IRCloggy #git 2011-01-01

Provider of IRC logs since 2005.

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

	2011-01-01
	← chrischambers left	00:03
	→ zombor joined	00:10
	→ chrischambers joined	00:10
	← {aaron} left	00:12
	Unknown[NF] → Unknown[OFF]	00:15
	← zombor left	00:16
	→ noahcampbell joined	00:16
	→ aidalgol joined	00:22
	→ timj joined	00:23
	← kylehayes left	00:23
tty1	charon, you around?	00:24
aidalgol	Where are the return codes for git documented?	00:25
cbreak	in the man pages for the respective commands	00:26
	usually there-s an --exit-code option	00:26
	like for git diff	00:27
patrikf	cbreak: uh, that's a rather special case...	00:27
	cbreak: what other commands have that option?	00:27
cbreak	which other case were you thinking of?	00:27
	← flo` left	00:28
patrikf	the common case is commands not having an --exit-code option	00:28
aidalgol	Do they all only return non-zero on an error, or do some return non-zero for some odd-but-not-bad cases.	00:28
	← DaveIngram left	00:28
cbreak	common case is that those that tell you something with exit codes have one :)	00:28
aidalgol	ah	00:29
	→ mbrumbelow joined	00:29
	← masonkramer left	00:29
cbreak	well, just read the man page	00:29
	→ masonkramer joined	00:29
	← ceej left	00:30
	← noahcampbell left	00:30
aidalgol	I was asking because I didn't see anything about error codes in the man pages.	00:31
	→ linusoleander joined	00:31
	← nakkor left	00:35
	→ nothingHappens joined	00:36
	→ nchaimov joined	00:36
	← mmw left	00:43
	← Buglouse left	00:50
	← justin-george left	00:54
	→ justin-george joined	00:54
	← Slix` left	00:58
	→ CarlFK joined	00:59
	← justin-george left	00:59
	← ch077179 left	01:01
CarlFK	how to I tag a revision? with "this is a stable version" so that someone can check out 'that' and not my current loopy version?	01:01
SethRobertson	Often people would have a stable branch and a working-branch	01:01
	However, you can certainly `git tag` to create a tag, but people would have to know about that tag name and ask for it	01:02
	You cannot normally reuse public tag names	01:02
	← anon33_ left	01:02
workmad3	I've always been a fan of the integration branch and master branch approach... developers do their work, integrate into the integration branch and the master branch is only updated from integration when its passed all tests and potentially some QA stuff depending on exact process	01:03
	← _inc left	01:04
CarlFK	workmad3: so master has the production code and bug fixes, new features get added to integration branch, which at some point gets merged with master?	01:05
workmad3	CarlFK: yah	01:05
SethRobertson	Some people do it that way. Also see gitflow for another workflow	01:05
CarlFK	cool. that makes the docs I am writing easier :)	01:05
workmad3	CarlFK: if integration only needs automated tests to pass to go into master, I'll actually get it set up so that my CI server will automatically do those merges ;)	01:05
SethRobertson	http://nvie.com/posts/a-successful-git-branching-model/	01:06
workmad3	I plan to get a setup soon with gitolite where I can even prevent merges to master from anyone other than the CI server	01:06
SethRobertson	That doesn't require gitolite	01:07
workmad3	SethRobertson: yeah... but I know how to do it with gitolite	01:08
	← linusoleander left	01:08
CarlFK	I bet you guys would like http://carlfk.blip.tv/file/4397842/ "Continuous Delivery by Jez Humble"	01:09
workmad3	CarlFK: it's on my coffee table downstairs ;)	01:09
SethRobertson	I don't believe in it myself	01:09
charon	tty1: sorta	01:09
	→ _inc joined	01:09
CarlFK	I think it is good to be able to do it, but then throttle it back	01:10
	→ Buglouse joined	01:16
	← tewecske left	01:18
	→ vitor-br joined	01:23
	← Dreamer3 left	01:26
Bird\|ghosted	workmad3, that actually sounds like the model I use at work. only, the version I deal with is recursive xD we have several levels of testing and integration going on	01:28
	← ExtraSpice left	01:29
	→ Martz^ joined	01:32
	← mbrumbelow left	01:45
	← saysjonathan left	01:47
	→ normanrichards joined	01:50
	← _inc left	01:55
	← unreal left	01:56
	→ unreal joined	01:57
	← slide left	02:00
	← TheFox21 left	02:02
	← RoboTeddy left	02:03
	← frakturfreak left	02:04
	→ noahcampbell joined	02:05
	→ AAA_awright_ joined	02:08
	← AAA_awright left	02:10
	AAA_awright_ → AAA_awright	02:11
	→ frakturfreak joined	02:17
	← sraue left	02:19
	→ Vampire0 joined	02:28
	→ justin-george joined	02:29
	→ sraue joined	02:29
	→ linusoleander joined	02:30
	← Vampire0_ left	02:31
	← blueyed left	02:37
	→ infid joined	02:43
	→ mbrumbelow joined	02:48
	← chrischambers left	02:49
	→ Angel-SL joined	02:50
	← d0k left	02:50
	← Angel-SL left	02:52
	→ harrisonk joined	02:56
	→ DaveIngram joined	02:57
harrisonk	hello	02:57
	← normanrichards left	02:58
harrisonk	I need help pushing a python file to gitorious	02:58
	I have the ssh keys imported	02:58
	← DaveIngram left	03:01
	→ simplechat joined	03:02
	→ t0rc joined	03:06
harrisonk	hello?	03:07
	← Buglouse left	03:08
Bird\|ghosted	harrisonk, is git push giving you an error message?	03:09
harrisonk	I don't know the command	03:10
	← kukks left	03:10
harrisonk	I tried the one that gitorious mentioned but it got me nowhere	03:10
	→ DaveIngram joined	03:11
	← Amorphous left	03:14
	← DaveIngram left	03:16
harrisonk	Bird\|ghosted: still there?	03:16
	← noahcampbell left	03:16
Bird\|ghosted	yeah	03:16
	→ Buglouse joined	03:16
Bird\|ghosted	you git commit'ed the file to your local repository, right?	03:16
harrisonk	Pardon?	03:17
	Sorry but I didn't understand	03:17
	→ rayners joined	03:18
	← Elfe left	03:18
	→ Elfe_ joined	03:18
harrisonk	Would it be "git push git://severdomainname.whatever/project/repo-name	03:19
	"	03:19
	?	03:19
	← Buglouse left	03:23
	← BiggFREE left	03:27
	→ Amorphous joined	03:28
	← davetoo left	03:29
Bird\|ghosted	harrisonk, you need to have stuff to push first :)	03:34
harrisonk	I have that	03:35
Bird\|ghosted	aka your HEAD must have changes that aren't in the repository you are pushing to	03:35
harrisonk	the repo is emty	03:35
	*empty	03:35
Bird\|ghosted	right, do you have a commit on your end?	03:35
harrisonk	and I have the file in a .git	03:35
Bird\|ghosted	ok. you have the file in your working copy	03:35
harrisonk	Hmm	03:36
	It's not there	03:36
	but....	03:36
Bird\|ghosted	o.O	03:36
harrisonk	I thought...	03:36
Bird\|ghosted	because you need to have it in your own repo before you can push it to someone else's	03:36
harrisonk	okay let's backup a few steps	03:36
Bird\|ghosted	hence, git add and git commit	03:36
	right. :)	03:37
	you have a git repository off somewhere else	03:37
harrisonk	yep	03:37
Bird\|ghosted	you have a Python file that you want to shove in that git repository	03:37
harrisonk	as in a site like gitorious?	03:37
Bird\|ghosted	do you have a working copy cloned from said git repository?	03:37
	(gitorious works. :)	03:37
harrisonk	I have the file I want to put in that repo	03:38
	I found the .git	03:40
	I got it so that I could push it to the server but I got this:	03:42
	No refs in common and none specified; doing nothing.	03:42
	Perhaps you should specify a branch such as 'master'.	03:42
	Everything up-to-date	03:42
	from running this: harrison@harison-desktop-Extream:~/build/climm$ git push http://git.gitorious.org/climm/climm-dev.git	03:43
SethRobertson	Try putting "master" after that command line. I am assuming here that the climm-dev repo is entirely empty (don't put master there without that)	03:47
harrisonk	yes the climm-dev is empty\	03:48
	*.	03:48
	got this: error: Cannot access URL http://git.gitorious.org/climm/climm-dev.git/, return code 22	03:49
