IRCloggy #git 2009-02-02

Provider of IRC logs since 2005.

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

	2009-02-02
	→ jrmuizel joined	00:00
	← Sigma left	00:00
	→ Flathead joined	00:01
	→ ben_h joined	00:02
	→ seangrove joined	00:03
	→ moon_unit joined	00:03
moon_unit	is there a way to check out a single file of a repository?	00:04
	→ caoliver joined	00:04
moon_unit	and	00:04
	set it up as a submodule	00:04
offby1	doubt it.	00:05
kevlarman	currently, no	00:05
	you can only clone the whole thing	00:05
offby1	submodle == commit, more or less	00:05
	s/submodle/submodule/	00:05
	→ plasticine joined	00:08
shd	moon_unit: a limited form of cloning can speed up the process, see shallow clone	00:08
	→ jmspeex joined	00:09
moon_unit	so i'm basically just trying to share a bunch of config files between projects	00:09
	← caoliver left	00:10
moon_unit	they're all in different directories in my projects	00:10
	is there any way to do that without setting up one repository per file?	00:10
	like say, config/s3.yml = shared_repo/s3.yml, lib/rake = shared_repo/rake, etc	00:12
	what's the cleanest solution	00:13
sshc	I have about 90 useless whitespace commits just to test server-side hooks. how do I erase those 90 commites from history?	00:13
Pieter	rebase -i	00:13
	or filter-branch	00:13
	or, if they're the top commits, reset --hard	00:14
	PYP	00:14
	← plasticine left	00:16
	← icwiener left	00:16
	← iskaldur left	00:18
	→ zaius joined	00:21
	→ lisyzavu joined	00:21
	→ plasticine joined	00:23
	← _Vi left	00:24
	→ dulia joined	00:27
	→ carllerche joined	00:27
	→ name joined	00:28
	← gwoo left	00:30
	→ fynn joined	00:31
fynn	What does the following error message mean:	00:31
	warning: 2 lines add whitespace errors.	00:32
dsop	fynn: that you have e.g. 2 times trailing whitespaces	00:32
	→ EmilMedve1 joined	00:33
doener	if you know which commit that was, you can use "git diff --check $commit^!" to see where the errors are	00:34
jayne	why is trailing whitespace at the end of a line bad?	00:34
doener	because it's useless and people might have setup their editor to automatically strip it	00:34
	leading to a lot of noise and totally useless conflicts	00:35
	you can configure which whitespace errors git complains about	00:35
jayne	why would anyone want their editor to make unsolicited changes to a file?	00:35
	→ reprore_ joined	00:35
fynn	dsop, doener: thanks.	00:35
dsop	jayne: because they are smart enough to know that whitespaces can cause unecessary trouble	00:36
	← EmilMedve left	00:36
jayne	hm. This seems to me like a case of the solution being its own problem	00:36
	→ pinholecamera joined	00:36
doener	jayne: even if it's not automatic, it can still cause problems. I might edit a line, and later change it back, not forget to add the useless whitespace	00:37
	← fynn left	00:37
pinholecamera	hey folks, i'm trying to build git on debian etch from source, brand new debian install... for some reason after i do ./configure (which appears successful), there is no make command, just says it's not found... any ideas why that might be?	00:37
Whtiger	it says make isn't found?	00:37
jayne	that doesn't seem like a problem that would be restricted to end-of-line whitespace. What about mid-line whitespace?	00:37
	→ plasticine__ joined	00:38
pinholecamera	Whtiger: yeah, it's really weird!	00:38
jayne	"if(" vs "if (", (x+y)/2 vs (x + y) / 2, etc.	00:38
dsop	jayne: go and fix your coding styles	00:38
doener	jayne: or two devs change code that is nearby, one replaces a line with "\t" with two empty lines (two separate his new code block from the surrounding code), the other keeps the \t	00:38
pinholecamera	Whtiger: says command not found, it's wild... never encountered that on a successful ./configure before	00:39
doener	jayne: that's coding style, and one way to do it should be enforced anyway. It's not useless but gives consistency to the code	00:39
pinholecamera	i've got the output in a pastebin here: http://pastebin.com/m3d50cdd2 ... looks a-okay to me	00:40
doener	pinholecamera: configure doesn't check for make IIRC, do you have make installed?	00:41
pinholecamera	doener, it's a default debian etch install, it should be installed by default, or at least was in past experiences for me	00:42
	but let me check	00:42
	← eddyp left	00:42
pinholecamera	hrmm, not seeing it so far, do you know the name of the package off the top of your head?	00:44
	← plasticine__ left	00:44
doener	make	00:44
pinholecamera	ahh thought there was more to it. damn, not installed, that'd do it, haha.	00:45
	weird that it's not installed on this machine, never run into that before	00:45
	thanks doener	00:47
	→ firstdate joined	00:48
	← jmspeex left	00:48
	→ jmspeex joined	00:49
	→ Dodji joined	00:49
	→ fynn joined	00:49
	→ plasticine__ joined	00:50
	← plasticine left	00:51
	→ wagle joined	00:51
fynn	doener: btw, is there a way to find the offensive commit if I don't know which it is?	00:51
	← firstdate left	00:51
	→ firstdate joined	00:53
	← radarek left	00:55
doener	fynn: I guess you could loop over a commit range and do the diff --check for each commi	00:57
	t	00:57
	← torarne left	00:58
fynn	doener: looping in something like a bash script, I assume?	00:58
	EmilMedve1 → EmilMedve	00:58
	← plasticine__ left	00:59
doener	yeah, for commit in $(git rev-list ....); do echo $commit; git diff --check $commit^! ; done	00:59
	or whatever	00:59
	← neunon left	00:59
doener	adding some checks to handle merge commits in a better way would be good I guess	01:00
	→ troter joined	01:01
Whtiger	HI	01:01
	er	01:01
	Irssi should tell me if I'm in the wrong channel	01:02
jayne	normally it does tell you your channel in the status line	01:02
	→ plasticine joined	01:02
fynn	doener: cool, thanks.	01:05
	← sgrimm left	01:06
	→ sgrimm joined	01:06
	← carllerche left	01:07
	→ carllerche joined	01:09
	← plasticine left	01:11
	← Niamor left	01:12
	← name left	01:12
	→ Niamor joined	01:13
	← carllerche left	01:14
	← fynn left	01:14
	→ plasticine joined	01:15
	← sshc left	01:20
	← firstdate left	01:20
	← bobesponja left	01:23
	← plasticine left	01:24
	→ plasticine joined	01:26
	← seangrove left	01:30
	→ seangrove joined	01:31
	→ nothingHappens joined	01:31
	← plasticine left	01:32
	← nothingHappens left	01:32
	→ therrg joined	01:32
	← travisjeffery left	01:33
	→ gretch joined	01:33
	← Ariens left	01:36
	← seangrove left	01:37
	→ seangrove joined	01:37
	← seangrove left	01:37
	← aziz left	01:41
	→ plasticine joined	01:41
	← parasti left	01:43
	← Niamor left	01:43
	→ Niamor joined	01:44
	→ mithro joined	01:46
	→ parasti joined	01:46
	← plasticine left	01:47
	→ nick_h[litage] joined	01:53
	→ plasticine joined	01:54
	→ qrush_ joined	01:56
	→ jchris joined	01:59
	← plasticine left	02:00
	→ kelvie joined	02:02
	← krh left	02:04
	← cgardner left	02:09
	← nick_h[litage] left	02:09
	→ nick_h[litage] joined	02:10
	← reprore_ left	02:11
matt	dumb question	02:11
	so i created a branch, changed the file	02:11
	→ plasticine joined	02:11
matt	i want to merge the branch back into master and commit to master (its already commited on the branch)	02:12
	← Flathead left	02:12
matt	how?	02:12
qrush_	git checkout master	02:12
	git merge otherbranch	02:12
deskin	matt: git checkout master && git merge thebranch	02:12
matt	ah	02:13
	ok thanks guys	02:13
	← johnw left	02:14
matt	do i need to recommmit after a merge?	02:14
kevlarman	no	02:14
deskin	only if you have to manually resolve merge conflicts	02:14
matt	nope. ok great thanks :)	02:15
	→ j416 joined	02:16
qrush_	matt: just check git log, you'll see the commit you merged	02:18
	← foca left	02:19
	← jerbear left	02:21
	← nick_h[litage] left	02:22
	→ nick_h[litage] joined	02:22
	→ fynn joined	02:24
	← plasticine left	02:27
	→ plasticine joined	02:28
	← dulia left	02:28
	← parasti left	02:28
	← GeertB left	02:34
	→ caseym joined	02:35
	→ GeertB joined	02:35
	← kukks left	02:35
j416	Does anyone have a neat way of storing branch descriptions? This has been discussed on the mailing list, but quite some time ago it seems. Keeping it in-branch kind of hinders merging (branch descriptions would fly all over the place)	02:36
	← ndim left	02:37
	→ ndim joined	02:37
	← plasticine left	02:38
	← toxx_ left	02:40
	→ plasticine joined	02:40
	→ mark[oz]_ joined	02:43
	← mark[oz]_ left	02:47
	← Niamor left	02:48
	→ Niamor joined	02:49
	→ mark[oz]_ joined	02:50
	← plasticine left	02:52
	→ scientes joined	02:54
	→ plasticine joined	02:54
fynn	OK, I have a weird situation here:	02:54
	"modified: foo/bar/baz"	02:55
	→ bondjuan joined	02:55
fynn	in `git status`. However, foo/bar/baz is a directory.	02:55
	Also, doing `git checkout foo/bar/baz` returns no feedback, but also doesn't change anything.	02:55
bondjuan	Hey guys, I'm kinda confused on how I can checkout a remote branch and pull it down to create a local branch on my computer?	02:55
fynn	What's up, and how do I fix it?	02:55
bondjuan	Is it considered a "tracking" branch?	02:56
	→ iskaldur joined	02:56
	← paulboone left	02:57
iskaldur	how good is the git eclipse plugin?	02:57
	← mark[oz] left	03:00
	← caseym left	03:01
Ilari	fynn: foo/bar/baz is submodule?	03:02
	bondjuan: Create local branch starting from apporiate remote tracking branch and the check that branch. There is shortcut for that: 'git checkout -b <newbranchname> <trackingbranch>'.	03:03
fynn	Ilari: possibly... I'm not even sure how to check that.	03:05
bondjuan	Ilari: thank you :)	03:06
	→ patmaddox joined	03:09
	→ bryanl joined	03:09
	← iskaldur left	03:09
	← plasticine left	03:09
	→ plasticine joined	03:11
	← Bass10 left	03:12
	← mark[oz]_ left	03:13
	→ mark[oz] joined	03:14
	← plasticine left	03:20
	→ cytrinox joined	03:20
	← schierbeck left	03:23
	→ martinpaul_ joined	03:23
	→ Flathead joined	03:23
gouki	git config --global user.name is user-based, right? I mean, if user A does it, it won't overwrite for user B on the same machine.	03:24
Ilari	fynn: 'foo/bar/baz/.git' exists?	03:24
	gouki: Correct (system-global would be --system).	03:24
deskin	gouki: you're correct; --global modifies ~/.gitconfig	03:24
gouki	deskin and Ilari thank you very much.	03:25
Ilari	gouki: And for obivious reasons, --system likely requires root...	03:25
	gouki: For setting options, that is.	03:25
fynn	Ilari: yes it does!	03:25
	← cytrinox_ left	03:25
Ilari	fynn: Then foo/bar/baz is a submodule.	03:26
	fynn: You can use 'git add' to stage submodule changes as usual...	03:26
	← wagle left	03:26
	→ plasticine joined	03:27
fynn	Ilari: what I actually want is to revert all changes.	03:29
	→ ben_h_ joined	03:32
	→ akitada_ joined	03:32
	← ben_h left	03:33
	ben_h_ → ben_h	03:33
	ben_h → ben_h_underscore	03:33
	← plasticine left	03:35
	← ben_h_underscore left	03:36
	→ ben_h joined	03:36
	← flaguy48 left	03:36
	← markelikalderon left	03:37
	→ mark[oz]_ joined	03:38
	→ mark[oz]__ joined	03:38
	→ girishr joined	03:39
	← mark[oz] left	03:39
	→ mark[oz]___ joined	03:39
	→ plasticine joined	03:40
	← martinpaul left	03:40
	→ ben_h_ joined	03:42
	← ben_h_ left	03:42
	← ben_h left	03:42
	→ ben_h joined	03:43
	← ankit9 left	03:45
	← plasticine left	03:48
	← nick_h[litage] left	03:50
	← mark[oz]_ left	03:55
	→ plasticine joined	03:57
	→ perezd88 joined	04:00
	→ jackdempsey joined	04:01
	→ dulia joined	04:04
	← thumper left	04:05
	→ thumper joined	04:06
	← jackdempsey left	04:06
	→ spearce joined	04:06
	→ ankit9 joined	04:06
	← mark[oz]__ left	04:06
	← akitada_ left	04:07
	← alezandro left	04:08
thom__	So I made a bunch of changes (added tons of files) to an archive on another machine, pushed the changes to the master repository on this machine, and now when I do git status it says all the newly added files are deleted. What do I need to do? git merge?	04:08
	→ travisjeffery joined	04:09
thom__	or git checkout to grab the new files?	04:09
Arrowmaster	thom__: git faq non-bare	04:09
Gitbot	thom__: Pushing to non-bare repositories is discouraged. See http://git.or.cz/gitwiki/GitFaq#non-bare	04:09
	← fynn left	04:10
	→ fynn joined	04:12
	→ bwwww joined	04:12
bwwww	hi guys. what man page explains the tree-ish notation?	04:13
thom__	hmmm, so what should I do in the future? a git pull from the master archive machine?	04:13
Arrowmaster	yeah	04:14
kevlarman	bwwww: probably git rev-parse	04:14
bwwww	kevlarman, yea, thats the one	04:15
thom__	is pushing to a branch safe?	04:15
bwwww	thanks	04:15
	← plasticine left	04:15
	← spearce left	04:16
	→ plasticine joined	04:18
	← girishr left	04:18
	← bwwww left	04:19
	← dulia left	04:19
	mark[oz]___ → mark[oz]	04:20
	→ Marmouset joined	04:23
	← plasticine left	04:23
fynn	basically, you're not supposed to push to non-bare repos, because you may be creating conflicts for the person working on them?	04:25
	← cthompson left	04:26
	← bryanl left	04:27
kevlarman	fynn: you're not supposed to push to non-bare repos because if you push into the checked out branch you'll trash the index	04:28
	→ plasticine joined	04:30
thom__	trash the index? I just did a git reset --hard because I didn't have any changes on this side so I think things are okay, but.... if I had just a git commit on this side it would have simply shown all the new files I had added as deleted, right?	04:31
kevlarman	not on your end, the other end	04:32
	→ Hydrogen joined	04:33
thom__	when I did a git status on the "pushed too" side it said all the new files i added on the other machine were deleted. Makes sense, I hadn't checked them out, there weren't there, so a commit would have just have recorded them as dleted in the HEAD, no?	04:35
	I think I resolved this correctly with the git reset, just trying to understand as a "learning moment"	04:36
	← plasticine left	04:39
kevlarman	thom__: see faq non-bare	04:39
Gitbot	thom__: Pushing to non-bare repositories is discouraged. See http://git.or.cz/gitwiki/GitFaq#non-bare	04:39
	→ spearce joined	04:39
	→ plasticine joined	04:43
	→ girishr joined	04:43
	← troter left	04:44
	→ jony_ joined	04:44
	→ bryanl joined	04:44
	→ troter joined	04:47
	← plasticine left	04:48
	→ aboodman joined	04:49
aboodman	Hello all, I realize preserving timestamps is a sore subject, and I've read the FAQ and http://kerneltrap.org/mailarchive/git/2007/3/1/240167. But I was wondering -- if I wanted to hack git to restore the last modified time on checkout, is that data already stored in the repository -- or would I have to store it?	04:52
	→ neunon joined	04:54
	→ plasticine joined	04:55
	← doener left	04:56
kevlarman	aboodman: wouldn't that screw badly with things like make?	04:56
offby1	aboodman: I assume you'd have to store it.	04:56
	dunno where	04:56
	→ sshc joined	04:57
aboodman	kevlarman: I have modified my build to plaace the build results in different directories for each branch	04:57
	so, no	04:57
sshc	I have a hook with private information (3 passwords) on a gitosis hosted git repository. Can a normal client see the hook?	04:58
kevlarman	aboodman: in that case, no, git only stores commit times	04:58
Ilari	sshc: No.	04:58
	sshc: Assuming only git:// or ssh:// (through gitosis) access...	04:58
	→ doener joined	04:58
sshc	ok	04:59
	→ cedricv joined	04:59
aboodman	thanks kevlarman	04:59
	← bryanl left	05:00
