IRCloggy #git 2010-06-22

Provider of IRC logs since 2005.

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

	2010-06-22
davr	they both do the same thing, one gets checked into the git repo, one does not	00:00
hagabaka	thanks	00:00
	← kenichi left	00:00
	→ tehbaut joined	00:01
hagabaka	.git/info/exclude is the one I want, since only my machine will have the file	00:01
	← tehbaut left	00:01
	← darwin_ left	00:01
	→ tehbaut joined	00:01
	← tehbaut left	00:01
	→ tehbaut joined	00:02
	← tehbaut left	00:02
	← drd20 left	00:03
	→ tehbaut joined	00:03
	← tehbaut left	00:03
	→ tehbaut joined	00:03
	← tehbaut left	00:03
	→ tehbaut joined	00:04
	← tehbaut left	00:04
	→ tehbaut joined	00:04
	← tehbaut left	00:04
	→ tehbaut joined	00:05
	← tehbaut left	00:05
	→ tehbaut joined	00:05
	← tehbaut left	00:05
	→ coolkehon joined	00:05
	→ tehbaut joined	00:06
	← tehbaut left	00:06
coolkehon	how do I update a git repo that I cloned ? I want to update it from the remote repo	00:06
	→ tehbaut joined	00:06
	← tehbaut left	00:06
mathepic	coolkehon: "git fetch" will download the new objects, "git merge origin" can merge them. You can do this in one step with "git pull	00:07
	← xlistener left	00:07
	→ tehbaut joined	00:07
	← tehbaut left	00:07
	→ Jordan_U joined	00:07
	→ tehbaut joined	00:08
	← tehbaut left	00:08
	→ stringoO joined	00:09
	→ tehbaut joined	00:09
	← tehbaut left	00:09
	→ tehbaut joined	00:09
	→ Aaaarg joined	00:13
	← werdan7 left	00:14
	← dreiss left	00:14
	← tanoku left	00:15
	← AngryParsley left	00:16
	← tedoc2000 left	00:17
	← damog left	00:17
	→ jrmuizel joined	00:17
	← mac_nibblet left	00:18
	→ tedoc2000 joined	00:18
	← tedoc2000 left	00:21
	← jtaby left	00:21
	→ csgeek joined	00:21
	← bthomson left	00:21
	→ bthomson joined	00:22
	→ werdan7 joined	00:22
	→ septic joined	00:23
	→ ivan joined	00:23
	← Jferro left	00:25
	← mathepic left	00:26
	→ sechrist_ joined	00:26
	← aziz left	00:28
	→ tedoc2000 joined	00:30
	← sechrist left	00:31
	sechrist_ → sechrist	00:31
	→ _numbers joined	00:35
_numbers	is thre a way to cherry-pick and preserve original author's name?	00:35
	they committed to the wrong branch	00:35
jhelwig	_numbers: Cherry picking does preserve the original author. It only changes the committer name.	00:35
_numbers	huh? git log only shows me the author and it says my name	00:36
	does it matter if there was a merge conflict that had to be resolved?	00:36
jhelwig	_numbers: git log --pretty=fuller	00:36
_numbers	ohh interesting	00:36
	unfortunately both say my name tho	00:36
	it may be because i had to fix a merge conflict before i could commit	00:36
jhelwig	Probably.	00:37
_numbers	maybe i just have to change my name in .gitconfig temporarily	00:37
jhelwig	_numbers: You shouldn't need to. You can tell commit to use different author/committer info.	00:38
	If the commit you're cherry picking is going to stick around, then it might be better to do `git cherry-pick -x ...`.	00:39
	Maybe not.	00:39
	← kingfishr left	00:39
	→ agile joined	00:39
_numbers	"Append the note only for cherry picks without conflicts."	00:39
jhelwig	Yeah, I was just going to explain the "maybe not" with that.	00:40
	→ kingfishr joined	00:40
jhelwig	_numbers: My recommendation would be to commit with --author='Other Person [email@hidden.address] when you commit after resolving the conflicts.	00:41
_numbers	jhelwig: ah thnx	00:41
jhelwig	It'll leave you as the committer, but show them as the author, which is probably a fairly accurate representation of reality.	00:42
	← frogonwheels left	00:42
	→ Jferro joined	00:43
engla	If Other Person did already appear in your history you can even abbreviate it as --author=Other	00:43
_numbers	oh sweet	00:44
	← Theaxiom left	00:46
	← sechrist left	00:50
	→ tedoc2000_ joined	00:50
	← tedoc2000 left	00:51
	← ignatp left	00:51
	→ Transformer joined	00:51
	← myosui left	00:53
	← Transformer left	00:53
	← tedoc2000_ left	00:56
	→ hohoho joined	00:56
	← Jferro left	00:56
	← Yuuhi left	00:57
	← carrus85 left	00:57
	← hohoho left	00:58
	← timj__ left	00:59
	→ adamv joined	00:59
	→ tedoc2000_ joined	01:01
	← rox` left	01:02
	→ hohoho joined	01:03
	← ponzao____ left	01:05
	← tedoc2000_ left	01:05
	→ bmac__ joined	01:06
	bmac__ → bmac	01:06
	← flaguy48 left	01:06
	← sh1mmer left	01:06
	← spaceonline left	01:07
	← agile left	01:08
TML	what's the difference between GIT_COMMITTER_DATE and GIT_AUTHOR_DATE?	01:08
	→ Heimidal_ joined	01:09
	← _numbers left	01:09
	← bmac_ left	01:09
	→ tedoc2000_ joined	01:10
mugwump	TML: commits have two attribution fields	01:11
	← csgeek left	01:11
mugwump	author is supposed to correspond to the person who wrote the change, commiter the person who applied it	01:11
	so that when you, eg, cherry pick, you don't "steal" the attribution from the original author	01:11
	← steveire left	01:12
	→ steveire joined	01:12
	→ dreiss joined	01:12
TML	mugwump: Got it, thanks. --pretty=fuller makes that a lot more clear	01:14
	→ agile joined	01:14
TML	Trying to turn an ad-hoc 'changlog' style history into a bunch of git commits - this is going to be a PITA :)	01:15
	Unode → Unode\|NA	01:15
	← beanluc left	01:16
mugwump	yes I've done that before	01:16
	→ justin-george joined	01:18
mugwump	http://git.catalyst.net.nz/gw?p=perl-history-massage.git;a=blob;f=chip-import.pl;hb=e14ec12bb8	01:18
	that was a script which ate the incremental patches left by Chip Salzenberg for Perl	01:18
	← Vampire0 left	01:19
mugwump	it started out a lot simpler, and grew features as the maintainers left more and more information	01:19
	→ csgeek joined	01:19
mugwump	it started out by just separating and aggregating changes by description	01:19
	and gradually gained features like author attribution, getting commit message information from mailing list archives using fuzzy matching, etc	01:19
	→ jtdowney joined	01:20
	← engla left	01:20
	→ hw1979 joined	01:23
	← tedoc2000_ left	01:24
	→ tedoc2000 joined	01:24
	→ VVSiz_ joined	01:26
	← spearce left	01:28
	← _VVSiz_ left	01:30
	← stamina_ left	01:31
	→ Vampire0 joined	01:32
	← mbroeker left	01:33
	→ chrislerum joined	01:34
	→ spearce joined	01:34
	← masterkorp left	01:35
	→ simplechat joined	01:36
	→ kazoo joined	01:36
	← spearce left	01:37
	→ spearce joined	01:38
	→ t0rc joined	01:39
	← madewokherd left	01:40
	→ madewokherd joined	01:41
	→ Nappy joined	01:42
	← spearce left	01:43
	← Nickeeh left	01:43
	← zenazn left	01:43
	← kazoo left	01:45
	← jarh left	01:45
	→ jarh joined	01:46
	→ srcerer joined	01:47
	shade_ → \shade\	01:49
	← bmac left	01:49
	← keyvan left	01:49
	→ bmac joined	01:50
	→ keyvan joined	01:50
	← keyvan left	01:52
	← gcristian_ left	01:52
	← mw left	01:52
	→ keyvan joined	01:53
	← herdingcat left	01:55
	→ thewoolleyman joined	01:59
	← Bass10 left	02:02
	← fmardini left	02:03
	→ anathematic joined	02:08
	← hohoho left	02:08
	← anathematic left	02:08
	→ fmardini joined	02:10
	→ anathematic joined	02:10
	→ Elfe joined	02:13
	← sagsousuke left	02:13
	← thewoolleyman left	02:14
	→ rlb3 joined	02:15
	← Elfe_ left	02:16
	→ welly joined	02:19
	← intripoon left	02:23
	← luke-jr left	02:23
	← cytrinox left	02:24
	← fmardini left	02:24
	→ cytrinox joined	02:26
	← drako left	02:27
	→ fmardini joined	02:28
	← harinath left	02:29
	→ sagsousuke joined	02:34
	← Durl left	02:34
	← Yuffster left	02:35
	← rlb3 left	02:35
	← Adaptee left	02:36
	→ Adaptee joined	02:37
	→ sixty joined	02:39
	→ Durl joined	02:40
	← jtdowney left	02:43
	→ harinath joined	02:44
	← sixty left	02:44
	← kukks left	02:46
	→ sixty joined	02:46
	← sixty left	02:46
	→ vire joined	02:48
	← vire left	02:56
	→ EddieRingle joined	02:56
	← Buglouse left	02:59
	← eletuchy left	02:59
	→ Transformer joined	03:00
	→ zenazn joined	03:00
	← Transformer left	03:01
	← zenazn left	03:02
	→ myosui joined	03:02
	← saiyr9000 left	03:03
	← apsysoev left	03:03
	← alester left	03:06
	← keyvan left	03:07
	→ keyvan joined	03:08
	→ ignatp joined	03:09
	← hw1979 left	03:10
	→ hw1979 joined	03:11
	→ herdingcat joined	03:11
	→ saiyr9000 joined	03:14
	→ jony joined	03:14
	← keyvan left	03:17
	→ Yuffster joined	03:17
	← hobodave left	03:18
	→ hobodave joined	03:19
	← ThiefMaster left	03:19
	→ ThiefMaster joined	03:23
	← hobodave left	03:24
	→ Transformer joined	03:24
	→ coderdad joined	03:26
	← Transformer left	03:26
	→ mw joined	03:28
	← jmcantrell left	03:30
	→ MattDiPasquale joined	03:32
	← mw left	03:33
	→ xlistener joined	03:33
	→ rayners joined	03:36
	→ keyvan joined	03:38
	→ rchavik joined	03:38
	← flazz left	03:41
	→ eletuchy joined	03:46
	← TDJACR left	03:47
	→ wjames5 joined	03:49
wjames5	hi, having trouble with a git remote i have on my own server, ssh keys seem to be failing	03:49
	ERROR:gitosis.serve.main:Repository read access denied	03:50
	fatal: The remote end hung up unexpectedly	03:50
	← t0rc left	03:50
	→ TDJACR joined	03:51
	← coppro left	03:51
wjames5	i run $ ssh-add -l on the workspace and it shows the path to the same public key that is in gitosis	03:51
	endeavour_ → endeavour	03:51
	→ coppro joined	03:51
wjames5	it stopped working out of the blue, seemingly	03:52
	← TDJACR left	03:53
wjames5	as a test if i ssh without -A i get no connection to agent, so ssh-agent seems to be working right. so what might cause gitosis to suddently freak. i cant clone or push	03:54
	→ TDJACR joined	03:56
	← tedoc2000 left	03:58
	→ hobodave joined	03:59
	→ hohoho joined	03:59
	← ignatp left	03:59
	→ codeshepherd joined	04:02
	→ phantomcircuit joined	04:03
	← MattDiPasquale left	04:07
	← phantomcircuit left	04:08
	→ phantomcircuit joined	04:09
	← traviscline left	04:10
	→ traviscline joined	04:10
	← chrislerum left	04:11
	← scarabx left	04:13
	→ frogonwheels joined	04:13
	← aresnick left	04:13
	← Durl left	04:14
	→ Durl joined	04:14
	→ ExtraSpice joined	04:14
	← coderdad left	04:15
	← harinath left	04:16
	← marc_kdab_ left	04:16
	→ scarabx joined	04:18
	→ consolers joined	04:18
	← endeavour left	04:19
	→ enherit joined	04:20
	← pengo left	04:22
	markatto → Tyler_Durden	04:29
	→ harinath joined	04:32
	← scarabx left	04:33
	← agile left	04:34
	→ scarabx joined	04:36
	→ resmo joined	04:39
	→ Phrogz joined	04:41
Phrogz	Why does git status incorrectly show my branch as differing from the master until I do a do-nothing push?	04:41
	http://pastie.org/1014691	04:41
	This happens consistently with this one repository; others in the network do not exhibit this behavior.	04:42
	(This one repository also happens to be unique in terms of the OS and git version involved.)	04:42
	→ codeshepherd_ joined	04:46
	→ agile joined	04:46
	← codeshepherd left	04:47
	codeshepherd_ → codeshepherd	04:47
	← snuxoll_ left	04:48
	→ snuxoll joined	04:48
	→ AdamVancouver joined	04:49
	→ radioman-lt joined	04:49
	Tyler_Durden → markatto	04:49
	← krawek left	04:50
	→ ige joined	04:51
	← ige left	04:51
	→ rajeshsr joined	04:55
	← Durl left	04:56
rajeshsr	hi all	04:56
	when i do a series of commit. How am i to retrieve the information about each commit?	04:56
consolers	git log for all your needs	04:57
Phrogz	rajeshsr: git log -3 for the last 3	04:57
rajeshsr	thanks!	04:58
	how do i refer to a particular commit and create a branch with that.	04:59
	← traviscline left	04:59
	← Twisted_Mentat left	04:59
	→ traviscline joined	05:00
rajeshsr	the thing is am confused with tags, a particular commit. How do you use a tag, if you can rollback to a commit?	05:01
	← jonmasters left	05:01
	← JackWinter left	05:02
	← traviscline left	05:02
	← Mage left	05:02
	← NfNitLoop left	05:03
	← eletuchy left	05:03
	← jony left	05:04
	→ traviscline joined	05:04
AdamVancouver	The tag will still point to that commit	05:06
	→ freimatz joined	05:07
Phrogz	rajeshsr: See the crazy long hash? That's your unique handle to a commit.	05:07
	rajeshsr: http://www.kernel.org/pub/software/scm/git/docs/git-branch.html	05:08
	→ marc_kdab_ joined	05:08
	← harinath left	05:09
	← wjames5 left	05:10
	→ apsysoev joined	05:12
	→ jonmasters joined	05:13
	← hyperair left	05:17
	→ hyperair joined	05:18
	→ Durl joined	05:19
	→ AAA_awright_ joined	05:20
	← codeshepherd left	05:20
	→ Zelphyr joined	05:20
	← Phrogz left	05:22
	→ devanjan joined	05:23
	← Durl left	05:23
	Zelphyr → Durl	05:23
	← AAA_awright left	05:24
	→ harinath joined	05:24
	← xlistener left	05:25
	← jrmuizel left	05:27
	→ codeshepherd joined	05:27
	→ johnw joined	05:28
	← stringoO left	05:29
	→ amerine joined	05:29
	→ Nickeeh joined	05:30
	← ChUrP left	05:32
	→ AndChat- joined	05:35
	← icefox left	05:36
	→ eletuchy joined	05:36
	← orafu left	05:36
	→ jrmuizel joined	05:36
	→ roop joined	05:37
	→ orafu joined	05:37
	→ necromancer joined	05:37
	← madewokherd left	05:37
	← AdamVancouver left	05:38
	→ sivakumart joined	05:41
	← sivakumart left	05:42
	→ sivakumart joined	05:43
	→ Vampire0_ joined	05:47
	← adamv left	05:48
	← hagabaka left	05:48
	← MasterLambaster_ left	05:50
	← Vampire0 left	05:50
	← loinclot_ left	05:51
	→ NfNitLoop joined	05:51
	← Cesario left	05:52
	← AndChat- left	05:52
	← kenneth_reitz left	05:52
	→ AdamVancouver joined	05:53
	→ sondreb joined	05:53
	→ lightcap joined	05:53
	→ Mage joined	05:54
	← Mage left	05:54
	→ Mage joined	05:54
Bombe	I have a commit that was cherry-picked from another branch. How do I find the original commit?	05:54
	→ AndChat- joined	05:56
	← henkboom left	05:56
	→ codeshepherd_ joined	05:57
	← sagsousuke left	05:57
	→ henkboom joined	05:58
	← AdamVancouver left	05:58
	← sondreb left	05:59
	→ Cesario joined	06:00
	→ sondreb joined	06:01
	← jfkw left	06:01
	→ MattDiPasquale joined	06:01
	← codeshepherd left	06:01
	codeshepherd_ → codeshepherd	06:01
	→ ph^ joined	06:02
	← doener left	06:02
	→ Reticenti joined	06:02
Reticenti	how would I clone just this directory? http://android.git.kernel.org/?p=platform/frameworks/base.git;a=tree;f=data/fonts;hb=HEAD	06:03
	← Durl left	06:03
	→ mtkd joined	06:03
	← AndChat- left	06:03
	→ AdamVancouver joined	06:03
	→ doener joined	06:04
	← amerine left	06:04
	← Yuffster left	06:05
	← AdamVancouver left	06:05
	→ AdamVancouver joined	06:06
	← necromancer left	06:06
rane_	you can't clone specific directories	06:08
	→ navetz joined	06:09
	← frank left	06:11
	← Reticenti left	06:12
	→ girishr joined	06:13
	← sondreb left	06:13
	→ cbreak_work joined	06:16
	← scarabx left	06:16
	→ candybar joined	06:16
	→ sondreb joined	06:18
	← leavittx left	06:18
consolers	no	06:19
	→ gebi joined	06:19
	→ jony joined	06:22
	← eletuchy left	06:22
	← lightcap left	06:23
	metze_away → metze	06:25
	→ damog joined	06:26
	← tomsdale left	06:27
	→ psankar joined	06:28
	← Textmode left	06:30
	← damog left	06:30
	← brosner left	06:30
	→ Reticenti joined	06:30
	→ Belna joined	06:30
	← seezer left	06:30
	→ seezer joined	06:30
	→ tildeequals joined	06:32
	→ broonie joined	06:33
	→ poseid joined	06:33
	→ ntoll joined	06:34
	← jrmuizel left	06:35
	→ tvw joined	06:37
	→ vu3rdd joined	06:38
	← jony left	06:38
	← keyvan left	06:40
	→ john__ joined	06:40
	→ lightcap joined	06:40
	→ sgh joined	06:40
	← teukka left	06:40
	→ teukka joined	06:40
	← john__ left	06:41
	→ keyvan joined	06:41
	→ PsWii60 joined	06:44
	← consolers left	06:44
	← PsWii60 left	06:46
	← hobodave left	06:46
	→ jony joined	06:48
	← dreiss left	06:49
	← gebi left	06:50
	→ urkud_ joined	06:51
	→ giskard joined	06:51
	← AdamVancouver left	06:51
	→ airborn joined	06:51
	→ ChUrP joined	06:52
	← nfrs left	06:52
	← Espen-_- left	06:52
	→ Espen-_- joined	06:53
	← rayners left	06:54
	→ dpino joined	06:54
	→ darwin_ joined	06:54
	← Mage left	06:56
	← codeshepherd left	06:58
	← Espen-_- left	06:58
	← lightcap left	06:58
	← poseid left	06:59
	→ thiago_home joined	06:59
	← harinath left	07:00
	→ alexpercsi joined	07:03
	→ midgetspy joined	07:03
alexpercsi	hi all. i need github access for my pc on windows, but i can't seem to find an ssh-keygen tool	07:04
	anyone know where i could find a good one?	07:04
midgetspy	is there a way I can change branches with outstanding changes and commit the changes to a different branch?	07:04
alexpercsi	midgetspy: git stash	07:04
	← dionoea left	07:04
	→ dionoea joined	07:04
	← plediii_ left	07:04
alexpercsi	then checkout another branch, then git stash apply	07:04
midgetspy	thanks	07:05
	and for your question, use putty	07:05
alexpercsi	midgetspy: no problem. not sure putty will work because i can't set the email address as a parameter	07:05
	in their ssh keygen tool i mean	07:05
midgetspy	I used putty for my github access... though I don't really remember the steps, I think I just followed a howto online	07:06
alexpercsi	alright midgetspy, i'll give it a closer look. thanks	07:06
DrNick	actually, you don't need stash for that, just checkout the new branch and commit	07:06
alexpercsi	DrNick: really? i thought that only works for untracked files	07:07
	→ DavidKlein joined	07:07
midgetspy	ya it won't let me switch branches with changes	07:07
DrNick	it'll work as long as none of the changed files are different in the new branch	07:07
alexpercsi	interesting	07:08
	→ lightcap joined	07:08
	adamholt → adamholt_away	07:09
	adamholt_away → adamholt	07:09
	→ skoop joined	07:09
	← lightcap left	07:10
	→ [Po]lentino joined	07:10
	← navetz left	07:11
	→ lucsky joined	07:12
	→ jankees joined	07:13
jankees	Hi Git people, I have been using git for a few months now and I must say I really like it	07:14
	→ dytrivedi_lunch joined	07:14
jankees	I have a question about submodules... I managed to create a submodule to another submodule, that works just fine	07:14
	→ pantsman joined	07:14
	dytrivedi_lunch → dytrivedi	07:14
	← phantomcircuit left	07:15
jankees	But what if I want a submodule of a subfolder in another repository (let's say repos/bin) is there a way to do this?	07:15
	← jony left	07:16
	→ lhz joined	07:16
	← hyperair left	07:17
	→ harinath joined	07:17
DrNick	not really	07:18
	wait, no, there isn't	07:18
	→ _Marek_ joined	07:18
_Marek_	hi all	07:18
	i'd like to ask whether it's possible to import several files which represent different versions of the same program into git. Program is a one-source-file program.	07:19
	→ hyperair joined	07:19
	← agile left	07:20
	← giskard left	07:20
	→ agile joined	07:21
jankees	DrNick: Where you talking to me? Is there no way to do this? I wonder how other people solve this problem	07:21
	dwmw2_gone → dwmw2	07:22
	→ codeshepherd joined	07:23
	← johnw left	07:23
	→ priidu_ joined	07:23
	→ jony joined	07:24
	← vinse left	07:27
	→ dj_tjerk joined	07:28
	deavidsedice → deavid	07:30
	← robinbowes left	07:31
	→ tizbac joined	07:32
	← Reticenti left	07:33
	← hyperair left	07:34
	← iizukaway left	07:34
	→ iizukaway joined	07:34
	→ steffkes joined	07:34
	Altreus → Altreus[hireme]	07:35
	Altreus[hireme] → Altreus	07:35
ptor	jankees: I don't know much about submodules, but if it's a bit intricate (as you indicate) you could ask on the mailing list. Sometimes, if it's not possible to do what you originally suggested, someone will come up with an alternative way to achieve the same goal.	07:39
	← timyhac left	07:39
_Marek_	basically i have program_v1.c program_v2.c program_v3.c and i'd like to import them as revisions of one file program.c	07:39
jankees	ptor: okay, thanks	07:39
	ptor: i will try to google a lot first	07:39
ptor	right	07:40
thiago_home	_Marek_: git init; mv program_v1.c program.c; git add program.c; git commit -m "Version 1"	07:40
	_Marek_: mv program_v2.c program.c; git commit -a -m "Version 2"	07:40
	_Marek_: etc	07:40
_Marek_	thiago, thanks :)	07:40
	→ hyperair joined	07:42
	→ mastro joined	07:42
	→ chrislerum joined	07:45
	→ lgd_ajax joined	07:46
	← chrislerum left	07:47
	← dahankzter left	07:49
	→ timyhac joined	07:51
	← Belna left	07:51
	→ Belna joined	07:52
	→ giskard joined	07:52
	← codeshepherd left	07:53
	→ nevyn^_ joined	07:55
	→ ige joined	07:56
	← freimatz left	07:56
	← brundle left	07:56
	→ freimatz joined	07:56
