IRCloggy #git 2009-11-06

Provider of IRC logs since 2005.

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

	2009-11-06
	← neunon left	00:00
	→ Mezner joined	00:01
	← adamm left	00:02
	← pflanze left	00:03
	← Paraselene_ left	00:03
	m-takagi_ → m-takagi	00:04
	→ r11t joined	00:05
	→ kardan_ joined	00:06
	← TeckniX left	00:06
	← hap left	00:08
	→ jchris joined	00:09
	← spaceonline left	00:10
	→ hap joined	00:10
	← jchris left	00:10
	→ Paraselene__ joined	00:11
	← radam left	00:12
	→ bigmack83_ joined	00:12
	→ bernardh joined	00:12
godsflaw	I have a repo that I broke off my main repo with filter-branch. I have run "reflog expire --expire=now --all", "gc --aggressive", and "prune" on this guy but it's still about 329Mb. I look at all versions in the history and there are no large blobs, but that damn pack file is huge. Is there anything that will give me a better view into this?	00:13
	jeckel → jeckel\|away	00:14
	jeckel\|away → jeckel	00:14
	→ nothingHappens joined	00:14
doener	godsflaw: did you also drop the filter-branch backups in refs/original?	00:14
	godsflaw: there's a recipe at the end of man git-filter-branch that has all necessary steps for the shrinking	00:14
Gitbot	godsflaw: the git-filter-branch manpage can be found at http://git.or.cz/man/git-filter-branch	00:14
godsflaw	well I cloned the result of filter-banch	00:14
	← ph^ left	00:14
godsflaw	so that stuff is gone	00:14
	← argonel left	00:14
doener	ok	00:15
	godsflaw: could you pastebin "ls -l .git/objects/pack"?	00:15
	→ WALoeIII joined	00:15
godsflaw	sure	00:15
	doener: it's small	00:16
	← ph^_ left	00:17
doener	it's 324MB ;-)	00:17
godsflaw	yeah	00:17
	only three commits that are very small	00:18
	so that is the old objects	00:18
doener	faq find-large	00:19
Gitbot	doener: Howto find large files. See http://git.or.cz/gitwiki/GitFaq#find-large	00:19
doener	godsflaw: did you try that script?	00:19
godsflaw	git log --name-only --follow --all -- .	00:19
	that shows a ton of junk	00:19
doener	godsflaw: hm, do you have tags?	00:19
godsflaw	yes	00:20
doener	godsflaw: did you properly rewrite those?	00:20
godsflaw	no	00:20
	=)	00:20
	I bet I did not	00:20
doener	(or: are they still meaningful? Maybe just drop them)	00:20
godsflaw	I might drop them	00:20
doener	subdirectory-filter often means that the tags become somewhat useless	00:21
godsflaw	we are not going to use them	00:21
	so you think this is our old tags	00:21
	← bigmack83_ left	00:21
	← bigmack83 left	00:21
	→ argonel joined	00:21
	← ixti left	00:21
doener	godsflaw: try "git log ... --branches", (instead of --all)	00:22
	godsflaw: then with --remotes, then with --tags	00:22
	← nsussman left	00:22
godsflaw	okay	00:22
doener	if only --tags shows the crap, you got your answer	00:22
	→ jschoolc_ joined	00:22
godsflaw	remotes is clean	00:23
	← kardan\| left	00:23
godsflaw	branches and tags produce a ton of info	00:23
	→ mattayers joined	00:23
godsflaw	I need the branches though	00:23
	and the other sub directories I broke out shrank	00:23
doener	godsflaw: hm, I took the "tons of junk" to mean "things that I thought I dropped"	00:23
	godsflaw: --branches definitely shouldn't show that kind of stuff	00:24
	→ onigiri joined	00:24
godsflaw	so --branches shows log history for stuff I removed with filter-branch	00:25
	same with tags	00:25
	→ vbgunz joined	00:25
doener	godsflaw: hm, that's weird. Especially since you cloned.	00:25
	godsflaw: that gives just a single branch that starts out from a remote tracking branch	00:26
godsflaw	let me message you the command I ran	00:26
doener	godsflaw: and you said that the remote tracking branches are clean	00:26
	makes no sense at all to me	00:26
	godsflaw: bold guess: you're in the wrong directory?	00:26
	→ rettub_ joined	00:27
	→ LoRe joined	00:27
godsflaw	doener: haha, nope	00:27
	← mattayers left	00:27
godsflaw	doener: but you will see if I am doing something silly in a sec	00:27
	← jschoolcraft left	00:28
	← drubies left	00:28
	← SRabbelier left	00:29
	Paraselene__ → Paraselene_	00:30
	← alezandro left	00:30
	← LiamH left	00:31
	← bernardh left	00:31
	← kimmie left	00:32
	← Ryback_ left	00:33
	← orafu left	00:33
	→ orafu joined	00:33
	→ archis joined	00:35
	← bentob0x left	00:36
	← archis left	00:36
	← Paraselene_ left	00:36
Mezner	Does anyone know how to clone a repo off a samba share in windows?	00:37
	jeckel → revans	00:37
selckin	c	00:37
	cp	00:37
	← relix left	00:37
	← rettub left	00:38
	← nothingHappens left	00:39
	← elTigre left	00:44
	→ muthu joined	00:45
	← nkuttler left	00:45
	← ilteris_ left	00:47
	← Yuuhi left	00:47
	→ ilteris_ joined	00:47
c_sphere	Is it valid that a bare repository has remotes defined (just for fetching, not merging)?	00:48
	→ lightcap joined	00:49
	← leaf-sheep left	00:50
	→ rchavik joined	00:51
	← hakunin left	00:51
	← zuez left	00:52
	← jaked left	00:53
lightcap	can I amend my last commit to not include files that were accidentally committed? (damn commit -a)	00:53
doener	lightcap: git reset HEAD^ -- path; git commit --amend	00:56
	→ DrNick2 joined	00:57
lightcap	doener: so if I want to remove two of the three files commited in that last commit, I'll need to add the one back that I wanted to keep, right?	00:57
doener	lightcap: no	00:57
lightcap	oh	00:57
doener	lightcap: you simply don't pass that one to the reset call	00:57
lightcap	ah!	00:57
	cool	00:57
	thanks!	00:57
doener	lightcap: "git reset <commit> -- path" only updates the given index entries	00:58
	→ agile joined	00:58
doener	lightcap: that's different from "git reset <commit>"	00:58
lightcap	sweet	00:58
doener	with "git reset HEAD^", the "git commit --amend" would make no sense at all	00:58
	→ nothingHappens joined	00:59
lightcap	right, it's really the same thing right?	00:59
	oh, no.	00:59
	git commit --amend is basically git rebase HEAD^	01:00
	your answer worked perfect, thanks a ton	01:00
doener	no, "git commit --amend" is like "git reset --soft HEAD^; git commit -c ORIG_HEAD"	01:00
lightcap	oh, I see.	01:01
	that makes more sense for sure.	01:01
doener	except for the fact that "git commit --amend" can handle merge commits, the other thing doesn't	01:01
	→ jchris joined	01:01
	→ Cainus joined	01:01
	→ hydrogen joined	01:02
	→ reprore_ joined	01:02
	← mjf left	01:02
	revans → revans\|away	01:03
Cainus	hey all... I'm looking for some git-newbie help.. I call "git pull" from one machine successfully everytime, but the other one is asking me to specify a branch to merge with... how do I know the branch name, and how do I specify it automatically like the other box?	01:03
c_sphere	Guess is that you wanted `git fetch`, not pull	01:04
	→ ustun_ joined	01:04
Cainus	but fetch doesn't appear to get the latest changes from the other machine	01:04
doener	Cainus: it does. "pull" is fetch+merge	01:05
	← xymbol left	01:05
doener	Cainus: fetch just (by default) updates the remote tracking branch.	01:05
	← ustun left	01:05
doener	Cainus: pull additionally merges things to your checked out branch	01:05
	→ xymbol joined	01:05
	ustun_ → ustun	01:05
	← DrNick left	01:05
doener	Cainus: and the error message from pull tells you how to setup the default for it	01:05
Cainus	right... but I can see on the github we interface that my changes are there... but when I do a git fetch, it doesn't get those changes	01:06
	← nothingHappens left	01:06
	← hydrogen left	01:06
	→ hydrogen joined	01:06
Cainus	errr github web interface	01:06
	sounds like I want a pull then, no?	01:07
c_sphere	Depends on what you want.	01:07
Cainus	I want to see those files on github in my local filesystem ;)	01:07
c_sphere	so do git merge origin/whatever	01:08
	← ustun left	01:08
doener	Cainus: git log origin/master	01:08
	→ ustun joined	01:09
doener	Cainus: that'll show the log of the remote tracking branch origin/master, which will confirm that "fetch" got the changes (assuming they're in the "master" branch head of the remote repo)	01:09
	(I'll simply assume that you only have one remote right now and that it's called origin)	01:09
Cainus	k so git log master doesn't have my lateste commit in the list	01:10
	it's got all of them but the latest	01:10
	← xymbol left	01:10
Cainus	github shows the latest commit though	01:10
doener	Cainus: because you didn't merge them yet	01:10
Cainus	"git merge master"?	01:10
doener	Cainus: "git pull" would do that automatically, or you can use "git merge origin/master", as c_sphere said	01:10
Cainus	git merge master -> already up-to-date	01:11
doener	Cainus: why would you do "git merge master" to merge "master" when the commits you want are in "origin/master"?	01:11
	← dan__ left	01:11
	DrNick2 → DrNick	01:11
doener	Cainus: if you do "git config --get remote.origin.fetch", you'll see:	01:12
	+refs/heads/:refs/remotes/origin/	01:12
	← ustun left	01:12
doener	which is the default refspec used when you fetch from the remote "origin"	01:12
	it means that all branch heads from the remote repo are fetched and get stored locally in remote tracking branches in refs/remotes/origin/*	01:12
Cainus	git config --get remote.origin.fetch does nothing... do I need to set some config var?	01:12
doener	Cainus: but "git log origin/master" worked? Or didn't you even try that?	01:13
	Cainus: seems like you never even added a remote then	01:13
	→ ustun joined	01:13
doener	Cainus: git remote add -f origin <url>	01:13
	Cainus: that adds the remote 'origin' and fetches from it	01:14
Cainus	yeah git log origin/master worked	01:14
	← hobodave left	01:14
	← thiago_rwc left	01:14
	→ Antaranian joined	01:14
doener	but then you should have a remote 'origin' and that should have 'fetch' set, unless you created that in some funky way	01:14
	could you pastebin your .git/config?	01:14
	← Antaranian left	01:15
	← dreiss left	01:15
Cainus	http://pastebin.com/m67041b7a	01:17
	→ hyperair joined	01:17
doener	Cainus: given that config, "git config --get remote.origin.fetch" should have shown "+refs/heads/:refs/remotes/origin/", as I said	01:17
	Cainus: anyway... you got the config entry...	01:17
Cainus	? it shows it now	01:18
	weird	01:18
	so now what?	01:18
doener	Cainus: so when you do "git fetch origin" (or just "git fetch", it defaults to 'origin' in your case), git will fetch all branch heads from the remtoe to update the remote tracking branches	01:18
	and then you can e.g. merge from them (or use log to see what you would merge or whatever)	01:19
	"git pull" is a convenience wrapper that does fetch+merge	01:19
Cainus	k git fetch origin doesn't get any changes to my local filesystem though	01:19
doener	it reads the config values branch.<name>.remote and branch.<name>.merge	01:19
	Cainus: it _does_, it puts the things into your git repo	01:19
	Cainus: and your git repo is obviously in your filesystem	01:20
	Cainus: it just doesn't merge	01:20
Cainus	and then I need to check them out somehow?	01:20
	sorry... I'm coming from P4 and svn, so I don't entirely get the local repo idea	01:21
doener	Cainus: so you can setup branch.<name>.remote and branch.<name>.merge to get defaults for "git pull" (just like the error message tells you)	01:21
	→ alezandro joined	01:21
	revans\|away → revans	01:22
Cainus	what should the values be though? how do I figure out the branch name?	01:22
jjuran	Cainus: git branch -r	01:22
Cainus	ah origin/master ?	01:23
doener	jjuran: that's actually useless there. branch.<name>.merge wants the ref of the remote repo, not the local remote tracking branch	01:23
	Cainus: well, if you want to fetch and merge the "master" branch head of the remote repo, it's:	01:23
	git config branch.<name>.remote origin	01:23
jjuran	doener: I suggested that as a way to find out what remote branches exist.	01:23
doener	git config branch.<name>.merge refs/heads/master	01:23
	jjuran: that's better done via "git remote show <name>"	01:24
	← muthu left	01:24
Cainus	I assume you mean <name> to be some var I swap into there... I don't know the value though	01:24
doener	Cainus: the name of your local branch head	01:24
	→ leaf-sheep joined	01:24
Cainus	yeah I understand the meaning of the var... I just don't know the value... is it apparent from my clone URL?	01:24
	→ MikeJ joined	01:25
	→ j416 joined	01:25
	→ elliottcable joined	01:25
	← hydrogen left	01:26
	→ jinzhu joined	01:26
	→ hydrogen joined	01:26
MikeJ	greetings #git, quick question, sometimes when doing a large commit with git svn dcommit it hangs, I assume doing some sort of rename detection.. any pointers how to disable this	01:26
Cainus	github tells me my only branch is master... is that what the value should be?	01:26
MikeJ	it successfully does the svn commit, then hangs after this	01:27
	it appears after the fetch	01:27
	or is there a way to hold off on a dcommit to not do the rebase between each svn commit	01:27
	→ pflanze joined	01:28
	→ PN joined	01:30
	← ustun left	01:30
	→ ustun joined	01:30
	← ProperNoun left	01:32
	← WebDragon left	01:35
Cainus	http://stackoverflow.com/questions/658885/how-do-you-get-git-to-always-pull-from-a-specific-branch solved it for me, but I'd like to understand what the heck I was doing wrong for it to break	01:35
doener	Cainus: you never setup branch.<name>.remote and branch.<name>.merge	01:36
	← hyperair left	01:36
	→ hydrogen_ joined	01:36
	← hydrogen left	01:36
doener	Cainus: those config settings are automatically setup when you create a branch head starting from a remote tracking branch, but not otherwise	01:36
Cainus	it did work previously though	01:37
doener	Cainus: then you somehow dropped those values, maybe you dropped the branch head and then recreated it from a specific commit (instead of a remote tracking branch), or you manually dropped the config settings, or whatever	01:37
Cainus	ah alright.... weird	01:38
	okay thanks for the help... that's got me going again ;)	01:38
	← j416 left	01:38
	← __iron left	01:39
	→ j416 joined	01:39
	← elliottcable left	01:41
	← Cainus left	01:41
	→ suman joined	01:43
	→ developernotes joined	01:43
developernotes	I've copied kdiff3.app/ to /usr/local/bin and also modified my ~/.gitconfig but it doesn't seem like git is able to launch kdiff3. Do you have to do something special with the pathing b/c kdiff3.app is a folder?	01:45
	→ muthu joined	01:48
jjuran	developernotes: Are you using it as a Mac application or a command-line program?	01:49
developernotes	jjuran: I would like to use it as a Mac application - GUI tool.	01:49
	I tried to update the entry to be as the following:	01:50
jjuran	developernotes: Putting the .app folder in a PATH directory won't help. You may as well keep it in /Applications or $HOME/Applications.	01:50
developernotes	cmd = "/usr/local/bin/kdiff3.app/Contents/MacOS/kdiff3"	01:50
	← davey left	01:50
	→ ablemike joined	01:52
	→ flazz joined	01:52
developernotes	jjuran: I moved it to /Applications but it still won't launch	01:52
	→ davey joined	01:53
jjuran	developernotes: I don't know how kdiff3 works. Applications like BBEdit come with separate command-line tools, bbedit and bbdiff, which are used to create single-document sessions inside the application process. If you run the executable inside Contents/MacOS, then the entire application process is your session. That might be how it's intended to work, though.	01:54
	developernotes: I would try running it directly from your shell before trying to configure it in Git.	01:55
developernotes	jjuran: ok, I'll keep plugging away at it.	01:55
	→ schambers joined	01:56
developernotes	jjuran: right, I can run it from terminal when I type 'open /Applications/kdiff3.app/'	01:56
	jjuran: I wonder if I need to include the 'open' command?	01:56
	jjuran: Nope, that didn't do it either.	01:57
	→ hyperair joined	01:58
jjuran	developernotes: 'open' will launch a normal Mac application.	01:58
	← ustun left	01:59
developernotes	jjuran: Yes, I know, I tried to add it to the cmd = "open /Applications/kdiff3.app/" entry in my .gitconfig	01:59
	jjuran: it didn't help.	02:00
jjuran	developernotes: No, 'open' belongs to the Mac universe, not the Unix one.	02:00
	→ johnnyg1 joined	02:00
	→ elliottcable joined	02:00
	← johnnyg1 left	02:01
	→ ustun joined	02:01
jjuran	developernotes: Hmmm, Contents/Resources/en has HTML docs.	02:01
	← qhoxie_ left	02:02
	← hydrogen_ left	02:02
developernotes	jjuran: ok	02:02
	← elliottcable left	02:03
jjuran	developernotes: Are you sure kdiff3.app even supports command-line scripting? I haven't found anything resembling a manual for this.	02:04
