IRCloggy #git 2010-03-16

Provider of IRC logs since 2005.

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

	2010-03-16
	← hobodave left	00:00
	← bentob0x left	00:00
	→ jamiew joined	00:01
	→ rburton- joined	00:03
yacoob	gebi, yup. I'm wondering how much should I rely on the fact that this approach works	00:03
	(using it to reduce lag for .git on NFS)	00:04
	→ rretzbach joined	00:04
rretzbach	(github) Hi, I just forked someones repo, made changes, commited and pushed. Then I've sent a pull request. Should I have created a special branch instead of using "master"?	00:05
dr_jesus	hey if I'm in the middle of a conflict resolution and I want to look at blobs from some other branch how do I do that without stashing and doing a checkout?	00:05
	← fool_ left	00:06
	← tedoc2000 left	00:07
	← Voziv left	00:07
	← Hackykid left	00:07
	← jrheard left	00:07
	→ fool_ joined	00:08
	← fool_ left	00:08
	→ fool_ joined	00:08
	← hyperair left	00:08
	← jasonn1234 left	00:08
wereHamster	dr_jesus: git show <otherbranch>:<path>	00:09
	→ nusse joined	00:09
nusse	hi there	00:09
	if have a lot of small changes in my working directory, everyone containing the same string, is it possible to ignore those changes?	00:10
	← struberg left	00:10
	→ hyperair joined	00:10
cbreak	git can not ignore changes	00:10
nusse	i'd like to add all other changes, but add -p will take ages	00:11
cbreak	you can add all changes	00:11
	and then remove those changes again	00:11
	... somehow	00:11
	maybe...	00:11
dr_jesus	wereHamster: thanks!	00:12
nusse	will take too long	00:12
	is there a way to decide automatically wether to include a hunk? (grep -qv evilstring)	00:13
Ilari	rretzbach: Well, if master contains only changes intended for upstream + what's already in upstream, its ok.	00:13
	→ jrmuizel joined	00:15
	← ehabkost left	00:15
jamiew	can someone commit/push to a submodule from the parent project?	00:16
rretzbach	Ilari: So as long as I don't work on that fork it's okay? Or when exactly is this method not good?	00:16
Ilari	rretzbach: When branch you request them to pull contains both changes meant for upstream and changes not meant to upstream.	00:17
	rretzbach: Because merges are all or none.	00:17
	← hipe left	00:17
	← dbb_ left	00:18
	→ friskd joined	00:18
	← josh0x0 left	00:23
	← bronson left	00:23
	← mssm left	00:23
	← amerine left	00:23
	← Theravadan left	00:28
	→ L_R_N joined	00:33
vinse	im a little confused. i did a merge that included a change to a file that i didnt want. so i did "git checkout HEAD^^ file" and when i look i can see i have the right version now, and git status tell me the file is modified	00:34
	but git diff has no output, what gives?	00:34
	← oliver3 left	00:35
	← LRN left	00:35
	← jas4711 left	00:35
	→ jmcantrell joined	00:36
wereHamster	the changes are already staged	00:36
	→ Swimming_Bird joined	00:36
wereHamster	git diff --cached	00:36
	← enherit left	00:37
	→ jas4711 joined	00:37
vinse	wereHamster: ah ok, thanks	00:39
	→ Theravadan joined	00:40
	→ guyzmo joined	00:41
	← pantsman left	00:41
	→ hobodave joined	00:41
mugwump	hachi: sub a sub-tree graft, you'd have to make a commit which has the tree you want first and graft that	00:44
	← raorn left	00:44
mugwump	log should follow the graft - perhaps there's another reachable commit which refers to the one you want to graft away? gitk is your friend	00:44
	→ raorn joined	00:44
	→ WALoeIII joined	00:44
	→ p0w3r3d joined	00:45
	← coldswitch left	00:47
hachi	mugwump: well that's the thing, since it's a subtree graft, I have to list both parents	00:47
	so it sees both, yes	00:47
	← Theravadan left	00:47
hachi	and it has to see both, otherwise it won't have the tree there to recognize the 'subtree' part of the graft	00:48
	or at least, I'll end up with a tree that looks like it only contains the grafted in files or something	00:48
	or, the most likely of anything... I don't actually /know/ what's going on, and insted I'm just wrong :)	00:49
	← loincloth left	00:49
	→ loincloth joined	00:49
	→ super3 joined	00:50
super3	hey guys	00:50
	← patthoyts left	00:52
mugwump	hachi, how well have you grasped the commit/tree/file DAG concept?	00:52
hachi	yes, but this whole subtree part throws me for a loop	00:53
mugwump	so, a subtree merge is where a branch is merged in in the history, but the changes from the branch were applied to a subtree of the target, not the root	00:53
	so in some respects it just looks like a regular merge	00:53
hachi	yeah, and this is what puzzles me... there is no explicit callout that the subtree was from the branch, there's just a magic hash matchup that tells it that is the case	00:54
	so, perhaps the first thing I should ask	00:54
	is there a way I can query the subtree at a treeish, and the tree at another treeish	00:55
	to see if they are actually the same	00:55
	before I go off grafting them together assuming that they are the same	00:55
mugwump	using git rev-parse commit:path	00:55
	I think rev-parse also has a pipe mode if you're doing it a lot	00:55
	→ ninjaneo joined	00:55
mugwump	well, guess that doesn't matter	00:56
hachi	well, is this going to give it to me per-file	00:56
	I'm about to try it	00:56
	so maybe I'll answer my own question	00:56
ninjaneo	I have cloned a repository (over ssh), I commit the changes, and push them, but is there a way to then fast forward the remote to the latest?	00:57
	← Cynos left	00:57
mugwump	ninjaneo: you mean the checkout attached to the remote?	00:57
	there's a hook for that	00:57
	← robo47 left	00:57
ninjaneo	yeah, because I just have a "git-shell", so no real ssh access	00:57
	but I would like to be able to update the remote (hosted website)	00:58
mugwump	mmm. git doesn't change the working copy when you push	00:58
ninjaneo	yeah I was hoping there was some way to tell teh remote to checkout the new files, something like an rsync	00:58
mugwump	because if there was a merge conflict the wrong end has to resolve it	00:58
	there is, post-checkout hook	00:59
	← gk left	00:59
	ninjaneo googles git post checkout hoook	00:59
mugwump	normally people say "you shouldn't do that", but I think git should try, and fail to push if there would be a merge conflict	00:59
	ninjaneo: faq non-bare	00:59
Gitbot	ninjaneo: Pushing to non-bare repositories is discouraged. See http://git.wiki.kernel.org/index.php/GitFaq#non-bare	00:59
ninjaneo	yeah, or allow the "merger" to control the conflict/attempt to resolve	00:59
mugwump	no need	01:00
	→ gk joined	01:00
mugwump	they can pull first, then push again	01:00
ninjaneo	I can make it bare, thanks for the tip. And i'll check out the link	01:00
hachi	mugwump: okay... the rev-parse hash is going to include things like commit messages, right?	01:00
	or... does it... beats me actually	01:00
mugwump	hachi: depends what you pass it	01:00
	to be explicit you can use ^{tree}	01:00
	at the end of the revision	01:00
hachi	oh sweeet	01:00
	yeah	01:00
mugwump	but normally that's implied by putting :/path after the rev	01:01
hachi	well, I can also pass the :	01:01
	← jamiew left	01:01
hachi	the : implies the path only, not the whole thing	01:01
	← jonshea left	01:01
	← tbf left	01:01
	← hobodave left	01:02
	eladmeid_ → eladmeidar_	01:03
	← super3 left	01:04
	eladmeidar_ → eladmeidar	01:04
	← ed1t left	01:04
	← anjen left	01:05
	→ seren joined	01:05
	← seren left	01:05
	→ seren joined	01:05
	→ anjen joined	01:06
ninjaneo	well thanks for the info guys, happy coding	01:06
	← ninjaneo left	01:06
	→ jinzhu joined	01:09
	← Chillance left	01:09
	← spaceonline left	01:09
hachi	mugwump: so, anytime there is a merge.. git takes the top of the tree in the branch being merged in, and compares it to every possible subtree in the merge point?	01:10
	→ ed1t joined	01:10
	← ed1t left	01:10
	→ ed1t joined	01:10
	← ed1t left	01:10
hachi	or is this more only when grafts happen? and doesn't this cause a scalability issue?	01:11
	course, maybe these things are stored in a btree, so the trees within the revisions can be looked up really fast, I dunno	01:11
	→ alexmchale_ joined	01:12
mugwump	hmm, well the hashes might not even be the same	01:12
hachi	I would assume that when the hashes aren't the same, the subtree magic doesn't work	01:12
mugwump	reading in all the tree objects for a revision isn't that expensive - see how long it takes to produce, eg 'git log --stat'	01:12
	← mjf left	01:13
	→ coldswitch joined	01:13
hachi	or are the matches done per-file?	01:13
mugwump	I don't really know tbh	01:13
	→ stringo0 joined	01:13
hachi	oh... blah... is this thing handled internally the same as a moved file?	01:13
mugwump	utsl	01:13
hachi	okie, no worries	01:14
	→ jasonn1234 joined	01:14
mugwump	you do have to specify to use subtree merge	01:14
	with -s or whatever	01:14
hachi	I have no way to specify that when I'm grafting	01:14
	there is no command, only the file	01:14
	← ab3 left	01:15
	→ ab3 joined	01:15
mugwump	hmm, match-trees.c:score_trees seems to be a function for that	01:15
	← jrmuizel left	01:16
mugwump	see http://repo.or.cz/w/git.git/commit/68faf6893	01:16
Gitbot	[git 68faf6893]: http://tinyurl.com/y8uh76e -- A new merge stragety 'subtree'.	01:17
	← p3rror left	01:17
	← seren left	01:17
	← godsflaw left	01:18
mugwump	ok, and it got folded into the recursive strategy. is it really automatic?	01:18
	the docs on the commit messages say, eg git merge -s subtree -Xtheirs other	01:18
	← stringo0 left	01:19
mugwump	searching git.git commit messages for your feature name can be very enlightening :)	01:19
	← Gullik left	01:20
	→ greg8si joined	01:20
	← offby1 left	01:20
	→ jrmuizel_ joined	01:21
	← jrmuizel_ left	01:21
	← friskd left	01:24
	← ab3 left	01:24
	← alexmchale_ left	01:25
	→ homa_rano joined	01:25
	→ stringo0 joined	01:26
	← jaql left	01:26
	← ultraviolet left	01:26
	→ dcampano joined	01:27
	→ jaql joined	01:28
hachi	mugwump: thanks	01:28
	this has been very enlightening	01:28
	← WALoeIII left	01:30
	← hyperair left	01:30
	→ hyperair joined	01:31
hachi	isn't there a command for constructing grafts, other than editing the grafts file?	01:32
	→ Kelly_zhou joined	01:33
hachi	oh, no... bad google match	01:33
	→ Lgb joined	01:34
	→ ultraviolet joined	01:34
	→ dbb joined	01:34
	→ joevano joined	01:37
	← novas0x2a left	01:37
	← Adaptee left	01:37
	← joevano left	01:38
	→ tvw joined	01:40
	→ Garen joined	01:41
	→ Bass10 joined	01:42
	→ joevano joined	01:43
	← joevano left	01:44
	← zodiak left	01:44
	← coldswitch left	01:46
	← gebi left	01:46
	← rretzbach left	01:47
	← bradhaydon left	01:47
	→ johnw joined	01:48
	→ joevano joined	01:48
	← joevano left	01:48
	→ joevano joined	01:48
	← joevano left	01:49
	→ pklingem joined	01:49
	→ alexmchale_ joined	01:49
	← jinzhu left	01:50
	→ jinzhu joined	01:51
	← p0w3r3d left	01:54
	← themill left	01:54
	→ joevano joined	01:54
	← tvw left	01:55
	← joevano left	01:55
mikelifeguard	How do you download just one file from a remote repo (master at HEAD)?	01:57
icefox	you can just fetch the object's that contains the HEAD object	01:58
	though just using a web interface is usually easier	01:58
mugwump	icefox, is that your git-achievements thing?	01:59
cbreak	git doesn't offer anything to download just one file	01:59
icefox	mugwump: yup	01:59
mugwump	I lol'd :)	01:59
icefox	fun hu :)	01:59
mugwump	I was wondering if you got XP for commits too	02:00
icefox	On reddit someone suggested putting it together with a git tutorial That would be a fun way to learn	02:00
	yah Author	02:00
	← richcollins left	02:01
icefox	Apprentice Author, Author, and Master Author	02:01
	Most all of the general commands let you go through the three stages	02:01
	mugwump: any amusing achievement you think should be added?	02:02
mugwump	heh, well I'm not sure how you'd do Historian	02:03
	→ jrmuizel joined	02:03
mugwump	for me it was writing a gfi importer and using git write-file directly	02:03
	er	02:03
	git hash-object -w :)	02:03
cbreak	"You are your father": rewriting history with git rebase -i	02:04
	"With the powers of the sand of time": git reset HEAD@{1}	02:04
icefox	mugwump: yah was thinking of adding something once you used update-index, hash-object and update-ref	02:04
mugwump	hmm, rebase -i is more like a Studious Author or something	02:04
	→ Textmode joined	02:04
mugwump	Plumber when you use X plumbing commands ? :)	02:05
icefox	already got rebase -i, it is Butcher	02:05
cbreak	"Master Negotiator": merging more than three heads at a time	02:05
icefox	Plumber is good, i'll add that	02:05
cbreak	"On the shoulders of giants": git pull --rebase	02:05
Textmode	morning all	02:06
mugwump	there's no achievement for merge?	02:06
icefox	mugwump: hmm no somehow missed that	02:06
mugwump	or mergetool ... hmm	02:06
icefox	What should it's name be?	02:06
mugwump	Peacemaker?	02:06
cbreak	"Everyone else hates you now": git push -f	02:07
	"... and no one noticed": pushing to a non-bare repo	02:07
	→ joevano joined	02:08
mugwump	that "hates you now" ... how about "Thug" ? :)	02:08
icefox	Thug is good :)	02:08
	→ catalis joined	02:08
cbreak	you award titles for each achievement?	02:08
	← sh1mmer left	02:08
icefox	cbreak: yah :) Achievements always have some silly name	02:09
	← joevano left	02:09
cbreak	in WoW they are stupid strings	02:09
	like, very stupid	02:09
icefox	heh	02:10
	yup	02:10
cbreak	you usually get them for making your own life more miserable than already	02:11
	icefox ads Plumber and Thug	02:11
	← domcat left	02:11
	← EricInBNE left	02:11
	→ friskd joined	02:12
mugwump	"Gypsie" for "git bundle" ?	02:12
cbreak	"Lost in Time": committing with detached head	02:12
mikelifeguard	How do you get a tarball or file with git-archive that doesn't have the pax header?	02:13
	← Determinist left	02:13
	← imajes left	02:13
icefox	already got bundle, called it delivery boy I think	02:13
mugwump	ah ok :)	02:13
icefox	added one for merge	02:13
	→ A_GWT_STARTUP joined	02:13
icefox	cbreak: whats the best way to detect that?	02:14
cbreak	"Get a Mac": turning on auto crlf	02:14
icefox	;D	02:14
cbreak	icefox: detecting when?	02:14
icefox	when commiting with a detached head	02:14
cbreak	git status says you're not on a head	02:15
	and HEAD doesn't start with ref:	02:15
	→ imajes joined	02:15
	→ loinclot_ joined	02:15
	← loincloth left	02:16
sitaram	bremner_: ping	02:16
	← A_GWT_STARTUP left	02:16
bremner_	sitaram: hi	02:16
	← hced left	02:17
	→ oliver3 joined	02:17
	→ a_GWT_startup joined	02:18
	← gambler left	02:19
	→ gambler joined	02:19
	→ sh1mmer joined	02:19
sitaram	bremner_: what TZ are you in? and how dependent are you on "personal branches" being available in every repo in your config?	02:21
bremner_	I'm in AST (GMT - 5, I think). I only use personal branches in a few repos	02:21
	→ warlock_mza joined	02:22
	→ coldswitch joined	02:22
	→ VVSiz_ joined	02:23
sitaram	bremner_: will you be around another hour or so; I may have something I want you to review and comment on	02:24
	← sh1mmer left	02:25
bremner_	sitaram: I should probably go to sleep before that, feel free to query me and I'll look at it in the morning?	02:25
	→ hummingkid joined	02:26
sitaram	bremner_: no problem... this is a biggish change so I plan to email some others also anyway, so no hurry	02:26
bremner_	sitaram: ok, I'll look for it.	02:27
	← flaguy48 left	02:27
	← VVSiz left	02:28
	→ fda314925 joined	02:28
ubuntum	in what cases might git stop watching changes?	02:28
	Git has stopped reporting changes for commit in git status	02:29
	← Tureba left	02:29
	← defectiv left	02:30
	← alexmchale_ left	02:32
ubuntum	very mysterious	02:32
	→ alexmchale_ joined	02:32
	← intripoon left	02:32
mugwump	a badly written .gitignore rule, perhaps	02:33
ubuntum	hmm	02:33
	→ anathematic joined	02:33
mugwump	also possible to do via .git/info/exclude	02:33
	← scarabx left	02:33
anathematic	hi I'm using git svn and I'm getting the error dcommitted on a detached HEAD because you gave a revision argument. when I push	02:33
	any ideas?	02:33
mugwump	there's index flags to ignore paths - see man git update-index	02:33
parasti	gitignore only applies to non-tracked files, though	02:33
ubuntum	thanks	02:33
bremner_	sitaram: in general should the same description that creates git-daemon-export-ok set permissions, or is this too much of a security risk?	02:34
	← dgl left	02:34
ubuntum	mugwump the thing is that it was working 15 minutes ago	02:35
	it just stopped when I changed to a branch	02:35
	I even tried going back to master	02:35
mugwump	did you use -f ?	02:35
ubuntum	even changes to master is not shown in the status	02:35
	-f for which command?	02:35
sitaram	bremner_: I would not like to mess with permissions needlessly... I assume you mean making the repo world-readable right? I dont want to assume there is no other way (such as putting apache in the same group as git) to do that and double guess the user	02:35
mugwump	switching branches	02:36
ubuntum	hmm no	02:36
	i just did checkout	02:36
mugwump	ok	02:36
	← friskd left	02:36
mugwump	and then you made some changes, and they're not showing up?	02:36
bremner_	sitaram: OK, just checking it wasn't an oversite.	02:36
ubuntum	yup not at all	02:36
mugwump	can you see them in 'git gui' ?	02:36
ubuntum	added a new file even	02:36
	ok just a sec	02:36
cbreak	icefox: "Let there be LIGHT": creating a commit without parent (the initial commit)	02:36
mugwump	or 'git diff HEAD' ?	02:36
ubuntum	git diff HEAD shows nothing	02:37
	→ offby1 joined	02:38
mugwump	even if you manually add to a file which is definitely in the repository (use git whatchanged for instance)?	02:38
ubuntum	gitk showing the past logs, but not since 5:30 pm (9:37 here)	02:38
mugwump	hmm	02:38
ubuntum	I am sure I did at least 1 commit since	02:38
mugwump	perhaps there's some silent internal error with the index or something	02:38
	→ friskd joined	02:38
mugwump	strace might show that up	02:38
cbreak	did you commit without being on a branch?	02:38
mugwump	can you see your commit with 'git log --all --since="4 hours ago"' ?	02:38
cbreak	commits done with detached head are gone	02:39
mugwump	also, try adding -g to that command	02:39
cbreak	check git reflog	02:39
mugwump	to search the reflog	02:39
ubuntum	git diff with a file name leads to zero line	02:39
	← pkw left	02:39
ubuntum	git logs shows but same as the gui	02:39
	← tildeequals left	02:39
ubuntum	ok just a sec	02:39
mugwump	what? 'git gui' does not show logs	02:40
cbreak	what about git diff HEAD~5?	02:40
	→ sh1mmer joined	02:40
ubuntum	git gui shows logs	02:40
cbreak	does git status show the branch you are on?	02:40
ubuntum	i mean git log with a file name gives me the same result as the gui	02:40
mugwump	'git gui' is the interactive commit tool	02:40
	used to be called git citool	02:40
	→ zodiak joined	02:40
