IRCloggy #git 2009-07-26

Provider of IRC logs since 2005.

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

	2009-07-26
aeruder	but branches aren't anything but labels	00:00
	→ eMyller joined	00:00
aeruder	all they are in the git db is just 'newbranch = <sha1>'	00:00
	← ludde2 left	00:00
Altreus	oh cool ok :)	00:01
tongueroo	k, this example more closely resembles the problem, gitk http://tinyurl.com/mbaku8 and gitlog http://tinyurl.com/nhk7t5	00:04
	so im on branch tung	00:04
	and i already have too merges in there	00:04
	← coldboot\|home left	00:05
tongueroo	g/too/two	00:05
	but i want to get rid of the first merge, the "merge branch 'master' into tung	00:05
aeruder	tongueroo: try this:	00:05
	git checkout tung && git rebase -i <sha1 tung 2>	00:06
	it'll open your text editor	00:06
	delete the merge	00:06
	save and quit	00:06
	i don't know how well rebase -i works with merges	00:06
	so inspect the results	00:07
tongueroo	i know what you mean, conflicts at each step after the merge and then im looking at fixing about 40 commits :/	00:07
aeruder	tongueroo: is your messed up repository something you can make public?	00:08
tongueroo	unfortunately no :(	00:08
	heh, that bad huh	00:08
aeruder	no, but it'd make helping easier	00:08
	can you make your test repo public?	00:09
tongueroo	sure	00:09
	let me find a place to put it	00:09
	→ bremner_ joined	00:09
tongueroo	i was thinking i would do something with git revert -m XX, last time i did that though, i screwed something up later on down, because git revert applies the reverse difference as a commit, and if you merge that branch later, it will apply the reverse commit once again later, not something i wanted at the time, i only wanted to get rid of the changes from that particular branch, not from branches that i was going	00:11
	to merge into later	00:11
	anyway, setting up the test repos	00:11
aeruder	tongueroo: revert is ok if it is already public	00:11
	← cmarcelo left	00:11
aeruder	but if you can still change history	00:11
	i'd do that	00:11
	← bremner_ left	00:13
	← markelikalderon left	00:14
	→ bremner_ joined	00:14
	→ MrN_ joined	00:14
	→ MisterN_ joined	00:14
	← kingemer left	00:15
	← prusnak left	00:15
	→ bremner__ joined	00:17
	← bremner_ left	00:24
tongueroo	alright, im having problems setting up a public repos, some apache issues i dont want to deal with right now, is a tarball of the repos good enough ? http://www.tongueroo.com/repos/test_merge.tgz	00:26
	→ bremner_ joined	00:26
	← psynaptic left	00:26
	→ corni_ joined	00:27
tongueroo	i have a feeling its going to be a rebase -i to change history and bunch of conflict resolution as the best way to go in this case :/	00:29
	→ aspotashev\|eeepc joined	00:30
	← MrN left	00:30
	← drewr left	00:30
aeruder	but honestly, assuming git rebase -i does what i think it'll do on your example, there is a strong possibility that you just are going to have to resolve some conflicts (especially if your patches after the merge are on top of/dependent on changes that were made by the merge)	00:30
	tongueroo: but i just tried	00:30
	i forgot -p	00:30
tongueroo	-p ?	00:30
	← MisterN left	00:30
aeruder	git rebase -i -p <sha1 of tung2>	00:30
	→ drewr joined	00:31
redbrain	haha i made git crash :P	00:31
tongueroo	ic	00:31
redbrain	but its not git's fault	00:31
	Initialized empty Git repository in /home/redbrain/gcc-dev/.git/	00:31
	error: git upload-pack: git-pack-objects died with error.	00:31
	fatal: git upload-pack: aborting due to possible repository corruption on the remote side.	00:31
	remote: fatal: Out of memory, realloc failed	00:31
	remote: aborting due to possible repository corruption on the remote side.	00:31
	fatal: early EOF	00:31
aeruder	redbrain: sounds like you ran out of memory	00:31
redbrain	my vps only has 256mb ram	00:31
tongueroo	interesting with the -p it shows the merge commits in rebase -i	00:31
redbrain	yeah thinking i might upgrade it now :)	00:31
aeruder	tongueroo: yea, -p is --preserve-merges	00:32
	it attemps to redo the merges as merges	00:32
	and as such you can just remove a merge commit from rebase -i	00:32
tongueroo	ic, but it still seems to keep the changes from the master merge, https://gist.github.com/494646a3262e8a93d3ac	00:33
	master 1, master 2, master 3, are all from the merge	00:33
	MisterN_ → MisterN	00:33
tongueroo	not sure if thats is what i should expect	00:33
	but not what i want :)	00:33
aeruder	tongueroo: that's what you should expect, your tung branch itself contains those	00:34
tongueroo	oh yeah	00:34
	duh	00:34
	i got confused	00:34
	k, cool	00:35
	so that does what i want	00:35
	← goulik left	00:35
tongueroo	sigh, this sample project has all the commits in nice different files so that went very smooth	00:35
	sigh, its going to be a mess in this project :/	00:35
aeruder	tongueroo: if it doesn't then rebase -i -p is more or less just reapplying the patches	00:35
	→ zenwryly joined	00:35
aeruder	but yea, you may have some conflicts to resolve	00:35
	that's just how it goes	00:35
tongueroo	alright, </gripe> , time to suck it up and do it	00:35
	→ imranh joined	00:35
	← parasti left	00:35
	← imranh left	00:36
tongueroo	i show make the committer fix this :P	00:36
	aeruder : thanks so much for your help	00:36
	g/show/should	00:36
aeruder	tongueroo: no prob, i've been waiting for the longest kernel compile in the world to finish anyway	00:36
	← bremner__ left	00:36
tongueroo	heh	00:36
aeruder	i really need to go through xconfig and remove a bunch of stuff	00:36
zenwryly	If I have a remote with some large and old history that i really don't need locally, how can I get rid of that history locally?	00:37
aeruder	zenwryly: you can do a shallow clone, but honestly, its just easier to keep it around	00:38
	← RobertMeta left	00:38
aeruder	at one point there was talk about having git fetch full history of the commit/tree objects and shallow fetch the blobs to have a little more useful of a shallow clone	00:39
zenwryly	aeruder: hmm, oh well, thanks for the answer	00:39
aeruder	don't know if anything ever came of it though	00:39
	← corni left	00:39
	→ MetaCosm joined	00:40
	← d0k left	00:41
	→ EricInBNE joined	00:42
zenwryly	aeruder: what about grafts and filter-branch?	00:43
	← bremner_ left	00:44
	← MisterN left	00:45
aeruder	zenwryly: you change the history you change every single sha1	00:46
	if you have a project where you'd like to publish a cut-off history	00:47
	then that's ok as long as everyone works off of it	00:47
	and you can keep the full history .git around and people can graft it on locally if they want to use it	00:48
	← MrN_ left	00:48
aeruder	but if you're not the one in control of the .git or you don't want to essentially have a 'new' .git and a 'history' .git then you're probably out of luck	00:48
zenwryly	aeruder: ok, more details. I will be the only one accessing either the local ore remote repos. I want to accrue changes locally, push periodically to an "archival" repo with full history, and then remove old history locally to save space. Would that push be impossible with the grafts approach?	00:50
	→ fbrunel joined	00:55
aeruder	zenwryly: once you create changed history you mess up all sha1's after the history change	00:55
	← fbrunel left	00:55
	→ camwest joined	00:55
aeruder	unless you shallow clone which essentially just creates a broken repository	00:56
	changed sha1's = no push	00:57
	→ owen1 joined	00:57
aeruder	shallow clone = no push	00:57
	you could do it via format-patch and such	00:57
zenwryly	aeruder: ok, thanks	00:58
	← loincloth left	00:58
mems	aeruder: any idea why one can't push from a shallow clone?	00:58
	← CIA-9 left	00:59
owen1	how to change my .bashrc to display info about the current branch i am in and whether there I modified files?	01:00
aeruder	but honestly if its that bad, you might be best off just starting a new repository with no history and just keep around an archive .git	01:00
owen1	(link to a guide online would be great)	01:00
aeruder	mems: dunno, just reading the man page	01:00
mems	aeruder: yea, I've read that too	01:00
aeruder	mems: shallow clones just suck right now, so unless someone comes along and uses them, they're just more or less useless	01:01
mems	so it's not a design limitation, it's just that it's not implemented (yet...)	01:01
	← corni_ left	01:01
aeruder	mems: there was some talk about cloning full trees/commits and just leaving out old blobs	01:02
	which gives you a much less broken repository	01:02
	but i don't know whatever came of that idea	01:02
mems	I guess shallow cloned repositories could just grab any objects they need in the future from a remote	01:02
	→ Cromulent joined	01:03
mems	at the time they're required	01:03
aeruder	mems: i kinda doubt that'll happen, not particularly in the git way	01:03
zenwryly	aeruder: I don't suppose you know of a DVCS that does support this style of operation?	01:04
	→ CIA-8 joined	01:04
mems	hmmm, I keep seeing this "git way" around the place	01:04
lordpil	owen1 __git_ps1 with git-completion.bash	01:04
aeruder	zenwryly: nope	01:05
owen1	lordpil: not sure what u mean, but google for the rescue..	01:06
lordpil	zenwryly are you storing binary files? why are you so concerned about space?	01:06
zenwryly	aeruder: well thanks again for your time and the excellent information	01:07
aeruder	lordpil: and if the space is that big of a concern, then he needs to just start a new .git and let people who need full history graft the two .git's together	01:08
zenwryly	lordpil: well I use git for my code VCS, but I need an archival backup system for my other documents and files, so I was hoping to keep the full history on the remote and limited history locally	01:08
	→ Theravadan joined	01:09
aeruder	zenwryly: sometimes for things like that i almost think centralized works better anyway (i.e. subversion or something)	01:09
mems	zenwryly: after a quick look at svk, I think it might support such things	01:09
	← jparrott left	01:09
lordpil	git is not a backup system :\|	01:09
mems	maybe for you	01:10
zenwryly	aeruder: ugh, svn is so space inefficient locally to start with	01:10
	← eMyller left	01:10
	← orafu left	01:10
	→ orafu joined	01:11
redbrain	awh svn is just crap :P where i used to work it was disgusting it is one SVN repos for EVERYTHING	01:11
	100GB was one	01:11
	→ marvil07 joined	01:11
mems	that's what partial checkouts are for...	01:11
redbrain	then we got a new cluster system for pearforce 10TB or data ontop of 160 cpus was just insane	01:11
	→ eMyller joined	01:11
redbrain	100TB*	01:11
	it was full of crap	01:11
aeruder	redbrain: believe me, if we used svn at work i'd be happy :)	01:12
lordpil	speaking of being full of crap	01:12
redbrain	haha yeah what are you using atm?	01:12
aeruder	redbrain: telelogic cmsynergy	01:13
redbrain	aeruder: what the hell is that :S	01:13
lordpil	i have a custom fast-import front-end where i had the terrible idea of trying to normalize newlines on conversion. i want to turn this off and rebuild the repo. but that means all revisions will change (with a small userbase of ~8 people). what's the best way to manage such a change?	01:13
zenwryly	lordpil: look, I've been in open source long enought to recognize the "when we can't do what's being asked, blame the querant" pattern. The fact is that archival backup semantics are very similar to DVCS semantics (history, push/pull) and I am exploring the options. It's worth exploring even if you don't think so.	01:14
aeruder	if you haven't heard of it, that's probably because you don't have 1000's of dollars to throw away on software that is slower and less featureful than almost any free VCS (and for cheap commercial ones)	01:14
	but it is "enterprise friendly"	01:14
	→ loincloth joined	01:14
aeruder	i think its just an expensive front to charge huge gobs of money for training	01:15
lordpil	zenwryly i was only saying it wasn't designed for backup. not that it isn't a good base for backup. right now you'd be better off using something else. but i'd love to see a real git backup system	01:15
	← namelessjon left	01:15
