IRCloggy #git 2011-11-07

Provider of IRC logs since 2005.

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

	2011-11-07
	← jamiemill left	00:00
	← ipalaus_ left	00:00
	← veganadian left	00:02
	→ cloudgroup joined	00:03
	→ jargon- joined	00:05
	→ RaceCondition joined	00:07
	→ EricInBNE joined	00:08
	← cloudgroup left	00:13
	← p3rror left	00:14
	→ gigamonkey joined	00:15
	→ jrnieder joined	00:16
	→ replore joined	00:18
	→ neurodrone_ joined	00:21
	← neurodrone left	00:21
	neurodrone_ → neurodrone	00:21
	→ CpuID2 joined	00:21
CpuID2	hey pals…. I've been going through some guides regarding pushing a copy of my git tree to a remote server (as opposed to doing a clone/pull from the target server), just deciding if I'm best using git init or git init --bare on the remote end prior to doing a git remote / git push remotetarget master ?	00:22
	← airborn left	00:22
jrnieder	hi CpuID2	00:22
CpuID2	one issue with git init is its considered "non-bare", so you get the "refusing to update checked out branch refs/heads/master" message on push	00:23
	→ airborn joined	00:23
jrnieder	"git init --bare" makes a repository with just the objects and refs	00:23
	without the files checked out	00:23
CpuID2	yea	00:23
	which defaults to the current working dir	00:23
jrnieder	so yeah, it's usually exxactly what one wants on a server	00:23
CpuID2	rather than creating a .git	00:23
jrnieder	"git init --bare foo.git"	00:23
CpuID2	ok so that gives me effectively what would be the contents of .git on a normal clone	00:24
jrnieder	if this repository is meant to be pushed to by multiple people, then "git init --shared --bare foo.git" and see gitcvs-migration(7)	00:24
	← easytiger left	00:24
CpuID2	basically… I'm looking to push out local copies of a package's source tree to boxes for deployment	00:25
	and then bootstrap/build within them	00:25
	but without hammering the pkg maintainers git server in doing so	00:25
jrnieder	ah, deployment :)	00:25
CpuID2	so on my staging server, i git clone first	00:25
jrnieder	I'm a fan of the "hook on the server side" model	00:25
CpuID2	and my staging box has ssh keys to all of the release targets	00:25
	:)	00:25
	initially i was going to bootstrap/build on the staging box	00:26
	but I've chosen to do that on the release targets for now	00:26
	→ easytiger joined	00:26
CpuID2	it does mean more builds, but meh	00:26
	theres benefits there as well	00:26
	yea i see whatcha mean about hooks on the server side	00:26
	(FYI, its freeswitch I'm pushing down)	00:26
	← kp^ left	00:26
sandstorm	is there any way to store repo on an ftpes account?	00:27
	← airborn left	00:27
jrnieder	sandstorm: sure, use your usual ftp client	00:27
CpuID2	so jrnieder in this scenario… do you see any problems with the mkdir /usr/src/freeswitch && cd /usr/src/freeswitch && git init route?	00:27
jrnieder	if your account has very limited space, I'd recommed using git-bundle(1)	00:27
sandstorm	jrnieder: can I do commits and checkouts over ftpes?	00:27
CpuID2	its not going to be a shared work environment on each release target	00:28
jrnieder	sandstorm: depends on lots of details --- how far away is your ftp server? is there a limited quota? any security requirements? why ftp and not something like git protocol, http, or ssh?	00:28
	← CannedCorn left	00:28
CpuID2	but for future releases i was going to just git push updates down to the targets (incrementally) then git reset --hard on them, rather than pushing full trees	00:28
	→ TheMoonMaster_ joined	00:29
sandstorm	jrnieder: very far, no limited quota, no security requirements(except tls) and how can I do via http? can I commit over http and so on ?	00:29
jrnieder	CpuID2: imho life is easier with just one central repo, with a hook to notify the targets to go fetch when when there are new changes to fetch	00:29
sandstorm	jrnieder: it is a shared hosting btw.	00:29
CpuID2	jrnieder: oh the issue here… is for initial new target deployments	00:29
	once they have a tree to work off i can have them pull down updates fine :)	00:30
jrnieder	CpuID2: I see	00:30
CpuID2	(or ill actually git push them from my staging box, due to the preference of not wanting my targets to have keys to the staging box)	00:30
jrnieder	anyway, others can say something about pushing to deploy	00:30
	I might be the only one who likes avoiding magic with worktrees :)	00:31
CpuID2	hehe :)	00:31
	in essence there will be a central repo here	00:31
	but i was avoiding running a git server to allow git:// clones	00:31
nevyn	gitolite!	00:31
CpuID2	the staging box will be the central repo	00:31
	which is built off doing a clone and keeping it updated from the upstream pkg.'s tree	00:32
jrnieder	sandstorm: regarding your questions -	00:32
	← thiago left	00:32
	← kermit left	00:32
jrnieder	a. commits happen locally, so that part of the workflow doesn't care what protocol you use :)	00:33
	→ kermit joined	00:33
sandstorm	jrnieder: can I push the commits to ftpes?	00:33
jrnieder	b. yep, I think fetch over ftps works, but it's been a while since I checked. Make sure you run "git update-server-info" at the appropriate times to make index information available to the client, since the server won't be smart enough to generate it on the fly	00:33
sandstorm	jrnieder: and do checkouts over there?	00:33
jrnieder	c. I don't remember if anyone implemented push over ftp/ftps	00:34
	but you can always use a normal file-synching program to just send the foo.git directory over there, and clients won't mind	00:34
sandstorm	hmm I see	00:35
jrnieder	you can certainly push over http	00:35
sandstorm	so file synching program will mimic pushes?	00:35
jrnieder	yep, yep	00:35
sandstorm	and pulls?	00:35
jast	yes, only much less efficiently	00:35
jrnieder	for pull, as I said, as long as you run "git update-server-info", the git client can handle it (slowly, as jast said)	00:35
sandstorm	I can push over http but since it is a shared hosting, they might not have set configurations for git to work	00:35
jast	the tricky part is updating the server info thingy	00:36
	with shared hosting it might be difficult	00:36
jrnieder	right, push over http requires either a webdav setup on the server (slow, complicated) or the "smart HTTP" cgi script	00:36
sandstorm	what is the use of server info ?	00:36
	I am very new to this repository thing	00:36
jrnieder	when git fetches, we don't want to just slirp up everything	00:36
jast	it's for "dumb HTTP" so that client can figure out what branches you've got and at what version they are	00:36
jrnieder	what jast said	00:37
jast	basically without the server info that HTTP transport doesn't really do you any good	00:37
sandstorm	got it	00:37
jast	the "smart HTTP", on the other hand, requires more stuff on the server to be set up	00:37
	← \|Agent left	00:37
sandstorm	so I will probably give up :)	00:37
jast	you need a git binary somewhere, and some CGI-related setup	00:37
	your project isn't, per chance, open source?	00:37
sandstorm	nope it is a small web project	00:38
	commercial one	00:38
jast	... there goes the free hosting alternative	00:38
sandstorm	we are 2 people so we wanted to use a nice versioning system	00:38
	→ glcrazy_ joined	00:38
	→ dv310p3r joined	00:38
sandstorm	to sync our works	00:38
jast	depending on how aggressive your respective firewalls are, you might be able to get by without a central repository	00:39
jrnieder	sandstorm: do you have ssh access to the shared host?	00:39
jast	since git is decentralized, in theory you can just swap stuff without a middle man	00:39
sandstorm	jrnieder: nope, ssh is disabled, it is a business hosting	00:39
offby1	sigh	00:39
sandstorm	unfortunately	00:39
	← k-man left	00:40
offby1	so "business" here must actually mean "lame"	00:40
sandstorm	hmm	00:40
	what would be the best practice to swap things in order to sync 2 peoples work ?	00:40
	copying over?	00:40
jrnieder	sandstorm: I'd exchange git bundles by email :)	00:40
sandstorm	oh very nice thinking sir	00:40
	is that possible?	00:40
jrnieder	or patches, to be able to comment on them	00:40
	yep	00:40
jast	bundles need some manual figuring out work	00:41
jrnieder	for bundles, see man git-bundle(1)	00:41
jast	automatic message the 'git-bundle' manpage can be found at <http://git.kernel.org/?p=git/git.git;a=blob_plain;f=git-bundle.html;hb=html>	00:41
jrnieder	for patches, see man git-format-patch(1)	00:41
jast	automatic message the 'git-format-patch' manpage can be found at <http://git.kernel.org/?p=git/git.git;a=blob_plain;f=git-format-patch.html;hb=html>	00:41
sandstorm	do patches include real work files?	00:41
jast	i.e. you need to tell git which commit range to pack into a bundle	00:41
sandstorm	I see	00:41
jast	textual patches, OTOH, are easily to read and thus review before applying them	00:41
sandstorm	ah very nice feature	00:41
jast	they're not so exciting if your repository involves binary files	00:41
sandstorm	for instance we create artwork, we create php files and etc. when we commit new files, the patch will include compressed versions of them right?	00:42
jrnieder	in a bundle, everything is compressed	00:42
CpuID2	hmm i have a feeling I'm rather close to where i want to be right now :)	00:42
jast	a patch is just a regular unified diff	00:42
sandstorm	for instance, .psd files etc	00:42
jrnieder	in a patch, everything is in a text form	00:42
CpuID2	i ended up just going git init && git config receive.denyCurrentBranch ignore	00:42
	on the target	00:42
jast	a bundle, on the other hand, is compressed	00:42
jrnieder	binary files get converted to some kind of textual representation	00:42
CpuID2	then doing a git remote … … && git push … master	00:43
jrnieder	of the delta	00:43
CpuID2	from my staging	00:43
sandstorm	so I will go with bundles I guess to sync our work ?	00:43
	ah I see	00:43
	interesting	00:43
	textual representation wow	00:43
CpuID2	but… now i have what looks like a lot of "deleted" files on my target according to git status, but they don't actually exist	00:43
jast	yeah, probably just base64-encoded or whatever	00:43
	← glcrazy_ left	00:43
	→ glcrazy_ joined	00:44
jast	CpuID2: automatic message an explanation of bare and non-bare repositories can be found here: http://bare-vs-nonbare.gitrecipes.de/	00:44
sandstorm	textual or not, patches or bundles will be serving for our needs very good features	00:44
	what happens if we concurrently changed same files and send bundles to each other?	00:45
	will they be merged and new patch comes out ?	00:45
CpuID2	reads	00:45
	\	00:45
jast	then at least one of you is going to have to merge the bundle locally, and create a new bundle that contains the merge	00:45
	→ bonhoffer joined	00:45
sandstorm	got it	00:46
jrnieder	or rather, when you fetch from the bundle, you'll see the actual history	00:46
	with a branch	00:46
jast	note that not the two bundles get merged, but rather devA's bundle gets merged to devB's local repository (including its history)	00:46
	fine distinction, though :)	00:46
jrnieder	and then you can use "git merge" on it	00:46
sandstorm	but if both of us merge it the product will be the same but different SHA1's ?	00:46
	← TheMoonMaster left	00:46
jrnieder	that's right, so one of you should do the merge	00:46
jast	if you merge it the same way, yeah	00:46
jrnieder	and send a bundle with it to the other	00:46
jast	usually it's best to have one person merge and the other just watch and look busy	00:46
sandstorm	hehehe	00:47
	I liked the watch and look busy part :P	00:47
	← glcrazy_ left	00:48
jast	if you trust the other developer to do good merges, you might want to think about convincing him to do the 'merge' part and focus on the 'look busy' part	00:48
sandstorm	I am trying to figure out the best practice complying with this 2 man scenario, that's why I am asking these, thanks alot for the information :)	00:48
	hmm so as far as I get, merging requires skills :)	00:48
	in terms of reliability and integrity	00:49
	maybe I could go with the merging part :(	00:49
jast	yeah, we regularly get folks trying to salvage their messed up merges	00:49
sandstorm	I see	00:49
jast	the problem with messed up merges is that they're hard to undo (mostly due to communication overhead, so in a 2-dev team it's less of an issue but still extra work)	00:49
sandstorm	I just started reading about repositories and git yesterday	00:50
	learning alot	00:50
jast	nice :)	00:50
	some people come here and try to solve problems without learning. doesn't work nearly as well.	00:50
	→ TheMoonMaster joined	00:50
jast	at least that's the impression I get	00:50
sandstorm	true, it is the extra work but I wonder how merges are done in huge projects such as linux kerel, it should be done elegantly	00:50
SethRobertson	read !book	00:50
jast	automatic message 'Pro Git' is a complete book about git that is available online for free (there's a print version, too): http://progit.org/book/	00:50
	well, I can explain it briefly	00:51
CpuID2	jast: ok i can pretty much confirm i want a non-bare repo, as i do want a working tree to be able to do stuff in on the destination box	00:51
	← caseymcg left	00:51
CpuID2	tried mkdir /usr/src/repo && cd /usr/src/repo && git init	00:51
bob2	tl;dr don't blindly use 'git pull'	00:51
CpuID2	then dropped that post-update in there	00:52
sandstorm	okay I can listen	00:52
jast	CpuID2: the last sections address how you can make that happen	00:52
CpuID2	(in hooks, chmod +x etc)	00:52
	it complained about receive.denyCurrentBranch on git push from my source, set it as ignore on the target	00:52
	it ran	00:52
	but i still don't have a working tree on the target?	00:52
	its definitely propagating data across though, looking at the output of git push on the source	00:52
	← RaceCondition left	00:52
	← sorin left	00:53
jast	sandstorm: there is a sort-of counterpart to merge. suppose devA has a few commits that he wants to get included upstream. before he sends it, he gets the latest changes from upstream, then uses this other tool to remove the need for the merge to involve any work, then sends the updated commits.	00:53
	it comes with its own gotchas, of course	00:53
CpuID2	interestingly enough on the target, git status shows them all as "deleted" files, like its awaiting a commit or something	00:53
jast	CpuID2: yeah, you need a hook script that syncs up the working tree with the new history	00:53
	← saschagehlich left	00:53
CpuID2	I've dropped the post-update script in there...?	00:53
	(from that URL)	00:54
jast	CpuID2: it's showing them as 'deleted' because they are new in the history and they're not in the working tree, i.e. it looks as if you checked out that new history and then deleted them	00:54
CpuID2	yea fair call	00:54
	can i verify if the hook was actually called?	00:54
sandstorm	jast: what if that commiting is trying to be done by 2 people at the exact time? they both got the latest codes but trying to do concurrent commits?	00:54
jrnieder	CpuID2: add "echo hellooooo there" to the hook :)	00:54
sandstorm	*same	00:55
CpuID2	ah :)	00:55
bob2	sandstorm, then one of them will end up with a very boring merge	00:55
jast	or have it create a dummy file	00:55
bob2	sandstorm, but you can just try and see for yourself	00:55
CpuID2	hah simple solution	00:55
	git reset --hard	00:55
	:)	00:55
sandstorm	hmm allright, I am learning :)	00:55
CpuID2	voila, working tree :)	00:55
jast	sandstorm: chances are still very good that if the submissions don't cover the same part of the codebase, they can be merged automatically	00:55
	→ Cipher-0 joined	00:55
jast	and if they can't, that's just rotten luck and one of them will have to redo the work :) or the maintainer will have to do it	00:56
	in the end the maintainer is always responsible anyway	00:56
sandstorm	ah I see	00:56
CpuID2	I'm still getting my head around git… but i am starting to see while its complexed in its own ways its very flexible :)	00:57
	(coming from an svn background)	00:57
sandstorm	so the maintainer will see 2 branches from the same codebase which were supposed to be done on top of eachother	00:57
	← tewecske left	00:57
jast	actually he'll probably see two branches that were supposed to be done "next to" each other	00:57
jrnieder	sandstorm: a good way to figure these things out is to experiment	00:57
	and see what happens	00:57
jast	yeah, experiments rock :)	00:58
sandstorm	allright, I need to try on a test repository :)	00:58
jast	luckily test repositories are cheap	00:58
CpuID2	ok awesome I'm satisfied with this now	00:58
sandstorm	thank you very much for your time and consideration	00:59
CpuID2	effectively not cloning down every time, just propagating changes :)	00:59
	jast and jrnieder: thanks for pointing me in the right direction/s :)	00:59
	→ flash3d8_ joined	00:59
jrnieder	yeah, thank you jast :)	01:00
jast	hey, what just happened	01:00
	→ qian joined	01:00
jast	three people thank me and my internet connection disappears :O	01:00
	→ sorin joined	01:00
sandstorm	thank you all, I really appreciate this community	01:00
	:P	01:00
	sorin → Guest73959	01:01
CpuID2	hahaha	01:01
jast	yeah, you're laughing	01:01
	→ luxurymo_ joined	01:01
jast	while I'll have to hold a funeral for my late connection	01:01
	→ timmywil joined	01:02
	← glcrazy left	01:02
CpuID2	hehehe	01:02
	→ aidenhong joined	01:02
CpuID2	jast: thanks for your help (since you missed it the first time ;))	01:02
	→ glcrazy joined	01:02
jast	no, I got it all right	01:02
	my connection was gone... but my IRC client is on another connection	01:02
	→ glcrazy_ joined	01:02
jast	in another castle, even	01:03
	anyway, you're welcome, guys	01:03
	tizzo → tizzo-afk	01:03
jast	still considering charging for the help at some point in the future ;)	01:03
CpuID2	ah :)	01:04
	nps	01:04
	irc bounces ftw	01:04
	→ lunchub joined	01:04
	→ kl4m joined	01:04
	← MUILTFN left	01:06
	→ PovAddict joined	01:06
	← bitkiller left	01:06
	← JoeyA left	01:06
PovAddict	I have a bunch of repositories with many common objects (they are the different attempts of conversions from the same SVN repository)	01:07
	is there any way to "merge" the packfiles together so they take less space?	01:07
	ie. end up with a single packfile with the objects of all	01:07
	← kriegerod left	01:07
kevlarman	PovAddict: if they really are the same, git repack will do it	01:08
CpuID2	ok heres a good question, whats the best method for using per-target SSH keys with git remote?	01:08
	.ssh/config ?	01:08
	i assume i can't really provide one in the url or as a param to git remote	01:09
PovAddict	CpuID2: yes, .ssh/config	01:09
CpuID2	np	01:09
PovAddict	kevlarman: how would I do that? if I copy the packfile of one repo into the objects/pack/ of another, and run git repack, it would just decide that none of the objects in that pack is referenced by refs in the repo	01:10
	← nudb2 left	01:10
kevlarman	PovAddict: why are you trying to share objects between different repositories?	01:10
mystik	s	01:10
kevlarman	(that tends to lead to some really unexpected things happening)	01:10
	→ endy_c joined	01:12
jrnieder	kevlarman: might be on a machine with limited disk space	01:12
	← flaguy48 left	01:12
CpuID2	hmm theres not something equivalent to rsync -e for git remote is there?	01:12
	that would be so much more preferable to using .ssh/config right now :)	01:12
SethRobertson	GIT_SSH?	01:12
jast	PovAddict: look at the alternates mechanism, and heed its warnings (documented for the -s option of git clone, for example)	01:12
CpuID2	o0	01:12
	that sounds promising	01:13
jrnieder	PovAddict: consider making them a single repository	01:13
CpuID2	aha :) http://www.reddit.com/r/git/comments/hdn1a/howto_using_the_git_ssh_variable_for_private_keys/	01:13
	perfect	01:13
	this will work, thx SethRobertson :)	01:13
PovAddict	kevlarman: they have lots of common data (in some cases there may be a single tree object differing between two), since they are different conversions of the same svn repo, with tweaked conversion settings	01:13
endy_c	Anybody here use bitbucket? How do I upload my already running local repo?	01:13
	→ cooloney joined	01:14
PovAddict	endy_c: I use bitbucket but only for Mercurial repos	01:14
jrnieder	PovAddict: what are these repos going to be used for? (just curious)	01:14
SethRobertson	git remote add bitbut, git push bitbut	01:14
endy_c	PovAddict: I'm trying for the first time bitbucket's git repo	01:14
	SethRobertson: thanks!	01:15
PovAddict	jrnieder: archival; I already had problems in the past with a converted repo missing stuff, realizing one month after, and not having kept enough information to track down what went wrong	01:15
	→ kerrick joined	01:15
	← sandstorm left	01:17
	→ maletor joined	01:17
	→ sandstorm joined	01:17
jrnieder	PovAddict: thanks	01:18
	→ shiba_yu36 joined	01:19
	← wbednarski left	01:20
	← glcrazy left	01:21
	← glcrazy_ left	01:21
	→ nyeates joined	01:21
kevlarman	PovAddict: then yeah just fetch all the branches into a single repo	01:22
nyeates	any recommendations on how to remove an api key from a file (but not the entire file), from many commits?	01:22
frogonwheels	nyeates: look at man git-filter-branch	01:23
jast	nyeates: automatic message the 'git-filter-branch' manpage can be found at <http://git.kernel.org/?p=git/git.git;a=blob_plain;f=git-filter-branch.html;hb=html>	01:23
	← sandstorm left	01:24
	→ CannedCorn joined	01:25
	← shruggar left	01:26
	← endy_c left	01:27
nyeates	k thx	01:28
	← nyeates left	01:28
	→ baburdick joined	01:28
	→ tiagobutzke joined	01:29
	← novettam_ left	01:31
	→ xiehong joined	01:31
	→ herdingcat joined	01:31
	→ envi_ joined	01:31
	Pilif12p → Pilif\|stufy	01:32
	Pilif\|stufy → Pilif\|study	01:32
	← ArseneRei left	01:34
	→ bonhoffer_ joined	01:35
	← Targen left	01:36
	→ thews joined	01:36
	→ caseymcg joined	01:38
	← envi_ left	01:38
	← bonhoffer left	01:38
	→ bonhoffer joined	01:38
	→ niloc132 joined	01:39
	← berserkr left	01:39
	← brian_g left	01:40
	← flash3d8_ left	01:40
	→ ziggles joined	01:41
	← bonhoffer_ left	01:41
	Pilif\|study → Pilif12p	01:42
	← bonhoffer left	01:44
	→ tomas_w joined	01:45
	→ andygrossman joined	01:45
	← andygrossman left	01:46
	← bigkm left	01:47
	→ eephillip joined	01:48
	← jrnieder left	01:50
	→ bonhoffer joined	01:52
	← Mannequin left	01:52
	→ cronin joined	01:55
	← luxurymo_ left	01:56
	← lisak left	02:00
	→ ouyes joined	02:01
	→ nightshift joined	02:09
	← Cipher-0 left	02:12
	← tomas_w left	02:15
	← nightshift left	02:15
	← the-fonz left	02:21
	→ alc joined	02:22
	← Guest73959 left	02:27
	← dreiss left	02:28
	→ Yuuhi`` joined	02:30
	← CannedCorn left	02:31
	→ wks joined	02:31
	← Yuuhi` left	02:32
	→ flaguy48 joined	02:33
	→ CannedCorn joined	02:33
	→ jrnieder joined	02:35
	→ regedit joined	02:38
	→ kimico joined	02:40
regedit	is there some general guideline about when to leave some file untracked and when to gitignore?	02:42
	i have files which are specific to the site at which the particular clone of the origin repo resides (think .htaccess files or path-specifig config files) should those be untracked or gitignored?	02:42
	← herdingcat left	02:43
jrnieder	regedit: not sure I follow	02:44
oc80z	me either	02:44
jrnieder	.gitignore is a good place to describe the files that your project's build process will create	02:44
regedit	hm leme try again	02:44
jrnieder	for example, it can list patterns like "*.o", and then contributors will not end up "git add"-ing any .o files unless they use --force	02:44
	→ glennpratt joined	02:45
jrnieder	one contraversial question is whether a project's .gitignore should include entries like ".*.swp" for generated files that are not due to the build process	02:45
regedit	this project is not your typical application / tool programmed in C / C++ / Java, it's actually just a website. php.	02:45
oc80z	right.	02:45
PovAddict	I think files like .*.swp should be in gitignore	02:45
oc80z	yeah	02:45
jrnieder	on one hand, this means it is less likely that someone will accidentally add those files; on the other hand, a list of all possible files of that kind can get a bit absurd	02:45
oc80z	them vim files, eh?	02:46
regedit	which means that the checked out working directory is the final shipped product	02:46
PovAddict	but something that is eg. a log file by a tool a certain developer uses, and not something used by the buildsystem, then that dev should add it to info/exclude	02:46
	which is local to his repo	02:46
oc80z	they are in my gitignore.	02:46
jrnieder	a principled answer would be to say, "always list those only in .git/info/exclude or the file mentioned in [core] excludesfile in ~/.gitconfig"	02:46
frogonwheels	.. or also local config is a good candidate for .gitignore.. where you have thingconfig.examplefile committed, and then thingconfig ignored	02:47
	jrnieder: or better yet, a global ignore file as appropriate.	02:47
	jrnieder: I mean for *.swp	02:47
jrnieder	frogonwheels: yeah, that's what I was vaguely gesturing at with the [core] excludesfile thing	02:47
frogonwheels	jrnieder: sorry - was typing it at the same time - so didn't read before <enter> :)	02:48
	→ ArseneRei joined	02:48
PovAddict	kevlarman, jrnieder: I won't bother merging the packfiles, I just found that rzip does wonders with the directory full of git repos	02:48
frogonwheels	oh .. no I mean I didn't read that far of your line jrnieder.. sorry	02:48
PovAddict	956MB -> 193MB	02:48
jrnieder	frogonwheels: yeah, np	02:48
	→ Vampire0 joined	02:48
jrnieder	there's also room for less principled approaches, like "we've received enough patches that add .swp files, so let's just list it in .gitignore to document that we are never going to have a source file with that name"	02:49
	→ variable joined	02:49
jrnieder	[/end side-rant :)]	02:50
frogonwheels	jrnieder: huh. yeah. f* add . ;)	02:50
	→ niboshi joined	02:50
PovAddict	and there's a small chance that LZMA will do even better; 4 minutes left for compression to finish	02:50
	← Vampire0_ left	02:51
PovAddict	wait, I took the above number from a stale ls -l :P	02:51
	rzip was actually 956MB -> 218MB	02:51
jrnieder	regedit: to answer your question --- frogonwheels's thingconfig.examplefile advice seems good	02:52
regedit	i was having an issue where I gitignored that kind of stuff but it was still being comitted, maybe because it was already added priorly...	02:53
PovAddict	yeah, gitignoring something won't remove it if you already added it	02:53
jrnieder	yep, .gitignore means "save me from accidentally adding it", not "ignore this file even if it is already being tracked"	02:54
regedit	so i have to make sure to .gitignore it before i do git add . right?	02:54
	← tiagobutzke left	02:54
jrnieder	regedit: nah, just use "git rm --cached"	02:55
	← maletor left	02:55
regedit	oh, ok	02:55
	where can i find #git logs ?	03:00
frogonwheels	regedit: it's in the topic	03:00
regedit	i was looking for it..	03:00
variable	hehe. git-svn lost to compiling gcc twice!	03:01
regedit	frogonwheels: i only see [1.7.7.2 \| http://git-scm.com \| Please skim http://jk.gs/gitchan.html before asking, or at least before giving up!]	03:01
variable	regedit: the second link	03:01
regedit	oh it's linked	03:01
variable	http://colabti.org/irclogger/irclogger_log/git?date=2011-11-07	03:01
regedit	sorry, thanks!	03:01
	← bonhoffer left	03:02
	variable modifies HKLM through regedit	03:03
	frogonwheels would have been more impressed if it had been through a vim plugin.. or 'reg' command directly ;)	03:04
	regedit has actually been having a hard time with environment variable s through CMD line 'REG' lately	03:04
PovAddict	eww	03:04
regedit	REG ADD HKCU\Environment /f /v HOME /d C:\foo\bar	03:05
	that part works	03:05
variable	frogonwheels: regedit would have to /nick for that to work	03:05
regedit	but when I REG DELETE HKCU\Environment /f /v HOME , the variable still works	03:06
	variable always works	03:06
regedit	I even developed a tiny little C++ executable to call windows api SendMessageTimeout() with flag 'WM_SETTINGCHANGE' for "Environment", didn't seem to help..	03:09
	it's like environment variables get cached or something	03:09
	← samrose left	03:10
ouyes	git is amazing	03:11
	← flavius left	03:12
	← PovAddict left	03:13
	← glennpratt left	03:13
offby1	variable: even on the Lord's Day? tsk tsk	03:21
	← CannedCorn left	03:22
	→ Targen joined	03:22
	← kerrick left	03:27
	← nicxvan left	03:29
variable	offby1: hrm?	03:29
	→ dreiss joined	03:35
CpuID2	hey guys… question, if i want to git pull and replace all local commits with whatever upstream has	03:37
	whats the best flag to use?	03:37
jrnieder	git fetch upstream	03:37
	git reset --hard upstream/master	03:37
	something like that	03:38
CpuID2	hmm upstream doesn't exist in this case	03:38
	what can i use to check what else its called?>	03:38
jrnieder	you mean, you use a url instead of a named remote?	03:38
	git fetch http://host/path/to/repo	03:38
	git log FETCH_HEAD	03:38
	git reset --keep FETCH_HEAD	03:39
	→ herdingcat joined	03:39
CpuID2	ive done a git clone git://xxx to get this repo	03:39
	by default does it name remotes?	03:39
