IRCloggy #git 2010-10-28

Provider of IRC logs since 2005.

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

	2010-10-28
	→ blarson joined	00:01
	→ antivigilante joined	00:01
	→ MetaCosm_ joined	00:02
	← hakunin left	00:03
	→ noktoborus_ joined	00:03
	→ whaley joined	00:03
	← pgib left	00:03
	→ airborn joined	00:04
	← blarson left	00:04
	← airborn left	00:05
	← friskd left	00:07
	← NEOhidra left	00:08
	← parasti left	00:08
	← ka1ser left	00:09
	→ _Vi joined	00:11
	→ Uierth_ joined	00:12
	← ajpiano left	00:13
	← Uierth left	00:15
	← jelmer left	00:16
	→ jelmer joined	00:17
	← artefon left	00:18
	→ rbarraud joined	00:18
	→ QinGW joined	00:18
	← Yuuhi left	00:20
	→ xiong joined	00:20
	→ QinGW1 joined	00:20
	← QinGW left	00:22
	→ QinGW joined	00:22
	← bousquet left	00:24
	← QinGW1 left	00:25
	← rbarraud left	00:25
	← ehabkost left	00:25
	→ scarabx joined	00:25
	→ coderdad joined	00:26
	→ mithridates joined	00:26
	→ xrogaan joined	00:27
	← p3rg2z left	00:27
	← Evious left	00:27
	→ rbarraud joined	00:27
	→ p3rg2z joined	00:27
mithridates	I deleted a branch, added to stage and committed and then pushed everything but I don't know why it is still in the branch list (on website) but in my git branch it doesn't show it anymore	00:28
AAA_awright	mithridates: Unless the remote is a --mirror you have to explicitly delete the branch I think	00:30
	git push (remote) local:remote, or to delete, simply git push (remote) local: iirc	00:30
mithridates	AAA_awright: explain more please	00:30
AAA_awright	mithridates: Did you take a look at the git-push man page?	00:31
mithridates	yes, but let me see again . you said --mirror?	00:31
	→ tty2 joined	00:32
mithridates	AAA_awright: I got ya , thank you	00:32
AAA_awright	mithridates: If that's your public repository you probably want mirror = true set, check your configuration	00:33
	← markt left	00:33
mithridates	AAA_awright: yes it's my public repository, ok tnx buddy	00:33
AAA_awright	That's how mine is setup, at least	00:34
	← tjholowaychuk left	00:34
	← peterjmag left	00:36
	← envi left	00:36
	← p3rg2z left	00:37
	→ bronson_ joined	00:38
	← wsclock left	00:41
mithridates	AAA_awright: mirror = true worked! thank you , I'm wondering is it normal that git log --graph doesn't show branches with dash like what progit.org shows ?	00:45
	← archcoder left	00:45
AAA_awright	No clue about git log anything	00:46
	→ xiangfu joined	00:46
AAA_awright	The man page is the first thing I go to before passing any arguments to got log	00:46
mithridates	AAA_awright: I did that, the problem is that the output that I get is different from what man page shows or progit.org shows	00:47
	→ BiggFREE joined	00:47
	← justin-george left	00:47
AAA_awright	Describe how it's different then	00:48
	→ rayners joined	00:48
mithridates	AAA_awright: the dashes in the beginning of each line that shows merged branches or different branches are not in my output	00:49
	for me it just shows one flat \| charachter in the beginning of each line and it doesn't represent merged branches	00:50
	→ Heimidal joined	00:50
mithridates	not just merged, even it doesn't represent different branches, everything is flat in one order	00:50
	→ timH joined	00:51
	← t0rc left	00:51
timH	can someone help a git noob?	00:52
	i've got Git from the bottom up and I'm trying to learn but I also want to be productive at the same time	00:52
	so I can't just drop everything (unless I can't find help)	00:52
mithridates	AAA_awright: do I upload the out put?	00:52
AAA_awright	mithridates: Try git log --graph --all	00:53
mithridates	AAA_awright: same result	00:53
AAA_awright	Are you sure you have merged branches to graph?	00:53
	er, merge commits?	00:53
	← MetaCosm_ left	00:54
	← leitao left	00:54
mithridates	AAA_awright: yes I am sure	00:54
Ilari	BTW: Looks like Github is on the list of the sites that Firesheep supports. Be careful. :-)	00:55
Eridius	Ilari: github's already fixed themselves	00:55
Ilari	Ah, HTTPS?	00:55
Eridius	they already use HTTPS for important things. They just now use a secure cookie to verify the non-secure one during HTTPs operations	00:56
	→ amerine joined	00:56
Eridius	http://github.com/blog/737-sidejack-prevention	00:56
Ilari	Ah, that's one way to do it... Require verification for sensitive operations.	00:56
	← bitkiller left	00:57
	→ Slix` joined	00:57
	← spearce left	00:57
mithridates	AAA_awright: http://paste.org/pastebin/view/24108	00:58
	← alvaro_o left	00:59
mithridates	AAA_awright: the output is totally different from what it is in websites or manuals	00:59
AAA_awright	mithridates: I don't see any merges in that log	00:59
Ilari	Does it also support repo.or.cz (even if the connection I'm using is wired)?	00:59
AAA_awright	It looks like a linear history	00:59
	← RobertLaptop left	00:59
mithridates	AAA_awright: but it is merged, "To see which branches are already merged into the branch you’re on, you can run git branch --merged"	01:00
	→ xxoxx joined	01:00
mithridates	AAA_awright: the first line shows it "$ git branch -v"	01:00
AAA_awright	mithridates: They both point to the same commit though	01:00
	It looks like	01:00
mithridates	AAA_awright: sorry the second command git branch --merge	01:01
timH	i need to merge one branch into another but when I clone the master and run "git branch" it only shows master	01:01
AAA_awright	mithridates: Try git log --graph --parents --all	01:01
timH	how do I get that other branch?	01:01
	→ tungtn joined	01:01
	← rbarraud left	01:02
	← Amorphous left	01:02
mithridates	AAA_awright: result is flat again, the only thing that is changed is hash ( it's longer now )	01:02
	→ phantomcircuit_ joined	01:02
AAA_awright	mithridates: Merges will show three hashes, the first is the commit SHA1, the next is the parents. Merges will have two or more parents.	01:03
	→ JEEBcz joined	01:04
	→ converge joined	01:04
	→ Yuffster joined	01:04
	← phantomcircuit left	01:04
	→ JEEB__ joined	01:06
mithridates	AAA_awright: $ git log --pretty=format:"%h %s" --graph this command has this out put for me http://paste.org/pastebin/view/24109	01:06
	← zain left	01:06
	← SimonP86 left	01:06
mithridates	AAA_awright: shouldn't be like this http://progit.org/book/ch2-5.html ?	01:06
	→ rbarraud joined	01:07
AAA_awright	mithridates: Yours will look like that if it's a linear history without any merge commits	01:07
mithridates	that's not a linear history! is that a linear history?	01:08
	← towski left	01:08
	← JEEB left	01:08
AAA_awright	mithridates: Run git log --parents, how many commits do you have with multiple parents?	01:08
timH	so I just needed to pull the other branch into master	01:09
	← JEEBcz left	01:09
	→ aaron01 joined	01:09
converge	i have a project with more than 200 files, to push it to a remote repository, do i need do git add filename and git commit for every single file, or have another method?	01:10
mithridates	AAA_awright: 8 commit with multiple parents	01:10
	← Entonian left	01:10
mithridates	converge: I guess you can use git commit -a -m "commit note"	01:10
converge	mithridates, i dont need to do git add before?	01:11
aaron01	If I want to backup a repository and maintain the remote tracking, what is the best way to do this? seems like clone --mirror isn't maintaining remotes	01:11
	→ mithro joined	01:11
mithridates	converge: -a will do that for you	01:11
AAA_awright	mithridates: Can you paste that?	01:11
converge	mithridates, cool, tks! =)	01:11
mithridates	converge: yw	01:12
	AAA_awright: yes sure	01:12
	→ kanru joined	01:12
	← converge left	01:13
	→ converge joined	01:13
mithridates	AAA_awright: http://paste.org/pastebin/view/24110	01:13
	→ RobertLaptop joined	01:13
AAA_awright	mithridates: Those all have a single parent, it's a linear history	01:13
	← rbarraud left	01:14
mithridates	AAA_awright: how did you recognize it? because there is just one hash?	01:14
	→ Adaptee joined	01:14
	→ tangyin joined	01:14
AAA_awright	Yeah. A merge commit has more than two parents, by definition.	01:14
Eridius	no, it has two or more	01:14
	→ zain joined	01:14
AAA_awright	*two or more	01:14
	More than one parent	01:14
	← alberto56 left	01:15
mithridates	ok let me commit something, now I've merged them so if I commit something that commit will have more than one parent and no matter in what branch I commit it	01:15
	is that right?	01:15
AAA_awright	mithridates: Thing about refs, including branches and tags, is they are only pointers to a particular commit	01:16
	→ Amorphous joined	01:16
	← Amorphous left	01:16
	→ Amorphous joined	01:16
mithridates	AAA_awright: would you ask it in another way, I don't understand it	01:17
Eridius	mithridates: commits are made independent of the knowledge of what HEAD is pointing to	01:17
mithridates	I got it	01:18
converge	mithridates, sorry i closed the windows "sem querer", the command is git -a -s 'comment' ?	01:18
AAA_awright	mithridates: Both of your branches point to the same commit, there is no diversion anywhere on the line that would produce a pretty grap	01:18
	*graph	01:18
	With connecting lines	01:18
mithridates	converge: git commit -a -m "comment" # check manual for git commit	01:19
	AAA_awright: it's getting more clear now	01:20
	← intripoon left	01:20
	← Heimidal left	01:20
	← phantomcircuit_ left	01:20
AAA_awright	mithridates: A branch is literally a 41-byte file in .git	01:21
mithridates	so how can I make a commit with multiple parents?	01:21
	→ phantomcircuit joined	01:21
AAA_awright	mithridates: Check out cat .git/refs/heads/master	01:21
mithridates	there is a SHA1 hash code	01:21
	93fdd76781e04df500e4f8b65c9e777eed10e6e8	01:21
	← Slix` left	01:22
AAA_awright	That's the branch. When you make a commit, it's going to change that to point to the new commit.	01:22
mithridates	and another one for the other branch	01:22
AAA_awright	mithridates: use git-merge	01:22
mithridates	cat .git/refs/heads/new 4975cdbe2d742d18d4deb9539814f6c9994fc765	01:22
AAA_awright	mithridates: Then try "git log new"	01:22
	I think	01:23
mithridates	you mean do I use git merge for commit ? like this "git merge --commit "comment" "?	01:23
AAA_awright	Something like that, take a look at the man page	01:23
	← Vampire0_ left	01:24
mithridates	AAA_awright: ok ok	01:24
	tnx buddy for the time that you spend	01:24
	→ Slix` joined	01:24
	← slakmagik left	01:26
	→ MetaCosm_ joined	01:28
	→ spliff joined	01:28
	← coderdad left	01:29
	← _Vi left	01:30
	← MetaCosm left	01:30
	← Pacman_Home` left	01:30
	spliff → drubies_	01:30
	← Slix` left	01:31
	→ Pacman_Home` joined	01:31
	→ Slix` joined	01:31
	→ nbrochu joined	01:32
	→ aspotashev joined	01:33
aaron01	So is the only way to make an 'identical' copy of a repo (and refs) to use something outside git (cp/rsync/etc)? Or is there a git way to do this?	01:33
	← shad left	01:34
Eridius	git clone --mirror perhaps?	01:34
aaron01	the remote refs change..	01:34
Eridius	--mirror maps all refs	01:34
	← SerajewelKS left	01:35
	→ SerajewelKS joined	01:35
	← SerajewelKS left	01:35
	→ SerajewelKS joined	01:35
	← Leemp left	01:35
	← rson left	01:35
aaron01	Hmm. From looking at it (with my limited git knowledge) it looks like the remote refs point to a commit hash, rather than to a remote location after using --mirror	01:35
	← dec left	01:36
Eridius	remote refs never "point to" a remote location	01:36
	→ rson joined	01:36
Eridius	non-symbolic refs (which in most repos, means all refs except for HEAD) always point to an object (e.g. a hash)	01:36
SethRobertson	However, it is true that you don't inherit the remotes	01:36
aaron01	my wording is off. So "git remote show origin" is different	01:36
Eridius	true, the new repo only has a single remote, origin, which points back to the source	01:36
	→ Leemp joined	01:36
	→ coderdad joined	01:36
	→ stliu joined	01:37
	← stliu left	01:37
	→ stliu joined	01:37
	→ Vampire0_ joined	01:38
	→ gretch joined	01:38
SethRobertson	Nor is it possible in the general case to "copy" the remotes even if you wanted to. remotes may be relative	01:38
aaron01	That's ok in this case..	01:38
SethRobertson	Or at least not absolute from your perspective	01:38
Eridius	aaron01: you could always just do a mirror clone, then copy .git/config from the old into the new repo	01:38
SethRobertson	And refetch	01:39
	In any case, `git clone --mirror` will make the data under version control identical. The metadata which is not under version control may be different	01:40
	← converge left	01:40
aaron01	Right. My main concern in using something 'outside' git is in case the repo is being changed while it is being copied. But just copying the .git/config seems safe enough	01:41
	→ ShooterMG joined	01:41
	← ShooterMG left	01:41
aaron01	Thanks Eridius / SethRobertson	01:42
SethRobertson	Note after copying .git/config you lose the mirror (or any) relationship between the backup and the original source	01:42
Eridius	aaron01: if you do run into issues with copying .git/config (e.g. if git does an incremental write instead of an atomic one), you may be able to use `git config --list` in the original repo and use that to reconstruct the file in the new	01:43
SethRobertson	Unless you tell the original about a remote named "original" which is a mirror that points to itself. I would experiment before doing this particular bit of recursive referencing	01:43
	← nakkor left	01:43
Eridius	actually, belay that. that'll print global and system config values too	01:43
	although I think you can use the --file switch to limit it to just .git/config	01:44
aaron01	The use case for this is that the original repo will either not exist or will be replaced	01:44
	→ dec joined	01:44
aaron01	So there doesn't need to be any maintained relationship between the mirrored copy and the original	01:44
Eridius	you may also want to copy over .git/info/exclude and any other important metadata (e.g. grafts), depending on if you care about that stuff	01:45
mithridates	AAA_awright: I did what you said, but it is still producing same result	01:45
	→ verto joined	01:45
mithridates	graph is not pretty	01:46
	oh yeah	01:46
	AAA_awright: :D it is pretty now	01:46
	→ patnotz joined	01:46
mithridates	I added --commit to the git merge <branch>	01:47
	← jelmer left	01:47
mithridates	yesss, I can see two SHA1 in front of that commit	01:48
Eridius	mithridates: you realize that --commit is only useful to override an earlier --no-commit flag, right?	01:48
AAA_awright	Eridius: To override a fast forward	01:48
	I think	01:48
	Or is that another option...	01:48
mithridates	I'm still wonder	01:49
Eridius	that's --no-ff	01:49
AAA_awright	Yep	01:49
mithridates	now what's the difference between this commit that a merge made with a commit that I've made by git commig -a -m "comment"	01:49
	?	01:49
Eridius	the merge makes a commit with 2 (or more) parents. `git commit` will only have 1 parent	01:50
AAA_awright	mithridates: git commit doesn't make merge commits...	01:50
mithridates	was that commit under the branch	01:50
	so git commit make a commit just about a new update in a branch	01:50
Eridius	both `git merge` and `git commit` will update the currently checked-out branch to point to the newly-created commit	01:50
mithridates	so they are not different? and the last one is the HEAD no matter if it's from git merge or git commit	01:51
	Eridius: but I did git commit but it didn't make git log --pretty work	01:52
	I mean it didn't assign two parents to that commit	01:53
Eridius	mithridates: err yes, like I said, `git commit` doesn't make a merge commit	01:53
	← bronson_ left	01:53
Eridius	you seem to be conflating making a merge commit with updating the current branch, when those are two completely separate actions	01:53
	→ psoo joined	01:54
	← drubies_ left	01:54
mithridates	Eridius: I got it now, it's so interesting :) thanks a lot Eridius and AAA_awright	01:54
	→ M1- joined	01:54
	→ RobertMeta joined	01:56
	→ checkers joined	01:59
checkers	hi, how do I show what was changed in an arbitrary past commit?	02:00
	← amerine left	02:00
	← Uierth_ left	02:00
Eridius	git show $COMMIT	02:00
	→ Uierth joined	02:01
	→ shad joined	02:02
	← Khisanth left	02:02
	→ quesada3 joined	02:05
	← M1- left	02:05
	← verto left	02:05
	→ sako joined	02:06
	← quesada2 left	02:07
	→ liu joined	02:08
	→ frogonwheels joined	02:08
	→ Khisanth joined	02:09
	→ rossand joined	02:10
	← RobertMeta left	02:10
	→ nullvariable joined	02:11
	← merlin83 left	02:12
	← liu left	02:14
	→ liu joined	02:14
	← phantomcircuit left	02:15
	← tty2 left	02:15
	← warthog9 left	02:20
	→ brntbeer joined	02:20
	→ RobertMeta joined	02:22
	← MetaCosm_ left	02:25
	← keystr0k left	02:25
	→ alberto56 joined	02:27
	→ Heimidal joined	02:28
	← aah left	02:28
	← liu left	02:29
	→ warthog9 joined	02:32
	← Bass10 left	02:35
vdv	what's the difference between "git branch master-copy master" and "git branch master-copy remotes/origin/master"? i.e. i understand that first creates a branch from local master branch, and second from remote one, but which important consequences have both?	02:39
	← andyburke left	02:40
	← Eaven left	02:40
	← kukks left	02:41
DrNick	the second sets the branch up to track the remote branch	02:42
SethRobertson	If master has diverged from remotes/origin/master, there will be a difference	02:42
	← aaron01 left	02:42
SethRobertson	But what DrNick said is also true	02:43
vdv	then in first case i must always rebase in master-copy after rebasing in master?	02:43
	← aspotashev left	02:43
vdv	and also in second case git push will upload branch to origin, and that doesn't happen in first case, right?	02:45
	→ thedeeno joined	02:46
tangyin	Eridius: Hello, Eridius.I'm a new one on git and can you tell me if my master branch is ahead of the origin/master, what should I do?	02:46
	← nbrochu left	02:47
	← msieradzki left	02:47
SethRobertson	vdv: You have to rebase in master-copy either way, either to get the changes from master or from remotes/origin/master	02:47
vdv	SethRobertson, how behaves git push in both cases?	02:48
	same?	02:48
SethRobertson	No, because master-copy is not tracking in the first case (though you could configure it to track) push will not work	02:48
	← Heimidal left	02:48
SethRobertson	You could merge/rebase to get the changes	02:48
Eridius	tangyin: if it's ahead of origin/master, that means you have local commits that aren't on the remote repo	02:49
	tangyin: if you want to update the remote repo to have your commits, you can `git push` them	02:49
	← smsfail left	02:49
	← sebersole left	02:49
	← welly left	02:49
vdv	SethRobertson, will git push create branch on origin if that branch is not listed in "git branch -a" as remotes/origin/mybranch?	02:49
tangyin	Eridius: But I want to push some code to origin,so how can I revert the ahead commit?	02:50
SethRobertson	vdv: Not unless you specifically ask it to, in one of several ways	02:50
	← Elfe left	02:50
	→ Elfe_ joined	02:50
	→ Unode joined	02:50
	← toddnine left	02:51
	→ sh1mmer joined	02:51
Eridius	tangyin: I'm not sure what you mean. Why would you revert the commit you're about to push?	02:51
vdv	SethRobertson, bit "git branch remotes/origin/mybranch; git push" will do that, right? (i.e. i'm trying to create a remote branch)	02:52
	→ timj_ joined	02:52
	← saurik left	02:52
SethRobertson	You want to say: `git push origin newbranch`	02:53
	→ saurik joined	02:53
	← kadoban left	02:53
tangyin	Eridius: I don't want to push these ahead commit, I just want to revert them.	02:54
SethRobertson	vdv: That will not establish a tracking relationship, of course. You can do that separately	02:54
	tangyin: I think you are using "ahead commit" in a way we do not understand	02:54
	tangyin: Are you trying to say you have several commits that have not been pushed, but you only want to push one (the last?) of those commits?	02:55
Eridius	tangyin: you want to throw away any local commits to make your local "master" branch identical to origin/master?	02:55
	← timj left	02:55
tangyin	Eridius: yes	02:55
	→ kadoban joined	02:56
tangyin	SethRobertson: I'm so sorry for my English :(	02:56
	→ chrisshattuck joined	02:56
SethRobertson	tangyin: No problem. `git reset --hard remotes/origin/master` will do what you want. You may want to make a backup `clone` of your directory beforehand	02:57
tangyin	SethRobertson: Yes, that's just I want.	02:57
	→ merlin83 joined	02:58
	← programble left	03:01
	← Dickhead left	03:03
vdv	i'm confused, i do changes in a branch, but same changes are in master	03:03
	→ dl9pf_ joined	03:03
vdv	*appear in master	03:03
SethRobertson	Without a merge or push?	03:03
vdv	yes	03:04
	→ Bass10 joined	03:04
	← dl9pf left	03:04
SethRobertson	You know that `git branch blah blah` does not change your current branch?	03:04
offby1	general-purpose git advice: 1) Read "Git For Computer Scientists" (http://eagain.net/articles/git-for-computer-scientists/); 2) Run "gitk --all"; 3) Learn about "git reflog"	03:04
SethRobertson	That is why people use the `git checkout -b` varient	03:05
	Check your branch with `git branch`	03:05
vdv	i do checkout after creating	03:05
	← gilaniali left	03:05
	← foocraft left	03:05
	← mbroeker left	03:05
vdv	or are "git checkout branch" and "git branch" not same?	03:06
SethRobertson	I would recommend running `gitk --all --date-order` yes	03:06
	If you mean "branch" literally, no	03:06
tangyin	SethRobertson: Thank you :)	03:07
vdv	so, i've created a branch with "git branch name" and then "git checkout name"	03:08
	that should be correct, no?	03:08
	← cannonball left	03:08
SethRobertson	Probably. only `git branch` will know for sure	03:08
vdv	i see my branch in the list	03:08
	marked with *	03:09
	when i do changes	03:09
	then i checkout master	03:09
	and git status shows changes	03:09
	→ Heimidal joined	03:11
vdv	i'm really confused	03:13
	git co -b behaves same	03:13
frogonwheels	vdv: you haven't committed the changes?	03:14
vdv	not, but git should say me that i have unstashed changes and don't let to switch branches	03:14
frogonwheels	vdv: when you checkout a branch, git will keep any changes you've made as long as that file hasn't changed between the 'checkout' you've left and the 'checkout' you arrive at	03:15
	← BiggFREE left	03:15
	← SRabbelier left	03:15
frogonwheels	vdv: the simplest example is when you make changes, and then want to commit them to a new branch .. you do git checkout -b <newbranch> then you commit	03:15
	vdv: it might be an option? but I quite like the default.	03:16
vdv	if i checkout some of the existing remotes make changes there, then turn back to master i don't see those changes in master	03:17
	*existing remote branches	03:17
	← automator left	03:17
frogonwheels	vdv: and you forgot to commit the changes you mean?	03:18
vdv	yes	03:18
	ups, no, if i forgot to commit git don't let to checkout master	03:19
	→ spearce joined	03:19
	→ t3i joined	03:19
frogonwheels	well you can always stash them right where you are -	03:19
	← spearce left	03:19
	→ nettrot joined	03:19
mithridates	oops , it's bad if you rebrase a commit which is already public. people will call you #$%@#@	03:19
frogonwheels	vdv: I like the default. If you want to change it, then if there's not a setting - there's probably a hook you could use	03:19
	← alnewkirk left	03:20
frogonwheels	mithridates: generally yes	03:20
	← coderdad left	03:20
frogonwheels	mithridates: you want to undo a rebase you just did?	03:20
	← zombor left	03:20
mithridates	frogonwheels: no, I just read rebase manual. but I'm interested to know how is undoing a rebase :)	03:21
	frogonwheels: would you tell me please?	03:21
	← andrewwatts left	03:21
frogonwheels	mithridates: if you've just done it .. git reset --hard HEAD@{1}	03:21
	← xiangfu left	03:21
frogonwheels	mithridates: ie put the branch you are on (presuming that's the on you rebased) back to the last commit it was on (ie the one before the rebase)	03:22
mithridates	the number means the last change? I mean the last commit?	03:22
frogonwheels	yep exactly	03:22
	mithridates: man git-reflog	03:22
Gitbot	mithridates: the git-reflog manpage can be found at http://git.or.cz/man/git-reflog	03:22
	← jschoolcraft left	03:22
frogonwheels	mithridates: works for (successful) merges too ..	03:22
	(which is nice in a merge, since with HEAD~1 you have two parents)	03:23
mithridates	neat, before starting to learn git I couldn't even imagine that universe has such a great system for programmers	03:23
	→ boombatower joined	03:23
	← boombatower left	03:23
	→ boombatower joined	03:23
	← antivigilante left	03:25
vdv	i don't understand, i'm in branch, i do changes, i don't want those changes show up in other branch when i switch to that branch, and that wasn't case before	03:25
