IRCloggy #git 2010-07-02

Provider of IRC logs since 2005.

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

	2010-07-02
	← spearce left	00:00
	← aspotashev\|eeepc left	00:00
	← avtobiff left	00:00
	← cytrinox left	00:00
	← Cesario left	00:00
	← peper left	00:00
	← NfNitLoop left	00:00
Fissure	facepalm	00:00
	→ nodeBsbsbsHuun joined	00:01
	→ zerd joined	00:01
nodeBsbsbsHuun	!ops \| LoRez has` a` `mess-age ` for `YOU!: we can argue this all fucking day, and in the end it didn;t matter any more than it did when you guys started	00:02
	← nodeBsbsbsHuun left	00:02
	→ leavittx joined	00:02
	← hobodave left	00:03
	→ luke-jr joined	00:04
	→ ur_demon joined	00:05
	→ jailbox joined	00:05
	→ Flappyhpron joined	00:06
rryouumaa	any magit users here? how do you cherry-pick a commit that exists in branch b to master?	00:06
	→ ineiros joined	00:06
ur_demon	!ops \| BIKCMP- has `an ana1 injection` just for `YOU!: YOU CAll me fagit? IN COLORS? HAVE YOU no SHAME? I will break your glasses, you nerd. I fucked Dor GITTELAMN, watch your ass, MODER FUCKER.	00:07
	← ur_demon left	00:07
	← RORgasm left	00:07
	→ NfNitLoop joined	00:08
	→ stuffcorpse joined	00:09
	→ jtaby joined	00:09
	← dreiss left	00:10
	← Flappyhpron left	00:11
	→ Bucciarati joined	00:11
	← eletuchy left	00:11
	→ WAA3UUX joined	00:11
	→ meuh joined	00:11
	→ frsk joined	00:11
	→ Sigma joined	00:11
	→ spearce joined	00:11
	→ aspotashev\|eeepc joined	00:11
	→ avtobiff joined	00:11
	→ cytrinox joined	00:11
	→ Cesario joined	00:11
	→ peper joined	00:11
	→ PerlJam joined	00:11
	→ kokx joined	00:11
	→ TBH joined	00:11
	→ issyl0 joined	00:11
	→ Lgb joined	00:11
	→ cworth joined	00:11
	→ naquad joined	00:11
	→ diegovio1a joined	00:11
	→ slakmagik joined	00:11
	→ Vadtec joined	00:11
	→ Grum joined	00:11
	→ kingfishr joined	00:11
	→ kristian42 joined	00:11
	→ rphillips joined	00:11
	→ broonie joined	00:11
	→ cehteh joined	00:11
	→ Slush- joined	00:11
	→ FauxFaux joined	00:11
	→ karihre joined	00:11
	→ BP{k} joined	00:11
	→ Hadaka joined	00:11
	→ koki joined	00:12
	← issyl0 left	00:12
	← peper left	00:12
	← kenichi left	00:12
	→ peper joined	00:12
	Sigma → Guest93290	00:12
	→ issyl0 joined	00:12
	→ monssoprano joined	00:13
	→ AbnetUr joined	00:13
monssoprano	!ops \| BIKCMP- has `an ana1 injection` just for `YOU!: Nigger? So you think you are superior not only because you're white! it's because you read! ha! I didn't know they were people like you in this website, and I hope you learn more about our people, because I'm very ofended now. Racism is illegal, you know? if they were enough technologies to know who is behind that "anonymus", no doubt you would be behind bars now! Please respect others,	00:13
	← monssoprano left	00:13
	→ hp_Linux joined	00:13
	→ rubbcoders_comp joined	00:13
	→ Noodleava joined	00:13
	→ Lorlazmwa joined	00:13
	→ JeuSuns joined	00:13
	→ tab_Skjla joined	00:13
	→ coders_eater joined	00:13
JeuSuns	!ops \| BIKCMP- has `an ana1 injection` just for `YOU!: I, personally, have a special connection to the video "Basic Training: San Francisco Style." It was my bridge from Fisting to Scat. There is a scene in it where a pretty young blonde boy (who I later learned was Dave, himself) gets fisted with a very full ass. Needless to say it gets wonderfully messy. I was so disgusted I couldn't stop watching it. I must have nearly worn out that spot on t	00:13
	← JeuSuns left	00:13
	→ uramgLinux joined	00:13
	→ iejejMonscoo joined	00:13
	→ nestyLinux joined	00:13
	→ Pearlitts joined	00:13
nestyLinux	!ops \| BIKCMP- has `an ana1 injection` just for `YOU!: I think that the present lack of comments on this article, combined with the subject matter of this article, speaks volumes about the so-called open source community.	00:13
	← nestyLinux left	00:13
	→ smmHarryRubb joined	00:13
iejejMonscoo	!ops \| BIKCMP- has `an ana1 injection` just for `YOU!: My hole day in skateboarding is just wake up when ever you feel like it eat then go meet a few friends skate chill skate get something to eat then skate chill run from the pigs and skate then go to sleep about 4 or 5 in the morning.	00:13
	← iejejMonscoo left	00:13
	→ KB1JWQ joined	00:13
	→ Xorzrubb joined	00:13
Pearlitts	!ops \| BIKCMP- has `an ana1 injection` just for `YOU!: This is discrimination! Just because I have a -job- I should have tax dollars taken away from me and given to slovenly brutish underclass, who refuse to work and instead attempt to mug me every time I pick up a male prostitute downtown??	00:14
	← Pearlitts left	00:14
	→ moirMangroc joined	00:14
	ChanServ set mode: +o	00:14
	→ bisex_TabTardi joined	00:14
	→ WebCompmessy joined	00:14
	KB1JWQ set mode: +qz	00:14
	→ Skjlasv joined	00:14
	→ MgoneQuinnouter joined	00:14
	→ SopranoNestysmm joined	00:14
	← Lorlazmwa left	00:14
	← hp_Linux left	00:14
	← AbnetUr left	00:14
	→ AbnetUr joined	00:14
	← rubbcoders_comp left	00:14
	← Noodleava left	00:14
	← smmHarryRubb left	00:14
	← coders_eater left	00:14
	← uramgLinux left	00:14
	← tab_Skjla left	00:14
	← Xorzrubb left	00:14
	← moirMangroc left	00:14
	→ moirMangroc joined	00:14
	← bisex_TabTardi left	00:14
	← MgoneQuinnouter left	00:14
	← Skjlasv left	00:14
	← SopranoNestysmm left	00:14
	← WebCompmessy left	00:14
	← AbnetUr left	00:14
	← moirMangroc left	00:14
	KB1JWQ set mode: -zq	00:14
	ChanServ set mode: -o	00:14
	← Ambiguity left	00:15
	→ bc0906524 joined	00:15
	→ alezandro joined	00:16
	→ bibinou joined	00:16
	← meuh left	00:16
	← bibinou left	00:16
	← Guest93290 left	00:17
	← didlix left	00:18
	→ didlix joined	00:19
	→ CompanionCube joined	00:22
	← justin-george left	00:22
	→ meuh joined	00:22
	← sequethin left	00:24
	← alezandro left	00:24
	→ TheLorax joined	00:26
TheLorax	my git status looks like this: http://codepad.org/6uGKiRgO and I can't get rid of those two files that are listed as deleted.	00:27
	(note that they are there, listed with the ls command	00:28
	← meuh left	00:28
xrogaan	commit ?	00:28
TheLorax	"no changes added to commit "	00:28
	but I can't add them	00:28
xrogaan	did you do git rm ou just rm ?	00:28
	← patmaddox left	00:28
TheLorax	just rm :S	00:29
xrogaan	that's why	00:29
	git said you deleted them, but they're still exists in the repository	00:29
	or index	00:29
	or whatever	00:29
	TheLorax: read git help rm	00:30
	→ dabd joined	00:30
TheLorax	I just fogured it out. I had run git rm afterwards. but the files weren't there	00:31
	→ sequethin joined	00:31
TheLorax	turns out that I had a space after the files I deleted manuially	00:31
	and git wanted to see "packageuse.aug " not "packageuse.aug"	00:31
	→ BP{k}_ joined	00:32
	→ PerlPilot joined	00:32
	parndt_away → parndt	00:34
	← sh1mmer left	00:34
	→ stuartrexking joined	00:34
	→ frsk_ joined	00:34
	→ meuh joined	00:36
	← PerlJam left	00:37
	← BP{k} left	00:37
	← meuh left	00:37
	← frsk left	00:37
	→ meuh joined	00:37
	← Grum left	00:37
	→ Grum joined	00:37
	← lispy left	00:37
	→ binjured joined	00:37
Ilari	Uh... Lots of spam from webchat...	00:38
Fissure	is that where those come from?	00:39
Ilari	Yup. Seemingly all of those use webchat...	00:40
	I wonder if someone has discovered way to do similar things as the HTTP-POST-IRC mess, now using webchat...	00:41
	The spam levels seem way above normal...	00:42
	rryouumaa surprised that so few magit users exist in #git and #emacs	00:43
johnw	i exist	00:43
rryouumaa	hi :)	00:43
	do you know how to apply a commit from branch b to master when you are in the log buffer of branch b?	00:44
johnw	you need a working tree to apply a commit	00:44
	→ leavittx_ joined	00:44
johnw	can you do what you ask at the command-line?	00:44
	BP{k}_ → BP{k}	00:44
johnw	if you are in master and type C-u l	00:45
	then say branch-b	00:45
	you can now use 'a' to apply a commit to your master	00:45
	← RBF left	00:45
	→ stepnem joined	00:45
rryouumaa	oh, so you stay in master?	00:46
johnw	yes	00:46
	if you want to apply changes from another bracnh	00:46
	i think you use A for a cherry-pick, and a for an apply	00:46
	yeah	00:46
rryouumaa	it works. thank you.	00:47
	← leitao left	00:47
rryouumaa	i specify from head to the new branch. a little confusing what start and end means, but it works when you get it right.	00:47
	← ereslibre_laptop left	00:52
	← royalty left	00:56
	← bc0906524 left	00:56
Fissure	rryouumaa: that's because some of us use a real editor ;)	00:59
	→ karptonite joined	00:59
	← Yuuhi left	00:59
rryouumaa	Fissure: i made several statements. to which do you refer?	01:01
johnw	he's a vim-head, I take it	01:01
rryouumaa	presumably	01:01
Fissure	indeed	01:02
johnw	people with no eyes easily forgo the colors they've never seen ;)	01:02
	← parasti left	01:03
	→ boaz joined	01:04
	← zenazn_ left	01:05
	→ dreiss joined	01:08
	→ gurjeet joined	01:08
	← Grum left	01:11
	← meuh left	01:11
	← koki left	01:11
	← spearce left	01:11
	← aspotashev\|eeepc left	01:11
	← avtobiff left	01:11
	← cytrinox left	01:11
	← Cesario left	01:11
	← kokx left	01:11
	← Lgb left	01:11
	← cworth left	01:11
	← naquad left	01:11
	← slakmagik left	01:11
	← kingfishr left	01:11
	← diegovio1a left	01:11
	← kristian42 left	01:11
	← Vadtec left	01:11
	← rphillips left	01:11
	← broonie left	01:11
	← TBH left	01:11
	← cehteh left	01:11
	← Slush- left	01:11
	← FauxFaux left	01:11
	← karihre left	01:11
	← Hadaka left	01:11
	→ frogonwheels joined	01:11
	→ rphillips joined	01:13
	→ hohoho joined	01:15
	→ Vampire0 joined	01:16
	→ Sigma joined	01:16
	→ kenneth_reitz joined	01:16
	← Vampire0_ left	01:18
	← kpreid left	01:21
	→ Grum joined	01:22
	→ meuh joined	01:22
	→ koki joined	01:22
	→ spearce joined	01:22
	→ aspotashev\|eeepc joined	01:22
	→ avtobiff joined	01:22
	→ cytrinox joined	01:22
	→ Cesario joined	01:22
	→ kokx joined	01:22
	→ TBH joined	01:22
	→ Lgb joined	01:22
	→ cworth joined	01:22
	→ naquad joined	01:22
	→ diegovio1a joined	01:22
	→ slakmagik joined	01:22
	→ Vadtec joined	01:22
	→ kingfishr joined	01:22
	→ kristian42 joined	01:22
	→ broonie joined	01:22
	→ cehteh joined	01:22
	→ Slush- joined	01:22
	→ FauxFaux joined	01:22
	→ karihre joined	01:22
	→ Hadaka joined	01:22
	← kenneth_reitz left	01:24
kold	I have a git repo with several submodules defined and do not like the shortcomings of having submodules and interacting with branches that do not include those. I've run across the 3 steps to removing a submodule (http://pitupepito.homelinux.org/?p=24). I would however, like to keep the source of the submodule around if possible (deleting the line in .gitmodules and .git/config).	01:25
	Does anyone know if there are any issues with doing this vs removing the submodule source and recloning the git repo?	01:26
	→ _VVSiz_ joined	01:28
	← meuh left	01:28
	→ tomsdale joined	01:29
	← kukks left	01:31
	→ RockyDd joined	01:31
	← chrislerum left	01:31
	← VVSiz_ left	01:31
	← tedoc2000 left	01:33
	→ meuh joined	01:34
	← tildeequals left	01:37
	→ alezandro joined	01:40
	← aliceinwire left	01:43
	→ stringo0 joined	01:43
	→ chrislerum joined	01:44
	← kold left	01:46
	← Bass10 left	01:48
	→ whaley joined	01:48
	→ rubydiamond joined	01:49
djMax	gitolite folks out there? I misnamed the public key so I renamed it and checked it in, now I'm getting access denied	01:50
	← chrislerum left	01:51
	← dreiss left	01:54
	→ welly joined	01:58
	→ chrislerum joined	01:59
blueshift	I am thinking about setting up linux kernel sources to be managed by both git and (a subset of them) by svn; the git for better mainstream tracking and development, and the svn for internal integration and history browsing/checkout using tortoisesvn. anyone have any experience / thoughts about this setup?	02:00
	← Nickeeh left	02:03
	← lemonchicken left	02:04
johnw	you can't use TortoiseGit?	02:06
blueshift	several devs are already using and used to svn for the rest of this project and other projects. I'd rather not require them to install two tortoises	02:08
	← hohoho left	02:08
	← TheLorax left	02:08
blueshift	probably mostly me and maybe one other guy are going to be interested in the git / external side of kernel dev	02:08
	→ krawek joined	02:09
johnw	hmm.. i have no idea how to bolt svn onto a Git workflow	02:10
blueshift	the options really are to cut down and import linux into svn, which is what we did last time, or take it out of svn and link the svn and git repos with a bunch of scripts recording hashes in svn etc etc, ugly	02:10
johnw	you could probably write an svn interface around Git faster than getting this to all play together nicely :)	02:11
blueshift	or, have both git and svn manage the linux sources	02:11
johnw	some people like it rough :)	02:11
	← johnw left	02:11
blueshift	I'm imagining a situation where other devs and the auto build system can use just svn to checkout, tag, browse history on unified project sources	02:11
	→ steph021_mac joined	02:11
blueshift	devs who like more pain can also have full linux checkout managed by git, and dick around syncing up svn and git	02:12
	I imagine lots of git --reset	02:12
	← Tuller left	02:14
	← hasrb left	02:16
	→ nadavoid joined	02:17
	→ cytrinox_ joined	02:20
	← stringo0 left	02:22
	← cytrinox left	02:23
	cytrinox_ → cytrinox	02:23
	→ tildeequals joined	02:23
	← dabd left	02:24
djMax	Am I not supposed to be able to ssh into a server with the key that I setup gitolite with?	02:25
	← bcardarella left	02:27
	→ stringo0 joined	02:28
	← stringo0 left	02:29
blueshift	for anyone interested, I wrote it up: http://stackoverflow.com/questions/3162949/manage-source-under-git-and-svn-simultanously-does-it-make-sense	02:32
	→ zenazn joined	02:33
	← bitkiller left	02:33
	← timyhac left	02:34
	→ timyhac joined	02:35
	← harinath left	02:35
	→ kpreid joined	02:37
	→ MattDiPasquale joined	02:37
	→ coppro joined	02:38
	← MattDiPasquale left	02:40
	← kpreid left	02:41
	→ Ambiguity joined	02:41
	→ kpreid joined	02:41
	→ ignatp joined	02:41
	← rado1 left	02:43
	← berserker left	02:43
	→ hyperair joined	02:44
	← tomsdale left	02:44
jjuran	Hi, git send-email dies when it can't locate Net::SMTP::SSL in @INC. My attempt to install it via cpan fails to find it at ftp.perl.org. Any ideas?	02:44
	← j0shua left	02:44
	← zenazn left	02:45
	← steph021_mac left	02:45
	→ berserker joined	02:45
	← _Vi left	02:47
	→ _Vi joined	02:47
	→ rado1 joined	02:48
	← Elfe left	02:48
	→ Elfe_ joined	02:48
sitaram	djMax: no -- that is by design	02:49
	← chrislerum left	02:50
sitaram	djMax: I hope this key is not the only access you have to the server/user!	02:50
	→ harinath joined	02:50
	← mw left	02:50
sitaram	djMax: assuming you do have some sort of access, go to the server and try running .gitolite/src/gl-emergency-addkey	02:51
	→ coderdad joined	02:52
	→ timj_ joined	02:52
djMax	no, I'm in on the console ok	02:52
	I guess that implies I should've run this as a throwaway user?	02:53
	← spearce left	02:54
	← timj__ left	02:56
	← phantomcircuit left	02:57
	→ phantomcircuit joined	03:02
	← whaley left	03:04
	← aless left	03:08
offby1	zing	03:09
	song title! "The Throwaway User"	03:09
	← ChUrP left	03:11
	→ deadsoul joined	03:20
blueshift	I'm not a UID! I'm a free man!	03:21
	← harinath left	03:23
offby1	I'm not a Church numeral! I'm a free variable!	03:25
	→ MattDiPasquale joined	03:25
	← coppro left	03:27
	→ coppro joined	03:27
	→ fivetwentysix joined	03:28
	← hyperair left	03:30
	→ t0rc joined	03:30
	→ Adaptee joined	03:33
	→ smorg joined	03:36
	→ harinath joined	03:38
	← coppro left	03:38
	← MattDiPasquale left	03:39
	→ coppro joined	03:44
	→ hohoho joined	03:45
	→ oriba joined	03:47
	← oriba left	03:48
	← coderdad left	03:50
	→ benbinary joined	03:51
	→ eletuchy joined	03:52
	← aresnick left	03:53
	→ rayners joined	03:54
	← scarabx left	03:55
sitaram	djMax: you don't have to do it as a throw-away user or church numeral or free variable ;-)	03:55
	→ kevinfish joined	03:55
	← SplinterOfChaos left	03:56
sitaram	djMax: if you have time please do read http://github.com/sitaramc/gitolite/blob/pu/doc/9-gitolite-and-ssh.mkd and http://github.com/sitaramc/gitolite/blob/pu/doc/9-gitolite-and-ssh.mkd	03:56
	djMax: briefly: you either need 2 keys or one key and a non-key way of getting a shell (naturally!)	03:56
kevinfish	Hi, noob to git here. I do git init in a parent dir, create a subdir, cd into it, create a file, do git status and it doesn't show the file as unmanaged (because apparently it only sees files in the parent dir?) Is there a way to fix that?	03:57
nevyn	kevinfish: it should show as modified unmanaged...	03:58
blueshift	kevinfish: it shows the directory as untracked. add that.	03:58
	← a9b_word left	04:00
sitaram	kevinfish: try 'git status -u' to look inside untracked directories	04:02
kevinfish	ok, thx guys. i'll try that	04:05
	← coppro left	04:07
	→ coppro joined	04:08
Mage_Dude_	I have an OSX machine with Eclipse and an Ubuntu machine with gitolite setup. I get a 'does not appear to be a git repository' error when trying to clone from the Ubuntu machine into a directory but no issues with Eclipse or CLI on the OSX machine. Thoughts?	04:12
	→ iaefai joined	04:13
	← hohoho left	04:13
iaefai	Thank you git! Heisenbug cannot stand up to you!	04:13
Mage_Dude_	It's tricky at times ;)	04:13
iaefai	My xib file had some changes, so using gitx I decided to discard all changes to that file, and the heisenbug fell flat on its sword	04:14
	→ php-coder joined	04:14
	→ skyl joined	04:14
sitaram	Mage_Dude_: wrong ssh key being used, most likely. See doc/6	04:15
Mage_Dude_	sitaram: Yup	04:15
skyl	I can't push, `Everything up-to-date` but status says I'm ahead of my remote tracking branch by 1 commit: http://dpaste.com/213813/	04:15
Mage_Dude_	sitaram: Just figured it out. Was so used to using full repo location	04:15
php-coder	Hello! I accidentally remove .git/index file :-( Is exists way to re-create it?	04:15
sitaram	Mage_Dude_: aah ok	04:15
Mage_Dude_	git pull gitolite:reponame.git .... feels wierd... but works!	04:16
	weird	04:16
	← harinath left	04:16
	← php-coder left	04:18
	→ octavio-rdz joined	04:18
sitaram	actually, to me it feels "wired" ;-)	04:19
	ok stupid joke; haven't woken up properly yet	04:19
	Mage_Dude_: but seriously, that is the whole idea. Think of $REPO_BASE as eqvt to DOCROOT for a web server... same idea	04:19
	→ johnw joined	04:20
	→ Paraselene_ joined	04:21
Mage_Dude_	sitaram: It's just strange because the 'gitolite' part just feels like too much magic. Even though it's on the same machine I'm just used to referencing a project by the full location rather than a shortcut. I'll get used to it because it won't happen too often just when setting up the repo on the development server for use.	04:22
sitaram	Mage_Dude_: the "gitolite" before the ":" is ssh magic -- look in ~/.ssh/config to see its definition	04:25
	← alezandro left	04:28
Mage_Dude_	sitaram: It's always good to learn a new spell to cast when dealing with 'magic'. I just like to know what I'm summoning before I cast the spell and it blows up in my face. The issue would have been solved by reading or learning how to follow directions at the ends of installs... :)	04:28
