IRCloggy #git 2008-09-02

Provider of IRC logs since 2005.

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

	2008-09-02
solo	parasti: k, thx. guess I'll have to resort to learning some svn-fu. :)	00:00
	← SRabbelier left	00:01
	Arrowmaster → Arrow	00:02
	Arrow → Arrowmaster	00:03
	← neoeinst left	00:03
	→ neoeinst joined	00:04
	← ben_h left	00:07
	→ ben_h joined	00:08
	→ loincloth joined	00:09
	← lea_w left	00:11
	← mithro__ left	00:12
	→ cilly joined	00:12
solo	parasti: actually, it looks like the format-patch \| am route got me what I needed. not as clean as git merge, but I guess that can't be helped. thx again!	00:13
	← solo left	00:16
	← kukks left	00:17
	→ Khisanth joined	00:20
	← ajonat left	00:22
	← loincloth left	00:27
	← roconnor left	00:27
	← sbeyer left	00:31
	← advany left	00:31
	→ capisce_ joined	00:37
	→ mithro__ joined	00:39
	← mithro__ left	00:39
	→ mithro joined	00:39
	← frando left	00:41
	← ShadeHawk left	00:43
	← eternaleye left	00:44
	→ aiur joined	00:45
	← powr-toc left	00:47
	→ eternaleye joined	00:47
	→ lea_w joined	00:48
	← capisce left	00:48
	← parasti left	00:49
	→ johnw joined	00:53
	← cbreak left	00:55
	← mhutch\|home left	00:59
	← Ademan_ left	01:00
	→ Ademan_ joined	01:01
	→ sztanphet joined	01:02
	→ b10h4ck3r joined	01:02
	→ polpak joined	01:03
	← eternaleye left	01:04
	→ loincloth joined	01:05
	← b10h4ck3r left	01:06
	→ markkalderon joined	01:06
	← markkald_ left	01:07
	→ jtoy joined	01:14
	← ehc left	01:17
	← codeshot left	01:17
	→ zachinglis joined	01:17
	← lea_w left	01:18
	→ kanru joined	01:19
	← kanru left	01:20
	← Henrique left	01:21
	← jmspeex left	01:21
	→ kanru joined	01:22
	stouset- → stouset\|away	01:23
	mlins → mlins\|away	01:24
	→ jmspeex joined	01:24
	← mbroeker left	01:25
	← davidfetter left	01:25
	→ paltman joined	01:26
	→ davidfetter joined	01:27
	mlins\|away → mlins	01:29
	← johnw left	01:30
	→ zachinglis_ joined	01:30
	mw → mw\|out	01:31
	→ johnw joined	01:31
	← zachinglis left	01:31
	← MK_FG left	01:32
	← Arrowmaster left	01:33
	→ leitgebj joined	01:35
	→ ciaran_ joined	01:37
	→ _VVSiz_ joined	01:38
	mlins → mlins\|away	01:39
	mlins\|away → mlins	01:40
	← mw\|out left	01:41
	← ciaran left	01:43
	→ Arrowmaster joined	01:43
	← xet7 left	01:43
	→ lea_w joined	01:48
	← rjent_ left	01:48
	← cilly left	01:54
	← dmhouse left	01:55
	← VVSiz_ left	01:55
	← VVSiz left	01:56
	→ bdiego joined	01:56
	→ VVSiz joined	01:56
	← mlins left	02:02
	← mithro left	02:09
	← mw-home left	02:11
	← neoeinst left	02:13
	← bobesponja left	02:14
	→ mlins joined	02:15
	← eMxyzptlk[away] left	02:16
	→ imyousuf joined	02:16
	→ rmh3093 joined	02:18
	← irotas left	02:20
	mlins → mlins\|away	02:25
	mlins\|away → mlins	02:26
	← mlins left	02:26
	→ rdavila joined	02:26
	← johnw left	02:30
	← lea_w left	02:31
	→ lea_w joined	02:40
	← yoh__ left	02:44
	← loincloth left	02:48
	→ loincloth joined	02:49
	← flaguy left	02:50
	→ MK_FG joined	03:02
	→ johnw joined	03:02
	→ lamont` joined	03:02
	← rdavila left	03:05
	→ harinath joined	03:07
	← lamont left	03:10
	→ cedricv joined	03:15
	← ferdy left	03:15
	← almostautomated left	03:24
	→ CodeOfficer joined	03:25
	lamont` → lamont	03:35
	← lea_w left	03:35
	← tongueroo left	03:36
	→ mithro joined	03:37
	← MK_FG left	03:40
	← zachinglis_ left	03:42
	← loincloth left	03:43
	← pygi left	03:46
	→ mario_ joined	03:46
	→ zachinglis joined	03:46
	← bdiego left	03:48
	mario_ → pygi	03:50
	→ AMorozov joined	03:51
AMorozov	Hi there!	03:51
	Could anybody suggest if I can compare two different subtrees in two revisions?	03:52
	An "obvious" syntax git diff -M <rev1>:some/path <rev2>:some/other/path doesn't work.	03:53
johnw	i think you'll have to clone the project twice, checkout those revisions, and then diff -r DIR1 DIR2	03:53
offby1	you're joking ... ?	03:53
AMorozov	johnw: :-(	03:53
johnw	use clone -s	03:53
offby1	how tedious	03:53
johnw	otherwise, how is he going to be able to recursively diff files?	03:53
	→ lea_w joined	03:55
AMorozov	Well, I certainly could do this (and well, I even don't need clone anything, just checkout & cp -pr), but well, it's not what I expected from a decent VCS.	03:55
	I guess there should be other solution. Hopefully...	03:55
johnw	um, git doesn't even allow you to compare different files, much less directories	03:56
	the git-diff command is for comparing commits	03:56
	what you're talking about is disparate filesystem objects	03:56
AMorozov	If exactly two paths are given, and at least one is untracked, compare the two files / directories.	03:57
	So at least certain "file comparsions" are possible.	03:58
johnw	hm.. i just tried 'git diff foo bar', where foo and bar are both in my repository, and it ignored me	03:59
AMorozov	... at least one is untracked.	03:59
johnw	the closest I can get is: diff <(git show HEAD:foo) <(git show HEAD:bar)	03:59
	but for this to work for directories would require checking out partial trees	04:00
	which brings us to my suggestion :)	04:00
	← imyousuf left	04:02
offby1	``git diff /tmp/hmm /tmp/me'' worked -- those files aren't in any repo	04:02
AMorozov	Certainly.	04:02
offby1	although honestly they were already in the fs, so I didn't gain anything over just plain "diff"	04:02
	← cedricv left	04:04
AMorozov	offby1: Actually I moved a subtree and need to compare it with its location in a previous revision.	04:04
johnw	git should have detected the file moves, can you just compare the commits?	04:05
AMorozov	So I would prefer to somehow use git instead of plain diff.	04:05
offby1	you could diff both the entire commits, and ignore the irrelevant bits ...	04:05
	johnw: oh.	04:06
AMorozov	johnw: I need to narrow the diff only to a subtree.	04:06
	Is it possible to ignore everything but ....	04:06
johnw	try using --relative	04:06
	i'm just look at man git-diff	04:07
Gitbot	johnw: the git-diff manpage can be found at http://git.or.cz/man/git-diff	04:07
AMorozov	johnw: probably it could help. Thank you!	04:07
	→ tjafk2 joined	04:09
	← zachinglis left	04:10
AMorozov	No it doesn't :-(	04:12
	→ irotas joined	04:12
johnw	:(	04:12
AMorozov	At least an "obvious" syntax git diff --relative <other_rev> doesn't help.	04:12
	Probably your initial suggestion is the fastest way to achieve the goal...But well....	04:13
johnw	i think --relative means: if you use git diff $commit1 $commit2, from within a subdirectory, it confines the diff to just files in that directory	04:13
	yeah, by now you'd be done :)	04:13
AMorozov	:-)	04:14
	→ spearce joined	04:15
AMorozov	You know there's a Russian satiric cartoon movie where a griffon teaches an ostrich to fly with a slogan: it's better to waste a day [in training] and then fly in five minutes.	04:20
	So I'm simply learning to fly just as that ostrich :-)	04:20
doener	AMorozov: git diff-tree rev1:path1 rev2:path2	04:23
	add "-p" for a patch, instead of the raw format	04:23
johnw	ah, very cool! thx doener	04:23
AMorozov	doener: hmm, I tried yesterday smth like that and well, it didn't work for me :-(. Anyway thank you a lot!	04:25
	← zobie left	04:25
AMorozov	Smth strange happenned yesterday :-)	04:25
	Thank you!	04:25
	← tjafk1 left	04:26
AMorozov	Ahh... I'm really stupid. Yesterday I mistyped the 'other/path' :-(	04:27
	Thank you all.	04:27
	← AMorozov left	04:27
	← johnw left	04:28
	← harinath left	04:45
	→ harinath joined	04:45
	← johan-s left	04:54
	← lea_w left	04:57
	→ lea_w joined	04:57
	→ cedricv joined	04:58
	→ xjrn joined	05:00
	← lea_w left	05:02
	→ digitalhobbit joined	05:03
	→ alenoosh joined	05:03
	← rmh3093 left	05:04
xjrn	if i dump in xyz.tgz and add a gig to a commit by accident, and i do rm -f xyz.tgz, will that space be reclaimable?	05:05
	err git-rm -fr xyz.tgz	05:05
doener	did you only "git add" the file, or also commit?	05:05
	← igorgue left	05:07
xjrn	yes commited	05:08
doener	did you drop the commit again (git reset)?	05:08
xjrn	nope it's still there, hanging out	05:08
doener	If you just remove the file in a follow-up commit, the space is of course not reclaimable	05:09
	← Ademan_ left	05:09
xjrn	right, yeah, rm'd, commited	05:09
doener	then the object is still required	05:09
xjrn	git-rm -f'd	05:09
	→ Ademan_ joined	05:09
xjrn	so now do exprt the new non-littered workspace, git-reset down to priori revision, commit new workspace (via stash?	05:12
	and then clone or something?	05:12
doener	if the latest commit is the one in which you deleted the file, and the previous one is the one in which you added the file, you can do: git reset --soft HEAD~2	05:13
	and then "git commit"	05:13
	that squashes the last two commits together	05:13
xjrn	having used git-rm is not a difference from leaving the file and then using rm and then soft-reset?	05:17
	→ ph^ joined	05:18
	← CodeOfficer left	05:20
	← markkalderon left	05:21
	← stouset\|away left	05:24
	→ rmh3093 joined	05:31
	→ jtier joined	05:32
jtier	Hello i am trying to install git on red 5 and i get these errors : http://pastebin.com/m35135352	05:33
	← madewokherd left	05:34
	→ marj joined	05:34
	→ Eludias joined	05:35
	→ priidu joined	05:35
cehteh	jtier: looks like you miss some perl stuff	05:36
	→ Beket joined	05:37
	← FH` left	05:42
	← stouset left	05:42
	→ stouset joined	05:45
	← paltman left	05:50
	← k776 left	05:53
	→ Xilon joined	05:53
	→ jonesy joined	05:56
jonesy	just installed Git from source on fresh OS X 10.5.4 Leopard install. When I try to clone a repo I get this error: "fatal: could not create work tree dir `<dir>`	05:57
	if I go back and install version 1.5.5.1 I don't seem to have this problem	05:57
	→ joern_ joined	05:59
jonesy	anyone?	05:59
	← Beket left	05:59
Xilon	I've got a problem with cloning/pushing to a remote repository. http://pastie.org/264351. This seems to only happen when I clone/push from/to that server. When cloning/pushin locally everything works, same goes for the server. Both are running git 1.6.0.1, but the remote has openssl 0.9.8e while I have openssl 0.9.7l. I read that there's a problem with openssl from fink on Mac OS, but I don't use fink.	05:59
	← Dreamer3 left	06:00
	← jtier left	06:00
Xilon	jonesy: cloning works fine (apart for that server^)	06:01
jonesy	I have three different servers I am trying to clone from, no matter what one I try I get that same error.	06:02
	one OS X server and two Linux servers	06:02
DrNick	run the clone in a directory you have write permission to	06:05
jonesy	I've tried it on my desktop, in my sites folder etc	06:06
	I am the owner of the folders I have tried it in	06:06
	← joern_ left	06:07
jonesy	most of the Git versions above 1.5.5.1 have given me this problem. Just reinstalled my machine tonight and hoped I could use the newest version	06:07
	Google doesn't seem to give me any results as to what is causing this	06:07
Xilon	I just tried cloning from github, works just fine (git 1.6.0.1, Mac OS Leopard 10.5.4)	06:10
jonesy	did you compile form source?	06:11
Xilon	yes	06:11
jonesy	are there any extra dependencies needed?	06:12
Xilon	not sure, I don't think so	06:13
	→ deavid joined	06:13
jonesy	I downloaded 1.6.0.1. then ./configure --prefix=/usr/local. then make and make install.	06:13
	did you do anything different?	06:13
	→ gitish joined	06:14
Xilon	no	06:14
	I did install git previously a long time ago, so if there were dependencies, I satisfied them back then	06:14
jonesy	ah	06:14
Xilon	It seems to depend on libz, curl and libcrypto (openssl), all of which, afaik, are installed by default on Mac OS Leopard	06:15
	It sounds more like a permission problem than something to do with git. Can you create files by other means in the same directory?	06:15
