IRCloggy #git 2010-06-27

Provider of IRC logs since 2005.

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

	2010-06-27
	→ Tuller joined	00:00
	← ChUrP left	00:01
	← jasonn1234 left	00:03
	← Paraselene_ left	00:04
	← mindworx left	00:04
	→ mindworx joined	00:06
	→ Paraselene_ joined	00:10
	→ Tyler-- joined	00:12
Tyler--	what causes: Permission denied (publickey,keyboard-interactive).	00:12
	→ MattDiPasquale joined	00:14
wereHamster	you not having the correct key nor being able to provide a password	00:18
	→ MicahElliott joined	00:25
	← JStout left	00:27
	→ orafu joined	00:27
	→ airborn_ joined	00:30
	← EricInBNE left	00:30
	→ EricInBNE joined	00:31
	→ chrislerum joined	00:32
	← Tuller left	00:35
	← mindworx left	00:38
	← airborn left	00:39
	← airborn_ left	00:39
	→ t0rc joined	00:42
	→ mindworx joined	00:43
	→ patmaddox joined	00:44
	→ resmo joined	00:44
	← little_owl left	00:46
	← bagoor left	00:46
	→ hobodave joined	00:48
	← mindworx left	00:48
	→ warlock_mza joined	00:53
	← patmaddox left	00:55
	→ patmaddox joined	00:55
	← MattDiPasquale left	00:57
	→ alex__c2022 joined	00:58
	← chrislerum left	01:04
	→ chrislerum joined	01:07
	← Tyler-- left	01:13
	→ Vampire0 joined	01:15
	← alex__c2022 left	01:21
	← MicahElliott left	01:23
	← chrislerum left	01:24
	→ Tyler-- joined	01:27
	→ VVSiz_ joined	01:27
	← ft left	01:28
	→ ft joined	01:29
	← skmidry__ left	01:36
	→ hobodave_ joined	01:36
	← warlock_mza left	01:36
	→ CodeRedLin joined	01:38
	← hobodave left	01:39
	← Tyler-- left	01:40
	→ chrislerum joined	01:45
	← CodeRedLin left	01:57
	→ byang joined	02:02
	← homiziado left	02:02
	→ Blackguard joined	02:03
	← Blackguard left	02:04
	← mithro left	02:05
	→ mithro joined	02:05
	→ Blackguard joined	02:05
	→ hasrb joined	02:16
hasrb	How do I ignore merges with the master branch in my commit log?	02:16
	for instance, I make a commit in the "user_login" that describes it and then checkout master and merge it	02:17
	I just don't want the "Merge branch 'user_login'" in my commit logs	02:17
	→ MicahElliott joined	02:19
	→ hohoho joined	02:20
	→ cytrinox_ joined	02:20
DrNick	hasrb: well, what do you want in your commit logs?	02:20
hasrb	DrNick: maybe my workflow is a little off. lets see...	02:21
	git checkout -b changes; make changes; git add . && git commit -a -m "updated x"; git checkout master && git merge changes	02:21
	so I want "updated x" in my logs	02:21
	cytrinox_ → cytrinox	02:21
	→ __skyl__ joined	02:21
hasrb	I dont' care about "Merge branch changes"	02:21
	make changes = edit files, not the make command	02:21
	or maybe that's normal? I guess I just don't see it a lot in other people's logs	02:23
	→ dreiss joined	02:24
	← MicahElliott left	02:25
	→ Elfe_ joined	02:26
	→ hyperair joined	02:27
	→ thomas_b joined	02:32
	← __skyl__ left	02:33
	→ alester joined	02:36
	← harinath left	02:39
	← patmaddox left	02:42
	→ patmaddox joined	02:43
	→ ajpiano joined	02:43
	→ ramusara joined	02:43
	→ walf443 joined	02:43
	← patmaddox left	02:47
	← ajpiano left	02:48
	← hohoho left	02:49
	→ ajpiano joined	02:50
	← walf443 left	02:52
	→ walf443 joined	02:52
	→ harinath joined	02:55
	← walf443 left	02:57
	→ walf443 joined	02:58
	← dreiss left	03:00
offby1	hasrb: there's a --no-merges option to "git log"	03:01
	that might be what you want	03:01
hasrb	offby1: yeah, I was mostly just thinking about on github when I'm working on a project with someone else, they don't really care about my merges with master since they don't explain much	03:01
