IRCloggy #git 2010-02-25

Provider of IRC logs since 2005.

WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

	2010-02-25
yclian	"start_point" I guess.	00:00
tiglionabbit	I have a bare git repository on my local machine that I have checked out in a few places. Some of those places are owned by root. When I push things as root, it creates objects in the bare repository that are owned by root. Then, when I try to push things as a non	00:00
	-root user, it fails because it doesn't have permission to deal with some of these objects	00:00
	what should I do?	00:00
	→ j416 joined	00:01
	→ fool__ joined	00:01
	← fool__ left	00:01
	→ fool__ joined	00:01
	← eletuchy left	00:03
	→ jrmuizel joined	00:03
	→ Azeroth-Work joined	00:05
	← fool_ left	00:05
	→ __iron joined	00:05
SpookyET	Damn it. The server has 1.6.5. I installed 1.7.0 in my home folder. How do I tell the hooks use to use the one in my home? It obviously doesn't read .bash_profile nor .profile. The path is never set	00:06
	← bdon left	00:06
	← yclian left	00:07
j416	SpookyET: what does "which git" say?	00:08
	→ pblittle joined	00:08
	→ ijuma_ joined	00:08
	← _iron left	00:09
	← dj_tjerk left	00:09
	← JackWinter left	00:10
	← fr0sty__ left	00:10
SpookyET	j416: It works from the shell	00:11
	j416: I need it to work when i push	00:11
	I tried PermitUserEnvironment yes in ~/.ssh/environment but it's not working	00:11
	→ JackWinter joined	00:12
j416	well, you have a path problem, most likely	00:13
	← Chillance left	00:13
j416	server finds git that is installed in a "proper" location first	00:13
	I would have guessed it works by setting a proper path in .bashrc but apparently not?	00:14
SpookyET	j416 it doesn't launch bash	00:14
j416	ok	00:14
SpookyET	so path is never set	00:14
j416	ah of course.	00:14
	← ben_alman left	00:15
	→ ccfontes joined	00:15
ccfontes	hello	00:15
	← AAABeef left	00:15
ccfontes	I just forgot my password for my repo in gitfarm. anyone knows the dudes help email there?	00:16
	← bibinou left	00:16
SpookyET	j416: ideas?	00:18
	← maattd left	00:19
	→ ben_alman joined	00:19
j416	SpookyET: how do you start git-daemon?	00:19
	→ chrisadams joined	00:19
	← mrchrisadams left	00:19
j416	oh sorry	00:19
	that's not what you're using right	00:20
	I'd guess you would have to set the global path.. hm.	00:20
	I really have no idea. Sorry	00:20
aleray	is it possible to comment/annotate a commit ?	00:20
j416	aleray: in recent versions, there is git notes	00:21
	man git-notes	00:21
Gitbot	j416: the git-notes manpage can be found at http://git.or.cz/man/git-notes	00:21
j416	I never use it though.	00:21
	← TeckniX left	00:21
	← pblittle left	00:21
j416	I prefer to keep my info in the commit.	00:22
	(ensures it stays there forever)	00:22
	git notes are stored in a separate branch, I don't know what happens when you push etc	00:22
aleray	j416, awesome. Could it be used to comment other people commits (and not only one's commits)?	00:22
	→ froschi1 joined	00:23
	← giallu left	00:23
	→ Garoth joined	00:24
j416	aleray: of course	00:24
	→ scarabx joined	00:24
j416	I don't see why it wouldn't.	00:24
	← froschi left	00:24
	← rudi_s left	00:24
aleray	j416, nice!	00:24
SpookyET	detached working trees with submodules are pain in the arse	00:26
	← DavidKlein left	00:26
	← Tureba left	00:26
	→ jMyles joined	00:26
	jMyles is about to becoming a first time git user.	00:26
jMyles	*become :-)	00:26
	← jrmuizel left	00:26
j416	jMyles: welcome. :)	00:26
jMyles	After looking at the difference between Subversion (whose name and double-entendre I love) and Git, I'm fairly certain that Git is preferable for me.	00:27
	← justinlilly left	00:28
aleray	I'm working on a git interface to enable a non-technical audience to collaboratively write text. The difference with a wiki is thatpeople don't have to work on a single shared copy and do there is no need to reach any consensus. if you want to have a look it's there http://brainch.stdin.fr/ (it is more than alpha right now)	00:28
SpookyET	jMyles: You didn't evaluate mercurial nor bazaar?	00:28
j416	aleray: cool	00:28
jMyles	j416: Thank you. :-) So, my first task is to create a revision system for a Drupal installation. A blog post on the topic (http://mechanicalrobotfish.com/posts/130-getting-git-together-drupal) is of some usefulness to me, but it assumes some basics about Git that I don't yet... er... get.	00:28
	SppokyET: Neither of these names are known to me.	00:28
j416	jMyles: I advise you to get familiar with git first	00:28
	→ mtheoryx joined	00:29
SpookyET	jMyles: Git's competition	00:29
	→ fr0sty_ joined	00:29
j416	jMyles: http://book.git-scm.com/ < there's a good start	00:29
	→ lispy joined	00:29
j416	there is lots of other good documentation as well.	00:29
aleray	j416, the interface is very unclear right now but I was thinking how awesome it would be to be able to annotate this for instance: http://brainch.stdin.fr/jerome/OngoingManifesto/history/	00:29
jMyles	j416: Sounds great! :-) I'd like to take a toe-dipping approach - most of the materials I have found are too dense for me to get in to.	00:29
j416	jMyles: I'd go head-first	00:30
aleray	I would to love to have your feedback if anyone of you wants to take a look	00:30
	← bkaney left	00:30
lispy	I think I created a branch (and I think I even pushed the branch), but now when I pull I get a funny error message. My friend can pull (and he gets the branch I created) but no funny error messages.	00:30
j416	it really helps to understand the basic concepts	00:30
	head-first into the git documentation that is. :)	00:30
lispy	What is the correct way to tell git that it should pull in the remote branch that corresponds to my local branch?	00:30
SpookyET	j416: Don't be a bible salesman. Let the man choose his religion.	00:30
j416	aleray: well, annotations are of course local as well	00:31
	SpookyET: excuse me?	00:31
	← JackWinter left	00:31
	→ zehrique joined	00:32
aleray	j416, what do you mean by "local" ?	00:32
j416	I'm just pointing him in the right direction (he is asking for help...)	00:32
	→ JackWinter joined	00:32
j416	jMyles: but of course, check other systems as well. Git may not be optimal for you.	00:32
	aleray: everything in git is local. Then you push/pull to/from the world.	00:32
	as for your app, I'm not sure how you are doing things.	00:33
	if you have access to a local git repo, you can do anything with it, including add notes, of course.	00:34
	← chrisadams left	00:34
j416	but you can't do things directly with a remote repo. You work locally, then update the remote, or have the remote fetch from you.	00:34
	← JoeConyers left	00:35
	→ jrmuizel joined	00:35
	→ sumangurung joined	00:36
aleray	j416, right now every is handled by the app. There is in fact only one system user, i just force the user name in the cmd line. I use git-python	00:36
j416	ok	00:37
	so everything is "local"	00:37
	← evanluck left	00:37
j416	as far as git is concerned	00:37
	how do you separate the different users?	00:37
	each user has one branch?	00:37
	→ evan_luck joined	00:38
	← evan_luck left	00:39
aleray	j416, each user have a set of repo. a bit like github I guess except I'm not using gitolite or so	00:39
j416	ah I see.	00:39
aleray	j416, it is very "alpha" :-)	00:40
j416	heh	00:40
	everything starts with very alpha :)	00:40
	← kennethlove left	00:41
SpookyET	Git submodules are a pain. If you want to modify them. You have to host your own repositories for the submodules. Otherwise, when someone else clones, they'll get an error.	00:41
lispy	Anyone have some ideas about my question?	00:41
	(I'm guessing it's a common question)	00:42
bob2	"funny error message" isn't very specific	00:42
	pastebin the whole error (and command line)	00:42
	← mrlarner left	00:42
davr	lispy: This is probably what you want: http://www.gitready.com/beginner/2009/03/09/remote-tracking-branches.html	00:43
	← jrmuizel left	00:44
	→ Dark_Shikari joined	00:47
jMyles	OK, page one is reasonable cognizable to be. I understand what a blog, tree, and commit are.	00:47
	← amerine left	00:47
Dark_Shikari	let's say I accidentally abort a rebase or am and "lose" a commit I made. what's the best way to find it? git fsck shows so many commits that there's no way I'll be able to figure out which hash is the one I want.	00:47
bob2	index	00:47
Dark_Shikari	there is no command named index	00:47
bob2	that's right	00:48
	git reflog	00:48
davr	fyi it stands for ref-log, not re-flog like I thought at first :P	00:49
Dark_Shikari	um, reflog PRUNES old commits	00:49
	I don't want to do that	00:49
	that would kill my commit	00:49
	→ stringo0 joined	00:49
davr	it does other stuff to, read man git-reflog especially the 'show' subcommand	00:49
Gitbot	davr: the git-reflog manpage can be found at http://git.or.cz/man/git-reflog	00:49
bob2	davr: haha	00:49
Dark_Shikari	oh, sweet, thanks	00:50
	"show" works	00:50
	thanks.	00:50
	← Dark_Shikari left	00:50
	← bentob0x left	00:50
	→ extofme joined	00:51
extofme	how does the kernel git repo do an "motd"?	00:52
	there doesnt seem to be a hook for git-upload-pack or anything	00:52
	→ flaguy48 joined	00:52
davr	extofme: dunno how they do it, but I noticed anything I 'echo' in my 'post-recieve' hook gets shown to the client whenever they push	00:53
extofme	yeah i know you can do it when people push; i want to "redirect" people to github that are using my personal server	00:54
lispy	davr: that feature is probably what I want. I guess it can't be applied post facto to my branch?	00:54
extofme	like an "this project is now hosted here..."	00:54
	→ fool_ joined	00:54
	← parasti left	00:54
	← fool_ left	00:54
	→ fool_ joined	00:54
extofme	with the command to change the origin config to another repo/server	00:55
davr	lispy: I don't know if there's a command to do it, but you can do it by hand by editing .git/config -- don't know exact lines offhand, make a new tracking branch & see what gets put in there	00:55
	FFForever-Away → FFForever	00:55
	← fool__ left	00:55
	→ RobertMeta joined	00:56
lispy	davr: for now I set branch.<branchname>.remote to be the path to the remote repo	00:56
	→ jrmuizel joined	00:57
	← lightcap left	00:57
	← extofme left	00:57
	← shruggar_ left	00:58
	→ extofme joined	00:58
	→ bousquet joined	00:59
lispy	davr: for whatever reason that fixed my immediate need. I'm still a git beginner and quite confused by it (I'm a darcs dev so that is what I'm used to...just using git at work)	00:59
	← MetaCosm_ left	01:00
	← shruggar left	01:00
	← dacat left	01:04
	← hyperair left	01:08
	← tschundeee left	01:11
Ilari	{davr,lispy}: In 1.7.0: --set-upstream of branch and push.	01:13
	← lispy left	01:14
	← jrmuizel left	01:14
SpookyET	What's the best way to deal with submodules that you want to modify?	01:16
	← aleray left	01:16
	→ erikvold joined	01:17
	→ jrmuizel joined	01:19
	← WaltzingAlong left	01:21
	← dreiss left	01:22
frogonwheels	SpookyET: i believe the whole point of submodules is that in their directory, they behave just like another git repository.	01:25
	SpookyET: the point of them being submodules is being able to Commit a specific revision of the the submodule as being the currently active commit	01:25
	← ccfontes left	01:25
frogonwheels	.. for the parent project.	01:26
SpookyET	frogonwheels: that works	01:26
	problem is cloning	01:26
	if you clone the parent project, and try to init the submodules, it will whine it can't find the reference	01:27
	← justin-george left	01:28
	← jrmuizel left	01:28
	rmancy → rmancy_lunch	01:28
frogonwheels	ok. my current knowledge of submodules is theoretical :) I'm looking in to using them.	01:29
	← hebz0rl left	01:33
SpookyET	frogonwheels: It seems that if you want to modify a submodule, you have to clone the third party project, host it somewhere where it is accessible to you and to those who will clone your superproject. Then add your clone of the third party project as a submodule.	01:34
	Then make sure you push the submodule whenever you make changes.	01:34
	pain in the ass.	01:34
	Especially if you want to use git push to deploy.	01:35
	I'm looking into subtree merging instead.	01:35
	→ kaseano joined	01:35
	← kingfishr left	01:35
frogonwheels	Maybe submodules just isn't the right tool for that job.	01:35
kaseano	hi =) I know a couple writers that could really use something like git	01:36
	would git work for that or is there something more computer-illiterate-friendly	01:36
frogonwheels	kaseano: it _could_ work. there are plenty of guis around that do most simple stuff	01:36
kaseano	I just love git I want to share it w the world haha	01:37
	is there a git gui	01:37
frogonwheels	kaseano: well.. git gui for a start	01:37
kaseano	cus I'm sure the github interface is fine for a writer, the terminal's prob the hard part	01:37
SpookyET	frogonwheels: They would be if they were implemented properly. AKA, changes to submodules are stored in parent's repository separately without conflicts.	01:37
kaseano	haha oh ok	01:37
	I'll check that out then	01:38
	→ chewbranca_ joined	01:38
	← fda314925 left	01:38
	← defectiv left	01:38
SpookyET	unless i'm missing something	01:38
frogonwheels	SpookyET: .. or maybe submodules are good at what they are doing, and you need a different mechanism - submodule overlays or something	01:38
	← tiglionabbit left	01:39
	→ hyperair joined	01:39
	→ jmcantrell joined	01:39
SpookyET	submodules are full git repositories themselves.	01:39
	← spaceonline left	01:39
	← chewbranca left	01:40
frogonwheels	yes. presumably the problem is that the submodule entry has to reference the git source tho', so when people clone the parent, it knows where to get it.	01:41
	I can see what you're trying to do.. you want a kind of joinfs kind of approach.. where you have code from an initial repository, and you want to track changes you've made to it.	01:42
	→ bousquet_ joined	01:43
	← bousquet left	01:43
	bousquet_ → bousquet	01:43
	← bousquet left	01:43
frogonwheels	.. and so the problem is the reference in the sub-module to your modified code doesn't correspond to a reference on the original host	01:43
	.. I really think you just need to host the close of the directory somewhere and reference the hosted clone.	01:44
	- at the moment, I actually think that's the way it should be done.	01:44
	→ jinzhu joined	01:45
	← kaseano left	01:45
SpookyET	frogonwheels: that's what I said. It seems that a lot of people say that submodules are utter suckage.	01:45
	frogonwheels: I found http://github.com/apenwarr/git-subtree	01:45
	looks interesting	01:46
	→ threeve joined	01:46
frogonwheels	That sounds more like what you want.	01:48
	it's basically a nice way of hosting a library in your main git tree	01:49
	.. being able to export it, and maintain it as well.	01:49
	.. which is, as I said, not to say that submodules are bad, just not always the correct solution :)	01:50
SpookyET	frogonwheels: the chances that you use a third party lib that requires no modification are slim	01:51
frogonwheels	SpookyET: .. but I'm prepared to host the modified library seperately.	01:52
SpookyET	That's the complicated part	01:52
	← bugfux left	01:52
	frogonwheels shrugs.	01:52
	→ infogulch joined	01:53
frogonwheels	I gotta host the main project locally anyway, so hosting 2 projects isn't really much extra work.	01:53
	.. and my libraries aren't on git originally anyway.	01:53
	← zehrique left	01:54
SpookyET	But what if you wanted to use git as deployment tull?	01:54
	git push web	01:54
	Now, it becomes complciated.	01:54
frogonwheels	notreally, if you're prepared to (say) chuck the forked library up on githup.	01:54
	anyway. I think git-subtree is more suited to what you need.	01:55
	← mjf left	01:55
SpookyET	frogonwheels: not if the stuff you want to version has private informat, paypal credentials for example.	01:56
frogonwheels	(I'm not arguing that git submodules are right for you, I'm putting forward that git submodules may not be 'broken')	01:56
SpookyET	s/informat/information	01:56
	FFForever → FFForever-Away	01:56
	← kpreid_ left	01:56
frogonwheels	you'd have to push the submodule seperately I guess.	01:57
mkl1	Hi, I have a branch that have about 10 descendent branches. Now I would like to move this branch, and all the descendent branches, to a new base, and I might need to do so quiet frequently.	01:57
	My "dirty" way is to rebase step by step, it takes about 15-20 rebases, but the branches is somehow complicated, and this seems error-prone to me.	01:57
	Is there a clean way to do so?	01:57
frogonwheels	.. but I don't know if that works at the other end.	01:57
	← imajes left	01:58
	→ digitalice joined	01:59
digitalice	hi	01:59
	how do i do in git, the same that SVN UP does? :)	02:00
