IRCloggy #git 2009-07-03

Provider of IRC logs since 2005.

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

	2009-07-03
	← neoeinstein left	00:00
	← majd left	00:02
	mtvee → mtvee\|away	00:02
	← johndoigiii left	00:03
	SwistakTheJester → SwisOff	00:04
	→ johndoigiii joined	00:05
	← joevandyk_ left	00:06
	→ joevandyk_ joined	00:10
	mtvee\|away → mtvee	00:11
	← rpag left	00:11
	→ sitaram joined	00:14
	← whee left	00:15
	← sergiodj left	00:15
jjuran	If I'm not mistaken, 5 hex digits won't yield a collision untl about a half million objects, on average. With 6 digits it's 8 million. With 8 digits, 2 billion.	00:15
	← ezyang left	00:16
	← hpoydar left	00:16
	makies → makies_bot	00:17
	← ericindc left	00:20
tomoj	don't you have to use birthday problem math there?	00:21
	→ lightcap joined	00:21
jjuran	tomoj: Am I supposed to divide the exponent by two instead of the number of keys?	00:23
tomoj	no, I think it's more complicated	00:23
	http://en.wikipedia.org/wiki/Birthday_attack	00:23
	← rushed left	00:23
tomoj	based on that approximation I get 1283 objects on average for 5 hex digits	00:23
	gonna run an experiment to test it	00:23
jjuran	Oops. IANAC.	00:24
tomoj	hmm.. I may be thinking about it wrongly	00:25
	→ whee joined	00:25
tomoj	if you just generate random 5-digit hex strings, it will only take you around 1283 to get a collision.	00:26
johndoigiii	lol	00:26
	sorry to set this off	00:26
	← ereslibre left	00:26
	← struberg left	00:28
	← qhoxie left	00:29
	→ standel joined	00:29
	← standel left	00:30
	→ alexyk_ joined	00:30
tomoj	but assuming sha1 is perfectly distributed, the first 5 digits of every hash IS a random 5-digit hex string, right?	00:30
	← alexyk left	00:30
jjuran	5 hex digits has 2^20 combinations, yielding 2^10 bits of security after dividing by 2. Where does the 1.25 coefficient come from?	00:31
tomoj	I was using the sqrt(pi/2 * n) approximation	00:32
	→ qhoxie joined	00:33
	← reprore left	00:34
	→ rpag joined	00:34
	← whee left	00:34
jjuran	Okay, so my revised numbers are a thousand objects for 5 digits, and each extra digit buys you a factor of four.	00:34
	→ JasonWoof joined	00:35
	← mithro left	00:35
	→ mithro joined	00:36
	← mw___ left	00:36
	← eldragon left	00:37
	← qhoxie left	00:37
	→ utu joined	00:37
tomoj	yup	00:37
	so less than 100000 objects for 8 digits	00:38
	→ neoeinstein joined	00:41
	← tmm1 left	00:42
johndoigiii	hmm, whats the easiest way to dump the commit id into a file? I'd like to test this out for a real project	00:42
	sorry, all of the id's in the log	00:42
tomoj	johndoigiii: maybe something like: git log --all --pretty=format:"%H" > ids.txt	00:46
	→ alexyk joined	00:47
johndoigiii	tomoj: git rev-list	00:47
	thats what i am using	00:47
tomoj	oh yeah, that's easier :)	00:47
	I guess you can't tell rev-list to list ALL the commits?	00:48
johndoigiii	--a	00:48
	all*	00:48
tomoj	well, I don't think that will list unreachable commits	00:48
	← alexyk_ left	00:49
johndoigiii	yeah	00:49
tomoj	but I guess those will be gone soon anyway and so don't really matter	00:49
johndoigiii	only 11256 commits for rails.git	00:49
	i dont think thats right	00:49
	→ khmarbaise joined	00:50
	← kennethlove left	00:53
	← d0k left	00:54
	← parasti left	00:55
	← klbate left	00:55
tomoj	yeah I got the same result	00:56
	← whaley left	00:56
	← WebDragon\|away left	00:57
	← WALoeIII left	01:00
johndoigiii	this is for rails.git http://dpaste.com/62708/	01:01
	with a rev list of 11,256	01:01
tomoj	7 digits is supposed to be good for 20000-ish on average	01:02
	but it's just average, so that sounds right	01:02
	← sitaram left	01:02
	← neunon left	01:04
	← __iron left	01:06
tomoj	git.git has 21287 commits with no duplicates for 7 digits	01:06
	→ altrux joined	01:08
	← lightcap left	01:08
	→ hpoydar joined	01:08
	makies_bot → makies	01:09
mugwump	use the birthday calculator	01:09
	http://my.instacalc.com/calc/91ca9ae67a434f82be1b84d85ea9e065	01:09
	eg, put in (16^7) for total_items	01:10
	← tongueroo left	01:10
	→ agenteo joined	01:10
	→ ericindc joined	01:10
tomoj	it says 19291 needed for 50% probability	01:11
	→ whee joined	01:11
mugwump	and then for something really spooky put in 57% for probability_of_match ...	01:11
tomoj	haha	01:11
	← toxx_ left	01:11
tomoj	is "needed for 50% probability" the same as "expected number you have to have to get a collision"?	01:12
mugwump	sure, you can read it like that	01:13
	it's more like "given X repositories that that many commits, 50% of them will have a collision"	01:13
tomoj	ah, I see	01:13
	← Bass10 left	01:13
mugwump	reduce the probability if you want to be more sure	01:14
tomoj	so rails with 11256 has about a 21% chance of needing 8+ digits	01:15
	← bcardarella left	01:18
	→ alexyk_ joined	01:19
	→ Marmouset joined	01:19
mugwump	er ... "4 out of 5 rails-sized projects will need 8+ digits"	01:20
	1 out of 5 I mean	01:20
johndoigiii	very cool	01:20
mugwump	I'm surprised it only has 11k commits tbh	01:21
	→ bcardarella joined	01:22
	← johndoigiii left	01:22
	→ thoraxe joined	01:22
	← sgrimm left	01:23
	← _Vi left	01:25
	← slyphon left	01:25
	← Superbaloo left	01:26
tomoj	linus's kernel tree needs 9 digits :O	01:26
	← binjured left	01:27
	← alexyk left	01:28
	→ jackdempsey joined	01:30
	→ jksz joined	01:31
	→ mw joined	01:34
	→ WALoeIII joined	01:35
	→ spaceonline joined	01:35
	← WALoeIII left	01:35
	← Chillance left	01:36
	→ jksz_ joined	01:38
	← joevandyk_ left	01:38
	← jksz left	01:38
	← icwiener left	01:38
	← Alinermedia left	01:39
	← alexyk_ left	01:42
	→ brizly1 joined	01:45
	← Octalot left	01:46
	← kinkerl left	01:47
	→ ehc joined	01:48
	← mw left	01:49
	← ehc left	01:49
	→ ehc joined	01:50
	→ hpoydar_ joined	01:52
	← hpoydar left	01:52
	mtvee → mtvee\|away	01:54
	mtvee\|away → mtvee	01:55
	← dreiss left	01:55
	→ flippo joined	01:56
	← aziz left	01:56
	← javatexan left	01:59
	→ bugfux joined	02:00
	← brizly left	02:01
	← psynaptic left	02:02
	← bugfux left	02:04
	← thoraxe left	02:04
	← utu left	02:05
	→ vals_ joined	02:06
	← jjuran left	02:07
	→ javatexan joined	02:09
loincloth	hey	02:09
	so	02:09
	i said "git push" and it spit out all that stuff about how i didin't specify where, etc	02:10
	but then it hung..	02:10
	and i wasn't sure what it was doing..	02:10
	so i killed it, but i did so JUST as it spit out info like it was actually doing a push	02:10
	→ blkf joined	02:10
loincloth	wait	02:10
	← loincloth left	02:10
flippo	That was special.	02:12
	lut4rp_AFK → lut4rp	02:13
	← univate_ left	02:14
	guitsaru → guitsaru\|away	02:14
	← flaguy48 left	02:14
	→ univate_ joined	02:15
	← kuadrosx left	02:17
	→ kuadrosx joined	02:17
	← whee left	02:18
	← tango_ left	02:21
	← khmarbaise left	02:21
	← krawek left	02:21
	→ krawek joined	02:22
	→ jeffrey__ joined	02:23
jeffrey__	Is there any easier way to generate the difference of two commit in Linux kernel source code than "git diff 42ca4fb69126dd9d0e25112edca1cacf846aa5c3 43644679a1e80f53e6e0155ab75b1093ba3c0365", since that the revision number is so long in Linux kernel source code.	02:23
	→ joevandyk joined	02:23
	← rhett left	02:24
	→ followumesh joined	02:26
followumesh	Hi	02:26
	← hpoydar_ left	02:26
	→ juanjoc joined	02:27
followumesh	Anyone knows how to jump between commits, like we do with --hard reset? thanks	02:27
	← spaceonline left	02:28
	→ rhett joined	02:28
	← joevandyk left	02:29
	→ tongueroo joined	02:29
	← followumesh left	02:30
	← rhett left	02:30
	← ehc left	02:30
	→ umesh joined	02:31
	← schambers left	02:32
	← JasonWoof left	02:32
smoofra	so aparetnly there's a cache of objects somewhere. anybody know wehre to find it?	02:33
	← Derander left	02:33
umesh	Hi, anybody knows how to jump to older commits?	02:34
	← sekimura left	02:34
kpreid_	jeffrey__: you can write any unambiguous prefix of those hashes, does that help? (e.g. perhaps git diff 42ca4 43644). any further refinement would require more information about where you're getting them from	02:34
	→ schambers joined	02:37
	← mnick left	02:37
	→ ben_h joined	02:38
jeffrey__	kpreid_: thank you. I tried, and it works with at least 7 digit prefix.	02:38
	→ sergiodj joined	02:39
	← javatexan left	02:41
	← neoeinstein left	02:42
	→ neoeinstein joined	02:42
	→ javatexan joined	02:43
	→ bugfux joined	02:44
	→ harinath joined	02:46
	→ mw___ joined	02:50
	→ tjafk2 joined	02:53
smoofra	hrm anybody know the diffirence between the SEEN and SHOWN flags?	02:53
	→ dreiss joined	02:54
	← blkf left	02:54
	← bugfux left	02:55
	→ bugfux joined	02:57
kanzure	so, can you remove something from a bare repository?	02:58
	← mw___ left	02:58
	→ alexyk joined	02:58
umesh	anybody knows how to older commits, AFAIK, we can do --hard reset, and ORIG_HEAD keeps the previous head of the branch, so if we reset multiple times we lose the track of the recent commit.	02:59
kanzure	I committed something stupid	02:59
	:(	02:59
umesh	*how to jump	02:59
	← MetaCosm left	02:59
	← drewr left	03:02
	← kukks left	03:02
	→ joevandyk joined	03:05
	lut4rp → lut4rp_AFK	03:06
	→ sitaram joined	03:07
	← tjafk1 left	03:10
	← fujin left	03:11
	← Strogg left	03:12
tomoj	umesh: if your problem is losing track of stuff: git reflog	03:12
	→ Strogg joined	03:13
kanzure	is there a way to rewind a bare repo?	03:17
	guitsaru\|away → guitsaru	03:18
	← djgera left	03:19
	← guitsaru left	03:21
offby1	what does "rewind" mean?	03:22
	→ kennethlove joined	03:26
	→ WebDragon joined	03:27
	→ bugfux_ joined	03:30
	← bugfux left	03:31
sitaram	you mean reset to a former state? not easily, unless you had the foresight to set core.logallrefupdates	03:32
	← marvil07 left	03:32
sitaram	but I probably misunderstood your question anyway	03:32
	→ sfwc joined	03:39
	→ hydrogen joined	03:40
	→ girishr joined	03:42
	← kennethlove left	03:44
	→ aspotashev\|eeepc joined	03:44
	← Mage_Dude left	03:46
	← schambers left	03:46
	→ Chouser joined	03:48
	← bugfux_ left	03:48
	→ bugfux joined	03:49
Chouser	is it bad for the same commit in two different branchs to have slightly different commit messages?	03:49
	→ Mage_Dude joined	03:49
offby1	if the messages are different, then they aren't the same commit.	03:50
Chouser	well, they're not the same commit anyway because if the branches are different, right?	03:50
	→ sergeykish joined	03:50
	→ schambers joined	03:51
offby1	wrong	03:51
	:)	03:51
Chouser	:-(	03:51
offby1	a "branch" is just a reference to a commit.	03:51
	You can have as many references to a given commit as you want.	03:51
	offby1 strongly recommends "Git For Computer Scientists"	03:51
offby1	'splains it all	03:51
	→ ankit9 joined	03:51
Chouser	ok, but isn't a commit a snapshot of the whole tree, not just a diff?	03:51
offby1	correct.	03:51
	it's a snapshot of the tree _plus_ a comment and author _plus_ pointers to other commits.	03:52
Chouser	so if the branches have any difference in their history (as you might expect of a branch), then the same patch applied to two branchs are different commits anyway, even if the message is identical	03:52
offby1	the other commits being its parents. There can be any number of parents, including 0; 1 is the most common number; 2 is pretty common; 3 or more is possible but I don't think I've ever actually seen it	03:52
	yes, the same patch applied to different branches yields different commits.	03:53
	easy enough to try that out as an experiment in a toy git repository.	03:53
Chouser	ok, so I phrased my original question incorrectly. :-)	03:53
	is it bad for the same patch applied to two different branchs to have slightly different commit messages?	03:53
offby1	even if the different branches happened to have the same tree -- if the branches indeed reference different commits, then the same patch applied to both will yield two different commits still.	03:53
	oh.	03:54
	ok, answering your question: I don't see why it would be bad.	03:54
Chouser	ok, great.	03:54
offby1	commit messages are up to you; use whatever you think will be clearest.	03:54
Chouser	I know git itself doesn't care, but I didn't know if other surrounding tools might	03:54
offby1	I _will_ say I can't offhand think of _why_ you'd want the message to vary.	03:54
	other tools won't care either.	03:54
	Now, us humans: _we_ might care :)	03:54
	offby1 poke sitaram	03:55
	→ w\|shimmy joined	03:55
offby1	surely you have an opinion?	03:55
	s/poke/pokes/	03:55
Chouser	in this case, we plan to have 'release' and 'bugfix' branchs	03:55
offby1	sure	03:55
sergeykish	Hello, I get "You must edit all merge conflicts and then ..." on rebase. But there is no unmerged files (git checkout HEAD^ file). What can I do?	03:55
Chouser	when a bug is reported that applies to both, it gets two tickets	03:55
	one for each branch	03:55
offby1	ah	03:55
	→ whaley_ joined	03:55
offby1	that would do it, wouldn't it.	03:55
	Chouser: I wouldn't worry.	03:55
Chouser	the bulk of the message in each case can be the same, but I want the ticket number to refer to the approriate ticket.	03:56
	offby1: ok, great, that's what I was hoping to hear.	03:56
	offby1: thank you very much!	03:56
	offby1 cackles evilly	03:56
Chouser	now I just have to write a tool so I don't have to manually edit that message each time :-P	03:57
	→ MikeStone joined	03:57
	→ jrmuizel joined	03:57
sergeykish	Wow, it fixed while I search answer	03:58
	What can it be?	03:58
offby1	Chouser: typically, people apply the patch to just one branch, and then merge that branch into the other.	03:58
	Or they might cherry-pick	03:58
Chouser	I plan to do something like a cherry-pick	03:58
sergeykish	And get again same thing	03:58
Chouser	...except I need to adjust the message on its way through	03:59
	it was the awkwardness of having to adjust that message that made me wonder if I was doing somethign vile	03:59
	sitaram just saw offby1's poke, is sorry he was away	04:00
offby1	you could cherry-pick and then "git commit --amend" to tweak the message	04:00
	sitaram: shameful	04:00
	→ majd joined	04:00
Chouser	offby1: yeah, I'm going to do something like that, but I don't want to edit the message manually.	04:00
offby1	should be doable	04:01
	← ankit9 left	04:01
Chouser	that is, I think I'll have the ticket number fixed automatically before popping up an editor so I can confirm it didn't screw up the regex	04:01
	that is, that the regex didn't screw up the message.	04:01
	→ ankit9 joined	04:01
sitaram	offby1: yes isn't it -- old men have no attention span these days!	04:01
	← ankit9 left	04:01
	← sjl left	04:02
	← scarabx left	04:02
