IRCloggy #git 2010-12-10

Provider of IRC logs since 2005.

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

	2010-12-10
	→ normanrichards joined	00:00
	← normanrichards left	00:01
	→ constantc joined	00:01
	← KU0N left	00:01
engla	jaalto: ok, show me if you have answer/questions that I can vote up for you. Else I guess they encourage signal over noise there (don't take it personally!)	00:01
	→ jeremybub joined	00:01
	← petran left	00:02
	← acts_as left	00:02
	→ KU0N joined	00:03
	← chrischambers left	00:04
jaalto	engla: I have none as I just registered in the hop eto comment a thread where people seek for advice about what Emacs package manager would be ideal. As I'm just writing one based on Git and lessons learned in debian I was hoping to give some insight to the original question, and more specifically to #13 about http://www.emacswiki.org/emacs/DELPS	00:04
Eridius	jaalto: you just registered? I think you can't comment until you accrue enough rep	00:05
	yep, you need 50 rep	00:05
	http://stackoverflow.com/faq	00:05
jaalto	Eridius; Yep, engla told me. The web pages have become a difficult place for us old ones.	00:05
	← constantc left	00:05
	← shennyg left	00:06
jaalto	engla: Thanks for clarification.	00:06
	← Simkin left	00:07
	→ aresnick joined	00:07
	← nothingHappens left	00:07
engla	oh even 50 rep	00:07
	→ Simkin joined	00:08
	→ shennyg joined	00:08
	← _psychic_ left	00:09
	← spearce left	00:09
	→ spearce joined	00:09
engla	jaalto: very welcome	00:10
	→ bronson joined	00:10
	→ chrischambers joined	00:10
	→ saulus joined	00:12
	→ TheMoonMaster joined	00:13
	← boombatower left	00:14
	→ MabZilla joined	00:14
	→ ShooterMG joined	00:14
	→ acts_as joined	00:15
	→ boombatower joined	00:16
saulus	hi, I have some questions about branches: I have some local git repository that contains some document that evolves over time. At some times I have to give a talk at some position and therefore want to have a branch <talk_at_somewhere>.	00:16
	← ChrisPitzer left	00:16
saulus	Now I started at some branch of <talk_at_A> because I forgot to create the new branch <talk_at_B>. Further I lost (on purpose) the HEAD by one commit at some time and created a new branch from there.	00:17
	Now I have the questions:	00:17
	← bcardarella left	00:17
	← craptex left	00:17
saulus	Was it ok how I got rid of my one change? (By git checkout HEAD~1 and git branch newname)?	00:18
	← davesag left	00:18
Eridius	you could have used `git branch newname HEAD^`	00:19
	also, if you're trying to record your talk at given points in time, you should use tags instead	00:19
	branches are meant for things that will change. Tags are meant for immutable points	00:19
	← therain16 left	00:19
saulus	Further I want to know if I do something loose if I delete some intermediate branch (lets say <talk_at_B_temp> and now Im already working in the branch <talk_at_C>? Since git confuses my with the message: error: The branch ' <talk_at_B_temp> ' is not fully merged.	00:19
Eridius	so you can use a branch to work on the talk if you need to make changes for a specific event, and then tag it when you're done	00:20
	that's a warning, it's telling you that you're deleting a branch that references commits that are not reachable from any other ref	00:20
	e.g. it's letting you know that you're potentially losing commits	00:20
	→ nikolica joined	00:20
saulus	ah Eridius, never heard about tags. Sounds really like what I need.	00:20
Eridius	you have a few options here	00:20
	→ Wyverald joined	00:20
Cogito	tags are good	00:20
Eridius	you could do all your development on master, and then just drop a tag whenever you give a talk	00:20
	you could use other branches if you need to make temporary changes (e.g. change things for one talk that you will just want to revert after)	00:21
Cogito	but if you need to do a talk specific change, maybe branch is best...	00:21
	→ svnlto joined	00:21
ericlee	hi how do I delete a branch with its contents?	00:21
saulus	ericlee: branch -d branchname	00:21
ericlee	saulus, thanks	00:22
saulus	Eridius: ok. Sounds like I will switch to tags soon. But back to the warning: What may I loose. I wrote about my scenario. Isnt the information of the to-deleting branch <talk_at_B_temp> contained in the following branch <talk_at_C>?	00:23
	To be even more precise: I created the talk_at_b_temp on accident and do want to get rid of the branchname. But the contents of the branch shall be included in talk_at_C	00:24
Cogito	where did you branch c from?	00:24
saulus	Cogito: from the <talk_at_B_temp>	00:25
Cogito	and now you want to delete b_temp	00:25
	what will c be attached to then?	00:25
	do you have a master branch?	00:26
saulus	oh, well. maybe Im a bit confused and the git message is from a different scenario.	00:26
	I just switched to C and deleted b_temp without a warning.	00:26
	← boombatower left	00:27
saulus	Cogito: that is what I dont understand. So a branch is not self contained but is on top of other trees, right? I thougt about it but am not sure.	00:27
	← acts_as left	00:27
Cogito	that is more or less my understanding	00:28
	the branch details the differences from the trunk	00:28
	→ mhfs joined	00:28
Cogito	you want a master branch (ie the trunk) that has your working copy of the talk	00:28
	then you tag every time you do a sepcifiic talk	00:29
mhfs	hey guys ... is there something wrong with this command? git --git-dir='/Users/mhfs/testrepo/.git' --work-tree='/Users/mhfs/testrepo' pull	00:29
Cogito	if you have to do talk specific changes, you branch for that talk, and tag it too	00:29
mhfs	I keep getting fatal: /usr/local/Cellar/git/1.7.3.3/libexec/git-core/git-pull cannot be used without a working tree.	00:29
Cogito	but go back to changing the master branch for the next talk	00:29
saulus	Ok. So then, if I want to change my "branches" to "tags", is it right, that I have to "rebase" (or is it "merge") my branches (starting from A to B to C ...) to the master? Then the master contains all in one kind-of-a-branch and I can tag it?	00:30
	← svnlto left	00:30
	← LiamH left	00:30
saulus	ok Cogito. It sounds really like "tags" is what I want and what I use "branches" for now.	00:31
	→ madewokherd joined	00:33
ericlee	hi, if I use git-svn, how do I do a fresh checkout?	00:33
	→ shennyg_ joined	00:36
	← pantsman left	00:39
	← shennyg left	00:40
	shennyg_ → shennyg	00:40
	← noahcampbell left	00:40
	yaotti_aw → yaotti	00:41
	→ acts_as joined	00:41
	→ devians[w] joined	00:42
	← Buglouse left	00:42
	← jeremybub left	00:43
	← robotarmy left	00:43
	← sftp left	00:44
	→ acts_as_ joined	00:45
	← acts_as left	00:45
	acts_as_ → acts_as	00:45
	← brntbeer_ left	00:46
	→ shiba_yu36 joined	00:48
	→ acts_as_ joined	00:48
	→ normanrichards joined	00:50
	← rzl left	00:50
	→ lresende joined	00:51
	← acts_as left	00:51
	acts_as_ → acts_as	00:51
	→ Buglouse joined	00:54
	← pedrobel_ left	00:54
	← tjholowaychuk left	00:55
	→ pedrobelo joined	00:55
	Unknown[NF] → Unknown[OFF]	00:56
	← normanrichards left	00:56
	→ frakturfreak joined	00:57
	← ShooterMG left	00:59
	← amerinese left	00:59
	→ frakturfreak_ joined	01:00
	← frakturfreak left	01:01
	← sako left	01:01
	→ Dreamer3 joined	01:02
	frakturfreak_ → frakturfreak	01:02
	← justin-george left	01:02
	→ zombor joined	01:02
	→ frogonwheels joined	01:03
saulus	so after lots of easy rebases I got the last problematic one: I am on master and want to rebase myBranch to it. There are some CONFLICTS that I resolved. Now I am on (no branch) in between a rebase. I solved the conflicts. Now 'git status' shows nothing and 'git rebase --continue' tells me No changes - did you forget to use 'git add'? - whats the problem here?	01:03
	→ normanrichards joined	01:04
	← spearce left	01:04
	→ QinGW joined	01:04
	← workmad3 left	01:05
	← priidu left	01:05
	← frakturfreak left	01:06
	← alvaro_o left	01:06
	→ hasrb joined	01:07
	← amerine left	01:07
	→ fedesilva joined	01:07
	→ amerinese joined	01:08
	→ fedesilva_ joined	01:10
saulus	ok, I skipped it. Seems to be fine, although I'm a bit frithtened what happened. Thank you Cogito and Eridius for your help and inspiration!	01:11
	← gh0stz left	01:12
	← fedesilva left	01:12
	→ xiangfu joined	01:13
	← fedesilva_ left	01:15
	← normanrichards left	01:16
	→ anti-pattern joined	01:17
	→ itamarjp joined	01:19
	← TheMoonMaster left	01:21
	← fualo left	01:21
	→ fualo joined	01:24
	→ BiggFREE joined	01:25
	← beatak left	01:26
	← foocraft left	01:27
	← gigamonkey left	01:28
	← mikeric left	01:28
	→ artefon joined	01:29
	← mhfs left	01:30
	← masonkramer left	01:34
	→ masonkramer_ joined	01:34
	→ radost_ joined	01:34
	→ sako joined	01:34
	→ wliao joined	01:35
	← radost_ left	01:35
	← aah left	01:36
	← radost left	01:36
	← MikhX left	01:36
	→ nothingHappens joined	01:37
	← lresende left	01:39
	← mase_wk left	01:40
	→ robotarmy joined	01:43
	→ kanru joined	01:44
	→ mase_wk joined	01:45
	← nothingHappens left	01:46
	← pervy_sage left	01:46
Cogito	no worries saulus	01:46
	hope it works well for you :)	01:47
	→ kevlarman joined	01:47
	← kenichi left	01:48
	← ExtraSpice left	01:49
	→ kpreid joined	01:49
kevlarman	i'm looking for some advice on how to get git-svn working neatly with a repo that looks something like this: http://pastebin.com/uzqu37Qj	01:50
	← oriba left	01:50
	← Squarism left	01:51
	→ rchavik joined	01:53
	← cpg left	01:53
	← Rotham left	01:54
	→ boombatower joined	01:54
	→ MattDiPasquale joined	01:57
	← Amorphous left	01:58
	→ kylehayes joined	01:58
kylehayes	I'm getting "cannot open .git/FETCH_HEAD: Permission denied" when doing a pull	01:58
rohanpm	kylehayes: check permissions on that file?	01:59
kylehayes	on the remote one or the local one?	01:59
rohanpm	local	01:59
	that would happen if, e.g., somebody was root while they did a fetch or pull in your repo	01:59
kylehayes	ah hah!	02:00
	← friskd left	02:00
kylehayes	ok checking	02:00
rohanpm	if that was indeed what happened, then there may be files scattered all under .git now owned by root and you should chown -R the whole thing	02:01
	← v0n left	02:02
kylehayes	rohanpm: hmm, not owned by root, owned by me	02:02
rohanpm	then strace and see why it fails	02:03
	→ fowlduck joined	02:03
	← smolyn left	02:03
kylehayes	rohanpm: you know what, I just looked at the remote one and that file is owned by root....it shouldn't be	02:04
rohanpm	hmm, I didn't know that a `git pull' would access FETCH_HEAD on the remote? (unless you're explicitly asking for FETCH_HEAD which would be unusual)	02:05
	← abetaha left	02:05
	← shennyg left	02:06
	→ shennyg joined	02:06
kylehayes	rohanpm: that would be unusual, I'm not doing that. changing now...we'll see if it fixes it	02:06
	→ friskd joined	02:06
rohanpm	this happens here as well when people do sloppy work on the production server	02:07
	→ bcardarella joined	02:09
	← bcardarella left	02:09
kylehayes	ugh	02:09
	that's frustrating	02:09
	oh wait, fetch_head is not on the remote	02:10
	→ Squarism joined	02:10
kylehayes	but the one on my local is owned by me	02:10
	← hal left	02:10
kylehayes	so you were saying to strace	02:10
	← yhager_ left	02:10
kylehayes	which evidently is not on our boxes	02:10
	→ Amorphous joined	02:11
	→ brendon2020 joined	02:11
	→ Deciphered joined	02:12
	→ montylounge__ joined	02:12
	→ envi joined	02:12
	← xiangfu left	02:13
kylehayes	rohanpm: after running strace, the command worked inside of it	02:13
	lol	02:13
	montylounge__ → montylounge	02:14
	← Evious left	02:15
	← kylehayes left	02:16
	← nikolica left	02:18
	← saysjonathan left	02:18
	← bronson left	02:20
	→ MikhX joined	02:21
	→ gh0stz joined	02:22
	← gh0stz left	02:22
	→ gh0stz joined	02:22
	← Squarism left	02:23
	→ Vampire0 joined	02:24
	→ bronson joined	02:24
	→ abetaha joined	02:24
	← itamarjp left	02:24
	← jarrod_ left	02:24
	→ Squarism joined	02:26
	→ fromap joined	02:26
	← Vampire0_ left	02:26
	→ KB1JWQ joined	02:33
KB1JWQ	I've added my .bashrc to git. I change the file, git add .bashrc, and... nothing. It doesn't show any changes.	02:34
	How would I begin effectively troubleshooting this?	02:34
	← vitor-br left	02:36
	→ _psychic_ joined	02:36
	← hakunin_ left	02:39
	→ Angel-SL joined	02:39
uau	KB1JWQ: what did you expect to "show changes" after "git add .bashrc"?	02:40
KB1JWQ	uau: Both before and after I ran the "git add", "git status" showed nothing.	02:41
	Er, rather, it showed other changed files.	02:42
	But not .bashrc	02:42
	→ brntbeer_ joined	02:42
	← Deciphered left	02:43
uau	KB1JWQ: i think that shouldn't happen	02:44
	it should show the file as changed if there was no ignore rule matching it	02:44
	and if there was an ignore rule then "git add" should fail	02:44
KB1JWQ	uau: There is an ignore rule that matches everything, with exceptions declared. .bashrc is one of them.	02:45
	So I'm not running into ignore issues.	02:45
	git version 1.7.3.3	02:45
	← MikhX left	02:45
	← engla left	02:48
uau	KB1JWQ: does "git diff .bashrc" show anything after changing it?	02:48
KB1JWQ	uau: No.	02:49
	http://pastebin.com/55sttLdJ	02:49
	→ hakunin joined	02:50
	← Squarism left	02:50
uau	what about "git show HEAD:.bashrc"?	02:50
	← noktoborus_ left	02:51
	← andrewwatts left	02:51
	← MabZilla left	02:52
	← acts_as left	02:52
	← Amorphous left	02:53
KB1JWQ	uau: ...I'm an idiot.	02:54
	~/.bashrc is a symlink to my dropbox folder.	02:54
	← gh0stz left	02:55
	→ constantc joined	02:56
	← saulus left	02:56
KB1JWQ	Apparently git doesn't follow symlinks.	02:56
	← aresnick left	02:57
	→ MabZilla joined	02:59
	← SimonP86 left	02:59
	← mohi666 left	02:59
uau	KB1JWQ: IIRC git stores the symlink path instead	03:00
KB1JWQ	Right, so time to adjust that a bit. :-)	03:01
	← ketas-l left	03:04
	→ khem_ joined	03:05
khem_	how do I use git to obtain https://github.com/char0n/ffmpeg-php ?	03:05
	this one i mean https://github.com/char0n/ffmpeg-php.git	03:06
KB1JWQ	!Dropbox/.bashrc	03:06
	Er.	03:06
	git clone !Dropbox/.bashrc	03:06
	Er...	03:06
	→ Amorphous joined	03:06
KB1JWQ	If I can stop failing: git clone https://github.com/char0n/ffmpeg-php.git	03:06
	→ kylehayes joined	03:07
	→ smolyn joined	03:08
	→ aresnick joined	03:08
Cogito	any idea why I might be getting error "cannot merge now" when using egit?	03:08
	← khem_ left	03:08
	← kylehayes left	03:09
	→ Elfe joined	03:09
	← Fullmoon left	03:11
	← Elfe_ left	03:12
	→ MikhX joined	03:14
	→ yhager_ joined	03:14
KB1JWQ	uau: Thank you for the help.	03:15
	← artefon left	03:15
	← jaeckel left	03:15
	→ jaeckel joined	03:16
	→ TrevorBramble joined	03:16
	→ acts_as joined	03:16
	← rchavik left	03:18
	← DGodNT left	03:20
	→ intripoon_ joined	03:20
	← montylounge left	03:22
	→ ketas-l joined	03:22
	← intripoon left	03:23
	← pedrobelo left	03:26
	← Titosemi left	03:30
	→ nothingHappens joined	03:31
	← chrischambers left	03:31
	← sraue left	03:32
	→ noktoborus_ joined	03:32
	→ rchavik joined	03:33
	← gusnan left	03:34
	→ normanrichards joined	03:35
	← constantc left	03:36
	← kukks left	03:38