developernotes	jjuran: it covers installation from source, I already have it built.	02:04
	← Mezner left	02:05
developernotes	jjuran: I have kdiff3 setup as a gui tool with git on the Windows side so I am aware that it works.	02:05
	→ dreiss joined	02:06
jjuran	developernotes: Found it. It's in chapter 2.	02:06
	→ jds joined	02:07
developernotes	jjuran: It reffers to all commands as just kdiff3.	02:07
	jjuran: in chapter 2.	02:07
	jjuran: how does that work?	02:09
jjuran	developernotes: If I run kdiff3.app/Contents/MacOS/kdiff3 foo bar, I get a diff in the GUI.	02:10
developernotes	jjuran: but are you able to launch it from executing 'git mergetool' on a single merge conflict?	02:12
jjuran	developernotes: Did you set mergetool.kdiff3.path?	02:13
	→ tomoyuki28jp joined	02:13
tomoyuki28jp	How can I get the current hash tag of git?	02:13
doener	define "current hash tag of git" please	02:14
	→ stringo0 joined	02:14
jjuran	developernotes: mergetool.foo.cmd is for writing a custom tool that takes its input from environment variables.	02:14
	← jaysern left	02:15
tomoyuki28jp	doener: Get the hash of the last commit.	02:15
	→ [JL]String\|Mobil joined	02:15
doener	tomoyuki28jp: the checked out commit? "last" in terms of time, isn't easy to get	02:15
	[JL]String\|Mobil → stringo0_	02:15
jjuran	tomoyuki28jp: git show-ref HEAD?	02:15
	← onigiri left	02:15
doener	tomoyuki28jp: "git rev-parse HEAD" would give the hash of the commit that HEAD resolves to	02:15
	→ redcuber joined	02:16
doener	(HEAD == what you have checked out)	02:16
developernotes	jjuran: [mergetool "kdiff3"]	02:16
	cmd = "open /Applications/kdiff3.app"	02:16
	← jds left	02:16
doener	jjuran: that would be "git show-ref -h HEAD"	02:16
tomoyuki28jp	'git rev-parse HEAD' gives me what I wanted. Thanks a lot!	02:16
	← redcuber left	02:17
doener	jjuran: without -h, HEAD isn't shown, so limiting to HEAD would result in no output	02:17
jjuran	developernotes: That won't work. Try: path = /Applications/kdiff3.app/Contents/MacOS/kdiff3 instead.	02:17
	developernotes: 'open' launches an application and returns immediately. No connection remains between the shell and the app.	02:18
	← sagsousuke left	02:19
developernotes	jjuran: When I ran that, I get the following:	02:21
	The merge tool kdiff3 is not available as '/Applications/kdiff3.app/Contents/MacOS/kdiff3 $PWD/$BASE $PWD/$LOCAL $PWD/$REMOTE $PWD/$MERGED'	02:21
	← tomoyuki28jp left	02:21
	→ VVSiz_ joined	02:24
developernotes	jjuran: Got it, the other pieces were leftover from the cmd entry, removing that lauched kdiff3.	02:24
	jjuran: Thanks for your help. :-)	02:24
	← kennethlove left	02:24
jjuran	developernotes: No problem. :-)	02:25
	← ablemike left	02:26
	→ ablemike joined	02:26
	← ablemike left	02:27
	← VVSiz left	02:31
	← developernotes left	02:31
	← stringo0 left	02:32
	→ hydrogen joined	02:34
	→ freerobby joined	02:36
	← suman left	02:37
	revans → revans\|away	02:39
	← engla left	02:39
	→ jigfox joined	02:40
	← markelikalderon left	02:42
	← hyperair left	02:42
	→ JoeConyers joined	02:43
	← jigfox_ left	02:44
	→ drubies joined	02:45
	← hydrogen left	02:45
	→ hydrogen joined	02:46
	← hydrogen left	02:46
	→ hydrogen joined	02:46
	← bugfux left	02:47
	← tstclair left	02:47
	→ jeremyhu\|yuffie joined	02:47
jeremyhu\|yuffie	I pushed some changes to my personal repo, but then I realized a commit message was wrong, so I did a 'rebase -i HEAD^' to fix the message, so now it's fixed locally. I want to push it to my repo, but 'git push jeremyhu +master' reports 'Everything up-to-date' even though HEAD and jeremyhu/master are different	02:48
	What am Idoing wrong? How can I correctly push the change?	02:48
	→ upul joined	02:50
Ilari	jeremyhu\|yuffie: Does 'git branch' show you are on master? Does 'git log' show the updated message?	02:52
jeremyhu\|yuffie	Ilari: yes, git log shows the updated commit message/hash, git log jeremyhu/master shows the old commit message/hash. I am on master.	02:53
	← MikeJ left	02:53
	← drizzd_ left	02:54
Ilari	jeremyhu\|yuffie: And 'git fetch jeremyhu' doesn't seem to do anything?	02:54
jeremyhu\|yuffie	Ilari: correct, git fetch jeremyhu does nothing.	02:55
	← tk left	02:55
doener	jeremyhu\|yuffie: and the first part? Does "git branch" show that you have master checked out?	02:56
	jeremyhu\|yuffie: or does "git log master" also show the fixed things?	02:56
jeremyhu\|yuffie	doener: ah interesting... it shows me on (no branch) for some reason...	02:57
doener	jeremyhu\|yuffie: you didn't finish the rebase	02:57
	jeremyhu\|yuffie: likely you forgot to run "git rebase --continue"	02:57
	jeremyhu\|yuffie: I guess you just did "git rebase -i HEAD^", change to edit, save, quit, edit, commit --amend	02:57
jeremyhu\|yuffie	><	02:57
	that was it, thanks	02:57
doener	rebase always puts you on a detached HEAD, and only updates the branch head once the rebase is finished	02:58
jeremyhu\|yuffie	ok, thanks for the help =)	02:58
	← jigfox left	03:00
	→ bdimcheff joined	03:00
	← kpreid_ left	03:00
	→ erikvold joined	03:00
	← ilteris_ left	03:01
	→ elliottcable joined	03:05
	← freerobby left	03:06
	← muthu left	03:07
	→ markelikalderon joined	03:08
	← patmaddox left	03:08
	→ hyperair joined	03:08
parasti	quit	03:10
	→ johnw joined	03:10
parasti	oops	03:10
johnw	join #haskell	03:10
	← parasti left	03:10
johnw	slow / key :)	03:10
doener	johnw: haha, you missed "04:10:10 <parasti> quit" right before you joined	03:11
johnw	hmm... must be the channel...	03:11
	→ ilteris_ joined	03:13
	← davido left	03:13
	→ kpreid_ joined	03:14
johnw	ah, found my /	03:14
	it was in the Boost channel	03:14
	→ davido joined	03:14
	→ cytrinox_ joined	03:20
	→ paltman joined	03:20
	← revans\|away left	03:20
	→ bcardarella joined	03:20
	← cytrinox left	03:21
	cytrinox_ → cytrinox	03:21
	← drubies left	03:22
	→ bibinou_ joined	03:23
	← bibinou_ left	03:24
	→ bibinou_ joined	03:24
	→ SjB joined	03:25
	← bcardarella left	03:26
	← bibinou left	03:26
	← SjB left	03:26
	bibinou_ → bibinou	03:27
	bibinou → cappert	03:27
	→ imajes joined	03:28
	cappert → bibinou	03:28
	← bibinou left	03:28
	→ bcardarella joined	03:28
	→ bibinou joined	03:29
	→ patmaddox joined	03:31
	← nsussman` left	03:32
	→ nsussman` joined	03:34
	→ secoif joined	03:39
	← elliottcable left	03:39
	← nsussman` left	03:40
	← john3909 left	03:40
	← godsflaw left	03:42
	← kukks left	03:43
	→ develCuy joined	03:45
	→ abms1116 joined	03:46
develCuy	I have to do a big "push" via ssh, but can't be completed because of memory limits for remote user. I have root access to that server, but don't know how to increase memory limit for my git user :(	03:47
jjuran	develCuy: Do smaller pushes?	03:47
develCuy	jjuran, how can I split my push?	03:48
jjuran	git push origin abc123:temp	03:48
develCuy	where abc123: temp is?	03:48
	← awnstudio left	03:48
lucs	Hey, that's my password!	03:48
jjuran	develCuy: abc123 is one of your commits along the way. 'temp' is just a temporary branch on the remote side.	03:49
	lucs: Sorry, I forgot I had it in my clipboard.	03:49
lucs	:)	03:49
develCuy	lucs, try a longer pass: 121212121212	03:49
jjuran	Then git push origin def456:temp, etc.	03:49
	← tedoc2000 left	03:50
	→ keystr0k joined	03:50
jjuran	Once the remote has all the commits you can git push origin master.	03:50
develCuy	just that? don't have to move/mix the commits from temp branch?	03:50
lucs	According to some study reported by Bruce Schneier, the most common password that was found in a sampling of many thousands (can't remember, maybe 50K), was 'password1'.	03:51
	→ tome joined	03:51
	← schmukkk left	03:51
lucs	a	03:51
	Oops	03:51
jjuran	develCuy: You don't need to do anything with the temp branch. The purpose is to get the commits to the remote.	03:51
lucs	At least, it has letters and digits :)	03:51
jjuran	lucs: When you require both letters and numbers (ruling out "password"), then yes, it's "password1".	03:51
	lucs lets the on-topic folks get on with their work.	03:51
tome	Hi, is it possible to clone a remote repository but only the master branch? I like how nicely git clone works, but I want new developers to only grab the master branch. Perhaps I should not be using clone?	03:52
	→ tjaway joined	03:52
	← tjafk left	03:53
develCuy	tome, I guess you can :)	03:54
tome	develCuy, is it easy to use something else?	03:55
	All I know is clone (newish git user)	03:55
	← WALoeIII left	03:55
develCuy	jjuran, I have tried what you said, and got	03:55
	error: unable to push to unqualified destination: temp	03:55
	The destination refspec neither matches an existing ref on the remote nor	03:55
	begins with refs/, and we are unable to guess a prefix based on the source ref.	03:55
	doing: git checkout -b temp	03:57
jjuran	develCuy: Try refs/heads/temp instead of temp.	03:57
	develCuy: That should work too.	03:57
develCuy	did not worked :(	03:58
	now trying with what you said	03:58
	← cedricv left	03:58
develCuy	:( still getting out of memory	03:59
	that commit is really big...	03:59
	with a file of 60MB	03:59
jjuran	develCuy: Text, or binary?	04:00
develCuy	binary	04:00
	→ elliottcable joined	04:00
jjuran	develCuy: Avoid checking in large binary files if possible. It's not one of Git's strong suits.	04:01
	← abms1116 left	04:01
	← gk left	04:01
develCuy	so I have to remove that file right?	04:01
tome	develCuy, you said you guess I can, how do you think I can do that? Or is what I am asking silly, and everyone usually just has all the remote branches cataloged (and showing up in git branch -r)	04:02
	← markelikalderon left	04:02
jjuran	develCuy: You'd have to remove the actual commit adding it and rebase the ones following it.	04:02
j416	jjuran: wouldn't he be able to just rsync over the objects dir?	04:02
	→ gk joined	04:03
develCuy	j416, I had that Idea, but my remote is a bare repo	04:03
	← squentin left	04:03
jjuran	j416: Probably, but are you sure the large file wouldn't pose an issue later on? Better not to add it at all if it can be avoided.	04:04
develCuy	tome: i'm not sure if you can accomplish with really basic skills, don't want to make you break something	04:04
	I'm gonna git clone till 1commit_before_dead	04:04
jjuran	develCuy: non-bare/.git/objects and bare.git/objects are the same structure.	04:05
	tome: I didn't see anything in man git-clone about transferring only the commits for a single branch.	04:06
Gitbot	tome: the git-clone manpage can be found at http://git.or.cz/man/git-clone	04:06
jjuran	Although git clone --branch lets you specify which branch to check out.	04:06
develCuy	jjuran, I guess a git pull origin branch will work	04:06
	he have to add remote repository manually	04:07
j416	jjuran: if it's already added... it will be there right	04:08
	I can imagine that if you can't even push, you would have trouble cloning or fetching from it later	04:09
	I suggest, increase memory at the remote	04:09
develCuy	j416, I'm trying to that, but ulimit -v is very low	04:09
	j416, I have root access but don't know where to increase the value :(	04:10
	or is there some way to wrap the git-receive-pack to call it with sudo?	04:10
	→ jlilly joined	04:10
jjuran	j416: Getting the blob transferred is not enough. If git tries to repack objects it will run out of memory again.	04:11
j416	I really have no idea about these things...	04:11
	jjuran: yeah	04:11
	that's what I meant	04:11
	when someone fetches or clones, it will die	04:11
	← jschoolc_ left	04:13
	→ jesterKing joined	04:13
cehteh	j416: 32bit os?	04:15
jesterKing	is it possible to give git explicitely ssh key like you would ssh -i /path/to/sshkey ?	04:15
j416	cehteh: ask develCuy	04:15
	j416 64-bit os	04:15
develCuy	32bit yep	04:15
cehteh	can you try to clean it up on a 64 bit box?	04:16
	jesterKing: iirc only by ssh config file	04:16
develCuy	cehteh, clean it up?	04:16
tome	An answer I decided on is that to get only the remote master branch tracked, I'd have to add the remote and then track master (or better, HEAD) manually.	04:16
	I don't see git clone doing that	04:17
jesterKing	cehteh: hmm, maybe setting GIT_SSH where it points to a script	04:17
cehteh	i just looking here, didnt read your problem .. but there is something big which might be removed from the repo	04:17
	jesterKing tests	04:17
cehteh	jesterKing: yes that may work	04:17
develCuy	I'm trying to pull into another repo, before big commit	04:19
	← krynnotaur left	04:19
develCuy	git pull origin dbeb	04:19
	does not works, got error	04:19
	oh, I forgot the ":" before refspec	04:20
	:( it is not working	04:21
	← eletuchy left	04:21
cehteh	it will certainly work on a 64 bit machine (but not really solve the problem)	04:21
	← r11t left	04:21
	← bcardarella left	04:21
	→ r11t joined	04:22
	→ bcardarella joined	04:22
develCuy	how can I pull a from first commit to commitN ?	04:22
	← r11t left	04:22
	→ r11t joined	04:22
	← bcardarella left	04:22
	← r11t left	04:22
	← lightcap left	04:22
	← keystr0k left	04:23
	→ r11t joined	04:23
	→ WALoeIII joined	04:23
	← r11t left	04:23
	→ bcardarella joined	04:23
	→ r11t joined	04:23
	← bcardarella left	04:24
	← r11t left	04:24
	→ r11t joined	04:24
	→ bcardarella joined	04:24
	offby1` → offby1	04:25
tome	Hi, when do the remote branches of a repository get updated in my local remote references? I know it happens at git clone, but when else?	04:26
	can't seem to find the docs for that	04:26
	← elliottcable left	04:27
	← bcardarella left	04:27
milli	tome: after a fetch or a pull (which is just a fetch and merge)	04:28
cehteh	develCuy: fetch and them merge by commit-sha1	04:29
tome	do you think that if I tracked a single branch, so if I didn't set up a repo with clone but rather with a branch --track, that a pull would Also still update all remote branches into my refs?	04:29
	← hydrogen left	04:29
	→ ablemike joined	04:30
	cehteh wonders if the fetch will already barf	04:30
milli	tome: git doesn't work like svn or bzr, you have to get the entire repo so you should always start with a clone	04:30
	tome: and after that, you can decide to just not track certain remote branches with a local tracking branch.	04:31
	in fact, the only local tracking branch you have after a clone is 'master'	04:32
	→ yaotti_aw joined	04:32
	yaotti_aw → yaotti	04:32
tome	milli, if the remote has hundreds of branches isn't it a lot of overhead to update them each pull?	04:33
	update = check if they need to be added to the list	04:33
	← spearce left	04:33
tome	i realize it's just list maint. no 'real' data is actually sent	04:33
	← vbgunz left	04:34
milli	tome: not usually... git is pretty fast	04:34
tome	milli, do the remote refs get removed from the local refs if they are removed on the remote?	04:35
milli	tome: keep in my that "git pull" only updates local tracked branches	04:35
	mind	04:35
tome	milli, but doesn't it have to refresh that list of all the possible branches	04:35
milli	tome: branch and tag refs are just small files containing the SHA1 of the commit they correspond to. It doesn't take long to update even thousands of those.	04:37
tome	lets be silly and say you had 100 remote branches each 10 characters long in name, so someone has to send/get 1000 characters at least to keep the lists updated.	04:37
	milli, gotcha, so they are updated, but their cost is minimal	04:37
	→ scarabx joined	04:39
	← aresnick left	04:39
tome	milli, thanks for your help. If the 'normal' way to do things is to clone and have git manage all those remote refs, then I'll use it that way	04:40
	← stringo0_ left	04:40
milli	tome: you'll find it's not an issue	04:40
	and you'll actually be happy when you have all those branches when your on a plane sometime....	04:40
tome	I thought it only updates the local branches?	04:41
milli	tome: was going to point you to start here if you want details on all this... http://www.kernel.org/pub/software/scm/git/docs/gittutorial-2.html	04:41
tome	gotcha	04:41
	→ eletuchy joined	04:41
develCuy	cehteh, thank you works	04:41
milli	tome: the key is to understand the object database. and as far as sending 1000 characters, that doesn't take very long these days. ;-)	04:42
	← secoif left	04:43
tome	milli, thanks for the link and for setting me down the right path with clone and not worrying about the refs. I'll read the link. Much appreciated	04:43
	→ aresnick joined	04:43
milli	tome: yw	04:43
	← eletuchy left	04:47
	← tome left	04:47
	→ bernardh joined	04:49
	← bdimcheff left	04:50
