IRCloggy #git 2011-02-12

Provider of IRC logs since 2005.

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

	2011-02-12
ayust	right, so what you probably really wanted to do was get rid of the local patch first, and then pull	00:00
pyther	so I would have wanted to revert a commit?	00:00
	← Pewpewarrows left	00:00
	← doener left	00:02
	← Mathman left	00:03
	→ sivy joined	00:03
	← benzado left	00:05
	← siyb left	00:05
	← dyer left	00:06
	← j4mie left	00:06
	→ Heimidal joined	00:06
sreich	what is the tool that shows up initially with git mergetool?	00:07
	after running git mergetool --tool=kdiff3, now I cannot get to whatever it had before, where I could select remote, local, or (a)bort, iirc	00:07
	← _psychic_ left	00:07
blaines	ayust: are there any other options other than git archive? Right now I'm going to use archive to copy the app to a location where tests and dependencies can be run, and then start the app from that dir vs. the repo dir	00:08
Eridius	sreich: do you have a config variable `merge.tool` set?	00:09
sreich	Eridius: ah, yes. does this happen automatically on first-run, or did I do this?	00:10
	→ theoros_ joined	00:11
Eridius	you must have set it. I've never heard of git mergetool setting that variable	00:11
	→ donCams joined	00:11
	← theoros left	00:11
donCams	hello	00:11
	← chrisshattuck left	00:11
	← eijk__ left	00:12
	→ eijk joined	00:12
	→ chrisshattuck joined	00:12
Eridius	sreich: looks like the default list is "[meld] opendiff kdiff3 tkdiff xxdiff [meld] tortoisemerge" (the positioning of meld depends on whether you're running Gnome)	00:12
	sreich: and then it adds vimdiff and merge, with order dependent upon whether your editor is vim	00:12
donCams	i'm using msysgit to connect to github	00:12
	and I always get permission denied <public key>	00:13
	I already added the public key to github	00:13
	am I missing something?	00:13
	← eletuchy left	00:14
	→ LouisJB joined	00:15
	← chrisshattuck left	00:15
	← jrmuizel_ left	00:15
	← oracal left	00:16
	← sivy left	00:17
	→ xiangfu joined	00:17
	← Caius left	00:19
sreich	Eridius: ah okay. thanks	00:20
	← pedrobelo left	00:20
ayust	blaines: why do you need another option?	00:21
	you can use archive for both...	00:21
	→ pedrobelo joined	00:21
	← jstemmer left	00:21
	← RaceCondition left	00:22
	→ Caius joined	00:22
blaines	ayust: I don't I'm just curious at this point	00:24
	→ siyb joined	00:26
	← Buglouse left	00:30
	→ JoeCoolDesk joined	00:37
	← jrdavid left	00:38
	→ jrdavid joined	00:39
	← jrdavid left	00:39
	→ jrdavid joined	00:39
	← jrdavid left	00:39
	→ jrdavid joined	00:40
donCams	argh Permission denied <public key>	00:43
	← normanrichards left	00:43
	← donCams left	00:44
	← nadavoid left	00:45
	→ cylence joined	00:45
	→ zeshoem joined	00:47
	← Dave^\| left	00:49
	theoros_ → theoros`	00:49
	theoros` → theoros	00:49
	← vgr left	00:51
	← kenichi left	00:51
subsonic-	how do I look at a file that's from a given tag?	00:52
jast	git show tag:path/to/file	00:52
subsonic-	jast: awesome	00:52
	← Heimidal left	00:53
	← eijk left	00:53
	← rayners left	00:56
	→ koo5 joined	00:56
	← Tuller left	00:57
JoeCoolDesk	What's the difference between a branch and a tag?	00:57
	→ eletuchy joined	01:00
SethRobertson	tag is fixed for all eternity, the head of a branch can move	01:01
	You can also annotate tags	01:01
	← kipras left	01:02
SethRobertson	branches have namespaces, tags do not	01:02
JoeCoolDesk	What do you use tags for?	01:02
SethRobertson	I use them to mark significant builds	01:03
	And release points	01:03
	→ sivy joined	01:04
subsonic-	yeah, me too	01:04
JoeCoolDesk	So I shouldn't tag if I don't mean it	01:04
	It's for marking release versions like 1.3.3.423292932039230932?	01:05
	← jfkw_ left	01:05
	← tjholowaychuk left	01:05
	← Teltariat left	01:05
subsonic-	JoeCoolDesk: it's up to you. But, the way I like to think about it is, "keep track of something that happened, but dont need to leave a branch oopen for it" I find it useful for "looking back" at old code keeping my created branches clean.	01:06
	← cylence left	01:06
SethRobertson	I guess. One problem with tags is there are no mulligans. If you find a bug after you pushed, you cannot easily change the tag location.	01:06
	← dreiss left	01:06
SethRobertson	Some people thus use branches to mark releases, but I find that problematic as well.	01:06
	→ jfkw joined	01:07
subsonic-	just like SethRobertson said, signifigent builds and perhaps release points.	01:07
	← pandeiro left	01:08
	→ dyer joined	01:09
	← dyer left	01:09
	→ dyer joined	01:09
JoeCoolDesk	How significant is significant?	01:09
jast	that's up to you	01:09
	if you don't want to push tags individually, you'll only want to tag stuff that is of interest to the general public	01:10
	or, let's say, your target audience	01:10
subsonic-	JoeCoolDesk: Well, recently, we dropped a whole member type and we want to clean up the code, so we tagged it "before-vendor-removal"	01:10
SethRobertson	My personal builds I do not tag. When I have builds which are built on the release build server and announced to the dev team, those get tagged (automatically)	01:10
JoeCoolDesk	Git pretty much runs my website	01:10
	→ jmorris joined	01:11
JoeCoolDesk	I develop on the beta branch and then I can push it to the live branch	01:11
	So I can tag every commit I push? What's the advantage of that?	01:11
	← jfkw left	01:12
	← aah left	01:12
subsonic-	JoeCoolDesk: it's up to you man. I wouldn't.	01:12
	← jmorris left	01:13
PerlJam	JoeCoolDesk: I think your second question answers yoru first question :)	01:13
subsonic-	Just think it as a way to "bookmark"	01:13
jast	just because you can do something doesn't mean it's a good idea :}	01:13
	→ Pewpewarrows joined	01:14
JoeCoolDesk	Yeah that's what I'm thinking	01:14
jast	git can be used as an obstinate deployment tool, but that doesn't necessarily make it a good idea :)	01:14
JoeCoolDesk	90% of the things people say are so great really are more work than worth it I think	01:15
jast	similarly, you can pierce your left eye with a rusty nail...	01:15
JoeCoolDesk	Not that everyone's abuzz about tags or anything	01:15
	→ drl joined	01:15
	← ki__ left	01:15
jast	most people use tags for marking releases period.	01:15
PerlJam	JoeCoolDesk: no, I think that what people say are so great is great for them and their situation isn't universal	01:15
	← mattly left	01:15
jast	if you don't have releases, you probably don't really need tags	01:16
	→ zorganae joined	01:16
	→ Heimidal joined	01:16
	← blaines left	01:16
	← kakos left	01:16
	ChanServ set mode: +v	01:16
zorganae	hi	01:17
	I'm trying to walk through the history of a branch with diff-tree trying to find a "zero diff" commit	01:17
	how do I check if a given commit exists?	01:17
	that is...	01:18
	what I am doing is "git diff-tree branch1^ branch2"	01:18
	how do I make sure "branch1^" exists?	01:18
JoeCoolDesk	I'm still trying to figure out if my history is repeated or if Giggle is simple displaying it oddly.	01:18
	← priidu left	01:18
	→ jvraoltc joined	01:20
	→ ThunderChicken joined	01:22
	← tewecske left	01:23
	→ mikepack joined	01:23
ayust	zorganae: git rev-parse --verify	01:27
	?	01:27
	← Mannequin left	01:28
	→ Mannequin joined	01:28
zorganae	ayust: it fails	01:29
	I guess I have to look at the exit value of git	01:29
	← ludde left	01:30
ayust	git rev-parse --quiet --verify branch^1	01:30
	t	01:30
	then look at the exit code	01:30
	nonzero = doesn't exist	01:30
zorganae	seems like a good idea	01:30
	let me try it	01:30
	:)	01:30
	← eletuchy left	01:31
	← pdelgallego left	01:31
	← leavittx left	01:31
	→ eletuchy joined	01:31
	← justin-george left	01:31
	← icwiener left	01:34
	→ normanrichards joined	01:35
	← radost_ left	01:36
	→ wliao joined	01:36
	← towski left	01:36
	← vitor-br left	01:37
	← utu left	01:38
	→ marienz_ joined	01:38
	← Yuuhi left	01:39
	→ explodes joined	01:40
explodes	Hello, new to git. Absolutely loving it. Way better than mercurial.	01:40
	Much more logical/easy to use.	01:41
	One question though, how do I modify a commit message of a commit that has not yet been pushed?	01:41
zorganae	ayust: I'm getting some other colateral issues... so I'm not able to finish this test now. But your idea seems good. Thanks :)	01:41
explodes	rebase could do it I guess...	01:41
PerlJam	explodes: if it's the last commit, you could do "git commit --amend"	01:41
	→ Mannequin1 joined	01:42
explodes	word	01:42
	← Mannequin left	01:42
explodes	What if its not?	01:42
dol-sen	then rebase it is	01:42
ayust	git rebase -i	01:42
	← smolyn left	01:42
ayust	or more specifically	01:42
	g	01:42
	→ t0rc joined	01:43
ayust	git rebase -i <sha of commit before the one you want to edit>	01:43
explodes	Right right, ok	01:43
ayust	and then follow the key in the comments	01:43
	→ eddic joined	01:43
explodes	Excellent. Its time to go, but thanks a lot.	01:43
	← explodes left	01:43
eddic	Is there a way to take a specific commit from on branch and apply it to another branch?	01:44
	→ JohnFord joined	01:44
	← JohnFord left	01:44
	← Eridius left	01:44
	← eletuchy left	01:44
	← d0k left	01:45
	→ eletuchy joined	01:45
zorganae	git cherry-pick	01:45
	(eddic: this was for you ;))	01:45
	← marienz left	01:46
eddic	I have read that this could break future merging?	01:46
	is this true?	01:46
	→ tangledhelix joined	01:46
	→ JohnFord joined	01:46
	← JohnFord left	01:46
	← tangledhelix left	01:46
	← Mannequin1 left	01:46
zorganae	I'll leave that for a more experienced person here...	01:46
	← jvraoltc left	01:47
SethRobertson	git cherry-pick will not break future merging	01:47
	← verto left	01:47
eddic	oh	01:48
	ok	01:48
	nevermind	01:48
	yes, that worked perfectly	01:48
	thanks:)	01:48
	← boombatower left	01:48
	→ eletuchy_ joined	01:49
	→ _Vi joined	01:49
	→ jhaddad joined	01:51
	← Schmallon left	01:51
	← eletuchy left	01:51
	← eddic left	01:52
jhaddad	i created a new repo, i've got about 30 commits, but they're all using the wrong email. how would I change the email in every commit in my history? I've never pushed to a remote.	01:52
ayust	jhaddad: man git-filter-branch	01:52
jast	jhaddad: the 'git-filter-branch' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-filter-branch.html [automatic message]	01:52
jhaddad	thanks, i'll look into it.	01:52
subsonic-	good job jhaddad	01:54
	→ vitor-br joined	01:54
jhaddad	subsonic-: thank you fine sir	01:54
subsonic-	haha	01:55
	marienz_ → marienz	01:55
	→ jfelchner joined	01:57
	→ brendon_ joined	01:58
	← tvw left	01:59
jhaddad	looks intense.	01:59
	← noahcampbell left	02:00
	→ justin-george joined	02:00
ayust	it's a powerful tool	02:01
	→ Behi joined	02:01
	← hvr left	02:02
	← Chillance left	02:03
jhaddad	seems like if I did it wrong without a backup i could be hosed	02:03
ayust	well	02:04
	until you garbage collect	02:04
	you can always just 'git reset --hard <original SHA>'	02:04
	so it's not that dangerous	02:04
	but having a backup never hurts	02:04
	and is a generally good idea	02:04
	→ silanus_ joined	02:05
jhaddad	attempting it now..	02:05
	← eletuchy_ left	02:05
	← silanus left	02:06
jhaddad	easier than I thought... git filter-branch --env-filter 'export GIT_AUTHOR_EMAIL="myemail@whatever.com"'	02:06
	→ glennhartmann joined	02:07
jhaddad	s	02:08
	→ blaines joined	02:09
	→ Schmallon joined	02:15
	← zorganae left	02:16
	← Araxia left	02:18
	→ smolyn joined	02:18
	← brendon_ left	02:19
	→ jrmuizel_ joined	02:19
	← jrmuizel_ left	02:20
	← jaeckel left	02:21
	→ jaeckel joined	02:22
	← jaeckel left	02:22
	→ jaeckel joined	02:22
	← nfiniteset left	02:22
	← dyer left	02:22
	← subsonic- left	02:23
	→ dyer joined	02:24
	← dyer left	02:24
	→ dyer joined	02:24
	← airborn left	02:25
	← _Vi left	02:27
	→ _psychic_ joined	02:29
	← Heimidal left	02:29
shruggar	sitaram, ping	02:29
	→ Heimidal joined	02:29
	→ Buglouse joined	02:29
	← Heimidal left	02:29
	→ dreiss joined	02:30
	→ ben_alman joined	02:30
	→ Yuffster joined	02:31
sitaram	shruggar: pong	02:35
	→ _Vi joined	02:35
shruggar	sitaram, I'm trying to follow the procedure described in "shell games" to manage hooks via the gitolite-admin repos. I'm having trouble with the "gl-install" part due to it (as might be expected) attempting to overwrite what I just put into the admin-dir with defaults	02:37
	← alvaro_o_ left	02:37
	← mw_ left	02:37
	← _psychic_ left	02:37
shruggar	the difference between what I'm doing and what is described is that I'm also trying to manage gitolite-admin hooks	02:37
	specifically that's causing problems due to this part: system("cp -R -p $ENV{GL_BINDIR}/../src $ENV{GL_BINDIR}/../doc $ENV{GL_BINDIR}/../hooks $GL_ADMINDIR");	02:38
	→ jmorris joined	02:39
	← mm_work left	02:39
sitaram	shruggar: two things to note to start with: (1) dont fiddle with gitolite-admin's post-update hook (2) dont fiddle with any repo's update hook. In short, don't fiddle with the hooks that gitolite comes with	02:39
	shruggar: there's also a line in there that will cause a "these two are the same file" problem; I haven't had a chance to look at it yet but its harmless	02:40
	← hyperair left	02:40
	→ rayners joined	02:40
	← amerine left	02:41
	← saysjonathan left	02:41
shruggar	sitaram, I admit I have fiddled somewhat: I added gl-post-update-pre-apply to add additional checks prior to the checkout/etc	02:42
sitaram	shruggar: what is the error you're getting. You said what line is causing it but not what error. IIRC that is the line that will complain about "these files are the same" or some such	02:43
	→ Vampire0_ joined	02:43
shruggar	sitaram, "remote: cp: cannot stat `/home/vcs/bin/../src': No such file or directory", as my $GL_BINDIR doesn't live anywhere near anything else	02:44
	I suppose this could all be solved by my just sticking the hooks in common/, as that will copy them to gitolite-admin too	02:45
	→ dl9pf_ joined	02:46
sitaram	shruggar: what install method did you initially use?	02:46
	and yes that would be the right way to do things	02:46
	← Vampire0 left	02:46
	← dl9pf left	02:47
	← _Vi left	02:47
	← coderdad left	02:47
sitaram	shruggar: it is possible that shell-games was not really tested for other install methods than "from-client", so the last line (gl-install) should probably be different. I'd have to test, but I probably dont have time right now	02:47
	← hobodave left	02:47
	← ormaaj left	02:47
	→ Vampire0 joined	02:48
	← Vampire0_ left	02:48
	→ tkil joined	02:49
	→ angelsl joined	02:50
shruggar	I initially used system-wide install, though I've screwed things up by now, probably. Originally that seemed to be overwriting my hooks because there's a line that says.... (looking...) ln_sf("$GL_PACKAGE_HOOKS/gitolite-admin", $hook, "gitolite-admin.git/hooks") if $GL_PACKAGE_HOOKS;	02:50
sitaram	shruggar: also read the doc on hook propagation	02:51
	shruggar: I need to go way for about 20-30 min; sorry	02:51
shruggar	np. I think the gist of it is "I shouldn't screw with so many things" :)	02:51
	← Elfe_ left	02:52
shruggar	so I tried setting GL_PACKAGE_HOOKS='' in my rc, since it seemed that was closer to what I actually wanted to be doing.... that's when gl-install started being noisy instead of silently overwriting my hooks	02:52
	← ch077179 left	02:52
	→ brendon_ joined	02:52