SpookyET	frogonwheels: to use submodules would require a tangled SSH mess. A) ssh to the server. (expected) B) ssh server to server. Annoying. Most likely have to setup ssh-agent manually	02:00
	← Arafangion left	02:00
	→ catalis joined	02:00
	→ imajes joined	02:01
Ilari	digitalice: Something like 'git pull' (there's no exact equivalent)?	02:01
	→ Arafangion joined	02:01
	rmancy_lunch → rmancy	02:01
	→ rsewill joined	02:02
	← altogether left	02:03
	← aziz left	02:05
	→ loincloth joined	02:05
	→ aziz joined	02:05
	→ kpreid_ joined	02:07
	← catalis left	02:08
	← rsewill left	02:08
	→ consolers joined	02:08
	→ kssddy joined	02:10
	→ JoeConyers joined	02:11
	→ urbanmonk joined	02:12
	→ justin-george joined	02:12
	→ kingfishr joined	02:12
	← hummingkid left	02:12
	→ altogether joined	02:14
	→ catalis joined	02:15
	← tedoc2000_ left	02:17
	← ttvd left	02:17
	→ ttvd joined	02:18
	→ tedoc2000 joined	02:18
	→ jaysern joined	02:18
	← digitalice left	02:20
SpookyET	oh, wow the 152 git executables still exist in git-core	02:20
	I guess git is just a wrapper	02:20
	← owen1 left	02:21
	→ tedoc2000_ joined	02:21
	← tedoc2000 left	02:22
	tedoc2000_ → tedoc2000	02:22
Ilari	SpookyET: Almost 100 of those are hardlinks to main git binary.	02:22
	→ mrlarner joined	02:24
	→ simplechat joined	02:25
Ilari	SpookyET: Here link count of main git binary is 104... There are 42 executables that are not hardlinked to multiple names...	02:26
	← Alagar left	02:26
	← hyperair left	02:26
	→ eletuchy joined	02:26
SpookyET	42 more to go...	02:26
	← spearce left	02:26
	← agib left	02:27
Ilari	SpookyET: Many of those 42 are not even written in C...	02:28
	→ VVSiz_ joined	02:28
	→ anjen joined	02:29
	← Arafangion left	02:29
	← nizuri left	02:29
	← tedoc2000 left	02:30
	← jaysern left	02:30
Ilari	And some of those that are can't be easily made builtin. Such as git-http-{fetch,push} (it would be needed to be merged to remote-curl group, not to git group).	02:31
	→ shtylman joined	02:32
	← VVSiz left	02:32
	← catalis left	02:33
	← Mage_Dude left	02:33
	← __iron left	02:33
	← magn3ts left	02:34
	← incluye left	02:35
	→ hced joined	02:37
	→ Cynos joined	02:38
consolers	"another myth is that git is huge with a hundred different programs (ls git * \| wc -l => 148). actually most of these are copies of each other with a different name and there are only like 75 different programs" -- at least in 2007	02:38
	→ catalis joined	02:39
consolers	if you dont preserve the symlinks your git installation becomes bigger than windows95	02:39
	→ Mage_Dude joined	02:39
consolers	if you dont preserve the *hardlinks your git installation becomes bigger than windows95	02:39
	→ binjured joined	02:40
	← catalis left	02:40
	← mtheoryx left	02:41
Cynos	Hi all, I'm having an issue that's new to me - I have a feature branch Foo created from master. In Foo two new files were created - A & B. Now, merging master into Foo deletes A & B.	02:41
cbreak	-rwxr-xr-x 104 root wheel 1,4M 16 Feb 03:52 git	02:42
	does that mean I have 104 hard links to this file?	02:42
Cynos	My guess is that at some point Foo got merged back into master in such a way that master is now more authoritative when it comes to the merge algorithm	02:42
	but I'm not sure of the actual cause (because I really want to avoid this happening again), and definitely not sure of the actual fix beyond a very painful manual merge.	02:43
	← JohannesSM64 left	02:43
Cynos	I've just been trying Git 1.6	02:43
	gah	02:43
cbreak	you can do git merge --no-commit	02:43
	and fix the removals away	02:43
Cynos	I've just been trying Git 1.7's new merge stuff on the off-chance, but using -s recursive -X ours has the same result.	02:43
	← naiad left	02:43
	← Mage_Dude left	02:44
Cynos	cbreak - yeah, I figure that's going to be the end result, but is my guess on the cause correct?	02:44
	→ Mage_Dude joined	02:44
	← AkumaStreak left	02:44
cbreak	in my opinion, which is not funded by anything besides user experience, it shouldn't delete them.	02:45
	← Mage_Dude left	02:49
	→ Mage_Dude joined	02:49
consolers	cbreak ya it'd be the `st_nlink' from the stat	02:50
	"whatever that is"	02:51
	→ jepster1 joined	02:51
	→ darkfaery joined	02:51
	→ Kelly_zhou joined	02:52
	← jepster left	02:53
	→ melter_ joined	02:54
	→ bdon joined	02:57
	← sumangurung left	03:00
	→ hekin joined	03:00
hekin	question, I specify 'foobar' should be ignored in the .gitignore, but with 'git status', I can still see it	03:00
Cynos	hekin - has it already been staged before you added it to .gitignore?	03:01
cbreak	is it tracked?	03:01
	git doesn't ignore things it tracks	03:01
hekin	Cynos: ahhhh	03:01
	cbreak: yes, great point	03:01
	→ naiad joined	03:04
frogonwheels	hekin: I spent a few minutes yesterday trying to hide some files.. before realising they were in git.	03:05
	← bdon left	03:05
hekin	frogonwheels: seems we have the same problem, but what am I supposed to do to 'untrack' these files?	03:05
frogonwheels	hekin: you can delete them from the repository at that point.	03:06
hekin	frogonwheels: you mean with 'git remove' ?	03:06
frogonwheels	hekin: it is _possible_ to rewrite the history without those files - but that's only workable if you're not sharing your tree anywhere :)	03:07
	hekin: I have a few files I might do that to once I figure out the commands to use	03:07
hekin	frogonwheels: thanks for the tips	03:08
Cynos	hekin - I use an alias like so in my ~/.gitconfig: forget = rm --cached	03:09
	→ rgr joined	03:09
Cynos	so "git rm --cached fileX fileY"	03:09
frogonwheels	Cynos: forget - nice alias	03:10
hekin	Cynos: but this will remove the files from the repository, which is shared by me and other guys	03:10
frogonwheels	hekin: ahh you just want to ignore them locally?	03:10
hekin	frogonwheels: yes	03:10
Cynos	errr	03:11
hekin	I don't want to remove them from the repository, I just don't want to track the changes	03:11
	sorry, guys, I didn't make myself clear enough	03:11
Cynos	so they're version controlled files	03:11
	but you don't want Git to track changes to them on your box?	03:11
hekin	yes	03:11
Cynos	um	03:12
frogonwheels	hekin: A in, You don't want to see them in git gui as changes?	03:12
hekin	yes	03:12
	IIRC, there's alias unstage	03:12
	but I don't remember the other part of the alias	03:12
	→ neunon joined	03:13
	← naiad left	03:17
	→ Joe_Conyers joined	03:17
	← JoeConyers left	03:18
	← Joe_Conyers left	03:18
cbreak	you can't track a file and at the same time not track it	03:19
Cynos	git reset -- stagedFileToNotStage	03:20
	but you'll have to do that each time	03:20
	→ JoeConyers joined	03:20
	← darkfaery left	03:22
	← cytrinox left	03:24
	← tliff left	03:24
	← jas4711 left	03:24
	→ tliff joined	03:24
	← binjured left	03:25
	→ cytrinox joined	03:25
	→ godsflaw joined	03:25
	frogonwheels → frogonwheels_	03:26
	→ Arafangion joined	03:26
	← jmcantrell left	03:26
	← kukks left	03:30
	← johnw left	03:31
	← Arafangion left	03:32
	← JoeConyers left	03:32
	→ JoeConyers joined	03:33
	→ agib joined	03:34
	← godsflaw left	03:34
	→ codeswing joined	03:38
	→ Arafangion joined	03:38
	→ mtheoryx joined	03:40
	← mtheoryx left	03:40
	→ mtheoryx joined	03:40
	← remi left	03:41
	→ jas4711 joined	03:41
	← coolcat left	03:42
	← mssm left	03:43
	← SlickT10 left	03:44
	→ ico2 joined	03:44
ico2	hi	03:44
	is there a way to remove the most recent couple of commits from a remote git tree? accidentally included sensitive info in the commit message	03:45
	← flaguy48 left	03:45
	← mtheoryx left	03:46
	→ TTilus joined	03:48
	→ fda314925 joined	03:48
ico2	i tried git reset, but it won't let me push	03:49
	← codeswing left	03:49
j416	ico2: reset your local branch and force-push it	03:49
	you need to do git push -f ( = git push --force)	03:49
ico2	trying j416	03:49
j416	be sure that no one has fetched from it	03:49
	and be sure that you want to make the remote look like yours	03:49
	← agib left	03:49
ico2	j416: thanks	03:50
j416	(if someone has fetched from it, they will be confused as to why the repo has changed)	03:50
	and they will have trouble with it	03:50
ico2	noone has fetched	03:50
j416	ok :)	03:50
	then it is fine	03:50
	← smuf left	03:51
j416	if it is sensitive info, you might want to prune the commit objects from the remote	03:51
	I'm not sure they are pruned automatically when you force-push	03:51
	maybe they are.	03:51
	→ tildeequals joined	03:53
	← mason-l left	03:56
	→ cloudhead joined	03:58
	→ hyperair joined	03:58
	← j416 left	03:58
	→ catalis joined	04:01
	→ merlin1978 joined	04:02
	← ph^_ left	04:02
	← Twisted_Mentat left	04:02
	← hced left	04:03
	→ ph^ joined	04:04
	← hipe_ left	04:07
	→ ehutzelman joined	04:08
	→ mason-l joined	04:11
	← mkl1 left	04:12
	← scarabx left	04:12
	→ t0rc joined	04:12
	← fr0sty_ left	04:15
	← mrlarner left	04:16
	← ehutzelman left	04:16
	← melter_ left	04:16
	→ scarabx joined	04:16
	→ fr0sty_ joined	04:17
	→ fool__ joined	04:19
	← fool__ left	04:19
	→ fool__ joined	04:19
	← aresnick left	04:20
	← altogether left	04:20
	→ dreiss joined	04:21
	→ timuruski joined	04:22
	← fool_ left	04:22
	← Cynos left	04:24
	← rmancy left	04:25
	← numberGrey left	04:26
	→ hipe joined	04:26
	→ flaguy48 joined	04:26
	→ munichlinux joined	04:30
	← extofme left	04:37
	→ aresnick joined	04:39
	← aziz left	04:42
	→ trs21219 joined	04:42
	← trs21219 left	04:47
	← disappearedng__ left	04:49
	← disappearedng left	04:50
	← aresnick left	04:50
	→ trs21219 joined	04:54
	← VVSiz_ left	04:55
	→ _VVSiz_ joined	04:55
	← consolers left	04:57
	→ tchan joined	04:58
	→ frogonwheels__ joined	04:59
	← frogonwheels_ left	05:00
	→ ArdRigh joined	05:01
	← mr_daniel left	05:03
	← trs21219 left	05:04
	→ disappearedng__ joined	05:04
	→ disappearedng joined	05:05
	→ mr_daniel joined	05:05
	→ SecretAgent joined	05:06
	→ justinlilly joined	05:10
	→ consolers joined	05:11
	← Bass10 left	05:12
	← SpookyET left	05:13
	← fr0sty_ left	05:16
	→ fr0sty_ joined	05:17
	→ bcardarella joined	05:19
	→ disappearedng_ joined	05:20
	← catalis left	05:21
	← smotchkkiss left	05:21
	← rgr left	05:23
	← cenuij- left	05:25
	→ agliodbs joined	05:25
agliodbs	if I have a merge conflict, how do I pull out the version of the file I had before the attempted merge?	05:25
	for just one file	05:25
	it's a very long SQL query, and the conflict markers are NOT helpful :-(	05:26
	← mason-l left	05:28
	← lanthan__ left	05:28
	← consolers left	05:28
	→ vu3rdd joined	05:29
	← t0rc left	05:29
	← pheaver left	05:31
	→ loinclot_ joined	05:31
	→ AkumaStreak joined	05:32
	← AkumaStreak left	05:33
	← loincloth left	05:33
offby1	agliodbs: I think there's a special syntax for it ... lemme look	05:34
	git show :1:filename shows the common ancestor, git show :2:filename shows the HEAD version and git show :3:filename shows the remote version.	05:35
	man git-merge	05:35
Gitbot	offby1: the git-merge manpage can be found at http://git.or.cz/man/git-merge	05:35
agliodbs	offby1: thanks	05:36
	offby1: I'm noticing that the conflict marker system only really works for procedural code	05:37
	not sure what I'd replace it with for SQL, though	05:39
	← threeve left	05:41
	→ mason-l joined	05:43
EvanR	i dont know how but my repo became out of sync with me	05:43
	→ loincloth joined	05:43
EvanR	can i git reset --hard on it?	05:43
	← loinclot_ left	05:43
EvanR	remote repo	05:43
agliodbs	out of sync with you?	05:44
	← dreiss left	05:44
EvanR	like its HEAD and my HEAD^, which i thought were the same commit, arent. they have a different hash	05:44
	so wheni tried to push, it failed	05:44
	no idea why that happened	05:45
	but i just need to reset --hard the remote repo and push again	05:45
	← fr0sty_ left	05:45
	→ rgr joined	05:45
offby1	by definition, HEAD^ and HEAD are different. Why would you think they're the same?	05:46
	← merlin1978 left	05:46
EvanR	because i committed once	05:46
	→ fr0sty_ joined	05:47
offby1	I would expect HEAD^ to be all zeroes in that case	05:47
	just tried it: git ss HEAD^ => fatal: ambiguous argument 'HEAD^': unknown revision or path not in the working tree.	05:48
	EvanR: so, you mean that your repo has exactly one commit object?	05:48
EvanR	it has exactly a lot of commit objects	05:49
	more than one	05:49
	→ roop joined	05:49
EvanR	different repos can have the same commit, but they wont all be HEAD	05:49
	its HEAD for me and HEAD^ for the not yet updated repo	05:50
	ok, git push --force may be the answer	05:51
offby1	oh, you're talking about two different repositories.	05:51
EvanR	yes	05:52
	my original question was typod	05:52
offby1	it's kind of a bad idea to "push" to a repo that has a checkout	05:52
	faq non-bare	05:52
	faq non bare	05:52
	faq nonbare	05:52
	sigh	05:52
EvanR	has a checkout?	05:52
offby1	the bot's supposed to present a URL describing that FAQ	05:52
EvanR	i dont think the repo is non bare	05:52
offby1	oh really?	05:52
EvanR	which means strictly speaking it has no HEAD	05:52
offby1	yeah	05:52
	so ....	05:52
EvanR	i was just trying to describe it	05:53
offby1	you must simply mean a particular ref is at one commit, whereas the corresponding ref on your box is at some other one	05:53
	what happens when you try a simple "push" without the --force?	05:53
	→ alesan joined	05:53
EvanR	first ill explain the situation again	05:53
	→ warthog9 joined	05:53
offby1	you said it "failed", but what exactly did it print?	05:53
alesan	hi... I'm moving my first steps with git... I've checked out a repository, now I've put some files in it and I would like to send it to the server	05:54
	how do I do that?	05:54
EvanR	there are two commits on my system, A and B, B after A. the remote repo should have A, but instead, it has C which is otherwise the same as A	05:54
bob2	alesan: commit and push	05:54
EvanR	so pushing B definitely doesnt work	05:54
bob2	alesan: bear in mind that push does NOT update the working copy on the other side	05:54
offby1	EvanR: what do you mean by "C which is otherwise the same as A"? Either they're the same, or they're not.	05:54
bob2	(unless it's on the same machine)	05:54
offby1	EvanR: maybe you need to "git pull" or "git pull --rebase"	05:54
alesan	bob2, the other side is gitorious...	05:54
EvanR	offby1: if i get pull, i will have a duplicate commit	05:55
offby1	EvanR: perhaps "--rebase" then.	05:55
bob2	alesan: ok	05:55
EvanR	my repo is correct	05:55
offby1	(and in fact, you might try "git pull" without --rebase; you might be surprised)	05:55
EvanR	i did that to test	05:55
offby1	s/surprised/pleasantly surprised/	05:55
	oh	05:55
EvanR	it merges the C into my repo	05:56
offby1	EvanR: I don't suppose this is a public server?	05:56
	← lemonchicken left	05:56
EvanR	yeah its public, ... gah	05:56
offby1	ah.	05:56
	then I can take a look	05:56
	it's so hard describing this stuff in text ... be nice if there were an easy way to draw circle-and-arrow diagrams, like in "Alice's Restaurant"	05:56
EvanR	also means if anyone pulled recently whatever i do will screw things up	05:57
	http://repo.or.cz/w/cantaveria.git im on branch audio	05:57
offby1	ok, hole on	05:57
alesan	bob2, so I have git add * and git commit	05:57
offby1	gonna clone that	05:57
	hope it's not too big	05:57
EvanR	its small	05:57
	→ project2501a joined	05:57
bob2	alesan: yes	05:58
	alesan: which git tutorial are you reading atm?	05:58
alesan	http://git-scm.com/ I am tring to get info out of this	05:58
offby1	EvanR: OK! I'm looking at gitk now.	05:58
	← hyperair left	05:58
bob2	ah	05:58
alesan	bob2, and the instructions I got from the gitorious page	05:58
offby1	EvanR: So I get the feeling that you want to avoid committing a merge commit. Is that right?	05:58
EvanR	yes, especially one with a commit that has the same message as another and is supposed to do the same thing	05:59
offby1	EvanR: you're _sure_ that git won't elide that commit?	05:59
EvanR	elide... elide	06:00
offby1	eliminate	06:00
	skip	06:00
	silently drop	06:00
EvanR	no, it considers it a non fast forward merge	06:00
offby1	hm	06:00
	so, I've got the repo from repo.or.cz; can you send me your patches?	06:00
	like 'git format-patch audio'	06:01
EvanR	if you and repo.or.cz agree to git reset --hard to e984b3b9fd63e that would be great	06:01
alesan	fatal: protocol error: expected sha/ref, got ' The git:// protocol is read-only.	06:01
	what could be the cause for this?	06:01
offby1	EvanR: y'know you _can_ force repo.or.cz to reset, I think	06:01
	that's "git push --force", as you suspected	06:01
	never done it though	06:01
bob2	alesan: you need to push to a ssh url	06:01
	or maybe webdav	06:02
	→ psankar joined	06:02
offby1	EvanR: only downside is you'll need to warn people who may already have dc07	06:02
EvanR	that commit has been there for a while though, and now that i think of it	06:02
