IRCloggy #git 2010-10-30

Provider of IRC logs since 2005.

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

	2010-10-30
	→ RobertLaptop joined	00:01
	→ envi joined	00:01
DrNick	carpii: git log B..HEAD	00:02
carpii	awesome, thanks :)	00:03
	← chrischambers left	00:03
DrNick	x..y is shorthand for ^x y, that is all commits reachable from y, excluding all commits reachable from x	00:03
carpii	if i cherrypick commits from A to B, and then merge A into B does it ignore the ones already cherrypicked >	00:04
DrNick	sort of	00:04
	cherry commits aren't the same commits as their originals, so you'll end up with a merge, but the merge won't (or shouldn't) conflict	00:05
carpii	ok, thanks again	00:06
	one last Q	00:06
	→ blaines joined	00:06
carpii	if i create a remote branch B off master, make commits to master and to B, merge master into B and then merge B into master	00:06
	is there a reason this would always create conflicts ? or ive just been unlucky tnite	00:06
	← AAA_awright left	00:07
DrNick	what do you mean by "remote branch B off master" ?	00:07
	← pedrobelo left	00:08
carpii	lets say i created a remote branch, and its up to date with master	00:08
	← DaveIngram left	00:08
carpii	i have a dev working on a remote branch, and occasionally he pulls my latest changes in from master	00:08
DrNick	by create a remote branch, you mean you created a branch locally and then pushed it to your central repository?	00:09
carpii	but when he commits everything back, and i try to merge his remote branch back into master, i seem to get conflicts more often than id expect	00:09
	yes, thats it	00:09
	→ whaley joined	00:09
	← mw left	00:09
	→ pedrobelo joined	00:09
carpii	pushed as a remote branch though	00:09
	theres a remote branch B, and myself and the dev has a local tracking branch B	00:09
bob2	is your expectation taking into account "file-content conflicts occur when people edit the same parts of the same file, also some people have terrible editors that churn line endings"? :)	00:10
	← joel_falcou left	00:10
DrNick	there's (potentially) six different branches involved in this scenario, and I'm not sure which you're talking about	00:10
	→ joel_falcou joined	00:11
carpii	bob, yes id expect those to conflict :)	00:11
	← blaines left	00:12
carpii	DrNick: its kinda hard to explain i guess. I just wanted to make sure its not a problem for him to be merging master into his branch, and then pushing back his commits to the remote one	00:12
	where 'his branch' is a local copy of the remote branch	00:12
	← hobodave left	00:13
	→ dreiss joined	00:18
	→ zackattack joined	00:18
zackattack	I'm on branch master, but my branch is ahead of origin/master by 9 commits. How do I save all my changes, and then make my working copy equivalent to origin/master? I have a bunch of changes that I don't want to implement until later ...	00:18
carpii	have the changes you dont want pushed, been committed yet ?	00:19
zackattack	yes	00:19
	← dr_win left	00:20
	← Dave^\| left	00:20
carpii	you work exclusively on master or something ?	00:21
	its normally a good idea to only merge branches into master when theyre ready to be pushed live	00:21
Eridius	go2: /m	00:22
	goddamnit	00:22
carpii	i think you can probably 'git reset <commit>'	00:22
	then git push	00:22
	then git reset back	00:23
zackattack	uh..	00:23
carpii	but check git status shows fewer commits, before you push, cos im not 100% sure	00:23
	→ madeofcode joined	00:24
madeofcode	advanced git q: i have 3 branches (master, feature1, feature1+2) can i get a diff of just feature 2 (i.e. (feature1+2 - feature1) differ with master) ?	00:25
carpii	feature1+2 is same as feature1 with a few more commits ?	00:28
	← cschneid left	00:29
	← bitkiller left	00:29
	→ lancepantz joined	00:29
	← hilem left	00:30
lancepantz	'git submodule update' should pull down submodules after cloning a repo, correct?	00:30
kokx	lancepantz: you first have to initialize them	00:31
	with git submodule init	00:31
	→ AAA_awright joined	00:31
madeofcode	carpii: yes	00:32
	s0ber_ → s0ber	00:32
madeofcode	carpii: but not neccessarily in order	00:32
	i.e. feature1+2 will have merged in feature1 a few times during it's development	00:33
lancepantz	thanks kokx, that did it	00:33
	← lancepantz left	00:33
	← wsclock left	00:34
	→ kanru joined	00:36
	→ bdiu joined	00:36
carpii	made, i dont know, sorry	00:37
	i just asked a similar question which might help partially	00:37
	carpii: is there a way of showing commits which are in current branch, but havent yet been merged into branch B ?	00:37
	DrNick: carpii: git log B..HEAD	00:37
	→ shiba_yu36 joined	00:38
	← shiba_yu36 left	00:38
	← tj left	00:39
	← dfr\|mac left	00:40
	← warthog9 left	00:41
	← parasti left	00:41
	← rudyl313 left	00:41
	→ tungtn joined	00:42
	← Evious left	00:43
	← tvw left	00:43
	→ cschneid joined	00:50
	← towski left	00:51
	→ nullvariable joined	00:52
	← zombor left	00:54
DrNick	carpii: if you were paying attention, you'd be able to guess that ^master ^feature1 feature1+2 would list the relevant commits	00:55
	← dreiss left	00:55
	→ mmoya joined	00:56
carpii	sorry, id just joined at the time. i didnt realise that was an answer to me	00:57
	← Schmallon left	00:58
	→ shiba_yu36 joined	00:58
	→ thiago joined	01:00
	→ caesarkw joined	01:00
	← brianpWins left	01:01
	← Amorphous left	01:03
	→ brntbeer joined	01:04
	→ thecoffman joined	01:04
	← yhager left	01:04
	→ yhager joined	01:04
	← caesarkw left	01:05
sitaram	adymitruk: I don's use any, but gitweb comes with git anyway. And some people like cgit -- enough to send patches to cgit to make it compatible with gitolite :-)	01:05
	well, gitweb conventions, actually	01:05
	← rikarends left	01:06
	← nevyn^ left	01:06
	→ nevyn^ joined	01:06
	← steph021 left	01:07
	← aah left	01:08
adymitruk	sitaram: thanks.. trying to configure gitweb right now	01:09
	← nevyn^ left	01:11
	→ Kira joined	01:11
	← warbee left	01:11
	→ jdonley83 joined	01:14
	← mmoya left	01:15
	← jdonley83 left	01:15
	← jer left	01:15
	Kira → Kal	01:17
	Kal → Kira	01:17
	→ Amorphous joined	01:19
	← Vampire0 left	01:23
	← aspotashev left	01:25
	← tstclair left	01:25
	← yhager left	01:26
	→ verto joined	01:29
	← comawhite\|laptop left	01:30
	← bdiu left	01:34
	← MikhX left	01:34
	→ MikhX joined	01:35
	→ bdiu joined	01:35
	→ comawhite joined	01:35
	← MikhX left	01:35
	→ MikhX joined	01:35
	← MikhX left	01:36
	→ Vampire0 joined	01:37
	→ mw joined	01:38
	→ towski joined	01:38
	← spearce left	01:38
	← tildeequals left	01:40
	← kipras left	01:40
	→ dharrigan joined	01:40
	← mw left	01:40
	← murr4y left	01:43
	→ brianpWins joined	01:43
	→ aresnick joined	01:44
	→ jrmuizel joined	01:46
	← carpii left	01:47
	← brianpWins left	01:48
	← abner left	01:49
	← jrmuizel left	01:49
	← slonopotamus left	01:50
	← nakkor left	01:52
	Kira → k_sze	01:52
	← thiago left	01:53
	→ jrmuizel joined	01:53
	→ abner joined	01:53
	→ rayners joined	01:53
	← dharrigan left	01:54
	→ alex__c2022 joined	01:54
	← bhosie left	01:57
	← abner left	01:58
	← alberto56 left	01:58
	← Elfe left	02:00
	→ Elfe_ joined	02:00
	→ quesada4 joined	02:02
	→ aspotashev joined	02:02
	← dukebag left	02:02
	← quesada3 left	02:04
	→ dharrigan joined	02:05
	→ yhager joined	02:07
	→ [1]intripoon joined	02:07
	→ jammur joined	02:09
	← intripoon left	02:10
	[1]intripoon → intripoon	02:10
	→ bhosie joined	02:11
jammur	I'm new to git and I think I'm misunderstand something. I created a new branch, checked out the branch, made a modification, switched back to the master branch, and when I ran git status, it showed the modification under the master branch as well. I though any modifications made in a branch were "sandboxed" inside that branch?	02:13
	← dharrigan left	02:15
Eridius	jammur: you didn't commit it	02:15
	it's still just a change in your working copy, and you can switch branches with that change there as long as switching to the new branch doesn't touch any of your modified files	02:16
	→ artefon joined	02:17
jammur	Eridius: Ok, that makes more sense. Thanks!	02:17
	→ dharrigan joined	02:21
	← andyburke left	02:23
offby1	jammur: 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"	02:24
	→ jherdman joined	02:26
	← dharrigan left	02:29
	→ abner joined	02:29
	→ dharrigan joined	02:30
	← dharrigan left	02:30
	→ dharrigan joined	02:30
	→ dukebag joined	02:38
	← dharrigan left	02:40
	← atpa8a left	02:40
	← whaley left	02:40
	← madeofcode left	02:45
	→ dharrigan joined	02:46
	← dharrigan left	02:46
	→ dharrigan joined	02:46
	← yngress left	02:46
	→ MikhX joined	02:47
	→ yngress joined	02:48
	← yngress left	02:49
	← intripoon left	02:49
	← k_sze left	02:49
	→ k_sze joined	02:50
	← Eridius left	02:51
	→ timj_ joined	02:52
	← k_sze left	02:53
	← aspotashev left	02:55
	→ k_sze joined	02:56
	← timj__ left	02:56
	← dharrigan left	02:56
	← eban left	02:58
	← jammur left	02:58
	← eft_ left	02:59
	→ Slix` joined	03:02
	→ dl9pf joined	03:03
	← dl9pf left	03:03
	→ dl9pf joined	03:03
	← dl9pf_ left	03:04
	← dukebag left	03:08
	← cschneid left	03:11
	← k_sze left	03:13
thiana	What effect does 'git init' have on a pre-existing repo? It says that it reinitialized the repo but nothing seems to change?	03:13
	→ jammur joined	03:14
	→ MmikeMRMA joined	03:15
	→ cschneid joined	03:16
	→ dharrigan joined	03:17
	← dharrigan left	03:17
	→ dharrigan joined	03:17
	→ cconstantine joined	03:17
cconstantine	http://www.kernel.org/pub/software/scm/git/docs/howto/revert-a-faulty-merge.txt	03:18
	→ k_sze joined	03:18
	→ rekeorbm joined	03:18
cconstantine	I'm trying to follow the directions in that link to recreate a branch for later merging	03:18
	and the --no-ff argument is rejected by 'git rebase'	03:19
	← robgleeson left	03:19
	Thell\|Away → Thell	03:19
cconstantine	I'm using git version 1.7.0.4	03:19
	← mbroeker left	03:19
	← MmikeRMRM left	03:19
cconstantine	so, I've reverted a merge that shouldn't have happened and I need to be able to merge that branch in later	03:19
	rekeorbm → mbroeker	03:19
cconstantine	is there fresh docs on how to do that?	03:19
	→ Wyverald joined	03:21
	→ gnufied joined	03:23
	← dharrigan left	03:23
	← verto left	03:23
cconstantine	no one here? On a friday night? I'm shocked!	03:26
	→ robgleeson joined	03:26
	← kanru left	03:28
	← wooden left	03:30
	→ Elfe joined	03:30
	← fr0sty__ left	03:31
	→ fr0sty__ joined	03:32
	→ quesada joined	03:33
	← Elfe_ left	03:33
Arafangion	Do we have a library of git hooks available, or are people generally expected to write their own?	03:34
	← thecoffman left	03:34
Arafangion	I would like a hook that automatically sync's two git repositories (as best as possible) every time a push or pull is performed.	03:35
	→ thecoffman joined	03:35
	← quesada4 left	03:35
	→ gilaniali joined	03:41
	← kukks left	03:41
	→ babusri joined	03:41
	→ wooden joined	03:41
	← wooden left	03:41
	→ wooden joined	03:41
	← yhager left	03:42
	← jammur left	03:43
	← babusri left	03:44
	→ babusri joined	03:44
	← gnufied left	03:46
	← artefon left	03:46
sitaram	Arafangion: you cannot do that safely in both directions. One has to be master, and any number of slaves	03:46
	← bhosie left	03:46
	← gilaniali left	03:46
sitaram	Arafangion: think about it; same problem as bare repos -- a potential merge in a non-human-intervention scenario	03:47
	→ gilaniali joined	03:47
sitaram	Arafangion: and finally, a pull does not have any hooks; the idea is you can as well replace the pull with a script that does pull+whatever you want. Only pushes have hooks, on the receiving side. Specifically, in this case, post-receive/post-update	03:48
Arafangion	Hmm.	03:48
	← phantomcircuit left	03:48
	→ dharrigan joined	03:49
	← dharrigan left	03:49
	→ dharrigan joined	03:49
	← Bass10 left	03:50
Arafangion	Sounds like I'd need to do a manual hook, in any case - set one up as "master", with the slave always pushing changes through, with forced updates. In the event of a possible loss of data, (as would be if a forced update was truely needed), then I could have it stow those in some new branch, and have it email me.	03:51
sitaram	Arafangion: I strongly suggest all pushes should only be to one server ("the master", if you will) and a post-receive hook on that does a "git push --mirror <slave url>"	03:52
Arafangion	Unfortunately, communication is one-way.	03:52
	Stupid VPN stuff.	03:53
sitaram	in which direction? rather, what's inside and what's outside, and which one do the devs have access to?	03:53
Arafangion	The master is accessible by /most/, but the slave is only accessible at the remote.	03:53
	I'll setup a cronjob or something so that the slave eventually updates.	03:54
	← wooden left	03:54
Arafangion	I think any solution is going to need someone to keep an eye on it, anyway, I might as well make it mostly automatic, without trying for perfection.	03:55
	→ wooden joined	03:55
	← wooden left	03:55
	→ wooden joined	03:55
sitaram	can the master reach the slave, or is it vice versa?	03:56
Arafangion	The master has no way of contacting the slave.	03:56
sitaram	ok, then dont use a hook on the master to 'push --mirror'. Use a cron job on the slave to fetch from the master instead. And tell devs to push to master	03:57
	→ slonopotamus joined	03:58
Arafangion	Yeah... I'll have to find some workaround - thanks for your help.	03:58
	← robgleeson left	03:58
sitaram	Arafangion: the only thing you need to guard against is I push a commit to master and you push to slave -- at that point, nothing automatic can deal with it, and you have to manually do stuff that should have been taken care of, normally, by git saying "sorry pal, pull first because the remote has moved on"	03:58
Arafangion	sitaram: I plan to make the push first do a pull, for good measure.	03:59
	→ robgleeson joined	03:59
sitaram	IOW, if you allow pushes to different servers, you end up postponing that "aah someone updated, we need to merge [and possibly resolve conflicts]" moment to much later. Not a good idea	03:59