shruggar	I think what I might do is write a separate gl-propagate-hooks script, that only does the one thing I want	02:53
	→ ProperNoun joined	02:53
	← ProperNoun left	02:53
	→ ProperNoun joined	02:53
shruggar	anyway, for now I'm going to bed. Thanks for putting up with my ramblings for a while :)	02:54
	← nadim left	02:57
	→ zombor joined	02:58
	← shruggar left	02:59
	← tildeequals left	02:59
	→ shiba_yu36 joined	03:01
	← drl left	03:01
	→ _Vi joined	03:03
	→ ch077179 joined	03:06
	→ hyperair joined	03:06
	→ boombatower joined	03:07
	← boombatower left	03:07
	→ boombatower joined	03:07
	← jhaddad left	03:09
	→ aah joined	03:10
	← yogione left	03:11
	← hyperair left	03:11
	→ Elfe joined	03:12
	← _Vi left	03:13
	→ leo2007 joined	03:14
	← scelis left	03:15
	→ drl joined	03:15
	→ kakos joined	03:17
	← ProperNoun left	03:18
	→ MikhX_ joined	03:18
	← MikhX left	03:19
	→ MikhX joined	03:19
	→ _Vi joined	03:20
	← jmorris left	03:20
	→ khaije\|minder joined	03:21
	→ Araxia joined	03:22
	← MikhX_ left	03:22
	← sivy left	03:28
	→ Mannequin joined	03:28
	← torvalamo left	03:28
	→ intripoon_ joined	03:30
	← leo2007 left	03:30
	→ jmorris joined	03:33
	← intripoon left	03:33
	→ gusnan joined	03:40
	← Amorphous left	03:40
	← mdel left	03:40
	→ lazyguru joined	03:44
	← normanrichards left	03:45
	→ coderdad joined	03:45
	→ Maletor joined	03:46
	→ mdel joined	03:52
	→ timj__ joined	03:52
	→ tcv_br joined	03:53
	← zorzar left	03:54
	→ Amorphous joined	03:55
	← glennhartmann left	03:55
	← Elfe left	03:55
	→ Elfe_ joined	03:55
	← tcv_br left	03:55
	→ ormaaj joined	03:56
	→ tcv_br joined	03:56
	← timj_ left	03:56
	← tcv_br left	03:57
	→ tcv_br joined	03:57
	← drl left	04:00
	← markelikalderon left	04:00
	← dreiss left	04:00
	← Textmode left	04:05
	→ Teltariat joined	04:08
	→ zorzar joined	04:08
	← kyv left	04:09
	← nb left	04:10
	→ stliu joined	04:12
	← stliu left	04:12
	→ stliu joined	04:12
	→ drl joined	04:13
	→ nb joined	04:15
	← Maletor left	04:17
context	kk so apparently git does have its limited	04:17
	limits*	04:17
	← brendon_ left	04:17
context	their is such a thing as too big a file	04:18
	→ LongBeach joined	04:18
context	4G of memory and 2G of page space and it cant gc ?	04:20
	the file is only 900M	04:20
	→ brendon_ joined	04:20
	→ TomyLobo joined	04:20
	← t0rc left	04:22
	← kukks left	04:22
	→ amerine joined	04:23
	← ThunderChicken left	04:26
	← Anubisss_ left	04:27
	← zeshoem left	04:30
	← gusnan left	04:33
	← tcv_br left	04:34
	← shiba_yu36 left	04:35
	← mdel left	04:37
	→ hyperair joined	04:39
	← Evious left	04:39
	← boombatower left	04:41
	→ kisielk_home joined	04:42
	→ Vik joined	04:43
	← _Vi left	04:47
	→ dr0id joined	04:48
	→ bradleyayers joined	04:49
	← zombor left	04:51
	→ int3h joined	04:53
	→ Zhonghua joined	04:55
	→ dreiss joined	04:56
	← coderdad left	05:00
	← Vik left	05:00
	→ harshu joined	05:02
	← pyther left	05:03
	→ yogione joined	05:07
	→ shiba_yu36 joined	05:07
	→ BiggFREE joined	05:08
	← TanEk left	05:11
	→ chrisshattuck joined	05:12
	→ normanrichards joined	05:14
	← jmorris left	05:18
	→ zombor joined	05:19
	→ blaines_ joined	05:19
	← blaines left	05:20
	→ arunpattnaik joined	05:28
	← zombor left	05:29
	← yogione left	05:31
	→ utu joined	05:32
	← blaines_ left	05:33
	→ towski joined	05:40
	→ jmorris joined	05:43
	← hyperair left	05:44
	← linusoleander left	05:46
	← bmahe left	05:48
	← harshu left	05:49
	← sreich left	05:52
	← chrisshattuck left	05:53
	→ maleknet joined	05:54
	→ mdel joined	05:55
	→ harshu_ joined	05:56
tango_	gasparri: "I temi posti sono tutti rilevanti: squilibri ed abusi che si registrano soprattutto nell'informazione radio-televisiva pubblica non possono essere elusi". chissà perché sono sicuro che non stia parlando del TG1	05:58
	(o di vespa)	05:58
	← rayners left	05:58
	← Mannequin left	05:59
	→ b4tm4n joined	05:59
	← dreiss left	05:59
	→ botanicus joined	05:59
dr0id	tango_: do you use github ?	06:00
	can the public activity be deleted ?	06:00
tango_	oops	06:01
	wrong channel guys	06:01
	dr0id: never tried so I have no idea sorry.	06:01
	btw, I think there is a #github channel, they might know more	06:02
	← Zhonghua left	06:02
	← arohner left	06:02
	← lazyguru left	06:02
	→ amondracorp joined	06:03
dr0id	tango_: they're dead today	06:03
tango_	poor things	06:03
	→ hyperair joined	06:07
	← arunpattnaik left	06:08
	← madewokherd left	06:11
	→ samhassell joined	06:11
	→ _Vi joined	06:12
	→ jvraoltc joined	06:12
	← hyperair left	06:14
	← amerine left	06:17
	→ hyperair joined	06:22
	→ struberg joined	06:22
	← b4tm4n left	06:30
	→ Heimidal joined	06:30
	← theoros left	06:34
	← project2501a left	06:36
	← int3h left	06:37
	← hyperair left	06:37
	→ hyperair joined	06:38
	← Buglouse left	06:38
	→ int3h joined	06:39
	← MikhX left	06:42
	→ iAsterisk joined	06:45
	← hyperair left	06:46
	← kisielk_home left	06:46
	→ MikhX joined	06:47
	ChanServ set mode: +v	06:48
	cxreg2 → cxreg	06:48
cxreg	what a pain. network wouldnt let me change nicks because i was "banned" in this channel	06:49
The_Tick	it's so you can't evad	06:49
	evade*	06:49
	for poor bans	06:49
cxreg	heh	06:49
	← vitor-br left	06:51
	→ Swimming_bird joined	06:52
	→ hyperair joined	06:53
	→ zeshoem joined	06:54
	→ pk___ joined	06:56
	← drl left	06:56
	→ iAsteris_ joined	06:57
pk___	i did a git reset --hard origin on my repository ..even then git status shows many modified files	06:57
	→ ExtraSpice joined	06:57
pk___	how do i reset the git repository completely	06:57
	→ Swimmin__ joined	06:58
	← Swimming_bird left	06:58
wereHamster	pk___: do you have crlf enabled?	06:58
pk___	dont know i am linux	06:59
wereHamster	which filesystem?	06:59
	← botanicus left	06:59
pk___	ntfs	06:59
wereHamster	which repository?	07:00
	← maleknet left	07:00
pk___	means?	07:00
	← hyperair left	07:00
wereHamster	is it the linux kernel?	07:00
	→ eletuchy joined	07:00
pk___	no no	07:00
	vlc repository	07:00
	← iAsterisk left	07:00
wereHamster	can you clone the repo to a linux filesystem and sii if it also shows modified files?	07:01
	← eft_ left	07:01
pk___	actually i have a slow internet connection	07:02
	so i cant really clone	07:02
	thats why i wanted to reset the existing repo	07:02
	← BiggFREE left	07:03
	← towski left	07:04
	← pk___ left	07:04
	← hagabaka left	07:05
	→ hagabaka joined	07:05
	← hagabaka left	07:05
	→ hagabaka joined	07:05
	→ Maletor joined	07:06
	→ pk___ joined	07:06
	→ DiGitalX joined	07:06
pk___	i cloned the repository	07:06
	and i issued a git status command	07:06
	and it shows all of them modified	07:07
wereHamster	pk___: my guess it's because you are on a case insensitive filesystem	07:07
	→ blaines joined	07:07
pk___	i am using linux and the drive is ntfs	07:08
wereHamster	yes, I know. You already told us	07:08
pk___	then what can i do now?	07:09
wereHamster	use something else than ntfs.. ?!?	07:09
	→ amerine joined	07:09
	← VladDrac left	07:09
pk___	but i have aother repository which is perfect	07:10
	→ VladDrac joined	07:10
pk___	on ntfs too	07:10
	→ drl joined	07:10
wereHamster	also vlc?	07:10
pk___	yes	07:11
	← dr0id left	07:11
pk___	oh sorry	07:11
	the other one is in ext4	07:11
	thanx :)	07:11
	→ nadim joined	07:13
	→ envi joined	07:15
	← struberg left	07:16
	→ iAsterisk joined	07:17
	→ project2501a joined	07:17
	← pratz left	07:18
	← iAsteris_ left	07:19
	← project2501a left	07:19
	← harshu_ left	07:20
	→ pratz joined	07:21
	← pk___ left	07:21
	→ iAsteris_ joined	07:22
	← MikhX left	07:23
	→ iAsteri__ joined	07:23
	→ harshu_ joined	07:23
	→ cesario joined	07:23
	← iAsterisk left	07:24
	← iAsteris_ left	07:26
JoeCoolDesk	How do I unadd files?	07:26
	← iAsteri__ left	07:27
	→ iasterisk joined	07:27
wereHamster	JoeCoolDesk: git reset <file>	07:27
	JoeCoolDesk: and git status	07:28
	→ Buglouse joined	07:29
JoeCoolDesk	Thanks	07:31
	← Maletor left	07:32
	← agile left	07:35
	← eletuchy left	07:41
	→ ki__ joined	07:43
CrazyGir	hello	07:45
	→ dr_win joined	07:46
	← jvraoltc left	07:46
CrazyGir	I'm not sure how, but I unexpectly got myself into a conflicted merge state	07:46
	the merge failed, so I'm not sure where to go from here	07:46
	how does one manually merge, is this the best way to resolve?	07:47
JoeCoolDesk	Is the source of the submodule saved in the parent repository?	07:52
	← Schmallon left	07:53
	→ thiago_home joined	07:54
	→ hyperair joined	07:55
	← harshu_ left	07:55
	→ harshu joined	07:56
	← dotCOMmie left	07:57
wereHamster	CrazyGir: man git-merge	07:57
jast	CrazyGir: the 'git-merge' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-merge.html [automatic message]	07:57
	→ dotCOMmie joined	07:58
	→ notze joined	07:58
wereHamster	there's a whole section about how to resolve merge conflicts	07:58
	ChanServ set mode: +v	07:59
notze	moinmoin	07:59
wereHamster	JoeCoolDesk: what do you mean by source?	07:59
	→ topi joined	07:59
	← topi left	07:59
	→ topi joined	07:59
notze	i have a problem pulling the latest git release i get the famous	08:00
	fatal: cannot pread pack file: No such file or directory	08:00
	what can i do here?	08:00
	i tried like git fsck --full git gc	08:00
	...	08:00
	how can i get further info what causes the problem?	08:00
	→ pfrenssen joined	08:01
	← blaines left	08:01
notze	lol moin pieter	08:01
	← Gryllida left	08:01
	→ randomus1r joined	08:03
	← randomus1r left	08:03
	→ XABQKXV joined	08:04
	→ tewecske joined	08:08
	← notze left	08:09
CrazyGir	ok, so I was not aware of git's ability to recognize different versions of files while working out the merge, and I've resulted in 4 local versions of the same files	08:10
	*file	08:10
	→ Gryllida joined	08:11
JoeCoolDesk	Werehamster, I mean the contents of the submodule, whatever it be.	08:14
	→ sudonotplease joined	08:14
	→ guardian joined	08:15
	→ moonk joined	08:17
	→ MikhX joined	08:17
	← moonk left	08:17
	← Heimidal left	08:23
	← amondracorp left	08:24
	← amerine left	08:24
	← Yuffster left	08:25
	→ joel_falcou joined	08:27
	→ eciur joined	08:27
	← xiangfu left	08:28
	← eciur left	08:28
	← guardian left	08:30
	← iasterisk left	08:32
	← zeshoem left	08:34
	→ doener joined	08:36
	→ lahwran joined	08:36
lahwran	how do I tell git that I deleted a file?	08:36
	← Silowyi left	08:36
lahwran	well, moved*	08:36
	← mikepack left	08:36
lahwran	I moved some files, did git add . at the root of the repo, and commited, but it doesn't seem to understand that the old files are gone	08:37
	→ xiangfu joined	08:37
lahwran	ah, found it partway down man git-rm	08:38
jast	ah: the 'git-rm' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-rm.html [automatic message]	08:38
mmattice	you didn't git mv them	08:38
	← dermoth_ left	08:38
doener	lahwran: "add <path>" will only add modified and new files, "add -u <path>" will add modified and deleted files, "add -A <path>" does both	08:38
lahwran	ah	08:39
	so add -A at the root of the repo and another commit will fix it?	08:39
	→ dermoth_ joined	08:39
	→ nevyn^ joined	08:39
selckin	git both the remove and add of the new file, and it should detect it's the same file and show as a move	08:39
	*stage	08:40
doener	lahwran: for example. But you'll likely want to amend the existing commit (unless you already pushed it)	08:40
lahwran	yeah, I didn't realize I had screwed up until I saw the error on github	08:40
	→ notze joined	08:41
	→ Utkarsh joined	08:41
	← Utkarsh_ left	08:44
	← nevyn^ left	08:45
	← hyperair left	08:50
	→ SIGBUS joined	08:54
	← utu left	08:56
	← topi left	08:57
	← joel_falcou left	08:59
	gonemad3 → workmad3	08:59
	→ protorom joined	09:03
	← urkud_ left	09:05
	→ srji joined	09:05
	→ _pingu joined	09:08
	← ormaaj left	09:11
	← babbage left	09:11
	→ priidu joined	09:12
	→ ormaaj joined	09:12
	→ anderse joined	09:13
	← sudonotplease left	09:16
	→ thermal joined	09:16
	→ sudonotplease joined	09:17
	→ maleknet joined	09:17
	→ guardian joined	09:17
	→ dozepih joined	09:18
	← thermal left	09:18
	← jmorris left	09:20
	→ deurk joined	09:21
	ChanServ set mode: +v	09:21
deurk	hello	09:21
	→ babbage joined	09:21
deurk	I use git to keep track of my dot files	09:21
	← thiago_home left	09:22
	aep just learned that this is a bad idea when doing git clean in thw wrong dir	09:22
deurk	and kinda made a mistake by commiting and pushing a few files in there that should not be part of the repository	09:22
	trouble is that the files are a few gigabytes...	09:22
	is there a way to remove them from the repository?	09:22
	→ Zephyrus joined	09:24
deurk	I did a git reset of the said commit	09:24
	but the bare repository (remote) still shows a size of a few gigabytes instead of few megabytes	09:25
	any way around it?	09:25
	→ leavittx joined	09:26
	← Behi left	09:29
	← dwmw2_AVF left	09:29
	← harshu left	09:29
	← pratz left	09:29
	→ rchavik joined	09:31
	→ radarek joined	09:32
	← rchavik left	09:32
	→ thiago_home joined	09:32
	→ j416 joined	09:33
protorom	there is a way... but I can't find it just now :-( it was something like: git filter-branch followed by git gc	09:33
	→ rchavik joined	09:33
	← maleknet left	09:34
	← rchavik left	09:34
	← thiago_home left	09:35
	→ boombatower joined	09:35
	→ rchavik joined	09:35
doener	deurk: see the end of man git-filter-branch, "CHECKLIST FOR SHRINKING A REPOSITORY"	09:36
jast	deurk: the 'git-filter-branch' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-filter-branch.html [automatic message]	09:36
protorom	@duerk have a look at git filter-tree's examples section	09:36
	oops ;-) filter-branch of course	09:36
	→ thiago_home joined	09:36
	→ Fullma joined	09:37
deurk	thx doener jast protorom	09:37
	:)	09:37
	→ ThomasLocke joined	09:40
	← dol-sen left	09:41
	→ harshu joined	09:43
	← JoeCoolDesk left	09:43
	→ pratz joined	09:43
	← DrNick left	09:43
	→ kumbayo joined	09:44
	→ DrNick joined	09:44
deurk	thx again, doing on git reset HEAD~1 followed by a git clone --bare file:///path did the trick	09:46
	← _pingu left	09:46
deurk	I was not using file:// previously so it hardlinked the repos thus kept the unneeded refs	09:47
	← deurk left	09:47
	→ khmarbaise joined	09:48
	→ chrischambers joined	09:49
	→ joel_falcou joined	09:50
	→ caruso_g joined	09:50
caruso_g	Hi all. I committed some patches I wanted to include into the next commit, I also made some edits before to realize I made the error in committing all. How do I revert the status to before the commit -a while keeping my edits?	09:52
	→ Chillance joined	09:52
	← robotarmy left	09:53
protorom	caruso_g: git reset HEAD~1	09:53
caruso_g	protorom: thanks mate. What is ~1? And (while looking at manual) which is the difference with HEAD^?	09:55
	→ Samual_ joined	09:55
	← ThomasLocke left	09:55
protorom	caurso_g: ~1 and ^ are the same - ~2 is the same as ^^ and ~3 is the same as ^^^... etc	09:55
	← Samual left	09:57
caruso_g	protorom: great, thanks a lot. But if I reset HEAD^^^ does Git keep my edits or I will loose ^^ and ^?	09:57
protorom	yes - the commits are still there - just hard to get to unless you created a new branch or tagged them first	09:58
	→ jogla joined	10:00
	→ RaceCondition joined	10:00
	→ urkud_ joined	10:00
protorom	you could also do something like git checkout -b fixme HEAD^ - then you incorrect commit is still available in your old branch, and you're back to where you were before your commit - then you can merge selectively from your broken commit	10:01
	(that was a really awkward sentence)	10:02
Pieter	notze: ?	10:02