mithridates	vdv: because you merged those branches together	03:26
	"I guess"	03:27
vdv	mithridates, no, i didn't do merge	03:27
	and rebase also	03:27
SethRobertson	vdv: Did you work out that you needed to commit eventually?	03:27
	What does `git status` say?	03:27
	→ ka1ser joined	03:27
	← ka1ser left	03:29
	→ ka1ser joined	03:29
frogonwheels	vdv: commit it.	03:29
	vdv: get into the habit of checking the status before you swap branches	03:29
vdv	i just did "git branch newbranch", then edited a file, then "git status" says that files is changed, but changes must be added, i.e. those are unstaged changes, then i do "git co newbranch", at this moment git always asked to stage and commit before checkout (or stash), but this time it just switches the branch and and changes are in other branch	03:30
frogonwheels	vdv: I told you why.	03:30
	← ka1ser left	03:30
	→ ka1ser joined	03:30
SethRobertson	As frogonwheels said, the changes "float" around until you commit them. Commit them, change branches, and you will not see the changes	03:31
vdv	frogonwheels, if don't commit git always asked to do that, and didn't let to switch	03:31
SethRobertson	It only refuses to let you switch if the files have diverged	03:31
	Trust us and commit.	03:31
frogonwheels	vdv: the reason is the files you modified are exactly the same in both branches. that's why it lets you.	03:31
vdv	ah	03:31
	→ fr0sty_ joined	03:32
	← rayners left	03:33
frogonwheels	s/why/when/	03:33
vdv	ok, thanks, i've never faced to such thing before	03:33
	← ka1ser left	03:33
mithridates	does this command "git reflog delete master@{2}" remove the two commit ago of master?	03:34
	→ ka1ser joined	03:34
SethRobertson	OK, this has now officially blown my mind. I can `git clone` a specific url if I use the IPv4 version of the hostname, but if I use the IPv6 version of the hostname, the clone hangs half-way through. This is an ssh problem but I'm not sure I believe ssh has such an ipv6 bug	03:35
frogonwheels	mithridates: that would jpresumably just remove the log entry (without rtfm)	03:35
bob2	SethRobertson: what makes you think it is a ssh problem?	03:35
	and not mtu or something	03:35
	← Phrogz_ left	03:35
frogonwheels	mithridates: if you want to remove just 2 commits ago - you need to use rebase -i (which rewrites history)	03:35
mariusSO	vdv: Unless switching a branch will clobber (alter) the files you are actively working on, git will switch branches without issue. If switching will clobber, either commit or stash your changes.	03:35
SethRobertson	strace shows the sending ssh reading the data but the receiving side never writes it. Now that I think on this more, it probably is a path-mtu problem.	03:36
	Though I have to ask--where are the error messages	03:36
mithridates	frogonwheels: I did rtfm but it says:"This command is to manage the information recorded in it ." I was wondering what kind of information is that	03:37
bob2	path mtu seems less likely	03:37
mithridates	frogonwheels: is it just log?	03:37
	→ bhosie joined	03:37
	← psoo left	03:37
bob2	what error would you like?	03:37
SethRobertson	ipv6 refuses to frag, so if someone is not honoring path mtu discovery, it would have this symptom	03:37
frogonwheels	mithridates: the reflog is just a log (it does help with keeping recent activity from being garbage collected)	03:37
mariusSO	mithridates: reflog is just a log of your actions in the repo. No need to delete those actions really :)	03:37
	→ nadavoid joined	03:38
mithridates	I got ya	03:39
mariusSO	mithridates: So, unless you really really need to expunge something from your repo, to make sure the objects are not there in the "db", just leave it alone :)	03:39
	→ aah joined	03:40
mithridates	mariusSO: ok ! , tnx , and if it was bothering me I have to (find the number of activity by reflog and put it infront of rebase -i)	03:41
	→ Eaven joined	03:41
fr0sty_	mithridates: 'bothering you' ? do you want the commit to go away or just the entry in the reflog?	03:42
mariusSO	Heh	03:42
mithridates	hey fr0sty_ , I wanted to remove some commits by rebase -i	03:43
mariusSO	mithridates: If you made a booboo in your commit history you can do 'git rebase -i HEAD~<n>, where <n> is the number of commits ago	03:43
	→ xiangfu joined	03:44
mithridates	yes I got it!	03:44
	→ JEEB joined	03:44
	← JEEB__ left	03:44
mithridates	oops, too much data for my little brain. I must go to bed.	03:45
	← xiong left	03:45
mithridates	good night	03:45
	thank you guys for all your helps	03:45
	← mithridates left	03:46
	mariusSO is a bit puzzled	03:46
mariusSO	:)	03:46
timH	i think git was created to drive me nuts	03:47
mariusSO	timH: Yet, here you are :)	03:48
	→ fivetwentysix joined	03:48
frogonwheels	timH: "Where do you want to drive your nuts today?"	03:48
timH	i mean, I just want to work on this project...	03:48
	guess I'll just nuke and start over	03:49
frogonwheels	timH: what's happened?	03:49
	timH: probably won't help.	03:50
timH	i mean, I just generally don't know what i'm doing.. i'm like day #1 with git	03:50
	:)	03:50
	but I wanted to merge two branches	03:50
	so I had one branch (master), I pulled the other branch	03:51
	had some conflicts, tried to resolve them, guess I accidentally committed them and then my git status said I was 2 commits ahead of master	03:51
	← bhosie left	03:51
frogonwheels	timH: where did you pull your other branch from?	03:51
fr0sty_	2?, hmmm	03:51
timH	i mean, you want the specific url?	03:52
	it was a branch off of the master	03:52
frogonwheels	timH: ok.. well it's a DVCS so all your branches are cloned into your local repo	03:52
fr0sty_	timh: history \|grep "^git" > pastbin	03:53
frogonwheels	timH: git branch -a	03:53
	← mdel_ left	03:53
timH	ahh	03:53
	i kept doing git branch and it would only show the current branch i guess?	03:53
fr0sty_	only shows local branches	03:53
frogonwheels	no - it ommits 'remote' branches.	03:53
timH	ah	03:54
frogonwheels	if you have a (remote) branch remotes/origin/mybranch then doing a git checkout mybranch will set up the local tracking branch for that remote branch	03:54
timH	well why didn't git merge <branchname> work?	03:54
	it has to be local?	03:54
frogonwheels	timH: git merge origin/<branchname> probably would though	03:55
fr0sty_	timH: sounds like it did work	03:55
	you got conflicts, right?	03:55
timH	that was from git pull	03:55
fr0sty_	pull = fetch + merge	03:55
frogonwheels	timH: what was your git pul cli?	03:55
	I mean what did you type?	03:56
fr0sty_	if you are in branch A and do 'git pull <remote> branchB' it will fetch branchB from the remote and merge it into branch A	03:56
timH	ok, I'm setting up a fresh master	03:56
fr0sty_	timH: history \| grep "^git" > pastebin	03:56
	show us what you actually did.	03:56
	→ mdel_ joined	03:57
	→ ajpiano joined	03:57
timH	can we just start from square 0.. i just did "git clone git://..... fspot_master"	03:57
	so now I want to do "git pull git://..... -b infobox"	03:58
frogonwheels	no	03:58
fr0sty_	-b?	03:58
timH	sorry	03:58
	i guess I was thinking merge	03:58
	isn't -b branch?	03:58
fr0sty_	'cd fspot_master; git remote add <name> url'	03:58
	is the pull url different?	03:59
timH	no	03:59
frogonwheels	fr0sty_: I think it's the same	03:59
timH	ok	03:59
thiana	If you're amending a commit, how do you remove a file from it?	03:59
fr0sty_	timH: 'you want to merge infobox into master?	03:59
timH	yes	03:59
fr0sty_	'git merge origin/infobox'	03:59
	→ gilaniali joined	03:59
	← Bass10 left	04:00
timH	ok, that's where I got to before.. 2 conflicts	04:00
	← rgr left	04:00
	← shamrock left	04:00
fr0sty_	edit the files to resolve the conflicts...	04:01
frogonwheels	timH: or use git mergetool	04:01
mariusSO	thiana: removing from the amending, or from the repo?	04:01
	→ alnewkirk joined	04:01
fr0sty_	but don't do anything else (like git add or git commit)	04:01
timH	ok, so one of the things in conflict is a directory?	04:01
thiana	mariusSO: The repo.	04:01
fr0sty_	thiana: git rm <filename>	04:02
mariusSO	thiana: 'git rm <file> && git commit --amend'	04:02
thiana	Thanks.	04:02
fr0sty_	or that...	04:02
	timH not possible.	04:02
	only files can conflict	04:02
	→ shiba_yu36 joined	04:02
fr0sty_	is this a public github repo?	04:02
timH	yes	04:03
fr0sty_	what is the url?	04:03
timH	git://git.gnome.org/f-spot	04:03
	well it shows "Unmerged paths" and under that is lib/TagLib/TagLib which is most definitely a directory	04:04
fr0sty_	stand by	04:04
	← alnewkirk left	04:05
	← ajpiano left	04:05
	→ alnewkirk joined	04:06
	→ ajpiano joined	04:06
	← tangyin left	04:07
wereHamster	submodule?	04:07
timH	yes	04:07
fr0sty_	that's a horse of a different color.	04:08
timH	sorry	04:08
fr0sty_	timH the conflict is both branches have modified the submodule commit.	04:09
	the commit the submodule points to, that is.	04:09
timH	honestly that makes zero sense to me	04:09
	:)	04:09
	i mean other than obviously in general software development terms	04:10
fr0sty_	why are you trying to merge these branches?	04:10
timH	well, I want the work that was done in infobox in master	04:10
	← Yuffster left	04:11
timH	i mean, I know git is better but I could already have this done in 3 different revision control systems :(	04:11
	→ Yuffster joined	04:11
timH	maybe I'll just abandon it for now	04:12
	→ jrmuizel joined	04:12
fr0sty_	timH: you need to do 'git submodule init'	04:13
	zz_rubydiamond → rubydiamond	04:13
	← alberto56 left	04:14
timH	ok	04:14
mariusSO	how would svn resolve externals at two different states in a merge?	04:14
	← Yuffster left	04:14
	→ Yuffster joined	04:14
	→ alberto56 joined	04:14
fr0sty_	possibly git submodule update as well	04:14
	← alberto56 left	04:15
fr0sty_	timH: are you sure master is the branch you want?	04:16
timH	fr0sty: at this point no	04:17
	but I mean, it's master, it's where everything needs to land I'm pretty sure	04:17
	anyway, I literally have a headache now and got to basically zero coding.. i'll leave it for someone else	04:18
	fr0sty: thanks	04:18
	← ignatp left	04:19
	→ cytrinox_ joined	04:21
fr0sty_	your welcome, i guess...	04:22
	if you want to just start coding, checkout infobox and have at it...	04:22
timH	fr0sty: but it's missing stuff that's in master right?	04:23
fr0sty_	yes,	04:23
	do you need any of it?	04:23
	← cytrinox left	04:24
	cytrinox_ → cytrinox	04:24
	rubydiamond → zz_rubydiamond	04:24
fr0sty_	generally if you are working on a feature you develop the feature first and then merge it back into the project	04:25
	← jrmuizel left	04:25
timH	right	04:27
fr0sty_	so develop first, merge second.	04:29
	← whaley left	04:30
	← Theravadan left	04:31
	← alnewkirk left	04:34
	→ hebz0rl joined	04:35
wereHamster	how would you 'merge first, develop second'	04:36
	?	04:36
	seems unpractical	04:36
timH	the code is basically finished, it's not mine, I'm just trying to get it into master	04:37
	I did some fixes that didn't need to be done	04:37
	and that's because this code wasn't in master	04:37
	← brntbeer left	04:39
	→ tangyin joined	04:40
frogonwheels	timH: I'm just gonna try this merge.. it can't be _that_ hard :)	04:42
timH	no, in fact it seems simple enough, even to me	04:42
	but i trying to learn 10 things at once	04:42
frogonwheels	have you done the git submodule init git submodule update stuff?	04:42
timH	i did	04:42
	→ towski joined	04:43
timH	i ended up nuking it all	04:43
	→ ignatp joined	04:43
timH	see, I don't know anything about the submodules in git	04:44
frogonwheels	timH: a submodule is just an external library - the main project stores a pointer to a particular commit in the library	04:45
timH	frogonwheels: i mean, of course conceptually I know what a submodule is.. :) i just didn't know they related to git in anyway	04:46
	→ TheColonial joined	04:46
TheColonial	hi everyone. i can't seem to find the command which shows a tag (or list of tags) that are applied to a revision. is this possible?	04:46
	i want to extract that as part of a script	04:46
fr0sty_	git tag --contains <rev> ?	04:47
	← Yuffster left	04:47
TheColonial	fr0sty_: thanks,i'll give that a shot!	04:48
	fr0sty_: that's perfect. thank you very much	04:49
fr0sty_	no problem.	04:49
	→ khmarbaise joined	04:50
frogonwheels	timH: hmm.. they deleted a submodule it appears?	04:52
timH	well, I think it got pointed from a local branch to a branch at another url	04:53
	← ignatp left	04:53
timH	scratch that	04:53
	it just went from one repo at github to another one	04:53
	← khmarbaise left	04:54
	→ freimatz joined	04:56
	→ krosenvold_ joined	04:56
frogonwheels	timH: ok clone with git clone --recursive <url> (this will do the submodules as well)..	04:57
	timH: then git merge origin/infobox	04:57
	timH: use git checkout --ours lib/TagLib to resolve the submodule conflict	04:58
	timH: and you could run git mergetool --too=kdiff3 and resolve the head conflict...	04:58
	timH: then git commit	04:58
	timH: and you're done.	04:58
timH	that's straight forward	05:00
	→ ct^^ joined	05:00
frogonwheels	timH: working out the TagLib thing took a bit of knowledge - but easy enough once you know.	05:00
timH	hmm, ok, give me a second, thanks	05:01
	then I need to create a patch	05:01
frogonwheels	timH: --tool=kdiff3	05:01
	timH: you need to send it to the developer?	05:01
timH	right	05:01
frogonwheels	timH: git am HEAD~1 after the commit	05:02
	no	05:02
	git format-patch HEAD~1	05:02
	sorry (am is the thing that applies patches)	05:02
	← madewokherd left	05:03
	← sftp left	05:03
	← ct^ left	05:03
	← scarabx left	05:03
	← OOPMan left	05:04
timH	frogonwheels: so that recursive step appears to be done by the autogen I guess?	05:05
	→ Pupeno joined	05:05
timH	i mean, it grabs all the submodules?	05:05
frogonwheels	timH: yeah - it clones all the submodules and checks out the appropriate commit	05:06
	... just means it's one command so you can go away and come back with a fully cloned repo including submodules.	05:06
timH	you missed a step in there	05:07
frogonwheels	?	05:07
timH	i had to do git remote add infobox ....	05:07
frogonwheels	nahh	05:07
timH	oh	05:08
frogonwheels	unless you're pulling from another url	05:08
	→ OOPMan joined	05:08
timH	i was still in the directory i executed the clone from	05:08
	when I tried to do the merge the first time	05:08
frogonwheels	:) ok I missed 'cd f-spot '	05:08
	→ Koteswara joined	05:08
timH	haha	05:09
	→ ShooterMG joined	05:09
timH	so after the "get checkout --ours ..." when I do the mergetool it still tries to have me resolve the "TagLib" directory	05:10
	→ justin-george joined	05:10
	→ Slix`` joined	05:10
	← ReleaseX left	05:10
frogonwheels	hold a sec.. gonna try that	05:11
	← Ogredude_ left	05:11
frogonwheels	timH: try ignoring that and skipping to the next one	05:11
	→ apsysoev joined	05:12
timH	then it still shows lib/TagLib/TagLib under unmerged paths	05:12
frogonwheels	arghh.. git checkout --ours lib/TagLib/TagLib	05:12
	sorry - missed a bit	05:12
	← Slix` left	05:12
timH	oh, I thought maybe it was like, recursive	05:12
	so, right now, since a merge is taking place there exists an "ours" and "theirs"?	05:13
	→ gasull1 joined	05:13
	→ nbrochu joined	05:14
timH	the command executes fine but still shows it under unmerged paths?	05:14
frogonwheels	hold a sec	05:14
	→ gasull joined	05:16
frogonwheels	timH: git reset HEAD lib/TagLib/TagLib	05:16
	methinks there are some usability issues with resolving submodule merges :\|	05:16
	← metaperl left	05:17
timH	http://pastebin.com/enktP2Cn	05:18
	that's what git status looks like now, can that be right?	05:18
fr0sty_	frogonwheels: a git diff will show what SHA the submodule should point at (ours and theirs) he can just checkout what he wants and add it.	05:18
frogonwheels	Look - try commiting what you've got and doing a git submodule update	05:18
	← gasull1 left	05:18
frogonwheels	fr0sty_: duh - of course thanks v much	05:19
	← Slix`` left	05:19
frogonwheels	fr0sty_: anyway , he wants (I believe) to use the one in the master branch.	05:19
timH	yes	05:19
	that much I know	05:19
	clear	05:19
	uggh	05:19
frogonwheels	timH: so commit, and it should be all good ( and git submodule update)	05:19
	timH: yeah that was a bit more painful than it should have been - the submodule was the pita	05:20
timH	so it says ahead by 2 commits?	05:20
frogonwheels	two?	05:20
fr0sty_	the infobox commit and the merge, I think...	05:20
frogonwheels	ooh. possibly then	05:21
	← Lgb left	05:21
timH	the log does show a "merge" right under the commit SHA1	05:21
	→ sh1m joined	05:21
timH	the merge isn't a SHA1 though it doesn't look like	05:21
fr0sty_	?	05:22
timH	Merge: 068daf6 57322a6	05:23
frogonwheels	http://pastebin.com/piDKcNDh	05:23
timH	it's right under the commit <sha1> in the log	05:23
frogonwheels	That's the short sha1s of the two parents for the merge	05:24
	→ amerine joined	05:24
timH	so just enough to make them unique?	05:24
frogonwheels	yep.	05:24
	or something like that	05:24
timH	hehe	05:24
	→ Lgb joined	05:24
timH	ok, so I made a patch	05:24
	wow	05:24
	thanks	05:24
frogonwheels	:) np	05:24
timH	i'm going to be utterly useless tomorrow	05:25
frogonwheels	hmm.. just had a thought	05:25
timH	oh?	05:25
frogonwheels	nah -it's ok	05:25
	really	05:25
	← Eaven left	05:25
timH	lol	05:25
	← ShooterMG left	05:25
frogonwheels	get some sleep :)	05:25
	← sh1mmer left	05:26
	sh1m → sh1mmer	05:26
timH	ok, thanks again	05:26
	timH → timH_away	05:27
	← sivy left	05:28
	→ khmarbaise joined	05:29
	← gasull left	05:32
	← hobodave left	05:32
	← hyperair left	05:33
	→ hobodave joined	05:33
	← Eridius left	05:33
	← dukebag left	05:33
	→ jelmer joined	05:34
	← jelmer left	05:34
	→ jelmer joined	05:34
	→ doener joined	05:37
	← hobodave left	05:37
	→ rpbertp13 joined	05:39
	→ yuriks_ joined	05:41
	← yuriks left	05:44
	→ robinr joined	05:44
	← amerine left	05:45
	→ urkud joined	05:46
	→ _ollie joined	05:47
	→ Pewpew joined	05:47
	→ goshawk joined	05:47
	→ muep joined	05:48
	→ gasull joined	05:49
	← timH_away left	05:49
	← Pewpewarrows left	05:51
	← khmarbaise left	05:53
	→ brakkus_droideka joined	05:54
	← eighty4 left	05:54
	→ sambarino joined	05:55
	← robinr left	05:55
	← _ollie left	05:55
brakkus_droideka	hey guys	05:56
	← adamm left	05:57
	← patnotz left	05:57
	→ happynoff1 joined	05:57
brakkus_droideka	i've put my public key in a git repository but it's still asking for my password when i try to git clone	05:58
	← gilaniali left	05:58
	→ eighty4 joined	05:58
	← nadavoid left	05:59
	→ gasull1 joined	06:00
	→ droideka joined	06:00
droideka	hey there	06:00
	back	06:00
	so any idea where i should start troubleshooting?	06:00
	http://pastebin.com/cNSb7x3s	06:01
	there is the output that i get	06:01
	← urkud left	06:02
	→ thiago joined	06:02
	→ Deesl joined	06:02
	← justin-george left	06:02
	← gasull left	06:03
	← brakkus_droideka left	06:03
	← checkers left	06:04
	← BP{k} left	06:04
	← BlindOracle left	06:04
	→ BP{k} joined	06:04
happynoff1	Hi ! I managed to make git work over HTTP. There is one last thing I want to fix. When I push, my password is asked twice. Why ? and how can I work around it ? thanks	06:05
	← drubies left	06:06
	← freimatz left	06:06
frogonwheels	droideka: either get the public repository URL - or you'll need to have permissions for the server	06:06
	→ BlindOracle joined	06:06
frogonwheels	happynoff1: pushing with what protocol? if it's http then shrug	06:07
happynoff1	yes it's http and it works well except for the password asking	06:07
frogonwheels	so git remote -v shows pushing with http?	06:08
	dav. ok	06:08
	or something	06:08
happynoff1	yes	06:08
	it's all http	06:08
	no ssl for now	06:08
frogonwheels	happynoff1: you're expecting it to ask once I gather?	06:09
happynoff1	yup. If I can find a way to not ask it but without using .netrc it's ok too :)	06:09
	→ freimatz joined	06:10
	frogonwheels shrugs	06:10
	→ gasull joined	06:10
	← kanru left	06:11
	→ hobodave joined	06:11
	← nettrot left	06:11
happynoff1	?	06:12
	happynoff1 → happynoff	06:12
	← gasull1 left	06:12
	← t3i left	06:12
	→ drubies joined	06:12
	← droideka left	06:13
	→ fredcooke joined	06:16
	→ cbreak_work joined	06:16
	→ slonopotamus joined	06:16
fredcooke	Hi, I'm trying to .gitignore all files/dirs except ones listed in .gitignore with ! preceeding them, but I'm having no luck. My exact config is :	06:16
	# comment	06:17
	*	06:17
	# comment	06:17
	!default	06:17
slonopotamus	so, let's try once again. anyone has ideas how to undo accidental git svn reset -r <very-old-revision> without refetching?	06:17
	← yhager left	06:18
doener	fredcooke: that should ignore everything but things called "default", but note that git won't recurse into directories unless they're called "default"	06:19
fredcooke	exactly, i have a directory called default with a single file inside called index.php	06:19
	and it gets ignored no matter what	06:19
	i also have !.gitignore in there	06:19
	← gasull left	06:19
fredcooke	otherwise nothing showed up	06:19
	the !.gitignore works fine	06:20
	but default /default default/ default/* etc etc do not work	06:20
doener	what is it now? default or index.php?	06:20
	ah, nvm	06:20
fredcooke	default, the directory, and all inside it, which is currently only index.php	06:20
	thanks for speaking up, btw! :-)	06:21
	→ t3i joined	06:21
doener	fredcooke: well, you un-ignore the directory, but not the file in it	06:21
	fredcooke: with the !default, git looks into that directory, finds index.php and sees that that matches *, so it ignores it	06:21
fredcooke	shouldnt the !default/* unignore all inside too ?	06:21
	or do i need two lines	06:21
	oen for the dir, one for its contents?	06:21
	didnt try that, hold one second	06:21
	→ psankar joined	06:22
	← psankar left	06:22
	→ psankar joined	06:22
	← freimatz left	06:22
	← goshawk left	06:22
bob2	fredcooke: * and .* appear to work fine	06:23
fredcooke	doener, thank you for making me look at it from the correct perspective! :-) two lines required, one for the directory, one for the files inside it	06:23
doener	fredcooke: I'd put '' and '!/' into the top-level .gitignore, and !* into default/.gitignore	06:23
	→ urkud joined	06:23
doener	fredcooke: because default/* won't recurse, i.e. doesn't match default/foo/file	06:23
fredcooke	oh	06:24
	and what is i dont want any git files in default/	06:24
	?	06:24
frogonwheels	doener: weird. doesn't work	06:24
fredcooke	doener, it does recurse	06:25
	The following works nicely:	06:26
	*	06:26
	!/default/	06:26
	!/default/*	06:26
	!.gitignore	06:26
	Thanks again! :-)	06:26
	→ philthno2 joined	06:27
	← TheColonial left	06:27
	→ airborn joined	06:28
	← Deesl left	06:28
	zz_rubydiamond → rubydiamond	06:29
	→ Deesl joined	06:29
	← Deesl left	06:30
	← fredcooke left	06:30
	→ _ollie joined	06:36
doener	frogonwheels: what doesn't work?	06:38
	→ envi joined	06:39
frogonwheels	doener: I couldn't even get what fredcooke claimed did the trick to work :(	06:39
	doener: oh yeah- sorry -It was in reference to the fredcooke conversation	06:39
doener	frogonwheels: fredcooke's "final" .gitignore will work for default/file but not for default/foo/file, guess he didn't try that	06:40
	frogonwheels: and for my approach, note that the order in the .gitignore file matters, * must come before !*/	06:40
frogonwheels	doener: sure. still didn't do it for me :(	06:41
	→ Mayank_ joined	06:42
Mayank_	Hi	06:42
frogonwheels	I made a clean project with .gitignore test default/index.php files	06:42
Mayank_	I needed some help with gt	06:42
doener	frogonwheels: worksforme http://git.pastebin.com/6p4ReJXh	06:42
Mayank_	git	06:42
	Mayank-Chopras-MacBook-Pro:first_app Mayank$ git remote add origin [email@hidden.address]	06:42
	fatal: remote origin already exists.	06:42
	Mayank-Chopras-MacBook-Pro:first_app Mayank$ git push origin master	06:42
	Permission denied (publickey).	06:42
	fatal: The remote end hung up unexpectedly	06:42
	I'm getting this error while trying to upload files to the repository	06:43
doener	off to work now	06:43
Mayank_	can anyone please help with this	06:43
frogonwheels	doener: bug that doesn't show index.php as untracked - or foo/index.php (if you added one)	06:43
	doener: sure	06:43
doener	frogonwheels: sure does, http://git.pastebin.com/vLfmKk0j	06:44
	→ Torsten joined	06:46
	← fivetwentysix left	06:46
	→ hyperair joined	06:47
	← Unode left	06:47
frogonwheels	Mayank_: if you want to access via a public url - use the git or http protocol otherwise read up on generating ssh private keys and setting up github with your public keys	06:48
	→ Kira joined	06:49
	Kira pokes nvie	06:50