	→ cilly joined	07:57
	→ codeshepherd joined	07:58
	← freshtonic left	07:59
	→ dahankzter joined	07:59
	→ psoo joined	08:00
	→ petrux joined	08:00
	→ nevyn^__ joined	08:00
	→ shruggar joined	08:01
	← nevyn^_ left	08:03
	→ Paraselene_ joined	08:08
	→ brundle joined	08:10
	→ froschi joined	08:10
	→ barismetin joined	08:12
	→ donttrustem joined	08:14
	→ tatsuya joined	08:14
	← raorn left	08:15
	→ raorn joined	08:15
	→ _Vi joined	08:15
	→ andreaaa joined	08:17
	← nevyn^__ left	08:17
	→ nevyn^_ joined	08:17
	← brundle left	08:17
	← anathematic left	08:17
donttrustem	hi guy's my dev who setup git has vanished and I have managed to workout how to clone and commit and push but when I logon to the main server to commit the changes to the production src I get this error git commit fatal: Not a git repository (or any of the parent directories): .git	08:18
	← andreaaa left	08:19
FauxFaux	It's probably a bare repo, not a working copy. You should only be pushing to a bare repo.	08:19
	→ andreaaa joined	08:19
	← pantsman left	08:20
donttrustem	I pushed the changes fine from my local copy	08:21
FauxFaux	To where?	08:21
	← thiago_home left	08:21
FauxFaux	git remote -v # will probably show somewhere as "origin (push)".	08:22
donttrustem	http://pastebin.com/Ld881D1N	08:22
	← jankees left	08:23
FauxFaux	One of us is misunderstanding something. ¬_¬	08:23
	← Aaaarg left	08:24
donttrustem	FauxFaux: OK .. I have a local clone of the directory	08:24
	I make my changes and then commit and then push to the remote server	08:24
	→ didlix joined	08:25
donttrustem	I now need to commit those changes to the prod source code but I am can't seem to do this .....	08:25
	← alexpercsi left	08:25
	→ necromancer joined	08:25
	→ hebz0rl joined	08:25
donttrustem	the issue is I didn't set this up so I am not sure how I commit on the prod server	08:26
	do I just do a commit	08:27
	[root@/datastore/directory.git]:git commit	08:27
	fatal: This operation must be run in a work tree	08:27
	← werdan7 left	08:28
donttrustem	FauxFaux: I am not sure what directory i need to be in to commit the changes	08:28
	← lgd_ajax left	08:28
	← yaotti_aw left	08:28
donttrustem	FauxFaux: http://pastebin.com/TMd5A9g2	08:29
	→ ThunderChicken joined	08:29
	→ poseid joined	08:29
	→ codeshepherd_ joined	08:29
	→ trivol joined	08:29
	← jony left	08:29
FauxFaux	donttrustem: directory.git is /not/ a working copy. It's just a block of changes. There are no files you recognise in directory.git.	08:29
	Committing etc. make no sense without a working copy.	08:30
	← raorn left	08:30
donttrustem	FauxFaux: OK ... so I need to commit to the directory them?	08:30
	→ rayners joined	08:30
FauxFaux	Commit moves changes from a working copy into the history. Your changes are already in the history or you couldn't have pushed them. Stop trying to commit stuff.	08:31
	← MattDiPasquale left	08:31
FauxFaux	What you probably want to do is find a checkout and git pull from it.	08:31
	→ werdan7 joined	08:32
donttrustem	FauxFaux: Ah so I now need to pull the changes .... I see, well I think I do	08:32
	FauxFaux: let me try an do a pull	08:32
	← necromancer left	08:32
	← codeshepherd left	08:33
	codeshepherd_ → codeshepherd	08:33
donttrustem	FauxFaux: thanks man ... that worked :)	08:33
	all changes are there :)	08:33
	FauxFaux: sorry for acting dumb but I have only been looking at for 30mins just got stuck	08:34
	← alezandro left	08:35
	← myosui left	08:36
	→ thierryp joined	08:37
	→ jony joined	08:37
	→ necromancer joined	08:38
	→ MattDiPasquale joined	08:38
	← MattDiPasquale left	08:40
	→ raorn joined	08:42
	→ Jaxan joined	08:44
	→ leavittx joined	08:45
	← codeshepherd left	08:45
	← project2501a left	08:47
	→ d0k joined	08:50
	→ saidi joined	08:54
	← agile left	08:54
	→ Aaaarg joined	08:55
	→ agile joined	08:56
	→ curtana joined	08:56
	← ChUrP left	08:56
	← AlexP left	08:59
	← herdingcat left	08:59
donttrustem	FauxFaux: hi sorry to be a pain but I have just tried to push another change from my clone to the remote server and I get this error [rejected] master -> master (non-fast-forward)error: failed to push some refs to 'ssh://mark@costaeye.com/datastore/directory.git'To prevent you from losing history, non-fast-forward updates were rejected Merge the remote changes before pushing again. See the 'Note about fast-forwards' section of 'git push --help' f	09:01
	details.	09:01
	FauxFaux: I am not sure how I do a merge	09:01
	← sivakumart left	09:03
FauxFaux	It wants you to pull the existing changes down before pushing new ones up.	09:03
donttrustem	thanks	09:03
	→ dajero joined	09:04
	→ hobodave joined	09:05
	← rayners left	09:05
	→ bluenovember joined	09:05
	→ TheNewAndy joined	09:06
	← freimatz left	09:06
	→ freimatz joined	09:07
	→ Lgb joined	09:07
	→ corni joined	09:07
	← roop left	09:07
shruggar	that message really should say "the 'NOTE ABOUT FAST-FORWARDS' section"	09:08
	← agile left	09:09
	→ FND joined	09:11
donttrustem	oops! I broke it http://www.costaeye.com/directory/	09:11
	→ agile joined	09:11
donttrustem	how do I rollback :S	09:11
bob_f	Haha.	09:11
FND	I often use `git log master..contributor/branch` -- is there a simple way (other than `git merge --squash`) to get a combined diff of all the "new" commits on that contributor's branch?	09:12
donttrustem	I found the problem :)	09:12
coppro	git diff master contributor/branch	09:12
	or maybe the other way round	09:12
FND	coppro: that only works if master hasn't moved on since the contributor branch was created	09:12
mapreduce	donttrustem: git help revert if you're talking about rolling back to a previous commit, otherwise git help reset and git help checkout	09:13
FND	otherwise you get a whole lot of unrelated chunks	09:13
	← justin-george left	09:13
coppro	oh, so you want to do it from the branch point then?	09:13
	hmm, I dunno	09:13
	→ project2501a joined	09:13
	← necromancer left	09:14
ptor	master...contributor/branch could do it	09:15
	as in 'git diff' and that range.	09:16
FND	ptor: that actually seems to work - I only ever tried with two dots, not three	09:16
coppro	ah, yes	09:16
	← enherit left	09:16
coppro	read the man page better then ;)	09:16
FND	hehe	09:16
	thanks guys	09:16
donttrustem	thanks guy's all fixed	09:16
ptor	FND: They're quite different in behaviour when there are checkins on both branches	09:17
	← mapreduce left	09:17
FND	ptor: which man page would I consult (Git has many, and I'm not always sure where to look... )	09:17
	← apsysoev left	09:17
coppro	man git diff	09:18
	→ strk joined	09:18
	← harinath left	09:18
strk	do tags get pushed ?	09:18
	or are them only local	09:18
ptor	FND: For the range stuff? Not certain.. all docs I ever saw on that (including external docs) only confused me. So I just keep my own rule-of-thumbs for them.	09:18
FND	coppro: already looking at that, I was thinking more in general though	09:18
shruggar	FND: man git-rev-parse	09:18
Gitbot	FND: the git-rev-parse manpage can be found at http://git.or.cz/man/git-rev-parse	09:18
coppro	start with the command you want	09:18
FND	that sounds like what I'm looking for	09:18
coppro	rev-parse is another good one	09:19
	FND is satisfied	09:19
FND	again, thank you	09:19
coppro	(note that git diff's use of ranges is different than git-rev-parse)	09:20
rajeshsr	am using git to manage the source code i have. A co-worker of mine uses something else. If I make a change here and send him the output of git diff, is that enough for him to patch it in his tree?	09:20
coppro	yes, but then you're Doing It Wrong	09:20
	set up the repostories as remotes to each other so that you can pull patches	09:21
shruggar	rajeshsr, usually, but git format-patch is the preferred method. Also, we have no idea, since that's entirely dependent on what /he/ is using	09:21
coppro	that's the wonderful thing about distributed reversion control	09:21
rajeshsr	shruggar, assume that, he has something that will make patching easier for him. All i need is to give him the diff..	09:22
shruggar	rajeshsr, we assume that, but we don't know what format his patching tool expects. Usually giving someone a diff is enough, but that is true in all situations, so git isn't even involved there.	09:23
ptor	rajeshsr: If he knows how to use GNU patch he shouldn't have any problems, after looking at the diff (e.g. will usually have to use 'patch -p1', for example)	09:23
strk	how should one share tags ?	09:23
	I see 'git tag' only affects the local .git, which doesn't get pushed	09:23
shruggar	rajeshsr, if you're working with binary files, you may need to say git diff --binary, but then we don't know if his patching tool supports that, for example. He can use "git apply", even if he doesn't use git for anything else, as well	09:24
	strk: git push --tags	09:24
rajeshsr	shruggar, he is using GNU patch. So a GNU diff is the expected output.	09:24
shruggar	strk: or, less-sledghammery: git push remote the-tag	09:24
	→ name joined	09:25
rajeshsr	shruggar, i see no output when i give: git format-patch	09:25
	name → Guest51002	09:25
shruggar	rajeshsr, then git diff --binary should be enough. Though really, what excuse exists anywhere in the world for not using git? :)	09:25
	← Guest51002 left	09:26
	→ Guest51002 joined	09:26
ptor	inertia	09:26
	Guest51002 → name	09:26
rajeshsr	haha, well, thats a different problem to solve! It is that he has been into comfort-zone of something else!	09:26
	Even am new to git to make him take it up!	09:27
	→ ncopa joined	09:27
rajeshsr	So, after 3 or 4 months, i can say what benefit he could get out of it..	09:27
	→ level09 joined	09:27
level09	what is exactly a conflict ?	09:28
	← Jaxan left	09:28
	← Aaaarg left	09:28
level09	sometimes I do minor changes and I still get a conflict	09:28
	why is that	09:28
ptor	rajeshsr: git format-patch <from>. E.g. if you keep your co-worker's current version in your local branch 'co-worker' then it'll be git format-patch co-worker (or -M co-worker to track moves). Replace with 'master' if you keep your common version in 'master' and you yourself are working from a branch. Then do the format-patch when checked out on your branch.	09:29
shruggar	level09, what command are you running when you get a conflict?	09:29
level09	merge	09:29
	git merge	09:29
	when i merge branches	09:29
	→ anathematic joined	09:29
	→ engla joined	09:29
	← engla left	09:30
	→ engla joined	09:30
ncopa	hi, i have an autobuild script that git pull's and build new packages and uploads those	09:30
	I'd like to: for each tag since last pull, do git checkut, build new packages and build isoimage. finally git pull and build+upload packages	09:30
shruggar	level09, a conflict occurs when both sides of the merge have edited the same part of the file, and those changes could not be automatically integrated. If you are seeing conflicts even when you think the changes involved are far away from eachother, it is possible you are saving files with the wrong line-endings	09:30
level09	shruggar: by part of file you mean line number / location ?	09:31
ncopa	any ideas how to do the for-each-tag-since-last-git-pull-do-git-checkout-and-something	09:31
rajeshsr	ptor, thanks! That creates a patch file. Isn't it same as git diff?	09:31
level09	shruggar: I think you're right, I get loads of conflicts when getting files into my win machine from linux	09:31
	← SRabbelier left	09:32
shruggar	level09, you could check output from "git log -p" to see if you have any diffs which look like a whole-file being deleted and re-written, though certain line-ending-related config options could hide this error	09:33
	→ roop joined	09:33
	→ harinath joined	09:33
level09	great thanks	09:33
saidi	you may use git config core.autocrlf false	09:34
	shruggar is pretty sure he despises the concept of core.autocrlf	09:34
level09	shruggar: git diff is not really that convenient , is there a better way to show changes between files	09:35
bob_f	Hi, I'm using gitolite but I can't find any reference to any hosts in my config file, so I can't figure out where I'm pushing to.	09:36
	But push works.	09:36
	← strk left	09:36
ptor	rajeshsr: Format is basically the same (i.e. would have to be applied with 'patch -p1' if applied at top, or just use 'git apply', which also works outside of a git repo). Main difference is that 'git diff' will make a diff against your last commit vs. the working tree (non-commited files), while git format-patch is only working on commited changes.	09:36
bob_f	I guess I could just push to see where it's going, but I can't figure out where this value is actually stored.	09:36
	Oh, wait, ~/.ssh/config ?	09:37
sitaram	bob_f: what does "git remote -v" say?	09:37
rajeshsr	ptor, oh, ok. Thank you very much.	09:37
bob_f	sitaram:	09:37
	origin gitolite:gitolite-admin (fetch)	09:37
	origin gitolite:gitolite-admin (push)	09:37
	→ jatt joined	09:37
sitaram	bob_f: ok then look for "host gitolite" and a few lines after that in ~/.ssh/config	09:38
bob_f	sitaram: Yep, got it. Thanks - feeling a bit worse for wear this morning. :)	09:38
	← name left	09:38
	→ masterkorp joined	09:38
	→ SRabbelier joined	09:39
bob_f	sitaram: And I can add users to gitolite-admin/conf/gitolite.conf for `repo gitolite-admin` and they'll then be able to push then too ?	09:39
sitaram	yup	09:39
bob_f	Or will they need to do some init stuff also ? This is something I remember not understanding when I set this up initially; how to set up other repo admins.	09:39
sitaram	bob_f: (1) you add them to your clone/conf/gitolite.conf (2) you git add/commit/push from your admin clone so the server gets the change (3) your friend does a "git clone git@server:gitolite" and makes changes and pushes	09:40
	← airborn left	09:41
	→ airborn joined	09:41
bob_f	sitaram: Great - thanks.	09:41
	← dajero left	09:41
	→ jtdowney joined	09:42
	← masterkorp left	09:44
	→ strk joined	09:45
strk	what's the best practice to mix git and CVS ?	09:45
	for single-user it seems fine here, but when sharing, I get different dates in Entries on each 'cvs update'	09:45
	ie: different hosts get different dates	09:45
	days of gaps, not seconds	09:45
thiago	strk: the best practice is to convert to Git :-)	09:46
	← hyperair left	09:46
spb	generally speaking, the best practice for any situation involving CVS is to ditch it as quickly as possible	09:46
thiago	and don't version-control the control files of another VCS	09:46
strk	thiago: not an option. I'm including a separate project, managed by CVS	09:47
	into our own development line	09:47
	so I've my version of the code, and want to track the CVS version	09:48
thiago	do not version-control the CVS control files	09:48
strk	for easily producing patches	09:48
engla	what a mess..	09:48
strk	thiago: if I do that then each developer will need to re-setup the CVS control files	09:48
	to be able to produce those patches for upstream	09:48
thiago	strk: correct	09:48
shruggar	convert upstream to git, run git-cvsserver? :)	09:49
strk	it's a pity, cause the only trouble so far are the dates in the Entry files	09:50
	→ hyperair joined	09:50
strk	for some unknown reason I get old dates when runnign cvs update on my system	09:50
	← jfine left	09:50
strk	'Entries', that is	09:50
	it's nice to have all references in the git repo for the rest	09:50
	Root, Repository, Tag...	09:50
thiago	that's because you're placing under version control files that are changed by the normal running of tools	09:50
	don't do that	09:50
ncopa	if i do git fetch && git merge, is there some way to detect if there are any new tags in remote repo before i do the git merge?	09:50
thiago	ncopa: git fetch will tell you the tags it fetched	09:51
	→ masterkorp joined	09:51
ncopa	but those affects otehr brances too	09:51
thiago	ncopa: correct	09:52
	ncopa: tags aren't in branches	09:52
ncopa	I'd like from a script do soemthing with each tag	09:52
thiago	ncopa: tags belong to the entire repository	09:52
ncopa	tags that affects this branch only	09:52
	lets say I'd like create a tarball of the repo and upload some place	09:53
shruggar	"tags referring to commits which are contained by this branch and no other branches" ?	09:53
	← steffkes left	09:54
ncopa	yes	09:54
	← walf443 left	09:54
	→ pellera joined	09:54
ncopa	since I'd like to checkout the tag	09:54
shruggar	ncopa, regarding "create a tarball": "git archive sometag"	09:54
ncopa	i need build stuff from tag before i package it	09:55
	→ walf443 joined	09:56
	← jhelwig left	09:56
ncopa	what im trying to do is build a new release of a linux distro from a script	09:57
	I'd like to just create a tag and push	09:57
	→ steffkes joined	09:57
ncopa	and the autobuilder generates new iso image	09:57
	the problem is if i tag and adds 3 more commits and then push	09:57
	then will the iso image not correspond to the tag	09:57
	← froschi left	09:57
	← hw1979 left	09:59
ncopa	i proably need do something with git rev-list	09:59
Ilari	ncopa: git rev-parse ?	09:59
shruggar	ncopa, how about whenever a new tag is pushed, check-out that tag, build it, etc? If you only want certain tags to be auto-built, then have a naming convention for tags	09:59
	← strk left	10:00
ncopa	i dont want manually checkout and build (which i do now)	10:00
	i want it be automatic	10:00
	so im working on that script	10:00
shruggar	ncopa, how about whenever a new tag is pushed, Have A Script check-out that tag, build it, etc? If you only want certain tags to be auto-built, then have a naming convention for tags	10:01
Ilari	ncopa: The script could also export a snapshot, unpack it somewhere and build.	10:01
	→ gnufied joined	10:02
ncopa	hum	10:02
	for each git push i already git pull and build the changes	10:02
Ilari	ncopa: Using git pull in scripts is a bad idea.	10:03
shruggar	ncopa, so do it for tags as well. What's the actual issue you're having?	10:03
ncopa	detect when there is new tag	10:03
	lunch. brb	10:04
	← walf443 left	10:04
shruggar	as was said before, git-fetch will tell you about new tags it fetched	10:05
Ilari	ncopa: Also in post-update hook, updated/new tags are visible in list of refs updated.	10:06
	New way to abuse remote helpers: As client-side "post-push" hooks. ):-)	10:08
	→ mapreduce joined	10:08
	→ walf443 joined	10:08
	→ tasslehoff joined	10:10
shruggar	I suppose having a real post-push makes as much sense if not more sense than post-commit	10:11
	→ mbroeker joined	10:11
	→ codeshepherd joined	10:12
tasslehoff	How can I fetch objects from one local repo to another? Setup a remote? None of the repos are bare.	10:12
shruggar	tasslehoff, bare repositories are only a convenience and are not required. Just set up a remote	10:13
	git remote add foo /path/to/repos	10:13
	→ maleknet joined	10:13
	← jtauber left	10:15
	← harinath left	10:16
	→ pallaert joined	10:17
	→ intripoon joined	10:19
tasslehoff	shruggar: thanks	10:20
	← FND left	10:22
	← devanjan left	10:23
	→ rajeshsr_ joined	10:24
tasslehoff	when I add the remote I get a warning about "no common commits", is that very bad?	10:25
	→ smuf joined	10:25
	→ tinto101 joined	10:25
shruggar	not necessarily, but it's probably not what you want	10:25
	tasslehoff, how were these repos created?	10:26
tasslehoff	shruggar: git svn clone :)	10:26
	shruggar: we're moving to git, but I still have to work with the svn server	10:26
	← mar77i left	10:27
shruggar	tasslehoff, I'm guessing that the repos weren't created using the same "git svn clone" string. For example, one may have been created with "git svn clone svn+ssh://foo/bar/baz" while another was created with "git svn clone file:///bar/baz"	10:27
tasslehoff	shruggar: yeah. one is actually trunk, while the other one is branches/foo.	10:27
	← tvw left	10:28
tasslehoff	hmm, when I say it out loud, this sounds sketchy	10:28
	→ ziro` joined	10:29
ziro`	what should i set my autocrlf to?	10:29