	→ pdelgallego joined	10:03
	→ reZo joined	10:06
	← ormaaj left	10:07
	→ pantsman joined	10:08
	← pantsman left	10:08
	→ pantsman joined	10:08
	→ ormaaj joined	10:08
	→ HG` joined	10:10
	← HG` left	10:11
	→ HG` joined	10:11
	← pantsman left	10:12
	→ OOPMan joined	10:13
	← ki__ left	10:14
	← leavittx left	10:15
	← sudonotplease left	10:15
	→ sudonotplease joined	10:16
	← workmad3 left	10:20
	← notze left	10:21
	← joel_falcou left	10:22
	→ notze joined	10:23
	← SIGBUS left	10:24
	← BearPerson left	10:25
	→ BearPerson joined	10:25
	→ jkp joined	10:26
	→ tcv_br joined	10:26
VladDrac	I've created a git repo, done some stuff and commits	10:26
	now I've created a bare clone out of it and rsynced it to a remote server, which will basically be my central repo	10:26
	is there a way to change my origin to this repo?	10:26
	saves me from cloning it and rebuilding it	10:27
thiago_home	VladDrac: yes, git config remote.origin.url	10:27
VladDrac	I shouldn't have asked "is there a way"	10:27
	of course there is	10:28
	with git there always is :)	10:28
	many thanks	10:28
	→ drizzd_ joined	10:29
doener	VladDrac: modern git versions also have "git remote set-url"	10:30
	← xiangfu left	10:32
	← drizzd left	10:32
	→ dwmw2_gone joined	10:35
j416	if I git clone any repo, are _all_ branches and all data cloned without exception?	10:36
	← egrouse left	10:36
	← h0st1le left	10:36
j416	(is there a setting to specify clone not to fetch objects for a certain branch etc?)	10:36
	→ h0st1le joined	10:37
thiago_home	j416: instead of cloning, init + remote add	10:39
	then you can change the fetch refspec	10:39
j416	thiago_home: what do you mean?	10:39
	← Utkarsh left	10:39
	→ Utkarsh joined	10:39
	← roger_padactor left	10:39
thiago_home	init, remote add, change the refspec, then fetch	10:40
j416	ah, sorry, I was wondering more in general	10:40
	→ egrouse joined	10:40
	← egrouse left	10:40
j416	if I clone, will everything be fetched?	10:40
	→ airborn joined	10:40
j416	(is there a setting someone can make so that that will not be the case?)	10:40
thiago_home	yes, if you clone, you get all branches	10:40
	all local branches	10:40
j416	so if I clone another repo, I can be sure that I have all its data	10:41
	and all tags?	10:41
	→ Bruce_Wayne joined	10:41
j416	I've been using git for quite a while now, but haven't really explored much of the remote features.	10:41
thiago_home	all tags and all local branches	10:41
j416	thanks.	10:41
	:)	10:41
	→ Chris64 joined	10:41
Chris64	hi all	10:41
	→ herlock joined	10:42
Chris64	is there a git client for android available?	10:42
j416	hi single one	10:42
Chris64	;)	10:42
	→ egrouse joined	10:42
	← egrouse left	10:42
j416	android. o_O never heard of git for android. wouldn't it be very cumbersome to develop on a.. phone? :D	10:43
	→ nevyn^ joined	10:43
j416	s/develop/do development/	10:43
Chris64	or just for reading, don't know :)	10:43
herlock	Hello. I would like to restore a previous commit. I mean, I would like to be able to work only on the files that were in the state of this commit. Without the changes made after that commit (and without the files that were created after that commit)	10:44
j416	Chris64: the closest solution _I_ can think of, is if you ssh to a computer that has git.	10:44
herlock	can you please give me a pointer to a doc that explain it ?	10:44
j416	herlock: you want either git revert or git reset	10:44
	herlock: man git-revert	10:44
jast	herlock: the 'git-revert' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-revert.html [automatic message]	10:44
j416	herlock: man git-reset	10:44
jast	herlock: the 'git-reset' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-reset.html [automatic message]	10:44
Chris64	j416: this sounds good, thanks :)	10:45
herlock	that's git reset or git revert ?	10:45
j416	'git revert' will create a commit recording the undoing of that commit. 'git reset' can make it look as if though that commit never happened.	10:45
	← notze left	10:46
j416	if you are collaborating and other people have fetched your branch, git revert would be appropriate. If you're on your own, git reset might be what you want.	10:46
	← lemonchicken left	10:46
malesca	herlock: Sounds like what you want is "git reset --hard mycommit", but read up first since you will lose every commit after that one (though they're technically not lost forever).	10:47
j416	(and any uncommitted changed with that, mind)	10:47
	s/changed/changes7	10:47
	s/7/\//	10:47
	bah	10:47
	:P	10:47
malesca	:)	10:47
	→ xzer joined	10:47
	→ slonopotamus joined	10:48
herlock	so I could do something like git revert	10:48
	← wliao left	10:48
herlock	oosp	10:48
j416	herlock: git revert [the sha1 of the commit you want to undo and record a revert commit for]	10:48
	herlock: git reset --hard [the commit before the commit you don't want]	10:49
	note again: be sure not to have any uncommitted changes you want to save.	10:49
	← sudonotplease left	10:50
	← Buglouse left	10:50
	← xzer left	10:50
j416	herlock: "git revert" is "record a commit that reverses the specified commit"; "git reset" is "reset my branch to the state at the specified commit"	10:50
	if that makes it more clear.	10:50
	→ sudonotplease joined	10:51
	← dr_win left	10:51
	→ egrouse joined	10:51
	← egrouse left	10:51
	→ dr_win joined	10:52
	→ Octalot joined	10:53
	→ ph^ joined	10:54
	→ GaVrA joined	10:54
	→ lemonchicken joined	10:54
	← Bruce_Wayne left	10:56
	→ RevoOf joined	10:56
herlock	ok, thank you guys :)	10:58
	→ maleknet joined	10:59
j416	you're welcome.	11:00
	→ tvw joined	11:02
	→ egrouse joined	11:02
	← egrouse left	11:02
	← slonopotamus left	11:02
	→ Utkarsh_ joined	11:02
	→ egrouse joined	11:02
	← egrouse left	11:02
	← reZo left	11:02
	→ slonopotamus joined	11:03
	← shiba_yu36 left	11:04
	← Utkarsh left	11:04
	→ timotei joined	11:04
	→ egrouse joined	11:04
	← egrouse left	11:04
timotei	hi. I had a series of commits which I was gonna git svn dcommit.	11:05
	but perl crashed in the meantime, leaving me with uncommited changes and stuff	11:05
	how can I resume the dcommiting?	11:05
	→ goshawk joined	11:06
	← stliu left	11:06
	→ egrouse joined	11:08
	→ Utkarsh joined	11:11
	→ Yuuhi joined	11:12
	→ bentob0x joined	11:12
	← maleknet left	11:12
	← Utkarsh_ left	11:13
	→ Geralt joined	11:14
Geralt	Hi, is it possible to apply only parts of a (stashed) commit to the working directory, like git add --patch or git checkout --patch?	11:14
	→ jstemmer joined	11:17
	← tvw left	11:18
j416	Geralt: I can't find such an option. Maybe because the stash is somewhat like a patch in itself.	11:18
	→ maleknet joined	11:18
j416	then you'd want to add --patch to git apply also, that is, apply a patch to the patch to apply... kind of	11:19
	eh :	11:19
	:)	11:19
	just speculating. You can do git stash save --patch, though. To only save some of it.	11:19
	or you can just apply the stash and commit only part of it through doing git add --patch first.	11:20
Geralt	j416: I already have these changes stashed, and yeah in the end I did it that way, git stash apply and then git checkout -p to discard what I don't want	11:20
	Is a stash not just a commit?	11:21
j416	Geralt: why not apply, then commit what you want to _save_, then reset?	11:21
	← samhassell left	11:21
j416	that's how I would do it	11:21
	a stash is a commit afaik.	11:21
	→ level09 joined	11:21
Geralt	j416: I have already stuff for another commit in the index and unfortunately I stashed a bugfix together with some debugging output away (and guess how long it took me to find out we the bug is suddenly back...) :-/	11:22
	we=why	11:22
j416	Geralt: have you found a way to selectively add changes from a specified commit?	11:22
	hm	11:22
Geralt	j416: not really, except for git cherry-pick --no-commit and then git add --patch	11:22
j416	I'd suggest you commit more often :)	11:23
	remember you can undo commits	11:23
Geralt	j416: I say that to myself all the time too...	11:23
j416	heh	11:23
Geralt	maybe one day I really get it into my muscle memory to do that :-)	11:23
j416	Geralt: one thing you _could do_,	11:24
	is	11:24
	commit whatever you have in the index now	11:24
	apply the stash	11:24
	then commit part of it	11:24
	← sudonotplease left	11:25
	→ sudonotplease joined	11:26
Geralt	j416: yes, I could have but I already discarded what I don't want, so I'm fine too and thanks for your help :)	11:26
timotei	so, it's a way to resume a stopped svn dcommit?	11:26
j416	ok. :)	11:26
timotei	that is, continue to apply the "commits"?	11:26
Geralt	And now I'll commit these changes right away	11:26
j416	Geralt: also, if you're doing bugfixes and everything on the same branch.. maybe you should have a look at this, I find it rather clever: http://nvie.com/posts/a-successful-git-branching-model/	11:26
	(not sure that is what you're doing, just hinting	11:26
	)	11:26
	→ curtana joined	11:28
	→ Fullmoon joined	11:29
	→ Dave^\| joined	11:32
timotei	http://blogs.gnome.org/diegoe/2009/03/18/saving-your-neck-when-git-svn-dcommit-fails/	11:32
	I've found the solution :D	11:32
	thanks anyway	11:32
Geralt	j416: I'm familiar with that model, but we don't have releases yet and we are only two people hacking on it, so development is pretty linear, with the usual merge/cherry-pick.	11:32
	← Caius left	11:32
j416	Geralt: ok. :)	11:33
	Geralt: I'm only one person, and I'm starting to adapt at least part of that model	11:33
	→ samhassell joined	11:33
j416	I think it's a good guide to have in mind anyway.	11:33
	← ch077179 left	11:34
	→ Caius joined	11:34
	← Caius left	11:34
	→ Badiss joined	11:42
	← harshu left	11:43
	→ harshu_ joined	11:43
	→ joel_falcou joined	11:45
	← pedrobelo left	11:47
	→ Titosemi joined	11:47
	→ pedrobelo joined	11:47
	← funfun left	11:51
	→ pyther joined	11:52
	← pyther left	11:52
	→ pyther joined	11:52
	→ pozic joined	11:52
pozic	Are there tools to go from a CVS repository to a Git repository with one command?	11:53
	← level09 left	11:54
	→ jstein joined	11:55
	→ funfun joined	11:55
jast	pozic: man git-cvsimport @	11:55
	pozic: the 'git-cvsimport' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-cvsimport.html [automatic message]	11:55
	→ bluenovember joined	11:56
jstein	hi, i dont understand how to merge from the manual. Can you help me? I have a repository on github https://github.com/jonasstein/fuzzer and a local one on my pc. "upliner" changed a line in the same file i was working on. How can i merge/push my stuff now?	11:57
	← timotei left	11:58
	← boombatower left	11:58
	→ Caius joined	11:59
	→ gmcdonald joined	11:59
	← slonopotamus left	11:59
	→ slonopotamus joined	11:59
	→ d0k joined	12:02
Geralt	jstein: fetch his changes (git fetch), merge it into your branch (git merge) then try again	12:02
	→ notze joined	12:04
	← Utkarsh left	12:05
	← iravid left	12:05
	→ iravid joined	12:06
	→ ixti joined	12:08
	← herlock left	12:09
	← urkud_ left	12:09
	← cesario left	12:10
	← OOPMan left	12:11
	→ ludde joined	12:15
	← maleknet left	12:17
	→ maleknet joined	12:19
	← jogla left	12:19
jstein	Geralt: i have only one local branch. Is it correct, that i need two? github ---> mixbranch <--- localmaster	12:20
Geralt	jstein: I'm not sure I understand your question. Can I see the output of 'git branch -a'?	12:23
	← Fullmoon left	12:23
jstein	sure: http://paste.debian.net/107402/	12:24
	→ dabaRs joined	12:24
Geralt	jstein: ok, so all you need to do is 'git pull && git push'	12:25
jstein	Geralt: http://paste.debian.net/107404/ i think i have to add the url to github, right?	12:29
	← slonopotamus left	12:30
	→ urkud_ joined	12:31
	→ cesario joined	12:31
	← nevyn^ left	12:32
Geralt	jstein: ok, then use git pull origin/master and if that fails git pull remotes/origin/master	12:32
	→ nevyn^ joined	12:32
	← ixti left	12:35
	→ hansu joined	12:36
	← notze left	12:36
	← DiGitalX left	12:36
	→ leavittx joined	12:37
	→ slonopotamus joined	12:38
	← tcv_br left	12:39
	← isionous left	12:41
	← priidu left	12:41
	← drl left	12:42
	→ saste joined	12:44
	← cesario left	12:45
jstein	$ git pull origin/master	12:45
	fatal: 'origin/master' does not appear to be a git repository	12:45
	fatal: The remote end hung up unexpectedly	12:45
	$ git pull remotes/origin/master	12:45
	fatal: Not a git repository (or any parent up to mount parent )	12:45
	Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).	12:45
	:(	12:45
	← int3h left	12:46
Geralt	jstein: is origin not your github repo?	12:47
cbreak	jstein: it's not git pull origin/master, it's git pull origin master	12:47
saste	http://pastebin.com/kUbq8X8T	12:47
	cannot get my head aound git-filter-branch...	12:47
cbreak	jstein: man git-pull	12:48
jast	jstein: the 'git-pull' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-pull.html [automatic message]	12:48
charon	jstein: there's a big difference between 'git {fetch,pull} origin master' ("please {get,merge} the master branch of the remote origin") and 'git merge origin/master' ("merge the last known state of master at origin, which is stored in refs/remotes/origin")	12:48
	jstein: in particular the former goes over the network and the latter doesn't	12:48
	(also, pull is roughly fetch+merge)	12:48
	finally, do note that 'git pull origin master' does not update origin/master for historical raisins.	12:49
	(bah, got my puns all mixed up)	12:49
cbreak	saste: use env-filter	12:49
	→ rns joined	12:49
cbreak	or fix your commit filter to do the proper things	12:49
jstein	i look in the .git/config if i can fix the serveraddress there	12:49
	→ Utkarsh joined	12:50
	→ DiGitalX joined	12:50
charon	jstein: most likely this doesn't have a thing to do with broken urls	12:50
	← pdelgallego left	12:50
cbreak	jstein: why fix a server address?	12:50
	it's not like you told git to even USE a remote	12:51
charon	(i.e., even if you had the correct origin url configured, 'git pull origin/master' would still fail as that is a nonsensical request)	12:51
cbreak	fix your commands	12:51
	→ priidu joined	12:52
saste	cbreak: http://pastebin.com/YhTmGAHh	12:54
cbreak	saste: don't commit tree!	12:55
	read the damn man page...	12:55
	→ drl joined	12:56
saste	cbreak: i'm doing that since one hour but imo the man is quite poor if you don't have already an intimate knowledge with git	12:58
	→ roger_padactor joined	13:00
	← dr_win left	13:00
cbreak	well... it's a man page, not a tutorial :)	13:01
jstein	ok the pull && push worked now, but i have a ugly HEAD line in source now. i thought git would automatic join the files if possible	13:01
	https://github.com/jonasstein/fuzzer/blob/master/fuzzyselect.py#L10	13:01
	saste: dont worry i have similar problems with the documentation	13:02
	← guardian left	13:02
	← j416 left	13:02
cbreak	join files?	13:03
	HEAD line?	13:03
	you're supposed to fix merge conflicts yourself	13:03
	there is no automated way to solve all of them	13:04
	you can use a merge tool that can help you if you figure out how :)	13:04
saste	jstein: well I did it with git rebase -i and git commit --amend --author.. I don't have time to dive for days in git docs...	13:04
	→ Zhonghua joined	13:04
cbreak	env filter is trivial	13:04
	just export your variables, then it'll do the rest	13:05
saste	cbreak: if you can point to the error I'm doing i'll be grateful	13:05
	cbreak: ehi maybe i'm just dumb...	13:06