Arafangion	sitaram: Will make their pushes much slower, but can't be helped.	03:59
sitaram	Arafangion: a non-human process has no concept of a pull. How will it resolve conflicts if any?	03:59
	other way is to let slave get commits only on a special branch which master does not get pushed directly, carry those across using cron, resolve manually on master -> dev fetch (humanly), and then update the real branch	04:00
	sorry if that was confusing	04:00
Arafangion	It would need to simply update the reference - it wouldn't do a merge or a rebase, which possibly results in lost data. I'd have to email a warning in that case.	04:01
	← alex__c2022 left	04:04
	← bdiu left	04:05
Arafangion	I think we're looking at this the same way - the difference is when one has to fix the issues. I simply make the potential loss visible and possible to fix up later, whereas your solution forces it to be dealt with at the time.	04:05
	← dharrigan left	04:06
	→ bdiu joined	04:06
	→ jrt05 joined	04:07
	← dl9pf left	04:07
	→ robinr joined	04:08
	→ kenneth_reitz joined	04:10
sitaram	yup. That's what I meant by "postponing that moment"	04:10
Arafangion	Most of the dev's are extremely new to git, I'd rather not make them frustrated with a complex setup just yet.	04:10
	They all use gui tools, for example.	04:11
	← robinr left	04:11
	→ bhosie joined	04:12
	← kenneth_reitz left	04:12
	← ignatp left	04:14
	→ dharriga1 joined	04:16
	→ cytrinox_ joined	04:20
	← xiong left	04:21
	← witquicked left	04:22
	→ rekeorbm joined	04:22
	← rekeorbm left	04:23
	← cytrinox left	04:24
	cytrinox_ → cytrinox	04:24
	← robgleeson left	04:24
	← mbroeker left	04:25
	→ simplechat joined	04:28
	→ childermass joined	04:28
	→ mmoya joined	04:28
	→ dukebag joined	04:32
	← dharriga1 left	04:33
	→ dharrigan joined	04:33
	← dharrigan left	04:33
	→ dharrigan joined	04:33
	→ urkud joined	04:34
	Thell → Thell\|Away	04:35
	← jrmuizel left	04:37
	← zackattack left	04:38
	→ gnufied joined	04:43
	← urkud left	04:45
	→ madeofcode joined	04:45
	← mmoya left	04:46
	→ yhager joined	04:48
	→ mithro joined	04:49
	← hyperair left	04:56
	→ urkud joined	04:57
	← Pacman_Home` left	04:58
	→ t3i joined	04:58
	→ ct^^ joined	05:00
	→ xiong joined	05:00
	← ct^ left	05:03
	← programble left	05:03
	→ kenneth_reitz joined	05:05
	← kenneth_reitz left	05:07
	← dharrigan left	05:08
	← urkud left	05:09
	← gnufied left	05:09
	→ sako joined	05:09
	← shennyg left	05:11
	→ shennyg joined	05:12
	→ hyperair joined	05:14
	← sako left	05:14
	← thecoffman left	05:17
	← SkramX left	05:17
	→ gnufied joined	05:18
	→ dharrigan joined	05:19
	← dharrigan left	05:19
	→ dharrigan joined	05:19
	← gnufied left	05:21
	← t3i left	05:22
	← Marquel left	05:24
	← dharrigan left	05:26
	→ thiago joined	05:28
	→ zombor joined	05:31
	← zombor left	05:32
	← brntbeer left	05:33
	→ erez joined	05:40
	← jfkw left	05:40
	→ yuriks_ joined	05:42
	← yuriks left	05:45
	→ Marquel joined	05:45
	→ dharriga1 joined	05:47
	← jherdman left	05:48
	← Slix` left	05:49
	← bhosie left	05:49
	→ Pacman_Home joined	05:52
	← xiong left	05:52
	← rayners left	05:53
	→ robgleeson joined	05:53
	← dharriga1 left	05:53
	→ cactux joined	05:57
	→ kanru joined	05:58
	→ amerine joined	05:59
	← amerine left	06:00
	→ amerine joined	06:00
	← amerine left	06:00
	→ gnufied joined	06:00
	→ amerine joined	06:01
	← gilaniali left	06:01
	→ dharrigan joined	06:04
	← cactux left	06:04
	← doener left	06:05
	→ doener joined	06:05
	→ gilaniali joined	06:05
k_sze	Does anybody know if there are GnuPG keysigning parties targeted specifically at Grails/Groovy/Spring Framework developers who also use Git/GitHub?	06:09
bob2	that sounds like a pretty narrow focus	06:10
slonopotamus	ah, those who also eat tuna on thirsdays?	06:11
	→ cactux joined	06:11
bob2	just go to a regular keysigning	06:11
	← Yuffster left	06:11
cactux	I clone a manifest and I set a envvar to the directory and after I execute "repo init -u $MANIFEST", but I get "Usage: repo init -u url [options]"	06:11
	The default.xml is in $MANIFEST. MANIFEST=/dir/manifest	06:12
	I ran the same command in my server and it works	06:12
	So I do not understand why it fails in my local machine	06:13
k_sze	bob2: yes, those who eat tuna on thirsdays and who also happen to live in Hong Kong.	06:13
	*Thursdays… =_=;	06:14
slonopotamus	cactux: are you sure you're talking on right channel?	06:16
	← dharrigan left	06:19
	→ maleknet joined	06:20
k_sze	What is wrong with lightweight tags?	06:25
	→ andyl joined	06:27
k_sze	Scott Chacon says in his slides that lightweight tags "can be moved around easily", making them "generally undesirable". Anybody knows what he means there? o_O	06:28
	→ Yuffster joined	06:29
	→ dharrigan joined	06:30
	← dharrigan left	06:30
	→ dharrigan joined	06:30
	→ Ogredude_ joined	06:30
	→ SIGBUS joined	06:34
	← dharrigan left	06:38
	← Yuffster left	06:38
	→ giles joined	06:47
giles	I sync repos using a cron job, there must be a better way?	06:47
	albel727 → buffy	06:48
buffy	giles: depends on what your repositories are. are they bare?	06:48
	← SIGBUS left	06:48
giles	one on github one on gitorious so i reckon	06:49
buffy	giles: hmm, well, you can put a post receive hook onto one of them, which will push updates to another and vice versa... but I'm not sure how stable that would be.	06:50
	← slonopotamus left	06:50
buffy	giles: out of curiosity, why different repos on different hosting?	06:51
giles	hmm maybe I will stick with cron then	06:51
	buffy: political issues	06:51
buffy	strange	06:51
	buffy → albel727	06:51
	← sadmac left	06:52
giles	albel727: my argument was:	06:52
	Well all this GitHub vs Gitorious vs something else is a bit pointless. One of the main features of git is that its de-centralized. You can host a clone of the repo on GitHub another on Gitorious another on Indefero.net and easily keep them all in sync easily using a cron job. Personal	06:53
	→ krosenvold_ joined	06:53
giles	Personally I think you would be bonkers not to host an "official" clone on your own server even if this is not the primary origin that people push to.	06:53
	→ sadmac joined	06:54
	→ dharrigan joined	06:54
	← dharrigan left	06:54
	→ dharrigan joined	06:54
albel727	giles: I see. well, how do you sync them with cron job, anyway? is there a repo on your server, which you push to both of these, or do developers use gitorious and github directly? if so, how do you automatically merge changes in cronjob, if someone pushes conflicting changes to them?	06:55
giles	I do it on my local machine yeah	06:55
	and good point.	06:55
albel727	giles: so, you do "pull github, pull gitorious, push github, push gitorious", and hope, that second pull wont fail? =)	06:56
	← Chillance left	06:56
giles	albel727: Well currently I am not commiting to gitorious	06:57
	its one of those "doh" moments	06:58
albel727	giles: hmm, I think your problems might go away, if you declare that gitorious repo is read only for everyone. then you can put a post-receive-hook in github, and gitorious would be always in sync.	06:59
giles	can you put a post-recieve-hook in github?	07:01
albel727	giles: I thought one can. But never tried, sorry.	07:02
doener	giles: http://help.github.com/post-receive-hooks/	07:02
	← sadmac left	07:03
giles	doener: ta	07:03
	→ sadmac2 joined	07:03
	sadmac2 → sadmac	07:03
giles	doener: I think this is limitied to sending http POST to a blog or something.	07:04
	→ OOPMan joined	07:06
	← giles left	07:06
	→ Sou\|cutter joined	07:08
albel727	giles: damn, I didn't occur to me, they might prohibit putting custom hooks (and for a reason, too). ok, then the best you can do, is write a small prog, which will receive that http ping and start syncing. I guess that and cron job (or running a poll among devs, which hosting they prefer, and keeping only that) are your only choices, sorry.	07:08
	← gilaniali left	07:08
	albel727 was late	07:08
	→ gilaniali joined	07:09
	→ tim_ joined	07:13
tim_	i have an unmerged path that is a directory for a submodule, what can I do?	07:13
	← yhager left	07:13
tim_	i've tried checkout --theirs, --ours	07:13
	to no avail	07:14
Sou\|cutter	I created a local branch, created a new remote branch based off of it, did some work, pushed again, did some more work, rebased with my local master no conflicts, and now my pushes to my remote branch are rejected	07:14
	← scarabx left	07:15
Sou\|cutter	any thoughts on what I did wrong and how I can untangle this?	07:15
	everything I've tried so far seems to create a conflict for every change I made	07:16
	← dharrigan left	07:16
	→ xiangfu joined	07:16
doener	tim_: make sure that there's no trailing slash on the path (tab-completion usually adds one)	07:17
k_sze	git-check-ref-format.html says: "2. They must contain at least one /. This enforces the presence of a category like heads/, tags/ etc. but the actual names are not restricted."	07:17
doener	Sou\|cutter: rebasing creates new commits, replacing the old ones you already pushed, so of course git complains	07:17
k_sze	What?! o_O	07:17
tim_	doener: you mean when I try to do a checkout --ours?	07:18
doener	tim_: Yup.	07:18
tim_	doener: there's not	07:18
doener	tim_: if that still doesn't work, use HEAD instead of --ours or MERGE_HEAD instead of --theirs	07:19
Sou\|cutter	doener: what you say makes sense in retrospect... I'm not sure how to fix it now, though. Should I create a new local branch off of master and cherry pick all my commits?	07:19
doener	Sou\|cutter: why rebase at all?	07:19
	Sou\|cutter: if it's a topic branch, just finish it and then merge it back into master, done	07:20
Sou\|cutter	doener: I forgot to mention that at some point in between I had been on master and pulled	07:20
tim_	doener: hallelujah, thanks!	07:20
	doener: git checkout HEAD FTW!	07:21
doener	Sou\|cutter: doesn't matter, for the problem at hand, that's not really different from having committed on master	07:21
	Sou\|cutter: (and it's obvious that master changed, as otherwise the rebase would have been a no-op ;-))	07:21
Sou\|cutter	doener: now when I try to merge back to master I'm getting conflicts for everything though.. I think. maybe I will double check, I'm kinda fried... been hacking on this project for like 16 hours	07:22
doener	Sou\|cutter: did you already reset the branch back to before the rebase?	07:22
Sou\|cutter	how do I do that?	07:23
doener	Sou\|cutter: checkout the branch, then run "git reset --hard <commit_the_branch_head_was_at_before_you_rebased>" (this drops all uncommitted changes!)	07:25
	Sou\|cutter: to find that commit, you might want to look at the reflog: git reflog show <branch_head>	07:26
	albel727 → albel727_away	07:27
	← gnufied left	07:27
	→ ijcd joined	07:28
Sou\|cutter	ok...	07:28
doener	I'll have to leave now... :-/	07:30
	← justin-george left	07:30
Sou\|cutter	doener: thanks for the help...	07:30
	← FabianB left	07:35
	→ hohoho joined	07:35
	→ giles joined	07:36
	→ d0k joined	07:37
	← hyperair left	07:38
	→ FabianB joined	07:39
	→ ifos joined	07:40
	← marc_kdab left	07:41
	comawhite → comawhite\|sleep	07:43
	→ frerich joined	07:44
	→ hyperair joined	07:46
	← thiago left	07:49
	← tim_ left	07:50
	→ rikarends joined	07:50
	← rikarends left	07:51
	→ dharrigan joined	07:52
	← dharrigan left	07:52
	→ dharrigan joined	07:52
	→ airborn joined	07:54
	← fda314925 left	07:55
	← iravid left	07:56
	albel727_away → albel727	07:56
	→ fda314925 joined	07:56
	← hohoho left	07:57
	← giles left	07:57
	→ rikarends joined	07:58
	← amerine left	07:58
	← rikarends left	07:59
	→ dsigurds joined	08:00
sl0ppy	git diff isn't showing me a diff of the files that are listed to be commited	08:01
	any idea how to get those included?	08:01
	→ iravid joined	08:01
mmattice	git diff --cached # will show just those	08:01
sl0ppy	oh ok cool thanks	08:02
	← ifos left	08:02
mmattice	I don't usually find showing all the diffs useful. showing what's staged and what's not is useful though	08:03
	→ yhager joined	08:03
	← dharrigan left	08:05
	→ pantsman joined	08:05
	← pantsman left	08:05
	→ pantsman joined	08:05
	← OOPMan left	08:06
	← hyperair left	08:11
	← towski left	08:14
	← yhager left	08:18
	→ gnufied joined	08:18
	→ ch077179 joined	08:19
	→ herdingcat joined	08:20
dsigurds	anyone know if gitosis supports namespacing repos by putting them in subdirectories?	08:20
	← gilaniali left	08:23
	← herdingcat left	08:23
	← gnufied left	08:25
	← shiba_yu36 left	08:28
sitaram	have you tried it and it didn't work? gitosis is a bit of "TIAS" (try it and see), no one actually supports it anymore	08:29
	← AlexP left	08:30
	→ priidu joined	08:32
	→ AlexP joined	08:33
	→ gebi joined	08:33
	← gebi left	08:33
	→ gebi joined	08:33
	→ boombatower joined	08:40
sl0ppy	how can i retract the last commit	08:41
	→ cas1 joined	08:41
	→ dels joined	08:42
	← cas1 left	08:42
albel727	sl0ppy: depends on whether you have pushed it already, or not, and whether you did already make some local changes in working tree on top of it, or not.	08:43
sl0ppy	actually	08:45
	it's not letting me merge	08:45
	it's saying local files will be overwritten	08:45
	but i want it to overwrite these files	08:45
albel727	sl0ppy: are these "local files" tracked?	08:46
	→ z1y joined	08:46
cbreak	they probably are not	08:50
	just delete them	08:50
albel727	sl0ppy: ^	08:51
sl0ppy	oh ok	08:51
	← Uierth left	08:51
sl0ppy	i can't do git-reset --hard?	08:52
	i have some commit's in the queue, but i can't push until i'm all merged	08:52
albel727	sl0ppy: git reset doesn't touch untracked files.	08:52
sl0ppy	oh	08:52
	oh these are under changed but not updated	08:52
sitaram	sl0ppy: the messge you got comes when (a) you have some local, UN-tracked files and (b) the branch you're merging in happens to want to create files with the same name	08:52
sl0ppy	Please, commit your changes or stash them before you can merge.	08:52