ubuntum	on branch master	02:40
mugwump	it's good for showing you the working tree vs. the index vs committed state	02:41
ubuntum	ohh I fired gitk as the gui tool	02:41
	did not know that one	02:41
mugwump	best thing since darcs record	02:42
ubuntum	i think it somehow added myfiles to "untracked"	02:42
	→ thedeeno joined	02:42
ubuntum	because in the gui tool my files is shown as untracked, not staged	02:42
	I wonder if something went fubar when I change to a previous branch	02:43
	shoot, what is the proper way to bring those untrack files without mesing up the tree?	02:43
mugwump	click on them	02:44
	git gui never changes your WC	02:44
ubuntum	ok	02:44
	← smuf left	02:44
ubuntum	shall I just commit in the gui?	02:44
mugwump	only if Balthar says it's ok	02:45
	→ entrosca joined	02:46
mugwump	you'll need chicken blood, some leaves from the ayahuasca shrub	02:46
ubuntum	yeah sure that would solve the misery in the world as well	02:46
	← loinclot_ left	02:47
	→ loincloth joined	02:48
	← imajes left	02:48
icefox	cbreak: added an achievement for comiting without a parent	02:49
	commiting	02:49
ubuntum	lost a branch totally :(	02:51
	→ imajes joined	02:51
	← parasti left	02:51
Textmode	thats got to be an achivement in git, too...	02:51
ubuntum	I agree	02:51
	I just did the unimaginable	02:51
	the branch is there but the some files in the branch has gotten deleted	02:52
	anyways my loss if your gain	02:52
	what would be the best way to delete an unwanted branch?	02:52
mugwump	look, if they were ever added, they will be in your object repository	02:53
	← pheaver left	02:53
	← eletuchy left	02:53
ubuntum	mugwump are you talking about deleting or loosing about a branch>	02:53
	← loincloth left	02:54
	→ loincloth joined	02:54
mugwump	I mean, if you think you have lost files, it's actually quite hard to do that	02:54
ubuntum	well when I change to that branch, the files are not there	02:55
mugwump	they'll be on another branch then probably	02:55
	try 'gitk --all -- path/'	02:55
ubuntum	yes I have them in the master	02:55
	but	02:56
mugwump	well, perhaps you need to merge master to your branch	02:56
cbreak	if they were deleted out of git, you could see them in the history	02:56
	→ kssddy joined	02:56
	← alexmchale_ left	02:56
ubuntum	they were not supposed to be in the master, but Git stopped tracking myfiles in the branch so I changed to Master.	02:57
	← alester left	02:57
cbreak	git doesn't stop unless you tell it to	02:58
ubuntum	maybe for some reason Git actually never switched to the Branch	02:58
	→ amerine joined	02:58
ubuntum	cbreak, that is a possibility but not in my case	02:58
	because I am the only one working here and I did not change anyting except my Python code	02:58
cbreak	git has no own will	02:59
ubuntum	sure	02:59
	← hummingkid left	02:59
cbreak	it doesn't even run unless you invoke a command	02:59
	also, if git stops tracking something	02:59
	it would still be in the history	02:59
ubuntum	yeah sure I understand all that	02:59
cbreak	so you could look at it with git show someoldrevision:path/to/file	03:00
ubuntum	the ebst explanation I have is that I deleted a branch because I realized that I did not need it	03:00
	since than nothing is same	03:00
	with -d option	03:01
	while I was in the master	03:01
	after the branch delete, git status showed me all the files I have for delete inj the next commit	03:01
	→ hobodave joined	03:01
ubuntum	I assumed that it was talking about the files in the deleted branch	03:01
	and since then I have no log or diff	03:02
	← darwin_ left	03:02
	← hobodave left	03:02
	← ProperNoun left	03:02
	→ DanRigby joined	03:03
ubuntum	When I committed right after branch deletion, it did not delete or remove any of my files so again I assumed Git took care of an unneeded deleted branch	03:04
	and it stopped showing files for delete in the git status	03:04
	→ loinclot_ joined	03:04
	← loincloth left	03:05
ubuntum	I am now stuck with my release version which I wanted it as a seperate branch :)	03:06
	→ ProperNoun joined	03:06
	← DanRigby left	03:06
	→ mdel joined	03:07
	← sh1mmer left	03:08
	→ hobodave joined	03:10
	→ bronson joined	03:12
	← hobodave left	03:15
	→ magneto joined	03:16
magneto	hello #git	03:17
	anyone got a link to a tutorial or other info on using git for backing up configuration files, like the /etc directory?	03:18
	← anathematic left	03:18
	→ bronson_ joined	03:18
	← bronson left	03:18
mikelifeguard	magneto: you probably want etckeeper	03:19
	← catalis left	03:19
mikelifeguard	it is packaged for many distros	03:19
	→ alexmchale_ joined	03:19
magneto	thx	03:19
rudi_s_	magneto: Be careful with git stash/git stash apply though, it can brake the access rights on files like passwd/shadow/..	03:19
	← coldswitch left	03:20
	→ cytrinox_ joined	03:20
	← madewokherd left	03:21
	→ defectiv joined	03:21
	← alexmchale_ left	03:22
	→ aryounce joined	03:22
	← cytrinox left	03:23
	cytrinox_ → cytrinox	03:23
	← magneto left	03:24
	→ catalis joined	03:25
	← amerine left	03:26
	← jasonn1234 left	03:27
	→ coldswitch joined	03:27
	← enyovelcora left	03:28
	← hyperair left	03:30
	← anjen left	03:32
	← loinclot_ left	03:34
	→ loincloth joined	03:34
	→ zjason` joined	03:38
	→ alexmchale_ joined	03:38
	→ anathematic joined	03:40
anathematic	I'm using git + svn at the moment and upon a git fetch I'm getting the error Checksum mismatch: blah/blah.rb	03:41
	:) any ideas?	03:41
	mikelifeguard → Mike\|\|gone	03:41
	← zjason left	03:42
	→ SingAlong joined	03:51
SingAlong	I get this "gitosis.serve.main:Repository read access denied" error when trying to push to my remote repo. I entered the passphrase wrong the first time i entered and I'm not able to use it.	03:51
	→ timj__ joined	03:52
	← Modius left	03:53
ubuntum	does anyone know if it is possible to complete git commands and branches in the console?	03:53
	→ Modius joined	03:53
	← dbb left	03:54
ubuntum	some auto complete	03:54
	← friskd left	03:54
	← engla left	03:55
	← timj_ left	03:55
SingAlong	anyone?	03:56
davr	ubuntum: yes, with bash completion	03:56
	ubuntum: I think 'apt-get install bash-completion' should be all you need in ubuntu	03:57
	SingAlong: so you're saying git is storing the incorrect password and you don't know how to enter the correct one?	03:58
	← schlortoo left	03:58
davr	SingAlong: I don't know about gitosis, maybe check inside .git/config or $HOME/.gitconfig or something to see if you can find where the password is stored	03:58
	← Twisted_Mentat left	03:59
	← zenwryly left	03:59
	← alexmchale_ left	04:01
	→ godsflaw joined	04:03
	→ mtkd_ joined	04:06
	→ magneto joined	04:06
	← aresnick left	04:08
	← mtkd_ left	04:09
	→ dbb_ joined	04:12
	← laj left	04:13
	← jailbox left	04:13
	→ ArdRigh joined	04:14
	← jmcantrell left	04:16
	← Bass10 left	04:16
	← jfkw left	04:17
	← frakturfreak left	04:21
	← ericlee left	04:22
	→ hipe joined	04:23
	→ ericlee joined	04:23
	← aspotashev\|eeepc left	04:24
	← disappearedng left	04:25
	← disappearedng_ left	04:26
	→ anathematic_ joined	04:27
	→ munichlinux joined	04:28
	← anathematic left	04:30
	anathematic_ → anathematic	04:30
	→ tedoc2000 joined	04:31
	→ brizly joined	04:33
	→ fool__ joined	04:34
	← fool__ left	04:34
	→ fool__ joined	04:34
	← warlock_mza left	04:34
	→ stantont joined	04:36
	← fool_ left	04:37
	→ disappearedng joined	04:40
	→ disappearedng_ joined	04:40
	← aryounce left	04:41
	→ mdel_ joined	04:41
	← mdel left	04:43
	← mdel_ left	04:47
	→ loinclot_ joined	04:50
	← loincloth left	04:50
	→ tildeequals joined	04:52
	→ jams joined	04:52
jams	hello, we have a repo where a few of the devs commited stuff with the username root. Is there an easy way to change this to reflect the actual dev name?	04:53
	→ sh1mmer joined	04:54
	← greg8si left	04:54
jams	the commits are rather old..towards the beginning of the repo history	04:54
DrNick	not easily, no	04:54
	or, rather, it's easy to do, but it has unpleasant ramifications	04:54
jams	hmmf ok	04:54
coldswitch	git filter-branch ?	04:54
DrNick	it would require rewriting the entire history using filter-branch from the point of the first root commit	04:55
jams	we have like 8 different "roots" and a few varaitions of the same dev name	04:55
	ubuntum → kursad	04:55
DrNick	which means you'd have to discards any existing clones	04:55
	→ mtkd_ joined	04:55
jams	i see..they would have to throw away what they have and reclone	04:56
DrNick	yeah	04:56
	← Swimming_Bird left	04:56
jams	thats doable	04:56
	i will look into the filter-branch as you suggested	04:57
	→ novas0x2a joined	04:57
jams	thank you	04:57
	and if it becomes a royal pita then we can live with the roots :)	04:57
DrNick	alternately, you could make them all push what they have to some central repo, rewrite that, then clone from that, discard all their "private" branches from the central repo	04:57
	and then clean each individual repo	04:58
	→ pheaver joined	04:58
jams	well in this case we do have one "central" place that everybody checks in and out of	04:58
	i should say push and pull from	04:58
	i don't suppose there is a way to have a pre-check to stop this from happening in the future?	04:59
DrNick	it's going to require some kind of flag day either way	04:59
	jams: oh, yeah, you can write arbitrarily complex hooks	04:59
	both on the client and server	05:00
anathematic	in .git/config what does the fetch parameter mean?	05:00
jams	know of any prebuilt hooks that would be a good starting place?	05:00
	← pklingem left	05:00
	← jonmasters left	05:01
	← hipe left	05:02
	← tedoc2000 left	05:02
	→ hipe joined	05:02
jams	oh look at that google kindly suggested "git filter-branch change author"	05:03
	again thanks I know have a good starting point vs no clue	05:03
	*now	05:03
anathematic	I'm trying to fetch the latest commits from subversion and I'm receiving the error Checksum mismatch: app/helpers/contact_helper.rb	05:04
	any pointers?	05:04
	→ tedoc2000 joined	05:04
	→ richcollins joined	05:05
	← hipe left	05:06
	→ hipe joined	05:07
anathematic	anyyyoneee	05:07
	→ pkw joined	05:07
	← tro left	05:07
	→ tro joined	05:09
	← thedeeno left	05:10
	← hipe left	05:11
	→ hipe joined	05:12
	← merlin83 left	05:13
	← SingAlong left	05:13
	→ Theravadan joined	05:13
	→ lightcap_ joined	05:15
	→ khelll joined	05:15
lightcap_	I'm totally confused about how to merge/rebase a feature branch...	05:16
	I have two commits on a branch called controller_data	05:17
	and master has 8 new commits...	05:17
	→ bradhaydon joined	05:17
lightcap_	how the heck do I merge/rebase changes from master into my feature branch...	05:17
	→ jonmasters joined	05:17
lightcap_	I thought I just did, but it was apparently on no_branch...	05:18
	← AAA_awright left	05:18
lightcap_	and now I show no changes on my feature branch... even after doing what I thought would resolve the conflicts.	05:19
	← mtkd_ left	05:19
	→ AAA_awright joined	05:20
	← hipe left	05:20
	→ hipe joined	05:20
lightcap_	and honestly, I'm not even sure if I should be rebasing or merging...	05:21
	→ merlin83 joined	05:22
	← merlin83 left	05:22
	→ merlin83 joined	05:22
	→ indyplanets joined	05:23
	→ eletuchy joined	05:23
	← tedoc2000 left	05:24
	← rburton- left	05:24
	← eno left	05:25
	→ cloudhead joined	05:26
	→ eno joined	05:26
johnw	lightcap_: if you're on "no branch": did you have conflicts?	05:27
lightcap_	I did, yeah.	05:27
	and I thought I resolved all of them.	05:27
johnw	you have to resolve them, "git add" those files, then do "git rebase --continue"	05:27
lightcap_	I guess mostly I'm wondering what the right workflow is...	05:28
	→ aryounce joined	05:28
lightcap_	johnw: even on "no_branch"?	05:28
johnw	no_branch is just an intermediate state	05:28
	→ amerine joined	05:28
johnw	it means you are in "mid-rebase"	05:28
	← _iron left	05:28
lightcap_	ok	05:28
	let me try that again...	05:28
	← hakunin left	05:28
lightcap_	so, from my feature branch I should do git rebase origin/master?	05:29
johnw	first do git rebase --abort	05:29
	then do what you justsaid	05:29
lightcap_	ok	05:29
	do I need to do a git fetch origin master first from my feature branch?	05:29
	← entrosca left	05:29
johnw	never hurts	05:30
lightcap_	okay, that didn't seem to do anything, so now I"ll do a git rebase origin/master	05:30
	→ hakunin joined	05:30
lightcap_	lots of conflicts...	05:30
	but I'm not surprised...	05:30
	→ zenwryly joined	05:31
	→ tedoc2000 joined	05:32
lightcap_	okay, once I resolve the commits and commit on no_branch then I do git rebase --continue?	05:34
johnw	don't commit on no branch	05:35
	just "git add" the files	05:35
lightcap_	oh..	05:35
	oops..	05:35
johnw	you want them in the index, not as a commit	05:35
lightcap_	what if I did commit	05:35
	can I just revert	05:35
johnw	yeah, do: git reset --soft HEAD^	05:35
lightcap_	ok	05:36
	← kukks left	05:36
	← pheaver left	05:37
	← tedoc2000 left	05:38
	← jrmuizel left	05:39
	← _nickel left	05:42
	→ scarabx joined	05:44
	← project2501a left	05:45
anathematic	I'm getting an Index mismatch: when trying to fetch the latest from my svn repo, ideas?	05:45
	→ loincloth joined	05:46
	← loinclot_ left	05:47
	→ codeshepherd joined	05:53
	← thomas- left	05:54
	← richcollins left	05:55
	← ari-_-e left	05:55
	← indyplanets left	05:57
johnw	anathematic: paste?	05:58
	← eletuchy left	05:58
anathematic	johnw: http://pastie.org/private/nsqhfkvhzt3gpwjkwbdrla	05:58
johnw	hmmm.. i've never seen that one before	05:59
	can you svn checkout?	05:59
	(with svn, that is)	05:59
anathematic	johnw: no strangely enough	05:59
	however I can git svn clone fine	05:59
	→ jrmuizel joined	06:00
	→ richcollins joined	06:00
	← richcollins left	06:01
	→ anu123 joined	06:01
	→ roop joined	06:03
	→ Jamer joined	06:03
	← roop left	06:03
	→ pheaver joined	06:05
	→ g0 joined	06:05
	← a_GWT_startup left	06:06
	→ EricInBNE joined	06:06
	→ Pupeno_ joined	06:07
	← anu123 left	06:07
	← Pupeno left	06:08
	Pupeno_ → Pupeno	06:08
	→ sde_antrix1 joined	06:10
	→ richcollins joined	06:11
	← stringo0 left	06:14
	← magneto left	06:15
	← jrmuizel left	06:15
	← amerine left	06:15
	→ lresende_ joined	06:18
	← lresende_ left	06:19
	← aryounce left	06:20
lightcap_	johnw: thanks for your help...	06:21
johnw	anytim	06:21
	e	06:21
lightcap_	johnw: it's a little confusing that after git rebase --continue it doesn't let you know that the files are added to the last commit...	06:21
	← lresende left	06:21
lightcap_	johnw: so I think everything is good now, with master rebased into my feature branch... but what do I do to merge that feature branch into master so I can push it up now?	06:22
	johnw: nevermind... a "git merge feature_branch" did the trick (I think)	06:25
johnw	yep	06:25
	it should have been a fast-forward merge	06:25
lightcap_	yeah it was	06:25
johnw	then you're all good	06:25
	sometimes, it's better to use --no-ff on git merge	06:26
	that way, if you need to you can revert the entire merge in one operation, or re-revert it to put it back	06:26
lightcap_	I still can't get used to all those "Merge branch 'feature_branch'" commit log entries... but oh well	06:26
	I guess they're good to keep track of wtf you just did.	06:26
	thank god for tests... I'd have no confidence in what I did if I didn't have tests :-)	06:27
	← Jamer left	06:27
lightcap_	johnw: are you suggesting to usually use --no-ff then?	06:27
	on a big merge anyhow	06:27
johnw	for a mainline published branch, it makes history clearer	06:27
	← eno left	06:27
johnw	you can see that this X commits came in with a certain feature implementation	06:28
	for working branches I don't bother	06:28
	otherwise, with fast-forward you lose the fact of the merge	06:28
lightcap_	really? there's a commit message saying the merge occured though	06:29
johnw	it looks like you made those X commits in sequence to master directly	06:29
	really?	06:29
lightcap_	yeah	06:29
johnw	git log shows a merge commit?	06:29
	hmm	06:29
	that doesn't sound like fast-forward then..	06:29
	← Boriskr left	06:29
lightcap_	ah, no	06:29
	it says merge made by recursive	06:30
	← wolog left	06:30
lightcap_	as if I know what that means :-)	06:30
johnw	so, --no-ff means there is always merge commit	06:30
lightcap_	ah	06:31
	so it's forcing that commit message...	06:31
johnw	and the merge commit itsef	06:31
lightcap_	that makes sense...	06:31
johnw	which maintains the historical fact of the merge	06:31
lightcap_	what's recursive?	06:31
johnw	in my master branch, I like to retain as much history as possible	06:31
	see: http://stackoverflow.com/questions/366860/when-would-you-use-the-different-git-merge-strategies	06:32
lightcap_	yeah, I don't it when there's a ton of those messages, but it does make it clearer. Of course, when they're local feature branches, no one else knows wtf that merge meant	06:32
	← munichlinux left	06:32
johnw	that's why I avoid unnecessary merge commits in working branches	06:32
	in unpublished working branches I even prefer rebase	06:32
	but it depends on what kind of history I want to retain	06:32
	→ eno joined	06:33
lightcap_	johnw: ah, it probably used recursive, because not only did I resolve conflicts I also had to change a few things that weren't conflicts to get the tests to pass	06:33
johnw	lightcap_: I also recommend http://www.newartisans.com/2008/04/git-from-the-bottom-up.html, if you haven't already	06:33
lightcap_	don't think I have.	06:33
	thanks!	06:33
	→ Boriskr joined	06:33
	← radhermit left	06:34
	→ aziz joined	06:34
	→ simplechat joined	06:34
	← ReinH left	06:34
	← iggy left	06:35
	← m-takagi left	06:36
	← siraj left	06:36
	← cmyers left	06:36
	← bob2 left	06:36
	← ProperNoun left	06:36
	← ivan left	06:36
	← bdash left	06:36
	← snuxoll left	06:36
	← relaxok left	06:37
	→ ReinH joined	06:38
	→ bob2 joined	06:38
	→ bdash joined	06:38
	→ cmyers joined	06:38
	→ siraj joined	06:38
	→ m-takagi joined	06:38
	→ snuxoll joined	06:38
	→ radhermit joined	06:39
	← hakunin left	06:40
	← brizly left	06:41
	→ ProperNoun joined	06:42
vinse	if i have branch bar that's a child of HEAD on branch foo, and i rebase foo, is bar affected?	06:42
	→ khmarbaise joined	06:43
	→ iggy joined	06:43
	← vals_ left	06:43
	→ ivan joined	06:43
	← ivan left	06:43
	→ ivan joined	06:43
	← kursad left	06:44
	m-takagi → m-takagi_	06:45
segher	vinse: nope	06:45
	← infogulch1 left	06:45
	← khmarbaise left	06:46
vinse	segher: ok, so the parent of bar is the commit that it was branched from, and that relationship never changes. yes?	06:48
	of course yes, never mind	06:48
	→ jokaro joined	06:51
	← xelados left	06:51
segher	vinse: "rebase some-branch" changes that branch, and nothing else	06:52
	← lightcap_ left	06:52
	→ Beaver joined	06:55
	← Beaver left	06:57
	→ loinclot_ joined	07:00
	← loincloth left	07:01
	← ArdRigh left	07:01
	← tg left	07:01