	→ Squarism joined	03:43
	→ M104 joined	03:43
	→ constantc joined	03:44
	→ rayners joined	03:44
	→ paladaxar joined	03:46
	← andyburke left	03:47
	→ constantc1 joined	03:48
	→ dgarr joined	03:49
	→ sraue joined	03:49
	← constantc left	03:50
	← constantc1 left	03:51
	→ gigamonkey joined	03:51
	→ frakturfreak joined	03:51
	← _psychic_ left	03:51
	← jsj0nes left	03:52
	→ timj_ joined	03:52
	→ eletuchy_ joined	03:52
	← eletuchy_ left	03:52
	→ _psychic_ joined	03:53
	← eletuchy left	03:53
	← frakturfreak left	03:53
	→ eletuchy joined	03:53
	← Cipher-0 left	03:55
	← timj left	03:55
	→ kingfishr joined	03:56
	← hasrb left	03:56
	→ constantc joined	03:57
	← airborn left	04:05
	← yhager_ left	04:08
	← Eridius left	04:10
	→ t0rc joined	04:10
	← abetaha left	04:12
	→ smolyn_ joined	04:12
	← smolyn left	04:12
	← nothingHappens left	04:12
	→ drl joined	04:13
	← sh1mmer left	04:13
	← boombatower left	04:13
	→ planet8 joined	04:18
planet8	What is the easiest way to see when a specific file was added to the repo?	04:19
Fangorn_	git log --oneline <filepath>	04:20
planet8	Fangorn, beautiful, thanks! :)	04:21
	← sivy left	04:22
	← Yuffster left	04:22
Fangorn_	planet8: if you have tig, you can use tig <filepath> also	04:24
	tig is a nice tool for exploring commit histories	04:24
	← Squarism left	04:24
crab	if you don't have tig, you'll have to do git log --reverse ;-)	04:25
Fangorn_	crab, i'm trying to get git log --oneline to also display the date...know how to do that?	04:25
	→ mdunham918_ joined	04:25
mdunham918_	how do i ungit a repository	04:25
	im tring to move it on my system and getting access denied	04:26
	→ amerine joined	04:26
crab	Fangorn: --oneline can't do it, but e.g. --format=%aD can	04:27
Fangorn_	ok	04:27
crab	Fangorn: see "PRETTY FORMATS" section in git-log(1)	04:27
	mdunham: i don't understand your problem or your question. what are you trying to do?	04:28
mdunham918_	i have a repo on my local disc	04:29
	and i want to repo it	04:29
	git has locked the folder entirely	04:29
	i was just playing with git to learn it	04:29
	← brendon2020 left	04:29
mdunham918_	and dont know it all the way now its annoying i cant move my stuff	04:29
	is there way to remove a repo	04:29
	← bronson left	04:29
crab	just delete it?	04:30
	← mmc left	04:30
	→ hasrb joined	04:30
	← shentino left	04:30
Fangorn_	planet8: git log --pretty=format:"%h %ai %s" <filepath>	04:31
mdunham918_	yes	04:32
	tried rm	04:32
	but it asks about 10000 access protected files	04:32
wereHamster	mdunham918_: rm -rf ?	04:32
planet8	Fangorn, U'r da man! Thanks! Very nice indeed!	04:33
Fangorn_	no problem, and like crab said, you can add --reverse to have it in ascending rather than descending order	04:33
	→ paladaxar2 joined	04:41
	← paladaxar left	04:41
	← MattDiPasquale left	04:44
	→ Yuffster joined	04:48
	→ dannas joined	04:49
	→ jrmuizel joined	04:52
	← jrmuizel left	04:52
	→ xiangfu joined	04:52
	← orafu left	04:53
	→ orafu joined	04:54
	← gebi left	04:54
	→ jony1 joined	04:54
	← t0rc left	04:57
	← d1b left	04:58
	→ shentino joined	04:59
	← shentino left	04:59
	→ shentino joined	04:59
	← orafu left	05:00
	→ orafu joined	05:01
	→ Deciphered joined	05:01
	← shentino left	05:02
	← BiggFREE left	05:04
	← aresnick left	05:05
	→ noahcampbell joined	05:06
	← orafu left	05:06
	→ orafu joined	05:07
	→ d1b joined	05:07
	← MikhX left	05:07
	← fowlduck left	05:07
	→ necromancer joined	05:08
	← M104 left	05:08
necromancer	hi, does anyone here run gitweb w/gitosis?	05:09
	← orafu left	05:10
necromancer	i'm trying to set it up on a subdomain on ubuntu	05:11
	on my ubuntu apache box that is :)	05:12
	→ orafu joined	05:12
	→ thiago_hel joined	05:13
	→ omer_ joined	05:14
	← planet8 left	05:17
	→ LaidBack_01 joined	05:17
	→ intripoon joined	05:19
LaidBack_01	hey, I'm using gitolite for a central repo. Currently I'm the only user, but it's likely that others will use this soon as well. That's good, i hope to get my group into version control at last.	05:19
	← dannas left	05:20
	← orafu left	05:20
	→ cytrinox_ joined	05:21
LaidBack_01	to that end, the boss is goign to want a nice web interface on it. I'd hazard a guess that more than a few of you are using gitolite... is there any webgui that you'd all recommend? Something that shows the diffs, commit messages, code, etc? I'm not looking for commit ability from the web side - that should all be handled by gitolite, so no webdav, but... monitoring would be the key.	05:21
	← cytrinox left	05:21
	cytrinox_ → cytrinox	05:21
	← intripoon_ left	05:22
	← dquestions left	05:22
	→ dquestions joined	05:22
	→ orafu joined	05:23
	← amerine left	05:23
	→ samlt joined	05:24
	→ aah joined	05:25
	→ M104 joined	05:27
sitaram	LaidBack_01: gitweb comes with git itself... and does all of that. cgit is another popular choice	05:28
	← necromancer left	05:30
	← noahcampbell left	05:31
	← stliu left	05:32
	→ Cromulent joined	05:32
	← zombor left	05:33
LaidBack_01	yeah, I just read your docs, and the rc file, saw the provisions for gitweb. just wasn't sure it was going to work till I read about it in your stuff there	05:33
	thanks for the fantastic mgmt tool, btw!	05:34
	← thiago_hel left	05:34
	← madewokherd left	05:35
	→ amerine joined	05:36
	← slide left	05:37
	→ srid joined	05:37
	← hobodave left	05:38
	→ zombor joined	05:38
	← hasrb left	05:39
	→ MetaCosm_ joined	05:41
	← M104 left	05:51
	→ slide joined	05:52
	← slide left	05:52
	→ slide joined	05:52
	← samlt left	05:52
	← intripoon left	05:54
	← Deciphered left	05:54
	→ intripoon joined	05:55
	→ Hunner joined	05:55
	← drl left	05:55
	← FabianB left	05:55
	← cylence left	05:55
	← alnewkirk left	05:55
	← towski left	05:55
	← TheUni left	05:55
	← tonyj left	05:55
	← jense left	05:55
	← dabbott left	05:55
	← aef left	05:55
	← dna left	05:55
	← dominikh left	05:55
	← jhelwig left	05:55
	← giallu left	05:55
	← xybre left	05:55
	← Tinned_Tuna left	05:55
	← cantoma left	05:55
	← KeeperOfTheSoul left	05:55
	← jas4711 left	05:55
	← reuss left	05:55
	← ianmcorvidae left	05:55
	← sfires left	05:55
	← divVerent left	05:55
	← cbreak left	05:55
	← candybar left	05:55
	← Cacheaway left	05:55
	← Juri_Agarin left	05:55
	← chitragupt left	05:55
	← avar left	05:55
	← sunn left	05:55
	← quadz\|2 left	05:57
	→ freimatz joined	05:57
	→ quadz\|2 joined	05:58
	→ petran joined	05:58
	→ roop joined	05:58
	← brntbeer_ left	05:58
	← petran left	05:59
	→ petran joined	05:59
	→ drl joined	05:59
	→ FabianB joined	05:59
	→ cylence joined	05:59
	→ alnewkirk joined	05:59
	→ towski joined	05:59
	→ TheUni joined	05:59
	→ tonyj joined	05:59
	→ jense joined	05:59
	→ dabbott joined	05:59
	→ aef joined	05:59
	→ xybre joined	05:59
	→ dna joined	05:59
	→ dominikh joined	05:59
	→ jhelwig joined	05:59
	→ Tinned_Tuna joined	05:59
	→ cantoma joined	05:59
	→ KeeperOfTheSoul joined	05:59
	→ jas4711 joined	05:59
	→ reuss joined	05:59
	→ ianmcorvidae joined	05:59
	→ sfires joined	05:59
	→ divVerent joined	05:59
	→ cbreak joined	05:59
	→ candybar joined	05:59
	→ Cacheaway joined	05:59
	→ Juri_Agarin joined	05:59
	→ sunn joined	05:59
	→ chitragupt joined	05:59
	→ avar joined	05:59
	← chitragupt left	06:00
	→ chitragupt joined	06:00
	← zombor left	06:02
	→ themarmot joined	06:03
	← kanru left	06:03
themarmot	can anyone tell me why several of my submodules(not maintained by myself) are suddenly not able to checkout a particular commit	06:03
	this seems to happen once every couple of months	06:04
	→ w0oter joined	06:04
	→ abetaha joined	06:04
themarmot	so i've gotta remove the offenders, the re-add them	06:04
mmattice	themarmot: some donkey has rebased them so they get a different hash?	06:04
	← LaidBack_01 left	06:05
themarmot	is the only way to fix that to do what I just explained above?	06:05
mmattice	the way to fix that would be with a 2x4 to their head	06:06
themarmot	lol	06:06
	i agree	06:06
	← rayners left	06:06
	← srid left	06:07
	← dgarr left	06:08
	→ lhz joined	06:09
	← smolyn_ left	06:09
	← OOPMan left	06:10
	← golodhrim left	06:10
	→ giallu joined	06:11
	→ thiago_hel joined	06:11
	← joel_falcou left	06:12
	→ kanru joined	06:13
	→ samlt joined	06:14
	← petran left	06:15
	→ leavittx joined	06:15
	→ hobodave joined	06:16
	→ stliu joined	06:16
	← stliu left	06:16
	→ stliu joined	06:16
	→ sambarino joined	06:20
	← Orbixx left	06:20
	→ Orbixx joined	06:20
	← abetaha left	06:21
	← Orbixx left	06:21
	← leavittx left	06:22
	→ MikhX joined	06:22
	→ boombatower joined	06:25
	→ phantomcircuit_ joined	06:25
	→ pervy_sage joined	06:26
	pervy_sage → svm_invictvs	06:27
	→ leavittx joined	06:28
	→ OOPMan joined	06:31
	← MetaCosm_ left	06:38
	← muep left	06:39
	ivey → ivey_off	06:40
	← _psychic_ left	06:43
	→ slonopotamus joined	06:45
	→ steph021 joined	06:49
	← jfkw left	06:52
	← eldragon left	06:54
	← MikhX left	06:55
	→ cstephens4 joined	06:56
	← samlt left	06:57
	← Textmode left	06:57
	← paladaxar2 left	06:57
	→ eldragon joined	06:57
	→ MikhX joined	07:01
	→ eagles0513875 joined	07:02
eagles0513875	how do i rebase	07:02
	→ devanjan joined	07:02
	← giallu left	07:03
	← cstephens4 left	07:06
	← Cromulent left	07:07
	← doener left	07:07
	→ doener joined	07:07
	→ wecing joined	07:08
wereHamster	with git rebase	07:08
eagles0513875	wereHamster: i am but it keeps chunking out the switches	07:09
	← slonopotamus left	07:09
	← wecing left	07:09
	→ skoop joined	07:10
	← skoop left	07:10
	→ skoop joined	07:10
	← Wyverald left	07:11
wereHamster	eh?	07:12
eagles0513875	wereHamster: i issue git rebase it wants some switches like -i or -f etc	07:15
	→ ludde joined	07:16
	← skoop left	07:17
wereHamster	yes, some commands tend to require certain options	07:17
	→ elmob joined	07:17
	→ pratz joined	07:17
	→ sh1mmer joined	07:20
eagles0513875	wereHamster: what options would i need	07:22
	im new to git and still learning the ins and outs of things	07:22
selckin	you still haven't said what you are trying todo	07:23
	← Buglouse left	07:23
eagles0513875	i need to rebase the LibreOffice code that im trying to run git pull on to update the code that i have	07:23
	← marc_kdab left	07:24
	→ Buglouse joined	07:24
wereHamster	eagles0513875: man git-rebase	07:24
jast	eagles0513875: the 'git-rebase' manpage can be found at http://git.or.cz/man/git-rebase [automatic message]	07:24
eagles0513875	ty	07:24
	← Mage left	07:27
	→ yhager_ joined	07:27
	→ Orbixx joined	07:29
	← devanjan left	07:30
	→ galderz joined	07:32
	← steph021 left	07:32
	← Vortex35 left	07:33
	→ nevyn^_ joined	07:33
	→ Rickardo1 joined	07:33
	← nevyn^_ left	07:34
	→ nevyn^_ joined	07:34
	→ devanjan joined	07:36
	← rchavik left	07:37
	→ Zhonghua joined	07:39
	→ Fullmoon joined	07:39
	→ Cromulent joined	07:40
	→ giallu joined	07:41
	→ saschpe joined	07:41
	← Fullmoon left	07:42
	← Modius left	07:43
	→ Modius joined	07:44
	→ Puck` joined	07:45
Puck`	hi everyone	07:45
	→ psankar joined	07:46
	← dreiss left	07:46
	→ Vortex35 joined	07:47
	← nuoHep left	07:47
	→ _ollie joined	07:49
	→ nuoHep joined	07:49
	← MikhX left	07:49
	→ pht joined	07:51
pht	hi	07:51
	← _ollie left	07:52
pht	is there a handy command that will tell me if current working directory ($PWD) is inside a git repo or not? (git status seems too noisy for me)	07:52
wereHamster	rev-parse --is-inside-work-tree	07:52
	pht: man git-rev-parse	07:53
jast	pht: the 'git-rev-parse' manpage can be found at http://git.or.cz/man/git-rev-parse [automatic message]	07:53
	→ skoop joined	07:53
	→ mmc joined	07:54
	→ vals_ joined	07:55
	← tango_ left	07:56
	→ inkedmn joined	07:57
pht	wereHamster, cool	07:57
	← pht left	07:57
	→ ph^ joined	07:57
	← ph^ left	07:58
	← eletuchy left	07:59
	→ ph^ joined	07:59
	→ dreiss joined	08:00
	→ lostern joined	08:00
inkedmn	dumb question about submodules: if I add a submodule to a project where I own/operate both the main project and the submodule (have write access to both repos), is it possible to make a change to a file in the submodule and commit that change back to the repo from whence it was checked out?	08:01
	→ teweWork joined	08:02
	→ justin-george joined	08:02
inkedmn	I realize I can do `git submodule update` to grab new changes that have been pushed to the remote - I'm just curious if I can push updates up to the repo from location where it's checked out as a submodule	08:04
	vals_ → tango_	08:06
	→ ntoll joined	08:06
	→ rchavik joined	08:06
	← yhager_ left	08:07
	← jahid_ left	08:07
	← rchavik left	08:07
	→ rchavik joined	08:08
	← vdv left	08:08
wereHamster	you can push the changes from the submodule itself	08:09
inkedmn	beauty	08:11
	thanks	08:11
	→ Schmallon joined	08:11
	→ Rai joined	08:11
	→ workmad3 joined	08:12
	→ opalka joined	08:15
	→ rchavik_ joined	08:16
	← tango_ left	08:16
	→ henrikbjorn joined	08:16
	← rchavik left	08:17
	→ MikhX joined	08:18
	→ Shoobdidoo joined	08:18
	← skoop left	08:20
	← dreiss left	08:20
	→ thierryp joined	08:20
	→ tango_ joined	08:21
	← hobodave left	08:21
	→ ejholmes joined	08:22
	→ priidu joined	08:22
	← Modius left	08:22
	← shamrock left	08:23
	→ guardian joined	08:25
	→ kirthi joined	08:27
Puck`	hi everyone, I'm pretty new to git, actually I'm just looking at it, I wish to use for webdevelopment and there's a question I have	08:27
	→ rane_ joined	08:27
	← kirthi left	08:28
KB1JWQ	Puck`: So ask it. :-)	08:28
Puck`	while I'm out there reading howtos and docs, I didn't find a clear answer on this: can git be used as a private version controller? I mean, can I restrict access to the git repository?	08:28
	← lostern left	08:28
KB1JWQ	Puck`: It can.	08:28
Puck`	that's swell, thank you KB1JWQ, I'll google on how to do that (:	08:29
KB1JWQ	Puck`: For instance, if you're using ssh as your transport, if you lack ssh credentials you can't commit to it.	08:29
	← Buglouse left	08:29
	→ resmo joined	08:29
	→ ixti joined	08:29
Puck`	great (:	08:31
	← psankar left	08:31
	← guardian left	08:32
	→ psoo joined	08:33
	→ itamarjp joined	08:33
yac	http://book.git-scm.com/4_rebasing.html	08:34
	^ are those diagrams drawed accordingly to some standard?	08:34
	rane_ → _rane	08:35
	→ linusoleander joined	08:36
	→ wipple joined	08:37
	← eagles0513875 left	08:38
	← fruight left	08:39
	→ _ollie joined	08:40
	← _ollie left	08:40
	← itamarjp left	08:41
	→ pantsman joined	08:41
drl	newbie question. What does this line mean: @@ -149,12 +128,11 @@ root index freq strong morph greek indo eng arti defin)	08:42
	→ joel_falcou joined	08:42
	→ airborn joined	08:42
	→ fruight joined	08:43
	← roop left	08:43
	← rchavik_ left	08:44
	→ vdv joined	08:44
	→ rchavik joined	08:45
	→ romankrv joined	08:46
	→ codetroll joined	08:48
	← themarmot left	08:50
	← opalka left	08:50
	← codespectator left	08:52
	→ dpino joined	08:52
	→ rseifert joined	08:53
	← rseifert left	08:54
	→ cilly joined	08:55
	← fruight left	08:57
	→ ayanes joined	08:58
	← ayanes left	08:58
	→ ayanes joined	08:58
	← workmad3 left	08:58
