IRCloggy #git 2010-06-03

Provider of IRC logs since 2005.

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

	2010-06-03
	← Yuuhi left	00:00
	← alezandro left	00:00
	→ stringoO joined	00:01
	← dreiss left	00:03
	← lresende left	00:06
	→ arss joined	00:08
	→ emmanueloga joined	00:08
hal	is anyone using meld to assist with merging?	00:09
	← arss left	00:09
	← theoros left	00:11
	← aafuentes left	00:11
	← Textmode left	00:12
	← rndstr left	00:13
	→ rndstr joined	00:14
	→ aafuentes joined	00:14
	→ XDS2010 joined	00:14
XDS2010	can i compile straight from git ?	00:15
	← pasky left	00:15
XDS2010	can i basically output the package with a non commited patch ?	00:15
	and can someone show me how	00:15
	its very important	00:15
mugwump	normally you compile from a checkout	00:16
cbreak	if it's not commited, it's not important	00:17
mugwump	not sure what you mean though	00:17
	→ stringo0 joined	00:17
	← icwiener_ left	00:17
mugwump	maybe you can explain what you are trying to achieve	00:17
	→ pasky joined	00:17
XDS2010	i dont have enough space to compile	00:17
mugwump	have you tried running 'git gc' ?	00:18
XDS2010	i suppose i could use my VPS	00:18
mugwump	that might free some space	00:18
cbreak	just delete stuff	00:18
	or get a new hard disk	00:18
XDS2010	i dont have enough space to compile	00:18
cbreak	I just ordered two 2TB for less than 120€	00:18
	they are very very cheap	00:19
XDS2010	some packages have this capability tho right ?	00:19
cbreak	git is not a package	00:19
	it's an SCM	00:19
	it's for development	00:19
XDS2010	argh	00:19
cbreak	every git repository contains the complete history of the project	00:21
mugwump	some commercial packages might do things like that, tools like ClearCase integrate heavily into the build process	00:22
	ie, they might have it as a tick-box on their feature list	00:22
	← luckiestmonkey left	00:22
hal	is anyone using meld to assist with merging?	00:22
	← ehabkost left	00:22
mugwump	hal: I know people that do, you found git mergetool ?	00:23
hal	mugwump: yes, that just sets the tool that is initiated when merging, doesn't it?	00:24
	← mrlarner left	00:24
	→ dreiss joined	00:24
mugwump	hal: you run it when a merge fails, and it starts the merge tool for you	00:25
	I just used it with emerge but you can configure it for meld	00:25
	git config --global merge.tool meld	00:25
	← q0rban left	00:26
	→ q0rban joined	00:26
	← shahri left	00:27
XDS2010	ok then can you guys walk me threw a build ?	00:27
hal	mugwump: yes, but I want to change the order of the arguments used when starting the merge tool	00:27
	it doesn't seem to take	00:27
XDS2010	there are some patches , but no confirmation if they work at all	00:28
	← q0rban left	00:28
XDS2010	im REALLY agravaited that git cant "export draft"	00:28
	← kpreid_ left	00:28
mugwump	XDS2010: stand back. explain what you are trying to do.	00:28
	→ q0rban joined	00:28
	← q0rban left	00:28
	→ q0rban joined	00:28
hal	currently I am using .... git config --global mergetool.meld.cmd meld '$BASE' '$LOCAL' '$REMOTE' '$MERGED'	00:28
XDS2010	well in mantis i can just export final	00:29
hal	but this is not added to .gitconfig	00:29
XDS2010	then >> download	00:29
mugwump	hal: you need to quote better	00:29
XDS2010	git doesn't have that	00:29
	git doesn't have that it would seem	00:29
hal	mugwump: why? What do you mean?	00:29
mugwump	git config expects a single argument for the value of an item	00:29
	eg git config --global mergetool.meld.cmd "meld '$BASE' '$LOCAL' '$REMOTE' '$MERGED'"	00:30
	it should show up in your ~/.gitconfig (or, more importantly, be visible on 'git config -l')	00:30
XDS2010	mugwump i want to export the final product with a uncommitted patch	00:30
	i dont want to have to keep compiling	00:31
	← corni left	00:31
XDS2010	:(	00:31
mugwump	ok... I think you're in the wrong place though	00:31
	← LRN left	00:31
	→ a9b joined	00:31
XDS2010	do any projects have this capability that anyone knows of ?	00:31
	→ LRN joined	00:32
	← Pupeno left	00:33
mugwump	"export the final product" what does that even mean	00:33
	← jmcantrell left	00:33
XDS2010	grrrrr	00:33
mugwump	well, ask a silly question...	00:34
XDS2010	yea i know	00:34
	→ steven_t joined	00:34
steven_t	hello	00:34
XDS2010	http://code.google.com/p/cyanogenmod/issues/detail?id=1575&colspec=ID%20Type%20Status%20Priority%20Version%20Model%20Owner%20Summary%20Stars%20Modified	00:34
steven_t	i get the impression that understanding git is not feasible without taking some CS courses first	00:34
XDS2010	see the bottom mugwump	00:34
steven_t	or i mean, understanding how it works	00:35
XDS2010	i mean it would be pretty silly to compile just for one patch	00:35
	but i'm pretty certain the patch works	00:35
mugwump	someone has to compile	00:35
XDS2010	< sigh > , damn	00:35
mugwump	but yeah you'll need a lot of disk space and patience to compile that :)	00:36
	Mantis probably isn't doing any compilation for its export functionality	00:37
	it looks php-focused	00:37
	steven_t: nah we teach it to people without that all the time here	00:38
	usually fine so long as they're a programmer of some sort	00:39
steven_t	teach what specifically?	00:39
mugwump	git	00:39
steven_t	oh sorry i thought i was in other channel, one where your comment would make less sense :D	00:39
mugwump	plenty of non-programmers use it too	00:39
steven_t	awesome :)	00:39
	→ Pupeno joined	00:39
	← dreiss left	00:39
mugwump	eg, I know of a project that get their translators to use it	00:39
steven_t	cool! ive read a lot of guides on git and i can do the basics, like commit, and barely branch, but after that i get lost	00:40
	so i figure the best thing to do is to learn how and why it works, not just how to use it	00:40
mugwump	there are some tech vids that might help, check the home page	00:40
steven_t	tech vids make my eyes blee	00:40
	d	00:40
mugwump	lots of resources there	00:40
	yeah, ok	00:40
	can be a bit slow	00:40
steven_t	time is definitely not something im keen on wasting.. :D	00:40
mugwump	so don't waste it by hurrying :)	00:41
steven_t	haste makes waste!	00:41
	(if ive said it once, ive said it a thousand times)	00:42
	← kenichi left	00:43
	→ hebz0rl joined	00:45
steven_t	well, ive followed this tutorial on how to use git to deploy my website,	00:49
	http://caiustheory.com/automatically-deploying-website-from-remote-git-repository	00:49
	but heres the thing.. i dont really understand what those git instructions do!	00:50
	i mean, they work, but im unsettled using something i dont understand	00:50
	← adymitruk left	00:50
steven_t	specifically i dont understand why you would create a --bare repo and then change it to no longer being bare.	00:51
	(although i also dont understand what a bare repo is.. i should go read some more)	00:51
	→ Digitalos joined	00:53
	← agrippa left	00:53
XDS2010	mugwump could i give you a tip and you could make a test build ?	00:53
hal	mugwump: thanks for your answer earlier. I think meld is working properly now, but I am struggling to use it...	00:54
XDS2010	i dont have nearly enough space	00:54
	I do on my vps tho	00:54
hal	mugwump: I am doing a 3 way merge, and there is a red highlighted section on each of the 3 files, but no arrow to take action on them	00:54
	do you have any idea how to resolve it?	00:54
	← jvraoltc left	00:55
	← floppyears left	00:57
	→ floppyears joined	00:58
	← yaotti left	00:59
XDS2010	ugh	01:00
	this isn't going to be fun	01:00
	→ scarabx joined	01:00
	← hebz0rl left	01:00
	← vinse left	01:01
	← Jonah left	01:02
	→ brett_h_ joined	01:02
	← brett_h left	01:03
	← harinath left	01:04
	→ unreal_ joined	01:05
	→ MattDiPasquale joined	01:06
	→ octaviordz joined	01:07
	→ luckiestmonkey joined	01:08
	← unreal left	01:08
	unreal_ → unreal	01:08
octaviordz	hi ppl	01:08
Sjors	hmmm	01:09
	I created a new local branch	01:09
	then I made a commit in that local branch, and now I want to create a remote branch with the same name, then send my commit there	01:09
octaviordz	I have a question ... I have a TFS repository and I wanted to migrate it to git, now I found out about svnbridge	01:09
	→ steph021_mac joined	01:09
Sjors	'git push' says "everything up-to-date", 'git push origin origin/branchname' says src refspec origin/noknotify does not match any	01:09
	→ Buglouse joined	01:10
octaviordz	so I used it to clone a project from that repository by git-svn .... but it faild in lets say r100 ...	01:10
Sjors	and 'git push origin noknotify' says "gitorious: thisrepoitory does only allow fast-forwards", remote rejecting the branchname -> branchname commit	01:10
	what do I do to create a remote branch?	01:10
octaviordz	so I cloned it again but now from r110 ...	01:10
odin_	octaviordz, clone sucks all resources and all revisions to your local repo	01:11
octaviordz	is there any way I could merge those repositories merging the svn-branches	01:11
odin_	octaviordz, there is no such thing as restarting a clone from	01:11
octaviordz	odin_ with git-svn you can clone a svn repository but just starting for a specific revision	01:13
odin_	octaviordz, ok yes with svn-git, that like chopping of any history older and XYZ	01:13
octaviordz	that is what I did ... but now I have two repositories one from revision 0-100 and other from revisino 110-500. (just as an example)	01:13
odin_	you mean 0-100 and 101-500 ?	01:14
	the problem is that you don't have the complete history	01:14
octaviordz	nop, for some reason I could not get revesion 101, 102, 103...etc	01:14
odin_	so GIT does not know how to go from 100 to 101 (or 110)	01:15
	← flaguy48 left	01:15
octaviordz	so I was wondering if there is any way I could merge the svn branche	01:15
	← aziz left	01:15
odin_	can you make a patch using pure SVN to go from r100 to r110 ?	01:16
octaviordz	as I see git does not really care about revision, am I right?	01:16
odin_	define care ?	01:16
	git sees each revision change as being a unique patch	01:16
octaviordz	well git does not handle a consecutive revision number like svn	01:16
jast	Sjors: git push origin branchname	01:17
odin_	yes it does, it just doesn't use the same number system	01:17
Sjors	jast: gitorious doesn't let me	01:17
jast	Sjors: yeah, noticed as soon as I wrote that :)	01:17
Sjors	oh it just let me \o/	01:17
odin_	git uses commit-id and SHA1 for the numbers, so a patch goes from one number to another, in a sequence	01:17
jast	do you have admin access to the repository? because this sounds like gitorious-specific configuration	01:17
	ah, great	01:17
	→ unreal_ joined	01:17
Sjors	"deny force pushing" in the settings	01:17
	toggled it, committed, then toggled it again	01:17
jast	shouldn't technically affect new branches, but whatever :)	01:18
odin_	octaviordz, what is it you want do do? you want one git repo going from r0 to r500 right ?	01:18
	octaviordz, have you thought about addressing the issues why it won't / can't clone r101 thru r110 ? can you make a pure SVN patch, checkout r100 in SVN and ask it to build a patch to get you to r110 and take a look at it	01:19
octaviordz	odin ....yes, that would be good ... but I do not care if I lost some history in the middle	01:19
	like I say for some reason (tfs o svnbridge) I could not get all revisions	01:19
	← unreal left	01:20
	unreal_ → unreal	01:20
odin_	octaviordz, but you will need a patch to cover those changes, can you make one ?	01:20
jast	and git-svn wasn't an option?	01:20
	oh, never mind	01:20
	→ harinath joined	01:20
octaviordz	ok, odin lets say I do that .... I have patches ... how can I tell git-svn (metadata) .. that now it is x revison	01:21
odin_	octaviordz, either you make one big patch to go from r100 to r110 or you need to access your r100-r500 repo as being the best you an get	01:21
	← spearce left	01:22
odin_	octaviordz, well you take your r0-r100 repo, you apply your patch, you commit it, ...	01:22
octaviordz	what I like about git-svn is that I can just do git svn fetch or git svn rebase ... and it looks for anything new in the svn repository	01:22
jast	do you actually want to keep the metadata? you only technically need it if you want to keep the original svn repository and your git-svn repository in sync	01:22
	oh, right	01:22
odin_	octaviordz, you then turn to your r110-r500 repo and ask it for format-patch for EVERY revision	01:22
jast	well, I'm pretty sure that it will be extremely tricky to fix that	01:22
odin_	octaviordz, you then go back to your r0-r100 repo and apply all those patches to it with 'git am'	01:22
	→ Phrogz joined	01:22
	→ textarcana_ joined	01:23
	← krunaldo left	01:23
odin_	octaviordz, then you should have one git repo with all history ? the only thing I don't know are if "tag" application is treated as a patch/change	01:23
jast	hmm, that might actually work	01:23
octaviordz	ok, and what about git-svn? I mean will it found at about those patches	01:24
jast	though you can use rebase instead of format-patch/am with a bit of trickery	01:24
odin_	octaviordz, but you must be able to create a patch in SVN (or whatever) to go from the one tree to the other (call it "wallpaper over the crack")	01:24
jast	git-svn should automatically rebuild the revmap once you add the additional commits	01:24
octaviordz	ok	01:25
odin_	octaviordz, maybe you want/need to verify the r100 tree and r110 trees in GIT are identical to in SVN/TFS	01:25
XDS2010	im catching so much flac with this build , would anyone mind building the source of a rom ?	01:25
	← Tiburon left	01:25
XDS2010	i need to incorporate 2 custom patches	01:25
	its very nerv racking	01:25
jast	XDS2010: wrong channel? or what are you talking about?	01:25
octaviordz	ok, thanks odin_	01:25
odin_	octaviordz, you don't want to be off by one by mistake, then test your patch is perfect, again in pure SVN/TFS, by applying it to r100 and then veify your r110 is the same as a real r110	01:26
	octaviordz, if you do that then you should be good to go with the patch	01:26
	← hyperair left	01:27
	→ seanmccann joined	01:27
	← seanmccann left	01:27
	→ jmcantrell joined	01:28
	→ entrosca joined	01:28
jast	octaviordz: basically, create repoA with old revisions and repoB with new revisions. go into repoB, git fetch repoA, git rebase repoA/branch	01:28
	not tested :)	01:28
	← jmcantrell left	01:28
odin_	jast, I am git noob, rebase is still a little like voodoo to maybe "format-patch/am" is safer :)	01:29
jast	rebase and am use the same mechanism internally	01:29
XDS2010	anyone here heard of mahalo ?	01:29
odin_	well me something #git, the documentation often talks of "index" in relation to modifications	01:30
	now I understand changed to the working tree (i.e. the real files on the file system in front of me)	01:30
	← locklace left	01:30
jast	basically, that command says this: remember all commits from current branch that are not in repoA/branch. hard-reset the current branch to repoA/branch. now apply the remembered changes.	01:30
odin_	but I have not quite understood the implications of how/when to make changed with the index are important	01:30
octaviordz	jast .. actually I did something like that but I got int repoA fetching from repoB O_o	01:30
	→ alezandro joined	01:31
odin_	so what is this mythical thing called the git "index" that the documentation talk about	01:31
jast	odin_: we also call it the staging area. it's used to tell git which versions of which files should be included in the next commit. you can use it to commit only some of the changes you've made to the files.	01:31
	→ jmcantrell joined	01:32
	→ locklace joined	01:32
jast	odin_: basically, git add says: use the current version of this file for the next commit. git commit -a says: use the current version of all files that I already know about. git add -p says: show me the individual changes in all the files and let me choose which hunks to stage	01:32
odin_	so if I make a directory and don't 'git add' then 'git status' doesn't know about it	01:32
	→ entropi joined	01:32
odin_	but if I put a file in that directory, suddenly 'git status' knows it is dirty	01:32
	← entrosca left	01:33
odin_	when I 'git add' then it updates the index, to add it to the commit ?	01:33
jast	if you never add something, it's not in the index	01:33
	→ codeshepherd joined	01:33
jast	once you add something, git starts doing stuff like checking whether the working tree has a different version than the index	01:33
m0	Hi, when someone refers to "Config	01:34
	http://sourceforge.net/apps/trac/sourceforge/wiki/Git%20hook%20script%20example Do I put in .git/config [hooks] mailinglist or hooks.mailinglist ?	01:34
	Is it by convention in git when we do firstword.secondword to denote a dictionary key ?	01:35
odin_	jast, ok thanks for clarification	01:35
	→ Morasique joined	01:35
jast	m0: we often call it foo.bar, and the "git config" tool accepts that notation, but in the actual config file it's [foo] bar. and foo.bar.baz is [foo "bar"] baz.	01:36
odin_	where is the best source of documentation about utility functions ? for example, I want to script some use of 'git', and I want to know the current branch and the last commit-id (on one line, on their own)	01:36
	← stringo0 left	01:36
jast	odin_: I don't know if there is any such thing. feel free to get pointers from #git until you get a feel for it, though. :)	01:36
	current branch: git symbolic-ref HEAD	01:37
	current commit: git rev-parse HEAD	01:37
odin_	thanks, is there a list of those things ? which man page?	01:37
	→ ruien joined	01:38
jast	"man git" has a list of all commands, roughly sorted with terse descriptions	01:38
	← Vampire0 left	01:38
XDS2010	anyone here heard of mahalo ?	01:38
m0	jast: is there a way to test hooks why its not working?	01:38
jast	the further down the list, the lower the level	01:38
	→ spicyWith joined	01:38
jast	m0: make it output something. if the output doesn't show up, the hook file is probably not executable or something, or it doesn't have the right filename, or it isn't in .git/hooks.	01:39
	(fwiw, in case you thought you had to, you don't need to configure a hook in .git/config unless the hook itself looks at that file)	01:40
m0	jast: I don't understand sorry :s I need to place hooks.mailinglist config	01:40
	jast: I thought I have to place that in the configure of the remote server	01:40
jast	m0: depends on the hook. I don't know which hook you are talking about.	01:41
	→ krunaldo joined	01:41
m0	jast: in debian it is located in /usr/share/doc/git-core/contrib/hooks/post-receive-email	01:41
	jast: which is http://sourceforge.net/apps/trac/sourceforge/wiki/Git%20hook%20script%20example	01:41
jast	oh, lemme have a look	01:41
	→ papna joined	01:41
jast	where did you put that hook in the remote repository?	01:42
	→ herdingcat joined	01:42
m0	jast: yea, in my prototype-web.git/hooks	01:42
	← codeshepherd left	01:42
jast	named "post-receive"? set to executable?	01:42
papna	How do I discard all my current changes rather than committing them?	01:42
jast	papna: completely and forever? git reset --hard	01:42
papna	jast: Thanks	01:42
odin_	is it possible to ask a remote repo to provide a format-patch between two tags on a branch ? i.e. I don't want the full network traffic for a clone, since I only need a small number of commits, for example I want to obtain a format-patch on repo linux-2.6-stable.git ....	01:42
m0	jast: I did chmod a+x post-receive-email and symlinked "ln -sf /usr/share/doc/git-core/contrib/hooks/post-receive-email hooks/post-receive"	01:42
odin_	on branch remotes/origin/master between tags v2.6.33 and v2.6.33.2	01:42
	→ hyperair joined	01:43
jast	m0: that should be fine. so you've added that [hooks] section and it's giving you trouble?	01:43
	→ lresende joined	01:44
	← lresende left	01:44
	← mbroeker left	01:44
m0	jast: I am going to add a print "sending email hook" in the start to see if I see it	01:44
odin_	I suppose the local tool is doing the asking "git" and it looks up all the commit-id between two tags and the remote repo streams them over ? maybe this is just not possible	01:44
m0	strange, there are a lot of echos in that file that I should have seen	01:45
	→ lresende joined	01:45
	→ codeshepherd joined	01:45
	← spicyWith left	01:45
	→ eli joined	01:45
m0	jast: it worked!	01:47
jast	what did? :)	01:47
	← hyperair left	01:47
m0	jast: the email =)	01:47
jast	ah, cool	01:48
XDS2010	anyone up to building for a tip?	01:49
	:(	01:49
	im all out of gas	01:49
ruien	?	01:50
XDS2010	ruien need to build a rom from source with 2 custom patches	01:51
	30$ mahalo tip in it for anyone who can help	01:52
ruien	build a rom..?	01:52
XDS2010	can i pm you ?	01:52
	O.O	01:52
ruien	sure	01:52
	→ Vampire0 joined	01:52
	← Ryback_ left	01:53
	→ altogether joined	01:54
m0	Now my next step is to create a hook that copies the changed files to the webserver :(	01:59
	I want to do a "production" branch where If I push changes to that branch it will update the webserver	01:59
	Lets see how easy it is with Git hooks	02:00