Ilari	aboodman: There are really two places you could store it: 1) In the commit object or 2) in some special file in tree. Even if one could define new commit header that indirects to listing of modification times, this isn't good idea because of reachability analysis wouldn't get it right...	05:01
	→ sverrej joined	05:01
	← cawel left	05:02
	→ tbrx joined	05:02
	← plasticine left	05:04
	→ zjason joined	05:04
deskin	I've been idly thinking of reworking git-svn to put git-svn-id: lines in git notes objects, but I wonder if this would be down the path of using notes for all commit metadata which doesn't really belong; wonder what using it for mtimes would look like	05:06
	→ plasticine joined	05:06
	→ SRabbelier joined	05:07
	← ekidd left	05:09
	→ tjafk1 joined	05:10
	→ schlort_ joined	05:12
	← plasticine left	05:12
	← Niamor left	05:13
	→ psankar joined	05:14
	→ Niamor joined	05:14
	← bondjuan left	05:14
	→ plasticine joined	05:16
	→ CardinalNumber joined	05:17
	→ ashleyw joined	05:19
	← perezd88 left	05:21
	schlort_ → schlort	05:23
	→ zobie joined	05:25
	→ plasticine__ joined	05:26
	← tjafk2 left	05:27
	← jrmuizel left	05:27
	← ProperNoun left	05:28
	→ jrmuizel joined	05:28
	→ perezd88 joined	05:29
	← perezd88 left	05:30
	← jrmuizel left	05:31
	← plasticine left	05:34
	→ bobesponja joined	05:35
	← travisjeffery left	05:36
	← plasticine__ left	05:37
	← fynn left	05:38
	← Niamor left	05:38
	→ jrmuizel joined	05:38
	→ Niamor joined	05:39
	→ ToxicFrog joined	05:40
	→ plasticine joined	05:40
	← schlort left	05:43
	→ fynn joined	05:45
	→ joshdavey joined	05:49
	← girishr left	05:50
	→ girishr joined	05:51
	← Hydrogen left	05:52
	← Beket left	05:53
	← jrmuizel left	05:56
	→ jrockway joined	05:56
	← bdimcheff left	05:56
	← zaius left	05:59
ieure	Okay, so is it me, or is git-svn just an utterly worthless piece of shit?	06:00
	Every time I’ve tried to use it, it’s been a massive pain in the ass, constantly doing the wrong goddamn thing and fucking up my repos.	06:00
Ilari	ieure: More like it does its best with SVN, but SVN is just too much... :-)	06:01
	← plasticine left	06:01
ieure	Ilari, I’m no fan of SVN, but the level of retarded bullshit I have suffered with Git is _nothing_ compared to SVN.	06:02
	I just dcommited some shit, and git-svn decided to commit it to a tag for some fucking reason.	06:02
kevlarman	because you had a tag checked out	06:03
	(or you merged a tag, which the man page tells you not to do for this very reason)	06:03
	→ plasticine joined	06:03
ieure	kevlarman, Nope. Unless git-svn was retarded enough to check out a tag for me.	06:04
	I did: "git svn clone --stdlayout $SVNPATH"	06:04
	Then worked.	06:04
	Then dcommitted.	06:04
	Did my work on branches, and specifically did "git checkout master" before I dcommitted.	06:05
	← curvature left	06:06
	→ curvature joined	06:06
	→ k776 joined	06:09
deskin	ieure: git-svn checks out whatever the most recent commit in svn was initially, wherever it may be	06:10
ieure	deskin, That qualifies as retarded.	06:10
	← k776 left	06:10
deskin	patches welcome	06:11
ieure	Where is that documented?	06:11
	← plasticine left	06:12
deskin	not sure it is documented in any docs which ship with git	06:12
	→ wagle joined	06:12
ieure	Awesome! So I quite possibly get a repo for a different branch/tag of my SVN repo every time I clone it with git-svn, and it’s undocumented behavior?	06:13
	Sign me the fuck up.	06:13
	→ reithi joined	06:13
	→ ludde joined	06:14
deskin	ieure: it's a bug if the git repo is different between clones; but yes, you'll get a different branch checked out. Of course, you're then able to checkout the branch you want to work on	06:15
	deskin sleep	06:16
DrNick	I take it you get the most recent SVN revision?	06:16
	← sshc left	06:16
ieure	I don’t really grok the Git lingo, but what I’m saying is: The files you get are for a random SVN path when you clone.	06:17
	Which seems to be what you’re saying. And I just cannot believe that anyone considers that helpful in any fucking way.	06:17
DrNick	I wasn't talking to you, I was talking to deskin	06:18
	→ gwoo joined	06:18
DrNick	you seem to have an attitude problem and aren't worth interacting with	06:18
ieure	My attitude problem is caused by the endless grief Git causes me every time I try to use it.	06:19
DrNick	then don't	06:19
	← davidfetter left	06:22
qrush	HEY DRNICK	06:22
	I HATE THE SOFTWARE THAT YOU SHOULD KNOW EVERYTHING ABOUT	06:22
	can't answer my questions? too bad, i'm gonna be an asshole anyway	06:22
	excuse me while i go beat my wife and throw my monitor out the window	06:22
	→ speakman joined	06:22
speakman	hi folks!	06:23
	← bdiego left	06:23
	→ plasticine joined	06:25
qrush	heyo	06:25
	→ a-priori joined	06:26
	→ ph^ joined	06:26
	→ travisjeffery joined	06:27
	→ Weaselweb joined	06:27
	→ muthu_ joined	06:28
	→ dulia joined	06:29
	← agenteo left	06:30
	→ iulian joined	06:31
	→ moriyoshi joined	06:32
	→ the6step joined	06:32
	← bobesponja left	06:33
	← plasticine left	06:33
	→ mansour joined	06:39
	→ plasticine joined	06:40
	→ robinr joined	06:45
	← plasticine left	06:45
	→ Tuomas joined	06:46
	← Tuomas left	06:50
	→ Tuomas joined	06:51
	dwmw2_gone → dwmw2	06:55
	→ plasticine joined	06:57
speakman	There's a thing with Git I do not really figure out how to handle; I have my very special patches in the kernel which I maintain myself. I've cloned the kernel source via Git and made my own branch (ts-2.6). But how do I now make sure everything new happening at my clone source to be merged/pulled into my own branch?	06:58
j416	speakman: git pull ?	06:59
	→ blarz joined	07:02
	← blarz left	07:02
speakman	will it by default pull the HEAD of the remote repo?	07:03
	← Aikawa left	07:03
j416	yeah	07:06
	← plasticine left	07:07
j416	it will update your copy of the remote repo	07:07
	then merge with it	07:07
	← a-priori left	07:08
	← neunon left	07:10
	← girishr left	07:10
	→ plasticine joined	07:10
speakman	is there any point with making my own branch named "master"?	07:10
	→ ashH joined	07:10
	→ moore_ joined	07:10
j416	how do you mean?	07:11
	you can name it to whatever you want. :)	07:11
	← mark[oz] left	07:11
speakman	HEAD is still what's count?	07:11
j416	I am not sure what you mean	07:12
	personally I try to keep all commits common to all branches in my master branch	07:12
speakman	I'm new to Git, there's alot of things to figure out, sry :)	07:12
j416	that way i can easily merge in master to any branch and get all "global" updates	07:12
	speakman: :)	07:13
	i have only used it for just over a month myself	07:13
	happy learning!	07:13
	HEAD is just a reference to the tip of the current branch	07:13
speakman	The Right Way <tm> would have been to NOT branch the master but just applying my patches to it and keep developing in that branch?	07:13
j416	it points to a commit	07:13
	→ Phlogi joined	07:14
j416	speakman: you could develop using only one branch, yes	07:14
Phlogi	I onced forked a git repot but just copying files and pushing them to another repo. Now I would like to get some changes from the original repo back into my forked version. Whats the easiest way? Can I import just some commits?	07:14
	← troter left	07:14
speakman	E.g; "git clone git://some.repo/kernel.git" and then just apply my patches and keep developing?	07:14
j416	but as you learn more, you will see that keeping multiple branches is very convenient.	07:14
speakman	Is there any reason to use branches even when not trying something "new" out?	07:15
j416	yes	07:15
	maybe you have a separate version of your project with only slight modifications	07:15
	that could be a separate branch	07:15
speakman	Where should critical bugfixes go to then?	07:16
j416	for example, I keep a debug branch that is in sync with the dev branch, but which contains code that spits out debug info	07:16
speakman	oh! nice point (y)	07:16
j416	a nice way to fix a bug would be to branch off, fix the bug in its own branch, test that it works, then merge it in again	07:16
	→ priidu_ joined	07:17
	← fynn left	07:17
j416	that way you could continue working on a functional project version if you, say, can't fix the bug directly	07:17
	← plasticine left	07:19
	← moore__ left	07:19
speakman	should I delete the branch when it's been merged?	07:19
j416	if you see no other use for it, I don't see a reason to keep it	07:19
	→ ceyusa joined	07:19
j416	(it would be troublesome to end up with hundreds of small branches in the end)	07:19
speakman	sounds true :)	07:20
j416	when you merge it, its history will be recorded in the branch you merge with anyway	07:20
	so you don't lose anything	07:20
	← robinr left	07:22
Phlogi	so I would add the original forked git repo as a local branch with git remote add foo foo_url ?	07:23
	→ plasticine joined	07:24
	← madewokherd left	07:24
	← spearce left	07:24
	→ arlenik_ joined	07:25
Phlogi	and now I fetched that... so how can I diff those two branches for example or how can I just merge single commits?	07:25
arlenik_	I added files with 'git add *.o', now how do I remove them?	07:25
ruskie	git rm *.o ?	07:26
j416	no	07:26
	← speakman left	07:26
	→ speakman_ joined	07:26
j416	git reset *.o	07:26
speakman_	(sorry, lost network connection)	07:26
j416	git status tells you this	07:26
	→ giallu joined	07:26
ruskie	ow yeah...	07:26
	I keep forgetting that bit	07:26
j416	git rm would delete your file from working dir aswell	07:26
	and from the repo	07:26
	← giallu left	07:26
ruskie	only if it's git rm -f isn't it?	07:27
j416	git reset resets your index	07:27
	→ giallu_ joined	07:27
j416	git rm removes the file from git alltogether	07:27
arlenik_	I get an error 'main.o has changes staged in the index'	07:27
j416	(or rather, records it as deleted)	07:27
Phlogi	I mean I don't want to merge the whole branches... just specific commits? Is that possible?	07:27
arlenik_	i only want to remove it from the index	07:27
	← giallu_ left	07:27
ruskie	Phlogi, cherry pick	07:27
DrNick	arlenik_: git rm --cached	07:28
ruskie	Phlogi, git cherry-pick --help	07:28
Phlogi	ok thanks	07:28
	→ plasticine_ joined	07:28
Phlogi	ruskie: does this hash value specifiy the commit completely or do I need to tell the branch too?	07:29
arlenik_	DrNick: i get usage error for git rm --cached	07:29
	nvm	07:29
ruskie	hashes are supposed to be unique all over the place from what I understand...	07:29
arlenik_	I have uncommented '*.[ao]' in my info/exclude file but git still tells me things like main.o are untracked	07:30
Phlogi	ruskie: thanks its works already :)	07:32
j416	Phlogi: the hashes are unique	07:33
Phlogi	nice :)	07:33
	and if I want to update my remote branch I do just git fetch foo bar again?	07:33
	→ bentob0x joined	07:33
j416	(otherwise git wouldn't work :))	07:33
	fetch yes	07:34
	→ girishr joined	07:34
	→ giallu joined	07:34
Phlogi	jup	07:36
	j416: how many of those hashes exist? :D its a hex value right?	07:36
j416	more than enough :)	07:36
	google "sha1" for info	07:37
	→ troter joined	07:37
	→ sandGorgon joined	07:37
	← troter left	07:38
arlenik_	I don't get it. why do i need to add stuff to the index? why can't I just make my changes and do commit?	07:39
DrNick	I don't get it. Why can't I make partial commits of only the changes I want to make right now? Why do I have to commit everything at once?	07:40
	→ troter joined	07:40
arlenik_	amusing	07:40
	← johan-s left	07:40
	→ shapeshed joined	07:41
	← plasticine left	07:41
arlenik_	so adding stuff to the index is like mini-commits?	07:41
DrNick	you stage things for commits, yes	07:42
arlenik_	does the index get cleared after each commit?	07:43
DrNick	the index is basically a flattened view of the entire working tree, so it doesn't it cleared, it gets updated to match (or, rather, you're creating a new commit from the contents of the index)	07:44
j416	arlenik_: the index contains everything that will be committed on the next commit	07:45
	→ au- joined	07:45
j416	when git status reports that your working dir is clean, it means everything matches your index	07:46
	when you alter a file, your working copy will be different from the copy stored in the index, so you update the index with git add	07:46
	every time you do git add <your file>, your file is copied to the index	07:47
DrNick	(more accurately, a new blob is created for the file contents, and the index is updates to point to that blob)	07:47
j416	and files (or individual changes) you don't add after editing them will simply not be included in the next commit	07:47
	← Casan left	07:48
j416	so at any time, when your working directory is reported to be clean, or you have files that are staged for commit	07:49
	it means git has stored a copy of that data	07:49
	untracked files and unstaged changes are not in git at all.	07:49
	what happens when you do a commit, is that a pointer is created to that data, and it is stored as part of your history	07:50
	simply explained.	07:50
	arlenik_: still there? :)	07:51
arlenik_	yes, i'm reading	07:51
	it's making more sense now	07:51
	so, how would i tell git to untrack a file that it's tracking right now?	07:51
j416	so say you only want to commit changes you made to file A, because the changes you have made to file B are not ready, and besides, they are totally unrelated to the changes in file A	07:51
	then you might choose to only stage A and commit	07:52
	→ naeu joined	07:52
j416	arlenik_: git rm <file> will remove the file from index, and from your working dir	07:52
	and in the next commit, it will be marked as deleted	07:53
	if you do git rm --cached, it will not touch your working copy, but still remove it from the index	07:53
	(cached = index)	07:53
arlenik_	ahh, okay. i see.	07:53
j416	and it will show up as untracked	07:53
	→ kervel joined	07:55
arlenik_	j416, DrNick, you've been a great help. Thank you for your time. I've got to go have fun with git now :-)	07:55
	← travisjeffery left	07:55
	← arlenik_ left	07:55
kervel	hello, we have a big project with lot of branches, and we need to keep track of "which commits need merging to master". with "git log" it works fine, except for commits that were already cherrypicked into master, they remain in the log. any solution ?	07:56
j416	you're.. welcome	07:56
	← therrg left	07:56
j416	:)	07:56
kervel	i found this comes close: http://kerneltrap.org/mailarchive/git/2008/7/23/2648344 but it needs rebasing	07:56
	→ tal67 joined	07:56
	← harinath left	07:56
	→ webben joined	07:57
webben	Hi, I'm trying to find out if git-svn supports generation of svn:mergeinfo metadata yet.	07:57
	→ ahupp\|fb joined	07:58
	← shapeshed left	07:58
	→ harinath joined	07:58
webben	I found this: http://marc.info/?l=git&m=121196641823189&w=2	07:58
	← Niamor left	07:58
webben	but I'm not sure where to look to see if something like it has been integrated and works.	07:59
	→ Niamor joined	07:59
DrNick	kervel: git cherry doesn't do what you want?	08:00
	→ shapeshed joined	08:00
	→ moore__ joined	08:01
webben	nvm ... grep -r "mergeinfo" git-1.6.1/* turns up nothing	08:01
	→ libc__ joined	08:03
	← harinath left	08:06
	→ meyering joined	08:08
	→ hdl2 joined	08:10
	← moore_ left	08:11
	← naeu left	08:12
	← libc_ left	08:13
	← muthu_ left	08:14
	→ hyperboreean joined	08:20
hyperboreean	hey guys, is there a way to remove the global git configs? a more elegant one, rather than doing git config user.name "" ?	08:21