	people have downloaded that commit	06:02
	→ greg8si joined	06:02
offby1	ah well	06:02
alesan	bob2, but gitorious page says: Your push url: [email@hidden.address]	06:02
EvanR	so i guess i need to pull	06:02
bob2	alesan: yes	06:02
EvanR	but before i do, i need to save my new commit, reset to before	06:02
	then commit again	06:02
	then push	06:02
bob2	alesan: did you add 'git://' to the beginning of that?	06:02
EvanR	god dammit	06:03
offby1	EvanR: that's confusing, but that's OK	06:03
alesan	bob2, no, what do you mean...	06:03
offby1	EvanR: just checking: you know about "gitk --all", right?	06:03
EvanR	but my new commit has B as the parent	06:03
offby1	hugely useful	06:03
EvanR	so a merge is unavoidable	06:03
bob2	alesan: pastebin your whole command + the whole error (http://paste.pocoo.org/)	06:03
EvanR	yes gitk --all	06:03
	← munichlinux left	06:03
offby1	EvanR: which commit are you referring to as B? I'm gonna tag it so I can keep track :)	06:03
	and A and C for that matter	06:03
EvanR	B is on my computer	06:03
offby1	ok	06:04
	if it's _only_ on your computer, then you can safely throw it out	06:04
	and make another commit with the same patch, but based where you want.	06:04
EvanR	ah	06:04
alesan	bob2, http://pastebin.ca/1809161	06:04
EvanR	how do i do that	06:04
offby1	well, first you might want to "git format-patch" to turn it into a genuine patch	06:04
	or save the affected files or something.	06:04
EvanR	ill do the patch to learn something	06:05
alesan	bob2, http://pastebin.ca/1809162	06:05
offby1	(strictly speaking none of that is necessary; the reflog preserves everything, but you might not know how to use the reglog)	06:05
	reflog	06:05
alesan	I have aded the gitorious instructions	06:05
bob2	alesan: run the commands they tell you to	06:05
EvanR	how would i use the reflog to get my new changes back	06:05
alesan	I did and that is the result :(	06:05
bob2	alesan: I don't think so	06:05
	alesan: cat .git/config	06:06
	→ girishr joined	06:06
alesan	well how can I start fron scratch?	06:06
	← scarabx left	06:06
offby1	EvanR: well, "git reflog audio" will show you ever value that the "audio" ref has ever had, even if nothing currently points at the commit	06:06
EvanR	right	06:06
offby1	which means that if you (say) reset, or rebase, that branch, and then change your mind, you can just reset it right back to where it was.	06:06
	very comforting :)	06:06
alesan	bob2, http://pastebin.ca/1809166	06:06
EvanR	and i would want to 'rebase' one of them to my future corrected HEAD	06:06
	→ taufiqkh joined	06:06
bob2	alesan: you ran some other command at some point	06:07
	alesan: git remote add origin [email@hidden.address]	06:07
alesan	fatal: remote origin already exists.	06:07
	→ munichlinux joined	06:08
bob2	because it poitns at the read only version	06:08
alesan	:(	06:08
bob2	git remote rm origin ; git remote add origin [email@hidden.address]	06:08
EvanR	gah	06:08
	← marc_kdab left	06:09
EvanR	offby1: how do i get patch to send you of my local repo	06:09
alesan	bob2, :) thank you very much	06:09
offby1	EvanR: one way is to give me a (read-only) URL to it	06:09
alesan	now it's work	06:09
offby1	if such a thing exists.	06:09
EvanR	first thing is to generate it	06:09
offby1	EvanR: another way is to run "git-format-patch REF", where ref is something that I already have (like remotes/origin/audio), then mail me the output	06:09
EvanR	git format-patch audio produces nothing	06:10
offby1	EvanR: if your repo is accessible, that'd be the easiest way: then I'd just get another branch	06:10
	well, "audio" is probably your _local_ branch	06:10
EvanR	yes	06:10
	is that not what you want	06:10
offby1	I'd try "git checkout audio && git format-patch --stdout remotes/origin/audio > send-to-offby1"	06:10
	← krawek left	06:11
offby1	but like I say: if your repo is accessible, then just a URL to it would work	06:11
EvanR	its not	06:11
	NAT	06:12
offby1	ah	06:12
EvanR	that command produces a very large patch starting at initial commit	06:12
offby1	really?	06:12
	hmm	06:12
EvanR	hold on	06:12
	typed it wrong	06:13
	well i have a patch	06:13
	ill pastebin it	06:13
offby1	great	06:13
	→ loinclot_ joined	06:14
EvanR	http://pastebin.ca/1809173	06:14
	→ pingveno joined	06:14
EvanR	i can see i did something horribly wrong with one of the two karplus commits	06:14
	cant be mine, because mine works!	06:14
offby1	naturally	06:15
	hole on, lemme grab that	06:15
	← loincloth left	06:16
offby1	OK, that added one commit: Sequencer simplified, basically working.	06:16
	now what?	06:16
EvanR	well look at karplus commit	06:16
	its somehow different from the corresponding one on the repo.or.cz one	06:16
offby1	er?	06:17
	I _have_ the repo.or.cz one	06:17
EvanR	yes	06:17
	?	06:17
offby1	everything I have is from repo.or.cz, except for _one_ commit, "Sequencer simplified, basically working."	06:17
EvanR	fast forward?	06:17
	ah	06:17
offby1	as I suspected, git ignored the first of your two patches	06:18
EvanR	the patch encompasses 1) the broken commit and 2) the new stuff	06:18
offby1	presumably because it wasn't needed	06:18
EvanR	alright then lets start over	06:18
	im going to run git push	06:18
offby1	ok	06:18
EvanR	(i figured the patch was for you see what my repo looked like)	06:19
offby1	you could just "git checkout audio && git reset --hard refs/remotes/audio" and then replay the patch	06:19
	then you'd have what I have	06:19
	which I suspect is fine	06:19
	when I said "refs/remotes/audio" I should have said "remotes/origin/audio"	06:20
EvanR	hrm	06:20
	well push fails. non fast forward, i would lose history, it wasnts to me make a merge commit	06:20
	you say i can apply this patch	06:20
	how to apply it	06:20
offby1	"git checkout audio && git am -3 the-patch-file"	06:21
EvanR	3?	06:21
offby1	yep	06:21
	don't ask :)	06:21
	when I did that, here's what I saw: http://gist.github.com/314290	06:21
EvanR	it failed	06:21
offby1	aw	06:21
	← kingfishr left	06:22
EvanR	http://pastebin.ca/1809180	06:22
offby1	hole on	06:22
EvanR	i think i should have reset farther or something	06:22
	instead of to audio HEAD	06:22
offby1	another idea: "git bundle create whatever remotes/origin/audio", then get me _that_ file	06:23
	that should send me the actual commit objects, not just a patch	06:23
EvanR	do you want the origin or mine?	06:23
	→ marc_kdab joined	06:23
offby1	on the other hand: perhaps the conflict will be easy to resolve	06:23
	EvanR: I don't understand your question	06:23
EvanR	what are you trying to get from me?	06:24
offby1	the exact state of your repo	06:24
	I assume it's very close to what I have, but not quite the same	06:24
EvanR	yeah why would a bundle of remotes/origin/audio be that	06:24
	← vu3rdd left	06:24
EvanR	origin is repo.or.cz	06:24
offby1	that command should give you everything that you have that remotes/origin/audio doesn't.	06:24
EvanR	oh	06:25
offby1	try "git bundle"; if the result isn't too big, send it to me somehow	06:25
	email would be fine	06:25
	if it's huge, then it was a bad idea	06:25
	← cloudhead left	06:25
EvanR	usage:...	06:25
	git bundle create <file> ...	06:26
offby1	well, git bundle create whatever remotes/origin/audio	06:26
	gotta go in a few minutes	06:26
EvanR	could not open remotes/origin/audio	06:26
offby1	drat	06:26
EvanR	:((	06:26
offby1	you specified some file name for "whatever"?	06:26
	e.g., "git bundle send-me-to-offby1 remotes/origin/audio" ?	06:27
EvanR	git bundle foo remotes/origin/audio	06:27
offby1	hmph	06:27
	guess neither of us can read man pages, then	06:27
EvanR	i think i can resolve this on my own, now that i realize i cant just reset the public repo	06:27
offby1	ok	06:27
	smack	06:28
	git bundle _create_ send-me-to-offby1 remotes/origin/audio	06:28
	offby1 rolls eyes	06:28
	← alesan left	06:28
EvanR	166k	06:28
	pm email addr?	06:29
offby1	but like I say I am probably outta here in a few minutes	06:30
EvanR	yeah	06:30
	well its sent	06:30
offby1	heh	06:31
	got it and fetched from it; but I don't think it added any commits!	06:32
	sigh	06:32
EvanR	i clearly have two karplus commits	06:32
	two different hashes	06:32
offby1	one's dc07	06:33
	the other one -- what are its parents?	06:33
EvanR	1cc38	06:33
	parent e984b3b	06:33
	← stringo0 left	06:33
offby1	interesting	06:33
EvanR	indeed!	06:33
	← greg8si left	06:33
offby1	so the two karplus commits are siblings?	06:33
EvanR	same parent	06:34
offby1	yeah	06:34
	→ hipe_ joined	06:34
	→ amerine joined	06:34
offby1	is one better than the other in some way?	06:34
EvanR	i have no idea!	06:34
	what is the difference?	06:34
	← hipe left	06:35
offby1	I mean, can you just ignore the 1cc38 one? Pretend you don't even have it?	06:35
	_I_ don't have it :)	06:35
EvanR	i just rewrote the sequencer based on it	06:35
offby1	anyway ...	06:35
	→ greg8si joined	06:35
EvanR	i guess i need to rebase	06:35
	locally	06:35
offby1	as threatened: I'm outta here. Good luck, and I'll leave my IRC client open, because I'm curious to see what you wind up doing.	06:35
EvanR	alright	06:36
	← munichlinux left	06:36
	→ vu3rdd joined	06:37
	Morasique → kitties	06:37
	kitties → Morasique	06:38
	→ loincloth joined	06:38
	← loinclot_ left	06:40
	← _VVSiz_ left	06:41
	→ VVSiz joined	06:42
	→ jokaro joined	06:43
	← Adlai left	06:45
	← taufiqkh left	06:46
	→ Adlai joined	06:46
	→ loinclot_ joined	06:47
	← loincloth left	06:49
	→ jjuran joined	06:49
	← warthog9 left	06:49
	← greg8si left	06:51
	→ psoo joined	06:52
	→ jrmuizel joined	06:53
EvanR	offby1: i reset to the common parent, pulled from repo.or.cz, checked out the four or five files from the reflog commit, then committed again with a copy of the original commit message	06:53
	→ loinclo__ joined	06:54
EvanR	then pushed	06:54
	things are sane again	06:54
	← loinclot_ left	06:56
	← Adlai left	07:05
	← adamm left	07:06
	→ sde_antrix joined	07:07
	→ Theravadan joined	07:08
	→ kingfishr joined	07:08
	← webchick left	07:09
	→ hyperair joined	07:15
	→ zjason` joined	07:17
	← ankit9 left	07:18
	← zjason left	07:19
	← project2501a left	07:22
	← pingveno left	07:24
	← Theravadan left	07:24
	→ thiago_home joined	07:24
	→ priidu joined	07:25
	→ numberGrey joined	07:25
	→ jaysern joined	07:27
	→ pingveno joined	07:28
	→ zjason`` joined	07:29
	← zjason` left	07:30
	→ munichlinux joined	07:30
	→ d0ugal joined	07:30
	← Pupeno left	07:32
	← ph^ left	07:33
	→ ph^ joined	07:34
	→ ph^_ joined	07:34
	→ Pupeno joined	07:36
	→ godsflaw joined	07:37
	← Pupeno left	07:37
	→ Pupeno joined	07:38
	← Textmode left	07:38
	← pingveno left	07:40
unijunction	i stashed my changes, then merged with a new commit, now i'd like to merge my changes	07:41
	← Mage left	07:41
	→ d1b joined	07:43
	→ Azeroth-Working joined	07:44
	→ bentob0x joined	07:45
	← godsflaw left	07:45
	← Azeroth-Work left	07:45
	→ godsflaw joined	07:45
	← kssddy left	07:46
tango_	wth: git log --follow -p git-gui/lib/shortcut.tcl	07:47
	doesn't give me anything	07:48
	← jaysern left	07:48
	← ph^_ left	07:49
	→ project2501a joined	07:51
	→ Azeroth-Work joined	07:52
	← Azeroth-Working left	07:52
	→ Axius joined	07:54
	→ giallu joined	07:54
	→ steffkes joined	07:57
	← project2501a left	07:58
	← thiago_home left	07:58
	→ mtkd joined	08:00
	→ mastro joined	08:01
	← marc_kdab left	08:01
	→ Azeroth-Working joined	08:02
	→ ph^_ joined	08:02
	← Azeroth-Work left	08:03
	← Axius left	08:04
	→ webus joined	08:05
	← fr0sty_ left	08:06
	→ fr0sty_ joined	08:07
	← girishr left	08:08
	→ Azeroth-Work joined	08:08
	→ ankit9 joined	08:08
	→ radarek joined	08:09
	← Azeroth-Working left	08:10
	→ jaysern joined	08:11
	→ Adlai joined	08:12
	← Azeroth-Work left	08:13
	→ Azeroth-Work joined	08:13
	→ killerchicken joined	08:13
	← Azeroth-Work left	08:15
	→ Azeroth-Working joined	08:15
	← JoeConyers left	08:16
	→ marcus__ joined	08:17
	→ Azeroth-Work joined	08:17
	← Azeroth-Working left	08:18
	← roop left	08:21
	→ roop joined	08:21
	← PuffTheMagic left	08:22
	→ thiago_home joined	08:24
	← loinclo__ left	08:25
	← mithro left	08:28
	→ Adlai` joined	08:29
	→ skoop joined	08:30
	← amerine left	08:30
	← Adlai left	08:31
	Adlai` → Adlai	08:31
	← timuruski left	08:33
	→ loincloth joined	08:35
	← niki left	08:35
	→ sumangurung joined	08:36
	← sumangurung left	08:36
	← jrmuizel left	08:36
	→ jrmuizel joined	08:36
	← jrmuizel left	08:37
	← loincloth left	08:39
	→ lucsky joined	08:39
	→ jelmer joined	08:40
	→ DavidKlein joined	08:41
ArjenL	tango_: Try it without 'git-gui/'	08:41
	← jaysern left	08:42
	← fool__ left	08:42
tango_	aha	08:42
ambuli	hi all, how can i revert back the changes to one commit before. ?	08:42
tango_	git log --follow -- lib/shortcut.tcl works	08:42
	→ ixti joined	08:43
tango_	thanks ArjenL. is that because git-gui is subtree-imported?	08:43
	→ galderz joined	08:43
ArjenL	tango_: Yeah, I think it is, dunno why though (haven't really looked into it)	08:43
	→ fool_ joined	08:44
	← fool_ left	08:44
	→ fool_ joined	08:44
	← ixti left	08:45
	→ TheNewAndy joined	08:45
	← thiago_home left	08:45
_rane	tango_: aaaaaaaaaaa, merrrrrge	08:46
tango_	_rane: sorry, merge what?	08:46
kkoehne	ambuli: You mean you just want to get rid of the last commit?	08:46
	← jjuran left	08:46
	→ _iron joined	08:47
kkoehne	ambuli: And you haven't shared your repo / pushed the commit up so far? then git reset --hard HEAD^ should work	08:47
_rane	tango_: at least my reaction plugin fix, it's broken	08:48
	→ sumangurung_ joined	08:48
	← sumangurung_ left	08:48
	← erikvold left	08:50
	→ gebi joined	08:51
ambuli	kkoehne, i have pused into remote repo also.	08:51
kkoehne	ambuli: Then reverting your commit is the only save thing to do (others might have pulled the commit from the remote repo). Do a git revert HEAD	08:52
	ambuli: This will still leave the commit in the history too	08:52
	← tildeequals left	08:54
	→ acoul joined	08:56
	← justin-george left	08:56
	← bcardarella left	08:57
	→ bcardarella joined	08:58
	→ ntoll joined	08:58
ambuli	kkoehne, so git revert HEAD will revert the last commit ?	08:59
	→ trtzrdb joined	09:00
kkoehne	ambuli: Yes, HEAD is a placeholder for the last commit	09:01
ambuli	kkoehne, i pulled the package, and ran git revert HEAD, and then i can i push to remote repo ?. git status -a does not show nothing to commit ?	09:01
kkoehne	ambuli: git revert HEAD did ask you for a commit message, right?	09:02
	qq	09:02
ambuli	yes it asked	09:02
kkoehne	ambuli: Just have a look at the history (git log): Should show you a new commit named 'Revert "My original commit message"'	09:02
	→ dj_tjerk joined	09:03
kkoehne	ambuli: So yes, if you push you'll push this second commit which just undoes all the changes of your first one.	09:03
	→ shruggar joined	09:04
ambuli	kkoehne, ok, so to push to master, git push origin master right ?	09:04
kkoehne	ambuli: Most likely :)	09:05
	→ Azeroth-Working joined	09:05
	← Azeroth-Work left	09:05
ambuli	kkoehne, but i tagged the commit which i want to revert back also. do i have to revert back tag as well ??	09:07
kkoehne	ambuli: Ouch. There is no easy way.	09:07
	ambuli: Have a look at the section "On Re-tagging" in the git tag man page	09:08
	ambuli: e.g. at http://www.kernel.org/pub/software/scm/git/docs/git-tag.html	09:08
	→ beni- joined	09:09
	← Azeroth-Working left	09:12
	→ Azeroth-Work joined	09:12
	→ tbf joined	09:12
	→ One-Man-Bucket joined	09:12
	m-takagi → m-takagi_	09:15
	→ traceback0 joined	09:16
traceback0	If you are on (no branch) and commit something and then do git branch <branch>--can you get those changes ?	09:17
	← ArdRigh left	09:18
	→ mrchrisadams joined	09:19
	← bcardarella left	09:19
	← fr0sty_ left	09:21
	→ fr0sty_ joined	09:22
One-Man-Bucket	I'm using git svn and I have 7 revisions in svn which i'm supposed to review. How can I create a diff on just those 7 revisions?	09:23
	→ JohnFlux joined	09:25
JohnFlux	Hey all	09:25
	how do I do a git blame on a different branch?	09:25
	I can't figure out the magical syntax to specify sha and filename to git blame	09:26
	and the man page doesn't say	09:26
	→ themill joined	09:27
JohnFlux	oh, just git blame SHA filename	09:27
shruggar	though it's one of those things where there's probably 800 ways to do it	09:28
	git blame SHA:relative-to-root-filename, for example	09:28
JohnFlux	could someone add an example to the git documentation :-)	09:29
	→ tobiassjosten joined	09:29
	→ tatsuya joined	09:32
	← numberGrey left	09:32
	→ Ramosa joined	09:32
	→ ixti joined	09:34
	← hyperair left	09:35
	→ tschundeee joined	09:35
	← JackWinter left	09:36
charon	shruggar: the : form most certainly doesn't work	09:36
	← anjen left	09:36
charon	shruggar: you'd be giving blame a blob, which doesn't really make sense as blobs don't have history	09:37