	← offby1 left	03:49
harrisonk	I got it I think	03:50
	I did	03:50
	Thanks	03:51
	Goodnight	03:51
SethRobertson	If you set gitorious up as a remote and then create tracking relationships, you will not need to put all of that stuff after push	03:51
	→ Buglouse joined	03:51
SethRobertson	Happy New Year	03:51
	harrisonk → harrisonk_away	03:51
	→ Textmode joined	03:55
	← timj left	03:55
	← Buglouse left	03:59
	harrisonk_away → harrisonk	03:59
	← jdav_gone left	04:01
	→ priidu joined	04:02
	→ Angel-SL joined	04:05
	→ jdav_gone joined	04:06
	→ tmandry joined	04:13
	→ eletuchy joined	04:13
tmandry	is there any way to view the difference between the HEAD of two different branches?	04:13
AAA_awright	There's only one head I thought	04:14
	8HEAD	04:14
	bah	04:14
	tmandry: A branch is a pointer to a single commit... Why not `git diff branch1 branch2`	04:15
tmandry	yeah.. idk the correct terminology.. I'll try that though	04:15
AAA_awright	tmandry: Technically, a branch is a 41-byte file... Look at the output of `cat .git/refs/heads/master`	04:16
	← priidu left	04:18
tmandry	$ git diff master 4.7-stable	04:19
	fatal: ambiguous argument '4.7-stable': unknown revision or path not in the working tree.	04:19
	→ Buglouse joined	04:19
	← WildPikachu left	04:20
	← ka1ser left	04:20
	← drizzd_ left	04:27
	← Bird\|ghosted left	04:42
	← rsewill left	04:51
	← linusoleander left	04:53
	← unixtippse left	04:56
	→ unixtippse joined	04:58
	← orafu left	05:02
	→ orafu joined	05:02
	→ offby1 joined	05:08
	→ maleknet joined	05:17
	→ Slix` joined	05:21
	→ frogonwheels_ joined	05:25
	← unixtippse left	05:25
	← kogent left	05:27
	← yhager_ left	05:27
	→ batrick joined	05:32
	← dr_win left	05:32
	→ dr_win joined	05:37
	← batrick left	05:39
	← t0rc left	05:45
	→ davetoo joined	05:47
	→ Kobaz joined	05:50
Kobaz	git pull Already up-to-date.	05:51
	git push To prevent you from losing history, non-fast-forward updates were rejected	05:51
	why would it reject my updates if my local copy is up to date with the remote?	05:51
SethRobertson	Have you rewritten history?	05:52
	rebase/reset?	05:52
Kobaz	nope	05:52
	oh weird... i just put in the full url and now it works	05:53
	i didn't used to need to do that	05:53
	git pull path-to-repo 1.0	05:53
	* branch 1.0 -> FETCH_HEAD	05:53
	now regular git pull says it's up to date	05:53
	→ batrick joined	05:53
SethRobertson	Perhaps there was a tracking branch configuration problem?	05:54
Kobaz	i did a cat of .git/config before and after, and nothing changed	05:54
SethRobertson	It was it was up to date before too.	05:55
	More likely your push is attempting to push branches other than your current checked out branch and your pull only merged the current branch, or something like that	05:55
Kobaz	oh yeah	05:56
	it's saying failed to push to master, but i;m working on 1.0	05:56
	okay that makes sense	05:56
SethRobertson	You could experiment to see if that is true, or just be fooled by my hand-waving	05:57
Kobaz	it definitly is complaiuning about master	05:58
	complaining	05:58
	Unknown[OFF] → Unknown[NF]	05:59
SethRobertson	Is it still complaining?	06:01
	← ^Mike left	06:02
	← rayners left	06:13
	← tmandry left	06:19
tty1	charon, well when you have a second let me know	06:20
	→ normanrichards joined	06:25
	→ madsy joined	06:28
	→ amerine joined	06:29
	→ fengshaun_ joined	06:31
fengshaun_	how do I reset my whole working directory to a specific commit?	06:31
	I have the commit ID	06:31
	← mbrumbelow left	06:34
crab	either "git reset --hard $sha1" or (more likely) "git checkout -b tmp $sha1"	06:34
fengshaun_	what is sha1?	06:35
f00li5h	it is a hash.	06:35
crab	the first will change where your current branch points, the latter will create a new branch at that commit, and you can switch back to your other branch whenever you're done.	06:35
	f00: it is the commit id	06:35
	f00li5h paws at f00	06:35
fengshaun_	thanks, but I have done git reset --hard by mistake, and pushed it!	06:36
	now, I want my working directory (which now thinks I'm a couple weeks behind) to be at the "most recent" change (a couple of commits ago)	06:37
davetoo	which regex syntax does git-svn grok?	06:38
crab	f00lish: oh, sorry.	06:38
	← Buglouse left	06:38
crab	feng: what is the most recent change? if you have the commit id, you can git reset --hard to it	06:38
fengshaun_	this is the log: http://gitorious.org/bmarks/bmarks	06:39
crab	otherwise do something like git reset --hard origin/master, or use git log -g to find the older commit in your reflog	06:39
fengshaun_	I want to undo that last push!	06:39
davetoo	n/mind, found it, sorry	06:40
fengshaun_	it's kind of a weird situation	06:40
	I want everything to go back to "added the skeleton for edit function: delete" in the second last push	06:41
	← sraue left	06:41
f00li5h	pushes ... aren't.	06:42
	find the revision of the commit you want to go back to	06:42
fengshaun_	ok	06:42
	then what?	06:42
	crab: did you check the log?	06:43
	ok, I did git revert, it seemed to fix the problem	06:47
	thanks a lot crab and f00li5h	06:47
f00li5h	pay it forward!	06:48
	← eletuchy left	06:48
	→ eletuchy joined	06:48
	→ sraue joined	06:54
	← amerine left	06:56
	Unknown[NF] → Unknown[OFF]	06:56
	→ tmandry joined	07:02
	→ kogent joined	07:19
	→ hobodave joined	07:29
	→ BiggFREE joined	07:53
	→ red_ruin joined	08:02
	→ krosenvo- joined	08:04
	→ napster_123 joined	08:16
	← Rockj left	08:18
	→ Rockj joined	08:19
	← kogent left	08:23
	← red_ruin left	08:23
	← hyperair left	08:26
	→ hyperair joined	08:27
	→ kogent joined	08:34
	← simplechat left	08:34
	← napster_123 left	08:39
	← fengshaun_ left	08:43
	← kogent left	08:49
aidalgol	Is `git mv foo bar' equivalent to `cp foo bar; git add bar; git rm foo'?	08:50
	→ parkie joined	08:55
tmandry	aidalgol: I suspect that the first version retains the history of the file before the move, and the second one does not	08:55
	→ codeboy joined	08:59
	→ kpreid joined	09:00
	← parkie left	09:07
	→ tewecske joined	09:13
	Angel-SL → angelsl	09:13
	→ noahcampbell joined	09:33
albel727	aidalgol: tmandry is wrong. unlike some other version control systems, in git those commands are equivalent. see https://git.wiki.kernel.org/index.php/GitFaq#Why_does_git_not_.22track.22_renames.3F	09:35
AAA_awright	Everyone should be required to know that (awesome feature) about Git before attempting to use	09:39
	→ OOPMan joined	09:41
Arrowmaster	although sometimes you wish you could force it to recognize something as a rename when it doesnt	09:41
	→ ThomasLocke joined	09:43
	← Slix` left	09:47
tango_	Arrowmaster: with log you can use -C and -M to try harder	09:47
	← noahcampbell left	09:47
	← tewecske left	09:48
	→ tewecske joined	09:49
	← OOPMan left	09:51
	→ joel_falcou joined	09:53
	→ s0ber joined	09:53
	← MikhX left	09:56
	← _Vi left	10:00
	→ doener joined	10:01
	→ khmarbaise joined	10:06
	→ unixtippse joined	10:06
	→ napster_123 joined	10:06
	→ OOPMan joined	10:06
	→ tizbac joined	10:11
	← imMute left	10:11
	← workmad3 left	10:12
	→ workmad3 joined	10:13
	→ _Vi joined	10:13
	→ ExtraSpice joined	10:17
	← ExtraSpice left	10:24
	→ salty-horse joined	10:25
	← mcnellis left	10:27
	→ ExtraSpice joined	10:27
	→ oracal joined	10:28
ribasushi	is there a way to ask git to use a different `diff` ?	10:30
	I want to pre-process files (strip some stuff) before I run a diff	10:31
AAA_awright	ribasushi: Check out git difftool --help	10:31
ribasushi	ahhh difftool, not diff	10:31
	damned docs :)	10:31
AAA_awright	Diff is the Git-specific format, yeah	10:32
	If it helps any it's listed under SEE ALSO in the diff page	10:32
	→ _inc joined	10:32
AAA_awright	The git-diff(1) page, rather	10:32