hal	is there any way to undo my last commit AND push to the remote branch?	02:01
	I am certain that noone else has used the branch since the change	02:01
	→ tye joined	02:02
tye	if I have the hash of a glob's contents, is there a good way to list any commits that introduced that glob? and how do I list all commits that contain a particular patch ID?	02:04
	→ Remixman joined	02:06
	← __skyl__ left	02:07
	← Bass10 left	02:07
drizzd	tye: you can have a look at the git find-large script	02:07
	man find-large	02:07
	faq find-large	02:07
Gitbot	drizzd: Howto find large files. See https://git.wiki.kernel.org/index.php/GitFaq#find-large	02:07
drizzd	it uses batch-check to find the path of a blob	02:08
	← Remixman left	02:08
drizzd	starting from that you can find the commit that introduced it using log	02:08
cehteh	mhm .. i should add my git-analyze-copyright to that wiki? ..	02:09
	→ Remixman joined	02:12
	→ hyperair joined	02:13
	→ pklingem joined	02:15
	← papna left	02:16
	→ Blackguard joined	02:18
	→ ExtraSpice joined	02:18
	← tildeequals left	02:20
	→ cytrinox_ joined	02:21
	← Remixman left	02:21
	← hal left	02:22
	← Yuffster left	02:22
	← cytrinox left	02:24
	cytrinox_ → cytrinox	02:24
	← ExtraSpice left	02:25
	← textarcana_ left	02:25
	← Digitalos left	02:25
	← siyb left	02:25
	← fmardini left	02:25
	← redondos left	02:25
	← tedoc2000 left	02:25
	← Mrcheesenips left	02:25
	← enherit left	02:25
	← Mage_Dude left	02:25
	← yashi left	02:25
	← SamothUK left	02:25
	← resno left	02:25
	← ft left	02:25
	← ivan left	02:25
	← merlin83 left	02:25
	← Cesario left	02:25
	← mooosh left	02:25
	← multi_io_ left	02:25
	← Borbus left	02:25
	← teuf left	02:25
	← Leemp left	02:25
	← tro left	02:25
	← eldragon left	02:25
	← avar left	02:25
	← uau left	02:25
	← tango_ left	02:25
	← niki left	02:25
	← squentin left	02:25
	← torben left	02:25
	← maxb left	02:25
	← MadCoder left	02:25
tye	thanks, drizzd	02:26
	← steph021_mac left	02:27
	→ consolers joined	02:27
	→ ExtraSpice joined	02:28
	→ CAAQZRE joined	02:28
	→ siyb joined	02:28
	→ fmardini joined	02:28
	→ redondos joined	02:28
	→ Mrcheesenips joined	02:28
	→ enherit joined	02:28
	→ Mage_Dude joined	02:28
	→ yashi joined	02:28
	→ SamothUK joined	02:28
	→ resno joined	02:28
	→ ft joined	02:28
	→ ivan joined	02:28
	→ mooosh joined	02:28
	→ merlin83 joined	02:28
	→ Cesario joined	02:28
	→ multi_io_ joined	02:28
	→ Borbus joined	02:28
	→ teuf joined	02:28
	→ Leemp joined	02:28
	→ tro joined	02:28
	→ eldragon joined	02:28
	→ avar joined	02:28
	→ uau joined	02:28
	→ tango_ joined	02:28
	→ niki joined	02:28
	→ squentin joined	02:28
	→ torben joined	02:28
	→ maxb joined	02:28
	→ MadCoder joined	02:28
	→ Digitalos joined	02:28
	← eldragon left	02:28
	→ meanburrito920 joined	02:28
	← Mrcheesenips left	02:29
	← hobodave left	02:29
	← Phrogz left	02:30
	→ Mrcheesenips joined	02:30
	← Mrcheesenips left	02:30
	→ Mrcheesenips joined	02:30
	← pklingem left	02:30
	→ eldragon joined	02:30
	→ jvraoltc joined	02:32
	→ Remixman joined	02:34
	→ johnw joined	02:36
	← hakunin left	02:37
XDS2010	anyone have the time to build a kernel with a custom patch ?	02:37
	→ tildeequals joined	02:37
jast	how is that related to git?	02:37
	→ KittyKatt joined	02:38
	→ hakunin joined	02:38
XDS2010	jast huh ?	02:38
	consolers reaching for suspend2 again, since 2.6.34 either cannot suspend for vbox or is much slower on ressume than a regular boot	02:38
jast	this channel is about git, not about compiling kernels	02:38
KittyKatt	Not sure if this is the place to ask a gitosis related question.	02:38
jast	KittyKatt: can't hurt to ask	02:38
KittyKatt	Will do.	02:38
consolers	xds /m me if you want	02:38
	→ Yuffster joined	02:39
KittyKatt	I'm using gitweb's $export_ok variable to look for a file in the bare git repositories that are in my gitosis system user's home.	02:39
	→ Phrogz joined	02:39
	← enherit left	02:40
KittyKatt	When I have the file there, gitweb displays it correctly, as it should.	02:40
	→ eclubb joined	02:40
odin_	is there a set of rules somewhere, over all the points to check for to see if a patch is a valid git patch ?	02:40
KittyKatt	But whenever I actually push to gitosis-admin.git, it deletes the file in all of the repositories.	02:40
	Anyway to prevent this?	02:40
odin_	for example, a normal patch is not a valid git patch, it has no From: line, etc...	02:40
	maybe there is already a tool that can take any patch and make it compliant with format-patch output ?	02:41
	← hyperair left	02:41
	← drizzd left	02:42
	← Phrogz left	02:42
jast	KittyKatt: I believe gitosis wants you to configure access in its own config file (gitweb=yes/no) and thus interferes, so you'd probably have to patch gitosis if you wanted to use a different mechanism	02:43
	odin_: look at git am's source code to see what it does with git patches	02:44
KittyKatt	Well, gitweb=(yes\|no) doesn't affect the way it displays at all.	02:44
	Perhaps because I am using git-daemon.	02:44
	Hm.	02:44
	→ hyperair joined	02:44
jast	gitweb and git daemon work independently from each other	02:44
KittyKatt	I shall see.	02:44
	← meanburrito920 left	02:45
	→ CodeTalks joined	02:45
	← MattDiPasquale left	02:47
	← eclubb left	02:47
KittyKatt	jast: See, putting a "gitweb = no" under the "[repo gitosis-admin]" section doesn't hide it in the overview page.	02:47
	At least for me it does not.	02:47
	→ eclubb_ joined	02:48
KittyKatt	I think it needs a gitweb.conf. I'm not sure.	02:48
jast	in gitweb.conf, make sure there is a line like $projects_list = "/home/git/gitosis/projects.list";	02:48
KittyKatt	gitosis doesn't update that file. ^^;	02:48
eclubb_	Hello. I have a workflow question.	02:48
jast	no, it doesn't, and usually it can't update it anyway	02:49
KittyKatt	My projects.list file is completely empty. At all times.	02:49
Ilari	KittyKatt: Does .gitosis.conf (its hidden) reflect changes pushed?	02:49
jast	that's, uh, weird	02:49
KittyKatt	Hm?	02:49
	→ CodeTalks_ joined	02:49
jast	after pushing updates to the admin repository, do they show up in /home/git/.gitosis.conf?	02:50
	← CodeTalks left	02:50
	CodeTalks_ → CodeTalks	02:50
	← CodeTalks left	02:50
KittyKatt	Yes.	02:50
	Of course. ^^;	02:50
	But my projects.list file has never had any information in it.	02:50
jast	well, you wouldn't believe how often something like that turns out to be the problem	02:50
	well, I guess for the default configuration it might not matter	02:50
KittyKatt	Hm...	02:51
	→ timj_ joined	02:52
KittyKatt	How to go about this?	02:53
jast	well, in my gitweb config, $export_ok is 'git-daemon-export-ok'	02:53
	→ MattDiPasquale joined	02:53
eclubb_	I forked a project (proj A) on GitHub so I could make some mods to my own copy (proj B) and use it as a base for several client projects. I then cloned proj B to do some work for a specific client (proj C).	02:53
KittyKatt	I would use gitweb's $export_auth_hook, but I'm not sure what to use in the perl subroutine.	02:53
jast	and then I set daemon=yes/no for public access in general, and gitweb=yes/no to control whether it shows up in the list	02:54
KittyKatt	...daemon.	02:54
	Right.	02:54
	One moment.	02:54
jast	I think gitweb=yes only makes sense if you also specify description and owner... otherwise I think it never adds the repository to the project list	02:54
	and then of course you'd have to tell gitweb about the projects list, as described above	02:55
KittyKatt	Yes, projects.list	02:55
	Hum.	02:55
eclubb_	I'm wondering what is the best way to get upstream changes from proj A through, proj B, adn then into proj C.	02:55
KittyKatt	Indeed.	02:55
	← jfkw left	02:55
jast	eclubb_: merge from A to B, then merge from B to C? too simple? ;)	02:55
	← timj__ left	02:56
KittyKatt	I think I'm going to look into defining a config variable in the repositories I want shown, and use the $export_auth_ok perl subroutine function to test if it is set.	02:56
	Does that sound like it could possibly work? >.>	02:56
	← jvraoltc left	02:56
jast	eclubb_: also, look at topgit for an alternative to managing sets of additional patches	02:56
	s/to/for/	02:56
	→ adamv joined	02:57
eclubb_	jast: That's actually what I'm doing now, but by the time it get's to proj C, I've lost the history of proj A and all I have in my history (apart from my own changes to proj C) is a merge entry for each merge from proj B.	02:59
jast	history doesn't get lost in merges	02:59
XDS2010	anyone want a quick job building a custom kernel ? all thats needed to do is download the source with git, apply 2 patches , and make.... needs to be on a 32bit machine & u can use my vps	02:59
eclubb_	jast: the individual commit history from A is gone.	02:59
XDS2010	from git	02:59
jast	no it isn't	03:00
	that's what merges are for: to keep all commits in the history	03:00
KittyKatt	What is the file "info/exclude" in a bare repository supposed to do? o.O	03:00
jast	KittyKatt: nothing. it's just part of the template for all new repositories.	03:00
KittyKatt	jast: I should reword the question. What CAN it do.	03:00
jast	in a bare repository, nothing.	03:01
KittyKatt	I see.	03:01
	KittyKatt is still looking for a way to stop gitosis from deleting the export_ok file. >.<	03:01
jast	eclubb_: you're probably either using something other than "git merge/pull" for merging, or you are looking at the history with a tool that hides part of the history of merges. the stuff should show up in gitk, for example.	03:02
	→ rryouumaa joined	03:02
KittyKatt	jast: Maybe...maybe I could make the file not writable by the git user?	03:02
	← eletuchy left	03:03
KittyKatt	chown it to someone else, perhaps.	03:03
jast	the exclude file? sure, but since the user isn't actually going to change the file, I don't see how it might make a difference	03:03
	← kukks left	03:03
	→ JonasF joined	03:04
	← doener left	03:04
KittyKatt	jast: Whenever I make changes to the gitosis-admin.git repository on my developing environment (i.e. working tree environment), and push to the server (bare repository environment), it deletes the git-daemon-export-ok file from all the repositories.	03:05
	← harinath left	03:05
KittyKatt	Thus making gitweb not display any projects.	03:05
jast	changing the permissions on a file doesn't prevent anyone from deleting it	03:05
	→ doener joined	03:05
	→ sid3k` joined	03:05
KittyKatt	Sure it does.	03:05
jast	no	03:05
KittyKatt	chmod a-w git-daemon-export-ok	03:05
	^^;	03:05
jast	deletion gets controlled by whether the containing directory is writable for the user	03:05
KittyKatt	...damn, forgot about that little factor.	03:06
	Hum-dee-dum...	03:06
jast	well, you could set daemon=yes in gitosis.conf to make it create that file, couldn't you?	03:06
	← j0rd left	03:07
	→ j0rd joined	03:07
KittyKatt	What exactly does the "daemon = yes" setting do?	03:08
	→ LotR joined	03:08
	→ Pev_away joined	03:08
	→ tedoc2000 joined	03:08
eclubb_	jast: Umm...right. I just pulled from A through B to C and all the commits are showing in C's history (like you said). It was very late last night when I last looked at this =) Anyway, it's good to have confirmation that I'm doing it the right way. Thanks for the help.	03:09
	→ eletuchy joined	03:09
jast	KittyKatt: AFAIK it creates the git-daemon-export-ok file and that's about it	03:09
KittyKatt	jast: I'll try that out.	03:09
	→ fool_ joined	03:10
fool_	hey guys what happen if you interupt a git commit ?	03:10
KittyKatt	jast: Awesome, that worked perfectly.	03:11
jast	fool_: usually the commit simply doesn't happen... depending on what you mean by interrupting	03:11
	← zimbres left	03:11
fool_	jast: ctrl-c ?	03:12
	or ctrl-z ?	03:12
jast	oh, it takes that long to create the commit that you actually have time to hit ctrl-c? :)	03:12
fool_	yeah :)	03:12
KittyKatt	jast: Perfect. Defining gitweb = no and daemon = no under [repo gitosis-admin] and daemon = yes and gitweb = yes at the top of the file under [gitosis] worked perfectly. Thanks for the help.	03:12
fool_	binary files	03:12
	big glob	03:12
Ilari	If you hit CTRL+C after it prints ID for commit, the commit has already been made...	03:13
jast	but the blobs get created while you do git add, so commit itself doesn't actually have to look at them again	03:13
	at any rate, commit only updates the branch after all objects have been created, so only two things can happen: there are some dangling objects (no problem, just some temporary waste of space) and the commit isn't recorded in the branch, or the commit goes through	03:14
fool_	well it print out commit ID already but stays there for 20 min now so i'm stuck at what to do next	03:14
	it's sucking up cpu like crazy too	03:15
Ilari	fool_: Ctrl+C it and look at the log?	03:15
	fool_: Or look at the log from anothe window?	03:15
fool_	Ilari: well that's why i ask first what happen if i interupt it	03:15
	lol	03:15
jast	autogc, perhaps?	03:15
Ilari	fool_: Well, open another shell and take a look at the log?	03:16
jast	well, the absolute worst thing that can happen if you interrupt is that you have to type the commit message again at some point	03:16
fool_	Ilari: the log has the commit already	03:16
jast	well, in that case it's safe to interrupt	03:16
fool_	so the question is should i just wait till it finish whatever it's doing or ???	03:16
jast	I assume it's trying to optimize the repository	03:16
	if you don't really care about that, just go ahead and interrupt	03:17
	← eclubb_ left	03:17
fool_	what do you mean optimize ?	03:17
	repack stuff ?	03:18
	← Blackguard left	03:18
mugwump	eg git gc - but it normally says that it's doing that	03:18
	→ mithro joined	03:18
	→ munichlinux joined	03:19
	→ harinath joined	03:20
XDS2010	anyone want a quick job building a custom kernel ? all thats needed to do is download the source with git, apply 2 patches , and make.... needs to be on a 32bit machine & u can use my vps , last chance	03:21
	→ __skyl__ joined	03:22
	→ jvraoltc joined	03:23
nanotube	XDS2010: well if you have a vps to do it on... why don't you do it yourself?	03:24
KittyKatt	brb	03:24
	← KittyKatt left	03:24
	← friskd left	03:24
	→ KittyKatt joined	03:25
johnw	yeah, seems like a strange offer	03:26
XDS2010	nanotube no time	03:27
	have work	03:27
	etc	03:27
	compensation available	03:28
	:)	03:28
	→ kpreid_ joined	03:28
	→ pklingem joined	03:28
me22	XDS2010: switch to gentoo, then the ebuild will do it	03:29
	← Buglouse left	03:29
XDS2010	eb?	03:30
	me22 wut?	03:30
jast	that would take even more time	03:30
nanotube	lol	03:30
jast	after all, a computer doesn't switch itself to gentoo	03:30
mugwump	yeah run gentoo on your phone	03:31
XDS2010	how much time does it take to build a standard kernel from source ? ( aprox )	03:32
nanotube	XDS2010: sorry, no time. try asking on ##linux	03:32
jast	depends on your system power. usually a couple of minutes.	03:33
nanotube	XDS2010: depends on speed of hw	03:33
jast	plus time you spend configuring it, of course	03:33
	and that depends largely on your level of experience and level of certainty about which features you're gonna use	03:33
mugwump	of course the important detail being omitted here is that this is a cyanogen kernel for android	03:34
johnw	what distro are you using XDS2010?	03:35
	→ hohoho joined	03:35
mugwump	it's not even a kernel patch	03:36
	← pklingem left	03:36
mugwump	but the tools are probably geared towards building the whole environment at once	03:36
XDS2010	johnw sw12	03:36
	← jensn left	03:37
johnw	well, you've got enough time to chat about it here	03:37
	→ henryubuntu joined	03:38
	← harinath left	03:38
mugwump	johnw: difference, one involves engaging brain...	03:38
	← aresnick left	03:39
johnw	i don't buy that he really doesn't have enough time, given how much time it would take to fully explain and get another person to do it	03:39
consolers	last time i measured, menuconfig took 1.5-3 times the actual build	03:39
	→ LongBeach joined	03:39
consolers	sure there may be other reasons	03:39
odin_	can I clone a local git repo as in copy in such a way as the new repo's upstream points to the original upstream ? (almost like a file copy)	03:40
consolers	worst case it may be like asking someone else to click "Accept" of a license agreement	03:40
johnw	nor would I want to deal with the tax ramifications of such "remuneration"	03:41
consolers	or is the vpn in nigeria??	03:41
johnw	vps, not vpn	03:41
consolers	i've got several offers to fly there to inspect mines	03:41
jast	odin_: making a physical copy is probably easiest	03:42
XDS2010	my verizon dial up connection doesn't help either	03:42
johnw	you may log in to his VPS only to discover that something has suddenly "broken", and he now wants you to pay him to fix it	03:42
odin_	jast, so there is nothing in .git/ which contains an absolute path or the toplevel dir name ?	03:42
jast	odin_: only in very exotic cases	03:42
XDS2010	johnw wth	03:43
	no one has to log into anything	03:43
consolers	nevermind xds	03:43
odin_	jast, I dont think I'm exotic, just trying to cut down network bandwidth and time	03:43
	→ jony joined	03:46
	← consolers left	03:47
	→ multigrain joined	03:51
	← henryubuntu left	03:52
	→ harinath joined	03:56
	← dgrift left	04:00
	→ dgrift joined	04:01
	← scarabx left	04:03
	← Mrcheesenips left	04:03
	→ robinr joined	04:05
	← jatan left	04:07
	← octaviordz left	04:08
	→ jatan joined	04:10
	→ benjo2 joined	04:11
	← cilly left	04:11
	→ cilly joined	04:13
	← luckiestmonkey left	04:13
	← brett_h_ left	04:16
	← cenuij left	04:17
	← jatan left	04:19
	→ sunzoje joined	04:20
	← jmcantrell left	04:20
	→ jatan joined	04:21
	← jvraoltc left	04:21
	→ magn3ts joined	04:22
	→ jvraoltc joined	04:23
	← jony left	04:24
	→ alireza_ joined	04:24
	← fool_ left	04:30
	→ schlortoo joined	04:30
	← drako left	04:32
	← codeshepherd left	04:34
	→ girishr joined	04:35
sunzoje	join #csharp	04:35
sitaram	command fell flat	04:35
frogonwheels	lol	04:35
	← Digitalos left	04:35
	→ codeshepherd joined	04:38
	← toxygen left	04:42
	→ toxygen joined	04:42
	← KittyKatt left	04:44
	→ radioman-lt joined	04:45
	← tar- left	04:45
	← yashi left	04:46
	← jjuran left	04:46
	→ kozmikreis joined	04:46
	← radioman-lt left	04:52
	→ giskard joined	04:52
	← giallu left	04:53
	← manyoso left	04:55
	← tildeequals left	04:57
	→ radioman-lt joined	04:57
	← urkud_ left	04:59
	← Twisted_Mentat left	04:59
	→ amerine joined	05:00
	← hyperair left	05:00
	→ hyperair joined	05:01
	← Yuffster left	05:01
	→ typemore joined	05:01
	← sunzoje left	05:01
typemore	)&( I just delted a branch that is important that I should not have delted.	05:01
	Is there a way I can recover it?	05:01
	I've delted it for less than 30 days. In fact, I've delted it for less than 30 minutes	05:01
	← adamv left	05:01
frogonwheels	typemore: if you've got a SHA1 then just checkout that sha1 and recreate the brancyh	05:02
typemore	i don't have it	05:02
frogonwheels	.. use reflog on HEAD to find it	05:02
	I'm not sure if reflog applies to deleted branches. somebody else might know.	05:02
typemore	"aa76466 HEAD@{3}: commit:"	05:02
	that's what I want	05:03
	how do I get that?	05:03
jast	yeah, git checkout -b thebranch aa76466	05:03
frogonwheels	:) Easy	05:03
jast	or HEAD@{3} instead of aa76466	05:03
	← dgs left	05:03
typemore	phew; i can discard that resignation email now; thanks everyone :-)	05:04
jast	awesome	05:04
	another life saved	05:04
	go go git!	05:04
	← stringoO left	05:05
	← doener left	05:06
	→ doener joined	05:08
hyperair	are git's aliases recursive?	05:09
	← a9b left	05:10
	← SunilThaha left	05:10
	← hasu left	05:10
	← Joeconyers left	05:10
johnw	you mean, can you use an alias inside an alias? try it	05:10
hyperair	i would, but my system's getting bogged down by too much i/o at the moment	05:11
johnw	you can chat, but you can't edit .gitconfig?	05:11
	→ villemv joined	05:11
hyperair	no wait, i think i might have tried it sometime back and failed miserably.	05:11
	more like i can edit .gitconfig, but i can't test it properly	05:12
	aha. it doesn't work.	05:13
	← Mage left	05:14
	→ Digitalos joined	05:14
	→ Mage joined	05:14
	→ keystr0k joined	05:15
keystr0k	Hey all... I just did a git pull and there was a merge conflict on an SWF file (compiled). How can I get the file in its original state before the merge conflict if latest version wasn't commited?	05:16
	Right now the file is "unmerged"	05:16
johnw	i believe it's git show :3:FILE	05:17
keystr0k	Normally I'd just clean up the file and re-add and commit... but where it's compiled it's a bunch of jibberish	05:17
johnw	you'll find this described in the "HOW TO RESOLVE CONFLICTS" section of man git-merge	05:17
Gitbot	johnw: the git-merge manpage can be found at http://git.or.cz/man/git-merge	05:17
	→ stringoO joined	05:17
	← lresende left	05:18
keystr0k	So do I want the originals, common ancestor, HEAD, or MERGE_HEAD version ?	05:19
	→ lresende joined	05:19
johnw	MERGE_HEAD, I believe	05:19
	← lresende left	05:19
keystr0k	sorry for my ignorance re: this btw	05:19