zenwryly	lordpil: oh, ok, thanks	01:15
lordpil	aeruder oh dear, telelogic is now ibm rational...	01:15
aeruder	lordpil: yea	01:15
lordpil	maybe soon they'll merge it into clearcase!	01:15
aeruder	basically the way we use it is the lock-edit model	01:16
	so you constantly are making calls to people to unlock files	01:16
lordpil	aeruder fwiw my organization at work uses clearcase, and i wrote a bidirectional interface with git. so i work out of git 90% of the time. it sucks, but it's better than nothing.	01:17
aeruder	lordpil: i do everything in git, i keep the synergy stuff on its own branch so i can merge people's work into mine	01:17
	and then one day i have a "cm synergy" day where i throw 100's of commits into cm synergy one patch at a time	01:18
	← justincarmony_ left	01:18
lordpil	yeah i don't have a good automated way to move commits into clearcase. it's only 'mostly automated' and requires some checking	01:18
aeruder	not ideal, but better than the way most people work where its just "oh bob is working on it right now, i guess i'm just out of luck until he's finished"	01:18
	lordpil: just updating the timestamps on special files in the cm synergy checkout can completely invalidate your checkout	01:19
	← marvil07 left	01:19
aeruder	its a fragile piece of junk, the software just can't wait until it gets a chance to lose your data	01:19
	</rant>	01:20
lordpil	honestly i think my org could switch to git in 6 months to a year, but though a few people are interested, not a single one has actually been willing to work on it other than me	01:20
	→ eletuchy joined	01:20
aeruder	lordpil: the problem with git is that it takes some actual interest in using a VCS	01:20
	if you're a random clicker or a person who uses a VCS because it is a requirement then they can fsck up git really really badly	01:21
lordpil	i'd be more than happy even if we switched to bzr, hg. hell, i might even take svn. but nobody cares	01:21
aeruder	lordpil: if they don't care about switching, then you'll just end up playing the "fix my git\|hg\|bzr\|svn repo for me" game	01:22
	← DavidKlein left	01:22
	← owen1 left	01:24
	← loincloth left	01:27
	← Cromulent left	01:31
	← here left	01:32
	← justin-george left	01:33
zenwryly	aeruder, lordpil: what about some sort of selective removal of blob objects?	01:35
	→ loincloth joined	01:36
	← name left	01:38
	← eMyller left	01:41
	→ Tricon joined	01:41
	→ eMyller joined	01:42
	→ Cromulent joined	01:43
zenwryly	aeruder, lordpil: If blob objects were removed in the local repo, would fetch (or something else) be able to retrieve them later if necessary?	01:44
aeruder	zenwryly: i'm guessing not	01:45
	← camwest left	01:45
	← Theravadan left	01:48
	← relix left	01:49
zenwryly	aeruder: how about git-fetch-pack, am I mis-reading it?	01:51
aeruder	zenwryly: generally you pass your top-level commits to the other repository and say 'this is what i got, give me what i'm missing'	01:52
	well, the other repo has no way of knowing that you manually deleted some blob that supposedly is there for one of the passed commits	01:52
	you are breaking your repository	01:52
	so expect fixing it to be as painful as the breaking process :)	01:52
zenwryly	aeruder: hehe, ok	01:53
	← rudi_s_ left	01:56
	← jchris left	01:56
	← defectiv left	01:57
	← fg3_ left	01:59
	→ fg3_ joined	02:00
	→ ajsharp joined	02:05
ajsharp	is there a way to set hooks for all git repositories?	02:06
	→ scarabx joined	02:08
	→ therrg joined	02:08
	← binrapt left	02:10
	→ wshimmy joined	02:11
	→ techsurvivor joined	02:13
	← laj left	02:13
	← spaceonline left	02:15
	← Cromulent left	02:18
	← seanmccann left	02:20
	← aziz left	02:23
	← w\|shimmy left	02:23
	→ cytrinox_ joined	02:26
	→ phira joined	02:27
phira	Hello git people!	02:28
	I have a currently live directory full of files	02:28
	I have git init and imported it all	02:28
	now I would like to create an external repo that will hold the contents of this dir	02:28
	that I can dev against	02:28
	should I close from the current dir to the repo, then change something on the current dir so it knows the repo is now its pull/push point?	02:29
	close/clone	02:29
aeruder	phira: you're probably best off cloning into your bare .git and then cloning from that back into a new checkout	02:31
	that's the easy/lazy way without having to setup config variables and such	02:31
phira	ok	02:31
	← ajsharp left	02:32
phira	will cloning back cause any problems given it's a live site?	02:32
	← eMyller left	02:34
	← iamcalledrob left	02:34
	phira makes a backup, then just tries	02:35
	→ eMyller joined	02:35
	← kukks left	02:37
	← cytrinox left	02:37
	← Tricon left	02:44
	← scientes left	02:46
	→ scientes joined	02:46
	← scarabx left	02:48
	← mithraic left	02:49
	→ mithraic joined	02:51
	→ tjafk2 joined	02:53
	→ eeanm joined	02:53
eeanm	my local branch is one commit ahead of the remote version of it, I can see so in gitk	02:54
	← lordpil left	02:54
eeanm	but I can't push to it, it says its up-to-date already	02:54
	→ lordpil joined	02:54
	Sho_ → Patch_	02:55
eeanm	ok I fixed it by using another clone, still wondering what the deal was though :)	02:57
aeruder	eeanm: how old is your git?	02:58
	git --version?	02:58
eeanm	1.6.0.2	02:58
	← loincloth left	02:59
aeruder	eeanm: hm, at one point you needed to explicitely fetch to update remote refs	03:00
	and they added something where when you pushed it updated the remote refs automatically later on	03:00
	although i think 1.6.0.2 had the auto-fetch in it	03:01
eeanm	well I fetched and what not	03:01
	I was git pull, git pushing, trying to see why it wouldn't accept this thing :)	03:01
	gitorious even made an embarassing log of it all	03:02
	http://gitorious.org/~eean/amarok/gcds-dynamic	03:02
	I kind of started out a bit messed up, I had made the clone on gitorious, and then fetched and branched off of it, but the local master that I had wasn't up-to-date	03:03
	I think that started the initial confusion	03:03
	anyways	03:03
	it works now I shouldn't mess with it :)	03:03
	← __iron left	03:05
	Patch_ → Sho_	03:07
	← flaguy48 left	03:07
aeruder	eeanm: i'm not a huge fan of git push's automatic semantics, i always explicitely push with full arguments like git push <remote> master:refs/heads/master	03:07
eeanm	why the refs/heads/master	03:08
	is there really ambiguity about that?	03:09
aeruder	well, the refs/heads/master is necessary if you are creating a new branch on the remote side, usually i leave off the refs/heads/ part	03:09
	← tjafk1 left	03:09
eeanm	ah ok	03:09
Arrowmaster	i dont think it is anymore	03:09
aeruder	Arrowmaster: ah, good to know, that always seemed a little annoying	03:09
	← drewr left	03:10
	← Sho_ left	03:12
	→ Sho_ joined	03:12
	← jjore left	03:17
	← Sho_ left	03:21
	→ Sho_ joined	03:22
	← phira left	03:29
	→ wulfmax joined	03:29
	← Sho_ left	03:33
	→ Sho_ joined	03:33
	← Sho_ left	03:36
	→ Sho_ joined	03:36
	→ beta1 joined	03:38
	← Sho_ left	03:39
	→ jaggz- joined	03:39
	→ Sho_ joined	03:40
	→ scientes_ joined	03:45
	← scientes left	03:47
	→ eMyller\|busy joined	03:48
	← SingAlong left	03:50
	← eMyller left	03:51
	← Sho_ left	03:52
	→ Sho_ joined	03:53
	→ iamcalledrob joined	03:53
	→ hydrogen joined	03:54
	→ bugfux_ joined	03:54
	← bugfux left	03:54
Arafangion	Is perl a major dependency of git? (I notice that not only does its scripts require perl, but also the binaries?)	03:57
	← wulfmax left	03:58
	← scientes_ left	04:02
	← Sho_ left	04:05
	→ Sho_ joined	04:05
	← hyperair left	04:08
	→ Chillance joined	04:09
	← Sho_ left	04:11
	→ Sho_ joined	04:12
	← Linkadmin left	04:14
	→ vstemen joined	04:15
vstemen	After upgrading to scmgit-1.6.0.2 on BSD, if I do a "git commit" of a new project I get a bunch of errors like	04:17
	Makefile:3: trailing whitespace.	04:17
	+DESTDIR ?=	04:17
	Makefile:23: trailing whitespace.	04:17
	+	04:17
	Then it just gives me a prompt back without committing.	04:17
	.	04:18
	Any Ideas?	04:18
	→ justin-george joined	04:19
	← Sho_ left	04:22
	← justin-george left	04:25
	← aspotashev\|eeepc left	04:25
kahmalo	vstemen: check your hooks, I think specifically pre-commit	04:30
	← eletuchy left	04:30
	→ eletuchy joined	04:31
	→ eletuchy_ joined	04:31
	← eletuchy left	04:31
	→ synusoida joined	04:36
	← Tommy[D] left	04:41
	→ Tommy[D] joined	04:41
	→ fridim_ joined	04:49
	← neoeinstein left	04:49
	← Tommy[D] left	04:50
	← jbwiv left	04:50
	→ Tommy[D] joined	04:51
	→ jbwiv joined	04:52
	← stuffcorpse left	04:53
	← psytek left	04:54
vstemen	kahmalo, there is a pre-commit hook. The last line of the script is	04:54
	exec git diff-index --check --cached $against --	04:54
	→ neoeinstein joined	04:55
vstemen	Is the default hook in git-1.6.0.2 broken?	04:55
	← krig_ left	04:55
	→ krig joined	04:55
	→ tazle_ joined	04:55
	← hydrogen left	04:55
	← Pewpewarrows left	04:56
vstemen	Yep, that's what is causing the error. That last line expands to	04:58
	git diff-index --check --cached 4b825dc642cb6eb9a060e54bf8d69288fbee4904	04:58
Gitbot	[git 4b825dc64]: http://tinyurl.com/55rga7 [tree]	04:58
vstemen	Which causes the errors I am getting.	04:58
	So, why is git installing a broken hook when I init the repo? What should I do, delete .git/hooks/pre-commit ?	05:00
	→ Linkadmin joined	05:02
vstemen	Well, looks like removing .git/hooks/* worked.	05:04
	thanks kahmalo for pointing me in the right direction	05:05
	It doesn't give me a very warm fuzzy feeling about gits stability with obvious bugs like that being the the release. I hope there are not many of these.	05:06
	fr → fr\|	05:08
	← hobodave left	05:08
	→ Flathead joined	05:09
	→ MetaCosm_ joined	05:10
	→ reprore joined	05:13
	← reprore left	05:22
	← Arrowmaster left	05:22
	→ reprore_ joined	05:23
	→ Arrowmaster joined	05:23
	← voker57 left	05:23
	→ voker57 joined	05:24
	→ Sigma joined	05:26
	→ jjore joined	05:26
	← MetaCosm left	05:27
	→ eMyller joined	05:34
	← sobersab1e left	05:34
	→ Theravadan joined	05:39
	→ seanmccann joined	05:44
	→ jason237 joined	05:44
	→ phira joined	05:45
jason237	hmm, setting a repository as an alternate for itself seems to make git gc do bad things...oops	05:45
	→ saurabhverma joined	05:46
	← sunoano left	05:50
	→ EruditeHermit joined	05:50
	→ yhager_ joined	05:51
	→ ben_h joined	05:52
yhager_	The source is under SVN. I want to work with git, but on two different locations, and use push-pull between the two. eventually to dcommit back to SVN. How should I build this?	05:52
	yhager_ → yhager	05:52
	→ stuffcorpse joined	05:56
Ilari	yhager: SVN and distribution doesn't mix very well...	05:57
	← eMyller\|busy left	05:58
yhager	I tried 'git svn clone' on A, then pushed to bare on B, and cloned it on C. I then push & pull to and from B, but I have the feeling I am doing something wrong (maybe I should use different branches?)	05:58
Ilari	yhager: The biggest problem is that SVN and merges don't mix. And pull is fetch + merge unless you specify otherwise.	06:00
yhager	Ilari: can you elaborate on that? I can either use SVN directly, or git + git-svn. I prefer the latter, if it works..	06:01
	joshthecoder → joshthecoder_afk	06:02
Ilari	yhager: git svn is quite limited due to limitations of SVN.	06:02
yhager	Ilari: ok, I see your point. So I guess I'll just do 'git svn clone' and 'git svn dcommit', and not try any of the "fancy" git suff..	06:04
Ilari	yhager: Yeah. At least better to do that way unless you know what you are doing...	06:04
yhager	Ilari: allright. Thanks.	06:04
	← MetaCosm_ left	06:06
	← jfxberns left	06:06
	← bgerber left	06:06
	← jksz left	06:06
	← harv left	06:06
	← mutex left	06:06
	← agile left	06:06
	← ryanakca left	06:06
	← moh left	06:06
	← Luke-Jr left	06:06
	← mithraic left	06:06
	← CIA-8 left	06:06
	← elliottcable left	06:06
	← xenoterracide left	06:06
	← slack left	06:06
	← m0 left	06:06
	← zwu left	06:06
	← azuk` left	06:06
	← jarh left	06:06
	← Thumper_ left	06:06
	← a-atwood left	06:06
	← blankthemuffin left	06:06
	← emias left	06:06
	← ruskie left	06:06
	← Fatal left	06:06
	← ciaranm left	06:06
	← deavid left	06:06
	← troyt left	06:06
	← tmz left	06:06
	← gebi left	06:06