jesterKing	ah, great	04:50
	got ssh stuff now set up to run from memorystick	04:51
develCuy	cehteh, jjuran, thank for you help. got the problem solved	04:52
	Blessings!	04:52
cehteh	:)	04:52
	← develCuy left	04:53
	→ harinath joined	04:54
	→ bcardarella joined	04:54
	→ urbanmonk joined	04:55
	→ tedoc2000 joined	04:55
	→ [1]awnstudio joined	04:59
	← bibinou left	05:00
	← bernardh left	05:01
	→ modsognir_ joined	05:02
	← rettub_ left	05:03
	← modsognir left	05:03
	modsognir_ → modsognir	05:03
	→ rettub_ joined	05:03
	← aresnick left	05:04
	← johnw left	05:04
	→ mithro joined	05:05
	→ solofight joined	05:05
	→ xymbol joined	05:07
	← univate left	05:08
	→ r11t_ joined	05:09
	→ univate joined	05:10
	→ bdimcheff joined	05:11
	← bdimcheff left	05:11
	← davey left	05:11
	← [1]awnstudio left	05:12
	→ bdimcheff joined	05:12
	← r11t left	05:13
	jesterKing → amino	05:14
	← jfkw left	05:18
	← bcardarella left	05:21
	← cloudhead left	05:22
	→ khmarbaise joined	05:27
	← davido left	05:27
	→ davido joined	05:28
	← khmarbaise left	05:29
	← ablemike left	05:30
	→ ablemike joined	05:30
	← tedoc2000 left	05:30
	← Sho_ left	05:31
	← petteri left	05:32
	← imajes left	05:34
	→ harinath1 joined	05:39
	← jpic left	05:41
	→ imajes joined	05:42
	← jlilly left	05:43
	← harinath left	05:44
	→ ReAn` joined	05:44
	← intripoon left	05:45
	← ReAn left	05:45
	← seancorfield left	05:46
	← xymbol left	05:52
	→ [1]awnstudio joined	05:53
	← reid06 left	05:56
	→ reid06 joined	05:56
	→ hakunin joined	05:56
	← paltman left	05:56
	amino → jesterKing	05:59
	→ martins joined	06:02
	→ psankar joined	06:04
	← merlin83 left	06:05
	← r11t_ left	06:05
	← ablemike left	06:06
	→ ablemike joined	06:06
	→ merlin83 joined	06:07
	← ablemike left	06:07
	→ ablemike joined	06:08
	→ saurabhverma joined	06:12
	← niki left	06:13
	← jinzhu left	06:14
	← jesterKing left	06:14
	→ jinzhu joined	06:14
	← piroko left	06:18
	← patcito left	06:19
	← ustun left	06:19
	→ iff_ joined	06:20
	→ roop joined	06:20
	→ alexyk joined	06:21
alexyk	I have a giant .git eating space. How can I clean it up, say I don't care for most of the old history?	06:22
	← bdimcheff left	06:22
jinzhu	alexyk: git gc ?	06:25
	hakunin → hakunin\|out	06:26
alexyk	jinzhu: ah ok	06:26
Fissure	alexyk: you can use grafts and filter-branch to remove old history	06:26
	you'll need to prune the reflogs for everything to be garbage-collected though	06:26
	git wasn't designed to make losing data easy :)	06:27
	→ petteri joined	06:27
alexyk	Fissure: hmm... say I'm pulling into master and don't care about the rest -- what's the simplest way to simplify?	06:29
	← agenteo left	06:30
	→ roop_ joined	06:31
Fissure	well, where do you want to prune the history	06:31
	you find the last commit you want to keep, and add a graft that removes its parent	06:31
	→ hobodave joined	06:32
	← iff left	06:32
Fissure	then run filter-branch to actually make that the new history	06:32
	← reprore_ left	06:32
Fissure	then master or whatever refs you filter-branch'd will reflect this truncated history	06:32
	← roop left	06:33
alexyk	nice	06:33
Fissure	and you just need to delete refs that refer to old history, expire the reflogs (some subcommand of git reflog, i forget exactly), and gc --prune	06:33
	although that gc will probably wind up dumping all the old history as loose objects	06:34
	it might be easier to just clone after deleting the refs	06:34
	since a clone will only transfer the objects it actually needs	06:34
	→ johnw joined	06:35
	→ piroko joined	06:36
	→ Aikawa_ joined	06:37
	← scarabx left	06:38
	→ jceb joined	06:39
	→ nkuttler joined	06:40
	← bry left	06:40
	← alexyk left	06:45
	← Aikawa left	06:45
	← nighthwk1 left	06:45
imajes	hey. so what does "Both added" mean in a git status output?	06:51
jjuran	imajes: I'm guessing it's a merge conflict?	06:57
	→ bigmack83 joined	06:57
	→ tk joined	06:58
imajes	jjuran: yeah, just trying to understand the exact nature of it	06:59
	← bambule left	06:59
jjuran	imajes: Sounds like you're merging two branches that each added a file in the same path.	06:59
	← justin-george left	07:00
imajes	yeah	07:00
	→ a389742 joined	07:00
imajes	weird.	07:00
	shrug.	07:00
	→ jigfox joined	07:00
	← flippo left	07:00
	→ elliottcable joined	07:01
wereHamster	Fissure: the git-filter-branch man page has a section about how to remove the old objects	07:02
	← dreiss left	07:04
	→ Tuomas joined	07:05
	← jigfox left	07:06
	← elliottcable left	07:06
	→ sde_antrix1 joined	07:06
	→ elliottcable joined	07:07
	ReAn` → ReAn	07:07
	→ dreiss joined	07:08
	roop_ → roop	07:09
	→ bambule joined	07:09
	→ ludde joined	07:10
	→ Weasel[DK] joined	07:13
	→ kkoehne joined	07:15
	→ bigmack83_ joined	07:16
	← comawhite left	07:19
	→ socratees joined	07:22
	← socratees left	07:27
	← WALoeIII left	07:28
	→ evl joined	07:33
	→ giallu joined	07:36
	← johan-s left	07:37
	← piyush__ left	07:37
	← johnw left	07:40
	← ablemike left	07:40
	→ seancorfield joined	07:40
	→ ixti joined	07:42
	← evl left	07:44
	→ harinath joined	07:46
	← jceb left	07:49
	→ drfree joined	07:50
	→ an44_ joined	07:52
	← bigmack83_ left	07:52
	← bigmack83 left	07:52
	→ jfo_ joined	07:54
	← elliottcable left	07:54
j416	when I do a force-push to github, I can't see the changes	07:59
	← bambule left	07:59
j416	is this some caching thing going on, or is there some error?	08:00
	ah nevermind	08:00
	it has the correct files	08:00
	it's the web interface that doesn't show the correct stuff - I guess it will get updated after a while	08:00
	← an44 left	08:00
	an44_ → an44	08:00
	← kristian52 left	08:02
	← harinath1 left	08:02
	→ TokyoDan joined	08:03
	← krawek left	08:04
	→ maattd joined	08:05
	← jfo_ left	08:06
	← maattd left	08:06
	← thm left	08:06
	→ ntoll joined	08:06
	→ jaked_ joined	08:07
	← sverrej left	08:08
	→ maattd joined	08:08
	← dreiss left	08:09
	→ johan-s joined	08:09
	← segher left	08:09
	→ galderz joined	08:11
	→ AAA_awright_ joined	08:13
	← kristian42 left	08:14
	← hyperair left	08:14
	→ hyperair joined	08:14
	→ shruggar joined	08:15
	→ ph^ joined	08:17
	← AAA_awright left	08:19
	→ __iron joined	08:19
	→ kristian42 joined	08:21
	→ MetaCosm_ joined	08:22
	← TokyoDan left	08:22
	← jaked_ left	08:22
	→ RobertMeta_ joined	08:22
	← ryoma left	08:22
	← RobertMeta left	08:23
	→ psynaptic joined	08:23
	→ wolog joined	08:24
	→ dpino joined	08:24
	← ben_h left	08:27
	→ sitaram joined	08:28
	→ pantsman joined	08:32
	→ aliceinwire joined	08:33
	← psynaptic left	08:33
	→ onigiri joined	08:34
	→ psynaptic joined	08:34
	← Tommy[D]_ left	08:36
	← pantsman left	08:38
	→ gaveen joined	08:39
	← MetaCosm left	08:39
	→ zoniq joined	08:40
	← kelvie left	08:42
cxreg	j416: i've noticed github can be slow to notice that too	08:44
j416	cxreg: but it will notice at some point right?	08:45
	it just hasn't processed it yet	08:45
cxreg	maybe, i'm actually not sure	08:45
j416	I guess we'll find out :)	08:45
cxreg	#github might know better	08:45
j416	nah	08:45
	they're dead	08:45
	tried :P	08:45
	oh rather	08:46
	it was another Q	08:46
	maybe I should... agh what the.	08:46
	I'll just wait and see if it solves itself :)	08:46
	→ elliottcable joined	08:46
j416	I changed my email on all commits of my repo (small one <10 commits) and force-pushed it to github. Now, I can see it when I'm logged in, but if I'm not logged in the whole repo is gone. Peculiar :)	08:47
	← ntoll left	08:48
	→ ntoll joined	08:48
	← onigiri left	08:50
	← an44 left	08:53
	← hyperair left	08:57
	iff_ → Iff	08:57
	→ sverrej joined	08:58
	→ drizzd joined	08:58
	→ relix joined	08:59
	→ bentob0x joined	09:04
	→ psoo joined	09:05
sitaram	j416: I've found that it's often the browser cache that is the problem because the URLs dont look like "query" URLs	09:05
j416	sitaram: oh, interesting	09:06
	will try to clear the cache	09:06
sitaram	when a URL doesnt have "?" in it, browsers cache it longer	09:06
	j416: well for firefox, "Ctrl-Shift-F5" would do that page	09:06
	no need to clear the whole cache	09:06
j416	sitaram: nope, even tried another browser (Safari which I never use)	09:06
	same	09:06
	it's on their side	09:07
sitaram	oh... if its public give me the URL and I'll see	09:07
j416	http://github.com/j416 :)	09:07
	I bet it will be fixed soon, it's not important	09:07
sitaram	ok what should it look like	09:07
j416	there should be 2 repos	09:07
	← sverrej left	09:07
sitaram	aah a repo itself -- no that takes time I know; it's an internal Q thing. Sorry I thought you meant you made a change to a page and it didnt show up	09:08
j416	ah, no	09:08
	I had a repo, and I filter-branched the whole thing to change my email	09:08
	and then i force-pushed it	09:08
	← maattd left	09:09
	← elliottcable left	09:09
	→ Darkas_ joined	09:10
	→ Teo`` joined	09:11
j416	(interesting new verbs that form here...)	09:11
sitaram	and the repo didn't show up? I see one called 'gitupdate'	09:11
	→ ciskje joined	09:11
sitaram	:)	09:11
j416	sitaram: not yet...	09:11
	the repo is called 'tidy'	09:11
	you should be able to see it if you tag that onto the url	09:12
	but it's the old one mixed with the new one	09:12
	← reid06 left	09:12
	→ maattd joined	09:12
	← hobodave left	09:12
sitaram	are you a paying customer? could this have gone into the "private" pool and so I can't see it?	09:12
	→ reid06 joined	09:12
j416	the latest commit is 29cc75e, but the link from the file itself (in the list) goes to 14847c1	09:12
	they have the same content, just email changed	09:13
	nope, I'm not paying. Everything is public.	09:13
sitaram	then there's a problem... how long ago did you push it?	09:13
	→ engla joined	09:13
j416	not long ago	09:13
	hmm	09:13
sitaram	yes i do see it	09:13
	email them maybe... something got stuck somewhere	09:15
j416	I bet it's just slow	09:15
	I'll wait a few days and see	09:15
sitaram	could be	09:15
	few hours should do	09:15
j416	ok :)	09:15
	→ evl joined	09:17
	→ ArdRigh joined	09:18
	dwmw2_gone → dwmw2	09:20
	→ elliottcable joined	09:20
	→ parolkar joined	09:21
parolkar	What do you do to quickly look at specific version of file in git repo, while you are working on a branch (not yet commited any changes)?. Not just the diff, but actual state of the file (with all lines included)	09:21
	→ yaotti_ joined	09:22
	← a389742 left	09:24
engla	parolkar: you can view any version of a file with "git show REV:filepath" where REV is the commit id, or something like HEAD, HEAD~5 etc	09:24
	→ a389742 joined	09:24
engla	however, git show does not give helpful error messages if the file doesn't exist in that revision you chose. filepath has to be from the top of the repo	09:24
j416	what if there was a command 'git peek' that allowed you to interactively browse your commits, expand, and pick a file to view (readonly of course)	09:25
	might be handy	09:25
parolkar	engla, Thanks , u made my day	09:26
engla	j416: git gui's blame interface can almost do that(?)	09:26
	→ kelvie joined	09:26
engla	I don't use it much but	09:26
j416	engla: I was thinking without the gui	09:27
	cli browsing interface	09:27
	something like git log --oneline;, but that allows you to select a commit and show the files in it	09:27
sitaram	j416: tig might, not sure	09:30
j416	nooo	09:30
	I want 'git peek'! :D	09:30
	← parolkar left	09:31
j416	but I will check tig.	09:31
	→ tbuehlmann joined	09:31
	→ jente joined	09:33
	→ lohrun joined	09:33
	← harinath left	09:35
	→ abms1116 joined	09:37
	→ chapter_29 joined	09:42
	→ markelikalderon joined	09:42
	← elliottcable left	09:42
	← yaotti_ left	09:43
	← jspiros left	09:44
	→ jspiros joined	09:45
	← jinzhu left	09:48
chapter_29	Hallo there. Does somebody know an escape sequence for # in the .gitignore file	09:48
	→ tono joined	09:48
chapter_29	i want to ignore files with the pattern #*# but # is also the comment sign for .gitignore	09:49
engla	chapter_29: you could add a slash /#*#, then it's only for files at the root of the repo	09:50
	else I don't know	09:50
	→ everzet joined	09:50
chapter_29	thank you, i'll try that.	09:51
engla	else the help page (git help ignore) man gitignore	09:51
Gitbot	engla: the gitignore manpage can be found at http://git.or.cz/man/gitignore	09:51
	← markelikalderon left	09:52
	m-takagi → m-takagi_	09:54
	→ Bidibule joined	09:55
Bidibule	hi all!	09:56
	anyone can help me with a littl issue ?	09:56
	← psoo left	09:56
doener	chapter_29: \#*#	09:56
	→ mattayers joined	09:58
Bidibule	I modified files in a branch but I was on the bad branch	09:58
	so I need to know how is it possible to transfert those modified (and not updated in git) files form my master to my other branch	09:59
	→ Axius joined	09:59
engla	you have not yet committed changes?	10:01
	that said, it is easy to move changes whether you have committed them or not	10:01
Bidibule	no yet commited	10:01
doener	then just checkout the other branch head	10:01
Bidibule	thanks for helping engla !	10:01
engla	you should just try "git checkout otherbranch"	10:01
Bidibule	yeah but it's straneg, git just delete my files !	10:02
	:/	10:02
engla	git will try to take the changes to the new branch	10:02
cehteh	you can stash your changes, switch branches and then stash apply them	10:02
	or use git checkout -m	10:02
	→ Chillance joined	10:02
chapter_29	douner: thank you...	10:03
doener	Bidibule: deleted your files? "git checkout" doesn't delete modified files.	10:03
Bidibule	right now git says I have 4 files changes but not updated	10:03
cehteh	it just refuses to switch branches	10:03
Bidibule	I already try the git checkout otherbranch	10:04
	→ Gargoyle joined	10:04
engla	it will fail if git cannot automatically merge	10:04
doener	Bidibule: and that did what?	10:04
cehteh	what does git branch say?	10:04
Bidibule	but then a git status says nothing changed	10:04
	beacuase I modify in my master branch	10:04
	(sorry I'm a bit new to git)	10:04
Gargoyle	anyone had any problems cloning a repo on ubuntu? (Specifically the gitosis repo)	10:04
	→ markelikalderon joined	10:04
doener	Bidibule: you modified files in your working tree	10:05
cehteh	Bidibule: it should just refuse to switch branches .. you are still on the old branch	10:05
doener	Bidibule: branch heads reference commits, which of course implies that uncommitted files can't be referenced by a branch head	10:05
cehteh	and if you switch with git checkout -m then caveat this might yield conflicts	10:05
Bidibule	yeah it says 'modified files cannot change branch'	10:05
doener	Bidibule: so where did you get the "git status says nothing changed" from then?	10:06
	cehteh would feel little saver with stashing	10:06
Bidibule	sorry I'm wrong...doeneer was right I cannot chenge branch	10:06
	change	10:06
doener	Bidibule: "git checkout" refuses to switch branches if you modified a file that also differs between the checked out branch and the branch you want to switch to	10:07
Bidibule	says 'local changes cannot switch changes'	10:07
	Yeah I know that	10:07
doener	Bidibule: as cehteh said, you can either stash the changes, then checkout and then use stash apply, or use "checkout -m"	10:07
Bidibule	I just want to "transfer" thos modified files to my new branch	10:07
doener	Bidibule: either way, things might apply cleanly, or you might get conflicts	10:08
Bidibule	How can i stash (git nOOb) ? :)	10:08
_rane	git stash	10:08
	git stash apply	10:08
	gets you far	10:08
doener	Bidibule: "checkout -m" is less typing, but "stash;checkout; apply" provides means to start over if you fail to resolve the conflicts correctly	10:08
Bidibule	great !	10:09
	Stash works fine !	10:09
	:D	10:09
	thanks boys	10:09
	→ elliottcable joined	10:10
	→ themeofn joined	10:10
Gargoyle	Another one, is there anything that would go wrong if I copied an entire repo from one server to another, and then changed the Master ip address/host on my local machines?	10:11
	← dsop left	10:11
	→ voker57 joined	10:12
cehteh	a lot, but nothing with the repos :)	10:12
	git repos are simple transferable, but ensure that the rest of your setup (registered remotes, ssh, ..) is correct	10:13
Gargoyle	cehteh: :o)	10:13
Bidibule	another little question....	10:13
	how can I undo a git add -A	10:13
	??	10:13
doener	git reset	10:13
Gargoyle	cehteh: cool. I'm a bit more worried why I can't seem to git clone form the server though!	10:13
Bidibule	need git cheatsheat !	10:14
	:D	10:14
doener	google	10:14
cehteh	Gargoyle: git:// ssh or http?	10:14
Gargoyle	cehteh: git://	10:14