codetroll	good morning good people.. At work we still use cvs and each time we install a javaproject the entire project is tagged as a version in cvs, making it easy to keep track of traceability and always be able to check exactly what has been installed in production etc. Is there a similar feature in git where I can tag the entire project tree as a version or similar???	08:59
	← Yuffster left	08:59
tango_	codetroll: in git you can only tag entire project trees	08:59
	→ paddy__ joined	09:00
codetroll	ok - how to do that?? (and how to find the tagged version again ;) )	09:00
	→ sebbi joined	09:00
tango_	man git tag	09:00
jast	the 'git' manpage can be found at http://git.or.cz/man/git [automatic message]	09:00
tango_	;-)	09:00
	ah lol	09:00
	man git-tag	09:00
jast	the 'git-tag' manpage can be found at http://git.or.cz/man/git-tag [automatic message]	09:00
tango_	jast should really be fixed for the dashy thing	09:00
codetroll	thank you :)	09:01
tango_	codetroll: and the git checkout if you want to check out a particular tag	09:01
	← carutsu_ left	09:01
	← dquestions left	09:01
tango_	beware that checking out a tag detatches your head and you should not do commit on top of that	09:01
	→ fruight joined	09:02
	→ dquestions joined	09:02
crashanddie	tango_, so you need to branch?	09:02
codetroll	cool..	09:02
tango_	crashanddie: depends on what you want to do	09:03
codetroll	and no I wouldn't commit on top of a tagged version. It would merely be to be able to get hold of what was installed/delivered etc at a specific point in time	09:03
	← Schmallon left	09:04
	→ jony2 joined	09:04
crashanddie	tango_, well, say I've got a tag named v2.4.1 which represents the commit that finalised version 2.4.1 of fooapp, current version is 2.6.x, I want to add a bugfix to that specific version and release it?	09:04
	→ davux joined	09:04
davux	hi :)	09:04
	← shiba_yu36 left	09:04
crashanddie	tango_, git checkout -b bugfix-v2.4.1 v2.4.1 ?	09:04
tango_	crashanddie: then you need to branch from v2.4.1 yes	09:04
davux	how do you reflect an "svn mv branch newname" with git svn ?	09:05
tango_	you could name it v2.4.1.x or v2.4.x depending on your scheme	09:05
davux	i lookes everywhere and I can't find it	09:05
	← jony1 left	09:05
davux	looked*	09:05
crashanddie	tango_, aye, but my process requires QA to validate before giving a new version number ;)	09:05
	but yeah, that would become 2.4.1.1	09:06
	just wanted to check I got it right, cuz I've imported 4 years worth of revisions and tagged them based on numbers versions, would've been bummed if it weren't possible to use those tags :P	09:06
tango_	so you call the branch v2.4.1.x or whatever and then v2.4.1.1 would be tagged when QA is happy with it	09:06
crashanddie	yeah, fair enough.	09:07
	→ opalka joined	09:07
tango_	you can also have a branch called like a tag, and then you can use tags/name vs heads/name to differentiate, but that's prone to generate some confusion	09:07
crashanddie	yeah, I usually avoid naming the branches the same as tags.	09:07
	→ Wyverald joined	09:11
	← dl9pf left	09:12
	→ pom joined	09:12
	→ dl9pf joined	09:12
	← dl9pf left	09:12
	→ dl9pf joined	09:12
	← linusoleander left	09:14
	→ nevyn^__ joined	09:15
	← asmega left	09:15
	→ olliey joined	09:16
Nickeeh	With these guidelines: http://tbaggery.com/2008/04/19/a-note-about-git-commit-messages.html anything else I may have to look for when writing commit messages?	09:17
	← nevyn^__ left	09:17
	→ nevyn^__ joined	09:17
	← nevyn^_ left	09:19
yac	i am looking for a tool which i can use to resolve conflicts when i have a list of (hundreds of) files which i know i can resolve by using "theirs" version	09:19
Nickeeh	yac: git checkout <file> --theirs ?	09:19
	← rchavik left	09:20
	← henrikbjorn left	09:21
yac	Nickeeh: im not sure if i can use it from the manpage. My git is after merge so the files in my working tree are staged for commit but in conflicted state	09:22
	romankrv → rrr	09:23
	← resmo left	09:23
	rrr → romankrv	09:23
	→ asmega joined	09:23
	→ roop joined	09:24
	← pantsman left	09:24
yac	Nickeeh: yep, thats it. thanks	09:26
	→ guardian joined	09:26
	← dquestions left	09:27
	→ dquestions joined	09:28
	→ codespectator joined	09:28
	← kpreid left	09:28
	→ lanthan_afh joined	09:29
	→ primalmotion joined	09:33
primalmotion	hi there	09:33
	I've got a small question. I've merged a branch and their was conflicts. I solved them but I screw up	09:33
	is there any git magic in order to cancel last merge ? (other than reset because I don't really remember the origin point of merging)	09:34
	In order to re-resolve my conflicts ?	09:34
	apanek → apanek-work	09:34
yac	hope you have not screwed up too much, i am fixing my screw up for ~10 hours now	09:35
primalmotion	yac: that's a big screw up :)	09:35
yac	maybe a git-revert may help you	09:35
	yep, it is	09:36
	→ spaceonline joined	09:36
primalmotion	yac: why don;t you revert and then retry ?	09:36
	→ resmo joined	09:36
primalmotion	that's what I do when I screw up ( and I often screw up ...) and that's why I ask my question about git magic to accelerate this process)	09:37
crashanddie	http://www.kernel.org/pub/software/scm/git/docs/user-manual.html#reverting-a-commit	09:37
davux	how do you know what branch a given commit belongs to?	09:38
	or what branches	09:38
primalmotion	crashanddie: yes I know that, I just wondering if there was a way to automatically revert to the last merge without searching the log	09:38
davux	gitk can know that but i don't have access to it	09:38
	→ kpreid joined	09:39
	← Wyverald left	09:40
	→ flavius joined	09:41
	→ albel727_2 joined	09:41
	← romankrv left	09:41
flavius	How can I review the changes that've been made upon "commit -a", i.e. diff'ing all the files about to be committed?	09:42
	... and show the diffs in the editor	09:42
	← albel727 left	09:43
davux	flavius: git diff --cached	09:44
	→ Anti-X joined	09:46
flavius	nope, not what I want (and it doesn't show anything either). I think I came across some parameters that made this possible: upon "commit -a" to also see in the editor all the diffs, so I can write an appropiate message	09:46
	→ eletuchy joined	09:46
	→ varuntv joined	09:47
flavius	Unfortunately I cannot recall the parameter, and I couldn't find anything in the manual	09:47
crashanddie	flavius, when you commit -a, you're already commiting	09:47
	← varuntv left	09:47
crashanddie	there's no review process involved	09:47
yac	primalmotion: its complicated i have several UML diagrams describing what happend and how to solve it	09:47
crashanddie	if git diff --cached doesn't return anything, it means there are no changes to the head	09:47
davux	flavius: add -p to your git commit command	09:48
primalmotion	yac: then good luck :)	09:48
flavius	crashanddie: git status shows lots of changes. I need the diff for every file shown there	09:48
yac	primalmotion: thanks, i think i am like half-way there	09:48
davux	flavius: how does git status mention the changed files?	09:48
flavius	davux: "unknown switch 'p'"	09:49
davux	are they about to me committed, or just modified?	09:49
flavius	davux: as "modified"	09:49
davux	so git diff	09:49
	← Gitzilla left	09:49
flavius	yeah, but how to do it for all of them at once?	09:49
crashanddie	it does?	09:49
davux	git diff	09:49
	← amerine left	09:50
davux	either git diff or git diff --cached will show them, depending on whether they're already staged for committing	09:50
	← albel727_2 left	09:50
flavius	davux: thanks, that did it. yet I am pretty sure there was something which made git put exactly that diff for everything in the editor when I was asked for a commit message	09:51
	← slide left	09:51
	→ gebi joined	09:51
	← kanru left	09:51
davux	flavius: that's -v, sorry, not -p	09:51
Puck`	so I created the git repository, I a reply back Initialized empty Git repository in /home/puck/1r.hu/.git/ . Next step would be to clone this on my machine, but I guess 1st I need to make it available. Do I need to configure apache for this, or git already know where this is?	09:52
davux	-p is for git log iirc	09:52
flavius	davux: yes, THAT was it	09:52
Puck`	*knows	09:52
davux	anybody know how to see the branches a given commit sits on?	09:53
	→ tatsuya joined	09:53
davux	i renamed a branch in svn using svn mv	09:54
	and i'm following the svn repo using git svn	09:54
	but i can't replicate the branch renaming into git	09:54
	← drl left	09:54
davux	i did git svn fetch though, and so i do have the commit's hash	09:55
	but i can't find it in any branch	09:55
	→ raorn joined	09:55
raorn	hi! i need help with merge	09:55
	http://www.kernel.org/pub/software/scm/git/docs/howto/using-merge-subtree.html - this describes how to merge one subtree into another	09:56
	← urkud left	09:56
raorn	i need to import single file foo.txt from a remote repository into bar/baz/foo.txt in my repo	09:56
	with ability to do further merges from "upstream"	09:57
	→ sivy joined	09:59
	→ ExtraSpice joined	10:00
	→ noyb joined	10:01
	← tatsuya left	10:03
	→ anathematic joined	10:04
	→ bluenovember joined	10:04
	→ Rai_ joined	10:04
	← Rai left	10:07
	← QinGW left	10:07
	→ Gitzilla joined	10:07
	→ drl joined	10:10
	→ engla joined	10:10
	← engla left	10:10
	→ engla joined	10:10
	→ dk39350 joined	10:14
	← drl left	10:14
	→ EgS joined	10:15
	→ workmad3 joined	10:16
	← boombatower left	10:18
	← dk39350 left	10:19
	→ rolfb joined	10:20
	→ paddy___ joined	10:20
	→ pratz_ joined	10:20
	← pratz left	10:21
	← paddy__ left	10:21
	← teweWork left	10:22
	→ teweWork joined	10:24
	→ tatsuya joined	10:25
	→ lundtor joined	10:25
	← teweWork left	10:25
	→ drl joined	10:26
	→ teweWork joined	10:27
	→ nevyn^_ joined	10:29
	← teweWork left	10:29
	→ Schmallon joined	10:29
	→ JoeCoolDesk joined	10:30
JoeCoolDesk	Hello friends.	10:30
	→ teweWork joined	10:30
	ChanServ set mode: +v	10:30
ludde	I've made a commit, but I realized that I want to split it up into two separate commits, how would i fix that (with rebase i assume)?	10:30
	← dquestions left	10:31
	← nevyn^__ left	10:32
	← teweWork left	10:32
	→ dquestions joined	10:32
	→ ereslibre joined	10:33
	← ereslibre left	10:33
	→ ereslibre joined	10:33
	→ teweWork joined	10:33
	← nicoulaj left	10:34
rudi_s	ludde: Not pushed yet? git reset HEAD~; git add -p; git commit; git add -p; git commit	10:37
	← mtx_init left	10:39
	← ejholmes left	10:39
	← MikhX left	10:39
ludde	rudi_s: I've made loads of commits after that commit i want to split	10:40
	I found git rebase -i and then edit	10:40
	it seems to do the trick	10:40
rudi_s	ludde: Oh, then rebase.	10:40
	→ curtana joined	10:40
rudi_s	But if you've pushed these commits, you'll break the history of everbody else.	10:40
	→ nicoulaj joined	10:41
ludde	it's my own private branch	10:41
rudi_s	ludde: Then it's fine.	10:42
	→ tvw joined	10:42
	← eletuchy left	10:42
	← priidu left	10:43
	← wliao left	10:43
	→ marc_kdab joined	10:44
	→ awilkins joined	10:46
	← omer_ left	10:46
tlonim	yeah..git rebase involves tampering with past, kind of like time travel :P	10:46
	→ amberovsky joined	10:47
engla	time travel is when you reorder commits	10:47
amberovsky	.	10:47
ludde	when i use squash with rebase, it opens up an editor with the empty file $@ (or maybe it was @$) several times	10:47
	← amberovsky left	10:47
ludde	what am i supposed to enter there?	10:47
	$@ is the name	10:48
	it's in the root of my project	10:49
	it's probably a git bug?	10:49
tlonim	engla: yeah..or delete or merge .. i meant any sort of tampering	10:49
	← workmad3 left	10:51
	Unknown[OFF] → Unknown[NF]	10:51
	→ Zephyrus joined	10:52
tlonim	ludde: how did you run it ? something like git rebase -i master~10 ?	10:52
	→ muep joined	10:53
	← anathematic left	10:53
ludde	git rebase -i <name_of_the_branch_i_branched_from>	10:54
	← muep left	10:54
	← elmob left	10:55
	→ priidu joined	10:55
	← JoeCoolDesk left	10:56
	→ henrikbjorn joined	10:56
	← gretch left	10:58
	→ muep joined	10:58
	← ketas-l left	10:59
	→ ketas-l joined	11:02
	← dr_win left	11:03
	→ dr_win joined	11:03
	→ d0k joined	11:04
	← rolfb left	11:05
	← nevyn^_ left	11:06
	→ rolfb joined	11:08
	→ Jippi_mac joined	11:09
	→ ramok joined	11:10
	← stliu left	11:11
	Jippi_mac → Jippi	11:12
	← ExtraSpice left	11:12
	← Cromulent left	11:13
	← FauxFaux left	11:14
	→ artefon joined	11:15
	→ zorzar joined	11:16
	→ dytrivedi joined	11:18
	← zorzar_ left	11:18
Cogito	If I have branched master, developed and merged updates from a remote master, then want to create a commit that has just my changes, i want to use rebase -i, correct?	11:19
	→ trupheenix joined	11:20
cbreak_work	that's one way	11:20
Cogito	I'm not sure how to only have my changes	11:20
cbreak_work	make a temporary working branch for that	11:20
Cogito	ok, how else could I do it?	11:20
cbreak_work	that way you won't destroy anything	11:20
	git filter-branch	11:20
	you'd write a filter program	11:20
Cogito	not at all sure of how to do that/what you mean	11:21
	I branch master than merge the updates to it?	11:21
	I need all my changes to be in one commit	11:21
	though I use many commits whilst developing	11:22
	dwmw2_gone → dwmw2	11:22
cbreak_work	oh, just use rebase... it's easier	11:23
	but be aware that rebase changes the current branch	11:24
	so if you want to keep it, branch off it	11:24
	← drl left	11:24
	← thiago_hel left	11:25
trupheenix	hi	11:25
	i am trying to track a remote branch	11:25
	→ ExtraSpice joined	11:25
	→ gusnan joined	11:25
trupheenix	but it is not listed in git branch -r	11:25
	however the branch has been created on the server	11:26
	so how do i track the remote branch?	11:26
	← dytrivedi left	11:26
cbreak_work	trupheenix: git fetch remote	11:26
	does it appear now?	11:26
	← Anti-X left	11:27
	← Zephyrus left	11:28
	→ Anti-X joined	11:31
	→ albel727 joined	11:32
mapreduce	git show 8174 seems to show me a particular file, in full, with no diff markers etc.	11:32
	git log 8174 shows nothing.	11:32
	← ExtraSpice left	11:32
doener	mapreduce: so 8174 is a blob	11:32
mapreduce	It turns out I was looking for a non-existent branch named 8174 (I tend to name branches after ticket numbers).	11:33
	Ok, I see.	11:33
	Is there a way of specifying that I want a branch (I realise that if there's a branch with the same name the branch will be shown)?	11:34
	→ salat joined	11:34
salat	hey..how do i push all my local branches to a remote repository?	11:35
	→ ExtraSpice joined	11:35
	← ramok left	11:36
	→ kai joined	11:37
kai	hi	11:37
	→ Ryback_ joined	11:37
	← Ryback_ left	11:37
	→ Ryback_ joined	11:37
kai	the current msysgit net installer seems to be missing libiconv2.dll, any idea where to get that from?	11:38
	← raorn left	11:39
	← SRabbelier left	11:43
	→ yotaff joined	11:44
	→ JoeCoolDesk joined	11:44
	→ vitor-br joined	11:45
	→ JohnBat26 joined	11:46
JoeCoolDesk	I have a problem with my repository being repeated in my git client	11:48
	← devanjan left	11:48
offby1	repeated?	11:50
JoeCoolDesk	Yes, repeated.	11:50
	rev1 rev2 rev3 rev1 rev2 rev3 rev4 rev5	11:51