	i want to only work with LF's but I want to ensure I don't accidently add CRLF's	10:29
shruggar	tasslehoff, there's the problem, then :) svn branches don't translate neatly into git branches like that. In some situations, you can clone the entire svn repository (specifying where branches, tags, etc, go) and then treat them as normal git branches	10:30
	tasslehoff, I've yet to meet a mature svn repository (out of all three I've looked at, so my own experience is rather minimal) where git-svn could process that sort of thing sanely	10:30
tasslehoff	shruggar: hm. I also want to add a remote to my colleagues checkout of the same branch, but I sense that will cause the same problem	10:31
	→ harinath joined	10:31
shruggar	tasslehoff, if you both cloned using the same string, you should be able to pass commits back-and-forth no problem	10:32
	→ ChUrP joined	10:34
Ilari	ziro`: input then	10:35
	→ tPl0ch joined	10:36
shruggar	tasslehoff, for example, I pass commits through a bare repository between work and home, since work and home both cloned via the same git-svn string	10:38
bob_f	sitaram: Sorry, juggling a bunch of things at once here. When I try to "git clone git@server:gitolite" I get "fatal: 'gitolite' does not appear to be a git repository".	10:38
	And same if I do "gitolite-admin", which is what the repo is called.	10:38
	← poseid left	10:40
Ilari	bob_f: Try 'gitolite:gitolite-admin'.	10:41
	bob_f: (weird games gitolite plays with ssh).	10:42
tPl0ch	Hi, I am trying to create a post-receive hook on our server. It is as simple as that: http://pastebin.com/hW0Ffq0T If I run it as the 'git' user, it works flawlessly, but when pushing via git+ssh from the local repo, it does not get executed.	10:42
bob_f	Ilari: Nope, same error.	10:43
Ilari	tPl0ch: 1) Don't run git as root. 2) Don't use pull in scripts, 3) if you cd in script, remember to unset $GIT_DIR.	10:43
bob_f	This is probably in the docs, let me check it up.	10:43
Ilari	bob_f: 'git clone gitolite:gitolite-admin' (literally that way) says that it does not appear to be git repository?	10:44
bob_f	Oh, wait.	10:44
	→ myosui joined	10:44
Ilari	bob_f: As opposed to things like access denied, or host not found.	10:44
bob_f	Nope, that dosen't work either - yeah, says it's not a repo.	10:44
	→ Aaaarg joined	10:44
Ilari	bob_f: What does 'ssh gitolite' (yes, 'gitolite' literally) say?	10:45
	→ devanjan joined	10:45
bob_f	Ilari: It says hello and lists repos and perms.	10:45
tasslehoff	and the final straw. it's not possible to use git-svn to create a bare repo, with support for pushing changes back up?	10:46
tPl0ch	Ilari, so, how do I update a clone in the post-receive hook?	10:46
	without pull?	10:46
Ilari	bob_f: 'git ls-remote gitolite:gitolite-admin' (again, literally).	10:46
	tPl0ch: fetch + reset.	10:47
shruggar	tasslehoff, "not possible" is too strong a word. "very inconvenient and annoying to set-up, and always requiring awareness of the set-up" is more accurate	10:47
tasslehoff	shruggar: :)	10:48
bob_f	£ git ls-remote gitolite:gitolite-admin	10:48
	fatal: 'gitolite-admin' does not appear to be a git repository	10:48
	→ MattDiPasquale joined	10:48
bob_f	(gitolite-admin is listed when I do "ssh gitolite" though).	10:49
	And I have R W perms.	10:49
Ilari	bob_f: Some very broken gitolite setup? If you didn't have permissions, it would print access DENIED message.	10:49
bob_f	Hrm.	10:49
Ilari	bob_f: Try 'git ls-remote gitolite:does-not-exist'?	10:50
bob_f	£ git ls-remote gitolite:does-not-exist	10:50
	fatal: 'does-not-exist' does not appear to be a git repository	10:50
	fatal: The remote end hung up unexpectedly	10:50
	→ ehabkost joined	10:50
Ilari	bob_f: Eeh... For me the error is: "fatal: remote error: R access for does-not-exist DENIED to Ilari".	10:50
	← MattDiPasquale left	10:51
bob_f	Hmmmm.	10:51
Ilari	bob_f: Wonder what the heck is going on. Based on 'ssh', there's gitolite on other end, but git connections look like they bypass gitolite.	10:51
bob_f	Yeah, weird, huh ?	10:51
shruggar	tasslehoff, the most annoying part of it: commits will ALWAYS be re-written when you dcommit, so if you "git push" to such a repository, a "git pull" right afterwards will try to merge. "git pull --rebase" /might/ work, but something like "git push && git remote update && git reset --hard origin/master" is more what is required	10:51
	tasslehoff, (something like that, anyway)	10:52
Ilari	bob_f: (ignore the "fatal: remote error " part, it is artefact of local setup).	10:52
bob_f	Ah okay.	10:52
	← airborn left	10:53
	→ airborn joined	10:54
Ilari	bob_f: Because error messages from gitolite are converted into protocol errors, and client prints prints protocol errors with 'fatal: remote error: " prefixed...	10:54
	Unode\|NA → Unode	10:54
	→ aliceinwire joined	10:55
sitaram	bob_f: putty?	10:55
bob_f	sitaram: Nope.	10:55
Ilari	bob_f: "ssh gitolite git-upload-pack does-not-exist". What does that complain about?	10:56
bob_f	Just £ ssh -V	10:56
	OpenSSH_5.1p1 Debian-5, OpenSSL 0.9.8g 19 Oct 2007	10:56
sitaram	bob_f: client OS and sever OS? (shooting in the dark here...)	10:56
tasslehoff	shruggar: hm. plan 1 is now to speed up the process towards !svn	10:56
bob_f	ubuntu-server and debian on my end.	10:56
	£ ssh gitolite git-upload-pack does-not-exist	10:57
	fatal: 'does-not-exist' does not appear to be a git repository	10:57
shruggar	tasslehoff, it is a good plan. Note that a new interface to svn is being actively worked on at the moment. I don't know if it will solve any of these problems, but I hardly see the point if it doesn't :)	10:57
	← jtdowney left	10:58
Ilari	bob_f: So that looks like it bypasses gitolite (but 'ssh gitolite' doesn't...). Hmm...	10:59
	← nevyn left	11:00
Ilari	bob_f: You ran 'ssh gitolite' and 'ssh gitolite git-upload-pack does-not-exist' from the same shell (just to rule out environment fscking things up?)	11:00
bob_f	Here's my ~/.ssh/config just in case: http://bpaste.net/show/7307/	11:00
	Ilari: Yes, same shell, no env changes.	11:00
tasslehoff	shruggar: is that interface mature enough for testing, or is it way-pre-alpha?	11:00
Ilari	bob_f: Looks sane...	11:01
	← project2501a left	11:01
	→ nevyn joined	11:02
shruggar	tasslehoff, I think last I saw there were certain aspects of it for which testing was requested, but I wouldn't call it mature enough that it's worth testing for actual use yet	11:02
rajeshsr	How does git distinguish between tag name and a branch name, when you do git branch newName tagName etc..	11:02
	?	11:02
sitaram	bob_f: run src/sshkeys-lint in your gitolite clone (prints USAGE message), then folow instructions, and report back	11:03
Ilari	bob_f: Bit of brute-forceish, but: Do 'ssh -vvvv gitolite' and save the full output to file. Then do the same (saving to other file) for 'ssh -vvvv gitolite git-upload-pack does-not-exist' and then diff the two files.	11:03
	FauxFaux has managed to lose a commit from reflog. /o\	11:03
	← Leemp left	11:03
bob_f	Ilari: Sure.	11:03
Ilari	bob_f: 'diff -u file1 file2' to diff them.	11:03
engla	git diff --no-index file1 file2 and you can use a more familiar diff	11:04
	→ MattDiPasquale joined	11:05
	→ Leemp joined	11:05
bob_f	Ilari: I can pastebin this safely ?	11:05
Ilari	bob_f: It shouldn't have any private information.	11:05
	← hyperair left	11:05
bob_f	Yeah, looks safe.	11:05
	http://bpaste.net/show/7309/	11:05
ncopa	back	11:05
	thanks all	11:06
	→ hyperair joined	11:06
	← nevyn left	11:06
	← Nickeeh left	11:06
	→ project2501a joined	11:06
	← saidi left	11:07
	← MattDiPasquale left	11:07
	← devanjan left	11:08
Ilari	bob_f: What the heck? Looks like 'ssh gitolite' did actual login?	11:09
	← ntoll left	11:10
Ilari	bob_f: Do the sshkeys-lint stuff...	11:12
	→ tvw joined	11:12
	→ nevyn joined	11:13
	← Mpenz left	11:15
	→ Mpenz joined	11:15
	← ige left	11:16
level09	when using reflog to revert, do I use checkout ? or reset ? or revert ?	11:18
	← andreaaa left	11:18
	← harinath left	11:18
wereHamster	what do you mean by 'revert'?	11:18
level09	like undo changes	11:19
	i have merged some branches and did some cleaning that messed up everything	11:19
Ilari	level09: 'reset --merge' probably then...	11:20
	level09: Or --hard (dangerous).	11:20
wereHamster	was the merge successful?	11:20
	or did it fail because of conflicts?	11:20
level09	failed	11:20
	→ radarek joined	11:20
wereHamster	reset --hard	11:21
	or checkout --merge -- .	11:21
level09	but if I want to go back into a specific point in my reflog	11:21
wereHamster	reset --hard or checkout	11:21
level09	okay thanks ! I'll try that	11:22
	→ aziz joined	11:22
	→ cytrinox_ joined	11:23
bob_f	Ilari: I'm gonna run to lunch, but this lint stuff should hopefully help me figure it out - thanks a lot for your help though, it's really appreciated.	11:23
	I'll let you know how it goes.	11:23
	← cytrinox left	11:23
	cytrinox_ → cytrinox	11:23
	← rajeshsr left	11:24
	rajeshsr_ → rajeshsr	11:24
	→ rajeshsr_ joined	11:24
	← LongBeach left	11:25
	← project2501a left	11:25
	→ oriba joined	11:26
	← dytrivedi left	11:27
	→ smuf_ joined	11:27
	→ flazz joined	11:28
	→ froschi joined	11:29
	← jony left	11:29
	← smuf left	11:30
	→ univate joined	11:31
	← masterkorp left	11:31
	← gnufied left	11:32
	→ harinath joined	11:32
	→ smuf joined	11:34
	← smuf_ left	11:34
	→ bentob0x joined	11:34
	→ gebi joined	11:36
	→ mathepic joined	11:36
	→ jony joined	11:37
	→ apsysoev joined	11:38
	← smuf left	11:38
	→ project2501a joined	11:39
sondreb	when i type "git rm <filename>" it deletes the file but then sends the message "Deletion of directory 'ubs/common/composer' failed. Should I try again? (y/n)?"	11:39
	why does it try to delete te folder?	11:40
mathepic	Whats the filename?	11:40
sondreb	C:\BuildSys\work\ubs\common\composer>git rm comp_resbuild.pm	11:41
	rm 'ubs/common/composer/comp_resbuild.pm'	11:41
	Deletion of directory 'ubs/common/composer' failed. Should I try again? (y/n)?	11:41
	→ jupko joined	11:41
jupko	http://pastie.org/1014987	11:41
	what's the problem with my local repository ?	11:41
	How can I correct it ? You can see the situation if you open the link I pasted.	11:42
	← zerd left	11:42
mathepic	sondreb: Thats odd.	11:42
	→ Yuuhi joined	11:43
	← ianmcorvidae left	11:43
	→ ianmcorvidae joined	11:43
wereHamster	jupko: git fsck --full	11:43
	→ zerd joined	11:43
	← wagle left	11:44
	← TheUni left	11:44
jupko	okay	11:44
wereHamster	jupko: and before you do that, can you check if the file .git/objects/e1/17cf9010e5973e9d65139424bde2e27cf7389c exists and if it is zero bytes long?	11:44
	→ TheUni joined	11:44
	→ [Po]lentino_ joined	11:44
	→ wagle joined	11:44
jupko	oops	11:44
	← psankar left	11:45
jupko	I have run fsck before the checking	11:45
	← [Po]lentino left	11:45
	[Po]lentino_ → [Po]lentino	11:45
jupko	4KB	11:46
	and it exists	11:46
	← marcog left	11:46
jupko	http://pastie.org/1014995	11:47
	→ marcog joined	11:47
jupko	you can see the output of fsck operation and the size of that object file.	11:47
	the question is still existing, how can I fix it ? :)	11:47
	and what can cause such a problem at all ? This will be my other question.	11:48
wereHamster	do you have a backup of that repo? or a clone that contains that object?	11:48
mathepic	You need to replace that object with one from a noncorrupt repo.	11:48
jupko	I can do a new clone because I have pushed the commits	11:48
	but I would like to avoid it next time so I would like to know what can cause such a problem.	11:49
wereHamster	filesystem corruption for example	11:49
jupko	only in this small git repository and no corruption in another places ? :)	11:49
	→ apsysoev1 joined	11:50
	← apsysoev left	11:50
jast	no corruption you have found so far	11:50
	← Lgb left	11:50
wereHamster	a block on the disk is usually 512 bytes, well within the 4kb file that is corrupt	11:51
jast	and it's actually quite likely to happen that way if, for example, you're working on that repository and you get a power failure in the middle of something	11:51
jupko	4KB is the page size on i686	11:51
wereHamster	page size, yes. But not block on the harddrive	11:51
jupko	ah okay I see	11:52
	I may know the problem.	11:52
	I am developing a driver	11:52
	and syscrash occured because of some bad lines.	11:52
	that I inserted	11:52
	and maybe it caused that.	11:52
	meanwhile I was in this repository.	11:52
	→ gnufied joined	11:52
jupko	so it's definitely worth to make a backup directory then.	11:52
	:-)	11:52
	→ Lgb joined	11:53
jupko	wereHamster: thank you.	11:53
jast	in theory git syncs stuff before updating a branch, but perhaps your filesystem did something nasty and ignored it	11:53
	← project2501a left	11:53
jupko	jast: no problem with filesystem at all	11:53
	→ NET\|\|abuse joined	11:53
	→ project2501a joined	11:53
	← jupko left	11:54
jast	well, the corruption has to have been caused by something :)	11:54
	← anathematic left	11:57
	→ anebi1 joined	12:00
	← mathepic left	12:01
anebi1	hi. i'm newbie in git and i want to ask if is possible to clone a repository under folder located in my current git repository. for example if my git repository is in this http path /git/php, i want to clone a repository under /git/php/vendor. is this possible somehow?	12:01
	← myosui left	12:02
jast	are you, perchance, talking about submodules?	12:03
	← project2501a left	12:04
anebi1	jast: if you call this submodule, then it is. i'm looking it from http url side. vendor is a subfolder in my gitrepo and i want to clone all content of remote under this folder.	12:05
	jast: i'm reading now for submodules :)	12:05
jast	oh, you want to clone only part of a repository?	12:05
	→ t0rc joined	12:05
jast	if so... well, tough luck. git doesn't support that.	12:06
	→ disappearedng joined	12:06
	← NET\|\|abuse left	12:06
anebi1	jast: the people moved the project from svn to git, i think i will need the full cloning if their repo. i have not checked yet how they have done this.	12:07
	but the content should go under vendor subfolder in my repo	12:08
	→ amirite joined	12:08
jast	sounds like a case for submodules all right... or the third-party git subtree tool, if you prefer that	12:10
Mpenz	sitaram: :) Thanks. I figured.	12:10
jast	some people don't like submodules very much, e.g. because you have to manually update the submodule and record the update whenever you need a newer version	12:11
anebi1	jast: i see. i will learn how submodules work and probably will go that way. thanks	12:12
	→ andreaa joined	12:13
	→ Aaaarg_ joined	12:14
	← rchavik left	12:14
	← Aaaarg left	12:17
	Aaaarg_ → Aaaarg	12:17
	← anebi1 left	12:17
	← ziro` left	12:17
tPl0ch	Hi, I am still trying to get a post-receive hook to run: http://pastebin.com/Bkj9ULm1. If I run that as the user 'git' on the server, it works flawlessly, but when I push from my local repository, the hook won't execute.	12:18
	← RobertLaptop left	12:19
	→ hugoxrosa joined	12:19
	→ anebi joined	12:22
	← d0k left	12:26
	← Aaaarg left	12:31
	→ stringoO joined	12:31
	← agile left	12:31
	→ d0k joined	12:32
	← freimatz left	12:33
	→ Aaaarg joined	12:34
	→ mw joined	12:35
	→ ntoll joined	12:37
	← gnufied left	12:38
	← priidu_ left	12:39
	← univate left	12:40
	→ IslandUsurper joined	12:40
	→ univate joined	12:41
	← codeshepherd left	12:45
	AAA_awright_ → AAA_awright	12:46
	→ codeshepherd joined	12:46
	← univate left	12:47
	← werdan7 left	12:48
	→ bdiego joined	12:50
	← anebi left	12:52
	→ anebi joined	12:53
	← keyvan left	12:53
	→ rlb3 joined	12:54
	→ LiamH joined	12:55
tPl0ch	Hi, I am still trying to get a post-receive hook to run: http://pastebin.com/Bkj9ULm1. If I run that as the user 'git' on the server, it works flawlessly, but when I push from my local repository, the hook won't execute.	12:55
	← disappearedng left	12:57
	→ MattDiPasquale joined	12:58
	← ajpiano left	12:58
	→ ajpiano joined	12:59
	→ _iksik joined	12:59
PerlJam	tPl0ch: how did you run it from the command line?	13:00
	← EricInBNE left	13:00
	← anebi left	13:01
PerlJam	tPl0ch: you do have execute permission set, yes?	13:01
	← iksik left	13:01
tPl0ch	PerlJam, yes, it's executable, and I did a su git && /path/to/hook/post-receive	13:01
	→ anebi joined	13:02
tPl0ch	it did everything right	13:02
	but whatever I do, it won't execute when pushing from my local repo	13:02
	← darkslategrey left	13:02
PerlJam	Are you sure you're pushing to the right place?	13:03
	→ Jferro joined	13:03
	← tommie left	13:03
tPl0ch	PerlJam, yes, since the hook does the fetch after I push from my local repo	13:03
	but only if I run it by hand	13:04
	my git version is 1.7.0.5	13:04
PerlJam	wait ... that doesn't make sense.	13:05
	→ jrmuizel joined	13:05
PerlJam	The hook does not execute when you push, yes?	13:05
tPl0ch	indeed	13:05
PerlJam	and how do you know you're pushing to the right remote?	13:06
	→ frakturfreak_ joined	13:06
	→ project2501a joined	13:06
tPl0ch	git remote show origin	13:06
Ilari	tPl0ch: And how do you tell it doesn't execute?	13:06
	← MattDiPasquale left	13:06
tPl0ch	Ilari, I added a 'touch PLEASE_RUN' to post-receive	13:06
	← stringoO left	13:07
	← ajpiano left	13:09
	→ disappearedng joined	13:10
	← tehbaut left	13:10
	← frakturfreak left	13:10
PerlJam	tPl0ch: hrm. I think this is one of those "devil in the details" kind of things. If you could generate a transcript of everything you did to make sure the script was executable, that you were pushign to the right remote, and that the remote isn't executing the post-receive, they might help	13:11
	← rgr left	13:12
bob_f	Ilari: Hi, just got back from lunch and been poking a bit more.	13:13
	← anebi left	13:13
bob_f	Ilari: If I unset ssh-agent env vars, doing "git clone git@gitolite:gitolite-admin" works - you see it is asking me for the passphrase for ~/.ssh/bob and not ~/.ssh/id_rsa	13:13
	← rlb3 left	13:14
Ilari	tPl0ch: Can you pastebin transcript of push?	13:14
	→ MattDiPasquale joined	13:14
	→ tehbaut joined	13:14
bob_f	Ilari: Oh, and I think I know why. When I run "ssh-add" it didn't ask me for the passphrase for ~/.ssh/bob (it only adds id_rsa and id_dsa files maybe ?) so I guess that key wasn't stored on the keychain and so wasn't ever being offered.	13:15
	I bet if I `ssh-add ~/.ssh/bob` it'll work.	13:15
	Ilari: Yep, that fixed it. :)	13:15
	→ ziro` joined	13:17
ziro`	guys, i'm really struggling with crlf	13:17
Ilari	bob_f: The setup I locally have generates stuff like that if I try to use the wrong key: "fatal: remote error: R access for gitolite-admin DENIED to ssh-15a87ce812b576846259cbb195fe042cb17f47bec8fda0dd9c3c507d".	13:17
ziro`	i've been using autocrlf on and just cloned on a fresh git install (no autocrlf) and my files have crlf's as far as i can see...	13:18
	→ RobertLaptop joined	13:18
Ilari	ziro`: Autocrlf false: "I have editors configured to load and save LF". autcrlf input: "I have editors configured to load LF, but they might save CRLF". autocrlf true: "My editors are not LF-friendly at all".	13:18
	← pellera left	13:19
ziro`	my stupid editor won't let me set the newline	13:20
	only for 'new files'	13:20
	i just want to strip crlf's where it finds them	13:20
	→ rlb3 joined	13:21
bob_f	Ilari: sshkeys-lint tells me that using a key that gives me a command line will give me the "blah blah does not appear to be a git repository" message, so I guess it's offering a key that has the right perms but is bypassing gitolite.	13:21
	→ pellera joined	13:21
bob_f	Ilari: Either way, I'm glad we got to the bottom of it - thanks a lot for your help, I'd still be banging my face on the keyboard without it. :)	13:21
ziro`	so to be honest, i don't mind if I get given LF'd files	13:21
	I just want to ensure against my editor saving with crlf or something	13:21
Ilari	ziro`: Set autocrlf to input then.	13:22
	→ poseid joined	13:22
ziro`	but i've apparently had it to true and i've still got files with crlf's in?	13:23