offby1	probably making dinner for your family or some other old-man thing	04:03
	or earning money!	04:03
	Squaaaaare!	04:03
sitaram	if I made dinner, they'd all be dead by now heh!	04:04
Chouser	offby1: ok, thanks for your help. Bedtime.	04:05
	← wshimmy left	04:05
	← majd left	04:06
	→ SjB joined	04:06
	→ justin-george joined	04:06
	← jrmuizel left	04:07
	offby1 nods	04:07
offby1	given that .in is far away, /me guesses it's breakfast or even lunchtime	04:08
	← hydrogen left	04:08
	→ majd joined	04:09
owen1	what is the best way to resolve conflict when merging branches?	04:12
	vimdiff?	04:13
offby1	I like kdiff3	04:13
	← juanjoc left	04:13
offby1	works on Windows, too	04:13
	→ upul joined	04:14
	← umesh left	04:15
	→ umesh joined	04:15
	← umesh left	04:15
owen1	vimdiff file_from_master file_from_branch how do i get the second file?	04:15
	← aspotashev\|eeepc left	04:15
offby1	dunno	04:16
	anyway it's best to mention _three_ files, not just two (the third is their common ancestor)	04:16
	owen1: try running "git mergetool"; chances are it'll Just Work™	04:16
owen1	it opened gvimdiff. how do i change it to vimdiff?	04:17
	→ hydrogen joined	04:18
sitaram	offby1: I'm at work actually, and they keep interrupting with stuff they want me to do; it's not fair!	04:21
	owen1: I like meld too	04:21
owen1	is it text-based?	04:22
sitaram	no	04:22
	you want text-based? "git mergetool -t vimdiff" should work I think	04:22
owen1	sitaram: i'll try it	04:22
	← Mage_Dude left	04:25
owen1	sitaram: works. i have 1 conflicted line. Am I suppose to delete it from one of the files and save it?	04:26
	→ j416 joined	04:29
j416	how can I get a hook to execute a script inside the repo?	04:29
	is there a variable holding the path to it, or should I use $GIT_ROOT/../ ?	04:29
owen1	sitaram: i tried to delete the line and closed vim. it asked me if the conflict was resolved, i said yes but i see a few weird files with all kind of extensions. what happened?	04:30
	→ ankit9 joined	04:31
	→ tlrobinson joined	04:31
owen1	and when i 'cat test' (this is the name of the conflicted file) i get the line with the conflict from both master and the branch.	04:31
	so i assume it failed.	04:31
	← majd left	04:34
sitaram	owen1: sorry;... will take a bit	04:38
	← upul left	04:38
	← sergeykish left	04:40
	← marc_kdab left	04:40
	→ sergeykish joined	04:41
	→ rs joined	04:42
	→ rakoth joined	04:44
	lut4rp_AFK → lut4rp	04:44
j416	hmm	04:47
	→ Lemao joined	04:47
j416	I would like to add a file using a hook, before commit	04:47
	is this possible?	04:47
	← stringo0 left	04:47
j416	I'm using git to manage a database-driven website, and I want to include dumps of the database for each commit	04:47
	s/dumps/a dump/	04:47
	i tried doing 'git add dump.sql' in pre-commit, but it doesn't seem to add the file	04:48
	ideas?	04:48
	→ scarabx joined	04:48
offby1	I'd be more prone to write a wrapper around "git commit" that simply does the add, and then commits	04:49
j416	hmm	04:49
	← rakoth left	04:50
	→ rakoth joined	04:50
offby1	back when I hung out in #svn, people would ask the equivalent question, and the answer was always either a) it's not possible for a pre-commit hook to modify the commit; or b) it's a _really bad idea_ to have the pre-commit hook modify the commit	04:51
	(forget which)	04:51
j416	ok!	04:52
	I wonder how I should wrap it... hm	04:52
	← Lemao left	04:52
offby1	oh, just a two-line shell script ought to do it	04:52
j416	yeah.. just trying to figure out where to put it	04:52
	← linoj left	04:54
offby1	j416: http://gist.github.com/139917	04:54
	untested, but you get the idea	04:54
	→ utu joined	04:55
j416	yeah... thanks	04:55
	I got that part	04:55
	rather, I was thinking of how to organise it all	04:55
	(so that I wouldn't forget to run the script instead of git commit, for instance..)	04:55
	maybe I should just dump manually	04:56
	yeah.. that seems good	04:56
	sorry :)	04:56
offby1	that "how do I remember to run the script" is indeed the trickiest part :-\|	04:56
	perhaps you could have your pre-commit hook _reject_ the commit if it doesn't find your sql dump file in it somewhere.	04:57
j416	offby1: that's clever	04:57
offby1	and zap you with a cattle prod for good measure	04:57
j416	haha	04:57
offby1	it'd only take a few times before you'd have formed the habit :)	04:57
j416	although	04:57
	I'm thinking, maybe a dump isn't needed for each and every commit	04:57
offby1	well, probably not :)	04:58
j416	I guess I could do it manually when needed instead	04:58
	it will be cleaner	04:58
offby1	that kind of thing sounds more like a backup than revision control	04:58
j416	i have written scripts for dump/restore so i just need to run those when needed	04:58
	yeh	04:58
	← dsop left	05:01
	← ankit9 left	05:03
MikeStone	I'm trying to install git locally on my new hostmonster account, and make is blowing up... I've been trying to google for a resolution to no avail... would this be the right place to seek help?	05:04
offby1	yep	05:04
	betcha you're missing some lbirary	05:05
	library	05:05
	either install it, or poke around in the Makefile to divine the magic incantation that tells make to skip that part	05:05
MikeStone	well, it's failing to make during the perl target it looks like	05:05
offby1	like, "make NO_CURL=1" or something	05:05
	hm, you probably need perl	05:05
	hard to imagine git without perl	05:05
MikeStone	well, I have perl, and I exported PERL_PATH=where it is installed	05:05
offby1	MikeStone: paste a transcript why dontcha	05:05
	also mention what OS it is	05:06
MikeStone	sure, one sec	05:06
	→ dsop joined	05:06
MikeStone	hmmm, what's the command to see the os again? I just have shell access, so not quite sure	05:06
Khisanth	uname -a :)	05:07
MikeStone	cool, thanks! hmmm.... just says GNU/Linux, 64 bit version	05:07
	perhaps just vanilla linux... not a particular distro....?	05:08
offby1	MikeStone: lsb_release -a	05:08
	maybe look in /etc/motd	05:08
MikeStone	not a command	05:09
offby1	it's a plain text file if it exists at all	05:09
MikeStone	/etc/motd not a file I have access to if it exists	05:09
	← bernd_ left	05:09
offby1	"motd" == "Message Of The Day"	05:09
	odd indeed.	05:09
MikeStone	well, it's hostmonster, they may have a custom linux build or something	05:09
offby1	"ls -l /" _might_ give a clue	05:10
MikeStone	ok, here's the make results with verbose turned on:	05:10
	make -C git-gui gitexecdir='/home6/invotati/Applications/git/libexec/git-core/\	05:10
	' all	05:10
	make[1]: Entering directory `/home6/invotati/Downloads/git/git-1.6.0.4/git-gui'	05:10
	make[1]: Leaving directory `/home6/invotati/Downloads/git/git-1.6.0.4/git-gui'	05:10
	make -C gitk-git all	05:10
	make[1]: Entering directory `/home6/invotati/Downloads/git/git-1.6.0.4/gitk-git\	05:10
	'	05:10
	make[1]: Nothing to be done for `all'.	05:10
	make[1]: Leaving directory `/home6/invotati/Downloads/git/git-1.6.0.4/gitk-git'	05:10
	make -C perl PERL_PATH='/ramdisk/bin/perl' prefix='/home6/invotati/Application\	05:10
	s/git' all	05:10
	make[1]: Entering directory `/home6/invotati/Downloads/git/git-1.6.0.4/perl'	05:10
	make[2]: Entering directory `/home6/invotati/Downloads/git/git-1.6.0.4/perl'	05:10
	make[2]: true: Command not found	05:10
	make[2]: *** [blibdirs] Error 127	05:10
	make[2]: Leaving directory `/home6/invotati/Downloads/git/git-1.6.0.4/perl'	05:10
	make[1]: *** [all] Error 2	05:10
	make[1]: Leaving directory `/home6/invotati/Downloads/git/git-1.6.0.4/perl'	05:10
offby1	wow, the web site sure looks sleazy	05:10
MikeStone	make: *** [all] Error 2	05:10
	[email@hidden.address] [~/Downloads/git/git-1.6.0.4]# uname -a	05:10
	Linux host380.hostmonster.com 2.6.28-9.15.intel.BHsmp #1 SMP Sat Apr 18 08:53:0\	05:11
	9 MDT 2009 x86_64 x86_64 x86_64 GNU/Linux	05:11
	which?	05:11
offby1	MikeStone: in the future, when we say "paste", we don't mean 'paste in the channel'; we mean 'paste on a paste site like gist.github.com'	05:11
MikeStone	ooooh, sorry	05:11
	I'm kind of a noob to irc... sorry!	05:11
offby1	.oO("true: Command not found"?!)	05:11
	offby1 smells a broken setup	05:11
	lut4rp → lut4rp_AFK	05:11
	← girishr left	05:11
MikeStone	yeah... it kinda seemed funky to me	05:11
	→ marc joined	05:12
offby1	well, if _I_ made a brain-dead distro like that, I wouldn't put _my_ name on it, either :-\|	05:12
MikeStone	I would try to figure it out myself, but I haven't touched a Makefile in years... kinda lost as to how to debug the problem	05:12
offby1	I think you need to talk to hostmonster	05:12
	→ marvil07 joined	05:12
	offby1 bets there is nobody there with a clue	05:12
	marc → marc_kdab	05:12
	→ Superbaloo joined	05:12
MikeStone	well, they seem to be pretty good so far	05:12
	some posts online indicated installing git on hostmonster would be a snap... maybe they had a host with a more standard distro or something	05:13
offby1	clearly not aimed at programmers, though	05:13
MikeStone	no, I think it is geared for people just putting junk up quickly... they have 1 click installers for the popular content management systems and blogs and whatnot	05:14
offby1	mm hmm	05:14
MikeStone	but they host rails pretty cheaply, and came recommended from a good friend	05:14
offby1	I will bet you there is a maximum of one person at that company with enough technical smarts to answer your question	05:14
MikeStone	I hope you are wrong, but I bet you are right	05:15
offby1	faq says it's CentOS	05:15
MikeStone	do you know what blibdirs is?	05:15
offby1	http://www.hostmonsterforum.com/showthread.php?t=3169	05:15
	MikeStone: blibdirs, if I recall correctly, is some directory in which perl expects to find binary something-or-others	05:15
	← scarabx left	05:16
tlrobinson	i'm looking for a way to merge a sub tree of one branch into another branch. the subtree merge seems to be merging a whole branch into a subtree of another branch, is that correct?	05:16
offby1	CentOS release 4.7 (Final) maybe	05:16
Chouser	git show --pretty=oneline <commit id> # prints a patch too	05:17
MikeStone	well, thanks for all the help, I'll figure it out eventually, I'm sure... sorry again about the paste ;-) I'll remember that for next time	05:17
Chouser	what's the right way to get just the first line of the commit message?	05:18
offby1	MikeStone: hmm, CentOS 4.7 isn't very old. It should be doable, somehow	05:18
MikeStone	yeah, I can't imagine they cripple the server THAT much...	05:18
offby1	Chouser: (didn't we discuss this already?) "git log --format=%xx" for some value of x :)	05:18
MikeStone	I wish they supported git out of the box though.... I don't want to go back to svn ;-)	05:19
Chouser	offby1: sorry if my brain is slipping, but I don't see how to get log to print just the one commit I want	05:20
offby1	Chouser: git log --format=%s SHA1 -1	05:20
	→ wsc joined	05:20
Chouser	ah, thanks	05:20
	→ jceb joined	05:21
Chouser	--pretty=format:%s	05:21
offby1	--format=%s works too in newer versions	05:21
	← Sonderblade_ left	05:22
	→ sgrimm joined	05:22
Chouser	oh, ok	05:23
	← zdennis left	05:23
	← johan-s left	05:24
	→ bernd_ joined	05:27
	← disappearedng left	05:28
sitaram	MikeStone: cat /etc/issue may tell you what distro it is	05:28
	← jackdempsey left	05:28
	sitaram wakes up	05:28
	← MikeStone left	05:30
	← cehteh left	05:32
wereHamster	good morning sitaram :)	05:32
	→ upul joined	05:35
	← SjB left	05:36
sitaram	wereHamster: good morning!	05:36
	I guess you saw Linus's reply... :)	05:37
	sitaram wonders if arafangion is around...	05:37
	→ bizhat joined	05:37
wereHamster	hm, linus's reply to what?	05:37
	← hydrogen left	05:38
wereHamster	re. nfs?	05:38
sitaram	aah... am I thinking of someone else...? there was this argument yesterday about what happens if you put a bare repo on an NFS/CIFS server and different people access it directly via mount	05:38
	yes nfs	05:39
wereHamster	reading it right now :)	05:39
sitaram	so it was good to hear that it will "just work" (modulo "git gc --prune", which is fine!)	05:39
	→ Moult joined	05:40
sitaram	and on a side note, it increasingly looks as if the cygwin port (which can do "git daemon", while msysgit cannot) can be really, really ignored. I always liked msysgit anyway, my windows users cant handle more complex installations!	05:40
	← utu left	05:41
	← j416 left	05:41
	← bizhat left	05:42
	→ giallu joined	05:42
	← sergeykish left	05:46
	→ ankit9 joined	05:46
	← Superbaloo left	05:49
	→ lightcap joined	05:53
	→ binjured joined	05:54
	← justin-george left	05:56
	← ankit9 left	05:59
	→ girishr joined	06:00
	→ ph^_ joined	06:00
	← mtvee left	06:00
	→ mtvee joined	06:00
	← stableable left	06:02
	→ girishr_ joined	06:05
	← krawek left	06:06
	SwisOff → SwistakTheJester	06:07
	→ saurabhverma joined	06:08
	← saurabhverma left	06:08
	→ Tuomas joined	06:10
	← altrux left	06:11
	→ ereslibre joined	06:11
	← marvil07 left	06:12
	→ ankit9 joined	06:13
	→ johndoigiii joined	06:13
	← ankit9 left	06:14
	← ph^_ left	06:14
	→ ankit9 joined	06:14
	← ankit9 left	06:14
	→ ankit9 joined	06:15
	→ lightcap_ joined	06:16
	← cloudhead left	06:17
	→ saurabhverma joined	06:17
	← saurabhverma left	06:18
	→ saurabhverma joined	06:19
	← saurabhverma left	06:19
	→ ppawel joined	06:20
	← k0001 left	06:20
	← kuadrosx left	06:21
context	clear	06:21
	ls	06:21
	oops	06:21
	→ psankar joined	06:23
	→ sea-gull joined	06:23
	→ k0001 joined	06:23
	→ drizzd joined	06:23
	← ereslibre left	06:24
	← k0001 left	06:24
	→ k0001 joined	06:25
	→ j416 joined	06:26
	← jceb left	06:26
	← gebi left	06:26
	→ gebi joined	06:26
	→ saurabhverma joined	06:28
	← saurabhverma left	06:29
	→ Mage_Dude joined	06:29
	← girishr left	06:31
	← j416 left	06:32
	← lightcap_ left	06:33
	← neoeinstein left	06:34
	→ fynn joined	06:36
	← alexyk left	06:38
fynn	So, I'm using a mergetool for the first time: vimdiff	06:38
	I have a LOCAL buffer on the left and a REMOTE one on the right.	06:38
	But what is the actual final version of the file that I'm going to commit?	06:38
	Basically, I'm asking what's the proper procedure to resolve a conflict through a mergetool that uses a split screen like this.	06:39
	← drizzd_ left	06:39
	← johndoigiii left	06:39
	← tongueroo left	06:41
joeytwiddle	fynn: i'm not an expert, but i imagine LOCAL is the one you want to fix	06:41
fynn	joeytwiddle: I thought so too, but it turns out to not be so simple.	06:43
	is there a way I can mark a file as unmerged?	06:43
	← lightcap left	06:43
fynn	I'd like to experiment with it a bit.	06:43
tomoj	fynn: I thought there were supposed to be three windows?	06:43