themill	I created a bare repo "git init --bare; chmod a+x hooks/post-update" and pushed "git push machine master" and browsing the remote repo with gitweb shows the contents of that push. But when I "git clone http://machine/project.git/" I am told "warning: remote HEAD refers to nonexistent ref, unable to checkout." What important step did I miss?	09:37
	→ JackWinter joined	09:37
	→ anjen joined	09:37
	← tbf left	09:39
	← JackWinter left	09:40
	→ JackWinter joined	09:41
	→ mtsirkin joined	09:41
shruggar	charon, ah, of course. Ignore me as per usual :D	09:41
	→ tbf joined	09:42
charon	JohnFlux: 'git blame -h' seems to be fairly concise, but i'm sure an example patch would be welcomed anyway	09:42
	← JackWinter left	09:43
	→ GarethAdams\|Work joined	09:43
charon	shruggar, JohnFlux: it's really not that hard: rev:path specifies a blob, while rev -- path specifies both a rev and a path. so all the commands that operate on both a revision and a path need the separated variant. the colon variant is only for commands that can use blobs directly like git-show, git-cat-file etc	09:43
	the only problem is that such reasoning is not in the manpages and i'm not sure where you'd put it	09:43
	→ abstortedminds joined	09:44
	→ hyperair joined	09:44
mtsirkin	any tips on keeping history in a state that will make git bisect work well?	09:45
	→ JackWinter joined	09:45
mtsirkin	looks like if e.g. build is broken in versions A..B I won't be able to bisect in that range without manual tweaking ...	09:46
One-Man-Bucket	hmm.. http://dl.dropbox.com/u/1704324/git-diff.png	09:46
	→ inffcs00 joined	09:47
charon	mtsirkin: if you know the build is broken because of always the same reason, you can have your 'git bisect run' script use 'git cherry-pick -n $build_fix' and compile with that	09:47
mtsirkin	charon, why would it be broken because of always the same reason?	09:48
	→ malesca joined	09:48
	← vu3rdd left	09:52
mtsirkin	charon, would be nice to attach bugfix to a range somehow, so that each bug gets its fix ...	09:53
charon	mtsirkin: well i don't know your use-case do i? you said your build was broken in A..B.	09:53
	you can use rebase -i to fix the entire range if you want	09:53
	and then bisect over the rebased copy	09:53
	← dj_tjerk left	09:53
mtsirkin	yes, this is what I end up doing.	09:53
	problem with rebase is it break merges	09:54
	break->breaks	09:54
charon	feel free to pick up the half-done 'rebase -i -p' rework from dscho's repo	09:54
mtsirkin	what does it do?	09:55
charon	-p is --preserve-merges	09:55
	but the current state of it is rather broken	09:55
	though it might be easier to make sure that your stuff compiles at every commit ;)	09:55
mtsirkin	well compiles is only half the story :)	09:56
	→ afloret joined	09:56
afloret	hello	09:56
mtsirkin	what I mean by breaks merges is that if someone merges my tree, and I rebase, we have a problem	09:56
charon	mtsirkin: oh, you should only rebase temporarily for the bisect	09:57
afloret	someone can tell me the command to see the latest commit revision with git ?	09:57
charon	you can do that on a detached HEAD or a scratch branch	09:57
afloret	git log ? with an option ?	09:57
mtsirkin	charon, then I have to redo it again for next bisect	09:57
	I thought about using git merge -s ours	09:58
charon	mtsirkin: i suspect you can make a bisection repo that backports fixes for known things, and use 'git replace' (or grafts) to put them in place	09:58
	→ chrisadams joined	09:58
charon	afloret: the useful half of the answer is 'git show'. the not so useful half is, "latest revision" is an extremely ill-defined concept in the context of a distributed vcs	09:58
	→ bluenovember joined	09:58
	← mrchrisadams left	10:00
	← disappearedng left	10:00
	← disappearedng__ left	10:01
	→ wagle joined	10:01
charon	afloret: so you might actually mean "the latest upstream commit" or some such, in which case the answer would be different.	10:02
afloret	charon: yes it's that the latest commit	10:02
	← anjen left	10:02
charon	afloret: try git fetch && git show @{upstream}	10:02
afloret	sorry, i start with git, normally i use clearcase or perforce	10:03
charon	if your git is too old for @{upstream}, you'll have to guess the branch; frequently it's origin/master	10:03
	← inffcs00 left	10:03
	→ inffcs00 joined	10:04
mtsirkin	charon, will git replace and/or grafts work well with others pulling my tree?	10:04
	← chrisadams left	10:05
charon	mtsirkin: they won't affect anyone else. i'd still recommend keeping them confined to a separate "bisection only" so as to not confuse yourself too much (and to keep the effects of any potential replace bugs limited)	10:05
	→ mrchrisadams joined	10:05
	← malesca left	10:05
afloret	charon: your commande give me two different informationq	10:05
mtsirkin	charon, well, what bothers me is duplication of effort this all causes	10:06
	← disappearedng_ left	10:06
charon	mtsirkin: duplication between you and others you mean?	10:06
mtsirkin	charon, what about git merge -s ours ORIG_HEAD?	10:06
afloret	charon: it seems that fetch give me the server and the branch were come from	10:06
mtsirkin	charon, yes ... everyone needs to bisect at some point - the best person to fix bisect for everyone would be the one that introduced the bug ...	10:07
afloret	charon: and show upstream seems give me the latest commit	10:07
	is right ?	10:07
	→ mbroeker joined	10:07
	← TheNewAndy left	10:07
charon	afloret: that's the point i was trying to make. git is distributed, so you have to answer "where am i looking for the latest commit?" yourself. you could mean the last commit on your current branch, or the last commit in your repo, or the last upstream commit on the "upstream" of your current branch, or the last commit in the remote repo overall. how should i know what the correct answer is? but what commands to use depends on this.	10:08
	→ TheNewAndy joined	10:08
charon	mtsirkin: dunno. i suspect (but haven't tested) that you'll need to use 'git bisect start <whatever> -- .' to filter the "do-nothing" merge. i wouldn't use it anyway, because you'd have to repeat that over and over with every bug fixed, causing a huge mess of a history.	10:09
	mtsirkin: replacements can be shared by manually making push/fetch specs for them, if that helps	10:10
	again i'd keep that confined to a "bisection only" repo to keep the original history sane/non-confusing	10:10
afloret	charon: ok, in fact i have done a clone from a git repository and i need to know the latest commit upstream on my clone	10:10
	← ixti left	10:10
mtsirkin	charon, history simplification seems to work well enough ... not sure I care about "mess" if it is not visible	10:12
charon	afloret: in that case 'git show @{upstream}' was the right answer, but note that 'gitk ..@{upstream}' is probably more useful	10:12
afloret	ok charon thanks for your answer :)	10:13
	← Kelly_zhou left	10:13
	→ hebz0rl joined	10:13
mtsirkin	charon, when you say "manually" does everyone who pulls need to do it too? If it's only me it might be ok ...	10:14
charon	mtsirkin: everyone. you need a fetchspec for refs/replace/* as those aren't usually fetched	10:15
JohnFlux	charon: could you just add a quick one line example to git blame --help	10:15
impulze	i'm in the middle of a rebase and i want to overwrite my current working dir with the contents of another commit	10:15
	← trtzrdb left	10:15
impulze	i.e. i want git-merge to just overwrite my whole working dir with "theirs"	10:15
	is that possible?	10:15
JohnFlux	charon: it would take much longer to check out the latest git, make a patch, figure out how to send it, etc, then for you to just add a one liner example	10:16
charon	JohnFlux: please send an email to [email@hidden.address] so i won't forget. i won't have time to hack on patches till at least tomorrow night, probably longer.	10:16
	(a one-liner is fine)	10:16
JohnFlux	thanks	10:16
mtsirkin	charon, for replacements - so what do I make a replacement of what? rebased tree a replacement of original one?	10:18
charon	mtsirkin: it's enough to make the tip of the rebase a replacement of the tip of what you were fixing	10:18
	→ jhelwig joined	10:18
JohnFlux	could someone do: git blame --help	10:18
	and tell me what they see for the small example?	10:19
	→ Paraselene__ joined	10:19
JohnFlux	I cannot understand what it's saying, so I'm wondering if it's a parsing error or something	10:19
charon	JohnFlux: that "small example" is just pointing you at the -S option to git-log as "also helpful if you're digging for history"	10:19
	i agree that it's pretty misleading	10:20
JohnFlux	I think the example text should be something different from "blame_usage" :-)	10:20
	→ disappearedng joined	10:21
	← Azeroth-Work left	10:22
	→ Azeroth-Work joined	10:22
	→ disappearedng__ joined	10:22
	→ fool__ joined	10:23
	← fool__ left	10:23
	→ fool__ joined	10:23
	→ JonathanT joined	10:25
	→ ixti joined	10:25
mtsirkin	charon, hmm so ... if someone merges in the middle or the rebased range, history will still me messy, won't it?	10:25
	or -> of	10:25
	→ mssm joined	10:25
charon	mtsirkin: well you'd have to rebuild the merge if you want an accurate replacement	10:26
	← orafu left	10:26
	→ orafu joined	10:26
	→ aatanasov joined	10:26
charon	or maybe i'm not understanding right... you mean, merges the middle of the rebased range to elsewhere?	10:26
	← fool_ left	10:26
JohnFlux	does this make sense:	10:27
	→ Determinist joined	10:27
JohnFlux	+ Say the line you are interested in has been changed in a minor way, masking the commit that really added that code.	10:27
	+ You can view the SHA that git blame gives for that line, then redo the blame starting from the commit before:	10:27
	+ git blame SHA~ myfile.txt	10:27
	→ trtzrdb joined	10:28
JohnFlux	I'm just thinking about how to improve the git blame documentation	10:28
	→ ludde joined	10:29
charon	yes, that's the standard mode of history drilling. git blame ... ; git show $sha; git blame $sha^ ...	10:29
ludde	i'm unable to delete a branch on the server, the 'git push origin :foobar' says error: unable to push to unqualified destination: foobar	10:29
charon	i think git gui blame has buttons to simplify that, but i've never been able to get comfortable with it	10:29
JohnFlux	charon: is it sha^ or sha~ ?	10:30
	I always forget	10:30
charon	same thing	10:30
JohnFlux	..	10:30
charon	^n is the n-th parent, defaulting to the first	10:30
	~n is the n-th first-parent ancestor, defaulting to the first	10:30
	so they overlap at ~ and ^	10:30
	← multani left	10:32
JohnFlux	I can't seem to understand what you're saying :-D	10:32
	first parent ancestor..	10:33
	JohnFlux scratches his head	10:33
	← trtzrdb left	10:33
	← hyperair left	10:33
JohnFlux	charon: the difference is only if there is a merge or something?	10:33
	charon: Okay I sent you a small email	10:34
charon	no. please read the "specifying revisions" section in man git-rev-parse. i'll be glad to answer questions after that, but i don't really have time for the stuff that is also in manpages	10:34
Gitbot	charon: the git-rev-parse manpage can be found at http://git.or.cz/man/git-rev-parse	10:34
JohnFlux	charon: np	10:34
charon	and yes, it sucks that this very important information is in the documentation for a scripting command that should not be needed in normal user operation. suggestions for where else to put it welcome.	10:35
	→ Azeroth-Working joined	10:38
	← abstortedminds left	10:38
	← Azeroth-Work left	10:39
	→ whitman joined	10:40
	→ ferdy joined	10:40
	← frogonwheels__ left	10:41
	→ frogonwheels joined	10:41
	← David-T left	10:44
	→ marc_m joined	10:44
	marc_m → marc_kdab	10:44
	→ trtzrdb joined	10:45
	→ Azeroth-Work joined	10:46
	← Azeroth-Working left	10:48
	← JohnFlux left	10:48
	← hipe_ left	10:50
	← TheNewAndy left	10:50
	→ hipe joined	10:50
	← Azeroth-Work left	10:50
mrchrisadams	hi guys, is anybody using gitosis here?	10:54
bob2	best to just ask	10:54
	→ syret joined	10:56
	← mtkd left	10:57
	→ hyperair joined	10:57
	← TDJACR left	10:57
	→ mtkd joined	10:59
mrchrisadams	bob2 I want to setup a postcommit hook to trigger a shell script on another server	11:00
	by caling ssh user@domain -c bash somescript.sh	11:00
	but I can't see anywhere what user git runs these hooks as	11:00
	so I'm not sure where to generate a keypair for the relvant user	11:01
Aides	your gitosis user of course	11:01
	which is 'git' for most of the cases	11:01
	what url are you using to push?	11:01
mrchrisadams	Aides you're right	11:01
	i push [email@hidden.address]	11:02
	but I don't see a git user in the home directory	11:02
charon	mrchrisadams: the very point of gitosis is to block shell access. so you'll probably have to go through 'sudo su - git' or similar to gain one.	11:02
mrchrisadams	where I would expect to see home/git, along with home/username	11:02
Aides	try grep ^git /etc/passwd	11:02
charon	(and as always, ~git is the home of user git)	11:02
	Aides uses /srv/git, for example	11:02
	→ warlock_mza joined	11:03
	← Haarg left	11:03
	← syret left	11:05
mrchrisadams	ah.. that makes sense. thanks Aides and charon!	11:05
	← giallu left	11:08
	← jfkw left	11:10
	→ ribasushi joined	11:10
	← JonathanT left	11:10
	→ giallu joined	11:11
	→ MetaCosm joined	11:12
	← giallu left	11:13
	→ giallu joined	11:13
	→ aziz joined	11:14
	← ico2 left	11:17
	← robinsmidsrod left	11:19
	→ lemonchicken joined	11:19
	→ killu joined	11:20
	→ TDJACR joined	11:22
	→ JonathanT joined	11:23
	← JonathanT left	11:25
themill	What do I need to do so that "git clone http://machine/project.git/" works rather than giving "warning: remote HEAD refers to nonexistent ref, unable to checkout."? Google tells me this is a common problem but never seems to present an answer.	11:25
	→ JonathanT joined	11:25
	→ ehabkost joined	11:25
	← TDJACR left	11:27
	→ remi joined	11:29
	→ dytrivedi joined	11:29
	← beni- left	11:29
	→ TDJACR joined	11:29
	→ Theravadan joined	11:30
	← hyperair left	11:30
	← riz2 left	11:31
	← chaoyi left	11:32
	← killerchicken left	11:32
	← rvsjoen left	11:34
	→ rvsjoen joined	11:35
	← TDJACR left	11:35
	→ agugsu joined	11:36
	← jepster1 left	11:37
	→ TDJACR joined	11:38
	← kinabalu left	11:38
	→ kinabalu joined	11:38
	→ maattd joined	11:40
	← TDJACR left	11:42
	→ Yuuhi joined	11:43
	→ d0k joined	11:44
	← psankar left	11:46
	← rgr left	11:46
	← webus left	11:47
	→ mithro joined	11:49
	← Theravadan left	11:49
	← urbanmonk left	11:49
	→ mklappstuhl joined	11:50
	→ project2501a joined	11:52
	→ TDJACR joined	11:52
	→ jmut joined	11:52
	→ hyperair joined	11:53
	→ niko joined	11:54
	← fool__ left	11:55
	← hekin left	11:55
	→ fool_ joined	11:57
	← fool_ left	11:57
	→ fool_ joined	11:57
	→ PureVirtual joined	11:57
	→ wereHamster joined	11:57
	← project2501a left	11:59
	← tghw left	11:59
	→ tghw joined	11:59
	→ k23z__ joined	12:00
	→ pinnacle joined	12:00
k23z__	hi, does anyone have a recommendation for some pdf git tutorial ?	12:00
	I want to read it offline	12:00
	and I want it to contain all the stuff I'll usually need	12:00
	anything like that around ?	12:00
	I don't want an entire book	12:00
	I'm not a sysadmin or something , I'm just a guy who wants to use git	12:00
	← inffcs00 left	12:01
	→ inffcs00 joined	12:02
	→ sagsousuke joined	12:02
	← tschundeee left	12:02
shruggar	k23z__, are you looking for a quick reference cheat sheet?	12:03
	→ rgr joined	12:04
k23z__	shruggar: yes but something I can learn from also	12:04
	← justinlilly left	12:04
k23z__	shruggar: a quick reference cheat sheet will refresh my memory(given that I have already covered those)	12:04
	I use svn ... I'm not a power user of svn atm ...	12:05
	but I do commit/status/log/checkout/branch , those kinds of things	12:05
	but that's about it	12:05
	I never did a merge for example	12:05
shruggar	http://book.git-scm.com/book.pdf	12:06
	→ Azeroth-Work joined	12:06
shruggar	I also always recommend new users read this: http://eagain.net/articles/git-for-computer-scientists/	12:06
	← segher left	12:07
	→ AAA_awright_ joined	12:10
	→ robinsmidsrod joined	12:10
	→ vinse_ joined	12:10
	← mrchrisadams left	12:10
	← kim-day left	12:10
	← twmi left	12:10
	← AAA_awright left	12:10
	← vinse left	12:10
	← tokkee left	12:10
	← DesertZarzamora left	12:10
	← fusepilot left	12:10
	← torben left	12:10
	← skyl left	12:10
	← ultraviolet left	12:10
	← Weiss left	12:10
	← przemoc left	12:10
	← Elfix left	12:10
	← mmattice left	12:10
	→ hekin joined	12:11
	→ torben joined	12:11
	→ mstormo joined	12:12
	← Azeroth-Work left	12:12
	→ DesertZarzamora joined	12:13
	→ Blackshark joined	12:14
	→ jarh joined	12:14
	→ tokkee joined	12:14
	→ maxb joined	12:14
	→ ozzloy joined	12:14
	→ maiOo joined	12:14
	→ EvanR-work joined	12:14