sitaram	sl0ppy: aah that's differen then; they are tracked files	08:53
albel727	sl0ppy: that explains it. you have changes, that you didn't do "git add" on.	08:53
sl0ppy	ah ok	08:53
	how can i get past this	08:53
sitaram	sl0ppy: if you really dont need those changes curently seen in "git status", I would do "git checkout -f" and then retry the merge	08:53
sl0ppy	i didnt change these files	08:53
	→ aspotashev joined	08:54
sl0ppy	i think someone else did though	08:54
	i pull'd down their changes	08:54
	actually	08:54
sitaram	what does	08:54
sl0ppy	ok i think i have it	08:54
albel727	sl0ppy: if they're under "changed but not updated", then it's you and only you who changed them.	08:54
sitaram	what does "git status" say? can you pastebin?	08:54
albel727	sl0ppy: or maybe you were playing around with non-hard git reset.	08:55
sl0ppy	i have a merge conflict	08:55
	can i just tell it to clobber my file?	08:55
	hrm	08:56
	Putr\|Away → Putr	08:56
sl0ppy	ok so i did a git pull, there's a merge conflict	08:56
	i just want to delete my local file	08:56
albel727	sl0ppy: "git checkout --theirs <file>"	08:56
sl0ppy	ok great thanks	08:57
	← madewokherd left	08:57
sl0ppy	wow i'm going in circles	08:57
z1y	Hi, I have a "commit" object and now I want to update the list of "parents" in that object. Which command can I use?	08:57
bob2	you'd need to make a new commit	08:58
sl0ppy	heh i did a pull again and those previous files i had to use checkout -f to override are back	08:58
bob2	what's the point?	08:58
sl0ppy	should i be using git rm?	08:59
	fatal: 'merge' is not possible because you have unmerged files.	08:59
_rane	is it possible to somehow symlink files from elsewhere to a git repo?	08:59
doener	z1y: that means history rewriting, filter-branch can do that, but why do you think you need that?	08:59
_rane	seems like git doesn't like symlinks	08:59
doener	_rane: no, git would track the symlink itself, you can only setup things the otherway around (symlink into the working tree)	08:59
	→ ph^ joined	09:00
doener	s/history rewriting/rewriting history/	09:00
_rane	I already have the stuff in my dropbox, and I'd just like to add that stuff into some git repo	09:00
z1y	doener: I want to learn how git works. I don't think that filter-branch works in my case ...	09:00
_rane	bleh	09:01
doener	z1y: why not?	09:02
sl0ppy	Please, fix them up in the work tree, and then use 'git add/rm <file>'	09:02
	anyone know what this means?	09:02
	i keep going in circles	09:02
_rane	fix something and then git add the fixed files	09:02
sl0ppy	i delete files, then i do a pull, same files come back	09:02
	i dont need to fix anything though	09:03
	i just want it to overwrite these files	09:03
_rane	need more context on what's happening	09:03
sl0ppy	i have 2 sets of commits	09:03
z1y	doener: because I want to update the list of parents for a single object. while "filter-branch" filters a branch not an object (?)	09:03
sl0ppy	i tried to push, and then it said i had to merge	09:03
	so i pull'd and there were conflicts	09:03
albel727	sl0ppy: what, you do you not know, how to resolve merge conflicts? you are supposed to a) fix conflicted file to your liking (in your case, that would be git checkout --theirs <file>) b) git add <file> c) after you're done with all conflicted files, do git commit. also read http://www.kernel.org/pub/software/scm/git/docs/user-manual.html#resolving-a-merge	09:03
sl0ppy	albel727: i already did my commits	09:04
z1y	doener: after the update, a new branch will be created, or broken	09:04
sl0ppy	also sorry i'm a little out of it, been up all night working and trying to commit this before i leave	09:04
	← nanotube left	09:04
bob2	z1y: yes, you'll be making a new branch	09:04
sl0ppy	albel727: i did my commits first	09:04
albel727	sl0ppy: for christ sake. now you're doing a "merge commit" - the commit, which fixes discrepancies between your and everybody's else work.	09:04
sl0ppy	yeah exactly	09:05
	i just want it to clobber these files on my local branch	09:05
doener	z1y: hm, indeed, only works for refs, not arbitrary commits	09:05
sl0ppy	i did git checkout -f	09:05
	but i keep ending up back in the same place	09:05
	→ urkud_ joined	09:05
	← noktoborus_ left	09:06
	→ psoo joined	09:07
albel727	sl0ppy: "clobbering files" is a change as any other in its own right. you're supposed to do a separate commit with that change. "git checkout -f" won't help you for obvious reasons.	09:07
doener	z1y: but anyway, commits (like all objects) are immutable, so you cannot change the parents of an existing commit object, you can only create a new one that differs from the original one only in its parents	09:07
	sl0ppy: git checkout --theirs -- <path>; check results; git add <path>	09:07
	→ nanotube joined	09:08
doener	sl0ppy: will tell git that the correct way to resolve the conflict is to use "their" version of the file	09:08
sl0ppy	doener: ok cool thanks, i'm also on reading some stuff on gitready	09:08
z1y	doener: ah, I see... Creating new commit... Sounds a challenge to me :D	09:08
albel727	doener: sl0ppy: and then git commit	09:08
sl0ppy	its weird, all these changes files i didnt edit at all	09:10
	am i recomitting someone's changes?	09:10
	→ noktoborus_ joined	09:10
	← urkud_ left	09:11
	→ roop joined	09:12
sl0ppy	ok got it!	09:13
	thanks for the help	09:13
	→ dr_win joined	09:14
	← babusri left	09:22
	→ project2501b joined	09:27
	← bdiu left	09:28
	→ Anti-X joined	09:31
	→ OOPMan joined	09:31
	← noktoborus_ left	09:34
	→ noktoborus_ joined	09:36
	→ nevyn^_ joined	09:37
	← airborn left	09:37
	→ parasti joined	09:38
	→ Schmallon joined	09:38
	← rpg left	09:38
	→ deepakk joined	09:40
	→ workmad3 joined	09:42
	deepakk → dkannan	09:43
	→ rpg joined	09:44
	← frerich left	09:50
	→ VoiDeT joined	09:53
VoiDeT	Where is the best place to get help with gitosis?	09:53
albel727	VoiDeT: nowhere. gitosis is no longer supported by anyone. you can try and ask here, ofc.	09:54
	(advise: migrate to gitolite)	09:55
	s/advise/advice/	09:55
	albel727 shoots himself	09:55
cbreak	luckily you miss to hit the trigger	09:55
VoiDeT	awesome thanks albel727	09:56
albel727	that's what I have a cyanide bottle for =)	09:56
sitaram	yes but your bullet hit the bottel just now, and it's all spilled!	09:56
	bottle	09:56
	hey can I borrow it; looks like I need it too!	09:56
VoiDeT	what other ones are out there that i should also check out?	09:56
	→ urkud_ joined	09:56
albel727	but cyanide is just as poisonous in air, as in drink, so...	09:56
gebi	VoiDeT: gitolite	09:56
	→ goshawk joined	09:57
sitaram	VoiDeT: other than gitolite? the others are all web-based. Gerrit is the best among them	09:57
	VoiDeT: gerrit is not just an access control tool; it will dictate your workflow -- it's main purpose is code review	09:57
VoiDeT	interesting	09:57
	← Wyverald left	09:58
	← boombatower left	09:58
	albel727 finally dies in convulsions from the toxic cyanide fumes.	09:59
VoiDeT	so im guessing you use gitolite albel727	09:59
	oh, too late, you died	09:59
sitaram	no he didn't; if he is using gitolite he can be resurrected :)	09:59
	from a mirror that he should have setup earlier	10:00
	→ boombatower joined	10:00
	albel727 gets resurrected by gitolite...	10:00
sitaram	:)	10:00
VoiDeT	hahaha	10:00
albel727	VoiDeT: well yeah, I do. but you'll be better off in asking sitaram (he's the author of gitolite =D)	10:00
VoiDeT	oh nice	10:00
sitaram	VoiDeT: the rest of them are not that good. Gitorious is hard to install, I'm told (I tried once, long before I wrote gitolite, failed). Gerrit I have never tried to install -- dont need code review and dont want my workflow changed	10:00
VoiDeT	Well gitosis was fine, until its post-update hooks broke now >_<	10:01
	albel727 hates sitaram. now all those wonderful toxic fumes got evaporated for nothing. =)	10:01
sitaram	albel727: actually, I very rarely tell people "gitolite is the best, use it". There's enough people doing that anyway ;-)	10:01
VoiDeT	well i	10:01
	→ _Unode joined	10:01
VoiDeT	well i'll look at getting gitolite up	10:01
bremner	huh. monotone made a release.	10:01
	sitaram thinks toxic fumes go well with rum and coke	10:01
sitaram	bremner: yeah but who wants to be stuck with one wife all ones life	10:01
	oh wait	10:01
	← dkannan left	10:01
VoiDeT	sounds like it can be installed on osx too	10:02
sitaram	VoiDeT: frogon_wheels (intentionally not highlighting) has even installed it on Windows :)	10:02
VoiDeT	snap	10:03
	sitaram assume VoiDeT is talking about gitolite, not monotone	10:03
VoiDeT	yep	10:03
	← k_sze left	10:03
sitaram	VoiDeT: but quickly though, let's see, for academic curiosity, what went wrong with your gitosis... did you upgrade python or something on the server lately?	10:04
	VoiDeT overwhelmed by branches atm	10:04
	← Veejay left	10:04
	→ corni joined	10:04
	← corni left	10:04
	→ corni joined	10:04
VoiDeT	i updgraded my ubuntu dist, which upgraded python setup tools, which then led to no where	10:04
	and now i can't seem to get the post-update hook triggering, and cbf with it	10:04
	→ chrischambers joined	10:04
VoiDeT	what branch do i go with sitaram, pu?	10:05
sitaram	VoiDeT: yes	10:05
	→ albel727_2 joined	10:05
sitaram	VoiDeT: I always prefer people use pu, ever since I created a test suite I have much more confidence in "pu", and master is really not that far behind anyway	10:05
	← albel727 left	10:06
sitaram	cbf?	10:06
VoiDeT	cbf = cant be farqd	10:06
albel727_2	(oh my. it is possible to install gitolite on windows? since when does windows have an adequate sshd?)	10:06
sitaram	VoiDeT: I do ask you spend some time reading doc/1 and doc/2 (install and admin docs)	10:06
	albel727_2: cygwin	10:06
	albel727_2 → albel727	10:06
VoiDeT	alrighty, will do	10:06
cbreak	step1: install vmware. step2: install linux :)	10:07
albel727	well, if only cygwin....	10:07
sitaram	VoiDeT: links off of http://github.com/sitaramc/gitolite/tree/pu/doc/ -- online docs, auto-rendered by github	10:07
	albel727 doesn't consider cygwin adequate	10:07
	→ hyperair joined	10:07
albel727	cbreak: lol	10:07
sitaram	albel727: neither do I. I only mentioned that because VoiDeT said it seems to work on OSX	10:07
	← mlq left	10:08
	→ mlq joined	10:08
	← mlq left	10:08
	→ mlq joined	10:08
VoiDeT	already there sitaram :)	10:08
	sitaram won't actually support use on Windows -- not as far as Windows specific issues are concerned	10:08
sitaram	VoiDeT: :)	10:08
cbreak	os x is a unix more so than most other platforms.	10:08
	windows on the other hand is it's own little species	10:08
sitaram	VoiDeT: theres also a migration document for gitosis migration; same link	10:09
cbreak	it always needs an "Extrawurst"	10:09
VoiDeT	well done sitaram	10:09
sitaram	VoiDeT: I've had a bit of practice at this now; been a year and 2 months	10:10
	:)	10:10
	→ icwiener joined	10:10
VoiDeT	there is a ton of config here, that im just going to close my eyes at for the time being	10:10
	i just need a place to push to for the time being	10:10
cbreak	make a local bare :)	10:10
sitaram	VoiDeT: defaults are fine dont worry	10:10
cbreak	once you scp it to the server, it's a full repo	10:10
VoiDeT	so its installed	10:10
sitaram	VoiDeT: which method did you use?	10:11
VoiDeT	now i dump some trusted keys and edit some configs	10:11
cbreak	gitolite is useless if all you want is a git repo to push to	10:11
VoiDeT	root	10:11
sitaram	ok	10:11
cbreak	it's useful to _restrict_ access to the repo	10:11
sitaram	cbreak: he was using gitosis, so I assumed he had some ACLs; but you're right -- may be overkill	10:11
VoiDeT	sitaram: so can i now clone down the config/key dir from gitolite like gitosis and admin from there?	10:12
sitaram	VoiDeT: are you sure you need this? if all you need is a place to push to, (one user), gitolite is overkill	10:12
	→ Wyverald joined	10:12
VoiDeT	well just for now, i will have more users in a few hours once its back up D:	10:12
sitaram	VoiDeT: not exactly... you do need to do some manual stuff; it's all explained in detail in the migrate doc	10:12
	VoiDeT: very detailed...	10:12
	sitaram writes more docs than code	10:13
VoiDeT	oki doki	10:13
albel727	lol	10:13
	albel727 wishes sitaram good luck	10:13
	sitaram wonders why he* needs luck and not VoiDeT :-)*	10:14
VoiDeT	ha	10:14
sitaram	just kidding... it'll all work. For people who don't skip reading the docs ;-)	10:14
albel727	because in the end, you'll be the one who will need to explain everything to VoiDeT (and everybody else) =D	10:14
VoiDeT	well	10:14
	the docs are pretty awesome albel727	10:15
albel727	VoiDeT: sure they are =)	10:15
VoiDeT	if i cant get it working with these docs, i think i should talk to a GP not sitaram	10:15
cbreak	grand parent?	10:15
sitaram	VoiDeT: I misunderstood your original question. I answered in terms of "what is required to complete the migration". Just to start using gitolite as it, what you said is fine -- just "clone down the admin repo" and so on...	10:15
	s/as it/as is	10:15
VoiDeT	ok awesome	10:16
sitaram	albel727: part of the reason for my documentation fever is my time zone; most of the time I'm not on when people ask	10:17
albel727	cbreak: general practitioner, probably	10:17
	← infinity0 left	10:17
	→ infinity0 joined	10:17
bremner	git punisher	10:17
VoiDeT	well, this leaves gitosis for dead	10:17
	is all i can say :\|	10:17
sitaram	:)	10:17
VoiDeT	looking at the features	10:18
sitaram	VoiDeT: I hate to preen but it left gitolite for dead 2 days after it first aired	10:18
VoiDeT	rip gitosis	10:18
albel727	sitaram: (btw, which tz is it? india?) surely that's only a very small part of the reason. I guess, even if you were online, you'd be torn apart by constant requests for explanation.	10:18
sitaram	albel727: yes, India	10:18
cbreak	sitaram: have you considered documenting via A.L.I.C.E.	10:19
sitaram	albel727: no it's not been that bad. Not since I stopped worrying about ssh problems ;-)	10:19