jrnieder	origin	03:39
CpuID2	or is it optional?	03:39
	ah	03:39
jrnieder	(see man git-remote)	03:39
	← alecthegeek left	03:39
jast	automatic message the 'git-remote' manpage can be found at <http://git.kernel.org/?p=git/git.git;a=blob_plain;f=git-remote.html;hb=html>	03:39
CpuID2	there we go :)	03:39
	→ mdel joined	03:39
CpuID2	git fetch origin	03:39
	git reset --hard origin/master	03:39
	:)	03:39
	all good now	03:40
regedit	what is the difference between writing "foo/" versus writing "/foo/" to a .gitignore file at the root of the working dir ?	03:40
	→ injekt joined	03:40
CpuID2	i assume thats the equivalent of git pull -f && git reset --hard	03:40
	in the event there were no local commits?	03:40
	→ alecthegeek joined	03:41
	← johnkpau_ left	03:41
jrnieder	CpuID2: I don't know what git pull -f does	03:41
injekt	How would one best go about displaying information about the first commit for a given file	03:41
jrnieder	try it?	03:41
CpuID2	oh i did	03:41
	it pulls down commits from upstream	03:41
	but it won't overwrite local commits	03:41
	→ Chib joined	03:41
CpuID2	i think git fetch origin works better there	03:41
jrnieder	jast: html docs moved to git.kernel.org/?p=git/git-htmldocs.git, master branch	03:41
thansen	I accidentally did a git pull and have conflicts as a result..is it possible/easy to get back to my pre-pull state?	03:42
jrnieder	thansen: man git-pull tells me "to cancel a conflicting merge, use git reset --merge"	03:44
jast	thansen: automatic message the 'git-pull' manpage can be found at <http://git.kernel.org/?p=git/git.git;a=blob_plain;f=git-pull.html;hb=html>	03:44
thansen	jrnieder: perfect, thanks	03:45
jrnieder	np	03:45
regedit	anyone please? what is the difference between "foo/" versus "/foo/" in a .gitignore file at the root of the working dir ?	03:45
jrnieder	regedit: man gitignore answers that in its cryptic way	03:45
jast	regedit: automatic message the 'gitignore' manpage can be found at <http://git.kernel.org/?p=git/git.git;a=blob_plain;f=gitignore.html;hb=html>	03:45
jrnieder	"If the pattern ends with a slash, it is removed for the purpose of the following description"	03:45
regedit	yeah i read that	03:46
	cryptic indeed	03:46
jrnieder	"If the pattern does not contain a slash /, git treats it as a shell glob pattern and checks for a match against the pathname relative to..."	03:46
	"Otherwise, git treats the pattern as a shell glob" that is not relative	03:46
regedit	i can't make heads or tails of most of that..	03:46
jrnieder	which part?	03:46
regedit	in which case does it follow which rules, and what may those rules be..	03:47
jrnieder	case 1: no '/' in the pattern (except a possible '/' at the end of the line)	03:47
regedit	"for the purpose of the following description" meaning for the purpose of clarity in the man page itself?	03:47
jrnieder	case 2: '/' in the pattern (ignoring a possible '/' at the end of the line)	03:48
regedit	by "relative" you mean relative to the working dir?	03:49
jrnieder	well, the rest of the sentence answers that question	03:49
	relative to the directory containing the .gitignore file	03:50
	meh, I'm being deliberately dense here, sorry about that	03:50
regedit	:P	03:50
jrnieder	a better answer to your question would be, "the relevant piece of documentation is completely unclear"	03:50
	"git log -S'purpose of the following' Documentation/gitignore.txt" tells me it's Junio's fault	03:50
	"I'd suggest writing a patch that makes it clear but possibly wrong and sending it to [email@hidden.address] and then the kind people on that list can correct your patch and give a clear answer and update the documentation at the same time"	03:51
regedit	i couldn't even write a wrong patch even if i tried	03:52
	let alone clear	03:52
jrnieder	rats, I was about to say "that's great --- writing a right patch is even better" :)	03:52
regedit	hah	03:52
	← niboshi left	03:52
variable	why does git-svn clone take _forever_ ?	03:53
jrnieder	anyway, the idea is that '*.jpg' will match jpeg files throughout the repository	03:53
	while '/*.jpg' will only match jpeg files in the toplevel directory	03:53
	if you want to match .jpg files in the Documentation/ directory and its subdirectories, you can use the pattern '*.jpg' in Documentation/.gitignore	03:54
	To match .jpg files in the Documentation/ directory/ and not in a subdirectory, you can use 'Documentation/*.jpg' in the toplevel .gitignore	03:55
	or '/*.jpg' in Documentation/.gitignore	03:55
CpuID2	ok next… :) sorry for all the questions, whats the most sane way to check if a directory is a git repo?	03:55
	← caseymcg left	03:55
regedit	jrnieder: i am copypasta'ing all of this into a personal man page for future ref	03:55
	→ niboshi joined	03:55
jrnieder	regedit: yay :)	03:56
	in case you missed it, that business about "that is not relative" up there was total nonsense	03:56
regedit	jrnieder: in fact, go right now and submit this patch "...that makes it clear but possibly wrong and sending it to [email@hidden.address] :P	03:57
	→ kerrick joined	03:57
jrnieder	hmph	03:57
	→ bindaasomatic joined	03:57
jrnieder	ok	03:57
	can I have your name and email to cc and for a Reported-by?	03:58
regedit	er wow are you serious? ok.....i guess..	03:59
	← timmywil left	03:59
jrnieder	or I can credit you as "regedit on irc", that's fine, too	03:59
regedit	er, how do i pm..	03:59
jrnieder	/msg jrnieder hello	03:59
	← aidenhong left	04:03
regedit	jrnieder: i believe any bit of useful info that anyone produces shouldn't go to waste in some irc log. if this little snippet about gitignore helps someone, it should be find'able for others too	04:04
	← Fretta left	04:05
regedit	as such, one day bots should parse IRC logs and construct entire wikis out of 'em	04:05
jrnieder	heh	04:06
	← jimubao left	04:06
jrnieder	maybe some day #git will be replaced by jast's bot on a web form and an "edit" button	04:07
regedit	hah - wait, edit button?	04:07
jrnieder	to change its responses	04:08
regedit	right	04:08
	is there such a thing as pulling down changes excluding specific files (possibly listed in the puller's .gitignores) ?	04:09
	or must a fetch/merge/pull be complete	04:09
chrisf	regedit: not that i know of.	04:09
regedit	so how can I possibly have the following workflow;	04:10
chrisf	regedit: you could do a merge, and then hand-resolve some files back to "your" premerge state.	04:10
	regedit: but that sounds like asking for pain.	04:11
regedit	1) there is an upstream bare repo for pulling from and pushing to	04:11
	← eeanm left	04:11
regedit	2) repo 'foo' wants to pull with the exception of some files, and repo 'bar' wants to pull including those files	04:11
	→ eeanm joined	04:12
regedit	3) profit	04:12
	any way of doing that?	04:15
jrnieder	regedit: which are these optional files?	04:16
regedit	jrnieder: Eclipse project files (.project, .buildpath, .settings/ folder etc)	04:16
jrnieder	oh, yuck, okay :)	04:17
regedit	developers who pull the repo should get those, but my production environment where the project (a php website) is deployed should not get those files	04:17
	← niloc132 left	04:17
jrnieder	write a makefile that only deploys the files you want deployed?	04:18
	→ CannedCorn joined	04:18
	→ macUzer joined	04:19
regedit	er, frankly i'm not so familiar with make and such tools... primarily because this isn't really the kind of project you need to compile / build / install etc, it's an as-is php website	04:19
cmn	then you know you have PHP available on the server, write a script in PHP that does it	04:19
regedit	cmn: what script? when will it be executed?	04:20
	→ Kazii joined	04:20
	← ArseneRei left	04:20
cmn	a deployment script to be executed whenever you want to deplay	04:20
	like on post-receive	04:20
jpalmer	regedit: he's telling you, git isn't the appropriate place for this. if you want to "deploy" things other than what is in git, you need to write teh appropriate deploy tools.	04:21
	→ kennethreitz joined	04:21
jpalmer	you can either write the logic into your existing php app (if these files exist, remove them) or you can write an actual deploy script that only deploys the files you want it to.	04:22
cmn	right, git is a VCS, so that's where the files are; in order to put the files in the right place, you need some sort of script that takes those files and puts them in the right place	04:22
	→ regedit_ joined	04:23
regedit_	oops	04:23
	did i miss anything	04:23
	i was saying;	04:23
	that's the weird thing about an as-is php website project managed by git; there's no 'deployment' process / script. a simply checked out wroking dir is all the website needs to run	04:23
cmn	I was saying: git is a VCS, so that's where the files are; in order to put the files in the right place, you need some sort of script that takes those files and puts them in the right place	04:24
regedit_	well this is news indeed...	04:24
cmn	you must have a deployment script, even if it consists just of 'GIT_TREE=/var/www/ git checkout -f'	04:24
	← regedit left	04:25
cmn	so just add delete those files after you've copied them to the right place	04:25
	← macUzer left	04:25
	← crakdmirror left	04:25
regedit_	cmn: but wouldn't I want to make use of the features of a VCS for my deployed website?	04:25
	it helps for example for hotfixes	04:25
cmn	you can make your deployment script be your post-receive hook	04:26
regedit_	keeping track of 'em by merging deployment hotfixes back to the main upstream repo	04:26
jpalmer	regedit_: seriously, git is for maintainging your files. it's not your deployment toolkit.	04:26
	← Chib left	04:26
regedit_	jpalmer: see my last question then	04:26
cmn	regedit_: there's nothing stopping you; just make that script part of your post-receive hook	04:27
	then you can deploy on each push, just remember that it's not "git" doing it, it's whatever script you wrote	04:27
regedit_	cmn: am googling about hooks now...	04:27
jpalmer	you can tie it in to your deployment toolkit, as cmn is explaining. but even for a 'simple web app' you should be using a deploy script, not git ;)	04:27
	← eephillip left	04:27
	→ crakdmirror joined	04:28
cmn	jpalmer: many HOWTOs tell you to use 'GIT_WOKRTREE=/var/www git checkout -f' as a deployment script	04:28
	the line there is a bit blurry, I'd say	04:28
EugeneKay	Many HOWTOs are just plain wrong	04:28
cmn	no doubt about that	04:28
regedit_	the git checkout -f just won't work for me actually	04:29
	as it does not perform deletions of removed files	04:29
cmn	ok, so you /do/ have a deployment script, however simple	04:30
	just delete whatever files you're not interested in	04:30
regedit_	another reason i'd want to use git as my deployment versus manually scripting something - why should i copy extra data all over again for every new fix/enhancement? i'd want something like git to just update the changed stuff only	04:30
cmn	but git doesn't have any 'deploy' command, so you need to script it	04:31
jpalmer	regedit_: you're asking here. So my assumption is you want the advice from people who are experienced in these things. If that assumption is wrong, let me just say: it's your web app. and your code. Do with it as you please, even if that means ignoring great advice.	04:32
regedit_	jpalmer: i was inviting anyone in the know to refute my arguments	04:32
	→ Turicas joined	04:32
jpalmer	and, we have.	04:32
jrnieder	now, now	04:33
	regedit_: how do you currently take care of deployment?	04:33
jpalmer	now the conversation is just starting to go in circles. you can write a deploy script.. or you can decide not to. but, I don't see a reason to keep asking the same questions	04:33
regedit_	jrnieder: i'm still setting things up manually, trying to figure out the best workflow	04:33
jrnieder	I see	04:34
	the post-receive hook receives a list of ref updates for every push	04:34
	the githooks manpage says a little about this	04:34
	so if one of your requirements is that only the changed files get updated, you can use that information to find out what files changed	04:35
regedit_	upon cmn's advice i'm looking to see whether some git hook can help me with the last-minute delete operations, which (sofar) is all i need to 'deploy'	04:35
	← injekt left	04:35
regedit_	right	04:35
	like i said, technically a plan-old checked out branch of this repo is (almost) all ready to go live	04:35
	it's just some files here 'n there i'd want to pluck out	04:36
jrnieder	sure, and that might be what you want to go with	04:36
	a plain git checkout, with some files plucked out	04:36
	as long as you're careful about the webserver configuration (so as not to serve the .git and so on), that can work fine	04:36
	← ISF left	04:36
jrnieder	note, though, that git is not made for this	04:36
	for example, when it updates files in the worktree, it follows the order of operations (1) remove changed files (2) write their new content out	04:37
	if someone browses there between steps (1) and (2), 404	04:37
	← madewokherd left	04:38
	→ Flogalen2 joined	04:38
regedit_	what you are saying would appear to advise one not to use the git working tree as the actual deployment, but rather make some deployment script/process which should perhaps copy all the content elsewhere where it will be used/served/deployed	04:39
	to which i presented 2 arguments;	04:39
	← Flogalen left	04:39
	← bradj left	04:40
regedit_	1) hotfixes 2) git has built in functionality for only updating/changing exactly and only whatever actually changed, i contrast to traditional copy commands which have no idea about versioning and diffs. shouldn't i prefer this?	04:41
jrnieder	well, in response to (1):	04:41
regedit_	*in contrast	04:41
jrnieder	I'm used to an edit/compile/test cycle when editing code	04:42
	← cronin left	04:42
jrnieder	when changes to two pages have to happen together to make sense, it avoids ever deploying a broken intermediate state	04:42
	→ jacknagel joined	04:42
jrnieder	however, I can understand that avoiding the 'compile' step would save some time	04:43
	isn't it possible to serve a development copy of the page from the source repo?	04:43
regedit_	a php website is a tiny bit different.. you can edit and test immediately (with a webserver running) and php files are rather independent units	04:44
	jrnieder: currenlty i was thnking of 3 repos on the server :D	04:44
jrnieder	ah, there are no separate development machines	04:44
regedit_	the upstream dev, a "sqa" repo for testing, and a production repo	04:44
jrnieder	that is what I was missing	04:44
regedit_	developers pull from the upstream dev repo	04:44
	and push changes to there	04:45
jrnieder	is sqa where the hotfixes happen?	04:45
	← kukks left	04:45
jrnieder	ah, nvm, i'll let you finish :)	04:45
regedit_	production is where hotfixes may happen	04:45
	(i'm done atm)	04:45
jrnieder	what is sqa for?	04:45
	→ angelsl joined	04:46
variable	jpalmer: poke	04:46
regedit_	since the upstream dev repo is bare, the sqa (quality assurance) pulls from the bare dev repo in order to checkout and test	04:46
jrnieder	I see	04:47
	and devs get the eclipse metadata in addition to the production code	04:47
regedit_	yes it would be ideal if developers pulled from the upstream dev repo which had the eclipse crap included	04:47
	→ beautifulmind joined	04:48
regedit_	whereas by the time stuff reaches (sqa or for sure) production repo, those crap files should not be there	04:48
jpalmer	variable: stop poking me!	04:48
variable	jpalmer: may I pm you then?	04:48
jpalmer	I was joking ;) but sure.	04:48
variable	jpalmer: hehe.	04:49
jpalmer	fair warning: I'm headed to bed in like 15 though ;)	04:49
jrnieder	regedit: I would be tempted to not track the eclipse stuff in the repo at all	04:49
	to have a separate repo for that and a script that new devs get to grab it	04:49
	→ stickyboot joined	04:50
jrnieder	but I'm probably making some wrong assumptions here --- e.g., when you add a new file to the site, does the eclipse metadata have to change for that?	04:50
	if so, my separate repo idea would be a pain in the neck	04:50
regedit_	jpalmer: http://www.youtube.com/watch?v=MXgr6SYYNZM&t=24	04:50
	← Swimming_Bird left	04:51
regedit_	jrnieder: i don't think the eclipse files change like that	04:51
	jrnieder: like when you add files or such	04:51
jrnieder	ah, that's good to hear	04:51
regedit_	they do change of you modify some eclipse specific settings and paths etc	04:52
	*if you	04:52
jrnieder	do those change very often?	04:52
	are those even sharable?	04:52
	I guess all the devs have a similar working environment	04:52
regedit_	actually, any changes that occur to those files - i don't ever want comitted / pushed to upstream	04:52
jrnieder	yeah, so then I'd suggest having a tarball (or git repo if they ever do change) with just those files	04:53
	and telling devs, "to get started, clone this repo and extract this tarball"	04:53
regedit_	hmm...	04:53
jrnieder	.gitignore can ignore them	04:53
regedit_	i see	04:54
	← fpcfan left	04:57
	→ caseymcg joined	04:58
	→ fpcfan joined	04:59
regedit_	ye i think i'll go with that jrnieder, devs will have to figure it out on their end... it's better this way, what if they don't like eclipse? ;)	05:03
	← dna left	05:03
	← kermit left	05:05
	→ kermit joined	05:05
regedit_	jrnieder: let's examine another little part of our gitignore research (for the kernel.org patch) ;	05:05
	← fpcfan left	05:06
regedit_	jrnieder: you said quote ".jpg will match jpeg files throughout the repository, while /.jpg will only match jpeg files in the toplevel directory"	05:06
	jrnieder: is the same true for directories?	05:06
	→ fpcfan joined	05:07
jrnieder	yep	05:07
	a '/' at the end of a pattern just means "match directories only" and does not affect the pattern in any other way	05:07
regedit_	will 'foo/' match any directory named foo, wherever this foo directory may be in the repo (/foo/, /dir1/foo/, dir2/dir3/foo/ )	05:07
	aha	05:07
jrnieder	but "foo" does not match a file named foo throughout the repo	05:08
regedit_	what would be an example of "any other way"	05:08
	→ eephillip joined	05:09
jrnieder	in fact, I don't think it's possible to write a gitignore pattern that will match a file named foo throughout the repo	05:09
	← d4rk1ink left	05:09
jrnieder	here, let me try something	05:09
regedit_	wait really? that's weird..	05:09
jrnieder	no, you're right	05:09
	the documentation's wrong	05:09
	gah	05:09
	I think the only way I'm going to find the rules here is by looking at the code :/	05:10
regedit_	or brute force testing? :)	05:10
	we should make a test repo, with all sorts of foo folders and bar files	05:11
	and .gitignore the hell out of it in every conceivable way	05:11
jrnieder	well, that's how I found out that you're right and the doc seems to be wrong...	05:11
	but dir.c is not so hard to read :)	05:11
	→ haseeb joined	05:12
jrnieder	anyway, to answer your question, yes, 'foo/' will match a directory named foo, wherever the foo is in the tree	05:12
regedit_	wow awesome	05:12
	← orafu left	05:13
	→ orafu joined	05:13
regedit_	and 'foo' will match a file or folder? or both?	05:13
jrnieder	both	05:13
	if you want just files, use "foo" followed by "!foo/"	05:13
regedit_	holy crap that is AWESOME	05:13
	it's a shame someone realy put good thought into .gitignore functionality, but didn't document it well..	05:14
jrnieder	yeah, thanks for the examples	05:14
	still working on the strawman manpage change :)	05:15
regedit_	awesome :D please show me when it's ready	05:15
	or, will i get it to my email?	05:15
	← codetroll left	05:15
	→ Chib joined	05:16
	→ Prasad joined	05:17
	→ dna joined	05:17
	→ the-fonz joined	05:17
	← eephillip left	05:18
EugeneKay	I think its a shame gitignore doesn't do regex	05:18
	← CannedCorn left	05:18
	← the-fonz left	05:22
	→ OOPMan joined	05:25
	→ aspotashev joined	05:27
	→ jesteracer joined	05:27
	→ mattalexx joined	05:31
	← regedit_ left	05:35
	← Kazii left	05:39
mattalexx	I just made some changes in a detached-head submodule, committed them, cd'ed to the root of the parent repo, rand "git submodule update". When I went back to the submodule, my changes were gone and the commit was gone. Did it really just do that? Usually git will make you add -f somewhere in order to delete work..	05:42
	→ aspotashev\| joined	05:43
	→ freimatz2 joined	05:44
	← aspotashev left	05:44
kevlarman	mattalexx: your commit is still in the reflog of that submodule	05:45
	mattalexx: and since submodules always have a detached head, you shouldn't be developing in them	05:45
	mattalexx goes to google "reflog"	05:45
kevlarman	mattalexx: man git reflog	05:45
jast	mattalexx: automatic message the 'git-reflog' manpage can be found at <http://git.kernel.org/?p=git/git.git;a=blob_plain;f=git-reflog.html;hb=html>	05:45
EugeneKay	mattalexx - submodules are supposed to be read-only, IIRC	05:45
	Don't work i them	05:46
	in them	05:46
mattalexx	kevlarman, Yes, I've heard that. But how should I work on them if I can't put them into a project and give them context?	05:46
	This one, for example, is a plugin. Without the parent repo, it doesn't work	05:47
kevlarman	mattalexx: 2 choices	05:47
	1) don't use submodules	05:47
	2) clone the repo on its own to do development on it	05:47
	( 1a) you may find git slave interesting)	05:48
	← haseeb left	05:48
mattalexx	kevlarman, So everytime I need to test, I push from the standalone repo and pull from the submodule>	05:48
	?	05:48
	Thanks, I'll look into it	05:48
	← mikepack left	05:50
	→ mikepack joined	05:51
mattalexx	kevlarman, Oh thank God for reflog. That was freaking me out a little bit.	05:52
kevlarman	mattalexx: git makes it very difficult to destroy your code accidentally	05:52
mattalexx	kevlarman, Yes, that's why I was only a little freaked out. I figured it would be somewhere.	05:53
	kevlarman, Plus I was using git add -p to add my changes, and the diff was still in my CLI scrollback buffer as a last resort.	05:54
f00li5h	mattalexx:	05:54
mattalexx	f00li5h, Yes?	05:54
f00li5h	o_O	05:54
	sorry, mistab	05:54
	← mikepack left	05:55
	→ phaedrix joined	05:55
	← aspotashev\| left	05:59
	← ziggles left	06:00
	→ hwrdprkns_ joined	06:00
	← hwrdprkns left	06:01
	← jacknagel left	06:01
	→ hwrdprkns joined	06:01
	→ tiglog joined	06:04
	← hwrdprkns_ left	06:05
	→ hwrdprkns_ joined	06:07
	→ d4rk1ink_ joined	06:07
	→ w09x joined	06:07
	→ marthinal joined	06:08
	→ ThomasLocke joined	06:09
	→ dSebastien joined	06:09
	← beautifulmind left	06:09
	→ hwrdprkns__ joined	06:09
	← dSebastien left	06:10
	← hwrdprkns left	06:10
	→ airborn joined	06:11
	→ drake01 joined	06:12
	← subbyyy left	06:12
	← hwrdprkns_ left	06:13
	← neurodrone left	06:14
	← eno left	06:17
	← lunchub left	06:19
	→ eno joined	06:19
	← sangeeths left	06:21
	→ lunchub joined	06:22
	→ Dave^\| joined	06:25
	→ harsh joined	06:27
	← baburdick left	06:28
	← gianlucadv left	06:29
	← Prasad left	06:29
	→ nyuszika7h joined	06:31
	→ sangeeths joined	06:33
	→ codetroll joined	06:34
	→ TrineLocke joined	06:34
	→ Prasad joined	06:35
	← Prasad left	06:35
	→ ketas-ts joined	06:36
	← gigamonkey left	06:36
	→ Prasad joined	06:36
	← harsh left	06:37
	← drev1 left	06:40
	← alc left	06:41
	← Turicas left	06:43
	→ Turicas joined	06:43
	← ceej left	06:44
	→ shruggar joined	06:45
	← Cogito left	06:47
	→ Cogito joined	06:47
	→ flyz_ joined	06:48
	→ na3r joined	06:49
	→ slonopotamus joined	06:50
	→ harsh joined	06:51
	← Turicas left	06:52
	← harsh left	06:52
	← __main__ left	06:54
	→ mishok13 joined	06:56
	→ __main__ joined	07:03
	← ouyes left	07:04
	→ harshpb joined	07:05
	← shruggar left	07:06
	→ hwrdprkns joined	07:10
	→ stoffus joined	07:12
	→ vdv joined	07:12
	spyraway → napdragon	07:12
	← TheMoonMaster_ left	07:13
	← slonopotamus left	07:13
	← hwrdprkns__ left	07:13
	← hwrdprkns left	07:15
	→ hwrdprkns joined	07:17
	→ amigojapan joined	07:17
	→ artib joined	07:18
	→ qian_pc joined	07:18
	→ ouyes joined	07:20
	→ cbreak-work joined	07:20
	← mishok13 left	07:22
	← lunchub left	07:24
	→ meebey joined	07:24
	→ stodan joined	07:25
	→ aidenhong joined	07:26
	→ ReekenX joined	07:29
	→ psoo joined	07:29
	← psoo left	07:29
	→ nicoulaj joined	07:31
	← dreiss left	07:32
	→ henrikbjorn joined	07:35
	← qian_pc left	07:39
	← caseymcg left	07:39
	→ flijten joined	07:39
	← imarcusthis left	07:39
	← amigojapan left	07:40
	← Prasad left	07:42
	→ the-fonz joined	07:43
	→ jceb joined	07:43
	→ ph^ joined	07:44
	← quadz left	07:45
	← lnostdal_ left	07:46
	→ GeorgeSebastian joined	07:47
	→ Prasad joined	07:47
	→ mishok13 joined	07:49
	← Chib left	07:50
	→ dangerousdave joined	07:52
	← ph^ left	07:52
	← dangerousdave left	07:52
	→ lnostdal_ joined	07:52
	→ ph^ joined	07:53
	← GeorgeSebastian left	07:53
	← Praise left	07:54
	→ Praise joined	07:54
	← vdv left	07:56
	← airborn left	07:56
	← mishok13 left	07:57
	→ ipalaus_ joined	07:57
	→ flyz- joined	07:59
	→ ph^_ joined	08:00
	← artib left	08:00
	dkannan\|away → dkannan	08:01
	→ Turicas joined	08:01
	← flyz_ left	08:02
	→ nax joined	08:03
	→ dc5ala joined	08:04
	← ph^ left	08:04
	→ vdv joined	08:05
	→ GeorgeSebastian joined	08:06
	← drake01 left	08:10
	← alecthegeek left	08:11
	← dkannan left	08:12
	→ qian_pc joined	08:13
	→ dkannan joined	08:13
	→ jhoepken joined	08:13
	← qian_pc left	08:13
	→ artib joined	08:14
	→ mishok13 joined	08:15
	→ lut4rp joined	08:15
	← lut4rp left	08:15
	→ lut4rp joined	08:15
	← lnostdal_ left	08:15
	→ qian_pc joined	08:15
	→ ngulden joined	08:15
	→ saschagehlich joined	08:17
	→ berserkr joined	08:18
	→ opalka joined	08:19
	← opalka left	08:19
	→ opalka joined	08:19
	← qian_pc left	08:21
	← lut4rp left	08:22
	→ JohnBat26 joined	08:22
	→ Dave^\|\| joined	08:22
	→ RaceCondition joined	08:23
	→ lut4rp joined	08:23
	← lut4rp left	08:23
	→ lut4rp joined	08:23
	← gerard0 left	08:24
	→ jamiemill joined	08:25
	← Dave^\| left	08:25
	→ garrys_game joined	08:26
	→ rendar joined	08:26
	→ heliom joined	08:27
	← stickyboot left	08:27
	← RaceCondition left	08:28
	→ lnostdal_ joined	08:29
	→ eijk joined	08:29
	→ Codif joined	08:29
	← harshpb left	08:30
	← heliom_ left	08:30
	← cooloney left	08:31
	← flyz- left	08:32
	→ alansaul joined	08:32
	→ cooloney joined	08:32
	← Prasad left	08:33
	← jamiemill left	08:33
	→ Prasad joined	08:33
	→ qian_pc joined	08:34
	← qian_pc left	08:34
	→ qian_pc joined	08:35
	→ psoo joined	08:36
	→ _kud joined	08:36
	→ teweWork joined	08:36
	→ jmah joined	08:37
	→ childbear joined	08:45
	→ Sitethief joined	08:46
	→ Spockz joined	08:46
	→ harshpb joined	08:47
	→ _Qman joined	08:50
	→ pixelboy_ joined	08:51
	→ safinaskar joined	08:51
	→ wbednarski joined	08:52
	→ thiago joined	08:54
	← ziro` left	08:55
	→ marekw2143 joined	08:55
	→ ziro` joined	08:56
	→ psakrii joined	08:56
marekw2143	hi, doing git add . does add all directories but no their subdirectories/subfiles	08:57
	← mitsuhiko left	08:59
thiago	are those subdirectories ignored?	08:59
marekw2143	thiago: no, my .gitignore has only one line: lib/	09:00
jrnieder	marekw2143: pastebin it and someone might have ideas	09:01
	→ shruggar joined	09:01
thiago	marekw2143: and the subdir in question isn't lib?	09:01
marekw2143	no	09:01
	← freimatz2 left	09:03
jrnieder	hm, let's see if this works: !repro	09:03
jast	hm: automatic message please paste (e.g. using http://git.pastebin.com/) a transcript of your terminal session in which the problem occurred so that we can troubleshoot more easily.	09:03
jrnieder	ah, yay	09:04
cbreak-work	marekw2143: type git status	09:04
marekw2143	http://paste.ubuntu.com/730798/	09:04
cbreak-work	does it show the directories you want to add but are not added as untracked?	09:04
	← BlindOracle left	09:04
	→ hurikhan\|Work joined	09:04
jrnieder	ah, submodules	09:04
	→ nudb2 joined	09:05
marekw2143	so both dirs: django-starter and django_jsonservice are added to repo, but their content isnt	09:05
	← babilen left	09:05
jrnieder	I don't think git add --recurse-submodules exists yet	09:05
	anyone know a good submodule tutorial?	09:05
cbreak-work	it makes no sense to recurse into submodules	09:06
jrnieder	sure it could	09:06
	→ mklappstuhl joined	09:06
jrnieder	git add --recurse-submodules would have to auto-commit or something, though	09:06
marekw2143	but it's such a common thing adding recursively all files/dirs - how to acheive that?	09:06
jrnieder	it would be weird	09:06
thiago	no	09:06
	not with submodules	09:06
cbreak-work	it makes no sense	09:06
	→ padams joined	09:06
cbreak-work	submodules are separate repositories	09:06
	they have a separate index	09:06
	everything about them is separate	09:07
marekw2143	it's strange	09:08
thiago	yes, it is	09:08
marekw2143	because when I entered into one directory, and then did: git add .	09:08
thiago	but that's how it is	09:08
marekw2143	then, all subdirectories and subfiles of that dir were added	09:08
	why didn't it do at first add ?	09:08
jrnieder	markew2143: submodules are an odd beast	09:08
	my advice is to learn about them first	09:09
marekw2143	why submodules?	09:09
jrnieder	because your repository has submodules	09:09
	they are not just subdirectories	09:09