	i mean you can see the diff though	03:01
	so i guess it does	03:02
	← walf443 left	03:03
	→ walf443 joined	03:03
	← ratc left	03:05
	← walf443 left	03:08
	← plediii_ left	03:10
	→ walf443 joined	03:10
	→ codeshepherd joined	03:12
	← walf443 left	03:15
	→ \mjd joined	03:15
	→ warthog9 joined	03:16
	← iruediger left	03:16
	← harinath left	03:18
	→ walf443 joined	03:18
	→ _2x2l joined	03:20
	← warthog9 left	03:22
	→ mindworx joined	03:23
	← walf443 left	03:24
	→ walf443 joined	03:25
scatterp	how would i figure out the svn url to get this branch http://github.com/jacek/imprudence/tree/hpa-backup ?	03:25
	→ jasonn1234 joined	03:25
DrNick	scatterp: the what?	03:26
scatterp	DrNick i want to download the hpa-backup branch ?	03:26
DrNick	ok, clone it	03:27
	→ eno joined	03:27
scatterp	DrNick i have to put a url like http://meerkat-viewer.googlecode.com/svn/trunk/	03:28
	oh i need a git program right	03:28
DrNick	you understand that git is not svn, right?	03:28
scatterp	not really no :)	03:28
	i do now	03:29
	→ univate joined	03:29
	← mindworx left	03:29
scatterp	so i need to download something to get it	03:29
doener	if you insist on using svn: http://svn.github.com/jacek/imprudence.git	03:30
scatterp	oh cool	03:31
	→ harinath joined	03:31
scatterp	doener will that get me the hpa-backup branch	03:31
	or will it get all the branchs ?	03:31
	→ warthog9 joined	03:32
doener	I have no clue about the svn stuff at github, I value my brain too much to use svn	03:32
scatterp	k	03:32
	ill test it out	03:32
	i just care about geting the code	03:32
	my codeings to bad to contribute	03:33
	→ mindworx joined	03:33
	← ramusara left	03:34
	← walf443 left	03:38
	← mindworx left	03:40
	→ walf443 joined	03:40
	→ MattDiPasquale joined	03:40
	← hasrb left	03:44
	→ mindworx joined	03:46
	← mindworx left	03:48
	← walf443 left	03:50
	→ walf443 joined	03:53
	→ maleknet joined	03:55
	→ happypolla joined	03:55
	→ OFICINA7 joined	03:59
	← codeshepherd left	04:00
	← harinath left	04:04
	→ codeshepherd joined	04:05
	← walf443 left	04:05
	→ walf443 joined	04:06
	← \mjd left	04:09
	→ \mjd joined	04:10
	← walf443 left	04:10
	→ walf443 joined	04:10
	→ multigrain joined	04:14
	← multigrain left	04:15
	→ ignatp joined	04:16
	← maleknet left	04:17
	← walf443 left	04:18
	→ harinath joined	04:18
	→ KiiK joined	04:18
	→ consolers joined	04:19
	← t0rc left	04:19
KiiK	"git:// http:// " what's the different between this 2	04:19
Ilari	KiiK: git:// is dedicated protocol for anonymous git repo fetching. HTTP is http (quite bad without serverside support).	04:20
KiiK	llari: any performance different ?	04:21
	→ maleknet joined	04:21
	← ignatp left	04:22
Ilari	KiiK: If HTTP is not backed by serverside CGI, then git:// has loads better performance.	04:22
	→ walf443 joined	04:22
Ilari	KiiK: If HTTP has server cgi backing it, then they are about equal (git:// might still be bit better).	04:22
	KiiK: The recomended protocol for pushing is ssh://	04:24
	← codeshepherd left	04:25
KiiK	llari: thanks	04:26
Ilari	KiiK: HTTP is mostly meant as backup for those that have overly strict firewalls.	04:26
scatterp	trying to check out a specific branch using git http://pastebin.com/fhX8m7km i get this error	04:27
	i am a little confused :S	04:27
Ilari	scatterp: You didn't cd to directory it created (and thus are outside repo).	04:28
scatterp	oh	04:29
Ilari	scatterp: And avoid spaces in paths... They may or may not cause trouble (depending on software).	04:29
