IRCloggy #git 2012-02-07

Provider of IRC logs since 2005.

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

	2012-02-07
monstaRtruck	y did i understand svn	00:00
EugeneKay	Are you specifying gitolite@hostname in your ssh url on the remote box?	00:00
monstaRtruck	but not git	00:00
jast	Bearcat: locally, what's the output of: git remote -v	00:00
monstaRtruck	well some stuff	00:00
Bearcat	what i'm wondering is if there's a conflict between teh way gentoo sets up git normally and gitolite. I mean do i want git starting up as a daemon as usual?	00:00
__machine	jast: thanks... where the named ref was N moves ago... that makes sense...	00:00
amgarchIn9	monstaRtruck: this is also a "number", 0087e8b180b8628bf986e7cdc16a3dcd1f6070e7, but a very long one. They are not ordered, that is the difference.	00:00
monstaRtruck	how do i find that number	00:00
jast	monstaRtruck: git is quite different. here's a document briefly explaining the differences: http://git-scm.com/course/svn.html	00:00
monstaRtruck	basicly right now i do diff head to find out the changes i made	00:01
EugeneKay	gitolite is not a daemon, it's a set of scripts called when you authenticate to SSH as the user 'gitolite'	00:01
	→ xiaotian joined	00:01
monstaRtruck	so i do then do revert head?	00:01
	→ tatsuya_o joined	00:01
	← _guns_ left	00:01
EugeneKay	It hooks in to SSH via authorized_keys command="" syntax	00:01
amgarchIn9	when in dobts: git diff \| patch -R	00:02
Bearcat	jast: when i do that in the gitolite admin dir on my client i see: origin [email@hidden.address] (fetch)	00:02
	origin [email@hidden.address] (push)	00:02
	→ anjen joined	00:02
monstaRtruck	ok well where do i start	00:02
jast	right, that looks as it should	00:03
Bearcat	jast: thanks	00:03
jast	monstaRtruck: what are you trying to achieve right now?	00:03
monstaRtruck	how do i find the revision number	00:03
	of before i made any changes	00:03
EugeneKay	Looks right. su - git and see if you have gitoltie stuff in the homedir	00:03
	← twisted` left	00:03
	→ axl_ joined	00:03
Bearcat	EugeneKay: i do	00:03
	→ tommyvyo joined	00:03
	← axl_ left	00:03
EugeneKay	And in the .ssh/authorized_keys file?	00:03
jast	monstaRtruck: note I'm not asking "how are you trying to achieve what you want to achieve"... I'm asking "what do you want to achieve" ;)	00:04
Bearcat	EugeneKay: yes	00:04
EugeneKay	Well.... that all sounds right	00:04
monstaRtruck	i want to revert	00:04
jast	hmm	00:04
	meaning: you have uncommitted changes that you want to get rid of?	00:04
monstaRtruck	to code before i made any changes	00:04
EugeneKay	I think he's referring to hg's revert, aka reset	00:04
monstaRtruck	no its all commited	00:04
jast	oh, so you want to throw away those commits?	00:04
monstaRtruck	but when i push live it might get mesed up	00:05
	like lets say something breaks	00:05
EugeneKay	git revert means "undo"	00:05
monstaRtruck	id want to quickly revert to revision number	00:05
jast	there are many meanings of "undo" :)	00:05
monstaRtruck	before i made changes	00:05
jast	and then do what? :)	00:05
monstaRtruck	and then its fixed	00:05
jast	there are many ways to revert, and they all do somewhat different things	00:05
monstaRtruck	or atleast i can say i didnt break it	00:05
EugeneKay	Bearcat - `ssh git@hostname info` + pastebin?	00:06
	← aidalgol left	00:06
charon	btw !revert	00:06
gitinfo	"Revert" is a heavily overloaded term. Do you mean: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard the uncommitted changes in the working tree [git checkout -- .]; c) undo committing [git reset --soft HEAD^, but type !rewriting_public_history] ?	00:06
	← baburdick left	00:06
jast	hey, who came up with 'checkout -- ' when there's the perfectly good 'reset --hard'? :}	00:06
	+.	00:06
Bearcat	EugeneKay: will do. I'll put some other info there too instead if pasting it here	00:06
EugeneKay	;-)	00:06
	→ Splash joined	00:07
RandalSchwartz	checkout probably respects dirty files	00:07
	or maybe even merges them	00:08
	→ shergill joined	00:08
Bearcat	EugeneKay: oh did you want me to run that from the client (i assume so)?	00:08
	← mandric left	00:08
monstaRtruck	well how do i find out the curent revision	00:08
jast	checkout doesn't merge anything	00:09
	the only difference is that it uses the files from the index instead of from HEAD	00:09
EugeneKay	Bearcat - well, using whatever SSH key you ran with gl-setup. Doesn't really matter "where" that is	00:09
jast	unstaged changes will still be clobbered	00:09
xil	hi, I did an interactive rebase and would like to retroactively specify -Xours. Do I have to abort and redo it or is there another way? "git rebase --continue -Xours" doesn't work	00:09
	← break57382 left	00:09
monstaRtruck	its not easy being this stupid	00:10
EugeneKay	xil - You can reset --hard back to where you were before the rebase, then rerun the rebase	00:10
amgarchIn9	monstaRtruck: git log	00:10
	← fly91 left	00:10
xil	EugeneKay: can't I just abort the rebase and rerun it?	00:10
	or is hard resetting and aborting not the same?	00:10
jast	xil: yes	00:10
EugeneKay	ctrl+c? I honestly dunno, I don't rebase :-p	00:10
	← jacekmigacz left	00:10
	→ baburdick joined	00:10
jast	rebase --abort and redoing it will work fine	00:11
	← baburdick left	00:11
jast	if the rebase is still in progress, that is	00:11
	→ forrestv joined	00:11
jast	if it isn't, you'll get an error :)	00:11
xil	jast: okay. Just checking that there isn't another way. It's harmless enough to abort and redo anyway	00:11
jast	yep, there isn't. :)	00:11
EugeneKay	It's really, really hard to "lose" a commit..... forgetting where you put it is another thing	00:11
monstaRtruck	only god can help me	00:12
	:~(	00:12
	← iocor left	00:12
Bearcat	EugeneKay: thanks for the help. paste is here: http://pastebin.com/z16eTsky	00:12
	→ ericmuyser joined	00:13
	→ jfisk1987 joined	00:13
RandalSchwartz	monstaRtruck - don't push unless you're sure you want to share it	00:13
jfisk1987	how could i force a push?	00:13
RandalSchwartz	that's what local testing is for	00:13
Bearcat	EugeneKay: i'm concered that git is running twice somehow..	00:13
EugeneKay	Bearcat - it looks like you have the (anonymous) git-daemon running. Service stop it and chkconfig(or wtf ever debian calls it) off	00:13
jfisk1987	im getting yelled at saying non-fast-forward updates are rejected	00:13
EugeneKay	That is COMPELTELY separate from gitolite	00:13
monstaRtruck	yeah but i want to learn	00:13
RandalSchwartz	after you share it, if you break something, you need to figure out how you broke it, and add more commits to fix that	00:13
EugeneKay	jfisk1987 - you need to fetch+merge(aka pull), then push	00:13
monstaRtruck	yeah but in svn i can just svn merge HEAD 123	00:14
EugeneKay	s/debian/gentoo	00:14
jfisk1987	eugeneKay: I want to ignore the changes from the server	00:14
monstaRtruck	.	00:14
EugeneKay	jfisk1987 - then you need to push -f to Force it	00:14
jfisk1987	i pushed a commit by mistake	00:14
	← checkers left	00:14
EugeneKay	See also !rewrite	00:14
gitinfo	[!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)	00:14
	← yuriks left	00:14
jfisk1987	thanks :)	00:14
	→ wuj joined	00:14
EugeneKay	And possibly !fixup as well	00:14
gitinfo	So you lost or broke something or need to otherwise find, fix, or delete commits? Look at https://gist.github.com/1612395 for full instructions. For hints type !fixup_hints in IRC. Remember: if you have pushed already, there are only a few things you can do without !rewriting_public_history (type that for more info)	00:14
	→ yuriks joined	00:15
EugeneKay	Bearcat - other than the git-daemon running, looks like gitolite is working 100% correctly. ;-)	00:15
	← auke- left	00:15
Bearcat	EugeneKay: thanks! Now why the heck is this loading twice...	00:15
EugeneKay	Master vs worker, probably. I've never touched daemon.	00:16
Bearcat	EugeneKay: The daemon has to be running for this all to work, doesn't it?	00:16
	→ xiangfu joined	00:16
	→ auke- joined	00:16
EugeneKay	Negatory. gitolite is explicitly non-daemon.	00:16
Bearcat	EugeneKay: ahh!	00:17
EugeneKay	<EugeneKay> gitolite is not a daemon, it's a set of scripts called when you authenticate to SSH as the user 'gitolite'	00:17
	<EugeneKay> It hooks in to SSH via authorized_keys command="" syntax	00:17
Bearcat	EugeneKay: well, geeze! Anthing to save me a few kb of memory :")	00:17
EugeneKay	It's more of separation of responsibilities. SSHd handles all the authentication+authorization and then tells gitolite who's calling.	00:18
	← alx- left	00:18
EugeneKay	You can, technically speaking, run gitolite outside of a SSH environment, or invoke it some way other than via authorized_keys, but those are beyond the scope of the man page	00:18
sitaram	EugeneKay: sshd handles authentication. Gitolite handles authorisation	00:18
	← barr5790 left	00:18
	← sudoquai left	00:19
EugeneKay	I meant authorization to the system. Auth to the repos is indeed gitolite. We are thus both correct.	00:19
jast	sshd handles part of authorisation, too	00:19
sitaram	EugeneKay: ok	00:19
	← glennpratt left	00:19
sitaram	Bearcat: the only other known form of authentication supported is http	00:19
jast	yeah, what everyone said.	00:19
	← martinjlowm left	00:19
Bearcat	ahh. So what would one need to run the git daemon for?	00:19
sitaram	one doesn't	00:19
EugeneKay	If you want to provide anonymous access outside of gitolite.	00:19
	Which is stupid, so turn it off.	00:20
Bearcat	oh! i see. ok	00:20
	yeah, no thanks.	00:20
nevyn	anonymous read access may be ok if that's what you want	00:20
	→ romeovs joined	00:21
	→ jasonbray joined	00:21
	→ jimubao joined	00:22
	← romeovs left	00:22
Bearcat	nah, this is just for my own personal use at this point. Really gitolite is overkill as it's just me using it, but i figured why not get it set up now in case that changes. Plus the security and learning more about keys and ssh was worth it alone.	00:22
xil	does anyone know a way for me to do a rebase so that all conflicts for all but the last commit in my current branch are resolved with -Xours, and it only asks me about the conflicts for the latest commit of mine?	00:25
	→ romeovs joined	00:26
xil	I was thinking of doing "git rebase -i -Xours develop" and then removing the last commit from the rebase, only to add it back on afterward and deal with the conflicts then	00:26
	but I don't know how to save the commit for adding on after the fact since a rebase will remove that commit from the history won't it?	00:27
	← Heimidal left	00:27
	← luigi left	00:27
CannedCorn	hrmm	00:27
	i'm not seeing my latest commits on github	00:27
	← jasonbray left	00:28
CannedCorn	why would that be the case	00:29
	it says i'm up to date when i push	00:29
Bearcat	whoo! i think it works!	00:30
CannedCorn	guys?	00:30
EugeneKay	CannedCorn - wrong branch?	00:31
CannedCorn	now it's there, nevermind	00:31
	strange	00:31
EugeneKay	Github site caching ;-)	00:31
CannedCorn	thanks EugeneKay	00:31
	← ipalaus left	00:32
	← Demego left	00:34
	→ sythe joined	00:35
	→ ShooterMG joined	00:35
	← Chib left	00:35
	← freezey left	00:36
	→ Chib joined	00:36
	← shergill left	00:36
	→ shergill joined	00:37
	→ jasonbray joined	00:37
	← jasonbray left	00:37
nevyn	xil: just do the rebase in a new branch and cherrypick the last commit over	00:38
	in fact make a new branch from HEAD^	00:39
	← MaximusColourum left	00:39
nevyn	then rebase that.	00:39
	then merge your current branch into it.	00:39
xil	"git branch -b newbranch HEAD^"?	00:40
	and then checkout newbranch and do the rebase?	00:40
	← adamm left	00:40
	→ LeMike joined	00:40
	→ heptadecagram joined	00:40
nevyn	yep	00:40
	← heptadecagram left	00:41
nevyn	xil: rebase doesn't delete the original commits that are rebased.. it just unlinks them from the dag.. they're still there.	00:42
xil	and then how do I incorporate the final commit, "checkout currentbranch; merge newbranch"?	00:42
nevyn	checkout newbranch merge newbranch	00:42
	← jagguli left	00:42
nevyn	checkout newbranch merge currentbranch	00:42
	then reset currentbranch to newbranch	00:42
	← dv310p3r left	00:42
	← giallu left	00:42
	→ jehoshua02 joined	00:42
xil	why not just merge back into currentbranch and delete newbranch?	00:43
	← LeMike left	00:43
jehoshua02	CareBear\: ping	00:43
nevyn	xil: because that does something totally different	00:43
xil	gotcha	00:43
	well thank you	00:43
	do I need to do a hard reset?	00:43
nevyn	xil: merging newbranch into currentbranch will take the diff from currentbranch to newbranch and apply it.	00:44
	which creates a new commit that contains the non-conflicted stuff from the rebase on currentbranch	00:44
	this is almost certainly not what you want.	00:44
	→ iocor joined	00:45
xil	that sounds correct	00:45
nevyn	that is what you want?	00:45
	given that it means you're going to have to resolve all the conflicts from the rebase AGAIN?	00:46
	← berserkr left	00:46
	← crisp left	00:47
	→ k776 joined	00:47
k776	Hello. How can I update git-svn to use SVN 1.7.2?	00:47
Bearcat	EugeneKay: Thanks a lot for your help today.	00:48
EugeneKay	Sure	00:48
	→ berserkr joined	00:48
	← jfisk1987 left	00:49
	← mabrand left	00:50
xil	nevyn: I haven't resolved the conflicts from the rebase. I don't want to resolve all of them one at a time, I'd rather resolve only the ones caused by the last commit	00:50
Bearcat	EugeneKay: i'm a little confused on a issue regarding pulling. Mind if i ask you about it?	00:53
	← k776 left	00:53
	← amgarchIn9 left	00:53
EugeneKay	!ask	00:53
gitinfo	yes, it's okay to ask questions here... in fact, you just asked one! ;) Pretty much any question is fine. We're not terribly picky.	00:53
	→ adamm joined	00:53
cmn	generally, if in doubt, don't pull	00:54
SethRobertson	And absolutely if there is any smidgen of a doubt or concern, do not push	00:55
	← ahstanford left	00:55
	→ Jedateach joined	00:56
	→ DofLeFou joined	00:56
	← romeovs left	00:56
	→ notola joined	00:56
	← notola left	00:56
Bearcat	EugeneKay: ok, so i have this gitolite repo all set up an awesome now. I have a .vim dir set up so i can pull it to other PCs in my network. Suppose i go to a clients house and i want to pull my .vim repo so i can work in vim in comfort. I make a backup of thier dir so i can restore it when i'm, done and then i ..what? Install git on their machine and then i'd only need a copy of my public key on a usb	00:57
	stick?	00:57
	← srcerer left	00:57
Bearcat	cmn, SethRobertson: i'll avoid the dirty jokes i'm sure you've heard before after statements like that...	00:57
EugeneKay	Home dir management is a tricky one, which I'm not really qualified to speak on	00:57
	→ paphio joined	00:57
bremner	!vcsh	00:58
gitinfo	https://github.com/RichiH/vcsh - a tool to manage dotfiles using git's fake bare repos to put more than one working directory into $HOME. Nothing to do with csh AFAIK.	00:58
SethRobertson	Bearcat: You may find that your .vim files are public enough to be given anonymous read/only access. I would never load a private key on someone else's computer	00:58
Bearcat	SethRobertson: you mean public? I don't need my private key for this.	00:59
cmn	the safest options is probably have your post-update script generate a tarball	00:59
SethRobertson	Bearcat: What do you need a public key for? The private key is what gives you access.	01:00
cmn	or zip or whatever, so you don't have to rely on git being on any machine	01:00
Bearcat	SethRobertson: i guess i'm confused. It was the public key that got copied over to the gitolite server, yes?	01:00
	← beatak left	01:00
tommcdo	Bearcat: yes, and the private key is what's required locally to match against the public key on the remote site	01:01
MacGyverNL	(On that note, if you're going to carry a USB stick you can also carry git on it)	01:01
SethRobertson	Yes, but your private key is what you use to prove you are you to gitolite	01:01
	← neurodrone left	01:01
tommcdo	Bearcat: why not just put your .vim directory on the USB stick? :P	01:01
	(jk, avoiding the problem)	01:01
Bearcat	tommcdo: well, yeah that's works, but it's not nearly as cool :"P	01:01
MacGyverNL	Bearcat: Basically, by using the private key you prove, to the gitolite server, that you possess the private key for that matching public key. The problem with this is that in order to prove this, the private key needs to be handled by the (potentially untrusted) client machine, in which case it can leak even though it's on the USB stick.	01:02
Bearcat	i don't care if it takes me 10 hours just to figure out how to use git to copy a directory, i'm gonna look cool! That's what counts! :P	01:02
MacGyverNL	Which is why we have things like smart cards and PKCS#11.	01:03
Bearcat	MacGyverNL: that makes sense	01:03
	gitinfo set mode: +v	01:04
MacGyverNL	I'm with SethRobertson on this one - your .vim files are probably public enough to be given anonymous read-only access, thus circumventing the problem of your private keys leaking.	01:04
Bearcat	oh..duh. Of couse. I'd need my private key on otherPersonsComputer/~/.ssh/	01:04
Jedateach	hello	01:04
EugeneKay	Use passphrase-protected priv keys	01:04
MacGyverNL	Well, not necessarily - you can load it from the stick - but it would still be accessible.	01:04
EugeneKay	And ssh-agent	01:04
Bearcat	yeah all my keys have passwords on them.	01:05
MacGyverNL	The point is that the moment you stick that USB-drive into an untrusted computer and key in your passphrase, basically, all bets are off.	01:05
Bearcat	MacGyverNL: nod nods	01:05
MacGyverNL	Now, I'm an extremely paranoid guy.	01:05
Jedateach	Can anyone help me resolve a issue setting up gitolite? - all working fine between workstation and server, but I want to clone repositories to different user accounts on the server (running cPanel by the way). When I try to do that, the ssh forced command doesn't appear to kick in, and I get asked for git@localhost's password.	01:05
	← Xizor left	01:06
MacGyverNL	So I tend to exaggerate and this will most likely not affect you in the slightest, but still, as long as you understand the issue I think you can make your own decisions.	01:06
	→ aldodelgado joined	01:06
	→ Heimidal joined	01:06
xil	nevyn: now my history looks really weird and convoluted, but I think that's because of a previous problem that got carried over weirdly. But it's okay because the state of things now is clean and stable. Thanks =]	01:07
Bearcat	thanks for the help folks	01:08
MacGyverNL	You're welcome, hope it helps.	01:09
jehoshua02	What is the common reason I'm asked for "password" instead of "passphrase" when writing to a repo?	01:10
cmn	you're probably using HTTP or ssh w/o a key	01:10
	← blorbx left	01:10
offby1	curse cmn and his fast fingers	01:10
cmn	:)	01:10
	← kenichi left	01:10
	← jehoshua02 left	01:11
	← shergill left	01:11
	→ dshap joined	01:12
	← ehsan left	01:12
	→ forsto joined	01:12
	← paphio left	01:12
	← jason237 left	01:12
dshap	hey guys - i just merged a "dev" branch into my master branch and in gitk i can see that they both indeed point to the same commit as expected - but i did git push to send everything to my remote (github) and it's not reflected there (i.e. on the github network it doesn't show master pointing to the same commit as dev)	01:13
	anyone have any ideas what could cause that behavior to occur?	01:13
	← vervic left	01:14
philllllijw	did you rebase at some point?	01:14
kevlarman	dshap: github is being slow atm	01:14
dshap	no rebasing	01:14
	kevlarman: o rly?	01:14
	← Chappo left	01:15
dshap	kevlarman: the weird thing - my coworker just did a git fetch on his master branch and then it told him his master was like 100+ commits behind origin/master, so it seems like the merge made it there but it's just not reflected in the github UI	01:15
	kevlarman: i guess it's just slow, as you suggest	01:15
	thanks	01:15
	→ shergill joined	01:15
kevlarman	dshap: now that sounds like a rebase	01:16
	→ aidalgol joined	01:16
kevlarman	unless it was that huge a merge i guess	01:16
dshap	kevlarman: i don't even know how to rebase	01:16
	← Jedateach left	01:16
dshap	kevlarman: it was a huge merge - we basically abandoned the master branch to build out v0.1 of our project	01:16
	kevlarman: now that it's finally out, wanted to bring master up to speed	01:16
	→ rchavik joined	01:16
	← rchavik left	01:16
	→ rchavik joined	01:16
	← Cam left	01:17
	→ Bass2 joined	01:18
	← AaronMT left	01:18
	← Bass2 left	01:19
	→ Bass2 joined	01:19
	← Bass10 left	01:21
	← kenperkins left	01:21
	→ Bass10 joined	01:22
	→ kerrick joined	01:22
	← kerrick left	01:22
	← berserkr left	01:23
	Hunner_ → Hunner	01:23
	→ Jedateach joined	01:23
	← johnhamelink left	01:24
	← Bass2 left	01:24
	ChanServ set mode: +v	01:24
Jedateach	hello - having gitolite trouble with ssh daemon asking for password when accessing on same server. Working fine between workstation and server.	01:24
	any way that the forced commands in authorised_keys could be bypassed?	01:24
	server is running cPanel, so I'm assuming that may be interfering in some way	01:25
EugeneKay	cPanel does all sorts of stupid things. Not beyond possibility that it's competely screwed up SSHd	01:25
Jedateach	:(	01:26
EugeneKay	`ssh gitolite@localhost info`, ctrl^c the password prompt, and paste the entry that is generated in /var/log/secure	01:26
	ssh -vv, that is	01:26
	← ScottE left	01:26
	← Bass10 left	01:26
	→ ChaoWang joined	01:26
	→ Bass10 joined	01:26
	← happyface left	01:27
cmn	do have the local account set up to use a key gitolite knows?	01:27
EugeneKay	Jedateach - pastebin, not PM ;-)	01:27
	← ericmuyser left	01:28
	→ eletuchy joined	01:29
Jedateach	ok, sorry	01:29
	1 sec	01:29
	http://pastiebin.com/?page=p&id=4f307e6f30321	01:29
EugeneKay	Last dozen lines tell the issue; none of your private keys had a matching pubkey on file with gitolite	01:30
	Add your localhost pubkey to gitolite-admin.git and retry ;-)	01:30
Jedateach	ok, thanks. thought I had that right, but perhaps not. I'll give it a go!	01:30
	← yuriks left	01:30
	→ yuriks joined	01:30
	← sythe left	01:30
	← erichynds left	01:31
	→ chrisdev joined	01:31
EugeneKay	I recommend not using id_dsas at all - sticking with one type of key makes life a lot easier	01:31
	And everything modern will accept a 4096-bit RSA key, SSL and SSH both	01:31
	EugeneKay disappears to dinner	01:31
Jedateach	ah, is dsa the default?	01:31
	→ sythe joined	01:31
EugeneKay	"yes"	01:31
Jedateach	I might have generated incorrecly	01:31
	y	01:32
kevlarman	EugeneKay: 4k already?	01:32
	→ porco joined	01:33
	← anjen left	01:33
	→ ehsan joined	01:33
	← iocor left	01:35
	← dshap left	01:36
	← akiress left	01:36
	→ lyddonb joined	01:37
	← DarkAR left	01:37
	← lacrymology left	01:37
	← Yuuhi left	01:38
	→ ericmuyser joined	01:38
	→ neurodrone joined	01:39
	→ jason237 joined	01:39
	← xil left	01:39
	← legnaleurc left	01:42
Jedateach	EugeneKay - that worked. Thanks for your time :)	01:43
SethRobertson	!thanks	01:44
gitinfo	Feeling thankful? Type "ExampleUser++", and ExampleUser will score karma points at http://carmivore.com (our preferred way to objectify self-worth). There's really no point to thanking me – instead, why not thank the person who made me help you?	01:44
Jedateach	EugeneKay++	01:44
	→ legnaleurc joined	01:44
	→ dv310p3r joined	01:44
	→ apok_ joined	01:44
	← apok left	01:46
	← Chillance left	01:47
	← Lummoxx left	01:47
	apok_ → apok	01:47
	← dreiss left	01:47
	→ ehsan_ joined	01:48
	← chrisdev left	01:48
	← nobitanobi left	01:48
	→ regedarek joined	01:49
	→ chrisdev joined	01:49
regedarek	hi, is possible to add change to last commit after i pushed to github this commit ??	01:50
	← ehsan left	01:50
SethRobertson	Possible? Yes. recommended? no	01:50
	!rewriting_public_history	01:50
gitinfo	Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)	01:50
	→ cooloney joined	01:51
	← tommcdo left	01:51
SethRobertson	regedarek: See also !fixup	01:51
gitinfo	regedarek: So you lost or broke something or need to otherwise find, fix, or delete commits? Look at https://gist.github.com/1612395 for full instructions. For hints type !fixup_hints in IRC. Remember: if you have pushed already, there are only a few things you can do without !rewriting_public_history (type that for more info)	01:51
	→ jasonbray joined	01:52
	← CannedCorn left	01:53
	→ nicxvan joined	01:54
	→ Chowzzf joined	01:54
aidalgol	How do I retrieve a file deleted in a previous commit?	01:54
	← jasonbray left	01:54
	→ zdk joined	01:54
RandalSchwartz	git checkout HEAD~:path/to/file	01:54
	← whyking left	01:55
RandalSchwartz	maybe HEAD~2	01:55
	→ jasonbray joined	01:55
SethRobertson	PREVIOUSCOMMITSHA^ anyway	01:55
	← jasonbray left	01:55
	→ jasonbray joined	01:56
aidalgol	Which option to git log lists affected files?	01:58
	← an0maly left	01:59
	← xiaotian left	01:59
	→ ehsan joined	01:59
	→ BSaboia joined	01:59
	→ BSaboia__ joined	01:59
	← BSaboia left	01:59
	← BSaboia__ left	01:59
	← jasonbray left	02:00
	→ herdingcat joined	02:01
	← ehsan_ left	02:01
	← Sigma left	02:02
	→ rurufufuss joined	02:02
	← xiangfu left	02:03
	→ gavin_huang joined	02:03
	→ jasonbray joined	02:03
	→ ehsan_ joined	02:03
	→ paphio joined	02:04
	→ dtgb joined	02:05
	← alberto56 left	02:05
cmn	several, depending on how you want your output	02:06
	--name-status --name-only --stat all give you that information	02:06
	→ srcerer joined	02:06
	← ehsan left	02:06
aidalgol	SethRobertson, RandalSchwartz: I get fatal: reference is not a tree:	02:07
	← Bearcat left	02:08
SethRobertson	aidalgol: what does this return? `git log -n 1 --oneline -- filename`	02:08
	→ BSaboia joined	02:08
	→ Vampire0_ joined	02:08
aidalgol	SethRobertson: The first seven digits of the commit hash in which the file was deleted and the commit message.	02:09
SethRobertson	Try `git checkout SHA^ -- filename`	02:09
	→ alberto56 joined	02:10
aidalgol	OK, that worked.	02:10
	Thanks!	02:10
	← jason237 left	02:12
	→ johnkpaul joined	02:13
	← Vampire0 left	02:13
	← macmartine left	02:13
	→ zomothesweeper joined	02:14
g0th	hi	02:14
	I just made a git add and git commit that I dont like	02:14
	how do I "undo" it?	02:14
SethRobertson	!fixup	02:14
gitinfo	So you lost or broke something or need to otherwise find, fix, or delete commits? Look at https://gist.github.com/1612395 for full instructions. For hints type !fixup_hints in IRC. Remember: if you have pushed already, there are only a few things you can do without !rewriting_public_history (type that for more info)	02:14
g0th	I did a backup of all files	02:14
	can I just do git checkout .?	02:14
SethRobertson	It will not undo the commit	02:15
g0th	so what is the "reverse" push command?	02:15
SethRobertson	fetch?	02:15
g0th	git fetch origin?	02:16
	that should do it?	02:16
SethRobertson	or just `git fetch` I'm not sure what this has to do with your first question. It isn't going to help	02:16
	← ehsan_ left	02:16
g0th	it will not undo the commit?	02:16
SethRobertson	Not at all	02:17
g0th	I could just delete my repository and git clone it?	02:17
	→ ehsan joined	02:17
SethRobertson	You could, or you could follow the link which includes specific instructions on several solutions to your problem	02:17
g0th	too complicated	02:18
	I tried	02:18
	it is 3:18am here	02:18
	I just try to fix my git stuff before going to befd	02:18
	I dont care about any local changes	02:19
bremner	I recommend going to bed	02:19
g0th	I already made backups of all (outside of the tree)	02:19
	I will do the git clone thing then	02:19
SethRobertson	Yes. If you are too tired to answer five questions from that link, you are to tired to enter the command correctly	02:19
g0th	I thought there was a simple command	02:19
SethRobertson	There is	02:19
g0th	can't you just tell me?	02:20
	← Lgb left	02:20
g0th	I don't get this "git clean" thing	02:20
SethRobertson	Mostly because it has nothing to do with undoing a commit	02:20
g0th	I just want to basically do "git push origin master" the other way round	02:20
	→ an0maly joined	02:20
g0th	but the whole page is about it	02:20
	← alberto56 left	02:21
	→ xiangfu joined	02:21
SethRobertson	You absolutely do not want to use git push. That is the worst command to use in this situation	02:21
	→ D-Chymera1 joined	02:21
g0th	yes I know :)	02:22
	but thanks for the hint considering I'm tired	02:22
	git reset?	02:22
SethRobertson	That is one of your options, with the right arguments.	02:22
g0th	ouff	02:23
	I don't understand the guide	02:23
	→ Lgb joined	02:23
g0th	I will read it again	02:23
	I am sure you understand what I want	02:23
	why do you let me suffer? ^^ ;)	02:23
SethRobertson	To improve the guide. If you don't understand it, I want to know why	02:23
	→ jason237 joined	02:24
SethRobertson	You can reveal each answer you give and we can try to find out where you are going wrong	02:24
	← D-Chymera left	02:25
	→ gnufied joined	02:25
g0th	Do you have a clean working directory?	02:25
	this is not clear at all	02:25
	→ gitfanboy joined	02:25
g0th	what does "clean working directory" mean	02:25
	← Jedateach left	02:25
SethRobertson	What does `git status` say?	02:25
	I will try to improve this	02:25
g0th	M board_player.js	02:26
	M misc.js	02:26
	M player/js/player_new.js	02:26
	and some ?	02:26
	basically I want to get rid of the "M"'s :)	02:26
	and not by pushing ;)	02:26
	← tizzo-afk left	02:26
SethRobertson	OK, then click that you have a dirty working directory	02:27
	→ spersaud joined	02:27
	→ alberto56 joined	02:27
g0th	but I don't care if everything else is deleted	02:27
	I mean the "?" stuff	02:27
	but it always is dirty then?	02:27
spersaud	when you add a directory .. are subdirectories updated ?	02:27
g0th	last question was if I had commited	02:27
	if I have commited it will always have "M"?	02:27
spersaud	how would you make sure all files and subdirs are posted to master ?	02:27
g0th	so I dont get this question	02:27
SethRobertson	g0th: You may have committed, but you also have uncommitted stuff.	02:28
g0th	oh the M is the uncommited stuff?	02:28
	I am tired ouff	02:28
SethRobertson	yes	02:28
g0th	ok now I get it	02:28
	the question should then read: do you have uncommited stuff?	02:28
	← stuntmachine left	02:29
g0th	maybe with some addition like "that you want to keep"	02:29
	← ajt left	02:29
	→ tizzo-afk joined	02:29
g0th	ok so it tells me I have to either commit them or "undo" them	02:30
	← mattly left	02:30
g0th	I can't leave any uncommited changes around	02:30
SethRobertson	OK, then "I want to discard it"	02:30
g0th	don't get why not but ok	02:30
	can't I just do git checkout .?	02:31
	← sythe left	02:31
	→ sythe joined	02:31
sythe	HEY	02:31
SethRobertson	That would work	02:31
sythe	Ok	02:31
g0th	but it isn't mentioned?	02:31
	which confuses me because I know git checkout but not those other commands	02:31
sythe	How can I replace a bare repo with a remotely cloned one?	02:31
SethRobertson	g0th: other options are mentioned if you click on the link saying you want to discard it	02:31
	← acl_ left	02:31
SethRobertson	sythe: mv repo othername; git clone URL	02:32
	← gitfanboy left	02:32
sythe	Uh	02:32
	→ joshsmith joined	02:32
sythe	Simplify those directions?	02:32
	→ weiyang joined	02:32
SethRobertson	Step one: move or delete the existing bare repo. Step two: clone the remote repo	02:33
g0th	if git diff returns empty	02:33
	is that equivalent to a clean directory?	02:33
	or do I also need to remove the "?" files?	02:33