cbreak	as I said, just set the variables, export them, and you're good	13:06
	no commit tree	13:06
	env filter is a very easy one	13:06
	commit filter is a complicated but powerful one	13:06
jstein	saste: do you want to write a better howto for newcomers?	13:06
cbreak	it's probably the most powerful of all filters	13:06
	→ DaveIngram joined	13:06
	← xiong left	13:06
jstein	saste: i bought the apress git book. It is quite good, but some things are still far beyond my uynderstanding	13:07
saste	jstein: maybe... but you know I have my own problems and my own docs to write... maybe a blog post..	13:07
	adamholt → adamholt_away	13:08
	→ isionous joined	13:08
	adamholt_away → adamholt	13:08
	→ g0bl1n joined	13:09
	→ Ruudjah joined	13:10
	← vandemar left	13:10
	→ _iron joined	13:10
	← tewecske left	13:12
	→ guardian joined	13:12
	→ j4mie joined	13:13
	← guardian left	13:13
	→ tewecske joined	13:13
	← dabaRs left	13:13
	← Araxia left	13:14
	← nevyn^ left	13:14
	← maleknet left	13:16
	→ vandemar joined	13:16
	→ MattDiPasquale joined	13:16
	→ nazgul102 joined	13:16
	← nazgul101 left	13:17
	← orafu left	13:17
	→ maleknet joined	13:18
	→ juangiordana joined	13:20
	→ kbedell joined	13:20
	← protorom left	13:21
	← j4mie left	13:23
	← sechrist left	13:24
	← MattDiPasquale left	13:24
	→ simplechat joined	13:24
	← kbedell left	13:29
	← drl left	13:31
	← pratz left	13:31
	← harshu_ left	13:32
	← maleknet left	13:34
	← RevoOf left	13:34
	→ kipras joined	13:35
	→ j4mie joined	13:35
	← urkud_ left	13:36
	→ pratz joined	13:37
jstein	thank you all for helping with my first merge	13:38
	← hansu left	13:39
	→ xiangfu joined	13:40
	← slonopotamus left	13:41
	→ drl joined	13:43
	→ mw_ joined	13:44
	→ sechrist joined	13:45
	→ notze joined	13:45
	→ guardian joined	13:46
	← TomyLobo left	13:47
	→ guardian1 joined	13:47
	→ _pingu joined	13:48
	→ pdelgallego joined	13:49
	← guardian left	13:50
	← drl left	13:50
	→ urkud_ joined	13:51
	← pratz left	13:53
	→ pratz joined	13:56
	adamholt → adamholt_away	13:58
	adamholt_away → adamholt	13:58
	← mastro left	14:00
	← j4mie left	14:00
	← JStoker left	14:02
	→ maattd joined	14:03
	→ drl joined	14:03
	→ gusnan joined	14:05
	→ markelikalderon joined	14:05
	→ Ulfalizer joined	14:07
	→ Mannequin joined	14:08
Ulfalizer	how can i get a list that maps branches to the branches they track?	14:08
charon	saste: did you get it fixed? the env-filter is just supposed to (re)set the variables documented in man git-commit-tree	14:09
jast	saste: the 'git-commit-tree' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-commit-tree.html [automatic message]	14:09
charon	saste: that being said, i think it's difficult to write a beginner's tutorial to one of the most advanced tools in git	14:09
Ulfalizer	i often give me local branches mnemonic names, and tend to forget what remote branches they track	14:09
	*my	14:09
charon	(not because env filters are hard, but because some of the more advanced things like changing parents can give very strange results)	14:10
saste	charon, jast: yes it was trivial (once you know the exact solution) ;)	14:10
	charon, jast: and my use case was pretty common	14:10
	charon, jast: I had some trouble at understanding the difference between the various kind of filters	14:11
charon	saste: poor jast just sent you a bot reply :)	14:11
	← kipras left	14:12
Ulfalizer	is this hard to do? i figured there'd be some simple command and that my rtfm skills weren't up to snuff :)	14:13
	→ ch077179 joined	14:14
cbreak	Ulfalizer: man git-branch	14:15
jast	Ulfalizer: the 'git-branch' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-branch.html [automatic message]	14:15
cbreak	do something like git branch -vv	14:15
	→ mastro joined	14:15
charon	or for script use, apply 'git rev-parse --symbolic-full-name $branch@{upstream}'	14:16
Ulfalizer	ah, didn't notice the tracked branch appeared within [] after the commit SHA. thanks!	14:17
	with git branch -vv that is	14:17
	it'd be neat if you could get a format where each line is "<branch> -> <tracked branch>", without the additional clutter	14:18
	perhaps git rev-parse can help with that	14:18
saste	charon: http://pastebin.com/B3WyK37m	14:19
	→ ThomasLocke joined	14:19
charon	Ulfalizer: git for-each-ref --format="%(refname:short)" refs/heads/ \| while read r; do echo "$r" "$(git rev-parse --symbolic-full-name --abbrev-ref "$r@{upstream}")"; done	14:21
	untested	14:21
	well, ok, it doesn't like that if a branch doesn't track anything	14:21
doener	git for-each-ref --format='%(refname) %(upstream)' refs/heads	14:22
	→ sedeki joined	14:22
Ulfalizer	doener: neat :)	14:22
charon	doener: oh wow. why did i try that and conclude it didn't work, yet it does :)	14:22
doener	charon: for-each-ref even knew %(upstream) before @{upstream} was introduced ;-)	14:23
charon	saste: you could submit a patch to make that an example in the manpage	14:23
	→ dr_win joined	14:23
charon	saste: still, i believe one cannot fully understand g-f-b before one has the git knowledge to simply read the script. i wonder if pseudocode in the manpage might help	14:24
	→ munhgauzen joined	14:25
	← Adaptee left	14:25
	← spaceonline left	14:25
	← Blackguard left	14:26
	→ linusoleander joined	14:26
	← anderse left	14:27
	→ iasterisk joined	14:27
	← munhgauzen left	14:27
	→ hyperair joined	14:28
	→ anderse joined	14:28
	→ Anubisss joined	14:29
	→ Blackguard joined	14:31
	→ munhgauzen joined	14:32
	→ evilgeenius joined	14:33
evilgeenius	how can delete all the changes I have made since I last commited?	14:33
	→ JStoker joined	14:33
cbreak	git reset --hard HEAD	14:34
	for example	14:34
	or git checkout HEAD -- .	14:34
	← HG` left	14:37
	→ jergen joined	14:37
evilgeenius	Is there a difference between those 2 commands?	14:37
cbreak	yes, many	14:38
	← JStoker left	14:39
	← goshawk left	14:40
	← bluenovember left	14:43
	← _pingu left	14:44
	→ nevyn^ joined	14:46
	→ bluenovember joined	14:48
	← bradleyayers left	14:48
	← joel_falcou left	14:50
	← bluenovember left	14:52
	→ JStoker joined	14:54
	← rns left	14:58
	→ bluenovember joined	14:58
	← caruso_g left	14:59
	← drl left	14:59
bluenovember	hey all. I messed around with my config mergetool and tortoisemerge path, and now when I run "git mergetool" I get "No files need merging" despite there clearly being conflicts. How can I fix this?	14:59
	→ nonnooo joined	14:59
bluenovember	(using msys git on windows, ideally wanting to use tortoisemerge as mergetool. my .gitconfig has no lines relating to merge in it now)	15:00
Chris64	this would be interesting to me too	15:01
cbreak	bluenovember: git config -l and git config -l --global	15:01
	→ cesario joined	15:02
bluenovember	cbreak, merge.tool=tortoisemerge and merge.conflictstyle=diff3	15:02
	nothing else under merge	15:02
	or mergetool	15:03
cbreak	well, remove those :)	15:03
bluenovember	heh kk	15:03
	→ kbedell joined	15:03
bluenovember	no change cbreak	15:04
cbreak	git status shows conflicts?	15:05
bluenovember	(git config --unset merge.tool etc, re-ran git-config -l, doesn't show anything relating to merge. re-ran "git mergetool" in conflicted state, "no files need merging"	15:05
cbreak	interesting. weird.	15:05
bluenovember	mmm actually shows staged changes, unmerged paths, and untracked (.backup, .base, .local, .remote files)	15:05
	→ Rogue joined	15:05
bluenovember	conflicts from rebasing a branch	15:05
	← cesario left	15:06
bluenovember	rebase --abort and redoing rebase results in the same state	15:06
	← sedeki left	15:06
	← Rogue left	15:07
	→ Rogue_Coder joined	15:07
	→ sivy joined	15:07
	← rchavik left	15:07
bluenovember	This all started when I removed tortoise\bin from path. Git was then giving me vimdiff. I couldn't get git to use tortoise again (despite setting path in $PATH and having merge.tool set) so reinstalled tortoise and nuked git config lines.	15:08
	← Rogue_Coder left	15:08
	← notze left	15:08
	→ Rogue_Coder joined	15:08
bluenovember	(I'm far too partial to the shotgun approach to solving problems)	15:08
	← sonnym left	15:09
	← Rogue_Coder left	15:09
	→ Rogue_Coder joined	15:09
	← nonnooo left	15:10
	→ HG` joined	15:10
	← Rogue_Coder left	15:10
	→ Rogue_Coder joined	15:11
	← munhgauzen left	15:11
	→ nevyn^1 joined	15:12
	→ drl joined	15:12
	← nevyn^ left	15:13
	→ RobertLaptop joined	15:14
	→ kukks joined	15:16
	→ icwiener joined	15:17
	← Geralt left	15:17
	← urkud_ left	15:18
	→ chrisshattuck joined	15:20
	← ch077179 left	15:20
	→ noonoooo joined	15:22
	← drl left	15:22
	→ jvraoltc joined	15:27
	→ notze joined	15:29
	→ ch077179 joined	15:31
	→ Maletor joined	15:32
	← g0bl1n left	15:34
	→ drl joined	15:34
	→ elfgoh joined	15:36
	← noonoooo left	15:36
elfgoh	Hi all, I am troubleshooting a git repo. git status shows that file 'a' has been modified. But git diff does not show any difference. Any ideas?	15:37
mnemoc	permissions?	15:38
elfgoh	mnemoc: no it isn't just discovered that it is due to windows line breaks :o	15:40
	→ piggz joined	15:41
mnemoc	diff should show those as ^M	15:41
cbreak	if he has weird line ending conversion settings, it can cause such problems	15:41
	→ moonk joined	15:43
piggz	hi, i have a local branch, there have been verios commits to it, and a couple of 'git pull origin master' to bring in the stuff from the master main repo.....git status says it is 700+ commtis ahead of the remote branch, can i rebase it or anything before pushing it to remove all the master commits?	15:43
elfgoh	mnemoc cbreak: ic. Those ^M only in 'git diff', not 'git diff a'	15:43
	Tks for the advice. Good day!	15:43
curtana	that was fun to debug!	15:44
cbreak	piggz: you can do git rebase origin/master	15:45
	→ urkud_ joined	15:45
piggz	cbreak: ah...i was trying ;git rebase orign master' .... why does one need the slash and the other not?	15:46
bluenovember	cbreak Chris64, curtana helpfully solved my issue. The merge was showing "nothing to do" as I was in a subdir of my git dir which had no conflicts. The issue with tortoisemerge not being found was due to me having a _line break character_ in my path. :\	15:46
	→ intripoon joined	15:46
cbreak	piggz: rebase rebases onto what you tell it. origin/master is a local tracking branch for origin's master	15:47
	← intripoon_ left	15:48
	← maattd left	15:49
	← moonk left	15:49
piggz	cbreak: does this look right: ?	15:50
	# On branch kexi-mobile-adam_pigg	15:50
	# Your branch and 'origin/kexi-mobile-adam_pigg' have diverged,	15:50
	# and have 736 and 9 different commit(s) each, respectively.	15:50
cbreak	why not?	15:50
piggz	cbreak: just wondering what the 736means	15:51
	→ goshawk joined	15:51
	← normanrichards left	15:51
cbreak	736 commits in one branch that the other does not have	15:51
	→ SimonP86 joined	15:53
	← elfgoh left	15:56
piggz	cbreak: this is what i did http://pastebin.com/zQuvvhpe	15:56
cbreak	do you like the result?	15:58
	→ maattd joined	15:58
	← maattd left	15:58
	→ nonnooo joined	15:58
	→ maattd joined	15:59
piggz	cbreak: i dont know :) ... the 736 seems like a large number and im not sure what it means...i have 16 local commits, so shouldnt they differ by 16?	16:00
	← larsmoil left	16:00
	← Titosemi left	16:00
	→ larsmoil joined	16:01
cbreak	piggz: well, look at it with a history viewer such as gitx or gitk	16:01
	after the rebase, they should have no commits you don't have	16:01
	(you did do a git fetch or git remote update before the rebase, right?)	16:01
	oh, what the hell	16:01
	you actually did rebase origin/master?	16:02
	you should have rebased onto your own upstream...	16:02
	→ Titosemi joined	16:02
cbreak	origin/master was just an example	16:02
piggz	cbreak: the admin has decided to delete the remote branch to push it again (........not quite getting the hang of git yet ;)	16:04
cbreak	well, you can rebase on the newly pushed version.	16:04
	you should have done git rebase yourremotenamehere/yourremotebranchnamehere	16:05
	→ OOPMan joined	16:05
	→ bradleyayers joined	16:05
	→ project2501a joined	16:05
piggz	cbreak: and how do i make the branch keep in sync with the remote master	16:05
cbreak	remote master? git rebase remotename/master	16:05
	obviously this makes you rewrite history	16:06
	so you will have to force push	16:06
	(and everyone else who uses the branch has to rebase instead of merge when he pulls it)	16:06
	piggz needs a book ;)	16:08
	← project2501a left	16:10
	← xiangfu left	16:11
	→ dreiss joined	16:15
	← cognominal left	16:16
	← sivy left	16:17
	→ tcv_br joined	16:17
	→ project2501a joined	16:17
	→ jstein_ joined	16:19
	← Badiss left	16:20
	← urkud_ left	16:20
	→ andrew_ joined	16:21
andrew_	hello all	16:21
	← acid left	16:21
	← jstein left	16:22
andrew_	im working in a repo that has a submodule. i've branched that repo and have started working in the new branch. when i pull on the new branch, it doesnt seem to be updating the submodule. how should i go about updating the submodule?	16:22
	← jettero_ left	16:24
	→ jettero joined	16:24
	← jettero left	16:24
	→ jettero joined	16:24
	← project2501a left	16:24
	← jettero left	16:25
	→ jettero joined	16:26
	← jettero left	16:26
	→ jettero joined	16:26
cehteh	andrew_: submodules are kindof manual .. you have to update them by yourself, while some git commands allow a --recursive option to recurse into submodules, rtfm for further info	16:27
	← spatulasnout left	16:28
	→ spatulasnout joined	16:29
	→ g0bl1n joined	16:29
andrew_	cehteh : can you clarify what you mean by manual in this context?	16:30
cehteh	do whatever you want to do by a command	16:30
	← angelsl left	16:30
andrew_	should i be concerned that the submodules in different branches of the same repo have different identifiers?	16:31
cehteh	enter the submodule, push/fetch there .. or commit the new submodule state on the level higher or whatever	16:31
	→ chrischambers_ joined	16:31
cehteh	identifiers?	16:31
	← MegaHerz left	16:31
	→ zombor joined	16:32
andrew_	https://github.com/jcdickinson/cecil/tree/203afaa5188f3dfcba3d01f91574f5caeab20460 vs https://github.com/jcdickinson/cecil/tree/762c44c57093abf12e580c96cb68fe4ddfca9661	16:32
	← Chris64 left	16:32
andrew_	the commit identifiers, uuids, whathaveyou	16:32
	← chrischambers left	16:33
	chrischambers_ → chrischambers	16:33
	→ ChUrP joined	16:33
cehteh	where are submodules there (i dont know how to use github in detail)	16:34
	→ elfgoh joined	16:34
andrew_	https://github.com/jcdickinson/Monoflector/tree/ui-unstable/External and https://github.com/jcdickinson/Monoflector/tree/master/External	16:35
cehteh	i still dont see submdules there :P	16:36
	→ vaddim joined	16:36
cehteh	https://github.com/jcdickinson/Monoflector/blob/ui-unstable/.gitmodules	16:36
	ah :)	16:36
andrew_	sorry, didnt know you were looking for that	16:36
	new learning and all	16:36
cehteh	so what do you try to do .. sync from upstream?	16:36
	→ cognominal joined	16:37
andrew_	i need to bring the submodule in the ui-unstable branch up to date with the submodule in master, and merge my changes to the submodule in the ui-unstable branch	16:37
cehteh	you cloned it already .. with git clone --recursive ?	16:37
andrew_	yus	16:37
cehteh	git submodule update --recursive	16:38
	when you want to switch branches/submodules you have to go into this submodule, do the desired changes there	16:38
	(git checkout possibly).	16:39