	→ broonie joined	12:14
	→ chrizz joined	12:15
	→ tpope joined	12:15
	→ daed joined	12:15
	→ Pev_away joined	12:15
	→ esc joined	12:15
	→ majoh joined	12:15
	→ dkr21 joined	12:15
	→ ruxpin joined	12:15
	→ metellius joined	12:15
	→ dyre joined	12:15
	→ MarcWeber joined	12:15
	→ runasand joined	12:15
	→ djpig joined	12:15
	→ trodrigues joined	12:15
	→ siprbaum joined	12:15
	→ Weiss joined	12:15
	→ ultraviolet joined	12:15
	→ cxreg joined	12:15
	→ chuck joined	12:15
	→ csgeek joined	12:15
	→ Elfix joined	12:15
	→ Cap_J_L_Picard joined	12:15
	→ LotR joined	12:15
	→ twmi joined	12:15
	← twmi left	12:15
	→ twmi joined	12:15
	→ com4 joined	12:15
	→ mebus joined	12:15
	→ dionoea joined	12:15
	→ bdash joined	12:15
	→ przemoc joined	12:15
	→ seanius joined	12:15
	→ kupsch joined	12:17
	← bentob0x left	12:17
	← dionoea left	12:18
	→ dionoea joined	12:18
	→ mrchrisadams joined	12:19
	→ Juice10 joined	12:19
	→ ebneter joined	12:19
	→ Bucciarati joined	12:19
	← trtzrdb left	12:22
	→ awx joined	12:26
	← k23z__ left	12:26
	→ mmattice joined	12:27
	→ skyl joined	12:27
	→ mEkmEk joined	12:27
	→ jjore joined	12:27
	← mr_daniel left	12:29
	→ tschundeee joined	12:29
	→ suzak joined	12:30
	→ maxter joined	12:30
	← JonathanT left	12:32
	→ JonathanT joined	12:33
	→ disappearedng_ joined	12:34
	← JonathanT left	12:35
	→ gitte joined	12:35
	→ corni joined	12:39
	→ nanoo joined	12:40
	→ Ryback_ joined	12:40
	→ SlickT10 joined	12:42
	← jinzhu left	12:46
	← froschi1 left	12:47
	← hekin left	12:48
	→ froschi joined	12:50
	← mbroeker left	12:53
	← sagsousuke left	12:54
	→ eli joined	12:54
	→ remi__ joined	12:57
	→ vu3rdd joined	12:58
	→ mbroeker joined	13:01
	← aziz left	13:06
	→ aziz joined	13:09
	→ parasti joined	13:12
	← jMyles left	13:13
	← inffcs00 left	13:17
	→ nizuri joined	13:18
	← JackWinter left	13:19
charon	themill: log in on that machine, go to the project and 'git symbolic-ref HEAD refs/heads/some_existing_branch'	13:20
	→ ksn joined	13:20
	→ CAAAN11 joined	13:21
	→ VAAGGID joined	13:21
themill	charon: Most excellent. I could have sworn I tried that already. Actually, using a newer git from Debian sid was able to do the clone straight off.	13:22
	(btw some_existing_branch was just "master"... I won't pretend to understand why that wasn't automagically done)	13:23
charon	that's odd, usually 'git init' sets HEAD to master	13:25
	maybe it was some extremely old version that didn't, i don't know the quirks of those	13:25
themill	1.5.6 isn't that old :)	13:25
	Anyway, seems solved now which is good.	13:26
	→ mst__ joined	13:27
	→ jjuran joined	13:28
	← mtsirkin left	13:29
	→ Pinaraf joined	13:32
Pinaraf	hi	13:33
	← mtkd left	13:33
Pinaraf	I've got a small problem migrating my workflow to qit, and I couldn't find any answer on the Internet :	13:34
	is there a «direct» equivalent of svn externals in git ?	13:34
	→ JackWinter joined	13:34
	→ mtkd joined	13:35
Pinaraf	I mean, when I clone my repository, I'd like it to clone a given git repository too	13:35
	in a subfolder	13:35
	← munichlinux left	13:35
	← warlock_mza left	13:35
charon	Pinaraf: git-submodules are sort of the equivalent, but you'll at least have to 'git svn update --init'	13:36
	(man git-submodule)	13:36
Gitbot	charon: the git-submodule manpage can be found at http://git.or.cz/man/git-submodule	13:36
	← roop left	13:36
Pinaraf	it says it references a given commit id	13:36
	→ munichlinux joined	13:36
mastro	can someone explain me why the git svn --commit-url option is discouraged ? I have 2 urls for the same SVN, one is public and one can only be used inside my company. i'm tracking the public url but it doesn't allow the COPY command so I want to make a single commit to the internal url and then switch back to the other one.. How could I do it?	13:36
charon	Pinaraf: also note that while externals can and usually do refer to the newest revision of the external, a submodule is always a fixed commit and to "update" it you must make a superproject commit that records the new submodule version	13:36
	this is arguably the right way to do it, but meh	13:37
	you could also look into the 'repo' tool that the android folks use	13:37
Pinaraf	charon: I just found git remote, could that be useful ?	13:37
	→ sumangurung joined	13:37
	→ psankar joined	13:38
mastro	reading the manual it say: Using this option for any other purpose (don’t ask) is very strongly discouraged. What's that supposed to mean???? "don't ask"?? why shouldn't I ask??	13:38
	→ justinlilly joined	13:39
charon	Pinaraf: no, that's just to manage the remotes you push/fetch	13:39
Pinaraf	isn't it possible to say «hey, fetch from that remote in that folder, thanks» ?	13:40
	← munichlinux left	13:41
charon	mastro: read the commit message that introduced it: ba24e74	13:41
Gitbot	[git ba24e74]: http://tinyurl.com/yfg8teq -- git-svn: add ability to specify --commit-url for dcommit	13:41
mastro	charon, i'm reading it.. it's dangerous.. but it is more dangerous if you don't explain how to use it and what could happen...	13:42
Pinaraf	charon: and there is no way to automatically update a submodule automatically ?	13:44
charon	mastro: i tend to side with eric... if git-svn(1) documented every possible way to shoot yourself in the foot with git-svn and explained why that happens, it would be unreadable	13:44
	Pinaraf: no	13:44
	→ Saur joined	13:44
mastro	charon, well.. now I need it and I don't know what's the proper way to use it and what could happen... so what should I do?	13:44
	→ frogonwheels_ joined	13:45
	→ munichlinux joined	13:45
	← frogonwheels left	13:46
	← seezer left	13:46
charon	mastro: so what's the problem? the manpage says the option was designed for the case where you have two different URLs to access the same repository.	13:46
	→ aresnick joined	13:46
charon	which seems to be your use-case.	13:46
mastro	charon, that's my case but I have to switch to the other repo for 3-4 commits then switch back to the main repo. and I tried locally using that commit-url parameter (with a test svn server) and it seems to ignore it	13:47
	→ seezer joined	13:48
	← dytrivedi left	13:48
mastro	charon, i've an svn locally: http://localhost/svn/repos/Foo and i want to commit on http://myHostAliasToLocalhost/svn/repos/Foo. but if i give it the commit-url it do the commit but in log i see the url with localhost	13:48
	← steveire left	13:49
	→ altrux joined	13:50
charon	mastro: in what log?	13:50
mastro	charon, let me pastebin it	13:50
	→ steveire joined	13:50
	← steveire left	13:50
	→ steveire joined	13:50
	← ben_alman left	13:51
	← disappearedng_ left	13:51
	→ auslander joined	13:52
	← auslander left	13:52
	→ auslander joined	13:52
mastro	charon, http://pastebin.com/dC9Y579L	13:52
	← radarek left	13:52
mastro	charon, i'd expect the last commit to say git-svn-id in "mastro" instead it keep using localhost	13:53
	→ catalis joined	13:53
	← psankar left	13:53
wereHamster	you commit to a different location, but still fetch from the old one	13:53
charon	mastro: on the contrary, those git-svn-id lines must match the URLs used in the config, otherwise git-svn would not be able to match config and commits	13:54
mastro	charon, so you are saying it is using the other one?	13:55
	charon, hum.. so if i put there a wrong url it should complain	13:55
	→ psankar joined	13:56
	→ ARos joined	13:57
	→ Soliton joined	13:58
mastro	charon, crossing my fingers	13:58
	→ NET\|\|abuse joined	13:58
	→ Bass10 joined	13:58
	→ Nies joined	13:58
	→ philsturgeon joined	13:58
philsturgeon	hey guys. I just wrote an article on using Git for deploying websites through various environments. could I get some feedback? http://philsturgeon.co.uk/news/2010/02/Deploying-websites-with-Git	13:59
	this is not intended as advertisment, as i am guessing most of you guys will know this already. I just want some feedback before the twittersphere piles onto it.	14:00
	→ jakemarsh joined	14:02
mastro	charon, it worked.. thanks for the support :)	14:02
ARos	Question for a generous soul: are there any tools for migrating a 'portion' of a repository to an empty repository? Specifically, I'm looking to migrate all files matching a specific glob or regex (and its git SCM metadata) to the new repository. Is this feasible?	14:02
hipe	ARos: absolutely!	14:02
	← justinlilly left	14:03
ARos	hipe: please oblige me	14:03
charon	philsturgeon: doesn't work. the push does not update the worktree (by design)	14:03
hipe	ARos: get really well acquinted with this one here: http://www.kernel.org/pub/software/scm/git/docs/git-filter-branch.html	14:03
charon	philsturgeon: please refer to the faq at tinyurl.com/gitfaq	14:03
PerlJam	philsturgeon: what charon said .	14:03
	→ radarek joined	14:03
philsturgeon	it seems to work for me	14:03
	i pushed changes, it updated the site	14:03
ARos	hipie: interesting... I was looking at git-fast-export\|git-fast-import as an alternative, but it complains about the lack of marks on the import side.	14:03
charon	philsturgeon: you'll want to make a post-receive hook that does something along the lines of cd "$GIT_DIR"; export GIT_DIR="$(pwd)"; export GIT_WORK_TREE=path/to/worktree; git reset --hard	14:04
	← _Vi left	14:04
hipe	ARos: several times, i decide that i need to move a sub-portion of my repository, with history, to a new repository. so the steps usually are 1) clone your existing repo, 2) move all the files you want in your new subrepo into a new folder, (or do that thing where you make it looklike a subfolder was the root folder the whole time) ...	14:04
PerlJam	philsturgeon: Prove it.	14:04
hipe	ARos: each time you perform these filter branch commands, be sure to select the --prune option. also at each step check your source files, a few times some strange things happened to me	14:05
charon	philsturgeon: i highly doubt it since the patches that ever implemented this were never merged, and even if you have them, you'd have to configure them first. if it really works (which i doubt) you have a post-update or post-receive hook in place.	14:05
ARos	hipe: would you provide an example 'git-filter-branch' statement that would make use of a glob or regex?	14:06
	→ ben_alman joined	14:07
charon	ARos: try --index-filter 'rm -f "$GIT_INDEX_FILE"; git reset -- "*.foo" \|\| true'	14:08
	untested, but that would be my angle of attack	14:08
	→ fool__ joined	14:08
	← fool__ left	14:08
	→ fool__ joined	14:08
ARos	charon: awesome!!!	14:09
	→ mklappst1hl joined	14:10
	← sumangurung left	14:11
teukka	is there a git command for asking the path (relative or absolute, does not matter) of the worktree root?	14:11
	← hipe left	14:11
	→ FilipeMaia joined	14:12
	→ hipe joined	14:12
charon	ARos: wait, made a mistake	14:12
philsturgeon	charon: http://pastie.org/842221 but it doesnt actually add the file	14:12
charon	needs to use $GIT_COMMIT	14:12
FilipeMaia	How can I checkout the revision just next to the current one? The opposite of HEAD~1?	14:12
philsturgeon	charon: this is what you mean about not updating the working tree right?	14:12
charon	ARos: as in, git reset $GIT_COMMIT -- "*.foo"	14:12
	← fool_ left	14:12
charon	philsturgeon: indeed. as i said, read the relevant FAQ article (look for "push" in the "why..." section)	14:12
	← mklappstuhl left	14:13
ARos	charon: ok, so you were missing a $GIT_COMMIT?	14:13
steffkes	hm .. search an windows gui with nice rendered git tree's any suggestions? already tried QGit & Git Extensions .. both not the best result for my test: http://progit.org/figures/ch5/18333fig0525-tn.png	14:13
parasti	teukka: git rev-parse --show-cdup	14:13
charon	ARos: yes	14:13
	ARos: the logic being that git-filter-branch doesn't affect HEAD, so whenever you want to refer to the currently filtered commit in a filter, that's $GIT_COMMIT	14:13
	← mklappst1hl left	14:14
teukka	parasti: thanks! i looked in the rev-parse manual but missed that... =)	14:14
	→ JoeConyers joined	14:14
	→ mklappstuhl joined	14:15
	← corni left	14:16
ARos	charon: while i certainly see that the index has been rewritten, I do not see files after checking out any of the newly-indexed commits	14:17
yhager	I made a few commits on my local master, and now I want to push them to a new branch at origin (and then track it locally). how can I push to a new branch?	14:17
	← NET\|\|abuse left	14:18
	→ NET\|\|abuse joined	14:18
teuf	yhager: git push origin master:remotebranchname	14:18
	→ Mezner joined	14:19
charon	ARos: what's the exact command you used?	14:19
	i.e. the full invocation	14:20
	← PureVirtual left	14:20
philsturgeon	charon: so jamming a git reset --hard HEAD in there will update the files?	14:20
	charon: i understand that hard resets can wipe local changes which is fine	14:20
yhager	teuf: thx	14:21
charon	philsturgeon: yes	14:21
ARos	charon: git-filter-branch --index-filter 'rm -f "$GIT_INDEX_FILE"; git reset $GIT_COMMIT -- "myglob" \|\| true'	14:21
	← nanoo left	14:21
	→ rapha joined	14:21
rapha	Hi!	14:21
philsturgeon	charon: quick check, do i just put that command into .git/hooks/post-receive ?	14:21
rapha	can git be configured to show megabytes instead of mibibits while pulling/pushing?	14:22
	← vu3rdd left	14:22
charon	ARos: works for me... did you run the first one already?	14:22
ARos	yes	14:22
	→ roop joined	14:22
charon	reset your repo to the old one	14:22
	i.e. pre-rewrite, which was stored in refs/original/refs/heads/<branchname>	14:23
	← pinnacle left	14:23
charon	and make sure you have a backup, i should have mentioned that. gfb can be a bit destructive.	14:23
	philsturgeon: it must be executable, so make sure it has +x permission and a #!/bin/sh line	14:23
	→ petrux joined	14:24
offby1	EvanR: glad to hear things are sane	14:24
ARos	charon: I reset to the old one and I still see no files. :) i didn't lose any data and can reclone anytime...	14:24
SlickT10	any one know how to change user information associated with a commit? I accidently commited, I dont don't know how many commits back, a revision but forgot to change the email address for my clone using git config user.email.	14:24
	→ threeve joined	14:26
charon	ARos: hmm. well, i've tested the one that has $GIT_COMMIT and that definitely works. the first one would have destroyed your history, hence the request to reset. is the upstream repo public?	14:26
	← catalis left	14:26
	← maattd left	14:26
	→ altrux_ joined	14:27
ARos	charon: sadly no... I'll continue to investigate, but I really appreciate both your and hipe's generosity. thanks!	14:27
charon	ARos: i'm taking a coffee break, but if you can't get it to work please post 'git log --graph --oneline --name-status' for both pre- and post-rewrite so i can get a picture of what's going on	14:28
	(or some other variant on the theme, if you find one that better shows the problem)	14:28
ARos	charon: thanks... will do!	14:28
	→ thm joined	14:29
steffkes	http://img522.imageshack.us/img522/5859/gitbranches.png these are the actual results .. any suggestions for other tools to try?	14:29
	← tbf left	14:29
thm	are there ways of making git support meta data like file permissions, owners, or special files like device files?	14:29
Rhonda	etckeeper can do that, IIRC	14:30
thm	sounds interesting	14:30
	← altrux left	14:30
	altrux_ → altrux	14:30
	← killu left	14:31
Rhonda	I don't know about special files, but from what I understood it definitely can work with owners and permissions.	14:31
steffkes	the "normal" git should already save permissions (alias chmod), no?	14:31
thm	it only keeps the x bit afaik	14:32
philsturgeon	charon: any guides on how to set this up? i have a post-recieve hook with the hard reset command and +x but its not doing a thing :(	14:32
	→ kukks joined	14:32
Rhonda	steffkes: Only the execute bit information of the user, nothing more.	14:32
	← jelmer left	14:32
	→ JonathanT joined	14:32
Rhonda	So it's a single bit vs. actually 9 that are stored in files.	14:32
	And that only for … well, files. :)	14:32
	→ jelmer joined	14:33
steffkes	old mode 100644, new mode 100755 .. this was the result of an "chmod 777" ?	14:33
	→ benalman joined	14:33
	← maxter left	14:34
	← jokaro left	14:36
Rhonda	steffkes: Like said, only mode that git keeps is execute permission on a file	14:37
	The rest is applied to umask.	14:37
	← ben_alman left	14:38
Rhonda	See also git config about core.sharedRepository	14:38
steffkes	Rhonda, ah .. okay :) still learning ;o	14:39
	perhaps you have an idea .. for windows gui-tools, rendering git branches like in the sample above?	14:39
offby1	steffkes: the best Git GUI I know of is plain old gitk, which works OK in windows	14:40
steffkes	offby1, ah lol .. clear .. the good old git gui .. mom, trying =)	14:40
	→ carlosleon joined	14:41
carlosleon	Hi there! o/	14:42
