IRCloggy #git 2012-04-11

Provider of IRC logs since 2005.

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

	2012-04-11
	→ brentdanley_ joined	00:00
milki	sounds like a copy/paste issue	00:00
Evanlec	Kre10s_: yea, you've got linebreaks in your key	00:00
Kre10s_	I used 'ssh-keygen -t rsa' to come up with this file.	00:00
	← koo3 left	00:00
Evanlec	Kre10s_: pastebin git.pub for us	00:00
	← brentdanley_ left	00:00
Kre10s_	oh... this is the private key.	00:01
Evanlec	;o	00:01
milki	but it says .pub	00:01
	!	00:01
	← ShooterMG left	00:01
	← eph3meral left	00:02
	← Drewch left	00:02
	→ dcarper joined	00:02
	→ koo4 joined	00:03
	← koo4 left	00:03
	← estevocastro left	00:03
	← marhaban left	00:04
	→ tabakhase joined	00:05
	→ VAAE8OJ joined	00:05
	← tabakhase left	00:05
	→ tabakhase joined	00:05
	→ fission6 joined	00:06
	← chrissbx left	00:06
offby1	Evanlec: nice try!	00:07
	← dcarper left	00:07
	← adam12 left	00:08
	← thiago left	00:09
	← joaoh82 left	00:09
	← dreiss left	00:10
	→ joaoh82 joined	00:10
Kre10s_	when I do 'dl-setup git.pub' git will be the name of the user with this pubkey... will this be the admin account?	00:11
	← alberto56 left	00:11
milki	yes, as identified by the pubkey	00:11
	but it is not the user you identity as your git remote url	00:11
	← mcluskydodallas left	00:12
Kre10s_	claro. those would be the git users the admin creates	00:12
	← BiggFREE left	00:12
milki	no, the git remote url is always the user gitolite is installed under	00:12
	the pubkeys are for the git users	00:12
	→ mikepack joined	00:13
	→ quazimodo joined	00:15
	← cespare left	00:16
	← brentdanley left	00:16
Kre10s_	ok so now I try to clone the admin repo: git clone gitolite@server:gitolite-admin	00:16
	← jameswarren left	00:17
	← tvw left	00:17
Kre10s_	but I cannot authenticate... the password is that of the linux gitolite user?	00:17
	→ jameswarren joined	00:17
	→ tvw joined	00:17
Kre10s_	or that of the git.pub key i generated?	00:17
	← iocor left	00:17
milki	what do you mean by cannot authenticate	00:18
Kre10s_	it always says wrong password for gitolite	00:18
	← neurodrone left	00:18
Kre10s_	maybe gitolite is not allowed to ssh	00:18
milki	is that the exact error messagE?	00:19
Kre10s_	Permission denied (publickey,password).	00:19
	fatal: The remote end hung up unexpectedly	00:19
	→ brentdanley joined	00:19
milki	ok, that means you arent providing the public key	00:19
	you either need to use .ssh/config or an ssh-agent	00:19
	← ehsan left	00:20
Kre10s_	they key generated to (localy) ~/.ssh/id_dsa.pub is not enough?	00:20
milki	you need to actually supply that to ssh somehow	00:21
	o...	00:21
	well, that should automatically be used then	00:21
	ssh should be asking you to unlock the key then	00:22
Kre10s_	how does ssh know what key to unlock?	00:22
milki	if you have an id_dsa or id_rsa, it will use that automatically if no other key is specified	00:23
	that would be the private key	00:23
Kre10s_	yes I have those in ~/.ssh/	00:23
	i have both id_dsa and id_dsa.pub	00:24
milki	so, is ssh asking you to unlock it then?	00:24
imMute	did you copy the public key to ~/.ssh/authorized_keys on the server?	00:25
	→ koo4 joined	00:25
milki	imMute: using gitolite	00:25
imMute	did you add it to the gitolite config then?	00:26
	← gv left	00:26
Kre10s_	ok. the server was not set up to allow ssh connections from the gitolite user...	00:27
	← pfFred left	00:27
Kre10s_	I have cloned the admin repository now.	00:27
milki	i didnt think ssh would give you that deny message if the user itself was not allowed to do ssh	00:28
	← williamcotton left	00:28
	← SeySayux left	00:29
	→ nikhgupta joined	00:30
nikhgupta	hey guys.. I have a repo in which i want to add a submodule. The problem is that the submodule is not contained in a separate directory	00:30
frogonwheels	nikhgupta: !?	00:31
	← ziggles left	00:31
nikhgupta	lets say if repo exists at: ".", then this submodule has directories: "./dir1" and "./dir2"	00:31
frogonwheels	nikhgupta: is this a new submodule or one that exists somewhere?	00:31
	→ master-lincoln joined	00:31
nikhgupta	the one that exists somewhere	00:31
frogonwheels	nikhgupta: you have to add it as subdir of the parent repo	00:32
nikhgupta	i need to add it to my root directory of the repo	00:32
	→ mcluskydodallas joined	00:32
nikhgupta	hmm	00:32
frogonwheels	nikhgupta: !xy	00:32
gitinfo	nikhgupta: [!UGFWIINI] You appear to be Using Git For What It Is Not Intended - https://www.google.com/search?q=UGFWIINI Let's step back for a minute - What are you actually trying to achieve? Are you sure git is the right tool for the job?	00:32
frogonwheels	hmm..	00:32
	← n1x left	00:32
frogonwheels	that wasn't quite what i meant.	00:33
nikhgupta	hehe.. i understand :)	00:33
frogonwheels	nikhgupta: back-up a step - what are you trying to achieve	00:33
	nikhgupta: describe your scenario. i'll be back in a mo'	00:33
	← gasull left	00:34
nikhgupta	thing is I have this plugin for a cms, kind of.. I made a git repo for the CMS.. Now, I want to add this plugin as a submodule and work on it.. However, the plugin system utilizes two directories which are both contained in the root folder of the cms.	00:34
	lets say, cms is in "./cms", then the plugin would be contained inside directories: "./cms/plugins/name" and "./cms/widgets/name.php"	00:35
frogonwheels	nikhgupta: symlinks ?	00:36
nikhgupta	my bad.. the second one is actually a single file that needs to be included within the plugin (its repo), i guess.. this second file is what is creating problem.	00:36
	yeah, i thought so, but then I don't think it will be distributable.. I would be using symlinks, definitely, if there are no other methods	00:36
frogonwheels	nikhgupta: either that or have a bundler for distribution	00:37
	nikhgupta: it sounds like a crappy design	00:37
nikhgupta	hmm.. bundler.. as in a shell script or something that automates the bundling right? (a git hook maybe?)	00:37
	→ subbyyy joined	00:37
	→ Fandekasp joined	00:38
nikhgupta	well, yes.. they just switched to a better version, but the earlier version on which this plugin of mine is based, has this structure.	00:38
	so i was sayin.. bundlers are like git hooks, if i am right?	00:39
	← yoklov left	00:39
nikhgupta	my bad.. found this: http://progit.org/2010/03/10/bundles.html	00:39
	thank you, frogonwheels, for directing me to bundles.. reading :)	00:40
	← nikhgupta left	00:43
	→ WillMarshall joined	00:44
	← delinquentme left	00:44
	→ barrbrain joined	00:44
	→ thiago joined	00:45
	← Zeelot left	00:45
	→ replore joined	00:47
Kre10s_	Ok. so I can now edit conf/gitolite.conf ... I see that there is already a repository testing with RW+ all...	00:49
	I cloned this repo, and added a README. did a git add, git commit, and tried a git push.	00:50
	I get 'No refs in common and none specified; doing nothing.'	00:50
	and 'Perhaps you should specify a branch such as 'master'.' .. so. what now.	00:51
	→ hobbyist joined	00:52
	→ charlenopires joined	00:52
	→ diegoviola joined	00:53
	→ neurodrone joined	00:53
milki	listen to it and specify a branch?	00:53
	← ipalaus_ left	00:53
milki	new repos dont have any branches	00:54
Kre10s_	seems i need to do git push origin master ... I'm used to the shorthand git push. what do i need to change to get that back?	00:54
	← charlenopires left	00:54
Kre10s_	oh. duh.	00:54
	the shorthand should work once there is one...	00:54
milki	yup	00:54
	← duckxx left	00:54
	← brngardner left	00:55
	→ Zeelot joined	00:55
	← timoxley left	00:56
	← hobbyist left	00:56
	→ blackmatrix_ny joined	00:56
	→ ipalaus_ joined	00:57
	← glennpratt left	00:58
	→ yoklov joined	00:58
	← metcalfc left	00:58
	→ gavin_huang joined	00:58
Kre10s_	ok. so now to add the git repositories to gitolite all a need to do is move them into the repositories/<reponame>.git and add an entry in gitolite-admin/conf/blabla commit push and try it?	00:59
	← bartman left	00:59
milki	its in the docs	01:01
	sounds about right	01:01
	except for blabla	01:02
	← JackWinter left	01:02
Kre10s_	yea...	01:02
	→ glennpratt joined	01:03
	→ alberto56 joined	01:03
	← SegFaultAX\|work left	01:04
	← AaronMT left	01:05
	→ bartman joined	01:06
	→ timoxley joined	01:06
	→ Vampire0_ joined	01:08
	→ neiraza joined	01:09
	← __class__ left	01:11
	← Vampire0 left	01:11
	→ n1x joined	01:11
	→ __class__ joined	01:12
	→ hobbyist joined	01:15
	→ xiangfu joined	01:19
	← glennpratt left	01:20
	← ColKurtz left	01:22
	← usmanity left	01:22
	← mikepack left	01:23
	→ ehsan joined	01:23
	← Bombe left	01:23
	→ duko_ joined	01:24
	← duko_ left	01:25
	← Vinnie_win left	01:25
	← gusnan left	01:25
	← kadoban left	01:26
	→ cordoval joined	01:26
cordoval	i am getting this while adding a submodule https://gist.github.com/6c06467a442741524918	01:26
	is it not from the top directory that i add the submodules?	01:27
	or should I cd to the low level directory and add it there?	01:27
	confused	01:27
	→ RainbowTables joined	01:27
	← frankaa112 left	01:28
	→ DofLeFou joined	01:28
	→ CEnnis91 joined	01:31
	→ NetRoY joined	01:31
	→ Bombe joined	01:31
SethRobertson	Sounds like you should remove the path from your .gitignore	01:31
	→ mikepack joined	01:32
SethRobertson	You can even do it temporarily. Once it is committed it can be ignored to your heart's content (but the ignoral will be ignored)	01:32
milki	lol	01:32
	← daaku left	01:32
	→ brngardner joined	01:33
	→ brentdanley_ joined	01:34
cordoval	or I could just -f	01:34
	force it	01:34
	i did that	01:34
	now it all works :D	01:34
	→ mgpcoe joined	01:34
cordoval	thanks guys	01:34
	SethRobertson: take it away on private if you have written some more on that cherry pick flow :D	01:35
	← cordoval left	01:35
SethRobertson	take it away on private?	01:35
	→ cordoval joined	01:35
	← el2ro left	01:35
SethRobertson	take it away on private? Anyway, I don't recommend the cherry-pick flow without some kind of support tool that AFAIK has not been written. You can always hire me to write it...	01:36
	dermoth_ → dermoth\|home	01:36
	← Nightfire034 left	01:37
	← mythmon left	01:38
	→ Vinnie_win joined	01:38
	← Bombe left	01:38
	→ SeySayux joined	01:39
	→ sj` joined	01:39
	← Dulak left	01:39
	← joaoh82 left	01:39
	← n1x left	01:39
milki	lol	01:39
	← alberto56 left	01:39
	← mrwalker left	01:40
	→ cooloney joined	01:40
	← calvados left	01:41
	← pragma` left	01:41
	→ pragma` joined	01:41
	← cordoval left	01:43
	← apok left	01:43
	← xiangfu left	01:44
	→ Bombe joined	01:44
	→ gchristensen joined	01:44
	→ mxweas_ joined	01:45
	→ recur joined	01:45
	→ joshsmith joined	01:46
	← Bombe left	01:49
	← Araxia left	01:49
	pragma` → pragma-	01:51
	→ cordoval joined	01:52
	→ el2ro joined	01:52
cordoval	can I overlap submodule directories?	01:52
	so the problem is not solved with -f basically I am telling the submodule to be added on the same folder of other submodule but that is not good	01:53
	how it is handled?	01:53
SethRobertson	You can create a shadow tree of symlinks pointing to the right place	01:54
	You could try to see if !vcsh could be convinced to work with submodules	01:55
gitinfo	https://github.com/RichiH/vcsh -- version control shell - manage dotfiles using git's fake bare repos to put more than one working directory into $HOME	01:55
SethRobertson	You could rearchitect your layout so that one can be a subdirectory or a peer of the other	01:55
	→ igotnolegs joined	01:55
cordoval	since i am using namespaces on the code I wonder if it is easier to put it into a nother folder	01:56
	yeah that is better i think i hope i dont cause more problems	01:56
	you rock SR	01:56
	→ zdk joined	02:00
	← whyking_ left	02:01
	→ iamjarvo1 joined	02:01
	← iamjarvo left	02:01
	← mxweas_ left	02:02
	→ weiyang joined	02:03
	→ dcarper joined	02:03
	← dcarper left	02:03
	→ dcarper joined	02:04
	← gchristensen left	02:04
	← TomyLobo left	02:05
	← master-lincoln left	02:05
	→ JackWinter joined	02:07
	← cordoval left	02:08
	→ anjibman joined	02:09
	→ BiggFREE joined	02:09
	gitinfo set mode: +v	02:09
anjibman	how to check the difference between the local and remote repo?	02:10
milki	git log local_branch..remote/remote_branch	02:10
cehteh	fetch first	02:11
milki	after fetching remote	02:11
	→ gchristensen joined	02:11
	← recur left	02:11
	← iamjarvo1 left	02:11
	← gavin_huang left	02:11
	→ arvind_k joined	02:12
anjibman	with git fetch origin/master I am getting 'origin/master' does not appear to be a git repository error msg	02:13
	← neersighted left	02:13
	→ xiangfu joined	02:13
	→ iamjarvo joined	02:14
offby1	anjibman: try "git fetch origin master"	02:14
	i.e., lose the /	02:14
anjibman	but I can see see log with git log master origin/master	02:14
	okie origin master works	02:15
	→ gavin_huang joined	02:15
anjibman	so i can see that fetch doesn't merge the updated content from the remote repo ... so what is the use of this fetch command then/	02:16
	← orafu left	02:16
	→ orafu joined	02:16
	→ neersighted joined	02:16
SethRobertson	You have to get the changes before you can merge them anywhere	02:17
	BTW, don'	02:17
	BTW, don't use the four argument version of fetch or pull. It will only end in tears. `git fetch origin` `git pull origin`	02:17
	→ mattsa joined	02:18
	← dcarper left	02:18
mattsa	github always creates it's gh-pages branch as a "new branch" that is - it has no previous commits (like it branches from a new repository). how is that accomplished?	02:18
	← neersighted left	02:18
	tizzo → tizzo-afk	02:18
SethRobertson	git checkout --orphan newbranch	02:19
	→ neersighted joined	02:19
mattsa	SethRobertson: thanks a bunch!	02:20
	→ master-lincoln joined	02:20
EugeneKay	Or !orphan, if your git is too old to have that	02:20
gitinfo	To create an orphaned branch(unrelated to existing history): git symbolic-ref HEAD refs/heads/orphan; rm .git/index; git clean -fdx; touch foo; git add foo; git commit -m 'Initial commit on branch orphan'	02:20
	← neersighted left	02:20
	← segher left	02:20
	← mikepack left	02:20
	→ neersighted joined	02:20
	→ aknewhope joined	02:21
	→ segher joined	02:22
	← neersighted left	02:22
	← mtx_init left	02:22
	→ DrDogwelder joined	02:22
	→ Bombe joined	02:23
	→ ChaoWang joined	02:24
	→ neersighted joined	02:25
	← anjibman left	02:27
	→ mtx_init joined	02:27
	→ rgwu joined	02:29
	→ apok joined	02:29
	← asakura left	02:29
	→ the_cybersphinx joined	02:30
	← ben225 left	02:31
	← thelazydeveloper left	02:31
	← cybersphinx left	02:33
	→ LucasTizma joined	02:33
	← aknewhope left	02:33
	→ limoges_ joined	02:34
	→ erujolc joined	02:34
	← nif left	02:35
	→ mxweas_ joined	02:35
	→ zendeavr joined	02:35
	→ intripoon joined	02:37
	→ asakura joined	02:37
	← infogulch_ left	02:39
	← intripoon_ left	02:40
	→ CannedCorn joined	02:40
	← erujolc left	02:41
	← BiggFREE left	02:41
	← w0ng left	02:42
	← mritz left	02:42
	→ w0ng joined	02:42
	← chronos left	02:43
	→ glennpratt joined	02:43
	→ kenperkins joined	02:44
	→ Davey joined	02:44
	← milk left	02:45
	← jameswarren left	02:45
	← replore left	02:45
	→ nif joined	02:45
	→ BiggFREE joined	02:45
	← diegoviola left	02:46
	→ milky joined	02:46
	← cannonball left	02:47
	→ daaku joined	02:47
	← brentdanley_ left	02:49
	← Rafusek left	02:50
	← Chibby_ left	02:50
	← Dreamer3 left	02:50
	→ Rafusek joined	02:51
	→ Dreamer3 joined	02:51
	← NetRoY left	02:52
	← arvind_k left	02:52
	← davidcoallier left	02:53
	← daethorian left	02:53
	← pppaul left	02:54
	→ Yuuhi` joined	02:54
	→ pppaul joined	02:54
	→ daethorian joined	02:54
	← smw left	02:55
	← Yuuhi left	02:56
	→ smw joined	02:56
	→ davidcoallier joined	02:56
	← acl_ left	02:58
	→ ben225 joined	03:01
bremner	commit A is reachable from HEAD, but not upstream. commit B is reachable from both. Commit A and B have the same patch id. Should git-cherry upstream HEAD report + A ?	03:01
	← eletuchy left	03:02
	← mattsa left	03:04
	→ mister_m joined	03:08
	← yoklov left	03:09
	← glennpratt left	03:11
	← mcluskydodallas left	03:11
	← DrDogwelder left	03:11
	→ curtp joined	03:11
	← cccaldas left	03:12
	→ mritz joined	03:13
	→ Gentlecat joined	03:13
	→ curtp_ joined	03:15
	← brngardner left	03:15
	→ aknewhope joined	03:15
	← wbednarski left	03:15
	← curtp_ left	03:16
	← curtp left	03:16
	→ fayaz joined	03:16
	→ kerrick joined	03:16
	← Fandekasp left	03:16
	→ traviscj joined	03:16
	→ Fandekasp joined	03:18
	→ ppppaul joined	03:24
	← krz left	03:25
	← Targen left	03:30
	→ jameswarren joined	03:31
	← akiress left	03:31
	→ akiress joined	03:32
	→ duckx joined	03:32
	→ n2diy joined	03:32
	← __class__ left	03:33
	→ __class__ joined	03:34
	← __class__ left	03:37
	→ brngardner joined	03:38
	→ __class__ joined	03:38
	← mxweas_ left	03:40
	← Fandekasp left	03:42
	← rgwu left	03:46
	← neiraza left	03:49
	→ jfelchner joined	03:50
	← pppaul left	03:50
	→ pppaul joined	03:51
	← RainbowTables left	03:52
	← timoxley left	03:53
	← nif left	03:54
	← kerrick left	03:56
	→ Fandekasp joined	03:57
	← __class__ left	03:58
	→ mo0nykit joined	03:59
	← brngardner left	03:59
	← aknewhope left	03:59
	← tharkun left	04:00
	→ __class__ joined	04:00
	→ aknewhope joined	04:00
	← Kre10s_ left	04:00
mo0nykit	Hi, suppose I have the following commits from earliest to latest: ABCDEFG. HEAD is currently on G. I want to squash the middle commits so that it will now look like AZG, where Z is BCDEF squashed together. How do I do that?	04:01
	→ ziggles joined	04:03
	← BiggFREE left	04:03
	← n2diy left	04:04
	→ hagabaka joined	04:04
	→ n2diy joined	04:05
	→ rurufufuss joined	04:07
	→ mxweas_ joined	04:08
	→ ferdna joined	04:08
offby1	mo0nykit: "git rebase -i A"	04:09
	bremner: I don't think so.	04:10
	← n2diy left	04:10
	→ satyanash joined	04:10
kini	offby1: won't that give him AZ, where Z is BCDEFG squashed together?	04:10
	or her, dunno	04:11
	← igotnolegs left	04:12
	← milli left	04:12
	← spaceonline left	04:12
	→ tharkun joined	04:12
mo0nykit	The only solution I could imagine is to branch off G, then get back to the original, reset to F, then rebase -i A	04:12
offby1	kini: depends on how he/she/it edits the list in the editor	04:12
kini	oh, true	04:13
	→ n2diy joined	04:13
kini	interactive rebase is nice	04:13
	→ milli joined	04:13
mo0nykit	offby1: Oh I got it. I'd just pick both F and G, then fixup the rest	04:13
	Thanks	04:13
offby1	\o/	04:14
	← mo0nykit left	04:15
	→ mo0nykit joined	04:16
	← BSaboia left	04:17
	← lyddonb left	04:17
	→ yoklov joined	04:18
	→ dcarper joined	04:18
	← CannedCorn left	04:19
	← ricky left	04:22
	← dcarper left	04:23
	→ mcluskydodallas joined	04:23
	← coyotebush left	04:25
	← hyperair left	04:26
	← gchristensen left	04:26
	→ gchristensen joined	04:26
	← aknewhope left	04:30
	← kukks left	04:30
	→ Targen joined	04:30
	→ sangi joined	04:31
	← kenperkins left	04:34
	← maqr left	04:35
	→ cjs joined	04:36
cjs	When I do a git log --status somefile, the status it shows me for each commit is only the changes to somefile, even if other files changed as well. How do I see all the files that changed for each commit touching somefile?	04:36
	← ben225 left	04:39
	← Bombe left	04:39
	→ beautifulmind joined	04:40
wereHamster	cjs: --full-diff	04:41
	→ igotnolegs joined	04:42
	← lonewulf` left	04:43
	→ maqr joined	04:44
	→ Bombe joined	04:44
	→ dr0id joined	04:45
	→ lonewulf` joined	04:45
cjs	Thanks.	04:45
	← Slix` left	04:46
	→ maleknet joined	04:47
	→ timoxley joined	04:47
	← mritz left	04:49
	→ k-man joined	04:49
k-man	is there a way to automaticaly git rm any files that are no longer on my filesystem (but still in the repo?)	04:49
	← ziggles left	04:50
	→ kerrick joined	04:50
wereHamster	k-man: git ls-files --appropriate-flags \| xargs git rm --cached	04:51
	→ cime joined	04:52
	→ spaceonline joined	04:53
	← maleknet left	04:54
k-man	wereHamster, thanks	04:54
	← stepnem left	04:55
	→ ziggles joined	04:55
k-man	wereHamster, it almost works too, except the files have spaces in the names and xargs can't cope with that	04:56
	→ ben225 joined	04:56
	→ henryC joined	04:56
	← nickcooley left	04:56
wereHamster	k-man: learn about the -0 xargs option	04:56
k-man	wereHamster, yeah i tried that, but afaict, it requires a null at the end of each file name but that does not seem to be what ls-files outputs ?	04:57
wereHamster	and the corresponding option in git ls-files ..	04:57
k-man	maybe i'm missing something	04:57
kini	heh	04:57
	→ nickcooley joined	04:57
k-man	wereHamster, ah - thats what I'm missing	04:57
	gitinfo set mode: +v	04:57
	→ stepnem joined	04:57
k-man	-z	04:57
	thanks	04:57
henryC	I did git add on a directory I meant to ignore. I haven't done git commit yet. Can someone tell me how to not add, and in fact ignore, that directory?	04:57
kini	henryC: git rm --cached dirname	04:58
	for ignoring, you'll need to edit .gitignore in the root of your repo	04:58
wereHamster	so you saw, 'oh, ls-files isn't producing the output I need' and it didn't occur to you that there might be an option for that?	04:58
henryC	kini: thanks	04:58
wereHamster	henryC: git reset <dir>	04:58
	← neersighted left	04:58
	→ joaoh82 joined	04:59
henryC	If I did "git add" for these files which had not been previously committed, and I want to ignore them, "git rm" or "git reset"?	04:59
	→ neersighted joined	04:59
kini	either will work if you haven't committed those files	04:59
wereHamster	try it and see	04:59
henryC	ok thanks	04:59
kini	git rm --cached will cause the next commit to not have those files, even if the current commit does	05:00
	git rm without the "--cached" will furthermore physically delete the files, but I guess you don't want that	05:00
henryC	right	05:00
	← thiago left	05:00
	← slide_ left	05:00
k-man	wereHamster, it did occur to me that it might be there - but it was named differently to what i excpected and i missed it the first time around	05:00
	→ slide_ joined	05:01
kini	k-man: don't worry about it, that's what this channel is for :)	05:01
	reminding people about what was in the man page all along	05:01
	← maqr left	05:02
k-man	is there a way to get git to add all untracked files except for any that should be ignored by .gitignore	05:03
	← limoges_ left	05:03
wereHamster	git add .	05:03
	or if you want untracked, git ls-files --other --exclude-standard	05:04
	it's all right there in the man pages, as kini said :)	05:04
	← Fandekasp left	05:04
	→ maqr joined	05:04
k-man	thanks	05:05
	→ Fandekasp joined	05:06
	→ maleknet joined	05:06
	← Orbitrix left	05:10
	→ shruggar joined	05:11
	← Gentlecat left	05:13
	← ppppaul left	05:13
	← Bass10 left	05:14
	→ _sundar_ joined	05:14
	→ LongBeach joined	05:17
	← lonewulf` left	05:17
	← mischat left	05:18
	← mo0nykit left	05:19
	→ ricky joined	05:20
	← yoklov left	05:21
	→ mo0nykit joined	05:21
	→ lonewulf` joined	05:22
	← master-lincoln left	05:25
	← n2diy left	05:26
	← mister_m left	05:26
	→ n2diy joined	05:27
	← Mowsh left	05:27
	ybit → fiddlybits	05:30
	fiddlybits → ybit	05:31
	← satyanash left	05:32
	← pppaul left	05:32
	← aalex left	05:33
	← subbyyy left	05:33
	← mo0nykit left	05:34
	→ caseymcg joined	05:34
	→ pppaul joined	05:34
	← ziggles left	05:34
	← brokenshakles left	05:34
	→ Dave^\| joined	05:35
	← sangi left	05:35
	→ dreiss joined	05:37
	→ JohnBat26 joined	05:40
	← madewokherd left	05:41
	→ glcrazy joined	05:41
	→ brokenshakles joined	05:41
	→ eletuchy joined	05:43
	← Targen left	05:43
	→ harshpb joined	05:43
	→ mklappstuhl joined	05:44
	→ chronos joined	05:47
	← chronos left	05:48
	→ Advocation joined	05:49
	← shruggar left	05:50
	← maleknet left	05:50
	← pppaul left	05:54
	→ pppaul joined	05:55
	← dnivra left	05:55
	← thursday left	05:55
	→ kadoban joined	05:57
	← lonewulf` left	05:58
	→ mikepack joined	05:58
	→ ajaykools joined	05:59
	← Advocation left	05:59
	→ jchan joined	05:59
	← LongBeach left	06:00
	← jayne left	06:00
	← jchan left	06:00
	→ lonewulf` joined	06:01
	→ flijten joined	06:01
	→ airborn joined	06:01
	→ jayne joined	06:02
	← ph^ left	06:02
	← duckx left	06:02
	← mcluskydodallas left	06:03
	← mikepack left	06:03
	← mklappstuhl left	06:04
	→ khmarbaise joined	06:04
	← dreiss left	06:05
	→ mishok13 joined	06:06
	← MissionCritical left	06:07
	← khmarbaise left	06:07
	← Davey left	06:08
	→ Dulak joined	06:09
	→ ThomasLocke joined	06:10
	→ Fullmoon joined	06:11
	→ maleknet joined	06:11
	← xiangfu left	06:11
	→ xiangfu joined	06:11
	← joaoh82 left	06:12
	→ cbreak-work joined	06:15
	→ arvind_k joined	06:15
	→ archdori joined	06:16
	← igotnolegs left	06:18
	← pppaul left	06:18
	→ pppaul joined	06:19
	→ dcarper joined	06:19
	← Fandekasp left	06:20
	← maleknet left	06:21
	→ Shusshu joined	06:22
	← dcarper left	06:23
	← diminoten left	06:25
	→ dpino joined	06:28
	← Werring left	06:30
	← mishok13 left	06:31
	→ Werring joined	06:31
	← neurodrone left	06:31
	→ Guest7979 joined	06:32
	→ ph^ joined	06:32
	← Guest27784 left	06:34
	→ Guest27784 joined	06:34
	→ MissionCritical joined	06:34
	← kerrick left	06:35
	→ maleknet joined	06:35
	← barrbrain left	06:35
	← DofLeFou left	06:36
	→ giallu joined	06:37
	→ vervic joined	06:38
	← mxweas_ left	06:38
	→ tsdh joined	06:38