wereHamster	JoeCoolDesk: you rebased your history, correct?	11:51
	offby1 admires wereHamster for (apparently) understanding JoeCoolDesk's problem	11:53
	wereHamster is just guessing	11:53
JoeCoolDesk	wereHamster, yes.	11:54
	wereHamster guessed right	11:54
	← henrikbjorn left	11:54
JoeCoolDesk	It's correct in git log though.	11:54
wereHamster	if it's correct in git log, where is it 'incorrect' then?	11:55
	→ leo2007 joined	11:56
leo2007	What's the quickest way to show the content of a commit?	11:56
wereHamster	leo2007: what do you mean by content?	11:57
JoeCoolDesk	wereHamster, in my git client, and pretty much every other client.	11:57
leo2007	the log and diff etc. possible also notes	11:57
	possibly	11:57
ptor	git show <sha1>, presumably	11:57
wereHamster	JoeCoolDesk: maybe the git clients are displaying two different brnaches which have the (almost) same comits	11:57
	leo2007: git show <sha>	11:57
leo2007	getting "fatal: bad object cde5721ffe534d0cb6c13f8fea9e91245d9b33c7"	11:58
JoeCoolDesk	wereHamster, I doubt that.	11:58
leo2007	I have tried exactly that.	11:58
wereHamster	leo2007: where did you get that hash from?	11:58
leo2007	from the log	11:58
wereHamster	leo2007: git cat-file -t cde5721ffe534d0cb6c13f8fea9e91245d9b33c7	11:58
leo2007	wereHamster: my bad. I was doing it in another git repo.	12:00
	it works now.	12:00
	sorry for the noise.	12:00
	→ kirthi joined	12:01
	← ereslibre left	12:02
	← paddy___ left	12:02
	→ paddy__ joined	12:03
	→ Cromulent joined	12:04
	← normanrichards left	12:05
	← dvaske left	12:07
	← ixti left	12:08
	→ tizbac joined	12:09
	→ Yuuhi joined	12:09
JoeCoolDesk	wereHamster, Giggle definitely shows repeation.	12:09
wereHamster	JoeCoolDesk: screen shot?	12:09
	← Anti-X left	12:10
JoeCoolDesk	Hold please.	12:11
	← davux left	12:14
	→ tjholowaychuk joined	12:16
	→ urkud joined	12:19
	→ yngress joined	12:21
	→ dvaske joined	12:23
	← airborn left	12:23
	→ henrikbjorn joined	12:26
	→ Anti-X joined	12:26
	← salat left	12:29
	→ mishok13 joined	12:30
	→ petercoulton joined	12:30
	← Anti-X left	12:33
	→ nevyn^_ joined	12:35
	→ skoop joined	12:39
	→ SianaGearz joined	12:41
	→ frakturfreak joined	12:43
	→ nevyn^__ joined	12:44
JoeCoolDesk	Giggle is by far the best GUI I've ever used for Git.	12:45
	wereHamster, for some reason when I pull and look at the local version with giggle, it seems fine, but the remote version is repeated.	12:45
	← nevyn^__ left	12:46
	→ nevyn^__ joined	12:46
	← nevyn^_ left	12:46
offby1	by far the best, except for when it gives the wrong answer	12:46
	other than that, though	12:47
Bombe	:)	12:47
	→ frakturfreak_ joined	12:49
	← frakturfreak_ left	12:49
	← frakturfreak left	12:49
	← JoeCoolDesk left	12:51
	yaotti → yaotti_aw	12:51
wereHamster	where's the screenshot?	12:52
	→ workmad3 joined	12:52
	← constantc left	12:53
offby1	http://live.gnome.org/giggle	12:54
wereHamster	I mean of his history..	12:54
offby1	shrug	12:54
	→ chrischambers joined	12:54
	yaotti_aw → yaotti	12:55
	→ Titosemi joined	12:56
Titosemi	hey guys, it's possibñe to make a merge in my master from another brach just until a desire commit with it sha1?	12:57
cbreak_work	just merge the sha	12:57
wereHamster	Titosemi: git checkout master && git merge <sha1>	12:57
Titosemi	wereHamster: thanks	12:58
	→ und0 joined	12:58
	offby1 is again amazed at wereHamster's mind-reading powers	12:58
	→ _ollie joined	12:59
	← Cromulent left	13:00
	→ jaql joined	13:00
Titosemi	and how can i see which commits are from a particular branch?	13:01
	← _ollie left	13:02
cbreak_work	not	13:02
	commits are not from a branch	13:02
	← KU0N left	13:02
cbreak_work	they can be in one	13:02
	you can ask git branch -a --contains <sha> for that	13:02
	but you can't know in which branch a commit was made	13:02
Titosemi	aham, ok	13:03
	← jony2 left	13:03
	→ KU0N joined	13:03
	← dvaske left	13:04
	→ nevyn^_ joined	13:05
	← nevyn^_ left	13:05
	← masonkramer_ left	13:05
	← nevyn^__ left	13:05
	→ laj joined	13:06
	→ nevyn^_ joined	13:06
Titosemi	so i can't have a view like this within gitx, isnt it? http://img254.imageshack.us/img254/2363/picture1od.png	13:06
cbreak_work	isn't that the normal view of gitx?	13:07
	if you show all branches?	13:07
	well, visual differences aside :)	13:08
	yaotti → yaotti_aw	13:08
Titosemi	hehehehe	13:08
	mine looks like this http://img829.imageshack.us/img829/5843/picturejl.png	13:08
cbreak_work	that's more or less the same	13:09
Titosemi	but you don't see two lines, one for master, and one for develop	13:09
	→ dels joined	13:09
Titosemi	also a line for each brach :)	13:09
cbreak_work	because master ends on the second line from the bottom	13:09
	each branch has one line	13:09
Titosemi	ok, that's a good news, because i plan to make some merge into master	13:10
	:)	13:10
	cbreak_work: thank you :)	13:10
cbreak_work	should all be fast forward	13:10
	np	13:10
Titosemi	really?	13:10
	i'm doing all the merge with no-ff	13:11
cbreak_work	:/	13:11
	I like fast forward	13:11
	my history is like a single line.	13:11
	→ kanru joined	13:11
Titosemi	i am following nvia branching model	13:11
	yesm that's what i dont want to :)	13:12
	← alnewkirk left	13:12
Titosemi	http://nvie.com/posts/a-successful-git-branching-model/	13:12
cbreak_work	why not?	13:12
	my branching model is successfull :)	13:12
Titosemi	hahaha	13:12
cbreak_work	I make a branch for each thing I want to do	13:12
Titosemi	my wasnt at all, that's why	13:12
cbreak_work	do it in the branch	13:12
	then rebase and fast forward it into my master	13:12
Titosemi	i'm liking this one	13:12
	→ JoeCoolDesk joined	13:13
	→ xiong joined	13:14
	→ archis joined	13:18
	← archis left	13:18
	→ archis joined	13:18
	← dquestions left	13:18
	→ dquestions joined	13:18
	→ coderdad joined	13:22
	← coderdad left	13:22
	→ coderdad joined	13:22
	→ shentino joined	13:23
	← justin-george left	13:23
	← storrgie_ left	13:24
ptor	I use --no-ff and --log to get an easy view of the actual changes (e.g. related to a number of problem reports) that get merged into the release branch. For other things, e.g. some "normal" feature work that I keep on the side for a while I use normal --ff.	13:24
	→ blalor joined	13:25
	→ nohk joined	13:28
	→ devanjan joined	13:28
	← lundtor left	13:29
	→ injekt joined	13:30
	→ storrgie_ joined	13:31
	← storrgie_ left	13:31
	→ storrgie_ joined	13:32
	← wipple left	13:33
	→ Cromulent joined	13:36
Titosemi	ok guys, so the merge from master comes from the tag v.02. isn't possible that this merge is at the same hight at the develop one? (i'm not sure if i explained myself) http://img253.imageshack.us/img253/5519/picture1tz.png	13:37
	→ m0 joined	13:37
engla	Titosemi: what do you want to do?	13:38
cbreak_work	Titosemi: tags are yellow. branches are green. current branch is orange	13:39
Titosemi	yes :)	13:39
cbreak_work	as you can see, the branch release-0.2 is tagged with v0.2	13:39
Titosemi	but i want the branch merge	13:39
cbreak_work	branch merge	13:39
	?	13:39
Titosemi	to be at the same hight as the v0.2	13:39
	← roop left	13:40
cbreak_work	you can use fast forward merging, yes	13:40
Titosemi	i've made a merge in the master. That is at the top of the history	13:40
cbreak_work	just git checkout master, git merge v0.2	13:40
	well	13:40
	not anymore	13:40
	since master is already ahead	13:40
Titosemi	hehehe, ok. How can undo that?	13:41
cbreak_work	git log -g	13:41
	then note the previous hash	13:41
	then git reset --hard thathash	13:41
	← und0 left	13:42
Titosemi	thanks	13:42
cbreak_work	might be HEAD@{1}	13:42
Titosemi	:)	13:42
cbreak_work	(Told you! :)	13:42
	→ bentob0x joined	13:42
	← jspiros left	13:43
	Unknown[NF] → Unknown[OFF]	13:43
	Unknown[OFF] → Unknown[NF]	13:44
	→ FauxFaux joined	13:44
	→ dvaske joined	13:44
	← jkp left	13:45
Titosemi	:) what does the -g? i dont get what it says in the man. Walk reflogs?	13:46
	→ jkp joined	13:47
	← codetroll left	13:47
	→ simplecloud joined	13:50
	→ amerinese_ joined	13:51
cbreak_work	Titosemi: man git-reflog	13:52
jast	Titosemi: the 'git-reflog' manpage can be found at http://git.or.cz/man/git-reflog [automatic message]	13:52
cbreak_work	it shows the same as git reflog, just diferently	13:52
	→ thiago_hel joined	13:54
	yaotti_aw → yaotti	13:54
	← amerinese left	13:54
Titosemi	thanks	13:55
	← dpino left	13:56
	→ chessguy joined	13:58
	→ zombor joined	13:59
	→ smolyn joined	14:00
	→ sftp joined	14:01
	→ lol_h joined	14:02
	→ SimonP86 joined	14:02
	→ jspiros joined	14:03
	→ zheng joined	14:04
	← artefon left	14:05
	← bentob0x left	14:05
	← thiago_hel left	14:05
	← Cromulent left	14:06
	→ Cromulent joined	14:10
	← kirthi left	14:11
	← jelmer left	14:12
	→ jelmer joined	14:13
	← cilly left	14:15
	→ MattDiPasquale joined	14:15
	← jspiros left	14:15
	→ jspiros joined	14:16
	→ cilly joined	14:17
	→ normanrichards joined	14:18
	← archis left	14:18
	← chessguy left	14:18
	← dnordstrom left	14:19
	→ dannas joined	14:20
	→ Yuffster joined	14:21
	← henrikbjorn left	14:22
	← JoeCoolDesk left	14:24
Titosemi	ho do i use difftool with a stage file?	14:24
	ok, got it	14:25
	← amerinese_ left	14:25
offby1	any time	14:25
	→ q0rban joined	14:27
	← ketas-l left	14:28
flavius	Upon merging two brancher, will the histories also be merged?	14:28
offby1	of course.	14:28
	what else could "merging" possibly mean?	14:28
flavius	I mean the logs	14:28
offby1	flavius: general-purpose git advice: 1) Read "Git For Computer Scientists" (http://eagain.net/articles/git-for-computer-scientists/); 2) Run "gitk --all"; 3) Learn about "git reflog"	14:28
flavius	offby1: thanks, will do	14:29
	→ simplechat joined	14:29
	← simplechat left	14:29
	→ simplechat joined	14:29
	→ sixteneighty joined	14:29
	← blalor left	14:30
	→ blalor joined	14:31
	→ Aleph_One joined	14:31
	→ ketas-l joined	14:31
	← rraasch left	14:32
	← paddy__ left	14:33
	← sambarino left	14:33
	→ ilteris__ joined	14:33
	→ mkramer joined	14:33
	← normanrichards left	14:36
	← MattDiPasquale left	14:36
	← vdv left	14:36
	→ HG` joined	14:36
	← flaguy48 left	14:37
	→ kukks joined	14:37
mastro	offby1, flavius actually i'm curious too... how do the output of git log handle it? does it only show one of the parent branch?	14:39
	→ shamrock joined	14:39
	← trupheenix left	14:39
SethRobertson	inkedmn: pnig?	14:40
	or even ping	14:40
	← nsussman left	14:41
	→ \mjd joined	14:41
	→ Yuffster_ joined	14:42
	← Yuffster left	14:45
	→ mfacenet joined	14:45
	← mfacenet left	14:45
	→ mfacenet joined	14:45
	→ heptadecagram joined	14:46
	← heptadecagram left	14:47
	→ LiamH joined	14:47
	→ NickyP joined	14:49
	→ nothingHappens joined	14:49
	→ jkp_ joined	14:49
	← dannas left	14:50
	← unrael_ left	14:50
	→ unrael_ joined	14:50
	← frogonwheels left	14:51
	← jkp left	14:52
	→ beatak joined	14:54
	← jkp_ left	14:54
	← mfacenet left	14:54
	→ TheMoonMaster joined	14:55
	← freimatz left	14:56
	→ jkp joined	14:56
	← workmad3 left	14:57
	← NickyP left	14:58
	← devanjan left	14:58
	← TheMoonMaster left	14:58
	→ ive joined	14:59
	← ive left	15:00
	← petercoulton left	15:01
	← LRN left	15:02
	→ cannonball joined	15:02
	← JohnBat26 left	15:03
	← jauler left	15:03
	→ LRN joined	15:03
	→ normanrichards joined	15:04
	→ Dave^\| joined	15:04
	→ agile joined	15:05
	← marc_kdab left	15:05
	← rgr_ left	15:06
	← kanru left	15:06
	← cbreak left	15:07
	← jshsu left	15:10
	→ jshsu joined	15:10
	Unknown[NF] → Unknown[OFF]	15:11
	→ kylehayes joined	15:11
	← psoo left	15:11
	← zheng left	15:12
	← xiangfu left	15:12
	→ ziou joined	15:12
	ivey_off → ivey	15:12
	← fruight left	15:13
	ChanServ set mode: +v	15:13
	→ cbreak joined	15:13
	→ Pewpewarrows joined	15:14
ziou	hi, i have a GIT clone of the linux kernel. I put my code into it, commit etc. sometimes, i push it to a "master public repos". Sometimes, I pull from linus's 2.6 GIT, sometimes the .x.y.z stable release (different GIT, also based on linus's GIT repo). Then it finds some differences which are not fast forwarded	15:16
	instead they're commited as real diffs	15:16
cbreak_work	git does not commit anything as diff	15:17
	→ heptadecagram joined	15:17
ziou	what i mean is that they're not being fast-forwarded	15:17
	even thus the changes I pull do not touch my local changes	15:17
	→ golodhrim joined	15:17
ziou	i remember reading that it's on purpose, but I can't find out why/how/if it can be disabled/if there's a better way to do what i do	15:18
	ivey → ivey_off	15:19
	→ __name__ joined	15:19
cbreak_work	of course you won't get fast forwards if you pull from different branches	15:20
	you do know what pull means, right?	15:20
	← Schmallon left	15:21
cbreak_work	it's a combination of fetch and merge	15:21
	ivey_off → ivey	15:21
	→ boombatower joined	15:21
	← boombatower left	15:21
	→ boombatower joined	15:21
cbreak_work	the merge part merges the branch you pull into your current branch	15:21
	← xiong left	15:21
	→ bhosie joined	15:22
	→ flaguy48 joined	15:22
cbreak_work	so every time you pull from an other branch, it'll give you a new merge commit, unless it can be fast forwarded (which is unlikely)	15:22
	← dvaske left	15:24
	← imajes left	15:24
	← Jippi left	15:24
	→ jstemmer joined	15:25
	← mdunham918_ left	15:26
	→ imajes joined	15:26
	← \mjd left	15:26
	→ hasrb joined	15:26
ziou	cbreak_work: lets say i follow linux-2.6/master. I branch it to linux-2.6/2.6.36. There is another git repos which has 2.6.36.2 (stable fixes). I want the fixes, so I pull from that branch. These changes won't be fast-forwarded. Can I change my branch to rebase itself on the 2.6.36.2 upstream so that it gets fast forwarded? With the knowledge, that 2.6.36.2 upstream has been merged from the same 2.6.36 base as me	15:27
	→ fruight joined	15:27
	← sonnym left	15:27
ziou	that is so that I can keep on pulling easily 2.6.36.3, .4 etc etc (as in it all gets fast-forwarded unless there are local changes that i made)	15:27
	← kylehayes left	15:28
	→ kylehayes joined	15:28
	← ivey left	15:29
	→ hohoho joined	15:29
cbreak_work	ziou: man git-rebase	15:29
jast	ziou: the 'git-rebase' manpage can be found at http://git.or.cz/man/git-rebase [automatic message]	15:29
	→ proycon joined	15:30
	→ yhager_ joined	15:30
	← urkud left	15:31
proycon	I have a git repository on github, and an SVN repository at work.. both already exists and now contain the same data.. I now want to make a local git repository that can read/write from/to both github and svn (through git svn)... is this possible?	15:31
	(the problem being that the repositories already exist, I'm not starting from scratch)	15:32
ziou	cbreak_work: ok - does it means GIT remember what you merged from somewhere? I though it was only using the unique hashs so that it could discover what's fast forwardable on it's own no matter the source, if they have common ancestors	15:33
	→ fedesilva joined	15:33
	→ ivey joined	15:34
	← hasrb left	15:34
	← kai left	15:34
cbreak_work	it does not store that	15:35
	→ unon joined	15:37
	← tjholowaychuk left	15:37