salty-horse	is there a way to ignore whitespace changes when staging (i.e. git-add)?	10:33
	specifically, dos/unix newline changes	10:33
	(similar to how git-diff -w works)	10:34
	← tizbac left	10:35
	← _Vi left	10:35
	→ chrischambers joined	10:41
	→ mbrumbelow joined	10:43
	← tty1 left	10:45
	→ _Vi joined	10:47
	← mbrumbelow left	10:49
	→ zorzar joined	10:52
	← zorzar_ left	10:53
	→ silv3r_m00n joined	11:05
silv3r_m00n	hi there	11:05
	I want to remove 2 files from git management	11:05
	how ?	11:05
jast	silv3r_m00n: man git-rm @	11:06
	silv3r_m00n: the 'git-rm' manpage can be found at http://git.or.cz/man/git-rm [automatic message]	11:06
silv3r_m00n	oh no	11:07
	that thing delete the whole file	11:07
	I just wanted to keep it out of git	11:07
jast	yes, that's why you use the --cached option	11:07
	→ linusoleander joined	11:10
	← lamont left	11:15
	← joel_falcou left	11:20
	← napster_123 left	11:28
	→ leo2007 joined	11:29
	← doener left	11:31
	→ doener joined	11:31
	← angelsl left	11:34
	→ angelsl joined	11:35
	← silv3r_m00n left	11:35
	→ airborn joined	11:36
	← dcorbin left	11:37
	→ impulze joined	11:50
impulze	hm how do i ignore changes for certain tracked files	11:51
	(without untracking them, obviously) :P	11:51
	seems update-index --assume-unchanged requires root or something	11:51
	at least it fails with "Unable to mark..."	11:51
leo2007	seems working here.	11:55
doener	impulze: a quick glance at the code suggests that the path you gave to update-index is not in the index#	11:56
	s/#$//	11:56
impulze	hm	11:57
	impulze verifies	11:57
jast	also, assume-unchange is volatile. for example, git reset unsets it	11:58
	just so you know.	11:58
impulze	:(	11:58
	nothing that can be arranged for permanent recording?	11:58
jast	the correct way is to not stage changes that you want ignored	11:58
impulze	i don't want them to appear in git diff though	11:58
charon	impulze: is this for config files of IDEs and such? the correct fix is usually to only track a template, not actually the configs themselves	12:07
impulze	i don't use any IDE no	12:08
charon	or is it for config files that you have to change for deployment?	12:08
impulze	no	12:09
	say a repository serves autogenerated files, like Makefile.in or somesuch	12:09
	and whenever i change it locally, it sees it as change, although i would never want to commit it	12:09
	the obvious fix is to ask the repository owner to not track generated files	12:09
	but that's not going to happen :P	12:09
charon	well yeah	12:11
	buy a supply of LART sticks from the hardware store	12:12
impulze	:)	12:12
	→ Schmallon joined	12:14
	→ noktoborus_ joined	12:14
	← noktoborus left	12:15
	→ Dreamer3 joined	12:16
	← eletuchy left	12:16
charon	(FWIW i have the same problem in one project, but it's getting even better: if i generate the whole configure toolchain on osx it doesn't work. but if i use the linux-generated not-supposed-to-be-in-repo files unchanged, it works. sigh.)	12:17
jast	o/~ don't use autotools o/~	12:18
impulze	the problem persists for any other build tool	12:18
charon	well, the problem persists for any other people who don't know the difference between source and distribution-tarball	12:19
	charon shrugs	12:19
impulze	that aswell	12:19
	so no easy fix then i assume	12:19
	always have an unused-commit in my tree which removes said files, i guess	12:19
	and always send patches rebased on that commit	12:20
charon	i wonder if we could leverage duy's work eventually to implement sort of a "default blacklist"	12:20
	(the series on negative-pathspecs)	12:20
jast	I wonder whether we should build a feature to persist assume-unchanged flags	12:23
impulze	! :)	12:24
	→ LotR joined	12:25
	noktoborus_ → noktoborus	12:25
LotR	what is an 'unused-commit' ?	12:26
impulze	i mean something that's not going to get used in a format-patch series	12:26
	because the dev wants his generated stuff to be tracked	12:26
	← masonkramer left	12:29
jast	actually, I think it might be possible to use skip-worktree for stuff like this	12:30
	roughly: set core.sparseCheckout to true; list patterns in .git/info/sparse-checkout for files that should be ignored	12:31
	→ thrillerboy joined	12:35
jast	but I can't get it to update the skip-worktree bits	12:36
	oh, I did it wrong	12:37
thrillerboy	Hi, Guys!! If I want to host a bare repo in our web server, to which our collegues can push changes, How to do it?? Are there any guides or documentation for that. Thanks.	12:37
jast	it's the other way round. basically, put '*' into .git/info/sparse-checkout, then use !foo patterns to ignore certain files	12:37
	finally, git reset --hard. this destroys uncommitted changes and removes those files from the work tree, then always pretends that they are exactly the same version as in the index	12:38
charon	thrillerboy: look up gitolite	12:41
	thrillerboy: or if you're forced to use http, look up the smart-http docs	12:41
	e.g. http://progit.org/2010/03/04/smart-http.html	12:41
thrillerboy	thanks a ton charon will look into it :)	12:42
impulze	jast: hm what would i have to invoke?	12:42
jast	impulze: three things. enable config variable. edit file. git reset --hard (caution advised).	12:43
	→ Casmoo joined	12:43
	→ guardian joined	12:43
impulze	jason237: i mean for the skip-worktree thing	12:43
	jason237: oops	12:43
	jast: ^	12:43
jast	impulze: that's what that is about	12:44
impulze	ah ok	12:44
	→ henrikbjorn joined	12:49
	← maleknet left	12:50
	→ Cromulent joined	13:05
	← codeboy left	13:08
	→ joel_falcou joined	13:11
	← thrillerboy left	13:11
	→ PBeck joined	13:12
PBeck	hi	13:12
	is it possible to get a log out of the sha information => git log a884 <= then should display the log entry?	13:12
rudi_s	PBeck: If you specify the SHA1 of the commit, then that works fine.	13:14
PBeck	rudi_s: you mean the whole sha?	13:14
cbreak	if it's unique enough	13:14
	you can use just the start	13:14
rudi_s	PBeck: No.	13:14
	Yep, as cbreak said.	13:15
PBeck	rudi_s: cbreak ok it's right. Is it possible to only get the one log entry?	13:16
cbreak	sure	13:16
PBeck	how?	13:16
cbreak	it's the first flag mentioned in the man page :)	13:16
	but you can just as well use git show <hash> if you only want that commit. Note that the default output format is different	13:17
PBeck	cbreak: yeah show will be the right - my mistake :)	13:19
	cbreak: which flag you mean?	13:19
	oh -n1	13:19
	→ Thumper_ joined	13:20
PBeck	thank you cbreak rudi_s	13:21
cbreak	PBeck: it's -1 :)	13:24
	then is a placeholder for a number	13:24
PBeck	or so :)	13:25
	oki :)	13:25
	→ KiiK joined	13:25
KiiK	copy & paste same code and save. then git diff will show all is diff.... how to avoid it ?	13:26
cbreak	don't do it?	13:26
	git diff will show differences, regardles with what means or intention you created them	13:27
bremner	KiiK: you mean you make two copies where you had one before?	13:27
	→ karpar joined	13:27
KiiK	open a file, copy all code, then paste it.. ( so the file is same code, no changes)	13:28
cbreak	there are changes	13:28
	git diff will show which	13:28
	(you can use --color-words to get more detail)	13:28
	it might be whitespace like line endings	13:29
bremner	sounds like a line ending problem. Or your editor does not preserve exactly when copying and pasting (spaces and tabs are another possibility).	13:29
	← JEEB left	13:30
KiiK	ahh, yes. line endings problem, thanks !!	13:30
	← henrikbjorn left	13:38
	← Mage left	13:38
	← KiiK left	13:39
	← lemonchicken left	13:41
	← tewecske left	13:43
	→ tewecske joined	13:45
	← BiggFREE left	13:47
	← workmad3 left	13:47
	→ eletuchy joined	13:55
	← karpar left	13:57
	← airborn left	13:59
	← eletuchy left	13:59
	→ banisterfiend joined	14:01