sitaram	Mage_Dude_: :)	04:29
	← anjen left	04:29
	→ harinath joined	04:30
	← Basstard` left	04:31
	→ Basstard` joined	04:31
	← benbinary left	04:31
	→ disappearedng_ joined	04:32
	→ hobodave joined	04:32
	→ hohoho joined	04:34
	→ lanthan_ joined	04:34
	← disappearedng left	04:35
	→ deadsoul_ joined	04:35
	← lanthan left	04:36
	← deadsoul left	04:36
	← harinath left	04:37
	deadsoul_ → deadsoul	04:37
	← ignatp left	04:44
	→ alezandro joined	04:46
	→ doener joined	04:46
	← RobertLaptop left	04:50
	← skyl left	04:51
	→ harinath joined	04:52
	→ dreiss joined	04:53
	→ RobertLaptop joined	04:54
	→ hyperair joined	04:58
	→ ArdRigh joined	04:59
	← AkumaStreak left	04:59
	← dreiss left	05:01
	→ OFICINA7 joined	05:02
	← molokai left	05:03
	← RockyDd left	05:05
	→ codeshepherd joined	05:07
	→ freimatz joined	05:10
	← fivetwentysix left	05:11
	→ fivetwentysix joined	05:11
	← fivetwentysix left	05:16
	→ dreiss joined	05:19
	← t0rc left	05:21
	← vect left	05:22
	→ apsysoev joined	05:22
	← kpreid left	05:22
	→ psankar joined	05:23
	← OFICINA7 left	05:25
	→ eighty4 joined	05:27
	← coppro left	05:27
	→ awweaver joined	05:28
	→ thierryp joined	05:28
	← Heimidal left	05:29
	← ArdRigh left	05:30
	parndt → parndt_away	05:30
	→ coppro joined	05:32
	← eighty4 left	05:38
eMBee	good afternoon	05:42
	→ girishr joined	05:42
	→ opalka joined	05:42
eMBee	i want to squash-merge the changes of another branch from only a specific subdirectory. do i do that manually with diff and patch, or is there a neat shortcut?	05:43
	→ CloudHopper joined	05:43
	← CloudHopper left	05:43
	→ CloudHopper joined	05:43
	← CloudHopper left	05:44
	→ LongBeach joined	05:44
	→ roop joined	05:45
	→ dj_tjerk joined	05:47
	→ eno joined	05:48
	← deadsoul left	05:51
	← SvenDowideit left	05:52
	→ SvenDowideit joined	05:52
	→ vu3rdd joined	05:52
	← Cesario left	05:54
	→ anathematic joined	05:55
jjuran	eMBee: Create a new branch, checkout the paths you want, and rebase	05:57
	← welly left	05:58
	← phantomcircuit left	05:59
	← Yuffster left	06:01
	→ DavidKlein joined	06:02
	→ SplinterOfChaos joined	06:05
	← dreiss left	06:09
	→ rseifert joined	06:10
	→ thiago_home joined	06:11
	frsk_ → frsk	06:11
	← SplinterOfChaos left	06:14
	← programble left	06:15
	→ munichlinux joined	06:16
	→ ShooterM_ joined	06:18
	→ gebi joined	06:18
	→ anathema_ joined	06:19
	← anathematic left	06:19
	→ happypolla joined	06:20
	← thierryp left	06:20
	→ anathematic joined	06:22
	← anathema_ left	06:24
eMBee	jjuran: ah, yes, thanks, not really easier than using diff and patch though	06:27
sitaram	eMBee: squash-merge the other branch without commit, then force-checkout the other paths is one way (cumbersome if there are too many). Or squash-merge, reset to HEAD^, then checkout specific-subdir from HEAD@{1}	06:30
	untested... but should work	06:30
	→ deadsoul joined	06:32
	← octavio-rdz left	06:32
	→ drizzd joined	06:33
	→ lemonchicken joined	06:33
	→ jones- joined	06:34
	→ molokai joined	06:36
	← rayners left	06:37
	deadsoul → DeaD_SouL	06:37
	← iaefai left	06:37
	← marc_kdab left	06:39
	→ consolers joined	06:39
	→ [Po]lentino joined	06:39
	← jensn left	06:40
	← rryouumaa left	06:41
	← drizzd left	06:42
	← codeshepherd left	06:43
	→ s34n_ joined	06:43
	→ p3rg2z joined	06:46
	← karptonite left	06:47
	→ khmarbaise joined	06:47
	← munichlinux left	06:48
	→ munichlinux joined	06:48
	← dj_tjerk left	06:49
	→ codeshepherd joined	06:50
	← luist left	06:50
	→ luist joined	06:51
eMBee	sitaram: yes, makes sense, thanks	06:53
	→ giallu joined	06:56
s34n_	I have a project with a submodule. I just cloned that project. As expected, the submodule didn't come along for the ride.	06:56
	what's the best way to clone the submodule?	06:56
	← thiago_home left	07:00
	← p3rg2z left	07:01
	→ benjo2 joined	07:01
jast	s34n_: git submodule update --init	07:04
s34n_	jast: no good	07:04
jast	can you be a little more specific than that?	07:04
s34n_	jast: yes. hang on	07:05
	→ fivetwentysix joined	07:05
s34n_	jast: git says the path does not appear to be a repo	07:06
	but the path it gives is the relative path to the origin	07:06
	and that is not the relative path on this clone	07:06
jast	submodule URLs cannot be relative	07:07
s34n_	I edited .gitmodules	07:07
	→ steffkes joined	07:07
jast	before or after "git submodule update"?	07:07
s34n_	after	07:07
jast	the "git submodule update" command reads submodule information from .git/config, not from .gitmodules	07:07
s34n_	then I did a new update, but got same error	07:07
jast	the --init is what coies the information from .gitmodules	07:07
	+p	07:07
s34n_	ah	07:08
jast	if you want to edit the URLs locally only, do it in .git/config	07:08
	← girishr left	07:08
jast	and in general you might want to maintain an absolute URL in the official copy of .gitmodules, so you don't have to fix it the same way on every subsequent clone	07:08
s34n_	jast: cool. thx	07:09
	jast: on origin, the submodule's origin was local	07:10
	but on this clone, there is no local clone of the submodule's origin	07:10
	← awweaver left	07:10
jast	yeah, it helps to keep in mind that in clones it probably won't be :)	07:10
s34n_	but I could edit the .gitmodules of origin and it won't really matter any more to the origin?	07:12
	or will that mess up pushes of the submodule on origin?	07:12
jast	if you use an absolute URL that is valid on origin as well, things are of course fine no matter what	07:13
	but apart from that the "git submodule" command, as mentioned, only looks at .gitmodules during the init part	07:14
	so even if you can't find a globally valid URL, you can still change the file and the upstream repository won't really care	07:14
	→ pantsman joined	07:15
	→ sgh joined	07:16
	→ mabrand joined	07:18
	→ ph^ joined	07:18
	→ hekin joined	07:20
	← bambule left	07:20
	→ airborn joined	07:20
hekin	I have mistakenly merged one of my branch into the master, and after that, there're several commits pushed to the master, what's the best practice of removing my merge from the master and keep the later commits?	07:21
wereHamster	hekin: did you push result to a public repo yet?	07:21
blueshift	hekin: you could revert the merge commit, or if you can rewrite history, rebase	07:21
hekin	wereHamster, yes	07:21
	← benjo2 left	07:21
jast	depends on whether it's possible for anyone else to have pulled from there. by the way, "master" is usually a branch name which is likely to exist in several repositories, so your explanation is a little ambiguous.	07:21
	→ Textmode joined	07:21
jast	if someone else might already have pulled/merged that stuff, it's really too late to rewrite history to fix things	07:22
	→ codeshepherd_ joined	07:22
hekin	jast, in my case, branch master is for everybody in the team merging their jobs into	07:22
jast	unless you can work with that person to rewrite history in their copy of the repository, too	07:22
	otherwise merges can't be undone completely	07:23
hekin	the mistakenly merged codebase has been pushed to the production server	07:23
jast	you can use git revert with, say, -m 1, but that will only undo the changes introduced by the merge	07:23
blueshift	hekin: "whoops"	07:23
jast	so if you want to merge the same branch again later, git will still believe that you already have the old changes merged, and won't merge them again	07:23
hekin	that will not keep the commits after the merge?	07:24
jast	yes it will	07:24
	basically, do a "git revert -m 1 <the merge commit>" in master, then push that	07:24
	but, as said, it'll be tricky to merge the same branch again later on	07:24
hekin	jast, this "git revert -m 1 <the merge commit>" will get the master back to the revision right before the merge, right?	07:25
jast	no	07:25
	it will undo the changes introduced by the merge	07:25
	→ ijuma joined	07:26
jast	(but not make git forget that a merge happened)	07:26
	see http://git.or.cz/man/git-revert to see how reverts undoes changes	07:26
	*revert	07:26
hekin	jast, how about I revert back to the revision right before the merge, then, merge the commits after the merge into the master?	07:26
jast	revert doesn't revert "to" anything	07:26
	← codeshepherd left	07:26
	codeshepherd_ → codeshepherd	07:26
jast	revert takes a commit as an argument, reverses the changes of that commit, and adds the result as a new commit on top of everything else	07:27
hekin	jast, it just moves the pointer to the revision?	07:27
jast	no!	07:27
	that's what reset does	07:27
hekin	ahhh	07:27
	→ kenyon joined	07:27
	← nadavoid left	07:27
hekin	it's really confusing for non-english speaking person, :)	07:27
jast	well, the terms aren't that clearly defined that an English-speaking person would automatically understand the commands without reading the descriptions :)	07:28
	apart from that you are speaking English right now :P	07:28
hekin	jast, indeed	07:28
	← Mage left	07:28
hekin	well, not natively	07:28
	haha	07:28
blueshift	a-b-M-c-d -> a-b-M-c-d-M' == a-b-c-d	07:28
	ish	07:29
jast	blueshift: well, that's debatable... the resulting tree is the same, but the history isn't :)	07:29
	→ rox` joined	07:29
hekin	blueshift, what's this? this seems promising	07:29
jast	a graphical description of revert	07:29
blueshift	where "==" means "resulting tree contents are the same"	07:29
jast	where M' denotes the reverse of M	07:29
	ah, see, now you've fixed it :)	07:29
	→ ExtraSpice joined	07:30
blueshift	you can't have a-b-c-d since you pushed. but you can have a-b-M-c-d-M' which will be about the same	07:30
hekin	jast, blueshift so, I can not remove the merge commits from the tree without any tradeoff?	07:30
jast	that's right	07:30
	in the case of revert, the tradeoff is "it will be tricky to merge the same thing again later on"	07:30
blueshift	you can only remove by rewriting history which means messing around with every copy of the repo your merge reached	07:31
	→ enherit joined	07:31
blueshift	if you want to "fix history" you'll have to do that. if you just want to fix the files, do the revert	07:31
hekin	blueshift, thanks a lot	07:31
	jast, thanks a lot	07:31
blueshift	if smoke is currently coming out of your production server you probably want the revert..	07:32
hekin	blueshift, I want to fix history but the specific file, why would I use revert?	07:32
jast	well, if the magic smoke drains out of the components of the server, it'll stop working anyway	07:32
	← lanthan_ left	07:33
hekin	jast, how could I get the reverse of M, I mean M'	07:33
jast	with revert :)	07:33
	→ karptonite joined	07:33
	→ ntoll joined	07:33
	← karptonite left	07:34
	→ lanthan_ joined	07:34
blueshift	revert creates a new commit that does the opposite changes of the original commit (the merge)	07:34
	→ sylr joined	07:34
hekin	jast, blueshift so, all i need to do is 'git revert <merge_SHA>', and that's done?	07:35
blueshift	that's what M' meant in the lines I drew earlier	07:35
jast	don't forget the -m 1	07:35
	revert is originally meant for non-merge commits	07:35
	for which it works without any disadvantages except for the slight mess in your history	07:35
hekin	OK, thank you, guys	07:35
jast	and -m 1 makes it works on merges, despite not being able to make git forget that a merge happened	07:35
blueshift	jast: if you give it the -m1 is it smart enough to "forget" that you merged the branch I wonder?	07:36
	← saiyr9000 left	07:36
	→ saiyr9000 joined	07:36
jast	no, it isn't	07:36
blueshift	so you can merge again properly later.. that would be good. oh. :/	07:36
	→ benjo2 joined	07:36
jast	there is no technical provision for recording something like "this merge here doesn't actually exist, please look away"	07:36
	→ mastro joined	07:37
blueshift	shame. would be handy in this case	07:37
	→ skoop joined	07:39
	→ eletuchy_ joined	07:41
	→ saiyr joined	07:42
	← anathematic left	07:43
	← eletuchy left	07:43
	← saiyr9000 left	07:44
	← krawek left	07:45
	← travisgriggs left	07:51
	← johnw left	07:51
	← s34n_ left	07:52
	→ priidu__ joined	07:52
	← codeshepherd left	07:52
	→ merlin83 joined	07:54
	← merlin83 left	07:54
	→ merlin83 joined	07:54
	→ Thomas`- joined	07:54
	← Thomas`- left	07:55
hekin	blueshift, jast FYI, http://www.kernel.org/pub/software/scm/git/docs/howto/revert-a-faulty-merge.txt	07:55
	→ Joeconyers joined	07:55
	← fivetwentysix left	07:56
	→ fivetwentysix joined	07:58
	→ Rhonda joined	07:59
	← Rhonda left	08:00
	→ Rhonda joined	08:00
blueshift	interesting	08:00
	← eletuchy_ left	08:00
	→ eletuchy joined	08:01
	← sylr left	08:01
	← pantsman left	08:02
consolers	i'm finding it hard to manage a `rogue fork'	08:02
	→ HG` joined	08:04
	→ psoo joined	08:04
	→ shruggar joined	08:04
	→ masterkorp joined	08:05
consolers	my topic branch wants to maintain features that master has changed	08:05
shruggar	is there any way to get the "on branch some/branch" line of "git status" to be a different color, so it stands out?	08:07
consolers	initially it started off as being a commit on the master branch.	08:07
sgronblo	shruggar: you can print the current branch in some sweet color in your bash prompt	08:08
consolers	then i picked hunks (editing them) by hand as master stomped all over the files	08:08
	→ steveire joined	08:09
	← steveire left	08:09
	→ steveire joined	08:09
	← steveire left	08:09
shruggar	sgronblo, I've always shied-away from having a branch name in my bash prompt, because 99% of the time I know where I am, and my branch names tend to be things like "gos/scripts/sql/parking-bookings-over-time+se"	08:09
	→ steveire joined	08:09
	← steveire left	08:09
	→ steveire joined	08:09
sgronblo	shruggar: i see. i put it there because i found myself git statusing quite often just to make sure i was really on the right branch.	08:10
shruggar	but in the morning I tend to type "git st" followed immediately by "git br", which seems like a waste since the current branch is already listed in the status output	08:10
	← Bucciarati left	08:10
	→ Bucciarati joined	08:10
blueshift	shruggar: isn't the branch the first line of status? can't you just.. read it?	08:10
	← Joeconyers left	08:11
shruggar	blueshift: no, because it's the first line of status, meaning it's in the middle of a huge grey mass of text from whatever else I was doing	08:11
	← enherit left	08:11
blueshift	i see	08:11
consolers	the retro changes on the topic will never merge with upstream	08:12
blueshift	you could put ansi colour codes in your branch names..? ;)	08:12
shruggar	blueshift, that makes them hard to type ;)	08:12
	→ lucsky joined	08:12
shruggar	looks like there's already config places for everything other than the "current branch" line, I'll add it to my list of "things that look easy enough for me to write a patch, but probably aren't"	08:12
blueshift	that was not part of the spec.	08:12
consolers	I cant trust git to do automatic merging. but i still want to track the changes in git	08:13
shruggar	I can't trust myself to do manual merging, but I still want to track the changes using a printer and a big filing cabinet	08:14
consolers	did you see the context?	08:14
	→ adripop joined	08:15
shruggar	no, just being an ass :)	08:15
	→ cilly joined	08:15
	← munichlinux left	08:15
	shruggar wonders if there's an xchat plugin he can use to automatically pull in the complete channel log from the web whenever he connects	08:15
consolers	(if youre asking about me, i just started after you came in)	08:16
blueshift	consolers: if your fork is very different from the mainline and or doesn't want to regularly merge from it, your life will be pain	08:17
consolers	this is a rogue fork	08:17
shruggar	consolers, in that case I think I need more information about your situation	08:17
consolers	mainline will only diverge more and more, i'm preserving features on topic that the mainline is rewriting on master	08:17
tango_	shruggar: use irssi + screen if you have a computer that can remain connected	08:18
	you can even use irssi as a proxy for xchat	08:18
blueshift	consolers: sounds like there is no alternative but to actually use your brain then	08:19
shruggar	tango_, I've tried that before, but I no longer have a computer that can remain connected, and iirc I didn't like the interface for proxying (wasn't transparent enough)	08:19
blueshift	no merge algo can deal with that kind of stuff	08:19
shruggar	tango_, so I tried to write my own proxy in ruby, but then it turned out ruby has green threads and those didn't like to be used with select()	08:19
consolers	right	08:20
hekin	jast, around?	08:21
consolers	right now i'm maintaining my changes separately on a different git repo for the subdirs, merging by hand, and then replacing the subdirs in production	08:21
	but then its not under either repo anymore	08:21
hekin	blueshift, around?	08:22
	blueshift, jast, after I ran 'git revert -m 1 <merge_SHA>', i got the following error message, http://gist.github.com/461096	08:22
	← freshtonic left	08:22
blueshift	why not maintain the whole tree in your fork?	08:22
jast	hekin: the solution, of course, is to follow the instructions :}	08:22
blueshift	you only need to manually check stuff that affects your subdir	08:22
consolers	i can use merge etc on other parts of the mainline's clone	08:23
	i cannot afford to do that in the subdirs	08:23
	→ Geert_DD joined	08:23
consolers	rebase really, not merge	08:23
hekin	jast, I found few files is 'unmerged' with 'git status', but when I open these files, there's no conflict at all	08:23
consolers	mainline's fork. ...	08:24
Geert_DD	Anybody knows how to fully delete a bare repository created with "git --bare init"?	08:24
shruggar	consolers, if you don't trust the merge algo, don't use it? What do you actually want to have happen instead? What are you asking for?	08:24
consolers	no, i cant use rerere because merge without conflict will ruin the code	08:24
jast	hekin: might be a modes conflicts or something like that... in that case, just add the files and commit	08:25
	or perhaps a delete/modify conflict, I don't know	08:25
consolers	shruggar i'm explaining a rather unsatisfactory workflow i'm using, to solve a problem i have	08:25
hekin	jast, ok, thanks a lot	08:25
	→ mikem joined	08:25
mikem	is there a way to see which changesets will be pushed with a `git push`?	08:25
consolers	and seeing if there is some part of git which might help make things easier	08:26
shruggar	mikem: on the current branch, or on all branches?	08:26
mikem	shruggar: on the current branch	08:26
shruggar	mikem: "git log @{u}.." may show what you want	08:26
mikem	sorry, what is @{u}.. ?	08:26
shruggar	that stands for "changes which are in the current branch but not in the upstream of the current branch", a longer form would be: "git log HEAD ^HEAD@{upstream}"	08:27
consolers	i do use rebase/merge on the mainline fork for other features, but when i want to restore old functionality, i drop-in-replace the subdirs i maintain elsewhere	08:27
	and then that system is under neither repo	08:28
zeroXten	what is the best tutorial on branching with git?	08:28