	→ tg joined	07:02
	→ rseifert joined	07:02
	→ fmardini joined	07:04
	← Garen left	07:04
	→ infogulch joined	07:08
	→ ixti joined	07:09
	← icefox left	07:10
	→ schlortoo joined	07:10
	→ chid joined	07:11
	← simplechat left	07:13
	→ lightcap_ joined	07:13
	→ davidw joined	07:13
	← infogulch left	07:14
	← Pupeno left	07:14
	davidw → Guest3290	07:14
jaalto	I'm trying to get permanently rid of BINARY file in the repository. Following some tutorial they suggested "git filter-branch". However, the object of the binary was not removed. What am I missing here: http://paste.lisp.org/display/96451 (See related discussion http://kerneltrap.org/mailarchive/git/2007/10/7/331568)	07:16
	→ dpino joined	07:16
	→ infogulch joined	07:16
johnw	reading...	07:17
jaalto	... two quotes are missing from --index-filter, but don't mind that (it's output from script).	07:17
johnw	btw, my git-scripts repo has a script called "flush" that very aggressively prunes out all objects not explicitly referenced	07:18
	← cloudhead left	07:18
johnw	http://github.com/jwiegley/git-scripts	07:18
	you cleared the stash?	07:19
	you cleaned .git/info/refs?	07:19
jaalto	I would have expecte that "git reflog expire..." and "git gc --aggressive --prune" would clear all	07:19
johnw	no	07:19
	jaalto reading git-flush	07:20
Bombe	The size of the pack file proves nothing. :)	07:20
	← anathematic left	07:20
	→ maxthoursie joined	07:20
jaalto	Bombe: ?	07:20
johnw	also, see the script git-unpack	07:21
jaalto	It's taking space. The only add was the BINARY.	07:21
johnw	it will remove all pack files	07:21
	then you can really see where size is taken up	07:21
	(then use git gc to repack)	07:21
jaalto	johnw: Nice scripts!	07:22
johnw	not all are from me, but many	07:22
	most are thanks to this channel, and discussions here	07:23
jaalto	johnw: Heh, I have tons of similar "commands" and still need more :-)	07:23
johnw	i have a lot of alias too	07:23
	like rc for rebase --continue, ra for rebase --abort, ca for commit --amend, etc.	07:24
jaalto	johnw: Myt "git" bassh function is currenly 3300 lines long and will grow after you're scripts are added	07:24
	your	07:24
	→ friskd joined	07:25
	→ kar8nga joined	07:26
jaalto	johnw: Exactly: rb = rebase, rbC (rebase --continue), rbA (rebase --abort), rbS (rebase --skip), rbL [N] (list patch number 000N), ....	07:27
	← maxthoursie left	07:27
	← kar8nga left	07:27
	← bauruine left	07:27
	→ maxthoursie joined	07:27
	→ munichlinux joined	07:27
	← BlindOracle left	07:30
	→ bauruine joined	07:31
	→ wolog joined	07:33
	→ ph^_ joined	07:33
	jelmer_ → jelmer	07:34
	→ BlindOracle joined	07:34
	← jokaro left	07:35
	← ph^_ left	07:38
	m-takagi_ → m-takagi	07:40
	→ ph^_ joined	07:42
	→ project2501a joined	07:42
	→ loincloth joined	07:42
	← loinclot_ left	07:42
	→ Garen joined	07:42
	→ psoo joined	07:42
	→ giallu joined	07:42
	→ priidu_ joined	07:43
	← ph^_ left	07:44
	→ mtkd_ joined	07:44
	→ dj_tjerk joined	07:44
	→ carl- joined	07:45
	→ dgl joined	07:46
	→ cedricv joined	07:47
	← g0 left	07:51
	→ DavidKlein joined	07:53
	← ProperNoun left	07:54
	→ _iron joined	07:55
	→ loinclot_ joined	07:56
	→ thiago_home joined	07:56
jaalto	johnw: I got it working, the magic was your scripts "prune" command that must be run in the end (git prune --expire 0). In case anyone is interested http://paste.lisp.org/display/96452	07:56
	← loincloth left	07:56
	← scarabx left	07:57
johnw	ahh	07:57
	→ trivol joined	07:57
	→ Pupeno joined	07:57
	→ gnufied joined	07:57
	← richcollins left	07:57
johnw	i remember spending a long time refining that	07:57
	← gnufied left	07:57
	→ tango_ joined	07:58
	→ ProperNoun joined	07:58
	← ProperNoun left	07:58
	→ ProperNoun joined	07:58
jaalto	johnw: I add "--expire 0", which your script did not have. Just to be sure.	07:58
	← r11t left	08:01
	→ ph^_ joined	08:02
	← mtkd_ left	08:02
	→ mtkd_ joined	08:03
	→ steffkes joined	08:04
	→ g0 joined	08:06
	→ maxter_ joined	08:08
	← ProperNoun left	08:09
	→ kumbayo joined	08:09
	← novas0x2a left	08:10
	← psoo left	08:10
	← ph^_ left	08:13
	→ rchavik joined	08:13
	← coldswitch left	08:13
	→ ncopa joined	08:14
ncopa	hi	08:14
	← johnw left	08:14
	→ ProperNoun joined	08:14
	→ entrosca joined	08:15
ncopa	what was the command to selective apply diff hunks?	08:15
	→ froschi joined	08:15
	→ psoo joined	08:15
	← drry left	08:18
wereHamster	git apply; git add -p ?	08:19
	← entrosca left	08:20
	→ trebor_dki joined	08:20
ncopa	thanks	08:20
	← bauruine left	08:21
	← Theravadan left	08:22
	→ kkoehne_ joined	08:23
	← defectiv left	08:24
	→ loincloth joined	08:24
	← loinclot_ left	08:25
	→ bentob0x joined	08:27
	→ d0ugal joined	08:28
	← ixti left	08:28
	→ struberg joined	08:29
	← thiago_home left	08:30
	← jelmer left	08:32
	← maxter_ left	08:32
	→ relaxok joined	08:32
	→ cilly joined	08:33
	→ ph^_ joined	08:34
	← Pupeno left	08:35
	→ lucsky joined	08:36
	→ ixti joined	08:36
	→ FDCS joined	08:37
	→ silence joined	08:39
	→ skoop joined	08:39
	silence → Guest51541	08:39
jaalto	How to check if rebase is in progress (from a script)?	08:39
	← rchavik left	08:39
	← Mage left	08:40
	→ entrosca joined	08:40
	→ mastro joined	08:40
	← froschi left	08:40
	→ gebi joined	08:40
	← entrosca left	08:40
	→ Pupeno joined	08:41
	← ixti left	08:41
	→ entrosca joined	08:41
	← stepnem left	08:44
	→ stepnem joined	08:44
	→ froschi joined	08:44
	← Guest51541 left	08:45
	← mtkd_ left	08:45
	→ zoniq joined	08:46
locklace	jaalto: iirc you can look for .git/rebase-apply	08:46
	→ mtkd_ joined	08:46
	← bradhaydon left	08:48
	→ ixti joined	08:49
	→ hipe_ joined	08:49
	← hipe left	08:49
	→ nicoulaj joined	08:49
	→ dereine joined	08:50
jaalto	locklace: ok	08:50
dereine	i'm writing a post-update script, any ideas how i can find out the name of the repositorie?	08:50
	← Boriskr left	08:52
	← lightcap_ left	08:53
wereHamster	dereine: a git repository doesn't have a name (other than the directory name it is stored in)	08:53
Saur	dereine: What do you mean by the name of the repository?	08:53
dereine	Saur: i'm using gitolite, and there each repo has a name	08:53
	damn i forgot to mention it	08:53
wereHamster	dereine: basename $GIT_DIR	08:54
	(should work)	08:54
	→ kkoehne joined	08:54
	→ mssm joined	08:54
dereine	thx!	08:54
	← dj_tjerk left	08:55
	→ lhunath joined	08:57
lhunath	$ gs set-tree trunk	08:57
	config --get svn-remote.svn.fetch :refs/remotes/git-svn$: command returned error: 1	08:57
	what's that?	08:57
	alias gs='git svn'	08:57
	and how do I debug it?	08:57
	→ Boriskr joined	09:03
	← mtkd_ left	09:04
	→ domcat joined	09:08
	→ shruggar joined	09:08
	→ avtobiff joined	09:09
	← zoniq left	09:10
	→ zoniq joined	09:10
	→ arialdo joined	09:14
arialdo	Hi. I'm able to sign my tags using git tag -u my_email but not with -s option. Git says "secret key not available"	09:15
	← zoniq left	09:15
	→ zoniq joined	09:15
	← FDCS left	09:15
arialdo	What's the reason? I googled with no results :\|	09:17
Saur	arialdo: Do you have a GPG-key for the email adress you use?	09:17
arialdo	Saur: yes. I even put a signingkey = my_gpg_id in .gitconfig	09:19
lhunath	load and make available your gpg-agent in the environment.	09:20
arialdo	Saur: using git tag -u myemail@domain I can sign tags	09:20
	← zoniq left	09:20
	→ novas0x2a joined	09:21
Saur	Well, a gpg-agent won't change anything, since it is only used to cache the pass phrase	09:21
Rhonda	arialdo: That's a bug in gpg, it only likes UIDs that are available in the secret key, which isn't required for a working environment.	09:21
	arialdo: You also need to use signkey in that cases.	09:21
	→ zoniq joined	09:21
Tass`	how to set pull to rebase by default?	09:22
Rhonda	arialdo: See the discussion in http://bugs.debian.org/497372	09:22
arialdo	Rhonda: Thanks! Do you mean I should simply solve my problem using -u myemail@domain?	09:23
Rhonda	arialdo: That's at least what I did, or rather, did put it down into the config to not have to hand it on the command line all the time.	09:24
	← gebi left	09:24
arialdo	Rhonda: Ok! Thanks again	09:24
	→ jailbox joined	09:24
	← AAA_awright left	09:25
	→ dytrivedi joined	09:25
	m-takagi → m-takagi_	09:27
	← bentob0x left	09:27
	Guest3290 → davidw	09:27
	→ doobry joined	09:27
	← davidw left	09:27
	→ davidw joined	09:27
	→ engla joined	09:29
	← zoniq left	09:29
	← engla left	09:29
	→ engla joined	09:29
	→ tatsuya joined	09:30
	→ zoniq joined	09:30
sitaram	dereine: you don't need the name for most purposes, except perhaps for reporting/logging. If you're using gitolite, the env var GL_REPO is set to the repo name without the $REPO_BASE at the start and the ".git" at the end	09:31
dereine	sitaram: thx! and thanks for this awesome software	09:32
sitaram	dereine: you're welcome!	09:32
	← krawek left	09:32
sitaram	dereine: basename $GIT_DIR will also work but if your repos are multi-level like foo/bar, it will only show you "bar". Depends what you need of course	09:32
	← mssm left	09:32
sitaram	GL_REPO on the other hand is essentially the full path minus the $REPO_PATH at the start and ".git" at the end	09:33
	← loincloth left	09:34
	→ loincloth joined	09:35
	→ AAA_awright joined	09:36
	→ eeanm joined	09:37
	→ sylr joined	09:37
sylr	Hi guys	09:38
eeanm	hey so I need to merge in a repo into a subdirectory of another repo	09:38
sylr	What is the good way for merging two identical branch (identical in the way that each diff between commits are the same) but they have not the same commit id ?	09:39
	-branch+branches	09:39
	← AAA_awright left	09:40
sitaram	sylr: just merge them; git should detect the sameness and avoid conflicts, AFAIK. Or if you can afford to change the SHA of one of them, rebase it on top of the other -- that will certainly de-dup the identical commits	09:41
sylr	sitaram, the branch are on two separte repo, so I push one branch on the other repo and I rebase it angaint the other branch ?	09:42
	→ jokaro joined	09:42
	← zoniq left	09:42
	→ AAA_awright joined	09:43
	→ zoniq joined	09:43
sitaram	sylr: yes. Although one usually does a fetch instead of a push, it comes to the same thing anyway	09:43
	sylr: but seriously, make a temp copy of one branch and try a merge first -- it may just work out!	09:44
	→ darkskiez joined	09:44
sylr	sitaram, I tried, it duplicates the common history	09:45
	← darkskiez left	09:45
	← struberg left	09:45
sitaram	sylr: that's not the same thing as "conflicts", and is usually harmless; although I guess you don't want it for perhaps aesthetic reasons.... rebase will work then	09:45
sylr	sitaram, esthetic is one thing, but the history duplicated it quite large, around 60000 commits, whats why I'm trying to not duplicate it	09:47
	thanks	09:47
	→ bauruine joined	09:47
	← multi_io left	09:47
	→ multi_io joined	09:48
sitaram	ok	09:49
	sylr: although you will find that the duplicated commits don't actually take up much space :) I would still call it aesthetics, as in "clutters up the gitk display" ;-)	09:50
sylr	sitaram, yeah :P	09:50
	I loke clean thinks too	09:51
	things	09:51
sitaram	like	09:53
	:)	09:53
sylr	yes, that too	09:54
	I love to patch my sentences	09:54
	sitaram, It worked thoroughly	09:56
sitaram	great	09:56
	→ loinclot_ joined	09:56
	← loincloth left	09:57
jaalto	The "rebase -i" doesn't obey "edit" marked from script. See tes http://paste.lisp.org/display/96455	09:57
	→ hced joined	09:57
	→ ssm2017 joined	09:58
ssm2017	hello	09:58
	← kkoehne left	09:58
	→ petrux joined	09:59
ssm2017	is there a way to auto increment tags, like i have : v0.0.7 and i would like to use a command to automatically tag v0.0.8	09:59
jaalto	What is diffrent here? In interactive EDITOR the 'pick' would also be changed to 'edit'	09:59
lhunath	ssm2017: write a script.	09:59
charon	jaalto: you can't run together the -i option	09:59
	→ thvdburgt joined	09:59
charon	jaalto: since it takes a backup suffix	09:59
	jaalto: changing to -ni should fix it	09:59
	→ bluenovember joined	10:00
sitaram	jaalto: is the print statement missing a semicolon>	10:00
	→ themill joined	10:00
sitaram	?	10:00
charon	(the option to _perl_, this has nothing whatsoever to do with git)	10:00
lhunath	ssm2017: alternatively, do the math yourself; I can't imagine it's that big of a deal.	10:00
jaalto	charon: No, plain "-i" is "in place editing" without backup.	10:00
	sitaram: Not required if last stement in block	10:00
sitaram	jaalto: which it is not	10:00
	charon sighs	10:01
sitaram	you have a "warn" right after...	10:01
ssm2017	lhunath, ok, thank you for attention, i thought that an option could do it for me like : git -coffee -sugars 1	10:01
jaalto	sitaram: Argh, right	10:01
	sitaram: thanks.	10:01
charon	jaalto: i'm 100% confident that -i must be separated by a space from the next option. rtfm.	10:02
sitaram	or make it -ni as charon said earlier	10:02
jaalto	charon: Nope :-) works in Perl 5.10	10:02
	← ssm2017 left	10:03
jaalto	charon: Dang. Correction: you're right.	10:03
charon	i was bitten too often by perl -ine '...' not working	10:04
	so i know :	10:04
	:p	10:04
sitaram	jaalto: you're double wrong actually, not only does it not work, once you switch to -ni or separate them, the "-n" will make the explicit while redundant :)	10:05
	jaalto working with the script, thanks fo all comments	10:06
	→ struberg joined	10:07
jaalto	sitaram: Yay, loose the "-n" and all works. Now making this take $i configurable for "git rbi N" command to edit any N commit.	10:08
	→ Mpenz joined	10:08
jaalto	charon, sitaram: thanks.	10:09
sitaram	you're welcome!	10:09
	→ vvsh joined	10:11
	→ loincloth joined	10:18
	← loinclot_ left	10:18
	→ stikonas joined	10:18
dereine	sitaram: can i asc you a silly ssh-key access question?	10:20
sitaram	sure	10:23
	but I may have to step out for an hour or so in a few min (as soon as my wife is ready!) so... ask anyway :)	10:24
wereHamster	someone else can answer if sitaram is away..	10:25
	→ gaveen joined	10:25
	→ jelmer joined	10:25
	→ mosno joined	10:26
dereine	sitaram: if i use ssh git@gitolite i get R W gitolite-admin	10:26
	← zoniq left	10:26
	→ zoniq joined	10:27
dereine	sitaram: if i execute the gl-easy-install i got http://pastebin.com/3KdKNvQ8	10:27
	← jelmer left	10:27
	→ jelmer_ joined	10:27
	jelmer_ → jelmer	10:28
sitaram	dereine: well it seems your ~/id_rsa key is also your gitolite key now (my best guess)	10:28
	→ galderz joined	10:29
sitaram	dereine: is this the same account/workstation from which you ran easy install?	10:29
dereine	sitaram: yes	10:29
sitaram	I suggest you run the sshkeys-lint script in src/ (run it without args to see what it needs and how to run it)	10:30
	→ loinclot_ joined	10:30
sitaram	(you will have to obtain your ~/.ssh/authorized_keys file from the server somehow)	10:30
	← loincloth left	10:30
sitaram	also try "ssh-add -l" and make sure all your keys show up or none of them do... in fact run through doc/6 :)	10:31
dereine	thx i will try it add	10:31
	← froschi left	10:33
	← will__ left	10:34
	L_R_N → LRN	10:36
	→ radarek joined	10:36
	→ luigi70 joined	10:36
	← Kelly_zhou left	10:36
luigi70	hello	10:36
	!list	10:37
	← struberg left	10:37
	← luigi70 left	10:38
	← tg left	10:39
Textmode	meep?	10:40
	→ Glenjamin joined	10:40
	← mosno left	10:40
Glenjamin	hi guys, can anyone tell me how to turn a local branch into a tracking branch?	10:41
	← jelmer left	10:41
	→ viniciusluiz joined	10:42
	← domcat left	10:42
Ilari	Glenjamin: 'git push --set-upstream <remote> <branchname>'?	10:42
Glenjamin	sounds plausible	10:42
	is that documented somewhere?	10:43
	its not mentioned in the man page for git push	10:43
	→ struberg joined	10:43
	← SRabbelier left	10:43
	→ Sonderblade joined	10:43
Glenjamin	or maybe its just my version actually	10:43
	→ mithro joined	10:45
	← munichlinux left	10:45
Ilari	Glenjamin: Yeah, appeared in 1.7.0.	10:45
	← zoniq left	10:46
	← g0 left	10:46
Glenjamin	and is there a switch to get branch to list my local branches and what they track?	10:46
	→ nadim joined	10:46
	→ zoniq joined	10:46
	→ alexmchale_ joined	10:46
Ilari	Glenjamin: 'git remote show <remote>' might show something useful...	10:46
Glenjamin	ah yes, thats the one	10:47
	thanks :)	10:47
	→ parigaudi joined	10:48
sitaram	"git branch -v -v" also works	10:48
	if you have multiple remotes and different branches are from different remotes that seems to be better	10:49
	→ tg joined	10:49
Glenjamin	good ol' guessable git commands	10:49
	ah yes, thats much nicer	10:50
shruggar	is there a way to list all commits Not made by a particular author? (I'm trying to see all commits made by someone other than me over the weekend)	10:50
wereHamster	tried --not --author=.. ?	10:51
sitaram	in fact "git remot show remote" seems to show many branches that are not actually tracked -- if the remote has a branch of the same name it assumes it is tracking it. Hmm weird	10:51
	→ schulty joined	10:51
mapreduce	drizzd: I tested the commands for mirroring an svn repository to git, and they worked. Will you be editing the page?	10:52
	→ frakturfreak joined	10:52
shruggar	wereHamster, seems to only take the --author part into account	10:52
drizzd	mapreduce: sure	10:52
	thanks for testing	10:52
	← stikonas left	10:53
	→ g0 joined	10:53
	→ stikonas joined	10:54
	→ blazzy joined	10:55
	← zoniq left	10:55
	→ zoniq joined	10:55
drizzd	mapreduce: hmm, do you have more than one branch in your subversion repo? and if yes, did they get mirrored?	10:56
	→ frakturfreak_ joined	10:57
	← frakturfreak_ left	10:57
	→ frakturfreak_ joined	10:57
Ilari	sitaram: BTW: I reinstalled git-daemon2 yesterday (had to reinstall whole system due to fs corruption). I used newest from pu-ssh-plus and for some reason gitolite-admin post-update hook didn't get installed (404).	10:57
	← pheaver left	10:57
	← zoniq left	10:58
	← frakturfreak left	10:58
	frakturfreak_ → frakturfreak	10:58
Ilari	sitaram: AFACIT, update hooks did get installed to all repos...	10:58
blazzy	I submitted a patch to a project viea email, the patch was applied. I pulled down the recent changes. Now my commit is in my local history twice. Is that to be expected?	10:59
	*via	10:59
	← engla left	10:59
sitaram	Ilari: pu-ssh-plus didn't install post-update hook? let me check...	10:59
Ilari	blazzy: Patches don't preserve commit IDs. Use rebase when doing patches...	10:59
wereHamster	blazzy: yes, that's to be expected if you submit the patch by mail	10:59
Ilari	sitaram: I used gl-install to install it (easy install is not possible).	11:00
	← VVSiz_ left	11:00
blazzy	so if it bothers me I should just rebase? Does it matter if I leave it as is?	11:00
	→ VVSiz joined	11:00
wereHamster	I'd rebase (or otherwise get rid of your local commit)	11:00
Ilari	blazzy: Merges and patches don't mix very well.	11:00