	→ Zhonghua joined	06:50
Mayank_	frogonwheels: im going through the RoR tutorial on the website and it does say to read up on ssh private keys so I'll read up on that ..thanks	06:50
	← hyperair left	06:52
	← sonnym left	06:53
	→ mtkd joined	06:54
	→ slyphon joined	06:55
	→ ph^ joined	06:55
	→ sonnym joined	06:55
	← eletuchy left	06:56
sejo	why does .git/branches exist? nothing is stored in there	06:56
	← theoros left	06:56
	→ _pingu joined	06:57
	→ ph^_ joined	06:57
	← ph^ left	06:57
	slyphon → slyphon_away	06:57
	← Anti-X left	06:57
	→ wagle_ joined	06:58
	← QinGW left	06:59
	→ QinGW joined	06:59
	← wagle left	06:59
	slyphon_away → slyphon	06:59
	→ Eaven joined	06:59
	← RobertMeta left	07:00
	← rvsjoen left	07:00
	→ dukebag joined	07:00
	→ RobertMeta joined	07:01
	→ rvsjoen joined	07:02
	→ Anti-X joined	07:03
	→ priidu joined	07:04
	← tty1 left	07:07
	→ Pew joined	07:07
	← dukebag left	07:07
	→ tty1 joined	07:07
	→ thierryp joined	07:08
	← Pewpew left	07:10
	→ devanjan joined	07:10
	→ nevyn^ joined	07:10
	← _pingu left	07:10
frogonwheels	sejo: http://www.siteground.com/tutorials/git/directory.htm appears it is deprecated	07:14
	← nullvariable left	07:15
boombatower	say i have two "modules" of code in one repository and I decide I want to split them into two repositories...what is the best way to do so and preserve a clean history. My plan was to just duplcate the repository and remove a module from each (different one) and perform rename operations, but then of course I will have all the commit history for the module that was removed	07:15
	I get the feeling git may have some scary way to do this	07:16
selckin	filter-branch	07:16
	→ nullvariable joined	07:16
	→ ch077179 joined	07:17
	← maguey_ left	07:18
boombatower	selckin: interesting, thx...i'll start playing with that	07:19
	git has everything :)	07:19
doener	frogonwheels: hm, your list of files is missing the second .gitignore	07:19
	← sonnym left	07:19
	→ josefrichter joined	07:20
	→ Theravadan joined	07:21
josefrichter	guys, I've made some changes only to realize I'm on different branch. can I somehow change the branch and commit on the right one? the files I was changing looked the same on both branches before I started to edit them.	07:21
selckin	yes. just do it	07:22
FauxFaux	cherry-pick or rebase.	07:22
frogonwheels	josefrichter: you've not checked in	07:22
selckin	he didn't commit yet, he can just change the branch and commit	07:22
boombatower	git status => nothing to commit (working directory clean) ; git filter-branch --tree-filter ´rm -f worker´ HEAD => Cannot rewrite branch(es) with a dirty working directory. ; I missing something?	07:22
frogonwheels	josefrichter: as selckin says - just checkout the right one - if the file is the same in both then it will let you	07:23
josefrichter	yep, not commited yet. i'm on branch but want to commit those changes on master. so I just git checkout master and I won't lose changes?	07:23
	→ sonnym joined	07:24
	← MikhX left	07:24
	→ cilly joined	07:25
josefrichter	it works. nice :-)	07:26
	← phaedrix left	07:27
	← towski left	07:28
	→ MikhX joined	07:28
	← nvie left	07:30
	→ galderz joined	07:31
	← thiago left	07:32
	→ Cap`Tain[Absent` joined	07:33
	← Cap`Tain[Absent` left	07:34
	← t3i left	07:34
frogonwheels	josefrichter: if it would need to overwrite a file to do the checkout, it would abort and not do the checkout	07:38
	(and tell you what you need to do to keep going)	07:38
josefrichter	ok, thanks	07:38
	→ hyperair joined	07:39
	→ dr_win joined	07:41
	→ fivetwentysix joined	07:41
	← urkud left	07:41
	← Heimidal left	07:45
	→ foocraft joined	07:46
	→ lzyy joined	07:46
	→ workmad3 joined	07:48
boombatower	rewrite subdirectory filter worked like a charm...not sure why removing a file gives me crap about being dirty	07:50
	← jacobat left	07:51
boombatower	ugh...man page has the ` and you need '	07:51
	→ jacobat joined	07:51
	→ Alias joined	07:52
Alias	Is there a git equivalent of "svn revert one-file-within-the-checkout"	07:52
	→ mpartipilo joined	07:52
boombatower	Alias: git checkout filename	07:52
Alias	checkout?	07:52
	Alias hits the docs	07:52
	cxreg wonders how gittogether went	07:53
boombatower	sweet removed directory from history...very cool	07:54
	hmm so it still keeps commit messages...just shows nothing as being changed?	07:55
	git filter-branch --index-filter 'git rm -r --cached --ignore-unmatch worker' HEAD	07:55
cxreg	boombatower: you probably wanted --prune-empty	07:55
boombatower	ah	07:56
	boombatower starts over	07:56
	→ rlotun joined	07:56
slonopotamus	meh.	07:56
	← lzyy left	07:56
	→ freimatz joined	07:58
	→ lzyy joined	07:58
boombatower	cxreg: much better thx	07:59
	← boombatower left	08:00
cxreg	np	08:00
	→ boombatower joined	08:02
	← boombatower left	08:02
	→ boombatower joined	08:02
	→ davzie joined	08:02
	→ kanru joined	08:02
	→ urkud joined	08:03
	→ nonno joined	08:04
	→ psoo joined	08:07
	→ eletuchy joined	08:11
	← foocraft left	08:13
	→ bentob0x joined	08:14
	← Pew left	08:14
	→ metaperl joined	08:15
	→ Pew joined	08:15
	→ phaedrix joined	08:15
	→ gebi joined	08:16
	← gebi left	08:16
	→ gebi joined	08:16
	→ khmarbaise joined	08:16
	→ tty2 joined	08:16
	← mpartipilo left	08:16
	← thedeeno left	08:17
	← drubies left	08:18
	→ thedeeno joined	08:18
	← tty1 left	08:18
	→ drubies joined	08:19
	← khmarbaise left	08:20
	→ Schmallon joined	08:23
	→ asmega joined	08:23
	← urkud left	08:25
	← davzie left	08:25
	← happynoff left	08:25
	← pedrobelo left	08:26
	← hyperair left	08:27
	← Mayank_ left	08:28
	→ davzie joined	08:30
	← GarethAdams\|Work left	08:34
	→ tatsuya joined	08:38
	← d0ugal left	08:38
	→ skoop joined	08:39
	← Eaven left	08:40
	→ dharriga1 joined	08:41
	→ hyperair joined	08:43
	→ cytrinox_ joined	08:45
	← krosenvold_ left	08:45
	← carutsu left	08:45
	→ icwiener joined	08:45
	← sigmonsays left	08:47
	→ dunpeal joined	08:47
	← cytrinox left	08:47
	cytrinox_ → cytrinox	08:47
dunpeal	Hey. Is there a Git command that shows not only the names of the remotes, and whether they're fetch/push, but also prints their actual URLs?	08:48
	→ akamike joined	08:49
	← dharriga1 left	08:49
doener	dunpeal: git remote -v	08:50
dunpeal	doener: heh, thanks.	08:52
	→ bushwakko joined	08:52
dunpeal	doener: OK, another question: I see both a (fetch) and a (push) URL; what does the distinction mean?	08:53
	→ sigmonsays joined	08:54
	← cytrinox left	08:54
	→ cytrinox joined	08:54
	→ Eaven joined	08:54
doener	dunpeal: that you can configure them individually	08:54
	dunpeal: can't come up with a use case right now, though	08:56
	→ marc_kdab joined	08:56
	← sdboyer left	08:56
	→ g0bl1n joined	08:56
dunpeal	doener: continuous / gauntlet build scenario.	08:56
	→ sdboyer joined	08:56
	← sdboyer left	08:56
	→ sdboyer joined	08:56
dunpeal	and yes, I see that git remote -v just shows remote.<name>.url as (fetch), and remote.<name>.pushurl as (push).	08:57
	→ curtana joined	08:57
	→ freshmilk joined	08:58
	dr_leppe1 → dr_lepper	08:59
	← boombatower left	09:00
freshmilk	Can anyone help out a noob? I've modified my .gitignore on my local machine, and successfully done a 'push', but when checking the .gitignore on the remote repo it hasn't been updated. How do I push the changes made to my local .gitignore to the repo?	09:00
	→ Ogredude_ joined	09:00
dunpeal	freshmilk: are you on the same branch?	09:01
	is the remote bare?	09:01
freshmilk	how do i check?	09:01
	what does bare mean? i only started with git 2 days ago!	09:02
	→ Aiden joined	09:02
dunpeal	reading some tutorial would be advisable	09:03
freshmilk	hold on, i've just done a 'git status' and it says 'changed but not updated' - gitignore	09:03
dunpeal	freshmilk: `git branch` should tell you which branch you're on	09:03
	Putr\|Away → Putr	09:03
dunpeal	`git status` will as well	09:03
	← Fullmoon left	09:03
dunpeal	freshmilk: the fact that you got that message from git status means that you didn't actually commit the changes to the .gitignore file.	09:03
	← rpbertp13 left	09:04
dunpeal	you need to understand the index and how to make commits.	09:04
	the short answer for now is that you need to do `git add .gitignore` before the commit, otherwise it won't get committed.	09:04
freshmilk	thanks dunpeal that did it.	09:04
Aiden	how do you add more keys to github?	09:04
freshmilk	there's a lot to learn with git, easy to get overwhelmed.	09:05
	thanks for the tip.	09:05
	→ Yuuhi joined	09:06
	→ boombatower joined	09:06
	→ kukks joined	09:06
	← Ogredude_ left	09:07
Aiden	nvm , found it	09:07
	← MacGyverNL left	09:08
	← dunpeal left	09:09
	← xiangfu left	09:10
	→ cytrinox_ joined	09:11
	← cytrinox left	09:12
	→ MacGyverNL joined	09:13
	← MacGyverNL left	09:13
	→ MacGyverNL joined	09:13
	← eletuchy left	09:13
	→ eletuchy joined	09:14
	→ bctrainers joined	09:14
	← bronson left	09:15
	← mtkd left	09:15
	← cytrinox_ left	09:16
	→ TuxCM joined	09:17
	→ Arthur joined	09:17
	← Theravadan left	09:18
	← eletuchy left	09:19
	← lzyy left	09:19
	→ Theravadan joined	09:20
	← Mage left	09:21
	→ Aiden\|TBBT joined	09:21
TuxCM	1.7.2.1 was OK, but as of 1.7.2.2, * t9101-git-svn-props.sh * started failing	09:24
	and as I use git-svn a lot to work on remote svn repo's, I don't dare to install	09:24
	← Aiden left	09:24
	→ consolers joined	09:28
TuxCM	git://gist.github.com/651002.git	09:29
	→ maguey_ joined	09:30
	← hyperair left	09:31
	→ d0ugal joined	09:33
	← Schmallon left	09:34
albel727_2	TuxCM: hmm. this would be better off reported to git mailing list (and maybe also to #git-devel)	09:37
	albel727_2 → albel727	09:37
TuxCM	I (sorry) don't have the time to follow git ML's	09:37
	I did so in the beginning	09:38
	sraue → sraue\|afk	09:38
TuxCM	I'm willing to report it, but it would be hard to do correct follow ups after that	09:38
	← maguey_ left	09:40
	sraue\|afk → sraue	09:40
	rubydiamond → zz_rubydiamond	09:40
albel727	TuxCM: well, if you can't help people in figuring out the problem, then I don't know. On your place, I'd definitely sent at least a single e-mail (with all the generic info, like what is your OS, etc) to git ml. Just say, that you're unable to follow, and hopefully people will be understanding.	09:41
TuxCM	ok, I'll do so	09:42
	→ cytrinox joined	09:43
	→ ch077179_ joined	09:44
albel727	TuxCM: Thanks. I think, info on where did you get that 1.7.2.2 distribution (was it a binary package, or self-compiled, and if so, how did you build that and what was the output of make) would be helpful too, but feel free to skip it, if you really busy.	09:45
	→ parasti joined	09:45
	← infinity0 left	09:45
	← ptor left	09:46
albel727	s/you really/you're really/	09:46
	→ infinity0 joined	09:46
	← infinity0 left	09:46
	→ infinity0 joined	09:46
	← ch077179 left	09:47
	← Haarg left	09:47
	← thiana left	09:47
	← Tommy[D] left	09:47
	→ Haarg joined	09:47
	→ thiana joined	09:48
	→ Tommy[D] joined	09:50
	← dr_win left	09:50
	→ msimkins joined	09:53
	zz_rubydiamond → rubydiamond	09:53
	→ allenwei joined	09:55
Aiden\|TBBT	We couldn't find a README for this repository, we strongly recommend adding one. For more details on what formats we support, visit github/markup	09:56
	why do i need this	09:56
	← Torsten left	09:56
	← josefrichter left	09:57
	→ cytrinox_ joined	09:57
Rhonda	It is displayed directly so it's considered to be the central documentation file.	10:00
	← cytrinox left	10:00
	cytrinox_ → cytrinox	10:00
	→ dharrigan joined	10:00
	← dharrigan left	10:00
	→ dharrigan joined	10:00
	← Eaven left	10:01
Aiden\|TBBT	is it a bad thing	10:01
	if i havent made it	10:01
	>_<	10:01
	Aiden\|TBBT → Aiden	10:01
	→ MetaCosm joined	10:01
	← MetaCosm left	10:01
	→ MetaCosm joined	10:01
	→ winterweaver joined	10:03
	← QinGW left	10:05
	← Kira left	10:06
	→ tvw joined	10:07
	← consolers left	10:09
	← ThiefMaster left	10:10
Rhonda	That's up to you if you consider lack of documentation a bad thing.	10:14
	← mithro left	10:14
	→ Fissure1 joined	10:15
	→ chrischambers joined	10:15
	← Fissure left	10:15
	→ blahdeblah joined	10:16
blahdeblah	Hi. Are there any tools available to set up automatic push from one repository to another without doing something old school like scheduling periodically from cron?	10:17
	← tangyin left	10:18
albel727	blahdeblah: why would you need that, exactly?	10:19
	← dharrigan left	10:20
	← gretch left	10:21
blahdeblah	albel727: So that my repositories stay in sync?	10:22
	Automatically, i might add...	10:22
albel727	blahdeblah: Completely non-specific. I'm asking, because if, e.g you want to push from one bare repo to another, then post-receive hook might be a good choice.	10:22
	→ ThiefMaster joined	10:23
blahdeblah	Sorry. They are two non-bare repos.	10:23
	← cehteh left	10:23
	← rson left	10:23
	← freshmilk left	10:24
albel727	...says blahdeblah and still provides no details on his use-case. =/	10:24
blahdeblah	albel727: I am fairly new at git, trying to migrate from svn, and saying the phrase "one central repo" gets you in trouble in this channel. So i'm trying to brainstorm ways i can get a near equivalent of one automatically updated central repository, but still do it "the git way".	10:25
	→ dean[w] joined	10:25
	→ chrisred joined	10:25
dean[w]	I just did a merge and pretty much the whole file conflicted but I want to pick the file from the branch I merged into the current branch	10:25
	← boombatower left	10:25
parasti	one central repo isn't quite the same as autopushing on commit	10:25
dean[w]	how can i do this, to save having to edit the file and merge so many conflicts	10:25
blahdeblah	parasti: tell me about it! :-)	10:26
albel727	blahdeblah: I'm pretty sure, most people here won't have any problem with "one central repo", and parasti tells the truth =)	10:26
	→ Schmallon joined	10:26
	← fivetwentysix left	10:27
	→ cehteh joined	10:27
TuxCM	albel727, mail sent. I actually checked again on all the three systems mentioned in the mail	10:29
blahdeblah	albel727: Let me see... What further information can i tell you about my scenario that would help clarify my question? I am working with puppet, whose configurations are written in a DSL. I have a server which has some shared and some independent code. When i do a commit & push from my laptop/desktop, i want the changes to be committed to the working copy on the server. What's the least-pain path to doing this?	10:29
	← bauruine left	10:30
albel727	TuxCM: yeah, I see it. thanks again. (btw, cc:-ing is a policy in git ml, so you'll see all responses alright)	10:31
TuxCM	fine	10:31
	← stliu left	10:34
	← Theravadan left	10:34
	rubydiamond → zz_rubydiamond	10:35
	→ dharrigan joined	10:36
	← rlotun left	10:36
	← Koteswara left	10:36
	← ayanes left	10:36
	→ ayanes joined	10:36
	→ awilkins joined	10:36
	→ chrisshattuck_ joined	10:37
	→ Koteswara joined	10:37
Aiden	if someone else wants my project on github, do they just use the git clone commando? Or some other?	10:38
	i have added there public keys	10:38
albel727	blahdeblah: ok, I didn't exactly understand what's that "puppet" is, and how exactly do you keep "shared and independent code" in single repository. But, I got an impression, that what your repo at the server needs to have a checked out working copy, and code from that working copy is used online. Am I completely off mark?	10:38
blahdeblah	albel727: No, pretty much spot on	10:38
	→ micha_ joined	10:38
	← micha_ left	10:39
blahdeblah	Pretend it's a Ruby on Rails web site and i want it automatically updated and live straight away. ;-)	10:39
	← chrisshattuck left	10:39
	chrisshattuck_ → chrisshattuck	10:39
albel727	blahdeblah: so, a working copy at server is used for production purposes. Ruby web site makes it completely clear.	10:39
blahdeblah	cool	10:39
	← jonesy left	10:39
albel727	blahdeblah: well, so this repo at server is THE repo, which all developers use to push their changes to?	10:40
	blahdeblah: if so, then faq non-bare	10:40
Gitbot	blahdeblah: Pushing to non-bare repositories is discouraged. See https://git.wiki.kernel.org/index.php/GitFaq#non-bare	10:40
	← cilly left	10:40
blahdeblah	Yes, although the only developer is me.	10:40
	→ cillybabe joined	10:40
	→ artefon joined	10:41
albel727	blahdeblah: read the part about updating checked out working copy with post-receive hook	10:41
	← dharrigan left	10:41
albel727	(post-update, to be specific)	10:41
blahdeblah	I know that non-bare repositories are discouraged, so i was trying to work out a way i could push to a bare repo, and have that automatically propagated to the working copy in production.	10:43
albel727	blahdeblah: that's why I pointed you specifically to read the part on "what if you still want to do it" =)	10:43
	there is a hook linked right on that page. with it, any change, introduced by push, that updates specific branch, will be immediately checked out.	10:44
blahdeblah	And is it OK for that specific branch to be the trunk/master/whatever the official git name for it is?	10:44
	→ sshc_ joined	10:44
albel727	blahdeblah: it can have any name	10:45
	→ giallu joined	10:45
	→ goshawk joined	10:45
albel727	blahdeblah: as long as it is the one, which is currently checked out	10:45
	← sshc left	10:45
blahdeblah	ok	10:45
	← aah left	10:45
	← cillybabe left	10:45
	→ cilly joined	10:46
blahdeblah	I would have thought this was a very common question - do people normally just schedule a frequent cron job instead of trying to be event-driven?	10:46
albel727	blahdeblah: of course, preferred practise might be creating specific branch like "live-server", check out it at server, and merge changes into it and push from clients. normally people who know about git hooks do it with post-update =)	10:46
	← Aiden left	10:46
albel727	s/practise/practice	10:47
blahdeblah	The server will never have anything except the master branch	10:47
albel727	(Personally I think it's the best way to do it, in this case)	10:47
blahdeblah	So i don't think it's necessary to have multiple branches.	10:48
	← Anti-X left	10:48
	← Arthur left	10:48
albel727	blahdeblah: well, it can have all number of branches (1..infinity) that it wants. it's only the existence of checked out branch, that matters. If you don't need multiple branches - don't use them.	10:49
	← thierryp left	10:49
	→ thierryp joined	10:49
blahdeblah	OK - no worries.	10:49
	I'm only learning git for this now because i realised i need to scale it out to multiple servers, each with private modifications. If it weren't for that, i'd still be perfectly happy with the old centralised tree view of the world that svn provides.	10:49
	→ Arthur joined	10:50
	→ zombor joined	10:50
albel727	blahdeblah: hmm, I see. well, bottomline: You're trying to use git as a deployment tool (which it is not). But if you still want to do it with git, this is a way (better that cron job, imo).	10:51
	s/better that/better than/	10:52
	→ Anti-X joined	10:52
	← MetaCosm left	10:52
	← zombor left	10:52
blahdeblah	Not really a deployment tool, because having a revision history on local modifications which are private to that server is essential. I'm using it as a way to share as much code/config as possible between my clients while still maintaining the privacy of each client.	10:53
	→ jojojo joined	10:53
	zz_rubydiamond → rubydiamond	10:53
	← jojojo left	10:55
	→ marcob joined	10:56
	← winterweaver left	10:57
albel727	blahdeblah: No, exactly like a deployment tool. Task of having revision history (which is solved by git) is in no way connected to unrolling specific revisions from specific git repositories to specific servers. See it that way: you don't really need git repositories on your servers. You might as well trigger a routine on your local computer, which will scp/rsync needed revision files there.	10:57
	blahdeblah: or alternatively, you'll have a server, which will have your central repo where you push to. and then it triggers scp/rsync to all other servers (deployment).	10:58
	→ ehabkost joined	10:58
	→ zombor joined	10:59
	→ bitkiller joined	11:00
	← bushwakko left	11:01
	← zombor left	11:02
blahdeblah	albel727: Perhaps i haven't explained myself well enough. There will still be local edits on those servers which must be revision-controlled.	11:02
albel727	blahdeblah: hmm. wait a sec. so you will do some development not only on your laptop, but directly on the server too?	11:03
blahdeblah	It's not that they're all identical copies of a public repo - they have private local configurations which will never be shared; they will also have some parts of the tree just pulled from a public server regularly.	11:03
	albel727: yes	11:03
	← kpreid left	11:03
albel727	blahdeblah: that totally changes everything.	11:03
blahdeblah	I figured i mustn't have explained it clearly... ;-)	11:04
	← icwiener left	11:04
blahdeblah	I have a bunch of clients who use the same application; they share a bunch of modules, but how those modules are put together is different in each case, and there are some private (i.e. non-published, proprietary to that client, etc.) changes as well.	11:05
	← awilkins left	11:05
	→ dharrigan joined	11:05
blahdeblah	So i need at least a working copy on my laptop, a public server from which they can pull public code, and a working copy on their servers. But i would like to be able to do the push to public, pull from public in an automated manner, as efficiently as possible.	11:06
	→ fivetwentysix joined	11:07
albel727	blahdeblah: now you completely confused me. do you have the need to push to their private repos, or not?	11:08
blahdeblah	I want to get changes that i push to my repo applied to their private working copies as quickly and as efficiently as possible.	11:09
	I don't care about pushing directly to the private repos, although i'm not averse to it if it happens to be the best way to solve the problem.	11:09
albel727	blahdeblah: hmm. now, if their servers are not related to you in any way (i.e they're customers, and you don't have superuser access to them), then them setting up cron job is the only scalable way.	11:11
	← dharrigan left	11:12
albel727	blahdeblah: you can have a post update hook wich will send them mail/other notification about you updating your public repo, which might trigger their pull jobs, but otherwise...	11:12
	← merlin83 left	11:13
albel727	blahdeblah: can the following be an alternative: on your laptop you have a git repo with one branch, incorporating tree, common to all servers, and a single branch per server with its local modifications. when you upgrade common branch, you rebase your per-server branches on top of it and push them to their respective repos. (the purpose is to get rid of editing on servers directly)	11:13
	← kanru left	11:13
blahdeblah	I have root access to the servers, but they are not strictly related to me or to my public repo. If i quit tomorrow because i'm sick of computers (not totally out of the question ;-), i want them to be able to function independently on their own.	11:13
	→ zorzar_ joined	11:14
	→ mett` joined	11:15
albel727	blahdeblah: well, in that case, it is only their authority/responsibility to have regular pulls from your repo. as I've said, you can make a post update hook, which will send them proper messages/will trigger their cron job/etc, but that's it.	11:15
	← zorzar left	11:16
blahdeblah	Maybe a regular cron job is the simplest & easiest way to sort it	11:17
	Incidentally, what does "rebase your per-server branches on top of it" mean?	11:18
	← jas4711 left	11:18
albel727	blahdeblah: yeah, looks like for your case a simple cron job would be ok. what is rebase? man git-rebase	11:18
Gitbot	blahdeblah: the git-rebase manpage can be found at http://git.or.cz/man/git-rebase	11:18
	blahdeblah pats the nice little Gitbot - good computer!	11:19
	← twopak left	11:20
albel727	blahdeblah: but "merge changes from common branch into each per-server branch" might be an even better way (I wanted to mention it too, it's just that I ran out of space due to irc line size constraint)	11:20
blahdeblah	:-)	11:21
	Merge just applies the same changesets to a different branch, doesn't it?	11:21
	→ twopak joined	11:22
albel727	blahdeblah: yep. why?	11:22
	twopak → Guest12623	11:22
blahdeblah	just trying to understand the difference between merge and rebase	11:22
albel727	blahdeblah: main difference is in form of history produced, rather than in the result obtained. the result will be the same, but the history will be kept linear, when using rebase.	11:24
blahdeblah	yep - makes sense	11:24
	Just been reading the rebase manual	11:24
	← Guest12623 left	11:25