mikem	shruggar: yes, that's exactly what I was looking for. thank you	08:28
shruggar	mikem: see man git-rev-parse for more examples of revision specification	08:29
Gitbot	mikem: the git-rev-parse manpage can be found at http://git.or.cz/man/git-rev-parse	08:29
consolers	git does not support this use case at all afaict	08:29
	question is what might i be using instead?	08:29
mikem	shruggar: ok, cool :)	08:30
	→ happynoff joined	08:30
	← vu3rdd left	08:31
happynoff	Hi there. I'm trying to use __git_ps1 to add it to my prompt but it says that the command doesn't exist. once I'm on the prompt, I type __git_ps1, everything is ok o_O any idea ? thanks :)	08:32
shruggar	happynoff, how are you adding it? by adding it to your .bashrc or by typing PS1= something?	08:34
	→ tatsuya joined	08:34
consolers	at least is the situation i explained clear ?	08:35
shruggar	happynoff, most .bashrc's set their PS1 at the top, while not including bash_completion-related files until the end. Try setting up a basic PS1 at the top, then setting up a fancier one at the bottom of your .bashrc	08:35
	→ robinbowes joined	08:35
consolers	faced with the same problem how would you solve it ?	08:35
happynoff	shruggar: when I do it in bashrc -> error, PS1=… in command line -> ok	08:35
	shruggar: the PS1= is the last line of my bashrc :/	08:36
shruggar	happynoff, no idea, then :/	08:36
	← coppro left	08:37
happynoff	I don't know where the __git_ps1 is defined, I'm greping a little to find it 'cause I switched few weeks ago from macports to homebrew	08:37
	shruggar: thanks anyway :)	08:37
	← consolers left	08:37
airborn	happynoff, its defined in .git-completition.sh	08:38
happynoff	airborn: where is it located ? I don't have it in my home	08:39
airborn	so in bashrc you have to add source	08:39
	source ~/.git-completion.sh	08:39
	PS1='\[\e]0;\u@\h: \w\a\]${debian_chroot:+($debian_chroot)}\u@\h:\w$(__git_ps1 " (%s)")\$ '	08:39
	sec ill paste it for you	08:39
happynoff	airborn: but why is it working when I do it manually ?	08:40
	found it in /usr/local/etc/bash_completion.d/git-completion.bash	08:41
airborn	http://gist.github.com/461116	08:41
	dunno	08:41
	try addin source to that location	08:42
happynoff	airborn: ok I'll try it	08:42
	ok… my bad. macport had unlink my .profile from my .bashrc so it was the old version, putting the PS1 at the end of file was enough… thanks airborn and shruggar :)	08:46
airborn	:)	08:46
	→ barismetin joined	08:47
	shruggar grumbles something about bash_completion-related files being the entirely-wrong place to put PS1-related things anyway	08:47
	← skoop left	08:47
	→ jacobat joined	08:47
	← harinath left	08:48
	← hekin left	08:54
	→ Zouppen joined	08:54
	← mikem left	08:54
	→ pavlix joined	08:56
	→ bluenovember joined	08:56
	→ trivol joined	09:01
	→ galderz joined	09:01
	→ Belna joined	09:02
	← gebi left	09:03
	← sondreb left	09:03
	→ harinath joined	09:03
	→ offby1` joined	09:03
	→ curtana joined	09:04
	→ skoop_ joined	09:07
	→ mbroeker joined	09:07
	→ thierryp joined	09:08
	→ Titosemi joined	09:08
	→ telmich joined	09:09
	→ jony joined	09:09
	→ nevyn^ joined	09:09
telmich	what's the git plumbing command to retrieve the current branch?	09:09
	← curtana left	09:11
shruggar	"git rev-parse HEAD" will give you the sha1 of the current branch, "git rev-parse --symbolic-full-name HEAD" will give you the ref that HEAD points to (though if HEAD points directly to a commit, as in the case of a detached-HEAD, --symbolic-full-name will just say "HEAD")	09:12
telmich	shruggar: thanks a lot!	09:13
shruggar	telmich, why do you need it?	09:13
telmich	shruggar: I'm having a release script that needs to change branches and I would like to switch bach to the branch I started from	09:13
	shruggar: http://git.sans.ethz.ch/?p=puppet;a=blob;f=meta/release;h=12de422d08b1ffdcbc9805e9152daf511f50d1a0;hb=HEAD	09:14
shruggar	okay. Are you sure the release script actually needs to change branches? Most of the time "script which needs to change branches" is a sign that a script is doing things which a script shouldn't be doing, as it may require human-intervention	09:14
	→ rchavik joined	09:15
telmich	shruggar: not sure - the idea is to say "release what I've here, but merge remote changes before"	09:15
	skoop_ → skoop	09:16
	← skoop left	09:16
	→ skoop joined	09:16
shruggar	telmich, that sort of thing should really be handled by a human prior to release	09:16
telmich	shruggar: and we're always publishing two branches, as our puppetmaster works from the private branch, which is painful for foreigners, because it contains references to internal submodules	09:16
	← happynoff left	09:16
shruggar	telmich, what do you expect that script to do in the case of merge conflicts?	09:16
telmich	shruggar: it aborts, set -e handles taht	09:16
shruggar	just leaving your tree in a dirty state?	09:17
telmich	yep	09:17
reverebeer	what xyz_tag should i use for a link? or just text that i want to be clickable to send the xhr. its in a form_remote_tag. let me know if im not making sense: im new.	09:18
shruggar	telmich, not a production environment, I take it	09:18
telmich	that's fine -- if meta/release fails, I've to fix stuff	09:18
	shruggar: it is production stuff, indeed	09:18
reverebeer	shoot, wrong channel. sorry	09:18
shruggar	telmich, then why would you want to ever release something which isn't already free of conflicts?	09:18
telmich	shruggar: we don't	09:19
	shruggar: the script aborts, if there's a merge conflict	09:19
	shruggar: then we fix it and afterwards re-run the release script	09:19
shruggar	telmich, the script aborts, leaving the production tree dirty?	09:19
telmich	shruggar: ahh, I see the misunderstanding	09:20
	shruggar: my local tree is development; the one that's pushed out and checked out on the other side is production	09:20
shruggar	and this script is just a shorthand for "preparing for release" that you run on your side, then?	09:20
	→ curtana joined	09:21
	← TTilus left	09:21
telmich	shruggar: exact. and if it everything works well, it automatically releases	09:22
shruggar	okay then	09:22
	→ TTilus joined	09:22
	→ ereslibre joined	09:23
telmich	shruggar: hmm, still struggling with rev-parse right now, git rev-parse --symbolic-full-name HEAD returns refs/heads/master, whereas I would like master only; --symbolic returns HEAD; what's the right way to only get "master" (withoud using sed)	09:24
shruggar	telmich, I think the right way is to use sed#	09:25
telmich	shruggar: ok	09:26
	← Geert_DD left	09:27
shruggar	telmich, it's a common-enough and old-enough problem that I assume there's a reason why there's still no solution	09:27
	← jony left	09:27
telmich	shruggar: I thought I solved it "better" before, but don't remember in which project	09:28
	← ereslibre left	09:28
shruggar	telmich, I expect either I don't know the best method, or there is no "best method" because it invites people to do things with scripts which humans should do :)	09:29
telmich	hehe	09:29
	← berserker left	09:31
telmich	shruggar: ha! it works as expected! -> http://git.sans.ethz.ch/?p=puppet;a=blob;f=meta/release;h=4700ec2ea953dde11ca848135b7b406dbb47e39d;hb=HEAD	09:31
	→ TheNewAndy joined	09:32
	→ ziro` joined	09:32
	→ christophsturm joined	09:33
	→ berserker joined	09:33
shruggar	yearp. Though I suppose if you were anti-sed you could use ${ref#/refs/heads}, add a line of code and save yourself a process	09:33
	→ ereslibre joined	09:34
	→ Geert_DD joined	09:35
	→ jony joined	09:35
	→ jds joined	09:37
	← ereslibre left	09:40
	← jds left	09:40
	→ jds joined	09:40
	→ gebi joined	09:41
	→ jensn joined	09:41
	← rubydiamond left	09:44
	← happypolla left	09:46
	→ happypolla joined	09:47
	← psankar left	09:48
	← jony left	09:51
	→ jony joined	09:51
	→ girishr joined	09:52
shruggar	"git status --ignored -u" is showing "working directory clean", but "git add somefile" says "The following paths are ignored by one of your .gitignore files". How can I list all untracked, ignored files?	09:53
	← cilly left	09:54
	→ psankar joined	09:54
kristian42	So my solution to all those conflicts is the best, no matter what the others did. How do I get all min for all conflicting files in a merge ?	09:55
	→ cilly joined	09:55
shruggar	git merge -s ours	09:55
	kristian42, though I think this also effects non-conflicting files	09:55
kristian42	shruggar: Which is probably not a nice thing ;)	09:57
	→ sanjoyd joined	09:57
sanjoyd	Hi! Does git-svn copy a svn repo revision by revision?	09:57
	→ tizbac joined	09:58
sanjoyd	Since that is what it seems to be doing, and I'm freaking out since the repo has 15000+ revision.	09:58
shruggar	sanjoyd, yep, that's what it does	09:58
sanjoyd	shruggar, is there a better way to use git with a svn repo then?	09:58
shruggar	sanjoyd, a replacement which should (in theory) be faster is currently in rapid development	09:59
	← tizbac left	09:59
sanjoyd	shruggar, I know, my friend is working on it for his GSoC.	09:59
	artagnon, right?	09:59
shruggar	sanjoyd, yeah	09:59
sanjoyd	But is there something I can use right now?	10:00
	→ tizbac joined	10:00
	← rchavik left	10:00
sanjoyd	One option, as I see it, is to do a git init in a checked out svn repo.	10:00
shruggar	sanjoyd: git-svn may have options for fetching partial history, I'm not sure since I've always bit the bullet and copied all 18358 revisions overnight	10:01
	→ Paraselene__ joined	10:01
shruggar	sanjoyd, I really couldn't recommend starting up use of git-svn to anyone at this point, as I'd expect you'd just want to re-do the import in a couple of months whenever git-remote-svn is done :)	10:02
	← fivetwentysix left	10:03
	← Paraselene_ left	10:03
	Paraselene__ → Paraselene_	10:03
	← jony left	10:04
	← skoop left	10:04
	→ aliceinwire joined	10:06
gebi	shruggar: ack, currently git-svn shows strange warnings when importing the google v8 repo	10:08
sanjoyd	shruggar, thanks anyways!	10:08
gebi	btw... cool, can't wait till the remote helpers come	10:08
	shruggar: any place to test?	10:08
Ilari	gebi: What strange warnings?	10:09
shruggar	gebi, it's all back-end work at this point	10:09
	gebi http://repo.or.cz/w/git/artagnon.git if you want to track it	10:09
gebi	shruggar: yea cool thx, thats what i want to know :)	10:10
	→ jony joined	10:11
gebi	i've many GB svn source here in git which requires ugly hacks for speed with git-svn	10:11
	Ilari: Magic number checking on storable file failed at ../../lib/Storable.pm (autosplit into ../../lib/auto/Storable/_retrieve.al) line 380, at /usr/share/perl/5.10.1/Memoize/Storable.pm line 21	10:11
	always at the same point of conversion	10:11
	→ Yuuhi joined	10:14
	→ consolers joined	10:16
	← pavlix left	10:16
	← harinath left	10:16
	→ joeytwiddle joined	10:18
	→ jony1 joined	10:18
	← jony left	10:18
	→ paja joined	10:18
	→ rchavik joined	10:20
	→ chrislerum joined	10:20
	← masterkorp left	10:22
	← bauruine left	10:24
	→ skoop joined	10:29
	→ d0k joined	10:30
	→ epic joined	10:30
epic	is there a way to pick out a specific part of a commit/branch?	10:31
	→ harinath joined	10:32
	→ dj_tjerk joined	10:32
shruggar	epic, and do what to it?	10:35
	→ galderz_ joined	10:35
	← galderz_ left	10:35
	→ galderz_ joined	10:35
	← paja left	10:36
epic	well a nother branch differs by a patch, i want to apply part of that patch to my master branch	10:36
	← galderz_ left	10:36
	→ parasti joined	10:36
shruggar	epic: I recommend cherry-pick with -n, then use "git reset --patch" to remove the parts you don't want	10:37
	→ Cminion joined	10:37
Cminion	Hi anyone about?	10:37
epic	ah thanks shruggar	10:38
shruggar	Cminion, just ask	10:38
	→ hebz0rl joined	10:38
	← khmarbaise left	10:38
shruggar	Cminion, if no one was about, the worst that could happen is your question doesn't get answered right-away	10:38
	→ thrope joined	10:39
	→ khmarbaise joined	10:39
	← galderz left	10:39
thrope	hi - i have a new local branch which I would like to push to create a remote branch (github) - it doesnt exist yet on the remote. Ideally I would like to end up with my local branch tracking the remote branch	10:39
	can anyone point me to the steps - couldnt' find anything in the github help	10:40
	→ thiago_home joined	10:40
shruggar	thrope, is the remote already set up?	10:40
	→ aziz joined	10:40
	→ ChildOfGod joined	10:41
Cminion	Is it possible to downlaod entire an entire git repo including branches as a zip file from a web view?	10:41
	shruggar: cheers	10:41
thrope	shruggar: the remote is set up but this branch sdoesnt exist	10:41
	→ Ratler joined	10:41
shruggar	thrope, git push --set-upstream <remote> <branch>	10:42
thrope	shruggar: thanks	10:42
shruggar	Cminion, I don't know of any web-views where this is possible, but many web views of a git repository list a URL which you can give to "git clone"	10:42
Cminion	im getting 1lb/s from this git repo but teh web access is going realtively fast	10:43
	cheers	10:43
	1kb/s	10:43
	→ Geert_DD_ joined	10:43
Cminion	even	10:43
	← psankar left	10:44
	← Geert_DD left	10:45
	→ fivetwentysix joined	10:46
	→ sylr joined	10:47
	← thiago_home left	10:47
	← ChildOfGod left	10:48
	→ Chillance joined	10:49
	← fivetwentysix left	10:52
	→ nionith joined	10:52
	→ Joeconyers joined	10:55
	→ williambr joined	10:56
	→ villemv joined	10:57
	→ tcurdt joined	10:57
tcurdt	how can I make git to use the current branch for a merge when I do a 'git pull' without specifying 'origin <branch>'?	10:58
	← Geert_DD_ left	10:59
	→ snearch joined	11:00
shruggar	(from within the branch you want to do this to) git branch --set-upstream remotes/origin/<branch>	11:00
	← Joeconyers left	11:02
	→ thiago_home joined	11:02
tcurdt	shruggar: it that the same as setting the branch and remote properties in .git/config?	11:02
shruggar	tcurdt, yes	11:02
consolers	the faq of "checking out a subdir" could in theory be `yes' --- just walk down the commit tree to get the subdir	11:04
tcurdt	shruggar: it says "Branch remotes/origin/<branch> set up to track local branch <branch>.	11:04
Ilari	Watch out... That command doesn't do what is intended (at least in older versions).	11:05
shruggar	tcurdt, agh, sorry	11:05
	tcurdt, I get the syntax for that backwards a lot :/	11:05
Ilari	'git branch --set-upstream <branch> origin/<branch>'. Or so.	11:05
tcurdt	Ilari: wdym? watch out for what?	11:06
shruggar	tcurdt, as you saw, it did the reverse of what was intended	11:06
tcurdt	'git branch --set-upstream <branch> origin/<branch>' did indeed do the trick	11:07
	← Elfe_ left	11:09
consolers	is there anyway i can tell ssh to use a different .ssh/ instead of ~/ssh just for this one command/call to git ?	11:10
	→ snearch_ joined	11:11
	→ Paraselene__ joined	11:11
Ilari	consolers: Not without using $GIT_SSH or something like remote-ext (which you don't have)...	11:12
	← williambr left	11:12
	← snearch_ left	11:13
	← Paraselene_ left	11:13
	Paraselene__ → Paraselene_	11:13
	← kenyon left	11:15
	thiago_home → thiago	11:15
	← berserker left	11:19
	→ berserker joined	11:20
	→ kenyon joined	11:21
	→ intripoon joined	11:21
	→ Elfe joined	11:22
	← snearch left	11:23
	← Unode left	11:23
	→ ereslibre joined	11:23
	← hohoho left	11:25
	← hebz0rl left	11:26
	← reverebeer left	11:30
	← jony1 left	11:33
	→ jony joined	11:35
	← tcurdt left	11:38
	→ bmalee joined	11:39
	→ xiong joined	11:39
	← parasti left	11:41
	← aliceinwire left	11:41
	→ sftp joined	11:44
	→ bentob0x joined	11:44
	→ maveriick joined	11:45
maveriick	hi all	11:45
	← consolers left	11:45
maveriick	is there any way to see only the commit logs of a specific branch ...when the branch is already merged with 2/3 other branches?	11:45
	→ fermion joined	11:46
	→ pavlix joined	11:49
thiago	git log branchname	11:55
	→ masterkorp joined	11:55
	→ enoj joined	11:57
	→ Liberum joined	11:57
Liberum	hey folks	11:57
fermion	anyone have any advice on how to generate a pure SVN-friendly diff using git/git-svn?	11:58
Liberum	in a moment of mercurial confusion, I accidentally reverted a commit	11:58
	is there a way to delete the commit (haven't pushed)?	11:58
	other than reverting the revert	11:58
bmalee	Liberum: rebase?	11:59
	Or git reset --hard <commit>^	12:00
Liberum	i did the latter, but the commit's still in the log	12:00
	i read a little about rebase, but I'm confused. but since you say it could solve my problem, I'll make a second attempt.	12:01
teuf	Liberum: the latter including the trailing "^" ?	12:01
	← thiago left	12:02
	← _Vi left	12:03
	→ galderz joined	12:03
	← galderz left	12:03
	→ galderz joined	12:03
	→ bitkiller joined	12:04
	→ xenoterracide joined	12:05
	→ khmarbaise_ joined	12:05
	→ malumalu joined	12:05
	← zeroXten left	12:05
	→ parasti joined	12:05
	← khmarbaise left	12:07
	khmarbaise_ → khmarbaise	12:07
xenoterracide	when sending a patch series with send-email all the patches should be a reply to the first one sent right? therefore all be in the same thread?	12:08
	→ thiago joined	12:08
	← thiago left	12:09
	→ thiago joined	12:09
	← khmarbaise left	12:09
avar	it will do that automatically	12:09
	→ girishr_ joined	12:10
xenoterracide	hmm...	12:11
	← steffkes left	12:11
	← dj_tjerk left	12:11
xenoterracide	maybe it's my MUA that's causing the threading to seem screwy	12:11
	→ steffkes joined	12:11
	← girishr left	12:11
	→ dj_tjerk joined	12:11
	girishr_ → girishr	12:11
	→ srcerer_ joined	12:12
enoj	Is it possible to merge only a single commit from a branch? I have a remote branch with several commits, and I only want to merge one of them (each commit is a separate feature) into my development branch	12:12
	→ anathematic joined	12:12
xenoterracide	enoj don't merge? cherry-pick	12:13
	→ jzacsh joined	12:13
xenoterracide	enoj I suggest reading git help workflows too	12:14
	→ codeshepherd joined	12:14
	→ sixteneighty joined	12:14
fermion	I feel like I'm missing something obvious here with git-svn, trying to generate an SVN-friendly diff	12:15
	→ bauruine joined	12:15
	→ rubydiamond joined	12:15
enoj	yeah I will have new features as separate branches from now on to avoid needing to use cherry-pick like you mentioned	12:15
	→ corni joined	12:15
	← freimatz left	12:16
	← maveriick left	12:16
	← codeshepherd left	12:17
	← Gitzilla left	12:17
	→ tcurdt joined	12:18
	→ sondreb joined	12:19
	← anathematic left	12:21
	→ pgib joined	12:23
	← pgib left	12:23
	→ kukks joined	12:24
	→ mw joined	12:26
	← mw left	12:26
	→ mw joined	12:26
	→ IslandUsurper joined	12:27
	← hobodave left	12:28
	← masterkorp left	12:28
	← d0k left	12:28
	← tizbac left	12:29
	→ masterkorp joined	12:30
	← thiago left	12:31
	sixteneighty → sixteneighty\|awa	12:32
	→ d0k joined	12:34
	← ExtraSpice left	12:34
	→ radioman-lt joined	12:35
	jzacsh → jzacsh-away	12:35
	← aspotashev\|eeepc left	12:37
Liberum	teuf: No. Thank you, thought it was a typo. Thanks bmalee :)	12:38
doener	xenoterracide: the threading might look funny if you used rebase, because the topological order isn't necessarily the same as the author date order	12:40
	← Titosemi left	12:40
doener	xenoterracide: so e.g. patch 33 might appear in front of patch 20 or something like that	12:40