	sitaram: And this system is so barebones I wouldn't be surprised if some needed command was missing and that caused the problem...	11:01
sitaram	Ilari: why is easy-install not possible? (you dont use any ssh keys at all? That is still something I have to fix...)	11:01
Ilari	sitaram: No sshd.	11:02
blazzy	Ilari, wereHamster, I guess I'll rebase then. thanks.	11:02
	→ zoniq joined	11:02
sitaram	Ilari: aah I see... perhaps I should upgrade pu-ssh-plus to the latest master, which has a way to install server-side and much simpler way to install... (although the "gl-setup" anciallry command may need some change for your environment)	11:03
	Ilari: will do it in a day or two... need to test	11:03
Ilari	sitaram: The only setup changes that are needed is to write gitolite.conf (needed anyway) plus adding command to HUP the daemon on gitolite-admin push (to make it reload SRP database).	11:04
	sitaram: That HUP was what I was needing hook passthrough for post-update anyway.	11:05
sitaram	Ilari: I only made hook passthru for "update" so far; I will add that also, bump pu-ssh-plus to that, and ping you	11:06
	(plus test of course)	11:06
Ilari	sitaram: I store server SRP database (for password authentication) inside gitolite-admin	11:06
sitaram	Ilari: SRP?	11:07
	sitaram forgets quickly.... is old!	11:07
Ilari	sitaram: Password authentication protocol.	11:07
	→ ntoll joined	11:09
sitaram	Ilari: aah http://srp.stanford.edu/	11:10
	→ textarcana joined	11:10
	← loinclot_ left	11:12
	→ loincloth joined	11:12
	← ringlej left	11:12
	← Lgb left	11:13
	→ Gullik joined	11:13
	← yev\|away left	11:14
	→ Chillance joined	11:14
	← TDJACR left	11:14
Ilari	The authentication I most often use with git-daemon2 is Unix authentication (currently supported authentications: Unix, Openpgp, SSH, SRP)...	11:15
	Oh yeah, and theres anonyomous as well...	11:15
	← kpreid_ left	11:16
	→ kpreid_ joined	11:16
	→ robo47 joined	11:17
drizzd	mapreduce: ok, my solution was wrong, because it only mirrors trunk.	11:17
sitaram	Ilari: and this SRP implementation is within git-daemon2? (perhaps via openssl of course)?	11:18
drizzd	mapreduce: To fix the issue you had with the default branch, execute 'git symbolic-ref HEAD refs/heads/trunk' in the mirror repo.	11:18
	→ TDJACR joined	11:18
Ilari	sitaram: And it can't happen in post-receive since post-receive runs first and HUP needs to occur after gitolite checkouts its admin repo.	11:18
	sitaram: Via gnutls.	11:18
	sitaram: Yes, its in git-daemon2. Namespace prefix: "srp-".	11:19
	← lhunath left	11:20
	← dytrivedi left	11:20
	→ froschi joined	11:20
sitaram	Ilari: ok... please bear with me for a day or two; I'll upgrade pu-ssh-plus to the point where you can install it purely server side (except you give it some dummy pubkey that doesn't get used later in your setup)	11:21
Ilari	sitaram: What would be useful is letting enter initial key ID (for openpgp or ssh key) and then give gitolite-admin RW privs for that user.	11:23
	→ codeshepherd_ joined	11:23
	← alexmchale_ left	11:25
	← codeshepherd left	11:26
	codeshepherd_ → codeshepherd	11:26
	← mastro left	11:26
	→ mastro joined	11:27
sitaram	Ilari: key ID? not sure what that is for ssh key. For pgp/gpg I sorta know what it is but not very familiar...	11:27
Ilari	sitaram: Like letting to name initial user 'ssh-cdf7ff90d88302b332e47aabed800ec6d99898b780efaae4200d578b' and not install public key.	11:28
sitaram	hmm	11:28
Ilari	sitaram: (obiviously needs switch for not installing key).	11:28
sitaram	let me get the basic stuff done and later I'll add that in... how will that long arg be passed?	11:29
	Ilari looks current gl-setup syntax...	11:29
Ilari	*looks up	11:29
	sitaram: 'gl-setup --no-key ssh-cdf7ff90d88302b332e47aabed800ec6d99898b780efaae4200d578b' (pick better option name...)?	11:30
	← codeshepherd left	11:31
	← riz2 left	11:31
mapreduce	drizzd: Thanks, the new instructions look good. Looks like you added the missing 'cd project' too.	11:32
	→ saimon joined	11:32
sitaram	Ilari: why bother; if it matches that pattern (or other similar patterns) then we go that way...	11:32
	→ HeR joined	11:33
	← mithro left	11:33
drizzd	mapreduce: yeah	11:33
HeR	hi	11:33
	i have a problem with git and mabye someone here can help me	11:34
	git log only converts some colors	11:34
	%C(yellow) ist just passed	11:34
	same as %(bold blue)	11:34
Ilari	sitaram: Oh, and that testing RW = @all is bit so and so with git-daemon2. With that, @all really means ALL.	11:34
	→ g0bl1n joined	11:35
	→ codeshepherd joined	11:36
Ilari	sitaram: git-daemon2 does not have concept of "any authenticated user". Only "any user" and "this user".	11:37
	← zoniq left	11:37
drizzd	HeR: you can see the raw output produced by git log using "git --no-pager log --color \| LESS= less", does that show all the color escape codes you expect?	11:37
HeR	no	11:38
	i triedgit log --graph --pretty=format:'%Cred%h%Creset -%C(yellow)%d%Creset %s %Cgreen(%cr) %C(bold blue)<%an>%Creset' --abbrev-commit --date=relative	11:38
	→ zoniq joined	11:38
	→ alexmchale_ joined	11:38
HeR	* ESC[31m5b46aceESC[m -%C(yellow)%dESC[m Bugfix: Fatal Error ... ESC[32m(22 minutes ago) %C(bold blue)<HeR>ESC[m	11:39
	one line...	11:39
Ilari	HeR: Yeah, yellow didn't seem to work (and neither did bold blue.	11:41
HeR	yes	11:41
	and %d	11:41
	← bronson_ left	11:41
HeR	it worked well on my computer	11:41
Ilari	HeR: Git versions?	11:41
HeR	then i moved it to another and it doesnt work now	11:41
	1.5.6	11:41
	oh and local	11:42
	1.6.3.3	11:42
Ilari	HeR: Those are quite old (especially the first).	11:42
HeR	i installed git from the ubuntu-repository	11:42
frakturfreak	which ubuntu release?	11:43
Ilari	HeR: Most repositories have obsolete versions of git.	11:43
HeR	ubuntu 9.10	11:43
Ilari	HeR: Okay, Debian testing has 1.7.0, which isn't terrible.	11:43
	→ ehabkost joined	11:44
	→ ideogon joined	11:45
frakturfreak	HeR: you better should enable the git-core ppa https://launchpad.net/~git-core/+archive/ppa	11:45
	HeR: Or get the last git over git	11:45
HeR	k i try it thanks	11:45
	→ corni joined	11:48
saimon	Hi all...I need some advice...I forked a branch off master (call it A). I did work on a feature in A. More work was done on master including a huge merge. When trying to rebase master into A, I'm getting lots of conflicts due to the merge. So I thought I could just create a patch of all my commits on A. However, things get more complicated because at some point I had set git config core.autcrlf and for whatever reason a lot of my files on A go	11:53
	lf's. I fixed the issue with a commit on A that converted a load of files back to just lf. If I use git diff commitsha1...commitsha2 (where sha1 is the first commit I made on A and sha2 is the last commit I made on A) this patch includes the commit with the crlf fixes. When I try to apply this patch to a new branch B (off master) it fails because of the crlf fixes. So after all that my question is how can I interactively create a patch with onl	11:53
	commits I want?	11:53
	← project2501a left	11:53
	→ cannonball joined	11:53
	← cedricv left	11:55
	← ideogon left	11:55
sitaram	does anyone know the eqvt diff options for "git log -1 -p -c HEAD" when HEAD is a merge commit? (the -1 is to show only one commit, the -p is show a diff/patch, and the "-c" is the interesting one but doesn't seem to work with diff)	11:56
	← trochala left	11:58
	→ Tureba joined	11:58
	← jinzhu left	11:59
sitaram	Ilari: any other formats of keys being supplied? (other than "ssh-[0-9a-f]{40}"?)	12:00
Ilari	sitaram: Actually, there are 56 digits there. Then there's 'openpgp-'<32 or 40 (and prerhaps 56 or 64) hex digits>, there's 'unix-'<username>... Those are the most important.	12:02
	← codeshepherd left	12:02
sitaram	ok thanks	12:02
Ilari	sitaram: I have never seen key that would generate openpgp- followed by 56 or 64 digits (the key I use generates 40), some old keys generate 32.	12:03
	→ parasti joined	12:03
Ilari	sitaram: Better to activate that mode with some special option (in case yet another mode appears).	12:04
	← Glenjamin left	12:04
	→ joevano joined	12:04
sitaram	maybe later; for now -- the number of users for this, worldwide, is a grand total of one :)	12:05
	→ darwin_ joined	12:06
cannonball	In a new repo, if I have done a ./configure && make, then a make distclean, there are still some files hanging around that are not tracked. I want to remove them. What's the best way? 'git checkout -f master' doesn't do it for me because it's already in master.	12:07
wereHamster	cannonball: git clean -dfx	12:07
cannonball	thx	12:07
	→ cedricv joined	12:09
	← zoniq left	12:09
	→ zoniq joined	12:10
	← stantont left	12:12
	→ codeshepherd joined	12:12
HeR	i tried to add the official repository to a suse system	12:12
	but there ist the errormessage:	12:12
	Erzeuge Zwischenspeicher für Repository 'Base git repository' [fertig]	12:12
	Fehler beim Aufbau des lokalen Zwischenspeichers:	12:12
	'repo2solv.sh' '-o' '/var/cache/zypp/solv/git/solv' '/var/cache/zypp/raw/git'	12:12
	Unknown checksum type: 7: sha256	12:12
	Warnung: Deaktiviere Repository 'Base git repository' aufgrund des obigen Fehlers.	12:12
	← joevano left	12:13
	→ trochala joined	12:13
	← alexmchale_ left	12:15
	→ joevano joined	12:15
	→ d0k joined	12:16
	← joevano left	12:16
	← hipe_ left	12:16
	→ hipe joined	12:17
	→ stantont joined	12:17
	← saimon left	12:18
	→ tvw joined	12:18
Ilari	Ah, V3 keys generate 32-hexdigit IDs, V4 generates 40-hexdigit IDs. There isn't V5 in any standard...	12:18
HeR	??	12:19
bremner_	HeR: try setting LANG=C first, and also it sounds more like a SuSE problem than a git one	12:19
	→ saimon joined	12:20
	→ joevano joined	12:20
	← rgr left	12:20
HeR	Bremmer: i searched after this problem, but there is nothing about it	12:21
	and LANG=C didnt worked	12:21
bremner_	you probably want something like "export LANG=C" if your shell is bash	12:22
HeR	yes it is, and i tried this	12:22
frakturfreak	HeR: LANG C <command>	12:22
	HeR: sounds strange	12:23
	← codeshepherd left	12:23
	→ codeshepherd joined	12:23
HeR	i also tried to install the package myself, but it hast a lot of dependencies	12:23
	← priidu_ left	12:24
	← saimon left	12:25
	→ TheNewAndy joined	12:25
	← joevano left	12:25
bremner_	well, like I said before, you will probably have better luck on a SuSE support channel, if it exists	12:25
	→ joevano joined	12:26
	← thvdburgt left	12:27
	→ tbf joined	12:27
	← joevano left	12:28
bremner_	sitaram: if you're around, does renaming repos require something fancy? or just mv directory, update gitolite.conf?	12:29
	→ josh0x0 joined	12:30
sitaram	bremner_: rename it at the OS level first, then change gitolite config and push	12:30
	→ hugoxrosa joined	12:30
bremner_	thanks	12:30
	→ rchavik joined	12:30
	← codeshepherd left	12:30
sitaram	bremner_: if you do it the other way, the push will cause an empty repo to be created; you have to needlessly overwrite it	12:31
	→ codeshepherd joined	12:31
	→ nanoo joined	12:32
	→ whitman joined	12:32
	→ struggleyb joined	12:35
	→ y10 joined	12:36
	→ auslander joined	12:36
	← kumbayo left	12:37
	→ fool_ joined	12:41
	← fool_ left	12:41
	→ fool_ joined	12:41
	→ yevlempy joined	12:42
	← fool__ left	12:44
	← DavidKlein left	12:44
	→ codeswing joined	12:48
	→ Yuuhi joined	12:51
	→ Bass10 joined	12:52
	→ djszapi__ joined	12:52
	→ Axil joined	12:52
	← djszapi__ left	12:52
Axil	anyone here can help me setup pubkey auth on my linux server... i followed the steps in http://progit.org/book/ch4-4.html	12:53
	but it didnt work	12:53
	is it something to do with my ssh configuration	12:53
	?	12:53
	→ luke___ joined	12:53
	→ mm_work joined	12:54
	← mm_work left	12:54
	→ mm_work joined	12:54
luke___	Hello!	12:54
Axil	hello luke	12:54
fryguy	Axil: yes	12:54
luke___	How can I avoid merge branch of 'master' helloworld:helloworld entries from the history ?	12:55
	It can be rather annoying all the time, when more than 100 people work on the same project.	12:55
	← dpino left	12:55
Axil	fryguy... can you tell me if your ssh_config has PubkeyAuthentication yes	12:56
	AuthorizedKeysFile .ssh/authorized_keys	12:56
charon	luke___: are you using full src:dst fetchspecs with pull? those probably don't do what you think	12:56
fryguy	Axil: it doesn't, go ask for help in your OS channel for how to get ssh set up correctly	12:56
luke___	charon: All I do is git pull	12:57
	sorry, git push	12:57
	→ jfkw joined	12:57
luke___	without any src:dst	12:57
	← HeR left	12:57
	charon ponders	12:59
	← dgrift left	12:59
charon	luke___: so the message is actually: Merge branch 'master' of <some_URL> ?	12:59
	← TheNewAndy left	13:00
luke___	yep	13:01
charon	the most frequent cause is that people make a handful of local commits and thereafter run 'git pull' to "update"	13:01
luke___	yes, this is the situation.	13:02
	Should I push before git pull or should I commit after git pull all the time ?	13:02
	→ gebi joined	13:03
charon	the first doesn't work and the second doesn't make much sense to ask	13:03
	in fact you should always check that you have no uncommitted work _before_ running git pull/merge	13:03
	unless you know what you're doing and/or are linus	13:03
	← bauruine left	13:04
	→ j416 joined	13:05
	→ snucko joined	13:05
luke___	k, so isn't there any solution to avoid this message in the log ?	13:05
charon	luke___: there are some options, but it depends on the workflow you want to use. if you just want to keep your history linear, see the --rebase option (but make sure you first understand what it does!). if you frequently get into a situation where you hold off on pushing because it's not ready yet, but at the same time want to keep up-to-date, use topic branches (man gitworkflows)	13:05
Gitbot	luke___: the gitworkflows manpage can be found at http://git.or.cz/man/gitworkflows	13:05
luke___	charon: do you mean that then I should use a separate branch during the new feature development, and merge it after it succeeded ?	13:07
charon	yes	13:07
	→ sanjoyd joined	13:08
luke___	it's weird for windows developers to do, I don't believe they will do this way even I agree with your advice.	13:09
	← d0ugal left	13:09
luke___	all they see, they used TortoiseSVN earlier, and they didn't experience it with that application.	13:10
charon	meh	13:10
luke___	I guess because of that SVN stores .svn folder in every folder.	13:10
charon	no, that's an entirely separate issue.	13:10
	→ elb0w joined	13:11
	→ rburton- joined	13:11
elb0w	How do I check what repo i have set as master	13:11
	← pkahle left	13:11
elb0w	in linux	13:11
luke___	k, but the reality is that SVN doesn't meet with this problem.	13:11
Ilari	elb0w: 'git remote show origin'?	13:11
charon	luke___: my not so humble nor polite opinion: git is like an electric screwdriver. you can pretend it's SVN and keep hammering in nails with it, and complain about the "problem" that its grip is not very suited to hammering. but that doesn't change the fact that you'd be much more efficient with git+screws than with svn+nails.	13:12
elb0w	hrm	13:12
	didnt work	13:12
charon	only, you'll have to overcome your mind block and learn how to use the screwdriver.	13:12
	→ pklingem joined	13:12
Ilari	elb0w: Also, 'git remote -v' to show all remotes you have and their URLs.	13:12
charon	luke___: the way to pretend it's a hammer, btw, is that --rebase option.	13:12
	(and you'll have to adjust in several other places)	13:13
	← luke___ left	13:13
	→ Ryback_ joined	13:14
	charon wonders if luke was offended by his opinion	13:15
	→ patthoyts joined	13:16
locklace	hehe	13:17
	→ carolyn joined	13:18
	← tbf left	13:19
	→ goodieboy joined	13:19
	← goodieboy left	13:20
	→ goodieboy joined	13:20
	← catalis left	13:20
goodieboy	i made some changes to master, but i'd like to move those changes to a different branch before committing. is that possible?	13:21
	← robo47 left	13:21
goodieboy	in other words, i forgot to create a new local branch before making my changes	13:21
	← galderz left	13:22
Aides	goodieboy: just switch to master and commit there	13:22
	→ Mezner joined	13:22
Aides	err, i meant just create a new branch and commit there	13:22
goodieboy	ok cool so it's that simple :)	13:22
Aides	yep	13:23
	← chid left	13:23
	→ priidu joined	13:23
	→ Bass2 joined	13:24
	← Pewpewarrows left	13:25
	← jefferai left	13:25
	← bdowning left	13:26
	← Bass10 left	13:26
	← Tommy[D] left	13:26
	← novas0x2a left	13:26
	← DrNick left	13:27
	← chewbranca left	13:27
	← dcampano left	13:27
	← kingfishr left	13:27
	← Gitzilla left	13:27
	← Mezner left	13:27
locklace	freenode fail	13:28
sitaram	Ilari: pu-ssh-plus updated, new-fangled-keys now addable via "gl-setup" program. This is a new "server-side" install mode built for distro packagers; last section of doc/0 has all the details. You should use this method to install... avoids all easy install, ssh, troubles as far as I am able to test	13:28
	← Ramblurr left	13:28
	sitaram wonders if that got truncated; should end with "I am able to test"	13:28
locklace	you're good	13:28
sitaram	thanks :)	13:28
	← davidw left	13:28
	← comps left	13:28
	← tildeequals left	13:28
	→ hyperair joined	13:28
	← hyperair left	13:28
	→ hyperair joined	13:28
	→ tildeequals joined	13:29
locklace	your client should be smart enough to break long lines anyway...	13:29
	→ Pewpewarrows joined	13:29
	← hced left	13:29
	← TheUni left	13:29
	→ DrNick joined	13:29
	← pkw left	13:29
	← dbb_ left	13:29
	← jjuran left	13:29
	→ novas0x2a joined	13:29
	← kssddy left	13:30
	→ rduplain joined	13:30
	→ jjuran joined	13:30
	← AkumaStreak left	13:30
	→ hummingkid joined	13:30
	→ bdowning joined	13:30
	→ dbb_ joined	13:30
	← rburton- left	13:30
	← Bass2 left	13:30
	→ Bass10 joined	13:30
	← tchan left	13:30
locklace	ok, time to officially rename this network to "freefall"	13:30
	← goodieboy left	13:30
	→ TheUni joined	13:30
sitaram	hmm not sure about that; I seem to recall being bitten once...	13:30
	→ Ramblurr joined	13:30
	← Ramblurr left	13:30
	→ Ramblurr joined	13:30
	← schulty left	13:30
	→ Gitzilla joined	13:31
	→ dcampano joined	13:31
	→ pkw joined	13:31
	→ LiamH joined	13:31
locklace	sitaram: which client?	13:31
sitaram	irssi	13:31
	→ chewbranca joined	13:31
	→ comps joined	13:32
locklace	sitaram: you want this: http://scripts.irssi.org/scripts/splitlong.pl	13:32
	→ altrux joined	13:32
	→ AkumaStreak joined	13:32
sitaram	aah	13:32
	→ jefferai joined	13:32
	→ jsquared joined	13:32
	→ kingfishr joined	13:32
sitaram	locklace: thanks!	13:33
	→ EO_ joined	13:33
	→ Tommy[D] joined	13:34
	→ hced joined	13:34
	← hummingkid left	13:34
sitaram	locklace: worked fine...	13:36
locklace	cool	13:38
	→ robo47 joined	13:39
	→ karpar joined	13:39
	→ bauruine joined	13:42
	→ hobodave joined	13:42
	→ TeckniX joined	13:42