jonesy	yup. I even tried to sudo the command when I was first thinking it was a permissions issue	06:16
	the weird thing is, if I go download Git 1.5.5.1, all of these problems go away	06:16
	→ thiago_home joined	06:17
Xilon	hmm maybe I should try git 1.5.5.1 and see if my problem goes away	06:17
	though I think it's openssl and not git	06:17
	Although I compiled git without curl on the server...	06:18
	That's only for http isn't it?	06:18
	→ Dreamer3 joined	06:18
jonesy	I believe so	06:18
	← spearce left	06:18
jonesy	found a tutorial on Github, seems the only dependency is gettext	06:19
Xilon	I have it installed	06:20
	→ phpcurious joined	06:21
jonesy	I am installing gettext now, hopefully that solves the problem	06:24
	→ lea_w joined	06:25
jonesy	nope, same issue	06:32
	← alenoosh left	06:33
phpcurious	hi, newbie question here. can i use git on windows?	06:34
Xilon	under cygwin, apparently	06:35
phpcurious	i tried msysgit ...	06:35
	but don't understand how to use it	06:35
	← gitish left	06:35
phpcurious	is it just under cygwin?	06:35
	← rmh3093 left	06:39
drizzd	phpcurious: it's in your start menu	06:40
	→ JenkingWinlker joined	06:40
jonesy	wonderful. Just tried the Git OS X Installer from Google Code and even that doesn't work	06:41
	same issue	06:41
	← polpak left	06:42
	→ Weaselweb joined	06:42
jonesy	back to Git 1.5.5.1 until I can find what is causing the problem	06:43
drizzd	jonesy: did you delete git-*?	06:45
jonesy	am I supposed to?	06:45
drizzd	if you install 1.6, yes	06:45
	there are no git-* commands in 1.6 any more	06:46
	only 'git <command>'	06:46
jonesy	let me recompile 1.6 again and give it a shot	06:46
drizzd	the command that fails in your case is `mkdir', so it should be something really trivial	06:47
Xilon	Why are the git-* binaries gone?	06:47
	I liked them when working with zsh - quicker to access history	06:48
Mikachu	where have all the good times gone?	06:48
drizzd	several reasons, there was a huge discussion about it on the git mailing list	06:48
Mikachu	Xilon: bind history-beginning-search-backward to some key	06:48
cehteh	no completion for zsh?	06:48
Mikachu	Xilon: it searches everything before the cursor, not just up to the first space	06:48
Xilon	in zsh you can for instance git-push <up> and it goes to the last git-push command	06:48
	oh, awesome	06:48
Mikachu	cehteh: he means typing git pu<up> to get the last git push line, but it will stop at the space and give you git whatever intsead	06:49
jonesy	I don't see git-*	06:49
cehteh	heh and why do people praise zsh so much? :)	06:49
jonesy	installed in /usr/local	06:49
drizzd	which git-clone	06:49
Mikachu	cehteh: ?	06:50
drizzd	(that's a command that will tell you where it is)	06:50
Xilon	cehteh: Umm, it has some nice features, perhaps?	06:50
Mikachu	cehteh: arguably because the default config binds basically no useful functions to keys :)	06:50
Xilon	why do people praise git so much? ;)	06:50
cehteh	Mikachu: lol	06:50
simmel	cehteh Because it's awesome! ;	06:51
	P	06:51
	Xilon Good point!	06:51
Mikachu	http://mika.l3ib.org/s/dot-zshrc	06:51
jonesy	hmmm, even which git-clone is returning nothing now	06:51
Xilon	git-clone doesn't exist in 1.6	06:52
thiago_home	Xilon: people complained that git was difficult and used "git-<tab><tab>" to prove it	06:52
simmel	Xilon Oh? What should one use instead?	06:52
thiago_home	Xilon: we know it not to be an argument at all, but go figure.	06:52
	simmel: git clone	06:52
Mikachu	simmel: 'git clone' :)	06:52
simmel	Oh	06:53
Xilon	"git clone"	06:53
simmel	= D	06:53
	I thought I missed something big = D	06:53
	→ bentob0x joined	06:53
jonesy	still a no go with 1.6	06:53
drizzd	jonesy: what do you mean?	06:54
Xilon	thiago_home: lots of binaries (aliases actually) = difficult?... great argument	06:54
drizzd	jonesy: 'returning nothing'	06:54
jonesy	I get the same error no matter what I do. Error: fatal: could not create work tree dir `dir`	06:54
drizzd	jonesy: did you delete git-* now?	06:54
jonesy	git-* doesn't exist in /usr/local/bin/	06:55
drizzd	that's why I said	06:55
	use 'which git-clone' to find out where it is	06:55
jonesy	which git-clone returns nothing. which git clone returns /usr/local/bin/git	06:56
	→ peritus_ joined	06:56
drizzd	ok, so git-clone is actually gone.	06:56
jonesy	if I got back to 1.5.5.1, everything works fine	06:56
	I even tried the GIt OS X Installer, gives me the same issue	06:57
	not sure what the hell i'm missing	06:57
Xilon	hmm interesting. 1.6.0.1 has a problem with unpacking while 1.5.5.1 doesn't	06:57
drizzd	old server maybe?	06:58
Xilon	git version is the same on both machines	06:59
	with 1.6.0.1 I got this error http://pastie.org/264351	06:59
	works fine with 1.5.5.1 on both machines though	06:59
drizzd	1.5.5.1 is the last version which used a script for clone	06:59
	← tango_ left	07:00
Xilon	So what could be the problem?	07:01
drizzd	jonesy: but 'git init' works?	07:01
Xilon	I googled it and it turns out that openssl from fink causes problems, but I'm not using anything from fink	07:01
jonesy	git init seems to work just fine	07:01
Xilon	I do have an older version of openssl than on the remote though	07:01
drizzd	jonesy: that's really weird. can you do an strace?	07:02
jonesy	I can git init, add files and commit	07:02
	→ johnw joined	07:02
jonesy	I will if you can tell me how to do it, I'm okay on the command line, not great yet	07:03
Xilon	interesting	07:03
	← Ademan_ left	07:03
Xilon	1.6.0.1 locally, and 1.5.5.1 on the remote works as well	07:03
	it's just 1.6.0.1 on both doesn't work	07:03
	← sverrej left	07:03
drizzd	Xilon: do you have to use --receive-pack?	07:03
	→ Ademan_ joined	07:03
Xilon	yeah because it's in my home directory	07:04
drizzd	maybe the relative path doesn't work	07:04
Xilon	I can set it up in the config though	07:04
drizzd	probably	07:04
	use /home/<whatever>/programs/usr/bin/git-receive-pack	07:04
	jonesy: strace <cmd>	07:05
Mikachu	Xilon: did you compile it yourself? did you give prefix both to make and to make install?	07:05
Xilon	I used ./configure	07:05
drizzd	so <cmd> would be whatever you did before that failed	07:05
jonesy	wtf, I think I found my problem.	07:06
	← phpcurious left	07:06
jonesy	the command I was running. git clone ssh://<server>/git/<repo>.git <folder>.dev/	07:06
	not sure why but I removed the trailing / on my local folder name <folder>.dev	07:07
	and now it's working	07:07
drizzd	ah, that makes sense	07:07
Xilon	it probably wanted to store in <folder>.dev/<repo>	07:07
drizzd	1.5.5.1 still used a script, which would strip the trailing slash	07:07
Xilon	at least that's how mv and similar utilities work	07:08
jonesy	I can't believe that is what has been causing the problem	07:08
Xilon	I think...	07:08
jonesy	it's working as it shoudl now	07:09
	← johnw left	07:14
	→ johnw joined	07:15
	← digitalhobbit left	07:16
Xilon	Still doesn't work	07:16
	oh well, I'll just use 1.5.5.1 on the remote, it's not like I'll be using it for long	07:16
	→ sverrej joined	07:16
Xilon	Does anyone have a hook to automatically generate doxygen documentation on receive?	07:20
	I'm wondering if it's a good idea to have a normal repo, as opposed to a bare repo, and push to it to generate documentation	07:21
drizzd	You could try running 'git unpack-objects' on the remote	07:21
	that seems to be failing	07:21
cehteh	it wont be a good idea but should be doable	07:21
	maybe you better trigger some checkout in a dedicated repo (which refers to the bare one via alternates) and build doxygen there, less fragile	07:22
Xilon	that is nicer	07:23
cehteh	anyways .. building website on push: http://lumiera.org/.git/hooks/post-receive	07:23
	thats also such a non-bare push repo setup	07:24
	ymmw ..	07:24
	mhm http://lumiera.org/.git/website/ .. interesting recursion, i didnt noticed it yet :)	07:26
	→ tango_ joined	07:26
	→ mako joined	07:26
	stouset → stouset\|away	07:27
	← jonesy left	07:28
	→ mcella joined	07:29
	← yango left	07:29
	→ yango joined	07:33
	→ BabelO joined	07:38
	→ anttih_ joined	07:38
	→ zerok joined	07:40
	→ hurikhan\|Work joined	07:43
	→ daaku joined	07:47
	→ charon joined	07:50
	→ ferdy joined	07:50
	→ AhtiK joined	07:52
	ciaran_ → ciaran	07:52
	→ ijcd_ joined	07:53
	← LostInTheWorld left	07:54
	→ huyx joined	07:55
	← Hermano left	07:56
	→ ijcd__ joined	07:57
	← AhtiK left	07:58
	← thiago_home left	08:04
Xilon	umm git 1.5.5.1 doesn't have --bare for init?	08:06
	oh, git --bare init, not git init --bare...	08:06
	← ijcd left	08:09
	→ elmex joined	08:12
	← ijcd_ left	08:13
	→ ajonat joined	08:13
	→ Saied joined	08:15
	← davidfetter left	08:15
Saied	what is the port of git?	08:16
	i found it, 9418	08:17
	← Saied left	08:17
	→ ebel joined	08:18
	→ macgerhard joined	08:20
	→ sf joined	08:21
	dwmw2_gone → dwmw2	08:24
	Ademan_ → Ademan	08:25
	← peritus_ left	08:34
	← statim___ left	08:39
	→ statim joined	08:39
	→ Sho_ joined	08:44
	← BarryCarlyon left	08:45
	→ paddor joined	08:47
paddor	What's the difference between git-annotate and git-blame?	08:47
	→ schlort joined	08:48
	→ peritus_ joined	08:50
	→ SRabbelier joined	08:50
charon	paddor: output format. git-blame has an option to behave like annotate, see man git-blame	08:50
Gitbot	paddor: the git-blame manpage can be found at http://git.or.cz/man/git-blame	08:50
	← sf left	08:53
jones-	Hello. I have a remote called "origin" and in .git/config it's "fetch" is +refs/heads/:refs/remotes/origin/. How do I change that so that I can also push tags?	08:53
	zbob__ → dkr	08:54
	← dkr left	08:54
anttih_	can't I rebase a git-svn repo if I did a clone with --no-metadata?	08:55
	→ dkr joined	08:56
paddor	charon: yes, but except of the output format? is one of the two more usful for special uses?	08:57
drizzd	jones-: you can use git push --tags to push tags	08:57
	← paddor left	08:58
drizzd	you could configure remote.origin.push = refs/tags/:refs/tags/	08:58
jones-	drizzd: thanks	08:58
drizzd	but then you would not push any heads by default	08:58
jones-	drizzd: I'll push --tags when I make a new tag.	08:58
drizzd	jones-: you can also explicitly push the tag: git push origin <tag>	08:59
jones-	Sounds fair enough :)	08:59
	drizzd: thanks	08:59
charon	paddor: just the output format.	09:00
Xilon	does zsh completion still work?	09:00
	seems it uses the git-* binaries instead of commands	09:01
	→ sf joined	09:02
ft	Xilon: zsh completion supports both. (you may want the version from zsh's cvs Repo; see: <http://zshwiki.org/home/docs/wikifaq> question 12).	09:02
Gitbot	ft: I'm sorry, there's no such object: http://zshwiki.org/home/docs/wikifaq.	09:02
	→ paddor joined	09:03
	→ BarryCarlyon joined	09:03
	→ saimon joined	09:03
	saimon pasted http://guide.opendns.com/?url=pastie.textmate.org%2Fpastes	09:03
anttih_	I don't get it, how can someone keep a svn mirror without the metadata in the commits?	09:04
saimon	So much for automation, heres the real link: http://pastie.textmate.org/private/q4sx37zckw0xg7h3kcxaw	09:05
	← bentob0x left	09:07
	→ redondos_ joined	09:07
saimon	Can anyone tell me what problem git has with that diff? Why can't it just merge it and why is it presenting that as a conflict?	09:08
	I'm getting this conflict when trying to push to a remote topic branch. For some reason (beyond me) they local and remote branches are continuously out of sync.	09:09
	they/the	09:09
charon	saimon: are you sure there is no whitespace difference?	09:11
	spaces vs tab or trailing spaces	09:12
saimon	charon: none whatsoever	09:12
	I'm sure	09:12
	just double checked it	09:12
	→ Ademan_ joined	09:12