	→ mabrand joined	06:39
	→ _tasslehoff_ joined	06:39
_tasslehoff_	Is msysgit on 64 bit windows 7 working (as) well (as can be expected)?	06:40
	← Dave^\| left	06:42
	→ pantsman joined	06:44
	← maleknet left	06:45
EugeneKay	Works fine for me.	06:45
	AFAIK the msysgit binaries are 32-bit only, but W7 doesn't mind. NEver has.	06:46
	← ivan` left	06:46
daaku	is there some built in/library to get a temp directory?	06:46
	→ ivan` joined	06:47
	→ marthinal joined	06:48
	→ mischat joined	06:49
	← airborn left	06:49
	← helgikrs left	06:50
_tasslehoff_	EugeneKay: good. buying a new pc at work, and its the first one we get with 64 bit Win.	06:50
	← WillMarshall left	06:50
	→ Sigma joined	06:50
	→ helgikrs joined	06:51
	→ dreiss joined	06:53
	→ mishok13 joined	06:54
	← vervic left	06:55
	← eletuchy left	06:56
	→ Element9 joined	06:57
	→ maleknet joined	06:58
	→ orkz joined	06:59
	← glcrazy left	07:01
	← vandemar left	07:01
	→ jceb joined	07:02
	→ slide joined	07:02
	→ Sitethief joined	07:03
	→ Dave^\| joined	07:03
	← albel727 left	07:03
	← slide_ left	07:05
	← harish_afk left	07:05
	→ harish_afk joined	07:05
	→ v_ joined	07:05
	← harish_afk left	07:05
	→ harish joined	07:06
	← ferdna left	07:06
	→ rendar joined	07:06
	← dhruvasagar left	07:08
	→ Chib joined	07:09
	→ chadi joined	07:09
	← Lujeni left	07:10
	→ dhruvasagar joined	07:11
	← orkz left	07:11
	→ rtuin joined	07:12
	→ vandemar joined	07:13
	← xiangfu left	07:15
	→ xiangfu joined	07:16
	gitinfo set mode: +v	07:16
tfnico_	I've been pondering on merging latest development in master into feature branches (as asked here http://stackoverflow.com/questions/6837500/in-git-is-it-possible-to-merge-from-the-master-to-a-branch-that-was-originally-c ) - is the answer correct? Are there no drawbacks to this approach?	07:16
	→ psoo joined	07:17
	← rtuin left	07:18
	← Zenith77 left	07:18
EugeneKay	git-merge really doesn't care what was branched when. It only cares about common ancestry and the two sides.	07:18
	It'll do its best, but ultimately it's up to you to sort out merge conflicts and make sure the results are sane.	07:18
	← ipalaus_ left	07:18
	← zeshoem left	07:19
	→ thelazydeveloper joined	07:19
EugeneKay	It usually does a pretty impressive job, so you are encouraged to !justdoit	07:19
gitinfo	If you go to the trouble of asking questions, please go to the trouble of taking the advice offered, Take backups (type "!backup" as a reply to learn more) if necessary -- or in fact, it's a good idea anyway. You can almost always censor log messages if you have to. We are literally helpless unless you tell us what we need or do what we ask. Please, Just Do It™.	07:19
kini	tfnico_: merging master into a feature branch is definitely a normal thing to do	07:20
	the only drawback is that if your feature branch goes on for a very long time without getting merged into master, you're going to have a lot of merge commits pulling master into your feature branch - which personally I don't consider a drawback, but apparently linus doesn't like that when people are making feature branches for the linux kernel	07:20
EugeneKay	The semantics of your history(whether you rebase feature branches, or merge master into tem before submitting for merge back into master, or use space monkeys) is really up to you and your workflow. git doesn't care.	07:20
kini	which is why git rerere exists (but that's advanced stuff so don't worry about it)	07:21
	→ netiru joined	07:21
tfnico_	kini: ok, thanks for that	07:21
	→ zeshoem joined	07:21
	← palyboy left	07:21
kini	EugeneKay: sure, but it's perfectly valid to ask questions in here about common practices	07:21
tfnico_	EugeneKay: arright, just wondering if there were some downsides down the road.. like if bisecting for bugs will get harder or something like that	07:21
	← Shusshu left	07:21
kini	which is, I gather, what tfnico_ is really asking about	07:21
	← brokenshakles left	07:21
	→ syphar joined	07:22
	→ sq-one joined	07:22
EugeneKay	My opinion is that a feature branch should be merged into master(or dev, which is what I use) once it's accepted. Do a test-merge and fix it up before you submit if it doesn't work cleanly.	07:22
	→ LiohAu_ joined	07:22
	→ Radium joined	07:23
kini	but what if the feature branch is not accepted? I mean, you usually merge master into a feature branch because you want to make sure it is not getting out of sync with master too much, but you haven't gotten approval to merge the branch into master, so you do the opposite	07:23
	← LiohAu_ left	07:23
	→ LiohAu_ joined	07:23
EugeneKay	Then you beat the maintainer over the head until it is accepted or you drop the feature branch. :-p	07:23
	It's gonna come down to how cmplex / modularizable your codebase is.	07:24
	← dstorrs left	07:24
	→ brokenshakles joined	07:25
EugeneKay	If feature A gets refactored halfway through to require something provided by new feature B(which was started after A diverged), then you should pull B into A, or from master to A after B has been merged.	07:25
	If that makes sense.	07:26
	← Goplat left	07:26
tfnico_	Background was that a colleague was always doing merge master -> featurebranch before merging the feature branch back to master. So we have merge commits that go "Merge branch 'master' into featureX" instead of the other way around. I got that a --no-ff when merging back to master will fix this, but in this case, the first merge is a bit of a waste anyhow.. Well discussion ensued on whether this w	07:26
	as the right approach, or if the feature branch should be rebased instead :P	07:26
	← dreiss left	07:26
EugeneKay	I say beat your colleague.	07:26
	← tonesfrommars left	07:26
tfnico_	already did that, now we're laying down some conventions for merging and branching :)	07:27
	thanks for the input. I hope it won't be necessary to merge that way very often, but good to know that it's still OK to do so.	07:28
	→ mxweas_ joined	07:28
	→ albel727 joined	07:29
	← mischat left	07:30
	← pantsman left	07:30
	← arvind_k left	07:31
	← ben225 left	07:32
	← ajaykools left	07:33
	→ ipalaus_ joined	07:34
	← tatsuya_o left	07:35
cxreg	"nested localized bash prompt commands" <- does that even make sense to anyone? o_O	07:35
EugeneKay	Not really.	07:36
cxreg	sometimes i wonder if ive gone off the deep end on features	07:36
	https://github.com/cxreg/smartcd/commit/ad855a68f2	07:36
	← zeshoem left	07:36
	→ Nikoli joined	07:36
EugeneKay	What's the one-liner to "warm up" a .git/ into disk cache?	07:37
cxreg	find .git -type f -exec cat {} > /dev/null \;	07:37
_ikke_	cxreg: Where is that usefull for?	07:38
	← aristidesfl left	07:38
_ikke_	that on_prompt	07:38
EugeneKay	There was one that used one of the plumbing commands. I don't think it was fsck.	07:38
cxreg	_ikke_: the use-case i imagined was related to history management	07:38
	→ zeshoem joined	07:38
cxreg	_ikke_: where you may want to 'history -a; history -n' on every prompt, but only in some directories	07:38
_ikke_	ah ok	07:39
cxreg	incidentally, localized bash hisory is another smartcd feature, which would play nicely with that	07:39
	← zeshoem left	07:39
cxreg	im having a feature asplosion o_o	07:39
	← timoxley left	07:40
cxreg	EugeneKay: oh hm, not sure then	07:40
	→ timoxley joined	07:41
cxreg	i would think gc and fsck are good ones for something like that	07:41
	→ martinjlowm joined	07:41
	→ maxm-- joined	07:41
cxreg	repack maybe	07:41
	← maxm- left	07:41
EugeneKay	My PS1 command is running git-status on first-entry to the repo, so that does something.	07:42
	← timoxley left	07:42
	← beautifulmind left	07:42
cxreg	for cache warming you mean?	07:42
EugeneKay	Don't think it does the entire objects store though	07:42
	Yeah	07:42
	→ dag- joined	07:42
cxreg	yeah .. status doesnt touch the whole repo	07:43
	jsut the index and a few objects	07:43
	git-status on first-entry to the repo sounds like a great use for smartcd! duck	07:43
EugeneKay	I parse the status output and keep it in some environment variables. Every 60 seconds, or whenever 'git' is invoked, the prompt command reloads the info	07:44
cxreg	right on	07:44
	→ arvind_k joined	07:44
	→ Ph03b3 joined	07:45
cxreg	you know for some crazy reason i thought i was in #xkcd	07:45
	i was wondering why this was so git-centric	07:45
EugeneKay	One of the vars is the $(git rev-parse --show-toplevel). That is checked on every PS1, and triggers a reload if it's changed.	07:45
	→ ajaykools joined	07:46
EugeneKay	The net effect is that I can move around and work in a large-ish repo without notable slowdown on every command.	07:46
	→ dc5ala joined	07:46
cxreg	nice	07:47
EugeneKay	I could speed it up even more by backgrounding the reloading of the info somehow(storing it in .git/repo_info?), but that seems like too much work.	07:47
	← tvw left	07:50
	← Kulrak left	07:50
	→ dvaske joined	07:50
	← Element9 left	07:51
	→ d0k joined	07:52
	← pppaul left	07:52
	→ Kulrak joined	07:52
	← Kulrak left	07:52
	→ Kulrak joined	07:52
	→ OOPMan joined	07:53
	→ pppaul joined	07:53
	← bauruine left	07:54
	→ dag-1 joined	07:55
	← welly left	07:56
	← ajaykools left	07:58
	← dag- left	07:58
	→ mklappstuhl joined	07:58
cxreg	sounds like you do put some time into your config, i sure would appreciate it if you ever had a chance to give me some feedback on my project :-)	08:00
_ikke_	Bevalt je laptop?	08:01
	→ PigeonFriend joined	08:01
_ikke_	wrong window, sorry	08:01
	← apok left	08:02
	→ dagon666 joined	08:03
	→ overbythere joined	08:04
	← Chaser left	08:05
	← tapas left	08:05
	→ dag- joined	08:08
	→ shruggar joined	08:08
	← sq-one left	08:08
	← dag-1 left	08:09
	← GodEater left	08:10
	← dr0id left	08:10
	→ Lujeni joined	08:11
	← iamjarvo left	08:12
	→ iamjarvo joined	08:12
	→ dag-1 joined	08:12
dagon666	do You guys know any good music streamer for apache ? I used jinzora till now but its getting a little bit outdated	08:13
	sorry wrong channel :)	08:13
	← ipalaus_ left	08:13
	← d0k left	08:14
	→ satyanash joined	08:15
	→ Desproges joined	08:15
	← dag- left	08:15
	→ d0k joined	08:15
	→ Gentlecat joined	08:16
	→ LimeBlast joined	08:16
EugeneKay	dagon666 - dunno what you mean by "for apache", but Subsonic is pretty good. It's GPL, but the canonical distribution requires a "donation" to run after 30 days. I ripped out the licensing code, and built a .war https://github.com/KHresearch/subsonic	08:17
	→ dag- joined	08:17
	→ dcarper joined	08:19
	← dag-1 left	08:20
	→ zheng joined	08:20
	→ iamjarvo1 joined	08:20
	→ pmatthews joined	08:20
	→ dag-1 joined	08:20
	← Radium left	08:20
	→ blacktulip joined	08:21
	← dag- left	08:21
	← iamjarvo left	08:22
	→ timoxley joined	08:22
	← bgerber left	08:22
	← gebura left	08:23
	← dcarper left	08:23
	← cmn left	08:23
	← dc5ala left	08:24
	← RichardBronosky left	08:24
	→ dc5ala joined	08:24
	← dc5ala left	08:25
	→ dc5ala joined	08:26
	← harshpb left	08:26
	→ wubofeng joined	08:27
	← wubofeng left	08:28
	→ kpreid joined	08:30
	→ mischat joined	08:30
	← mischat left	08:30
	→ dag- joined	08:30
	→ master-lincoln joined	08:30
	→ jlebrech joined	08:30
	→ mischat joined	08:30
	→ pyhead joined	08:31
	← dag-1 left	08:31
	→ gyre007 joined	08:31
	→ tatsuya_o joined	08:31
	→ marhaban joined	08:32
	← zheng left	08:32
	← Lujeni left	08:33
	→ bgerber joined	08:34
	→ francisfish joined	08:34
	← pppaul left	08:34
	→ pppaul joined	08:35
	→ sq-one joined	08:36
	→ cmn joined	08:36
	← EditThat left	08:36
	→ marcogmonteiro joined	08:37
	→ EditThat joined	08:39
cjs	/join #ubuntu	08:39
	Oops! Sorry.	08:39
	→ jmd joined	08:40
	→ ipalaus_ joined	08:40
	← sq-one left	08:40
overbythere	cjs: ~_~	08:40
	false advertising was that?	08:40
	← blackmatrix_ny left	08:41
cjs	Yes, I'm a very clever spambot. :-)	08:41
overbythere	:p	08:41
Dulak	/join #ubuntu	08:41
	It's not working.	08:41
kini	it's funnier when people leave out the slash by mistake rather than putting a space in front of it by mistake	08:41
	→ dag-1 joined	08:41
kini	then it really does look like spam	08:41
overbythere	I was on a network one time where the users thought they just had to say the channel name to get there	08:42
	I don't really understand some people's thinking	08:42
	→ Orbitrix joined	08:42
	← maleknet left	08:42
	→ dag-2 joined	08:43
	← neersighted left	08:44
	← dag- left	08:44
	← hagabaka left	08:44
	← dag-1 left	08:45
	→ berserkr joined	08:47
	→ S0lign0c joined	08:47
	→ noisebleed joined	08:47
	← noisebleed left	08:48
	→ noisebleed joined	08:48
	→ neersighted joined	08:48
	→ tvw joined	08:48
	→ _inc joined	08:48
	← xiangfu left	08:48
	→ xiangfu joined	08:49
	gitinfo set mode: +v	08:49
	← gyre007 left	08:49
S0lign0c	hi	08:49
gitinfo	S0lign0c: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.	08:49
S0lign0c	Can I make a git diff and edit with vim in the same time ?	08:49
jmd	S0lign0c: If you have multitasking operating system, then I suppose you can.	08:50
S0lign0c	:D, yes edit with the change, (+-), with diff colors	08:51
	?	08:51
	gitinfo set mode: +v	08:51
	→ tcurdt joined	08:52
	← arvind_k left	08:53
	→ martinjlowm-lptp joined	08:53
	→ dr_lepper joined	08:54
	← dr_lepper left	08:54
	→ dr_lepper joined	08:54
	→ maleknet joined	08:56
	← dag-2 left	08:56
	→ hyperair joined	08:57
	→ dag- joined	08:59
	→ WillMarshall joined	09:00
	← master-lincoln left	09:02
EugeneKay	S0lign0c - sounds like you're after "vimdiff"	09:03
	→ gldn joined	09:04
	← gldn left	09:04
	← dag- left	09:04
pmatthews	or mergetool	09:04
	← maleknet left	09:05
	→ slide23 joined	09:05
	→ master-lincoln joined	09:05
	→ arvind_k joined	09:06
	→ bielewelt joined	09:06
	→ tqmz joined	09:06
	→ slide_ joined	09:08
	← slide left	09:09
	← slide23 left	09:10
	← mklappstuhl left	09:11
jmd	Is there a way I can make git pull default to the --rebase option ?	09:12
	← timoxley left	09:12
	BinGOs uses an alias for that.	09:13
_ikke_	jmd: man git config	09:13
gitinfo	jmd: the git-config manpage is available at http://jk.gs/git-config.html	09:13
_ikke_	jmd: pull.rebase	09:13
BinGOs	purl = pull --rebase	09:13
	← pppaul left	09:14
	→ Fandekasp joined	09:15
	← _inc left	09:15
Dulak	You can just set pull.rebase to true, why use a custom alias?	09:15
	→ gjoseph joined	09:15
	→ pppaul joined	09:15
BinGOs	because I don't like wtf gotchas	09:16
	← lausi left	09:17
	→ lausi joined	09:17
	gitinfo set mode: +v	09:17
BinGOs	horses for courses, jumpers for goal-posts, each to their own. I just prefer to have them as two distinct operations.	09:18
pmatthews	also if you set the pull.rebase, can you specify the -p flag?	09:18
	→ maleknet joined	09:18
	→ djszapi joined	09:18
	← archdori left	09:18
	← marhaban left	09:18
	← harish left	09:19
Desproges	about vimdiff, what the git equivalent to "svn diff -rPREV" ? (on a working directory clean )	09:19
djszapi	Hey! What is the reason behind that I can use the first 6-7 hexa value of a commit hash to identify the commit ? Why can I drop the remaining hexa values ? Could someone please help me with understanding that ?	09:19
	← ndim left	09:19
Grum	djszapi: because sha1sums are quite unique and you'd need a million files on average to get a collision on sha's shorter than 8chars	09:20
_ikke_	djszapi: Because in a single repo, the first values are enough to be unique	09:20
Grum	and git will scream at you when it has a collision :)	09:21
	← Salonce left	09:21
Grum	or rather; when the shortened sha you use is not unique	09:21
	if you find a sha collision I think people pay good money for that :D	09:21
Dulak	pmatthews: don't think so, for that the alias would work, but pull.rebase is the same as 'git pull --rebase'	09:22
Desproges	it's not hexa (16 digits) but 36 digits, (so there is something like 36^6 possibility)	09:22
	→ intripoon_ joined	09:22
Grum	^26 right? :p	09:22
	err 16 >.>	09:23
djszapi	Grum I can write a repo generator app where it collides :)	09:23
	so you said you would pay ? :)	09:23
	→ replore joined	09:23
Grum	djszapi: erm good luck doing that	09:23
	← fission6 left	09:23
	→ dag- joined	09:23
Grum	and I wont pay, some institute on this world will I think :D	09:23
djszapi	just joking.	09:23
Desproges	a-z0-9 Grum	09:23
	← dhruvasagar left	09:23
Desproges	oh now sorry you right	09:24
	0-9a-f	09:24
Grum	i'm confused how :p	09:24
	ay yeah indeed :)	09:24
FauxFaux	djszapi: https://blog.goeswhere.com/2010/12/git-set-commit-id/	09:24
Grum	djszapi: just saying right know there is no known sha collision (at least last i checked :D)	09:24
	→ flavius joined	09:24
Desproges	anyone for my question ? about vimdiff, what the git equivalent to "svn diff -rPREV" ? (on a working directory clean )	09:25
	← intripoon left	09:25
EugeneKay	djszapi - !treeish	09:25
gitinfo	djszapi: What the heck is a "tree-ish"? Read 'man gitrevisions' and http://book.git-scm.com/4_git_treeishes.html	09:25
Grum	Desproges: what does that do?	09:25
FauxFaux	"HEAD" specifies the previous commit, if that's what you're after.	09:25
	→ dhruvasagar joined	09:25
_ikke_	FauxFaux: previous?	09:25
FauxFaux	Most recent?	09:26
EugeneKay	FauxFaux - "current" commit. index would be your "next" commit.	09:26
	→ _Qman joined	09:26
Desproges	svn diff -rPREV returns the diff between the last and the previous commit (if we didn't modified anything)	09:26
FauxFaux	I would call the one in the index the current commit.	09:26
EugeneKay	It's not a commit yet :-p	09:26
Grum	Desproges: git diff HEAD^..	09:26
	← arvind_k left	09:27
djszapi	Grum: thanks for the information.	09:27
	Grum: it means that md5 could also be truncated in certain cases out there ?	09:27
Grum	djszapi: no, git is special	09:27
	and it keeps track of if its unique	09:28
	→ oehman joined	09:28
Grum	so unless you build that system yourself, don't shorten hashes :)	09:28
	→ _inc joined	09:28
Grum	(and md5 has known collissions)	09:28
EugeneKay	djszapi - hash shortening is provided for convenience only, and is expanded to the full one. If a similar hash is encountered git whines about ambiguity and requires you to specify a few more digits	09:28
Grum	also FauxFaux; that is foul .... reallly really foul	09:28
Desproges	thanks Grum. In fact I want something more .. if I do "svn diff -rPREV path/to/file" I get the difference with the last modification of that file	09:29
EugeneKay	SHA1 has known weaknesses and a collision may be found, but none has yet to be encountered AFAIK.	09:29
	I expect git 2.0 or 3.0 to migrate to SHA-2 or even SHA-3	09:29
djszapi	Grum: yes, obviously, but I do not see the differences between this solution and others out there.	09:29
Grum	Desproges: git diff HEAD^.. -- path/to/file :P	09:29
djszapi	I mean it is not unusual there are probably no collisions too chancy.	09:29
Grum	EugeneKay: yeah it will eventually migrate but that will be a pain in the ass	09:29
Desproges	that does not work Grum	09:29
Grum	Desproges: lemme see	09:30
EugeneKay	Grum - there is a repo version number which can be incremented. git-migrate(or such) to rebuild all repos with the new version, et voila!	09:30
Grum	it does work Desproges	09:30
	but you misphrased your question i think	09:30
	→ JeanMertz joined	09:30
Desproges	ok, I reformulate	09:30
Grum	I think you are looking for the 'last change of a particular file irregardless of which commit it is in'	09:31
	which unless that file is changed in the last commit, will not work for that file :)	09:31
Desproges	how can I replace : " git diff `git log /path/to/file \| head -n2 && tail -n1 \| awk '{print $2}'` /path/to/file	09:32
Grum	it works -- why replace it? :/	09:32
EugeneKay	Turn it into a git extension. vim ~/bin/git-foo, then run `git foo` ;-)	09:33
	← djszapi left	09:33
Desproges	because it's a long command line	09:33
EugeneKay	Or, if you're esoteric like me, replace your `git` binary with a wrapper.	09:33
Grum	Desproges: long is bad? :(	09:33
Desproges	yeah I can do a script for taking a file and a number, but I thought there were something to do it already	09:34
	← replore left	09:34
Desproges	well, 2 git command for one result :/	09:34
EugeneKay	What are you trying to get, exactly? The last diff of $FILE ?	09:34
Desproges	git diff `git log`	09:34
Grum	EugeneKay: yeah	09:34
charon	what's wrong with: git log -p -1 $file	09:34
EugeneKay	man git-log, look at the diff-generating options	09:35
gitinfo	the git-log manpage is available at http://jk.gs/git-log.html	09:35
	→ TomyLobo joined	09:35
Grum	charon: too short	09:35
Desproges	oh, very nice charon I think that's it !	09:35
	→ TNNUK joined	09:35
Grum	lovely, i didnt realize -<n> was added .. onmnom	09:35
	gitinfo set mode: +v	09:35
Desproges	mhhh, charon, have you some magic to delete the log line ?	09:36
charon	Grum: that's probably been there since day 1...	09:36
Grum	i call lies :P	09:36
charon	Desproges: huh?	09:36
Grum	charon: he doesnt want the header	09:36
	← tcurdt left	09:36
charon	also note that in a mergy history, the -1 just picks a random* commit	09:36
	* ok you got me, not random but date-ordered modulo topology	09:36
Grum	really random not ^1? :p	09:36
	=D	09:36
EugeneKay	-N works just fine with !lol	09:37
gitinfo	git config --global alias.lol "log --oneline --graph --decorate"	09:37
charon	still, if you don't know the dates of your commits by heart and have merged two lines of history that both change the file, the choice will seem arbitrary	09:37
Desproges	date ordered is fine for me	09:37
	and I m not understanding yet the topology thing	09:37
TNNUK	With git 1.7.5.4; In the past when I've done "git checkout <branch>" where <branch> matches the name of a remote branch, it has automatically set a new local branch to track the remote. For some reason this has stopped happening and I have to manually tell it to track the remote, but only with one repository. Any ideas?	09:37
EugeneKay	The lol graph makes grepping history easy	09:37
Grum	Desproges: just dont do merges ;D	09:37
	FauxFaux high-fives Grum.	09:38
charon	Desproges: what's wrong with the header? you should probably configure your pager correctly, and then just fire up 'git log -p $file' so you can scroll down for further changes	09:38
	← pppaul left	09:38
TNNUK	i.e. other repos in the same gitolite setup still work just fine	09:38
	→ ndim joined	09:38
	LiohAu_ → LiohAu	09:39
	→ pppaul joined	09:39
Grum	TNNUK: maybe you have multiple remotes that provide the same branch and thus it doesn't know which to pick?	09:39
	→ aristidesfl joined	09:40
EugeneKay	TNNUK - !repro the output of 'git branch -avv' and 'git remote -v'	09:40
gitinfo	TNNUK: Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.	09:40
Desproges	What I m looking for is something like " git diff -l 3 /path/to/file" to see the differences of that file between the current one, and 3 modifications before	09:40
	← cbreak-work left	09:40
	← maqr left	09:40
	← dr_win left	09:41
	← _inc left	09:41
Desproges	so maybe what I wrote before can be good in a little script : " git diff `git log /path/to/file \| head -n2 && tail -n1 \| awk '{print $2}'` /path/to/file "	09:41
EugeneKay	Desproges - I would just do a log --graph -- $FOO and then plug the SHAs into diff manually, in all honesty.	09:41
	→ cbreak-work joined	09:41
EugeneKay	Less edge cases	09:41
charon	Desproges: except the && makes no sense there	09:41
	← zivester left	09:41
	→ dr_win joined	09:41
speakman	!aliases	09:41
charon	also, in a properly organized git history you'll want to fire up gitk or similar anyway, and look at the commits themselves.	09:41
EugeneKay	speakman - !triggers	09:41
gitinfo	speakman: See http://jk.gs/git/bot/trigger.php	09:41
Desproges	giving one or more file to checks, and an extra param "$prev" to handle the number	09:41
speakman	thanks	09:42
Desproges	yes charon, - && + \|	09:43
	→ pfFredd joined	09:43
Desproges	git diff `git log /path/to/file \| head -n2 \| tail -n1 \| awk '{print $2}'	09:43
	→ arvind_k joined	09:44
	→ sdeboey joined	09:44
	→ Superlokkus joined	09:47
	→ whyking_ joined	09:47
	← master-lincoln left	09:47
	← _tasslehoff_ left	09:48
	→ pro_metedor joined	09:48
	← weiyang left	09:48
	→ timoxley joined	09:49
	→ cortana2 joined	09:50
	→ papegaaij joined	09:50
charon	Desproges: i think i can properly word my gut feeling now. take a step back and think about why you want to compare the file N revisions back	09:50
speakman	Are there any "consensus" regarding "downstream" fork management with Git? I still can't really figure if I should rebase my local customized fork, or just merge continously from upstream. Ideas? Or better; Links?	09:50
EugeneKay	speakman - !best_practices	09:51
gitinfo	speakman: There is no one right answer for git best practices, but a consensus from #git is available at http://sethrobertson.github.com/GitBestPractices/	09:51
charon	Desproges: i suspect it's because those N revisions are tightly coupled, which would mean you don't separate commits at the right place	09:51
EugeneKay	speakman - how you merge / rebase / etc is up to you and the guy you're submitting patches / pull requests / etc to.	09:51
	← cbreak-work left	09:51
	← txomon left	09:51