	→ kenperkins joined	02:34
SethRobertson	g0th: No, git status should say nothing is changed or only new files	02:34
	→ ajt joined	02:34
	→ ehsan_ joined	02:34
	← shergill left	02:34
g0th	hmm but status is the same	02:34
	after git checkout ...	02:34
SethRobertson	So perhaps you should actually follow the documented instructions	02:34
g0th	you said checkout should work	02:34
	I thought it was about uncommited changes	02:35
	there are no uncommited changes	02:35
sythe	SethRobertson: Uh...give them to me in linux terminal commands?	02:35
SethRobertson	sythe: What is the path to your bare git repo. What is the URL you wish to clone?	02:35
spersaud	I was using git add .. and I assumed all files in a sub dir would be added to repo	02:36
sythe	hm	02:36
spersaud	but its not the case ...	02:36
sythe	What permissions would I need to see the directories in a git repo in a web browser?	02:36
	755?	02:36
	← ehsan left	02:36
spersaud	how do I make sure all files in a subdir are recursively added ?	02:36
sythe	Or 777?	02:36
	→ acl_ joined	02:36
sythe	Oooh	02:37
	How can I find the location of a git repo?	02:37
SethRobertson	sythe: 755, but probably you need gitweb.cgi	02:37
	spersaud: It is the case, unless the directory or files are gitignored	02:37
	sythe: Someone has to tell it to you.	02:38
	→ Jedateach joined	02:38
sythe	Like...how can I figure out if http://example.com/git is the same as /home/sythe/public/git	02:38
g0th	git reset --hard @{u}	02:38
	→ notola joined	02:38
g0th	would this command have worked from the start?	02:38
SethRobertson	sythe: There is no general way to discover this, other than perhaps making a change in one place and seeing if it magically appears in both places.	02:38
g0th	was that all I had to do?????	02:38
	→ f0i joined	02:39
SethRobertson	g0th: quite so	02:39
sythe	SethRobertson: Well...IDK which is which	02:39
g0th	man	02:40
SethRobertson	Certainly there is no way for me to find out	02:40
	← apok left	02:40
g0th	why not just write it at the beginning?	02:40
sythe	How would I find out?	02:40
	→ apok joined	02:40
SethRobertson	ask the person who set it up	02:40
	→ cosgroveb joined	02:40
	← xiangfu left	02:40
	tizzo-afk → tizzo	02:40
sythe	-.-	02:40
	← weiyang left	02:41
	← flavius left	02:42
	← lyddonb left	02:42
g0th	thx anyway	02:43
	← notola left	02:43
	← preinheimer left	02:44
	→ preinheimer joined	02:45
	← Jedateach left	02:45
	← aidalgol left	02:45
	→ spaced0g joined	02:45
cosgroveb	i have a sort of weird situation... i got into a case where my local branch local topic_branch has commits a, b, c, and d... and my remote topic_branch has a and c... tried pushing and got "! [rejected] HEAD -> topic_branch (non-fast-forward)"... it tells me to do a git pull.. but of course that doesn't work :(	02:45
	should i push force?	02:45
	← apok left	02:46
cosgroveb	if the problem even makes sense the way i have stated it	02:46
	→ Cam joined	02:46
philllllijw	how did you manage to do that	02:46
	← dtgb left	02:47
SethRobertson	cosgroveb: This is !rewriting_public_history You could use `git rebase @{u}` to make it acbd and then push that	02:47
gitinfo	cosgroveb: Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)	02:47
	← abetaha left	02:48
	→ macmartine joined	02:48
	→ tga joined	02:48
tga	greetings	02:48
	← brian_g left	02:48
	← spersaud left	02:49
tga	I could use a bit of help figuring out a proper workflow	02:49
	→ lyddonb joined	02:50
RandalSchwartz	tga - "cp file.c file.c.good"	02:50
tga	I have a central repo with master and work in progress branches	02:50
	so I pulled the wip branch locally and started commiting changes to it	02:50
RandalSchwartz	do you have to share work on the work-in-progress?	02:50
	are they topic-oriented?	02:50
tga	when the origin/wip branch gets modified by someone else, how do I update my local repo and files?	02:51
SethRobertson	tga: I suggest reading !best_practices	02:51
gitinfo	tga: There is no one right answer for git best practices, but a consensus from #git is available at https://gist.github.com/1540906	02:51
tga	it's my own repo, trying to set things up so I can work with someone else on it	02:51
RandalSchwartz	yeah, best practices	02:51
tga	alright, that's probably what I need	02:51
SethRobertson	Also read !book	02:51
gitinfo	There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !vcbe and !parable	02:51
RandalSchwartz	oh - by the way, my slides are up now	02:51
	the ones that go with the talks	02:51
SethRobertson	The ones I already reviewed for you around xmas or a different set?	02:52
RandalSchwartz	... http://www.slideshare.net/RandalSchwartz/introduction-to-git-11451326	02:52
	more or less the same	02:52
	the video is also up though	02:52
	tizzo → tizzo-afk	02:52
RandalSchwartz	... http://vimeo.com/35778382	02:52
philllllijw	RandalSchwartz++	02:53
	SethRobertson++	02:53
sythe	Hey...I was just told to use SVN because it's "Easier"	02:53
cosgroveb	SethRobertson should be the only person working on this branch but agree that rewriting public history == bad idea even though the remote server allows... in your suggestion what is @{u}? :)	02:53
sythe	Is that true?	02:53
	XD	02:53
RandalSchwartz	troll!	02:53
philllllijw	sythe, if you just stay away from rebase and any other fancy commands, I don't see why it should be difficult	02:53
	→ xiangfu joined	02:54
	→ jrnieder joined	02:54
bremner	sythe: you are welcome to discuss svn in #svn	02:54
sythe	I'm trying to hook it up to Redmine, lol	02:54
	Massive headaches.	02:54
RandalSchwartz	git is a more powerful tool	02:54
philllllijw	our team has had many issues with git+continuous intengration tools though	02:54
sythe	I love Git, actually...IMO, way better than CVS and SVN	02:54
RandalSchwartz	if you often cut yourself with powerful tools... I can see their point	02:54
	← maletor left	02:55
sythe	Between it and Mercurial, it's a lot tougher to pick a clear-cut winner	02:55
philllllijw	just dont run your git server on windows	02:55
sythe	Except..then you see Github.	02:55
RandalSchwartz	except for speed.	02:55
sythe	And Git wins.	02:55
RandalSchwartz	and for avoiding the idea of "revision #456"	02:55
	which can never end well	02:55
	← tango_ left	02:55
RandalSchwartz	wait - is it bzr that has artificial version numbers, or hg?	02:56
	one of them has it, and it's bad	02:56
	← Nibato left	02:56
SethRobertson	!isgitbetter?	02:56
gitinfo	we think so, but if you don't, please go away and use $YOURFAVOURITESYSTEM. See http://whygitisbetterthanx.com/ among many many other places.	02:56
	← johnkpaul left	02:56
	→ shiba_yu36 joined	02:56
	← kenperkins left	02:57
	tizzo-afk → tizzo	02:59
	→ Jedateach joined	02:59
	← jwmann left	03:00
	← monstaRtruck left	03:00
	dhruvasagar_ → dhruvasagar	03:01
	ChanServ set mode: +v	03:02
Jedateach	I run a server that has a unique user for every website I build (cpanel). I need to pull and push to a local git repository, and github. The dilemma I'm facing is how to do this without creating public/private keys for every user.	03:02
	← lyddonb left	03:03
	→ Nibato joined	03:03
	→ lyddonb joined	03:03
RandalSchwartz	Jedateach - not clear... do they need to push/pull too?	03:03
	if it's all you, why do you need more keys?	03:03
Jedateach	I'm thinking I could add github and my local server to ssh_config, so that every user has access… but I would prefer that some users don't have access to the keys.	03:04
SethRobertson	Jedateach: Yes, it sonds like you can just upload the public key to each unique user and then git to user@host.	03:04
	→ xil joined	03:04
SethRobertson	Jedateach: So the website makes git commits?	03:04
	→ themachine joined	03:04
	← themachine left	03:04
Jedateach	not necessarily	03:04
xil	hi. So when there are merge conflicts I can use "git mergetool" to fix them. Is there a way to do that just for files in the working tree? Or even in the index?	03:04
SethRobertson	Jedateach: Please explain your workflow	03:05
RandalSchwartz	you can't have a merge conflict except in the working tree.	03:05
SethRobertson	xil: You can use `git status` to see which files are bad and then edit them in your editor	03:05
	← jason237 left	03:05
xil	SethRobertson: but I want to edit them in a diff style	03:06
philllllijw	I think I may start using git bash. I get made fun of by linux people because I use gui	03:06
	← cYmen left	03:06
Jedateach	SethRobertson: develop the website locally which includes submodules from git, push to private repository, create new cPanel account (user), clone into that account from private repo	03:06
jrnieder	xil: you might be able to cobble something together using "git difftool"	03:06
SethRobertson	xil: You typically search for the conflict marker and then you have a diff to edit	03:07
xil	well I'm using meld as my mergetool, and would like that for this as well if possible. I could do it manually but I'm wondering if there is a shortcut	03:07
	→ jwmann joined	03:07
cosgroveb	if i am on a branch, git push force will only affect the branch, correct? (no lectures on rewriting history please - personal project :( )	03:07
	← Spockz left	03:07
SethRobertson	Jedateach: Yes, so each cpanel user gets your pubilc key loaded in .ssh/authorized_keys, you then `git init` in some directory for that webserver, and then you push from your local system to the cpanel server	03:07
xil	jrnieder: will difftool allow me to actually edit the files?	03:07
nevyn	cosgroveb: it depends what push.default is set to.	03:08
jrnieder	xil: not sure, never tried it	03:08
	← cccaldas left	03:08
RandalSchwartz	sounds like using clone as a deploy tool	03:08
	that's gonna be a problem	03:08
xil	oh lol. git difftool seems to give me the option to activate my mergetool	03:08
	← Cam left	03:08
SethRobertson	cosgroveb: see push.default in man git-config	03:08
gitinfo	cosgroveb: the git-config manpage is available at http://jk.gs/git-config.html	03:08
	→ paulreinheimer joined	03:08
xil	that's awesome	03:08
RandalSchwartz	git is NOT a deploy tool	03:08
Jedateach	SethRobertson: it just seems a hassle to create keys for every user. I'm coming from SVN, where all I needed was a password.	03:09
RandalSchwartz	Jedateach - see above	03:09
SethRobertson	Jedateach: Yes, that is why I gave you a workflow to NOT create keys for each user	03:09
RandalSchwartz	don't deploy with clone. use some sort of install tool that can give proper ownership	03:09
	→ cYmen joined	03:09
RandalSchwartz	people really like abusing git, I guess. :)	03:10
	→ weiyang joined	03:10
Jedateach	RandalShhwartz: I'm just not educated in the ways of deploy tools.	03:10
	what do you suggest?	03:10
SethRobertson	RandalSchwartz: "proper ownership" is the wrong argument. I would say, proper permissions, or atomic updates, or synchronized database updates, or...	03:10
	→ amigojapan joined	03:11
	tizzo → tizzo-afk	03:11
	→ SegFaultAX joined	03:11
cosgroveb	nevym SethRobertson thanks... set to "current" so i should be OK	03:11
SethRobertson	Jedateach: A lot of people talk about Capistrano	03:11
	!thanks	03:11
gitinfo	Feeling thankful? Type "ExampleUser++", and ExampleUser will score karma points at http://carmivore.com (our preferred way to objectify self-worth). There's really no point to thanking me – instead, why not thank the person who made me help you?	03:11
	← jwmann left	03:11
nevyn	abusing clone can be convienient. particularly if your app is git aware in some way (eg the app can do things with branches)	03:12
	← preinheimer left	03:12
	← csherin left	03:12
	← cooloney left	03:13
	→ nobitanobi joined	03:13
Jedateach	SethRobertson: oh, so were you saying I should modify cpanel to include a common public key for every new user set up?	03:14
SethRobertson	Jedateach: Yes, that is one of your options.	03:14
Jedateach	ok, I think that discussion helped me a bit. I'll research capistrano / deployment tools etc.	03:14
	SethRobertson++	03:15
	← alberto56 left	03:15
Jedateach	RandalSchwartz++	03:15
	← knittl left	03:15
	→ cooloney joined	03:15
	← paphio left	03:16
	RandalSchwartz bows	03:16
	RandalSchwartz hits head on desk	03:16
RandalSchwartz	ouch!	03:16
	← Deck` left	03:16
	← Destos left	03:16
	← Marquel left	03:17
	→ Kichael joined	03:17
	← vereteran left	03:17
	→ themachine joined	03:18
	← wereHamster left	03:18
	→ LongBeach joined	03:19
	← hyperair left	03:20
	→ mandric joined	03:20
	→ hyperair joined	03:22
	→ Prime326 joined	03:22
	→ intripoon_ joined	03:24
	← nobitanobi left	03:26
	→ nobitanobi1 joined	03:26
	← xil left	03:26
	← intripoon left	03:28
	← nobitanobi1 left	03:29
	← legnaleurc left	03:31
	→ Chibby_ joined	03:31
	→ aidalgol joined	03:32
	← jasonbray left	03:33
	→ legnaleurc joined	03:33
	← plusk left	03:33
	← themachine left	03:33
	→ freezey joined	03:34
	← chrisdev left	03:36
	→ knittl joined	03:36
	← knittl left	03:36
	→ knittl joined	03:36
	→ wereHamster joined	03:37
	→ Deck` joined	03:37
	→ Marquel joined	03:37
	→ vereteran joined	03:37
	← legnaleurc left	03:37
	← vereteran left	03:38
	→ abetaha joined	03:39
	→ igotnolegs joined	03:40
	→ legnaleurc joined	03:41
	→ mattly joined	03:41
	← mattly left	03:41
	→ mattly joined	03:41
	→ vereteran joined	03:42
	← ketas left	03:42
	← wuj left	03:43
	← tiago left	03:43
philllllijw	RandalSchwartz, I'm only 15min into your intro video but I hope it will go into a little more detail about how the trees, etc work. Preferably with pictures	03:45
	you keep mentioning "sha1 of tree and parent commit" but I dont know what you mean by parent commit	03:47
offby1	when two commits love each other very much ...	03:47
philllllijw	aww	03:48
	← igotnolegs left	03:48
SethRobertson	philllllijw: See !doc	03:48
gitinfo	philllllijw: A list of useful documentation of all kinds is here: http://git-scm.com/documentation -- or try the wiki at http://git.wiki.kernel.org/. Also try typing "!book" "!cs" "!bottomup" "!parable" "!best_practices" or "!vcbe" or "!designers" here in IRC. !book is probably the most helpful.	03:48
nevyn	so... in view of my want to abuse clone to do releases does anyone know of any tooling for this? (something like jenkins that'	03:48
	s git aware)	03:48
SethRobertson	philllllijw: cs bottomup and parable might be where you want.	03:48
	And of course book	03:49
philllllijw	!bottomup	03:49
gitinfo	'Git from the bottom up' starts with explaining the building blocks of git and proceeds to tell you how they fit together. http://ftp.newartisans.com/pub/git.from.bottom.up.pdf	03:49
philllllijw	cool	03:49
SethRobertson	Should we add !IntroToGit to the list?	03:49
gitinfo	Randal Schwartz has an introduction to git video http://vimeo.com/35778382 and slides http://www.slideshare.net/RandalSchwartz/introduction-to-git-11451326	03:49
	→ L42y joined	03:49
	← Dreamer3 left	03:49
RandalSchwartz	yeah	03:49
	→ Dreamer3 joined	03:49
RandalSchwartz	it should go on the wiki too	03:49
	haven't had time to add that	03:49
SethRobertson	Err, is the wiki back?	03:50
RandalSchwartz	oh - I thought it was	03:50
	← orafu left	03:50
RandalSchwartz	I could be wrong :)	03:50
	→ orafu joined	03:50
	→ Transformer joined	03:50
	← hyperair left	03:51
	→ Milossh joined	03:51
	→ hyperair joined	03:51
	→ ketas joined	03:52
	→ jayeffess joined	03:52
	← ajt left	03:52
	← scelis left	03:53
	← Transformer left	03:53
	← mattly left	03:54
	← BSaboia left	03:55
	← LongBeach left	03:55
	→ LongBeach joined	03:55
	→ nixmaniack joined	03:55
	← LongBeach left	03:56
	← OOPMan left	03:56
	→ levifig joined	03:56
	→ paphio joined	03:56
	← mcluskydodallas left	03:56
	→ mcluskydodallas joined	03:57
	→ amoffat joined	03:58
	← aidalgol left	03:58
amoffat	hi, sometimes checking out a branch doesn't say "your branch is behind <remote tracking branch> by X commits" when it IS behind by X commits. anyone have any ideas on why that is?	03:59
	→ dfr\|mac joined	03:59
amoffat	sometimes it does say it though, which is really weird	03:59
	← macmartine left	03:59
SethRobertson	talk to us when it happens & after a fetch	04:00
amoffat	SethRobertson, it's happening right now, and I just fetched	04:01
nevyn	amoffat: is the branch configured with a suitible tracking branch?	04:01
amoffat	SethRobertson, there's a diff between my master and origin/master, but my local branch says nothing about how many commits its behind	04:01
SethRobertson	git branch -vv : does you branch have an upstream listed in []? what is it?	04:01
	← Jedateach left	04:02
amoffat	nevyn, SethRobertson. oh..well, no, it doesn't look like it's being tracked. doh	04:02
jrnieder	amoffat: you can fix it up with "git branch --set-upstream master origin/master"	04:02
	or, with a more intuitive UI, by running "vim .git/config"	04:03
amoffat	jrnieder, perfect, now i can see how much behind my local branch is. thanks guys :)	04:03
RandalSchwartz	editing .git/config more intuitive? :)	04:03
jrnieder	than --set-upstream, yes :)	04:03
	I never remember the order of the arguments	04:04
	→ tango_ joined	04:04
	← cosgroveb left	04:04
	← Milossh left	04:05
	→ cooloney_ joined	04:05
	→ JohnBat26 joined	04:06
	← cooloney left	04:06
	→ aidalgol joined	04:07
	← Trezkel left	04:08
	← sythe left	04:08
	→ guns joined	04:09
	→ panthrodeb joined	04:09
	→ dfr\|mac_ joined	04:10
	→ LongBeach joined	04:10
	← LongBeach left	04:10
	→ scientes joined	04:11
	← jayeffess left	04:11
scientes	when i have a tracking branch, how do i pull it in, and then rebase all my commits at master/HEAD ?	04:11
philllllijw	won't a rebase "pull it in"?	04:12
	← zomothesweeper left	04:12
nevyn	scientes: so where does your work live?	04:12
scientes	in master, but this branch has a change for deployeement reasons	04:13
	← dfr\|mac left	04:13
scientes	aha, yes git rebase worked	04:13
	i initially did merge, which was clearly wrong	04:14
nevyn	rebase is what you want	04:14
	and you want git fetch or git remote update	04:14
	not git pull	04:14
RandalSchwartz	people "pull" far too often	04:14
jrnieder	can use "git pull --rebase" as a shortcut, I suppose	04:15
philllllijw	whats a pull for vs a fetch?	04:15
jrnieder	pull = fetch + merge	04:15
RandalSchwartz	fetch = bring remotes into your origin/*	04:15
	← lonewulf` left	04:15
RandalSchwartz	pull = that plus merge it into your local branches	04:15
philllllijw	so how do you merge them in?	04:15
	why wouldnt you want to merge it in?	04:15
RandalSchwartz	maybe you just want to view the changes	04:15
jrnieder	when I was starting out, I always did things in that order	04:16
	git fetch	04:16
	git log HEAD..FETCH_HEAD	04:16
scientes	they are just branches of the same repository	04:16
jrnieder	email to complain about things	04:16
	then git merge FETCH_HEAD	04:16
philllllijw	a branch is just 41 text bytes!	04:16
RandalSchwartz	I read that somewhere. :)	04:16
philllllijw	RandalSchwartz, good video	04:16
RandalSchwartz	the new one? yes	04:17
philllllijw	quite long. Might have to finish watching it tomorrow	04:17
RandalSchwartz	yeah - lots of stuff	04:17
	and I talked quite fast	04:17
nevyn	philllllijw: because you don't know what crazy shit upstream have done to break the newfeature your customer really wanted you just wrote.	04:17
RandalSchwartz	imagine if I'd talked normal speed	04:17
	← kukks left	04:17
	→ glennpratt joined	04:17
scientes	yuck	04:18
	that makes pushes non-forward-commits	04:18
jrnieder	rebase does, yes	04:18
scientes	so I have to do it some other way	04:18
	than having a "production" special patch	04:18
jrnieder	oh, yeah	04:18
scientes	if i controled the remote it could do that as a hook	04:18
jrnieder	see the FAQ entry about config-files	04:18
	!configfiles	04:18
gitinfo	the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.	04:19
jrnieder	that's assuming the production magic can be made into data instead of code	04:19
scientes	like production usually will auto reset --hard HEAD	04:19
	→ nobitanobi joined	04:19
scientes	but you could also do a rebase in there	04:19
	← Fandekasp left	04:20
	← amigojapan left	04:20
	→ newbie999\|3 joined	04:20
	← iamjarvo left	04:22
EugeneKay	Deploying with "reset --hard HEAD" is so terrible for so many reasons	04:22
	→ stringoO joined	04:22
	→ macmartine joined	04:22
	← newbie999 left	04:23
RandalSchwartz	I prefer a workflow of "git archive ..." \| (tar xvfz - -C /dest/dir)	04:23
	then it's not about the workdir	04:23
	← lyddonb left	04:24
EugeneKay	Indeed.	04:24
RandalSchwartz	in fact, when I "git push origin master:live", it runs something like that with an rsync	04:24
	← mcluskydodallas left	04:24
RandalSchwartz	with the side effect that "live" always tracks what's deployed live	04:24
	... https://gist.github.com/1714235	04:25
	→ Fandekasp joined	04:25
	← nicxvan left	04:27
wereHamster	RandalSchwartz: git archive \| tar will not delete files. Which is the reason some people prefer checkout/reset	04:27
RandalSchwartz	see my rsync	04:27
	and rsync is also magical in terms of things it doesn't copy and things it doesn't delete.	04:28
	← jimubao left	04:28
RandalSchwartz	rsync --filter='. -' source dest <<EOF	04:28
	is my new go-to copier	04:28
	→ zomothesweeper joined	04:29
	→ beautifulmind joined	04:29
RandalSchwartz	I can be very precise in what to consider in source	04:29
	and what to ignore in dest	04:29
	→ Chappo joined	04:29
	← panthrodeb left	04:30
sitaram	RandalSchwartz: too bad rsycn can't take tar file as input :-)	04:30
	rsync*	04:30
	→ newbie999 joined	04:31
RandalSchwartz	true	04:31
	but as you see, I can map that	04:31
	→ duckxx joined	04:31
RandalSchwartz	what'd be nice is to take a "git fast-export" and map it	04:32
sitaram	yeah but the frugal in me wishes you could avoid the /tmp/XX.$$ in there	04:32
	← newbie999\|3 left	04:32
sitaram	oh no; that sneds history also. Why would you want that?	04:32
RandalSchwartz	not if you just send the final commit	04:32
	→ csherin joined	04:32
RandalSchwartz	git fast-export HEAD	04:32
	no history, right?	04:32
sitaram	hmm yeah I see now	04:33
	I had not realised that	04:33
	← spaced0g left	04:33
RandalSchwartz	so you could build git-archive from that	04:33
	→ queequeg1 joined	04:33
	→ duckx joined	04:34
	→ davidrussell joined	04:34
philllllijw	what kind of rebase workflow do you guys use? Do you just decide to always delete your branches once it is merged so you dont accidentally forget that you pushed it out or merged it?	04:34
EugeneKay	I don't. Rebasing causes cancer, AIDS, and rapes kittens.	04:35
RandalSchwartz	I rebase when I know I've never pushed the commits	04:35
	I like linear history when I can	04:35
philllllijw	I agree but I find myself losing track of what Ive done with each branch so Ive just been avoiding it	04:37
	← duckxx left	04:37
RandalSchwartz	you're probably pushign too often then	04:37
	→ diogogmt joined	04:38
philllllijw	I merge into master when I think I've got a working version of code. But after testing sometimes Im proven wrong	04:38
	← zomothesweeper left	04:39
	← Heimidal left	04:39
RandalSchwartz	might be useful to make a test merge branch then	04:40
	→ Heimidal joined	04:40
philllllijw	yes, that is in the works	04:40
RandalSchwartz	git checkout -b trial HEAD	04:40
	git merge master	04:40
	test - test - test	04:40
philllllijw	we're trying to get our continuous integration system set up but git on windows is terrible	04:40
RandalSchwartz	and if that works, then do it for real	04:40
	otherwise, delete trial, start over	04:40
philllllijw	anyway, i'll finish this video tomorrow. gnight	04:41
	← aldodelgado left	04:42
	← duckx left	04:42
wereHamster	philllllijw: you need to work on less things in parallel.	04:42
	→ shiba_yu_ joined	04:43
philllllijw	less branches at once?	04:43
wereHamster	if you work on one thing, then finish it, merge, then start the other thing you'll never have the problem of knowing which branches are valid and which not	04:43
	→ ajt joined	04:43
philllllijw	that's really not possible	04:44
	← shiba_yu36 left	04:44
wereHamster	can you elaborate?	04:44
	→ harshpb joined	04:44
RandalSchwartz	s/less/fewer/	04:44
	← ormris left	04:44
philllllijw	I worked on 4 different branches just today	04:44
	→ jeff2 joined	04:44
philllllijw	RandalSchwartz, I realized 'less' sounded ridiculous after writing it	04:45
	← gchristensen left	04:45
philllllijw	wereHamster, I have one branch per "fix ticket" essentially	04:45
wereHamster	but you fix the ticket, then it's done. right?	04:45
philllllijw	wereHamster, it has to get tested first	04:45
jrnieder	phillllllijw: one thing I sometimes do is to email out a patch and then delete the corresponding branch	04:45
	gitinfo set mode: +v	04:46
philllllijw	jrnieder, how do you email a patch?	04:46
jrnieder	once the patch is in others' hands, I can always pick it up again and reapply it	04:46
wereHamster	philllllijw: git brnach -m ticket-xyz-in-testing	04:46
jrnieder	phillllllijw: I use "git format-patch" to format it nicely	04:46
	and then cp/paste with $editor_of_choice in mutt	04:46
	others seem to be fond of "git send-email" or "git imap-send", or use attachments	04:47
philllllijw	i'm on windows for the time being	04:47
jrnieder	s/mutt/$mailer_of_choice/ then	04:47
philllllijw	im guessing those commands won't do much for me	04:47
jrnieder	I don't see why they shouldn't	04:47
	→ TL joined	04:48
jrnieder	anyway, I said I throw away the branch when I email out the changes	04:48
	→ pidus joined	04:48
jrnieder	but sometimes I push out the changes instead	04:48
philllllijw	I dont know who I would email the patch to	04:48
jrnieder	same principle applies --- once the branch is in others' hands I trust, I don't need my local copy any more	04:49
wereHamster	the person doing the testing..?	04:49
jrnieder	attach to ticket tracker?	04:49
philllllijw	ah, maybe that's the disconnect. We don't currently have local testing set up	04:49
	which is stupid but... temporary	04:49
	i'll have to look into the patch thing some more. I'm not very familiar with how that's normally done	04:50
	→ jason237 joined	04:50
jeff2	I'm trying to `git push -u origin master`, but am getting "ERROR: Permission to .. denied to XXX", even though my username is set correctly in ~/.gitconfig and .git/config - it is still picking up "XXX" from somewhere. What configuration option am I missing?	04:50
	← glennpratt left	04:50
	→ shiba_yu36 joined	04:51
	→ kryl99_ joined	04:51
jrnieder	jeff2: it gets your username from the OS	04:51
bob2	pretty sure there's no where to put a username in ~/.gitconfig	04:51
	← pidus left	04:51
bob2	http://bpaste.net/ .git/config	04:51
RandalSchwartz	probably a .ssh/config problem :)	04:51
jeff2	jrnieder: As in $USER? It is getting some other name, not $USER..	04:52
bob2	the answer is more or less 'make ssh work', though	04:52
jrnieder	and your ssh key through the usual channels (~/.ssh/*, $SSH_AUTH_SOCK)	04:52
	ah, listen to merlyn, he's making more sense :)	04:52
	← legnaleurc left	04:53
jrnieder	you can override the username in a url by adding <username>@ before the hostname	04:53
	← aidalgol left	04:53
jeff2	I've added User in ~/.ssh/config, but there's no mention of user XXX anywhere in that config	04:53
	→ akiress joined	04:53
jeff2	cool that might work, let me try	04:53
jrnieder	but probably best to nip the underlying configuration problem in the bud like randalschwartz suggests	04:54
jeff2	well, my URL is: [email@hidden.address] - so the user is already set to 'git' right?	04:54
	→ sbell joined	04:54
	→ shiba_y__ joined	04:54
	→ themachine joined	04:54
	← shiba_yu_ left	04:54
	→ lyddonb joined	04:54
jrnieder	jeff2: yes, so it's probably the ssh key rather than the username that's giving you trouble	04:54
	→ kerrick joined	04:55
bob2	(that'd probably be clearer if you hadn't redacted the error)	04:55
	← shiba_yu36 left	04:56
jeff2	definitely an SSH error. will investigate	04:56
	→ Trezker joined	04:59
	→ plasticle joined	04:59
	← plasticle left	05:00
	← norton- left	05:02
	→ c1de0x joined	05:05
	← mandric left	05:05
	← jeff2 left	05:06
	← Splash left	05:08
	← lyddonb left	05:08
	← harshpb left	05:08
	← cilly left	05:09
sitaram	RandalSchwartz: http://dpaste.org/MI91I/ -- I call it "tar--delete"	05:11
	EugeneKay: ^	05:12
EugeneKay	Scary	05:13
sitaram	EugeneKay: not really; it's actually doing something pretty simple	05:14
	→ cilly joined	05:14
	← acl_ left	05:14
sitaram	I just went overboard on the comments as usual	05:14
EugeneKay	/ # ; comment-ception	05:14
	//*	05:14
	← JohnBat26 left	05:15
	sitaram didn't get it...	05:15
	→ harshpb joined	05:15
sitaram	explanation for us low-IQ folks please	05:16
EugeneKay	Probably for the best	05:16
sitaram	aah well ok	05:16
	← diogogmt left	05:16
RandalSchwartz	interesting	05:16
	← neurodrone left	05:19
	→ dreiss joined	05:21
	→ jeff2 joined	05:23
	gitinfo set mode: +v	05:24
	→ acl_ joined	05:24
	← wagle left	05:24
jeff2	never knew SSH caches public keys in memory, but rebooting fixed my problem	05:24
	← unkmar left	05:24
	← ajt left	05:27
	← vandal- left	05:29
wereHamster	jeff2: it's the ssh agent who does	05:30
	← nobitanobi left	05:31
	→ Splash joined	05:31
	→ disappearedng joined	05:32
	→ nobitanobi joined	05:32
bob2	no need to reboot	05:33
	← TL left	05:34
	← Linc left	05:35
	→ wagle joined	05:35
sitaram	ssh-add -D # or similar, from memory	05:36
	← ericmuyser left	05:37
	← freezey left	05:38
	← kaste left	05:40
	→ pidus joined	05:41
wereHamster	man ssh-add, from memory :P	05:42
	← ceej left	05:47
tga	woot, figured out branching and set up my process	05:48
	tga has seen the light	05:48
EugeneKay	Is it a disco ball or a laser pointer or some sort of trance rave?	05:48
	→ freimatz joined	05:48
	← jrnieder left	05:49
tga	bright white light, once you see it there is no going back	05:49
	there is life before dvcs and life after dvcs	05:49
milki	what happens during dvcs?	05:50
EugeneKay	The same things that happen during potato	05:51
	← matthiaskrgr left	05:52
	→ matthiaskrgr joined	05:54
	→ krz joined	05:54
	→ shergill joined	05:56
sitaram	you get peeled then boiled?	05:57
	→ dfr\|mac joined	05:57
milki	you are supposed to peel them before boiling them?	05:57
sitaram	or after I guess; after is easier if you wait for it to cool down	05:58
	← f0i left	05:58
wereHamster	why the extra work to peel them?	05:58
	it's not like the skin is poisonous ...	05:58
sitaram	well if you buy 'em from the farmers market they need to be cleaned well. But this is getting serious and I can't think of something funny to say	06:00
	milki: hey, thinking about that packaging doc, I just realised that the actual content was in the "install" doc since Feb 2010 (2 years ago). What I said the other day was merely when it moved to its own doc.	06:00
milki	oo	06:00
	→ nrichards joined	06:00
sitaram	I wrote so much I forgot when I wrote what!	06:01
	→ f0i joined	06:01
	← dfr\|mac_ left	06:01
milki	hehe	06:01
sitaram	which in the end could be the real reason to use a VCS :)	06:02
milki	git blame!	06:02
sitaram	yup	06:02
	→ jceb joined	06:03
kevlarman	blame git!	06:04
	← gnufied left	06:05
	← Targen left	06:05
	← Chibby_ left	06:05
	→ Chibby_ joined	06:08
	← dfr\|mac left	06:08
	← sbell left	06:09
	← guampa left	06:09
	→ maleknet joined	06:09
	→ JB3181 joined	06:11
	← squentin left	06:11
	→ gnufied joined	06:11
	→ squentin joined	06:12
	← daed left	06:12
sigs	it's not a real vcs anyway	06:13
	← eno left	06:14
	← hyperair left	06:14
	→ daed joined	06:14
sigs	(don't blame me / for my son Stan / he apt-get installed git / and now he's off to join a clan)	06:14
	*the clan, apparently	06:14