	then go back to the higher level .. git diff will show you that you changed the submodule, you need to commit (git add + commit) that there too	16:39
	if you want to publish this you possibly have to fix the .gitmodules too	16:40
andrew_	good deal. that command worked	16:40
	anything special to pushing my changes to the main branch submodule?	16:41
	← drl left	16:41
cehteh	overall the submodule is a bit pita to use, the backend support in git is ok but the frontend is bit mindbending, explicit and manual	16:41
andrew_	hehe im getting that	16:42
cehteh	just do as usual for every change, that is push the submodule repo up and dont forget to commit the new submodule state in the upper level and push that up too	16:42
	→ workmad3 joined	16:43
andrew_	thanks for the assist, much appreciated	16:43
cehteh	gtg	16:44
	→ yogione joined	16:45
	← dreiss left	16:46
	← vaddim left	16:46
	→ vaddim joined	16:46
	→ dekebal joined	16:48
	→ vaddimk joined	16:48
	← workmad3 left	16:48
	← elfgoh left	16:49
	→ sternocera joined	16:50
sternocera	Hi. How do I apply a patch to my working copy, and then apply another patch on top of that?	16:50
	← vaddim left	16:52
bremner	sternocera: perhaps man git-apply is what you seek	16:52
jast	sternocera: the 'git-apply' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-apply.html [automatic message]	16:52
	← vaddimk left	16:52
	← kpreid left	16:54
	→ kpreid joined	16:55
	→ drl joined	16:55
	← curtana left	16:55
	← yogione left	16:56
	← maattd left	16:58
	→ rajeshsr_ joined	16:59
	→ n3m joined	17:00
	→ sivy joined	17:01
	← LongBeach left	17:01
	ChanServ set mode: +v	17:02
	→ urkud_ joined	17:03
	← airborn left	17:04
	← guardian1 left	17:04
	← drl left	17:05
	→ Textmode joined	17:06
	→ curtana joined	17:06
	→ pgib joined	17:07
	→ sourcemaker joined	17:07
	← andrew_ left	17:09
	→ j416 joined	17:10
	→ lanthan joined	17:10
	← mw_ left	17:12
sternocera	bremner: My concern is that doing two git applys in succession will not have the desired effect.	17:12
	→ gilligan_ joined	17:14
	→ Yuffster joined	17:15
bremner	sternocera: can you be more specific? I don't understand what your worry is. I also don't understand why you don't just try it, and see how it works.	17:15
	→ kipras joined	17:15
	← zombor left	17:16
sternocera	Does applying two patches have any effect other than independently applying each set of changes?	17:16
bremner	well, if the conflict, the order will matter.	17:16
	← madsy left	17:18
	→ drl joined	17:18
sternocera	ah yes	17:19
	but they're atomic operations anyway	17:19
	→ navatwo joined	17:19
	← jvraoltc left	17:20
navatwo	Hi, I cannot seem to get git to work with remotes. All I am trying to do is merge a pull request from github. Everytime I checkout to my upstream repos it says im in detached HEAD mode. I have fetched the upstream repos, and pulled it into the branch. When I check out it says that I'm not on any branch. Output: http://paste.villavu.com/show/410/	17:22
bremner	sternocera: When you say patch do you mean commit? Patches are just text files with add-delete lines. They can easily fail halfway through.	17:22
	← j0ran left	17:22
j416	navatwo: that is normal.	17:23
navatwo	I can't push to the upstream remote though.	17:23
j416	navatwo: you cannot commit to a remote branch directly	17:23
	you create a local branch, commit to that, then push to upstream.	17:23
	← hagabaka left	17:23
	→ alama joined	17:24
j416	checking out the upstream directly lets you work on that, but your commits are not referenced by any branch ( = detached HEAD)	17:24
	→ Araxia joined	17:24
navatwo	so if I wanted to commit to the upstream, and I have my local branch, how would I do it then?	17:25
j416	so when you move to another branch, you have to either remember the sha1 of your commit in detached HEAD mode, or check reflog.	17:25
	← sivy left	17:25
	→ MattDiPasquale joined	17:25
j416	navatwo: http://help.github.com/git-cheat-sheets/	17:25
	"Creating a tracking branch"	17:25
sternocera	bremner: I just mean a patch file.	17:26
	← curtana left	17:26
alama	thinking of setting up a cron script to do git stash && git apply every hour	17:27
	→ hagebake joined	17:27
j416	navatwo: or http://stackoverflow.com/questions/520650/how-do-you-make-an-existing-git-branch-track-a-remote-branch	17:27
bremner	sternocera: well, I don't think git-apply is atomic, I'm not sure. But since it doesn't create a commit, you have a chance to inspect the results.	17:28
	→ sivy joined	17:28
navatwo	whats the premise behind tracking?	17:28
j416	please elaborate	17:29
sternocera	bremner: okay, thanks.	17:29
j416	/clarify	17:29
	← pfrenssen left	17:29
	→ mdpatrick joined	17:29
navatwo	j416: I have my local branch tracking the upstream, how would I go about committing to the upstream then. I still get fast-forward erros	17:30
j416	commit to your local branch	17:30
	← MattDiPasquale left	17:30
j416	push to remote	17:30
	navatwo: faq non-ff	17:30
jast	navatwo: Your push would lose changes on the remote: please see https://git.wiki.kernel.org/index.php/GitFaq#non-ff [automatic message]	17:30
navatwo	I've followed that and had no help.	17:31
j416	what happened?	17:31
navatwo	it just gives the errors.	17:32
	I feel like forcing the commit is ridiculous too.	17:32
j416	"the errors"?	17:32
navatwo	just says fast forward	17:32
j416	doubt that.	17:32
navatwo	non*	17:32
j416	did you do git pull?	17:33
navatwo	http://paste.villavu.com/show/411/	17:33
	yes.	17:33
	says its up to dete	17:33
	date*	17:33
j416	in that case your commits have diverged	17:33
	you need to either lose your changes locally and make your local reflect the remote	17:33
	or	17:33
	force-update remote	17:33
	→ level09 joined	17:33
j416	or, create a new tracking branch for the remote, merge that, then push	17:33
	s/merge that/merge your branch into that/	17:34
	← chrisshattuck left	17:35
j416	navatwo: you should read about how git works internally. git object model and branches.	17:35
	it would help.	17:35
	it's not that complicated. well invested time.	17:35
navatwo	... now `upstream/master` is considered ambiguous.	17:36
j416	because you created a local branch with the same name perhaps.	17:36
	try refs/remotes/upstream/master	17:36
	→ kumbayo_ joined	17:38
navatwo	You were right about that, I deleted the branch.	17:38
	j416: http://paste.villavu.com/show/412/	17:38
	I do not see what I did wrong there..	17:38
	:s	17:38
	→ towski joined	17:39
navatwo	It even states that it is a fastforward.	17:39
	:\|	17:39
j416	weird	17:40
	dunno	17:40
	how does the respective logs look?	17:40
navatwo	I cannot commit to my personal repos either	17:40
j416	why?	17:40
navatwo	Same issue, says its a non-ff	17:40
	→ engla joined	17:40
	→ chrisshattuck joined	17:41
	→ Schmallon joined	17:41
	← engla left	17:41
	→ engla joined	17:41
tango_	navatwo: commits don't talk about fast-forwardness	17:41
	← kumbayo left	17:41
navatwo	the merge isn't in the logs.	17:41
	:s	17:41
	→ madewokherd joined	17:42
	→ tildeequals joined	17:42
tango_	navatwo: from the pastebin above, what did you was to checkout the upstream/master branch to a detached head	17:42
	navatwo: that is not a merge	17:42
j416	navatwo: it shouldn't be, you never merged branches, you just merged in detached HEAD state	17:42
tango_	you are currently working out of any reference branch	17:42
j416	tango_: it's not?	17:42
tango_	oh wait	17:43
	didn't see the merge	17:43
	ok, I see the merge then	17:43
j416	afaik you can merge into detached head if you like	17:43
tango_	navatwo: you can push HEAD to master	17:43
	← sourcemaker left	17:43
tango_	true, true	17:43
j416	just as you can commit to detached HEAD	17:43
tango_	navatwo: you can do something like git push upstream HEAD:master but keep in mind that your merge is locally not attached to any branch, so it will be lost when you switch	17:43
j416	ah, that's it	17:43
	tango_: thanks, didn't spot that	17:43
	must be tired	17:44
	:)	17:44
tango_	it's hard to spot, took me a while	17:44
	← lirakis_afk left	17:44
tango_	navatwo: also, you merged cohen-master with upstream/master, while your push was to try to push master to upstream/master	17:44
	navatwo: so you're trying to do at least 3 different things	17:44
	→ gebi joined	17:45
	← gebi left	17:45
	→ gebi joined	17:45
j416	tango_: he did't merge it to upstream/master, he merged it into detached HEAD	17:45
	upstream/master remains unchanged	17:45
	j416 hints	17:45
tango_	yea, but the detached HEAD was a checkout of upstream/master. upstream/master will remain unchanged, but I was talking more about the actual codebase	17:45
	→ _pingu joined	17:45
j416	yep	17:45
	← yhager left	17:45
	→ merlin1991 joined	17:45
navatwo	So: $ git merge cohen-master upstream/master ?	17:46
j416	navatwo: so what you want to do now is likely, as tango_ says: git push upstream HEAD:master	17:46
	← mdpatrick left	17:46
j416	then checkout master and reset it to reflect the new upstream/master.	17:46
	→ tvw joined	17:46
navatwo	That did it..	17:47
j416	perfect. :)	17:47
navatwo	Hmm, how would I avoid detached then.	17:47
tango_	navatwo: but it's not going to leave any trace of what you did there	17:47
j416	navatwo: you should read about git object model and branching	17:47
tango_	navatwo: when you want to merge, you should want to merge on one of your local branches first, for example cohen-master	17:47
merlin1991	is there a way to "reverse lookup" the commit id of the workingtree, I have a tarball of the working tree in the past and the repository, now I want to know wich commit that tarball belongs to	17:47
tango_	navatwo: so you would git checkout cohen-master ; git merge upstream/master ; git push upstream cohen-master:master	17:47
	← _pingu left	17:48
navatwo	tango_: so, create local branch, merge it with the upstream and the remote, then push to the upstream?	17:48
	← drl left	17:48
tango_	navatwo: or even merge in one of your existing branches	17:49
	depends on what you want	17:49
navatwo	yeah	17:49
	depends who it is from, ofc.	17:49
	:)	17:49
tango_	if you don't want local history of the branch you CAN do that in a detached head	17:49
navatwo	Thanks tango_ and j416	17:49
j416	you're welcome	17:49
navatwo	tango_: I feel like thats a poor idea to get into a habit with.	17:49
tango_	navatwo: it is. and there is nothing wrong with having merges in local branches. it all depends on what you want	17:50
j416	navatwo: google to find how branches work, and how commits are chained together	17:50
	→ zombor joined	17:51
	→ mdpatrick joined	17:52
	→ pantsman joined	17:52
	← pantsman left	17:52
	→ pantsman joined	17:52
	← n3m left	17:52
	→ simosx joined	17:53
	← simosx left	17:53
	→ simosx joined	17:53
	← simplechat left	17:53
	← simosx left	17:53
	← dekebal left	17:54
	→ n3m joined	17:57
	→ lirakis joined	17:57
	← level09 left	17:58
	← bla left	17:59
	← madewokherd left	18:01
	→ drl joined	18:02
	→ posciak joined	18:02
	→ hobodave joined	18:03
	→ galexy joined	18:03
	← gilligan_ left	18:03
	← khmarbaise left	18:04
	← galexy left	18:04
	← ping-pong left	18:04
	→ madewokherd joined	18:05
	← madewokherd left	18:07
	← mdpatrick left	18:09
	→ madewokherd joined	18:09
	← merlin1991 left	18:09
	← tildeequals left	18:10
	← Maletor left	18:10
	← radarek left	18:10
	→ khmarbaise joined	18:10
	→ mdpatrick joined	18:10
	→ jaakkos joined	18:11
	← sivy left	18:12
	← EricInBNE left	18:13
	← drl left	18:14
jaakkos	probably offtopic but until i find a better channel, could someone give pointers for how to set up gitweb to use UNIX access control, as it is on the machine where the repo is hosted? the basic idea is to make gitweb access control identical to 'git over ssh'.	18:15
	→ dekebalus joined	18:18
	← pdelgallego left	18:18
	← niko left	18:18
bremner	jaakkos: gitweb access is controlled by unix permissions, but it can't be identical to git over ssh, because it is read only	18:19
	← jstein_ left	18:19
	→ bla_ joined	18:19
cbreak	isn't gitweb just a bunch of scripts run by apache?	18:20
	→ dtrott joined	18:20
cbreak	so they will probably use apache's unix permissions	18:20
bremner	one script, but yes	18:20
	koo5 → javaanse_jongens	18:21
	→ TanEk joined	18:22
	→ LongBeach joined	18:22
	→ coderdad joined	18:23
	→ niko joined	18:25
	← bluenovember left	18:26
	→ akosikeno joined	18:27
	→ drl joined	18:27
	→ joel_falcou joined	18:27
	→ muneeb joined	18:28
	→ chrischambers_ joined	18:30
	→ Bass10 joined	18:31
	← chrischambers left	18:31
	chrischambers_ → chrischambers	18:31
	→ pdelgallego joined	18:32
	← Bass10 left	18:32
	→ treak007 joined	18:32
	→ Bass10 joined	18:33
	→ radarek joined	18:34
	→ jvraoltc joined	18:35
	← kbedell left	18:35
dtrott	I want to re-write my local history so all commits to a particular file are at the end (call the file A).	18:38
	Normally just a simple interactive rebase, but some of my commits are to multiple files, is there a way I can easily break those commits so that all commits to (A) are at the end of the history ?	18:38
	← RaceCondition left	18:38
cbreak	no	18:38
bremner	dtrott: you can use "edit" within the interactive rebase	18:39
	→ si_keno joined	18:39
cbreak	there are a few non-easy ways	18:39
	→ xiong joined	18:39
dtrott	for example ?	18:40
cbreak	what bremner said	18:40
	or using git-filter-branch with a index or tree filter	18:40
	← speakman left	18:40
	→ speakman joined	18:40
	← radarek left	18:40
	→ RaceCondition joined	18:40
	← iravid left	18:41
	→ iravid joined	18:42
jaakkos	bremner, cbreak: ro should be ok, but indeed, what i'd like to achieve is that users must log in to gitweb with the SSH credentials and they see what they could see with git-over-ssh.	18:43
cbreak	impossible	18:43
	how would you provide an ssh public key to a web page?	18:43
	upload it?	18:43
jaakkos	i'm sure we would be ok supporting only password-based authentication	18:44
	for gitweb.	18:44
	i think this probably boils down to writing some custom authentication module to gitweb... it appeared to support something like that	18:45
cbreak	oh, right	18:45
	ssh supports passwords still too	18:45
jaakkos	cbreak: excuse me for my lack of RTFM, but i'm interpreting this like there is no authentication in gitweb at all unless i implement one?	18:46
cbreak	I have no idea.	18:46
	→ ping-pong joined	18:47
bremner	jaakkos: correct.	18:47
	→ __name__ joined	18:47
	← __name__ left	18:47
	→ __name__ joined	18:47
	→ JPT joined	18:47
	→ kuhrt joined	18:50
	→ wtfness joined	18:51
	← ChUrP left	18:51
	← jkp left	18:51
jaakkos	it's getting to the point where i should just try things myself, but if i protect git directories with, say, apache's htaccess, would that have the desired effect except it's htaccess instead of local UNIX credentials?	18:51
cbreak	no	18:52
	→ Buglouse joined	18:52
cbreak	what user do you think the gitweb script runs as?	18:52
TanEk	does someone knows what is the usage of commit->util?	18:54
	can I use it to store anything I want?	18:54
cbreak	what's that?	18:54
TanEk	in the git source	18:54
jaakkos	perhaps www-data or root, but i fail to see the connection between htaccess (working or not) and the user involved, as this works nicely with ordinary directory structures.	18:55
	→ Guest64225 joined	18:55
	→ yhager joined	18:56
	← Guest64225 left	18:56
