IRCloggy #git 2011-06-09

Provider of IRC logs since 2005.

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

	2011-06-09
	← budsondixon left	00:00
	→ Pupeno joined	00:01
	← ralc left	00:02
	→ angelsl joined	00:03
	→ budsondixon joined	00:04
	→ flowbee joined	00:05
	← banisterfiend left	00:05
flowbee	how do i add all modified (previously commtted files) to the index	00:06
	→ segher joined	00:06
	→ stepnem joined	00:08
	← AntelopeSalad left	00:09
	← budsondixon left	00:09
canton7	add -u	00:09
	(or commit -a)	00:09
	add . works as well for the current dir. subtle differences though, look em up	00:10
	also add -A if you want to delete stuff as well, tho i'd advise against it	00:10
	→ banisterfiend joined	00:10
	← tvw left	00:10
	← f0i left	00:12
	← banisterfiend left	00:12
	→ AntelopeSalad joined	00:13
	→ budsondixon joined	00:13
	← ceej left	00:13
	← gusnan left	00:13
	← saschagehlich left	00:14
offby1	huh, what's the risk there?	00:14
	← gebbione left	00:14
	← kenichi left	00:14
	← notslad left	00:15
	→ normanrichards joined	00:16
	← fedesilva left	00:16
	← budsondixon left	00:17
canton7	i'm trying to remember... it might have been on the mailing list someone was talking about removing it at some point?	00:18
offby1	reading the man page now	00:19
	→ ISF_ joined	00:19
	← cesario left	00:20
	← Agos left	00:20
	→ pdtpatrick__ joined	00:20
	← ns5 left	00:20
	→ ns5 joined	00:20
ns5	is it possible to check out a copy of sources files from a remote repo? .git dir is not needed, only the source code	00:21
canton7	you can clone the repo with --depth of 1 if you want, but you're using git for the transfer, therefore git repos have to be involved	00:21
	→ budsondixon joined	00:22
	← ISF left	00:22
pdtpatrick__	how can u see the files that were last checked in?	00:22
canton7	define "checked in"?	00:22
pdtpatrick__	committed	00:23
canton7	so the files that were in the last commit? git show	00:23
pdtpatrick__	lets say files were committed last week and that was the last commit. How can i see what those files were ?	00:23
canton7	my bad, that's the diff, not the actual files.	00:24
	actually that's what you wanted isn't it? the changes that were committed?	00:24
	canton7 is getting a bit too tired for this	00:24
	→ darknop joined	00:24
pdtpatrick__	found it	00:25
	git log --stat --summary	00:25
	→ arpegius joined	00:25
darknop	how do I get just the contents of this bluecurve-gtk-themes directory under the livecd branch?: http://git.fedorahosted.org/git/bluecurve-gtk-themes	00:25
arpegius	how can i ignore a directory's contents but not the directory itself?	00:25
canton7	arpegius git doesn't track directories	00:25
arpegius	. and * and foo/* all seem to include the directory foo	00:25
	← towski left	00:26
arpegius	ah	00:26
canton7	stick a .gitignore in the directory, though, and track the .gitignore	00:26
	← budsondixon left	00:26
canton7	pdtpatrick__ git show and git log are much of a muchness in this regard	00:27
	← urkud_ left	00:27
	→ PHadm joined	00:27
	← jfkw left	00:28
canton7	darknop git clone --branch livecd git://git.fedorahosted.org/livecd	00:29
	ChanServ set mode: +v	00:29
canton7	ignore that extra whitespace	00:29
	→ urkud_ joined	00:29
	← Murr left	00:29
canton7	darknop there doesn't appear to be a livecd branch though? so the command i gave you will fail	00:30
PHadm	is there a way to undo a git fetch?	00:30
canton7	PHadm why would you want to?	00:30
	→ budsondixon joined	00:30
darknop	git://git.fedorahosted.org/livecd	00:31
PHadm	is that a "no"	00:31
darknop	isn't thta the livecd branch?	00:31
	http://git.fedorahosted.org/git/?p=livecd;a=summary	00:31
canton7	darknop, looks like a repo to me. you can see the branches in the repo at the bottom, under "heads"	00:32
	← serverhorror left	00:32
canton7	darknop so git clone git://git.fedorahosted.org/livecd	00:32
offby1	PHadm: mostly "no".	00:32
	PHadm: why would you want to, anyway?	00:33
canton7	darknop scrap that, git clone git://git.fedorahosted.org/bluecurve-gtk-themes	00:33
	→ banisterfiend joined	00:33
PHadm	because the dev company we were using is supposed to be cleaning up a bit of a mess. If I fetch now and it introduces meyham, I want to step back and wait for them to finish housecleaning	00:34
canton7	offby1 can't find that reference, so i'm going to go back on what i said about add -A	00:34
	← budsondixon left	00:34
offby1	phhow could a fetch introduce mayhem? Are you confusing "fetch" with "pull"??	00:35
canton7	PHadm copy/paste the .git dir? :P	00:35
	PHadm at its worst, a git fetch will fetch tons of objects that might be completely useless, and just take up space. they'll get pruned eventually though, so no lasting damage	00:35
PHadm	I like that idea!!!!	00:36
canton7	PHadm sometimes the old-school solutions are the best :)	00:36
PHadm	so true canton7!	00:37
diegoviola	old-school++	00:37
	← xiangfu left	00:37
pdtpatrick__	is there a command to run to tell u where the .git/conf file is ?	00:37
canton7	one of the first things that convinced me that git >> svn was you could copy-paste your repo locally before a potentially screwy operations and you were completely safe	00:38
offby1	eww	00:38
	pdtpatrick__: git rev-parse --git-dir	00:38
	→ offby1 joined	00:38
	→ budsondixon joined	00:39
bremner	canton7: copy-paste?	00:39
	← offby1 left	00:39
	→ offby1 joined	00:39
canton7	yeah should have cloned it, i know. these were early days	00:39
offby1	he's translating it into language that Da Yoot can understand	00:39
bremner	so, like, you totally tweet .git	00:40
	← PHadm left	00:40
	← chessguy left	00:40
offby1	tewtally	00:41
canton7	heh, if you can stop me strangling twitter first	00:41
	← Pupeno left	00:42
canton7	one guy i helped on here decided to express his gratitude by tweeting about it. seriously. now my nick appears on twitter :(	00:42
	← mcluskydodallas left	00:43
	← budsondixon left	00:43
	← canton7 left	00:47
	← blackboard left	00:47
	→ budsondixon joined	00:47
	→ blackboard joined	00:47
	← blackboard left	00:48
	→ centipedefarmer_ joined	00:48
	→ fedesilva joined	00:50
offby1	aw	00:50
	← brian_g left	00:50
	→ Opettaja joined	00:51
	← budsondixon left	00:51
c00kiemon5ter	aaw	00:52
	← Yuuhi left	00:52
	→ saschagehlich joined	00:53
	→ sterlo joined	00:54
	← SimonP86 left	00:55
	→ cooloney joined	00:56
	→ budsondixon joined	00:56
	← squiggy left	00:57
	← shiba_yu36 left	00:58
	← blorbx left	00:59
	← budsondixon left	01:00
	← centipedefarmer_ left	01:01
	→ centipedefarmer_ joined	01:01
	← SomeGuy left	01:03
	→ Ownatik joined	01:04
	→ budsondixon joined	01:04
	← budsondixon left	01:05
	→ budsondixon joined	01:05
	← dnuke left	01:06
	→ farktronix joined	01:08
	→ milli joined	01:09
	← normanrichards left	01:09
	← budsondixon left	01:09
	← freelock left	01:10
	← pdtpatrick__ left	01:11
	→ t3i joined	01:12
	→ amerine_ joined	01:12
	← chrisdev left	01:13
	← ns5 left	01:13
	→ tearsong joined	01:13
	→ budsondixon joined	01:14
	→ xiangfu joined	01:16
	← RobertLaptop left	01:16
sitaram	that's not gratitude, that's gratitude 2.0 ;-)	01:16
	← tearsong left	01:18
	→ jfkw joined	01:19
	→ tearsong joined	01:19
	→ Rogi_ joined	01:19
	→ shiba_yu36 joined	01:20
	← johnw left	01:21
	→ towski joined	01:21
	← tjholowaychuk left	01:22
	← ColKurtz left	01:22
	→ chrisdev joined	01:23
	← Rogi_ left	01:23
	→ evorocks joined	01:23
	← Rogi left	01:23
	← towski left	01:24
	← berserkr left	01:26
	→ DrDogwelder joined	01:26
	← ksawicki left	01:26
	← DrkShadow left	01:27
	← chrisdev left	01:27
	← dan23232 left	01:27
	→ xenoterracide joined	01:27
	← tearsong left	01:28
xenoterracide	how would I cherry-pick from commit X to the tip of that branch	01:28
	← kpreid left	01:29
	→ SavoSuperCool joined	01:29
	→ tearsong joined	01:29
	← mr_mustard left	01:29
banisterfiend	xenoterracide: git cherry-pick <commit X>	01:30
	→ normanrichards joined	01:32
xenoterracide	yeah but I was pretty sure you can do a range of commits	01:32
	← t3i left	01:33
	← hyperair left	01:34
	← drev1 left	01:35
	← LekeFly left	01:35
banisterfiend	xenoterracide: if you want ot do a range of commits then use rebase --onto	01:36
	→ mcluskydodallas joined	01:36
	← spearce left	01:36
	→ dbgster joined	01:36
	→ qwp joined	01:36
qwp	i dont git git	01:37
	git it?	01:37
	→ gavin_huang joined	01:38
	→ ajt joined	01:40
	→ Vampire0_ joined	01:40
	← Vampire0 left	01:41
	← __name__ left	01:41
	← DoctorDarren left	01:41
	← dr_win left	01:43
	→ glaksmono_ joined	01:43
glaksmono_	guys, how do you create a branch from the master repo?	01:44
	→ bindaas joined	01:45
	→ dr_win joined	01:45
ries	glaksmono_: git branch someBranchName	01:45
	git checkout someBranchName	01:45
	← offby1 left	01:48
	← kst left	01:48
	← Dr4g left	01:49
	→ kst joined	01:49
	→ drev1 joined	01:51
	→ bonhoffer joined	01:52
	← alberto56 left	01:53
	← tystr left	01:54
	→ dnuke joined	01:54
	← glaksmono_ left	01:54
	→ webPragmatist joined	01:54
	← centipedefarmer_ left	01:55
	← webPragmatist left	01:55
	← tearsong left	01:57
	→ dwmw2__ joined	01:59
	→ offby1 joined	01:59
	← dwmw2_ left	01:59
	← boaz left	02:00
	→ boaz joined	02:01
	→ hyperair joined	02:01
	← mcluskydodallas left	02:02
	→ DarkHouse joined	02:03
	← german__ left	02:05
	← ap3mantus left	02:05
	→ jamison joined	02:05
	← kukks left	02:06
	→ ap3mantus joined	02:06
	→ neurogeek joined	02:08
	← darknop left	02:09
	→ erlnoob joined	02:10
	← boaz left	02:10
	→ f0i joined	02:10
	→ boaz joined	02:10
	← dbgster left	02:10
neurogeek	Hi all, question: I have a bare repo, served using apache+dav. Im trying to setup the post-receive email hook, but the hook is not being executed. Do I have to set some other variables in the config?	02:11
	besides the [hooks] section?	02:11
	→ boombatower joined	02:11
	← boombatower left	02:11
	→ boombatower joined	02:11
	← urbanmonk left	02:12
	← thomas_b left	02:13
cehteh	neurogeek: chmod +x it?	02:14
	→ _VVSiz_ joined	02:14
neurogeek	nah, the perms are good	02:14
	← boaz left	02:14
	→ minter joined	02:14
	→ boaz joined	02:15
	← erichammond left	02:15
	← ki__ left	02:15
	← banisterfiend left	02:16
	← kst left	02:16
	← bitkiller left	02:16
	→ zwick joined	02:16
	→ hobodave joined	02:16
	→ kst joined	02:16
	ChanServ set mode: +v	02:17
	→ thomas_b joined	02:17
	← pyrony left	02:17
minter	Hi folks - I have a corrupted git repository that I can't seem to recover from. Anything I try to do, I get a "failed to read object" error that I can't seem to work around.	02:17
	← VVSiz_ left	02:18
	← minter left	02:19
	→ minter joined	02:19
	→ ceej joined	02:19
bonhoffer	i am playing with branching for the first time -- i switched branches and played around with files, but when i go back to master, the files are still modified, when i do checkout i get a bunch of lines like: M test/test_helper.rb	02:19
	ChanServ set mode: +v	02:19
bonhoffer	i thought when i checked out the previous branch, my modifications wouldn't be there	02:19
minter	Sorry, xIRC was being weird. If anyone replied to the problem with "failed to read object" I missed it.	02:20
	minter:(git)teamsnap.broken[master]/$ git co user_files_paperclip fatal: failed to read object 4406bafc1128e184fe002b68fa347faf79ea795f: Invalid argument	02:20
	→ orospakr joined	02:20
	→ tshauck joined	02:21
minter	Any ideas on how to recover from that "failed to read object" error?	02:23
	→ slavyXS joined	02:23
	← saschagehlich left	02:24
	→ banisterfiend joined	02:27
	→ sebersole joined	02:28
	← SavoSuperCool left	02:29
	→ fakewaffle joined	02:29
	→ \yrlnry joined	02:29
	→ SavoSuperCool joined	02:29
	← mkanat left	02:30
	→ dfr\|mac joined	02:30
	← zwick left	02:31
	← sebersole left	02:32
	→ vivekp joined	02:33
	→ RobertLaptop joined	02:33
	← RobertLaptop left	02:36
	→ silouck joined	02:36
	→ RobertLaptop_ joined	02:36
silouck	i have done some cahnges in the local git repo, but i wold like to have all like its in master, how can i do?	02:36
	→ kpreid joined	02:39
	← minter left	02:40
	→ intripoon_ joined	02:40
	← neurogeek left	02:41
	← orospakr left	02:42
	→ orospakr joined	02:42
	→ pyrony joined	02:42
	← kpreid left	02:43
	→ kpreid joined	02:43
	← intripoon left	02:43
	← dreiss left	02:50
	← normanrichards left	02:51
	← Ownatik left	02:54
	← antw left	02:54
	→ jrock20041 joined	02:56
jrock20041	Is it possible to merge an existing svn repo into an existing git repo?	02:56
	← banisterfiend left	02:57
	← hobodave left	02:57
	→ banisterfiend joined	03:00
	← vivekp left	03:00
	← budsondixon left	03:01
	← cooloney left	03:01
	→ jwmann joined	03:02
	→ cooloney joined	03:03
	← orospakr left	03:04
	← offby1 left	03:04
	→ orospakr joined	03:04
	→ normanrichards joined	03:04
	← mgaunard left	03:05
	→ caseymcg1 joined	03:06
SethRobertson	jrock20041: permaently (never going back to svn?)	03:09
jrock20041	SethRobertson: I have one repo I would like to go back to but more likely never	03:10
	← DrDogwelder left	03:11
SethRobertson	I would say use git-svn to import into git, then you can use `git-subtree merge' or `adding the was-svn repo as a remote and merge or create local branches'. Going back...much more challenging	03:12
jrock20041	I got the ones moved over from svn that the git repo was empty but all that is left are ones where both are not	03:12
	→ sebersole joined	03:12
SethRobertson	I don't understand that	03:13
jrock20041	oh so I could do like I did for moving them to empty git, I just need to do the git-subtree merge	03:13
	→ gretch joined	03:13
SethRobertson	yes	03:13
	→ stliu joined	03:14
jrock20041	I do not have a git-subtree	03:14
	command	03:14
	let me google that for a mac	03:14
SethRobertson	It is a technique, not a command	03:15
	google is the right choice	03:15
	← banisterfiend left	03:15
jrock20041	oh	03:15
chrisf	technically, a merge strategy [-s subtree]	03:15
	man git merge	03:15
jast	automatic message the 'git-merge' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-merge.html	03:15
	→ dl9pf joined	03:16
	← dl9pf_ left	03:17
jrock20041	Hmm this might be harder than I think	03:18
	So I am in the directory that I cloned the files for my git repo	03:18
	I did the git svn command that brought in the information from the svn repository	03:19
	→ mipata_1 joined	03:19
jrock20041	Now when I brought the svn stuff in, I did not put it to a branch	03:19
	← sterlo left	03:19
	← oc80z left	03:20
	→ mgaunard joined	03:20
	→ oc80z joined	03:21
	← preflex left	03:21
	→ m0 joined	03:24
jrock20041	Should I have brought in my svn stuff into a branch?	03:24
	→ offby1 joined	03:24
SethRobertson	It should be on a local git branch, yes	03:24
	← offby1 left	03:25
jrock20041	ok so there is my mistake	03:25
	→ banisterfiend joined	03:25
	← veganadian left	03:25
	← pyroscope left	03:25
	→ Zygo joined	03:25
	→ preflex joined	03:26
	← banisterfiend left	03:26
	→ muratux joined	03:26
	→ versatiletech joined	03:26
	← caseymcg left	03:26
	→ sterlo joined	03:27
	← sterlo left	03:27
	← diegoviola left	03:30
	← Ryback_ left	03:31
	→ candybar joined	03:31
	← hyperair left	03:32
	→ budsondixon joined	03:32
jrock20041	So I would then run: git merge -s subtree branchname	03:33
	→ banisterfiend joined	03:34
	← qwp left	03:34
	← aef left	03:34
chrisf	jrock20041: most likely that is all you will need, yes.	03:34
jrock20041	Just tried and it did not work	03:35
	→ offby1 joined	03:35
	← mandric left	03:35
	← offby1 left	03:36
	→ offby1 joined	03:36
	← banisterfiend left	03:36
	→ dan23232 joined	03:36
	← bonhoffer left	03:37
	← pyrony left	03:37
	→ dreiss joined	03:37
jrock20041	Well I guess I will try again tomorrow. Thanks for all the help guys	03:37
	← jrock20041 left	03:37
	→ aef joined	03:38
	← ries left	03:38
	→ mandric joined	03:38
	← fcuk112 left	03:39
	→ banisterfiend joined	03:39
	→ Transformer joined	03:39
	→ herdingcat joined	03:40
	→ xerora joined	03:41
	← ceej left	03:42
	← Transformer left	03:43
	→ corra joined	03:45
	→ krz joined	03:45
	← caseymcg1 left	03:46
	→ bbrandon joined	03:47
	→ secretrobotron joined	03:49
secretrobotron	i seem to be having a bunch of trouble with submodules	03:49
	anyone care to help?	03:49
versatiletech	I'm trying to figure out a workflow for the following situation. I have a large group of HTML files for a website. At the current stage it just has filler text. I want to start working on the same set of files to insert content from word docs. At the same time I'd like to keep the first group of files separate. I don't know if a branch for this makes sense. Then I'd like to keep the files I'm adding content to separately. I may have fixes to the CSS o	03:50
	HTML that I might do on either group of files, and may add or remove files. I'd like both groups to be synchronized, so if I make a change to the CSS in the filler'd text group, I'd like to merge those changes to my HTML with content group. Any ideas?	03:50
	→ Merwin joined	03:51
	→ Corren joined	03:51
	← Corren left	03:54
	→ hyperair joined	03:54
	← RobertLaptop_ left	03:55
versatiletech	Or should I just have the two groups in two different folders and use a diff program to keep them in sync and just commit both folders to the repo?	03:55
SethRobertson	secretrobotron: My advice is that submodules suck and you will find it much easier to think about and use gitslave, though it is not always better for some workflows. But don't ask about asking, just ask	03:55
	→ RobertLaptop joined	03:55
	→ project2501b joined	03:56
secretrobotron	SethRobertson: i've had some recent trouble with them. we've been trying to add a submodule in place of a directory that was there by the same name	03:56
	now, conflicts when merging are... impossible to overcome	03:56
SethRobertson	versatiletech: If you will not be changing the filler text, branches and merging sounds good for you	03:56
secretrobotron	git simply says "error: unable to index file xxxxxxxx"	03:57
	when i try to add it	03:57