charon	maybe publish the repo, write down a few commands to reproduce and send mail to the list	09:13
saimon	I'm baffled as to why git every now and again loses sync between the local and remote topic branch. I practically always push to my remote branch after every commit.	09:13
Pieter	or upload the three files necessary for the merge?	09:13
	saimon: perhaps someone else rebases them sometimes?	09:14
	saimon: that might get you these kind of conflicts	09:14
Xilon	ft: got a link maybe? I can't find the actual file (http://zsh.cvs.sourceforge.net/zsh/zsh/Functions/Completion/?hideattic=0)	09:14
Pieter	saimon: look at the history with gitk --all	09:14
	← huyx left	09:14
saimon	Pieter: Nope, it's my own private branch. I'm the only one that touches it.	09:14
	← redondos_ left	09:14
Pieter	saimon: and you don't use rebase yourself?	09:15
saimon	Pieter: yes but between master and my local topic branch i.e. my workflow is like this:	09:15
ft	Xilon: http://zsh.cvs.sourceforge.net/checkout/zsh/zsh/Completion/Unix/Command/_git	09:15
Pieter	saimon: or, perhaps you did a commit --ammend and tried to push that?	09:16
saimon	git checkout topic1; code...; git commit -a && git push; rinse/repeat; then: git checkout master; git pull origin master; git merge topic1; git push; git checkout topic1; git rebase master	09:16
Xilon	ft: thanks	09:17
saimon	Pieter: no ammend	09:17
	can't look at gitk right now because tcl complains about a missing init.tcl	09:17
Pieter	git log --graph?	09:17
charon	saimon: that "git rebase master" looks a bit suspicious	09:18
Xilon	yay it works :)	09:18
saimon	Pieter: I'm 1.5.5.3 log doesn't seem to have --graph option	09:18
Pieter	saimon: after that git rebase, you'll have to use git push -f to update your remote branch	09:18
	→ _slacker_ joined	09:18
saimon	charon: How suspicious?	09:18
	-f?	09:18
	ahh	09:19
	cause once it's rebased, the remote ref is no longer an ancestor of the local one?	09:19
charon	why would you rebase if you already merge into master? that's mixing the merge and rebase workflows	09:19
saimon	charon: just got used to doing that in case my partner merged anything into master	09:20
charon	but it's asking for trouble	09:21
	after the merge, you'll have all changes from topic1 already in master	09:21
	the rebase generates a new set of commits with the same changes, but no history relationship whatsoever, not merged to master at any point	09:21
	then the next time around, you tell it to merge the _same_ changes again	09:21
	while i understand your frustration about an apparently nonsensical conflict, it's just a bad idea in the first place	09:22
	if the topic branch needs fixes, you can just commit them on the topic branch and merge to master again	09:22
saimon	so to pull in changes on master. I should just git merge master in my topic branch?	09:23
charon	yes	09:23
saimon	so when would I need to do a rebase? (I'm still a bit confused about what rebase does exactly=	09:24
	→ koke joined	09:24
charon	(i found Tv's video on topic branches very helpful, http://blip.tv/file/1114793/)	09:25
	→ drizzd_ joined	09:25
saimon	thanks I'll have a look see.	09:25
charon	in email workflows for example	09:25
saimon	thanks guys...Time to rehone my git workflow	09:25
charon	say, you wrote a patch series for git.git but in the meantime there have been changes to the master code that cause conflicts when you apply the series	09:26
	← pygi left	09:26
	← Ademan left	09:26
charon	since patches can't express merges, you'd have to rebase your series and resend it	09:26
	Ademan_ → Ademan	09:26
charon	there are other uses of course	09:26
	← macgerhard left	09:26
saimon	ok...I see...so in my worklfow, I'd basically only do a rebase if I came across a conflict when merging into master?	09:28
	→ levente joined	09:29
charon	if you have the power to do the merge to master yourself, you should do that instead of a rebase	09:29
	(i.e. checkout master; merge topic)	09:29
saimon	and just fix any conflicts?	09:30
	← levente left	09:30
saimon	so in general don't rebase?	09:30
	→ Kanniball joined	09:30
	→ levente joined	09:33
levente	hi, I need help on gitosis	09:34
[GarethAdams]	only rebase with branches that are never getting pulled by anyone else	09:34
levente	Does enyone have some time here?	09:34
charon	saimon: i hate absolute "don't"s, but yes, if a merge works just as well you should prefer that	09:34
	and never ever rebase anything you've published unless you've warned them three times in writing	09:34
saimon	charon: :)	09:34
	← sf left	09:35
saimon	[GarethAdams]: got it. thanks...	09:35
	ok i'ts now clear...	09:35
	watching tv's talk...	09:35
	→ Yuuhi joined	09:35
	← ajonat left	09:35
charon	the bit about topic branches starts at 32min iirc, though the rest is also interesting	09:36
	→ CodeWar joined	09:37
[GarethAdams]	I actually have `git pull` set up to do a rebase rather than a merge on most of our projects, since we all pull/push to github rather than from each other	09:37
levente	Can anyone help me setting up a repository with gitosis?	09:37
[GarethAdams]	it keeps the commit tree a lot more tame	09:38
Weaselweb	is there a way to download that video?	09:40
	← schlort left	09:40
CodeWar	is there an equivalent to hg's inotify?	09:41
	← redondos left	09:41
	→ redondos joined	09:41
	[GarethAdams] → GarethAdams	09:42
	← drizzd left	09:42
charon	Weaselweb: there's a link on eagain.net/blog	09:42
Weaselweb	charon: ah, ok. thanks	09:43
	← ph^ left	09:43
	→ parasti joined	09:44
	→ ph^ joined	09:44
	← ph^ left	09:44
	charon wonders if he should write a topic on "my upstream has been rebased" recovery for git-rebase.txt	09:45
	→ ph^ joined	09:45
charon	it might serve as a good warning. the current one is very tame and buried in "notes" four screenfuls down	09:45
	← ph^ left	09:49
	→ lionr joined	09:50
	→ sf joined	09:50
lionr	hey all -- hmm, if you've gotten into a state with submodules where you updated the superproject before the submodule (which means you can't now clone the submodule) - how do you best fix that?	09:50
	→ ph^ joined	09:51
	← daaku left	09:53
	← CodeWar left	09:54
	→ Kbyte joined	09:54
	← johnw left	09:55
	← ph^ left	09:56
	→ markkalderon joined	09:57
	→ ph^ joined	09:57
	← aiur left	09:58
	← tango_ left	09:59
	← ph^ left	10:03
	→ ph^ joined	10:04
	→ jonas_k9 joined	10:05
	← jtoy left	10:07
	← mithro left	10:07
	← Xilon left	10:15
	dwmw2 → dwmw2_lunch	10:17
	← markkalderon left	10:19
	→ mithro joined	10:29
	→ ereslibre joined	10:30
	→ macgerhard joined	10:32
macgerhard	hi guys, how can I rewrite my repository so only a certain range of commits are kept, everything else is discarded?	10:32
	→ botanicus joined	10:32
	← morphir left	10:33
sander_	When I and the repository have changed a file.. how can I choose to use the whole file of my or the repositories version?	10:39
	→ tango_ joined	10:40
	→ cilly joined	10:40
Mikachu	sander_: you mean when you merge?	10:40
sander_	Mikachu, yep	10:40
Mikachu	you can say git show :1:path/to/file > path/to/file, the 1 can be 2 or 3 too	10:41
	i can never remember which is yours and which is the other branch's though :)	10:41
	(and maybe there's an easier way (i know there is git checkout --ours/--theirs in #next))	10:41
sander_	fatal: ambiguous argument ':3': unknown revision or path not in the working tree.	10:42
	Use '--' to separate paths from revisions	10:42
	← levente left	10:43
Mikachu	it works here	10:44
	→ kukks joined	10:44
sander_	git show :3:config ?	10:44
Mikachu	yes	10:44
	← sf left	10:45
Mikachu	does git ls-files -u list the file?	10:45
macgerhard	Mikachu: any ideas on how to rewrite a repo and only keep a range of commits?	10:45
sander_	Mikachu, no.	10:45
Mikachu	sander_: then you're not in a merge conflict	10:45
	→ Lash\| joined	10:48
	→ jlemieux______ joined	10:49
	jlemieux______ → jlemieux	10:49
	→ clairvy joined	10:50
	→ advany joined	10:51
	→ sf_ joined	10:53
	← sf_ left	10:54
	→ bentob0x joined	10:55
macgerhard	when creating a patch with git format-patch, how can I make sure that the patch will include commit comments?	10:56
parasti	by looking at the generated patch? it includes the commit message by default	10:57
sander_	Mikachu, thanks.. I figured it out.	11:00
	How can I set an default umask when pushing to a rep?	11:00
	I'm almost finnished with setting up a bare repository.. with sticky group bit.. just have an issue with users writing files which aint group writeable.	11:01
Mikachu	did you use the shared switch?	11:02
sander_	No :-p	11:02
Mikachu	--shared=foo	11:02
sander_	And foo should be?	11:02
bremner	!man git-init	11:03
Gitbot	bremner: the git-init manpage can be found at http://git.or.cz/man/git-init	11:03
macgerhard	parasti: I ran git format-patch -8 and it doesn't include the commit messages. Subject includes only the first line from the commit message.	11:04
parasti	macgerhard: well, I doubt anyone would like the subject to contain the entire commit message	11:06
	← clairvy__ left	11:07
macgerhard	parasti: ok, I'll tell you what I'm trying to achieve, maybe you can offer a better suggestion because this is becoming so frustrating!	11:07
	I have a bunch of commits that refer to a particular directory which I want to move in it's own repo. I have tried using git filter-branch, but because that directory has changed name, it didn't seem to work very well. What am I missing?	11:08
	→ sf_ joined	11:09
	← FunkeeMonk left	11:10
	← Soliton left	11:10
	→ Soliton joined	11:11
	← mako left	11:11
parasti	macgerhard: are you having any trouble with the format-patch route? it seems like a good way to do what you want to do	11:13
macgerhard	parasti: yep. First of all, it doesn't include the full commit messages. Then I got to the patch where the directory has renamed, I'm getting this error: http://pastie.org/264436	11:16
parasti	macgerhard: it does include the full commit messages -- first line in the subject, the rest in the body of the patch	11:18
macgerhard	parasti: hmm, I can't see them in the body. Let's say that it does include them in the body, when I run git apply patch_file, I still have to manually enter the commit message & then commit	11:19
parasti	macgerhard: you don't usually use git-apply on git-format-patch output	11:20
	macgerhard: use git-am instead	11:20
	macgerhard: as for the renames, you need to tell format-patch to detect them	11:20
	macgerhard: and you probably need --binary too	11:20
macgerhard	let me rtfm for git am	11:21
	→ whee joined	11:24
	→ scroogey joined	11:25
scroogey	how can i checkout a branch w/o the whole history?	11:25
bremner	scroogey: how do you mean checkout? maybe git-archive	11:25
scroogey	yeah just the latest revision w/o the history backlog	11:26
bremner	!man git-archive	11:26
Gitbot	bremner: the git-archive manpage can be found at http://git.or.cz/man/git-archive	11:26
scroogey	aha oki :)	11:26
	thought so already	11:26
bremner	scroogey: watch the trailing / on --prefix	11:26
	← whee left	11:27
	← cedricv left	11:27
	→ whee joined	11:28
	→ iulian joined	11:29
	→ cedricv joined	11:32
	← botanicus left	11:34
anttih_	how do I mirror a SVN repo without the metadata in the commits? --no-metadata doesn't seem to work.	11:34
	← jlemieux left	11:35
macgerhard	parasti: I ran this git format-patch -M --binary <sha> & then git am, same errors, am doesn't handle the renames very nicely	11:37
sander_	I get: fatal: Entry 'config/database.yml' not uptodate. Cannot merge..... The problem is that we have 2 versions of config/database.yml ..how should we solve that?	11:42
	..I do "git pull" and get the message abow.. that file is added to .gitignore	11:43