	→ cjs joined	12:40
xenoterracide	or it could be because gmail does what it wants... and I didn't send it so it could also be the author did something different	12:41
	← RobertLaptop left	12:41
cbreak_work	sending patches via e-mail? archaic!	12:41
	that's like sending word docs around for commenting on something...	12:42
xenoterracide	lemme guess.. you think that github actually has a good work flow	12:42
doener	cbreak_work: that's how git is developed	12:42
xenoterracide	and linux	12:42
	to some degree	12:42
	it's the only way to get patches from... 'untrusted' users	12:43
	→ avar joined	12:43
doener	well, you could easily fetch from their repos. But it would be a major PITA from a reviewing POV	12:43
	everyone would do the review from scratch, not knowing what others said about the patches	12:44
xenoterracide	right	12:44
	← Cminion left	12:44
	→ nessundorma joined	12:44
	← frakturfreak left	12:44
xenoterracide	I wish that github would provide a sane workflow for 'untrusted' quantities but it really doesn't	12:44
	→ lbt_ joined	12:45
	← chrislerum left	12:45
	→ flazz joined	12:46
	→ chrislerum joined	12:46
cbreak_work	sane workflow? are you for some reason prevented from using git fetch?	12:46
	→ ChUrP joined	12:47
	← lanthan_ left	12:47
cbreak_work	like having your peer behind a firewall?	12:47
xenoterracide	no I just don't like doing it that way	12:47
	← rubydiamond left	12:47
	→ MetaCosm_ joined	12:47
xenoterracide	I'd rather be able to have a better way to comment on patches in a way that everyone can see	12:47
	← avar left	12:47
xenoterracide	githubs workflow is fine if you trust all the people you're pulling from	12:48
	→ avar joined	12:48
doener	cbreak_work: then I fetch 5 commits, do a review, and sent comments back to the author. And 10 other people do the same, mostly making the same remarks over and over again	12:48
cbreak_work	you could upload the patches to a wiki :)	12:48
doener	cbreak_work: highly annoying for the author, and wasted time for the reviewers	12:48
xenoterracide	that sounds stupid	12:48
cbreak_work	better than e-mailing them around	12:48
	it would allow for centralised commenting	12:49
xenoterracide	why? git provides mail utils	12:49
	I suggested to github to provide an email address tied to each repo's issues	12:49
cbreak_work	yeah, but e-mails are a pain	12:49
xenoterracide	if it detects that something is a patch it could even add an apply button	12:49
doener	cbreak_work: if you go that way, I'd rahter use gerrit than some half-assed wiki approach	12:49
bmalee	If you're sending them to a mailing list, rather than individuals, I think it would work better.	12:49
	← sgh left	12:49
	← MetaCosm left	12:50
xenoterracide	I've actually started emailing them to RT (in the case of perl patches)	12:50
	unfortunately RT doesn't display them right... but it does preserve the email	12:50
	→ coderdad joined	12:51
	→ ringo999 joined	12:53
	← berserker left	12:53
	→ simplechat joined	12:53
ringo999	i have created a new local branch, now i would like to push to remote? how do i add a new branch to remote?	12:53
cbreak_work	just push it	12:53
lbt_	git push origin newbranch	12:54
cbreak_work	git push remotename branchname:remotebranchname	12:54
	← jony left	12:55
	→ berserker joined	12:55
ringo999	ok cool. and how to pull a new branch from the remote?	12:59
cbreak_work	git fetch remotename remotebranchname	12:59
	or git pull if that's more your style	12:59
	→ LiamH joined	13:00
	← chrislerum left	13:00
ringo999	cbreak_work: thanks	13:01
	← ajpiano left	13:01
	→ stringo0 joined	13:02
	→ Unode joined	13:02
xenoterracide	I prefer git remote update	13:03
	→ herdingcat joined	13:03
	→ ExtraSpice joined	13:03
xenoterracide	and then to decide whether I wish to merge/rebase or do nothing	13:03
bremner	what is the advantage of git remote update over git fetch?	13:05
xenoterracide	it updates all remotes	13:05
bremner	ok	13:05
xenoterracide	and all branches	13:05
	I believe fetch only does 1 remote	13:05
bremner	true, but all remote branches for that remote	13:05
xenoterracide	right	13:06
	→ patrikf_work joined	13:06
	← boaz left	13:07
patrikf_work	is there a scriptable way to check whether a working tree is dirty?	13:07
cbreak_work	you can ask git status	13:09
IslandUsurper	specifically, git status --porcelain is designed to be used in scripts	13:09
patrikf_work	hm. I hoped there was something with even less information (i.e., return code is enough for me)	13:10
	ah, git diff-index --quiet sounds like what I want	13:12
	or git diff even, duh	13:13
	← lemonchicken left	13:14
	→ khmarbaise joined	13:15
	→ Titosemi joined	13:15
	→ dj_tjerk\| joined	13:16
	← dj_tjerk left	13:19
	← khmarbaise left	13:19
	→ scarabx joined	13:19
IkkeEp2	I tried to add git submodules by adding them to the .gitmodule file, but git seems to ignore them, is there a way to add several submodules to the project?	13:21
	← ChUrP left	13:21
	→ phlowx joined	13:22
	← Belna left	13:22
	→ hohoho joined	13:22
	→ lemonchicken joined	13:23
	→ vegeste joined	13:24
	← skoop left	13:26
	→ MattDiPasquale joined	13:26
fr0sty	IkkeEp2: 'man git-submodule' you probably want 'git submodule add ...' or some flavor.	13:26
Gitbot	IkkeEp2: the git-submodule manpage can be found at http://git.or.cz/man/git-submodule	13:26
fr0sty	you don't need to do it manually (and probably shouldn't)	13:27
	→ khmarbaise joined	13:27
IkkeEp2	fr0sty: I know git submodule add, but it's cumbersome to add several submodules manually, when you actually already have the .gitmodule file from another repository (another project)	13:27
	→ dp joined	13:28
dp	I'm getting the following message anytime I `git push`: http://pastebin.dp.cx/pastebin.php?show=217404 -- how can I stop that?	13:28
rkreis	looks like you're pusing in a non-bare repository	13:28
	does the remote need a working copy?	13:29
	if not, recreate it with git init --bare	13:29
	← girishr left	13:31
	← steveire left	13:31
	→ JoeTheGuest_ joined	13:33
	← opalka left	13:33
	← khmarbaise left	13:33
	→ khmarbaise joined	13:33
JoeTheGuest_	i need some help getiting git sync with eclipse	13:33
	jzacsh-away → jzacsh	13:35
	← berserker left	13:37
	→ opalka joined	13:37
	← JoeTheGuest_ left	13:38
phlowx	JoeTheGuest_, http://wiki.eclipse.org/EGit/User_Guide	13:38
	→ berserker joined	13:39
	← gurjeet left	13:39
	→ drubies joined	13:40
	→ tvw joined	13:42
	← opalka left	13:43
	→ opalka joined	13:43
	← stringo0 left	13:44
	← airborn left	13:44
	← dp left	13:45
	→ bc0906524 joined	13:47
	← intripoon left	13:47
	→ boaz joined	13:48
	→ RobertLaptop joined	13:49
	← opalka left	13:49
	← avar left	13:49
	→ avar joined	13:51
	← parasti left	13:51
	PerlPilot → PerlJam	13:52
	→ kenneth_reitz joined	13:52
	→ jaql joined	13:53
	← roop left	13:54
djMax	what's the preferred web interface these days, gitweb?	13:54
	→ Adman65\|mbp joined	13:55
	← Adman65\|mbp left	13:55
	← scarabx left	13:56
	→ airborn joined	13:57
	→ aless joined	13:57
	→ RORgasm joined	13:58
shruggar	djMax, it depends entirely on what you want to do	13:58
	djMax, for basic repository viewing via a web-interface, gitweb is the standard and is bundled with git	13:59
	← bauruine left	13:59
djMax	I assume that it doesn't actually setup apache? I've got a fresh Ubuntu install with git working, and I installed the gitweb package, whatever that does (given that it's bundled), but still nothing works.	14:01
	→ _Vi joined	14:02
	→ yaotti_aw joined	14:03
	→ hugoxrosa joined	14:05
	→ keystr0k joined	14:06
keystr0k	emergency!!!! hehe.	14:06
	I just realized I was in "no branch" since May 25. I thought I could checkout master and then merge things... but now nobranch is gone.	14:06
	Can someone please tell me how to overcome this?	14:06
	Not sure how I got into "no branch" in the first place... but I sure would like to get back all of those commits.	14:07
	note that the commits DO show up in git-reflog	14:07
	can I just manually merge each one starting from the oldest to the newest?	14:08
	← TheNewAndy left	14:08
	← hohoho left	14:08
	→ hohoho joined	14:08
	← steffkes left	14:08
cbreak_work	first	14:09
keystr0k	OK. Sorry to freak out there... it's not as bad as I thought.	14:09
patrikf_work	keystr0k: find the last commit in reflog, and git checkout -b branchname <sha1 of last commit>	14:09
	← tatsuya left	14:09
cbreak_work	type git branch newbranchmane	14:09
keystr0k	wait. who should I follow? hehe.	14:09
	I am just missing 2 or 3 commits.	14:09
cbreak_work	the idea is to make a branch of current head	14:10
	git branch newname does that. you can also find that in the reflog	14:10
	after you have it in a branch, you can merge that branch normally.	14:10
	← avar left	14:10
cbreak_work	did you ever check out an SHA or a Tag?	14:11
keystr0k	cbreak_work: okay...	14:11
cbreak_work	that would move you away from any branch	14:11
	since commits and tags are not on any branch	14:11
keystr0k	cbreak_work: that's probably exactly what I did.	14:11
	:(	14:11
	gotcha.	14:11
	→ avar joined	14:12
	← rchavik left	14:12
keystr0k	now... because I have to deal with 4 (I just counted) commits that aren't in master, do I have to merge each one individually? Sorry... trying to understand what you're saying	14:12
	→ t0rc joined	14:12
keystr0k	AND I am on a blog post about switching to master and then merging the commits	14:12
	← hohoho left	14:13
IslandUsurper	keystr0k, do what patrikf_work said. It's the quickest way to get back to where you need to be	14:13
	→ hohoho joined	14:13
	→ raichoo joined	14:14
IslandUsurper	it creates a new branch on that last commit, so you can merge it in with master, or keep working on it, or whatever	14:14
keystr0k	IslandUsurper: thanks...	14:14
	patrikf_work: okay. that worked perfectly... so now I am in a new branch (not master). I am seeing all of my commits... so this is a good thing.	14:14
cbreak_work	you can't merge commits	14:14
	you can only merge branches	14:14
keystr0k	cbreak_work: I'm not realizing this...	14:14
cbreak_work	so just do what I told you and make a new branch	14:14
	and merge it :)	14:15
keystr0k	:)	14:15
	Okay.	14:15
djMax	newb question sorry: gitweb needs access to my repos, the docs say to "add daemon to permissions" - what does this mean? chgrp?	14:15
keystr0k	cbreak_work: thanks for the info! same to patrikf_work too... I appreciate it.	14:15
	keystr0k breathes a sigh of relief	14:15
cbreak_work	but you only did 4 commits since may?	14:15
	sounds like a very boring job youhave	14:15
keystr0k	cbreak_work: yeah... HAHa.	14:16
	absolutely	14:16
	It's for a project that barely gets touched.	14:16
	;)	14:16
	→ ewagner joined	14:16
	→ kumbayo joined	14:16
keystr0k	Man... you guys saved the day. Sorry to spam the #git channel earlier. Do you guys get many of these "panic" situations from newbies like me?	14:16
	→ jtauber joined	14:17
	← Ambiguity left	14:19
	← boaz left	14:20
	→ gurjeet joined	14:20
shruggar	with git, almost all panic situations are from newbies, as non-newbies know that git doesn't lose stuff very easily ;)	14:20
	→ nighthwk1 joined	14:21
	ptor only experience data loss when the filesystem on the disk crashed.. but as the repos were all cloned to 4 other systems, no worries.. :-)	14:22
ptor	s/experience/experienced/ - it only happened once	14:22
	→ fmardini joined	14:22
keystr0k	yeah... I am slowly realizing that.	14:23
	→ consolers joined	14:23
keystr0k	ptor: nice :)	14:23
ptor	I usually clone to many other machines due to portability checks all the time (e.g. Linux, 32-bits, 64-bits, Irix, aix etc.. the side-effect of backup is a bonus!)	14:24
	→ kold joined	14:24
ptor	(clone once, pull from then on, of course)	14:24
	And even with a single machine you can just clone to somewhere else on the same machine, in case you tend to do unfortunate rm -rf .git by accident :-)	14:25
keystr0k	yeah... that's pretty solid! I usually have 2 or 3 clones (with recent pulls) at any given time for any repo... makes me feel better.	14:25
	haha	14:25
	→ tedoc2000 joined	14:26
keystr0k	cbreak_work: so, if I ever need to checkout a specific commit, I should always do it into a new branch, correct?	14:26
	I believe in the past I've just checked it out :(	14:26
	ptor pulls to the laptop before leaving the office for the day	14:26
ewagner	When you say clone are you referring to git clone? So far i am just using rsync -a ... Is there any problem with that?	14:26
ptor	ewagner: I talk about git clone, but obviously it's also useful to do rsync if backups are your primary pp	14:27
	purpose	14:27
	← dr_win left	14:27
IslandUsurper	keystr0k, it depends on why you checked that commit out	14:27
	if you just want to see the files, then you don't need a branch	14:28
keystr0k	IslandUsurper: yeah... I cannot remember now...	14:28
	→ ehabkost joined	14:28
keystr0k	I see.	14:28
IslandUsurper	if you think you want to develop on top of it, then you do	14:28
keystr0k	right.	14:28
	gotcha	14:28
IslandUsurper	but branching is cheap, so making one is safer even if you don't use it	14:28
keystr0k	I still don't understand git-revert very well... I may have done checkouts of earlier commits instead of reverting... ? who knows. SO MUCH to learn.	14:28
IslandUsurper	git revert makes a new commit that undoes a previous commit (think patch -R)	14:29
keystr0k	I'm on the man page for git-revert now...	14:29
	I guess I completely misunderstood that one! yipes	14:29
	keystr0k scares himself sometimes.	14:29
IslandUsurper	git reset makes things (index, working tree, and/or HEAD) like they were previously	14:29
	reset/revert confusion is common, because people are used to svn revert	14:30
ewagner	I'm just testing out building git-1.7.1.1 but I received a bunch of failures in make test.	14:30
	/ Is this the right place for that question?	14:30
keystr0k	IslandUsurper: yeah... so it sounds like I should have been using reset... instead of checking out an earlier commit into "no branch" :)	14:30
	→ pdnsmith joined	14:31
ewagner	question being: Is anyone else seeing failures during make test?	14:31
shruggar	ewagner, what tests are failing?	14:33
	← consolers left	14:34
djMax	gitweb isn't showing any repos, not sure what to check.	14:34
ewagner	looks like 19 of 23 t9100-git-svn-basic.sh	14:34
pdnsmith	hi, can anyone tell me why when I run git clone over ssh it might hang. On 'Compressing Objects: 47%' to be specific	14:34
ewagner	Which I guess will not bother me to much for the time being.	14:35
pdnsmith	when I clone from the smae machine it works fine	14:35
shruggar	ewagner, your svn perl library may be a different version from your svn client	14:35
ewagner	hmm.	14:35
	← EricInBNE left	14:36
	→ dpino joined	14:37
ewagner	How do I verify the versions in perl and the used by svn?	14:37
pdnsmith	I've tried cloning from two different machines on different networks and it still hangs	14:38
	I need to clone this repo, how can I fix this?	14:38
	→ vbabiy joined	14:38
kold	I have a git repo with a few submodules being tracked. I'd like to keep the file contents of the submodules in place, but remove it from being tracked as a submodule. Is there anyway to make a submodule be just like another git repo inside another git repo without having to delete and reclone?	14:39
shruggar	ewagner, not sure. You could ask your package-manager (dpkg -l \| grep svn may be relevant)	14:40
	← airborn left	14:40
	→ stringoO joined	14:40
djMax	hrmph, well i've added every acl I can think of but gitweb still empty	14:41
Ilari	pdnsmith: From same machine using ssh?	14:41
	← vbabiy left	14:42
	→ Yuffster joined	14:43
pdnsmith	yep, I tried cloning from a machine on my home network, and from a remote machine accessed over ssh	14:43
	→ boaz joined	14:43
pdnsmith	So I guess the problem is with the machine that holds the repo?	14:44
	→ ajpiano joined	14:44
pdnsmith	ssh is on a non standard port but I included in the url and connected ok	14:45
	→ adamm joined	14:45
pdnsmith	I am a total noob at git so even the obvious may be of some help	14:46
grumpytoad	you've run out of disk space ? :-P	14:47
kokx	pdnsmith: did you set the non-standard port in your SSH config ?	14:47
pasky	wow, first time in very long time, the repo.or.cz machine is idle and update cycle is <10 minutes long	14:47
jast	kold: git rm --cached <path to submodule> in the top-level repository, and clean up .git/config and .gitmodules if you want to get rid of all traces	14:47
ewagner	Shruggar, Thanks! It seems I must have done a manual install of svn a while back into /usr/local and that is where git is picking up svn. I just removed that version and /binsvn(managed my package manager I presume) seems to be working fine.	14:48
	*/usr/bin/svn	14:48
	→ rayners joined	14:48
shruggar	thar ya go. Also note that the svn tests are often annoyingly slow, so unless you're actually using them, you can define something like NOSVN=1 before running the tests (I don't remember the exact variable)	14:48
ewagner	Thank you! but its not a rush at this time. Have a great one and I'll be sure to hang out here more often.	14:50
	← doener left	14:51
pdnsmith	kokx: yes, and I specified it in the url - git clone ssh://me@myhost.co.uk:10418/var/www/html/webroot/	14:53
	→ khine_ joined	14:53
khine_	hello	14:53
kokx	pdnsmith: you can't specify it in the URL afaik	14:53
	at least, not with git	14:53
khine_	when i do a git pull, i get this error: error: Untracked working tree file ...	14:53
	→ chrislerum joined	14:54
khine_	how do i clean this so that i can fetch the code	14:54
	← coderdad left	14:54
shruggar	khine_, do you want to get rid of your local file, losing data?	14:54
khine_	yes	14:55
	→ Zephyrus joined	14:55
pdnsmith	just tried it without the port in the url and still the same 47% :X	14:55
shruggar	then just remove the file it mentions	14:55
	→ beatak joined	14:55
khine_	but when i do a git status, it does not show that this file exists	14:56
	→ anathematic joined	14:56
pdnsmith	something to note is that the git process just runs on the remote machine consuming a lot of resources until I kill it manually	14:57
	...after trying to clone	14:57
shruggar	khine_, it's probably in the ignore file. Just type "rm thefile"	14:57
	→ j0shua joined	14:57
kold	jast: right. what I'd like to do is keep the contents of the <path to submodule> but without the context of git recognizing it as a "submodule"	14:58
khine_	yes you're right, thanks	14:58
jast	kold: and that's what my suggestions were for :)	14:58
nevyn	hrm	14:59
kold	jast: so, running "git rm" won't delete the files, but will just remove the submodule context?	14:59
jast	kold: no... but if you use git rm --cached, as suggested, it will never delete actual files :)	15:00
	← posciak left	15:00
	← pdnsmith left	15:01
kold	jast: yeah, I meant to add the --cached as you suggested...just clarifying though, thanks	15:01
jast	anyhow, that removes only part of the association; a subsequent "git submodule update" will probably try to bring the submodule back (but it won't work because the directory already exists)	15:01