	→ eno joined	06:15
	← eno left	06:15
	→ eno joined	06:15
	→ ThomasLocke joined	06:17
	→ michaelh joined	06:18
	← r126t left	06:19
	→ r126t joined	06:20
	→ AmirBehzad joined	06:21
	← JB3181 left	06:21
	← L42y left	06:22
	→ L42y joined	06:22
	→ swombat_ joined	06:23
	← dennda_ left	06:23
	→ dennda_ joined	06:23
	← paphio left	06:23
	→ Pupeno_W_ joined	06:23
	→ JB3181 joined	06:24
	← swombat left	06:25
	→ maletor joined	06:25
	← Pupeno_W left	06:25
	Pupeno_W_ → Pupeno_W	06:25
	← fisted left	06:25
	← Splash left	06:26
	← shiba_y__ left	06:27
	→ fisted joined	06:28
	→ marthinal joined	06:28
	→ shiba_yu36 joined	06:29
	→ t4nkd joined	06:29
	← vdv left	06:29
	→ vdv joined	06:29
	→ anjen joined	06:30
	← MestreLion left	06:31
	← shiba_yu36 left	06:32
	← tga left	06:32
	→ ajaykools joined	06:33
	← bgerber left	06:33
	← L42y left	06:34
	→ Dave^\| joined	06:34
	→ tga joined	06:34
	← guns left	06:34
	← t4nkd left	06:34
	→ L42y joined	06:35
	← mikepack left	06:35
	← thursday left	06:35
	→ jargon- joined	06:35
	→ jhoepken joined	06:36
	← dreiss left	06:37
anjen	So I've been maintaining a RoR project for a while, but I haven't had any reason (til now) to branch the project. I've finally gotten to the point where we're movinng off ruby 1.8.6, and in order to make sure that my upgrade doesn't futx things, I'd like to branch the repo	06:37
	→ bgerber joined	06:37
anjen	so, I've created a branch (r187_upgrade)....Am I correct that "checkout" cimply changes the branch on my local copy (until I push something to the remote)?	06:38
	← eletuchy left	06:38
anjen	*simply, not cimply	06:38
	← jhoepken left	06:39
milki	yes	06:39
	← themachine left	06:39
anjen	ok, great, thanks. next question if I may	06:40
	→ jacekmigacz joined	06:40
sitaram	well... for now, yes you may. Not sure about later	06:40
anjen	so now I am on the r187_upgrade branch. if I push, it wil send to a new branch at github of the same name, correct?	06:40
	skylord5816 → Xd-2PnUc	06:40
	sitaram is kidding of course	06:40
sitaram	the first time, only if the push explicitly names it	06:41
anjen	that is,git push origin r187_upgrade	06:41
sitaram	'git push' won't. 'git push origin r187_upgrade' will. Later pushes, 'git push' will	06:41
	(assuming all default config settings)	06:41
	Xd-2PnUc → skylord5816	06:41
anjen	sitaram: right, understood	06:41
	last question:	06:41
	← albel727 left	06:42
	← skylord5816 left	06:42
	sitaram waits	06:42
anjen	if I modified source code on the local dev environment prior to creating the branch, is there any issue with doing what I noted above? Or do I need to comomit and push the original to the branch first?	06:42
	→ mishok13 joined	06:43
anjen	(essentailyl to populate the github repo branch)	06:43
	*essentially	06:43
sitaram	modified but did not add/commit? no issue. When you 'git checkout -b some-new-branchname' it'll carry the changes across	06:43
	→ themachine joined	06:43
	→ pro_metedor joined	06:43
anjen	yeah, I've modified it (unfroze a rails 2.3.2 app) then refroze at 2.3.5	06:44
	← themachine left	06:44
anjen	but I haven't commited or anything else yet	06:44
sitaram	brb	06:44
	→ flijten joined	06:45
sitaram	should work fine, IIUC	06:45
anjen	oh....methinks *maybe I understand. git checkout -b some-new-branchname will create a new branch AND change to it. So 1 step two ops. easier in the long run.	06:45
sitaram	yes	06:46
	milki never gets around to doing -b >.<	06:46
sitaram	it will do that, and cause your subsequent commits to go on that branch because that is now your current branch	06:46
anjen	understood	06:46
	thanks for the help sitaram :)	06:47
	→ martinjlowm-lptp joined	06:47
sitaram	ywyyw	06:47
	you're welcome!	06:47
	(macro fail!)	06:47
milki	lol its a macro?	06:47
	>.<	06:47
	← pro_metedor left	06:48
sitaram	yeah; an irssi macro. I'm a bad typist so it makes sense to have frequent things be macros	06:48
	→ duckxx joined	06:48
milki	haha	06:48
	i like typos	06:48
	so im okay with it	06:49
sitaram	think of it like what hallmark did to the actual sentiment of wishing someone well/happiness/etc	06:49
	without the price :)	06:49
	→ dreiss joined	06:50
	→ dermoth\|home joined	06:51
	← btanaka left	06:52
	← anjen left	06:53
EugeneKay	And glitter	06:55
	→ TrineLocke joined	06:56
	→ senny joined	06:57
	→ Zer000 joined	06:58
	→ pilotbub joined	06:58
Zer000	hey guys, I need to share a repository with the world but my full name is in the initial commit. How can I change that	06:58
	← davidrussell left	06:58
sitaram	changing your name is hard	06:59
	oh you meant change the commit... hmm	06:59
Zer000	No i mean my name is showing up as the person who made the (initial) commit	07:00
	this repository is new, so maybe there's a way to redo everything?	07:00
sitaram	you don't want credit? anyway I think if you use the new "--orphan" option to checkout it can be done	07:00
EugeneKay	!orphan	07:01
gitinfo	To create an orphaned branch(unrelated to existing history): git symbolic-ref HEAD refs/heads/orphan; rm .git/index; git clean -fdx; touch foo; git add foo; git commit -m 'Initial commit on branch orphan'	07:01
EugeneKay	Or you can filter-branch it	07:01
Zer000	thanks ok.	07:01
	← Zer000 left	07:02
	← grkvlt left	07:02
	→ grkvlt joined	07:02
	→ amigojapan joined	07:02
pilotbub	building git atm looking forward to struggling to accomplish basic tasks with it and giving up after a few painful minutes	07:02
EugeneKay	"building" ?	07:02
	→ h0bbit joined	07:02
EugeneKay	What's wrong with the packaged version?	07:03
pilotbub	no binaries for my platform	07:03
	→ hh__ joined	07:03
	gitinfo set mode: +v	07:03
hh__	is it possible to add a file and to say don't track version history for this file?	07:04
	→ scottj joined	07:04
	→ ericmuyser joined	07:04
sitaram	zendeavor: I think filter-branch is likely to work better; the orphan idea is much more painful	07:04
hh__	for instance, i just have a image file that i don't want to have a history of, but would like it to be tracked for changes	07:05
scottj	if I accidently commit to master instead of a branch how would I 1) move those commits into a branch and out of master or 2) just get rid of the commits from master so that when I pull from origin it works fine?	07:05
	→ ipalaus joined	07:06
	→ vanna^^ joined	07:06
pilotbub	git is just one of those things i know i will hate but must endure	07:07
	→ W8uiE5 joined	07:07
	← vanna^^ left	07:08
_ikke_	pilotbub: What is the reason you hate it? Because you are forced to use it?	07:09
hh__	scottj: i think you just need to delete the commit, it shouldn't affect origin	07:10
	→ airborn joined	07:10
_ikke_	hh__: You can't just 'delete' a commit ;)	07:10
	← Prime326 left	07:10
hh__	well you can 'reset' the last known place HEAD was or whatever	07:10
pilotbub	yes its all the rage these days if i want to hack around with software theres no avoiding it	07:10
scottj	hh__: is there a way to move the commits to a branch before deleting them?	07:10
_ikke_	scottj: rebase	07:11
	scottj: How many commits are you talking about?	07:12
scottj	_ikke_: two	07:12
	→ Dreamer3__ joined	07:12
hh__	git rebase -i master~2	07:12
	← Dreamer3 left	07:12
scottj	_ikke_: I've used rebase to reapply my commits in a branch after pulling in from master, not sure how to use it to move commits from master to a branch	07:13
hh__	if it's the last 2 commits, i think that should do it	07:13
	← dreiss left	07:14
scottj	_ikke_: oh, that's for removing the commits from master, not moving them to a branch before doign that right?	07:14
_ikke_	scottj: No, rebase is for moving them	07:15
scottj	reset is for deleting?	07:15
	"deleting"	07:15
_ikke_	'deleting'	07:15
	yeah	07:15
	→ arvind_khadri joined	07:15
	← arvind_khadri left	07:15
	→ arvind_khadri joined	07:15
scottj	do I run that rebase from inside the new branch? do I branch from master before or after resetting?	07:15
	← TrineLocke left	07:16
hh__	_ikke_: you know if git has a way to add a file to track, but never maintain it's history? kind of like cloning with shallow?	07:16
	scottj: run it from inside the branch you want to move master's last 2 commits to	07:16
_ikke_	hh__: I know you can set assume-unchanged for that file.	07:16
	scottj: What does git log master~2..HEAD return?	07:17
	← acl_ left	07:17
_ikke_	hh__: First tell what you want to achieve?	07:17
hh__	well for instance my project has a twitter/facebook icon images, i really don't care to keep a history of them, but i want to include with my repo to be complete	07:18
	→ dc5ala joined	07:19
	← jceb left	07:19
scottj	_ikke_: my two commits	07:19
_ikke_	scottj: Ok, and on which branch are you currently?	07:19
	hh__: There is a project that allows you to externally store such files	07:19
	→ cbreak-work joined	07:20
scottj	_ikke_: both master and my feature branch (branched from master after those two commits)	07:20
_ikke_	hh__: git-annex	07:20
	← macmartine left	07:20
	← Bass10 left	07:20
pilotbub	this process is driving me headlong into a suicide net just like a chinese apple slave	07:21
_ikke_	scottj: git rebase master~2 <feature-branch>	07:21
	pilotbub: What platform are you using?	07:21
pilotbub	10.5	07:22
	→ hyperair joined	07:22
_ikke_	10.5 what?	07:22
	← hyperair left	07:22
	→ hyperair joined	07:22
pilotbub	ppc	07:22
IcePic	macbuntu? =)	07:22
EugeneKay	pilotbub: !refund	07:23
gitinfo	pilotbub: If you are not satisfied with git for whatever reason, you are entitled to a full refund of the purchase price, and are invited to use another VCS. Elsewhere.	07:23
EugeneKay	;-)	07:23
	pilotbub golf claps	07:23
_ikke_	EugeneKay: He says he has no choice but to use git ;)	07:23
	→ eighty4 joined	07:23
scottj	opps, is there a way to revert git reset --hard HEAD~2 ?	07:24
_ikke_	pilotbub: But instead of trying to fight it, mayby try to embrace it	07:24
EugeneKay	scottj - man git-reflog	07:24
gitinfo	the git-reflog manpage is available at http://jk.gs/git-reflog.html	07:24
_ikke_	scottj: git reset --hard @{1}	07:24
scottj	_ikke_: thanks	07:24
pilotbub	here's a query: why doesn't git release a UB pkg seems like a no-brainer to this end user	07:25
	← eighty4 left	07:25
	← stringoO left	07:25
	charon .oO( UB? )	07:26
	← csherin left	07:28
IcePic	I heard "why doesn't the whole world accomodate my personal needs?", but that's just me.	07:28
_ikke_	IcePic++	07:29
pilotbub	IcePic are you psychotic	07:29
IcePic	yes	07:29
	← dyek left	07:29
	← jakob left	07:29
	← eean left	07:30
	← mishok13 left	07:30
pilotbub	PowerPC started it all lest ye forget	07:30
IcePic	I would like for the git coders to make a unix program that works on unices, and not spend time making a debi..ubun...mintOS package all day.	07:30
	→ jakob joined	07:30
EugeneKay	Cool story, pilotbub.	07:30
	→ aidalgol joined	07:30
EugeneKay	If you're having an issue compiling git.git for $PLATFORM, I suggest you post to the mailing list for help and/or with a patch to make it work.	07:30
	→ _iron joined	07:31
EugeneKay	The last major PC manufacturer to use PPC abandoned it what... 5 years ago?	07:32
pilotbub	lol last PC manufacturer	07:32
IcePic	I think IBM still has some kind of POWER line, but yes. noone sells PPCs in stores since long ago.	07:32
EugeneKay	So it's very much a case of "nobody has done it because nobody cares anymore. Feel free to do it yourself"	07:33
IcePic	but if anyone wants a DonQuixote-cause to chase, my m68k runs bsd, and by GOD noone makes precompiled git packages for it, would you believe the madness?	07:33
charon	last i heard the Makefile worked on way more arcane archs/OSes like AIX on god knows what processor	07:33
pilotbub	maybe i will	07:33
	EugeneKay sticks head under rock at mention of the word "makefile"	07:34
pilotbub	if the bloody thing would compile	07:34
cbreak-work	Sony uses PowerPC CPUs in their PlayStation	07:34
IcePic	if it in fact doesnt, that would explain why you cant find precompiled packages for it. O_o	07:35
pilotbub	its stalled on gettext but terminal is still responsive... only time will tell	07:35
IcePic	uhu	07:36
	← nobitanobi left	07:36
	← jakob left	07:36
	→ apok joined	07:37
EugeneKay	Ah, PS3. Forgot about that thing.	07:37
	← regedarek left	07:37
	← kennethreitz left	07:37
EugeneKay	Sadly Sony decided to piss off the linux hacking community, so I doubt anybody has gotten to gitting on it	07:37
pilotbub	im still waiting for the n64 controller usb dongle with mac support	07:38
	← kadoban left	07:38
EugeneKay	The official controller of people with three hands	07:38
	← ipalaus left	07:38
	← W8uiE5 left	07:38
	← JB3181 left	07:38
	→ jakob joined	07:39
	← ajaykools left	07:41
	→ kaste joined	07:41
	→ eean joined	07:42
	→ eletuchy joined	07:42
	← jakob left	07:43
	→ gusnan joined	07:44
	← nixmaniack left	07:45
	← beautifulmind left	07:45
	→ giallu joined	07:47
	→ jutaro joined	07:48
	← jutaro left	07:49
	→ ph^ joined	07:49
	→ ph^_ joined	07:51
	← h0bbit left	07:51
	← aidalgol left	07:51
	← wagle left	07:52
	→ Spockz joined	07:53
	← eletuchy left	07:53
	← ph^ left	07:53
	→ jakob joined	07:53
	→ h0bbit joined	07:55
	→ OOPMan joined	07:57
sitaram	pilotbub: what is "UB"?	07:59
	sitaram hates duckduckgo-ing for 2-letter abbrevs; too many false matches	07:59
	→ mishok13 joined	08:00
charon	sitaram: it dawned on me at some point that it must be the mac os "unified binary" stuff that has ppc and x86 executable images in the same file	08:00
	→ bjoernh joined	08:01
	→ jceb joined	08:01
sitaram	aah Apple stuff. I shoudn't have asked :)	08:02
	thanks charon	08:02
	→ ph^ joined	08:02
	→ gldn joined	08:03
	← ph^_ left	08:04
	← Spockz left	08:04
	← jakob left	08:05
	← no1home left	08:05
tga	if I have a bunch of changes in one branch how can I also merge them into another branch _with the exception of one file_?	08:05
	→ paphio joined	08:06
	← lapistano left	08:06
tga	other than manually copying and replacing that file, which I'm considering as an option	08:06
_ikke_	merge -n	08:06
	→ dreiss joined	08:06
	→ lapistano joined	08:06
_ikke_	wait	08:06
	← Chibby_ left	08:07
_ikke_	merge --no-commit	08:07
charon	tga: check out the other branch and 'git show --pretty=email $commit -- $all_other_files_affected \| git am'	08:07
	← airborn left	08:08
	→ tiglog joined	08:08
	→ jakob joined	08:08
tga	funny, so that turns changes into individual patches	08:09
	that I can then apply	08:09
	pilotbub is still building...	08:09
	→ eletuchy joined	08:09
tga	_ikke_, and what you're suggesting is to merge everything, then remove the file I don't want before commiting	08:09
	right?	08:09
orzo	i dont see how you would specify $commit when you are merging a branch	08:09
charon	tga: well, except for the $all_other_files_affected that is roughly equivalent to 'git cherry-pick $commit'	08:09
	tga: at least semantically	08:09
	← kerrick left	08:09
	← tiglog left	08:09
charon	orzo: a branch is a pointer to a commit. you can use the branch name to "name" the commit on any command line that expects a commit (usually spelled "commit-ish" in the manuals)	08:10
	→ tiglog joined	08:10
	← X-tonic left	08:10
	→ leavittx joined	08:10
charon	but yeah, in tga's case having several commits that must be "merged" (it won't be true a merge even if you fake one!!) complicates "my" strategy	08:11
	← jceb left	08:11
orzo	maybe git should let you specify files to the merge command	08:12
	→ thiago joined	08:12
	← h0bbit left	08:12
_ikke_	orzo: git merge merges branches, not files	08:13
	→ jceb joined	08:13
	→ Milossh joined	08:14
	→ LeMike joined	08:15
charon	orzo: the problem is that such "fake" merges, if you created them (it's certainly possible), potentially mess up every merge after them	08:15
	← LeMike left	08:15
charon	because a merge is a promise that you brought in all changes from all sides	08:15
	→ teweWORK joined	08:15
	← pilotbub left	08:16
charon	for example, there would be no way to say "and now also merge the changes that i skipped previously"	08:17
	→ caseymcg joined	08:17
	→ toabctl joined	08:17
	→ h0bbit joined	08:18
	→ batlock666 joined	08:19
orzo	maybe it would help to have a way to a way to retroactively edit a commit to be split up into multiple commits for each affected file and then do a cherry pick	08:19
	← Lgb left	08:20
orzo	i suppose such a thing could be done in a way to maintain compatibility with clones that didn't do the separation too	08:21
_ikke_	orzo: git rebase	08:21
	→ Sitethief joined	08:21
	← Element9 left	08:22
	→ Lgb joined	08:22
	→ Daniel joined	08:22
	→ stodan joined	08:22
orzo	it could be implmeneted so that you could push the seperation to a repo that hasnt done it	08:23
	heh	08:23
	← leavittx left	08:26
	→ psoo joined	08:26
	← gabalafou left	08:26
	→ ipalaus joined	08:26
	← gitinfo left	08:27
	→ BiggFREE joined	08:27
	→ stoffus joined	08:28
	← etcetera left	08:28
	← oc80z left	08:28
	← dotemacs left	08:28
	← CEnnis91 left	08:28
	← necromancer left	08:28
	← pl0sh left	08:28
	← arnihermann left	08:28
	← Dr4g left	08:28
	← qknight left	08:28
	→ dpino joined	08:28
	← jefferai left	08:28
	← jmcantrell left	08:28
	→ jefferai joined	08:28
	← Blackshark left	08:29
	← cytrinox left	08:29
	← NimeshNeema left	08:29
	→ qknight joined	08:29
	← MetaCosm left	08:29
	← lyddonb_ left	08:29
	← dkannan left	08:29
	← garrys_game left	08:29
	← Harzilein left	08:29
	→ vervic joined	08:29
	→ oc80z joined	08:29
	→ cytrinox joined	08:29
	→ dkannan joined	08:29
	← acrocity left	08:30
	→ Blackshark joined	08:30
	← jakob left	08:30
	→ gabalafou joined	08:30
	→ jakob joined	08:30
	← duckxx left	08:30
	→ duckxx joined	08:31
	→ acrocity joined	08:31
	→ gitinfo joined	08:31
	ChanServ set mode: +o	08:31
	← h0bbit left	08:31
	→ etcetera joined	08:31
	→ oso96_2000 joined	08:31
	→ pro_metedor joined	08:32
	→ glcrazy joined	08:32
	← dpino left	08:32
	→ vpopov joined	08:34
	→ dpino joined	08:34
	← tatsuya_o left	08:35
	→ soc42 joined	08:35
	→ necromancer joined	08:35
	← duckxx left	08:36
	→ garrys_game joined	08:36
	← Daniel left	08:38
	← ericmuyser left	08:39
	← glcrazy left	08:39
	← amigojapan left	08:40
	→ barr5790 joined	08:40
	→ eijk joined	08:41
	← scientes left	08:42
	→ jacobrask joined	08:44
jacobrask	I'm having some issues with git status performance. It's a pretty new and small repository, but with a handful of submodules. Will submodules slow down git status in a repository? git status . is way faster	08:44
	→ whitman joined	08:45
	← etcetera left	08:45
	→ johnhamelink joined	08:45
jacobrask	it only takes about 2 seconds now but the index file is only 57K so it should be instantaneous	08:45
_ikke_	Git status has to check every submodule too, so it could be a problem	08:45
	→ LeMike joined	08:46
	→ Dr4g joined	08:47
	→ snearch joined	08:47
EugeneKay	Buy a SSD / more ram.	08:48
jacobrask	so git-update-index and things like that would probably not do any difference	08:48
	well it's running in a virtual machine with my development environment	08:48
EugeneKay	And/or run 'git status' less(preferably, ditch your PS1 script or make it asynchronous)	08:48
_ikke_	VM's aren't well-known for their IO speed	08:48
	→ dotemacs joined	08:49
	→ rendar joined	08:49
EugeneKay	Out of sheer boredom, I wrote such an async-ish PS1-git into my .bashrc: https://github.com/EugeneKay/scripts/blob/master/HOMEDIR/.bashrc	08:50
jacobrask	then that's probably the issue, I'm getting a new work station in a couple of weeks and can allocate some more ram to the vm	08:51
	→ caseymcg1 joined	08:51
EugeneKay	It reloads whenever 'git' is run, or every 5 minutes.	08:51
	← caseymcg left	08:51
jacobrask	nice	08:51
EugeneKay	Or if you change repos, but not dirs within repos.	08:51
	← caseymcg1 left	08:51
EugeneKay	That latter bit only matters as far as the untracked count is concerned, because it works from ., rather than the top-level of the repo.	08:52
	I'm waffling between removing the untracked counter altogether and moving it to counting the whole thing.	08:52
	← hyperair left	08:52
	← merlin83_ left	08:52
	→ tcurdt joined	08:52
tga	I just love the git messages	08:53
	You are in 'detached HEAD' state.	08:53
	← Milossh left	08:53
_ikke_	!detached	08:53
gitinfo	A detached head occurs when you are not currently checked out into a local branch, or in other words there is no way for your commits to change a reference. This can happen if you check out a tag, remote tracking branch, or SHA; or if you are in a submodule or you are in the middle of a merge or rebase that is conflicted.	08:53
EugeneKay	I think the plumbing could use a reinvention(2.0, perhaps?), but it makes sense once you look at it from a historical perspective.	08:54
wereHamster	what for?	08:54
	→ Shusshu joined	08:54
EugeneKay	The fact that checkout does a half-dozen different things	08:54
	→ caseymcg joined	08:54
EugeneKay	"add" and "rm" don't mean the opposite thing	08:55
	→ cespare joined	08:55
wereHamster	two, right?	08:55
	→ MetaCosm joined	08:55
	→ pl0sh joined	08:55
EugeneKay	All the lil things that aren't exactly intuitive, but have to be memorized	08:55
cespare	I just ran into the recent gitdir -> superproject change. Is there a document I could read that describes the motivation behind it/	08:55
EugeneKay	There's checkout foo, checkout -b, checkout --orphan, checkout --....	08:56
cespare	(for submodules)	08:56
	→ NimeshNeema joined	08:56
	→ ilius joined	08:57
_ikke_	EugeneKay: Isn't that the porcelain instead of plumbing?	08:57
	→ vandal- joined	08:57
EugeneKay	_ikke_ - indeed it is. I need sleep.	08:58
wereHamster	EugeneKay: but you wouldn't propose a separate commadn for each of those things, right?	08:58
	→ etcetera joined	08:58
	→ arnihermann joined	08:58
	→ Daniel joined	08:58
EugeneKay	wereHamster - I dunno, I'd have to sit and design it, and I'm not a UI designer.	08:59
	I can't draw a pretty lady, but I can point out an ugly one to you.	08:59
	→ tvw joined	08:59
	ChanServ set mode: +v	08:59
	→ shruggar joined	09:01
	← Araxia left	09:01
	→ lyddonb_ joined	09:02
	→ h0bbit joined	09:02
tga	when doing a merge can I just specify the commit to use from the other branch so I don't pick up the latest one?	09:02
	→ Milossh joined	09:02
Daniel	Hey guys, does anyone know the reason why "--follow" isn't the default behavior when doing a "git log <filename>" or "gitk <filename>"?	09:02
	← Milossh left	09:04
	← yuriks left	09:04
	← DrHouse\|TARDIS left	09:05
	→ robotmay joined	09:05
_ikke_	tga: What do you want to achieve?	09:06
	→ flavius joined	09:06
	→ Sigma joined	09:09
	→ break57382 joined	09:09
	← h0bbit left	09:10
tga	_ikke_, merge with a certain commit from the other branch	09:11
	not with the latest one	09:11
	I ended up doing a reset and setting up a new branch	09:11
	then merging with that	09:11
	← ph^ left	09:12
tga	or at least that's what I'm trying to do, I'm not sure it's working	09:12
	→ canton7 joined	09:12
	→ anddam joined	09:12
	→ merlin83_ joined	09:12
anddam	I checked out a branch, made some changes and committed, I thought the commit would be put into the branch but I'm in a detached state	09:12
cespare	Where can I find documentation for "gitdir"? (A file named .git that points to the git directory)	09:13
anddam	in order to put the commit into the branch I want do I need to use git merge?	09:13
thiago	anddam: git commit puts it in a branch	09:15
	anddam: the problem is that you were not in a branch to begin with	09:15
	when you say "I checked out a branch", you didn't. You did something wrong there.	09:15
shruggar	cespare: man gitrepository-layout	09:16
gitinfo	cespare: the gitrepository-layout manpage is available at http://jk.gs/gitrepository-layout.html	09:16
	→ soulcake joined	09:16
cespare	shruggar: thanks!	09:17
anddam	thiago: I wonder why I was detached, I did a git checkout remotes/origin/gh-pages	09:17
shruggar	hm, though it doesn't appear to have any more information than what you just said	09:17
thiago	anddam: that's why	09:17
	anddam: you did not check out a branch. Remote refs don't count as branches.	09:17
shruggar	cespare: what else do you want to know?	09:17
anddam	thiago: what should I've done instead?	09:17
cespare	shruggar: Nothing :) I couldn't find information about it and I was wondering what the specification was	09:18
thiago	anddam: checked out a branch	09:18
cespare	I'm glad that it's that simple.	09:18
	→ LimeBlast joined	09:18
	→ Element9_work joined	09:18
thiago	anddam: only stuff in refs/heads (a.k.a. "local branches") count as branches	09:18
cespare	shruggar: this change confuses a library I'm using so I wanted to patch it up	09:18
thiago	anddam: if you don't have a branch yet, create one to track that remote's branch	09:18
anddam	thiago: now I do, local gh-pages branch	09:19
	so I have to merge my 21c3b76 into gh-pages branch	09:19
thiago	anddam: cherry-pick it	09:19
tga	alright, I tried a few things and I'm still confused	09:19
anddam	thiago: why not merge?	09:19
tga	so again, if I have one single commit in another branch that I want to apply to head	09:20
thiago	anddam: in this case, it might not make a difference	09:20
tga	how can I only get those changes?	09:20
anddam	in fact it did work	09:20
	thanks	09:20
	thiago: so the issue was checking out a remote branch	09:20
	s/branch/ref	09:20
thiago	yes	09:21
tga	format-patch and am?	09:21
thiago	you can't check out anything but local branches. If you try to check out anything else, you go into "detached head" state.	09:21
	that allows you to be in that commit, but you cannot modify the ref by committing.	09:21
anddam	I see	09:22
	→ h0bbit joined	09:22
	→ tqmz joined	09:22
anddam	thiago: when I checked out origin/gh-pages local gh-pages was created, tho'	09:22
	I didn't use -b	09:22
thiago	anddam: I don't know. There are certain shortcuts that can create a local branch tracking a remote of the same name.	09:23
anddam	gh-pages merges with remote gh-pages	09:23
	master merges with remote master	09:23
thiago	anddam: I never remember what they are and I end up creating a branch called "--track"	09:23
anddam	is this where local branch and remote refs are connected?	09:23
	that is when I pull from remote the local branch get merged	09:23
thiago	that "merges with" means that the remote's branch is the upstream for this local branch	09:24
	if you "git pull" with no arguments, it will merge that remote branch	09:24
anddam	no, I definitely got confused by the terminology	09:24
thiago	if you "git push" with no arguments and push.default is set to "upstream", it will push the local branch into the upstream	09:24
	→ Milossh joined	09:24
anddam	I have to go through a complete tutorial, not just taking the bit I need at the moment	09:25
	thanks again, bye	09:25
	KyleXY_ → KyleXY	09:25
	← anddam left	09:26
	← canton7 left	09:26
	← Vile left	09:27
	← disappearedng left	09:27
	← ehsan_ left	09:27
	→ eijk_ joined	09:27
	← eijk left	09:28
Daniel	Trying this again before moving on to my next question. Does anyone know the reason why "--follow" isn't the default behavior when doing a "git log <filename>" or "gitk <filename>"?	09:29
_ikke_	Daniel: No clue	09:30
	→ moremorefun joined	09:30
_ikke_	Maybe because it does extra work which costs time	09:30
	← smw left	09:32
Daniel	Hm, could be. But I think more times then not people would want the full history. In my eyes it would make more sense to make --follow default and have something like --fast or --short to just take the quick way out.	09:32
cmn	it does extra work on each iteration, only works for one file, it's a hack and it's very un-gittish	09:32
	← L42y left	09:32
Daniel	How is it un-gittish? Just because it works on files and not commits?	09:34
cmn	it works on a file rather than a source tree	09:34
	rather, it focuses on what happened to one single file and ignores anything that might have happened around it	09:35
	→ nixmaniack joined	09:35
Daniel	But it's a quick and effective way to find out what someone did in a particular file. And I don't know about you, but the code I read and the code I produce ends up in files.	09:36
	→ d3vic3 joined	09:37
shruggar	I keep mine on disks	09:37
	← nixmaniack left	09:38
	→ johnathanb joined	09:38
cmn	the code i write tends to be distributed around several files	09:38
tga	if I have a master project and sub-projects that only have different database and settings files	09:39
cmn	!config_files	09:39
gitinfo	[!configfiles] the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.	09:39
shruggar	until git log -L is completed, it's all useless anyway :)	09:39
	git log -L is more gittish than git log --follow	09:39
tga	what's a sane way of keeping the code in one repo and only the db and settings in each project?	09:39
cmn	what options is -L?	09:39
	→ nixmaniack joined	09:40
mjt	i remember I used git rev-parse to find a "pretty-printed" "place" where a given commit is -- like sometag+Nrevs. But can't see how to do that from the manpage. It was something very simple, with an option referencing to a branch or remote or something like that... What it was? :)	09:40
charon	cmn: my eternal todo item	09:40
cmn	I see	09:40
shruggar	cmn: "line-level history". eg: "show me the history of this function, no matter what file it was in"	09:40
charon	cmn: you give it a line range and it finds all commits that touched this range, adjusting it as it goes back	09:40
cmn	so blame on steroids?	09:40
charon	it's like blame and log combined	09:40
	→ achadwick joined	09:40
shruggar	yeah, blame on steroids	09:40
tga	blame the log	09:40
	→ dr_win joined	09:41
shruggar	I told charon I would look at it and try to clean it up, turns out I forgot how bad I was at code-reading	09:41
	← Daniel left	09:42
	→ Daniel joined	09:42
	← maletor left	09:43
	← nixmaniack left	09:43
charon	cmn: here's what it looks like on git log -L'/do_askpass/,/^}/':prompt.c https://gist.github.com/6173cc1a84f747321a44	09:44
	shruggar: don't take so much blame, it's my fault for never working on it :(	09:44
	it's mostly okay, but in lots of little corner cases it segfaults	09:44
	→ FernandoBasso joined	09:45
	→ nixmaniack joined	09:45
cmn	cool	09:45
	← weiyang left	09:48
	→ tatsuya_o joined	09:48
	← xiangfu left	09:49
	← d3vic3 left	09:49
	← tqmz left	09:49
	→ tqmz joined	09:50
	← tqmz left	09:50
	← abetaha left	09:50
	→ tqmz joined	09:50
	→ mbroeker joined	09:51
	→ cvprog joined	09:51
shruggar	charon: don't worry, it's just added to the pile of git things I haven't finished :) I foolishly thought, "I know, I'll just try to clean up someone else's work! That'll keep me focused!"	09:53
	AAA_awright_ → AAA_awright	09:54
shruggar	I /could/ blame my hard drive crashing last time I was working on it, but really I'd only gotten as far as rebasing, which I'd need to do again anyway	09:55
	→ vervic_ joined	09:55
	← vervic_ left	09:56
	→ vervic_ joined	09:56
	← cvprog left	09:56
	ChanServ set mode: +v	09:57
	→ dr_lepper joined	09:57
Daniel	shruggar: Do you have any documentation on how it is intended to work?	09:57
	← dr_lepper left	09:57
	→ dr_lepper joined	09:57
	← ChaoWang left	09:57
	← vervic left	09:57
	vervic_ → vervic	09:57
	→ beautifulmind joined	09:58
	→ Mikachu joined	09:58
	← arvind_khadri left	09:59
	→ cvprog joined	09:59
	→ Vile joined	10:00
	← aspiers left	10:01
	← johnhamelink left	10:02
	→ jatt joined	10:02
	→ johnhamelink joined	10:03
	→ hyperair joined	10:04
	← dorkmafia left	10:04
	← dreiss left	10:05
	← tqmz left	10:05
	→ cncMaker joined	10:06
	→ tqmz joined	10:06
	gitinfo set mode: +v	10:07