	How does .gitignore works?	11:44
	→ GNUix joined	11:44
GNUix	how do I stop zsh from trying to autocorrect git config to git .config	11:46
	→ pygi joined	11:46
	← JenkingWinlker left	11:47
	← ph^ left	11:47
	→ ph^ joined	11:49
Soliton	sander_: sounds like that file is checked in.. .gitignore is there to ignore uncommitted files.	11:53
macgerhard	how do I get format-patch to handle renames? I've used -M but that doesn't seem to cut it. The error that I'm getting is http://pastie.org/264436	11:53
	← deavid left	11:55
	← sf_ left	11:58
	→ yofel__ joined	11:58
parasti	macgerhard: does using -M -M (double -M) make any difference?	11:59
	← SRabbelier left	11:59
	→ SRabbelier1 joined	11:59
	→ cbreak joined	12:00
	SRabbelier1 → SRabbelier	12:00
macgerhard	parasti: nope :(. But I have the .dotest directory. Can I fix the patch somehow?	12:01
	→ boink__ joined	12:01
	← yofel_ left	12:02
	→ Weasel[DK] joined	12:02
	→ cannonball joined	12:02
	← grahal left	12:03
	→ reuss joined	12:04
	← bentob0x left	12:05
reuss	doesn't split work with git add -i anymore? ... it's not in the "short-options" Stage this hunk [y/n/a/d/j/J/e/?]? - but in the long descriptions it's still there: s - split the current hunk into smaller hunks	12:05
Mikachu	reuss: it works, but it needs enough context lines	12:08
macgerhard	reuss: if you don't have split, it means that that hunk cannot be split further	12:08
charon	reuss: 'split' only works if there are context lines in the middle of the hunk	12:08
	← siprbaum left	12:08
charon	damn, Mikachu wins	12:08
Mikachu	but you can use e instead!	12:08
	← paddor left	12:08
reuss	hrm, okay - it was two lines seperated by a blank line, and then another line - just assumed it could split it on the blank line ...	12:09
macgerhard	Mikachu: care to jump in on my problem with renames & git am?	12:09
Mikachu	macgerhard: not particularly	12:09
macgerhard	Mikachu: is it a hairy one?	12:09
charon	reuss: a blank context line, or a blank add/remove line?	12:10
	(i'd consider it a bug in the former case, it should be able to split on any kind of context)	12:10
tango_	question: should I just hope that jakub gets back to my refmark patch, or should I resend to the ml?	12:10
	← ph^ left	12:11
	→ joern_ joined	12:13
	← raz left	12:13
	boink__ → raz	12:13
reuss	charon: uhm, just a blank line - i think	12:14
	→ lorandi joined	12:15
	→ paltman joined	12:15
	→ siprbaum joined	12:16
charon	reuss: works for me...	12:17
	i.e. it correctly splits at blank context lines, and you can't blame it for not treating added (+) blank lines special	12:17
	← joern_ left	12:20
	← redondos left	12:21
	→ redondos joined	12:21
	dwmw2_lunch → dwmw2	12:29
	→ ph^ joined	12:29
	→ tusho joined	12:32
tusho	are there online docs for Git.pm anywhere?	12:32
	→ cmarcelo joined	12:32
sztanphet	only google knows	12:32
	← marj left	12:32
	→ up_the_irons2 joined	12:32
tusho	sztanphet: I did try, you know.	12:33
	Thanks for the sarcasm, though.	12:33
charon	'perldoc Git' ?	12:34
macgerhard	Mikachu: cheers for making me think a bit harder on my issue. I've sorted it by applying the changes manually, all other patches worked afterwards :). Cheers!	12:36
	→ MK_FG joined	12:38
	← MK_FG left	12:38
	← Judofyr left	12:40
	→ Judofyr_ joined	12:40
	→ mlins joined	12:41
	→ EmilMedve joined	12:42
	← paltman left	12:44
	→ jlemieux______ joined	12:44
	jlemieux______ → jlemieux	12:44
	→ priidu_ joined	12:45
	→ deavid joined	12:46
	← zerok left	12:46
	→ DrewR joined	12:47
	← jonas_k9 left	12:47
jones-	Hello. I have a new local git repo: I did git-init, did some work, then did git-add and git-commit and now everything is fine. Now I want to create a directory in a svn repo and use git-svn to sync. How can I proceed?	12:48
	→ mbroeker joined	12:48
jones-	Maybe I'll just edit my .git/config	12:49
	→ jptix joined	12:50
	← priidu left	12:50
	← tusho left	12:55
jones-	Hmm. What does this mean?	12:56
	$ git-svn dcommit	12:56
	Unable to determine upstream SVN information from HEAD history.	12:56
	Perhaps the repository is empty. at /usr/bin/git-svn line 424.	12:56
	→ mw\|out joined	12:56
	mw\|out → mw	12:56
parasti	jones-: you should read an introduction to git-svn or at least man git-svn	12:58
Gitbot	jones-: the git-svn manpage can be found at http://git.or.cz/man/git-svn	12:58
jones-	I've been looking at that.	12:58
	It's not clear to me how to behave if you first have a git repo and then you want to publish it to svn.	12:58
	I have other repos that I were ported from svn to git, but can't do the other way around :)	12:58
parasti	you never want to publish it to svn, that's all :P	13:00
jones-	parasti: heh, I wish.	13:00
Mikachu	you have to first somehow create the svn repo, then clone that with git-svn, then you can commit and svn dcommit	13:01
	→ bobmcw joined	13:01
jones-	Mikachu: ok thanks.	13:01
gebi	jones-: git-svn clone the svn repo and rebase your branch on top of the git-svn branch, then git-svn dcommit should work just fine	13:10
	→ JohnFlux joined	13:10
	→ chrisb joined	13:10
jones-	gebi: good idea, I'll try	13:10
JohnFlux	I've committed 3 times in a row, because I kept making stupid mistakes	13:10
	is it possible to now merge these three commits, before I push upstream?	13:10
	so upstream see it as just a single commit	13:11
broonie	JohnFlux: rebase -i or merge --squash	13:12
gebi	something like: git reset HEAD^^; git commit --amend	13:12
	or reset HEAD~3; git commit	13:13
	→ grahal joined	13:15
JohnFlux	so, which is the best way ? :-)	13:15
	merge --squash sounds like the friendliest way	13:15
mbroeker	JohnFlux, and check svn ls http://your.repo.com/projects/project/trunk; whether your files are in the svnroot or the trunk folder.	13:16
JohnFlux	mbroeker: what's svn got to do with it?	13:16
	are you in the right channel? :)	13:16
mbroeker	ah mr jones ;)	13:17
	JohnFlux, sorry, wrong topic	13:17
JohnFlux	hehe :)	13:17
	stouset\|away → stouset	13:17
	← G_SabinoMullane left	13:17
	← GNUix left	13:17
JohnFlux	hmm, I don't understand the help for merge --squash	13:17
	git rebase -i doesn't give any output.. I don't know if it's changed anything	13:18
	git log still shows my 3 commits, so I guess not	13:18
ehamberg	git rebase -i HEAD~3	13:18
JohnFlux	ehamberg: thanks	13:18
	awesome :)	13:19
ehamberg	you probably want to "squash" the two last commits	13:19
JohnFlux	as well as rebase?	13:19
	the rebase has 3 lines like "rebase xxxxxx comment.."	13:20
	→ elux joined	13:20
JohnFlux	when it's prompting for the commit message	13:20
	→ sergiodj joined	13:20
elux	how can i fetch/update a submodule?	13:20
JohnFlux	which all looks good. but the commit message also says:	13:20
	# If you remove a line here THAT COMMIT WILL BE LOST.	13:20
	does that mean I can't change the commit message?	13:20
elux	when i git pull in the directory it tells me in the status the files have changed	13:20
ehamberg	JohnFlux: don't you have three lines with "pick [sha]"?	13:21
JohnFlux	I'd like to change the commit message to something else...	13:21
	ehamberg: yep I have that	13:21
ehamberg	JohnFlux: you will be prompted afterwards	13:21
	change "pick" to "squash" for the two last commits	13:21
	they will be "squashed" into the first. when you save, you will be prompted for a commit message	13:21
JohnFlux	'use commit' means 'use commit message' ?	13:21
ehamberg	don't care about the message for now.	13:22
	your $EDITOR will be launched again when you save and exit.	13:22
JohnFlux	ah okay, so I'll make the first one pick, and the last 2 squash	13:22
ehamberg	yes	13:22
JohnFlux	ah I see	13:22
	now I can change the comment :)	13:22
ehamberg	good :-)	13:23
JohnFlux	that's so awesome	13:23
	once I get used to this :)	13:23
ehamberg	yeah, i use it all the time :)	13:23
	you just need to do it once to see how it works.	13:23
JohnFlux	so, if I made all three pick instead, what would the difference be?	13:23
ehamberg	nothing :)	13:23
	→ vbabiy joined	13:24
	→ G_SabinoMullane joined	13:25
saimon	How is it possible to amend older commit messages (i.e. not the lastest one) without using git rebase -i?	13:29
	(and without having to edit the code)	13:29
Pieter	you can use git-filter-branch	13:29
saimon	Pieter: I was sure git could do it :) thanks	13:30
	→ sf_ joined	13:30
Pieter	but or use something like git checkout commit-to-be-changed && git commit --amend && git rebase --onto HEAD commit-to-be-changed master or so	13:31
saimon	Pieter: I want to avoid using rebase as this topic branch tracks a remote topic branch. WIll this come up with problems when pushing to the remote branch?	13:33
	← mithro left	13:33
Pieter	saimon: yes. but using filter-branch will cause the same problems	13:33
	either way you're rewriting history	13:33
	→ markkalderon joined	13:34
saimon	I see....i.e. isn't there a low level command that takes a commit object and allows you to edit it's message directly without modifying history?	13:34
Pieter	no	13:35
	the message is part of the history, you can't ever change it	13:35
saimon	ok...	13:35
	thanks	13:35
	→ dotsintacks_ joined	13:35
	← j_engelh left	13:35
	← EmilMedve left	13:38
	→ EmilMedve joined	13:38
	← ph^ left	13:38
	← leitgebj left	13:39
	→ ph^ joined	13:39
	← dotsintacks left	13:41
	← scroogey left	13:43
	→ joern- joined	13:43
	→ warlock_mza joined	13:44
	← Judofyr_ left	13:46
	→ Judofyr joined	13:47
	→ mithro joined	13:51
	← Judofyr left	13:51
	← reuss left	13:53
	← mbroeker left	13:53
	→ aroben joined	13:57
	→ Catfish joined	13:57
	→ qrush joined	13:57
	← irotas left	13:58
	→ paltman joined	14:02
	← redondos left	14:03
	→ redondos joined	14:03
	→ tjafk1 joined	14:03
	→ loincloth joined	14:03
	← kanru left	14:04
	→ irotas joined	14:04
	← jds left	14:04
	← stouset left	14:04
	→ kanru joined	14:06
	→ spearce joined	14:08
	← chrisb left	14:08
	← warlock_mza left	14:08
	← tjafk2 left	14:11
pasky	how can i make git fsck report only real errors?	14:14
parasti	git fsck \| grep 'real error'?	14:16
	← sf_ left	14:20
	→ Zarin joined	14:21
	cbrake_away → cbrake	14:21
	→ leitgebj joined	14:22
	→ gberdyshev joined	14:22
	→ chrisb joined	14:25
	← lionr left	14:25
	← Sho_ left	14:26
	→ up_the_irons3 joined	14:27
	← DrewR left	14:29
	gberdyshev → gberdyshev_	14:29
	→ brett_h joined	14:29
	gberdyshev_ → gberdyshev	14:29
brett_h	anyone know anything about "PROPFIND request failed" when doing a git-svn clone? http://pastebin.com/m2fb06ec0	14:30
	← EmilMedve left	14:30
	← loincloth left	14:31
	→ warlock_mza joined	14:31
	→ loincloth joined	14:31
	→ rmh3093 joined	14:33
	← priidu_ left	14:33
	← warlock_mza left	14:33
	← zumbrunn left	14:34
	→ EmilMedve joined	14:35
	← loincloth left	14:35
	→ loincloth joined	14:36
	→ alley_cat joined	14:36
	→ mheld joined	14:37
	← loincloth left	14:37
	→ loincloth joined	14:37
	← loincloth left	14:38
	→ loincloth joined	14:38
	→ priidu joined	14:38
	→ DrFrasierCrane joined	14:38
	→ loinclot_ joined	14:39
	← ysis left	14:40
	← loincloth left	14:40
	→ stouset joined	14:40
	loinclot_ → loincloth	14:40
	→ roderyk joined	14:40
	← roderyk left	14:41
mheld	anybody familiar with the textmate bundle for git?	14:41
	← DrFrasierCrane left	14:42
	← loincloth left	14:42
	← pygi left	14:42
	→ loincloth joined	14:43
	← loincloth left	14:43
	← up_the_irons2 left	14:43
	→ aiur joined	14:44
	← up_the_irons3 left	14:45
	→ zerok joined	14:47
	→ kinzunikonzikonz joined	14:49
kinzunikonzikonz	hey peoplle	14:49
	→ loincloth joined	14:51
JohnFlux	ehamberg: I have another commit that I want to merge with my previous commit	14:51
	→ warlock_mza joined	14:51
JohnFlux	ehamberg: i haven't yet committed	14:51