marekw2143	ohh, true !	09:09
	I've forgotten - my subprojects have separate .git directory, yes	09:10
cbreak-work	everything about them is separate	09:10
	← Prasad left	09:11
	→ Prasad joined	09:11
	← ReekenX left	09:11
	→ phluks joined	09:12
	→ kriegerod joined	09:12
marekw2143	uff, thank you all for help	09:12
	← alansaul left	09:12
	← w09x left	09:12
	→ _md joined	09:13
	→ Prateekchy joined	09:15
	→ ReekenX joined	09:15
marekw2143	do you know some public free git server which allows private repositories (not visible to others)?	09:16
	→ whitman joined	09:16
	← lnostdal_ left	09:17
phluks	marekw2143: www.assembla.com - though I dont have extensive experience with it.	09:17
m1sc	marekw2143: bitbucket?	09:17
marekw2143	ooo, nice	09:17
	→ BlindOracle joined	09:17
marekw2143	which one is better ?:)	09:17
	→ Shusshu joined	09:18
	→ beautifulmind joined	09:19
	← yann2 left	09:19
	→ drake01 joined	09:20
	→ freimatz2 joined	09:21
	← _kud left	09:22
phluks	bitbucket looks shinyer :)	09:23
	→ PigeonFriend joined	09:24
	→ gusnan joined	09:25
	→ babilen joined	09:28
marekw2143	yeah, and I think it's more reliable than assembla in offering free plan	09:29
	← ReekenX left	09:30
	← Prateekchy left	09:32
	→ ReekenX joined	09:32
	→ _kud joined	09:32
	→ Trezker joined	09:34
	→ canton7 joined	09:34
	→ khmarbaise joined	09:37
	→ achadwick joined	09:37
	← the-fonz left	09:37
	→ envi_ joined	09:38
	→ jamiemill joined	09:38
	→ evildmp joined	09:39
	→ tvw joined	09:40
	→ iknite joined	09:40
	← ReekenX left	09:41
	→ solofight joined	09:42
	→ muv joined	09:42
solofight	people git not available for mac os lion version ? http://code.google.com/p/git-osx-installer/downloads/list?can=1&q=&colspec=Filename+Summary+Uploaded+Size+DownloadCount+UploadedBy	09:42
	or am i looking at the wrong place ?	09:43
	→ tatsuya_o joined	09:43
rue	Homebrew, at least	09:43
	→ alecthegeek joined	09:44
solofight	rue, @ me ?	09:44
	→ ReekenX joined	09:44
rue	solofight: Ah, yes. Dunno about other methods	09:45
thiago	solofight: install the snow leopard one	09:45
	→ dharrigan joined	09:46
	← dharrigan left	09:46
	→ dharrigan joined	09:46
	← Codif left	09:46
	← harshpb left	09:46
solofight	thiago: you mean the same dmg file for snoe leopard will work for lion as well ?	09:46
	← canton7 left	09:46
thiago	solofight: yes	09:46
	→ Codif joined	09:46
thiago	solofight: "snowleopard" in the name indicates the minimum version	09:46
solofight	thiago: oh, ok ! but shouldn't it say that snowleopard and above or something like that !	09:47
thiago	in the file name? no	09:47
	→ dr_lepper joined	09:47
	← dr_lepper left	09:47
	→ dr_lepper joined	09:47
	→ brian_g joined	09:48
solofight	thiago: people get confused because there is a installer for leopard and then snow leopard which makes them look for a installer for lion	09:48
	?	09:48
	← jrnieder left	09:48
	← fisted_ left	09:49
	→ harshpb joined	09:49
thiago	I don't see an installer for leopard and one for snow leopard, for the same Git version	09:49
solofight	thiago: ohhhh	09:50
thiago	they're all Leopard, except for the ones from the past 4 months	09:50
solofight	thiago: are you a mac user ? one can upgrade their OS for free of charge ? or it costs ?	09:50
thiago	I'm not a Mac user	09:51
solofight	thiago: ok thank you very much for the help. will try what you have sugested	09:52
	→ fisted joined	09:52
solofight	thiago:++	09:52
	thiago: ++	09:52
	← replore left	09:52
solofight	Gitzilla: thiago ++	09:52
	→ iPHP joined	09:52
solofight	:(	09:52
sitaram	solofight: http://www.stickycomics.com/computer-update/	09:53
solofight	thank you rue	09:53
	hey sitaram :) thats a good one	09:54
	← Praise left	09:54
	→ bonhoffer joined	09:54
	→ Praise joined	09:55
	← eean left	09:55
solofight	sitaram: good site btw, thanks for letting me know	09:56
	solofight taking a break and browsing more comics @ stickycomics	09:56
	→ eean joined	09:57
	Silex- → Silex	09:58
sitaram	:)	09:59
EugeneKay	Sneeze	10:00
	→ akamike joined	10:01
	→ kalusn joined	10:01
	→ glcrazy joined	10:02
	→ glcrazy_ joined	10:02
	→ endy_c joined	10:03
	← drake01 left	10:03
	← xiehong left	10:04
endy_c	What is the pro/con of using SSH compared to HTTPS?	10:04
	→ alecat joined	10:04
thiago	endy_c: ssh works	10:04
	← qian_pc left	10:04
wereHamster	both work	10:04
thiago	using the new magic http, yes	10:05
	the old dumb transport was really bad	10:05
FauxFaux	endy_c: SSH is likely to be significantly easier to set-up, especially if you already have unix aware users.	10:06
	← opalka left	10:06
cmn	if your uses don't know about SSH and SSH keys, then HTTPS might be easier for them, because they just have to remember a user/pass combination	10:07
endy_c	ok got it, thanks all	10:07
nevyn	cmn: if you have a smallish number of users. then just setting it up for them is probably less grief than dealing with passwords	10:09
	I really hate passwords	10:09
	← nudb2 left	10:09
	← ReekenX left	10:09
	→ ReekenX joined	10:09
cmn	sure, if you can go around the office and do it for them and then tell them it works magically, that might be the easiest option	10:10
solofight	nevyn: i can manage them for you :)	10:10
	→ opalka joined	10:10
solofight	cmn: cant you route all the ssh requests with a same key or something ?	10:11
FauxFaux	We have both SSL client certs and ssh keys here. Explaining to users that there's two different kinds of magic and only one of them breaks on a regular basis (the SSL ones, obviuosly) is great fun.	10:11
	solofight exploring the possibility of routing an ssh request via a middleman	10:11
cmn	solofight: set up the same key for all users? sure, but then you can't use permissions	10:11
solofight	cmn: then identify the key based on their system login id and attach them dynamically with their ssh request	10:12
	?	10:12
	using username and password is not recommended by several experts	10:12
	when it comes to ssh	10:13
cmn	that sound extraordinarilly contrived; you want to MITD all your users so you can dynamically allocate them a key?	10:13
	MITM*	10:13
solofight	endy_c: howd you even compare ssh with https !	10:14
nevyn	FauxFaux: we don't do certificate or key management very well at a organisational level	10:14
	we're barely competent at password management	10:14
	→ guardian joined	10:14
EugeneKay	Post it notes under keyboards ftw	10:14
cmn	solofight: oh, I see, you want to have them login to a computer with user/pass and that computer would have their "real" keys in it... I don't think that would solve the problems with SSH passords	10:15
solofight	cmn: yeah the level of stupidity/usability depends on your situation	10:15
	← wanders left	10:15
solofight	correct	10:15
guardian	hello, i'm using git rebase -i to edit a commit. how can i undo commited changes with git ammend? i'm trying to edit a commit so that commit doesn't modify a file but i want to keep other changes	10:15
FauxFaux	nevyn: Basically the only benifit of an enterprise (1,000+ person) organisation is that you have to not have your entire IT team spending their entire life on passwords. And then they have a Domain and misconfigure it, woop.	10:15
solofight	cmn: why not ?	10:15
endy_c	solofight: I have this option for accessing bitbucket	10:16
FauxFaux	guardian: 'e'dit the commit, checkout HEAD # to revert the file you don't want, then rebase --continue.	10:16
nevyn	FauxFaux: dude.. I work for a 60,000 seat organisation	10:16
	← _kud left	10:16
FauxFaux	Pwned.	10:16
thiago	guardian: git reset -p HEAD~	10:16
nevyn	FauxFaux: user passwords are under control it's passwords for things like root, oracle, and tibco that suck	10:16
guardian	git reset -h HEAD~ path/to/file would work?	10:17
	git reset -p HEAD~ path/to/file would work?	10:17
	sorry for typo	10:17
thiago	guardian: yes	10:17
solofight	endy_c: oh so you mean advantages of using ssh over https or vice versa when it comes to pushing code to a server.. ?	10:17
thiago	if it's the entire file, you can drop the -p	10:17
endy_c	solofight: yep	10:17
guardian	it's the entire file	10:17
thiago	guardian: reset it like that, commit --amend	10:18
	guardian: then decide what to do with the changes	10:18
	← envi_ left	10:18
cmn	thiago: they'd still have to use a password, which makes it attackable from a package-timing front	10:19
thiago	cmn: they who?	10:19
cmn	your users	10:19
	thiago doesn't have users	10:20
	→ Chib joined	10:21
cmn	thiago: sorry, I meant solofight	10:21
solofight	endy_c: i think turn around time comes into play. stay here for deep view point from experts	10:21
	→ alansaul joined	10:21
guardian	thiago: thx; would have FauxFaux suggestion worked as well?	10:21
thiago	guardian: no	10:21
	← teukka left	10:22
	→ teukka joined	10:22
thiago	guardian: revert would have created a new commit	10:22
	guardian: unless you rebase again and squash the two commits together	10:22
guardian	ok thx	10:23
	i didn't want a new commit; just rebase -i and remove a file's changes before push	10:23
	→ _kud joined	10:23
	← aidenhong left	10:25
solofight	cmn: cant you make them as thin clients who logs into a central machine the moment they switch on and then they connect through ssh from that machine ?	10:26
	cmn: if i understood your problem correctly ^	10:26
endy_c	solofight: yep, will definitely stay around to listen	10:26
cmn	that wouldn't help against the timing attacks	10:26
	when you use a password with SSH, it sends a packet for each letter (because humans are slow) so looking at the time difference between those packets, you can reduce which passwords they may be using	10:27
	solofight reading more about time attacks	10:27
thiago	no it doesn't	10:27
	well, not the login password	10:28
	if it's after the connection was established, then yes, one packet per letter	10:28
cmn	hm, I thought I read that some people had managed to do this with the SSH login pw	10:28
solofight	cmn: nooo, am talking about using an rsa key with 1024 or 2048 bits for ssh	10:28
cmn	ah, right, you want them to use that and change that key from under them at some point	10:29
	→ mrsrikanth joined	10:29
thiago	the login password isn't sent until you press the Enter key	10:29
solofight	thiago: a typical example of what you said - sudo ing in the remote machine after gaining ssh access as normal user ?	10:29
thiago	solofight: yes	10:30
solofight	thiago: aint the communication encrypted ?	10:30
thiago	solofight: it is	10:30
FauxFaux	"OpenSSH_5.8p1 Debian-7ubuntu1, OpenSSL 1.0.0e 6 Sep 2011" doesn't send a packet until you return the password here, cmn, according to wireshark.	10:31
	← MacGyverNL left	10:31
cmn	thiago: ok, the internet says you're right :) it looks like I misunderstood the attack	10:31
solofight	lets say for each character user has pressed the encryption algo sends 1024 bits => aint that a huge number to decode to guess a password ?	10:31
thiago	it is	10:31
	I never claimed you could decrypt the session	10:31
	that was cmn	10:31
solofight	alright	10:32
cmn	I didn't say you could decrypt the session, I thought you could guess the password	10:32
FauxFaux	solofight: The implication is that you can work out how much stuff someone is typing in by how many packets there are; which you certainly can with my irssi setup.	10:32
thiago	a "timing attack" could try to guess the password from the timing of the packets, applying some heuristics based on how fast you can type it	10:32
cmn	(which does sort-of amount to the same...)	10:32
solofight	so cmn timing attack are useless even performed unless they are using a super computer to decrypt your codes	10:32
FauxFaux	solofight: a) Supercomputers can't even remotely break any sensible encryption algorithm, as far as we know. b) No, any information about anything counts as a cryptographic breach.	10:33
solofight	FauxFaux: ok so its going to be a guess with hell lot of possibilities... what happens after that ?	10:33
cmn	solofight: no, because the timing of the packets (after logging in) put together with a typing heuristic gives you a smallish (relatively) amount of keypresses that the user might have done	10:34
	→ MacGyverNL joined	10:34
	← MacGyverNL left	10:34
	→ MacGyverNL joined	10:34
FauxFaux	solofight: I only use three passwords, either 'a', 'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa' or 'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa'. My username is [email@hidden.address] and I use ssh. You have one attempt to get the right password, or I'll shoot you in the head. What do you need?	10:34
	← evildmp left	10:35
solofight	FauxFaux: number of key strokes	10:35
	→ Vile joined	10:36
FauxFaux	Which, if ssh wasn't implemented sensibly, you could get via. this timing attack, without breaking the encryption, right?	10:36
	→ sweb joined	10:36
	→ Wooga joined	10:36
solofight	If i know the pool of possible values - yes ?	10:36
FauxFaux	Exactly.	10:37
	In fact, you can probably get it even with the current ssh implementation, as the time taken to type the longer password is significant and it's obvious when the communication is paused for password exchange.	10:37
cbreak-work	FauxFaux: http://www.huffingtonpost.com/2011/10/27/iphone-4-accelerometer-spy_n_1035193.html	10:37
	sidechannel attacks ftw	10:37
FauxFaux	Genius. My favourite recent one was that recent SCHNEEEIR post about usb ports.	10:38
cmn	nice	10:38
	← marekw2143 left	10:38
solofight	FauxFaux: but how does a attacker get the possible list of values ? by tracing the users activities online or something ?	10:38
	→ NetHawk joined	10:38
cbreak-work	http://www.schneierfacts.com/?	10:38
cmn	I remember some people managed to do the same by pointing a laser to a laptop screen, this is cooler	10:38
	← bonhoffer left	10:39
	solofight googling for irssi setup which FauxFaux mentioned	10:39
FauxFaux	cbreak-work: http://syhw.posterous.com/two-amusing-side-channel-attacks	10:40
	→ nudb2 joined	10:41
	→ TAsnw joined	10:41
	→ w0bni joined	10:42
	→ wanders joined	10:42
cbreak-work	neat	10:42
	if they had used FireWire, they could have just read the key schedule and so on out of the memory...	10:42
solofight	FauxFaux: where do you get all these updates ? i mean is there a single source to update myself ?	10:42
FauxFaux	Tee hee.	10:42
cbreak-work	schneier has a newsletter	10:42
FauxFaux	And a blog that posts significantly faster than I can search it for stories.	10:43
solofight	cbreak-work: http://www.schneierfacts.com/ or this http://syhw.posterous.com	10:44
	← wbednarski left	10:44
FauxFaux	SPEAKING OF SECURITY, WHERE'S MY SHA-256 PORT OF GIT?	10:44
TAsnw	hey guys. I have an svn repo, and a public git svn checkout. Is it possible to pull from the public git-svn checkout and then git svn dcommit from my copy? It feel non trivial because the commit hashes are different (different checkout paths, the public http:// vs my ssh+svn://). Has anyone stumbled across anything similar? Thanks.	10:44
FauxFaux	You can fetch then rebase --onto or cherry-pick the commits across, but there's no automation assistance afaik.	10:45
	← khmarbaise left	10:45
solofight	thanks guys for knowledge upgrade :)	10:46
	good day to all	10:46
TAsnw	FauxFaux, I'm trying to avoid the initial fetch from svn	10:46
	which is super slow	10:46
	← solofight left	10:46
TAsnw	because git-svn iterates over all the commits.	10:46
	→ jorn joined	10:46
jorn	hiho, i'm stuck in a 100 % cpu git pull... any idea what to do?	10:47
	← Praise left	10:47
	→ Praise joined	10:48
jorn	oh ah, it just finished... still weird, took 5 minutes for a 21 mb repo!?!	10:49
sweb	what is problem : ssh_exchange_identification: Connection closed by remote host	10:49
	fatal: The remote end hung up unexpectedly	10:49
	?	10:49
cmn	jorn: do you have many references?	10:49
	→ empity joined	10:50
	← alecthegeek left	10:50
jorn	cmn: 1	10:50
	so i'd say no	10:50
cbreak-work	sounds like ssh/network problems	10:51
	← tiglog left	10:51
	→ Kazii joined	10:51
cmn	jorn: that's odd; that kind of thing tends to happen with many branches/tags; next time it happens, try to attach gdb	10:51
jorn	cmn: could the problem be caused by very many small files?	10:53
	(the other guy checked in the whole eclipse .metadata folder)	10:53
cmn	could be, are you runnign on NFS or some other slow FS?	10:53
	→ Ashfame joined	10:53
	← Praise left	10:54
	→ Praise joined	10:54
jorn	locally	10:55
	clone also takes very long	10:55
	i attached gdb, what then?	10:55
	← Turicas left	10:56
cmn	if it's a recent-ish git, it could be due to git checking that all is connected, which could take a long time with many new objects	10:56
	jorn: when it runs at 100% CPU again, look at the backtrace	10:56
jorn	(get a lot of warnings that no debug info is available, as not compiled in)	10:56
	→ alecthegeek joined	10:56
cmn	you'll get that if you're using the binaries from your distro	10:56
	← Ashfame left	10:57
jorn	close() <- write_entry() <- checkout_entry() <- unpack_trees() <- cmd_clone() <- handle_internal_command() <- main()	10:58
	← Chib left	11:00
	→ Chib joined	11:01
	→ pimpin joined	11:01
pimpin	Iop	11:01
	Someone to give an hand to a beginner please ?	11:02
	→ workmad3 joined	11:02
cmn	jorn: hm, it looks like it's doing the checkout there	11:03
	→ tiagobutzke joined	11:03
pimpin	My issue could come from the old GIT version (1.4.4.4) installed in the server I use	11:03
	→ EnderMB joined	11:03
pimpin	but I can't update it	11:04
FauxFaux	pimpin: WHAT IS THE NATURE OF YOUR GIT EMERGENCY?	11:04
	← alnewkirk left	11:04
	← iknite left	11:04
pimpin	FauxFaux: not a real emergency	11:04
	→ DarkAR joined	11:05
cmn	you still haven't told us what's wrong	11:05
EnderMB	Hey all. I have a bunch of settings files and folders that I want to ignore, all with a full-stop as a prefix (e.g. .settings, .logs, etc). I see that it's easy enough to stop a file with a certain extension, but how do I ignore all files and folders with a certain prefix?	11:05
pimpin	just a problem I can't solve since this morning	11:05
FauxFaux	EnderMB: Wildcards work anywhere, '.*' will be fine.	11:05
EnderMB	FauxFaux: Awesome, thanks for the help!	11:06
cmn	jorn: if you check at different times, you might see a pattern (or the 'perf' tool)	11:06
pimpin	yes sorry… tks for helping me.	11:06
	← nudb2 left	11:07
FauxFaux	pimpin: Arrrrgh! Your question is entirely useless. Fill in the blanks: "I tried [X] in [Y]. I got [A] but I was expecting [B]. The simplest way I can think to reproduce this is [C]." Reply on one line. If it doesn't fit, use a pastebin. Do not even think of speaking before fully filling in the question.	11:07
	→ wbednarski joined	11:07
	→ setmeaway joined	11:08
	→ noisebleed joined	11:09
	← noisebleed left	11:09
	→ noisebleed joined	11:09
pimpin	I clone a rep, make changes, commit on local, push and look at my files pushed : they are unchanged. -> I look at log and see my commit well pushed but my files are already overwritten on the server with the old version.	11:10
FauxFaux	pimpin: You're pushing to a non-bare repository?	11:10
	← RichiH left	11:10
pimpin	that could be right, I do not specify option when creationg repo	11:11
FauxFaux	Pushing to a non-bare repository is discouraged as it makes no sense because it makes it look like your files are unchanged; it's not even worth working out why, just don't do it.	11:11
pimpin	I look for this option in the doc	11:11
FauxFaux	man git init	11:12
jast	automatic message the 'git-init' manpage can be found at <http://git.kernel.org/?p=git/git.git;a=blob_plain;f=git-init.html;hb=html>	11:12
FauxFaux	Except that the bot is broken. (jast)	11:12
pimpin	tks U very much FauxFaux	11:12
	→ pidus joined	11:13
	→ iknite joined	11:13
pimpin	is ti possible to transform my repo in a bare one ?	11:14
FauxFaux	git clone --bare oldrepo oldrepo.git	11:14
pimpin	Tks	11:15
	← cenuij left	11:17
	→ soc42 joined	11:17
	← JohnBat26 left	11:20
	→ sattu94 joined	11:23
	→ drev1 joined	11:23
	→ Aaaarg joined	11:24
louiz	(I like how pimpin managed to tell us 4 times that he had something to tell us, without ever telling what it is)	11:24
FauxFaux	This is why that alias exists. (/argh) :)	11:24
pimpin	Sorry I'm not used to IRC… and wondered if it was the right place to ask help for beginers	11:24
	→ pfsmorigo joined	11:24
louiz	aha, that’s an alias :D I thought that you were very comprehensive and took a lot of your time to write this sentence :D	11:25
	← rendar left	11:25
cmn	there's also !ask	11:25
FauxFaux	No, it comes up much more frequently than you'd think. ¬_¬	11:25
jast	automatic message yes, it's okay to ask questions here... in fact, you just asked one! ;) Pretty much any question is fine. We're not terribly picky.	11:25
louiz	:)	11:25
	← torarne left	11:25
	← kimico left	11:25
	← elyob left	11:25
cmn	there's also !ask	11:26
jast	automatic message yes, it's okay to ask questions here... in fact, you just asked one! ;) Pretty much any question is fine. We're not terribly picky.	11:26
cmn	sorry, wrong window	11:26
pimpin	tks guys	11:26
louiz	you’re welcome :o	11:26
pidus	I had made some changes to a file, committed and tried to push, forgetting to pull. There were conflicts, I fixed them, tried to push again but now it neither lets me push - 'To prevent you from losing history, non-fast-forward updates were rejected	11:26
	Merge the remote changes (e.g. 'git pull') before pushing again', nor pull - You have not concluded your merge (MERGE_HEAD exists).	11:26
	Please, commit your changes before you can merge, nor commit - 'fatal: cannot do a partial commit during a merge.'.	11:26
	← ipalaus_ left	11:26
	→ ipalaus_ joined	11:27
louiz	did you add the file that were resolved from the conflict?	11:28
	← drev1 left	11:28
	→ gwasshoppah joined	11:28
jmah	pidus: did you commit after fixing the conflicts?	11:28
louiz	he said he cannot commit	11:28
	← jorn left	11:29
jmah	ah	11:29
	→ cybersphinx joined	11:29
	← bauruine left	11:30
jmah	pidus: add your resolutions, commit, then push.	11:30
	→ elyob joined	11:30
	← pidus left	11:31
	→ pidus_ joined	11:31
	← pidus_ left	11:31
	→ pidus_ joined	11:31
	← the_cybersphinx left	11:31
louiz	pidus_: are you there?	11:31
pidus_	sigh…my connection had died. Wonder if I missed a reply	11:31
	← ph^_ left	11:31
	← alecthegeek left	11:32
louiz	+louiz \| did you add the file that were resolved from the conflict?	11:32
	and	11:32
	jmah \| pidus: add your resolutions, commit, then push.	11:32
pidus_	louiz: I did do an add	11:32
louiz	hrm	11:32
pidus_	and when I do commit it says "fatal: cannot do a partial commit during a merge."	11:33
	← wanders left	11:33
	← alansaul left	11:33
	← akamike left	11:33
	← dr_lepper left	11:33
	← nax left	11:33
	← meebey left	11:33
	← stoffus left	11:33
	← kerrick left	11:33
	← wks left	11:33
	← CpuID2 left	11:33
	← frogonwheels left	11:33
	← Ingmar left	11:33
	← Samoi left	11:33
	← [M]ax left	11:33
	← stamina left	11:33
	← Samual left	11:33
	← dwmw2_gone left	11:33
	← iivvoo left	11:33
	← Daniel0108 left	11:33
	← avtobiff left	11:33
	← phix left	11:33
	← SethRobertson left	11:33
	← jspiros left	11:33
	← milli left	11:33
	← oso96_2000 left	11:33
	← gebi left	11:33
	← armenb left	11:33
	← Dreamer3 left	11:33
	← twem2_ left	11:33
	← plopix left	11:33
	← twem2 left	11:33
	← pasky left	11:33
	← necromancer left	11:33
	← tokkee left	11:33
	← Tianon left	11:33
	← zibri left	11:33
	← Bombe left	11:33
	← AlexanderS left	11:33
	← ruskie left	11:33
	← EugeneKay left	11:33
	← Ratler left	11:33
	← JeffJohnson left	11:33
	← quentusrex left	11:33
	← ketas left	11:33
	← IcePic left	11:33
	← lamont left	11:33
	← nevyn left	11:33
	← deever left	11:33
	← Cerise left	11:33
	← chrisf left	11:33
	← aeris- left	11:33
	← comawhite left	11:33
	← Xeross left	11:33
	← Aikar left	11:33
	← jeremy_c left	11:33
	← cwillu_at_work left	11:33
	← kanru left	11:33
	← jeffhung left	11:33
	← xou left	11:33
	← cbz left	11:33
	← Fissure left	11:33
	← wolfric left	11:33
	← fahadsadah left	11:33
	← shachaf left	11:33
	← Caelum left	11:33
	← wereHamster left	11:33
	← RichardBronosky left	11:33
	← ciaranm left	11:33
	← mike_perdide left	11:33
	← tuxx- left	11:33
	← Kiall left	11:33
	← LotR left	11:33
	← district left	11:33
	← rdz left	11:33
	← zykes- left	11:33
	← joschi left	11:33
	← jfkw left	11:33
	← cantoma left	11:33
	← madduck left	11:33
	← awx left	11:33
	← prudhvi left	11:33
	← jakob left	11:33
	← dsop left	11:33
	← bfrog left	11:33
	← Fiouz left	11:33
	← tg left	11:33
	← GeorgeSebastian left	11:33
louiz	wow	11:33
	→ CpuID2 joined	11:33
	← CpuID2 left	11:33
	→ CpuID2 joined	11:33
jmah	you didn't do enough adding then	11:34
	→ Samual joined	11:34
jmah	check "git status"	11:34
	→ Dreamer3 joined	11:34
	→ ponch joined	11:34
	← elyob left	11:34
	← flijten left	11:35
	→ Samoi joined	11:35
	← Samoi left	11:35
	→ Samoi joined	11:35
	← shruggar left	11:35
	→ Bombe joined	11:35
	→ envi_ joined	11:35
	→ frogonwheels joined	11:36
pidus_	jmah: You are right "On branch master # Your branch and 'origin/master' have diverged, # and have 1 and 15 different commit(s) each, respectively." But I never touched the files that it says require commit	11:36
cmn	pidus_: don't try to commit just the one file, that won't wokr	11:36
	→ jeffhung joined	11:36
pidus_	cmn: you mean, do -a	11:36
cmn	sort of	11:36
	→ _aeris_ joined	11:36
jmah	or just "git commit"	11:36
	→ deever joined	11:37
	→ armenb joined	11:37
	→ meebey joined	11:37
	→ xou joined	11:37
	→ phix joined	11:37
	→ Aikar joined	11:37
pidus_	so, are these 15 files the ones that i fetched after trying to push	11:37
	→ bfrog joined	11:37
	→ cbz joined	11:37
	← bfrog left	11:37
	→ bfrog joined	11:37
cmn	if git diff gives you something you like (preferably nothing), then add -u then commit	11:37
	→ Fiouz joined	11:37
	→ gebi joined	11:37
	→ district joined	11:37
	→ jfkw joined	11:37
	→ stamina joined	11:37
	→ mike_perdide joined	11:37
	→ nevyn joined	11:37
	→ wereHamster joined	11:37
	→ ciaranm joined	11:37
	→ jeremy_c joined	11:37
	→ IcePic joined	11:37
	→ pasky joined	11:37
	→ twem2 joined	11:37
	→ wolfric joined	11:37
	→ Caelum joined	11:37
	→ nardev joined	11:38
	→ tokkee joined	11:38
	→ dsop joined	11:38
pidus_	cmn: git diff doesn't give me anything	11:38
cmn	those files would be the ones that are different between your branch tip and the remote's branch tip	11:38
	→ Guest93405 joined	11:38
	→ dr_lepper joined	11:38
	← dr_lepper left	11:38
	→ dr_lepper joined	11:38
	← Fiouz left	11:38
	→ tuxx- joined	11:38
	→ zibri joined	11:38
cmn	then add -u should be what you need	11:38
	→ Fiouz joined	11:38
	← tatsuya_o left	11:38
	← Fiouz left	11:39
	→ RichardBronosky joined	11:39
	→ Fiouz joined	11:39
	→ kanru joined	11:39
	→ alecthegeek joined	11:39
	→ floubidou joined	11:39
	← Fiouz left	11:39
	→ Fiouz joined	11:40
	← Fiouz left	11:40
	← Praise left	11:40
	→ jakob` joined	11:40
	→ Kiall joined	11:40
	→ Fiouz joined	11:40
pidus_	cmn: But the commit shows changes to those 15 files, something that I didn't do and wouldn't want to show in the commit :(	11:40
	→ tatsuya_o joined	11:40
	→ canton7 joined	11:41
	← Fiouz left	11:41
	→ akamike joined	11:41
	→ AlexanderS joined	11:41
	→ dwmw2_gone joined	11:41
	→ Praise joined	11:41
	→ tg joined	11:41
	→ fahadsadah joined	11:41
	← floubidou left	11:41
	→ Fiouz joined	11:41
	→ stoffus joined	11:42
	→ twem2_ joined	11:42
	→ Xeross joined	11:42
	→ joschi joined	11:42
cmn	what do you mean by "shows changes"	11:42
	→ shachaf joined	11:42
	→ Lucio85 joined	11:42
	→ lamont joined	11:42
	← tatsuya_o left	11:43
	→ milli joined	11:43
	→ comawhite joined	11:43
	→ tatsuya_o joined	11:43
	→ EugeneKay joined	11:43
	→ Cerise joined	11:43
	← Lucio85 left	11:43
pidus_	cmn: it shows delete/modified/new on a number of files that I never touched. I guess this is because I tried to commit and push to remote repo without doing a pull.	11:44
cmn	if the push didn't succeed, it didn't change anything anywhere	11:44
	→ kimico joined	11:44
cmn	that output might come from the merge	11:44
	← safinaskar left	11:45
cmn	I don't remember right now exactly what commit tells you when you commit a merge. Does 'git show' show any output?	11:45
	← alecthegeek left	11:45
	→ JeffJohnson joined	11:45
	← JeffJohnson left	11:45
	→ JeffJohnson joined	11:45
	→ Dave^\| joined	11:45
	→ plopix joined	11:45
	← soc42 left	11:45
	→ sanduz2 joined	11:45
	→ prudhvi joined	11:46
	→ ketas joined	11:46
	→ madduck joined	11:46
	→ ruskie joined	11:47
pidus_	cmn: show shows a single commit - 'Merge remote-tracking branch 'origin/master'' and "Conflicts" in the one file where there was a conflict before add	11:47
	→ quentusrex joined	11:48
	← quentusrex left	11:48
	→ quentusrex joined	11:48
pidus_	cmn: Can i revert this commit, and do another merge and push (this time without the conflict)	11:48
cmn	'show' only ever shows one commit; if it doesn't show a diff for anything other than the conflicted files, you're fine, you can push	11:48
	→ nax joined	11:48
	→ rdz joined	11:48
cmn	the conflict will still be there	11:48
	→ Ratler joined	11:48
	→ wks joined	11:48
	→ awx joined	11:48
	→ cantoma joined	11:48
	← harshpb left	11:49