blahdeblah	Does it make any difference in terms of the size of the local repo? If most of the changes are on the public repo (which i expect they will be), then it seems to me that rebase would result in more efficient use of space in the long run...	11:26
albel727	blahdeblah: well yeah, there are some other peculiarities in usage of rebase (like rebasing public history leading to horrendous result), but some people prefer to rebase their local changes on top of upstream. and rebase is a bit more natural to cvs/svn users .	11:26
blahdeblah	Yeah - rebase reads like merging the changes from trunk into a branch	11:26
albel727	sorry, have to go for a min. brb.	11:27
blahdeblah	no problem	11:27
	Thanks a lot for taking the time to understand my situation, albel727.	11:27
doener	merge merges, creating a single new commit. Rebase creates a whole new line of history, with completely new commits (all untested, of course)	11:28
albel727	blahdeblah: you're welcome. =) it would be have been even better if I could actually advise something good, though. =/	11:30
doener	also note that unlike e.g. svn, git's merge is about merging lines of history, that changes are merged is just a side-effect, while svn's merge is just about merging changes	11:30
blahdeblah	albel727: at the moment, i'll settle for "good enough". ;-)	11:31
albel727	blahdeblah: but speaking of size - no, I expect size differences to be negligible. the only difference is that merge sometimes (most of the time) creates an additional "merge commit", but that is negligible.	11:32
blahdeblah	doener: My carbon-based semantic parser failed on your last comment.	11:32
	→ icwiener joined	11:33
albel727	blahdeblah: I don't see why one would be worrying about size of a repository anyway, unless you're keeping big binary files in it.	11:33
blahdeblah	Old habits die hard :-)	11:33
	Same reason that it's still useful to run JavaScript compressors - they're only small files, but every little bit counts!	11:34
	→ awilkins joined	11:34
albel727	=)	11:34
	blahdeblah: doener means, that svn doesn't actually have a concept of "lines of history". branch in svn - is no more than a versioned copy in a subdirectory. it doesn't keep history of merges (which change came from where, etc). git does.	11:35
	← Adaptee left	11:37
blahdeblah	Yeah - OK. I think i understand. In this environment, though, lines of history are at best irrelevant and at worst annoying distractions.	11:37
FabianB	albel727: since svn 1.5 it has basic branch tracking, but it's quite simple and only helps a little for merging	11:38
	→ liu joined	11:38
albel727	FabianB: Didn't know that. Thanks =)	11:39
	→ nsdk joined	11:40
	→ intripoon joined	11:41
	→ cannonball joined	11:44
	← cannonball left	11:44
	→ cannonball joined	11:44
	→ jcdny joined	11:44
	→ kpreid joined	11:44
albel727	blahdeblah: well, I agree with you on that. But then, again, it's because you use git improperly. I'd say, you actually need deployment/software update system, but you try to re-purpose VCS to do it instead.	11:45
	→ artemb joined	11:46
	← TuxCM left	11:46
artemb	Hi! There is a project on github whith two branches. I have a clone of the master. How can I switch to the other branch?	11:46
parasti	artemb: what you have is a clone of the repo, not a clone of one branch (master)	11:47
	artemb: so just switch to the other branch	11:48
	→ roop joined	11:48
parasti	artemb: list with "git branch -r", switch with "git checkout" (e.g., git checkout -b otherbranch origin/otherbranch)	11:48
	← jelmer left	11:48
parasti	artemb: or perhaps "git checkout --track origin/otherbranch" as that sets up the config correctly	11:49
artemb	ok, cool. Thanx, parasti	11:50
	→ [1]xxoxx joined	11:50
blahdeblah	albel727: The application i'm working with (http://www.puppetlabs.com/) is itself the deployment tool, so your summation of the situation is kind of ironic to me. Git's role in the process is to manage the configuration & code of the deployment server. Sharing code is what git is good at, so it doesn't feel like a bad fit to me...	11:51
	parasti: What's the difference between checkout --track and clone?	11:52
	man git checkout	11:52
	Where's the bot gone?	11:52
	man git-checkout	11:52
Gitbot	blahdeblah: the git-checkout manpage can be found at http://git.or.cz/man/git-checkout	11:52
blahdeblah	heh	11:52
	← goshawk left	11:53
	← xxoxx left	11:53
	[1]xxoxx → xxoxx	11:53
parasti	blahdeblah: checkout is a local operation	11:53
	→ chronos joined	11:54
	→ xiangfu joined	11:54
blahdeblah	So does "upstream" not necessarily imply a remote repo?	11:54
parasti	blahdeblah: also, checkout has nothing in common with clone	11:54
	blahdeblah: so the "difference" is "everything"	11:54
	← Arthur left	11:54
	→ Arthur_ joined	11:55
albel727	blahdeblah: haha. ironic, indeed. =) well, git might be ok doing that, but making deployment server bootstrap/update itself would be quite beautiful, imo =)	11:55
blahdeblah	albel727: Incestuous, more like... ;-)	11:55
albel727	haha	11:56
	blahdeblah wonders why there's nothing about "upstream" in the FAQ that is so liberally discussed in the manual pages	11:56
parasti	blahdeblah: first you have to realize that very few operations actually interact with a remote repo	11:56
Rhonda	I need to reset a ref for a branch in a bare repository. I think git symbolic-ref would be the tool, but I'm unable to figure out what argument it requires.	11:57
parasti	blahdeblah: what Git does is fetch a remote repo's branches into its own repo, and all the operations happen on those copies	11:57
	→ radioman-lt joined	11:57
	→ shamrock joined	11:57
parasti	blahdeblah: these can be listed with git branch -r (for "remote-tracking branch")	11:57
	→ vaishalitrivedi joined	11:58
parasti	blahdeblah: even your local master is created by clone off of the copy called origin/master	11:58
	blahdeblah: and origin/master is updated by git fetch (which is used also when you do git pull)	11:58
	Rhonda: didn't you ask this some days ago...	11:59
albel727	blahdeblah: yeah, I see the problems here, like "what to do, if some update brokes the updater and it will be unable to download subsequent fixes?", but otherwise, that's e.g how windows updates work =)	12:00
blahdeblah	parasti: So there is a considerable difference between tracking a remote repo, and the origin?	12:00
	→ _pingu joined	12:00
	← seanius left	12:00
Rhonda	parasti: Likely, though in there it was for the default branch so I got away with using git reset	12:00
blahdeblah	albel727: yeah, not to mention that puppet doesn't do revision history.	12:00
parasti	blahdeblah: origin really means nothing, it is the default name for the remote configured by clone	12:01
Rhonda	parasti: But your memory is great - it must has been more than a week ago.	12:01
parasti	blahdeblah: it refers to a remote repo, there's no difference	12:01
	Rhonda: anyhow, git symbolic-ref acts on symbolic refs (refs pointing to refs, such as HEAD)	12:01
	Rhonda: what you probably want is git update-ref (git update-ref -h)	12:01
Rhonda	Ah, thanks.	12:02
blahdeblah	I think my brain is full for today. Thanks a lot for the help and understanding, parasti and albel727. Good night to you all!	12:02
albel727	blahdeblah: well, it shouldn't revision its source code itself. that's what your public puppet repo is for. but it should keep program configuration preserved/revisioned in some database, shouldn't it?	12:02
	blahdeblah: oh, good night, then =)	12:02
	→ rhct joined	12:03
	← artemb left	12:03
	← workmad3 left	12:03
	→ mtkd joined	12:03
	← mtkd left	12:03
	→ mtkd joined	12:04
Rhonda	parasti: Thanks, that did the trick! :)	12:04
	← Uierth left	12:05
	← nsdk left	12:05
	→ d0k joined	12:06
	→ jonesy joined	12:07
	→ seanius joined	12:07
	→ joe85 joined	12:09
	← mtkd left	12:09
	→ Ryback_ joined	12:10
	← Ryback_ left	12:10
	→ Ryback_ joined	12:10
	← Arthur_ left	12:11
	← Chaser_ left	12:11
	→ sebersole joined	12:12
	→ Chaser_ joined	12:12
	→ Arthur_ joined	12:13
	→ jaql joined	12:16
	← jcdny left	12:17
	← sebersole left	12:18
	→ sebersole joined	12:19
	← sebersole left	12:19
	→ sebersole joined	12:19
	→ twopak joined	12:19
	← sebersole left	12:19
	twopak → Guest31708	12:19
	→ sebersole joined	12:20
	→ rson joined	12:22
	→ lvianna joined	12:22
	→ scarabx joined	12:22
	sitaktif → sitaktif__	12:25
	← icwiener left	12:25
	→ richardbronosky joined	12:26
	← jaql left	12:26
	← thierryp left	12:26
	← vaishalitrivedi left	12:27
	→ mbroeker joined	12:28
	→ smuf joined	12:29
	← allenwei left	12:30
	→ thierryp joined	12:31
	sitaktif__ → sitaktif_	12:31
	→ jaql joined	12:32
	→ IslandUsurper joined	12:32
	← nonno left	12:33
	→ blee_ joined	12:35
	← _pingu left	12:35
	→ kenneth_reitz joined	12:36
	Putr → Putr\|Away	12:37
	← Anti-X left	12:38
	← blahdeblah left	12:40
Aleric	How do I delete remote branches again?	12:40
jast	git push --delete remote branch	12:40
Aleric	Thank you jast :)	12:40
	I thought it was with a ':' somewhere :/	12:40
albel727	Aleric: yeah it is possible with : too	12:41
	→ daedra joined	12:41
albel727	Aleric: --delete option is only a recent addition, actually	12:41
daedra	how might I create a bare repository from an existing one?	12:41
albel727	Aleric: old school way is git push <repo> :<branch>	12:42
	→ TSK joined	12:42
albel727	daedra: man git clone --bare	12:42
	→ thiago joined	12:43
daedra	albel727: ok I've run git --bare init	12:43
	→ dvaske joined	12:44
daedra	inside my_project.git	12:44
	→ mplb joined	12:44
albel727	daedra: or alternatively, just "git init --bare" on remote server, then in existing repo do "git remote add <remotename> <url>", and then git push <remotename> <branchname> will do its trick.	12:44
	daedra: ("git --bare init" == "git init --bare", in case you're wondering, so just go on with git remote add..., etc)	12:45
	→ TomV-415 joined	12:45
albel727	daedra: if that is intended to become THE bare repo, where you'll push to, then convention is to call this repo "origin" (hence, git remote add origin <url>)	12:46
	← Pew left	12:47
	← robinbowes left	12:50
TomV-415	I could use a bit of help with git - I am using git with heroku, and to push an app to heroku, the branch must be 'master'. Right now, I need to roll back the app on heroku... so I need to 'roll back' the 'master' branch.	12:50
	then push it to heroku..	12:50
	Seems easy, but not really clear how to revert and then push to a remote host.	12:51
	← TSK left	12:51
TomV-415	Can anyone help me?	12:51
jaql	are there any git based wiki's that would be good to use for something like site documentation?	12:51
	any good git based wikis*	12:51
	← shiba_yu36 left	12:51
daedra	albel727: is it possible to add a local "remote server"	12:52
	i.e. just a directory on the same machine?	12:52
albel727	daedra: sure, why not	12:52
daedra	how do I git remote add origin?	12:53
albel727	daedra: just use direct path, instead of <url>	12:53
daedra	ah direct path. Got it	12:53
	thanks	12:53
albel727	daedra: both relative and absolute paths work. there is also possibility to do the same using "file://<path>" url (but there is a subtle difference from just specifying path).	12:54
TomV-415	If I do git reset 50200324d25477975cf7b4774bf02d7..... then commit, will that effectively set the master to the state of 50200... ?	12:55
	→ shiba_yu36 joined	12:55
TomV-415	so I can push master to heroku?	12:55
daedra	albel727: hmm I get fatal: '/my/abs/path/bare_repo.git': unable to chdir or not a git archive	12:55
	fatal: The remote end hung up unexpectedly	12:56
	wagle_ → wagle	12:56
TomV-415	or is 'git revert' a better choice?	12:56
daedra	after setting the remote origin to /my/abs/path/bare_repo.git	12:56
albel727	TomV-415: yes, git reset is a way, but it is not recommended (and push will fail without additional options). git revert is a better choice indeed.	12:56
	→ pgib joined	12:56
albel727	daedra: is that what git push tells you?	12:56
daedra	yes	12:57
TomV-415	albel727: what options do I need to assure the push does not fail?	12:57
	→ jason237 joined	12:57
	→ hyperair joined	12:57
albel727	TomV-415: it won't fail with git revert. with git reset it is --force option, but you might lose some changes from other developers. see faq non-ff	12:58
Gitbot	TomV-415: Your push would lose changes on the remote. See https://git.wiki.kernel.org/index.php/GitFaq#non-ff	12:58
	→ flazz joined	12:58
	← scarabx left	12:58
	→ jcdny joined	12:58
	→ tsb joined	12:59
albel727	daedra: you're sure, that bare_repo.git is an initialized bare git repo, and it resides at exactly that path? if yes, pastebin me .git/config from both of these repos, and also your exact "git remote add" command.	13:00
TomV-415	albel727: I'm currently the only one working on this, so no need to worry about other developers, just need to get the code on heroku back to a earlier state. It looks like revert is good for going back one commit, but I need to go back quite a bit farther..	13:00
tsb	How can you update an exiting remote branch? My git pull doesn't work since the IP has changed.	13:00
	← richardbronosky left	13:00
TomV-415	tsb: you can delete and add remotes using git remote	13:00
	← intripoon left	13:00
albel727	TomV-415: well, I'm not sure about Heroku workings (and how it will react to that) but yeah git reset + git push --force should help you.	13:01
	→ vp2010on joined	13:01
SethRobertson	Depending on your version of git. You might need to adjust the git-config on the remote system	13:01
	→ coderdad joined	13:01
SethRobertson	I suggest making a backup with clone to prevent unintended disaster	13:02
	`git reset --hard SHA` or git reset --hard HEAD^^^^^^^^^^^^` or whatever also works	13:02
	← mplb left	13:02
tsb	SethRobertson, you aren't talking to me right?	13:03
	TomV-415, okay, I'll have a look thanks	13:03
SethRobertson	TomV-415	13:03
albel727	tsb: git remote set-url <remote> <newurl>	13:03
	tsb: man git-remote	13:03
Gitbot	tsb: the git-remote manpage can be found at http://git.or.cz/man/git-remote	13:03
	← Chaser_ left	13:04
TomV-415	SethRobertson: Thanks! Good idea to get the clone first!	13:04
	→ Chaser_ joined	13:05
daedra	albel727: my fault.. I was specifying the wrong path	13:05
	apologies :)	13:05
	→ mtkd joined	13:05
albel727	daedra: no problem. glad it worked =)	13:05
daedra	thanks	13:05
TomV-415	SethRobertson: Since I just created a local branch, will that survive the reset hard?	13:05
	→ verto joined	13:06
TomV-415	SethRobertson: (so I can continue to fix the problem.)	13:06
	→ merlin83 joined	13:07
	→ yngress joined	13:07
	← mtkd left	13:07
SethRobertson	TomV-415: It will survive	13:07
TomV-415	SethRobertson: Thanks for your help!	13:07
	albel727: thanks for your help!	13:08
albel727	TomV-415: you're welcome =)	13:08
	← fivetwentysix left	13:09
	→ q0rban joined	13:10
TomV-415	FYI: for heroku, the --force flag did the trick!	13:11
	→ Slix` joined	13:13
	→ corni joined	13:14
	← corni left	13:14
	→ corni joined	13:14
	← jason237 left	13:14
	← smuf left	13:15
	← mmendez534 left	13:15
	← dvaske left	13:15
	→ dvaske joined	13:15
	→ zombor joined	13:15
	← kpreid left	13:15
	→ brntbeer joined	13:16
	→ Anti-X joined	13:18
	← tungtn left	13:18
	→ pgregory joined	13:19
pgregory	hi, heeeeelp	13:20
	→ jason237 joined	13:20
pgregory	some moron with a pick axe wiped out our power just as I'd pressed return on a git rebase operation!	13:20
	now I'm back up, but git is saying that the repo is "Not a git repository"	13:20
	is there any way to recover this?	13:20
	→ koltroll joined	13:21
pgregory	the .git folder is there	13:21
	← AlexanderS left	13:21
	← radioman-lt left	13:21
jast	lesson learned: use a different filesystem/journalling setting next time ;)	13:21
	→ kpreid joined	13:21
jast	is HEAD still there?	13:21
pgregory	jast: yeah right, on Windows, how exactly?	13:21
jast	ah well	13:21
	out of luck, I guess	13:21
pgregory	yes HEAD is there.	13:21
	→ smuf joined	13:21
jast	does it have some kind of usable content	13:22
	?	13:22
pgregory	but it still says not a repository	13:22
	I'd just done a git stash, so all my recent work is buried in the repo, I need to get access to that stuff.	13:22
jast	my guess is that the contents of HEAD got clobbered	13:23
	so, about that question? :)	13:23
	← TomV-415 left	13:24
	→ AlexanderS joined	13:24
pgregory	jast: so, you don't thik it's possible to recove anything from this?	13:25
	all I need is the stash	13:25
	→ kylehayes joined	13:26
jast	I never said that. would you please answer the question?	13:26
	← roop left	13:28
pgregory	which question?	13:28
	→ fivetwentysix joined	13:28
	← xiangfu left	13:28
jast	152207 < jast> does it have some kind of usable content	13:28
	→ galderz_ joined	13:28
	← fivetwentysix left	13:28
jast	like, you know, text or a sha1	13:28
pgregory	it looks like rubbish, if by 'it' you mean HEAD	13:28
	→ gilaniali joined	13:28
jast	in that case, overwrite the contents with exactly this: ref:refs/heads/master	13:29
	or whatever other branch you have	13:29
	ideally whatever you had last checked out	13:29
	→ kjy112 joined	13:29
	← galderz left	13:29
jast	then, to be sure that there are no traces of the rebase left, git rebase --abort	13:29
	galderz_ → galderz	13:29
kjy112	hello, how do you remove one of your repositories?	13:29
jast	whoops	13:29
albel727	pgregory: try "git-symbolic-ref HEAD refs/heads/master"	13:29
jast	actually, "ref: refs/heads/master"	13:29
SethRobertson	kjy112: rm -rf directory-name; or is this question more subtle?	13:30
albel727	SethRobertson: damn, I was going to make this exact joke, but you were faster =)	13:31
kjy112	oops sorry i thought this channel is github	13:31
	→ mfacenet joined	13:31
	→ drubies_ joined	13:31
	→ tama joined	13:31
pgregory	jast: doing that has at least allowed git to identify the repository, it now says error: bad signature fatal: index file corrupt	13:32
SethRobertson	git fsck --full	13:32
jast	rm .git/index; git reset	13:32
	← tsb left	13:32
	→ icwiener joined	13:33
	dwmw2_gone → dwmw2	13:34
	→ lzyy joined	13:34
pgregory	jast: everything I do results in error: Invalid HEAD	13:34
	← ph^_ left	13:34
	← MikhX left	13:34
	→ MikhX joined	13:34
	→ bhosie joined	13:36
	→ whaley joined	13:36
jast	pgregory: did you use master as in my example or a different branch?	13:37
	→ yhager joined	13:37
pgregory	jast: master	13:37
	I've managed to get access to the stash, and get a diff out of it.	13:37
	→ icwiener_ joined	13:37
pgregory	so I'm now re-cloning from the origin, hopefully I can re-apply the patch and get back to business.	13:37
jast	oh, so we're not going to recover the rest... okay	13:38
	← Slix` left	13:38
pgregory	jast: I'm happy to try, but it was basically the changes in that stash that I really needed.	13:39
	← mbroeker left	13:39
pgregory	if you have any other ideas, I'm all ears, good to know these things anyway.	13:39
	← daedra left	13:39
	← icwiener left	13:39
jast	well, I think the branch might be broken, too	13:39
	does 'git show master' work without errors?	13:40
pgregory	fatal: ambiguous argument 'master': unknown revision or path not in the working	13:40
	tree.	13:40
	Use '--' to separate paths from revisionsfatal: ambiguous argument 'master': unknown revision or path not in the working	13:40
	tree.	13:40
	Use '--' to separate paths from revisions	13:40
	→ roop joined	13:40
jast	what does 'git rev-parse refs/heads/master' say?	13:40
pgregory	same	13:41
albel727	pgregory: "git show master --"	13:41
pgregory	but with refs/heads/master	13:41
jast	right, so the branch ref is likely clobbered, too	13:41
pgregory	fatal: bad revision 'master'	13:41
jast	well, how about: 'git show master@{0}'	13:41
	← dvaske left	13:42
pgregory	same, with the different name again.	13:42
	→ dvaske joined	13:42
	← shiba_yu36 left	13:42
albel727	pgregory: "git show refs/heads/master@{0}"	13:43
pgregory	just to be clear, I wrote the exact text 'ref: refs/heads/master' into the HEAD file in .git.	13:43
jast	yeah, that part should be fine	13:43
	but apparently the thing that stores what master itself points to is broken, too	13:43
	have a look at .git/logs/refs/heads/master. gibberish or textual contents?	13:44
pgregory	hmm, clutching at straws, but I just copied .git/ORIG_HEAD to .git/HEAD and it seems to be better.	13:44
	gitk at least comes up now	13:44
jast	what's in that now?	13:44
	← corni left	13:45
pgregory	it's a commit id	13:45
jast	ah	13:45
	well, that won't get us the branch back, I'm afraid	13:45
pgregory	refs/heads/master looks like gibberish	13:46
jast	how about logs/refs/heads/master	13:46
pgregory	sequence of ids	13:46
albel727	I guess "copy orig_head refs/heads/master", and "git update-ref refs/heads/master refs/heads/master@{0}" will do the trick (as long as reflog is ok)	13:46
	→ shiba_yu36 joined	13:46
getxsick	is it possible to have a branch with few changes which is visible for other branches (e.g. master) but when someone pulls the master, these changes are not pulled?	13:47
	← verto left	13:47
pgregory	albel727: shall I give that a go?	13:47
albel727	jast: I'd do that in your place, but jast is the main advisor to you here, so I dare not to interfere.	13:47
	← dvaske left	13:48
jast	yeah, go with that	13:48
	← kjy112 left	13:48
	→ dvaske joined	13:48
jast	I didn't think of that. elegant way of saving more manual work.	13:48
pgregory	fatal: Log .git/logs/refs/heads/master is corrupt.	13:49
albel727	jast: sorry, I wanted to say pgregory: in previous msg.	13:49
getxsick	hm, anyone?	13:49
albel727	pgregory: that was the update-ref output, I guess?	13:49
pgregory	albel727: yes	13:49
	→ Bass10 joined	13:49
albel727	pgregory: well, then there is no helping it.	13:50
pgregory	albel727: ok, thanks for all the efforts, I'll rebuild from origin and apply my changes.	13:50
albel727	pgregory: wait	13:50
jast	getxsick: 'visible for other branches'?	13:50
pgregory	sincerely thanks for trying.	13:50
getxsick	jast: i mean to see these changes in e.g. master branch	13:50
albel727	pgregory: meaning, there is no helping it but looking int logs/refs/heads/master and picking correct sha	13:50
jast	getxsick: if they are in master, they are in master.	13:50
albel727	pgregory: open that log, and pick/memorize the second sha in the latest line.	13:51
getxsick	jast: i have branch b1 and master. i do some changes in b1, then i want to see them also in master. i can simply merge it but then if someone would pull master he get these changes either. how to avoid it?	13:51
albel727	pgregory: then open .git/refs/heads/master, and write that sha down here	13:52
jast	getxsick: well, you have to choose. either see them in master (in which case they are part of master) or don't (in which case they are not part of master and won't get merged in)	13:52
	→ mbroeker joined	13:53
getxsick	jast: so there is no chance to make it visible from master but forbid to pull these changes ?	13:53
pgregory	albel727: that's exactly what's alredy in refs/heads/master	13:53
	← brntbeer left	13:53
jast	getxsick: nope	13:53
getxsick	;/	13:53
jast	pgregory: well, in that case you are probably done	13:53
pgregory	it seems to be logs/refs/heads/master that is corrupt, the last line is gibberish	13:53
jast	ah	13:53
getxsick	jast: anything like svn:external or something?	13:53
jast	getxsick: http://git.or.cz/man/git-submodule	13:53
albel727	pgregory: well, remove that last line then.	13:53
pgregory	albel727: ok	13:54
	albel727: update-refs now returns safely	13:54
	or at least cleanly	13:54
albel727	pgregory: well, that should be it. see if you lost some changes from master. if not - then we're done, probably.	13:55
jast	does 'git show' now give you the most recent commit that you had?	13:55
pgregory	jast: it gives the commit just before the rebase	13:56
	→ rgr joined	13:56
albel727	pgregory: did you do git rebase --abort, btw?	13:56
pgregory	jast: so the rebase hasn't been applied, but if I try to it just comes back saying everythign is 'D'	13:56
	→ akahn joined	13:56
pgregory	albel727: No rebase in progress?	13:56
getxsick	jast: ok, but this is only for whole files. i would like to keep some changes like "foreing"	13:56
	← bhosie left	13:57
jast	getxsick: it's for directories into which you can map other repositories	13:57
albel727	pgregory: you're already trying to do another git rebase? what does "git status" say?	13:57
jast	just like svn externals, only that we don't support floating versions	13:57
pgregory	albel727: Not on any branch, but otherwise, as expected.	13:58
albel727	pgregory: oh, so do git checkout master, first.	13:58
pgregory	albel727: done, now rebase?	13:58
albel727	pgregory: yeah, now try it.	13:58
jast	sure, if you want to :)	13:58