j416	hyperboreean: rm ~/.gitconfig	08:21
	if you are looking to unset a single key using git config	08:22
Pieter	or git config --unset --global user.name	08:22
j416	you will find the flag --unset in its man-page :)	08:22
hyperboreean	thanks guys, I think removing git config will do it :)	08:25
j416	hyperboreean: note that the git config files are plain text files in ini-style format	08:25
	→ jmesnil joined	08:25
j416	you can edit them as they are if you want	08:26
hyperboreean	yes, I have edited some of them ... just I didn't know that I have a global one :)	08:26
j416	git config is just a command-ish way of doing just that	08:26
	→ moore_ joined	08:26
j416	ok :)	08:26
	→ vuf_ joined	08:27
	← moore__ left	08:28
	→ johan-s joined	08:29
	← shapeshed left	08:30
ben_h	hi folks	08:31
	i've got a repo with a bunch of commits - i'd like to undo the latest two commits, i.e. move the HEAD back to HEAD^2, but leave the changes from those two commits in the working copy	08:32
	remove two commits without touching the files on-disk in effect.	08:32
	is there an easy way to do that?	08:32
Pieter	git reset HEAD^2	08:33
	→ derRichard joined	08:35
telmich	morjen	08:37
	← shoe left	08:38
	→ foca joined	08:38
	→ pierre- joined	08:39
	→ shoe joined	08:40
ben_h	jeez, that _is_ easy :)	08:43
	thanks Pieter	08:43
	bbl all	08:43
	→ naeu joined	08:43
kervel	i tried git log --cherry-pick and git cherry for showing A..B log that should omit everything from B that was cherrypicked in A, and it still shows the cherrypicked changes	08:44
	← ben_h left	08:44
	martinpaul_ → martinpaul	08:46
	→ charon joined	08:47
	→ hurikhan\|Work joined	08:47
vuf_	I have a detached HEAD, and want to move master to where I'm at ... how would I do that?	08:48
Pieter	vuf_: git branch -f master	08:49
vuf_	Pieter: right, that escaped me ... thanks	08:49
	I was looking for a named version of reset	08:50
	← charon left	08:52
	← derRichard left	08:53
	→ rraasch joined	08:55
	→ sverrej_ joined	08:55
	→ sjn joined	08:55
	→ lebreeze joined	08:58
	← lebreeze left	08:58
	← Marmouset left	09:00
	→ lebreeze joined	09:01
	← lebreeze left	09:02
	→ dwave joined	09:03
	→ charon joined	09:04
	→ comphappy joined	09:04
	← j416 left	09:05
	→ muthu_ joined	09:05
	→ arekm joined	09:06
comphappy	if I put a system link in my local git, when i do git add will it see as system link, or the files inside of it	09:06
	arekm wonder why git send-email can't be used from outside of repository. I often work with repositories on one machine and would like to send-email from another	09:06
	← tal67 left	09:07
	← sverrej left	09:07
DrNick	comphappy: a what?	09:08
charon	arekm: it uses the repo's .git/config too. you could make a patch to make that optional	09:08
	moore_ → moore33	09:09
comphappy	DrNick: I have my local git project repo at ~/triageproject	09:09
	I have another project i want to put in that git repo at ~/workspace/otherproject	09:09
	→ __name__ joined	09:10
comphappy	can I put a system link in ~/triageproject that points to ~/workspace/otherproject	09:10
	or is that just asking for trouble	09:10
DrNick	oh, you mean a symbolic link?	09:10
	no, you can't, git will just store the symlink itself	09:10
comphappy	sorry... yes I need some more coffee	09:11
	ok that is what I though	09:11
	← __name__ left	09:11
	→ name joined	09:11
moore33	comphappy: If you want to include one project in another, you need submodules or to read about doing a subtree merge.	09:11
comphappy	the problem with sub-modules is that trac cannot work with them	09:12
	→ tal67 joined	09:12
	← plasticine_ left	09:14
	← zjason left	09:18
	← benjo2 left	09:19
	→ samjam joined	09:22
	← comphappy left	09:22
	→ solydzajs joined	09:23
	→ derwolf joined	09:24
	→ mark[oz] joined	09:24
	← sverrej_ left	09:25
	→ sverrej_ joined	09:26
	← mithro left	09:27
	← jchris left	09:32
	→ benjo2 joined	09:35
	← name left	09:35
	→ name joined	09:37
	← moore33 left	09:40
	→ moore33 joined	09:41
	→ Catfish joined	09:42
Catfish	Heya	09:42
	Anyone use emerge? I've gotten stuck trying to do a merge over ssh...	09:42
	After using a or b to select a difference, are you supposed to just manually remove the conflict markers, or is there a command?	09:43
	I can't seem to find it...	09:43
	→ cmarques joined	09:43
	→ psychoschlumpf joined	09:47
psychoschlumpf	hi	09:47
	is there a way to only allow a user to push his changes to the master repo if all tests run fine on it?	09:48
	like with svn's pre-commit-hooks	09:48
	→ simoncpu joined	09:49
simoncpu	hello	09:49
thiago	you can use pre-recieve hooks	09:49
simoncpu	i'm experimenting with a workflow that uses both cvs and git	09:49
	how can i ignore CVS/* directories?	09:49
thiago	which technically are a misnomer: it's after the reception, but before the update	09:49
	simoncpu: echo CVS >> .gitignore	09:49
psychoschlumpf	thiago: will a pre-recieve hook check the whole bunch of commits or every single commit?	09:50
simoncpu	but there are lots of CVS directories... =)	09:50
thiago	psychoschlumpf: it will do whatever you write in the script	09:50
	simoncpu: but they are all called "CVS"	09:50
simoncpu	ah...	09:50
	lemme try	09:50
	very coooool	09:50
psychoschlumpf	ok: the other way around: will git call the post-receive hook for the whole push or for every single revision which is pushed in one push?	09:50
	simoncpu gives thiago a beer	09:50
thiago	psychoschlumpf: I think you should read the documentation of the pre-receive hook instead	09:51
psychoschlumpf	thiago: okay, thanks	09:52
thiago	there's a hook that is called for every single commit in the pack	09:53
	but that's not the pre-receive hook	09:53
psychoschlumpf	just read it	09:53
	pre-receive hook is called once and gets the single commits on stdin	09:53
thiago	the pre-receive hook receives a list of branch changes in stdin. It's up to you to decide what to do.	09:54
	you can test every single commit introduced, or you can test just the finals, or you can do anything in-between	09:54
	→ uhu01 joined	09:58
uhu01	hi, is there a command to condense three commits to one? I think i red it some time ago, but I couldn't find a hint anywhere	10:00
teuf	uhu01: git rebase -i helps you to do that	10:00
	→ markelikalderon joined	10:00
tokkee	uhu01: git rebase -i -> squash the commits.	10:00
uhu01	thx	10:01
	← dulia left	10:04
	← moriyoshi left	10:05
	→ Clebrate joined	10:06
	→ moore_ joined	10:06
Clebrate	Does anybody here speak Scandinavian?	10:07
thiago	I can read Norwegian	10:07
Clebrate	I need to translate "to fork a project"	10:07
	thiago: God bless you :)	10:07
thiago	from a Scandinavian language or to?	10:07
	← pierre- left	10:07
Clebrate	To.	10:08
Pieter	the 'from' part would be easy	10:08
	← moore33 left	10:08
Clebrate	heheh	10:08
	yeah...	10:08
	→ pierre- joined	10:08
jnl	thiago: any scandinavian language in particular, or just.. scandinavian in general? :)	10:08
thiago	jnl: --> Clebrate	10:09
Pieter	\|> deflection	10:09
thiago	Clebrate: you can try the "å forke" verb...	10:09
	:-P	10:09
Clebrate	hehe	10:09
	<Rounin> Å forke...	10:10
	<Rounin> Eneste forskjellen er at o'en kanskje er litt lengre, og tonefallet er helt ødelagt	10:10
thiago	that's norwegian	10:10
Clebrate	<Rounin> Synkende tone frem til midten av o'en, og så stiger det på et øyeblikk... Akkurat som i "bønner" og "venner"	10:10
	<Rounin> Vi har jo ord som "forgrene" (branch) osv., men forking er forking	10:10
	from some other channel	10:10
Pieter	that's norwegian	10:11
Clebrate	he was kinda joking though	10:11
	yup	10:11
Pieter	oh	10:11
	I was too late :	10:11
	:)	10:11
thiago	why are you talking about tones?	10:11
Clebrate	dunno	10:12
thiago	and he's saying that "forking is forking"	10:12
Clebrate	i think he's a vocal coach	10:12
Pieter	he's a hardcore forker	10:12
thiago	"the only difference is that the 'o' can be a bit longer, and the tonefall is entirely destroyed. The sinking tone from the middle of o and lasts for an instant, like in 'bønner' and 'venner'. We have words like 'forgrene' (branch), but forking is forking"	10:13
	← Niamor left	10:13
	→ Niamor joined	10:14
Pieter	in Dutch we just use 'branch' for 'branch'	10:14
	or, at least I do	10:14
thiago	I would too in Portuguese, even though it sounds nothing like Portuguese	10:14
Pieter	most tech-talk is very English-centric anyway	10:15
Clebrate	:D	10:15
	those are my two favorite languages	10:15
Pieter	compared to German or French, where a lot of stuff is translated	10:15
Clebrate	oh and arabic	10:15
thiago	Portuguese has only one word with the /tʃ/ sound and was imported from Italian	10:15
Clebrate	portuguese and dutch that is	10:15
	← markelikalderon left	10:16
Pieter	:) Dutch is very easy if you're a native speaker	10:17
	just like with any other language	10:17
	← mark[oz] left	10:17
	← solydzajs left	10:19
	← ankit9 left	10:19
	→ mark[oz] joined	10:21
	→ Kbyte joined	10:21
	→ ben_h joined	10:21
	← ahupp\|fb left	10:27
	← tango_ left	10:28
	→ tango_ joined	10:28
	← moon_unit left	10:28
	→ aguerreiro joined	10:29
	← bgmarete left	10:32
	→ timebomb joined	10:34
timebomb	heya im trying to convert a mercurial repo to git and fast-export fails. anyone come across this error before? http://pastie.org/377208	10:35
	→ _Vi joined	10:35
	→ flow joined	10:36
	→ tanek joined	10:37
	← tal67 left	10:37
thiago	that means the reader side closed	10:37
	check if git-fast-import created a crash log in your target .git	10:37
	← sandGorgon left	10:37
	← ben_h left	10:38
	→ ben_h joined	10:38
timebomb	http://gist.github.com/4e25308d7a5e5d1b086e	10:39
	→ ankit9 joined	10:39
	→ trupheenix joined	10:40
timebomb	says mark :0 not declared but obviously i do not know how to deal with that	10:40
trupheenix	how do i correct spelling mistakes inside the message for a commit after the commit is done?	10:41
thiago	git commit --amend	10:41
trupheenix	thiago: older commits?	10:41
speakman_	(missed the scandinavian discussion, but I'm from Sweden if it's to any help)	10:41
thiago	git rebase -i oldestcommit^	10:41
	speakman_ → speakman	10:41
thiago	change to "edit"	10:41
	remember that you can't change commits that other people already have	10:41
speakman	wont a spelling fix change the sha1?	10:42
thiago	it will	10:42
	that's why you can't change commits that other people already have	10:42
timebomb	any idea how to deal with that thiago ? hg verify says everything is fine ...	10:44
	→ j416 joined	10:44
thiago	what does git say?	10:44
	ah, you pasted it	10:44
	it's a bug in the fast-export	10:45
	→ tal67 joined	10:46
timebomb	known bug? i couldn't find anything about it but maybe i was just searching for the wrong thing	10:47
thiago	I don't know if it's known	10:47
	but it's telling fast-import to use mark 0, but it didn't define that mark	10:48
	therefore, it's a bug	10:48
	→ jaeckel joined	10:48
Pieter	I'd guess it can't find a commit's mark number and just returns 0	10:51
timebomb	mh	10:51
	→ pinholecamera_ joined	10:54
	← pinholecamera left	10:58
	pinholecamera_ → pinholecamera	10:58
	← ankit9 left	10:59
	← tamiko left	11:02
	→ tamiko joined	11:02
	→ schierbeck joined	11:02
	← gretch left	11:02
	→ greyface joined	11:04
	← schierbeck left	11:06
	← mark[oz] left	11:06
	→ gouki_ joined	11:06
	← aguerreiro left	11:08
	← sverrej_ left	11:09
	→ kraymer joined	11:09
	← flow left	11:12
	→ sverrej joined	11:12
	← plediii left	11:14
	← Phlogi left	11:16
	→ kbingham joined	11:16
	→ Phlogi joined	11:17
	← gouki left	11:22
	→ mark[oz] joined	11:23
	→ eddyp joined	11:28
	← gouki_ left	11:32
	→ gouki_ joined	11:32
	← Flathead left	11:32
	→ madnashua joined	11:32
	← hurikhan\|Work left	11:32
	← muthu_ left	11:32
telmich	hmm, why can't I git push --all --tags?	11:35
Voker57	use --mirror maybe	11:36
Pieter	I think --all implies --tags?	11:36
Voker57	nope	11:36
	→ hurikhan\|Work joined	11:37
telmich	oh, yes...strange, though	11:37
	I would not thing --mirror function really excludes me from naming both manually	11:38
	← troter left	11:39
timebomb	hm hg-to-git doesn't work either :S	11:40
	→ Yuuhi joined	11:42
	→ lorandi joined	11:43
	→ grahal joined	11:45
	← j416 left	11:47
	← moore_ left	11:49
	→ moore_ joined	11:49
	← trupheenix left	11:52
	→ harinath joined	11:52
	→ radarek joined	11:54
	→ bryanl joined	12:02
	→ johan-s_ joined	12:03
	← hurikhan\|Work left	12:05
	→ hurikhan\|Work joined	12:07
	← iulian left	12:10
	→ schlort joined	12:12
	← johan-s left	12:13
	← eddyp left	12:13
	→ Hendu joined	12:15
	→ cemerick joined	12:16
	→ slava joined	12:16
slava	i have a patch that renames a file, only changing its case, and now i'm having trouble pulling this change onto a mac (case-insensitive file system	12:16
	is this a known bug?	12:17
shd	i think it's a feature	12:18
	→ simosx joined	12:18
slava	a feature?	12:18
	yeah, because it encourages people to switch to linux instead of using inferior operating systems, right, haha	12:19
shd	right :-)	12:19
au-	noob question.	12:19
shd	no seriously, the project that uses such names is at fault here..	12:20
au-	git can't merge 2 files if it doesn't have to commit history right?	12:20
shd	au-: right, but patching is different	12:20
slava	shd: regardless of coding practices and naming conventions, git shouldn't refuse to pull a patch that renames a file changing case	12:20
	→ johan-s joined	12:20
slava	in this instance it was renaming foo.TXT to foo.txt	12:20
au-	can I do it with patching? /me looks up 'patch'	12:20
slava	one would think this woudl be pretty harmless	12:20
shd	slava: better mail your report to git mailing list	12:21
	← mark[oz] left	12:21
slava	I will, but their answer will probably be similar to yours	12:21
	where I work, refusing to fix a bug would get you fired ;-)	12:21
	→ Sho_ joined	12:22
shd	slava: fortunately it's an open source project so we can argue what is a bug and what is not..	12:22
au-	it's an "unintended feature"	12:22
slava	shd: if 'git status' reports that there are no uncommitted changes, and git pull only has to fast forward without doing a merge, and it fails, that's clearly a bug	12:23
	→ ashleyw_ joined	12:24
	→ moore__ joined	12:25
Arjen	foo.txt and foo.TXT are the same file on osx, hence the error	12:25
	→ Kathrin-25^away joined	12:25
slava	sure, but there's no time when both exist at once	12:26
Arjen	Just create a case-sensitive HFS+ disk image and use that	12:27
	There has been lots of threads about this on the mailing list	12:27
	→ iulian joined	12:28
	← webben left	12:30
slava	sounds like something that should be fixed, then	12:30
Arjen	slava: What's the exact error message? Something like "fatal: destination exists ..." ?	12:30
	← ashleyw left	12:31
slava	I don't have it on hand anymore, but roughly 'untracked file foo.TXT would be overwritten by merge"	12:32
	but the file was checked in and git status did not report uncommitted changes	12:32
	let me reset the repo and pull agian	12:32
	error: Untracked working tree file 'basis/io/encodings/japanese/CP932.txt' would be overwritten by merge.	12:33
	gouki_ → gouki	12:33
	← moore_ left	12:34
	→ mark[oz] joined	12:35
Arjen	It's know, and apparently hard to fix well	12:36
	s/know/known/	12:36
	← madnashua left	12:37
	← mark[oz] left	12:37
	← johan-s left	12:37
	← johan-s_ left	12:38
	→ vbabiy joined	12:39
	→ johan-s joined	12:39
	← EmilMedve left	12:40