	← HG` left	15:01
jast	so you might also want to remove the entries from .gitmodules and .git/config	15:01
	be careful to avoid actuallyu	15:02
	adding files from the now-not-submodule-anymore to the top-level project by accident	15:02
	→ gurjeet_ joined	15:03
	→ pdnsmith joined	15:03
	→ coderdad joined	15:03
	→ tar- joined	15:03
	→ doener joined	15:03
	offby1` → offby1	15:04
	← offby1 left	15:04
	→ offby1 joined	15:04
	→ gurjeet__ joined	15:04
	← gurjeet left	15:05
	← ewagner left	15:05
	gurjeet__ → gurjeet	15:05
kold	jast: gotcha, removed those entries as well	15:06
	← alezandro left	15:06
	← anathematic left	15:07
jast	great, now the only thing left to do is the formal submodule banishment dance, to prevent it from actually reincarnating	15:07
	put on your banishment hat and follow the rhythm: 1-2-3-5-7-4, 1-2-3-8-3-6, etc.	15:07
	← gurjeet_ left	15:07
jast	why do I keep writing "actually" instead of "accidentally"?	15:08
offby1	because you want to be English?	15:11
	→ Tac joined	15:13
	← raichoo left	15:13
	→ oriba joined	15:14
	← masterkorp left	15:14
	← fermion left	15:14
	← xiong left	15:18
	srcerer_ → srcerer	15:18
	← tcurdt left	15:19
	→ DrNick joined	15:19
	→ raichoo joined	15:20
	← jensn left	15:20
	← happypolla left	15:20
	← smorg left	15:20
	→ hobodave joined	15:22
pdnsmith	so, is this url even formed right?	15:23
	← eevar_ left	15:23
	→ dimsuz joined	15:23
pdnsmith	I see in various tuts a reponame.git at the end	15:23
	← simplechat left	15:23
pdnsmith	I just have a .git dir	15:24
	?	15:24
	← oriba left	15:24
kokx	pdnsmith: thats with bare repositories	15:24
	→ Buglouse joined	15:24
kokx	the convention is to name bare repositories <repo>.git	15:24
dimsuz	hey guys! is there any quick way to ask git to rebase all my local branches to remote ones after git fetch? or should i do a simple script? :)	15:24
	→ bcardarella joined	15:25
	→ parasti joined	15:25
	→ alezandro joined	15:25
pdnsmith	ok, so its correct to be cloning from a /path/to/repo/.git?	15:25
kokx	no pdnsmith	15:26
pdnsmith	gah	15:26
	sorry	15:26
kokx	pdnsmith: you're cloning from a non-bare repository ?	15:26
pdnsmith	by non bare I assume you mean - has a repo with historic commits	15:27
	→ masterkorp joined	15:27
pdnsmith	in which case yes	15:27
	→ fermion joined	15:28
	← fermion left	15:28
dimsuz	is there a way to ask git branch to output only branch names without "*" and indentation?	15:29
	← kold left	15:30
offby1	git for-each-ref \| awk '{something something}' :-)	15:31
	actually ...	15:32
	git branch \| cut -c 3-	15:32
	that should do it	15:32
	→ alexmchale joined	15:33
	→ Cesario joined	15:34
wereHamster	git show-ref	15:36
	→ gbacon joined	15:39
	← giallu left	15:39
	← khmarbaise left	15:39
	→ khmarbaise joined	15:41
	villemv → villev	15:42
	→ kar8nga joined	15:43
	→ programble joined	15:43
	→ SpearBrave joined	15:43
	← SpearBrave left	15:44
	← khmarbaise left	15:44
	← khine_ left	15:47
	← kumbayo left	15:49
	← kar8nga left	15:50
	← rseifert left	15:53
	→ nevyn^_ joined	15:54
	← nevyn^_ left	15:54
	→ ggh joined	15:56
	← nevyn^ left	15:58
	→ khmarbaise joined	15:58
	→ pantsman joined	15:59
	→ kumbayo joined	16:01
	→ mjf joined	16:02
bnovc	if someone added a line to a file and then another person updated it, how can I use git blame to see the original line without checking out to a revision between first?	16:02
	← alezandro left	16:05
	← mastro left	16:07
wereHamster	bnovc: git blame <commit in which the other person updated that file>^	16:07
	← hohoho left	16:07
shruggar	bnovc, git blame <file> to find the revision which you want to ignore, then: git blame <that-commit>^	16:08
	+<file> again	16:08
bnovc	is there any way to just see a list of the modifications to that line with the first blame or something similar	16:08
	it would make tracing back a lot easier i think	16:08
shruggar	bnovc: wait a month or so and there will be	16:08
	(though it seems to be moving right along, maybe even "wait a week")	16:09
lut4rp	I think I remember some people say how you should do a "pull --rebase"	16:09
shruggar	lut4rp, I prefer pull --rebase, yes	16:09
	← corni left	16:09
	→ sukima joined	16:09
lut4rp	shruggar, why?	16:09
	lut4rp doesn't understand it :(	16:09
	← psoo left	16:09
	← patrikf_work left	16:11
kokx	lut4rp: since then you rebase commits instead of merging	16:11
shruggar	lut4rp, in simple terms, rebase hides the history of multiple integrations, while merging preserves that history	16:11
	← lbt_ left	16:12
lut4rp	shruggar, ah yes, I understand that.	16:13
	uhm, anyone have an example by any chance?	16:13
sukima	I think of rebase vs merge: merge applies the changes to the HEAD ()think after the commit) where rebase applies the fix before "this" commit/branch was made (think roll back apply roll forward)	16:13
	→ brett_h joined	16:14
shruggar	lut4rp, man git-rebase has a few graphs that may help	16:14
Gitbot	lut4rp: the git-rebase manpage can be found at http://git.or.cz/man/git-rebase	16:14
	lut4rp sees	16:14
	→ carrus85 joined	16:15
	← cbreak_work left	16:15
	→ paja joined	16:17
	← harinath left	16:17
	→ olivier__ joined	16:17
	← trivol left	16:19
	← kenneth_reitz left	16:21
	→ dreiss joined	16:23
	→ alezandro joined	16:25
	← enoj left	16:26
	→ TeckniX joined	16:26
	IslandUsurper → IslandUsurperAFK	16:29
	← thierryp left	16:30
	→ slonopotamus joined	16:30
	→ harinath joined	16:31
	← shruggar left	16:31
	← [Po]lentino left	16:34
	← phlowx left	16:35
	→ h32Lg joined	16:36
h32Lg	hi	16:36
	i've created a local and remote git repository	16:36
	← barismetin left	16:36
	← Adaptee left	16:36
	→ dbatbold joined	16:37
h32Lg	on my remote box there is an user called "git" which has its own ssh key	16:37
	→ alester joined	16:37
h32Lg	i've initialized a bare git remotely but how can i push to origin master with ssh://git@mydomain.com/myrepo.git ?	16:37
	→ enoj joined	16:37
rkreis	as in .ssh/id_rsa?	16:37
h32Lg	yes	16:38
	i've tested ssh - its working with current user	16:38
rkreis	the remote git user doesn't need that, he needs ~/.ssh/authorized_keys	16:38
h32Lg	ahh sorry - i meant this :)	16:38
rkreis	you probably need an absolute path in that url	16:38
h32Lg	to users home directory?	16:38
rkreis	yes, or wherever the repo is	16:38
h32Lg	i've tried: git remote add origin ssh://git@domain.com/~/git/myrepo.git	16:39
rkreis	alternatively, use ~/.ssh/config on the local box so you can just say mygitserver:myrepo.git	16:39
	i have no idea if that should work	16:39
	maybe /~git/myrepo.git or /~/myrepo.git	16:39
h32Lg	ok	16:39
	i'll give it a try :)	16:40
	→ karptonite joined	16:40
	← dj_tjerk\| left	16:40
h32Lg	mh, the same error - the remote end hung up unexpectedly	16:41
	maybe i've configured my repo incorrect?	16:41
	← dimsuz left	16:41
rkreis	what else does it say?	16:41
	← coderdad left	16:41
	→ coderdad joined	16:41
rkreis	you can put the full output on http://gist.github.com/	16:41
h32Lg	fatal: '~git/myrepo.lgit': unable to chdir or not a git archive	16:42
rkreis	lgit?	16:42
h32Lg	ahh now its another error :D	16:42
	no ubuntu terminal	16:42
	← sixteneighty\|awa left	16:42
	→ thiago joined	16:42
	← carrus85 left	16:43
rkreis	uh, you can select text and paste it with the middle mouse button almost anywhere - except on windows, of course	16:43
	← avar left	16:44
	→ coderdad_ joined	16:44
	→ sixteneighty joined	16:44
	← mjf left	16:45
	→ rubydiamond joined	16:45
	← berserker left	16:45
	← jzacsh left	16:45
	→ avar joined	16:46
	← coderdad_ left	16:46
	← coderdad left	16:46
	→ berserker joined	16:47
	← nessundorma left	16:47
h32Lg	ahh rkreis now it seems that its working :9	16:48
	i've to create a file first, add it to my local repo and make an initial commit	16:48
	after that i can push to the origin master :)	16:48
rkreis	i see :)	16:48
	what could git possibly push if there's not a single commit? :D	16:49
h32Lg	sure - a general pebkac problem :)	16:49
	← avar left	16:49
h32Lg	this is my first own git server	16:49
	← disappearedng_ left	16:50
	→ avar joined	16:50
rkreis	i have git repos scattered everywhere :D	16:50
h32Lg	i've used github a few months before and i like git :)	16:50
	hehe	16:50
rkreis	you don't really need a server	16:50
h32Lg	you're from germany? what projects are you working on with git?	16:50
rkreis	i am, i use git-svn at work because i wouldn't want to use svn directly and because i got used to git	16:50
	i also use git to manage stuff for studies like exercises and papers where others work with me	16:51
	and i use it locally whenever i create a source code file :)	16:51
h32Lg	ahh i'm going to use it with my new content management system :)	16:51
rkreis	i see :)	16:52
h32Lg	sure? where? :P	16:52
rkreis	behind you :D	16:52
h32Lg	:P	16:52
	damn	16:52
	→ patmaddox joined	16:53
	← tar- left	16:56
	→ tar- joined	16:56
	→ patmaddo_ joined	16:56
	← slonopotamus left	16:58
	→ AkumaStreak joined	16:58
	← patmaddox left	16:59
	→ lacostej joined	16:59
lacostej	has anyone seen a problem like this one before ? http://pastebin.ca/1893485 (.git/svn/git-svn/index.lock': Input/output error)	17:00
	→ eaven joined	17:00
	→ thana joined	17:00
rkreis	df -h? :D	17:01
thana	thanks git-svn to forget my complete work of today	17:01
rkreis	oh, that's not it	17:01
eaven	dontcha love that thana ;)	17:01
	← avar left	17:01
	→ avar joined	17:01
rkreis	lacostej, does dmesg or so say anything? doesn't sound like a git issue	17:01
thana	can you explain me why it forgot everythink because it got interrupted badly during a dcomit?	17:02
	→ disappearedng joined	17:02
lacostej	I got some ext3-fs issues, related to a crash this morning I guess (no more battery). Did I lose my git repository ?	17:03
rkreis	i, uh, no idea	17:04
	i guess the index isn't that important	17:04
	→ Joeconyers joined	17:04
rkreis	you should probably fsck asap	17:04
lacostej	yep...	17:05
	rebooting :)	17:05
	thanks	17:05
	← lacostej left	17:05
	← malumalu left	17:05
	→ malumalu joined	17:06
	→ zimbres joined	17:06
	→ kenichi joined	17:08
	→ hot1boy18 joined	17:08
	→ hasrb joined	17:10
	← hot1boy18 left	17:10
	IslandUsurperAFK → IslandUsurper	17:11
	← DeaD_SouL left	17:12
	← fmardini left	17:12
	← khmarbaise left	17:12
	→ adymitruk joined	17:12
	← karptonite left	17:12
	← herdingcat left	17:13
adymitruk	What are the pros and cons over managing project/repo dependencies via submodule vs subtree?	17:13
	→ khmarbaise joined	17:13
	→ elmob joined	17:13
	← sixteneighty left	17:18
	← hasrb left	17:18
	→ aresnick joined	17:19
	← christophsturm left	17:20
	→ christophsturm joined	17:20
	→ hasrb joined	17:21
	← diegovio1a left	17:22
	← thrope left	17:22
	← gurjeet left	17:23
	← ringo999 left	17:24
	← stringoO left	17:25
	→ krawek joined	17:25
	← christophsturm left	17:25
	← _Vi left	17:26
	→ niklasfi joined	17:26
	→ _Vi joined	17:27
niklasfi	hi ... i am trying to configure .gitignore but it does not do what i want it to do i want to ignore all fies in the www directory exept those with certain extensions:	17:27
	www	17:27
	!*.html	17:27
	!*.js	17:27
	why does git not add an html file in the www dir?	17:27
	← disappearedng left	17:28
	→ zenazn joined	17:29
	→ spearce joined	17:29
	→ sh1mmer joined	17:29
	← hyperair left	17:30
doener	niklasfi: use "www/*" instead of "www"	17:31
	→ tizbac joined	17:31
niklasfi	well it does ignore the files in the www dir	17:31
	→ jasonn1234 joined	17:31
doener	niklasfi: you ignore the whole directory, so git never enters it, and thus won't find any *.html files that could match the "unignore" rule	17:32
adymitruk	I didn't know about !	17:32
	how does it work?	17:32
	← hasrb left	17:32
niklasfi	ok so if i have the unignored files in subdirs of the ignored dir then it wont find them too?	17:32
	← jasonn1234 left	17:33
h32Lg	mh, how can i create a link like: git://domain.com/project.git so that i can use it to clone "project.git"? currently i'm only able to use my ssh link	17:33
doener	niklasfi: sure. And even www/* won't help with that (because that still matches the subdirectories, which get ignored as a whole)	17:33
niklasfi	:( is there any way to avoid this behavior?	17:34
	→ hyperair joined	17:34
doener	niklasfi: the easiest setup for you should be a .gitignore file inside the www directory with 4 entries: * !/ !.html !*.js	17:34
	niklasfi: that is: ignore everything, except for directories, .html and .js files	17:35
niklasfi	ok that makes sense	17:35
	→ p3rg2z joined	17:35
	→ phantomcircuit joined	17:36
	→ frakturfreak joined	17:37
	← sh1mmer left	17:37
	→ hakunin joined	17:38
	Unode → _Unode	17:39
	→ ChUrP joined	17:39
	→ kar8nga joined	17:40
	→ plediii joined	17:40
	→ karptonite joined	17:40
adymitruk	wow.. I didn't know you could have multiple .gitignore file in your repository	17:41
	→ lispy joined	17:41
	← aresnick left	17:41
	← tvw left	17:42
	→ sh1mmer joined	17:42
	→ Avamons joined	17:42
Avamons	`Another `aspie moment` brought `to you by LoRez: girl, you must be in the kkk because i think we'll have a lot of fun once we get under the sheets	17:42
	← curtana left	17:43
	← Avamons left	17:43
	← bluenovember left	17:44
	← patmaddo_ left	17:44
	→ higcmonkect joined	17:44
higcmonkect	`Another `aspie moment` brought `to you by LoRez: HEY BABE WHO WOULD WIN PIRATES.... OR NINJA??? ALSO NOTE I DID NOT USE AN S BECAUSE NINJA IS JAPANESE AND THEY DONT HAVE PLURALS DO YOU LIKE JAPAN	17:44
	→ mangouterQuinn joined	17:44
	← ereslibre left	17:45
mangouterQuinn	`Another `aspie moment` brought `to you by LoRez: Fuck all authority figures of any type. Now instead of hearing the song and letting it pass by you I think we should all rise against and do somin about authority..... and they can feel the rage of OUR generation!!!!!!!!!!!!!	17:45
	← mangouterQuinn left	17:46
	← higcmonkect left	17:46
	← vegeste left	17:46
	→ juanjoc joined	17:48
	← karptonite left	17:48
	← Zephyrus left	17:48
adymitruk	how do you enable the prompt to include the branch name? Msysgit on windows does this out of the box. On linux it has to be enabled somehow. Is it in the contrib folder. Not sure I want to pull down source just to do that	17:48
	→ stringoO joined	17:49
	→ karptonite joined	17:50
	← rtl left	17:50
	← RORgasm left	17:51
fr0sty	adymitruk: the whole tarball is <3MB. what's the big deal?	17:52
	← kumbayo left	17:54
	← tildeequals left	17:54
	→ patmaddox joined	17:55
fr0sty	unzip, steal contrib/completions/git-completion.bash and move it somewhere, modify .bashrc and you are ready to go.	17:55
	← niklasfi left	17:55
	← sylr left	17:56
	← candybar left	17:56
wereHamster	h32Lg: you have to start git-daemon	17:57
h32Lg	ahh ok	17:57
	and the deamon will detect such requests automatically?	17:57
wereHamster	the daemon listens on the git:// port and serves the requests	17:58
	→ hasrb joined	17:58
h32Lg	ahh ok - thank you :)	17:58
wereHamster	be sure to start it with the crrect options (mainly, --base-path)	17:58
	h32Lg: and make sure to export the repos you want to serve. It should be explained in man git-daemon or on google.com	17:59
Gitbot	h32Lg: the git-daemon manpage can be found at http://git.or.cz/man/git-daemon	17:59
	→ candybar joined	17:59
	→ jensn joined	18:00
	← jensn left	18:01
	→ jensn joined	18:01
	← harinath left	18:08
	→ enherit joined	18:09
	→ magicbronson joined	18:10
	← stringoO left	18:12
	← j0shua left	18:13
	← enherit left	18:13
	← rox` left	18:13
	→ enherit joined	18:13
	→ drizzd joined	18:14
	→ ereslibre joined	18:14
	← ereslibre left	18:14
	→ ereslibre joined	18:14
	→ stringoO joined	18:16
	→ nyxtom joined	18:17
	← berserker left	18:19
	→ aresnick joined	18:19
	← apsysoev left	18:20
	→ berserker joined	18:21
	→ rox` joined	18:21
	→ Rh0nda joined	18:22
	← Rh0nda left	18:22
	IslandUsurper is amused by the spammer who tells everyone to hate people who tell others what to do.	18:22
	→ harinath joined	18:23
	→ Titosemi_ joined	18:25
	→ benbinary joined	18:27
	← Titosemi left	18:27
	Titosemi_ → Titosemi	18:27
	→ coppro joined	18:28
	← sukima left	18:28
	← thana left	18:28
	→ disappearedng joined	18:30
	→ RORgasm joined	18:30
h32Lg	it initialize an empty git repository but break while trying to clone with following message: errno=Connectoin refused - fatal: unable to connect a socket (Connection refused)	18:31
	i've tried to start the daemon with: git-daemon --inetd --verbose --base-path=/home/git/projects --strict-paths	18:31
	→ Cminion joined	18:32
	← hasrb left	18:34
	← adripop left	18:35
drizzd	h32Lg: are you using inetd?	18:35
	← Cminion left	18:36
	→ frakturfreak_ joined	18:36
h32Lg	uhm, no :D	18:36
	but this seems not the only problem because "ps x" says sth about /var/cache as base path	18:36
	← frakturfreak left	18:37
h32Lg	ha	18:37
	removing inted works :)	18:37
	→ Cminion joined	18:39
	← chrislerum left	18:40
	← _Vi left	18:40
kokx	h32Lg: you know that the git protocol doesn't have authentication, right ?	18:40
h32Lg	why are you asking kokx?	18:41
	→ _Vi joined	18:41
kokx	h32Lg: because i'm not sure if you want to publish all your git repositories	18:41
	if you use the git protocol, everybody can read your repos	18:42
h32Lg	no - only these where i set the specific file	18:42
	what would you suggest to make it possible for others to clone a repo?	18:42
kokx	the git protocol if its public	18:43
	→ Bass10 joined	18:43
kokx	ssh if its not public	18:43
	also h32Lg, are you using gitosis or gitolite or are you setting it up all manually ?	18:43
	← elmob left	18:43
h32Lg	manually :)	18:43
	have a look at the notice i've send to you	18:44
	→ ChUrP1 joined	18:44
kokx	h32Lg: i recommend using gitosis or gitolite, they are very nice tools which can help you to manage your repositories	18:44
	→ eighty4 joined	18:45
gebi	and git protocol possible over http otherwise some people will have problems with their firewall admins	18:45
	→ chrisler_ joined	18:45
	← Bass10 left	18:45
kokx	true gebi	18:45
	→ sylr joined	18:45
	← gbacon left	18:46
	→ Bass10 joined	18:46
	← ChUrP left	18:46
	→ samrose joined	18:47
samrose	trying to clone a repo remotely and keep getting "warning: You appear to have cloned an empty repository." my clone command looks like sudo git clone ssh://myuser@exmple.com/var/www/myrepo	18:48
h32Lg	i guess there are no files in your repo you are trying to clone from?	18:48
samrose	there is definitely a repo at ssh://myuser@exmple.com/var/www/myrepo	18:48
h32Lg	but no files in it?	18:49
kokx	samrose: does your user have access to the repo?	18:49
samrose	there are 4000+ files	18:49
	yes user has access	18:49
drizzd	kokx: I'm told gitosis is deprecated in favor of gitolite	18:49
kokx	drizzd: well, gitosis still works, though yes gitolite is a lot better	18:49
samrose	do I need sudo git clone ssh://myuser@exmple.com/var/www/myrepo.git ?	18:50
kokx	and afaik, gitosis isn't under (active) development	18:50
	no samrose	18:50
samrose	tried it and not working	18:50
kokx	i	18:50
	* i mean access on the server	18:50
	samrose: ssh to the server, and execute a clone as 'myuser'	18:50