	→ achadwick joined	09:51
charon	Desproges: if that's the case, perhaps you can learn about the history rewriting tools (commit --amend, rebase -i) to clean up such chains of half-done commits	09:52
EugeneKay	My preference is to branch where I branch from, then merge back in when I'm done with my branch. If I need to merge something in from the current 'master'(or 'dev') to make my feature work, then I do so. Other than that, KISS.	09:52
charon	Desproges: a month or a year on from now, that will make your history much easier to understand for yourself	09:52
	→ cbreak-work joined	09:52
EugeneKay	speakman - git-flow is a pretty good model that a lot of people seem to use, if not the shell commands themselves.	09:52
Grum	one of the true powers of it :D	09:52
Desproges	charon: today, it was just for seing the evolution and to check nothing was forgotten during merges. I agree I've lot of things to learn (and last time I tried to make a rebase -i , I didn't make it well , but I progress :)	09:53
	→ fisted joined	09:53
	slide_ → slide	09:55
TNNUK	EugeneKay: http://pastebin.com/YHig3zFH Cheers :)	09:55
	→ zivester joined	09:55
	← timoxley left	09:56
TNNUK	Grum: only one remote. Tried with a fresh clone and had the same problem so it's not a local problem. Other people have the same issue too	09:56
EugeneKay	TNNUK - you probably want to use the third syntax of checkout; `git checkout -b foo origin/foo`.	09:56
Desproges	and also using --amend would be really interessing for me	09:56
TNNUK	EugeneKay: yes, that syntax does work. But the syntax I used in the paste worked until a couple of days ago	09:58
	EugeneKay: I can use the other syntax, but it doesn't sit well with me that something has broken	09:58
EugeneKay	TNNUK - there was some discussion about changing the default behaviour on the mailing list a few weeks back, I think this might have been affected. Poke at the changelog for git.git.... or don't use bleeding-edge ;-)	09:59
	← dr_win left	09:59
TNNUK	EugeneKay: it works fine with another repo on the same gitolite setup, so I doubt it would be that	09:59
shruggar	I get a warm fuzzy feeling when I can use git to prove something correct :) After moving some changes into a branch, then merging that branch back into the master, git rev-parse HEAD@{2}: HEAD: gives two identical hashes, proving that I didn't change anything but the graph. Yum :)	10:00
EugeneKay	Probably little to do with gitolite, more to do with your local config	10:00
	→ _tasslehoff_ joined	10:00
EugeneKay	Pastebin your .git/config from the borked repo and the non-borked one ?	10:00
TNNUK	EugeneKay: same problem for other people on different machines too	10:01
	One min	10:01
EugeneKay	Well that is odd, then.	10:02
	← Desproges left	10:02
Dulak	What is @{u}?	10:02
EugeneKay	Dulak - !treeish	10:02
gitinfo	Dulak: What the heck is a "tree-ish"? Read 'man gitrevisions' and http://book.git-scm.com/4_git_treeishes.html	10:02
Dulak	Thanks	10:03
speakman	EugeneKay: the problem comes when you manage a lot of customization which is not of interrest at all for upstream project. One could use quilt and version control a bounch of patches, but it seems a bit old-fashioned.	10:03
	→ timoxley joined	10:03
	← ChaoWang left	10:04
	← pppaul left	10:04
EugeneKay	speakman - eg, your fork should incorporate changes from upstream, but not the other way around?	10:04
	→ pppaul joined	10:04
TNNUK	EugeneKay: the .git/config's are the same :/ (appart from the branch names etc being different obviously)	10:04
EugeneKay	speakman - I would treat 'upstream' as a feature branch, and just merge it in to my primary branch regularly.	10:04
	TNNUK - well, I have nfi. Sorry.	10:04
TNNUK	EugeneKay: thanks anyway :)	10:05
	← madsy left	10:05
EugeneKay	speakman - This is in fact what I do with my fork of Subsonic, where 'upstream' is git-svn.	10:05
	→ beautifulmind joined	10:06
	← timoxley left	10:06
	→ subh joined	10:06
speakman	EugeneKay: Good point! Must run, bbl	10:07
EugeneKay	Is that a threat?	10:08
	← Guest27784 left	10:10
	→ geb joined	10:10
	← pppaul left	10:11
	→ pppaul joined	10:12
	→ slide23 joined	10:13
	→ dixel joined	10:13
	← slide left	10:15
	gitinfo set mode: +v	10:15
dixel	Hello! Need some help: git log HEAD^ HEAD shows all history instead of last two commits - what can I do to fix this?	10:15
	← ehsan left	10:16
EugeneKay	!dots	10:16
gitinfo	In the log family (git-log, gitk, etc.) A..B means "everything in B but not in A" [formally: ^A B] and A...B means "everything in A or B but not in both" [formally: A B --not $(git merge-base --all A B)]. An empty "side" of the dots implies HEAD, so 'git log master..' is very different from 'git log master'!	10:16
EugeneKay	Use .. instead of a space between them.	10:17
dixel	EugeneKay: thx =)	10:17
EugeneKay	You can also use HEAD^.. meaning "from HEAD^ to now". See also !treeish	10:17
gitinfo	What the heck is a "tree-ish"? Read 'man gitrevisions' and http://book.git-scm.com/4_git_treeishes.html	10:17
	EugeneKay tosses gitinfo a !botsnack	10:17
	← cjs left	10:17
EugeneKay	Oh wtf, he doesn't work in /me? Bad bot!	10:17
	← Fandekasp left	10:18
shruggar	does he even respond to !botsnack ?	10:19
gitinfo	Om nom nom	10:19
	→ dcarper joined	10:19
	← dixel left	10:19
	→ Desproges joined	10:20
	→ Fandekasp joined	10:20
	→ tespezzo joined	10:20
	← tespezzo left	10:20
	→ eijk joined	10:21
	← maleknet left	10:22
	→ hroi joined	10:22
hroi	hi	10:22
gitinfo	hroi: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.	10:22
hroi	I just joined google codes.	10:22
	→ harish joined	10:22
hroi	and set up a project with git...	10:23
	← pfFredd left	10:23
hroi	I'm wondering how I can now just push all my current code in there to get started...	10:23
	→ falena joined	10:23
shruggar	(does google code support git?)	10:23
	← dcarper left	10:24
TNNUK	EugeneKay: I think it's info/refs in the repo :/	10:24
	→ ajaykools joined	10:24
TNNUK	EugeneKay: Some of the branches are missing from info/refs, there are some that are in there, which work	10:25
EugeneKay	Interesting.	10:25
FauxFaux	git add 70mb.txt && git commit -m "lol" literally hangs my q9600 / 4gb / ext4 / linux 3.2 (lolbuntu) machine for a couple of minutes. :s	10:26
	hroi: git remote add origin https://google/code/url/for/project && git push -u origin master	10:26
	→ f0i joined	10:26
pro_metedor	git submodule update	10:26
	fatal: Needed a single revision	10:26
	Unable to find current revision in submodule path 'xxx'	10:26
	whats going on?	10:27
	← dag- left	10:27
EugeneKay	pro_metedor - !repro	10:27
gitinfo	pro_metedor: Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.	10:27
	← satyanash left	10:28
pro_metedor	I cannot get a submodule: this is the output https://gist.github.com/2358500	10:28
hroi	FauxFaux: is origin my my local project source code folder?	10:28
	FauxFaux: how does git know what to copy over?	10:28
_ikke_	pro_metedor: The people who updated the submodule forgot to push it or something like that	10:28
TNNUK	EugeneKay: git update-server-info in the remote repo fixed the info/refs file, but it looks like that was a symptom of the same problem	10:29
	Becuase it still doesn't work	10:29
FauxFaux	hroi: No, run those in the repo; 'origin' is just a (standard) name. 'master' is the name of the branch to copy onto google code.	10:30
	→ Bartzy joined	10:31
Bartzy	A bit unrelated to git itself, but maybe someone knows: When doing a pull request from a branch to another branch in GitHub, comments on the code on commits there vanishes once a user force pushes new commit history to the source branch	10:32
	Can I somehow see the comments? This is our code review process, so it's pretty lame that I c	10:32
	Can I somehow see the comments? This is our code review process, so it's pretty lame that I can't see the original comments I had on the code.	10:32
FauxFaux	#github	10:32
Bartzy	OK, thank you.	10:33
	→ maleknet joined	10:34
	← WillMarshall left	10:35
	← jceb left	10:36
	→ sattu94 joined	10:40
hroi	all repository packages, could do with a little less typing...	10:42
	how about... git syncup myfolder http://location ... (end of strory)	10:43
	← netiru left	10:43
hroi	"git remote add origin ..." I ask, why the "remote" isnt it obvious?... why the "add" isnt that obvious too?	10:44
	← rue_XIV left	10:45
	→ rue joined	10:45
	← Guest7979 left	10:46
	→ honschu joined	10:47
	← honschu_ left	10:48
	→ robotmay joined	10:48
hroi	FauxFaux: git remote add origin https://google/code/url/for/project && git push -u origin master ... doesnt work for me...	10:49
	FauxFaux: I guess I have to do something first... maybe register each file i want to be part of my git.	10:50
	← iamjarvo1 left	10:51
	← kadoban left	10:51
	← brokenshakles left	10:53
	→ jceb joined	10:53
	← davidcoallier left	10:54
	→ brokenshakles joined	10:54
	← enmand left	10:54
	→ DarkAR joined	10:55
	→ davidcoallier joined	10:55
canton7	hroi, yes, you push commits. If you haven't made any commits, git won't push anything	10:56
	hroi, !book is a good place to start with git	10:56
gitinfo	hroi: There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !gitt !vcbe and !parable	10:56
	→ iocor joined	10:58
	→ j0ran1 joined	11:00
	← j0ran left	11:01
	← vdv left	11:01
	← zdk left	11:02
hroi	canton7: thanks, when i have time to read a book... man git, is interesting... especially the title "stupid content tracker" ... but I guess the user must be very knowledgable.	11:04
gitinfo	canton7: the git manpage is available at http://jk.gs/git.html	11:04
	← pppaul left	11:04
	→ sbell joined	11:04
	← tvw left	11:04
	→ tvw joined	11:04
canton7	hroi, yeah. Sadly git isn't really something you can pick up without reading some sort of comprehensive guide	11:04
	← sbell left	11:04
TNNUK	EugeneKay: argghhh; I had directory names matching the branch names	11:05
	TNNUK is so stupid	11:05
	→ pppaul joined	11:05
EugeneKay	!dashes	11:05
gitinfo	Use a double-dash(--) to separate refs from paths, especially when dealing with ambiguous file names. Ex: git checkout master -- origin (check out the file "origin" from branch "master")	11:05
	← hyperair left	11:05
	← j0ran1 left	11:05
TNNUK	EugeneKay: thank you for your help	11:06
EugeneKay	I suppose a 'ls' would have told me that in your pastebin :-p	11:06
TNNUK	EugeneKay: indeed. Sorry for wasting your time :-D	11:06
	← TNNUK left	11:06
	EugeneKay shrugs	11:06
	← guns left	11:07
EugeneKay	My other activity for the evening is getting stoned on pain pills and wishing my knee would STOP FUCKING HURTING.	11:07
	So, waste away!	11:07
	← caseymcg left	11:08
	← davidcoallier left	11:08
hroi	ok I just did a $ git commit -a ... then $ git push ... git reports everything up to date... an nothing new shows up on the remote server.	11:09
	← LiohAu left	11:10
EugeneKay	'git branch -avv'	11:10
	And !repro	11:10
gitinfo	Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.	11:10
	← oehman left	11:12
	→ oehman joined	11:12
	← oehman left	11:12
	→ matthieu6371 joined	11:14
	← matthieu6371 left	11:15
hroi	EugeneKay: thanks, here is the terminal... https://gist.github.com/2358672	11:15
	→ Peste_Bubonica joined	11:16
hroi	EugeneKay: avhrr2vw is a new directory with some things in it.	11:16
EugeneKay	Can you please run: git branch -avv, and git remote -v, and add to the Gist?	11:16
hroi	EugeneKay: I simply want to push all of this up to google	11:16
	ok	11:16
	← marthinal left	11:16
hroi	EugeneKay: https://gist.github.com/2358679	11:17
EugeneKay	try: git push origin master	11:18
hroi	EugeneKay: ok finally git is doing something.. uploading it seems	11:19
EugeneKay	Yup.	11:20
	→ luke__ joined	11:20
EugeneKay	When it's done, create a new commit and then try: git push	11:20
	It should automagically know which remote(origin) and branch(master) you want.	11:20
	→ j0ran joined	11:20
	← Fandekasp left	11:21
hroi	EugeneKay: autmagically :) like that word... but didnt feel like automagical	11:21
EugeneKay	The reason it didn't work the first time is because the original you cloned from didn't have any history to start with - normally, it "just works" with git push.	11:21
	← pppaul left	11:23
	→ pppaul joined	11:23
hroi	EugeneKay: why doesnt anything show up on google codes source or download page?	11:24
cmn	I'm fairly certain you have to put tarballs by hand	11:24
hroi	EugeneKay: no, I see it now	11:24
	cmn: EugeneKay: can I symbolic link some local directories into the local source directory... and then git add that?	11:26
EugeneKay	No, git would track it as a symlink	11:26
	You can mv them into your repo and then symlink where they used to be to get the same sort of effect.	11:26
	← dv310p3r left	11:27
hroi	EugeneKay: problem is I have a local directory called simply src/ and has a number of apps under it that I want all to be part of the same system.	11:28
	← pppaul left	11:28
	→ pppaul joined	11:28
	← maleknet left	11:30
	← pyhead left	11:31
	→ pyhead joined	11:32
	→ enmand joined	11:32
	→ oriba joined	11:32
	→ hyperair joined	11:33
	← pppaul left	11:33
	→ master-lincoln joined	11:33
	→ pppaul joined	11:34
	→ storrgie joined	11:35
	← storrgie left	11:35
	← pyhead left	11:36
	← gchristensen left	11:36
	← JackWinter left	11:39
	→ vervic joined	11:40
	→ iamjarvo joined	11:40
	← master-lincoln left	11:40
	→ stroll2 joined	11:42
	→ litwol joined	11:42
litwol	hello	11:42
gitinfo	litwol: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.	11:42
litwol	O_O	11:42
	oh nice bot	11:42
	→ maleknet joined	11:42
	→ EyesIsMine joined	11:42
litwol	so i have a git branch checkout conflict. in master branch i have tarball extracted in sites/all/modules/contrib/ctools, in my work branch i replaced this tarball extract with git submodule. chances are i did it halfassed cause now i have a conflict... when i try to checkout master branch ("git checkout master") it aborts once merge hits the ctools directory.	11:43
	→ JackWinter joined	11:43
	→ kerrick joined	11:44
bremner	In http://paste.debian.net/162848/, 5ee29e4 is cherry-picked from 4220da4 and I verified they have the same patch-id. git cherry origin/master HEAD nonetheless reports 4220da4. Any ideas what could be going on?	11:44
litwol	it seems however, it is entirely possible to switch branch from master to work branch, while master has tarball checkout and work branch has git submodule	11:44
	← berserkr left	11:44
litwol	so the tarball goes away and submodule comes in (when switching from master to work branch)	11:44
	but it is not possible to go in reverse (conclusion via aborted checkout)	11:44
	← _Qman left	11:44
	← Ph03b3 left	11:45
litwol	i have not tried any rebasing or whatever else is possible because i've never used more advanced techniques before, it is entirely possible they may be the solution to my problem.. right now i just dont know and too scared to screw up	11:45
	→ Kre10s joined	11:45
	→ berserkr joined	11:45
	← ajaykools left	11:46
litwol	hopefuly some one here can advise what next step to try to resolve this, so i can reverse above process (ie checkout master branch, going from work branch which has gitsubmodule to master branch which has conflicting directory regular tarball extract which is tracked in repo)	11:46
	→ marhaban joined	11:46
Kre10s	I added a pibkey to keydir altered the config to allow RW on a repository, set up the remote in my local repo, but git pull asks for gitolites password not the private keys... why could this be?	11:47
	→ palyboy joined	11:47
	→ shruggar1 joined	11:48
bremner	ssh setup problem	11:48
	Kre10s: see the ssh troubleshooting doc on the gitolite site	11:48
	← gjoseph left	11:49
sitaram	bremner: do either of those commits have a + or a - sign in front?	11:49
	→ Superlokkus_ joined	11:49
bremner	sitaram: a + for 4220da4	11:50
sitaram	bremner: and I suppose if you reverse the arguments you get the other one with a + in front?	11:50
	← dhruvasagar left	11:50
	← shruggar left	11:50
	← pppaul left	11:51
	← Superlokkus left	11:51
	Superlokkus_ → Superlokkus	11:51
	← _tasslehoff_ left	11:51
bremner	sitaram: if I reverse the arguments I get nothing. Is that unexpected?	11:52
	→ pppaul joined	11:52
sitaram	bremner: to me, yes but that doesn't necessarily mean much :)	11:52
	→ dhruvasagar joined	11:52
Kre10s	ahh I see the problem: remote: could not symlink gitolite-hooked	11:52
	→ pyhead joined	11:52
Kre10s	I guess the .ssh/authorized_keys isn't being updated.	11:53
sitaram	bremner: is it possible that one of them is reachable from both heads and the other is not?	11:53
	← jbub left	11:53
bremner	sitaram: yes, that seems to be the case; 5ee29e is reachable from both, but 4220da4 only from HEAD	11:55
sitaram	bremner: yup; I just noticed	11:55
bremner	isn't that the normal case though?	11:55
	→ pfFredd joined	11:55
bremner	or maybe git-cherry does not expect upstream to be merged in?	11:55
	← OOPMan left	11:55
	→ jbub joined	11:56
	→ Qoidnef joined	11:56
sitaram	bremner: I don't think that's normal but the point is that this is logical if you imagine the algo as "drop common ones then report what's left", but the dropping happens only pairwise	11:56
	→ binman-rt joined	11:56
sitaram	bremner: all the times I have used 'git cherry' have been on divergent refs	11:57
	→ reels joined	11:57
	← pppaul left	11:57
	→ pppaul joined	11:57
bremner	sitaram: ok, I think this is enough of a clue for me to figure it out. Probably I have to script up a loop with git rev-list and git patch-id.	11:57
	thanks.	11:57
sitaram	bremner: cherry origin/master v1.4rc3-1	11:58
	← j0ran left	11:59
	→ ajaykools joined	11:59
sitaram	that's the last commit before you merged both sides	11:59
	you're welcome!	11:59
	→ j0ran joined	11:59
	tizzo-afk → tizzo	12:00
	← binman-rt left	12:00
sitaram	bremner: if you rebased HEAD on top of origin/master, things would even out. I think	12:00
	but you've got merges; rebase is... unrepdictable with merges :)	12:01
	Kre10s: if it couldn't symlink something, the Unix permissions are borked.	12:02
	← Qoidnef left	12:02
	→ jim0203 joined	12:03
	← j0ran left	12:04
	→ michale joined	12:04
litwol	How do you guys resolve branch checkout aborts when branch being checked out has tracked code added after tarball export, while branches being switched from has that same directory tracked as submodule?	12:04
	→ cannonball joined	12:04
	→ mikewintermute joined	12:05
	jlebrech → jlebrech\|afk	12:05
	← ph^ left	12:06
bremner	sitaram: your cherry works as expected, with a - in front of 4220d	12:06
Bartzy	Weird stuff - I had 2 branches on github, merged the via a pull request both, and now the commit history is not ordered right	12:07
	It is ordered from old to new:	12:07
	→ gchristensen joined	12:07
Bartzy	commit of branch 1 -> commit of branch 2 -> merge commit of branch2 -> merge commit of branch1. Why is that ? :\	12:07
	→ djszapi joined	12:07
	← djszapi left	12:08
sitaram	bremner: which means my guess about the "algorithm" is correct or close enough	12:08
	:)	12:08
	→ _tasslehoff_ joined	12:09
	→ ph^ joined	12:09
	← hroi left	12:10
	← ph^ left	12:10
	→ davidcoallier joined	12:11
	→ n1x joined	12:12
	→ ph^ joined	12:12
	→ oehman joined	12:12
pro_metedor	hot to discard all changes up to the last commit ?	12:13
	how*	12:13
Bartzy	How do I see the parent branch of my branch ?	12:14
Silex	pro_metedor: you want to make it so your history only has one commit (the last one) ?	12:14
pro_metedor	nope, I just want to fall back to last commit	12:15
Silex	oh, "git reset --hard"	12:15
	but be cautious, that really lose changes	12:15
	← n1x left	12:16
pro_metedor	I want the files that are one the "modified" section, to be brought back as they are on repo	12:16
	→ netiru joined	12:16
pro_metedor	I have some submodules attached so I need to be careful for them	12:16
	← jceb left	12:17
pro_metedor	I know there is something with --cached word that works	12:17
	→ j0ran joined	12:17
	→ orkz joined	12:18
	← orkz left	12:18
	← _Vi left	12:18
	→ gusnan joined	12:18
	← DarkAR left	12:19
cbreak-work	pro_metedor: git reset --hard works.	12:19
Kre10s	bremner, sitaram thanks.	12:19
Silex	you can also git checkout -- filename	12:19
cbreak-work	it will nuke all changed files back to the revision you specify, like HEAD	12:19
	→ dcarper joined	12:20
	→ orkz joined	12:20
	← ajaykools left	12:20
cbreak-work	Bartzy: branches don't have parents	12:20
	do you mean upstream?	12:20
Silex	btw, what's the diff between "git reset --hard -- file" and "git checkout -- file" ?	12:20
Bartzy	cbreak-work: I mean the hash the branch was started from.	12:21
cbreak-work	Silex: one works, the other doesn't	12:21
Bartzy	cbreak-work: Is it possible to get that?	12:21
	→ jokar joined	12:21
	← dwmw2_gone left	12:21
cbreak-work	reset --hard is reset in branch mode, it does not take single file arguments	12:21
	Bartzy: no	12:21
	Bartzy: git does not track that.	12:21
Silex	cbreak: oh, interesting. Thanks	12:22
cbreak-work	Bartzy: you can get the merge base relative to some other branch if you want	12:22
	jokar hello all	12:22
jokar	i need help about git	12:22
_ikke_	jokar: !just_ask	12:22
gitinfo	jokar: You can just ask your question. If anybody knows the answer, they will answer soon (most of the time)	12:22
jokar	i'm a beginner in git	12:22
	→ dwmw2_gone joined	12:23
jokar	what should i do?	12:23
_ikke_	!book	12:23
gitinfo	There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !gitt !vcbe and !parable	12:23
_ikke_	Read that	12:23
jokar	gitinfo: excuse me can i send personal message?	12:23
Bartzy	cbreak-work: Hmm.. how ?	12:23
_ikke_	jokar: gitinfo is a !bot	12:24
gitinfo	jokar: [!gitinfo] I am an IRC bot which responds to certain keywords to provide helpful(?) information to humans. Please see http://jk.gs/git/bot for more information about how to use me.	12:24
	← dcarper left	12:24
jokar	hehe	12:24
	_ikke_ :can i send pm?	12:24
cbreak-work	Bartzy: man git-merge-base	12:24
gitinfo	Bartzy: the git-merge-base manpage is available at http://jk.gs/git-merge-base.html	12:24
_ikke_	jokar: Nope, please keep it in here	12:24
	→ cccaldas joined	12:24
jokar	_ikke_: i have some persoanl question?	12:25
_ikke_	jokar: Personal in what sense	12:25
cbreak-work	he wants to know your age.	12:25
jokar	_ikke_: i have some question about git and a book	12:25
cbreak-work	or bank account number	12:25
jokar	nope	12:25
cbreak-work	or hobbies	12:26
_ikke_	jokar: So why is that personal?	12:26
jokar	_ikke_: i want translate a book about git in my native language	12:26
	but i didn't know anything about it	12:27
	is it good for developer	12:27
_ikke_	jokar: https://github.com/progit/progit	12:27
	jokar: What is your native language?	12:27
	← orkz left	12:27
jokar	_ikke_: persian	12:27
	← michale left	12:28
_ikke_	farsi?	12:28
jokar	yeah	12:28
_ikke_	I can recommend progit	12:28
	and it's open source, so you can easily provide a translation	12:29
jokar	progit book?	12:29
_ikke_	yes	12:29
jokar	_ikke_: how git help developer?	12:29
cbreak-work	it's an SCM, it manages source code	12:30
jokar	it have any interface?	12:30
cbreak-work	it's a multi-parallel temporal/spatial source manager	12:30
	yes	12:30
	it has a very nice command line interface	12:30
jokar	can you show me a output	12:30
cbreak-work	and a number of third party graphical user interfaces	12:30
	you can install it yourself	12:31
	← subh left	12:31
cbreak-work	here's a short tutorial: http://schacon.github.com/git/gittutorial.html	12:31
	→ jsoa joined	12:31
	→ dbgster joined	12:31
	← mgpcoe left	12:31
	→ timoxley joined	12:32
	→ jceb joined	12:32
jokar	cbreak-work: thank you	12:33
	→ ajaykools joined	12:33
_ikke_	jokar: See, that's why we keep it in this channel	12:33
jokar	ok	12:33
	→ _inc joined	12:34
	← netiru left	12:37
	→ snearch joined	12:38
	→ chronos joined	12:39
	← cooloney left	12:39
	← maleknet left	12:41
	tizzo → tizzo-afk	12:43
	← hobbyist left	12:43
litwol	How do you guys resolve branch checkout aborts when branch being checked out has tracked code added after tarball export, while branches being switched from has that same directory tracked as submodule? This error happens: error: The following untracked working tree files would be overwritten by checkout:	12:43
	→ Bass10 joined	12:43
FauxFaux	That line had a lot of words not in the right order.	12:43
	→ glennpratt joined	12:44
litwol	yey finally an aknowledgement	12:44
	jlebrech\|afk → jlebrech	12:44
litwol	FauxFaux: gladly will reformat if i get some tips how it should be	12:44
FauxFaux	litwol: 1) Why are you trying to switch branches with uncommitted files?	12:45
litwol	FauxFaux: everything is committed.	12:45
	→ alberto56 joined	12:45
litwol	i can give full description if you're up for it	12:45
	with exact (not so long) list of steps taken to get to current state (ie, the conflict)	12:46
FauxFaux	You said that you have untracked files. Why are they not committed? If they were committed then you wouldn't get the message?	12:46
	→ macabre joined	12:46
litwol	no no	12:46
	everything is tracked	12:46