scatterp	oh they will	04:29
	i was thinking just to get the src code i want	04:29
	and rename the folder after	04:30
	is that ok ?	04:30
	→ codeshepherd joined	04:30
Ilari	scatterp: Yes.	04:30
scatterp	ok so now i did that i didnt see much erm activity	04:31
	← walf443 left	04:31
scatterp	i am trying to compile that branch	04:31
	so i am kinda missing the step where a folder appears with that branch name ??	04:31
	← happypolla left	04:33
Ilari	scatterp: This is not SVN.	04:34
	scatterp: Checkout the correct branch (see 'git branch -r').	04:34
	← chrislerum left	04:35
scatterp	ok	04:35
	→ walf443 joined	04:36
scatterp	im still lost	04:37
	→ anathematic joined	04:37
	→ brian joined	04:40
	← brosner left	04:41
	← codeshepherd left	04:44
	← walf443 left	04:45
	← alester left	04:48
	→ walf443 joined	04:48
	→ codeshepherd joined	04:50
	← consolers left	04:53
	→ keyvan joined	04:56
	→ scatterp joined	04:56
	← adamm left	04:57
	← walf443 left	04:58
	← KiiK left	04:58
	→ MicahElliott joined	04:59
	→ walf443 joined	05:01
	brian → brosner	05:02
	← walf443 left	05:11
	→ walf443 joined	05:15
	← codeshepherd left	05:16
	→ jrmuizel joined	05:20
	→ Yuffster joined	05:22
	← walf443 left	05:23
	→ walf443 joined	05:24
	→ fmardini joined	05:25
	→ xiong joined	05:26
xiong	Okay, now is the time I really want git to work for me. I don't want to back out of a whole commit but I want to restore a few files to their state at the time of an earlier commit. Can it be done?	05:27
DrNick	git checkout $COMMIT -- $FILES	05:31
	← walf443 left	05:33
	→ walf443 joined	05:35
	→ bodom joined	05:37
xiong	DrNick, Let me look at that. I want the current files the way they are, too, so I've moved them into another folder.	05:38
	→ smorg joined	05:39
xiong	DrNick, 'man git-checkout' doesn't indicate that at all. Not saying you're wrong, just saying I don't see it.	05:39
DrNick	third paragraph in the description section	05:40
	also the last line in the synopsis	05:40
	← walf443 left	05:43
xiong	It does work; thank you.	05:44
	→ happypolla joined	05:44
	→ walf443 joined	05:45
	→ niki joined	05:45
xiong	Manpage uses the word 'paths'; I'm thinking 'directories' and wanting files.	05:45
Leemp	How would i delete a tag from a remote repo? I assumed i could delete it locally, and push that, but that did not work. Any thoughts?	05:47
DrNick	git push --delete refs/tags/$TAG	05:49
	← walf443 left	05:49
	→ walf443 joined	05:50
	← walf443 left	05:51
	→ walf443 joined	05:52
xiong	Thanks again, DrNick++. This is why I'm using git.	05:53
	← walf443 left	05:56
	← OFICINA7 left	05:56
	→ marcial5 joined	05:57
	← bodom left	05:57
	→ walf443 joined	05:58
	→ scarabx joined	05:59
	← hyperair left	06:01
	→ codeshepherd joined	06:09
	← walf443 left	06:10
	→ walf443 joined	06:11
	→ SvenDowideit joined	06:16
	← scarabx left	06:18
	← walf443 left	06:18
	→ walf443 joined	06:19
	← codeshepherd left	06:19
	← jasonn1234 left	06:20
	→ mjf joined	06:25
	→ ExtraSpice joined	06:25
	← Blackguard left	06:27
	→ codeshepherd joined	06:27
	← walf443 left	06:28
	→ walf443 joined	06:30
	← bobrik left	06:32
scatterp	heres my git branch -r how can i merge origin/hpa-backup + origin/login-manager ?	06:33
	← walf443 left	06:34
	→ walf443 joined	06:35
	← jrmuizel left	06:35
	→ g3d joined	06:41
	→ nchaimov joined	06:42
	← codeshepherd left	06:43
	← walf443 left	06:47
	← \mjd left	06:47
	← anathematic left	06:48
cxreg	into each other? or into master?	06:48