tkil	jaakkos -- there are some efforts for doing fine-grained access control using just git. one of the main devs is usually in here, in fact. gitlosis or something?	18:57
cbreak	gitolite	18:58
	but that's via ssh	18:58
jaakkos	perhaps has to do with gitosis?	18:58
	← drl left	18:58
cbreak	it heavily relies on ssh public key authentication for user identification and authentication	18:58
tkil	ah. sorry for the misinformation.	18:58
	→ Curtman joined	18:59
	← wtfness left	18:59
cbreak	I think there were attempts to do alternate authentication, but I have no idea how far they are	18:59
jaakkos	perhaps it is still possible to protect the whole gitweb with some standard apache method... then i could, perhaps, set up multiple gitwebs	18:59
cbreak	https://github.com/sitaramc/gitolite	18:59
	→ rns joined	18:59
bremner	jaakkos: gitweb is cgi, no idea how that interacts with apache authentication.	19:00
cbreak	jakob: you can protect access to the script	19:00
	so that people who aren't authenticated to apache can't run the script.	19:00
	← JPT left	19:01
	→ kota joined	19:01
	→ maattd joined	19:01
jaakkos	cbreak: still, htaccess cannot be used for even that?	19:02
cbreak	no idea	19:02
	that's apache stuff	19:02
jaakkos	yeah, gotta try it out	19:03
cbreak	the problem with using .htaccess for a git repo is that you do not access the git repo	19:03
	you access the script	19:03
	the script runs git commands	19:03
	and they don't give a second thought about even considering .htaccess for anything	19:03
jaakkos	yep... intuitively i would say i have to protect the script directory	19:04
	→ guardian joined	19:04
	← kuhrt left	19:05
	→ ChrisBuchholz joined	19:07
	← Schmallon left	19:09
	← ddilinge1 left	19:10
	→ drl joined	19:11
	← ChrisBuchholz left	19:11
	← kumbayo_ left	19:11
	← joel_falcou left	19:11
bremner	jaakkos: maybe look at https://github.com/sitaramc/gitolite/blob/pu/doc/http-backend.mkd	19:12
	→ ddilinger joined	19:12
	← Rogue_Coder left	19:12
	← Buglouse left	19:12
	→ LosBorg joined	19:13
	← LosBorg left	19:14
	ChanServ set mode: +o	19:14
	wereHamster changed the topic to: 1.7.4.1 \| Homepage: http://git-scm.com \| Can't talk on the channel? http://jk.gs/gitvoice/ \| Everyone asleep or clueless? Try [email@hidden.address] \| Channel log http://j.mp/gitlog \| Mailing list archives: http://j.mp/gitlist \| Gits on git: http://j.mp/gittalks \| Pastebin: http://gist.github.com/ \| http://tinyurl.com/GitSurvey2010Results	19:14
	ChanServ set mode: -o	19:14
	→ jkp joined	19:15
	→ wtfness joined	19:16
jaakkos	bremner: thanks	19:16
	→ Buglouse joined	19:17
	→ LosBorg joined	19:17
	→ radarek joined	19:17
	← notze left	19:18
	← bradleyayers left	19:18
	← eMBee left	19:22
	← dKingston left	19:22
	← tcv_br left	19:23
	→ msierks joined	19:24
	← tvw left	19:24
	→ deryl joined	19:24
	→ eMBee joined	19:25
msierks	would someone be able to assist me in setting up a remote git repository with webdav/apache ?	19:25
deryl	I have a problem. I forgot to set git config user.name and user.email so my localhost information was used instead. I thought, to correct the info, I did a git commit -C <commit_hash> --amend --author="My Name [email@hidden.address] && git push but that appears to be incorrect	19:25
	→ joel_falcou joined	19:26
deryl	What am i doing incorrect?	19:26
cbreak	deryl: that doesn't work this way	19:26
	--amend is only for the last commit	19:26
	you want to use git filter-branch --env-filter	19:26
	or if you like manual labor, git rebase -i	19:26
deryl	ok, let me man that	19:26
	→ si_keno_to joined	19:26
deryl	hehe no, I'm really weak on my git-fu so the less I have to touch at the moment w/o screwing my repo the better	19:27
	← si_keno_to left	19:27
	→ si_keno_to joined	19:27
	→ gnoel joined	19:28
	← marc_kdab_ left	19:28
	→ kbedell joined	19:28
	→ amerine joined	19:29
deryl	ouch. looks like this won't just modify the user.email portion. it will have completely different object IDs	19:29
cbreak	of course	19:30
	because the commit hashes contain everything	19:30
	they are a hash of the commit object	19:30
	which contains the user ID	19:30
deryl	ahh ok. makes sense	19:30
cbreak	if you find a way to change the content of a commit without changing it's hash	19:30
	tell me	19:31
deryl	hehe	19:31
cbreak	so I can get a lot of money :)	19:31
deryl	share?	19:31
cbreak	80:20	19:31
	← drl left	19:31
deryl	70:30	19:31
cbreak	ok.	19:31
deryl	hey if you make a mil, I get 300G I can wipe out all my debts with that! :)	19:31
	divorces kill ya ;)	19:32
cbreak	that's human interaction.	19:32
deryl	hehe	19:32
	← towski left	19:32
cbreak	become a nerd and never leave your room. cheaper.	19:32
deryl	hah! think I've done that already	19:32
	direct response to divorcing the ol bat	19:33
	← eMBee left	19:33
	→ eMBee joined	19:33
deryl	ok, I mean its my onw repo so i can leave it as it is. Just don't like exposure of the home information like that. But I've since fixed the user.name and user.email stuff so all future commits will be correct.	19:34
	thanks for the help cbreak	19:34
	← radarek left	19:34
cbreak	rewriting history is not something uncommon with git	19:34
	← srji left	19:34
cbreak	so in your situation, I'd do a filter-branch	19:34
	well, unless the repository is already public	19:34
deryl	looks a tad bit off-putting. (ok scary is the word)	19:34
cbreak	then you should not change the past	19:35
deryl	yeah the repo is on github	19:35
	→ sivy joined	19:35
	→ sreich joined	19:36
	→ mikepack joined	19:36
	← nonnooo left	19:36
	→ gdmfsob joined	19:37
	← envi left	19:38
	→ marc_kdab_ joined	19:38
	← mishok13 left	19:39
	← g0bl1n left	19:40
	→ iband joined	19:42
Curtman	Sorry for the dumb noob question.. When I create a new file, and git-add it.. Why doesn't it exist when I git-diff?	19:45
	→ drl joined	19:45
c00kiemon5ter	cause it has been indexed	19:46
bremner	Curtman: git diff --cached	19:46
c00kiemon5ter	use 'git diff --cached'	19:46
Curtman	Thanks..	19:46
	→ salty-horse joined	19:46
salty-horse	hey. is there a way to maintain a commit/object with some local changes that you never want pushed, but want to exist on every branch? maybe an extension someone made?	19:47
wereHamster	salty-horse: no.	19:48
TanEk	someone knows how to build git with debug info? make -g?	19:49
	← dyer left	19:50
	← drl left	19:50
	→ emptyflask joined	19:51
	→ tgunr joined	19:52
	← bentob0x left	19:55
	← bla_ left	19:56
	→ torvalamo joined	20:01
	← yhager left	20:01
	← gusnan left	20:02
	→ towski joined	20:03
c00kiemon5ter	TanEk: -g needs to passed to gcc	20:03
	→ hollyw00d joined	20:03
	← towski left	20:04
c00kiemon5ter	I havent build git from source, if it comes with a configure file there may be an option to enable it there	20:04
	← khaije\|minder left	20:05
c00kiemon5ter	if not you can either edit the makefile or if the makefile reads environmental variables do something like: CC="-g" make	20:05
	→ ben225 joined	20:06
	← _Vi left	20:07
cmn	the Makefile passes -g to the computer by default	20:07
TanEk	yes	20:07
	I just didn't know gdb was not giving the name in the stack by default	20:07
cmn	c00kiemon5ter: CC overrides the compiler binary, you mean CFLAGS	20:07
	← goshawk left	20:08
TanEk	so I thought the debugging infos were not generated by default, but actually they are	20:08
	→ acolin joined	20:09
	← acolin left	20:09
	→ bberg joined	20:10
	← treak007 left	20:10
c00kiemon5ter	cmn: depends on how the makefile is build	20:11
	→ drl joined	20:11
wereHamster	CC == compiler is pretty much the standard ...	20:11
c00kiemon5ter	I've seen makefiles with no CFLAGS, and makefiles that do 'gcc ${CC} ${defaults}'	20:11
	← saste left	20:11
	← msierks left	20:12
	← urkud_ left	20:13
	← ben225 left	20:14
	→ urkud_ joined	20:14
	→ madsy joined	20:15
	→ salty_horse joined	20:16
	← drl left	20:18
	← salty-horse left	20:19
	→ Heimidal joined	20:20
	← Yuffster left	20:21
	← mastro left	20:21
	← amerine left	20:22
	← no_maam left	20:22
	← HG` left	20:23
	→ HG` joined	20:24
	→ kuhrt joined	20:26
	→ Saur joined	20:30
	← Saur left	20:31
	→ m4co joined	20:31
	← anderse left	20:31
	→ drl joined	20:31
m4co	what is the difference of fetch vs pull? why would I use fetch?	20:31
wereHamster	m4co: pull = fetch + merge	20:32
	and you would use fetch if you only want to fetch but not merge	20:32
m4co	but if I'm fetching for a branch other than master	20:33
	← doener left	20:33
m4co	from*	20:33
	isn't that the same as pull?	20:33
wereHamster	no	20:33
m4co	actually can I pull from another branch into master, as in merging?	20:33
	sorry I'm a bit confused here	20:33
	→ ben225 joined	20:34
m4co	imagine the following: git pull origin another_branch <- won't that just fetch and merge another_branch changes?	20:34
wereHamster	you can use 'git pull . <branch>' instead of 'git merge <branch>', but I see no reason to ever use that	20:34
	m4co: yes	20:34
m4co	ok and if I do git fetch another_branch	20:34
	is that the same thing?	20:35
wereHamster	that will give you an error	20:35
	It's either 'git fetch origin' or 'git fetch origin another_branch'	20:35
m4co	and git fetch origin another_branch is a "temporary" pull until I merge?	20:35
wereHamster	no. fetch does not merge	20:36
	fetch.. uhm.. only fetches	20:36
m4co	that's what I meant, I'd have to merge later right?	20:36
wereHamster	if you want to do so.. yes	20:36
m4co	most likely fetch is used to test things?	20:36
	→ mastro joined	20:36
	→ Maletor joined	20:36
m4co	it's just when I think of "merge" that always means two branches	20:37
wereHamster	no. Fetch is used if you want to fetch stuff, but not use that stuff right away. Maybe later you want that but not right now	20:37
	← Maletor left	20:37
m4co	but how can I use them? I'd have to merge?	20:37
wereHamster	or checkout	20:37
m4co	github was explaining another way to handle pull requests now they changed it and I'm all lost	20:37
	before it was a matter of checkout out in a new branch, test it, if it works merge to master	20:38
	← bberg left	20:38
wereHamster	I'm pretty sure github has a step-by-step guide how to handle pull requests. So just follow that	20:38
m4co	now it's add origin, fetch (??) merge into that branch and then merge into master	20:38
	I've done the "fetch" part but I can't see the changes	20:38
wereHamster	where are you looking?	20:39
m4co	http://help.github.com/pull-requests/	20:39
wereHamster	I mean for the changes	20:39
m4co	"Merging a pull request"	20:39
	→ yogione joined	20:39
c00kiemon5ter	it's in the far end	20:39
	it's still the same thing	20:39
m4co	I want to create a branch with the changes, look and then merge into master...so much simple that way	20:40
wereHamster	m4co: and what is it that you don't understand?	20:40
	← drl left	20:40
m4co	what I don't understand is why I can't see the changes	20:40
	with this "fetch" method	20:40
wereHamster	m4co: it's better to merge first and then look at the result	20:40
c00kiemon5ter	cause you havent merged yet	20:40
wereHamster	m4co: how did you name the remote?	20:40
m4co	but merge what? the commit?	20:40
c00kiemon5ter	the fetch	20:40
m4co	you mean merge the fetch into master?	20:40
c00kiemon5ter	yes, that's what he does in the example	20:41
m4co	why would I merge into master instead of it's own branch?	20:41
wereHamster	m4co: just follow the example in the help page and replace names/urls where appropriate	20:41
m4co	I mean, what's the benefit..I want to test it before merging	20:41
wereHamster	m4co: no. test after merging	20:41
	salty_horse → salty-horse	20:41
c00kiemon5ter	m4co: with test, you mean create another branch and merge there ?	20:42
m4co	c00kiemon5ter, yes, makes more sense to me to create a branch with the fork to test it	20:42
	the help was explaining that way before	20:43
	→ Mage joined	20:43
c00kiemon5ter	so, git checkout master # go to the master branch	20:43
m4co	now it's saying to fetch and merge into master and then test it, why not test it in a different branch and if it goes well merge into master?	20:43
c00kiemon5ter	git checkout -b mergetest # create another branch and checkout there	20:43
wereHamster	m4co: of course you can do that	20:43
c00kiemon5ter	that branch will be used to merge and test	20:43
m4co	c00kiemon5ter, go on..	20:44
c00kiemon5ter	git pull git://github.com/foo/bar/that	20:44
wereHamster	m4co: but there's no reason to create the branch if all you do is merge + test + maybe push if test passes	20:44
c00kiemon5ter	^ get the sources you want to merge, fetch them and merge them in one step	20:44
	^^^^ this can be replaced with	20:45
	^^^^ git fetch git://github.com/foo/bar/that	20:45
	^^^^ git merge bar/that	20:45
	so at this point you've created a new branch, and merged there the remote sources	20:45
	if it's all good and well you can checkout to master and merge mergetest branch	20:46
wereHamster	c00kiemon5ter: at no point did your commands create a new branch	20:46
tkil	how do i ask git-svn which local commit matches a remote revision? i see it in the log output, but i don't see how to get at it later...	20:46
c00kiemon5ter	if not, resolve w/e conflicts and continue	20:46
	← HG` left	20:46
	→ Maletor joined	20:46
c00kiemon5ter	wereHamster: git checkout -b mergetest # that's a new branch	20:46
	→ GNUtoo\|laptop joined	20:47
wereHamster	tkil: git log --grep=@<your revision>	20:47
m4co	I think I'm starting to understand better	20:47
GNUtoo\|laptop	hi, is there a --subject-prefix=	20:47
wereHamster	c00kiemon5ter: oh, I missed that	20:47
GNUtoo\|laptop	option	20:47
	I don't have it	20:47
tkil	wereHamster -- thanks, will try that.	20:47
GNUtoo\|laptop	I mean I sent a mail with that option and the [PATCHv2] didn't made it to the message subject	20:47
m4co	wereHamster, why there's not reason to create a branch for the fork?	20:48
	← Maletor left	20:48
GNUtoo\|laptop	I don't want to do the same mistake again	20:48
	how can I fix that?	20:48
m4co	What I don't understand is the reason to fetch into master	20:48
	← Samual_ left	20:48
	→ utu joined	20:48
tkil	wereHamster -- spiffy, thanks. i had to not use the "r" prefix (as is traditional for svn revs) but other than that, it worked perfectly. thanks!	20:48
m4co	also, instead of "fetch" and then "merge" can't I just "pull git://github.com/foo/bar" instead ?	20:49
	into master?	20:49
c00kiemon5ter	m4co: fetch followed by a merge is the same as pull, when you fetch you create a temp branch with the fetched sources, in github's exxample that branch is named kneath/error-page	20:50
	→ Yuffster joined	20:50
m4co	ooohhhh when I fetch a new branch is created?	20:50
	← rajeshsr_ left	20:50
m4co	and is it automatically checked out there?	20:50
c00kiemon5ter	m4co: http://sprunge.us/DZMS	20:50
	that's the same thing, the github example is doing	20:50
GNUtoo\|laptop	I'll refrase	20:51
	hi, I tried the --subject-prefix= but it seem not to exist on my git version, how can I emulate it?	20:51
	I've git version 1.7.0.4	20:52
m4co	ok, now it makes more sense	20:52
	I guess the remote is good for a quick fetch <remote> or pull <remote> instead of full URL	20:52
	however, if it's the same thing I fail to understand the use of fetch	20:52
	if it won't merge and I can't see the changes, what is it for?	20:53
wereHamster	m4co: there is no reason to create a branch just so you can merge and test the changes in the pull request	20:53
	→ drl joined	20:53
	← jkp left	20:53