banisterfiend	k	14:01
	do i have to add .gitignore to .gitignore?	14:01
	if i want to ignore .gitignore?	14:01
cbreak	no	14:01
Cro_Crx	lol	14:01
	you can't	14:01
cbreak	use .git/info/exclude instead	14:01
	sure you can ignore it	14:02
	but that's not the purpose of .gitignore	14:02
	← PBeck left	14:02
	→ dj_tjerk joined	14:03
	→ lemonchicken joined	14:05
	→ Bird\|ghosted joined	14:08
Tommy[D]	How can i get the last change to a file (or a specific change before that) with git (e.g. file a was changed some time ago and i just want to see that change or a previous change to that file)	14:09
cbreak	Tommy[D]: git log -p -- filenam	14:09
Tommy[D]	and a change before the last one?	14:11
teuf	cbreak's command will show all changes, try it	14:11
Tommy[D]	ah, missed the later lines containing previous diffs, thanks	14:15
	→ jmd joined	14:19
	→ henrikbjorn joined	14:24
	→ drl joined	14:27
	← justin-george left	14:40
salty-horse	does anyone have a script that sets all the GIT_*_DATE env variables when given a commit?	14:40
	→ justin-george joined	14:41
	← sraue left	14:41
	← zenom left	14:41
	→ darthdeus joined	14:42
	→ shine-neko joined	14:42
shine-neko	hey hey	14:42
	← hobodave left	14:42
	→ airborn joined	14:44
	→ RichiH joined	14:46
	← justin-george left	14:46
	→ bluegene joined	14:46
	→ _pingu joined	14:47
	← drl left	14:48
RichiH	i am pondering if i want to start managing my photos in git. we are talking a total size of a few dozen GB, expected to grow quicker rather than slower. pics are between 500 kiB and 30 MiB in size. i will most likely never change any single picture; i am more likely to simply create new versions when editing, etc. does it make sense to keep that in git or am i crazy? i am pondering git annex, as well	14:48
CarlFK	change any single picture = create new versions when editing,	14:48
	so.. what?	14:49
	I am guesing you mean "copy pic to new file, edit new file"	14:49
RichiH	i mean open pic, edit, save as new file	14:50
	→ Mage joined	14:50
	→ d0k joined	14:50
RichiH	i.e. git would always see exactly one version of any given file	14:50
_pingu	I've got 50 perlscripts in a dir. Now I want to makes versions for some of them. Each is a separate project. What is a nice way to organize that?	14:51
RichiH	_pingu: the really clean answer is 50 separate git repos. the really sloppy one is one flat directory for all of them. if you expect to offer any of the files to anyone else other than you at any time in the future, i would tend to split the repos. if it's your own stuff, do whatever is easiest	14:53
tango_	_pingu: zit	14:53
	_pingu: you can keep them in the same directory and track each one separately	14:53
	← malfy left	14:53
_pingu	RichiH: it's only for me. What do you think about a branch forach and add + commit only the script for this branch?	14:54
tango_	_pingu: http://git.oblomov.eu/zit (and yes it's a shameless plug since I'm the writer)	14:54
	→ sraue joined	14:54
	→ rseifert joined	14:54
	→ malfy joined	14:55
_pingu	tango_: intresting	14:55
	→ t0rc joined	14:56
shine-neko	I say a concern for a clone http://pastebin.com/6Bj5AuJE	14:58
tango_	_pingu: if the scripts are all in the same directory, you can spare a few blocks by pre-creating .zit (à la RCS)	14:59
shine-neko	Can anyone help me?	14:59
CarlFK	RichiH: sounds like you want rsync. the only thing git will offer is a place to log comments to commits, like "removed red eye" from the 8 pics that are part of the commit	15:00
	← malfy left	15:00
	← _pingu left	15:02
	→ Pupeno joined	15:03
CarlFK	there is git rm and mv, is there a git cp to dupe a file?	15:05
	so that you don't have to cp a b; git add b	15:05
jast	no, but you can write an alias that does just that :)	15:07
shine-neko	CarlFK, you talking to me?	15:08
	← leo2007 left	15:09
CarlFK	shine-neko: no. I was asking anyone. to see if there was something better than SaveAs	15:10
shine-neko	ha okay	15:10
CarlFK	jast: I was hoping there would be a history of the 'branch' ... so that later you would see that b came from a. this is all just pondering RichiH's workflow	15:12
shine-neko	you had no idea about my problem?	15:12
CarlFK	shine-neko: "aborting due to possible repository corruption on the remote side."	15:14
	fix that ;)	15:14
	I have no idea how. I don't have much git foo	15:14
shine-neko	:s	15:16
jast	there is never any explicit renaming/copying history. any such thing is done by way of auto-detection.	15:16
CarlFK	jast: what about git mv?	15:17
	→ malfy_ joined	15:17
jast	git mv is exactly equivalent to git rm + git add	15:18
	the other way around, though :)	15:18
CarlFK	huh. that makes me a little bit sad :)	15:19
	here i figured there was a few bits of history being saved in case I ever need it.	15:20
jast	remember: git tracks content, not filenames	15:20
	git log -M will almost certainly display the renames anyway	15:20
	← joel_falcou left	15:24
	→ napster_123 joined	15:24
	→ newneo joined	15:26
newneo	hi all	15:26
	→ notmyname joined	15:28
	← notmyname left	15:28
	← nothingHappens left	15:31
	← henrikbjorn left	15:31
newneo	can someone explain me how to work with submodule. I use a submodule in a project, then I have to modify files in this module	15:32
	but when I clone my git repo the submodule is still link on the official github of course	15:32
	so how can I include my changes knowing that I can't push on the official repo of the module ?	15:33
_ikke_	Yeah, indeed. When you are going to alter the submodule, you should link the submodule to a repository where you ahve access to	15:33
cbreak	make your own repo and point the submodule info to that	15:33
	be aware that who ever clones your parent repo also needs to be able to access the submodules	15:33
newneo	ok so I have to fork all submodules I use ?	15:34
_ikke_	newneo: Yeah, if you are planning to alter them	15:34
newneo	ok I didn't know that	15:34
cbreak	you can also just not change them	15:34
newneo	I use 3 submodules, 1 which I don't alter, but I have to change the 2 others	15:35
_ikke_	newneo: Then you can just leave that one alone, and change the remote urls of the other 2	15:36
cbreak	then you need to provide the new changed submodules to your cloners	15:36
	→ JEEB joined	15:38
	← Textmode left	15:38
newneo	thank you for your help	15:38
	← vitor-br left	15:40
	→ Zephyrus joined	15:41
	← frogonwheels_ left	15:43
	← gigamonkey left	15:44
	← thomasg left	15:44
	→ gigamonkey joined	15:44
	← ptor left	15:45
	→ vitor-br joined	15:45
newneo	now if I need to remove my submodule for including the file into my project, is it enough if I remove the submoduel line in .git/config and .gitmodules ?	15:48
cbreak	hmm?	15:50
	why do you need to remove the submodule?	15:50
	include which file?	15:50
newneo	I changed files of the submoduel so I can't keep it update with the official repo, so I want to just include files of this submodule into my main project	15:51
cbreak	I'd just fork the submodule	15:51
	→ AAA_awright_ joined	15:52
newneo	but I would have to do that before starting changing my files no ?	15:52
	it's a bit late now	15:52
cbreak	no	15:52
	you can fork it at any time	15:52
	just change the URL in your .gitmodules, run git submodule sync	15:53
	and make sure you push to the new location from your submodule	15:53
	→ dr4g joined	15:53
cbreak	obviously, the new location has to be accessible to everyone who has access to the master	15:53
newneo	ok so 1st I fork it on github, then I change the url and git submodule sync	15:54
dr4g	ls -l	15:54
	oops :)	15:54
newneo	ok I will try this way thank you	15:54
dr4g	hey guys, how can i get git to ignore a config file change	15:54