Ilari	ziro`: Just be aware that changing line ending style of existing repository is bit nontrivial.	13:23
ziro`	so now I have to convert them all?	13:23
Ilari	ziro`: CRLFs in working tree (autocrlf true does that) or CRLFs in repository (bad)?	13:23
ziro`	yeah, I know it's non trivial, it's just fecking annoying since i've had autocrlf set to true and i've still managed to get crlf's in there :S	13:23
	ok heres what I did	13:24
	→ augustl joined	13:24
ziro`	I was working on my local working tree with autocrlf true	13:24
tPl0ch	PerlJam, Ilari http://pastebin.com/AH7Rmx6Y	13:24
	← project2501a left	13:24
ziro`	i installed git on my mac, (default autocrlf = false?), cloned and copied the file over to my windoze machine and opened it in notepad and had line breaks (notepad only renders crlf's, not lf's)	13:25
	therefore I'm judging I have crlf's in my repository	13:26
Ilari	ziro`: 'git show HEAD:path/to/file \| od -t x1' should show hexdump of file as it is in repository. CRLF is 0d 0a, LF is 0a.	13:27
	← tildeequals left	13:27
ziro`	lemme check	13:27
	→ gbacon joined	13:27
	← maleknet left	13:28
	← jrmuizel left	13:28
	→ maleknet joined	13:28
	← radioman-lt left	13:29
	← poseid left	13:29
	← hohoho left	13:30
	→ tildeequals joined	13:30
	← codeshepherd left	13:30
ziro`	0d 0a :(	13:31
	→ dytrivedi joined	13:31
Ilari	ziro`: Hmm... How did those CRLFs get into the repo? Autocrlf should translate them to LFs.	13:32
	→ j0shua joined	13:32
ziro`	yeah :/ it's defo set to true though	13:32
	maybe i screwed something up	13:32
	← augustl left	13:33
ziro`	whats the best way to go about fixing this?	13:33
	← maleknet left	13:36
	→ tomsdale joined	13:36
	← oriba left	13:38
	← tasslehoff left	13:38
	→ gnufied joined	13:39
	→ aresnick joined	13:40
	→ jrmuizel joined	13:41
	→ alezandro joined	13:43
	← t0rc left	13:43
	→ cannonball joined	13:45
mapreduce	Can I git log in the opposite chronological order?	13:45
	← intripoon left	13:45
engla	mapreduce: git log --reverse --date-order <more options> can that be a start?	13:45
	← marc_kdab_ left	13:46
	→ Phrogz joined	13:46
	← coppro left	13:46
mapreduce	Thanks.	13:48
	→ warlock_mza joined	13:50
	→ coderdad joined	13:52
	→ JackWinter joined	13:52
	→ masterkorp joined	13:53
	← rlb3 left	13:53
	→ akahn joined	13:55
	← akahn left	13:55
	← gebi left	13:56
	← hobodave left	13:56
	→ hobodave joined	13:56
	→ maleknet joined	13:56
	→ Joeconyers joined	13:56
	→ damog joined	13:57
	→ rlb3 joined	13:57
	← rlb3 left	13:58
	← MattDiPasquale left	13:58
	← Jferro left	13:58
	→ Jferro joined	13:59
	→ MattDiPasquale joined	13:59
	← MattDiPasquale left	13:59
	→ pmd joined	13:59
	→ rlb3 joined	13:59
	← masterkorp left	13:59
	→ MattDiPasquale joined	13:59
	→ t0rc joined	14:00
	← jrmuizel left	14:00
	← csgeek left	14:00
	→ felipec joined	14:00
	→ Bass10 joined	14:01
	→ myosui joined	14:01
	→ eletuchy joined	14:01
	→ stringoO joined	14:01
	← hobodave left	14:01
	→ jrmuizel joined	14:01
ziro`	ok i just used a tool to convert my crlf's to lf's	14:02
	i only did it on text files (js, css, php, phtml, log) etc	14:02
	← Bass10 left	14:03
	← sgh left	14:03
	→ Bass10 joined	14:04
	→ scarabx joined	14:04
	→ necromancer joined	14:04
	← maleknet left	14:06
	→ tedoc2000 joined	14:06
	→ madewokherd joined	14:07
	← rlb3 left	14:08
	→ rgr joined	14:08
	→ ergates joined	14:09
ergates	er, hullo	14:09
	← dytrivedi left	14:10
ergates	I wonderde if anyone can help. I'm trying to write a shell script that can have a date passed to it, the script can then download a tarball from before that date from a remote repository without having to do a git checkout	14:10
	→ icefox joined	14:10
ergates	I've got git-archive getting a head rev in this way, and gitweb can be pointy clickied, but the only way i can think of passing a date to git-archive is with git-rev-list and that seems to need a local repo	14:11
shruggar	ergates: what alternative do you envision where one would need to do a git checkout?	14:11
	→ erichynds joined	14:12
	→ csgeek joined	14:12
ergates	shruggar: sorry? You mean, why would I not want to git checkout?	14:13
	→ Zephyrus joined	14:13
shruggar	ergates: No, I mean why would you need to do a checkout in order to get an archive	14:13
	→ spaceonline joined	14:13
shruggar	ergates: Or in order to determine which tag you want	14:13
ergates	shruggar: ah! No, i don't want to do a checkout, but unless I do a checkout I can't find the revision number I need - chicken and egg type thing	14:14
	← rgr left	14:14
	← tvw left	14:14
	→ rlb3 joined	14:14
shruggar	ergates: Why would you need a checkout in order to find it?	14:14
	← damog left	14:14
ergates	shruggar: that is what I'm hoping to find out here - I can't seem to find a method	14:15
shruggar	ergates: And I'm asking what you're talking about, because I can think of no process for finding a ref which would involve checkout	14:16
	do you perhaps mean "clone"?	14:18
	← jrmuizel left	14:18
	← scarabx left	14:18
	← roop left	14:19
ergates	well, I'm trying to do this from a command line, hoping it will examine the remote repository, find the ID that corresponds to that date, pass it to git archive. The 'remote' line in the git-rev-list obviously doesn't work - i was just trying stuff out:	14:20
	git archive --format=tar -v --remote=git://www.aleph1.co.uk/yaffs2 $(git rev-list -n1 --remote=git://www.aleph1.co.uk/yaffs2 --before="2010-05-30 23:59:59" master) > yaffsdate.tar	14:20
shruggar	a combination of man git-ls-remote, and man git-fetch-pack --depth=1 may be able to let you examine the metadata for the remote tags without downloading absolutely everything about it, but that may still attempt to download one complete level of history	14:22
Gitbot	shruggar: the git-ls-remote manpage can be found at http://git.or.cz/man/git-ls-remote	14:22
shruggar	not sure how --depth plays with tags	14:22
	→ rseifert joined	14:22
ergates	thanks - I'll have a read	14:22
	→ Yuffster joined	14:22
shruggar	git doesn't lend itself well to being used in a non-distributed manner	14:23
	alternatively, if git-web is available, you may be able to use that to get the information you need	14:23
	← TheNewAndy left	14:23
	← trivol left	14:24
ergates	shruggar: yes, I installed git-web, and thought about using wget, but that relies on a manual eyeball of the ID , and what my task atm is that i'm updating scripts that used to run cvs pserver commands	14:24
Phrogz	shruggar!	14:25
shruggar	Me?	14:25
	→ etnt joined	14:25
ergates	anyway, thanks shruggar	14:25
shruggar	ergates, I'm sure there's something parsable in git-web output. rss feeds, for example,	14:25
	→ trivol joined	14:25
	→ codeshepherd joined	14:25
Phrogz	For years...many, many years...I had a test file on my server called "shruggar's ridge" which I guess you wanted to know how to have an outset border with a large gap between the up and down. So I made this file showing how to wrap two elements to achieve the effect.	14:26
etnt	how do I update my submodule to the latest version ?	14:26
Phrogz	I just deleted it last week. :)	14:26
ergates	ok, ta	14:26
shruggar	Phrogz, damn, 'cause NOW I need it!	14:27
	Phrogz, j/k, I gave up on the design end long ago and I strictly work on back-end stuff now :)	14:27
	← MattDiPasquale left	14:27
	← tinto101 left	14:27
Phrogz	:)	14:27
	← ergates left	14:27
	→ MattDiPasquale joined	14:27
shruggar	seriously though, I that must have been at least five years ago	14:28
	← MattDiPasquale left	14:28
	→ nighthwk1 joined	14:28
	← csgeek left	14:29
	← codeshepherd left	14:29
Phrogz	Yeah.	14:29
	Phrogz has a dirty tmp directory :)	14:29
	← rlb3 left	14:29
	← thierryp left	14:29
shruggar	for me that would have been six or so hard drives, or at least formats, ago	14:29
	→ codeshepherd joined	14:29
shruggar	still elements of the same .bashrc though...	14:30
	→ panzi joined	14:30
dennda	What is the inverse of HEAD~1? (aka, not previous commit, but next commit. chronologically)	14:30
panzi	what is the git equivalent to "hg update -r tagname"	14:30
dennda	panzi: -r?	14:31
	frakturfreak_ → frakturfreak	14:31
dennda	panzi: aka, go to tag name? git checkout tagname	14:31
panzi	ic, thx	14:33
	→ MattDiPasquale joined	14:33
Phrogz	Can someone explain this behavior: http://gist.github.com/448541 (Why does my status show dirty after a pull, and then get clean after a do-nothing push?)	14:35
	← codeshepherd left	14:36
shruggar	Phrogz, I don't see your status showing dirty in that log	14:37
	→ spearce joined	14:37
Phrogz	shruggar: I mean git status shows that I'm ahead of the master (which I'm not), and then after the do-nothing push status says I'm fine.	14:37
	← vu3rdd left	14:37
Phrogz	s/master/origin/	14:38
shruggar	Phrogz, that I don't know. Can you do a: git rev-parse HEAD before/after each "status" ?	14:38
Phrogz	OK, 30s	14:38
	shruggar: http://gist.github.com/gists/448541 (reload)	14:40
ziro`	is it possible to remove a directory from git but not from my working tree ( but obviously unindex it too)	14:41
shruggar	ah, I think I see. You did "git pull origin planning", which may not update refs/remotes/origin/planning (I can never remember the specifics of that)	14:41
	Phrogz, then when you puhed, refs/remotes/origin/planning was updated	14:42
	→ sagsousuke joined	14:43
Phrogz	Ah, OK, sounds plausible. Thanks.	14:43
shruggar	I tend to use "git remote update" followed by either "git merge" or "git rebase" depending on what I want to do	14:43
	→ hobodave joined	14:43
shruggar	(or "git reset --hard", for that matter)	14:43
	← Phrogz left	14:44
	← jatt left	14:45
	→ csgeek joined	14:47
	→ ajpiano joined	14:48
	← thomas_b left	14:48
	→ project2501a joined	14:48
	→ Alien_Freak joined	14:49
	← csgeek left	14:49
	→ thomas_b joined	14:50
	→ MetaCosm joined	14:51
	→ tar- joined	14:51
	→ Jaxan joined	14:53
	← mw left	14:54
	→ mw joined	14:54
	← Alien_Freak left	14:55
	→ Alien_Freak joined	14:56
	→ irc2samus joined	14:58
	← apsysoev1 left	15:00
	→ jepler joined	15:01
	← ExtraSpice left	15:01
	→ hohoho joined	15:02
divVerent	is git hash-object "officially" allowed to be used in scripts?	15:02
	i.e. can I rely on it staying?	15:03
	I basically need a md5sum-like functionality (simple file hashing) in a shell script that is in a git repository	15:03
	← hohoho left	15:03
	→ hobodave_ joined	15:03
divVerent	can I rely on "git hash-object filename" to return a hash for the file in all foreseeable future?	15:03
shruggar	divVerent, yes, you can rely on it	15:03
wereHamster	yes	15:03
	→ rgr joined	15:04
divVerent	also, sicne when does it exist?	15:04
	was it in 1.6? 1.5?	15:04
	← hobodave left	15:05
wereHamster	divVerent: man git -> GIT COMMANDS	15:05
	← mithro left	15:06
shruggar	7672db20c2060f20b01788e4a4289ebc5f818605	15:06
Gitbot	[git 7672db20c]: http://tinyurl.com/3ad35hy -- [PATCH] Expose object ID computation functions.	15:06
divVerent	2005... ok	15:06
wereHamster	that would be tag v0.99 and later	15:06
divVerent	then I can assume that about every install of git today has it	15:06
wereHamster	yes	15:06
divVerent	BTW, I wonder what a blob hash actually is	15:07
	→ jrmuizel joined	15:07
divVerent	it doesn't seem equal to the SHA1 of just the file content	15:07
	← pellera left	15:07
divVerent	it probably is SHA1 of file content with some extra header prepended	15:07
shruggar	the blob contains a small amount of meta-data	15:07
divVerent	but well, for my application it doesn't matter what is prepended, any hash does :P	15:07
	← girishr left	15:07
wereHamster	divVerent: http://progit.org/book/ch9-2.html	15:08
	divVerent: scroll down to 'Object Storage'	15:08
divVerent	ah, I see	15:08
	→ NET\|\|abuse joined	15:09
	→ marc_kdab_ joined	15:09
	← fmardini left	15:09
shruggar	I didn't know they were individually zlib-deflated	15:09
divVerent	=> "blob 16\000what is up, doc?"	15:09
	hm...	15:09
	← hyperair left	15:09
divVerent	is "blob 16\000what is up, doc?" valid too, and yields another hash? ;)	15:10
	or "blob 016"	15:10
	i.e. can you, maliciously, create two "seemingly equal" objects with different hash?	15:10
	or is it a requirement that the header uses no leading zeroes, no extra spaces or other crap :P	15:10
shruggar	divVerent, try it and find out. Report it as a bug if git doesn't complain	15:11
wereHamster	I think it is, but you'd have to consult the source code to be sure	15:11
divVerent	so you'd consider that a bug? ok	15:11
shruggar	I'd consider it an extremely minor bug	15:11
	→ malumalu joined	15:11
	← rajeshsr left	15:11
divVerent	know a commandlone tool to JUST deflate?	15:12
	(not prepend header like gzip does)	15:12
	← Jaxan left	15:12
	→ hyperair joined	15:12
	← frakturfreak left	15:13
shruggar	divVerent, I tend to use perl or php. You may be able to just gzip and remove the header, though	15:13
	← pallaert left	15:14
	→ adamv joined	15:14
	← mmattice left	15:15
	→ Nemurenai joined	15:15
	→ rlb3 joined	15:15
	← cbreak_work left	15:16
divVerent	ok, such stuff is not accepted by git	15:16
	nice, strict parsing of that :P	15:16
	← andreaa left	15:16
	→ webus joined	15:17
	← radarek left	15:17
	← rlb3 left	15:17
	← parasti left	15:19
	→ nadavoid joined	15:20
	→ hohoho joined	15:21
	→ mmattice joined	15:21
Evanlec	hey, im wondering how you web developers use git in your workflow...do you commit locally and run a local testing env, pushing only at the end of the day, wondering if anyone has tips on git for web development	15:21
	→ priidu_ joined	15:22
	→ rox` joined	15:22
shruggar	commit locally and run a local testing environment, pushing whenever a feature is done, with a separate script to do releases which we run outside of peak-time	15:24
Evanlec	shruggar: yea, i guess that is the best way	15:24
Rhonda	commit locally with a local testing environment, pushing to an integration test environment and from there get it deployed to the proper place - if you have a very big picture	15:25
shruggar	the script does various things like automatic tagging and permission tweaking as well	15:25
Evanlec	shruggar: i guess i just need to setup an optimal local testing enviornment, i find it time-consuming keeping the files and data in sync etc	15:25
	any tips on optimizing that aspect?	15:25
	← warlock_mza left	15:26
Evanlec	im considering using mysql replication...but maybe thats overkill?	15:26
	← Belna left	15:26
Evanlec	or set it up for remote connection	15:27
shruggar	Evanlec, we generally don't bother keeping catalog images in-sync (just have a fallback image to display), and we keep two databases which everyone uses: one is the "development" database (which uses git to version snapshots, so we can always get back to a particular point when bisecting), and one is the "live-sync" database (a weekly snapshot of the live database in case we for some reason need to test with a lot of data)	15:27
	→ ExtraSpice joined	15:28
	← nadavoid left	15:29
myosui	Let's say I have a branch "dev" with 100 commits. My master has got the first 50 commits of "dev" branch. Now I want to merge into master the commits "70-100" from "dev". How can I do it?	15:29
Evanlec	shruggar: okay, so you're working with data thats going to be atleast a week old then?	15:29
shruggar	myosui, so you want to discard commits 50-60?	15:29
wereHamster	myosui: create a temp branch with just commits 70-100 and merge that into master	15:29
shruggar	Evanlec, for development purposes, yeah. Only about once a year does an issue come up where we care about very-recent data	15:31
	→ bcardarella joined	15:31
	← aliceinwire left	15:31
Evanlec	shruggar: i see, well for one application i work on (the timeclock system) i tend to need fairly up-to-date data	15:31
myosui	wereHamster, It's about the yesterday issue. I have this main branches: trunk/kernel, trunk/ws, trunk/migrations, trunk/base, trunk/admin	15:32
Evanlec	well, i need up-to-date data if theres a current problem	15:32
	← skoop left	15:32
myosui	wereHamster, each one of this branches has got their own files. Now I would like to have all of this files in master.	15:32
	→ rayners joined	15:32
	← gnufied left	15:32
	→ raichoo joined	15:33
Evanlec	shruggar: and yea, keeping catalog (item) images in sync is crazy lol, i discovered this before	15:34
shruggar	Evanlec, it's rare that we've encountered a data-dependent problem that wasn't either a data-issue (ie: needs to be corrected in-place anyway) or a problem with an import script (ie: if we need the bad data, we just run the import script locally)	15:34
myosui	wereHamster, the problem here is that when I initialize the branches I deleted the files that were at master in the first place. Now if I try to perform a merge to master I loose the deleted files. I would like to have a way to perform merges from the trunk branches without calling the commits related to the files deletion process.	15:34
Evanlec	shruggar: problems with import scripts are common here	15:34
shruggar	Evanlec, so just run the import script locally ;)	15:35
wereHamster	myosui: uhm, yeah, that's what you get for messing with the history like that..	15:35
Evanlec	shruggar: much of the code i work on takes raw data, processes it and puts it into the db	15:35
wereHamster	myosui: try this: merge --no-commit <...>; checkout HEAD -- <the files>; git commit	15:35
tPl0ch	http://pastebin.com/AH7Rmx6Y <- Why is my post-receive hook not being executed?	15:35
	→ tobiassjosten joined	15:35
Evanlec	shruggar: yea, i really need to get off windows, its a crappy testing environment, thankfully my boss agreed to let me install linux	15:36
shruggar	Evanlec, I miss that kind of work. converting things into other things is my favorite type of thing :)	15:36
	← traviscline left	15:36
	→ kumbayo joined	15:36
	→ traviscline joined	15:36
	→ parasti joined	15:36
	← traviscline left	15:37
	→ traviscline joined	15:37
	← traviscline left	15:38
Evanlec	shruggar: yea, well the other option is to connect to the live db remotely, but i just ralized that kind of kills the point of 'testing' lol	15:38
	→ traviscline joined	15:39
Evanlec	i've done it both ways	15:39
	but connecting remotely can kill performance	15:39
	← traviscline left	15:39
	→ frakturfreak joined	15:40
Evanlec	shruggar: its fun work yea, just wish it wasn't all php	15:41
shruggar	Evanlec, ick. Okay, I also miss that work because it wasn't in PHP :)	15:41
Evanlec	shruggar: haha, yea, im sick of php	15:42
	← t0rc left	15:44
	→ kipras joined	15:44
	← froschi left	15:45
	IslandUsurper → IslandUsurperAFK	15:47
Evanlec	its getting better though (trying to make myself believe that lol)	15:47
	→ Jaxan joined	15:47
	← jeremy_c left	15:48
	← tomsdale left	15:48
	← steffkes left	15:51
	→ bcardarella_ joined	15:51
	← bcardarella left	15:52
	bcardarella_ → bcardarella	15:52
	→ richard421 joined	15:52
	→ kpfleming joined	15:52
	→ froschi joined	15:52
	← erichynds left	15:53
	→ thatdavidmiller joined	15:54
	← rox` left	15:55
richard421	It looks like 'git format-patch' may generate shortened sha1 (7 chars) that is ambiguous. For example, when I tried 'git am' I got 'error: short SHA1 db8e898 is ambiguous.' Does anyone know how to increase the default length for the shortend sha1?	15:55
	→ kristian42 joined	15:56
shruggar	shortened sha1s are meant to be generated unambiguously. Though, of course, this can only examine the repos in which the patch was generated	15:57
	← simplechat left	15:58
	→ rox` joined	15:58
	← psoo left	15:59
richard421	Ironically the patch file is created in the same repo! I am just testing out the procedure.	15:59
	→ gebi joined	15:59
	← harinath left	16:00
richard421	There are about 200,000 files in the repo. Perhaps that is reaching the uniqueness limit of a 7-char sha1	16:02
mapreduce	git log trunk tells me that refname: 'trunk' is ambiguous.	16:03