	← teuf left	06:06
	← runasand left	06:06
	← yzlin left	06:06
	← stegano left	06:06
	← univate left	06:06
	← panfist left	06:06
	← frsk left	06:06
	← jonmasters left	06:06
	← TDJACR left	06:06
	← eikke left	06:06
	← ahf left	06:06
	← dorileo left	06:06
	→ sunoano joined	06:08
Ilari	yhager: Basically merges + git-svn => trouble.	06:09
yhager	Ilari: I see. I guess I was expecting too much from (git-)svn..	06:10
Ilari	SVN just plain doesn't support merges in sense Git uses them. No, not even SVN 1.5+	06:11
	Also, atomicity guarantees are much weaker. Git has atomic branch updates (strictly stronger requirement than atomic commits) while SVN only has atomic changesets (strictly weaker requirement than atomic commits).	06:15
	In fact, the atomicity guarantees are not sufficient to implement git svn in race-free manner.	06:15
EruditeHermit	hi, can anyone help me with git push?	06:18
	I created a branch to my repository and now I want to push that branch to a remote repository but I keep getting the message "Everything up to date"	06:18
	and it doesn't do anything	06:19
rubyist	git push origin/new_branch	06:19
Ilari	EruditeHermit: You have to specify that branch to be pushed since it doesn't already exist. Also, note that first non-option to push is URL or nickname of the repository to push to, the following is names of stuff to push or mapping rules.	06:20
	rubyist: Doesn't look sane.	06:20
EruditeHermit	error src refspec origin/highlighting does not match any	06:21
rubyist	Ilari: always worked for me :P	06:21
EruditeHermit	so if I have 2 branches master and highlighting is it origin/highlighting?	06:22
Ilari	EruditeHermit: 'git push origin highlighting'?	06:22
EruditeHermit	Ilari: there should be a space?	06:22
Ilari	EruditeHermit: Yes.	06:22
EruditeHermit	error src refspec origin does not match any	06:22
	oh	06:23
	it worked with just git push <repo> highlighting	06:23
Ilari	EruditeHermit: What did you actually try to run? 'git push origin origin highlighting'???	06:23
EruditeHermit	no	06:23
	I tried git push <repon_url> origin highlighting	06:24
	→ priidu joined	06:24
EruditeHermit	as well as git push <repo_url> origin/highlighting the first time	06:24
Ilari	EruditeHermit: The proper syntax to push to origin is 'git push origin highlighting'.	06:24
EruditeHermit	oh	06:24
Ilari	EruditeHermit: origin is also repository designation.	06:24
EruditeHermit	so origin replaces the repo_url	06:24
	ah	06:24
	that makes more sense	06:24
Ilari	EruditeHermit: For extra confusion 'origin' may also be local branch name, tag name, remote tracking branch (that's actually fairly common) or something even more exotic.	06:26
	← Bass10 left	06:26
EruditeHermit	awesome	06:26
	not going to remember that until I hit all those problems	06:27
	=p	06:27
	Ilari: so one more question	06:27
Ilari	EruditeHermit: What push tries to interpret those names as depends on their position in command line.	06:27
	← priidu left	06:27
	→ priidu joined	06:28
EruditeHermit	so now I have my local repository with 2 branches	06:28
	where in the file structure will I see the different files associated with each branch?	06:28
Ilari	Repository URL can be replaced by remote name in all user-level commands except in clone (for obivious reasons).	06:28
	EruditeHermit: That is, where are the versions stored?	06:29
EruditeHermit	yeah	06:29
Ilari	EruditeHermit: Inside the .git directory (heavily compressed)...	06:29
EruditeHermit	say I want to test both versions of the program	06:29
	do I need to checkout each branch do different places?	06:29
Ilari	EruditeHermit: man git-archive?	06:30
Gitbot	EruditeHermit: the git-archive manpage can be found at http://git.or.cz/man/git-archive	06:30
	→ Cromulent joined	06:30
	→ mutex joined	06:32
	→ agile joined	06:32
	→ harv joined	06:32
	→ jksz joined	06:32
	→ bgerber joined	06:32
	→ jfxberns joined	06:32
	→ MetaCosm_ joined	06:32
	→ ahf joined	06:32
	→ deavid joined	06:32
	→ azuk` joined	06:32
	→ zwu joined	06:32
	→ a-atwood joined	06:32
	→ teuf joined	06:32
	→ dorileo joined	06:32
	→ blankthemuffin joined	06:32
	→ ruskie joined	06:32
	→ eikke joined	06:32
	→ jarh joined	06:32
	→ m0 joined	06:32
	→ Fatal joined	06:32
	→ yzlin joined	06:32
	→ TDJACR joined	06:32
	→ slack joined	06:32
	→ runasand joined	06:32
	→ univate joined	06:32
	→ Thumper_ joined	06:32
	→ tmz joined	06:32
	→ jonmasters joined	06:32
	→ emias joined	06:32
	→ ciaranm joined	06:32
	→ frsk joined	06:32
	→ panfist joined	06:32
	→ stegano joined	06:32
	→ gebi joined	06:32
	→ troyt joined	06:32
	→ xenoterracide joined	06:32
	→ moh joined	06:32
	→ Luke-Jr joined	06:32
	→ ryanakca joined	06:32
	→ elliottcable joined	06:32
	→ mithraic joined	06:32
	sunoano → Guest47510	06:32
	→ CIA-78 joined	06:32
	← reprore_ left	06:32
EruditeHermit	Ilari: so from the examples in the man page all of them seem to want to output to zipped archives	06:35
	Ilari: what if I just want to create a directory with it?	06:35
Ilari	EruditeHermit: Create tar archive and pipe it to tar (hint: see -C option to tar).	06:37
EruditeHermit	Ilari: so when looking at the git directory with a filemanager, the files you see are always the mater branch?	06:40
	master*	06:40
	→ jjuran joined	06:40
Ilari	EruditeHermit: The working copy is from currently checked out branch.	06:41
	EruditeHermit: Use either 'git branch' or the shell prompt thingy to display what it is.	06:41
EruditeHermit	oh	06:41
	nice	06:41
	I don't have to create a new archive then	06:41
	I can just checkout the branch I want	06:42
	→ hobodave joined	06:42
	← Guest47510 left	06:42
	→ sunoano joined	06:43
	→ Yewbacca joined	06:44
	← YewbaccaZzz left	06:45
	← Cromulent left	06:48
	→ YewbaccaZzz joined	06:49
	← Yewbacca left	06:50
	← mithro left	06:53
	→ mrbendig joined	06:56
	← seanmccann left	07:00
	→ sprasanna joined	07:02
	← bobesponja left	07:05
	→ Weaselweb joined	07:08
	→ mithro joined	07:12
	← mithro left	07:17
	→ drizzd joined	07:20
	← YewbaccaZzz left	07:22
	→ pascal2 joined	07:25
	→ juanjoc joined	07:25
	← cloudhead left	07:26
	→ sprasanna_ joined	07:28
	← techsurvivor left	07:29
	→ [Po]lentino joined	07:35
	← tongueroo left	07:36
	→ marvil07 joined	07:36
	← Chillance left	07:37
	→ jceb joined	07:39
	→ mithro joined	07:39
	← juanjoc left	07:42
	← sprasanna left	07:43
	← jaggz- left	07:43
	← sprasanna_ left	07:52
	→ angerman joined	07:52
	→ Sigma[Mtp] joined	07:53
	← pascal2 left	07:55
	→ seanmccann joined	07:59
	← seanmccann left	07:59
	→ porcnick joined	08:00
	← bugfux_ left	08:01
	→ bambule joined	08:04
	→ sprasanna joined	08:05
	← Sigma left	08:05
	→ reprore joined	08:07
	→ lordpil- joined	08:10
	→ kumbayo joined	08:11
	← brizly1 left	08:11
	← marcoecc left	08:12
	→ brizly joined	08:12
	→ lack_ joined	08:15
	→ Sho_ joined	08:17
	← Theravadan left	08:18
	← phira left	08:18
	→ doener joined	08:23
	← cpg left	08:24
	lack_ → lack	08:26
	← reprore left	08:27
	← lordpil left	08:27
	← eMyller left	08:29
	→ Tobin joined	08:30
	→ patmaddox joined	08:31
	← beta1 left	08:31
	← Sho_ left	08:33
	→ Sho_ joined	08:33
	→ ludde joined	08:33
	← ben_h left	08:35
	→ viric joined	08:38
	← bambule left	08:40
	→ Yewbacca joined	08:40
	← fg3_ left	08:43
	→ markelikalderon joined	08:45
	← marvil07 left	08:47
	→ phira joined	08:50
	→ kumbayo_ joined	08:50
	→ mithro_ joined	08:55
	← EruditeHermit left	08:56
	→ sycofly joined	08:58
	→ mbroeker joined	08:59
	← iamcalledrob left	08:59
	← kumbayo left	09:00
	← jfxberns left	09:01
	← mithro left	09:02
	→ iamcalledrob joined	09:02
	← iamcalledrob left	09:03
	← neoeinstein left	09:04
	→ relix joined	09:04
	→ neoeinstein joined	09:05
	→ adeodatus joined	09:09
	← adeodatus left	09:10
	← jjuran left	09:10
	← saurabhverma left	09:13
	→ patrikf joined	09:14
	→ Grahack joined	09:14
	← Grahack left	09:15
	→ jmspeex joined	09:17
	← kpreid_ left	09:21
	→ kpreid_ joined	09:23
	→ UUncia joined	09:24
	→ jjuran joined	09:24
	→ psynaptic joined	09:26
	→ khmarbaise joined	09:28
	→ saurabhverma joined	09:29
	→ fg3_ joined	09:32
	← EvanDotPro_ left	09:34
	→ EvanDotPro_ joined	09:34
	→ solydzajs joined	09:35
	← EvanDotPro_ left	09:35
	→ EvanDotPro_ joined	09:35
	→ ciskje joined	09:35
	← saurabhverma left	09:35
	← phira left	09:37
	← mrbendig left	09:37
	← Sho_ left	09:42
	→ Sho_ joined	09:42
	→ shruggar joined	09:43
	→ kinkerl joined	09:43
	→ yann joined	09:50
	← markelikalderon left	09:52
	← akitada left	09:56
	→ akitada joined	09:57
	← Sho_ left	09:59
	← eletuchy_ left	10:06
	→ eletuchy joined	10:06
	→ gitte joined	10:09
	← ciskje left	10:15
	→ hyperair joined	10:19
	← mutex left	10:20
	→ d0k joined	10:20
	→ mutex joined	10:21
	→ davegb3 joined	10:21
	→ DavidKlein joined	10:21
	→ ben_h joined	10:23
	← eletuchy left	10:23
	← jmspeex left	10:23
	← ben_h left	10:23
	→ ben_h joined	10:24
	→ __iron joined	10:25
	→ corni joined	10:28
	← viric left	10:31
	→ klbate_ joined	10:32
	← akitada left	10:32
	→ sitaram joined	10:34
	→ quizme joined	10:37
quizme	i accidentally deleted a whole directory	10:37
	← sycofly left	10:37
	→ GeertB joined	10:37
quizme	how can i check out from HEAD only that directory plus everything under it ?	10:38
	nm	10:39
	→ Yuuhi joined	10:39
wereHamster	quizme: git checkout HEAD -- <directory>	10:39
quizme	wereHamster thank u.... lol funny name	10:39
	← Tobin left	10:41
	← klbate left	10:42
	→ Cromulent joined	10:49
	← JEEB left	10:52
	→ parasti joined	10:54
	→ volodya joined	10:55
	← hyperair left	10:58
	→ chris2 joined	11:01
	→ JEEB joined	11:01
	→ akitada joined	11:03
	→ name joined	11:09
	→ EruditeHermit joined	11:10
	← EruditeHermit left	11:10
	← kinkerl left	11:11
	← emias left	11:15
	→ emias joined	11:15
	→ justin-george joined	11:15
	← volodya left	11:16
	← mithro_ left	11:22
	→ mithro_ joined	11:22
	← justin-george left	11:24
	← ^Einstein left	11:24
	← brizly left	11:25
estan	hrm. i accidently did a commit when in "Not currently on any branch" state, i'm not sure how i ended up in that state. now i'd like that commit to be on my branch, which i checked out..	11:25
teuf	you can use git cherry-pick SHA1 if you know the sha1 of this commit	11:27
wereHamster	estan: git cherry-pick <shaofthatcommit>	11:27
estan	yea i was thinking that should work. it will become a new commit on the branch though, with its own sha, right?	11:27
	→ brizly joined	11:28
teuf	yep	11:28
estan	hm. okay.	11:29
	any idea how i could have ended up in that state btw?	11:30
	because i'm looking at my bash history and i can't see that i did anything that would cause it.	11:30
parasti	estan: git reflog might have a clue	11:30
estan	okay, i'll look that up.	11:31
wereHamster	estan: that state is called 'detached head', you can read more about it here: http://sitaramc.github.com/concepts/detached-head.html	11:32
	← akitada left	11:33
	← Cromulent left	11:33
estan	wereHamster: yea i was actually looking at that URL just yesterday just out of curiosity (since someone pasted it here).	11:34
	i should read it again i guess.	11:34
	what i've done now after realizing i did a commit on detached head is just `git checkout my_branch`.	11:35
	and git reflog show now shows:	11:35
	→ Hendu joined	11:35
estan	9316906... HEAD@{0}: checkout: moving from 563ff11ecb920af32bc290de0f73225e22e68dc4 to tables-gsoc	11:35
	563ff11... HEAD@{1}: commit: Doc fix.	11:35
	0044504... HEAD@{2}: commit: Silly mistake of storing the position the table layout instead of	11:35
	but i don't get it, since that 0044504 commit ended up on the tables-gsoc branch, but the 563ff11 one didn't (ended up on a detached head)..	11:36
	i didn't do anything in between..	11:36
	i'll go read that url for a while.	11:37
	→ drewr joined	11:37
doener	estan: that part of the reflog that show how you ended up on a detached HEAD.	11:37
	estan: look for another "checkout: " line	11:38
estan	hm yes i found it now, 7383362... HEAD@{6}: checkout: moving from tables-gsoc to 73833625f8709d02ad611210d39daca0156814b4	11:40
	but i have no idea why it happened, and it's so weird, because there's a commit after this checkout: line that still ended up on the tables-gsoc branch (shows up in `git log' on that branch), and one commit that didn't.	11:40