offby1	.oO("mom"?)	14:42
	← NET\|\|abuse left	14:42
steffkes	offby1, mom like moment as in "one moment" ;o	14:42
	save time *g	14:42
offby1	steffkes: I figured.	14:42
	steffkes: You do realize, though, that "mom" is short for "Mother"	14:42
	mom == mama, more or less	14:43
	← munichlinux left	14:44
carlosleon	does anybody has an example of how to push using git over http? something like git http-push origin master ?	14:44
	→ becomingGuru joined	14:45
wereHamster	use a http:/// type url	14:46
	but push over http requires a fairly recent git version IIRC	14:47
Ilari	1.6.6...	14:47
carlosleon	I guess I have the last version	14:47
	but, where should I put the http:// url ? :-/	14:48
Ilari	carlosleon: As URL remote points to.	14:48
	→ Jeena joined	14:48
	← zuez left	14:48
SlickT10	any one know how I can fix a mistyped email in my early commits. I have it fixed so my next commits will be correct. But what about old ones? Is there a way to say that anyone that has the name 'john doe' also has 'email@example.com'?	14:49
Ilari	carlosleon: (And in latest version, 'git remote set-url <remote> <url>' should change where remote points to).	14:49
carlosleon	mmm ok...	14:49
Ilari	SlickT10: filter-branch (I think there's example of that). But beware it changes IDs of all commits.	14:50
Jeena	Hi, our internal product name changed from foo to bar and I would like to change the name of the repository, do I just mv foo.git bar.git or is there more I have to change?	14:50
Ilari	carlosleon: To check the current url: 'git remote -v'.	14:50
	→ cannonball joined	14:50
Ilari	Jeena: Plain git or is gitosis/gitolite involved?	14:50
	benalman → ben_alman	14:50
carlosleon	Ilari: is already set to the url I want to push	14:50
	→ spearce joined	14:51
Jeena	I don't know what gitosis is, so I suppose plain git	14:51
Ilari	carlosleon: Then just try pushing (there's a lot to go wrong, but...)?	14:51
	Jeena: Then just mving it should be enough (and changing the URL references on client end).	14:52
carlosleon	it returns me a 22 error code :-/	14:52
Ilari	carlosleon: Can you pastebin complete output?	14:52
	→ killu joined	14:52
Jeena	ok thanks	14:52
	← simplechat left	14:53
	← snucko left	14:53
carlosleon	Ilari: http://pastie.org/842276	14:53
steffkes	offby1, not really that better with gitk, no? http://img197.imageshack.us/img197/9458/gitbranches02.png	14:53
Ilari	carlosleon: You can't push to github over HTTP (pushes there are over SSH).	14:54
carlosleon	so should I type git http-push origin master?	14:55
	← altrux left	14:56
	→ Adlai` joined	14:56
	→ altrux joined	14:56
	→ spaceonline joined	14:56
	← Adlai left	14:56
	Adlai` → Adlai	14:56
	← killu left	14:57
	→ smuf joined	14:57
	→ killu joined	14:57
Ilari	carlosleon: No. You can not push there over HTTP!	14:57
	→ cubuanic joined	14:58
	→ britg joined	14:58
offby1	steffkes: it's the only one I use :)	14:59
steffkes	offby1, i actually try .. for the fellows ;)	15:00
	→ nanoo joined	15:00
offby1	steffkes: how did you get the diagrams in the left column?	15:00
	→ mklappst1hl joined	15:00
offby1	Those look nicer than GraphViz	15:00
steffkes	offby1, i have no idea .. there are taken from here: http://progit.org/book/ch5-3.html#largemerging_workflows	15:01
	← Ramosa left	15:03
carlosleon	Ilari: so, am I condemned? :'(	15:03
Ilari	carlosleon: You can't use SSH?	15:03
	← mklappstuhl left	15:03
offby1	ah	15:03
	→ ceez joined	15:03
carlosleon	Ilari: I can not :(	15:04
	← ankit9 left	15:04
	← ceez left	15:04
carlosleon	I'm at the office, behind ugly firewall </3	15:04
	→ ceez joined	15:05
offby1	steffkes: those graphs from progit _look_ nice, but I don't actually understand what they're trying to tell me :-\|	15:05
carlosleon	and I've been trying lot of diffrerents tricks without success	15:05
	← ceez left	15:05
	→ rahmu joined	15:05
offby1	carlosleon: if your office prevents you from SSH-ing out, then your employer doesn't want you to be productive.	15:06
	they might as well forbid you from using the telephone or taking bathroom breaks.	15:06
carlosleon	LOL	15:06
offby1	or wearing eyeglasses, or ...	15:06
	→ kim-day joined	15:06
RandalSchwartz	and if your office prevents you from ssh'ing out, but still permits https connections, they're just fooling themselves. :)	15:06
carlosleon	offby1: we don't use git in the office. It's something more personal	15:06
steffkes	offby1, you do not? they are ne best formatted ones i've found in the need :/	15:06
RandalSchwartz	socat to the rescue. :)	15:06
offby1	i guess!	15:06
	proxy!	15:07
steffkes	offby1, okay .. they seemed sometimes a little bit "constructed" .. but it's okay for the tutorial :D	15:07
carlosleon	I can access to pages that, imho, should be forbidden	15:07
	RandalSchwartz smiles at the contortions he's done over the years to get through aggressive firewalls	15:07
offby1	carlosleon: of course you can.	15:07
carlosleon	does my office sucks? :-/	15:07
Ilari	carlosleon: And if its something "personal", use bundles and upload them from somewhere with SSH access?	15:07
offby1	carlosleon: another option is to push to a USB stick, take the stick home, and then get its data onto a server from home.	15:08
RandalSchwartz	yeah, sneakernet those bundles.	15:08
offby1	Ilari: jinx, more or less	15:08
	RandalSchwartz: you too	15:08
	jinxers!	15:08
carlosleon	offby1: that's what, in fact, I'm doing	15:08
offby1	carlosleon: you too!	15:08
carlosleon	hahahahaha, I'm just trying to make a little push for a community project that is needing it urgently	15:09
	← giallu left	15:10
	→ gk joined	15:10
	← psankar left	15:10
Ilari	carlosleon: And no others with push access online?	15:11
	→ tildeequals joined	15:11
carlosleon	actually, I was trying to clone the repo but it didn't allow me. Then I used an HTTP Tunnel program (Windows XP) then I could do it. But I just worked for forking it; didn't work for pushing :-/ very odd	15:11
	Ilari: can they push and appear me as the commiter?	15:11
offby1	carlosleon: you might be able to send patches via email, too, the way our forefathers did.	15:11
	Uphill, both ways, in the snow	15:11
Ilari	carlosleon: Bundles will even preserve commit IDs (which impiles preserving author/committer).	15:12
	offby1 pictures himself preserved in brine	15:12
offby1	eww	15:12
	← tschundeee left	15:14
	← mst__ left	15:14
	← kpreid_ left	15:15
	→ kpreid_ joined	15:15
carlosleon	at least, does anybody knows a free ssh account with a little cuota ? In that way I can do the push :-/	15:16
	Ilari: how is that thing about the bundles? as we say in spanish "¿Con qué se come eso?"	15:16
	→ t0rc joined	15:17
	→ bentob0x joined	15:17
	← hipe left	15:17
	→ ceez joined	15:18
offby1	carlosleon: a "bundle" is a bunch of commit, tree, and blob objects in a single file. It's a way to transport actual commits without a network	15:18
	← ceez left	15:18
	→ Haarg joined	15:18
offby1	you get the bundle to some other repo, and then you can fetch from it	15:18
Ilari	Bundles can also contain tags (including tag objects)...	15:19
	→ ceez joined	15:19
offby1	carlosleon: if you don't mind spending $6/month, you can get your own Xen slice at http://prgmr.com/xen/	15:19
	carlosleon: actually, github is free, and gives you ssh access.	15:20
	repo.or.cz might be too	15:20
	→ mvanholstyn joined	15:21
	→ simNIX joined	15:21
	← unijunction left	15:21
	← becomingGuru left	15:22
	→ Pupeno_ joined	15:22
	→ giallu joined	15:22
EvanR-work	repo.or.cz is free, i like it because its a very simple site	15:23
	no passwords ;)	15:23
charon	philsturgeon: did you read http://git.wiki.kernel.org/index.php/GitFaq#Why_won.27t_I_see_changes_in_the_remote_repo_after_.22git_push.22.3F ?	15:23
EvanR-work	and has this thing 'mob branch'	15:23
philsturgeon	charon: i got it working :)	15:23
charon	philsturgeon: there's even a premade post-update hook.	15:23
	ok.	15:23
	← Araneidae left	15:24
philsturgeon	thanks though charon	15:24
charon	ARos: any luck?	15:24
EvanR-work	offby1: i fixed that problem last night...	15:24
carlosleon	the problem in here, is that I can not go out through ssh :-/ #epic #fail	15:24
ARos	charon: yes	15:24
offby1	EvanR-work: yeah, I saw	15:25
ARos	charon: I used the following command: 'git-filter-branch --tree-filter 'rm -f $(ls -1 \| grep -v <myfileregex> \| xargs)' HEAD'	15:25
	→ mst__ joined	15:26
	← JackWinter left	15:26
ARos	charon: that command did exactly what i wanted (or at least i think so from what i've seen thus far).	15:26
charon	ARos: ok, that works too, even though it's much slower than an index filter. i still wonder why the reset recipe didn't work, but meh.	15:26
	← Pupeno left	15:26
	Pupeno_ → Pupeno	15:26
ARos	hehe	15:26
	yer	15:26
	→ geekcloud joined	15:28
	← radarek left	15:28
	← loops left	15:28
	← Ryback_ left	15:28
	→ joeytwiddle joined	15:30
	← kpreid_ left	15:30
	← fr0sty_ left	15:32
	→ rekoerbm joined	15:35
	→ JackWinter joined	15:35
	← rekoerbm left	15:35
	← steveire left	15:37
	→ unijunction joined	15:37
	Guest26987 → fahadsadah	15:38
	← fahadsadah left	15:39
	→ fahadsadah joined	15:39
eli	sitaram: ping	15:41
	→ steveire joined	15:41
	← steveire left	15:41
	→ steveire joined	15:41
	offby1 wonders if eli is investigating gitolite	15:42
offby1	although sitaram is expert in many matters, gitolite appears to be his specialty	15:42
eli	Yes, I am.	15:42
	→ SpookyET joined	15:42
offby1	it's like I can READ your MIND	15:42
sitaram	eli: hi	15:42
eli	sitaram: What's the quickest way to send you a patch?	15:43
	← unijunction left	15:43
sitaram	not fedex	15:43
	← fahadsadah left	15:43
sitaram	:D	15:43
	[email@hidden.address]	15:43
	→ reprore_ joined	15:43
eli	(I am about to start a test, which I still didn't finish writing, so no time to dig up...)	15:43
	→ Tank` joined	15:43
	← geekcloud left	15:43
eli	And how do I get an email with the patch from git? (I remember some format-patch thing, will it just work?)	15:44
	← Garen_ left	15:44
sitaram	eli: format-patch is good, but a plain git diff output is also good enough for manually applying, which is what I will be doing anyway	15:45
	eli: format-patch preserves your authorship and commit message also of course	15:45
eli	Ah, OK.	15:45
	→ Garen joined	15:46
	offby1 has an unpleasant Deja Vu	15:46
	← t0rc left	15:46
eli	I'll probably have time in a few hours -- american students don't ask too many questions during exams for some reason.	15:46
	→ _Mica_ joined	15:46
	→ bdimcheff joined	15:46
sitaram	offby1: ??	15:47
	→ fahadsadah joined	15:47
	→ k23z__ joined	15:48
offby1	sitaram: oh, a few months ago eli was contributing to my IRC bot, which lives on github; he'd never used git before, and it was excruciatingly difficult describing to him what he needed to do. I've had this problem before with other undoubtedly-smart people	15:48
k23z__	what does git mean when it asks me	15:48
	Stage this hunk [y/n/a/d/s/?]?	15:48
	wtf is a hunk ?	15:48
offby1	e.g., a co-worker who has a very open mind about git, nevertheless said (quite truthfully) that ever time I told him what command to run, it wouldn't work :)	15:48
teuf	k23z__: a part of a diff/patch	15:48
offby1	k23z__: a hunk is a bit of a patch	15:48
	teuf: jinx	15:48
teuf	:)	15:48
	→ HG` joined	15:48
offby1	k23z__: the bit it's just displayed to you.	15:48
k23z__	yeah	15:48
offby1	k23z__: so you want "y" for "yes", "n" for no, etc.	15:48
k23z__	and what exactly is it asking me ?	15:48
	what does it want from me ?	15:49
sitaram	offby1: :)	15:49
offby1	k23z__: do you want to "apply" that hunk to the file	15:49
k23z__	I hit y	15:49
	yeah sure I want to fucking apply it	15:49
offby1	k23z__: I'm guessing this all started because you typed "git add -p"?	15:49
k23z__	why else would I do git add --interactive	15:49
offby1	k23z__: I suggest "git add -p" instead of "git add --interactive"; the latter is too complex for my tiny brain, but I can understand the former	15:49
	→ tschundeee joined	15:50
k23z__	I sure hope I don't need to return to some previous version because I don't know if I'll be able to visualize previous versions before reverting with them	15:50
	← tschundeee left	15:50
k23z__	I'll rely on github web interface then ...	15:50
	as I don't know enough git to know how to peek at previous revs	15:50
	→ kumbayo joined	15:50
	← philsturgeon left	15:50
offby1	k23z__: I have two pieces of advice for you: 1) run "gitk --all"; it makes your repo pretty clear; 2) read "Git For Computer Scientists" (http://eagain.net/articles/git-for-computer-scientists/)	15:51
EvanR-work	k23z__: if you know the rev number, youre all set ;)	15:52
	→ malumalu joined	15:53
	← ph^_ left	15:53
	← tildeequals left	15:54
EvanR-work	git show <rev>, shows the differences between it and the parent, git checkout <rev> puts you there, you can get back with git checkout master (or whatever branch)	15:54
	← NetHawk left	15:55
	← ThiefMaster left	15:56
	→ bcardarella joined	15:57
	→ fr0sty_ joined	15:57
	→ pingveno joined	15:57
	→ unijunction joined	15:57
EvanR-work	and an alternative to gitk --all is git log --graph --oneline ;)	15:58
	→ jmcantrell joined	15:58
EvanR-work	but gitwebs graphic log seems to hvae the sanest graph	15:58
	→ kpreid_ joined	15:59
	→ tedoc2000 joined	15:59
jmcantrell	I've been working on a script for a while now that makes working with multiple scm's easier (for me anyway). I could use some critique, if anyone would be so kind. http://github.com/jmcantrell/sourcery	16:00
	→ NetHawk joined	16:00
WebDragon	so a misunderstanding of the direction has led me to abandon the develop branch and the feature branch off it, and start anew from master with a new devel branch, however SOME of the changes from develop+feature I want to cherry-pick over into devel before deleting the old branches, now that I better grok where to go.	16:01
	I know there's a cherrypicking function of sorts in git; how do I selectively pick changes from develop and feature branches into the current devel twig where I'm doing it properly	16:01
	→ urbanmonk joined	16:02
	→ nullvariable joined	16:04
	→ loincloth joined	16:05
	← pingveno left	16:06
Thumper_	WebDragon: you can cherry-pick from gitk - just right click on the commit	16:06
	← k23z__ left	16:07
WebDragon	O_O	16:07
	oo I like that	16:07
	← ixti left	16:09
	→ sumangurung joined	16:10
	→ methanol joined	16:11
	→ maattd joined	16:11
	← mklappst1hl left	16:11
	← nanoo left	16:11
	→ mklappstuhl joined	16:13
	← psoo left	16:13
	→ kumbayo_ joined	16:13
	← fool__ left	16:13
	← giallu left	16:15
	→ fool_ joined	16:15
	→ codeswing joined	16:15
	← fool_ left	16:15
	→ fool_ joined	16:15
	← saccade_ left	16:15
	← kumbayo left	16:16
	← tchan left	16:16
	→ tchan joined	16:18
	→ ThiefMaster joined	16:18
	→ LRN joined	16:20
	← kumbayo_ left	16:20
	→ ankit9 joined	16:21
	← f00li5h[HireMe] left	16:25
	← rahmu left	16:27
	→ catalis joined	16:28
	→ webchick joined	16:28
	← maattd left	16:28
cbreak	git can't clone mercurial projects, correct?	16:28
	→ Thell joined	16:29
Thell	Hey all, quick question, the only url I have for an svn repo is https, how do I tell git svn to use anonymous/anonymous as user/password?	16:30
	← skoop left	16:31
	← afloret left	16:31
	← sumangurung left	16:31
wereHamster	http://anonymous:anonymous@host.tld/path ?	16:31
	just use the same url as you would with svn	16:31
steffkes	for what .. if you have no pwd-protection	16:31
Thell	gotcha :)	16:31
	ahhhh, using --username for git-svn is good too :)	16:34
	← whitman left	16:34
	→ uchimata_ joined	16:34
	→ lightcap joined	16:35
	→ stkrzysiak joined	16:36
	→ pheaver joined	16:37
	← lightcap left	16:37
	← uchimata left	16:38
	→ nanoo joined	16:38
	← Pinaraf left	16:38
	→ tvw joined	16:38
	→ bibinou joined	16:39
	→ amerine joined	16:39
	→ lightcap joined	16:40
	← roop left	16:41
	← EricInBNE left	16:42
	→ til joined	16:42
	→ [reaVer] joined	16:44
	← reaVer left	16:45
	→ tmske joined	16:45
	← mvanholstyn left	16:46
tmske	hi, I'm doing a git svn rebase but I get a merge conflict, how can I tell git to take the file from svn and overwrite mine?	16:46
	← ben_alman left	16:46
	→ _Vi joined	16:48
	← methanol left	16:48
wereHamster	tmske: git checkout --theirs -- <file>	16:49
WebDragon	hmm I can cherry-pick the entire commit from another branch, but can I _selectively_ cherry-pick only certain LINES from a particular FILE in a particular branches' commit ?	16:50
wereHamster	not directly, just do the cherry-pick, reset HEAD~1 and then use add -p	16:50
	← unijunction left	16:50
	← lightcap left	16:50
	→ mvanholstyn joined	16:51
wereHamster	or diff \| patch and then add -p	16:51
WebDragon	also cherry-picking from gitk results in similar file permissions overrides that I'm attempting to preserve with the setgitperms.perl from contrib/hooks	16:52