	→ happylife joined	11:49
cmn	with git reset --hard HEAD^ you can put the worktree, index and HEAD to the commit before the merge	11:49
	← ponch left	11:49
cmn	but if you merge again, it will still conflict	11:49
	→ LotR joined	11:49
pidus_	and what with the 'Merge remote-tracking branch 'origin/master'?	11:49
	← heptadecagram left	11:49
	→ GeorgeSebastian joined	11:49
	→ Fissure joined	11:49
	← Praise left	11:50
	→ Praise joined	11:50
	→ SethRobertson joined	11:51
pidus_	this never appeared in a commit.	11:51
	← EricInBNE left	11:51
	← artib left	11:51
cmn	have you ever done a merge before?	11:51
	→ artib joined	11:51
	→ neurodrone joined	11:52
	→ chrisf joined	11:52
pidus_	cmn: yes. A fetch, followed by a merge. I do it before any commit/push	11:52
	forgot to do it this time round	11:52
	→ kerrick joined	11:52
cmn	then your repo must have low traffic; if the merge is a fast-forward (i.e. the history did not diverge) then it will just update the branch	11:53
	→ nixmaniack joined	11:53
cmn	if the history does diverge, you need to do a real merge	11:53
	→ Daniel0108 joined	11:53
cmn	and that's the default message for a merge	11:53
	← hyperair left	11:54
pidus_	oh ok	11:54
	→ fermo joined	11:54
	← tiagobutzke left	11:55
pidus_	and by real merge you mean git merge origin/master?	11:55
	→ jspiros joined	11:56
	← albel727 left	11:57
	→ leen^\| joined	11:57
	→ Tianon joined	11:58
	← Tianon left	11:58
	→ Tianon joined	11:58
	← Dave^\| left	11:58
	→ albel727 joined	11:58
	← GeorgeSebastian left	11:58
cmn	pidus_: a real merge is a merge that can't be fast-forwarded	11:58
	that is, where there are different commits in one and the other	11:59
	instead of one just being ahead of the other	11:59
	→ rendar joined	11:59
pidus_	cmn: ok. Not sure if I'm getting all of it. Got to do some reading of help.	12:00
canton7	http://sandofsky.com/images/fast_forward.pdf	12:00
	← shiba_yu36 left	12:01
	← endy_c left	12:01
	← empity left	12:01
wereHamster	canton7: nice document. is it yours?	12:02
	← Praise left	12:02
	→ Praise joined	12:02
canton7	wereHamster, sadly not. someone else (can't remember who) found it after a quick google	12:02
pidus_	thanks cmn and canton7. Got the fast_forward thing.	12:03
canton7	I liked it so I bookmarked it :)	12:03
	→ shiba_yu36 joined	12:03
	← artib left	12:03
pidus_	Just one last question - lame though again it prolly is. If you pull, just before committing and assuming that there are no changes between the time you pulled and you commit, ever push will be ff right? Guess I missed out on the conflict thing here!	12:04
	*every	12:04
	→ shiba_yu_ joined	12:05
	← shiba_yu36 left	12:05
	→ keller joined	12:06
	→ alexs joined	12:06
	← alexs left	12:06
	→ alexs joined	12:06
	← Praise left	12:06
canton7	pidus_, yup that's right. If you try and push something that would require the remote to do a non-ff merge, the push is rejected. You then have to pull, and merge those changes in yourself, ending up with something that the remote can ff merge	12:06
	→ lisak joined	12:06
	→ Praise joined	12:06
	← marthinal left	12:07
sweb	if i want to use hooks	12:07
	file must be without .sample ?	12:07
	← ziro` left	12:07
canton7	sweb, yup	12:07
sweb	canton7, not work :/	12:07
canton7	(must also be executable with a valid shebang)	12:07
pidus_	canton7: Cool. I guess I got it now. Thanks a ton and sorry for bugging you all with my noob questions!	12:07
	→ wanders joined	12:07
	← Kindari left	12:08
	← d4rk1ink_ left	12:08
canton7	pidus_, no worries. I'd advise reading e.g. !book cover-to-cover, as git's something that it's not too easy to learn using the "learn how to solve problems as you come to them" technique :)	12:09
jast	pidus_: automatic message 'Pro Git' is a complete book about git that is available online for free (there's a print version, too): http://progit.org/book/	12:09
	← _kud left	12:09
	→ zykes- joined	12:09
pidus_	canton7: Will postively read it	12:10
	← ipalaus_ left	12:10
pidus_	cmn: Thanks	12:10
	→ GeorgeSebastian joined	12:11
	← mklappstuhl left	12:11
	← squentin left	12:11
	→ erik_tr joined	12:12
	← alexs left	12:13
	← DarkAR left	12:13
	← pidus_ left	12:13
	→ pidus joined	12:13
	→ Kindari joined	12:16
	→ novettam joined	12:17
	→ weiyang joined	12:17
	← GeorgeSebastian left	12:18
	→ ajt joined	12:19
sweb	i use this introduction but when i commit and push from my client. Public files and git not updated	12:19
	http://joemaller.com/990/a-web-focused-git-workflow/	12:19
	but hub repository update	12:19
canton7	sweb, try putting an echo statement in the hook. Does that get displayed? Can you run the hook directly from the shell?	12:20
	→ oso96_2000 joined	12:21
	→ niboshi_ joined	12:21
	← herdingcat left	12:24
	→ squentin joined	12:24
	← psakrii left	12:24
	→ tiagobutzke joined	12:24
	→ ph^ joined	12:25
	← novettam left	12:26
	← lacrymology left	12:26
	← kmc left	12:28
	← TAsnw left	12:28
	← kalusn left	12:29
	→ GeorgeSebastian joined	12:29
	→ novettam joined	12:29
	← _md left	12:30
sweb	canton7, problem is for hook permissions. my problem is solved. ty	12:30
canton7	:)	12:30
	→ the_cybersphinx joined	12:32
	→ DarkAR joined	12:35
	← tatsuya_o left	12:35
	← GeorgeSebastian left	12:36
	← cybersphinx left	12:36
	→ mitsuhiko joined	12:36
	→ alnewkirk joined	12:36
	→ alexs joined	12:36
	← alexs left	12:36
	→ alexs joined	12:36
	← wolfric left	12:36
	→ wolfric joined	12:36
	→ tatsuya_o joined	12:36
sweb	how to pass my password to git push in linux command line	12:36
	?	12:36
crashanddie	sweb, you can't.	12:37
	→ subbyyy joined	12:37
	→ _md joined	12:37
	→ kmc joined	12:37
canton7	sweb, what protocol are you pushing over? http(s) or ssh?	12:37
sweb	ssh	12:37
canton7	look into setting up ssh key authentication	12:38
thiago	configure ssh so you don't have to type a password	12:38
	← bindaasomatic left	12:40
	→ bindaasomatic joined	12:40
sweb	i must password in this line url = ssh://gituser@gitserver/path/to/.git/	12:41
	← stodan left	12:41
sweb	thiago, ^	12:41
wereHamster	must?	12:42
canton7	sweb, google for something like "ssh without password". There are millions of hits	12:42
	this isn't a git problem	12:42
	→ regedit joined	12:47
	← ben_alman left	12:47
	← Samoi left	12:47
	→ johnkpau_ joined	12:48
regedit	any files in the same directory as the .git folder is part of the working tree?	12:49
	→ oriba joined	12:49
	→ GeorgeSebastian joined	12:49
	← dv310p3r left	12:49
	→ Samoi joined	12:49
	← na3r left	12:51
	→ spatulasnout joined	12:51
phluks	Yep	12:51
	→ p3rror joined	12:51
phluks	unless they are in .gitignore	12:52
canton7	(or .git/info/exclude)	12:52
	← Codif left	12:52
	→ c00kiemon5ter joined	12:53
	← c00kiemon5ter left	12:53
	→ c00kiemon5ter joined	12:53
regedit	however i seem to remember it being advised not to have the actual project files in the same place as the .git folder, but rather make a subfolder for it. true?	12:53
	← c00kiemon5ter left	12:53
	← jesteracer left	12:54
phluks	I disagree. However, it depends on the layout of your project. It may make more sense.	12:54
	← GodEater left	12:55
cmn	the working tree is more about the files that git cares about... what do you mean by project files in a subfolder?	12:55
phluks	regedit: If your git repo _is_ the project or if you have more projects in your git repo fx.	12:56
wereHamster	regedit: if your project files are in /foo/bar, then the .git dir is /foo/bar/.git	12:56
regedit	right	12:57
	← GeorgeSebastian left	12:57
regedit	thanks guys	12:59
	← regedit left	12:59
	← glcrazy_ left	13:00
	→ Cipher-0 joined	13:01
	→ merkosh joined	13:03
	→ Ingmar joined	13:03
	→ avtobiff joined	13:03
	→ GodEater joined	13:03
	→ [M]ax joined	13:03
	→ leo2007 joined	13:03
	← teweWork left	13:04
	← msponge left	13:04
	→ ramusara joined	13:06
	→ d4rk1ink joined	13:07
	→ alberto56 joined	13:07
	→ khmarbaise joined	13:08
	← Chib left	13:08
	→ teweWork joined	13:08
	→ lacrymology joined	13:08
phluks	Has anybody been working with a dynamic or "masterless" branching strategy workflow?	13:09
wereHamster	explain..	13:10
	→ GeorgeSebastian joined	13:10
phluks	.. fx where one releases dynamic branches in stead of the more "normal" static release branch	13:10
cmn	it sounds like more trouble than it's worth	13:11
phluks	Or in other words you dont have a specific branch for releases	13:11
	It may be	13:11
	but it seems very flexible	13:11
	And hard to explain :)	13:11
cmn	if there is no canonical stable branch, how do people know what code is stable and which may crash their machine?	13:12
phluks	Exactly. Theres a problem with the philosophy of the branches.	13:12
	where they come from and where they are going	13:12
	My current solution is to have a separate repo for the branch config	13:13
cmn	why not in the same repo?	13:13
	→ cburroughs joined	13:13
phluks	where each branch has a dir with files with the current state of say dev, test and live	13:14
	→ _kud joined	13:14
phluks	cmn: it gives problems with merges	13:14
	→ verto joined	13:14
phluks	the config is common to all branches	13:15
	ideally Id like to have a "global" option in git	13:15
cmn	hm, ok, I thought you meant it was some sort of human-readable text	13:15
phluks	but Im not sure that makes sense	13:15
	← JeffJohnson left	13:15
phluks	cmn: it is, but it is structured in dirs, like a sort of database	13:16
	← albel727 left	13:16
phluks	"global" with respect to branches in this case	13:16
	→ JeffJohnson joined	13:16
	← JeffJohnson left	13:16
	→ JeffJohnson joined	13:16
cmn	for which fields in particular?	13:16
phluks	cmn: fields?	13:18
	← GeorgeSebastian left	13:18
cmn	options, whatever, which parts of the config do you need to have defined for all branches?	13:18
	← johnkpau_ left	13:20
	→ tswr joined	13:20
phluks	I got interupted by Project manager - just a second	13:20
	→ Dave^\| joined	13:21
	← leen^\| left	13:21
	← _kud left	13:23
	← SpecialEmily left	13:23
	→ _kud joined	13:23
	opalka → opalka\|lunch	13:23
tswr	Hi, all! I am writing a watch dog. What is the best way to check that the repository is available? When trying to clone the http repository, if it is down, git-remote-http hangs for a very long time. Is there a way to setup the timeouts for clone?	13:23
	← jceb left	13:23
canton7	why not just create your own http request?	13:24
	← Daniel0108 left	13:24
	← magcius left	13:24
phluks	back	13:24
	→ _kud_ joined	13:25
tswr	canton7, well, it looks like writing a small limited version of helper, but with timeouts... looks a bit unnatural, though quite possible	13:25
phluks	cmn: its the meaning of the branchs - actually. What they are for. Feature, release branch on dev, test or live	13:25
	→ magcius joined	13:25
phluks	cmn: what features are on what releases and so on	13:26
	s/branchs/branches/	13:26
canton7	tswr, request a small file that you know should always exist, like info/refs. use curl or something, then you have absolute control	13:26
tswr	canton7, i'd also like to check that push/pull works ok for specific users	13:27
cmn	phluks: that sounds like something that should live on a branch of its own and never get merged	13:27
	← subbyyy left	13:27
phluks	cmn: yes, but i cannot checkout a branch on the live server, that does not contain the code :)	13:28
	Or have I misunderstood?	13:28
canton7	tswr, there's http.lowSpeedLimit and http.lowSpeedTime, might be worth looking at	13:28
	← weiyang left	13:28
	← _kud left	13:29
tswr	canton7, I tried using them, but it looks like they refer only to the speed after connect was successful	13:29
	:)	13:29
	→ Suresht joined	13:29
	← gwasshoppah left	13:29
	← leo2007 left	13:29
canton7	tswr, I'd suggest checking that the remote is online first with your own http request, then attempting to clone/push whatever you need to. It's a bit hacky, but it'll work	13:30
	→ GeorgeSebastian joined	13:30
	→ erichynds joined	13:31
tswr	canton7, it will work when I'm lucky :) the server can go down right after the first http request :)	13:31
phluks	We are trying a solution that is basicly a dynamic version of this http://nvie.com/posts/a-successful-git-branching-model/	13:31
canton7	tswr, what are the chances? Given that such an event will only cause the watching test to slow down significantly (but not fail), I'd say that was acceptable	13:31
phluks	In stead of develop and master we have numbered releases.	13:31
	Ah, cake -> away	13:32
	→ jsick3 joined	13:32
cmn	phluks: you don't need to checkout a branch, you should be able to do "git checkout descriptions -- '*'"	13:32
	→ chris_99 joined	13:32
tswr	canton7, well, yes that's what I'll probably do if I won't find anything better	13:33
	→ albel727 joined	13:33
tswr	canton7, thank you!	13:33
	→ Daniel0108 joined	13:33
	→ ilteris_ joined	13:34
	← workmad3 left	13:34
	← teweWork left	13:35
canton7	tswr, sorry I couldn't find something neater :P	13:35
	← roflin left	13:35
	→ ipalaus_ joined	13:35
	→ roflin joined	13:36
tswr	canton7, :)	13:37
	← GeorgeSebastian left	13:37
	← tswr left	13:37
	→ harshpb joined	13:37
	← henrikbjorn left	13:38
	→ bartek joined	13:39
	→ tommyvyo_ joined	13:40
	→ weiyang joined	13:41
	← sweb left	13:41
	→ RichiH joined	13:41
	→ sweb joined	13:42
RichiH	what's the sanest way to find out what variant of http://www.kerneltrap.com/mailarchive/linux-netdev/2009/4/29/5598314/thread#mid-5598314 made it into the kernel tree and in which tags it exists? the subject of the email does not appear in git log.	13:42
	→ timmywil joined	13:43
	← timmywil left	13:43
	← tatsuya_o left	13:43
	→ glennpratt joined	13:44
FauxFaux	RichiH: git log --all --grep ?	13:44
	→ tatsuya_o joined	13:44
	→ soc42 joined	13:47
phluks	cmn: ...ah, that could be the solution I am looking for. Thnx I'll look into that	13:47
RichiH	FauxFaux: neat; that's a _lot_ faster than searching the log via less. now that i have the commit id, how do i find all tags which include that commit?	13:47
	→ flavius joined	13:48
	← nardev left	13:48
cmn	git tag --contains <tag>	13:48
	← dwmw2_gone left	13:48
	→ dwmw2_gone joined	13:48
	→ hyperair joined	13:49
	→ AaronMT joined	13:50
	← kvanderw left	13:50
	→ ISF joined	13:50
	→ heptadecagram joined	13:50
	← sattu94 left	13:51
	← nax left	13:52
	← akamike left	13:52
	→ teweWork joined	13:52
	→ niloc132 joined	13:56
	← erik_tr left	13:56
	← khmarbaise left	13:58
	← sanduz2 left	13:58
	← neurodrone left	13:58
	napdragon → spyraway	14:00
	→ timmywil joined	14:01
	← wks left	14:01
	→ kR15T0 joined	14:02
	→ Flawless joined	14:03
	→ CannedCorn joined	14:03
	ChanServ set mode: +v	14:03
	← jspiros left	14:04
Flawless	Hello guys. I'm working on a git project, and I'm trying to learn git, but I'm not terribly experienced with it yet. I stashed some changes while working on another problem, and have now un-stashed them	14:04
	However, when I do "git status", I get this:	14:04
	http://p.adora.dk/P2219.html	14:05
	what does "both modified" mean? And how come I can't see my changes the files with "git diff"?	14:05
	← avinashhm left	14:05
	→ jspiros joined	14:05
	→ dv310p3r joined	14:05
	← envi_ left	14:05
louiz	because there’s a conflict between what you changed AFTER you stashed something, and what the stashed things are	14:06
	→ envi_ joined	14:06
	→ d0k joined	14:06
	→ _kud joined	14:06
Flawless	Yes, I got a number of conflicts on unstash. I manually resolved them all	14:06
	so what now?	14:06
louiz	if you manually resolved them, you have to git add each of them	14:06
kR15T0	how can I remove the commit before the last one using rebase on a master branch?	14:06
louiz	Flawless: git status says: # (use "git add/rm <file>..." as appropriate to mark resolution)	14:07
	← mgpcoe left	14:07
louiz	you have to do "git add somefile" to say “this file is resolved”	14:07
	← flavius left	14:07
louiz	once everything is resolved, commit	14:08
Flawless	ah, cool. It's just that I'm wary of doing "git add", since I've seen in the past that the files are then added, but do not show up in the diff	14:08
	which is.. confusing	14:08
louiz	to see the files that have been added, do git diff --cached	14:08
Flawless	ah, cool	14:09
louiz	"git diff" just shows the changes that have not yet been added	14:09
	← _kud_ left	14:09
louiz	and git diff HEAD shows both	14:09
Flawless	Yeah, that's a key difference from (say) SVN that I'm still trying to get used to	14:09
	cool	14:09
louiz	yeah, that’s also different from hg	14:09
	since there’s no staging area in hg	14:10
	← Dave^\| left	14:10
	← lisak left	14:10
	→ DarkAR_ joined	14:10
	← keller left	14:10
	← weiyang left	14:10
	→ leen^\| joined	14:10
	← Dreamer3 left	14:11
	← jspiros left	14:11
	← timmywil left	14:11
	← roflin left	14:11
	← Daniel0108 left	14:11
	← JeffJohnson left	14:11
	← d4rk1ink left	14:11
	← Samoi left	14:11
	← oriba left	14:11
	← alexs left	14:11
	← novettam left	14:11
	← oso96_2000 left	14:11
	← zykes- left	14:11
	← wanders left	14:11
	← Praise left	14:11
	← Tianon left	14:11
	← kerrick left	14:11
	← chrisf left	14:11
	← SethRobertson left	14:11
	← LotR left	14:11
	← Ratler left	14:11
	← quentusrex left	14:11
	← ketas left	14:11
	← madduck left	14:11
	← prudhvi left	14:11
	← plopix left	14:11
	← EugeneKay left	14:11
	← milli left	14:11
	← lamont left	14:11
	← twem2_ left	14:11
	← stoffus left	14:11
	← tg left	14:11
	← fahadsadah left	14:11
	← Xeross left	14:11
	← Cerise left	14:11
	← AlexanderS left	14:11
	← zibri left	14:11
	← tuxx- left	14:11
	← dsop left	14:11
	← wolfric left	14:11
	← ciaranm left	14:11
	← xou left	14:11
	← deever left	14:11
	← _aeris_ left	14:11
	← Kiall left	14:11
	← kanru left	14:11
	← MacGyverNL left	14:11
	← Flawless left	14:11
	← ReekenX left	14:11
	← alecat left	14:11
	← freimatz2 left	14:11
	← jmah left	14:11
	← Cogito left	14:11
	← nyuszika7h left	14:11
	← variable left	14:11
	← reuss left	14:11
	← coredumb left	14:11
	← troyt left	14:11
	← blast_hardcheese left	14:11
	← torbenh3 left	14:11
	← traviscj left	14:11
	← etcetera left	14:11
	← Nickeeh left	14:11
	← Marquel left	14:11
	← Internet13 left	14:11
	← kenyon left	14:11
	← Drakevr left	14:11
	← metze left	14:11
	← Arrowmaster left	14:11
	← avar left	14:11
	← j0ran left	14:11
	← mino left	14:11
	← Ksg89 left	14:11
	← karihre left	14:11
	← teuf left	14:11
	← FDFlock left	14:11
	← yogurt_truck left	14:11
	← DjSlash left	14:11
	← preflex left	14:11
	← magnusol left	14:11
	← Delerius left	14:11
	← eeadc left	14:11
	← nb left	14:11
	← meder left	14:11
	← TheFriscoFainter left	14:11
	← iff left	14:11
	← didlix left	14:11
	← mediaslave left	14:11
	← Grum left	14:11
	← d1b left	14:11
	← jrockway left	14:11
	← haarg left	14:11
	← naderman left	14:11
	← deavid left	14:11
	← eregon left	14:11
	← OPK_ left	14:11
	← dropdrive left	14:11
	← rphillips left	14:11
	← electrogeek left	14:11
	← koki left	14:11
	← jaeckel left	14:11
	← ska-fan left	14:11
	← tychoish left	14:11
	← youam left	14:11
	← morgajel left	14:11
	← Corey left	14:11
	← BlindOracle left	14:11
	← beautifulmind left	14:11
	→ Cipher-0_ joined	14:11
	→ soc42_ joined	14:11
	← w0bni left	14:11
	← PuffTheMagic left	14:11
	→ innociv_ joined	14:11
	← felher left	14:11
	← spacebat left	14:11
	→ jwmann1 joined	14:11
	→ JEEBcz joined	14:11
	→ hwrdprkns_ joined	14:11
	← BeerPerson left	14:11
	← mattp_ left	14:11
	← mrchrisadams left	14:11
	← charon left	14:11
	← shennyg left	14:11
	← pronoy left	14:11
	→ brian_g_ joined	14:11
	← mtkd left	14:11
	← Adys left	14:12
	← CannedCorn left	14:12
	→ mrchrisadams_ joined	14:12
	→ Adys joined	14:12
	← jwmann left	14:12
	→ wbednars_ joined	14:12
	← Cipher-0 left	14:12
	← flaguy48 left	14:12
	→ mattp_ joined	14:12
	→ LiohAu__ joined	14:12
	→ psoo_ joined	14:12
	← brian_g left	14:12
	← mirage335 left	14:12
	← dcelix left	14:12
	→ charon joined	14:12
	← Rejinka left	14:12
	← mitsuhiko left	14:12
	← eMBee left	14:12
	← psoo left	14:12
	← mika left	14:12
	← ngulden left	14:12
	← hobodave left	14:12
	← ricky left	14:12
	→ estan_ joined	14:12
	← shachaf left	14:12
	← Kazii left	14:12
	→ hobodave joined	14:12
	← RichardBronosky left	14:12
	→ chuck001 joined	14:12
	→ milki_ joined	14:12
	→ childbear2000 joined	14:12
	→ tiagobutzke_ joined	14:12
	← Cipher-0_ left	14:12
	→ felher joined	14:12
	→ w0bni joined	14:12
	← niboshi left	14:12
	← _Qman left	14:12
	← milki left	14:12
	← envi_ left	14:12
	← tiagobutzke left	14:12
	← phix left	14:12
	← tizzo-afk left	14:12
	← juvenal left	14:12
	← markvandenborre left	14:12
	← errr left	14:12
	tiagobutzke_ → tiagobutzke	14:12
	← childbear left	14:12
	→ bitslip_ joined	14:12
	← heptadecagram left	14:13
	← berserkr left	14:13
	← fir31ight left	14:13
	← johntramp left	14:13
	→ youam joined	14:13
	→ torbenh3 joined	14:13
	← ilteris_ left	14:13
	← clone1018 left	14:13
	→ alanp joined	14:13
	← sweb left	14:13
	← crass left	14:13
	← DarkAR left	14:13
	← Dave^\|\| left	14:13
	← chuck00 left	14:13
	← atdiehm left	14:13
	← nDuff left	14:13
	← Cacheaway left	14:13
	← Gitzilla left	14:13
	← kermit left	14:13
	← tommyvyo left	14:13
	tommyvyo_ → tommyvyo	14:13
	→ Cipher-0 joined	14:13
	→ hwrdprkns__ joined	14:13
	→ clone1018 joined	14:13
	→ _Qman joined	14:13
	← hwrdprkns left	14:13
	← spatulasnout left	14:13
	← mace left	14:13
	← jjurach left	14:13
	→ ricky joined	14:13
	← soc42 left	14:13
	← kmc left	14:13
	← ajt left	14:13
	← nixmaniack left	14:13
	← setmeaway left	14:13
	← wbednarski left	14:13
	← tvw left	14:13
	← guardian left	14:13
	← Yuuhi`` left	14:13
	← cyphase left	14:13
	← Exposure left	14:13
	← _Vi left	14:13
	← Error404NotFound left	14:13
	← chrizz left	14:13
	→ Rejinka joined	14:13
	← jpalmer left	14:13
	← serj left	14:13
	← bitslip left	14:13
	← mtx_init left	14:13
	← dotCOMmie left	14:13
	← Araxia left	14:13
	← estan left	14:13
	← alanp_ left	14:13
	← Guest96808 left	14:13
	← harv left	14:13
	← RobertLaptop left	14:13
	← puff left	14:13
	← DrNick left	14:13
	← alnewkirk left	14:13
	← mishok13 left	14:13
	→ BearPerson joined	14:13
	→ alnewkirk joined	14:13
	Araxia_ → Araxia	14:13
	→ AABD3HY joined	14:13
	→ shachaf_ joined	14:13
	← LiohAu left	14:13
	→ xou_ joined	14:13
	← BearPerson left	14:13
	→ BearPerson joined	14:13
	← ttvd left	14:13
	← ccushing left	14:13
	→ setmeaway2 joined	14:13
	LiohAu__ → 64MAA169I	14:13
louiz	:/	14:13
	← pl0sh left	14:13
	← _Qman left	14:13
	← JEEB left	14:13
	→ etcetera joined	14:13
	← rchavik left	14:13
	← s0ber left	14:13
	← toobluesc left	14:13
	→ Yuuhi`` joined	14:13
	← Fissure left	14:14
	← phaedrix left	14:14
	← Slush- left	14:14
	← lqi left	14:14
	← dr_lepper left	14:14
	← localhost left	14:14
	← kst left	14:14
	→ cybersphinx joined	14:14
	← the_cybersphinx left	14:14
	← eijk left	14:14
	→ chrizz joined	14:14
	← zivester left	14:14
	← mdpatrick_ left	14:14
	→ timmywil joined	14:14
	← tgunr_ left	14:14
	← innociv left	14:14
	→ juvenal joined	14:14
	← jeremy_c left	14:14
	← jeffhung left	14:14
	← dilfridge left	14:14
	← charon left	14:14
	→ charon joined	14:14
	→ heptadecagram joined	14:14
	← thews left	14:14
	→ dilfridge joined	14:14
	← sontek left	14:14
	← SJS left	14:14
	→ Error404NotFound joined	14:14
	← jstemmer left	14:14
	← bobdobbs left	14:14
	→ harv joined	14:14
	→ kylebk joined	14:14
	→ markvandenborre joined	14:14
	→ RichardBronosky joined	14:14
	→ Marquel joined	14:14
	→ s0ber_ joined	14:14
	s0ber_ → s0ber	14:14
	→ guardian joined	14:14
	→ ajt joined	14:14
	→ phix joined	14:14
	→ errr joined	14:14
	→ mace joined	14:14
	→ mika joined	14:15
	← mika left	14:15
	→ mika joined	14:15
	→ jeremy_c joined	14:15
	→ spacebat joined	14:15
	→ phaedrix joined	14:15
	→ envi_ joined	14:15
	→ spb_ joined	14:15
	→ johntramp joined	14:15
	→ ISF_ joined	14:15
	→ fir31ight joined	14:15
	→ niboshi joined	14:15
	→ mtkd joined	14:15
	→ naderman joined	14:15
	← spb left	14:15
	← naderman left	14:15
	→ naderman joined	14:15
	→ serj joined	14:15
	spb_ → spb	14:15
	→ dotCOMmie joined	14:15
	← crakdmirror left	14:15
	→ dcelix joined	14:16
	← vmiklos left	14:16
	← Gunni left	14:16
	→ _Vi joined	14:16
	→ berserkr joined	14:16
	→ mtx_init joined	14:16
	← hwrdprkns_ left	14:16
	← ph^ left	14:16
	← TML left	14:16
	← dec left	14:16
	← ahf left	14:16
louiz	Is there a way to know, BEFORE pushing which commit will be pushed? For example I want to review all the commit that will be pushed IF I do "git push origin master", can I do that?	14:16
	→ crakdmirror joined	14:16
	→ Dreadnaught joined	14:16
	← Dreadnaught left	14:16
	→ Dreadnaught joined	14:16
louiz	(if somebody sees the question between all these netsplits…)	14:16
	→ mdpatrick joined	14:16
	→ sontek joined	14:16
	← svm_invictvs left	14:17
	→ Exposure joined	14:17
	→ RobertLaptop joined	14:17
	→ atdiehm joined	14:17
	→ dec joined	14:17
	← Unknown[NF] left	14:17
	→ tizzo-afk joined	14:17
	← teweWork left	14:17
	→ kmc joined	14:17
	→ ph^ joined	14:17
	← Kulrak left	14:17
	← sycofly left	14:17
	→ nyuszika7h joined	14:17
	→ zivester joined	14:17
	→ eijk_ joined	14:17
	→ Grum joined	14:17
	→ koki joined	14:17
	→ davec__ joined	14:17
	→ envi_\|2 joined	14:17
	→ jspiros joined	14:17
	→ Flawless joined	14:17
	→ roflin joined	14:17
	→ JeffJohnson joined	14:17
	→ d4rk1ink joined	14:17
	→ Samoi joined	14:17
	→ oriba joined	14:17
	→ alexs joined	14:17
	→ novettam joined	14:17
	→ oso96_2000 joined	14:17
	→ zykes- joined	14:17
	→ wanders joined	14:17
	→ Praise joined	14:17
	→ Tianon joined	14:17
	→ kerrick joined	14:17
	→ chrisf joined	14:17
	→ SethRobertson joined	14:17
	→ LotR joined	14:17
	→ Ratler joined	14:17
	→ madduck joined	14:17
	→ ketas joined	14:17
	→ prudhvi joined	14:17
	→ plopix joined	14:17
	→ Cerise joined	14:17
	→ EugeneKay joined	14:17
	→ milli joined	14:17
	→ lamont joined	14:17
	→ Xeross joined	14:17
	→ twem2_ joined	14:17
	→ stoffus joined	14:17
	→ fahadsadah joined	14:17
	→ tg joined	14:17
	→ AlexanderS joined	14:17
	→ kanru joined	14:17
	→ zibri joined	14:17
	→ tuxx- joined	14:17
	→ dsop joined	14:17
	→ wolfric joined	14:17
	→ ciaranm joined	14:17
	→ _aeris_ joined	14:17
	→ MacGyverNL joined	14:17
	hubbard.freenode.net set mode: +v	14:17
	→ alecat joined	14:17
	→ freimatz2 joined	14:17
	→ Cogito joined	14:17
	→ variable joined	14:17
	→ reuss joined	14:17
	→ coredumb joined	14:17
	→ traviscj joined	14:17
	→ Nickeeh joined	14:17
	→ Internet13 joined	14:17
	→ mino joined	14:17
	→ kenyon joined	14:17
	→ yogurt_truck joined	14:17
	→ Drakevr joined	14:17
	→ metze joined	14:17
	→ avar joined	14:17
	→ Ksg89 joined	14:17
	→ karihre joined	14:17
	→ teuf joined	14:17
	→ FDFlock joined	14:17
	→ DjSlash joined	14:17
	→ preflex joined	14:17
	→ magnusol joined	14:17
	→ eeadc joined	14:17
	→ nb joined	14:17
	→ meder joined	14:17
	→ TheFriscoFainter joined	14:17
	→ iff joined	14:17
	→ didlix joined	14:17
	→ mediaslave joined	14:17
	→ d1b joined	14:17
	→ jrockway joined	14:17
	→ morgajel joined	14:17
	→ haarg joined	14:17
	→ deavid joined	14:17
	→ eregon joined	14:17
	→ OPK_ joined	14:17
	→ rphillips joined	14:17
	→ electrogeek joined	14:17
	→ jaeckel joined	14:17
	→ ska-fan joined	14:17
	→ tychoish joined	14:17
	→ Dreamer3 joined	14:17
	← envi_\|2 left	14:17
	← didlix left	14:17
	← mediaslave left	14:17
	← d1b left	14:17
	← rphillips left	14:17
	→ mediaslave joined	14:17
cmn	louiz: update the remote-tracking branches, then git log branch..remote/branch will tell you what commits are different	14:17
	→ Araxia_ joined	14:17
	← cmn left	14:17
	→ jmah joined	14:17
	→ ttvd joined	14:17
	→ cmn joined	14:17
	← dilfridge left	14:17
louiz	cmn: ah yeah, great, thanks	14:17
	← brian_g_ left	14:17
	→ BlindOracle joined	14:18
	← traviscj left	14:18
	← fahadsadah left	14:18
	← nb left	14:18
	→ DrNick joined	14:18
	← yogurt_truck left	14:18
	← haarg left	14:18
	← zivester left	14:18
	← Dreamer3 left	14:18
	→ gigamonkey joined	14:18
louiz	(one sentence is Excess flood, ok…)	14:18
	← ISF left	14:18
	→ ahf joined	14:18
	→ sycofly joined	14:18
	→ cyphase joined	14:18
	→ Unknown[NF] joined	14:18
	→ Corey joined	14:18
	← OPK_ left	14:18
	→ eMBee joined	14:18
	→ dilfridge joined	14:18
	→ zivester joined	14:18
	→ yogurt_truck joined	14:18
	→ Dreamer3 joined	14:18
	→ OPK joined	14:18
	→ Kiall joined	14:18
	→ mitsuhiko joined	14:18
	→ d1b joined	14:18
	→ TML joined	14:18
	→ svm_invictvs joined	14:18
	← svm_invictvs left	14:18
	→ svm_invictvs joined	14:18
	→ didlix joined	14:18
	→ jeffhung joined	14:18
	← ls\|james left	14:18
	→ kermit joined	14:18
cmn	louiz: or rather, branch...remote/branch will tell you what's different	14:18
	→ deever joined	14:19
	→ rphillips joined	14:19
	← rphillips left	14:19
	→ rphillips joined	14:19
	→ SJS joined	14:19
louiz	Ok, I find this ... or .. notation quite hard to remember :p	14:19
	→ traviscj joined	14:19
	→ shennyg joined	14:19
	→ quentusrex joined	14:19
	→ j0ran joined	14:19
	→ pronoy joined	14:19
	← pronoy left	14:19
	→ pronoy joined	14:19
	→ Delerius joined	14:19
	→ troyt joined	14:20
	→ ph^_ joined	14:20
	← armenb left	14:20
	→ workmad3 joined	14:20
	→ Gunni joined	14:20
cmn	louiz: yes, it's a bit tricky	14:20
	← ricky left	14:20
	→ ricky joined	14:20
	→ thews joined	14:21
	→ Arrowmaster joined	14:21
	← codetroll left	14:21
	→ jstemmer joined	14:22
	→ ilteris_ joined	14:22
	→ vmiklos joined	14:22