albel727	lol	10:19
sitaram	cbreak: what is A.L.I.C.E? Some AI like Elisa?	10:19
cbreak	much more advanced	10:19
	you can modify it via xml I think	10:19
sitaram	albel727: also there are a couple of regulars (were_Hamster, Il_ari) who help out a lot too	10:20
albel727	sitaram: yeah, they sure do.	10:20
sitaram	cbreak: not this: http://www.alice.org/ I assume	10:21
	← VoiDeT left	10:21
albel727	sitaram: probably this: http://alice.pandorabots.com/	10:21
cbreak	http://alice.pandorabots.com/	10:21
	indeed	10:21
	→ VoiDeT joined	10:21
sitaram	yeah I got to http://www.alicebot.org/about.html	10:21
albel727	(now, who didn't dream about talking with fake captain kirk? =D)	10:22
VoiDeT	checking out my authorized_keys file now, i see "/usr/local/bin/gl-auth-command"	10:22
	is that right for it to be and absolute path?	10:22
	→ _ollie joined	10:22
	→ tvw joined	10:22
sitaram	VoiDeT: yes	10:23
	← OOPMan left	10:23
	← aspotashev left	10:23
sitaram	VoiDeT: avoids all sorts of problems, actually	10:23
	see GIT_PATH in conf/example.gitolite.rc for why :-)	10:24
VoiDeT	ah k	10:24
	nah :D	10:24
albel727	methinks, that most of the problems, that people can't solve themselves with documentation, will require more insight, than can be provided by a bot.	10:24
bremner	albel727: is because of your mother that you think that?	10:25
sitaram	VoiDeT: may not be a problem in your case, but that code is generalised so it looks like you dont need it	10:25
albel727	bremner: I beg your pardon? what was that about my mother now?	10:25
sitaram	bremner: LOL! Literally!	10:25
	← hyperair left	10:25
	← roop left	10:26
sitaram	albel727: Why do you suppose someone mentioned your mother?	10:26
bremner	albel727: was a joke reference to a famous bot, no disrespect to your family	10:26
	sitaram loves AI pseudo-Freuds!	10:26
sitaram	bremner: I'm sure he knew and was egging you on :)	10:26
	← z1y left	10:27
albel727	sitaram: bremner: oh, that explains it. =D which one was that famous bot, so I won't stay ignorant further?	10:27
bremner	sitaram: perhaps, but I have my serious personality overlay in this channel	10:27
	albel727: eliza	10:27
albel727	ah, automated terapist, I see =)	10:27
sitaram	albel727: just to be clear, my response to your response was also what I suspect Eliza would have said, if you happened to be talking to two of them at the same time :-)	10:28
cbreak	eliza is very simple	10:29
	it usually answers with some question	10:29
sitaram	like my wife actually	10:29
albel727	lol	10:29
	albel727 is finally able to get all the humor in those jokes	10:29
sitaram	"why do you think we dont need a new refrigerator?"	10:29
VoiDeT	fatal: 'gitolite-admin' does not appear to be a git repository - but i think this might be some error due to me still having some gitosis stuff around the place	10:30
sitaram	VoiDeT: wrong key	10:30
	← _ollie left	10:30
sitaram	VoiDeT: you're probably using a key that gets you a shell, or the key you used in gl-setup was already in the authorized_keys file as a "shell" key	10:30
	→ _ollie joined	10:31
VoiDeT	hmmm	10:31
sitaram	VoiDeT: FMTYEWTK about this: http://github.com/sitaramc/gitolite/blob/pu/doc/ssh-troubleshooting.mkd	10:31
	← infinity0 left	10:31
VoiDeT	what i did was get my git user ssh-keygen, and used that for gl-setup	10:31
	albel727 googles fmtyewtk	10:32
	→ infinity0 joined	10:32
	← infinity0 left	10:32
	→ infinity0 joined	10:32
	→ yuvankumar joined	10:32
	albel727 successfully googled fmtyewtk, and now lols non-stop	10:33
sitaram	VoiDeT: ssh-keygen on the server? or your workstation	10:33
VoiDeT	server	10:33
sitaram	from which userid? the same one? it's weird but it should work to clone from there, if that is what you did	10:33
	in fact gitolite can run all from one userid, as long as you're ssh-ing to yourself of course	10:34
VoiDeT	well this was under my git user on the server	10:34
	i ran the ssh-keygen, then gl-setup with that public key	10:34
	but im guessing i have to add in my clients key?	10:34
	or can i just copy the public key from the server to the client	10:35
	→ dr_win_ joined	10:35
sitaram	VoiDeT: wait wait... you'd have to copy the private key also if you want to do it that way	10:35
	VoiDeT: guess you'll need to understand a little about ssh based on that question: try http://github.com/sitaramc/gitolite/blob/pu/doc/gitolite-and-ssh.mkd	10:36
VoiDeT	ha	10:36
sitaram	albel727: actually, that is FMTYEWTK :)	10:36
	sitaram has recieved lots of compliments for that* doc specifically, more than others*	10:36
	albel727 understands why =D	10:36
sitaram	something about the flow of it I guess; I'm rather pleased with how it turned out	10:37
	← dr_win left	10:38
VoiDeT	argh	10:38
	sorry	10:38
	totally misread	10:38
	← tango_ left	10:40
sitaram	VoiDeT: dont feel too bad; ssh is a pain. Accounts for 75% of my support load, despite being barely 10% of the code and also being 95% similar to how any such tool (gitosis, gitorious, etc) do it	10:40
VoiDeT	no no, i understand how authorized keys etc work, just that i misread it and was a robot	10:41
	albel727 didn't have any problems with ssh at all, since he read gitolite-and-ssh.	10:41
sitaram	albel727: :)	10:41
albel727	s/since/ever since/	10:41
	→ Yuuhi joined	10:41
VoiDeT	now im getting the following users (pubkey files in parens) do not appear in the config file: imac(imac.pub)	10:42
	so is this as simple as adding in imac as an admin of gitolite repo?	10:42
	→ p3rg2z joined	10:42
	→ dwmw2 joined	10:42
	→ intripoon joined	10:43
albel727	(no, I'm telling the truth. It made me understand much more about ssh, than I ever expected to get from such a short doc)	10:43
sitaram	VoiDeT: yes indeed	10:44
	→ tango_ joined	10:44
VoiDeT	this has to be done serverside sitaram, because i havent made a clone yet	10:44
sitaram	VoiDeT: even there -- do an actual clone via ssh, despite it loops back to the same user	10:45
	← dels left	10:45
VoiDeT	alrighty	10:45
sitaram	VoiDeT: otherwise there are other methods but this is the cleanest	10:45
	VoiDeT: let me explain: a normal file:/// based clone will fail when you push because the update hook is expecting gitolite info that is not present. Then you have to bypass that. Can be done (there's an env var that lets you do that ;-) but it's simpler to do it the canonical way	10:46
VoiDeT	bah	10:47
	well if i didnt stuff up my install, i bet i would be done and sleeping by now	10:48
	but: warning: You appear to have cloned an empty repository.	10:48
	at least i can clone something now :D	10:48
sitaram	VoiDeT: should not have happened to the admin repo... that's bad	10:49
VoiDeT	argh	10:49
	I may as well just start over? :D	10:49
sitaram	VoiDeT: well, you can, but if you're talking about sleep go ahead and we'll do it tomorrow. It is possible to recover from pretty much anything, but its your choice	10:50
VoiDeT	oh no no, sleep means giving up	10:50
sitaram	if you want to start over, be sure to read: http://github.com/sitaramc/gitolite/blob/pu/doc/1-INSTALL.mkd#_cleaning_out_a_botched_install	10:50
	sitaram is full of links :)	10:50
VoiDeT	hahahaha nice	10:50
sitaram	VoiDeT: thats only the "robot" way. As I said, it's possible to recover from pretty much anything. E xcept an empty admin repo -- not sure how you got that	10:51
	can you paste your install session, if that window is still up?	10:51
	(the one where you ran "gl-setup)	10:52
	(the one where you ran "gl-setup")	10:52
VoiDeT	mmmm	10:52
	not really :(	10:52
	→ jspiros joined	10:52
VoiDeT	ill try again	10:52
sitaram	VoiDeT: wait wait... when you added imac.pub you got a message saying something	10:52
VoiDeT	yep	10:53
sitaram	VoiDeT: that means the admin repo was ok at that time; you would not have have got that if it wasn't	10:53
VoiDeT	well i dunno what happend :(	10:53
sitaram	oh well, start again and keep a transcript to paste for me in case I spot something	10:53
	what does "ssh git@server info" say?	10:54
VoiDeT	done	10:54
	got my clone on	10:54
	smooth sailing from here i am guessing	10:54
	← yuvankumar left	10:54
sitaram	my clone == gitolite-admin clones from server to workstation? and has actual data (like conf/gitolite.conf). Then yes	10:54
VoiDeT	:)	10:55
	its working sitaram	10:55
sitaram	you can grab stuff from your gitoSIS-admin clone, follwo the migrate doc to convert the config file, copy keys over (but be careful if you have files with @ in the name; see doc), and so on... done	10:55
	follow*	10:55
VoiDeT	so do i have to add it in via repos? or can i do it via users	10:56
sitaram	add what in?	10:56
VoiDeT	say i want user1 to own 100 repos, can i just list a group of users, and a group of repos for all of those users	10:56
	like gitosis	10:56
sitaram	VoiDeT: the syntax is totally different but yes you can. A lot more actually	10:57
	→ intripoon_ joined	10:57
	sitaram thinks gitosis syntax is Windows INI style; fine for disparate pieces of info like in ~/.gitconfig, but not for more homogenous stuff like ACLs	10:57
sitaram	VoiDeT: just use the conversion program to convert your gitosis conf file; dont do it manually	10:58
	if its long	10:58
VoiDeT	ooo	10:58
	better acl	10:58
sitaram	yup	10:59
VoiDeT	looks good	10:59
	but what if i want basic, instead of repo - then all acl. Can i just have a list of users, with a particular set of permissions, then all repos for the users with those permissions	11:00
	for example, i have 10 admins, who can RW on everything, i just want to list them, and all the repos in the system	11:00
sitaram	@admins = alice bob eve <newline> repo @all <newline> RW = @admins	11:01
VoiDeT	:\|	11:01
	VoiDeT 's brain exploded	11:01
VoiDeT	thats awesome	11:01
	→ kipras joined	11:01
sitaram	you can use @all for users also, but repo "@all <newline> RW = @all" makes gitolite sort of pointless	11:01
	RW+ if you want rewind capability (force push)	11:01
VoiDeT	so http://gist.github.com/655203 is fine	11:03
sitaram	awesome? you aint seen nothing. Try http://github.com/sitaramc/gitolite/blob/pu/doc/3-faq-tips-etc.mkd#_simpler_syntax	11:03
	yes its fine	11:04
	→ Veejay joined	11:04
	← p3rg2z left	11:04
	← intripoon_ left	11:06
VoiDeT	muhahaha	11:06
	its working fine	11:06
sitaram	:0	11:06
albel727	what's so funny about it? =)	11:06
VoiDeT	its too easy :D	11:06
sitaram	did you check out the example I linked just above (talking of brain exploding)?	11:06
VoiDeT	installation would of been a breeze if i had a coffee before hand too	11:07
sitaram	now you see why I laugh when people say "svn has ACLs, and git doesn't	11:07
	""	11:07
VoiDeT	sitaram: thats exactly what i was after	11:07
	→ shahid_ joined	11:07
	sitaram is curently in a big fight with some svn-heads on -- of all places -- linkedin!	11:07
VoiDeT	so this is better? http://gist.github.com/655203	11:08
sitaram	VoiDeT: you're missing a "repo @personal" line just before the last one	11:09
	VoiDeT: your current line 5 merely defines a macro; doesnt use it	11:09
jim	shoot em the url of Linus' google talk and gitcasts... and say good nite :)	11:09
	← psoo left	11:09
VoiDeT	the svn sucks talk?	11:10
jim	I think so, and cvs sucks more	11:10
albel727	jim: linus's google talk?	11:10
VoiDeT	sorry sitaram, i dont follow you	11:10
albel727	VoiDeT: "RW =" line requires "repo <somerepo>" line before it	11:10
sitaram	VoiDeT: ok let me parse that out for you. A line like "@foo = bar baz" is like a #define in C. (Sort of, not 100% but for now that will do)	11:11
jim	VoiDeT, it's kinda interesting...	11:11
	→ MantiS joined	11:11
sitaram	VoiDeT: so you haven't defined what repo you're specifying RW permissions for; you have to do that with a "repo @personal" or osme such line	11:11
jim	he calls himself a git a lot, and says large groups of programs and people are stupid :)	11:11
VoiDeT	ahhh of course	11:11
	so more like http://gist.github.com/655203	11:12
	→ rikarends joined	11:12
	← MantiS left	11:12
albel727	yeah, that's more like it	11:12
	→ intripoon_ joined	11:12
VoiDeT	so good	11:12
jim	this git from the bottom up thing is interesting too tho	11:12
albel727	jim: progit is no worse	11:13
	← iravid left	11:13
jim	what's progit?	11:13
albel727	jim: http://progit.org/book/	11:13
sitaram	VoiDeT: works	11:14
	→ zorzar_ joined	11:14
	→ iravid joined	11:14
VoiDeT	works amazing	11:14
	what is gitosis? i forgot :D	11:14
jim	hali's sister :)	11:14
albel727	jim: and progit talks about the same things as "git from the bottom up", in chapter 9 "internals"	11:15
VoiDeT	thanks so much albel727 and sitaram	11:15
	← intripoon left	11:16
sitaram	VoiDeT: you're welcome!	11:16
jim	in what order?	11:16
	dwmw2 → dwmw2_gone	11:16
albel727	VoiDeT: it's more like thanks to sitaram and sitaram again (for writing gitosis), but heh =)	11:16
sitaram	alphabetical :)	11:16
albel727	s/gitosis/gitolite	11:16
sitaram	albel727: ouch	11:16
	:)	11:16
	albel727 is sorry	11:16
VoiDeT	:D	11:16
	hahahahahaha	11:16
	albel727 shoots himself again	11:17
	← zorzar left	11:17
	→ intripoon joined	11:17
akheron	I've never tried gitolite, maybe I should	11:17
	gitosis has worked well	11:17
albel727	why wait till it'll break? =) break it now and try gitolite =P	11:18
	← Mage left	11:18
sitaram	:)	11:18
	akheron: jokes apart, I used to tell people "dont upgrade unless your gitosis is broken". But lately I have heard more than a few people say that a python or python-setuptools upgrade has broken gitosis	11:19
akheron	oh?	11:19
albel727	akheron: (but I guess, sitaram had enough explanation for today, so maybe try it tomorrow =D)	11:19
sitaram	akheron: so yeah I agree with albel727 now	11:19
	dwmw2_gone → dwmw2	11:19
akheron	actually, tv used to work in the same company as I	11:19
	he has a bad habit of starting projects and leaving them unmaintained... :(	11:20
	← pantsman left	11:20
albel727	take your revenge on him - migrate to gitolite!! =D	11:20
	dwmw2 → dwmw2_LHR	11:20