	is there a better way than doing git-commit -a; git-rebase -i ..	14:51
	← peritus_ left	14:52
cbreak	you mean like git commit -a --amend?	14:52
	← loincloth left	14:52
	→ loincloth joined	14:53
JohnFlux	cbreak: awesome	14:54
	cbreak: thanks	14:54
	← loincloth left	14:55
	→ rdavila joined	14:56
mheld	anybody here do web development work using git?	14:56
	→ charon_ joined	14:57
bmalee	mheld: sort of; I use git-svn.	14:57
mheld	ah	14:58
	→ aCiD2 joined	14:58
	← charon left	14:58
	charon_ → charon	14:58
JohnFlux	bmalee: how are you finding it?	14:58
	← cedricv left	14:58
	→ johs joined	14:58
bmalee	JohnFlux: miles better than svn, for a start.	14:59
	← lea_w left	14:59
	→ loincloth joined	15:00
	→ Ryback_ joined	15:01
	← aiur left	15:02
bmalee	JohnFlux: I've had no major problems, and e.g. branching has been very helpful.	15:02
	→ aiur joined	15:02
	→ schlort joined	15:03
	← kinzunikonzikonz left	15:03
	→ akahn joined	15:05
akahn	If my co-worker cloned a git repo from me the other day, how can he get the changes that I've made since then?	15:05
bmalee	akahn: 'git pull' should default to the repo it was cloned from.	15:05
akahn	bmalee: okay, i'll give that a shot, thanks. does that give him a preview of what's going to change or does it just do it?	15:06
	→ macgerhard_ joined	15:06
	BarryCarlyon → Perl_	15:06
	Perl_ → BarryCarlyon	15:06
	→ davidfetter joined	15:06
johs	It applies your changes to his working directory also.	15:06
	If he just wants to get the changes without doing that, he can do "git fetch".	15:06
mheld	what workflow would you use for web design/development?	15:07
	with a live site and a dev site to test the upcoming live stuff	15:07
johs	That is, it also merges.	15:08
	← cebewee left	15:08
aCiD2	I just did "git merge some-branch" then realised I didn't actually want to merge this branch, what's the corerct way to undo - git reset HEAD?	15:08
akahn	johs: ok, cool.	15:09
	← macgerhard_ left	15:10
	← macgerhard left	15:10
	→ macgerhard_ joined	15:10
	← warlock_mza left	15:12
	← mithro left	15:12
	→ mithro_ joined	15:12
	→ sf_ joined	15:14
aCiD2	Ah, got it - git-reset --hard ORIG_HEAD	15:14
	← elux left	15:18
	→ eternaleye joined	15:18
	→ juanjoc joined	15:19
	→ kinzunikonzikonz joined	15:22
	→ zumbrunn joined	15:22
	← kinzunikonzikonz left	15:22
	→ cyclist joined	15:23
johs	I'm curious about the "git am"-feature. I have a repository B cloned from A. I do some changes in B, make a patch with format-patch. Then "git am" is used to apply it to A. So far so good, but if I later pull from A in B, the patch I sent is applied to B again, along with a merge commit.	15:23
	Is there some clever way to avoid this?	15:23
	→ bdiego joined	15:24
aCiD2	I assume A can't pull from B?	15:25
johs	That's the idea.	15:25
	← kanru left	15:26
	→ jackdempsey joined	15:27
johs	Of course, I can just toss the commit from B after sending the patch...	15:27
	Is that how it's done?	15:27
	Well, it works, anyway.	15:28
	That is, "git reset" in B, and then pull.	15:28
Tv	johs: 1) use bundles 2) use rebasing in B	15:28
johs	Oh, cool.	15:29
	Thanks.	15:29
cyclist	which command can i use to find the object ids of the history of a file? and can they be used as input to git-pack-objects?	15:32
	← redondos left	15:32
	→ redondos joined	15:32
cyclist	actually i want to do the latter in order to understand the pack format	15:32
cbreak	can you not use git log to get all commits that affect a file?	15:33
	→ zumbrunn_ joined	15:33
cyclist	doh. right	15:33
cbreak	and then use git cat-file to get the tree from the commit	15:34
spearce	yea, but then you are really taking about doing something more like `git rev-list --all -- $path \| while read c; do git rev-parse $c:$path;done`	15:34
cbreak	and again to get the file hash out of the tree	15:34
cyclist	is the resulting list suitable for feeding into git-pack-objects?	15:35
johs	Tv: Do people use "git bundle" to submit patches, by the way?	15:35
	Tv: Or is it only intended for, as the man-page says, things like sneakernets.	15:36
	← joern- left	15:36
Tv	johs: it's kinda rare to actually need it	15:36
	johs: what's the difference between submitting patches and sneakernet, really..	15:36
	it's like a git push you can save as a file	15:36
	nothing more, nothing less	15:36
spearce	cyclist: yes	15:36
Tv	in the real world, most projects that would actually need to use bundles for cooperation can just set up a pushable repo	15:37
johs	Yes, and that is no problem for me either. I was just curious about how people actually manage with patches by email. :)	15:37
spearce	Tv: i'm actually working on a web based code review system that uses bundles as the means of injecting new series into the review queue.	15:38
johs	That is, how the patch-submitter would normally work with his repository to avoid the logspam that ensues from pulling commits you've submitted.	15:38
	→ crab joined	15:38
crab	hi.	15:38
	→ neoneurone joined	15:38
johs	Just curious, is all. :)	15:39
crab	what's the best option for non-ssh push-to-central-repo support?	15:39
johs	Thanks for your tip.	15:39
	← zumbrunn left	15:39
offby1	crab: if "central" and you are on the same box, then "git push /path/to/repository.git" works fine :)	15:39
charon	johs: i actually just rewind and rebuild git.git after building cruft for a while	15:39
crab	offby1: gee, thanks.	15:40
charon	that is, reset --hard origin/master and merge the two or so "eternal" local branches	15:40
offby1	crab: otherwise, I dunno. Maybe email or ftp a bundle to the central box, then (somehow) log in and "pull" from the bundle ... ?	15:40
	← neoneurone left	15:40
	← sverrej left	15:41
johs	charon: Yeah, that works, I guess.	15:41
	→ kinzunikonzikonz joined	15:41
johs	charon: Thanks to you as well.	15:41
	→ joern_ joined	15:41
	← kinzunikonzikonz left	15:42
offby1	crab: don't give up because nobody else answered. I too am curious to hear a better answer than mine	15:43
	← Lash\| left	15:43
	→ kanru joined	15:43
	→ kinzunikonzikonz joined	15:43
	→ macgerhard joined	15:44
cyclist	i have just the .git directory and i'm doing a "git checkout master" and it prints the directories and doesn't do anything else! wtf	15:45
	→ mbroeker joined	15:45
	← mlins left	15:45
	→ polpak joined	15:45
	→ digitalhobbit joined	15:45
	← sgrimm left	15:46
doener	cyclist: the list tells you that you have changes to those files. In this case: you delete them all	15:46
	cyclist: "git reset --hard" now to throw those local changes away	15:46
	→ quizme joined	15:47
cyclist	no no. I just copied the .git directory from my project to another location in order to experiment. now i want to checkout	15:47
quizme	hello, what is the prefix option for? should i set it to /usr or /usr/local ? I'm confused.	15:47
cbreak	I left it at the default which seemed to be /usr/local	15:47
Soliton	cyclist: which is the same as deleting your working tree.	15:48
Weaselweb	^^	15:48
cyclist	aw right	15:48
mbroeker	cbreak, install it in /usr/local and it won't interfere with your distros package manager and it's easier for you to remove or upgrade it later.	15:49
cbreak	mac os x has no package manager	15:49
quizme	cbreak: so to install: make configure; ./configure; make all doc; make install install-doc install-info;	15:50
cbreak	but I agree, local is the best place	15:50
redondos	what about fink?	15:50
cbreak	I don't like fink	15:50
quizme	why is /usr/local the best place ?	15:50
	← kinzunikonzikonz left	15:50
doener	quizme: if you can live without info pages, use the quick-install-doc target which installs the prebuilt man pages	15:51
quizme	mbroeker: oh i see	15:51
doener	quizme: building the docs takes a few ages	15:51
offby1	nod	15:51
doener	... assuming you install from git.git, not sure if the tarballs have prebuilt docs	15:51
	→ kinzunikonzikonz joined	15:52
quizme	doener: so, make install quick-install-doc; ?	15:52
doener	yep	15:52
quizme	ok doener thanks	15:52
doener	and no "make doc" either	15:52
cbreak	I use /usr/local because I know all in it is installed by me, and not some system related stuff	15:52
quizme	doner: your name rhymes with boner btw	15:52
	cbreak: ok got it, thanks	15:53
	thanks everybody	15:53
	→ by007 joined	15:54
quizme	i think i understand RSA encryption now...	15:54
	← qrush left	15:54
	← saimon left	15:55
quizme	i don't need to do the quick-install-doc do i (if i use the internet for reading docs) ?	15:55
	← andres left	15:55
quizme	make install worked, but make quick-install-doc barfed on me	15:56
	does anybody have a better article than this one: http://toolmantim.com/article/2007/12/5/setting_up_a_new_remote_git_repository ?	15:59
	for setting up a remote git repository ?	15:59
	← joern_ left	16:00
	← ben_h left	16:00
quizme	what's the /var directory for? Why put your remote git repositories there ?	16:00
	← macgerhard_ left	16:00
	← advany left	16:01
thiago	quizme: /var is for data	16:01
cyclist	spearce: thanks. your git-rev-list\|while.. git-rev-parse thing is just what I needed	16:01
	← bobmcw left	16:01
thiago	lots of servers are in /var/www or /var/lib/www or even /var/lib/apache	16:01
	though recently many distros started using /srv for that	16:01
	← kinzunikonzikonz left	16:01
quizme	thiago: ohh	16:02
	thiago: var = data . . . . . . that's deep	16:02
spearce	cyclist: technically its not quite complete as pack-objects takes an optional file name after each object id. the file name is a hint to help it hash objects for the same file into the same bucket so it can consider them for delta compression. but the path name is irrelevant if you are only submitting objects for the same file.	16:02
	← Kanniball left	16:02
	→ qrush joined	16:02
thiago	quizme: I used to have /mnt/junk	16:02
quizme	thiago: then you saw the light of /var	16:03
cyclist	ok i'll remember that	16:04
	→ Fullmoon joined	16:04
	sztanphet → sztanpet	16:05
	→ bobesponja joined	16:06
	→ andres joined	16:06
	andres → zz_andres	16:06
	zz_andres → andres	16:06
	← JohnFlux left	16:07
	← macgerhard left	16:07
	← bobesponja left	16:07
	→ macgerhard joined	16:08
	→ lchavez joined	16:08
	← chrisb left	16:09
	→ lea_w joined	16:09
lchavez	Hello .. i would like ask if it is posible to use git on a server .. using https or ssl ... but it should be definitely a secure way ..	16:10
	oh i misstype me .. sorry ..	16:11
spearce	cyclist: what are you trying to learn by poking pack-objects? have you read technical/pack-format.txt ?	16:12
cyclist	yes. I want to see the deltas in raw text form, for one	16:12
spearce	well, the deltas are not stored in text format. ;-)	16:14
	→ Beket joined	16:14
	→ LiamH joined	16:14
cyclist	i know that they are compressed. but what is compressed? diffs?	16:14
	← sf_ left	16:14
spearce	a binary delta format is compressed with zlib.	16:15
	the binary delta format is based upon two commands: "insert" with an sequence of bytes to insert at the current position, and "copy" with a source offset and a length to copy from the base object.	16:15
cyclist	i imagined so. something like "very compressed diffs"	16:15
spearce	the delta format works on any file format, but depends on being able to find blocks of common bytes to copy from the base.	16:16
	gberdyshev → gberdyshev_	16:17
	← rmh3093 left	16:18
cyclist	Yes obviously. In a patch, the "-" lines and the " " lines are completely useless. The "-" lines could be replaced by a "remove N lines" command, and have only the "+" lines.	16:18
spearce	yea, its even more compressed than a diff because there is no context and no pre-image data in the delta. the "-" sort of lines in a diff are handled in a git delta by simply not copying that region from the base.	16:18