cehteh	well in any way that means you misconfigured something ..	10:14
	→ _grepper joined	10:15
cehteh	check git-daemon setup, firewall?	10:15
elliottcable	Hey, people. Using `git name-rev` to get the name of a tag or commit	10:15
	→ GarethAdams\|Work joined	10:15
elliottcable	as `git name-rev --tags --name-only --always --no-undefined HEAD`	10:15
	← martins left	10:15
_grepper	Does anyone know if it's possible to recover from this error: "fatal: index file smaller than expected"	10:15
	?	10:15
elliottcable	the problem is, all of my tag names come out with an appended ~0	10:15
	oops, "^0"	10:15
	how can I prevent this?	10:15
engla	elliottcable: do you want to use git describe ?	10:16
	← themeofn left	10:17
engla	elliottcable: git describe (by default) names a commit from your tags. name-rev is just terribly correct and gives the commit names.. the appended ^0 means that you mean the commit pointed to by the tag	10:18
	← __iron left	10:18
elliottcable	hm	10:18
	engla: I need it to either return a tag, if the current commit is the tip of a tag; else, the commit ID.	10:18
	engla: I’m going to be appending it to a thing as a version number of sorts; my tags are of the form "v1" or "v5"	10:19
kampasky	hmm, is there a clever easy way to make a git fetch wrapper that will call a callback in the same way as post-receive hook during push?	10:19
doener	elliottcable: that's not what your name-rev call gives you either	10:19
kampasky	i can't think of anything more elegant than ugly perl script comparing hashed git for-each-ref outputs	10:19
	even diff won't work, will it?	10:19
elliottcable	so I want it to be either Thing-x86_64-Type-v3 or Thing-x86_64-Type-a81f3e if not built from a tagged (stable) revision	10:19
kampasky	hmm, wait	10:19
	i could use comm	10:19
doener	elliottcable: (and your name-rev is like 'describe --contains', not just 'describe')	10:19
elliottcable	doener: it is, except for the ^0	10:19
engla	elliottcable: git describe does this, it makes names like v17-147-gab9061d	10:19
	elliottcable has never used describe, and will manpage it now	10:19
kampasky	i dont think i ever used comm in an actual shell script before... ;)	10:19
	→ sverrej joined	10:20
doener	elliottcable: well, "Thing-x86_64-Type-a81f3e" isn't "else, the commit ID"	10:20
Bidibule	Thanks all byt the way !	10:20
elliottcable	doener: how so?	10:20
doener	elliottcable: the (abbreviated) commit id is just "a81f3e"	10:21
elliottcable	no, I added the other stuff	10:21
	→ sprasanna joined	10:21
elliottcable	the exact code is `TAG=$(git name-rev --tags --name-only --always --no-undefined HEAD)` followed by `NAME="ArchLinux-$ARCH-$TYPE@$TAG"`	10:21
doener	elliottcable: and that could end up as: ArchLinux-x86-foo@v5~43^2~54	10:22
	($ARCH being x86 and $TYPE being foo, obviously)	10:22
engla	describe's format is more predictable	10:23
elliottcable	yeah…	10:23
	v5~43^2~54? how so?	10:23
	so far, it only returns v5 if it’s on a tag	10:23
doener	elliottcable: git checkout v5~43^2~54	10:23
elliottcable	it literally remembers what you last checked out? o_O	10:23
_grepper	FWI, upgrading git to 1.6.5.1 got rid of my problem above	10:23
	"member:fatal: index file smaller than expected"	10:24
doener	elliottcable: no, you could have done "git checkout v5~43^2~40; git checkout HEAD~14" as well	10:24
	elliottcable: it's just the checked out commit that matters	10:24
	→ dsop joined	10:25
elliottcable	hm	10:25
	anyway, had never heard of -describe, lemme manpage that real quick	10:25
engla	man git-describe	10:25
Gitbot	engla: the git-describe manpage can be found at http://git.or.cz/man/git-describe	10:25
elliottcable	oh meh	10:25
	most recent tag that is reachable from a commit?	10:25
	I only want the tag if it is directly on top of a tag, if you know what I mean	10:26
doener	elliottcable: same thing, IIRC "git describe --contains" even calls name-rev. Without --contains, it's "backwards" (as you just said)	10:26
c_sphere	if [[ "$(git rev-parse $tag)" -eq "$(git rev-parse HEAD)" ]]; then echo You are on $tag; fi;	10:26
	elliottcable continues reading, then	10:26
engla	elliottcable: v17-147-gab9061d means.. 147 commits after v17, commit id ab9061d	10:26
elliottcable	engla: yeah, caught that	10:27
doener	for some definition of "after" ;-)	10:27
elliottcable	engla: use git name-rev to generate something like that string for another project, must remember to go back and replace that with git describe	10:27
	ahhah!	10:28
	--candidates might do it!	10:28
	→ Aikawa joined	10:28
	→ rraasch joined	10:28
elliottcable	… or (reads another line) --exact-match o o	10:28
engla	..in some way 147 commits on top of v17	10:28
doener	elliottcable: with 0, yeah, same as --exact-match, but that just fails if there's no match	10:28
	← jmspeex left	10:29
doener	elliottcable: --always doesn't work there	10:29
	not sure if that's a bug	10:29
	→ yaotti_ joined	10:29
	← jetienne_ left	10:29
	→ DiaPo joined	10:30
	→ adelfin joined	10:30
c_sphere	or just check for the existence of ~ or ^ in a git describe output; if there is any, you can't be on a tag	10:31
adelfin	hi	10:31
elliottcable	hm	10:31
	yeah, seems like a bug	10:31
	would be perfect if it did o_o	10:31
engla	c_sphere: iff on a tag, $(git describe) = $(git describe --abbrev=0) (?)	10:31
adelfin	I have a little problem with git-trac-commit script	10:32
doener	engla: if you do that check, you can as well do "git describe --exact-match \|\| git rev-parse HEAD"	10:32
adelfin	It seems that can't find revs	10:32
doener	elliottcable: best bet seems to be TAG=$(git describe --exact-match 2>/dev/null \|\| git rev-parse HEAD)	10:33
c_sphere	hmpf	10:33
doener	elliottcable: maybe add --short to rev-parse	10:33
elliottcable	yeah	10:38
	that works great	10:38
	thanks!	10:38
	← Aikawa_ left	10:39
	→ intripoon joined	10:40
	← wolog left	10:41
	← upul left	10:42
	→ mineo joined	10:42
	← Theravadan left	10:44
	← chapter_29 left	10:44
	→ wolog joined	10:45
	→ bibinou joined	10:46
	← j416 left	10:46
	→ j416 joined	10:47
	→ jfkw joined	10:47
	← jmibanez left	10:53
	← abms1116 left	10:53
	→ JoseLobato joined	10:54
	→ aspotashev\|eeepc joined	10:55
	← Bidibule left	10:56
	→ Bidibule joined	10:56
	← Axius left	11:02
	→ ph^_ joined	11:06
	← maattd left	11:08
	→ Escafane joined	11:08
	← cantoma left	11:10
	→ cantoma joined	11:10
	→ crbhome_ joined	11:11
	← JoseLobato left	11:13
	→ hyperair joined	11:13
kampasky	would people be interested in CIA reporting commits in git here? or too much noise?	11:16
doener	too much noise I'd guess. Unlike SVN, you usually get dozens of commits at once with git	11:17
	so that would flood the channel everytime that Junio pushes	11:17
	but that's just a guess	11:18
jast	too much noise.	11:18
	but that's just a guess. ;)	11:18
	← brizly left	11:18
kampasky	yeah	11:18
	ok	11:18
	→ brizly joined	11:18
	→ Axius joined	11:18
	VVSiz_ → VVSiz	11:21
	← ArdRigh left	11:21
	← a389742 left	11:21
	← Gargoyle left	11:22
	→ froschi joined	11:24
	→ parasti joined	11:25
	← patmaddox left	11:26
	← lohrun left	11:26
	← hyperair left	11:27
	→ hyperair joined	11:27
	→ jschoolcraft joined	11:29
	← crbhome left	11:31
	← voker57 left	11:33
	← warthog9 left	11:33
	← jefferai left	11:33
	→ tvw joined	11:34
	← Natch\| left	11:36
	→ transplant joined	11:36
	→ warthog9 joined	11:36
	→ jefferai joined	11:37
	← vinse left	11:38
	→ hugoxrosa joined	11:39
	→ bigmack83 joined	11:47
	← bibinou left	11:47
	→ bibinou joined	11:48
	→ lorandi joined	11:51
	→ cilly joined	11:52
	← froschi left	11:54
	← adelfin left	11:55
sitaram	doener: I need some help on how to make something happen on old gits. At least 1.5.4.3 doesn't like this: cd bare-repo; GIT_WORK_TREE=some-directory git checkout -f	11:57
	← aliceinwire left	11:58
sitaram	doener: what would be a good, safe, alternative? git archive \| tar -C some-dir -xf - # seems to work fine	11:58
	(or any better ways)?	11:58
	(no worries about over-writing on that some-dir; it's what we want)	11:59
	→ a389742 joined	11:59
	→ RickXy joined	12:00
	→ voker57 joined	12:00
	← doener left	12:01
	→ doener joined	12:01
	→ colossus joined	12:03
	← sprasanna left	12:03
	→ khmarbaise joined	12:06
	← voker57 left	12:06
	← leachbj left	12:07
	← rraasch left	12:08
	← _grepper left	12:09
	m-takagi_ → m-takagi	12:09
	← tetha left	12:10
	→ tetha joined	12:10
	→ froschi joined	12:11
colossus	Hi All. Does anybody known an good issue tracker/ Code Browser WebGui for git?	12:11
	← Axius left	12:12
Rhonda	colossus: sd :)	12:13
	→ gitster` joined	12:14
Rhonda	colossus: http://syncwith.us/sd/ <-- not sure about the WebGui part though	12:14
	→ d0k joined	12:14
	← tbuehlmann left	12:15
colossus	Searching like this: http://www.indefero.net/	12:16
	→ tbuehlmann joined	12:17
	← markelikalderon left	12:17
	← transplant left	12:18
	← johan-s left	12:22
	→ Geralt joined	12:23
	← [1]awnstudio left	12:25
	→ Paraselene_ joined	12:25
	→ ehabkost joined	12:25
	→ johan-s joined	12:27
	← gitster left	12:29
	← GarethAdams\|Work left	12:29
	← colossus left	12:29
	← khmarbaise left	12:30
	← ia left	12:34
	← a389742 left	12:40
	→ a389742 joined	12:40
	← cilly left	12:41
	→ squentin joined	12:41
	← Geralt left	12:43
	← Cesario left	12:46
	← GDG left	12:47
	→ tfarina joined	12:48
	→ ia joined	12:50
	→ voker57 joined	12:51
	→ cilly joined	12:52
	← flazz left	12:53
	→ Axius joined	12:54
	← a389742 left	12:57
	→ WebDragon joined	12:57
	← RobertLaptop left	13:01
	→ aliceinwire joined	13:01
	← sdboyer left	13:03
	← kpreid_ left	13:03
	← roop left	13:04
	← solofight left	13:04
	→ ncopa joined	13:04
ncopa	hi	13:04
	are there some convenient way to get the number of commits since last tag?	13:05
shruggar	I believe so	13:05
ncopa	forexample git describe gives: v1.9.0-279-gdcc6c60 and i want 279	13:05
shruggar	ah, not sure then	13:06
ncopa	'git describe \| cut -d- -f2' works as long as there are no - in tags	13:06
lack\|work	shruggar: By the way, regarding our conversation yesterday, I think this may be a more-complete test: http://pastebin.com/m5a457de6	13:07
	Important difference: I added a 'test -e bar/new' to ensure the file is created in the proper location as a result of the '-p2'	13:07
	→ dgoodwin joined	13:08
shruggar	lack\|work, quite. I left that part out because it was late and I couldn't remember where it was intended to create the file :)	13:08
dgoodwin	hey guys, anyone familiar with the git config push.default option (i think it got removed in recent git releases, wondering if it was replaced with something else)	13:08
	trying to figure out how to make git push (no arguments) push only the current branch	13:09
lack\|work	dgoodwin: set it to "tracking"	13:09
	dgoodwin: Then read more in man git-config ;)	13:09
Gitbot	dgoodwin: the git-config manpage can be found at http://git.or.cz/man/git-config	13:09
	← zoniq left	13:10
shruggar	ncopa, I've been looking for a better way than this, but: git log --oneline tag..HEAD \| wc	13:10
lack\|work	dgoodwin: Or maybe "current", depending on what you really want the behaviour to be.	13:10
shruggar	lack\|work, and really the "find" should be removed, once the "test" is there	13:10
	→ zoniq joined	13:11
dgoodwin	lack\|work: i can see it on the web, but it's not in my local man page (git 1.6.2.5) and seems to have no effect	13:11
shruggar	lack\|work, I'd say remove the "find" line (and the && that leads to it, obviously), then send that test to the mailing list	13:11
lack\|work	shruggar: Sure, though I thought that was more information for the person watching the test output than a real "test"	13:11
	Cool, I'll do that.	13:11
dgoodwin	lack\|work: but i see it's in newer git man page so it just must be more recent than my distro, ignore me =] thanks	13:12
shruggar	lack\|work, well, it was, but only because I couldn't remember where the -e needed to point to :)	13:12
lack\|work	shruggar: One last thing, before I send this to the list -> What's the proper testbed for actually running a test like this myself?	13:12
shruggar	lack\|work, under the "t" directory of git.git	13:13
	lack\|work, and pass it -v when you're running it, otherwise it'll just say something like "failed"	13:14
	shruggar runs off to lunch before any more can be asked of him	13:15
lack\|work	dgoodwin: Ah yes. I think you need 1.6.3 for that.	13:16
	→ murr4y joined	13:16
lack\|work	shruggar: Thanks very much!	13:16
ncopa	thanks shruggar	13:16
murr4y	hi! is there any caveats i should be aware of when having a git repo inside another git repo? or should it maybe not be done at all?	13:16
	→ bry joined	13:16
lack\|work	murr4y: If you want the outer repo to know about the inner repo, learn about man git-submodule	13:17
Gitbot	murr4y: the git-submodule manpage can be found at http://git.or.cz/man/git-submodule	13:17
murr4y	ok thanks - so by default it doesn't?	13:18
lack\|work	murr4y: By default it will just show up as an untracked directory.	13:19
	You can .gitignore it if you don't want the outer git to worry about it.	13:19
murr4y	ah, yes, i see	13:19
	thanks alot :)	13:19
lack\|work	murr4y: Glad to help!	13:19
	← flaguy48 left	13:27
	← hummingkid left	13:30
	→ docelic joined	13:31
	→ abms1116 joined	13:31
docelic	Hey folks, when I do git rm, does it completely remove the thing or I can still access old history/versions?	13:31
spb	old versions are still accessible	13:34
doener	that doesn't affect existing commits.	13:34
spb	history will always be there unless you go back and rewrite it	13:35
doener	would be pretty stupid if deleting a file would break your whole version control history	13:35
docelic	Right. So if I deleted something a couple commits ago, how do I bring it back?	13:36
	→ schmukkk joined	13:36
	→ Tommy[D]_ joined	13:37
doener	docelic: for example: git checkout <old_commit> -- path/to/file	13:37
	→ alexyk joined	13:38
	→ a389742 joined	13:39
docelic	oh yes, that worked. Very nice, and thanks for your hints	13:39
	← jente left	13:40
	→ freerobby joined	13:41
	← freerobby left	13:41
	→ akahn joined	13:41
	← agile left	13:41
bentob0x	gosh, I just did a git reset HEAD^ but I don't want it	13:41
doener	so? just reset back	13:42
	→ flazz joined	13:42
bentob0x	git reset --soft patchid?	13:42
doener	no, patch ids won't work with reset... AFAIK there's no command that accepts patch ids	13:43
bentob0x	followed by git reset HEAD?	13:43
	it just did work here :)	13:43
doener	bentob0x: I doubt that you actually used a patch id	13:44
	bentob0x: patch ids are generated by, you guessed it, "git patch-id". They don't identify commits	13:45
bentob0x	I did a git reflog and it showed me 183acc9 HEAD@{1}: commit: blah so I did a git reset --soft 183acc9	13:45
	and it worked	13:45
	→ harinath joined	13:45
bentob0x	ah sorry	13:45
	I meant sha1 id	13:46
	or commit id	13:46
	my mistake	13:46
	thx doener	13:46
lack\|work	So what's the proper way to submit a new test case to the mailing list? As a patch against git.git? As a standalone script? Or does it matter?	13:46
	→ ustunozgur joined	13:46
doener	bentob0x: but doing "git reset --soft" followed by "git reset" is pretty pointless	13:46
	bentob0x: first you tell it to keep the index, then you tell it to throw it away anyway	13:46
ustunozgur	hi, I'm trying to clone a local repo to another folder, as a bare repo, but it creates an empty git repo instead	13:47
	← cilly left	13:47
ustunozgur	Initialized empty Git repository in /home/ustun/Dropbox/deneme2/	13:47
doener	bentob0x: just "git reset 183acc9" would have done (or "git reset HEAD@{1}")	13:47
ustunozgur	The command I'm using is: git clone --bare . ~/Dropbox/deneme2	13:47
doener	ustunozgur: it always says that, because it first has to create the empty repo before it can populate it	13:47
jast	lack\|work: patches are probably preferred, if the things in Documentation/SubmittingPatches are adhered to	13:47
doener	ustunozgur: did you actually try to look into the new repo?	13:48
dsop	hmm I have a .gitattributes with export-ignore pattern in my root directory of the working directory but it seems that it doesn't work	13:48
lack\|work	jast: Makes sense, thanks :)	13:48
dsop	I added test/ export-ignore to the attribute file	13:48
	and I have directory test/templates/ etc	13:48
doener	dsop: needs --worktree-attributes to use the .gitattributes files from the worktree	13:49
ustunozgur	doener, hmm I guess you are right,	13:50
doener	dsop: without that option, it uses .gitattributes from the tree it archives	13:50