	→ jmcantrell joined	13:43
	→ Mezner joined	13:46
	→ spearce joined	13:46
	← arialdo left	13:47
	← karpar left	13:47
	→ tpolzin joined	13:47
	← comps left	13:49
	← dcampano left	13:50
	→ comps joined	13:50
	← frakturfreak left	13:50
	→ drubies joined	13:50
	→ frakturfreak joined	13:51
Scotto_	how would i push changes to a live version of a website?	13:51
	does that involve having a repo on the live server?	13:51
	any siggestions getting this part of my workflow figured out would be appreciated	13:52
	*suggestions	13:52
	← hobodave left	13:52
	← tg left	13:53
	→ tg joined	13:53
	← Axil left	13:53
charon	Scotto_: rsync, or a repo on the remote server with a post-receive hook that calls checkout -f or reset --hard (not pull)	13:55
	→ Plouj joined	13:56
Plouj	hi	13:56
	what does the arrow in "remotes/origin/HEAD -> origin/master" mean?	13:56
	→ drry joined	13:56
Ilari	Plouj: Presumably in this case symbolic reference (symbolic link).	13:57
Plouj	oh, sorry I mean in the output of git branch -v	13:57
	← bauruine left	13:57
Plouj	git branch -a	13:57
Ilari	Plouj: I guessed this is from remote branch display.	13:58
Plouj	ok	13:58
	so what is the meaning of this symbolic link?	13:58
	→ tbf joined	14:00
	→ d0ugal joined	14:00
	→ sagsousuke joined	14:00
charon	42, obviously	14:00
	Plouj: you can use origin/HEAD (and since it's HEAD, also just origin) to refer to origin/master	14:01
	so in a sense, it's the default branch of the remote repo	14:01
Plouj	so, nothing to do with my local master being a tracking branch of the remote master?	14:01
charon	no.	14:01
	← sagsousuke left	14:02
	→ sagsousuke joined	14:02
	→ icefox joined	14:02
	→ dcampano joined	14:03
	→ jasonn1234 joined	14:03
	← stikonas left	14:04
	→ Swimming_Bird joined	14:04
Scotto_	how do revert changes? i just botched a dos2unix and i need to NOT commit the changes	14:07
	→ ringlej joined	14:08
pklingem	I'm trying to copy a directory from a local branch into the current branch, should I use git fetch for this?	14:08
	→ Determinist joined	14:10
sylr	hi	14:10
	is it possible to remotely rename a branch with git protocol ?	14:10
krh	you can push the branch under its new name	14:11
	and then delete to old one	14:12
	→ smuf joined	14:12
	→ Vortex34 joined	14:12
sylr	krh, ok, thanks	14:12
	→ Ramon_ joined	14:12
	← fool_ left	14:13
Plouj	where is the meaning of a remote-tracking branch documented?	14:14
	← Haarg left	14:15
	→ fool_ joined	14:15
Ramon_	how to ignore already versioned files?	14:15
	← yevlempy left	14:16
	← jokaro left	14:16
	→ yevlempy joined	14:16
Plouj	oh in git-branch	14:17
	← loincloth left	14:17
wereHamster	Ramon_: add it to .gitignore and git rm --cached it	14:18
	→ bauruine joined	14:19
	← y10 left	14:20
	→ y10 joined	14:21
	→ tedoc2000 joined	14:22
Plouj	howcome git checkout -b remote/branch is different from git checkout --track remote/branch?	14:23
	← niko left	14:24
charon	Plouj: briefly put, because -b takes an argument but -t doesn't. the first creates a new branch with full name refs/heads/remote/branch starting at HEAD, while the second uses some DWIMmery to create a new branch 'branch' starting at (and tracking) remote/branch	14:24
	charon hopes he got the -t semantics right since he never uses that	14:25
shruggar	Plouj, I think in modern git, they're the same	14:25
	wait, no, I'm wrong	14:25
	yeah, what charon said, I think	14:25
Plouj	o_O	14:26
shruggar	git checkout --track remote/branch is more like (in modern git): git checkout -b branch remote/branch	14:26
charon	indeed, that's a nice way of putting it	14:26
	→ stamina_ joined	14:27
	← whitman left	14:27
	← tedoc2000 left	14:29
	→ aryounce joined	14:31
Plouj	I guess 1.6.5 is old	14:32
wereHamster	<v1.6.5>	14:32
Gitbot	[git v1.6.5]: http://tinyurl.com/yjeuqhw -- v1.6.5	14:32
wereHamster	about six months old	14:33
	→ MrGando joined	14:33
	→ Haarg joined	14:34
MrGando	Hello , I'm having problems with git , I'm getting "bash: git-upload-pack: command not found" when I try to clone, but I CAN execute git-upload-pack from the console, it's already in my Path, any idea ???	14:34
	← carolyn left	14:36
	→ carolyn joined	14:37
wereHamster	it's not in the standard path. When you log in bash executes .bashrc or .profile where you probably modify your PATH, when git creates an ssh tunnel these files are not loaded.	14:38
	→ kbingham joined	14:38
MrGando	wereHamster: I can actually type git-upload-pack in the console and it's executed	14:39
	so I think it's in the path...	14:39
wereHamster	try doing this: ssh user@host git-upload-pack	14:40
	← rseifert left	14:40
	→ icwiener joined	14:41
	→ hummingkid joined	14:42
	→ jelmer joined	14:43
MrGando	wereHamster: Can I specify the path to the git-upload-pack in the remote server ?	14:44
	→ tchan joined	14:44
wereHamster	MrGando: man git-push	14:45
Gitbot	MrGando: the git-push manpage can be found at http://git.or.cz/man/git-push	14:45
wereHamster	or man git-clone	14:46
Gitbot	wereHamster: the git-clone manpage can be found at http://git.or.cz/man/git-clone	14:46
wereHamster	yep, it's right there...	14:46
	→ altogether joined	14:47
	→ chaoflow joined	14:48
	← rchavik left	14:49
	← j416 left	14:49
	← viniciusluiz left	14:49
	← khelll left	14:50
	← tg left	14:50
	→ loincloth joined	14:51
	→ tg joined	14:51
	→ christophsturm joined	14:52
chaoflow	git svn dcommit fails for me for a local (file:///) svn repo: "No such file or directory: Can't create directory '/tmp/tmpziV9zb/repos/testpackage/db/transactions/6-6.txn': No such file or directory at /usr/lib/git-core/git-svn line 570"	14:54
	git-svn version 1.7.0.2 (svn 1.6.9)	14:54
	it seems the on-disk repo format changed	14:55
	→ tedoc2000 joined	14:55
	← ringlej left	14:55
	← tedoc2000 left	14:55
	← jelmer left	14:56
	→ tedoc2000 joined	14:56
	→ kukks joined	14:57
	← joeconyers left	14:57
	← tpolzin left	14:57
	→ byang joined	14:58
	→ schierbeck joined	14:59
	← carolyn left	14:59
	← jasonn1234 left	15:00
	← byang left	15:01
	← sde_antrix1 left	15:01
	→ carolyn joined	15:01
	← EricInBNE left	15:02
	← tedoc2000 left	15:03
	→ tedoc2000 joined	15:03
	→ jasonn1234 joined	15:05
Scotto_	anyone ever run into dos2unix changing flv files? or other binary files?	15:06
Plouj	is that a git command?	15:06
Scotto_	Plouj: no	15:07
	Plouj: im converting the line endings before tracking with git	15:08
Plouj	well, then I don't see why it wouldn't change whatever files you point it at	15:08
	maybe those flv files have line endings	15:08
	→ joeconyers joined	15:08
	agib\|away → agib	15:11
	→ viniciusluiz joined	15:11
	← jasonn1234 left	15:11
	→ level09_ joined	15:11
	← level09_ left	15:12
wereHamster	of course they have. There is about 1/256 chance that a byte in that file is a newline	15:12
	→ rgr joined	15:12
	← Tommy[D] left	15:12
	→ coderdad joined	15:13
Scotto_	its ashame that theres no tool to convert all non-binary file line endings	15:13
	← christophsturm left	15:13
charon	maybe once you can come up with an oracle that correctly tells binary from text files, someone will write such a tool	15:14
	→ christophsturm joined	15:14
Scotto_	charon: looks like someone wrote an improved dos2unix which includes a --skip-binary flag	15:14
	i havent tried it yet though	15:14
	→ Tommy[D] joined	15:15
charon	it'll still only be a heuristic	15:15
Scotto_	so i gues si need to just write something to focus on only the files extensions i know are non-binary	15:15
	→ tjgillies joined	15:16
charon	as a crude counterexample, the format that .torrent files are stored in is a key-value format where every byte is significant since there are byte counts all over the place. but they do not necessarily have to look like a binary file (a common heuristic is checking for 0 or <32 bytes). [they typically do have some of these bytes in the 'value' payload, but that's beside the point]	15:17
	← agib left	15:17
wereHamster	better to use file extensions to differentiate between binary and text (IMHO)	15:18
Scotto_	yeah thats what im goign to do	15:18
	i have a set group of extensions which i have no doubt about	15:18
wereHamster	maybe use file --mime-type	15:19
Scotto_	i was just goign to use -name '*.txt'	15:19
	→ xiong joined	15:21
Plouj	Scotto_: you have now advanced to the next level of Unix shell usage!	15:21
	← ph^_ left	15:21
xiong	I'm starting to wonder why git isn't part of the OS.	15:21
Plouj	git kernel module?	15:21
	← GarethAdams\|Work left	15:22
xiong	Why is version control considered a programmer's tool? What prevents me from putting a repo in, say, my mail folder?	15:23
wereHamster	UGFWIINI prevents you	15:23
	→ coldswitch joined	15:23
xiong	Or perhaps more salient, that folder full of business letters.	15:23
	→ hobodave joined	15:24
xiong	Is UGFWIINI a bad thing!?	15:24
	← hummingkid left	15:27
xiong	Yah, Plouj: maybe not a kernel module per se but a bag on nautilus. There could be little traffic signal dots next to each and every file and folder: checked in, not checked in. Something like that. And a button you click that brings up a little window you can scribble your commit message, then commit.	15:27
	← viniciusluiz left	15:28
	→ catalis joined	15:28
Plouj	:(	15:29
	→ j416 joined	15:29
	→ makx joined	15:29
makx	hello dudes	15:29
parasti	you mean a file manager plugin, not "part of the OS"	15:29
makx	how do i exclude a patch from git	15:29
	speaking of git://git.openvz.org/pub/linux-2.6.32-openvz	15:30
	i need the openvz patch minus the 2.6.32.10	15:30
	they didn't rebase on top of it	15:30
	but just merged it in	15:30
xiong	parasti, I don't 'mean' anything yet. I just have this sudden rush in my head.	15:30
wereHamster	makx: what is 'the 2.6.32.10' ?	15:30
	→ bcardarella_ joined	15:31
makx	stable upstream linux-2.6 release	15:31
	wereHamster:	15:31
	^^	15:31
xiong	One Ring to Rule Them All. But yeh, more realistically, a nautilus plugin might be very doable.	15:31
charon	xiong: i think there's some work on nautilus integration of git-cheetah	15:31
	not sure if it's usable though	15:31
wereHamster	makx: checkout whatever you want, merge in whatever else you want. Done	15:31
makx	so in this repo a git log 7f5e918e62cbc9ac27c2f47d3c3dd4b86f67ff0e..	15:31
	shows also the 2.6.32.10 commit	15:31
	wereHamster: what are you waffling?	15:32
wereHamster	I have no idea how the history of that repo looks like	15:32
	← bcardarella left	15:33
	bcardarella_ → bcardarella	15:33
wereHamster	if you just want to exclude one patch, git revert is what you want. However that doens't work on merge commits	15:33
xiong	charon, I'll have to look in on that, when I have a mo. The only tricky part of something like that is deciding which folders get repo'd. You don't want one big repo for your whole FS (I guess); you don't want one repo per folder (I guess).	15:33
biz	what's the best way to check for a clean working tree in a scripted fashion? Should I check the exit code of git status or is there something better suited?	15:33
charon	xiong: well, it's still "only" git integration in the sense that it lets you use git repos from your file manager. (as i should be, really)	15:34
makx	wereHamster: I'd need this whole patchset rebased on newer 2.6.32.10	15:34
xiong	Dunno. Maybe with the wide availability of terabyte drives, it's reasonable to have one giant repo at / and then have a rule such that when you create a repo lower down the tree, that folder is excluded by default from the parent repo.	15:34
	→ JEEB joined	15:34
	← sh1mmer left	15:35
makx	repo can be seen on http://git.openvz.org/?p=linux-2.6.32-openvz;a=summary	15:35
	← codeshepherd left	15:36
wereHamster	what are you refering to by 'this patchset'?	15:36
	→ xperience joined	15:38
	← xperience left	15:38
	← Plouj left	15:38
	← hipe left	15:39
	→ hipe joined	15:39
	→ Tuttle joined	15:41
Tuttle	Is there some quick subcommand the status of all files in the cwd or tree? I mean to indicate for each existing file whether it is tracked, modified, ignored, untracked etc. ?	15:42
wereHamster	git status or git ls-files (with various options)	15:42
	← gaveen left	15:43
	→ beni- joined	15:43
xiong	I'm stoked. I just made my entire development perl executable folder a repo. Now I can tool with it as much as I like and trust I can get back to something that works, no matter how I mess it up.	15:43
	Considering how many hours I spent building the thing, that's not a trivial win.	15:44
	← ixti left	15:44
	→ amerine joined	15:44
	← hipe left	15:44
	→ lightcap_ joined	15:45
Textmode	night all	15:45
	← Textmode left	15:46
Tuttle	wereHamster: i doubt git-status does what i need. git ls-files seemed to require numerous options to get near to what i expected	15:46
	→ anjen joined	15:47
	→ malumalu joined	15:47
wereHamster	create an alias which executes ls-files multiple times	15:47
Tuttle	hm, that's ugly. but thanks	15:47
	→ pkahle joined	15:47
	← lightcap_ left	15:48
	← ambuli left	15:49
	→ aniket joined	15:49
	→ dpino joined	15:49
	→ fool__ joined	15:50
	← fool__ left	15:50
	→ fool__ joined	15:50
	→ cloudhead joined	15:50
	← ryanakca left	15:50
	← struberg left	15:51
	deavid → _deavid_old	15:51
	→ ixti joined	15:51
fred	is there a nicer (for scripting purposes) alternative got "git status \| grep -q 'working directory clean'" ?	15:52
	*alternative to	15:52
	← fool_ left	15:53
	← zoniq left	15:53
	← malumalu left	15:53
	→ zoniq joined	15:54
	→ DavidKlein joined	15:54
	→ deavid joined	15:55
wereHamster	fred: look at the git-rebase script	15:55
	→ sh1mmer joined	15:55
	→ stringo0 joined	15:55
	→ mklappstuhl joined	15:56
	→ gaveen joined	15:56
	← Tuttle left	15:56
	← ncopa left	15:56
	← _deavid_old left	15:59
	← sh1mmer left	16:00
	← skoop left	16:00
	← auslander left	16:00
	← gaveen left	16:01
	→ linduxed joined	16:03
	→ auslander joined	16:03
	→ xperience joined	16:03
	← auslander left	16:03
	→ auslander joined	16:03
linduxed	i just did a "git svn fetch" and then "git svn rebase" whereafter i got "foo/bar/foo.bar: needs update"	16:04
	← xperience left	16:05
	← y10 left	16:05
	← sagsousuke left	16:05
linduxed	i thereafter did git update-index --refresh	16:05
	→ kmap_ joined	16:05
linduxed	that only resulted in the same message	16:05
	→ schulty joined	16:05
	← tjgillies left	16:05
	← schierbeck left	16:05
	→ webchick joined	16:05
wereHamster	'needs update' means that your working tree is dirty. Either commit or stash or throw the changes away	16:06
	→ defectiv joined	16:06
	← kmap left	16:06
	→ tamm0r joined	16:06
linduxed	is there a way to check what the diff between the fetched changes and the local changes are?	16:06
wereHamster	git diff <whatyoufetched>	16:06
	kmap_ → kmap	16:07
	← tamm0r left	16:07
linduxed	done	16:08
	thx	16:08
	← parigaudi left	16:08
	→ ph^_ joined	16:08
	← aryounce left	16:09
	→ sh1mmer joined	16:10
	← carl- left	16:12
	→ gmcinnes joined	16:12
gmcinnes	Hi all. Stupid question. When merging two branches on my machine and I get a conflict, which is "local" and which is "remote"? I always forget, and I seem to remember there's a trick to it.	16:13
aniket	can someone help me with this problem:	16:13
	http://gist.github.com/raw/334157/b321cd74520c67e48c683859ae8cea1547719f3c/gistfile1.txt	16:13
	→ WebDragon joined	16:13
aniket	Alternative Link: git://gist.github.com/334157.git	16:14
	← kmap left	16:14
	← beni- left	16:14
altogether	gmcinnes: http://www.gitready.com/advanced/2009/02/25/keep-either-file-in-merge-conflicts.html	16:14
	→ kmap joined	16:14
	← friskd left	16:14
altogether	gmcinnes: dunno if that's current, just googled	16:15
	→ Voziv_ joined	16:15
gmcinnes	altogether: ah ha! thanks. Thought I could only apply those merge strategies to the whole merge.	16:15
Voziv_	Hey i got a question about the exclude file..... if i ignore a path such as a /bin directory of a project, can I specifcally make a type of file like db3 files still stay included?	16:16
	→ hakunin joined	16:17
wereHamster	inside bin/.gitignore: * and !*.db3	16:17
steffkes	"An optional prefix ! which negates the pattern; any matching file excluded by a previous pattern will become included again. If a negated pattern matches, this will override lower precedence patterns sources. "	16:17
charon	aniket: not sure i understand your question right, but if it is "how do i copy history from A to B where neither is a clone of the other", simply use 'git remote add' to create a remote and then fetch from it	16:17
Voziv_	ah oki	16:18
	→ tjgillies joined	16:18
	← agile left	16:18
Voziv_	so will that work in my .git/info/exclude file as well?	16:18
	← tg left	16:18
	→ aryounce joined	16:18
altogether	gmcinnes: does that page answer your problem? on careful read, it seems to just let you choose to keep theirs or yours, and not help tell you which is which when looking at a file with a confilct	16:18
	← bauruine left	16:18
	→ tg joined	16:18
aniket	charon: ok i'll try that	16:18
	← aryounce left	16:19
steffkes	Voziv_, it's just the same .. add .gitignore and work there	16:19
	→ laj joined	16:19
wereHamster	Voziv_: if a directory is excluded git won't even descend into it, so bin/ and !bin/*.db3 won't work (at least that's how I remember it_)	16:19
	← ixti left	16:19
Voziv_	ok	16:19
wereHamster	maybe bin/* and !bin/*.db3 will work though	16:20
	→ malumalu joined	16:20
gmcinnes	altogether: yeah, it does. I think. But I'm still curious about the other problem. "git merge foo" with conflicts, then "git mergetool" gives me a LOCAL and REMOTE which one points to the branch foo?	16:21
	schulty → schulty\|away	16:21
	→ parigaudi joined	16:22
	← patthoyts left	16:22
	← coldswitch left	16:22
	→ johnw joined	16:22
	→ gaveen joined	16:22
altogether	my guess is foo would be remote, can you find out by experiment?	16:23
	gmcinnes: see, i don't use mergetool, i just look at the file myself	16:23
	← tbf left	16:24
	← joeconyers left	16:25
	← g0 left	16:25
	← tstclair left	16:26
	← webchick left	16:26
	schulty\|away → schulty	16:26
	→ SRabbelier joined	16:27
	← MrGando left	16:27
	schulty → schulty\|away	16:28
	schulty\|away → schulty	16:28
	schulty → schulty\|away	16:28
	schulty\|away → schulty	16:28
	← tjgillies left	16:29
	← nusse left	16:30
	→ kenichi_ joined	16:31
	→ joeconyers joined	16:31
	→ tbf joined	16:31
vinse	gmcinnes: i agree with altogether, i'm pretty sure foo would be remote	16:33
	← stringo0 left	16:33
	← linduxed left	16:36
	→ dj_tjerk joined	16:39
	→ schlorto1 joined	16:39
	← schlorto1 left	16:39
	→ ajpiano joined	16:39
	→ elmomalmo joined	16:39
	→ schnable joined	16:40
	← psoo left	16:40
schnable	what is the best way to branch a repo and have that branch become master in a new repo? I don't need the history of the old repo in the new one.	16:40
Voziv_	it works great now, thanks guys :)	16:41
	→ tjgillies joined	16:41