johnw	common ancestor is a historical version, prior to HEAD	05:19
	← MattDiPasquale left	05:19
keystr0k	which is 3...	05:19
johnw	HEAD is what you're trying to merge into	05:19
	yeah, try 3	05:19
keystr0k	thanks johnw	05:20
	← juice_d left	05:21
doener	hm, latest version wasn't committed? Why did git allow the merge then? Or did you do "git checkout -m <other_branch>" to get the conflict?	05:21
johnw	hi doener	05:21
	← harinath left	05:21
	← chrismsnz left	05:21
doener	hi johnw	05:21
	and given a recent git version, if you want to replace the file in the working tree, you should be able to use "git checkout --ours/--theirs <path>" instead of messing with "git show :<stage>:<path>"	05:22
keystr0k	doener: the latest version was just sitting there. I didn't realize that it was in a dirty state... I did a pull and it did that.	05:23
doener	keystr0k: hm, merge should have complained and aborted	05:23
keystr0k	so now I'm left with a dirty file... and I cannot easily see how it dirtied it... yeah. I thought it would have done that too...	05:23
	git show :3:FILE may have worked.	05:23
johnw	git checkout --ours FILE is more sense-making	05:24
doener	keystr0k: anyway, if you're ok with just using the version from "upstream", git checkout --theirs <path> should be fine	05:24
	→ loincloth joined	05:24
	→ jony joined	05:24
doener	(:3: is the "theirs" version, so I assumed that to be the one you want...)	05:24
	← loinclot_ left	05:24
keystr0k	doener: I see...	05:25
johnw	HEAD is ours? ah, I see now	05:25
keystr0k	so is there any way to get the file in it's modified but uncommited state?	05:25
	before git-merge had its way with it?	05:25
	→ meanburrito920 joined	05:25
keystr0k	or is that what :3 is getting me?	05:25
doener	keystr0k: no stage 3 contains the version from the branch you're merging	05:26
	s/no/no,/	05:26
johnw	:2: should be his file?	05:26
doener	hm, well, stage 2 is the version from HEAD, but if he really had uncommitted changes (I'm still not sure if git really allows that), those aren't in stage 2, AFAIK	05:27
johnw	Git doesn't make a temporary commit for the sake of attempting the merge?	05:27
keystr0k	oh darn	05:28
	→ vu3rdd joined	05:28
doener	git usually simply refuses to do the merge	05:28
	keystr0k: had you run "git add" on the uncommitted version?	05:28
keystr0k	WOW.	05:28
	I have been mis-speaking.	05:28
	I DID commit the changes...	05:28
johnw	ah ha	05:28
keystr0k	I'm so sorry!	05:28
doener	ok, sanity restored	05:28
keystr0k	;)	05:28
	Phew	05:28
	lol	05:29
	← benjo2 left	05:29
	→ tildeequals joined	05:29
doener	my brain was just trying to enumerate the special cases for merge, in which it allows a merge even with uncommitted changes. And that hurts ;-) (It's a pretty cool feature, but I can never recall the special cases, except for the one that's bad bad bad for users with a "commit -a" habit)	05:30
keystr0k	haha.	05:30
	doener: I appreciate you racking your brain... hehe...	05:31
	→ opalka joined	05:31
	← cli_ left	05:31
doener	don't think of it as a voluntary action, I just can't help it g	05:31
keystr0k	heh	05:31
	→ cli_ joined	05:32
	doener wanders off to get some coffee to get awake	05:32
	→ jjuran joined	05:32
	← sagsousuke left	05:34
	← magn3ts left	05:34
	→ treak007 joined	05:35
	→ harinath joined	05:35
	→ dytrivedi joined	05:36
	→ timClicks joined	05:37
	→ roop joined	05:38
	→ urkud_ joined	05:38
timClicks	is there an option to use git format-patch and have it generate a single patch, rather than patch per commit?	05:38
	← nanotube left	05:42
mugwump	timClicks: use git diff	05:43
jast	no. if you want a diff over a range of commits, use git diff commit1 commit2 >file	05:43
	← bcardarella left	05:43
	→ bcardarella joined	05:43
timClicks	mugwump jast: ty	05:43
frogonwheels	timClicks: then use git apply to apply the patch.	05:46
	→ hasu joined	05:47
	← eletuchy left	05:47
	← bcardarella left	05:48
	→ rseifert joined	05:53
	← hyperair left	05:53
	→ sunil joined	05:54
	← loincloth left	05:56
	→ loincloth joined	05:56
	← schmick1 left	05:58
	← giskard left	05:58
	← loincloth left	06:04
infid	trying to edit the commit message before last but 'git rebase -i HEAD~2' is saying fatal: needed a single revision. invalid base	06:07
	→ takeshita_kenji joined	06:08
takeshita_kenji	Hello. I'm having a problem caused by a machine failure during a commit and push.	06:09
	When I try to do a "git pull", I get this: object 25f89f222346009414e4f4458c17796343e37326 is corrupted	06:09
	When I try to do a "git status", I get this: fatal: index file smaller than expected	06:09
	How can I fix this without having to copy over the entire 20 GB repository?	06:09
	← robinr left	06:10
	← rado1 left	06:11
takeshita_kenji	I also get "invalid HEAD" errors.	06:11
	Google doesn't seem to have much useful information here.	06:13
	← madewokherd left	06:14
	← jjuran left	06:14
	← hasu left	06:15
	→ jjuran joined	06:15
	← Pupeno left	06:16
takeshita_kenji	Anyone?	06:16
	→ schmick1 joined	06:17
	← meanburrito920 left	06:18
takeshita_kenji	I don't even know if anyone's ever seen this. Google sure is useless here.	06:18
	→ cbreak_work joined	06:18
	→ keyvan joined	06:19
	→ aspotashev\|eeepc joined	06:20
	← diegoviola left	06:20
mugwump	takeshita_kenji: usually you just need to remove the corrupted file	06:22
	you can point git to another object using git replace	06:22
	if requireed	06:22
takeshita_kenji	What can I do about this?: fatal: No such ref: HEAD	06:23
	→ Adaptee joined	06:23
mugwump	take a look under .git, is there a 'HEAD' or 'head' file ?	06:23
takeshita_kenji	Yes.	06:23
	It contains this: ref: refs/heads/master	06:24
	The fire refs/heads/master is empty.	06:24
	← merlin83 left	06:24
mugwump	ok, you can remove that file	06:25
	→ prgmatic joined	06:25
mugwump	it should show up on git reflog refs/heads/master or git fsck --lost-found	06:25
	← jvraoltc left	06:25
takeshita_kenji	notice: HEAD points to an unborn branch (master)	06:25
	This is all that's printed after that: fatal: object 25f89f222346009414e4f4458c17796343e37326 is corrupted	06:26
	On the remote end, that object is not corrupted.	06:26
	← altogether left	06:26
	← Remixman left	06:26
takeshita_kenji	Can I just copy it over?	06:27
mugwump	normally local corrupt objects will be found in .git/objects/25/f89f...	06:27
takeshita_kenji	Oh.	06:27
	What do I do about this one?	06:27
mugwump	you can remove that if it says it's bad	06:27
takeshita_kenji	Okay.	06:27
mugwump	especially if you've already confirmed it's ok on the other end	06:28
	→ eMyller joined	06:28
eMyller	'lo people	06:28
	error: cannot run : No such file or directory	06:29
	error: There was a problem with the editor ''.	06:29
	Please supply the message using either -m or -F option.	06:29
	why am i getting this error message? :\	06:29
mugwump	eMyller: probably your $EDITOR or some other variable is set to empty string	06:29
eMyller	nano (the editor git aways used) is just fine	06:29
	← toxygen left	06:30
eMyller	mugwump: where is that set? .<shell>rc?	06:30
	← prgmatic left	06:30
	→ toxygen joined	06:30
mugwump	could be, try : env \| grep EDITOR - that will tell you if it's the environment	06:30
	git var -l \| grep EDITOR - tells you what git thinks	06:30
eMyller	yea, it's empty	06:30
mugwump	should be unset, empty makes no sense	06:31
eMyller	also empty.	06:31
mugwump	empty = no output?	06:31
eMyller	gonna set it in zshrc, thanks a lot :)	06:31
	no, EDITOR=	06:31
	and GIT_EDITOR=	06:31
takeshita_kenji	Okay, now it says this: fatal: object 549b8312f3e36af97bdda2484f566b76a3260c02 is corrupted	06:31
mugwump	right, yeah that's not so good though git could probably behave better	06:31
	← herdingcat left	06:31
takeshita_kenji	(That's with git pull.)	06:32
mugwump	takeshita_kenji: do you have shell access remotely? if so try git fsck --full	06:32
eMyller	mugwump: editor is working again, thanks :)	06:32
	→ mtkd joined	06:32
takeshita_kenji	Trying it.	06:33
	← timClicks left	06:33
	← toxygen left	06:34
takeshita_kenji	The repository is quite large, so it'll take a while for the check to run.	06:35
	← floppyears left	06:36
	→ floppyears joined	06:36
	← girishr left	06:36
	→ ph^ joined	06:38
	→ lucsky joined	06:38
	→ merlin83 joined	06:40
	← merlin83 left	06:40
	→ merlin83 joined	06:40
	→ toxygen joined	06:43
	← amerine left	06:43
	← rayners left	06:43
	→ tvw joined	06:44
	← multigrain left	06:45
	→ yashi joined	06:46
	→ dj_tjerk joined	06:47
	→ bentob0x joined	06:47
	→ multigrain joined	06:47
	→ nayankk joined	06:49
nayankk	Hello Git!..	06:50
coppro	same to you, bub	06:50
	→ giallu joined	06:50
	→ trivol joined	06:52
mugwump	takeshita_kenji: if it succeeds, the easiest option is to re-clone	06:53
takeshita_kenji	I wouldn't say that'd be the easiest option.	06:53
	← eMyller left	06:53
takeshita_kenji	Last time I tried that, I had to clone and tar the resulting clone, then split it into 50 MB files, then copy the 50 MB files over, reassemble, and extract.	06:53
	Even after that, there was some corruption that had to be handled by a pull.	06:54
mugwump	otherwise you can repair single objects using, eg "ssh host 'GIT_DIR=/path/to/foo.git git cat-file blob BLOBID' \| git hash-object -w -t blob --stdin"	06:54
takeshita_kenji	That'd probably be easier.	06:54
	Ilari wanders in...	06:54
mugwump	replace 'blob' with whatever type of object is broken	06:54
	→ Pupeno joined	06:54
mugwump	that sounds quite painful	06:54
takeshita_kenji	(The whole non-bare repository is about 40 GB.)	06:54
	← villemv left	06:55
takeshita_kenji	(Bare is 20 GB.)	06:55
mugwump	hmm	06:55
	← Pupeno left	06:55
mugwump	that's pretty big really	06:55
takeshita_kenji	You can imagine how painful it was to convert from Subversion to git, even with git-svn.	06:55
mugwump	sure	06:55
takeshita_kenji	Okay, the remote machine is all fine and good.	06:55
Ilari	takeshita_kenji: Firstly, you need to move the corrupt object aside: 'mv .git/objects/54/9b8312f3e36af97bdda2484f566b76a3260c02 .git/549b8312f3e36af97bdda2484f566b76a3260c02.bak'.	06:55
takeshita_kenji	No errors.	06:55
mugwump	^^ Ilari's comment is also right, I suggested using rm - but that will be required for the repair of the object to work	06:56
	← fahadsadah left	06:57
mugwump	what works a lot better is to split up the svn repository into many git repositories	06:57
	eg, using svn-fast-export-all	06:57
takeshita_kenji	Unfortunately, even one directory in this repository is nearly all of the repository.	06:57
mugwump	repositories that size are very unweildy	06:57
takeshita_kenji	It's for my webcomic.	06:57
coppro	O_o	06:57
	no wonder	06:57
	all those binary files	06:57
mugwump	right	06:57
Ilari	takeshita_kenji: After that, use 'git fsck' (no --full might make it bit faster) to find if there are more corrupt loose objects.	06:58
	→ scarabx joined	06:58
	→ willem joined	06:59
Ilari	With Binary files, the nature of file makes a difference. Uncompressed JRSRs (~6MB apiece) are much nicer than compressed JRSRs (~900kB apiece)...	06:59
	← tedoc2000 left	07:00
	→ fahadsadah joined	07:00
takeshita_kenji	They're pretty much all PNGs.	07:00
coppro	Ilari: Why is that?	07:00
	← willem left	07:00
	→ willem joined	07:00
takeshita_kenji	When I tried running that, I got: fatal: git cat-file 549b8312f3e36af97bdda2484f566b76a3260c02: bad file	07:00
	The file does exist remotely.	07:00
Ilari	coppro: The uncompressed ones delta nicely. Also, if you want to version OpenDocument files, use the flat format.	07:00
	takeshita_kenji: Copy that file from remote, it should be compatible.	07:01
	← treak007 left	07:01
Ilari	takeshita_kenji: The '54/9b8312f3e36af97bdda2484f566b76a3260c02' file.	07:01
takeshita_kenji	Right.	07:01
coppro	ah, that makes sense. As for ODs, what do you mean flat? The uncompressed version?	07:01
Ilari	.fodt, .fods and those. They are single uncompressed XML files.	07:02
coppro	ah	07:02
	didn't know that	07:02
	neat	07:02
	← scarabx left	07:03
Ilari	coppro: I have got repo that has 120 JRSRs (~6.5MB apiece) plus 90 versions of .fods file that in the end is about 100kB. Total packed size is about 11.7MB.	07:05
coppro	wow, nice	07:05
	← project2501a left	07:06
cbreak_work	text compresses well	07:06
coppro	yeah	07:06
cbreak_work	extra verbose text like xml especially so	07:06
thiago_home	Ilari: I wish all .fod* files were that good	07:06
	→ project2501a joined	07:06
thiago_home	OOo often mis-saves .fodp	07:06
	when you reopen, there's formatting lost	07:06
Ilari	Eh, its supposed to be native format...	07:07
thiago_home	"supposed" being the key word here	07:07
	→ kingfishr joined	07:07
thiago_home	I also noted that it has whitespace issues	07:07
	takeshita_kenji wrote a Python script to make the copy fix easy.	07:07
Ilari	Well, from the fact that project has JRSR files, one can guess what type of project it is... :-) And its not compressiblity. Just deflating them compresses to about 1MB apiece.	07:08
	← keystr0k left	07:09
	→ petrux joined	07:09
	← typemore left	07:09
	← harinath left	07:10
Ilari	Majority of that 6.5MB is ASCII-coded binary gunk.	07:10
	→ radarek joined	07:10
cbreak_work	It's too bad people don't use LaTeX more instead of "Word Processors"	07:11
	→ giskard joined	07:11
cbreak_work	documents would be a lot more beautiful	07:11
	and there would be much less to read of them	07:11
	← aspotashev\|eeepc left	07:12
takeshita_kenji	Okay, now we have some more information.	07:13
	→ priidu joined	07:13
takeshita_kenji	Some objects in pack have checksum mismatches, among other things.	07:13
Ilari	Uh, oh...	07:13
	takeshita_kenji: How did you get that repository corrupted?	07:14
takeshita_kenji	System hardlocked.	07:14
	(I fixed that problem earlier today.)	07:14
	It must've locked right when the git commit was going.	07:14
	← stringoO left	07:14
takeshita_kenji	It's hard to recover from something like that.	07:15
Ilari	That doesn't explain corrupt packs...	07:15
	← tchan left	07:15
takeshita_kenji	Well, I did do a repack a short while before the hardlock, so maybe the filesystem didn't flush everything yet.	07:15
	← Mage left	07:16
	→ stringoO joined	07:16
	→ tchan joined	07:16
	← Tricon left	07:16
Ilari	Corrupt packs are a lot nastier to repair than corrupt loose object files.	07:17
takeshita_kenji	The system became unstable after a massive upgrade (Gentoo) a few weeks ago. It hardlocked like that for a little while until I booted without two of the RAM sticks in.	07:17
	→ Tricon joined	07:17
Ilari	Ah, RAM corruption. That would explain it...	07:18
takeshita_kenji	It's been stable enough now, but I have to deal with the mess it caused.	07:18
	How can I fix the pack files	07:18
	→ hyperair joined	07:18
takeshita_kenji	*?	07:18
	← coppro left	07:19
Ilari	takeshita_kenji: One way would be to move the corrupt pack aside (move it aside! not delete!) and then run recovery unpack on it (if you leave it in place, it won't unpack).	07:19
hyperair	recovery unpack?	07:20
	what's that?	07:20
Ilari	takeshita_kenji: (git unpack-objects -r <moved-aside.pack). After that, 'git fsck --full'.	07:20
takeshita_kenji	Yeah, what's that?	07:20
	Oh.	07:20
	← rryouumaa left	07:20
	→ g3d joined	07:20
	← bentob0x left	07:21
	← dytrivedi left	07:21
	← kpreid_ left	07:21
takeshita_kenji	Ah, now I remember why I had to manually copy it over: error: inflate: data stream error (incorrect data check)	07:21
Ilari	takeshita_kenji: And don't delete the pack even after doing that, there may be more data salvageable after getting some missing objects.	07:21
takeshita_kenji	Yep, I get a bunch of those when running unpack-objects.	07:23
	Is that to be expected?	07:23
Ilari	takeshita_kenji: Can you pastebin segment from output?	07:23
takeshita_kenji	http://pastebin.com/49Z1A8mz	07:24
	→ sgh joined	07:24
	← munichlinux left	07:24
	← Heimidal left	07:25
takeshita_kenji	Well, it failed. I updated the Pastebin entry to show everything.	07:25
Ilari	takeshita_kenji: Updated? I only see four pairs or errors...	07:26
takeshita_kenji	Refresh the page.	07:26
	"fatal: Out of memory, malloc failed"	07:26
	→ mastro joined	07:26
	← hyperair left	07:27
Ilari	takeshita_kenji: Well, you can try using 'git fsck --full' to see how much you are missing...	07:27
	→ nevyn^ joined	07:27
	← __skyl__ left	07:28
	→ gaveen joined	07:28
	→ galderz joined	07:29
	← multigrain left	07:29
takeshita_kenji	There are 12341 missing blobs.	07:30
	← cloudhead left	07:30
	→ DavidKlein joined	07:30
Ilari	takeshita_kenji: Eek.	07:31
	→ villemv joined	07:31
	→ ProperN[out] joined	07:32
	← ProperNoun left	07:33
	ProperN[out] → ProperNoun	07:33
	← alireza_ left	07:33
	→ hyperair joined	07:34
	→ d0ugal joined	07:34
	→ juice_d joined	07:34
takeshita_kenji	It appears that the old script I used to do the reassembly is still there.	07:34
	It would be nice if a simple tar < *.part worked, but there are still data errors doing that.	07:35
	→ froschi joined	07:35
	→ NET\|\|abuse joined	07:36
	→ d0k joined	07:37
	← NET\|\|abuse left	07:38
	→ aspotashev\|eeepc joined	07:40
	→ WHL joined	07:40
WHL	Gratis Spotify Premium account: http://www.gratispremium.com/?ref=433384	07:41
	← WHL left	07:42
	→ LuckySMack joined	07:44
LuckySMack	adding the line "tmp/" to my .gitignore file will ignore any files/changes in the tmp/ dir right? when adding files and commiting, git status still shows changes to those files.	07:45
	→ villemv_ joined	07:46
doener	LuckySMack: it will ignore untracked files in any directory called tmp/	07:46
	→ sh1mmer joined	07:46
	→ priidu_ joined	07:46
doener	LuckySMack: .gitignore has no meaning for tracked files	07:46
	← priidu left	07:46
	→ shruggar joined	07:47
	→ pantsman joined	07:47
LuckySMack	ok, then is there a way to make git ignore changes in certain directories then? I have another directory as well I want git to ignore that it still seems to be tracking. or do I have to not use "git add ." and add individual files instead?	07:49
cbreak_work	just don't track it...	07:49
	you can stop tracking stuff with git rm --cached	07:49
	← villemv left	07:49
cbreak_work	git add will not add ignored files unless you force it to	07:49
	← jowls left	07:51
	← hyperair left	07:53
	→ hyperair joined	07:53
	→ consolers joined	07:55
LuckySMack	ok so "git rm" will untrack a file until you add it again with "git add", but "git rm --cached" will remove it and not re track it unless you force it with "git add"?	07:56
	← dj_tjerk left	07:56
LuckySMack	I thought this was the reason for .gitignore, so make add ignore files unless calling them directly	07:56
	*adding them directly	07:56
DrNick	you can't ignore files that are already a part of the repository	07:57
	→ pigeon joined	07:58
DrNick	and git rm deletes files as far as git is concerned; whether they're still on disk afterwards is determined by the presence or absence of the --cached parameter	07:58
LuckySMack	well, then whats the best way to have certain files not change when updating file locally to a remote production branch? for example I have a database.php which has the database setting, which are different locally	07:58
DrNick	"Don't Do That"	07:58
	← jjuran left	07:58
	← stringoO left	07:59
LuckySMack	yea i saw that git rm deletes the files. I didnt want that	07:59
	→ Paraselene__ joined	08:01
	→ Swistaczek joined	08:01
wereHamster	LuckySMack: you don't track files which are different between branches or repos	08:05
	you ignore them, and instead provide a template which the user can rename and adjust	08:05
	← d0ugal left	08:05
	→ naeu joined	08:05
	→ forvan joined	08:05
	→ d0ugal joined	08:06
	← cilly left	08:09
	→ drako joined	08:09
	→ Nickeeh joined	08:09
	← takeshita_kenji left	08:10
	← entropi left	08:10
	→ takeshita_kenji joined	08:12
	← takeshita_kenji left	08:12
	← Tricon left	08:12
	→ ntoll joined	08:13
	→ jjuran joined	08:17
	→ _tydeas_ joined	08:17
_tydeas_	Hello ppl	08:17
	← Nickeeh left	08:20
	← ilteris_ left	08:21
	→ ilteris_ joined	08:22
	→ raichoo joined	08:23
_tydeas_	I am going to mirror a project from code.google.com to github	08:23
	← pantsman left	08:24
_tydeas_	the author file will be created after i have `git svn fetch` ?	08:24
	I am following this proccedure http://www.fnokd.com/2008/08/20/mirroring-svn-repository-to-github/	08:24
	← ilteris_ left	08:26
	→ GodEater_ joined	08:28
	← GodEater_ left	08:28
	→ GodEater_ joined	08:28
	← GodEater_ left	08:28
	← radarek left	08:29
	→ Belna joined	08:31
charon	_tydeas_: no, the author map is a mapping from svn usernames to Realname [email@hidden.address] and must be provided in advance	08:32
	← aspotashev\|eeepc left	08:32
charon	(otherwise, git-svn will substitute a dummy)	08:32
_tydeas_	what do you mean?	08:32
	i fetch and then create the .authors file like it says http://technicalpickles.com/posts/creating-a-svn-authorsfile-when-migrating-from-subversion-to-git/ ?	08:33
charon	for example, in many svns my username is 'charon', so if you just run 'git svn clone svn://example/com/' all my commits will seem to come from a user 'charon <charon@repo_uuid>'	08:33