	and no other checkout: between them.	11:41
doener	estan: did you by chance use "git rebase"?	11:41
wereHamster	estan: are you sure the commit in the reflog and on the branch are the same (have the same sha1)? Maybe you cherry-picked it or something	11:41
doener	estan: what does "git branch --contains 0044504" say?	11:41
estan	doener: hum. it says nothing.	11:42
	so that means it was on a detached head too then?	11:42
	how come it shows up in `git log' on the branch?	11:42
doener	estan: "git log tables-gsoc" shows it?	11:43
	→ resmo joined	11:43
resmo	hi	11:43
	is there anything like "git ls" to see which files are tracked?	11:44
estan	oh wait!	11:44
	sorry i'm stupid, no it doesn't show.	11:44
doener	resmo: ls-files	11:44
estan	so both those commits were accidently made on a detached head.	11:44
	still not sure how i ended up there though.	11:44
doener	estan: did you use "git rebase"?	11:44
estan	i mean, what caused the 7383362... HEAD@{6}: checkout: moving from tables-gsoc to 73833625f8709d02ad611210d39daca0156814b4	11:44
	→ Cromulent joined	11:46
estan	doener: hm. yes i did an interactive rebase to change a commit message.. i can see it right after the checkout: line in the reflog.	11:46
	doener: could that have caused it?	11:46
	(and if so, why?)	11:46
doener	estan: rebase works on a detached HEAD, and only updates the branch head (and checks it out again) once the rebase is finished	11:47
	estan: i.e. you probably never finished the rebase	11:47
estan	doener: ah, i'm starting to remember now.	11:47
	what's the correct way to amend a commit? i probably never did git rebase --continue or something?	11:47
doener	estan: likely. After amending the commit, you need to "rebase --continue", so the other commits are rebased onto the amended one	11:48
estan	i think i only did the `git commit --amend`.	11:48
	right.	11:48
	how do i get out of this mess? ;) just cherry pick those commits on the detached head into my branch?	11:48
	the thing is that i already cherry-picked them into master and `git svn dcommit'ed them from there.	11:49
	(before realizing they were made on the detached head).	11:49
doener	estan: you probably still have the rebase status stuff. rebase --continue might create a mess now, so I'd do a rebase --abort, eventually do a new rebase, to amend the commit again, and then cherry-pick the other commits	11:49
estan	(using git-svn against KDE SVN).	11:49
doener	estan: ugh	11:50
estan	yea, they're moving to git eventually.	11:50
doener	estan: then probably something liek: rebase --abort, git svn rebase, check history, cherry-pick	11:50
	estan: or something like that	11:50
estan	hm. yea. i'll see if i can sort it out.	11:51
	(only using master as a 'transport' out into KDE SVN, even have it in a separate working directory using git-new-workdir from contrib/).	11:51
	→ asksol joined	11:52
	→ pisi joined	11:52
	← kpreid_ left	11:52
	→ kpreid_ joined	11:53
estan	all this because i was vain and wanted to fix a spello in a commit msg ;)	11:53
	← Hendu left	11:54
pisi	Is there some easy/documented way of putting a git commit hash into autoconf version tags, without manual file-editing before generating the configure script?	11:54
	Something that would help with nightly builds...	11:55
	← Yewbacca left	11:55
	← resmo left	11:55
doener	pisi: http://gitorious.org/~flameeyes/pulseaudio/flameeyes-pulseaudio/blobs/master/git-version-gen	11:55
	pisi: I guess there might be a newer version of that script somewhere... It's just what google threw at me	11:56
	→ ludde2 joined	11:56
estan	doener: ah just aborting the rebase and then rebasing against SVN actually solves the problem, as i'll be getting those two commits from there (since i git svn dcommited them to SVN before realizing they were on a detached HEAD).	11:58
	all well now.	11:58
pisi	doener: ok, this is not much different that i have currently. But does a more meaningful version tag indeed. thanks	11:59
	← pisi left	11:59
	→ eletuchy joined	12:00
	← gitte left	12:02
	← GeertB left	12:03
	← [Po]lentino left	12:03
	→ aziz joined	12:05
	← patmaddox left	12:05
	← ludde left	12:08
	← hummingkid left	12:10
	→ moonflux joined	12:15
	← khmarbaise left	12:18
	→ saurabhverma joined	12:21
	→ aspotashev\|eeepc joined	12:22
	← eletuchy left	12:24
	→ eletuchy joined	12:25
	← TDJACR left	12:29
	→ khmarbaise joined	12:29
	← saurabhverma left	12:30
	← ben_h left	12:32
	← patrikf left	12:32
	→ TDJACR joined	12:36
	← ThaDon left	12:36
	← mithro_ left	12:40
	← Cromulent left	12:42
	← dodo_the_last left	12:42
	→ Cromulent joined	12:43
	→ scarabx joined	12:43
	→ dodo_the_last joined	12:44
	→ ben_h joined	12:51
	← ben_h left	12:52
	← synusoida left	12:53
	→ ben_h joined	12:58
	← Flathead left	13:00
	← ben_h left	13:01
	→ kukks joined	13:02
	→ jfxberns joined	13:03
	→ hydrogen joined	13:03
	→ bremner_ joined	13:06
	→ eMyller joined	13:07
	→ mithro joined	13:09
	→ Yewbacca joined	13:09
	→ jfxberns_ joined	13:11
	→ Bass10 joined	13:12
	← xkr47 left	13:12
	← bremner_ left	13:13
	← EricInBNE left	13:13
	→ bambule joined	13:14
	← UUncia left	13:15
	← jjuran left	13:16
	→ xkr47 joined	13:20
	→ mithro_ joined	13:22
	→ rudi_s joined	13:23
	→ SjB joined	13:24
	→ bremner_ joined	13:26
	→ bremner__ joined	13:27
	← bremner__ left	13:27
	→ bremner__ joined	13:28
	→ neunon joined	13:29
	← jfxberns left	13:29
	← mithro left	13:31
	→ ciskje joined	13:36
	← Sigma[Mtp] left	13:36
	→ Sigma joined	13:36
	← hydrogen left	13:36
	← ppawel left	13:43
	← bremner_ left	13:44
	← therrg left	13:46
	← jceb left	13:47
	← bremner__ left	13:48
	← SjB left	13:48
	← sunoano left	13:50
	→ bremner_ joined	13:51
	← bremner_ left	13:52
	→ bremner_ joined	13:52
	← bambule left	13:57
	→ bambule joined	13:58
	← ashH left	13:58
	← Cromulent left	14:04
	→ Octalot joined	14:04
	→ djgera joined	14:08
	← bremner_ left	14:08
	→ Alinermedia joined	14:08
	→ sunoano joined	14:08
	→ kevin_ joined	14:09
	← mbroeker left	14:09
	→ jparrott joined	14:10
	→ hyperair joined	14:11
	← ChezaWho left	14:14
	→ sprasanna_ joined	14:14
	→ khmarbaise_ joined	14:15
	→ mbroeker joined	14:17
	→ Kermos joined	14:18
	← fridim_ left	14:19
Kermos	hey all, anyone know of a way to restrict git clone access? Got my repository running nicely on my server but don't want to give anyone in the world access to it.	14:19
cbreak	remove read access	14:19
Kermos	well there are some people who do need read access	14:20
	just not everyone :)	14:20
	or anyone rather	14:20
spb	don't run git-daemon for that repo	14:20
	require people to use git+ssh	14:20
Kermos	hmmm ok I'll give that a try	14:21
	→ eMyller\|busy joined	14:21
Kermos	ok when I try this, (git-daemon turned off), git clone ssh+git://repoaddress, I get asked for my ssh password and then get this error:	14:23
	fatal: '/orion.git': unable to chdir or not a git archive	14:23
Ilari	Kermos: Is orion.gti in filesystem root?	14:24
	Kermos: Corresponding cd would be 'cd /orion.git'.	14:24
Kermos	yea	14:24
	← sprasanna left	14:24
Kermos	oh	14:24
	hmm...	14:24
	← bambule left	14:24
	← jfxberns_ left	14:24
	→ jfxberns joined	14:25
Ilari	Kermos: I suspect 'ls /orion.git' on server would complain about nonexistent file...	14:25
Kermos	it would	14:25
	that path works fine with git-daemon	14:25
Ilari	Kermos: Fix the path in URL then.	14:25
Kermos	but I guess not without	14:25
Ilari	Kermos: git-daemon has --base-path	14:25
Kermos	ahh, gotcha	14:25
Ilari	Kermos: SSH with (restricted) shell doesn't.	14:26
	← eMyller left	14:26
Kermos	had to adjust permissions too it seems after adjusting the path :)	14:27
	that works now, thanks	14:27
Ilari	Kermos: Depending on access requirements, traditional Unix groups may not be sufficient.	14:28
Kermos	Well the only people that would have any user account on the server are members on the team that are ok to have read access. I just don't want the repository exposed to the whole world.	14:30
	← cedricv left	14:31
	← khmarbaise left	14:31
Ilari	Kermos: They all should also have write access?	14:32
Kermos	no, commits go through me	14:33
	so I only need write access	14:33
	← eMyller\|busy left	14:33
cbreak	there are acls	14:36
	→ eMyller joined	14:36
Kermos	well the repository is configured using gitorious with only me listed for write access	14:36
	so I'd expect that to be ok as far as that goes	14:37
Ilari	Kermos: If it asks for password, you very likely screwed up somewhere.	14:37
	→ ppawel joined	14:39
Kermos	it doesn't if I use [email@hidden.address] to clone as my path	14:39
	→ saurabhverma joined	14:39
Kermos	or rather git@server:repository, sorry	14:40
	→ AkumaStreak joined	14:41
	→ Kruger joined	14:41
	→ ashH joined	14:42
Ilari	Kermos: Yeah, better to use such URLs.	14:42
Kermos	I did give git my public key, which I know is why it doesn't ask for a password if I use the git@server url	14:43
	ok	14:43
	I'll use that then, prefer that anyway as it doesn't expose the whole path to the repo	14:43
	→ orb joined	14:44
Kermos	thanks very much, help is much appreciated :)	14:44
	→ legendulo joined	14:47
orb	Within the last year, have there been any advances in people wanting to split a commit?	14:49
	→ MetaCosm joined	14:51
	← kahmalo left	14:52
cehteh	orb: huh?	14:53
orb	Like, when the topmost commit has changes to foo.c and bar.c, and one wants to amend HEAD to only have foo.c's changes and put bar.c's changed into a new commit on top of HEAD.	14:55
wereHamster	git reset HEAD~1 -- bar.c; git commit --amend; git reset HEAD@{1} -- bar.c; git add bar.c; git commit	14:56
orb	a reflog hack, heh :)	14:57
	nice	14:57
wereHamster	maybe there's even an easier way, I don't know	14:57
orb	So far did it via git reset --soft HEAD^; git stash; git stash apply; git add foo.c; git commit; git add bar.c; git commit;	14:58
	→ saurabhverma1 joined	15:00
	→ kahmalo joined	15:00
	→ juanjoc joined	15:00
kahmalo	git cvsimport nowadays runs git repack -a -d. Should it run git gc --auto instead?	15:01
	→ Base2 joined	15:02
	← legendulo left	15:02
kahmalo	near the end of git cvsimport, it runs git-count-objects and checks whether there are more than 1024 kilobytes of unpacked objects	15:02
Base2	Hey all, on my macbook pro, when I run "vim", my arrow keys work fine. In git, when I "git commit" and vim pops up, my arrow keys don't work. Any ideas?	15:03
orb	kahmalo: I don't think so. A fresh import should ideally be completely packed, like a git clone	15:03
bremner	Base2: is your EDITOR set to "vi" or "vim"	15:04
orb	Base2: could be related to $TERM one way or another	15:04
Base2	export EDITOR="/opt/local/bin/vim"	15:04
kahmalo	orb: when running git cvsimport incrementally, it the git repack -a -d often leaves the amount of unpacked objects still over the 1024 kilobyte limit, so it'll repack on the next run too	15:04
	perhaps because of reflogs in my local branches, I don't know	15:04
Base2	Also, when I do svn commit, vim is fine	15:04
kahmalo	whereas git gc --auto would just skip the repack.	15:04
	→ goulik joined	15:05
wereHamster	orb: erm, no. Import from foreign vcs is usually done commit by commit, which leaves you with lots of loose objects	15:05
	← Kruger left	15:05
Base2	Strange, I have two vim binaries... thats the wrong one... never mind	15:05
orb	wereHamster: but it does autopack on git-svn clone after 1000 or so	15:05
Base2	orb: ya and makes git-svn so painfully slow on a huge svn repo	15:06
orb	Base2: nah I think the larger problem with svn is network latency ;)	15:06
wereHamster	orb: it=git-svn, not git core. The vcs import script has to tell git to repack the repo after it is done	15:06
	← MetaCosm_ left	15:07
	← chris2 left	15:09
	← sergiodj left	15:19
	← saurabhverma left	15:19
	→ marvil07 joined	15:19
kahmalo	quite a lot seems to be missing from gitrepository-layout(5): info/attributes, rr-cache, MERGE_RR, FETCH_HEAD, MERGE_HEAD, ORIG_HEAD at least	15:21
	← Kermos left	15:22