ustunozgur	but the problem here is I think it is being smart and not copying files?	13:50
	because the size of the bare repo is small	13:50
dsop	doener: okay, but it should work if I put it into .git/info/attributes	13:50
doener	dsop: maybe :)	13:51
	ustunozgur: define "small". How did you check, and what did you compare it to?	13:51
bentob0x	I see	13:52
	thx doener	13:52
doener	bentob0x: you	13:52
	're welcome	13:52
	doener hates this keyboard	13:52
bentob0x	sometimes you panic a bit when you loose a patch :)	13:52
	→ hummingkid joined	13:52
dsop	doener: is there a way to debug this?	13:53
	→ mbroeker joined	13:53
doener	dsop: the dumb approach: strace, see if it tries to open the file you expect it to. Smart way: No idea.	13:54
j416	has anyone tried this bugtracker 'SD'?	13:54
	is it worth trying?	13:55
doener	dsop: the attributes code seems somewhat cryptic to me, but I haven't spent more than ~5 minutes on reading it...	13:55
	← a389742 left	13:55
ustunozgur	doener, ok, so the warning was not important as you said on the local machine. but the problem is, I'm putting into a dropbox folder, and on the other computer, it still says "you appear to have cloned an empty directory".	13:56
	→ johnnyg joined	13:56
	← psankar left	13:56
	→ ankit9 joined	13:57
	← johnnyg left	13:57
doener	ustunozgur: but the repo you're cloning actually has some branch heads, right?	13:58
ustunozgur	Yes. I tried --no-hardlinks in case it didn't copy the files.	13:58
	→ ahaczewski joined	13:58
	← ahaczewski left	13:58
ustunozgur	I'm trying to share a folder over Dropbox	13:58
	→ jlilly joined	13:58
ustunozgur	and pull from my other computer	13:58
	→ mm_work joined	13:59
	→ a389742 joined	13:59
doener	ustunozgur: what does "git branch" say in the original bare repo?	14:00
	← alexyk left	14:00
ustunozgur	master only	14:00
doener	hm, that's at least not empty	14:00
	→ kpreid_ joined	14:01
	→ drubies joined	14:02
	→ Messi joined	14:02
doener	hm, no idea right now, and I need to get back to work	14:02
	← Messi left	14:03
ustunozgur	ok, thanks.	14:03
	← DiaPo left	14:03
	→ catphish joined	14:04
	← rchavik left	14:05
	→ everzet_ joined	14:07
	← kpreid_ left	14:08
reaVer	CoD Modern Warfare 2: http://static.arstechnica.com/nbfl-tower.jpg	14:09
	→ paltman joined	14:09
	→ TeckniX joined	14:09
	→ Natch joined	14:09
	→ paltman_ joined	14:10
zomg	I don't get why people bother whining about all the mw2 stuff	14:10
	You haven't even played it yet, so how do you think you know it'll suck :p	14:11
	→ jas4711 joined	14:11
	→ blackxored joined	14:12
blackxored	how do I push tags, forgot?	14:12
	→ cenuij joined	14:12
zomg	put your hand next to the tag, then apply force	14:12
	(no idea)	14:12
lack\|work	blackxored: 'git push tag <tagname>'	14:13
	→ tstclair joined	14:13
blackxored	lack\|work, thanks I was using git push tags without results	14:13
	→ harinath1 joined	14:14
	← everzet left	14:15
	everzet_ → everzet	14:15
	→ hobodave joined	14:16
	← Axius left	14:16
	← harinath left	14:17
	← a389742 left	14:17
	← elliottcable left	14:18
	← urbanmonk left	14:19
	→ vbgunz joined	14:20
	→ bdimcheff joined	14:20
	→ cilly joined	14:20
	← vbgunz left	14:21
	→ kpreid_ joined	14:21
	← abms1116 left	14:21
	← everzet left	14:21
	← alezandro left	14:23
	→ madewokherd joined	14:23
	→ icwiener joined	14:25
	← paltman left	14:25
	paltman_ → paltman	14:25
	← cilly left	14:26
	→ cilly joined	14:26
	→ irc2samus joined	14:26
	← Darkas_ left	14:27
	→ crankharder joined	14:27
crankharder	is there a fast way to get a list of all the unique user names that have committed to a repo?	14:28
	→ petrux joined	14:28
bentob0x	can you clone a repository with all the branches right away?	14:28
	→ aresnick joined	14:28
	← nenne left	14:28
teuf	bentob0x: that's the default mode of operation of git clone	14:28
bentob0x	it only creates references to the remote branches but I can't check them out	14:29
	→ kukks joined	14:29
	→ sagsousuke joined	14:29
	→ RobertLaptop joined	14:29
engla	crankharder: git shortlog does that, but also lists every commit ever (made by each person). for example git shortlog -e -n (normally used with a commit range like v1..v2)	14:31
	crankharder: (perhaps it is not fast, I only have a very small repo)	14:32
	← jlilly left	14:32
dsop	hmm git-archive doesn't open .git/info/attributes if I run it from a subdirectory	14:32
	→ suman joined	14:33
engla	crankharder: git shortlog -e -n \| grep -E "<.+@.+>" \| sed -e 's/ (.*$//g'	14:34
	→ parasti_ joined	14:35
	← parasti left	14:35
	← kkoehne left	14:36
	→ maravillas joined	14:37
	← Thumper_ left	14:37
	→ drewolson joined	14:39
crankharder	engla: thank you	14:39
	→ Thumper_ joined	14:39
	← bigmack83 left	14:41
	→ krynnotaur joined	14:41
	← hyperair left	14:42
	→ cannonball joined	14:43
	→ _ikke_ joined	14:43
	→ scarabx joined	14:43
	AAA_awright_ → AAA_awright	14:43
	→ LiamH joined	14:44
_ikke_	When I run git difftool, what is version A en what is version B?	14:45
	→ der-onkel joined	14:45
	→ sh1mmer joined	14:46
	→ alexyk joined	14:46
der-onkel	Hi,, perhaps some can give ma a hint for [group ... in gitosis conf	14:46
	Can i use an singel [group ... ]section to define a group of developer and use this group as memeber informations for other groups?	14:48
	→ spearce joined	14:49
	→ vbgunz joined	14:50
bgerber	I can not seem to get the archive of messages to the git list for the last 24 hours. We had a power outage and it appears I was unsubscribed.	14:50
	→ name joined	14:50
bgerber	How may I get them. My memory is not working properly	14:51
sitaram	bgerber: see channel topic	14:51
_ikke_	/topic	14:52
	← s0ber left	14:52
	→ jmspeex joined	14:52
bgerber	thanks	14:53
der-onkel	ok i got it just use the @ sign	14:53
	→ elTigre joined	14:53
	← flazz left	14:54
bgerber	x	14:55
	→ DavidKlein joined	14:56
	pilif → pilif\|away	14:56
	→ s0ber joined	14:58
	→ khmarbaise joined	14:58
bgerber	that get the threads but I want to down load all for a 24hour period. There has to be a better way than clicking one at a time.	14:59
	I know I have done it. I just can not remember	14:59
	→ a389742 joined	15:00
	← DavidKlein left	15:00
	→ DavidKlein joined	15:00
	→ jmcantrell joined	15:00
	← der-onkel left	15:01
	→ nighthwk1 joined	15:01
	→ harinath joined	15:02
	← brendon_work left	15:02
	→ Turias joined	15:02
	→ loincloth joined	15:03
	→ brendon_work joined	15:03
	→ hyperair joined	15:03
	← harinath1 left	15:04
	← jas4711 left	15:04
	← intripoon left	15:05
	← sh1mmer left	15:06
	← sergiodj left	15:07
	← reaVer left	15:09
	→ corni joined	15:09
jmcconnell	any idea why a successful push into a svn cloned repo would result in it getting corrupted (missing tree, now failing fsck)?	15:10
	← gaveen left	15:12
	→ Haarg joined	15:13
doener	jmcconnell: grafts?	15:13
jmcconnell	thats a new term to me	15:13
doener	so probably not	15:13
Thell	anyone happen to know of a public repo of a project that uses -alpha -beta -rc tags and has done so for a full version?	15:14
	Thell is just wanting to test a script	15:14
doener	Thell: isn't that a matter of running "git commit" and "git tag" a few times?	15:14
	Thell: i.e. should be easy enough to setup a test repo	15:15
Thell	More than a few :)	15:15
	figured git clone would be fastest	15:15
doener	Thell: for x in -alpha -beta -rc{1,2,3,4} ""; do git commit --allow-empty -m "Commit $x"; git tag v1.0$x; git commit --allow-empty -m "Work on $x"; done	15:17
	or for numbered alphas/betas: for x in -{alpha,beta,rc}{1,2,3,4} ""; do git commi	15:18
	t --allow-empty -m "Commit $x"; git tag v1.0$x; git commit --allow-empty -m "Wor	15:18
	k on $x"; done	15:18
	gah, stupid newlines...	15:18
Thell	might just be faster than google :P	15:18
	← mbroeker left	15:18
Thell	Guess I'll need to add an inner loop for each stage's commits as well... thanks doener :)	15:19
	→ patcito joined	15:19
doener	Thell: well, it does one commit after each tag, as if that isn't enough ;)	15:20
	→ reaVer joined	15:20
Thell	lol... 5 per or 10 would be more testable, or random from 5-10	15:20
	← s0ber left	15:21
	← a389742 left	15:21
	← joeytwiddle left	15:22
	← johan-s left	15:22
	→ s0ber joined	15:23
	→ harinath1 joined	15:23
	→ jlilly joined	15:24
	→ ganba joined	15:24
	← aspotashev\|eeepc left	15:24
	→ sergiodj joined	15:24
	← harinath left	15:26
	→ tstclair_ joined	15:27
	← tstclair left	15:27
	← tstclair_ left	15:28
	→ tstclair joined	15:28
	← ankit9 left	15:28
	→ [reaVer] joined	15:28
	→ Travis-42 joined	15:28
Travis-42	Can I make a patch from staged files without committing them?	15:29
c_sphere	git diff --cached	15:29
Travis-42	ah that makes sense, thanks c_sphere	15:29
c_sphere	of course that has no commit log	15:30
	← tstclair left	15:30
Travis-42	it's good enough for my purposes, thanks	15:30
	→ tstclair joined	15:30
	tstclair → tstclair_test	15:31
	tstclair_test → tstclair	15:31
	← UncleCJ left	15:32
	→ UncleCJ joined	15:32
c_sphere	p a	15:32
	→ jas4711 joined	15:32
	← giallu left	15:34
Altreus	How can I temporarily prune a commit from the history to see whether it is my problem? While keeping intact the commits since it	15:35
	→ reprore joined	15:36
teuf	Altreus: git revert ?	15:36
Altreus	teuf: hmm - really?	15:36
	to go from A--B--C--D--E to A--B--C--E while holding onto D so I can put it back later?	15:37
	← yaotti_ left	15:37
PerlJam	Altreus: git revert will commit a new rev that undoes the effect of some previous rev	15:38
jast	revert adds a new commit on top that reverses the changes in the given commit	15:38
	→ spaceonline joined	15:38
	← evl left	15:38
jast	I think actually removing the commit temporarily is much more hassle than that	15:38
Altreus	I see, then I can remove F	15:39
PerlJam	Altreus: why not use git-bisect ?	15:39
Altreus	PerlJam: It's new and scary to me	15:39
PerlJam	Altreus: excellent! You should use it.	15:39
	→ joeytwiddle joined	15:39
	← reaVer left	15:39
Altreus	>_<	15:39
	← Bidibule left	15:39
Altreus	okay	15:39
	Altreus baby steps	15:39
Altreus	PerlJam: Because there are only three revisions. I can bisect that manually :P	15:40
	→ magicrobotmonkey joined	15:41
Altreus	It's more a case of if my commit didn't do it, something somewhere is royally buggered	15:41
PerlJam	Altreus: better to learn the new and scary with an easyish task so that you can understand it better when it's harder.	15:41
Altreus	The problem is...	15:41
	I haven't changed revision since the bug started	15:41
PerlJam	but, if you're pressed for time, it's better to learn when you're not :)	15:42
Altreus	Nor have I changed the code	15:42
	so I don't have a 'good' revision	15:42
	So pulling out D above is purely conjecture anyway and might be rubbish	15:42
magicrobotmonkey	is there a bug in git-fetch where it ignores --quite?	15:43
	er, --quiet	15:43
	→ ankit9 joined	15:44
	→ ablemike joined	15:45
	→ tedoc2000 joined	15:46
	← khmarbaise left	15:47
	→ flazz joined	15:47
	→ ihoka joined	15:48
	← kpreid_ left	15:51
	→ ctusar joined	15:52
	→ sh1mmer joined	15:53
	← Thumper_ left	15:53
	→ Thumper_ joined	15:53
bentob0x	can you diff two stashes between themselves?	15:53
c_sphere	yes	15:54
	→ giallu joined	15:54
bentob0x	nice	15:54
	just tried after asking ... sorry ^^	15:54
	← tango_ left	15:55
bentob0x	<3 git	15:55
	→ corp186 joined	15:55
	→ dirk2 joined	15:55
	→ evl joined	15:55
corp186	is there anywhere on the web where people have collections of git hooks?	15:56
	hooks that other people can grab and use?	15:56
	← brendon_work left	15:58
	← galderz left	15:58
	→ kpreid_ joined	15:58
	← nevyn left	15:59
	← sde_antrix1 left	15:59
	← sverrej left	15:59
	← pflanze left	16:00
	offby1 is afraid that, if he were to grab a hook, he'd hurt his hand	16:00
offby1	they're sharp and pointy, y'know	16:00
	→ nevyn joined	16:00
	← ncopa left	16:01
	← madewokherd left	16:01
	← Aikawa left	16:02
	→ javatexan joined	16:02
	← alexyk left	16:02
	→ markoa joined	16:03
	→ khmarbaise joined	16:04
	→ mejla joined	16:04
	→ futoase joined	16:04
mejla	is there any way how to get full summary when using "git merge --log"? (Not only the first line...)	16:05
	→ bcardarella joined	16:05
	→ brendon_work joined	16:05
	→ MarcWeber joined	16:06
markoa	Hi folks. Does default scope evaluate it's conditions every time, or it's a class method that's constructed just once during Rails boot time? Like if I pass Time.now as a parameter, will it always re-evaluate that or use a single Time object?	16:06
	→ Aikawa joined	16:06
markoa	err wrong channel sorry	16:06
	← paltman left	16:07
	← sagsousuke left	16:08
	← mrtazz left	16:08
	← jarh left	16:08
	← Tanguy left	16:08
	← ayaka_t left	16:08
	← goneri left	16:08
	← tro left	16:08
	← wagle_home left	16:08
	← prh left	16:08
	← tg left	16:08
	← emias left	16:08
	← uau left	16:08
	← jrockway left	16:08
	← tswicegood left	16:08
	→ Tanguy joined	16:08
	→ mrtazz joined	16:08
	→ jrockway joined	16:08
	→ jarh joined	16:08
	→ tg joined	16:08
	→ emias joined	16:08
	→ tro joined	16:08
	← Kudos left	16:08
	← w-m- left	16:08
	← levifig left	16:08
	← Tank` left	16:08
	→ tswicegoe joined	16:08
	→ Kudos joined	16:08
	→ prh joined	16:08
	→ levifig joined	16:08
	← jefferai left	16:08
	← Pug\|Lagged left	16:08
	← scrooloose left	16:08
	← roo9 left	16:08
	← gernot left	16:08
	← jjore left	16:08
	← multani left	16:08
	← imarcust2is left	16:08
	← gionn left	16:08
	← Jippi left	16:08
	← Rhonda left	16:08
	← lfranchi left	16:08
	← ilteris left	16:08
	← ruphy left	16:08
	← Tronic left	16:08
	← lu_zero left	16:08
	← ricky left	16:08
	→ uau joined	16:08
	→ PugMajere joined	16:08
	→ w-m- joined	16:08
	→ roo9 joined	16:08
	ilteris_ → ilteris	16:08
	→ ruphy joined	16:08
	→ Tronic joined	16:08
	← prh left	16:08
	← mpr left	16:08
	← milli left	16:08
	← lamont left	16:08
	← ft left	16:08
	← ttelford left	16:08
	← zenwryly left	16:08
	→ imarcusthis joined	16:08
	→ jefferai joined	16:09
	→ lu_zero joined	16:09
	→ ft joined	16:09
	→ milli joined	16:09
	→ lamont joined	16:09
	→ wagle_home joined	16:09
	← galador left	16:09
	← Griffon26 left	16:09
	← zjason left	16:09
	← Garoth left	16:09
	← tmm1 left	16:09
	← Ilari left	16:09
	← Wraithan left	16:09
	← kyron left	16:09
	← Nafai left	16:09
	← eMBee left	16:09
	← TheUni left	16:09
	→ troyt joined	16:09
	→ Garoth joined	16:09
	→ eMBee joined	16:09
	→ Rhonda joined	16:09
	→ Jippi joined	16:09
	→ gionn joined	16:09
	→ goneri joined	16:09
	→ mps joined	16:09
	→ zjason joined	16:09
	→ kyron joined	16:09
	→ Nafai joined	16:09
	→ Ilari joined	16:09
	→ scrooloose joined	16:09
	→ multani joined	16:09
	→ gernot joined	16:09
	→ Griffon26 joined	16:09
	→ TheUni joined	16:09
MarcWeber	How to jump to a commit on the current branch which has a specific commit date such as 2 weeks ago? HEAD@{2 weeks} doesn't have to be on the same branch eg when you rebased some commits when pulling over and over again	16:09
	→ jjore joined	16:09
	gionn → Guest51722	16:09
	w-m- → Guest9259	16:09
	→ tmm1 joined	16:09
c_sphere	can use $(git rev-list --until="2 weeks ago" -1)	16:09
	→ Wraithan joined	16:10
	← wagle_home left	16:10
	→ prh joined	16:10
	→ Tank` joined	16:11
	→ maattd joined	16:11
	→ Darkas joined	16:11
engla	MarcWeber: HEAD@{2.weeks.ago} is good?	16:11
	(you may use periods instead of spaces)	16:11
	→ partha joined	16:11
	→ ayaka_t joined	16:12
	→ zenwryly joined	16:12
engla	ah ok by commit date, nevermind	16:12
	Guest51722 → gionn	16:12
c_sphere	of course by commit date. you'd get in hell if you tried to use the authoring date.	16:12
	gionn → Guest87518	16:13
engla	problem is that @{2.weeks.ago} uses the reflog date, which may be something completely different	16:13
	or not even exist (you can delete the log whenever you want)	16:13
	→ Ryback_ joined	16:13
	→ flazz_ joined	16:13
partha	whats the quick way to find all the conflicted files in git ?	16:13
	→ Hendu joined	16:13
MarcWeber	c_sphere: That does work. engla: I'm not on the main branch then. I'm on the commits I had on top of the remote that time.	16:14
	← flazz_ left	16:14
partha	I could grep ">>>>" across the tree but I am sure theres a nicer way	16:14
	→ comboy_ joined	16:14
	→ flazz_ joined	16:14
j416	partha: git status ?	16:14
engla	MarcWeber: you can use the reflog of a specific branch, like master@{yesterday} but it has the same problem --it's by reflog, not commit date	16:14
	→ galador joined	16:15
partha	j416: shows me all the modified and new files	16:15
j416	yeah	16:15
	then I don't know. sorry	16:15
	→ alezandro joined	16:15