	← flexd left	14:22
	→ teweWork joined	14:22
	← ph^ left	14:22
	← nyuszika7h left	14:22
	→ dropdrive joined	14:22
	← Arrowmaster left	14:22
	→ Arrowmaster joined	14:22
	← Cogito left	14:22
	→ ls\|james joined	14:22
	→ jjurach joined	14:22
	→ crass joined	14:22
	ricky → Guest72597	14:22
	← shadowmaster left	14:22
	→ Fissure joined	14:22
	→ shadowmaster joined	14:22
	→ nyuszika7h joined	14:23
	→ Cogito_ joined	14:23
	→ weiyang joined	14:23
	→ nb joined	14:23
	← bburhans left	14:23
	→ toobluesc joined	14:23
	← docgnome left	14:23
	→ blast_hardcheese joined	14:24
	← tabakhase left	14:24
	← alnewkirk left	14:24
	→ bburhans joined	14:24
	→ pl0sh joined	14:24
	→ Sico joined	14:25
	← SicoWork left	14:25
	→ localhost joined	14:25
	← _kud left	14:25
	d1b → db	14:25
	→ fahadsadah joined	14:25
	→ _kud joined	14:25
	Sico → SicoWork	14:25
	→ tabakhase joined	14:25
	db → Guest74077	14:26
	→ mirage335 joined	14:26
	→ haarg joined	14:26
	→ Daniel0108 joined	14:26
	← Guest74077 left	14:26
	→ ceej_ joined	14:26
	→ GunArm joined	14:26
	→ docgnome joined	14:27
	→ Gitzilla joined	14:27
	→ nixmaniack joined	14:27
	→ armenb joined	14:27
	→ nDuff joined	14:27
	→ lisak joined	14:27
	→ d1b joined	14:27
	← nyuszika7h left	14:27
	→ flavius joined	14:27
	→ flaguy48 joined	14:28
	← auke- left	14:28
	← tabakhase left	14:28
	→ tabakhase joined	14:28
	→ nyuszika7h joined	14:29
	→ mishok13 joined	14:29
	→ CannedCorn joined	14:29
	→ fedesilva joined	14:29
	→ flexd joined	14:30
	→ auke- joined	14:30
	← jeremy_c left	14:30
	→ jeremy_c joined	14:30
	← nixmaniack left	14:30
	← sitaram left	14:31
	← Prasad left	14:31
	→ sitaram joined	14:31
	← sitaram left	14:31
	→ sitaram joined	14:31
	→ nixmaniack joined	14:32
	→ Slush- joined	14:32
	→ axl_ joined	14:32
louiz	Is there a way to rename a committer name in a whole repository (all the concerned commits), WITHOUT breaking the repository for people who already cloned it etc?	14:32
cmn	no	14:32
louiz	ok	14:32
	that’s what I thought	14:33
FauxFaux	(By design!)	14:33
louiz	(Yeah, I understand)	14:33
matthiasgorgens	Does "git grep" also look in stashes?	14:33
cmn	the committer name is part of the data in the commit which is used to hash it, so if you change it, you're changing the hash etc etc	14:33
	→ joshsmith joined	14:33
louiz	yep, right	14:33
cmn	it's actually a feature	14:33
	→ flijten joined	14:33
	← fir31ight left	14:34
phluks	cmn: It seem to me that I'll have to create my own commits (mktree, commit-tree) if I want to commit to an unmerged branch or, do you have a fix for that too :)	14:34
louiz	And if I do that though, what will happen for people with their clone? Will they have to rm the directory and reclone it completely?	14:34
	→ eletuchy joined	14:35
louiz	what will happen if they try to pull the modified repository?	14:35
cmn	they'll merge two completely different histories that somehow produce the same worktree	14:35
	→ eephillip joined	14:35
canton7	louiz, if they merge, they'll join together the old history (the one in their repo) and the new one, and you'll end up with a third history which is a merge of the other two. If they rebase, everything will be fine (I think)	14:35
	someone please confirm on my rebase statement :P	14:36
cmn	they'd have to use --onto	14:36
	but it's doable	14:36
	→ Zariel joined	14:36
	→ muneeb joined	14:36
	→ neurodrone joined	14:37
canton7	or if they don't have any local commits which weren't rewritten (or any local changes), they can fetch; reset --hard origin/branch	14:37
	← lnostdal left	14:37
	← nixmaniack left	14:37
louiz	I’m worried only for people using my repository just to update their code and use my software. I can tell all the devs to reclone if necessary. So will they just be able to pull (without conflict)?	14:37
cmn	phluks: you can do a partial commit ignoring that magic file and then checkout that other branch and commit that file there	14:37
	→ fir31ight joined	14:37
	→ drev1 joined	14:38
louiz	actually I’ll try myself	14:38
	→ ccushing joined	14:38
	← flavius left	14:38
	→ luke-jr joined	14:39
	→ keller joined	14:39
luke-jr	I want to make a private git repo public, but I need to audit it for passwords in the entire history-- is there a simple way to list all the objects in a repository, so I can 'git show' each one?	14:39
	→ teknoid joined	14:40
	← teknoid left	14:40
	← MacGyverNL left	14:41
	← glcrazy left	14:41
	→ MacGyverNL joined	14:42
	← MacGyverNL left	14:42
	→ MacGyverNL joined	14:42
	← troyt left	14:43
canton7	luke-jr, git rev-list --objects --all	14:44
luke-jr	thx	14:44
	← drev1 left	14:44
phluks	cmn: hmm, Im not sure I follow you. The way I see it I need to not switch branches. Like the nice checkout.	14:45
cmn	well, that checkout invocation works, but committing only ever really happens on the active branch	14:45
Zariel	If the first commit in my branch history has incorrect eol how I can I fix it so it cleanly applies to other branches?	14:46
phluks	yes	14:46
	← workmad3 left	14:46
	→ drev1_ joined	14:46
	← kpreid left	14:46
	→ troyt joined	14:47
	→ Jackneill joined	14:47
	← Jackneill left	14:47
	→ Jackneill joined	14:47
	← ilteris_ left	14:48
	← kR15T0 left	14:49
	← bindaasomatic left	14:50
	joshteam_mustEMA → joshteam	14:50
cmn	phluks: checkout when switching branches will keep your local modifications, so you can change the descriptions when you're committing on whatever branch; checkout descriptions branch and commit the descriptions there	14:50
	→ jonatasnona joined	14:52
	→ tiglog joined	14:52
	→ ilteris_ joined	14:54
	→ kpreid joined	14:54
	← cooloney left	14:55
	→ ceej joined	14:55
	← muneeb left	14:56
	→ kvanderw joined	14:56
	→ arvind_khadri joined	14:57
	← ISF_ left	14:57
	→ psakrii joined	14:57
	← oc80z left	14:57
	← troyt left	14:57
	← Praise left	14:57
	→ oc80z joined	14:57
	← ceej_ left	14:57
	→ ph^ joined	14:57
	→ Praise joined	14:57
	← eeadc left	14:57
	→ kR15T0 joined	14:57
	← yorick left	14:57
	→ aidenhong joined	14:57
	→ troyt_ joined	14:57
	→ drl joined	14:57
	→ jpalmer joined	14:57
	← arvind_khadri left	14:57
	→ arvind_khadri joined	14:57
	→ yorick joined	14:58
	← ph^_ left	14:58
phluks	cmn: Ah, well. That would work. I was beginning to look into add, write-tree and then update the desctiption ref to the new tree-object.	14:58
	ah, I would need a commit too :)	14:58
	← kerrick left	14:59
	→ eeadc joined	14:59
	→ scottschulthess joined	14:59
scottschulthess	i'm getting	14:59
	fatal: could not read '/etc/git-commit-template': No such file or directory	14:59
	→ ISF joined	14:59
scottschulthess	when doing "git commit"	15:00
	→ muneeb joined	15:00
scottschulthess	any ideas where i can get that file	15:00
	→ cannonball joined	15:00
cmn	phluks: yeah, git will probably get confused otherwise	15:00
	→ dobby156 joined	15:01
dobby156	hi	15:01
	← Cogito_ left	15:01
	← nyuszika7h left	15:01
dobby156	when i do a git branch -a I am fairly sure it's list is out of date	15:02
	how do I get it to fetch the list of remote branches	15:02
	thank you	15:02
	→ Cogito joined	15:02
cmn	dobby156: git fetch	15:02
frogonwheels	dobby156: git ls-remote ?	15:02
	← bjornbjorn left	15:02
cmn	if you have several remotes, git remot update or git fetch --all	15:02
	→ herdingcat joined	15:03
	→ nwp__ joined	15:03
	herdingcat → ericlee_home	15:03
frogonwheels	dobby156: git ls-remote will list the branches on the remote, but won't update locally.	15:03
	→ bjornbjorn joined	15:03
frogonwheels	dobby156: git doesn't update stuff locally until you fetch	15:03
	→ Swimming_Bird joined	15:04
canton7	scottschulthess, git config commit.template ?	15:04
guardian	i have a placeholder file in my repo; which is dummy content aimed to be replaced by generated content. once the file is added to git, is .gitignore the right solution to prevent other developers from the team to override that placeholder content ?	15:04
canton7	guardian, no. .gitignore will not ignore files which are tracked	15:04
cmn	dobby156: when you say out-of-date, do you mean that you have branches that have been delted? if want to get rid of those, use git fetch --prune	15:04
	← ph^ left	15:05
	→ hashpuppy joined	15:05
guardian	canton7: so what's the solution?	15:05
	is there kind of lock?	15:05
	← kpreid left	15:05
FauxFaux	No.	15:05
hashpuppy	new git user. i did "git branch feature1" and then git add .; git ci -m "adding to feature1". but it added it to master. how do i rollback this commit from master and commit it to feature1 branch	15:05
frogonwheels	guardian: you should have dummy content as blah.sample or something	15:05
	.. and don't commit that file.	15:05
cmn	hashpuppy: branch creates a branch, it doesn't switch you to it	15:06
frogonwheels	you can use sparse-checkout to ignore it if you really need to.	15:06
canton7	hashpuppy , 'git branch' creates a new branch, but doesn't check it out. Use 'git branch feature1; git checkout feature1', or 'git checkout -b feature1' for short	15:06
hashpuppy	cmn: ha. yeah, i learned that the hard way	15:06
	→ elyob joined	15:06
FauxFaux	hashpuppy: git checkout feature1 && git cherry-pick master && git branch -f master master~ && git checkout master	15:06
	→ gerard0 joined	15:06
canton7	hashpuppy, git reset --soft HEAD^; git checkout feature1; git commit should do what you want	15:07
FauxFaux	Mmm, that's a waste.	15:07
canton7	or FauxFaux's way :P	15:07
hashpuppy	which way is preferred? as i'll spend sometime learning to decipher it	15:07
FauxFaux	git branch -f feature1 master && git reset --hard HEAD~ # is what I'd do, but I told myself I'm not going to use reset unless absolutely necessary. ¬_¬	15:07
	I suggest my way as it doesn't involve dealing with uncomitted stuff, so is conceptually simpler. Basically, it involves reset. :p	15:07
hashpuppy	your first suggestion?	15:08
FauxFaux	Uh, doesn't involve reset.	15:08
	Yes.	15:08
canton7	(mine uses a reset --soft, so much less dangerous than a reset --hard. FauxFaux uses fewer commands. It's up to you, really)	15:08
hashpuppy	thanks	15:08
canton7	well, dangerous if you screw up the command. it's not that dangerous	15:08
FauxFaux	It involves asking if you have uncomitted changes, at least.	15:08
canton7	true	15:09
	(as always, there are lots of solutions which will work just fine)	15:09
hashpuppy	FauxFaux: that was almost magical. thanks. i will need to figure out how that works now.	15:09
	← Praise left	15:10
	← CannedCorn left	15:10
	→ Praise joined	15:10
	← Praise left	15:10
	→ Praise joined	15:10
	← childbear2000 left	15:10
	← frogonwheels left	15:11
canton7	FauxFaux, is there any advantage in doing "git branch -f master master~; git checkout master" as opposed to "git checkout master; git reset --hard HEAD~" ?	15:11
	→ the_metalgamer joined	15:12
Silex	it doens't do the same thing?	15:12
	← eletuchy left	15:12
Silex	hum, actually it does... or not... errr	15:12
	Silex is confused and shuts up	15:12
	→ talexb joined	15:12
canton7	to be fair, if there are uncommitted changes, they'll be safe with FauxFaux's version	15:12
	← opalka\|lunch left	15:13
FauxFaux	Yeah, as I said, I try and avoid trying to get people to understand reset. :)	15:13
	← rdz left	15:13
	← jspiros left	15:14
Silex	except if I'm ignorant, I find reset much more understandable than "branch -f"	15:14
	→ jspiros joined	15:14
dobby156	cmn: thanks for the answer	15:15
canton7	intuitively, branch is for creating/modifying branches. I don't tend to think of it as a branch moving tool. Maybe that's just me, though	15:15
dobby156	I just wanted a list of all the branchs on the remote server	15:15
FauxFaux	Silex: Mmm. Probably because the two-argument form of branch isn't very common.	15:15
dobby156	which it has done	15:15
	→ ngulden joined	15:15
dobby156	cmn: just another quick question if I don a git pull (of branch B) whilst in branch A it will try and merge them	15:16
	or does it just switch branches	15:16
	thank you	15:16
CareBear\	pull = fetch + merge	15:16
	this is the first line in the man page	15:16
hashpuppy	so reset will wipe out uncomitted changes?	15:16
dobby156	CareBear\: thank you	15:16
CareBear\	hashpuppy : reset will point your branch at somewhere else than it is now	15:17
FauxFaux	hashpuppy: Read http://progit.org/2011/07/11/reset.html . It does hundreds of things, most of which you don't want or need.	15:17
CareBear\	hashpuppy : --hard will drop uncommitted changes	15:17
	→ cwillu_at_work joined	15:17
hashpuppy	thanks. i def asked my question at the right time	15:17
	→ nixmaniack joined	15:17
	→ drake01 joined	15:18
	← drake01 left	15:18
	→ drake01 joined	15:18
FauxFaux	"The right time", also known as working hours in Europe. ¬_¬	15:18
	→ dfr\|mac joined	15:18
	← drake01 left	15:18
CareBear\	hashpuppy : when in doubt, commit	15:18
	→ cakehero joined	15:18
	← angelsl left	15:19
CareBear\	hashpuppy : as long as you have committed, your work can usually always be retrieved	15:19
canton7	(then, even if you throw away the commit by accident, you can still get it back. If you throw away uncommitted changes by accident, they're gone)	15:19
CareBear\	(of course not if you rm -rf)	15:19
	← Parkotron left	15:19
	← tizzo-afk left	15:19
	→ angelsl joined	15:19
	← muneeb left	15:20
	← dfr\|mac left	15:20
	← kenperkins left	15:20
	← drl left	15:21
	→ Kotoc joined	15:21
	→ dfr\|mac joined	15:21
	← frsk left	15:21
	ChanServ set mode: +v	15:22
	→ flyz_ joined	15:22
Kotoc	looking to do a non-root, local dir install of git on redhat .. anyone know how? I don't have make/gcc,cc either :(	15:22
	→ drake01 joined	15:23
	← drake01 left	15:23
	→ drake01 joined	15:23
	← vdv left	15:23
PerlJam	Kotoc: Is this an embedded system of some sort, or just severely impoverished for no good reason?	15:23
	← drake01 left	15:24
	→ Know1edge joined	15:24
FauxFaux	Kotoc: Get an identical redhat machine on which you do have root, build everything inside a prefix, copy the entire thing over.	15:24
Know1edge	Hey guys, im trying to get my boss to learn to use git…he's on a mac, what's a pretty nice gui git client for him to start on?	15:24
	→ charco joined	15:24
	← flyz left	15:24
charco	Hello all :)	15:24
	← ipalaus_ left	15:25
	→ airborn joined	15:26
charco	I want a friend to pull from my machine, but I just don't know how.. I don't want to give everyone access to the repo using git daemon to serve the repo over http, and I don't want to give the other user full ssh access.. Is there an easy way to achieve this?	15:26
	→ mdunham918_ joined	15:27
canton7	charco, git-daemon doesn't use http. You can use apache to give him http access (it's a pain and pretty ugly), or you can set him up a shell account, but set his shell to git-shell (very restricted_	15:27
FauxFaux	charco: You can create a new user with git-shell, and they can use that user, or you can install gitolite (which is probably packaged and hence easier).	15:27
wereHamster	charco: so you want to allow someone else to access your data without giving him permisison to access your data?	15:27
	← drev1_ left	15:27
PerlJam	charco: I'd go with gitolite :)	15:27
FauxFaux	M-M-MEGAJINX.	15:28
canton7	you can also send him a patchset over email, using git-format-patch and friends	15:28
	haha :)	15:28
	← freimatz2 left	15:28
FauxFaux	git bundle	15:28
wereHamster	I'm curious how you manage to do that..	15:28
phluks	Know1edge: im not mac-dude, but if you cannot make the linux-stuff run on your mac, you can try to check out smart git: http://www.syntevo.com/smartgit/index.html	15:28
mdunham918_	hello i'm working on building a github tracker so if a project I'm following has a new release i will be notified. The way I'm thinking of doing this is by tracking the tags on a project if a new tag is created in the repo then i would know it has been updated would this be the best way of knowing when a project has a new release?	15:28
	→ codebeaker joined	15:28
	← weiyang left	15:28
FauxFaux	She wants bob to have access to the code without bob having shell access or charlie getting access, wereHamster.	15:28
PerlJam	mdunham918_: github does have an RSS feed	15:29
mdunham918_	sweet	15:29
	← Araxia left	15:29
	Araxia_ → Araxia	15:29
Know1edge	phluks: I can make the linux stuff run on my mac, it's just my boss isn't that "tech savvy"	15:29
charco	I am a guy	15:29
mdunham918_	where do i find it PerlJam	15:29
FauxFaux	No, your name is Alice, sorry.	15:29
wereHamster	mdunham918_: if you define a 'release' as 'a new tag' then yes. But not all projects use tags	15:29
FauxFaux	mdunham918_: Github also has ways to notify you about loads of things. See the admin interface and #github.	15:29
mdunham918_	RSS might be a better way	15:29
charco	canton7, that would be nice, but I prefer just giving pull access	15:29
	← Suresht left	15:29
PerlJam	mdunham918_: what FauxFaux said	15:29
charco	wereHamster, I want to give him permissions, but not to my whole computer	15:29
phluks	Know1edge: I am not a mac-dude :) - but Smartgit looks shiny	15:30
canton7	Bob and Alice are such strong conventions that we're not going to break them, just because of someone's real gender :P	15:30
mdunham918_	thanks guys	15:30
charco	git-shell is enough to pull ?	15:30
FauxFaux	Yes.	15:30
Know1edge	thanks phluks	15:30
phluks	Know1egde: ur welcome :)	15:30
wereHamster	charco: is that a one time thing or do you want that to work in the future?	15:30
PerlJam	charco: I'd still go with gitolite.	15:31
charco	wereHamster, it's a one time thing, but I want to know if there's a clean way if I need to do this in the future	15:31
	→ ecthiender joined	15:31
FauxFaux	canton7: The conventionts SHOULD NOT be violated, it is RECOMMENDED that any violators are shot.	15:31
wereHamster	charco: if it's one time, just fire up git-daemon	15:31
charco	PerlJam, I think gitolite is overkill	15:31
wereHamster	if it's recurring, install gitolite	15:31
	or use git-shell	15:31
charco	I think I will go with git-shell, if I can't get it going, I will go with git-format-patch and send it via email..	15:32
	Thank you all :)	15:32
FauxFaux	charco: gitolite from packages is, like, four operations.	15:32
canton7	the git mailing list (and therefore the development of git) functions very well using primarily emailed patches	15:32
FauxFaux	That's for write, though; this is for read.	15:32
	→ spearce joined	15:32
canton7	Well, when you think that one alternative to the git mailing list system is lots of pull requests from lots of contributers, it ends up much the same	15:33
	→ swombat joined	15:33
canton7	but yes	15:34
	→ regedit joined	15:34
	→ Textmode joined	15:34
	← Praise left	15:35
	→ dr_lepper joined	15:36
	← dr_lepper left	15:36
	→ dr_lepper joined	15:36
	→ Praise joined	15:36
regedit	i have "/foo/bar/" and "!/foo/bar/exception.txt" in my .gitignore file, the exception doesn't seem to work - git whines about that file (/foo/bar/exception.txt) being ignored when i try to git add it	15:36
	← cyphase left	15:37
canton7	make the first like /foo/bar/*	15:37
	*line	15:37
	→ avinashhm joined	15:37
	← wbednars_ left	15:37
	← Cogito left	15:38
canton7	"unignore" lines only match files which are explicitely ignored. if you ignore a folder, git doesn't consider each of the files inside for ignoring, and so the unignore rule never kicks in	15:38
	→ cyphase joined	15:38
	→ Cogito joined	15:38
regedit	ok thanks!	15:38
	← Error404NotFound left	15:39
	← harshpb left	15:39
	→ harshpb joined	15:39
	← mdunham918_ left	15:39
	→ mikewintermute joined	15:40
dobby156	Hi sorry another question, I am having a problem with branchs, i do a git fetch origin otherbranch	15:40
	and I get this	15:40
	→ MUILTFN joined	15:40
dobby156	* branch otherbranch -> FETCH_HEAD	15:41
	ok	15:41
	but when I do a git branch it isn't listed there	15:41
	→ ^slacker^ joined	15:41
dobby156	and i can't switch into it	15:41
	← oriba left	15:41
	← pidus left	15:41
	→ mklappstuhl joined	15:42
offby1	that's because you didn't fetch it properly.	15:42
	← canton7 left	15:42
	← harshpb left	15:42
offby1	just do "git fetch"; that should get all branches; then you try checking it out again	15:42
	→ harshpb joined	15:43
	→ tvw joined	15:45
	← ixti left	15:45
	→ kai joined	15:45
	→ Suresht joined	15:45
	← Suresht left	15:45
	→ Suresht joined	15:45
kai	hey folks	15:45
	→ ixti joined	15:46
	→ bindaasomatic joined	15:46
dobby156	offby1: thank you	15:46
	← Vampire0 left	15:47
kai	if I want to write a hook that takes the references a user just pushed to the repository, makes sure those pass the unit tests and only commits them if they do pass, what hook would I want to use?	15:47
	→ Vampire0 joined	15:47
	spyraway → spyro	15:48
	→ antoniojasr joined	15:48
kai	and can I check out a tree with the user's sha1 even if that's not in the repo yet?	15:48
	→ jaequery joined	15:49
talexb	kai I think you want a pre-commit hook for running the user tests.	15:49
PerlJam	kai: what you're attempting sounds misguided to me (good idea, applied poorly)	15:49
	→ stodan joined	15:49
	→ n8o-mba joined	15:49
kai	PerlJam: why misguided? I think continous integration that checks after the tree is broken is the wrong approach	15:49
talexb	kai I'm not sure who you'd want to check our a tree as you've described .. if you set up a hook to be triggered by a user's command, you are running as that user.	15:49
	← stoffus left	15:50
codebeaker	kai: what happens when your team run the tests, anyway … and then have to wait for another test run to do their commit?	15:50
	one of the best things about git is that it's fast to commit	15:50
	don't ruin that for your team	15:50
kai	codebeaker: my tests currently take 0.11s	15:50
PerlJam	kai: see talexb and codebeaker :)	15:50
	→ kukks joined	15:50
kai	it takes me hours to fix breakages to the tree	15:50
	← arvind_khadri left	15:51
codebeaker	putting road blocks in place for your team isn't the answer	15:51
	education, and support is	15:51
kai	ok, so here's the thing	15:51
codebeaker	unless you don't respect them	15:51
	→ nyuszika7h joined	15:51
kai	it's an academic software project. I'm in no position to fire people who don't play along, because they're working in other departments at other universities	15:51
	→ ew0 joined	15:51
	→ SegFaultAX joined	15:51
kai	I do control the repository, though, so that's my only lever	15:52
	← ew0 left	15:52
	← tuxx- left	15:52
	→ jrm2k6 joined	15:52
codebeaker	also kai, regarding your point about CI being "too late" - just treat the CI as the "working" branch, and have the CI merge anything to master that passes your standard requirements	15:52
	that's much cleaner, imo	15:52
talexb	kai What codebreaker said.	15:52
kai	I know this is trying to fix a people problem by technical measures, but it's all I have	15:52
talexb	kai It's the wrong approach.	15:52
kai	hmm	15:52
	→ tuxx- joined	15:53
kai	can I restrict who's allowed to push to master?	15:53
	or do I need to set up personal repositories that the CI system pulls from?	15:53
	← SegFaultAX left	15:53
codebeaker	you could create a pre-commit hook to reject commits on master	15:53
	→ baburdick joined	15:53
	← lamont left	15:53
bremner	gitolite can enforce that kind of thing	15:53
codebeaker	and encourage them to push on a "review" branch	15:53
	→ SegFaultAX joined	15:54
PerlJam	kai: Also, by never letting the branch get broken you miss the opportunity for publicly humiliating the person(s) who broke it! ;)	15:54
kai	PerlJam: well, given it's a private repo, I don't really have that	15:54
	← muv left	15:54
PerlJam	kai: sure you do ... it's just that your version of "public" is smaller than the general idea of public	15:54
	→ erik_tr joined	15:55
	← baburdick left	15:55
kai	well, it's too small, I guess. at least the satisfaction of complaining about the broken branch doesn't make up for the productivity loss of having to spend a day or two to fix the branch	15:56
	← neurodrone left	15:56
codebeaker	kai: would it be unreasonably time consuming to "review" branches?	15:56
	← ericlee_home left	15:56
codebeaker	and have your committers work on pull request/ something based workflow	15:56
	→ ArseneRei joined	15:56
codebeaker	ala-github (yes, I'm aware some of the Git beards frown on the pull req. methodology)	15:57
bremner	no, we just frown on thinking it requires github ;)	15:58
kai	codebeaker: well, the politics here are a bit touchy. I think I can sell a CI setup as a 'quality improvement' step, I doubt I can sell the 'you don't get commit rights anymore because you keep breaking the tree'	15:58
	→ lamont joined	15:58
bremner	kai: just make a seperate branch for CI then	15:58
kai	even though this'd be the single step that would fix my problems :)	15:58
codebeaker	kai: having worked in a team where this was a problem, put everybody on a mailing list, and mail broken-build notifications from the CI to everyone	15:59
	and when it becomes an issue (soon) start taking names, and bringing people up on it	15:59
regedit	why am i getting errors when pushing to a bare remote? "fatal: remote end hung up unexpectedly"..."error in sideband demultiplexer"..."died of signal 13"..."failed to push some refs to '/home/foo/bar/'" and yet the push seems to have been successful, i even pulled to another repo to check the history - looks ok	15:59
codebeaker	and remember `git revert` is politically insenstive	15:59
	→ imarcusthis joined	16:00
	→ ipalaus_ joined	16:00
codebeaker	have your CI revert the commit if it doesn't pass	16:00
	← Sitethief left	16:00
	→ ofus joined	16:00
	← kennethreitz left	16:00
	ChanServ set mode: +v	16:01
	← jwmann1 left	16:02
	← Tianon left	16:02
	→ jwmann joined	16:02
	→ Tianon joined	16:03
	← Tianon left	16:03
	→ Tianon joined	16:03
	← gusnan left	16:03
	← erik_tr left	16:03
	→ tjh joined	16:04
	jwmann → blah	16:04
	blah → jwmann	16:04
	→ ScottE joined	16:04
	← phluks left	16:04
	→ kennethreitz joined	16:05
	→ giallu joined	16:05
	← giallu left	16:05
	→ giallu joined	16:05
	→ CannedCorn joined	16:07
	→ ericdc joined	16:07
	→ tewecske joined	16:08
	← tewecske left	16:08
	→ workmad3 joined	16:08
	← CannedCorn left	16:09
	→ Flogalen joined	16:09
	← mrsrikanth left	16:10
kai	codebeaker: thanks for the input, I'll see if I can get the CI push working revisions to a "curated" branch	16:10
	→ nsevn joined	16:10
	→ kenperkins joined	16:10
codebeaker	kai: sorry if you feel "shot-down" but you can't be "that guy" - plus there's more tactful ways to embarrass your hopeless users :)	16:10
	and CI rejecting anything that fails tests seems reasonable to me :)	16:10
kai	codebeaker: no, I think you're right	16:11
	← Flogalen2 left	16:11
	← tiglog left	16:11
	→ Ryback_ joined	16:11
	← Guest93405 left	16:12
kai	basically all I'm trying to fix is that I need to spend too much time to work on a tree that doesn't run the tests after a fetch/rebase	16:12
	→ necromancer joined	16:12
kai	and your suggestion gives me a tree that will work like that	16:12
regedit	i'm getting these errors when pushing to a bare remote http://pastebin.mozilla.org/1376656 and yet the push seems to have been successful (i pulled from the bare to another repo, history looks ok) any help please?	16:13
	→ Guest64080 joined	16:13
	← mattalexx left	16:13
kai	I don't actually care if that tree is called "master" or "bob"	16:13
	← gebi left	16:13
	← cbreak-work left	16:13
nsevn	hey- i'm a new git user and i'm having trouble figuring out the right way to organize my project in git	16:13
	my project combines three projects that are in other git repositories	16:14
	→ Drakonite2 joined	16:14
nsevn	each project will be in its own subdirectory at the root of my project	16:14
	→ Patabugen joined	16:14
	← Guest64080 left	16:14
	← axl_ left	16:14
kai	nsevn: do those subprojects change often?	16:14
nsevn	submodules don't seem to make sense because i'll be making a lot of local changes to each project	16:14
kai	ah :)	16:14
	← ecthiender left	16:14
	← antoniojasr left	16:15
nsevn	and different repositories don't seem to make sense because very often a commit will relate to all three projects at once	16:15
	→ intripoon_ joined	16:15
nsevn	but i do want to be able to merge changes from the external projects occasionally	16:15
	→ Drakonite3 joined	16:15
nsevn	i feel like i'm looking for some kind of "merge into directory" command	16:16
Patabugen	Hey. I've made some changes (including some file renames) and commited them. I then realised I should have done a merge first. What's the best way to sort everything out?	16:16
	← Drakonite left	16:16
	← Drakonite3 left	16:16
	→ axl_ joined	16:16
nsevn	where i can track each project as a local branch and then merge it into the corresponding subdirectory in my master branch when i need to	16:16
	does something like this exist?	16:16
Patabugen	I've been playing with git-diff and patch, but that's getting confused because of the file renames	16:16
	← MUILTFN left	16:16
	→ ehsan joined	16:17
nsevn	Patabugen: i think you want to do "git reset --soft HEAD^" to undo the commit	16:17
SethRobertson	nsevn: Perhaps you should consider gitslave as an alternative to git-submodule, depending on your development workflow it may be better (or worse).	16:18