hyperair	ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEA3kAM93zX6eN9OH7BjWS9kclC+QKky0VdoUxxHnNytySsSj9RRybMVdyePSL/8rNsSa/9bTEq4+NslNCfH32dwbQouyAstFrKmOaG1ejRkjikjvY/VJEPCLuZGY9iKFoiZhSr2sZi4ewg2FUuYi8Zv5Lxa7io7qJucyd3z7xOhgbcfmtzx21lQryR2jClJuBZvykxTHNFCnh03K5j9yAqnIfjfX72xOpVSaYhN8TAHnAokIXkQ2Nm/PS6ThFTfa/TgZpmYEdMdt1X3do5y6igimxufE5RWw3OfvVWfCbRDSKN5qDD+sa9+eoD5w+EF8EoQ9+hwqxQIkZ1L5PRrrHj/Q== General shell key	15:22
cbreak	interesting.	15:23
wurble	Why would git fail over an sshfs mount?	15:23
	← quizme left	15:28
	→ bambule joined	15:30
hyperair	oh shit, did i accidentally paste that in here?	15:30
	!@#$	15:30
	hyperair wonders how that even got in here	15:31
hyperair	i don't remember middle clicking or anything of that sort	15:31
	→ Rich_Morin joined	15:32
	→ Cromulent joined	15:38
	→ reprore joined	15:41
	← saurabhverma1 left	15:41
	→ fabiand1 joined	15:42
fabiand1	Hey. is there a german git channel? :)	15:43
orb	Can't use English?	15:43
	→ ben_h joined	15:43
fabiand1	sometimes - liek now - it's even quite complicated to explain in my first language ..	15:44
	fabiand1 → fabiand-prop	15:44
fabiand-prop	well. I want to reduce N commits to one.	15:47
Ilari	hyperair: At least its pubkey... :-)	15:47
	fabiand-prop: sounds like interactive rebase.	15:47
hyperair	Ilari: yeah, at least it's just that.	15:48
	hyperair was setting up a gitorious account	15:48
fabiand-prop	ok ..	15:49
	→ Theravadan joined	15:49
	← eMyller left	15:50
hyperair	fabiand-prop: you want interactive rebase	15:50
	fabiand-prop: that one allows squashing of commits.	15:51
fabiand-prop	hyperair: yes. i'm currently reading it ..	15:51
	hyperair: i'm looking for a way to automate this ..	15:51
	→ eMyller joined	15:51
hyperair	ah	15:51
	automate.	15:51
	that sounds hard. is that even possible?	15:51
	wait, i think if you change the EDITOR environment variable to a custom script, then git rebase -i will call up that editor, whereby you can get it to squash the commits.	15:52
	← davegb3 left	15:52
Ilari	Well, it is theoretically possible if commits to squash are expressible by range expression.	15:53
hyperair	hmm?	15:53
	how?	15:53
	hmm come to think of it, a graft should be able to squash commits quite well.	15:53
Ilari	Except that commit messages...	15:54
hyperair	yes, true	15:54
	but he wants it automated. how do you automate that?	15:54
Ilari	No easy way. But at least in some cases it is possible to automate (then there are cases where it isn't).	15:56
hyperair	hmm filter-branch allows you to drop commits it seems	15:57
	which is basically squashing them, similar to using grafts	15:57
	no way to specify the message though	15:57
fabiand-prop	hmm i figured out a way	15:57
hyperair	how?	15:57
fabiand-prop	i create a branch A, do my N commits in A, then, creating a new branch B, taking the diff between master and A and commiting it within B - so i just drop every single commit inbetween ...	15:59
hyperair	hmmm that would work.	15:59
	← Cromulent left	15:59
	joshthecoder_afk → joshthecoder	15:59
	→ bobrik joined	15:59
hyperair	i'd use filter-branch with a --msg-filter and a --commit-filter	16:00
	← Linkadmin left	16:03
fabiand-prop	hyperair: hm ok. i might look into that aswell :)	16:03
hyperair	=)	16:04
fabiand-prop	hyperair: thanks for now :)	16:04
hyperair	you're welcome	16:04
	← corni left	16:05
	→ ludde joined	16:07
	→ ROBOd joined	16:07
ROBOd	hello guys	16:07
	i am trying to push my changes to a remote git repo and i get rejected	16:07
	← ben_h left	16:07
ROBOd	! [rejected] mdl19-tinymce3 -> mdl19-tinymce3 (non-fast forward)	16:08
wereHamster	ROBOd: faq non-ff	16:08
Gitbot	ROBOd: Your push would lose changes on the remote. See http://git.or.cz/gitwiki/GitFaq#non-ff	16:08
ROBOd	wereHamster: thanks	16:08
	i did a rebase from a different branch	16:08
	→ loincloth joined	16:09
	← neunon left	16:09
	→ jelmer____ joined	16:10
	→ bugfux joined	16:12
	← priidu left	16:13
	← fabiand-prop left	16:16
	← moonflux left	16:16
	→ moonflux joined	16:16
	→ chestnut joined	16:17
	← ludde2 left	16:18
	→ snearch_ joined	16:19
wurble	Is there a way I can do git status only on the current directory?	16:20
Ilari	wurble: There's no way to limit status to some directory. Maybe closest would be diffing with --name-only (and ls-files to get untracked files).	16:22
ROBOd	i added a remote repo and its own branches don't show up in git branch -r. what can i do?	16:24
	git remote show .... does show me the remote branches	16:24
	but still, i cannot do git checkout -b whatever remote-repo/wanted-branch	16:24
	→ volodya joined	16:25
Ilari	ROBOd: 'git fetch remote-repo'.	16:25
aeruder	ROBOd: so you've just done the git remote add command?	16:25
	← bobrik left	16:25
ROBOd	aeruder: yes	16:25
aeruder	ROBOd: so now run git fetch <remote>	16:25
Ilari	ROBOd: Next time, use -f to remote add and it will run that fetch for you.	16:25
ROBOd	worked other times...	16:25
	aha, thanks	16:25
aeruder	git remote add (without -f) just basically edits your .git/config	16:26
	→ Cromulent joined	16:28
	→ ericindc joined	16:30
	← scarabx left	16:32
	→ dreiss joined	16:33
	← stepnem left	16:36
	→ intripoon joined	16:37
	← bugfux left	16:38
	← kevin_ left	16:41
	← moonflux left	16:42
	→ bugfux joined	16:42
	← mika left	16:42
	→ jceb joined	16:42
	→ tvw joined	16:43
	← volodya left	16:44
	← kumbayo_ left	16:47
	→ kumbayo__ joined	16:47
	← Cromulent left	16:50
	→ fridim_ joined	16:51
	→ ryandotsmith joined	16:51
	→ arcticpenguin joined	16:52
ryandotsmith	hey! when i do a git diff production on my master branch i see heaps of data	16:52
	→ cpg joined	16:53
ryandotsmith	but when i do a git merge production on master, it says everything is up to date . . . .	16:53
	any ideas on what is going on ?	16:53
	← ciskje left	16:53
Ilari	ryandotsmith: Master is ahead of production branch.	16:54
	fr\| → fr	16:54
ryandotsmith	ok, what do i need to do to get production merged into master ?	16:55
	→ Cromulent joined	16:55
arcticpenguin	Hello, I'm wondering if anyone can point me to a gitweb resource.... I've followed the instructions from a few sites out there and I have the page rendering via Apache virtual hosts, but apache throws a Permission Denied error when trying to get into my /srv/git/repository/project.git - it says 'itweb.cgi: Can't cd to /srv/git/repositories/sofia.git: Permission denied'	16:55
	← Cromulent left	16:58
biz	arcticpenguin: If you don't run the gitweb.cgi through suexec, you need to allow the Apache user to read the repository	17:02
arcticpenguin	thanks.. i just did that and it worked	17:03
	← arcticpenguin left	17:04
Ilari	ryandotsmith: It is already merged.	17:06
	ryandotsmith: (well, if you got the branches in that command right).	17:06
	→ _grepper_ joined	17:06
ryandotsmith	actually, it is not. There are still changes in production that i want in master	17:06
	← snearch_ left	17:06
Ilari	ryandotsmith: 'git log master..production' lists all commits that are in production but not in master.	17:07
	← eMyller left	17:07
	→ EruditeHermit joined	17:08
	→ flazz joined	17:08
EruditeHermit	hi, when I try to do a git merge, I get does not point to a commit	17:08
	I am doing git merge <repo url>	17:08
Ilari	EruditeHermit: Merge does NOT take URLs.	17:08
EruditeHermit	oh	17:08
	→ eMyller joined	17:09
	→ tongueroo joined	17:09
	→ cemerick joined	17:11
Ilari	EruditeHermit: To merge commits from other repositories, you need to fetch them first. Pull does this before merging or rebasing the commits.	17:12
	EruditeHermit: Additonally, one needs to specify what branch to merge, as there can be multiple.	17:13
	← ryandotsmith left	17:13
	← kanzure left	17:15
	→ kanzure joined	17:15
	← ryanakca left	17:15
	→ ryanakca joined	17:16
	← chestnut left	17:16
	→ gurami joined	17:17
EruditeHermit	Ilari: yeah I just realized that =p	17:18
	thanks	17:18
gurami	hello. I've setup gitweb and now I'm trying to setup passwordless logins to myserver, but I get public key permission denied message :[	17:19
	can anyone tell me what permissions to setup on $home and ./ssh? please...	17:20
jast	.ssh needs to be 700 and all files in it 600	17:20
orb	They don't strictly need to be. Only id_dsa pub shall be 0600.	17:20
gurami	oh thanks! I've been looking for it about 1 hour.	17:20
	I'll try that thanks guys	17:21
	;-)	17:21
	← dorileo left	17:21
	→ _Vi joined	17:21
jast	orb: actually I think sshd checks most other files too if StrictModes is enabled (which it is almost everywhere)	17:22
gurami	no strict mode is disabled in my case	17:22
	← name left	17:23
	→ name joined	17:23
jast	in that case it doesn't really matter... at least sshd won't care	17:25
	← eMyller left	17:25
gurami	it does not login :\|	17:25
	I'll look in ssh -vvv now	17:26
jast	I was about to suggest that :)	17:26
	← solydzajs left	17:26
	← psynaptic left	17:26
gurami	:)	17:26
	← bugfux left	17:27
	→ bugfux joined	17:27
	→ eMyller joined	17:28
	→ kinkerl joined	17:33
	→ Linkadmin joined	17:34
	← eMyller left	17:35
	← Yewbacca left	17:36
	→ jmspeex joined	17:36
gurami	here please look http://paste.uni.cc/20115	17:38
	← fridim_ left	17:38
	→ eMyller joined	17:38
	→ volodya joined	17:39
	← sprasanna_ left	17:40
gurami	I can't decipher it. it's really annoying I've all permissions right, I've deleted know hosts file, generated rsa keys from scratch	17:41
	:[	17:41
tongueroo	hey guys, is it possible to trace back with git log and see filter for merges that happened, even if the committer alter the typical "merge 'branchname' into..." message	17:41
	i guess i would look for commits with more than one parent ?	17:41
	still confuses me :P	17:41
	→ priidu joined	17:41
hyperair	git log --graph	17:42
	← thiago_home left	17:42
	← _Vi left	17:42
	→ _Vi joined	17:42
	→ xelados joined	17:43
xelados	how do I change to a previous commit/version?	17:44
	It's been a long time since I've touched git.	17:44
	→ thiago_home joined	17:46
	← tvw left	17:46
hyperair	commit --amend	17:47
	← eMyller left	17:47
hyperair	or rebase -i	17:47
xelados	I just did a checkout on the hash and it seemed to work.	17:48
	→ eMyller joined	17:48
agib	If I'm working on a topic branch with a bunch of new directories and files, but want to go back to my master and do some things there, how do I clean my working tree again?	17:49
EruditeHermit	Ilari: how do you cancel local cached changes?	17:50
agib	is it get reset --hard HEAD?	17:51
	eMBee → sm	17:51
	sm → eMBee	17:51
	→ arcticpenguin joined	17:52
	→ elmob joined	17:53
	→ ramlev joined	17:53
agib	anyone?	17:53
orb	anyone can give me a pizza.	17:53
eb_	agib: that won't delete the new files and dirs	17:53
agib	eb_: ah... do you know what will?	17:53
EruditeHermit	agib: I want to do exactly the same thing =p	17:53
eb_	yes, sec	17:53
	i think git clean	17:53
	plus some options tho	17:53
ramlev	orb: if you can fetch it yourself in denmark, then yes	17:53
orb	ramlev: can I resort to a git fetch? :)	17:54
agib	hmm	17:54
eb_	git clean -d -f	17:54
kahmalo	xelados: git checkout with a commit ID detaches the HEAD. If you now edit files and use git commit, the resulting commit won't be on any branch.	17:54
	← Theravadan left	17:54
ramlev	orb: not implemented in the danish pizza houses	17:54
orb	cramps..	17:54
EruditeHermit	I keep getting you have local changes to file cannot switch branches	17:55
	even after the git clean -d -f	17:55
orb	-dfx perhaps.	17:55
eb_	clean only removes untracked files	17:55
	so, you additionally should do a 'git reset --hard HEAD'	17:55
xelados	kahmalo: Oh, that's no good. What's the proper way to reverse current changes and revert to my last commit?	17:55
wereHamster	EruditeHermit: git stash	17:55
kahmalo	xelados: It depends on whether you want the branch to remember that the changes had been made and reverted.	17:56
xelados	I don't want it to remember that.	17:56
EruditeHermit	wereHamster: stash is cool =)	17:57
	eb_: thanks that worked	17:57
agib	eb_: thank you	17:57
eb_	np	17:57
kahmalo	xelados: Then check out the branch and use git reset --hard with the commit id. This discards all local changes in files; if you don't want that, use something other than --hard.	17:57
xelados	"HEAD is now at [hash]"	17:58
	Good?	17:58
	← ramlev left	17:58
kahmalo	xelados: I think git reset always displays that. git branch, or git status, should show whether HEAD is still detached or now on a branch.	17:59