	or whatever vim calls them	06:43
fynn	tomoj: hm, good point, I may have screwed things up with my custom settings ;)	06:44
	anyways, how do I mark a file as unmerged again?	06:44
wereHamster	fynn: git checkout -m -- <file>	06:44
fynn	thanks.	06:45
	← giallu left	06:46
	→ tongueroo joined	06:47
tomoj	I get LOCAL on the left, REMOTE on the right, and another in the middle	06:48
	→ Derander joined	06:48
fynn	tomoj: yup	06:50
	exactly, just found it in the source:	06:50
	http://github.com/gitmirror/git/blob/606475f3178784e5a6b3a01dce1a54314345cf43/git-mergetool--lib.sh	06:50
	← ericindc left	06:50
Gitbot	[git 606475f31]: http://tinyurl.com/ndnztk -- Remove filename from conflict markers	06:50
fynn	I wrote my own custom difftool and mergetool	06:51
	but I reused my difftool as the mergetool	06:51
	← vals_ left	06:51
fynn	and forgot to add $REMOTE to the mergetool one.	06:51
	← sgrimm left	06:54
fynn	wereHamster: btw, "git checkout -m -- file.py" doesn't seem to mark the file as unmerged.	06:54
	← AkumaStreak left	06:55
fynn	The file appears as "modified" (and staged) in git status, not unmerged.	06:55
	→ thiago_home joined	06:56
	← ankit9 left	06:56
	→ ankit9 joined	06:56
	→ roop joined	06:57
	→ johndoigiii joined	06:57
wereHamster	does mergetool only pick up files that are marked as unmerged?	06:58
	→ thhp_ joined	06:59
	→ Fullmoon joined	07:00
	→ mnick joined	07:00
	← __bert__ left	07:00
fynn	yup.	07:00
	→ sgrimm joined	07:01
	← bugfux left	07:02
tomoj	is git leaving the .orig file or is that my mergetool?	07:02
	← tongueroo left	07:03
	← bobesponja left	07:03
drizzd	tomoj: see the mergetool.keepBackup config option	07:04
	← joevandyk left	07:04
	← sgrimm left	07:04
tomoj	drizzd: thanks	07:04
	→ ssmithstone joined	07:05
	→ ciskje joined	07:05
	→ bugfux joined	07:05
	← Korren left	07:06
ssmithstone	I have a svn repos which i've cloned with git svn clone i've created a branch called sprint_1 on the svn and started work with it on my local git with git checkout i've done the work dcommit'd back to the svn into the brances/sprint_1 , now i want to put this on to the trunk so i do checkout master then git merge sprint_1 which pulls the changes into the master branch however when i go to do a git svn dcommit the u	07:08
	← ankit9 left	07:08
	→ ankit9 joined	07:09
tomoj	anyone know about getting emerge mergetool to use emacsclient? I found a bit of discussion about it on the mailing list but no conclusion	07:09
	→ khmarbaise joined	07:09
	← ankit9 left	07:09
	← sunoano left	07:09
	→ ankit9 joined	07:09
	← ankit9 left	07:09
thiago_home	ssmithstone: don't use git merge with git svn	07:09
	it won't work	07:09
ssmithstone	thiago_home: what command will work ?	07:10
thiago_home	ssmithstone: none	07:10
	well, actually, git merge --squash	07:10
	that will work	07:10
	← bugfux left	07:10
	→ khaase joined	07:10
	→ sunoano joined	07:11
tomoj	guess I need to build a custom mergetool as well :(	07:11
	← ciskje left	07:11
	← khaase left	07:11
gebi	ssmithstone: git merge works, but only if you make a fast forward merge, (you have to rebase your branch on top of trunk)	07:12
	→ ciskje joined	07:12
thiago_home	a squash will work too	07:13
ssmithstone	thiago_home: cheers that worked a treat just what i wanted	07:13
thiago_home	there are two problems with a normal merge:	07:13
	→ rraasch joined	07:13
thiago_home	one is that SVN just can't hope to represent the level of information that Git has	07:13
	second is that git-svn scans the commits to find out which branch to submit to, and it'll find the wrong branch	07:13
	thiago_home → thiago	07:14
	→ david joined	07:14
	david → Guest15533	07:14
ssmithstone	cheers for the help	07:15
	← GodEater left	07:15
	→ justin-george joined	07:15
	← ssmithstone left	07:15
	→ GodEater joined	07:16
	→ giallu joined	07:17
wereHamster	or rebase would also work	07:18
	joshthecoder → joshthecoder_afk	07:19
	← sitaram left	07:19
	→ galderz joined	07:20
	← agenteo left	07:23
	← yinzen left	07:23
	← laj left	07:23
	← kingfishr left	07:25
AAA_awright	Is there a way, or any planned way, to discard all history until the very most recent? Actually, not keep any data in the repository at all? The purpose would be to speed updates/pulling from a server, even better then rsync could.	07:25
	→ tango_ joined	07:30
	→ girishr joined	07:32
	→ j416 joined	07:33
drizzd	AAA_awright: it's possible to do shallow clones	07:34
	→ __bert__ joined	07:34
	→ karmi joined	07:36
	← justin-george left	07:36
	→ struberg joined	07:37
Mage_Dude	drizzd: Can't you just checkout a tag/hash? And only pull file versions associated with that tag/hash?	07:41
epic	how do i truncate all my local patches to 1 big one?	07:41
	← mtvee left	07:42
drizzd	Mage_Dude: first of all, git only allows you to access published tags or branches for security reasons	07:44
	if you fetch any one of those, git will download the associated history and all trees	07:44
	→ johan-s joined	07:45
AAA_awright	drizzd: Yeah, I know about that, but can you /keep/ it shallow? Is there any way to prune objects not used by the current checkout?	07:45
	fr\| → fr	07:45
Ilari	AAA_awright: Transfer diffs? With binary diffing enabled so it can handle binary files?	07:45
drizzd	AAA_awright: I don't think so	07:45
Ilari	AAA_awright: And 'git apply' works without repository.	07:46
AAA_awright	Not exactly what I had in mind	07:47
	A client only needs to tell a server I have revision x, and the public repository server simply sends the compressed changes, more efficent then rsync.	07:48
	→ fujin joined	07:48
	→ jmesnil joined	07:49
drizzd	AAA_awright: yes, in theory you are correct. Git cannot currently do that.	07:50
	← dreiss left	07:51
drizzd	OTOH, if you really benefit from the diff, the overhead of getting the versions in between should not be too great.	07:51
	← Derander left	07:51
	tbf\|zzz → tbf	07:51
drizzd	s/diff/delta	07:51
	→ ceyusa joined	07:51
	drizzd has to leave	07:52
	→ ikrabbe joined	07:52
	→ priidu joined	07:53
	→ Superbaloo joined	07:56
AAA_awright	What about git --git-dir=.gitnew clone --shallow 1 .git	07:57
	← rs left	07:57
AAA_awright	Then rm .git, mv .gitnew .git, though that is a bit of a hack...	07:57
	← thiago left	07:58
wereHamster	imho, discarding history won't make pulling much faster. commits and trees make up a very small fraction of the actual data.	07:59
	→ lbt joined	07:59
	→ ponch joined	08:01
	← rakoth left	08:02
	→ naeu joined	08:02
	→ aatanasov joined	08:02
AAA_awright	I know how efficent git is at packing objects, but if you are never going to use the history (only the public repo would keep history to resolve diffs), why heep it around? My intended target is for fairly large binary files, anyways.	08:02
	← Pewpewarrows left	08:03
	→ shruggar joined	08:03
	→ struberg1 joined	08:03
Ilari	AAA_awright: The server can give client a tag that client replays on next update...	08:04
	← aatanasov left	08:04
	→ robinr joined	08:05
AAA_awright	You could, but I still like the idea of this with Git because you can commit changes and push them to a mirror.	08:06
	→ kinkerl joined	08:06
ponch	Hi all. Is there a way to convert a single trivial commit into a standard patch as would be created by the standard diff program? I mean without the git extensions such as sha indication etc...	08:06
AAA_awright	git diff HEAD HEAD^ (right?)	08:07
struberg1	or git diff hash1 hash2	08:07
	→ aatanasov joined	08:07
Ilari	ponch: Or 'git format-patch <commit>^!'.	08:07
ponch	but i don't want the lines beginnig with "diff --git ..." and "index ..."	08:08
Ilari	You can't disable sha indication. Standard patch should be able to digest it anyway.	08:08
AAA_awright	I thought those were just comments	08:08
wereHamster	AAA_awright: what about git archive?	08:09
	← j416 left	08:09
	← drizzd left	08:09
ponch	yes i know that, but sometimes i send git diff to people worrying about those lines and i'd like to know if is there a way to disable them	08:11
Ilari	ponch: AFAIK, no, there isn't.	08:11
ikrabbe	I have a development host (D) and a deployment host (I), where D can reach I, but I cannot reach D. So I want to push my changes from D to I. But I get that warning that I need to set receive.denyCurrentBranch=ignore.	08:11
AAA_awright	Maybe for something else... hm... but no. Just some way to keep an absolute minimum amount of data in the repository.	08:11
ponch	ok thanks for the clarification	08:11
ikrabbe	is there another way for doing so ?	08:11
Ilari	ponch: Extensions like copies, renames, binary patching... Those aren't compatible with ordinary patch and aren't even on by default.	08:12
ikrabbe	or is there a way that I don't need to reset --hard my host (I) everytime I push into its current HEAD?	08:12
	→ saurabhverma joined	08:12
	→ __iron joined	08:13
Ilari	ikrabbe: Set up suitable hook. See man githooks	08:13
Gitbot	ikrabbe: the githooks manpage can be found at http://git.or.cz/man/githooks	08:13
Ilari	ikrabbe: And as hint: Push is not a commit.	08:13
ikrabbe	I know, I first commit, then I push	08:13
charon	ikrabbe: said hook is also described in ,faq non-bare	08:14
Gitbot	ikrabbe: Pushing to non-bare repositories is discouraged. See http://git.or.cz/gitwiki/GitFaq#non-bare	08:14
ponch	Ilari: yes I know that, but I have this need only for trivial patches changing one or two lines per file	08:14
	Ilari: I think I'll use sed and grep	08:15
	← fynn left	08:16
	→ rakoth joined	08:16
Ilari	ponch: But AFAIK, the stuff that's on by default IS.	08:17
	← cedricv left	08:17
Ilari	ponch: Perhaps even future versions of patch could deal with full set of git extensions to that format... But some extensions are bit nasty to support fully.	08:19
	→ cedricv joined	08:20
	← ben_h left	08:20
Ilari	ponch: I tested: patch does not bork on diff --git and index lines...	08:22
	→ JohnFlux joined	08:23
ikrabbe	ok, I understand, that pushing to non-bare repos is discouraged. But my action isn't too uncommon. Maybe I should setup a bare repo near the deployment copy. Actually those hooks don't satisfy me too. Are there other solutions for that topic ?	08:23
	→ Weaselweb joined	08:23
JohnFlux	when I do "git branch -a" how do I know what branches are local and which are remote?	08:23
	hmm I'm confused	08:24
	if I do "git branch -r" it shows me branches which have been deleted on remote	08:24
	← mithro left	08:24
JohnFlux	I basically want to delete any local branches that I've deleted the remote branch for	08:24
wereHamster	JohnFlux: branches that start with remotes/ are remote tracking branches	08:24
ikrabbe	JohnFlux, git branch -a shows remotes/XXX/branch for remote branches	08:24
bremner	JohnFlux: man git-remote	08:24
JohnFlux	hmm I think i need a newer version of git	08:24
Gitbot	JohnFlux: the git-remote manpage can be found at http://git.or.cz/man/git-remote	08:24
wereHamster	JohnFlux: and take a look at git remote prune	08:24
charon	ikrabbe: sorry for being unclear. i know that the faq entry states you "shouldn't" do that, but you have the exact use-case and the faq _also_ describes the hook required for that.	08:25
	← struberg left	08:25
wereHamster	JohnFlux: or you can use git branch -a --color	08:26
	→ solydzajs joined	08:26
	← saurabhverma left	08:26
ponch	Ilari: yes, true. but there isn't a way to remove those lines, apart using sed and grep. I was asking only this.	08:27
	→ karmi_ joined	08:27
Ilari	ponch: Just removing the diff --git line might actually break patch.	08:28
	→ Teo` joined	08:28
ponch	Ilari: just remove the ^index line, and the "--git " token	08:28
	← johndoigiii left	08:28
ponch	Ilari: I'm not saying to work this this normally, but only in rare occasion	08:29
	← sfwc left	08:31
Ilari	ponch: Some stupid patch-lookalike on remote side?	08:31
ponch	Ilari: yes	08:32
Ilari	ponch: And if it is, just doing those modifications won't necressarily make it work.	08:32
ponch	Ilari: true, and I now understand why there isn't a way to do it	08:32
Ilari	Normal patches can be stream-processed. Doesn't work for some git extensions (without making backup files in some cases).	08:34
	→ saurabhverma joined	08:34
	← saurabhverma left	08:34
Ilari	That's the main difficulty in supporting git extensions in patch utility...	08:35
	And the index lines are effectively unsupportable (good thing that they can be mostly ignored) by anything except git. But OTOH, not even git apply can always use them (they require access to repository).	08:36
	→ voker57 joined	08:37
	← rakoth left	08:40
	← aatanasov left	08:40
	→ fujin_ joined	08:40
	lut4rp_AFK → lut4rp	08:43
	← Mage_Dude left	08:43
	← ph^ left	08:46
	← Guest15533 left	08:47
	← fujin left	08:50
	→ bentob0x joined	08:51
	← karmi left	08:52
	← JohnFlux left	08:53
ikrabbe	charon. yes I've read that. somehow the hook solution doesn't satisfy me, as its far to complex, though its stated to be easy (the mentioned link). I will think about a better solution, like pushing into a non active branch.	08:53
	← ceyusa left	08:53
	← cpg left	08:54
charon	ikrabbe: if you know you will never have uncommitted changes in that repo (i.e., never work in it directly) you can also just use 'git reset --hard' as the hook action	08:54
	→ DavidKlein joined	08:54
ikrabbe	charon, yes, but maybe there is a smarter solution to that topic ... git reset --hard, sounds so ... well ... hard.	08:55
charon	:)	08:55
	there's also a receive.denycurrentbranch = updateinstead patch floating on the list	08:56
ikrabbe	:D	08:56
	how about a git deploy	08:56
	→ ceyusa joined	08:56
tomoj	ikrabbe: I'm currently setting up capistrano on my host :)	08:57
charon	ikrabbe: http://article.gmane.org/gmane.comp.version-control.git/110251/match=updateinstead	08:57
	← khmarbaise left	08:57
charon	ikrabbe: and http://article.gmane.org/gmane.comp.version-control.git/121334/match=updateinstead for a success story	08:57
	→ rudi_s joined	08:57
	→ kobani joined	08:57
	← Marmouset left	08:57
	→ khmarbaise joined	08:57
	← reaVer left	08:58
	← ponch left	08:59
	→ ponch joined	08:59
	→ mat69 joined	09:00
	← jds left	09:01
mat69	is there a way to merge specific commits to clean up a local git branch before pushing upstream --> e.g. commit nr. 1 with nr. 12 (using nr. here to show that a lot were inbetween) as these are on the same thing?	09:02
	→ ams joined	09:02
charon	mat69: man git-rebase, interactive mode, "squash" action	09:03
Gitbot	charon: the git-rebase manpage can be found at http://git.or.cz/man/git-rebase	09:03