	so what i instead do is create an author map with a line saying 'charon = Thomas Rast <...>'	08:34
	→ benjo2 joined	08:34
charon	and then run git svn clone -A author_map svn://..../	08:34
_tydeas_	check above link ^ this is what i am going to do	08:34
charon	(of course you can also do a two-step init/fetch)	08:34
	yeah, so instead of: git svn init -T http://svn.codehaus.org/jruby/trunk/jruby/	08:34
	make an .authors file (or whatever you want to call it) in the format i said	08:35
	then do: git svn init -A .authors -T http://svn.codehaus.org/jruby/trunk/jruby/	08:35
	← Swistaczek left	08:35
	← consolers left	08:35
	← resmo left	08:35
_tydeas_	but how can i the authorfile from an svn repo?	08:35
charon	you can't, svn doesn't have this information	08:35
	it's there precisely to provide information that SVN doesn't	08:35
_tydeas_	aha	08:35
charon	well, there's a neat hack, let me check	08:36
	→ dpino joined	08:36
charon	http://thomasrast.ch/git/git-svn-conversion.html under "generating the author map" there's a three-line shell hack that generates a dummy list	08:36
	containing all usernames in your svn repo	08:36
	if you want to start from that instead of a blank file	08:36
	→ `Assassin joined	08:37
	← froschi left	08:38
`Assassin	Is the developer of pianobar here?	08:38
wereHamster	`Assassin: what does pianobar have to do with git?	08:39
`Assassin	Irrelevant.	08:39
	← uau left	08:39
	→ uau joined	08:39
	← ExtraSpice left	08:40
	→ unknown joined	08:40
	← mooosh left	08:41
	→ aspotashev\|eeepc joined	08:41
	← codeshepherd left	08:41
	→ oriba joined	08:41
_tydeas_	charon: i article you gave my says `It assumes that you want to do a one-way conversion, i.e., after the conversion you do not want to use the SVN repository any more.`	08:42
	this is not had to do with the updating	08:42
	→ kpreid_ joined	08:45
	→ mooosh joined	08:46
	→ dytrivedi joined	08:47
	→ codeshepherd joined	08:48
	→ aspotashev joined	08:49
	← aspotashev\|eeepc left	08:49
	← johnw left	08:51
	→ froschi joined	08:51
	← kpreid_ left	08:52
	→ Belna_ joined	08:53
	→ carranca joined	08:53
	← oriba left	08:54
	→ smer joined	08:54
	→ kpreid_ joined	08:54
	→ tatsuya joined	08:54
	← Belna left	08:56
	← fahadsadah left	09:00
	→ textarcana_ joined	09:00
	→ jony1 joined	09:01
	← jony left	09:01
smer	newbie question: how do I create a repos online via ssh (want to work on with others!)?	09:02
rudi_s	smer: If you trust them, just setup a shared ssh account, upload the repository there (scp for example) and clone from it/push to it, git clone user@server:path/to/repo	09:05
	→ jds joined	09:06
	← freshtonic left	09:06
rudi_s	If you don't, setup up a public repository (http, git server, ..) and let others clone from you. If they send you patches or pull requests, you merge them and push to your repo.	09:06
	→ curtana joined	09:08
	→ fahadsadah joined	09:08
	← codeshepherd left	09:08
smer	@rudi_s Thanks, sounds good. Will after cloning from server every commit be commited to server online then?	09:09
charon	_tydeas_: huh?	09:09
	_tydeas_: my comments on the author map apply regardless of whether it's a one-way conversion	09:10
	i merely wrote the linked article in the context of such	09:10
	→ radarek joined	09:11
	← hyperair left	09:11
	→ hyperair joined	09:12
	→ bluenovember joined	09:12
	→ [Po]lentino joined	09:13
	→ voxpelli-laptop joined	09:13
	→ emanuelez joined	09:14
rudi_s	smer: Old commits are. New ones not immediately, only after a git push to the server.	09:15
emanuelez	hello... I'm having problems with a rebase and CRLF :( I use linux but other developers use windows. if i try to rebase from my personal branch to master (remoted to the public repo) i get file staged in the moddle of the rebase! any hint?	09:16
	→ resmo joined	09:17
	→ codeshepherd joined	09:17
	→ Titosemi joined	09:18
_tydeas_	charon: have not understood Custom layouts	09:20
charon	_tydeas_: is your svn repo organized in trunk/ branches/ tags/ as most of them are?	09:21
_tydeas_	charon: it's not mine but i suppose	09:21
cbreak_work	emanuelez: sounds like you have a conflict. resolve it.	09:21
	git rebase should tell you what to do on rebase conflicts	09:21
	essentially, your options are git rebase --abort, or fixing it, git add the fix, git rebase --continue	09:22
charon	_tydeas_: then don't worry about that section	09:22
	→ ispeak joined	09:22
_tydeas_	charon: it orgnized	09:22
	charon: use -s?	09:22
emanuelez	cbreak_work: I wish that was the case :( http://pastebin.org/302431	09:22
charon	_tydeas_: yes	09:23
	g3d → posciak	09:23
smer	@rudi_s Thanks, I'm trying...	09:23
rudi_s	np (btw. no @ in irc, use nick: ..)	09:23
_tydeas_	charon: so i have author-map	09:24
	charon: now time to clone ?! :)	09:24
	think this will take time	09:24
cbreak_work	emanuelez: you can't fix the conflict?	09:24
	← yason left	09:25
	← kpreid_ left	09:25
_tydeas_	charon: should i keep metadata?	09:25
charon	_tydeas_: if you do a conversion, no. if you want to work with the svn repo, yes.	09:26
	→ kpreid_ joined	09:26
	← sh1mmer left	09:26
emanuelez	cbreak: did u see my pastebin? it's not a conflict. as u can see in both branches i don't have any file in the stage area, everything is committed. thei try to merge and it tells me there's a couple of files in the staging area... it tells nothing about conflicts. it's a problem with CRLF	09:26
	← froschi left	09:26
emanuelez	cbreak_work: : did u see my pastebin? it's not a conflict. as u can see in both branches i don't have any file in the stage area, everything is committed. thei try to merge and it tells me there's a couple of files in the staging area... it tells nothing about conflicts. it's a problem with CRLF	09:26
_tydeas_	charon: svn is on google code and i want to make a read-only mirror to git hub	09:26
	← smer left	09:26
_tydeas_	so i think no metadata right?	09:27
	→ Swistaczek joined	09:27
cbreak_work	emanuelez: line 59 and 60	09:27
charon	_tydeas_: right	09:27
cbreak_work	files that git can't merge during a rebase are left in the working directory for you to resolve their conflicts	09:28
	→ kumbayo joined	09:30
_tydeas_	git svn clone -A author-map --no-metadata -s http://... will do i quite clone?	09:30
	where there should be listing the copy	09:30
emanuelez	cbreak_work: yes, i know... i do that all the time, but now if I go to the mentioned file it does not even have the <<< stuff. so even if i resolve the conflists, add the files and rebase --continue it will come with the same error again	09:30
cbreak_work	what does git diff say?	09:30
charon	_tydeas_: note that with -s, you need to remove the trunk/ part of the URL	09:31
_tydeas_	Using higher level of URL: http://yii.googlecode.com/svn/trunk => http://yii.googlecode.com/svn	09:31
	got this	09:31
	from utpout	09:31
emanuelez	cbreak_work: well, the situation evolved since that pastebin. so now i have some proper diff.	09:32
_tydeas_	*output	09:32
charon	ah well, then maybe git-svn fixed it by itself	09:32
emanuelez	let me do another pastebin with the current situation	09:32
_tydeas_	this is a prompt or the git understood ?	09:32
	← roop left	09:32
	← urkud_ left	09:32
	← frakturfreak left	09:32
_tydeas_	no the command finished and nothing appears in the folders	09:33
	← tvw left	09:34
emanuelez	cbreak_work: http://pastebin.org/302528	09:34
	→ frakturfreak joined	09:34
	← frakturfreak left	09:34
	← ph^ left	09:35
cbreak_work	and git diff is empty?	09:35
	→ ph^ joined	09:35
	→ salisbury joined	09:35
emanuelez	if I do git add : warning: CRLF will be replaced by LF in src/main/java/hudson/plugins/synergy/SynergySCM.java	09:36
_tydeas_	charon: http://pastie.org/990258	09:36
	charon: check please	09:36
	← LuckySMack left	09:36
cbreak_work	maybe you have CRLF documents in the repo	09:37
	and yours has autocrlf turned on	09:37
	→ airborn joined	09:37
emanuelez	cbreak: that's what i was trying to say :)	09:37
cbreak_work	you can always turn it off	09:37
airborn	hello	09:37
charon	_tydeas_: well, use some reading skills. "Author: (no author) not defined in author-map file" you must define every author when using an author map. (note that (no author) means that there are revisions that you have no permissions to read, but that's svn territory)	09:38
	→ alexpercsi joined	09:38
alexpercsi	hi all	09:38
	is there any way i can revert changes to specific files from a specific commit?	09:39
cbreak_work	I doubt you mean revert.	09:39
emanuelez	cbreak_work: core.autocrlf=input	09:39
cbreak_work	try checkout	09:39
alexpercsi	i know how to rebase interactive and edit a commit	09:39
	but i want to remove changes i made in that commit to certain files	09:39
cbreak_work	alexpercsi: do you really, really want to rewrite history?	09:39
charon	alexpercsi: git revert -n $sha1 && git reset && git add $file && git commit && git reset --hard	09:39
alexpercsi	thanks	09:40
	the answer is yes, i do want to rewrite history	09:40
cbreak_work	then use git rebase -i	09:40
alexpercsi	this was a hasty commit to make something available to someone	09:40
cbreak_work	and edit the commit you want to edit	09:40
charon	in that case using my way (and then rebasing) is kind of round-about	09:40
alexpercsi	and i'm on a topic branch so the rest of my code is safe	09:40
	rebase -i?	09:40
	is that rebase interactive?	09:41
cbreak_work	yes	09:41
alexpercsi	and how can i pull out those changes? I'm liable to miss a few if i do it by memory and gitx does not show them as changed	09:41
cbreak_work	once you're in the edit mode of that commit, you can just git checkout -- filename to get the original...	09:41
	why pull?	09:41
	← jony1 left	09:41
	→ jony joined	09:41
alexpercsi	sorry, i misspoke	09:41
	pull as in take them out	09:42
	remove them	09:42
cbreak_work	all you need to do is undo the changes	09:42
	easiest way is by checking out the version from before you did them	09:42
alexpercsi	alright, i'll try that	09:42
	thanks all :)	09:42
	← Swistaczek left	09:42
_tydeas_	charon: so add Author: before each author?	09:43
charon	_tydeas_: no, define the "(no author)", as in: (no author) = none [email@hidden.address]	09:44
_tydeas_	why that?	09:47
	will everything me assigned to no author now?	09:47
	sorry for newb questions pretty new to git and subverion systems	09:47
	← resmo left	09:49
	← naeu left	09:51
charon	[11:38] <charon> _tydeas_: [...] (note that (no author) means that there are revisions that you have no permissions to read, but that's svn territory)	09:53
	→ roop joined	09:53
	→ girishr joined	09:55
_tydeas_	charon: check please http://pastebin.com/6yZ1Q12g	09:55
	charon: is this an error? what can i do?	09:56
charon	_tydeas_: can you pastebin your .git/config please?	09:56
	and which git version do you use?	09:56
	→ icreatures joined	09:57
	← icreatures left	09:58
	→ icreatures joined	09:58
	← ereslibre left	09:58
_tydeas_	charon: http://pastebin.com/L2f3yQPH	09:58
	git version 1.6.6.1	09:59
	← icreatures left	09:59
	→ ereslibre joined	10:00
	← ereslibre left	10:00
	→ ereslibre joined	10:00
	→ khmarbaise joined	10:01
charon	_tydeas_: not sure if this affects git-svn but you have two copies of 'branches' and 'tags' in the config, try removing the duplicate	10:01
	if that doesn't help, see if 1.7.1 has the same bug	10:01
_tydeas_	if i remove this duplicate what can should i do next?	10:03
	to continue clone	10:03
	→ ive joined	10:03
charon	git svn fetch	10:06
	← alezandro left	10:08
_tydeas_	the repo is created under whereIwouldLikeRepoToBe/svn/.git	10:08
	can i just mv the svn/* ../	10:08
	will work right?	10:08
charon	no, because * doesn't include .git	10:09
	you have to move that too	10:09
_tydeas_	ok thanks	10:11
	→ drizzd joined	10:11
	← kpreid_ left	10:12
	→ kpreid_ joined	10:13
	← kpreid_ left	10:18
	→ kpreid_ joined	10:18
s0enke	in a git pre-commit hook i use "git diff-index --name-only --cached --diff-filter" to get the files that are staged. but that can be bypassed by git commit -a or git commit <file>. Is there a way to get ALL files that are going to be commited?	10:20
wereHamster	maybe ls-files ?	10:21
	→ gebi joined	10:21
	← mithro left	10:24
	→ codeshepherd_ joined	10:24
	→ alecs joined	10:26
alecs	hi!	10:26
	i have pulled a revision from a repository	10:26
	→ whatdoesitwant joined	10:26
	← codeshepherd left	10:27
	codeshepherd_ → codeshepherd	10:27
alecs	i have added some files to gitignore, and now when i want to pull, i get "error: Untracked working tree file 'my file name' would be overwritten by merge. Aborting "	10:27
	← willem left	10:27
	→ froschi joined	10:27
	← radioman-lt left	10:29
	→ parasti joined	10:29
	← kpreid_ left	10:30
	→ rolfb joined	10:30
	→ kpreid_ joined	10:32
	→ ehabkost joined	10:37
	→ radioman-lt joined	10:39
	← radioman-lt left	10:41
	→ steveire_ joined	10:41
	← steveire_ left	10:41
	→ steveire_ joined	10:41
	→ dkastner joined	10:42
	→ ereslibre_laptop joined	10:43
	← ereslibre_laptop left	10:43
	→ ereslibre_laptop joined	10:43
	← ereslibre left	10:43
	→ mbroeker joined	10:44
	→ ExtraSpice joined	10:44
	← steveire left	10:44
	← emanuelez left	10:45
	← ruien left	10:46
	← vmiklos left	10:47
	← whatdoesitwant left	10:47
	→ twoface joined	10:48
twoface	Is it possible to clone a repository without getting .git folders?	10:49
	→ Tricon joined	10:49
	→ aziz joined	10:50
	→ vmiklos joined	10:50
	→ MaNI joined	10:51
	← voxpelli-laptop left	10:51
MaNI	how to delete all local files that are not part of the main repository?	10:51
Arrowmaster	git clean	10:51
	← vmiklos left	10:52
	→ jtauber joined	10:52
ptor	twoface: It sounds like what you want is a snapshot of the files, not a clone of a git repo. If so, look at 'git archive --remote=<repo>' (man git-archive)	10:53
Gitbot	twoface: the git-archive manpage can be found at http://git.or.cz/man/git-archive	10:53
twoface	thanks	10:55
MaNI	Arrowmaster: thanks	10:56
	→ rchavik joined	10:56
	→ munichlinux joined	10:58
	← MaNI left	10:59
	← airborn left	10:59
	← twoface left	11:00
	→ vmiklos joined	11:01
	→ Yuuhi joined	11:02
wereHamster	alecs: that means that the files are untracked in your local repository but tracked in the branch you are trying to merge. And git stops because it doesn't want to overwrite your local files (because, once overwritten, git would have no way to restore the original files which would mean possible data loss)	11:02
	alecs: either delete or move aside the local files	11:02
	← dytrivedi left	11:03
GarethAdams\|Work	..or commit them	11:03
	oh, it's the newly ignored files that are conflicting, ignore what I said	11:04
	← gaveen left	11:07
	→ Nickeeh joined	11:10
	ereslibre_laptop → ereslibre	11:10
	← benjo2 left	11:11
_tydeas_	some using github i have some Fork Queues marked as `Will likely not apply cleanly`	11:11
	what can i do?	11:11
	→ corni joined	11:12
	→ radioman-lt joined	11:14
	← kumbayo left	11:14
AAA_awright	_tydeas_: If it conflicts ask whoever made the patch to pull and fix the patch?	11:15
	← hyperair left	11:15
wereHamster	_tydeas_: fetch the commits and merge locally	11:15
_tydeas_	AAA_awright: ask to pull and fix the patch?	11:16
	i have not understood why this happened	11:16
AAA_awright	Eh, same thing. Someone needs to pull the latest updates.	11:16
	← corni left	11:16
	← ispeak left	11:17
	→ corni joined	11:17
	← Hydroxide left	11:18
_tydeas_	to pull and merge locally i will have to have write access right?	11:19
	→ benjo2 joined	11:19
	← radioman-lt left	11:19
_tydeas_	so i can push?	11:19
_ikke_	_tydeas_: You cannot just push to a normal repository	11:20
_tydeas_	what can i do?	11:21
	→ psoo joined	11:22
_tydeas_	Is there an example how to resolve such a problem? I am worried about how to keep the author details for the commit. Don't want to be me how will pull local and merge	11:24
	← RobertLaptop left	11:27
wereHamster	_tydeas_: you fetch the commit in question to your local repository, then run 'git cherry-pick <commit>'	11:28
	← codeshepherd left	11:28
	→ TheNewAndy joined	11:29
	→ gaveen joined	11:29
	← psoo left	11:30
	→ tobiassjosten joined	11:30
	→ simplechat joined	11:30
	← benjo2 left	11:30
_tydeas_	in question?	11:32
wereHamster	the commit in the fork queue which you want to apply	11:32
	[Po]lentino → [Po]lentino\|away	11:34
	← emmanueloga left	11:36
_tydeas_	which command to use `git fetch ???? `	11:38
wereHamster	git fetch <repo which has the commit> <branch from that repo which contains that commit in its history>	11:38
	→ dytrivedi joined	11:39
	→ radioman-lt joined	11:40
	→ bentob0x joined	11:40
_tydeas_	got master -> FETCH_HEAD	11:40
	used the other users repo and master branch	11:40
	now run git cherry-pick but which commit?	11:40
wereHamster	the commit in the fork queue which you want to apply	11:41
	← Tricon left	11:41
_tydeas_	cherry pick will push it too?	11:41
wereHamster	no, cherry-pick does not push. But git push does push. (how surprising)	11:42
_tydeas_	http://pastebin.com/X0hWMbg9 what i got from the cherry	11:43
	says something about patch	11:43
	and this is the full proccedure	11:44
	and this	11:44
	http://pastebin.com/u0mJM25b	11:44
wereHamster	_tydeas_: read lines 2-6 in your first pastebin	11:45
	→ HG` joined	11:46
_tydeas_	don't understand what i have to do	11:46
wereHamster	_tydeas_: git help tutorial, or http://progit.org/. Read it.	11:46
	→ hebz0rl joined	11:50
	→ ilteris_ joined	11:52
	← eli left	11:52
	← alecs left	11:54
	← jtauber left	11:54
	← carranca left	11:54
	→ jtauber joined	11:54
	← munichlinux left	11:56
_tydeas_	didn't managed to resolve it	11:56
	i have some both modified files	11:56
	should i add them?	11:56
	← hohoho left	11:59
	← curtana left	12:00
	→ loincloth joined	12:01
	→ codeshepherd joined	12:01
	→ codeshepherd_ joined	12:02
	← bentob0x left	12:04
	→ codeshepherd__ joined	12:04
	→ bentob0x joined	12:06
	← codeshepherd_ left	12:06
	← codeshepherd left	12:06
	← nevyn^ left	12:06
	→ codeshepherd_ joined	12:06
	← codeshepherd_ left	12:07
	← codeshepherd__ left	12:08
	← PuffTheMagic left	12:08
	→ brett_h joined	12:08
	→ PuffTheMagic joined	12:08
	→ codeshepherd joined	12:08
	← `Assassin left	12:08
	← ntoll left	12:10
	→ nevyn^ joined	12:10
	← eno left	12:10
	→ voxpelli-laptop joined	12:11
	← jtauber left	12:12
	← dytrivedi left	12:12
	→ eno joined	12:12
	← Titosemi left	12:12
	← codeshepherd left	12:13
	← ereslibre left	12:16
	← TheNewAndy left	12:16
	→ khannz joined	12:18
khannz	Hello, everyone. Can someone tell me, what I'm doing wrong? My trouble-log: http://gist.github.com/423799	12:18
	→ Titosemi joined	12:19
	→ codeshepherd joined	12:19
	JEEBcz → JEEB	12:20
	→ TheNewAndy joined	12:20
	← JEEB left	12:21
	→ JEEB joined	12:21
	→ ereslibre joined	12:21
	← ereslibre left	12:21
	→ ereslibre joined	12:21
wereHamster	khannz: you can't push through anonymous http.	12:21
	khannz: log into github and then copy the http url from the url field	12:21
	→ codeshepherd_ joined	12:21
	← codeshepherd left	12:22
khannz	wereHamster: even if i'm configured token for github?	12:22
wereHamster	which token? ANd where configurde?	12:22
	← mbroeker left	12:23
	← parasti left	12:23
	← loincloth left	12:23
	→ loincloth joined	12:23
khannz	wereHamster: 1 sec	12:23
	← priidu_ left	12:24
khannz	wereHamster: watch "Token config" section here http://help.github.com/git-email-settings/	12:24
	→ mbroeker joined	12:25
wereHamster	git doesn't have any special code for the github website. It doesn't know what to do with github.token, it ignores that config field completely	12:25
	→ flazz joined	12:25
mnemoc	they have a helper for that iirc	12:26
wereHamster	github.token/github.user is only used in some tools which are specific to the github website.	12:26
	← codeshepherd_ left	12:26
khannz	wereHamster: thanks for explanation! now what i need to do with R+W-access-availiable-url?	12:27
wereHamster	khannz: log into github and then copy the http url from the url field	12:27
khannz	wereHamster: copied, but what's next?	12:27
wereHamster	git config remote.origin.url <the url>	12:28
khannz	wereHamster: aha! thank you, good person!	12:28
	Working great now! Thanks, community and wereHamster again =)	12:32
	← khannz left	12:32
	→ Nightsorrow joined	12:34
	→ coderdad joined	12:34
	→ Joeconyers joined	12:35
	→ codeshepherd joined	12:37
	← raichoo left	12:37
	→ codeshepherd_ joined	12:40
	→ priidu joined	12:40
	→ hasu joined	12:41
	← codeshepherd_ left	12:42
	→ codeshepherd_ joined	12:42
	← codeshepherd left	12:42
	codeshepherd_ → codeshepherd	12:42
	← rsteig left	12:43
	→ manyoso joined	12:43
	→ curtana joined	12:45
	← TTilus left	12:45
	→ eli joined	12:46
	→ codeshepherd_ joined	12:46
	← codeshepherd left	12:46
	codeshepherd_ → codeshepherd	12:46
	→ IslandUsurper joined	12:46
	→ rsteig joined	12:47
	← coderdad left	12:47
	→ mw joined	12:47
	← mw left	12:47
	→ mw joined	12:47
	→ coderdad joined	12:47
	→ yopp joined	12:48
yopp	Hi!	12:48
	is it possible to migrate from gitosis to gitorous?	12:49