	i assure you	12:46
	→ jrots joined	12:46
litwol	let me elaborate	12:46
FauxFaux	Git disagrees.	12:46
	→ brentdanley_ joined	12:46
	← brentdanley_ left	12:46
litwol	1) i started with project repo (ie, origin/main). into it i introduces a directory with files via tarball extract. then i added those files to git (so entire directory and new files are tracked)	12:47
	2) i created new branch (work branch). in that branch i removed above tarball extract dir+files and replaced with git submodule. committed the work branch just fine.	12:48
	3) tried checking out back into master branch, conflict happened.	12:48
Bartzy	git reset is awesome	12:48
	→ thiago joined	12:49
litwol	FauxFaux: admittedly, to get submodule /contents/ into my repo i had to run git submodule update.... could it be that files pulled via submodule update are what being conflicted ?	12:49
	← _inc left	12:49
	← snearch left	12:50
charon	litwol: from the POV of the super-repo, the files in the sub-repo are untracked. (it merely tracks the fact that there's a submodule and its tip sha1)	12:51
	→ mercutio22 joined	12:52
	→ TeckniX joined	12:52
	→ diegoviola joined	12:52
litwol	charon: i c	12:52
charon	litwol: admittedly the submodule interface is rather rough if you remove/rename submodules. perhaps you can link up with the relevant devs on the mailing list to help improve it	12:52
	← dwmw2_gone left	12:52
	→ netiru joined	12:52
litwol	actually	12:52
	i didnt think of this before...	12:52
	but ...	12:52
	litwol tries --force	12:53
mercutio22	I would like to checkout an older version of my code because the newest changes are broken. How would I do that?	12:53
	→ maleknet joined	12:53
EugeneKay	mercutio22 - many different ways. git checkout -b oldversion $OLDSHA will give you a new branch, named oldversion, starting at $OLDSHA	12:53
	← papegaaij left	12:54
EugeneKay	!fixup may be a good read for you, s well as the !book	12:54
gitinfo	So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions. For hints type !fixup_hints in IRC. Remember: if you have pushed already, there are only a few things you can do without !rewriting_public_history (type that for more info)	12:54
	There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !gitt !vcbe and !parable	12:54
litwol	this is nice	12:54
	← jokar left	12:54
litwol	by employing --force i was able to coerse git into submission with the /desired/ endresult	12:54
	so overall i am happy	12:54
	→ BSaboia joined	12:55
mercutio22	EugeneKay: thanks for the links	12:55
	→ Ryback_ joined	12:55
	→ dwmw2_gone joined	12:55
mercutio22	EugeneKay: but yes, I want a new branch with an oldversion so I guess I will do the #OLDSHA thing	12:56
	→ babilen joined	12:57
litwol	FauxFaux: charon: thx guys. you got me thinking closer towards the solution and now all is resolved. much appreciated :)	12:57
	→ revagomes joined	12:59
	← kerrick left	13:02
	→ DarkAR joined	13:02
	← litwol left	13:03
	← pppaul left	13:05
	→ ph^_ joined	13:05
	→ neurodrone joined	13:05
	← kaen left	13:05
	← ph^ left	13:05
	→ pppaul joined	13:06
maxorator	I want to move the whole root dir into one new dir in root dir, is it possible to make git understand those are not new files (so their history and diffs would still be intact)?	13:07
	← ben_alman left	13:07
FauxFaux	maxorator: It'll work it out automagically. man git log -M	13:08
gitinfo	maxorator: the git-log manpage is available at http://jk.gs/git-log.html	13:08
	← Chib left	13:09
	← ybit left	13:10
	← pppaul left	13:10
	→ pppaul joined	13:11
	tizzo-afk → tizzo	13:11
	→ saidi joined	13:12
	→ AaronMT joined	13:13
	← pppaul left	13:14
	→ pppaul joined	13:14
	→ ben_alman joined	13:15
	→ aalex joined	13:15
	→ ybit joined	13:17
	→ alexisabril joined	13:17
	← neurodrone left	13:18
	→ sbell joined	13:19
	→ fission6 joined	13:19
	tizzo → tizzo-afk	13:20
	changedNicks → upasana	13:20
	upasana → upasna	13:20
	← resmo left	13:20
	→ resmo joined	13:21
	← vervic left	13:21
	→ Contex joined	13:21
	← saidi left	13:21
	→ Element9 joined	13:22
	→ OOPMan joined	13:22
	← EricInBNE left	13:22
	→ dv310p3r joined	13:25
	← _tasslehoff_ left	13:25
	→ _Vi joined	13:26
	→ CannedCorn joined	13:27
	→ mrwalker joined	13:28
	→ mikos joined	13:28
	← alberto56 left	13:29
	→ wbednarski joined	13:30
	← sattu94 left	13:30
	→ williamcotton joined	13:31
	← sbell left	13:32
	→ sbell joined	13:32
	← kermit left	13:32
	→ kevwil joined	13:33
	← aalex left	13:33
	→ Vortex34 joined	13:34
	→ kermit joined	13:34
	← fayaz left	13:34
	→ t4nkd joined	13:34
	→ chrisgeorge joined	13:35
	← netiru left	13:35
	← Vortex35 left	13:36
	← sbell left	13:37
	→ lyddonb joined	13:37
	→ ddenis joined	13:37
	→ RichardBronosky joined	13:38
	→ juanfer joined	13:39
ddenis	hi guys. How do I pass a rename threshold to git merge ? For git log and diff there is -M which works fine for me (-M80), but I for git merge I tried using -X rename-threshold=80 which still results in a conflict saying files were deleted on one side and deleted on teh other	13:39
	→ macabre^ joined	13:40
	→ satyanash joined	13:40
	→ brngardner joined	13:41
	← oehman left	13:41
ddenis	sigh, I should read the output carefully before asking	13:42
	merge.renamelimit option is what I am looking for	13:42
	← iocor left	13:42
	→ ISF joined	13:42
	→ dcarper joined	13:43
	← macabre left	13:43
	→ iocor joined	13:44
	→ smw_ joined	13:45
	← stkrzysiak left	13:48
	→ ehsan joined	13:48
	← Maccer left	13:48
	← hyperair left	13:48
_ikke_	Ok, weird	13:49
	n/m	13:49
	← xiangfu left	13:51
ddenis	hm, there is still one issue - the rename of one file is correctly detected by git log --name-status, but not by git merge with the same rename threshold	13:52
	← dcarper left	13:53
	← maleknet left	13:55
	→ arti_t joined	13:56
	→ Maccer joined	13:56
	→ mister_m joined	13:56
	→ Shusshu joined	13:57
	← DarkAR left	13:57
	→ _inc_ joined	13:57
	← ehsan left	14:00
	→ master-lincoln joined	14:00
mercutio22	is there a way for me to find which was the last commit that modified a particular file?	14:02
FauxFaux	git log -1 file	14:02
	← CannedCorn left	14:03
	← gavin_huang left	14:05
	→ nimesh_accenture joined	14:05
nimesh_accenture	how do i get a list of remote branches in git?	14:06
_ikke_	git branch -r	14:06
	→ gavin_huang joined	14:06
nimesh_accenture	no i want to clone a remote branch "git://gitorious.org/gstreamer-omap/gst-ducati.git"	14:06
	but not the master branch	14:07
	← gavin_huang left	14:07
	← zivester left	14:07
	← gchristensen left	14:07
_ikke_	git ls-remote	14:08
	→ gavin_huang joined	14:08
	→ zivester joined	14:08
nimesh_accenture	so what would be my command? git ls-remote git://gitorious........ ?	14:08
_ikke_	yes	14:09
	tizzo-afk → tizzo	14:10
nimesh_accenture	i'm getting this:	14:11
	fatal: http://gitorious.org/gstreamer-omap/gst-ducati.git/info/refs not found: did you run git update-server-info on the server?	14:11
	what does that mean?	14:11
_ikke_	Is it a dumb http server?	14:11
	(ie, not git enabled)	14:12
	hmm	14:12
	gitorious	14:12
nimesh_accenture	no git:// is blocked from where i work so i replaced git:// with http://	14:12
_ikke_	You can't just do that	14:12
	Not if they don't offer the repo via http	14:12
	← thelazydeveloper left	14:12
nimesh_accenture	git clone via http works	14:13
	→ netiru joined	14:13
nimesh_accenture	so the exact command that i gave was :	14:14
	git ls-remote http://gitorious.org/gstreamer-omap/gst-ducati.git	14:14
_ikke_	I can't clone the repo via http	14:14
	← williamcotton left	14:14
	← arti_t left	14:15
nimesh_accenture	oh.. i didn't try on that particular repo... let me check	14:16
variable	Are there pre-svn-dcommit and post-svn-dcommit hooks ?	14:16
	→ marthinal joined	14:16
_ikke_	nope	14:17
	→ zokko joined	14:17
zokko	hi guys	14:17
gitinfo	zokko: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.	14:17
	← ph^_ left	14:17
	→ Xeon06_ joined	14:17
nimesh_accenture	the url is git.gitorious. ....	14:18
	→ coyotebush joined	14:18
Xeon06_	Hey guys	14:18
	← Kre10s left	14:18
nimesh_accenture	so now it works	14:18
Xeon06_	I made a little mistake	14:18
zokko	i need to push all branches from local git repository to gitolite, should I use remote origin add?	14:18
variable	_ikke_: :( Is it possible to work around that? Do I need to make an alias and a script?	14:18
	→ gchristensen joined	14:18
nimesh_accenture	thx ikke	14:18
Xeon06_	I merged a branch into master and PUSHED, and I shouldn't have. Is there any way to undo it?	14:18
	← frogonwheels left	14:21
	← Xeon06_ left	14:21
	→ apok joined	14:21
	→ Xeon06_ joined	14:22
_ikke_	Xeon06_: !rewrite	14:22
gitinfo	Xeon06_: [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)	14:22
	→ duckxx joined	14:23
	← zivester left	14:23
	→ mneorr joined	14:24
	→ zivester joined	14:24
	← fission6 left	14:26
	→ eijk_ joined	14:26
	← eijk left	14:26
	← whyking_ left	14:27
	← dkannan left	14:27
zokko	guys, how can i remove(delete) some branch ?	14:27
	→ tktiddle joined	14:27
_ikke_	zokko: git branch -d <name>	14:27
tktiddle	if i delete some files locally will git delete them from the repo too?	14:28
	← jmd left	14:28
_ikke_	tktiddle: Not unless you commit it	14:28
cmn	tktiddle: see what git status says	14:29
	← JohnBat26 left	14:29
tktiddle	so i can delete a file then do "git commit -am 'deleted a file'" ? and when i push it th file will be deleted on the server as well?	14:30
	← zivester left	14:30
cbreak-work	tktiddle: no	14:31
	git stores history.	14:31
	→ d2dchat joined	14:31
cbreak-work	pushing only pushes new commits, new snapshots.	14:31
	the file will still be in the old ones.	14:31
cmn	deleting stuff from the history is the opposite of what you want to do with git	14:32
	or any VCS	14:32
tktiddle	OK but it will no longer show on the head rith? aI don't have to use "rm" to do what I described..	14:32
	→ rockets joined	14:32
	→ zivester joined	14:33
	← lausi left	14:33
	→ lausi joined	14:33
	→ ehsan joined	14:33
	→ savr joined	14:34
	← zivester left	14:34
savr	git add .	14:34
	isn't adding files o.0	14:34
	→ herdingcat joined	14:34
	→ lausi_ joined	14:34
	← lausi_ left	14:34
cbreak-work	savr: which files is it not adding?	14:34
	it is only supposed to add .	14:34
	→ zivester joined	14:34
savr	everything frankly... I think it is the git server	14:35
	let me test	14:35
	← herdingcat left	14:35
cbreak-work	if you want to add everything	14:35
	use git add -A	14:35
	that will add everything in the current folder	14:36
	→ DarkAR joined	14:36
savr	aha	14:36
	→ eph3meral joined	14:37
savr	including subdirectories?	14:37
cbreak-work	no. But files in subdirectories	14:38
	→ bigkitty joined	14:38
bigkitty	hi all	14:38
gitinfo	bigkitty: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.	14:38
savr	pardon?	14:38
canton7	savr, git doesn't track directories	14:38
savr	including the files in subdirectories ?	14:39
zokko	how about re-initalizing some repository under gitolite?	14:39
	i mean, i'd love to empty current repo and fill it with new data?	14:39
savr	seems like it was a laggy server	14:39
Xeon06_	_ikke_: I don't really care about history, I just want to "unmerge" my branch from master	14:39
	→ tty_ joined	14:39
cbreak-work	zokko: just make a new repository	14:39
zokko	cbreak-work: but i want it to have same name	14:40
cbreak-work	savr: git does not track directories, so it adds only files. All files, wether directly or in a subdir.	14:40
_ikke_	Xeon06_: Is the merge the last thing you did? (except for the push)	14:40
savr	:D	14:40
cbreak-work	zokko: then force push new history	14:40
zokko	maybe there is some way to do it without deleting and creating?	14:40
savr	what I wanted to hear	14:40
	thanks	14:40
	← smw_ left	14:40
	← savr left	14:40
Xeon06_	_ikke_: In master, yes. I then went in another branch and committed some other unrelated things before noticing my mistake	14:41
_ikke_	Xeon06_: ok	14:41
	so if you're on master, git log -1 shows the merge?	14:41
	→ arietis joined	14:42
Xeon06_	_ikke_: It shows the last commit in the branch I merged into master	14:42
	← ISF left	14:42
	→ sbell joined	14:43
	← dagon666 left	14:44
_ikke_	Hmm?	14:45
Xeon06_	_ikke_: There no message explicitely saying I merged a branch, but the latest commit in my master branch is the latest commit in the branch I merged into it	14:46
	← tty_ left	14:47
	← traviscj left	14:48
	→ hobbyist joined	14:49
	← d3vic3 left	14:49
ddenis	how can I display a content of a merge conflict? git show on the merge shows the whole change, but I want to see only my conflict resolution	14:49
	→ d3vic3 joined	14:50
Xeon06_	_ikke_: Oh yes, okay git log says merge and the two hash tags	14:51
	→ tty_ joined	14:51
	← tty_ left	14:51
	← pfFredd left	14:52
zokko	guuys	14:52
	← juanfer left	14:52
zokko	ive successfully push-ed all branches. how can i push tags?	14:52
	→ juanfer joined	14:52
FauxFaux	I'll give you one guess.	14:52
	→ pfFredd joined	14:53
	→ tty_ joined	14:54
tty_	hi. i am running gentoo linux on a web server. i want to use this server also for hosting code/etc with git. 2-3 people will access/mantain the code/etc. i have installed git and i am trying to set up multiuser access. i keep getting "Exception caught during execution of ls-remote command" when trying to import to eclipse. i have also created a new linux user but i see no need for using git? am i wrong? i've also read somewhere about using ssh keys. any	14:54
	on this? thanks	14:54
_ikke_	Xeon06_: Is that the last commit?	14:54
zokko	erm?	14:54
	→ Radium joined	14:54
	shruggar1 → shruggar	14:54
shruggar	tty_: !gitolite	14:55
gitinfo	tty_: Want to host as many git repos (and users!) as you like, on your own server, with fine-grained access control? You want gitolite: https://github.com/sitaramc/gitolite - Documentation: http://sitaramc.github.com/gitolite/master-toc.html	14:55
	→ XOXO1 joined	14:55
	→ tfcoding joined	14:55
Xeon06_	_ikke_: yep! But it's been pushed	14:56
	→ stkrzysiak joined	14:56
_ikke_	Xeon06_: Well. To undo it, you can do git reset --keep HEAD^	14:57
	→ ph^ joined	14:57
zokko	for i in `git branch -r\|awk -F "/" '{print $2}'`; do git push origin tags/$i:refs/tags/$i; done	14:57
_ikke_	Xeon06_: Then git push -f to force it onto the remote	14:57
zokko	doesnt work, how should i push those tags? :/	14:57
_ikke_	git push -f origin master	14:57
	→ thirdknife joined	14:58
	← Sitethief left	14:59
	← coyotebush left	14:59
	← pyhead left	14:59
	→ marduk joined	15:01
	← Xeon06_ left	15:01
	← marduk left	15:01
	→ whyking_ joined	15:01
	→ bizarrobrian joined	15:02
	← f0i left	15:02
	← mneorr left	15:02
	→ Xeon06_ joined	15:02
	→ joaoh82 joined	15:02
	← juanfer left	15:03
	→ juanfer joined	15:03
Xeon06_	_ikke_: I don't think it worked. Now when I do git log I see all the commits from my other branch that I had merged	15:03
tty_	anyone knows how to have a multiuser git server on linux? read/write access.	15:04
t4nkd	If I did $git checkout -b my_local origin/her_remote -- made some changes, committed those, how do I push my changes back to origin/her_remote	15:04
	← whyking_ left	15:05
FauxFaux	git push origin her_remote	15:05
	← hobbyist left	15:05
t4nkd	cool ty	15:05
PerlJam	tty_: Just use gitolite.	15:06
t4nkd	I had to do `git push origin origin/her_remote' -- that sounds right?	15:06
FauxFaux	No.	15:06
t4nkd	Your command resulted in 'src refspec her_remote does not match any.'	15:07
PerlJam	t4nkd: you needed "git push origin my_local" originally I do believe	15:07
FauxFaux	I really have no idea what the default settings do, do I. Apart from the fact that I clearly have no idea what I'm talking about; git push origin my_local:her_remote # will probably work, and your command will probably have done nothing. But, y'know.	15:08
	THE DEFAULTS ARE TERRIBLE.	15:08
	→ ShooterMG joined	15:08
bizarrobrian	hi guys, we are migrating our repos to a different server...is there an easy way to force everyone's remote urls to point to the new server? (outside of a dns redirect which we can't do)	15:08
PerlJam	bizarrobrian: tell them to update their urls	15:09
t4nkd	Sorry, I'm not implying that you don't know what you're talking about, really	15:09
	← beautifulmind left	15:09
	→ hotspants joined	15:09
bizarrobrian	PerlJam: yea, I just didn't know if I could push it out rather than having them do it manually	15:09
	→ ShooterM_ joined	15:09
	→ sq-one joined	15:09
FauxFaux	t4nkd: No, I was serious. The default for what that does is being changed, because it confuses everyone (especially me). I really shouldn't answer questions about push as I changed it ages ago and have completely forgotten how to use the old way.	15:09
	← ShooterMG left	15:10
FauxFaux	man git config push.default tracking	15:10
gitinfo	the git-config manpage is available at http://jk.gs/git-config.html	15:10
t4nkd	Oh, true that, well damn	15:10
	→ hobbyist joined	15:10
	→ alberto56 joined	15:11
	← ThomasLocke left	15:11
	← rockets left	15:11
	→ beatak joined	15:11
	tizzo → tizzo-afk	15:11
	← tty_ left	15:11
	→ mikepack joined	15:12
	tizzo-afk → tizzo	15:13
	← iocor left	15:13
	← juanfer left	15:13
	← cbreak-work left	15:14
	← milki left	15:14
	→ milki joined	15:14
	→ qt-x joined	15:15
	→ ceej joined	15:15
	→ mklappstuhl joined	15:16
	dwmw2_gone → dwmw2	15:16
	→ chrisshattuck joined	15:17
	← TomyLobo left	15:17
	→ juanfer joined	15:17
t4nkd	Well, shit, I just changed the push.default to tracking	15:18
	And that seemed to be fine	15:18
	→ jaimedp joined	15:18
	← el2ro left	15:18
	← wbednarski left	15:19
	← Xeon06_ left	15:19
	← digiv left	15:21
	→ ehks joined	15:21
	← flijten left	15:21
	→ leavittx joined	15:21
	← joaoh82 left	15:22
	← hobbyist left	15:22
	← pppaul left	15:22
	→ pppaul joined	15:23
	→ hyperair joined	15:24
	→ Nightfire034 joined	15:26
	← rchavik left	15:27
	→ kenperkins joined	15:28
	← _sundar_ left	15:28
	→ traviscj joined	15:30
	← pmatthews left	15:32
	→ hagabaka joined	15:33
	← hagabaka left	15:33
	→ hagabaka joined	15:33
	← arvind_k left	15:34
	← babilen left	15:34
	→ troydm joined	15:35
troydm	ok git status show following:	15:36
	modified: bundle/vundle (untracked content)	15:36
	now how do i untrack that bundle/vundle	15:36
	which by itself contains another repository	15:36
	doing git rm bundle/vundle says error: Not owner	15:36
	→ kukks joined	15:37
	← ajaykools left	15:38
	← tsdh left	15:39
	→ yoklov joined	15:39
	→ cakehero joined	15:39
	→ duckx joined	15:40
troydm	ok found it git rm --cached	15:40
	it helped	15:40
	→ neurodrone joined	15:42
	← axl_ left	15:42
	← duckxx left	15:42
	← pppaul left	15:42
	← juanfer left	15:42
	← ipalaus_ left	15:42
	→ pppaul joined	15:43
	→ orkz joined	15:43
	← Poapfel left	15:44
	← troydm left	15:44
	→ Poapfel joined	15:44
	→ juanfer joined	15:45
	→ axl_ joined	15:45
	← luke__ left	15:47
	← kpreid left	15:49
	← brngardner left	15:50
	→ _iron joined	15:50
	→ Tocacar joined	15:53
	macabre^ → macabre	15:55
	→ wbednarski joined	15:56
	← dhruvasagar left	15:57
duncanm	hey offby1, it's been a long time	15:57
	→ williamcotton joined	15:58
	→ dhruvasagar joined	15:59
	→ kpreid joined	15:59
	→ joaoh82 joined	16:00
	→ digiv joined	16:01
	← angelsl left	16:01
	digiv → digiv\|meeting	16:01
	← juanfer left	16:02
bigkitty	how to install git in centos?	16:02
	should i build it from source?	16:02
	← berserkr left	16:02
bigkitty	what is the best solution to install git for centos?	16:03
FauxFaux	Ask the centos people. It's trivial to build from source if you have standard libraries and perl available, though. Possible without.	16:03
	→ ziggles joined	16:03
	→ juanfer joined	16:03
	→ Zenith77 joined	16:04
	← robotmay left	16:05
	← pro_metedor left	16:05
	← jaimedp left	16:05
	→ pyhead joined	16:05
	→ pro_metedor joined	16:05
	← pro_metedor left	16:05
	← overbythere left	16:05
	← kevwil left	16:06
	← thirdknife left	16:07
	← tfcoding left	16:07
	← spaceonline left	16:07
	← cakehero left	16:08
	→ airborn_ joined	16:09
	← envi_ left	16:09
	← eeadc left	16:10
	← nb left	16:10
	tizzo → tizzo-afk	16:11
	→ envi_ joined	16:11
	← ShooterM_ left	16:12
	← mishok13 left	16:12
	→ ShooterMG joined	16:12
	→ SimonNa joined	16:14
	← XOXO1 left	16:14
	→ cccaldas_ joined	16:15
	→ glcrazy joined	16:16
	→ nb joined	16:16
	→ iocor joined	16:16
	← psoo left	16:16
	← pppaul left	16:17
	← dc5ala left	16:17
	→ pppaul joined	16:18
	→ ipalaus_ joined	16:18
	← apok left	16:18
	← cccaldas left	16:19
	← gavin_huang left	16:20
	→ soulcake joined	16:20
	→ mritz joined	16:20
offby1	la la la	16:21
	duncanm: that's all I know how to say	16:21
	← parcs` left	16:22
	← pppaul left	16:23
	← aristidesfl left	16:24
	→ pppaul joined	16:24
	← mklappstuhl left	16:25
	→ macmartine joined	16:25
	→ fisted_ joined	16:25
	→ SkAzZ joined	16:26
	← lausi left	16:26
	← fisted left	16:26
	→ babilen joined	16:26
	← pppaul left	16:27
	→ pppaul joined	16:27
	← reels left	16:28
	→ Carraway joined	16:28
	← giallu left	16:28
	→ notola joined	16:28
	← ipalaus_ left	16:28
	→ coyotebush joined	16:28
	→ ipalaus_ joined	16:30
	← achadwick left	16:31
	→ felgru joined	16:31
	← master-lincoln left	16:32
	→ lausi joined	16:32
	← pppaul left	16:32
	→ pppaul joined	16:32
	→ freeed joined	16:33
	← williamcotton left	16:33
	← ipalaus_ left	16:33
	← Desproges left	16:34
	→ SegFaultAX\|work joined	16:34
	← jceb left	16:35
	→ parcs` joined	16:35
	→ apok joined	16:35
	→ troulouliou_dev joined	16:36
	← alberto56 left	16:36
	← noisebleed left	16:36
troulouliou_dev	hi "git push origin :mybranch" , is the real way to delete a remote branch on the public repo ?	16:37
	→ Araxia joined	16:37
	← w0ng left	16:37
	→ w0ng joined	16:37
FauxFaux	troulouliou_dev: Or git push --delete origin mybranch, if you'd prefer.	16:37
	→ n1x joined	16:37
	← Samual left	16:38
	→ codewrangler joined	16:38
	← teotwaki left	16:38
troulouliou_dev	FauxFaux, and if i deleted a branch localy with git -d -r origin/mybranch ; is there a command to push this action ?	16:39
	→ mklappstuhl joined	16:39
	← matthiaskrgr left	16:39
troulouliou_dev	git branch -d -r origin/mybranch i mean	16:39
FauxFaux	Don't fiddle with things inside origin/, it'll just confuse things.	16:39
	← stkrzysiak left	16:40
	← netiru left	16:40
	→ stkrzysiak joined	16:41
	← yoklov left	16:42
	→ dzonder joined	16:42
troulouliou_dev	ok	16:43
	thanks FauxFaux	16:43
	→ thelazydeveloper joined	16:43
	← pyhead left	16:45
	← mikewintermute left	16:45
	← Radium left	16:45
	→ Per4u3e joined	16:45
	→ Samual joined	16:45
	→ ins\ joined	16:46
	→ dreiss joined	16:46
	→ matthiaskrgr joined	16:47
	← kpreid left	16:48
	→ usmanity joined	16:49
	→ rook2pawn joined	16:49
rook2pawn	i added Thumbs.db to my .gitignore and it a /img/foo/Thumbs.db still shows up in git status	16:50
	→ storrgie joined	16:50
rook2pawn	i then created a .gitignore_global and added Thumbs.db there, and git status still showed this file	16:50
Evanlec	gentlemen, whats the proper procedure for setting up a git repo where one user is desginated as "administrator" and must sign-off on commits before they're allowed to be merged ?	16:50
	← storrgie left	16:50
cortana2	rook2pawn, did you previously commit it? if so you have to tell git to stop tracking it (git rm it then git commit)	16:50
Evanlec	rook2pawn: if the file was being tracked before you made the .gitignore then it will still be tracked	16:50
rook2pawn	Evanlec: simply make that users's branch the one to go to production with or pre-production	16:51
Evanlec	rook2pawn: thus, you have to do what cortana2 said	16:51
rook2pawn	Ah ok	16:51
cortana2	Evanlec, the exact procedure depends on your requirements. where i work each developer has a git repo and pushes their work out in branches, asks for a review, and i merge and push to the main repo if acceptable	16:51
Evanlec	rook2pawn: whats to stop someone from making commits to that branch?	16:51
rook2pawn	Evanlec: so what im saying is suppose you have 3 developers, one of whom has the final word	16:51
	Evanlec: each developer has his own branch of each other's repos	16:52
Evanlec	cortana2: yes thats the kind of setup I want	16:52
rook2pawn	when each dev makes a change, eachh other developer can pull those changes into their local branch of that developers changes	16:52
cortana2	but you could have one repo for review requests and have everyone push to different branches there, rather than each developer having their own repo, for instance	16:52
	there are a lot of variations on this theme	16:52
	→ GodEater joined	16:52
	→ kpreid joined	16:52
	← GodEater left	16:53
	→ GodEater joined	16:53
Evanlec	humm	16:53
	← w0ng left	16:53
cortana2	hm. are you asking about how you can _prevent_ another user from pushing to your main repo?	16:53
Evanlec	i'd prefer one repo	16:53
	→ storrgie joined	16:53
rook2pawn	then when the "master" dev makes teh call, he/she can pull those changes from the local branch of an approved dev's work into his own main branch	16:53
Evanlec	with branches for each user...	16:53
	cortana2: yea i guess	16:53