mat69	charon: thank you! will try that :)	09:03
	→ psynaptic joined	09:03
fujin_	git reset --soft HEAD~12; git commit -a is one I tend to use when I can't be arsed interactive rebasing	09:03
	fujin_ → fujin	09:04
shruggar	or just "git rebase -i the-upstream-branch", it's pretty intuitive.	09:04
charon	fujin: that would squash all 12 of them	09:04
	→ intripoon joined	09:04
	→ reaVer joined	09:04
	← psankar left	09:04
fujin	ya, thus the lazyfactor	09:04
charon	you shouldn't act lazy when pushing upstream	09:04
	→ jds joined	09:04
charon	besides, -a and --soft are sort of contradictory	09:05
fujin	yeah, no	09:05
	→ justin-george joined	09:05
charon	(not that they contradict each other, but the intent of --soft is to keep the original commit but -a says to also grab anything that was unstaged)	09:05
shruggar	ick, so you just squash 12 random commits into one?	09:05
fujin	Occasionally, yeah.	09:06
	→ saurabhverma joined	09:06
charon	let's hope you never need to revert one of the original 12	09:07
fujin	well considering I just wrote and commit all 12 and then decided that I wanted to squash them into one	09:07
	I'd say that's unlikely	09:07
shruggar	I find rebase -i to be a good sanity check to make sure the changes I'm committing are actually what I wanted. Like the commit-level version of doing git add --patch instead of git add	09:07
fujin	seriously though, thanks for trying to tell me how to operate git	09:07
charon	well, if you're offended by good advice maybe you shouldn't give bad advice to someone clearly not asking for what you told him	09:08
shruggar	okay, I'm separating us three before a fight breaks out :)	09:08
fujin	dude, that wasn't good advice	09:09
	you told me shit I did not want to here and presumed I didn't know the intended behaviour	09:09
	they are a similar enough operation that it was worth mentioning	09:09
	go back to being asleep or clueless	09:09
shruggar	I think interactive rebasing is fun :)	09:12
fujin	yeah, me too	09:13
shruggar	it makes me feel a little powerful after I do it, like I could control the weather or something	09:13
	→ psankar joined	09:13
fujin	totally :]	09:13
	→ Octalot joined	09:15
	→ the_undefined_ joined	09:15
the_undefined_	is there a command listing all "untracked" files? Ideally within a given path?	09:15
	← fujin left	09:16
the_undefined_	this is the closest I have gotten to: git clean -n scripts/	09:16
	just wondering if there is an alternative that does not have any human-readable text mangled with it	09:16
shruggar	the_undefined_, what is your goal? Do you need to pipe a clean list to a script, or see them in a.. okay, that answers that	09:16
	→ sitaram joined	09:16
	→ fujin joined	09:16
charon	the_undefined_: git ls-files -o -- scripts/	09:17
the_undefined_	shruggar: pipe the list into a script I'm writing	09:17
	thx charon !!	09:17
	works perfectly	09:17
	← justin-george left	09:18
	← ponch left	09:19
	→ sadleder joined	09:19
	→ name joined	09:25
	→ dys` joined	09:25
	→ doener joined	09:28
	lut4rp → lut4rp_AFK	09:30
	→ voker57_ joined	09:30
	← dys left	09:31
	→ mcella joined	09:31
	← jeffrey__ left	09:31
	→ WinterWeaver joined	09:32
	dys` → dys	09:33
	→ JohnFlux joined	09:36
	→ icwiener joined	09:37
	← Moult left	09:38
	→ Hendu joined	09:39
	→ jony joined	09:43
	← jony left	09:43
	→ _daniel_ joined	09:45
	→ aatanasov joined	09:45
	← flippo left	09:47
	← voker57 left	09:48
	→ Moult joined	09:52
	→ psoo joined	09:53
	→ Gabriel joined	09:55
	← sitaram left	09:56
	→ elmob joined	09:58
	→ cedric_ joined	10:01
	→ Pewpewarrows joined	10:01
	→ sgrimm joined	10:02
	→ yann\|work joined	10:03
yann\|work	there is no option to git-apply to accept line offsets ?	10:04
	← sea-gull left	10:05
	← aatanasov left	10:05
	→ shilin joined	10:07
shilin	hi, 'git grep -c abc' - give me all files, even with zero abc, how to make it output only files that contain 'abc' ?	10:08
	← mat69 left	10:10
tomoj	shilin: man git-grep	10:10
Gitbot	shilin: the git-grep manpage can be found at http://git.or.cz/man/git-grep	10:10
	→ mat69 joined	10:12
	→ HG` joined	10:12
	→ aeiou joined	10:13
	← bentob0x left	10:14
aeiou	I have a dir (projectsettings/) that i want git to ignore, but i already added it via git add *	10:14
	do i add it to .gitignore then git-rm ?	10:14
	i dont want to delete it from the hdd just from being versioned	10:14
charon	aeiou: git rm --cached	10:15
	→ Chillance joined	10:15
aeiou	thanks!	10:16
shilin	been there before asking :), but still don't see the right option, in git community book -c is enough for what I am looking for	10:17
	→ bonhoffer joined	10:17
	← robinr left	10:18
	→ peti joined	10:19
bonhoffer	rubymine is asking for "origin name" what is that?	10:19
aeiou	are there any git repository viewers that run on php?	10:19
	i just want to see the directory tree and differences	10:20
w\|shimmy	(hint: git grep -c --all-match abc)	10:20
bonhoffer	i am using gitosis with a clone url of git@myserver:foo.git	10:20
aeiou	i dont need any fancy ticket systems	10:20
	→ Cynner joined	10:20
peti	Hi. Does anyone know whether Git can access a repository over (anonymous) FTP instead of HTTP?	10:20
	← Cynner left	10:21
doener	shilin: doesn't happen for me	10:21
Ilari	peti: AFAIK, yes.	10:21
	peti: But fetch only, no push.	10:21
w\|shimmy	shilin: its --all-match, in addition to -c	10:22
peti	Ilari: Very cool. Do you by any chance know whether I can configure an external program that's to-be-used for fetching? I.e. can I tell Git to call nc-ftp?	10:22
shilin	I see now, for some reason I considered that to be a synonim for -c when looked at help in my terminal, my bad, thank you!	10:23
	→ Dvyjones joined	10:23
	→ jaeckel joined	10:23
Ilari	peti: Nope. It uses curl to do the transfers.	10:23
peti	Ilari: Hmm. That's a shame. I am behind some mad firewall that requires a special FTP client to get through. I don't think libcurl can do that.	10:24
	→ parasti joined	10:24
	→ bentob0x joined	10:24
Ilari	peti: Probably just active/passive mode thing?	10:25
peti	Ilari: I don't know for sure. All that knowledge is super secret. :-\|	10:26
	→ linopolus joined	10:26
doener	w\|shimmy: hm? "git grep -c abc" and "git grep -c --all-match abc" do the same for me	10:26
linopolus	Hi	10:26
peti	Ilari: Anyway, it's simple enough to find out ... I'll just try it.	10:26
	← linopolus left	10:26
Ilari	peti: The difference between active and passive is that one opens the data connection from server to client and the other opens it client to server.	10:27
peti	Ilari: Yes, I know the FTP protocol.	10:27
w\|shimmy	doener: I get a full list :0 with just -c.	10:28
doener	w\|shimmy: git version? mine's git version 1.6.3.3.385.g60647	10:29
	makies → makies_bot	10:29
	→ klbate joined	10:29
w\|shimmy	git version 1.6.3.3.467.g98a79	10:30
doener	w\|shimmy: that's next I guess?	10:30
w\|shimmy	doener: yeah, from some time in the last week or two	10:31
	← bonhoffer left	10:31
	→ cipio joined	10:31
JohnFlux	what's the magic command to enable colors?	10:32
	and why isn't it on by default :-)	10:33
	it's so hard to find the command via google	10:34
	← ia left	10:34
	→ ia joined	10:34
	← cedric_ left	10:34
JohnFlux	ah color.ui that's the key	10:34
doener	w\|shimmy: hm, now that's interesting... If I redirect stdout, then I see the :0 lines	10:35
cipio	so for some reason, rubymine can connect to my git clone and pull it down - -but command prompt gives me: "fatal: The remote end hung up unexpectedly"	10:35
	in windows	10:35
doener	w\|shimmy: http://git.pastebin.com/m359b2659	10:36
w\|shimmy	doener: my git grep is paged, is that default?	10:36
	→ Yuuhi joined	10:36
doener	w\|shimmy: if I make the terminal window small enough, I get it paged as well. less is started with -F, so big window => less quits	10:39
	w\|shimmy: but even with the small window, I don't see the :0 lines	10:40
w\|shimmy	doener: interesting. git log builtin-grep doesn't have anything that looks related recently either	10:41
peti	Ilari: I tried it, but unfortunately Curl doesn't get past the firewall. Apparently, I really need that specially patched ftp client program. :-( Anyway, thank you for your help!	10:41
Ilari	peti: You tried that active/passive (probably there's way to set it)?	10:42
peti	Ilari: Yes, I did, but it didn't make a difference.	10:43
doener	w\|shimmy: hm, if I don't redirect stdout, I see just git, sh and less being execve'd (strace)	10:45
	w\|shimmy: but if I redirect stdout, I see grep (non-git) being executed	10:45
	← fujin left	10:45
Ilari	peti: Get connection dump (strace, tcpdump, wireshark?) somehow and figure out what ncftp is doing to get past the firewall...	10:45
peti	Ilari: I don't have root privileges, so I cannot tcpdump anything. Strace might work, though, I could try that.	10:46
Ilari	peti: Extracting the connection data from strace is not trivial.	10:47
doener	w\|shimmy: and of course it never happens when you use --cached or a specific tree (as an external grep wouldn't find that)	10:48
peti	Ilari: Yes, and to make matters worse, there's no strace on that machine either.	10:48
	→ fujin joined	10:49
doener	w\|shimmy: Ah! Found out what's going on	10:50
	← cytrinox left	10:51
Ilari	peti: wget --passive-ftp doesn't work?	10:51
doener	w\|shimmy: I have colors enabled for git-grep, which causes it not to use an external grep	10:51
peti	Ilari: No.	10:52
w\|shimmy	doener: I think I do too; but I'm only seeing git being execve'd	10:53
Ilari	peti: Wonder what the heck that firewall is doing then...	10:53
doener	w\|shimmy: maybe you forgot to use strace with -f?	10:53
w\|shimmy	doener: of course I did	10:53
Ilari	peti: But if you could hack libcurl to make ncftp handle the connections, that could work.	10:53
peti	Ilari: The admins probably hardcoded some magic port number or something. It's security by obscurity.	10:54
w\|shimmy	doener: yep, external grep called	10:54
	← karmi_ left	10:54
doener	w\|shimmy: yeah, and with "git grep --no-color -c abc" I see the :0, too	10:55
	w\|shimmy: and you should not see them with "git grep --color -c abc"	10:55
peti	Ilari: I would like Git to call that external program (curl, wget, ncftp, whatever) without libcurl being involved at all. It baffles me that this is not possible, because it's actually quite simple to implement.	10:55
w\|shimmy	doener: yep	10:56
doener	(or just --no-ext-grep)	10:56
Ilari	peti: Nope, it isn't simple to implement.	10:57
peti	Ilari: Huh? What's difficult about calling an external program to fetch a file from an URL?	10:57
Ilari	peti: The dumbfetcher code is quite tightly bound to libcurl.	10:58
	→ cytrinox joined	11:00
peti	Ilari: I don't know ... I never really looked at the code. To be honest, I don't know libcurl very much either. Maybe what I'm trying to do is really simple, but I just don't know about it. I guess, I'll be reading some documentation. Again, thanks for your help.	11:00
	→ d0k joined	11:01
	→ ntoll joined	11:02
	→ the_undefined joined	11:03
peti	Take care, guys.	11:03
	← peti left	11:03
	→ moonflux joined	11:05
ikrabbe	curl is very capable of doing everything to HTTP	11:05
	→ mat69_ joined	11:06
	→ spaceonline joined	11:08
	→ aziz joined	11:12
	← girishr left	11:14
	→ elmo2k3 joined	11:17
	← sgrimm left	11:19
	← mat69 left	11:20
	← the_undefined_ left	11:22
	→ UUncia joined	11:22
	← disjunctor left	11:24
	← elmo2k3 left	11:27
	→ elmo2k3 joined	11:27
	← voker57_ left	11:28
	← girishr_ left	11:28
	→ voker57_ joined	11:28
	← spaceonline left	11:29
	← cipio left	11:30
	→ bonhoffer joined	11:30
	← elmob left	11:32
	← Moult left	11:32
	→ corni joined	11:33
	→ juanjoc joined	11:33
	→ savonarola joined	11:34
	→ mbroeker joined	11:34
	← savonarola left	11:34
	→ savonarola joined	11:34
	← upul left	11:35
	← savonarola left	11:35
	← priidu left	11:36
	→ hpoydar joined	11:38
	makies_bot → makies	11:42
	← Hendu left	11:43
	→ aatanasov joined	11:43
	→ sergeykish joined	11:44
	→ savonarola joined	11:45
	← savonarola left	11:50
	→ Fatal joined	11:51
Fatal	is there a way for me to see all deleted files throughout the history of my tree?	11:51
	worded that wrong, I bet there is, so the question should probably be; how can I find files that I deleted a bunch of commits ago	11:52
doener	Fatal: git log --all --diff-filter=d --name-only --pretty=format: \| sort -u	11:52
Ilari	Fatal: 'git log --name-status --diff-filter=D'?	11:53
doener	Fatal: ehrm, --diff-filter=D (upper case)	11:53
Fatal	doener/Ilari: thanks	11:54
tomoj	Fatal: did you pick that nick so that messages to you would look like errors? :D	11:55
Fatal	tomoj: sadly no	11:55
	→ marvil07 joined	11:55
Fatal	tomoj: but that's the main reason why I stopped idling in here :)	11:55
	→ Hendu joined	11:56
	← scientes_ left	11:57
	→ disjunctor joined	11:57
	← bonhoffer left	12:02
	← LylePerl left	12:04
	→ ankit9 joined	12:04
	← ankit9 left	12:04
	→ sergeyki1h joined	12:05
	← Hendu left	12:05
	→ jrmuizel joined	12:06
	← jrmuizel left	12:06
	→ kbingham joined	12:07
	→ mcella_ joined	12:08
	← mcella left	12:08
	← hummingkid left	12:10
	→ taher joined	12:11
	→ Cromulent joined	12:14
	lut4rp_AFK → lut4rp	12:16
	→ ankit9 joined	12:17
	← sadleder left	12:22
	→ bonhoffer joined	12:22
	← ankit9 left	12:22
	← rraasch left	12:22
	→ ankit9 joined	12:22
	← DavidKlein left	12:23
	→ elmex_ joined	12:23
	→ alanhaggai joined	12:23
	shruggar considers changing his nick to "Warning"	12:24
	→ sitaram joined	12:25
	→ Saviq joined	12:25
	← dys left	12:25
Saviq	guys how do I get diff of two latest revisions?	12:25
jast	you mean "git show"?	12:26
	→ sukut joined	12:26
	← taher left	12:26
Saviq	or "git diff", I want to create a diff of last two commits	12:26
	← ankit9 left	12:26
	← giallu left	12:26
shruggar	git diff HEAD~2..HEAD ? does that not work?	12:27
	← sergeykish left	12:27
sukut	is there a way to see a version of a file without having to checkout?	12:28
Saviq	shruggar: it probably does :) I'm completely new to git	12:28
shruggar	git show commit-sha:path/to/file/relative/to/root/of/work/tree	12:28
	to sukut ^	12:29
	← teuf left	12:29
Saviq	hmm I cloned from git://github.com/philn/pygobject.git (http://github.com/philn/pygobject/tree/wakeupfd) but I don't see the commits made in this branch... do I need to switch somehow after cloning?	12:29
	→ teuf joined	12:29
	← dkr21 left	12:29
	→ dkr21 joined	12:29
	→ jrmuizel joined	12:30
shruggar	Saviq, not sure I understand the question, but "checkout" is how you switch between local branches. You may want to look through the "gittutorial" manpage	12:31