getxsick	jast: let me show you a small usecase. in some files there are specific configs which depend on the repository location. unfortunately, these files are not only config files but also with some code which has to be maintain and sync. so i was thinking to make a special branch just to change a config data, but to avoid to sync it to other repos as it would break their configs	13:59
	→ xiong joined	13:59
	→ alberto56 joined	14:00
jast	the usual way to deal with stuff like that is to keep only a template for the config file in the repository	14:00
albel727	getxsick: I'm not sure I've completely understand your problem, but you may want to look at this: thomasrast.ch/git/local-config.html	14:00
	getxsick: http://thomasrast.ch/git/local-config.html	14:00
	dl9pf_ → dl9pf	14:01
	← dl9pf left	14:01
	→ dl9pf joined	14:01
	→ Kira joined	14:01
pgregory	albel727: and jast: magic, all seems well with the world again.	14:02
	→ aah joined	14:02
jast	cool.	14:02
pgregory	thanks sooooo much, saved my life!	14:03
	well, maybe that's a bit melodramatic, but saved a lot of pain at least :)	14:03
	← avinashhm left	14:03
albel727	pgregory: not so sure we helped that much. git repos are hard to kill. anyway, you're welcome =)	14:03
pgregory	albel727: that's really good to know, as I've just finishes persuading the powers that be at my company that we should switch completely.	14:04
	although as they are currently running CVS, it's not that hard a sell :)	14:04
albel727	lol	14:04
getxsick	albel727: yep. this is it!	14:04
pgregory	at least to anyone with half a brain.	14:04
	→ avinashhm joined	14:04
albel727	pgregory: too much of a requirement nowadays =P	14:04
	→ RoosterJuice_ joined	14:05
pgregory	:)	14:05
	albel727 is not sure, how much of a brain he has himself =D	14:05
pom__	Can I combine several commits into a single one?	14:05
getxsick	btw. interesting that git doesn't have it as a standard. something like 'global branch' or something	14:05
fr0sty	pom__: man git-rebase and look at --interactive	14:06
Gitbot	pom__: the git-rebase manpage can be found at http://git.or.cz/man/git-rebase	14:06
pgregory	pom__: let me take a stab at this, yes.	14:06
	but you should never do it if you've already pushed.	14:06
jast	pom__: sure, but note that it's usually not advisable if someone else has already obtained the separate commits	14:06
pom__	jast: They have not. It's only on a local branch.	14:06
RoosterJuice_	can I create a git repo of a subdirectory that is part of a bigger git repo? for example if I have a django app as part of a django project, can I make a repo of that subdirectory the same way?	14:07
pgregory	pom__: git rebase --interactive then	14:07
albel727	pom__: yeah, then git-rebase --interactive, or maybe even man git-merge --squash is the way	14:07
Gitbot	pom__: the git-merge manpage can be found at http://git.or.cz/man/git-merge	14:07
	← metaperl left	14:07
	→ xiangfu joined	14:07
jast	rebase -i is one of the supremely awesome tools in git :)	14:08
	albel727 concurs	14:08
	← ymasory left	14:09
	← freimatz left	14:10
pom__	I have a little trouble following exactly. "git rebase -i" outputs help. I guess it is because there is nothing to be commited currently.	14:10
	→ sivy joined	14:11
getxsick	HEAD^0 is the same as HEAD^ ?	14:11
parasti	pom__: it's because that is not valid syntax	14:11
jast	pom__: git rebase -i <commit before the first one you want to change>	14:12
getxsick	ah, it's detaching	14:12
jast	getxsick: HEAD^ = HEAD^1	14:12
pom__	To end up here, I did the following. 1) Started new branch. 2) Edited and committeed intermittently (some commits produce breakage) 3) My final merge I'd like to have as a single commit since the intermediate ones might produce non-working codebase.	14:12
getxsick	jast: and HEAD^0 = detach HEAD, right?	14:12
jast	that's one use case, yes	14:13
fr0sty	getxsick: HEAD^0 is invalid	14:13
jast	fr0sty: no	14:13
parasti	pom__: it's a good idea to check the manpage for a command you just learned of, because you won't always get copy paste solution in here	14:13
getxsick	fr0sty: ? i see it's working	14:13
	→ corecode joined	14:13
jast	pom__: if you just want to merge this stuff to another branch, just use git merge --squash	14:13
corecode	hey	14:13
	fr0sty shakes head	14:13
fr0sty	getxsick: 'git checkout HEAD' will detach.	14:13
corecode	i'm a bit unsure how i best merge another repo into a subtree	14:13
jast	fr0sty: no	14:14
corecode	-Xsubtree=debian/ doesn't do what i want	14:14
parasti	fr0sty: git checkout HEAD^0 will, HEAD will do nothing	14:14
jast	the goggles! they do nothing!	14:14
	fr0sty needs coffee, sleep, something...	14:14
	← steph021 left	14:14
parasti	come to think of it, my head usually does nothing, maybe I should try head^0	14:15
	← pgib left	14:15
	← g0bl1n left	14:15
	← pgregory left	14:15
jast	corecode: you could just use the third-party tool git-subtree :)	14:15
	the subtree merge strategy tries to make the merged stuff fit structurally with what's already in the target branch	14:16
	← devanjan left	14:16
	← psoo left	14:16
getxsick	how to change author of already comitted change?	14:16
	→ pigdude joined	14:17
jast	latest commit?	14:17
getxsick	jast: no	14:17
jast	well, then you'll have to use something like interactive rebase	14:17
	set the commit in question to 'edit', then when it stops at that point, git commit --amend --author="A B [email@hidden.address]	14:17
infinity0	any maintainers of git-svn in here?	14:17
corecode	jast: can i do it with plumbing?	14:17
jason237	is there a way to cherry-pick a sequence of commits in a single command?	14:18
	← icwiener_ left	14:18
corecode	no	14:18
	→ icwiener_ joined	14:18
	← artefon left	14:18
	→ dvaske_ joined	14:18
corecode	for loop or xargs -n1	14:18
jast	jason237: recent versions of git support cherry-picking ranges	14:18
jason237	jast: ah, so I see, thanks	14:18
	→ fowlduck joined	14:19
jast	corecode: yes, with git read-tree --prefix=	14:19
	that doesn't automatically become a merge, though	14:20
	*merge commit	14:20
corecode	yes, that's what i'm looking at	14:20
	do a merge -s ours after?	14:20
jast	no	14:20
kylehayes	A developer on my team is getting the following error from EGit when attempting git status: "fatal: unable to create '[path]/.git/index.lock' : File exists	14:21
jast	git merge -s ours --no-commit <other branch>, then the read-tree incantation, then commit	14:21
	→ boombatower joined	14:21
	← boombatower left	14:21
	→ boombatower joined	14:21
_ikke_	I have a php file which git sees as binary	14:21
	← dvaske left	14:21
jast	kylehayes: either another operation is still running or a previous command crashed. in the latter case, tell him to remove that file.	14:21
pom__	jast: --squash was the one I needed. Thanks all!	14:22
corecode	jast: thanks	14:22
kylehayes	jast: he does remove the file...is it due to him using the command-line and egit at the same time maybe?	14:22
jast	corecode: make sure to check that read-tree gives you the right result :}	14:22
_ikke_	I have a php file which git sees as binary. How can I tell git to treat the file as just a text file?	14:22
	→ winterweaver joined	14:22
	← shiba_yu36 left	14:22
jast	kylehayes: possibly. whenever an operation finishes it removes the lockfile, so it's only likely to happen if both are literally running at the same time. tell him not to do that. :}	14:22
	_ikke_: http://git.or.cz/man/gitattributes	14:23
kylehayes	jast: haha, excellent. Thanks!	14:23
	→ converge_ joined	14:23
_ikke_	jast: But how comes it sees the file as binary in the first place?	14:23
jast	_ikke_: the detection is not particularly smart	14:24
	I don't recall the details right now, though	14:24
	→ jas4711 joined	14:25
	← gilaniali left	14:25
	← aah left	14:26
[exa]	_ikke_: can you post a sample of the PHP? and check with xxd whether there is no utf8 mark on the beginning?	14:26
	← dvaske_ left	14:27
avinashhm	hi, is it possible to reset to 1 commit previos .. something like HEAD-1 ... ??? i tried 'git reset --hard HEAD-1' .. .not working .. any ideas ??	14:27
_ikke_	[exa]: There doesn't seem to be a utf8 mark, just the contents of the file at the first position	14:27
	avinashhm: HEAD^1	14:27
jast	avinashhm: HEAD~1 or simply HEAD~	14:28
_ikke_	oosp	14:28
	← jason237 left	14:28
avinashhm	_ikke_, jast .. thanks a lot .. :-)	14:28
jast	'-' is a valid character in branch names, that's why we use the weird squiggly instead	14:28
	^1 is the same as ~1, but the same is not true for ^2 and ~2	14:28
	→ dvaske_ joined	14:28
	→ MattDiPasquale joined	14:28
avinashhm	jast, is it ... ?? I actually thought it was ..	14:29
_ikke_	[exa]: This is the output of the file: http://pastie.org/1255587	14:29
	← sonnym left	14:29
jast	^1 = first parent, ^2 = second parent. ~1 = first parent, ~2 = first-first grandparent	14:30
	and so on	14:30
	→ cytrinox_ joined	14:30
	→ dunpeal joined	14:30
	→ dels joined	14:30
	← esc left	14:30
avinashhm	so to reverse 3 commits , what should we use ??	14:30
	→ esc_ joined	14:31
dunpeal	avinashhm: git revert	14:31
jast	I don't think that was the question :)	14:31
_ikke_	head~3	14:31
[exa]	_ikke_: seems perfectly valid... is there some weird UTF stuff in that?	14:31
dunpeal	heh, OK	14:31
	On bare repo /repo, I want to get the contents of file bar/foo.txt on reference ref. How?	14:31
_ikke_	[exa]: Doesn't seem so	14:32
avinashhm	so 4 commits head~4 and so on right ???	14:32
	→ burg joined	14:32
corecode	dunpeal: git show or so?	14:32
_ikke_	[exa]: There is a commit message where someone mention he had deleted some weird characters on the end, maybe that was the problem	14:32
	← esc_ left	14:32
jast	avinashhm: yes, amazingly ;)	14:33
avinashhm	jast, :-) ... hehe .. it took me a while to sink ...	14:33
dunpeal	corecode: hm, no workie	14:33
	→ FabianB_ joined	14:33
burg	Question: I used 'git svn clone -r9000' to start working on an SVN repository. Is it possible to do 'git svn blame' on prior revisions than the one I started the clone with? (e.g: 8000 ?)	14:33
dunpeal	avinashhm: guess how you refer to 5 commits back?!	14:33
	→ pgib joined	14:34
	← cytrinox left	14:34
	cytrinox_ → cytrinox	14:34
dunpeal	burg: no.	14:34
avinashhm	dunpeal, git reset head ~5 ..	14:34
	:-) ..	14:34
	→ dharrigan joined	14:34
dunpeal	burg: you're not importing the information, how is Git supposed to know?	14:34
	← FabianB left	14:34
burg	dunpeal: i was hoping it would use native SVN BLAME to do that	14:34
	→ rayners joined	14:35
burg	dunpeal: that's what I thought was "git svn blame" purpose was	14:35
dunpeal	burg: maybe it does, but if you want the history in git, you need to import it.	14:35
	→ esc joined	14:36
	← esc left	14:36
	→ esc joined	14:36
dunpeal	git svn blame might work against a server, but server may be gone.	14:36
	jast: saw my question? ;)	14:36
jast	dunpeal: git show ref:bar/foo.txt	14:36
	→ mtkd joined	14:37
	→ nizze joined	14:37
nizze	Hello all!	14:37
	← mtkd left	14:37
nizze	How can I remove these http://imgur.com/Mr4KB	14:37
	→ mtkd joined	14:37
jast	these what? the blue circles? :)	14:37
nizze	No	14:38
	The right side branch	14:38
	or sequence of commits	14:38
jast	looks like you're viewing several branches at the same time	14:39
nizze	Yes	14:39
	that's gitk --all	14:39
infinity0	how do i access the current commit message from a pre-commit hook?	14:39
	← sako left	14:39
	→ jmcantrell joined	14:39
sitaram	nizze: Ctrl-F5 :)	14:40
corecode	infinity0: git log?	14:40
	or so	14:40
nizze	It worked, but what happened?	14:40
jast	infinity0: it's run before the commit message is known. use the prepare-commit-msg hook if you want to muck around with the message.	14:40
infinity0	jast: ah ok thanks	14:40
nizze	Is ctrl-f5 different from plain f5	14:40
	?	14:40
	← giallu left	14:40
jast	yes, see the menu	14:41
	one is update, the other is reload	14:41
	→ spearce joined	14:41
jast	though I have no idea what the difference is	14:41
	← skoop left	14:41
jast	badly named if you ask me	14:41
	← dunpeal left	14:41
	→ sako joined	14:41
	← joe85 left	14:42
_ikke_	[exa]: I found a blob where there were a few 0 bytes on the end	14:43
nizze	Okay thanks guys	14:43
jast	all right, I gotta go. have a nice day, everyone.	14:44
nizze	Bye, thanks	14:44
	← burg left	14:45
	← rayners left	14:46
	← dharrigan left	14:47
	→ t0rc joined	14:48
	← lzyy left	14:50
	← xiangfu left	14:51
	→ sftp joined	14:52
	← jmcantrell left	14:54
	← Anti-X left	14:54
	← kylehayes left	14:57
infinity0	is there a pre-merge hook or equivalent?	14:57
	i've got a git repo mirroring an svn repo, and i want to reject merges of commits that haven't been svn-dcommited (because that screws with git-svn)	14:58
	← jas4711 left	14:58
	FauxFaux only just noticed that git push --all didn't push tags. :/	14:59
	← hobodave left	14:59
	→ Anti-X joined	14:59
	→ hobodave joined	14:59
	← SianaGearz left	14:59
	→ tj joined	15:01
fr0sty	infinity0: no pre-merge hook, sorry.	15:01
	tj → tjholowaychuk	15:01
	← dvaske_ left	15:02
infinity0	oh :( is there any other way to reject merges?	15:02
	→ dvaske joined	15:02
fr0sty	post-commit can write a nasty message...	15:02
infinity0	lol	15:03
	→ brajesh joined	15:03
fr0sty	you might be able to use the commit-msg hook	15:03
infinity0	hmm and then run git reset --hard HEAD^ lol	15:03
	→ rpg joined	15:03
	← hobodave left	15:04
fr0sty	it would have the branch-names in the message and you could investigate the commits there to see f they are dcommitted and fail there.	15:04
	← dvaske left	15:04
fr0sty	I'm not sure how reliably you can parse that template message though...	15:04
	→ dvaske joined	15:04
infinity0	i can get the parent SHA1s of the commit and check if they have the git-svn-id thing	15:04
	just need a hook to do it.. i'll try commit-msg but neither pre-commit nor prepare-commit-msg were called for automatic merges using "git merge"	15:05
	← koltroll left	15:05
	← teukka left	15:05
	← dvaske left	15:05
	→ dvaske_ joined	15:05
	← lamont left	15:05
infinity0	no, it bypasses commit-msg too :/	15:06
albel727	infinity0: hmm. are you sure you had set the executable bit on hook?	15:07
Kira	If I edit blah.txt, 'git add' it, and modify blah.txt again, is there a way I can recover what is in the index, discarding the later modification using a git command?	15:07
infinity0	yeah they're all executable	15:07
	it runs for "normal" commits	15:07
	just not "git-merge" with no conflicts	15:07
parasti	because there is no commit, it's a fast forward by default	15:08
	← liu left	15:08
infinity0	parasti: it's not a ff, it says "Merge made by recursive."	15:09
parasti	and you have a merge commit?	15:09
infinity0	yes	15:09
parasti	hm	15:09
	→ Theravadan joined	15:10
	← converge_ left	15:11
	← sambarino left	15:12
	→ carutsu joined	15:12
albel727	Kira: try "git checkout -- <file>"	15:13
	← marc_kdab left	15:13
	→ sonnym joined	15:13
	← kpreid left	15:14
	← ch077179_ left	15:15
	← cbreak_work left	15:16
fr0sty	infinity0: what is in your commit-msg? also, you are not passing --no-verify, right?	15:17
	← carutsu left	15:17
	→ wsclock joined	15:17
fr0sty	what happens if your commit-msg hook is: 'echo rejecting commit && false'	15:18
	→ rayners joined	15:18
	→ goshawk joined	15:18
Kira	albel727: looks like 'git checkout -m -- <file>' would do it.	15:18
	→ carutsu_ joined	15:19
	← akamike left	15:20
	← malfy left	15:20
	→ malfy joined	15:20
	← malfy left	15:20
	→ malfy joined	15:20
albel727	Kira: -m doesn't make any sense in this context. but if it works for you, who cares.	15:21
fr0sty	Kira, not necessary to use the 'merge' argument. it might actually do something other than you want.	15:21
	it sounded like you wanted to discard the later changes instead of merging with them.	15:21
	→ mxmdrone joined	15:21
albel727	Kira: well, actually, it might even hurt, now that I think of it. if you indeed wanted to do what I and fr0sty are thinking, then don't use -m.	15:23
	→ Yuffster joined	15:24
	→ CoinOpeBoy joined	15:24
Kira	ah, looks like I misunderstood that last sentence about -m in 'git help checkout'.	15:24
	→ beatak joined	15:25
infinity0	fr0sty: atm, just "echo ----commit-msg:$@" and yeah no --no-verify	15:25
	→ t3i joined	15:25
	Putr\|Away → Putr	15:26
	→ psoo joined	15:26
	→ fr0sty__ joined	15:27
	→ kipras joined	15:29
	← psankar left	15:30
	→ tizbac joined	15:30
	← fr0sty_ left	15:31
	→ aidan joined	15:31
aidan	I've got a folder in my git repos, and I want to make the contents of that folder the base of the repository. Is this possible?	15:32
albel727	infinity0: I'd advise against plain echo. maybe git merge redirects stdout > /dev/null. try echoing to a file, or return non-zero, so git aborts.	15:32
	→ err_ok joined	15:33
fr0sty	infinity0: add false to the end of the script so the commit will be rejected. Also, take albel727's suggestion and write to, or just touch a file	15:33
infinity0	oh ok i'll try that thanks	15:33
	← err_ok left	15:34
	← drubies_ left	15:34
	← icwiener_ left	15:35
	← sh1mmer left	15:36
	→ richardbronosky joined	15:36
	→ hobodave joined	15:37
infinity0	albel727, fr0sty: no, still nothing	15:38
	← Leemp left	15:38
	→ warbee joined	15:38
	← tama left	15:38
	← rayners left	15:38
	→ kanru joined	15:39
	→ rayners joined	15:39
aidan	I was thinking about moving everything manually, like git mv basefolder/foo ../, git mv basefolder/bar ../	15:39
fr0sty	aiden: man git-filter-branch and look at --subdirectory-filter	15:40
Gitbot	aiden: the git-filter-branch manpage can be found at http://git.or.cz/man/git-filter-branch	15:40
	→ icwiener joined	15:40
fr0sty	infinity0: what is your git version?	15:41
	→ GarethAdams\|Work joined	15:41
infinity0	git version 1.7.1	15:42
aidan	nice thanks	15:42
infinity0	i can send you my test script if you want	15:42
albel727	infinity0: sure, pastebin it.	15:43
	← OOPMan left	15:43
	← yuriks_ left	15:43
	→ yuriks_ joined	15:43
	→ ymasory joined	15:44
	← EricInBNE left	15:44
	← nuoHep left	15:45
infinity0	albel727: http://pastebin.com/An0My9ZD that's two files, the test script and the commit-msg hook	15:46
	ignore the "fuck up the history" parts, that's to test a bug in git-svn	15:46
	albel727 looks	15:46
	→ Uierth joined	15:47
infinity0	(you'll need svn btw)	15:47
	← kanru left	15:47
	← Zhonghua left	15:48
	→ nevyn^_ joined	15:49
	← thiago left	15:49
fr0sty	infinity0: what does 'git log --oneline --graph --decorate' show after you run that script?	15:49
	← rvsjoen left	15:49
	→ dukebag joined	15:50
	→ teukka joined	15:50
	← teukka left	15:50
	→ teukka joined	15:50
	← zain left	15:50
	→ boucman joined	15:50
	→ zain joined	15:50
infinity0	fr0sty: http://pastebin.com/CxdAq5M4	15:51
	← sako left	15:51
	← nevyn^ left	15:51
	→ rvsjoen joined	15:51
	← nevyn^_ left	15:52
	→ nuoHep joined	15:52
	→ bhosie joined	15:52
fr0sty	infinity0: where is the commit-msg file?	15:52
	ah, missed the ln lines...	15:53
	and it is executable, right? you should be seeing output every time you commit...	15:54
boucman	hello everybody	15:54
	→ khmarbaise joined	15:55
infinity0	fr0sty: yes, i am, except for the auto-merge	15:55
boucman	i'm trying to push some changes through git-svn, but I get an error message	15:55
infinity0	i guess git just isn't calling the hook for that	15:55
boucman	Le système de fichiers ne contient pas cet élément: File not found: transaction '47242-1', path '/changelog' at /usr/lib/git-core/git-svn line 572	15:55
	i've been googling a little but couldn't find anything relevant... anybody have a clue to help me investigate ?	15:55
	→ aspotashev joined	15:56
fr0sty	what if you did prepare-commit-msg instead?	15:56
	← mtkd left	15:56
	← vital left	15:57
	← zain left	15:57
	→ vital joined	15:57
infinity0	fr0sty: same	15:57
	← brajesh left	15:58
	albel727 can't seem to make prepare-commit-msg working for him, too, for some reason.	15:58
	→ _Vi joined	15:58
	→ lamont` joined	15:58
	→ LiamH joined	15:58
albel727	s/can't seem/seem unable to/	15:59
	albel727 shoots himself	15:59
fr0sty	infinity0: take the post-merge does reset --hard HEAD^ route, then.	15:59
	→ OOPMan joined	15:59
infinity0	yeah i'll try that	15:59
	not sure if it's "safe", but i'll try it	15:59
fr0sty	the commit is still in the reflog if you need it.	16:00
	← ka1ser left	16:02
	→ Heimidal joined	16:02
	← Heimidal left	16:02
	→ coreyballou joined	16:03
	→ Heimidal joined	16:03
	→ zain joined	16:04
	← cballou left	16:04
	coreyballou → cballou	16:04
	→ kpreid joined	16:04
	→ aah joined	16:04
infinity0	cool, thanks	16:04
	← _ollie left	16:05
	→ jason237 joined	16:05
	→ jsemar joined	16:05
	→ nevyn^ joined	16:06
	→ Leemp joined	16:06
	→ Dave^\| joined	16:09
	→ twomashi joined	16:09
	→ silver_hook joined	16:09
silver_hook	Hullo.	16:09
twomashi	Hola	16:10
silver_hook	Noob question after RTFMing, but being an idiot due to IANAC IAAL PEBKAC problem:	16:10
armenb	when I commit changes to a git repository, what does git use to represent those changes? does it use diff? or something else?	16:10
silver_hook	Do I have to add files that I changed every time I edit them?	16:10
fr0sty	armenb: read 'git for computer scientists' or 'git from the bottom up'	16:11
armenb	fr0sty: thanks!	16:11
fr0sty	they have good explanations of how git handles content.	16:11
	→ adamm joined	16:11
PerlJam	silver_hook: you're adding changes, not files.	16:11
silver_hook	PerlJam: Ah, OK. That explains it.	16:11
	← twomashi left	16:12
patrikf	silver_hook: only when you want to commit these changes	16:12
	→ steph021 joined	16:12
patrikf	silver_hook: and if you generally don't make changes that you don't want to commit, there's add -u and commit -a	16:12
	← t3i left	16:12
silver_hook	OK, and if I use 'git commit -a' then it adds only changes of files that are already tracked or every file in the dir?	16:13
PerlJam	the former	16:13
armenb	silver_hook: only files that are tracked (i.e. "git add"-ed)	16:14
silver_hook	Sweet.	16:14
armenb	silver_hook: you have to "add files [you] changed every time" because git doesn't commit directly from the working tree to the repository	16:16
silver_hook	Aaaaaand, therefore the difference between staged and unstaged changes is that the former were added to git, but not commited; while the latter were neither, right?	16:16
armenb	silver_hook: you manually register/confirm your changes in "the index" (think of it as a staging area)	16:16
	before you actually commit.	16:16
patrikf	silver_hook: correct. the term "staged" means what you mean with "added"	16:16
silver_hook	OK, and if I committed before adding changes, I basically just spammed my log?	16:17
patrikf	silver_hook: you usually say changes are "staged" or "added to the index" (which is the file that represents "staging area")	16:17
	→ khmarbaise_ joined	16:17
patrikf	silver_hook: commit won't let you make empty commits by default, so you must've done something (check with log -p or --stat)	16:17
silver_hook	patrikf: Hmmm, OK.	16:18
	Thanks for explaining :]	16:18
	← chrisshattuck left	16:18
	← philthno2 left	16:19
	← khmarbaise left	16:19
	khmarbaise_ → khmarbaise	16:19
	→ cytrinox_ joined	16:19
	← erez left	16:20
	→ marc_kdab joined	16:20
	Fissure1 → Fissure	16:22
	→ chrisshattuck joined	16:22
	← icwiener left	16:23
	← cytrinox left	16:23
	cytrinox_ → cytrinox	16:23
	→ tama joined	16:24
	← nevyn^ left	16:24
	→ necromancer joined	16:25
	→ gilaniali joined	16:27
	→ bauruine joined	16:27
	→ Noldorin joined	16:27
	→ jim joined	16:27
Noldorin	hi. can i control what Git does with my EOLs when i clone a branch?	16:27
	← rayners left	16:29
jim	hi... say I have a repo with a remote, but the address of the remote later changes... would I be able to edit the remote address somehow to accomodate the new location?	16:29
necromancer	yeah just git remote -r <name>	16:29
	i'm sorry i think it's git remote rm <name>	16:30
	and git remote add <name> <address>	16:30
	then git push <name>	16:30
jim	if I'm pushing to the remote	16:30
	if it's read-only, I'd pull?	16:30
	← ari-_-e left	16:31
armenb	wow. "git from the bottom up" is awesome. why the hell haven't i found this document before?	16:31
	→ corni joined	16:31
	→ rpbertp13 joined	16:31
	← priidu left	16:31
	→ nevyn^ joined	16:31
jim	armenb, url?	16:32
	→ kenichi joined	16:32
	→ sako joined	16:33
	← khmarbaise left	16:33
	→ _ollie joined	16:34
	IslandUsurper → IslandUsurperAFK	16:34
	← marcob left	16:35
	← dvaske_ left	16:35
	→ dvaske joined	16:35
	← jason237 left	16:35
	→ ka1ser joined	16:37
	→ joel_falcou joined	16:37
	← _ollie left	16:38
	← drako left	16:38
	→ dharriga1 joined	16:38
	← AlexP left	16:39
	→ jason237 joined	16:39
fr0sty	Noldorin: man git-config and look at core.eol	16:39