	→ joern_ joined	16:20
cyclist	this is exactly what i wanted to learn by poking objects :)	16:20
	← BarryCarlyon left	16:20
	← Zarin left	16:20
cyclist	I am still going to poke a few to see it in action	16:21
	← sergiodj left	16:23
spearce	http://www.gelato.unsw.edu.au/archives/git/0608/25286.html	16:24
	← tango_ left	16:24
	← crab left	16:24
spearce	there's a newer version of that but i can't dig up the url. its mostly correct.	16:24
quizme	how do you specifiy a different username on a remote git server when doing 'git remote add' ?	16:26
spearce	quizme: scp style syntax: "git remote add foo quizme@server:/path/to/repo.git"	16:27
	← cbrake left	16:27
	→ jerbear joined	16:27
quizme	spearce: thanks, i'll try that.	16:28
	→ macgerhard_ joined	16:28
akahn	Is there a way to let my coworker pull from my repo without him having to use my login/password and without making a user for him on my machine?	16:29
blowery	akahn: you should be able to use the git protocol. it doesn't have any auth afaik.	16:29
	← macgerhard_ left	16:29
	← litage left	16:29
blowery	akahn: man git-daemon	16:30
Gitbot	akahn: the git-daemon manpage can be found at http://git.or.cz/man/git-daemon	16:30
	← gberdyshev_ left	16:30
	→ JanK joined	16:30
	→ schacon joined	16:31
JanK	is there a clean way to include a svn tracked framework in my git project?	16:31
quizme	spearce: what's the syntax if the protocol is specified as ssh://myserver.com:/ ... i tried fred@ssh://myserver.com, but that didin't work.	16:32
	→ amigaoxxxx joined	16:32
spearce	quizme: then its like http as in ssh://user@myserver.com	16:32
JanK	i'd use svn:externals if i'd use svn for the project	16:32
	← joern_ left	16:33
akahn	blowery: i thought i was using the git protocol. on his machine we ran git pull dev:/Users/akahn/emtest master, but that didn't work, if we put in akahn@dev instead, and i entered my password that worked.	16:33
	and i had started git-daemon	16:33
blowery	akahn: hrm. dunno? i don't use it personally.	16:33
quizme	spearce: thanks, i'll try again :)	16:33
akahn	ok. i think the --export-all option may do what i need	16:34
	→ kumbayo joined	16:34
	→ ajonat joined	16:34
	← by007 left	16:35
quizme	in "push origin master" is "origin" a reserved word in git? or instead can you call "orgin" "home_base" or "the_womb" ?	16:36
spearce	its just a default name for a remote. it can be any valid URL or remote name in .git/config.	16:36
quizme	ok thanks.	16:37
	i think i'll call it "mother_ship" then.	16:37
	→ joern_ joined	16:38
	← mithro_ left	16:38
	→ mithro_ joined	16:39
	→ shodges joined	16:40
	← joern_ left	16:40
	→ abenea joined	16:40
	← cyclist left	16:41
quizme	is there some insane compression going on with git? cuz it's unbelievably fast	16:41
	← priidu left	16:41
	→ steve_j joined	16:41
Tv	quizme: there's a built-in time machine	16:41
	quizme: your operations are really being run in a quantum computer on the 23rd century	16:41
	← steve_j left	16:41
quizme	tv: sweet.	16:42
Tv	the git basic design is very smart	16:42
	i am a linus fanboy	16:42
	← macgerhard left	16:42
Tv	(though packing many-files or big-files repos is a pain)	16:42
	→ rmh3093 joined	16:42
	→ priidu joined	16:43
	← lea_w left	16:44
quizme	i'm pretty impressed	16:44
	it's oodles of times faster than svn	16:44
	it kinda makes the svn guys look like slackers	16:44
Tv	oh it gets better	16:45
	they started writing articles on why dscms are bad for you ;)	16:45
	→ sgrimm joined	16:46
	→ gberdyshev joined	16:47
	→ bryanl joined	16:47
	→ mlins joined	16:47
	→ joern_ joined	16:47
	← joern_ left	16:48
	→ lea_w joined	16:48
	← bartocc left	16:48
	← lea_w left	16:48
	← bryanray left	16:49
eMBee	good evening	16:49
	→ bts- joined	16:49
	→ joydivider joined	16:52
	← gk left	16:52
bgerber	When I do a pull from a cloned repo I am seeing this error	16:52
	error: Could not read 66e24d206a835be951e16a7a877c5701e274e1c8	16:52
	Does this mean my update is bad?	16:52
	← aiur left	16:53
bgerber	I see this happen 2-3 times a week on my git pull of the repo.	16:53
	What does this really mean?	16:53
	← amerine left	16:54
	← Yuuhi left	16:58
	← rmh3093 left	16:59
	← Kbyte left	17:00
	→ pygi joined	17:00
	→ com4 joined	17:01
	stouset → stouset\|away	17:02
	← joydivider left	17:02
	→ joydivider joined	17:03
	← sgrimm left	17:13
	→ sgrimm joined	17:13
	→ CodeOfficer joined	17:14
	→ thiago_home joined	17:15
	→ rmh3093 joined	17:20
	← charon left	17:21
	← mcella left	17:22
	→ sverrej joined	17:22
gebi	is there a way to let gitweb fill out the cloneurl itself?	17:23
	→ bryanray joined	17:24
	← shodges left	17:24
	→ kbingham joined	17:24
	→ arpegius joined	17:27
	→ warthog9 joined	17:29
	→ tango_ joined	17:29
	alb → albertito	17:30
arpegius	looking to use github for my projects. i'll have a lot of one-off whatever.com sites that i'd love to keep in a branch 'projects'. this works out fine w/ svn, as i can switch my trunk to projects/branches/projectName... git doesn't seem lean toward this method. branches in a git project should be really connected with the codebase, right? each one of my whatever.com projects would be entirely different from one another, so i've all but	17:30
	← sztanpet left	17:31
albertito	arpegius: the line got interrupted in "i've all but"	17:33
arpegius	... abandoned thoughts of using git. any recommendations? :)	17:33
	basically, does it make more sense for me to stick w/ svn	17:34
	→ LostInTheWorld joined	17:34
quizme	git is awesome	17:34
	arpegius should ask questions correctly the first time	17:34
	eMBee is using git for his scripts. one branch per script, no common root branch, but a merge brahc to tie them all together for common checkout in a bin directory	17:34
schacon	arpegius: why not keep each project in a separate repo?	17:35
arpegius	5 private repos for the micro plan on github	17:35
eMBee	to keep things in seperate repos you need seperate directories, that's to much for single file scripts, especially if they are supposed to be in your path	17:35
arpegius	so i want to use it for in-house version control, not public projects. private repos are a must.	17:36
schacon	and you don't want them as just subdirectories in a single repo?	17:37
	stouset\|away → stouset	17:37
	→ Cacheaway joined	17:37
eMBee	oh, you want multiple projects in one branch? why?	17:38
schacon	you _can_ create separate branches for each project in one repo, but I think that may be more complex than you want	17:39
arpegius	right.	17:39
	its a cinch w/ svn	17:39
eMBee	not really, just creating the branch is slightly tricky	17:39
arpegius	when i'm in my trunk i just switch to repo/branches/projectToWorkOn	17:39
	w/ svn	17:40
	→ gitish joined	17:40
albertito	arpegius: you can use different directories, different branches, or just ditch github and have all the repos you want =)	17:40
	← hurikhan\|Work left	17:41
	← eikonos left	17:41
	→ hurikhan\|Work joined	17:41
arpegius	i guess i can make it work. just gotta wrap my head around git a bit more	17:42
	← _dontpanic_ left	17:42
arpegius	until it makes more sense than svn :)	17:42
schacon	arpegius: that's true, but that's essentially the same as having X different repos in svn - if they're totally seperate projects that can't be merged, then they're not really branches	17:42
	→ vintrepid joined	17:42
albertito	arpegius: branches are often used or the same project, but you can have totally unrelated branches in the same repository	17:42
	← qrush left	17:43
schacon	if you _really_ want to do that, you can create a bunch of empty branches that don't share a root and import the projects one by one	17:43
	http://book.git-scm.com/5_creating_new_empty_branches.html	17:43
	← polpak left	17:43
schacon	then just 'git checkout whatever_project' to switch between them	17:43
	→ chrisb joined	17:43
	← ebel left	17:43
	eMBee wonders when git is going to get a simple shortcut for that	17:44
schacon	eMBee: in general, I think most people don't think it's the best thing to do in most circumstances	17:44
	→ _dave_h_d joined	17:45
schacon	in this case, it is much better practice and makes more sense to have one repo with subdirs for each project or multiple repos, but they're just doing this to get around github or having to have multiple svn repos setup	17:46
	← ijcd__ left	17:46
schacon	even the way they're using it in svn makes more sense to have multiple svn repos, imho	17:46
_dave_h_d	hi all i have a php doc file headers, can i get git to write a revision number instead of * @version $Id:$	17:47
schacon	_dave_h_d: you probably don't want to - there is no simple id number like in svn	17:47
	it would be a 40 char sha	17:47
_dave_h_d	so get rid of that line?	17:48
schacon	i would - it's better to use 'git log file' to see history on it	17:48
arpegius	anyone running git server? i've got a 256mb slice on slicehost and just running apache and mysql takes the majority of the resources. didn't want to set up a source control server on there as well if i could avoid it.	17:48
schacon	or 'git blame file' to see who's edited it	17:48
arpegius	then i could get around github's private repo limit :)	17:48
_dave_h_d	schacon: so can i get a revision history for the file?	17:49
schacon	yup - 'git log file'	17:49
eMBee	schacon: true, but it can be convenient to have everything in one repo for backup and cloning purposes	17:49
albertito	arpegius: I use a git server that is launched by inetd, it doesnt't really use many resources. But you don't even need that if you are the only one using it, you can just use ssh	17:49
arpegius	reading http://blog.commonthread.com/2008/4/14/setting-up-a-git-server now... i see thats true albertito	17:50
albertito	arpegius: and if you want to have more than one user, you can use gitosis and handle everything over ssh	17:50
	← JanK left	17:50
	→ ehc joined	17:51
albertito	arpegius: yes, and you don't even need a git user, you can use your own and set up a "repos" directory in your home or something like that	17:52
	arpegius strokes chin stubble	17:52
	← ehc left	17:53
	→ bobmcw joined	17:54
	→ ehc joined	17:55
quizme	thanks everybody for your help. good night!	17:57
	← quizme left	17:57
	→ flaguy joined	17:58
	← zrajm left	17:59
	→ bobmcw_ joined	18:00
	← vintrepid left	18:00
	← kanru left	18:01
	← hurikhan\|Work left	18:02
	→ bobmcw__ joined	18:02
	→ gberdyshev_ joined	18:06
	← spuk- left	18:08
	← bobmcw__ left	18:08
	← lchavez left	18:09
	→ foca joined	18:10
	→ cbrake joined	18:10
	← bobmcw left	18:12
	→ thannoy joined	18:14
	← bobmcw_ left	18:16
	dwmw2 → dwmw2_gone	18:19
	→ ijcd joined	18:19
	← ijcd left	18:19
	→ ijcd joined	18:19
	← gberdyshev left	18:20
	→ eikonos joined	18:21
	← ajonat left	18:23
	→ JanK joined	18:24
JanK	shouldnt git-svn convert svn branches to git branches?	18:25
	and tags to tags	18:25
vmiklos	svn has no tags	18:25
	svn tags are in fact branches.	18:26
JanK	hmm, ok than shouldnt it create branches?	18:26
albertito	JanK: see the -t and -b options	18:26
JanK	i used --stdlayout (or similiar) and it recognized the folder structure (at least i have the trunk in the checked out folder now)	18:27
Soliton	maybe you want git branch -r?	18:27
	← Ryback_ left	18:28
	→ capisce joined	18:28
	→ DrewR joined	18:29
JanK	Soliton: yep :/	18:29
	← _dave_h_d left	18:31
JanK	hmm, ok all branches and tags are listed as "remote branches"	18:32
	but i'm wondering which branch my master is	18:32
	as it isnt trunk	18:33
	← CodeOfficer left	18:37
Soliton	git log -1 probably tells you in the git-svn-id line.	18:37
	← kergoth`work left	18:40
	← capisce_ left	18:40