cortana2	ok well... if you use ssh to contact your repos then it's a matter of setting up filesystem permissions so that only you can write to the main repo	16:53
	→ kevwil joined	16:53
	→ robotmay joined	16:53
	→ w0ng joined	16:53
Evanlec	okay, so i should use unix rights instead of gitosis or gitolite?	16:54
cortana2	there's a thing called 'gitolite' which exists to make this easier... you can specify who can read/write to various repos by specifying it in a file rahter than having to screw around with the filesystem permissions manually	16:54
	← storrgie left	16:54
	← orkz left	16:54
rook2pawn	Evanlec: you could delete the branch or screw it up, it doesnt matter, the "master" dev wont have to pull that branch into his own	16:54
cortana2	or is it called gitosis these days?	16:54
Evanlec	cortana2: how do i make it where you can see the "signed-off by: somebody" in the commit logs	16:54
cortana2	ah	16:54
	← francisfish left	16:54
cortana2	i don't use that	16:54
	← gusnan left	16:54
rudi_s	cortana2: !gitosis	16:54
gitinfo	cortana2: 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	16:54
Evanlec	the "master" branch is on the server	16:55
	i mean	16:55
cortana2	thanks rudi_s	16:55
rudi_s	np	16:55
Evanlec	thats where the repository is hosted, from a bare git repository in a repo.git dir	16:55
	→ Kaisyu joined	16:55
	→ francisfish joined	16:55
rook2pawn	ok, so say you have 3 developers, 1 of which is the authority, call them A, B, C where C is the "authority dev"	16:55
Evanlec	okay, gitolite then	16:55
cortana2	Evanlec, but i think when you commit you add -s and it adds the message. perhaps you do that when you commit the merge	16:55
	it would imply you use --no-ff too	16:55
rook2pawn	A has a branch of B and C. B has a branch of A and C. C has a branch of A and B	16:55
Evanlec	rook2pawn: yea, that is our setup, 3 people, 2 are developers , the 3rd is company owner -- but not a programmer -- he just wants to be able to approve changes before they go live	16:56
	← jlebrech left	16:56
rook2pawn	Yes, so C simply git pull origin "devA branch"	16:56
	→ aristidesfl joined	16:56
	← pppaul left	16:56
rook2pawn	then the changes A made to his branch get placed on C's plate	16:57
Evanlec	rook2pawn: from C's local machine?	16:57
rook2pawn	yes	16:57
	then if C likes it	16:57
Evanlec	but then C still has to push it	16:57
	← aristidesfl left	16:57
	→ pppaul joined	16:57
	← Per4u3e left	16:57
rook2pawn	what	16:57
	C has to push what	16:57
	→ Mowsh joined	16:57
cortana2	Evanlec, it's possible to set it up so that A can fetch from C's local repo on his/her machine	16:57
Evanlec	the branch	16:57
	err	16:57
	the commit, idk	16:57
	← apok left	16:58
cortana2	but i think it's more usual that C would push their work to their public repo for review	16:58
troulouliou_dev	is it possible to pull from origin/master to all the local branch in one command ?	16:58
cortana2	this workflow is explained in the git user guide i think	16:58
	← revagomes left	16:58
Evanlec	that seems kinda silly though, we dont want access to eachother's local machine, only the server	16:58
	→ apok joined	16:58
rook2pawn	if C likes the cahnges of A, then C can pull that branch after inspecting A on his local (C) machine	16:58
	← _inc_ left	16:58
	← sdeboey left	16:59
Evanlec	rook2pawn: C pulling branches to his local machine does not make them live on the server (web and git server)	16:59
cortana2	Evanlec, see chapter 4 of http://schacon.github.com/git/user-manual.html#public-repositories	16:59
	→ macabre^ joined	16:59
	→ stkrzysiak_ joined	16:59
	← stkrzysiak_ left	16:59
	← macabre left	16:59
	← francisfish left	16:59
	← stkrzysiak left	17:00
	← dzonder left	17:00
	→ stkrzysiak joined	17:00
Evanlec	cortana2: hmm, dont want this to be public	17:00
rook2pawn	Evanlec, that's right, obviously your deploy machine will have its own git repo that simply pulls	17:00
Evanlec	only accessible to users with SSH acces	17:00
cortana2	Evanlec, it doesn't mean "public" like that unless you set it up like that	17:01
rook2pawn	and the deploy machine is linked to pull from C, and you pull that switch when you want to	17:01
Evanlec	rook2pawn: on the single server we have, the master repo is in a bare repo dir (project.git/), and there's a working copy clone that is the code in production (via web)	17:01
cortana2	e.g., here on the server we have /srv/repos/project/integration.git (which is the main repo) and /srv/repos/project/$USER.git (which each user pushes to when they want their work to be merged). see/	17:02
Evanlec	cortana2: oh, each user has their own repo...hmm	17:02
cortana2	filesytem permissions mean that each USER repo is only writable by the user, and the integration repo is only writable by me and the other people who review	17:02
	← pppaul left	17:02
cortana2	yeah. you could just have one repo for reviews, and users push to different branches	17:02
rook2pawn	the chapter you should read is book.git-scm.com/3_distributed_workflows.html	17:02
Evanlec	hmm	17:03
cortana2	IME that requires more from users because they have to actually do their work on different branches and not step on each other's toes	17:03
	→ samba_ joined	17:03
	→ pppaul joined	17:03
Evanlec	well there's like I said, only 3 people right now using git on this project	17:03
rook2pawn	the whole point is that people get updates from other people but those updates dont go their own work, they go to a local named branch (remotes)	17:03
cortana2	yeah but that's no reason not to set it up properly :)	17:03
rook2pawn	3 people is PLENTY to do it this way	17:03
Evanlec	right	17:03
cortana2	yeah. it's never too early to get your infrastructure sorted out properly	17:04
Evanlec	"this way" meaning seperate repos, or seperate branches for each user?	17:04
rook2pawn	and then after they inspect it on their own machine they can merge that result back into their own local master	17:04
	basically for 3 dev's we are talking about 6 total branches	17:04
	→ kadoban joined	17:04
rook2pawn	4 devs (a situation i had last year) we had 12 branches	17:05
Evanlec	and 1 repository?	17:05
rook2pawn	4 repositories.	17:05
	3 were forks	17:05
Evanlec	no, not for 4	17:05
rook2pawn	1 was master	17:05
Evanlec	for 3	17:05
	→ JohnBat26 joined	17:05
rook2pawn	oh for 3 its 1 master, 2 forks	17:05
Evanlec	3ppl you had 3 repositories on server?	17:05
rook2pawn	on github	17:05
	→ cakehero joined	17:05
Evanlec	hmm	17:05
	→ BSaboia__ joined	17:05
Evanlec	we'r not using github at the moment...	17:05
rook2pawn	basically if you have k devs, each dev has k-1 branches	17:05
Evanlec	i could try and pitch that but	17:06
	→ yoklov joined	17:06
	← pfFredd left	17:06
Evanlec	right now its just git over ssh to our web/git/etc server	17:06
cortana2	IMO that is overcomplicating it... i didn't like people pulling from each other unless they were actaully working together on the same thing	17:06
rook2pawn	cortana2: yes we had 48 hours in last year's node knockout with 4 devs on the same thing	17:07
Evanlec	yes simple is better, especially since the owner is brand new to git	17:07
rook2pawn	every minute counted	17:07
cortana2	it results in them incorporating some other random developer's work without that work necessarily being ready	17:07
rook2pawn	we were on IRC and i would be like OK guys check out my branch	17:07
cortana2	Evanlec, check out those two links... once you've read them then these ideas should seem much less complex	17:07
Evanlec	rook2pawn: development with this project is not nearly that fast paced	17:07
	→ dwmw2_gone joined	17:07
rook2pawn	they would checkout my branch and when the leader liked it, and then made a change, the other devs pulled from him, and by transitivity, got my changes	17:07
Evanlec	cortana2: k	17:07
	← coyotebush left	17:07
	← w0ng left	17:08
rook2pawn	the nice thing is that changes are trasitive, if C likes B and then C also makes his own change	17:08
Evanlec	rook2pawn: did these people checkout your branch over ssh? you gave them ssh access or using git daemon?	17:08
rook2pawn	then A pulls from C, A now has B's work	17:08
	← dwmw2 left	17:08
	← notola left	17:08
rook2pawn	no, everyone connected to github	17:08
	and obviously its all free	17:08
Evanlec	Oh right	17:08
	humm	17:09
	→ w0ng joined	17:09
	← BSaboia left	17:09
rook2pawn	it literally took me half an hour to get this setup	17:09
	→ coyotebush joined	17:09
rook2pawn	git remote add and read that chapter	17:09
Evanlec	boss wouldnt want this on public github, but might shell out $ for private github	17:09
rook2pawn	yeah, i pay $8 or so for 5 private repos	17:09
Evanlec	rook2pawn: k	17:09
	rook2pawn: not bad	17:09
	← jameswarren left	17:09
	→ TomyLobo joined	17:10
	← tyatpi_ left	17:10
	rook2pawn thinking of making a 5 minute video on how to set this all up	17:10
	← segher left	17:10
	← jrots left	17:11
	← w0ng left	17:11
	→ revagomes joined	17:11
Evanlec	rook2pawn: github docs dont explain this?	17:11
	→ brngardner joined	17:12
rook2pawn	they do but i think videos explain it without having to read? it just makes it easier perhaps	17:12
	and to also realize the sheer power this way has over every other setup	17:12
Evanlec	rook2pawn: yea, you should do that	17:12
rook2pawn	the other setups actually don't make ANY sense now , i mean, 0 sense	17:13
	← pppaul left	17:13
	→ segher joined	17:13
	← sq-one left	17:13
Evanlec	rook2pawn: is this what i should be reading/following? http://book.git-scm.com/3_distributed_workflows.html	17:13
	→ pppaul joined	17:13
	→ w0ng joined	17:13
	→ FernandoBasso joined	17:13
rook2pawn	yes	17:13
Evanlec	k	17:13
rook2pawn	especially the part after "small closely knit group"	17:14
	← segher left	17:14
rook2pawn	that is the crux of it	17:14
	→ melvinram joined	17:14
	→ FACEFOX joined	17:14
	← FACEFOX left	17:14
	→ segher joined	17:14
rook2pawn	learn git remote and its options so you can successfully see you are creating remotes that are actually hooked up and wired	17:15
	and git merge is essential here as well	17:15
	→ FACEFOX joined	17:15
	← FACEFOX left	17:15
	← LucasTizma left	17:15
Evanlec	rook2pawn: okay, but lets say we dont have github, how can I pull from dev B's local repo	17:15
	?	17:15
rook2pawn	i have no idea, sorry	17:15
Evanlec	have him run git daemon	17:15
	damnit	17:15
melvinram	I have a git repo, I made some changes and did git add . . What I actually needed to do was stash my changes, create a new branch, unstash and commit into that branch. How do I undo my git add . ?	17:15
	→ LucasTizma joined	17:16
	→ elijah joined	17:16
	← stkrzysiak left	17:16
Evanlec	melvinram: git reset i think	17:16
rook2pawn	git reset HEAD	17:16
Evanlec	or that	17:17
	→ machine2 joined	17:17
melvinram	will that remove the changes I staged?	17:17
PerlJam	Evanlec: you can use local file system operations, direct ssh, or you can setup gitolite (for instance)	17:17
melvinram	I don't want to loose my changes.	17:17
	I just want it to be unstaged	17:17
	will git reset HEAD do that?	17:18
PerlJam	melvinram: just create the new branch and commit the changes already in the index to that branch	17:18
	→ ISF joined	17:18
PerlJam	melvinram: git checkout -b foo; git commit; # done.	17:18
Evanlec	PerlJam: local file system operations? I dont think that would work...	17:18
	PerlJam: and i dont think that dev B wants to give me SSH access to his personal machine	17:19
PerlJam	Evanlec: personally, I like gitolite.	17:19
Evanlec	PerlJam: alright i'll look into gitolite then, it solves this issue?	17:19
melvinram	PerlJam: I can do that? I can just stage a file in one branch and commit in a new branch?	17:19
	→ joebodo_ joined	17:19
Evanlec	PerlJam: yea, are you sure git will let him do that since his index is dirty?	17:20
PerlJam	melvinram: unless I've missed or misunderstood something, yes.	17:20
	→ alberto56 joined	17:21
PerlJam	melvinram: you don't stage anything "in a branch"; It's just staged for commit. Which branch you commit to can be changed by checking out the appropriate branch	17:21
melvinram	interesting	17:21
	that is pretty cool	17:21
	← bgerber left	17:21
	→ dr_win joined	17:22
	← cortana2 left	17:22
Evanlec	PerlJam: yea thats neat, i did not know one could do that	17:22
	PerlJam: if you have unstaged changes, it will complain if you try switch branches then won't it?	17:23
	← segher left	17:23
PerlJam	Evanlec: yes.	17:23
Evanlec	kk	17:23
PerlJam	Evanlec: but that's what stash is for :)	17:23
	← iocor left	17:23
Evanlec	righto	17:23
duncanm	offby1: heh	17:23
Evanlec	I usually use stash for that	17:23
joebodo_	I have forked an opensource project and have made changes that the project owner would like to take. I also have personal changes that I would like to keep. How can I manage both personal and project changes. Should I rebranch my own copy and make changes there?	17:23
	→ eeadc joined	17:23
Evanlec	but, staging and creating new branch + commit is an alternative that I didnt know about	17:23
	← ISF left	17:24
	← revagomes left	17:24
PerlJam	Evanlec: yeah, it's common because of confusion around branches	17:24
	→ sdeboey joined	17:24
Evanlec	PerlJam: yea, git's lightweight branches are awesome	17:24
	→ oohahh joined	17:24
	→ frnkkpp joined	17:24
PerlJam	thinking things like "I'm in a branch" perpetuates the confusion.	17:25
Evanlec	PerlJam: I tried mercurial for a while, and realized it totally lacked lightweight branching like git and I said F this	17:25
	→ hobbyist joined	17:25
Evanlec	to branch in mercurial you have to make an entire copy of the tree	17:25
	→ tyatpi_ joined	17:25
PerlJam	yep, just like svn	17:26
Evanlec	which seems stupid to me, copying directories is what we did before version control systems	17:26
	← sdeboey left	17:26
PerlJam	(I assume hg has the same optimization that svn has so that it doesn't actually copy the whoel thing)	17:26
Evanlec	version-control is designed to eliminate this clutter	17:26
	PerlJam: oh ok, so yea at least you're not wasting disk space, but still adding clutter	17:26
	← yoklov left	17:26
	→ yoklov joined	17:27
	← shruggar left	17:27
	gitinfo set mode: +v	17:27
	→ dzonder joined	17:27
Evanlec	I love it, our codebas is filled with index.php, index_before_JK.php, index_old.php, index_blah.php type filename copies	17:27
	makes me cringe	17:27
	→ mikewintermute joined	17:27
Evanlec	and want to cry	17:27
	← mikewintermute left	17:27
PerlJam	yeah, I still can't believe people develope that way.	17:28
Evanlec	yea, its appauling, probably i wasnt suprised because its php ;p	17:28
	→ aristidesfl joined	17:28
	→ superdmp joined	17:29
Evanlec	the boss then decided he wanted to approve changes before they go live, so he made a dev/ folder and staging/ folder	17:29
	basically just more copies of the codebase	17:29
	← LucasTizma left	17:29
Evanlec	ugh	17:29
	← tatsuya_o left	17:29
	← tktiddle left	17:29
	← tvw left	17:29
superdmp	I have made a new branch locally, and some changes in it that I have committed; now I wish to push them to github	17:29
PerlJam	at least he was trying to move in the right direction even if he was a little misguided	17:29
	← marcogmonteiro left	17:30
superdmp	of course github doesn't know yet about my branch	17:30
	→ tatsuya_o joined	17:30
superdmp	so I can't just push, can I?	17:30
PerlJam	superdmp: sure you can, you just have to be a little more verbose about it	17:30
superdmp	I have to tell git that I need a new remote branch- correct?	17:30
Evanlec	PerlJam: yea, i explained git to him, sent him a video and some reference material, he seems to be quite interested now thank god	17:30
PerlJam	superdmp: "git push -u origin branch"	17:30
	Vampire0_ → Vampire0	17:30
superdmp	where 'branch' is the name of the new branch?	17:31
	gitinfo set mode: +v	17:31
Evanlec	superdmp: yar	17:31
ntelford	is there a way to revert all the commits from one branch that exist within another? i.e. "revert all intersecting commits"?	17:31
superdmp	fantastic, thanks all	17:31
PerlJam	superdmp: after that, a plain "git push" should work	17:31
superdmp	this must be the most immediately helpful IRC channel of all the ones I use	17:32
PerlJam	ntelford: that's an odd request. What happened? :)	17:32
Evanlec	ntelford: git doesnt really work like that, those commits aren't copied between branches (duplicated)	17:32
	← Fullmoon left	17:32
	← Nightfire034 left	17:32
ntelford	PerlJam, someone merged a branch to master that wasn't ready to be merged and we need master to be clean so we can apply hot-fixes	17:32
Evanlec	ntelford: commits shared by branches are simply commits before the branch was made	17:32
PerlJam	ntelford: was it a fast-forward merge?	17:33
ntelford	PerlJam, I don't believe so	17:33
	← zivester left	17:33
	← hobbyist left	17:33
PerlJam	ntelford: then you can reset master to just before the merge commit	17:33
kevlarman	ntelford: you can revert the merge commit	17:33
	but then future merges won't bring in those commits	17:33
ntelford	PerlJam, what about all the other merges/commits that have taken place since then?	17:33
Evanlec	revert or reset, either one	17:33
PerlJam	ntelford: if it was a fast foward merge, you can do the same thing, but you'll have to check th reflog for the right commit.	17:33
ntelford	hmm	17:33
	← tatsuya_o left	17:34
PerlJam	ntelford: Hmm.	17:34
Evanlec	ntelford: oh, well in that case you'll have to cherry-pick those new commits into a temporary branch first i think	17:34
kevlarman	ntelford: to remerge it you'll have to revert the revert commit, then do another merge	17:34
	← Superlokkus left	17:34
	← PigeonFriend left	17:34
Evanlec	I dont know actually	17:34
	→ Superlokkus joined	17:35
PerlJam	ntelford: start by reverting the merge commit and see where you need to go from there.	17:35
	→ Radium joined	17:35
	→ tewecske joined	17:35
	→ pikpik joined	17:36
	← mister_m left	17:36
	→ phishy joined	17:36
	→ LucasTizma joined	17:38
	→ berserkr joined	17:38
charon	ntelford et al: after reverting a merge, you'll probably want to rebase the topic (that was merged earlier) to retain your sanity. otherwise git will think the topic was already merged, and get all confused	17:38
	→ jrots joined	17:39
	→ williamcotton joined	17:39
	→ Russell^^ joined	17:40
cmn	see http://opensource.apple.com/source/Git/Git-26/src/git-htmldocs/howto/revert-a-faulty-merge.txt	17:40
charon	heh. i wanted to make a "but but but apple is evil" comment, but it's actually an email from linus.	17:41
	→ FACEFOX joined	17:41
kevlarman	charon: just reverting the merge before re-merging seems easier	17:41
	kevlarman has some really rebase-happy coworkers	17:41
charon	kevlarman: my point is that after the revert, git will be confused.	17:41
cmn	yeah, it's the first link that comes up when you ask google for the title	17:41
charon	trust me, or go try.	17:42
	kevlarman: it's different if you make the merge disappear (e.g. with reset), but reverting it does not change the effects the merge has on topology, which causes problems later.	17:42
kevlarman	charon: such as?	17:42
PerlJam	charon: good point	17:42
charon	assuming topic has not been developed since the merge, every commit in topic is now an ancestor of master	17:42
	→ pantsman joined	17:43
	← pantsman left	17:43
	→ pantsman joined	17:43
charon	so a subsequent merge will think it's fully merged already, and do nothing	17:43
	if you throw out the assumption it gets even funkier	17:43
kevlarman	charon: that's what the second revert is for	17:43
charon	with merge trying to apply only the later half of the history	17:43
kevlarman	right before a second merge	17:43
	→ mysphyt joined	17:43
charon	kevlarman: merge only considers the heads you give it, and the merge-base. merge-base only considers the shape of history, it completely ignores reverts	17:44
	← magnus_ left	17:44
	→ iocor joined	17:44
kevlarman	charon: i know	17:44
	→ arti_t joined	17:45
	← fo_x86 left	17:45
	→ magnus_ joined	17:45
kevlarman	charon: so you have a bad merge, then a revert, then more work on both branches, then a revert of the original revert commit, then a second merge	17:45
Evanlec	charon: what do you mean by 'rebase the topic' what is the topic referring to?	17:45
	← dr_win left	17:46
charon	Evanlec: the branch that was merged	17:46
superdmp	what is it called when I want to checkout from the remote an old version of the code, for a particular commit?	17:46
	→ dr_win joined	17:46
charon	kevlarman: ah, i see. ok, that would work. perhaps the history ends up being somewhat confusing, but meh.	17:46
Evanlec	charon: ok	17:47
FauxFaux	superdmp: "git checkout nameofcommit"; you'll !detach head	17:47
gitinfo	superdmp: [!detached] A detached HEAD occurs when your HEAD does not point at a branch. New commits will NOT be added to any branch, and can easily be !lost. This can happen if you a) check out a tag, remote tracking branch, or SHA; or b) if you are in a submodule; or you are in the middle of a c) am or d) rebase that is stuck/conflicted. See !reattach	17:47
	← alexisabril left	17:47
	← krosenbl left	17:47
	← Kaisyu left	17:47
	→ ghvandoorn joined	17:47
superdmp	so pointing at a branch means pointing at the latest version of that branch? rather than just being anywhere within it?	17:48
	frnkkpp → frnkkpp_on_drugs	17:48
cbreak	a branch is a label for one singular commit	17:48
charon	superdmp: it literally refers to the branch name, as opposed to a specific commit by sha1	17:48
ntelford	is it possible to have a branch that looks like it pushes to two remote branches?	17:48
	→ krosenbl joined	17:48
FauxFaux	ntelford: Nope.	17:48
superdmp	ok, thank you	17:48
	→ FreeArtMan joined	17:49
mysphyt	Hey, folks. My post-update hook doesn't seem to be firing on a remote repo after I push to it. It's named post-update and is readable and executable to all. Any guesses where to start looking for the problem?	17:49
ntelford	hmm	17:49
	→ jameswarren joined	17:49
	← mischat left	17:50
	← pantsman left	17:50
	→ bgerber joined	17:50
	← ajpiano left	17:51
EugeneKay	mysphyt - insert an 'echo "This is the remote post-update hook"' at the top of the hook. It might be running, but exiting poorly.	17:51
	→ n3m joined	17:51
	← dvaske left	17:51
	→ ajpiano joined	17:51
	← rurufufuss left	17:52
	← mercutio22 left	17:52
	→ voidx joined	17:52
	← toothr left	17:52
	→ sdeboey joined	17:52
	← phishy left	17:52
charon	superdmp: note that a branch itself is really a pointer to its tip ("newest") commit, so if HEAD points at a branch, then it can also be used to refer to that commit in contexts where a commit can be specified (i.e. almost everywhere)	17:53
	← w0ng left	17:53
mysphyt	EugeneKay--ha! I left the default command in the script, and it looks like that's causing it to fail. I had my debug message right after that, assuming it would be fine. Which, of course, it wasn't.	17:53
	→ mercutio22 joined	17:53
mysphyt	I always seem to find I'm just a bit dumber than I expect. :P	17:53
superdmp	charon: cheers	17:53
	← magnus_ left	17:53
	→ mcluskydodallas joined	17:53
EugeneKay	"You can try to make a thing foolproof, but they will always build a better fool"	17:53
	→ w0ng joined	17:54
	→ dstorrs joined	17:54
	→ magnus_ joined	17:54
mysphyt	EugeneKay Now I just have to figure out why "exec git update-server-info" kills the script--and, in fact, killed my SSH session somehow when I ran it directly.	17:54
bremner	exec	17:55
charon	mysphyt: that's because exec by definition replaces the calling process with the callee	17:55
	← kpreid left	17:55
mysphyt	charon--as `man exec` just showed me. Again. Always just a little stupider than I'd expected. sigh	17:55
	← michaelmartinez left	17:55
EugeneKay	:-D	17:55
	→ Nightfire034 joined	17:55
	→ lacrymology joined	17:56
	→ slide joined	17:56
	← pppaul left	17:57
	← magnus_ left	17:57
	← marhaban left	17:58
	← slide23 left	17:59
	→ pppaul joined	18:00
	→ magnus_ joined	18:01
	← cakehero left	18:01
	← melvinram left	18:02
	← mercutio22 left	18:03
	← dr_win left	18:05
	→ mercutio22 joined	18:05
	→ dr_win joined	18:05
	→ SkAzZ1 joined	18:06
	← hyperair left	18:07
	→ fission6 joined	18:07
	← voidx left	18:07
	→ shruggar joined	18:07
	→ stkrzysiak joined	18:07
	← SkAzZ left	18:07
	← stkrzysiak left	18:07
	→ jceb joined	18:07
	→ b1lly joined	18:10
	→ rockets joined	18:10
	← dr_win left	18:11
	← kenperkins left	18:11
	← elijah left	18:11
	→ DerekPioneer joined	18:12
	→ voidx joined	18:12
	← leavittx left	18:13
b1lly	I just created a repo on the github website and now im configuring my git thru bash, do i need to mkdir the same name as the repo name?	18:13
	→ segher joined	18:13
	← segher left	18:13
	→ trafalgar joined	18:14
	← trafalgar left	18:14
	← pppaul left	18:14
	→ marcogmonteiro joined	18:14
PerlJam	b1lly: you mean github has all of the code?	18:14
	← shruggar left	18:14
	← jim0203 left	18:14
b1lly	nah	18:14
	my desktop has all of the code right now	18:14
	i wanna do my first commit	18:15
	im setting up for the first time	18:15
	← alberto56 left	18:15
	→ jim0203 joined	18:15
PerlJam	b1lly: git add remote origin <URL for your github repo>; git push -u origin master	18:15