SethRobertson	secretrobotron: You have a path which is a submodule on one branch and a directory on another? Yes, that is going to cause problems.	03:57
	→ ExtraSpice joined	03:57
SethRobertson	I would suggest making it a submodule on the branch as well. It can be a different branch in the submodule	03:58
versatiletech	SethRobertson: well in the second group of files the filler text would be replaced with real content. Does that mean if I use the branching method, I'd get a crap load of conflicts each time I merge the filler text'd branch?	03:58
secretrobotron	SethRobertson: i didn't think of that	03:58
	make it a submodule on the branch first	03:59
	wlil try. hold on	03:59
	← bronson left	03:59
SethRobertson	versatiletech: Treat the fillter text as the branch and the real text as "master". Merge from fillter to master. As long as you do not change the fillter text, you will not have any conflicts	03:59
	secretrobotron: No guarantees of me still being awake.	03:59
	→ caseymcg joined	04:00
	← erlnoob left	04:01
secretrobotron	that's ok, trying now :) thanks for the idea	04:01
	→ Corren joined	04:02
	← Corren left	04:03
	← budsondixon left	04:03
	← OOPMan left	04:04
	→ cytrinox_ joined	04:05
	→ Anxi80 joined	04:05
	← slavyXS left	04:05
	← cytrinox left	04:06
	cytrinox_ → cytrinox	04:06
jmcantrell	what's the best way to export the latest revision to a directory?	04:08
	← troyt left	04:08
	← mipata_1 left	04:09
	← _grepper left	04:09
	→ troyt joined	04:09
	← project2501b left	04:09
	→ budsondixon joined	04:09
	→ diegoviola joined	04:10
secretrobotron	SethRobertson: thank you kindly :) seems to have worked.	04:10
	so... so strange.	04:10
versatiletech	SethRobertson: ok, so lets say I did a CSS bug fixes to the branch, and a couple of HTML structural changes. Master has the same files but with real content (different images, copy, etc). Let's say that on master I also make changes to the CSS file so now both the CSS file in both branches have diverged. How would I handle merging in the CSS changes in to master?	04:10
secretrobotron	jmcantrell: git clone, delete .git file?	04:10
\yrlnry	cp -r, delete .git file?	04:10
	→ Savoxis joined	04:10
	Savoxis → is	04:11
	→ _grepper joined	04:12
	← mandric left	04:12
	← tshauck left	04:12
	← cooloney left	04:13
	→ pen_ joined	04:14
	→ chx joined	04:14
	← pen_ left	04:14
quuxman	how do I email the commit message in the post-commit hook if the script receives no parameters?	04:14
chx	what would be the best way to revert an old commit? i mean i can do git show abcde \|patch -p1 but that aint elegant.	04:15
	*Rp1	04:15
quuxman	chx: wouldn't that be just `git checkout VERSION`	04:15
	← fedesilva left	04:15
chx	quuxman: i want to keep all commits just revert one.	04:15
quuxman	ah, yes, that would be nice	04:16
\yrlnry	if the one you want to get rid of is the most recent, just use git reset --hard HEAD~1.	04:16
	← secretrobotron left	04:16
\yrlnry	If you want to get rid of one that is not the most recent, git-rebase -i is probably the easiest way.	04:16
	← _grepper left	04:16
chx	uh	04:17
	→ versatiletech_ joined	04:17
chx	even if it's say 20-40-100 commits ago ?	04:17
	http://www.kernel.org/pub/software/scm/git/docs/git-revert.html isnt this what i want :) ?	04:17
	← dnuke left	04:17
	← eletuchy left	04:17
\yrlnry	Probably.	04:18
	→ pyrony joined	04:18
	← versatiletech left	04:19
	versatiletech_ → versatiletech	04:19
	← [intra]lanman left	04:19
chx	thanks :)	04:20
	← chx left	04:20
quuxman	anyway, seems natural to have the commit message easily available from the post-commit hook, but I can't find it in the docs, or a decent example post-commit	04:20
	→ ki__ joined	04:21
	→ OOPMan joined	04:23
	← juangiordana left	04:24
	→ Behi joined	04:27
	→ HaydenMuhl joined	04:27
	→ bolt joined	04:27
	→ leo2007 joined	04:28
selckin	think the first arg is a file with it?	04:28
	→ erichammond joined	04:28
	← caseymcg left	04:29
	→ sangi joined	04:31
	→ slavyXS joined	04:31
	← \yrlnry left	04:39
	← ki__ left	04:40
	← versatiletech left	04:42
quuxman	that's for pre-commit	04:42
	← sebersole left	04:42
quuxman	selckin: post-commit has no params	04:42
	→ e6nian joined	04:43
quuxman	oh duh... I could just use the pre-commit hook	04:43
	→ cooloney joined	04:43
e6nian	how to update a bare git repo?	04:43
	→ pyroscope joined	04:43
e6nian	git fetch only fetch the refs of the repo	04:44
	→ fedesilva joined	04:46
	← fakewaffle left	04:47
	→ dr0id joined	04:47
	→ mikepack joined	04:51
	← xiangfu left	04:52
	← dfr\|mac left	04:52
	← preflex left	04:53
	← amerine_ left	04:56
	← stringoO left	04:58
	← kst left	04:58
	→ akent joined	04:58
	← Opettaja left	04:58
	→ kst joined	04:58
	→ preflex joined	04:59
	→ dfr\|mac joined	05:01
	→ caseymcg joined	05:03
	← slavyXS left	05:04
	→ ThomasLocke joined	05:04
	→ Lenoj joined	05:06
	← leo2007 left	05:07
	ChanServ set mode: +v	05:07
	→ banghouse joined	05:07
e6nian	git reset --soft refs/remotes/origin/master not working	05:08
	why?	05:08
	fatal: ambiguous argument 'refs/remotes/origin/master': unknown revision or path not in the working tree.	05:08
	→ nevyn^_ joined	05:09
	→ dnuke joined	05:09
	← lunks left	05:10
	→ lunks joined	05:11
	→ cesario joined	05:15
	← ISF_ left	05:15
	→ gnufied joined	05:17
	→ BiggFREE joined	05:17
	← cherro left	05:17
	← madewokherd left	05:20
	→ jonshea joined	05:21
	→ airborn joined	05:21
	← e6nian left	05:21
	← dfr\|mac left	05:26
	→ dfr\|mac joined	05:27
	← boaz left	05:27
	← kst left	05:27
	→ kst joined	05:29
	→ subsume joined	05:30
	→ boaz joined	05:30
	→ mysphyt joined	05:30
	← mysphyt left	05:31
	← krosenvold left	05:33
	→ Chibby_ joined	05:33
	→ _W_ joined	05:34
	← jwmann left	05:35
	→ compnerd joined	05:37
	→ TrineLocke joined	05:40
	← troyt left	05:40
	→ neverov joined	05:40
	← neverov left	05:41
	← dan23232 left	05:41
	→ Zenopus_ joined	05:45
	← Zenopus left	05:46
	← compnerd left	05:46
	→ namemon joined	05:47
	← keyo left	05:47
	→ compnerd joined	05:47
	→ Zenopus joined	05:48
	→ slavyXS joined	05:49
	← Zenopus_ left	05:49
	→ yarvin joined	05:52
	→ workmad3 joined	05:53
	← compnerd left	05:55
	→ kukks joined	05:55
	→ compnerd joined	05:56
	← lunks left	05:58
	← compnerd left	05:59
	→ Deesl joined	06:00
	← dfr\|mac left	06:01
	← RobertLaptop left	06:02
	→ alnewkirk joined	06:02
	→ compnerd joined	06:03
	→ thiago_home joined	06:03
	← nevyn^_ left	06:03
	→ nevyn^_ joined	06:04
HaydenMuhl	Oh, man. I didn't know you could merge more than two branches together at once.	06:08
	→ KindOne joined	06:08
	← yarvin left	06:13
	→ Zenopus_ joined	06:14
	← budsondixon left	06:14
	← Zenopus left	06:15
	Zenopus_ → Zenopus	06:15
	→ ben225 joined	06:16
	← Maletor left	06:16
	→ gkmngrgn joined	06:16
	← gkmngrgn left	06:17
	→ tewecske joined	06:17
	→ gkmngrgn joined	06:17
	→ RobertLaptop joined	06:17
	→ budsondixon joined	06:19
	← nevyn^_ left	06:20
	← is left	06:21
	← SavoSuperCool left	06:22
	→ is joined	06:22
	→ warthog9 joined	06:23
	→ dvaskot joined	06:25
	← gkmngrgn left	06:25
	→ glaksmono joined	06:26
glaksmono	guys, how do you create a branch from the master repo?	06:26
	→ jutaro joined	06:26
	→ Maletor joined	06:27
	→ jceb joined	06:29
	→ slavy joined	06:29
	← slavy left	06:29
	→ JohnBat26 joined	06:30
thiago_home	glaksmono: git branch name-of-the-branch first-commit-of-the-branch	06:30
	← JohnBat26 left	06:30
	← Maletor left	06:32
	→ slavy joined	06:32
glaksmono	thiago_home: what's first-commit-of-the-branch? ?	06:32
	is that the master is?	06:33
KindOne	trying to "sh ./makexpi.sh" within cygwin, for "https-everywhere", I am using Windows with 1.7.4-preview20110204, and I am getting "./makexpi.sh: line 43: git: command not found" with this line 43 "if [ "$1" != "uncommitted" ] && [ -n "$(git status src -s)" ] ; then" any suggestions ?	06:33
	→ khmarbaise joined	06:33
	← Fudge left	06:34
	← slavy left	06:34
thiago_home	glaksmono: it is the SHA-1 of the first commit you want on that branch, or the name of the branch that is at that SHA-1	06:34
	if you want to create a branch that starts at the commit where origin/master is, just type that	06:35
	→ slavy joined	06:35
	← slavy left	06:35
	← idoru left	06:36
	→ Maletor joined	06:36
	← namemon left	06:36
	→ namemon joined	06:37
	→ erlnoob joined	06:37
	→ SavoSuperCool joined	06:38
	← dreiss left	06:38
masterkorp	hello, good morning	06:39
	→ leo2007 joined	06:40
	← xerora left	06:40
	← khmarbaise left	06:40
	→ bkozal joined	06:40
	← thiago_home left	06:40
glaksmono	thiago_home: so how do i find that SHA-1 thing? can u give example?	06:41
	can anyone help me? :(	06:42
	→ slavy joined	06:42
	→ slavy60xs joined	06:44
	← gavin_huang left	06:44
	→ xiangfu joined	06:45
	← slavy left	06:45
	→ troyt joined	06:46
	← metcalfc left	06:49
cbreak_work	glaksmono: ask a question	06:49
	→ Murr joined	06:50
	← mikepack left	06:50
	← jfkw left	06:50
	← Vile left	06:51
	← Maletor left	06:52
cbreak_work	glaksmono: just do a git log, and pick the commit that you want to be the newest in your new branch	06:52
	the new head	06:52
	← workmad3 left	06:54
	→ eean joined	06:54
	→ idoru joined	06:55
	← budsondixon left	06:59
	← Modius left	06:59
	→ yarvin joined	06:59
	→ buddybro_ joined	07:01
	→ dpino joined	07:02
	→ Maletor joined	07:02
	← troyt left	07:03
	← subsume left	07:04
	→ ksinkar joined	07:05
	→ Fudge joined	07:05
	→ radarek joined	07:05
	→ budsondixon joined	07:06
	← kukks left	07:07
	← xiangfu left	07:08
	← fedesilva left	07:08
	← SimonNa left	07:09
	→ Badiss joined	07:11
	← dvaskot left	07:11
	→ tvw joined	07:12
	→ airborn_ joined	07:13
	→ ben225_ joined	07:14
	← dr0id left	07:14
	→ rolfb joined	07:14
	← rolfb left	07:14
	← drev1 left	07:14
	← sangi left	07:15
	← ben225_ left	07:15
	← Lenoj left	07:15
	→ sangi joined	07:16
	→ slavy joined	07:16
	→ ben225_ joined	07:17
	← boombatower left	07:17
	→ kukks joined	07:20
	← slavy left	07:20
	← sangi left	07:22
	→ sangi joined	07:22
	← buddybro_ left	07:22
	→ boombatower joined	07:22
	→ slavy joined	07:23
	→ legumbre_ joined	07:23
	← mgaunard left	07:23
	→ Zenopus_ joined	07:24
	← legumbre left	07:25
	← slavy left	07:25
	← Zenopus left	07:25
	→ d0k joined	07:25
	← Zenopus_ left	07:28
	← boombatower left	07:29
	← dnuke left	07:29
	← corra left	07:29
	→ Zenopus joined	07:30
	← Tommy[D] left	07:30
	→ boombatower joined	07:31
	→ Tommy[D] joined	07:31
	→ mishok13 joined	07:34
	→ eijk joined	07:35
	→ frederick joined	07:36
	← ben225_ left	07:37
	→ slavy joined	07:37
	→ ben225_2 joined	07:37
	← ben225_2 left	07:37
	← ben225 left	07:38
	← slavy left	07:38
	→ blast_hardcheese joined	07:38
	→ mgaunard joined	07:40
	→ jmd joined	07:42
	→ slavy joined	07:42
	← slavy left	07:43
	← Fudge left	07:43
	→ slavy joined	07:44
	← slavy left	07:44
	→ cilly joined	07:44
	← cilly left	07:44
	→ cilly joined	07:44
	→ ruby_on_tails joined	07:45
ruby_on_tails	how can i check the repo url ?	07:45
	git remote shows origin, i want the repo url	07:45
	← namemon left	07:45
jmd	git remote -v	07:45
	→ slavy joined	07:46
	← slavy left	07:46
	→ pcav joined	07:46
	→ saschagehlich joined	07:46
	→ aminpy joined	07:47
	→ slavy joined	07:47
ruby_on_tails	jmd: i want to change it i should use remote rename ?	07:47
	← slavy left	07:47
	→ Ricki joined	07:48
	→ ricki_ joined	07:48
	← ricki_ left	07:48
jmd	I normally delete it and re-add it.	07:48
ruby_on_tails	ok	07:48
	→ slavy joined	07:48
	← slavy left	07:49
	→ Jippi joined	07:50
	← Murr left	07:52
jmd	I want to copy (not merge) an entire file from one branch to another. How do I do that?	07:53
FauxFaux	git checkout branch path/to/file # or git show branch:path/to/file > new/path/to/file	07:54
	→ nfr joined	07:54
	→ capncodewash joined	07:55
	← Jippi left	07:55
	→ Fudge joined	07:55
	→ strk joined	07:56
	← nfr left	07:56
	→ delx joined	07:56
delx	Hello	07:56
	I have creatd a git repository on my server (git init --bare) and I clone it through SSH. It says "It seems you have cloned an empty repository." So far ok. But "git branch" doesn't even show a master branch and I can push at all to my server! What's wrong here?	07:56
strk	in a git-svn setup (started from just looking at trunk/) can I add a branch tracking a different SVN branch ?	07:56
jmd	FauxFaux, Thanks	07:56
FauxFaux	delx: Normal procedure is: git init locally; commit; git remote add origin user@host:path/to/bare; git push -u origin master	07:57
cbreak_work	delx: empty repositories do not have a branch	07:57
	← NetRoY left	07:57
cbreak_work	delx: just commit something	07:57
	(in your clone)	07:57
	after that, git push as FauxFaux said	07:58
	you don't need to add a new remote if you cloned	07:58
	→ namemon joined	07:58
	→ slavy joined	07:58
	→ \|kukks\| joined	07:59
delx	FauxFaux, cbreak, now the fololowing:	07:59
	error: src refspec master does not match any.	07:59
	error: failed to push some refs to 'ssh://git@host/home/git/matrix.git/'	07:59
	I tried ssh git@host and that works	08:00
	→ Badababuba joined	08:00
delx	and I have write-privileges in /home/git/matrix.git/ !	08:00
_VVSiz_	guys, is there a way to enter comments in .config file?	08:00
FauxFaux	delx: You completely ignored what I said and tried to push from your cloned repository, didn't you?	08:00
	→ kristofferR joined	08:00
delx	FauxFaux, no! I did an init - committed something, added the remote origin and pushed it!	08:01
cbreak_work	delx: you should read what FauxFaux wrote	08:01
	how did you push?	08:01
	→ SimonNa joined	08:01
cbreak_work	how did he tell you to push?	08:01
	do you spot the difference?	08:01
delx	git push -u origin master	08:01
cbreak_work	yep. that's the one.	08:01
FauxFaux	What does 'git status' say the current branch is?	08:01
delx	I'm not stupid :) I did it the way!	08:01
	git status -> "On branch master"	08:02
	← kukks left	08:02
delx	git branch -a -> Nothing	08:02
FauxFaux	Okay, I'll admit that you probably have an interesting problem, but I have no idea what the solution is. ¬_¬	08:02
cbreak_work	did you do something stupid like git init in an existing repo?	08:02
jmd	cbreak, Of course he didn't. He just said that he ISNT stupid.	08:03
cbreak_work	I am a genius and I still do stupid things.	08:03
	it's just that I am usually able to fix them :)	08:04
	_VVSiz_ → VVSiz	08:04
	← SimonNa left	08:04
	→ dwmw2___ joined	08:05
jmd	That's ironic, because I'm stupid, but I do ingenious things.	08:05
	← dwmw2__ left	08:05
	→ davzie joined	08:06
	→ dr0id joined	08:06
	← f0i left	08:06
	← delx left	08:07
	← cilly left	08:08
	← Spockz left	08:09
	→ PigeonFriend joined	08:09
	← boombatower left	08:09
	← OOPMan left	08:11
cbreak_work	anyway	08:12
	just clone, commit, push	08:12
	that should work	08:12
	← slavy left	08:12
	← Maletor left	08:12
	→ delx joined	08:14
	→ SimonNa joined	08:14
	→ jas4711 joined	08:14
	← delx left	08:15
	→ sebbi joined	08:15
	→ jatt joined	08:17
	← NetHawk left	08:18
	→ Leonidas_ joined	08:19
	← Leonidas_ left	08:19
	→ Leonidas_ joined	08:19
	Leonidas_ → Leonidas	08:19
	→ NoNoNo joined	08:19
	→ slavy joined	08:20
	→ OOPMan joined	08:20
	← KindOne left	08:22
	← erichammond left	08:22
	→ KindTwo joined	08:22
	→ gebi joined	08:24
	KindTwo → KindOne	08:24
	← slavy left	08:25
	→ KindTwo joined	08:25
	← KindTwo left	08:25
	→ KindTwo joined	08:25
	← diegoviola left	08:26
	← KindOne left	08:27
	KindTwo → KindOne	08:27
	→ _Qman joined	08:31
	→ gebbione joined	08:31
	→ ch077179 joined	08:34
	← verto_ left	08:36
	→ krosenvold joined	08:38
	← muratux left	08:38
	→ slavy joined	08:38
	→ jjido joined	08:39
	→ Jippi joined	08:40
	← slavy left	08:42
	→ d00dlez joined	08:42
	→ rendar joined	08:42
	→ tatsuya joined	08:43
	→ Yuuhi joined	08:44
	→ opalka joined	08:47
	← opalka left	08:47
	→ opalka joined	08:47
	→ verto_ joined	08:49
	→ elyob joined	08:52
	← gebbione left	08:52
	← envi_laptop left	08:52
	→ g0bl1n joined	08:52
	→ pizza_ joined	08:53
	← HaydenMuhl left	08:58
	→ Agos joined	09:03
	→ curtana joined	09:06
	→ deuterium joined	09:11
	← lanthan left	09:11
	→ stermi joined	09:11
stermi	hi all, I'm new to git and I need some explanation	09:12
	← strk left	09:12
stermi	what does gitolite do? It's an amministrative tool for git? It check for permissions on clone/push?	09:12
	← Jippi left	09:12
	→ Jippi joined	09:13
Rhonda	It is an ssh gate keeper. It helps you administer ssh keys and push permission	09:13
dogmatic69	stermi: it sort of allows you to set up a git server similar to github	09:13
	← namemon left	09:13
stermi	have I to put all my repositories to /home/git/repositories?	09:13
FauxFaux	Yes.	09:13
Rhonda	You can configure the path, but besides that, yes.	09:14
	Actually you configure the path at installation time of gitolite and stop caring about it from there, you don't have to "put" anything in there yourself.	09:14
FauxFaux	Should get jast to reply to "!gitolite" with "Gitolite is a tool for teaching users how to use and debug problems with ssh keys".	09:14
	→ harshu_ joined	09:14
stermi	ok, so I need to clone gitolite-admin for the new project, modify it, and push on server?	09:14
Rhonda	Exactly.	09:15
stermi	so after it, if someone form another pc try to do git clone [email@hidden.address] it will be denied?	09:15
	→ hendersonj joined	09:16
FauxFaux	Unless you grant them access, yes.	09:16
stermi	sorry for stupid question but I'm trying to understand how to do that :D	09:16
	→ jutaro1 joined	09:16
	← jutaro1 left	09:17
Rhonda	No question is stupid (enough to be laughed at by someone even more stupid)	09:17
sitaram	stermi: you do not need gitolite until you come to the point of managing multiple users and their access to multiple repos	09:17
FauxFaux	gitolite-admin's config has a list of users that are allowed access.	09:17
stermi	yes FauxFaux I already installed but I want to be sure about what I tought :D	09:18
	sitaram, I'm trying to build a good env to develop with my collegues	09:18
sitaram	ok	09:18
stermi	and I've bought a VPS ( for also other pourpose ) so I want to restrict access to it :P	09:19