	→ kraymer_ joined	12:40
	← kraymer left	12:40
	→ plasticine joined	12:41
	→ parasti joined	12:43
imarcusthis	&win 2	12:45
	meh	12:45
	→ cannonball joined	12:45
	← plasticine left	12:46
	← slava left	12:46
	→ KRiSPY joined	12:47
wereHamster	what's the best way to check from a script whether the working tree is dirty or not?	12:47
tokkee	fonseca: Hi!	12:47
doener	wereHamster: regarding the index or HEAD?	12:47
tokkee	fonseca: I'm getting the following when using tig 0.13 with LC_ALL=C: http://osprey.tokkee.org/~tokkee/tmp/tig-0.13-lc_all=c.png	12:48
	fonseca: Any ideas what might be wrong?	12:48
wereHamster	doener: HEAD I assume. I want to see if I can safely do git-rebase	12:48
	← reithi left	12:48
au-	okay so I have 2 different branches that I need to merge. I have no history of the files	12:49
	← KRiSPY left	12:49
au-	does git have any tools to make this easy?	12:49
	← bryanl left	12:49
tokkee	fonseca: It works fine with e.g. en_us.UTF-8.	12:49
	→ mikl joined	12:49
doener	wereHamster: I'd probably copy the checks from git-rebase.sh	12:50
	→ awarde joined	12:50
	← greyface left	12:51
doener	wereHamster: maybe using --quiet with diff-index command	12:51
	← girishr left	12:51
	→ jel joined	12:51
jel	Hey guys. Anyone know how to fix a "fatal: bad object HEAD" error?	12:52
	← jony_ left	12:54
	→ greyface joined	12:55
	cbrake_away → cbrake	12:55
wereHamster	doener: thanks, the rebase script has indeed these checks	12:55
	← arekm left	12:55
	← sverrej left	12:56
wereHamster	or, well, I could call git-rebase directly and let it fail if the working tree is dirty..	12:56
deskin	DrNick: (from ~7 hours ago) you get the most recent revision from SVN, mapped to trunk/branch/tag however you specify, and git checks out a local master branch at that commit	12:57
	jel: what does the contents of .git/HEAD look like?	12:58
	→ solydzajs joined	12:58
	→ mcella joined	13:01
	← GeertB left	13:01
tokkee	fonseca: That seems to happen with any non-UTF-8 locale, both using ncursesw 5.5 and a "5.7+20090124" snapshot.	13:02
	← timebomb left	13:03
	← johan-s left	13:07
jel	deskin: "refs/heads/master". That refs file is present.	13:07
deskin	'refs/heads/master', not 'ref: refs/heads/master' ?	13:07
jel	deskin: sorry, yes. ref:...	13:07
deskin	what does git rev-parse refs/heads/master say? What is the content of .git/refs/heads/master ?	13:08
jel	deskin: .git/refs/heads/master is just a hash. The tree used to work before I zipped them all up and moved them to a windows machine.	13:10
	rev-parse gives: fatal: ambiguous argument 'refs/head/master': unknown revision or path not in the working tree.	13:10
thiago	is there a .git/refs/heads/master file?	13:12
deskin	hmm; windows aside, copying the repo to another machine ought to work	13:12
jel	thiago: there is. It contains a hash, which looks normal as far as I can tell	13:12
thiago	why didn't you push to another machine instead?	13:12
	→ KRiSPY joined	13:12
deskin	my best advice at this point is to run git fsck --lost-found, look through the dangling commits listed with git show <hash>, and when you find the right one, git update-ref refs/heads/master <hash>	13:13
	I'm wondering if it's a permissions issue, but that's just a shot in the dark	13:13
	are you running cygwin? Can you strace git rev-parse and pastebin it somewhere?	13:14
	→ thm joined	13:14
jel	deskin: ok. Thanks. At least I know it should be fixable. Gotta go now. Any chance it's related to the remote repos ssh aliases not being set yet?	13:14
deskin	no, that has nothing to do with locating objects	13:14
jel	deskin: cygwin, yes. msys gave the same error. I'll give that a shot later.	13:14
	deskin: ok, thanks	13:14
deskin	alright, good luck	13:15
	← ceyusa left	13:15
	→ ekidd joined	13:18
	→ cmarcelo joined	13:20
	← thm left	13:24
	→ Acry joined	13:24
	← Acry left	13:25
	→ foutrelis joined	13:28
	→ cthompson joined	13:31
	→ davidfetter joined	13:32
	→ EmilMedve joined	13:33
	→ bryanl joined	13:33
	→ ProperNoun joined	13:34
	→ a-priori joined	13:35
	← naeu left	13:36
	→ naeu joined	13:37
	← samjam left	13:38
	→ jas4711 joined	13:39
	→ markelikalderon joined	13:40
	← CardinalNumber left	13:40
	→ EmilMedve1 joined	13:47
	← cmarcelo left	13:48
	← Niamor left	13:48
	→ cmarcelo joined	13:49
	→ Niamor joined	13:49
	→ boto joined	13:49
	← EmilMedve left	13:49
	→ cawel joined	13:50
	→ metaperl_work joined	13:50
metaperl_work	what git command lists the latest pushes/commits?	13:50
	→ imexil joined	13:50
ciaran	log	13:50
	← psankar left	13:50
ciaran	:/	13:50
	← Kathrin-25^away left	13:50
	← ashleyw_ left	13:51
metaperl_work	thank you	13:51
imexil	Hi, I've got a free floating commit which is not attached to any branch. How can I now apply this to, let's say master?	13:51
	→ ashleyw joined	13:52
Voker57	git cherry-pick	13:52
	→ emss joined	13:53
imexil	yes of course ... thank you Voker57 :)	13:53
metaperl_work	git log doesnt show me which files were pushed with each commit	13:54
doener	-p\|--name-only\|--name-status\|--stat\|...	13:54
nothingmuch	metaperl_work: git whatchanged	13:54
	→ Araneidae joined	13:55
Araneidae	How do I view the previous revision of a file, please?	13:56
nothingmuch	git show decafbad:path/to/file	13:56
	similarly git checkout can be used to go back to the previous revision	13:57
metaperl_work	wow	13:57
	thank you nomis	13:57
	i mean nothingmuch	13:57
Araneidae	Well, it's quite painful to find the decafbad bit	13:57
	Is there a way to say simply last nth version?	13:57
nothingmuch	read on git-rev-parse	13:57
	HEAD~N	13:57
Araneidae	$ git blame sstrip.sh~1	13:58
	fatal: cannot stat path scripts/build/tools/sstrip.sh~1: No such file or directory	13:58
parasti	metaperl_work: note that "push" is what happens when you use "git push"	13:58
nothingmuch	git blame HEAD~1:sstrip.sh	13:58
	i think	13:58
	→ dantrell joined	13:58
dantrell	Let's say there's repos foo and bar. Bar is a much smaller version of foo but maintains the same format. I want to create a patchset for foo to use to update to bar's changes. Is there no way to do this for an outside repo?	13:58
	→ ceyusa joined	13:58
doener	nothingmuch: without the :	13:59
	→ gitte joined	13:59
nothingmuch	metaperl_work: note that doener's answer is also for you	13:59
doener	nothingmuch: blame takes revision and name, not just a blob	13:59
nothingmuch	ah right	13:59
	if it were show it'd probably need the colon though, right?	13:59
doener	blob's have no history, so it'd be pretty hard for blame to do anything useful	13:59
	yep, show <blob> (i.e. with colon) shows the blob	14:00
Araneidae	Unfortunately I think `git blame HEAD~1 filename` is giving me the filename as of the previous HEAD version.	14:00
doener	show <commit> <path> shows the commit with the diff limited to <path<	14:00
Pieter	Poor little Bobbie Blob.. he only knows what's happening now, but not what hap -- what was I saying?	14:00
Araneidae	Not so useful when I'm trying to track back through the history of one particular file	14:00
nothingmuch	Araneidae: what are you trying to find out?	14:00
Araneidae	I'm trying to find when a particular pattern was in this one file	14:00
	→ wildmind joined	14:00
Pieter	Araneidae: git log -S'searchmatch' -- path/to/file	14:01
Araneidae	There's a line that isn't in the file in head, and isn't in early versions	14:01
wildmind	I pushed some changes to a remote branch. The changes appear in the index on the remote but are not commited ... how come?	14:01
Pieter	wildmind: did you read faq non-bare ?	14:02
Gitbot	wildmind: Pushing to non-bare repositories is discouraged. See http://git.or.cz/gitwiki/GitFaq#non-bare	14:02
moore__	How about git-log -n2 file?	14:02
	→ johan-s joined	14:02
Araneidae	Pieter, that was helpful. Odd, though. The offending line doesn't seem to be in my history! Hmm.	14:03
	→ Kathrin-25^away joined	14:03
wildmind	Pieter: thanks	14:03
Araneidae	At least, assuming that my -S'pattern' would have picked up all commits containing matching lines...	14:04
Pieter	Araneidae: it only shows commits in which that pattern was added or removed	14:05
	not in which it moved	14:05
	Araneidae: you can also try a 'git log -p' and search through that with less if you want a quick fix	14:05
	;)	14:05
Araneidae	I have a horrible feeling the offending line was in a branch which isn't in my history (I'm searching a `git svn` history)	14:06
	Actually, `git-log -p` is exactly what I wanted! Thanks.	14:06
moore__	Araneidae: previous revision of file: git-rev-list --max-count=2 HEAD foo.c \| tail -1	14:06
Pieter	moore__: -1 --skip=1	14:07
	→ bobmcw joined	14:07
Araneidae	Unfortunately I use git very infrequently, so every time I do I'm painfully rusty!	14:07
	← ekidd left	14:07
Pieter	yeah, that happens.. I have the same with hg/bzr or even svn	14:08
moore__	Pieter: cool	14:08
	EmilMedve1 → EmilMedve	14:09
nothingmuch	svn can drive me nuts... svn log -n... no worky, svn log -l... nope, svn log --last? urgh!!! then I read the help	14:10
	(that's git style, svk style and then darcs style...)	14:10
	→ muthu_ joined	14:10
	← hegge left	14:11
	→ hegge joined	14:11
Araneidae	Ah well. Most of my work is with svn -- I know to avoid working with branches!	14:12
	← EiNZTEiN left	14:12
Pieter	yeah, that results in having to write very clean code all the time, it's a very good practice :)	14:12
Araneidae	The other thing I really have to watch with svn: as soon as any directory structure changes ... commit as soon as possible!	14:14
dantrell	Let's say there's repos foo and bar. Bar is a much smaller version of foo but maintains the same format. I want to create a patchset for foo to use to update to bar's changes. Is there no way to do this for an outside repo? I don't want to pull in the outside repo (which is really big) just to patch it.	14:17
Voker57	git diff	14:18
moore__	dantrell: git format-patch and git-am can be used to shuttle patches between dissimilar repos.	14:20
	→ esparkman joined	14:20
	→ ekidd joined	14:21
	→ aziz joined	14:22
	← joshdavey left	14:22
	→ girishr joined	14:23
	← wildmind left	14:23
	← a-priori left	14:24
dantrell	moore__: I have been looking at git format-patch but I don't see how to diff foo and bar and get a single patch from that.	14:25
nothingmuch	Araneidae: git-svn is a real life saver	14:25
	it makes svn bearable	14:25
	→ joshdavey joined	14:26
moore__	dantrell: Are there revisions of foo and bar where their contents are the same?	14:26
	→ xl0 joined	14:27
dantrell	moore__: Nope.	14:27
	→ Bass10 joined	14:27
nothingmuch	you can [ab]use git cherry-pick	14:27
Araneidae	nothingmuch, I guess I ought to try using it for my svn development! So far I've only used it to browse external repositories, when it's very useful. But then I'm used to svn and its quirks.	14:27
moore__	dantrell: What does it mean to update one to the other then?	14:28
	→ plediii joined	14:28
tro	what could I have done to my repo or its config that git gui doesn't notice about half the changes made to the working copy? i have to run git update-index --really-refresh to have them show up in git gui	14:28
moore__	You could just dump all the files from bar into foo and commit the result.	14:28
tro	i'm using msysgit	14:28
xl0	Hi, guys. Could someone please tell me, what's the right way to set up a bare repository to track an other repo?	14:29
	Git pull doesn't work on bare repositories, and even if it worked, how would it handle the branches besodes master?	14:30
parasti	xl0: use git fetch instead	14:31
dantrell	Voker57: Wait, I think git diff does it...maybe...	14:31
parasti	xl0: pull is fetch + merge	14:31
	→ francois___ joined	14:31
	→ krh joined	14:31
	→ jschoolc joined	14:31
	← francois___ left	14:32
jschoolc	is there an easy way to rollback/revert to a commit and blow away everything that happened after that, committed or not?	14:32
dantrell	moore__: Say foo has a/1 a/2 b/1 c/1 and bar has a/3 b/2 d/* e/*. There are pretty much more conflicts. I want what bar has in foo.	14:32
nothingmuch	jschoolc: git reset --hard	14:32
dantrell	much no conflicts*	14:32
nothingmuch	dantrell: you can use git fetch and then git cherry-pick	14:32
	→ dmlloyd joined	14:33
moore__	dantrell: cp -r from bar into foo and commit the result. The diff isn't interesting, right? If you want to preserve the history in bar's commits, then you need to start from some common point in foo and bar.	14:34
	← _Vi left	14:34
moore__	dantrell: cp -r the working tree of bar, of course.	14:34
jschoolc	nothingmuch: does reset take a sha to reset to?	14:34
	← ben_h left	14:35
nothingmuch	jschoolc: git help rev-parse... like any other command	14:35
	sha1, ref name, an expression involving ~ or ^ or @...	14:35
dantrell	moore__: I just want to create a patch to send to the foo maintainer. So I need the diff.	14:35
	← girishr left	14:36
	→ GeertB joined	14:36
moore__	dantrell: Do what I described, commit the result in foo, then use git-diff.	14:36
	dantrell: Rather, use git-format-patch for the canonical result.	14:37
	→ samjam joined	14:37
moore__	dantrell: Otherwise, use plain old diff.	14:37
	→ ph^_ joined	14:39
	← qrush_ left	14:39
xl0	parasti: And how should I set up the origin remote?	14:39
	→ greister joined	14:39
greister	how to use git clone the tag?	14:39
	→ spearce joined	14:39
dantrell	moore__: Thanks.	14:39
	← imexil left	14:41
	← Araneidae left	14:42
	→ sverrej joined	14:43
	← metaperl_work left	14:43
	→ vladimirprieto joined	14:43
vladimirprieto	hi	14:53
	← ph^ left	14:53
	← Niamor left	14:53
	→ Niamor joined	14:54
vladimirprieto	i'm thinking about installing git on the server, but i have some doubs about. is there's anybody that can answer a specific question on the use of git?	14:54
	→ bartman joined	14:54
xl0	parasti: Just git fetch does not seem to work. What am I doing wrong? ;)	14:54
	vladimirprieto: Don't ask to ask, just ask.	14:55
vladimirprieto	ok	14:55
	this is my problem: i got a few programmers working on a web project. this project is already working and it had been using on differents customers on the same server.	14:56
	every custumer has its own directory www.myserver.com/customer	14:57
	every customer directory is an instance of the project we are programing, but every instance is not exactly the same. so, the question is: git can copy/export the project for productions WITHOUT touching some files in it (cus that files make identify every project)?	14:59
	← ekidd left	14:59
	→ soul9 joined	15:00
soul9	hi	15:00
	tryiong to rebase:	15:00
	→ qrush__ joined	15:00
soul9	$ git rebase origin/master	15:00
	could not detach HEAD	15:00
	← vuf_ left	15:00
xl0	vladimirprieto: yes, if you ask it the right way.	15:00
soul9	anyone got a clue? ☺	15:01
	→ kukks joined	15:01
	→ jrmuizel joined	15:02
vladimirprieto	x10: thanks for answer, but what do you mean exactly	15:02
xl0	vladimirprieto: First, you have explicitly add newly created files to be tracket by git.	15:03
	→ bwalk joined	15:03
xl0	And you can even select which files to include in a commit, if there are multiple changed files.	15:04
	← shoe left	15:04
	→ ekidd joined	15:04
vladimirprieto	ok	15:05
	→ EiNZTEiN joined	15:05
gitte	soul9: without more information, no.	15:05
vladimirprieto	x10: i think i understand.	15:05
	→ EiNZTEiN_ joined	15:05