c00kiemon5ter	yes, it aliases the url to a name you specify, so you can easily refetch repull or w/e that remote's sources	20:53
m4co	wereHamster, but I won't merge before testing in that branch	20:53
	for example, I create a branch, pull, test, and then merge into master	20:53
wereHamster	m4co: why not pull directly into master?	20:54
m4co	to keep things organized, shouldn't the forks have it's own branches until it merges into master?	20:54
	→ tildeequals joined	20:54
	← gnoel left	20:54
wereHamster	m4co: the forks do have their own branches. But if you get a pull request all you want to do is merge the changes into master, right?	20:55
m4co	test them, then merge	20:55
wereHamster	no. merge first then test	20:55
c00kiemon5ter	if you plan merging into master then you just pull/fetch+merge into master. you dont have to create a branch to test the merge, but you can if you want.	20:55
m4co	test into their branch	20:55
c00kiemon5ter	wereHamster: he means branch, merge, test. then merge to master	20:55
wereHamster	m4co: why do you insist on testing in a separate branch? There's absolutely no reason for that	20:56
	← babbage left	20:56
m4co	I must be misunderstanding the branch concept then	20:56
	to me, all forks have their own branches	20:56
	I pull that fork into a branch then merge into master	20:56
c00kiemon5ter	:P the only reason for that is if the merge has conflicts and you accidentally reset --hard to a wrong hash, you loose some of your work. that's an accident, so just to avoid accidents you may want to make a new branch for the merge. cant thiink of another reason	20:57
	→ Maletor joined	20:57
	→ project2501a joined	20:57
m4co	ok say I merge directly into master	20:58
	then I test, something's wrong	20:58
wereHamster	c00kiemon5ter: you shuoldn't reset --hard. There's reset --merge. And why would you use a hash directly and not ORIG_HEAD or HEAD@{1} ?	20:58
m4co	I just checkout to undo?	20:58
wereHamster	or.. just HEAD if there were conflicts	20:58
c00kiemon5ter	wereHamster: dunno, :)	20:58
wereHamster	m4co: you reset --hard ORIG_HEAD	20:58
	← drl left	20:59
	← nevyn^1 left	20:59
m4co	see, if I had a branch for it to test the fork I could simply just not merge it :)	20:59
	→ nevyn^ joined	20:59
m4co	does my example make sense?	20:59
wereHamster	what does that buy you? Instead of 'git reset --hard ORIG_HEAD' you'll have to do 'git checkout master && git branch -D <the test branch>'	20:59
c00kiemon5ter	yes, but with that logic one would create a new branch for every commit he does	20:59
m4co	isn't checkout master && branch -D "safer" ? or just easier?	21:00
wereHamster	no, it's not safer at all	21:00
m4co	ok now I think I understand	21:01
	but how about the fetch then, why fetch instead of pull if I have to merge after fetching anyway?	21:01
	you mention it's good in case I don't want to merge it yet, right?	21:01
	→ tcv_br joined	21:01
wereHamster	m4co: ask the github staff why they chose this way	21:01
Ulfalizer	i wonder if 'git status' and other operations could be made much faster if the fs had a special timestamp for directories that changed with the contents of files as opposed to just when creating/removing files. that way it would only have to check a single timestamp.	21:01
m4co	but then how can I see the changes? should I checkout to a different branch?	21:01
Ulfalizer	it would be recursive, of course	21:02
c00kiemon5ter	m4co: what merge does is it creates a new commit after all files and changesets are there. in order to go back to a previous state in your tree and get rid of the new changes you would git reset to that state.	21:02
Ulfalizer	probably wouldn't add that much overhead either if you just enabled it for a few directories	21:02
m4co	c00kiemon5ter, and fetch doesn't create a new commit?	21:02
c00kiemon5ter	fetch no	21:02
m4co	fair enough, and does it create a branch?	21:03
c00kiemon5ter	fetch doesnt change anything to your code	21:03
	merge does	21:03
wereHamster	m4co: it does if you tell it to	21:03
m4co	a temp branch	21:03
cmn	m4co: if you just want to see the difference you can use git diff master fork/branch	21:03
	they're not temproray in the usual sense, if there are published branches in the remote, they will be created	21:03
Ulfalizer	'git status' is sometimes annoyingly slow at work :/	21:03
m4co	c00kiemon5ter, but if it doesn't change anything what is the point, how can I test it? that's what I don't understand	21:03
wereHamster	Ulfalizer: a large working tree?	21:03
	→ mcmat joined	21:04
	← MikhX left	21:04
	← salty-horse left	21:04
Ulfalizer	wereHamster: yes	21:04
wereHamster	m4co: you merge whatever you fetched into master, then test	21:04
	← pdelgallego left	21:04
wereHamster	Ulfalizer: make it smaller..	21:04
bremner	m4co: I think the concept you are missing is http://www.gitready.com/beginner/2009/03/09/remote-tracking-branches.html	21:04
m4co	but that's pulling so why the extra step	21:04
wereHamster	m4co: I can't read minds. Ask the github staff	21:04
c00kiemon5ter	m4co: fetch will "fetch" the remote sources. it wont do a thing more. after you've fetched as cmn said you can either git diff the branches or merge, and see how that goes	21:04
	→ pdelgallego joined	21:05
c00kiemon5ter	merge will give you a new commit so you can git diff that	21:05
	or git show probably	21:05
Ulfalizer	wereHamster: i don't think it's much larger than it would have to be, if i'm getting you right	21:05
bremner	m4co: fetching updates your local mirror of the remote repo	21:05
m4co	ok if I can diff the branches it makes more sense	21:05
Ulfalizer	it's a big project	21:05
wereHamster	Ulfalizer: then you have a slow harddisk. Get a ssd	21:05
Ulfalizer	...	21:06
linusoleander	Is it somehow possible to jump to a specific commit?	21:06
	I've used git reset --hard, is that the way to go?	21:06
	I want to be able to jump back	21:06
wereHamster	linusoleander: git checkout <sha>	21:06
m4co	cmn, so I can git diff master fork/branch after fetching fork/branch ?	21:06
wereHamster	linusoleander: don't use reset for that purpose	21:06
linusoleander	wereHamster: Perfect!	21:06
	Thanks!	21:06
cmn	m4co: you usually do fetch on a remote	21:06
m4co	yea	21:06
cmn	but once you've fetched and the branch is there	21:07
	then you can diff between them	21:07
	use git branch -a to show remote branches	21:07
m4co	oh we have remote branches too	21:07
cmn	yeah, they're very usefull	21:07
m4co	so the fetch would create a remote branch?	21:07
c00kiemon5ter	:P :D	21:07
	→ jmorris joined	21:07
cmn	yeah, fetch will create or update the remote branches	21:08
m4co	fetching created a remote branch, yes?	21:08
	aha :D	21:08
	← Maletor left	21:08
m4co	imagine my confusion when I didn't know about remote branches	21:08
	made no sense	21:08
	now it does :P	21:08
	I was mixing local branch with remote branch	21:09
cmn	git is distributed, so you can have branches from many places	21:09
m4co	having a local branch to pull if you can just fetch a remote branch makes no sense indeed	21:09
	← Zhonghua left	21:09
	→ webPragmatist joined	21:09
webPragmatist	how do i checkout a specific revision of a file?	21:09
wereHamster	webPragmatist: and put it where?	21:10
m4co	thanks guys wereHamster cmn c00kiemon5ter	21:10
mcmat	hi. I don't really understand git-remote. It seems to me that doc and reality are different: i clone the linux git repo (git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git linux-2.6.git), then add a remote (git remote add --tags drm-intel git://git.kernel.org/pub/scm/linux/kernel/git/ickle/drm-intel.git). Then i do git fetch, but nothing happens, and git branch -r still only show me origin/master.	21:10
m4co	and thanks bremner	21:10
wereHamster	mcmat: git fetch drm-intel	21:11
	→ drl joined	21:11
wereHamster	mcmat: or 'git remote update' to fetch from all remotes	21:11
mcmat	wereHamster: i also tried but same result (or lack of)	21:11
	← _iron left	21:11
mcmat	werehamster: i tried git fetch drm-intel, but not git remote update	21:11
	i try	21:12
GNUtoo\|laptop	I'll refrase a third time	21:12
cmn	mcmat: if you just fetch, your local working copy won't be modified (look at what we just talked about with m4co)	21:12
	you need to merge	21:12
GNUtoo\|laptop	how can I send a patch with [PATCHv2] in the subject line?	21:12
wereHamster	cmn: that's not what he wants	21:12
m4co	I can checkout to a remote branch too right?	21:13
wereHamster	GNUtoo\|laptop: --annotate	21:13
	m4co: it'll detach your HEAD	21:13
	← jmorris left	21:13
m4co	what does that mean?	21:14
	← OOPMan left	21:14
m4co	it would be temporary right? like a different HEAD (the remote branch)?	21:14
	← priidu left	21:14
GNUtoo\|laptop	wereHamster, ok thanks a lot	21:14
wereHamster	m4co: faq detached. Or man gitglossary -> detached head	21:14
jast	m4co: the 'gitglossary' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/gitglossary.html [automatic message]	21:14
webPragmatist	wereHamster: overwrite the smae file	21:14
	wereHamster: theoretically just reverting a single file	21:15
wereHamster	webPragmatist: git checkout <rev> -- <path>	21:15
webPragmatist	whats -- do btw?	21:15
wereHamster	webPragmatist: it tells git to parse <path> as a path and not a revision	21:15
webPragmatist	oh i see	21:16
	i thougth it metnt "back" or minus one	21:16
wereHamster	it's not required usually but use it anyway	21:16
	→ priidu joined	21:16
	← priidu left	21:16
mcmat	wereHamster: tried git remote update. It tells me "Fetching origin, Fetching drm-intel", but still, git branch -r lists only origin/HEAD and origin/master. drm-intel branches do not appear (though they exist, since i can see them with git remote -v show drm-intel	21:17
Ulfalizer	conventionally -- as an argument to a unix command marks the end of the "-x"-style flag arguments, to prevent what follows it to be interpreted as such a flag	21:18
	→ _Vi joined	21:18
Ulfalizer	handy e.g. if the command accepts a file and you happen to have a file whose name would otherwise be interpreted as a flag	21:18
	i guess git might use it in the sense of "what follows is a path, so do not attempt to interpret it as anything else"	21:21
	→ Garii joined	21:21
	← sneakret left	21:22
Ulfalizer	*from being interpreted as	21:22
	← project2501a left	21:22
	← kbedell left	21:23
	← m4co left	21:26
	← tkil left	21:26
	Ulfalizer wonders why git has moved a bunch of heads from refs/heads to logs/refs/heads	21:26
wereHamster	Ulfalizer: logs are the reflogs	21:27
	← iasterisk left	21:27
	→ Maletor joined	21:30
Ulfalizer	wereHamster: why did git remove the refs from refs/heads though? 'git branch' lists branches that have reflogs in logs/refs/heads but no ref in refs/heads.	21:31
wereHamster	Ulfalizer: packed refs probably	21:32
	Ulfalizer: don't ever look into the .git directory if you don't understand the internals...	21:33
Ulfalizer	just curious	21:34
	is refs/heads/<foo> to be understod as a symbolic notation in 'git push <remote> branch:refs/heads/<foo>' then? i mean, if no refs/heads/<foo> is actually guaranteed to exist for the branch..	21:35
	← marc_kdab_ left	21:35
Ulfalizer	+o	21:35
	← dekebalus left	21:35
	← guardian left	21:37
	→ MikhX joined	21:38
wereHamster	eh?	21:38
	← Heimidal left	21:39
wereHamster	how/where git stores the refs is an internal implementation detail. If you use the git commands (even plumbing commants) you don't have to worry about these details	21:39
	→ Samual joined	21:40
Ulfalizer	so git makes sure the right thing happens even if the remote has packed the refs and so has no ref for the branch in refs/heads?	21:40
wereHamster	of course	21:40
	even if the refs are packed they haev names such as refs/heads/foo	21:41
	so 'the remote has packed the refs and so has no ref for the branch in refs/heads' does not make sense	21:42
Ulfalizer	is that a special case for refs/heads, or would git handle any path the same?	21:42
	→ s0ber_ joined	21:42
Ulfalizer	i'm a bit unsure how other paths would interact with porcelain commands..	21:42
wereHamster	which paths?	21:42
Ulfalizer	substituting foo/bar/baz for refs/heads	21:43
wereHamster	uhm.. is one a pathspec and the other one a ref?	21:43
Ulfalizer	but perhaps you can't put arbitrary paths	21:43
wereHamster	or are both refs?	21:43
	← s0ber left	21:43
	s0ber_ → s0ber	21:43
	→ priidu joined	21:44
Ulfalizer	i was thinking something along the lines of 'git push origin <branch>:foo/bar/baz'	21:44
	what would that do? is it guaranteed to create the file .git/foo/bar/baz on the remote, or is it just a symbolic notation? is it even allowed?	21:44
wereHamster	try and see	21:45
Ulfalizer	don't think i have any remote handy to experiment with :/	21:45
	at least not one i should experiment with	21:45
wereHamster	git init --bare /tmp/ulfalizer; git push /tmp/ulfalizer master:foob/ar/bz	21:45
Ulfalizer	guess i could set one up, unless my laziness wins out	21:45
wereHamster	then cd /tmp/ulfalizer && git show-ref	21:46
	← pantsman left	21:50
	← mikepack left	21:50
	→ radarek joined	21:52
	← tildeequals left	21:52
	← Zephyrus left	21:53
	← priidu left	21:53
Ulfalizer	wereHamster: is it best to view the contents of the .git/refs directory as implementation details that often mirror what e.g. git-show-ref would show you but need not do so?	21:53
	i.e., refs/heads/foo might not actually be a file in .git/refs/heads/foo	21:54
wereHamster	Ulfalizer: it's best to never look at it directly and instaed use git for-each-ref	21:54
	the commands exist for a reason	21:54
	and the reason is not because someone was bored...	21:54
	← DiGitalX left	21:55
Ulfalizer	that's probably something that might trip up a lot of beginners who might expect refs/heads/foo to be a filesystem path with no special interpretation	21:55
	heck, it got me confused :)	21:55
	← Bartman017 left	21:55
	→ priidu joined	21:56
wereHamster	beginners are not encouraged to look into the .git directory. And if they do and don't understand what they are seeing it's their problem, not ours	21:56
Ulfalizer	wereHamster: i hope the tutorials/manual makes that clear :/	21:57
	← priidu left	21:57
wereHamster	no tutorial I know of tries to describe the .git directory contents	21:57
	→ priidu joined	21:58
Ulfalizer	i think i read one. or at least one who phrased it along the lines of "heads are stored as files in the .git/refs/heads directory"	21:58
	← posciak left	21:58
Ulfalizer	but perhaps that was just a bad tutorial	21:59
wereHamster	then it's wrong and you should tell its author	21:59
	← sudonotplease left	22:01
	→ sudonotplease joined	22:01
Ulfalizer	are there any guarantees whatsoever about the structure of .git/? no - not going to depend on it, just curious :)	22:02
	→ Silowyi joined	22:04
	← d0k left	22:05
JoeHazzers	my git hooks don't seem to be running properly... i.e. they don't seem to be running at all after i push to them. what file should i put hooks in for after pushing to it from a remote?	22:05
	→ notze joined	22:07
	→ jkp joined	22:08
	→ DiGitalX joined	22:08
	← nanotube left	22:10
	← pedrobelo left	22:13
	← j416 left	22:15
	← jkp left	22:16
	→ armut joined	22:16
armut	hi, i need help	22:16
	← LongBeach left	22:17
NfNitLoop	armut: What's your problem? (as a side note: Don't ask to ask. See: http://jeff.jones.be/technology/articles/how-to-ask-for-help-on-irc/ )	22:18