sitaram	stermi: I suggest http://sites.google.com/site/senawario/home/gitolite-tutorial (this is a good tut that shows gitolite install using the 'non-root' method)	09:19
stermi	thank you sitaram this let me understand something more about new users and permissions	09:20
	but I don't get something. How can I create new repos?	09:21
	with normal git commands?	09:21
Rhonda	No	09:21
	You add them to the gitolite-admin's config file.	09:21
	and push	09:21
	<stermi> ok, so I need to clone gitolite-admin for the new project, modify it, and push on server?	09:21
stermi	and who will do changes?	09:21
Rhonda	You asked that before. :)	09:21
	← jmd left	09:22
stermi	after the push gitosis will automatically create new repos?	09:24
	→ erichammond joined	09:24
sitaram	I would be very surprised if it did	09:25
	:-)	09:25
selckin	gitoLite does	09:25
stermi	Ok so I don't understand how it work sitaram (thank you for your great tool!)	09:25
sitaram	stermi: erm that was a joke; you said "gitosis will ..." while talking about gitolite	09:25
	← intripoon_ left	09:25
stermi	aajjaja	09:26
	hahahahaha sorry :D	09:26
	→ redGoat joined	09:26
stermi	tomorrow I studied both :P	09:26
	→ intripoon joined	09:26
	→ dragoonis joined	09:28
Rhonda	Did you travel back in time to today?	09:28
	→ Leemp2 joined	09:29
stermi	damn I need coffe	09:29
	← dec left	09:30
	→ dec joined	09:30
	→ shiba_yu_ joined	09:32
	← shiba_yu36 left	09:32
	→ dbgster joined	09:33
stermi	ok last stupid question of the day: if I remove from my gitolite-admin users and repositories, they will be removed and deleted from my server ?	09:33
	repositories will be deleted ?	09:33
	← jjido left	09:33
stermi	I mean I don't have to touch my server anymore right? I can do everything editing my gitolite-admin	09:33
	and pushing changes	09:34
	← Leemp2 left	09:34
	← glaksmono left	09:34
	→ brian_g joined	09:35
sitaram	stermi: search for "deleting a repo" in http://sitaramc.github.com/gitolite (this is the gitolite doc main page/master index; you can browse interesting entries there or search for stuff using Ctrl-F)	09:35
	→ babilen joined	09:35
	→ antw joined	09:36
	← \|kukks\| left	09:39
stermi	I have last question (I don't find it on help page). Is there some problem using gitosis with other tools like tortoisegit for windows? My collegue use it (and windows <.<)	09:39
	→ kukks joined	09:39
	→ jjido joined	09:39
dogmatic69	stermi: i would recommend using git the terminal, and not with silly gui's	09:41
albel727	why would it be a problem? (and it's _gitolite_. don't insult it by calling it gitosis).	09:41
	← mishok13 left	09:41
	← NoNoNo left	09:41
	→ MonsieurLu joined	09:41
stermi	damn me another time :D	09:41
	anyway yes, I was talking about gitolite	09:42
albel727	well, the trickest part might be with setting up ssh keys for tortoisegit properly.	09:42
stermi	dogmatic69, to push changes I will use my own pc with linux and I'll do by command line git	09:42
albel727	b/c tortoisegit uses plink, iirc.	09:42
	→ mishok13 joined	09:42
sitaram	stermi: as far as any client is concerned it just uses ssh pubkey based access. It does not even know that gitolite is on the server	09:42
stermi	I was talking about normal git push/clone	09:42
albel727	so, you'll have to convert openssh public keys to plink public keys (or the other way round)	09:42
stermi	thatnk you sitaram anyway I will do some test totday	09:42
	→ NoNoNo joined	09:43
sitaram	stermi: but as albel727 says, putty/plink's keyformats may have some issues. Look in that same master index for "putty"	09:43
	← d0k left	09:44
stermi	ok	09:45
	← ch077179 left	09:45
	← davzie left	09:46
	→ ch077179 joined	09:47
	← jjido left	09:48
	← NoNoNo left	09:50
	← radarek left	09:50
	→ LekeFly joined	09:51
	← Zenopus left	09:52
	← dbgster left	09:53
	→ Zenopus joined	09:53
	→ radarek joined	09:56
	→ intripoon_ joined	09:58
	← hyperair left	09:58
	→ freshmilk joined	09:59
	← krz left	09:59
	← intripoon left	10:00
speakman	When having an internal fork of an upstream project, and keep rebasing upon upstream changes - will internal "release" tags pointing to commits not available under any head still be reachable?	10:00
FauxFaux	Tags are immutable; they'll keep the unrebased bits of the history alive. gitk --all.	10:01
	← pipoun left	10:01
	← mishok13 left	10:02
	→ mishok13 joined	10:02
speakman	FauxFaux: Not sure I really understand	10:03
	← leo2007 left	10:04
bremner	!cs	10:04
jast	automatic message 'Git for Computer Scientists' explains the basic structures used by git, which is very helpful for understanding its concepts. http://eagain.net/articles/git-for-computer-scientists/	10:04
bremner	speakman: ^	10:04
speakman	But since my previous release was my "patchset" upon an older upstream release, and my next release will be my patchset upon a newer upstream release.	10:04
	bremner: thanks!	10:04
	← VVSiz left	10:07
	← aminpy left	10:07
	→ stamina joined	10:07
	→ Zenopus_ joined	10:08
	← Zenopus left	10:08
	Zenopus_ → Zenopus	10:08
speakman	bremner: FauxFaux: Something like this: http://pastebin.com/RRYf9ZiS	10:08
FauxFaux	Yes, that's exactly what happens.	10:08
speakman	"git gc" won't remove any objects required by myfirstrelease when I've git rebased later on to make mysecondrelease?	10:09
	(I can see git gc removing objects when there is no references at all to it though)	10:09
	Which won't be the case as long as my release tags points to it,right?	10:10
	→ aminpy joined	10:10
FauxFaux	Yeah, a tag makes something alive/reachable, as in, not applicable for garbage collection.	10:10
speakman	great thanks!	10:10
	→ ralc joined	10:11
speakman	so - is this approach (all in all) an acceptable way of maintaining local changes to an upstream project?	10:11
	→ Leemp joined	10:12
FauxFaux	It's what I do, yes. Some people like quilt instead of git but they're heretics.	10:13
	← babilen left	10:13
speakman	:D	10:15
	← jophish_ left	10:17
	→ k-man joined	10:18
	← urkud_ left	10:18
	→ mklappstuhl joined	10:20
	→ chittoor joined	10:22
	→ davzie joined	10:22
	← chittoor left	10:22
	← arpegius left	10:23
	→ tore__ joined	10:24
	→ slavy joined	10:25
	← SimonNa left	10:25
	← slavy left	10:25
ichilton	Was speaking to someone from Github at a conference yesterday and asking him about some problems i'm having with git submodules. He advised me to checkout 'lightweight submodules' and described them as submodules done right. However, i'm not finding much with searching for it - any ideas?	10:27
	→ fauxhawk joined	10:27
	→ michas joined	10:27
ichilton	there seems to be something called lightweight checkouts, but there doesn't seem to be much info around them either...	10:27
bremner	ichilton: the only alternative I know about is gitslave, but I haven't heard it called lightweight	10:28
	→ marcob joined	10:28
ichilton	humm ok - thanks	10:28
	→ aspotashev joined	10:29
ichilton	i'm using git submodules to bring the php framework & modules and my core app module into my application repo - but they dont really play nicely with branching...	10:29
	→ babilen joined	10:31
	→ verto__ joined	10:31
	← intripoon_ left	10:32
	← verto_ left	10:33
	→ intripoon joined	10:33
	→ Spockz_ joined	10:34
djh	I'm having an issue with git taking a very long time to do a fetch when it's run via sudo. Only error I can see if I strace it is "Warning: No xauth data; using fake authentication data for X11 forwarding." - anyone know if there's a fix for this?	10:34
	← ch077179 left	10:35
	← Anxi80 left	10:35
	→ Anxi80 joined	10:35
	← dr0id left	10:35
bremner	well, somebody's got to ask. Why sudo?	10:35
FauxFaux	djh: Unset DISPLAY.	10:35
djh	bremner: Well, sudo -u user to be more accurate	10:35
Fudge	hi im getting this message, how do i resolve it fatal: You have not concluded your merge (MERGE_HEAD exists).	10:36
FauxFaux	Conclude your merge.	10:36
Fudge	how do i do that FauxFaux	10:36
FauxFaux	Commit.	10:36
	→ dr0id joined	10:37
michas	Hi, I'm trying to build git on HP-UX. make now complains, it cannot find zlib.h. If I point --with-zlib to the directory containing the correct include directory it does not change anything. what might be my problem?	10:38
djh	FauxFaux: That does indeed work, ty.	10:38
FauxFaux	config.log gives more details on why configure failed.	10:39
	← caseymcg left	10:41
Fudge	thanks FauxFaux	10:42
	← brian_g left	10:42
michas	configure is not complaining. but I had a typo in the path. now make complains about git-compat-utils.h line239: error 1000: Unexpected symbol ")"...	10:42
	→ brian_g joined	10:42
	→ namemon joined	10:43
FauxFaux	djh: Another solution is probably fiddling with xauth or ~/.Xauthority or sudo settings wrt. environment variables, either directly or for ssh-agent etc., but it's all far too much work.	10:43
michas	does git need a special compiler like gcc?	10:43
djh	FauxFaux: It does sound that way	10:43
FauxFaux	It's built with gcc on all platforms, I'm not sure if there's an official statement that patches won't be accepted for other compilers.	10:43
	← redGoat left	10:45
	← stermi left	10:46
speakman	FauxFaux: what would you think of this approach, to keep the number of internal commits to be rebased as low as possible; http://pastebin.com/9WMsaizq	10:46
	→ stermi joined	10:46
FauxFaux	speakman: Editing commtis as you move them around leads to unhappy, see git clone git://git.goeswhere.com/git/putty.git && cd putty && gitk --all and try and work out what's happened between the rebases.	10:48
	→ cjs joined	10:49
	← tore__ left	10:50
	→ muratux joined	10:51
	→ canton7 joined	10:51
	← muratux left	10:52
	→ muratux joined	10:52
	→ jophish_ joined	10:52
	← stliu left	10:53
	← muratux left	10:53
	→ muratux joined	10:54
	← muratux left	10:54
	→ muratux joined	10:55
	→ linusoleander joined	10:55
dogmatic69	in a pre-commit hook, how can i tell if there is just files being deleted?	10:57
	← cjs left	10:58
	← linusoleander left	10:59
	→ leo2007 joined	10:59
	← dr0id left	10:59
cbreak_work	how about looking at the diff?	11:00
	← namemon left	11:01
	→ namemon joined	11:01
	→ dr0id joined	11:02
dogmatic69	cbreak_work: how would i get that in the hook?	11:02
	→ Ownatik joined	11:02
	← jophish_ left	11:04
cbreak_work	dogmatic69: by running git diff? :)	11:05
	→ johnson09 joined	11:05
	← dl9pf left	11:05
dogmatic69	cbreak_work: that gives stuff in 'Changed but not updated:' not 'Changes to be committed:'	11:05
	→ Sonderblade joined	11:06
	← Spockz_ left	11:06
	← Sonderblade left	11:06
	→ Sonderblade joined	11:07
	← antw left	11:09
	→ vmiklos joined	11:09
	→ antw joined	11:09
	← lanthan_afh left	11:10
	→ ramusara joined	11:11
	← bkozal left	11:13
	→ dl9pf joined	11:14
	← dl9pf left	11:14
	→ dl9pf joined	11:14
	→ jophish_ joined	11:14
	← apassi left	11:19
	→ apassi joined	11:21
	← harshu_ left	11:21
	→ urkud_ joined	11:22
	← banghouse left	11:23
	← yarvin left	11:25
	← apassi left	11:27
	→ apassi joined	11:29
	→ Kontext joined	11:31
Kontext	hello ppl	11:31
	→ SimonP86 joined	11:31
	← namemon left	11:31
Kontext	I've made some changes to my local project (git repository)	11:31
	if i type git status I get this	11:31
	new file: administrator/compone/ mplampla	11:31
	Changes not staged for commit:	11:31
	modified: administrato/ mpla mpla	11:32
	→ namemon joined	11:32
Kontext	then I did a git commit -m "test"	11:32
	and then I try to update my remote repository (github)	11:32
	by typing git push grocery master	11:32
	and I get this Everything up-to-date	11:32
FauxFaux	Enough context, more point.	11:32
	← aminpy left	11:33
Kontext	the point is that everything is not up to date	11:33
	isn't this the right command to update my remote repo	11:34
	with the changes I've made to my local repo;	11:34
	→ Jordano joined	11:34
Jordano	hi	11:34
Bombe	Is “grocery” your github remote, Kontext?	11:34
Kontext	yes	11:34
Bombe	And has that commit you did gone to the master branch?	11:35
Jordano	I'm cloning with git the android kernel but it's stucked on Counting objects, can somebody help me to find what's wrong ?	11:36
FauxFaux	The number has stopped going up, or it's spending ages going up/	11:36
Kontext	I did a git commit -m	11:36
	so I believe so	11:36
Bombe	Kontext, what branch does git status say you’re on?	11:37
wereHamster	Kontext: what does 'git branch' show you?	11:37
Jordano	yesterday night was working, but I had to delete everything.	11:37
	and now it does not download anything	11:37
Kontext	* master	11:37
	* On branch master	11:37
	here's me repo: https://github.com/Agis-	11:37
Bombe	Kontext, ah, whatever, take a hammer: gitk --all	11:37
	That should make clear what's happening.	11:38
	← bbrandon left	11:38
Kontext	i'm not very familiar with git, I get the scren now	11:40
	← herdingcat left	11:40
	← p3rror left	11:41
	← SimonP86 left	11:41
Kontext	what should I check for?	11:41
	← gretch left	11:42
Kontext	Local uncommited changes, not checked in to index	11:42
	master-remotes/grocery/master test	11:42
	→ SimonP86 joined	11:44
	→ xiangfu joined	11:45
	→ Spockz joined	11:46
	← LekeFly left	11:46
	← muratux left	11:47
	→ cannonball joined	11:49
	→ xiangfu` joined	11:49
	→ yarvin joined	11:49
	← xiangfu left	11:51
	→ dbgster joined	11:52
	← yarvin left	11:54
	→ p3rror joined	11:55
	→ rostayob joined	11:55
rostayob	how can I remove the last commit, preserving the files as they are?	11:56
	→ gusnan joined	11:56
selckin	git reset HEAD^, if i understand you correctly	11:56
rostayob	selckin: so reset won't touch the files, it would just remove the commig?	11:57
	*commit	11:57
selckin	define "the files"	11:58
	→ harshu_ joined	11:58
selckin	it will not remove those changes from the workdir	11:58
banisterfiend	selckin: it will	11:59
	selckin: git stash	11:59
	git stash && git reset HEAD^	12:00
rostayob	selckin: I've used reset --hard before, and it just reverts everything to the state you were in	12:00
selckin	well, did i tell you to use --hard?	12:00
	→ JohnBat26 joined	12:00
	← urkud_ left	12:00
	→ jutaro1 joined	12:00
rostayob	selckin: no, I just want to be sure that I won't lose data :P	12:00
selckin	then make a backup first	12:01
	→ gavin_huang joined	12:01
banisterfiend	ah sorry selckin was right	12:01
rostayob	selckin: yeah I'd like a command that doesn't require me to do that, I'll try reset now	12:01
selckin	rostayob: if you're unsure, make a backup, should only take a few seconds, good way to learn & try out new things risk free	12:01
	→ bkozal joined	12:02
rostayob	selckin: yeah reset works, thanks	12:02
	→ urkud_ joined	12:02
	← shiba_yu_ left	12:02
	← sangi left	12:03
	→ NetHawk joined	12:05
	→ alberto56 joined	12:09
	→ versatiletech joined	12:09
	← Kontext left	12:10
	→ rolfb joined	12:12
	← cyphase left	12:13
	← elyob left	12:13
	→ SimonP86_ joined	12:15
	→ jsick3 joined	12:15
	→ banghouse joined	12:15
	→ piti joined	12:17
	→ SimonP86__ joined	12:17
	→ eikaas_ joined	12:17
	← SimonP86 left	12:18
	← eikaas left	12:19
	← jutaro1 left	12:20
	← deuterium left	12:20
	← SimonP86_ left	12:20
	→ cherro joined	12:21
	→ nazgul101 joined	12:22
	← gusnan left	12:24
	← Ownatik left	12:26
	← _Qman left	12:27
	← mgaunard left	12:31
	→ lunks joined	12:31
	← normanrichards left	12:31
	← namemon left	12:32
	→ babbos joined	12:32
	→ namemon joined	12:32
	→ Jdub71 joined	12:33
babbos	If i want to execute a post-update hook only if a specific branch is pushed what do i do?	12:33
	← aspotashev left	12:33
	→ aspotashev joined	12:34
	→ jutaro1 joined	12:35
	← jutaro1 left	12:35
	→ __name__ joined	12:36
	← __name__ left	12:36
	→ __name__ joined	12:36
	← ap3mantus left	12:36
	→ akerl joined	12:37
	→ Granjow joined	12:37
	→ Ownatik joined	12:37
Granjow	Good morning :)	12:37
	→ ap3mantus joined	12:37
Granjow	I have a project in my git repository which I would like to put into a separate repository now. Is this possible with preserving the directory structure?	12:37
patrikf	babbos: man githooks, post-update, second paragraph.	12:38
jast	babbos: automatic message the 'githooks' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/githooks.html	12:38
	→ replore_ joined	12:38
Granjow	eh, with preserving the commit history	12:38
	→ Fly9 joined	12:38
selckin	Granjow: google "git split repo"	12:38
	→ Dave^\| joined	12:38
	→ sonnym joined	12:38
	→ centipedefarmer_ joined	12:39
Granjow	selckin: whew, thank you :)	12:39
	← alberto56 left	12:41
babbos	patrikf some more help please? maybe an example? I read the paragraph many times	12:41
	← Spockz left	12:42
	→ PabloM joined	12:43
	← akerl left	12:44
	→ mgaunard joined	12:45
	→ fedesilva joined	12:46
	→ alberto56 joined	12:46
	→ RaceCondition joined	12:46
	→ nd` joined	12:46
patrikf	babbos: your hook will receive the changed refs as arguments.	12:47
babbos	so i'll work with $1, $2... testing that right now...	12:48
	→ s0undt3ch joined	12:49
	← nd` left	12:49
	← Fly9 left	12:50
	→ achadwick joined	12:50
	→ akerl joined	12:50
	→ LiamH joined	12:51
	→ vdv joined	12:52
	← ap3mantus left	12:52
Granjow	selckin: How much of the history does git change there? My repository is still 100 MB (quite a lot for text files only and 60 commits); 99 MB are in the .git subdirectory.	12:53
	s/change/keep/	12:53
	→ mika joined	12:53
	← dbgster left	12:53
Granjow	Seems to me that there is still lots of old stuff?	12:53
selckin	Granjow: it will keep everything for x day	12:54
	Granjow: man git-gc probably has some to remove stuff	12:54
jast	Granjow: automatic message the 'git-gc' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-gc.html	12:54
	← mklappstuhl left	12:54
	→ bitkiller joined	12:55
Granjow	git gc --prune="0 days" and my repository is still 100 MB.	12:56
selckin	needs more then that	12:56
	← ramusara left	12:57
FauxFaux	Granjow: Bottom of man git filter-branch	12:57
jast	Granjow: automatic message the 'git-filter-branch' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-filter-branch.html	12:57