vladimirprieto	x10: thank you very much	15:05
	→ ashleyw_ joined	15:08
	→ Dvyjones joined	15:10
	→ ben_h joined	15:11
xl0	Found what I was looking for.	15:14
	git --bare clone	15:14
	git remore --mirror add origin git://...	15:14
	git remote update now updated all the branches properly.	15:15
	Why isn't git remore --mirror add dont by git clone automatically? ;)	15:15
ekidd	xl0: git clone --mirror is probably what you want.	15:15
gitte	xl0: in general, it does not make any sense at all.	15:16
	→ marvil07 joined	15:16
gitte	xl0: --mirror overwrites the local copies of the tracked branches.	15:16
xl0	gitte: It makes a lot of sence when --bare is used.	15:16
	ekidd: It is not yet in the man pages. ;)	15:17
gitte	xl0: not necessarily. As I said, it _overwrites_ the local copies of the tracked branches.	15:17
	← rraasch left	15:17
xl0	gitte: What else would I want a bare repository update to do?	15:17
	→ lebreeze_ joined	15:18
gitte	xl0: if you only have one upstream, right.	15:18
	← lebreeze_ left	15:18
soul9	gitte: ?	15:18
	gitte: ask..	15:18
	← Niamor left	15:18
	→ Niamor joined	15:19
gitte	xl0: but Git is _decentralized_.	15:19
soul9	i'd be happy to help you help me	15:19
gitte	soul9: was that really all that Git said? I cannot believe that.	15:19
soul9	yes	15:19
	it was	15:19
xl0	gitte: I understand it quite well. Probably yes, it should not be the default, but git clone --mirror seems to be a good idea.	15:20
soul9	git rebase origin/master	15:20
	rewwinding...	15:20
	could not detach HEAD	15:20
	← ashleyw_ left	15:20
	→ ashleyw_ joined	15:21
gitte	soul9: so what is your current branch right now ("git branch")?	15:21
soul9	WIP2	15:21
	→ Weasel[DK] joined	15:21
gitte	soul9: could you try "git checkout origin/master^0"?	15:21
	← kanru left	15:22
	← ashleyw left	15:23
soul9	gitte: is the ^0 deliberate?	15:23
gitte	soul9: yes.	15:23
soul9	(if it is, what does it mean?)	15:23
	← Hendu left	15:24
soul9	sec, it's my co-worker who's having problemms, so i'm copying his repo over..	15:24
gitte	soul9: it means that I want to detach the HEAD, not check out "origin/master" itself.	15:24
	← ekidd left	15:24
	→ bdiego joined	15:26
soul9	hmm	15:26
	ok	15:26
	← lorandi left	15:27
	→ MarcWeber joined	15:27
	→ acoc joined	15:27
MarcWeber	When renaming/moving a lot of files which is the best strategies to follow changes made to another repo? Will git still apply the changes to the new location or will I get delete /modified conflicts all the time?	15:28
soul9	gitte: nice, ok, that gave me something to work on	15:30
	→ rraasch joined	15:30
	← xl0 left	15:30
gitte	soul9: I am sure I don't understand... no problem there?	15:31
	← jrmuizel left	15:31
soul9	gitte: no, that command gave me a usable error message	15:31
gitte	soul9: ah, okay. Good, then.	15:32
soul9	gitte: yeah, it was: untracked local file would be overwritten..	15:32
gitte	soul9: that's what I expected, actually.	15:32
soul9	yeah, well, it wasn't that obvious to me ;-)	15:33
	i hope i remember ^0	15:33
gitte	soul9: it happened to me, recently, that's why...	15:33
soul9	ah, i see	15:33
gitte	soul9: although it happened with rebase -i, and that command does not hide the error message.	15:33
	← vladimirprieto left	15:33
soul9	i see	15:33
	what is the difference between rebase and rebase -i?	15:34
gitte	soul9: the latter is interactive...	15:34
soul9	i don't understand exactly..	15:34
frsk	Try for yourself and see the difference?	15:34
soul9	what would interactive mean in the context of a rebase	15:34
	i did	15:34
	hmm	15:34
gitte	soul9: it lets you choose the order, and lets you skip commits.	15:35
soul9	it was a long time ago, when i wasn't a git wiz yet ;-)	15:35
	i see	15:35
	well, i don't want that	15:35
	i'd like git to give me a usable error tho	15:35
	is there a rebase --verbose or something?	15:35
	☺	15:35
	→ krawek joined	15:36
	→ hugo_ joined	15:38
	→ akitada joined	15:38
	← greister left	15:38
	→ foca_ joined	15:39
gitte	soul9: not that I know of. It uses "git checkout -q" to detach the HEAD, and that suppresses warnings _and_ errors.	15:39
	soul9: I mean the messages, of course.	15:39
	← mahesh left	15:41
	← Kathrin-25^away left	15:42
	→ Kathrin-25^away joined	15:43
Knirch	if I accidently did git reset HEAD~1, how can I quick and easy get that changeset back?	15:43
gitte	Knirch: should be a matter of "git reset HEAD@{1}".	15:44
wereHamster	how do I get the output of 'request-pull' into an email? I can't pipe it to imap-send, so it leaves me with manual copy&paste. But the lines are >80 chars so my MUA wraps those	15:44
	← foca left	15:44
gitte	Knirch: does not work if you had uncommitted but staged changes, though.	15:44
Knirch	gitte: thanks, just what I was looking for	15:44
wereHamster	and I see that the pull requests sent to the kernel mailing list aren't wrapped, how do they do it?	15:45
Knirch	all the rev-parse fairy dust is confusing me, I should read that page once a day until it sticks	15:45
	→ ashleyw joined	15:45
	→ khisanth_ joined	15:47
	← hugo_ left	15:48
	← akitada left	15:48
Knirch	hmm, I was toying around with apply.whitespace, had it at fix, git add file ; git commit -mtest echoed rewrite file (93%), resetting to HEAD~1 and changing apply.whitespace to error and adding/committing again still rewrites the file	15:49
	khisanth_ → Khisanth	15:49
Knirch	is the resolution cached somewhere?	15:49
	← tango_ left	15:49
	← qrush__ left	15:50
Knirch	guess I'm just confusing myself, nevermind	15:53
	← ashleyw_ left	15:53
	← adam[pac] left	15:54
	← EiNZTEiN left	15:55
	EiNZTEiN_ → EiNZTEiN	15:55
	← dwave left	15:55
	→ skorpan joined	15:57
	→ adam[pac] joined	15:58
	→ Ryback_ joined	16:00
	← Sho_ left	16:01
	→ lresende joined	16:03
	→ willb joined	16:03
	→ jrmuizel joined	16:03
	→ cib3r joined	16:07
	→ Ratler_ joined	16:08
	← au- left	16:15
	← mikearr left	16:15
	← johan-s left	16:16
	→ bgmarete joined	16:16
	← ben_h left	16:17
cib3r	How do I make git exclude files that are already on svn? I can get it to ignore files that don't yet exist, but not already existing files...	16:19
	→ withanx joined	16:19
	← marvil07 left	16:19
withanx	if I committed changed files and I wanted to pull them out of the commit, back to a changed state, how would I do that?	16:20
	→ bantic joined	16:20
	→ paulboone joined	16:21
	→ jesselucas joined	16:21
	← bgmarete left	16:21
	→ mikearr joined	16:21
	← foca_ left	16:23
	→ rubydiamond joined	16:23
	→ greyface_ joined	16:23
	→ akitada joined	16:23
	→ ekidd joined	16:26
	→ dwave joined	16:26
vuf	withanx: git reset HEAD^, assuming it is the latest commit	16:28
	or, you could make your additional changes and do "git commit --amend" to replace the latest commit	16:28
	→ rwaite joined	16:30
offby1	cib3r: I imagine you'll have to "git rm" them, _then_ add them to .gitignore	16:30
cib3r	offby1, would that remove them from svn as well?	16:30
offby1	once you dcommitted, yes.	16:31
	→ jchris joined	16:31
cib3r	no way to do it without removing from svn?	16:31
withanx	HEAD^ is the last commit on current branch?	16:31
offby1	I can't think of any way. You're suggesting that you want your git repository to be permanently out of sync with your svn repository, which seems weird.	16:31
	withanx: I think so, yes	16:32
vuf	withanx: HEAD is the latest, HEAD^ is the one before	16:32
withanx	aha	16:32
	offby1 is off by one, as befits his nick	16:32
vuf	hehe	16:32
withanx	so git reset HEAD^ would stage everything from the last commit?	16:33
thiago	no	16:33
withanx	or unstage	16:33
	uncommit	16:33
	heh	16:33
thiago	it would simply move the index and the branch back to the last commit	16:34
	→ icwiener joined	16:34
thiago	leaving the checkout intact	16:34
	← greyface left	16:34
	→ bcardarella joined	16:34
withanx	would you lose all the data from the last commit, or would they show as modified?	16:34
	← paulboone left	16:34
bcardarella	Was 'git-svn create-ignore' removed?	16:35
thiago	the checkout is intact	16:35
	meaning the files are not modified: they are left exactly as they were committed	16:35
bcardarella	Because I'm on version 1.5.2.4 of git-svn and I'm getting an error	16:36
	→ jpeeler joined	16:37
	← naeu left	16:38
Weaselweb	bcardarella: "an error" doesn't say anything about your problem, so you have to post your error message to get someone helping you	16:39
	→ korpios joined	16:40
	gitte CURSES the genius that made "mount -o iocharset=iso8859-15" ignore the _explicit!_ option	16:40
	← jpeeler left	16:40
	→ jpeeler joined	16:41
	← jpeeler left	16:41
bcardarella	Weaselweb: good point :p... lemme pastie: http://pastie.org/377429	16:41
Circuitsoft	Is it possible to plug different diff modules into git?	16:41
	(filetype-specific diff)	16:41
	← rraasch left	16:41
	→ rindolf joined	16:42
	→ jpeeler joined	16:42
rindolf	Hi all.	16:42
	← jpeeler left	16:42
Weaselweb	bcardarella: that's not an error message, it's the usage message for you how to call git-svn (despite the fact, the actual git-svn command was ommited) indicating, you typed something wrong	16:42
rindolf	Using git cvsimport - is there any way I can set up an initial date to restrict the import from?	16:42
bcardarella	Weaselweb: I typed 'git-svn create-ignore' in my repos root directory	16:43
	→ chem_ joined	16:43
	← mikearr left	16:43
	← ekidd left	16:43
	→ mikearr joined	16:43
	→ jpeeler joined	16:43
	← jpeeler left	16:43
	→ mithraic joined	16:44
	→ jpeeler joined	16:44
Weaselweb	bcardarella: mh, according to the usage message you posted, you're git version doesn't support "git-svn create-ignore", but i'm not firm with features in specific versions. maybe anotherone may help you here	16:45
	→ bgmarete joined	16:45
bcardarella	Weaselweb: okay, I'll see if I can find a better version... currently I'm on 1.6.0.4 from MacPorts.	16:46
	→ agile joined	16:46
Weaselweb	bcardarella: some minutes ago you said, you're using version 1.5.2.4	16:46
bcardarella	Weaselweb: that's the version of git-svn	16:46
	If they're both supposed to have the same version number then maybe that's the issue	16:47
Weaselweb	mh, i would expect they have the same version, but i'm not sure	16:47
	← jpeeler left	16:47
	→ ekidd joined	16:48
	→ jpeeler joined	16:48
	← jpeeler left	16:48
	→ jpeeler joined	16:49
	→ jashmenn joined	16:49
	← jpeeler left	16:49
	← chem_ left	16:49
bcardarella	Weaselweb: yeah, that was the issue. Wrong version of git-svn. Thanks!	16:49
Weaselweb	you're welcome	16:49
	← fields left	16:50
	→ jpeeler joined	16:50
	← plediii left	16:50
	← jpeeler left	16:50
	→ rlorandi joined	16:51
	→ johan-s joined	16:52
	→ jpeeler joined	16:52
	← hurikhan\|Work left	16:53
rindolf	Using git cvsimport - is there any way I can set up an initial date to restrict the import from?	16:53
	← jtimberman left	16:54
	← sgrimm left	16:55
Weaselweb	rindolf: maybe "-d" from cvsps will help you	16:56
rindolf	Weaselweb: ah.	16:56
	→ qrush_ joined	16:57
	→ rafael_lorandi joined	16:57
	→ travisjeffery joined	16:58
	← giallu left	16:59
	← moore__ left	16:59
	← ToxicFrog left	16:59
	→ chem_ joined	17:00
	→ moore__ joined	17:00
	← Kathrin-25^away left	17:02
	→ Kathrin-25^away joined	17:03
	→ alley_cat joined	17:03
	← chem_ left	17:03
	→ foca joined	17:05
	← scientes left	17:08
	→ paulboone joined	17:08
	→ chem_ joined	17:09
	→ ashleyw_ joined	17:09
	→ lois[pac] joined	17:09
	← ekidd left	17:10
	→ david_koontz joined	17:10
	→ scientes joined	17:12
jel	sigh by default, Visual IRC keeps a tiny buffer, and doesn't log.	17:12
	I was talking to someone earlier about fixing "fatal: bad object HEAD".	17:12
	Can someone point me in the right direction (and explain what would cause this)?	17:13
	← rlorandi left	17:13
	→ cgardner joined	17:15
	← jelmer left	17:17
	→ hurikhan\|Work joined	17:17
	→ _raz_1 joined	17:17
	← _raz_1 left	17:18
	→ Hydrogen joined	17:18
	→ jelmer joined	17:19
Knirch	jel: your looking for deskins suggestion?	17:19
	you're	17:19
jel	Knirch: yep, that was him.	17:19
	Knirch: if you have it logged, that'd be great.	17:19
	(or just remember)	17:19
	chem_ is away: (automatically dead ) [BX-MsgLog Off]	17:20
	chem_ is idle, automatically dead [bX(l/on p/off)]	17:20
Knirch	[14:13] < deskin> my best advice at this point is to run git fsck --lost-found, look through the dangling commits listed with git show <hash>, and when you find the right one, git update-ref refs/heads/master <hash>	17:20
jel	deskin: don't suppose you're still around?	17:20
	→ stepz joined	17:20
jel	Knirch: thanks :)	17:20
deskinm	I am, but busy with $work; PM me and I can try to get to it	17:21
	← speakman left	17:21
deskinm	jel: ^^	17:22
	→ sgrimm joined	17:22
jel	deskinm: no worries, I was rushing earlier and didn't realise your instructions were so complete. That should do, I'll ask the channel in general if I get stuck :)	17:22
	deskinm: thanks again :)	17:22
stepz	I'm trying to figure out how to get my existing git repository with linear history into subversion, can anyone recommend how to approach that?	17:22
deskinm	sure	17:22
stepz	found a thread in the mailinglist archives that should be a solution, but I can't figure out for the life of me, what I ahve to rebase on what and merge with what	17:23
jel	git fsck --lost-found shows lots of missing blobs and a few (but still numerable) dangling blobs. Is that normal?	17:24
	← ashleyw left	17:24
kevlarman	stepz: easiest way is probably to make the first commit by hand (using regular svn), then git svn init, checkout the svn branch, and merge your history (make sure it's a fast forward), then dcommit	17:25
	← mithraic left	17:25
	→ bobesponja joined	17:26
	→ moore_ joined	17:26
stepz	how do I merge my history?	17:26
kevlarman	git merge <branch>	17:26
	(it might require some rebasing to convince git that the branches share a common history)	17:27
	← moore__ left	17:27
stepz	that rebasing I'm trying to figure out	17:28
	the first commit of the git repo and of the empty checkout have nothing in common	17:29
	→ halogenandtoast joined	17:30
kevlarman	it's really an empty checkout?	17:30
halogenandtoast	what is the difference between git update-index --assume-unchanged and git rm --cached? (and what do they do)	17:30
	← dantrell left	17:30
stepz	yep, no files in it	17:30
kevlarman	stepz: could probably use a graft	17:31
stepz	basically, I suppose I have to something like this: http://gist.github.com/56994	17:31
	← chem_ left	17:31
	→ Tv joined	17:31
stepz	ok, git-graft looks exactly like what I think will work	17:33
	thanks for the tip	17:33
kevlarman	stepz: echo $(git rev-parse B) $(git rev-parse a) > .git/info/grafts	17:33
	→ plediii joined	17:33
kevlarman	(and possibly filter-branch to permanently rewrite your history)	17:33
	→ softdrink joined	17:34
	← dwave left	17:34
	→ tclineks joined	17:35
stepz	I have to run for now, I'll see if it works tomorrow	17:35
	→ fallonetwo joined	17:35
stepz	thanks for the help	17:35
fallonetwo	hey folks, how goes it	17:35
	← Kathrin-25^away left	17:35
tclineks	could someone point me to some resources regarding my git-svn situation described here: http://dpaste.com/115727/ ? i'm looking at http://duncan.mac-vicar.com/blog/archives/282 and http://kerneltrap.org/index.php?q=mailarchive/git/2008/12/14/4416604 wondering if anything else out there would be illuminating	17:36
	→ zacharyc joined	17:36