cncMaker	hi im new to git i want to apply a patch to a clone	10:07
	← herdingcat left	10:07
_ikke_	cncMaker: What kind of patch?	10:07
charon	Daniel: log -L? it doesn't exist yet	10:07
cncMaker	i got the clone inside and the patch abc.patch	10:07
shruggar	"git apply"	10:08
	→ MadDogg joined	10:08
	← MadDogg left	10:08
cncMaker	"git apply abc.patch "	10:09
	shruggar: that way	10:09
	git --help has no command apply	10:11
	→ arvind_khadri joined	10:11
_ikke_	man git apply-patch	10:12
gitinfo	the git manpage is available at http://jk.gs/git.html	10:12
_ikke_	wait	10:12
	man git-apply	10:12
gitinfo	the git-apply manpage is available at http://jk.gs/git-apply.html	10:12
cncMaker	thanks	10:13
	← porco left	10:14
	← Milossh left	10:15
	← kryl99_ left	10:16
	→ smw joined	10:17
cncMaker	im folowing this http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Git and i am confused	10:17
	http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Git#from_patches from patches	10:18
	→ psakrii joined	10:18
frogonwheels	cncMaker: Which bit is confusing?	10:19
_ikke_	That's when you have the patches as e-mails	10:19
cncMaker	"git am --signoff patch.mbox" i got the colne and a file names "abc.patch"	10:19
	← forsto left	10:19
frogonwheels	cncMaker: 'apply' applies a _patch_ 'am' applies a commit sent as email (usually generated with format-patch)	10:20
	cncMaker: so if 'abc.patch' has extra header info with comitter and the comit message, then use 'am' otherwise, use 'apply' and commit it.	10:21
cncMaker	so in the main folder of the clone "git am abc.patch"	10:21
_ikke_	cncMaker: No	10:21
	cncMaker: git apply abc.patch	10:21
cncMaker	frogonwheels: i will paste the patch one moment	10:21
frogonwheels	_ikke_: !? it depends if abc.patch is a patch, or an format-patch	10:21
	cncMaker: try it and see. you can always roll it back.	10:22
cncMaker	frogonwheels: http://paste.pocoo.org/show/547202/	10:23
_ikke_	It's an e-mail	10:23
cncMaker	it came by a mail into a zip	10:24
frogonwheels	_ikke_: am should probably work, but you might not like the commit message	10:24
_ikke_	frogonwheels: You mean cncMaker probably	10:24
frogonwheels	_ikke_: sorry yeah. in a rush	10:25
cncMaker	B)	10:25
frogonwheels	cncMaker: just use git am as described	10:25
	cncMaker: suck-it-and-see. you can always roll it back (especially if you have a clean working dir to start with)	10:25
	→ noisebleed joined	10:25
	← noisebleed left	10:25
	→ noisebleed joined	10:25
	← smw left	10:25
frogonwheels	gl gg	10:26
cncMaker	my i-net is slow woudt not like to load 150MB once more	10:26
	→ ph^ joined	10:26
	→ d3vic3 joined	10:26
Daniel	charon: Yes I know. That's why I asked how it's intended to work (key word: intended). It's not a bad idea to know what you want to build before you build it. Writing up some documentation for others to read and give feedback on isn't a bad start.	10:26
cncMaker	thanks i try it out	10:26
	← cncMaker left	10:27
	→ faber joined	10:27
	→ ajt joined	10:29
charon	Daniel: the code exists, it's just rather rough. it basically runs through history updating a set of line ranges, assigning blame to commits as it goes and updating the ranges accordingly	10:30
	→ cortana2 joined	10:32
	← Skyx-mobile__ left	10:32
	← apok left	10:34
	→ iivvoo joined	10:35
	← Swathe left	10:35
	← d3vic3 left	10:36
	← h0bbit left	10:37
Daniel	charon: So it's like a blame but instead of looking at lines it looks at line ranges?	10:37
	→ Swathe joined	10:37
shruggar	it's like a repeated blame	10:38
	and instead of looking at a particular file, it looks at content	10:38
	very gittish	10:38
	← ilteris_ left	10:39
	→ Harzilein joined	10:39
	← Harzilein left	10:39
	→ Harzilein joined	10:39
	← Harzilein left	10:39
	→ Harzilein joined	10:39
	← Harzilein left	10:39
	→ Harzilein joined	10:39
	← Harzilein left	10:39
	→ Harzilein joined	10:39
	← Harzilein left	10:39
	→ Harzilein joined	10:39
	← Harzilein left	10:39
	→ kryl99_ joined	10:40
	→ h0bbit joined	10:42
	← tga left	10:42
Daniel	Something just dawned on me, when doing "gitk <filename>" I don't see changes in other files than the one I supplied. What I want is something that shows me all commits that have effected that partocular file, but I want to see everything in those commits.	10:42
	→ Skyx-mobile__ joined	10:43
	→ albel727 joined	10:44
cmn	gitk $(git log --oneline -- file) might do the trick	10:44
FauxFaux	Daniel: gitk $(git log --format=%h -- file) # bah!	10:44
cmn	indeed, what he said	10:44
jast	gitk $(git rev-list HEAD file)	10:46
	shorter! ;)	10:46
	← OOPMan left	10:46
cmn	but that doesn't work if you've a file named HEAD ;)	10:47
	→ brianpWins joined	10:47
Daniel	luckilly I don't =)	10:47
brianpWins	i think I accidentally made the master branch of a remote repo named something other then master. and now it won't let me delete it. is that possible ?	10:47
	← caseymcg left	10:48
cmn	renaming remote branches is hard, as you'd have to delete it on the old name	10:48
	→ whyking joined	10:48
Daniel	I'm still getting commits that haven't changed the file I'm interested in though.	10:48
cmn	or you mean that you pushed for the first time with some other name?	10:48
brianpWins	cmn: thats another issue, or maybe the initial issue. I deleted that branch locally	10:48
	cmn: the first push i did was git push origin demo1:demo1	10:49
cmn	deleting a branch locally doesn't affect he remote	10:49
brianpWins	then later i deleted locally and then tried to delete remote with: git push origin :demo1 but it says (deletion of the current branch prohibited)	10:49
cmn	so you pushed a local branch demo1 to the remote and named it demo1 there	10:49
	you're pushing to a non-bare remote	10:49
	probably, anyway, I don't see why a bare remote would forbit that	10:50
	fobid*	10:50
	→ X-tonic joined	10:50
cmn	the issue there is that you'd be deleting the branch to which the worktree belongs, and that's bad	10:50
	why are you pushing to a non-bare remote?	10:50
brianpWins	cmn: created the repo on github with the "new repo" button. the added it as origin and did my first push	10:51
	→ Cromulent joined	10:51
	← h0bbit left	10:51
cmn	ok, then what they mean is not "current" but "default" branch, I guess	10:51
	push the branch with the name you do want, change the default branch to be that one and then you should be able to delete demo1	10:52
brianpWins	cool. how do i change the default branch?	10:52
cmn	clickety click on the website	10:52
	it's on the admin part of the repo	10:53
	→ angelsl joined	10:53
Daniel	Well, I've seen the error in my thinking at least, I'll get back to you when I've read up on the subject some more.	10:54
	→ h0bbit joined	10:55
brianpWins	hmmm thats not working so well. i think i'll just blow away the remote repo and push it as new	10:56
	→ swombat joined	10:56
	← swombat_ left	10:57
	→ Ryback_ joined	10:57
Daniel	Now for my second question. It's regarding the renaming of files, which is quite common when refactoring code. So far it seems somewhat random how git handles this, sometimes it understands that the file has been renamed, sometimes it thinks the file has been deleted and a new one created. The 'fix' suggested for this is to do the actual rename in a separate commit and nothing else. But code needs	10:57
	to be changed to compile when moving files, other files need their imports changed and so on or else the code won't compile. How should one handle this?	10:57
	→ iocor joined	10:59
	→ Daniel_692385 joined	10:59
	← tiglog left	10:59
	ChanServ set mode: +v	11:00
	← gandy93 left	11:00
	→ swombat_ joined	11:00
	← swombat left	11:01
	← Daniel left	11:01
	← moremorefun left	11:01
shruggar	Daniel: you can try tweaking the "rename" threshold, or telling git to be more-thorough in rename detection	11:02
	→ txomon joined	11:02
	← xerora left	11:02
	← dc5ala left	11:03
	← swombat_ left	11:03
	→ swombat joined	11:04
cmn	if you're trying to figure out where some code originally came from, blame -C -C -- file might be more useful than following logs	11:04
	→ d0k joined	11:04
shruggar	indeed	11:04
	→ leavittx joined	11:04
shruggar	git doesn't track /file/ renames because /file/ renames are common but unimportant. Not granular-enough to be useful.	11:05
	← Matrixiumn left	11:05
	← ph^ left	11:05
cmn	it's like the web 2.0 of VCSs, it's the user-generated content that counts :)	11:06
	shruggar gives cmn the "lamest description of git ever" award	11:06
	← h0bbit left	11:06
	← txomon left	11:07
cmn	in that specific sense, there is more to it	11:07
	← chased_ left	11:07
	→ babilen joined	11:07
	→ twisted` joined	11:08
	← Daniel_692385 left	11:09
	→ Daniel joined	11:09
	ChanServ set mode: +v	11:09
Daniel	True that, but when reviewing changes in gerrit it's much easier when it interprets changes like that as renames rather than delete and add. Although as I understand Gerrit has its own sneaky way of enterpreting git.	11:09
	→ h0bbit joined	11:09
	→ gandy93 joined	11:11
	← zdk left	11:12
	← break57382 left	11:12
cmn	there's a difference between detecting partial renames and showing a diff	11:12
	a diff are the literal changes you have to do on a set of files	11:13
	← LotR left	11:13
	→ ph^ joined	11:13
	→ LotR joined	11:14
Daniel	cmn: Yes, and those are the changes I'm interested in. But to see the diff i Gerrit it needs to understand that the file has been renamed, otherwise the only 'diff' you see is a file deletion and a new file added, you can't see the changes between the two. Granted, it's more of a Gerrit problem, but I figure that if Git can't figure it out properly then what chance does Gerrit stand.	11:15
	→ OOPMan joined	11:16
	→ dc5ala joined	11:16
	→ smw joined	11:16
cmn	that's exactly what a diff is	11:17
	that deletion and addition are the literal changes you need to do	11:17
	if it's a partial rename, the UI should show it as that for easier reviewing	11:17
Daniel	cmn: If that's how git should look at it then nothing should ever be a rename?	11:18
cmn	that's how git sees things, yes	11:18
	git doesn't see the diffs either really	11:18
Daniel	cmn: So how come git shows things as a rename sometimes when they have been moved? Isn't that a bug in that case?	11:18
cmn	git sees two states of a worktree	11:18
	bug?	11:19
	← gnufied left	11:19
	→ berserkr joined	11:19
shruggar	git infers renames based on how much of the content is the same. If git doesn't think it's a rename, it may not actually be a rename, regardless of what command you used to tell git about the change.	11:20
	that's the point	11:20
Daniel	cmn: You said that git doesn't understand renames, but when I do a "git status" and have such changes in my staging area, it sometimes states rename and sometimes it states add/delete	11:20
	← h0bbit left	11:20
cmn	git doesn't track renames	11:20
	I never said it didn't understand them	11:20
	all of that information is generated on the fly according to heuristics	11:21
	if the file is small enough of different enough, it won't consider it as a rename or copy	11:21
	→ dorisabayon joined	11:21
Daniel	cmn: Ok, so how do I change the heuristics?	11:21
	→ gnufied joined	11:21
	→ dharrigan joined	11:22
	← dharrigan left	11:22
	→ dharrigan joined	11:22
cmn	you hack on git	11:22
IcePic	Daniel: what would be the end-result of that?	11:22
cmn	if you want to change the level at which git will say something is a rename or copy, you tell it	11:22
	look at the -C and -M options in log	11:22
	and related commands	11:22
Daniel	IcePic: A faster way of seeing what has changed by looking at my staging area for one.	11:24
	→ h0bbit joined	11:24
	← Fandekasp left	11:25
	dorisabayon → Fandekasp	11:26
	← smw left	11:28
	← gnufied left	11:29
	→ gnufied joined	11:29
Daniel	cmn: But that wouldn't effect how my staged changes are displayed in a git status, how would I tweak that?	11:31
IcePic	Daniel: but would it be a problem if the diff says "1000 lines removed from file Source, 1000 lines added to file Destination" as opposed to "renamed the 1000-line file from Source to Destination" ? I appreciate that one would be shorter, but the end result would be the same	11:31
	← nixmaniack left	11:31
	→ murr5y joined	11:33
	← murr4y left	11:33
	← beautifulmind left	11:34
cmn	man git config has some options, though I can't see status in there	11:34
gitinfo	the git-config manpage is available at http://jk.gs/git-config.html	11:34
	← nimrod10 left	11:34
cmn	don't have time to look at the source now	11:34
	→ EricInBNE joined	11:34
	→ nimrod10 joined	11:34
	← h0bbit left	11:35
	← jonesy left	11:37
	← dr_win left	11:37
	← koo5 left	11:37
	← murr5y left	11:39
	→ murr4y joined	11:39
	← twisted` left	11:40
	→ twisted` joined	11:40
Altreus	Is there a quick way of getting the root of the workin gtree?	11:42
	→ Milossh joined	11:42
jast	Altreus: git rev-parse --show-toplevel	11:43
	→ gavin_hu_ joined	11:46
	← gavin_huang left	11:47
Altreus	:) thanks	11:48
	← toabctl left	11:48
	← eno left	11:48
	→ eno joined	11:50
	→ aspiers joined	11:50
	→ Demego joined	11:50
	→ ajaykools joined	11:53
	← vpopov left	11:54
	→ h0bbit joined	11:56
	← paphio left	11:57
	← teweWORK left	11:57
	→ Dave^\|\| joined	11:59
	← iocor left	11:59
	← Samoi left	12:00
	→ toabctl joined	12:02
	→ vpopov joined	12:02
	← Dave^\| left	12:02
	→ juvenal joined	12:04
	→ dorisabayon joined	12:04
Daniel	IcePic: Well, if only 40 lines were actually changes then I'd only want to see those 40 lines as changed, regardless of if the file was moved or not. But if Gerrit gets it wrong it doesn't show me those 40 lines, but rather both files in their interity being changed, which makes it very difficult to do a proper code review.	12:05
	← h0bbit left	12:05
Daniel	cmn: No worries mate, thanks for your help.	12:06
	← psakrii left	12:08
	→ h0bbit joined	12:08
	← Fandekasp left	12:09
	← krz left	12:09
	← harshpb left	12:09
	← vpopov left	12:10
	→ JeanMertz joined	12:11
	→ Guest45348 joined	12:11
	← dpino left	12:12
	← Guest45348 left	12:12
cespare	Is there a way to get git to tell me where the gitdir is? It may, of course, be at `git rev-parse --show-toplevel`/.git, but it might be somewhere else if $GIT_DIR is set or if using the gitfile format (wherein .git points to a gitdir elsewhere)	12:14
	→ brian_g joined	12:16
	→ smw joined	12:16
	← robotmay left	12:16
	← eno left	12:16
	→ robotmay joined	12:16
cortana2	cespare, git rev-parse --git-dir	12:17
wereHamster	cespare: git rev-parse --git-dir ?	12:17
cespare	it was neck and neck guys, very nice	12:17
cmn	cespare: the man git rev-parse manpage is full of wonderful options	12:17
gitinfo	cespare: the git-rev-parse manpage is available at http://jk.gs/git-rev-parse.html	12:17
cespare	cmn: awesome, thanks	12:18
	→ eno joined	12:18
	→ shiba_yu36 joined	12:18
	→ revagomes joined	12:18
	← johnathanb left	12:19
	gitinfo set mode: +v	12:19
JeanMertz	Hey guys, I created a new empty branch using http://book.git-scm.com/5_creating_new_empty_branches.html	12:19
	However, the stupid person that I am, I hadn't commited all changes yet, and they seem to be gone now. Is there any way to get them back?	12:19
	→ johnathanb joined	12:19
cmn	seem to be gone from where?	12:20
	what did you do to lose the changes?	12:20
	→ sattu94 joined	12:20
	← h0bbit left	12:20
	← redrocket left	12:20
	← revagomes left	12:21
JeanMertz	cmn: I was on the master branch, I had uncommited changes, I then used the example from the link above to create a new "clean" branch, (detached from HEAD). This seemed to work, but when I switched back to master, all uncommited changes where gone	12:21
	← sattu94 left	12:21
JeanMertz	cmn: git clean -fdx seemed to start deleting all files	12:22
wereHamster	JeanMertz: that's what it does	12:22
cmn	gah, those instructions are very old school	12:22
	→ sattu94 joined	12:22
wereHamster	you should have committed your changes. Now they are gone.	12:22
	← sattu94 left	12:22
	→ h0bbit joined	12:22
wereHamster	and since you never told git about them, it can't restore them for you.	12:23
JeanMertz	Yeah, stupid of me to fire that command I know. I already figured I lost the files. But then I went ahead and checked the .git directory, and it's 1,5GB, which indicates to me that the files are somehow still there?	12:23
	→ revagomes joined	12:23
cmn	not the unstaged changes	12:23
wereHamster	JeanMertz: did you ever 'git add' them?	12:23
shruggar	Unix rule: be wary of anything with a -f in it	12:23
	→ sattu94 joined	12:23
cmn	and take backups before you do anything the Internet tells you to	12:23
JeanMertz	wereHamster: I can't remember, but I do know that all files without the files I lost where only a few MB.	12:24
wereHamster	JeanMertz: then good luck with git fsck	12:24
cmn	that'd be one version only	12:24
JeanMertz	I agree with you all, I am to blame for this, and I should never do such things when you just woke up, just wondering if there was any last hope for me, heh.	12:24
cmn	git fsck --unreachable will give you objects, among them will be the staged changes	12:25
	unstaged changes won't exist anywhere in .git	12:25
ndim	Going forward... is there any new, clean, robust, user-proof way to create an empty parent-less branch (replacing those hardcore "rm -f .git/index" etc instructions)?	12:26
	← bjoernh left	12:26
cmn	checkout --orphan	12:26
	← eno left	12:26
	→ canton7 joined	12:26
wereHamster	cmn: that only replaces two of the commands	12:26
JeanMertz	cmn: Thanks, so now I get a huge list of unreachable blobs, time to google what to do with them	12:26
	← techstein left	12:26
	← smw left	12:26
	→ dpino joined	12:26
wereHamster	you still need git clean	12:26
cmn	depending on what you really want to do yes	12:27
wereHamster	JeanMertz: git show them, store them in appropriate files	12:27
	← Chappo left	12:27
	→ eno joined	12:27
	← eno left	12:27
	→ eno joined	12:27
cmn	if you want a branch that's really different git rm --ignore-unmatched * should be a good replacement for rm .git/index	12:28
JeanMertz	wereHamster: ouch, any way to find out what their original name where? if I git show them I get a huge blob of binary data, because most of what I lost where images and/or zip files	12:28
wereHamster	no	12:28
JeanMertz	wereHamster: okay thanks, anything else I can do with them besides git show? restore? open in image viewer/zip file to check what it is?	12:29
cmn	ndim: there's no great way to create a different project inside a repo because that's not what you're meant to do	12:29
wereHamster	JeanMertz: there is nothing more that git can assist you with	12:29
ndim	cmn: Except for github web pages.	12:29
	And other cases.	12:29
cmn	which is a github thing, and not a git thing	12:29
	← scottj left	12:29
	← dim left	12:30
cmn	and it's a one-time thing at any rate	12:30
	← ajaykools left	12:30
	→ DarkAR joined	12:30
JeanMertz	wereHamster: Okay, thank you for the help all.	12:31
	← soc42 left	12:31
cmn	creating a detached history is something very project- or provider-specific, not something you do on your usual workflow	12:32
	→ L42y joined	12:32
	← h0bbit left	12:34
shruggar	is there a way to tell git "checkout all mode changes, leave code changes as they are" ?	12:34
	← hh__ left	12:35
	→ bauruine joined	12:35
	→ jonesy joined	12:37
	← Daniel left	12:37
	→ h0bbit joined	12:37
	← LeMike left	12:40
canton7	shruggar, there might be a way using core.filemode... Set it true, checkout the files with the wanted permissions, set, it false, set out the files with the wanted content...	12:40
	s/set out/checkout	12:41
	← eno left	12:41
cmn	ls-tree -r \| cut \| xargs chmod sounds easier than that	12:43
	→ redrocket joined	12:43
	→ eno joined	12:43
JeanMertz	Does any of you know how I can iterate over each file using "git fsck --unreachable" and do "git show blob hash here > blob hash here.png"?	12:43
cmn	or I might be underestimanting the complexity	12:43
	for obj in $(git fsck --unreachable \| cut -f 2); do git cat-file -p $obj > $obj.png; done or close to that	12:44
wereHamster	git fsck --unreachable \| while read something sha; do git show $sha > $sha; end	12:45
cmn	that might be -f 3, actually; and you'd want to grep for blob, I guess	12:45
	← canton7 left	12:45
cmn	the format is "unreachable $type $hash", so you might need an extra arg there, I think	12:46
	cmn should try to use read more often instead of trying to shove cut everywhere	12:47
	→ ko1 joined	12:48
	→ k1o joined	12:48
JeanMertz	Hmm wereHamster it doesn't fire (with extra arg), I have to ctrl+c out of it, but I can't see if there's a missing end/semicolon somewhere?	12:48
	← h0bbit left	12:49
	← Cromulent left	12:49
wereHamster	done instead of end	12:49
	→ d3vic3 joined	12:50
	→ erichynds joined	12:51
	→ h0bbit joined	12:51
JeanMertz	Holy crap, wereHamster and cmn you two rock! :) You just helped me restore 1,5GB of photo's/pixelmator files. Now I still need to manually change the extension of all the files that aren't png's, but that shouldn't be too much work. Thank you very much guys. Now I'm off to add a sticky-note to my screen "Never do anything fancy in git before you Commit!"	12:52
cmn	shouldn't most of those files be in that other branch?	12:53
	→ dnjaramba joined	12:53
	→ cannonball joined	12:53
JeanMertz	It was a small job for a client, so I got lazy and never comitted anything after the initial commit. And then they came to me and said "we want something else", so I figured I'd create a quick headless branch to show off some stuff, which in return destroyed all the old files, which would've been fine if I hadn't been storing all event photo's in that directory.	12:54
	So it doesn't matter where they are now, I'm moving them to a safe location as we speak.. type	12:55
	← dc5ala left	12:55
	← dv310p3r left	12:55
cmn	"headless branch"... you mean parentless branch or something	12:56
JeanMertz	yes I do	12:57
	→ uu joined	12:57
	← dnjaramba left	12:57
JeanMertz	In my git GUI it said "detached from HEAD", so that's where my "headless branch" is coming from, heh	12:57
	← eno left	12:57
cmn	that sentence is wrong	12:58
	→ dnjaramba joined	12:58
cmn	unless you dereference HEAD in your mind	12:58
JeanMertz	Hmmm. It's not showing up anymore so not sure what it said exactly, but I'm pretty sure "detached from HEAD" was in there.	12:59
	→ eno joined	12:59
cmn	it might have said "detached HEAD"	12:59
	that would be right	12:59
	← bartman left	13:02
	→ chrisdev joined	13:03
	← eno left	13:06
	← h0bbit left	13:07
	→ eno joined	13:07
	tizzo-afk → tizzo	13:07
	→ crisp joined	13:08
	→ iocor joined	13:08
	→ beautifulmind joined	13:10
	← uu left	13:11
	→ bartman joined	13:11
	← alta189 left	13:13
	← Dreamer3__ left	13:14
	→ h0bbit joined	13:14
	← pidus left	13:15
	← cirbri left	13:15
	→ johnkpaul joined	13:16
	→ cccaldas joined	13:17
	→ digitalfiz joined	13:17
	→ pidus joined	13:17
	← digitalfiz left	13:18
	→ digitalfiz joined	13:18
	← Swathe left	13:18
	← Kulrak left	13:19
	→ guampa joined	13:19
	← dougsland left	13:19
	← digitalfiz left	13:19
	→ Swathe joined	13:20
	→ dougsland joined	13:20
	← johnkpaul left	13:22
	→ teweWORK joined	13:22
	→ isaak joined	13:23
	→ digitalfiz joined	13:23
	← eno left	13:25
	← pidus left	13:25
	← batlock666 left	13:26
	→ eno joined	13:26
	→ batlock666 joined	13:26
	→ Splash joined	13:29
	→ Bass10 joined	13:29
	← tatsuya_o left	13:30
	← revagomes left	13:30
	→ revagomes joined	13:30
	← h0bbit left	13:30
	→ h0bbit joined	13:32
	← eno left	13:33
	← OOPMan left	13:33
	→ neurodrone joined	13:34
	→ yshh joined	13:34
	→ eno joined	13:34
	← dorisabayon left	13:35
	→ jonatasnona joined	13:36
	→ Fandekasp joined	13:36
	← jonatasnona left	13:37
	→ t4nkd joined	13:38
	→ jonatasnona joined	13:38
	→ Meme joined	13:38
	← jonatasnona left	13:38
	→ jonatasnona joined	13:38
	← h0bbit left	13:38
	← jonatasnona left	13:41
	→ jonatasnona joined	13:41
	← iocor left	13:41
	← paulreinheimer left	13:42
	← Meme left	13:42
	→ andrewa2 joined	13:42
	→ QaDeS joined	13:43
	← fruight left	13:43
	← kryl99_ left	13:43
	← DarkAR left	13:43
	← jonatasnona left	13:43
	→ jonatasnona joined	13:44
	← jonatasnona left	13:44
	→ jonatasnona joined	13:44
	← jonatasnona left	13:45
	→ gjer984 joined	13:45
	← ilius left	13:45
	→ jonatasnona joined	13:45
	← revagomes left	13:45
	← AmirBehzad left	13:46
	→ revagomes joined	13:46
	← eno left	13:47
	ChanServ set mode: +v	13:48
	→ eno joined	13:48
	→ iamjarvo joined	13:49
	→ kadoban joined	13:50
	→ h0bbit joined	13:51
	→ oneiroi joined	13:51
oneiroi	Having to support a couple legacy svn repo's wanting to use working copy checkouts using git-svn seeing this: Read access denied for root of edit when cloning using a read-only user, any ideas?	13:52
	← revagomes left	13:53
	→ revagomes joined	13:53
	← joshsmith left	13:53
FauxFaux	oneiroi: Just clone -r HEAD if you don't need the history; much faster and less broken.	13:54
	→ krz joined	13:55
oneiroi	FauxFaux: thanks giving that a try now, sooner I can get shot of these legacy repos the better imo	13:55
	→ dennis\|osw joined	13:55
	→ iocor joined	13:56
	gitinfo set mode: +v	13:56
dennis\|osw	hey	13:56
	→ jose2 joined	13:56
	→ fruight joined	13:57
	→ stuntmachine joined	13:57
	← keller left	13:57
dennis\|osw	How would you go about version controlling a mysql database?	13:58
cmn	depending on what you want to do, the answer may be "you don't, you take backups"	13:59
dennis\|osw	i guess structural changes should be followed always, but content only in some tables.	13:59
FauxFaux	dennis\|osw: http://sourceforge.net/projects/dbdeploy/ or whatever is built into your platform.	13:59
	← jacekmigacz left	13:59
cmn	you should not track content	13:59
	that's what the database is for	13:59
	→ keller joined	13:59
	← rchavik left	13:59
SethRobertson	/clear	14:00
dennis\|osw	well, i have config settings in my software. If i change them in a dev stage, i want to know about it	14:00
cmn	so track the db schema and the code using it	14:00
	→ Guest95996 joined	14:00
dennis\|osw	The config settings is db content. Also, i use dynamic "labels" if you will that may change during development	14:01
cmn	what?	14:02
	→ ilius joined	14:02
	← eno left	14:02
cmn	that sounds like a deployment/bootstrap issue	14:02
dennis\|osw	There are certain settings that are stored in a database table. That's content, not structure	14:02
	cmn: not quite sure what you mean by that?	14:02
cmn	ok, so it's a deployment issue; make your deployment script make sure they're there	14:02
	← ilius left	14:03
	→ eno joined	14:03
	→ AaronMT joined	14:04
	← h0bbit left	14:04
cmn	I mean that those config settings sound like stuff you need to get into the DB for the app to work correctly	14:04
	← ipalaus left	14:04
	→ Jackneill joined	14:04
cmn	which is a deployment issue	14:04
dennis\|osw	Ah, no not really... software behaviour can vary depending on the value of those settings.	14:04
	For example, the database will contain information wether or not a certain module is active	14:05
	or, certain module configuation values	14:05
	besides, also the entire content is dynamic, so database content, that can be altered during development.	14:06
cmn	so they're setting for the actuall app?	14:06
	→ h0bbit joined	14:06
dennis\|osw	yes	14:06
	it's a cms	14:06
	maybe i should have made that clear earlier, to avoid confusion	14:06
	if so, sorry :)	14:07
	← jonesy left	14:07
	→ alta189 joined	14:07
	← jose2 left	14:07
	→ bartek joined	14:07
	→ fpighi joined	14:08
	→ josemota joined	14:08
	← Guest95996 left	14:09
	← dpino left	14:09
	→ vpopov joined	14:09
dennis\|osw	Most common online answer i read is dump the db in a sql file and track that, but that's not really what i want	14:10
fpighi	does anyone know how I could do:	14:10
	git log TAG..HEAD	14:10
	and force the rev list selector to start from the first commit if TAG is not in the repo?	14:10
cmn	why do you change application settings in development?	14:11
	← cvprog left	14:11
	← vpopov left	14:11
	← ajt left	14:12
	→ heptadecagram joined	14:12
thiago	fpighi: impossible	14:12
	fpighi: if you tell it a commit, it will start from that commit	14:12
	fpighi: if you tell it some garbage that it cannot resolve to a commit, it cannot continue	14:12
	→ alan_w joined	14:13
	→ ajt joined	14:13
fpighi	thanks thiago	14:13
	I'll have to do it in another way then	14:13
L42y	exit	14:13
dennis\|osw	cmn: it can contain module activation, it can contain certain module settings, etc. Besides, the database also contains textual values, etc.	14:13
	← L42y left	14:14
fpighi	I'm trying to invoke git bundle lastbackup..HEAD, but the first time it executes I will have to check if the tag exists first. Unless someone knows a one-line way of doing it	14:14
	← nrichards left	14:14
cmn	textual values for what? it sounds like you're mixing development choices with user preferences	14:15
	← teweWORK left	14:15
	→ dfr\|mac joined	14:16
	→ lyddonb joined	14:18
	→ dv310p3r joined	14:19
	← Lgb left	14:20
	← eno left	14:20
	← h0bbit left	14:20
	→ tatsuya_o joined	14:20
	→ codewrangler joined	14:21
	← Milossh left	14:21
	← tatsuya_o left	14:21
	→ pepr joined	14:21
	→ tatsuya_o joined	14:21
	→ eno joined	14:21
	→ Lgb joined	14:21
	ChanServ set mode: +v	14:22
dennis\|osw	cmn:wn	14:22
	sorry, premature enter ;)	14:23
	→ preinheimer joined	14:24
dennis\|osw	cmn: Most simple example I can think off... customer wants a custom banner that he can edit in the backend. We create a custom label that this customer can edit. A label is a value in a language label table.	14:24
josemota	/part/wc	14:25
dennis\|osw	This example is not config setting, but will do for my example on why content tracking will be required in my case (as far as i think)	14:25
pepr	Hi, I do have a lot of tiny examples in subdirectories -- one example, one subdirectory of the ..../tutorial/. I would like to keep them all in the tutorial/ directory, but treat the changes separated for each of the examples with respect to the git. But I would also like to find quickly what example was changed. Something like a common git log for subdirectories. What approach do you recommend?	14:25
	→ glcrazy joined	14:25
	→ h0bbit joined	14:25
	→ gchristensen joined	14:26
	→ ephelion joined	14:26
	← maleknet left	14:26
	← keller left	14:26
EugeneKay	pepr: man git-log, it supports a path argumet	14:27
gitinfo	pepr: the git-log manpage is available at http://jk.gs/git-log.html	14:27
	← cooloney_ left	14:27
	→ keller joined	14:27
	→ isaak_ joined	14:28
	→ oriba joined	14:28
	→ tommcdo joined	14:29
	→ sudoquai joined	14:30
pepr	EugeneKay & gitinfo: OK, I probably did not describe it correctly. The first goal is how to keep many subdirectories in the tutorial/. Should each subdirectory have its own git repository?	14:30
	← whyking left	14:30
EugeneKay	I don't see a reason to - git supports as many subdirectory levels as your filesystem can handle	14:31
	← isaak left	14:31
	← revagomes left	14:32
	→ ceej joined	14:32
pepr	EugeneKay: I already have about 100 subdirs. When updating one example, the commit is done for all of them. Is it OK?	14:33
EugeneKay	Sure - git commits happen at the repo level. Commits which dont affect a certain subdir just.... don't affect a certain subdir.	14:34
	→ revagomes joined	14:34
	→ ipalaus joined	14:34
cmn	dennis\|osw: it still sounds like something you need to do during deployment; if customer A wants one picture and customer B another one and you need to put that into the database, a config file for each customer with those values (in directories or in branches) and a deployment script that knows where to set each value is what you're looking for	14:34