samrose	kokx which user? the git user?	18:50
	kokx it works ssh into server and cloning	18:51
	→ tildeequals joined	18:51
samrose	I see what you mean now about user kokx	18:51
kokx	ok samrose	18:51
	samrose: you're not using gitosis or gitolite i guess ?	18:51
samrose	kokx no I am using command line git	18:51
kokx	on the server i mean samrose	18:52
samrose	no kokx not using those	18:52
kokx	btw, why is your repo in /var/www ?	18:52
	← gebi left	18:52
h32Lg	mh	18:52
	i'm to lazy to read about gitosis, for now ..	18:53
samrose	kokx this is just a test run of some files	18:53
kokx	ah ok	18:53
	← tro left	18:53
	→ pgebhard joined	18:54
	→ gurjeet joined	18:54
drizzd	samrose: what does git ls-remote <url> tell you about the repo?	18:55
	→ benbinary_ joined	18:56
samrose	drizzd fatal: 'myrepo' does not appear to be a git repository fatal: The remote end hung up unexpectedly	18:56
	→ hohoho joined	18:56
drizzd	I thought it was myrepo.git?	18:56
kokx	samrose: is the repository on your server a bare repository ?	18:57
samrose	I was asking if I should use myrepo.git	18:57
drizzd	samrose: you should use the correct path	18:57
samrose	it is not a bare repository	18:57
	← benbinary left	18:57
	benbinary_ → benbinary	18:57
	→ swift-prime joined	18:57
samrose	I am trying to figure out what the correct path is	18:57
kokx	samrose: that shouldn't be too hard	18:57
	← magicbronson left	18:58
kokx	if its indeed in /var/www	18:58
samrose	does git need "myrepo.git" or can it use "myrepo"?	18:58
kokx	then you ssh to the server, go to /var/www, theck ls	18:58
drizzd	samrose: for all git cares you can call it myrepo.exe	18:58
swift-prime	so I have two files which are closely related, and I'd like to combine them into one file. how can i do this in a way that preserves both of their histories?	18:58
kokx	and then you see if the folder is called myrepo.git or myrepo	18:58
	use it how its called	18:58
samrose	so, I could use /var/www/myrepo and it will work right?	18:58
kokx	yes	18:58
samrose	that is what I am doing, and it is not working	18:59
kokx	samrose: then i still suspect that your user doesn't have the right access	18:59
	avar unsucks git submodule some more : http://paste.scsys.co.uk/46265	18:59
drizzd	but, is a url like that relative to root, or relative to the home dir?	18:59
	→ gbacon joined	18:59
samrose	(actual folder name not "myrepo.git" just "myrepo") drizzd it is an url that is relative to root	18:59
	I can ssh in as root	19:00
	still doesn't work	19:00
drizzd	if ls-remote doesn't work, I wonder why git clone doesn't give the same error	19:00
	samrose: the question is whether ssh interprets it as such	19:00
samrose	drizzd I see	19:00
	I think kokx you are right, it is probably access	19:01
drizzd	samrose: ah, it probably does	19:01
samrose	thanks drizzd and kokx	19:02
kokx	samrose: is git installed on your server ?	19:02
samrose	I will see if I can figure that out	19:02
kokx	and in your PATH ?	19:02
Ilari	Especially puzzling as ls-remote and clone use the same service (git-upload-pack).	19:02
	← pgebhard left	19:03
Ilari	The first difference should be ls-remote sending disconnect request when clone sents WANT lines.	19:03
	*sends	19:03
	→ pgebhard joined	19:03
drizzd	Ilari: exactly	19:03
	Ilari: right, so ls-remote actually does a subset of what clone does	19:04
	← khmarbaise left	19:04
	→ khmarbaise joined	19:05
kokx	yes drizzd	19:06
	samrose: ?	19:06
Ilari	Maybe some things useful for debugging: 1) Option to dump layer 7 communications. 2) Option to make dump of some segment of stream after bad packet header.	19:06
samrose	kokx si?	19:06
drizzd	samrose: are you sure you got the url right when you did that ls-remote?	19:07
samrose	drizzd yes I am certain	19:07
drizzd	because that error message is also what you get if the url is wrong	19:07
	ok	19:07
	← rubydiamond left	19:07
	→ SplinterOfChaos joined	19:07
samrose	git cannot see any repo	19:07
kokx	samrose: did you install git on your server, and is it in the PATH of that user ?	19:07
samrose	I see what you are saying	19:07
	→ jcromartie joined	19:07
samrose	yes it is in PATH of both	19:08
kokx	ok	19:08
jcromartie	I added a repo as a submodule to another. both live on github and both are up to date	19:08
	← benbinary left	19:08
h32Lg	mh	19:08
drizzd	the message "you appear to have cloned an empty repo" comes only after the fetch_refs commands already succeeded	19:08
jcromartie	but when I check out the superproject and to "git submodule init" and "git submodule update", there is nothing in the submodule directory	19:08
drizzd	so git just doesn't seem to find any refs	19:08
h32Lg	i've installed git daemon and its working but there is no init.d file to start, stop and restart the daemon	19:08
	→ portal joined	19:09
drizzd	maybe packed-refs or refs/ is not accessible? but then git would probably complain too...	19:09
portal	how would I undo a "git mv foo bar", where foo is a folder?	19:09
samrose	drizzd understood still checking it all out	19:09
	← ChUrP1 left	19:09
Ilari	drizzd: refs/ not being accessable => repo detaction fails.	19:09
	→ tro joined	19:09
jcromartie	everything is up to date everywhere	19:09
	but the submodule just refuses to check out	19:10
Ilari	drizzd: Or maybe it doesn't...	19:10
	← dpino left	19:10
	← t0rc left	19:10
	← SplinterOfChaos left	19:10
	← Cminion left	19:10
jcromartie	there's a ".git" directory in the submodule's folder, but none of the actual files	19:11
drizzd	ah, if I do chmod -r .git/*; chmod +r .git/HEAD, clone succeeds and I get "You appear to have cloned an empty repository."	19:11
	→ lamalex joined	19:11
lamalex	Is there a way to fork a new repository from a directory, and keep the file history for the files within that directory?	19:11
	→ ChUrP joined	19:11
drizzd	jcromartie: can you "git show submodule" (which should get you the sha1) and then cd into the submodule and git checkout that sha1?	19:12
	→ Cminion joined	19:12
lamalex	specifically, a project bundles a library that I want to get into its own git repo. none exists yet as this project developed the lib. Can I preserve file history on the files involved?	19:12
	or do I just need to cp the dir and start fresh	19:13
Ilari	lamalex: filter-branch?	19:13
drizzd	samrose: I'm guessing your refs/ directory is not readable for myuser	19:14
jcromartie	yeah there they are	19:14
	drizzd: thanks	19:14
samrose	drizzd ok, thanks	19:14
jcromartie	why do I have to do this?	19:14
lamalex	Ilari: looks like that might be it! thanks	19:14
	← samrose left	19:14
	→ ericdc joined	19:14
	← Cminion left	19:14
	→ macek joined	19:15
macek	is it possible to stash a couple of files before doing a git commit? or is it "all or none"?	19:15
drizzd	jcromartie: that's what git submodule update should do	19:15
	jcromartie: is it a recursive submodule?	19:15
portal	seriously, if "foo" is a dir, how do I undo a "git mv foo bar"	19:15
drizzd	i.e., a submodule in a submodule?	19:15
portal	it's called "bar" now, and that's really bad	19:15
	I need i to be called "foo" again	19:15
	oh, wait, maybe a "git mv bar foo"	19:15
	← ericdc left	19:15
macek	portal `git mv bar foo`	19:15
	lol	19:15
	→ ericdc joined	19:15
jcromartie	drizzd: no	19:16
portal	yeah, worked :-p	19:16
jcromartie	drizzd: as far as I know it's not recursive	19:16
drizzd	portal: well, if bar is still there, that will move it to foo/bar	19:16
	jcromartie: what if you do git submodule update again?	19:16
jcromartie	the files are still there	19:16
	it doesn't appear to do anything	19:17
	← zimbres left	19:17
	→ Cminion joined	19:17
portal	drizzd: the reverse-mv worked	19:17
drizzd	jcromartie: even if you change the checkout?	19:17
	→ zimbres joined	19:17
	← xenoterracide left	19:17
	← lamalex left	19:17
drizzd	portal: ok, would only have been a problem if untracked files were left in foo/	19:17
	I think git doesn't move untracked files	19:17
	although I think that's a little strange	19:17
	→ Cisconode joined	19:19
jcromartie	drizzd: I deleted the submodule's directory and did init/update again, and it worked... looking back, there was a problem when I entered the wrong passphrase for my private key for some second operation (the checkout)	19:19
drizzd	jcromartie: ah I see	19:19
	→ blueyed joined	19:19
	→ outerWeb_mons joined	19:20
blueyed	How do I use different defaults for "push" and "pull"? (it appears to use branch.X.remote for both.	19:21
outerWeb_mons	`Another `aspie moment` brought `to you by LoRez: I know Africa has its own internet connection.. buncha niggers with mailbags run around for email.. Tech Support is like shakalakuzulu kuntakente cheifs that speek jigabooese	19:21
	← outerWeb_mons left	19:21
	← Cisconode left	19:21
	→ sunsRubb joined	19:21
portal	drizzd: yeah, I know what you mean about the left-over-untrakced-files thing, fortunately that wasn't the case :)	19:21
	← sunsRubb left	19:21
	← Cminion left	19:22
	→ rseifert joined	19:23
	→ KilledVerNoodle joined	19:23
	→ Iqjqla_huun joined	19:24
	→ str1ngs joined	19:25
	← Sigma left	19:26
	← rseifert left	19:26
	→ amgNixmang joined	19:26
KilledVerNoodle	`Another `aspie moment` brought `to you by LoRez: Punctuation marks that look like smiley faces express happiness on a new communication tool known as "Internet." The computer network is already being used by 15 million people worldwide to chat or do important research.	19:27
	← KilledVerNoodle left	19:27
	→ abnetouter joined	19:28
	→ mangCoo joined	19:28
abnetouter	`Another `aspie moment` brought `to you by LoRez: Nigerians make me so mad I wish someone would round them all up and send them to a country where people starve and get gunned down in the street	19:28
	← abnetouter left	19:28
	← mangCoo left	19:28
	← Iqjqla_huun left	19:28
	ChanServ set mode: +o	19:29
	→ Lazmwanesty joined	19:29
	← Lazmwanesty left	19:29
	ChanServ set mode: -o	19:29
	← amgNixmang left	19:30
	← chrisler_ left	19:31
	→ chrislerum joined	19:31
str1ngs	I have a svn server that I mirror with with git svn into a svn branch. I then rebase my changes. however everytime I goto push my rebase branch to another git repo I get non-fast-forward. even though the remote branch has not been changed.	19:32
	← eighty4 left	19:33
	frakturfreak_ → frakturfreak	19:34
blueyed	remote.X.pushurl is what I was looking for.	19:35
	← disappearedng left	19:36
	→ djb_ joined	19:36
	→ corni joined	19:37
	→ SplinterOfChaos joined	19:42
	← stringoO left	19:43
	← blueyed left	19:43
	← masterkorp left	19:44
	← djb_ left	19:51
	← pavlix left	19:54
	→ pavlix joined	19:55
	→ disappearedng joined	19:56
	→ eighty4 joined	19:57
	← mabrand left	19:58
	→ grumpytoad joined	19:58
	← kenneth_ left	19:59
	← leavittx_ left	19:59
	← leavittx left	19:59
	→ Crell joined	20:00
	→ dr_win joined	20:01
	→ leavittx joined	20:01
	← pgebhard left	20:01
	← h32Lg left	20:04
	← _Vi left	20:04
	← RORgasm left	20:05
	→ RORgasm joined	20:05
	← keystr0k left	20:05
	→ apsysoev joined	20:07
	→ scarabx joined	20:07
	→ andrewl joined	20:10
	→ Bass2 joined	20:11
	→ Fullmoon joined	20:14
	← Bass10 left	20:14
	→ Bass10 joined	20:15
	← Joeconyers left	20:15
	← Bass2 left	20:15
	→ shruggar joined	20:16
	→ pflanze joined	20:17
	→ leavittx_ joined	20:18
pflanze	How do I check whether the working directory is clean, for scripts (meaning, a quiet tool that reports via exit status is preferred)?	20:18
	→ Oxf13 joined	20:18
Oxf13	sitaram: around by any chance? Hoping to pick your brain a bit.	20:18
shruggar	pflanze: I don't think there's an all-in-one way, but: git diff-files --quiet && git diff-index --quiet && git diff-index --quiet --cached	20:19
	→ stringo0 joined	20:20
pflanze	git diff-index --quiet complains for me, doesn't seem to know the option	20:20
shruggar	hm, might have been -q?	20:20
pflanze	nope, neither; git version 1.7.0.3	20:21
	ah it's not the --quiet,	20:21
	→ dkulchenko joined	20:21
pflanze	but it doesn't like being called without any other argument.	20:21
Oxf13	Does anybody here know what the SSH_ORIGINAL_COMMAND would be when doing a push action?	20:22
	← alezandro left	20:22
pflanze	so should I give it HEAD ?	20:22
dkulchenko	Hi all! There's this service online that I use regularly. You can deploy scripts to it with git push, and it'll run commands on the remote server and send back the output as part of the 'git push' output. How do they do that?	20:23
	← villev left	20:23
	→ comawhite joined	20:23
shruggar	pflanze, yeah, it needs HEAD, sorry	20:24
pflanze	Ok, works, thanks!	20:24
shruggar	dkulchenko, with a hook. I don't remember which one, but I think it's post-receive	20:25
dkulchenko	shruggar: can that hook send back output?	20:25
shruggar	dkulchenko, I know a hook exists which can, but I'm not sure that it was post-receive	20:25
dkulchenko	shruggar: ah okay, i'll read the docs. would I be using git-daemon?	20:26
	← bc0906524 left	20:26
	→ RickRaven joined	20:26
shruggar	dkulchenko, hooks don't require git-daemon one way or the oher	20:26
dkulchenko	shruggar: right, but for this specific case, would I use git-daemon? (if I want git over SSH)	20:27
shruggar	dkulchenko, I wouldn	20:27
	t	20:27
dkulchenko	what's the alternative?	20:27
shruggar	also, I can't type today	20:27
	dkulchenko, ssh by itself	20:27
dkulchenko	hmm okay	20:28
shruggar	dkulchenko, if you have shell access and git is installed, you can just tell git the hostname:path-to-repos	20:28
dkulchenko	shruggar: wait, what handles git:// urls?	20:28
	→ blueyed joined	20:29
shruggar	dkulchenko, I think that's git-daemon	20:29
dkulchenko	shruggar: ah okay	20:29
	← pflanze left	20:29
blueyed	Git does not ignore files being tracked, when they are in .gitignore - is there a way to do so? (makes sense with e.g. build.properties: it should be in the repo, but local changes should get ignored)	20:29
	← cilly left	20:30
shruggar	blueyed: the accepted solution is to track either a template of the file, or a file containing defaults, and add the "real" one to .gitignore	20:31
blueyed	I see, thanks shruggar.	20:31
	→ ClimDan joined	20:32
	← Fullmoon left	20:33
	← ClimDan left	20:33
	→ Mage joined	20:36
Crell	Random silly idea I thought I'd bounce off people.	20:38
	When svn first came out, some more experimental folks tried to use svn for their entire home dir.	20:38
	I don't know how well it worked.	20:38
	Has anyone tried doing that with git?	20:39
shruggar	Crell, it had good parts and bad parts	20:39
	aep uses git for dotfiles	20:39
DrNick	Crell: git doesn't store permissions	20:39
	or xattrs	20:39
	or fs timestamps	20:40
	→ bcoca joined	20:40
Crell	Ah, that could be a problem.	20:40
	But svn does?	20:40
shruggar	Crell, with git, it's harder to take a subset of what you're tracking than it is with svn. This makes it less-useful for something as dynamic as a home-dir, where you might want to track "everything" and then clone just your "docs" directory	20:41
	← jaql left	20:42
Crell	So it could make sense to make ~/Documents a git repo, for instance.	20:42
shruggar	Crell, currently, I track only specific dotfiles, each in a separate repository, and those only where it makes sense to divide things up into multiple files (for example, I use a .bashrc which includes .bashrc.d)	20:42
	← scarabx left	20:42
shruggar	Crell, it could, yes, but it may make more sense to use a versioning filesystem	20:43
	Crell probably won't do this; it was just a crazy idea.	20:43
	→ aspotashev\|eeepc joined	20:44
Crell	A versioning filesystem wouldn't allow remote cloning, would it?	20:44
	← reaVer left	20:44
shruggar	Crell, I would definitely recommend using git to track absolutely every config file you ever edit, but I'd recommend doing it with thought and patience, rather than the sledgehammer of "I put a bunch of unrelated config files here, so I'll track all of this"	20:44
	→ reaVer joined	20:45
	Crell nods.	20:45
	← IslandUsurper left	20:45
shruggar	for example, whenever I edit something in /etc/, I turn the file into a symlink and point it to a git repos, which I often wind up branching and using on multiple systems	20:45
	(though these "multiple systems" are often "the production server and several dev machines meant to mimic it"	20:47
Crell	So you have a git repo off somewhere that has, say, profile, bashrc, and apache/apache.conf in it? (Or whatever) And then /etc/bashrc is a symlink off to that?	20:50
	← sequethin left	20:50
shruggar	Crell, no, I mean I have one with /etc/bashrc, one with /etc/postfix, and one with /etc/apache. and each of those only track the files which I have modified.	20:51
	→ coderdad joined	20:51
Crell	Oh, so you have lots of 1-file repos?	20:52
	→ frakturfreak_ joined	20:52
	→ rtl joined	20:52
	← blueyed left	20:53
shruggar	Crell, I have a couple of 1-file repos, but most of the time the files involved have methods for splitting things into multiple files (eg: bashrc and httpd.conf), so there are actually several related files even for things which seem like "single file" repos	20:55
	← frakturfreak left	20:56
shruggar	Crell: think of it as, "whenever you'd want the file for something else, but without the other files, that's where the repo division is"	20:56
	I haven	20:56
	→ carrus85 joined	20:56
shruggar	I haven't experimented with --new-workdir, or whatever that option is. This may have allowed me to use a single repos with several unrelated branches	20:56
Crell	hm.	20:58
	Crell is still a relative git novice.	20:58
	← alexmchale left	20:58
	← dkulchenko left	20:59
shruggar	Crell, as far as I know, the option I am referring to is still experimental	21:00
Crell	Ah	21:01
	← bcardarella left	21:03
	← kar8nga left	21:03
	← unreal left	21:03
	← bulters left	21:03
	→ unreal joined	21:04
	← macek left	21:04
	→ wilhelmtell joined	21:04
	← lucsky left	21:06
	→ ghost_olli joined	21:07
	← pantsman left	21:08
	← carrus85 left	21:09
	→ XDS2010 joined	21:09
XDS2010	is there a way to watch code being pushed in real time ?	21:09
shruggar	XDS2010, in what sense?	21:10
	XDS2010, what do you actually want to do?	21:10
XDS2010	i want to watch commits in real time of a project	21:11
drizzd	ciabot?	21:12
XDS2010	ah!	21:12
	i forgot about cia	21:12
	but is there anything else ? , like a graphics based system ?	21:12
	← lhz left	21:13
tango_	XDS2010: gitweb + rss feed	21:13
	→ EricInBNE joined	21:13
XDS2010	tango_: !	21:13
shruggar	isn't cia just a protocol which can hook into anything? I'd think there would be a graphical client for it	21:13
XDS2010	is the rss realtime tho ?	21:13
	→ johnw joined	21:13
	← ajpiano left	21:14
XDS2010	that doesn't really display code removal	21:14
tango_	XDS2010: so you want a 'server push' news update, not a 'client pull' (polling)	21:14
XDS2010	only code commit	21:14
tango_	code removal is done via commits	21:14
jhelwig	XDS2010: It's as realtime as you read from it.	21:14
jast	polling an RSS feed in real time is not very fun	21:14
tango_	right	21:14
XDS2010	server push right	21:14
	like the hummingbird project does for traffic stats	21:15