partha	np	16:15
	thanks :)	16:15
engla	partha: git diff --cached will show "unmerged" or so for conflicted files	16:15
	← comboy left	16:15
engla	but there is a better way	16:15
	I only don't know it	16:16
	← tvw left	16:16
partha	ah cool	16:16
	lemme try that	16:16
engla	git ls-files -u shows unmerged files	16:16
	← elTigre left	16:16
	→ ricky joined	16:17
	→ krawek joined	16:17
	Guest87518 → gionn	16:18
	gionn → Guest1557	16:18
	← flazz left	16:18
	→ mjf joined	16:18
	← name left	16:19
	→ keystr0k joined	16:19
	flazz_ → flazz	16:19
keystr0k	Is there a parameter for git commit that works like -a but ONLY does deleted files... I don't want to commit a lot of new / changed files, but want to automatically commit all deleted files without having to specify them.	16:20
	→ jeckel joined	16:20
	→ lfranchi joined	16:21
	→ dreiss joined	16:22
	Guest1557 → gionn	16:22
	← jmspeex left	16:22
	gionn → Guest52669	16:23
partha	grep >>> worked better :)	16:23
corp186	how do I roll back a commit to a remote repository?	16:24
	→ bdiego joined	16:24
bremner	keystr0k, what about git add `git ls-files -d`	16:24
	→ sdwrage joined	16:24
	← saurabhverma left	16:24
sdwrage	hey all	16:24
	→ archis joined	16:24
	← bentob0x left	16:24
	→ exygen joined	16:24
wereHamster	corp186: git push -f <remote> <commit you want to reset to>:<branch you wnat to reset>	16:25
bremner	corp186, or you might want man git-revert	16:25
Gitbot	corp186: the git-revert manpage can be found at http://git.or.cz/man/git-revert	16:25
	← archis left	16:25
corp186	wereHamster: ahhh, I didn't realize the src needed to be set properly	16:25
	normally I'd use git revert	16:25
	but I'm just setting up a repo, and doing some test commits	16:26
keystr0k	bremner: well let me just try that... :(	16:26
	oops :)	16:26
wereHamster	corp186: revert creates a new commit, that push command does something completely different!	16:27
corp186	wereHamster: I know	16:27
	that's why I won't use that command once the repo is in use	16:27
bremner	keystr0k, you need to do a commit afterwards...	16:27
	→ elliottcable joined	16:27
	Guest52669 → gionn	16:28
keystr0k	bremner: I'm getting a fatal pathspec........ did not match any files.	16:28
	gionn → Guest77476	16:28
	→ jwshen joined	16:29
	← patcito left	16:31
	← dreiss left	16:31
bremner	keystr0k, hmm, maybe I misremember the output format of ls-files -d. what does a line of output from git ls-files -d look like?	16:31
	← catphish left	16:32
	Guest77476 → gionn	16:32
	gionn → Guest67437	16:33
	← j416 left	16:33
	xenoterracide → xenoterracide_	16:34
	← mejla left	16:34
	→ mejla joined	16:35
	→ patcito joined	16:36
keystr0k	bremner: looks like	16:36
	mypath/files/filex	16:37
	mypath/files/filey	16:37
	→ qhoxie_ joined	16:37
	Guest67437 → gionn	16:38
bremner	keystr0k, my bad you want git rm `git ls-files -d`	16:38
	gionn → Guest57674	16:38
keystr0k	bremner: cool. I figured I had to change that... I still get a fatal pathspec issue... could be because of a space in a filename.... hmmm	16:38
	→ wagle_home joined	16:39
	← hakunin\|out left	16:40
doener	keystr0k: git ls-files -d -z \| xargs -0r git add -u	16:41
keystr0k	doener: wowzers.	16:41
	→ hakunin joined	16:41
	Guest57674 → gionn	16:42
	gionn → Guest45734	16:43
	→ Aikawa_ joined	16:43
	← EricInBNE left	16:44
keystr0k	doener: xargs: illegal option -- r	16:44
	→ Diapo joined	16:45
keystr0k	I did just find this... looks a little hokey... for x in `git status \| grep deleted \| awk '{print $3}'`; do git rm $x; done	16:45
	→ quellhorst joined	16:45
	← kpreid_ left	16:45
quellhorst	how can I git commit --amend something 5 commits back?	16:46
doener	keystr0k: git ls-files -d -z \| xargs -0 -I{} git add -u {}	16:46
hyperair	keystr0k: you could use git clean	16:46
	quellhorst: rebase.	16:46
doener	hyperair: he wants to remove files from the index that have been removed from the working tree	16:46
hyperair	quellhorst: look into interactive rebasing.	16:46
doener	hyperair: so "git clean" won't do	16:46
keystr0k	yeah	16:47
hyperair	doener: ah sorry, i was looking at the one-liner script	16:47
keystr0k	this worked well: for x in `git status \| grep deleted \| awk '{print $3}'`; do git rm $x; done	16:47
	about to try doener	16:47
	s last suggestion	16:47
hyperair	...?	16:47
	i thought you wanted to git add?	16:47
	not git rm?	16:47
corp186	I have branches a and b, and I want to merge a series of commits starting from tag1 ending at tag2 from b to a, how do I do that?	16:47
hyperair	oh wait	16:47
	nevermind	16:48
	whoops	16:48
keystr0k	hyperair: sorry... maybe I mispoke...	16:48
	haha	16:48
	hyperair misread	16:48
	Guest45734 → gionn	16:48
doener	hyperair: "git add -u" works to stage deletions	16:48
hyperair	doener: i see.	16:48
	← wagle_home left	16:48
	gionn → Guest17570	16:48
	← wolog left	16:48
	→ wagle_home joined	16:48
keystr0k	Okay bash gurus... which is more efficient???	16:49
	git ls-files -d -z \| xargs -0 -I{} git add -u {}	16:49
	or for x in `git status \| grep deleted \| awk '{print $3}'`; do git rm $x; done	16:49
	same results	16:49
doener	keystr0k: that grep/awk monster should fail if you have spaces in your filenames	16:49
c_sphere	git ls-files -d -z \| xargs -0 git add -u	16:49
keystr0k	doener: let me test	16:49
c_sphere	preferably even xargs -0n1	16:49
	er	16:49
	-0r	16:49
	> git ls-files -d -z \| xargs -0r git add -u	16:49
doener	c_sphere: his xargs doesn't know -r	16:49
c_sphere	uh that's dumb :-)	16:50
doener	c_sphere: already had that as my first suggestion ;-)	16:50
	← ia left	16:50
c_sphere	well if you don't have -r you cannot reasonably use xargs	16:50
	← petrux left	16:50
c_sphere	or does `git add -u` at least do nothing?	16:50
doener	c_sphere: with -I{}, it also won't do anything if there's no input to xargs	16:50
c_sphere	hm.	16:50
keystr0k	git add -u adds new files... I just want to rm deleted files... hang on	16:51
	← cilly left	16:51
c_sphere	what sort of xargs is that, bsd?	16:51
doener	c_sphere: it's just crap that it will run "git add -u <path>" over and over	16:51
c_sphere	keystr0k: if you use `git rm -f` they will already be added/rm'd.	16:51
keystr0k	c_sphere: yeah, unfortunately files may be deleted via a web interface	16:51
	← Creap left	16:52
keystr0k	doener: you're right, that longer command fails on paths with spaces.	16:52
	doener: thanks for pointing that out'	16:52
	And the winner is: git ls-files -d -z \| xargs -0 -I{} git add -u {}	16:52
	Guest17570 → gionn	16:52
	← elliottcable left	16:53
doener	keystr0k: hm, given your incapable xargs, this is probably better:	16:53
	→ pflanze joined	16:53
doener	keystr0k: git ls-files -d -z \| xargs -0 git rm --cached --quiet	16:53
	gionn → Guest13117	16:53
c_sphere	rm without any file arguments is not quite so quiet either.	16:53
	← imajes left	16:53
Ilari	c_sphere: your xargs lacks -r?	16:53
keystr0k	doener: that worked too	16:53
c_sphere	Ilari: no, keystr0k's.	16:53
keystr0k	how do I check if my xargs lacks -r?	16:53
Ilari	Ah.	16:53
	→ kpreid_ joined	16:54
doener	c_sphere: hm, true. But at least it runs "git rm ..." only once ;-9	16:54
c_sphere	keystr0k: see if running xargs -r gives an error :p	16:54
keystr0k	c_sphere: illegal option...	16:54
	I'm on a Mac :/	16:54
	ehe	16:54
doener	keystr0k: 17:44:13 <keystr0k> doener: xargs: illegal option -- r	16:54
c_sphere	git ls-files -d -z \| perl -0e 'system "git", "add", "-u", @ARGV' or something.	16:54
	.. if $#ARGV> 0 ;)	16:55
doener	git ls-files -d \| git update-index --stdin --remove	16:56
	why use anything but git? bla, bla, plumbing, bla, bla... This looks like a scripting job -> plumbing ok	16:57
	c_sphere wonders if git supports n in filenames	16:57
quellhorst	I always pull and push from the master branch on github... how do I get rid of the message "You asked me to pull without telling me which branch yo"	16:57
	Guest13117 → gionn	16:58
	← giallu left	16:58
quellhorst	do I need to add branch.master.remote ?	16:58
keystr0k	thanks doener c_sphere and bremner	16:58
quellhorst	to my .git/config	16:58
	gionn → Guest3081	16:58
doener	keystr0k: haha, mistyped the command in here, must be "--remove --stdin", not "--stdin --remove"	16:59
Ilari	Idea: 'git rm --deleted'. Untracks all files that no longer longer existing in working copy.	16:59
	→ ia joined	16:59
	→ giallu joined	17:00
doener	c_sphere: it escapes that as "foo\nbar" (including the quotes), and update-index handles that output from ls-files just fine	17:00
c_sphere	nice.	17:00
quellhorst	n/mind	17:00
	→ patmaddox joined	17:00
	← quellhorst left	17:00
	← Aikawa left	17:00
doener	Ilari: hm, wouldn't that belong to "git add"? Though "git add --deleted" for sure looks funny :-)	17:00
keystr0k	looks funny but seems to make sense	17:01
	heh	17:01
	→ s0ber_ joined	17:02
	Guest3081 → gionn	17:02
	gionn → Guest66354	17:03
	jeckel → revans	17:03
	← markoa left	17:04
	← suman left	17:04
	→ ereslibre joined	17:06
	→ r11t joined	17:06
	← mejla left	17:07
	→ NET\|\|abuse joined	17:08
	Guest66354 → gionn	17:08
	← VVSiz left	17:08
	→ VVSiz joined	17:08
	gionn → Guest84971	17:08
	→ Jinzhu joined	17:09
	← vbgunz left	17:11
sitaram	doener: I need some help on how to make something happen on old gits. At least 1.5.4.3 doesn't like this: cd bare-repo; GIT_WORK_TREE=some-directory git checkout -f	17:11
	→ kraymer joined	17:12
c_sphere	sitaram: so don't use old gits while doing that ;)	17:12
	→ elliottcable joined	17:12
	Guest84971 → gionn	17:12
sitaram	doener: using git archive seems to work fine, would that be ok as a substitute?	17:12
	gionn → Guest47768	17:13
sitaram	c_sphere: I wish I could control what gits my users (running gitolite) want to use	17:13
	← kpreid_ left	17:13
doener	sitaram: to do what? Just getting a copy of HEAD^{tree} into some directory?	17:14
sitaram	doener: yes	17:14
	→ camwest joined	17:14
doener	sitaram: "git archive ... \| tar ..." would do, but you might get into tar portability hell	17:14
	→ Sigma joined	17:14
	← elliottcable left	17:15
	→ elliottcable joined	17:15
	← s0ber left	17:15
sitaram	doener: between Linux and something else or withi Linux itself?	17:15
	within*	17:15
doener	sitaram: between different tar programs (gnu, star, bsd, whatever) [and I don't actually know tar well enough to really tell which options are the bad ones)	17:16
sitaram	ok	17:16
magicrobotmonkey	can someone try the --quiet flag for git-fetch and tell me if it works for them?	17:16
sitaram	any other alternatieves within git itself?	17:16
	← mattayers left	17:16
doener	sitaram: I'd probably use a temporary GIT_INDEX_FILE and read-tree	17:17
	Guest47768 → gionn	17:17
Ilari	Also, --work-tree and --git-dir might get around the issue.	17:17
sitaram	doener: which means I'd have to make a repo on the destination side	17:17
Ilari	Not sure if that works...	17:17
sitaram	Ilari: tried those; didnt work	17:17
	gionn → Guest68785	17:17
doener	sitaram: hm? With "git checkout" you obviously need a repo, too	17:18
sitaram	doener: I have one, but it's bare... I use GIT_WORK_TREE to make it checkout to some other directory	17:18
	doener: works fine on all recent gits	17:18
	→ rudi_s joined	17:19
	→ vbgunz joined	17:19
	← bibinou left	17:20
	→ Theravadan joined	17:20
doener	sitaram: GIT_INDEX_FILE=/tmp/tmp_index GIT_WORK_TREE=/tmp git read-tree --prefix=git/ -u HEAD	17:21
	sitaram: worked fine for me, has put the stuff into /tmp/git	17:21
sitaram	doener: ok; I'll try it on 1.5.4.3 (and maybe a few others); thanks!	17:22
doener	sitaram: (of course I tried with a non-ancient git)	17:22
	Guest68785 → gionn	17:22
sitaram	:)	17:23
	gionn → Guest22896	17:23
	← magcius left	17:23
	→ robinsmidsrod joined	17:24
	→ magcius joined	17:24
doener	sitaram: of course, the /tmp/tmp_index thing should be less static (i.e. use mktemp)	17:24
	sitaram: and don't forget to drop that file afterwards	17:24
sitaram	yup	17:24
	→ justin-george joined	17:25
	→ wolog joined	17:25
sitaram	doener: thanks... this should do the trick I think	17:25
	Guest22896 → gionn	17:27
	gionn → Guest61800	17:27
	→ patrikf joined	17:28
	← justin-george left	17:29
	← markc left	17:29
	→ mapreduce joined	17:29
	→ s0ber joined	17:30
mapreduce	If you do a git checkout nameofsomebranch then a git status, why would git tell you that you're not currently on any branch?	17:30
	→ icwiener_ joined	17:30
Ilari	mapreduce: 1) It wasn't a local branch. 2) Checkout failed?	17:30
	← icwiener left	17:31
mapreduce	The branch was named origin/something - does that make it not a local branch? I still don't know why it would not report that you're on that branch.	17:31
	Would one need to create a local branch and link them or something akin to that?	17:32
	← shruggar left	17:32
	Guest61800 → gionn	17:32
	gionn → Guest27163	17:33
	→ bartel joined	17:33
magicrobotmonkey	yea origin/* are branches in the remote	17:33
engla	mapreduce: yes. git checkout -b something origin/something	17:33
bartel	is there a way to have an email sent whenever a repo is pushed to??	17:33
magicrobotmonkey	you probably want to git checkout --track origin/branch	17:33
engla	magicrobotmonkey: that's better	17:34
	← ciskje left	17:36
	← reprore left	17:36
	Guest27163 → gionn	17:37
	gionn → Guest90174	17:37
	→ reprore joined	17:37
mapreduce	--track sounds interesting.	17:37
	← ablemike left	17:38
	→ til joined	17:39
	→ GarethAdams\|Work joined	17:39
	→ jaysern joined	17:40
	← elliottcable left	17:40
magicrobotmonkey	if you come from a non-centralized source countrol, it can help ease the transition	17:40
	← nkuttler left	17:41
	← s0ber_ left	17:41
doener	magicrobotmonkey: uhm, s/non-// ?	17:41
	← evl left	17:42
	→ catphish joined	17:42
	Guest90174 → gionn	17:42
	← catphish left	17:42
	→ evl joined	17:42
	← pflanze left	17:43
	gionn → Guest49139	17:43
magicrobotmonkey	yea	17:43
	← ereslibre left	17:44
	→ kennethlove joined	17:46
	→ djszapi joined	17:47
	Guest49139 → gionn	17:47
	← RobertLaptop left	17:47
	gionn → Guest28180	17:47
	→ _Mica_ joined	17:48
djszapi	Can I change author name easily in a history ?	17:49
	apart from rebase separataly every commit	17:49
	it would be useful e.g. the author name changed meanwhile and we can see in the history that author under two names.	17:49
	it would be nice to do one instead of two author name	17:49
engla	technically there is no easier way, you have to rewrite every commit. but you can make the process automated and easy with filter-branch	17:50
	→ RobertLaptop joined	17:50
djszapi	hm ?	17:51
	← engla left	17:51
	→ jackdempsey joined	17:52
wereHamster	filter-branch with env-filter and set GIT_AUTHOR_{EMAIL,NAME}	17:52
	← bcardarella left	17:52
	→ bernardh joined	17:52
	Guest28180 → gionn	17:52
djszapi	hm, it's not a separate branch	17:52
	there is 1 master only.	17:53
	← jackdempsey left	17:53
	gionn → Guest48104	17:53
wereHamster	mayb you should take a look at what filter-branch does ...	17:54
	parasti_ → parasti	17:55
djszapi	okay, I'm looking at.	17:55
	← Escafane left	17:56
	→ malumalu joined	17:56
	→ kpreid_ joined	17:57
	Guest48104 → gionn	17:57
	gionn → Guest94738	17:57
	→ lohrun joined	17:58
	← _Mica_ left	17:59
	← evl left	18:01
	← lohrun left	18:01
	revans → revans\|away	18:02
	Guest94738 → gionn	18:02
	← capnkernul left	18:02
	gionn → Guest38716	18:03
	→ justin-george joined	18:03
	revans\|away → revans	18:03
	← justin-george left	18:03
	→ justin-george joined	18:04
djszapi	maybe I need using sed.	18:05
	git-filter-branch --env-filter "export GIT_AUTHOR_NAME='New name'; export GIT_AUTHOR_EMAIL='New email'" HEAD	18:06
	I found such an example	18:06
	but It doesn't change a given author	18:06
	← cenuij left	18:06
	→ dreiss joined	18:06
	← revans left	18:06
djszapi	http://theelitist.net/2009/10/12/git-change-revision-author	18:07
	Guest38716 → gionn	18:07
	→ elTigre joined	18:07
	← jschoolcraft left	18:07
	gionn → Guest36943	18:07
	← maattd left	18:08
	← Alien_Freak left	18:09
	→ maattd joined	18:09
	← malumalu left	18:10
	← djszapi left	18:11
	→ malumalu joined	18:11
	→ agile joined	18:12
	→ Alien_Freak joined	18:12
	Guest36943 → gionn	18:12
	gionn → Guest4569	18:13
	← ganba left	18:13
	→ flaguy48 joined	18:15
	→ tvw joined	18:15
	→ onigiri joined	18:16
	Guest4569 → gionn	18:17
	gionn → Guest84822	18:17
	→ elliottcable joined	18:19
	← froschi left	18:19
	← nighthwk1 left	18:20
	← erikvold left	18:21
	→ godsflaw joined	18:22
	Guest84822 → gionn	18:22
	gionn → Guest20610	18:23
	← dpino left	18:25
	→ imajes joined	18:25