xelados	no branch...	18:00
kahmalo	You didn't check out the branch then, before resetting?	18:00
xelados	nope..	18:00
	Okay, there.	18:00
orb	.oO( git config alias.guillotine checkout HEAD )	18:01
xelados	It works now. Thanks.	18:01
	← agib left	18:01
	→ saurabhverma joined	18:01
	→ davegb3 joined	18:03
	→ markelikalderon joined	18:04
	← qrush left	18:05
	→ pisi joined	18:08
wereHamster	orb: why guillotine? checkout HEAD is a noop	18:08
	you probably mean checkout HEAD^0	18:09
pisi	How to change the gpg binary that git uses to sign tags? I have several and as normal command line signing works but git fails, I assume it is using the wrong version	18:09
orb	wereHamster: inconceivable git behavior! :)	18:09
	yeah.	18:09
pisi	(I use a smart card for signing and it has worked at least once successfully before, but I can't remember what I did differently last time)	18:10
	→ marc joined	18:10
	marc → marc_kdab	18:10
	← ericindc left	18:11
wereHamster	pisi: the binary doesn't seem to be configurable, but try adjusting PATH so that your prefered binary is picked first	18:11
kahmalo	pisi: it seems to just search for gpg in PATH. relevant functions: do_sign in builtin-tag.c, start_command in run-command.c	18:11
	→ bobesponja joined	18:11
	← _Vi left	18:12
	→ fridim_ joined	18:13
	→ _Vi joined	18:13
	→ hummingkid joined	18:13
pisi	I have it in my path, doesn't work. have to debug ... the bridge to smart cards is dumb and complicated in gpg, unfortunately...	18:15
	→ agib joined	18:16
	← flazz left	18:18
wereHamster	pisi: try setting GIT_TRACE=true to see what git is doing	18:18
	→ qrush joined	18:18
	← tomaw left	18:18
	← bugfux left	18:19
	→ bugfux joined	18:19
	→ tomaw joined	18:20
	→ mokasin joined	18:21
	→ patrikf joined	18:22
mokasin	Hi. I've a beginner's question. How do you resolve best a file rename conflict. Szenario Alice renames a file to foo and changes some lines, Bob changes the name into bar and also changes some lines. Merge: Get shows now a conflikt and both files are in the folder... how to go on?	18:23
Ilari	mokasin: Edit the files to state you want and add/rm them.	18:24
tokkee	mokasin: ... and rethink your file naming scheme scnr ;-)	18:25
mokasin	okay, lets say I decide to call the file foo. but how do I include easily the changes from bar into foo?	18:25
rudi_s	mokasin: diff and patch; or your editor if it has diff support.	18:27
mokasin	@tokkee, and rethink the developers names ;). okay, I guessed so. so theres no integrated command: take this filename with all changes of both in it? that's a pity	18:29
	← ijuma left	18:29
	→ ijuma__ joined	18:29
wereHamster	mokasin: or take a look at man git-merge-file	18:29
Gitbot	mokasin: the git-merge-file manpage can be found at http://git.or.cz/man/git-merge-file	18:29
mokasin	hey, that's what I was looking for	18:30
	→ giallu joined	18:31
mokasin	perfect. thx	18:31
	← reprore left	18:31
tokkee	mokasin: I did not refer to using 'foo' and 'bar' but rather to two developers renaming the same file to different names at the same time - that sounds rather broken ;-)	18:32
mokasin	yeah, I suppose ;-). but hey, for what I have a CVS that have to handle the mess I create ;)	18:34
	→ stepnem joined	18:37
	← Base2 left	18:38
	← eMyller left	18:38
	← bugfux left	18:39
	→ eMyller joined	18:40
	← dodo_the_last left	18:43
	→ tvw joined	18:44
	→ bcardarella joined	18:47
	→ joce_ joined	18:50
	→ ajonat joined	18:53
	→ eean joined	18:55
	→ trefn joined	18:56
trefn	is there a way to only show commit by 1 author with git log?	18:56
	→ chris2 joined	18:56
trefn	i want to see my first commit on a group project	18:56
	→ aarongibralter joined	18:56
kahmalo	git log --author=<pattern>, says the manpage	18:56
	← agib left	18:57
trefn	thanks, didn't see it.	18:57
	← aarongibralter left	18:58
	→ agib joined	18:58
	→ jjuran joined	19:01
almostautomated	anyone know of a windowed diff util that allows setting regex ignored like CLI diff? (*nix) ?	19:03
joce_	almostautomated: Have you looked at SourceGear's DiffMerge?	19:05
thiago_home	almostautomated: diff with regex?	19:06
	almostautomated: what are you talking about?	19:07
	← Akufen left	19:07
almostautomated	thiago_home: ie: diff -I 'regex'	19:07
	like that but visual instead of console	19:07
thiago_home	why do you want that?	19:07
	thiago_home has never used that option	19:07
thiago_home	I'm curious	19:07
almostautomated	thiago_home: I use it to ignore lines that I _know_ will be different	19:08
thiago_home	right, but why do you want to ignore lines you did change?	19:08
almostautomated	ie:: comparing two different patches that _should_ make the exact same changes but come from different bases	19:08
thiago_home	ah, the context lines?	19:09
almostautomated	no... the index lines	19:09
wurble	Why do I get "untracked working tree file foo.c would be overwritten by merge" when I try to checkout a branch bar, when bar contains foo.c? foo.c is unmodified.	19:09
thiago_home	wurble: it's untracked, not unmodified	19:10
doener	almostautomated: maybe you want interdiff instead of diff?	19:10
almostautomated	thiago_home: when you do git diff -p	19:10
	interdiff?	19:10
	almostautomated never heard of that one	19:10
doener	almostautomated: http://man.he.net/man1/interdiff	19:10
almostautomated	LMFAO::: http://people.redhat.com/twaugh/interdiff/	19:10
wurble	thiago_home: untracked by which branch?	19:10
almostautomated	doener: thanks... will read.	19:11
thiago_home	wurble: the currently checked out one	19:11
wurble	thiago_home: I believe I see it in the log --name-only messages for branch bar.	19:11
thiago_home	wurble: right, the file is in that branch	19:11
	→ corni joined	19:11
thiago_home	wurble: the problem is that it is not in the current branch	19:11
	← tvw left	19:12
	← bcardarella left	19:12
	→ Short_some_thing joined	19:12
	← volodya left	19:12
	→ resmo joined	19:12
	→ tvw joined	19:12
wurble	thiago_home: Thanks.	19:13
	← juanjoc left	19:13
	← eMyller left	19:14
wurble	How would I diff my current file with the one in the branch?	19:15
thiago_home	is it tracked already?	19:15
	i.e., git add ?	19:16
	and committed	19:16
patrikf	wurble, git diff branch:file file	19:16
wurble	thiago_home: No. Not in this branch.	19:16
thiago_home	patrikf: that works? Cool.	19:17
	thiago_home didn't know	19:17
wurble	patrikf: I get ambiguous argument 'foo.c': unknown revision or path not in working tree	19:17
	← asksol left	19:17
patrikf	wurble: well, it works for me	19:18
wereHamster	wurble: try git diff <branch> -- <file>	19:19
aeruder	wurble: i think that error suggests it can't find foo.c	19:19
	is that really what your file is called?	19:19
	→ Akufen joined	19:19
aeruder	and keep in mind, the branch: syntax needs to have the relative path to the file	19:19
	← lu_zero left	19:19
aeruder	i.e. if its at /a/foo.c, you need branch:a/foo.c	19:19
orb	... except when there is a : in it	19:19
aeruder	(relative to the root of the git repo)	19:19
wurble	wereHamster: That works but it just shows the file as deleted (it's not)	19:21
	← hyperair left	19:21
	→ ryandotsmith joined	19:21
wereHamster	wurble: are you sure your current HEAD contains that file?	19:22
wurble	aeruder: Yes I'm sure	19:22
almostautomated	doener: thanks, but that doesn't look like what I'm looking for. Prob is: 'diff -w A B == same' 'diff -w A^ B^ == same' diff -w of ( diff -w A^ B and diff -w B^ A ) != same when I do it in the script, but does when I do it manually.	19:23
wurble	wereHamster: My current HEAD doesn't seem to contain the file. That is my problem. I want to know if what I have in my working directory is different that in the branch I want to to switch to.	19:23
	→ k_shaurya joined	19:24
	← k_shaurya left	19:24
almostautomated	Trying to isolate the difference between my scripts diff output and manual output, yet am failing to see the context even with diff -w -y	19:24
wereHamster	wurble: why is it untracked in your current HEAD but included in the other branch?	19:24
wurble	The directory the file is in is the same as the name of my branch which may be compounding problems.	19:24
	wereHamster: Not sure. Isn't that what would happen if I checkout -b bar; add foo.c; checkout master?	19:25
	→ dwmw2_gone joined	19:25
	→ aliceinwire joined	19:25
	← ryandotsmith left	19:26
wereHamster	wurble: assuming you commit after adding the file, yes that would lead to such situation	19:26
	← mithro_ left	19:26
wereHamster	er, wait, it wouldn't. If one file is included in one branch but untracked in another one, git removes it when switching branches	19:27
wurble	wereHamster: So assuming I retained the file in my working directory by some other means, how would I verify that it is the same as in the other branch?	19:27
	i.e. That I didn't update it.	19:27
	→ galderz joined	19:28
	← markelikalderon left	19:28
	← galderz left	19:28
wereHamster	wurble: temporary add the file to the index and run git diff	19:28
	← _grepper_ left	19:28
wurble	wereHamster: Against the other branch?	19:29
wereHamster	yes	19:29
	git diff <otherbranch> -- <yourfile>	19:29
	← gurami left	19:30
wurble	wereHamster: Thanks, that did it.	19:30
	Is there an easy way to merge the changes?	19:30
mems	Does anyone have any solutions for migration svn:externals to submodules? I'm migrating a subversion repo with various svn:externals cross-references using git-svn, and would like to use submodules as a replacement.	19:31
wereHamster	wurble: where would you like to have the resulting commit? On your current branch or on the branch which already contains that file?	19:31
	← saurabhverma left	19:32
wurble	wereHamster: branch which already contains it. I actually just want the current branch copy, uncommited, on the branch that already contains it. I guess I could just copy it and then checkout the branch and overwrite.	19:32
	→ _grepper_ joined	19:32
	→ stringo0 joined	19:34
	→ Grahack joined	19:34
	← Grahack left	19:35
wereHamster	wurble: git add the file, git checkout -m <otherbranch>; resolve the conflicts (if any); git commit	19:35
wurble	wereHamster: Thanks	19:35
	→ justin-george joined	19:35
	← patrikf left	19:36
	→ Base2 joined	19:37
	← Base2 left	19:37
EruditeHermit	if I do git checkout experimental	19:39
	→ ryandotsmith joined	19:39
EruditeHermit	and then do git merge master	19:39
	→ Base2 joined	19:39
EruditeHermit	it says Already up to date even though I know master has new changes	19:39
wereHamster	EruditeHermit: does 'git log experimental..master' list any commits?	19:40
EruditeHermit	no	19:41
	← Base2 left	19:41
wereHamster	there you go, master doesn't contain any commits that aren't already in experimental	19:41
wurble	What's the best way to commit a single file on a different branch when I have other uncommited changes (that I do not want to commit yet) on the current branch?	19:41
	→ mithro_ joined	19:42
EruditeHermit	wereHamster: but there are changes to a file in master that aren't in the other branch	19:42
wereHamster	EruditeHermit: did you commit those changes on master?	19:42
EruditeHermit	yes	19:42
	if I do git commit -a on master it says nothing to commit	19:43
	← ryandotsmith left	19:43
	→ Pewpewarrows joined	19:44
wereHamster	wurble: I'd use git stash	19:44
	EruditeHermit: does git diff master experimental -- <file> show the differences you are missing?	19:44
	marvil07 → marvil07-lunch	19:45
wurble	wereHamster: And then apply on the other branch?	19:45
EruditeHermit	wereHamster: yes	19:45
wereHamster	wurble: yes	19:45
	EruditeHermit: maybe the change got lost in a merge conflict	19:46
EruditeHermit	this is a pretty new repo	19:46
	haven't had any merge conflicts	19:46
	→ spaceonline joined	19:46
	← cemerick left	19:47
	← kumbayo__ left	19:48
wereHamster	EruditeHermit: do you know in which commit these lost changes were introduced to master?	19:48
EruditeHermit	nope	19:48
	→ MrN joined	19:48
	→ p0w3r3d joined	19:49
	→ eMyller joined	19:52