doener	sure it does, as the merge commits combine the histories by means of having multiple commits	15:37
	s/commits/parents/	15:37
proycon	hmm.. I think I'm onto something now... git svn is smarter than I thought	15:37
doener	... assuming that we're talking about histories, not plain changes.	15:38
proycon	I just did a git svn init in the github clone	15:38
	→ hebz0rl joined	15:38
	→ rhl joined	15:39
	→ Buglouse joined	15:40
	→ marc_kdab joined	15:40
	← resmo left	15:42
ziou	cbreak_work: i've read the man page but i still don't really understand what's happening. I made a drawing that makes it clearer, if you don't mind helping my newbiness	15:43
	http://goo.gl/CDpl4 (PNG image)	15:43
	← jspiros left	15:44
	→ Textmode joined	15:44
cbreak_work	going home soon, sorry	15:44
	→ lresende joined	15:44
cbreak_work	git-for-computer-scientists should help, google it, maybe :)	15:45
ziou	well, or anyone else of course, the PNG is really making the issue clear	15:45
	← galderz left	15:45
	→ tjholowaychuk joined	15:45
	← Textmode left	15:46
	← cbreak_work left	15:46
	→ Textmode joined	15:46
proycon	hmm.. no luck :( I get Unable to determine upstream SVN information from HEAD history	15:47
	← muep left	15:47
	→ jspiros joined	15:48
	→ abetaha joined	15:48
ScottO_	how would i get the history for a single file?	15:49
	← Yuffster_ left	15:52
	← imajes left	15:53
tango_	ScottO_: git log filename	15:53
	→ fsk joined	15:53
	← fsk left	15:54
	→ madewokherd joined	15:55
	→ TheMoonMaster joined	15:56
	→ _psychic_ joined	15:56
	← simplecloud left	15:56
	→ lchm joined	15:56
	→ jmil joined	15:58
	← Aleph_One left	15:58
	← OOPMan left	15:59
	→ alvaro_o joined	15:59
	yaotti → yaotti_aw	16:00
	→ psoo joined	16:00
	→ thiago_hel joined	16:01
	← thiago_hel left	16:01
	→ thiago_hel joined	16:01
	thiago_hel → thiago_arn	16:02
blalor	I just did a "git svn rebase" while on master. one update came in, with commit id be21c. the output of "git diff be21c" is empty, however.	16:03
	"git log -p" shows a diff in one file, tho. What am I doing wrong?	16:04
	← lchm left	16:04
	→ ttn joined	16:05
awilkins	blalor, Have you made any changes to master that are not in the remote SVN branch?	16:05
blalor	no. doing all my work on a local branch	16:05
	← ttn left	16:06
	← ayanes left	16:07
awilkins	So the revision be21c is the same as your current HEAD checkout?	16:07
	← friskd left	16:07
	← marc_kdab left	16:07
blalor	I'd done essentially the same change (minus some whitespace difference, I think) on my "localdev" branch and then stashed it, without committing to "localdev"	16:07
	should be; how can I check?	16:07
	← jspiros left	16:07
	→ eletuchy joined	16:08
	← ziou left	16:09
awilkins	If `git diff be21c` returns nothing than your current tree is identical to be21c	16:09
blalor	"cat .git/refs/heads/master" returns be21c44ca51f5fcc1b781f6491c7dda031f94254 (the long whole sha1)	16:09
	→ jspiros joined	16:09
	→ miker__ joined	16:09
	→ sonnym joined	16:09
	← miker__ left	16:09
blalor	hrm. ok. so how do I see the contents of just that commit? `git show be21c`?	16:10
awilkins	svn rebase takes all the remote changes and replays your local commits after them - so it always commits new revisions,	16:10
	← yhager_ left	16:10
	→ goldins73 joined	16:12
	← goldins73 left	16:12
awilkins	Since it's the last commit on your local branch, try git diff HEAD^	16:12
	→ roop joined	16:12
	→ rgr joined	16:13
blalor	ok, I misunderstand "git diff"; i'm thinking of it as a svn revision and basically want "svn diff -c <revision>".	16:13
thiago_arn	you can do a similar thing:	16:13
	git diff sha1~..sha1	16:13
mapreduce	git show <revision>	16:13
thiago_arn	equivalent to svn diff -r revision-1:revision	16:13
mapreduce	sha1 rather than revision, sorry.	16:14
	← simplechat left	16:14
awilkins	or bzr diff -r 1..2 :-)	16:14
blalor	ack. thanks, thiago_arn. "git diff sha1~..sha1" == "git diff sha1^" ?	16:14
thiago_arn	there's also the not-widely-known git diff sha1^!	16:14
	blalor: no	16:14
	sha1~ == sha1^	16:14
	"git diff sha1~" compares the current tree to sha1~	16:15
	"git diff sha1~..sha1" compares sha1~ to sha1	16:15
blalor	isn't that what I typed? doesn't "sha1^" compare the predecessor to sha1 to sha1?	16:15
awilkins	Squiggle is "-1" ? So is ^ "-1..0" ?	16:15
thiago_arn	it's not what you typed	16:15
	← saschpe left	16:16
thiago_arn	no	16:16
	~ is "first parent"	16:16
	→ OOPMan joined	16:16
thiago_arn	^ is also "first parent"	16:16
	~2 is "first parent's first parent"	16:16
awilkins	Right, so ~ == ^	16:16
thiago_arn	^2 is "second parent"	16:16
awilkins	Erk	16:16
thiago_arn	and ^! is not ^	16:16
	→ ymasory joined	16:16
awilkins	Excellent, learning things :-)	16:16
	→ Jippi joined	16:16
	blalor goes to soak his head	16:17
	← gebi left	16:17
thiago_arn	X^! is the same as X~..X (if X has only one parent)	16:17
	← HG` left	16:18
	← kylehayes left	16:18
	→ gebi joined	16:18
	← gebi left	16:18
	→ gebi joined	16:18
awilkins	Ah, I thought when you were using ! it was to convey excitement at the not-widely-known-ness :-)	16:18
thiago_arn	ah, no, sorry	16:19
	→ pigdude joined	16:19
	thiago_arn usually writes them so that it's not ambiguous	16:19
thiago_arn	apparently I forgot this time	16:19
	← dels left	16:19
awilkins	Heh.... the pling is not on the community book page about treeishes	16:20
	→ friskd joined	16:20
thiago_arn	awilkins: you can see it in "man git-rev-list"	16:20
jast	awilkins: the 'git-rev-list' manpage can be found at http://git.or.cz/man/git-rev-list [automatic message]	16:20
blalor	awilkins: pling?	16:20
thiago_arn	make that rev-parse	16:20
	→ pratz joined	16:20
	→ sandbags joined	16:21
	← sandbags left	16:21
	← pratz_ left	16:21
	→ sandbags joined	16:21
	← sandbags left	16:21
	→ sandbags joined	16:21
	→ MattDiPasquale joined	16:22
	→ dlistex joined	16:23
	→ xelister joined	16:25
awilkins	blalor, '!' == pling (very old slang we used to use about 25 years ago :-) )	16:25
xelister	I change tabs/ident format in a file	16:25
	← codespectator left	16:26
xelister	how to do it so that diff doesnt show like if entire file changed with milion - and milion +	16:26
	i.e. spaces -> tab conversion	16:26
	→ airborn joined	16:26
blalor	awilkins: guess I'm not that old. I'd call it a "bang"; part of sh-bang (#!)	16:26
	→ JonMarkGo joined	16:26
	← roop left	16:26
	← thiago_arn left	16:27
JonMarkGo	I just did a git reset --hard a2.... to reset back to a previous revision, how do I now commit and push this reverted code?	16:27
dtolj	Does git honour the http_proxy string? its set but after trying to clone a project it times out unable to connect to socket? I am trying to go thru proxy	16:28
	→ kylehayes joined	16:28
	jaredhoyt[work] → jaredhoyt	16:29
	→ Schmallon joined	16:30
	yaotti_aw → yaotti	16:30
	→ radost joined	16:31
	→ Joghurt joined	16:31
	← dlistex left	16:32
	→ xanderbeedle joined	16:33
xanderbeedle	Is it possible to interactively squash a merge?	16:34
	← skoop left	16:34
	→ ch077179 joined	16:35
	→ Zephyrus joined	16:38
	→ SegFaultAX joined	16:38
	← friskd left	16:39
	→ kalusn joined	16:39
SegFaultAX	I have to work against an SVN repo, and one of my committers just fucked a file that I had changes to. Now I am getting a merge conflict. How can I rebase to his version and then just manually edit my changes back in?	16:39
	← Shoobdidoo left	16:40
	← rolfb left	16:40
SegFaultAX	Sorry about the language, 3 days of work just got flushed. :/	16:40
apanek-work	SegFaultAX: stack your changes, pull his, unstack your changes and merge them	16:41
	At least that's one way to do it	16:42
	→ bauruine joined	16:42
	→ icwiener joined	16:43
SegFaultAX	apanek-work: I can't figure out how to let me merge his stuff in.	16:44
	→ carutsu_ joined	16:45
SegFaultAX	How can I just pull his version/	16:45
apanek-work	SegFaultAX: Have you committed your changes?	16:45
SegFaultAX	Yes.	16:45
	And git svn rebase, now are merge conflict stage.	16:45
	Now at*	16:45
apanek-work	SegFaultAX: You could also push your changes to another branch, pull his changes and then merge the two branches in git	16:46
	← JonMarkGo left	16:46
apanek-work	Which is probably way easier than working against the SVN repo directly.	16:46
SegFaultAX	apanek-work: I am already at the merge stage, how can I just use his version.	16:46
apanek-work	SegFaultAX: git svn revert or something?	16:47
SegFaultAX	apanek-work: No. How do I make his file win the conflict?	16:47
	I don't want to revert the whole rebase. That does me no good, I'll be out of sync with the trunk.	16:48
apanek-work	SegFaultAX: git mergetool ?	16:48
	→ SRabbelier joined	16:48
apanek-work	Anyways, gotta go - weekend time :P	16:48
	← fedesilva left	16:49
	← abetaha left	16:51
	→ radost_ joined	16:54
	→ fedesilva joined	16:55
	→ diwwin joined	16:55
	← Cogito left	16:56
	← Zhonghua left	16:56
	← radost left	16:57
	→ hasrb joined	17:00
	← Angel-SL left	17:00
	← sebbi left	17:00
	→ TheFox21 joined	17:00
	→ t0rc joined	17:01
	→ codespectator joined	17:02
	→ nevyn^__ joined	17:02
	yaotti → yaotti_aw	17:02
	→ friskd joined	17:03
	← heptadecagram left	17:03
	← nevyn^__ left	17:03
	→ nevyn^__ joined	17:03
	← diwwin left	17:03
	← nevyn^_ left	17:04
	← teweWork left	17:04
	← crashanddie left	17:05
	→ abetaha joined	17:05
	→ ChUrP joined	17:08
	← kalusn left	17:08
	q0rban → q0rban\|lunch	17:08
	→ fredmorcos joined	17:09
	← zangetsu left	17:09
	← SianaGearz left	17:11
	← abetaha left	17:11
	→ abetaha joined	17:13
	→ TheMoonM_ joined	17:14
	← TheMoonMaster left	17:15
	→ [Po]lentino joined	17:15
	→ hobodave joined	17:16
	Unknown[OFF] → Unknown[NF]	17:16
	→ khmarbaise joined	17:17
	→ ilogger2 joined	17:18
	→ joel_ joined	17:18
	→ saysjonathan joined	17:19
	→ amerine joined	17:21
	→ Fullmoon joined	17:21
	→ yhager_ joined	17:21
	→ noahcampbell joined	17:21
	→ diwwin joined	17:24
	→ kpreid_ joined	17:26
	→ urkud joined	17:27
	kpreid_ → kpreid	17:30
	→ kenichi joined	17:30
ScottO_	is it possible to merge only a single commit into a feature branch from master?	17:30
	→ newls3 joined	17:30
jast	git cherry-pick	17:30
	→ jarrod_ joined	17:30
ScottO_	ill look into it thanks	17:31
	→ zangetsu joined	17:31
	← diwwin left	17:32
	→ shennyg joined	17:33
ScottO_	jast: that works the other way around? like if i made the change in my feature branch and wish to merge to the master?	17:33
newls3	[rejected] master -> master (non-fast-forward)	17:35
davr	cherry-pick doesn't merge, it copies the commit	17:35
	→ kalusn joined	17:35
ScottO_	oh cool	17:35
	→ Eiler joined	17:35
newls3	To prevent you from losing history, non-fast-forward updates were rejected	17:35
jast	newls3: http://git.or.cz/gitwiki/GitFaq#non-ff	17:35
newls3	thx	17:35
	dang! that was quick jast	17:35
	right to the problem	17:36
ScottO_	so i should be able to just checkout master and then cherry-pick the commit from the feature branch	17:36
davr	newls3: it's a very very very common thing to people who are new to git. I almost pasted that link, but he beat me by two seconds :P	17:36
newls3	ahh! cool	17:36
	i just use git for minor heroku pushes and was like wtf!	17:36
	never seen that with the apps i put up there	17:37
	learn something new everyday	17:37
	← kalusn left	17:37
blalor	actually, newls3 & jast: https://git.wiki.kernel.org/index.php/GitFaq#non-ff	17:38
	the redirect loses the anchor	17:38
jast	oh? worked for me iirc	17:38
	yeah, still does	17:38
	well, lemme change the alias anyway	17:38
blalor	maybe browser-dependent; Didn't work for me w/ Safari	17:38
newls3	New one after I did that	17:39
	[remote rejected] master -> master (pre-receive hook declined)	17:39
	looks like I have a sucky commit	17:40
jast	heroku is some kind of hoster, right? it's something they are doing	17:40
newls3	yeah they are a hoster	17:40
jast	the hook ought to have given a more detailed message somewhere above that line	17:41
	→ Anti-X joined	17:41
xanderbeedle	When doing a merge, is it possible to do an interactive squash, similar to interactive rebase?	17:41
jast	xanderbeedle: not really. merges are all-or-nothing.	17:42
	you can use a series of cherry-pick -n invocations	17:42
xanderbeedle	Fair enough	17:42
	Is there any kind of reference to the point when a branch was made?	17:43
jast	or do a merge --squash, then revert individual commits, then use interactive rebase to squash that together	17:43
	no	17:43
xanderbeedle	Then, I could interactively rebase my feature branch before merging	17:43
jast	you can figure out the current merge bases, though	17:43
	git merge-base -a <branch 1> <branch 2>	17:43
kevlarman	is there any way to get git-svn to work nicely with an svn repo that has 2 branches at the root (/trunk and /stable), and (annoyingly) a bunch of files outside of those 2 branches?	17:45
jast	kevlarman: not really	17:45
davr	yes, if you don't mind ignoring the files outside those two branches	17:46
kevlarman	well, i do	17:46
xanderbeedle	jast: so that ancestor will normally be the commit before the branch took place?	17:46
jast	that's pretty far off common conventions, and it's a bit difficult to support every possible layout	17:46
davr	are they in another directory outside those two branches, or are they on the same level?	17:46
kevlarman	but i can use svn for the files outside if it comes to that	17:46
jast	xanderbeedle: no, it's the most recent thing that both branches have in common, i.e. if you merged in the meantime (squashed merges don't count), it's usually that most recent merge	17:46
davr	in another directory you can do it (just make it another branch in git, doesn't matter that it has nothing in common with the others), on the same level...I don't know if that's possible	17:47
kevlarman	davr: both actually	17:47
	→ spearce joined	17:48
davr	you MIGHT be able to make another git branch pointing at the root svn folder, and then tell it to ignore the svn folders that contain the other branches, but I'm not sure if that would work	17:48
xanderbeedle	jast: in my model, I work in feature branches that get deleted after a merge	17:48
	i imagine a rebase would throw things out?	17:49
	→ aresnick joined	17:49
jast	xanderbeedle: come again?	17:49
	→ varuntv joined	17:49
xanderbeedle	nevermind, i'm not talking sense :P	17:49
jast	okay :)	17:49
kevlarman	davr: switching branches like that would do a ton of damage anyway	17:50
	davr: how would i get it working with just the 2 branches?	17:50
	it's not obvious from the man page (-T/-t/-b don't really fit here)	17:50
davr	create the git-svn repo just pointing at one of the branches, tell it there are no branches or tags. then manually add the second one	17:50
jast	it might be possible to use -T twice	17:51
	→ dreiss joined	17:51
kevlarman	O_o	17:51
	→ foocraft joined	17:51
jast	i.e. -T branch1 -T branch2	17:51
	→ khmarbaise_ joined	17:52
kevlarman	jast: what do the refs look like in that case?	17:53
davr	for manually adding branches, you basically put in the .gitconfig to map svn paths to git refs	17:53