	← Clebrate left	17:36
fallonetwo	i know with .git/info/excludes you can tell the add process what to skip over... is there a way to say explicitly "only add/check for changes in these files and directories" instead of the reverse, the implicit excluding?	17:37
	← jaeckel left	17:38
tclineks	fallonetwo: i think you could ignore * then add !(pattern) lines as well	17:39
	← Niamor left	17:39
fallonetwo	it'd nice to mitigate if my client decides to create a bunch of directories without telling me	17:39
	→ Niamor joined	17:39
	← rindolf left	17:41
skorpan	in git, is the "head" the "latest commit" in a specific branch?	17:41
kevlarman	skorpan: the current branch specifically (if you're on a branch)	17:42
Weiss	or for a bare repository, a reference to the branch which will be checked out on a non-bare clone	17:43
	(i think?)	17:43
	← lisyzavu left	17:43
skorpan	so it's the current branch, not some commit on the current branch?	17:44
fallonetwo	tclineks: hrmm, do you know of any good docs on the !(pattern) stuff for git exclude?	17:44
tclineks	using git-svn i was tracking a branch at /foo/branches/bar (cloned with --stdlayout --prefix=svn/ (svnurl)/foo) It was moved to /tmp-foo/branches/bar then /foo/trunk/foo -- wanting to guide git-svn to keep history correctly	17:44
	→ qogezil joined	17:44
kevlarman	skorpan: it's the most recent commit in whatever you have checked out (the current branch, or no branch in the case of a detached HEAD)	17:45
skorpan	oh, okay. thanks	17:45
Weiss	skorpan: cat .git/HEAD to see - it's a symbolic reference	17:45
	or usually is (?)	17:46
skorpan	yeah, it was	17:46
	← jel left	17:47
kevlarman	tclineks: i had a similar problem with svn mv, but i ended up just switching to git completely so i can't help	17:47
	← plediii left	17:47
tclineks	kevlarman: i know it's possible, just having to dive a little deeper than i'd like. might make a good blog post	17:47
	→ lucsky joined	17:48
	← lucsky left	17:48
	→ Kathrin-25^away joined	17:48
	→ qrush__ joined	17:51
	← tal67 left	17:52
fallonetwo	tclineks: so something like this to ignore all files except the explicitly defined ones? http://pastebin.com/m44aa52d6 i seem to be only getting admin/ and the *.php files there	17:56
tclineks	!admin/*	17:57
	probably	17:57
fallonetwo	ah, i see	17:57
	→ keymone joined	17:57
kevlarman	!admin/ != !/admin/ iirc	17:57
keymone	hi	17:57
	can i specify more that one remote for a branch?>	17:57
kevlarman	if you have the same folders it might match both of them	17:57
keymone	i just want git pull to sync with few remotes	17:58
fallonetwo	hrmm, that doesn't seem to work, it just sees the *.php files now	17:58
kevlarman	keymone: git pull = fetch + merge, and merge tends to fail on non-trivial merges when merging more than one branch	17:59
	← hurikhan\|Work left	18:00
	← Niamor left	18:01
	→ hurikhan\|Work joined	18:01
keymone	kevlarman: so is it possible? merging can fail anyway..	18:02
	→ Niamor joined	18:02
fallonetwo	is there something i'm missing that needs to go on the end of the !pattern exceptions? i'm not really seeing anything in the man pages	18:02
	← willb left	18:07
	→ willb joined	18:07
	→ drizzd_ joined	18:07
	← leachim6 left	18:08
	← softdrink left	18:08
	→ plediii joined	18:09
	← halogenandtoast left	18:10
	→ leachim6 joined	18:10
	← hdl2 left	18:12
	→ wwalker joined	18:12
	← qrush_ left	18:13
wwalker	I used to use git diff to see the changes in my local files, but I just did and it returns nothing. I have recently run git -add -u. Does that make git not see any differences? If so, how do I get git to show me the diff between current state and last committed?	18:13
	← jmesnil left	18:13
PerlJam	wwalker: sounds like you want git diff --staged	18:14
	← fallonetwo left	18:14
wwalker	PerlJam: thank you	18:15
	→ lebreeze joined	18:17
wwalker	PerlJam: --cached was what I was looking for (found by searching the git-add man page for stage :)	18:17
	thanks	18:17
	← cedricv left	18:18
moore_	wwalker: --cached diffs between the index and a named commit.	18:19
	→ cedricv joined	18:20
wwalker	moore_: or HEAD if nothing given, which is the behavior I need. I have staged the commit but then wanted to review the changes I was about to commit (I know I should have done that before I ran git add -u, but...)	18:20
	→ nick125_ joined	18:21
	← lebreeze left	18:21
PerlJam	wwalker: --cached and --staged are synonyms	18:21
	← muthu_ left	18:21
	← Kathrin-25^away left	18:21
	→ Kathrin-25^away joined	18:21
	→ daaku joined	18:21
	→ muthu_ joined	18:21
	→ Jc2k joined	18:22
	→ ToxicFrog joined	18:22
	← bcardarella left	18:23
	← dmlloyd left	18:23
	→ fynn joined	18:23
	→ hyperbor1ean joined	18:23
	→ jashmenn_ joined	18:24
	→ ceej joined	18:25
wwalker	PerlJam: my version (1.5.5.1) says --staged is invalid option	18:25
	→ dmlloyd joined	18:25
	← drizzd left	18:25
	← daaku left	18:26
	← jashmenn left	18:26
	← bdiego left	18:26
	← cthompson left	18:26
	← tanek left	18:26
	← Eridius left	18:26
	← alexkane left	18:26
	← hyperboreean left	18:26
	→ pogma joined	18:26
	→ Leonidas joined	18:26
	← willb left	18:27
	→ thiago_home joined	18:27
	→ Fearliss joined	18:28
canidae	is there some magic trick to clone a --bare repository over http? it looks for /info/refs, but that does not appear to exist in --bare repos	18:28
	→ willb joined	18:29
	→ andreaja joined	18:29
	← Tv left	18:29
pinholecamera	hrmm, still not having luck with those !pattern exceptions, there must be something simple i'm missing	18:30
	→ nick_h[litage] joined	18:30
	← hyperbor1ean left	18:30
	← Hydrogen left	18:30
	← qogezil left	18:31
	← harinath left	18:31
	→ qelyrupaz joined	18:31
	→ hyperboreean joined	18:31
thiago_home	canidae: did you run the command you need to run for http serving to work?	18:33
	git update-server-info	18:33
canidae	argh, i always forget that one	18:33
	thanks	18:34
	← Kbyte left	18:34
fonseca	tokkee: The problem you have was fixed in http://repo.or.cz/w/tig.git?a=commit;h=495fd0788ecb4aeda6ee3b47fbfc000f70368948	18:35
	tokkee: I will release a tig-0.14 with many regression fixes one of the coming days	18:36
	← rafael_lorandi left	18:36
	→ lorandi joined	18:37
pinholecamera	is there some kind of delimeter at the end of the !pattern excludes?	18:37
	← qelyrupaz left	18:37
	→ drobbins joined	18:38
drobbins	is it normal for "git am" to choke on certain mail headers and abort?	18:38
	and is there any way to deal with this issue?	18:38
	→ leqy joined	18:38
	← cmarques left	18:39
	← Kathrin-25^away left	18:40
	← plediii left	18:40
	→ Kathrin-25^away joined	18:40
	→ WALoeIII joined	18:42
	→ torarne joined	18:43
	→ krawek_ joined	18:43
	→ daaku joined	18:43
	→ bdiego joined	18:43
	→ cthompson joined	18:43
	→ tanek joined	18:43
	→ Eridius joined	18:43
	→ alexkane joined	18:43
	→ japhb joined	18:43
	→ dkr21 joined	18:43
	→ Kudos_ joined	18:43
	→ jim_c joined	18:43
	→ Fissure joined	18:43
	→ Toad joined	18:43
	→ dirker joined	18:43
	→ Toad__ joined	18:45
	→ w00t joined	18:45
	← lorandi left	18:46
	→ lorandi joined	18:46
	← mick_laptop left	18:47
	→ Sigma joined	18:51
	→ ceej_ joined	18:51
	← skorpan left	18:51
	← dirker left	18:51
	← Fissure left	18:51
	← dkr21 left	18:51
	← jim_c left	18:51
	← cthompson left	18:51
	← japhb left	18:51
	← Kudos_ left	18:51
	← daaku left	18:51
	← Toad left	18:51
	← alexkane left	18:51
	← tanek left	18:51
	← bdiego left	18:51
	← Eridius left	18:51
	→ knitt1 joined	18:55
knitt1	hi! how can i add a second root-commit?	18:55
drizzd_	knitt1: git symbolic-ref HEAD newroot	18:55
	or simply "git remote add" of an unrelated repository	18:56
	← icwiener left	18:56
	→ icwiener joined	18:56
	→ purestorm joined	18:57
	← Thargor left	18:57
purestorm	Can someone point me to a resource explaining how git svn rebase could ever fail when importing a new repository?	18:57
knitt1	drizzd_: thanks. i know i can pull another repo, but i wanted to do it in-place	18:59
drizzd_	purestorm: I'm not a git-svn user, but that sounds awfully vague	18:59
	→ dkerschn1r joined	19:00
	→ Thargor joined	19:00
parasti	purestorm: rebase can't work in an empty repository with nothing checked out. you want either git svn clone or git svn fetch	19:00
	← moore_ left	19:00
	→ japhb joined	19:00
	→ moore_ joined	19:01
purestorm	parasti: I have an empty svn repository with only trunk/tags/branches directories. I want to import my git stuff into that svn repository. How could rebase fail? The local patches should be conflict-less, should they not be? I'm trying to understand what's going on here...	19:01
	→ Kudos joined	19:01
	→ daaku joined	19:02
	→ bdiego joined	19:02
	→ cthompson joined	19:02
	→ tanek joined	19:02
	→ Eridius joined	19:02
	→ alexkane joined	19:02
	→ dkr21 joined	19:02
	→ Kudos_ joined	19:02
	→ jim_c joined	19:02
	→ Fissure joined	19:02
	→ dirker joined	19:02
	← krawek left	19:02
	→ ahupp\|fb joined	19:03
	→ rindolf joined	19:04
rindolf	Hi all.	19:04
parasti	purestorm: ok, then I misunderstood what you were doing	19:04
rindolf	git cvsimport -d :pserver:anoncvs@sourceware.org:/cvs/src/ -p -d,2009-01-01 gdb seems to be stuck at {{ cvs rlog: Logging src/etc }}. What can I do?	19:04
	← Thargor left	19:04
purestorm	parasti: Do you have an idea?	19:04
parasti	purestorm: like drizzd_ says, you'll have to provide more details... what you want to do, what commands you used, and what errors you got	19:06
	← leqy left	19:06
	← ceej left	19:07
	→ dkr21_ joined	19:07
purestorm	OK, sorry, I have to run. Evidently, this would take some time to understand :) Thanks, though.	19:07
	← purestorm left	19:07
	→ jaalto joined	19:07
	→ nirihy joined	19:07
	← bentob0x left	19:07
	→ Tv joined	19:08
	Toad__ → Toad	19:10
	← name left	19:10
	→ name joined	19:11
mutex	hmmm	19:11
	← Kudos_ left	19:11
	← dkr21 left	19:11
	← jim_c left	19:11
	← daaku left	19:11
mutex	how do i extract a patchset for applying to a source tree ?	19:11
	git-format-patch creates everything in email format	19:11
thiago_home	that's it	19:12
gitte	mutex: Have you found the "git diff" command yet?	19:12
	← rindolf left	19:12
thiago_home	mutex: the email format is the second best format available	19:13
	→ Thargor joined	19:13
mutex	gitte: yes of course, but maybe I have missed something	19:13
	because I can't seem to get it to produce proper merge commit diffs	19:13
	← moore_ left	19:13
	→ warthog9 joined	19:13
gitte	mutex: how should that go? A diff for a _merge_? Render me puzzled.	19:14
	→ moore_ joined	19:14
	← lorandi left	19:15
	→ rlorandi joined	19:15
	gitte thinks that somebody fell in love with a horrible name: "primer"	19:16
canidae	how "dangerous" is it to set up a repository with --enable=receive-pack? eg. can that completely ruin a repo?	19:16
	s/that/someone/	19:17
	← dirker left	19:17
	← Fissure left	19:17
	← cthompson left	19:17
	← alexkane left	19:17
	← tanek left	19:17
	← bdiego left	19:17
	← Eridius left	19:17
krh	5~	19:17
gitte	canidae: you mean git daemon?	19:17
canidae	yeah	19:17
	libc__ → libc	19:17
canidae	s/set up/share/ :)	19:17
gitte	canidae: you might have said that right away...	19:17
	→ bdiego joined	19:17
	→ cthompson joined	19:17
	→ tanek joined	19:17
	→ Eridius joined	19:17
	→ alexkane joined	19:17
	→ Fissure joined	19:17
	→ dirker joined	19:17
gitte	canidae: I can wreck your whole work if you let me push to your git daemon.	19:18
	← bdiego left	19:18
gitte	canidae: I can introduce backdoors, viruses, you name it.	19:18
	canidae: I can create fake branches, make up committer names, fake work in your name.	19:18
	canidae: and I can delete everything.	19:18
	→ tanek__ joined	19:18
	← nirihy left	19:18
canidae	that sounds about my definition of "dangerous"	19:19
gitte	canidae: so... "dangerous"? For some interpretations of "dangerous", the answer is "yes, absolutely".	19:19
	canidae: you should have known that before asking, though, as you'd allow write permissions without authentication.	19:19
	→ nyzaquvoc joined	19:19
	→ bdiego joined	19:20
canidae	basically i'd like to set up a repo <anonymous> can push to, and rather let me pick the stuff i want. sending them patches per mail has proven less than efficient	19:20
	→ ice799 joined	19:21
gitte	canidae: why don't you set up a repository on repo.or.cz and initialize the "mob" branch?	19:21
mansour	I don't understand when do I need to create repo with the --bare option. One site says, it's for repo without working directory. What does that mean? can someone explain ?	19:21
	← bdiego left	19:22
gitte	mansour: you had any exposure to Subversion or something like that?	19:22
canidae	i guess i could do that, yeh	19:22
mansour	yes, I did sirk with SVN for a while.	19:23
	but all from eclipse	19:23
gitte	mansour: so you know what a working directory is.	19:23
mansour	I did work with SVN	19:23
	→ mithraic joined	19:24
mansour	well, my understanding it's the dir holding the main project	19:24
	← nyzaquvoc left	19:24
	→ helo_ joined	19:24
mansour	gitte: am I right ?	19:24
gitte	mansour: close. It only holds a particular revision of your project.	19:24
	→ tuqagiw joined	19:25
gitte	mansour: to get another revision into your working directory, you have to ask the repository, right?	19:25
mansour	hum, ... ok ? so what is the deffrence between working and none working then ?	19:25
	yeah, u r right ?	19:25
	← tuqagiw left	19:25
mansour	yep true	19:25
	← johan-s left	19:25
mansour	how does this compare to "none-working" dir ?	19:26
gitte	mansour: with Git, the repository happens to be much more closely coupled to the working directory, it does not sit on a remote server, but in a subdirectory named .git/ (at least that is the common case).	19:26
	→ reli joined	19:26
gitte	mansour: for some applications, you do not need a working directory, it can even be undesirable, as with a "public" repository people can clone.	19:27
mansour	Ok, I see what u mean ? yes, when I create it with --bare, it has lots of files	19:27
gitte	mansour: there, you do not want any working directory, as the repository _already_ contains all the revisions, and you want people to clone it, not inspect a working directory.	19:27
	→ simosx_ joined	19:27
gitte	mansour: no, that is not it. If you create it with --bare, you do not have a .git/ subdirectory holding the repository.	19:28
	mansour: instead, you have the contents of the repository right there, in your current directory.	19:28
mansour	ok, let me make sure I understand.	19:28
gitte	mansour: because you asked not to have any working directory, the repository can live there.	19:28
mansour	I C.	19:28
helo_	i comitted a new file, made some changes, did a "git stash" and then did a "git checkout HEAD~3"... then i did a "git unstash", wanting to switch back to the latest commit, so did "git checkout HEAD", but the file i stashed was pulled into my (branchless) HEAD~3	19:29
	← mithraic left	19:29
	← dirker left	19:29
	← Fissure left	19:29
	← cthompson left	19:29
	← alexkane left	19:29
	← tanek left	19:29
	← Eridius left	19:29
helo_	and apparently it wants me to merge the new file before changing branches... i'd rather just leave HEAD~3 untouched...	19:29
mansour	So, when I create it with bare it has no project structure (dir heirarichy). When without bare, it greates the repo in .git. I noticed they are the same structure.	19:29