VoiDeT	it'd be funny if github used gitolite :)	11:21
	← intripoon_ left	11:21
albel727	(not sure that it would be much of a revenge, though. a weak insult, probably =D)	11:21
sitaram	akheron: dont blame him. If he had actually repsponded to my increasingly desperate emails there would not be a gitolite today :)	11:21
akheron	he also has a bad habit of not responding to anyone's emails :D	11:21
sitaram	VoiDeT: no theirs is all ruby on rails, a custom ssh daemon, and so on. There's a presentation on it somewhere	11:21
	akheron: I'm glad it wasnt personal then ;-)	11:22
akheron	no, I believe it wasn't	11:22
sitaram	I'm sure it wasn't; I was joking :)	11:22
akheron	tv doesn't even respond to emails of our CEO, and tv owns 1/4 of the company!	11:23
sitaram	oh wow...	11:23
akheron	not a big business though, but anyway	11:23
cbreak	you sure he knows how to use an e-mail client?	11:23
akheron	:)	11:23
	→ bluenovember joined	11:23
	← rikarends left	11:23
sitaram	would it be impolitic to ask what company? (feel free to ignore)	11:23
akheron	http://www.inoi.fi/	11:23
sitaram	aah	11:24
akheron	tv left to the US a few years ago	11:24
cbreak	they arrested him for terrorism! that's why he never replies!	11:24
akheron	I don't know what he does there, something related to Big Web Things	11:24
	← shahid_ left	11:24
cbreak	I heard americans do that all the time with computer professionals	11:24
akheron	cbreak: he tweets, maybe you can do that from the jail?	11:25
cbreak	are those tweet thing the same pseudo random noise as SMS?	11:25
sitaram	cbreak: only security professionals ;-) and only if they crack a "big corporation"s allegedly secure PDF viewer :)	11:25
	← nevyn^_ left	11:30
	→ bitkiller joined	11:39
	→ rikarends joined	11:40
	← Mage_Dude left	11:40
	→ shahid_ joined	11:40
	← rikarends left	11:43
	→ Mage_Dude joined	11:46
	→ Dave^\| joined	11:48
	→ dharrigan joined	11:51
	← dharrigan left	11:51
	→ dharrigan joined	11:51
	→ Dave^\|\| joined	11:58
	← dharrigan left	11:59
	→ aspotashev joined	11:59
	→ leen^\| joined	12:00
	← Dave^\| left	12:01
	← Dave^\|\| left	12:01
	← infinity0 left	12:04
	→ infinity0 joined	12:05
	← mithro left	12:07
	← thomas_b left	12:09
	← nanotube left	12:10
	→ thomas_b joined	12:10
	→ ShadeHawk joined	12:10
ShadeHawk	How to mark that code and ideas were originally by some other author, but were rewritten by me (what signoff like annotation to use)?	12:11
	← boombatower left	12:12
wereHamster	ShadeHawk: you put the original author into the commit message	12:13
	→ boombatower joined	12:14
	← boombatower left	12:14
	→ boombatower joined	12:14
ShadeHawk	wereHamster: but the feature were, like, totally rewritten. Not one shred of code remains as is, and one commit (or two) exploded into long patch series	12:14
wereHamster	Originally-by:	12:15
	→ xiong joined	12:15
wereHamster	Inspired-by:	12:15
	Suggested-by:	12:15
	ShadeHawk: run git log \| grep -i "by: "\|grep -vi -E '(Signed-off-by\|acked-by\|tested-by\|noticed-by)' inside git.git or the linux kernel	12:15
	← boombatower left	12:15
ShadeHawk	git log \| grep -i "by: "\|grep -vi -E '(Signed-off-by\|acked-by\|tested-by\|noticed-by)' \| sed -e 's/:.*$//' \| sort \| uniq	12:17
wereHamster	see, you can do it yourself ;)	12:18
	← project2501b left	12:19
	→ mithro joined	12:20
	→ babusri joined	12:23
	→ project2501b joined	12:24
	→ vector_xyz joined	12:29
ShadeHawk	though I don't have kernel repository, so I could test it only on git repository	12:29
vector_xyz	hey guys i have a developer who cloned our master branch... so he has a copy of it now on his pc... he made changes... how can i get those changes now from him so i can review it before merging it or so... if he does git push... he will push it already to our master but i want it to be reviewed... and then added	12:30
	so im not sure which commands i need for that...	12:31
	on github its simple because you can do it all through the web... but through command line i got no idea	12:31
	→ dfr\|mac joined	12:32
wereHamster	vector_xyz: he could push into a different branch than master	12:32
	or push into a separate repo. Or send you the patches via email	12:32
vector_xyz	so each dev could have his own branch and that would workout fine ?	12:32
wereHamster	yes. But usually each dev has his own repo	12:33
	← infinity0 left	12:33
vector_xyz	i see were using a paid service so each repo sadly is another 10-20$ :x so i guess branch could be an alternative	12:33
drizzd	ShadeHawk: for the kernel http://fpaste.org/0Q2g/	12:34
wereHamster	vector_xyz: sure, if you pay for it then it could get expensive. But setting up a gitolite repo is fairly easy	12:34
	→ shiba_yu36 joined	12:35
vector_xyz	ah i see so i could just host it ourselfs :)	12:35
ShadeHawk	vector_xyz: can you access his PC and repository via SSH? if not, but you can access his PC, can he configure git-daemon? if not, let him push to separate branch in central repository "git push HEAD:for-you"	12:35
	drizzd: thanks!	12:36
	← thiana left	12:37
	← dfr\|mac left	12:38
	← project2501b left	12:40
	← flaguy48 left	12:42
	→ nicoulaj joined	12:44
	→ project2501b joined	12:45
	→ nanotube joined	12:46
wereHamster	lol, laughed-at-by:	12:47
	or signed-off-and-morning-tea-spilled-by:	12:48
	← robgleeson left	12:48
	→ murr4y joined	12:48
wereHamster	umm-i-can-live-with-that-by:	12:48
	→ dfr\|mac joined	12:48
	→ robgleeson joined	12:48
	→ Chillance joined	12:49
	→ igorw joined	12:50
	Moo\|beernothanks → Moo\|sprintingUK	12:51
	← Wyverald left	12:53
	← project2501b left	12:54
	← antgel left	12:54
	→ goldbergus joined	12:54
bremner	stared-at-in-horror-by:	12:55
	← aspotashev left	12:55
albel727	haha	12:56
	→ project2501b joined	12:58
wereHamster	nose-poked-at-pile-o-crap-by:	13:02
	← shiba_yu36 left	13:04
	→ alberto56 joined	13:05
	→ dark joined	13:05
	← mithro left	13:06
albel727	hehe. are we having "think up the most ridiculous analogue of reviewed-by:" holiday today, or what? =D	13:07
	← foocraft left	13:08
dark	git show HEAD^:file shows it with less. is it configurable? (I would like to show it in emacs, using emacsclient)	13:08
albel727	dark: yeah, it is	13:08
	GIT_PAGER environment option, and something about it in man git-config	13:08
Gitbot	albel727: the git-config manpage can be found at http://git.or.cz/man/git-config	13:08
dark	ok, now it is more like a problem on how make emacsclient accept something from stdin..	13:09
	but I see I can do git show .. > /tmp/file	13:09
	dwmw2_LHR → dwmw2_gone	13:11
	→ foocraft joined	13:11
	→ aspotashev joined	13:12
	→ tizbac joined	13:12
	→ shiba_yu36 joined	13:14
	← aresnick left	13:15
	→ Bass10 joined	13:17
	→ converge joined	13:18
	← goldbergus left	13:18
	→ mithro joined	13:19
converge	hi, i have a remote repository 'master', to update my local repos. with the remote repos i did: git checkout master; git pull master, but git pull master looks wrong..	13:19
	somebody can help me?	13:19
	← tungtn left	13:21
	← shiba_yu36 left	13:22
	→ OOPMan joined	13:23
	← kipras left	13:23
	← frakturfreak left	13:24
	→ goldbergus joined	13:24
wereHamster	converge: you didn't explain your problem	13:28
	until you do that, nobody can help you	13:28
	← jtauber left	13:29
converge	ok, i find a problem now, will fix it first	13:30
	im still kind of lost with git	13:30
wereHamster	converge: http://progit.org	13:31
converge	i was taking a look on this book, i will buy it	13:32
	im changing from svn, there is a lot of differences	13:32
	→ airborn joined	13:33
	← plopix left	13:35
	→ plopix joined	13:35
	← rson left	13:46
	→ icwiener_ joined	13:46
	← robgleeson left	13:47
	← icwiener left	13:47
	→ thiana joined	13:52
	← thiana left	13:52
	→ Thell joined	13:53
	→ thiana joined	13:53
	→ robgleeson joined	13:54
	→ Thell_ joined	13:54
	← EricInBNE left	13:54
	→ hilem joined	13:55
	→ Thell\|busy joined	13:55
	dfr\|mac → dfr\|bus	13:55
	→ Thell\|b3 joined	13:56
	← rpg left	13:56
	← Thell\|Away left	13:56
	→ Thell\|b4 joined	13:56
	← Thell left	13:57
	→ Thell\|b5 joined	13:57
	→ Thell joined	13:58
	← Thell_ left	13:59
	→ Thell_ joined	13:59
	← Thell\|busy left	13:59
	→ Thell\|busy joined	14:00
	← Thell\|b3 left	14:01
	← Thell\|b4 left	14:01
	← rvsjoen left	14:01
	← Thell\|b5 left	14:02
	← Thell left	14:02
	→ rvsjoen joined	14:03
	← Thell_ left	14:03
	← goldbergus left	14:04
	← sftp left	14:05
	← foocraft left	14:07
	→ goldbergus joined	14:09
	→ foocraft joined	14:10
	→ josh0x0 joined	14:12
	← goldbergus left	14:14
	→ sftp joined	14:14
	→ brothers joined	14:15
	→ theoros joined	14:18
	→ jtauber joined	14:18
	← bauruine left	14:19
	→ goldbergus joined	14:20
	← kpreid left	14:20
	← Bombe left	14:21
	→ bentob0x joined	14:21
	→ kpreid joined	14:21
	← airborn left	14:21
	→ Bombe joined	14:21
	→ airborn joined	14:22
	→ masonkramer joined	14:26
igorw	converge: you can also read it online for free	14:26
	→ virgil joined	14:27
impulze	is there a guy that can manage git-reset -p instead of git gui which does git-add -p?	14:27
	→ artefon joined	14:27
	→ bauruine joined	14:28
	← project2501b left	14:28
impulze	*gui	14:28
	;)	14:28
	← sftp left	14:30
	→ rgr joined	14:31
	← hilem left	14:32
	→ rikarends joined	14:33
dark	it's awesome that there are publishers that accept to release the content under some creative commons license (such as that pro git book). the book itself also seems good.. =)	14:33
	← dfr\|bus left	14:36
	← aspotashev left	14:38
	→ Diverdude joined	14:38
Diverdude	What is git?	14:38
wereHamster	Diverdude: git-scm.com	14:38
	also in the topic	14:38
Diverdude	a version control system like SVN?	14:39
wereHamster	Git is a free & open source, distributed version control system designed to handle everything from small to very large projects with speed and efficiency.	14:39
dark	also http://en.wiktionary.org/wiki/git	14:39
wereHamster	straight from the homepage	14:39
	why don't you read it...?	14:39
Diverdude	so why is git better than vn?	14:39
	svn	14:39
wereHamster	Diverdude: http://whygitisbetterthanx.com/#svn	14:39
	straight from google 'why is git better than svn'	14:40
	if you'd type your questinos into google instead of irc, you'd get the same answers..	14:40
	→ bosdfs joined	14:46
	← shahid_ left	14:47
	← __iron left	14:47
converge	i removed a lot of .svn directories inside my project and did: git add .; git commit; git push origin master; but the svn files still are in the remote respos. , there is another way to do it properly?	14:48
wereHamster	you need to use git add -u .	14:48
	otherwise git won't update deleted files in the index	14:48
	also, use git status	14:48
	→ EricInBNE joined	14:49
wereHamster	also, next time use git rm -f '**/.svn' .svn	14:50
converge	can i do it on windows?	14:51
wereHamster	sure	14:51
converge	cool =)	14:52
	→ yhager joined	14:52
	← ph^ left	14:52
	← goldbergus left	14:53
dark	I use -A to add deleted files. but it also add non-deleted files.. (it's --all). (but that's because nearly all adds I do is git add -Av ., I actually have an alias, git a, to it)	14:53
	→ oriba joined	14:54
	→ icwiener joined	14:54
	→ fedesilva joined	14:55
	← icwiener_ left	14:55
	→ __name__ joined	14:55
	← __name__ left	14:55
	→ __name__ joined	14:55
	→ goldbergus joined	14:58
	→ shiba_yu36 joined	14:58
	→ scarabx joined	14:58
	→ ph^ joined	15:00
sgronblo	could somebody confirm for me that the staging area contains a snapshot and not a delta set?	15:01
wereHamster	ack	15:01
	er.. confirmed	15:01
	→ icwiener_ joined	15:01
sgronblo	all right	15:01
	← icwiener left	15:02
	← virgil left	15:03
	→ tangyin joined	15:03
igorw	when I have a topic branch that was merged into master, is there a way to get a diff AFTER the merge has been performed	15:04
wereHamster	diff of what?	15:04
igorw	basically the equivalent of git diff master topic	15:04
	→ icwiener joined	15:05
igorw	as it was before the merge	15:05
wereHamster	you can pass any two commits to gi tidff	15:05
	so just pass <master before the merge> and <topic before the merge>	15:05
	→ sftp joined	15:06
	← tangyin left	15:06
	← goldbergus left	15:07
albel727	igorw: e.g. if merge have just happened, then it is "git diff master^1..master^2"	15:08
igorw	so is there an easy way to find the merge commit on master that points to topic?	15:08
sgronblo	is there any logic to how the caret parents are counted? which one is ^1 and which one is ^2	15:08
albel727	sgronblo: yes, there is	15:08
wereHamster	the first parent is ^1, the second parent ^2 and so on	15:08
	→ project2501b joined	15:09
sgronblo	yes... but which one is the "first" and which is the "second"?	15:09
albel727	sgronblo: ^1 is the branch you're currently on (i.e the one, into which you merge changes into)	15:09
	← icwiener_ left	15:09
igorw	s0ber: if you're on master and do git merge topic1 topic2; ^1 is master, ^2 is topic1 and ^3 is topic2	15:09
sgronblo	albel727: i guess you mean the branch which was merged into, only. it probably has nothing to do with what branch you are on when you write your carets	15:10
albel727	(well, not exactly "branch", rather a commit, which was pointed at by the branch into which you were merging changes)	15:10
	← project2501b left	15:10
igorw	my problem is basically this; we use github compare links that are linked to from tickets	15:11
	after the merge, it turns blank	15:11
	http://github.com/phpbb/phpbb3/compare/develop...rxu:ticket/9867	15:11
	so it's harder to review those changes after they have been merged	15:11
wereHamster	sure, because now the diff between devlop and the ticket is empty	15:12
igorw	of course, it makes sense	15:12
	it'd be nice if I could use a script to show the changes after the merge	15:13
	← rvsjoen left	15:13
wereHamster	replace 'develop' with the commit befor the merge	15:13
	→ goldbergus joined	15:14
igorw	and how to find it in an automated way?	15:14
wereHamster	walk through the history, look which merge has ^2 == ticket/XXXX and then extract ^1	15:14