selckin	git reflog expire=now --all; git gc --prune=now, some toher stuff	12:57
	→ hyperair joined	12:57
	→ shiba_yu36 joined	12:57
	→ DoctorDarren joined	12:58
Granjow	selckin: This shrinked my repository to a size of 100 MB	12:59
	FauxFaux: One moment, I'll try the other stuff there	12:59
selckin	are you running it on the original?	13:00
Granjow	selckin: No, on the clone of the clone of the splitted clone	13:01
	→ dwmw2____ joined	13:01
	← dwmw2___ left	13:02
Granjow	Does this mean something special? http://paste.debian.net/119320/	13:03
banisterfiend	selckin: your name reminds me of the seljuk sultanate of rum	13:03
	← harshu_ left	13:03
	→ envi_laptop joined	13:03
banisterfiend	selckin: http://en.wikipedia.org/wiki/Seljuk	13:04
	→ AaronMT joined	13:04
	→ Murr joined	13:04
selckin	banisterfiend: read it backwards	13:05
	→ icwiener joined	13:05
banisterfiend	nikcles	13:05
	→ drubies joined	13:06
	→ juvenal joined	13:06
	→ german__ joined	13:06
	→ Danielpk joined	13:06
	→ nlisgo joined	13:08
	→ bartek joined	13:09
Granjow	Tried to delete the refs/remotes/origin/master now with git update-ref, but it seems that this was not a good idea.	13:09
selckin	git remote del origin	13:10
	→ jutaro1 joined	13:10
	← jutaro1 left	13:12
	→ daniel-s joined	13:13
daniel-s	hi	13:13
	if I say "git push origin master"	13:13
	I'm updating the master branch	13:13
	on origin	13:13
	with my most recent commit of my current HEAD?	13:13
ndim	you update the remote master with your local master	13:14
	"git push origin master" is short for "git push origin master:master"	13:14
Granjow	selckin: Whatever I try, I cannot get it any smaller. Also removed the origin, size did not change.	13:15
selckin	can you prove that it should be smaller?	13:15
	← RobertLaptop left	13:16
	← Deesl left	13:17
Granjow	selckin: No. But should it really be that big for 60 commits on C++ files?	13:18
	← versatiletech left	13:18
selckin	pritty sure git is right, and you commited something big at one point	13:18
	← s0undt3ch left	13:18
Granjow	selckin: can I check this?	13:19
	→ versatiletech_ joined	13:19
wereHamster	yes you can	13:19
Granjow	wereHamster: how can I check this?	13:19
	;)	13:19
wereHamster	I can say that without even knowing your question, because git can anything!	13:20
	(no, I did not read your question, I read the backlog from bottom up)	13:20
Granjow	I really changed text files only according to gitk.	13:20
	→ ap3mantus joined	13:21
wereHamster	Granjow: make a git clone file:///path/to/the/repo	13:21
	is that clone big, too?	13:21
Granjow	wereHamster: The exact same size.	13:21
	100 MB as well.	13:21
selckin	did you try git repach -ad, after the other too?	13:21
wereHamster	Granjow: there's a git find-large script on the wiki	13:21
	→ ramusara joined	13:21
wereHamster	run that in your repo, it'll show the largest objects	13:21
Granjow	wereHamster: Or wait, does it need to be file:///, or can it also be like /tmp/somewhere?	13:22
	← centipedefarmer_ left	13:22
Granjow	selckin: Yes, several times	13:22
selckin	then run the script	13:23
Granjow	file:/// takes a little bit longer until I get the 100 MB repository	13:23
	ok	13:23
	→ cyphase joined	13:24
Granjow	ah. It for example points to a 18 MB .xcf file that is definitely outside of the directory I split. As all other files are as well.	13:25
	→ zwick joined	13:25
	→ centipedefarmer_ joined	13:25
	→ JakeyChan joined	13:25
JakeyChan	Do you have document for gitignore ?	13:25
FauxFaux	man gitignore	13:26
jast	automatic message the 'gitignore' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/gitignore.html	13:26
Granjow	Why are they still there? I just did git log --names-only, and tere is no .xcf file at all.	13:27
	→ s0undt3ch joined	13:27
selckin	this a public repo?	13:28
JakeyChan	*	13:29
	!project/	13:29
	I want to ignore all file exception the project directory, is it wright ? in the project have many sub-direction	13:29
	sub-directory	13:29
	→ cakehero joined	13:30
Granjow	selckin: Nope. The one I want to split will become a public repository.	13:30
	→ normanrichards joined	13:30
	← johnson09 left	13:30
	→ LekeFly joined	13:31
	→ broonie joined	13:32
	← Jdub71 left	13:33
broonie	Has the algorithm used to decide what to send when doing a fetch got less efficient recently?	13:33
	→ herdingcat joined	13:33
broonie	I've been noticing fetches seem to be transferring substantially more data (eg, over 200M for a daily -next update).	13:34
	← SimonP86__ left	13:34
	← verto__ left	13:36
banisterfiend	broonie: did you upgrade git recently?	13:38
	← hyperair left	13:39
Granjow	git log --name-only --follow --all -- *.xcf shows me the commit 29dc3f08557d3e4b0542a29cfd9bafea5cf46ed1.	13:41
	But this commit is not in «git log» at all.	13:41
	Why?	13:41
	← joel_falcou left	13:42
Granjow	And how can I really get rid of it?	13:42
	→ TeckniX joined	13:42
wereHamster	Granjow: 'git log' displays the log of "HEAD"; git log --all displays logs of all refs	13:42
	Granjow: git branch --contains 29dc3f08557d3e4b0542a29cfd9bafea5cf46ed1	13:42
	that'll tell you in which branch that commit is	13:43
Granjow	wereHamster: No output there. How can I see all refs?	13:43
	→ dr_lepper joined	13:43
	→ stermi_ joined	13:43
stermi_	hi all :)	13:43
wereHamster	Granjow: git describe 29dc3f08557d3e4b0542a29cfd9bafea5cf46ed1 ?	13:44
	Granjow: or git branch --all --contains 29dc3f08557d3e4b0542a29cfd9bafea5cf46ed1	13:44
Granjow	wereHamster: The commit is in git log --all indeed. git describe says: fatal: No tags can describe '29dc3f08557d3e4b0542a29cfd9bafea5cf46ed1'.	13:44
	→ freelock joined	13:44
stermi_	Example: I add a key to to gitolite for example emanuele.pub and I add emanuele to manage my repositories	13:45
wereHamster	Granjow: so in which branch is it?	13:45
stermi_	I want to use this to access to my git repos from my windows client	13:45
Granjow	wereHamster: no --all option for git branch ... git 1.7.5.3	13:45
stermi_	which is the link from my git on windows and the key on server? the username on windows?	13:45
wereHamster	Granjow: git branch -a --contains 29dc3f08557d3e4b0542a29cfd9bafea5cf46ed1	13:45
Granjow	wereHamster: No output.	13:46
	→ abhatnag joined	13:46
Granjow	Hm	13:47
sitaram	stermi_: http://sitaramc.github.com/gitolite/doc/gitolite-and-ssh.html (this link explains in great detail how gitolite (and similar programs) use ssh)	13:47
Granjow	I cloned the original repository with /path. I'll try file:///path now.	13:47
	← stermi_ left	13:48
	→ subsume joined	13:48
wereHamster	didn't I say you should do that lke 30 minutes ago?	13:48
	→ [intra]lanman joined	13:48
Granjow	wereHamster: From the original repository? I did it, but from the one I already run filter-branch on.	13:49
	→ verto__ joined	13:49
	← cooloney left	13:49
	→ SimonP86 joined	13:50
	← airborn_ left	13:51
	→ sou_ joined	13:51
Granjow	Oh. With file:/// the size of the repository has changed to ... 100 MB. I really feel stupid now.	13:52
selckin	without tried to be smart and makes hardlinks	13:53
	→ somebody__ joined	13:53
	← somebody__ left	13:53
	← xiangfu` left	13:54
	← is left	13:55
	← SavoSuperCool left	13:55
	← versatiletech_ left	13:55
Granjow	But how can I get rid of those files now.	13:55
wereHamster	which files?	13:56
Granjow	The large ones which do not exist anymore but still exist.	13:56
wereHamster	so do they exist or not?	13:56
	→ elyob joined	13:57
Granjow	wereHamster: The .xcf which is in a completely different directory and only appears with git log --name-only --follow --all -- *.xcf	13:57
wereHamster	if you want to remove files from the history, use git filter-branch	13:57
Granjow	I followed the steps in the answer marked as answer here: http://stackoverflow.com/questions/359424/detach-subdirectory-into-separate-git-repository	13:58
	Using file:/// for the source repository.	13:58
	→ Ryback_ joined	13:59
	← Ryback_ left	13:59
	→ Ryback_ joined	13:59
Granjow	I filtered cworkspace/, and the xcf file is in Graphics/	13:59
	as before, git branch -a --contains 29dc3f08557d3e4b0542a29cfd9bafea5cf46ed1 does not return anything	13:59
selckin	you know that won't change the original right	13:59
	← gavin_huang left	14:00
	← alnewkirk left	14:00
	→ f0i joined	14:00
Granjow	selckin: I know. What I want to do is create a separate repository from the cworkspace/ subdirectory. This worked partially, it shows me the files of the cworkspace/ subdir and only the commits from there, but it should be rather 2 MB than 100 MB.	14:01
	→ Mannequin joined	14:01
Granjow	(The splitted repository is too big. The original one should stay as is for now.)	14:02
	→ v0n joined	14:02
	→ boaz_ joined	14:02
	← jutaro left	14:02
	→ gozoner joined	14:04
	→ ceej joined	14:05
	← michas left	14:05
	← JohnBat26 left	14:05
	← centipedefarmer_ left	14:06
	← shiba_yu36 left	14:07
broonie	banisterfiend: Probably.	14:09
	→ eikaas joined	14:09
Granjow	This is such a waste of time	14:10
	→ RobertLaptop joined	14:10
selckin	try providing more information and exactly what you did etc	14:11
	← eikaas_ left	14:11
	→ mandric joined	14:12
Granjow	What I did is this: http://paste.debian.net/119330/	14:13
	← cakehero left	14:14
	← sebbi left	14:15
Granjow	selckin: Anything that looks suspicious?	14:16
	← freelock left	14:16
selckin	well why aren't you checking the size of .git	14:16
Granjow	selckin: Because the size of .git is the size of the current directory minus 1 MB	14:17
	→ psoo joined	14:18
cm\|tmp	How do I ignore all directories in a specified directory with git?	14:18
	→ alextheger joined	14:19
cm\|tmp	.gitignore*	14:19
_rane	try adding */ to .gitignore in that directory	14:19
	→ madewokherd joined	14:19
selckin	Granjow: try starting from scratch and doing it properly	14:20
	→ joel_falcou joined	14:20
cm\|tmp	That worked, thanks	14:20
patrikf	Granjow: just do a fresh clone	14:20
selckin	Granjow: you also never ran gc after the proper expire	14:20
patrikf	Granjow: or follow the checklist in the filter-branch manpage (but it's easier to just do a clone)	14:21
erlnoob	hi, I can't find the semantics of doing just "git fetch" ... while the man page for doing just "git push" is pretty clear (i.e, defaults to current branch's remote)	14:21
Granjow	Okay, you tell me step by step what I have to do. What do I clone now. The original repository?	14:21
	← urkud_ left	14:22
	← d00dlez left	14:22
	→ notjohn joined	14:22
patrikf	no, the filtered	14:22
Granjow	Which one? The one from the pastebin?	14:22
patrikf	you should know.	14:22
	→ shiba_yu36 joined	14:22
	← meuh left	14:23
	→ alex__c2022 joined	14:23
Granjow	patrikf: Okay, I did. Now what?	14:23
	← preflex left	14:24
patrikf	check that the clone has everything you need, check that it's as big or small as expected and if that's the case throw away the one you cloned it from	14:24
Granjow	It has everything I need, but also a lot that I do _not_ need, as all other clones before.	14:24
FauxFaux	Well, the stuff hasn't been filtered, then.	14:25
patrikf	then you didn't run filter-branch properly	14:25
FauxFaux	Caused by ^ ¬_¬	14:25
SethRobertson	Did you filter the tags too?	14:25
patrikf	causilty ftw!	14:25
	causality*	14:25
Granjow	patrikf: Then I don't know how to run it properly. How do I?	14:25
FauxFaux	You work it out by reading the man page.	14:26
Granjow	SethRobertson: I don't know. I used git filter-branch --subdirectory-filter cworkspace/slowmoGPU/ -- --all as suggested by the man page.	14:26
patrikf	so if you said it has stuff you don't need, what did you refer to?	14:26
Granjow	FauxFaux: No. I did everything as written in the man page, and it does not work.	14:26
	← jatt left	14:27
selckin	do you still have the fiels in the workdir?	14:27
Granjow	patrikf: git log --name-only --follow --all -- *.xcf shows me a commit with files like Graphics/svg/x3/X3.xcf	14:28
	which is definitely not in the cworkspace/ directory	14:28
	→ Eko joined	14:28
Granjow	selckin: I cannot see them, only with the git log command.	14:28
	→ meuh joined	14:29
SethRobertson	Granjow: Do you have tags?	14:29
patrikf	Granjow: then you probably wanted to use filter-branch --prune-empty	14:29
	→ preflex joined	14:29
Granjow	patrikf: Still 100 MB.	14:29
patrikf	if not asked to, filter-branch will not remove commits that end up being no-ops.	14:29
Granjow	SethRobertson: There is one, yes.	14:30
	→ nevyn^_ joined	14:30
patrikf	SethRobertson has a good lead, too.	14:30
SethRobertson	Granjow, then I point you to the checklist for shrinking a repository: You really filtered all refs: use `--tag-name-filter cat -- --all` when calling `git-filter-branch`	14:30
Granjow	patrikf: I run git filter-branch --prune-empty on the cloned filtered clone.	14:30
	← opalka left	14:30
	→ thierryp joined	14:31
	→ jutaro joined	14:31
Granjow	SethRobertson: I can do this now, right? Or must I do it when applying the subdirectory-filter?	14:31
	← stermi left	14:31
	← bkozal left	14:31
patrikf	the latter	14:31
	(for the tag part, don't know if prune-empty also removes commits that were empty already before the filtering)	14:32
Granjow	So I use this command: git filter-branch --subdirectory-filter cworkspace/slowmoGPU/ -- --all --prune-empty --tag-name-filter cat	14:33
SethRobertson	Granjow: I don't think using those arguments are every wrong but I will admit to have not studied the problem	14:33
Granjow	(on the clone of the original repository)	14:33
	okay?	14:33
	→ Tac joined	14:33
patrikf	shuffle it so that "-- --all" is the last part	14:34
Granjow	git filter-branch --subdirectory-filter cworkspace/slowmoGPU/ --prune-empty --tag-name-filter cat -- --all	14:34
patrikf	because -- separates arguments to filter-branch and arguments to rev-list	14:34
	sounds good	14:34
	← ceej left	14:35
Granjow	Okay. 103 MB. Next step? Remove the remote origin?	14:35
	← silouck left	14:35
patrikf	clone again and check whether the clone has everything you want and lacks everything you don't want	14:35
	→ ceej joined	14:36
Granjow	patrikf: Done. Has everything I want and everything I don't want as well. Still 103 MB.	14:36
patrikf	okay, please give an example for something you don't want	14:37
	→ versatiletech joined	14:38
	← versatiletech left	14:39
	→ versatiletech joined	14:39
	← thierryp left	14:39
Granjow	patrikf: Hm. Now the .xcf has gone. git-find-large finds the largest file with 75 kB.	14:40
	→ thierryp joined	14:40
Granjow	Can I try git gc --aggressive now?	14:40
	← BiggFREE left	14:41
patrikf	that won't do anything on a fresh clone, I think	14:41
	→ fakewaffle joined	14:41
patrikf	hmh	14:41
Granjow	So it is possible that 60 commits on simple C++ files result in a 100 MB history?	14:41
	→ jfkw joined	14:41
Granjow	100 MB _compressed_ history	14:41
	I mean, I do work a lot, but not that much.	14:41
	← slavyXS left	14:42
	← slavy60xs left	14:42
	← namemon left	14:42
SethRobertson	Granjow: Seems unlikely but it depends on the file size. You can look at each reachable commit to try and find how big it is	14:42
	→ dfr\|mac joined	14:43
Granjow	SethRobertson: Is there an easy way to do so? Can I output all diffs, for example?	14:43
SethRobertson	`git log -p --all \| wc` should do it.	14:44
	← robgleeson left	14:44
Granjow	The git-find-large says the top10 range from 75 kB to 25 kB.	14:44
	→ whyking joined	14:44
Granjow	66236 255781 2198520	14:44
	→ ericbarnes joined	14:44
Granjow	(output from git log -p --all \| wc)	14:44
patrikf	and how do you check the git repository's size?	14:45
Granjow	du . -hs.	14:45
	or cd .git; du . -hs	14:45
	138 MB at the moment in the .git directory.	14:45
	← freshmilk left	14:46
patrikf	wait a minute	14:46
	could git just have hardlinked the packfile?	14:46
	crap, the filter-branch manpage even has that	14:47
Granjow	I always used file:///	14:47
	← mishok13 left	14:47
patrikf	oh	14:47
	← mgaunard left	14:47
	→ judu joined	14:47
selckin	Granjow: have to run expire before the gc, with all the options listed earlier	14:47
	← ksinkar left	14:47
patrikf	but a fresh clone wouldn't benefit from that	14:47
	← rolfb left	14:48
Granjow	So I'll just live with it?	14:49
	← judu left	14:49
	→ judu joined	14:49
Granjow	What happens if I commit files in the cworkspace directory plus a large one somewhere outside this directory. Will this be included?	14:50
judu	Are there any gitolite and ruby users?	14:51
Granjow	.git/objects/* contains some files that are > 2 MB	14:51
	← thierryp left	14:51
Granjow	Can't be this either though. I used to check in a time tracking db, but its size is only 50 kB. And it is not listed in git-large-files.	14:52
judu	I need help with a ruby hook, using rvm to supply ruby, and gitolite doesn't know it…	14:52
patrikf	Granjow: so what are these objects? (check with git show)	14:52
	← jutaro left	14:52
	→ deqer joined	14:52
	← gnufied left	14:53
deqer	hey guys	14:53
	it's me again	14:53
Granjow	And according to wc, the third column is the number of bytes (git log -p --all), so 2 MB is < 100 MB.	14:53
	→ Modius joined	14:53
	→ HG` joined	14:53
	ChanServ set mode: +v	14:53
selckin	send me the repo	14:53
Granjow	patrikf: git show what? Git show filename of the .git/objects/8e/c7e8fa340444fbe751b35eb6dcebf520471c84 file?	14:53
abhatnag	Granjow: go 'git cat-file -p 8ec7e8fa340444fbe751b35eb6dcebf520471c84' to see what's in that object	14:54
	← subsume left	14:54
	→ ShooterMG joined	14:54
sitaram	judu: what type of hook is this? (pre-receive, post-receive, ...etc)?	14:55
Granjow	abhatnag: It is a file from Graphics/ which is such kind of not inside cworkspace/.	14:56
judu	it's an update hook	14:56
	← nevyn^_ left	14:56
	← LekeFly left	14:56
	→ mishok13 joined	14:57
judu	My problem is that gitolite cannot find the gems I want	14:57
	← Modius left	14:58
Granjow	abhatnag: To be precise: It is a: JPEG 3034x2024 3034x2024+0+0 8-bit DirectClass 2.198MB 0.000u 0:00.009 (last	14:58
	← meuh left	14:58