	khmarbaise_ → khmarbaise	17:53
davr	look at the very bottom of the git-svn help page	17:53
jast	kevlarman: no idea :) I only know for the case of multiple -b arguments, but that won't work for your case	17:54
	→ kipras joined	17:54
	→ genomega joined	17:54
	← urkud left	17:54
	→ khmarbaise_ joined	17:57
kevlarman	jast: well git didn't error on 2 -T's	17:58
	→ aah joined	17:58
kevlarman	let's see what happens in an hour or 2 :(	17:58
jast	good sign :)	17:58
	← khmarbaise left	17:59
	→ MikhX joined	17:59
	khmarbaise_ → khmarbaise	17:59
kevlarman	so far it looks like it's only folloing trunk	17:59
jast	it follows the other one starting from the first revision it appeared in	18:00
	← varuntv left	18:00
	→ Luyang joined	18:03
	→ khmarbaise_ joined	18:03
kevlarman	these look weird	18:03
	$ ls .git/refs/remotes/	18:03
	trunk trunk@277 trunk@95 trunk@95-- trunk@97	18:03
	→ g0bl1n joined	18:04
	→ eletuchy joined	18:04
	→ fr0sty joined	18:05
jast	that kind of thing happens if trunk got replaced a bunch of times	18:05
	→ marc_kdab joined	18:05
kevlarman	fetch = stable:refs/remotes/trunk <- and that looks wrong	18:05
	← khmarbaise left	18:06
	khmarbaise_ → khmarbaise	18:06
	→ varuntv joined	18:06
	→ justin-george joined	18:06
	← varuntv left	18:06
	→ radost joined	18:07
kevlarman	and it keeps going back to r1 for some reason	18:07
	→ floppyears joined	18:08
	← aresnick left	18:08
	← marc_kdab left	18:08
	← fr0sty left	18:08
	→ marc_kdab joined	18:09
	→ tro joined	18:10
	→ bhosie joined	18:10
jast	ah well, so the handling of two -T arguments perhaps isn't that good. try changing it to fetch = stable:refs/remotes/stable.	18:10
	→ pheaver joined	18:11
	→ HG` joined	18:12
kevlarman	jast: that was the only fetch line	18:12
	→ jkp joined	18:13
	← khmarbaise left	18:13
	→ rolfb joined	18:13
	← rolfb left	18:13
jast	hmm.	18:14
	I'd say just make it two fetch lines, but apparently the history is too chaotic for git-svn to grasp anyway	18:14
kevlarman	yeah this might be annoying	18:14
	lots of index mismatches fetching the second branch	18:15
	can't wait for these guys to just switch to git :(	18:15
	→ urkud joined	18:17
_psychic_	Which is recommended more: gitosis or gitolite?	18:17
	→ ShadeHawk joined	18:17
	→ mlq_ joined	18:18
albel727	_psychic_: gitolite, ofc.	18:18
	← mlq_ left	18:19
_psychic_	Kinda seems like a successor to gitosis.	18:19
	→ aresnick joined	18:19
albel727	_psychic_: gitosis is near dead... no, scratch that, it IS dead. gitolite is kinda (only) successor (that I know of).	18:20
_psychic_	i c	18:21
	cool, thanks albel727	18:21
albel727	np =)	18:21
blalor	what about ges? I tried that earlier today and gave up when the readme failed to get me up and running.	18:21
	← tro left	18:21
blalor	it SOUNDS interesting…	18:21
	_psychic_ plans to use gitolite + the chaw for now	18:22
kevlarman	jast: well, it sort of worked to use 1 -T and add stable afterwards, but i'm kinda worried it won't stay that way	18:24
	Anti-X → torvalamo	18:25
	→ steph021 joined	18:25
	→ mtx_init joined	18:25
kevlarman	Couldn't find revmap for svn+ssh://svn.example.com/repo/trunk/...	18:26
	W:unknown path/rev in svn:mergeinfo dirprop: /trunk:96	18:26
	Found merge parent (svn:mergeinfo prop): 1b53b7532327114a5ad03fcd9d08a7cffddc9787	18:26
	Couldn't find revmap for svn+ssh://svn.example.com/repo/trunk/...	18:26
	^from fetching the second branch	18:26
	→ keystr0k joined	18:27
	← aresnick left	18:28
	→ gh34 joined	18:29
	→ itamarjp joined	18:32
	→ andrewwatts joined	18:32
	→ vitor-br joined	18:33
	→ t0rc joined	18:36
	→ khmarbaise joined	18:38
	← dreiss left	18:39
	→ verto joined	18:39
	← marc_kdab left	18:41
	→ KU0N joined	18:41
	→ pdelgallego joined	18:41
	→ hakunin joined	18:43
	→ tro joined	18:43
kevlarman	btw while i'm here, is there anything resembling git-hg that doesn't suck yet?	18:44
	→ bronson joined	18:44
	← bronson left	18:46
	→ laidback_01 joined	18:47
	→ radost_ joined	18:48
laidback_01	hey, I'm using gitolite, and I've just about gotten gitweb running with it. I'm at the point where I'm getting a permission denied error "index.cgi: Can't cd to /home/git/repositories: Permission denied, referer: http://coderepo.krh.int/"	18:48
drizzd	kevlarman: does hg-to-git suck?	18:48
laidback_01	now, the perms for that directory: drwx------ 7 git git 7 Dec 10 05:09 repositories	18:48
	→ smolyn joined	18:48
drizzd	wow, it's rather short	18:49
laidback_01	so... should I be opening the permissions up a bit? I know that gitolite documentation says not to muck with the directory, but doe that include the perms changes?	18:49
drizzd	laidback_01: I am guessing your websever is not running as user 'git'?	18:50
kevlarman	drizzd: i haven't seen that one yet	18:50
drizzd	kevlarman: it's in contrib/hg-to-git	18:50
	but I don't know more about it either	18:50
laidback_01	drizzd, nope, it's as www, but I forgot about changing this in the .gitolite.rc -> $REPO_UMASK = 0027; # gets you 'rwxr-x---'	18:51
	← radost left	18:51
	← radost_ left	18:51
laidback_01	seems like I'll have to manually change the perms to match that mask. that's fine then.	18:51
bremner	laidback_01: yeah, you need to open up the permissions for gitweb	18:51
laidback_01	:q!	18:51
kevlarman	drizzd: does it work in both directions?	18:51
	→ Vortex34 joined	18:51
drizzd	kevlarman: I just skipped through the docs, but I doubt it	18:51
kevlarman	then it won't work for me	18:52
drizzd	kevlarman: maybe hg has the corresponding git-to-hg?	18:52
	→ dvaske joined	18:53
kevlarman	drizzd: there's hg-git, but import scripts don't work for actual daily use	18:53
	→ muep joined	18:53
	→ giallu joined	18:53
	← tro left	18:54
kevlarman	(hg-git is a bit more than that, but using it in place of git-hg is painful at best)	18:54
	← kpreid left	18:54
	→ jtaby joined	18:54
	← urkud left	18:55
	→ thiago_arn joined	18:55
	→ jrmuizel joined	18:57
	← thiago_arn left	18:59
	→ thiago_arn joined	18:59
	thiago_arn → thiago_home	18:59
	← jkp left	18:59
	→ skoop joined	19:00
	← skoop left	19:00
	→ skoop joined	19:00
	← andrewwatts left	19:00
	→ Heimidal joined	19:01
	→ dnordstrom joined	19:01
	→ dreiss joined	19:01
	→ sh1mmer joined	19:02
	→ gbacon_ joined	19:04
	→ Cromulent joined	19:05
	← eletuchy left	19:05
	← foocraft left	19:05
	→ Evious joined	19:06
	→ carutsu_ joined	19:07
	→ khmarbaise_ joined	19:08
	→ urkud joined	19:09
	→ nicoulaj joined	19:09
	← khmarbaise left	19:09
	khmarbaise_ → khmarbaise	19:09
	→ kalusn joined	19:09
	→ foocraft joined	19:11
	→ dbrock joined	19:11
	← carutsu_ left	19:12
	ChanServ set mode: +v	19:13
dbrock	is it possible to tell git to ignore future changes to a file that is already in the repository?	19:13
	→ carutsu_ joined	19:14
kevlarman	dbrock: no	19:14
dbrock	kevlarman: okay, thanks	19:14
	(the reason I ask is that I wanted to add an example configuration file so that people get it when they clone, but changes are ignored)	19:14
	→ proycon joined	19:14
	→ zhobbs joined	19:14
	← HG` left	19:15
DrNick	git mv config config.example	19:15
	← thiago_home left	19:16
dbrock	DrNick: yeah, that's what I just did :)	19:16
	simple and good solution	19:16
	→ tro joined	19:16
	→ thiago_home joined	19:16
	→ doener joined	19:17
	← Cromulent left	19:17
	→ tvw joined	19:18
	→ shentino joined	19:18
	← shentino left	19:18
	→ shentino joined	19:18
	→ andyburke joined	19:18
	→ samlt joined	19:19
	← carutsu_ left	19:19
	→ henrikbjorn joined	19:20
	→ mdpatrick joined	19:20
	→ dark joined	19:20
	→ jshsu joined	19:21
ScottO_	isn't cherry-picker a derogatory term?	19:21
dark	does git 1.7.3.3 have support for aliases running on the current dir? (I'm using a patched 1.7.3.2 for that)	19:21
	← samlt left	19:21
	← thiago_home left	19:22
	← khmarbaise left	19:22
	→ petran joined	19:22
davr	ScottO_: the only other way i've ever heard it used is that machine with a little basket that you sit in that raises you up	19:23
ScottO_	oh maybe im thinking about some other fruit picker	19:24
	← mdpatrick left	19:24
	→ mdpatrick joined	19:26
	← torvalamo left	19:26
	→ thiago_home joined	19:26
	→ Rai joined	19:27
	→ Yuffster joined	19:27
	→ GodEater joined	19:30
	← GodEater left	19:30
	→ GodEater joined	19:30
	→ HrznDefeated joined	19:31
	→ eletuchy joined	19:32
	→ mattz_ joined	19:33
	← smolyn left	19:33
	→ carutsu_ joined	19:33
	ChanServ set mode: +v	19:33
HrznDefeated	Is anyone familiar with the svn-to-git php script/tool?	19:33
	→ torvalamo joined	19:33
bremner	yikes.	19:34
	← kalusn left	19:34
HrznDefeated	Yikes? Scary script?	19:34
	→ kalusn joined	19:34
bremner	well, what does it do that git-svn does not?	19:34
HrznDefeated	Well it uses git-svn...	19:35
	→ mkramer joined	19:35
HrznDefeated	I guess it does some analysis for you and also helps in identifying branches (and making the correct choice) on repositories that are not vanilla	19:35
	afk- bbl	19:36
	← eletuchy left	19:37
	← newls3 left	19:38
	← jarrod_ left	19:38
proycon	I have a git repository (cloned from github), and a svn repository (central work repository). I want to sync this git repository with svn, in such a way that I can read and write from/to both github and svn. I don't mind doing this manually, my local working copy is to be the proxy. Is this possible?	19:38
	cause I've been fiddling for hours without much luck :/	19:39
kevlarman	proycon: there's an example of how to get that working in the git-svn man page	19:40
	← dbrock left	19:40
proycon	there is?	19:40
	hmm	19:40
	→ mfacenet joined	19:40
	← mfacenet left	19:40
	→ mfacenet joined	19:40
proycon	can't really see it,, the problem is that I want to sync with a remote git repository and a remote svn	19:42
	→ robotarmy joined	19:42
kevlarman	proycon: see the basic examples section	19:42
proycon	ok, I'm looking at it now	19:43
	thanks	19:43
	hmm, from what I see it doesn't really cover what I want	19:45
patrikf	proycon: what you want is not exactly the most intelligent setup	19:46
proycon	what would be a better recommendation?	19:46
	I basically want to use github, and my work's internal svn	19:46
patrikf	choose either git OR svn	19:46
proycon	and I want to use git	19:46
kevlarman	proycon: the last example is exactly your use case	19:46
	← skoop left	19:47
	→ hohoho joined	19:47
patrikf	you can work on SVN repos using git, but as soon as you're publishing the git repo you're in trouble	19:47
proycon	hmm, yes, that's the trouble I find myself in yes	19:47
bremner	proycon: here is an example of where I do something similar, but read-only access to svn http://paste.debian.net/102071/	19:47
proycon	bremner: thanks	19:48
kevlarman	patrikf: it works fine as long as you don't publish anything that isn't in svn yet	19:48
bremner	proycon: basically, I clone the git repo, then do some easy magic to get git-svn connected to a branch again	19:48
proycon	I do want people to be able to commit to svn too	19:48
	I'll just have to sync it then, that's ok	19:49
patrikf	kevlarman: right... but doesn't render that publishing the git repo useless?	19:49
	← carutsu_ left	19:49
patrikf	s/render that/that render/	19:49
proycon	and I also want people to be able to use github (both for reading and writing)	19:49
	I'm still studying that that last example	19:50
kevlarman	patrikf: you can still publish branches off of it	19:50
	and the main benefit is other people not having to git svn clone again	19:50
bremner	proycon: I don't see why commiting to svn wouldn't work in my case, I just haven't done it.	19:50
	→ troutwine joined	19:50
	→ kar8nga joined	19:51
proycon	let me see if I get your snippet	19:51
	← kenichi left	19:53
	→ rsewill joined	19:54
	→ marc_kdab joined	19:54
	→ pedrobelo joined	19:54
	← pdelgallego left	19:54
	← rsewill left	19:55
	→ ablemike joined	19:55
ablemike	hi all. looking for the best way to do this and I am not certain of the terms i should be searching for.	19:55
	→ rsewill joined	19:55
ablemike	I would like to step back and forth through revisions without creating branches	19:56
	almost like checking out the hash	19:56
	sha-1 i mean	19:56
	← robotarmy left	19:56
ablemike	whats the best way to do this?	19:56
	← itamarjp left	19:56
	→ andrewwatts joined	19:56
patrikf	ablemike: "git checkout <hash>"	19:57
proycon	bremner: hmm, no luck yet	19:57
	git svn info -> Unable to determine upstream SVN information from working tree history	19:57
	that is if I do a git svn info after your commands	19:57
ablemike	patrikf: i know i can do that	19:57
proycon	I should do a fetch first perhaps	19:57
ablemike	but i have to look up the hash everytime	19:57
bremner	proycon: you have to start with a valid git-svn repo	19:57
proycon	ah ok, I started from the github clone	19:58
	that makes sense yeah	19:58
patrikf	ablemike: you can use any of the methods listed in man git-rev-parse	19:58
jast	ablemike: the 'git-rev-parse' manpage can be found at http://git.or.cz/man/git-rev-parse [automatic message]	19:58
	← andrewwatts left	19:58
proycon	trying again	19:58
	→ itamarjp joined	19:59
	→ shennyg_ joined	19:59
	→ andrewwatts joined	19:59
	→ SimManiac joined	19:59
SimManiac	hi	19:59
	→ codespectator joined	20:00
	← mkramer left	20:00
	← shennyg left	20:01
	shennyg_ → shennyg	20:01
ablemike	patrikf: thanks but still a little fuzzy :)	20:02
	i know how to list the revisions	20:02
	I guess I was hoping for a shortcut to typing the sha-1 each time	20:02
	← itamarjp left	20:03
	← gbacon_ left	20:03
proycon	bremner: ok, and now I can add the github remote in the normal way?	20:04
	→ pantsman joined	20:04
	← pantsman left	20:04
	→ pantsman joined	20:04
ablemike	proycon git remote add something_here git@github....	20:05
patrikf	ablemike: yes. it's all in that manpage, in the "specifying revisions" section, especially "~" and "^" should be of use to you	20:05
	→ beatak joined	20:05
ablemike	something_here would be your identifier	20:05
proycon	yeah	20:06
	that part I know	20:06
ablemike	patrikf: putting on glasses and looking closer ...	20:06
	→ mkramer joined	20:06
	← mkramer left	20:06
ablemike	so wait ... i could do like: git checkout master-1 ?	20:07
	or git checkout master@{5}	20:08
	← codespectator left	20:08
kevlarman	ablemike: git checkout master~5 is probably what you want	20:10
	master@{5} is the state of master 5 changes ago, which won't be what you expect after something like a reset or rebase	20:11
ablemike	yeah i grok!	20:11
	awesome	20:11
	@{5} was more like a pure path of master and not the merges	20:12
	→ digitalfiz joined	20:12
kevlarman	~ always follows the first parent btw	20:12
	← hohoho left	20:12
	→ smolyn joined	20:12
kevlarman	master^1/master^/master~1 are all equivalent, master^2 will be the second parent, etc.	20:12
charon	ablemike: no you don't understand. master~5 means "5 first-parents back from current master". OTOH master@{5} means "5 steps back of what you did to your local master". see man git-reflog for the latter mechanism	20:13