	git show-ref \| grep trunk tells me I have refs/heads/trunk and refs/remotes/trunk	16:03
	I'm guessing it's the refs/remotes/trunk one I should be deleting, as I have no remotes.	16:04
shruggar	richard421, hm, then not sure. What does git rev-parse --short db8e898[the full name of the commit you mean] give?	16:04
mapreduce	git reflog delete refs/remotes/trunk tells me that refs/remotes/trunk is Not a reflog. Any suggestions?	16:04
wereHamster	richard421: --full-index	16:04
	(to git format-patch)	16:05
	← panzi left	16:05
shruggar	mapreduce, if I understand you correctly, you're not trying to delete a reflog, you're trying to delete a ref	16:06
	← rox` left	16:06
mapreduce	shruggar: Yes.	16:06
	→ avelino joined	16:06
shruggar	mapreduce, so don't say "git reflog delete" :)	16:06
avelino	hi, help pls: To ssh://git@192.168.1.3/var/www/html/git/test.git ! [rejected] master -> master (non-fast forward)	16:06
mapreduce	shruggar: The error message clued me in on that, but I couldn't see how to go about it.	16:07
	→ plq joined	16:07
shruggar	mapreduce: however, I suspect your actual problem is a /tag/ named "trunk"	16:07
	← midgetspy left	16:07
mapreduce	git branch -d refs/remotes/trunk complains.	16:07
	Hmm.	16:07
plq	hi, is there a way to reverse a merge operation? i mistakingly messed up a public repository	16:07
wereHamster	avelino: faq non-ff	16:07
Gitbot	avelino: Your push would lose changes on the remote. See https://git.wiki.kernel.org/index.php/GitFaq#non-ff	16:07
shruggar	mapreduce, but I am wrong, so ignore me	16:07
mapreduce	Yes, I can't find a tag with that name.	16:08
shruggar	mapreduce, try "git remote rename trunk somenewname"	16:08
mapreduce	Could ot rename config section 'remote.trunk' to 'remote.somenewname'	16:09
	s/ot/not/	16:09
	git remote -v comes back with nothing, so I assume I have no remotes.	16:10
shruggar	I don't know how you got yourself into that situation and I don't know why git-log is complaining about it. git update-ref -d refs/remotes/trunk	16:10
mapreduce	That's the one.	16:10
	I've been told that one at least twice before, thanks.	16:10
shruggar	what did you do to get something called refs/remotes/trunk, though?	16:11
mapreduce	Something to do with using git svn without a --prefix.	16:11
shruggar	I figured git-svn was to blame	16:11
mapreduce	For some reason my co-worker committed to svn instead of git today even though we made the 'final move' a month ago.	16:11
	So I had to pull it through.	16:11
shruggar	muscle memory, I'd assume	16:12
	hobodave_ → hobodave	16:12
mapreduce	I'm not sure he has either. :)	16:12
	→ rox` joined	16:12
shruggar	I'd tell him "no!", hit him with a rolled-up newspaper, and demand he re-commit through git instead	16:12
	← rgr left	16:12
shruggar	if I ever get people to move from SVN, I want it to be permanent :)	16:12
	→ AAABeef joined	16:12
mapreduce	Ok, I cannot push from the git-svn repo to bare (bare was the one that had the duplicate trunk a moment ago).	16:13
shruggar	what's the error?	16:14
avelino	help error: insufficient permission for adding an object to repository database ./objects fatal: failed to write object	16:14
mapreduce	non-fast forward, I know what that means etc.	16:14
sitaram	bob_f: you around?	16:14
	← NET\|\|abuse left	16:14
avelino	How to give this permission?	16:14
	→ harinath joined	16:14
mapreduce	It claims not to be able to fast-forward. I don't really care about the trunk branch, we use master as a central one now. So I created a branch named deleteme in bare, and found that I can't push from the git-svn repo to that via git push bare deleteme	16:14
	src refspec deleteme does not match any.	16:15
shruggar	avelino, are you the user who created the repository?	16:15
sitaram	bob_f: you around?	16:15
Ilari	sitaram: You want summary of what the problem was or something else?	16:15
sitaram	bob_f: your ssh-agent problem is documented in doc/6 [Ilari: if I understood correctly from backlog]	16:15
Ilari	sitaram: Ah...	16:16
	← dpino left	16:16
sitaram	bob_f, Ilari: look for "ssh-agent problems"	16:16
	and the word "quirk"	16:16
	← giskard left	16:16
sitaram	:)	16:16
bob_f	sitaram: Hello.	16:16
	sitaram: Ah cool, it is documented - I was going to suggest that might be a good idea but didn't know if it was in the scope of the docs or not.	16:17
sitaram	bob_f: I feel stupid; I knew exactly what it was but couldnt identify it then (but I am currently out of town in long meetings with big people; some reduction of brain function is expected!)	16:17
Ilari	That's just insane behaviour IMO. Ignoring explicit instructions from user without saying so.	16:18
	→ gnufied joined	16:18
	← eletuchy left	16:18
sitaram	Ilari: who do I complain to? ;-)	16:18
	← sagsousuke left	16:18
Ilari	sitaram: Openssh doesn't take bugreports? :-)	16:18
richard421	wereHamster: --full-index is what I am looking for. thanks!	16:19
	← Jaxan left	16:19
Ilari	git-remote-gits would spew error message and fall back to internal signing code in that case (use key not present in agent).	16:19
bob_f	sitaram: Ha, that's okay - the time I spent trying to solve that problem was time I didn't spend having to do some work on trac, so it's not so bad really. ;)	16:20
	You guys gave me enough to point me in the right direction, we got there in the end.	16:20
	← level09 left	16:21
	→ psoo joined	16:22
avelino	yes shruggar	16:22
	shruggar git clone OK	16:23
	→ krawek joined	16:23
shruggar	avelino, chmod u+rw -R .git should do it, but I don't know how you got into that situation in the first place, so I hesitate to blindly say that	16:24
avelino	shruggar "chmod u+rw -R .git" in server or client?	16:25
Evanlec	probably by pushing as a user without permissions on the remote	16:25
shruggar	avelino: what were you doing when you got the error?	16:26
sitaram	Ilari: sorry for that comment; I didn't mean it literally... just too lazy	16:26
	→ jdroid- joined	16:26
	← richard421 left	16:27
avelino	shruggar "git push master" in client	16:27
shruggar	avelino, are you the user who created the repository you are pushing into?	16:28
	← Nemurenai left	16:28
	← parasti left	16:28
	← barismetin left	16:29
	→ st0w joined	16:29
avelino	shruggar error full http://pastebin.com/5aRDdZQh	16:29
jdroid-	can someone help me understand the following: mkdir test ; cd test ; git init ; git add . ; echo 'word' > test.txt; git add test.txt; git commit -m'first commit'; git branch testbranch; echo 'weeee' >> test.txt'; git checkout master ; cat test.txt	16:30
avelino	shruggar "git clone ssh://thiago@192.168.1.3/var/www/html/git/lol.git" I edited the files I wanted and want to give a push	16:30
shruggar	avelino, but are you the owner of 192.168.1.3/var/www/html/git/lol.git ?	16:30
jdroid-	the last part 'cat test.txt' will show the change i made in testbranch, even though im' in master. Why is that the case?	16:30
	← thatdavidmiller left	16:31
adamv	jdroid-: You made a change to test.txt but did not commit it. Unstaged changes stay around between branch switches.	16:31
tPl0ch	jdroid-, wouldn't you need to use git branch -b?	16:31
shruggar	avelino, or, more to the point, is "thiago" the owner of /var/www/html/git/lol.git on 192.168.1.3	16:31
jdroid-	adamv That's what I thought. But, if I add the change, it goes into whatever branch I have checked out. Correct?	16:33
	adamv does that also mean checking out the file will reset the file back to the version in whatever branch I'm in or does it go back to the version from testbranch?	16:33
	← tPl0ch left	16:34
	← j0shua left	16:34
jdroid-	tPl0ch: Yes. I am helping a friend learn it so Im showing him the small steps first.	16:34
	oh, and he's gone...	16:34
	← trivol left	16:34
krunk-	how do I pull from a windows share? git pull Z:\Projects\proj development	16:35
	← shruggar left	16:35
	krunk- got it	16:36
	→ giskard joined	16:37
avelino	shruggar Not in this group of user thiago	16:37
	→ frank_ joined	16:37
	← bcardarella left	16:39
	IslandUsurperAFK → IslandUsurper	16:41
	→ dl9pf_ joined	16:43
	→ bcardarella joined	16:43
	← bcardarella left	16:44
	← tstclair left	16:45
	← webus left	16:46
	→ eletuchy joined	16:46
	→ dreiss joined	16:46
	← disappearedng left	16:47
	→ rajeshsr joined	16:48
	← Joeconyers left	16:49
	← mtkd left	16:49
	→ Joeconyers joined	16:49
	← Zephyrus left	16:51
	← petrux left	16:51
	→ jtaby joined	16:54
	← Adaptee left	16:55
	← coderdad left	16:55
	→ coderdad joined	16:56
	→ smuf joined	16:59
	← bdiego left	17:00
	← nighthwk1 left	17:00
	→ disappearedng joined	17:00
	→ justin-george joined	17:00
	← coderdad left	17:01
	← jony left	17:01
	→ OFICINA7 joined	17:02
	← spaceonline left	17:03
	→ whatdoesitwant joined	17:03
	← adymitruk left	17:04
	← powr-toc left	17:05
	← ehabkost left	17:05
	← malumalu left	17:05
	→ malumalu joined	17:06
	→ rajeshsr__ joined	17:07
	→ ehabkost joined	17:07
dl9pf_	hi guys ! i see a tricky issue with git if patchnames with ":" are used. e.g. mkdir test: ; cd test: ; git clone -l ../../git -> will fail as it expects the ":" in test: to be part of an ssh url	17:08
	← rajeshsr left	17:08
	rajeshsr__ → rajeshsr	17:08
dl9pf_	s/patchnames/pathnames/	17:08
	← whatdoesitwant left	17:08
	← rox` left	17:09
	→ jeevan_ullas_ joined	17:12
	→ rox` joined	17:12
	→ pantsman joined	17:13
	→ chrowe_ joined	17:13
	→ dlonie joined	17:13
davr	easy solution: stop doing something silly like using : in pathnames	17:13
	← chrowe_ left	17:14
	→ Paraselene__ joined	17:14
	← gnufied left	17:15
AAA_awright	How do large implementations (say, github) store multiple clones of a repository from multiple users without duplicating data?	17:15
PerlJam	AAA_awright: why do you think they don't duplicate data?	17:16
	← dlonie left	17:16
	→ juanjoc joined	17:16
doener	AAA_awright: alternates, or something similar	17:16
	→ adymitruk joined	17:16
	← Paraselene_ left	17:16
	Paraselene__ → Paraselene_	17:16
AAA_awright	PerlJam: Because generally you can clone/push without copying megabytes of data	17:17
gebi	AAA_awright: the sha1 sums of git objects lead to automatic dedublication and github uses a custom backend	17:18
	funny times is we are one day able to produce sha1 collisions :)	17:18
AAA_awright	Ha	17:18
	Are there any notes about their backend at all? I couldn't find any	17:19
gebi	their blog is a nice read, though more about their infrastructure	17:20
doener	dl9pf_: use a file:// url instead of a plain path.	17:21
dl9pf_	davr: no option - wasn't my choice	17:21
	→ jtibu joined	17:22
dl9pf_	doener: tnx, hmm you mean mkdir test: ; cd test: ; git clone -l file:///foo/bar/git ?	17:22
davr	AAA_awright: IIRC github reimplemented the git server from scratch, and it's not open source	17:23
	← jdroid- left	17:23
	→ apsysoev joined	17:23
doener	dl9pf_: yep	17:23
dl9pf_	doener: nop, thats not working	17:23
AAA_awright	davr: Yeah, where did you hear that?	17:24
doener	dl9pf_: and that means what exactly?	17:24
	eats your cat?	17:24
	ah, the hardlinking won't work	17:24
dl9pf_	iirc its transport.c:860 -> if ":" is in -> not local, thus "-l" is pointless and it goes the ssh way	17:24
	→ radioman-lt joined	17:26
	← jeevan_ullas_ left	17:26
davr	AAA_awright: don't know where i first heard it, but this mentions some of it: http://github.com/blog/530-how-we-made-github-fast	17:26
	→ jeevan_ullas_ joined	17:26
AAA_awright	Ah thanks	17:27
dl9pf_	http://pastebin.ca/1888631	17:27
	thats the result	17:27
davr	AAA_awright: looks like some parts are open, some parts are not ("a proprietary library called Chimney...proprietary script called Gerve")	17:27
	← rajeshsr left	17:27
doener	dl9pf_: that's not using file://	17:27
	→ rajeshsr joined	17:27
dl9pf_	ok, sec another run	17:27
	→ andreaa joined	17:28
	→ zenazn joined	17:28
	← deavid left	17:28
dl9pf_	is it 2 or 3 times / ? file:// or file:/// ?	17:29
	→ disappearedng_ joined	17:29
	← disappearedng_ left	17:29
	→ disappearedng_ joined	17:29
	→ deavid joined	17:29
doener	file:// + absolute path, thus: file:///path/to/repo	17:29
engla	file:// is the protocol then / is the root	17:29
	← rox` left	17:31
dl9pf_	good - so this seems to work... hmm what about adding file:// if "-l" is used to clone ?	17:31
	as "--local" means "on filesystem" per definition	17:31
doener	hm? -l does hardlinking, and doesn't work with file://	17:32
	(as I said a few minutes ago)	17:32
	so if -l would imply file:// it wouldn't ever work	17:32
	← disappearedng left	17:33
dl9pf_	http://pastebin.ca/1888638	17:33
	→ rox` joined	17:33
dl9pf_	ok, then not file:// but !ssh	17:33
	without file, it will try to do ssh as it thinks about the ":" being !local url	17:34
doener	dl9pf_: in your test case, just using an absolute path should also work	17:34
	dl9pf_: but of course that still fails when the path to the repo itself contains a :	17:34
dl9pf_	hmm, lets see	17:34
	← frank_ left	17:35
dl9pf_	yes, true	17:36
doener	hm, actually, absolute paths should always work (there's a "slash comes before colon" test)	17:36
	I wonder why that failed for me	17:36
	← segher left	17:36
dl9pf_	hmm, you mean transport.c is_local ?	17:37
doener	yup	17:37
dl9pf_	there's "return !colon \|\| ( slash && slash < colon ) \|\| dos...	17:38
	→ bartek joined	17:38
dl9pf_	so !colon wins - right ?	17:38
	→ segher joined	17:38
bartek	Hi there. If I just did a git pull from master on an existing local branch - How can I revert the changes? It merged master into my branch, and .. I don't want that (wasn't thinking when I pulled :( )	17:38
	→ mtkd joined	17:38
doener	dl9pf_: which is uninteresting, because we're having a colon in the path	17:38
wereHamster	bartek: git reset --hard HEAD@{1}	17:39
	← necromancer left	17:39
	→ fai\|safe joined	17:39
	← fai\|safe left	17:39
bartek	phew, thanks wereHamster	17:39
	→ fai\|safe joined	17:40
	← fai\|safe left	17:40
	← jeevan_ullas_ left	17:40
	→ jeevan_ullas_ joined	17:40
	→ thiago_home joined	17:42
	→ kenichi joined	17:43
	← Chaser left	17:44
CaMason	my git gui text has turned super-large. It does this occasionally, and I need to restart to fix it	17:44
	Anyone know of a fix without a restart?	17:44
	→ frank_ joined	17:45
	← jeevan_ullas_ left	17:46
	→ jeevan_ullas_ joined	17:46
	→ Hanumaan joined	17:47
	→ Chaser joined	17:48
	→ gnufied joined	17:50
	← ChUrP left	17:50
	← bluenovember left	17:52
	← felipec left	17:52
	→ ChUrP joined	17:53
	← Fandekasp left	17:53
myosui	wereHamster, There is an easy thing to perform exactly what I'm looking for. git checkout -b tmp hash_of_the_deletion_last_commit; git push origin tmp:my_branch/tmp; git checkout -t -b admin origin/admin; git merge --no-commit -s ours origin/my_branch/tmp; git commit -a; git merge --no-commit origin/my_branch. Now I only merge the changes to the files present in my_branch after the initialization of the branch.	17:54
	← jeevan_ullas_ left	17:54
	→ Fandekasp joined	17:55
wereHamster	myosui: why the push?	17:55
	← mastro left	17:56
myosui	wereHamster, Just to stay with that branch in repos for historical reasons.	17:56
	← nevyn^_ left	17:57
myosui	wereHamster, But you're right, we don't need the push. But it works pretty well. Now I can perform all the merges that I want that git doesn't bother with the older commits that are not present in admin. Super!	17:57
	→ damog joined	17:59
myosui	wereHamster, I must thank you for your insight. It have been very helpfull. Even yesterday. Thanks a lot!	17:59
	→ voxpelli-laptop joined	18:00
	→ rlb3 joined	18:01
	→ Shurakai joined	18:03
	→ sh1mmer joined	18:03
	← voxpelli-laptop left	18:03
	→ g3d joined	18:03
	← sh1mmer left	18:03
	→ voxpelli-laptop joined	18:03
	← rajeshsr left	18:04
	← tatsuya left	18:05
	→ smik joined	18:07
	→ tstclair joined	18:07
smik	How to list all the files that changed from HEAD^ to HEAD	18:07
	→ Guaguasi[work] joined	18:08
gebi	my first (and pretty ugly one) would be: git diff ... \|lsdiff	18:09
DrNick	git diff --name-only HEAD^ HEAD	18:09
gebi	ah, good to know :)	18:09
smik	DrNick: That is so good to know	18:11
	→ Stoqtech joined	18:11
smik	DrNick: can't git status do something similar?	18:11
	← JackWinter left	18:11
Stoqtech	how do i remove a branch?	18:11
DrNick	smik: nope	18:11
Stoqtech	on github	18:11
	→ sgh joined	18:11
smik	DrNick: thank you anyway	18:11
	← jtibu left	18:13
smik	Anyone here knows of a FTP cliend (text based) that can take multiple files name as input and upload them automagically	18:13
	← rlb3 left	18:13
avar	lftp?	18:13
teuf	smik: lftp, ncftp should be able to do that	18:13
smik	teuf: avar : And man lftp could provide me everything that I would need to know about them right?	18:14
	← gebi left	18:15
	→ jtibu joined	18:15
avar	smik: try it	18:15
teuf	or help once lftp is running	18:15
	→ coderdad joined	18:15
wereHamster	Stoqtech: git push origin :<the branch name>	18:15
adymitruk	git svn clone fails on a fatal error.. anyone else had this happen? (3000 revisions into the process :(	18:15
	→ ChUrP1 joined	18:16
smik	avar: teuf I'll figure it out. thank you. Basically I wanted to update my website which is Git monitored. I wanted something so that I can 1) get the list of the modified files and then upload them on my server.	18:16
	Anyway thanks for the help	18:16
Stoqtech	wereHamster is the colon important?	18:16
wereHamster	yes	18:16
Stoqtech	in my case it should be git push origin :master ?	18:16
	ok thanks	18:16
	it says deletion of current branch prohibited	18:17
	:(	18:17
adymitruk	-f	18:17
Stoqtech	?	18:17
adymitruk	'force' option	18:17
	is it a bare repo that you're pushing to?	18:18
	← bartek left	18:18
	← jbd left	18:18
Stoqtech	no adymitruuk	18:18
	← ChUrP left	18:18
adymitruk	you need '-f' then	18:18
	.. or '--force' if you prefer	18:18
wereHamster	Stoqtech: it is a bare repo if you push to github	18:18
	← adamv left	18:18
	→ tazz joined	18:19
	← tazz left	18:19
Stoqtech	ok wereHamster :p then it is	18:19
	i dont know what im talking about	18:19
	→ Alagar joined	18:21
Stoqtech	thanks the -f seems to have worked	18:21
	← gnufied left	18:22
	← Hanumaan left	18:23
	← paja left	18:25
	← OFICINA7 left	18:26
	→ OFICINA7 joined	18:28
Guaguasi[work]	Hi. I'm attempting to convert an svn repository to git and I'm having quite the time of it. I have a production code base that is not at the same commit as the repository (and has files of different revisions strewn about) and a development code base that is up to date and has commits pending. Unfortunately the svn repo wasn't branched or anything. My problem is that I want the fubar'd production code to be the master branch and then the	18:29
	current repo/dev code to be the development branch, but I'm at a loss on how to do that. I used this as a starting point: http://tinyurl.com/ye7syzp	18:29
	→ Durl joined	18:29
	→ codetroll joined	18:29
	← bentob0x left	18:31
	→ rajeshsr joined	18:31
	← sgh left	18:31
	→ sgh joined	18:32
Stoqtech	git push origin :master hasn't worked	18:32
PerlJam	Guaguasi[work]: the production and development code bases .. are they in git?	18:32
Stoqtech	i still have all my commit history	18:32
	i want to wipe my github clean and repush the first commit	18:33
	← apsysoev left	18:33
Guaguasi[work]	It's all svn currently, I'm trying to get them over to git	18:33
	→ thierryp joined	18:33
OFICINA7	can somebody help me here?	18:34
adymitruk	Stoqtech: Get the repo looking like you want locally, then do a git push -f origin master	18:34
	that would do it	18:34
OFICINA7	I'm trying to have 2 repos with the same local folder	18:34
Guaguasi[work]	I wouldn't have a problem if I just moved the dev to git, but the production code is what's killing me	18:34
Stoqtech	ok adymitruk thanks	18:34
PerlJam	Guaguasi[work]: things get easier if you have git repos. I'd probably start by converting the svn repos to git, then try to merge them together and fix them up.	18:34
	Guaguasi[work]: or, transition dev to production first, then try to migrate to git.	18:35