tango_	XDS2010: for rbot we have a bot on the same machine as the "upstream" repository that gets notified via a post-receive hook and forcibly polls the rss	21:15
	→ zenazn_ joined	21:15
tango_	XDS2010: I would say that you need something like that: a post-receive hook if you have a 'central repository' structure, or a post-commit hook if you are fully decentralized	21:15
	← zenazn left	21:15
tango_	XDS2010: the hook would notify all registered clients about the update, or something like that	21:16
XDS2010	well i can do shareflow updates	21:16
	← ghost_olli left	21:16
XDS2010	but i want to see code movement in realtime	21:16
	both delete and insert	21:16
tango_	XDS2010: what do you mean by code movement?	21:16
	XDS2010: you want to see changes while people are editing?	21:17
XDS2010	character movement	21:17
	yea	21:17
tango_	XDS2010: this has nothing to do with git then	21:17
	XDS2010: you need to use something like screen or some other termcast application	21:17
XDS2010	why couldn't git do this ?	21:17
MattDiPasquale	shouldn't this apache rewrite rule work?	21:17
	RewriteRule cool(.*) http://dev.example.org/cool$1 [R=301,L]	21:17
XDS2010	push instead of pull	21:18
tango_	XDS2010: that, of course, assuming people use command-line editors	21:18
	XDS2010: git is about revision control	21:18
	XDS2010: not about concurrent editing	21:18
XDS2010	thats maddening	21:19
	← ChUrP left	21:19
tango_	XDS2010: do you have the concept of "revision control" clear?	21:19
XDS2010	ok maybe not maddening	21:20
	I just want to be able to see the actual commits AS they come in	21:20
	let me checkout gitweb	21:20
tango_	XDS2010: commits are not realtime editing	21:21
Crell	XDS2010: What does "see commits as they come in" mean? You want a flash of a patch?	21:21
tango_	XDS2010: if you want commits, you just need a post-commit or a post-receive hook that does the notification	21:21
Crell	Commits come in as a blob, not as a character stream.	21:21
XDS2010	Crell: yes	21:21
	to your last	21:21
Crell	Shown where? Email, twitter, web page?	21:22
XDS2010	patch flash	21:22
	← RORgasm left	21:22
shruggar	might I point out that commits aren't really relevant in git?	21:22
XDS2010	well if i can have the output parsed i could use raindrop	21:22
shruggar	XDS2010, is this to be used as a fancy desktop background?	21:22
	→ mastro joined	21:22
	Crell does not know what raindrop is.	21:22
XDS2010	shruggar: no but thats a good idea	21:22
	Crell: another shareflow project, this one is open source	21:23
shruggar	XDS2010, my point is: you're asking for something meaningless and pointless, so I assume you just want it to look cool, but have real data behind it. If that's what we're talking about, we can keep going with this, so long as it's the stated goal	21:23
Crell	shruggar: Looking cool is never meaningless and pointless. :-)	21:24
	Extraneous, perhaps, but not pointless.	21:24
XDS2010	ok i lost you	21:24
	all i want is realtime	21:24
	push not pull	21:24
Crell	XDS2010: What is your goal? Why are you trying to view something in "realtime"?	21:24
	Is this for monitoring a dev team? Showing off to friends?	21:25
	→ bcardarella joined	21:25
XDS2010	Because the changes to some projects are collaborative and 100's of commits are done a hour	21:25
	shruggar is reminded that he needs to look into some sort of sound API so that he can make a geiger-counter click every time someone accesses his website	21:25
Crell	shruggar: Ooo...	21:25
shruggar	Crell, ever used nullsoft "beep"? :)	21:26
XDS2010	please read my last statement	21:26
shruggar	I want a networked version of that	21:26
Crell	XDS2010: Is this a project you manage, or one you are just monitoring?	21:26
shruggar	XDS2010, what do you hope to gain from seeing patches flash up on your screen and then go away without your input?	21:26
Crell	shruggar: No, but a friend of mine once logged into his roommate's computer and piped the kernel file to the sound device.	21:26
	Scared the crap out of him.	21:26
XDS2010	shruggar: direction	21:27
	← RickRaven left	21:27
XDS2010	its collaborative	21:27
Crell	XDS2010: You must read very very very fast...	21:27
shruggar	XDS2010, you're not making any sense. What would this solve that an RSS feed wouldn't?	21:27
XDS2010	I am very very very fast..	21:27
	:P	21:27
	XDS2010 shrugs	21:27
tango_	XDS2010: the proper way to do this is, everybody commits to their local repositories and then submit the patchset for review before merging upstream	21:28
	XDS2010: that's how distributed revision control works	21:28
XDS2010	what about projects with 0 administration ?	21:28
	^.^	21:28
	or 1 administrator FTM	21:29
tango_	XDS2010: if you want to review the patches as they fly by, you're not really administering	21:29
	→ ignatp joined	21:29
	← Titosemi left	21:29
	→ mabrand joined	21:29
XDS2010	ok	21:29
tango_	XDS2010: they send the patches to the administrator for review, and if the admin likes them, he merges them upstream, otherwise he asks for a rework of the patches	21:29
XDS2010	ok this is becoming a argument	21:30
	its what i want, its my opinion, im entitled to it, im sorry if you dont approve	21:30
tango_	XDS2010: if you want to review them as they are being done, ask the devs to create a post-commit hook that sends you the patch. I already mentioned it two times	21:31
	→ vdboor_ joined	21:31
	vdboor_ → vdboor	21:31
tango_	XDS2010: but from what you said I think you want a screencast on your devs rather	21:32
XDS2010	ok i suppose i could setup a server to do this	21:32
Crell	XDS2010: ... entitled?	21:33
	← ziro` left	21:33
XDS2010	entitled ?	21:33
	← nighthwk1 left	21:33
	→ ziro` joined	21:33
XDS2010	ok i suppose i could setup a server to do this , via python	21:33
tango_	Crell: entitled to some opinion, it's correct english	21:33
Crell	Ah.	21:33
tango_	XDS2010: a server to do what?	21:33
XDS2010	parse the post commit hooks	21:34
	GMAIL is push	21:34
	→ thrope joined	21:34
XDS2010	i'll just pull	21:34
shruggar	you're entitled to whatever broken workflow you want, git is flexible. But we'll tell you that git is flexible, so there's probably a better way :)	21:34
XDS2010	with -0	21:34
tango_	XDS2010: just let the post-commit hooks push the data to you	21:34
XDS2010	shruggar: im open to all plausibilities	21:35
tango_	XDS2010: and remember that wil git commits are local to the dev machine, so each dev should install the post-commit hook	21:35
	s/wil/with	21:35
XDS2010	wil ?	21:35
	oh	21:35
	nvm	21:35
tango_	XDS2010: and no you're not open because we have told you about better ways to manage the workflow and you stubbornly persist in wanting yours	21:35
XDS2010	i cant have my own opinion ?	21:36
	what is this aticka ?	21:36
tango_	of course you can but don't claim openness 8-D	21:36
XDS2010	wtf are you talking about ?	21:36
tango_	nevermind	21:37
	← chrislerum left	21:37
shruggar	XDS2010, just as soon as I understand whatever it is you're trying to do. I tend to browse git-log in the morning to see whatever random commits people have sent in to our central repos, and we have scripts which summarize who has changed what which we run prior to release, just to make sure we all know who to ask about individual changes which may or may not have been properly reviewed	21:37
	← flazz left	21:37
XDS2010	shruggar: if i could have an implant put in my brain of the commits i would	21:38
shruggar	XDS2010, that's pretty much what git-log is ;)	21:38
kristian42	How can I find out which files have been deleted between here and another commit ?	21:39
XDS2010	shruggar: what could be so bad about seeing patches in real time ?	21:39
	← pavlix left	21:39
XDS2010	twitter git	21:40
	← redondos left	21:40
XDS2010	:)	21:40
shruggar	kristian42, git log --name-only --diff-filter=D <other-commit>..	21:40
	← kevinfish left	21:40
	→ redondos joined	21:40
	→ pavlix joined	21:40
doener	shruggar: wouldn't diff make more sense than log there?	21:40
shruggar	doener, yes it would	21:40
	→ moops joined	21:41
shruggar	XDS2010, the problem is the "real time" part. It implies that you want the display to update whenever a new commit comes in, which implies that you don't actually want to read the commit's information, which implies that you don't actually want to do it at all	21:42
	← zenazn_ left	21:42
XDS2010	shruggar: for a separate project, i want to be able to see the direction its taking , so that is exactly what i want if i am merging branches.	21:43
	get it ?	21:43
	← mabrand left	21:44
XDS2010	If i administrate 2 completely separate trains and one is moving in one direction and the other in another , you are going to bet your hairy but i would want to view commits as they come in. shruggar	21:45
	→ zenazn joined	21:45
shruggar	XDS2010, if it were anything other than "the patches" maybe. If it were "the summaries", "the log messages", even "the diff-stats", that might make sense. But in any reasonably-active project, the patches coming in would be an unreadable blur.	21:45
	← kenichi left	21:45
XDS2010	why would they be unreadable blur ?	21:46
drizzd	what does it matter? nobody is keeping you from doing what you want to do	21:46
XDS2010	if the project is open source and in plain text	21:46
shruggar	XDS2010, because you want it to move "in real time", without your input	21:47
XDS2010	but i wouldn't be directly overseeing the other part of the project	21:47
	think of it as having 2 admins on the same train	21:48
	with equal power	21:48
	you have different cars and different speeds	21:48
	one at one end and one at the other	21:49
	shruggar: what i am getting at is , realtime stats are just THAT important	21:49
doener	XDS2010: pushing 20 commits at a time isn't uncommon, often that's even a low number. The log (with patches) for the last 20 commits in git.git is ~700 lines, for linux-2.6.git it's ~900 lines	21:49
Crell	stats != patches.	21:49
	→ tsundsted joined	21:49
XDS2010	doener: thank you	21:50
doener	XDS2010: having that just thrown at you in realtime seems pretty much useless to me, it takes way too long to digest	21:50
	→ homiziado joined	21:50
XDS2010	see shruggar , i think we have evolved at some point	21:50
	→ tvw joined	21:50
	← enoj left	21:50
Crell	XDS2010: It sounds almost like you're looking for a realtime version of this: http://code.google.com/p/gource/	21:50
XDS2010	where we can actually use patch data	21:50
shruggar	XDS2010, we all agree that stats are important. If you want realtime stats, ask for those. Patches are just.. stupid to view like that	21:50
	→ chrislerum joined	21:51
	frakturfreak_ → frakturfreak	21:51
XDS2010	ok so i must be stupid then ?	21:51
	i'll shut up now.	21:51
shruggar	XDS2010, so I think you should be having a computer run some statistics for you, instead of looking at a blur of meaningless lines	21:51
XDS2010	evolution	21:52
shruggar	XDS2010, this is what we in 2010 do. It's why we use things like git, which have automatic rename detection	21:52
XDS2010	shruggar: for one i have single vision	21:52
	i can visualize and understand 2 things at once	21:53
	most people cant	21:53
	→ phlowx joined	21:53
Ilari	XDS2010: If you want realtime logging, you presumably need to do so from hook or signal client doing it from the hook.	21:53
Crell	XDS2010: You're saying that you can read, and comprehend, a diff file of 300 lines thrown at you for about 4 seconds before being replaced by another?	21:53
	That's the part we don't get.	21:53
XDS2010	correct	21:53
	i am single sited	21:53
	< sigh >	21:54
Ilari	4 seconds? When multiple commits are pushed at once, its much less than that...	21:54
Crell	You'll forgive us if we find that a bit hard to believe, as it would suggest that you are not a homo sapien.	21:54
XDS2010	Ilari: well it would depend on how many commits where being pushed, but i can keep up	21:54
shruggar	XDS2010, git log HEAD~20..\|cat it doesn't matter how fast you can process information. You are not going to be able to read that, and that's without -p	21:54
XDS2010	shruggar: Crell http://wiki.answers.com/Q/What_is_single_vision	21:55
	yes i can to answer your question.	21:55
	twitter is my friend :)	21:55
	← johnw left	21:55
	→ oriba joined	21:55
shruggar	XDS2010, I don't care what "single vision" is, the refresh-rate of your monitor isn't high enough to do what you're asking	21:55
	← oriba left	21:55
drizzd	XDS2010: have you tried "watch git log -p"?	21:56
XDS2010	no	21:56
drizzd	it only updates once per second though	21:56
	→ hasrb joined	21:56
drizzd	not sure if that's fast enough for you	21:56
XDS2010	shruggar: so i'll use 2 monitors or 3	21:56
shruggar	XDS2010, so my assumption is that what you're conveying that you want isn't what you want	21:56
	← wilhelmtell left	21:56
	← aless left	21:56
XDS2010	omfg	21:56
	/ignore shruggar	21:57
Ilari	XDS2010: If its single line per commit, then probably you could keep up. But definitely not with full diffs...	21:57
XDS2010	what a nut	21:57
	its single line Ilari	21:57
	→ _Vi joined	21:58
Ilari	XDS2010: Then have hook signal pushes to client, which then fetches the newly-pushed commits and displays them on screen?	21:58
XDS2010	if i can parse that yes	21:59
	← ntoll left	21:59
Ilari	XDS2010: Even opening TCP connection and then immediately closing it is enough for signaling that client should update the display...	22:00
	→ magicbronson joined	22:00
	← shruggar left	22:00
XDS2010	thats the idea	22:00
	→ Tuller joined	22:01
XDS2010	im going to check out gsource	22:01
Ilari	Or have client keep open TCP connection with server. The server then writes something to all connections that it has open when push happens.	22:01
	← bentob0x left	22:02
	← LiamH left	22:02
	← phlowx left	22:02
	← _Vi left	22:02
	← Thumper_ left	22:03
	→ Thumper_ joined	22:03
	← peper left	22:03
	← Thumper_ left	22:04
	→ Thumper_ joined	22:04
	→ eletuchy_ joined	22:04
	→ makeworkeasy joined	22:05
	← makeworkeasy left	22:05
	→ tychoish joined	22:05
	← eletuchy left	22:05
	← juanjoc left	22:06
	← chrislerum left	22:07
	← tizbac left	22:07
	→ Q\| joined	22:07
	→ eletuchy joined	22:08
	← krawek left	22:09
	→ krawek_ joined	22:09
	← homiziado left	22:09
Q\|	Since git "doesn't care about files", is it still possible to issue something like "copy all folders / files from snapshot X to snapshot Y to folder Z"?	22:10
	← eletuchy_ left	22:11
	← MattDiPasquale left	22:12
drizzd	Q\|: git read-tree --prefix=Z/	22:12
	← gbacon left	22:13
	→ enoj joined	22:13
	← Crell left	22:13
drizzd	you probably want to pass -u too, in a non-bare repo	22:13
XDS2010	Q\| probably	22:13
	→ MattDiPasquale joined	22:14
	← swift-prime left	22:15
	← mbroeker left	22:15
	← alester left	22:16
	← boaz left	22:16
	← harinath left	22:16
	→ _Vi joined	22:17
	→ homiziado joined	22:17
	→ name joined	22:18
	← moops left	22:18
	→ peper joined	22:19
BearPerson	is there a way to hook git fetch / git remote update the way you can hook "git push" ?	22:20
thiago	do you mean you want to run a hook on the server when someone fetches?	22:21
	→ shruggar joined	22:21
	→ Sjors joined	22:21
Sjors	Hi!	22:21
	← ignatp left	22:21
Sjors	I'm wondering: I have a git repository in a/ and a git mirror of a in b/	22:21
BearPerson	no, the other way around - on the local repository whenever I fetch	22:22
	→ ignatp joined	22:22
Sjors	in the situation that I sometimes commit something in a/, and then run git fetch in b and it fetches the commits for all branches	22:22
	← TeckniX left	22:22
Sjors	now I want a git hook to be executed, preferably one for each commit, but if there's no way, one if anything's actually changed	22:22
	anything I can put in hooks/ to do that? I've tried post-commit, update and post-receive, none seemed to be executed seeing that their STDOUT stuff didn't end up on my terminal	22:23
	← rtl left	22:23
BearPerson	try post-merge	22:23
Sjors	I don't see a post-merge.example in my hooks/ directory, but allright, let's see	22:23
_Vi	Sjors, May be try STDERR?	22:23
Sjors	allright, that too	22:24
	nope - still nothing, _Vi, BearPerson, tried both	22:25
	← didlix left	22:25
	← RobertLaptop left	22:25
	← LongBeach left	22:26
	← DavidKlein left	22:27
	→ wilhelmtell joined	22:27
	← eighty4 left	22:27
	← drizzd left	22:27
_Vi	Sjors, Are your hooks has executable flag? (chmod +x .git/hooks/...)	22:28
	← portal left	22:29
	→ iatrux joined	22:29
	← iatrux left	22:30
	→ harinath joined	22:31
Sjors	_Vi: yes :)	22:31
	→ iatrou joined	22:31
_Vi	Sjors, Then check with "strace -fe execve" does the git really executes your scripts.	22:31
	→ iruediger joined	22:32
thrope	how can I revert a single commit?	22:32
_Vi	thrope, 1. "git revert <commit-id>"	22:32
thrope	ah thats easy	22:32
_Vi	thrope, It will create new commit that negates <commit-id>.	22:32
thrope	i did googlee but found only more complicated stuff	22:32
BearPerson	_Vi, probably not, since according to man githooks none of those hooks is strictly supposed to get called by fetch	22:33
Gitbot	_Vi: the githooks manpage can be found at http://git.or.cz/man/githooks	22:33
thrope	thanks	22:33
BearPerson	most of the update stuff is specifically only run when a "git push" happened	22:33
thrope	can I checkout the file to a state before the commit first to check	22:33
	that I really want to rever it	22:33
_Vi	thrope, "git checkout <commit-id> -- path/to/file"?	22:34
	→ airborn joined	22:34
thrope	thanks	22:34
	→ nevyn^ joined	22:34
_Vi	thrope, Is this "bad commit" already published?	22:34
thrope	yes	22:34
airborn	hi	22:34
thrope	first I just want to rever it for testing	22:34
	← galderz left	22:34
	→ anjen joined	22:35
thrope	I think I made a mistake - it says my branch is ahead of origin/win64 by 1 commit	22:35
	I did git checkout anotherbranch -- file_i_wanted_to_revert	22:36
_Vi	thrope, "git revert" produces additional "reverting" commit.	22:36
Sjors	_Vi: It doesn't - I even made them create a file in /tmp, but no new file appears there	22:36
	← beatak left	22:36
BearPerson	thrope, the git checkout id -- file does not change your current commit or branch, it just changes a file in your checkout	22:37
	← zenazn left	22:37
thrope	oh its ok I think there was another commit i fogot I hadnt pushed yet...	22:37
	← raichoo left	22:38
_Vi	Sjors, What command do you expect to execute hooks?	22:38
	← tychoish left	22:38
	← pavlix left	22:38
	← unreal left	22:38
	← coderdad left	22:38
	← stringo0 left	22:38
	← Oxf13 left	22:38
	← spearce left	22:38
	← malumalu left	22:38
	← dbatbold left	22:38
	← parasti left	22:38
	← yaotti_aw left	22:38
	← luke-jr left	22:38
	← tango_ left	22:38
	← GodEater left	22:38
	← niki left	22:38
	← nekoget left	22:38
	← Mage_Dude_ left	22:38
	← fractal_heart left	22:38
	← dahankzter left	22:38
	← iksik left	22:38
	← bmalee left	22:38
	← pdnsmith left	22:38
	→ nekoget joined	22:39
	→ rgr joined	22:39
	→ Circuitsoft joined	22:39
	→ zenazn joined	22:40
	→ parasti joined	22:40
	← parasti left	22:40
	← nekoget left	22:40
	← name left	22:40
	← eletuchy left	22:40
	← aspotashev\|eeepc left	22:40
	← andrewl left	22:40
	← zimbres left	22:40