	← trivol left	16:42
	→ webchick joined	16:42
	→ balooga joined	16:43
	← giallu left	16:43
	← Voziv_ left	16:45
	← ntoll left	16:46
	← nicoulaj left	16:47
	← sh1mmer left	16:49
	→ s0enke joined	16:50
s0enke	hi	16:50
	is it possible to clone only specific branches or does clone imply a complete "checkout" of the remote repo?	16:50
	→ lhz joined	16:51
	← MetaCosm_ left	16:51
	→ richcollins joined	16:51
	← MetaCosm left	16:51
	→ bentob0x joined	16:52
	← zodiak left	16:52
Weiss	s0enke: at least as I understand it, a clone starts from wherever "HEAD" points and works down to the bottom of the tree, so you only get one branch	16:52
	← steffkes left	16:53
Weiss	schnable: similar answer to above, although your question isn't quite clear to me. make sure "HEAD" points at the new branch, then the next clone from there will have that as "master"	16:53
s0enke	Weiss: well the remote originally is an imported svn repo, so i'm not sure where head points to	16:54
Weiss	schnable: note that you don't have to call your "main" branch "master", but if you do want that, you can rename the branches	16:54
	← tbf left	16:54
Weiss	s0enke: "git symbolic-ref HEAD" will tell you	16:54
	http://git.bitwiz.org.uk/?p=libdrm.git;a=summary - example of a Git repo with no "master" branch and HEAD pointing to something else	16:55
	← maxthoursie left	16:55
schnable	Weiss: that will include history tho, I basically want to take a snapshot of my code and start a new repo from it; simple cp of the files is not cool cause I have submodules	16:55
Weiss	schnable: ah.. in that case, some magic with git filter-branch --parent-filter I think (that's what I've used (once) in the past to discard history)	16:56
schnable	im experimenting with a clone --depth 1	16:56
	← Ramon_ left	16:57
schnable	which i may not be able to push from, but maybe i can cp the output of that so I get .gitmodules but still in a state where i can do git submodule init	16:57
	Weiss: what is git filter? i dont have that command	16:57
Weiss	I've never used submodules before, so I'll defer that to someone with better understanding :)	16:58
badboy_	it's filter-branch, schnable	16:58
Weiss	schnable: git filter-branch	16:58
schnable	gotcha	16:58
	→ Theravadan joined	16:58
	→ jasonn1234 joined	16:59
balooga	Hi folk. I just set up an account at github, and have made my first successful commit. Unfortunately, I did 'touch README', then 'git add README', and then began editing the README file. So now README on github is empty. README on my local disk contains text. But I cannot find out how to commit this changes, because 'git diff' doesn't return any changes.	16:59
	→ hipe_ joined	16:59
Saur	balooga: What does git status say about the README file?	17:00
schnable	clone --depth 1, then setting new remote, and pushing to that seems to work	17:00
	← cloudhead left	17:00
	← eladmeidar left	17:01
	← gmcinnes left	17:01
	→ eladmeidar joined	17:01
	← tjgillies left	17:02
	← Mezner left	17:02
	← nanoo left	17:03
balooga	schnable: sorry, NEWS & LICENSE have this problem. 'got status' lists some untracked files. Nothing else.	17:03
	err, 'git status'	17:03
	← fmardini left	17:03
	→ crubb joined	17:04
crubb	hey guys	17:04
	just getting started on GIT, is there a svn:externals equivalent w/ git?	17:04
fred	git submodules (sort-of)	17:05
	← comps left	17:06
	← Pupeno left	17:06
crubb	cheers fred	17:06
	← webchick left	17:06
altogether	after i run git init, .git/config has ignorecase = true. i'm on a mac which i'd think should be case sensitive, so why does git want to ignore case?	17:07
	→ enherit joined	17:07
crubb	altogether: are you sure there won't ever be a windows checkout? ;)	17:08
altogether	no, there could always be that some time	17:08
	← tokkee left	17:09
	→ aresnick joined	17:09
	→ balooga1 joined	17:09
	→ tokkee joined	17:09
altogether	crubb: i see what you mean - i got to this question troubleshooting something - changed a file name from GemFile to Gemfile, but git didn't notice the change	17:11
	→ tjgillies joined	17:11
	← balooga left	17:11
parasti	altogether: doesn't OSX default to case insensitive file system (but the FS supports case sensitivity itself)?	17:11
cbreak	HFS+ preserves case	17:11
altogether	windows is a bad thing, or has that already been acknowledges hereabouts	17:11
	→ posix4e joined	17:11
cbreak	but it considers letters only differing in case the same	17:11
posix4e	So if i committed a whole bunch locally, is their a good way to role it into one big megacommit before pushing it ?	17:12
cbreak	so you can not have README and readme in the same folder, for example	17:12
	posix4e: why do you want to?	17:12
	→ comps joined	17:12
cbreak	to make it less easy to understand?	17:12
	→ daharon joined	17:13
posix4e	cbreak: our changlist is customer faxing	17:13
	*facing	17:13
altogether	posix4e: know about rebase?	17:13
posix4e	altogether: i haven't done it yet	17:13
	→ r11t joined	17:13
cbreak	so you want your customers to understand it less?	17:13
altogether	posix4e: it's what you want	17:13
posix4e	cbreak: you are being rude and unhelpful	17:13
cbreak	you can squash commits with git rebase -i	17:13
	but you are rude to your customers if you do that	17:14
	← posix4e left	17:14
cbreak	posix4e: it's not me who's rude here	17:14
	→ _nickel joined	17:14
cbreak	note that rebase -i rewrites history	17:14
altogether	he/she's gone nevermind	17:14
cbreak	so if you pushed it somewhere, it will conflict	17:14
altogether	was about to warn the same :)	17:15
cbreak	hope that he reads the man page then...	17:15
altogether	ja, and the definition of rude	17:15
cbreak	I think the best method is to create a merge commit with an overview of the changes that are merged	17:16
daharon	Hey guys, we have a bunch of repos that span two different departments at work. We are using gitosis to serve up the repos, but it's getting a bit messy. We have projects that are separated out into many repos for security reasons. Any tips on managing gitosis so that the repos can be partitioned out by department?	17:16
cbreak	and not squash everything together	17:16
	← balooga1 left	17:16
daharon	I thought about just creating a new user for gitosis, but we want to share all the keys	17:16
cbreak	share the keys?	17:17
altogether	parasti: as to your q, i can only guess - installed SL with all defaults, and don't know if FS actually supports it	17:17
cbreak	make all people have the same key?	17:17
daharon	cbreak, what I mean is, I don't want to create several keys for a single user.	17:17
	→ sh1mmer joined	17:17
	← r11t left	17:17
	← sh1mmer left	17:17
cbreak	you don't need to create any key	17:17
	every user should have or create his own	17:17
	→ r11t joined	17:17
	→ jstephenson joined	17:18
cbreak	I don't know gitosis itself, but in SSH, which I think it is based on	17:18
	a key identifies a user	17:18
	← bcardarella left	17:20
bremner_	daharon: the answer to almost all gitosis questions is: look at the gitolite docs, it might do what you want :)	17:20
	← priidu left	17:21
	→ carrus85 joined	17:21
	← enherit left	17:23
	← godsflaw left	17:23
altogether	i guess if my problem is 'change README to readme without messing with future windows checkouts or other scripts that just might care about case' then my solution is 'git rm README && git touch readme && git add readme'. i guess.	17:24
	→ coldswitch joined	17:25
cbreak	my tip: do not touch case.	17:25
	cbreak uses all lowercase for filenames in programs, with only few exceptions	17:26
altogether	cbreak: usually me too, just ran across a special case was troubleshooting	17:26
	→ MetaCosm joined	17:26
	← locklace left	17:27
	← _sh3 left	17:27
	→ WebDragon\|away joined	17:27
cbreak	Hmm... funny... it seems the exceptions are about 90% of my new project :/	17:27
	ah well, at least it's consistent :)	17:27
altogether	now you're being rude	17:27
	← WebDragon left	17:27
	WebDragon\|away → WebDragon	17:27
	→ locklace joined	17:27
altogether	don't wonder if i'm kidding. i am.	17:28
	→ _sh3 joined	17:28
	← dereine left	17:28
	← _nickel left	17:28
	→ aryounce joined	17:28
	→ enherit joined	17:28
	← gebi left	17:28
	← joeconyers left	17:29
	→ aryounce_ joined	17:29
	← tjgillies left	17:29
	→ jamiel joined	17:30
	→ viniciusluiz joined	17:30
jamiel	Hi all, I seem to be stuck in a merge conflict loop i can't get out of ... I made changes in a local branch with my master being a git svn repository	17:30
	I merged my changes into master but had a conflict, sorted it out, did a git svn fetch and git svn rebase and am now stick with:	17:31
	"Merge conflict during commit: File or directory 'www/trunk/lib/api' is out of date; try updating: The version resource does not correspond to the resource within the transaction. Either the requested version resource is out of date (needs to be updated), or the requested version resource is newer than the transaction root (restart the commit). at /usr/libexec/git-core/git-svn line 570"	17:31
	everytime i try dcommit	17:31
	← merlin83 left	17:31
jamiel	i've tried making a fresh branch, then reset back to the last svn-id commit	17:31
	and a rebase into my branch	17:31
	← hced left	17:32
	← aryounce left	17:32
	aryounce_ → aryounce	17:32
jamiel	I get the same conflicts, sort them out ... checkout master , then git svn fetch and tried git merge backup --ff .. it adds all the new files but a dcommit gives same error	17:32
	← Swimming_Bird left	17:32
	→ hced joined	17:33
	← tetha left	17:33
	← g0bl1n left	17:33
	→ jelly-bean joined	17:34
	→ zodiak joined	17:34
jelly-bean	can i make it so when i do git pull from any branch it attempts to pull for ALL branches, not just the one i am currently on?	17:34
	→ friskd joined	17:34
	→ joeconyers joined	17:34
	→ Mizar-jp joined	17:35
	← tvw left	17:35
	← d0ugal left	17:35
	→ tbf joined	17:35
	← r11t left	17:35
	→ tetha joined	17:35
	→ merlin83 joined	17:36
	← merlin83 left	17:36
	→ merlin83 joined	17:36
	→ r11t joined	17:37
	→ PointMan joined	17:38
	→ bradhaydon joined	17:38
	← gaveen left	17:39
	→ tjgillies joined	17:40
	→ kumbayo joined	17:40
	← JackWinter2 left	17:41
	← tbf left	17:41
	← r11t left	17:41
	→ cj joined	17:41
cj	hurm. I accidentally git rm'd something and committed it... how do I remove that file from that commit?	17:41
	→ r11t joined	17:41
	→ webchick joined	17:41
altogether	jelly-bean: i actually have not tried this yet, about to - but looks good for you http://www.gitready.com/advanced/2009/03/11/easily-manage-git-remote-branches.html	17:42
	jelly-bean: another possibility for you is hub	17:43
	← hakunin left	17:43
	→ JackWinter joined	17:43
	← PointMan left	17:43
	→ Pupeno joined	17:45
	→ dreiss joined	17:46
	→ gnufied joined	17:47
	← joeconyers left	17:48
	→ PointMan joined	17:48
	→ PointMan_ joined	17:49
	← PointMan left	17:49
	PointMan_ → PointMan	17:49
	← anjen left	17:49
	← PointMan left	17:50
aniket	does git1.5.3 have a command like "git checkout --ours xxx.txt" ?	17:50
	← petrux left	17:50
	← radarek left	17:51
	← catalis left	17:51
cbreak	what would that do?	17:51
	oh, index checkout	17:52
	if it doesn't, you can check out from the commits instead	17:52
	like git checkout HEAD -- file.txt	17:52
	→ hakunin joined	17:53
	→ xperience joined	17:53
	→ gaveen joined	17:53
altogether	aniket: if after conflicts, maybe this http://www.gitready.com/advanced/2009/02/25/keep-either-file-in-merge-conflicts.html	17:53
	← richcollins left	17:53
	→ joeconyers joined	17:54
aniket	altogether: ya i tried this command but it does not work in git v1.5.3 i think	17:54
	← Determinist left	17:54
altogether	aniket: but you saw what it says at bottom of page? about earlier gits?	17:54
	→ MetaCosm_ joined	17:55
jamiel	How in my current branch, how can I see the difference from a previous revision?	17:55
aniket	altogether: ok, thanks!	17:55
altogether	jamiel: git diff sha1-of-preivious-revision	17:55
jamiel	doesn't seem to work :( I get an empty file	17:55
	→ krawek joined	17:56
jamiel	Maybe I'm going too far back?	17:56
	→ jceb joined	17:56
altogether	jamiel: then they're identical? seriously, sure you're not on that commit?	17:56
	jamiel: no, there's no limit	17:56
jamiel	Even if it was before the branch was created?	17:56
	→ tstclair joined	17:57
	→ Lumbendil joined	17:57
	themill → themill_	17:57
Lumbendil	hi	17:57
	→ tbf joined	17:57
altogether	jamiel: ok, what did you mean by 'previous revision'?	17:57
Lumbendil	anyone arround here uses PortableGit?	17:58
	→ agib joined	17:58
	→ lresende joined	17:58
fr0sty	jamiel: what does 'git log --oneline <previous_commit>..HEAD' show?	17:59
	← viniciusluiz left	17:59
aniket	altogether: can you also tell what does this command do: git reset -- xxx.txt	17:59
	themill_ → themill	18:00
Lumbendil	I don't seem to manage it to get my public keys... :S	18:00
	← joeconyers left	18:01
	→ domcat joined	18:01
aniket	can someone tell what does this command do: git reset -- xxx.txt	18:01
	← krawek left	18:01
	→ hebz0rl joined	18:01
altogether	aniket: dont worry i'm sure everyone saw you ask me 3 seconds ago :) - was experimenting, it does depend on what you have in your index at the time	18:02
	← wolog left	18:02
	→ godsflaw joined	18:03
	← gaveen left	18:03
	→ cloudhead joined	18:04
charon	aniket, altogether: it doesn't really depend. it fills the index entry for xxx.txt with the contents of HEAD	18:04
	→ riz2 joined	18:04
charon	the effect is to unstage any changes to the file that were staged before	18:04
	← godsflaw left	18:04
kenichi_	hello. if i branchA from master, <commits>, then branchB from A, <commits>, merge B to A, merge A to master, then delete A before B, will i get "is not fully merged" when i try to delete B?	18:05
charon	(the slight twist is that if the file was newly added, you need git rm --cached instead of git-reset since there is nothing in HEAD to copy to the index)	18:05
	→ godsflaw joined	18:05
altogether	charon: ok ja, that's clearer, thanks	18:05
	← froschi left	18:06
	→ priidu joined	18:07
	← tatsuya left	18:08
Lumbendil	anyone knows how to use public key auth with PortableGit? it doesn't seem to find the key, or something like that :S	18:08
	← altrux left	18:08
	← ebneter left	18:09
	← tjgillies left	18:09
	→ ebneter joined	18:09
	← yevlempy left	18:09
	→ brizly joined	18:09
cbreak	git doesn't search the key	18:10
	→ yevlempy joined	18:10
cbreak	it's SSH who does	18:10
	if you have it at a non-standard location, just use ssh-add	18:10
	← tbf left	18:10
	← WebDragon left	18:11
	← TDJACR left	18:11
Lumbendil	i put it in the ssl folder :S	18:12
	well, seems i managed to read the file with openssl now	18:12
	let's see if I manage to get it working...	18:12
	← schnable left	18:12
	→ RdeWilde joined	18:12
cj	Lumbendil: http://wiki.debian.org/SSLkeys#PEMkeys.28SSLcertificates.29	18:13
	→ joeconyers joined	18:13
	→ snaka joined	18:14
	→ simNIX joined	18:14
	← przemoc left	18:15
	agib → agib\|away	18:15
	agib\|away → agib	18:15
	→ przemoc joined	18:16
	→ tjgillies joined	18:17
	← doobry left	18:17
Lumbendil	ssh-add doesn't connect to the agent :S	18:18
	← joeconyers left	18:18
cj	Lumbendil: apt-get install keychain ; eval `keychain --eval`	18:18
	→ TDJACR joined	18:18
	→ gaveen joined	18:19
cj	if you're not using debuntu or similar, you can grab keychain from...	18:19
	← spearce left	18:20
cj	http://agriffis.n01se.net/keychain/	18:20
Lumbendil	I'm on windows, trying to get PortableGit working XD	18:20
	agib → agib\|away	18:20
cj	Lumbendil: ah, are you using cygwin?	18:21
Lumbendil	nope, there is a binary on the web of a portable win compilation	18:21
cj	okay... does it come with an ssh-agent ?	18:21
Lumbendil	yup	18:21
	but it doesn't seem to work	18:21
cj	have you run it? the ssh-agent, that is?	18:22
Lumbendil	i've started it up with ssh-agent	18:22
	and then ssh-add doesn't find it	18:22
	:(	18:22
	agib\|away → agib	18:22
cj	Lumbendil: you'll need to get the environment variables that ssh-agent spits out exported	18:22
	← shruggar left	18:23
Lumbendil	i read that it exports them :S	18:23
cj	once ssh-agent is running, you can add the keys to the agent with 'ssh-add c:\path\to\id_rsa'	18:23
	oh, good. now do an ssh-add	18:23
Lumbendil	no agent found	18:23
	→ joeconyers joined	18:24
Lumbendil	though i can see the exports	18:24
	→ WebDragon joined	18:24
cj	so the environment variables aren't exported	18:24
Lumbendil	damned shell, i can't copy from it	18:24
cj	right-click, chose 'mark'	18:24
Lumbendil	output of ssh-agent:	18:24
cj	highlight	18:24
	hit enter	18:24
	it's now in your copy buffer	18:24
Lumbendil	already tried, ain't working	18:24
cj	nice. windows is a pita ;)	18:25
	→ alester joined	18:25
Lumbendil	SSH_AUTH_SOCK=/tmp/ssh-pEDFId2100/agent.2100; export SSH_AUTH_SOCK;	18:25
	and the same with the PID	18:25
	it's a "unix shell" inside a windows XD	18:25
	gonna try something	18:25
	i oficially hate the windows shell	18:26
cj	check %TEMP%\ssh-ppEDFId2100 ?	18:26
	← WebDragon left	18:26
cj	or %TEMP_DIR% or whatever windows calls it...	18:27
vinse	Lumbendil: i'm getting the same feeling about working in solaris -- it's exactly like a regular *nix machine, except slightly different in a million tiny undocumented ways!	18:27
	← joeconyers left	18:27
	← johnw left	18:28
vinse	like, hey, it seems like sed isn't working like i'm used to, what version is it ... oh it's a version that won't even tell you what version it is!	18:28
Lumbendil	xD	18:28
cj	vinse: prefix all of your commands with 'g' and you should be good ;)	18:29
	is there even a gsed?	18:29
Lumbendil	felt that way the day a windows machine with java 1.6 didn't want to run a .jar wich worked on another windows on the same university	18:29
	← coldswitch left	18:29
	← oliver3 left	18:29
Lumbendil	cj, the instance is there	18:30
	together with all the other ones	18:30
	so... export problem, right?	18:30
	:(	18:30
	→ spearce joined	18:31
	← zoniq left	18:32
	→ zoniq joined	18:32
Lumbendil	great, ssh-add worked	18:32
	← snaka left	18:32
Lumbendil	or at least, no error prompt	18:32
	→ richcollins joined	18:33
Lumbendil	yay, back to the ol' permission denied mate...	18:35
	→ calebr joined	18:35
	← webchick left	18:35
	→ froschi joined	18:36
Lumbendil	using github, if that gives any hint :S	18:36
	← tjgillies left	18:37
	→ joeconyers joined	18:37
	← pkw left	18:37
	→ pkw joined	18:38
	← pkw left	18:38
	← carolyn left	18:38
	← schlortoo left	18:39
	→ viniciusluiz joined	18:40
	→ carolyn joined	18:41
	← cilly left	18:41
Lumbendil	& I can fetch, so it ain't working for some other reason... :S	18:41
	→ schlortoo joined	18:41
	→ niko joined	18:42
	→ wolog joined	18:42
cj	Lumbendil: what's ssh-add -l say?	18:44
	← calebr left	18:44
cj	Lumbendil: also, does the other end recognize your ssh key?	18:44