	→ apsysoev joined	18:35
	→ wjames6 joined	18:36
Stoqtech	i still see all my commit history	18:37
wjames6	hey - i have a git server, and it has a rsa key from my laptop, but i dont have my latop - so i want to add another key. circle jerk, how can i add the second key manually (since i have root on the server)	18:37
Stoqtech	nevermind	18:37
Guaguasi[work]	PerlJam: That's the issue, though, I'm not sure how I would get the production code fixed. I'm at the point where I have a clean local git repository I can push to a new remote one, but then I would screw myself with the production code, wouldn't I?	18:37
	→ ereslibre joined	18:38
	→ skoop joined	18:39
PerlJam	Guaguasi[work]: dunno. It's unclear to me.	18:40
	← thierryp left	18:43
	← Stoqtech left	18:44
	→ coppro joined	18:44
	→ jhelwig joined	18:44
	← jhelwig left	18:44
	← thiago_home left	18:45
	← ereslibre left	18:45
	→ dl9pf joined	18:45
	← dl9pf left	18:45
	→ dl9pf joined	18:45
	← squentin left	18:45
	← tobiassjosten left	18:45
	← dl9pf_ left	18:46
	→ jhelwig joined	18:46
	→ thiago_home joined	18:47
	→ Nickeeh joined	18:47
avelino	How to give permission to a user in a repository?	18:47
Guaguasi[work]	PerlJam: Basically, following the guide I linked, I have a perfect git repository with all the latest commits. The production code, however, is nowhere near that and I'm unsure how I would move that over to work with the new git repo. I'm considering creating the dev repo, cloning the old commit I need, and just copy/pasting the files from the production code and pushing that to the master branch. But one, I don't know if that's possible	18:48
	, and two, if I were to do that, how would I merge or rebase or whatever the branches so I can continue to work properly?	18:48
	→ Stoqtech joined	18:48
	← voxpelli-laptop left	18:49
Stoqtech	sorry my computer overheated and died	18:49
	so i did git push -f origin master and it did overwrite everything	18:49
	but the commit history is still there	18:49
	← harinath left	18:49
	← Shurakai left	18:50
	← drizzd left	18:52
cbreak	Stoqtech: after a git push -f, the history in the branch you push to will be the same as in the branch you are	18:52
	→ vinse joined	18:52
	→ drizzd joined	18:53
	→ benmanns joined	18:53
	← rayners left	18:54
	← engla left	18:54
	← rajeshsr left	18:55
Stoqtech	aah cbreak	18:56
	so i can delete the history locally	18:56
	how can i do that?	18:57
cbreak	it's called history rewriting. Don't do it if you don't know what you're doing	18:57
	Stoqtech: man git-rebase -i	18:57
Gitbot	Stoqtech: the git-rebase manpage can be found at http://git.or.cz/man/git-rebase	18:57
cbreak	the consequences of not knowing what you do can be ... severe	18:57
Stoqtech	they cant be that severe cbreak...im just playing	18:57
	i can always delete the whole repository and start again	18:58
	→ masterkorp joined	18:58
cbreak	the severity isn't about your repo	18:58
	but the repo of those who clones/fetched from it	18:58
	← plq left	18:58
	adamholt → adamholt_away	18:58
	adamholt_away → adamholt	18:58
Stoqtech	yeah cbreak..no-one is pulling from my repo at the moment	18:59
	→ rlb3 joined	18:59
	→ ereslibre joined	19:00
	→ GorillaPatch joined	19:01
	→ francisfish joined	19:03
	← frsk left	19:04
	← avelino left	19:04
Stoqtech	im trying git branch -d branchname but it wont let me as im in the repository	19:05
	how can i delete it?	19:05
	← smik left	19:05
	← curtana left	19:05
	→ frsk joined	19:05
	← _Vi left	19:06
	→ _Vi joined	19:07
	→ smik joined	19:07
cbreak	Stoqtech: check out some other branch?	19:07
	← Nickeeh left	19:07
	→ zyphlar joined	19:08
Stoqtech	i dont understand, sorry	19:08
cbreak	type git status	19:08
zyphlar	hello fellow gitters	19:08
cbreak	it will tell you on which branch you are	19:08
Stoqtech	im on master	19:08
	← rlb3 left	19:08
cbreak	which branch do you want to delete?	19:08
Stoqtech	master	19:08
zyphlar	where would i go to find out what "<filename> needs update" means after having typed "git reset <commit> <filename>" ?	19:08
cbreak	you can't delete the branch you are on	19:09
	delete some other branch	19:09
Stoqtech	hmm :(	19:09
	← Gitzilla left	19:09
Stoqtech	all i want to do is remove all the history and re-do my first commit	19:09
cbreak	why do you want to delete master?	19:09
	start a fresh repository	19:09
	→ alester joined	19:09
Stoqtech	hmmm there must be a way of just re-initialising or something	19:09
cbreak	maybe you can rm -rf .git, and then just git init	19:09
Stoqtech	ok :)	19:10
	thats an idea	19:10
cbreak	don't type rm -rf / by accident	19:10
Stoqtech	i didnt :)	19:10
	but thanks	19:10
zyphlar	any idea on what "<filename> needs update" means?	19:11
cbreak	zyphlar: try git status. what does it say?	19:11
	→ rajeshsr joined	19:12
zyphlar	changes to be committed: new file <filename>	19:12
	changed but not updated: deleted <filename>	19:12
	(also <filename> doesn't actually exist according to my webserver leading me to wonder)	19:13
	i'm trying to "restore from backup" about 5 files from 2 commits ago	19:13
coppro	checkout HEAD^2:filename	19:14
zyphlar	why would google lead me to git reset?	19:14
cbreak	you can use git reset too	19:14
tango_	the changes you mention show that you (1) created the file (2) git added it (3) removed the file	19:14
cbreak	git checkout HEAD^2 -- filenames	19:14
	or git reset --hard HEAD^2 -- filenames	19:14
zyphlar	that didnt happen though	19:14
cbreak	oh, sorry	19:15
	that's wrong	19:15
	HEAD~2	19:15
	not ^2	19:15
tango_	git checkout is the way to go, I would say	19:15
zyphlar	hmm k	19:15
	will that create a new commit containing only the "new file" ?	19:15
	← rajeshsr left	19:15
	→ ehabkost_ joined	19:16
cbreak	only if you git commit it	19:16
	→ ryah joined	19:16
	→ LongBeach joined	19:16
cbreak	you probably used git reset without --hard, so it didn't touch the working dir	19:16
ryah	i want a command to find out what files a particular author has touched	19:17
zyphlar	k	19:17
ryah	or alternatively - given a file a command that lists the authors who have touched it	19:17
zyphlar	error: pathspec 'c1089ec69fd9bbbe019d5cc72ca3f4e502338f34:files/3512/6332/7344/dv-a1.jpg' did not match any file(s) known to git.	19:17
	not supposed to use a colon?	19:17
cbreak	ryah: man git-blame	19:18
Gitbot	ryah: the git-blame manpage can be found at http://git.or.cz/man/git-blame	19:18
	→ gebi joined	19:18
cbreak	ryah: man git-log --author	19:18
Gitbot	ryah: the git-log manpage can be found at http://git.or.cz/man/git-log	19:18
	→ freeride joined	19:19
	← ehabkost left	19:20
	→ shruggar joined	19:20
zyphlar	git checkout <commit> -- <filename> worked	19:22
	is there a way to do multiple filenames? just separated by spaces?	19:22
OFICINA7	please, how can I have 2 repo (dev and pro) and 2 branches (master and pro)????	19:22
cbreak	zyphlar: try.	19:22
zyphlar	i am le scared ;)	19:23
	thanks though great help	19:23
cbreak	OFICINA7: by creating them	19:23
	but you should maybe use different names	19:24
Stoqtech	cbreak, thanks your idea worked	19:24
OFICINA7	please, from the begining, how many folders should I have for this app?	19:24
	1 for dev + 1 for pro?	19:25
	or can I use the same "myapp" folder?	19:25
cbreak	depends	19:25
	← freeride left	19:25
cbreak	how many folders do you want?	19:25
	one repo is one folder.	19:25
	obviously, a repository can not be the same as an other repo	19:26
OFICINA7	anyone familiar with #heroku ?	19:26
	are you* familiar with #heroku ?	19:26
cbreak	one repository can create more branches than you will ever need.	19:26
	I have no idea what that is	19:26
OFICINA7	ok, that's another github...	19:27
	so i can't have 2 repos inside the same .git folder?	19:27
	→ parasti joined	19:28
	← lhz left	19:28
OFICINA7	cbreak	19:28
cbreak	no. of course not.	19:28
OFICINA7	ok, new thing the noob's just learned	19:28
cbreak	that's like asking if you can have two houses in the same house.	19:28
OFICINA7	ok	19:29
cbreak	you can use submodules to have repos _inside_ the other	19:29
OFICINA7	I need to put my development files inside [email@hidden.address]	19:29
cbreak	but they are trouble	19:29
OFICINA7	and my production files inside [email@hidden.address]	19:29
	← tar- left	19:29
OFICINA7	say my dev repo has 2 branches... master and production	19:30
	I type git commit -a -m "sending to production"; git push dev master	19:31
	and my files go to my development repo, right?	19:31
	→ Rayne joined	19:31
cbreak	why?	19:32
Rayne	if i rebase a repo i just have to pull from it to get all "clients" rebased, too? or does the rebase only affect the repo?	19:32
cbreak	why do you make a new commit about sending stuff somewhere?	19:32
	that's stupid	19:32
OFICINA7	let me show you my app	19:32
	http://coolapp-dev.heroku.com/	19:32
	this is the development URL	19:33
	http://coolapp.heroku.com/ and this is production	19:33
	each one is a repo	19:33
	I want to do my nightly changes to my dev repo, and see it in the URL	19:33
	→ sylr joined	19:33
	→ Adaptee joined	19:33
OFICINA7	later, I say "yeah, this is good for production" and I want to dump one repo into another	19:34
	how can I do that?	19:34
	← wjames6 left	19:34
OFICINA7	the production repo has no intention of having many branches, it's just to put the production files	19:34
	← ntoll left	19:35
	← jrmuizel left	19:35
OFICINA7	because when I put my files there, it automatically deploy them to the production website	19:35
wereHamster	OFICINA7: git clone [email@hidden.address] cd coolapp-dev; git push [email@hidden.address] HEAD:master	19:35
	→ HG` joined	19:35
	← MattDiPasquale left	19:36
OFICINA7	wereHamster: can I have 2 repo inside the same .git?	19:39
	git remote add dev [email@hidden.address] git remote add pro [email@hidden.address]	19:40
	← Durl left	19:40
wereHamster	then just git fetch dev master; git push pro FETCH_HEAD:master	19:40
cbreak	you can't. a .git folder IS a repository	19:40
	it does not contain repositories.	19:40
	← ryah left	19:41
	→ MattDiPasquale joined	19:41
	→ Durl joined	19:41
OFICINA7	wereHamster what's a FETCH_HEAD anyways?	19:41
wereHamster	it's whatever you fetched last	19:42
	← ehabkost_ left	19:43
	← resmo left	19:45
OFICINA7	I see	19:45
	I'm experimenting it	19:45
	I could first push to my pro for the first time	19:46
	→ Nemurenai joined	19:46
	← gbacon left	19:46
	← whaley left	19:47
OFICINA7	wereHamster: how should I use the merge command to send files locally from dev to pro?	19:47
	← ericdc left	19:47
OFICINA7	checkout master; merge pro; ?	19:47
	→ sixteneighty joined	19:47
	→ gbacon joined	19:47
	← zyphlar left	19:47
cbreak	don't push to non-bare repositories.	19:48
	just a note :)	19:48
	← ereslibre left	19:48
	→ ereslibre joined	19:49
	→ ericdc joined	19:50
OFICINA7	cbreak please help me do a merge?	19:50
	→ friskd joined	19:50
OFICINA7	git checkout master; git merge pro; right??	19:51
cbreak	OFICINA7: man git-merge	19:52
Gitbot	OFICINA7: the git-merge manpage can be found at http://git.or.cz/man/git-merge	19:52
cbreak	is anything unclear?	19:52
	someone in here has a link to a git workflow webpage	19:52
	← gebi left	19:52
cbreak	I think it could help you, but I am not that one :/	19:52
OFICINA7	I can't merge =/	19:52
cbreak	why?	19:53
	→ Acorn joined	19:53
OFICINA7	because I'm a dumb noob	19:53
Acorn	Hi, I'm just starting with git and am having problems cloning a ssh git repo, I can connect fine with putty	19:54
	← ereslibre left	19:54
Acorn	I get "Network error: Cannot assign requested addressInitialized empty Git repository in .."	19:54
	"fatal: The remote end hung up unexpectedly"	19:55
	what does that mean?	19:55
cbreak	what did you do?	19:55
Acorn	i'm using tortoisegit, right clicked in a folder and did clone	19:55
	put the ssh url in the field and browsed to the private key	19:56
cbreak	what url?	19:56
Acorn	it's privately hosted	19:56
	ssh://username@domain.de/~/repo.git	19:57
cbreak	my current working hypothesis is that either the repo you want to clone from is wrong	19:57
	(maybe it's empty?/not initialized?)	19:57
	or that your URL is wrong	19:57
	← raichoo left	19:58
OFICINA7	weird...	19:58
	being on MASTER, and typing GIT MERGE PRO doesn't send MASTER files to PRO	19:58
	however	19:58
Acorn	if I connect with putty to domain.de with the private key and login with username I get to a folder with the repo.git folder	19:58
jast	OFICINA7: merge works the other way around	19:58
OFICINA7	being on PRO and typing GIT MERGE MASTER brings MASTER files to PRO	19:59
cbreak	OFICINA7: git is case sensitive	19:59
jast	merge always works on the currently checked out branch	19:59
OFICINA7	cbreak I'm just collouring in irc	19:59
cbreak	merge merges the named thing into your current branch	19:59
jast	because that's the only way it can allow you to fix conflicts	19:59
	→ resmo joined	19:59
OFICINA7	cbreak, jast, yeah that was the big bug	20:00
	← lucsky left	20:00
OFICINA7	=/	20:02
	now another basic thing	20:02
	say my PRO branch is really nice	20:02
	→ raichoo joined	20:02
OFICINA7	and I have ruined my MASTER branch	20:02
	how do I say "send PRO to MASTER"?	20:03
cbreak	Acorn: weird. try absolute path?	20:03
wereHamster	OFICINA7: which master?	20:03
cbreak	OFICINA7: you can't send stuff to other branches	20:03
wereHamster	and what is pro?	20:03
cbreak	at least not on the same repo	20:03
OFICINA7	PRO and MASTER are my branches	20:03
wereHamster	git checkout master; git merge -s ours master; or git checkout master; git reset --hard pro;	20:04
	← tstclair left	20:04
	← Alagar left	20:04
OFICINA7	cbreak I only need a second repo to be known, so I can push local files to a second repo	20:04
cbreak	you can't push files, only commits	20:05
	← HG` left	20:05
OFICINA7	cbreak thx[	20:05
	wereHamster	20:05
	so...	20:05
	→ ry joined	20:05
ry	gitolite question	20:06
	→ adamv joined	20:06
ry	lost my keypair, trying to add another key, using gl-emergency-key does not work	20:06
OFICINA7	be inside master, and type, git reset --hard pro --> "bring me PRO commit and don't ask me why"	20:06
	← masterkorp left	20:06
OFICINA7	wereHamster right?	20:06
ry	create mode 100644 keydir/ry@imac.pub /usr/local/bin/gl-emergency-addkey: line 40: src/gl-compile-conf: No such file or directory	20:06
	I ran gl-compile-conf manually, altho that line does worry me	20:07
wereHamster	OFICINA7: yes, however you shoujlb be aware of the consequences	20:07
Acorn	cbreak, same error	20:07
OFICINA7	wereHamster ok, I'm listening	20:07
ry	for some reason it keeps adding my OLD key, not my new key	20:07
OFICINA7	wereHamster: http://coolapp-dev.heroku.com/ and http://coolapp.heroku.com/	20:08
wereHamster	OFICINA7: -> google, or man git-reset	20:08
Gitbot	OFICINA7: the git-reset manpage can be found at http://git.or.cz/man/git-reset	20:08
OFICINA7	wereHamster thank you so much	20:08
wereHamster	OFICINA7: also, if you use reset --hard, you can run into faq non-ff later	20:08
Gitbot	OFICINA7: Your push would lose changes on the remote. See https://git.wiki.kernel.org/index.php/GitFaq#non-ff	20:08
jast	hooray for gitbot :)	20:08
OFICINA7	so I should just be aware	20:09
	→ crescendo joined	20:09
OFICINA7	that if I am bringing my master code back in time by using another branch as a source, I should remember I have new files and should push them to my repo under the master branch?	20:10
crescendo	Hello, I have a process-oriented question: if I have a project that includes 5+ open source projects, what makes the most development sense? A single repository containing all of them, or multiple individual repositories?	20:10
	← benmanns left	20:11
	→ oskude joined	20:11
wereHamster	crescendo: multiple repos if the projects, and submodules or subtree merge	20:11
	→ pflanze joined	20:11
	← pantsman left	20:11
	← rseifert left	20:11
pflanze	Is it possible to do multi-line greps with git-grep?	20:11
jast	OFICINA7: most importantly, if anyone else has a copy of the old master branch, it's pretty much impossible to automatically reconcile the two versions in a meaningful way. if nobody else has a copy of the broken history, there's nothing to worry about, though.	20:11
OFICINA7	Gitbot are you actyally a bot?	20:11
jast	OFICINA7: yes he is	20:11
	← snucko left	20:12
	→ impulze joined	20:12
jast	pflanze: Try And See(tm) ;)	20:12
OFICINA7	then he's the nices bot, he speaks almost humanly	20:12
pflanze	I'd like foo.bar to match if . spans say up to 5 lines	20:12
oskude	hello, is it ok to have only * and !.gitignore and exact files i want in .gitignore ?	20:12
OFICINA7	jast, so what's the correct way to make changes?	20:12
Acorn	how do I set my openssh key for gitgui?	20:14
OFICINA7	master -> development-with-no-head-ache, adding-something -> development-with-headache, and pro -> production?	20:14
pflanze	jast: git grep -C5 'foo.*bar' does not do that. So I tried and didn't find a way, but this doesn't mean I haven't missed a possibility (some flag or don't know).	20:14
Acorn	(in windows 7)	20:15
	do I need to store it in my user folder somewhere?	20:15
	← sixteneighty left	20:15
	→ sgh_ joined	20:16
	← sgh left	20:16
	→ JackWinter joined	20:20
	→ smuf_ joined	20:21
jast	pflanze: '.' may default to matching everything except for newlines; it does in perl	20:21
	← malumalu left	20:21
	← oskude left	20:21
pflanze	jast: and how would you say newlines, then`	20:21
	?	20:21
	← francisfish left	20:21
	→ dpino joined	20:21
	→ francisfish joined	20:21
pflanze	Or, how would you say "not default"? ;)	20:22
jast	though in POSIX grep . should match anything	20:22
	POSIX regexps, that is	20:22
	← ry left	20:22
	→ tstclair joined	20:23
	→ akahn joined	20:23
	← smuf left	20:23
akahn	What would be the most straightforward way to find out if the current branch has been pushed to a remote?	20:23
wereHamster	git branch -vv	20:24
	→ p3rg2z joined	20:24
Acorn	where should the git ssh config stuff go in the installation folder in windows?	20:24
	→ bambule joined	20:24
jast	pflanze: hmm, the original grep searches per line, I suspect git copied that behaviour	20:24
wereHamster	to 'a' remote or to its currently configure d remote?	20:24
	← frank_ left	20:24
	← kpfleming left	20:24
wereHamster	akahn: ^^^	20:24
Acorn	are, the root of the git folder worked \o/	20:24
	→ smuf joined	20:24
	← airborn left	20:25
	→ pantsman joined	20:25
Acorn	It's using my current windows username as the user, why? I set the user setting in ssh_config	20:25
JackWinter	i have a local wine repo that i track wine dev from, have some of my own branches to. normally do a git fetch and a git rebase origin. i've also put a copy of a local branch on git://repo.or.cz/wine-rt.git, now when i try to push i get this http://pastebin.org/351952 how to fix?	20:26
	← francisfish left	20:26
jast	Acorn: but other settings in the same file work?	20:26
	for the same host, I might add	20:26
Acorn	jast, well it's translating the host into the right address	20:26
	← psoo left	20:27
	→ smuf__ joined	20:27
	← smuf_ left	20:27
jast	hmm.	20:27
Ilari	JackWinter: Rebase rewrites history.	20:27
jast	I don't expect any glaring errors, but can you paste the config file anyway (feel free to censor things out), just to be sure?	20:28
	pastebin, that is	20:28
	→ tobiassjosten joined	20:28
JackWinter	Ilari: ah i think i remember, i used -f or some such to force the update. no one else has write access to the repo, so think i'll try that again	20:29
	→ Belna joined	20:29
Rayne	does git use file system features (except hardlinking/linking)? i ask this because i want to initialize a repo on a ntfs partition (with linux + ntfs-3g)	20:29
jast	JackWinter: in short, the two versions of newrt have different changes on them. I suggest looking at their respective histories (e.g. gitk newrt origin/newrt) to see what's going on.	20:29
	← smuf left	20:30