	which means naturally that the incl_header.php file doesnt' get included until I go back and re-chmod 644 the file after cherrypicking	16:52
	I really wish I knew C well enough to write patches -- I'd much prefer git to have a way to enable/disable full permissions-tracking so it can be properly used for website development	16:53
	← acoul left	16:53
wereHamster	what besides the executable bit do you need to store?	16:53
jast	it would be a rather large operation to add that	16:53
WebDragon	wereHamster: read perms for group/world for one thing	16:53
wereHamster	checkout as the user/group who will use it	16:54
	← HG` left	16:54
WebDragon	wereHamster: I run localhost-only devel vhosts for all my clients so I can preflight scripts before uploading to the devel staging server and thence to the live server	16:54
	and as soon as git resets the world-read bit, I can't preview the dam files anymore :)	16:54
	I cherrypicked the change to the header include file into the current branch, only to have the commit get applied (I thought it would stage it in the index, not actually commit the thing) and the header promptly disappear from the devel server website	16:55
wereHamster	cherry-pick -n or reset HEAD~1 after the cherry-pick (like I suggested)	16:56
WebDragon	but it _committed_ the change	16:56
	well I did it from the gitk window	16:56
	right clicking the commit	16:56
wereHamster	that's what you get for using a gui ;)	16:56
WebDragon	Thumper_: had suggested that earlier :)	16:56
	LOL	16:56
rapha	is there any way to get git to measure pull/push progress in MB instead of MiB?	16:56
WebDragon	ok you got me there	16:56
	Thumper_ ducks	16:56
	← aatanasov left	16:57
tmske	wereHamster: thanks!	16:57
wereHamster	rapha: patch the source...	16:57
WebDragon	do I need to reset --soft HEAD~1 or just reset HEAD~1 ?	16:57
wereHamster	HEAD~1 (you need to unstage the changes)	16:57
rapha	wereHamster: thanks, I just could not get my mind to stretch to that	16:58
	WebDragon is annoyed that WE have to use MiB and the manufacturers still get to use MB .. since it was their fuckup they should be the ones to have to use the whack-ass abbrevs	16:58
WebDragon	meanwhile MB can be used to stand for what it always did. 1024K	16:59
rapha	i don't want the abbreviation to change, i want the measurement to change	16:59
WebDragon	it's not OUR fault people are brainless screwheads	16:59
wereHamster	rapha: wait, you want git to display MiB and measure in 1000's ?	17:00
WebDragon	rapha: ahh calculation	17:00
	← nanoo left	17:00
	WebDragon thinks MiB should be the 1000's measure and MB should be the 1024k-measure	17:00
rapha	WebDragon: no, git displays MiB and measures MiBs. I want it to display MBs and measure MBs.	17:00
	← d0ugal left	17:01
wereHamster	17:59:35 <rapha> i don't want the abbreviation to change, [...]	17:01
	wereHamster is confused	17:01
WebDragon	rapha: that's what I assumed you had meant originally	17:01
rapha	yeah, let's call that a `misunderstanding'	17:01
WebDragon	you want to switch standards from MiB to MB	17:01
	and display the resultant calculations accordingly	17:01
rapha	MiB/MiB -> MB/MB is what i want	17:01
	yeah	17:02
	→ andrewa2 joined	17:02
wereHamster	send a patch to the mailing list (the address is in the topic)	17:02
rapha	well, to be more exact, i wanted to know if there's an option for it	17:02
	since there isn't i'll just a ccept my fate	17:02
wereHamster	no	17:02
	← ankit9 left	17:02
rapha	thanks :)	17:03
	→ unijunction joined	17:04
	→ Tureba joined	17:06
	→ ben_alman joined	17:07
	→ rhebus joined	17:07
	← [reaVer] left	17:09
	← rgr left	17:10
	→ davidw joined	17:11
	← bibinou left	17:11
	→ consolers joined	17:13
	→ Jeena_ joined	17:13
	→ jiunec joined	17:14
	→ reaVer joined	17:14
	← kpreid_ left	17:15
	→ kipras joined	17:15
	← Jeena left	17:16
	Jeena_ → Jeena	17:16
	→ ProperNoun joined	17:16
	→ ankit9 joined	17:17
	→ seanmccann joined	17:17
	← reaVer left	17:18
rhebus	I'm working with a local clone of a repository under winxp. I want to clear local changes so I can merge in a remote branch, so I did a "git reset --hard HEAD", but afterwards there are still local changes! The changes relate to LF vs CRLF. The repository was created on a linux system but I think it contains crlf endings in the offending files. What can I do?	17:19
offby1	rhebus: oh that.	17:19
	← Adjective left	17:19
rhebus	i've tried "git config core.autocrlf input"	17:20
offby1	rhebus: see http://offby1-whining.blogspot.com/search/label/git	17:21
	→ reaVer joined	17:21
	← reaVer left	17:21
	→ reaVer joined	17:21
	→ disappearedng_ joined	17:21
	← kipras left	17:21
	→ kipras_ joined	17:21
	← tchan left	17:22
rhebus	offby1: thanks very much, that fixed it :)	17:23
	→ defectiv joined	17:24
	→ saccade_ joined	17:25
	→ tchan joined	17:25
	→ dj_tjerk joined	17:26
	→ nanoo joined	17:27
	← marc_kdab left	17:27
jmcantrell	I've been working on a script for a while now that makes working with multiple scm's easier (for me anyway). I could use some critique, if anyone would be so kind. http://github.com/jmcantrell/sourcery	17:27
	→ plod joined	17:27
	← britg left	17:27
	→ anjen joined	17:29
	→ coolcat joined	17:29
	→ alley_cat joined	17:29
	→ methanol joined	17:31
	← tobiassjosten left	17:31
	← tetha left	17:32
	→ tetha joined	17:33
	→ philsturgeon joined	17:34
	← petrux left	17:35
	← ryan[WIN] left	17:35
	← philsturgeon left	17:36
	← Adlai left	17:37
wereHamster	jmcantrell: send that to the mailing list	17:37
	→ jrmuizel joined	17:37
	→ toxx_ joined	17:38
	← methanol left	17:38
	FFForever-Away → FFForever	17:39
	← Determinist left	17:39
	→ priidu_ joined	17:41
	→ Adjective joined	17:41
	← disappearedng_ left	17:41
	← mst__ left	17:41
	→ priidu__ joined	17:41
	← chewbranca_ left	17:43
	← thm left	17:43
jmcantrell	wereHamster: ok	17:44
	← ProperNoun left	17:45
	← SpookyET left	17:45
	→ LiamH joined	17:46
	→ RobertLaptop_ joined	17:46
	← nizuri left	17:47
	→ mst__ joined	17:47
	← RobertLaptop left	17:48
	→ ajpi joined	17:49
	← killu left	17:50
	→ descention joined	17:50
	← ajpiano left	17:50
descention	I'm having trouble with submodules. I have the .gitmodules file with the path and url, but I think my repo still considers the directory part of the main project. I'm using mysysgit/tortoiseGit on windows.	17:52
	← jelmer left	17:52
	← bentob0x left	17:53
	← andrewa2 left	17:53
descention	the submodule is not listed in .git/config, but I'm not sure if it needs to be.	17:53
	→ psoo joined	17:54
	← jmut left	17:54
	→ segher joined	17:54
	→ Ryback_ joined	17:56
	→ SpookyET joined	17:57
	← bambule left	17:57
	→ yann joined	17:57
	← jrmuizel left	17:58
	→ christophsturm joined	17:58
	→ lhz joined	17:58
	→ tehbaut joined	17:58
	← tchan left	17:58
	→ mrlarner joined	17:59
	→ tchan joined	17:59
	→ jrmuizel joined	18:01
	→ David-T joined	18:02
	← agugsu left	18:02
	← mrlarner left	18:03
	→ maxter joined	18:03
	← hebz0rl left	18:05
FFForever	if I move a file do I have to delete/readd it in git?	18:05
davr	generally yes	18:06
wereHamster	FFForever: man git-mv	18:06
Gitbot	FFForever: the git-mv manpage can be found at http://git.or.cz/man/git-mv	18:06
	→ krawek joined	18:06
FFForever	thx	18:06
	→ nessundorma joined	18:07
	← shruggar left	18:07
FFForever	one last thing how can I update my source?, do I need to reclone it?	18:08
wereHamster	FFForever: man git-pull	18:08
Gitbot	FFForever: the git-pull manpage can be found at http://git.or.cz/man/git-pull	18:08
	← davidw left	18:09
ARos	Would someone kindly translate 'git-filter-branch --tree-filter 'rm -f "somefileglob*"' HEAD' into a 'git-filter-branch --index-filter' statement? Thus far no luck...	18:10
	← krawek left	18:11
	→ carrus85 joined	18:11
	← agliodbs left	18:11
	→ webus joined	18:12
	→ jaql joined	18:13
	jiunec → cenuij	18:14
	← cenuij left	18:14
	→ cenuij joined	18:14
	← godsflaw left	18:14
	→ ixti joined	18:16
	← smuf left	18:16
	→ killu joined	18:17
	← loincloth left	18:18
peper	what was the config to force remote branch deletion when you get deletion of the current branch prohibited	18:18
	← codeswing left	18:19
	← consolers left	18:19
	← Adjective left	18:19
	→ maattd joined	18:20
	→ sh1mmer joined	18:21
	→ magn3ts joined	18:22
	→ marc_kdab joined	18:22
	← nanoo left	18:22
	→ sh1m joined	18:22
	← magn3ts left	18:23
	→ magn3ts_ joined	18:23
	← sh1mmer left	18:23
	sh1m → sh1mmer	18:23
	→ bcardarella_ joined	18:23
	← One-Man-Bucket left	18:25
	→ ph^_ joined	18:25
	→ krawek joined	18:25
	← gitte left	18:27
	← sh1mmer left	18:28
	→ agliodbs joined	18:29
	→ sumangurung joined	18:29
	← magn3ts_ left	18:30
	← krawek left	18:30
davr	Is it possible to just get the contents of a subdirectory with the new "sparse checkout" in git 1.7.0? (like how you can checkout just a subdirectory in SVN, and you only get the contents of that dir, not the whole folder structure)	18:31
	← saccade_ left	18:33
	→ numberGrey joined	18:33
	← aresnick left	18:33
	← agliodbs left	18:34
	→ pantsman joined	18:35
wereHamster	peper: prohibited by who?	18:35
Thell	anyone know of any blog posts showing how people are usinggit-notes (and what they are using it for)	18:36
	?	18:36
	← tehbaut left	18:37
	← WebDragon left	18:37
Thell	kinda like a use-case type of thing?	18:37
	→ corni joined	18:38
	← galderz left	18:38
	→ f00li5h[HireMe] joined	18:38
	→ sh1mmer joined	18:38
	← inetic left	18:39
	← sh1mmer left	18:39
	← vinse_ left	18:39
	→ Mark__T joined	18:40
	→ krawek joined	18:40
	ajpi → ajpiano	18:41
	→ sumangurung_ joined	18:41
	→ mklappst1hl joined	18:41
	→ dreiss joined	18:42
	→ aresnick joined	18:42
	← bcardarella_ left	18:43
	→ sh1mmer joined	18:43
	← sumangurung left	18:44
	sumangurung_ → sumangurung	18:44
Thell	ya know... that isn't the easiest google search... ("git-notes"\|"git notes") returns lots of random notes :P	18:44
ciaranm	you need to use wolfram alpha instead. it knows the different meanings of 'notes'!	18:44
wereHamster	Thell: does the notes patchset not contain a descripitno?	18:44
Thell	there's a man page... didn't see a use case type of doc	18:45
	Thell looks again	18:45
	← mklappstuhl left	18:45
	→ tehbaut joined	18:45
	→ ilteris joined	18:45
	→ Theravadan joined	18:46
	← mastro left	18:46
	→ mrlarner joined	18:47
davr	it's a pretty simple tool IMO. that's like asking "are there any posts showing a use-case for notepads?"	18:47
	it's for whatever you want :P	18:47
	→ thiago_home joined	18:48
	→ bblfish joined	18:48
Thell	lol... I'm interested in see how people are using. ie: tying into ticket trackers? As web page data? To populate wiki? As transient note to author type of things... etc.	18:49
bblfish	hi, just wondering what the policy people have for placing files such as .idea/workspace.xml into git - ( an intelliJ directory)	18:50
	→ godsflaw joined	18:50
	→ chrisadams joined	18:51
Thell	bblfish: from what I've seen, things that are per user should not be in the project. if workspace.xml is like visual studios .sln files though then that just depends on the project.	18:51
	That just my limited experience though	18:52
	→ leitao joined	18:52
bblfish	mhh, probably a question I should ask on #intellij	18:52
leitao	how do I unmark the file to a commit that I just selected with "git-add" ?	18:52
Morasique	git reset HEAD file	18:53
Thell	leitao: git checkout -- path/to/file or git reset -- path/to/file two different cases though	18:53
	the reset will keep your contents, the checkout will actually check out the HEAD version of that file	18:54
	← catalis left	18:54
leitao	Thell: i need to keep my content. Just remove one file from the commit	18:55
	thanks!	18:55
	← mrchrisadams left	18:55
	→ tbf joined	18:56
leitao	Thell: it says, iscsi.c: locally modified. Also, It seems to be marked to be commited	18:56
Thell	hmmm, guess I forgot the HEAD like Morasique stated:: as per 'git status':: (use "git reset HEAD <file>..." to unstage)	18:58
	→ Sara-25-sexy joined	18:58
	→ Chillance joined	18:59
	→ loincloth joined	18:59
	← sde_antrix left	18:59
	← jrmuizel left	19:00
	→ vinse joined	19:00
	→ nessundorma_ joined	19:02
	→ brizly joined	19:03
	→ viniciusluiz joined	19:04
	→ magn3ts joined	19:05
	→ justin-george joined	19:05
	← nessundorma left	19:06
	nessundorma_ → nessundorma	19:06
viniciusluiz	pessoal uma pergunta... tenho a linha	19:06
	git remote add origin ssh://repo@10.155.22.154/~/adap_sig.git	19:06
	so que preciso que ela aponte para a porta 10022	19:06
	como faco isso ????	19:06
	→ jrmuizel joined	19:06
	← ARos left	19:07
wereHamster	viniciusluiz: this is an english-only speaking channel	19:08
	→ ijuma__ joined	19:08
	← Sara-25-sexy left	19:08
viniciusluiz	hi people... hello people .. sorry ..! =/	19:08
	← SlickT10 left	19:10
	→ bousquet joined	19:10
	← ijuma left	19:10
	← loincloth left	19:11
tango_	viniciusluiz: ssh://repo@ip:port ?	19:12
	← carlosleon left	19:13
viniciusluiz	tentei assim mas nao deu em nada	19:13
	aparece o seguinte	19:13
	$ git push origin master	19:13
	ssh: connect to host 10.155.22.154 port 10022: No route to host	19:13
	fatal: The remote end hung up unexpectedly	19:13
	→ loincloth joined	19:15
wereHamster	I understand the git output, but not the other messages you have written. They seem to be in a english dialect that I don't understand	19:16
	← maxter left	19:16
unijunction	whats the git console command for reset amster branch to here?	19:17
wereHamster	anyway, 'No route to host' should be self-explaining	19:17
	unijunction: checkout master; reset [options] HEAD@{1}	19:18
	→ d0ugal joined	19:19
unijunction	my master is pointing at a commit behind what i consider to be my most recent	19:19
	if i git reset commitid and reload gitk it looks like it hasn't moved	19:19
	→ RdeWilde joined	19:19
unijunction	and somebody else fetching will miss it	19:19
wereHamster	what did 'git reset commitid' print to stdout?	19:21
unijunction	nothing	19:21
	was clean	19:21
wereHamster	are you sure you are on 'master	19:21
	?	19:21
unijunction	but, messing around some got the desired effect	19:22
	i think i may not hav been	19:22
	there was some messy stuff about emrging	19:22
	and stupid people with windows newlines	19:22
	grr :-/	19:22
	makes merging hard/stupid	19:22
	→ fool__ joined	19:22
unijunction	but yeah, cheers	19:22
	← fool__ left	19:22
	→ fool__ joined	19:22
wereHamster	the windows people should use autocrlf	19:23
	← priidu__ left	19:24
Thell	and repo maintainers should use git log -p to ensure no mixed line endings end up in the repo...	19:25
	→ mtheoryx joined	19:25
	→ binjured joined	19:25
	← priidu_ left	19:27
	← priidu left	19:27
	→ riz2 joined	19:27
	← fool_ left	19:27
cxreg	http://profile.ak.fbcdn.net/object/79/42/n2204603590_28769.jpg	19:29
	→ bambule joined	19:31
	← sumangurung left	19:31
	→ bugfux joined	19:32
	← mtkd left	19:32
	← zmyrgel left	19:34
RdeWilde	Smart kid	19:35
	→ fynn joined	19:35
fynn	Hey. I "git add"ed a new file foo.py. Then I did "git reset --hard". Now foo.py is completely gone.	19:36
	Any way to recover it?	19:36
Ilari	fynn: No easy way. Its in object database somewhere...	19:36
bblfish	I found this for intellij http://jetbrains.net/devnet/docs/DOC-1186	19:37
fynn	Ilari: hm, any leads? foo.py contains a ton of new code.	19:37
bblfish	the problem is if I remove those files then I can't checkout another branch	19:37
Ilari	fynn: 'git fsck --full'?	19:37
fynn	how do I even look for it?	19:37
	← chrisadams left	19:38
	→ kpreid_ joined	19:39
	← webus left	19:40
	→ khmarbaise joined	19:46
	→ erikvold joined	19:53
	← imajes left	19:53
	→ imajes joined	19:53
	← imajes left	19:53
	→ imajes joined	19:54
	← tatsuya left	19:56
	← jrmuizel left	19:57
	→ sh1m joined	19:58
	← JonathanT left	19:58
	→ codeswing joined	19:59
	→ tildeequals joined	19:59
	← magn3ts left	19:59
	→ ja2 joined	20:00
	← ntoll left	20:00
	← sh1mmer left	20:01
	sh1m → sh1mmer	20:01
	→ Twisted_Mentat joined	20:02
	→ ntoll joined	20:03
	→ catalis joined	20:05
	→ r11t joined	20:06
	← _Vi left	20:06
	← fr0sty_ left	20:06
	→ _Vi joined	20:06
	→ fr0sty_ joined	20:07
	→ Zephyrus joined	20:08
	→ sumangurung joined	20:09
	→ pingveno joined	20:10
	← mtheoryx left	20:10
	→ matthieu joined	20:11
	← tildeequals left	20:11
	← maattd left	20:12
	← bluenovember left	20:12
	→ bdimcheff_ joined	20:13
	→ Adlai joined	20:15
	← hyperair left	20:16
	→ sh1m joined	20:17
	← sh1mmer left	20:17
	sh1m → sh1mmer	20:17
	← bdimcheff left	20:17
	bdimcheff_ → bdimcheff	20:17
	→ magn3ts joined	20:18
	← christophsturm left	20:21
	← sh1mmer left	20:22
	← pingveno left	20:22
	→ jackdempsey joined	20:22
	→ Mike_MnD joined	20:23
	→ Mage joined	20:23
	→ nanoo joined	20:23
Mike_MnD	how can i clone from ssh port other then the default 22	20:24
badboy_	host.example.com:[port] ?	20:25