cbreak	(git init first)	18:15
PerlJam	(assuming I understand you correctly)	18:15
	← Shusshu left	18:15
	→ pppaul joined	18:15
	← Bartzy left	18:16
b1lly	lets say	18:16
cbreak	http://book.git-scm.com/3_getting_a_git_repository.html	18:16
b1lly	my files are like 10gb big right now	18:16
	→ segher joined	18:16
b1lly	is it gunna take forever to push that	18:16
FauxFaux	!large	18:16
gitinfo	[!large_files] Git isn't yet great at large files(larger than RAM). Look at !annex for solutions. You can find them (after gc) with: git verify-pack -v .git/objects/pack/pack-*.idx \| grep blob \| sort -k3nr \| head \| while read s x b x; do git rev-list --all --objects \| grep $s \| awk '{print "'"$b"'",$0;}'; done	18:16
cbreak	git does not work well with such filesizes	18:16
	← arti_t left	18:17
cbreak	consider using something else than git, or maybe bup/git-annex	18:17
	→ xr1rr joined	18:17
b1lly	!annex	18:18
gitinfo	Storing lots of binary files in git causes repo balloon. git-annex and git-media are two solutions to work around this by keeping the blobs out of the repo. http://git-annex.branchable.com/ and https://github.com/schacon/git-media	18:18
	← xr1rr left	18:18
	→ xr1rr joined	18:18
	→ michaelmartinez joined	18:21
	→ hyperair joined	18:21
	→ mishok13 joined	18:23
oohahh	I got gitgot running last night to email me daily reports of my repositories. Seems okay. Do you have a multi-repo git manager that you prefer?	18:23
	← frnkkpp_on_drugs left	18:24
	→ frnkkpp joined	18:24
FauxFaux	Interesting tool. Not sure why you'd care, though; surely you either work on lots of repo at the same time all the time, at which point you use git !slave, or you work on everything independently,which git is fine for?	18:24
gitinfo	[!gitslave] gitslave (http://gitslave.sf.net) is useful to add subsidiary git repositories to a git superproject when you control and develop on the subprojects at more or less the same time as the superproject, and furthermore when you typically want to tag, branch, push, pull, etc. all repositories at the same time.	18:24
	→ dr_win joined	18:25
	← syphar left	18:25
oohahh	Hmm, haven't set up super/sup projects before. I'll check into that.	18:26
	→ khmarbaise joined	18:26
	→ kpreid joined	18:26
	→ kenperkins joined	18:26
	→ mulinux joined	18:26
	← Zenith77 left	18:27
	← dhruvasagar left	18:27
	← blacktulip left	18:27
	→ dhruvasagar joined	18:27
	← gchristensen left	18:28
	→ gchristensen joined	18:29
	← voidx left	18:29
	→ melvinram joined	18:30
	→ madsy joined	18:30
	← madsy left	18:30
	→ madsy joined	18:30
	→ Destos joined	18:31
	→ xalba joined	18:31
	→ ColKurtz joined	18:32
	→ p0t4t0 joined	18:32
	← brngardner left	18:32
	← xr1rr left	18:33
	→ guns joined	18:34
	guns → Guest11592	18:34
	→ cespare joined	18:34
	→ Chaser joined	18:34
	← Guest11592 left	18:34
	→ xalba_ joined	18:35
	→ harshpb joined	18:36
	→ shruggar joined	18:38
	→ brunosaboia__ joined	18:38
	→ alberto56 joined	18:39
	← samba_ left	18:40
	← xalba_ left	18:41
	← ColKurtz left	18:42
	← BSaboia__ left	18:42
	← charon left	18:42
	→ charon joined	18:43
	→ diminoten joined	18:43
	→ callen joined	18:43
	← dbgster left	18:43
	← charon left	18:43
	→ charon joined	18:43
callen	I have an existing repo with some fixtures in it, and I want to switch a particular file inside this existing repo to a git-annex type thing where it stores the data but doesn't attempt to track the data	18:43
	→ master-lincoln joined	18:43
callen	is that possible without creating a "git annex" repo from scratch?	18:44
	is there a native way to tell git to pseudo-ignore a file?	18:44
charon	mysphyt: as they say: good judgement comes from experience, and experience comes from bad judgement. i don't see how "stupid" comes into the picture ;)	18:44
	→ Zenith77 joined	18:45
	← harshpb left	18:45
mysphyt	charon: True enough--but for me, one definition of stupidity is ingenuity applied to bad judgment. I'm often not just wrong--I'm wrong in excitingly creative and unexpected ways. :)	18:45
	→ harshpb joined	18:46
	← joebodo_ left	18:46
	← btx94 left	18:46
	→ btx94 joined	18:46
	→ marhaban joined	18:47
	← btx94 left	18:48
	→ ben225 joined	18:48
	← marhaban left	18:48
	→ btx94 joined	18:48
	← btx94 left	18:48
	→ marhaban joined	18:48
	→ freedrull joined	18:49
	← marhaban left	18:49
	→ marhaban joined	18:49
	→ orkz joined	18:49
	← mercutio22 left	18:50
	→ marhaban1 joined	18:50
	← marhaban left	18:50
	→ mattalexx joined	18:50
	← marhaban1 left	18:50
	→ marhaban joined	18:51
mattalexx	What's the command that will output the contents of a file as it is in the index?	18:51
	← marhaban left	18:51
	← jrots left	18:51
	→ marhaban joined	18:51
freedrull	i squashed around 80 or so commits on branch A with git rebase -i HEAD~80, now i want to rebase on branch B, but when i do, it goes through each commit i squashed as if that single commit is still the individual commits. does that make sense?	18:52
	→ txomon joined	18:52
	← txomon left	18:52
	→ txomon joined	18:52
	← marhaban left	18:52
	→ marhaban joined	18:52
	← marhaban left	18:53
	← juanfer left	18:53
	← mcluskydodallas left	18:53
	→ marhaban joined	18:53
	→ aristidesfl1 joined	18:53
	→ guampa joined	18:55
	← envi_ left	18:55
	← rrn left	18:55
offby1	yes.	18:55
	tizzo-afk → tizzo	18:55
	← aristidesfl left	18:55
mattalexx	I'm surprised, I though that would be a common command but I can't even find it on the internet anywhere.	18:56
	Probably not using the correct terms	18:56
	← jceb left	18:56
	→ mercutio22 joined	18:56
	→ dvance joined	18:56
	→ marhaban1 joined	18:56
	→ cakehero joined	18:56
	← marhaban1 left	18:57
	→ ColKurtz joined	18:57
	→ marhaban1 joined	18:57
	← shruggar left	18:57
	→ rrn joined	18:57
	← marhaban1 left	18:58
	→ marhaban1 joined	18:58
	← marhaban left	18:58
	← Nightfire034 left	18:58
offby1	mattalexx: I suspect there's some syntax for "git show".	18:58
	← marhaban1 left	18:58
	→ marhaban joined	18:59
callen	does anyone know of an intelligent way to dumbly store large files in git that don't involve git-annex?	18:59
	dumbly meaning, I don't want it version controlled, just the contents carried along with everything else.	18:59
EugeneKay	yes, git-media. :-p	18:59
callen	and I don't want to have to purge all the history everytime I change it just to keep the repo size reasonable.	18:59
SethRobertson	!large-files	18:59
	→ marhaban1 joined	18:59
SethRobertson	!largefiles	18:59
	← marhaban left	18:59
hyperair	!hugefiles	19:00
EugeneKay	!big	19:00
gitinfo	[!large_files] Git isn't yet great at large files(larger than RAM). Look at !annex for solutions. You can find them (after gc) with: git verify-pack -v .git/objects/pack/pack-*.idx \| grep blob \| sort -k3nr \| head \| while read s x b x; do git rev-list --all --objects \| grep $s \| awk '{print "'"$b"'",$0;}'; done	19:00
	→ juanfer joined	19:00
EugeneKay	!annex	19:00
gitinfo	Storing lots of binary files in git causes repo balloon. git-annex and git-media are two solutions to work around this by keeping the blobs out of the repo. http://git-annex.branchable.com/ and https://github.com/schacon/git-media	19:00
callen	dude.	19:00
	what did I just say.	19:00
offby1	callen: when you ask "how can I store something in git, but I don't want it version controlled" ... you're asking for complexity and pain	19:00
callen	18:59 < callen> does anyone know of an intelligent way to dumbly store large files in git that don't involve git-annex?	19:00
mattalexx	!ginormouslygiganticfile	19:00
EugeneKay	-annex != -media ;-)	19:00
callen	offby1: it's a dumb fixture.	19:00
	→ eoh joined	19:00
	← marhaban1 left	19:00
callen	offby1: of sparse data.	19:00
	→ Fullmoon joined	19:00
	← TeckniX left	19:00
EugeneKay	If you want a different solution feel free to write it yourself; that's what we know of.	19:00
callen	offby1: it really needs to NOT be version-controlled, just stored.	19:00
hyperair	git help media	19:00
	→ marhaban joined	19:00
hyperair	callen: stuff it in .gitignore	19:00
	→ alexisabril joined	19:00
hyperair	¬_¬	19:00
callen	hyperair: it has to be stored with the rest of the repo :\|	19:01
hyperair	then you don't want git.	19:01
	simple as that	19:01
	← marhaban left	19:01
offby1	callen: perhaps write a pre-commit hook that rejects commits that modify or delete it	19:01
bremner	so both git-annex and git media do what you want.	19:01
EugeneKay	callen - "I want to keep my stuff in version control but I don't want to version control it" ?	19:01
bremner	why not use one of them?	19:01
	→ marhaban joined	19:01
callen	EugeneKay: it's a dumb fixture of sparse data.	19:01
	EugeneKay: it's needed to make the code work.	19:01
hyperair	auto-download	19:01
	symlink	19:01
callen	EugeneKay: being sparse data, if any changes are made to the schema, the entire 700mb file gets invalidated and replaced.	19:01
kevlarman	callen: if it's sparse and uncompressed it'll diff well	19:01
callen	I'd rather the basic git instrument work alone.	19:02
	← marhaban left	19:02
EugeneKay	callen - sounds like a job for a hook+rsync ;-)	19:02
	→ joshteam joined	19:02
	← mulinux left	19:02
callen	kevlarman: no it won't.	19:02
	tizzo → tizzo-afk	19:02
callen	kevlarman: just trust me, it's the pathological case for git's diffing behavior.	19:02
	→ marhaban joined	19:02
kevlarman	callen: so it's compressed?	19:02
callen	I just need to the repo to hold onto the fucking file and not diff it. that's it. I tried git-annex and it's way overdone.	19:02
	kevlarman: I'm anything but in the mood to explain the parts that are already understood.	19:02
	→ marhaban1 joined	19:03
	← marhaban left	19:03
callen	kevlarman: I'd rather explore a solution to making git forcibly never diff the file and just "hold onto it" like a blind symlink but still cart the file around	19:03
hyperair	callen: hang on, you want the repo to hold on to the latest version of the file, and this file may change.	19:03
	← iamjarvo left	19:03
hyperair	so you don't want it to be version-controlled	19:03
	but shipped around with the rest of the repo	19:03
callen	hyperair: I want the equivalent of a purge everytime it gets replaced/changed.	19:03
	→ iamjarvo joined	19:03
callen	hyperair: yes.	19:03
kevlarman	callen: that's not possible by definition with git	19:03
hyperair	callen: you want make and wget. ¬_¬"	19:03
	→ aristidesfl joined	19:03
	← marhaban1 left	19:03
hyperair	callen: and .gitignore.	19:03
kevlarman	callen: you can use git-annex or git-media to prevent people from having to download every version	19:03
	← ColKurtz left	19:03
	→ marhaban joined	19:03
kevlarman	or you can store it outside git	19:03
SethRobertson	Or git-media, or git-annex, or a separate repo which you can rewrite every time you change something	19:04
callen	git-annex didn't work out, as I said. I'll take a look at git media if I can find it	19:04
kevlarman	(actually both of those store it outside git-media)	19:04
	outside git*	19:04
	← marhaban left	19:04
hyperair	callen: actually what kevlarman said was a good idea. how about submodules?	19:04
callen	I don't want it version-controlled period	19:04
	gitinfo set mode: +v	19:04
	→ marhaban joined	19:04
kevlarman	that wasn't me	19:04
SethRobertson	Wasn't that me?	19:04
EugeneKay	!UGFWIINI	19:04
callen	shoving it into another repo just shifts the data explosion to a different repo.	19:04
gitinfo	You appear to be Using Git For What It Is Not Intended - https://www.google.com/search?q=UGFWIINI Let's step back for a minute - What are you actually trying to achieve? Are you sure git is the right tool for the job?	19:04
callen	EugeneKay: thank you, your desire to be as unhelpful as possible is noted.	19:05
hyperair	callen: put it in a submodule, commit --amend each time so that there's only ever one commit..	19:05
kevlarman	but yes if you really want you can stick it in a submodule that gets rebased on every change	19:05
	← marhaban left	19:05
callen	EugeneKay: I would say that at this point, it's clear you have nothing to contribute and you go back to whatever you were doing before bothering me.	19:05
	← marcogmonteiro left	19:05
	← aristidesfl1 left	19:05
	ChanServ set mode: +o	19:05
	→ jrots joined	19:05
	EugeneKay set mode: +q	19:05
	→ marhaban joined	19:05
	→ marcogmonteiro joined	19:05
EugeneKay	And I'd say that you are unwilling to accept the reality of the impossibiltiy of your question.	19:05
offby1	EugeneKay: that sure is an awkward trigger.	19:05
freedrull	so you can't squash a bunch of commits, then treat it like one commit when you're rebasing another branch onto this one? why not?	19:05
EugeneKay	offby1 - there's also the ! doinitrong shortcut	19:06
	← marhaban left	19:06
	→ marhaban joined	19:06
	EugeneKay set mode: -o	19:06
kevlarman	!doinitrong	19:06
gitinfo	[!UGFWIINI] You appear to be Using Git For What It Is Not Intended - https://www.google.com/search?q=UGFWIINI Let's step back for a minute - What are you actually trying to achieve? Are you sure git is the right tool for the job?	19:06
	→ mcluskydodallas joined	19:06
kevlarman	yeah i like that one better :D	19:06
offby1	freedrull: I'm not 100% sure what you're doing, but I think the answer is: no, you cannot	19:06
	← marhaban left	19:06
	→ marhaban joined	19:07
EugeneKay	callen - we've given you the best info we can. It's up to you to step back and figure it out; we won't spoon feed you how to run rsync.	19:07
SethRobertson	freedrull: You have not provided enough information, for instance, are the commits also on the other branch?	19:07
offby1	freedrull: you might be able to see how the two branches differ, then reset the second branch to the rebased one, then apply those diffs.	19:07
superdmp	I want to checkout a remote branch that's not currently in the local repository - I could pull then checkout the branch, but I assume there is a way of doing it one go	19:07
	← marhaban left	19:07
superdmp	git pull origin <branch> ?	19:07
	← Dreamer3 left	19:07
	→ marhaban joined	19:07
	← cakehero left	19:07
offby1	freedrull: that way both branches will now share the big squashed commit.	19:07
	I've never done that though	19:08
	→ Dreamer3 joined	19:08
	← lyddonb left	19:08
SethRobertson	superdmp: Do not use the four argument version of fetch or pull. It will only lead to tears	19:08
cmn	superdmp: pull doesn't mean what you think it means	19:08
superdmp	cmn: that would not surprise me	19:08
	← marhaban left	19:08
	→ marhaban1 joined	19:08
SethRobertson	superdmp: YOu can `git fetch origin` (if you are not up to date). Then you can `git checkout branchname` to get a local tracking branch or `git checkout origin/branchname` to get a read-only-ish remote tracking branch	19:08
cmn	you're looking for fetch	19:08
offby1	X doesn't mean what you think it means, for a depressingly large number of possible values of X	19:08
superdmp	thanks again	19:09
	← marhaban1 left	19:09
	→ marhaban joined	19:09
	← marhaban left	19:10
freedrull	SethRobertson: branch A(with the squashed commits) has the squashed commit. the commits in this squashed commit have no commits on branch B. branch A originally was forked from branch B, but branch B has since added more commits not on branch A, so i want to rebase branch B onto branch A.	19:10
	→ marhaban joined	19:10
	→ skammer joined	19:10
freedrull	with the idea of eventually merging branch A into branch B. (if it helps, branch B is my 'development' branch and branch A is a feature branch)	19:10
	← skammer left	19:10
	← marhaban left	19:10
SethRobertson	freedrull: But basically if you have really squashed 80 unique commits previously exclusively on branch B and then want to merge B into A, or rebase B on top of A, then git will not know about the previous 80 unique commits. Period.	19:10
	→ marhaban joined	19:11
	→ snearch joined	19:11
	← marhaban left	19:11
	offby1 suspects that a simple illustration would be useful right around now	19:11
	→ whitman joined	19:11
	→ marhaban joined	19:11
	→ brngardner joined	19:12
SethRobertson	So you want the squashed commit on top of the commits on branch B. That is fine. How many commits does `git log A..B` return?	19:12
	← marhaban left	19:12
freedrull	SethRobertson: ok...but that's what i want i think? when i'm on branch a, i type git rebase B but it seems to be going through each commit like it still exists...	19:12
	→ marhaban joined	19:12
freedrull	offby1: haha yeah is there something i can draw on...	19:12
	← marhaban left	19:13
	→ voidx joined	19:13
	→ marhaban joined	19:13
SethRobertson	freedrull: Answer the question: how many commits do you see with `git log A..B`?	19:13
	→ sq-one joined	19:13
	← marhaban left	19:13
freedrull	SethRobertson: about 70....	19:14
	why?	19:14
	→ marhaban joined	19:14
SethRobertson	I thought you squashed them?	19:14
	Those 70 commits are going to be replayed on top of the commits on B.	19:14
	← marhaban left	19:14
freedrull	these aren't the squashed commits, this shows the new commits on branch B that are not in A yet. i want to rebase B onto A so i can get these 70 commits into A and replay my work ontop of them	19:14
	→ marhaban joined	19:14
SethRobertson	So you want to do: `git checkout B; git rebase -p A` correct?	19:15
	← marhaban left	19:15
freedrull	`git checkout A; git rebase B;`	19:15
	→ marhaban joined	19:15
freedrull	other way around	19:15
	← harshpb left	19:16
	← marhaban left	19:16
SethRobertson	That is what I thought, but not what you said the second time. Anyway, when you do that, the 70 commits you mention are replayed on top of all of B's commits.	19:16
	→ marhaban joined	19:16
	→ revagomes joined	19:16
offby1	freedrull: semi-seriously: if you find it's hard to make your problem understood, a little shell script that reproes it would be perfec.t	19:16
SethRobertson	That will put A's unique commits on top of B's commits.	19:16
	← marhaban left	19:17
	→ marhaban joined	19:17
	← coyotebush left	19:17
	← sdeboey left	19:17
freedrull	SethRobertson: ok thats what i want. i think it is working ok...	19:18
	→ hobbyist joined	19:18
SethRobertson	Great. Glad I could clarify your confusion.	19:18
freedrull	offby1: sorry next time ill try to make a readable paste. just couldn't think of a good way to explain it without drawing. but a sequence of shell commands would probably help...	19:18
	→ ColKurtz joined	19:18
SethRobertson	You can always use `gitk A B` and then draw arrows on top	19:19
	← alberto56 left	19:19
	← n3m left	19:19
	→ Targen joined	19:20
	← kenperkins left	19:20
	ChanServ set mode: +o	19:20
EugeneKay	callen - you are asked to keep it civil, and to understand that your use case is not well served by git's object model. There are several possible solutions, the best of which do not involve /usr/bin/git.	19:20
	← tureba left	19:20
	→ eletuchy joined	19:20
	EugeneKay set mode: -q	19:20
	→ n3m joined	19:21
	EugeneKay set mode: -o	19:21
freedrull	SethRobertson: ah ok...thanks i will next time	19:21
	← marhaban left	19:21
	→ kenperkins joined	19:21
	→ dbgster joined	19:21
	→ pandeiro joined	19:21
	← akiress left	19:22
	← freeed left	19:22
	→ whyking_ joined	19:22
	→ estevocastro joined	19:22
	→ tureba joined	19:22
	→ Nightfire034 joined	19:22
	← juanfer left	19:22
	digiv\|meeting → digiv	19:23
	← FreeArtMan left	19:23
	← Vortex34 left	19:23
	→ lyddonb joined	19:24
	→ alberto56 joined	19:24
	→ DannyZB joined	19:24
	→ krz joined	19:24
DannyZB	guys : I use GitGui over mac .. how do I remove files from a branch + make sure they never get to the repository ?	19:24
	← robotmay left	19:25
xalba	Hello. New to git. "git commit" and "git commit -a" should make differents commits, is it possible that in my installation "git commit" calls "git commit -a" ?	19:26
m1sc	DannyZB: if those files have been tracked earlier, they are in the repo. (if that's the question) http://help.github.com/remove-sensitive-data	19:26
	→ sandstrom joined	19:26
	← sandstrom left	19:26
	→ sandstrom joined	19:26
	← hobbyist left	19:26
offby1	xalba: unlikely. However I suspect there are many cases where those two commands will do the same thing	19:27
cbreak	DannyZB: first step: start the terminal and navigate to the repository	19:27
	after that, git rm the file, enter it in .gitignore, then commit.	19:27
sandstrom	How can I reset the repository from a remote?	19:27
cbreak	this will NOT affect past history, only future	19:27
	and other users can of course override it	19:27
	sandstrom: not at all. You have to do it locally	19:28
	with git reset --hard or similar. Read the man page for details	19:28
offby1	xalba: for example, I think if you edit a file, then run "git add" on it, then "git commit" and "git commit -a" will both commit that file	19:28
	← troulouliou_dev left	19:28
sandstrom	cbreak: thanks!	19:28
xalba	But "git commit" shouldn't commit unstaged changes	19:29
offby1	correct.	19:29
	← bartek left	19:29
offby1	"git commit -a" however will.	19:29
DannyZB	cbreak: thx	19:30
	→ bartek joined	19:30
PerlJam	xalba: Are you seeing the same effect from both versions of git commit ?	19:30
	→ toothr joined	19:30
xalba	I think so, I was trying "git commit" with unstaged changes to see "no changes added to commit (use "git add" and/or "git commit -a")"	19:31
	but it actually did commit unstaged changes	19:31
offby1	xalba: !repro	19:32
gitinfo	xalba: Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.	19:32
offby1	or it didn't happen :)	19:32
	← lyddonb left	19:32
	→ cbx333 joined	19:33
xalba	I'll try to reproduce it	19:33
	← cbx333 left	19:33
	→ hobbyist joined	19:33
offby1	that's ideal.	19:34
	Either you've found a bug (unlikely but possible) or you're misunderstanding something; the repro will be useful in either case.	19:34
	→ pokeswap joined	19:34
	→ juanfer joined	19:34
	← sandstrom left	19:37
	→ aristidesfl1 joined	19:37
	→ mechanyancat joined	19:37
	→ eletuchy_ joined	19:37
	← aristidesfl1 left	19:37
mechanyancat	git filter-branch --index-filter 'git rm --cached --ignore-unmatch tests/fixtures/fixture_name.json' HEAD	19:37
	→ j416 joined	19:38
	→ cybersphinx joined	19:38
	← brunosaboia__ left	19:39
mechanyancat	freedrull: ^^ you led me down the right path, thank you.	19:39
	← aristidesfl left	19:40
	← eletuchy left	19:40
	→ alvaro_o joined	19:41
	← alvaro_o left	19:41
freedrull	mechanyancat: what did i do?	19:41
	← the_cybersphinx left	19:41
mechanyancat	freedrull: helped me find the vernacular needed to discover the solution to the problem.	19:41
	freedrull: namely, the concept of purging the file completely from history.	19:42
	→ teweWORK joined	19:42
	→ johnkpaul joined	19:42
johnkpaul	I've made a few commits on master that I should have been making in a branch	19:42
offby1	gasp	19:42
johnkpaul	I have made a new branch from master, now how do I move the master pointer back to the original commit	19:42
offby1	git reset --hard ORIGINAL-SHA	19:42
	easy as pi	19:43
	have you pushed master since you screwed up?	19:43
	If not, you're golden	19:43
johnkpaul	no	19:43
	I haven't	19:43
	cool	19:43
offby1	\o/	19:43
johnkpaul	:)	19:43
	so doing that won't lose the commits that I made	19:43
	← hobbyist left	19:43
johnkpaul	because I have them in a branch	19:43
	right? I'm still wrapping my head around how git works	19:43
offby1	correct	19:44
	← mabrand left	19:44
offby1	even if you forgot to make the new branch, and then reset, those commits _still_ aren't lost, but it takes a bit of extra work to resuscitate them.	19:44
	→ oehman joined	19:44
johnkpaul	ah ok	19:44
	they're in the .git directory, but nothing is pointing to them	19:44
offby1	exactly!	19:44
	but in reality there _is_ something pointing at them, called the "reflog".	19:45
	← mysphyt left	19:45
offby1	and it exists more or less for just that reason: so that you don't lose commits.	19:45
johnkpaul	oh ok, that's where I start not knowing how it works	19:45
	and the reflog is local to my repo, right?	19:45
offby1	correct.	19:45
johnkpaul	so if I run a git show on a sha that no branch has access to	19:46
	it'll still work?	19:46
offby1	actually you sound like you're more clued-in than 95% of people who describe themselves as "new to git"	19:46
	yes, it'll work; Try It And See™.	19:46
johnkpaul	ha, I will	19:46
	eh, I read the free book	19:46
	progit	19:46
offby1	well, there you go.	19:46
johnkpaul	but I don't really use these features enough to consider myself not new	19:46
	→ robotmay joined	19:46
offby1	it'll come	19:46
johnkpaul	svn I knew like the back of myhand	19:46
	yeah, definitely	19:46
	thanks for your help offby1!	19:47
	← voidx left	19:47
offby1	yep, I was a svn whiz too	19:47
	I've forgotten most of it by now though.	19:47
johnkpaul	yeah, I'm forgetting most of it	19:47
	← FernandoBasso left	19:47
johnkpaul	except the horrors of tree conflicts	19:47
	because the pain is permanently burned into my head	19:47
kevlarman	johnkpaul: forgetting how svn works is the first step to learning git :P	19:48
	→ aspotashev joined	19:48
	offby1 still remembers the working copy getting hosed every so often	19:48