	→ _graham_ joined	18:41
	→ priidu_ joined	18:42
	→ alb joined	18:43
	→ malahal joined	18:45
malahal	I did some 'rm -rf *' from the working directory... Now it says, no branch, what does that mean?	18:46
	git branch	18:46
	* (no branch)	18:46
	master	18:46
	origin	18:46
	v2.6.26-rc1	18:46
	→ tongueroo joined	18:47
Pieter	you did a checkout on non-branch, which detached your head	18:47
malahal	Pieter, that is correct. How do I checkout a tag now?	18:47
Pieter	git checkout -f tagname	18:48
malahal	Pieter, git checkout 'tag' would do?	18:48
	Pieter, thanks for the force option.	18:48
	→ lea_w joined	18:48
RandalSchwartz	checking out a tag also results in a detached head, if I recall	18:49
	as in, you're not in a place where you can make a commit	18:49
Pieter	yes	18:49
	well, you can, but it won't change the tag	18:49
RandalSchwartz	that must be relatively new	18:49
Pieter	at least since 1.5.3, I think	18:50
RandalSchwartz	in Olden Tymes, detached heads were basically view-only	18:50
	← JanK left	18:50
malahal	Pieter, RandalSchwartz I won't be making any commits....	18:51
RandalSchwartz	"then you're part of the problem, not part of the solution" :)	18:51
	or maybe just a guy... "fear of commitment" :)	18:52
	→ BarryCarlyon joined	18:54
	→ warlock_mza joined	18:55
	← priidu left	18:55
	→ sbeyer joined	18:55
	← digitalhobbit left	18:57
Pieter	I think it would be nice if "git commit" would show a warning after having committed on a headless checkout	18:57
	stouset → stouset\|away	18:58
	→ david_koontz joined	18:59
	← koke left	19:01
	← foca left	19:01
	→ kergoth`work joined	19:02
	→ archis joined	19:02
sbeyer	Pieter: it shows Not currently on any branch.	19:02
	Pieter: in the commit message editor	19:02
	Pieter: more attention please ;)	19:02
	:)	19:02
Pieter	sbeyer: yeah, but that doesn't work with commit -am"hello"	19:03
	← _slacker_ left	19:03
	← albertito left	19:04
	stouset\|away → stouset	19:04
	← lea_w left	19:05
	← loincloth left	19:07
	→ koke joined	19:08
	mlins → mlins\|away	19:08
	← koke left	19:09
	← chrisb left	19:18
	→ koweywork joined	19:24
	← alley_cat left	19:27
	← jdavis_ left	19:30
	← kbingham left	19:30
	← gitish left	19:30
sbeyer	Pieter: commit -a is evil ;)	19:31
	→ elSid joined	19:31
sbeyer	Pieter: if you use the index and do git status, then you'll see if you're not on a branch :)	19:32
	← theoros left	19:32
	← ferdy left	19:32
elSid	hey, I'm getting an error which says that a file would be overwritten by merge. cannot merge.	19:32
	I want this file to be overwritten by the merge	19:32
	I've tried deleting the file and checking it out again, but doesn't fix	19:33
	...is there a force option?	19:33
sbeyer	elSid: you could just remove the file ;)	19:34
Tv	elSid: maybe you checked it out from the index, which means it could still be modified when compared to HEAD	19:34
sbeyer	elSid: usually this message only comes up, if you have an untracked file in the working tree, that will be overwritten	19:34
alb	elSid: the file is untracked by git, and git is afraid to overwrite it. If you are sure, just remove it and then do the merge	19:34
Tv	how can it be untracked by git if he said he "checked it out again"?	19:35
	→ robinr joined	19:35
alb	Tv: I missed that part	19:36
	elSid: is the file tracked by git? =)	19:36
elSid	when I deleted it the first time, I just deleted it locally	19:36
	should i git rm filename ?	19:36
	and then re checkout	19:36
	← paltman left	19:36
elSid	yes, it's tracked by git	19:37
	→ ferdy joined	19:37
elSid	I made a change on a server's local repository, and forgot, then made changed on my comp local directory and pushed them	19:37
	← mountie left	19:38
Tv	elSid: get to a point where git status doesn't list it, then try the merge	19:38
	← flaguy left	19:40
	← warthog9 left	19:41
	← hidek left	19:41
	← Fullmoon left	19:42
	alb → albertito	19:42
elSid	..I reset the head, deleted the file and re pulled,	19:42
	seemed to work	19:42
	← elSid left	19:42
	→ elSid joined	19:43
	← Weaselweb left	19:44
	→ pjhyett joined	19:46
rillian	does msysgit work ok with putty/pagent keys?	19:48
	← jmspeex left	19:48
robinr	rillian: I think so, if plink is putty	19:52
rillian	I believe so	19:52
	TortoiseSVN manages to find it somehow, so I've never had to find out :)	19:53
	→ ShadeHawk joined	19:53
robinr	msysgit asks you about it I think	19:53
	→ morphir joined	19:53
	→ ebel joined	19:54
	← pygi left	19:54
	← mheld left	19:54
MaddHatter	rillian: yes, it works if you set GIT_SSH	19:55
rillian	ok, thanks	19:59
	I'll give it a try	19:59
	← rmh3093 left	20:01
	← warlock_mza left	20:01
	→ DiscoMur2 joined	20:05
	← DiscoMury left	20:05
	→ DrNick2 joined	20:05
	← DrNick left	20:05
	DrNick2 → DrNick	20:05
	← janm left	20:07
	← DiscoMur2 left	20:07
	→ spuk- joined	20:09
	← mithro_ left	20:09
	→ mithro_ joined	20:10
pasky	what is Giuseppe Bilotta's nick again?	20:10
tango_	pasky, me 8-D	20:10
pasky	ah	20:10
	can you explain the last hunk, please?	20:10
	← BarryCarlyon left	20:10
tango_	pasky, yes. we keep the ^{} at the end of indirect tags so that format_ref_marker can differentiate	20:11
	→ ph^_ joined	20:12
tango_	pasky, (see my $indirect = ...)	20:12
	that was jakub's idea as opposed to the git get type call	20:12
	zerok → zerok\|away	20:13
ShadeHawk	pasky, tango_: get_git_references saves extra info which is later stripped in format_ref_marker anyway	20:14
	pasky, tango_: namely the 'type' part, e.g. "tags" in "tags/v1.5.0" from "refs/tags/v1.5.0" is stripped also	20:15
tango_	hm no it's not	20:15
	→ jpwatts joined	20:15
tango_	unless that's a recent change	20:15
gebi	is there a way to let gitweb fill out the cloneurl itself?	20:16
ShadeHawk	gebi: yes, see gitweb/README (or gitweb/INSTALL); there is gitweb configuration variable for that	20:17
	→ rmh3093 joined	20:20
	→ pumpkin joined	20:21
pumpkin	how can I merge two git repos into one? I have a client and a server of a program that I'd like to track simultaneously, and I'd like to retain histories	20:22
	→ boodle joined	20:22
	← akahn left	20:22
	← rmh3093 left	20:22
boodle	can someone please explain how to do a revert?	20:23
	→ rmh3093 joined	20:23
Pieter	pumpkin: just pull one into the other	20:23
	boodle: git revert <sha>	20:23
gebi	ShadeHawk: cool :)	20:24
	thx	20:24
boodle	complains "fatal: Dirty index: cannot revert"	20:24
pumpkin	git shower	20:24
sbeyer	boodle: what do you mean by revert? Reversing the changes of a commit?	20:26
	← EmilMedve left	20:26
sbeyer	boodle: Or cleaning your working directory?	20:26
boodle	sbeyer: really want to "rollback" my local master to a prior github commit	20:26
sbeyer	boodle: Take a look if "git help reset" might help you... I think git reset --hard <sha> is the right choice for you, but I may be wrong	20:27
boodle	sbeyer: ty	20:27
sbeyer	boodle: and perhaps #github can help you ;)	20:27
boodle	oops :) ok!	20:27
	→ boink__ joined	20:28
	hachi_ → hachi	20:29
	→ loincloth joined	20:30
	→ amerine joined	20:32
	→ mountie joined	20:32
	→ flazz joined	20:35
flazz	how do i clear the index?	20:35
	← pumpkin left	20:36
	← priidu_ left	20:37
	→ sf_ joined	20:38
	← statim left	20:39
	→ statim joined	20:40
	← jlemieux left	20:40
sbeyer	flazz: I'm currently not sure if there is some porcelain for that, but git read-tree HEAD could help you, if you look for a way to remove the changes in the index	20:40
	← raz left	20:40
	boink__ → raz	20:40
flazz	i guess i could touch all the files	20:40
doener	sbeyer: uhm, "git reset"?	20:40
sbeyer	doener: without touching the working tree?	20:41
	but perhaps flazz is meaning looking for something different	20:41
	→ warthog9 joined	20:41
doener	yeah. "git reset" == "git reset --mixed" == "Resets the index but not the working tree"	20:41
sbeyer	-meaning	20:41
	doener: ok :-)	20:41
doener	flazz: what exactly do you mean by "clearing the index"?	20:42
sbeyer	doener: it's so obvious that I didn't think of it	20:42
	;-)	20:42
	mlins\|away → mlins	20:42
	← mlins left	20:43
doener	sbeyer: well, there's still the possibility that he wants "rm .git/index" ;-)	20:43
sbeyer	doener: yes	20:43
	→ alb joined	20:43
sbeyer	doener: but I couldn't imagine ;)	20:44
	→ codeshot joined	20:46
	← boodle left	20:47
	→ k776 joined	21:00
	← thannoy left	21:02
	← spearce left	21:02
	→ thannoy joined	21:02
	← albertito left	21:03
	← Eludias left	21:03
	← cannonball left	21:05
tango_	pasky, so what's your take on the refmark patch?	21:09
	→ IdahoEv joined	21:14
	← ph^_ left	21:15
IdahoEv	How do I create a repo that is a clone of another repo, but only up through a certain commit (where that commit is several commits older than HEAD). I don't want the newer information in this repo at all.	21:15
Tv	IdahoEv: init and fetch just what you want	21:16
	← iulian left	21:16
	→ archis_ joined	21:16
Tv	IdahoEv: or clone and throw away unwanted	21:16
	← koweywork left	21:16
thiago_home	IdahoEv: clone, then reset the heads to the commits you want	21:16
IdahoEv	"throw away unwanted" is done with git-reset?	21:17
thiago_home	for the current branch, yes	21:17
	if you want to change a branch without checking it out (including remote branches), you can use git update-ref	21:17
IdahoEv	okay, thanks muchly	21:17
	← nkallen left	21:18
	← lorandi left	21:22
	→ ajonat joined	21:22
	← anttih_ left	21:25
sander_	How do I merge a file?	21:27
	stouset → stouset\|away	21:27
thiago_home	gut merge <commit>	21:27
	git	21:27
Tv	you don't merge files, you merge commits (with their full history)	21:27
sander_	git merge "test" gives me: You are in the middle of a conflicted merge.	21:27
ShadeHawk	git-merge-file / merge / rcsmerge	21:28
	or git-mergetool (after configuring it, of course)	21:28
	stouset\|away → stouset	21:28
ShadeHawk	sander_: do you want to _merge_ file, or to _resolve_ a merge?	21:29
drizzd_	sander_: use git status	21:29
sander_	git status tells me the file is unmerged	21:29
	I've modified it as I want	21:29
	I want to commit!	21:29
ShadeHawk	git add <file>	21:29
	should mark it as resolved	21:30
thiago_home	what did git merge say when you got the confict?	21:30
ShadeHawk	(pre-commit hook still checks for conflict markers)	21:30
sander_	Cool. now it works	21:30
	thiago_home, You are in the middle of a conflicted merge.	21:31
thiago_home	sander_: before that.	21:31
	sander_: you did a git merge and it said there was a conflict.	21:31
	sander_: what else did it say?	21:31
sander_	Automatic merge failed; fix conflicts and then commit the result.	21:31
thiago_home	was that all?	21:32
ShadeHawk	Do you have account on KernelTrap? If you have, could you increment counter for Git User's Survey 2008 announcement in "Popular mailing list messages"? TIA	21:32
sander_	Auto-merged app/models/user_observer.rb	21:32
	← Ademan left	21:32
sander_	CONFLICT (content): Merge conflict in app/models/user_observer.rb	21:32
	← ereslibre left	21:32
sander_	thiago_home, that was all.. ANd I didnt know how to fix the conflict..	21:33
	But I know now :-)	21:33
	← archis left	21:33
	→ sergiodj joined	21:33
thiago_home	which version of git is that?	21:33
	I thought it had instructions to git add	21:33
sander_	1.5.6	21:33
	I did an "git pull".. when I got the conflict	21:34
drizzd_	thiago_home: I don't think merge tells you to use add	21:34
sander_	But git status gave me the thing to do.	21:34
	← vbabiy left	21:34
thiago_home	drizzd_: right	21:34
	it's just git rebase then	21:34
	← amigaoxxxx left	21:34
sander_	Things like this makes it abit hard to learn git.. But i'm trying ;-()	21:35
thiago_home	but it's reports like yours that help improve it	21:35