	← jtibu left	20:30
	← GorillaPatch left	20:30
Acorn	jast, actually, I think the settings aren't working as the user was something that could be interpreted as a domain name, so it was probably actually connection to it as a domin	20:30
	where should the settings file be?	20:31
jast	Rayne: ntfs isn't an ideal choice but it will probably work. easy to find out, at any rate. :)	20:31
	← g3d left	20:31
	← codetroll left	20:31
jast	Acorn: no idea. try your system profile directory...	20:31
Acorn	C:\Program Files (x86)\Git\ssh ?	20:31
JackWinter	jast: the local one is the one is use to track the official wine and to do my work on, the repo.or.cz is the one i created for publishing something. so think it will be ok to force the update	20:31
akahn	wereHamster: hm, i'm not sure	20:31
Guaguasi[work]	Say I have a branch 10 commits and I'm trying to create another branch starting from commit 5 and ignoring the rest, is that possible?	20:31
jast	Guaguasi[work]: "starting from commit 5" meaning that it only has the commits that are newer than 5?	20:32
Guaguasi[work]	jast: no, commits 1-5	20:32
	← kristian42 left	20:32
Acorn	jast, found it! let's see if it works	20:32
	→ saikat joined	20:33
	← ziro` left	20:33
Acorn	jast, damn, no address associated with name	20:33
saikat	when you guys fork a repository, is it common practice to add that repo as just another remote on your local clone of the original repo?	20:34
	← DavidKlein left	20:34
Acorn	jast, I put the ssh_config file where the known_hosts file was	20:34
saikat	it seems like the only way to do things if i want to see what changes i have made in my fork of the repo vs. the original repo	20:35
Acorn	jast, do I have it the right way round, the host is the alias you give it and the hostname is the url?	20:35
	← froschi left	20:36
	→ flaguy48 joined	20:37
	← Belna left	20:37
	adamholt → adamholt_away	20:38
	adamholt_away → adamholt	20:38
	→ stamina_ joined	20:40
	→ masterkorp joined	20:41
	→ bulbous joined	20:42
	→ johnw joined	20:43
Acorn	when I try and ssh it gets to line 4 in my ssh_config file and says it's a bad number 'Port [22]'	20:43
	what's wrong with it?	20:43
cbreak	numbers have no [] around	20:44
	→ jrmuizel joined	20:44
Acorn	Oh!!!	20:44
	I've been putting everything in square brackets because the guide I was following had them	20:44
	doh	20:44
	→ squentin joined	20:45
	← squentin left	20:45
	→ squentin joined	20:45
cbreak	in man pages, [this] means that this is optional.	20:45
	usually.	20:45
	→ \mjd joined	20:45
bulbous	Hey all -- got a tagging question, have looked all over, can't find the answer. I have tagged a commit using git tag -- then pushed it using git push --tags -- then I can get it on a different repository using git fetch --tags. All that works great. when I delete the tag from the origin using git push origin:refs/tags/tagname -- that appears to work also, mostly... new clones don't have the removed tag, but any clones that have already fetc	20:45
	that tag don't delete it when they do a new fetch	20:45
	← Nemurenai left	20:46
bulbous	hopefully that made sense -- after deleting the tag 'testtag' on the remote, I would expect that a new: git fetch --tags would result in testtag being deleted on my local repository, but that doesn't appear to be the case.	20:46
jast	bulbous: yeah, tags are not really revision-controlled. git doesn't have a a mechanism for removing tags that no longer exist on upstream.	20:47
cbreak	tags are local.	20:47
	← urkud_ left	20:47
\mjd	I don't understand how tag objects are shared between repositories. I want to tag a certain revision with a tag object T. then I want to git-push origin and have T get topied to origin, and elsewhere I want to git-fetch origin and have it tell me where T is now.	20:47
cbreak	if you want to remove them, you have to do it local.	20:47
bulbous	I was afraid of that, ahh well, thanks :)	20:47
Acorn	damn, it's still asking me for a password, and it asks me for a passphrase for my ssh key even though I didn't set one	20:47
	← cilly left	20:48
bulbous	not a huge deal, just a minor nuisance	20:48
\mjd	Sometimes I move T with git-tag -f T HEAD and then when I try to push the change to origin, it just says "everything up-to-date".	20:48
cbreak	tags are immutable	20:48
	you can not change where they point to	20:48
jast	Acorn: especially the latter shouldn't happen... and it certainly shouldn't ask you for both password and passphrase	20:48
bulbous	\mjd: you need to do git push --tags	20:48
\mjd	bulbous: thanks, that's what i wanted.	20:48
	→ bobrik joined	20:48
Acorn	jast, I just pressed enter when it asked me for a passphrase, and then it asked me for a password for the user account (there is none)	20:48
jast	ah	20:49
\mjd	git fetch --tags was what I neeed. bulbous: thanks again	20:49
jast	well, it shouldn't ask for a passphrase, but if it think it needs one, that makes sense	20:49
bulbous	\mjd: glad i could help a bit :)	20:49
jast	+s	20:49
cbreak	jast: encrypted private key?	20:50
Acorn	jast, also, why do I need to use the -F flag? why doesn't it find the config file autmatically?	20:50
cbreak	man ssh-add	20:50
jast	Acorn: use the -F flag where?	20:50
cbreak	maybe	20:50
Acorn	jast, git bash when using the ssh command	20:50
jast	cbreak: he said he didn't set a passphrase for the key	20:50
bulbous	\mjd: you can automatically push tags also if you add the following lines to your remote:	20:50
	push = refs/heads/:refs/heads/	20:50
	push = +refs/tags/:refs/tags/	20:50
jast	well, I suspect that the file is not in the right place	20:50
cbreak	hmm.	20:50
\mjd	bulbous: thanks.	20:51
jast	as I said, I'd try the user profile directory	20:51
	→ Ins\|de joined	20:51
	← Adaptee left	20:51
Acorn	the thing is, when I tried to connect it saved their ssh key to a know_hosts file, and that's where I put my files	20:51
	jast, that's where it is, .ssh folder in the user profile directory	20:51
	→ Adaptee joined	20:52
Acorn	and it's finding it there	20:52
Ins\|de	hello, after rebasing an already pushed commits (yeah, i know it's dangerous, but no one else has cloned or pulled the repo) how can i rewrite the new history ?	20:52
cbreak	you did	20:52
	rebase -i is one of the commands that rewrite history	20:53
	→ beatak joined	20:53
jast	Acorn: if it's finding it there, what do you need -F for?	20:54
beatak	what is the command to see the list of files that are different between two branches?	20:54
	← kumbayo left	20:54
cbreak	beatak: man git-diff	20:54
Acorn	jast, oh right, it's not finding it then :(	20:54
Gitbot	beatak: the git-diff manpage can be found at http://git.or.cz/man/git-diff	20:54
jast	beatak: hint: --name-only	20:54
\mjd	Ins\|de: You can push the rewritten history by using the -f option to git-push.	20:55
crescendo	is there a way to do a "git pull" test, of sorts, that only shows what would change?	20:55
beatak	Thanks gurus	20:55
cbreak	crescendo: git fetch instead	20:55
Ins\|de	\mjd, ok, thanks!	20:55
cbreak	and then diff to the remote tracking branch	20:56
	→ Twisted_Mentat joined	20:56
\mjd	crescendo: git fetch will acquire the commits without trying to update your local branches. Then you can use (for example) git diff master origin/master to see what will change.	20:56
crescendo	Aha, perfect. thanks.	20:56
jast	crescendo: you can git fetch followed by something like gitk ...@{u}	20:56
	(yeah, cryptic, I know)	20:56
	→ urkud_ joined	20:56
\mjd	crescendo: I also like git show-branch master origin/master	20:56
	→ necromancer joined	20:56
flaguy48	Acorn: use ssh -v -v to get a lot more info about what is going on.	20:57
	← beatak left	20:57
rane_	ssh -vv	20:57
	→ psoo joined	20:57
jast	in fact, make that "a LOT more"	20:57
crescendo	\mjd, hrm. I'm just putting together a plan for a production "pull", essentially. we haven't even initialized our git repo yet, but we just rsync'd it to the development server and are preparing to import it as "production"	20:58
	from there, I want to test a pull of the same code, to verify that I have something I can roll back to later if this breaks.	20:58
	...without breaking production today.	20:58
cbreak	how can you rsync something that doesn't exist?	20:59
\mjd	crescendo: You can git fetch the remote stuff, then check it out and test it.	20:59
Acorn	it says PEM_read_Privatekey failed	20:59
\mjd	crescendo: git fetch origin; git checkout origin/master; test	20:59
crescendo	cbreak, rsync'd current production to development, will import from development	20:59
Acorn	and something about it being type unknown	20:59
crescendo	\mjd, and I should initialize the repo as normal?	21:00
\mjd	crescendo: initialize which repo?	21:00
cbreak	if you check out a remote tracking branch, be careful what happens	21:00
	you might get a new local tracking branch	21:00
	or a detached head	21:00
	depending on what git version you use	21:00
crescendo	\mjd, the git repo that will reside on the development server, but that only exists as an rsync'd duplicate of production at the moment	21:01
Acorn	what could cause reading the private key to fail?	21:01
	← bobrik left	21:01
jast	Acorn: what did you use to create the private key?	21:01
	→ bobrik joined	21:01
	→ Gitzilla joined	21:01
Acorn	puttygen	21:01
\mjd	crescendo: if you rsynced it from an existing repo, then that should include the .git directory, which is itself the repository. no need to initialize it.	21:01
jast	Acorn: putty generates a different key format than openssh uses	21:02
	→ n3kl joined	21:02
crescendo	\mjd, there is no repo at all right now. we are creating one for the first time. :)	21:02
jast	though I think that conversely, putty can use openssh keys	21:02
cbreak	crescendo: if you ever want to push to it, you have to git init --bare a new one	21:02
\mjd	crescendo: Oh, sorry. yes, initialize the repo as usual, and then put everything in it.	21:02
Acorn	but tortoisegit asks me if I want to use a puttykey	21:02
cbreak	if you don't, then having an rsynced clone is no problem	21:02
	← jtaby left	21:02
crescendo	cbreak, we're going to be "pull"ing. does that change things?	21:02
jast	well, I don't know what tortoisegit actually does in the background	21:02
cbreak	pull NEEDS a working dir	21:03
jast	perhaps it uses plink instead of openssh	21:03
n3kl	Can I rewrite the author history of a repo?	21:03
crescendo	cbreak, the rsync clone is only so I can import a clean set of files without things like "cache/", etc	21:03
cbreak	push NEEDS no working dir	21:03
Acorn	it has a "load putty key" checkbox	21:03
\mjd	n3kl: yes. git-filter-branch will do that.	21:03
	→ jtaby joined	21:03
jast	n3kl: sure. see if the examples in http://git.or.cz/man/fit-filter-branch won't help you. :)	21:03
n3kl	Nice, thanks	21:03
jast	whoops, link broken	21:04
	← ajpiano left	21:04
jast	I'm sure you can fix it yourself	21:04
n3kl	yup	21:04
	fit gilter branch	21:04
jast	exactly	21:05
	→ unreal_ joined	21:05
	Unode → Unode\|NA	21:05
	← unreal left	21:06
	unreal_ → unreal	21:06
crescendo	So, my process so far is; 1) copy production code to test server, 2) remove transient files from test server, 3) initialize git repo on test server, 4) import test server files into git, and to deploy to production, 5) perform "git pull ..." on production server, pointing at repo on test server. Am I correct?	21:07
	→ DrNick2 joined	21:08
n3kl	How can I see what arguments or env variables are available to the hook scripts?	21:09
	→ tobiassjosten_ joined	21:09
	← pflanze left	21:09
n3kl	I want to track ever ip pulling data	21:09
crescendo	...or should I host the initial "production" repository on the production server, and do "git push" on the test server (after branching and modifying, obviously)	21:09
	← corni left	21:09
	→ bcardarella joined	21:09
	← DrNick left	21:09
jast	n3kl: apart from what's documented, create a hook script that just outputs $@ and the output of set :)	21:09
	then you'll know	21:10
	→ karptonite joined	21:10
n3kl	ah, too easy	21:10
	here I am looking for documentation	21:10
jast	(the docs are in http://git.or.cz/man/githooks, by the way)	21:10
n3kl	looking at that one	21:11
	DrNick2 → DrNick	21:11
	← tobiassjosten left	21:11
	← gbacon left	21:12
	← stringoO left	21:15
	→ engla joined	21:15
	← engla left	21:15
	→ engla joined	21:15
	→ kristian42 joined	21:16
	→ SlickT10 joined	21:17
n3kl	thanks	21:20
akahn	Is there a way to know what commit introduced the changes that are conflicting with my rebase?	21:20
	so that I have more information than 'HEAD'?	21:20
thiago_home	no	21:22
	but you can find out by examining the file log and guess:	21:23
jast	git blame and look at the conflicting lines :)	21:23
thiago_home	git log -p ..MERGE_HEAD -- filename	21:23
	git log -p MERGE_HEAD.. -- filename	21:23
	← JackWinter left	21:23
	→ programble joined	21:23
akahn	thanks	21:24
	→ mathepic joined	21:26
akahn	Is it just me, or if I'm fixing a conflict in the middle of a rebase, my version of the file could be not totally up to date, because my later changes to the file haven't been merged in yet?	21:27
	← \mjd left	21:28
jast	akahn: that's right. rebase deals with your commits one by one.	21:28
akahn	ok. thanks	21:28
	for a second i thought "why is that method being used? i removed that!"	21:29
	← skoop left	21:29
	← smik left	21:29
thiago_home	akahn: if you're fixing in the middle of a rebase, then that's very correct	21:30
	your file is not up to date until the rebase is finished	21:30
	→ HG` joined	21:31
	→ smik joined	21:31
	← raichoo left	21:32
	← tstclair left	21:33
	→ theHakon joined	21:34
	adamholt → adamholt_away	21:34
	→ nevele joined	21:34
	adamholt_away → adamholt	21:34
akahn	right	21:35
nevele	anyone see this error before? https://gist.github.com/aacdde1eb53ff03447c1	21:37
	← _Vi left	21:37
	→ spaceonline joined	21:37
	← resmo left	21:37
nevele	or, more specific.. how can i revert a branch back to a specific commit?	21:37
	→ _Vi joined	21:38
nevele	seems i horked something in my remote repo, get that error everytime i try to push	21:38
	← locklace left	21:38
jast	nevele: local git fsck doesn't complain?	21:38
	→ airborn joined	21:39
	← mtkd left	21:39
	← _Vi left	21:39
nevele	jast: in the local repo, just do a "git fsck" ?	21:39
OFICINA7	#git how do I revert my changes to my last commit ?	21:39
jast	yep. complaints about dangling objects don't matter.	21:39
OFICINA7	#git how do I revert my changes to my last commit, I haven't commited yet?	21:39
jast	OFICINA7: not pushed yet, you mean?	21:39
wereHamster	OFICINA7: reset --hard	21:39
theHakon	i just installed git and it seems to be somewhat incomplete, init/commit/checkout and most stuff works, but if i try 'log' or 'diff' it either tells me 'broken pipe' or it gives no output at all	21:39
	→ _Vi joined	21:39
Ilari	nevele: 'git fsck --full' might be needed to get full check.	21:40
jast	theHakon: broken pager?	21:40
wereHamster	theHakon: what's your PAGER set to?	21:40
nevele	jast, how about a dangling blob?	21:40
jast	nevele: doesn't matter	21:40
nevele	jast: same error	21:40
OFICINA7	jast: I had my a, b and c files, then I commited	21:40
jast	nevele: you mean, git fsck outputs the same error message?	21:41
OFICINA7	now I created d, e, f ... z files, and changed the a, b and c	21:41
jast	OFICINA7: and you want to add that to your previous commit?	21:41
OFICINA7	but I didn't type git commit	21:41
nevele	no, same error as the jist when i try to push to remote	21:41
jast	nevele: yeah, fsck wasn't supposed to fixed anything, just making sure that the problem is completely on the remote end	21:41
nevele	tried a full fsck as well	21:41
jast	--full is the default these days	21:41
OFICINA7	jast, no, I'm sincerely sorry for changing the files and I want to make them go back to the way they were	21:41
theHakon	jast, wereHamster, no it says 'Broken pipe'	21:41
wereHamster	theHakon: what says broken pipe?	21:42
nevele	so does that mean i am basically nuking the remote?	21:42
	→ level09 joined	21:42
jast	theHakon: does this work? git --no-pager log	21:42
nevele	and pushing fresh from teh local	21:42
level09	is there an equivalent to --dry-run for git pull ?	21:42
	I want to try it without actually applying the mege	21:42
	j*merge	21:42
wereHamster	level09:pull = fetch+merge, so just do the fetch	21:42
jast	level09: git fetch and you're free to compare your branch and origin/<branch>	21:42
	e.g. gitk ...@{u}	21:42
	(literally)	21:42
	← bobrik left	21:43
theHakon	jast, git --no-pager log works fine thx :)	21:43
wereHamster	theHakon: echo $PAGER	21:43
level09	wereHamster: fetch is updating my remote branchs, is there a way to simulate the merge without applying it ?	21:43
	← ExtraSpice left	21:43
jast	theHakon: that indicates that your default pager is broken. consider changing your shell profile/rc file to set PAGER to something standard like /usr/bin/less	21:43
theHakon	wereHamster, it just outputs a blank line	21:43
nevele	so nothing can be done on the remote end?	21:43
wereHamster	level09: there's no point in 'simulating' the merge	21:43
OFICINA7	jast, I regret changing current files and creating new files, how do I bring my code back to the last time I commited?	21:43
	→ locklace joined	21:43
wereHamster	level09: just do the merge, if you like the result, keep it, if not, undo the merge	21:44
level09	jast: is there any better tool than git diff ? it doesn't seem so explanatory	21:44
	wereHamster: thanks, I was hoping for something similar to git push --dry-run	21:44
mathepic	level09: Does	21:44
jast	OFICINA7: so you want to obliterate the commit and uncommitted changes? I'm still not quite sure what you want :)	21:44
level09	it shows failures in case they exist	21:44
theHakon	jast, ok i did a 'export PAGER=less' is that the way to go?	21:44
mathepic	level09: Try "git difftool"	21:44
jast	theHakon: should probably work	21:44
	level09: I didn't suggest diff, actually. I suggested gitk which visualizes stuff for you.	21:45
theHakon	jast, could i also set git to use --no-pager by default?	21:45
wereHamster	theHakon: man git-config	21:45
Gitbot	theHakon: the git-config manpage can be found at http://git.or.cz/man/git-config	21:45
OFICINA7	jast: I commited and I was happy, then I created a lot of files, and now I am not happy anymore, I want to be happy again	21:45
level09	jast: gitk doesn't work on centos :P	21:45
	or does it ?	21:45
mathepic	jast: It shouldn't be broken because PAGER isn't set, though. I've ever set it and Git commands work fine without a pipe error.	21:45
	never set it*	21:45
jast	level09: tig is pretty cool, if you've got it installed	21:45
	basically a curses equivalent of gitk	21:45
level09	let me look that up :)	21:46
jast	OFICINA7: does that mean that the answer to my previous question is yes?	21:46
OFICINA7	jast: I haven't commited again because I am unhappy	21:46
jast	mathepic: perhaps there's a wonky system default there	21:46
OFICINA7	jast: I didn't understand the terms :P	21:47
theHakon	wereHamster, ok thx	21:47
	→ kukks joined	21:47
jast	OFICINA7: so basically you have the committed changes that you don't like, and some additional changes are not committed. out of that, what do you want to keep?	21:47
theHakon	jast, yeah i tried tig on an other box before and really liked it	21:47
OFICINA7	nothing, I just with things were back the way they were	21:48
	← chrowe left	21:48
jast	OFICINA7: all right, then just git reset --hard HEAD^ (= destroy uncommitted changes and go back to the commit before the most recent one)	21:48
	the things we mentioned before apply if you've already pushed the bad commit	21:48
OFICINA7	jast: this includes the files I created?	21:48
jast	if you added them, yes. if you didn't, they'll just stay there.	21:49
	→ brosner joined	21:49
	← brosner left	21:49