kevlarman	(you can relearn after you understand git, but i've found you can't learn git while you know svn :P)	19:48
	← lacrymology left	19:48
johnkpaul	haha, well sadly at the moment I Have to use both, but much more git than svn	19:49
	and greatfully not git-svn	19:49
kevlarman	johnkpaul: git-svn is much more tolerable than svn	19:49
	← erichynds left	19:49
johnkpaul	eh, I tried, I found it very difficult to deal with all of the random errors	19:49
	→ stkrzysiak joined	19:49
	← Nightfire034 left	19:50
johnkpaul	although I didn't put that much time into figuring out what caused them	19:50
	← dropdrive left	19:50
	gitinfo set mode: +v	19:50
	← _Vi left	19:50
	← pandeiro left	19:51
	← glcrazy left	19:52
	→ CannedCorn joined	19:52
	→ glcrazy joined	19:53
johnkpaul	awesome, just pushed the corrected branch and master state, and it looks like I had the branch the whole time	19:53
	git is amazing sometimes	19:53
EugeneKay	Is there a way to convince git to change the executable bit without chmod-ing the file in question? Eg, windows git.	19:53
	← OOPMan left	19:53
johnkpaul	windows has an equivalent of chmod, no?	19:53
wereHamster	EugeneKay: you mean just in the index?	19:53
EugeneKay	wereHamster - yes	19:53
wereHamster	EugeneKay: update-index, probably	19:54
EugeneKay	man git-update-index	19:54
gitinfo	the git-update-index manpage is available at http://jk.gs/git-update-index.html	19:54
EugeneKay	johnkpaul - tell that to `git` ;-)	19:54
	→ pandeiro joined	19:54
EugeneKay	wereHamster - --chmod, yup. <3	19:54
	→ Davey_ joined	19:54
	→ ctrochalakis joined	19:55
	→ syphar joined	19:55
	← enmand left	19:55
xalba	<offby1><gitinfo>: Perhaps I'm misunderstanding but it seems to me that "git commit" is doing the same as "git commit -a"	19:56
	→ marhaban joined	19:56
offby1	xalba: without a !repro it'll be impossible for us to figure out what's going on	19:56
gitinfo	xalba: Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.	19:56
	← lonewulf` left	19:57
	← marhaban left	19:57
	→ marhaban joined	19:57
xalba	sorry I don't know what a !repro is	19:57
gitinfo	Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.	19:57
	← marhaban left	19:57
xalba	I pasted it using https://gist.github.com/	19:57
EugeneKay	xalba - the !bot knows all	19:58
gitinfo	xalba: [!gitinfo] I am an IRC bot which responds to certain keywords to provide helpful(?) information to humans. Please see http://jk.gs/git/bot for more information about how to use me.	19:58
	→ marhaban joined	19:58
	← marhaban left	19:58
	→ marhaban joined	19:58
	← yatiohi left	19:59
	← marhaban left	19:59
	→ marhaban joined	19:59
	→ lonewulf` joined	19:59
freedrull	it seems git rebase -p branch leaves you on a detached head with a new merge commit?	19:59
	← iamjarvo left	20:00
	← marhaban left	20:00
	→ marhaban joined	20:00
	← marhaban left	20:01
	→ sandstrom joined	20:01
	← sandstrom left	20:01
	→ sandstrom joined	20:01
	→ marhaban joined	20:01
	← mercutio22 left	20:01
SethRobertson	freedrull: Not unless you have a conflict or were on a detached head to start with	20:01
	Probably you had a merge conflict	20:01
	rebase conflict	20:01
	← t4nkd left	20:01
	← traviscj left	20:01
	← marhaban left	20:01
	→ marhaban joined	20:02
	← jrots left	20:02
	← marhaban left	20:02
	→ marhaban1 joined	20:02
offby1	xalba: give us the URL or else we won't be able to see it.	20:03
	→ marhaban joined	20:03
	← marhaban1 left	20:03
	→ gjoseph joined	20:03
	← nb left	20:04
	← marhaban left	20:04
	→ marhaban joined	20:04
	→ jrots joined	20:04
	← marhaban left	20:04
	→ schnatta joined	20:04
	← mcluskydodallas left	20:04
xalba	git://gist.github.com/2361975.git	20:05
	→ pantsman joined	20:05
	← pantsman left	20:05
	→ pantsman joined	20:05
	← orkz left	20:05
	← pandeiro left	20:06
	→ nb joined	20:06
gjoseph	hi there	20:06
gitinfo	gjoseph: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.	20:06
	← lonewulf` left	20:06
gjoseph	any idea why git config --get svn-remote.svn.fetch :refs/remotes/git-svn$; echo $? would return/output 1 for a given repo, and 0 for another ? In both cases, the property does not exist.	20:06
	if the pattern regex was invalid, it should exit with 2 according to man. if i don't pass a pattern, both exit with 0	20:07
	→ OOPMan joined	20:07
	← stkrzysiak left	20:08
	→ mcluskydodallas joined	20:08
gjoseph	sorry, property exists in both cases, same value, and doesn't match the pattern	20:08
SethRobertson	xalba: You specifically named the file you wanted to commit!	20:08
	→ Stone44 joined	20:09
offby1	oh, is that it?	20:09
	I'm as surprised as xalba	20:09
gjoseph	gist coming up	20:09
	→ teknoid joined	20:09
offby1	xalba: yeah, SethRobertson is right. The man page (!man git-commit) says that if you specify the file's name, it ignores the index	20:10
	← juanfer left	20:10
SethRobertson	Basically that evicts everything from the index, adds the named file, recommits, and then readds things back to the index. I have a nasty suspicion that it doesn't always readd everything back to the index 100% properly, but that is another issue	20:10
offby1	so that explains that	20:10
	kludge-a-licious	20:10
	→ Nightfire034 joined	20:10
	← iocor left	20:10
offby1	NEXT!	20:10
	→ lonewulf` joined	20:11
	← Ryback_ left	20:11
xalba	aaahhh thank you!!	20:11
gjoseph	woops, sorry for the overlap	20:12
	← Ard1t left	20:12
gjoseph	—> https://gist.github.com/2362113	20:12
	→ Subodh joined	20:13
	← marthinal left	20:14
	← n1x left	20:14
	← JohnBat26 left	20:14
	→ juanfer joined	20:14
offby1	all hail the power of the transcript	20:15
	gitinfo set mode: +v	20:15
Subodh	I have two commits with SHA1 1234, and SHA2 of 3456. I want to git format-patch these two commits into one patch file how do I do that?	20:16
	→ Harzilein joined	20:16
Harzilein	hi	20:16
gitinfo	Harzilein: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.	20:16
	→ the_cybersphinx joined	20:16
	gitinfo set mode: +v	20:17
Harzilein	can i just get a named branch for whatever that bf6f is?: http://paste.debian.net/162917	20:17
	→ illusi joined	20:17
Harzilein	(git reflog \| head -4)	20:17
	← alberto56 left	20:17
	← illusi left	20:18
cbreak	Harzilein: you can try git name-rev bf6f7d7340255f85997460bc80d432f742410bc6	20:18
Harzilein	hmm...	20:18
	remotes/origin/master~1	20:18
	→ illusi joined	20:18
Harzilein	so what happened there? (it's actually our production checkout, so not on a particular dev's machine)	20:18
	→ windsurf_ joined	20:19
windsurf_	if i want to use git for deploying my website, i was thinking I'd ssh into my host and clone my repo but then my .git will be exposed publicly	20:19
	thoughts?	20:19
	i've tried rsync instead from local to remote but that isn't preserving permissions properly and ftp is too slow	20:20
	← cybersphinx left	20:20
bremner	!web	20:20
gitinfo	[!website] Git is not a website deployment tool, but can sometimes play one in sufficiently simple/lax environments with a little help. One example of help is: http://toroid.org/ams/git-website-howto and another approach: https://gist.github.com/1714235	20:20
	← j416 left	20:21
	← DannyZB left	20:21
gjoseph	sorry to reiterate, not sure this went through: any idea why git config --get svn-remote.svn.fetch :refs/remotes/git-svn$; echo $? would return/output 1 for a given repo, and 0 for another ? In both cases, the property does not exist. — https://gist.github.com/2362113	20:21
	→ btanaka joined	20:22
	→ marcogmonteiro_ joined	20:23
kevlarman	gjoseph: look at your output again, it exists when it returns 0	20:23
	← Subodh left	20:24
SethRobertson	windsurf_: More specifically, you can either request the web server to not permit .git, or you can also ask git to checkout the working directory to another path than the one the .git repo is at	20:25
	← marcogmonteiro left	20:25
	marcogmonteiro_ → marcogmonteiro	20:25
gjoseph	kevlarman: dammit. yeah i just noticed. wth was i thinking ? wonder if "at some point" it did behave differently. seems git svn create-ignore was unhappy only about the 1.1 branch	20:26
	(but thanks!)	20:26
	→ t4nkd joined	20:26
	← fission6 left	20:27
	neersighted → neersighted\|AFK	20:27
	airborn_ → airborn	20:27
	→ psoo joined	20:28
	← syphar left	20:28
	→ syphar joined	20:28
	← robotmay left	20:29
	→ marhaban joined	20:31
	← marhaban left	20:32
	→ marhaban joined	20:32
	neersighted\|AFK → neersighted	20:33
	← marhaban left	20:33
	→ marhaban1 joined	20:33
	→ phishy joined	20:33
	← phishy left	20:33
	← marhaban1 left	20:34
	→ hobbyist joined	20:36
	← gartenstuhl left	20:36
	← airborn left	20:36
	→ iamjarvo joined	20:37
	→ gartenstuhl joined	20:37
	← whitman left	20:38
	→ s3a joined	20:38
	← hobbyist left	20:38
	← gartenstuhl left	20:39
s3a	Let's say I modify an EXISTING java file and I want to commit the change, do I still use git add?	20:39
	or is git add just for when i created a new java file that was not in the project already?	20:39
	← eijk_ left	20:40
	→ eijk_ joined	20:40
	← melvinram left	20:40
	→ brentdanley_ joined	20:40
bremner	every change	20:41
	there are shortcuts like git commit -a, but in the long run, adding every change works out better	20:41
	← felgru left	20:41
	→ stkrzysiak joined	20:41
	digiv → digiv\|afk	20:42
	→ gryzzly joined	20:42
	→ felgru joined	20:42
	→ airborn joined	20:42
s3a	bremner, alright. thanks :)	20:43
	← sbell left	20:43
	→ maurer joined	20:43
	→ sbell joined	20:43
maurer	If I've git added a file, and don't want to stage it into this commit, how do I do that?	20:43
	(e.g. I have typed git add and didn't mean it, what's the "oops" command)	20:43
	→ tatsuya_o joined	20:44
	← fling left	20:44
gryzzly	Hi. Is there any place where I can find documentation explaning how to perform manual merge beyond the basic example? I am getting something like "merged common ancestors" and googling for it doesnt lead to anything that looks like a trustable reference. Thansk!	20:44
	← Stone44 left	20:45
	← jrots left	20:45
	← sbell left	20:48
offby1	maurer: "git status" will tell you.	20:48
	I think it's "git reset HEAD file"	20:48
	gryzzly: are you saying that you're running "git merge" and getting an error?	20:49
	← jsoa left	20:50
gryzzly	offby1: nope, I am doing git merge, I then get a conflict. I want to resolve the conflict manually, without running a mergetool. It's more for learning purpose and understanding, actually.	20:50
	← airborn left	20:50
	← mishok13 left	20:50
offby1	oh that's easy	20:52
	just edit the files	20:52
	make 'em look the way you want	20:52
	then run "git add" on them	20:52
	then "git commit"	20:52
	that's it	20:52
	← martinjlowm-lptp left	20:53
	→ drizzd joined	20:54
gryzzly	offby1: Sure :-) I'd like to understand though, what "\|\|\|\|\|\|\| merged common ancestors" means	20:54
	→ martinjlowm-lptp joined	20:54
	← chrisshattuck left	20:55
superdmp	is there anything harder than git reset --hard? I want it to get rid of untracked files too	20:55
gryzzly	so, i know what it means when code is separated by "<<< HEAD // some code from HEAD here ======== // some code from merge branch here >>>> 123123hash "	20:55
	← xalba left	20:56
gjoseph	superdmp, err re-clone the repo ? :D	20:56
PerlJam	superdmp: git clean -dxf # DANGER WILL ROBINSON	20:56
gryzzly	when merging files you haven't worked on on a huge project, you can't always make sense of different lines and also testing in the middle of huge merge isn't always possible as building won't work. so it would be nice to understand what git is trying to suggest	20:56
	← Fullmoon left	20:56
maurer	offby1: I found it, ant i tis that command, but git status doesn't tell me.	20:56
	*and it is	20:56
superdmp	PerlJam: danger because it will delete files?	20:56
Harzilein	cbreak: thanks	20:57
superdmp	good!!	20:57
Harzilein	any idea how i could go reverse the order in this?: http://www.zorched.net/2008/04/14/start-a-new-branch-on-your-remote-git-repository/	20:57
	← Sigma left	20:57
	→ robotmay joined	20:57
	→ likkabe joined	20:57
	← windsurf_ left	20:57
	← gjoseph left	20:57
Harzilein	as in, i created a branch pointing to that earlier commit, now i don't want to do new commits, just push the branch "metadata"	20:57
	← bartek left	20:58
	← cannonball left	20:58
	→ DCN joined	20:58
Vinnie_win	In git fetch upstream; git checkout master; git pull upstream master; Is the "fetch" necessary? It seems redundant.	20:58
PerlJam	Harzilein: git checkout SHA1; git checkout -b new_branch; git push -u origin new_branch; # assuming I understand you	20:58
	← fahi left	20:59
	← chronos left	20:59
	← else- left	20:59
	← baggles left	20:59
PerlJam	Harzilein: where SHA1 is the commit where you want to start the branch	20:59
	→ zorzar_ joined	20:59
	← NimeshNeema left	20:59
	→ baggles joined	20:59
	→ fahi joined	20:59
	← zorzar left	20:59
	← waldvogel left	20:59
	← jaequery left	20:59
offby1	gryzzly: well, do you know, fundamentally, what a merge conflict is?	20:59
	→ jaequery joined	20:59
	← likkabe left	20:59
Harzilein	PerlJam: i already have the branch. i'll try adding -u to see if that causes the new branch to get pushed	21:00
	→ Slix` joined	21:00
	→ lisbakke joined	21:00
	→ else- joined	21:00
	gitinfo set mode: +v	21:00
lisbakke	is there a way to have githooks in the repository so that when someone runs git clone they will have the githooks?	21:00
PerlJam	lisbakke: no	21:00
offby1	lisbakke: I don't think so :-(	21:00
lisbakke	:(	21:00
offby1	dunno why that is	21:00
	seems like it'd be convenient	21:00
lisbakke	so, remote hooks? :\|	21:01
PerlJam	offby1: seems like it'd be a security nightmare to me.	21:01
lisbakke	is there a way to make a remote hook that can differentiate a git clone vs. the git clone issued by gerrit?	21:01
PerlJam	(but, yes, it could be convenient :)	21:01
lisbakke	(http://code.google.com/p/gerrit/)	21:02
	← martinjlowm-lptp left	21:02
offby1	lisbakke: my hunch is "no"	21:02
PerlJam	lisbakke: why?	21:02
gryzzly	offby1: sure, I do :-) I also used SVN and Mercurial for long enough time. I just never seen this kind of "tag" saying "merging common ancestors". It's something Git specific and I can't find documentation about it.	21:02
lisbakke	PearlJam: Because we want people to submit code reviews through gerrit instead of github pull requests. To do this they have to run gerrit-clone instead of git clone from the very start.	21:03
	→ lwhalen joined	21:03
lwhalen	hey, can someone help me recover a hosed gitolite installation?	21:03
	← hagabaka left	21:03
offby1	gryzzly: maybe I'd better make a conflict of my own, to check; my understanding was that git's conflict markers are pretty much the same as svn's	21:03
lwhalen	I accidentally blew away /var/gitolite/.gitolite/	21:03
lisbakke	PerlJam I mean :)	21:03
lwhalen	while I have a local copy of gitolite.conf, how can I tell the server to recompile it?	21:04
Harzilein	PerlJam: that worked. before/after: http://paste.debian.net/162922	21:04
PerlJam	lisbakke: give them a "setup" script that does the gerrit clone and sets up the hooks you need	21:04
Harzilein	PerlJam: thanks :)	21:04
PerlJam	Harzilein: pay it forward :)	21:05
SethRobertson	lwhalen: Without knowing the ins and outs, make a backup of what is left, reinstall, and then do what you would normally do	21:05
offby1	gryzzly: ok, now: _where_ are you seeing this "merging common ancestors" message?	21:05
	← falena left	21:05
	← mcluskydodallas left	21:05
offby1	gryzzly: here's what my little toy conflict looks like: http://ix.io/2ic	21:05
lwhalen	SethRobertson: will that save my repositories?	21:05
	→ sbell joined	21:05
SethRobertson	lwhalen: Remember the bit where I told you to make a backup? Yeah. That.	21:06
lwhalen	haha!	21:06
lisbakke	PerlJam: People are used to git clone and then submitting pull requests. A "setup" script requires that they have read some docs... if they read the docs they'd know they should be using gerrit. But I'd like a way to prevent them from making the mistake of git cloning in the first place	21:06
lwhalen	I have the backup, thankfully :-)	21:06
	→ subh joined	21:06
SethRobertson	lwhalen: sitaram will probably be on later and of course is the man if you want less drastic "help"	21:06
	← macabre^ left	21:06
	→ cybersphinx joined	21:06
lwhalen	I've got two dozen cranky devs howling for my head right now	21:07
	I've got to get this up and running asap :-p	21:07
	→ macabre joined	21:07
SethRobertson	Well, then take a second backup and try to reinstall.	21:07
	← qt-x left	21:08
kevlarman	lwhalen: how did you hose it?	21:08
lwhalen	gitolite was installed in /var/gitolite, on a /var mount	21:08
	← DCN left	21:08
PerlJam	lisbakke: they have to type some command one way or another, so I don't much see the difference if it's "git clone", "gerrit clone", or "<insert your command here>"	21:08
lwhalen	/var filled up, so thinking I was clever I created a new LVM just for gitolite	21:09
PerlJam	lisbakke: Step 0: type "set-me-up" # runs gerrit clone and configures the repo	21:09
	← the_cybersphinx left	21:09
lwhalen	mount /dev/vg3/git /newgit; rsync -avrt /var/gitolite/* /newgit; rm -rf /var/gitolite/*; umount /newgit; mount /dev/vg3/git /var/gitolite/	21:09
	rsync -avrt does NOT grab dotfiles. Much to my chagrin.	21:10
	→ subrosa joined	21:10
lisbakke	PerlJam: Yeah. But what I'm saying is that people go to the project and immediately git clone it, by habit, and begin working. They need to be warned, somehow, that they are wasting their time and should be running gerrit-clone.	21:10
	← dstorrs left	21:10
PerlJam	lisbakke: that's what README files are for :)	21:11
kevlarman	lwhalen: easiest would be to reinstall gitolite from scratch, push gl-admin.git, then have all your devs repush their repositories i think	21:11
	tizzo-afk → tizzo	21:11
lisbakke	yeah :) it's in there. i guess if someone didn't read the README then it's their fault..	21:11
kevlarman	lwhalen: since they by definition have (nearly) full copies	21:11
lwhalen	we've got 30+ repos	21:11
	→ NimeshNeema joined	21:11
	← snearch left	21:12
kevlarman	lwhalen: well, i suppose you could create a commit with no real changes in gl-admin	21:12
	and push that to make it rerun all the hooks	21:13
	← eoh left	21:13
	→ eoh joined	21:13
	← d0k left	21:14
	← oehman left	21:14
	← TomyLobo left	21:15
	→ lyddonb joined	21:15
	→ orkz joined	21:15
	→ dstorrs joined	21:16
	→ voidx joined	21:16
	← khmarbaise left	21:16
	← machine2 left	21:16
	← arietis left	21:16
	← psoo left	21:17
	← lisbakke left	21:17
	← lyddonb left	21:17
	→ tvw joined	21:18
	← voidx left	21:18
	← robotmay left	21:19
	estevocastro → estevo\|afk	21:19
	← gryzzly left	21:20
	→ ignacio joined	21:20
	→ mcluskydodallas joined	21:20
	← sandstrom left	21:20
	→ melvinram joined	21:21
	← melvinram left	21:21
	→ machine2 joined	21:22
	→ _Vi joined	21:22
lwhalen	hrm	21:23
	on the reinstall, I'm getting:	21:23
	bad reponame 'CREATOR/.*' or you forgot to set $GL_WILDREPOS	21:23
	← onicrom left	21:23
	← w0ng left	21:24
	→ w0ng joined	21:25
	← Maccer left	21:25
	→ webflo joined	21:26
	→ Maccer joined	21:26
	← pantsman left	21:27
	← ins\ left	21:29
	← sq-one left	21:31
	← johnkpaul left	21:31
	→ RobW joined	21:31
	→ Licenser joined	21:32
Licenser	aloa :)	21:32
	← mrwalker left	21:34
	← ghvandoorn left	21:34
	→ griotspeak joined	21:34
	→ voidx joined	21:35
charon	Licenser: you neatly dodged the bot's !hi autoreply	21:36
gitinfo	Licenser: [!welcome] Welcome to #git, a place full of helpful gits. If you have a question, feel free to just go ahead and ask—somebody should answer shortly. For more info on this channel, see http://jk.gs/git/ Take backups (type !backup to learn how) before taking advice.	21:36
Licenser	charon thanks :)	21:36
	← revagomes left	21:36
Licenser	so will you teach your bot aloa as greeting? ^^	21:36
charon	file it with jast as a feature request	21:36
Licenser	teehee	21:37
	← Carraway left	21:37
	← RobW left	21:37
Licenser	what I wanted to ask, I have a submodule and would like to update it to a new tag but I kind of don't find out how	21:38
	→ dropdrive joined	21:38
Vinnie_win	can I roll a commit back on both my local and remote repos with "git reset --hard HEAD^" ?	21:39
	→ EricInBNE joined	21:39
	← krz left	21:39
	→ lyddonb joined	21:39
	← griotspeak left	21:40
	→ krz joined	21:40
	→ alberto56 joined	21:41
	← alberto56 left	21:42
	→ txomon\|home joined	21:42
	← txomon\|home left	21:42
	→ txomon\|home joined	21:42
	← jefferai left	21:43
offby1	that only does it locally.	21:44
	You can then do "git push -f", but: !rewrite	21:44
gitinfo	[!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)	21:44
	→ jefferai joined	21:44
	→ John______ joined	21:44
	→ el2ro joined	21:45
	← jim0203 left	21:45
	→ ipalaus_ joined	21:45
	→ alberto56 joined	21:47
	← alberto56 left	21:47
	→ enmand joined	21:48
	← Destos left	21:48
	→ kristjan joined	21:49
	→ marhaban1 joined	21:50
	← marhaban1 left	21:51
	→ marhaban joined	21:51
	→ marhaban1 joined	21:52
	← marhaban left	21:52
	← koo4 left	21:52
	← glcrazy left	21:52
	→ koo4 joined	21:52
	→ jim0203 joined	21:52
	→ iocor joined	21:53
	← felgru left	21:54
	← dbgster left	21:54
	← gchristensen left	21:55
	← John______ left	21:56
	← marhaban1 left	21:56
	→ eoh\| joined	21:56
	← eoh left	21:57
	→ marhaban joined	21:57
	← eoh\| left	21:57
	← voidx left	21:58
	→ eoh joined	21:58
	← t4nkd left	21:58
	← orkz left	21:58
	→ yhager joined	21:58
	→ voidx joined	21:59
	← ignacio left	22:00
	→ marhaban1 joined	22:01
	→ axl__ joined	22:01
	← marhaban left	22:01
	→ ghvandoorn joined	22:01
	← lwhalen left	22:01
	← axl__ left	22:02
	→ acl_ joined	22:03
	← juanfer left	22:03
	→ axl__ joined	22:03
	← axl_ left	22:04
	axl__ → axl_	22:04
	← dr_win left	22:04
	→ krzkrzkrz joined	22:04
	→ dr_win joined	22:05
	← glennpratt left	22:05
	← SkAzZ1 left	22:05
	← eicto left	22:05
	← marhaban1 left	22:06
	← krz left	22:07
	krzkrzkrz → krz	22:07
	← michaelmartinez left	22:08
	← AaronMT left	22:08
	→ fission6 joined	22:08
	← w0ng left	22:08
	→ w0ng joined	22:09
	← chrisgeorge left	22:09
	← timoxley left	22:09
	← _iron left	22:10
	→ chrisgeorge joined	22:11
	← martinjlowm left	22:11
	← koo4 left	22:11
	← kermit left	22:11
	← dv310p3r left	22:12
	→ kermit joined	22:12
	← estevo\|afk left	22:12
	→ es joined	22:12
	← macabre left	22:15
	← rendar left	22:16
	← iocor left	22:16
	← dr_win left	22:16
	→ koo3 joined	22:17
	→ dr_win joined	22:17
	← dpino left	22:18
	← lyddonb left	22:18
	← Russell^^ left	22:19
	→ infogulch_ joined	22:19
	← voidx left	22:20
	← marcogmonteiro left	22:20
	→ tidalwav1 joined	22:20
tidalwav1	Seems like there was a change in a recent Git release that prompts for a commit message when a recursive merge is made with no conflicts? Git never used to prompt for the message, how do I get back the previous behavior?	22:21
	← koo3 left	22:21
	← dstorrs left	22:22
	→ dstorrs joined	22:24
	→ robotmay joined	22:24
	← w0ng left	22:25
	→ coyotebush joined	22:25
	→ w0ng joined	22:26
	→ koo3 joined	22:26
CareBear\	so	22:26
	→ the_drow joined	22:26
CareBear\	$ git status	22:26
	fatal: failed to read object d5b7e80f64e63ec67f00e417a4938b539e4bdc87: Invalid argument	22:26
the_drow	hi, I kinda don't get what to do with git add -i	22:26
CareBear\	the_drow : p	22:26
the_drow	with hg record everything is simple and clear for me. I'm not much of a git user though	22:26
CareBear\	the_drow : then you get to select and/or edit hunks before you add them to the index	22:27
the_drow	I want to select what to add to commit	22:27
FauxFaux	CareBear\: Filesystem has corrupted that object. Backup, move it out of the way, git fsck, pray. That, or give up and reclone.	22:27
	← koo3 left	22:27
the_drow	now what do I choose? patch? update? and what do I do after that?	22:27
CareBear\	FauxFaux : no clone, the object is 0 bytes	22:27
	the_drow : I said: p	22:27
the_drow	link to RTFM is also ok	22:27
	← mastro left	22:27
CareBear\	man git-add	22:27
gitinfo	the git-add manpage is available at http://jk.gs/git-add.html	22:27
	← Bass10 left	22:27
FauxFaux	CareBear\: Yeah, 0 bytes is corrupt.	22:27
CareBear\	FauxFaux : I have two 0-byte objects	22:28
FauxFaux	Uh oh. Still, move-and-pray?	22:28
CareBear\	I renamed them	22:28
	bad object HEAD	22:28
	← syphar left	22:28
	→ koo3 joined	22:29
	→ dvide joined	22:29
the_drow	oh for crying out loud. hg prints the diff it wants me to decide about. am I supposed to guess?	22:29
	still don't get it	22:29
CareBear\	the_drow : choose files first	22:29
	the_drow : you can also do all that in the command: git add -p file.c	22:29
mattalexx	the_drow, git add -p	22:29
CareBear\	the_drow : did you select p yet?	22:29
the_drow	nope	22:30
	← Licenser left	22:30
CareBear\	fourth time: select p	22:30
the_drow	done	22:30
FauxFaux	CareBear\: What did you /do/ to your filesystem? Dropbox? Virtual machine shared disc? SMB/NFS? Laptop ramdisc?	22:30
CareBear\	now you get a list of files	22:30
FauxFaux	CareBear\: cat .git/HEAD; if it's ref: foo/bar, go look in foo/bar etc.	22:30
CareBear\	FauxFaux : neither, but the filesystem was full at one point	22:30