nsevn	Patabugen: then "git stash save" to save all your changes	16:18
SethRobertson	nsevn: gitslave (http://gitslave.sf.net) is useful when you control and develop on the subprojects at more of less the same time as the superproject, and furthermore when you typically want to tag, branch, push, pull, etc all repositories at the same time.	16:18
	nsevn: git-submodule is better when you do not control the subprojects or more specifically wish to fix the subproject at a specific revision even as the subproject changes upstream	16:18
nsevn	Patabugen: then do your merge and restore your changes with "git stash apply"	16:18
	→ mikepack joined	16:18
Patabugen	nsevn ahh thank you! I'll give that a go!	16:19
	← Drakonite2 left	16:19
	← Cope left	16:19
nsevn	SethRobertson: thanks, i hadn't heard of gitslave, i'll give it a look :)	16:20
	← variable left	16:20
	← tommyvyo left	16:20
	← ouyes left	16:21
	→ MUILTFN joined	16:21
regedit	will git tell me "error: failed to push some refs to '/home/foo/bar'" if a post-receive hook fails on the receiving end of the push (such as not reading its STDIN or wtvr) ?	16:21
thiago	regedit: no	16:21
	regedit: post-receive failure does not count as Git error	16:22
	regedit: pre-receive failure does	16:22
SethRobertson	It does print the output of the script, though	16:22
thiago	regedit: but you get the log	16:22
	→ tommyvyo joined	16:22
	← Praise left	16:22
	→ ShadesofGraylin\| joined	16:22
regedit	i can blame most of the errors here http://pastebin.mozilla.org/1376656 on post-receive failure. but the last error still bothers me "error: failed to push some refs to '/home/foo/bar'"...	16:23
	→ Praise joined	16:23
regedit	why did that error happen and how come it looks as if the entire push DID succeed?	16:23
	(i pulled from the pushed repo and checked the history, looks like the push was ok)	16:23
	← verto left	16:24
	→ Kulrak joined	16:24
	← Kulrak left	16:24
	→ Kulrak joined	16:24
	← nwp__ left	16:24
	← Dreadnaught left	16:24
	→ bob_ joined	16:24
	← tuxx- left	16:25
	← ISF left	16:25
	→ tuxx- joined	16:25
jwmann	Can anybody explain this for me? http://d.pr/2jYn	16:26
	→ alvaro_o joined	16:26
	→ platzhirsch joined	16:27
yrlnry	You pushed to origin, but you should have pushed to upstream.	16:27
	jwmann^^	16:27
regedit	thiago: any idea? ^^^	16:27
	← ipalaus_ left	16:27
jwmann	yrlnry: upstream is not my repo	16:27
yrlnry	jwmann: what exactly is it that you want to have explained?	16:28
jwmann	my master is a forked repo of upstream/master	16:28
	I pulled from upstream	16:28
	and somehow I'm AHEAD by 24 commits.	16:28
	→ gebi joined	16:29
yrlnry	jwmann: When it says "upstream/master", it means the local branch named "upstream/master" that is supposed to be tracking the actual master branch on the "upstream" remote.	16:29
Patabugen	nsevn: That's still giving me trouble because of the renames. It seems to be doing the rename before it does the modifications. Any ideas?	16:30
yrlnry	jwmann: when you "git pull upstream master" it gets the changes from the upstream master and pulls them into the local "master" branch and moves "master" forward 24 commits. But it doesn't change the "upstream/master" ref that is supposed to be tracking the upstream repo; that stays where it was.	16:30
	So master is now ahead of "upstream/master" by 22.	16:30
	24.	16:30
	→ codetroll joined	16:30
yrlnry	It is not, of course, ahead of the actual upstream master.	16:30
	Did that make sense?	16:30
	→ Liquid-Silence joined	16:31
	← alvaro_o left	16:31
	← traviscj left	16:31
	→ dcarper joined	16:31
	→ Sico joined	16:32
	→ ReekenX joined	16:32
	← SicoWork left	16:33
	→ poppiez joined	16:33
	← ks_Marian left	16:34
	← bburhans left	16:36
	← ShadesofGraylin\| left	16:36
	→ marvin__ joined	16:36
	← _aeris_ left	16:37
Patabugen	Can I stash some changes, but when I apply them tell git to forget about the file renames and put the things back where they were?	16:37
	← platzhirsch left	16:37
	← JeffJohnson left	16:37
	ChanServ set mode: +v	16:37
	→ Prasad joined	16:37
	← mquin left	16:37
	→ _aeris_ joined	16:37
	→ bburhans joined	16:38
marvin__	hi	16:38
	→ Hounth joined	16:38
	→ dreiss joined	16:38
	→ JeffJohnson joined	16:38
	← JeffJohnson left	16:38
	→ JeffJohnson joined	16:38
talexb	Patabugen, Your question's not really clear.	16:38
	← bburhans left	16:38
talexb	If you want to pop a stash, then rename files after that, go ahead and do that.	16:39
Patabugen	The problem I'm in is that I've made some changes including some file renames. I've then realised I should have done this after merging from master	16:39
marvin__	I got some vendor linux kernel code, I checkout the mainline tree and unpacked the vendor code into the working tree.	16:39
Patabugen	I've tried using git-diff and patch, and stashing the changes first. But it all fails because I've got renamed files	16:39
	so it's not merging the file, then moving it, it's just creating the old one again	16:39
marvin__	how to tell git diff to ignore the mode changes?	16:39
	or how to tell git to restore only the mode of all files it knows already off?	16:40
Patabugen	So I'm thinking if I can stash the changes leaving the files where they were before, the merge will be better and I can move them afterwards	16:40
talexb	Patabugen, It sounds like you know what you have to do.	16:40
Patabugen	I don't know how to apply the stashed changes without the renames though	16:41
	← _md left	16:41
talexb	Patabugen, How about starting a branch, popping the stashed changes, renaming the files back to what they should be, then merging that branch into the master?	16:42
Patabugen	I guess I could	16:42
	I think it'll be easier just to make all the changes again	16:43
	→ bburhans joined	16:43
	→ _md joined	16:43
Patabugen	they weren't hard, just lots of re-naming :)	16:43
	← giallu left	16:43
Patabugen	In the process I have learned how to move files without git losing their history though... so it's been worth the time	16:43
talexb	Right .. git mv is your friend.	16:43
	← yuriks_ left	16:44
	→ yuriks_ joined	16:44
	yuriks_ neves uses git mv	16:44
yuriks_	:[	16:44
bremner	git mv is a placebo	16:44
Patabugen	Will git mv also keep the history, even if I make changes in the same commit?	16:44
	→ traviscj joined	16:44
bremner	rename detection does not depend on git mv at all afaik	16:44
	→ baburdick joined	16:44
	→ drev1 joined	16:44
Silex	git mv is a shortcut for "cp oldname newname; git add newname; git rm oldname" afaik	16:45
bremner	that matches my understanding	16:45
jwmann	yrlnry: No that's doesn't make sense	16:46
	yrlnry: Do you think you could explain it? :(	16:46
bremner	I'm not even sure if it matters if the add and remove happens i the same commit.	16:46
Patabugen	I know you lose the history if you rename and edit in the same commit	16:47
	This article says if you do it in two commits you're OK http://blog.bauermann.eng.br/2008/01/03/take-care-when-renaming-files-in-git/	16:47
regedit	thiago: i deactivated (renamed) the post-receive hook and now pushes work just fine, otherwise i get "error: failed to push some refs to '/home/foo/bar'". given the fact that it doesn't even say [rejected] or such, i'm assuming this is a direct result of post-receive failures...	16:47
bremner	Patabugen: I suspect it depends how much you edit.	16:48
	→ jceb joined	16:48
Patabugen	I don't think it matters, though I've not specifically looked I have noticed that history was lost	16:48
	← jaequery left	16:49
	← AaronMT left	16:49
	→ jaequery joined	16:49
bremner	Patabugen: have you noticed renames detected with e.g. 85% written after them?	16:49
	→ msponge joined	16:49
bremner	those are combined edit+renames.	16:49
	→ AaronMT joined	16:49
Patabugen	Yeah, but then a git log -- path/to/file shows a shortened history	16:49
bremner	anyway, it's probably a good idea to do the edits in a seperate commit, independently of how it interacts with the rename detection heuristics	16:50
	JEEBcz → JEEB	16:50
	← JEEB left	16:50
	→ JEEB joined	16:50
Patabugen	I think I will be doing so from now, it's caused quite a pain to have not done it this afternoon	16:50
	→ flavius joined	16:51
	ChanServ set mode: +v	16:52
bob_	Why is there no 1.7.7.2 at http://code.google.com/p/msysgit/downloads/list ?	16:52
	← dreiss left	16:52
jast	I think the msysgit maintainers don't build every single release	16:53
	anyway, they're human and don't always build everything instantly	16:54
bob_	@jast thanks	16:54
poppiez	hi, might be a stupid question: I have a reposity that hosts my main website. here all my JS libs, default CSS classes etc. are included. now, I've created a new repo that I want to use as a submodule. for the repo to work, I have to include libs that my main website already uses. how would that work in practice?	16:54
	as I see it, I'll have files that overlap	16:54
	and if I provide a build script in my repo uses as a submodule, all the JS will be bundled as a single file	16:55
	→ bitkiller joined	16:55
	← Praise left	16:56
	→ wbednarski joined	16:56
	→ Praise joined	16:57
	← EugeneKay left	16:58
	← dharrigan left	16:59
	← mklappstuhl left	16:59
	← EnderMB left	16:59
	→ EugeneKay joined	16:59
	← Cipher-0 left	17:00
	→ phlipper joined	17:01
	← dr_lepper left	17:01
	→ tshauck joined	17:01
	← Praise left	17:02
	→ Jay_Levitt joined	17:02
	← harshpb left	17:03
	→ harshpb joined	17:04
	→ SicoWork joined	17:04
	← whitman left	17:04
	← Sico left	17:04
	→ vdv joined	17:04
	← marvin__ left	17:05
	→ Praise joined	17:05
	← psoo_ left	17:09
	← ArseneRei left	17:09
	← Praise left	17:09
	→ Praise joined	17:10
	← Praise left	17:10
	→ Praise joined	17:10
regedit	is there a linux command for "do nothing" ?	17:10
	← teweWork left	17:10
mmattice	true	17:11
	→ arvind_khadri joined	17:11
	← arvind_khadri left	17:11
	→ arvind_khadri joined	17:11
regedit	hm that works, thanks!	17:11
	→ Galaxor joined	17:11
thiago	:	17:11
	← MacGyverNL left	17:11
	← jrm2k6 left	17:12
	← baburdick left	17:12
	← mishok13 left	17:12
thiago	: works everywhere, but true is a Linuxism	17:12
	→ MacGyverNL joined	17:12
	← MacGyverNL left	17:12
	→ MacGyverNL joined	17:12
jast	the BSDs have true as well	17:12
	← scottschulthess left	17:13
	→ teweWork joined	17:13
regedit	thiago: it seems that a post-receive hook failure does not really stop a push from completing, but a (misleading) error "error: failed to push some refs to '/home/foo/bar'" still shows	17:13
	→ baburdick joined	17:13
regedit	it is (apparently) false; all refs have been successfully pushed	17:13
mmattice	thiago: : is a shell builtin. not a 'command'	17:14
regedit	how do i submit a bug to git?	17:15
Galaxor	Hi. I had a repo I was working on. I decided I wanted to share it, so I made a bare copy with git clone --mirror. It's all good, except that I had a development branch checked out in my work tree when I made the clone. Therefore, now, when you make clones off of the bare repo, you don't get master, you get my dumb development branch.	17:15
jast	regedit: post to the mailing list, see link in topic	17:15
Galaxor	Is there a way to change the default branch in the bare repo?	17:15
jast	Galaxor: you need remote shell access or something	17:15
	if you have that: git symbolic-ref HEAD refs/heads/otherbranch	17:15
Galaxor	jast: Aha. Glorious, thank you very much.	17:16
	→ alphacow joined	17:16
	→ mandric joined	17:17
	→ Raging_Hog joined	17:18
regedit	lol "true command- Do nothing, successfully"	17:18
	← Hounth left	17:18
	← wbednarski left	17:19
	← MUILTFN left	17:19
	← Liquid-Silence left	17:20
	← hwrdprkns__ left	17:20
	← Nikoli left	17:20
	→ Antisoche joined	17:21
	→ Liquid-Silence joined	17:21
SethRobertson	poppiez: have you figured it out?	17:22
	→ ipalaus_ joined	17:23
	← bob_ left	17:23
	→ hwrdprkns_ joined	17:23
	← flavius left	17:23
	→ MUILTFN joined	17:24
	← airborn left	17:24
	→ kerrick joined	17:25
	← Liquid-Silence left	17:26
	← kenyon left	17:26
	← shachaf_ left	17:26
	← ipalaus_ left	17:26
	→ kenyon joined	17:27
	→ shachaf joined	17:27
	→ rhl joined	17:27
	← hwrdprkns_ left	17:27
	→ innociv joined	17:28
	→ marvin__ joined	17:28
	→ MANCHUCK joined	17:28
	← tshauck left	17:28
	← MUILTFN left	17:29
	→ ipalaus_ joined	17:29
	ChanServ set mode: +v	17:29
marvin__	any ideas what I can do about the wrong file modes?	17:29
MANCHUCK	how do I "merge" multiple submodules into 1 directory	17:29
	← innociv_ left	17:30
	→ kedare joined	17:30
	→ kenichi joined	17:31
	→ DofLeFou joined	17:31
	→ psoo joined	17:31
	← ReekenX left	17:32
	← mirage335 left	17:32
	→ maletor joined	17:32
	← nicoulaj left	17:32
	→ hwrdprkns joined	17:32
	→ infogulch joined	17:32
MANCHUCK	or better yet how can i have multiple submodules be in the same path?	17:32
	→ MUILTFN joined	17:34
	← ipalaus_ left	17:34
	← omn left	17:35
	← kR15T0 left	17:35
	← mandric left	17:36
	→ Araxia_ joined	17:36
	← babilen left	17:36
jast	MANCHUCK: not possible	17:37
PerlJam	MANCHUCK: perhaps you could explain in more detail why you want what you want?	17:37
jast	you can create a new repository and merge two others together in it, then use that as a submodule	17:37
	not terribly maintenance-friendly, though	17:38
MANCHUCK	jast, I am using doctrine and I pulleld it in as a submodule however that has its own submodules	17:38
jast	oh.	17:38
	well, you can nest submodules recursively	17:38
	← hwrdprkns left	17:38
	→ gusnan joined	17:38
	→ yakko joined	17:38
	← MUILTFN left	17:38
	→ ipalaus_ joined	17:39
	← charco left	17:39
MANCHUCK	how do I do that>	17:39
marvin__	okay. try mailing list then	17:39
	← niloc132 left	17:39
	← teweWork left	17:39
	← _kud left	17:39
	← jceb left	17:39
	← Octalot left	17:39
yakko	hey guys, 3 commits ago I did some stuff and I regret, how can I remove that commit's sha1 while still keeping the latest commits?	17:40
	← JSHANAB_ left	17:40
	← gebi left	17:40
marvin__	not possible, yakko	17:40
yakko	marvin__, since it's a one file thing, can I unstage those somehow?	17:41
jast	uh what? of course it's possible	17:42
yakko	I don't mean to cherry pick	17:42
jast	it's only recommended if you haven't shared that commit yet, nor the newer ones	17:42
	← baburdick left	17:42
yakko	I just wanna kill a commit ( I have it's sha1)	17:42
	→ hwrdprkns joined	17:42
jast	yes, but have you published that commit or the newer ones?	17:43
yakko	I'm the only developer in the project	17:43
jast	if you haven't, we can make it like the commit never happened	17:43
	all right	17:43
	the solution is git rebase. see the section 'interactive mode' in its manpage.	17:43
marvin__	git rebase, for unpublished trees. but wont this change the other commit ids?	17:43
	→ baburdick joined	17:43
jast	interactive rebase is pure magic. :)	17:43
	sure will	17:43
yakko	I can checkout to a later one and remove it	17:43
	→ luxurymode joined	17:44
	→ airborn joined	17:44
marvin__	git rebase to the previous commit id that you want to remove	17:45
jast	look, just use interactive rebase. remove the messed up commit from the list. done.	17:45
	← psoo left	17:45
MANCHUCK	oh its just running git submodule update --recursive --init	17:46
	→ pidus joined	17:46
	→ dreiss joined	17:47
	← whodevil left	17:47
	→ psoo joined	17:47
	← asakura left	17:47
	← iknite left	17:47
	← ngulden left	17:47
yakko	marvin__, so "git rebase 9346327462398478293" will remove those changes?	17:48
jast	no.	17:49
	→ mirage335 joined	17:49
jast	reading the section about interactive rebase will remove those changes. indirectly. :)	17:49
	→ asakura joined	17:49
	→ MUILTFN joined	17:49
jast	git rebase <ID of broken commit> just won't do anything at all.	17:49
marvin__	correct	17:50
yakko	what am I missing?	17:50
marvin__	do an interactive rebae	17:50
	← iXeno left	17:50
	→ tystr joined	17:50
jast	the part where you read about interactive rebase.	17:50
	← tystr left	17:51
	← roflin left	17:51
	→ Turicas joined	17:51
	→ babilen joined	17:51
	→ tystr joined	17:52
	→ roflin joined	17:52
	→ canton7 joined	17:52
	← traviscj left	17:53
	→ Dreadnaught joined	17:53
	← Dreadnaught left	17:53
	→ Dreadnaught joined	17:53
	→ EricDalquist joined	17:54
	← Kulrak left	17:54
	ChanServ set mode: +v	17:54
	← benrush left	17:54
	→ cytrinox joined	17:54
	← kerrick left	17:55
yakko	there were people asking me something IRL	17:55
	ok interactive rebase, I'll give it a read	17:55
	← padams left	17:55
marvin__	git rebase -i commit-id~1 or something	17:56
	← workmad3 left	17:56
	← arvind_khadri left	17:58
	flyz_ → flyz	17:58
	→ Chillance joined	17:58
	← achadwick left	17:58
	→ ksawicki joined	17:58
	← iPHP left	18:00
EricDalquist	Is there a way to grab a folderfrom project A and move it into project B?	18:00
marvin__	I got some vendor linux kernel code, I checkout the mainline tree and unpacked the vendor code into the working tree.	18:00
	I got some vendor linux kernel code, I checkout the mainline tree and unpacked the vendor code into the working tree.	18:00
	how to tell git diff to ignore the mode changes?	18:00
PerlJam	EricDalquist: just like you would with any other folder.	18:00
marvin__	or how to tell git to restore only the mode of all files it knows already off?	18:01
	← Suresht left	18:01
marvin__	i want to find out what they did change. but git diff is showing me all the file modes changes as diffs. probably the tar was packed incorretly	18:02
	→ PySloan joined	18:03
	← dobby156 left	18:04
	→ solofight joined	18:04
	← ksawicki left	18:04
	→ mandric joined	18:05
	→ ksawicki joined	18:05
bremner	marvin__: look for core.filemode in man git-config	18:05
jast	marvin__: automatic message the 'git-config' manpage can be found at <http://git.kernel.org/?p=git/git-htmldocs.git;a=blob_plain;f=git-config.html;hb=master>	18:05
solofight	people when git is being installed it asks me to close other applications (the first screen of the installation process displays it)- any idea why ?	18:05
	thought i didnt close any application it installed successfully	18:05
marvin__	if git doesn't have that functionality, it would be great if somebody add it	18:05
	← EricDalquist left	18:05
bremner	solofight: that sound like a windows/msysgit thing, not git per se	18:06
canton7	solofight, what are you using to install git? chances are it wasn't made by the git devs	18:06
bremner	jinx :)	18:06
	← madduck left	18:06
	← envi_ left	18:06
solofight	bremner, yes windows git	18:06
	canton7: am installing git on windows	18:07
	→ big_dog joined	18:07
	→ gebi joined	18:07
	← _md left	18:07
	→ madduck joined	18:08
	→ mishok13 joined	18:08
marvin__	excellent. filemode is my friend. thanks :-)	18:08
	← tatsuya_o left	18:09
	← codebeaker left	18:10
	← big_dog left	18:12
	← Jay_Levitt left	18:12
yakko	thanks for the rebase people people :D	18:12
	I'd never use that command, dead simple	18:12
	:DDDDDDD	18:12
	I was so afraid	18:13
	← baburdick left	18:13
	→ prasadkr joined	18:13
	→ baburdick joined	18:13
marvin__	this is also why you probably want to work on a seperate branch, but master :-)	18:13
	→ traviscj joined	18:14
	← vdv left	18:14
	← Prasad left	18:14
	← stodan left	18:14
	→ big_dog joined	18:14
	→ soc42__ joined	18:14
	← marvin__ left	18:15
	→ EricInBNE joined	18:16
	← soc42_ left	18:16
	prasadkr → Prasad	18:17
	→ amitayh joined	18:17
	→ niloc132 joined	18:18
	← fermo left	18:18
	← big_dog left	18:18
	← cyphase left	18:18
	← poppiez left	18:19
	← Galaxor left	18:19
talexb	:( Just tried git-stitch-repo, and saw 'git-fast-import died of signal 9' after about 20 minutes. Swap was at 14.5G last time I looked.	18:20
	→ big_dog joined	18:20
	← amitayh left	18:22
	→ codebeaker joined	18:23
	← nsevn left	18:23
	→ k_89 joined	18:24
k_89	what hashing algo does git use .. and does it use it per file basis??	18:25
	← infogulch left	18:25
talexb	k_89, I think git uses SHA1 for hashing.	18:25
k_89	talexb, k .. thanks	18:26
talexb	Each commit has a hash value .. not sure about whether each file has its own hash value as well.	18:26
EugeneKay	Yes.	18:26
	Everything is hashed.	18:26
k_89	it should .. how else would it check for changesi n files	18:26
	→ shruggar joined	18:27
	← big_dog left	18:27
k_89	hmmm... i'm gonna make a git like thing (without rollbacks) for dbs in php	18:28
canton7	k_89, it's more than that. !bottomup	18:28
jast	k_89: automatic message 'Git from the bottom up' starts with explaining the building blocks of git and proceeds to tell you how they fit together. http://ftp.newartisans.com/pub/git.from.bottom.up.pdf	18:28
	→ _kud joined	18:28
	← guardian left	18:28
EugeneKay	k_89 - make sure to look up libgit2 and php-git on Github.	18:30
	→ gragoery joined	18:30
k_89	EugeneKay, k .. thanks	18:30
	← codebeaker left	18:30
	→ cyphase joined	18:30
	← gragoery left	18:31
	← shruggar left	18:31
	← niloc132 left	18:32
	← berserkr left	18:32
	← k_89 left	18:32
	← _kud left	18:33
	← cannonball left	18:35
	→ doshi joined	18:35
	← PigeonFriend left	18:37
solofight	people apon seeing people not replying as soon as i said i installed git on windows i understand i am in the wrong room ? if correct please point me to the right direction ?	18:37
	my question was	18:37
	people when git is being installed it asks me to close other applications (the first screen of the installation process displays it)- any idea why ?	18:37
	thought i didnt close any application it installed successfully	18:37
	→ CannedCorn joined	18:37
	← CannedCorn left	18:38
talexb	solofight, I'm not that familiar with Windows, but it doesn't always behave well during installations. Contention of the registry, perhaps?	18:38
	← tvw left	18:39
	→ Jay_Levitt joined	18:40
solofight	talexb: i have around 12 programs opened and i was still able to install it succesfully	18:40
	?	18:40
	64MAA169I → LiohAu	18:40
talexb	solofight, No idea. Ask Microsoft.	18:40
solofight	talexb: ok thanks for your time	18:40
bremner	err, whomever makes git-for-windows	18:40
cbreak	solofight: windows is weird	18:41
canton7	solofight, that's the juristiction of the msysgit team. It's probably just for safety. Lots of programs seem to do it	18:41
bremner	where if FauxFaux when we need him to defend windows	18:41
FauxFaux	Watching Harry Potter and talking to my significant other.	18:41
	→ Cipher-0 joined	18:42
	spyro → spyraway	18:42
	← dr_win left	18:43
	← baburdick left	18:43
	→ baburdick joined	18:43
	← Patabugen left	18:45
	← Swimming_Bird left	18:46
	← eeadc left	18:46
	→ t4nkd joined	18:46
	→ eeadc joined	18:46
	→ ColKurtz joined	18:47
	← msponge left	18:49
	→ _iron joined	18:50
	→ kerrick joined	18:50
	→ sedwards joined	18:50
	← drev1 left	18:51
	← swombat left	18:51
	← Prasad left	18:51
solofight	canton7: exactly am wondering for long time now	18:51
	← innociv left	18:51
	→ variable joined	18:51
	→ drev1 joined	18:51
variable	with git-svn how do I "pull" updates ?	18:51
canton7	git svn rebase	18:52
	← psakrii left	18:52
	→ flavius joined	18:52
variable	canton7: thanks	18:52
	← sedwards left	18:52
	→ sedwards joined	18:52
	→ innociv joined	18:52
	← sedwards left	18:53
canton7	solofight, from googling, it seems to be to avoid a (remote) chance of a shared dll/exe/whatever being used by another program during installation	18:54
	→ sedwards joined	18:54
	← sedwards left	18:54
	→ sedwards joined	18:54
canton7	(if it is being used elsewhere, a reboot will be required to complete the installation)	18:54
	← sedwards left	18:54
	→ sedwards joined	18:54
	← sedwards left	18:54
	→ sedwards joined	18:54
canton7	iirc, ext uses a system where currently-in-use files can be modified, and new programs get the new version, programs using it prior to the change get the old version... So this issue doesn't occur. It's also why linux updates don't require a restart (except for kernel upgrades)	18:55
solofight	canton7: yeah lot of people say lot of things. when i googled found out that its to save memory which is required in huge amounts while installation process which also makes sense ?	18:55
canton7	could, could. I suspect that's less of an issue nowadays, though	18:56
	→ icwiener joined	18:56
canton7	I suspect the main motivation is "everything else says it, and we might as well be on the safe side"	18:56
	→ abyssal joined	18:56
	← sedwards left	18:57
	→ ProLoser\|Work joined	18:57
solofight	canton7: a shared dll/exe being used ? you mean replacing a shared dll/exe when its being used by other programs ?	18:57
canton7	possibly. On that point, I'm parroting what I've read elsewhere :P	18:58
solofight	canton7: you say ext file systems maintain multiple copies of files ? old for programs which are already using it and new one for upgraded/newly installed programs ?	18:58
	← doshi left	18:58
canton7	http://it.toolbox.com/blogs/locutus/why-linux-can-be-updated-without-rebooting-12826	18:58
	that's where I first read it, I think	18:58
	→ doshi joined	18:59
solofight	canton7: thank you. will read now	18:59
canton7	I remember another article decribing how to recover a file that you'd just deleted, but was still referenced by something... or other. Can't remember the details off the top of my head	18:59
	← abyssal left	19:00
canton7	this looks similar: http://linuxshellaccount.blogspot.com/2008/08/recovering-deleted-files-by-inode.html	19:00
jast	ext2/3/4 don't allow modification of files that are already open for writing by another application	19:00
	but it's possible to delete a file that's open for reading	19:00
canton7	jast, open for reading, then	19:00
	aha	19:00
jast	I'm not sure about modifying a file that's open for reading	19:01
	most of the time, updates don't modify the existing file but just clobber it and write a new one	19:02
canton7	it appears at least executables are loaded into memory prior to execution, so changing the file on-disk doesn't affect the one that's executing	19:02
jast	yep	19:03
	→ storrgie joined	19:03
	← p3rror left	19:03
	← MUILTFN left	19:03
	→ ner0x joined	19:03
	← _Vi left	19:03
jast	works the same way on windows... but the file is still open in both cases, and windows won't let you change/delete it	19:03
ner0x	Is there a way to ignore a directory in a file tree?	19:03
canton7	man gitignore	19:04
jast	automatic message the 'gitignore' manpage can be found at <http://git.kernel.org/?p=git/git-htmldocs.git;a=blob_plain;f=gitignore.html;hb=master>	19:04
	→ _Vi joined	19:04
	← EricInBNE left	19:04
	→ CannedCorn joined	19:04
	→ iXeno joined	19:05
ner0x	canton7: I used git ignore and it doesn't seem to be working.	19:05
canton7	ner0x, is the file tracked?	19:05
	git won't ignore tracked files/folders	19:05
ner0x	canton7: Brand new files	19:06
	→ desaila joined	19:06
ner0x	Actually, adding t/files seemed to do the trick, odd.	19:06
canton7	ner0x, what's the contents of your .gitignore, where is it, and where are the files?	19:06
	hmmm	19:06
	← kerrick left	19:06
	→ guy___brush joined	19:06
ner0x	base/t/files is what I was trying to get rid of. Added t/files to base/.git/info/exclude and it seemed to work.	19:06
	← hurikhan\|Work left	19:06
desaila	if you're on a local branch, and you do something like, "git pull origin master", does it automatically update your branch's code as well?	19:07
	← Cogito left	19:07
	→ dcarper_ joined	19:07
canton7	desaila, it will update all local-tracking branches, and merge origin/master into the local branch	19:07
	→ Cogito joined	19:07
canton7	ahem, remote-tracking branches. the ones that begin with "origin/"	19:07
	← dcarper left	19:08
desaila	ah	19:08
	→ alecthegeek joined	19:08
desaila	so how do pull origin/master into my local branch without creating a new branch?	19:08
	or is that a merge operation?	19:08
canton7	a pull is a fetch plus a merge	19:08
	it's a shortcut, if you like	19:08
	← heinrich5991 left	19:08
jast	'git pull origin master' merges into the currently checked out branch	19:08
desaila	^	19:08
	→ phishy joined	19:08
desaila	that's what i thought jast	19:08
	→ p3rror joined	19:09
canton7	hmm, i should have said "merge origin/master into the currently-checked-out branch", not "local branch"	19:09
	clarity fail	19:09
	← doshi left	19:09
	→ MUILTFN joined	19:10
	→ heinrich5991 joined	19:10
	← Vile left	19:10
	→ aspotashev\| joined	19:10
	← sheriff left	19:10
	→ UnsignedCode joined	19:10
	→ swombat joined	19:11
	→ alanp_ joined	19:11
	→ kerrick joined	19:11
	→ vdv joined	19:11
UnsignedCode	hi all, running gitosis on Ubuntu for personal projects… do I have to connect to the server every time I want to 'create' a new project from Windows, and create it there first, edit the gitosis-conf file etc...	19:12
	← kriegerod left	19:12
	→ alecthegeek_ joined	19:12
kevlarman	UnsignedCode: no, in fact you shouldn't ever do that	19:12
	← Kotoc left	19:12
kevlarman	UnsignedCode: clone the gitosis-admin repo	19:12
	← alecthegeek left	19:13
	alecthegeek_ → alecthegeek	19:13
kevlarman	and edit there	19:13
	→ carutsu joined	19:13
kevlarman	when you push, the server's configuration updates	19:13
UnsignedCode	but i need to do something on the server for every project?	19:13
kevlarman	UnsignedCode: also, !gitosis	19:13