Granjow	To be precise, I didn't even know this file still exists	14:58
sitaram	judu: if you didn't use gitolite how would it find the gems? Do it the same way except call the hook "update.secondary" instead of "update" (because gitolite already uses the "update" hook)	14:58
	→ meuh joined	14:58
abhatnag	Granjow: sorry I picked up on this late, what's the problem?	14:59
	→ subsume joined	14:59
	← v0n left	15:01
	← Jordano left	15:01
Granjow	abhatnag: The problem is: I have a big repository and a project in cworkspace/slowmoGPU which will eventually be OpenSource. I want to split this into a new git repository. It consists of text files only, but when I «git filter-branch» it, its size is at 100 MB (60 commits, total size of all files is 1 MB).	15:01
	← bartek left	15:01
	← radarek left	15:02
abhatnag	Granjow: have you tried a visual tool to locate the big files? what OS are you on?	15:02
	← daniel-s left	15:02
	← cesario left	15:02
selckin	did you actually do the expire+gc with all the options after you did the filter w/ tags?	15:03
	→ nevyn^_ joined	15:03
	→ bartek joined	15:03
	→ v0n joined	15:03
Granjow	abhatnag: I'm on linux. When I'm using the git-find-large script from https://git.wiki.kernel.org/index.php/GitFaq it reports that the biggest file is 75 kB. And the JPG file from above definitely should not be in the repository anymore since it was in a subsubdirectory of the Graphics/ subdirectory which is not a subdirectory of cworkspace.	15:03
	→ Spockz joined	15:04
Granjow	selckin: I just did git reflog expire --expire=now --all && git repack -ad again: 141 MB.	15:04
judu	sitaram: I am already using update.secondary. When using the system ruby distribution, no problem, but I want to use rvm (ruby version manager). I think that gitolite does not load the proper environment.	15:05
abhatnag	Granjow: so you're just trying to locate which files are the largest? Depends on what distro you're on, but try getting something like 'Disk Usage Analyzer' (prepacked with yum based distros) on your slowmoGPU	15:05
	→ Corren joined	15:05
	← dr0id left	15:05
judu	Here is what I get when I push : http://paste.pocoo.org/show/403505/	15:05
selckin	Granjow: you're not running gc	15:05
Granjow	Oh.	15:05
	git gc --prune=now ... 3.7 MB.	15:05
	← Spockz left	15:05
canton7	ncdu is a good curses tool	15:05
selckin	Granjow: i told you to run that like as my 2nd line	15:05
Granjow	That's a lot better.	15:06
	selckin: I run it like 100 times today.	15:06
selckin	Granjow: and didn't see it in the pastebin which i said again	15:06
Granjow	selckin: http://paste.debian.net/119337/ ;)	15:07
	→ tore_ joined	15:07
	→ _psychic_ joined	15:07
selckin	yes you where right all along, we where all wrong	15:07
	← nevyn^_ left	15:07
	→ devth_ joined	15:08
Granjow	I'll examine my history now ... write everything down ... and try again to check whether it works again :D	15:08
	← budsondixon left	15:08
Corren	ok, I created a local branch, made a bunch of commits, merged that into master	15:08
	and now I need to revert out all of those changes	15:08
	how would you do that?	15:08
Granjow	selckin: I did not say anything like that.	15:08
sitaram	judu: I dont know anything about ruby, but if you're trying to install ruby software using a hook you're probably doing something very very wrong	15:08
canton7	Corren you've answered your own questio :) git revert <sha1 of merge commit>	15:09
	← ruby_on_tails left	15:09
	← JakeyChan left	15:09
Corren	yea, but when I look at the commits on the github page, I see all of my commits, not just the merge	15:09
	→ schnatta joined	15:09
judu	sitaram: I am not trying to install ruby software in my hook. But to set the right environment, I have this in my bash/zshrc : [[ -s "/usr/local/rvm/scripts/rvm" ]] && . "/usr/local/rvm/scripts/rvm"	15:10
Granjow	selckin: What I meant to say is that I was amazed that it worked this time and not all other times before.	15:10
canton7	Corren, hmm? If you revert the merge commit, it will revert all changes made by the merge, and so the branch which culminated in the merge	15:10
Corren	rgr	15:10
judu	Is there a way to tell gitolite to initialize its environment with this command ?	15:11
Corren	now, does revert roll back the entire repo, or just extract that commit?	15:11
sitaram	judu: that statement looks fine, assuming your update.secondary hook is in shell	15:12
	→ centipedefarmer_ joined	15:13
judu	sitaram: So, you're saying that I should have an update.secondary written in bash, that sets its environment using that statement and then calls a ruby script ?	15:13
	← fedesilva left	15:14
	→ drev1 joined	15:14
	← EricInBNE left	15:15
sitaram	judu: I have no idea what you're doing so I can't honestly answer that sorry	15:15
	can only guess	15:15
	← Scorpion44a left	15:15
	← tore_ left	15:16
sitaram	judu: if I reduce your question to "how do I ensure such-and-such env var is set and then such-and-such program is run from the update hook" then yes that is what you need to do	15:16
	← cbreak_work left	15:17
	← evorocks left	15:20
	→ corra joined	15:20
	← verto__ left	15:20
	→ chrisdev joined	15:21
	← dpino left	15:21
jpalmer	Let me start by saying, I'm not a developer. so excuse the possibly retarded question. Is there a way to have a file populated with the git commit hash, as it's being commited? (seems like the hash would be generated after the commit, so I suspect I know the answer	15:21
selckin	not really	15:21
canton7	there is a workaround though. i think it's to do with one of the gitattrbutes filters? so the commit id is written to the file as it's checked out, rather than as it's committed?	15:22
jpalmer	We have an application stack that.. after it's installed.. we don't have a good way to determine what version was installed from git. so I'm trying to figure out a decent way to be able to identify versions.	15:22
	← german__ left	15:22
	← _psychic_ left	15:22
selckin	it's a really bad idea anyway, your build system should be doing it if needed	15:22
	→ _psychic_ joined	15:22
selckin	so put it a file at build time and include it	15:23
	→ iori joined	15:23
judu	sitaram:That's exactly what I needed. Thanks a lot, I was not at all in the "bash script to set env then call my ruby script" scope!	15:23
	← kpreid left	15:23
jpalmer	selckin: so you're thinking a file (named verstion.txt for the sake of discussion) and as we rlease ne versions of the application stack, we manually update the file?	15:24
	→ jutaro joined	15:24
jpalmer	wow, can't type: "as we release new"	15:24
	← boaz left	15:24
	boaz_ → boaz	15:24
	→ boazsender joined	15:25
	← psoo left	15:26
	→ halfline_ joined	15:26
	← j4mie left	15:28
	← Corren left	15:29
	→ abhishek joined	15:29
	← compnerd left	15:29
	← Murr left	15:29
	→ alvaro_o joined	15:30
	→ sebersole joined	15:31
	← abhatnag left	15:31
	→ Raging_Hog joined	15:31
	→ compnerd joined	15:32
	→ fedesilva joined	15:32
	→ verto__ joined	15:32
	→ mkramer joined	15:32
	→ namemon joined	15:33
	→ tystr joined	15:33
selckin	well usually you have a build step, like compiling etc, which is fully automated and scriptable, and capable of auto creating the file and including it	15:34
	← xxtjaxx left	15:34
	→ pandeiro joined	15:36
	← banisterfiend left	15:37
	→ cilly joined	15:38
	→ banisterfiend joined	15:38
	← alvaro_o left	15:39
	→ Enchilada joined	15:40
	ChanServ set mode: +v	15:40
cbz	who can i speak to about getting a permanent voice?	15:41
_rane	so what do you need a voice for?	15:41
	← buddybro left	15:41
wereHamster	cbz: register your nick	15:41
cbz	My nick is registered.	15:41
	→ stermi joined	15:42
	← jonshea left	15:42
cbz	_rane: so i don't have to keep requesting it on the website.	15:42
Granjow	At the end it was just very few commands for extracting the directory: http://paste.debian.net/119351/	15:42
wereHamster	are you logged into your freenode account?	15:42
	→ thiago_home joined	15:42
Granjow	cbz: Did you also identify? /msg nickserv identify PASSWORD	15:42
	← alip left	15:42
	← Leemp left	15:43
_rane	i don't have a voice and i can talk	15:43
	weird stuff!	15:43
Granjow	selckin, wereHamster, patrikf -- thanks a lot for your help and patience!	15:43
	_rane: No, I cannot read you.	15:43
Enchilada	Whenever I download something from GitHub, or wherever in the Git community, why does the working copy always accompany the .git folder? Why isn't only the .git folder transmitted since .git has all the data anyway?	15:43
	I don't get it. Seems a waste of space.	15:43
	→ Leemp joined	15:43
Granjow	:-]	15:43
	← subsume left	15:43
Enchilada	waste of internet bandwidth, that is	15:43
wereHamster	Enchilada: git clone --bare	15:43
selckin	Enchilada: why do you think thats not allready the case?	15:44
stermi	sitaram, I've installed msysgit+tartoise svn imported pub key created with msysgit tool, updated gitolite-admin but it wont work :'(	15:44
wereHamster	Enchilada: also, only the contents of the .git folder are transfered, then unpacked into your working tree	15:44
cbz	Enchilada: the working copy isn't transmitted separately	15:44
_rane	Enchilada: I think the working copy is created based on .git so it's not actually bandwidth wasted	15:44
wereHamster	Enchilada: so no bandwidth is wasted	15:44
Enchilada	ah ok	15:44
	thanks	15:44
wereHamster	did you really expect Linus to be that stupid? :)	15:44
	→ soc42 joined	15:45
sitaram	stermi: http://sitaramc.github.com/gitolite/doc/ssh-troubleshooting.html (this page and stuff linked from it is about all the help I can give you at this time, sorry)	15:46
	← devth_ left	15:46
stermi	anyway I damn hate windows	15:46
	← crakdmirror left	15:46
stermi	I don't understand how people can develop with it	15:46
	← zwick left	15:47
selckin	stermi++	15:47
	← mastroDani left	15:47
	→ alip joined	15:47
stermi	selckin, that's mean you quote me?	15:48
	→ zwick joined	15:48
	← zwick left	15:48
	← Ownatik left	15:48
	→ crakdmirror joined	15:48
	→ dpino joined	15:49
	← hendersonj left	15:50
	← jceb left	15:50
	→ Spockz joined	15:51
	→ d0k joined	15:51
	→ abhishek_ joined	15:51
	→ halfl_ne_ joined	15:51
	→ kenichi joined	15:52
	→ fedesilva_ joined	15:53
	← verto__ left	15:53
	→ mastroDani joined	15:53
	← fedesilva_ left	15:53
	→ amerine_ joined	15:53
	← abhishek left	15:54
	← halfline_ left	15:54
	← jophish_ left	15:54
	← cenuij left	15:55
	→ jophish_ joined	15:55
	→ shurane joined	15:56
	→ jmd joined	15:56
	← fedesilva left	15:56
stermi	the only possible thing it's when my friend use git on windows it's not using (taking automatically) the right ssh key	15:56
	← Behi left	15:57
	→ hendersonj joined	15:57
	← corra left	15:57
selckin	msysgit+plink+pagent work fine	15:58
sitaram	friends dont let friends use windows	15:58
	← jutaro left	15:58
	selckin is considering bringing a hd from home to be able to use linux :(	15:58
Enchilada	Would I therefore be correct in guessing that it's enough to back up the hidden .git folder when backing my shit up?	15:58
sitaram	USB stick will do, selckin	15:59
Enchilada	Since the rest has already been commited.	15:59
	← judu left	15:59
selckin	sitaram: if this gets any slower the outcome might not be pritty	15:59
	→ tjholowaychuk joined	15:59
	← pyrony left	15:59
sitaram	aah	15:59
	→ PHadm joined	15:59
selckin	usb3 gogo :(	15:59
	→ mikepack joined	16:00
stermi	sitaram, friend don't let friends use also mac, but it's harder	16:01
	← alextheger left	16:01
	→ spearce joined	16:02
selckin	mac is fine	16:02
	ChanServ set mode: +v	16:02
selckin	can only have policy objects, its good to dev on	16:02
PHadm	hello all... contest with Git. So far, git 1; me 0. Trying to use .gitignore to exclude settings.php. .gitignore is ignoring the ignore	16:03
stermi	said a mac user :D	16:03
selckin	i'm not	16:03
	PHadm: git will never ignore a tracked file	16:03
PHadm	thats the point.. I don't want to track it	16:03
	← RaceCondition left	16:04
jmd	selckin, What if it's in .gitignore ?	16:04
selckin	jmd: never.	16:04
	→ SimonNa joined	16:04
	← dfr\|mac left	16:04
	← centipedefarmer_ left	16:04
selckin	PHadm: is it tracked allready or not?	16:04
	← kenichi left	16:05
	→ hyperair joined	16:05
	→ kenichi joined	16:05
PHadm	yes, it already is	16:05
wereHamster	PHadm: git rm --cached settings.php && git commit -m 'untrack settings.php'	16:05
	→ zwick joined	16:05
	← kenichi left	16:06
PHadm	perfect.. thank you	16:06
	→ verto__ joined	16:06
	← jophish_ left	16:07
	→ Buglouse joined	16:07
	→ vivekp joined	16:07
	→ urkud_ joined	16:08
	→ ctshryock joined	16:08
	→ nevyn^_ joined	16:08
Granjow	Is it true that gitosis has been replaced by gitolite?	16:09
	← jfelchner left	16:09
wereHamster	yes	16:09
	!gitosis	16:09
jast	automatic message gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite	16:09
	→ RaceCondition joined	16:09
Granjow	Aye, thank you. So I will switch my server to gitolite.	16:10
	← PHadm left	16:10
	→ tearsong joined	16:10
	← __name__ left	16:10
	← vivekp left	16:11
	← zwick left	16:11
	← ctshryock left	16:11
	→ Swimming_bird joined	16:12
	→ jonshea joined	16:12
	→ jimbo-jones joined	16:12
	→ dan23232 joined	16:14
	← capncodewash left	16:14
	→ jophish_ joined	16:14
	→ capncodewash joined	16:17
	→ cesario joined	16:17
Granjow	wereHamster: It seems that gitolite just works(TM) o.O That's very cool.	16:17
	→ Murr joined	16:19
	← xenoterracide left	16:19
	← herdingcat left	16:20
	← PabloM left	16:20
	→ akosikeno joined	16:21
	← sou_ left	16:21
	→ herdingcat joined	16:21
	← capncodewash left	16:22
	← Badiss left	16:22
	→ ksawicki joined	16:22
	→ bgy joined	16:23
bgy	hi	16:23
	how could I cancel a pushed commit?	16:23
wereHamster	bgy: you push one of its ancestors to the remote repo	16:23
bgy	wereHamster: like git push origin HEAD^	16:24
	?	16:24
wereHamster	no, git push origin HEAD^:master	16:24
	→ lectrick joined	16:24
bgy	wereHamster: ok thnak you!	16:25
	→ pyrony joined	16:25
	← Spockz left	16:25
	→ metze joined	16:25
	← tvw left	16:26
	← verto__ left	16:26
	→ btanaka joined	16:26
	→ Maletor joined	16:26
jmd	bgy, But if anyone has pulled from that repo, they're going to have problems.	16:27
	→ Corren joined	16:27
	→ Ownatik joined	16:27
bgy	jmd: yeah I know that, in fact I don't advise to do that but the question was asked to me and I was wondering what is the correct way to do	16:29
	→ fedesilva joined	16:29
	→ centipedefarmer_ joined	16:30
	← mkramer left	16:32
	→ cakehero joined	16:32
	← nevyn^_ left	16:32
jmd	What's the equivalent of mercurial's "transplant" command?	16:36
	← tewecske left	16:37
	→ xerora joined	16:38
	→ verto__ joined	16:38
pasky	jmd: it seems hg transplant is akin a bit to git rebase (-i) and a bit to git cherry + git cherry-pick	16:38
	← RaceCondition left	16:39
	→ nakkor joined	16:39
	← nakkor left	16:39
	→ kenichi joined	16:40
	← lunks left	16:41
	q0rban → q-rban	16:41
	→ kedare joined	16:41
	→ verto_ joined	16:41
	← ExtraSpice left	16:41
	← pandeiro left	16:41
	→ chittoor joined	16:42
	→ smolyn joined	16:43
	→ lunks joined	16:43
	← verto__ left	16:43
	← shurane left	16:43
	→ PabloM joined	16:43
	→ Vaati joined	16:43
	← verto_ left	16:44
	→ Travis-42 joined	16:44
	← smolyn left	16:44
	← urkud_ left	16:45
	ChanServ set mode: +v	16:45
Vaati	hello all	16:46
	← centipedefarmer_ left	16:46
	← Enchilada left	16:46
Vaati	... meaning 7 people ... ;-)	16:46
	thats an ingenious method of stopping spam, I must admit	16:46
	→ dnuke joined	16:47
	canton7 is confused...	16:48
Vaati	about what?	16:48
canton7	aah i get you. nah, everyone in here can talk. i think the +v thing is to do with people using online irc?	16:49
Vaati	oh	16:49
	← pcav left	16:49
Vaati	nope, im using irssi...	16:49
	weird	16:49
	I thought it was some other form of +m that was going on here..	16:49
sitaram	Vaati: I see 812 nicks online on this channel right now	16:50
Vaati	for 800 some people, it is quiet though	16:50
	;-)	16:50
	→ rgr joined	16:51
offby1	I thought you got +v only if your nick was "registered" ...	16:51
	→ spaceonline joined	16:51
Vaati	offby1: do /topic	16:52
Arrowmaster	no you goto the webpage in the topic and pass a captcha or something and it +v's you	16:52
offby1	huh, I've never done that	16:52
Vaati	so, can I complain about disfunctional git repositories here ? lol..	16:52
	→ pandeiro joined	16:52
	← achadwick left	16:52
	← g0bl1n left	16:53
bremner	offby1: you are a registered ... nick	16:53
Vaati	I am too, but its not registered to me, i'm borrowing this nick	16:54
	← babilen left	16:55
	→ troessner joined	16:55
	← spaceonline left	16:55
	→ ExtraSpice joined	16:56
	← Vaati left	16:56
	← kedare left	16:56
	→ kedare joined	16:57
offby1	so I'm half-right: you get "voiced" either if you're registered, or if you've filled in the captcha	16:57
	← troessner left	16:57
	← btanaka left	16:57
	→ ki__ joined	16:57
	→ jrock20041 joined	16:57
jmd	I can't get cherry-pick to accept more than one commit	16:58
	← Travis-42 left	16:59
	→ n1xNc0d3 joined	17:00
	← Ownatik left	17:01
	← devth left	17:01
	→ devth joined	17:02
	→ RaceCondition joined	17:02
	← stamina left	17:03
	← RaceCondition left	17:03
	→ urkud_ joined	17:03
jrock20041	Is there a tutorial on how to merge an existing svn repo into an existing git repo?	17:04
	→ gdmfsob joined	17:04
	← pandeiro left	17:05
	← davzie left	17:06
	← Murr left	17:06
	← mishok13 left	17:06
	→ radarek joined	17:06
	→ stamina joined	17:06
	→ SimonP86_ joined	17:07
	← pyrony left	17:07
offby1	jmd: I don't think it's intended to take more than one.	17:07
	jmd: if you have more than a couple of commits, maybe you should be rebasing instead of cherry-picking	17:08
jmd	the man page says 'Given one OR MORE existing commits ...'	17:08
	(my emphasis)	17:09
	→ SimonP86__ joined	17:09
	← SimonP86 left	17:10
	→ gnufied joined	17:10
	→ Rotham joined	17:11
	→ german__ joined	17:12
	← SimonP86_ left	17:12
	→ spaceonline joined	17:12
	← hendersonj left	17:13