	I mean .git and the repo --bare	19:30
	I C now	19:30
	but here's the catch.	19:30
	← rubydiamond left	19:30
mansour	gitte an I right up to here ?	19:30
gitte	mansour: I am waiting for the "catch".	19:31
mansour	ok,	19:31
	→ plasticine joined	19:31
	→ mithraic joined	19:31
mansour	now, I created a project , and did "git init", then If i need to put it where athers can see it, then I need to push only what ?? .git ??	19:32
	gitte thinks that jidanni definitely has way too much time on his hands, but does not use it for useful stuff.	19:32
gitte	mansour: umm. How do you want to push _directories_?	19:32
doener	gitte: but apparently he got a crystal ball ;-)	19:32
	← muthu_ left	19:33
mansour	I created, a project and it was cool "assume, it never really happened". And I want to put it on my remote repo	19:33
	→ gammons joined	19:33
gitte	doener: he has mushrooms, cooked to a nice mash, where I have the organ using 25% of my oxygen intake.	19:33
mansour	do I just copy .git with ssh to the server ?	19:33
gitte	mansour: I suggest reading up on "publishing your work" or some such in the Git manual.	19:34
	← ceyusa left	19:34
mansour	ok, I will. thank you.	19:34
	You answered the main question I have, and I can dig further now. thak you gitte	19:34
gammons	Hey I have a (hopefully not too n00b) question. I have a local branch, called local/release, and on my origin server, I have a branch called origin/release. I'd like to push the changes I made in my local branch to the remote branch. I've been calling 'git push origin release', but the changes don't seem to be making it to the remote branch. Anything obvious I am doing wrong?	19:34
	→ agib joined	19:36
	← nick_h[litage] left	19:36
helo_	where can i find out what "* Unmerged path somefile.cpp" when i try to "git stash pop" means?	19:36
thiago_home	helo_: it means there was a conflict	19:37
gitte	gammons: yes, you should rename your branch to "release" with "git branch -m local/release release". No need for an extra hierarchy for local branches.	19:37
	helo_: just look for "<<<<<<" in somefile.cpp.	19:38
gammons	gitte: hah! that worked! thanks a log	19:39
	s/log/lot	19:39
	actually log is funnier	19:39
	→ henry_kleynhans joined	19:41
	ph^_ → ph^	19:41
	→ resmo joined	19:43
	← simosx left	19:43
	← charon left	19:45
	→ eddyp joined	19:47
	→ ahupp\|fb_ joined	19:49
	← mikl left	19:49
	← cgardner left	19:50
	← qrush__ left	19:51
	← ashleyw_ left	19:52
	→ felipec joined	19:54
	← plasticine left	19:57
	← korpios left	19:57
	→ krawek__ joined	19:57
	← henry_kleynhans left	19:59
	→ sdboyer-laptop joined	19:59
	← bryanl left	20:02
	← hurikhan\|Work left	20:03
	← LotR left	20:04
	→ hurikhan\|Work joined	20:05
	← sdboyer-laptop left	20:07
	→ NSaibot joined	20:08
	→ shapeshed joined	20:10
	← ahupp\|fb left	20:10
	→ plediii joined	20:10
	← gammons left	20:11
	→ bryanl joined	20:12
	← scientes left	20:13
	← rwaite left	20:13
	← plediii left	20:13
	← krawek_ left	20:14
	→ earcar joined	20:14
	→ plediii joined	20:14
	← hurikhan\|Work left	20:16
mansour	gitte, u there ?	20:16
gitte	Sort of.	20:17
	→ hurikhan\|Work joined	20:17
	→ cgardner joined	20:17
	← rlorandi left	20:17
mansour	cool, I looked around and couldn't find a clear doc that gives me what I want.	20:17
	→ dramsay joined	20:17
mansour	again, I created a project and decided to put it on a remote, pushing doesn't work as there's no repo on the remote yet. So do I create it with --bare on the server, or do I just copy ".git" dir	20:18
	?	20:18
	→ charon joined	20:18
	→ bieneff joined	20:18
	← ahupp\|fb_ left	20:19
	← thumper left	20:20
	→ foca_ joined	20:21
	→ thumper joined	20:21
gitte	mansour: create it with --bare.	20:23
acoc	if I git mv a file, does the new file retain the history of the previous file	20:23
gitte	mansour: or use one of the many providers that make it easy for you.	20:24
	acoc: history in Git is revision-based, not file-based, so your question does not make full sense to me.	20:24
	simosx_ → simosx	20:24
	→ nick_h[litage] joined	20:25
	→ dirker joined	20:25
	→ cthompson joined	20:25
	→ tanek joined	20:25
	→ Eridius joined	20:25
	→ alexkane joined	20:25
	→ Fissure joined	20:25
	← bgmarete left	20:25
	← hurikhan\|Work left	20:25
acoc	gitte, if I were to do a git log of the new file, would it contain the previous commits or only the last one that moved it	20:25
mansour	I C, gitte thank you, I will	20:26
gitte	acoc: huh? If you commit the result, the parent commit of the new one will be the current HEAD, so I really do not get your question. What's it with the files?	20:26
	→ Marmouset joined	20:27
acoc	gitte, if you were doing a diff of that new file from an older revision, would it be able to see that it's location changed an compare it to the file as it was in it's origional location	20:28
	→ ahupp\|fb joined	20:29
krh	acoc: yes, use git log --follow (and don't mind gitte, he's being deliberately obtuse ;)	20:29
MarcWeber	gitte: I think acoc means: When renaming a file is there a command to conviniently follow the renamings to get all patches containing that file?	20:29
	I've checked, qgit does'nt recognize this	20:29
	gitte → obtuse-gitte	20:29
kevlarman	-M?	20:30
	← foca left	20:31
	← jashmenn_ left	20:31
acoc	krh, thanks that was helpful	20:34
	→ docwhat joined	20:34
canidae	uh. someone got "user.name=Foo Bar", although in logs it just show "Foo". he's using cygwin, anyone heard about this before?	20:36
	Ratler_ → Ratler	20:37
	→ robinr joined	20:38
	→ dreiss joined	20:39
obtuse-gitte	canidae: it's probably even unquoted, right?	20:40
	canidae: so make sure that "Foo Bar" is quoted in .git/config	20:41
canidae	aight, we can try that	20:41
	← plediii left	20:41
	→ ijcd joined	20:42
	→ ephebo joined	20:43
canidae	didn't help :(	20:44
	→ johnw joined	20:44
	← Tuomas left	20:45
	← vuf left	20:46
	→ mczepiel joined	20:46
	← mczepiel left	20:46
	← jrmuizel left	20:46
	→ jrmuizel joined	20:47
	→ qrush_ joined	20:48
	→ ekidd joined	20:48
	→ plediii joined	20:48
	← johnw left	20:49
obtuse-gitte	canidae: recorded commits will not change...	20:49
canidae	he tried new commits	20:50
	or so he said, but he ran off now. i'll check that again when he gets back (which may take some hours)	20:50
obtuse-gitte	Check if "git var -l" shows the correct name.	20:50
canidae	thanks for the help, though :)	20:50
	obtuse-gitte runs off, too	20:51
	← obtuse-gitte left	20:51
	→ HanLecter joined	20:55
	← MarcWeber left	20:55
	→ earcar_ joined	20:56
	→ GlycaDS joined	20:58
	→ jasber joined	20:58
	← travisjeffery left	20:59
	← kraymer_ left	20:59
	← GlycaDS left	21:00
	← Kathrin-25^away left	21:00
drobbins	what is the best way to figure out why I can't push (a non-fast-forward merge)?	21:01
eddyp	I have multiple branches and two remotes; I want that push to one of the remotes to push two branches and the other to push just one; how do I configure th elocal repo to specify the to branches I want pushed for remote 1?	21:01
	→ Kathrin-25^away joined	21:01
doener	drobbins: I like "gitk --left-right local...remote"	21:01
	← earcar_ left	21:02
drobbins	doener: I just merged some changes in, but I don't understand why this broke things.	21:02
eddyp	the remotes are origin and puorcz	21:02
	puborcz	21:02
drobbins	doener: how about on cmdline?	21:02
doener	eddyp: set remote.origin.push and remote.purocz.push	21:02
eddyp	puborcz is the one which I want to contain two branches	21:02
drobbins	doener: just applied some patches and did a merge, not sure why it's not a basic "here are some recent commits.."	21:03
eddyp	doener: I figured out that part, but how do I specify the two refspecs for puborcz?	21:03
doener	drobbins: same deal. Though I'd probably just use "local..master" (two dots) with git log	21:03
drobbins	doener: maybe it's due to some of the commits being chronologically earlier	21:03
doener	drobbins: telling the left and right side apart in the text log isn't something my brain likes	21:03
drobbins	ok, maybe I just need to pull and figured it was trickier than it really was.	21:04
doener	drobbins: the log command then tells you what's missing from the local branch	21:04
drobbins	looks like I'm missing a commit that is upstream, that I pushed from another box.	21:04
	doener: excellent	21:04
	doener: while I have you here....	21:04
	doener: I've been getting inline patches in emails...	21:04
eddyp	doener: now I have : "refs/heads/mob:refs/heads/mob", but I can't figure out the format to add the master, too	21:04
doener	eddyp: add a second remote.<whatever>.push line	21:05
	eddyp: if you use "git config", use the --add option	21:05
	drobbins: save as mbox/maildir, feed to "git am"	21:06
eddyp	doener: that was the missing bit, thanks	21:06
drobbins	doener: see, I did that... and it said that it was an invalid patch or some such thing, error on line 28 or thereabouts (unfortunately don't have the error in front of me right now)	21:06
	→ keystr0k joined	21:06
drobbins	doener: are my submitters generating these patches the wrong way if I'm getting that error?	21:07
doener	ah, the header thing, saw that while flying through my scrollback	21:07
drobbins	yep	21:07
	← dmlloyd left	21:07
doener	hard to tell without the email at hand, at least for me	21:08
	→ dmlloyd joined	21:08
doener	I don't recall hearing of such problems caused by headers. The mail body, yes, but not headers.	21:09
	→ plediii_ joined	21:09
	← earcar left	21:14
	← markelikalderon left	21:14
	→ joevano joined	21:14
	← drobbins left	21:15
	→ drobbins joined	21:15
	← drobbins left	21:18
	← Voker57 left	21:19
	← esparkman left	21:19
	→ drobbins joined	21:19
	← cmarcelo left	21:19
	→ Voker57 joined	21:20
	→ cmarcelo joined	21:20
	→ Murr joined	21:23
	← derwolf left	21:25
	→ esparkman joined	21:26
	← cmarcelo left	21:28
	← seezer left	21:32
	← warthog9 left	21:32
	→ warthog9 joined	21:33
	→ seezer joined	21:33
keystr0k	I've been working on a branch... can I pull into my master without merging, just automatically assume the branch is the version of the file(s) I want.	21:36
	?	21:36
	← ia left	21:36
keystr0k	I guess I need my master branch to look exactly like my "testing" branch.	21:36
jayne	there is an "ours" merge strategy	21:37
	→ ia joined	21:37
drizzd_	keystr0k: if you don't want your local changes, I'd just throw them away and create a new branch based on the remote master	21:37
	← pierre- left	21:37
	→ bradly joined	21:38
keystr0k	drizzd_, well, all of the data/branches are local... It's just a repo on my local machine. I made changes to branch X and I need to make sure master looks just like branch X.	21:38
	← plediii left	21:39
drobbins	doener: had to reboot due to keyboard not responding	21:39
	doener: I have a paste for you	21:39
	doener: this is what I am experiencing: http://www.privatepaste.com/16Y3Yrw2Sl	21:39
doener	and line 34 is?	21:40
drobbins	doener: those patches were saved by mutt into a maildir	21:40
drizzd_	keystr0k: git checkout master; git reset --hard X;	21:40
keystr0k	drizzd_, aaaah.	21:40
drobbins	doener: one moment.	21:40
keystr0k	drizzd_, thanks. that looks good.	21:40
	→ johnw joined	21:40
keystr0k	I knew about git reset hard but I didn't know you could reference a different branch.	21:41
	← resmo left	21:41
drizzd_	keystr0k: if X was based on master, git merge X would do the trick too (that's the normal mode of operation)	21:41
drobbins	doener: http://www.privatepaste.com/e40rBMLy7b see line 34 (I'm assuming it means line 34 from the top of the mail header)	21:42
keystr0k	drizzd_, okay... yeah, it is based on master... I'll do a normal merge and if the results aren't desired I'll try the other method.	21:42
	← torarne left	21:44
	← priidu_ left	21:45
	metze_away → metze	21:45
	← GeertB left	21:46
drobbins	doener: any ideas? does it look like a git bug of some kind?	21:47
doener	drobbins: no, seems to skip the header before starting to count the lines	21:47
drobbins	ok	21:47
	let me get the patch body then.	21:47
doener	hm, or maybe not	21:48
	doener is confused right now	21:48
drobbins	hrm.	21:48
doener	I shouldn't mess with my terminal colors while working	21:48
	black cursor on black background just doesn't work	21:48
drobbins	maybe mutt is word-wrapping the inline attachment before it saves it.	21:48
	I notice that everything is word-wrapped so that it meets the standard 78 or so column limit	21:49
	→ stevecrozz joined	21:50
doener	using "s" or "C" in mutt? never seen that messing with my mails	21:50
stevecrozz	can someone please help me troubleshoot this error: warning: remote HEAD refers to nonexistent ref, unable to checkout.	21:50
	← Dvyjones left	21:50
stevecrozz	I've published a git repository over http, and I can't get 'git clone' to work with it	21:50
drobbins	doener: looks like it's word-wrapping.	21:51
	doener: I'm assuming that the patch can't be word-wrapped or it b0rks git, understandably.	21:52
doener	The header on line 11, 12 looks funny	21:52
	though that's far mor than 78 chars	21:52
	← docwhat left	21:53
	← Catfish left	21:54
	← Weaselweb left	21:54
stevecrozz	"error: warning: remote HEAD refers to nonexistent ref, unable to checkout." -- how do we fix this ?	21:54
doener	drobbins: btw, which git version is that?	21:54
RandalSchwartz	stevecrozz - fix the remote head. :)	21:55
drobbins	doener: 1.6.1.2 and 1.6.1.0	21:55
stevecrozz	thanks RandalSchwartz... but how?	21:55
RandalSchwartz	is it a bare repo?	21:55
stevecrozz	no, its got an initial commit	21:55
RandalSchwartz	no - bare means "does it have a working tree"	21:56
stevecrozz	i was able to do: "git push upload master" from my project repo	21:56
RandalSchwartz	err. the opposite of that	21:56
stevecrozz	that worked	21:56
	→ cubuanic joined	21:56
	RandalSchwartz waits for bare vs non-bare to be answered	21:56
doener	did you run "git update-server-info" after the push?	21:56
stevecrozz	I'm very new to git... I'm not sure how to answer the question	21:57
RandalSchwartz	that's not needed if it's bare	21:57
stevecrozz	not i didn't run that	21:57
RandalSchwartz	stevecrozz... is there a "checkout"?	21:57
doener	RandalSchwartz: hm? That's always required for http	21:57
RandalSchwartz	local files besides just the .git?	21:57
	oh - didn't realize this was http	21:57
	when did he say that?	21:57
stevecrozz	i started a project here, built a git repo, then uploaded it to my web server....	21:57
	RandalSchwartz scrolls back	21:57
doener	about 7 minutes ago	21:57
RandalSchwartz	ok, missed that	21:58
	then yes, you need the update-server-info step	21:58
	but eventually, you should get away from http transport	21:58
drobbins	doener: did some research, and it appears to be an issue with the email as sent out by the upstream developer.	21:58
doener	drobbins: mind bouncing that to me?	21:58
stevecrozz	RandalSchwartz: this is a brand new project... i can move away now if need be	21:58
	you prefer git:// ?	21:59
	→ markelikalderon joined	21:59
drobbins	doener: the email? sure, what's your addy?	21:59
doener	[email@hidden.address]	21:59
RandalSchwartz	you make it sound like it's me personally that prefers it. :)	21:59
	but yes, for best functionality, you want git-over-ssh protocol	21:59
	→ abbe joined	21:59
RandalSchwartz	that's what everthing is optimized for	22:00