	← RickXy left	18:26
	→ smotchkkiss joined	18:26
	← Hendu left	18:26
	Guest20610 → gionn	18:27
	→ s0ber_ joined	18:27
	gionn → Guest4516	18:27
	← futoase left	18:28
	→ RickXy joined	18:29
	← David-T left	18:31
	→ hakunin_ joined	18:31
	→ vinse joined	18:31
	← hakunin left	18:32
	Guest4516 → gionn	18:32
	→ paltman joined	18:33
	gionn → Guest89468	18:33
	← aresnick left	18:34
	→ djgera joined	18:34
	← paltman left	18:34
	← Jinzhu left	18:35
	← jlilly left	18:35
	→ jlilly joined	18:35
	← elliottcable left	18:37
	Guest89468 → gionn	18:37
	gionn → Guest16193	18:38
	← s0ber left	18:39
	← mineo left	18:39
	← bry left	18:40
	dwmw2 → dwmw2_gone	18:41
	← djgera left	18:41
	→ lhz joined	18:42
	Guest16193 → gionn	18:43
	gionn → Guest98398	18:43
	← ijuma left	18:44
lucs	Um, someone /kick gionn until he fixes his rename stuff?	18:44
	→ sverrej joined	18:45
_rane	tip of the day: turn renames/joins/parts/quits off on big channels	18:45
	→ bugfux joined	18:45
lucs	_rane: Is that something my client can do? I'll look it up.	18:45
	← sh1mmer left	18:45
_rane	probably	18:46
lucs	Ok, thanks.	18:46
Alien_Freak	how do I remove.. or change the remote location of a submodule?	18:46
_rane	lucs: you have irssi	18:46
	/ignore #git JOINS PARTS QUITS NICKS	18:46
	→ jelly-bean1 joined	18:46
_rane	should do it	18:46
lucs	Aha. Thanks again :)	18:47
	Guest98398 → gionn	18:47
	gionn → Guest97548	18:48
	← RobertLaptop left	18:48
	← sverrej left	18:48
	→ niki joined	18:48
sitaram	I have them go all into an 8-line window at the top	18:48
lucs	sitaram: You do that with irssi?	18:49
sitaram	not sure how I managed it though; irssi is still a mystery to me	18:49
lucs	Same here :)	18:49
sitaram	lucs: yes -- I can send you my config, FWIW	18:49
Mikachu	sitaram: make sure you don't include your nickserv password ;)	18:49
	→ ilteris__ joined	18:49
sitaram	Mikachu: :)	18:49
lucs	That would be nice, yes: [email@hidden.address]	18:49
sitaram	sure give me 5 min	18:50
lucs	Take your time my friend.	18:50
	→ ProperNoun joined	18:50
	← PN left	18:51
	→ qrush joined	18:52
	Guest97548 → gionn	18:53
	gionn → Guest528	18:53
	← flazz left	18:53
	→ flazz joined	18:53
	← corni left	18:53
	← kpreid_ left	18:54
	rettub_ → rettub	18:55
	Guest528 → gionn	18:57
	→ kpreid_ joined	18:57
	→ sitaram_ joined	18:57
	gionn → Guest82990	18:58
	← sitaram_ left	18:58
	→ drizzd_ joined	18:59
	→ sitaram_ joined	19:00
	← sitaram_ left	19:00
	→ resmo joined	19:02
	Guest82990 → gionn	19:03
	→ eletuchy joined	19:03
	← jlilly left	19:03
	gionn → Guest87690	19:03
	← hyperair left	19:03
	→ muszek joined	19:04
sitaram	done;	19:04
	← zoniq left	19:05
	→ tango_ joined	19:05
lucs	Thanks :)	19:05
	← alezandro left	19:07
	→ mbroeker joined	19:07
	→ jlilly joined	19:07
	Guest87690 → gionn	19:07
	→ ganba joined	19:08
	gionn → Guest65475	19:08
	→ ankit9_ joined	19:08
	→ alezandro joined	19:10
	Guest65475 → gionn	19:13
	← ankit9 left	19:13
	gionn → Guest87005	19:13
	→ sh1mmer joined	19:14
	← drubies left	19:14
	← hugoxrosa left	19:14
	→ hugoxrosa joined	19:15
	← reprore left	19:16
	Guest87005 → gionn	19:17
	→ Arrowmaster` joined	19:17
	gionn → Guest90779	19:18
	→ bambule joined	19:18
	← Arrowmaster left	19:19
	Arrowmaster` → Arrowmaster	19:19
	→ Thumper__ joined	19:19
	→ PHLAK joined	19:19
	← sh1mmer left	19:21
PHLAK	I'm looking for a way to include one repository in 2 others, is this possible? IE - I have 3 repositoried, A, B and C and I would like repo A to have repo C as part of it, and repo B to also have C as part of it but A and B are independant of each other	19:22
	← drizzd left	19:22
	Guest90779 → gionn	19:22
PHLAK	is that what a submodule is?	19:22
	gionn → Guest1462	19:22
	→ drubies joined	19:22
	← drubies left	19:22
	→ nsussman joined	19:23
	← jelly-bean1 left	19:23
	← Thumper_ left	19:25
	← ankit9_ left	19:27
	Guest1462 → gionn	19:27
	gionn → Guest54161	19:28
	→ jschoolcraft joined	19:31
	Guest54161 → gionn_	19:32
	gionn_ → gionn	19:33
	← akahn left	19:33
	→ tech13 joined	19:34
	→ bentob0x joined	19:34
	← flazz left	19:35
	→ flazz joined	19:36
	← Theravadan left	19:38
	→ ghostman joined	19:38
	→ Paraselene___ joined	19:39
ghostman	involved?	19:39
	Natch → Natch\|	19:39
	→ ijuma joined	19:40
	← Paraselene_ left	19:42
	Paraselene___ → Paraselene_	19:42
	← flazz left	19:42
	→ flazz joined	19:43
	← jayne left	19:44
	→ jayne joined	19:44
	→ firecrow8 joined	19:44
firecrow8	hi all, git submodule question	19:44
	we have a git repo which includes unit test and a deploy folder	19:45
	we only want to bring the deploy folder in as a submodule in a seperate repo	19:45
	is this possible	19:45
	→ Theravadan joined	19:45
	Thumper__ → Thumper_	19:47
	← icwiener_ left	19:47
	← khmarbaise left	19:49
	← r11t left	19:52
cbreak	you want to rewrite history?	19:52
ghostman	gentoo got a shitty root shell! ;)	19:55
	← bambule left	19:55
c_sphere	really?	19:55
	did they put \e[0;33m in front, or what?	19:55
ghostman	really!	19:55
cbreak	maybe you could clone your tree, then do a git-filter-branch --subdirectory-filter	19:55
	to get the submodule	19:55
	and do an --index-filter to remove the submodule-folder from the main repo	19:56
	but that's a total rewrite of history	19:56
ghostman	Will anybody know about POSIX shown on root shell before doing env-update on gentoo? Any answer?	19:56
c_sphere	1. Why would I need env-update, 2. why is not theirs halfway posix-compliant?	19:57
	→ bambule joined	19:59
	← ganba left	19:59
ghostman	lol	19:59
	← flazz left	19:59
ghostman	german servers are defenetly nazi-occupied!	20:00
firecrow8	cbreak, thanks that would work	20:00
	I think I'm just going to make all the testing and other stuff another repo with the subfolder as a submodule	20:01
	← tech13 left	20:01
cbreak	rewriting history is dangerous	20:02
	make backups	20:03
	and be aware that all the other repositories will get out of sync	20:03
firecrow8	thanks cbreak	20:03
cbreak	np :)	20:03
	← dirk2 left	20:04
firecrow8	btw, I'm in gentoo too, all the best with your install	20:04
	→ flazz joined	20:04
	← firecrow8 left	20:07
	← vital left	20:07
	← kukks left	20:08
	← relix left	20:08
	→ vital joined	20:08
ghostman	lol	20:09
	→ kukks joined	20:09
	→ jaysern_ joined	20:09
	← tbf left	20:09
	→ ratdeptrai joined	20:10
	→ David-T joined	20:13
	← kpreid_ left	20:13
	← aidan left	20:14
	← reid06 left	20:14
	→ tekky joined	20:14
	→ ceez joined	20:14
ghostman	hey, what`s up, funny guys?	20:15
	→ kpreid_ joined	20:16
	→ jaysern__ joined	20:17
	→ ramsey joined	20:17
	→ ruediger joined	20:17
	Leonidas1 → Leonidas	20:17
ruediger	hi	20:17
ramsey	I created a new repository to act as my "origin" with the following: mkdir foo.git; cd foo; git init ...	20:18
	tswicegoe → tswicegood	20:18
	← magicrobotmonkey left	20:18
ramsey	then I set git config receive.denycurrentbranch refuse	20:18
ruediger	I have a problem with a git repo. I killed git pull and now when trying to execute git pull again I get the following message http://codepad.org/wIOz8gJ2 how can I fix this?	20:18
ramsey	but when I try to create a new branch to switch to it (i.e. "git branch repo"), I get this error: fatal: Not a valid object name: 'master'.	20:19
	What's wrong with my repo?	20:19
	tswicegood: I decided to ask in here :-)	20:19
tswicegood	ramsey: well, I'm guessing you have an empty repo :-)	20:21
	→ PN joined	20:21
	→ jaysern___ joined	20:21
ramsey	tswicegood: obviously	20:21
	← jaysern_ left	20:21
tswicegood	ramsey: you can't create a new branch if there's nothing in your repository	20:21
	← lack\|work left	20:22
	→ gk__ joined	20:22
ramsey	tswicegood: but when I try to "git push origin master" from the remote repo (which should push to this newly created one), then I get this error: ! [remote rejected] master -> master (branch is currently checked out)	20:22
	→ _grepper joined	20:22
	→ aidan joined	20:22
ramsey	tswicegood: how can I switch to another branch on the origin if master doesn't exist?	20:22
	← gk left	20:22
	← camwest left	20:23
Ilari	ramsey: Presumably one can commit to create master and then rename the branch.	20:24
tswicegood	ramsey: it's in that should work (just did for me)	20:24
	→ jaysern_ joined	20:24
tswicegood	ramsey: you could try git push origin master:remote-master, then rename remote-master to master on your remote repo	20:25
Ilari	ruediger: 'mv .git/objects/a9/ab0dcfec2585a3840df647b51b18d5734f2fdc .git/objects/a9ab0dcfec2585a3840df647b51b18d5734f2fdc.bak'.	20:25
	→ drubies joined	20:25
	← drubies left	20:25
ruediger	Ilari: thx!	20:25
tswicegood	oh duh, that's what your receive.denycurrentbranch refush is gonna cause	20:25
	→ drubies joined	20:26
tswicegood	ramsey: why are you setting denycurrentbranch?	20:26
	← bernardh left	20:26
Ilari	tswicegood: Well, it would have been push-to-HEAD otherwise.	20:26
	ruediger: That should at least fix that error. Don't know if there are more errors in that repo.	20:26
ruediger	Ilari: at least git pull worked :)	20:27
tswicegood	Ilari: you have to have a HEAD though	20:27
	← jschoolcraft left	20:27
Ilari	ruediger: Did that file you moved away get recreated? If so, good.	20:27
	tswicegood: HEAD always exists.	20:28
	← tono left	20:28
ramsey	tswicegood: All right, so I just touched a file, added, and committed it. Then I created a new branch and switched to it. Then I was able to push from the remote.	20:28
	:-)	20:28
ruediger	Ilari: missing commit a9ab0dcfec2585a3840df647b51b18d5734f2fdc says git fsck	20:28
	→ ustun joined	20:29
tswicegood	ramsey: what are you trying to accomplish with this?	20:29
	Ilari: even in a repo that doesn't have a commit? What would HEAD point to?	20:29
Ilari	ruediger: Ugh. You have other repositories that would contain that commit?	20:29
	→ juan_arandaalvar joined	20:29
ruediger	Ilari: I could ssh to the server. Shall I copy .git/objects/a9/ab0dcfec2585a3840df647b51b18d5734f2fdc to my local repo?	20:30
Ilari	ruediger: Yes.	20:30
	ruediger: Two non-corrupt object database files with same name are always interchangeable (but .pack/.idx come in pairs).	20:31
	← Turias left	20:31
	← vital left	20:31
Ilari	tswicegood: HEAD is symbolic reference, and like symbolic links, these can be dangling.	20:32
ustun	What does --no-hardlinks option for git clone do? It says optimize the cloning process by copying files under directory. What does it optimize there?	20:32
	→ suman joined	20:32
tswicegood	Ilari: ahh, gotcha	20:33
ustun	also I don't understand the -l option, it says the files are hardlinked to save space. What does that mean? Does that mean if I delete the original repo, the clone will no longer work b/c some of its files were in the other repo?	20:33
	dwmw2_gone → dwmw2	20:33
ustun	as links?	20:33
	→ vital joined	20:34
Ilari	ustun: You know what Unix hard link is?	20:34
ustun	Ilari: not exactly	20:34
	← lorandi left	20:34
cbreak	same file, different names	20:34
	← kraymer left	20:35
Ilari	ustun: Well, nuking the original won't break the clone because it only unlinks only the name there. The name in new repo will remain.	20:35
ramsey	tswicegood: I was trying to set up an origin that I would push everything into	20:36
	tswicegood: I got it working	20:36
	← PHLAK left	20:36
Ilari	ustun: On Unix, one file can have multiple names (across directories on same filesystem). The file won't be deleted until all names have been unlinked.	20:36
ruediger	Ilari: thx git fsck stopped complaining :)	20:36
Ilari	ruediger: Good idea to have backups.	20:36
ustun	Ilari: hmm, ok I think I got it, it is like the release-retain cycle :)	20:36
Ilari	ustun: Reference count.	20:36
	→ rlarson82 joined	20:37
	← jaysern__ left	20:37
ustun	ah, yes	20:37
ruediger	Ilari: yeah.	20:37
	← jaysern___ left	20:38
ustun	now, back to my original problem: I have a folder on my Linux machine, and I created a bare repository in the Dropbox folder. Then, on my mac, I cloned the bare repo, but it complains "warning: You appear to have cloned an empty repository."	20:38
	← jaysern_ left	20:39
	← ruediger left	20:40
Ilari	ustun: The files were not transferred correctly?	20:40
	→ niki_ joined	20:40
cbreak	did you init it?	20:40
ustun	The .git directory is there, but smaller on Mac, 28 MB vs 16 MB. and when I issue "git fetch", it says "remote end hung up unexpectedly	20:40
cbreak	did you add any files to the repo?	20:40
ustun	no, I didn't init it, it was already inited in the original repo. I had added several files, committed them.	20:41
Ilari	ustun: And I remember someone had severe problems putting Git repo on Dropbox (so make sure you have backup!)	20:41
	← niki left	20:41
cbreak	hmm... is it even possible to add files to bare repos directly?	20:41
	← ProperNoun left	20:42
Ilari	It is. But not exactly easy.	20:42
	→ ProperNoun joined	20:42
ustun	http://www.yashh.com/links/2008/sep/23/using-git-with-dropbox/ I'm following this guide	20:42
	maybe the problem is with dropbox, different clients linux and mac	20:43
Ilari	ustun: By "committed" you mean "pushed"?	20:43
ustun	this guy seems to be a mac only person	20:43
	Ilari: no, I just committed to the local repo git commit	20:44
	→ hydrogen joined	20:44
cbreak	so you have a bare repository	20:45
	← NET\|\|abuse left	20:45
cbreak	how did you add files to it?	20:45
	→ jaysern_ joined	20:45
cbreak	git add won't work I think, since you have no working directory	20:45
	→ NET\|\|abuse joined	20:45
ustun	I first had a local git repository on ~/localrepo. I was using it for the last month	20:45
cbreak	the usual way is by pushing or pulling	20:45
ustun	then, I cloned it into dropbox as a bare repo	20:45
	git clone --bare ~/localrepo ~/Dropbox/remoterepo	20:46
	then, on my other computer, I did:	20:46
	git clone ~/Dropbox/remoterepo ~/localrepo2	20:46
	Not a bare repo this time	20:46
	but	20:46
cbreak	uhm... you cloned a different remoterepo	20:46
	~/ is different on the two computers	20:47
ustun	why?	20:47
	hmm yes :)	20:47
	one is /Users/ustun	20:47
	the other /home/ustun	20:47
	hmm	20:47
cbreak	~/Dropbox/remoterepo is on one computer only, if you clone from it on the other	20:47
	it won't be there	20:47
	← suman left	20:48
	← ceez left	20:48
ustun	cbreak: why? Both computers have ~/Dropbox/remoterepo	20:49
	Dropbox syncs them	20:49
	→ galderz joined	20:50
	← WebDragon left	20:51
ustun	Btw, when I first cloned as a bare repo, it said : Initialized empty Git repository ....	20:53
	even though it is not empty	20:53
	this is normal, however, right?	20:53
tpope	yes	20:53
	→ WebDragon joined	20:54
ustun	OK, one more thing: On my second computer, the one I cloned from the bare repo, the folder is empty except the git folder.	20:54
	When I do "git pull", it says "Your configuration specifies to merge the ref 'master' from the	20:55
	remote, but no such ref was fetched.	20:55
	When I do git br, no output is given	20:55
tpope	I've never done this dropbox magic but ordinarily your bare repo would look like ~/Dropbox/remoterepo.git	20:55
	← ijuma left	20:56
tpope	and then inside of it would be directories like refs, objects, etc, just like in a .git dir	20:56
	→ elliottcable joined	20:56
ustun	OK, it is like that actuall.	20:57
	actually*	20:57
	and when I clone that remote repo from my second computer, the .git dir is filled too.	20:57
	but the working dir is empty	20:57
	← jaysern_ left	20:58
	← irc2samus left	20:58
	← PN left	21:00
cbreak	can you clone from the bare directly?	21:00
	without this dropbox thingie doing magic stuff?	21:01
	→ WALoeIII joined	21:01
ustun	I can clone in the first computer	21:01
	→ henkboom joined	21:02
cbreak	so it's probably this dropbox thing that fails?	21:02
ustun	cbreak: maybe, let me rsync the bare repo to my second computer, and see whether I can clone from it	21:03
henkboom	Hi, I'm trying to import a monotone database into a git repository, but after I import it using fast-import the repo still looks bare, and "git log" outputs "fatal: bad default revision 'HEAD'".	21:03
cbreak	you can clone via ssh	21:03
	if you want to	21:03
henkboom	"du" tells me that the .git directory has grown though, is there another step I need to do after fast-import?	21:03
charon	henkboom: see what 'git branch -a' says	21:04
henkboom	charon: no output	21:04
	→ Sho_ joined	21:04
charon	hrm. it's possible that the fast-import resulted in lots of commits but no branches pointing at them. but i don't know anything about monotone, so i'm not sure what could trigger that or how to fix it	21:04
	→ RobertLaptop joined	21:05