sukut	shruggar: I did try that, but it says "fatal: ambiguous argument my-giant-commit-hash:path/to/my/file"	12:31
Saviq	shruggar: try cloning git://github.com/philn/pygobject.git	12:31
	according to http://github.com/philn/pygobject/commits/wakeupfd/ it should have two commits from philn	12:32
Ilari	sukut: Does it say the same for 'git show my-giant-commit-hash'?	12:32
Saviq	and I don't seem to have it here after cloning :/	12:32
	→ Bass10 joined	12:32
sukut	shruggar: no, it shows the commit log and patch	12:33
shruggar	Saviq, they seem to be in the "wakeupfd" branch	12:33
Saviq	yeah, how do I switch... sorry, but I'm really not yet familiar with git	12:34
shruggar	Saviq, seriously check out the git tutorial :) git co --track -b wakeupfd remotes/origin/wakeupfd if I recall correctly	12:35
Saviq	shruggar: yeah got it :) thanks	12:36
sukut	shruggar: sorry, it did work, my mistake, the path was wrong, thanks	12:36
shruggar	sukut: try "git log --name-only my-giant-commit-hash" to verify the- yay, I was right! :)	12:36
Saviq	shruggar: though git doesn't undertand 'co'	12:36
shruggar	Saviq, woops.. me and my aliases	12:36
Saviq	:)	12:36
	← elmex left	12:37
	elmex_ → elmex	12:37
sitaram	shruggar: you have an alias that expands to "yay, I was right! :)" ? cool!	12:37
shruggar	sitaram, no, no.. I don't type that /nearly/ enough to need an alias	12:37
sitaram	:)	12:38
sukut	ha ha ha	12:38
	→ hydrogen joined	12:39
	→ sgrimm joined	12:39
	← _daniel_ left	12:40
	← Saviq left	12:45
	→ sea-gull joined	12:46
	→ lorandi joined	12:49
	← jrmuizel left	12:50
	→ Bennid joined	12:51
	→ ericindc joined	12:52
	← Bennid left	12:52
	← Pewpewarrows left	12:53
	→ Bennid joined	12:53
	→ giallu joined	12:53
	← mcella_ left	12:54
	Will\|work → Harrison	13:00
shruggar	in my git manpages, there are a lot of blocks which look like: .ft C\nsome example commands\n.ft I assume the ".ft" parts are some formatting thing which isn't being translated correctly when I "make doc"... anyone know what might be causing it?	13:01
charon	shruggar: mess with the ASCIIDOC* Makefile variables and pray	13:01
	Harrison → Will\|work	13:03
	← hpoydar left	13:05
struberg1	JGit question: how to checkout without using the GitIndex?	13:05
	← hydrogen left	13:05
	← corni left	13:06
sitaram	shruggar: for what it is worth, asciidoc drives me nuts sometimes. I have one machine on which it all works fine, so I build only on that machine now :)	13:08
	→ mw joined	13:08
shruggar	sitaram, I don't use asciidoc, I use git, so all I really care about is how to fix it, however it's generated :/	13:09
doener	shruggar: just use "make quick-install-doc" in a git.git clone	13:11
sitaram	shruggar: me too; which is why when I found one place that worked I'm sticking to it; sorry for not being helpful	13:12
	→ alienbrain joined	13:12
doener	shruggar: it's not worth messing with asciidoc to build the docs yourself. Well, unless you're working on the docs	13:12
	→ voker57__ joined	13:12
	→ aspotashev\|eeepc joined	13:12
	→ rhett joined	13:15
	→ stringo0 joined	13:15
sitaram	shruggar: you know what, I've got those too now... dang!	13:17
charon	i can't find a working combination of ASCIIDOC8 and DOCBOOK_XSL_172 either... doener's advice is probably best	13:17
shruggar	it's some sort of information virus, we should stop talking about it before we're all infected!	13:18
	whaley_ → whaley	13:18
	← sergiodj left	13:18
	→ Lemao joined	13:19
	← shilin left	13:19
	← Lemao left	13:19
	→ hpoydar joined	13:21
	→ laj joined	13:22
	→ patrikf joined	13:22
charon	aha! ASCIIDOC8=1 ASCIIDOC_NO_ROFF=1 works here, with asciidoc 8.4.4 and docbook-xsl 1.74.0	13:22
	of course _NO_ROFF is only mentioned in Documentation/Makefile	13:23
	but that one also has a nice table mapping docbook-xsl versions to configuration variables	13:23
	of course it still formats the "Rule" blocks in gitworkflows as "Example". duh.	13:24
shruggar	make quick-install-doc works for me. I wasn't previously aware of it	13:25
	← jaeckel left	13:28
	← jksz_ left	13:28
sitaram	doener: last time I did what you said, I got manpages that were _ahead_ of the actual (stable) version, because that only picks up the latest "origin/man". May not be a big deal though...	13:29
	← ceyusa left	13:29
	→ hummingkid joined	13:30
	→ dodo_the_last joined	13:30
	→ priidu joined	13:31
sitaram	(although it is possible to get precisely what you want but just checking out a specific commit and copying the files!)	13:31
	→ krawek joined	13:31
	← voker57_ left	13:34
doener	sitaram: or by using "make quick-install-doc DOC_REF=$what-you-want"	13:34
sitaram	doener: wow -- that works? I took a look at the makefile and thought origin/man was hardcoded -- I guess I don't know makefiles that well!	13:34
	doener: thanks!	13:34
	→ kukks joined	13:35
doener	sitaram: just be careful with it, if you use "DOC_REF=master", I guess it'll just copy the source code to the man dir	13:35
sitaram	doener: I know... it's just doing some plumbing to directly take the tree from a ref and pump it out to some place :-)	13:38
	← schambers left	13:39
	← Chillance left	13:40
	→ leehambley joined	13:40
leehambley	how do I reset a single file to a revision that is head~2 ?	13:40
	reset says "dan't use reset with paths"	13:40
	I have the sha1 of the version i wanna go back to	13:40
sitaram	you can use checkout	13:41
	→ ben_h joined	13:41
shd	leehambley: git checkout revision -- file	13:41
	leehambley: it's not a "reset" however.. you may want to consider using git revert also	13:42
sitaram	shd: if it's a single file he wants to do this to, a checkout is the only thing that makes sense (to my mind)	13:42
leehambley	ohh, thanks - I don't mind about `breaking` the history	13:42
	thanks	13:42
sitaram	(assuming other files were also affected by commits in between)	13:43
	← marvil07 left	13:43
shd	sitaram: assuming HEAD~2 changes more than one file	13:43
leehambley	sitaram: shd that checkout was perfect, thanks	13:43
sitaram	shd: right, thats what I wanted to say	13:43
	← struberg1 left	13:44
	← marc_kdab left	13:46
	→ MisterN joined	13:49
	→ jrmuizel joined	13:50
	→ schambers joined	13:51
	← psankar left	13:52
	→ schambers1 joined	13:53
	← harinath left	13:55
	→ Chillance joined	13:56
	← sitaram left	13:56
	← aeiou left	13:57
	← offby1 left	13:57
	← solydzajs left	14:00
	→ harinath joined	14:00
	→ hugoxrosa joined	14:02
	← w\|shimmy left	14:02
	→ solydzajs joined	14:03
	→ boto joined	14:04
	→ solydzajs_ joined	14:07
	→ wshimmy joined	14:07
	→ mefesto joined	14:08
	← __bert__ left	14:08
	→ __bert__ joined	14:08
	← schambers left	14:09
	← bentob0x left	14:10
	→ bentob0x joined	14:10
	← disjunctor left	14:12
	← knittl left	14:13
	→ mithro joined	14:14
	← dodo_the_last left	14:14
	← aziz left	14:16
	← harinath left	14:18
	→ dodo_the_last joined	14:19
	← solydzajs_ left	14:19
	← solydzajs left	14:22
	← ikrabbe left	14:24
	← tim_jenssen left	14:24
	→ MetaCosm joined	14:27
	→ spaceonline joined	14:28
	→ jackdempsey joined	14:29
	← ericindc left	14:33
	→ ericindc joined	14:35
	→ DaveIngram joined	14:40
	← tbf left	14:41
	→ disjunctor joined	14:43
	← aatanasov left	14:43
	→ rogg joined	14:46
	← Strogg left	14:47
	rogg → Strogg	14:47
	← roop left	14:48
	→ offby1 joined	14:50
	→ c0dem4gnetic joined	14:50
	lut4rp → lut4rp_AFK	14:50
	→ knittl joined	14:51
	schambers1 → schambers	14:51
	← moonflux left	14:51
	→ lamalex joined	14:52
lamalex	Can anyone give me a hand wrapping my head around rebase?	14:52
offby1	ha	14:52
	→ DavidKlein joined	14:52
offby1	sure	14:52
	← WinterWeaver left	14:52
offby1	hand, head, heart!	14:52
	← yann\|work left	14:52
lamalex	:)	14:53
	offby1 struggled with that	14:53
offby1	got a specific question?	14:53
	← johan-s left	14:54
	→ reprore joined	14:54
lamalex	I'm trying to get a branch ready for merging, but dont really want all of my mini commits appended to master's log	14:54
	i tried doing git rebase -i master	14:54
	but it failed :(	14:54
offby1	how?	14:55
	what'd it do?	14:55
	that was the right thing to do, by the way	14:55
	given the problem you just stated.	14:55
lamalex	Cannot 'squash' without a previous commit	14:55
offby1	ah, that's easy	14:55
	I bet you replaced _all_ the "pick"s with "squashes", including the first, right?	14:55
wshimmy	lamalex: rebase -i <whatever-you-just-tried-to-rebase-onto>^	14:55
lamalex	offby1: i squashed all but the last one	14:56
offby1	ok, that's backwards	14:56
	try squashing all the the _first_	14:56
lamalex	last meaning most recent	14:56
	can I squash them all and then give a new commit message?	14:56
offby1	more or less, yes.	14:57
	the caveat is that you have to have at least one "pick" _before_ the first "squash"	14:57
	otherwise there will be no commit to "apply" to master.	14:57
shruggar	"git diff HEAD cron-verify" produces no output, but "git br -d cron-verify" says "error: The branch 'cron-verify' is not an ancestor of your current HEAD."	14:57
offby1	shruggar: probably the two branches have the same tree "by coincidence"	14:58
	shruggar: I always run "gitk --all" when confused	14:58
teuf	wouldn't git merge --squash do the same when you want to squash all commits from a branch ?	14:58
offby1	teuf: could be; I never use that feature, but perhaps I should	14:58
shruggar	also: if I "git merge cron-verify", "git-log -p" shows a huge patch	14:58
wshimmy	shruggar: git cherry cron-verify will tell you if there are different histories	14:58
lamalex	teuf: hm maybe	14:58
offby1	shruggar: ugh	14:58
teuf	(which I'm not sure is a great thing to do for non-trivial merges)	14:58
	→ steveire joined	14:58
	← Cromulent left	14:59
lamalex	I'm not sure rebasing is exactly what I wnat to do, I'm used to bzr which does this for me more or less	14:59
	← sergeyki1h left	15:00
offby1	lamalex: from what you said -- "[I] dont really want all of my mini commits appended to master's log" -- rebasing is _exactly_ what you want	15:00
lamalex	ok good	15:00
	at least I know what I want to do	15:00
offby1	that's the canonical use case	15:00
	I suggest you make a toy repository and play with rebase, to get a feel for what it does	15:00
lamalex	probably a good idea, but I need to merge this branch :)	15:01
	no time for toys	15:01
	so I can squash them all then put my new commit message at the top? or do I edit the first's commit message and keep it as pick?	15:02
	← bcardarella left	15:02
jast	you can simply edit the message after squashing them all. that will probably be easiest	15:02
	(using git commit --amend)	15:02
lamalex	so i just mark them all as squash, save, git commit --amend?	15:03
wshimmy	lamalex: or mark the first as edit	15:03
charon	actually it'll concatenate the entire series of commit messages from all squashed commits, and offer that for editing anyway	15:03
teuf	lamalex: but the history is so uninteresting that you want to get rid of it for merging ?	15:03
jast	I've got a faster solution for this particular situation, by the way	15:03
	← MetaCosm left	15:03
lamalex	teuf: it gets rid of the history?	15:04
	that sucks	15:04
	→ Tommy[D] joined	15:04
jast	well, history as in, it replaces the series of small commits by a big one, thus altering your local history accordingly	15:04
steveire	Anyone know about gitorious? Someone cloned my repo, and I got a message that I had made another clone of it at this link: http://gitorious.org/~steveire/grantlee/tracking_repository_for_5431. I visited, and I get a message saying the repo is being created. Any idea what's going on?	15:04
lamalex	i mean the history in this case isn't very interesting	15:05
	so thats not a big deal	15:05
jast	yeah, well, that's the reason why you wanted to get rid of it in the first place, isn't it	15:05
	we wouldn't be doing all this stuff otherwise	15:05
	→ Korren joined	15:05
lamalex	so are any of you familiar with bzr?	15:05
teuf	history has to be really uninteresting to be dropped imo	15:06
	← Korren left	15:06
jast	and the fastest way to do it is this: 1) git reset --hard to kill all uncommitted changes, 2) git reset --soft <parent commit of what you want to consolidate into one>, 3) git commit	15:06
	another similarly fast way involves using git merge --squash	15:07
lamalex	http://paste2.org/p/300857	15:07
	← priidu left	15:08
lamalex	so that's a bzr log, when a branch gets merged, the branches commits go inline so you can see what happened in the branch	15:08
jast	...	15:08
	that's just a question of visualization	15:08
	gitk displays it a bit like that if you simply use git merge	15:08
	or git log --graph	15:09
	← sunoano left	15:10
lamalex	jast: ah	15:10
	is there a way to view only first level commits?	15:11
jast	sure	15:11
	→ kumbayo joined	15:11
jast	for example, git log --first-parent	15:11
offby1	lamalex: all I know about bzr is: it's so slow it makes me want to kill myself	15:12
jast	oh, hey, are we starting a flamewar?	15:12
lamalex	ha	15:12
offby1	and: the Emacs project will probably switch to it :-(	15:12
lamalex	thats funny because git makes me feel the same way	15:12
teuf	yay SVN FTW!!!	15:12
offby1	huh	15:12
	→ sunoano joined	15:12
offby1	now, if you'd said "so _confusing_ that I want to kill myself", I'd have understood that :)	15:12
jast	all you losers are using total crap systems. sccs is the only way to go!	15:13
	→ scarabx joined	15:13
charon	lamalex: 'rebase -i' isn't exactly the best benchmark ;)	15:13
	→ kuadrosx joined	15:13
offby1	jast: real men don't use revision control.	15:13
shruggar	charon: for either average speed or average confusion level of git.	15:14
	I want everything to be like rebase -i <3	15:14
	← crankharder left	15:15
	← DaveIngram left	15:16
	← dmlloyd left	15:16
	→ dmlloyd joined	15:16
	→ DaveIngram joined	15:19
	→ crankharder joined	15:19
	← DaveIngram left	15:20
	← leehambley left	15:21
	← reprore left	15:21
	→ reprore_ joined	15:22
	← kobani left	15:22
	→ DaveIngram joined	15:22
	← ams left	15:22
	→ nickjohnson joined	15:24
nickjohnson	How can I see the raw tree objects Git stores?	15:24
	Eg, without any pretty printing or formatting at all	15:24
jast	git cat-file tree <tree sha1 name>	15:24
nickjohnson	Thanks!	15:25
	→ corni joined	15:25
	← DaveIngram left	15:25
nickjohnson	How does it format trees?	15:25
	It looks like size, name, hash tuples?	15:25
	→ DaveIngram joined	15:25
	makies → makies_bot	15:25
	← elmo2k3 left	15:26
nickjohnson	Though I don't see obvious terminators	15:26
	Oops, nm	15:26
	→ adeodatus joined	15:26
	← patrikf left	15:28
	← alienbrain left	15:29