igorw	so I guess there's no low level git commands for this stuff	15:15
wereHamster	git rev-list would be the low-level command	15:15
	→ rvsjoen joined	15:17
impulze	hm i wonder if i could get git-gui to use sync stuff from my index to workingdir rather than the opposite	15:17
	any ideas how to?	15:18
wereHamster	'git checkout -- .' would do that	15:18
igorw	wereHamster: ok, thanks	15:18
albel727	igorw: man git-log . I guess, something like git log --format="%H %P" \| grep `git-rev-parse topic` might help you	15:18
Gitbot	igorw: the git-log manpage can be found at http://git.or.cz/man/git-log	15:18
wereHamster	albel727: IMHO it's better to use rev-list --parents instead of the custom format	15:19
	→ mmc joined	15:19
	← josh0x0 left	15:20
	← alberto56 left	15:20
	→ irc2samus joined	15:21
albel727	wereHamster: yeah, I agree. In fact, it looks like I've tried to emulate git rev-list --parents =D. "git rev-list --parents -1 <topicbranch>" it is, then.	15:21
	← shiba_yu36 left	15:21
albel727	(oh, no it's wrong)	15:22
	→ LiamH joined	15:22
mmc	let's say I have these 2 commits: 1/ functionality with lot's of comments 2/ commit stripping the comments. Now I want to have this: 1/ functionality 2/ commit adding those comments (so that 1/ can be published, and 1/ can be used privately). So, I want to have A' = A+B and B' = -B How to do it? Can this be done only by making a backup out-of the repo directory?	15:23
	← fedesilva left	15:23
wereHamster	mmc: rebase and edit/split the commits	15:23
albel727	igorw: "git rev-list --children -1 <topicbranch>"	15:24
mmc	wereHamster: I guess I cannot obtain "-B" with only git commands.	15:24
wereHamster	-B? git diff -R <B>	15:24
mmc	hm, maybe yes	15:24
	ok	15:24
	→ ct^ joined	15:25
albel727	mmc: and what happened to good old git revert? just apply it to B, and voila.	15:25
	→ king313 joined	15:27
albel727	(so, basically, you just need to revert commit 2 yielding commit 3, and then squash commits 1 and 2 together)	15:27
mmc	albel727: very nice, indeed	15:27
irc2samus	guys is there something like svn:external in git? what should I read about? :)	15:27
king313	Hi. How, if I have added a modified file to the next commit, with "git add ." I can remove it	15:27
	← ct^^ left	15:27
king313	so it will not be uploaded in the next commit and push?	15:27
albel727	king313: you want, git rm --cached <file>, probably.	15:28
irc2samus	king313: git reset	15:28
king313	thanks guys!	15:28
albel727	king313: whoops, no. git reset would be better	15:28
	← ct^ left	15:28
	→ zombor joined	15:29
	→ warthog9 joined	15:30
	← theoros left	15:30
	→ ct^ joined	15:31
	← fr0sty__ left	15:31
	→ Yuffster joined	15:32
	→ fr0sty__ joined	15:32
	→ alberto56 joined	15:34
	→ kukks joined	15:37
	→ fedesilva joined	15:37
irc2samus	btw I'm looking for something like svn:external in git, that is, to be able to include an external repo within my own	15:41
	→ Kira joined	15:42
	→ slonopotamus joined	15:42
	← sh1mmer left	15:45
	fr0ggie → fr0ggie`	15:45
igorw	irc2samus: there's submodules, they work a bit differently though	15:45
irc2samus	igorw: hmm ok I'll read about those thanks!	15:46
	in any case, do you think they could help me include a separate repo within my own? I need to have a library deployed along with my code and I would like it to be updated when the author updates instead of me keeping it fresh	15:47
	→ chrisshattuck joined	15:47
	→ SkramX joined	15:48
	← kupsch left	15:49
	→ kupsch joined	15:51
	→ ddevaal joined	15:51
	← fedesilva left	15:52
	← Diverdude left	15:52
	→ sako joined	15:54
	→ parkin joined	15:54
	→ nonno joined	15:54
parkin	we currently have a publishing server with bare git repositories, which we push/pull too over ssh - this works great, but now i have a requirement for permissions / security - how do I allow / restrict access to git repos?	15:56
igorw	irc2samus: it's probably easier to have a separate deploy script that checks out those dependencies	15:56
sako	hey guys, i want to get the git branch to show up in my prompt.. here is what i have so far: http://dpaste.com/267176/	15:56
igorw	because submodules are bound to a specific revision	15:56
sako	i get no errors but the branch are not showing when i navigate to a git repo	15:56
	→ programble joined	15:57
sako	can someone help me figure this out?	15:57
	← irc2samus left	15:57
Ilari	parkin: Gitolite?	15:57
sako	i think this is the line thats messed up export PS1="\[${COLOR_GREEN}\]\w $(__git_ps1 " (%s)")> \[${COLOR_NC}\]";	15:57
doener	sako: use ' not "	15:57
sako	doener: i have some weirdness going on with '	15:58
doener	sako: with " the expressions are evaluated right away and the result gets hard-coded into your prompt	15:58
parkin	Ilari: havent heard of that before	15:58
sako	hmm	15:58
parkin	will have a quick look	15:58
	what are the benefits of it?	15:58
sako	doener: i am on a mac and if i dont use " my colors are messed up they show up as \e2345	15:59
	can i use both in bash?	15:59
doener	sako: no clue about that... then put a \ before the $() subshell to stop it from being executed	15:59
sako	-bash: /Users/svarozian/.bashrc: line 58: syntax error near unexpected token `('	16:00
	← warthog9 left	16:01
sako	export PS1="\[${COLOR_GREEN}\]\w \$(__git_ps1 " (%s)")> \[${COLOR_NC}\]";	16:01
doener	sako: escape the inner quotes	16:02
	← xiangfu left	16:02
	→ madewokherd joined	16:04
	← werdan7 left	16:04
	→ dfr\|bus joined	16:05
	→ thiago joined	16:06
	→ aspotashev joined	16:06
	← dfr\|bus left	16:07
	→ Diverdude joined	16:08
	→ iruediger joined	16:10
	← cschneid left	16:11
	← parkin left	16:11
	→ ofeone joined	16:13
	← bryanl left	16:13
sako	well that worked thanks doener	16:13
	← robgleeson left	16:13
sako	im assuming because that made it an expression that is evaluated later?	16:14
	just trying to understand what	16:14
	\$ does	16:14
wereHamster	otherwise the __git_ps1 would be evaluated when you set PS1	16:15
sako	i see	16:16
wereHamster	but you wnat it evaluated when the shell evaluates PS1	16:16
	→ theoros joined	16:16
	← igorw left	16:16
	→ robgleeson joined	16:16
	→ cschneid joined	16:17
	← goldbergus left	16:17
	← mithro left	16:17
	← slonopotamus left	16:19
sako	guys so if i have a branch x... do i do git pull origin x	16:20
	if i want to pull in changes from origin/x into local x?	16:20
	i dont have to explicitly say git pull origin/x x?	16:20
	or do i?	16:21
	→ goldbergus joined	16:21
sgronblo	you can set up branch x to track origin's x	16:23
	→ parkin joined	16:23
	→ langhun joined	16:24
	→ marc_kdab joined	16:24
	→ antgel joined	16:24
dark	can't a git alias be executed at the current directory? (it is always exec at the git root directory). or, is some way to pass the current directory?	16:25
	← chrisshattuck left	16:25
	← EricInBNE left	16:26
	← __name__ left	16:26
offby1	dark: can you be more specific?	16:26
dark	I have two aliases: env = !env and pwd = !pwd	16:27
	← bosdfs left	16:27
dark	git pwd always shows the root directory (which is usually good, for most aliases)	16:28
	git env \| grep PWD also shows as PWD=rootdirectory	16:28
wereHamster	dark: pwd = !$PWD/(git rev-parse --show-prefix)	16:28
	← langhun left	16:28
offby1	curious.	16:29
wereHamster	oh, that won't work	16:29
offby1	I didn't know you could make the alias mechanism run shell commands	16:29
dark	sh: -c: line 0: syntax error near unexpected token `git' , sh: -c: line 0: `$PWD/(git rev-parse --show-prefix)'	16:29
wereHamster	there was a patch to add a special syntax for aliases to get the pwd of the directory where the user is	16:29
	→ mithro joined	16:29
wereHamster	but I don't know if that was merged or what happened to it	16:29
	dark: 18:29:12 <wereHamster> oh, that won't work	16:30
offby1	manual indeed says "Note that shell commands will be executed from the top-level directory of a repository, which may not necessarily be the current directory."	16:30
dark	pwd2 = !echo $PWD/$(git rev-parse --show-prefix) runs, but shows /current/root/	16:30
wereHamster	again, that won't work	16:30
	I was wrong, and I'm sorry	16:30
offby1	dark: you might consider writing a _shell_ alias that does some stuff, then invokes the git alias.	16:30
	→ evilgeenius joined	16:31
dark	yes.. what I want to do is: git a is the same as git add -Av ., but I want also a git a something being the same as git add -Av something	16:31
offby1	s/shell alias/shell function/	16:31
sitaram	parkin: http://github.com/sitaramc/gitolite#_main_features	16:31
dark	offby1, yes, or a shell script..	16:31
offby1	(shell functions are easier to understand than shell aliases, in my humble opinion )	16:31
dark	but I find the syntax git something somethingelse to be nice	16:31
evilgeenius	My git index file has become corrupt. How can i replace it without loosing the changes to the files im working on?	16:31
	→ ymasory_ joined	16:32
wereHamster	dark: see if $GIT_PATHNAME_PREFIX is available in the alias	16:32
	← oriba left	16:33
dark	no.. I'm in a directory called src. git env \| grep src shows nothing	16:33
wereHamster	dark: or prefix the alias with '!!' instead of '!'	16:33
dark	sh: !env: command not found	16:34
wereHamster	then none of the patches were merged	16:34
	← masonkramer left	16:35
dark	ah, other people actually proposed it. hmm.	16:35
wereHamster	see [email@hidden.address] and [email@hidden.address]	16:35
	offby1 wonders if that was eli doin' the proposin'	16:35
dark	what? where?	16:36
offby1	wereHamster: how do you turn those things into email messages? gmane?	16:36
wereHamster	dark: those are message-ids. You can use gmane to search the messages	16:36
	offby1: yeah, eli	16:36
offby1	his unique style comes through, yay, even unto his message-ids	16:38
	← jast left	16:38
wereHamster	offby1: you know him?	16:39
	→ brntbeer joined	16:39
	→ jast joined	16:39
	→ fedesilva joined	16:39
dark	I was unable to locate those message ids, but I can find this discussion on various places, like http://kerneltrap.org/mailarchive/git/2010/5/5/29773	16:40
	in fact, I will try to apply this: http://kerneltrap.org/mailarchive/git/2010/5/5/29807	16:40
	← antgel left	16:40
sejo	hey all I'm following http://gitref.org/branching/	16:40
dark	I suspect that if the patch applies cleanly, it will be like changing a single line in the ebuild..	16:41
sejo	but when trying out the merge I get this: CONFLICT (delete/modify): test.py deleted in HEAD and modified in change_def. Version change_def of test.py left in tree.	16:41
	Automatic merge failed; fix conflicts and then commit the result.	16:41
	Thell\|busy → Thell\|busy\|Away	16:41
	Thell\|busy\|Away → Thell\|busy	16:41
	Thell\|busy → Thell	16:41
sejo	what does that mean? Can't git detect the namechange anymore?	16:41
dark	but people replied saying the patch is wrong..	16:42
wereHamster	sejo: "fix conflicts and then commit the result."	16:42
sejo	wereHamster: yes i know that, the problem is following the doc it states that there should not be a confict	16:42
	however there is one	16:42
offby1	wereHamster: I "know" him in the IRC sense. He's one of the PLT guys (they created the "Racket" language, which I use)	16:42
wereHamster	sejo: where does it say there should not be a conflict?	16:42
sejo	Well, it will just figure it out. Notice that I had no merge conflicts and the file that had been renamed now has the 'HiWorld' class name change that was done in the other branch. Pretty cool.	16:43
	git merge change_class	16:43
	Renaming hello.rb => ruby.rb	16:43
	Auto-merging ruby.rb	16:43
	Merge made by recursive. ruby.rb \| 6 ++---- 1 files changed, 2 insertions(+), 4 deletions(-)	16:43
	that's stated in the doc	16:43
wereHamster	sejo: there's no 'test.py' in that references. So you ditdn't follow it. You did something else which caused the conflict	16:44
	← ijcd left	16:44
sejo	wereHamster: I did the same just in python	16:44
wereHamster	no, you ditn't	16:44
	where does the reference tell you to delete test.py?	16:45
	← childermass left	16:45
sejo	master: http://paste.pocoo.org/raw/283552/ branche change_def: http://paste.pocoo.org/raw/283553/	16:46
	I did not delete	16:46
wereHamster	test.py deleted in HEAD and modified in change_def	16:46
sejo	I did git mv test.py code.py	16:46
wereHamster	well, that's the same as a delete	16:47
	→ saladin joined	16:47
	← kanru left	16:47
sejo	the doc does the same?	16:47
	$ git mv hello.rb ruby.rb	16:47
theoros	it's a shortcut for a regular mv, a git rm, and a git add	16:48
	→ project2501b joined	16:48
wereHamster	sejo: I see. You changed the files too much (they are completely different now)	16:48
	→ FsTluxe joined	16:48
wereHamster	so git doesn't recognize the rename and can't merge the changes with the correct file	16:48
	sejo: see how the reference just renames one line of the ~7 line file	16:49
theoros	you can change the default 90% limit can't you?	16:49
dark	it looks like in my source here, this patch is already applied.. (git-1.7.2.2.tar.bz2)	16:49
sejo	wereHamster: ok thanks!	16:49
	← ofeone left	16:49
	→ bhosie joined	16:49
	← simplechat left	16:49
drizzd	theoros: you can for diff, but not for merge I believe	16:50
	← nonno left	16:50
drizzd	and for merge, the limit is less than 90% anyways	16:50
	← project2501b left	16:52
	→ nonno joined	16:52
	→ ofeone joined	16:52
	← evilgeenius left	16:53
sejo	wereHamster: can it also be that it's the root-commit?	16:55
	→ whaley joined	16:57
dark	in the end, the !! commit was accepted or not?	16:58
	*patch	16:59
	→ antgel joined	17:00
	← leen^\| left	17:01
	→ Dave^\| joined	17:01
wereHamster	sejo: no	17:02
	→ simplechat joined	17:02
	← simplechat left	17:02
	→ simplechat joined	17:02
wereHamster	sejo: it's because the two files are completely different	17:03
	→ alnewkirk joined	17:03
	← parkin left	17:04
	← FsTluxe left	17:07
	→ towski joined	17:07
	→ yngress joined	17:10
	→ parkin joined	17:10
	→ bdiu joined	17:11