	← RobertLaptop left	17:14
canton7	jmd it's pretty new. do you have a recent version of git?	17:14
jmd	what's new?	17:15
	→ dreiss joined	17:15
	Vampire0_ → Vampire0	17:15
canton7	cherry-pick taking a range of commits	17:15
jmd	I'd be very surprised if my man page was newer than my git binary.	17:15
	← deqer left	17:15
canton7	fair play, just making sure you weren't reading the latest man page off the new, but with an old binary	17:16
jmd	git version 1.7.2.3	17:16
	→ kpreid joined	17:16
	→ pantsman joined	17:16
	← pantsman left	17:16
	→ pantsman joined	17:16
canton7	s/new/net/	17:16
	→ Murr joined	17:17
	→ jutaro joined	17:17
	← Zenopus left	17:17
	→ deqer joined	17:18
	→ Zenopus joined	17:19
	← Rotham left	17:19
	→ centipedefarmer_ joined	17:20
	← leo2007 left	17:21
	→ kingfishr joined	17:21
	← Leemp left	17:21
	→ Leemp joined	17:22
	→ yhager joined	17:23
	← curtana left	17:24
	→ Ownatik joined	17:25
	→ metcalfc joined	17:25
	← deqer left	17:25
	← n1xNc0d3 left	17:25
	→ drizzd joined	17:26
	→ RobertLaptop joined	17:26
	→ deqer joined	17:27
	← kpreid left	17:27
	← tatsuya left	17:27
	→ kpreid joined	17:28
	→ Xilent joined	17:29
	→ zwick joined	17:30
	← kedare left	17:30
	← mandric left	17:31
	→ cenuij joined	17:32
	← cenuij left	17:32
	→ cenuij joined	17:32
	→ mandric joined	17:32
	→ fedesilva_ joined	17:32
	→ gusnan joined	17:32
	← jutaro left	17:33
	→ hobodave joined	17:33
	→ alcuadrado joined	17:33
alcuadrado	Can I push everything except one commit?	17:33
	One of the devs modified an specific config file, and in order to pull he committed it locally	17:34
	now he has to push everything except that commit	17:34
albel727	that was a stupid idea. undo the merge commit, and make that dev aware of git stash.	17:35
	→ remi_ joined	17:35
	→ HG`` joined	17:35
	← remi_ left	17:35
	→ remi_ joined	17:35
cbreak	alcuadrado: you can push parts of the history	17:35
albel727	crap. 4chan folk has spoiled me. I've lost my politeness.	17:36
	← remi_ left	17:36
cbreak	but only if it's up to a limit	17:36
	→ maurer_nopriv joined	17:36
	← stermi left	17:36
cbreak	you can not leave away a commit in the middle	17:36
	← fedesilva left	17:36
alcuadrado	cbreak, great! how?	17:36
	fedesilva_ → fedesilva	17:36
cbreak	just reset away the commit you want to not push	17:36
	(put it into a new branch before)	17:36
Granjow	I've got two users in gitolite. How can I read as the second user? Or, generally, how exactly am I identified? Just by the SSH key I use for the connection?	17:37
	→ btanaka joined	17:37
	← soc42 left	17:37
cbreak	Granjow: yes	17:37
	← mastroDani left	17:37
alcuadrado	btw, is there a way to say to git "leave this file in this revision?"	17:37
	→ pyrony joined	17:38
cbreak	alcuadrado: no	17:38
	→ lightcap joined	17:38
	← ksawicki left	17:38
cbreak	git does not do per-file history	17:38
	→ nevyn^_ joined	17:38
	← HG` left	17:38
alcuadrado	yep, I know that, but who knows :)	17:38
	thanks cballou	17:38
	cbreak,	17:38
Granjow	cbreak: I know that I can set the used SSH key in ~/.ssh/config for an IP or hostname. Can I also somehow override this temporarily for a clone/push?	17:38
cbreak	Granjow: with ssh-add maybe	17:39
albel727	alcuadrado: first off, you need to "git reset --hard <that commit sha you don't want to push>". then git branch <newbranch>. then "git reset --hard <that_commit>~1" then git pull. if all the history is ok, then you can push. a file from a specific revision can be obtained with git checkout <sha> -- <file>.	17:39
cbreak	I would do it differently	17:40
alcuadrado	thanks albel727 :)	17:40
cbreak	I'd first git branch backup	17:40
	→ n1xNc0d3 joined	17:40
	→ Chillance joined	17:40
cbreak	then git reset --hard commitbeforetheevilcommit	17:40
	then git push	17:40
	← angelsl left	17:40
Danielpk	Guys, can i push to a no-bare repository?	17:40
albel727	hmm. that probably is safer, yes. you forgot a pull, incidentally. I guess, git reflog has spoiled me too.	17:41
	Danielpk: !bare	17:41
jast	Danielpk: automatic message an explanation of bare and non-bare repositories can be found here: http://bare-vs-nonbare.gitrecipes.de/	17:41
	→ dannas joined	17:41
cbreak	Danielpk: only with lots of care	17:41
	Danielpk: basically, do NOT touch the branch that is checked out	17:42
	→ thakis joined	17:42
cbreak	or who ever uses that branch will get into trouble	17:42
Danielpk	cbreak: i want use it to deploy	17:42
thakis	hi folks. is there an easy way to ask git log which files i changed in the last 2 weeks?	17:42
	→ yota joined	17:42
	← deqer left	17:42
cbreak	Danielpk: git is not for deploying	17:42
	→ mastroDani joined	17:42
cbreak	thakis: ask git diff	17:42
	thakis: man git diff --name-status or what ever	17:43
jast	thakis: automatic message the 'git-diff' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-diff.html	17:43
Danielpk	cbreak: humm i know he isnt, but i cant use Capistrano so...	17:43
cbreak	with the range you want	17:43
	← nlisgo left	17:43
albel727	Danielpk: just read the doc I gave you.	17:44
	→ deqer joined	17:44
thakis	cbreak: i couldn't find a way to tell git diff to only show diffs by me (as in --author)	17:44
	→ jstemmer joined	17:44
cbreak	albel727: hmm, that seems different than faq non-bare	17:44
jast	albel727: automatic message Pushing to non-bare repositories is discouraged: please see https://git.wiki.kernel.org/index.php/GitFaq#non-bare	17:44
Danielpk	albel727: oi will read thx ;)	17:44
cbreak	thakis: that's what git log is for	17:45
albel727	cbreak: of course it does. ) jast wrote it, if I'm not much mistaken. I think, non-bare faq is too hard on a newbie.	17:45
thakis	cbreak: i couldn't find a way to let git log show a list of files joined over several commits	17:45
	hence my asking here	17:45
cbreak	thakis: yeah, it can't really do joins	17:45
captainramen	great news everyone	17:45
cbreak	thakis: but you can use a --format that only has the names	17:46
captainramen	i've finally convinced my shop to convert to the git religion	17:46
cbreak	and then pipe it through sort, maybe	17:46
adymitruk	captainramen: what were they using before?	17:46
	← ramusara left	17:46
	cbreak guesses folders with dates in the name	17:47
Granjow	cbreak: This did not do the trick, but I could add a new Host in ~/.ssh/config: http://help.github.com/multiple-ssh-keys/	17:47
	That's quite nice, it even works :)	17:47
cbreak	yeah, but it's a hack	17:47
Granjow	Why?	17:47
cbreak	if you use ssh-add, you have to unload the other key, and not use the ssh config stuff at all	17:48
	Granjow: because it abuses some fake host name for key selection :)	17:48
	→ zxc___ joined	17:48
captainramen	adymitruk: subversion, what else	17:49
	→ mando joined	17:49
captainramen	adymitruk: although before that they were tacking on the date to the end of filenames	17:49
Granjow	cbreak: Yet it works, and I don't need to ssh-add again when I want to test next time :) But, I already have an entry for my server (the other user) in my .ssh/config; when I now unload the ssh agent, wouldn't it just try the key in there?	17:49
	I mean, how can I force it to use the ssh-add-ed?	17:49
cbreak	it'll use it if no special one is configured I suppose	17:50
	→ Travis-42 joined	17:50
cbreak	I never added override keys to ssh config	17:50
	→ psoo joined	17:50
	← thakis left	17:50
Granjow	So you're doing it via ssh-add only? (no ~/.ssh/config?)	17:50
	← centipedefarmer_ left	17:52
	← mastroDani left	17:52
	← Travis-42 left	17:52
	→ Vile joined	17:52
adymitruk	captainramen: rofl! too funny	17:52
	← herdingcat left	17:54
cbreak	Granjow: yes. I only use one key though	17:54
	→ anjen joined	17:54
Granjow	cbreak: Okay. And thanks for the hints, I think I now understand what ssh-add does :)	17:55
	→ Rogi joined	17:55
cbreak	it saves you from having to type in your pass poem every time you want to use the key	17:55
	or leave it on disk unencrypted	17:55
	cbreak is partially paranoid	17:56
adymitruk	I've got the nasty problem of someone wanting to store excel files in the same repo as the code	17:56
	→ opalka joined	17:57
adymitruk	I know that excel is a compressed folder structure with xml data	17:57
cbreak	adymitruk: for git it's just a file that changes every commit	17:57
	that means: big repository, slow cloning, unresolvable merge conflicts	17:57
adymitruk	I'm wondering if there is a smudge clean script out there to store excel as xml and allow git not to treat these files as binaries	17:58
cbreak	make him a submodule or one of those git slave things	17:58
	and keep his filth out of the main repo	17:58
	← maurer_nopriv left	17:58
	→ mastroDani joined	17:58
jrock20041	Is there a tutorial on how to merge an existing svn repo into an existing git repo?	17:58
	→ mcluskydodallas joined	17:58
adymitruk	cbreak: unfortunately it is something that has to be stored beside the code. it's test cases	17:59
	cbreak: I will get them off of excel in a while for those.. probably just use html	18:00
cbreak	use csv	18:00
	excel can read that	18:00
	and write it too	18:00
	I think...	18:00
albel727	it can	18:00
Granjow	cbreak: Yes, that's true. I'm atm too lazy to change to ssh-add (because the current way works), but the next time I have to set up my PC ... :)	18:01
	← Murr left	18:01
	← btanaka left	18:02
	← marcob left	18:02
	← alberto56 left	18:04
	→ Fretta joined	18:06
captainramen	so, i need to train these people in the ways of git. probably in a conference room on osx	18:07
	is there some super simple online tutorial page i can just load up and follow?	18:07
	← bgerber left	18:07
	← zxc___ left	18:09
cbreak	you should just show general workflow	18:10
	→ zxcp joined	18:10
	ChanServ set mode: +v	18:10
	← cakehero left	18:10
	→ Murr joined	18:10
zxcp	hey guys im trying to set up git on my ubuntu server with this guide: https://help.ubuntu.com/community/Git .... everything works fine until im running "git clone [email@hidden.address] im getting Permission denied (publickey).	18:10
	← aspotashev left	18:10
	→ veronatii joined	18:10
	→ minter joined	18:10
	ChanServ set mode: +v	18:11
minter	Hello - anyone around who might be able to help me figure out how to recover from a corrupted repository?	18:11
	→ joeytwiddle joined	18:12
	→ mma joined	18:12
	→ shurane joined	18:12
	→ bgerber joined	18:13
captainramen	cbreak: only thing i'm not sure of is how to create a merge conflict so I can show them how to fix it	18:15
	cbreak: since it's damn nigh impossible with git =)	18:15
cbreak	just clone the repositoy	18:15
	and add a new function to a class	18:15
	→ justin-george joined	18:15
cbreak	at the same spot	18:15
captainramen	cbreak: thanks!	18:15
	→ threeb joined	18:15
cbreak	easy to solve conflict :)	18:15
selckin	just change the same line twice and try to merge	18:16
	← babbos left	18:16
captainramen	if i rebase one against the other i will still get the conflict right	18:16
	← nevyn^_ left	18:17
cbreak	yes	18:17
captainramen	awesome	18:17
cbreak	might look different though	18:17
minter	When I attempt to do things like switch branches or run "git gc" or "git fsck", I'm getting errors like this:	18:17
	minter:(git)teamsnap.broken[master]/$ git co user_files_paperclip fatal: failed to read object 4406bafc1128e184fe002b68fa347faf79ea795f: Invalid argument	18:17
	I'm stumped as to how to get past this bad object.	18:18
	→ boombatower joined	18:18
	← boombatower left	18:18
	→ boombatower joined	18:18
	← spaceonline left	18:19
cbreak	minter: weird. I've not seen that so far.	18:20
	can you clone the repo?	18:20
bremner	can you fsck (or yourOS equivalent) ?	18:20
minter	I can clone the remote repo, but then all my stuff in the local branches is inaccessible.	18:20
adymitruk	cbreak: I think tab separated is better as this data is very textual and you can find commas in the data	18:20
cbreak	separate them with what ever you want...	18:21
	as long as it's text, git should be able to deal with it reasonably well	18:21
	minter: I mean, can you clone the local one	18:22
minter	Hm, lemme try that.	18:22
zxcp	hey guys im trying to set up git on my ubuntu server with this guide: https://help.ubuntu.com/community/Git .... everything works fine until im running "git clone [email@hidden.address] im getting Permission denied (publickey).	18:22
	← noisebleed left	18:22
	← erichammond left	18:23
minter	No, cloning the broken repository gives the same error - "remote: error: Could not read 440cf08c2207cf3aa1132966fcec10b2ca4a601e"	18:23
bremner	!gitosis	18:23
jast	automatic message gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite	18:23
	← mma left	18:23
cbreak	minter: can you find that file?	18:23
minter	Where would it be?	18:23
cbreak	it should be inside .git/objects	18:23
minter	K - lemme check.	18:23
	← drizzd left	18:24
cbreak	zxcp: also, can you log in with that ssh key on the server?	18:24
minter	cbreak: So if the object is '440cf08c2207cf3aa1132966fcec10b2ca4a601e', should it live in .git/objects/44/0cf08c2207cf3aa1132966fcec10b2ca4a601e ?	18:25
	I don't see anything similarly-named in .git/objects/44	18:25
cbreak	that's what I think, yes.	18:25
	weird	18:25
	you wouldn't happen to be using that crappy drop box thing, would you?	18:26
minter	I don't think so - just straight git on OS X 10.6	18:26
cbreak	very weird.	18:26
	→ irc2samus joined	18:26
minter	Yeah, I'm stumped as to why it's trying to read this object that I can't find.	18:27
	← rostayob left	18:27
zxcp	cbreak: yes	18:28
	no problems ssh to the server with my ordinary account	18:30
	→ bthesorceror joined	18:30
cbreak	no	18:30
	that's not what I mean	18:30
	with the git account	18:30
	→ nevyn^_ joined	18:31
	→ gitn00b joined	18:31
	q-rban → q0rban	18:32
minter	cbreak: I tried to send an email to the git email list and it hasn't seemed to show up in the archives yet - do you know if there are steps for a first-time poster to use on that list?	18:32
zxcp	cbreak: no, i cant	18:32
	→ smolyn joined	18:32
cbreak	I've never looked at that list	18:32
	zxcp: then something is wrong with your ssh setup, somehow	18:33
	can't help you more	18:33
	get gitolite :)	18:33
	← lightcap left	18:33
zxcp	"Installing the gitosis package will create the user gitosis and the home directory for it at /srv/gitosis, which contains the repositories and configuration information" from the guide im follow	18:33
PerlJam	zxcp: don't use gitosis, use gitolite	18:34
zxcp	but i cant see the user gitosis	18:34
cbreak	gitosis is ancient and as far as I heard unmaintained since ages	18:34
zxcp	alright i'll try gitolite	18:35
	thank you :)	18:35
	→ btanaka joined	18:37
	→ hacim joined	18:38
hacim	i see this in my logs a lot: git-daemon[16931]: The remote end hung up unexpectedly	18:38
	what causes that?	18:38
cbreak	hacim: all kinds of things	18:39
	network cable unplugged for example	18:39
	→ soc42 joined	18:39
	→ morganick joined	18:39
hacim	cbreak: ok, lets assume the nic is stable on the machine :)	18:41
	→ drake01 joined	18:41
	← Xilent left	18:41
hacim	because I know it is	18:41
cbreak	mitm attack	18:42
	→ ctshryock joined	18:42
	→ dv310p3r joined	18:43
patrikf	connection loss on the client and ^C probably have the same effect	18:43
	← ctshryock left	18:43
	→ zivester__ joined	18:44
cbreak	or the server just terminated for some reason like lack of permission/memory/motivation/...	18:44
	→ Slush- joined	18:44
	← Maletor left	18:44
patrikf	why would it log that the remote end closed the connection then?	18:45
	doesn't sound very plausible.	18:46
	← rndbit` left	18:46
Granjow	zxcp: gitolite is running before you even found documentation for gitosis. As I saw today.	18:47
	← shiba_yu36 left	18:47
	→ Maletor joined	18:47
	← bthesorceror left	18:47
	→ __name__ joined	18:47
	← __name__ left	18:47
	→ __name__ joined	18:47
	→ LongBeach joined	18:48
	→ g3d joined	18:49
	← g3d left	18:49
	→ g3d joined	18:49
g3d	hi, when I invoke git show --pretty=format:%b I'm not only getting message body, but diff as well. How can I supress diff output?	18:50
	→ rndbit joined	18:50
	← erlnoob left	18:51
	← minter left	18:52
albel727	wow. that must be a bug, I guess.	18:53
	you can try git log -1 --pretty=%b	18:53
	← frederick left	18:53
g3d	albel727: yeah... ok thanks	18:54
	→ ymasory joined	18:54
	→ blorbx joined	18:54
	legumbre_ → legumbre	18:56
albel727	g3d: incidentally, --quiet option helped me with git show.	18:57
	it's documented in man git-diff-tree	18:57
jast	automatic message the 'git-diff-tree' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-diff-tree.html	18:57
	← alip left	18:58
g3d	albel727: interesting... thanks	18:58
	→ kipras joined	18:58
	→ erlnoob joined	18:58
	→ jjuran joined	18:59
	→ mkanat joined	18:59
	→ alip joined	18:59
	→ gkmngrgn joined	18:59
NooRotic	what's the difference between git stash pop and git stash apply ?	19:02
	prolly should just use the googles.....my bad	19:02
	← SimonNa left	19:02
	→ poseid joined	19:03
thiago_home	NooRotic: pop also removes the entry from the stash list	19:04
	→ Spockz joined	19:04
thiago_home	so you can't apply it again	19:04
	→ yairgo joined	19:04
	→ explodes joined	19:04
	→ Modius joined	19:04
explodes	Is there some way to edit my user image?	19:05
	Someone on our team has a custom one...	19:05
	Mine is auto-generated it looks like.	19:05
	← Granjow left	19:06
	← jjuran left	19:06
NooRotic	thankz thiago	19:06
	makes sense	19:06
	← gitn00b left	19:06
	→ KB1JWQ joined	19:06
	→ ph^ joined	19:07
	← ph^ left	19:07
	→ ph^ joined	19:07
	→ Textmode joined	19:09
	← Modius left	19:09
	→ Modius joined	19:10
	→ NightMonkey joined	19:10
	→ erichammond joined	19:11
NightMonkey	Hi, . Are there any cons to accessing a repo via Local and* via ssh (via a single 'git' user)?	19:11
	KB1JWQ → Corey	19:11
explodes	How do I undo a pull?	19:12
	→ hobodave_ joined	19:12
	nevyn^_ → nevyn^	19:12
explodes	> git pull origin HEAD, meant to do git pull origin develop	19:12
jmd	Exposure, reset	19:12