EugeneKay	Unless you're running into performance issues you should just leave it as it sits. I don't think splitting your repo up will gain you anything	14:35
cmn	keeping your DB in version control isn't going to make things any better	14:35
EugeneKay	And splitting it would more than likely exacerbate performance problems, not solve.	14:35
	→ mick_home_ joined	14:35
	← albel727 left	14:36
	← h0bbit left	14:36
mick_home_	hi everyone, I'm trying to just have one branch (one person makes updates on it and I merely pull their changes). I would like both of us to be working off "master"	14:36
jonatasnona	keeping the DB in version control is not a good idea	14:36
	trust me	14:37
pepr	EugeneKay: Then, I should probably prefix my commit messages with the subdirectory name to be able to recognize the changes in the log. Just some extra discipline, right?	14:37
mick_home_	somehow when I did a sync - I see: "master" and "origin/master"	14:37
	→ spaced0g joined	14:37
SethRobertson	Keeping the baseline of a DB under version control is good. Keeping the tactical changes is not	14:37
	→ teweWORK joined	14:37
EugeneKay	Sure. Or rely upon log's path sensitivity ;-)	14:37
	→ TeckniX joined	14:37
_ikke_	mick_home_: origin_master is a so called remote tracking branch	14:37
mick_home_	SethRobertson: makes life hard when your config is in the db though :-/	14:38
	_ikke_: I have one person checking into some branch (no idea where) and I haven't been able to pull yet	14:38
	← eno left	14:39
mick_home_	when I use gitk to see what is up I see: o master - remotes/origin/master	14:39
	← vdv left	14:40
mick_home_	any ideas	14:40
	?	14:40
cmn	on what exactly? that seems fine	14:40
	→ albel727 joined	14:40
	← dharrigan left	14:40
cmn	you seem confused about decentralised VCSs	14:40
mick_home_	cmn: I can't pull their changes	14:40
cmn	what error do you get?	14:40
	→ h0bbit joined	14:40
	→ eno joined	14:41
	→ dharrigan joined	14:41
	← dharrigan left	14:41
	→ dharrigan joined	14:41
mick_home_	cmn: could be, I've used: sccs, rcs, cvs and svn	14:41
cmn	git works differently from them	14:41
mick_home_	so this is still rather new to me	14:41
cmn	what error do you get when you try to pull?	14:41
mick_home_	the only stuff that I've done in git is stuff that I just use for myself	14:41
	← glcrazy left	14:42
mick_home_	nothing, I just get an older version - now they can't "push" - it hangs for them	14:42
cmn	so the problem isn't pulling but pushing?	14:42
thiago	what is the last thing they see before it hangs?	14:42
mick_home_	tortoisegit on their side, I'm on linux	14:42
thiago	tell them to use the shell	14:43
	→ MariusKarthaus joined	14:43
	→ axl_ joined	14:43
mick_home_	ok, I guess I'll give it a go	14:43
	→ Kulrak joined	14:43
	← Kulrak left	14:43
	→ Kulrak joined	14:43
cmn	are seeing it hang or is tortoisegit simply not reporting progress?	14:43
mick_home_	I guess the workflow would be: git commit -m "the commit message"	14:44
	then: git push	14:44
	or do I need to do: git push master	14:44
dennis\|osw	cmn: ok so you say you would create a database patch that you deploy seperate from committing files?	14:44
mick_home_	cmn: tortoisegit sits for days	14:44
	and I can't pull any updates	14:45
jonatasnona	mick_home_, try git push origin master	14:45
cmn	dennis\|osw: not a patch, if you need a specific value in a field, then that's what you should write to it	14:45
	→ ShadesofGraylin\| joined	14:45
mick_home_	ok, I'll try that	14:45
	thanks guys	14:45
cmn	mick_home_: if there are no updates in the common repo, then there's nothing for you to pull, that's not the issue	14:45
dennis\|osw	cmn: not sure how you would go about that, considering it could be 100 values that may have changed...	14:46
mick_home_	she has made a lot of changes	14:46
	that I don't see	14:46
cmn	mick_home_: you just said she hadn't managed to push	14:46
mick_home_	I see them on her end and they are committed	14:46
cmn	so that's on her end	14:46
mick_home_	right, pushing hangs for days	14:46
	yup	14:46
cmn	until the push completes, there's nothing new in the remote repo	14:47
mick_home_	I need to wait for her to get back - sigh	14:47
	is there another UI for a windows person that someone would recommend	14:47
cmn	there's egit if they use eclipse	14:47
	and gitk and git-gui both work under msysgit	14:48
mick_home_	it is mainly for selenium tests	14:48
	← DofLeFou left	14:48
	→ DofLeFou joined	14:48
mick_home_	tortoisegit seemed like a good idea	14:48
	:)	14:48
	→ kevwil joined	14:49
	→ anjen joined	14:49
cmn	dennis\|osw: are those configuration values you need to care about?	14:49
pepr	mick_home_: I second for the Git GUI and gitk from the msysgit. You can also freely mix them with console commands that are not supported graphically.	14:49
	← freimatz left	14:49
cmn	it sounds like those are user preferences, which belong in backups	14:50
mick_home_	cmn: since you are talking about dbs and version control - how would you track things like drupal (all of the development config is stored in the db)	14:50
	→ joshsmith joined	14:51
mick_home_	module development is tied to db updates	14:51
	← iocor left	14:51
	← h0bbit left	14:51
cmn	track what parts?	14:51
mick_home_	I've seen a few suggestions and all seem to have drawbacks	14:51
pepr	mick_home_: I did use cvs earlier and also tried Tortoise for the others. In my opinion, tortoise makes it "too easy" and it obscures the essential procedures.	14:51
	→ ajaykools joined	14:51
	← rurufufuss left	14:51
cmn	do you want to track your choices or do you want to develop drupal?	14:51
mick_home_	cmn: I've done some drupal development in the past (modules)	14:52
	→ beatak joined	14:52
mick_home_	I don't currently though - I just always had issues w/ figuring out the best way to have the db also in a decent state	14:53
	← DofLeFou left	14:53
	→ john2496 joined	14:53
mick_home_	I used to use svn hooks to do mysqldumps before commits	14:53
	→ DerekPioneer joined	14:53
DerekPioneer	i have my main working directory at site.com with .git inside of it	14:54
	← dnjaramba left	14:54
DerekPioneer	i created a subdomain at dev.site.com	14:54
	i then used ln -s to link the .git over there	14:54
	how do i then make a new branch that tracks the master	14:54
	← Fandekasp left	14:54
cmn	you might be looking for the new-workdir script	14:55
	← frogonwheels left	14:55
mick_home_	is there a similar thing to svn hooks in git?	14:55
cmn	mick_home_: what db updates do you have to do in order to do development?	14:55
	← FernandoBasso left	14:55
cmn	yes, there are hooks	14:55
	→ h0bbit joined	14:56
	← revagomes left	14:56
mick_home_	making sure certain configuration parameters are set	14:56
cmn	to anything in particular?	14:57
	← kermit left	14:57
	→ diegoviola joined	14:57
mick_home_	can't really think of exact things (I'm not doing anything related at the moment - so it is difficult to think about that atm)	14:57
	→ v0n joined	14:57
dennis\|osw	cmn: it would enhold added & changed module config settings, added modules, added & changed language labels & label values, added & changed content, etc.	14:58
mick_home_	it isn't important anyhow, I was just curious	14:58
	← demet8 left	14:58
mick_home_	ya	14:58
	basically	14:58
cmn	content?	14:58
	for module dev?	14:58
	→ jacekmigacz joined	14:58
	← dougsland left	14:59
EugeneKay	pepr - no PMs, please, and !thanks ;-)	14:59
gitinfo	Feeling thankful? Type "ExampleUser++", and ExampleUser will score karma points at http://carmivore.com (our preferred way to objectify self-worth). There's really no point to thanking me – instead, why not thank the person who made me help you?	14:59
mick_home_	you do module dev for sites, that in turn need to be under some type of control	14:59
	→ silv3r_m00n joined	14:59
silv3r_m00n	what do these commands do https://gist.github.com/1760096 ?	14:59
	especially git symbolic-ref HEAD refs/heads/gh-pages ?	14:59
	→ dougsland joined	15:00
	← eno left	15:00
cmn	but my question is why you need to keep the site's content in your development environment	15:00
EugeneKay	cmn - QA? :-p	15:00
	→ sjefen6 joined	15:00
cmn	that's testing, though	15:00
	if you need real data, plug the real db in	15:01
	code and content are different things	15:01
EugeneKay	Tell that to the guy who built a wiki on git	15:02
pepr	EugeneKay: (this was because I am learning the tool for IRC :)	15:02
mick_home_	just you do realize that when configuration changes between dev and prod (think things like js minification) - there are differences (like that neat bug that I found in IE that limits the amount of css files that you can have)	15:02
	→ eno joined	15:02
mick_home_	those changes are stored in the db	15:02
EugeneKay	!config	15:02
gitinfo	[!configfiles] the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.	15:02
	→ iocor joined	15:02
mick_home_	module preferences are stored in the db	15:03
	ya, which is why there are a bunch of drupal modules that I see as pretty hack-ish that do that	15:03
	→ Vortex34 joined	15:03
mick_home_	namely "features" can be abused to do that	15:03
	but god forbid that you foget to checkin a change	15:04
	everything breaks and you get an impedince mismatch	15:04
	→ marlun joined	15:05
mick_home_	EugeneKay: for web development who only has things stored in a config file? nobody that I know	15:05
	← snearch left	15:05
mick_home_	except for db params and a few other things	15:05
dennis\|osw	cmn: why can't content not be part of a development, if you don't want all things to be hardcoded, but lateron changable by the end user?	15:06
	can't not? ehr... you get what i mean ;)	15:06
	← beautifulmind left	15:06
	← h0bbit left	15:06
	← ephelion left	15:07
marlun	If I just made a "git merge branch" and git said "Updating <sha1>...<sha2>" and I want to undo the merge, can I do "git reset --hard <sha1>"?	15:07
cmn	if the content is tracked alongside the content, it /is/ hardcoded	15:07
	→ ISF joined	15:07
	→ zweb joined	15:07
	← gchristensen left	15:08
dennis\|osw	If you develop a module with a form, the form has labels. Maybe the client will want to rename the values later, or translate it later. That means that you need dynamic variables stored in a database. If you add things here, it is content, yes...	15:08
	→ dpino joined	15:08
	← silv3r_m00n left	15:08
wald0	how i can modify an existing commit (without create a new one) ? im reading about rebase but i dont quite get it (it creates a new commit)	15:08
	← pepr left	15:08
	← ISF left	15:08
	→ revagomes joined	15:09
dennis\|osw	but this content is part of and required for the development of this module.. and the current working copy doesn't have these values to work with..	15:09
	sorry not working copy, master version...	15:09
	→ ISF joined	15:09
	← eno left	15:10
	→ h0bbit joined	15:10
dennis\|osw	cmn: not sure what you mean by "if tracked, it is hardcoded"	15:11
cmn	each commit is hardcoded	15:11
dennis\|osw	depends on your definition of hardcoded ofcourse	15:11
mick_home_	by that definition - everything is hardcoded ;)	15:11
cmn	so if you deploy things as you have them in a particular commit, your content would get rolled back if there'd been updates	15:11
	mick_home_: huh?	15:12
	← zweb left	15:12
dennis\|osw	i think of hardcoded as not changable in a GUI but need to open script files	15:12
mick_home_	I see your point	15:12
	← neurodrone left	15:12
	→ eno joined	15:12
	← eno left	15:12
	→ eno joined	15:12
cmn	dennis\|osw: what does it matter where you have to change it?	15:12
dennis\|osw	cmn: if you roll back the entire module, there will be no need for the content either	15:13
cmn	entire module?	15:13
dennis\|osw	argh :)	15:13
	or whatever you've changed...	15:14
	→ glennpratt joined	15:14
	← ShadesofGraylin\| left	15:14
	→ mortenvp joined	15:15
cmn	what is this content you're speaking off if it disappears with the module?	15:15
	→ neurodrone joined	15:15
cmn	when you say content in a CMS, I understand the stuff people have put on it	15:15
dennis\|osw	that was related to my example of implementing a module with a form that contained language labels	15:15
	→ dnjaramba joined	15:16
	→ glcrazy joined	15:16
cmn	and what would happen if you need to rollback the latest commit because you made a change?	15:16
	what happens to the changes people have made since you last deployed?	15:16
dennis\|osw	isn't a change a new commit?	15:16
	→ johnkpaul joined	15:16
	← neurodrone left	15:16
dennis\|osw	cmn: no idea really.. i guess it would be hard to roll back changable data, i get your point on that	15:17
	← diegoviola left	15:18
cmn	so every time anybody does a change, you want to dump the whole db into your repo?	15:18
dennis\|osw	it would have to look at the name of a certain label (in this example), and not the value, but i guess that's hard	15:18
	← oneiroi left	15:18
cmn	this is precisely why you separate code from data and the MVC model exists	15:19
dennis\|osw	cmn: no that's another thing, i would like the full schema, but only content of certain tables to be tracked...	15:19
	→ oneiroi joined	15:19
	← oneiroi left	15:19
	→ oneiroi joined	15:19
	← oneiroi left	15:19
	→ csherin joined	15:20
cmn	why are those tables special	15:20
	?	15:20
	← ipalaus left	15:20
	← h0bbit left	15:20
dennis\|osw	cmn: then how would you go about custom editable multi language values without making it table content?	15:21
	→ ehsan joined	15:21
dennis\|osw	cmn: and how would you go about creating modules that may be disabled in a production enviroment, but will be added in a working environment, and on/off is based on a setting that also the end user can turn on/off in a cms?	15:22
	without making it table content?	15:22
	i'm not asking to be a smartass, but because i'm seriously interested.. :)	15:23
cmn	do have a source for those multi-language values?	15:23
dennis\|osw	it would be a table language with column name - value, and you have a table language_international with same columns, and the code checks where to get the values from	15:24
	→ LV-426 joined	15:24
dennis\|osw	(it's more columns than that, but this is enough info i guess)	15:24
cmn	so it's l10n data	15:25
	→ pidus joined	15:25
	← ehsan left	15:25
	→ jwmann joined	15:25
	→ h0bbit joined	15:25
LV-426	hello, is it possible to have git run 'indent' as a pre-hook for diif operations ? the idea is to be able to track a remote repo with different local code formatting preferences.	15:26
	→ nixmaniack joined	15:26
dennis\|osw	cmn: not familiar with the term, let me read up quickly :P	15:26
cmn	localisation, data you use to speak to the user in their language	15:26
	→ g105b joined	15:26
	← D-Chymera1 left	15:26
dennis\|osw	cmn: yeah read it, eh, i guess so yes	15:26
	→ flyz joined	15:26
	→ D-Chymera joined	15:27
	← eno left	15:27
	← ajaykools left	15:28
	← swombat left	15:28
	jacobrask → Creap	15:28
	← judaz left	15:28
	→ btanaka joined	15:28
	→ eno joined	15:29
	→ jonesy joined	15:29
DerekPioneer	when i use git-new-workdir i see one branch called "master"	15:29
	← Creap left	15:30
DerekPioneer	how do i edit that master and have it sync with my old work dir	15:30
	→ astrostl joined	15:30
cmn	does that come from you or does the customer decide what's appropriate?	15:30
	→ ephelion joined	15:30
	→ freezey joined	15:30
	→ kermit joined	15:30
	← kermit left	15:30
cmn	DerekPioneer: huh? didn't you want to have a different branch?	15:30
DerekPioneer	cmn yeah i did	15:31
	i ran this "git branch --track dev master"	15:31
	in the new work dir	15:31
	now i'm on the branch dev	15:31
	i want to push dev to master and then push master back to the old work dir	15:31
	→ ipalaus joined	15:31
dennis\|osw	cmn: appropriate in what way? you mean if we decide for our customers what content is appropriate?	15:32
DerekPioneer	or i guess	15:32
cmn	I mean if those strings come from you or does the user decide what they should be	15:32
DerekPioneer	i just want to push dev back to the master in work dir	15:32
	→ n8o-mba joined	15:32
dennis\|osw	cmn: that varies	15:32
	← cooldude left	15:32
	→ fly9 joined	15:32
	← glcrazy left	15:33
	→ gchristensen joined	15:33
	→ cakehero joined	15:33
cmn	DerekPioneer: so you want to merge dev into master	15:34
DerekPioneer	that depends, when i use new-workdir is master synced up with the old working directory?	15:34
dennis\|osw	cmn: usually it is part of a design that can be discussed with the client, but we design only 1 language, so on multiple languages we define a default value for languages we know, and the user changes or adds translations	15:34
cmn	the workdirs are separate from branches	15:34
dennis\|osw	cmn: they would usually do that after development	15:34
	← ephelion left	15:35
DerekPioneer	i have site.com and dev.site.com . how do i edit dev.site.com and then synch up those changes with site.com	15:35
	what's the right workflow supposed to be?	15:35
	← gnufied left	15:35
cmn	that sounds like it belongs in a different project; the way many projects deal with that is regular dumps of the l10n data	15:35
	as it's generally not version-tracked	15:36
	DerekPioneer: once you're happy with the dev branch (the branch, not whatever changes you've done on top and you haven't comitted) you merge it into your master branch	15:36
	← JackWinter left	15:37
	→ airborn joined	15:38
	→ JackWinter joined	15:38
cmn	once the l10n data arrives at the customer, do you have to change it yourself as well?	15:38
	or the changes happen at the customer site?	15:39
dennis\|osw	cmn: still, somewhere in the process, the language label is created in the database. And this label is - translated or not - already table content. Seperatiting it from the development of the change seems unlogical to me	15:39
	← h0bbit left	15:39
cmn	label being something that the user sees?	15:39
dennis\|osw	cmn: i guess that would occur in an acceptance environment	15:39
	→ RaceCondition joined	15:40
	→ macmartine joined	15:40
	→ kermit joined	15:40
	← kermit left	15:40
RaceCondition	I'm trying to do `git diff commit1..commit2 -- path` and getting "usage: git diff [--no-index] <path> <path>" -- why?	15:41
	← josemota left	15:41
RaceCondition	I just want to get the diff between commit1 and commit2 for only the path(s) I specify	15:41
	→ gnufied joined	15:41
	→ kermit joined	15:41
	← jonatasnona left	15:41
dennis\|osw	a label in this case is technically a language table entry, name - value... but without this, the change may not make sense so is part of the change. Regardless of the value of this language label, it needs to exist, and the existence depends on it's entry in a table, which is considered content.. don' tknow how to explain it better. ;)	15:41
cmn	loose the dots	15:41
	← gchristensen left	15:42
	→ axl__ joined	15:42
	← eno left	15:43
_ikke_	In a rebase conflict, what does deleted by us mean?	15:43
	→ mandric joined	15:43
shruggar	deleted by them, I think :)	15:43
	← isaak_ left	15:44
_ikke_	Ok	15:44
	hehe	15:44
	→ starlays joined	15:44
	→ Targen joined	15:44
shruggar	it's from the perspective of the "original" commits being applied to the new "base" commits	15:44
	→ eno joined	15:44
dennis\|osw	cmn: But apart from multi-lingual, there are also modular changes. We maybe create for example a "Slideshow module". This can be turned on/off. There is a table "modules" with a filed "activated" that is 1 or 0. This is technically also content the way you define it (or git would define it), but still it's a required change that can't be seperated from the project/commit, in my opinion.	15:44
	→ h0bbit joined	15:44
dennis\|osw	filed=field (column)	15:45
	← axl_ left	15:45
cmn	whether to activate a module or not would fall under user preferences in my book	15:45
	I don't quite get how the code is dependent on these fields existing or being set	15:46
	the fields should exist or the DB would complain, I guess	15:46
	but that's a schema change, not a change that adds content to a db	15:46
dennis\|osw	whether it is activated may be a user preference, but the ability to activate it is not	15:47
	→ cooldude joined	15:47
cmn	so you need to register it in some table?	15:47
dennis\|osw	so you think the code is broken if i say we use a "modules" table, and this contains the modules, there descriptions and whether or not it is activated?	15:48
	← angelsl left	15:48
dennis\|osw	cmn: yes, the module is in a table and the active modules will be checked and stored in the code in a modules array. The code will check if a module is active before it is shown	15:49
	← gusnan left	15:49
RaceCondition	any way to cherry-pick only for specific paths?	15:49
shruggar	dennis\|osw: generally one stores "migrations", ie: automated scripts for creating necessary database entries. /not/ the database itself (I only read one line and may have utterly mis-guessed what was being discussed)	15:50
	→ JameshC joined	15:51
	→ mikepack joined	15:51
dennis\|osw	shruggar: i guess this will be eventually the end conclusion	15:51
	shruggar: although it will probably not fit my company processes	15:52
cmn	I wouldn't consider it broken, just annoying	15:52
	← starlays left	15:52
	→ starlays joined	15:52
dennis\|osw	cmn: haha :)	15:52
cmn	but it's something that the deployment system should deal with	15:52
	it's not a reason to keep a the modules table tracked alongside the code	15:53
	← stodan left	15:53
	→ gchristensen joined	15:53
dennis\|osw	cmn: i guess the biggest issue will still be language related then	15:54
	→ Dragooon joined	15:54
dennis\|osw	cmn: mostly because front end html/css developers change them as they go in there building process	15:54
cmn	yeah, there's no good solution to that, unfortunately	15:54
dennis\|osw	they are not familiar enough with deployment scripts, etc.	15:54
RaceCondition	what output of `git diff` should I use to be later able to apply that diff to a bunch of files?	15:54
	→ alx- joined	15:54
cmn	have you thought of using gettext or something similar?	15:55
RaceCondition	--raw?	15:55
cmn	use format-patch	15:55
	→ spiette_ joined	15:55
	→ alex__c2022 joined	15:56
	← dfr\|mac left	15:56
	offby1 and RaceCondition should get together and start an Error Gang	15:57
RaceCondition	there's also a dude called SegFault in #python	15:58
	← stoffus left	15:58
	→ jonatasnona joined	15:58
	→ abetaha joined	15:58
dennis\|osw	cmn: is that gettext line referring to me?	15:58
	→ ephelion joined	15:58
cmn	eyah	15:58
	← h0bbit left	15:58
	← gandy93 left	15:58
cmn	yeah*	15:58
	← astrostl left	15:59
dennis\|osw	cmn: haven't thought of it.. will read up on it	15:59
cmn	I don't know how/if it works with HTML, but for programming languages, it's very helpful	15:59
	← Skyx-mobile__ left	15:59
	← eno left	15:59
	→ astrostl joined	15:59
dennis\|osw	cmn: no harm in reading up on it to find out :)	16:01
	→ jasonbray joined	16:01
dennis\|osw	got loads to think about anyway, this is one more on the list	16:01
	→ kennethreitz joined	16:01
dennis\|osw	thanks a lot for your input	16:01
	→ eno joined	16:01
dennis\|osw	cmn++	16:01
	→ ScottE joined	16:02
dennis\|osw	does that give you karme points? don't see anything happening! :)	16:02
wald0	how i can modify an existing commit (without create a new one) ? im reading about rebase but i dont quite get it (it creates a new commit)	16:02
dennis\|osw	karma*	16:02
cmn	the bot doesn't show it, but they're up on carmivore.com	16:02
	→ PigeonFriend joined	16:02
	→ billy_ran_away joined	16:02
cmn	carmivore.com/user shows some stats	16:02
	← cbreak left	16:03
	→ mabrand joined	16:03
dennis\|osw	on other note, what is a good irc client nowadays? i come from the mirc period, but now it's not freeware anymore? :O	16:03
FauxFaux	irssi. ¬_¬	16:03
cmn	I use smuxi	16:04
	← jatt left	16:04
	→ h0bbit joined	16:04
	← arvind_khadri left	16:04
	tizzo → tizzo-afk	16:05
dennis\|osw	FauxFaux: your reply seems to contain sarcasm or special personal benefits >:)	16:05
shruggar	most people use irssi… I can't bring myself to be one of them	16:06
	→ Destos joined	16:06
FauxFaux	dennis\|osw: It's a good suggestion, it's just basically the exact opposite of mirc. :p	16:06
shruggar	I use Adium, which is not an IRC client.	16:06
	xchat was my preferred client, but I don't think that has an OSX version	16:07
haarg	there is	16:07
	it's terrible	16:07
shruggar	Adium is "good enough" as a basic interface, for me. No client has ever been satisfactory	16:08
	→ paszo2008 joined	16:09
	→ davidrussell joined	16:09
	→ neurodrone joined	16:09
haarg	i just switched to using macirssi yesterday and so far it's better than the alternatives. had to compile it myself though to get the perl plugin working.	16:09
	← Element9_work left	16:09
	→ techstein joined	16:10
	← ephelion left	16:11
dennis\|osw	argh, i can't work with that... i'm a windows user! ;)	16:12
	← glennpratt left	16:13
dennis\|osw	it lacks any form of design ;-)	16:13
	← jonatasnona left	16:13
	→ vipershadow joined	16:13
	← flijten left	16:14
	→ jonatasnona joined	16:14
cmn	I'd expect that from an OSX user	16:14
	→ viller joined	16:14
cmn	but from an user of the OS that went from ugly grey boxes to fisher-prize	16:15
	that seems out of place	16:15
shruggar	I once wrote an ftp "client" (as a test, mind you) which just created FIFOs for whatever input/output was needed. I recommend doing that for IRC as well.	16:15
	← cbreak-work left	16:15
	→ eXp`iRc\|37982 joined	16:16
	← eXp`iRc\|37982 left	16:16
	← h0bbit left	16:16
	→ Element9_work joined	16:16
	← jonatasnona left	16:16
	→ ephelion joined	16:17
	→ dcarper joined	16:17
	← Sitethief left	16:17
	→ cbreak joined	16:19
	→ icwiener joined	16:19
	← batlock666 left	16:19
	← LotR left	16:19
	→ LotR joined	16:20
	→ texasmynsted1 joined	16:20
	← eno left	16:21
	→ tiagobutzke joined	16:21
	← tiagobutzke left	16:21
	→ spal joined	16:22
	← shiba_yu36 left	16:22
	← andrewa2 left	16:22
wald0	it is not possible to modify a commit without create a new one ?	16:22
	→ eno joined	16:22
	→ h0bbit joined	16:22
	→ kenperkins joined	16:23
cmn	no	16:23
spal	Is the first diagram accurate here? http://progit.org/book/ch2-2.html I thought "git add newfile" is going to stage the file as well. But the diagram indicates that adding a new file would put it into unmodified state.	16:23
cmn	wald0: you might find notes useful to annotate commits after the fact	16:23
shruggar	wald0: what is your use-case?	16:24
	← preinheimer left	16:24
cmn	spal: the file is staged, but it's considered unmodified because there's nothing to compare against	16:24
spal	so it is staged as well as unmodified. The diagram doesn't seem to capture that.	16:25
	→ khmarbaise joined	16:25
cmn	staged in that diagram means "modified since the last commit and staged to be included in the next one"	16:26
wald0	shruggar: debian packaging with git-buildpackage, it converts the patches in a special branch which every patch is a commit, this state is converted back too (every commit is converted to a series of patches), so i dont know how to modify a patch if is equivalent to a commit	16:26
cmn	...staged so the changes are included in the next one	16:26
	→ wagle joined	16:26
	→ Kingsy joined	16:26
	→ tomasm- joined	16:27
	← pro_metedor left	16:27
	→ gltest_ joined	16:27
Kingsy	got a quick question about adding a remote to git using ssh, I am assuming something like this is correct --> git remote add remoteName [email@hidden.address] yeah?	16:27
	but how do you sepcify a path? and or a port for the ssh connection?	16:28
	say the path I wanted to remote to was /srv/http/blah/	16:28
	← oc80z left	16:28
	→ oc80z joined	16:28
tomasm-	hi, is it possible to create a dump of content from a specific revision (not just diff between revisions, but the whole files, but only those that changed), suitable for a zip file, so I don't overwrite local changes? i want a way to 'backup' the old 'live' version for easy restoration if needed	16:28
shruggar	Kingsy: [email@hidden.address] to specify a port, you need to edit your .ssh/config	16:28
dennis\|osw	i'm out, tnx again cmn, shruggar, will reconsider the way we commit db changes vs deployment script. Not sure if it'll be beneficial/doable, but will consider it.	16:28
Kingsy	shruggar: ok thanks	16:28
	→ hackoo joined	16:28
	← ephelion left	16:29
wald0	shruggar: but if git can't edit/modify commits, this is a PITA	16:29
	← gltest left	16:29
shruggar	tomasm-: git archive is what you want	16:29
cmn	Kingsy: or you can use a proper ssh url like ssh://user@host:port/path	16:29
	← dennis\|osw left	16:29
cmn	man git clone has a lot of examples	16:29
gitinfo	the git-clone manpage is available at http://jk.gs/git-clone.html	16:29
	← hackoo left	16:29
	→ hackoo joined	16:30
Kingsy	so git remote add remoteName ssh://user@host:port/path/to/folder is also correct?	16:30
cmn	yes, that's another way to specify ssh urls	16:30
Kingsy	cool thanks	16:30
	→ sythe joined	16:30
	← hackoo left	16:31
	→ hackoo joined	16:31
	→ davidcoallier joined	16:31
offby1	jast: what are the odds that you can make the gitbot work in #github too? I always want to type e.g. !repro there, and a always disappointed	16:32
gitinfo	jast: Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.	16:32
	← hackoo left	16:32
offby1	s/and a/and am/	16:32
	← cilly left	16:32
	→ hackoo joined	16:32
	→ TikaL joined	16:33
tomasm-	shruggar, it's a LOT of files in the tree, can I get just those that have changed between revisions? ie to 'undo' a newer revision	16:33
	→ number_number joined	16:33
	← hackoo left	16:33
	→ hackoo joined	16:34
shruggar	tomasm-: can you describe your situation again? What do you want to do which every-day-use git commands don't let you do?	16:34
	→ alvaro_o joined	16:34
	→ CEnnis91 joined	16:34
	← hackoo left	16:34
	← alvaro_o left	16:34
	← dharrigan left	16:34
	← psoo left	16:35
	← vipershadow left	16:35
	← ShooterMG left	16:36
	→ ephelion joined	16:36
	← h0bbit left	16:36
number_number	if I want to commit to my github repo from two different machines.. do I need to put my token containing .gitconfig on both.. or do I generate a seperate token on each machine?	16:36
	→ kukks joined	16:36
tomasm-	shruggar, i just want to generate a zip file of the PREVIOUS revision's files, at least the subset of files that changed, so I can just unzip the files to revert (now that i think about it, I can just revert the tree, but I may want to have zip files for a certain 'state', without it being every last file)	16:37
	→ gusnan joined	16:37
	→ PySloan joined	16:37
Kingsy	hmm I think I am missing something, I am trying to setup my first git repo.. I have a repo on my local machine, and I created a repo with git init --bare on my server.. I then made a change to my local repo and did git push origin master it pushed fine but I cant see any change on my remote repo? what happened?	16:37
cmn	number_number: that's not what the github token is for	16:37
Kingsy	git status on the remote repo says "fatal: This operation must be run in a work tree" so I am assuming I have missing something "big"	16:37
jast	offby1: sorry, code assumes one channel only	16:37
cmn	Kingsy: you're not in a worktree	16:38
	or not in something git recognises as one	16:38
	oh, you're trying to find the files on the server	16:38
	→ dharrigan joined	16:38
	← dharrigan left	16:38