wurble	wereHamster: Okay, after stash applying on the other branch I have unmerged conflicts on other files (which I don't want to commit, I will restore them when I return to my original branch). checkout on those files doesn't work, complaining that they are "unmerged"	19:52
	→ beni- joined	19:53
wurble	Never mind. I solved it by adding it and then resetting it.	19:53
	Thanks again	19:53
	← agib left	19:53
wereHamster	wurble: after git stash apply, you should be able to git add the file you want to commit and then run git commit, ignoring all the other files that have conflicts	19:54
wurble	wereHamster: It didn't let me do that.	19:55
beni-	hi guys. i have a big file (video project) in an old revision, but I won't need that file not anymore. is there a way to remove the file out of that old revision?	19:55
wurble	The file was present on my branch. Maybe that is an extra problem.	19:55
	wereHamster: The whole process was more painful than I would have liked. When I returned to my original branch I stash applied and then tried to merge but then had to again resolve conflicts with the file that I updated.	19:56
wereHamster	EruditeHermit: that is weird. Maybe run git blame on the file in master to find out which commit introduced the changes, and then check if the same commit is in the experimental branch	19:57
wurble	This use-case happens often to me. I wish there was a better way to handle it.	19:57
EruditeHermit	wereHamster: ok thanks	19:57
wereHamster	beni-: not without rewriting history	19:57
beni-	wereHamster: what does that mean?	19:58
	← jceb left	19:58
wereHamster	beni-: start reading http://www.kernel.org/pub/software/scm/git/docs/user-manual.html#fixing-a-mistake-by-rewriting-history	19:59
	→ jceb joined	19:59
beni-	wereHamster: ok thanks, i'll have a read through	19:59
	→ Marmouset joined	19:59
wereHamster	wurble: that is to be expected. If merging A into B results in conflicts, then also merging B into A will result in conflicts	20:00
	← p0w3r3d left	20:01
	← arcticpenguin left	20:02
	klbate_ → klbate	20:02
	← elmob left	20:03
	→ hammu joined	20:04
	→ scarabx joined	20:04
	← hammu left	20:04
	ijuma__ → ijuma	20:07
	← eean left	20:08
	← mokasin left	20:08
	← robinr left	20:09
	← ROBOd left	20:09
	← jjuran left	20:11
	← marvil07-lunch left	20:13
	→ marvil07-lunch joined	20:13
	marvil07-lunch → marvil07	20:14
	← _grepper_ left	20:18
	→ tlm joined	20:18
tlm	i'm doing an interactive rebase; i marked one old commit with `edit', split it into several commits, and ran `git rebase --continue'. now i'm getting merge conflicts with changes that were not part of that commit. have i done something wrong?	20:21
	→ dodo_the_last joined	20:22
	MrN → emy_	20:22
	emy_ → MrN	20:22
plopix	why is ' +$' evil in source code?	20:23
	git complains about it. is there a way to let git accept it?	20:24
	← jceb left	20:24
kahmalo	man git-config: core.whitespace, apply.whitespace, color.diff.whitespace	20:25
	← jelmer____ left	20:25
Gitbot	kahmalo: the git-config manpage can be found at http://git.or.cz/man/git-config	20:25
parasti	disable whatever nazi hook you have enabled	20:25
	→ robinr joined	20:26
	→ binrapt joined	20:26
binrapt	Urgh I just misspelled a word in a commit and already pushed it onto the main server - what can I do to fix the text?	20:27
	It's the log message I mean	20:27
	Not the content of the code	20:27
	I really want to fix it, one can hardly understand what it means now	20:28
kahmalo	amend and force push. then talk with everyone who pulled the original	20:28
	it doesn't matter whether the change is in the commit message or files	20:29
parasti	or undo it with "git revert", then "git cherry-pick -n" the original commit and recommit with proper message	20:29
	cpg → cpg\|biab	20:29
kahmalo	a commit note would be another option, but it seems uncertain whether that feature will ever be in a released git	20:30
	→ ryandotsmith joined	20:30
	→ Theravadan joined	20:31
binrapt	kahmalo: How does a force push work?	20:33
	Normal push gives me errors :(	20:33
	← dreiss left	20:33
binrapt	I amended it successfully and my local history is alright now	20:34
kahmalo	git push -f or with a + in front of the refspec. disables the fast-forward check. but the server may have a hook that requires fast-forward anyway.	20:34
	← David-T left	20:37
binrapt	It's repo.or.cz	20:39
	→ David-T joined	20:39
binrapt	Worked! Thanks	20:39
	← aliceinwire left	20:43
	← mbroeker left	20:45
	← Alinermedia left	20:46
	← Weaselweb left	20:47
	→ bugfux joined	20:49
	← k0001 left	20:49
	← khmarbaise_ left	20:50
	← bugfux left	20:51
	← beni- left	20:51
	→ mika joined	20:52
	→ k0001 joined	20:53
	← Theravadan left	20:54
priidu	is it possible "git reset --hard HEAD" for remote git repo after pushed up (currently I made small script what ssh to remote server and exec "git reset --hard HEAD"	20:55
orb	use push -f	20:55
priidu	orb, thank, so simple :-)	20:55
orb	then again I wonder why you would want reset in a bare	20:56
	← robinr left	20:57
priidu	i'm updating web server data in git, after updated locally i push up to remote git repo (kind of abusing git as rsync etc)	20:58
	← ludde left	20:58
	→ ^Einstein joined	21:01
orb	that's for sure...	21:02
	pushing into a non-bare still requires reset	21:02
wereHamster	priidu: take a look at http://toroid.org/ams/git-website-howto	21:03
doener	orb: hm? Why should push require reset?	21:03
orb	doener: because the wc is still at the old HEAD	21:03
	doener: pushing only updates the head, not the wc or index.	21:04
priidu	orb, yep just tested, i still had to "git reset --hard HEAD" at remote server	21:04
	→ robinr joined	21:04
doener	orb: ah, pushing to a non-bare repo and to the ref referenced by HEAD	21:04
	orb: didn't even think of that case, as it's so, uhm, non-useful ;-)	21:05
orb	doener: don't tell me	21:05
priidu	wereHamster, thanks for the link, will try it out	21:06
	← Soliton left	21:07
	→ iaj joined	21:08
iaj	hello there :)	21:08
	← resmo left	21:08
	← jonmasters left	21:08
Ilari	Hmmh... The shim transport layer doesn't seem to consider extensibility at all... :-(	21:09
	→ lu_zero joined	21:09
iaj	got a little question.. after I git pulled - it shows me how many insertions..deletions there are.. but how can I actually view the changes that have been made?	21:09
orb	shim, what, where? In git? Is not that a kernel thing?	21:09
uau	i'm trying to work around around git's sucky submodule functionality and support shallow clones of submodules	21:09
	← __iron left	21:10
	→ jonmasters joined	21:10
	→ dreiss joined	21:10
Ilari	orb: In git.	21:10
uau	should "git clone --depth 1 `git config --get submodule.NAME.url` DIRNAME" do about the same as "git submodule update DIRNAME", assuming the repo doesn't exist yet?	21:10
	→ __iron joined	21:10
uau	and assuming you have run git submodule init DIRNAME	21:10
orb	ilari: http://www.shim6.org/ :)	21:11
	→ rdeshpande joined	21:11
rdeshpande	hi all	21:11
Ilari	uau: Those diffstats are from merge into current branch.	21:11
	← mems left	21:11
	← wurble left	21:12
uau	Ilari: did you mean to address that to someone else?	21:12
rdeshpande	can someone explain the difference between committer and author? i was looking at a commit in teh source tree clearly committed by person A, but with git cat-file I saw the author as person A but committer as person B.	21:12
	→ Soliton joined	21:12
Ilari	uau: OOps.	21:13
	iaj: Those diffstats are from merge into current branch.	21:13
	← jmspeex left	21:13
doener	rdeshpande: If I send a patch to e.g. Junio, and he creates a commit which ends up in the git.git repo, then I'm the author, but Junio is the committer	21:13
	rdeshpande: ... he creates a commit from my patch	21:14
	← bitprophet left	21:14
iaj	Ilari, huh??	21:14
rdeshpande	doener: does the same apply for, say a merge between branches? woudl the person that perform the merge be the committer but the initial writer of the code be the author?	21:14
doener	rdeshpande: no, a merge creates a new commit, and that would (most likely) have the author being the same as the committer	21:15
	gernot_ → gernot	21:15
doener	rdeshpande: a merge commit has two or more parents, so the original commits are reachable through it. And those have the respective authors set	21:16
rdeshpande	hrm	21:16
	im trying to figure out how this came about in our source tree	21:16
doener	rdeshpande: rebasing changes the committer, but keeps the author	21:17
	← MrN left	21:17
doener	rdeshpande: because rebasing creates new commits, but reuses the changes from the original commits	21:18
rdeshpande	ah	21:18
	← Soliton left	21:18
	→ jmspeex joined	21:18
	→ Soliton joined	21:19
	→ Soliton_ joined	21:19
	← Soliton_ left	21:19
rdeshpande	doener: thanks	21:19
	← lack left	21:19
	→ tbr joined	21:24
	← corni left	21:25
	← tbr left	21:27
	→ jmspeex_ joined	21:28
	→ juanjoc joined	21:28
	→ s0ber_ joined	21:28
	← bambule left	21:29
	← juanjoc left	21:30
	← orb left	21:32
	← jmspeex left	21:33
	← scarabx left	21:34
	→ markelikalderon joined	21:34
	← rdeshpande left	21:34
	← eMyller left	21:35
	← robinr left	21:37
	→ Gitzilla joined	21:37
	jmspeex_ → jmspeex	21:38
	→ eMyller joined	21:38
almostautomated	why would a git diff -w -p piped to apply produce a 'garbage' patch vs a git diff -p work fine ? What determines 'garbage'?	21:41
	← s0ber left	21:41
Ilari	almostautomated: What was -w again?	21:42
	almostautomated: Ignore space change+	21:42
	s/\+/?/	21:42
doener	Ilari: yep, all whitespace changes	21:42
	→ Base2 joined	21:42
	← stuffcorpse left	21:42
	→ stuffcorpse joined	21:43
almostautomated	except one, I believe and that is EOL @ EOF... That is something I've been reading on the mail list about	21:43
	→ EricInBNE joined	21:43
doener	Ilari: or rather: all whitespace. Opposed to -b which does something like s/\s+/ /g	21:43
	→ pigeon joined	21:43
	← trefn left	21:44
	← binrapt left	21:44
doener	i.e. with -b "abc def" is treated the same as "abc def", but different from "abcdef"	21:44
	with -w, all three are considered to be the same	21:44
	cpg\|biab → cpg	21:45
almostautomated	I wonder if that would make apply think the diff produced patch is garbage	21:45
	← justin-george left	21:46
osfameron	evening	21:46
almostautomated	It essentially would be... in regards to patching something	21:46
Ilari	Idea about implementing extensiblity to external transports: add command that returns capabilities. And fetch and list would need to be capabilities as well, because some transports might not be able to easily implement them.	21:46
osfameron	I've just done 'git rebase' to pull from a topic branch, and git status tells me "Your branch and 'origin/release_2-0' have diverged, # and have 37 and 2 different commit(s) each, respectively." How do I check that everything is OK and ready to git push?	21:46
Ilari	E.g. smart transports.	21:47
	osfameron: That doesn't look OK.	21:47
osfameron	Ilari: oh :-(	21:47
doener	osfameron: rebase to pull from a topic branch?	21:48
Ilari	osfameron: At least not for pushing to release_2-0.	21:48
doener	osfameron: that sounds like you did: git checkout release_2-0; git rebase <topic>	21:48
osfameron	doener: yes	21:48
doener	osfameron: which is pretty much the wrong way around	21:48
osfameron	yay!	21:49
	I should 'git merge' instead?	21:49
doener	osfameron: that has put the commits that are in release_2-0 which aren't in <topic> on top of <topic> and called the result release_2-0	21:49
osfameron	or git rebase from within the topic branch?	21:49
doener	osfameron: first, you want to restore the old state of release_2-0	21:49
osfameron	ok	21:49
	how do I do that?	21:49
	delete the branch and create a new tracking branch?	21:50
doener	osfameron: check its reflog (git reflog show release_2-0), find the pre-rebase entry (should be @{1}) and reset to there	21:50
osfameron	git reset release_2-0@{1} ?	21:51
doener	osfameron: yup	21:51
	osfameron: probably you want --hard, though	21:51
	osfameron: that kills uncommitted changes, so be careful...	21:51
	osfameron: then, you can choose between simply merging the topic branch, or rebasing the topic branch first, and then merging it	21:52
osfameron	seemed ok without --hard (as in, git log looks vaguely sane)	21:52