_rane	git reset --hard @{1}	19:12
	← opalka left	19:12
	← Spockz left	19:12
	→ tore_ joined	19:13
	← hacim left	19:13
	← nevyn^ left	19:13
	← hobodave left	19:14
	← gnufied left	19:15
	→ spaceonline joined	19:15
	← warthog9 left	19:16
	← normanrichards left	19:16
	→ normanrichards joined	19:16
	← btanaka left	19:16
	← ymasory left	19:16
	→ warthog9 joined	19:17
	NightMonkey hears crickets.	19:17
	→ gnufied joined	19:17
	→ berserkr joined	19:17
NightMonkey	Does anyone have a setup with both Local and SSH access, together?	19:17
bremner	NightMonkey: is the repo bare?	19:18
NightMonkey	bremner: Sorry, "bare" in what sense?	19:18
bremner	NightMonkey: !bare	19:18
jast	NightMonkey: automatic message an explanation of bare and non-bare repositories can be found here: http://bare-vs-nonbare.gitrecipes.de/	19:18
NightMonkey	bremner: Ah, thanks. Reading...	19:18
	→ alberto56 joined	19:19
	← osfameron left	19:19
	← stamina left	19:19
	→ sgronblo joined	19:19
	→ osfameron joined	19:19
	→ stamina joined	19:19
	← shurane left	19:20
explodes	explodes: !bare	19:20
jast	explodes: automatic message an explanation of bare and non-bare repositories can be found here: http://bare-vs-nonbare.gitrecipes.de/	19:20
	→ cwillu_at_work joined	19:20
	→ tokkee joined	19:20
	→ Cacheaway joined	19:21
	← warthog9 left	19:21
	← airborn left	19:21
jmd	Why do we talk about non-bare repositories?	19:21
	← _W_ left	19:21
jmd	Shouldn't they be called "clothed"?	19:21
	← n1xNc0d3 left	19:22
	→ airborn joined	19:22
	→ _W_ joined	19:22
NightMonkey	Ah, I'm gonna have to look up what "git init --shared" is, now, too.	19:23
	!shared	19:23
cbreak	NightMiles: it's in man git-init	19:23
jast	NightMiles: automatic message the 'git-init' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-init.html	19:23
	→ RichardBronosky joined	19:23
cbreak	do NOT confuse it with --shared of git clone	19:23
	→ Slix` joined	19:23
	→ ISF joined	19:24
	→ legumbre_ joined	19:24
NightMonkey	cbreak: Yep, reading the man page now, thanks.	19:24
	→ RaceCondition joined	19:25
	← zwick left	19:25
	→ mhagger joined	19:25
	← legumbre left	19:25
	→ helvete joined	19:27
erlnoob	hi, how do I unset upstream on a branch?	19:28
	← Murr left	19:29
	→ zwick joined	19:31
NightMonkey	bremner: So, we're starting with a bare repo. :)	19:32
	It will be a SSH-created repo ala Pro Git, but a user would like Local access, too.	19:32
	(dedicated git user, git-shell, etc.)	19:33
	→ jackdempsey joined	19:33
bremner	NightMonkey: they have to clone	19:33
	err. wait. What do you mean local access? editing the files on that host?	19:34
	→ caseymcg joined	19:34
NightMonkey	bremner: A user with a shell account would like to locally interact with the repo.	19:34
	bremner: They're beefing that they don't want to do ssh from the same host that the repo lives on...	19:35
bremner	just have them run git clone file:///something	19:35
	and set the permissions appropriately	19:35
	it should be fine.	19:35
patrikf	see the --shared argument to git init	19:36
	← iori left	19:36
	→ rolfb joined	19:36
NightMonkey	bremner: Ah, so I'd have to open the permissions on the "git" user's home directory to allow all access to, say, a "gitdev" group which has the developers in it?	19:38
bremner	correct	19:38
	and presumeably write as well.	19:39
NightMonkey	patrikf: Does --shared just do this permissions stuff for me, or does it do something different to the repo's metadata?	19:39
bremner	but in practice, cloning over ssh and cloning over file are pretty much the same, so you may not want to bother	19:39
NightMonkey	bremner: OK, but in order to allow the file:// protocol to work, I'd need to open the permissions on the repo, right?	19:40
bremner	the developer may have had the idea they could just "cd ~git/repo && vim foo.php" which isn't possible with a bare repo (or with svn for that matter)	19:40
	NightMonkey: correct	19:40
	← poseid left	19:40
	← akosikeno left	19:40
	→ malumalu joined	19:40
jrock20041	Is there a tutorial on how to merge an existing svn repo into an existing git repo?	19:41
	← fedesilva left	19:41
	legumbre_ → legumbre	19:41
	→ nevyn^ joined	19:41
	← gnufied left	19:41
	← anjen left	19:41
NightMonkey	bremner: Thank you!	19:41
	← spaceonline left	19:43
patrikf	NightMonkey: you need to set the permissions properly for ssh, too.	19:44
	→ gnufied joined	19:44
patrikf	unless you're using a special setup such as gitolite.	19:44
NightMonkey	patrikf: SSH is already working. But, do you mean that if I blow open permission for Local to work, SSH+git will not?	19:45
	← drake01 left	19:45
	NightMonkey shudders.	19:45
patrikf	I did not mean that. it can happen if you don't know what you're doing and do it wrong	19:45
	← soc42 left	19:45
	→ jjuran joined	19:46
NightMonkey	patrikf: Like changing the permissions on ~/.ssh/ or such.	19:46
mhagger	jrock20041: are the two repo related? E.g., did the git repo come from an earlier git-svn conversion of the SVN repo?	19:46
NightMonkey	patrikf: (clumsily, that is)	19:46
	→ booginga joined	19:46
patrikf	NightMonkey: do you use gito(lite\|sis) or normal ssh?	19:46
bremner	yeah, so init --shared is crucial there so that git sets the permissions on newly created objects and refs.	19:46
NightMonkey	patrikf: normal ssh. I looked at gitolite, and it looks neat, but this is for a team of two, maybe 3 devs at present.	19:47
patrikf	well if you use normal ssh it's no difference for permissions than local access.	19:47
	unless you've set it up in a weird way	19:47
	i.e. if you have one of them working, the other will, too.	19:47
	→ fedesilva joined	19:48
	← stepnem left	19:48
	→ marc321 joined	19:48
	← gnufied left	19:48
patrikf	but you should've used --shared=group or similar already then, or it wasn't fully working anyway.	19:49
NightMonkey	patrikf: Thanks. I was hoping that there would be some nasty corner case that would be brought up, just because I'd like to keep it simple, one protocol. ;)	19:49
	← jackdempsey left	19:49
NightMonkey	patrikf: But, I guess I have to tell them that "Git is so flexible!" ;)	19:50
patrikf	its fully transparent to the user. the only thing that changes is the URL	19:50
	→ atyz joined	19:50
	← helvete left	19:50
NightMonkey	It would seem, from a workflow perspective, it would be easier for everyone to just ssh. Plus, you could do some nice DNS tricks that make migrations simpler.	19:51
	← Raging_Hog left	19:51
patrikf	sure, it's up to you to make it a policy to use ssh even on the same machine	19:52
NightMonkey	patrikf: Well, maybe later down the road. Anyway, thank you for helping me understand.	19:53
	→ stepnem joined	19:53
	← erlnoob left	19:54
	← OOPMan left	19:54
	← jrock20041 left	19:54
	← marc321 left	19:55
	→ ksawicki joined	19:55
	→ ironfroggy joined	19:56
ironfroggy	can anyone remind me how to update master to origin/master ignoring whatever master currently is?	19:57
_rane	git reset --hard origin/master	19:57
ironfroggy	thanks	19:57
	← mhagger left	19:57
	→ g0bl1n joined	19:58
	→ batrick joined	19:59
	→ Chaser joined	19:59
	→ neunon joined	19:59
	← neunon left	19:59
	→ neunon joined	19:59
	→ jutaro joined	20:00
	← zwick left	20:02
	→ frooh joined	20:07
	← frooh left	20:07
	→ frooh joined	20:07
frooh	what's the best way to find the newest tag contained in a given branch?	20:07
	← jmd left	20:07
	→ versatiletech_ joined	20:08
cbreak	frooh: man git-describe	20:08
jast	frooh: automatic message the 'git-describe' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-describe.html	20:08
frooh	oh awesome	20:08
cbreak	or look at git log --decorate	20:08
	← versatiletech left	20:08
	versatiletech_ → versatiletech	20:08
	→ babilen joined	20:08
	→ OOPMan joined	20:09
	→ Nikoli joined	20:10
frooh	very cool	20:10
	cbreak: thanks :-)	20:10
cbreak	np	20:11
	← p3rror left	20:11
	→ juangiordana joined	20:11
	← juangiordana left	20:12
	← Dave^\| left	20:12
	← ksawicki left	20:12
	← Vile left	20:13
	→ juangiordana joined	20:13
	→ Vile joined	20:14
	← frooh left	20:15
	← g0bl1n left	20:16
	← jjuran left	20:16
	→ Pistos joined	20:18
Pistos	How can I dump what a file looked like at a given SHA1?	20:18
	Besides git co && cat file	20:18
cbreak	Pistos: man git-show	20:20
jast	Pistos: automatic message the 'git-show' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-show.html	20:20
Pistos	Oh here: git show SHA1:file	20:20
cbreak	Pistos: man git-cat-file -p	20:20
jast	Pistos: automatic message the 'git-cat-file' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-cat-file.html	20:20
patrikf	git show with the <treeish>:<path> syntax	20:20
Pistos	Thanks, guys.	20:20
	Just needed to websearch just a little bit beyond what I did before coming here.	20:21
	→ lanthan joined	20:21
	← Slix` left	20:22
	← veronatii left	20:23
	→ p3rror joined	20:24
	← nevyn^ left	20:25
	→ tvw joined	20:26
	→ Spockz joined	20:26
	← mando left	20:27
	← mikepack left	20:29
	→ Zenopus_ joined	20:29
	→ spaceonline joined	20:30
	← RobertLaptop left	20:30
	← sonnym left	20:31
	← Zenopus left	20:31
	Zenopus_ → Zenopus	20:31
	← mastroDani left	20:31
	→ nevyn^ joined	20:32
	→ havenn joined	20:32
	← explodes left	20:33
	← Spockz left	20:33
	← dannas left	20:33
	→ vulpy_ joined	20:35
	→ ben225 joined	20:36
	→ fcuk112 joined	20:36
	→ mklappstuhl joined	20:37
	→ mastroDani joined	20:38
	← zxcp left	20:42
	→ squiggy joined	20:42
	→ RobertLaptop joined	20:43
	→ agitgit joined	20:43
	← nevyn^ left	20:43
	← agitgit left	20:44
	→ workmad3 joined	20:44
	→ tewecske joined	20:44
	← spaceonline left	20:44
	→ lmarcetic joined	20:44
	→ muddo joined	20:45
	hobodave_ → hobodave	20:45
lmarcetic	hi. i compiled a file, and then commit the thing. how do i permanently remove the binary from the repo?	20:45
	← Pistos left	20:45
lmarcetic	i'm the ultimate noob btw	20:45
_rane	sounds accurate	20:46
lmarcetic	all the more reason i need help, don't you think?	20:46
	→ dbgster joined	20:46
_rane	you've not provided enough information on what you have done and want to do	20:47
	← mcluskydodallas left	20:48
	← morganick left	20:48
	→ mcluskydodallas joined	20:49
lmarcetic	compiled the thing, git add *, git commit -m 'fixed indents', git remote add origin blah, git push origin master, AND THEN: rm a.out. now i want it to disappear elsewhere (as if it never was commited if possible)	20:49
	→ jankoprowski joined	20:50
	← alex__c2022 left	20:51
jankoprowski	Hi	20:51
	→ LekeFly joined	20:51
	→ lmarcetic_ joined	20:51
jankoprowski	What is the difference (when I'm on branch 'experimental') between 'git merge master' and 'git rebase master'	20:51
	← lmarcetic_ left	20:51
	← lmarcetic left	20:51
cbreak	lmarcetic: git commit --amend after staging the removal	20:51
	then force push	20:51
	be sure to understand the meaning of history rewriting	20:52
	→ lmarcetic joined	20:52
lmarcetic	where was I?	20:52
	→ yarvin joined	20:52
lmarcetic	remote add origin thing, disregard that, that failed :-)	20:52
	so, how do I uncommit or at least remove a.out now?	20:52
	→ alex__c2022 joined	20:53
	← fakewaffle left	20:53
jankoprowski	lmarcetic: git revert	20:53
	← drubies left	20:54
lmarcetic	jankoprowski: git revert what?	20:55
	← jsick3 left	20:55
charon	lmarcetic: revert nothing. do what cbreak said	20:55
jankoprowski	lmarcetic: commit You want to revert	20:55
	You mast pass SHA1 of this commit as parameter	20:55
lmarcetic	i didn't hear what cbreak said	20:55
	oh	20:56
charon	lmarcetic: including (very important!) understanding the meaning of history rewriting	20:56
cbreak	lmarcetic: git commit --amend after staging the removal	20:56
	then force push	20:56
	be sure to understand the meaning of history rewriting	20:56
	← neunon left	20:56
	→ lightcap joined	20:56
jankoprowski	What is the difference (when I'm on branch 'experimental') between 'git merge master' and 'git rebase master'?	20:56
charon	jankoprowski: abstractly speaking, if B is $(git merge-base master HEAD) then merge combines the changes sets 'git diff B..master' and 'git diff B..HEAD' and builds this combined set on top of B, then calls it a merge commit	20:56
	← bartek left	20:56
cbreak	jankoprowski: http://eagain.net/articles/git-for-computer-scientists/	20:57
charon	jankoprowski: while rebase rebuilds the entire chain of individual changes "done by" the commits B..HEAD on top of curent master	20:57
	← versatiletech left	20:58
jankoprowski	charon: I get it - in rebase I modify existing commits, in merge I do not.	20:58
	→ versatiletech joined	20:58
	← ironfroggy left	20:58
lmarcetic	cbreak: i still see it in github history :-(	20:59
jankoprowski	is operation 'git push' block remote repository?	20:59
charon	jankoprowski: well, except as the author of the eagain article eloquently said: you cannot modify, you can only rewrite and forget	20:59
lmarcetic	also jankoprowski	20:59
	← replore_ left	20:59
	→ nevyn^ joined	20:59
cbreak	lmarcetic: was the force push successfull?	20:59
	(compare hash of git log with the hash you see there)	20:59
charon	jankoprowski: so rebase rewrites. if you also forget depends on whether you have more refs pointing at commits in the current branch	20:59
lmarcetic	cbreak: yes	20:59
jankoprowski	ok	21:00
	charon: is 'git push' somehow block remote repository?	21:00
cbreak	lmarcetic: then you don't see it anymore	21:00
	→ Anubisss joined	21:00
	← Ownatik left	21:01
	← rolfb left	21:01
charon	jankoprowski: huh?	21:01
	← ap3mantus left	21:01
lmarcetic	cbreak: got ti right here, can even download.	21:01
cbreak	no	21:01
lmarcetic	cbreak: i didn't put in --force first though	21:01
cbreak	you just said that the hash is the new one	21:01
	so the old version of the history is not there anymore	21:02
jankoprowski	charon: If two 'git push' commands will be run in the same time - are they queued or made in the same time?	21:02
lmarcetic	cbreak: revert 'blah' is in the history as the third commit. the second commit, however, still has a.out	21:02
	← hobodave left	21:02
	→ hobodave joined	21:03
cbreak	lmarcetic: why the hell did you do a revert?	21:03
	that was dumb	21:03
	→ brainbox1100_ joined	21:03
	← vulpy_ left	21:03
	← brainbox1100_ left	21:03
cbreak	you'll have to rewrite the history with git rebase -i now	21:03
lmarcetic	fatal: Needed a single revision	21:03
cbreak	yes...	21:04
	→ ap3mantus joined	21:04
cbreak	read the man page	21:04
lmarcetic	Invalid base?	21:04
cbreak	lazy users...	21:04
lmarcetic	new users	21:04
cbreak	you'll kill your repository if you use rebase without reading the man page	21:04
	it's a very sharp knife in git's arsenal	21:04
	(unless you also read git reflog :)	21:04
	← nazgul101 left	21:04
charon	jankoprowski: it's atomic per ref iirc	21:05
lmarcetic	this man page is huge	21:06
	← ap3mantus left	21:06
lmarcetic	i'm just gonna remove my repo via some web interface and start anew	21:06
jankoprowski	charon: so hmm ... I'm just thinking to merge incoming revision to another branch in hook.	21:06
	charon: And I'm guessing is it possible or not.	21:07
charon	lmarcetic: you could start by not listening to two people's contradictory advice at the same time [scnr]	21:07
	← chittoor left	21:07
charon	jankoprowski: possible, but usually not desirable because a merge may conflict and require manual intervention	21:08
	← Jippi left	21:08
jankoprowski	charon: That is what I exactly need.	21:08
	charon: Check is merge will work - in dry mode.	21:08
	charon: I do not know is there something like dry merge in git :	21:09
charon	no. try merging; if it conflicts, back out	21:09
	← german__ left	21:09
jankoprowski	charon: Can I do this on the same bare repositry where push coming?	21:09
	← urkud_ left	21:09
jankoprowski	charon: I want to do this while push comming.	21:10
charon	jankoprowski: you cannot merge at all in a bare repo	21:10
jankoprowski	charon: wow - I did not know :	21:10
	charon double checks	21:10
lmarcetic	charon: ok i just recommited all of history -_- please tell me how to unmake a dir a repo	21:11
	← brian_g left	21:11
	→ sonnym joined	21:11
cbreak	jankoprowski: merging requires a working directory to resolve conflicts	21:11
	← tewecske left	21:12
cbreak	no working directory equals no merging	21:12
charon	jankoprowski: indeed, needs a worktree to even run	21:12
	→ ksawicki joined	21:12
cbreak	lmarcetic: unmake?	21:12
charon	(in theory it could run with just an index, but nobody cared enough to make that possible)	21:12
jankoprowski	ok - so while push i need some extra working directory to merge just after push?	21:12
cbreak	lmarcetic: git does not track directories	21:12
	jankoprowski: no	21:13
	← tore_ left	21:13
cbreak	jankoprowski: pushing does never require a merge	21:13
charon	lmarcetic: if you mean "how do i forget all git metadata about a repository", then the dangerous answer is 'rm -rf .git'	21:13
lmarcetic	charon: and there's no safe answer	21:13
	?	21:13
charon	mv .git /some/place/save/dont_lose_this.git	21:13
	argh, misspelled "safe".	21:13
lmarcetic	XD	21:13
	← HG`` left	21:14
	→ urkud joined	21:15
	← LekeFly left	21:15
	→ lastk joined	21:15
charon	jankoprowski: i think you have an XY problem by the way. what are you really trying to solve?	21:15
lmarcetic	yay! clean history	21:16
jankoprowski	charon: I try to merge things send to review when they pass	21:16
lmarcetic	charon: thanks	21:16
	← lmarcetic left	21:16
jankoprowski	And when the merge fail - ask developer to rebase their branch and push again but this time - merge automatically	21:17
	← jutaro left	21:17
jankoprowski	idea is there is 'master' is read-only, and only review system can merge to it	21:17
	and only branches applied by reviewers will be merged (by review system) to master.	21:18
	← ksawicki left	21:18
	← jstemmer left	21:18
charon	jankoprowski: sounds like you are reinventing gerrit. but yeah, merge in a temporary tree, and if it conflicts just back out.	21:18