jast	ablemike: the 'git-reflog' manpage can be found at http://git.or.cz/man/git-reflog [automatic message]	20:13
	← mtx_init left	20:13
	→ hasrb joined	20:14
	ablemike 's head is swimming	20:14
kevlarman	is jast the new gitbot or something? :P	20:14
jast	I'm filling in :)	20:14
charon	kevlarman: also, but the real jast is still there.	20:14
	→ bluegene joined	20:14
	← bluegene left	20:15
charon	ablemike: to put that in some real context: assume you have no local changes of any sort in your repo, and run 'git fetch && git merge origin/master' (or equivalently 'git pull') after a while	20:16
	← mfacenet left	20:16
charon	ablemike: this would usually (if upstream behaved) result in a fast-forward merge	20:16
ablemike	right	20:16
	single line	20:16
charon	suppose this fast-forward was of linear history, and gave you 5 new commits	20:16
	now master~5 is 5 steps back from the new state, i.e., your old state	20:17
ablemike	ok	20:17
	but if there was a branch	20:17
	and it had 5 commits	20:17
	and merged in	20:17
charon	OTOH master~3 is a commit that you never had checked in your repo	20:17
ablemike	wait	20:18
charon	furthermore master@{1} is "the previous state of the local master", which would be master before the merge	20:18
ablemike	holyshit i am so lost	20:18
charon	argh, i'm losing words	20:18
	→ mtx_init joined	20:18
ablemike	lol	20:18
charon	<charon> OTOH master~3 is a commit that you never had checked out in your repo	20:18
	→ robotarmy joined	20:18
	→ DaveIngram joined	20:19
charon	read http://marklodato.github.com/visual-git-guide/index-en.html if you still don't understand what ~ does	20:19
	→ sako joined	20:19
	← hasrb left	20:20
blalor	git a complex beast, but the diagrams are the best I've ever seen for an SCM	20:20
	→ hasrb joined	20:21
charon	i'd go further and say git is a very simple data model hiding under an unfortunately very complex UI	20:21
blalor	or that	20:21
	→ bdesk joined	20:21
kevlarman	it's not that complex, it just exposes a lot more of its low level guts to those who want it than other scms	20:21
charon	most of the diagrams are of the data model, hence simple ;)	20:22
	→ jackdempsey joined	20:22
	→ pdelgallego joined	20:22
bdesk	if i want to make a repo on a server, and i want to allow everyone with an account on the server to have read and write access, then what is the best way to set this up?	20:22
kevlarman	bdesk: best way is not to use accounts	20:23
	← robotarmy left	20:23
bdesk	kevlarman: everyone connected to the project already has an account on the server...	20:23
	→ banjiewen joined	20:23
charon	bdesk: ... but gitolite. second best way is to read the description of core.sharedRepository in man git-config	20:23
jast	bdesk: the 'git-config' manpage can be found at http://git.or.cz/man/git-config [automatic message]	20:23
	→ mfacenet joined	20:23
	← mfacenet left	20:23
	→ mfacenet joined	20:23
bdesk	kevlarman: are you going to suggest gitolite or somethingn instead?	20:23
	→ bluegene joined	20:23
	← bluegene left	20:23
	→ bluegene joined	20:23
	← bluegene left	20:23
kevlarman	bdesk: yes, and charon beat me to the second best way	20:24
	charon creeps back under his rock	20:24
proycon	bremner: Your 'git show-ref origin' yields an empty string here, I don't think it's supposed to right?	20:24
bdesk	kevlarman: ok well i am too sucky of a sysadmin to do gitolite so maybe i'll try the second way	20:25
	→ Buglouse joined	20:25
kevlarman	bdesk: the second way is more work	20:25
	→ bluegene joined	20:25
bdesk	...	20:25
	then i'll just make everything globally read-write and the users can clone directly using ssh i guess?	20:26
	bdesk: man git-config	20:26
jast	bdesk: the 'git-config' manpage can be found at http://git.or.cz/man/git-config [automatic message]	20:26
bdesk	lol	20:26
kevlarman	the issue is files not staying world writable	20:27
	→ icwiener joined	20:27
bdesk	so you are saying that gitolite is the easiest way other than paying some git hosting company?	20:27
	gitolite has soooo much documentation and i don't know how to linux ;_;	20:27
bremner	proycon: the concept is you have already pushed that branch, and you are cloning somewhere else. So origin/blah should exist at that point.	20:28
	→ jpitts joined	20:28
jast	well, just init your repositories on the server with git init --bare --shared=all	20:28
	all done!	20:28
	← jpitts left	20:28
proycon	bremner: ah wait, I first pull github and then do those things?.. that makes sense yeah	20:28
bdesk	jast: thanks i'll try that. not sure if you're being sarcastic but i'm assuming not	20:28
	→ jpitts joined	20:29
jast	try it and you'll see :)	20:29
eean	bdesk: yea you might want to use gitorious or something if you are using linux as a verb :D	20:29
bdesk	gitorious is easier?	20:29
jast	yeah, well, I gotta go. got some linuxin' to do. :P	20:29
	gitorious hosts open source stuff for you for free	20:29
bdesk	o	20:29
jast	as do repo.or.cz and github.com	20:29
eean	as does github	20:29
bdesk	well i'm using github but now i need something private	20:30
	and i have a server	20:30
jast	right. in that case, just follow my elaborate recipe.	20:30
eean	github is pretty cheap for small stuff	20:30
bdesk	i'm not gonna get addicted to the github crack	20:30
	→ Saviq joined	20:31
	→ cstephens4 joined	20:32
	ChanServ set mode: +v	20:32
	← yhager_ left	20:33
jpitts	if i want to track a svn repo with git svn, and I have multiple folders for branches and tags ( horrible, I know ) do I have any options WRT getting them all into a git repo?	20:33
jast	jpitts: you can use -b and -t multiple times on the intial command line	20:33
	← henrikbjorn left	20:35
jpitts	oh hey it even say that, under the stdlayout section, which I was ignoring because I was convinced mine wasnt :) thanks	20:35
proycon	bremner: hmm... still no luck, am I doing something wrong still? http://pastebin.com/9DSiffg6	20:35
troutwine	Is this the proper channel to ask questions about gitosis issues?	20:36
	← keystr0k left	20:38
jast	troutwine: yes, but 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	20:38
bremner	proycon: "github" is a remote; in my case "origin/sourceforge" is a branch	20:38
proycon	ah ok, so I first sync the remote into a branch	20:38
	thanks, will give that a try	20:38
laidback_01	bdesk, hey, I just used this setup last night and yesterday. Between this and sitaram's documentation, gitolite came together well. I stopped after I got giolite and gitweb working correctly - no need for me to use svn and/or trac. http://blog.august8.net/2010/11/03/gitolite-with-svntrac-repositories/	20:39
	← jpitts left	20:39
bdesk	laidback_01: ty	20:40
laidback_01	that recipe is for debian based setups, but it's easy to adapt - I used it for gitolite installation on freebsd without issue.	20:40
	yep	20:40
	→ SianaGearz joined	20:41
	→ jarrod_ joined	20:42
	→ MattDiPasquale joined	20:43
	→ thelazydeveloper joined	20:44
	→ dl9pf_ joined	20:45
	→ mintry_ joined	20:45
	→ jhaddad joined	20:46
	→ LRN joined	20:47
	← torvalamo left	20:50
	← jackdempsey left	20:51
	→ FabianB_ joined	20:52
	← mintry_ left	20:53
troutwine	jast: I did not know that. Thank you.	20:53
	→ torvalamo joined	20:55
	→ mc1 joined	20:56
	← jshsu left	20:57
	→ jshsu joined	20:57
	← mc1 left	20:58
HrznDefeated	Hello, I'm trying to clone an SVN repository into Git (with the intent to migrate from SVN to Git). I seem to be having some trouble getting git-svn to work. I've done it exactly how it suggests in the "Pro Git" (and a couple other ways) but I either get errors or it just stops doing the conversion with several revisions left to go.	20:58
	← LRN left	20:59
HrznDefeated	What kind of resources are available to help with more difficult migrations? (Links? Experts?)	20:59
	← Luyang left	20:59
HrznDefeated	Google has been a run around for me so far...	21:00
charon	HrznDefeated: most likely a pastebin of your actual errors would be far more helpful than generic questions	21:01
	also, see how new your git is	21:01
	→ LRN joined	21:02
	→ piotr joined	21:02
	→ fredmorcos joined	21:02
HrznDefeated	charon: Got an error I'll send right over (revmap related?). But I am also interested in just general resources...	21:03
	→ linusoleander joined	21:03
	→ Chillance joined	21:05
charon	i wrote this one some time ago: http://thomasrast.ch/git/git-svn-conversion.html	21:06
	→ benjf joined	21:07
HrznDefeated	Reading it now	21:08
	→ mmw joined	21:09
charon	however, git-svn is still the centerpiece (for now...) so if that fails, you have to fix the failure first :)	21:10
tangledhelix	i have a submodule inside a repository and i wanted to fetch it from a new url. i removed its entry in .gitmodules, but now it complains no matter what i do, and i can't 'git submodule add' the new URL, it says it already exists. i don't see anything like 'git submodule rm' in the help. any suggestions?	21:12
	→ Eridius joined	21:12
SethRobertson	`git rm` the old submodule mount point. YOu might have to `git rm --cached`	21:12
	tangledhelix: Are you happy with submodules or can I promote an alternative?	21:13
tangledhelix	i've been happy so far, but i'm listening	21:13
	(context: i use this for vim plugins)	21:13
SethRobertson	Oh, I think I have already promoted gitslave to you (or someone else who is using git with vim plugins)	21:14
tangledhelix	must have been someone else	21:14
SethRobertson	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.	21:14
	For you it would trade being able to quickly and trivially get up to the top of all of the vim plugins with having an add-on command	21:15
tangledhelix	i'll bookmark that and check it out. in this case though, these are other people's code mostly, not stuff i develop on	21:15
	← cstephens4 left	21:16
	→ mmc joined	21:16
tangledhelix	SethRobertson: the only reason i keep them as submodules is so i can fetch changes from the master if the author updates	21:16
	SethRobertson: but gitslave may come in handy on other projects. thanks! also, git rm did exactly what i needed	21:16
SethRobertson	Right, so you would be able to say `gits pull` and have everything come up to the latest commit	21:16
tangledhelix	ah ok. that does sound useful	21:17
	→ przemoc joined	21:19
	→ Luyang joined	21:20
	→ chrischambers joined	21:21
	→ bcardarella joined	21:21
	← joel_ left	21:21
	← mmc left	21:22
	→ icwiener_ joined	21:22
	← proycon left	21:23
	→ HG` joined	21:23
	→ drizzd_ joined	21:24
	→ lightcap joined	21:24
	→ juanjoc joined	21:24
	← icwiener left	21:24
	← icwiener_ left	21:26
	→ monkeytail joined	21:26
monkeytail	is it possible to clone only a specific file from a github project	21:27
	→ stringlau joined	21:27
	← LRN left	21:27
bdesk	monkeytail: i think git is not really file-oriented	21:27
	→ foo joined	21:28
monkeytail	bdesk: how would you handle it?	21:28
	← dreiss left	21:29
foo	I am trying to remove a dir so git doesn't track it with git reset the/dir/not/to/track - but I get this error: fatal: Failed to resolve 'HEAD' as a valid ref. - hmm, thoughts? I just did git add . - do I need to git commit -a -m ?	21:29
monkeytail	clone the entire rep and delete the not wanted dirs and fiels?	21:29
	→ itamarjp joined	21:29
bdesk	monkeytail: yeah i would clone the repo and then just pick out the file i want	21:29
cehteh	monkeytail: clone entire repo but only check that file out	21:29
	ChanServ set mode: +v	21:29
stringlau	I cloned from git using plain git:// command, now I want to push something so I need ssh authentication (I already have an account in the remote git), is there a way to reconfigure my local git folder so it behaves as if I cloned it using ssh://username@... ?	21:29
cehteh	you can do a shallow clone if thats ok for you	21:30
	→ Warning- joined	21:30
monkeytail	cehteh: ah.. i'll try that	21:30
	→ boombatower joined	21:30
	← boombatower left	21:30
	→ boombatower joined	21:30
	→ LRN joined	21:31
HrznDefeated	charon: still working on that output.	21:31
Arrowmaster	stringlau: just edit .git/config and change it	21:32
HrznDefeated	charon: So general question-- is git-svn usually a one shot run (run once and the whole repo smoothly comes in?)? Be it right or wrong it appears that I do a "git svn fetch" and then if goes through several SVN revs and then quits. I then run again ("git svn fetch") and it does several more revs and quits.	21:33
	and I can repeat this process several times	21:33
charon	HrznDefeated: and that's with git 1.7.x?	21:33
	i dimly recall some bugs along those lines	21:33
	but yes it's supposed to do it all in one go	21:34
	← Warning- left	21:34
HrznDefeated	Yeah I'm running version 1.7.3.1.msysgit.0	21:34
	→ tewecske joined	21:35
HrznDefeated	One go- that's what I thought... hmmm	21:35
Luyang	HrznDefeated: how to update Git for windows (!=msysgit)	21:35
HrznDefeated	huh?	21:35
	← dark left	21:35
HrznDefeated	I should not use msysgit?	21:35
	→ mmc joined	21:35
	← monkeytail left	21:35
charon	you absolutely should. (well, you should be using ordinary git on a real OS, but that's not a useful answer most of the time)	21:36
HrznDefeated	What's the alternative? cygwin is painful	21:36
bdesk	lunix	21:36
HrznDefeated	charon: Yeah my hands are tied	21:36
Luyang	wow git svn initial clone really takes time	21:37
mmattice	it pulls every revision, yeah, it takes a while	21:37
charon	HrznDefeated: then again (as i even noted in my little writeup) i heard rumours of windows-only problems, and then there's the performance issue, so even linux in a VM may be better than straight msysgit for large git-svn operations	21:37
HrznDefeated	charon: Hmm...	21:38
	→ mrico joined	21:38
HrznDefeated	Well I don't have a linux box handy but I'll think about that for a minute.	21:39
	On another note:	21:39
	← bluegene left	21:39
	← Luyang left	21:40
HrznDefeated	Our svn repo is a bit messy. Roughly 7500 commits and there are several projects (8-10 at least) in the one repo. Is there a recommended way to migrate? I mean should I be looking to create a git repo for each project? If so how do I make several git repos out of one svn...	21:40
	→ alberto56 joined	21:40
Arrowmaster	hehe charon knows all about that kinda mess	21:41
SethRobertson	It would be best if you split projects into their own repo, I think the concensus would be	21:41
	→ bluegene joined	21:41
HrznDefeated	Makes sense as an end result.. but how do I get there?	21:41
	← alberto56 left	21:42
SethRobertson	You could convert to git and then (assuming each project was in its own directory tree) use git-filter-tree on copies to remove the other projects	21:42
	← jtaby left	21:42
SethRobertson	Are the projects related? Meaning are there dependencies on each other?	21:42
HrznDefeated	I see. Create several clones and then filter each clone so that it becomes specialized?	21:42
	→ jtaby joined	21:42
SethRobertson	HrznDefeated: Yes	21:42
	← linusoleander left	21:43
Arrowmaster	you can also run git-svn multiple times and point it to different locations on the svn repo	21:43