	Belna_ → Belna	12:52
charon	yopp: the repos, of course, just clone --mirror the old repo and then push --mirror to the new one, repeat for every repo. the auth data is probably slightly harder	12:52
yopp	uh/	12:52
	I have about 50 repos	12:53
krunaldo	script it?	12:53
yopp	um.	12:55
	→ calin joined	12:55
	→ raichoo joined	12:55
	→ Yan_nick joined	12:56
	→ altogether joined	12:57
	→ aschiwi joined	12:59
	← aschiwi left	12:59
	→ bcardarella joined	13:01
	← Yan_nick left	13:01
	→ Yan_nick joined	13:03
	← Nickeeh left	13:03
	← nevyn^ left	13:04
	→ nevyn^ joined	13:04
	→ frakturfreak_ joined	13:05
	→ alezandro joined	13:07
	← Yan_nick left	13:07
	→ jensn joined	13:08
	→ scarabx joined	13:09
	← calin left	13:09
	[Po]lentino\|away → [Po]lentino	13:09
	→ calin joined	13:10
	→ mase_home joined	13:10
	→ nanotube joined	13:10
	→ q0rban\|lunch joined	13:11
	← felipec left	13:11
	→ JeffJohnson joined	13:12
JeffJohnson	howdy	13:12
	→ Textmode joined	13:12
JeffJohnson	how can I reset an branch directory(that has commited branch changes) to the master version of this directory?	13:12
wereHamster	checkout master -- directory/	13:13
	← sylr left	13:13
	→ sunbright joined	13:14
	← q0rban left	13:14
JeffJohnson	wereHamster: thx, whats "--" in the cmd?	13:15
wereHamster	a delimiter so taht git knows to treat master as a branch name (actually, a <tree-ish>) and directory/ as a pathspec	13:15
JeffJohnson	ok thx	13:16
	→ coderdad_ joined	13:16
	← gebi left	13:18
	← coderdad left	13:18
JeffJohnson	it not really works :) I still have the same files in my directory	13:18
infid	i'm trying to edit the commit message before last but 'git rebase -i HEAD~2' is saying 'fatal: needed a single revision. invalid base'. Anyone have any idea why?	13:19
	← altogether left	13:20
charon	perhaps you only have two commits in total?	13:21
doener	JeffJohnson: do you mean that nothing was changed, or that new files haven't been removed?	13:21
	← roop left	13:21
JeffJohnson	doener: new files are still there	13:21
	doener: the dir still looks like the HEAD version :)	13:22
	← opalka left	13:22
	→ jherdman joined	13:22
	← codeshepherd left	13:22
	→ hohoho joined	13:23
doener	JeffJohnson: so I take it that the difference between master and HEAD is that only new files have been added, but no existing files have been modified? Because you basically answered my "this or that" question saying "both"...	13:23
	→ paddor joined	13:24
	→ urkud_ joined	13:24
	→ hohum joined	13:25
infid	charon: yeah i do only have two commits total	13:25
doener	JeffJohnson: anyway, yeah, checkout only adds/replaces files. You could use "git reset master -- directory/", which should remove the respective index entries, causing the files in the working tree to be untracked. And then you could use "git clean -f directory/" to get rid of those untracked files	13:26
infid	charon: it's my initial commit message i want to edit	13:26
	← d1zzy left	13:26
	← Nightsorrow left	13:26
	← jony left	13:26
	← girishr left	13:27
	→ jony1 joined	13:27
charon	infid: unfortunately that's a bit trickier because git-rebase can't rebase onto nothingness. you need to do something along the lines of making a new branch at HEAD^, use git commit --amend, then rebasing (in this case cherry-picking will do) the old commits except the root on top of it.	13:27
JeffJohnson	doener: thx :)	13:27
	→ danmatthews joined	13:27
JeffJohnson	doener: hrm but I cant commit the reseted dir, how do I do this?	13:28
	← bin10101 left	13:28
doener	JeffJohnson: hm? Just "git commit" should do, the reset command already adjusted the index	13:28
JeffJohnson	doener: mmh its works with git commit -a, but not with git commit $path	13:28
	hrm	13:28
	→ TTilus joined	13:29
	→ MattDiPasquale joined	13:30
	→ Nemurenai joined	13:30
infid	charon: hmm sounds like it'd be easier to start over almost	13:30
charon	really?	13:31
doener	git checkout HEAD^; git commit --amend; git rebase --onto HEAD master^ master	13:31
charon	if you're confident with reflog and a detached HEAD, that translates to: git checkout HEAD^ && git commit --amend && git cherry-pick master && git branch -f master	13:31
doener	starting over can't be a lot easier than that	13:31
infid	ok	13:31
charon	(hmm, i shouldn't change my mind mid-sentence and not use the reflog after all)	13:31
	→ TeckniX joined	13:32
charon	ah, tack on 'git checkout master' in my scheme	13:32
	← scarabx left	13:33
	→ d1zzy joined	13:33
	← q0rban\|lunch left	13:33
	← calin left	13:34
	→ Nickeeh joined	13:36
	→ bin10101 joined	13:36
paddor	How to get the content of a file from a specific date on STDOUT? (without checking it out in the working tree)	13:36
	← nayankk left	13:36
	← Determinist left	13:37
patrikf	paddor: git show <tree-ish>:path	13:37
	← slide left	13:37
patrikf	paddor: and you can map a date to a commit with git rev-list -n 1 --before/--after (if you want to script it, otherwise you'd probably just use git log)	13:39
	← Digitalos left	13:39
	→ t0rc joined	13:40
	→ drubies joined	13:41
	← paddor left	13:41
	→ paddor joined	13:42
	← Nickeeh left	13:42
	→ tizbac joined	13:42
	→ yrlnry joined	13:43
sgronblo	hmm, how do you move a branch again?	13:45
paddor	patrikf: Thanks. Isn't there a feature in Git which allows something like HEAD@{yesterday} ?	13:45
_ikke_	sgronblo: What do you mean by moving a branch	13:45
yrlnry	I was reading somewhere that after a failed merge, the index contains four instances of a conflicted file: two versions from the parent commits, a version from the common ancestor, and a version with conflict markers. What happens when the merge has more than two parents?	13:45
	← raichoo left	13:45
	← paddor left	13:45
yrlnry	sgronblo: checkout branch; reset commit	13:46
sgronblo	_ikke_: not like git branch -m. I made some commits to master instead of creating a topic branch for them first. then checked out the topic branch where master was, then want to move master back to where git-svn is.	13:46
yrlnry	sgronblo: checkout master; reset git-svn	13:46
	→ paddor joined	13:46
yrlnry	sgronblo: or in theory you can use update-ref.	13:47
	→ cannonball joined	13:49
	→ schulty joined	13:49
infid	paddor: yeah, the 'date spec', see half way down http://book.git-scm.com/4_git_treeishes.html	13:50
	→ raichoo joined	13:51
	← Lgb left	13:53
sgronblo	yrlnry: should i use reset --hard?	13:53
yrlnry	sgronblo: it doesn't much matter, but if you do, make sure you have nothing important in the working tree first.	13:55
	→ Lgb joined	13:56
sgronblo	yrlnry: since i already checked out the topic branch at the tip, and havent made any changes, they should remain even if i reset --hard git-svn after checking out master?	13:57
	← npouillard left	13:57
yrlnry	yes, you can always return to the topic branch when you want it.	13:58
	← Adaptee left	13:58
sgronblo	but if i do plain reset (--mixed) then i will have the changes made by the three commits in the topic branch, as unstaged changes right?	13:59
	→ altogether joined	14:00
	→ kukks joined	14:00
sgronblo	yrlnry: ok i think it worked as expected with --hard. thanks for the help.	14:01
yrlnry	You're welcome.	14:01
infid	paddor: you might also find git grep a faster way to find what you're looking for	14:02
patrikf	paddor: ... or the "pickaxe" feature (git log -S)	14:03
	→ chevdor joined	14:03
patrikf	(or git log -p <file>)	14:04
	frakturfreak_ → frakturfreak	14:05
XDS2010	could anyone do a compile ?	14:06
	→ masterkorp joined	14:06
	→ Mpenz joined	14:06
	→ aresnick joined	14:08
infid	XDS2010: can you be a little more specific	14:08
	← hohoho left	14:08
XDS2010	looking for someone willing to compile a kernel with 2 patches that are not commited , compensation available	14:08
	→ brez- joined	14:09
	← paddor left	14:09
XDS2010	infid ^'	14:10
	→ jrmuizel joined	14:10
	← bin10101 left	14:10
	→ bin10101 joined	14:11
	infid is sorry i asked	14:12
	→ bin10102 joined	14:12
	← alexpercsi left	14:14
	→ parasti joined	14:14
	→ munichlinux joined	14:14
	→ paddor joined	14:15
	→ altrux joined	14:15
	← bin10101 left	14:16
	→ ProperN[out] joined	14:17
wereHamster	XDS2010: are you too lazy to download the kernel sources or install a compiler or why do you ask here?	14:17
	← ProperN[out] left	14:17
	→ ProperN[out] joined	14:17
doener	gcc 0-day? :-D	14:17
	→ ntoll joined	14:18
	→ q0rban joined	14:19
	← q0rban left	14:20
	← ProperNoun left	14:21
	ProperN[out] → ProperNoun	14:21
	← rseifert left	14:21
	→ tedoc2000 joined	14:21
	→ hebz0rl_ joined	14:21
	← sgh left	14:22
	← hebz0rl left	14:23
	→ Buglouse joined	14:24
	← aresnick left	14:24
	→ aresnick joined	14:25
	→ _iron joined	14:25
	← _iron left	14:25
	← coderdad_ left	14:27
	← drubies left	14:27
	← munichlinux left	14:27
	→ coderdad joined	14:28
	→ paja joined	14:28
	→ engla joined	14:28
	← bin10102 left	14:29
	→ chrowe joined	14:29
	← simplechat left	14:29
	→ bin10101 joined	14:29
ereslibre	is it recommended to run "git gc" on server end, or is it useless ?	14:30
	→ jfkw joined	14:31
	→ bin10102 joined	14:31
	→ kumbayo joined	14:32
	→ styx-tdo joined	14:33
	→ mw_ joined	14:33
	← mw_ left	14:33
	→ mw_ joined	14:33
styx-tdo	hi, can anyone help me? maybe i am too stupid, but i cannot find "fork" on githug	14:33
	*github	14:33
	← mw left	14:33
	mw_ → mw	14:33
ereslibre	styx-tdo: go to the project you want to fork, right top there is a "Fork" button	14:34
	← bin10101 left	14:34
	← ph^ left	14:34
	← TheNewAndy left	14:34
bdrewery	lol 'githug'	14:34
ereslibre	bdrewery: hugging gits since 2010 ;)	14:35
	→ LiamH joined	14:35
frsk	..and someone has actually reserved githug.com :p	14:35
styx-tdo	errr.. javascript solved this grmbl	14:35
engla	styx-tdo: you won't be happy if you want to use github without javascript	14:36
styx-tdo	engla: yes. this is so.. well.	14:37
	w3m-friendly.	14:37
	ereslibre is even surprised github loads without js xD	14:38
engla	I want a cgit frontend to github	14:38
	← vu3rdd left	14:41
charon	if only git were distributed so you had free choice of hosting provider...	14:41
engla	I still want a cgit frontent to github	14:42
	→ hohoho joined	14:42
	→ _iron joined	14:43
wereHamster	and I want a pink pony!	14:44
	→ simplechat joined	14:44
	← raichoo left	14:44
	→ muszek joined	14:44
muszek	hi	14:44
	← shennyg left	14:45
	→ rayners joined	14:45
	→ ramusara joined	14:47
	→ shennyg joined	14:47
	← shennyg left	14:47
muszek	I want to interactively part a part of a file. The thing is that a part that I want to add is neighboring another change, which I don't want to add. when git asks me whether I want to add this hunk, it offers me several options. "s" (for "split") is not among them. However, in the help thingy (after I choose "?"), I get a "s" option... http://pastebin.ca/1876658	14:47
	how can split the hunk?	14:48
	← ahf left	14:48
	→ winterweaver joined	14:48
	← kpreid_ left	14:49
patrikf	muszek: "s" is only available if git can determine a point to split the hunk (i.e. an unchanged line between changes). use "e" to edit the patch manually	14:49
muszek	patrikf: thank you	14:49
	→ ahf joined	14:50
muszek	I'm done editing... I should just :wq (write and quit) in vim?	14:50
	← mastro left	14:51
charon	patrikf: yes. it will check if the edited patch applies, and if so, automatically accepts this new hunk	14:52
patrikf	muszek: what charon said	14:53
charon	oh, sorry.	14:53
patrikf	nvm	14:53
	charon grabs more coffee	14:53
patrikf	happens to the best ;-)	14:53
muszek	thanks	14:54
	→ ph^ joined	14:54
	← bcardarella left	14:54
	→ kar8nga joined	14:55
	→ clayjar joined	14:55
	← aresnick left	14:55
	→ neha__ joined	14:55
	← jony1 left	14:55
	← _tydeas_ left	14:55
clayjar	Hello, a question regarding use of color under SLES 10. I haven't had any problem with Slackware, Debian/Ubuntu, but I see Escape codes such as "ESC[1m" everywhere on SLES 10. Colors work fine for directories and in VIM, but not Git.. any ideas?	14:56
SamothUK	clayjar: probably your less setting?	14:57
clayjar	declare -x LESS="-M -I" declare -x LESSCLOSE="lessclose.sh %s %s" declare -x LESSKEY="/etc/lesskey.bin" declare -x LESSOPEN="lessopen.sh %s" declare -x LESS_ADVANCED_PREPROCESSOR="no"	14:58
	See anything unusual?	14:58
SamothUK	clayjar: I need -R for colours to work on mine	14:58
	garlic:[ 3:56PM]%echo $LESS	14:58
	-aeimnRsx8	14:58
	like so	14:58
neha__	hi all, i am using git on windows, when i use 'git add .' command to add the files. The problem is the sub-folders contains spaces, which causes error message 'fatal: pathspec folder_name_before_space did not match any file', how to solve the issue.	14:58
	→ transplant joined	14:59
styx-tdo	say, git commit DOES commit to the repo, right?	14:59
SamothUK	from the manpage : -R or --RAW-CONTROL-CHARS Like -r, but only ANSI "color" escape sequences are output in "raw" form. Unlike -r, the screen appearance is maintained cor- rectly in most cases. ANSI "color" escape sequences are sequences of the form:	14:59
mmattice	quote the paths	14:59
clayjar	SamothUK, that worked. Thanks!	14:59
charon	styx-tdo: of course, but people who have to ask that usually assume "the repo" lives on some remote server (as with svn) whereas with git the whole repo lives in the .git	14:59
	(i.e. locally)	14:59
styx-tdo	eh.. so, how does one upload stuff to the bloody server?	14:59
charon	man git-push	14:59
Gitbot	charon: the git-push manpage can be found at http://git.or.cz/man/git-push	14:59
charon	also goes for unharmed servers btw	15:00
	→ hukl joined	15:00
SamothUK	clayjar: cool :)	15:00
	← Joeconyers left	15:00
	← clayjar left	15:00
styx-tdo	aaaaaaaaaaaaaaahhhhhhhhhhhhhhh	15:01
	push!!!!!!!!!!!!!!!1	15:01
patrikf	charon: the server might get aggressive though when you start pushing it around	15:01
	→ stringoO joined	15:01
	→ raichoo joined	15:02
	→ hyperair joined	15:02
charon	styx-tdo: and before you fall into the next trap, note that pull is not the opposite of push. rather, fetch is, and pull=fetch+merge.	15:02
	→ Swistaczek joined	15:02
charon	patrikf: we should also implement 'git shove' ;)	15:02
styx-tdo	hehe	15:02
	thanks, charon	15:02
	→ Jferro joined	15:03
transplant	or you can login and fetch	15:03
	← Jferro left	15:03
hyperair	is there any way to set global gitattributes?	15:04
	← giallu left	15:04
hyperair	i don't feel like copying the same file around in .git/info/attributes	15:04
	→ kergoth joined	15:04
hyperair	(i'm using it to specify a custom tool for merging a certain type of file)	15:04
davr	git config --global	15:04
patrikf	charon: hah. if i ever need an alias for pull -f...	15:04
kergoth	is there a way to rename a branch on a git repository that many folks pull from, without forcing everyone to adjust their local branches to work? i assume not, i don't think you can just symlink the old branch to the new for compatibility or something	15:05
davr	maybe that's different from attributes tho...not sure	15:05
	→ meanburrito920 joined	15:05
patrikf	charon: er, push -f	15:05
	← hasu left	15:06
	← paddor left	15:07
	→ Jferro joined	15:07
	→ Terabyte joined	15:08
	← Terabyte left	15:08
	← transplant left	15:08
	← mtkd left	15:08
	→ mtkd joined	15:08
charon	hyperair: no, but maybe there should be one by analogy with core.excludesfile	15:09
	← simplechat left	15:09
	→ saysjonathan joined	15:09
	← kergoth left	15:09
	→ mastro joined	15:10
	← jherdman left	15:10
	← werdan7 left	15:11
	← neha__ left	15:11
	→ irc2samus joined	15:12
	← meanburrito920 left	15:14
	→ jherdman joined	15:14
	← cbreak_work left	15:15
	→ CarlFK2 joined	15:15
	→ werdan7 joined	15:16
	→ Titosemi_ joined	15:16
	→ ProperN[out] joined	15:16
	← ProperN[out] left	15:17
	→ ProperN[out] joined	15:17
	← nevyn^ left	15:17
	← Titosemi left	15:17
	Titosemi_ → Titosemi	15:17
	← winterweaver left	15:17
	→ nevyn^ joined	15:18
hyperair	charon: yeah, there should be. but there isn't	15:18
	← nevyn^ left	15:19
	→ nevyn^ joined	15:19
	← ProperNoun left	15:20
	ProperN[out] → ProperNoun	15:20
	→ XDS2010__ joined	15:20
	← hohoho left	15:21
	← XDS2010 left	15:21
	← CarlFK2 left	15:22
	← radioman-lt left	15:24
	← bin10102 left	15:24
	→ bin10101 joined	15:24
	→ radioman-lt joined	15:24
	← brez- left	15:26
	← posciak left	15:26
	→ meanburrito920 joined	15:26
	→ bin10102 joined	15:27
	→ bcardarella joined	15:27
	← bin10102 left	15:31
	→ jvraoltc joined	15:31
	← bin10101 left	15:32
	→ bin10101 joined	15:33
	← bin10101 left	15:34
	→ bin10101 joined	15:37
	→ Heimidal joined	15:38
	→ NET\|\|abuse joined	15:38
	← Swistaczek left	15:39
	→ dj_tjerk joined	15:40
	← bin10101 left	15:41
alezandro	what's the right command to bring a file in my working dir back to it's state as of last commit, throwing away any uncomitted changes?	15:41
offby1	git checkout HEAD -- that-file	15:42
alezandro	offby1: thanks	15:43
	→ AAABeef joined	15:44
	→ JackWinter joined	15:44
	← meanburrito920 left	15:45
offby1	not obvious, I'm afraid	15:45
	→ sagsousuke joined	15:46
yrlnry	Why not reset --hard HEAD ?	15:47
	Oh, a file.	15:47
	Never mind!	15:47
	← project2501a left	15:47
	→ dytrivedi joined	15:48
	← nevyn^ left	15:48
	→ giallu joined	15:50
	→ kipras joined	15:51
	→ Determinist joined	15:51
	→ coderdad_ joined	15:52
	← JackWinter left	15:53
	→ shade__ joined	15:53
	→ kahmalo joined	15:53
	← XDS2010__ left	15:53
	→ XDS2010__ joined	15:53
	← ajpiano left	15:54
kahmalo	I've been having some difficulty with git grep, in git version 1.7.1. Seems like it doesn't find strings in some binary files where GNU grep does find them.	15:54
curtana	maybe it doesn't search binary files?	15:55
kahmalo	I tried the -a and --text options too.	15:55
curtana	huh	15:55
	← hebz0rl_ left	15:56
	→ WebDragon joined	15:56
	← coderdad left	15:56
	→ domcat joined	15:57
	→ project2501a joined	15:57
kahmalo	Hm, now it seems to find the matches after all... I'll need to test a bit more.	15:57
	→ meanburrito920 joined	15:57
	← LRN left	15:57
	→ ajpiano joined	15:58
	→ LRN joined	15:58
	→ adamv joined	15:59
	← dj_tjerk left	15:59
	→ Nickeeh joined	16:00
	→ RobertLaptop joined	16:00
	→ Jonah joined	16:00
	→ elmob joined	16:00
	→ hobodave joined	16:01
	→ JackWinter joined	16:01
	IslandUsurper → IslandUsurperAFK	16:02
	← EricInBNE left	16:03
	→ yann joined	16:03
	← JackWinter left	16:04
	← mase_home left	16:04
	→ dkobia joined	16:06
	← giskard left	16:07
	← voxpelli-laptop left	16:08
	→ elyob joined	16:09
	→ JoseLobato joined	16:10
	← trivol left	16:11
	→ entrosca joined	16:11
	← priidu left	16:12
	← shade__ left	16:12
	→ jaloga joined	16:13
	← Nemurenai left	16:14
	← RobertLaptop left	16:15
	← kar8nga left	16:15
	← JoseLobato left	16:16
	← jaloga left	16:17
	← entrosca left	16:20
	← urkud_ left	16:20
	→ mattikus joined	16:20
	→ hal joined	16:25
hal	github has the concept of forking. What is the native git equivalent of this, please?	16:26
Rhonda	clone	16:26
	← tildeequals left	16:27
hal	Rhonda: thank you. Does clone automatically track the original repo/branch?	16:27
Rhonda	Yes, it will set git up that a simple git pull will pull from that original remote.	16:28
	→ kenichi joined	16:30
	← ntoll left	16:30
	← project2501a left	16:31
	→ CodeTalks joined	16:32
ReinH	hal: the original becomes the "origin" remote and master is set up to track it.	16:33
yrlnry	I was reading somewhere that after a failed merge, the index contains four instances of a conflicted file: two versions from the parent commits, a version from the common ancestor, and a version with conflict markers. What happens when the merge has more than two parents?	16:33
	Investigation of git-rev-parse suggests that there are never more than the four instances.	16:33
	So what happens?	16:33
	→ giskard joined	16:33
	← XDS2010__ left	16:33
ReinH	yrlnry: like in an octopus merge?	16:33
	→ resmo joined	16:33
	→ lresende joined	16:33
	← lresende left	16:33
	← CodeTalks left	16:34
	→ lresende joined	16:34
yrlnry	ReinH: like in whatever happens when you say git merge and supply 43 commits on the command line.	16:34
	← gaveen left	16:34
	← shruggar left	16:34
ReinH	yrlnry: octopus won't allow a manual merge	16:35
	man git-merge	16:35
Gitbot	ReinH: the git-merge manpage can be found at http://git.or.cz/man/git-merge	16:35
ReinH	and read the merge strategies section	16:35
yrlnry	man git-merge says "git merge ... <commit>..."	16:35
	→ shade__ joined	16:35
Gitbot	yrlnry: the git-merge manpage can be found at http://git.or.cz/man/git-merge	16:35
hal	thank you Rhonda and ReinH	16:36
	← meanburrito920 left	16:36
ReinH	and read the merge strategies section	16:36
	→ slonopotamus joined	16:36
ReinH	hal: yw	16:36
yrlnry	Okay, thanks.	16:36
	IslandUsurperAFK → IslandUsurper	16:36
	→ XDS2010 joined	16:37
doener	yrlnry: the octopus merges actually does a series of 3-way merges, so you never have more than 3 versions (the version with the conflict markers is just in the working tree)	16:38
	← kahmalo left	16:38
yrlnry	Also thanks.	16:39
	← d0ugal left	16:39
	→ Yuffster joined	16:39