	→ anathematic joined	06:48
	→ walf443 joined	06:51
	← mjf left	06:51
	→ project2501a joined	06:51
	← MicahElliott left	06:52
	← anathematic left	06:53
	→ JdGordon joined	06:54
	→ eletuchy joined	06:54
JdGordon	I've got a local repo here with 2 remotes (ne for upstream and one for github)... how do I create a remote branch on hitbud from a local branch?	06:54
	→ doener joined	06:55
JdGordon	$ git branch --track github_newparser github/newparser <- no worky	06:55
	→ tvw joined	06:55
	→ [reaVer] joined	06:55
	← [reaVer] left	06:55
	→ [reaVer] joined	06:55
JdGordon	upstream is origin, github is the github remote	06:55
scatterp	cxreg into master	06:57
cxreg	scatterp: git merge origin/hpa-backup; git merge origin/login-manager	06:57
	← coppro left	06:57
cxreg	what's so hard about that? :)	06:57
scatterp	its kinda unlogical but ok	06:57
cxreg	why?	06:57
scatterp	its the lack of any real confirmation that git has done something	06:58
	that messes with me	06:58
cxreg	you want fireworks and clowns?	06:58
scatterp	cxreg i think i already did that will doing it again be ok ?	06:58
cxreg	sure, you can merge as many times as you want	06:58
	if there's nothing unmerged, it will just do nothing	06:58
scatterp	cxreg "your local file system was altered" would work	06:58
cxreg	check git log	06:59
scatterp	how do i get out	07:00
	after?	07:00
	→ aless_ joined	07:00
scatterp	got it	07:01
	← mithro left	07:02
	← aless left	07:03
	→ LongBeach joined	07:04
	→ rajeshsr joined	07:07
	→ hyperair joined	07:12
	→ bababfds joined	07:12
	→ blinky joined	07:18
	→ wolog joined	07:18
	→ DavidKlein joined	07:21
	→ MicahElliott joined	07:24
	→ marc_kdab joined	07:27
	→ thiago_home joined	07:28
	→ Titosemi joined	07:31
	← Yuffster left	07:32
	→ lhz joined	07:33
	→ priidu joined	07:38
	← hyperair left	07:40
	← tvw left	07:44
	→ Acorn joined	07:44
	← MattDiPasquale left	07:45
	→ drizzd joined	07:50
	→ justin-george joined	07:54
	← justin-george left	07:54
	→ airborn joined	08:00
	← marc_kdab left	08:01
	← MicahElliott left	08:04
	→ raichoo joined	08:05
	→ [Po]lentino joined	08:07
	→ candybar joined	08:08
	→ jstemmer joined	08:12
	→ mbroeker joined	08:14
	← airborn left	08:17
	→ rox` joined	08:17
	→ pantsman joined	08:18
	→ werdan7 joined	08:19
	→ poseid joined	08:19
	→ giskard joined	08:25
	← raichoo left	08:38
	→ ziro` joined	08:38
	→ pcwick joined	08:40
	← orafu left	08:43
	← rox` left	08:50
	→ bobrik joined	08:53
	→ sylr joined	08:55
	→ simplechat joined	08:56
	→ issyl0 joined	09:03
	→ airborn joined	09:05
issyl0	Hi there.	09:05
pragma_	issyl0: what is the meaning of that?	09:06
issyl0	I'm having a few problems when I do git push.	09:06
	Er, let me pastebin.	09:06
	http://issyl0.co.uk/repository/giterror1.txt There's the error I'm getting when I try to push.	09:06
wereHamster	issyl0: permission problems?	09:08
issyl0	wereHamster: not that I can tell.	09:08
	It worked five minutes ago when I pushed osomething else!	09:09
	→ gebi joined	09:09
	→ Jaxan joined	09:13
	← smorg left	09:14
	→ tizbac joined	09:16
	→ disappearedng joined	09:20
	→ justin_ joined	09:22
	→ tobiassjosten joined	09:26
	← airborn left	09:28
	← keyvan left	09:30
	→ keyvan joined	09:33
	→ engla\|swe joined	09:34
	→ anathematic joined	09:35
	← keyvan left	09:39
	→ mohamed_yosry joined	09:40
	→ ph^ joined	09:40
	← poseid left	09:40
	← Titosemi left	09:42
	→ bagoor joined	09:42
	→ markbao_ joined	09:44
	→ ThiefMaster` joined	09:45
	← ThiefMaster` left	09:45
	→ ThiefMaster` joined	09:45