doener	osfameron: yeah, but "git status" won't	21:52
osfameron	oh I see, lots of "modified" files	21:52
doener	osfameron: --hard (vs. the default --mixed) is just about the working tree being reset, too	21:52
osfameron	ok, well I don't have any uncommitted changes so I'm OK	21:53
doener	osfameron: you can just do "git reset --hard" now, to drop all those uncommitted changes	21:53
osfameron	untracked files don't get reset do they?	21:53
	← Base2 left	21:53
doener	osfameron: no, they don't. "git clean" can drop them	21:53
	← stuffcorpse left	21:53
osfameron	no that's cool, I'm happy to keep them for now (notes and logfiles and such)	21:53
	→ stuffcorpse joined	21:53
osfameron	unless... can you git clean a directory?	21:54
doener	osfameron: anyway, the plain "merge" may obviously give you a non-linear history (which isn't any problem per se). If you insist on having a linear history, use the rebase+merge approach	21:54
	osfameron: you can tell clean to operate only on a certain directory (and its subdirectories), yes	21:54
osfameron	I think I'd settle for stuff working for now :-) so I may go for just merge	21:54
doener	osfameron: don't use the rebase approach if the topic branch is public though	21:55
osfameron	ah no, there was only one other developer on this repo, and he didn't use that branch	21:56
	→ jelmer____ joined	21:56
	→ trefn joined	21:57
osfameron	ok I'm getting: error: Entry 'cgi-bin/admin.cgi' would be overwritten by merge. Cannot merge.	21:58
	that doesn't sound good...	21:59
	→ sergiodj joined	21:59
osfameron	that file existed since before the topic branch, dunno why it'd think they were two separate files	21:59
doener	osfameron: did you do the "git reset --hard" thing? Otherwise, you of course still have those uncommitted changes, and git complains about them	21:59
osfameron	doener: yeah I did	21:59
	git status became cleaner (modulo the files untracked in this branch)	22:00
doener	osfameron: ah, you did a --mixed reset, the file might be left as untracked	22:00
osfameron	I did 'git reset --hard' as you'd suggested	22:00
doener	osfameron: that doesn't drop untracked files (as you correctly said)	22:01
	→ urbanmonk joined	22:01
doener	osfameron: the --mixed reset doesn't touch the working tree, but the index. So files can become untracked	22:02
	← Ilari left	22:02
	→ Ilari joined	22:02
	osfameron considers renaming his repo/ and checking out a clean version	22:02
	→ robinr joined	22:02
	← offby1 left	22:03
	→ Cromulent joined	22:04
urbanmonk	is it possible to have setup gitweb to require authorization on some projects but not others?	22:04
	→ _grepper_ joined	22:05
almostautomated	doener: did more reading/studying while trying the 'format-patch'->'apply' version of whitespace fixing. There were some discussions about whitespace at EOF, and indeed, that was the cause of my earlier diff -w A^ B != diff -w B^ A issue (when diff -b A = B and diff -b A^ = B^). I think there is a work-around that'll work for it though...	22:06
Ilari	osfameron: Reclone usually solves nothing...	22:06
	← trefn left	22:07
	cpg → cpg\|biab	22:07
	← sunoano left	22:08
osfameron	Ilari: I thought I might have broken stuff... I certainly don't underestand what I'm doing now	22:08
	→ trefn joined	22:08
Ilari	osfameron: Make a backup?	22:08
	→ pjackson joined	22:09
	← stuffcorpse left	22:09
pjackson	Can I get format-patch to do its thing without mail headers?	22:09
	→ stuffcorpse joined	22:10
Ilari	pjackson: No. But you don't have to mail the stuff.	22:10
	→ sunoano joined	22:10
	← fridim_ left	22:11
	← Sigma left	22:12
osfameron	Ilari: ok, it's all checked in (and pushed :-)	22:12
	Ingmar → hacchi	22:12
osfameron	but I'm very confused	22:12
	hacchi → register	22:13
	← schambers left	22:13
	register → hacchi	22:13
osfameron	I'm in release branch; git merge <topic>; fix 1 conflicting file; git add <conflicting file>; git merge # again; and now I have this error about cgi/admin.cgi would be overwritten	22:13
doener	osfameron: merge again? You have to commit the first one first	22:14
osfameron	doener: I thought to resolve conflict you just had to 'git add' ?	22:14
	→ offby1 joined	22:14
doener	osfameron: that updates the index, yes. But it doesn't commit	22:15
	osfameron: just imagine that you have two conflicting files	22:15
corecode	http://gist.github.com/155937 how can that happen?	22:16
Ilari	osfameron: It just marks it resolved. But conflicted merge is by defintion nontrivial, and nontrivial merges require commit to do.	22:16
	← stuffcorpse left	22:16
doener	osfameron: you fix the first one and add it. If that would implicitly commit, you'd commit the still unmerged version of the other file	22:16
	→ stuffcorpse joined	22:16
	← Short_some_thing left	22:16
corecode	did the repo change while it was cloning?	22:16
doener	corecode: http clone?	22:17
	← tbf left	22:17
	→ jamessharpe joined	22:17
osfameron	ok, so I've committed that conflicted file - am I ok to try merge again?	22:17
	← offby1 left	22:17
doener	osfameron: you committed the merge, not that file	22:17
	hacchi → ingmar	22:17
corecode	doener: no, local	22:17
doener	osfameron: merging again would just tell you that there's nothing to merge	22:18
	← Altreus left	22:18
corecode	doener: but there are updates happening in parallel	22:18
doener	corecode: using a plain path, or file:// ?	22:18
corecode	plain path	22:18
osfameron	doener: nope, I removed changes from index and committed just that file	22:18
corecode	is there a difference?	22:18
	→ offby1 joined	22:18
Ilari	corecode: Yes.	22:19
doener	osfameron: you still committed the merge. You just told git that merging that branch results in changing that file only	22:19
osfameron	doener: eeeek!	22:19
	or is that a good thing?	22:19
corecode	would this be an effect?	22:20
doener	osfameron: unless you want the merge result to be that only that one file changed, no, it's not a good thing	22:20
	← k0001 left	22:20
osfameron	doener: I thought you told me I had to commit that file first?	22:20
	ingmar → hacchi	22:21
Ilari	osfameron: How did you commit only that one file?	22:21
doener	osfameron: no, I told you that you need to commit, and not to run "git merge" again	22:21
	→ rhyhann joined	22:21
rhyhann	Hi	22:21
osfameron	ah... this isn't the same as conflict in rebase, where you have to restart the rebase (with --continue) A merge is just OK once you've resolved the conflicts?	22:22
	→ j416 joined	22:22
osfameron	Ilari: um, I unstaged everything and then added in that one file again	22:22
rhyhann	Can I merge changes in a repositoly from another one that is not a fork, nor a clone ?	22:22
iaj	how can I view the changes that have been made (e.g. the files that have been updated) in last git pull?	22:22
doener	osfameron: "rebase --continue" doesn't restart the rebase, it continues it	22:22
	→ k0001 joined	22:22
	hacchi → ingmar	22:23
Ilari	osfameron: Unstaged? If its via reset, it aborts the merge.	22:23
doener	osfameron: and a merge just needs to be committed, and not tried again (unless you aborted it, of course)	22:23
osfameron	doener: um, yes, I meant "restart" as in "continue" as opposed to "start over", but not very clear, sorry	22:23
	ah ok	22:23
doener	Ilari: "git reset <path>" doesn't abort the merge AFAIK	22:23
osfameron	how can I tell if I aborted my merge or not?	22:24
Ilari	osfameron: Does git log list it as merge (merge line)?	22:24
osfameron	Ilari: I don't see anything about merge there: just commit ...\nAuthor:...\nDate:\n.... and the description	22:25
Ilari	osfameron: Thats not merge then...	22:25
	← jonmasters left	22:25
osfameron	ok... so I think I want to back out that commit and start over...	22:25
	osfameron sighs	22:25
	← doener left	22:26
osfameron	doener, Ilari: thanks for the pointers ;-)	22:26
Ilari	osfameron: reflog and reset.	22:26
osfameron	--hard ?	22:26
	→ [JL]String joined	22:26
	[JL]String → stringoO	22:26
	← T_X left	22:27
Ilari	osfameron: Yeah...	22:27
	→ jonmasters joined	22:27
	→ T_X joined	22:27
	→ offby1` joined	22:29
	← nitrate left	22:30
	← stuffcorpse left	22:31
	→ stuffcorpse joined	22:31
	→ trefn_ joined	22:31
	→ Rolenun joined	22:32
	← trefn left	22:32
	→ pint_size joined	22:33
	← offby1 left	22:33
	← pint_size left	22:33
	offby1` → offby1	22:33
	← urbanmonk left	22:34
Rolenun	does anyone know if there is an irc channel for git# ? :)	22:34
	→ wolog joined	22:34
Ilari	Rolenun: git#? The C# reimplementation?	22:35
Rolenun	yep	22:37
	→ bugfux joined	22:38
Ilari	I guess this channel is the most apporiate (no specific channel exists).	22:38
Rolenun	Ilari, thanks, I was just wondering. I dont have any questions about it. Just doing my homework before I become involved with the project ;)	22:40
	← tvw left	22:41
	← csgeek left	22:45
	← Linkadmin left	22:45
	← eMyller left	22:46
	← jamessharpe left	22:47
	← stuffcorpse left	22:50
	→ stuffcorpse joined	22:51
	← chris2 left	22:51
	→ cemerick joined	22:53
	→ csgeek joined	22:54
	→ tlm` joined	22:55
	← cemerick left	22:55
almostautomated	There. With DIFF1 being diff -w of A^ B and DIFF2 for B^ A:: If [ -z "$(git diff A B)" ] && [ -z "$(diff -w $DIFF1 $DIFF2 \| grep -E "[<>] [+-]\s+[:print:]+")" ]; then echo "Whitespace fixing worked."; fi	22:57
	← robinr left	23:00
	← twas left	23:02
	→ TML joined	23:02
	ryandotsmith → ryandotsmith\|awa	23:04
	ryandotsmith\|awa → ryandotsmith	23:05
	← Cromulent left	23:05
	← giallu left	23:06
	← intripoon left	23:07
almostautomated	actually that should have been $(git diff -b A B)	23:08
	← tlm left	23:10
	← Octalot left	23:10
	→ obiwan_chinobi joined	23:11
almostautomated	and I guess that second diff could be tightened up with -b instead of -w as well	23:11
aziz	how can I amend the previous commit and remove files from it?	23:12
aeruder	aziz: git rm <files> ; git commit --amend	23:14
	just make the changes like you were gonna make another commit, and the amend makes the changes into the previous commit	23:14
aziz	k thanks	23:15
	← TML left	23:15
	→ TML joined	23:15
	← priidu left	23:15
	→ Cromulent joined	23:20
	→ Linkadmin joined	23:21
	← AkumaStreak left	23:23
	← jonmasters left	23:23
	← TML left	23:24
	→ AkumaStreak joined	23:24
elspeth	hm	23:24
	i need to do a merge	23:24
	→ brizly1 joined	23:24
elspeth	how do i specify a mergetool?	23:24
aziz	hm, that wasn't actually a good idea. I didn't actually want to remove the files.	23:24
	→ TML joined	23:25
aziz	I wanted to not have them in the previous commit.	23:25
elspeth	automerge failed, need to do a manual one, shouldn't be too hard, just can't remember the invocation to say 'i want to use textmate to merge this manually'	23:25
	→ jonmasters joined	23:26
aziz	Anyway, I rolled back the last commit: git reset --soft HEAD^	23:27
	→ bitprophet joined	23:29
almostautomated	is --quiet and --exit-code broken on 1.6.3 for git diff ?	23:30
	← k0001 left	23:31
	← tlm` left	23:32
	→ w\|shimmy joined	23:33
	→ marvil07_ joined	23:34
	→ ^Einstei1 joined	23:34
	← ^Einstein left	23:35
	← j416 left	23:36
	→ strangy joined	23:36
	← marvil07 left	23:36
	← thiago_home left	23:37
	→ k4jcw joined	23:37
rudi_s	elspeth: man git-mergetool	23:37
Gitbot	elspeth: the git-mergetool manpage can be found at http://git.or.cz/man/git-mergetool	23:37
k4jcw	Good evening. I'm reading the git docs as we speak, but is there way to refresh the current clone after the master has been updated?	23:38
	← djgera left	23:38
	→ hydrogen joined	23:38
rudi_s	k4jcw: git fetch and git merge or git pull (which does the fetch/merge)	23:38
	→ Base2 joined	23:39
k4jcw	Ah. Excellent, thank you, rudi_s.	23:39
rudi_s	np	23:40
	← brizly left	23:40
elspeth	rudi_s: thankee	23:40
rudi_s	np	23:40
	← DavidKlein left	23:41
	→ lightcap joined	23:43
	← lightcap left	23:44
	← EruditeHermit left	23:46
	→ lightcap joined	23:46
	← wshimmy left	23:46
	← angerman left	23:47
	← d0k left	23:48
	← bugfux left	23:51
	→ jjuran joined	23:53
	← lightcap left	23:53
	→ lightcap joined	23:53
	← shruggar left	23:54
	← lightcap left	23:56
	← iaj left	23:57
	→ iaj joined	23:59
	→ cloudhead joined	23:59

Logs Search ←Prev date Next date→

Channels Documentation