sander_	There is too many commands. and I dont know where to start all the time	21:35
drizzd_	sander_: you just learned how to merge, that's not so bad	21:35
ShadeHawk	sander_: IIRC Junio just added description of conflict format, and how to resolve conflict to git-merge manpage	21:35
thiago_home	git developers all know that you use git add when you've fixed conflicts, so they won't even pay attention to the message from git-merge	21:36
	(or lack thereof)	21:36
drizzd_	sander_: any SVN user would go green with envy	21:36
	Sput thinks git-pull at least used to give that message though	21:37
sander_	I'm actually all fresh.. git is my first real experience with a version control system.	21:37
Sput	sander_: oh good, not spoiled by $NONDVCS	21:37
	:)	21:37
ShadeHawk	Wohooo,... 883 responses (not counting 'test' channel) in Git User's Survey 2008....	21:37
	→ krawek joined	21:38
drizzd_	sander_: not polluted you mean	21:38
	← jackdempsey left	21:38
	→ flask- joined	21:38
sander_	Yeah.. My dad is using git too much.. so he convinced me to learn it too.	21:39
sf_	sander_: your dad works at trolltech?	21:39
sander_	No.. He works at birdstep.	21:39
sf_	ah	21:40
sander_	IN the .us part.	21:40
sf_	bad guess ;-)	21:40
	→ flaguy joined	21:40
sander_	Pretty nice.. I just got the deployment app capistrano.. which already have git support.. working with ruby on rails	21:41
	← gberdyshev_ left	21:41
sander_	git push; cap deploy:update; cap deploy:restart .. and a new version is running in production mode on my server. :-)	21:42
	fReAkY[t] → freaky[t]	21:42
	→ amystrat joined	21:42
ShadeHawk	sander_: did you fill the survey yet? ;-)))	21:42
amystrat	is there a way to revert a file to a previous state based on a commit tag?	21:43
sander_	ShadeHawk, which one?	21:43
	Have I missed something?	21:43
ShadeHawk	http://tinyurl.com/GitSurvey2008	21:44
	amystrat: "git checkout <tag> -- <file>" or "git show <tag>:<file> > <file>"	21:44
tango_	ShadeHawk, oh you're jakub 8-D	21:45
	← kumbayo left	21:45
amystrat	ShadeHawk, thanks	21:46
tango_	ShadeHawk, so, any chance for the refmark patch to go inline? I'm really itchy to patchbomb the ml with my other improvements to gitweb ;)	21:46
ShadeHawk	(sorry, somehow I lately use signature without "ShadeHawk on #git" there)	21:47
tango_	I should probably start usign a similar signature 8-)	21:47
ShadeHawk	tango_: the problem is that gitweb doesn't have maintainer, although Junio pays attention to what people who contribute significantly to it says	21:47
tango_	ShadeHawk, I know. he's waiting for you and/or pasky to ack the patches	21:48
	btw is there a reason why gitweb doesn't have a maintainer?	21:48
ShadeHawk	(but if you do "/WHOIS ShadeHawk" you can see that 'Jakub Narebski' is my realname)	21:48
tango_	ShadeHawk, yes that's how I just realized it was you 8-D	21:49
mutex	has anybody tried gitbuilder yet ?	21:49
	← arpegius left	21:50
	→ Gitzilla joined	21:50
	zerok\|away → zerok	21:51
	→ chrisb joined	21:51
	← amystrat left	21:51
	← ebel left	21:52
	→ novas0x2a joined	21:53
Pieter	ah, ShadeHawk is Jakub	21:53
	freaky[t] → fReAkY[t]	21:53
	← jerbear left	21:55
Mikachu	rockbox has a wiki page with ircname <-> realname mappings... :)	21:55
	fReAkY[t] → freaky[t]	21:56
	← joydivider left	21:57
	→ lea_w joined	21:59
	→ guille_ joined	21:59
guille_	hi	22:00
	→ CodeOffi_ joined	22:00
	→ doener_ joined	22:00
	ShadeHawk → jnareb	22:01
jnareb	just for you... ;-)	22:02
tango_	lol	22:02
	← brett_h left	22:02
Tv	/whois jnareb	22:04
	← leitgebj left	22:05
novas0x2a	i'm trying to exclude everything in a test dir, except for things with a . in them. i added "tests" (or ./tests, or /tests) and "!tests/." to info/exclude, but it didn't seem to work as i expected. did i do something wrong?	22:06
	→ jchris joined	22:06
	← zerok left	22:07
	← grahal left	22:07
jchris	say I want to make my master branch look just like some other branch, and I don't care about any conflicts (just want the other branch to be the master now). is there an easy way?	22:08
	CodeOffi_ → CodeOfficer	22:08
	← cilly left	22:08
RandalSchwartz	git branch -d yourbranch	22:09
	git checkout -b yourbranch master	22:09
	oops... backwards	22:09
	you want "master" to be a clone of "otherbranch"?	22:09
jchris	do you mean git branch -d master ?	22:09
jnareb	jchris: git checkout <my master branch>	22:09
jchris	yeah. master got lame	22:10
	new branch is awesome	22:10
RandalSchwartz	why do you care about the letters "m a s t e r"?	22:10
jnareb	jchris: git reset --hard <some other branch>	22:10
jchris	using it to deploy	22:10
RandalSchwartz	just ignore it	22:10
jnareb	jchris: or "git branch -f -m"	22:10
	→ cilly joined	22:10
jchris	easier than changing the others around	22:10
RandalSchwartz	if "deploy" depends on "m a s t e r", then that's the part that will end up in thedailywtf.com	22:10
	as in, fix that first./	22:10
	← bryanl left	22:11
jchris	plus the team is following master (much to my chagrin)	22:11
RandalSchwartz	again. broken	22:11
	fix that	22:11
	← _graham_ left	22:11
	← doener left	22:11
RandalSchwartz	create a practice of being able to announce "this is the good SHA"	22:11
	like using a tag or something	22:11
	branch names aren't special	22:12
	they're local.	22:12
jchris	<RandalSchwartz say I've got a bunch of weblinks to master on github that I can't change at this point. but I want master to be good	22:12
jnareb	jchris: or push <some other branch>:<master>	22:13
Gitbot	[git master]: http://tinyurl.com/68cfzf -- Merge branch 'maint'	22:13
jchris	are you saying that forcing it to be the other branch's content will hurt people when they pull?	22:13
	← stouset left	22:14
	← deavid left	22:14
jchris	jnareb: reset --hard <other branch> has the feel I'm going for	22:14
	→ deavid joined	22:14
jchris	sadly... ! [rejected] master -> master (non-fast forward)	22:16
RandalSchwartz	as it should be, as it would mess up your followers	22:16
	maybe with some sort of -f	22:17
	but then you'll get angry emails	22:17
	no matter what you do. :)	22:17
	you could do a "git merge -s ours" splice	22:17
jchris	so I basically have to merge and hand clean conflicts until I get a fast-forward?	22:17
RandalSchwartz	but I consder that... "crude"	22:18
	jnareb → ShadeHawk	22:18
jchris	I guess when you're working with distributed scm, you have to be more diplomatic	22:19
	← rillian left	22:19
jchris	in this regard, svn is fascist (but simpler) :)	22:19
RandalSchwartz	it's much easier when you don't care about losing information. :)	22:20
andres	jchris: well, if youre not really caring you could do something like git diff a..b\|git apply	22:20
jchris	I guess I don't really care about losing information	22:20
andres	But that screws merge tracking information...	22:20
jchris	this repo has seen a lot of abuse (unrelated projects in it) so the semantics I'm looking for might be best accomplished by creating a new repository and losing all history :)	22:22
	rm -rf .git/ and relax	22:23
schacon	this seems like a dumb question, but what is the difference between 'git blame' and 'git annotate'? simple usage output looks the same to me	22:23
jchris	thanks for the help and perspectives, y'all	22:25
drizzd_	schacon: I don't really know, but git blame is supposed to be faster	22:27
	schacon: do you know git gui blame?	22:27
schacon	no, I hadn't seen that - that's pretty cool	22:29
	I have never really used git-gui a lot	22:29
	← rmh3093 left	22:30
	← lea_w left	22:31
	← thannoy left	22:31
	← sf_ left	22:32
	→ rmh3093 joined	22:32
	← Weasel[DK] left	22:34
flask-	I'm trying to run a cherry-pick and it's failing with "error: Error building trees". What does this mean?	22:37
	→ kbingham joined	22:38
	← BabelO left	22:38
	→ shodges joined	22:41
drizzd_	flask-: that should not be all it says	22:41
flask-	drizzd_: first it listed the two files that had merge conflicts, each as having "needs merge"	22:41
	drizzd_: then that first file listed three time as "unmerged" with three different commit id's	22:42
	drizzd_d: then again for the second file (3 times as "unmerged"...)	22:42
ShadeHawk	schacon: git-annotate used to be different implementation	22:42
	← elSid left	22:42
drizzd_	but no other error message? that's strange	22:42
ShadeHawk	schacon: nowadays it simply git-blame with git-annotate parameters and output	22:42
	flask-: three different stages?	22:43
	← leachim6_ left	22:43
	→ elSid joined	22:44
	→ johnw joined	22:44
drizzd_	I can't find any code path that would give this error message without reporting a different error before, at least not in the current version	22:44
	up_the_irons → up_the_irons2	22:45
	→ madewokherd joined	22:45
	← juanjoc left	22:46
	← up_the_irons2 left	22:46
drizzd_	flask-: which version is it?	22:46
flask-	1.5	22:47
	1.5.5.4	22:47
	→ up_the_irons joined	22:48
	→ jackdempsey joined	22:49
	sander_ → Snadder	22:49
	alb → albertito	22:50
drizzd_	hmm, what does git status say?	22:50
flask-	it says that those two pages are changed but not updated	22:52
	hmm, maybe i need to add them	22:52
drizzd_	flask-: you probably can't cherry pick on a dirty tree	22:52
flask-	there ya go	22:52
	it worked	22:52
	← ferdy left	22:52
flask-	i forgot to add the	22:52
	them*	22:52
drizzd_	ok	22:52
	→ WALoeIII joined	22:53
	→ pygi joined	22:53
drizzd_	the error message was a little misleading, I think	22:53
	→ lenary joined	22:55
	← cmarcelo left	22:57
	→ morphir_ joined	22:57
	← rdavila left	22:59
	→ ferdy joined	22:59
	← lenary left	23:01
	→ rdavila joined	23:02
	← Adri2000 left	23:02
	← jtong left	23:05
	← shodges left	23:07
	← iggy left	23:07
	← LiamH left	23:10
	← loincloth left	23:10
flask-	drizzd_: Another question: I created a new branch 'branch_b', checked it out, and then cloned the repo. The clone repo's current branch is now listed as 'master'. Why wouldn't it be 'branch_b' ?	23:10
	← mithro_ left	23:11
ciaranm	clone doesn't clone things	23:11
flask-	?	23:11
ciaranm	clone doesn't make exact copies of everything	23:12
	it only copies some things	23:12
	clone's a rather bad name for it really	23:12
ShadeHawk	'origin' or 'origin/HEAD' would be 'branch_b' in clone	23:12
	← robinr left	23:12
ShadeHawk	unless you use mirror clone, 'master' on clone is its own local branch	23:12
flask-	hmm	23:17
	I did accomplish this, though, on another clone of the same repo	23:17
	i just have no idea how I did it	23:17
ShadeHawk	i.e. if cloned (original) repository has for example 'master' and 'branch_b', and 'branch_b" is current head at the time of non-mirror cloning	23:18
	the clone would have remote-tracking branches 'origin/master', 'origin/branch_b', and 'origin/HEAD' which would by symbolic ref pointing to 'origin/branch_b'	23:19
	'master' is there purely local branch...	23:19
flask-	i know what you're saying	23:25
	→ chrisb` joined	23:25
	← pygi left	23:26
	← Beket left	23:26
	← tongueroo left	23:26
	chrisb → Guest49560	23:26
	chrisb` → chrisb	23:27
	→ Zarin joined	23:28
	→ tongueroo joined	23:29
	← rdavila left	23:30
	← morphir left	23:30
	← morphir_ left	23:30
	→ Ademan joined	23:30
	← Guest49560 left	23:31
	→ mithro_ joined	23:34
	← p4tux left	23:36
	→ BarryCarlyon joined	23:36
	← Khisanth left	23:37
	← davidfetter left	23:44
	← parasti left	23:45
	← LostInTheWorld left	23:45
	→ perezd88 joined	23:46
	← flazz left	23:47
	← ShadeHawk left	23:47
	→ jlemieux______ joined	23:48
	jlemieux______ → jlemieux	23:48
	mw → mw\|out	23:50
	← novas0x2a left	23:51
	← Zarin left	23:52
	→ chrisb` joined	23:54
	← jpwatts left	23:54
	← chrisb left	23:56
	← flask- left	23:57
	→ iggy joined	23:58

Logs Search ←Prev date Next date→

Channels Documentation