newneo	how a fork can also still be update with the official repo ?	15:54
dr4g	git commit -a ... shows me a list of the files and i don't want to commit in a specific file	15:54
	advice ?	15:54
cbreak	newneo: git rm --cached the file to remove it from the repository	15:54
newneo	dr4g: use .gitignore	15:54
	← AAA_awright left	15:54
dr4g	hmm thanks	15:54
cbreak	you should not commit config files in the first place, so it's better to remove them from the repo	15:55
	→ RoboTeddy joined	15:56
	→ petisnnake joined	15:58
	→ workmad3 joined	15:59
petisnnake	Hi, I'm very new to git. If I clone my repo from github my latest changed files (4 of them) aren't being cloned. I seem to have created a fork somehow. How can I merge my fork with the master branch on my server and overwrite anything on my master branch with the files in the fork?	16:01
cbreak	what do you mean?	16:02
	every git clone is a fork	16:02
newneo	petisnnake: look the url définition in your .git/config for origin, you should put your fork URL	16:03
cbreak	you have to fetch/pull to get new commits from other repositories, or push to send them to others	16:03
	don't edit .git/config	16:03
	unless you know what you are doing. which you seem to not to.	16:03
petisnnake	hmmm	16:03
	→ p3rg2z joined	16:05
	← primalmotion left	16:05
	→ timotei joined	16:06
timotei	hi	16:06
	does anyone know what a SVN repo must have so I can use git-svn ?	16:06
	→ albel727_2 joined	16:06
cbreak	it must exist?	16:07
timotei	I have a SVN with plain authentication , http://	16:07
	well, did:	16:07
	git svn init http://url	16:07
	and then	16:07
	git svn rebase	16:07
petisnnake	Well my main problem is that I made a commit / push of 4 files. If I clone my repo, those files aren't cloned for some reason (instead their old versions are). If I browse my repo on github the files aren't changed there either. But I can see the diff's if I click the string looking like f5db6db0682ac800b4f4 next to a 'commit' link	16:08
timotei	cbreak: and got:	16:08
	fatal: ambiguous argument 'HEAD': unknown revision or path not in the working tree.	16:08
	→ primalmotion joined	16:08
	← albel727 left	16:10
cbreak	petisnnake: do a git log	16:11
	do a git status	16:11
timotei	and:	16:11
	$ git svn fetch -rHEAD	16:11
	RA layer request failed: PROPFIND request failed on '/svn/ic2011_repo': PROPFIND of '/svn/ic2011_repo': 403 Forbidden (http://svn3.xp-dev.com) at D:\bin\Git/libexec/git-core/git-svn line 1773	16:11
	the repo has 18 revs now	16:11
cbreak	timotei: why don't you just git svn clone the repo?	16:12
timotei	hmm	16:13
	trying that. (I thought there is just git clone, but not git svn clone ^_^)	16:13
	and I get the same: RA layer request failed: PROPFIND request failed on '/svn/ic2011_repo': PROPFIND of '/svn/ic2011_repo': 403 Forbidden (http://svn3.xp-dev.com) at D:\bin\Git/libexec/git-core/git-svn line 1773	16:13
cbreak	it's in the man page	16:13
	well, your subversion obviously doesn't have a HEAD revision	16:14
	which isn't too surprising I guess	16:14
	because subversion uses numbers for revisions	16:14
timotei	well... yeah	16:15
	this is what the repo contains: http://pastebin.com/h00wj1zy	16:15
newneo	cbreak: is it possible to have a fork with a branch which is always update with the init repo ?	16:20
cbreak	what's an init repo?	16:20
timotei	cbreak: I've created a public repo: http://svn3.xp-dev.com/svn/gitsvn_test/ . Do you get the same errors trying to clone that?	16:20
newneo	I call init repo the repo which the fork is based on	16:21
cbreak	timotei: works	16:21
timotei	git svn clone?	16:22
cbreak	newneo: repositories are independent	16:22
	timotei: yes	16:22
timotei	the mysysGit must be the problem :(	16:22
	msys*	16:22
cbreak	windows is always a problem :)	16:22
	← jmd left	16:22
timotei	not really :D	16:22
newneo	cbreak: ok but we can't add a function (instead of a cron job of course) for pulling new commits on init repo to the fork	16:23
cbreak	pulling has to be interactive	16:23
	unattended merging is not generally possible	16:23
timotei	cbreak: I'd change on linux but the project is windows related.	16:23
cbreak	newneo: man githooks if you want to try to hack something together, but I bet it'll bit you in the ass	16:24
jast	newneo: the 'githooks' manpage can be found at http://git.or.cz/man/githooks [automatic message]	16:24
newneo	thank you	16:24
	← primalmotion left	16:29
newneo	is it a good practice to create an upstream branch in a submodule. The upstream branch will be link on the official repo, and the master branch will be link with my fork ?	16:33
borior	is there any way to edit a commit message for a hanging tagged commit?	16:33
newneo	borior: git commit --amend http://stackoverflow.com/questions/179123/how-do-i-edit-an-incorrect-commit-message-in-git	16:35
borior	newneo: no, that doesn't answer my question	16:38
timotei	cbreak: well, do you have any idea on what to do?	16:38
borior	i have a commit from some time in the past, which may or may not have a parent, for which I'd like to edit the commit message	16:38
cbreak	no idea.	16:39
borior	I'm not sure this is easily achievable -- or at least not with porcelain commands	16:39
cbreak	borior: every commit but the initial commit has a parent commit	16:39
	but this is not relevant	16:39
	what matters for you is if they have child commits	16:40
	git rebase -i can edit such commits, but that will be history rewriting	16:40
borior	cbreak: no, that's not true	16:40
cbreak	with all the associated problems	16:40
	that is true. go read the docs.	16:40
shine-neko	I say a concern for a clone http://pastebin.com/6Bj5AuJE	16:41
	you had no idea about my problem?	16:41
borior	cbreak: no, it's not true. I can create a commit with git commit-tree that doesn't have a parent"	16:41
cbreak	so? it's an initial commit.	16:41
borior	cbreak: not in any meaningful sense	16:41
cbreak	in the sense that it has no ancestors...	16:41
	so it's initial	16:41
borior	right, thanks for that	16:41
	can I edit the commit message for such a commit?	16:42
cbreak	read what I wrote	16:42
	history rewriting has consequences	16:42
borior	i know exactly what those consequences are	16:42
cbreak	then just use rebase -i and live with them :)	16:42
	use --root if you have to	16:43
	→ zenom joined	16:44
	← t0rc left	16:45
	→ icwiener joined	16:49
	→ justin-george joined	16:50
timotei	cbreak: hmm, using git in cygwin rather than msys git, I get a different error lol	16:53
cbreak	just as a comparison, I am on OS X	16:53
timotei	Initialized empty Git repository in /cygdrive/d/git/.git/ URL access forbidden for unknown reason: access to 'http://svn3.xp-dev.com/svn/asd' forbidden at /usr/lib/git-core/git-svn line 1920	16:53
	oh	16:53
	xD	16:53
cbreak	I used git svn clone http://svn3.xp-dev.com/svn/gitsvn_test/	16:54
	← Cro_Crx left	16:55
	→ marsilainen joined	17:10
marsilainen	hi	17:10
	← kpreid left	17:10
marsilainen	I'm adding a new file and commiting it	17:10
	← Cromulent left	17:10
marsilainen	now, this file is a shell script and it has executable permissions set	17:10
	→ kpreid joined	17:10
marsilainen	when I commit it, and then after I do a git status it tells me that I have modified the file	17:11
	← timotei left	17:11
marsilainen	but there is no diff	17:11
	I believe that it has committed the file with rw- permissions instead of rwx and that is the difference it can now see	17:11
	← angelsl left	17:11
marsilainen	this has happened at least twice now	17:11
	am I doing something wrong?	17:12
cbreak	the only permission git tracks is x	17:12
	just add it again and commit --amend the last commit	17:12
	(unless you pushed)	17:12
marsilainen	ok	17:12
	is there any reason this should be happening?	17:12