	← DaveIngram left	15:29
	→ alexyk joined	15:29
	→ DaveIngram joined	15:30
	→ dreiss joined	15:30
	← alexyk left	15:31
	← crankharder left	15:32
	→ crankharder joined	15:33
	← whaley left	15:37
	→ whaley joined	15:39
	← reprore_ left	15:39
	← glommer left	15:39
	→ reprore_ joined	15:39
	→ mtvee joined	15:40
	→ voker57 joined	15:41
	→ SlipperyChicken joined	15:42
	joshthecoder_afk → joshthecoder	15:45
	→ almagro joined	15:49
	→ alexyk joined	15:49
	→ voker57_ joined	15:49
	→ bradly joined	15:50
	dwmw2_gone → dwmw2	15:50
	→ bgamari joined	15:50
	→ robwilkerson joined	15:55
	← voker57__ left	15:55
	← robwilkerson left	15:55
	← reprore_ left	15:56
	→ reprore_ joined	15:56
	→ Fazer2 joined	15:57
	← nickjohnson left	15:57
Fazer2	hi, I've reinstalled Windows and installed Git again	15:58
	← lamalex left	15:58
	← samjam left	15:58
Fazer2	I tried to open an existing (old) repository, but its .git/config file was filled with "NULL"	15:58
	so I found a backed up version of it and managed to load the repository	15:59
offby1	ugh	15:59
Fazer2	but git now doesn't see any changed in unstaged area	15:59
	any changes	15:59
	← bonhoffer left	16:00
Fazer2	is it a problem that now the repository is on a different partition (with E label, earlier was C)	16:01
	?	16:01
offby1	shouldn	16:01
	t be	16:01
	→ sitaram joined	16:01
Fazer2	hm, I just did "verify database" in git gui and got "error: bad signature, fatal: index file corrupt"	16:03
	← fluxusps left	16:03
	→ bugfux joined	16:04
	→ Alinermedia joined	16:04
	← naeu left	16:05
	→ Derander joined	16:05
sitaram	ouch...	16:07
	← tango_ left	16:07
	← voker57 left	16:07
	→ patmaddox joined	16:07
sitaram	of course, unless you "git add"ed something and then subsuquently lost of changed the file, this isn't a big deal -- there are ways to get back a sane index	16:08
	like git checkout	16:08
	→ altrux joined	16:08
sitaram	and reset	16:09
	← stuffcorpse left	16:09
sitaram	the real question would be why it got corrupt in the first place...	16:09
	offby1 blames society	16:09
	→ stuffcorpse joined	16:10
Fazer2	because of a messy move from one partition to another? and a news windows installation	16:10
	new	16:10
	I foung git reset --HARD could do the trick	16:10
	ouch, sorry for all the typos ;-)	16:11
	heh, .git/index file is filled with "NULLNULLNULL..."	16:12
	sitaram agrees with offby1; it is society. A society using too much Windows that is :-)	16:12
sitaram	.oO(or maybe it got power; I heard power corrupts...!)	16:13
jast	if the index file (.git/index) is corrupted you can simply delete it... git will create a new one as soon as it needs to	16:14
Fazer2	ok	16:14
	← almagro left	16:14
jast	as you noticed, git reset does about the same thing, so it doesn't matter now anyway	16:15
	← warthog9 left	16:15
	→ bx2 joined	16:15
	← galderz left	16:16
Fazer2	but I didn't do git reset, just read about it	16:16
jast	oh.	16:17
	← jmesnil left	16:17
Fazer2	ok, I've deleted the index file and now I have all files in unstaged area, as if I have changed every line	16:18
	← c0dem4gnetic left	16:18
stringo0	noob question - what's an intro to the site called? (For example a flash widget with just buttons?)	16:18
	← mat69_ left	16:18
Ilari	Fazer2: 'git reset' (no arguments).	16:18
stringo0	(that was not a git question ;p)	16:18
	→ quinter joined	16:18
	← ciskje left	16:19
charon	stringo0: annoying ;)	16:19
	← psoo left	16:19
stringo0	fine - but what's it called? I want to use the right terminology :P	16:19
	← doener left	16:19
sitaram	init	16:19
stringo0	something like this - http://nuclearfallout.net/	16:19
	→ doener joined	16:19
sitaram	next page is called add	16:19
	then commit and push :-)	16:20
	→ Lemao joined	16:20
	→ char00les joined	16:20
	← Lemao left	16:20
	→ MetaCosm joined	16:21
	← juanjoc left	16:21
stringo0	-_-	16:22
Fazer2	ok, git reset helped :-)	16:22
stringo0	funny ;p	16:22
	← bugfux left	16:23
offby1	stringo0: it's called "a pain in the ass"	16:24
Ilari	Aarggh... Got 2 years of vendor development as single drop (version control? what version control?). Maginitude of my changes being +13k, -38k... Thats going to take a while to merge... :-(	16:24
	← sukut left	16:25
	→ karmi joined	16:25
Fazer2	a merge request from hell	16:25
offby1	heh	16:25
	← ppawel left	16:25
	← hpoydar left	16:25
	→ hpoydar joined	16:26
Fazer2	http://torvalds-family.blogspot.com/2009/06/happiness-is-warm-scm.html very related	16:26
	"so far it's been such a nice thing that I thought I'd better post while in a good mood. Before somebody sends me the merge request from hell."	16:27
	← Derander left	16:27
Ilari	What I'm doing: Rebase my development (its linear) on top of new version and then commit that as merge of tips. But I'm at patch #6, while my changes span ~130 patches.	16:28
	Also darn. I need new backup strategy for it. It no longer fits on standard 9cm floppy disk... :-/	16:31
stringo0	offby1: I'm trying to figure out what it's called - widget doesn't sound right	16:32
	trying to figure out the name of a module rather, I think I'll just call it a module	16:32
	(not actually thinking of implementing an intro flash annoyance)	16:32
Fazer2	spash?	16:32
	splash	16:32
stringo0	there!	16:32
	splash! Thanks :P	16:32
Fazer2	5 dollars :-P	16:33
	← reprore_ left	16:33
	→ reprore_ joined	16:34
	← tono left	16:34
	→ btanaka joined	16:35
	→ tono joined	16:35
offby1	stringo0: I'm very glad to hear that you're not actually thinking of implementing it; and annoyed that I didn't think of "splash" myself, first.	16:36
	← shruggar left	16:36
Ilari	Yay. Looks like XHTML2 is pretty much dead.	16:37
	→ Moult joined	16:37
afief	does the number of branches in a local repository affect the speed of git operations?	16:38
	← steveire left	16:38
wereHamster	.. and long live html5 without the <video> tag	16:39
tomoj	what, did something just happen? (re html?)	16:40
Ilari	Announcement that XHTML2 group expires at end of the year and won't be continued.	16:40
	→ cpg joined	16:41
	← hpoydar left	16:41
	→ hpoydar joined	16:41
	→ voker57__ joined	16:42
Ilari	AFAIK, <video> didn't get axed.	16:42
offby1	afief: I wouldn't think so	16:42
afief	offby1: thanks	16:42
offby1	I'm just guessing	16:43
tomoj	found this: http://news.zdnet.co.uk/internet/0,1000000097,39670329,00.htm	16:43
afief	offby1: I didn't see anything that would suggest so either, but thought I'd ask here before making any wrong assumptions	16:43
Ilari	Some operations need to look branches up (filesystem dependent slowdown). Some iterate all branches.	16:43
	→ bobesponja joined	16:44
	→ hash joined	16:44
afief	Ilari: you mean for example the git-branch command where stuff is listed?	16:44
Ilari	afief: That's one example. git for-each-branch is another.	16:45
afief	Ilari: I don't know the for-each-branch yet :)	16:45
Ilari	Oops for-each-ref	16:45
	hash → hash_g	16:45
Ilari	Actually, there might be some linear slowdowns. But still, its probably not very noticeable even with huge numbers of branches.	16:46
charon	i guess the most noticeable slowdown is in the phase of fetching where the remote side advertises its branch heads	16:47
	so you probably don't want to fetch from a repo that has thousands of refs	16:47
Ilari	afief: How many k of branches are you talking about?	16:47
	← scarabx left	16:47
afief	Ilari: I was thinking about 2k	16:48
charon	(then again even then it's only a few KBs, so meh)	16:48
Ilari	afief: That's not probably very bad...	16:48
	← giallu left	16:48
hash_g	hi all.	16:48
	how to take some changes from one commit from one branch to other branch?	16:48
	not whole commit just some part	16:48
afief	hash_g: git-cherry-pick I think	16:49
offby1	yup	16:49
	← rhett left	16:49
afief	Ilari, charon: thanks guys, I learn something new everytime I come here	16:49
jast	cherry-pick will give you the whole patch. you can use cherry-pick -n and then use something like git add -i to unstage some of those changes, perhaps.	16:50
	← sitaram left	16:51
	← tono left	16:52
	→ johndoigiii joined	16:52
	← psynaptic left	16:54
hash_g	thanks:) will try that	16:54
	→ psynaptic joined	16:59
	← reprore_ left	16:59
	→ reprore joined	16:59
	← candybar left	17:00
	← bobesponja left	17:00
	→ warthog9 joined	17:02
	→ patrikf joined	17:03
Fazer2	"They told me - "Son, you're special. You were born to do great things". You know what? They were right." - 1989, Linus Torvalds	17:04
	← voker57_ left	17:04
	← alexyk left	17:05
	← ben_h left	17:05
	→ icwiener_ joined	17:06
	→ bobesponja joined	17:06
	→ sergiodj joined	17:08
	→ Ryback_ joined	17:09
	→ Derander joined	17:12
	← icwiener_ left	17:13
	← johndoigiii left	17:13
	← reprore left	17:13
	← the_undefined left	17:13
	→ reprore_ joined	17:13
offby1	wait, that's what they told _me_	17:17
	→ aziz joined	17:17
	← sgrimm left	17:18
	→ alexyk joined	17:20
	← icwiener left	17:21
	→ bugfux joined	17:25
wereHamster	offby1: so, have you done any great things yet?	17:25
	← HG` left	17:25
	→ drizzd joined	17:27
Fazer2	wereHamster: he has spoken to us	17:27
	→ eno__ joined	17:27
	→ SterlingH joined	17:28
hash_g	add -i is great !!! <3	17:28
	← Moult left	17:28
	→ candybar joined	17:30
	← bugfux left	17:31
	→ marvil07 joined	17:32
	→ metalfan_ joined	17:32
metalfan_	hi	17:32
	ive set GIT_EDITOR=vim, but commit -a still starts nano	17:32
	?	17:32
wereHamster	metalfan_: did you export GIT_EDITOR?	17:33
	→ spearce joined	17:34
metalfan_	nope	17:34
	what does that do again?	17:34
wereHamster	it sets that env.variable to all processes	17:34
metalfan_	thx, that worked	17:35
	← eno left	17:38
	← hash_g left	17:38
	→ tongueroo joined	17:38
	← alexyk left	17:39
	→ alexyk joined	17:39
	← sergiodj left	17:39
	→ giallu joined	17:40
ericindc	how can i do a diff off the same file but on different branches?	17:40
wereHamster	ericindc: diff branch1 branch2 -- file	17:41
	git diff ...	17:41
ericindc	wereHamster: ok, i'll try that. thanks.	17:41
	→ bcardarella joined	17:47
	→ flippo joined	17:49
	→ ph^ joined	17:56
	← reprore_ left	17:58
	← saurabhverma left	18:00
	← quinter left	18:01
	→ broonie_ joined	18:03
	← broonie left	18:03
	← kuadrosx left	18:03
	← alanhaggai left	18:04
	← alexyk left	18:04
	→ struberg joined	18:04
	→ kuadrosx joined	18:05
	← bcardarella left	18:07
	← marvil07 left	18:07
	→ hpa joined	18:08
	← Chillance left	18:08
hpa	Odd placement of <<<<'s during merge: http://pastebin.ca/1483228	18:08
	→ Chillance joined	18:08
spearce	what is odd about that hpa?	18:09
wereHamster	hpa: the first, second or third one?	18:09
spearce	looks normal to me	18:09
hpa	Third one	18:09
spearce	its an empty region in head, but inserted region in the other side	18:09
hpa	The first <<<'s should have been at line 114	18:09
wereHamster	your side has added nothing, the other side has added some text	18:09
spearce	possibly head deleted from there	18:10
	while other side modified there	18:10
hpa	I don't think that's what happened, though	18:10
spearce	oooh, i see, you're right, it probably should have been up at 114	18:10
wereHamster	ah, I see. maybe git didn't recognize the context correctly because of the changed whitespace?	18:11
spearce	yuck. post to list, someone who knows the diff code well like gitte should look at that case	18:11
	← Chillance left	18:11
	→ Chillance joined	18:11
	← chrizz left	18:12
	→ corni_ joined	18:13
daed	weird	18:13
	← broonie_ left	18:13
	→ broonie joined	18:13
	→ jchris joined	18:14
	→ MattRich joined	18:16
	← krawek left	18:16
	→ krawek joined	18:16
	→ alexyk joined	18:17
	→ dodo_the_last1 joined	18:17
jackdempsey	hey all, quick process/workflow question........currently if i have a clone of someone's git repo, and i want them to pull in an update from me, i usually create a branch in my repo for the feature, let them pull in things when they can, and then i'll update later and get that change into my master from pulling from them	18:19
	is this extra work? should i just create the change on my master branch, let them pull it in, and be done with it?	18:19
	← broonie left	18:19
	→ broonie joined	18:19
spearce	no, its usually easier to make a branch for them, so its a stable point that won't move around	18:20
	besides, creating a branch is just `git branch for-soandso`	18:20
jackdempsey	just thinking, what happens if they have a bunch of updates i don't, then they apply my commit on top of that....it should still have the same sha (unless they rebase)......so, me creating a branch for it is a but superfluous, no?	18:20
	ok good, so then i'm doing it right :-)	18:20
spearce	not like its hard	18:20
jackdempsey	yeah def	18:20
	ok cool, thx spearce	18:20
	← Derander left	18:20
spearce	and when you update back, and merge into your main work branch, `git branch -d for-soandso` should work, because it was merged in. if they amended a patch or rebase or something, it'll fail, and you can figure it out from there	18:21
jackdempsey	ah yeah, good way to catch that, cool	18:21
	← ericindc left	18:21
	→ stefanb85 joined	18:25
stefanb85	hi all	18:25
	→ Pewpewarrows joined	18:25
stefanb85	I'm a little disturbed about a particular behavior of git merge	18:25
	I ran into a situation when a file under conflict has missing content from the two sources	18:25
	so the resulting merge would contain invalid code	18:26
	you can take a look here:	18:26
	http://syslinux.zytor.com/wiki/index.php/GSoC2009	18:26
	sorry	18:26
	http://pastebin.ca/1483228	18:26
	this one :)	18:26
	and see at the end, the last difference	18:26
	is wrong	18:26
	→ defectiv joined	18:27
defectiv	what's the git version of svn revert?	18:27
wereHamster	stefanb85: hpa just posted the very same code :)	18:27
defectiv	i have "changes to be commited" and "changed but not update", and i just want to revert those files.	18:27
stefanb85	wereHamster, ah, ooops :)	18:27
wereHamster	defectiv: take a look in the 'git status' output	18:27
stefanb85	wereHamster, I told him eariler about this issue	18:28
hpa	wereHamster: Sorry, it's stefanb85's code, not mine :)	18:28
stefanb85	hpa, actually the merge is from your code + my code	18:28
	← corni left	18:28
defectiv	wereHamster: i just quoted from the git status output, hence my problem.	18:29
	so now i want to revert what was shown in git status.	18:30
	→ sergiodj joined	18:30
defectiv	brb...	18:30
	← metalfan_ left	18:30
wereHamster	defectiv: git checkout -- <file> to unstage changes	18:31
	er, wrong. that will revert changes in the working directory	18:31
	→ dmlloyd_ joined	18:31
defectiv	git reset --hard ?	18:31