jast	of course you can always force git to destroy all files that aren't tracked by using git clean (which is a dangerous operation, of course)	21:49
OFICINA7	oh that helps	21:49
	!	21:49
	how do I git clean?	21:50
	-n or -f?	21:50
	I just want my unadded files to be gone	21:50
	→ resmo joined	21:50
	← airborn left	21:50
	← tobiassjosten_ left	21:50
	→ airborn joined	21:51
jast	reading a manual page once in a while can't hurt. it's actually a very short one, too. ;P	21:51
OFICINA7	jast: I'm a beginner =/	21:52
aep	and that means you cant read?	21:52
OFICINA7	I could do what I wanted, just commited the changes and deleted that branch	21:52
	aep: that means I'm dumb enough not to know what to look for	21:53
aep	man git-clean	21:53
	← radioman-lt left	21:53
Gitbot	aep: the git-clean manpage can be found at http://git.or.cz/man/git-clean	21:53
aep	git manpages are quite good	21:53
	→ mar77i joined	21:53
	← nevele left	21:54
jast	well	21:54
	they are definitely complete, but I have higher standards for "good"	21:54
aep	better then all the modern crap that only has zillion pages html crap with no grouping...	21:55
OFICINA7	aep: the word commit isn't even listed on that page	21:55
aep	indeed, it is not. git clean doesnt have anything to do with commits	21:55
FauxFaux	They're good apart from when they start using terminology that you kind of understand but don't. Like the reset --x documentation.	21:55
jast	that's because git clean doesn't deal with commits	21:55
OFICINA7	aep: see, I didn't know that	21:55
aep	ok, good point	21:56
OFICINA7	so, I have this cool.txt file	21:56
	witch has "cool" as it's content	21:56
jast	FauxFaux: yeah, that sort of thing, or "it also accepts options that are accepted by <some other command>" (which is also kind of a UI problem: some commands have way too many options)	21:56
OFICINA7	I DO COMMIT IT	21:56
	then I change it's content to "I HATE WHAT I DID"	21:56
bulbous	jast: after reading your posts to OFICINA7, I realize that I'm not sure how to revert changes that I have pushed to the server, I've used git reset to undo a local commit -- what about stuff that has been pushed? I looked around a bit online, found 'git revert' -- but when I try that I get a conflict -- not sure why...	21:56
FauxFaux	jast: Or anything that talks about refspecs. ARRRRRRGHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH ARGHA RGH RGHARGHARGH.	21:56
aep	yeah, the overview kindof lacks topical subsections	21:56
	← necromancer left	21:56
OFICINA7	so this is why I WILL NOT COMMIT	21:57
jast	bulbous: git revert applies the reverse of the change introduced by the original commit, so yes, conflicts can happen. you'll just have to fix them and commit the result.	21:57
OFICINA7	how can I undo my changes?	21:57
wereHamster	OFICINA7: git checkout -- cool.txt	21:57
aep	bulbous: unclean work dir?	21:57
OFICINA7	wereHamster: ok, if I did that to a thousand files?	21:58
jast	well, didn't we mention git reset --hard which undoes all uncommitted changes to tracked files?	21:58
bulbous	aep: nope, I did a clean	21:58
wereHamster	I did, a while ago	21:58
jast	dream team for complete destruction of everything: git reset --hard, then git clean	21:58
	← naderman left	21:58
OFICINA7	jast: so I should use both commands?	21:58
jast	you don't have to, but it sounds like that's what you want	21:58
OFICINA7	say I both created files and updated files	21:59
aep	bulbous: never seen a direct conflict aftzer a revert. odd. but even then, just fix them shrug	21:59
bulbous	jast: i could see conflicts -- but if i do git pull, change one file, git commit, git push, git revert -- seems like I shouldn't have a conflict	21:59
OFICINA7	jast: say I also deleted a lot of files, I regret everything I did	21:59
	IslandUsurper → IslandUsurperAFK	21:59
jast	bulbous: that makes sense indeed	21:59
wereHamster	OFICINA7: reset --hard; git clean -dfx	21:59
jast	OFICINA7: if they were tracked, reset --hard will restore them	21:59
aep	bulbous: does git revert itself whine about the conflict?	21:59
	→ rlb3 joined	22:00
bulbous	aep: yep -- here's the message:	22:00
	Automatic revert failed. After resolving the conflicts,	22:00
	mark the corrected paths with 'git add <paths>' or 'git rm <paths>'	22:00
	and commit the result.	22:00
OFICINA7	jast, wereHamster: wouldn't there be a way to commit the changes, and delete the latest commit?	22:00
aep	oO never seen that	22:00
	← _Vi left	22:00
wereHamster	OFICINA7: for what purpose?	22:00
aep	reverting a commit in the middle?	22:00
	ie not HEAD	22:00
OFICINA7	that would seem a little more patchy	22:01
	→ _Vi joined	22:01
jast	"patchy"?	22:01
wereHamster	'patchy'?	22:01
OFICINA7	packed	22:01
	patch-like	22:01
jast	'packed'?	22:01
wereHamster	'packed'?	22:01
bulbous	aep: i do git log, and then do git revert to the 2nd commit i see in the list	22:01
jast	'patch-like'?	22:01
aep	OFICINA7: do you actually want git-commit --ammend ?	22:01
wereHamster	'patch-like'?	22:01
	:)	22:01
OFICINA7	lol	22:01
	nobody understands me	22:01
aep	bulbous: yeah, that would conflict of course.	22:02
jast	bulbous: if the first and second commit changed the same parts of the same files, conflicts are to be expected	22:02
OFICINA7	in a nutshell..	22:02
wereHamster	I think you don't even know what you want..	22:02
OFICINA7	you guys use rails?	22:02
wereHamster	go on	22:02
jast	I made that mistake once ;P	22:02
bulbous	aep: jast: I was thinking that revert would revert to a specific revision, not revert the revision passed in -- I've just migrated from subversion land, so my thinking isn't quite right yet -- thanks much :)	22:02
OFICINA7	I generated a lot of files, changed a lot of files, and now everything is broken	22:03
	I don't even know where to start from	22:03
aep	bulbous: aah, no it doesn't	22:03
OFICINA7	what do I do?	22:03
	→ stringoO joined	22:03
wereHamster	OFICINA7: git reset --hard; git clean -dfx	22:03
OFICINA7	thx	22:03
	→ bkeating_ joined	22:03
aep	bulbous: you can reset the head to that specific revision. but that only works if no one pulled yet	22:04
	→ naderman joined	22:04
	← naderman left	22:04
	→ naderman joined	22:04
bulbous	aep: is there a different command for that?	22:05
bkeating_	I made a branch to develop on and would like to update it with the latest HEAD. How best to do this? This isn't a merge, right? I could just 'pull origin master' while in my dev branch?	22:05
	← hebz0rl left	22:05
aep	bulbous: yeah, git-reset. but what i guess you want is git-diff <otherref> \| patch and then commit	22:05
	← masterkorp left	22:06
jast	bkeating_: pull does actually use merge, too	22:06
	so in that case it wouldn't make a difference unless you have local changes in master	22:06
bulbous	aep: yeh ok thanks much :)	22:06
bkeating_	jast: never knew. thanks.	22:06
bulbous	think i got a handle on it now :)	22:06
jast	pull = fetch+merge	22:06
	← rlb3 left	22:06
aep	bulbous: git diff 123abc... \| patch -R && git commit -m "revert crap i did. i'm so sorry. lulz"	22:07
jast	we explain that formula a lot :)	22:07
bulbous	aep: nice I like it	22:07
OFICINA7	wereHamster it says refusing to clean	22:07
jast	bulbous: there's a somewhat obscure way to create a commit that completely reverts the tree to some previous state	22:08
wereHamster	is that all it sais?	22:08
bkeating_	ultimately, this new dev branch will be merged back into master, i just want to keep it updated so there is little drift between the two	22:08
OFICINA7	wereHamster I should pass -n or -f	22:08
jast	bulbous: git read-tree --reset <old commit>; git commit; git reset --hard	22:08
	← Rayne left	22:08
OFICINA7	fatal: clean.requireForce defaults to true and neither -n nor -f given; refusing to clean	22:08
wereHamster	OFICINA7: well.. didn't I write git clean -dfx ?	22:09
aep	jast: that sounds dangerous	22:09
	→ brosner joined	22:09
bulbous	jast: that would be the equivalent of aep's patch right?	22:09
aep	no	22:09
jast	aep: only if you have uncommitted changes	22:09
	which will obviously disappear	22:09
OFICINA7	wereHamster thanks	22:09
aep	jast: i'm more thinking about other peoples trees beeing broken	22:09
	← naderman left	22:09
jast	yeah, well, it's kind of a hard revert	22:09
	← eno left	22:10
jast	but it doesn't really break anything	22:10
	← sgh_ left	22:10
aep	so they will merge the conflict and bring back the reverted changes :P	22:10
jast	why?	22:10
	this is creating a new commit that has the same contents as an old one	22:10
bulbous	ahh ok -- so it would be the same if noone else had gotten a newer revision then?	22:10
	→ iruediger joined	22:10
aep	because their not pushed commits dont apply to the new head	22:10
jast	it's basically the same as git revert on all commits after the specified one, rolled into one commit	22:11
aep	bulbous: if no one has pulled yet, you can safely do what jast said	22:11
jast	even if someone has pulled, it's still safe	22:11
	they'll just be annoyed at the huge bunch of conflicts	22:11
	especially if you revert a huge chunk of the history	22:11
aep	well, you'll get conflicts anyway	22:11
	regardless how you revert	22:11
jast	at any rate, it's a revert and not a reset	22:11
	so it won't confuse git merge	22:11
	→ naderman joined	22:11
	← naderman left	22:11
	→ naderman joined	22:11
aep	ooh now i see what read-tree does	22:11
jast	hehe :)	22:12
	I said it was obscure	22:12
	← Acorn left	22:12
jast	but it's extremely useful if you want to revert "everything"	22:12
bkeating_	jast: can i bug you with that last question i posted?	22:12
aep	yeah sounds like a good idea. thanks for the tip :)	22:13
jast	bkeating_: the last thing I saw from you wasn't a question :)	22:13
	← iruediger left	22:13
	adamholt → adamholt_away	22:13
	adamholt_away → adamholt	22:13
jast	but if you've got another one, sure, just ask	22:14
	→ eno joined	22:14
	IslandUsurperAFK → IslandUsurper	22:15
	← RobertLaptop left	22:16
bulbous	aep, jast: thanks much -- I'm not 100% on what read-tree does, but I'm gonna read up a bit about it	22:17
jast	basically it reads the contents of a commit into the index	22:17
	it's a low-level thing that's part of checkout, too	22:17
	→ masterkorp joined	22:17
bkeating_	jast: whoops. well Im asking wether a 'merge' is what I want when the branch will ultimately end up being merged the other way (into master)	22:18
	← [Po]lentino left	22:18
bulbous	ahh ok -- so it'd be like pulling, then cloning to a different directory, then replacing the contents of the current directory with the new clone, then committing?	22:18
bkeating_	i know that'll give me the updates i want; but im not sure if that breaks VC in the long run	22:18
	← naderman left	22:18
jast	bkeating_: that's one thing you can use. if you're not sharing the branch until it gets merged into master, git rebase master will make for cleaner history but might cause more conflicts along the way. it's a matter of preference. many people do use rebase for private dev branches.	22:18
	→ adamm joined	22:19
jast	it doesn't break anything if you use merge, that's for sure	22:19
	merge is meant to be all-powerful :)	22:19
	← masterkorp left	22:19
	→ masterkorp joined	22:19
bkeating_	rebase, great -- have never used it but got a breadcrumb now to learn. ;)	22:19
	its not private... i own the entire project (im the only developer) i just want my vc logs to track correctly and cleanly.	22:19
jast	you don't want to use rebase on commits that you've already published somewhere	22:20
bkeating_	not this weird master merged into dev then dev merged into master, etc.	22:20
jast	unless you want to severely inconvenience everyone who's got a copy of those commits	22:20
	← IslandUsurper left	22:20
aep	bulbous: sort of. except the index is not the same as your working dir	22:21
bulbous	aep: ahh didn't realize that -- it's a directory in .git	22:22
	aep: ?	22:22
aep	a file	22:22
jast	the index is what we also call the staging area	22:22
aep	i have no clue what it actually contains	22:22
jast	it's where git keeps track of what to include in the next commit	22:22
bkeating_	how come gitready.com is not mentioned in the documentation section on the git site?	22:22
aep	but its the same thing where git-add puts stuff into	22:22
	→ naderman joined	22:23
	← naderman left	22:23
	→ naderman joined	22:23
jast	bkeating_: it might be referred in the git wiki. feel free to contribute a patch to the git site, though. :)	22:23
	aep: it's basically about the same thing as a tree, but with additional cached information to speed things up, such as inodes (I think)	22:23
bulbous	aep: that makes sense then, so read-tree sort of pre-stages it	22:23
	← LiamH left	22:24
bkeating_	jast: that may be more dangerous than anything at the moment and im learning simple things like rebase, still ;) but i'll keep it in mind.	22:24
	← akahn left	22:24
jast	bulbous: that's right	22:24
aep	bulbous: yeah. so the next commit will contain the exact tree that your commit had from the hash you did read-tree on	22:24
	pretty smart idea.	22:24
	it doesnt even destroy your workdir (although it wont apply anymore)	22:25
bulbous	jast, aep: ok think i finally got a handle on it. I am loving git, the learning curve is well worth it	22:25
	right	22:25
	→ daaku joined	22:25
jast	you can actually even do the same thing in one command, but it's even more low-level that way:	22:26
	wait, you can't, it's still two commands	22:26
aep	you have to commit :P	22:26
	← ChUrP1 left	22:26
	← tango_ left	22:26
aep	unless your command includes magic hackery to move commits around	22:27
	that would be scary	22:27
	→ tango_ joined	22:27
jast	git update-ref refs/heads/master $(git commit-tree <old commit>^{tree} -p $(git rev-parse HEAD))	22:28
	I guess you could argue that those are three commands	22:28
	→ saimon_ joined	22:28
jast	that's about the most low-level way you can do it	22:28
saimon_	ack...just instinctively ctrl-c'd a git commit -a and now (after a git reset --soft HEAD^) I can stage any files. There must be a lock file somewhere but I can't see any.	22:29
jast	.git/index.lock, probably	22:29
saimon_	can/can't	22:29
	jast: nope...checked...it's not there.	22:29
jast	what's the error message?	22:30
	← bcardarella left	22:30
	← sylr left	22:30
	← daaku left	22:30
	→ bcardarella joined	22:31
	← bcardarella left	22:31
saimon_	jast: there is none. I do a git add path/to/file it returns and doesn't complain but when I do a git status it's not staged (shown as modified but not staged)	22:31
jast	hmm. that's weird. you could remove the index and start over staging everything, I suppose...	22:31
	no guarantees that that'll fix it, though	22:31
	← smik left	22:31
saimon_	jast: If I remove the index, I don't lose my changes just the fact that they were staged right?	22:32
jast	saimon_: yep	22:33
saimon_	jast: I'll try that ..thx	22:33
jast	removing the index effectively means that everything will be staged as it was after the last commit	22:33
	← coderdad left	22:34
	← st0w left	22:34
saimon_	jast: argh...now every single file is marked as deleted :(	22:34
	→ coderdad joined	22:34
jast	git reset should fix that	22:34
	(no args!)	22:34
saimon_	jast: ok	22:35
	phew	22:35
	:)	22:35
	cool...I can stage again...	22:35
aep	jast: what does that ^{tree} do?	22:35
wereHamster	aep: man git-rev-parse	22:35
Gitbot	aep: the git-rev-parse manpage can be found at http://git.or.cz/man/git-rev-parse	22:35
jast	aep: in this case it's probably optional... it dereferences a ref to a tree object	22:35
aep	jast: commit-tree wants the object on stdin btw	22:35
	hm	22:36
jast	aep: it wants the commit message on stdin	22:36
aep	ah!	22:36
	aep missread that	22:36
saimon_	jast: next time I know it's better to never cancel a git command as there probably a way to undo it.	22:36
jast	saimon_: this shouldn't really happen most of the time. you probably just canceled at the very worst possible moment. :)	22:37
	but yes, a commit is easy to undo	22:37
saimon_	jast: yeah...what does a git reset no args do exactly?	22:37
jast	it returns the index to the state after the most recent commit	22:37
	and doesn't touch the working tree	22:37
saimon_	ok cool...nice to know...god I love git	22:38
	← coderdad left	22:38
jast	join the club :)	22:38
saimon_	jast: many thanks	22:38
jast	any time	22:38
	especially when I should be doing other things... like now	22:38
	← tizbac left	22:40
	← alester left	22:42
	→ ajpiano joined	22:43
	← irc2samus left	22:44
	← hugoxrosa left	22:50
	← TDJACR left	22:53
	→ Textmode joined	22:53
	→ TDJACR joined	22:56
	← level09 left	22:56
	← dpino left	22:57
	→ rlb3 joined	22:58
	← rlb3 left	22:58
	← Ins\|de left	22:58
	→ rlb3 joined	22:59
	→ freshtonic joined	22:59
	← cannonball left	22:59
	← SlickT10 left	23:00
	→ RobertLaptop joined	23:01
	→ mjf joined	23:03
	← MattDiPasquale left	23:04
	→ MattDiPasquale joined	23:05
	← d0k left	23:05
	← coppro left	23:06
	← pantsman left	23:06
	← MattDiPasquale left	23:08
	← myosui left	23:08
	← mbroeker left	23:10
	← smuf__ left	23:11
	→ scarabx joined	23:13
	→ jmcantrell joined	23:14
	→ brizly joined	23:14
	coyo → AwayYote	23:14
	← andreaa left	23:15
	← brizly1 left	23:16
MarcWeber	http://dpaste.com/210373/ I'm getting this error: emote: fatal: packfile ./objects/pack/pack-a5d97b75d74ad47018af158a985307c504304831.pack cannot be accessed	23:17
	Which is the way to recover? I tried deleting the remote branch and recreating it. The problem still exists. Using a different repo which already has some commits can fetch the branch easily.	23:18
	→ iruediger joined	23:19
	← airborn left	23:22
wereHamster	MarcWeber: permission probljems?	23:22
	← mjf left	23:22
MarcWeber	We always accessed that account using ssh only (same account). So I don't think so.	23:23
	→ MattDiPasquale joined	23:23
	dwmw2 → dwmw2_gone	23:23
MarcWeber	The file does not exist at all	23:24
	← Paraselene_ left	23:24
	→ EricInBNE joined	23:26
	← apsysoev left	23:27
	← shruggar left	23:29
	→ consolers joined	23:29
	← jmcantrell left	23:29
	→ jmcantrell joined	23:30
MarcWeber	Running git fsck shows the error "too many open files'	23:30
	← dj_tjerk left	23:31
	→ LiamH joined	23:32
	← adymitruk left	23:33
	→ t0rc joined	23:33
	← didlix left	23:33
	→ jmcantrell_ joined	23:35
	→ sebastorama joined	23:37
	← rlb3 left	23:37
	← aziz left	23:37
	← psoo left	23:38
	← jmcantrell_ left	23:38
	← jmcantrell left	23:38
	← adamv left	23:39
	→ lectrick joined	23:39
lectrick	"error: You have local changes to 'app/controllers/home_controller.rb'; cannot switch branches." How do I stash the current changes so I can commit them to a different branch?	23:39
MarcWeber	lectrick: git stash ?	23:39
lectrick	how do I use git stash? is it like a temporary branch?	23:40
	MarcWeber ^	23:40
	→ coppro joined	23:40
	→ tinto101 joined	23:40
	← rox` left	23:40
MarcWeber	lectrick: type it into the console or ask google. Or try man git-stash	23:40
Gitbot	lectrick: the git-stash manpage can be found at http://git.or.cz/man/git-stash	23:40
	→ fishey joined	23:41
	← locklace left	23:41
	← Tommy[D] left	23:41
	← juanjoc left	23:41
	→ locklace joined	23:42
	→ Tommy[D] joined	23:42
	→ bodom joined	23:42
	← Yuuhi left	23:42
cYmen	can I somehow get an overview of a repository with branches and tags and such?	23:46
FauxFaux	gitk --all; git branch -a	23:47
	← justin-george left	23:49
	← AAABeef left	23:50
offby1	cYmen: log --graph --pretty=oneline --abbrev-commit --decorate	23:50
cYmen	apparently I created several tags by experimenting with "git tag $command" :)	23:50
	→ justin-george joined	23:52
	← justin-george left	23:53
	AwayYote → coyo	23:53
	→ anathematic joined	23:53
MarcWeber	wereHamster ulimit -n problems :)	23:53
	← bkeating_ left	23:53
	← stepnem left	23:54
	← bodom left	23:56
	← kipras left	23:57
	← lectrick left	23:58
	← parasti left	23:59

Logs Search ←Prev date Next date→

Channels Documentation