cbreak	because you changed the file after adding it?	17:12
marsilainen	I'm commiting through the egit extension in eclipse if that makes any difference	17:12
cbreak	(set it executable)	17:12
marsilainen	I didn't touch it	17:12
	it was executable before I did the commit	17:12
cbreak	doesn't matter	17:13
	what matters is when you add the file	17:13
	git commits the stuff you add	17:13
marsilainen	true	17:13
	ok	17:13
cbreak	not the stuff that's in your working directory	17:13
marsilainen	hmm	17:13
	I guess I must have made it +x after the add then	17:13
	ok, well thanks	17:13
	→ foocraft joined	17:15
shine-neko	I say a concern for a clone http://pastebin.com/6Bj5AuJE	17:15
	→ Revision17 joined	17:22
	→ Drakevr joined	17:23
	→ lamont joined	17:24
	← OOPMan left	17:25
	← tewecske left	17:33
	→ tewecske joined	17:35
	→ joel_falcou joined	17:36
	← AAA_awright_ left	17:37
	← napster_123 left	17:39
	→ Cromulent joined	17:41
	← dr4g left	17:43
	→ mmw joined	17:48
	→ ^Mike joined	17:49
	→ nvictor joined	17:49
nvictor	help	17:50
	i can't clone a repo folder	17:50
	→ henrikbjorn joined	17:50
nvictor	http://pastebin.com/ZqLwuATc	17:50
	→ munificent joined	17:50
munificent	simple noob question: i have a bunch of uncommitted changes. i've decided i don't want to commit them to master and want to create a branch to shunt them into. how does switching branches work with uncommitted changes?	17:51
	all the docs i read seem to assume everything is committed already. :(	17:51
patrikf	nvictor: can you clone via git:, ssh: or http: instead?	17:51
^Mike	munificent: stash them, create a branch, then unstash them	17:52
patrikf	munificent: checkout -b foo	17:52
	generally, if there are no conflicts, git lets you switch branches even with uncommitted changes	17:52
	if you want to create a branch off your current HEAD, that will always work	17:53
munificent	HEAD is what's already committed?	17:53
patrikf	HEAD is what you've currently checked out and are working on	17:53
	usually, this is a branch (e.g. master), or a commit (in case of a detached HEAD)	17:54
munificent	some of that is over my head, but i think it makes sense.	17:54
	if i do a checkout, will that leave uncommitted changes in my working dir?	17:54
cbreak	if you switch branches, yes	17:55
	if you do it in file mode, then it will overwrite stuff	17:55
patrikf	or use -f	17:55
cbreak	if you use -f, all betts are off anyway :)	17:55
patrikf	again, in general, git commands won't throw away your data, unless you use too much --force or it's the command's sole purpose	17:55
munificent	haha, yeah forcing doesn't seem like the smartest option for my current level of experience. :)	17:56
	that's what i figured, but it doesn't hurt to check first. thanks!	17:56
nvictor	patrikf: can you help with that ssh command?	17:58
	→ TuomasT joined	17:59
	← RoboTeddy left	18:00
TuomasT	Can I change the e-mail address put into my git histories?	18:00
mnemoc	TuomasT: yes	18:01
	→ fowlduck joined	18:02
TuomasT	mnemoc: and after I changed it with git config --global user.email "xx@yy", can I change it in earlier git commits?	18:02
	← guardian left	18:03
mnemoc	TuomasT: yes, see git-filter-branch	18:03
	TuomasT: you construct a new branch with new mangled commits	18:04
	→ RoboTeddy joined	18:05
	← nvictor left	18:06
	← RoboTeddy left	18:09
	→ noahcampbell joined	18:10
	← CarlFK left	18:11
	→ zombor joined	18:12
	← munificent left	18:16
	← Pupeno left	18:31
	→ joshsmith joined	18:32
	→ Pupeno joined	18:33
	→ guardian joined	18:33
	← Pupeno left	18:33
	← joshsmith left	18:33
	← normanrichards left	18:37
	← guardian left	18:38
	← workmad3 left	18:39
	→ Sigma joined	18:39
	→ guardian joined	18:42
	← henrikbjorn left	18:49
	→ avinashhm joined	18:49
	← avinashhm left	18:49
	← unixtippse left	19:00
	→ unixtippse joined	19:00
	→ MikhX joined	19:02
	← p3rg2z left	19:03
	← linusoleander left	19:07
	→ linusoleander joined	19:07
	→ hobodave joined	19:15
	→ aah joined	19:16
	← mmw left	19:16
	→ drizzd joined	19:21
	← khmarbaise left	19:22
	← airborn left	19:22
	← Cromulent left	19:22
	→ drizzd_ joined	19:23
	← drizzd left	19:26
	→ OOPMan joined	19:28
	→ khmarbaise joined	19:28
	← linusoleander left	19:29
	→ linusoleander joined	19:29
	→ rajgopal joined	19:30
	← OOPMan left	19:32
	← rajgopal left	19:33
	← doener left	19:33
salty-horse	I'm trying to use filter-branch to remove a specific commit by its sha1. should this work? http://pastebin.com/L96fLA69	19:36
infid	at work we only use svn and do a lot of work out of trunk and branches. i'm the only one considering using git with the svn repos. should i encounter many problems if i'm the only one using git-svn and can git-svn handle svn feature branches, merge tracking, etc?	19:37
	or is svn-git just an expirimental tool not suited yet for the real world?	19:38
	← MikhX left	19:39
salty-horse	infid, it's ok for the real world. I haven't tried doing merges with it, but it can understand svn merges properly and displays them as git merges	19:40
	there's also a bit of a problem setting svn properties	19:40
	→ Buglouse joined	19:40
	→ workmad3 joined	19:41
salty-horse	infid, that tool is useful if you'd like to work on topic branches locally, and then you commit to svn as a "rebase" on the trunk	19:41
	→ gilaniali joined	19:44
jast	salty-horse: looks fine on first glance. just try it... filter-branch keeps a backup anyway	19:44
davetoo	sigh	19:45
	I need to find a small but really branchy svn archive to experiment with; I"m not sure that ignore-paths is doing what I want/what I think it's doing	19:45
	← workmad3 left	19:47
	→ OOPMan joined	19:48
	← sivy left	19:50
salty-horse	jast, it doesn't seem to have any effect. let me be specific if I may: http://pastebin.com/Q8LxZZzX	19:50
	← gilaniali left	19:50
infid	salty-horse: ok	19:50
	← Zephyrus left	19:51
salty-horse	infid, keep an svn checkout around for doing svn-only stuff like setting properties. do actual commits from the svn-git clone	19:52
	jast, still with me? :)	19:54
infid	salty-horse: ok but with branches i cloned, are you saying to just svn dcommit to the branch, then do the merge on the svn side?	19:55
salty-horse	infid, yes, but I'm not aware of what git-svn can do. perhaps you should test it. git-svn may work well with doing merges :)	19:56
	the manual says not to do merges :)	19:57
	(but I may not understand it)	19:57
infid	it still probably safest to do it from the svn side	19:58
	but then i'm not yet seeing any reason to not just use svn all the time if i still have to keep an svn checkout around	19:58
	← newneo left	19:58
infid	i guess if i learn enough git to see what it can do in general that svn can't i'll know if it's worth using git-svn or not	19:59
salty-horse	infid, you can work on a "perfect" series of commits before committing them to svn all at once	19:59
	← TuomasT left	20:00
	→ Zephyrus joined	20:00
infid	hmm yeah	20:01
	thatd keep my svn repo cleaner i suppose	20:01
	→ jensn_ joined	20:04
	← Schmallon left	20:04
	→ Schmallon joined	20:05
	→ loincloth joined	20:05
salty-horse	infid, you can also use git to manage several tasks at once. with svn, you'd have to use different directories, and svn likes to take up a lot of space :)	20:06
	infid, if your network/svn server is slow, it's also much faster to dig into the history when you have a local git clone of the repository	20:06
	and don't forget you can start a gitweb server and discuss your change with others before making a real commit	20:07
	→ MikhX joined	20:07
infid	if i am using a git clone of an svn branch i still have to keep the git branch up to date with trunk. using git has the mental overhead as well as 'space' overhead	20:08