armut	i deleted a directory, i want to restore it	22:18
NfNitLoop	git checkout directoryName.	22:18
	→ robotarmy joined	22:19
armut	git error: pathspec 'dino' did not match any file(s) known to git.	22:19
NfNitLoop	had you checked in 'dino' previously?	22:19
armut	i deleted as: git -rf dino	22:19
	no	22:19
NfNitLoop	then how can git restore it?	22:20
armut	just added	22:20
NfNitLoop	aah.	22:20
	Hmmm.	22:20
armut	git add dino	22:20
	→ jstein_ joined	22:20
NfNitLoop	I think you may be out of luck.	22:20
armut	:)	22:20
NfNitLoop	since you never committed it, the files didn't get saved in history.	22:20
	they WERE saved in your index.	22:20
	but then you removed them with 'git rm'.	22:20
armut	yes never commited	22:20
NfNitLoop	You might be able to find them in your object cache? I dunno. Anyone more knowledgeable than I on this paying attention? :)	22:21
	← webPragmatist left	22:21
armut	weher can i find cache?	22:21
NfNitLoop	it's in .git, but I don't know enough about how the index works to give you more info.	22:23
	waite in here for a while and see if someone else can help. :)	22:23
	← Leonidas left	22:23
armut	ok thanks	22:24
NfNitLoop	wait, even. I don't know why I thought it needed an E. :)	22:24
	← DiGitalX left	22:26
	← joel_falcou left	22:26
patrikf	armut: fsck can show you all objects that have no active reference, yours will be among them	22:26
	(git fsck, that is)	22:26
	it might take some time and/or grepping to find it, though	22:27
NfNitLoop	unfortunately, those objects probably don't include useful things like... file names or paths. IIRC.	22:27
patrikf	yeah.	22:27
	← bgerber left	22:27
	← utu left	22:28
patrikf	oh, you're looking for a whole directory, right.	22:29
armut	i get fsck list, blobs and trees... i think trees are directories	22:29
patrikf	AFAIK, if you add to the index, only blobs are created, but not trees	22:29
	(trees just reference blobs together with a filename)	22:29
	you might be able to filter out the ones you want with the mtimes of the objects in .git/objects	22:30
	→ bgerber joined	22:30
	← mcmat left	22:31
	→ Heimidal joined	22:32
	ChanServ set mode: +v	22:32
jstein_	a friend made his own branch how can i merge it to master again? https://github.com/jonasstein/fuzzer/compare/master...Branch_53d39f3cd686c382f383a3f3f5bb2e533120a49f#diff-2	22:32
	→ HG` joined	22:33
	→ TheBuzzSaw joined	22:34
armut	objects like a maze !! :)	22:34
cbreak	jstein_: with merge	22:34
	fetch it into your repo, then just merge normally	22:34
	← GNUtoo\|laptop left	22:34
jstein_	git merge master foo?	22:34
cbreak	git merge foo, while on master	22:34
jstein_	fetch, or pull?	22:34
cbreak	fetch	22:34
	pull is fetch and merge	22:35
jstein_	then pull is better now?	22:35
cbreak	well... I guess you could just pull it...	22:35
jstein_	ok	22:35
cbreak	but if you fetch and then merge it	22:35
jstein_	thx	22:35
cbreak	you know what you're doing	22:35
jstein_	not really ;-)	22:35
	← gebi left	22:35
	← coderdad left	22:36
	← __name__ left	22:38
	← notze left	22:38
	← _ollie left	22:42
	→ nanotube joined	22:42
	← ben225 left	22:43
	← sternocera left	22:44
	→ ben225 joined	22:45
	← HG` left	22:50
	← TheBuzzSaw left	22:51
jstein_	cbreak: it worked thanks. but now i have a HEAD line in the source. i think a good editor would manage that merge automatic, right?	22:52
cbreak	head line?	22:52
jstein_	i will try it with emacs magit	22:52
cbreak	if there's a conflict, you have to resolve it	22:53
jstein_	https://github.com/jonasstein/fuzzer/blob/master/canvas.py#L88	22:53
cbreak	it's not the editor that has to do that, it's you	22:53
	where?	22:53
jstein_	but the conflict is so easy i think the editor could autoguess	22:53
cbreak	guessing is a terrible idea	22:54
	because guesses can be wrong.	22:54
	it's much better to let humans deal with conflicts	22:54
	but I don't see a conflict in there	22:54
jstein_	uh sorry my friend just pushed a version without HEAD	22:54
cbreak	honestly	22:54
	don't commit stuff that have conflicts :)	22:54
	resolve them	22:54
	THEN commit	22:55
	← tweakism left	22:57
	← rns left	22:57
	← Unknown[NF] left	22:59
	← tg left	22:59
	← jvraoltc left	22:59
	← priidu left	23:00
	→ tg joined	23:00
	← drl left	23:01
	→ bousquet joined	23:02
jstein_	how	23:02
	i will need 2 local repository right?	23:02
cbreak	why?	23:03
jstein_	one with my data and one for merging?	23:03
cbreak	why?	23:03
	I don't get what you say	23:03
jstein_	i suppose even after months with git. the git system is far beyond my understanding	23:04
	← phantomcircuit left	23:04
	→ phantomcircuit_ joined	23:04
	← phantomcircuit_ left	23:04
	→ phantomcircuit joined	23:04
	← ketas left	23:06
	← ketas-l left	23:06
	→ ghhhhhhh joined	23:07
	→ blaenk joined	23:09
	→ ketas joined	23:09
	→ rns joined	23:10
	← ghhhhhhh left	23:10
	← ThomasLocke left	23:10
	→ arberber joined	23:11
	→ boombatower joined	23:11
	← boombatower left	23:11
	→ boombatower joined	23:11
	ChanServ set mode: +v	23:11
	← Mpenz left	23:11
jstein_	i simulate my git upload and you correct me if i am wrong please	23:11
	git clone git:....	23:12
	→ Mpenz joined	23:12
jstein_	hackhack	23:12
arberber	Suppose I have two machines on a network. Machine A has a git repo, and I have a login on this machine. Nonetheless, I want to work on machine B. Is there a way to just clone it over the network without having to use some sort of git server?	23:12
jstein_	git commit -a -m "worked"	23:12
cbreak	arberber: yeah.	23:12
	if you have ssh access	23:12
jstein_	arberber: ssh	23:12
	arberber: or just by filesystem	23:12
	← alama left	23:13
jstein_	git merge origin master	23:13
arberber	er, sorry	23:13
jstein_	git push	23:13
	fini. is that right?	23:13
cbreak	no	23:13
	why merge?	23:13
arberber	what i mean to say is, john has it in his home directory.	23:13
jstein_	arberber: you can do things like git clone /data/project /mywork	23:14
arberber	on A. I, phil, also have a login on A. I want to clone _from john_ to machine B	23:14
jstein_	like copy command	23:14
cbreak	arberber: do you have ssh access?	23:14
	→ drl joined	23:14
arberber	yes, i do have ssh access to machine A	23:14
	but of course, i'm not john	23:14
cbreak	then git clone ssh://server/../john/path/to/repo.git	23:15
arberber	hmm	23:15
cbreak	assuming that your home directory is inside the same folder as john's	23:15
	and that you have read access to it	23:15
arberber	my home directory is inside the same folder as john's? what do you mean by that?	23:15
cbreak	you have a home directory, right?	23:16
Ulfalizer	wereHamster: re. the .git discussion, even the git manpage says things like "<tag> a valid tag name (i.e. the contents of $GIT_DIR/refs/tags/<tag>) :/	23:16
	+"	23:16
cbreak	on the machine the git repository is on	23:16
arberber	well, of course, i'm in /home/philip/, and he's in /home/john/	23:16
cbreak	arberber: well, then your home directory is in the same folder as john's	23:16
	← radarek left	23:16
arberber	but i mean to clone it from machine A to machine B	23:16
cbreak	so?	23:17
Ulfalizer	as well as "<head> a valid head name (i.e. the contents of $GIT_DIR/refs/heads/<head>)."	23:17
cbreak	as long as you can access it, it's ok	23:17
arberber	i don't understand why us being in /home/ is relevant, but ok.	23:17
	→ CDomovoy joined	23:17
cbreak	arberber: did you not read what I wrote?	23:17
patrikf	just give the full path if in doubt	23:17
cbreak	I wrote: ... ssh://server/../john/path/to/repo.git	23:17
patrikf	and I find ssh syntax way more logical - server:/home/john/foo.git	23:18
arberber	oh, i see.	23:18
cbreak	so the ../john switches to his home directory from yours	23:18
	but only if you are in the same folder	23:18
arberber	i apologize. you just meant due to the way you wrote the path.	23:18
	i understand.	23:18
cbreak	exactly	23:18
	← jstein_ left	23:18
cbreak	patrikf: my ssh runs on port 2222	23:18
	patrikf: :P	23:18
arberber	alright	23:18
patrikf	cbreak: yeah, so you can't use it, but the rest of us can... ;-)	23:19
cbreak	!!!! unfair !!!!	23:19
patrikf	or actually you can	23:19
arberber	let me try all of this. thank you cbreak.	23:19
	← CyberDomovoy left	23:19
patrikf	cbreak: with a proper .ssh/config	23:19
cbreak	nah, I changed /etc/services	23:19
	← ben225 left	23:19
patrikf	what?	23:19
	← zombor left	23:19
patrikf	that's, like, evil!	23:19
cbreak	(probably a bad idea, since now I have to specify the port for every standard server...)	23:19
patrikf	indeed.	23:19
cbreak	but ah well :)	23:19
	patrikf sleeps well knowing cbreak doesn't administrate any systems he has to work on :P	23:20
cbreak	I am not a sysadmin	23:20
	I am software engineer	23:20
	I don't have to maintain stuffjust release it	23:20
	→ jstein joined	23:20
	← jstein left	23:20
	→ jstein joined	23:20
Teltariat	...	23:20
Ulfalizer	cbreak: are you the kind who churn out uncommented buggy code at a billion loc/s and make my life a living hell? :(	23:21
cbreak	no	23:21
	I make code with beautiful APIs, well documented doxygen comments	23:22
	and a UI that makes our users weep	23:22
	and beg for forgiveness	23:22
	→ Ghent joined	23:22
	→ Unknown[NF] joined	23:23
Ulfalizer	software engineers should have more empathy for their fellow coders. "the code is the documentation".. yeah, right, as if intentions are obvious from your crappy code :P	23:23
Ghent	I'm sorry, I'm having real issues trying to find this answer on my own... I'd like to change hosts from github to unfuddle, is there a way I can import my git repo into unfufddle and it keeps all the history and such?	23:23
cbreak	oh, but it is obvious	23:23
	you just have to read the comments	23:23
Ulfalizer	cbreak: i've started to consider those a luxury in other's code :)	23:23
cbreak	or if you are so inclined, read the unit tests	23:23
patrikf	Ghent: probably you just push there...	23:24
cbreak	Ghent: just git remote add the new repo	23:24
	and then push	23:24
Ghent	ok, pretty new with git too... I'll se if I can work that out :)	23:24
cbreak	force it if you have to	23:24
	→ EricInBNE joined	23:24
	← Maletor left	23:26
	← robotarmy left	23:26
	← RobertLaptop left	23:27
Buglouse	exporting GIT_DIR='git://url' GIT_WORK_DIR='/path', should work with just 'git clone' ?	23:28
cbreak	wth are you trying to do?	23:28
Buglouse	Trying to pull from git://url into a defined dir.	23:29
	using env variables	23:29
	→ robotarmy joined	23:29
ddilinger	Buglouse: but why using env variables	23:29
	Buglouse: why not on the command line	23:29
cbreak	first, cd into the repository	23:29
	then pull	23:29
	← blaenk left	23:29
cbreak	if you are not in the repo, you have to set GIT_DIR and so on correctly	23:29
	to _LOCAL_ directories obviously	23:30
	← armut left	23:30
	→ blaenk joined	23:30
cbreak	GIT_DIR and GIT_WORK_TREE at least	23:30
	← maattd left	23:31
	→ priidu joined	23:32
	← robotarmy left	23:33
Buglouse	got it GIT_DIR GIT_WORK_TREE, ok. but want to know if GIT_DIR could work with an external repo, or perhaps another var.	23:34
	and if i have GIT_WORK_TREE defined, why i have to specify it for git clone.	23:35
	← ddilinger left	23:35
	← bousquet left	23:35
	→ robotarmy joined	23:35
patrikf	Buglouse: first you say you want to pull, now you want to clone	23:36
	and no, you can not set GIT_DIR to something remote	23:36
	what do you want to do?	23:36
	→ bousquet joined	23:37
Buglouse	sorry, clone. setting GIT_WORK_TREE='/path' clones into path but then HEAD is put in to the dir defined by git://url.	23:37
	→ maattd joined	23:37
	← vsayer left	23:37
patrikf	yes, clone is a special case, just leave GIT_WORK_TREE and specify the target directory on the command line	23:38
cbreak	Buglouse: no	23:38
Buglouse	just have vars make my life easier, use them so i can just 'git checkout' 'git clone' without specifying anything more.	23:38
	→ vsayer joined	23:38
cbreak	Buglouse: do you know what you have to put into GIT_DIR?	23:38
Buglouse	well.. before i was using GIT_DIR='/path/.git' but no it is unset.	23:39
cbreak	well, you have to put in the path to the .git directory of your repository	23:39
	→ sternocera joined	23:39
patrikf	Buglouse: that doesn't make any sense.	23:39
cbreak	and then you have to put the path to your working tree into GIT_WORK_TREE	23:39
	and THEN you can use git commands outside a git repository	23:39
Buglouse	understand. just wanted to know why it was not working as expect with clone, patrikf states that it is a _special_ case.	23:40
	← Octalot left	23:40
cbreak	it's not special	23:40
Buglouse	and with clone /path is to be specified.	23:41
cbreak	of course...	23:41
	because clone doesn't care about the current git repository	23:41
	→ Veejay joined	23:41
patrikf	(that makes it a special case in that regard IMO ;-))	23:41
Buglouse	didn't know, cbreak. Thought was putting wrong thought to the words from man.	23:41
	^w/e	23:42
	→ normanrichards joined	23:42
Buglouse	thanks	23:42
cbreak	hmm... I guess it's special if you look it at that	23:42
	but it would be rather stupid if it were not	23:42
	← leavittx left	23:42
cbreak	and cloning would require a repository :)	23:42
	← pozic left	23:43
	→ alama joined	23:43
	→ Octalot joined	23:43
Buglouse	again, GIT_WORK_TREE works somewhat without path to clone. just not with HEAD	23:44
patrikf	that would indeed make little sense as clone is just a shortcut command.	23:44
	Buglouse: I don't get that sentence	23:44
	← bousquet left	23:44
cbreak	Buglouse: you have no idea what you are talking about	23:44
patrikf	and neither have I ;-)	23:45
Buglouse	in part	23:45
	→ utu joined	23:46
Buglouse	i meant that if i set the var=/path; using git clone git::/url, contents are copied to /path except for a few.	23:46
patrikf	oh, now I get it, git clone does indeed respect GIT_WORK_TREE	23:48
	just not GIT_DIR	23:48
	→ stigrett joined	23:48
	→ yhager joined	23:48
Buglouse	ture	23:48
patrikf	Buglouse: note that if you need to set GIT_DIR/GIT_WORK_TREE, you have either a very special use case (think <0.1%) or you're doing something wrong (and possibly making your life more difficult than it need be)	23:49
arberber	Er. If I initialize a repository on a machine, does it not automatically create the .git file from which one can clone the repo?	23:50
Buglouse	patrikf: Just trying to make a script, setting GIT vars i can use other cmds without specifying paths.	23:50
patrikf	Buglouse: "cd"?	23:50
	arberber: .git is typically not a file, but a directory; what makes you think it doesn't?	23:51
Buglouse	patrikf: with using as few pushd as possible.	23:51
Ulfalizer	arberber: note that .git won't be listed without e.g. the -a flag to ls	23:51
	← yhager left	23:53
	← SimonP86 left	23:54
	hagebake → hagabaka	23:54
	← Araxia left	23:54
	→ Araxia_ joined	23:55
	← stigrett left	23:55
cbreak	arberber: the .git directory IS the repository	23:56
	← The_Tick left	23:56
cbreak	that's where all the history ends up	23:56
	→ RobertMeta joined	23:56
	→ The_Tick_ joined	23:56
	← RobertMeta left	23:56
	→ RobertMeta joined	23:56
	← DaveIngram left	23:57
	→ AAA_awright_ joined	23:57
	→ jvraoltc joined	23:58
	← HopsNBarley left	23:59
	← MetaCosm left	23:59
	→ coderdad joined	23:59
	→ HopsNBarley joined	23:59
	← HopsNBarley left	23:59
	→ HopsNBarley joined	23:59
	← AAA_awright left	23:59

Logs Search ←Prev date Next date→

Channels Documentation