	→ Alagar joined	20:26
	← leitao left	20:26
	→ hyperair joined	20:26
Thumper_	Mike_MnD: add port info in .ssh/config for that host?	20:26
	→ JonathanT joined	20:27
	→ sockmonk joined	20:27
	→ sh1mmer joined	20:27
	→ priidu__ joined	20:29
	→ priidu joined	20:29
	→ priidu_ joined	20:30
	→ nanooo joined	20:32
	→ radarek joined	20:33
thiago_home	Mike_MnD: or use ssh://hostname:port/path	20:34
	← nanoo left	20:34
	→ johnw joined	20:36
	← codeswing left	20:37
	← DavidKlein left	20:37
	← matthieu left	20:39
	← hyperair left	20:41
	← lhz left	20:41
	→ r0ns joined	20:42
	→ drizzd joined	20:43
	FFForever → FFForever-Away	20:44
	→ brendon_work joined	20:44
	→ dabd joined	20:44
	→ Logjam joined	20:45
	→ maattd joined	20:47
	← r11t left	20:48
	← Mike_MnD left	20:48
	→ r11t joined	20:48
	← mithro left	20:52
	→ [Nies] joined	20:52
	→ justinlilly joined	20:53
	← r11t left	20:54
	← JoeConyers left	20:54
	← Nies left	20:56
	→ AkumaStreak joined	20:57
	→ patthoyts joined	20:57
	← viniciusluiz left	20:58
	← d0k left	20:58
	← d0ugal left	20:59
	← brizly left	21:01
	← fynn left	21:01
	← jspiros left	21:01
	← sumangurung left	21:01
	→ jspiros joined	21:03
	→ abasit joined	21:05
	→ blytkerch joined	21:05
	→ WebDragon joined	21:06
	← _iron left	21:06
WebDragon	how do I ask git config to show me ALL the options, even if they haven't been changed from the defaults, and aren't physically set anywhere?	21:06
	← blytkerch left	21:06
abasit	any idea why "git pull" wont work? I get this error: "git: 'pull' is not a git-command. See 'git --help'."	21:06
badboy_	o_O	21:06
	→ _iron joined	21:07
badboy_	and what does git --help say?	21:07
WebDragon	abasit: what version?	21:07
abasit	git version 1.6.4.2	21:07
	→ jelly-bean joined	21:07
WebDragon	did you install all the git packages for your distro?	21:07
jelly-bean	our git repo has grown to 100mb+	21:07
	→ mjf joined	21:07
jelly-bean	can i chop it in half so it becomes two dif repos	21:08
	one big and one little with just last few commits	21:08
abasit	im not sure... im using a hosting service...	21:08
	→ tildeequals joined	21:08
	← nanooo left	21:09
abasit	git help shows pull as a valid command	21:09
	← corni left	21:10
abasit	any ideas?	21:10
	← _Mica_ left	21:12
abasit	WebDragon?	21:12
WebDragon	abasit: shell is bash ?	21:13
	abasit: try (git help pull)	21:14
	→ jrmuizel joined	21:14
WebDragon	heck your versions' even newer than mine.	21:15
	git-1.6.2.5-1.fc11.i586	21:15
drizzd	abasit: somebody had that same problem yesterday	21:16
	← abasit left	21:16
	← ehabkost left	21:16
WebDragon	for my distro's release, the actual command as referenced by (git) is in /usr/libexec/git-core/git-pull - is it missing, do you think, on your system? or is git somehow confused as to where its executables are stored?	21:17
	maybe it's mis-installed?	21:17
	← magn3ts left	21:17
	WebDragon wanders off since he has work to do and can't wait for answers in the meantime, but will still be checking back periodically	21:18
jelly-bean	any way to truncate git repo history/commits	21:21
	← mvanholstyn left	21:22
	← malumalu left	21:22
WebDragon	how do I check one file out of an earlier branch tip? git checkout master -- path/to/file ?	21:25
	and bring it into the current branch mind you	21:25
	anyone?	21:25
	→ davidw joined	21:27
	← kipras_ left	21:27
	davidw → Guest76185	21:28
	→ EricInBNE joined	21:31
	← remi__ left	21:31
WebDragon	yeap nevermind, that was indeed it :)	21:31
	→ c_axis joined	21:32
	← c_axis left	21:32
	← tbf left	21:33
rtyler	spearce: how well tested is the JGit with the Gerrit 2.1 snapshots?	21:39
	← justinlilly left	21:39
rtyler	I'm pretty consistently running into "fatal: The remote end hung up unexpectedly" after a few days of operation	21:39
	(when pushing)	21:40
	← ceez left	21:40
rtyler	it claims to have had an internal server error, but sshd_log doesn't show any exceptions	21:40
	→ tatsuya joined	21:41
spearce	blargh	21:41
	it doesn't get enough testing. it clearly needs more.	21:41
rtyler	anything I can help you with to find out what's going on?	21:42
	is the android gerrit running the snapshot? or still on 2.0?	21:42
spearce	its running 2.1.2-rc1, built last night	21:42
	we're also running that internally	21:42
	→ endpoint_david joined	21:43
	Guest76185 → davidw	21:43
	← davidw left	21:43
	→ davidw joined	21:43
	← RdeWilde left	21:44
endpoint_david	what's a good way to resolve multiple binary conflicts? specifically, I've got a batch of conflicting images, would like to review the version from each commit before making a decision on each	21:44
	→ saccade_ joined	21:44
endpoint_david	ISTR a <digit>:path/to/file syntax for some command that let you show the version of the file for various index states	21:45
	but I can't recall it right now	21:46
	← saccade_ left	21:46
	→ justinlilly joined	21:47
spearce	rtyler: jgit changed a lot... but i can't see anything that could cause a hang up like that. what is that from? fetch? push?	21:48
rtyler	push	21:49
endpoint_david	I assume that if there was a binary conflict that the contents would have to be different, not just two commits both adding the same binary file	21:49
rtyler	`git push gerrit master`	21:49
	spearce: I'm building the latest head now	21:49
	→ posix4e joined	21:49
	← catalis left	21:50
	← jelly-bean left	21:53
	→ jasonlife joined	21:53
jasonlife	Is there a good practice document for centralized workflow?	21:53
	→ EvanR joined	21:54
jasonlife	I'm not clear how can I handle the branches using centralized workflow..	21:54
	← David-T left	21:56
	→ David-T joined	21:56
jmcantrell	jasonlife: you would just have a bare repo that everyone pushes/pulls to/from	21:57
	← psoo left	21:58
endpoint_david	ok, I think I can use the output of `ls-files -u` to retrieve the current stage/blobs	21:58
	← jrmuizel left	21:58
	← Mezner left	21:58
jasonlife	jmcantrell: Does shared repo should have only master? no branches in shared repo?	21:58
	← mrlarner left	21:59
jmcantrell	jasonlife: no, it's just like the developer repos, just no working copy	22:00
	← Zephyrus left	22:00
jmcantrell	jasonlife: you don't have to push everything to the central repo, but you can	22:01
jasonlife	jmcantrell: then it will be like cvs .. right?	22:02
	in terms of using it..	22:03
jmcantrell	jasonlife: if you mean that there will be a central repo, then yeah, but that's about the extent of the similarities	22:04
	→ DavidKlein joined	22:04
jmcantrell	jasonlife: you're familiar with the idea of a dvcs?	22:04
	← ph^_ left	22:05
jasonlife	jmcantrell: yeah, I'm currently using cvs, and thinking about migrating to other SCM..	22:05
jmcantrell	jasonlife: well, cvs is not distributed. you have to start thinking differently if you're jumping to a dvcs	22:06
	→ ph^_ joined	22:06
	→ jrmuizel joined	22:07
jasonlife	jmcantrell: If I push something stupid into central repo, and find out it some time later, is it possible to remove my changes from the central repo without affection any other pushes made by someone else?	22:07
	← segher left	22:07
	← ntoll left	22:07
	← auslander left	22:08
jasonlife	without affecting	22:08
jmcantrell	jasonlife: not if they've pulled your changes	22:08
	but you shouldn't be pushing to your central repo until you're satisfied with your dev repo	22:08
jasonlife	But if I push into central, nobody can't push until they pull my changes.. rigth?	22:09
	→ segher joined	22:09
jmcantrell	right	22:09
jasonlife	yeah.. I shouldn't have to push my stupid, but sometime it happens..	22:09
cbreak	you can push a revert comit	22:10
	one that undoes your changes	22:10
	but the history will remain	22:10
	← priidu_ left	22:10
	← priidu left	22:10
	← priidu__ left	22:10
jasonlife	oh..	22:10
jmcantrell	which is actually the "right" way	22:11
	← jrmuizel left	22:11
cbreak	the "propper" way, yes	22:11
endpoint_david	I'd like to see a "local-only" branch flag that basically prevents a push for that particular branch name	22:11
cbreak	but if you slandered your boss in the commit message, that won't help you	22:11
jmcantrell	heh	22:11
	← drizzd left	22:11
endpoint_david	or perhaps excludes unless explicitly used in a refspec	22:11
	← robinr left	22:11
cbreak	endpoint_david: branches are not usually pushed by default. I am not sure what is required for them to be pushed, but I think it's the existence of a remote branch that is called the same	22:12
	endpoint_david stuck on 1.6.0.1 right now; don't bother suggesting the upgrade, not my call... :-)	22:12
	← ja2 left	22:13
	← jaql left	22:13
jmcantrell	yeah, i think if you have push.default = matching, it will do that on a "git push"	22:14
	→ jrmuizel joined	22:14
jmcantrell	I just read through Spolsky's tutorial on hg. It's very simple and neat, but I feel like I would be missing something, coming from git.	22:15
	← mklappst1hl left	22:16
	→ hipe joined	22:17
	→ abasit joined	22:18
	← abasit left	22:18
	← Mark__T left	22:21
	→ ntoll joined	22:21
posix4e	SOo... I am super new to git I wanna push from my local branch foo to a remote branch called origin/foo	22:21
	← jrmuizel left	22:22
posix4e	but when i try i get unabel to chdir or not a git archive	22:22
endpoint_david	posix4e: what is the command you're using?	22:22
jmcantrell	posix4e: what does "git remote" show?	22:22
posix4e	git remote shows origin	22:23
	i am trying git push origin/foo	22:23
rtyler	spearce: running the latest snapshot now, i'll let you know how it goes	22:23
	← yann left	22:23
rtyler	(I was able to push after the upgrade)	22:23
spearce	rtyler: thanks. we've done a ton of pushes on the internal server today, nobody has complained yet.	22:24
	of course now that i say that... give it 5 minutes and someone will say something	22:24
endpoint_david	posix4e: the first argument to push is the remote name, so it'd likely be 'origin' in this case; the second is the branch name	22:24
rtyler	:)	22:24
endpoint_david	so if you're pushing the branch named 'foo' to the 'origin', you'd: git push origin foo	22:24
	not origin/foo	22:24
	← radarek left	22:25
posix4e	src refspec foo does not match any	22:25
	← bambule left	22:25
posix4e	ahh i see src:dst	22:26
	→ jrmuizel joined	22:26
posix4e	sweet i did my first push	22:26
	hooray	22:26
	rtyler throws confetti at posix4e	22:27
	← JonathanT left	22:27
endpoint_david	note that src can be any ref, including an explicit sha1; comes in handy sometimes for more complicated workflows	22:27
	→ radarek joined	22:28
	← rapha left	22:29
	← jrmuizel left	22:29
	→ jrmuizel joined	22:29
	→ christophsturm joined	22:30
	→ tbf joined	22:32
AkumaStreak	arg	22:33
	i'm so used to using the same 3 git commands that i forgot all the others	22:33
	if i git log a file	22:33
	to see revision history	22:33
	how do i checkout one of those revisions	22:34
	← ph^_ left	22:34
AkumaStreak	(go back in time marty mcfly style)	22:34
	→ dgl joined	22:34
rtyler	spearce: we're starting to move into a "dual master" kind of set up with our branching, i.e. the "master" branch is where current 1.xx development is occurring, and there is a 2.xx kind of branch where far newer development is occurring (isn't ready to be merged into master since we push the site from master)	22:35
	→ maxter joined	22:35
rtyler	spearce: any tips for managing these changes with Gerrit?	22:35
ndim	AkumaStreak: "git checkout 666deadbeef"	22:36
	← magcius left	22:36
AkumaStreak	won't that checkout a bunch of files associated with a commit tho	22:37
ndim	AkumaStreak: Just be aware that your branch name will not point to the now checked out commit, so you might want to "git branch foobar" to call it a (new) branch "foobar".	22:37
	→ fool_ joined	22:37
	← fool_ left	22:37
	→ fool_ joined	22:37
spearce	rtyler: push to the different branches using different refs/for/foo names? refs/for/master, refs/for/next, ...	22:37
AkumaStreak	i am google'ing. i want something like "git checkout BoostButton.pngFOR79046259efbaf4c459d7af5f54a6dfd74514773e"	22:37
ndim	Or just do the stuff with gitk and mouse clicking, if that helps you better understand what you are doing.	22:37
rtyler	spearce: should I then be creating that "next" branch in the gerrit repository to signify that something has been "merged"?	22:38
	← radarek left	22:38
rtyler	(my vocabulary in this area is lacking)	22:38
ndim	AkumaStreak: Oh, you want a single file in that rev... hmm. no idea.	22:38
	:)	22:38
	→ b14ck joined	22:38
	ndim leaves again.	22:38
b14ck	I accidentally deleted a file without using git rm. How can I check it back out real quick then git rm it?	22:38
	← steffkes left	22:38
spearce	rtyler: yea, you have two branches in gerrit then, "master" and "next". changes get uploaded to one of them, and when they merge, they merge into that branch.	22:38
rtyler	spearce: that's so crazy it might just work! ;)	22:39
	→ magcius joined	22:40
	← bdimcheff left	22:40
	← fool__ left	22:40
	← jakemarsh left	22:41
	← defectiv left	22:41
	→ Fnord^ joined	22:41
	← davidw left	22:41
	→ Moc joined	22:41
	← Logjam left	22:41
Fnord^	Hi, is there any sort of VCS web frontend that lets users make actual code changes via html forms and such?	22:41
	There's tons of browsers, but...	22:42
Moc	how do I merge 2 branch ? I have 2 branch that conflict when trying to merge. I fix the conflict in the branches, but if I update 1, I get the conflict about the other... so I need to merge them together	22:42
	← Tureba left	22:43
cbreak	you can use git merge	22:44
	← dgl left	22:44
cbreak	Moc: look at man git-merge, there's a section "HOW TO RESOLVE CONFLICTS"	22:45
Gitbot	Moc: the git-merge manpage can be found at http://git.or.cz/man/git-merge	22:45
	← jmcantrell left	22:45
	← alley_cat left	22:45
Moc	the 2 branch together, there is no conflict	22:45
	← ben_alman left	22:46
Moc	I'll just rebuild a new git repo from scratch like I used to do	22:49
cbreak	that sounds quite wasteful :)	22:49
	← til left	22:49
rtyler	spearce: wahoo! it works	22:49
	rtyler is totally in love with gerrit	22:50
Moc	I've made scripts that export everything, I then merge everything back into a new repo	22:50
	← Theravadan left	22:50
Moc	like by trying stuff, I did an error, trying to revert a merge, and it say I can't do it	22:50
	→ magn3ts joined	22:51
	→ __iron joined	22:52
	← cubuanic left	22:52
	← cehteh left	22:53
	← _iron left	22:55
	→ mrlarner joined	22:56
	→ jelly-bean joined	22:56
	← ixti left	22:56
	← maxter left	22:57
	→ maxter joined	22:58
	← jrmuizel left	22:58
	← tvw left	22:59
	← ntoll left	23:00
jelly-bean	one of the ./.git/pack/pack-*.pack files in our repo has ballooned up to 525mb	23:01
	← oliver3 left	23:01
	→ d0ugal joined	23:01
	→ warthog9 joined	23:02
	→ altrux_ joined	23:03
	→ hipe_ joined	23:03
AkumaStreak	OH	23:04
	duh	23:04
	git checkout [commit-ref] [filename]	23:04
	git continues to rule	23:04
	← bcardarella left	23:04
	← hipe left	23:05
	← justin-george left	23:05
	← altrux_ left	23:05
	← altrux left	23:06
	→ jrmuizel joined	23:07
	← pheaver left	23:08
	→ ilogger2 joined	23:50
cbreak	you can rewrite history	23:50
	→ hipe_ joined	23:50
cbreak	b636bc7ca: man git rebase -i	23:50
	or just nuke it away from the server and commit it yourself	23:50
	b636bc7ca: nuking away with man git-reset	23:51
Gitbot	b636bc7ca: the git-reset manpage can be found at http://git.or.cz/man/git-reset	23:51
b636bc7ca	thank	23:51
	s	23:51
cbreak	if you use rebase -i, you can set the commit for edit	23:51
	or if it's the last, just edit it with git commit --amend --author=somename	23:51
	anyway, all in man pages	23:51
	if anyone has already pulled that commit, you're in trouble	23:52
	→ nessundorma_ joined	23:52
	nessundorma_ → nessundorma	23:53
b636bc7ca	cbreak: So if I rebase the upstream, it will cause issues with people who've pulled the commit?	23:53
cbreak	yes	23:53
b636bc7ca	blerg	23:53
cbreak	all history rewriting will	23:53
	you have to do it on every repository that pulled	23:54
johnw	b636bc7ca: the downstreamers will need to reset to the new HEAD	23:54
b636bc7ca	ok, I can tell them to do that	23:54
johnw	reset --soft origin/XXX is the safest bet	23:54
	then if they had any changes somehow, they will appear in their index/working tree	23:55
b636bc7ca	We've just been using the commits to the upstream as part of employee reviews, and I don't want to 'cheat' for this guy.	23:55
johnw	in that case, reset --hard should be OK	23:55
jelly-bean	johnw: that script chopped my git repo from 500mb+ to only 878kb! ???	23:56
	→ hipe joined	23:57
jelly-bean	and the .git/objects dir is empty now mostly	23:57
	oh nvm its not	23:58
	→ hipe__ joined	23:58
	→ Mike\|\|gone joined	23:59
	← hipe left	23:59
	→ mithro joined	23:59
	← hipe_ left	23:59

Logs Search ←Prev date Next date→

Channels Documentation