	← loincloth left	20:08
	← _Vi left	20:08
davetoo	git has too many ways to do it :)	20:09
	→ _Vi joined	20:09
infid	i've noticed at home i can handle a lot more mental overhead than i can at work. it's usually best to keep thigns as simple as possible when you have deadlines and people looking over your shoulder not having time to watch you fumble around tryin to remember how to get things to interoperate :[	20:09
	but it was worth investigating. i'll probably just not use git or svn unless it's the only thing i'm using per project	20:10
	← banisterfiend left	20:10
	→ rsewill joined	20:11
	← jensn_ left	20:12
	→ gilaniali joined	20:13
infid	my K.I.S.S-o-meter has served me well in the past, waving back and forth until landing on things like not using soap, most gui tools, network file systems, and other forms of abstractions that usually only serve to complicate things vs the simpler alternatives	20:14
salty-horse	infid, I'm using git to manage my changes before I push it into some weird revision control system at work. even when I only use it for personal management, I'm getting a lot of benefit out of it. is the mental overhead due to you being new to git? that can pass. once you get over it, there's no reason not to use it. if you don't yet feel ready to use it as a general purpose tool, practice more at home	20:14
infid	yeah i always practice at home first	20:14
	Unknown[OFF] → Unknown[NF]	20:15
infid	i just need to use git by itself for a while until i'm used to it. i'm new to it	20:15
salty-horse	once you get more fluent with git, you can figure out a workflow that's suitable for you	20:15
	I'd also suggest reading the Pro Git book. it helped me understand it better than any other guide	20:15
infid	thanks	20:17
salty-horse	np :)	20:20
infid	i love vcscommand.vim	20:21
	→ AAA_awright joined	20:21
infid	since if i do switch back and forth from my svn/git-svn checkouts it's seamless	20:21
	→ Cromulent joined	20:24
infid	i take it that git, like svn, doesn't do anything with file permissions/ownership? eg if i git clone or pull, it converts the files to be owned by whatever user i'm running the commands as?	20:26
	→ marioxcc joined	20:26
davetoo	That is something that the docs are confusing me about.	20:27
	It talks about "file modes" and "permissions" and I dont think they mean what i think they mean	20:27
	→ jensn_ joined	20:29
	← jensn_ left	20:30
	bremner engages mindreading mode	20:30
bremner	nope, I have no idea what you think, sorry.	20:30
	← ^Mike left	20:31
davetoo	"Git stores the mode of its content in the tree, but it doesn’t store any permissions data, which means it’s not good for backing up directories in which permissions are important."	20:31
	from the peepcode pdf; doesn't make sense to me.	20:31
	Maybe the author meant s/permissions/ownership/ ?	20:32
bremner	ah, sorry. don't know that document. It think it misses "executable" before mode.	20:32
	if that is something you paid for, you should complain.	20:32
davetoo	US$9.00	20:32
	:)	20:33
	Once I figure out how Git actually behaves, I'll have a leg to stand on, there.	20:33
	→ TuomasT joined	20:34
	→ sivy joined	20:35
	→ dermoth\|home joined	20:37
	← AAA_awright left	20:37
	← dr_win left	20:37
	← davetoo left	20:38
	→ loincloth joined	20:38
	← foocraft left	20:41
	← TuomasT left	20:42
	← joel_falcou left	20:47
	← salty-horse left	20:47
	→ disappearedng joined	20:48
disappearedng	Hey if I want to just pick out a few changes applied to certain files in another commit should I use git cherry pick	20:49
	← sivy left	20:50
	→ AAA_awright joined	20:51
	→ nixbox joined	20:57
nixbox	hi all	20:57
marioxcc	helo	20:57
	*ll	20:58
nixbox	i was looking up some online email archives and found a patch to a driver. How do I create a real patch out of it. The patch is not available as an attachment and it was submitted using the git format patch command	20:58
marioxcc	nixbox: what do you mean by real patch?	21:00
	git-format-patch makes real patches	21:00
	→ mmw joined	21:01
nixbox	marioxcc: i mean how can i apply this patch on a tree as right now it is in an email archive, do i have to make some specific file out of the email contents and then use a git command or some other command to apply it?	21:01
marioxcc	yes	21:02
	git apply for instance	21:03
nixbox	marioxcc: thanks	21:05
	← ptl left	21:06
	→ napster_123 joined	21:11
	→ sivy joined	21:13
	← linusoleander left	21:13
madsy	I don't quite understand what 'origin' is with relation to git push. Is it the default repo name when making a new repo?	21:18
	→ linusoleander joined	21:20
	→ weeifuh joined	21:20
	→ dimsuz joined	21:21
	→ normanrichards joined	21:22
Bird\|ghosted	madsy: origin is the name git uses for the the remote repository when pushing or pulling	21:23
	so you have things like origin/master	21:23
	← petisnnake left	21:23
Bird\|ghosted	which translates to "the master branch on the repository you're pulling from"	21:23
madsy	So origin is a branch name for remotes?	21:24
Bird\|ghosted	right :) remote branches will be origin/	21:25
madsy	So the branch names on the actual remote is not origin. It's just the name you use to refer to remotes?	21:26
	← sivy left	21:26
Bird\|ghosted	yeah	21:26
madsy	Because if I look at my github repo for instance, its branch name is master.	21:26
	→ amerine joined	21:26
madsy	But if I push a with checkout set to a new branch, I actually push that branch	21:27
	So it looks more like a repo name to me, not a branch name	21:27
	Which is the source of my confusion. I'm not exactly sure what a "repository name" is	21:28
	→ sivy joined	21:30
jast	origin/foo is a local mirror of 'foo' in 'origin'. you can see all of them using git branch -r.	21:31
	← Mage left	21:32
	← Buglouse left	21:32
	→ Mage joined	21:33
	← Zephyrus left	21:40
	← weeifuh left	21:41
	→ TheFox21 joined	21:43
	← sivy left	21:43
	← noahcampbell left	21:48
	→ workmad3 joined	21:55
	→ ptl joined	21:55
	← oracal left	22:00
	→ txomon joined	22:02
txomon	hello!, I have been using git for a long time, and now, I am trying to convince the investigation group I am in to move from svn to git. The main problem is that I don't know EXACTLY which are the main points that make git better	22:03
	← ptl left	22:04
	→ ptl joined	22:04
	→ oracal joined	22:04
	→ nakkor joined	22:04
	← nakkor left	22:05
dimsuz	hey guys. noob question. i just rebased my local branch against some remote and now i'm going to do some work and then tomorrow i'm going to rebase against another remote which already contains commits i rebased today. will it rebase ok, in the light of some new commits i'll introduce today locally?	22:06
	hope that which i just said is an understandable sentense :)	22:07
	← gilaniali left	22:07
rudi_s	txomon: http://whygitisbetterthanx.com/	22:07
madsy	txomon: Easier branching/merging to name one thing. With subversion, I was scared stiff when making branches.	22:08
	txomon: In git, it is a vital part of my workflow	22:08
	← workmad3 left	22:09
jast	txomon: find some advocacy at whygitisbetterthanx.com	22:09
	whoops, too late. sorry.	22:09
txomon	rudi_s: Thank you!	22:09
rudi_s	txomon: np	22:09
txomon	jast: Thank y	22:10
jast	that's what I get for reading every second line	22:10
txomon	madsy: I will read that	22:10
	jast: ^^	22:10
rudi_s	txomon: Another important feature is local work, I can create patches locally and then submit them upstream.	22:10
jast	yeah, that's one of the really big ones for me	22:11
albel727_2	dimsuz: yeah, it should be ok. git will detect duplicate commits, and skip them.	22:11
	albel727_2 → albel727	22:11
	→ EricInBNE joined	22:11