	ThiefMaster` → ThiefMaster	09:45
	→ keyvan joined	09:46
	→ k3yvn joined	09:47
	→ rox` joined	09:49
doener	issyl0: the previous push likely just didn't try to write to that directory	09:50
issyl0	doener: hmmm?	09:50
doener	issyl0: objects/40 contains loose objects, the "40" is a prefix of the 40 character sha1	09:50
issyl0	Ah right.	09:50
	← keyvan left	09:50
	→ aziz joined	09:50
doener	issyl0: so if the previous push didn't transfer any objects with a sha1 starting with 40, it didn't touch that directory	09:50
issyl0	Well why did it decide to then?	09:50
doener	issyl0: well, _this_ push needs to transfer such an object	09:51
	issyl0: I'm just saying that it's easily possibly that one push works, while a different one fails due to broken permissions	09:51
issyl0	OK.	09:52
	:)	09:52
doener	issyl0: check the owner of objects/40, I bet it's wrong	09:52
issyl0	Heh.	09:55
	I'm going to start again. :)	09:55
	Thanks for your help.	09:55
	← bagoor left	09:56
	→ corni joined	09:58
	→ francisfish_away joined	09:58
	→ psoo joined	09:59
	→ radarek joined	10:07
radarek	hello, do you know any tools to create diagram with nice graphs of commits (I'm making a presentation about git) - preferred for OSX but it doesn't have to	10:08
	← tobiassjosten left	10:09
	← gebi left	10:13
	← harinath left	10:15
	← mohamed_yosry left	10:15
cbreak	radarek: looked at gitx and graphviz=	10:16
radarek	cbreak: can graphiz use some custom themes? (those showed in gallery are ugly :))	10:18
cbreak	themes?	10:18
	it doesn't have themes	10:18
	you can control everything	10:18
	→ consolers joined	10:18
radarek	cbreak: even how elements looks like (color, border, radius etc)?	10:20
cbreak	yes...	10:20
radarek	ok, so I'll look at this	10:20
	thanks	10:20
cbreak	I got this in my .gitconfig:	10:20
	alias.graphviz=!f() { echo 'digraph git {	10:20
	node [shape=record];' ; git log --pretty='format: "%h"[label="{%ci\|%s}"]' "$@" \| sed 's/\|$[^}]\{0,32\}$[^}]/\|\1/g' ; git log --pretty='format: %h -> { %p }' "$@" \| sed 's/[0-9a-f][0-9a-f]/"&"/g'; echo '}'; }; f	10:20
	I don't claim that it's good :)	10:20
	→ phlowx_ joined	10:20
cbreak	but it did something, ages ago when I wrote it	10:20
j0rd	i in-correctly have initiated two .git directories in my code path. /home/www/domain.tld and /home/www/domain.tld/htdocs . I've been commiting to both by accident. I want all my commits to go into the lowest parent directory. Can I merge these two repositories history? Or should I white the higher child and re-commit to parent?	10:22
cbreak	j0rd: look at man git-merge with subtree	10:23
	I've never used it, but it might do what you want	10:23
doener	http://www.kernel.org/pub/software/scm/git/docs/howto/using-merge-subtree.html	10:24
	note that it's somewhat ugly, because you have a rename in the history...	10:25
	I'd probably use filter-branch to either build a clean history from the start, or at least rewrite the history for "htdocs" to move everything in it down into a "htdocs" directory, and then merge that new histry	10:26
	/ry$/ory$	10:26
	→ dr_win joined	10:26
	→ hyperair joined	10:27
	→ kipras joined	10:27
	→ radioman-lt joined	10:30
	→ harinath joined	10:30
	→ raichoo joined	10:32
	← niki left	10:36
	→ parasti joined	10:37
	→ poseid joined	10:39
	← poseid left	10:39
	→ mithro joined	10:43
	← hyperair left	10:43
	→ mithro_ joined	10:45
	← mithro left	10:47
	→ frank_ joined	10:48
	→ disappearedng_ joined	10:49
	→ Bennid joined	10:51