	→ kpreid__ joined	21:06
henkboom	charon: ok, thanks anyways, I'm going to ask in the monotone channel next, it might be that my export needed to be done differently	21:06
	← ustun left	21:06
	→ ustun joined	21:08
	← hap left	21:10
	← Tuomas left	21:11
	← kpreid_ left	21:13
	← juan_arandaalvar left	21:17
	← agile left	21:17
	→ relix joined	21:18
USA_NO_1	if i modified a file but the origin master modified that file also and i cannot do a git pull, what's the appropriate next step? is there a wawy to visually see the two diffs and merge? or git clean if i want to remove my cahnges and update to master?	21:18
	pardon typos... super lag	21:18
	← ntoll left	21:18
wereHamster	USA_NO_1: git clean removes untracked file	21:19
henkboom	charon: I tried again and now I have branches :/	21:19
wereHamster	if you want to discard changes, checkout or reset --hard	21:19
henkboom	how do I make one of them be the head?	21:19
charon	henkboom: git checkout $branch	21:19
henkboom	ah, thanks	21:19
wereHamster	if you want to merge, commit first and then merge. Or you can commit and pull --rebase, or stash, pull, stash pop	21:20
	← jwshen left	21:20
charon	henkboom: note that unlike in some other scms i could mention, git's HEAD is always the current commit. it does not carry any notion of "new" or the like.	21:20
	→ bram_ joined	21:20
henkboom	so there is no one revision which is considered to be the "head"?	21:20
Ilari	henkboom: Each branch has tip. HEAD is tip of current branch.	21:21
henkboom	okay, thanks	21:21
	→ r11t joined	21:21
	→ level09 joined	21:21
bram_	hey everyone. simple question: I have a clone of a git repos, and I did some commits in that clone. Now I would like to make a patch that has all these commits (with the commit messages and everything) in order to apply it elsewhere. What's the best way to do this? I was reading about format-patch, but I'm not sending through email, nor do I want one file PER commit.	21:21
Ilari	bram_: Bundles (git bundle)? Those preserve commit IDs too.	21:22
Mikachu	bram_: so you want one patch with all the commit messages just stacked on top of eachother?	21:23
bram_	Mikachu: yes	21:23
tpope	bram_: format-patch --stdout > filename puts them in a single file. it's email like but you're not limited to sending via email	21:23
Mikachu	um, okay, try this, (git log origin/master..; git diff origin/master ) > mypatch	21:23
	← hugoxrosa left	21:24
bram_	hehehe, 3 solutions!	21:24
	HELP!	21:24
	→ icwiener joined	21:24
tpope	mikachu's gives one monster patch with a series of commit messages attached. Ilari's is a weird binary format not really suited to patches	21:24
bram_	tpope's one looks the easiet...	21:24
Mikachu	tpope: i asked if that's what he wanted and he said yes	21:25
bram_	well, I do prefer readable patches :)	21:25
tpope	you asked what he wanted not what he needed	21:25
bram_	tpope: and then to apply?	21:25
tpope	bram_: git am < filename	21:25
bram_	am = apply ?	21:25
Mikachu	tpope: there's a limit to how much i care ;)	21:26
tpope	am stands for apply-mbox (it's email like, as I said)	21:26
bram_	oh, cool	21:26
	well, ok, then only one question remains:	21:26
	what will happen if my patch gets applied to origin and then I pull again in the clone?	21:26
Ilari	bram_: You get duplicate commits (with differing IDs)	21:27
tpope	if you pull --rebase you'll get the right result	21:27
bram_	btw, I suppose you mean format-patch origin --stdout > file, right?	21:27
	awesome, that is exactly what I want!	21:27
	btw, the case is: on-server changes of a repository, but we don;t want to push from the server	21:28
Ilari	Note that rebase doesn't handle nonlinear history too well...	21:28
tpope	usually when sending patches though, I'd recommend creating them on a separate branch, and tossing that once they get accepted	21:28
bram_	well, the people doing the accepting and the people doing the editing are kind of the same set of people, so it's not that dangerous	21:28
	→ jrmuizel joined	21:29
bram_	we just really want to avoid having to push from the servers, which would mean having priv/pub keypairs there	21:29
tpope	if it's a close knit group of people I'd wager it'd be easier just to have some centralized repo hosted elsewhere where you can push your changes	21:29
ustun	cbreak: I cloned it via ssh, it worked, then I found out that there was a problem with dropbox syncing, once that was resolved, everything works now. I host my bare repo on ~/Dropbox	21:30
tpope	but your use case sounds a little strange. hard to say	21:30
cbreak	ok :)	21:30
USA_NO_1	if i am missing a particular file that the origin/master has but is not in my local clone (due to a rm) can i fetch it?	21:31
Ilari	USA_NO_1: 'git checkout origin/master -- <file>'.	21:31
	→ PN joined	21:31
USA_NO_1	werd thx	21:31
bram_	tpope, llari, ...: thanks!	21:31
	→ LRN joined	21:32
	→ jschoolcraft joined	21:32
	← bentob0x left	21:32
	→ corni joined	21:33
	← ihoka left	21:37
	→ meway joined	21:38
	→ nadavoid joined	21:38
meway	hello	21:38
	your website registration sucks balls	21:38
	← meway left	21:40
	→ Signum joined	21:40
Blackshark	how do i force overwrite my fork with the current state of the master repository?	21:41
	→ segher joined	21:41
	← ProperNoun left	21:42
Signum	Blackshark: git reset --hard ?	21:42
	Blackshark: If you mean you want to overwrite the contents of your working directory with the most recent state in your local repository.	21:42
	Signum wonders why he can't commit single hunks via git-cola any more	21:43
cbreak	just make a new branch	21:43
Blackshark	Signum: i have forkey another project from github. then checked out my fork to the working dir and added the official as the master. now then i try to merge the upstream with the master i get conflicts which i want to ignore and overwrite aggresivly	21:45
	← galderz left	21:46
Signum	Blackshark: I understand. Not sure. I've such an issue last week and git refused to do that, too.	21:47
	← brendon_work left	21:48
	← bram_ left	21:49
	→ brendon_work joined	21:49
	← brendon_work left	21:50
	← krynnotaur left	21:50
charon	Blackshark: why bother overwriting? you can just start a new branch from the upstream	21:52
	← jmcantrell left	21:53
charon	merging the upstream in this situation would be actively wrong, since a merge means you took changes from both sides	21:53
ghostman	what is fuckin git?	21:54
cbreak	ghostman: nothing. git doesn't even have a boyfriend yet.	21:56
	← javatexan left	21:57
ghostman	but a girlfriend?	21:57
	← LiamH left	21:57
	← exygen left	21:57
	ChanServ set mode: +o	21:57
	Mikachu set mode: +b	21:57
	Mikachu kicked ghostman (Mikachu)	21:57
	← smotchkkiss left	21:59
	← flazz left	22:00
tek0	odd	22:01
	→ flazz joined	22:01
	→ SRabbelier joined	22:01
	→ krynnotaur joined	22:01
Mikachu	i grepped my log and he's only ever trolled	22:02
	Mikachu set mode: -o	22:02
doener	hooray for Germans :-(	22:02
	← ehabkost left	22:02
	← malumalu left	22:02
	charon offers .ch sanctuary to doener ;)	22:03
	← Sigma left	22:03
doener	charon: maybe next year ;-)	22:04
	← bambule left	22:04
	← ctusar left	22:04
	← sdwrage left	22:06
	← blackxored left	22:06
charon	doener: on what occasion?	22:06
	apart from fleeing from trolls of course ;)	22:06
doener	charon: nothing in particular. I'll (hopefully?) finish university in January. Who knows where I'll get a job	22:07
c_sphere	doener: uh I can give you some	22:08
charon	ah	22:08
	doener: academia is nice cough	22:08
c_sphere	in academia you have to think so much -.-	22:08
charon	i consider that a good thing :P	22:08
	→ parasti_ joined	22:08
	→ thiago_home joined	22:09
	← yann left	22:09
	← maravillas left	22:09
c_sphere	Then again they are like bureaucrats.	22:09
	→ agile joined	22:09
c_sphere	Of all the important acedemic CS papers, only a few made it to RFCs, let alone widely-spread industry protocols ;)	22:09
	→ tjholowaychuk joined	22:09
	← parasti left	22:10
	← LotR left	22:10
charon	oh, there are some results that should be most important to every CS student	22:11
	← tjholowaychuk left	22:11
	→ jparrott joined	22:11
charon	for example, there's some paper on how sokoban is pspace-complete	22:11
c_sphere	hah	22:11
	"I knew that since I was a child, I had in my blood. Just looking at the time told me it's P"	22:11
doener	if it's not P = NP or P != NP, I don't care :-p	22:11
c_sphere	similarly, Atomix should be in the same class as Sokoban	22:13
Thell	doener: thanks for earlier... helped me write up a test segment for our version generator script to illustrate the output values... starting at line 379 in http://npp.pastebin.com/m632dd266	22:14
	→ Creap joined	22:14
	→ chapter_29 joined	22:14
doener	Thell: I liked my version better, was just one line ;-)	22:15
Ilari	Sokoban: PSPACE-complete.	22:15
	← qrush left	22:15
charon	the amazing bit about that paper is that they construct a level crossing	22:16
Thell	If anyone is on Win with msysgit available from cmd.exe the repo is available from git://github.com/almostautomated/npp-community.git and the script is in the v4 branch.	22:16
charon	i.e., a sokoban setup that you can only cross east-west or north-south, but not otherwise	22:16
Thell	lol @ doener, no doubt!	22:16
Ilari	Wikipedia says Atomix is also PSPACE-complete.	22:17
Thell	although, that didn't test anything ;) And one things for sure, bash > batch _everyday_ of the week in _every_ way.	22:17
	→ LotR joined	22:18
Thell	I would've written it with powershell except I couldn't expect XP users to have it installed, nor Vista users to have it updated. :/	22:18
	Thell shrugs	22:18
	← Signum left	22:18
Thell	At least it is written now, and seems to be 'fairly' robust.	22:19
level09	I'm new to GIT, just got a few questions about the workflow	22:20
c_sphere	You define the workflow mostly.	22:21
	→ Bass10 joined	22:21
Thell	I've been using it for a while and still have questions about the workflow...	22:21
level09	:)	22:21
	well I think its how I'm using it	22:21
	anyway i'll just ask	22:21
	← Bass10 left	22:22
level09	for a single user use, do I need branches ?	22:22
	isn't enought just to create commits ?	22:22
	→ Bass10 joined	22:23
	← loincloth left	22:23
	→ tchan1 joined	22:24
	→ tek joined	22:24
	→ davegb3 joined	22:24
Thell	branches are good	22:24
	← breitz left	22:24
	← joeytwiddle left	22:24
Thell	even for the single dev	22:24
	hakunin_ → hakunin	22:24
Thell	IMO	22:24
	→ djgera joined	22:24
	→ EricInBNE joined	22:25
	← djgera left	22:25
	→ patthoyts joined	22:26
level09	okay , one more question .. what do I do when I reach a stable release ? just create a new commit ?	22:26
Nugget	you have a beer and celebrate.	22:27
Thell	amen!	22:27
c_sphere	crap, now you make me going to the store just for beer :/	22:27
Thell	level09: you'll more than likely want to tag it.	22:27
	← alezandro left	22:28
Thell	level09: If you haven't found it yet via google you might want to... Pro Git	22:28
level09	Thell: I found smart git	22:29
	still i'm a command line fan	22:29
	← Guest9259 left	22:29
	→ g0bl1n joined	22:29
level09	ah	22:29
	thats a book :D	22:29
Thell	:D	22:29
	← tchan1 left	22:30
	→ tchan1 joined	22:30
	→ ozzloy joined	22:31
	← tchan left	22:32
	tchan1 → tchan	22:32
	← d0k left	22:32
	→ GPSIN joined	22:36
	← GPSIN left	22:37
	→ s0ber joined	22:38
	→ YertleTheTurtle joined	22:38
	← hydrogen left	22:39
	→ ijuma joined	22:40
YertleTheTurtle	I seem to have gotten myself stuck... I tried to checkout another branch in my local repo, and then where i tried to switch backto the original branch it says i have modified files (which I didn't modify). Running 'git reset --hard' says 'HEAD is now at .... ' but when I run git status again it still says that file is modified. Any help getting me back to my original branc would be greatly apprecated.	22:40
	← jepler left	22:40
patthoyts	You can force it: git checkout -f master	22:41
Ilari	YertleTheTurtle: Sounds like you are using msysgit and some CRLFs have gotten into the actual repository.	22:42
	→ jepler joined	22:42
YertleTheTurtle	patthoyts, will try that.	22:42
	← relix left	22:42
YertleTheTurtle	Ilari, your correct. I thought I had that all worked out between my windows and linux boxes.	22:42
	tstclair → tstclair-out	22:43
	← tstclair-out left	22:43
	← tekky left	22:44
	← til left	22:45
	← s0ber_ left	22:45
	← Diapo left	22:46
	← flaguy48 left	22:49
	→ awnstudio joined	22:50
	→ relix joined	22:50
	← drewolson left	22:51
	← DavidKlein left	22:51
	→ joeytwiddle joined	22:51
	→ Geralt joined	22:53
	← Geralt left	22:53
	← keystr0k left	22:53
	→ ArdRigh joined	22:54
	→ flaguy48 joined	22:54
	← spaceonline left	22:54
	← TeckniX left	22:55
	← agile left	22:55
	← lhz left	22:55
	→ __iron joined	22:56
	← awnstudio left	22:57
	← mbroeker left	22:58
	→ awnstudio joined	22:58
patthoyts	If I have two branches (master and dev) and following some work on master I merge onto dev, fix some conflict and commit. How can I post a patch file for this to the mailing list?	22:58
	← chapter_29 left	22:58
patthoyts	git format-patch -M master is what I'd normally use but this dumps out all the commits that were merged from master, not the fact of the merge.	22:58
	← Bass10 left	22:58
	→ adamm joined	22:59
	→ c_korn joined	23:00
	← hobodave left	23:00
patthoyts	gitweb's commitdiff seems to look about what I expect, maybe I can work out what it does.	23:00
c_korn	hello, I want to use git to backup a directory. can git make a new commit each time I change or create a file in that directory ?	23:01
c_sphere	unfortunately git does not produce mergediffs.	23:01
	patthoyts: so... it would skip the one	23:01
	→ jjlee joined	23:05
	← tom left	23:05
	→ lgonze joined	23:06
	← level09 left	23:06
patthoyts	looks like 'git show HEAD~1' might be ok. It seems to produce a diff with two columns of plusses or minuses and mentions the two parent commit ids in a header.	23:07
lgonze	Gitmasters, I am a clueless newbie, and I can't figure out why this process leaves me ahead of orgin/master by 1 commit: (1) clone (2) checkout (3) modify a file (4) git commit -a	23:07
jjlee	playing with git-filter-branch, I expected that if I "git log HEAD~3..HEAD~1" exits successfully, then so would "git-filter-branch --env-filter 'filter goes here' -- HEAD~3..HEAD~1"	23:08
	not so, though -- why?	23:08
patthoyts	lgonze: you now have one commit in your local repository that is not in the remote repository.	23:08
	So you would need to push your commit to publish it in the remote.	23:09
lgonze	my understanding of the commit is that it would have pushed it up to the remote. what am I not understanding?	23:09
	git -push?	23:09
patthoyts	git push remote	23:09
lgonze	fatal: 'remote' does not appear to be a git repository	23:09
c_sphere	man git-push	23:09
Gitbot	c_sphere: the git-push manpage can be found at http://git.or.cz/man/git-push	23:09
c_sphere	do that.	23:09
jjlee	(obviously, 'filter goes here' is replaced with a shell script that works when I don't specify the final two args)	23:09
lgonze	c_sphere: fair enough	23:09
	thx all	23:10
	← elliottcable left	23:10
jjlee	instead, git-filter-branch complains: "Which ref do you want to rewrite?"	23:12
	→ intripoon joined	23:13
	← m104 left	23:14
	← Haarg left	23:14
	→ muthu joined	23:15
	→ Octalot joined	23:16
	→ ryoma joined	23:16
	← tvw left	23:17
	← ph^_ left	23:18
	← nadavoid left	23:18
	← g0bl1n left	23:19
	→ thiana joined	23:19
	→ jmcantrell joined	23:20
	← jlilly left	23:20
	← rudi_s left	23:21
	← ph^ left	23:21
	← muszek left	23:25
	→ flazz_ joined	23:26
	← flazz_ left	23:26
	← Creap left	23:26
	→ pantsman joined	23:27
	→ javatexan joined	23:29
	← scarabx left	23:33
	→ leachbj_ joined	23:33
	→ Sigma joined	23:36
	→ Aikawa joined	23:37
	← Aikawa_ left	23:37
	→ yaotti_ joined	23:38
	← tek left	23:39
	← fualo left	23:40
	← Sigma left	23:40
	← SRabbelier left	23:41
	← flazz left	23:42
	← ustun left	23:42
	→ ustun joined	23:42
	← niki_ left	23:45
	→ niki_ joined	23:46
	→ drubies_ joined	23:48
	→ hacim joined	23:48
hacim	does anyone know how I can exclude .gitignore and .gitmodules from git-archive generated archives?	23:49
	→ s0ber_ joined	23:49
hacim	something like this in .gitattributes: Set export-ignore .gitignore	23:50
	→ yaotti__ joined	23:54
	← corni left	23:55
	→ sycofly joined	23:57
	← cannonball left	23:58
	← patrikf left	23:59
	→ bcardarella joined	23:59

Logs Search ←Prev date Next date→

Channels Documentation