	→ dharrigan joined	16:38
cmn	that's not what the repo is for	16:38
ephelion	When you initialize a "bare" git repo, there IS no working tree	16:39
cmn	git log -p --all will tell you whether they've arrived	16:39
Kingsy	ah yup	16:39
	so git log shows that it worked great :)	16:39
	← csherin left	16:40
Kingsy	however perhaps I am missing the point with this...	16:40
cmn	possibly	16:40
	a bare repo is for synchronisation	16:41
	it's not where you work on files	16:41
Kingsy	I wanted to have a development system in which the remote server was "production" and the local machine was dev, you make changes each time you commit.. whenever you complete a change in full you push to the production server, (which is live)	16:41
	→ h0bbit joined	16:41
tomasm-	shruggar, I think what I'm looking for is "git archive -o backup.tar master $(git diff --name-only) # To run before I 'commit'	16:41
cmn	have your post-receive hook run a deployment script	16:42
	← barr5790 left	16:42
Kingsy	cmn - what that for me?	16:42
cmn	something like git archive master \| tar xf - -C /var/www/whatever	16:42
	← keller left	16:42
cmn	yeah	16:42
	→ duckxx joined	16:42
Kingsy	umm that doesnt make sense from what I know about this so far.. isnt this the whole point of git? so all devs have a copy of the same files?	16:43
	→ canton7 joined	16:43
	→ keller joined	16:43
	→ ephelion_ joined	16:44
cmn	yes	16:44
Kingsy	by pushing and fetching? from what I think your saying it sounds like a repo does something different	16:44
cmn	a bare repo is different from a non-bare repo	16:44
Kingsy	oh ok .. so I shouldnt have created a bare repo then	16:44
number_number	cmn: my bad, I've missed a trick I guess.. when I try "git clone [email@hidden.address] I get 'permission denied'.. which I suspect is because I've identified myself as myname in ~/.gitconfig.. but I might be way out	16:44
	← keller left	16:44
	← ephelion_ left	16:44
cmn	number_number: that's down to your ssh config	16:44
	→ ephelion_ joined	16:44
	→ mordof joined	16:44
Kingsy	cmn: what type of repo would you recommend for my scenario?	16:44
RandalSchwartz	my pre-recieve deploy hook - https://gist.github.com/1714235	16:44
mordof	how would i see all the files tracked in my repo?	16:44
	← ephelion left	16:45
Kingsy	i assume there is a "standard" repo of some sort?	16:45
cmn	Kingsy: do what Randal says	16:45
Kingsy	that does what I need?	16:45
shruggar	mordof: git ls-files	16:45
	← ephelion_ left	16:45
cmn	depends what you define as standard	16:45
	→ ephelion_ joined	16:45
RandalSchwartz	season to taste, of course	16:45
mordof	shruggar: thanks	16:45
cmn	the non-bare repo is the one you work on	16:45
RandalSchwartz	I'm copying just a portion of the files	16:45
	← ephelion_ left	16:45
mordof	when i do a git rm it deletes the files out of the directory as well correct?	16:45
cmn	a bare repo is one where you synchronise	16:45
	→ ephelion joined	16:45
RandalSchwartz	sitaram suggested a "git-fast-export to rsync" filter :)	16:46
ephelion	Kingsy: If you are going to use git for web development, I highly recommend reading Joe Maller's post: http://joemaller.com/990/a-web-focused-git-workflow/	16:46
RandalSchwartz	I like that	16:46
	but it's not done yet. :)	16:46
	← BiggFREE left	16:46
Kingsy	RandalSchwartz: hmm I don't really understand that.. does that mean you git repo is in a seperate dir to your "live site" and then you use that so when its done receiving a change it copys the relevent files from the repo into the live site?	16:46
RandalSchwartz	Kingsy - precisely	16:47
	← eno left	16:47
cmn	a thing to remeber is that your web server isn't a developer	16:47
RandalSchwartz	that hook is on my server	16:47
	I edit on my laptop	16:47
	→ keller joined	16:47
RandalSchwartz	when I "git push origin master", it just copies to the server, but no deploy	16:47
Kingsy	ok I see...	16:47
RandalSchwartz	when I "git push origin master:live", that triggers the hook	16:47
	→ JohnBat26 joined	16:47
Kingsy	ahhh right I see	16:47
RandalSchwartz	the hook takes a "git-archive" of what I pushed to live, and then populates the live site	16:47
	as a side effect, the "live" branch is always precisely what's live	16:48
	so I can diff against that to see what my next changes will do	16:48
	→ eno joined	16:49
Kingsy	RandalSchwartz: hmm I don't really know how to set that up	16:49
RandalSchwartz	!man githooks	16:49
gitinfo	the githooks manpage is available at http://jk.gs/githooks.html	16:49
cmn	!website has the instructions	16:49
gitinfo	Git is not a website deployment tool, but can sometimes play one in sufficiently simple/lax environments with a little help. One example of help is: http://toroid.org/ams/git-website-howto	16:49
	← SeySayux left	16:49
cmn	kinda	16:49
	→ netzapper joined	16:49
mordof	so..	16:50
	hmm	16:50
netzapper	has anybody embedded git (or specifically jgit) into an application to be used as a software update feature?	16:50
mordof	i'm running the same thing right now, this is a but odd	16:50
cmn	I think that one uses a lame checkout -f, but you just need to get the idea	16:50
FauxFaux	netzapper: So you want all of your clients to have every previous version of your app on disc, and two copies of the latest version?	16:50
	→ AzizLight joined	16:50
AzizLight	Hi everybody	16:50
	← johnathanb left	16:50
netzapper	FauxFaux: damn. Good point.	16:50
AzizLight	how can I get the last modification date of file using git please?	16:51
	← revagomes left	16:51
FauxFaux	AzizLight: git log --format="%ae %cd" -1 -- path/to/file	16:51
	← mastro left	16:51
ephelion	Could you also use git whatchanged <file>?	16:52
FauxFaux	Uh, %ad, not %ae.	16:52
	← bgerber left	16:52
cmn	same difference	16:52
	ephelion: ^^	16:52
	→ SeySayux joined	16:52
cmn	except whatchanged doesn't limit itself to the last commit, as FauxFaux's command does	16:52
Kingsy	thanks for the links.. very very cool	16:52
ephelion	cmn: Gotcha.	16:53
	← Heimidal left	16:53
	← sythe left	16:53
	→ sythe joined	16:53
AzizLight	thanks guys	16:53
	→ bgerber joined	16:53
	→ mastro joined	16:54
	→ churp joined	16:54
	→ ShooterMG joined	16:55
	← h0bbit left	16:55
	← babilen left	16:55
	← number_number left	16:56
	→ pandeiro joined	16:56
	← iocor left	16:57
	← icwiener left	16:58
	→ codeshah joined	16:58
	→ icwiener joined	16:58
EugeneKay	It amuses me that my Windows desktop has a newer git version than my linux boxen	16:58
	→ Yuuhi joined	16:59
	← dr_lepper left	16:59
mick_home_	so, I did a: git push origin master - from the windows box and it just sits there	16:59
	→ eijk__ joined	16:59
mick_home_	the codebase doesn't warrant the time that it sits there and does nothing	16:59
	:-/	16:59
canton7	mick_home_, what's the protocol? git:// ?	17:00
	← eijk_ left	17:00
Kingsy	exit	17:00
	← Kingsy left	17:00
mick_home_	canton7: yes	17:00
RandalSchwartz	exit: not found :)	17:00
	→ dr_win joined	17:00
mick_home_	haha	17:00
canton7	mick_home_, that has issues under windows where it hangs	17:00
mick_home_	alternatives?	17:00
canton7	mick_home_, ssh or http(s)	17:00
	ideally ssh	17:00
mick_home_	not really on windows	17:01
	ssh on windows can suck	17:01
canton7	msysgit coems with openssh, which works fine as an ssh client	17:01
	→ h0bbit joined	17:01
mick_home_	ah true	17:01
	as a client :)	17:01
	← whitman left	17:01
RaceCondition	if a new SVN branch has been created upstream, how to make it visible to git-svn?	17:01
canton7	for a server... I've seen it work fine with cygwin + openssh, and copssh	17:02
RaceCondition	`git svn rebase` seems to have no effect, i.e. `git branch -a` still doesn't list the newly created SVN branch	17:02
mick_home_	so, how do I go about changing it from git:// to ssh?	17:02
	← Pupeno left	17:02
	← tomasm- left	17:02
	← robotmay left	17:03
mick_home_	do I do a new `git pull` and then cp the changed files over to that and then do a push?	17:03
cmn	man git remote	17:03
gitinfo	the git-remote manpage is available at http://jk.gs/git-remote.html	17:03
canton7	mick_home_, git remote set-url	17:03
cmn	there's set-url there	17:03
mick_home_	the url then after that?	17:04
	→ dr_win_ joined	17:04
	← dr_win left	17:04
canton7	git remote set-url <remote_name> <remote_url>	17:04
	→ Pupeno joined	17:04
mick_home_	what is remote_name?	17:05
	as in, master?	17:05
canton7	the name of your remote. Probably 'origin'	17:05
	← EricInBNE left	17:05
mick_home_	would it start w/ git+ssh://	17:06
canton7	probably ssh://user@host/path, or the other syntax is user@host:path. See the "Git URLs" section of man git-clone	17:06
gitinfo	the git-clone manpage is available at http://jk.gs/git-clone.html	17:06
	→ DarkAR joined	17:06
	→ chmac joined	17:07
canton7	bear in mind that git:// and ssh are different protocols, and you won't be able to talk to a server serving just git:// by speaking over ssh	17:07
	→ tomasm- joined	17:07
	→ boombatower joined	17:07
	→ jeremyw joined	17:07
EugeneKay	They're the same protocol; git:// is a native TCP transport, ssh:// does it over a SSH tunnel	17:07
jeremyw	Hello all.	17:07
mick_home_	canton7: so, I'll need to reconfigure git on the server? (i have ssh and git running)	17:08
	→ i42n joined	17:08
tomasm-	is it possible to have 'macros' in a repository file that get updated, etc when you make changes to a file, so you know what version it is? something automatic.	17:08
jeremyw	I've been looking around online and I still can't figure out how to mark a local branch such that "git push" does not push it to remote. (I'm trying to keep a local branch as private until I get things in order.)	17:08
	← viller left	17:08
canton7	mick_home_, if you're currently set up to server over git:// (you have git-daemon running), and you haven't set it up to serve over ssh, then yes	17:08
jeremyw	I started looking at refspecs and I think that's what I need to create/edit but not knowing them that well, I figured I'd ask for help.	17:08
	← sythe left	17:09
	→ LeMike joined	17:09
	← eno left	17:09
mick_home_	canton7: ok, I'll have to look into what changes I need to make	17:09
	→ Q\| joined	17:09
chmac	I have a hosted git repository on codebasehq.com. I want to have /server1/etc/ in the online repository, and then "checkout" in svn terms that into /etc/ on my machine. How do I do that with git?	17:10
FauxFaux	chmac: You can't.	17:10
	← h0bbit left	17:11
	→ eno joined	17:11
EugeneKay	Sure you can.	17:11
	→ revagomes joined	17:11
EugeneKay	chmac: man git-archive, use <path> and pipe it to tar -x	17:11
gitinfo	chmac: the git-archive manpage is available at http://jk.gs/git-archive.html	17:11
shruggar	that would be closer to "export" in svn terms	17:12
	when faced with a similar situation, I checkout the whole thing, then use symlinks	17:12
EugeneKay	Eh, I dunno the svn verbage. But it's a solution :-p	17:12
chmac	shruggar: Interesting, that's certain a possibility.	17:12
	→ MANCHUCK joined	17:13
jeremyw	Can someone confirm my question didn't just get overlooked due to channel chatter? I know ethics is to be patient and not reask but with flurry of talk around the time I asked, would be nice to know it was seen.	17:13
chmac	EugeneKay: Thanks, I'll look into that.	17:13
EugeneKay	Another(and the one I use with my etckeeper repos) is to use branches for each machine	17:13
canton7	jeremyb, look at push.default in man git-config	17:13
gitinfo	jeremyb: the git-config manpage is available at http://jk.gs/git-config.html	17:13
	← youngblood left	17:13
	→ diegoviola joined	17:13
jeremyw	canton7: Thanks.	17:13
chmac	EugeneKay: Didn't know about etckeeper, interesting, thanks.	17:13
EugeneKay	It's waaaaay better than trying to screw it up yourself	17:14
canton7	you probably want 'upstream' -- that will make git push only the current branch when you run "git push". With any value of push.default, "git push" will not push a branch to a remote which does not yet exist on the remote	17:14
tomasm-	hi, how can i compare a file that isn't within my repository against any revisions IN the repository? ie to see whether it matches a version/commit or not?	17:14
	→ cassovlet joined	17:14
EugeneKay	man diff ;-)	17:14
	← diegoviola left	17:14
EugeneKay	Looking for it within the history.... I dunno, actually.	17:14
	→ eijk joined	17:15
jeremyw	canton7: That seems right but I end up with my local branch, untracked or not, being created on my remote which is why I started looking around. I wish there was an easier way to just say "this is local only, do not push ever". ;)	17:15
	← d3vic3 left	17:15
	→ h0bbit joined	17:15
canton7	jeremyb, there's a thread on the ML about implementing something -- similar to hg's phases -- but no way to say "never push this" yet	17:15
	→ iocor joined	17:16
mordof	will using git rm remove a file from my working directoy? i simply want to stop tracking a file	17:16
jeremyw	canton7: Thanks for the heads up. I'll have to be a little more careful when pushing then.	17:16
	← eijk__ left	17:16
	← TikaL left	17:16
	← Davey left	17:16
canton7	jeremyb, with push.default set to upstream, git will only implicitely push the current branch, as I say.	17:17
	mordof, git rm --cached to untrack the file but not delete it	17:17
tomasm-	how do I print out (to stdout) a specific revision of a file, without modifying the current version?	17:17
mordof	canton7: ty	17:17
canton7	tomasm-, git show revision:filename	17:17
	← jeremyw left	17:17
mick_home_	hmm, so I was reading some docs on it and it doesn't say that you need to do anything differently on the server - yet it doesn't recognize my repo :(	17:18
	→ Bass2 joined	17:18
	→ johnathanb joined	17:18
EugeneKay	The server needs to have git installed	17:19
mick_home_	I tried to do a new: git clone ssh://me@git.myhostname.org/my-reponame	17:19
	when i do: git clone git://git.myhostname.org/my-reponame <-- this works	17:19
EugeneKay	You need to do user@hostname:/path/to/repo/	17:19
	← Bass2 left	17:19
mick_home_	aah	17:19
	→ Davey joined	17:19
	← Davey left	17:19
	→ Davey joined	17:19
EugeneKay	Same sort of syntax rsync takes	17:19
mick_home_	so :/srv/git/my-reponame	17:19
	ok, let re retry	17:20
EugeneKay	You may be interested in !gitolite	17:20
gitinfo	Want to host as many git repos (and users!) as you like, on your own server, with fine-grained access control? You want gitolite: https://github.com/sitaramc/gitolite	17:20
	← gldn left	17:20
	← pdtpatrick left	17:20
chmac	EugeneKay: etckeeper definitely looks like the way to go. Can I have it automatically pushed (is that the right verb?) to the codebase repository on every commit?	17:20
canton7	unless you ssh to [email@hidden.address] assuming git's home is at /srv/git -- then you can do git clone git@hostname:my-reponame.git	17:20
	→ Bass2 joined	17:20
mick_home_	ok, no ":" necessary	17:20
	→ kushou joined	17:20
	← eno left	17:20
mick_home_	but it worked	17:20
EugeneKay	chmac - sure, build a post-commit hook to invoke git-push	17:20
mick_home_	I just needed the full path	17:21
chmac	EugeneKay: Ok, I'll research both of those topics. Thanks a lot.	17:21
EugeneKay	Not a particularly GOOD idea, but you can. :-p	17:21
	chmac: man githooks	17:21
gitinfo	chmac: the githooks manpage is available at http://jk.gs/githooks.html	17:21
	→ tewecske joined	17:21
	← vandal- left	17:21
	ChanServ set mode: +v	17:21
	← teweWORK left	17:21
	← paszo2008 left	17:22
	→ devth joined	17:22
	← Element9_work left	17:22
	→ eno joined	17:22
mick_home_	canton7: git clone ssh://me@hostname/srv/git/my-reponame <-- worked for me	17:22
	← philips left	17:22
	← LimeBlast left	17:22
	← johnathanb left	17:23
	← JohnBat26 left	17:23
	→ kenichi joined	17:24
	→ philips joined	17:24
	← tomasm- left	17:24
	← senny left	17:25
	→ csherin joined	17:26
	← h0bbit left	17:28
	← soulcake left	17:31
	← tchan left	17:31
	← jasonbray left	17:32
	← yshh left	17:32
	← giallu left	17:32
mick_home_	where do I submit the bug regarding "git push" + git:// on windows?	17:33
	→ ericmuyser joined	17:33
	→ ephelion_ joined	17:33
	→ vpopov joined	17:33
	← ephelion left	17:33
	ephelion_ → ephelion	17:33
	→ tchan joined	17:33
	→ MistuhKurtz joined	17:34
cmn	it's a known bug, but not one that's going to get a lot of attention	17:34
	pushing over git:// is not something you should do	17:34
	← tcurdt left	17:35
	→ h0bbit joined	17:35
	← berserkr left	17:35
	← vpopov left	17:35
	← chmac left	17:35
	← davidcoallier left	17:35
	← lapistano left	17:36
	→ lapistano joined	17:37
	← RaceCondition left	17:37
	← abetaha left	17:37
	→ amstan joined	17:37
	→ Heimidal joined	17:37
amstan	so let's say i git add something	17:38
	then delete the file	17:38
	is there a way to recover the file from the index?	17:38
cmn	delete as in only remove from the filesystem?	17:38
	→ Jay_Levitt joined	17:38
amstan	yes	17:38
cmn	checkout -- file	17:38
amstan	perfect, i wasn't sure that checkout worked with index	17:38
cmn	unless you tell it otherwise	17:39
	← fruight left	17:39
amstan	i always used it with work with the commits	17:39
	← Bass10 left	17:39
	← Bass2 left	17:40
	→ Bass10 joined	17:40
	← Bass10 left	17:41
	→ Bass10 joined	17:42
	← ericmuyser left	17:42
	→ fruight joined	17:42
	→ caseymcg joined	17:42
	← achadwick left	17:43
	→ yshh joined	17:43
	← niko left	17:43
	← yshh left	17:44
	→ isaak joined	17:45
	→ kerrick joined	17:45
	→ Raging_Hog joined	17:45
	← h0bbit left	17:47
	← pidus left	17:47
	← jason237 left	17:47
	→ nrichards joined	17:48
	← loincloth left	17:48
	→ h0bbit joined	17:49
	← AzizLight left	17:49
	→ abetaha joined	17:50
	→ dorkmafia joined	17:50
	→ maletor joined	17:51
	tizzo-afk → tizzo	17:51
	→ urbanmonk joined	17:51
	← urbanmonk left	17:52
	→ urbanmonk joined	17:52
	→ aidalgol joined	17:54
	→ jason237 joined	17:54
	→ apok joined	17:58
	← apok left	17:58
	→ dreiss joined	17:58
	← stuntmachine left	17:58
	→ OOPMan joined	17:59
	← ephelion left	17:59
	← urbanmonk left	17:59
	← bauruine left	18:00
tango_	sitaram: am I wrong or in gitolite there is currently no support for multiple keys per user, and using groups is the only way around that?	18:00
	← h0bbit left	18:01
	→ stuntmachine joined	18:01
cmn	you are wrong	18:01
	[email@hidden.address]	18:02
	← amstan left	18:02
	→ PovAddict joined	18:02
PovAddict	hi	18:02
	what tool do you suggest for large files in git? not as a normal repo, more like what git-annex does	18:02
cmn	!annex	18:03
	hm, the bot doesn't like that	18:03
	← gnufied left	18:03
cmn	annex is one option, git-media is another	18:03
	← pandeiro left	18:03
PovAddict	git-annex sucks for me, I've only been experimenting and I almost-lost data twice	18:03
	→ h0bbit joined	18:04
PovAddict	cp foo foo-copy; git annex add foo foo-copy; (now both are symlinks to the same internal object, yay, deduplication for free)	18:04
	git annex unannex foo	18:04
	now 'foo' is again the original data, and 'foo-copy' is a broken symlink	18:04
	→ whyking joined	18:04
	← brianpWins left	18:05
	← leavittx left	18:05
cmn	then send a bug report	18:05
	though why you'd want to do that...	18:05
PovAddict	first I have to replace these broken symlinks with the correct data :/	18:05
	→ psoo joined	18:05
	← bartek left	18:05
	← mishok13 left	18:06
PovAddict	cmn: I have lots of folders with photos, I ran "git annex add" on some of them, some had duplicated files, then I unannexed them all (why does it matter why? isn't unannex is supposed to be the reverse of add?) and it only recovered one of the duplicates, the other set is now broken links	18:07
	← Splash left	18:07
	→ bartek joined	18:07
cmn	the example is what I wondering about	18:08
	→ niko joined	18:08
cmn	it didn't give any information as to the usefulness of the method	18:08
	→ Skyx-mobile__ joined	18:08
PovAddict	suppose I annex-add'ed it accidentally	18:08
cmn	I've no idea whether unannex is the opposite of add, maybe it isn't	18:08
	→ preinheimer joined	18:09
	← shruggar left	18:10
	← X-tonic left	18:10
	→ dunpeal joined	18:10
	→ robotmay joined	18:11
	← matthiaskrgr left	18:11
dunpeal	Hey. Is there a good alias for `git push <some remote> $current_branch`?	18:11
	Basically, a shortcut to pushing only the current head to whatever remote you specify, or the default if you don't.	18:12
	← jceb left	18:13
	← flyz left	18:13
	→ matthiaskrgr joined	18:13
cmn	man git config push.default	18:13
gitinfo	the git-config manpage is available at http://jk.gs/git-config.html	18:13
	← OOPMan left	18:14
	← PySloan left	18:14
	← spiette_ left	18:15
	→ soulcake joined	18:16
	← h0bbit left	18:16
	← kenichi left	18:17
	→ codepoet_ joined	18:17
	← PovAddict left	18:17
	← Demego left	18:17
	← techstein left	18:18
	→ whitman joined	18:20
	→ Chaser joined	18:20
	→ flyz joined	18:20
	← soulcake left	18:20
	← caseymcg left	18:21
	← eno left	18:22
	→ notola joined	18:22
	→ wingie joined	18:22
	← wingie left	18:22
	← jfelchner left	18:23
	→ soulcake joined	18:23
	→ eno joined	18:23
	→ caseymcg joined	18:24
	← iocor left	18:24
	← PigeonFriend left	18:25
	→ X-tonic joined	18:25
	→ techstein joined	18:25
	→ vpopov joined	18:26
	→ caaakeeey joined	18:28
	← notola left	18:28
	→ OOPMan joined	18:28
	← caseymcg left	18:28
caaakeeey	not particularly a git question - but im not sure where else to ask, lets say im writing an open source library/framework, and a app that uses it, whats the best way of structuring things before im at the stage where it makes sense to use releases and package managers?	18:30
	→ amstan joined	18:30
cmn	that depends on how close they are	18:31
amstan	so i would like to get a precommit hook that outputs some of my files to pdfs and saves them in the repository, but i don't want it to do it each time, only if it's necessary	18:31
	as in, if the file changed this commit	18:31
	is there a way to do that?	18:31
PerlJam	caaakeeey: the best way is the one that you understand and works for you :)	18:31
amstan	i already have a makefile that can accept file names and stuff for exporting	18:31
cmn	amstan: look at the default pre-commit hook	18:31
	it only looks at changed files	18:31
SethRobertson	sitaram: I don't understand what the purple (github) is supposed to represent	18:32
caaakeeey	maybe abetter example would be to say i have multiple apps that are using this library	18:32
	← ipalaus left	18:32
caaakeeey	and whenever i work on an app, im likely to make changes to the framework/library as it finds it's feet	18:32
	← vpopov left	18:32
	← noisebleed left	18:32
	← soulcake left	18:33
SethRobertson	sitaram: In your recent http://sitaramc.github.com/gitolite/pictures.html	18:33
	It doesn't particularly matter to me, but if I don't understand it, others may not either	18:33
amstan	cmn: i don't really see anything of this sort there	18:35
cmn	caaakeeey: then they're not that separate; you can later use git-subtree to separate out he library if you need to	18:35
	→ preflex_ joined	18:35
SethRobertson	caaakeeey: You might find that !gitslave works well here. Or not.	18:35
gitinfo	caaakeeey: gitslave (http://gitslave.sf.net) is useful to add subsidiary git repositories to a git superproject when you control and develop on the subprojects at more or less the same time as the superproject, and furthermore when you typically want to tag, branch, push, pull, etc. all repositories at the same time.	18:35
cmn	amstan: it's in line 31	18:36
	you'll want to adjust the filter to M	18:36
	but that's how you get the name of the changed files since the last revision	18:37
	← eno left	18:37
	← preflex left	18:37
caaakeeey	cmn, yeah they aren't that separate, but the multiple apps that use it in the same way certainly are	18:37
	preflex_ → preflex	18:37
	→ hrutu joined	18:37
	→ yshh joined	18:37
amstan	cmn: you mean git diff --cached --name-only --diff-filter=A -z HEAD?	18:37
cmn	yes	18:38
amstan	cmn: and then s/A/M?	18:38
sitaram	SethRobertson: just something external that has the gitolite source; see command #2 in the bullets below	18:38
amstan	cmn: they both output nothing	18:38
cmn	except you need $against sometimes	18:38
	← gusnan left	18:38
sitaram	SethRobertson: normally github, but could also be googlecode since I push a copy there also	18:38
	→ eno joined	18:38
	← hrutu left	18:39
	→ hrutu joined	18:39
	→ h0bbit joined	18:39
SethRobertson	sitaram: Ah, I thought it had something to do with user "alice", not where the directory came from.	18:39
	← hrutu left	18:39
cmn	amstan: can you come up with a reproducible test case?	18:39
	→ ezzzy joined	18:39
ezzzy	hi	18:39
	guys, doing a git add . gives me this error: fatal: CRLF would be replaced by LF in vendor/assets/stylesheets/blueprint/plugins/fancy-type/screen.css.	18:39
SethRobertson	sitaram: Since it was in the column for Alice's workstation	18:40
amstan	cmn: yeah, i got it now, i forgot to git add	18:40
cmn	there you go then	18:40
	→ iocor joined	18:40
amstan	cmn: so this gives me a list of files changed, do you know anything offhand that i can use to filter those files and remap them to another folder? when calling make i have to give it the pdf files not the original files	18:40
	i guess grep and sed would do it	18:41
cmn	sed or some bash scripting	18:41
	→ shruggar joined	18:41
	← alex__c2022 left	18:42
	→ DofLeFou joined	18:43
	→ alex__c2022 joined	18:43
	← keller left	18:43
	← airborn left	18:45
	← h0bbit left	18:46
	← shruggar left	18:46
	→ shruggar joined	18:46
	← Chaser left	18:47
	← ezzzy left	18:48
	→ RaceCondition joined	18:48
	← alex__c2022 left	18:49
	← Ryback_ left	18:49
	→ keller joined	18:49
	→ airborn joined	18:49
	← eletuchy left	18:50
	← RaceCondition left	18:50
	→ RaceCondition joined	18:51
	→ herriojr joined	18:51
	← boombatower left	18:51
	← shruggar left	18:51
	← RaceCondition left	18:51
	→ diogogmt joined	18:52
	← LeMike left	18:53
caaakeeey	thanks for your help guys :)	18:53
	← caaakeeey left	18:53
	→ shruggar joined	18:53
	→ yrlnry joined	18:53
	← n8o-mba left	18:53
	gitinfo set mode: +v	18:54
	← khmarbaise left	18:54
	→ magnus_ joined	18:54
	← mbroeker left	18:55
	← MANCHUCK left	18:56
herriojr	hey, so I have a project with some submodules, and even though there are no changes, it keeps saying something is modified in the submodule (new commits), but when I "cd" into the submodule, I get "nothing to commit (working directory clean)"	18:56
	→ MANCHUCK joined	18:56
herriojr	every attempt at reverting, so it doesn't recognize any changes fails	18:56
dunpeal	cmn: thanks	18:56
	← dunpeal left	18:56
	← whyking left	18:58
	→ VladGh joined	18:58
	→ Chaser joined	18:59
	→ n8o-mba joined	18:59
	← Vile left	19:00
	→ codepoet__ joined	19:01
	← LV-426 left	19:01
	← codepoet__ left	19:01
	→ Splash joined	19:01
	→ mishok13 joined	19:01
	← codepoet_ left	19:02
	→ di-five joined	19:02
	← dnjaramba left	19:02
	→ dnjaramba joined	19:03
	← stuntmachine left	19:03
	← Q\| left	19:03
	→ Ratchet486 joined	19:04
	→ dp_ joined	19:06
	→ Invader joined	19:06
	→ jasonbray joined	19:07
	ChanServ set mode: +v	19:07
	→ demet8 joined	19:07
	→ h0bbit joined	19:08
	→ lonewulf` joined	19:08
	→ Rogi joined	19:08
	→ stuntmachine joined	19:09
Deck`	what does git rev-parse 'HEAD@{1}' mean ?	19:10
RandalSchwartz	!man git-rev-parse	19:10
gitinfo	the git-rev-parse manpage is available at http://jk.gs/git-rev-parse.html	19:10
RandalSchwartz	and in fact, that manpage also describes what HEAD@{1} is	19:11
Ratchet486	I have a repository. I am using it to track sql scripts. (the scritps are for a MSSQL db and I am using Microsoft SQL Server Management Studio to edit them.)Today I noticed that they are being tracked as binary files. They are text. I dont know why they are being tracked as binary files.	19:11
	Can anyone tell me how to track them as text	19:12
	→ cybersphinx joined	19:13
	→ mantic0 joined	19:13
	← mantic0 left	19:13
	← dp_ left	19:14
	→ mantic0 joined	19:14
Ratchet486	I tried to google this already. I am having a difficalt time finding info on it	19:14
CareBear\	sorry, did any of what I just said get through?	19:14
Deck`	RandalSchwartz i found it, just it doesnt work, says "ambiguous argument 'HEAD@{1}': unknown revision or path not in the working tree."	19:15
	← isaak left	19:15
CareBear\	Ratchet486 : discover why they are considered binary, I guess ms uses ucs2, don't discover by googling, but by studying the actual bytes in the files	19:15
	← the_cybersphinx left	19:15
Ratchet486	CareBear no that is the first message I saw from you	19:15
	Oh	19:16
	← kerrick left	19:16
Ratchet486	Once I know that I can ...	19:16
	and what is ucs2?	19:16
	unicode of some sort?	19:16
	→ apok joined	19:17
CareBear\	yes	19:17
	→ glennpratt joined	19:17
	← johnhamelink left	19:17
	→ aspotashev joined	19:18
	← kenperkins left	19:18
	← fisted left	19:18
Ratchet486	So once I know that I will be able to do what?	19:18
	← h0bbit left	19:19
	← macmartine left	19:19
	→ jehoshua02 joined	19:19
	→ infogulch joined	19:19
jehoshua02	Anybody here familiar enough with gitolite to help with some troubleshooting?	19:20
	← kushou left	19:21
	→ eletuchy joined	19:21
	← Omni\|AFK left	19:22
	→ Omni\|AFK joined	19:22
canton7	jehoshua02, ask away, and anyone who knows will chip in	19:22
	→ fisted joined	19:22
	← fisted left	19:22
	→ fisted joined	19:22
jehoshua02	canton7: thanks. First things first, how do I know the master key works?	19:22
	Well, how do I verify that my gitolite install is working?	19:23
	Omni\|AFK → Omni\|Work	19:23
canton7	jehoshua02, you mean the key you supplied when you first installed gitolite? It works if you can clone the gitolite-admin repo over ssh :P	19:23
jehoshua02	Yesh.	19:23
amstan	oh man! now you're telling me?	19:24
Omni\|Work	How does git handle tracking branches when you mention the full repo URL on the push command line?	19:24
	→ guilhermecortere joined	19:24
amstan	i can't actually add or delete files from the git index in pre-commit	19:24
	so there's no way to have it export the pdfs just before the commit	19:24
Deck`	how to get old revision ?	19:24
	← marthinal left	19:25
	→ Element9 joined	19:25
amstan	Deck`: git checkout revisionhash	19:25
Deck`	assuming that git rev-parse HEAD return the current one	19:25
	← fisted left	19:25
	← dpino left	19:25
	→ h0bbit joined	19:25
canton7	Deck`, define "get". You can have a peek in read-only mode, create a branch at some past point, reset the current branch head to a past revision, or make the contents of the next commit the same as an arbitrary previous commit	19:25
	→ zweb joined	19:25
	→ fisted joined	19:25
	→ chris_zweber joined	19:25
amstan	Deck`: what i told you is peek, you shouldn't commit after that	19:25
RandalSchwartz	do you want the parent of the current commit?	19:25
	← zweb left	19:26
	← chris_zweber left	19:26
RandalSchwartz	that's "git checkout HEAD~"	19:26
	→ marthinal joined	19:26
	← DarkAR left	19:26
	→ zweb joined	19:26
amstan	RandalSchwartz: oh man, you can just HEAD~? i wrote HEAD~1 for the longest time like an idiot!	19:26
RandalSchwartz	yeah	19:26
	← dnjaramba left	19:26
RandalSchwartz	"I have a video that says that" :)	19:26
	→ alex__c2022 joined	19:27
	→ dnjaramba joined	19:27
	→ ericmuyser joined	19:28
	→ break57382 joined	19:28
	→ tktiddle joined	19:29
	← shruggar left	19:29
	← eletuchy left	19:29
	→ fr0stbyte joined	19:29
	→ eletuchy joined	19:29
	← Chaser left	19:29
	→ gusnan joined	19:30
jehoshua02	When I ssh into the git user, it asks for the passphrase, then the password?	19:30
	→ vandal- joined	19:30
	← fr0stbyte left	19:30
	→ Chaser joined	19:30