	yup that worked	18:32
	← dmlloyd left	18:32
	dmlloyd_ → dmlloyd	18:32
wereHamster	defectiv: that will undo all changes in your working tree and also reset the index. A highly destructive command btw	18:32
	← dodo_the_last left	18:32
stefanb85	wereHamster: for comparison, here is the git diff (which behaves correctly) http://pastebin.ca/1483252	18:33
	→ naeu joined	18:33
	← char00les left	18:33
wereHamster	stefanb85: is that only whitespace or also code changes?	18:33
stefanb85	whitespace (indentation) + small portions of code	18:34
	I converted some statements to macro calls	18:34
defectiv	reset index? huh?	18:34
	← Linkadmin left	18:34
	→ Linkadmin joined	18:35
jackdempsey	defectiv: the index is where you put changes to be commited	18:35
	in svn you just commit something and it goes up to your server	18:35
wereHamster	stefanb85: spearce suggested hpa to send a mail to the mailing list. I don't know if he's done that yet. Next time though, try to do whitespace changes in a separate commit, this way code changes won't get lost in between houndreds of lines of diffs	18:35
jackdempsey	in git there's this index or staging area	18:35
	when you run git commit, changes in the index are committed to your repo	18:35
	if the index is new to you, its definitely something you'll want to read up on	18:36
defectiv	the index is where things are staged, right?	18:36
wereHamster	yes	18:36
	← dodo_the_last1 left	18:36
defectiv	what is "destructive" about this command then? it seems analogous to svn revert.	18:36
stefanb85	wereHamster, in this case the whitespace changes came from a branch (hpa's), and the code changes from the other (my branch, which was based on older code, which was not yet affected by whitespace changes at that time)	18:37
	→ leehambley joined	18:37
jackdempsey	defectiv: --soft --mixed and --hard all do a varying level of reverting of work/changes	18:37
	git help reset to look at the examples	18:37
	--hard does the most severe changes (basically anything changed is gone, working directory, index, etc)	18:38
	others will leave your working directory alone and just undo the index	18:38
	play around with it in a sample repo just to get comfortable	18:38
	→ AkumaStreak joined	18:38
	→ apeiros_ joined	18:39
wereHamster	defectiv: git tries very hard not to lose work. It refuses to overwrite unknown files etc. Using reset --hard will revert everything in your working directory, even if if means that git overwrites something which it can't recover later	18:39
	→ bonhoffer joined	18:39
wereHamster	some other commands (checkout, push) have a --force flag which means a similar thing, it irretrievably destroys files or history	18:40
	← drizzd left	18:42
	← ph^ left	18:42
	mtvee → mtvee\|away	18:43
	→ Urgleflogue joined	18:43
	→ Airwulf joined	18:43
	→ ph^ joined	18:43
Airwulf	hey	18:43
	how can I remove a file added with 'git add ...' from the staging area?	18:44
	→ juanjoc joined	18:44
wereHamster	Airwulf: git reset <file>	18:44
	mtvee\|away → mtvee	18:44
	→ ericindc joined	18:44
	← mtvee left	18:45
	← Fazer2 left	18:45
Airwulf	ah I think I have to ask my question in a different way	18:46
	I changed something in a file. After finish I do 'git status' and see. Oh sh*t I forgot to commit the older changes	18:46
	how can I add all the older changes without the file I just have changed	18:47
jackdempsey	so the file has old and new changes in it?	18:48
	or the old changes are in other files	18:48
Airwulf	yes the old changes are in other files	18:48
jackdempsey	and you want to add those, and commit those, and remove the changes from a new file from the index?	18:49
wereHamster	Airwulf: git add <otherfile> for each file you want to commit; git commit; git add <thisfile>; git commit	18:49
Airwulf	my solution is: copy the one file to temp, reset the file, add/commit older changes, copy the new file back and make another add/commit	18:49
	→ jayzer joined	18:49
jackdempsey	ah	18:49
	yeah sounds like you want git stash	18:50
	also, git add -p is useful	18:50
Airwulf	wereHamster, thats my problem. the otherfiles are more than 20	18:50
jackdempsey	and you don't want to specify them by hand right?	18:50
wereHamster	Airwulf: git add -u; git reset <thisfile>; git commit will commit all files but <thisfile>. Is that what you want?	18:50
	lut4rp_AFK → lut4rp	18:51
	← broonie left	18:51
jackdempsey	nice wereHamster didn't know about -u thx	18:51
	← AkumaStreak left	18:52
	← Urgleflogue left	18:52
Airwulf	jackdempsey, that's it I don't want to specify them by hand	18:53
jackdempsey	right, so what wereHamster said is probably great for you	18:53
	add everything, remove the one you don't want added	18:53
	commit those in the index and then you're golden	18:53
	if you don't know git stash i'd recommend looking into it as well as it can be useful at times	18:53
Airwulf	add everything and remove the one I don't want from staging area was my idea. But how can I remove one file from staging area?	18:54
wereHamster	Airwulf: git reset <file>	18:54
	wereHamster is repeating himself	18:54
	→ bizhat joined	18:56
	→ broonie joined	18:56
	← bizhat left	18:56
	→ icwiener joined	18:56
	← leehambley left	18:59
	← cpg left	18:59
	→ cipio joined	18:59
	→ cpg joined	19:01
	← alexyk left	19:03
	← schambers left	19:05
	← broonie left	19:07
	→ bcardarella joined	19:07
	← DaveIngram left	19:08
	← bx2 left	19:10
	← thhp_ left	19:10
	← wagle left	19:11
	→ broonie joined	19:11
	→ alexyk joined	19:12
	→ neoeinstein joined	19:12
	← bonhoffer left	19:12
	→ marvil07 joined	19:13
	← alexyk left	19:13
	→ schambers joined	19:14
	← Airwulf left	19:16
	→ radarek joined	19:19
	→ bizhat joined	19:21
	← cipio left	19:22
	→ \ask joined	19:22
	→ bonhoffer joined	19:23
struberg	spearce, ilari hi! Is there an example how to do a checkout without using GitIndex?	19:23
	→ johan-s joined	19:24
	→ drizzd joined	19:24
	← marvil07 left	19:24
	← hugoxrosa left	19:24
spearce	no, i thought i wrote one for you though that used dircache?	19:25
struberg	hoops, have to check my mails, wait txs :)	19:26
	← bizhat left	19:26
	← johan-s left	19:26
	← CIA-7 left	19:28
	← ph^ left	19:29
Chouser	cherry-pick can require a (usually interactive) conflict resolution, right?	19:30
	→ hugoxrosa joined	19:31
	→ CIA-7 joined	19:32
struberg	spearce I'm afraid I think we only talked about the use of DirCache.writeTree which I now already use for all the Index handling	19:32
	spearce what I mean: the WorkDirCheckout still needs the GitIndex. Any other way to do this?	19:33
spearce	no, we have to write that code over using DirCache and TreeWalk :-(	19:33
struberg	btw, today I did the first full maven release cycle with the maven-scm-provider-jgit :)	19:33
	ah oki, I just did like to make sure I'm not coding bullshit ;)	19:34
	← voker57__ left	19:35
	→ voker57__ joined	19:35
offby1	Chouser: I would assume so -- it's applying a patch, and there's always the risk that it won't apply	19:35
	← voker57__ left	19:35
radarek	Chouser: yes	19:35
Chouser	ok, thanks. I was hoping to make this multi-step thing a single command, but there's a cherry-pick in the middle	19:36
radarek	Chouser: what you are trying to do?	19:37
	→ saurabhverma joined	19:37
Chouser	automate as much as possible the steps needed to create and update tickets and patchs for a release-bugfix branch based on tickets and patches for the development branch	19:38
	→ tango_ joined	19:38
radarek	Chouser: uh, it always is hard thing:)	19:39
Chouser	so I want to point at a dev ticket and have my tool extract the commit from a comment, create a new ticket based on the old one, rebase (cherry-pick) the old patch onto the release branch, format as a patch and attach to the new ticket for review and approval.	19:40
	simple, right? :-)	19:40
radarek	Chouser: yeah :)	19:41
Chouser	my plan had been to do this on a single local working copy, but I think if I require two working copies (one for each branch) I won't need to move either off of their respective branches.	19:41
	→ BuZZ-dEE joined	19:42
Chouser	that way if the cherry-pick needs some conflict resolution, I can save off the ticket numbers, comments, and whatever other bits I need to pick up after the user (me!) has resolved the conflict.	19:42
	it may all be a tall order, but I have irrational optimism and long holiday weekend. Who's going to stop me? :-)	19:43
	→ rhett joined	19:45
	← adeodatus left	19:47
	← hpoydar left	19:52
	→ hpoydar joined	19:52
	← \ask left	19:53
	→ d0k_ joined	19:56
	← broonie left	19:56
	cpg → cpg\|biab	19:58
	→ khmarbaise_ joined	20:02
	← DavidKlein left	20:02
	← d0k left	20:03
	→ eno joined	20:03
	→ vasandgvd joined	20:04
	← khmarbaise left	20:04
	← Gabriel left	20:05
	→ tbf joined	20:05
	← SlipperyChicken left	20:07
	← saurabhverma left	20:09
	← mbroeker left	20:10
	← eno__ left	20:14
	→ drewr joined	20:14
	← aziz left	20:16
	→ aziz joined	20:16
	← vlt left	20:18
	← bonhoffer left	20:19
	← vasandgvd left	20:20
	← bentob0x left	20:21
	→ nekrad1 joined	20:21
nekrad1	hi folks, is there any way to efficiently use existing pack files from other repos (w/o ever touching them) ?	20:22
	→ Sho_ joined	20:25
	→ memiux joined	20:26
	← drizzd left	20:30
	← memiux left	20:31
	→ steveire joined	20:32
	← steveire left	20:32
	→ abw_ joined	20:32
	← abw_ left	20:32
	→ bx2 joined	20:33
	lut4rp → lut4rp_AFK	20:37
	← Weaselweb left	20:38
	← __iron left	20:38
	→ __iron joined	20:39
	← jrmuizel left	20:39
	← dfr\|work left	20:40
	← Krynnotaur left	20:41
	→ dfr\|work joined	20:43
	→ tvw joined	20:44
	→ majd joined	20:47
	← ericindc left	20:48
	→ drizzd joined	20:49
dereine	is it better to user git merge repo or git pull . repo	20:52
patrikf	dereine: pull = fetch + merge	20:53
BuZZ-dEE	hello, i want to submit my project to my sourceforge git account, but don't know how can i do that. i have read this https://sourceforge.net/apps/trac/sourceforge/wiki/Git	20:53
	and this https://sourceforge.net/apps/trac/sourceforge/wiki/Shell%20service	20:54
	can anyone help me	20:54
	?	20:54
patrikf	dereine: and pull is usually used for remote repositories, so anything that isnt "."	20:54
	→ dodo_the_last joined	20:54
	→ alip joined	20:54
	→ Krynnotaur joined	20:55
dereine	patrikf: thx nice	20:55
	i think local you don't need fetch :)	20:55
	← corni_ left	20:55
	→ corni joined	20:56
	← hpoydar left	20:56
patrikf	dereine: also, what you referred to as "repo" really is a branch	20:56
	→ hpoydar joined	20:56
patrikf	just to brush up your terminology :)	20:56
	cpg\|biab → cpg	20:56
	ryanakca_ → ryanakca	20:56
offby1	BuZZ-dEE: exactly what are you doing, exactly what do you see, and exactly how does that differ from what you expected to see?	20:57
	I've never used git on SourceForge but I'm guessing it's not terribly different from github	20:57
	← spaceonline left	20:57
	← Fullmoon left	20:58
BuZZ-dEE	offby1: i have used the Interactive Shell Service and then i have done this http://pastie.org/533653	21:00
	→ fundamental joined	21:00
offby1	Is that _exactly_ what you typed? I.e., you really typed "USERNAME"?	21:00
BuZZ-dEE	offby1: no	21:01
	;)	21:01
offby1	oh.	21:01
	← lbt left	21:01
BuZZ-dEE	my resl username and project	21:01
offby1	it'd probably be more useful to see _exactly_ what you typed, excluding passwords	21:01
	if you're going to edit history before showing it to me, that makes it less useful	21:01
dereine	patrikf: ups yes, i'm not too familiar with git yet :(	21:01
offby1	also, it appears you transcribed that information into pastie; it'd be better to show me an actual transcript of your shell session	21:02
BuZZ-dEE	offby1: okay ;)	21:02
patrikf	dereine: I guess you read one of the tutorials? never hurts to read another one. ;-)	21:02
	← mnick left	21:03
Mage_Dude__	Anyone have advice for creating public keys for Win machines so Eclipse can use git on a remote Linux server?	21:03
BuZZ-dEE	offby1: http://pastie.org/533656	21:03
	→ djgera joined	21:03
	← corni left	21:04
offby1	BuZZ-dEE: well, that's a slight improvement. But I still cannot see a) your shell prompt -- that is sometimes useful; b) the responses to those commands.	21:04
	→ corni joined	21:04
offby1	I really think it'd be a lot easier for you to simply copy and paste from your terminal, rather than transcribing.	21:04
BuZZ-dEE	offby1: okay	21:04
dereine	patrikf: i' currently reading gitready, quite cool stuff i didn't userd yet	21:04
offby1	Mage_Dude__: get the PuTTY suite, run PuTTYgen	21:05
BuZZ-dEE	offby1: http://pastie.org/private/nkqmdzxu5lzby1fppwvg	21:07
	offby1: http://pastie.org/private/f7cpcj7pljwnpkihflnkzq	21:07
	← fundamental left	21:07
	← s0ber left	21:08
	← hugoxrosa left	21:08
offby1	BuZZ-dEE: ok, now we're getting somewhere.	21:08
	hold on	21:08
	→ s0ber joined	21:09
offby1	hmm	21:09
	you seem to be doing everything they told you to do :)	21:09
BuZZ-dEE	what can i do?	21:10
offby1	It's possible that you simply need to make a dummy commit ... I'm not sure.	21:11
	let me try, since I too have a SF account	21:11
BuZZ-dEE	okay :)	21:12
offby1	OK, I'm trying this out just using local directories, and I'm seeing the same thing as you!	21:12
	That means the sf instructions are wrong.	21:13
	I suspect you simply need to make a single commit.	21:13
	Let's see if that's right ...	21:13
BuZZ-dEE	how?	21:13
offby1	yep	21:13
	touch it ; git add . ; git commit -m "first commit"	21:13
	you don't actually _have_ to do this	21:13
	you can wait until you have some real files.	21:13
	But this will get you going.	21:13
	← struberg left	21:14
Mage_Dude__	Gitosis question: When I add a key to the cloned version of the admin and then push to the master, it doesn't seem to show the keys on the system I pushed to?	21:14
offby1	BuZZ-dEE: you might see a big scary warning, too	21:14
BuZZ-dEE	offby1: where i have to do that? in my local shell?	21:14
offby1	Mage_Dude__: what do you mean by "show"?	21:14
	BuZZ-dEE: yes, the same shell whose transcript you showed me	21:14
BuZZ-dEE	okay	21:15
Mage_Dude__	offby1: I added 'key.pub' to the keydir and put the entry to .conf file. Committed and pushed changes to remote gitosis-admin but the keys are not present on the remote system?	21:15
	Mage_Dude__ → Mage_Dude	21:15
offby1	Mage_Dude: how do you know they're not present?	21:18
Mage_Dude	offby1: The directory doesn't contain them on the remote system.	21:18
BuZZ-dEE	offby1: is that a good message to me?	21:21
	http://pastie.org/private/wta8xawmzasrcnqr2a5g	21:21
	that was all i have to do?	21:22