dark	is it possible to have, in, git show something:file, this "file" be relative to current directory?	17:11
offby1	as far as I know, yes	17:12
	Try It And See™.	17:12
dark	i'm trying	17:12
	← GarethAdams\|Work left	17:12
	← saladin left	17:13
wereHamster	no, <sha>:<path> is always relative to the root	17:14
dark	I think one can do some magic with git rev-parse --show-toplevel and pwd	17:15
	← thiago left	17:16
	→ orafu joined	17:16
sejo	wereHamster: ok get it working bet needed 75% the same	17:16
	← simplechat left	17:16
	→ thiago joined	17:17
	← keystr0k left	17:18
	← cschneid left	17:18
	→ zain joined	17:18
	→ cschneid joined	17:19
	→ ninjaneer joined	17:19
	← aspotashev left	17:19
wereHamster	sejo: in the reference the files are only 65% same	17:19
	← goldbergus left	17:21
sejo	i know but i have much less lines	17:21
	so added lines :p	17:21
	now it worked	17:21
	→ icwiener_ joined	17:22
	← parkin left	17:23
	← icwiener left	17:23
	→ nevyn^_ joined	17:24
	→ goldbergus joined	17:26
robinbowes	Anyone help me cloning a project onto github?	17:26
	→ ExtraSpice joined	17:27
robinbowes	I want to make a copy of the pssh codebase from git://aml.cs.byu.edu/pssh.git into a repo on github	17:28
	→ frakturfreak joined	17:28
	← whaley left	17:28
	← Gitzilla left	17:28
wereHamster	robinbowes: clone that repo, push it to github	17:28
robinbowes	I'm trying that	17:29
wereHamster	use --mirror when cloning and --all when pushing	17:29
	→ battleman2004 joined	17:29
	→ bitsnack joined	17:29
robinbowes	According to github, I need to do:	17:29
	git remote add origin [email@hidden.address]	17:29
	But I get this error:	17:30
	fatal: remote origin already exists.	17:30
bitsnack	i'm an absolute git-noob and I don't know how to delete a commit	17:30
	a already pushed commit	17:30
bremner	bitsnack: man git-revert	17:30
Gitbot	bitsnack: the git-revert manpage can be found at http://git.or.cz/man/git-revert	17:30
	← fedesilva left	17:31
bitsnack	it's my initial commit	17:32
wereHamster	robinbowes: use some other name for the remote	17:32
	maybe 'github'	17:32
	so you can git push github --all --tags	17:32
robinbowes	Ah, OK	17:32
	robinbowes has never quite got my head round git	17:33
wereHamster	robinbowes: progit.org	17:33
robinbowes	Good, is it?	17:33
dark	robinbowes, cat .git/config	17:33
	from the root	17:34
wereHamster	dark: why?	17:34
bitsnack	is there a command to "deinit" a repository?	17:34
wereHamster	bitsnack: rm -rf .git	17:34
dark	to see the remote origins..	17:34
bitsnack	just delete the .git directory?	17:34
	oh, okay	17:34
robinbowes	Hmm, I did a clone using --mirror	17:34
bitsnack	thanks	17:34
	→ Mage joined	17:34
robinbowes	That gave me something that doesn't look particularly useful	17:35
wereHamster	dark: git remote show	17:35
	robinbowes: --mirror creates a bare repo IIRC	17:35
dark	:) nice	17:35
robinbowes	Yes, seems to. You suggested I use it!	17:35
dark	I think bare repositories are more used for gitweb, stand-alone servers, etc	17:36
	← thiago left	17:37
robinbowes	Hmm: error: --all and --tags are incompatible	17:37
wereHamster	robinbowes: I did, because when you want to create a mirror on github then that's the easiest way	17:37
	← goldbergus left	17:37
wereHamster	try git push github --mirror	17:37
robinbowes	I just did git push github --all	17:38
wereHamster	that only pushes branches, but not necessarily all tags	17:38
robinbowes	OK, did git push github --mirror	17:39
	http://github.com/robinbowes/pssh	17:39
	OK, that looks good	17:40
wereHamster	robinbowes: http://support.github.com/discussions/repos/684-how-to-create-repository-in-mirror-mode	17:40
robinbowes	OK, I'll try and figure out how to use that	17:41
	Thanks	17:41
	→ Gitzilla joined	17:42
	→ hyperair joined	17:42
	→ goldbergus joined	17:42
	← Schmallon left	17:42
	← ninjaneer left	17:44
	→ warthog9 joined	17:47
	← goldbergus left	17:49
	← alnewkirk left	17:51
	→ Schmallon joined	17:51
	→ dreiss joined	17:52
	← ymasory_ left	17:52
	→ psoo joined	17:53
	← MikhX left	17:53
	→ Heimidal joined	17:53
	→ MikhX joined	17:53
	→ alnewkirk joined	17:53
	→ goldbergus joined	17:55
	← bhosie left	17:56
	← nonno left	17:58
	brntbeer → brnt\|burnt	17:58
	→ anjen joined	17:58
	→ larsl joined	17:59
	← anjen left	17:59
	→ jfkw joined	18:00
	← nairb774 left	18:00
larsl	Hello, I'm wondering if there's a way to refer to the commit that is the first one in the history of the current branch after the last merge of another branch.	18:00
wereHamster	larsl: no	18:00
	→ FsTluxe joined	18:01
wereHamster	well, no easy way. You'd haev to script it	18:01
	← dreiss left	18:02
	← dsigurds left	18:02
larsl	OK... is there a way to find the latest common ancestor of two commits?	18:03
	→ fedesilva joined	18:03
	← alnewkirk left	18:04
wereHamster	merge-base	18:04
	→ alnewkirk joined	18:04
	← tvw left	18:05
	← babusri left	18:06
	← zombor left	18:07
	← scarabx left	18:09
larsl	Found a way to work around it. filter-branch and rebase to the rescue.	18:10
	→ nairb774 joined	18:13
	← OOPMan left	18:14
theoros	?	18:15
	← krosenvold_ left	18:18
	← iravid left	18:18
	comawhite\|sleep → comawhite	18:18
	← rikarends left	18:18
	→ rikarends joined	18:19
	→ iravid joined	18:19
	→ ignatp joined	18:19
	→ whaley joined	18:24
	→ andreaa joined	18:24
	← nairb774 left	18:27
	← shennyg left	18:31
	← fedesilva left	18:32
	→ Adaptee joined	18:32
	→ project2501b joined	18:35
	← project2501b left	18:36
	→ svnlto joined	18:38
	← parasti left	18:38
	→ nairb774 joined	18:39
	← envi left	18:42
Thell	It'd be nice to have a git rebase --sign HEAD~3 -vs- git rebase --interactive HEAD~3 -> 'edit' to each -> git commit --amend -s -> save -> git rebase --continue -> repeat.	18:43
	hmmm, wonder if I could just create an alias to do that.	18:43
	→ rpg joined	18:44
	← bitsnack left	18:45
	→ shennyg joined	18:46
	← alberto56 left	18:47
	→ alester joined	18:50
	→ Guest10739 joined	18:52
	Guest10739 → gdoteof_home	18:52
	← nanotube left	18:54
	← theoros left	18:54
	→ theoros joined	18:54
	→ fedesilva joined	18:59
	→ nanotube joined	18:59
	→ mmc1 joined	18:59
	→ justin-george joined	19:01
	← mmc left	19:01
	← vital left	19:02
	→ vital joined	19:04
	→ krosenvold_ joined	19:06
	→ mcdebugger joined	19:06
	→ torsthaldo joined	19:07
	→ childermass joined	19:10
	→ steph021 joined	19:10
	← mmc1 left	19:10
	← converge left	19:14
robinbowes	wereHamster: yt?	19:15
	← xiong left	19:18
	→ justin_ joined	19:19
	→ zombor joined	19:20
	← tizbac left	19:22
	MmikeMRMA → MmikeDOMA	19:22
	← Aleric left	19:23
	→ phantomcircuit joined	19:23
	← maleknet left	19:24
	→ chrisshattuck joined	19:25
	← zombor left	19:25
bitkiller	anyway to do something like 'git add -p' line by line?	19:26
	← erez left	19:27
bitkiller	in other words, I want to split a block, but have no option	19:27
Thell	bitkiller: git add --interactive will let you do that	19:28
wereHamster	robinbowes: ?	19:28
robinbowes	Hi	19:28
DrNick	-p is a just a shortcut into --interactive	19:28
	see the split command	19:28
robinbowes	So, I followed http://support.github.com/discussions/repos/684-how-to-create-repository-in-mirror-mode	19:29
	I now appear to have an "empty" dir	19:29
wereHamster	then you didn't follow it	19:30
	→ karl_popper joined	19:30
	→ mmc joined	19:30
robinbowes	My github repo is populated	19:30
	But nothing in the local dir	19:30
wereHamster	nothing at all?	19:31
robinbowes	Well, there's a .git dir	19:31
bitkiller	Thell, can't find an option in the interactive mode	19:31
	← ijuma_ left	19:32
robinbowes	"git remote show" shows "mirror" & "origin"	19:32
	→ malumalu joined	19:32
bitkiller	DrNick, i don't have the split command for the block in question (if i understood what you mean)	19:32
	→ meatcar joined	19:32
robinbowes	Like I said, my github repo is populated: http://github.com/robinbowes/pssh	19:33
Thell	bitkiller: '5: patch'	19:33
bitkiller	Thell, this is the same as I am doing: git add --patch	19:33
	← graphitemaster left	19:33
Thell	yep	19:34
bitkiller	Thell, then I have no option to split the block	19:34
Thell	you do	19:34
bitkiller	No I don't	19:34
	← chrisshattuck left	19:34
Thell	when you get to the hunk you want to split, you can do split or even edit	19:34
	→ xiong joined	19:35
bitkiller	I can edit, but cant split 'automatically'	19:36
robinbowes	I'll try it again	19:36
Thell	bitkiller: sometimes the internal logic can't find a logical place to split and it'll tell you so.	19:36
	← FsTluxe left	19:36
	→ chrisshattuck joined	19:36
	→ gilaniali joined	19:37
	← robgleeson left	19:37
bitkiller	Thell, ok, i see...i'd say the internal logic is kind of flawd	19:38
	→ erez joined	19:38
bitkiller	thanks anyway	19:38
Thell	sure	19:38
	← goldbergus left	19:38
	← whaley left	19:38
	→ robgleeson joined	19:38
	→ flaguy48 joined	19:39
	← LRN left	19:40
	→ LRN joined	19:41
	→ njhjhhhhhj joined	19:42
	→ goldbergus joined	19:42
	← karl_popper left	19:42
	← gdoteof_home left	19:44
	→ tizbac joined	19:45
	→ ijuma_ joined	19:46
	← artefon left	19:46
	→ robinr joined	19:47
	← Tac left	19:48
	→ Tac joined	19:50
	← Adaptee left	19:50
	← nanotube left	19:51
	→ dharrigan joined	19:52
	← dharrigan left	19:52
	→ dharrigan joined	19:52
	← mithro left	19:52
	← zenom_ left	19:53
robinbowes	wereHamster: I did this: http://pastie.org/1260655	19:54
	I deleted and re-created the github repo, ie. it was empty before I ran those commands	19:55
	It is now populated as a mirror of the pssh repo	19:55
	http://github.com/robinbowes/pssh	19:55
	→ corni_ joined	19:56
	← chrisshattuck left	19:56
	← dharrigan left	19:58
	→ stamina_ joined	19:59
	← corni left	19:59
	← psoo left	20:01
	← erez left	20:01
	→ chrisshattuck joined	20:01
	← xiong left	20:04
	← stamina_ left	20:04
	→ zombor joined	20:05
	→ mithro joined	20:05
	→ nanotube joined	20:07
	→ dl9pf joined	20:07
	← dl9pf left	20:07
	→ dl9pf joined	20:07
	← robinr left	20:09
	← disappearedng_ left	20:09
	← kukks left	20:09
	→ erez joined	20:15
	→ mbroeker joined	20:17
	← icwiener_ left	20:21
	← njhjhhhhhj left	20:26
	← Heimidal left	20:26
	→ sh1mmer joined	20:27
	← andreaa left	20:29
	← joeytwiddle left	20:30
	← fedesilva left	20:32
	→ Morasique joined	20:35
	→ tildeequals joined	20:36
	→ joeytwiddle joined	20:37
	→ fedesilva joined	20:39
	→ dl9pf_ joined	20:41
	← gilaniali left	20:41
	← dl9pf left	20:42
	← Schmallon left	20:44
	→ Schmallon joined	20:46
	← chrisshattuck left	20:50
	← Anti-X left	20:51
	→ jmcantrell joined	20:54
	→ robinr joined	20:55
	← phryk left	20:55
	→ amerine joined	20:55
	→ Anti-X joined	20:58
	→ whaley joined	20:58
	→ bradg0620 joined	20:59
	← robinr left	20:59
	→ frakturfreak_ joined	21:00
	← mastro left	21:01
	→ hilem joined	21:01
	← zombor left	21:01
	→ scarabx joined	21:02
	← fedesilva left	21:02
	← frakturfreak left	21:02
	→ gilaniali joined	21:03
	← ch077179 left	21:03
	→ opalka joined	21:06
	← alester left	21:07
	← amerine left	21:09
	→ ipalindromei joined	21:09
	← bentob0x left	21:09
	→ disappearedng joined	21:10
	→ kukks joined	21:10
	← Schmallon left	21:10
ipalindromei	I'm getting this really crazy issue with git 1.7.3.2 on a Mac. It's reporting a png file modified, but the hash value is the same as the previous commit. I tried a commit -av but it reports "no changes added to commit". reset --hard doesn't work either.	21:11
	← nicoulaj left	21:11
ipalindromei	any pointers would be appreciated.	21:11
	→ amerine joined	21:12
	→ corni__ joined	21:12
Ilari	ipalindromei: 'git update-index --refresh' doesn't help?	21:12
ipalindromei	reports the file as "needs update"	21:12
	← urkud_ left	21:13
	← brothers left	21:13
Ilari	ipalindromei: In which section does git status report file as modified? Not staged for commit?	21:13
	→ MikhX_ joined	21:13
	→ mastro joined	21:13
ipalindromei	"Changed but not updated:"	21:13
Ilari	ipalindromei: What does 'git ls-files -s' report as hash of that file?	21:14
	→ dharrigan joined	21:14
	← dharrigan left	21:14
	→ dharrigan joined	21:14
ipalindromei	here's one really goofy part of it. The file is named www/img/downArrow.png, but the modified report is "modified: www/img/DownArrow.png" (note caps difference)	21:14
	two files match a grep -i downarrow: 6885... www/img/DownArrow.png, and c385a... www/img/downArrow.png	21:15
	← corni_ left	21:15
	← childermass left	21:15
	← MikhX left	21:15
	MikhX_ → MikhX	21:16
	→ urkud_ joined	21:16
	→ jargan joined	21:16
	← bburhans left	21:17
	← jast left	21:17
	→ bburhans joined	21:17
	← bburhans left	21:17
	→ bburhans joined	21:17
Ilari	ipalindromei: What does 'git add www/img/downArrow.png' complain about?	21:17
	← Derander left	21:17
ipalindromei	no complaint, exits with 0. But git status still reports "changed but not updated"	21:17
	← Morasique left	21:18
	→ Derander joined	21:18
ipalindromei	as further information, ls www/img/ \| grep -i downarrow only reports downArrow.png in the directory.	21:18
	→ BlueSherpa joined	21:20
	→ _iron joined	21:20
	← gilaniali left	21:20
	→ gilaniali joined	21:21
BlueSherpa	I have a 14mb sql file checked into git. It says the file is Unchanged, but it looks like the file was in conflict (because of the various <<<< and >>>> markers spread through the file). Is there a way to fix the conflict and remove the markers or do I have to manually go through the file?	21:21
	← whaley left	21:22
	← Anti-X left	21:22
cschneid	Is there a good way to manage large binary files inside git? I dont' really want to version them, but I do want them easily available to anybody who grabs the project	21:23
	frakturfreak_ → frakturfreak	21:23
cschneid	versioned shell script "grab_assets.sh" and files hosted on s3 perhaps?	21:23
	→ Schmallon joined	21:23
BlueSherpa	cschneid: I think there's a way to link external projects in git	21:23
	← dharrigan left	21:23
BlueSherpa	that way grabbing the main project also pulls the others	21:23
cschneid	BlueSherpa: example large file is a virtual machine image. Or photoshop files w/ mockups.	21:24