doener	that an octopus merge gets "aborted" when there are conflicts is IIRC even unrelated, as the file is still left unmerged in the index	16:40
	→ AkumaStreak joined	16:41
	← forvan left	16:43
	← fmardini left	16:44
	→ icwiener joined	16:45
	← ive left	16:47
	→ urkud_ joined	16:47
	→ ive joined	16:47
	→ project2501a joined	16:48
	→ spearce joined	16:50
	← luke-jr left	16:51
	→ luke-jr joined	16:51
yrlnry	ReinH, doener: So it is definitely the case that the index never contains more than four versions of the same file, as implied by the section of git-rev-parse(1) that begins "A colon, optionally followed by a stage number (0 to 3)..."?	16:51
	→ RobertLaptop joined	16:52
doener	yrlnry: 1-3 are normally only there if 0 is not there and vice versa	16:52
	→ JackWinter joined	16:52
	← coderdad_ left	16:53
doener	yrlnry: unmerged files have entries in stages 1-3, "regular" tracked files have just stage 0	16:53
	→ luckiestmonkey joined	16:53
	→ coderdad joined	16:53
ReinH	yrlnry: correct	16:54
	← ramusara left	16:54
yrlnry	Why is the index so big even after git reset?	16:55
	What's in there?	16:55
ReinH	"so big"?	16:55
yrlnry	37880 bytes, instead of the 0 bytes I was expecting.	16:55
	→ coderdad_ joined	16:56
yrlnry	Oh, I see.	16:57
	It doesn't record just the stuff I added.	16:57
	← tobiassjosten left	16:58
yrlnry	it records the state of the last commit, plus whatever I added.	16:58
	Oh, it has to do that because when i make a commit, the index is turned into the tree for the new commit.	16:58
	Okay, thanks.	16:58
	→ WebDragon\|away joined	16:58
	← coderdad left	16:58
	← coderdad_ left	16:58
	← slonopotamus left	16:59
	← WebDragon left	17:00
	→ priidu joined	17:00
yrlnry	Oh, git-reset just copies the tree from the current HEAD to the index.	17:01
	→ vu3rdd joined	17:01
	← JackWinter left	17:01
	→ pantsman- joined	17:01
	→ slonopotamus joined	17:03
	→ Swistaczek joined	17:04
ReinH	:)	17:04
	makes sense when you think about it...	17:04
	← pantsman- left	17:05
	→ pantsman- joined	17:05
	← pantsman- left	17:05
yrlnry	How does git-status know so quickly what has changed between the working tree and the index? It can't be hashing every file in the working tree because that would be too slow. I thought it used the mtimes on the files to narrow down the search, and hashed only the files whose mtimes has changed, but that seems not to be the case.	17:07
	→ pantsman joined	17:08
wereHamster	yrlnry: why do you think that's not the case?	17:08
yrlnry	It might be the case. I set the mtime on a modified file to January and it still shows up in the git-status output. So it could still be using the mtime, if the index records the original mtime of the file.	17:09
	← bluenovember left	17:09
yrlnry	What is it doing?	17:10
	← HG` left	17:10
yrlnry	And how can I get git to show me the exact contents of the index, including the multiple stages of multiple-stage files?	17:10
	← mastro left	17:10
	→ ChUrP joined	17:10
	← schulty left	17:10
yrlnry	git-show-index.	17:10
	Whee.	17:10
wereHamster	the exact contents of the index? cat .git/index	17:10
	← alezandro left	17:11
yrlnry	What an asshole.	17:11
	→ alezandro joined	17:11
	← t0rc left	17:12
	← Nickeeh left	17:12
wereHamster	well, it depends on what you want to know from the index. And show-index does not display the contents of .git/index, but of a pack index	17:12
	→ spicyWith joined	17:13
wereHamster	there's ls-files, or the different rev-parse syntaxes, such as :<x>:<path>	17:13
	← hukl left	17:14
nanotube	wereHamster: maybe instead of 'cat' you should have used 'more' :P	17:15
	← dytrivedi left	17:15
wereHamster	or hexdump	17:15
nanotube	hehe	17:16
	← slonopotamus left	17:16
	← curtana left	17:17
	← resmo left	17:17
	← spicyWith left	17:18
	← radarek left	17:18
doener	yrlnry: git ls-files -s	17:18
	yrlnry: gets you the mode bits, the object id, the stage number and the path. And shows all stages	17:19
yrlnry	doener: thanks.	17:19
	← paja left	17:19
	→ tvw joined	17:20
	→ gp5st joined	17:21
	← spearce left	17:21
	→ resmo joined	17:22
	→ malumalu joined	17:22
gp5st	hey! so, i forget what the command is called, but I have a repo that i've made accessable via http (read-only, no dav). How do I update the point at which the person cloning over http gets as the head?	17:22
	sorry if that wasn't as clear as I'd like it to be:(	17:22
	← ereslibre left	17:22
doener	update-server-info	17:23
	→ paja joined	17:23
	WebDragon\|away → WebDragon	17:23
gp5st	ooo, thanks doener	17:23
	→ WebDragon\|away joined	17:24
hal	how do you undo a change that you have already pushed to the remote, please?	17:24
doener	git revert <commit>	17:24
rolfb	git reset --hard doesn't touch not tracked files?	17:24
doener	adds a new commit that reverts the changes introduced by the given commit	17:24
	rolfb: s/tracked/untracked?/ ?	17:24
rolfb	not tracked/untracked yes	17:25
	:)	17:25
doener	ah, heh, my brain skipped the "not", somehow...	17:25
	rolfb: IIRC it might touch them.	17:25
hal	what is the difference between git revert and git reset --hard sorry, I don't understand	17:25
rolfb	"it might"	17:26
	:)	17:26
	→ mrlarner joined	17:26
doener	rolfb: e.g. you might have file "foo" which is untracked, and you do "git reset --hard master~5" and master~5:foo exists. IIRC reset --hard will then overwrite the untracked file	17:26
	→ kumbayo_ joined	17:26
hal	ok, git reset disgards uncommitted changes	17:27
doener	hal: reset resets the branch head to a given commit, so you might drop commits from the history	17:27
	hal: revert adds new commits	17:27
rolfb	doener: ah	17:27
	but all files that are not added to the index will survive	17:28
	unless they have counterparts	17:28
	← WebDragon left	17:28
	→ vinse joined	17:29
doener	rolfb: yup. And "not add to the index" part is important. They don't have to be committed for git to kill them ;-)	17:29
	i.e. there's no guarantee that the reflog will help you in restoring them	17:29
rolfb	crud	17:29
	→ spearce joined	17:29
rolfb	k	17:29
	i'll create a backup directory	17:29
	should have been able to make a dry run	17:30
	:P	17:30
doener	what I mean is: echo 123 > foo # "foo" being a new file; git add foo; git reset --hard # bye bye "foo"	17:30
	← kumbayo left	17:30
rolfb	yup	17:30
	i figured	17:30
yrlnry	It turns out that the stuff I wanted to know about the structure of .git/index is defined in cache.h in the git source code.	17:31
	→ _tydeas_ joined	17:31
yrlnry	The comments imply that it uses not only the mtime, but also "dev/ino/uid/gid/size", or at least the least significant 32 bits of each.	17:32
	I have not looked closely at most of the code though.	17:32
	← spearce left	17:32
	→ spearce joined	17:33
	→ Nickeeh joined	17:33
	← elmob left	17:34
	← priidu left	17:35
	→ priidu joined	17:36
	→ enherit joined	17:36
hal	ok, I think I want to undo the changes made last commit, and the commit before last, using the revert method	17:36
	← gp5st left	17:36
hal	could someone help me through this please? I don't want to make a mistake	17:36
wereHamster	git revert HEAD~2; git revert HEAD~2	17:38
	← jds left	17:38
wereHamster	er, wrong. git revert HEAD; git revert HEAD~2	17:38
doener	wereHamster: 2? I'd expect 1	17:38
	← Jferro left	17:39
doener	yeah, for depedent changes that's better. For independent changes, HEAD~1 HEAD~1 should also be fine	17:39
hal	wereHamster: thank you - isn't it git revert HEAD~1; git revert HEAD~1 ?	17:39
wereHamster	of course, I was wrong :)	17:39
hal	oh great - I am even understanding it too! :)	17:40
	thank you very much wereHamster	17:40
	→ aresnick joined	17:40
_tydeas_	Using github.com as a repo a user makes forks of my project. One of the user commits was an empty commit. It tried to make a pull request on that but don't appear in my dashboard. Is this maybe a reason that i can not apply the pull requests he made after?	17:40
	→ radarek joined	17:41
	→ krawek joined	17:42
	→ iravid joined	17:42
	← hyperair left	17:43
	← shade__ left	17:43
	← timj_ left	17:44
	→ shade__ joined	17:44
_tydeas_	i command --depth until which commit i want to fetch changes?	17:47
	← batjack left	17:51
	→ pheaver joined	17:51
	→ lhz joined	17:51
	→ __iron joined	17:51
	→ batjack joined	17:54
	→ ispeak_ joined	17:54
	→ kpreid_ joined	17:54
	← _iron left	17:55
	→ smuf joined	17:55
	→ khmarbaise_ joined	17:55
	→ ereslibre joined	17:56
	← ereslibre left	17:56
	→ ereslibre joined	17:56
	← bentob0x left	17:56
	← sunil left	17:56
	→ gaveen joined	17:56
	← altrux left	17:57
	← iravid left	17:57
	← khmarbaise left	17:58
	khmarbaise_ → khmarbaise	17:58
	← werdan7 left	17:58
	← shade__ left	17:59
	← frakturfreak left	17:59
	← petrux left	18:00
	→ frakturfreak joined	18:00
	→ werdan7 joined	18:00
	→ jowls joined	18:01
	← bcardarella left	18:01
	→ ribasushi joined	18:02
	← Heimidal left	18:03
	[Po]lentino → [Po]lentino\|away	18:05
	← kpreid_ left	18:05
	→ flaguy48 joined	18:07
	→ kar8nga joined	18:07
	← galderz left	18:08
	← resno left	18:08
	→ drubies joined	18:08
_tydeas_	Is this way of thinking correct. I have a commit in remote repo that cloned by mine, and i want to cherry-pick ( with -x ). Are the following steps correct: remote add and then cherry-pick?	18:09
	→ nevyn^_ joined	18:10
_tydeas_	As i can see something must happen in the middle of this 2 actions but can not understand what	18:11
cbreak	you have to git fetch	18:11
	just adding a remote does not do any network ops	18:11
	fetch will get the stuff from the remote	18:11
	and if you cloned from there	18:11
	you do not need to add the remote	18:11
	since it is already added for you under the name origin	18:11
	→ bcardarella joined	18:12
cbreak	(git remote -v)	18:12
	→ marcelinollano joined	18:12
	→ theoros joined	18:14
	→ AAA_awright_ joined	18:15
	→ dfas` joined	18:15
	← kar8nga left	18:17
	← kumbayo_ left	18:17
	← alezandro left	18:17
	← Swistaczek left	18:17
	← radioman-lt left	18:17
	← ph^ left	18:17
	← LiamH left	18:17
	← parasti left	18:17
	← jensn left	18:17
	← eno left	18:17
	← tatsuya left	18:17
	← uau left	18:17
	← alester left	18:17
	← salisbury left	18:17
	← Cesario_ left	18:17
	← berserker left	18:17
	← doki_pen left	18:17
	← AAA_awright left	18:17
	← niko left	18:17
	← codejunky left	18:17
	← baccenfutter left	18:17
	← mapreduce left	18:17
	← mnemoc left	18:17
	→ jarh joined	18:17
	→ timj joined	18:17
	→ parasti joined	18:17
	→ kar8nga joined	18:18
	→ uau joined	18:18
	→ tatsuya joined	18:18
_tydeas_	cbreak: i don't want to fetch i think i want to cherry-pick it because i want the user details	18:18
	← ispeak_ left	18:18
cbreak	you don't want to cherry pick?	18:18
	→ alezandro joined	18:18
_tydeas_	i want to cherry pic	18:19
	→ tildeequals joined	18:19
cbreak	it is technically impossible to cherry-pick commits you do not have	18:19
	so you will have to fetch them	18:19
_tydeas_	so how can i fetch a specific commit	18:19
	→ ispeak_ joined	18:19
cbreak	unless you already have them	18:19
	just fetch all of them	18:19
	→ wanders joined	18:19
	→ eno joined	18:19
	→ mnemoc joined	18:19
cbreak	_tydeas_: man git-fetch	18:19
Gitbot	_tydeas_: the git-fetch manpage can be found at http://git.or.cz/man/git-fetch	18:19
cbreak	fetch is NOT merge	18:20
	→ berserker joined	18:20
	→ Pieter_ joined	18:20
	→ KeeperOfTheSoul joined	18:20
cbreak	it does not merge anything, it just updates your local tracking branches	18:20
	→ LiamH joined	18:20
	→ tokkee joined	18:20
	→ mapreduce joined	18:20
	→ codejunky joined	18:20
	→ Slush- joined	18:21
_tydeas_	fatal: Commit b4152d0606a2d201f10e2542bc7c8f57e99b0c10 is a merge but no -m option was given.	18:21
	what about this when i try to cherry pick?	18:21
	→ Goneri joined	18:21
_tydeas_	where can i get the mainline?	18:21
	→ siprbaum joined	18:21
cbreak	cherry picking merge commits :/	18:21
	I don't know that's a good idea	18:22
	→ bartman joined	18:22
	← luke-jr left	18:22
wereHamster	_tydeas_: use git merge b4152d0606a2d201f10e2542bc7c8f57e99b0c10; insted of cherr-pyck	18:22
	→ luke-jr joined	18:22
cbreak	wouldn't that merge the branch?	18:22
	→ Binary0010_ joined	18:22
_tydeas_	wereHamster: i want to keep user history and want to keep this commit sha1	18:22
cbreak	including previous changes?	18:22
	← marcelinollano left	18:23
cbreak	cherry-picking will not preserve the sha	18:23
	and it will not preserve history	18:23
	→ Cesario_ joined	18:23
	→ baccenfutter joined	18:23
	→ hyperair joined	18:23
_tydeas_	sorry didn't want to say history	18:23
	→ jensn joined	18:23
	→ kumbayo_ joined	18:23
cbreak	there's only one way to get changes while preserving their sha1:	18:24
	a fast forward merge	18:24
	→ doki_pen joined	18:24
	→ Swistaczek joined	18:24
	→ ph^ joined	18:24
	← project2501a left	18:24
	→ Vojto joined	18:24
_tydeas_	yes i want to preserve it with user details so i can merge other pull-requests based on this commit	18:24
	did i made my self clear?	18:24
	→ salisbury joined	18:25
	→ cloudhead joined	18:25
	→ rubydiamond joined	18:25
Vojto	Hey there, can anyone help me? How do I clone older version of repository on github -- like 02b0e6dcc129096a4d32	18:25
_tydeas_	man clone	18:25
wereHamster	Vojto: you clone the repo, then checkout the commit you want	18:25
	→ radioman-lt joined	18:25
Vojto	wereHamster: I just do git checkout 02b0e6dcc129096a4d32 ?	18:25
	→ charon joined	18:25
cbreak	git clone gives you all history	18:25
wereHamster	Vojto: yes	18:25
cbreak	not just the tip	18:26
	→ project2501a joined	18:26
cbreak	be aware that doing that will detach your head.	18:26
	→ Joeconyers joined	18:26
	← Binary0010 left	18:26
	← Binary0010_ left	18:26
cbreak	so don't commit until you know what you're doing	18:26
Vojto	thanks a lot	18:26
	→ Binary0010 joined	18:26
Vojto	I just need to play with older version of repo. I thought checkout is for switching to branches.	18:26
	← luckiestmonkey left	18:26
cbreak	it's for switching to something. you can switch to any commit	18:26
IslandUsurper	Vojto, if you do want to make commits based on that commit, go ahead and make a branch there	18:27
	→ HG` joined	18:27
cbreak	there's a second form that only gets you files, and does not switch.	18:27
Vojto	thanks guys, you are awesome	18:27
	WebDragon\|away → WebDragon	18:28
	← rayners left	18:29
	← redondos left	18:30
	→ niko joined	18:30
	→ redondos joined	18:30
_tydeas_	when merged a specific commit i got an other commit sh1 even with --ff	18:31
	← rchavik left	18:31
cbreak	as I said	18:32
_tydeas_	no sorry	18:32
cbreak	true fast forward merges are the only way to get same sha1	18:32
	← ehabkost left	18:32
cbreak	and that way you get the whole history, not only one commit	18:32
	(because that's the definition of an ff merge)	18:32
_tydeas_	cbreak: yes that's what happen	18:33
	→ ehabkost joined	18:33
cbreak	it is impossible to not get the history, while retaining the sha	18:33
	because the sha is generated from the history, among others	18:33
	→ coderdad joined	18:34
	← gaveen left	18:34
	← ispeak_ left	18:35
	← urkud_ left	18:35
	← cannonball left	18:36
	← niko left	18:36
	→ niko joined	18:37
	→ urkud_ joined	18:37
	← steven_t left	18:40
	→ benjo2 joined	18:42
	← niko left	18:42
	→ g0 joined	18:42
	→ niko joined	18:43
g0	hi. I ran git rebase, made changes but didn't commit them (:P) and did git rebase --continue is there a way to revert the rebase?	18:44
	← multi_io_ left	18:45
bdrewery	using the reflog you can checkout the original ref	18:45
	man git-reflog	18:45
Gitbot	bdrewery: the git-reflog manpage can be found at http://git.or.cz/man/git-reflog	18:45
g0	bdrewery, i thought of reflog, but it didn't occur to me to checkout :P Thanks!	18:46
cbreak	do not checkot	18:47
	→ sparr joined	18:47
bdrewery	or just git reset --hard ref	18:47
cbreak	if you check out something from the reflog, you will get a detached head	18:47
	you have to do a reset, if you want to change your branch head	18:47
	so git checkout master	18:47
sparr	I cloned a repository, made a change, committed the change locally. How can I revert back to the unchanged version?	18:47
cbreak	and after that, a hard reset to the sha	18:47
	sparr: do you mean revert? Probably not. If you mean reset, it's git reset --hard HEAD^	18:48
	→ Theravadan joined	18:48
	← altogether left	18:48
	← weary left	18:48
	← mqt left	18:48
	← niko left	18:49
	→ johnw joined	18:49
	← keyvan left	18:49
	→ niko joined	18:49
	← m104 left	18:49
	→ keyvan joined	18:50
_tydeas_	[dmtrs@linux svn]$ git svn rebase	18:51
	Unable to determine upstream SVN information from working tree history	18:51
	what does it mean?	18:51
	→ giskard_ joined	18:51
johnw	ah, I've seen this many times before	18:52
	trying to remember...	18:52
	← villemv_ left	18:52
cbreak	are you on a branch with svn meta?	18:52
_tydeas_	no meta	18:52
johnw	ah	18:52
	it requires it to know which branch to communicate with	18:53
	← mtkd left	18:53
_tydeas_	johnw: svn branch?	18:54
johnw	well, svn path is more accurate, but yeah	18:54
	← giskard left	18:54
ribasushi	is this the right place to ask about gitk weirdness ?	18:54
	← niko left	18:54
	→ niko joined	18:55
	→ adymitruk joined	18:55
	→ [Po]lentino joined	18:56
	← dpino left	18:56
	← saysjonathan left	18:56
sparr	cbreak: thanks, I was missing the "^"	18:56
	← [Po]lentino\|away left	18:57
_tydeas_	johnw: how can i do it?	18:57
johnw	how can you do what?	18:57
_tydeas_	svn rebase	18:57
	with this error	18:57
	i get	18:57
johnw	without metadata, you can't; it wouldn't make any sense	18:57
	→ jailbox joined	18:58
	→ Mage joined	18:58
	→ Jferro joined	18:59
	→ luckiestmonkey joined	19:00
sjohnson	let's say you make 3 changes on different files, thereby having 3 modified files	19:01
	is there a way to git checkout everything to the last commited state, without doing git checkout 3 times?	19:02
	← drako left	19:02
cbreak	sjohnson: what?	19:02
	why would you check out three times anyway?	19:02
	you want to reset away your local changes?	19:03
sjohnson	yeah	19:03
cbreak	git reset --hard	19:03
sjohnson	obviously i woudln't want to check out three times, hence the question	19:03
cbreak	git checkout -- file.1 file.2 file.3	19:03
broonie	sjohnson: If you want to preserve other local changes you can specify as many files as you like in the command line	19:03
cbreak	there are a few other ways	19:03
sjohnson	i suppose if i had 15 modified files	19:04
	and i wanted to "start fresh" (from the last commit)	19:04
	it would make sense to find an easier way than file.1 ... file.15	19:04
	seems git reset --hard is the way	19:04
	→ m104 joined	19:04
	→ altrux joined	19:04
	→ razvand joined	19:05
	→ gbacon joined	19:05
sjohnson	cbreak: thanks	19:06
	will try it out	19:06
	← WebDragon left	19:06
	← kumbayo_ left	19:08
	← ive left	19:08
	→ hachi joined	19:09
	← pantsman left	19:09
	← kristian42 left	19:10
hachi	hi all, when writing a post-receive hook sometimes the 'oldrev' is 000000, how can I find what the real starting revision of the received data is from git?	19:10
	← niko left	19:10
	→ theoros_ joined	19:10
sjohnson	hachi: do you mean 00000000 ?	19:12
wereHamster	rev-list $newrev \| tail -n1 ?	19:12
	← bcardarella left	19:12
	← theoros left	19:12
wereHamster	or rev-list --reverse -1 $newrev	19:12
	→ pantsman joined	19:12
hachi	except the 00000 is to indicate a new branch	19:13
	→ kristian42 joined	19:13
hachi	not that a single rev was applied	19:13
	so your solution will only tell me the last rev... not all of the revs received	19:13
	← NetHawk left	19:13
wereHamster	if oldrev is 0000.., then `rev-list $newrev` are the commits which have been received	19:14
	→ niko joined	19:14
hachi	sjohnson: I find that in practice git rarely gives me 000000 without it being full 0	19:14
sjohnson	hachi: how many zero's are there?	19:14
	you have said four, five, and six 0's now. which is it?	19:15
hachi	it doesn't matter, read between the lines	19:15
	I mean an oldrev that is fully zeros	19:15
	I don't feel like typing all the zeroes though, can you fault me for that?	19:15
	→ NetHawk joined	19:16
hachi	wereHamster: there is a common history with other branches though, and those haven't been received	19:16
	← giskard_ left	19:16
hachi	I'm trying to find the new commits	19:16
	if I say git rev-list I'm going to run over my entire history again to the start of time	19:17
wereHamster	ah, I see.	19:17
	→ montylounge joined	19:17
	→ s34n joined	19:17
	→ WebDragon joined	19:19
	← ajpiano left	19:19
	← HG` left	19:20
	← niko left	19:22
sjohnson	hachi: sorry, just was curious how many exactly there were	19:22
	← project2501a left	19:22
	← enherit left	19:22
	→ ajpiano joined	19:22
	→ _chevdor joined	19:23
	← ph^ left	19:23
_tydeas_	johnw how can i get only the metadata from an svn repo?	19:26
	→ giskard joined	19:26
johnw	you have to clone again	19:26
	← chevdor left	19:26
_tydeas_	entire?	19:27
	or will clone only metadata?	19:27
	→ halogenandtoast joined	19:28
	→ diegoviola joined	19:28
halogenandtoast	Is there a way to revert only specific lines in a file?	19:28