	← viniciusluiz left	18:45
	← themill left	18:45
	→ lightcap joined	18:47
	→ webchick joined	18:47
Lumbendil	ssh-add -l didn't add anything	18:47
	tnx, gonna make it add the file	18:47
	it should, it does through eclipse	18:47
	← domcat left	18:47
	← joeconyers left	18:47
	← richcollins left	18:48
cj	ssh-add -l lists the files your agent knows about ;)	18:48
	→ psoo joined	18:49
	← tildeequals left	18:49
Lumbendil	yup, read about it in a link, though I didn't think about that	18:49
	→ viniciusluiz joined	18:49
Lumbendil	now i've to get it to automate all this...	18:49
	since colleges are waaay to lazy to do all this... -.-	18:49
	→ tjgillies joined	18:49
Lumbendil	btw, silly question, how do you change where github stores everything?	18:50
ToxicFrog	How do you mean?	18:50
Lumbendil	the folder it outputs everything you pull	18:51
	→ koltroll joined	18:51
	→ scode joined	18:51
	→ Axius joined	18:51
cj	git clone <remote repository uri> <local repository path>	18:51
Lumbendil	ook	18:52
	← trochala left	18:52
	→ bcardarella joined	18:53
scode	Ok. I have a fork on github that i've been working on. The only thing I've done is to make some commits to my master, and pull from upstream/master (where upstream is a remote pointing to the upstream repository). Now I find that my local master (and my origin/master) has diverged from upstream/master in ways other than my local changes, and I have no clue why. Is there some good way to figure out which commits are missing in one branch relative to anothe	18:53
	→ joeconyers joined	18:53
Ilari	scode: 'gitk master...upstream/master'?	18:55
	→ Mezner joined	18:55
	→ t0rc joined	18:56
	→ meanburrito920_ joined	18:57
scode	Ilari: I'll give it a shot (once I've built it with gitk support). But I was hoping to just generate a flat list of commits that are in one branch but not the other. Since I have done no surgery and done no complex merges, I would expect the difference to boil down to that.	18:57
	→ maglos_ca_ joined	18:58
	→ michael_scheibe joined	18:59
	← viniciusluiz left	18:59
	→ Alagar joined	18:59
	→ xris joined	18:59
	→ justin-george joined	19:00
	← tjgillies left	19:00
	← novas0x2a left	19:00
xris	I'm trying to figure out how to make --name-status a default for git log, but I can't seem to find the right set of things to use in ~/.gitconfig.	19:00
	← Axius left	19:00
	← michael_scheibe left	19:01
	hipe_ → hipe	19:01
scode	Ilari: That does show commits to upstream/master that are supposedly merged into my master, yet they aren't according to 'git diff' (and according to 'diff -uNr' on checkouts of the branches).	19:02
	← schulty left	19:02
	← christophsturm left	19:03
	← maglos_ca_ left	19:03
scode	(and 'git merge upstream/master' claim's i'm up to date)	19:03
	→ fool_ joined	19:04
	← fool_ left	19:04
	→ fool_ joined	19:04
	→ viniciusluiz joined	19:04
Ilari	scode: 'git rev-list --left-right master...upstream/master'?	19:06
	→ maglos_ca joined	19:06
	→ viniciusluiz1 joined	19:06
	← fool__ left	19:06
	→ ruz joined	19:06
	← viniciusluiz left	19:07
ruz	merged a branch, now can not find way to see the diff of this merge :(	19:07
	→ tjgillies joined	19:07
scode	Ilari: Empty output for "master..upstream/master"; I see my own commits when I do it for "upstream/master..master".	19:08
	.... and some commits by others apparantly.	19:09
Lumbendil	anyone knows of something like "eval" in windows console? i know it doesn't belong here... :S	19:09
	→ domcat joined	19:10
Lumbendil	but I'm trying to make a .bat file for a portable git to set itself up and i'd need that	19:10
	← joeconyers left	19:10
makx	grr wereHamster the answer was really simple	19:10
	git diff v2.6.32.10...HEAD	19:10
	now what the hack three dots are special	19:11
	← webchick left	19:11
	← viniciusluiz1 left	19:12
	→ viniciusluiz joined	19:12
	→ krawek joined	19:13
	→ themill joined	19:14
	→ tildeequals joined	19:14
	→ kursad joined	19:14
scode	Ilari: Ok. "git checkout upstream/master ; git fetch ; git checkout master" - now I"m back to expected state. I have no idea how I broke it though (and 'git fetch' produced no output btw, so I'm surprised it made any difference at all).	19:15
	← viniciusluiz left	19:17
	→ joeconyers joined	19:17
	← hakunin left	19:17
	← krawek left	19:18
	→ Blaine joined	19:18
	→ hakunin joined	19:18
	→ viniciusluiz joined	19:18
	→ krawek joined	19:20
	→ sagsousuke joined	19:21
	→ radarek joined	19:21
	→ naeu joined	19:23
	← Lumbendil left	19:23
	← Tureba left	19:24
	→ jar_corefile joined	19:26
	← krawek left	19:26
	→ pboucher joined	19:27
	← codeswing left	19:28
pboucher	I was wondering if this would be a good place to ask for help on designing a git setup?	19:28
	→ thiago_home joined	19:29
	← Pupeno left	19:30
	→ WebDragon joined	19:31
locklace	pboucher: state your case	19:31
pboucher	I've got an app with which I work that has a plugin API. I write a lot of plugins against this API which are self contained tools.	19:31
	← Wicked left	19:31
pboucher	I also download third party tools for this API.	19:31
	→ krawek joined	19:32
pboucher	I would like to have a repo that represents all my tools that I can have hooks that make it so when anything in master changes, my changes are copied out to my production area.	19:32
	But if I only have one repo with a personal dev branch and a master branch that automatically copies to production space with a hook and all my plugins are gobbed in this one repo, how do I version individual plugins?	19:33
	→ Axius joined	19:34
pboucher	...	19:34
	Would it be better to have one repo per plugin?	19:34
	But then, how to make it so that modifications in the master branch are pushed out to my production environment? I'd need to setup hooks in all repos.	19:34
	← jspiros left	19:35
	← vinse left	19:35
	→ jspiros joined	19:35
pboucher	The workflow I'd like is that each dev has their personal branch space and there's the master branch.	19:35
	→ vinse joined	19:36
pboucher	Any commits that are added on top of the master branch will be deployed to the production environment.	19:36
	→ Wicked joined	19:36
pboucher	Being rather new to git I'm a bit stumped as to how to set this up.	19:36
	← jdmason left	19:36
	→ nanoo joined	19:37
icefox	:) /me submits my first patch to git	19:37
	← vvsh left	19:37
	← krawek left	19:37
	→ krawek joined	19:37
locklace	pboucher: maybe have a look at http://nvie.com/git-model	19:37
	← hakunin left	19:38
	← carolyn left	19:38
	← DrNick left	19:39
	→ hakunin joined	19:39
fr0sty	pboucher: also, consider git submodules. have each plugin be a submodule in a superproject.	19:40
	← hakunin left	19:40
	→ hakunin joined	19:40
fr0sty	when you have changes to your plugin, you would update the submodule, commit to the superproject and have that pushed to your production area (or pulled from your production area if you prefer).	19:40
	I can, unfortunately, give you little help beyond that suggestion.	19:41
pboucher	Can one commit to submodules "through" a super project? That way I could administer the multitude of small per plugin repos but any dev would only need to deal with the super repo?	19:42
	→ DrNick joined	19:42
	← krawek left	19:42
pboucher	locklace: Thanks for the link. Interresting reading for sure!	19:42
	Don't know how I'd have one repo with all the the plugins having all those branches but I'll investigate.	19:43
locklace	the chapters on submodules and subtree merge in http://progit.org/book/ (not to mention all the others) are worth reading too	19:43
	← mm_work left	19:44
	→ jdmason joined	19:44
	→ Cryp71c joined	19:48
	← aniket left	19:49
	→ rm-f`whoami` joined	19:51
	→ archis joined	19:52
	→ project2501a joined	19:52
	← t0rc left	19:53
	← radarek left	19:54
	→ richcollins joined	19:55
jamiel	Hi guys, I have a clean git svn repo ... I have just done a branch called 'dev' ... now, I am plagued with merge conflicts because we are a team of 5 working on the same code... what is the best practice here ... development in my dev branch, just before I want to commit, do a git svn fetch ... sort out any conflicts ... do a git svn dcommit ... then how do I get that it into my master? checkout master and git svn fetch?	19:55
	← archis left	19:55
	→ radarek joined	19:56
cbreak	if you work on the same code, you will get conflicts	19:57
thiago_home	before dcommitting, you must always rebase	19:57
	then fix any conflicts	19:57
	→ _nickel joined	19:59
	← zoniq left	20:00
	→ patthoyts joined	20:01
	→ kipras joined	20:01
	→ tbf joined	20:01
	← altogether left	20:01
	← spearce left	20:02
	→ webchick joined	20:03
	← Cryp71c left	20:03
	→ spearce joined	20:03
	← maglos_ca left	20:03
	← Axius left	20:05
	← webchick left	20:06
	→ Twisted_Mentat joined	20:07
	← meanburrito920_ left	20:08
	agib → agib\|away	20:08
	→ sh1mmer joined	20:08
	→ Mage joined	20:08
	← Mage left	20:08
	→ Mage joined	20:08
	→ pantsman joined	20:10
	→ meanburrito920_ joined	20:10
	← lightcap left	20:11
	← siprbaum left	20:12
	→ rudi_s joined	20:13
	← Mezner left	20:17
	← rudi_s_ left	20:17
	← ToxicFrog left	20:17
	→ Mezner joined	20:17
	→ webchick joined	20:18
	← parigaudi left	20:18
	→ siprbaum joined	20:19
	← siprbaum left	20:19
	→ siprbaum joined	20:19
	← viniciusluiz left	20:22
	→ Laibsch joined	20:22
	→ qubodup joined	20:23
qubodup	hello	20:23
	I'm a non-developer, who uses git to pull the latest revisions of some softwares	20:23
	often, git complains about files being different. for example 'configure'	20:23
Mage	it will indeed do that	20:23
badboy_	`git checkout -- file` first	20:23
qubodup	I want git to just pull and take everything from HEAD. How would I do that? or is what I just said senseless?	20:24
Laibsch	I am now on lucid. Did interactive git rebase change? I believe that when I stopped to edit a commit, I would be presented with an unclean working directory, ready to commit. Now, "git status" turns up empty. Or is this a problem in lucid?	20:24
Bombe	qubodup, git fetch && git reset --hard origin/master	20:24
	← patthoyts left	20:24
	← spearce left	20:24
	← richcollins left	20:25
	← wolog left	20:25
qubodup	Bombe: thx	20:25
	→ seanmccann joined	20:25
qubodup	badboy_: you too, it worked for now, but Bome's way seems simple	20:25
badboy_	yep	20:26
	→ wolog joined	20:26
	→ Axius joined	20:26
	← Alagar left	20:26
	→ spearce joined	20:26
	→ wwwald joined	20:26
	→ Ilari_ joined	20:28
	→ dotsintacks joined	20:29
	→ ToxicFrog joined	20:29
dotsintacks	is there an option for git status to show me ignored files (i.e., ignore .gitignore temporarily)?	20:30
Mage	mv .gitignore .gitfoobar && git status	20:30
	← Ilari left	20:30
	Ilari_ → Ilari	20:30
	← tobier left	20:30
Rhonda	dotsintacks: set core.excludefile to something different.	20:31
	→ t0rc joined	20:31
	← froschi left	20:31
	← Axius left	20:31
Rhonda	… not sure if it's possible to set config options on the commandline?	20:31
	→ krawek joined	20:31
badboy_	git config core.excludefile ...	20:32
	or add --global if this should be added to ~/.gitconfig	20:32
Rhonda	badboy_: That's not for just a single call, that's storing it.	20:32
	← joeconyers left	20:32
badboy_	ah, thought you wanted to store it =D	20:32
Rhonda	The question rather was to temporarily set it. :)	20:32
	→ tobier joined	20:32
badboy_	ok.	20:33
	→ mjf joined	20:33
	← daharon left	20:34
	→ coldswitch joined	20:34
	← pantsman left	20:34
	→ jrmuizel joined	20:35
	← yhager left	20:35
	→ yhager joined	20:35
	→ pantsman joined	20:36
	← sh1mmer left	20:38
	yevlempy → yev\|away	20:38
	← jamiel left	20:39
	→ sh1mmer joined	20:41
	← sh1mmer left	20:42
	← ajpiano left	20:43
	← ehabkost left	20:43
	← pklingem left	20:43
	← DrNick left	20:43
	→ ehabkost joined	20:45
	← xperience left	20:46
	← yev\|away left	20:47
	← sagsousuke left	20:47
	← d0k left	20:47
	← pboucher left	20:48
	← enherit left	20:50
	→ __iron joined	20:51
	← RdeWilde left	20:51
	← webchick left	20:52
	← auslander left	20:53
	← _iron left	20:54
	← amerine left	20:55
	→ tvw joined	20:55
	← jar_corefile left	20:55
	← Mezner left	20:56
	← dotsintacks left	20:56
	← Bass10 left	20:57
xris	I'm trying to figure out how to make --name-status a default for git log, but I can't seem to find the right set of things to use in ~/.gitconfig. Does anyone know if this is even possible? (or maybe have a link to some documentation)	20:58
	← cannonball left	20:59
	→ webchick joined	20:59
	← jdmason left	20:59
	→ giorgian joined	20:59
	← webchick left	20:59
	→ WebDragon\|away joined	21:00
	→ webchick joined	21:00
wereHamster	xris: to see which various config options exist, man git-config	21:00
Gitbot	xris: the git-config manpage can be found at http://git.or.cz/man/git-config	21:00
	→ yev\|away joined	21:01
	← priidu left	21:01
	→ priidu joined	21:01
	← josh0x0 left	21:01
	→ stkrzysiak joined	21:01
	→ nanooo joined	21:02
	← drubies left	21:02
	← matthieu_ left	21:02
Laibsch	let's assume I have a linear history a-b-c. How do I create a branch with a-c from that? (IOW, it should not contain a-b).	21:03
	← WebDragon left	21:03
	WebDragon\|away → WebDragon	21:04
Bombe	Laibsch, git checkout -b new-a a; git cherry-pick c	21:04
Laibsch	c is a range of commits	21:04
	not just one single commit	21:04
	sorry for being unclear	21:05
	→ jdmason joined	21:05
Bombe	Then maybe you want to take a look at man git-rebase	21:05
Gitbot	Bombe: the git-rebase manpage can be found at http://git.or.cz/man/git-rebase	21:05
	← nanoo left	21:05
xris	wereHamster: yeah, couldn't see much in there for this specific issue (unless --name-status has some other name in config)	21:05
Bombe	Laibsch, especially the section about interactive rebasing.	21:05
wereHamster	Laibsch: do you want to remove one commit from a list of commits?	21:05
giorgian	hi. where can I find the return values of git commands?	21:05
Laibsch	I've done interactive rebasing before, but right now I don't see how that would apply.	21:05
przemoc	giorgian: $?	21:06
Bombe	Laibsch, just remove commit(s) b from the editor shown when rebasing interactive.	21:06
Laibsch	wereHamster: depending on the perspective, I want to add the b-c range of commits to rev A or I want to strip the a-c range from HEAD.	21:06
	← lucsky left	21:06
giorgian	przemoc: I mean, the possible values	21:07
wereHamster	Laibsch: you should be able to do that either with interactive rebase like Bombe said, or by using rebase --onto ..	21:07
Laibsch	Bombe: I see what you mean. But that's not so convenient. a-b is about 700 commits.	21:07
	wereHamster: I'll have a look into --onto	21:07
	thanks	21:08
giorgian	git commit, with nothing to be added, returns 1; do other errors also return 1, or different codes?	21:08
przemoc	giorgian: source code can say more, but generally all you want to know is whether $? == 0, i.e. success	21:08
giorgian	przemoc: no	21:08
	I'll look the source code, ok, but success is not enough: I don't consider nothing to be added so much of a trouble...	21:09
	← ehabkost left	21:09
	← tvw left	21:09
	← brizly left	21:10
	← project2501a left	21:10
giorgian	(I've put /etc under git, with an automatic script for apt, so that apt commands do git add . ; git commit, but when nothing changed I get an error, which makes apt fail...)	21:10
	→ project2501a joined	21:10
	← mastro left	21:10
przemoc	giorgian: you can use etckeeper	21:11
	BTW	21:11
giorgian	przemoc: ok, I'll try it	21:12
	thanks	21:12
	→ a_GWT_startup joined	21:12
Laibsch	--onto was just what I was hoping for. thanks	21:12
	I would have never guessed	21:13
	← malumalu left	21:14
	← webchick left	21:15
	→ slashbeast joined	21:18
slashbeast	Hi there!	21:18
	→ d0ugal joined	21:18
slashbeast	I just wonder, sometime when I clone remote repo, I can type 'git branch' and I see only master. but I know that there is also devel branch, git checkout devel and work	21:19
	why it not showing?	21:19
	→ xiong_ joined	21:19
	agib\|away → agib	21:19
	← xiong left	21:20
	xiong_ → xiong	21:20
wereHamster	git branch -a	21:20
	→ lightcap joined	21:21
	→ DrNick joined	21:22
	← rduplain left	21:23
	← domcat left	21:24
	← corni left	21:24
	← t0rc left	21:24
	← lightcap left	21:27
slashbeast	wereHamster: thanks	21:27
	;-)	21:27
	→ webchick joined	21:27
	← slashbeast left	21:27
	← zodiak left	21:27
	← _nickel left	21:28
	← bcardarella left	21:29
	← Blaine left	21:30
	→ victoryoalli joined	21:31
	Mike\|\|gone → mikelifeguard	21:33
	← victoryoalli left	21:34
	← cloudhead left	21:34
	← coldswitch left	21:35
	→ Pupeno joined	21:36
	← LiamH left	21:38
	← hugoxrosa left	21:39
	→ idamon joined	21:41
	← kumbayo left	21:42
	← nanooo left	21:43
	← WebDragon left	21:43
	← Theravadan left	21:44
eli	Is there a way to see an annotated commit without the referenced object?	21:44
	→ zodiak joined	21:45
eli	(which `git show' seems to always do)	21:45
	→ Theravadan joined	21:45
	← meanburrito920_ left	21:45
	→ webchickenator joined	21:45
hachi	ugh... git server doesn't allow ~ in the url when you have user dirs turned off	21:46
eli	eli: `git cat-file -p'	21:47
	eli: thanks	21:47
	→ trochala joined	21:47
	→ digitalice joined	21:47
	← jmcantrell left	21:47
digitalice	hey	21:48
	error: Untracked working tree file 'portal/media/css/amarillo' would be overwritten by merge.	21:48
	what should i do?	21:48
	← kenichi_ left	21:49
	→ flaguy48 joined	21:49
jast	digitalice: if you don't have any changes of your own in that file, just delete it and try again	21:49
	← webchick left	21:49
digitalice	ok :D	21:49
	← psoo left	21:49
	← Ilari left	21:50
	→ Ilari joined	21:50
digitalice	jast: its full of that kind of files	21:51
	it just a testing version	21:51
	← dpino left	21:51
digitalice	[root@linode1 portal]# git checkout	21:51
	Your branch is behind the tracked remote branch 'origin/master' by 35 commits,	21:51
	and can be fast-forwarded.	21:51
	← hooopy left	21:51
	→ hugoxrosa joined	21:51
digitalice	how do i do the checkout?	21:51
	← icwiener left	21:51
	← textarcana left	21:53
	→ Giles joined	21:53
	→ _nickel joined	21:53
digitalice	jast?	21:54
	→ eletuchy joined	21:54
Giles	on Everyday GIT http://www.kernel.org/pub/software/scm/git/docs/everyday.html it talks about Repository Administration	21:54
	← BlindOracle left	21:55
Giles	and in the section cvs-style share it we are told to put the repo dir in /home	21:55
	in this style would you do git clone user@hostname:../devo.git	21:56
	or what?	21:56
	← digitalice left	21:56
	→ lightcap joined	21:56
	← tstclair left	21:57
	→ kumbayo joined	21:57
	→ none joined	21:58
	→ posix4e joined	21:58
Giles	or can it only be used in the way with git-deamon?	21:59
hachi	when you rev-parse a tree (not a revision) is that the content of the tree, or just the structure of the tree?	21:59
Ilari	Giles: git-daemon has nothing to do with user@hostname URLs...	22:00