Gitbot	Noldorin: the git-config manpage can be found at http://git.or.cz/man/git-config	16:39
	→ orafu joined	16:39
Noldorin	ok cheers	16:39
fr0sty	jim: just edit .git/config and change the url	16:40
	← akahn left	16:40
fr0sty	armenb: because you never asked... :-)	16:40
	→ zxww joined	16:40
	← zxww left	16:40
Noldorin	hmm what's the git equivalent of hg update?	16:41
	of bzr update	16:42
	← hobodave left	16:42
	← _Vi left	16:42
	← dvaske left	16:43
	← andyl left	16:43
	→ dvaske joined	16:43
	→ AlexP joined	16:43
	→ _Vi joined	16:43
	← awilkins left	16:43
	← winterweaver left	16:44
	← _Vi left	16:44
	→ Evious joined	16:44
	← whaley left	16:44
	→ karamorf joined	16:46
	← Theravadan left	16:46
	→ _Vi joined	16:46
	→ khmarbaise joined	16:47
	← eldragon left	16:47
	← majoh left	16:47
	← _Vi left	16:47
	← kpreid left	16:47
sako	what does bzr update do/.	16:48
	just updates or merges as well?	16:48
	→ _Vi joined	16:48
	→ axiom joined	16:48
	axiom → majoh	16:49
	→ _ollie joined	16:49
	← dharriga1 left	16:50
	→ towski joined	16:50
	→ eldragon joined	16:50
	vp2010on → vpon	16:51
	← dvaske left	16:51
	necromancer → necromancer`brb	16:51
	→ dvaske_ joined	16:51
	← Schmallon left	16:52
Noldorin	sako, it fetches changes on the remote branch and updates the current branch with them	16:52
	hmm...i'm getting a lot of errors cloning this git repo, of the form:	16:52
	error: git checkout-index: unable to create file docs/perl/SWF::Action.html (Inv	16:52
	alid argument)	16:52
	← _Vi left	16:53
	← vpon left	16:53
sako	Noldorin: i think you want git pull	16:53
	as the "equivalent", though not sure if it is 100% equivalent in functionality	16:53
	→ Theravadan joined	16:54
sako	git pull fetches and merges	16:54
Noldorin	sako, that's what i thought, but if i for example delete local files and want to refresh the branch from the server, what do i do?	16:54
	in bzr it's just bzr update	16:54
	← _ollie left	16:55
	← nevyn^ left	16:55
sako	hmm	16:57
	you might need to do a git checkout	16:57
	← dvaske_ left	16:57
sako	i know to revert a file to what is on the server.. i just run git checkout filename	16:57
	not sure if that works for files you deleted locally	16:57
	→ dvaske joined	16:57
jason237	if you want to set your branch to the server branch, use git reset --hard origin/branch	16:57
Noldorin	jason237, ah, thanks	16:58
sako	nice i learned something new	16:58
	← mxmdrone left	16:58
	← khmarbaise left	16:58
	→ jherdman joined	16:58
sako	jason237: whats --hard do?	16:58
jason237	--hard updates the working tree as well	16:58
	← spearce left	16:58
	← gorlak1 left	16:59
jason237	note, that will throw away any local commits	16:59
Noldorin	i see	16:59
jason237	well, they'll still be in the repo, just not in your branch anymore	16:59
	→ davidfetter joined	16:59
	← davzie left	17:00
Noldorin	jason237, hrmm. so i want to clone a remote branch, but how do i give it clone.eol before it clones?	17:00
	→ Zephyrus joined	17:00
	← dvaske left	17:01
	→ drako joined	17:01
	← drako left	17:01
	→ drako joined	17:01
	→ dvaske joined	17:01
	← smuf left	17:01
jason237	well, I think core.eol only affects the working copy	17:01
	← bhosie left	17:01
	→ khmarbaise joined	17:01
Noldorin	jason237, yeah. the problem is that when it clones it uses native (windows) EOL	17:02
	but i want unix EOL	17:02
	i.e. lf rather than crlf	17:02
jason237	I'd try git checkout -f or git reset --hard HEAD to see if they update the EOL convention	17:02
	if not, move .git out of the way, blow away the working tree, move .git back and then try those commands again	17:03
	(or one of them)	17:03
Noldorin	will do	17:04
	→ jschoolcraft joined	17:05
	→ ludde joined	17:06
	→ programble joined	17:07
	→ gp5st joined	17:08
	← khmarbaise left	17:08
	← ludde3 left	17:08
	IslandUsurperAFK → IslandUsurper	17:08
	← chrisred left	17:08
	→ _ollie joined	17:09
gp5st	when i do git ls-files --unmerged it shows me the folders corresponding to submodules, (while in a submodule), how do i know what's unmerged?	17:09
	← drako left	17:09
	→ stalker joined	17:09
	→ icwiener joined	17:09
gp5st	i think it was my .gitignore file, tis' the only one that had >>>>hash in it	17:10
	but..fatal: 'commit' is not possible because you have unmerged files. :-\	17:10
	any ideas?	17:10
	→ koltroll joined	17:12
Noldorin	jason237, what do you mean blow away the working tree?	17:12
	ijuma__ → ijuma_	17:12
	← jason237 left	17:12
	← _ollie left	17:14
	← majoh left	17:15
	→ axiom joined	17:15
	axiom → majoh	17:15
	← Arthur_ left	17:15
	→ spearce joined	17:17
	← tstclair left	17:17
	→ metaperl joined	17:18
	→ Arthur joined	17:18
davidfetter	anybody got resources re: migrating from perforce?	17:18
	→ justin-george joined	17:19
	→ Buglouse joined	17:19
	→ pestilen1e joined	17:19
	← aidan left	17:20
pestilen1e	really dumb question: how can i clone a old version of something in a git repo? there is only one branch, but the "git log" shows many revisions	17:20
	(and there are no tags)	17:21
Bombe	pestilen1e, git checkout <rev>	17:21
	← tvw left	17:24
	← tatsuya left	17:24
	→ g0bl1n joined	17:25
pestilen1e	Bombe: i don't understand. i don't see any way to specify e.g. date in the man page	17:25
	only branch	17:26
	→ solar_sea joined	17:26
	→ nsdk joined	17:26
IslandUsurper	pestilen1e, clone is not checkout. clone grabs all of a repo's history. checkout puts a specific revision in the working tree	17:26
Bombe	pestilen1e, funny. My man page for git-checkout also mentions “tree-ish”.	17:27
	pestilen1e, a “tree-ish” is anything that can be reduced to a single commit. That might be the name of a tag, the name of a branch, or the ID of a commit itself.	17:27
	→ khmarbaise joined	17:27
Bombe	pestilen1e, git log tells you lots of commit IDs. Just “git checkout” one of those. :)	17:27
	← MattDiPasquale left	17:28
pestilen1e	Bombe: ok, that's what i was looking for :)	17:28
	IslandUsurper: are you saying i have all of the revisions if i did a clone?	17:28
	→ khmarbaise_ joined	17:28
	necromancer`brb → necromancer	17:28
IslandUsurper	pestilen1e, yes. that's why checkout works	17:29
pestilen1e	well i'll be.	17:29
IslandUsurper	and log, and diff, and...	17:29
	→ MmikeMRMA joined	17:29
	← galderz left	17:29
MmikeMRMA	Oh, I love the existance of this channel!	17:29
IslandUsurper	welcome to the wonderful world of distributed version control	17:29
MmikeMRMA	How do I revert to commited changes in just one file? I did a commit, and then I changed dozen of files, but now I want to 'undo' the changes to just one file?	17:30
	→ rayners joined	17:30
cbreak	MmikeMRMA: man git-checkout	17:31
Gitbot	MmikeMRMA: the git-checkout manpage can be found at http://git.or.cz/man/git-checkout	17:31
solar_sea	git's fun. I gave up on it after no single tutorial in an hour of reading explained me the difference between the head and the master indeces, but its' fun nevertheless :)	17:31
cbreak	just check out that file	17:31
	solar_sea: that's because there are no indices in git :)	17:31
MmikeMRMA	I love you guys! :)	17:32
	→ sh1mmer joined	17:32
solar_sea	cbreak, egit/jgit said they have, go figure	17:32
	← khmarbaise left	17:32
	khmarbaise_ → khmarbaise	17:32
	→ drako joined	17:33
cbreak	imposters. both of them!	17:33
	→ Phrogz joined	17:33
	← roop left	17:34
	pestilen1e → pestilence	17:34
Phrogz	I have a local repo. I want to create a remote repo (over ssh) that becomes its master (so that "git pull/push" just works), but I want to preserve my existing history. What are the steps?	17:34
	→ _ollie joined	17:35
	← _ollie left	17:35
Phrogz	(I have a bare repository init'd remotely, but of course I can blow that away if needed).	17:35
	← pestilence left	17:35
	→ pestilence joined	17:35
	← alip left	17:35
	← apsysoev left	17:35
SethRobertson	have you set up the remote as a "remote" yet?	17:35
cbreak	there are two ways: first, clone and then just copy the clone over	17:36
	→ alip joined	17:36
cbreak	or the other way, do what SethRobertson implied and add it as remote, then push	17:36
SethRobertson	(clone as a bare mirror)	17:36
cbreak	no	17:36
SethRobertson	(push as a mirror)	17:36
cbreak	not as mirror	17:36
SethRobertson	why not?	17:36
cbreak	mirror only very rarely is what you want	17:36
SethRobertson	But isn't this that case?	17:36
cbreak	no	17:36
	Phrogz waits for the answer to settle out	17:36
SethRobertson	You don't want to keep it a mirror, but for initial population?	17:36
cbreak	no	17:37
SethRobertson	You say no but I don't hear an explanation	17:37
cbreak	you didn't ask for one :)	17:37
	but the reason is basically in the man page	17:37
	mirror pushes _everything_ (almost)	17:37
SethRobertson	I don't think he has any other remotes	17:38
	But perhaps --all would be better	17:38
cbreak	indeed.	17:38
SethRobertson	And --tags	17:38
cbreak	well, more importantly	17:38
Phrogz	So, I git-clone, then copy the clone to the remote location...and then how do I make the remote location the new master/source/whatever?	17:38
cbreak	add a --set-upstream	17:38
	nah, if you have a bare, you don't need to clone	17:39
	just add it as remote	17:39
	then push the stuff you want there	17:39
Phrogz	Forgive me, but your word "just" implies that I have knowledge that I think I do not.	17:39
cbreak	Phrogz: man git-remote	17:39
Gitbot	Phrogz: the git-remote manpage can be found at http://git.or.cz/man/git-remote	17:39
cbreak	use that command to add the bare you created on the server	17:40
	pick a nice name	17:40
	→ MattDiPasquale joined	17:41
Phrogz	cbreak: Neither clone nor remote seem to accept a --set-upstream; where is that to be applied?	17:41
cbreak	push	17:41
	← gilaniali left	17:43
	← rossand left	17:44
	← Arthur left	17:45
	← corni left	17:45
	lamont` → lamont	17:45
	→ Arthur_ joined	17:45
	← rpbertp13 left	17:46
	→ rpbertp1_ joined	17:47
	← rpbertp1_ left	17:47
	→ josefrichter joined	17:48
	→ andyburke joined	17:49
	→ nevyn^ joined	17:49
	← bentob0x left	17:50
	← dvaske left	17:51
	→ dvaske_ joined	17:51
	← zain left	17:52
	→ Athox joined	17:52
	← nevyn^ left	17:53
jim	is it usually the case there would be a repo per project, and maybe more than one repo? are there ways to do more than one project per repo?	17:55
	→ pantsman joined	17:55
	→ khmarbaise_ joined	17:56
	← Anti-X left	17:56
cbreak	maybe	17:56
	usually there''s one repository per project per user	17:56
	so with five users and three projects, you have 15 repositories	17:56
	← t0rc left	17:56
cbreak	possibly one additional per project as "central" one	17:56
	← khmarbaise left	17:57
	khmarbaise_ → khmarbaise	17:57
	→ foocraft joined	17:58
jim	also, would it be the case that per user, there would be a bare repo on a public facility and a nonbare copy of it on the user's home machine (typically that is)	17:58
	← carutsu_ left	17:58
cbreak	depends	17:58
jim	ok, so that arrangement is not necessarily typical?	17:58
	→ zain joined	17:59
	← Athox left	17:59
	← khmarbaise left	18:02
	← CoinOpeBoy left	18:02
	→ khmarbaise joined	18:03
	← koltroll left	18:03
	← zain left	18:03
	→ ph^ joined	18:03
	→ gilesgoatboy joined	18:04
	→ robinbowes joined	18:04
	← joel_falcou left	18:06
gilesgoatboy	is there an easy way I can say, hey git, tell me the difference between this file on branch foo and the very same file on branch bar?	18:07
cbreak	jim: you can do it how ever you want. Afaik, one or more central repository are popular	18:07
	we use that workflow	18:07
	→ joel_falcou joined	18:07
cbreak	gilesgoatboy: yes.	18:07
gilesgoatboy	how do I do it? git df branchname filename?	18:08
cbreak	gilesgoatboy: man git-diff	18:08
Gitbot	gilesgoatboy: the git-diff manpage can be found at http://git.or.cz/man/git-diff	18:08
cbreak	search for git diff [--options] <commit>..<commit> [--] [<path>...]	18:08
gilesgoatboy	hmm, manpages are nice, but I just typed my guess on the command line and it worked :-)	18:09
	→ Thell joined	18:09
	← dvaske_ left	18:09
	→ dvaske joined	18:10
	→ zain joined	18:10
	← gilesgoatboy left	18:10
	← nbrochu left	18:10
	→ whaley joined	18:11
Thell	Quick question: merge resulting in multiple conflicted files, I've fixed all, but realize I want to redo _one_ of the file resolutions before continuing. How do I reset that one file back to the conflicted state so mergetool will pick up that it is conflicted?	18:11
	→ Anti-X joined	18:11
	← ajpiano left	18:11
albel727	Thell: try "git checkout -m -- <file>"	18:13
Thell	'-m' <== that rings a bell!	18:13
albel727	Thell: merge. =) hope it will keep ringing after you read on it in man git-checkout	18:13
Gitbot	Thell: the git-checkout manpage can be found at http://git.or.cz/man/git-checkout	18:14
meder	if i add a remote repository like github on my local git repo, is that basically treating it like a centralized repo?	18:14
	im looking to setup a centralised repo	18:14
	→ jas4711 joined	18:14
Thell	thank you, that was it. I re-read git-merge a few times trying to remember what it was.	18:14
	Thell shrugs	18:14
Thell	Thank god for irc and helpful people ;)	18:15
cbreak	meder: no	18:15
	← dvaske left	18:15
	→ dvaske_ joined	18:15
cbreak	if you add a remote then you just treat it as remote	18:15
	← khmarbaise left	18:15
meder	cbreak: can you point to a guide covering central repo and covering merges /pushes to it from local ones?	18:15
albel727	Thell: I found it myself just two hours ago. So looks like it's all thanks to god =D	18:15
	← gp5st left	18:15
Thell	:D	18:16
cbreak	meder: there's nothing in git that lets you make central repositories	18:16
	it's just a workflow decision	18:16
meder	well	18:16
karamorf	if I want to ignore all files in a directory but actually have the directory in the repo (so a tmp directory basically), what do I need to put in the .gitignore file? '*' seems to include the directory...	18:16
cbreak	if you say a repository is central, then it is	18:16
meder	right.	18:16
cbreak	git won't care	18:16
meder	that's what im looking for, since i know people have used this workflow	18:16
	use this*	18:17
cbreak	it's just like any other remote	18:17
meder	are there any special things to consider when making the central one? in terms of having group permissions to write, etc	18:17
zombor	just use gitolite	18:17
	← dvaske_ left	18:18
	→ dvaske joined	18:18
	→ alnewkirk joined	18:18
cbreak	meder: depends on how it is used	18:18
	via ssh?	18:18
meder	yes	18:18
cbreak	from different user accounts?	18:18
meder	yes	18:18
cbreak	or all the same one with gitolite?	18:18
meder	different user accts	18:18
cbreak	do you need per-branch permissions? or can all have full control?	18:18
meder	the latter	18:18
cbreak	then make sure you create a shared bare repository	18:19
	→ blarson joined	18:19
cbreak	meder: man git-config, look for sharedRepository	18:19
Gitbot	meder: the git-config manpage can be found at http://git.or.cz/man/git-config	18:19
meder	i see	18:19
cbreak	meder: man git-init --shared	18:19
Gitbot	meder: the git-init manpage can be found at http://git.or.cz/man/git-init	18:19
	← Buglouse left	18:20
	→ bhosie joined	18:20
	← warbee left	18:20
meder	cbreak: bare with me since ive only used cvs ( UGH ).. if user A commits to the central repo and user B has subsequent commits, the user wont be asked to update his/her local repository, right? or does it matter the command used? eg push vs merge?	18:21
	user B for the latter	18:21
cbreak	don't	18:21
meder	don't what?	18:21
cbreak	make a --bare repository	18:21
	no one should commit to the central repository	18:21
	it's not meant to be used for commiting	18:22
	← bryanl left	18:22
zombor	http://toroid.org/ams/git-central-repo-howto	18:22
	read that	18:22
	→ koltroll joined	18:23
	→ jstemmer joined	18:24
	→ dunpeal joined	18:24
dunpeal	Hey. I have a hefty post-receive script, and I don't want my users to have to wait for it to complete. So I created it as hooks/post-receive.heavy, and put this in hooks/post-receive:	18:24
	hooks/post-receive &	18:25
	err:	18:25
	hooks/post-receive.heavy &	18:25
	→ ghost_olli joined	18:25
dunpeal	now, that should put the process in the background, so the user who pushes shouldn't have to wait for it. except it doesn't.	18:25
	How do I do it?	18:25
	→ industrial joined	18:26
rson	karamorf: i had the same problem, unfortunately my solution was to specify the type of file to ignore because in my case i could do that. hopefully you'll find a better solution.	18:26
	→ dvaske_ joined	18:26
	← dvaske left	18:26
	→ ch077179 joined	18:26
alnewkirk	/j #perl	18:26
industrial	1) If I make a git submodule it will always take the latest version to clone and use and not the version I added when I committed the .gitmodule? 2) can I have submodules under submodules? eg create a .vim repo with plugin submodules but it being a submodule to eg a dotfile repo. How does syncing everything work then? :P	18:26
	← bhosie left	18:27
cbreak	1) no	18:27
	2) you can put recursive submodules in each other until you die of old age	18:27
	→ bhosie joined	18:27
	→ patmaddox joined	18:28
industrial	so if the submodule updates then mine doesnt, and I have to commit changed happened in the submodule to my own repo to use them?	18:28
	changes*	18:28
cbreak	the parent knows exactly which commit of the submodule it WANTS	18:28
	and parents always get what they want	18:28
	← sraue left	18:29
cbreak	so you have to recommit the parent with an updated child, when you want to update the child	18:29
	obviously, that's usually exactly what you want anyway	18:29
industrial	ah, right, thats what I was thinking.	18:29
	time to break up my configs \o/	18:29
karamorf	rson: did the same thing, luckily only .html files were being generated in this dir.	18:29
	← avinashhm left	18:30
	→ avinashhm joined	18:30
	→ jfkw joined	18:32
	← alberto56 left	18:32
	← dvaske_ left	18:32
	→ dvaske joined	18:32
	→ Laibsch joined	18:33
Laibsch	Hi, is it possible to set denyNonFastforwards to for all branches but one?	18:34
	Hi, is it possible to set denyNonFastforwards to true for all branches but one?	18:34
cbreak	maybe with gitolite	18:34
	← industrial left	18:35
dunpeal	any clue about my question?	18:35
	← koltroll left	18:35
dunpeal	I tried detaching with nohup, git still waits for it to finish.	18:35
	Laibsch: you can only that with a hook, either one you wrote or someone else's (gitolite?)	18:36
Laibsch	I see	18:36
	→ sraue joined	18:37
Laibsch	thanks. I am not admin of the server. Merely a user	18:37
	→ SEJeff_work joined	18:37
	→ giallu joined	18:37
SEJeff_work	So I've got 2 repositories I'm trying to combine into 1 locally and preserve the history	18:37
	I tried adding the second one as a remote using a file:// uri. How do I merge the second remote into the current repo?	18:38
_ikke_	Just git pull remotename branchname	18:39
albel727	dunpeal: try putting it in (), like that "(nohup hooks/post-receive.heavy &)"	18:40
	→ khmarbaise joined	18:41
	← kipras left	18:42
	→ artagnon joined	18:42
	← msimkins left	18:42
dunpeal	albel727: doesn't help :(	18:43
	→ msimkins joined	18:43
artagnon	Given the path of a file, how do I get the corresponding blob's SHA1 restricting the search to just one given commit?	18:44
	← q0rban left	18:44
albel727	dunpeal: hmm. maybe call "true" on a line after "()"?	18:45
	→ akahn joined	18:45
	← spearce left	18:45
	→ robinr joined	18:45
dunpeal	albel727: I had to redirect both stdout and stderr.	18:45
	that was the trick.	18:45
wereHamster	dunpeal: I once used a message queue for such heavy post-receive work	18:45
dunpeal	and & of course.	18:45
	wereHamster: looks like I'm good with "1>/dev/null 2>&1 &".	18:46
albel727	dunpeal: hmm, tty detaching. I see.	18:46
dunpeal	job just runs in the background, the push returns immediately.	18:46
	→ q0rban joined	18:47
	← Laibsch left	18:48
	→ timvisher joined	18:48
	→ spearce joined	18:49
jim	does the post receive you're doing affect either your working dir or the local repo?	18:49
dunpeal	jim: nope, it basically just sends an email.	18:49
	→ sgh joined	18:49
meder	ive read http://toroid.org/ams/git-central-repo-howto but im still confused on when i git push to the central repo. i commit locally and push. do i rebase or merge?	18:51
	← bhosie left	18:51
meder	that is, conflicts between multiple users	18:51
	user A pushes to central repo	18:51
	user B pushes, gets Merge the remote changes before pushing again. See the 'non-fast forward'	18:51
	→ bhosie joined	18:52
	← mfacenet left	18:53
jim	well, my take would be, you'd commit locally as many times as you want while you're working on some specific issue, or are testing it	18:53
	once you KNOW it works, push.	18:54
	← blarson left	18:54
meder	oh	18:54
	git pull, then push back	18:54
	well, for small cases this works	18:54
	← ghost_olli left	18:54
jim	you would branch to be working on a specific issue locally, then merge the branch when you know it works, then push	18:55
	(that would be another way)	18:55
	← rayners left	18:55
	→ erez joined	18:57
	← bhosie left	18:58
	→ blarson joined	18:59
	→ bhosie joined	18:59
albel727	artagnon: git ls-tree <treeish> <files>?	19:00
	← Pupeno left	19:01
	← przemoc left	19:02
	← yngress left	19:02
	→ przemoc joined	19:04
	← rpg left	19:04
	→ Pupeno joined	19:04
	→ rossand joined	19:04
	← aspotashev left	19:05
	→ itrebal joined	19:06
	← jdav left	19:07
	→ ghost_olli joined	19:10
	→ aspotashev joined	19:11
	← ghost_olli left	19:11
	← jas4711 left	19:12
artagnon	albel727: Right, thanks :)	19:12
	← envi left	19:12
	→ jdav joined	19:13
	← Anti-X left	19:13
albel727	artagnon: yw =) keep up good work on svn fast-import.	19:14
	← solar_sea left	19:16
	← dvaske left	19:17
	→ dvaske_ joined	19:17
	← mbroeker left	19:17
	← Thell left	19:17
artagnon	albel727: Thanks :)	19:18
	→ WebDragon joined	19:18
	→ MetaCosm joined	19:18
	→ msimkins_ joined	19:18
	→ __name__ joined	19:19
	← __name__ left	19:19
	→ __name__ joined	19:19
	→ Anti-X joined	19:19
	← Noldorin left	19:22
	← msimkins left	19:22
karamorf	if I use --intent-to-add (git add flag), does that mean its just adding the path to the repository and not the contents?	19:22
	→ joeyh joined	19:23
wereHamster	yes	19:23
joeyh	hmm, mkdir foo; git mv * foo fails in what seems a needlessly annoying way. Is there some easy workaround?	19:23
	← bhosie left	19:23
	← thierryp left	19:24
	→ DaveInMoody joined	19:24
joeyh	ok, -k .. wasn't expecting "DTRT" to just be an option..	19:24
Phrogz	git push --set-upstream gitbox #=> error: unknown option `set-upstream'	19:25
	Is that new in 1.7 or something?	19:25
	← dvaske_ left	19:25
	necromancer → necromancer`brb	19:25
	→ dvaske joined	19:26
	← Yuffster left	19:26
	→ hobodave joined	19:27
	← dunpeal left	19:27
	Phrogz goes to install 1.7 to find out	19:28
	jdav → jdav_gone	19:28
	→ Phoebus joined	19:28
karamorf	rson: figured it out. put '*' in the .gitignore file for the directory, then `git add -f dir/.gitignore`. dir gets added and all of the files get ignored.	19:29
joeyh	though -k really seems a slightly too big hammer when all i want is for git mv to skip bad parameters	19:29
rson	karamorf: i suppose the -f is what does the trick, thanks.	19:30
	→ phantomcircuit joined	19:30
	← kenneth_reitz left	19:31
	← VadtecWork left	19:31
	→ VadtecWork joined	19:31
	→ craq joined	19:32
	→ [algo] joined	19:32
[algo]	hi	19:32
craq	arg. i pulled latest for a project and it said i need to merge, i go to mergetool and there's no conflicts.. i try and save and it says cannot merge?	19:33
karamorf	rson: or better yet, in the root .gitignore you could put '!.gitignore'	19:33
[algo]	I have git working copy. I want a directory from another git repo in it. How can I ... ?	19:33
	→ kipras joined	19:33
[algo]	in SVN that'd be svn:export	19:33
craq	at this point, i dont care about the changes i made to this file but i already committed.. how can i just force it to use the remote copy easily?	19:33
wereHamster	[algo]: man git-export	19:33
[algo]	man git-export	19:33
	No manual entry for git-export	19:33
wereHamster	[algo]: sorry, man git-archive	19:33
Gitbot	[algo]: the git-archive manpage can be found at http://git.or.cz/man/git-archive	19:33
[algo]	ghm	19:33
	→ urkud joined	19:34
[algo]	thank you	19:34
wereHamster	you got me confused there with the svn export..	19:34
[algo]	ummm ?	19:34
	I mean include another repo into my copy	19:34
	maybe submodule..	19:34
	I need to use external library in my sources	19:34
	sorry	19:35
	svn:external	19:35
	svn propset svn:externals...	19:35
	→ bcardarella joined	19:36
	← Phrogz left	19:37
wereHamster	[algo]: ah, that's something different	19:37
	you canto use 'svn:external' to point to a git repo	19:37