madsy	txomon: Also, since you can commit without access to the central repo (having such a thing is even optional), you can commit work without having internet access.	22:12
	→ robotarmy joined	22:12
txomon	im comparing with subversion	22:12
	cant I do that in subverion !?	22:12
madsy	No, "commiting" in subversion requires access to the central server.	22:13
	In git, "committing" and "pushing to an external repo" are two seperate things.	22:13
dimsuz	albel727: great. thank you!	22:13
	git really rocks :)	22:14
	→ ^jamie^ joined	22:14
madsy	txomon: Which means that in git, you can commit work to your local repo while you're on a plane trip for instance. And then push it to the external repo (if you have one) much later.	22:14
txomon	I always have worked with Git, that's why I cant make my mind to needing to have network...	22:15
	← oracal left	22:16
madsy	I agree that git's way is much more intuitive and usable.	22:16
	→ robinbowes joined	22:17
	→ jfkw joined	22:19
madsy	txomon: The reason why subversion differs, is that it isn't distributed. You don't have a local repository. "commiting" there is equivalent to reflect the changes on the server.	22:20
	At least if I have understood svn correctly :)	22:20
txomon	oki	22:20
	I thing that the web (http://whygitisbetterthanx.com) has all I need	22:21
	← ^jamie^ left	22:22
	← robotarmy left	22:28
	→ noahcampbell joined	22:30
txomon	¬¬ I hate subversion...	22:32
	← amerine left	22:32
txomon	what a hard ( & strange) thing to learn	22:32
	→ amerine joined	22:32
	→ sivy joined	22:34
madsy	Hm, both things I mentioned was key points on the whygitisbetter-site :)	22:35
divVerent	odd... "git gc" made a repo larger	22:35
	from 190 to 215 MB	22:35
	well, it was not --aggressive :P	22:35
	← linusoleander left	22:38
	← sraue left	22:40
	← TheFox21 left	22:41
	← levifig left	22:42
	← dimsuz left	22:42
	← Mage_Dude left	22:42
txomon	divVerent: ¿Have you thought about removing unnecesary files?	22:45
	(big files also included)	22:45
	for example I worked programming in VHDL for University and my repo passed from a 10 MB repo to a 50 MB... I made "git gc" and did nothing	22:47
	I had to delete large files that were automaticly generated by the ISE	22:47
wereHamster	the lesson to learn: never store generated files in the repo	22:48
	→ Mage_Dude joined	22:49
	→ masonkramer joined	22:51
	→ ketas joined	22:52
	→ sraue joined	22:54
	← hyperair left	22:54
txomon	yep	22:55
	I used this fantastic script	22:55
	→ jrmuizel joined	22:55
	→ hyperair joined	22:56
	→ robotarmy joined	22:56
txomon	(im searching it)	22:58
	→ jsoft joined	22:59
	← jaeckel left	23:00
jsoft	I have a repository gizmo locally. I want to put it on another box I have via ssh every now and then, basically making the remote copy the 'master' one.	23:00
	How do I do this?	23:01
	← hyperair left	23:01
wereHamster	git remote add <name> <other box> && git push <name> master	23:01
jsoft	hmm	23:02
	← guardian left	23:03
	← jrmuizel left	23:03
	← noahcampbell left	23:06
	← rseifert left	23:07
txomon	and here it is	23:09
	https://gist.github.com/762090	23:09
	→ workmad3 joined	23:10
	→ amerinese joined	23:10
	← normanrichards left	23:11
	→ hyperair joined	23:11
Arrowmaster	divVerent: if you are trying to shrink the size of your repo try git repack -adf	23:12
	← disappearedng left	23:15
	← ThomasLocke left	23:20
	← zombor left	23:25
	→ Determinist joined	23:29
	→ ZoeB joined	23:30
	→ temp01 joined	23:30
temp01	would it be fine if I remove the global [user] config from .gitconfig?	23:31
ZoeB	hi! I cloned my desktop's repository on my laptop, and committed a bunch of new changes on my laptop, in various branches. how can I transfer all the new commits across all branches back to the desktop again?	23:31
temp01	I just want each repo to have its own settings set by me (diff email addresses)	23:31
	don't want to accidentally commit with the global info in some repo	23:32
	→ serard joined	23:32
serard	hello	23:32
Arrowmaster	temp01: sure just remove it from .gitconfig and set it in each repos .git/config, if you forget to set it in a .git/config then you should get either an error or a warning when trying to commit	23:33
temp01	nice, thanks	23:33
	hmm, is it possible that it might screw up some tools that don't handle that case well?	23:34
serard	Currently I have checked out a branch. I have made modifications to files which I want to commit on another branch which doesnt exist yet. How can I create that remote branch and commit my files changes to that new branch ?	23:34
Arrowmaster	temp01: it could, for example i had to add advice.implicitIdentity=false to my /root/.gitconfig to hide the warning when using etckeeper	23:35
serard	I have created locally my new branch, but when I checkout this new branch, all my files are changed, so I cannot commit them to this new branch :/	23:35
temp01	Arrowmaster: hmm, I guess I'll just remove it for now and see if I run into any troubles	23:36
wereHamster	serard: git checkout -b <new brach> && git commit	23:36
serard	thanks :)	23:38
ZoeB	I got as far as "not git pull laptop" (laptop being the name of the remote repository in my config)	23:38
	← ketas left	23:40
madsy	serard: Changes has to be commited before you switch between branches. If you want to update either one, you most likely want to perform a merge.	23:42
wereHamster	madsy: wrong	23:42
madsy	Oh?	23:42
	Git complains here if my working directory has uncommited files when I do git checkout	23:42
wereHamster	it can do so in certain situatinos	23:42
	but if you want to create a new branch with the uncommitted changes, you can always do that	23:43
	← brosner left	23:43
khmarbaise	Hi committed two changes and pushed (to github9 the two changes separately..now i would like to combine the last two committs (as far as i know via git rebase HEAD~2) is that a good idea ?	23:43
madsy	Ah, sure. I didn't know what it was the case for serard	23:43
	I thought he already had a branch	23:43
wereHamster	'which I want to commit on another branch which doesnt exist yet' would be a pretty strong indication that the branch doesn't exist yet	23:44
madsy	Sorry, I read wrong :)	23:44
	← justin-george left	23:44
madsy	khmarbaise: Why would you do that?	23:45
	← fowlduck left	23:45
wereHamster	khmarbaise: that's a bad idea, because you already pushed those commits	23:45
khmarbaise	madsy: because i would like to comine the last two commits .....	23:45
	wereHamster: So no possibility ?	23:46
madsy	khmarbaise: Why do you want to combine them?	23:46
wereHamster	it is possible, but not recommendede	23:46
	← Cromulent left	23:46
khmarbaise	wereHamster: Because they have the same log message and do the same ...(enhanced doc)	23:46
wereHamster	khmarbaise: you basically say 'screw you' to anyone who fetched those commits	23:46
ZoeB	or am I missing the point and I'm not supposed to do that?	23:47
khmarbaise	wereHamster: Yes in this case i do...cause i commited a few minutes ago...	23:47
	← albel727 left	23:48
wereHamster	reset --soft HEAD^ && git commit --amend	23:48
	→ albel727 joined	23:48
wereHamster	then push -f. And don't forget to inform any person who fetched those commits	23:48
	good luck finding out who did that..	23:48
khmarbaise	wereHamster: First thanks for the help. Currently i have only a single watcher... ;-)	23:49
wereHamster	# of watchers != # of people who can potentially fetch form the repo	23:49
khmarbaise	wereHamster: I know...Of course there is a chance that someone has fetched already..but i don't believe (hope so..)...	23:51
	Thanks.	23:51
	← _inc left	23:52
	← mmw left	23:54
	→ _inc joined	23:54

Logs Search ←Prev date Next date→

Channels Documentation