diegoviola	hi	19:28
	how do i checkout a file from head?	19:28
	or restore a file	19:28
johnw	you have to clone teh entire thing	19:28
diegoviola	that was deleted	19:28
johnw	why did you clone it without metadata?	19:28
	that's for one-way conversion only	19:28
_tydeas_	talking here hours before	19:28
cbreak	halogenandtoast: man git-checkout	19:29
Gitbot	halogenandtoast: the git-checkout manpage can be found at http://git.or.cz/man/git-checkout	19:29
cbreak	the -p option might suit you	19:29
s34n	Is it possible/advisable to have multiple git repos that pull from each other?	19:29
halogenandtoast	cbreak: Yeah I was looking at the man page for that, but didn't think of using the -p option	19:29
wereHamster	sure	19:29
halogenandtoast	thanks	19:29
s34n	so I don't have to remember which one to push from?	19:30
wereHamster	s34n: if you don't want to remember urls, use remotes	19:30
cbreak	s34n: don't push to non-bares.	19:31
halogenandtoast	cbreak: that worked, thanks	19:31
s34n	I can just sync up at any time by pulling into whichever repo I'm working in	19:31
cbreak	don't pull into bares.	19:31
	the decision wether to pull or push is more or less made by the type of the destination	19:31
	imho.	19:31
	of course, there are ways around	19:31
	→ project2501a joined	19:33
s34n	so I can clone A to B, then set B as a remote of A and pull both directions?	19:33
_tydeas_	should i git init before i clone?	19:33
wereHamster	no, git clone creates the repo for you	19:33
cbreak	clone is a superset of init	19:34
_tydeas_	so if i use git svn clone -A author-map http://yii.googlecode.com/svn/	19:34
	← lemonchicken left	19:34
	← tvw left	19:34
_tydeas_	i will be later able to rebase?	19:34
	→ pklingem joined	19:36
johnw	_tydeas_: did you use --no-metadata before?	19:37
_tydeas_	yes	19:38
johnw	ok, just drop that then :)	19:38
_tydeas_	what?	19:38
johnw	that option	19:38
	← halogenandtoast left	19:38
_tydeas_	and reclone?	19:38
johnw	yes	19:38
	→ niko joined	19:38
	← pklingem left	19:40
	→ enherit joined	19:40
	← sunbright left	19:41
	← JeffJohnson left	19:41
	→ MetaCosm joined	19:43
	← enherit left	19:45
redondos	what's a straightforward way of getting a list of just the hashes of the commits between two commits?	19:45
	← rubydiamond left	19:45
	→ lemonchicken joined	19:45
redondos	i want to get that list, and cherry pick all of them. maybe there's an easy way of doing that. it's not a rebase, since the current branch and the one I'm getting the commits from are not related in the history	19:46
theoros_	rev-list	19:46
redondos	ty	19:46
theoros_	you can use rev-list plus xargs -n 1	19:46
	→ rseifert joined	19:47
wereHamster	redondos: if you want to apply a list of commits to a branch, that can be done with rebase	19:48
	← froschi left	19:48
	← niko left	19:48
adymitruk	Git rebase --onto new start end	19:49
	→ niko joined	19:49
redondos	using rebase, mustn't the two branches have a common point in history?	19:49
	← vu3rdd left	19:49
	← tatsuya left	19:49
	→ JeffJohnson joined	19:50
	← JeffJohnson left	19:50
	→ JeffJohnson joined	19:50
wereHamster	no	19:50
	rebase literally takes a list of commits a..b and applies them onto another commit c	19:51
	→ bcardarella joined	19:51
redondos	fabulous. now how do i feed that list of commits to rebase?	19:51
wereHamster	what adymitruk said... ?	19:51
redondos	gotcha	19:51
theoros_	git rebase --onto a b c, means take the list of commits reachable by c, but not those reachable by b, and apply them to a	19:51
redondos	thanks all three.	19:51
	→ enherit joined	19:52
adymitruk	If you're going to want the initial commit, use the --root option	19:53
	← d0k left	19:53
adymitruk	ot	19:53
	the start commit is omitted otherwise	19:53
	← aresnick left	19:53
	→ roger_padactor joined	19:54
	← mbroeker left	19:54
roger_padactor	how would i change the user to use to fetch the repo from a server i used a different user that a should have	19:54
wereHamster	you mean the user in the url?	19:55
	→ nikolavp joined	19:55
	→ Nemurenai joined	19:55
roger_padactor	well for the clone ssh i didn't put in a username and it defaulted to the one on my computer which i coincidentally had on the server so i used it. but i shouldn't have	19:56
	→ tr3buchet\|afk joined	19:56
wereHamster	git config remote.origin.url <url with the correct username>	19:56
nikolavp	can someone help me with this input/output error	19:56
	http://gist.github.com/424376	19:56
	← krh left	19:57
nikolavp	i cannot commit anything. Every other sub command works as expected	19:57
	tr3buchet\|afk → tr3buchet	19:57
wereHamster	nikolavp: does 'git add -u; git commit' work?	19:57
roger_padactor	wereHamster: whats the correct format username:ssh://url/path/to/git	19:57
	→ ehabkost_ joined	19:58
	← rolfb left	19:58
wereHamster	roger_padactor: man git-fetch	19:58
Gitbot	roger_padactor: the git-fetch manpage can be found at http://git.or.cz/man/git-fetch	19:58
wereHamster	-> GIT URLS	19:58
s34n	I copied some files into a subdirectory of my working tree. It turns out that the directory I copied already had a .git	19:58
nikolavp	no wereHamster it is the sae	19:58
	← ehabkost left	19:58
nikolavp	same*	19:58
s34n	I deleted the .git from the subdirectory, but I still can't git add the subdirectory	19:59
WebDragon	I've got a commit on the tip a test branch that I want to remove, however I want to keep the changes applied by that commit, in the working tree, so I can make more modifications to them and then later re-enter a commit. so should I just use --amend ?	19:59
	→ mar77i joined	19:59
theoros_	git reset HEAD^	19:59
wereHamster	s34n: any errors?	19:59
theoros_	or git reset --mixed HEAD^ if you'd like to make sure	19:59
	→ jmcantrell joined	19:59
WebDragon	theoros_: that was to me?	19:59
theoros_	WebDragon: yes	20:00
	"Resets the index but not the working tree (i.e., the changed files are preserved but not marked for commit) and reports what has not been updated. This is the default action."	20:00
s34n	wereHamster: no errors. it just doesn't add anything	20:00
wereHamster	nikolavp: maybe your editor is weird, try setting GIT_EDITOR=vi or so	20:00
WebDragon	cheers	20:00
	→ priidu_ joined	20:00
s34n	git status tells me nothing to commit (working directory clean)	20:00
cbreak	more weird than vi? :)	20:01
	what's the name of the subdirectory?	20:01
s34n	ah. I get an error if I cd into the subdir and git add *	20:02
	it tells me that the contents are "in a submodule" that matches my current directory	20:02
nikolavp	thank wereHamster that worked out :)	20:02
	→ mbroeker joined	20:02
nikolavp	wereHamster: can i read what is the current editor in another terminal	20:02
	so i can see what is not working correctly	20:03
	echo $GIT_EDITOR is not set	20:03
wereHamster	nikolavp: echo $VISUAL $EDITOR $GIT_EDITOR	20:04
	→ drako joined	20:04
wereHamster	I think there is a way to get the current editor that git would use, but I don't know how	20:04
	→ rayners joined	20:04
	← jmcantrell left	20:04
nikolavp	only $EDITOR is set and it is vi	20:05
theoros_	trace?	20:05
	← drako left	20:05
	→ drako joined	20:05
wereHamster	nikolavp: git var GIT_EDITOR, that's what git will use	20:06
theoros_	strace your git commit and paste it i suppose	20:06
	→ mattwynne joined	20:06
nikolavp	it is somehow set to editor :O	20:06
theoros_	GIT_EDITOR=editor ?	20:06
nikolavp	yeah	20:06
theoros_	uhh	20:06
mattwynne	So I'm trying to use git-svn with my friend	20:06
theoros_	so what does that even mean	20:07
mattwynne	we're getting into quite a mess when we try to push / pull between our git repos as well as SVN	20:07
nikolavp	maybe ubuntu has some weird defaults -.-	20:07
mattwynne	any advice?	20:07
	→ robinr joined	20:07
wereHamster	mattwynne: ditch svn, migrate completely to git	20:07
mattwynne	wereHamster: not possible unfortunately, we're working with a publisher who is stuck on SVN	20:08
	but we both want to use git	20:08
sjohnson	what does "git reset --soft" do, if it doesnt update the index or change your modified files?	20:08
theoros_	what kind of mess are you getting into?	20:08
	← project2501a left	20:08
theoros_	sjohnson: i wonder that myself, often	20:08
	→ felipec joined	20:08
adymitruk	Changes the current commit pointer. You will have additional changes that reflect the difference of where you're now	20:11
theoros_	that isn't --soft	20:11
s34n	if I type git submodule, it tells me there is "No submodule mapping found in .gitmodules for path [path to my imported directory]"	20:11
theoros_	that's --mixed	20:11
adymitruk	Ah	20:11
	→ altogether joined	20:12
s34n	so it thinks there is a submodule, and blocks my ability to fold it into the current repo	20:12
theoros_	maybe it's meant to be sort of a git reset --dry-run?	20:12
sjohnson	thats what i thought too	20:12
adymitruk	So what	20:12
sjohnson	kinda like, does dick all	20:12
Caelum	how do you apply a git patch	20:12
wereHamster	Caelum: git apply <patch> ?	20:12
	→ mattikus_ joined	20:12
	← mattwynne left	20:12
s34n	but the submodule def seems corrupted (probably something to do with rm .git in that dir)	20:12
wereHamster	or git am [-3] <patch>	20:12
	← robinr left	20:13
Caelum	wereHamster: thank you!	20:13
	→ project2501a joined	20:13
	← mattikus left	20:14
	← resmo left	20:14
Caelum	huh, I already applied this patch, but I don't remember applying it	20:14
wereHamster	that's not git's fault..	20:14
Caelum	not blaming git	20:15
wereHamster	:)	20:15
	← unreal left	20:15
	→ unreal_ joined	20:15
	→ giskard_ joined	20:16
	unreal_ → unreal	20:16
	← ExtraSpice left	20:16
	← nevyn^_ left	20:17
	→ giskard__ joined	20:17
	← giskard left	20:18
	→ pcc1 joined	20:18
	→ tvw joined	20:18
	Caelum reapplies it properly	20:18
pcc1	it there an easy way to check the existence of a ref in perl (perhaps using Git.pm)? this needs to check unpacked and packed refs	20:18
	→ mattwynne joined	20:19
	_chevdor → chevdor	20:20
roger_padactor	Im confused i cloned my repo and locally i added a new dir i did a git add . then a commit but it says nothing to commit. i just created the dir?	20:20
	← giskard_ left	20:20
wereHamster	roger_padactor: git tracks content. And directories with no files in them have no content	20:21
DrNick	roger_padactor: git doesn't support empty directories and nobody has cared enough to fix it	20:21
roger_padactor	hahaha ok that would explain it :)	20:21
mar77i	argh, my git tree is falling into pieces every time... :(	20:21
Arrowmaster	just put an empty .gitignore in the dir	20:21
	if you really need it to track an empty directory	20:22
	← malumalu left	20:22
	← dkobia left	20:23
	← urkud_ left	20:24
	← razvand left	20:24
johnw	touch ".keep" inside empty directories you want to track	20:25
	→ justin-george joined	20:25
johnw	then check that empty file in, and voila	20:25
_ikke_	johnw: Why not just an empty .gitignore file	20:26
johnw	only because i tend to have only one of those per project	20:26
	but sure, any filename works	20:26
	→ shade__ joined	20:26
	← niko left	20:26
	→ giskard joined	20:27
	← giskard__ left	20:28
	← domcat left	20:28
	← smuf left	20:28
	shade__ → \shade	20:29
	\shade → \shade\	20:30
yrlnry	Why can't git track an empty directory? What technical limitation prevents one from making an empty tree?	20:30
wereHamster	not a technical limitation	20:30
	→ urkud_ joined	20:30
	→ giskard_ joined	20:30
	→ Bass10 joined	20:30
johnw	it's part of the design	20:30
wereHamster	just nobody willing to do the work	20:30
johnw	git tracks content, not containers	20:30
adymitruk	Just to give critics something to talk about :)	20:31
johnw	and people something to ask here :)	20:31
AAA_awright_	I thought trees and blobs were all types of containers	20:31
	AAA_awright_ → AAA_awright	20:31
yrlnry	But git has tree objects. A tree has a list of (name, mode, sha1) triples. Why can't it have an empty list?	20:31
wereHamster	it can	20:31
	← Bass10 left	20:32
yrlnry	So why can't there be an object in a git repository that represents a tree with no triples, and why can't some other tree contain that object under some name, say "my_empty_dir">	20:32
	?	20:32
	← giskard left	20:33
	→ Bass10 joined	20:33
wereHamster	because nobody cares that much to write the patch to enable it. There's lots of places in the code that needs to be checked to make sure nothing will break	20:34
yrlnry	Why wasn't it enabled from the beginning?	20:34
wereHamster	because git tracks content	20:35
ereslibre	hi there. does bare repos need maintainance (git gc) ?	20:35
johnw	here, let me say it again: because git tracks content	20:35
	← giskard_ left	20:35
AAA_awright	The git FAQ iirc says you can just place a .gitignore file that matches everything, since obviously there is going to be non-tracked content in the directory	20:35
yrlnry	git also tracks trees. Why doesn't it track empty trees?	20:35
johnw	good idea	20:35
	→ dgs joined	20:36
	← univate left	20:36
AAA_awright	So the answer is, then, because Git was designed to track content not files, and it enforces that, not that it's a technical limitation	20:36
ereslibre	yrlnry: read the git manual at git-scm.com, each commit is a pointer to a tree, and trees point to other trees and to blobs (possibly other things i'm missing), you can't add an empty dir	20:36
wereHamster	yrlnry: no, it dosen't track trees. It stores them in the object database, but only for the purpose to save names and permissions of blobs	20:36
	→ domcat joined	20:36
	← domcat left	20:36
ereslibre	tracking an empty dir makes just no sense	20:37
yrlnry	if it makes so little sense, why is it a FAQ?	20:37
	Who are all these crazy confused people who want to track empty directories?	20:37
ereslibre	yrlnry: people who come from svn and cvs	20:38
	← DavidKlein left	20:38
sjohnson	anyone here read pro git?	20:39
wereHamster	yrlnry: git was initially written to manange the linux kernel source. Source means content. So at the beginning there was no reason to store empty directories. They are usually only needed by the build system or such, and those systems can create the directory as-needed.	20:39
	← AlexanderS left	20:39
_ikke_	sjohnson: I did	20:39
ereslibre	yrlnry: in the dvcs world (the most common: git, hg, bzr) only bzr allows you to track empty dirs	20:39
sjohnson	_ikke_: did you read it online or just buy the book? i have a feeling that the book might be updated, so it's better to read it online.	20:39
johnw	what would a patch to create an empty dir look like?	20:40
_ikke_	sjohnson: Read it online	20:40
	→ giskard joined	20:40
yrlnry	I'm still missing something crucial here. The git data structures have no trouble representing an empty tree, so what's the problem? Was it written in at the begnning as a special case, or what?	20:40
	← jrmuizel left	20:40
sjohnson	_ikke_: cool thanks	20:40
	← Jferro left	20:40
	ereslibre still wonders if he has to setup a cron to run "git gc" on server side repos (bare ones)	20:41
	→ tjholowaychuk joined	20:41
johnw	yrlnry: oh, they do? how do you know that? I've never seen an empty tree	20:41
wereHamster	yrlnry: lots of git code probably assumes that trees can't be empty. So one would need to review large parts of the source to make sure nothing breaks. And nobody ever bothered to do the work	20:41
	← mbroeker left	20:41
	→ stephenjudkins joined	20:42
tjholowaychuk	im writing a CI server, I was wonder what the best approach is for checking out a specific revision? git pull origin master && git checkout rev ? or something like git pull origin master rev ?	20:42
johnw	tjholowaychuk: the former	20:42
wereHamster	ereslibre: I run it manually, about once a year :) But my repos don't have that much traffic	20:42
ereslibre	wereHamster: thanks :)	20:43
sparr	wereHamster: directories are, themselves, content. having to put a placeholder file to preserve an empty directory is a silly throwback to things like pkzip in the early 1990s.	20:43
wereHamster	tjholowaychuk: git remote update; git checkout <commit-ish>	20:43
stephenjudkins	tjholowaychuk: `git fetch` to fetch all changes from the server; `git pull` is roughly equivalent to `git fetch && git merge origin/[BRANCH]`	20:43
ereslibre	wereHamster: thought only had to be ran on tracking repos	20:43
yrlnry	johnw: I don't know, but werehamster claimed at 16:31 that that was the case.	20:43
tjholowaychuk	johnw: is there a way around creating the anonymous branch? otherwise I have to checkout master again after	20:43
	→ jrmuizel joined	20:43
johnw	tjholowaychuk: use git checkout -b NAME REV	20:43
	make your own branch	20:43
	← Nemurenai left	20:44
wereHamster	yrlnry: it's theoretically possible to create an empty tree object, just like you can create an empty blob object. But I never claimed that git knows how to handle that	20:44
sparr	johnw: what does a patch look like to create an empty file? (not rhetorical, I actually don't know)	20:44
johnw	sparr: I think it's just like an executable bit change	20:44
	that's kind of content-less too	20:45
_ikke_	sparr: wereHamster said it, it's not just one single patch, it needs to be fixed globaly	20:45
	← tjholowaychuk left	20:45
	← berserker left	20:45
wereHamster	sparr: http://sprunge.us/TZEZ	20:45
	→ vitor-br joined	20:45
	← adymitruk left	20:45
	→ Jferro joined	20:45
wereHamster	sparr: why do you want a checkout have an empty directory? For what purpose?	20:46
sparr	mode 100644	20:46
	what is the "100"?	20:46
	← luckiestmonkey left	20:47
sparr	which flags are those?	20:47
	→ giskard_ joined	20:47
	→ luckiestmonkey joined	20:47
wereHamster	git internal flags, one bit is used to indicate a submodule for example	20:47
	← [Po]lentino left	20:47
WebDragon	one bit indicates whether or not you were caffeinated when you made that commit	20:47
_ikke_	lol	20:48
WebDragon	one bit's there just in case the other bit was poisoned	20:48
_ikke_	the last bit is for checking the other bits	20:48
WebDragon	<that last one read in Peter Lorre's voice>	20:48
	← hyperair left	20:48
_ikke_	WebDragon: Peter Lorre doesn't ring a bell	20:49
	← giskard left	20:49
WebDragon	that's what you get for being too young	20:49
	→ Twisted_Mentat joined	20:49
	→ giskard joined	20:49
WebDragon	http://en.wikipedia.org/wiki/Peter_Lorre	20:50
	← gbacon left	20:50
	← d7777777 left	20:50
WebDragon	he did this one sanitarium skit with Abbott and Costello that was priceless	20:50
Ilari	sparr: 100644 is 10 + 0644, that is regular file that is not executable.	20:50
	→ berserker joined	20:50
	← ereslibre left	20:51
WebDragon	Lorre played the role of Joel Cairo in The Maltese Falcon (1941) and portrayed the character Ugarte in the film classic Casablanca (1942)	20:51
	← Belna left	20:52
	← ivan left	20:52
	← giskard_ left	20:52
	← adamv left	20:53
	← khmarbaise left	20:53
WebDragon	if you've ever seen either of those two Bogart films, you've seen Lorre	20:54
	← giskard left	20:55
	→ ivan`` joined	20:55
	→ mjf joined	20:55
	→ dkobia joined	20:56
_ikke_	WebDragon: Nup	20:56
	← felipec left	20:56
	← NET\|\|abuse left	20:57
WebDragon	had a very distinctive voice. got typecast as a sinister foreigner a lot in movies	20:57
_ikke_	i've read that	20:58
WebDragon	s/movies/early Hollywood movies/	20:58
	← jherdman left	20:58
WebDragon	Lorry was the epitome of 'that creepy weird guy'	20:58
IslandUsurper	_ikke_, seen Disney's Aladdin? do you remember the line where the Genie lists what he won't do for a wish? "I can't bring people back from the dead" <-- Peter Lorre impression	20:58
WebDragon	:)	20:58
	→ cannonball joined	20:58
WebDragon	IslandUsurper: hey, that's right	20:58
_ikke_	IslandUsurper: I've seen that movie, but can't remember that line anymore	20:59
	IslandUsurper: Maybe youtube has that line	20:59
	← IslandUsurper left	21:00
	→ adymitruk joined	21:00
	← radioman-lt left	21:00
	→ giskard joined	21:01
WebDragon	http://www.youtube.com/watch?v=84K6Zvvem-Y&feature=player_embedded#!	21:01
	← kar8nga left	21:01
	→ schulty joined	21:01
_ikke_	WebDragon: already found that one	21:02
	← jowls left	21:02
	← schulty left	21:04
	← glebd left	21:05
	→ hyperair joined	21:06
	→ eletuchy joined	21:07
	→ paddor joined	21:07
	← Joeconyers left	21:08
	→ felipec joined	21:09
	→ gbacon joined	21:09
	← lhz left	21:11
	← paddor left	21:11
WebDragon	http://www.youtube.com/watch?v=l9Gb96IjvsI&feature=player_embedded#! <-- this is actually what I was referring to -- this is from an old radio show	21:11
	→ paddor joined	21:11
WebDragon	has the 'just in case the blue pill was poisoned' line I was referring to	21:12
	← paddor left	21:12
_ikke_	which part?	21:12
jefferai	sitaram: stupid question -- how do you delete repos? both in the case of wildcard (e.g. can a user delete their own repo?) and non-wildcard (do you just wipe it out in $REPO_BASE and remove the config, or is there an automated way?)	21:12
	← rseifert left	21:12
adymitruk	Rm .git -r -f	21:13
jefferai	adymitruk: via gitolite	21:13
	← muszek left	21:14