Giles	indeed I gave a ssh URL	22:00
	→ jamiel joined	22:00
	← bluenovember left	22:00
	← levifig left	22:01
	→ t0rc joined	22:01
	← gaveen left	22:01
	← jdmason left	22:01
Giles	Ilari: what I am asking about is the setup that is explained in http://www.kernel.org/pub/software/scm/git/docs/everyday.html	22:02
	← jamiel left	22:02
	→ jdmason joined	22:03
	← bradhaydon left	22:04
Giles	Ilari: its seems to suggest you put repos under /home e.g. /home/devo.git	22:04
	→ josh0x0 joined	22:04
	← loincloth left	22:05
	→ webchick__ joined	22:05
	← _nickel left	22:05
	← werdan7 left	22:05
	← aziz left	22:05
	← trochala left	22:06
ndim	Giles: I would use "git clone /home/devo.git" in that case.	22:06
	← jceb left	22:06
Giles	if I don't want to setup a 'Git user' as you would when using stuff like gitolite or gitosis but I want to use a git group	22:07
	→ levifig joined	22:07
Giles	should I put my repos in /pub/scm	22:07
	and then clone them using git clone user@hostname:/pub/scm/project.git	22:08
	→ jonmason joined	22:09
	← webchickenator left	22:09
	← jdmason left	22:09
	→ sixteneighty joined	22:09
	→ gebi joined	22:10
	← qubodup left	22:11
	← hugoxrosa left	22:11
ndim	Giles: For ssh access, user@host:/path/to/repo.git, yes, or if you're working locally anyway, /path/to/repo.git.	22:11
	← xris left	22:12
	← t0rc left	22:12
	→ Tureba joined	22:12
ndim	Giles: And the path where you place the repos is your choice as sysadmin.	22:12
Giles	ndim: okay, I was just asking if there was any standard conention?	22:12
	s/conention/convention/	22:12
	→ werdan7 joined	22:13
	→ sgh_ joined	22:13
ndim	I'd say the standard convention is to keep stuff short. And using a single common special "git" user has the advantage that you can just use relative pathes in user@host:path-to/repo.git and that will then probably be resolved relative to wahtever the git user's $HOME will be.	22:14
	← levifig left	22:14
ndim	So if you want to use a shared repo with different users... I would definitively avoid /home/repo.git. /home/scm/repo.git or /git/repo.git or something similar would probably make most sense.	22:16
	→ amerine joined	22:16
Giles	ndim, yeah but prehaps you could create /home/giles/ and /home/fred/ and symlink them to /pub/scm	22:16
Ilari	Giles: Also if you are looking for multi-user setup, also look into Gitolite.	22:16
	→ bibinou joined	22:16
Giles	Ilari: it seems to not make sence to me, run git as a single user in a multi user setup?!	22:17
	why not create multiple users accounts on the machine.	22:17
	← Chillance left	22:17
ndim	Well, for some reason it makes sense for github to "run git as a single user in a multi user setup" with probably a few 100,000 users...	22:18
Giles	rather than a single user that has its shell set to a wrapper script that then works out which user you are.	22:18
ndim	...so it can't be all that bad. :)	22:18
Giles	ndim: yeah I don't really know the implications of 100,000 user accounts on a unix box	22:19
	→ Chillance joined	22:19
Giles	I can see that gitolite scales better	22:19
	but in my setup I just want 1 or maybe 2 users to access one repo.	22:20
	← simNIX left	22:20
ndim	Giles: All your users will eventually end up doing IO on the same repo. At which point you are merging the users's IO into a single IO stream does not really matter.	22:20
	→ gsgsgs joined	22:20
Ilari	Giles: Gitolite can do both MAC (mandatory access control) / DAC (discretory access control) type setups.	22:20
Giles	I don't know what that means?	22:20
	← gsgsgs left	22:21
ndim	Giles: For a 1 or 2 user setup, just use /home/giles/git/meh.git, and if you actually change the location afterwards, for your 1 or 2 users it is not a problem to change their "git remote"s afterwards.	22:21
Ilari	Giles: Either admin can control what repos exist and what permissions the have, or let users create them & give access to others.	22:21
	← naeu left	22:21
Giles	okay thanks	22:22
	I am off to sleep	22:22
	← Giles left	22:22
	← kumbayo left	22:25
	← xiong left	22:26
	← hced left	22:27
	← TeckniX left	22:29
	← segher left	22:29
	→ maxb joined	22:29
	→ seezer joined	22:30
	← yev\|away left	22:30
	← dcampano left	22:30
	← webchick__ left	22:31
	→ joevano joined	22:32
	← fda314925 left	22:32
	← Will\| left	22:32
	→ fda314925 joined	22:33
	→ Will\|\| joined	22:33
	→ binjured joined	22:33
	← Theravadan left	22:34
	← lhz left	22:34
	→ segher joined	22:34
	→ textarcana joined	22:35
	→ doobry joined	22:35
	← _Limb left	22:35
	← avtobiff left	22:36
	→ Limb joined	22:36
	← imajes left	22:36
	→ imajes joined	22:36
	→ avtobiff joined	22:37
	← jelly-bean left	22:37
	← GodEater left	22:37
	→ numbah joined	22:37
	← posix4e left	22:37
	→ posix4e joined	22:37
	← jY left	22:37
	→ eco joined	22:38
eco	I have some uncommited changes but I'd like to switch over to a different branch. Do I stash these changes or is this something git does automatically?	22:39
_rane	you need to stash or commit	22:39
	← DavidKlein left	22:39
eco	Alright. Thanks.	22:39
_rane	git stash	22:39
	→ jY joined	22:39
	→ t0rc joined	22:39
_rane	guess you knew that	22:39
	← doobry left	22:40
drizzd	eco: you can use git checkout -m to merge the changes over to another branch, without stashing them first	22:41
	← wwwald left	22:41
drizzd	also means you may not be able to reproduce the original changes if something goes wrong	22:41
eco	Interesting. Not quite what I need but good to know it's there.	22:41
	← chuck left	22:42
drizzd	eco: trying to change branches without -m is safe, but may fail	22:42
	→ chuck joined	22:42
	← David-T left	22:42
	← robo47 left	22:42
drizzd	(if the merge is non-trivial)	22:42
	← dwmw2_gone left	22:43
hachi	why does filter-branch complain if my working tree is dirty?	22:43
	→ David-T joined	22:43
drizzd	hachi: probably because it can change the index and therefore potentially invalidates your changes?	22:44
	← mklappstuhl left	22:44
hachi	why does it change the index, I thought it operated on revisions	22:44
	and not my working tree	22:44
drizzd	if it didn't change the index, you would have to manually fix it up	22:45
hachi	filter-branch?	22:45
	it never changes my HEAD or index	22:45
	→ mklappstuhl joined	22:45
	→ dwmw2_gone joined	22:45
hachi	well, it could change my HEAD if I'm doing a filter-branch on the same branch that I have checked out	22:45
drizzd	I'm pretty sure it does	22:45
	← t0rc left	22:46
drizzd	exactly	22:46
hachi	the operation I'm doing right now will not change my index or HEAD at all	22:46
drizzd	you think	22:46
	← imarcusthis left	22:46
hachi	I'm doing a filter-branch on a different branch than the one I'm on	22:46
ndim	filter-branch like rebase needs a working dir to do its work in.	22:46
	← sixteneighty left	22:46
	→ imarcusthis joined	22:47
drizzd	ndim: I don't think it uses the normal working dir though	22:47
	← radarek left	22:47
	← alester left	22:47
	← JackWinter left	22:47
	← tychoish left	22:47
	→ JackWinter joined	22:47
	→ tychoish joined	22:47
ndim	drizzd: That would be curious, but could be true.	22:47
drizzd	--tree-filter uses .git-rewrite/t, for example	22:47
	← jrmuizel left	22:48
	← hakunin left	22:48
drizzd	ndim: it would potentially remove untracked files otherwise	22:48
ndim	Good point.	22:48
	← hebz0rl left	22:49
drizzd	hachi: in git, branches are not strictly disjoint	22:49
	← dwmw2_gone left	22:49
	→ magn3ts joined	22:49
drizzd	just because you are a different branch does not mean you do not share any history, in particular	22:49
	HEAD may point to a commit that is also in a different branch	22:49
	→ hebz0rl joined	22:49
drizzd	so your argument is flawed	22:49
	← wolog left	22:49
	← meuh left	22:49
	← kipras left	22:51
drizzd	besides, what's the big deal? that's what git stash is for	22:51
mugwump	hachi: yeah if you're not using --tree-filter that's a bit of a bug really	22:51
	← karihre left	22:51
	→ dwmw2_gone joined	22:51
parasti	drizzd: eh. if you rewrite history on one branch, other branches referencing the original commits won't be affected...	22:51
	→ meuh joined	22:51
	→ karihre joined	22:51
	← up_the_irons left	22:52
	← tbf left	22:52
drizzd	parasti: you got me there	22:52
	→ meanburrito920_ joined	22:52
	← radhermit left	22:53
	← coderdad left	22:53
	→ radhermit joined	22:54
	← lightcap left	22:54
	← mklappstuhl left	22:54
hachi	drizzd: my big point would also be that I was not in any branch	22:55
drizzd	mugwump: or --index-filter	22:55
hachi	I was on 'initial commit', sorta like being in a bare repo, except I hadn't set bare = yes yet	22:55
	← iggy left	22:55
	→ madewokherd joined	22:55
	← ProperNoun left	22:56
	→ iggy joined	22:56
hachi	and the graph could be completely disconnected between two branches, like I have here :)	22:56
	→ dcampano joined	22:56
drizzd	hachi: sure, parasti already invalidated my point	22:56
	nevertheless	22:56
	it seems kind of pointless	22:56
	← aryounce left	22:56
cbreak	totally disconnected branches seems weird	22:56
	why are they in the same repository?	22:56
drizzd	especially since in your case, the solution is simply to set bare = yes, no?	22:56
hachi	because I'm preparing to graft it together	22:57
cbreak	yuck	22:57
	← meuh left	22:57
hachi	the filter-branch is cjust a couple steps before I start grafting	22:57
cbreak	unclean! impure!	22:57
hachi	mugwump caused it all... his work on the perl5 history inspired me to do this insanity at work	22:57
	→ meuh joined	22:57
yacoob	cbreak, the word you're looking for is 'infidel' 8)	22:57
	→ yev\|away joined	22:58
cbreak	exactly! thanks	22:58
hachi	YOU WILL NEVER KEEP ME DOWN MAN :D	22:58
	I will rise up against the man!	22:58
yacoob	standard tar+feather rules apply.	22:58
cbreak	I guess you can Hide secret documents from the evil officials in a different branch	22:58
	yacoob has one repo with 4 different related, but never going to be merged branches	22:58
hachi	steganography with git	22:58
cbreak	they'd never figure out how to use git	22:58
	← josh0x0 left	22:58
yacoob	config files, 4 different flavours	22:59
	← tghw left	22:59
	→ tghw joined	22:59
cbreak	config files in git? :/	22:59
hachi	I will hide my data in the low order bits of the SHA-1 fields	22:59
	← raorn left	22:59
	← Bombe left	22:59
cbreak	now that sounds like it requires CPU power :)	22:59
	← dwmw2_gone left	22:59
hachi	yeah... or maybe something weak... like MD5 :D	22:59
	← seanmccann left	23:00
cbreak	I heard from some competitor of freenet	23:00
	that apparently just inserts random data into a network	23:00
	→ dwmw2_gone joined	23:00
cbreak	and if you want to download something	23:00
	you download a list of random blocks you have to xor together to get the original	23:00
	really weird	23:01
	→ Bombe joined	23:01
hachi	there was the whole PAR file thing on usenet	23:01
	I loved the idea	23:01
cbreak	par is just checkblocks	23:01
yacoob	a friend of mine has written steganography lib that was using tcp headers to send the message	23:01
cbreak	current freenet has similar thing: FEC.	23:02
hachi	no, cause you could extend it out past missing a single block	23:02
	for each file you couldn't grab, get that many par files	23:02
	← LotR left	23:02
hachi	and you have enough of the polynomial to reconstruct the data	23:02
cbreak	you need to download 50% of the data and you can reconstruct the file in freenet	23:02
	→ LotR joined	23:02
Mage	i read up on a stegenography library that makes it seems like your data is completely random but is actually a secret message	23:02
cbreak	doesn't matter which 50%	23:02
hachi	cbreak: wow, that's kinda nifty	23:03
	← tg left	23:03
cbreak	(in reality, it does matter since it does FEC on a band level)	23:03
	← avtobiff left	23:03
cbreak	but it's almost like that	23:03
	→ avtobiff joined	23:03
	→ tg joined	23:03
cbreak	Mage: good AES makes your data look random	23:03
	← darwin_ left	23:03
	← radhermit left	23:03
Mage	cbreak: yeah that's what i was referring to	23:03
	:p	23:04
	→ ajpiano joined	23:04
Mage	bbl	23:04
	→ mosno joined	23:04
	→ ProperNoun joined	23:04
	→ darwin_ joined	23:04
	→ raorn joined	23:05
	← yev\|away left	23:05
	→ lightcap joined	23:06
	→ up_the_irons joined	23:07
	← meuh left	23:07
	→ yev\|away joined	23:07
	← lamont left	23:08
	→ meuh joined	23:08
	← ph^_ left	23:08
	→ radhermit joined	23:09
Ilari	Its also possible to have random "garbage" where if one knows the key, one can extract message. Different keys can extract different parts and without knowing the key, its not possible to determine what keys pull out data, how much they pull out and indeed is there any valid keys.	23:09
	← amerine left	23:10
	→ lamont joined	23:10
	→ dpsm joined	23:12
	← ajpiano left	23:13
dpsm	Hi All, I would like a suggestion for the following situation... I have received a zip file with an 'initial' project version an started to track changes on a git repo... the problem is that i eventually get updates on other zip files and need to merge the new versions with my own changes. Any suggestions ?	23:14
	← mosno left	23:14
dpsm	I have two public branches one to track the ziped file changes and another with my own commits...	23:15
	← ReinH left	23:15
rudi_s	dpsm: Sounds good, just merge the zipped file branch into master every time you get a new zip file.	23:16
	→ ReinH joined	23:16
	→ aryounce joined	23:17
	← smuf left	23:18
dpsm	rudi_s: the thing is that my changes are much less than the content that changes on the ziped files... so wouldn't the merges be painfull ?	23:18
	← JEEB left	23:19
	← twmi left	23:19
hachi	oh yeah, my other question	23:19
	← meuh left	23:19
	← j416 left	23:19
	→ JEEB joined	23:19
	← dwmw2_gone left	23:20
	→ meuh joined	23:20
hachi	when I do a `git log` on a file, there is a merge point for this file a little ways down in the history (3 or 4 revisions)	23:20
	it says this information in the log	23:20
	→ twmi joined	23:20
	← twmi left	23:20
	→ twmi joined	23:20
hachi	but I've tried a bunch of the options in the manpage, I can't find anything that gets git to follow the merge for information	23:21
wereHamster	dpsm: you'll only get merge conflicts if you both modified the same sections in files	23:21
	← raorn left	23:21
rudi_s	dpsm: Git is pretty smart when merging, so I don't think it's a problem. If you have to do the same merge again and again have a look at git rerere.	23:21
	← ReinH left	23:22
	→ raorn joined	23:22
	→ ReinH joined	23:23
dpsm	thanks everyone =]	23:23
	← dpsm left	23:23
rudi_s	np	23:23
	→ Guest20296 joined	23:23
	→ jmcantrell joined	23:23
	→ sh1mmer joined	23:24
	← bibinou left	23:26
	→ juan_arandaalvar joined	23:26
	← thiago_home left	23:26
	← tokkee left	23:27
	← sgh_ left	23:27
	→ ntoll joined	23:28
	← Ryback_ left	23:28
hachi	is the general accepted way of handing out a grafts file to simply say "here, create or append your .git/info/grafts file with these lines" ?	23:29
	← krawek left	23:29
	→ krawek joined	23:29
	← raorn left	23:29
	← dj_tjerk left	23:29
	→ raorn joined	23:29
	mikelifeguard → Mike\|\|gone	23:30
	→ rado1 joined	23:31
	← stantont left	23:31
rado1	what do the "..." mean in this command git filter-branch ... C..H ?	23:31
	→ dwmw2_gone joined	23:31
	→ sgh joined	23:32
	→ sh1m joined	23:33
rudi_s	rado1: man git-rev-parse and scroll down to 'Specifying ranges'.	23:33
Gitbot	rado1: the git-rev-parse manpage can be found at http://git.or.cz/man/git-rev-parse	23:33
	← juan_arandaalvar left	23:33
	→ daharon joined	23:33
rado1	rudi_s, thanks	23:34
crubb	to create something which might be referred to a git repository i'll need to git clone --bare ?	23:35
daharon	Hey guys, I'm trying to have two or more gitosis users serving up repos. I have one gitosis user whose keydir is a symlink to the public keys, and it works. But when I create the second and do the same symlink, I get asked for a passphrase when I try to clone. Any tips?	23:35
rudi_s	crubb: What are you trying to do?	23:35
	→ Theravadan joined	23:36
crubb	integrate GIT into trac using GitPlugin, it states the Test.git above needs to be the .git repository directory (i.e. the one containing the file HEAD and config), not the working tree directory that holds the .git/ folder.	23:36
	← sh1mmer left	23:36
rudi_s	crubb: Yep, then use git clone --bare.	23:36
	→ AdmiralSausage joined	23:36
	← Haarg left	23:36
crubb	rudi_s: cheers	23:36
	← dwmw2_gone left	23:37
	← BP{k} left	23:37
	← meuh left	23:38
	→ BP{k} joined	23:38
	→ meuh joined	23:38
	← sh1m left	23:38
	→ greg8si joined	23:38
	→ dwmw2_gone joined	23:39
	← AdmiralSausage left	23:40
	→ catalis joined	23:40
	→ AdmiralSausage joined	23:40
	← sgh left	23:41
	→ maxter joined	23:42
	← dwmw2_gone left	23:42
	→ sh1mmer joined	23:42
	→ dwmw2_gone joined	23:43
	← lightcap left	23:43
	→ mssm joined	23:43
	← hobodave left	23:46
	← ntoll left	23:46
	→ alexmchale_ joined	23:48
	← meuh left	23:49
	← maxter left	23:49
	← BP{k} left	23:49
	→ BP{k} joined	23:49
	→ meuh joined	23:49
	← AdmiralSausage left	23:52
rtyler	spearce: is there any way I can dicard drafted comments?	23:52
spearce	painfully click delete on each one	23:52
	:-(	23:52
rtyler	spearce: I accidentally hit "DOne" for somebody else's comment on diff, and I can't Edit it or discard it	23:52
	I also can't add more comments now ><_	23:52
spearce	not logged in?	23:52
rtyler	nope, I am	23:52
	→ agile joined	23:52
	m-takagi_ → m-takagi	23:53
spearce	did you click on the done line to expand it?	23:53
	done replies are automatically folded closed to save space	23:53
rtyler	spearce: it's stuck open	23:53
	no JS errors though	23:53
	← a_GWT_startup left	23:54
rtyler	going to try to signout and sign in	23:54
	→ EricInBNE joined	23:54
	→ levifig joined	23:55
rtyler	spearce: so in the unified diff view, the little right-facing triangle in the left margin indicating the line	23:55
	spearce: as soon as I click on a line, it disappears	23:55
	spearce: wondering if that has anything to do with this	23:55
spearce	that's the current line cursor...	23:55
	agib → agib\|away	23:55
rtyler	I can't move the blue line at all though	23:55
	← mssm left	23:56
spearce	huh, it should move when you click on a line. if not, bug.	23:56
	agib\|away → agib	23:56
	rtyler is a bit irritated with this	23:56
rtyler	got gsql I can use to nuke this drafted "Done"	23:56
	+?	23:56
	→ pklingem joined	23:56
spearce	delete from patch_comments where change_id = 123 and status = 'd' and author_id = youraccountid;	23:57
rtyler	spearce: aha, it's a problem with chrome	23:57
spearce	youraccountid is on your settings page	23:57
rtyler	spearce: logged in with FF, and I was able to discard the draft fine	23:57
spearce	rtyler: please file a bug, i should have caught this, and should fix it	23:57
	← d0ugal left	23:58
	← zodiak left	23:59

Logs Search ←Prev date Next date→

Channels Documentation