	← Fullmoon left	21:43
Arrowmaster	multiple times i mean in multiple repos	21:44
HrznDefeated	Seth- for the most part no. Except there are a couple of tangled projects for example we have a product that has a 2d and a 3d version. They share a bunch of code but then have exclusive code. In svn one is branch.. but which one is called the master in an individual repo??	21:44
SethRobertson	Even with the bizarre tag/branch/head directory tree of SVN? I didn't know that	21:44
	→ Squarism joined	21:44
	← mrico left	21:44
SethRobertson	Oh, so the different projects are on different branches, not just different directory trees? That is...more problematic.	21:45
HrznDefeated	Arrow- so I would then run git-svn once for each individual project to create its own git repo directly from the svn?	21:45
	← LRN left	21:45
SethRobertson	No, now that I think of it, untangling that in git isn't hard after all. You just make two clones as before, delete the unwanted branch, and if necessary rename the left-over branch back to master.	21:46
HrznDefeated	Seth- for example in the svn, One project is called VC2D and is located at /trunk/VC2D and the other is called VC3D and branched from VC2D and is located at /trunk/VC3D	21:46
	→ LRN joined	21:46
	← dvaske left	21:46
	→ kvonh joined	21:47
SethRobertson	That is fine, then. Unless you want to preserve the branched relationship so that fixes in one can be merged back and forth	21:47
HrznDefeated	We do want to preserve that	21:47
	Since they are both in active development	21:47
SethRobertson	Then you do not want to split those into different repos	21:47
HrznDefeated	And have common code	21:47
SethRobertson	The common code could, in theory, be a third repo and then use submodules or gitslave to create a superproject	21:48
HrznDefeated	Ok so, same repo. Now what about the master/branch business?	21:48
	→ yhager_ joined	21:48
HrznDefeated	Good point but it probably isn't modular enough for that...	21:48
	→ Pausenbrot joined	21:48
	← ablemike left	21:48
SethRobertson	What about it? the /trunk/ thing doesn't imply branches	21:49
	→ andrewwatts_ joined	21:49
HrznDefeated	Right so in the new git repo-- VC2D would be the master and VC3D would be a branch? Even though both are under active development and being released periodically?	21:49
SethRobertson	No, they would be different directories, just like they are now. Could you convert them so that they were on different branches? Umm. I have to believe so, but it would be a bit tricky	21:50
	← andrewwatts left	21:50
	andrewwatts_ → andrewwatts	21:50
	← bdesk left	21:51
HrznDefeated	In SVN we regularly merge between the two branches (VC2D and VC3D)	21:51
SethRobertson	If they are branches now they will remain branches when converted to git.	21:52
	But /trunk/ doesn't imply branches. /branches/ implies branches	21:53
	← sh1mmer left	21:53
HrznDefeated	Yes that's confusing	21:53
SethRobertson	If you are using svn to pretend they are branches are merge different directories to each other, git doesn't do that.	21:53
	If you are using a non-standard disk layout, well your conversion task may be challenging	21:53
HrznDefeated	So I guess what we did was this... In SVN we had /trunk/VC2D and we branched and put the branch at /trunk/VC3D perhaps reasoning that this was a new offshoot that would be shipping and would eventually replace the VC2D.	21:55
SethRobertson	However, if they are currently not branches and you want to turn them into branches, there is a way forward. However, the common code would have to exist on both branches.	21:55
HrznDefeated	So we should have put that branch at /branches/VC3D instead?	21:55
SethRobertson	That is the "standard" subversion naming scheme, yes.	21:55
	But as I said, if you cannot fix this post-facto in subversion, you can fix it in git with sufficient effort	21:56
troutwine	My new gitolite installation is prompting for a password post gl-setup, but none of the suggestions in the troubleshooting guide help. Would someone be so kind as to debug this issue with me?	21:57
HrznDefeated	SethRobertson- yes I see that was an error. So we will have to work around that I guess	21:57
SethRobertson	However, the trick is going to be to ensure that the two subversion "branches" are full merged before you convert to git, so that we can instruct git to forge a fake merge so that future changes will be from that base	21:57
kevlarman	troutwine: make sure the permissions on .ssh are correct	21:57
troutwine	kevlarman: They are, on both server and client. "drwx------"	21:58
	→ Schmallon joined	21:58
HrznDefeated	Seth- you would not recommend migrating before pending merges are completed? Or just in this special case of trunk confusion?	21:59
	← Pausenbrot left	21:59
	← juanjoc left	22:00
SethRobertson	Well, certainly with "classic" subversion I would strongly suggest finishing any merges due to its inability to record merges properly and thus not being able to tell git about the merges. Whether more recent subversions (e.g. 1.5 and newer) that have better merge support record sufficient information I cannot say.	22:00
troutwine	kevlarman: That was covered in the troubleshooting document: http://goo.gl/rzLey	22:01
SethRobertson	The safest option is to finish the merge, certainly, or at least to do some test conversions to see how it works. You should do test conversions no matter what, and try to script the entire conversion process	22:01
	→ otzi joined	22:02
	← mfacenet left	22:02
kevlarman	troutwine: see if ssh's logs have anything to say about why it's rejecting your key	22:02
	← Schmallon left	22:02
Arrowmaster	use ssh -vvv and see if its even trying to use your key	22:03
	→ TheMoonMaster joined	22:03
	← kar8nga left	22:03
	→ Cogito joined	22:03
HrznDefeated	Seth- Thanks for the advice. Unfortunately we may be months out from finishing the merges	22:03
	What exactly needs scripting in the conversion process?	22:04
	troutwine is getting a lot of ssh login attempts. :/	22:05
troutwine	kevlarman: Ah, the account is locked.	22:05
	Arrowmaster: It is.	22:05
	D'oh.	22:05
	→ nkuttler joined	22:05
	→ chessguy joined	22:09
	→ martinciu joined	22:09
	← fredmorcos left	22:09
troutwine	Well, I unlocked the git account. Still no luck and nothing in auth.log to go on.	22:09
	→ v0n joined	22:09
HrznDefeated	Seth: Any thoughts on what needs scripting (as you mentioned) in the SVN -> Git process?	22:10
nkuttler	when I "git svn init" + "git svn fetch -r foo" I get "fatal: refs/remotes/trunk: not a valid SHA1". any clues? when i fetch witout specifying a revision there is no problem (except that the fetch takes hours). v1.7.3.3	22:11
troutwine	kevlarman, Arrowmaster: The gitolite user clearly has my public key in its authorized_keys.	22:12
kevlarman	troutwine: and you're trying to run a command that authorized_keys allows?	22:12
	← martinciu left	22:14
	→ ahammond joined	22:14
troutwine	kevlarman: I assume. The authorized_keys were created by gitolite, but the command is gl-auth-command and I'm attempting to clone the admin respository as suggested in the installation instructions.	22:14
	ChanServ set mode: +v	22:14
	→ martinciu joined	22:15
ahammond	hi, I have some changes in master that I shelved and would like to move to a branch "serial", how do I do that please?	22:15
kevlarman	troutwine: try ssh -vvv gl-auth-command as Arrowmaster suggested	22:15
	← yhager_ left	22:16
SethRobertson	HrznDefeated: The entire process from the initial conversion to the clones, the `git filter-tree`s, etc.	22:16
	Do it manually writing down each step and then retry from scratch once you did it manually using the script	22:16
troutwine	kevlarman: https://gist.github.com/736918	22:17
	→ bdiu joined	22:17
	→ ThiefMaster joined	22:17
HrznDefeated	Seth: So the main script value being in the filter portion (after everything is in git)	22:17
kevlarman	troutwine: well there you go	22:17
	your key isn't in openssh format	22:17
nkuttler	hm. the revision i tried to fetch was a git svn tag. i can checkout the previous revision without problems. go figure..	22:18
troutwine	kevlarman: Oh, you're right. My eyes skipped that. That makes, though. I'm using that same key to login to the system as my own user.	22:18
	Rather, "That makes no sense, though. I'm..."	22:19
Arrowmaster	are you using windows?	22:19
troutwine	Linux.	22:19
	← vitor-br left	22:19
	← martinciu left	22:19
troutwine	I scp'ed my ~/.ssh/id_rsa.pub over to the system's /tmp when installing gitolite, as per the instructions.	22:20
	← kvonh left	22:20
	→ eletuchy joined	22:20
	← bdiu left	22:20
	→ pratz_ joined	22:20
kevlarman	troutwine: ssh is complaining about your local key	22:21
	not the remote one	22:21
troutwine	Yeah. I'll generate a new one.	22:21
kevlarman	sshd refuses to give any information about why it's failing as a security feature	22:21
troutwine	Makes sense.	22:22
	→ martinciu joined	22:22
troutwine	How could I have created a non-openssh key using openssh's ssh-keygen?	22:22
ahammond	how do I move a stash between branches?	22:23
kevlarman	ahammond: apply it on top of your new branch	22:23
ahammond	kevlarman: so I do git checkout newbranch; git stash pop?	22:23
kevlarman	yeah more or less	22:23
	i'd use apply to keep it around until you make sure you got the results you expected	22:24
khaije\|amalt	a couple questions, if i commit a file from outside of the working directory, what path does it have? also is it possible to specify the path a/o the branch it should commit to?	22:24
ahammond	kevlarman: thanks!	22:24
	→ vitor-br joined	22:25
Arrowmaster	khaije\|amalt: you cant commit a file outside the working directory	22:25
khaije\|amalt	Arrowmaster: git --git-dir path/.git --bare commit -m "from anywhere"	22:25
	→ paradigm joined	22:26
Arrowmaster	try it and see	22:26
	i bet it blows up spectacularly	22:27
	← HG` left	22:27
ahammond	kevlarman: I'm getting "no valid stashed state found". The stash seems to be tied to the branch.	22:27
khaije\|amalt	? i just did, there wasn't any apparent problem	22:27
Arrowmaster	well you only commited what was already in the index, did you try adding to the index like that?	22:28
	→ sonnym joined	22:29
SethRobertson	HrznDefeated: The main benefit is repeatability without errors, so that you can do test conversions to fine tune everything and see if, for example, you need to finish up the merges before you do the real conversion.	22:29
khaije\|amalt	Arrowmaster: yup, same prefix with an : add somefile	22:30
troutwine	kevlarman: I generated a new ssh key. Still no dice.	22:31
	→ eijk_ joined	22:31
	→ yhager_ joined	22:31
	→ coderdad joined	22:32
eijk_	Hi! How do I configure a certain branch to always allow forced pushes to origin?	22:32
troutwine	kevlarman: Again, not an RSA1 keyfile. Ergh.	22:32
	← lightcap left	22:34
	→ hyperair joined	22:34
	→ matthijs joined	22:34
kevlarman	troutwine: pastebin the public half	22:34
	← Saviq left	22:35
	← yhager_ left	22:35
troutwine	kevlarman: https://gist.github.com/736918#file_id_rsa.pub	22:35
	→ fredmorcos joined	22:35
	→ sh1mmer joined	22:36
kevlarman	troutwine: and id_rsa looks similar?	22:36
	(except about 4x as long)	22:36
troutwine	kevlarman: Yep.	22:36
	And, again, I _can_ use the above public key to login to my user on the same system. No problem.	22:37
kevlarman	any idea where the -----BEGIN comes from?	22:37
troutwine	kevlarman: None.	22:38
	→ strixv joined	22:40
troutwine	kevlarman: Okay, well, thank you for the help. I have to be off, I'm afraid.	22:40
	← troutwine left	22:40
	← martinciu left	22:40
charon	HrznDefeated: sorry, was away for a while. do you still need help?	22:41
	← Squarism left	22:42
HrznDefeated	charon: Yes but I'm not sure what question to ask just yet. Thanks	22:44
	→ Pupeno joined	22:44
	→ redondos joined	22:44
	← doener left	22:44
SethRobertson	inkedmn: ping?	22:46
ahammond	kevlarman: git checkout -b new_branch solved my problem	22:47
	← TheMoonMaster left	22:47
	→ martinciu joined	22:48
HrznDefeated	Seth: I understand about the value of repeatability with the scripting. It seems like most of the scripting would be massaging the repo once it's already in git, is that correct? Or can the script be of value in the "git svn fetch" portion?	22:48
SethRobertson	HrznDefeated: Oh sure, for instance, if subsequent svn activities occur before the conversion process is completely specified. eg. if you decide to finish the merge process before final conversion would be one example	22:50
	← martinciu left	22:51
SethRobertson	But it is of course your code.	22:52
	→ martinciu joined	22:52
	→ Mage joined	22:54
	→ Tianon joined	22:55
	← Tianon left	22:55
	→ Tianon joined	22:55
	→ kpreid joined	22:58
	→ ivan joined	22:58
	← bcardarella left	22:59
	→ dreiss joined	23:00
	→ nadavoid joined	23:00
	← Tianon left	23:01
	← eletuchy left	23:01
	→ Exposure joined	23:01
	→ Tianon joined	23:02
	← Tianon left	23:02
	→ Tianon joined	23:02
	→ ft joined	23:04
	→ ReinH joined	23:05
	← HrznDefeated left	23:08
	→ HrznDefeated joined	23:08
	→ alberto56 joined	23:09
	← bhosie left	23:09
	→ bdiu joined	23:09
	→ g0 joined	23:09
	→ Linc joined	23:10
	→ gilaniali joined	23:12
	← pantsman left	23:13
	← mmw left	23:14
	← shentino left	23:14
	← thiago_home left	23:14
	← SimManiac left	23:16
	aah → him	23:17
	him → us	23:18
	us → dougbarn	23:19
	dougbarn → aah	23:19
	← bdiu left	23:20
	→ quadz joined	23:20
	← matthijs left	23:21
	← mmc left	23:24
	← shennyg left	23:25
	→ tchan joined	23:25
	→ zombor joined	23:27
	→ bdiu joined	23:28
	← nicoulaj left	23:28
	→ strk joined	23:30
	→ nd___ joined	23:30
strk	so, we found git clean not removing dirs with no write permissions	23:31
	that's with -xdf flags	23:31
	any idea how to force force force ?	23:31
	← MattDiPasquale left	23:31
	← g0 left	23:31
bremner	strk: fix your permissions?	23:31
	→ mmc joined	23:32
bremner	e.g. with chmod -R on unixy things	23:32
strk	thanks :/	23:32
	I know how to fix permissions	23:32
bremner	you wouldn't believe how many people dont :)	23:32
	← Heimidal left	23:32
bremner	so what's your proposal, that get sets the write permission on the containing directory?	23:33
	s/get/git/	23:33
	← jhaddad left	23:35
	→ jhaddad_ joined	23:35
	← jhaddad_ left	23:35
	→ jhaddad joined	23:36
	← Cogito left	23:36
	← jhaddad left	23:38
	→ jhaddad joined	23:38
	← g0bl1n left	23:40
	← jrmuizel left	23:40
	← jhaddad left	23:40
	→ jhaddad joined	23:41
	→ aresnick joined	23:43
	← gilaniali left	23:44
	ChanServ set mode: +v	23:48
HrznDefeated	hmm	23:48
	→ carutsu_ joined	23:49
	→ joshsmith joined	23:50
	→ eletuchy joined	23:50
	← rsewill left	23:51
	→ wipple joined	23:52
	→ guardian joined	23:53
	← MikhX left	23:54
	← eletuchy left	23:54
	← bdiu left	23:58
	→ tuxillo joined	23:59
tuxillo	hi	23:59
	← GodEater left	23:59
tuxillo	how do I see the commits introduced by a merge?	23:59

Logs Search ←Prev date Next date→

Channels Documentation