WebDragon	hahhahhaha "that's the first robin I've heard this spring" -- I forgot that line	21:14
	← johnjay left	21:16
	← yrlnry left	21:16
WebDragon	the moment's shortly around 5min into it	21:16
	→ johnjay joined	21:17
	→ kpreid_ joined	21:17
	← kpreid_ left	21:18
	← chrowe left	21:18
	← engla left	21:18
	→ kpreid_ joined	21:19
	→ mbroeker joined	21:19
	← nuoHep left	21:20
	← priidu_ left	21:20
	← priidu left	21:20
	← flazz left	21:21
	← kpreid_ left	21:21
	← ehabkost_ left	21:22
	← giskard left	21:22
	→ ispeak_ joined	21:23
	← WebDragon left	21:24
	← pheaver left	21:24
	→ nuoHep joined	21:25
	← altogether left	21:25
	→ paddor joined	21:26
	→ GodEater joined	21:27
	← GodEater left	21:27
	→ GodEater joined	21:27
	← __iron left	21:29
	→ giskard joined	21:29
	← nikolavp left	21:30
_eric	what is the most efficient command to print what branch you're currently in?	21:31
	the 'git branch \| grep "^*"' thing rubs me the wrong way	21:32
jast	git symbolic-ref HEAD is the most low-level one, I suppose	21:32
_eric	sweeet	21:32
	← raichoo left	21:33
	→ giskard_ joined	21:36
	← Nickeeh left	21:37
	← yopp left	21:37
	→ marioxcc joined	21:38
marioxcc	hi all	21:38
	← giskard left	21:38
	← paddor left	21:38
marioxcc	i just did try to make a merge (there are some conflicts) but innmediatly i realized i don't want to merge yet	21:38
	is there a simple way to undo the merge (as git rebase --abort, for example)	21:38
jast	git reset --hard	21:38
marioxcc	oh	21:39
	yes	21:39
	why didn't I try that before?	21:39
	lol :)	21:39
	jast: thanks	21:39
	→ glebd joined	21:40
	→ jds joined	21:40
	→ Philwelch joined	21:41
	→ airborn joined	21:42
	← ToxicFrog left	21:43
	← ericdc left	21:45
	← jds left	21:46
	← altrux left	21:48
	← bcardarella left	21:48
	← ChUrP left	21:48
	← textarcana left	21:49
	textarcana_ → textarcana	21:49
	← urkud_ left	21:51
	→ sixteneighty joined	21:53
	← ispeak_ left	21:53
	← giskard_ left	21:54
	← vitor-br left	21:55
	marioxcc → marioxcc-AFK	21:55
	← irc2samus left	21:56
	← mbroeker left	21:56
	← sixteneighty left	21:57
	→ Transformer joined	21:59
	→ giskard joined	22:01
	← Transformer left	22:01
	→ hasu joined	22:04
	marioxcc-AFK → marioxcc	22:04
infid	git is the bee's knees. the wit's t*ts. the nerd's words. the bit's commits.	22:05
	i feel like a repository surgeon now	22:05
	← unknown left	22:05
	← marioxcc left	22:06
infid	a repo man if you will	22:06
	→ marioxcc joined	22:06
	→ EricInBNE joined	22:08
	→ sixteneighty joined	22:12
	← LRN left	22:12
s34n	I have a subdirectory in my repo which git seems to think is a submodule.	22:13
	How do I convince git that it is not a submodule?	22:13
	so that I can add it?	22:14
	→ giskard_ joined	22:14
	← giskard left	22:14
	→ LRN joined	22:14
adymitruk	Edit your .gitmodules file	22:15
s34n	adymitruk: there is no such file	22:15
	← johnw left	22:16
	→ giskard joined	22:16
wereHamster	s34n: do you have any other changes in the working tree or index?	22:17
	→ Wally joined	22:18
Wally	hey, can someone tell me how I go about doing "Svn update" using git	22:18
s34n	wereHamster: I barely init-ed the repo, then cloned it and noticed that the clone was missing a subdir	22:18
adymitruk	Check your cundith	22:18
	Config	22:18
s34n	wereHamster: so everything is brand new with only one commit	22:18
Wally	cundith :P	22:19
	→ zimbres joined	22:19
wereHamster	s34n: try deleting the index (rm .git/index), then git reset --hard and then try to add the dir again	22:19
	← giskard_ left	22:19
adymitruk	Swype keyboard on a nexus one :)	22:19
wereHamster	Wally: git pull	22:19
s34n	wereHamster: thx	22:19
	→ chris____ joined	22:19
adymitruk	Git fetch	22:19
Wally	cheers	22:19
s34n	actually, I just deleted .git and re-init-ed	22:20
	← tizbac left	22:20
Wally	git fetch svn?	22:20
s34n	that worked	22:20
chris____	hello everyone. does anyone have a second to discuss a workflow issue with me?	22:20
s34n	chris____: just ask the questions	22:20
chris____	thanks, s34n - will do	22:21
adymitruk	Any android git clients out there?	22:21
	Fooling for it always gets me the git repo obviously	22:21
	Googling	22:22
	→ giskard_ joined	22:22
mugwump	adymitruk: never heard of one - there was an iPhone app written in Obj-C by a github staffer	22:22
	← giskard left	22:23
chris____	I normally work on low-volume traffic WordPress websites but have a new client with 4 very high-traffic WordPress websites. In the past, I could safely download a copy of the MYSQL database from the production server, load it locally, develop locally, then restore the production database with the one I had edited locally. but with a lot of traffic, I cannot pull a copy of the DB from prod to local, edit and restore the remote db because I wi	22:23
	have already overwritten changes to the prod db which happened while I was editing...see the dilemma?	22:23
	anyone advise me on best practices here?	22:24
	← giskard_ left	22:24
	← Wally left	22:25
	← lresende left	22:26
NfNitLoop	chris____: not quite sure what that has to do with git...	22:26
chris____	how to keep my local (development) copy of the MYSQL db synced with the production copy?	22:26
wereHamster	-> #mysql ?	22:26
chris____	well, this is all within a git-repo environment	22:26
NfNitLoop	#WordPress?	22:26
	git doesn't know/care about MySQL.	22:27
	Or Wordpress.	22:27
	It just does source code.	22:27
chris____	didn't mean to be off target guys, just didn't know if there was some way I could be using Git for all of this	22:27
	← benjo2 left	22:27
chris____	thanks for the direction. I'll try mysql or wordpress channels.	22:27
wereHamster	use git for what? To store the mysql database?	22:27
NfNitLoop	chris____: #WordPress will likely have the best ideas re: WordPress development workflow.	22:28
chris____	thank you guys!	22:28
NfNitLoop	chris____: generally, you want to save every modification you make to a DB in development as a script that can be run to upgrade your production DB when it goes live.	22:28
_tydeas_	any ideas on this errors? http://pastie.org/991253	22:29
	← masterkorp left	22:30
	→ ericdc joined	22:30
	→ giskard_ joined	22:31
_tydeas_	the local repo has been constructed with `git svn clone -A author-map code.google.com/project	22:31
	→ icwiener_ joined	22:31
	← icwiener left	22:32
	← m104 left	22:33
NfNitLoop	_tydeas_: I think the first time you do a push, you have to specify it as master:master	22:34
	← mattikus_ left	22:35
NfNitLoop	because by default it tries to make sure there's an existing master to push to. (?)	22:35
	→ ph^ joined	22:35
_tydeas_	same error	22:35
	the clone is FROM SVN REPO	22:35
	→ mbroeker joined	22:35
	← airborn left	22:36
wereHamster	NfNitLoop: no, master is the same as master:master	22:36
	_tydeas_: it simply means that you don't have any branch named 'master'	22:36
adymitruk	any way to rewrite one commit really early in history?	22:36
wereHamster	adymitruk: rebase -i	22:36
_tydeas_	what can i do?	22:37
	git branch master	22:37
	← jrmuizel left	22:37
_tydeas_	chekout it ?	22:37
	I say it again this is a svn repo cloned WITH metadata	22:37
adymitruk	but that won't respect any merging and branching	22:37
NfNitLoop	_tydeas_: what does 'git branch' say?	22:37
	→ textarcana_ joined	22:38
_tydeas_	nothing	22:38
adymitruk	I want to preserve the history structure	22:38
_tydeas_	adymitruk: you want to go back in time?	22:38
adymitruk	But with one commit really early on altered	22:38
_tydeas_	you want to go back to a previous commit ? do you mind losing that commit ?	22:39
	← stephenjudkins left	22:39
adymitruk	I cannot lose the commit. It had other important things in there	22:40
	← NetHawk left	22:41
adymitruk	I basically need a utility to rewrite all the branches and their complete histories on top of the new altered commit	22:42
jjuran	adymitruk: man git-filter-branch	22:42
Gitbot	adymitruk: the git-filter-branch manpage can be found at http://git.or.cz/man/git-filter-branch	22:42
	← chevdor left	22:43
	→ EddieRingle joined	22:43
	← hyperair left	22:43
_tydeas_	how can i determine the refspec my repo has	22:44
wereHamster	_tydeas_: git branch -a	22:44
	← LiamH left	22:44
	← rayners left	22:44
	← coderdad left	22:44
	← jayne left	22:44
_tydeas_	got this remotes/git-svn so git push origin master:remotes/git-svn ?	22:45
	→ galderz joined	22:45
	← galderz left	22:45
	→ galderz joined	22:45
wereHamster	git push origin remotes/git-svn:refs/heads/master	22:46
	for example..	22:46
	→ ProperN[out] joined	22:46
	→ NetHawk joined	22:46
	← ProperN[out] left	22:46
	→ ProperN[out] joined	22:46
adymitruk	Will git filter-branch reach all nodes seen with gitk --all?	22:47
_tydeas_	i `git push origin remotes/git-svn` and killed this when you said about git push origin remotes/git-svn:refs/heads/master. Is this going to be a problem?	22:47
	← radarek left	22:47
wereHamster	adymitruk: it can, if you tell it to	22:48
	_tydeas_: define 'problem'	22:48
	though probably not	22:49
	← giskard_ left	22:49
	← ProperNoun left	22:49
	ProperN[out] → ProperNoun	22:49
_tydeas_	duplicates etc	22:49
adymitruk	What do I have to do? Just specify all branches and tags?	22:49
	→ ToxicFrog joined	22:49
	→ jayne joined	22:49
	← schmick1 left	22:49
	← Unode\|NA left	22:50
wereHamster	adymitruk: git filter-branch --one-of-the-filters (such as --commit-filter) <what to filter, such as HEAD or --all>	22:50
_tydeas_	no how can i git svn rebase	22:50
	and wok.	22:51
wereHamster	git svn rebase <enter> ?	22:51
_tydeas_	http://pastebin.com/3SQM0r1g	22:51
	result this	22:51
	← Determinist left	22:51
wereHamster	_tydeas_: git checkout -b master origin/git-svn	22:51
adymitruk	I want to remove a giant dir called libs which has a lot of DLLs	22:51
	It was introduced years ago	22:52
	← kipras left	22:52
wereHamster	adymitruk: then use te index filter, the man page has an example	22:52
	← TeckniX left	22:52
adymitruk	I would like to put that history into a sub module hosted on git hub instead	22:52
_tydeas_	don't work http://pastebin.com/ENn51QSU	22:52
adymitruk	So the code would still be good	22:53
_tydeas_	i don't have a master branch	22:53
adymitruk	It I wouldn't need to filter any thing lin the source code	22:53
	← sixteneighty left	22:53
wereHamster	_tydeas_: git checkout -b master unofficial-yii-framework-mirror/origin/git-svn	22:53
_tydeas_	Did you intend to checkout 'unofficial-yii-framework-mirror/unofficial-yii-framework-mirror/origin/git-svn' which can not be resolved as commit?	22:54
	this this time	22:54
wereHamster	_tydeas_: git branch -a	22:54
	← s34n left	22:54
	→ khaije\|selen joined	22:54
_tydeas_	i don't have a master branch	22:54
	this is for branch -a remotes/git-svn remotes/origin/master	22:55
	→ saimon joined	22:55
_tydeas_	git checkout -b remotes/origin/master remotes/git-svn ?	22:55
	→ s34n joined	22:55
wereHamster	_tydeas_: git help tutorial	22:56
_tydeas_	won't help	22:56
wereHamster	onece you are done with it, git checkout -b master remotes/git-svn	22:56
	alright, I'll be gone then	22:56
_tydeas_	i don't understand which point of git help tutorial will. sorry for beeing annoying	22:58
	→ madewokherd joined	22:58
khaije\|selen	Hi all, I'm setting up gitolite and I'm wondering can I define refex groups, such as @sensitivebranch = /refs/tags/ secretlab	22:59
_tydeas_	what's the diff remotes/master remotes/git-svn	22:59
	← AkumaStreak left	23:00
	← cannonball left	23:00
	→ florianb joined	23:00
saimon	hi...I've got an issue I'm not sure how to solve: I've just added a submodule to my project. Adding the submodule went well. However, this submodule has a number of submodules itself. After I did: git submodule update --init --recursive in the main submodule, git status now reports that certain files within the nested submodules are now dirty. The reason is because I have git config core.autocrlf=input. I'd like to just ignore these changes but	23:00
	matter what I try they still appear as dirty. Within each dirty nested submodule, I've tried git reset , HEAD git reset --hard HEAD. I'm a bit stuck right now.	23:00
	→ freshtonic joined	23:00
	← florianb left	23:00
_tydeas_	wereHamster: can you tell me which point of git help tutorial suites my ? please	23:01
	← orafu left	23:02
	→ orafu joined	23:02
	→ Determinist joined	23:03
	→ jrmuizel joined	23:03
	→ Heimidal joined	23:04
	← adymitruk left	23:05
_tydeas_	when trying to create branch i get	23:06
	http://pastebin.com/eH6Xw18Y	23:06
frogonwheels	_tydeas_: I _think_ git branch master HEAD should disambiguate it	23:08
	← aspotashev left	23:08
frogonwheels	Is there no master branch?	23:09
	→ dj_tjerk joined	23:09
frogonwheels	_tydeas_: hmm.. what does git branch -a give you?	23:09
	← EricInBNE left	23:09
_tydeas_	no master	23:09
	created one with HEAD	23:09
	← Vojto left	23:09
_tydeas_	* master	23:09
	remotes/origin/master	23:09
	remotes/git-svn	23:09
	remotes/origin/master	23:09
frogonwheels	oooh that's exciting - two origin/master branches? umm.. wtf? Hopefully I'm misunderstanding.	23:10
	→ EricInBNE joined	23:10
frogonwheels	gg	23:10
	kids fighting in the kitchen :)	23:10
_tydeas_	what can i do?	23:11
wereHamster	frogonwheels: one is a local branch (refs/heads/remotes/origin/master), the other a remote branch (refs/remotes/origin/master)	23:11
	← hakunin left	23:12
	← aziz left	23:12
wereHamster	if he had colors enabled, he'd see that one is in green color and the other one is red	23:12
	or, one normal color and the other one red	23:13
_tydeas_	for git reflog i get	23:13
	edbd496 HEAD@{0}: checkout: moving from remotes/origin/master to master	23:13
	edbd496 HEAD@{1}: checkout: moving from remotes/origin/master to edbd4964017b3c854baf5c5c3e9d73b87570e479^0	23:13
	a13410b HEAD@{2}: checkout: moving from master to remotes/origin/master	23:13
	← galderz left	23:14
_tydeas_	can i hard reset this mess?	23:14
	← MattDiPasquale left	23:14
	← Titosemi left	23:15
	→ DrAk0 joined	23:16
	← drako left	23:16
_tydeas_	don't know what should i do	23:18
	← vmiklos left	23:18
saimon	guys...Can someone spare a minute or so into explaining how I can reset autocrlf changes within nested submodules?	23:19
	← ph^ left	23:19
s34n	I setup a remote to a repo that was originally cloned from mine. I want to pull from it.	23:19
	→ vmiklos joined	23:20
	→ adymitruk joined	23:21
s34n	but it tells me I have to specify a branch for the pull	23:22
	← ironfroggy left	23:22
	← lucsky left	23:22
s34n	I did a set-head -a on the remote	23:23
saimon	I wonder wether deleting .git/index would do it...	23:24
s34n	and git remote show says the remote branch is master	23:24
wereHamster	set-head only affects the local repo	23:24
	→ Bac0n joined	23:24
	← pantsman left	23:24
wereHamster	s34n: you can't change HEAD in the remote repo from your local host	23:24
saimon	nope :(	23:25
	→ radarek joined	23:25
	← Jferro left	23:25
saimon	ok so I've tried to delete the index, git checkout /path/to/file, git stash, git reset HEAD, git reset --hard HEAD....Google isn't helping much.	23:26
Arrowmaster	saimon: what command are you trying? just 'git pull'?	23:28
	er	23:28
	s34n not saimon	23:28
saimon	:(	23:28
s34n	Arrowmaster: git pull [remotename]	23:29
	wereHamster: I'm trying to set which branch of the remote is used for pulls, etc.	23:29
	wereHamster: I thought that set-head would do that for me	23:30
Arrowmaster	git config branch.master.merge refs/heads/master	23:31
	oh and git config branch.master.remote [remotename]	23:31
	→ roastb joined	23:31
	← _tydeas_ left	23:31
	← roastb left	23:32
Arrowmaster	saimon: what kinda autocrlf changes?	23:32
	→ roastb joined	23:33
saimon	Arrowmaster: e.g. "warning: CRLF will be replaced by LF in FABridge.js"	23:33
s34n	it says "Local ref configured for 'git push'"	23:33
	does that mean I can't pull?	23:33
saimon	Arrowmaster: after I did this: git submodule update --init --recursive	23:33
Arrowmaster	saimon: are you on windows?	23:34
saimon	Arrowmaster: osx	23:34
	Arrowmaster: so 2 of my nested submodules now have files in them that are dirty. I had core.autocrlf set globally to input (as per github suggestion)	23:35
Arrowmaster	saimon: why do you have autocrlf enabled then?	23:35
	ok	23:36
saimon	Arrowmaster: http://help.github.com/dealing-with-lineendings/	23:36
	← AAABeef left	23:36
Arrowmaster	well you need to either disable it or if you have access to those submodules repos just commit and let them get removed	23:37
	← Haarg left	23:37
Arrowmaster	since they likely shouldnt be there	23:37
Bac0n	i have problem setting up tracking branch with different name then remote "git checkout -b testing origin/server"	23:38
Arrowmaster	wow that github page is completely wrong	23:38
saimon	Arrowmaster: I don't have access to them...I've set autocrlf to false in the parent project but I guess I'll need to disable it globally. But once I do, how can I get rid of those changes? i.e. will git reset --hard HEAD within each nested sbumodule do the trick?	23:38
	→ Remixman joined	23:38
Arrowmaster	saimon: it should	23:38
saimon	Arrowmaster: really...someone should tell them...	23:38
	→ Haarg joined	23:39
Arrowmaster	i will	23:39
saimon	Arrowmaster: will try that then thanks..	23:39
Arrowmaster	i know tekkub, one of their support monkeys so ill yell at him	23:39
Bac0n	setting up tracking with same name works "git checkout -b server origin/server"	23:39
	it say "Branch testing set up to track remote branch server from origin."	23:40
saimon	Arrowmaster: that did it thanks...So what is the correct setting? false? (my logic says line endings should be handled bu editors, not git)	23:40
Bac0n	but git push only say "Everything up-to-date"	23:40
wereHamster	Bac0n: man git-config -> push.default	23:41
Gitbot	Bac0n: the git-config manpage can be found at http://git.or.cz/man/git-config	23:41
	← Paraselene__ left	23:42
	← tedoc2000 left	23:42
	← roastb left	23:44
	← gbacon left	23:44
Bac0n	wereHamster: thanks alot .... worked	23:47
Arrowmaster	saimon: correct setting on osx would be false or input, the problems only show up when somebody commits crlf line endings to a repo, the proper solution is just about always to remove them	23:47
sitaram	jefferai: there is no code within the main code of gitolite to remove a repo and there never will be... at least for non-wildcard repos. For wildcards repos you can it up using admin-defined commands; see http://github.com/sitaramc/gitolite/blob/pu/doc/admin-defined-commands.mkd#rmrepo	23:48
	← ajpiano left	23:48
	← icwiener_ left	23:48
	← adymitruk left	23:48
Arrowmaster	saimon: you could change the safecrlf option to warn and it would just give a message like before but not block anything	23:48
	→ ajpiano joined	23:48
saimon	Arrowmaster: ok so I had it set to input but since I can't control external nested submodules I think I prefer to leave it set at false.	23:48
	Arrowmaster: ah that's nice to know..thanks	23:49
Arrowmaster	saimon: if you dont plan on making any commits to a repo just turning off safecrlf for it will make those messages completely go away	23:49
	← tvw left	23:49
Arrowmaster	although i think it will still show unstaged changes on a git status	23:50
	← thomas_b left	23:50
Arrowmaster	so turning it off for those is probably best	23:50
saimon	Arrowmaster: thanks...yeah I'm just going to turn it off...	23:51
Arrowmaster	but the recommendation of setting autocrlf to input is a good one for any repo you make commits to	23:51
sitaram	khaije\|selen: should just work... didn't it?	23:51
	→ Unode\|NA joined	23:51
	→ thomas_b joined	23:52
	Unode\|NA → Unode	23:52
Arrowmaster	if i could program in C i would probably have submitted a patch by now for a better way to keep crlf line endings out of repos	23:52
khaije\|selen	sitaram: I didn't try since I wasn't sure and it has potentially huge implications	23:52
	→ ajpi joined	23:53
khaije\|selen	But I'll add it to my nodes, gitolite is very impressive cheers!	23:53
	Err... "nodes" should be "notes" :)	23:54
sitaram	khaije\|selen: I just noticed it's not documented; thanks for reminding me :)	23:54
	→ kpreid_ joined	23:54
	→ pklingem joined	23:55
khaije\|selen	sitaram: it is implicitly, (I saw it after I asked), the .conf document says the format "accepts [zero or more refexs]"	23:56
	← ajpiano left	23:56
sitaram	aah	23:56
	→ Cynos joined	23:57
sitaram	no that is not a refex group; that's plain refex list in the R/RW/... line	23:57
Cynos	Hi all, just a quick question, I just committed something on one branch that I need on another branch and I can't merge them just yet. If I cherry pick the commit from branch A onto branch B, will it mess things up when I later merge A into master and then B into master?	23:58
khaije\|selen	oh I see, so I'm glad I asked then!	23:58
	→ welly_ joined	23:58
sitaram	the section under "GROUPS" is generic enough (does not imply that you can use it only for repos and users) but maybe an example would help	23:58
plediii	is there a way to recover changes lost due to a checkout?	23:59

Logs Search ←Prev date Next date→

Channels Documentation