	FauxFaux : it's not impossible that I tried to do a commit when there was no space free	22:31
FauxFaux	I very much doubt that git would have fiddled with HEAD after failing to write some objects. Maybe it has a bug, though.	22:31
CareBear\	the_drow : you should get a list of files?	22:31
	← koo3 left	22:31
	← Dave^\| left	22:31
the_drow	thanks. woot. s is nice. hg record doens't have that and it should :)	22:31
CareBear\	the_drow : select the files you want, then it'll iterate over the hunks in each file	22:31
	FauxFaux : unclear, the problem seems to be a few months old (when I last poked at this repo)	22:32
	→ envi_ joined	22:32
CareBear\	yes, HEAD master d5b7e80f64e63ec67f00e417a4938b539e4bdc87 which is one of the two 0 objects	22:32
	← frnkkpp left	22:32
CareBear\	how to identify commit objects?	22:33
	fsck	22:33
FauxFaux	git rev-list --all # if it'll play.	22:33
CareBear\	a bunch of dangling blobs and trees	22:33
	about one per file	22:33
the_drow	ok I now have two commits, how do I merge them into one commit and tell git to repeat that on the remote?	22:33
CareBear\	the_drow : git rebase -i HEAD^	22:34
FauxFaux	CareBear\: Completely ignore dangling output from fsck, it's irrelevant.	22:34
CareBear\	the_drow : make that git rebase -i HEAD^^ actually	22:34
	← drizzd left	22:34
	← CannedCorn left	22:34
CareBear\	the_drow : change pick for second commit to squash (if you want to combine commit messages) or fixup (if you want to use the first commit message and discard the second)	22:34
	→ koo3 joined	22:34
	← Davey_ left	22:34
CareBear\	nay, rev-list is also unhappy	22:35
	← mikepack left	22:35
CareBear\	--reverse also doesn't do it	22:36
	→ mgpcoe joined	22:37
FauxFaux	find .git/objscts -exec git cat-fil..	22:37
	(brb)	22:37
CareBear\	thanks	22:37
	.git/objects $ ls -1 /\|git cat-file --batch	22:38
	07/bd35f4b3e6f189f08bdb3ad826619c429d0a2b missing	22:38
the_drow	thanks	22:38
CareBear\	it says missing for all objects :\	22:38
	← Sargun_Screen left	22:39
kevlarman	CareBear\: probably because of the /	22:39
tidalwav1	anyone on my question from before?	22:39
CareBear\	kevlarman : yes you're right, it needs object identifiers and not filenames	22:40
sitaram	SethRobertson, kevlarman: lwhalen has apparently left, but what he needed was just "gl-setup -q admin-name.pub" (using the correct pubkey), followed by "git push -f" from the most recent clone of the admin repo	22:40
	SethRobertson, kevlarman: he did not mention losing ~/repositories, only ~/.gitolite (because instead of "/var/gitolite/" he gave "/var/gitolite/*" to the rsync command and lost dot files)	22:41
	(he also needs to recreate ~/.gitolite.rc but that is not so hard for most sites)	22:41
	← envi_ left	22:41
the_drow	ok now after the rebase when I push it asks me to merge with remote which I don't want. I want to overwrite with my rebase. How can it be done?	22:41
	← brngardner left	22:41
	← guampa left	22:42
CareBear\	the_drow : please pastebin exact output from your git push command?	22:42
	← kristjan left	22:42
	← teweWORK left	22:42
acl_	tidalwav1: try adding "--no-edit" to your git merge command (ref: http://git-blame.blogspot.com/2012/04/git-1710.html)	22:43
	← DarkAR left	22:43
the_drow	CareBear\: http://pastebin.com/XBiCcsYa	22:43
	← tewecske left	22:43
	→ iocor joined	22:44
tidalwav1	acl_: Thanks very much, that explains it	22:44
	acl_: found it in the manpage for merge right before you told me :)	22:44
acl_	tidalwav1: np :)	22:44
the_drow	CareBear\: -f? :P	22:45
CareBear\	FauxFaux : ah no I remember now, I was working on the code, with an SSD, until the battery drained	22:45
	← flaguy48 left	22:45
	← jave left	22:45
	← tidalwav1 left	22:45
CareBear\	the_drow : yes, you can push -f but please see about !rewriting	22:45
gitinfo	[!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)	22:45
	← master-lincoln left	22:45
FauxFaux	CareBear\: Uh oh.	22:45
the_drow	CareBear\: if I am working alone currently, is this a non-issue?	22:47
CareBear\	FauxFaux : all right, I think I've got the most recent commit. let's see	22:47
	the_drow : yes	22:47
	← tyatpi_ left	22:48
CareBear\	FauxFaux : reflog is empty	22:48
FauxFaux	You can just reset --hard to a commit.	22:48
CareBear\	no thanks - then I will lose uncommitted changes	22:48
the_drow	awesome, thanks	22:48
CareBear\	the_drow : you're welcome	22:48
FauxFaux	You can just reset --something to a commit, then.	22:48
the_drow	now I thought that git was the king of rewriting history	22:49
	→ jave joined	22:49
CareBear\	the_drow : it is	22:49
the_drow	why it can't just repeat what I've done everywhere	22:49
CareBear\	the_drow : you can, but you will get different commits	22:49
	the_drow : because the commit includes timestamps	22:49
the_drow	I mean, I get that with mercurial. I just tell everyone what I've done and expect them to do the same.	22:49
eph3meral	so, I've noticed that when I git pull --rebase origin master while in my feature branch	22:49
	and then keep working	22:49
	← ceej left	22:49
eph3meral	and then master changes a little more, and then I rebase again to get the new changes	22:49
	I have to go through the ENTIRE rebase process starting from my original branch point, which is hell after a topic branch matures for a while	22:50
	is there a way around this?	22:50
the_drow	CareBear\: so they will be a different rev?	22:50
	→ ceej joined	22:50
CareBear\	eph3meral : sync often with upstream	22:50
	the_drow : the commit will be different because it has a different timestamp (and presumably also a different committer)	22:51
	the_drow : "rev" isn't a great term in git lingo :)	22:51
the_drow	so, howcome it is not repeatable?	22:51
CareBear\	the_drow : second time: commits include among other things timestamp and committer	22:51
	→ eletuchy joined	22:51
the_drow	so change the commiter?	22:52
CareBear\	no	22:52
	send them their commit	22:52
	sorry	22:52
	send them your commit	22:52
	→ SkAzZ joined	22:52
the_drow	that is, push to their repository so they will push?	22:52
CareBear\	push to a repository which they pull --rebase from is one way	22:53
	sending a patch via email is another	22:53
	see man git send-email	22:53
gitinfo	the git-send-email manpage is available at http://jk.gs/git-send-email.html	22:53
CareBear\	push is slightly more reliable for a few corner cases where email doesn't work	22:53
	→ timoxley joined	22:53
bremner	yeah, sending 8 bit patches with git-send-email is chancy	22:54
the_drow	ok, got it. I think. Why not just keep the committer and the timestamp?	22:54
	← dzonder left	22:54
CareBear\	bremner : git am can not take commits which have patch-looking content in the commit message	22:54
	the_drow : define keep?	22:55
	← eletuchy_ left	22:55
the_drow	that is that the changeset will have the same timestamp and commiter but it will be a different changeset	22:55
CareBear\	changeset is also not a git term	22:56
	→ lyddonb joined	22:56
CareBear\	the_drow : I don't understand your question I'm afraid	22:57
	→ brngardner joined	22:57
the_drow	too tired to phrase it	22:57
	2:00 am here	22:57
	time for bed I guess.	22:57
	← lonewulf` left	22:57
the_drow	I'll come back tomorrow. promise.	22:57
	ttyl	22:57
	← Superlokkus left	22:58
CareBear\	the_drow : git considers that two people working on the same file and changing the same parts of it and committing it independently should lead to a conflict	22:58
the_drow	oh right	22:58
CareBear\	there is actually an exception to this	22:58
the_drow	oh?	22:58
CareBear\	when doing rebase, if a local commit is identical in contents to a commit fetched from the remote, then the local commit will be discarded	22:59
	(and the fetched commit used)	22:59
the_drow	nice	22:59
CareBear\	(it will then have been applied before all local commits)	22:59
	← fission6 left	23:00
	→ lonewulf` joined	23:00
	→ hotwings joined	23:00
	→ kaen joined	23:01
	← kaen left	23:01
	→ kaen joined	23:01
the_drow	git is powerful, but it is kinda dangerous. it's the reason we choose hg at work. and for ppl who have worked only with svn or worse, with sourcesafe it's a progress	23:01
	← neersighted left	23:01
the_drow	anyway	23:01
CareBear\	it's really difficult to lose data with git	23:01
the_drow	I'm kinda crashing here.	23:01
	← cime left	23:01
the_drow	I'll be back	23:01
eph3meral	CareBear\, what do you mean by "sync" ?	23:01
the_drow	promise :)	23:01
eph3meral	CareBear\, when you say "sync often with upstream"	23:01
	isn't that the purpose of git pull --rebase origin master?	23:02
CareBear\	the_crow : commits can appear to be lost, but any unused commits get stored for some 90 days before they are garbage collected	23:02
eph3meral	to sync my current branch with master?	23:02
hotwings	hi.. how would i download the 10 newest commits, each into their own directory? i cant find anything to tell git to 'clone commit hash <id>' like you can do with mercurial for example (using hg clone -r <revision #>)	23:02
	→ neersighted joined	23:02
CareBear\	eph3meral : another way to say sync: get your commits into upstream as quickly as possible	23:03
	hotwings : every clone has every commit	23:03
	→ apok_ joined	23:03
eph3meral	CareBear\, ahh, well... way to avoid actually answering the question, but thanks for trying, drink coke, play again	23:04
CareBear\	(unless git clone is otherwise instructed)	23:04
	→ v0n joined	23:04
CareBear\	eph3meral : the point is that if your branch is very long then yes you do have to handle conflicts over and over again	23:04
hotwings	err.. so i would download a fresh clone.. then reverse the latest commit.. copy the dir, reverse the next one, copy the dir, etc etc ?	23:04
CareBear\	hotwings : lord no	23:05
	hotwings : why do you want separate directories?	23:05
	← coyotebush left	23:05
	→ welly joined	23:05
hotwings	something broke and im trying to figure out which commit did the breaking	23:05
CareBear\	hotwings : let's start there. I sense an XY problem. you want X and think Y is how to do it, so you ask for Y. but Y is not the best way to get X.	23:05
	hotwings : for this you should use git bisect	23:05
	indeed, XY problem	23:05
	hotwings : check out man git-bisect but there is also good info about bisect in !book	23:06
gitinfo	the git-bisect manpage is available at http://jk.gs/git-bisect.html	23:06
hotwings	i believe one of the last 10 commits caused the break, i just dont know any way to test each of the last 10 commits	23:06
CareBear\	!book	23:06
gitinfo	There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !gitt !vcbe and !parable	23:06
	← apok left	23:06
	apok_ → apok	23:06
CareBear\	if you can't positively test for the error how can you isolate it?	23:06
	if you can test for the error automatically, then you just have to give a few git commands and then git will find the breaking commit for you	23:07
	it is pretty kick ass	23:07
hotwings	i can test for the error	23:07
CareBear\	perfect	23:07
	← SkAzZ left	23:07
eph3meral	CareBear\, that's only when rebasing, if I were to do a regular merge from master in to my topic branch, that wouldn't be the case... right?	23:08
hotwings	the break is in an output plugin for VDR.. so i start vdr -> test for break.. if its broken, i want to try the previous commit.. recompile, start vdr -> check , and so on	23:08
	← chrisgeorge left	23:08
CareBear\	hotwings : http://progit.org/book/ch6-5.html scroll down to Binary search	23:08
	← duckx left	23:09
CareBear\	eph3meral : possibly, I'm unsure, I don't do much merging	23:09
	eph3meral : I do believe that git "remembers" merge conflict resolutions under some circumstances	23:09
	→ madewokherd joined	23:09
CareBear\	eph3meral : perhaps check out man git-rerere although I don't remember exactly what it does :)	23:10
gitinfo	the git-rerere manpage is available at http://jk.gs/git-rerere.html	23:10
	← Nightfire034 left	23:10
CareBear\	hotwings : you can't automate the test? ok, then just go through bisect manually	23:10
	hotwings : please read the documentation now	23:10
	← s3a left	23:10
	← kevwil left	23:11
	← tatsuya_o left	23:11
	← schnatta left	23:11
eph3meral	CareBear\, nice, now that seems to answer the question, thanks :)	23:11
	→ tatsuya_o joined	23:12
	→ smax joined	23:12
smax	git rebase --abort # Didn't restore my branch to the way it was before running git rebase master	23:12
	I don't want to resolve the conflicts myself	23:12
CareBear\	smax : then who will do it?	23:12
hotwings	no, the test cant be automated because i have to see if the output plugin is working or not (ie: black screen only, or video is working)	23:12
CareBear\	hotwings : you can do it manually. did you check out the docs?	23:13
smax	don't worry about it.	23:13
	how do I restore to a particular commit	23:13
CareBear\	smax : what happened when you did rebase --abort?	23:13
	← macmartine left	23:13
smax	Stuff not worked.	23:14
	how do I restore to a particular commit	23:14
CareBear\	wrong	23:14
	please put the output from git in a pastebin	23:14
	→ robotmay_ joined	23:14
smax	it's not git, it's the project.	23:14
	← robotmay left	23:14
hotwings	yes ive read it just now but i dont see how to compile <commit hash> inbetween each test	23:14
	← ph^ left	23:14
	→ tarepanda26 joined	23:14
CareBear\	hotwings : use your code	23:14
	hotwings : if you don't need to compile it then do whatever else to test it	23:15
	hotwings : you only need to name a known good and a known bad commit	23:15
	← stkrzysiak left	23:15
smax	ffs	23:15
	it's a simple question.	23:15
CareBear\	smax : it's possibly an XY question	23:16
hotwings	im not sure what code you mean.. the vdr plugin code (vdr-softhddevice) that im trying to test the last 10 commits of or?	23:16
CareBear\	smax : I want to understand what happened with your rebase	23:16
smax	the hell should I know.	23:16
CareBear\	smax : "Stuff not worked." doesn't tell me, output from git could	23:16
eph3meral	smax, git reset --hard SOMECOMMITHASHGOESHERE	23:16
	i believe	23:16
	→ ryden joined	23:16
smax	eph3meral: thank you.	23:16
	sheesh.	23:16
eph3meral	smax, git reset --hard HEAD && git clean -d -f	23:16
CareBear\	eph3meral : do you know the XY axiom?	23:16
eph3meral	the second command may be dangerous if you have files you have .gitignore'd but don't want deleted	23:16
CareBear\	hotwings : are you sure that the error is within th elast 10 commits?	23:17
	→ AaronMT joined	23:17
ryden	Hi, is it possible to have a repository with some submodules and use all of them as if it'd be a single repository?	23:17
CareBear\	ryden : short answer no not really	23:17
ryden	Awww.	23:17
hotwings	CareBear\ - about 95% certain, yes	23:17
	→ flaguy48 joined	23:17
CareBear\	hotwings : no use. must be 100%	23:17
	← mritz left	23:17
CareBear\	hotwings : find the first good commit	23:17
ryden	Thanks CareBear\	23:17
CareBear\	ryden : what aspects are important to you?	23:18
	hotwings : it's perfectly fine if the known good commit is actually pretty old	23:18
ryden	CareBear\, what do you mean?	23:18
	→ lunks joined	23:18
	← soulcake left	23:18
CareBear\	hotwings : since it's a binary search some hundred commits will not make a big difference in the number of iterations needed	23:19
hotwings	CareBear\ - definitely within the last 15.. i believe within the last 10.. but last 15 im 100% sure	23:19
CareBear\	ryden : you probably had some particular aspects of git repositories in mind when you asked the question, but the question is very very general, so the answer must be no. but if you specify what you had in mind perhaps the answer will be yes.	23:19
lunks	Hi, I'm using git-svn and for some reason the svn tags are not being translated as tags, but rather branches. Is there something I should do or did I do something wrong? My config file: http://pastie.org/3771204	23:20
	← ajpiano left	23:20
ryden	CareBear\, I have some code that is shared between multiple applications, and also some code that some developers might not see.	23:20
CareBear\	hotwings : great! git checkout master && git bisect start && git bisect bad && git bisect good HEAD~15	23:20
hotwings	CareBear\ - ok, and i do that after i do `git clone git://projects.vdr-developer.org/vdr-plugin-softhddevice.git` right?	23:20
	← tarepanda26 left	23:20
CareBear\	hotwings : no	23:21
	hotwings : did you read the man page?	23:21
hotwings	i read the "binary search" section you recommended	23:21
CareBear\	hotwings : after git bisect good HEAD~15 you test. if works, run git bisect good. if not works, run git bisect bad.	23:21
	hotwings : before that I linked to man git-bisect	23:21
gitinfo	the git-bisect manpage is available at http://jk.gs/git-bisect.html	23:21
smax	git reset --hard HEAD && git clean -d -f # fucked my branch up. thanks. You're lucky I backed it up to a new branch.	23:21
	← smax left	23:21
CareBear\	smax : this is also completely typical for an XY problem	23:22
hotwings	and the plugin binary _doesnt_ need to be recompiled at all? that seems really strange	23:22
CareBear\	hotwings : of course it does. you run git bisect inside your repo where the bug will exist.	23:22
eph3meral	how do I add -y / --no-prompt to my ~/.gitconfig	23:22
	i don't see any docs on this	23:22
CareBear\	hotwings : if you need to clone that vdr-plugin-softhddevice.git for something else then go ahead	23:22
eph3meral	i see it in the git-mergetool man page, but there's in information on what the setting is called in .gitconfig	23:22
	→ eicto joined	23:22
	← v0n left	23:23
	→ master-lincoln joined	23:23
	→ WillMarshall joined	23:23
	→ ajpiano joined	23:23
	← aspotashev left	23:23
	← superdmp left	23:24
	← Element9 left	23:24
eph3meral	also, my google searches are coming up a bit dry	23:24
CareBear\	smax : you asked a straightforward question about how to do something and eph3meral answered exactly how to do it. you can NOT blame eph3meral for you having asked what turned out to be the wrong question. XY problem: you want to do X and think Y is the way to do it, so you ask for help with Y. but Y is actually not the right way to do X. please ask about X.	23:24
eph3meral	lots of close results	23:24
	not quite what I want	23:24
CareBear\	eph3meral : see if there's a config option mentioned in the text for the command line option	23:25
	eph3meral : if not, I don't think there's a way to configure it. but you can make an alias which includes it, using git alias	23:25
	← thomas_b left	23:26
	← hgb left	23:26
	→ robotmay joined	23:26
	→ marhaban joined	23:26
	← marhaban left	23:27
	→ marhaban joined	23:27
	→ soulcake joined	23:27
	← gerard0 left	23:28
	→ dzonder joined	23:28
	← marhaban left	23:28
	← jweinberg left	23:28
	→ marhaban1 joined	23:28
	→ jweinberg joined	23:28
	gitinfo set mode: +v	23:28
	← marhaban1 left	23:28
jweinberg	I have a local repository that I want to push into a directory onto a remote repository, how do I do that while preserving history?	23:28
	← robotmay_ left	23:28
	→ marhaban joined	23:29
CareBear\	no can do	23:29
thiago	git is not a deployment tool	23:29
CareBear\	a repository is a repository	23:29
	→ kevwil joined	23:29
thiago	you can push to a repository, but git will not update anything because of the push	23:29
	→ robotmay_ joined	23:29
thiago	if you want to do that, you can write a script or use one of the many existing ones	23:29
	← marhaban left	23:29
	→ marhaban1 joined	23:29
CareBear\	I like http://joemaller.com/990/	23:30
jweinberg	I'm not trying to deploy, this is the remote I want to make it live on	23:30
	← robotmay_ left	23:30
	← marhaban1 left	23:30
hotwings	CareBear\ - thanks for your help. im going to see if i can fumble through this and figure out which commit caused this breakage. i guess i just have to recompile the plugin after each 'git bisect [good\|bad]'	23:30
	→ swappoke joined	23:31
	← es left	23:31
	→ robotmay_ joined	23:31
	→ marhaban joined	23:31
	→ MestreLion joined	23:31
	→ estevocastro joined	23:31
CareBear\	hotwings : yes exactly, you have to recompile, but if the Makefiles are decent they should only compile what actually changed between each step	23:31
thiago	jweinberg: then remote repositories don't have directories	23:31
	← robotmay left	23:31
	→ robotma__ joined	23:31
	→ marhaban1 joined	23:32
	← marhaban left	23:32
	← mklappstuhl left	23:32
	← robotma__ left	23:32
CareBear\	jweinberg : the closest you can come is to work remotely and: create a subdirectory in the repo, add the repo you want as a submodule, commit, git submodule init && git submodule update	23:32
	← robotmay_ left	23:32
CareBear\	jweinberg : but you'll still not be able to push into it. see joemaller.com/990 for one way to get around that	23:32
	← marhaban1 left	23:32
hotwings	CareBear\ - i use ccache here so that should happen. but either way the plugin is very small and compiles in a matter of seconds anyways	23:33
	→ marhaban joined	23:33
	→ robotmay joined	23:33
CareBear\	hotwings : so you should find the problem quickly. great!	23:33
	← eijk_ left	23:33
	→ eletuchy_ joined	23:33
hotwings	CareBear\ - yup.. just need to wait for others to finish watching their movie so i can take the tv over and test :)	23:33
	← marhaban left	23:33
	← pokeswap left	23:33
	→ marhaban1 joined	23:33
	← codewrangler left	23:34
lunks	I can't get this tag thing with git-svn :/	23:34
MestreLion	I have this scenario: upstream/master----A1---A2 (A2 is a "linear extension" of A1). Now master have moved and I want to rebase both A1 and A2 to the new master, as upstream/master(new)------A1'----A2'. Is this possible in a single command?	23:35
lunks	Could it be that the tags are wrong in the svn server itself?	23:35
	→ unclepiss joined	23:35
	← eletuchy left	23:35
	← Unclemantis left	23:36
CareBear\	MestreLion : do you have A1 and A2 on a separate branch locally?	23:36
	← lunks left	23:36
	← sbell left	23:36
CareBear\	MestreLion : and is your remote called upstream or origin?	23:36
MestreLion	CareBear\: yes, A1 and A2 are branches	23:36
CareBear\	they are separate branches?	23:36
	→ sbell joined	23:36
MestreLion	CareBear\: upstream	23:36
	→ n1x joined	23:37
	← ketas left	23:37
	→ maestrojed joined	23:37
MestreLion	CareBear\: define "separate" ? yes, they are different branches, but A1 is a direct ancestor of A2	23:37
CareBear\	if they are different branches you need one command per branch	23:37
	git rebase upstream/master A1 && git rebase A1 A2	23:38
	← iocor left	23:38
	← marhaban1 left	23:38
	← neurodrone left	23:38
	→ coyotebush joined	23:38
MestreLion	humm, no need of --onto ?	23:38
	← cccaldas_ left	23:38
	→ blackmatrix_ny joined	23:39
	→ hgb joined	23:39
MestreLion	yes, of course not...	23:39
	that's simple enough, thank you CareBear\	23:40
	← unclepiss left	23:40
CareBear\	MestreLion : you're welcome	23:40
	→ Guest87934 joined	23:40
	← sbell left	23:41
	← ldlework left	23:41
	← alexisabril left	23:41
	→ Unclemantis joined	23:41
	Unclemantis → Guest28280	23:42
	→ ldlework joined	23:42
maestrojed	I am trying to revert one file that was part of a previous commit. "git checkout 765ef0e -- folder/file.php" I get no error, no message. Doesn't seem like the file reverted. Is my command wrong?	23:44
	← brngardner left	23:44
	← rockets left	23:44
	tizzo → tizzo-afk	23:44
	→ thomas_b joined	23:44
CareBear\	maestrojed : the term revert only applies to commits, and means creating a new commit which negates the changes of the reverted commits. you checked out a file from a previous commit, and if there is no error then you get no error message.	23:45
	← eph3meral left	23:45
CareBear\	meastrojed : confirm using git status and git diff	23:45
	→ superdmp joined	23:45
	← maurer left	23:45
maestrojed	CareBear ok, no doubt my terminology is off :). I would be fine with this all being a new commit. I just want my old file back. Git status shows no modified files, nothing staged, etc.	23:46
	← dhruvasagar left	23:47
	→ gdoteof joined	23:47
	← ehsan left	23:47
CareBear\	meastrojed : are you sure that 765ef0e has your previous state of the file?	23:47
	maestrojed : perhaps 765ef0e is the commit that changed the file (incorrectly) to it's current state?	23:48
	maestrojed : that would explain why git status shows no changes	23:48
	maestrojed : try git checkout 765ef0e^ dir/file	23:48
	→ rchavik joined	23:49
	← rchavik left	23:49
	→ rchavik joined	23:49
	← spyro left	23:49
maestrojed	CareBear\ thats makes sense, I think you are right	23:49
	→ spyro joined	23:49
	← Slix` left	23:49
	→ dhruvasagar joined	23:50
	→ mister_m joined	23:53
	← rrn left	23:53
	← acl_ left	23:53
	→ envi_ joined	23:54
	← blackmatrix_ny left	23:54
	← d2dchat left	23:55
	← kenperkins left	23:56
	← soulcake left	23:56
jweinberg	http://www.nomachetejuggling.com/2011/09/12/moving-one-git-repo-into-another-as-subdirectory/ is what	23:56
	is what I was looking for	23:56
	may have phrased it poorly I guess	23:57
	← kpreid left	23:57
	← n1x left	23:57
	→ rrn joined	23:58
	← envi_ left	23:59

Logs Search ←Prev date Next date→

Channels Documentation