jast	UnsignedCode: automatic message gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite	19:13
UnsignedCode	no way to say "dump this project over there"	19:13
kevlarman	UnsignedCode: you need to update gitosis-admin for every project	19:13
	→ joch_n joined	19:14
	← baburdick left	19:14
canton7	UnsignedCode, you can set up permissions for entire folders, with the create permissions (gitolite only I think?), then just push new projects to the folder	19:14
UnsignedCode	that's the kind of thing i'm looking for.	19:14
	for the short term, there's only me working on these things.	19:14
canton7	the line from my gitolite config file:	19:14
	← p3rror left	19:15
EugeneKay	gitolite has CREATOR syntax you can abuse	19:15
	→ baburdick joined	19:15
	← alanp left	19:15
canton7	repo projects/.*\nC = canton7\nRW+ = canton7	19:15
carutsu	I have an alias that works in the command line but not when added to my .gitconfig file	19:15
	here's it pristine = !git clean -fxd -e \*.kdev4 -e .kdev4/ && git reset --hard	19:15
UnsignedCode	ok, so steps to get rid of gitosis and get gitolite set up? :)	19:15
carutsu	how do I debug it?	19:15
EugeneKay	Gitolite has a migrate doc	19:16
UnsignedCode	ooooh.	19:16
EugeneKay	http://sitaramc.github.com/gitolite/migr.html	19:16
UnsignedCode	magic.	19:16
	→ doshi joined	19:16
UnsignedCode	thanks very much.	19:16
	i'll have a poke around.	19:16
	does gitolite use keys like gitosis or passwords?	19:17
	(or configurable)	19:17
kevlarman	keys	19:17
	→ workmad3 joined	19:17
UnsignedCode	thanks. i'll give this migration a go.	19:17
canton7	yeah, the concept's pretty much identical	19:17
	← doshi left	19:17
	→ guardian joined	19:18
	← Know1edge left	19:18
	→ Know1edge joined	19:19
	← gusnan left	19:19
	← stepnem left	19:21
	← nixmaniack left	19:21
	← lindenle left	19:22
	→ stepnem joined	19:22
	← workmad3 left	19:23
	← NetHawk left	19:23
	→ kevwil joined	19:23
	→ p3rror joined	19:24
	→ rzec joined	19:24
	→ marekw2143 joined	19:25
marekw2143	what's the difference between push and pushurl?	19:26
ProLoser\|Work	push url?	19:27
	i don't know of a command that has pushurl	19:27
canton7	neither does google	19:27
	← p3rror left	19:28
ProLoser\|Work	i know of git push <url> branch	19:28
ner0x	So let me git this straight. Git will add all the files to .git/objects even if they are untracked?	19:28
marekw2143	nono, but there's config parameter when doing: git config remote.origin.pushurl	19:28
ner0x	Just as a hash? Or what?	19:28
	ProLoser\|Work wonders if ner0x's pun was deliberate	19:28
canton7	ner0x, no. git ignores untracked files	19:28
jmah	ner0x: if you "git add" a file, a blob for it gets added to objects and it becomes tracked.	19:29
ProLoser\|Work	ner0x: have you worked with svn?	19:29
jmah	(even though it may not be in a commit yet)	19:29
canton7	marekw2143, aaah. I believe that's for if you want to push somewhere different to where you pull from	19:29
ProLoser\|Work	but if you remove it before committing it i think it's taken out of the objects	19:29
ner0x	Well, when I do git status, the files are not there. Git add . also does not add them. But when I upload the entire directory to my backup server, it's making a .git/objects file for each... Very odd.	19:29
marekw2143	canton7: thanks	19:30
canton7	it'll sit there for x days, won't it?	19:30
marekw2143	and what is remote.origin? is it the default remote repository, e.g. the one that is used when doing: git push	19:30
ProLoser\|Work	ner0x: why don't you just git push to a backup server?	19:30
	marekw2143: usually	19:30
ner0x	ProLoser\|Work: More than just the app that has to get backed up, it's just one folder located in centralized folder.	19:31
canton7	ProLoser\|Work, it remains in objects if you un-add an untracked file	19:31
ProLoser\|Work	but i think that's more convention than anything else	19:31
jmah	marekw2143: "origin" is the default remote name for an upstream repository, like "master" is the default branch name (but there's nothing special about the name)	19:31
ProLoser\|Work	what's in a name?	19:31
marekw2143	jmah: so when doing push then always name of remote repository should be specified?	19:31
ProLoser\|Work	would a rose by any other name still smell as sweet?	19:31
	→ p3rror joined	19:31
	← yakko left	19:31
ProLoser\|Work	probably if it was called shit-tulips	19:32
	→ sudanix joined	19:32
canton7	marekw2143, look at push.default	19:32
	← swombat left	19:32
ner0x	It's adding 1 byte files under .git/objects for each object, but not tracking it.	19:32
canton7	ner0x, when you do what? add the file?	19:32
jmah	marekw2143: each branch can be configured to track a certain upstream branch	19:33
	if you "git push -u myremote mybranch", the -u sets the default upstream for mybranch to myremote/mybranch	19:33
ner0x	canton7: After I've done a git add .	19:33
jmah	then plain "git push" will use that by default	19:33
canton7	jmah, only when push.default is set to upstream, which it isn't by default	19:33
	ner0x, that's right. the object is created when you add the file to the index	19:34
ner0x	canton7: But it is indeed still not tracked?	19:34
	← angelsl left	19:34
canton7	(the object is compressed. I suspect the min filesize on yoru filesystem is 1 byte)	19:34
	ner0x, it isn't tracked until you make a commit, no	19:34
ner0x	But should show up in git status?	19:35
canton7	hmm? Yes. Git status shows all (well, except ignored) files	19:35
	that are changed	19:35
	(or added, removed, untracked, etc)	19:35
ner0x	canton7: These files do not show up in git status.	19:36
	But an object file is created for them.	19:36
	Most odd.	19:36
	← soc42__ left	19:36
marekw2143	hmm, what's upstream branch?	19:36
	is it the branch which was branched ?	19:37
canton7	ner0x, what are you doing exactly? step by step?	19:37
ner0x	candybar: Well, I'd like to remove all files from t/files from being tracked, currently they are showing up in git status.	19:37
	canton7: ^ sorry	19:37
canton7	marekw2143, depends on the context tbh. In the context of push.default, it's the branch you set as the upstream branch, using push -u or branch --set-upstream. Git also sets the config automatically when you create a local branch based off a remote one	19:37
	ner0x, git rm --cached t/files ?	19:38
	→ eletuchy joined	19:38
ner0x	canton7: Okay, good so far. Now add t/files to .git/info/exclude?	19:38
canton7	commit the removal	19:39
	→ whitman joined	19:39
canton7	then add t/files to .git/info/exclude	19:39
ner0x	Alright. Done.	19:39
	→ aisipos joined	19:39
ner0x	All of the git object remain. :-/	19:40
	.git/objects have not gone away	19:40
	→ gusnan joined	19:40
jmah	ner0x: have they been committed?	19:40
	(the t/files)	19:40
ner0x	jmah: I never once have I committed them after seeing them in git status.	19:40
canton7	ner0x, the file's object in .git/objects won't go away until git gc cleans it up, which will be another 90 days or so	19:40
ner0x	canton7: Can I do it manually?	19:41
marekw2143	so for example git push origin master:satellite/master will push changes in local master to the master branch on what?	19:41
	→ wbednarski joined	19:41
marekw2143	that example is from git push help	19:41
canton7	ner0x, if the files were ever tracked, they'll always remain in .git/objects	19:41
jmah	canton7: blobs don't have a date, won't they go away immediately?	19:41
ner0x	canton7: They were never tracked.	19:41
canton7	jmah, let me test..	19:41
jmah	ner0x: try "git prune --dry-run"	19:41
	← flyz left	19:42
canton7	jmah a straight gc doesn't clear them. a git gc --prune=now did, though	19:42
ner0x	jmah: Tons and tons of blobs listed.	19:42
	→ macmartine joined	19:42
	← jonatasnona left	19:42
canton7	prune, by default, takes 90 days iirc	19:42
kevlarman	30	19:42
canton7	ner0x, git gc --prune=now	19:42
	→ Swimming_Bird joined	19:42
canton7	hmm, 14 according to the man page, kevlarman	19:43
kevlarman	weird	19:43
	i would have expected 30 just like the reflog	19:43
canton7	well, that's the value of gc.pruneexpire, which I assume is what we're after	19:43
ner0x	canton7: Looks like that worked. objects are now 1.2 MB, which sounds about right.	19:44
canton7	the reflog (gc.<pattern>.reflogexpire is 90 days, kevlarman	19:44
	← baburdick left	19:44
canton7	(which is where I got my 90 days from, under much the same assumption)	19:44
ProLoser\|Work	hey	19:45
	git stash apply --index 1 isn't working	19:45
	what's the proper order if i want the stash at index 1 (not index 0)	19:45
ner0x	canton7++ # thanks	19:45
	jmah++ # thanks	19:45
canton7	kevlarman, or actually, reflogexpireunreachable is probably what we're after, in which case you're right on that one	19:45
	→ baburdick joined	19:45
jmah	ProLoser\|Work: "git stash apply stash@{1}"	19:46
ProLoser\|Work	oh that's the stash name?	19:46
	i thought i just use the 1	19:46
jmah	this way works, I'd love to hear if there's another way	19:46
canton7	"If the --index option is used, then tries to reinstate not only the working trees changes, but also the indexs ones. However, this can fail, when you have conflicts (which are stored in the index, where you therefore can no longer apply the changes as they were originally"	19:46
	kvanderw → kvanderw\|lunch	19:47
canton7	so --index isn't what you're after	19:47
	→ shruggar joined	19:47
	← p3rror left	19:48
	→ spatulasnout joined	19:49
	→ mysphyt joined	19:49
	→ Kazii joined	19:49
	← PySloan left	19:52
	← wbednarski left	19:53
	← harshpb left	19:53
	← n8o-mba left	19:54
	← OOPMan left	19:54
	← guardian left	19:55
	← solofight left	19:55
	→ poppiez joined	19:56
	← pfsmorigo left	19:57
	← kenichi left	19:58
	← aspotashev\| left	19:58
	→ infogulch joined	19:58
	→ Nikoli joined	19:58
	← infogulch left	19:58
	→ retoryter joined	19:59
	← retoryter left	19:59
	← alecthegeek left	20:00
	→ tvw joined	20:00
	→ marthinal joined	20:00
	← rzec left	20:00
	← dfr\|mac left	20:01
	← mysphyt left	20:02
	→ n8o-mba joined	20:02
	→ p3rror joined	20:02
	← koo6 left	20:03
	→ boombatower joined	20:04
	← ipalaus_ left	20:04
	→ alecthegeek joined	20:05
	→ omn joined	20:05
	← Spockz left	20:06
	→ codebeaker joined	20:07
	← kerrick left	20:08
	← pidus left	20:09
	→ OOPMan joined	20:09
	← ner0x left	20:09
	kvanderw\|lunch → kvanderw	20:10
	→ FylGood joined	20:11
	→ workmad3 joined	20:11
	→ infogulch joined	20:11
	← dc5ala left	20:12
	← agile left	20:14
	→ giallu joined	20:14
	← baburdick left	20:15
	→ srji joined	20:15
	← ketas left	20:15
	→ ketas joined	20:15
	← comawhite left	20:15
	→ comawhite joined	20:15
	← oc80z left	20:15
	→ baburdick joined	20:15
	← p3rror left	20:16
	← nyuszika7h left	20:16
	← UnsignedCode left	20:17
	→ oc80z joined	20:17
	→ arnpr joined	20:17
arnpr	is there any support for gitolite here? it keeps asking my new user about a password, I already copied his public key and added him to admin repo	20:18
canton7	yeah, we handle gitolite queries. sitaram is normally lurking for the really complex questions, too :)	20:19
	the new user... does he have his private key set up correctly? permissions right? is he using sudo? what does ssh -vvv git/gitolite:host give?	20:19
arnpr	canton7: how can I know if his key is correctly setup? he made it with ssh-keygen	20:20
	I added it to the keydir/ and made a new project and gave him RW+	20:20
canton7	arnpr, well these are the most common ssh problems. chances are the key problems are on his end	20:21
	→ agile joined	20:21
	← the_metalgamer left	20:22
canton7	if he's being asked for a password, then either gitolite hasn't accepted his key, or he's set his side up wrong. have a look in <gitolite-home>/.ssh/authorized_keys. Don't mod anything, but see if his public key's in there	20:22
	if it is, the problem is almost certainly on his end	20:22
	→ nyuszika7h joined	20:23
	→ ldlework joined	20:23
ldlework	If I have a repository that contained a/ and b/ and I want to take b/ and all history concerning it and break it out into a seperate repo, that's pretty much flagrant wishful thinking?	20:23
canton7	git filter-branch --subdirectory-filter	20:24
ldlework	wo	20:24
	w	20:24
	← lamont left	20:24
canton7	more common than you'd think :P there are lots of SO posts on it, some tuts, and the man page gives good examples	20:24
	← codebeaker left	20:24
ldlework	canton7, what happens, if you know offhand, if a commit touched both? Is the commit present in the new history?	20:24
canton7	I believe it just takes the original repo, and rewrities it to be the subdirectory only. you need to play another trick to remove a/ from the original repo	20:25
	→ thedeeno joined	20:25
	→ dfr\|mac joined	20:25
canton7	(not a complex trick, though)	20:26
	← BlindOracle left	20:26
	→ Vile joined	20:26
	← alecthegeek left	20:26
jmah	ldlework: I did this when converting a complex repo from svn	20:26
canton7	the first two examples on the man page will cover you	20:26
jmah	you'll need a good understanding of git	20:27
	(if you want to go the whole way)	20:27
	← eletuchy left	20:27
jmah	in summary, use filter-branch --subdirectory-filter to make a new repo with that dir	20:28
	← dcarper_ left	20:28
jmah	then I made a file mapping the commits to the trees	20:28
	then in the parent repo, make a script that for each commit, copies the tree of the subdir, deletes the files, and makes a submodule ref with the corresponding commit (looking up the commit corresponding to the tree)	20:29
_ikke_	I'm running behind a firewall now, but I have access to a VPS. My plan was to make an ssh tunnel to forward the port to the server: ssh -f user@host -L 9418:github.com:9481 -N. I added a remote with the following url: git@localhost:username/repo.git. But when I try to push to it, it asks for a password for git@localhost. Why is that?	20:29
canton7	jmah, was that necessary? surely filter-branch --index-filter 'git rm --cached --ignore-unmatch a' HEAD will trips a/ from the origin repo? aaah you were setting it up as a submodule	20:29
jmah	indeed.	20:29
	not sure if Idlework wants to go that far	20:30
CareBear\	_ikke_ : user@host is the short version of scp URLs	20:30
	_ikke_ : try git://localhost/username/repo.git	20:30
	_ikke_ : I might just push to the VPS and let the VPS push to github	20:31
	→ verto joined	20:31
	→ alecthegeek joined	20:31
_ikke_	CareBear\: I would still need to create a tunnel to push to the VPS	20:31
arnpr	canton7: we remade the key and re-added it to the repo, and problem persist	20:31
	I checked the authorized_keys file but don't know what I should be looking for, canton7	20:31
canton7	arnpr, have you tried answering any of the questions I asked you?	20:32
_ikke_	CareBear\: And user@host normally works for ssh	20:32
	← ThomasLocke left	20:32
	→ btanaka joined	20:32
canton7	arnpr, in the authorized_keys in gitolite's home dir, you're after a line which begins "command=/some/path/gl-auth-command <his username>",somestuff ssh-rsa <his public key	20:33
	← alecthegeek left	20:33
canton7	arnpr, it's far, far more likely that the problem is on his end, though	20:33
CareBear\	_ikke_ : git runs transparently over ssh	20:33
	→ ThomasLocke joined	20:33
CareBear\	_ikke_ : if you can ssh to the vps and set up a tunnel then git can also ssh to the vps and transfer stuff	20:33
_ikke_	CareBear\: right	20:34
CareBear\	_ikke_ : port 9418 does not use ssh	20:34
	_ikke_ : port 9418 is git:	20:34
_ikke_	CareBear\: Got it	20:34
CareBear\	_ikke_ : git over ssh always uses port 22 like every other ssh	20:34
	← TrineLocke left	20:34
_ikke_	yup	20:34
	Still need to use 443	20:34
	22 is blocked too	20:34
	→ kerrick joined	20:34
CareBear\	that's fine	20:34
	git will run ssh	20:34
	so you just add a stanza in .ssh/config	20:35
_ikke_	ok	20:35
canton7	arnpr, quick checklist for you, on his machine: 1. does ~/.ssh/id_rsa exist? 2. does it have 600 permissions? 3. does .ssh have 700 permissions? 4. what does "ssh -vvv git@your_machine" give? (I'm assuming the gitolite user is git. Correct if it isn't)	20:35
	← ofus left	20:35
	→ lamont joined	20:36
CareBear\	echo -e 'host myvpson443\n port 443\n hostname myvpspublicip' >> ~/.ssh/config && git add remote vps myvpson443:repo.git	20:36
	_ikke_ : ^	20:36
	← workmad3 left	20:36
CareBear\	_ikke_ : then, git push vps master will end up using ssh to port 443 on the vps	20:36
	dilfridge → dilfridge_	20:37
	dilfridge_ → helium-4	20:37
	helium-4 → dilfridge	20:37
	← saschagehlich left	20:37
	dilfridge → helium-3	20:37
	← alphacow left	20:37
	helium-3 → dilfridge	20:37
	← srji left	20:38
	→ khmarbaise joined	20:38
	→ dirkle joined	20:39
	→ srji joined	20:41
	← dv310p3r left	20:42
	← nyuszika7h left	20:43
ldlework	jmah, I don't want submodules. I want a/ and b/ to be seperate repositories, both taken from the original repo	20:43
	← lamont left	20:43
ldlework	I don't need the original repo any longer, but each of the two new repos need their relevant history	20:43
canton7	ldlework, nice and easy, then. follow the example from man git-filter-branch	20:43
jast	ldlework: automatic message the 'git-filter-branch' manpage can be found at <http://git.kernel.org/?p=git/git-htmldocs.git;a=blob_plain;f=git-filter-branch.html;hb=master>	20:43
ldlework	Is it enough to run filter-branch on each subdirectory to get two new?	20:43
canton7	yes	20:43
ldlework	Nice.	20:43
	thank you	20:44
	→ ReekenX joined	20:44
jmah	just duplicate the repo first	20:44
	← sudanix left	20:44
canton7	bear in mind that filter-branch will operate on the repo you run it in	20:44
	so, yeah, duplicate it furst	20:44
	← baburdick left	20:45
infogulch	whoa my whole git directory is a friggin mess, but github is fine. how do i completely reset it to match exactly what it is on github	20:45
	→ baburdick joined	20:45
canton7	infogulch, git fetch. for each branch you want to reset, 'git reset --hard origin/branchname'	20:46
infogulch	ty	20:46
canton7	well, checkout the branch first, then reset	20:47
infogulch	oh ok	20:47
canton7	(as reset operates on the current branch)	20:47
infogulch	it's not working...	20:47
canton7	ok, so what are you running, and how do you know it isn't working?	20:47
	← Kazii left	20:47
infogulch	i only have one branch, "master"	20:47
	i'm running: git reset --hard origin/master	20:48
	and i know it's not working because when i run git status it shows a bunch of untracked files	20:48
	→ lamont joined	20:48
	← psoo left	20:48
infogulch	unless that would be fixed with something else	20:48
canton7	git won't touch untracked files. use git clone -dfx to nuke those (git clean -dxn to test what will be removed first)	20:48
	ack	20:48
	git clean	20:48
	← ThomasLocke left	20:49
canton7	can't type tonight	20:49
infogulch	ah i see	20:49
	→ LongBeach joined	20:49
	← mishok13 left	20:49
	→ ipalaus_ joined	20:50
ldlework	canton7, git clone original separated; cd separated; git filter-branch --subdirectory-filter b HEAD --prune-empty; git remote rm origin; git reset --hard; rm -rf .git/refs/original/; git reflog expire --expire=now --all; git gc --aggressive --prune=now; ?	20:51
canton7	no! git clone with that syntax will usee hardlinks	20:51
ldlework	okay, adding --no-hardlinks	20:51
canton7	either use file:/// syntax, or just us a box-ordinary copy -- that's safest :P	20:52
	*bog-ordinary	20:52
infogulch	whew much better	20:52
_ikke_	CareBear\: I'll just use https	20:52
ldlework	Not sure what a box-ordinary copy is	20:53
	→ stickyboot joined	20:53
	→ DAAN5U1 joined	20:53
	→ ahmedkotb joined	20:53
canton7	ldlework, you'll also probably want git filter-branch --subdirectory-filter a --prune-empty -- --all. Then you can clone it, which is the easiest way of cutting out the crap	20:53
	sorry, I mean bog-ordinary. as in, cp -R	20:54
	← srji left	20:55
	→ metcalfc joined	20:56
infogulch	heh, i suppose dropbox really doesn't like symlinks xD	20:56
	← marthinal left	20:56
_ikke_	infogulch: I don't have any problems with symlinks in dropbox	20:57
	← LiohAu left	20:57
	→ the_metalgamer joined	20:57
infogulch	i'm symlinking one github repo into a subdir of another	20:58
	on multiple computers...	20:58
	lets just say it wasn't pretty and that i'm glad i pushed to github last night	20:58
arnpr	canton7: worked like a charm, think it was the name of his pub key	20:59
canton7	arnpr, sweet!	20:59
arnpr	which wasnt id_rsa, changed that and eprmissions, worked	20:59
canton7	ssh problems usually boil down to 2 or 3 things. those are the main 2	21:00
	← Jackneill left	21:02
	→ nyuszika7h joined	21:05
	← the_metalgamer left	21:05
	→ alecthegeek joined	21:07
	← kerrick left	21:08
	← FylGood left	21:10
	→ workmad3 joined	21:13
	← marekw2143 left	21:15
	← GunArm left	21:16
	← baburdick left	21:16
	← mastro left	21:16
	→ mishok13 joined	21:17
	→ baburdick joined	21:17
	← alecthegeek left	21:18
	← dfr\|mac left	21:18
	→ GunArm joined	21:18
	← cburroughs left	21:18
	→ the_metalgamer joined	21:19
	→ pantsman joined	21:19
	← pantsman left	21:19
	→ pantsman joined	21:19
	← heptadecagram left	21:20
	→ mklappstuhl joined	21:20
	← cortana` left	21:20
	← GunArm left	21:21
	← iXeno left	21:22
	← Raging_Hog left	21:22
	→ dfr\|mac joined	21:22
	→ GunArm joined	21:23
	← dfr\|mac left	21:23
	← GunArm left	21:23
	→ mithridates joined	21:24
mithridates	does git use mmap to read/write/search data for its objects?	21:25
	my eyes are so tired to check the code :(	21:25
FauxFaux	"git grep mmap" in git.git returns loads of things including in sha1_file, the internals of hash-object.	21:26
	← khmarbaise left	21:27
	← workmad3 left	21:27
CareBear\	_ikke_ : why https?	21:27
	← Shusshu left	21:27
	← phishy left	21:28
	← arnpr left	21:28
	← btanaka left	21:28
	← bindaasomatic left	21:28
	← poppiez left	21:29
	→ jonnyATroot joined	21:29
	→ jacknagel joined	21:29
	← Turicas left	21:30
	← jacknagel left	21:30
	← kl4m left	21:30
	→ neurodrone joined	21:31
	→ alex__c2022 joined	21:31
_ikke_	CareBear\: Because I then can directly connect to github in stead of using my vps	21:31
	→ dfr\|mac joined	21:32
mithridates	FauxFaux: what are the internals of hash-object? what do u mean by hash-object? it's a new terminology in git!	21:32
FauxFaux	mithridates: hash-object is the thing that creates entries in the object database, in .git/objects. man git-hash-object	21:32
_ikke_	mithridates: git hash-object is a git command	21:32
jast	mithridates: automatic message the 'git-hash-object' manpage can be found at <http://git.kernel.org/?p=git/git-htmldocs.git;a=blob_plain;f=git-hash-object.html;hb=master>	21:32
jmah	mithridates: hashing is kinda core to git	21:32
	why are you wondering about mmap?	21:33
mithridates	I'm making a system similar to git's db, I do hashing for names, and I want to write/read data on files so I was wondering how git handles these stuff	21:33
	I asked in #c what's the most efficient way to do that, they said mmap so I thought git should use that too for its objects	21:34
	→ cooloney joined	21:34
	← flavius left	21:35
	← dfr\|mac left	21:35
cbreak	do you even have a clue what mmap is?	21:35
	→ dfr\|mac joined	21:35
mithridates	I think I know, it takes a chunk of memory and write it to a file	21:35
	that's the impression that I get from its name and header of its man page	21:36
cbreak	nope	21:36
mithridates	so I guess I can put my structs or whatever there	21:36
	so what is it?	21:36
cbreak	it's a method to map a file into memory	21:36
ProLoser\|Work	git stash pop like git stash apply + drop?	21:36
cbreak	instead of reading it	21:36
	← neurodrone left	21:36
canton7	ProLoser\|Work, yes	21:36
cbreak	ProLoser\|Work: yes	21:36
	→ Wilto joined	21:36
ProLoser\|Work	cool	21:37
mithridates	cbreak: oh cool, but how does it handle data types? does it make everything char * ?	21:37
	← Wilto left	21:37
cbreak	mithridates: it has advantages over simply reading a file sequentially if you do random access or the file is big	21:37
	no	21:37
	it does not handle data types	21:37
	← babilen left	21:38
cbreak	as the name says, it maps raw memory	21:38
	-> if you store things like integers, floats or even pointers, you're in deep trouble :)	21:38
	← airborn left	21:38
	← leen^\| left	21:39
mithridates	right, so whats its advantage to simple open the file and read it? does it just save the time of for each line going to the disk, getting data and read it to memory? and load the whole file at once?	21:39
	→ jtr__ joined	21:40
	→ booginga joined	21:40
	→ youngblood joined	21:40
	← mishok13 left	21:41
	→ eletuchy joined	21:41
mithridates	copy everything from the location of file on disk to memory only one time?	21:41
	like if I have 2kb file, it makes 2kb space on memory and drag and drop all data from the disk on that space?	21:41
	→ frogonwheels joined	21:42
	→ kpreid joined	21:42
	← Davey left	21:43
	→ oriba joined	21:43
	→ p3rror joined	21:44
	← stickyboot left	21:44
	← flijten left	21:44
	← Cipher-0 left	21:44
	→ airborn joined	21:44
	← CannedCorn left	21:45
	← airborn left	21:45
mithridates	I think I don't make sense right now, thanks anyway	21:45
	→ airborn joined	21:45
	→ stickyboot joined	21:46
	← baburdick left	21:46
canton7	well tbh, this is #git, not #c. cbreak knows a lot of C, but, of the regulars, he's probably in the minority. basically, don't be too surprised if you don't get answers to very off-topic questions :)	21:46
	← airborn left	21:47
cbreak	mithridates: it uses the Virtual Memory subsystem	21:47
	so it can map in and out pages as needed	21:47
	-> you use less memory, and random access is faster	21:48
	→ Kazii joined	21:48
cbreak	(and/or easier)	21:48
	page size is 4kb on Mac OS X	21:48
	I think that's similar on Linux and Windows	21:48
	→ airborn joined	21:48
	← gigamonkey left	21:49
cbreak	also, block size of new hard disks is also 4kb	21:49
CareBear\	mithridates : yes, you got it. when you use mmap, the system drags and drops the file	21:49
	that's exactly how it works	21:49
	← __class__ left	21:49
CareBear\	"it's a unix system. I know this!" (bonus points: name the film!)	21:49
	← bartek left	21:50
	→ baburdick joined	21:50
_ikke_	Jurasic Park	21:50
CareBear\	points for you	21:50
	→ alecthegeek joined	21:51
	← sangeeths left	21:51
	← estan_ left	21:51
	← boombatower left	21:52
jmah	CareBear\: keep in mind, mmap doesn't work on systems without a mouse (like servers) because the dragging is crucial	21:52
_ikke_	lol	21:52
	→ NetHawk joined	21:52
	← ldlework left	21:52
	→ koo6 joined	21:53
_ikke_	Sometimes when you can't find the mouse pointer, it's the system dragging page files	21:53
CareBear\	mithridates : mapping is a concept of exposing something in a different way	21:54
	← airborn left	21:54
CareBear\	mithridates : memory mapping is the concept of exposing something as if it was memory, even though it is not memory	21:54
jmah	mithridates: it's kinda like "mounting" a file into memory	21:54
CareBear\	mithridates : files can be memory mapped, and hardware devices can also be memory mapped	21:54
	→ airborn joined	21:54
mithridates	thanks a lot guys, I logged this conversation and I will read it later, I must go. Thanks a lot!	21:55
talexb	Crap .. did a reset --hard and lost the merge updates I'd done ..	21:55
cbreak	more specifically: the memory of devices can.	21:55
CareBear\	mithridates : this is significant, because accessing memory is very easy for programs (as you know, there is a mov instruction for this)	21:55
canton7	talexb, have you come across the reflog?	21:55
CareBear\	cbreak : not neccessarily	21:55
cbreak	talexb: merge updates?	21:55
	→ s0ber_ joined	21:55
canton7	(assuming you did the merge, then undid it...)	21:55
talexb	canton7, Not often enough :/	21:55
	← alecthegeek left	21:55
jmah	talexb: was your stuff (ever) committed?	21:55
talexb	jmah Yes.	21:56
cbreak	I wonder if he meant "Merge conflict resolutions" when he said "merge updates"	21:56
CareBear\	mithridates : so instead of a program needing to deal with the concept of bytes stored inside a file, with mmap the operating system takes care of all I/O, and makes it look like the entire file is stored somewhere in memory	21:56
jmah	what was the reset command you used?	21:56
talexb	I committed, then did another pull, which made me realize the work I'd done may not have been complete.	21:57
	← pimpin left	21:57
talexb	To reverse the merge, I did a hard reset.	21:57
	That appeared to have deleted the merges that I'd done.	21:57
	→ btanaka joined	21:57
CareBear\	git reflog	21:57
talexb	git reset --hard HEAD~	21:57
	← s0ber left	21:58
	s0ber_ → s0ber	21:58
jmah	talexb: run "git reflog", find a place that looks better, and git reset --hard to that (e.g. "git reset --hard HEAD@{1}")	21:58
	→ amalloy joined	21:58
	→ lnostdal_ joined	21:58
	← fisted left	21:58
	← ipalaus_ left	21:58
	→ nicxvan joined	21:59
	→ xeno__ joined	21:59
	← Jay_Levitt left	21:59
	→ Jonny_ joined	21:59
	← traviscj left	21:59
	Jonny_ → Guest84035	21:59
	→ rzec joined	22:00
talexb	jmah++ Wow. Cool. Thank you.	22:00
	→ niloc132 joined	22:00
mediaslave	I have a .htaccess file that needs to belong in a repo in a certian state. I also need to change this file locally, is there a way to have git ignore my changes to this file?	22:00
	→ alecthegeek joined	22:00