	stuff that wouldn't change often. Version control sucks at binary file diffs... and I certainly wouldn't want a full history of a file	21:24
	unless I asked for it.	21:24
	→ iipalindromeii joined	21:25
BlueSherpa	cschneid: well, if you want them in git, you can check them into their own repository where they just sit	21:25
	→ nicoulaj joined	21:25
iipalindromeii	Ilari: back, sorry about that. Got disconnected.	21:25
BlueSherpa	cschneid: otherwise, it sounds like you're trying to include links inside the repository that points to files that reside outside of the repository	21:25
cschneid	BlueSherpa: yeah. This is more of a workflow question than a git-specific question I suppose.	21:26
	was wondering how other people handle it	21:26
BlueSherpa	I've been including mockups in git	21:26
	→ Anti-X joined	21:26
BlueSherpa	not VMs tho	21:26
cschneid	basically I want to avoid a giant clone, due to a long version history of a file	21:26
	→ whaley joined	21:26
Ilari	iipalindromeii: Perhaps that's because filesystem is case-insensitive?	21:26
BlueSherpa	makes sense	21:26
cschneid	BlueSherpa: yeah, we've been making heavy use of Vagrant, with custom VMs... but they get big quickly (several hundred megs), and we're all remote	21:27
BlueSherpa	cschneid: I think it's a question of whether you're archiving the large binaries	21:27
Ilari	iipalindromeii: Maybe try 'git rm --cached www/img/DownArrow.png', 'git add www/img/downArrow.png'? But that would be rename...	21:27
cschneid	BlueSherpa: I wonder if you can set submodules to do a shallow clone	21:27
	by default	21:27
BlueSherpa	good question	21:27
cschneid	so we can still have the full history (disk is cheap), but without the network overhead	21:28
BlueSherpa	something like - don't follow links or what not	21:28
cschneid	BlueSherpa: more of "grab the newest version of each file, but no history please"	21:28
BlueSherpa	cschneid: that's where git externals would come in	21:29
	just a sec, I'm looking at the command	21:29
	← ipalindromei left	21:29
	iipalindromeii → ipalindromei	21:29
cschneid	BlueSherpa: git externals? is that the command name?	21:29
BlueSherpa	I think that's the feature	21:30
	← urkud_ left	21:30
BlueSherpa	I googled, git external	21:30
	came up with one link so far http://6brand.com/git-svn-externals-rails-plugins.html	21:30
	cschneid: looks like the clone command supports a depth option for adjusting the amount of history you grab	21:31
cschneid	yeah	21:31
ipalindromei	Ilari: yes, it's case insensitive. My issue is that I need to discard the changes.	21:31
cschneid	I think the shell script approach may be best. So workflow for a fresh copy of a repo is: git clone host:repo	21:31
ipalindromei	Ilari: I'm working with a detached head	21:31
cschneid	then ./fetch_assets.sh	21:31
	easy enough for a small team	21:31
BlueSherpa	cschneid: have you thought of using git hooks?	21:31
cschneid	is there an after_clone hook?	21:31
	I've only barely scratched hooks	21:31
ipalindromei	hmm. Actually, your git rm --cached seemed to do the trick. I guess I could commit that then leave the commit dangling somewhere. Lose references to it.	21:32
BlueSherpa	dunno, but .... git init --bare	21:32
	that'll create all of the various hook examples	21:32
ipalindromei	Ilari: okay, I'll work with that. Thanks for the help!	21:32
	→ hobodave joined	21:32
BlueSherpa	I don't see any clone hooks, but maybe the manual will have more available	21:33
	cschneid: yea, it says post-checkout hook is run after a clone	21:33
cschneid	BlueSherpa: thanks for looking into this. I have a few ideas for going forward.	21:35
BlueSherpa	cool, good luck with it =)	21:35
	cschneid _away	21:35
	← meatcar left	21:36
	→ icwiener joined	21:38
ipalindromei	Okay, still not working. If I remove the files, commit that change, then rebase and drop the change...I immediately get the problem again.	21:39
	the problem is that the file is checking out as www/img/downArrow.png, but git status is looking for www/img/DownArrow.png	21:40
	any idea where I could start to look in the code to see where this issue is occuring?	21:40
	→ z0mb joined	21:41
	← justin_ left	21:41
	← sh1mmer left	21:41
	→ gonemad3 joined	21:42
	← workmad3 left	21:42
	→ fismoll8 joined	21:43
fismoll8	hello everyone. I am pulling from an svn repo using git-svn. Did two commits, created two patches via git format-patch HEAD~2...but when I run git apply --check my_patches, I get error: patch failed: trunk/vimfiles/ftplugin/latex-suite/mathmacros-utf.vim:26	21:44
	error: trunk/vimfiles/ftplugin/latex-suite/mathmacros-utf.vim: patch does not apply. What is going on? I have inspected the patchfiles, and they look fine.	21:44
	→ kipras joined	21:45
	← gilaniali left	21:45
	← whaley left	21:46
	→ whaley joined	21:46
	→ little_owl joined	21:47
	← hilem left	21:48
	→ hilem joined	21:49
	← hilem left	21:49
	→ hilem joined	21:50
	← ShadeHawk left	21:54
	← z0mb left	21:55
	→ fedesilva joined	21:56
Thell	I'm trying to get a --msg-filter to work on adding sign-offs, could someone take a look at http://git.pastebin.com/uFGwWEQT and see if there is an obvious reason this logic works from the command line but not in the sourced script?	21:58
	looks like it should be simple enough to do, but something is causing it to be different than @ the prompt	21:59
	→ One3yed joined	21:59
	← AAA_awright left	21:59
little_owl	What is a minimal IQ to understand Git? If it a high value, I may stop learning it in advance ... :-(	22:00
	→ AAA_awright joined	22:00
Thell	my test from the command line is: git log -1 \| cat && if $(git log -1\|grep -q -E '^[[:space:]]*Signed-off-by: '); then echo "Would exit"; else echo; echo "Signed-off-by: "; fi	22:00
larsl	little_owl: Your IQ must be higher than the number of active branches in your work repository.	22:01
Thell	lol	22:01
larsl	Which gives quite some headroom.	22:01
little_owl	larsl: :-D Ok, I can continue ....	22:01
	← fedesilva left	22:03
	← One3yed left	22:06
	→ zombor joined	22:07
	→ Micki joined	22:08
little_owl	I use a git-cvs module. Is it possible to "merge" during the import more CVS modules into one Git reposity?	22:09
	← Micki left	22:11
larsl	little_owl: You should be able to fetch different CVS modules to different branches and then merge locally. Look at the git-cvs man page, I haven't done any CVS work in a while (thankfully).	22:11
	← cwillu_at_work left	22:11
	← jones- left	22:11
little_owl	larsl: Thanks for a hint!	22:12
	→ cwillu_at_work joined	22:12
avar	right, import it into git, then merge	22:13
	→ jones- joined	22:13
	→ bhosie joined	22:14
	← chrischambers left	22:15
	Putr → Putr\|Away	22:16
	← gonemad3 left	22:18
Thell	wooot!	22:19
golodhrim	Happy scary Halloween	22:19
Thell	Figured out the problem with my filter branch :D Should have used double quote instead of single around the grep	22:19
	← VoiDeT left	22:21
little_owl	So I have to import every cvs module to its' own git repository, than create a "main" git repository and "git pull" every imported git repository to "main" one. Am I right? Do I lost a possibility to synchronize the code with the original cvs repository?	22:23
larsl	little_owl: You should be able to import them to different branches in the same repository.	22:25
	→ envi joined	22:26
	← scarabx left	22:27
little_owl	larsl: And synchronize a git branch with the related cvs module ... nice	22:27
	← whaley left	22:28
	← mdel_ left	22:28
	← _iron left	22:29
	← brnt\|burnt left	22:33
	← goshawk left	22:36
	← Fullma left	22:37
	← zombor left	22:39
	→ ludde3 joined	22:43
	→ frogonwheels joined	22:43
	← ofeone left	22:43
	→ SIGBUS joined	22:44
	← SIGBUS left	22:44
	← tizbac left	22:44
	← ludde left	22:45
	← kipras left	22:46
	← jmcantrell left	22:46
	← kukks left	22:49
	dwmw2_gone → dwmw2_BOS	22:49
	← ipalindromei left	22:50
	← malumalu left	22:50
Thell	anyone3on who has setup an alias in msysgit that takes arguments?	22:51
	→ BiggFREE joined	22:52
Thell	I'm trying to setup a filter-branch alias and can't seem to get the correct escaping to have the single quote get parsed correctly for the cmd portion	22:53
	→ drizzd_ joined	22:53
	← programble left	22:56
	← drizzd left	22:56
wereHamster	foo = "filter-branch ... ' ... "	22:57
	does that not work?	22:57
	→ keyvan joined	22:58
	← fismoll8 left	22:58
Thell	nope, have to do foo = \"!sh -c '...' \" to get virtually any git command alias to run, let alone trying to get an arg in there	22:58
keyvan	Hi, anyone know how to git clone to particular SHA/commit?	22:58
	← wooden left	22:58
	→ programble joined	22:59
	← programble left	22:59
	→ programble joined	22:59
Thell	keyvan: if I undertand correctly you can't unless it is in refs/heads on the remote	23:00
keyvan	i believe it is, although i dont fully understand refs/heads	23:00
wereHamster	keyvan:clone one of the branches which includes that commit, then checkout that commit	23:00
keyvan	i gotta pull every version change of node to find out why its incompatible with socket...... stuck on this issue haha	23:01
wereHamster	keyvan: clone the whole repo, then use bisect	23:01
	→ wooden joined	23:01
keyvan	wereHamster: that sounds like it may work, reading on bisect now	23:01
	wereHamster: this is exactly what i needed. thanks	23:03
	← Elfe left	23:04
keyvan	wereHamster: (assuming i understand it correctly but it looks like it automates what i need :D)	23:04
	→ jmcantrell joined	23:06
	← programble left	23:06
	→ programble joined	23:06
	← wooden left	23:06
	→ EricInBNE joined	23:07
	→ fisted_ joined	23:07
	← kpreid left	23:07
	← EricInBNE left	23:08
	→ EricInBNE joined	23:09
	← rhct left	23:09
	← basicxman left	23:09
	→ rhct joined	23:09
	→ petran joined	23:12
petran	Has anyone ever encountered an "error: git-pull died of signal 4" error while git pulling?	23:13
	→ Elfe joined	23:13
	← ari-_-e left	23:13
wereHamster	signal 4 is SIGILL here	23:14
	→ friskd joined	23:15
	← ExtraSpice left	23:15
	← Schmallon left	23:16
fisted_	when merging an upstream branch, and a conflict occurs, will the offending file, which i edit to resolve the conflicts, already contain all other non-conflicting changes to it, caused by the merge?	23:16
wereHamster	yes	23:17
petran	wereHamster: what do you mean? what is SIGILL?	23:17
wereHamster	petran: "git-pull died of signal 4" - signal 4 is SIGILL	23:17
	petran: http://en.wikipedia.org/wiki/SIGILL	23:17
petran	wereHamster: thanks, reading it	23:18
	← little_owl left	23:18
wereHamster	4 is sigill on linux and mac. But I don't know about other platforms	23:19
fisted_	wereHamster: woudl you mind taking a short glance at http://penenen.de/weird.png (the actual code is irrelevant). on the left it shows the real differences between me (red) and upstream (green) (i diff'ed those two files manually there). on the right side you can see what a merge does to that file (note that stuff outside the conflict markers is thill the old version :O)	23:19
	← hilem left	23:20
petran	wereHamster: illegal instruction! (it's on a Mac) that makes me think there might be something wrong with the git installation... hrm...	23:21
fisted_	s/thill/still/	23:21
wereHamster	petran: ehwo did you install git?	23:21
	fisted_: are you sure the diff is between the two versions you're trying to merge?	23:22
	→ mmoya joined	23:22
petran	wereHamster: I didn't install it. it's for another guy I am supporting. he said he has 1.7.0, that it was working before and now it doesn't... I wonder if he installed anything recently that could mess with git	23:22
	← king313 left	23:22
fisted_	wereHamster: i'm sure (double checked) that they both reflect the most recent version in their master branches. what i cannot ensure is that my fetch/merge actually pulls the latest thing	23:23
	but actually it should	23:23
	from a clean working directory, i added a remote branch, and fetched it, and merged it into my master branch	23:23
wereHamster	fisted_: how did you fetch and merge it? And how did you generate the diff?	23:25
	→ TSK joined	23:25
	← priidu left	23:26
	← sraue left	23:27
fisted_	wereHamster: starting with a clean working directory, i said 'git remote add upstream git://...', then 'git fetch upstream', then 'git merge upstream/master'	23:28
	wereHamster: i generated the left diff by manually calling diff on copies of that file from my repo, and the upstream repo	23:28
	i'm currentyl cloning upstream to make sure again	23:29
	← quadz left	23:29
	→ quadz joined	23:29
	← opalka left	23:30
fisted_	made sure.	23:30
	← bhosie left	23:31
fisted_	i'm probably missing something essential to merging	23:31
	wereHamster: incase you're not short on bandwidth and/or time, could you try to clone my repo and reproduce it?	23:31
	← corni__ left	23:32
fisted_	my repo: git://github.com/fisted/tw_tournament.git - upstream repo: git://github.com/oy/teeworlds.git	23:32
	← brosner left	23:33
fisted_	what i did in order to merge: http://codepad.org/CEVxhz3X	23:34
	→ little_owl joined	23:34
	← nevyn^_ left	23:35
	← ajpiano left	23:35
	→ sraue joined	23:35
petran	wereHamster: do you thing the SIGILL problem can be only caused by a bad installation, or could it be also caused by something wrong in the repo the user is trying to pull from?	23:36
wereHamster	petran: that has nothing to do with the contents in the repo	23:36
	← little_owl left	23:36
petran	wereHamster: i guess it doesn't even get the chance to contact the repo.. it dies before that, doesn't it?	23:37
	← icwiener left	23:39
	→ sixteneighty joined	23:39
	← chrizz left	23:39
wereHamster	fisted_: I haev no idea. Ask again in a few hours (or g=during the week), or ask on the mailing list	23:43
	→ ajpiano joined	23:44
	→ dharrigan joined	23:50
	← dharrigan left	23:50
	→ dharrigan joined	23:50
	→ fedesilva joined	23:50
	← d0k left	23:51
	← friskd left	23:55
	← jmcantrell left	23:55
	→ sh1mmer joined	23:58

Logs Search ←Prev date Next date→

Channels Documentation