23:54
jast for single-user stuff just using git via ssh and gitweb for web interface should be fine for you23:54
peterhil What OS? When installing from FreeBSD ports you can install quite a lot of dependencies to git ranging from git-web to SVN support etc23:54
f00li5h peterhil: gitorious is a bundled solution-on-a-website thing too, sin't it?23:55
peterhil It is, but you can find the sources with a little digging23:55
And you can then install it on your server23:55
→ banisterfiend joined23:55
→ Tac_Home joined23:56
← thiago left23:56
f00li5h peterhil: sneaky23:56
peterhil wait a sec, I can find the link23:57
← mino left23:57
← akkad` left23:57
peterhil http://cjohansen.no/en/ruby/setting_up_gitorious_on_your_own_server23:58
And getting the sources: git clone git://gitorious.org/gitorious/mainline.git gitorious23:58
f00li5h peterhil: very handy!23:59
thanks a bundle23:59
peterhil The first link has quite comprehensive guide23:59
f00li5h peterhil++23:59

Logs Search ←Prev date Next date→ Channels Documentation