jankoprowski	charon: I implementing gerrit functionality on top of reviewboard	21:19
	→ brian_g joined	21:19
	→ bbrandon joined	21:21
	→ cesario_ joined	21:22
	← ISF left	21:22
	← Ryback_ left	21:22
	← legumbre left	21:23
jankoprowski	charon: I love ReviewBoard but this one of two features which is missing	21:23
	→ legumbre joined	21:23
	→ v4n_ joined	21:23
	← cesario left	21:24
	cesario_ → cesario	21:24
cbreak	jankoprowski: just let the revierwer merge in their own clones	21:24
	that's faster and easier anyway	21:24
	→ linusoleander_ joined	21:24
	→ LekeFly joined	21:25
	← v4n_ left	21:26
	← yarvin left	21:26
	→ Murr joined	21:27
	← RaceCondition left	21:27
	← abhishek_ left	21:27
	→ SethRobertson joined	21:27
jankoprowski	cbreak: ok	21:28
	← erichammond left	21:28
jankoprowski	cbreak: I will use Your advice - but here is the thing	21:28
	If ReviewBoard can not merge	21:29
	and send e-mail to developer asking about rebase	21:29
	I want to "short" time between rebase and merge	21:29
	Here is the thing	21:29
	While "push", this after rebasing	21:29
cbreak	time between rebase and merge?	21:29
	either rebase or merge	21:29
	← LongBeach left	21:30
jankoprowski	cbreak: developer can not merge to 'master' because 'master' is read-only	21:30
	← cannonball left	21:30
jankoprowski	they can rebase their own branch and send one more time to review.	21:30
cbreak	they have their own master	21:30
jankoprowski	They can not push their own master	21:30
	because of gitolite	21:30
cbreak	they don't have to	21:30
	they can push it to their own branch or own repo	21:30
jankoprowski	cbreak: and this gives him nothing because building machine compile cody only from 'center' repository which is gitolite repo	21:31
	→ ap3mantus joined	21:31
	→ Samual_ joined	21:31
jankoprowski	so	21:32
	they must rebase their own branch	21:32
	and push it again	21:32
cbreak	they are clearly not supposed to use the building machine	21:32
	they don't have to rebase	21:32
	as I said above	21:32
	merge or rebase	21:32
	← g3d left	21:32
cbreak	both just adds the disadvantages of both	21:32
	for relatively small gain	21:32
jankoprowski	cbreak: You don't get it aren't you?	21:32
cbreak	more than you apparently	21:33
jankoprowski	I do not think so.	21:33
	Do You know how I use rebase?	21:33
cbreak	of course...	21:33
jankoprowski	To assure that merge will be fast-forward	21:33
cbreak	you don't need to merge if you use rebase	21:34
jankoprowski	I must fast-forward 'master' branch pointer	21:34
cbreak	because you can rebase your history onto the tip of the branch you want your changes to be in	21:34
jankoprowski	to ocompile this new stuff	21:34
cbreak	after you've done that, there's no need to merge anymore	21:35
	you can just reset your target branch to the new head	21:35
jankoprowski	really?	21:35
cbreak	that works without a working directory	21:35
	→ farktronix_ joined	21:35
cbreak	fast forwarding is NOT merging	21:35
	← tearsong left	21:35
cbreak	fast forwarding is what happens when you push, for example	21:36
	→ btanaka joined	21:36
	← Textmode left	21:36
jankoprowski	But I always use --no-ff	21:36
	even if merge is fast-forward	21:36
cbreak	then you merge.	21:36
	← linusoleander_ left	21:36
jankoprowski	yes	21:36
	← farktronix left	21:36
	farktronix_ → farktronix	21:36
jankoprowski	I need this	21:36
cbreak	and add the disadvantages of merge on top of the disadvantages of rebase	21:36
jankoprowski	greate	21:36
cbreak	at least some of them	21:36
jankoprowski	fast-forward, even with --no-ff - if does not conflict - they still does not conflict	21:37
cbreak	but such a merge is also guaranteed conflict free	21:37
jankoprowski	and this is the point	21:37
cbreak	fast forward and no-ff are mutually exclusive	21:37
jankoprowski	I need to ask about rebase	21:37
	← urkud left	21:38
jankoprowski	to do merge --no-ff without conflicts on reviewboard	21:38
	ok, sorry I must go to sleep now because roommate going sleep	21:38
	thank You for discussion	21:38
	buy	21:38
cbreak	night	21:38
	← jankoprowski left	21:39
	← psoo left	21:39
	q0rban → q-rban	21:39
	← atyz left	21:40
	→ urkud joined	21:40
	← ericbarnes left	21:41
	← PabloM left	21:43
	→ helvete joined	21:43
	← ThomasLocke left	21:44
	← gkmngrgn left	21:44
	← TrineLocke left	21:45
	← btanaka left	21:47
	← [intra]lanman left	21:48
	→ SimonNa joined	21:49
	← Murr left	21:50
	→ versatiletech_ joined	21:51
	← booginga left	21:51
	← pizza_ left	21:52
	→ a\|i joined	21:52
a\|i	how can I update only 1 of the git submodules?	21:52
thiago_home	cd submodule; run command to update	21:53
a\|i	what's the command?	21:53
	git submodule update?	21:53
	→ rzec joined	21:53
rzec	is it possible to use git revert on a specific file instead of reverting every file in the commit?	21:54
	← a\|i left	21:54
	→ [intra]lanman joined	21:55
	← versatiletech left	21:55
	versatiletech_ → versatiletech	21:55
	→ warthog9 joined	21:55
offby1	not directly, but sorta	21:55
cbreak	rzec: man git-checkout, use it in file mode. or git revert -n and then only commit the file you want	21:55
jast	rzec: automatic message the 'git-checkout' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-checkout.html	21:55
	offby1 nods	21:55
	→ jophish joined	21:57
	→ german__ joined	21:57
	→ robgleeson joined	21:57
	→ Murr joined	21:58
	→ saml joined	21:58
	→ BiggFREE joined	22:00
	→ merkosh joined	22:02
	← rzec left	22:02
	→ denysonique joined	22:03
	← denysonique left	22:03
	→ denysonique joined	22:03
	← _psychic_ left	22:04
	← dv310p3r left	22:04
	← banghouse left	22:05
	← versatiletech left	22:05
	← helvete left	22:05
	← cilly left	22:06
	→ EricInBNE joined	22:06
	← Murr left	22:08
	→ Leemp2 joined	22:10
	→ debiminho joined	22:10
	← debiminho left	22:10
	← Leemp left	22:11
	→ gebbione joined	22:11
	← jophish left	22:11
	← Leemp2 left	22:11
	→ Leemp3 joined	22:12
d1zzy	is there a way to safely rename a ref? (I can't just move files around, for example if I have packed-refs enabled)	22:12
	← nullvariable left	22:12
d1zzy	git update-ref seems to only set/remove refs, not rename them	22:12
	← jonshea left	22:12
wereHamster	to rename a ref, create the new one and delete the old one	22:12
	→ Murr joined	22:13
d1zzy	that won't rename it	22:13
thiago_home	if it's a branch, git branch -m	22:13
d1zzy	in the case of annotated tags	22:13
	creating a new one will create a new tag object	22:13
	different SHA1	22:13
thiago_home	you can't rename an annotated tag	22:13
d1zzy	I want to keep the original tag object	22:13
	thiago_home: hmm	22:13
thiago_home	the name is part of the tag	22:13
wereHamster	d1zzy: eh? git update-ref refs/tags/new refs/tags/old; git update-ref -d refs/tags/old	22:13
d1zzy	thiago_home: oh you're right, silly me	22:13
	← dbgster left	22:13
d1zzy	it's n the tag contents	22:13
wereHamster	so you want to rewrite a tag	22:14
	or, amend	22:14
d1zzy	then I guess, related question	22:14
	why is the tag name part of the tag contents?	22:14
wereHamster	because you made it so	22:14
thiago_home	because tags aren't meant to be renamed	22:14
d1zzy	do software use the name inside the contents without the ref?	22:14
thiago_home	tags are final	22:14
	→ bonhoffer joined	22:15
	→ da1l6 joined	22:15
thiago_home	they are done once and done correctly	22:15
d1zzy	what they point to is final and I don't want to modify that	22:15
	← bindaas left	22:15
thiago_home	you don't rename, you don't move tags	22:15
wereHamster	d1zzy: which software?	22:15
d1zzy	and that is the basic feature of a tag in general in SCMs	22:15
	→ dsop joined	22:15
	← [intra]lanman left	22:15
d1zzy	wereHamster: I don't know, any software interfacing with git repos	22:15
wereHamster	d1zzy: the tag message is just like a commit message	22:15
d1zzy	but it seems interesting that tha name is part of the(immutable) contents	22:15
wereHamster	you can put anything in there you want	22:15
d1zzy	it is like if they want to allow softare to find tag names without their refs	22:15
	wereHamster: yes but the tag name is a special header, it's not part of the message	22:16
wereHamster	the tag message doesn't necessarely have to be the version. You can write a normal comment there	22:16
	no, tag name is not a special header	22:16
	→ lekefly_ joined	22:16
d1zzy	thiago_home: what purpose does that serve? (can't rename, move tags)	22:16
	wereHamster: then we can ignore it and support tag renaming :)	22:17
wereHamster	what do you mean by 'tag renaming'? Rename the ref? Or rewrite the tag message?	22:17
thiago_home	because tags are supposed to sign a release cryptographically	22:17
wereHamster	totally different things	22:17
	← LekeFly left	22:18
thiago_home	the SHA-1 of the tag signs the exact tag	22:18
d1zzy	wereHamster: the tag name	22:18
thiago_home	so the name is part of it	22:18
wereHamster	d1zzy: you can do that. just use update-ref to rename it	22:18
d1zzy	thiago_home: wereHamster just said the name is just an arbitrary string in the contents and not special :)	22:18
wereHamster	no, I did not	22:18
	← mklappstuhl left	22:18
d1zzy	(d1zzy) wereHamster: yes but the tag name is a special header, it's not part of the message	22:18
	(wereHamster) no, tag name is not a special header	22:18
	so which is it?	22:19
wereHamster	what do you mean by tag name? refs/tags/foo ?	22:19
cbreak	d1zzy: git cat-file -p it and look yourself	22:19
d1zzy	wereHamster: I mean the "tag <foobar>" line in the git cat-file -p <tag> output	22:19
thiago_home	the name of the tag is a special header	22:19
d1zzy	cbreak: I've been doing that	22:19
wereHamster	d1zzy: oh that, that's part of the tag object, and thus immutable	22:20
	← saml left	22:20
	→ andrewpthorp joined	22:20
	← p3rror left	22:20
d1zzy	yes, I undestand that everything there is immutable, d0h	22:20
	but	22:20
	why was it added there in the first place	22:20
	what purpose does it serve to have the tag name stored in its contents?	22:20
wereHamster	because tags are supposed to be immutable	22:21
d1zzy	when you say that, you mean certain aspects about them are immutable	22:21
	most importandly where they point at	22:21
wereHamster	no, all aspects should be immutable. All of git expects that tags don't change	22:21
d1zzy	why was it decided that the tag name should be immutable?	22:21
wereHamster	if you want a 'movable' tag, use a branch	22:21
	→ ejo joined	22:21
d1zzy	I'm not asking how it is, that is clear to me now, I'm asking why it is so	22:22
	a branch would also allow to update where it points to	22:22
	which is not good	22:22
wereHamster	I'm telling you the why: tags are supposed to be immutable	22:22
d1zzy	yes but when deciding how to design the tags (in git) it was decided to put certain information in it that is immutable	22:22
ejo	Hi. My project structure is like: /virtualenv/MyProj/myproj. I accidentally ended up with .git in both MyProj and myproj dirs. What's the best way to fix the situation and get back to just one .git?	22:22
d1zzy	why some information like the tag name was added while other stuff was not	22:22
wereHamster	which other stuff wasn't added?	22:23
	← lunks left	22:23
d1zzy	say the current branch in the repo when the tag was created (which is irrelevant, but why was it considered tha tthe tag name is relevant?)	22:23
bonhoffer	if there is a branch on github that is a remote branch of mine and i want to look for any recent changes, what is the best workflow?	22:24
d1zzy	wereHamster: anyways, it's a rather pointless discussion, things are as they are, just seem strange a little bit that the name was included :)	22:25
cbreak	d1zzy: the name of a tag has certain relevance	22:25
wereHamster	d1zzy: what would be the usecase where you would want to change the name?	22:25
cbreak	if you sign a tag with your gpg key as v1.0-alpha	22:25
	and some random guy renames it to v1.0-release	22:26
	that changes the appearance of the tag quite a bit	22:26
bonhoffer	i was thinking to pull in the remote and then do a diff	22:26
cbreak	it influences git describe too	22:26
d1zzy	wereHamster: for example if I reorganize things in a project, I move certain tags under "deprecated/" but keep their last component name	22:26
cbreak	bonhoffer: no need to pull	22:26
d1zzy	cbreak: ok	22:27
cbreak	do a git remote update	22:27
	and then git diff/git log at your heats content	22:27
bonhoffer	cbreak: ok, thanks	22:27
	would gitk be one way to see the difference?	22:27
	or mergetool?	22:27
cbreak	no	22:27
	gitk is a tool to see the history	22:27
	maybe it shows the patches of a commit, I don't know	22:27
	git diff is the tool to see differences	22:28
	git log -p can also do it	22:28
bonhoffer	thanks, so i'll focus on the remote update	22:29
	← muddo left	22:29
cbreak	that's probably the least of your problems :)	22:29
	→ diegoviola joined	22:29
cbreak	git remote update will fetch your remote tracking branches	22:29
	← gebbione left	22:29
	← gusnan left	22:30
bonhoffer	how do i remove a remote?	22:31
	→ kelvie_ joined	22:31
patrikf	did you even try to find out?	22:31
cbreak	bonhoffer: man git-remote	22:31
jast	bonhoffer: automatic message the 'git-remote' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-remote.html	22:31
bonhoffer	yes	22:31
	← babilen left	22:32
cbreak	look for remove	22:32
bonhoffer	i googled it, but didn't see a huge harm in asking . . .	22:32
	cbreak: thanks	22:32
	← andrewpthorp left	22:32
	← airborn left	22:32
	← namemon left	22:33
	← d0k left	22:33
	→ budsondixon joined	22:34
bonhoffer	cbreak: easy, thanks	22:34
	← Agos left	22:35
	← eijk left	22:35
bonhoffer	cbreak: do you have a good reference to understand the output of git log -p?	22:35
cbreak	no	22:36
bonhoffer	for example, the difference between +++ and ---	22:36
cbreak	it's more or less normal diff syntax	22:36
FauxFaux	Plusses mean added, minuses mean removed. ¬_¬	22:36
bonhoffer	FauxFaux: yeah, but three, what does that mean?	22:37
	→ mando joined	22:37
FauxFaux	It's suspicious that they only appear on the filename line, isn't it.	22:37
	← v0n left	22:38
bonhoffer	very	22:38
	so should i look for ++ too?	22:38
wereHamster	bonhoffer: http://en.wikipedia.org/wiki/Diff#Unified_format	22:38
bonhoffer	wereHamster: thanks!	22:38
	← threeb left	22:39
	→ nevyn joined	22:40
	← warthog9 left	22:43
	→ mklappstuhl joined	22:44
	← MonsieurLu left	22:44
	← cesario left	22:44
	→ pizza_ joined	22:45
	← envi_laptop left	22:45
irc2samus	guys I was reading this article: http://stevenharman.net/blog/archive/2011/06/09/git-pull-with-automatic-rebase.aspx which claims that merge-commits break git bisect, is that the case? anyone can explain a bit more about it?	22:45
	→ ostolvis joined	22:45
	← LiamH left	22:46
	→ joel_falcou_ joined	22:48
	← joel_falcou left	22:48
	← cyphase left	22:50
	→ cyphase joined	22:51
	→ Epicide joined	22:52
	← kipras left	22:53
	← mkanat left	22:54
	→ ksawicki joined	22:55
	→ keyo joined	22:55
	← marc_kdab_ left	22:55
	← joel_falcou_ left	22:56
	← thiago_home left	22:56
wereHamster	not true	22:57
	→ veronatii joined	22:57
	← ksawicki left	22:57
wereHamster	he doesn't even explain why merges break bisect...	22:57
	→ helvete joined	22:58
	→ muddo joined	23:00
	← nevyn^ left	23:00
	← irc2samus left	23:01
	← RobertLaptop left	23:01
	← f0i left	23:04
	← amerine_ left	23:05
	← hobodave left	23:05
	→ lunks joined	23:06
	← kenichi left	23:07
	← ap3mantus left	23:07
	→ marc_kdab_ joined	23:08
	← Murr left	23:10
	← ExtraSpice left	23:12
	← normanrichards left	23:12
	← Ricki left	23:12
	← german__ left	23:13
	→ N0va` joined	23:13
	← icwiener left	23:14
	← boombatower left	23:15
	→ envi_laptop joined	23:16
	← muddo left	23:17
	← Sonderblade left	23:17
	← notjohn left	23:18
	← radarek left	23:18
	← workmad3 left	23:19
	→ joel_falcou joined	23:21
	← ki__ left	23:22
	→ nevyn^ joined	23:22
	← TeckniX left	23:22
	→ p3rror joined	23:24
	← dan23232 left	23:24
	← Danielpk left	23:24
	→ Murr joined	23:25
	→ german__ joined	23:25
	→ avocado joined	23:26
	→ Slix` joined	23:26
	→ replore joined	23:27
	→ erichammond joined	23:28
	← replore left	23:29
	← Morphous left	23:29
	← malumalu left	23:31
	← Chillance left	23:31
	→ agile joined	23:34
	→ muddo joined	23:36
	← fedesilva left	23:36
	← Murr left	23:37
	← pantsman left	23:38
avocado	i have a remote repo with a post-receive hook that just has "git checkout -f" in it but when i do something like "git push web sillybranch", the hook only checks out master. anyone know where i can read the branch from in that post-receive hook?	23:39
	→ ki__ joined	23:39
	→ stringoO joined	23:40
	← Yuuhi left	23:40
	→ dbgster joined	23:42
	→ ap3mantus joined	23:42
	← tvw left	23:42
	← brian_g left	23:42
keyo	avocado: why not just use the right branch	23:42
	→ brian_g joined	23:43
	→ Murr joined	23:43
	← brian_g left	23:44
	→ brian_g joined	23:44
	→ rendar_ joined	23:45
	← rendar_ left	23:45
	← mando left	23:47
	← Tac left	23:47
	→ Morphous joined	23:47
avocado	i don't know how to do that on bare repos	23:47
	← rendar left	23:48
	→ apocalyptiq joined	23:49
wereHamster	avocado: man githooks	23:50
jast	avocado: automatic message the 'githooks' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/githooks.html	23:50
	→ replore joined	23:51
	← replore left	23:52
	← Murr left	23:55
	→ yota_ joined	23:55
	← yota_ left	23:56
	→ amouge joined	23:57
	→ boombatower joined	23:57
amouge	hi all.. Trying to use git on my centOS 5.5 server and getting this error: git: error while loading shared libraries: libiconv.so.2: cannot open shared object file: No such file or directory	23:57
	→ aidalgol joined	23:57
	← N0va` left	23:57
amouge	libiconv is installed	23:57
	can anyone help me diagnose this problem?	23:57
wereHamster	amouge: how did you install git?	23:58
amouge	wereHamster: i believe from source	23:58
	i cant remember now :/	23:58
	its been a while	23:58
	← ShooterMG left	23:58
wereHamster	compile it again	23:58
	→ bpalmer joined	23:59

Logs Search ←Prev date Next date→

Channels Documentation