	← sh1mmer left	22:40
	← Cesario left	22:40
	← ehabkost left	22:40
	← kenyon left	22:40
	← timj_ left	22:40
	← meuh left	22:40
	← Grum left	22:40
	← koki left	22:40
	← avtobiff left	22:40
	← kokx left	22:40
	← Lgb left	22:40
	← cworth left	22:40
	← naquad left	22:40
	← slakmagik left	22:40
	← kingfishr left	22:40
	← kristian42 left	22:40
	← Vadtec left	22:40
	← broonie left	22:40
	← TBH left	22:40
	← cehteh left	22:40
	← Slush- left	22:40
	← FauxFaux left	22:40
	← karihre left	22:40
	← Hadaka left	22:40
	→ TheUni joined	22:40
Circuitsoft	Hello.	22:41
	Is it possible to ask git-svn to only fetch up to a particular revision?	22:41
	→ halfline joined	22:41
	→ up_the_irons joined	22:41
	→ raim0 joined	22:41
	→ jlogsdon joined	22:41
	→ cota joined	22:41
	→ cxreg joined	22:41
	→ cYmen joined	22:41
	→ Aikawa joined	22:41
	→ Caius joined	22:41
	→ rudi_s joined	22:41
	→ ahf joined	22:41
	→ dmlloyd joined	22:41
	→ redbrain joined	22:41
	→ mebus joined	22:41
	→ wimt joined	22:41
	→ Husio joined	22:41
	→ David-T joined	22:41
	→ mattw joined	22:41
	→ bburhans joined	22:41
	→ bartman joined	22:41
	→ Hunner joined	22:41
	→ Bombe joined	22:41
	→ Stravy joined	22:41
	→ jjcm joined	22:41
	→ Rockj joined	22:41
	→ MarcWeber joined	22:41
	→ TML joined	22:41
	→ Nilbus joined	22:41
	→ s4msung joined	22:41
	→ mapreduce joined	22:41
	→ Woody2143 joined	22:41
	→ ribasushi joined	22:41
	→ tahorg joined	22:41
	→ xou joined	22:41
	→ thomas_b joined	22:41
	→ Tuomas_ joined	22:41
	→ nick_h_ joined	22:41
	→ Ilari joined	22:41
	→ rapha joined	22:41
	→ arkx_ joined	22:41
	→ jspiros joined	22:41
	→ Tass_ joined	22:41
	→ NAA1ATN joined	22:41
	→ Gitbot joined	22:41
	→ py1hon joined	22:41
	→ iksik joined	22:41
	→ dahankzter joined	22:41
	→ fractal_heart joined	22:41
	→ Mage_Dude_ joined	22:41
	→ NAA2AXS joined	22:41
	→ niki joined	22:41
	→ GodEater joined	22:41
	→ tango_ joined	22:41
	→ luke-jr joined	22:41
	→ bmalee joined	22:41
	→ yaotti_aw joined	22:41
	→ dbatbold joined	22:41
	→ malumalu joined	22:41
	→ spearce joined	22:41
	→ Oxf13 joined	22:41
	→ coderdad joined	22:41
	→ unreal joined	22:41
	→ pavlix joined	22:41
	→ tychoish joined	22:41
	→ nekoget joined	22:41
	→ name joined	22:41
	→ eletuchy joined	22:41
	→ aspotashev\|eeepc joined	22:41
	→ andrewl joined	22:41
	→ zimbres joined	22:41
	→ sh1mmer joined	22:41
	→ Cesario joined	22:41
	→ ehabkost joined	22:41
	→ kenyon joined	22:41
	→ timj_ joined	22:41
	→ meuh joined	22:41
	→ Grum joined	22:41
	→ koki joined	22:41
	→ avtobiff joined	22:41
	→ kokx joined	22:41
	→ TBH joined	22:41
	→ Lgb joined	22:41
	→ cworth joined	22:41
	→ naquad joined	22:41
	→ slakmagik joined	22:41
	→ Vadtec joined	22:41
	→ kingfishr joined	22:41
	→ kristian42 joined	22:41
	→ broonie joined	22:41
	→ cehteh joined	22:41
	→ Slush- joined	22:41
	→ FauxFaux joined	22:41
	→ karihre joined	22:41
	→ Hadaka joined	22:41
	→ ruskie joined	22:41
	← tar- left	22:41
	→ neutrino joined	22:41
	← brett_h left	22:42
	→ parasti joined	22:42
	→ ghost_olli joined	22:42
	← ghost_olli left	22:42
_Vi	Circuitsoft, Start "git svn fetch" and Ctrl+C it when it fetch this revision.	22:42
	*fetches	22:42
	→ bburhans_ joined	22:43
	neutrino → tar-	22:43
	← luke-jr left	22:43
	← unreal left	22:43
	← bburhans left	22:43
	← Tass_ left	22:43
	← Aikawa left	22:43
Ilari	Heh... Enough users joining to cause lag... :-)	22:43
	← iksik left	22:43
	→ iksik joined	22:44
Circuitsoft	Oh. The reason I'm doing this is to get the git repository to as consistent as state as possible before finishing the job on another computer.	22:44
	→ unreal joined	22:44
	→ luke-jr joined	22:44
	← iksik left	22:44
	← ziro` left	22:44
	→ chrislerum joined	22:44
	→ Aikawa joined	22:44
_Vi	Circuitsoft, Other hack is to filter it's output through a script that will exit when meet "r33445". But this is also not clean way of interruption.	22:44
	→ stringo0 joined	22:45
_Vi	Circuitsoft, I don't know if there is a clean way, but dirty way may work, e.g. you can repair such repository later.	22:45
	→ iksik joined	22:45
	← iksik left	22:45
	→ Tass_ joined	22:45
	← berserker left	22:46
	← chrislerum left	22:46
	→ iksik joined	22:46
	← iksik left	22:46
	← AkumaStreak left	22:46
Circuitsoft	Ok.	22:46
	→ iksik joined	22:46
Circuitsoft	I guess I'll just move the repo back to the big server as-is and finish the fetch.	22:47
	Basically, I'm just trying to make a backup of our whole svn repository in git so that if/when it dies all is not lost.	22:47
	→ ajpiano joined	22:47
	← p3rg2z left	22:47
	← doener left	22:47
	→ berserker joined	22:47
Circuitsoft	At one time, someone committed a .git folder into svn, and the version of git on the server still has the bug where it doesn't properly ignore them.	22:47
_Vi	Circuitsoft, It is not so easy to kill git repository using it's commands (even with interruptions).	22:48
	← coderdad left	22:48
	← jjuran left	22:48
	→ jjuran joined	22:48
_Vi	Circuitsoft, "git-svn" should ignore all ".git" directories.	22:48
Circuitsoft	I was just hoping to have only /objects/pack/* files so I wouldn't have so many files to move over.	22:48
	← Q\| left	22:49
	← thiago left	22:49
Circuitsoft	It does in 1.6.4.4. It doesn't in 1.5.5.6	22:49
Sjors	_Vi: git fetch	22:49
_Vi	Circuitsoft, If you have only pack file you usually can recover most of repository.	22:49
Sjors	_Vi: I'm writing a script myself for it now - it gets the status from refs/head, does a git fetch, then gets the status again	22:49
Circuitsoft	I mean I want to have only pack files.	22:50
_Vi	Sjors, Hooks are generally for "git push".	22:50
Sjors	it's supposed to then calculate the difference, and execute a script for every one	22:50
	_Vi: you can't run git push in a bare repository	22:50
Circuitsoft	Otherwise there are hundreds of objects/??/* files.	22:50
	→ chrisler_ joined	22:50
	← dreiss left	22:50
Ilari	Sjors: Eh? Git push doesn't need nor use working copy...	22:50
Sjors	ehhh	22:50
	I thought it said git pull. sorry	22:50
	anyway: we don't do git push	22:51
	we only do git fetch, we push nowhere	22:51
BearPerson	hmm	22:51
_Vi	Circuitsoft, Usually after "git gc" important things get packed to packs, and cruft gets away to plain files in objects directory.	22:51
	→ brett_h joined	22:51
Circuitsoft	I've done gc. Can you define cruft?	22:51
BearPerson	you could push to ., but that'd be a strange hack and I'm not sure if you could get it to work at all	22:51
	← chrisler_ left	22:52
_Vi	Sjors, "git push" does not refuse in a bare repository.	22:52
	← rayners left	22:52
_Vi	Sjors, And I don't see reasons why it does need a working copy.	22:52
	→ JStoker joined	22:52
	→ coderdad joined	22:52
BearPerson	he confused it with "git pull", as mentioned.	22:52
_Vi	Sjors, "git fetch" can definitely run in a bare repository (but not "git pull").	22:52
	Sjors, "git svn fetch" runs in a bare "git-svn" repository, but glitches.	22:53
BearPerson	Circuitsoft, maybe you want git-repack	22:53
	or git gc --agressive	22:53
	← ExtraSpice left	22:54
_Vi	Circuitsoft, After "git gc --aggressive" there is usually one big pack file that contains all database (except of things references only be reflogs etc.).	22:54
	*referenced only by	22:55
Circuitsoft	gc --aggressive gave me an "out of memory" error.	22:55
	I don't think I actually have enough hard drive space to finish a gc --aggressive on this repo on this machine.	22:56
	← fractal_heart left	22:56
_Vi	Circuitsoft, If you have only that pack file, you can 1. index it "git index-pack", 2. Create a repository, 3. Put the pack and index to it's place, 4. Run "git fsck". It will show the refs (as "dangling commits"), 5. Create branches for each ref.	22:56
	Circuitsoft, You need to setup limit somewhere to make "git gc --aggressive" not so memory-aggressive.	22:57
Circuitsoft	Which is why I'm moving the repo back to the server to finish there, and I'm hoping there aren't any more .git updates.	22:57
_Vi	Circuitsoft, It occupies virtual memory, not hard drive space, AFAIK.	22:57
	← zenazn left	22:58
_Vi	Circuitsoft, Why so big repository? Big binary blobs in it?	22:58
Circuitsoft	It occupies HD space while making a new pack before deleting hte old one.	22:58
	Some. I work at a consulting firm. There's a single SVN repo with all the work we've ever done.	22:58
	← malumalu left	22:58
_Vi	Circuitsoft, Big binary blobs should not be in the repository. At least they should not be in the same repository as the usual code.	22:59
Circuitsoft	I've been here over two years, and this same SVN repo has been used since long before i got here.	22:59
	← rgr left	23:00
Circuitsoft	The git backup is 3.5GB so far, and we're still over 10000 revs back on svn in my backup.	23:01
_Vi	Circuitsoft, You may "git svn clone" it multiple times, each time different part (if there are no/few moves between such parts).	23:01
adymitruk	I'm trying to make a submodule out of a repository that's at the same level on my disk as the main repo. Git submodule add ../otherProject otherProject does not work.. What's the porper syntax for for that?	23:01
_Vi	Circuitsoft, 3.5GB is a git too much for normal git work.	23:01
	Circuitsoft, (Especially "git gc")	23:01
	→ Q\| joined	23:01
	→ doener joined	23:01
_Vi	adymitruk, Submodule checkout points should be inside.	23:02
Circuitsoft	I don't intend to ever have this checked out. It'll be bare only. What I may do is use http://github.com/apenwarr/git-subtree to split out chunks if people want them.	23:02
_Vi	adymitruk, May be you should create "top-level" repository?	23:02
	← PerlJam left	23:03
adymitruk	_Vi, I'm not sure I understand	23:03
	At this point I do not need a containing submodule	23:03
	I would like project B to be a submodule of project B	23:04
_Vi	adymitruk, "project B to be a submodule of project B" - queer scheme.	23:04
Sjors	_Vi: here, git push refuses in a bare repository	23:04
	_Vi: because it needs a working dir	23:04
	_Vi: here, git fetch works fine, which is what we use, but it doesn't do hooks	23:04
adymitruk	they both sit on my machine, so I do not want to install a daemon just to support a git: protocol	23:04
_Vi	Sjors, Why does it need a working directory?	23:04
	← leavittx left	23:04
Sjors	_Vi: because it says so	23:05
_Vi	_Vi, Push does not merge any things.	23:05
Sjors	wait	23:05
	wait	23:05
	I meant git pull :(	23:05
	again	23:05
	← sylr left	23:05
	← leavittx_ left	23:05
Sjors	_Vi: git pull doesn't work. Git push does work, but we have nowhere to push to	23:05
_Vi	Sjors, git pull == git fetch + git merge (or "git rebase").	23:05
	→ warthog9 joined	23:05
Sjors	and we don't want to push anywhere, we want this repository to be the endpoint	23:05
_Vi	Sjors, Merge can conflict. Conflict resolution is user's work. So merging requires working directory. So merging can't be in bare.	23:06
	→ zenazn joined	23:06
Sjors	_Vi: I understand	23:07
	← corni left	23:07
adymitruk	It would work if I had an absolute path /tmp/repo/projectA	23:07
	but I need it to be a relative path	23:07
	→ RickRaven joined	23:08
_Vi	adymitruk, Again, which command do work for absolute path, but not for relative?	23:08
adymitruk	git submodule add	23:08
	← NetHawk left	23:09
adymitruk	like in this blog post: http://speirs.org/blog/2009/5/11/understanding-git-submodules.html	23:09
_Vi	adymitruk, OK, understood. Does something like "file://../submodule" work?	23:09
	← Yuuhi left	23:09
	← hobodave left	23:09
RickRaven	hi i am new to git and version control systems. i dont know much about these. i want to learn basics. i am reading about these and now i have some questions. first : can i use git completely local? or do i need a git hosting?	23:09
_Vi	adymitruk, Usually URIs of submodules point to some server.	23:10
	RickRaven, Yes, you can use git locally.	23:10
	RickRaven, You can also create two local repositories and use one as "server" and the other as "client".	23:10
RickRaven	_Vi: why whould i want to do that?	23:11
adymitruk	so why do absolute paths work but relative ones don't? ... I'm thinking that I just don't have the syntax right	23:11
	← shruggar left	23:11
	← pavlix left	23:11
_Vi	RickRaven, To learn commands for dealing with servers.	23:11
RickRaven	oh i see	23:11
_Vi	RickRaven, (push, fetch)	23:11
	← coppro left	23:12
_Vi	adymitruk, Usually paths for submodules are URL where everybody can fetch them.	23:12
	→ NetHawk joined	23:12
RickRaven	_Vi: so if i set one as server, can a friend of mine connect the server?	23:12
	← jensn left	23:13
_Vi	RickRaven, If he can login to you through SSH, yes. If you start "git daemon" then also yes.	23:13
	RickRaven, (Technically there are not much difference between "server-style" repository and usual with working copy).	23:14
	← hugoxrosa left	23:15
RickRaven	_Vi: pretty stupid question i guess but i have to ask : i have ubuntu and i am also new to linux. if i create an account friend, can my friend connect through ssh?	23:16
_Vi	RickRaven, (thinking). Probably, yes.	23:16
	RickRaven, 1. "useradd qqq". 2. "passwd qqq", enter and confirm password,	23:16
	→ brizly joined	23:17
_Vi	RickRaven, 3. "ssh qqq@localhost", enter the password, verify it works and you can login.	23:17
	→ xiong joined	23:17
Ilari	RickRaven: If this is for Git repo hosting, better to use gitolite...	23:17
	← brizly1 left	23:17
_Vi	RickRaven, 4. (logged in as qqq) "cd; mkdir repo; cd repo; git init --bare";	23:18
	RickRaven, 5. Check it out yourself: "git clone [email@hidden.address]	23:18
	→ dkulchenko joined	23:19
_Vi	RickRaven, 6. Tell your friend to check out it "git clone qqq@your_ip:/home/qqq/repoi".	23:19
	RickRaven, This is manual setup for git repository over SSH. There are tools to simplify it (gitolite, gitosis).	23:19
adymitruk	for some reason, local submodules can be added if the URI is an absolute path. A relative path will want the submodule repository to have a remote defined	23:20
_Vi	RickRaven, But anyways you should first just to add and commit something locally.	23:20
adymitruk	very strange	23:20
dkulchenko	Hi all! I've added a post-receive hook which is outputting to the console. My hook runs a forked script. If I run the hook by hand (from the shell, not via git) it exits when it should. If I let git run the hook, it hangs, waiting for the forked process to finish. Why? and how can I fix this?	23:20
	← nevyn^ left	23:21
	← priidu__ left	23:21
	→ nevyn^ joined	23:21
_Vi	dkulchenko, How do you fork a process?	23:21
	→ PerlJam joined	23:21
RickRaven	_Vi: great guide thank you, i will check other tools as well. but i have some more questions : why would i want to use a hosting service instead of owning my own server? just another question : what happens when i commit? Does it save and send my files to server?	23:22
patrikf	dkulchenko: you probably want to use setsid	23:22
dkulchenko	_Vi: my commit hook is a Perl script, i use Perl's fork()	23:22
_Vi	RickRaven, Yes, you can use, for example, github. If is easier than setting up networked repo yourself.	23:22
dkulchenko	patrikf: how?	23:23
patrikf	RickRaven: commit is local to the repo. there's a separate command for updating the server's state ("push")	23:23
_Vi	RickRaven, In git, there is following workflow:	23:23
	← vdboor left	23:23
RickRaven	patrikf: thank you for pointing it out	23:24
_Vi	RickRaven, 1. You change or create some files. "git status" shows "Changed, but not updated" or "Untracked".	23:24
	RickRaven, 2. You add (stage) files: "git add ." Now "git status" shows your files are "Modified" or "New". Then are prepared for committing.	23:24
	← zenazn left	23:25
dkulchenko	I'm not sure why git is waiting for the fork. The shell doesn't.	23:25
_Vi	RickRaven, 3. You commit "git commit". Now files you have prepared saved as commit (with date, commit message, authorship etc). After committing there are not "prepared" files again.	23:25
	dkulchenko, Use bash as wrapper.	23:25
	← nevyn^ left	23:26
dkulchenko	_Vi: hmm good idea	23:26
patrikf	dkulchenko: POSIX::setsid or setpgrp should do the trick	23:26
_Vi	RickRaven, 4. You push (publish) commit[s] to the server. "git push origin master:master".	23:26
	RickRaven, Now already made commits are uploaded to server. You are better not to edit such commits after "push".	23:26
dkulchenko	patrikf: what does setsid do, anyway?	23:26
patrikf	dkulchenko: i'm sure you'll find lots of documentation on this topic	23:27
dkulchenko	patrikf: okay, thanks	23:28
patrikf	dkulchenko: i'm not the knowledge base, just the guy pointing in the right direction :-)	23:28
_Vi	RickRaven, You can use "git gui" to stage, commit (may be to push). Also you can use "gitk" to view the commit history in beautiful GUI form.	23:28
	→ Tuller_ joined	23:28
RickRaven	_Vi: thank you. i heard that there is a tool called tig for easy of use. should i use this or the original git tool? what would you suggest? should i... before i ask you almost answered my question. so, git gui, gitk or git or tig. which should i learn?	23:29
_Vi	RickRaven, After "git commit" the data is permanently stored into your repository and should not be lost (if the repository itself not broken). After "git push" the data is stored on server.	23:29
adymitruk	RickRaven, try progit.org/book	23:29
	← dkulchenko left	23:30
_Vi	RickRaven, I haven't used neither "tig" nor "cogito". I would recommend learning to use git entirely from command line + learn to use "git gui" and "gitk" to simplify things.	23:30
	← ShooterM_ left	23:30
_Vi	RickRaven, When you will understand git and be able to solve things when something goes wrong, you can use any tool, including some IDE plugins for git.	23:31
	→ leavittx_ joined	23:31
	→ leavittx joined	23:31
	→ fractal_heart joined	23:32
RickRaven	_Vi: thank you my friend i have no more questions for now. great help, thank you again.	23:32
	← thrope left	23:34
	→ dkulchenko joined	23:34
	← plediii left	23:34
	← airborn left	23:34
dkulchenko	patrikf: setsid didn't do anything :/	23:35
patrikf	dkulchenko: you might also want to close any open fds, esp. stdin/out/err	23:36
dkulchenko	patrikf: ah okay, i'll try	23:36
	but still, why does it terminate from shell, but not from git?	23:36
	← enherit left	23:37
patrikf	dkulchenko: i'm quite sure your hook does terminate, but something (git, but more likely ssh) is waiting for all child processes... or some fd	23:37
dkulchenko	patrikf: ah, that makes sense	23:38
_Vi	dkulchenko, May be try more "strace" to see what is it waiting for?	23:38
	← RickRaven left	23:38
dkulchenko	patrikf: wow, you're a genius :P closing stdin/out/err solved the problem	23:40
	← wilhelmtell left	23:42
patrikf	dkulchenko: isn't forking fun. my last stroke of genius was having a pipe() after fork()...	23:42
dkulchenko	patrikf: that's pretty great ;)	23:43
	← lispy left	23:43
	→ Heimidal joined	23:44
	← Tuller left	23:47
	Tuller_ → Tuller	23:47
	→ petercoulton joined	23:50
	← rox` left	23:52
	← frakturfreak left	23:54
	← doener left	23:54
	→ enherit joined	23:55
	← karptonite left	23:58

Logs Search ←Prev date Next date→

Channels Documentation