stevecrozz	hmmm.. i want to avoid creating shell accounts for everyone who needs to use the git repo....	22:00
RandalSchwartz	isn't that what gitosis is about?	22:00
	→ agenteo joined	22:00
stevecrozz	i have no idea... im a super noob, is it?	22:01
RandalSchwartz	... http://scie.nti.st/2007/11/14/hosting-git-repositories-the-easy-and-secure-way	22:01
	that might be outdated	22:01
	but it says what I was about to say	22:01
stevecrozz	ok	22:01
RandalSchwartz	... It manages multiple repositories under one user account, using SSH keys to identify users. However, users do not need shell accounts on the server, instead they will talk to one shared account that does not allow arbitrary commands	22:01
doener	stevecrozz: for read-only access git:// using the daemon is fine. For push, you'll want ssh, maybe in combination with gitosis or similar things	22:02
stevecrozz	ok..	22:02
	→ RoPP joined	22:03
drobbins	doener: gimme a few mins and I'll get it out to you	22:04
doener	no need to hurry	22:04
	← ekidd left	22:04
drobbins	just need to pretend	22:04
	(that I'm hurrying) - so I don't forget :)	22:04
doener	heh, I know how that works :-)	22:05
stevecrozz	hmm gitosis has an ubuntu package in the official repo	22:05
	ill try it	22:05
	← cemerick left	22:05
	← Murr left	22:07
	→ Murr joined	22:07
	dwmw2 → dwmw2_gone	22:08
	→ dwave joined	22:09
	← jesselucas left	22:11
	← esparkman left	22:13
	← ferdy left	22:14
	← markelikalderon left	22:14
drobbins	doener: bounced. and I found out that the patch was sent via kmail, which could explain the mangling.	22:14
	← charon left	22:14
	← cannonball left	22:15
	→ ferdy joined	22:15
doener	drobbins: yep, the patch definitely looks broken, just from the huge context in the first hunk	22:15
drobbins	doener: ok, cool, good to know. thank you for your help.	22:16
	← robinr left	22:16
	← vbabiy left	22:17
doener	xenoterracide: there's a section about sending patches with KMail in Documentation/SubmittingPatches	22:17
	← shapeshed left	22:17
xenoterracide	doener: interesting but I've a more interesting issue now (don't know that I like kmail anywho	22:17
	← tanek__ left	22:18
doener	drobbins: hm, maybe you meant gmail?	22:18
	← bieneff left	22:18
xenoterracide	git format-patch -1 ea11098821f1495bd499e6926dd3a4232da839b7 --stdout -M funtoo/funtoo.org I'm trying that... but the patch being sent is this one d9aefeeb which happens to be the latest head on funtoo/funtoo.org, maybe I've missed something?	22:19
	metze → metze_away	22:19
doener	xenoterracide: you're passing two commits to format-patch	22:20
	xenoterracide: -1 limits it to the most recent one	22:20
	xenoterracide: just drop the funtoo/funtoo.org at the end	22:21
xenoterracide	ah	22:21
	← bryanl left	22:21
doener	and if you want multiple single commits, use --no-walk instead of -1	22:22
drobbins	sounds like we need a HOWTO for format-patch submissions	22:22
	→ Hydrogen joined	22:22
doener	basically, format-patch works like "log", with one exception, if you give just a single commit (and no -1, --no-walk or whatever), it defaults to $commit..	22:23
	so "git format-patch foo" and "git log foo" differ, the former is like "git log foo.." (WRT the commit range it operates on)	22:24
	← dwave left	22:24
	→ Flathead joined	22:24
doener	but "git format-patch foo..bar" and "git log foo..bar" work on the same commits	22:25
	← NSaibot left	22:25
drobbins	will test in a bit.	22:27
	→ mark[oz] joined	22:27
xenoterracide	doener: hmm... can't find -1 in the man page	22:28
	or no-walk	22:29
	← bobmcw left	22:29
	← radarek left	22:29
	← stevecrozz left	22:29
doener	--no-walk is in the rev-list man page, -1 would be -n 1 if one wasn't lazy	22:30
	cbrake → cbrake_away	22:30
xenoterracide	doener: why does the SubmittingPatches guide (for git) tell you to use the remote-branch?	22:30
doener	-n is also in the rev-list man page	22:30
xenoterracide	be nice if they were in the format-patch man page	22:30
	since that's the command I'm running	22:30
doener	xenoterracide: then the whole rev-list man page would be in there... and the diff manpage as well	22:31
	would be slight overkill	22:31
xenoterracide	yeah I suppose	22:32
	← foca_ left	22:32
xenoterracide	doener: at least I know where to look now	22:32
doener	xenoterracide: it uses origin/master because it then operates on origin/master.. == origin/master..HEAD ==> all commits in the history of the checked out commit that aren't in upstream	22:32
	it's just a shortcut	22:33
kansan	i forked a git project. now im within my checkout dir ... and i want to pull the commits from the mainline of the project (not my forked version).... how can i do this?	22:33
	i forked via github that is	22:33
doener	kansan: add the mainline repo as a remote (git remote add <name> <url>)	22:33
	← Ryback_ left	22:33
DrNick	git remote add -f mainline git://mainline	22:33
	→ aroben joined	22:34
	→ purestorm joined	22:35
kansan	ok i did that	22:36
	now what do i do	22:36
doener	with -f or without?	22:36
purestorm	Hi, so back to my issue. I have a git repository and now I want to use a subversion repository as the central server. I found this how-to: http://code.google.com/p/support/wiki/ImportingFromGit. Now, I get conflicts on git svn rebase and I do not understand how this could happen, my git repository contains a linearized of patches, right?	22:37
	← mcella left	22:37
	← mark[oz] left	22:38
	→ perezd88 joined	22:39
	→ mark[oz] joined	22:39
	← Phlogi left	22:40
	→ dmq joined	22:42
	dmq wonders if its a bug that git diff -w doesnt ignore missing newline at eof.	22:43
	← keystr0k left	22:43
	→ bdiego joined	22:43
	→ rogg joined	22:43
johnw	dmq: sounds ilke it tome	22:43
	→ mcella joined	22:46
	← mcella left	22:46
	← felipec left	22:47
	← Kathrin-25^away left	22:48
	→ Kathrin-25^away joined	22:49
kansan	say i have a git repo that has in the root: foo/ ... now i decide i want to yoink out foo and put it in a brand new repo where its root. can i do this?	22:49
	← aroben left	22:49
	→ sjzzalx joined	22:50
	→ krawek joined	22:50
sjzzalx	Hi all. When I try to git svn clone, I get http://pastebin.com/m6fe99d5a on password prompt. I don't like this. Anyone know how to fix it?	22:50
	→ aroben joined	22:52
dmq	that looks like you have a slightly broken perl install actually.	22:52
	you need to install the perl Term::Readkey module somehow.	22:52
	→ esparkman joined	22:52
sjzzalx	dmq: Ok, awesome, thanks. : )	22:53
dmq	its a little weird tho, as afair that module should be around already.	22:53
	→ chuck joined	22:54
dmq	unless your vendor has shipped a broken perl.	22:54
chuck	hey, i've got a rather complex problem, but I figure i'll ask, as git might be able to do it	22:54
dmq	oh, no, wait, im wrong.	22:54
	its not a core module.	22:55
	but i you do need to install it.	22:55
chuck	I want to write a fork of something that is using SVN, and trunk/ is always ready-to-run, so I want to use the latest SVN. how would I go about doing that in git so I could easily merge in the latest changes upstream?	22:55
	← EiNZTEiN left	22:56
dmq	ah right, Term::ReadLine is core.	22:56
	since 5.002	22:57
	→ lordpil- joined	22:57
Marmouset	chuck, cannot you use git-svn ?	22:57
chuck	Marmouset: would that work for this?	22:57
sjzzalx	dmq, thanks, I got it installed and that problem's gone now actually. But now I get error: More than one value for the key svn-remote.svn.fetch: :refs/remotes/git-svn	22:57
	. :( brb, gotta restart pidgin	22:57
	← sjzzalx left	22:57
chuck	I have no idea what git-svn is all about, but I'm open to try it	22:57
	← bantic left	22:57
	→ sjzzalx joined	22:57
	→ flaguy48 joined	22:58
Marmouset	chuck, git-svn enables you to create a git repo from a svn one. try git-svn clone followed bu the depot svn utl	22:58
	url	22:58
dmq	that part i cant really help with sjzzalx.	22:58
Marmouset	From the man page : git-svn - Bidirectional operation between a single Subversion branch and git	22:58
	← Weasel[DK] left	22:58
Marmouset	chuck, I use that to track svn repos	22:58
chuck	Marmouset: would that be able to have the upstream code in a branch of something so I can merge	22:59
sjzzalx	dmq, I see, thanks.	22:59
	← krawek__ left	22:59
dmq	im kinda curious why git svn is using Term::ReadKey and not Term::ReadLine.	23:00
	:-)	23:00
	→ aroben_ joined	23:00
	← sjzzalx left	23:01
	→ sjzzalx joined	23:01
Marmouset	chuck, you get a git repo. So you do whatever git let you do :)	23:01
	← SRabbelier left	23:02
chuck	so would I do something like: git svn clone http://whatever/svnroot/whatever	23:02
	then make a branch for my fork?	23:02
	that seems backwards	23:02
Marmouset	git-svn not git svn	23:02
chuck	maybe I can use git pull?	23:02
	yeah w/e	23:03
Marmouset	check git-svn options	23:03
	you can tell it what is the branch/trunk/tag directory and so on	23:03
	→ LotR joined	23:05
	← Voker57 left	23:05
	→ markelikalderon joined	23:08
	← reli left	23:10
	← Murr left	23:11
	← EmilMedve left	23:11
	← solydzajs left	23:13
	← lordpil left	23:13
	→ pathall joined	23:14
	← jmspeex left	23:14
pathall	hi, i would like to merge several repos under an umbrella repo. so i have app1, ... app5, and i want to create umbrella/{app1,...app10} . what command should i be looking at?	23:14
	oops, s/5/10/ :)	23:15
	→ vbabiy joined	23:15
	← aroben left	23:17
	→ a-priori joined	23:17
purestorm	How could a git rebase from an external repository to an empty repository fail? I thought that a repository contained conflict-free patches.	23:18
klmann	pathall: this is how i did it: go to the repos app1...5 and move the root content into a new created direcotry (mkdir app_n; git mv * app_n). then merge the repositories to the umbrella-repo	23:18
	→ Ariens joined	23:19
	→ qrush__ joined	23:19
	← helo_ left	23:19
vomjom	purestorm, i don't even think it makes sense to rebase there	23:20
	maybe you misunderstand what rebase does?	23:20
purestorm	I have to :/ it is a svn repository.	23:20
	← qrush_ left	23:20
purestorm	Or I am understanding something wrong.	23:20
	I have a git repository	23:20
pathall	klmann: ah good idea, thanks	23:20
	← aroben_ left	23:21
purestorm	and now I want to push everything to a subversion repository following http://code.google.com/p/support/wiki/ImportingFromGit	23:21
vomjom	purestorm, did you do svn clone?	23:21
klmann	pathall: but i have to warn you. i moved from svn to git and i had this idea of one umbrella-repo too, but it showed that it is not as flesible as multiple git repositories	23:22
	*flexible	23:22
pathall	klmann: i just want to squeeze a bunch of stuff under one app on github because of the private app limit	23:22
vomjom	http://gitready.com/intermediate/2009/01/31/intro-to-rebase.html <--- this is a good explanation of what rebase does by the way	23:22
	→ urbanmonk joined	23:22
purestorm	vomjom: https://gist.github.com/801115105eaeeaf3c018	23:22
klmann	pathall: all right, then its ok ;)	23:23
RandalSchwartz	oof - what an evil web page	23:23
vomjom	purestorm, is that what you did?	23:23
purestorm	vomjom: The paste, yes.	23:23
	and it fails	23:24
RandalSchwartz	why do people think it's cute to change the annotations for "this is a link" to something really unfamiliar	23:24
vomjom	purestorm, sorry, no idea why it's failing then. what message are you getting?	23:24
purestorm	error: patch failed: Sequential/SConscript:1	23:25
	error: Sequential/SConscript: patch does not apply	23:25
	Hrm, this could be linked to a merge later on...	23:25
	→ fejo joined	23:26
purestorm	or a branch	23:26
	hm, is it possible to rebase something back in time to something back in time?	23:26
sjzzalx	purestorm, git reset?	23:27
	← LotR left	23:27
sjzzalx	not sure if it does everything you want but it will put your repo back to an old commit	23:27
	→ krawek_ joined	23:30
	→ radarek joined	23:30
purestorm	hrm, OK	23:30
	Well, I'll resolve it by hand then.	23:31
	Thanks, nevertheless.	23:31
	← Sigma left	23:32
chuck	oh good god, git-svn clone downloads every single revision T^T	23:33
	← a-priori left	23:33
klmann	chuck: yeah, it actually does what it says ;)	23:33
pathall	hmm, klmann, what if i just cloned the dirs without making the subdirectory? like, git clone /home/blah/app1/.git /home/blah/umbrella/app1	23:33
chuck	klmann: how do I grab the latest revision? xD	23:33
klmann	pathall: you would not have the history of the imported repository in the main repo	23:34
	pathall: ... i think. actually, i didnt try it	23:34
pathall	klmann: oh, clone doesn't keep the history?	23:34
klmann	klmann: man git-svn. i dont know the command	23:34
Gitbot	klmann: the git-svn manpage can be found at http://git.or.cz/man/git-svn	23:34
klmann	chuck: yes, but it does not merge the hostory in the umbrella-repo	23:35
	pathall: so if you use git log in the umbrella-repository, you will not see the changes in the other repositories	23:35
chuck	klmann: how do I just grab the latest code	23:35
pathall	klmann: yep, you're right, just checked that. thx	23:35
	← alley_cat left	23:35
klmann	chuck: i dont know	23:36
	← boto left	23:36
chuck	.-.	23:36
	→ Guest58011 joined	23:37
Guest58011	Hello everyone. quick question: I have 2 local repos I want to merge.. can I do this? checkout a branch or tag from a diff local repo?	23:38
	→ tvw joined	23:38
	→ aroben joined	23:38
shd	Guest58011: git pull /path/to/other/repo	23:39
Guest58011	and I can point to a SHA1 in that repo correct?	23:39
	or tag?	23:39
	git pull /path/to/other/repo/tag\|\|sha1?	23:39
shd	you can do it like this:	23:39
	cd /path/to/other/repo	23:40
	git checkout -b newbranch SHA1/tag	23:40
	cd /path/to/first/repo	23:40
	git pull /path/to/other/repo newbranch	23:40
	→ aroben_ joined	23:40
shd	hmm.. i think git pull /path/to/other/repo SHA1/tag might work as expected..	23:40
Guest58011	awesome! thank you very much. git rules	23:41
shd	please try the latter first..	23:41
Guest58011	will do	23:41
	← CrazyTux1 left	23:41
shd	Guest58011: often, what people do is track the other repo as a remote	23:42
purestorm	Gah, now I got a segmentation fault in git svn dcommit. Am I lost now, my newer commits do not show any more and I am stuck with an unclean repository state.	23:42
shd	Guest58011: then you can simply do: git merge SHA1	23:43
	Guest58011: branch name isn't really important	23:43
	→ aroben__ joined	23:44
	← WALoeIII left	23:44
Guest58011	ok the fetch is working taking a while..	23:45
	← akitada left	23:46
	← krh left	23:47
plopix	why is a fast-forward merge failing on that? http://pastebin.com/d3fe9e50c	23:48
	→ synsol joined	23:49
	← aroben_ left	23:50
	→ aroben_ joined	23:50
	→ moore__ joined	23:51
	← purestorm left	23:51
	← krawek left	23:51
	← tvw left	23:52
	← moore_ left	23:52
Circuitsoft	Is it possible to plug filetype-specific diff modules into git?	23:54
	→ softdrink joined	23:55
	→ grover joined	23:56
	→ krawek__ joined	23:56
Arrowmaster	Circuitsoft: http://www.kernel.org/pub/software/scm/git/docs/gitattributes.html#_generating_diff_text	23:57
	← aroben left	23:58
	← agile left	23:58
grover	if one does "git pull git://git..../proj.git master" do you pull in that tree's lightweight or heavyweight tags?	23:58
Guest58011	Can I get the SHA1 of a tag?	23:59
Circuitsoft	Guest58011: git rev-parse	23:59
	<tagname>	23:59
	← ph^ left	23:59

Logs Search ←Prev date Next date→

Channels Documentation