jehoshua02	I'm pretty sure it shouldn't do that, but I'm not sure why it's doing that.	19:30
tktiddle	If i add new files to my .gitignore how can I remove them from the origin without deleting them from my local branch?	19:30
	← vandal- left	19:30
amstan	so here's my pre-commit hook: http://pastebin.com/W6syz52z	19:31
	← break57382 left	19:31
	→ break57382 joined	19:31
amstan	why does it choke on the git rm? i get the following: fatal: Not a git repository: '.git'	19:31
	← billy_ran_away left	19:31
	← guilhermecortere left	19:31
offby1	amstan: put "set -x" somewhere in the script, which should cause it to echo each command. That might help.	19:32
	put it at the top.	19:32
	← glennpratt left	19:32
	→ berserkr joined	19:32
amstan	offby1: yep, git rm chokes it	19:33
	can i do git rm from inside a pre-commit hook?	19:33
	← ericmuyser left	19:33
	← jehoshua02 left	19:33
amstan	this guy seems to have it working just fine: http://stackoverflow.com/questions/4764304/git-export-mysql-database-on-commit	19:33
	why can't i do it though?	19:33
	it's like an enviroment variable isn't working	19:34
	if i cd ../.. just before it then it works	19:34
kevlarman	amstan: use GIT_WORKTREE in your hook if you need anything relative to that	19:35
	← eletuchy left	19:35
	← h0bbit left	19:37
	← X-tonic left	19:38
	→ Vile joined	19:38
	→ barr5790 joined	19:39
	→ hrvojem joined	19:39
	← cortana2 left	19:39
	→ h0bbit joined	19:39
	→ jehoshua02 joined	19:40
	← mastro left	19:40
	← flippo left	19:40
	→ vandal- joined	19:41
	← diogogmt left	19:41
	→ churp_ joined	19:42
Omni\|Work	I need to update remote refs on a push to a repo specified on the command line.	19:42
	← churp left	19:42
Omni\|Work	I've had a really hard time making this happen.	19:42
	← icwiener left	19:43
	→ icwiener joined	19:43
	→ becomingGuru joined	19:43
	← break57382 left	19:43
Omni\|Work	Is there some trick to it I should be aware of?	19:44
	→ break57382 joined	19:44
	← spaceonline left	19:44
	← Chaser left	19:45
kevlarman	amstan: err, it's GIT_WORK_TREE	19:45
	→ canton8 joined	19:45
	← hrvojem left	19:46
amstan	kevlarman: yeah, i did this just before the command, it seemed to work: unset GIT_DIR	19:47
kevlarman	amstan: that's a bad idea	19:47
amstan	kevlarman: my script works perfectly now!, i'm so happy	19:47
	kevlarman: why?	19:47
kevlarman	amstan: because git uses that	19:47
	→ felipec joined	19:47
	← felipec left	19:48
amstan	kevlarman: otherwise git add doesn't work if i cd	19:48
	→ felipec joined	19:48
amstan	kevlarman: and i have to cd, otherwise make won't like me	19:48
	and my variables will look like ass if i have to get GIT_WORK_TREE in all of them	19:48
kevlarman	amstan: you know you can use pushd/popd or spawn a subshell?	19:48
amstan	so, what if i save the variable, unset it, do the git add, then set it back?	19:49
	← cehteh left	19:49
kevlarman	amstan: that's an ugly hack	19:49
amstan	but like.. it's working fine at the moment	19:49
	kevlarman: and spawning a subshell isn't?	19:49
kevlarman	amstan: you're relying on behavior that might break	19:49
	→ pdtpatrick joined	19:50
	→ btx94 joined	19:50
amstan	i'm going to have something like pre-commit.bootstrap then a real pre-commit that's being launched in a subshell	19:50
	that seems weird	19:50
kevlarman	amstan: you don't need a separate hook for a subshell	19:50
amstan	ok, so what do you propose?	19:51
	← astrostl left	19:51
kevlarman	amstan: try a script that does 'pwd;(cd /etc;pwd);pwd'	19:51
amstan	here's my script: http://pastebin.com/ir7ENJfm	19:51
	← h0bbit left	19:51
amstan	kevlarman: i can't, because git add also needs to be executed from that folder	19:51
	→ drizzd joined	19:51
amstan	and in that folder it'll give me that .git not found error	19:51
	← jacekmigacz left	19:52
amstan	oh..	19:52
kevlarman	amstan: replace cd with pushd	19:52
	instead of unset, use popd	19:52
	← vervic left	19:52
kevlarman	and modify the path to git add	19:52
	→ spaceonline joined	19:53
kevlarman	(to eagle/pdf/*.pdf)	19:53
amstan	kevlarman: what about git rm?	19:53
kevlarman	amstan: same	19:53
	also your seds can be replaced with basename	19:54
Omni\|Work	More reliably replaced.	19:54
	→ leavittx joined	19:54
Omni\|Work	Using sed to split apart paths... :-P	19:54
amstan	Omni\|Work: i was hoping all of that was easier about 2 hours ago	19:54
	Omni\|Work: yet here i am	19:54
	Omni\|Work grins.	19:55
kevlarman	amstan: doing things the "right" way is always easier in the long run	19:55
	→ linusoleander joined	19:55
amstan	i think the main problem is that makefiles should know about git stuff	19:55
	Omni\|Work shakes his head.	19:56
jehoshua02	I'm on a fresh CentOS 6.2 minimal install. I've just installed gitolite. I'm trying to clone the admin repo, but it keeps asking me for "password" instead of "passphrase" (which I left empty for the gitolite admin key).	19:56
kevlarman	amstan: not assuming they're being run from the folder they're in is a good start	19:56
Omni\|Work	Whenever I've seen a make script rely on details of the VC system, it's almost always been bad.	19:56
	← canton8 left	19:56
kevlarman	Omni\|Work: exporting version info is a valid reason to rely on git	19:56
amstan	Omni\|Work: yeah, fine	19:56
	Omni\|Work: but still, they're simple in concept, but when they get more complicated everything falls apart	19:57
Omni\|Work	kevlarman: Yes, that is one good exception to that rule.	19:57
amstan	Omni\|Work: like the fact that the main behavoir is to give it what you want it to make, instead of what you just updated	19:57
kevlarman	Omni\|Work: hell i've seen makefiles that export version info from git in projects that use svn :D	19:57
	← dnjaramba left	19:57
	→ h0bbit joined	19:57
jehoshua02	Okay, so, I created a test user, copied the private key for the gitolite-admin and tried `git clone git@server:gitolite-admin`.	19:57
	→ dnjaramba joined	19:58
jehoshua02	It asks for passphrase, I hit enter, then it asks for git password. Any ideas?	19:58
Omni\|Work	jehoshua02: Did you multiply the factors in your private key together properly?	19:58
kevlarman	jehoshua02: you need the passphrase to decrypt the key	19:58
	← faber left	19:59
kevlarman	Omni\|Work: obviously everyone should be able to do 2048 bit multiplication in their head	19:59
	:D	19:59
Omni\|Work	kevlarman: Yes! It's really easy if the numbers involved are expressed in base 2!	19:59
	→ snearch joined	20:00
	← aspotashev left	20:00
Omni\|Work	Then it's just lots of shifting and adding.	20:00
amstan	now i just need to tell me exporter program to pdf to also include git details in there	20:00
kevlarman	Omni\|Work: but then you have to convert to a higher base to type it out in a reasonable amount of time	20:01
jehoshua02	Omni\|Work: I generated the gitolite-admin keys like this: ssh-keygen -t rsa	20:01
Omni\|Work	Well, base16 is the obvious choice there. nods sagely	20:01
	→ gavin_huang joined	20:01
jehoshua02	empty for no passphrase.	20:01
Omni\|Work	Though, base64 is also good.	20:02
	← gavin_hu_ left	20:02
Omni\|Work	jehoshua02: Interesting. Is the public key stored in the ~/.ssh/id_rsa file?	20:02
	Err, private key.	20:03
jehoshua02	id_rsa is the private key. id_rsa.pub is the public key, both in ~test/.ssh	20:03
Omni\|Work	That's interesting. Maybe someone has stupidly and insanely disabled RSA based logins on the server?	20:03
Rogi	is there a tool thats as good as github's compare?	20:04
Omni\|Work	Or, PublicKeyAuthentication I believe the option is.	20:04
jehoshua02	Omni\|Work: The default is yes on CentOS 6.2 minimal.	20:04
	→ Lummoxx joined	20:04
kevlarman	jehoshua02: you encrypted the key	20:04
Omni\|Work	Rogi: My favorite tool is a rocket launcher, mostly because it's so much fun to make things explode.	20:04
	→ sythe joined	20:05
kevlarman	jehoshua02: so now you need to supply the decryption password to use the key	20:05
	← Heimidal left	20:05
Omni\|Work	kevlarman: jehoshua02 says (s)he generated it with an empty passphrase.	20:05
jehoshua02	he.	20:05
Omni\|Work	Don't like to assume. :_)	20:06
kevlarman	Omni\|Work: and the output says otherwise	20:06
	Omni\|Work nods.	20:06
Omni\|Work	I would agree.	20:06
	← merlin83_ left	20:06
jehoshua02	I need passphrase-less keys because another application on the server will maintain the gitolite-admin for me.	20:06
	← ISF left	20:07
Lummoxx	hi all, anyone have a link handy that can show me git version compatibility (eg: version on server older than client)?	20:07
	→ jacekmigacz joined	20:07
Omni\|Work	What happens if you simply type 'ssh git@server'?	20:07
cmn	Lummoxx: it's not a problem	20:07
	← alex__c2022 left	20:07
kevlarman	Lummoxx: you have to go really far out of your way to break compatibility	20:07
	Lummoxx: older than .9	20:07
	or something stupid like that	20:07
	← mick_home_ left	20:07
cmn	the wire protocol supports extensions	20:08
Lummoxx	heh, ok, excellent. Thanks!	20:08
cmn	and both instances agree on what features to use	20:08
kevlarman	Lummoxx: the only reason you particularly need new git on the server is if you want to use smart http	20:08
	(dumb http is always a bad idea)	20:08
Lummoxx	makes sense.	20:08
	→ Demego joined	20:09
Lummoxx	Thanks for the info, I appreciate it.	20:09
	← dnjaramba left	20:10
	→ dnjaramba joined	20:10
tktiddle	If i add new files to my .gitignore how can I remove them from the origin without deleting them from my local branch?	20:11
	← aidalgol left	20:11
	← h0bbit left	20:11
	← toabctl left	20:11
yrlnry	Suppose I have a series of commits. I would like to break this into two parallel branches, one containing all the changes to the files in set X, and the other containing all the changes to the files in set not-X. Is there a convenient way to do this?	20:12
	→ rzec joined	20:13
	← snearch left	20:13
	→ RaceCondition joined	20:13
rzec	What is the best way to handle someone who works on windows and git showing that they changed every file because the line endings changed?	20:14
canton7	rzec, read http://help.github.com/line-endings/ and !crlf	20:14
gitinfo	rzec: to fix problems with line endings on different platforms, check out http://line-endings.gitrecipes.de/	20:14
	← abetaha left	20:15
canton7	tktiddle, to stop git from tracking files, without deleting the files, git rm --cached <file>	20:16
	← kevwil left	20:16
	→ abetaha joined	20:16
	→ h0bbit joined	20:16
	← roxlu_ left	20:17
	← n8o-mba left	20:17
Omni\|Work	When I do 'git push <repo_url> <branch>' it doesn't create the branch in the remote repo. And if the branch already exists in the remote repo, it doesn't update it.	20:18
	← Dragooon left	20:18
canton7	Omni\|Work, what does it output?	20:19
	→ RaptorX joined	20:19
Omni\|Work	'Up to date' or something like that.	20:20
	← Lgb left	20:20
Omni\|Work	Acts like there's nothing to push.	20:20
	→ ipalaus joined	20:20
Omni\|Work	It tends to work if I do just a plain old 'git push' without specifying a repo URL.	20:20
canton7	Omni\|Work, does it work if you add the url as a remote, or is it strictly a problem when a url is used with git push?	20:21
	← robotmay left	20:21
Omni\|Work	I haven't tried adding it as a remote in the 'normal' way.	20:21
canton7	(i.e. git remote add tempremote <repo_url>; git push tempramote <branch>)	20:21
Omni\|Work	I've done it by editing the config file by hand, and that didn't seem to help much, but I suspect I didn't do it right.	20:21
	canton7: I will try that next time I have the issue and come back here to ask more questions if it doesn't work.	20:22
	→ Lgb joined	20:22
Omni\|Work	Thanks!	20:22
canton7	Omni\|Work, cool. That'll settle whether it's a branch-pushing issue in general, or just something to do with using a url with push	20:22
Omni\|Work	I partly want to know because I've been asked to write a script which will purposely abandon a branch and force the branch ref to a new spot on master.	20:23
	Which is an evil thing to do, and I explained the nature of the evil, but that's what's wanted anyway. :-)	20:24
	→ merlin83_ joined	20:24
	→ rtconner joined	20:24
	→ fedesilva joined	20:25
	→ Heimidal joined	20:25
	← h0bbit left	20:25
	← di-five left	20:25
	→ johnhamelink joined	20:26
	← zweb left	20:26
	← dnjaramba left	20:27
	→ ericandrewlewis joined	20:27
	→ dnjaramba joined	20:28
	← codeshah left	20:28
	← tqmz left	20:29
	→ tqmz joined	20:29
	→ n8o-mba joined	20:29
	→ h0bbit joined	20:31
	← tatsuya_o left	20:31
	→ glcrazy joined	20:33
	← sythe left	20:34
	→ sythe joined	20:34
	← shergill left	20:35
	→ Retistic joined	20:35
Retistic	i'm new to git	20:35
	i just called igt revert and window cam up in console. i edited the revert commit message, but i'm not sure how to save now	20:35
	→ icwiener_ joined	20:36
	→ no1home joined	20:36
cmn	you're in an editor	20:36
Retistic	cmn: yeah, how do i save it and execute the commit	20:36
cmn	if you haven't configured any, git launches vim by default	20:36
	ESC ZZ	20:36
cbreak	on windows? :)	20:36
	← i42n left	20:36
Retistic	cmn: ah i hate vim, never used it. i use gedit	20:37
cmn	then you should tell git that	20:37
Retistic	cmn: how can i do that?	20:37
cbreak	export EDITOR=gedit	20:37
Retistic	cbreak: ubuntu	20:37
	← icwiener left	20:37
cbreak	in your .profile or what ever	20:37
amstan	i hate it that vim doesn't let itself killed by ctrl+c either	20:37
cmn	put that in your .bashrc somewhere	20:37
amstan	so it's just the worst thing for begginers and people not familiar with it	20:37
cmn	that's not what's supposted to happen	20:37
cbreak	amstan: it just protects you from not having to read the manual	20:37
amstan	cbreak: yes, but i don't want to use it at all	20:38
	cbreak: yet some programs insist of launching it, and then i have no way of closing it	20:38
cbreak	you can use killall vim in an other console :)	20:38
cmn	C-c is meant to be catchable by the application	20:38
	→ zweb joined	20:38
amstan	cbreak: yeah, which is really annoying	20:38
cmn	amstan: have you told your system you don't want vim?	20:38
cbreak	or close it properly	20:38
	:wq	20:38
amstan	cmn: i did	20:38
	cmn: and it works in my profile	20:39
	cmn: but then i do visudo and it complains about vim not being installed	20:39
cmn	then that's a bug of whatever application, not of vim	20:39
	right	20:39
	→ soidexe joined	20:39
Retistic	okay i pressed ^x and it prompted to save, so the revert worked. how do i find the file to change the editor?	20:39
cbreak	visudo uses some stupid program that is not vim on the system I used today	20:39
cmn	visudo /is/ asking your system to use vi to edit sudoers	20:39
soidexe	may I ask about github here?	20:39
cbreak	half the file was filled with control characters when I was done :/	20:39
cmn	Retistic: it's ~/.bashrc	20:39
	← glcrazy left	20:40
cmn	soidexe: if it's github-specific, it should go to #github	20:40
hyperair	cbreak: it uses EDITOR or VISUAL	20:40
amstan	soidexe: sure, but there's also a #github channel	20:40
	→ txomon joined	20:40
	← txomon left	20:40
	→ txomon joined	20:40
cbreak	hyperair: not the one I set up	20:40
amstan	soidexe: if it applies..	20:40
hyperair	hmm that's werid	20:40
cbreak	hyperair: wonder if sudo preserves that env var	20:40
hyperair	ah	20:40
	maybe not	20:40
cbreak	I used sudo visudo	20:40
hyperair	try sudo env	20:40
	← dnjaramba left	20:40
cmn	cbreak: if you start it as vi, it behaves as vi instead of vim	20:40
amstan	cbreak: yeah, depends on the sudo settings, and editor wasn't included in there by default	20:40
	← vandal- left	20:40
amstan	cbreak: not in arch	20:40
	→ dnjaramba joined	20:40
	← felipec left	20:40
Retistic	cmn: i don't see any export line	20:41
cbreak	cmn: it had two lines of shortcuts in the window bottom	20:41
	← QaDeS left	20:41
cmn	Retistic: add it	20:41
cbreak	I don't think that was anything even related with vi	20:41
cmn	cbreak: oh, that's nano	20:41
cbreak	what ever it was, I didn't like it :)	20:41
Retistic	just add "export EDITOR=gedit" anywhere in the file?	20:41
cmn	yes	20:41
hyperair	cbreak: man sudoers and search for env_editor	20:41
cmn	except the first line	20:41
Retistic	cmn: thank you very much for your help	20:41
soidexe	How can I identify which branch this commit belongs to? https://github.com/kutzi/jenkins/commit/239add9781a26f2157ba602a419189c255a87421	20:41
cmn	soidexe: git branch --contains <hash>	20:42
	→ Xizor joined	20:42
	← h0bbit left	20:42
cbreak	on OS X I never had that problem	20:42
	(never had to change sudoers either...)	20:42
	← RaceCondition left	20:42
soidexe	cmn: doesn't work..	20:42
SethRobertson	soidexe: !repro (hint, you are wrong if you used it properly)	20:43
gitinfo	soidexe: Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.	20:43
cmn	maybe you need to add -a	20:43
hyperair	cbreak: sudoers mentions why.	20:43
	cbreak: it's a security hole blah blah	20:43
	so yes, it's a security feature to ignore VISUAL/EDITOR.	20:43
soidexe	cmn: thanks! -a helped	20:43
	→ vervic joined	20:43
cbreak	hyperair: ... makes sense I guess	20:43
cmn	helpes as in what?	20:43
cbreak	but it could easily make it not a security hole	20:44
	visudo that is	20:44
cmn	in case VISUAL='rm -rf/' ?	20:44
cbreak	by using the user of the invokiing user	20:44
hyperair	cbreak: by the way, sudoers on ubuntu says that it uses /usr/bin/editor by default, which is configurable	20:44
cbreak	and just running the app on a temporary file	20:44
	→ shergill joined	20:44
cbreak	hmm.	20:44
	sudoers has an insults property.	20:45
hyperair	oh yeah it's fun. =p	20:45
	/usr/bin/editor links to /etc/alternatives/editor which is configurable using update-alternatives --config editor	20:45
	← spaceonline left	20:46
cmn	whenever I look at that I'm amazed debian doesn't hit the symlink depth limit	20:46
	→ codeshah joined	20:46
hyperair	what's the limit anyway?	20:46
	→ aristidesfl joined	20:47
hyperair	i've only ever hit the limit with recursive symlinks	20:47
aristidesfl	hi	20:47
cmn	no idea, 5 or 10 or so, I think	20:47
hyperair	but that's because they're recursive and can never resolve to anything anyway	20:47
Omni\|Work	Is there a simple command that simply tells me which branch is currently checked out?	20:47
aristidesfl	what's the difference between the reset —hard andhe checkout?	20:47
SethRobertson	Omni\|Work: No, you need to look for the * in git-branch output	20:48
	→ tangledhelix joined	20:48
SethRobertson	aristidesfl: A lot, depending on exactly how you use the commands. They can be the same in certain specialized circumstances.	20:48
	→ h0bbit joined	20:49
	→ Milossh joined	20:49
	← Heimidal left	20:49
aristidesfl	SethRobertson: is there a difference between "checkout HEAD~" and "reset —hard HEAD~" ?	20:49
	→ s0ber_ joined	20:50
SethRobertson	aristidesfl: Yes! Very much so.	20:50
aristidesfl	SethRobertson: does the checkout only change the working directory?	20:50
SethRobertson	aristidesfl: One changes where the current branch pointer points, the other shows you a read-onlyish version of the previous commit	20:50
	aristidesfl: Only is such a strong word. It also puts you on a detached head	20:50
	→ khmarbaise joined	20:51
	→ Heimidal joined	20:51
aristidesfl	SethRobertson: except if I checkout the head like when I checkout a branch?	20:51
	← s0ber left	20:51
	s0ber_ → s0ber	20:52
cmn	when you checkout HEAD, the symref is dereferenced, so you're telling git to checkout the current branch	20:52
	HEAD~ is not a branch, it's a commit	20:53
aristidesfl	cmn: I know	20:53
SethRobertson	aristidesfl: If you say `git checkout local-branch-name` then it puts you on the other branch. Everything else will put you on a detached head. Entirely unlike `git reset` which will change where the HEAD points to (depending on arguments)	20:53
aristidesfl	you didn't understand my question	20:53
	→ vandal- joined	20:53
aristidesfl	SethRobertson: what are the implications of behing on a detached head? what happens if I stage and commit while on a detached head?	20:54
	→ jmho joined	20:54
cmn	the same as if you weren't	20:54
	but no branch ref will be updated when/if you commit	20:54
SethRobertson	aristidesfl: It will "work", but if you then change to another ref, that commit you just made will be dangling and you cannot easily get back to it. In two weeks, if you do nothing to it, it will go away	20:54
	→ unclemantis joined	20:55
SethRobertson	You can `tag` or `branch` to create a ref to point at the detached/dangling commit and it will be "normal" then	20:55
unclemantis	If I made changes and I don't want to commit any of them in a branch and go back to the way the branch was before i made any changes do i use git reset --hard HEAD	20:55
SethRobertson	unclemantis: That could work. See !fixup for full details	20:56
gitinfo	unclemantis: So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions. For hints type !fixup_hints in IRC. Remember: if you have pushed already, there are only a few things you can do without !rewriting_public_history (type that for more info)	20:56
	← h0bbit left	20:56
unclemantis	i have not used git add . or git commit -am yet	20:57
aristidesfl	SethRobertson, cmn: so basically if I commit while on a detached it's like I'm creating a ghost branch which if not named in 2 weeks will be deleted?	20:57
	→ skylord5816 joined	20:57
	← skylord5816 left	20:57
	→ skylord5816 joined	20:57
	← skylord5816 left	20:57
	→ skylord5816 joined	20:57
SethRobertson	unclemantis: The full details covers that case, but in general your proposal will work.	20:58
cmn	you're created unnamed history	20:58
	→ LucasTizma joined	20:58
	← nixmaniack left	20:58
aristidesfl	cmn which can be converted to a branch?	20:58
SethRobertson	unreferenced history would be more technical	20:58
	→ nixmaniack joined	20:58
SethRobertson	Which could be explicitly referenced by pointing a branch ref at it, yes	20:59
	→ shruggar joined	20:59
cmn	I almost wrote unreferenced, but it's referenced by the reflog	20:59
aristidesfl	so what is the most common use for this?	20:59
cmn	throw-away changes, tests	21:00
aristidesfl	not the branch, but the checkout previous commit	21:00
SethRobertson	aristidesfl: Let me clarify this for you: when you are on a detached head, treat it like it was read-only. Yes, you can technically commit to it, but...don't.	21:00
cmn	inspecting earlier commits	21:00
aristidesfl	ok thanks for the precious help :)	21:00
	sythe → sythe-afk	21:01
	← zweb left	21:02
	→ themachine joined	21:02
	← vandal- left	21:03
	→ h0bbit joined	21:03
	← vervic left	21:04
	← erichynds left	21:04
	← Jackneill left	21:06
	→ zulax joined	21:08
Omni\|Work	I find the concept of detached heads to be rather creepy.	21:09
	Almost as bad as headless workstations.	21:09
	← sudoquai left	21:09
cmn	why would you have a headless workstation?	21:10
	← v0lksman_ left	21:10
	← khmarbaise left	21:10
	→ MS515631 joined	21:10
Omni\|Work	Because you operate it via ssh.	21:10
cmn	that wouldn't qualify as a workstation for me	21:10
Omni\|Work	Workstation is just the place where you do work. You could have a laptop which is basically a terminal and ssh into your workstation.	21:11
	→ techhelp joined	21:11
cmn	right, so if I don't sit at it, I don't call it a workstation	21:11
	Omni\|Work nods.	21:11
Omni\|Work	I can see that point of view. But then again, I have a head.	21:12
	→ kenichi joined	21:12
	← h0bbit left	21:13
	→ urbanmonk joined	21:13
	← oc80z left	21:13
	→ oc80z joined	21:15
	← psoo left	21:15
	Omni\|Work writes a small script to generate keyclick noises in a poisson distribution.	21:15
	→ vandal- joined	21:16
SethRobertson	And when you checked it into git you had a problem?	21:16
kevlarman	Omni\|Work: just get a model m or cherry blue keyboard and it'll do that for you :D	21:16
	→ soulcake joined	21:16
	→ acl_ joined	21:16
	← tangledhelix left	21:16
	SethRobertson claims model M's rule. ← typed on a model M	21:16
	← cccaldas left	21:17
kevlarman	SethRobertson: right now i'm using crappy rubber dome keyboard :(	21:17
	cherry brown at home	21:17
unclemantis	i am diggin this http://sethrobertson.github.com/GitFixUm/#uncommitted_everything	21:18
yrlnry	"cherry brown" sounds like the pseudonym of a porn actress known for anal scenes.	21:18
	→ alex__c2022 joined	21:18
unclemantis	i do have an issue. Please kill the green. It makes everything else look red when i look away :(	21:18
kevlarman	(don't recommend those actually, too soft, if i was buying a new one today i'd get clear or black)	21:18
SethRobertson	That is a feature	21:18
	→ jfelchner joined	21:18
	→ h0bbit joined	21:18
	← MS515631 left	21:19
cmn	think of it as saving on drugs	21:19
	(but yeah, that green is horrible)	21:19
	→ adammokan joined	21:19
unclemantis	I am currently at http://sethrobertson.github.com/GitFixUm/#uncommitted_everything	21:20
	and trying to figure out what I should be doing :(	21:20
	will git reset --hard do what i want?	21:20
	← heptadecagram left	21:20
unclemantis	undo all changes since the branch was created?	21:20
cmn	created?	21:20
	that's a fuzzy concept	21:21
	← dnjaramba left	21:21
SethRobertson	unclemantis: Since the branch was created? No. It will undo all uncommitted changes made to files which have been committed.	21:21
aristidesfl	unclemantis: that will reset to the last commit	21:21
SethRobertson	unclemantis: As it says, this will not touch untracked or ignored files	21:21
	→ dnjaramba joined	21:21
unclemantis	as you can see, i just don't want to F up	21:22
SethRobertson	unclemantis: Once you are done with that, if you want to then get rid of commits you have made, restart.	21:22
	← btanaka left	21:22
SethRobertson	(restart the question session)	21:22
unclemantis	i will give git reset --hard a try	21:22
	← shruggar left	21:22
	→ btanaka joined	21:22
unclemantis	what do i have to loose? I have a backup in github and i have a backup on my live server	21:23
aristidesfl	SethRobertson: restart?	21:23
	→ kenperkins joined	21:23
Omni\|Work	I want to be able to specify the hardness of my reset on the Mohs scale.	21:23
SethRobertson	From the top, yes	21:23
unclemantis	damn it i don't know if that worked or not	21:23
	there are still some TOUCHED files	21:24
	← matthiaskrgr left	21:24
SethRobertson	What does "TOUCHED" mean? git-status is more technical in its terminology	21:24
unclemantis	i know	21:25
	→ zomothesweeper joined	21:25
	→ matthiaskrgr joined	21:25
	← jasonbray left	21:25
unclemantis	i am going to pull	21:26
	the heck with this	21:26
	says it is up to date but yet	21:26
SethRobertson	yet?	21:26
Omni\|Work	I think it means the file is a little odd.	21:26
SethRobertson	Perhaps, but some clarity is needed.	21:26
	← n8o-mba left	21:27
SethRobertson	pulling will not help, BTW, if you input to the pull is bad	21:27
unclemantis	sigh the files that where created and i did not add or commit are freaking in my develop branch and when i did a pull it said that it is up to date	21:27
	i am all confursed	21:27
Omni\|Work	Perhaps the word 'TOUCHED' should be replaced with 'schizophrenic'.	21:27
SethRobertson	unclemantis: Did you read the document, or my email, when I said that? (16:21:43) SethRobertson: unclemantis: As it says, this will not touch untracked or ignored files	21:28
unclemantis	ya, that's it :)	21:28
	sigh ok	21:28
SethRobertson	The very paragraph you were reading when you found out about git-reset told you what to do.	21:28
unclemantis	so that explains the untracked files that are now MERGED into my develop branch from my categories branch	21:28
	← h0bbit left	21:28
SethRobertson	they are likely not merged	21:28
	Since they were not committed	21:28
cmn	if they're untracked, they can't be merged	21:28
SethRobertson	They are just there. Floating in space	21:29
unclemantis	but they are physicaly there	21:29
	→ statusfailed joined	21:29
unclemantis	i can see them in the ls	21:29
cmn	they may exist on the filesystem whilst the branch is active	21:29
	that doesn't mean they're in the branch	21:29
	→ mick_home_ joined	21:29
mick_home_	hi everyone	21:29
SethRobertson	!hi	21:29
gitinfo	[!welcome] Welcome to #git, a place full of helpful gits. If you have a question, feel free to just go ahead and ask—somebody should answer shortly. For more info on this channel, see http://jk.gs/git/ Take backups (type !backup to learn how) before taking advice.	21:29
mick_home_	I'm trying to make master on the server the same as a working directory and toss out all of the old changes	21:30
	→ whyking joined	21:30
statusfailed	I accidentally have been committing to a detached HEAD, so i've made a branch with "git branch detached <hash>", how can I make HEAD point at this branch?	21:30
Omni\|Work	I'm more of a silly git.	21:30
unclemantis	gonna try git clean -ndx	21:30
mick_home_	I've done a fresh checkout, git rm *, and git commit && git push remote	21:30
SethRobertson	statusfailed: git branch newbranch	21:30
	← break57382 left	21:31
SethRobertson	then git checkout newbranch	21:31
	Or probably git checkout -b newbranch	21:31
statusfailed	SethRobertson: oh yeah, I did the checkout too	21:31
unclemantis	that is giving me a list of files that would be removed	21:31
statusfailed	oh, without the -b	21:31
	→ n8o-mba joined	21:31
unclemantis	how do i remove created files and revert changes of existing files back?	21:31
SethRobertson	statusfailed: The -b is a shortcut for automatically creating the branch	21:31
mick_home_	when I copied the working directory it says that it can't do that because there is data in the repository and it gives me something regarding: receive.denyCurrentBranch	21:31
statusfailed	SethRobertson: Oh I see! So is it now "HEAD"? Do I need to merge it with master or something?	21:32
unclemantis	so i would do git clean -fdx	21:32
SethRobertson	unclemantis: YES!!! Did you keep reading? Please read the ENTIRE paragraph	21:32
	← fairuz_ left	21:32
	→ fairuz_ joined	21:32
mick_home_	that I need to change that (can't read the whole thing because windows sucks at creating terminals)	21:32
unclemantis	SethRobertson i just don't want to completely freak this up	21:32
SethRobertson	statusfailed: `git branch` will tell you whether it is HEAD. Look for the *	21:32
aristidesfl	if I do git reset branch, will I delete the current branch?	21:32
cmn	mick_home_: you're trying to change the branch from under the working directory which is invariably bad news	21:32
	← GodEater left	21:32
SethRobertson	aristidesfl: No. You may trash the contents, but the branch will still exist.	21:32
mick_home_	I just want to make the current directory head	21:32
	and get rid of what is in there	21:33
SethRobertson	mick_home_: Then...commit?	21:33
unclemantis	SethRobertson it shows files that will be removed. it shows nothing about the files that where changed	21:33
mick_home_	it fails	21:33
	commit works	21:33
SethRobertson	unclemantis: You already deleted the files which were changed. There is only one part left.	21:33
mick_home_	just: git push origin fails	21:33
	→ xerora joined	21:33
unclemantis	i have not run git clean -fdx yet	21:34
cmn	not just	21:34
mick_home_	cmn: I'm listening	21:34
	:)	21:34
cmn	it tells you you're trying to do something bad to the remote repository and if you really want to do it	21:34
SethRobertson	unclemantis: If you are not sure you want to delete those untracked files, then stash them, reading the paragraph ABOVE the one you are at. If you are sure, then delete them.	21:34
	→ notjohn joined	21:34
cmn	then you should set a config variable	21:34
	which git tells you about	21:34