	you'd have to convert the git repo to svn	19:37
Nightwolf	i tried to install egit on the ubuntu eclipse as described here: http://www.vogella.de/articles/EGit/article.html it is listed as installed but it doesn't show up. i'm currently downloading another eclipse version, but maybe someone here knows the issue?	19:37
	← jobermayr left	19:37
	→ deadnoralive joined	19:37
deadnoralive	Hi, is there a channel for git-python?	19:38
	→ bhosie joined	19:38
	← erez left	19:39
	← __name__ left	19:42
	← Zephyrus left	19:42
	← dvaske left	19:43
	← joeyh left	19:43
	→ dvaske joined	19:43
	← ph^ left	19:44
	← craq left	19:44
	→ theoros joined	19:44
	→ TSK joined	19:44
	→ ph^ joined	19:45
	← dwmw2 left	19:46
	← [algo] left	19:46
	→ ajpiano joined	19:47
	→ bdiu joined	19:48
	← dvaske left	19:48
	→ dvaske joined	19:48
	→ DrAk0 joined	19:51
	← DrAk0 left	19:51
	→ DrAk0 joined	19:51
	← Elfe_ left	19:51
	← DrAk0 left	19:52
	← metaperl left	19:52
	→ engla joined	19:52
	← frakturfreak left	19:53
	→ dwmw2 joined	19:53
	← dvaske left	19:54
	→ dvaske_ joined	19:54
	→ guest3456 joined	19:54
	→ Elfe joined	19:55
	→ _ollie joined	19:55
	← drako left	19:55
	← _ollie left	19:55
	← dvaske_ left	19:56
	→ dvaske joined	19:56
guest3456	having problems with git add	19:57
	heres the paste http://gist.github.com/652205	19:57
wereHamster	is SiteIniCreator a git repo?	19:57
	is that folder ignored?	19:57
guest3456	not ignored	19:58
	→ kumbayo joined	19:58
	← nullvariable left	19:58
	→ nullvariable1 joined	19:58
	nullvariable1 → nullvariable	19:58
guest3456	is git case sensitive?	19:58
wereHamster	yes	19:58
guest3456	maybe 10 command prior i did this: http://gist.github.com/652209	19:58
	with lowercase s in Sites	19:59
	let me paste my whole workflop	19:59
	← chronos left	20:00
albel727	guest3456: do you have any files in SiteIniCreator?	20:00
guest3456	http://gist.github.com/652211	20:01
	yes i do	20:01
	← dvaske left	20:01
	← aspotashev left	20:02
	→ dvaske joined	20:02
	→ warbee joined	20:02
	← chrisshattuck left	20:03
	← cilly left	20:03
	← spearce left	20:04
	← dvaske left	20:04
	← MikhX left	20:04
	→ dvaske joined	20:04
	→ MikhX joined	20:05
	→ pedrobelo joined	20:06
	← primalmotion left	20:08
	← timvisher left	20:09
	← MikhX left	20:10
	← dvaske left	20:10
	→ MikhX joined	20:10
	→ dvaske joined	20:10
	← SEJeff_work left	20:11
	→ drubies_ joined	20:11
	← pedrobelo left	20:11
	← dwmw2 left	20:11
	→ DarkUranium joined	20:12
DarkUranium	hey	20:12
	will adding "doc" to gitignore only affect ./doc or is it recursive?	20:12
	I mean, will it also affect ./foo/bar/doc?	20:12
	← dvaske left	20:12
	→ dvaske_ joined	20:12
wereHamster	yes	20:13
cbreak	DarkUranium: it will, indeed	20:13
	you can call it /doc though	20:13
	→ mw joined	20:13
	→ dr_win joined	20:13
	→ pedrobelo joined	20:13
	← MetaCosm left	20:13
DarkUranium	cbreak: / means root of the git repo, I take it?	20:13
wereHamster	no	20:13
frooh	why do people always include the .git in git urls when, apparently, it's never actually required?	20:13
wereHamster	the dir where the gitignore is	20:13
DarkUranium	ah	20:14
	root of the repo in this case then =P	20:14
cbreak	frooh: because it's part of the folder name	20:14
DarkUranium	hmm	20:14
	are wildmarks supported?	20:14
	*~	20:14
frooh	cbreak: but every server out there knows to ignore it?	20:14
wereHamster	DarkUranium: man gitignore	20:14
Gitbot	DarkUranium: the gitignore manpage can be found at http://git.or.cz/man/gitignore	20:14
cbreak	frooh: no	20:14
	why ignore it?	20:14
	it's part of the folder name	20:14
DarkUranium	cbreak: I meant for stuff like foo.c~	20:14
	cbreak: the backups that editors make	20:15
frooh	cbreak: my point is, with gitosis, gitolite, and githib, if the repo is foo.git, I can do clone server:foo	20:15
	and it works great	20:15
	so why even mention the .git part?	20:15
cbreak	you can't	20:15
wereHamster	DarkUranium: he wasn't talking to you..	20:15
DarkUranium	yeah, I noticed that =P	20:15
frooh	I...can?	20:15
cbreak	if a repo is foo.git, then it's name if foo.git	20:15
frooh	interesting.	20:15
	well, works for me. I was just curious anyway	20:15
cbreak	there could potentially be foo, foo.gi and foo.git2 repositories too	20:15
	→ witquicked joined	20:16
albel727	DarkUranium: look .git/info/exclude	20:16
cbreak	if you would ignore stuff, which would you use?	20:16
	← dvaske_ left	20:16
	→ dvaske joined	20:16
frooh	ok, so it is probably special casing in the code then	20:16
	because it does work	20:16
	← khmarbaise left	20:16
	← Phoebus left	20:16
	→ fr0sty_ joined	20:16
wereHamster	what does work?	20:16
doener	cbreak: when you ask git to clone /bla/foo it will look for the repo in foo.git and foo/.git	20:16
albel727	DarkUranium: there has to be an example of default ignores for C project. *~ is a part of it.	20:17
	← pedrobelo left	20:17
cbreak	doener: and it won't look in /bla/foo?	20:17
	that's stupid	20:17
doener	cbreak: well, and foo itself	20:17
	→ brntbeer joined	20:17
frooh	wereHamster: what doener said, git clone server:foo (Which does the same as server:foo.git)	20:17
	→ pedrobelo joined	20:17
frooh	I just like leaving the .git off and wondered why other people don't do the same :-)	20:17
cbreak	where does it look when I ask for blah/foo.git.git?	20:17
frooh	no clue	20:18
cbreak	using wrong names seems dumb to me	20:18
frooh	I wouldn't name a repo that :-)	20:18
	exactly	20:18
	← pedrobelo left	20:18
cbreak	no idea why people would do that	20:18
DarkUranium	thanks wereHamster, albel727	20:18
cbreak	so why clone foo if the folder is foo.git? dumb...	20:18
	← jdav_gone left	20:18
frooh	well, you say dumb, I say brief	20:18
	frooh won't ask next time :-)	20:19
	→ pedrobelo joined	20:19
	albel727 wholeheartedly agrees with cbreak. minus 3 letters is not brief, it's dumb.	20:19
cbreak	if you want foo, call the repo foo	20:19
frooh	albel727: four	20:19
cbreak	not foo.git	20:19
	. is not a letter :)	20:19
frooh	touche	20:19
	-4 keystrokes	20:19
doener	cbreak: http://git.pastebin.com/KnPQCiqg	20:19
albel727	<sarcasm> oh, that completely changed everything </sarcasm> =)	20:19
IslandUsurper	mmm, pedantry	20:19
cbreak	you're in a channel frequented by programmers	20:20
	of course we're pedantic :)	20:20
IslandUsurper	and it's delicious	20:20
frooh	the point is that if I can make the url simpler, I can remember it better	20:20
cbreak	why remember it at all?	20:20
frooh	git@foo:project.git vs foo:project	20:20
cbreak	and if you forget the .git in a git repo	20:20
albel727	is it hard to remember to append .git?	20:20
doener	One upside I see is that you can ignore bare vs. non-bare	20:20
cbreak	... well	20:20
	you can ignore that anyway	20:21
	you can call non bare repos foogit.git or bare ones just foogit	20:21
	← DaveInMoody left	20:21
	← engla left	20:21
NfNitLoop	or you can just foogit about the whole thing.	20:21
frooh	NfNitLoop: nice	20:22
albel727	=)	20:22
NfNitLoop	(Ok. Yeah, that was bad.) :p	20:22
doener	well, I like the directories to have the "sane" names, but I usually don't care whether I clone a bare or a non-bare repo	20:22
cbreak	not like you have to	20:22
NfNitLoop	I think people use x.git to show that, yes, this URL is to a git repository. As opposed to just a project x.	20:23
frooh	right	20:23
albel727	oh my god. this strace shows completely horrible thing.	20:23
frooh	that makes sense	20:23
cbreak	but leaving away file endings is like ignoring the exe in a nekid-young-girl.jpg.exe	20:23
	it's wrong	20:23
	and it might bite you in the ass	20:23
frooh	well, if you use win32	20:23
cbreak	maybe	20:23
frooh	:D	20:24
albel727	git clone /foo tries /foo.git before /foo	20:24
	this is insane	20:24
NfNitLoop	albel727: Oh, it does? That's ... unexpected.	20:24
cbreak	... let's make a revolution	20:24
	→ jdav_gone joined	20:24
tgunr_	pretty cool vis of GIT http://bit.ly/a9V90U	20:24
albel727	NfNitLoop: doener provided this http://git.pastebin.com/KnPQCiqg	20:24
cbreak	but that behaviour sounds fun	20:24
frooh	this is the last time I ask a question in here out of curiosity	20:24
cbreak	you can set up fake repositories	20:24
	the real is foo, fake is foo.git	20:25
	→ frakturfreak joined	20:25
cbreak	and then, good luck cloning from foo :)	20:25
	tgunr_: I don't like lybian pages	20:25
albel727	cbreak: yeah, that's it =)	20:25
	→ sholden joined	20:25
	← dvaske left	20:26
	albel727 thinks, it should be reported as a bug/security vulnerability	20:26
	→ dvaske joined	20:26
sholden	Hi everyone. I'm trying to set up git-http-backend on dreamhost, but I'm running into a bit of trouble. I'm using basic auth, and it seems to be working. After that, it tries to get /hellojava/git/info/refs and hits 404	20:26
	tgunr_ shurgs, don't look?	20:27
sholden	my server log is showing [Thu Oct 28 13:18:10 2010] [error] [client 173.236.172.31] Request not supported: '/home/sshconnectiongit/git.sshconnection.com/'	20:27
cbreak	albel727: what's the semantics of that?	20:27
albel727	cbreak: sorry?	20:27
sholden	GIT_CURL_VERBOSE is showing a 404	20:27
cbreak	albel727: what is the meaning of a bunch of balls and the lines in between them?	20:27
	trying to imagine how my repos look like	20:27
wereHamster	sholden: does that folder exist?	20:28
drizzd_	sholden: so your repo is called "git"?	20:28
	← tama left	20:28
albel727	cbreak: you mean tgunr_'s youtube link?	20:28
	← whaley left	20:28
cbreak	albel727: no, the one you posted	20:29
sholden	I created a directory ~/git/ehllojava.git and in there i did git --bare init-db	20:29
cbreak	or was that someone else?	20:29
sholden	am i doinitwrong? sorry, im a git newb, just trying to get up and running and learn my way around	20:29
cbreak	I mean the movie anyway :)	20:29
wereHamster	sholden: and how did you try to clone it?	20:29
	→ WebDragon\|laptop joined	20:29
sholden	GIT_CURL_VERBOSE=1 git clone http://git.sshconnection.com/hellojava.git	20:29
	→ [algo] joined	20:29
[algo]	does that look correct ?	20:30
albel727	cbreak: I only reposted doener's strace. ly-link was tgunr_'s. I'm trying to figure out the meaning of the video myself, anyway =)	20:30
[algo]	git submodule add http://github.com/marijnh/CodeMirror.git www/modules/test/codemirror	20:30
wereHamster	sholden: are you using smart http?	20:30
	→ ignatp joined	20:30
sholden	i can post a gist of my git-http-backend.cgi and .htaccess if that helps...	20:30
	wereHamster: trying to...	20:30
cbreak	but then who posted the youtube link?	20:30
	[algo] → wereGrandma	20:30
	← zombor left	20:30
wereHamster	sholden: sorry, I don't know anything about smart http	20:31
	→ zombor joined	20:31
	wereGrandma walks around	20:31
cbreak	can't find it in the channel log anymore :/	20:31
albel727	cbreak: <tgunr_> pretty cool vis of GIT http://bit.ly/a9V90U	20:31
sholden	here's my configs: http://gist.github.com/652266	20:32
	well, my cgi and .htaccess	20:32
cbreak	oh, that's youtube?	20:32
	thanks albel727	20:32
	← WebDragon left	20:32
	WebDragon\|laptop → WebDragon	20:32
	← xiong left	20:33
	→ Eridius joined	20:33
albel727	cbreak: oh, I got it. this is just a graph of working tree. as per this link http://code.google.com/p/gource/wiki/Screenshots	20:34
cbreak	oh!	20:34
	now it makes sense	20:34
	← ph^ left	20:34
cbreak	git's graph looks boring. maybe they haven't heard of folders :/	20:35
	← joel_falcou left	20:35
	→ ari-_-e joined	20:35
guest3456	sorry was afk	20:35
	any idea on my problem?	20:35
	← doener left	20:35
	→ ty_ joined	20:35
guest3456	here it is again: http://gist.github.com/652205	20:35
	→ thiago joined	20:36
albel727	guest3456: no idea. I was unable to reproduce it in my testing repo. I don't see, how it could happen at all.	20:36
	← bmalee left	20:36
guest3456	im thinking cause of case sensitivity	20:36
	→ bmalee joined	20:36
guest3456	cause earlier i added "sites/" not "Sites/"	20:36
	here was the full work flow: http://gist.github.com/652211	20:36
NfNitLoop	guest3456: ah, yes, likely.	20:37
	Try renaming Sites/ to sites/	20:37
	that may fix the problem.	20:37
guest3456	hrmm ok	20:37
	ultimately i want uppercase	20:37
wereHamster	guest3456: try avoiding case insensitive filesystem	20:37
guest3456	:)	20:38
	← giallu left	20:38
albel727	guest3456: do you have case insensitive FS? which one is it? FAT? =)	20:38
guest3456	im on winxp NTFS	20:38
	→ Phrogz joined	20:39
guest3456	renaming the directory to "sites" in windows explorer makes "git status -s" show nothing	20:39
	thumbsup	20:39
ty_	howdy all. I'm new to git and having a heck of a time with a post-update script :(	20:39
guest3456	now how can i have it be "Sites/"	20:39
wereHamster	guest3456: git rm --cached -r sites; mv sites Sites; git add Sites	20:39
	← rvsjoen left	20:40
Phrogz	Based on previous advice, I cloned a repo, moved it to a remote location, added that loc as a remote and used set-upstream to the new remote name. But now when I "git push" I get "refusing to update checked out branch...non-bare repository..."	20:40
	I'm actually using the remote as a bare repository; how can I tell Git that it is?	20:40
	← TSK left	20:40
	← nsdk left	20:40
guest3456	excellent thanks	20:41
NfNitLoop	Phrogz: is it? did you do clone --bare?	20:41
	→ rvsjoen joined	20:41
Phrogz	NfNitLoop: Nope; either I wasn't told to--or missed--that option.	20:41
guest3456	well i think htats it for now :) thanks guys	20:41
	→ dharrigan joined	20:41
	← dharrigan left	20:41
	→ dharrigan joined	20:41
	→ giallu joined	20:42
	← giallu left	20:42
	→ giallu joined	20:42
Phrogz	NfNitLoop: Is there a way to hack this in after the fact, or must I clone and copy again?	20:42
NfNitLoop	Phrogz: ah. Try a clone --bare. :)	20:42
	→ Devastator joined	20:42
Phrogz	(It takes a long time to push the full repo to the remote.)	20:42
NfNitLoop	You can, on the same file system, do a clone --bare.	20:42
	It should be very fast.	20:42
Phrogz	Ah, nice. :)	20:42
NfNitLoop	(It'll do hardlinking if it can.)	20:42
Phrogz	(duh)	20:42
NfNitLoop	you may even want to clone --mirror.	20:42
	which implies bare, and puts all of your branches into the refs/heads	20:43
	so that subsequent clones will get them.	20:43
	Which I just discovered today while trying to set up a mirror. It's handy!	20:43
	→ testX joined	20:44
	← zombor left	20:44
Phrogz	That worked great, thanks	20:44
	← jonesy left	20:45
	← bhosie left	20:46
	← IslandUsurper left	20:47
	← koo5 left	20:50
	← dvaske left	20:50
	→ dvaske joined	20:50
	← dharrigan left	20:51
	→ Mage joined	20:52
ty_	I'm trying to use a post-update hook so that when I push changes to a bare repo on a server it will automatically update the development working copy of that repo. The post-update script I've created works fine if I execute it via SSH on the server, but gives me a 'fatal: Not a git repository: '.'' error when pushing to the bare repo and I don't understand why. http://gist.github.com/652296	20:52
	← ymasory left	20:54
wereHamster	ty_: google 'git website howto'	20:54
	→ tama joined	20:55
	→ msimkins joined	20:55
	← dvaske left	20:55
	→ dvaske joined	20:56
	→ eletuchy joined	20:56
ty_	I was trying to follow along with http://joemaller.com/990/a-web-focused-git-workflow/ more or less	20:56
	← xxoxx left	20:56
NfNitLoop	ty_: maybe the current working directory isn't set when running it remotely?	20:57
ty_	the script cd's to the proper directory, if I add a 'pwd' to the script it prints the correct location	20:57
	← jaql left	20:57
	← msimkins_ left	20:57
	→ ericdc joined	20:58
	← eletuchy left	20:58
	→ s0ber_ joined	20:58
	← drubies_ left	20:59
NfNitLoop	ty_: does working/copy/.git exist?	21:00
ty_	yes, and the script works if I execute it on the server directly	21:00
NfNitLoop	oh, as you said. Sorry, I think I'm low on caffeine. :)	21:00
ty_	np	21:00
	→ jrmuizel joined	21:01
	← s0ber left	21:01
	← stalker left	21:01
NfNitLoop	I haven't actually written triggers, so I'm not sure why that's not working for you.	21:01
SethRobertson	gitosis back from the dead? What does this mean for gitolite? Improvements all around we can hope.	21:01
	← dvaske left	21:01
	→ dvaske_ joined	21:01
	→ battleman2004_ joined	21:01
	← jcdny left	21:01
	← jherdman left	21:02
	→ dharrigan joined	21:02
	← dharrigan left	21:02
	→ dharrigan joined	21:02
	→ yairgo joined	21:02
	← dvaske_ left	21:03
	→ dvaske joined	21:03
wereHamster	SethRobertson: why would you think so?	21:03
davidfetter	mmm...zombie gitosis	21:03
	</homer>	21:04
SethRobertson	The latest email on the git mailing list	21:04
wereHamster	SethRobertson: I expect the first reply will tell him to switch to gitolite ;)	21:04
SethRobertson	Oh, I didn't read it fully. He was describing the problems, not that he was going to fix them	21:04
yairgo	If i clone a repository and someone deletes a branch on the remote, after I fetch and do git branch -a it still shows the remote branch. Is there a way to sync the remote so that this branch doesn't show up locally anymore?	21:05
wereHamster	yairgo: git remote prune	21:05
	or manually delete the branch	21:05
	← necromancer`brb left	21:06
	→ artefon joined	21:08
	← cannonball left	21:09
	← Guest31708 left	21:09
	← kumbayo left	21:09
Ilari	å	21:10
	→ tvw joined	21:10
Ilari	oops.	21:10
	→ jcdny joined	21:11
	→ aspotashev joined	21:12
davidfetter	thåt so?	21:12
ty_	well I added an unset GIT_DIR to the script and it seems to work now. I don't know what that does or why I need it... but ok	21:12
	jams__ → jams	21:13
albel727	ty_: why just not use already existing scripts?	21:14
ty_	didn't find one that did what I wanted	21:14
albel727	ty_: faq non-bare	21:15
Gitbot	ty_: Pushing to non-bare repositories is discouraged. See https://git.wiki.kernel.org/index.php/GitFaq#non-bare	21:15
ty_	I'm not	21:15
	← rossand left	21:15
ty_	I'm pushing to a bare repo that executes a script that changes working directories and pulls those changes into a non-bare repo	21:15
yairgo	wereHamster: thanks	21:16
	→ nicktoye joined	21:16
albel727	ty_: that's just for the sake of the script provided under the link. quote: "The latest draft post-update hook for this is at http://utsl.gen.nz/git/post-update, which deals with almost all cases, apart from where there is already a conflicted merge on the remote side"	21:16
	← kenichi left	21:16
	← jams left	21:16
	← alnewkirk left	21:17
	→ SvenDowideit_ joined	21:17
	← q0rban left	21:17
ty_	I've never seen those scripts. probably would've helped if I had :p	21:18
albel727	ty_: oh my. now I've read your use-case.	21:18
	→ dreiss joined	21:18
albel727	ty_: so you have a bare repo, which modifies another non-bare repo on hook? =)	21:18
ty_	yeah exactly	21:19
	I'm pushing local changes to the bare repo and I want those changes to be reflected in the non-bare copy that runs on our development server.	21:19
	← dr_lepper left	21:19
ty_	which appears to be working now, though we'll see what happens when it's not just me pushing changes I guess ;)	21:19
	← balgarath_ left	21:20
	→ balgarath joined	21:20
albel727	ty_: why couldn't it have been a single non-bare repo on development server? =) anyway, I guess it would be better, if in hook of bare repo you do a simple push to nonbare repo, which will have the hook I provided.	21:21
	← karamorf left	21:21
	→ brosner joined	21:21
ty_	because pushing to the non-bare repo was frowned upon everywhere I looked. in particular it wouldn't let me push to the development branch because it was already checked out.	21:21
	→ drizzd joined	21:22
albel727	ty_: oh, that's avoidable. some git config option, IIRC	21:22
	→ mscotta joined	21:22
	← jstemmer left	21:22
ty_	ah. well my first look at git was monday when I started this new job and was expected to have a dev environment up asap... so my knowledge has a few eh... holes ;)	21:22
	← guest3456 left	21:22
albel727	=)	21:22
	→ Slix` joined	21:23
mscotta	hi! I've a problem compiling git-1.7.3.2	21:23
esc	ty_, albel727 : receive.denyCurrentBranch	21:23
albel727	esc: yeah, just found it in man =)	21:23
	← artefon left	21:23
	← dharrigan left	21:24
	→ primalmotion joined	21:24
	← robinr left	21:24
	← shadowmaster left	21:24
albel727	ty_: so, "git config receive.denyCurrentBranch false" would have helped you =)	21:24
ty_	there were a few complexities as well that made pushing to the non-bare repo even trickier. there are changes to certain config files that have to persist on the development working copy. plus some other developers will be making changes directly to that working copy because they aren't willing to work on a command line or in a local environment while they mess with stuff	21:24
	I remember seeing that denyCurrentBranch option, but it looked to me like it would nuke local changes	21:24
	← ignatp left	21:24
mscotta	I got the message "git-compat-util.h:152: fatal error: openssl/ssl.h: Nessun file o directory"	21:25
	← patmaddox left	21:25
albel727	ty_: thomasrast.ch/git/local-config.html	21:25
	→ MOAR-f00li5h joined	21:25
	← drizzd_ left	21:25
	→ patmaddox joined	21:25
albel727	ty_: http://thomasrast.ch/git/local-config.html	21:25
	→ WebDragon\|laptop joined	21:26
	→ shadowmaster joined	21:26
ty_	nice article, that'll almost certainly help down the road	21:26
	← ari-_-e left	21:26
	← fr0sty_ left	21:26
	← Pupeno left	21:26
	← blarson left	21:26
	← icwiener left	21:26
	← sako left	21:26
	← marc_kdab left	21:26
	← cballou left	21:26
	← OOPMan left	21:26
	← cehteh left	21:26
	← jim left	21:26
	← phaedrix left	21:26
	← slonopotamus left	21:26
	← timj_ left	21:26
	← Pacman_Home` left	21:26
	← zenom_ left	21:26
	← AAA_awright left	21:26
	← crash\ left	21:26
	← SvenDowideit left	21:26
	← tgunr_ left	21:26
	← mlq left	21:26
	← IRSeekBot left	21:26
	← f00li5h left	21:26
	← jkp_ left	21:26
	← Fullma left	21:26
	← halfline left	21:26
	← tchan left	21:26
	← joeytwiddle left	21:26
	← nchase left	21:26
	← thear left	21:26
	← d1b left	21:26
	← Iff left	21:27
	← cods left	21:27
	← remi left	21:27
	← cwillu_at_work left	21:27
	← Utoxin left	21:27
	← d1zzy left	21:27
	← abuzor left	21:27
	← rduplain left	21:27
	← davr left	21:27
	SvenDowideit_ → SvenDowideit	21:27
albel727	eh. what was that blazingly fast scroll?	21:27
	→ ari-_-e joined	21:27
	→ fr0sty_ joined	21:27
	→ Pupeno joined	21:27
	→ blarson joined	21:27
	→ icwiener joined	21:27
	→ sako joined	21:27
	→ jim joined	21:27
	→ marc_kdab joined	21:27
	→ cballou joined	21:27
	→ OOPMan joined	21:27
	→ cehteh joined	21:27
	→ phaedrix joined	21:27
	→ slonopotamus joined	21:27
	→ timj_ joined	21:27
	→ Pacman_Home` joined	21:27
	→ zenom_ joined	21:27
	→ Utoxin joined	21:27
	→ AAA_awright joined	21:27
	→ crash\ joined	21:27
	→ tgunr_ joined	21:27
	→ mlq joined	21:27
	→ IRSeekBot joined	21:27
	→ jkp_ joined	21:27
	→ Fullma joined	21:27
	→ halfline joined	21:27
	→ tchan joined	21:27
	→ joeytwiddle joined	21:27
	→ nchase joined	21:27
	→ abuzor joined	21:27
	→ thear joined	21:27
	→ d1b joined	21:27
	→ Iff joined	21:27
	→ cods joined	21:27
	→ remi joined	21:27
	→ cwillu_at_work joined	21:27
	→ d1zzy joined	21:27
	→ rduplain joined	21:27
	→ davr joined	21:27
	→ js_ joined	21:27
	→ nick_h joined	21:27
	→ docgnome joined	21:27
	← WebDragon\|laptop left	21:27
albel727	and again	21:27
	← joeytwiddle left	21:27
ty_	mmm join/leave spam	21:27
VadtecWork	its called a netsplit...ride the wave!	21:28
albel727	cool	21:28
mscotta	hi! I've a problem compiling git-1.7.3.2	21:28
	→ ignatp joined	21:28
ty_	anyway time to head outta here for the day. thanks for the valuable info albel727, esc, and others.	21:28
albel727	yw	21:29
	bue	21:29
	s/bue/bye	21:29
	← WebDragon left	21:29
esc	ty_: np	21:29
	→ joeytwiddle joined	21:29
	← ty_ left	21:29
	→ dharrigan joined	21:29
	→ LeLutin joined	21:29
	← patmaddox left	21:30
	← urkud left	21:31
	← LeLutin_ left	21:31
	← thiago left	21:32
mscotta	VadtecWork: can you help me with a problem compiling git-1.7.3.2	21:33
	← battleman2004_ left	21:33
	→ battleman2004_ joined	21:34
nicktoye	hi, can I ask a quick question. I'm new to Git and learning at the moment. If I already have a file tracked, and then I modify it, before I commit it, I need to add it again?	21:34
esc	nicktoye: indeed, add it to the index/stage	21:35
	nicktoye: you can check that it worked with 'git status'	21:35
	→ joshsmith joined	21:35
nicktoye	ok thanks. BTW, I am on Chapter 3 of 'Getting Good with Git' by Andrew Burgess	21:35
	← kipras left	21:36
esc	nicktoye: haven't read that one yet, thanks for the ref	21:37
	← steph021 left	21:37
nicktoye	hey no problem, by the way I am using Colloquy for IRC, how do you quote people back?	21:39
	← psoo left	21:39
	← flazz left	21:39
	← rvsjoen left	21:40
	→ urkud joined	21:41
	← curtana left	21:41
	→ rvsjoen joined	21:41