Mage_Dude	I don't understand how using 'git push' doesn't update origin repo? It doesn't seem that I should have to manually add the keys to both the admin system and the remote system just add to admin repo, push, updated?	21:22
offby1	Mage_Dude: my hunch is: you're either looking in the wrong directory, or even the wrong computer :-) If "git push" didn't die, then it worked.	21:23
	← Chillance left	21:23
offby1	BuZZ-dEE: your latest paste appears to be the tail end of the output from a "git commit" command.	21:24
BuZZ-dEE	offby1: yes its only the end	21:24
offby1	BuZZ-dEE: I don't understand what you want to know about it, though.	21:24
	← warthog9 left	21:25
Mage_Dude	offby1: Shouldn't be wrong system. Using 'git remote show origin' asks for passphrase (the key for this system hasn't taken either) returns correct IP for correct machine and correct repo.	21:25
BuZZ-dEE	offby1: i want to submit my project to my sf-git and then i want import that in eclipse using the gitplugin for eclipse	21:25
	← hpoydar left	21:26
	→ hpoydar joined	21:26
offby1	BuZZ-dEE: you'll probably want to do "git push"	21:27
	to get your code onto sourceforge	21:27
	or, rather, to get it into the sourceforge repository that you're sharing, as opposed to the "private" one that you've been manipulating via the shell	21:28
Mage_Dude	Can you 'sudo cd' into a locked directory?	21:28
BuZZ-dEE	yes, if my code then is on sourceforge	21:28
offby1	Mage_Dude: I doubt I can help	21:28
BuZZ-dEE	offby1: into the sourceforge repository	21:28
offby1	Mage_Dude: what is a "locked directory"?	21:28
BuZZ-dEE	offby1: i want to use git instead of svn	21:29
	← hpoydar left	21:29
BuZZ-dEE	with eclipse	21:29
	but my SF Account is new, so i have to submitt my project too	21:29
Mage_Dude	Under the 'git' user which contains the gitosis main repo the directory structure has a 'repositories' directory. Trying to 'cd' into the directory gives 'permission denied'	21:29
	→ hpoydar joined	21:29
offby1	Mage_Dude: who are you logged in as?	21:32
	(anyway I don't think the public keys live under 'repositories')	21:32
Mage_Dude	offby1: As a matter of fact, the gitosis main repo is bare, and has no working directory. So there is no /keydir/ directory, but gitosis.conf does show that the change (adding another user) was made to the gitosis.conf file.	21:33
	← lorandi left	21:33
Mage_Dude	offby1: So changes are being made, I just must have my keys in the wrong place. I do have to manually add keys to the server system for them to be used?	21:34
offby1	nope	21:35
	take a look at ~git/.ssh/authorized_keys	21:35
	← sea-gull left	21:35
offby1	that's where I'd expect the public keys to wind up	21:35
	→ ajonat joined	21:35
	→ eldragon joined	21:37
	← hpoydar left	21:37
	→ hpoydar joined	21:37
	→ warthog9 joined	21:38
Mage_Dude	offby1: Hmmmm now we're getting somewhere. The authorized_keys file looks like a mess. Multiple entries for a single user, some span lines (not good)	21:38
offby1	that's perfectly normal	21:39
	that file is meant for SSH to read, not you.	21:39
Mage_Dude	offby1: Even though it says autogenerated by gitosis, do not edit, can I still edit it or do I need to learn certain gitosis commands to remove users?	21:39
	offby1: I've got 7 lines for one key and it starts with ---- BEGIN SSH2 PUBL..... then has the key.	21:40
	← radarek left	21:41
	spearce didn't know OpenSSH accepted --- BEGIN SSH2 --- style keys in authorized_keys	21:41
spearce	i thought you needed to convert those first	21:41
	← name left	21:41
offby1	Mage_Dude: I'm 98% sure there's a way to remove users without having to edit that file	21:41
	← kumbayo left	21:41
Mage_Dude	spearce: I bet you do (or just need to generate correct keys)	21:42
offby1	Everything should be doable by just editing your local copy of the admin repository, committing, and pushing	21:42
Mage_Dude	offby1: I think I can just replace the .pub with the correct key and it should update it on git push	21:42
offby1	I would think so.	21:42
spearce	ssh-keygen -i can do the convert in format, its fairly trivial	21:43
Mage_Dude	Are the first 28 characters supposed to be the same for all keys?	21:43
offby1	wouldn't surprise me	21:44
spearce	largely, iirc those first few characters contain the key type, "ssh-dss" or "ssh-rsa"	21:44
offby1	yeah, mine all look the same for the first buncha characters	21:44
spearce	despite it being at the start of an OpenSSH key line anyway, its embedded in the binary packet	21:45
	← tvw left	21:45
	← bcardarella left	21:45
	→ smotchkkiss joined	21:45
	← intripoon left	21:46
Mage_Dude	It doesn't matter if the keys are in different formats right? As long as they are all valid ssh keys...should be fine?	21:47
	→ mat69_ joined	21:47
offby1	should be	21:47
	← hpoydar left	21:49
Mage_Dude	Ok, looks better now. But git push still asks for passphrase from remote machine	21:49
	→ hpoydar joined	21:50
offby1	well, that's the usual ssh trouble	21:51
	it suggests that your private key isn't loaded into ssh-agent	21:52
Mage_Dude	Restart of sshd?	21:52
	cpg → cpg\|brb	21:53
	← rhett left	21:53
	→ leehambley joined	21:56
offby1	doubt that'd help	21:57
	I can't remember the last time I needed to start ssh	21:57
	← btanaka left	21:58
	← tongueroo left	21:59
Mage_Dude	I'm sure this would be easier if everyone was on Linux. But, of course, people 'want' what they think they need which leaves me to support and maintain the main linux based webserver, repos and backup in addition to those who want to work with XP/Vista/Win7 and OSX....client overkill?	21:59
	← hpoydar left	22:00
	← Alinermedia left	22:00
	← ntoll left	22:03
	← WebDragon left	22:03
	← smotchkkiss left	22:05
	→ the_undefined joined	22:06
	← candybar left	22:06
	← MattRich left	22:06
	← Mage_Dude left	22:07
	← the_undefined left	22:08
	← icwiener left	22:11
offby1	actually, git-over-ssh isn't that bad on windows; msysgit does a decent job. Has its own ssh client and everything	22:11
	the most recent download, though, is missing some DLLs	22:12
	so get an older one	22:12
	← gerberb left	22:14
	← patmaddox left	22:14
	← stefanb85 left	22:15
	→ bugfux joined	22:17
	→ camwest joined	22:17
	jackdempsey → jack-away	22:18
	→ Chillance joined	22:19
	← bx2 left	22:22
	← BuZZ-dEE left	22:23
	cpg\|brb → cpg	22:24
	← boto left	22:24
	← bugfux left	22:26
	→ wahjava joined	22:27
	← bradly left	22:30
	→ wagle joined	22:31
	→ mnick joined	22:34
	← Bennid left	22:35
	← Tuomas left	22:35
	→ candybar joined	22:36
	← Ryback_ left	22:37
	→ tongueroo joined	22:40
	← candybar left	22:41
	→ cloudhead joined	22:44
	→ bgerber joined	22:46
	→ ericindc joined	22:46
	→ Mage_Dude joined	22:47
	← mat69_ left	22:47
	→ candybar joined	22:48
	→ zimbres joined	22:53
	→ kennethlove joined	22:53
zimbres	hi, I would like to know a web-based free software hosting system with git support	22:54
	.	22:54
	← karmi left	22:55
	→ bugfux joined	22:56
	← jchris left	22:56
	→ jchris joined	22:56
offby1	github.com	22:56
	gitorious	22:57
	repository.or.cz	22:57
	oops	22:57
	repo.or.cz , rather	22:57
	← bugfux left	22:58
	→ edt joined	22:58
	← apeiros_ left	22:58
edt	something has gone wrong with my git tree. There are now hundreds of unreconzied files in the tree (which _ARE_ garbage). How do I tell git to remove them from the tree? (I tried git reset --hard v2.6.30)	23:00
Arrowmaster	git clean?	23:01
patrikf	edt: git clean	23:01
	← patrikf left	23:01
	→ jlanghorst joined	23:02
offby1	I do "git clean -dxf", which is a bit more dangerous, but works better :-)	23:03
	← jchris left	23:03
	→ jchris joined	23:03
	← jchris left	23:05
	→ jchris joined	23:07
	→ scarabx joined	23:08
	← sunoano left	23:09
	→ \ask joined	23:10
edt	how do I endup with untracked files doing:	23:11
	git clean -f	23:11
	← fr left	23:11
	← MarcWeber left	23:11
edt	git branch 30.1 v2.6.30	23:11
	git pull git pull git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.30.y.git	23:11
	git merge edac	23:11
	git merge btrfs	23:11
	at this point I have about 30 untracked files	23:12
	→ sunoano joined	23:12
edt	why?	23:12
bremner	edt: use a pastebin for that many lines please	23:12
	edt: did you check after git clean -f?	23:12
edt	I' try again	23:12
bremner	edt: check man git-clean, particularly the -x flag	23:13
Gitbot	edt: the git-clean manpage can be found at http://git.or.cz/man/git-clean	23:13
	→ fr joined	23:13
edt	bremner -x is interesting but the files are not build files...	23:14
	→ MarcWeber joined	23:14
	Tommy[D] → Tommy[D]_	23:14
	← wahjava left	23:15
edt	bremner need clean -fdx	23:15
	thanks	23:15
	← khmarbaise_ left	23:16
	← djgera left	23:17
b14ck	Can anyone recommend a good document for git beginners who want to get started managing their project with git?	23:17
	← dodo_the_last left	23:19
	← corni left	23:19
offby1	there's lots	23:20
	look on git-scm.org	23:20
	← Chillance left	23:20
	→ slyphon joined	23:20
	→ Chillance joined	23:20
	→ dodo_the_last joined	23:21
b14ck	thanks	23:24
	← juanjoc left	23:25
	→ reprore joined	23:25
b14ck	edt, http://github.com/comradeb14ck/wildpbx/blob/6487f0ca841f3549d3fa9de279c237794eaac853/libraries/agi/c/wpbx-cagi-internals.c line 426	23:26
Mage_Dude	Anyone use Eclipse on Windows and can successfully pull from a non-github repository?	23:26
b14ck	offby1, thanks	23:26
	← rettub_ left	23:27
	→ rettub_ joined	23:28
	← d0k_ left	23:28
	→ Tommy[D] joined	23:29
	Tommy[D] → Guest9318	23:29
offby1	Mage_Dude: I'm 99% certain that github versus not-github is irrelevant	23:29
	← Yuuhi left	23:29
offby1	if it works from the command line, it ought to work in eclipse	23:29
	← Guest9318 left	23:30
Mage_Dude	On a *nix based (non-mac) system, using 'git clone [email@hidden.address] still asks for passphrase even though user is in gitosis.conf and key.pub is in keydir/ On Windows, rather than ask for the user passphrase, Eclipse asks for the gitbot passphrase (which doesn't exist) rather than using the key provided by the user executing the command.	23:31
	← scarabx left	23:31
	← ajonat left	23:31
offby1	I suspect that both gitosis.conf and keydir are irrelevant.	23:31
	get a simple "ssh 192.168.2.28" "working" (i.e., failing with a complaint that GIT_ORIGINAL_COMMAND wasn't set in the environment, or something) first	23:32
	→ Tommy[D] joined	23:32
	Tommy[D] → Guest55945	23:33
	← edt left	23:33
Mage_Dude	I can ssh into the system just fine from win/nix/mac, but git isn't accepting win/nix/mac non-CLI clone commands. By ssh 192.168.2.28 'working' what do you mean other than just being able to ssh in?	23:34
offby1	mostly that.	23:34
	I'd be surprised, though, if you were able to get a shell by doing "ssh [email@hidden.address]	23:35
Mage_Dude	Yes, it fails with "ERROR:gitosis.serve.main:Need SSH_ORIGINAL_COMMAND in environment." because gitbot has no shell access.	23:36
	→ AkumaStreak joined	23:36
offby1	OK, that says to me that it is indeed working	23:36
	→ Artissimo joined	23:36
offby1	and you can get that behavior on Windows?	23:36
	(you might need to spell "ssh" as "plink.exe" or something :-)	23:37
	Guest55945 → Tommy[D]	23:37
	← Tommy[D]_ left	23:37
Mage_Dude	offby1: Eclipse provides no option for that kind of command	23:38
Artissimo	http://code.google.com/p/git-osx-installer/	23:38
	Is that the package needed for setting up a git server?	23:38
Mage_Dude	Artissimo: I've installed it earlier (thanks) but still having configuration issues and pulling from the main server repos	23:39
offby1	Mage_Dude: I don't even know what that means. I'm not suggesting you try Eclipse just yet; I'm suggesting you try running a command-line ssh client on Windows (in case you haven't already)	23:39
	← nekrad1 left	23:39
Artissimo	Mage_Dude: On leopard?	23:39
tomoj	Artissimo: you can compile git yourself if you like	23:40
Mage_Dude	Artissimo: 10.5.something? Yes? (Sorry I don't know my Mac 'codenames') for the OS's...	23:40
	offby1 can never remember the Debian/Ubuntu code names either, so once he learns 'em he sticks 'em into /etc/issue	23:42
Mage_Dude	offby1: Trying putty (using [email@hidden.address] still asks for passphrase	23:43
tomoj	Mage_Dude: it's supposed to ask for the passphrase, isn't it?	23:43
Mage_Dude	offby1: I don't even care that much about branding names. It's all Windows, Mac, Linux...much past that...boring details (for the most part) :)	23:43
	tomoj: No. By giving it a copy of the public key it should accept the connection without challenge.	23:44
tomoj	your key is passphrase-less?	23:44
offby1	tomoj: ideally, no: ideally one types one's passphrase into an ssh agent _once_, when you log in, and then don't need to type it again (until you log out)	23:44
bremner	MadCoder: err really gitbot? out of the box shouldn't that be gitosis@ or git@	23:44
offby1	bremner: I _think_ it's up to you	23:44
bremner	Mage_Dude: ^^	23:44
tomoj	Mage_Dude: so you really mean the password?	23:44
bremner	sorry M a d C o d e r	23:45
Mage_Dude	brpmner: Out of the box? No I followed a gitosis tut (scientist one I think...) and it allows you to create any user you want.	23:45
bremner	okey	23:45
	← majd left	23:45
	← mw left	23:47
tomoj	Mage_Dude: I mean, is it saying "Enter passphrase for ..." or "Password:"	23:47
Mage_Dude	tomoj: Right	23:48
	→ kwatz joined	23:48
tomoj	which one? :)	23:48
Mage_Dude	tomoj: Actually, on Nix it asks for passphrase, Win it asks for password	23:48
	→ mw joined	23:48
tomoj	so nix is working, win is offering the wrong key or you don't have the key in there	23:48
	Mage_Dude: maybe try giving your windows ssh client a -v flag?	23:49
	offby1 laughs cruelly	23:49
offby1	have fun figuring out a) how to do that in the first place; b) where the output goes when it's invoked by eclipse	23:50
	← bgerber left	23:50
tomoj	well I meant plink	23:50
	→ Derander joined	23:51
	→ bgerber joined	23:52
	← camwest left	23:52
	→ marcot joined	23:54
	→ gerberb joined	23:54
	← gerberb left	23:55
marcot	Hello. Does git push ssh://example.com/ uses the configuration in .ssh/config? It doesn't seem to be reading here.	23:55
bremner	marcot: yes, it should	23:55
marcot	bremner: Sorry, I'm missing something here.	23:56
offby1	marcot: it's possible that GIT_SSH is set to something other than "ssh"	23:57
	marcot: this is a good job for "strace" (or "ktrace" if you're on OS X or FreeBSD)	23:57
marcot	offby1: no, it's working.. I just noticed I was using another host this time, which wasn't set in .ssh/config.	23:57

Logs Search ←Prev date Next date→

Channels Documentation