	← Guest84035 left	22:00
amalloy	i can't figure out the right set of parameters to pass to rebase. i want to rewrite a topic branch so that it applies cleanly (without merges) to master, and then update master to point at the new rebased head	22:00
	← oriba left	22:00
	→ fisted joined	22:00
canton7	mediaslave, advice is the commit e.g. .htaccess.sample, and ignore .htaccess.	22:00
	← OOPMan left	22:01
amalloy	i can do this with a lengthy multi-command process, but i'm sure if i just said the right magic word to rebase it would do this in one step	22:01
frogonwheels	mediaslave: can't you just put the .htaccess stuff in the config file of the server?	22:01
jast	amalloy: with the topic branch checked out, git rebase master; then go to master and git merge topicbranch	22:01
	← airborn left	22:01
	→ jtdowney joined	22:01
	→ kerrick joined	22:01
	← innociv left	22:01
canton7	mediaslave, (I've had some luck with smudge filters myself, but it's hacky)	22:01
	dilfridge → helium-3	22:01
mediaslave	frogonwheels: Yes I can and do. This is just an example. There are other files like this that I may need to do it with	22:01
frogonwheels	mediaslave: otherwise canton7's advise is the best. Have something that copies the default default.htaccess to .htaccess if it doesn't exist	22:01
mediaslave	I see...	22:02
frogonwheels	mediaslave: well have a process that copies all those other file 'defaults' if they don't exist.	22:02
amalloy	jast: right, that's the three-step (four, really, if i'm going to check topic back out in a minute) that i'm trying to avoid	22:02
frogonwheels	mediaslave: yes, there IS a way. but it's not something you should be relying on if you have a choice.	22:02
mediaslave	Okay, thanks. I was kinda hoping there was a file in the .git folder that I could mess with...	22:02
	→ airborn joined	22:02
mediaslave	frogonwheels: okay, thanks.	22:03
jast	amalloy: okay, you can reduce it to two steps if you really feel that you need to do that: after the first step (the rebase), git push . topicbranch:master. you'll notice that this is slightly hacky.	22:03
jmah	amalloy: after the rebase, you can "git update-ref refs/heads/master HEAD"	22:03
	← jonnyATroot left	22:03
mediaslave	I can use capistrano to do the defaults. idea...	22:03
frogonwheels	mediaslave: you CAN use the sparse checkout feature ..	22:03
jmah	which sets master to HEAD, but keeps the current branch checked out	22:03
jast	I recommend against using update-ref	22:03
canton7	amalloy, you can use rebase --onto, I believe, to avoid checking out the feature branch first	22:03
	→ babilen joined	22:04
jast	canton7: but it'll still checkout master at some point	22:04
canton7	jast, for sure	22:04
jmah	jast: it requires caution, but I use it a bunch, why do you recommend against it?	22:04
jast	so he has to go back to the other branch after that	22:04
	→ OOPMan joined	22:04
jast	jmah: because it requires caution, unlike the solution I suggested	22:04
canton7	jast, it saves a checkout command if he starts on master	22:04
	→ sangeeths joined	22:04
jast	canton7: I don't know that he does	22:04
	helium-3 → dilfridge	22:04
canton7	jast, then it won't help, of course :) That's why I adorned my comment with "to avoid checking out the feature branch first"	22:05
jast	canton7: also, re the list of links you suggested earlier... I'm kind of divided on including the website howto since it uses the problematic checkout -f trick	22:05
	dilfridge → helium-3	22:05
jast	the "problematic" part being that it doesn't delete files that got removed upstream	22:06
canton7	jast, it's a fair point. It's a common scenario and the link gets posted a lot, though. I guess it's between you being able to put whatever warning you like next to the link, and hoping that whoever hands the link out in the channel knows to supply a warning	22:06
	→ boombatower joined	22:06
	← sangeeths left	22:06
	→ ldlework joined	22:06
canton7	well, that against having the link be seen as "official"	22:06
	canton7 doesn't know which is best	22:07
jast	canton7: yeah, I'd like to solve it properly which would involve providing a warning, an explanation and instructions on alternatives. at that point I'd essentially have to write a new article. :)	22:07
	← helium-3 left	22:07
	→ dilfridge_ joined	22:07
canton7	heh, too true	22:07
	← tommyvyo left	22:07
	← joshsmith left	22:08
jast	generally I think I might have to setup some kind of system where people can submit content for triggers	22:08
frogonwheels	jast: I use update-ref quite often, but every now and again, I update-ref my checked out branch :( ... which is something you really have to avoid doing!	22:08
	→ __class__ joined	22:08
jast	with an overview page	22:08
	← Textmode left	22:08
canton7	that could work	22:08
jast	frogonwheels: yeah, that's why I recommend the "push to ." trick instead	22:08
	← luxurymode left	22:08
jmah	jast: what's this "push to ." trick? can I subscribe to your magazine?	22:09
canton7	hmm, couldn't you pull the branch -f trick as well?	22:09
	dilfridge_ → helium-3	22:09
	← avinashhm left	22:09
jast	that's more unsafe than push ., isn't it?	22:09
	helium-3 → dilfridge_	22:09
	dilfridge_ → helium-3	22:10
jast	jmah: the magazine is this channel. ;) "git push . src:dest" locally changes dest to match src, but only if it's possible as a fast-forward	22:10
canton7	no idea. throwing it out there to see what the more experiences gitters say about it	22:10
frogonwheels	jast: I know of that trick but haven't used it. I should use it for that specific circumstance.	22:10
	← xarragon left	22:10
	→ xarragon joined	22:10
jast	it should also prevent you accidentally updating the checked out branch, but I haven't tested that	22:10
frogonwheels	jast: It will surely, 'cause you can't push (by default) to a checked out branch	22:11
jast	yeah, I'm just not sure whether this is tested for local pushes. it should be.	22:11
	→ linusoleander joined	22:12
	helium-3 → dilfridge	22:12
	← Guest72597 left	22:13
	← xeno__ left	22:13
frogonwheels	jast: tested. behaves as expected	22:13
jast	good.	22:14
	→ ricky joined	22:14
	→ ipalaus_ joined	22:14
	← kedare left	22:14
frogonwheels	jast: and now that I've done it.. I'll remember to use it more likely than not ;)	22:14
jast	my work here is done. ;)	22:15
	← the_metalgamer left	22:16
	← guy___brush left	22:16
	→ dv310p3r joined	22:16
	← vdv left	22:16
	← alecthegeek left	22:17
	← baburdick left	22:17
	→ baburdick joined	22:18
	← whitman left	22:18
	← cytrinox left	22:18
	← ramusara left	22:19
	← w0bni left	22:19
	← Ryback_ left	22:19
	← gusnan left	22:20
	→ CannedCorn joined	22:20
	← jtr__ left	22:20
	← p3rror left	22:21
	← d0k left	22:22
	→ avinashhm joined	22:22
	ChanServ set mode: +o	22:23
	jast changed the topic to: 1.7.7.2 \| http://git-scm.com \| Please skim http://jk.gs/git/ before asking, or at least before giving up!	22:23
	jast set mode: -o	22:23
	← mklappstuhl left	22:24
	→ ESphynx joined	22:24
ESphynx	Hey guys, something I've discovered recently, when working on a side branch... it's very useful to do: 1. git rebase master 2. git rebase origin/sidebranch	22:25
	Is there any operation that kind of automates that?	22:25
	← Arrowmaster left	22:25
	ChanServ set mode: +v	22:26
	ChanServ set mode: -v	22:26
jmah	shell history?	22:27
	:)	22:27
ESphynx	hehe	22:27
	I just see find git a bit dumb sometimes :\|	22:27
	→ c00kiemon5ter joined	22:27
ESphynx	what feels like the natural workflow to me is never default	22:27
	← c00kiemon5ter left	22:27
	→ c00kiemon5ter joined	22:27
jmah	well, some people like to rebase a lot, some don't	22:28
	← chris_99 left	22:28
ESphynx	I like obvious histories... seeing merges all over when no merge really happened is annoying	22:28
	jsut because someone did something else and pushed to the repo before me shouldn't clutter the history with merges	22:29
	git rebase master : this brings him stuff tha happened on the master branch... then git rebase origin/sidebranch lets me push it to the central repo's sidebranch in a chronological order...	22:30
	no rebase option would let me do this in a single operation, is kind of my question?	22:30
jast	the problem with automatic consecutive rebases is that each rebase might involve you fixing a series of conflicts... and it can get pretty confusing if it seamlessly goes from one rebase to the next during that	22:30
	→ wbednarski joined	22:30
ESphynx	I guess so.	22:30
jmah	git rebase master && git rebase origin/sidebranch	22:30
ESphynx	I was kind of happy when I figured out that is what I needed to do.	22:31
	note that here... nothing NEW was on origin/sidebranch	22:31
	i'm just doing the rebase on master as if it happened 'after' the current sidebranch	22:32
	in fact it's really confusing to me that there is no one operation for this	22:32
jmah	why isn't just "git rebase master" enough then?	22:32
ESphynx	i.e. origin/sidebranch is the same as 'sidebranch' BEFORE the operation	22:32
	→ thursday joined	22:32
ESphynx	jmah: because then I cannot push it to origin/sidebranch ...	22:33
	← ProLoser\|Work left	22:33
jast	you're setting yourself up for a lot of pain if you ever try to merge any of this stuff	22:33
ESphynx	Am I?	22:33
jast	the reason it's different is because the second rebase takes the new commits from master and rewrites them with new commit IDs	22:33
	i.e. they're now slightly different on master vs. your side branch	22:34
ESphynx	yes.	22:34
	← joch_n left	22:34
jast	not to mention that if they contained any merges on master, those merges are now gone in the side branch	22:34
ESphynx	jast: I just tried doing a git merge sidebranch from master, and there's just 2 conflicts easily fixed...	22:35
	← ipalaus_ left	22:35
jast	well, it won't always cause pain	22:35
ESphynx	sidebranch here is temporary while performing the work.	22:35
	← amalloy left	22:36
ESphynx	what i'm trying to do with my 2 rebases... is bring in the stuff that happened on master, but make it as if it happened after whateve I have right now.	22:36
jast	this stuff gets pretty unpredictable. it depends on the degree to which the new stuff on sidebranch and the new stuff in master touch the same parts of code.	22:36
	← c00kiemon5ter left	22:36
	→ ipalaus_ joined	22:36
ESphynx	jast: But I expect most conflicts to be fixed as part of those rebase...	22:37
jast	well, I suppose you could use interactive rebase and reorder the commits to your liking	22:37
	that gives you maximum flexibility	22:37
	→ envi_ joined	22:38
ESphynx	yeah I know... pretty tricky to figure out though :P although I started using interactive rebases recently as well...	22:38
jmah	"bring in the stuff that happened on master, but make it as if it happened after whateve I have right now" means merge to me	22:38
ESphynx	but to me 'git merge' is a feature I'd do away completely with :P	22:38
	← timmywil left	22:38
	← ReekenX left	22:38
ESphynx	jmah: do this without any ugly 'Merge' commit	22:38
	→ nicksagona joined	22:38
jast	see, it's a question of aesthetics	22:38
	if you don't like the look of merge commits, you sign up for more hassle :)	22:39
ESphynx	I do not.	22:39
	they get super confusing when you try to look at the history of what happened there.	22:39
	→ jrm2k6 joined	22:39
ESphynx	especially with git log.	22:39
FauxFaux	git log --graph --oneline	22:40
	→ Arrowmaster joined	22:40
canton7	(add --decorate for more prettyness)	22:40
ESphynx	yeah... but most of the times, the code don't even touch	22:40
	so why the merges?	22:40
cbreak	git config --global alias.graph 'log --graph --decorate --oneline --all'	22:40
jast	well, it may not be a syntactical merge, but it's a semantic one	22:41
cbreak	merges have many advantages over rebases	22:41
	→ nardev joined	22:41
canton7	(or use gitk, if you're not on cli-only, and you prefer it)	22:41
cbreak	I often do it to bring a multi-commit feature branch into master	22:42
jast	a merge says: this joins two lines of development. I (git) have no way to determine whether they influence each other in any way, so that's all I'm gonna say about them.	22:42
cbreak	(I rebase onto master, and then --no-ff merge that branch)	22:42
jast	a more flexible model would require a different design and new ideas	22:42
	(let's not talk about the overkill darcs does in that area)	22:42
ESphynx	But rebase seems pretty good...	22:43
	--ff-only and rebases make me happy.	22:43
	I just wish git was friendlier to those who prefer this model ;)	22:43
jast	well, at any rate, this is the first time I've ever seen anyone do this two-part rebase thing	22:43
cbreak	you can do it if you want.	22:43
jast	and I've been active in this channel for several years	22:43
ESphynx	jast: that fact is interesting :) I thought people would do this.	22:43
cbreak	you can add --ff-only as default	22:43
	ESphynx: see man git-config merge.ff	22:44
jast	ESphynx: automatic message the 'git-config' manpage can be found at <http://git.kernel.org/?p=git/git-htmldocs.git;a=blob_plain;f=git-config.html;hb=master>	22:44
ESphynx	yeah I'm very happy with git...	22:44
jast	it's possible that you can prevent the extra rebase by structuring your workflow slightly differently	22:44
	I'm not up to thinking about it right now, though ;)	22:44
ESphynx	I just wanted to hear your opinion on this workflow: P	22:44
	→ rosco_y joined	22:44
cbreak	imho, rewriting public history is dumb	22:45
	your master is public	22:45
	so rebasing onto master and then something else is dumb.	22:45
jast	so after you've done those two rebases, what is it you want to do?	22:45
cbreak	-> logic 101 :)	22:45
ESphynx	cbreak: well no because that is on the sidebranch...	22:45
jast	push them to a remote copy of sidebranch?	22:45
	→ drizzd joined	22:45
cbreak	ESphynx: master is master	22:45
ESphynx	cbreak: the poitn of the 2nd master is to 'not' rewrite the public history of the sidebranch	22:45
	of the 2nd rebase*	22:45
cbreak	you would rewrite public history of master	22:46
jmah	I use rebases for trivial changes, or for rebase -i with a series of fixups etc. on a branch	22:46
ESphynx	cbreak: How so?	22:46
jast	thinking about it, you could do a squash merge of master into sidebranch	22:46
cbreak	you rebased onto master	22:46
ESphynx	cbreak: mater doesn't change... this is all on the sidebranch	22:46
cbreak	so you potentially got master commits	22:46
jast	if your sense of aesthetics allows for that ;)	22:46
cbreak	then you rebase those onto some other branch	22:46
	the master commits might not be on that branch, so they would get rewritten	22:46
	-> bad	22:46
	← envi_ left	22:46
ESphynx	i don't see the issue here. master always goes forward, the history of master does not change.	22:47
cbreak	it does.	22:47
	← booginga left	22:47
ESphynx	it sure does not!	22:47
	← baburdick left	22:47
cbreak	let's try it with an example	22:47
ESphynx	git log master always shows the same commit hash IDs	22:47
cbreak	let's say you have a new commit on master	22:47
	(Commit A)	22:48
ESphynx	ok	22:48
cbreak	you rebase onto master	22:48
	now your history contains A	22:48
ESphynx	history of sidebranch ?	22:48
	→ baburdick joined	22:48
cbreak	then you rebase onto some other branch	22:48
	that commit A now gets rewritten	22:48
ESphynx	right the origin's sidebranch	22:48
	it does on the sidebranch.	22:48
jast	yeah, so A has gotten rewritten on sidebranch... but not on master	22:48
cbreak	you now have a new commit that was A	22:48
	so masters history was rewritten	22:48
	-> commit duplication	22:49
jast	I don't follow	22:49
	← kevwil left	22:49
jast	yeah, it's duplicated, but that's just like a cherry-pick	22:49
cbreak	it's like implicit cherry picking	22:49
ESphynx	cbreak: yes... but the whole sidebranch is temporary and will be deleted at the end.	22:49
jast	I don't personally like cherry-picking all that much, but it's a viable thing to do in general	22:49
ESphynx	yes... I love cherry-picking.	22:49
cbreak	in general cherry picking is a stupid thing to do	22:49
jast	so what happens to the stuff on the branch when you delete it? I'm curious...	22:49
	← jtdowney left	22:49
cbreak	it provokes the same problems rebase has	22:49
	without even preserving basic ordering	22:50
ESphynx	jast: I rebase it onto master at the end.	22:50
	jast: and then merge it into master.	22:50
	and then delete sidebranch	22:50
cbreak	your history will be a tangled mess	22:50
ESphynx	cbreak: it will be perfectly clear. as if everything happened chronologically.	22:50
jast	hmm	22:50
cbreak	you'll have to retest every commit after every rebase if you want to take it exactly	22:50
	→ jtdowney joined	22:50
cbreak	because every rebase can introduce errors in the history	22:51
	because it creates NEW history that was never tested	22:51
jast	but you don't want to force-update sidebranch at any part during the progress, right?	22:51
	*process	22:51
ESphynx	jast: right I never force update.	22:51
cbreak	I only rebase private branches for good reasons:	22:51
	to keep the potential problems of rebasing to a minimum	22:51
	interactions between code can lead to invalid results without creating merge conflicts	22:52
ESphynx	cbreak: I consider these sidebranches as private branches with a backup.	22:52
cbreak	if they are private, then you don't need to rebase onto the remote	22:52
	you can just force push	22:52
ESphynx	they are not e.g. a development branch.	22:52
jast	if you're fine with doing a bunch of "rebase --skip"s during the final rebase (before you delete sidebranch), you can do "git merge --squash master" instead of the double rebase	22:52
cbreak	saving you the potential problems when rewriting master commits	22:52
ESphynx	true... but they are not totally private.	22:53
	other people can use them e.g. for trying it out.	22:53
jast	what's your take on that? I think it's probably going to be the least effort.	22:53
ESphynx	but read-only by them	22:53
cbreak	that doesn't matter	22:53
	→ joshsmith joined	22:53
cbreak	as long as they don't base work on them, you can rebase	22:53
jast	short of using more complicated commands instead of the double rebase	22:53
cbreak	and force push	22:53
jast	such as cherry-picks with an explicit range	22:53
ESphynx	cbreak: was afraid they'd have trouble on a pull?	22:53
cbreak	they won't pull	22:53
	if they just try out, they would just check out a new branch	22:54
	← joshsmith left	22:54
ESphynx	well yeah they do pull.	22:54
cbreak	and when they are done, throw it away	22:54
ESphynx	for quicker update.	22:54
cbreak	quicker update?	22:54
ESphynx	they might try it out a few times	22:54
	after I fix some thing :P	22:54
cbreak	they might not even make a branch	22:54
	just git checkout origin/foo	22:54
ESphynx	so they keep the brnach around	22:54
cbreak	and stay in detached head	22:54
ESphynx	they could.	22:54
	jast: what will that --squash do?	22:55
	→ traviscj joined	22:55
	← regedit left	22:55
ESphynx	basically make it one commit?	22:55
cbreak	squash will just make one single merge commit	22:55
	like a normal merge	22:55
jast	ESphynx: it creates a standard, non-merge commit that just happens to contain the changes a merge would otherwise introduce	22:55
ESphynx	bringing master up to date...	22:55
cbreak	but without the second parent	22:55
ESphynx	but as a single commit...	22:55
jast	yeah	22:55
ESphynx	as opposed to a bunch of cherry-picked commits (which I like)	22:55
	→ whyking joined	22:55
jast	well you can cherry-pick ranges	22:56
cbreak	it's basically like svn's merges :)	22:56
ESphynx	git rebase master && git rebase origin/sidebranch is easier :P	22:56
jast	so, roughly, instead of the double rebase you might do: git cherry-pick `git merge-base master sidebranch`..master	22:56
	← alex__c2022 left	22:56
	← tiagobutzke left	22:56
jast	this is very easy to abstract away in a script, unlike the double rebase	22:56
ESphynx	ah.	22:56
jast	(I haven't tested this, mind you)	22:57
ESphynx	jast: I guess that was the answer I was looking for :P	22:57
jast	okay, so my mind isn't completely asleep yet	22:57
	← mikewintermute left	22:58
jast	I still think there's a chance this will, at some point, create a hell of conflicts for you... but I can't make any kind of reliable prediction about it.	22:58
ESphynx	ithanks	22:59
cbreak	rebase --squash will lose your commits	22:59
ESphynx	jast: it seems to have worked well this far :P	23:00
jast	ESphynx: just giving you my perspective. the decision is, of course, all yours.	23:00
ESphynx	no easy way to `git merge-base` on cmd.exe eh? :)	23:00
	appreciated :P	23:01
jast	I don't think so	23:01
	but msysgit includes bash	23:01
canton7	one of the msysgit install options is whether you want to have git in your PATH	23:01
	I think that's what FauxFaux uses	23:01
ESphynx	yeah, but I usually stick to cmd.exe without bash :P	23:01
cbreak	what do you need merge-base for?	23:01
ESphynx	cbreak: cherry-pick ranges :P	23:02
	FauxFaux high-fives ESphynx.	23:02
jast	git itself uses bash for some of its commands, anyway	23:02
ESphynx	or just a git log `git merge-base`	23:02
	← metcalfc left	23:02
cbreak	ESphynx: you can do that with normal ranges maybe	23:02
	ESphynx high-fives FauxFaux back	23:02
	← Samoi left	23:02
jast	yeah, if you can figure out the right starting point for the range	23:02
cbreak	so master..foo is the same as (merge-base master foo)..foo	23:02
	→ Samoi joined	23:03
ESphynx	anyways, I got the code rebased like I wanted, so I will resume my work :P thanks folks.	23:03
	is it?	23:03
jast	cbreak: good point	23:03
ESphynx	master..foo is the same ?	23:03
cbreak	yes.	23:03
ESphynx	oh. interesting.	23:03
cbreak	master..foo means in foo but not master	23:03
jast	so, git cherry-pick sidebranch..master	23:03
cbreak	that would cherry-pick what is in master but not the sidebranch	23:04
ESphynx	So: git cherry-pick sidebranch..master ?	23:04
jast	it's a bit difficult to wrap your head around that since the syntax seems to suggest that it means something else	23:04
	← Amorphous left	23:04
ESphynx	cbreak: being in the sidebranch ?	23:04
cbreak	well, with git's history interpretation of "not in sidebranch"	23:04
	← kvanderw left	23:04
jast	doesn't matter where you are	23:04
cbreak	ESphynx: independent of where HEAD is	23:04
	but careful	23:04
jast	it all comes down to reachability	23:04
cbreak	if you repeat that cherry pick, you'll get duplicate commits :)	23:05
jast	basically sidebranch..master takes all commits reachable from master, and subtracts the one reachable from sidebranch	23:05
ESphynx	oh.	23:05
cbreak	"not in blah" is decided with history order	23:05
ESphynx	the rebase addresses that though? right?	23:05
cbreak	and not something like patch-id	23:05
	← traviscj left	23:05
jast	this leaves the one that master added after it split off from sidebranch	23:05
	in this case, anyway	23:05
ESphynx	i know I've done it a few times	23:05
	← MUILTFN left	23:05
ESphynx	multiple double-rebases	23:06
cbreak	well, cherry picking the same thing might just fail	23:06
	← ^slacker^ left	23:06
cbreak	if it can't apply the same thing multiple times	23:06
	or create empty commits	23:06
	or create no commits	23:06
	I've not actually tried that	23:06
ESphynx	ah.	23:06
cbreak	be expected to be surprised... or not. :)	23:06
	→ Spockz joined	23:06
jast	yeah, I didn't think about that	23:06
ESphynx	I'll stick to what has served me well then :P	23:06
jast	it might create a conflict situation for each time it re-applies commits	23:07
	→ kevwil joined	23:07
	→ workmad3 joined	23:08
cbreak	rebase has the same problem	23:10
	It will give you a weird conflict, and you have to rebase --skip to resolve it	23:10
jast	if the two commits have the same patch-id I think it would skip it	23:10
	← linusoleander left	23:10
cbreak	usually those irritate me since I search for an actual conflict first... :)	23:10
jast	in other cases... not so much :)	23:10
	→ neurodrone joined	23:11
cbreak	patch-id is a heuristic	23:11
jast	I know	23:11
	if the sidebranch commits touch a different part of the codebase, it'll work	23:11
	if they don't, all bets are off	23:11
cbreak	if it works it's nice. But I think it's prefering false positives over missed negatives	23:12
	← Cogito left	23:12
cbreak	a reasonable thing to do	23:12
jast	indeed. it's just a hash of the diff, actually, afaik	23:12
ESphynx	ah that's the patch-id?	23:13
	→ Cogito joined	23:14
	→ zarac joined	23:14
cbreak	see man git-patch-id	23:14
jast	automatic message the 'git-patch-id' manpage can be found at <http://git.kernel.org/?p=git/git-htmldocs.git;a=blob_plain;f=git-patch-id.html;hb=master>	23:14
zarac	Hey folks	23:14
jast		23:14
	whoops :)	23:14
	hey to you too	23:14
	→ mysphyt joined	23:15
zarac	I did a bad commit (included some files i did not want to). Is reset the right command to use if i just want to undo the last commit (leaving the working directory as is)?	23:15
	; )	23:15
	← glennpratt left	23:16
	← nicksagona left	23:16
jast	zarac: actually you can just fix things up and use git commit --amend to redo the bad commit	23:16
zarac	I included the build files. ; )	23:16
jast	so the fix will probably involve git rm --cached	23:16
zarac	ah, that was my next question ; )	23:16
	thanks a bunch	23:17
	→ envi_ joined	23:17
	← baburdick left	23:18
	→ baburdick joined	23:18
	← Spockz left	23:19
	→ madewokherd joined	23:19
	→ vals_ joined	23:20
	← tango_ left	23:20
	← kermit left	23:20
	← rzec left	23:20
	→ kermit joined	23:20
	→ Amorphous joined	23:20
	vals_ → tango_	23:21
	← axl_ left	23:21
	← juvenal left	23:21
	→ subbyyy joined	23:21
	← koo6 left	23:21
	→ sirdancealot joined	23:22
	← ixti left	23:24
	→ lucasvo03 joined	23:25
	← jmah left	23:25
	→ juvenal joined	23:25
	← lucasvo03 left	23:27
	← giallu left	23:27
	→ lucasvo joined	23:27
	← madewokherd left	23:28
	← lucasvo left	23:29
	→ krazyj joined	23:30
	→ lucasvo_2 joined	23:30
krazyj	is it possible to make a git pull also do force a pull of all submodules, too?	23:30
	← lucasvo_2 left	23:30
	→ lucasvo_3 joined	23:30
jast	well, sort of... but the interface is kind of screwed up in that the new versions in the submodules won't be checked out automatically	23:31
lucasvo_3	Are there any tips on importing huge svn repositories? I'm trying to convert an svn repository with over 12k commits and about a 100 tags and it's been taking several hours already with no end in sight.	23:31
jast	so it's "git merge --recurse-submodules; git submodule update --recursive"	23:32
	→ [ESphynx] joined	23:33
jast	lucasvo_3: not really. some folks are working on a tool to do imports faster, but I don't think it's ready for use yet. I'm not following its development very closely, though.	23:33
krazyj	jast: so if i were to make that into a pull hook, every time i 'git pull' remote master, it'll pull+merge the submodules from their repos, too?	23:33
	→ madewokherd joined	23:33
	← mysphyt left	23:34
	← tabakhase left	23:34
	← mattp_ left	23:34
jast	krazyj: actually, you might want/have to use "git pull; git submodule --merge --recursive" instead. I haven't personally done anything like that, so it's going to involve a bit of experimenting.	23:34
krazyj	ah ok	23:35
	→ tabakhase joined	23:35
	→ mattp_ joined	23:35
krazyj	jast: basically, i'm using a plugin library that i will make no edits to and just want to keep it up to date with what the creators have posted	23:35
	and make it easy for all of my coworkers to stay on a similar page	23:35
	← ESphynx left	23:35
	[ESphynx] → ESphynx	23:35
	→ tizzo-afk joined	23:36
	← sontek left	23:36
	← JP left	23:36
	← akraut left	23:36
	← etcetera left	23:36
	→ akraut joined	23:36
	→ JP joined	23:36
	→ etcetera joined	23:36
	← envi_ left	23:36
	← tabakhase left	23:36
	→ tabakhase joined	23:36
	→ sontek joined	23:37
jast	krazyj: right. in that case there's another way you can do it that is more likely to never mess anything up: "git pull; [fix potential conflicts here and commit if necessary]; cd submodule; git fetch; git checkout <the branch> (probably master); cd ..; git add submodule; git commit -m "Update submodule blah blah blah"	23:37
	not terribly pretty, I guess	23:37
	→ mklappstuhl joined	23:38
krazyj	jast: ahh, so basically i have to manage the submodule updates, i guess?	23:38
lucasvo_3	jast, do you know if github has their own method or are they using the same tools?	23:38
	yeah they seem to do.	23:39
jast	krazyj: yeah	23:39
krazyj	jast: thanks for the help dude!	23:39
jast	krazyj: you're welcome	23:40
	← Flogalen left	23:40
	→ ISF joined	23:41
	← Fandekasp left	23:42
	← fedesilva left	23:42
	← wbednarski left	23:43
	← dfr\|mac left	23:43
	← workmad3 left	23:44
	← airborn left	23:45
	← krazyj left	23:45
	← drizzd left	23:46
	→ macRover joined	23:46
	→ saschagehlich joined	23:47
	← ipalaus_ left	23:51
	→ alnewkirk joined	23:51
	← kevwil left	23:52
	← sirdancealot left	23:52
	← pantsman left	23:57
	→ sirdancealot joined	23:58
lucasvo_3	can I delete all modifications that are older than three months?	23:58
	← novettam left	23:58
jast	not without a lot of side-effects	23:59

Logs Search ←Prev date Next date→

Channels Documentation