	→ Chani joined	21:43
_rane	you don't quote people	21:43
	→ werdan7 joined	21:44
Chani	how do I undo my last (local) commit? I can git checkout HEAD^ to move to the previous commit, but hte branch still points to hte one I want to delete	21:44
_rane	try tab for nick autocomplete	21:44
	→ iPad3Gslippyd joined	21:44
	← pigdude left	21:44
Eridius	Chani: git rset HEAD^	21:44
	err reset	21:44
Chani	Eridius: thanks :)	21:44
Eridius	if you want to throw away the local changes too, use --hard	21:44
nicktoye	ok, so if I want to reply to someone	21:44
Eridius	nicktoye: just start typing their name and hit tab	21:44
nicktoye	Eridius: like that?	21:45
	→ nadavoid joined	21:45
Eridius	yep	21:45
nicktoye	cool	21:45
	if I have a blank txt file and I have 2 branches including my master. I add text in the master version, and then switch to the other branch, and then add text there. then I switch back to master and run git merge [branch] - I get this	21:45
	CONFLICT (content): Merge conflict in README.txt	21:45
	Automatic merge failed; fix conflicts and then commit the result.	21:45
	← MattDiPasquale left	21:46
	→ alnewkirk joined	21:46
	← dharrigan left	21:46
	← DarkUranium left	21:47
	← ch077179 left	21:47
	← ignatp left	21:49
	← blee_ left	21:49
	← brntbeer left	21:49
	→ blee_ joined	21:49
	→ MattDiPasquale joined	21:50
mscotta	Anyone who ever compiled git?	21:51
nicktoye	anyone know the answer to my problem?	21:51
mmattice	large group of people there, I'm sure.	21:51
Eridius	nicktoye: you didn't ask a question	21:51
	← dvaske left	21:52
	→ erez joined	21:52
	← sako left	21:52
	→ dvaske joined	21:52
	← sonnym left	21:52
mscotta	Eridius, trying compiling git, I got the message "git-compat-util.h:152: fatal error: openssl/ssl.h: Nessun file o directory"	21:52
rson	mscotta: what distro?	21:53
nicktoye	Eridius: sorry, the question was why do I get a conflict?	21:53
mscotta	rson, Lubuntu	21:53
rson	mscotta: do you have openssl-dev installed (just a guess)	21:53
itrebal	nicktoye: try opening up the file and seeing what is committed, edit it until you're satisfied, and then commit it	21:54
albel727	nicktoye: because you edited the same lines differently in both branches. what else did you expect?	21:54
wereHamster	mscotta: or set NO_OPENSSL=1	21:54
nicktoye	albel727: I don't know, I'm new at this. Ok, so that makes sense. Because I have two files that are both different, they won't merge into one file?	21:55
	→ Frog joined	21:55
albel727	nicktoye: git is not a human. it can't miraculously understand, what your program means, and choose to keep one line over another, so he asks you to do it.	21:55
	← Frog left	21:55
	← erez left	21:55
albel727	s/so he/so it/	21:55
	← tizbac left	21:55
	albel727 feels that git is alive =D	21:56
nicktoye	albel727: Ok, so if the 2nd branch had the same lines, but added an extra line, then that would work. Let me try and get back to you.	21:56
	→ dharrigan joined	21:56
	← dharrigan left	21:56
	→ dharrigan joined	21:56
	← iPad3Gslippyd left	21:56
	← foocraft left	21:57
	→ slippyd joined	21:57
	→ hilem joined	21:57
	→ Yuffster joined	21:58
mscotta	rson, thanks! No error with libglobus-openssl-dev but why "apt-get build-dep" did not install it?	21:58
	← g0bl1n left	21:59
rson	mscotta: not a clue, i don't use a debian based distro so i'm not familiar with apt and friends	21:59
kandjar	msysgit is really annoying; somehow, it doesn't let me rebase.	21:59
	when I try to rebase and squash or reorder the commits; it freezes...	21:59
	→ skoop joined	21:59
kandjar	does anyone know a work around to that issue?	22:00
	← wereGrandma left	22:00
mscotta	rson: ok! Thank you very much!	22:00
drizzd	kandjar: do you have binary files in your repo?	22:01
nicktoye	hmm, ok, so I have master with README file that is blank, and I have authorization branch with README file that has a single line. I am in master and I run the merge command and it says it is already up to date. But it can't be can it?	22:01
	← eno left	22:02
Chani	nicktoye: what merge command? :)	22:02
	← sgh left	22:02
nicktoye	Chani: git merge authorization	22:02
Chani	huh	22:02
rson	did you commit the file in authorization?	22:02
nicktoye	yes	22:03
	← ajpiano left	22:03
rson	and the readme in master?	22:03
nicktoye	yes	22:03
	both git status in branches say working directory clean	22:04
	→ mplb joined	22:04
	→ patmaddox joined	22:04
	→ eno joined	22:04
albel727	nicktoye: you're doing it in a single repo, I hope? and readme is really empty, when you checkout master, and has a single line, when you checkout authorization? and yet merge does nothing?	22:05
	← pgib left	22:05
	← testX left	22:06
nicktoye	albel727: yes, well the file says it is 4kb in my finder.	22:06
	but it is empty in master branch, but has a single line in authorization	22:07
	← coderdad left	22:07
	← akahn left	22:07
	← tvw left	22:07
	← rgr left	22:07
	→ tvw joined	22:07
	→ frakturfreak_ joined	22:07
	← dels left	22:07
theoros	i think this is a bug in git	22:07
	→ rgr_ joined	22:08
theoros	it comes up now and then	22:08
	← quesada3 left	22:08
albel727	nicktoye: and current branch is master, when you do git merge authorization?	22:08
	← gebi left	22:08
kandjar	drizzd: in the repo or in the commits i m trying to rebase?	22:08
	→ dels joined	22:08
	← shamrock left	22:08
	→ shamrock joined	22:08
theoros	nicktoye: what happens if you git checkout back and forth?	22:08
	and then try the merge again	22:08
nicktoye	albel727: yes it is in master and I did to git merge authorization - so this could be a bug. As long as it is, then I'll note it and try again	22:08
kandjar	drizzd: as a matter of fact, I do.	22:08
	→ quesada3 joined	22:08
drizzd	kandjar: from the branch you are trying to rebase, can you do "time git format-patch --stdout --ignore-if-in-upstream <upstream>.." where <upstream> is the branch you are rebasing against	22:09
nicktoye	theoros: it says Already up-to-date	22:09
drizzd	and compare the time to the same command without --ignore-if-in-upstream	22:09
	(it's ok to cancel if it takes a long time	22:09
	) oh, and probably you should add >/dev/null	22:10
albel727	drizzd: what does "git log --graph --oneline --decorate --all" show you?	22:10
	drizzd: oops, sorry	22:10
	← frakturfreak left	22:10
drizzd	kandjar: argh, let me try that again	22:10
nicktoye	ok, I am going to start again.	22:10
albel727	nicktoye: what does "git log --graph --oneline --decorate --all" show you?	22:10
drizzd	"time git format-patch --stdout --ignore-if-in-upstream ..<upstream>"	22:10
nicktoye	apples-imac:systems nicktoye$ git log --graph --oneline --decorate --all	22:10
	* 56f7389 (HEAD, master) resolved conflict	22:10
	\|\	22:10
	\| * 47dc739 (authorization) added content from authorization branch for readme file	22:10
	* \| 1045f00 added content from master for readme	22:10
	\|/	22:10
	* 015e367 Initial commit	22:11
albel727	nicktoye: next time, use pastebin	22:11
nicktoye	ok sorry	22:11
	→ jonesy_ joined	22:11
kandjar	drizzd: I m actually trying to do: 'rebase -i'	22:11
drizzd	kandjar: doesn't make a difference	22:12
albel727	nicktoye: you have actually already merged and resolved conflict in favor of empty readme.	22:12
kandjar	ok if im in current, and I want to do git rebase -i master	22:12
	u want me to do: git checkout current ;	22:12
nicktoye	albel727: oh have I? i'll have another crack at it. practice makes perfect.	22:12
kandjar	"time git format-patch --stdout --ignore-if-in-upstream ..master"	22:12
	← pedrobelo left	22:12
kandjar	?	22:12
drizzd	kandjar: actually, remove the dots as well	22:13
theoros	nicktoye: git checkout master; git reset --hard 1045f00; git merge authorization	22:13
drizzd	time git format-patch --stdout --ignore-if-in-upstream master >/dev/null	22:13
	→ pedrobelo joined	22:13
drizzd	kandjar: or are you doing rebase -i HEAD~n ?	22:13
albel727	nicktoye: use the "git log --graph" line I told you to understand how your history evolves. it'll help you.	22:13
	→ scarabx joined	22:13
theoros	(or just use gitx if you're on mac os x since it's less painful)	22:13
	← goshawk left	22:13
drizzd	kandjar: ah, never mind	22:14
nicktoye	theoros: http://pastebin.com/1QGUZK0D	22:14
kandjar	wow it's dumping like crazy... :)	22:15
theoros	nicktoye: now git diff and you'll see the conflict	22:15
drizzd	kandjar: anyways, if that's the problem, you are suffering from a problem that's already fixed in git.git's pu branch, but I don't know when the fix will be released	22:15
nicktoye	albel727: yes I am using git log in this book	22:15
drizzd	kandjar: like I said, add >/dev/null	22:15
	← skoop left	22:15
kandjar	ok done in 52s	22:16
albel727	nicktoye: plain "git log" and "git log --graph --oneline --decorate --all" are like completely different. the former doesn't tell you anything about shape of your history/relations between branches.	22:16
	← jrmuizel left	22:17
kandjar	should it hang or go through?	22:17
drizzd	kandjar: it should terminate	22:17
kandjar	the test you ask me to do?	22:17
	it did	22:17
nicktoye	albel727: yes I know, I've done the latter too, but I am only on ch3, so baby steps. :)	22:17
kandjar	in 52s	22:17
	← yairgo left	22:17
drizzd	kandjar: but 52s is rather long, did you wait for the rebase to finish for so long?	22:18
kandjar	drizzd: so that mean that the bug which got fixed?	22:18
	← itrebal left	22:18
kandjar	last time I had the issue; I let it sit for hte night	22:18
drizzd	kandjar: well, it can take longer depending on the number of changes to binary files you have	22:18
	and depending on the size of those files	22:18
nicktoye	ok, and now if I want to update the authorization branch I checkout that branch and git merge master	22:19
	← joshsmith left	22:19
kandjar	with /dev/null /	22:19
drizzd	so yes, this performance problem will likely be fixed in the next release	22:19
kandjar	real 0m1.252s	22:19
	user 0m0.010s	22:19
	sys 0m0.030s	22:19
	→ joshsmith joined	22:19
kandjar	with /dev/null it's pretty fast.	22:19
albel727	nicktoye: I consider it a failure on book's part. nothing helps better that log --graph to understand what your actions lead to. it should have been introduced much sooner.	22:19
kandjar	the output flooding was the slow part :)	22:19
	← Chani left	22:19
albel727	nicktoye: yes, that will make autorization and master point to the same commit.	22:19
drizzd	kandjar: right...	22:19
nicktoye	albel727: the book does say --graph	22:20
kandjar	even in the prerelease 1.7.3 it wont be fixeD?	22:20
	Putr → Putr\|Away	22:20
nicktoye	so gitx then? better than Terminal?	22:20
albel727	nicktoye: ... --oneline --decorate --all? =)	22:20
kandjar	Git-1.7.3.1-preview20101002.exe ??	22:20
drizzd	kandjar: well, apparently this is not your issue anyways	22:20
kandjar	mmm	22:20
nicktoye	albel727: yep, thanks. :)	22:20
kandjar	oh. so what's causes mine then :(	22:21
albel727	nicktoye: gitx is definitely better on mac os. depends on whether you're a terminal maniac, though =)	22:21
drizzd	kandjar: I would propose to analyze with perf, but that probably won't work on windows...	22:21
	← joshsmith left	22:21
nicktoye	albel727: well I'm a web designer	22:21
	← Ilari left	22:22
albel727	nicktoye: ... so, you're a terminal maniac, then. I get it, no need to elaborate further. =P	22:22
kandjar	true,	22:22
	Ilari_antrcomp → Ilari	22:22
kandjar	I guess I could rebase by hand..	22:22
	→ Schmallon joined	22:23
kandjar	cherry pick each commit I don't want to rebase, then do a git reset soft	22:23
nicktoye	albel727: gosh no. I'm kind of forcing myself to use it as it's quite powerful apparently.	22:23
kandjar	and one massive commit.	22:23
albel727	nicktoye: I was kidding =)	22:23
nicktoye	albel727: I know.	22:23
	→ Ilari_ joined	22:23
	Ilari → Ilari_antrcomp	22:23
	Ilari_ → Ilari	22:23
kandjar	geeezz... as much as I like git... as much I HATE msysgit!	22:24
xrogaan	what's the difference ?	22:25
	← artagnon left	22:25
kandjar	drizzd: do you know if it's safe to install a newer version on top of the existing one? or should I remove the existing one first?	22:25
drizzd	kandjar: hmm, I still wonder what else the problem could be	22:25
albel727	nicktoye: it's really simple. regularly you'll need to use 2-3 commands. proper gui-s make it easier. e.g you may want to look in direction of man git-mergetool, to resolve merge conflicts with your favorite merger/differ app, etc. also there are git gui, gitx, etc...	22:26
Gitbot	nicktoye: the git-mergetool manpage can be found at http://git.or.cz/man/git-mergetool	22:26
drizzd	kandjar: but there is no way to find out unless you run some profiling tool	22:26
	→ graphitemaster joined	22:26
graphitemaster	ffs!	22:26
kandjar	FYI: $ git --version	22:26
	git version 1.7.0.2.msysgit.0	22:26
drizzd	kandjar: it's safe to install a new version on top of the existing one	22:26
graphitemaster	dammit git fucked up	22:26
	graphitemaster could kill git right now	22:26
graphitemaster	JESUS, that is bullshit	22:26
wereHamster	yeah, blame the tool and not the user	22:27
	wereHamster blames rm for deleting his files	22:27
drizzd	xrogaan: msysgit is the windows port of git	22:27
	graphitemaster made about idk 600 changes, forgot to pull so he could not push it to master, so dumb him pulled and lost everything :(	22:27
kandjar	gonna try the 1.7.3 prev then..	22:27
	→ ignatp joined	22:27
	→ sonnym joined	22:27
nicktoye	albel727: well I have Versions for SVN, and I like that. So as a designer I want beautiful GUI or bare bones old school stuff. :)	22:27
wereHamster	graphitemaster: what? You pulled and lost evrything? That's not possible	22:27
kandjar	graphitemaster: if you committed stuff in git, it's not lost.	22:27
nicktoye	but at the moment, I just need to get through this book, to get a taster for the commands.	22:28
xrogaan	drizzd: I know	22:28
	← Anti-X left	22:29
	← mplb left	22:29
graphitemaster	http://pastebin.com/MGh4mAqP	22:29
	^:(	22:29
	http://github.com/graphitemaster/Kill--Field-Engine/commit/c2840c431d8854c2ffa5aeedf004070636abfe22	22:29
	latest change, and it's that stupid merge	22:30
wereHamster	graphitemaster: and what did you lose?	22:30
	of course it's a merge, pull = fetch+merge	22:30
drizzd	xrogaan: well, that is the only difference...	22:30
	← blarson left	22:30
graphitemaster	ohh fewhh, never mind	22:30
	I got it, okay wow my heart was beating about 6000000000 times a millisecond there	22:30
	graphitemaster absconds	22:31
	→ blarson joined	22:32
	→ coderdad joined	22:32
	← d0k left	22:32
	→ joshsmith joined	22:32
	← coderdad left	22:33
	← Dave^\| left	22:33
	← deadnoralive left	22:34
	← nicktoye left	22:35
	→ xiong joined	22:38
	→ ProLoser joined	22:39
ProLoser	how can i move a submodule from one folder to another?	22:39
	git mv doesn't work	22:40
	→ EricInBNE joined	22:40
	← dreiss left	22:41
	→ sako joined	22:41
	→ envi joined	22:42
	← tjholowaychuk left	22:42
	← ehabkost left	22:42
kandjar	so far so good; first part of the rebase worked after the upgrade.	22:42
	← dharrigan left	22:45
	→ dr_lepper joined	22:45
	← bcardarella left	22:46
	→ dharrigan joined	22:46
	← Evious left	22:46
	← shennyg left	22:46
frogonwheels	ProLoser: delete the submodule and re-add it in one commit I would say	22:47
	ProLoser: that's pretty much all that a git mv does	22:47
	→ shennyg joined	22:47
	← Schmallon left	22:48
	→ Evious joined	22:48
	→ warlock_mza joined	22:49
	→ brntbeer joined	22:50
getxsick	hi guys	22:51
	→ Ogredude_ joined	22:51
	→ ajpiano joined	22:51
fr0sty_	getxsick: yo	22:52
getxsick	fr0sty_: oh, nice to see you again	22:52
	fr0sty_: i spent some time on reading manuals	22:52
	and I would like to sum it up (the whole workflow)	22:52
	maybe i will draw a new diagram	22:54
	← silver_hook left	22:54
	← dels left	22:55
	→ converge joined	22:55
	← Yuuhi left	22:56
	← ProLoser left	22:57
	→ bowenl2 joined	22:57
	← warbee left	22:58
bowenl2	to act as though I completely re-cloned a repo, disregarding the changes i've made to my working company, what commands should i use?	22:58
	s/company/copy/	22:58
	→ zombor_ joined	22:59
	← zombor_ left	22:59
	→ zombor_ joined	22:59
	→ MikhX_ joined	22:59
bowenl2	git checkout --force	23:01
	causes git pull to fail the next time another working copy is updated	23:02
	→ ProLoser joined	23:02
ProLoser	i want to undo the last commit i did	23:02
	permanently	23:02
	← MikhX left	23:02
	MikhX_ → MikhX	23:02
Eridius	ProLoser: git reset --hard HEAD^ (watch out, this will throw away uncommitted changes too). Warning: don't do this if you've already pushed the commit	23:03
	→ kenichi joined	23:03
	← pantsman left	23:03
ProLoser	okay, now i did that, i want to pull another fork into mine, but i want to use rebase, as there are no conflicts	23:04
Eridius	git pull --rebase?	23:04
	← xiong left	23:04
ProLoser	strange, it doesn't show up when i do -h	23:04
	okay	23:04
Eridius	the manpage si your friend	23:05
	→ MetaCosm joined	23:05
	→ alberto56 joined	23:06
fr0sty_	bowenl2: what do you want to get rid of? just the working copy changes?	23:07
	or do you have commits you want to throw away?	23:07
bowenl2	yes, every change	23:07
getxsick	fr0sty_: http://img34.imageshack.us/img34/9421/repox.png what do you thinkg about it?	23:08
bowenl2	i want to re-clone it without actually deleting and recloning	23:08
fr0sty_	bowenl2: git reset --hard <remote>/<branch>	23:08
theoros	getxsick: oh god	23:08
fr0sty_	that will reset your branch to the remote state.	23:08
	← Ryback_ left	23:08
bowenl2	eek thanks	23:09
getxsick	theoros: ?	23:09
fr0sty_	bowenl2: or just delete all your local branches.	23:09
bowenl2	i didn't realize i needed origin there	23:09
fr0sty_	'git branch' and then 'git branch -D <branchname>	23:09
theoros	getxsick: those are some pretty bad graphs. what are you graphing? general workflows with git or a specific one? for general there's http://progit.org/book/ch5-1.html	23:09
	← chovy left	23:09
getxsick	theoros: very specific	23:09
	theoros: it's a continuation from the talk yesterday	23:10
theoros	i see	23:11
	← brosner left	23:11
getxsick	theoros: if you would like to join to the discussion, i can say some words	23:11
fr0sty_	theoros: they have people working at a remote site with only one person authorized to connect over the vpn.	23:11
	← dr_win left	23:11
fr0sty_	the 'superuser' will be marshalling all the commits to be pushed across to 'main server'.	23:12
	that's what I understand at least...	23:12
getxsick	yep	23:12
	+ all changes pushed to the 'main server' has to be done as a superuser (i mean about Author tag here)	23:12
	fr0sty_ has to bow out...	23:12
theoros	somewhat like http://progit.org/figures/ch5/18333fig0503-tn.png	23:12
	just with more restrictions	23:12
	fr0sty_ comes back for one sec...	23:13
getxsick	theoros: no sure	23:13
fr0sty_	only one 'Author'?	23:13
	← boucman left	23:13
fr0sty_	can there be multiple committers?	23:13
getxsick	fr0sty_: yes, that's why i want to rebase it	23:13
	fr0sty_: multiple commiters on 'repoB'	23:13
	→ boucman joined	23:13
fr0sty_	why go through the trouble?	23:13
getxsick	but one commiter on the main server from all changers through VPN	23:13
fr0sty_	why?	23:14
	what are you trying to accomplish?	23:14
getxsick	repoB is a different company	23:14
	and mainserver is a different	23:14
	← bowenl2 left	23:14
fr0sty_	you do realize that rebasing rewrites all of your commits, right? so they are a separate history.	23:14
theoros	why not just give access to the repository and only the repository?	23:14
fr0sty_	where does 'Author' enter into this.	23:14
	← dharrigan left	23:14
	← RoosterJuice_ left	23:14
getxsick	fr0sty_: this is a restriction :(	23:14
fr0sty_	getxsick: your 'superuser' can push commits authored by anyone...	23:15
getxsick	they just want to see in git log/gitk author as a name of superuser	23:15
	fr0sty_: yes	23:15
	fr0sty_: they all commits pushed by 'superuser' has to be visible as his own	23:15
	only on 'mainserver'	23:15
fr0sty_	getxsick: then save everyone some pain and set-up everyone to commit with --author="superuser".	23:15
getxsick	on repoB it's more than OK to keep multiple authoring	23:16
fr0sty_	then the committers can be themselves...	23:16
	getxsick: you are requiring your team to maintain parallel histories. this is, uh, suboptimal...	23:16
	← Lgb left	23:16
getxsick	fr0sty_: hmm, if everyone will commit as 'superuser' in repoB history all changes will be as 'superuser', am i right?	23:16
fr0sty_	getxsick: each commit has an 'Author' and a 'Committer'	23:16
getxsick	fr0sty_: i know, i don't like it either	23:16
	hmm	23:17
	ok	23:17
	wait	23:17
	let me check what is what	23:17
	← Heimidal left	23:17
getxsick	probably i'm confusing definitions here	23:17
	← icwiener left	23:17
getxsick	when i do `git log` i see 'Author' right?	23:17
	← converge left	23:17
fr0sty_	you can force the 'Author' to be someone else if necessary, but please, please, please don't have someone filter-branch all your commits to push them...	23:18
getxsick	fr0sty_: ok, so what is 'Commiter' then?	23:18
fr0sty_	'git log --format=full	23:18
	← Phrogz left	23:18
fr0sty_	or 'git show <commit>'	23:18
	← c3l left	23:19
getxsick	'git sshow <commit>' shows only Author	23:19
FauxFaux	git cat-file commit <commit>	23:19
getxsick	anyway i see the point	23:19
	ok	23:19
	so the question is	23:19
fr0sty_	ack, need --format=full on that too.	23:19
	→ Lgb joined	23:19
getxsick	i would like to allow all commiters on repoB to use his own Author and Commiter	23:19
	← mscotta left	23:19
FauxFaux	The question is "Why do you press return so often?". ¬_¬	23:19
	→ harleypig joined	23:20
getxsick	FauxFaux: sorry :)	23:20
wereHamster	again!	23:20
	→ dharrigan joined	23:20
	← dharrigan left	23:20
	→ dharrigan joined	23:20
fr0sty_	don't, please don't,...	23:20
	I will say no more,	23:20
	← ari-_-e left	23:20
	fr0sty_ hangs it up for the day...	23:20
	wereHamster too	23:20
getxsick	if 'superuser' pulls all changes from them to his cloned repository, can he rebase these pulled commits to change Author/Commiter and push it to 'mainserver' ?	23:21
	← mw left	23:22
	→ albel727_2 joined	23:22
harleypig	On github there are three separate repositories for vimoutliner, none appear to be official. Assuming the other two repositories don't go too far astray I'd like to merge changes made to those other repos. Is there a tutorial for using multiple repos for remote repos?	23:23
	← MetaCosm left	23:23
	← brntbeer left	23:24
	← albel727 left	23:24
	← zain left	23:25
	← SvenDowideit left	23:25
	→ SvenDowideit joined	23:25
	← thear left	23:25
	← MattDiPasquale left	23:26
	→ battleman2004__ joined	23:26
	→ battleman2004 joined	23:28
	← battleman2004_ left	23:29
	← battleman2004__ left	23:31
	→ zain joined	23:32
	← dharrigan left	23:32
	zombor_ → zombor	23:33
	→ ari-_-e joined	23:33
	← richardbronosky left	23:34
	→ madewokherd joined	23:35
	MOAR-f00li5h → f00li5h	23:36
	← LiamH left	23:38
	→ MetaCosm joined	23:41
	← MetaCosm left	23:41
	→ MetaCosm joined	23:41
	← golodhrim left	23:42
	→ bhosie joined	23:42
	→ golodhrim joined	23:42
	← kenichi left	23:42
	→ dharriga1 joined	23:43
	→ cognominal joined	23:43
ProLoser	i'm a complete moron. I did a rebase pull of another dev's work and it's a mess, my commit is the head commit, i'm thinking of reverting back to before his commits and putting my commit back in, how wouldst i do this?	23:44
	checkout the last commit before his and then... cherry pick i'm guessing?	23:44
	← Bigcheese left	23:44
ProLoser	do i like cherry pick onto an old commit a newer commit, or does cherry picking only work for branches	23:44
	in other words, before i try this, should i checkout a past commit or reset to it?	23:45
theoros	before you do anything, tag your current commit	23:46
ProLoser	tag it?	23:46
theoros	yes, tag it	23:46
	with a tag	23:46
	→ ralf_e joined	23:46
bob2	or cp -a the whole repo	23:46
theoros	so that you have a ref to get reachability	23:46
bremner	well there is the reflog	23:46
frogonwheels	ProLoser: git reset --hard HEAD@{1} will undo your last operation for commit/rebase/merge	23:47
theoros	the reflog is a scary place	23:47
frogonwheels	ProLoser: (which is the reflog)	23:47
theoros	i refer tagging stuff to make sure i can easily get to things	23:47
	and then deleting the tag when i'm in working order	23:48
	prefer*	23:48
	← dvaske left	23:49
	← Kira left	23:49
	→ dvaske joined	23:49
	→ drubies_ joined	23:53
	← slippyd left	23:55
	← beatak left	23:57
	→ brntbeer joined	23:57
ProLoser	okay it's tagged	23:57
	so now what?	23:57
	→ bhowmisu joined	23:59

Logs Search ←Prev date Next date→

Channels Documentation