	and you quoted	21:34
mick_home_	it cuts off though	21:34
	← Heimidal left	21:34
	← n8o-mba left	21:34
mick_home_	it says set it on the remote server	21:35
unclemantis	i am sure i want to delete the files. I am talking about the files that i CHANGED.	21:35
cmn	get a better terminal	21:35
	← Invader left	21:35
mick_home_	where is the config stored on the server?	21:35
cmn	yes, you need to set it on the remote	21:35
mick_home_	cmn: windows sucks	21:35
cmn	it's stored in the repo	21:35
mick_home_	ah ok	21:35
SethRobertson	unclemantis: What does `git status` say? Anything except untracked files?	21:35
	→ GodEater joined	21:36
	← GodEater left	21:36
	→ GodEater joined	21:36
	← revagomes left	21:36
	← dv310p3r left	21:36
	← marthinal left	21:36
	→ n8o-mba joined	21:36
unclemantis	https://gist.github.com/c672d912a051447dfb03	21:36
mick_home_	hmm, I just did a git rm where the server is	21:36
	that might help	21:36
statusfailed	SethRobertson: Ah, all sorted. Thanks for getting me on the right track!	21:37
	mick_home_ tries again	21:37
	→ Heimidal joined	21:37
SethRobertson	unclemantis: So, you do not have any uncommitted changes EXCEPT for the files git knows nothing about. You care 1) ignore them 2) commit them 3) delete them or 4) stash them.	21:37
	statusfailed: !thanks	21:37
gitinfo	statusfailed: Feeling thankful? Type "ExampleUser++", and ExampleUser will score karma points at http://carmivore.com (our preferred way to objectify self-worth). There's really no point to thanking me – instead, why not thank the person who made me help you?	21:37
cmn	mick_home_: how would that help?	21:38
	← amstan left	21:38
cmn	does that have anything to do with the error message git gave you?	21:38
	← t4nkd left	21:38
	← jonesy left	21:39
statusfailed	SethRobertson++	21:39
	→ eletuchy joined	21:39
unclemantis	i guess delete them	21:39
	← zulax left	21:39
unclemantis	ok i think i am back to normal	21:40
	→ tatsuya_o joined	21:40
unclemantis	SethRobertson sorry to sound like a whiny but	21:40
	i really should find a local college and take a course on subversioning	21:40
bremner	subversionin?	21:41
SethRobertson	You should read !book	21:41
gitinfo	There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !vcbe and !parable	21:41
cmn	is that keeping versions of versions?	21:41
	→ diegoviola joined	21:41
	→ giallu joined	21:42
	→ teweWORK joined	21:43
aristidesfl	"<SethRobertson> aristidesfl: No. You may trash the contents, but the branch will still exist. " ok but if the head will be moved to other branch, there will be no reference for the old branch?	21:45
	→ ericmuyser joined	21:45
cmn	a branch /is/ a reference	21:46
	← fairuz_ left	21:46
SethRobertson	The old commits may be unreferenced. The branch "name" will still exist.	21:46
	← icwiener_ left	21:46
	← jehoshua02 left	21:47
unclemantis	SethRobertson thank you	21:47
	← tommcdo left	21:48
SethRobertson	!thanks	21:48
gitinfo	Feeling thankful? Type "ExampleUser++", and ExampleUser will score karma points at http://carmivore.com (our preferred way to objectify self-worth). There's really no point to thanking me – instead, why not thank the person who made me help you?	21:48
unclemantis	SethRobertson++	21:48
	→ Davey_ joined	21:49
unclemantis	did it work?	21:49
	← bgerber left	21:49
	→ Elisha joined	21:49
	→ h0bbit joined	21:49
SethRobertson	Let us assume so	21:49
unclemantis	:)	21:49
cmn	wow, you're really racking up the karma	21:50
	ChanServ set mode: +v	21:50
SethRobertson	Yes, but the most important question is left unanswered. Am I number 1?	21:50
RandalSchwartz	you are number 6, I am number 2.	21:50
	"but who is number 1!"	21:51
cmn	is there a global tracker somwehere?	21:51
Elisha	hello people, I have problem pushing to the server ... 411, err=22, I tried setting up postBuffer, but still not working	21:51
	← notjohn left	21:51
	→ bgerber joined	21:51
SethRobertson	cmn: Not that I have found.	21:51
	Elisha: Please !repro	21:51
gitinfo	Elisha: Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.	21:51
	→ gitter joined	21:51
	← gusnan left	21:52
Elisha	< 1.7.2.5 debian https://gist.github.com/998bdd7a6d3ac184e9a6	21:52
	→ DrHouse\|TARDIS joined	21:52
cmn	well, you're number one in my local ego tracker	21:52
	← giallu left	21:53
	→ fairuz_ joined	21:53
	← h0bbit left	21:53
	SethRobertson blushes	21:53
cmn	and for the record, I mean my small ruby script that asks carmivore	21:54
SethRobertson	Elisha: What command did you use	21:54
Elisha	git push	21:54
	→ spiette_ joined	21:54
SethRobertson	Elisha: `git remote -vv` What scheme are you using? ssh? http?	21:54
	→ giallu joined	21:55
SethRobertson	cmn: Oh, I thought you were telling me that you thought I had the biggest ego around. Having the most karma is almost as good.	21:55
	← stuntmachine left	21:55
cmn	it's actually called "ego watch", because if I ever put it on the 'net, it'd be where you go to watch how big your ego is allowed to be	21:56
	← starlays left	21:57
Elisha	SethRobertson: http	21:57
	← dcarper left	21:58
	→ h0bbit joined	21:58
	→ notjohn joined	21:58
SethRobertson	Elisha: Well, either your client or server gits need to be upgraded, or the network in between is trying to do some ad injection or is otherwise changing the stream.	21:58
RandalSchwartz	maybe a bad ssh	21:59
SethRobertson	I know that was completely imprecise, but I'm not sure how to be more precise	21:59
	→ macmartine joined	21:59
RandalSchwartz	ssh that-host date	21:59
	should print just the date	21:59
SethRobertson	Since it is not using ssh, that is irrelevant	21:59
RandalSchwartz	some people break their login scripts	21:59
	oh - sorry	21:59
	git proto?	21:59
SethRobertson	http	21:59
Elisha	its on the same machine SethRobertson, but there's a proxy in between maybe that's it	21:59
RandalSchwartz	this is rapidly spiraling down :)	21:59
	← linusoleander left	22:00
cmn	the proxy could certainly be causing problems, does curl know about it?	22:00
SethRobertson	Elisha: If possibly, avoid using http:// for a local host communication	22:00
	→ Samoi joined	22:00
	← Samoi left	22:00
	→ Samoi joined	22:00
	← bartek left	22:01
	← fedesilva left	22:02
	→ Heimidal_ joined	22:02
	→ cdrane joined	22:02
	← btanaka left	22:02
	← zomothesweeper left	22:02
	→ vervic joined	22:03
	← _iron left	22:03
	→ dcarper joined	22:03
	← Heimidal left	22:03
mick_home_	i feel cursed	22:03
	i've removed the repo - figured I'd start from scratch, new repo and all	22:03
	same error	22:04
	wt...f	22:04
	sigh	22:04
cmn	doing the same thing is going to have the same consecuences	22:04
	see Einstein re insanity	22:04
	← Heimidal_ left	22:04
	→ Heimidal joined	22:04
	← oc80z left	22:04
	← soulcake left	22:04
	→ Heimidal_ joined	22:05
	← Heimidal left	22:05
mick_home_	cmn: under config, what do I need to set it to?	22:05
	receive.denyCurrentBranch = ....	22:05
	← mishok13 left	22:05
mick_home_	the error didn't really say	22:05
cmn	man git config	22:05
gitinfo	the git-config manpage is available at http://jk.gs/git-config.html	22:05
mick_home_	it just said Access Denied	22:05
	→ tapout joined	22:06
cmn	but make sure you know what that means	22:06
SethRobertson	You have to change the server configuration to allow/deny on the server.	22:06
	→ oc80z joined	22:06
cmn	yeah, instead he tried recreating a non-bare repo	22:06
SethRobertson	Oh dear	22:06
	I'll just let you handle this one, then.	22:06
tapout	hello, i'm trying to use git gui (msysgit). Whenever I go to clone .. ssh://root@host.com/root/git/ ... it complains	22:06
	fatal: Couldn't find remote ref HEAD	22:06
	← txomon left	22:07
tapout	the weird thing is, if i reboot into ubuntu, no issues are found	22:07
mick_home_	tapout: welcome to my world	22:07
SethRobertson	tapout: Is this a brand new repo?	22:07
tapout	yes	22:07
	← shergill left	22:07
mick_home_	linux -> np	22:07
	→ diogogmt joined	22:07
SethRobertson	tapout: Once you push your first commit up, that will stop happening	22:07
	→ frogonwheels joined	22:07
mick_home_	windows -> omgwtfdiediediestabstabstab	22:07
tapout	SethRobertson, it won't even clone so I can do it	22:07
	← cannonball left	22:08
	mick_home_ tries this again	22:08
SethRobertson	tapout: Well...can you push your first commit from linux and the proceed from there on windows?	22:08
cmn	your problem isn't one that has to do with OS	22:08
	← gitter left	22:08
cmn	you're trying to do things git considers bad	22:08
	which is why it's complaining	22:08
tapout	oh because it's brand new, there isn't a HEAD.. so i should push on linux and it will have a head.. ok trying	22:08
	← wald0 left	22:08
	← jwmann left	22:09
	→ gittersly joined	22:09
	← h0bbit left	22:09
	← gittersly left	22:09
tapout	Perhaps you should specify a branch such as 'master'.	22:09
	→ wald0 joined	22:09
tapout	did they change something in git? hrmm	22:10
	or maybe i stopped doing it correctly, first time i've experienced this	22:10
	→ gitterslys joined	22:10
SethRobertson	tapout: git push origin master	22:10
	→ Heimidal joined	22:10
SethRobertson	This is as it always has been	22:10
mick_home_	ok failed again	22:10
SethRobertson	Actually, git push -u origin master	22:10
	what error?	22:10
cmn	mick_home_: what are you actually doing?	22:11
SethRobertson	mick_home_: As in !repro	22:11
gitinfo	mick_home_: Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.	22:11
	← Heimidal_ left	22:11
mick_home_	now I just get access denied	22:11
	→ zsero joined	22:11
	← dcarper left	22:11
cmn	when you do what?	22:11
	gitinfo set mode: +v	22:11
cmn	as in yeah, post a transcript	22:12
	← cdrane left	22:12
mick_home_	ok	22:12
tapout	yay... thank you!	22:14
	← RaptorX left	22:14
SethRobertson	!thanks	22:14
gitinfo	Feeling thankful? Type "ExampleUser++", and ExampleUser will score karma points at http://carmivore.com (our preferred way to objectify self-worth). There's really no point to thanking me – instead, why not thank the person who made me help you?	22:14
	→ bronson joined	22:14
	→ gitter2sly joined	22:14
	gitinfo set mode: +v	22:15
mick_home_	ok, I guess after I rm'ed the repo from /srv/git/myrepo then I created another (mkdir /srv/git/myrepo && cd !$ && git init && chmod -R 777 /srv/git/myrepo)	22:15
gitter2sly	hello	22:15
mick_home_	I now can't clone the repo	22:16
	access denied	22:16
canton7	mick_home_, did you git init --bare ?	22:16
	→ h0bbit joined	22:16
mick_home_	--bare <-- shit	22:16
	forgot	22:16
	← tgunr left	22:16
	← Raging_Hog left	22:16
cmn	wtf? I mentioned bare vs non-bare as soon as you came in	22:17
	it's git's error message	22:17
	← vervic left	22:17
gitter2sly	I have only one branch, but I have commits that are not in it	22:17
	and none is marked as unreachable	22:17
cmn	tags? remote branches?	22:18
gitter2sly	I'd like to get rid of them but I don't see how	22:18
	I don't have any tags	22:18
RandalSchwartz	remote branches?	22:18
	git branch -avv	22:18
tapout	should git gui (msysgit) take forever to push?	22:18
	we're talking like 5 minutes already	22:19
gitter2sly	when I do git branch -a gives me	22:19
	← VladGh left	22:19
gitter2sly	*master	22:19
cmn	tapout: that depends on how much you're pushing and which protocol	22:19
tapout	ssh and only the YII framework (192 files)	22:19
cmn	and how crappy your uplink is	22:19
gitter2sly	git rev-list master gives me 93 lines	22:19
	→ shergill joined	22:19
gitter2sly	git rev-list --all gives me 173 lines	22:19
RandalSchwartz	probably HEAD history	22:20
	or master history	22:20
gitter2sly	and git fsck --unreachable gives me nothing	22:20
	← mandric left	22:20
RandalSchwartz	oh wait, those aren't in git rev-list --all ?	22:20
gitter2sly	yes they are	22:20
RandalSchwartz	gitter2sly - how about "git rev-list master@{2}"	22:20
	→ tgunr joined	22:20
RandalSchwartz	vs "git rev-list master"	22:21
cmn	if they're in the reflog, they're reachable, but rev-list --all shouldn't show them	22:21
RandalSchwartz	hmm	22:21
	→ lotsofcode joined	22:21
mick_home_	weird, git pull -v --progress "origin"	22:21
	← Jay_Levitt left	22:21
gitter2sly	RandalSchwartz, this tells me	22:21
cmn	how about git log --graph --annonate --oneline --all ?	22:22
mick_home_	says: Access Denied then it grabs it and says "success"	22:22
gitter2sly	fatal: Log for master only has 2 entries	22:22
cmn	mick_home_: you should probably look into using !gitolite	22:22
gitinfo	mick_home_: Want to host as many git repos (and users!) as you like, on your own server, with fine-grained access control? You want gitolite: https://github.com/sitaramc/gitolite	22:22
	← diogogmt left	22:22
	← n8o-mba left	22:22
mick_home_	well it did work	22:22
	but not sure why it says access denied	22:22
	← g105b left	22:23
cmn	no idea, w/o a transcript, we can't say	22:23
mick_home_	I can't even see anything under /var/log/messages	22:23
gitter2sly	RandalSchwartz,what does this syntax mean?	22:23
	master@{2]	22:23
	← jacekmigacz left	22:23
RandalSchwartz	the second most recent version of "master"	22:23
	I was thinking the reflog had some more commits	22:24
	← ajt left	22:24
gitter2sly	how's that? master has versions?	22:24
cmn	mick_home_: why would that be under the messages logfile?	22:24
RandalSchwartz	it has a path through time	22:24
	← skylord5816 left	22:24
RandalSchwartz	!man git-ref-log	22:24
	→ diogogmt joined	22:24
RandalSchwartz	!man git-reflog	22:24
gitinfo	the git-reflog manpage is available at http://jk.gs/git-reflog.html	22:24
mick_home_	it started working and the error went away	22:24
	just dooing "git pull origin"	22:25
	→ n8o-mba joined	22:25
gitter2sly	ok	22:25
mick_home_	cmn: access denied - figured if it was fs related it would write to syslog	22:25
	← ph^ left	22:25
	→ Dmole joined	22:25
RandalSchwartz	try "git name-rev --all"	22:25
	see if it shows any additional names	22:25
	← shergill left	22:25
gitter2sly	now, with one given commit, how can I know how it is reachable ?	22:25
cmn	that doesn't seem very scalable	22:25
	→ shruggar joined	22:26
thiago	gitter2sly: what exactly do you want to do?	22:26
mick_home_	i g/g thanks guys	22:26
	← adammokan left	22:26
	→ amokan joined	22:26
	← joshsmith left	22:27
gitter2sly	that gives me several names	22:27
	original/refs/heads/master~4	22:27
mikepack	How can I update my (github) fork to pull in changes from the parent repo?	22:27
	(without deleting/recreating the fork...obviously)	22:27
gitter2sly	undefined	22:27
	and master~32	22:28
	← phishy left	22:28
gitter2sly	and the same with other numbers	22:28
cmn	mikepack: it won't do it automatically	22:28
gitter2sly	I removed a file using filter-branch	22:28
cmn	mikepack: your fork is your own repo	22:28
	if you want to update it, you push to it	22:28
gitter2sly	that created a branch	22:28
mikepack	cmn: I figured, do I set it up as a remote repo?	22:28
gitter2sly	and I'm trying to get rid of the orl one	22:29
	→ mcluskydodallas joined	22:29
gitter2sly	old	22:29
RandalSchwartz	aha - so you have origina!	22:29
	← diogogmt left	22:29
cmn	mikepack: yes	22:29
RandalSchwartz	git branch -D original/refs/heads/master	22:29
mikepack	cmn: I want to pull upstream changes	22:29
RandalSchwartz	but once it's gone, it's GONE	22:29
	so be careful	22:29
cmn	mikepack: so pull upstream changes	22:29
	← h0bbit left	22:29
mikepack	cmn: Damn you're smart :)	22:29
cmn	keep in mind that git builds a network of repos so there's no need to keep your up to date constantly	22:30
gitter2sly	RandalSchwartz, yes you seemed to have a good track	22:30
	→ diogogmt joined	22:30
cmn	if you need changes from upstream you can just grab them from there	22:30
gitter2sly	that's ok I can recreate it	22:30
cmn	and when you have differences, you can push them to your repo	22:30
RandalSchwartz	I've been hacking git since 1997. :)	22:30
	err... 2007	22:30
	grr	22:30
	← mick_home_ left	22:30
RandalSchwartz	no wait... april 2005	22:30
	wow, has it been that long	22:31
cmn	wanted: 10+ years of git experience	22:31
RandalSchwartz	my git google tech talk was in 2007	22:31
	and the most recent one a few weeks ago	22:31
gitter2sly	:)	22:31
kevlarman	RandalSchwartz: was it as awesome as linus's git tech talk?	22:31
RandalSchwartz	linus talked about what git wasn't	22:31
gitter2sly	error: branch 'original/refs/heads/master' not found	22:32
RandalSchwartz	I followed up and talked about what git was	22:32
	← n8o-mba left	22:32
RandalSchwartz	oh - it should be in refs/original	22:32
	maybe git branch -D refs/original/heads/master	22:33
	I dunno. I just go into .git and delete manually :)	22:33
canton7	gitter2sly, read the very bottom section of man git-filter-branch	22:33
gitinfo	gitter2sly: the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html	22:33
gitter2sly	branch not found	22:33
	where can I find it in .git?	22:34
cmn	under its name	22:34
canton7	gitter2sly, the line from the manpage says git for-each-ref --format="%(refname)" refs/original/ \| xargs -n 1 git update-ref -d, but make sure you also follow the next steps	22:34
SethRobertson	`find .git -name master` will find it (among others) unless it is packed	22:34
	→ bauruine joined	22:34
RandalSchwartz	how did you end up running filter-branch anyway?	22:35
	what purpose?	22:35
	→ h0bbit joined	22:35
	← rzec left	22:35
	← giallu left	22:35
	→ rcreasey joined	22:36
	→ Salonce joined	22:36
tapout	git gets to: Writing objects: 30% (656/2186), 1.38 MiB \| 2.74 MiB/s and it doesn't update after that	22:36
	what's up?	22:36
gitter2sly	.git/logs/refs/heads ?	22:36
	the one in logs?	22:36
rcreasey	random question: if I define a bunch of remotes to a local git repository, is there any way to push those references to another remote so that if someone then clones that repository, they can clone the defined remotes I had in that repository?	22:37
	→ EricInBNE joined	22:37
canton7	rcreasey, they can clone with --mirror i guess	22:37
	← MANCHUCK left	22:37
rcreasey	okay, so if my origin is say github and the other remotes i have are mirrorA, mirrorB and mirrorC on my local repo	22:38
cmn	or commit a config file	22:38
rcreasey	how do I push those to github	22:38
	← teweWORK left	22:38
	← diogogmt left	22:39
	← mansour left	22:39
diegoviola	git push origin [branch name] ?	22:40
	→ diogogmt joined	22:40
cmn	push what?	22:40
Retistic	hod can you check how many commits a local branch is ahead of another local branch?	22:40
	how*	22:40
	→ ajt joined	22:40
cmn	with status, branch -vv also tells you	22:40
gitter2sly	a file containing a password	22:40
	→ thursday joined	22:40
	→ n8o-mba joined	22:41
cmn	though you do have to set up the tracking	22:41
gitter2sly	oups that is an answer to an old question:)	22:41
cmn	rev-list brancha..branchb will give you the one side, branchb..brancha will give you the other	22:41
Retistic	cmn: is there a way to compare 2 branches though, other than git diff	22:41
	← Heimidal left	22:41
Retistic	that just shows the last commit for each	22:42
cmn	what do you want to compare about those branches?	22:42
	then each has one differing commit	22:42
	→ ISF joined	22:42
	← lyddonb left	22:42
	← aristidesfl left	22:42
cmn	brancha...branchb should show both of those	22:42
	← tewecske left	22:42
Retistic	essentially i have a local master and another branch. i didn't delete the other branch and i want to make sure i merged everything into master and that there are no other changes on that branch	22:42
	← dnjaramba left	22:43
	→ chx joined	22:43
gitter2sly	ok I'm kind of lost	22:43
	← churp_ left	22:43
cmn	so that command told you there are	22:43
chx	hi. git branch --contains <commit> only lists local branches not remote branches. is it possible to find remote branches? they have fetched , they are just not checked out.	22:44
cmn	you might find git log --graph --oneline --annotate brancha...branch more understandable	22:44
	chx: add -a	22:44
	→ dnjaramba joined	22:44
Dmole	is there a way to "hide" a branch in gitolite?	22:44
	→ LeMike joined	22:44
cmn	no	22:45
	← h0bbit left	22:45
cmn	not in any straightforward way, at least	22:45
chx	cmn: but will that show remote branches ..?	22:45
cmn	that's a git limitation	22:45
	← queequeg1 left	22:45
cmn	chx: that's what -a means	22:45
	→ mandric joined	22:45
chx	cmn: oh	22:45
	→ xnite joined	22:45
	← koo4 left	22:45
cmn	Dmole: you can tell git to limit itself to a namespace, but not to outright hide one	22:46
	← ko1 left	22:46
	← k1o left	22:46
	← zsero left	22:46
cmn	patches or efforts to advance that issue are welcome	22:46
	→ linusoleander joined	22:46
Dmole	cmn: I was thinking something like this "config gitolite-options.deny-repo = 1"	22:47
	→ ko1 joined	22:47
	→ queequeg1 joined	22:47
cmn	that name suggests hiding a repo, not a branch	22:47
	← n8o-mba left	22:47
	→ koo4 joined	22:47
	→ k1o joined	22:48
Dmole	cmn: yes I was hoping there was something like that for branches	22:48
cmn	if you don't want particular people knowing about a repo, don't add them to the access list	22:48
	no, as I said, it's a git thing, and efforts to provide git-upload-pack with a filtering mechanism are welcome	22:49
	← diogogmt left	22:49
	→ diogogmt joined	22:49
	← ThomasLocke left	22:49
	← MartinCo left	22:50
	→ MartinCo joined	22:50
chx	cmn: git log --graph --oneline --annotate branch -a fatal: ambiguous argument 'branch': unknown revision or path not in the working tree.	22:51
gitter2sly	ok I don't have anything in .git that is named master	22:51
chx	cmn: git log --graph --oneline --annotate -a fatal: unrecognized argument: --annotate	22:51
gitter2sly	but I still have that weird branch	22:51
cmn	chx: huh?	22:51
	← mordof left	22:51
cmn	that was for someone else	22:52
	← Salonce left	22:52
	← john2496 left	22:52
chx	git version 1.7.8.4	22:52
	oh.	22:52
	sorry	22:52
	← eletuchy left	22:52
canton7	chx, the -a was meant for you, but not the "git log"	22:52
	→ Salonce joined	22:52
Dmole	cmn: hmm maybe it would make more sense to add something like "-R refs/tags/v[0-9] = @staff" and have it limit the first check~	22:52
chx	ah ha!	22:53
	i see.	22:53
	← rendar left	22:54
	← diogogmt left	22:54
	→ h0bbit joined	22:54
	← Shusshu left	22:54
	← chrisdev left	22:54
gitter2sly	git reflog gives me a bunch of HEAD{x}: checkou: moving from ...	22:54
	can this log prevent some commits to be considered as orphans ?	22:55
	→ diogogmt joined	22:55
	← chx left	22:55
	→ john2496 joined	22:55
	← oc80z left	22:56
cmn	that's the reason for its existance	22:56
	it's to stop you from destroying your data	22:56
	← crisp left	22:56
cmn	Dmole: but gitolite can't control that	22:57
Dmole	cmn: yet :)	22:57
gitter2sly	ok	22:57
cmn	right, as I said, efforts yada yada	22:57
	→ oc80z joined	22:57
	→ alberto56 joined	22:57
gitter2sly	so how can I get rid of it?	22:57
	→ mastro joined	22:57
canton7	gitter2sly, read the bottom of man git-filter-branch, like I told you	22:58
gitinfo	gitter2sly: the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html	22:58
gitter2sly	sorry I didn't catch that one	22:59
	← diogogmt left	23:00
	← leavittx left	23:00
chilversc	what I'd like to do is shut git gui's gc warning up, that thing really annoys me	23:00
	→ diogogmt joined	23:01
	← kermit left	23:01
	← spal left	23:02
	← martinjlowm-lptp left	23:02
	← john2496 left	23:03
	→ joshsmith joined	23:03
gitter2sly	ouch	23:03
	I cloned the repo to some other place, and I think that's it	23:04
	← h0bbit left	23:04
	← whitman left	23:04
	← alan_w left	23:05
gitter2sly	so the clone isn't really a clone	23:05
	← Vile left	23:05
	→ alazar1 joined	23:06
	→ Heimidal joined	23:06
	← davidrussell left	23:08
	← diogogmt left	23:08
	→ h0bbit joined	23:08
	→ aristidesfl joined	23:08
cmn	-v?	23:09
	→ diogogmt joined	23:10
	→ kevwil joined	23:10
	← michaelh left	23:12
	← drizzd left	23:12
	→ DarkAR joined	23:13
	← macmartine left	23:13
	← themachine left	23:14
	← cakehero left	23:15
	← amokan left	23:15
gitter2sly	other question: I'm not sure how to specify the rev-list to commands like filter-branch	23:15
	← diogogmt left	23:15
gitter2sly	can I just type: git filter-branch --index-filter 'git rm ...' 28c0100 ?	23:16
	where 28c0100 is the id of one commit ?	23:16
	→ diogogmt joined	23:16
gitter2sly	I want to delete a file in all commits older than this one	23:16
	← texasmynsted1 left	23:16
	→ adammokan joined	23:17
cmn	you'd have to rebase --onto the rest of the branch on top of the rewritten one	23:17
	← alazar1 left	23:17
	→ w09x joined	23:17
	← fly9 left	23:17
gitter2sly	really? I thought it was possible to use filter-branch only on a subset of rev	23:17
	not sure where I read what though...	23:18
cmn	filter-branch acts on a branch or set of branches	23:18
Deck`	i have some trouble with git via http. I set up the repo in /home/user/proj.git and changed owner of files - "chown -R www-data proj.git". So everything works well except hooks. I noticed that it's not run. Should i solve my problem with apache and rights for files or git doesnt care?.	23:18
gitter2sly	I see	23:18
cmn	but the manpage can say a lot more	23:18
	← digitalfiz left	23:18
cmn	Deck`: are you using the dumb http protocol?	23:19
	→ bjohnson8591 joined	23:19
	← h0bbit left	23:19
Deck`	cmn i should. because i have to share access	23:19
	← alex__c2022 left	23:19
cmn	huh?	23:20
	the dumb http protocol is never a good solution	23:20
	look into !gitolite	23:20
gitinfo	Want to host as many git repos (and users!) as you like, on your own server, with fine-grained access control? You want gitolite: https://github.com/sitaramc/gitolite	23:20
Deck`	cmn i.e i use git protocol to public access (read-only) and http protocol for (read-write) for developers	23:20
	← diogogmt left	23:20
cmn	so are you using the smart or dumb http transport?	23:21
	→ diogogmt joined	23:21
	← tapout left	23:21
	← gavin_huang left	23:22
	← Heimidal left	23:22
Deck`	cmn dump	23:22
	→ Heimidal joined	23:22
Deck`	dumb*. googled	23:22
	→ h0bbit joined	23:22
cmn	then you better switch to the smart transport	23:23
	that'll give you hooks for sure, and much better speeds	23:23
Deck`	ye. will read the arcticle i found. i even didnt know about smart one	23:23
	← bjohnson8591 left	23:23
	← beatak left	23:23
cmn	the smart on is the actual git protocol	23:23
	the dumb http protocol is a horrible horrible thing	23:24
	← diogogmt left	23:24
	→ diogogmt joined	23:25
	← Destos left	23:26
	→ cdrane joined	23:28
	ChanServ set mode: +v	23:28
	cdrane → Obromios	23:29
	→ fayimora joined	23:29
	← diogogmt left	23:30
aristidesfl	"<SethRobertson> aristidesfl: No. You may trash the contents, but the branch will still exist. " ok but if the head will be moved to other branch, there will be no reference to the old branch?	23:30
cmn	didn't you already ask that?	23:30
	→ diogogmt joined	23:30
	→ eletuchy joined	23:31
aristidesfl	cmn: guess I did, but I lost the answer	23:31
	sory	23:31
cmn	when you say "head", do you mean HEAD or the branch tip?	23:31
	← lotsofcode left	23:31
aristidesfl	cmn I meant branch tip	23:31
cmn	if you move the branch tip, the old one will exist in the reflog for a while	23:31
aristidesfl	with reset, the head will still point to the same branch tip	23:32
cmn	depends on which options and parameters you give it	23:32
aristidesfl	reset —hard branch	23:32
cmn	that will make the currently active branch point to whatever "branch" points to	23:33
	← johnkpaul left	23:33
RandalSchwartz	--hard not -hard	23:33
	← Splash left	23:33
cmn	and update the index and the files int the workdir	23:33
	← ajt left	23:33
	← cnorman left	23:33
	← h0bbit left	23:34
	← diogogmt left	23:34
	→ ajt joined	23:34
SethRobertson	aristidesfl: (16:46:04) SethRobertson: The old commits may be unreferenced. The branch "name" will still exist.	23:34
	← w09x left	23:34
	→ diogogmt joined	23:34
	← DarkAR left	23:34
	→ drev1 joined	23:34
	sythe-afk → sythe	23:34
aristidesfl	SethRobertson: so basically all the commits done it that branch will be lost and the branch will look like "new"	23:35
cmn	RandalSchwartz: that's an m-dash, it's an increasingly common occurrence 'round these parts, some clients apparently think it's helpful to make two dashes into an m-dash	23:35
	← canton7 left	23:35
gitter2sly	hey	23:35
	← stepnem left	23:35
RandalSchwartz	damn clients	23:35
gitter2sly	with rebase I lost all the timestamp information	23:35
	si there any way to keep it?	23:36
RandalSchwartz	timestamp?	23:36
	it's still in each commit	23:36
cmn	it's kept unless you tell git otherwise	23:36
SethRobertson	aristidesfl: They are unreferenced (assuming no other tag or branch points at them) and can be found for 2+ weeks, unless you work hard at running commands to flush them.	23:36
RandalSchwartz	what did you filter-branch for, anyway?	23:36
	I missed that from before.	23:36
	or maybe you didn't answer me	23:36
cmn	removing a password	23:36
SethRobertson	aristidesfl: I have a reference on how to find them again	23:36
cmn	he answered later, you were away	23:36
RandalSchwartz	ok thanks	23:36
	yeah, I'm doing real work here, some of the time	23:37
	RandalSchwartz grins	23:37
	→ stepnem joined	23:37
RandalSchwartz	so if it's just to rm a file, that won't affect the commit timestamps	23:37
	those should be just brought forward.	23:37
	→ h0bbit joined	23:37
cmn	I think you're confusing people	23:38
	or maybe I am	23:38
RandalSchwartz	heh	23:38
	it's quite possible	23:38
gitter2sly	actually the rm didn't change the time	23:38
	RandalSchwartz has to leave in two minutes anyway	23:38
RandalSchwartz	well - the filter branch where you rm'ed the password	23:38
gitter2sly	but then I wanted to get the file back for all commits newer than a given one	23:38
	so I rebased it	23:38
	← eijk left	23:39
cmn	a rebase doesn't affect the author time	23:39
gitter2sly	and now all the commits of the new branch are shown "1 minute ago"	23:39
	← diogogmt left	23:39
gitter2sly	weird then...	23:40
	→ diogogmt joined	23:40
	→ johnkpaul joined	23:40
gitter2sly	what could change the time?	23:40
	oh I know	23:41
	I had to do a merge while rebasing	23:41
	that's the reason, isnt it?	23:41
	I did git rebase --onto 12355 68090 HEAD	23:42
	then I got a conflict, so I did git add myfile	23:43
	git rebase --continue	23:43
	← tktiddle left	23:43
gitter2sly	do I have any solution to keep my times?	23:44
cmn	no idea if --continue changes author time, --amend shouldn't	23:44
aristidesfl	hunk-by-hunk basis what does this means?	23:44
	← diogogmt left	23:44
	← anjen left	23:44
	← ericandrewlewis left	23:45
cmn	are you asking what a hunk is?	23:45
	→ diogogmt joined	23:46
	→ Goplat joined	23:47
	← h0bbit left	23:47
	← johnkpaul left	23:47
	← kevwil left	23:48
	→ macmartine joined	23:50
	→ h0bbit joined	23:52
	← OOPMan left	23:52
	← abetaha left	23:54
	→ rchavik joined	23:54
	← rchavik left	23:54
	→ rchavik joined	23:54
aristidesfl	cmn: yes	23:54
cmn	a hunk is each of the parts of a diff	23:54
	← airborn left	23:54
	→ OOPMan joined	23:54
cmn	so it's a very specific way of saying "piece by piece"	23:55
	← diogogmt left	23:56
nevyn	gitter2sly: rebase.. will make new commits. by definition	23:56
	→ diogogmt joined	23:56
cmn	but the author and commit dates are different values	23:56
nevyn	right.	23:56
cmn	doesn't rebase try to keep the author date?	23:57
nevyn	one would hope so.	23:57
cmn	that's the issue	23:57
	the dates seem to have reset	23:58
	→ Obromios_ joined	23:58
	← freezey left	23:58
nevyn	well...	23:58
	→ kryl99_ joined	23:59
	→ Obromios__ joined	23:59

Logs Search ←Prev date Next date→

Channels Documentation