	← disappearedng left	10:53
	← aliceinwire left	10:55
j0rd	cbreak: thx. I'll take a look	10:57
	← Jaxan left	10:57
	→ Jaxan joined	10:58
	→ hyperair joined	11:00
	→ gebi joined	11:00
	← mithro_ left	11:01
	← Jaxan left	11:03
	→ ChUrP joined	11:07
j0rd	to save myself some pain, might it just be easier to remove the child and then re-add/commit to the parent?	11:09
	i don't think i'm comfortable enough with git to attempt this merge	11:09
cbreak	you'd lose all history	11:09
	← henkboom left	11:09
j0rd	and preferably, i'd like to spend my time working, instead of fighting with git	11:09
	ya, i know i'll lose it	11:09
	→ aspotashev\|eeepc joined	11:10
	→ icwiener joined	11:10
j0rd	but it should work fairly simple after that i immagine	11:10
cbreak	git allows you to keep it	11:10
	but it's tricky as described above	11:10
j0rd	especially when i'm dealing with around 8430 files :)	11:11
	backing it up and attempting	11:12
	we'll see	11:12
	if it takes me more than 20 minutes, gonna just lose my history and move on	11:12
	→ henkboom joined	11:12
cbreak	number of files should not matter	11:12
j0rd	unless i have same files, which i do	11:13
	i assume	11:13
cbreak	how can you have the same files when one is in a subdirectory of the other?	11:13
	→ icwiener_ joined	11:13
j0rd	i might have commited them from the root and i may have commited them from the child as well	11:14
	i wasn't paying attention	11:14
	granted i understand this is my mess up	11:14
	i just didn't notice i had a .git in my child folder...	11:15
cbreak	I don't think that's possible	11:15
	if you have a .git in your child, everything in that child belongs to the inner .git	11:15
j0rd	i guess we'll find out. please excuse my ignorance with git. still learning	11:15
	← icwiener left	11:15
cbreak	the outer doesn't know about it... I think...	11:16
j0rd	gotcha	11:16
cbreak	you can easily check that by the way	11:16
j0rd	i know. just gonna try the merge and see what happens	11:16
cbreak	I don't think straight merging will work	11:17
	ah anyway, good luck :)	11:17
	← anathematic left	11:17
FauxFaux	If you have cpu time to burn you can make merge attempt to resolve renames on stupid numbers of files, very useful.	11:17
j0rd	not this sub-tree merge strategy link i have?	11:17
	→ anathematic joined	11:22
j0rd	"Merge with strategy ours failed." on step 2 of "how to use the subtree merge strategy"	11:22
	=(	11:22
	→ bagoor joined	11:22
	→ unreal_ joined	11:32
j0rd	can i export the commit history from one repo and merge it into the other somehow?	11:33
	like export the history from my child. remove the .git folder. Import it into my root somehow altering the filepaths?	11:33
	unreal_ → unreal	11:33
Rhonda	j0rd: One approach might be git format-patch and then git am	11:34
	In the files that git format-patch exports you can easily tweak the filepaths, if needed.	11:35
cbreak	in your situation, you'll first have to git filter-branch to move the subdirectory contents	11:35
j0rd	any non-kernel.org howtos with that	11:35
	kernel.org docs are fairly technical and confusing. they assume i know what the hell I'm doing	11:35
	i'll try and find some alternate documentation, but if you know of any that would be wonderful	11:36
	cbreak learnt git from man pages	11:36
cbreak	well, man pages and http://eagain.net/articles/git-for-computer-scientists/	11:36
j0rd	i'd like git for dumbasses	11:37
	:)	11:37
cbreak	there is a git-for-svn-users	11:37
j0rd	but thank you	11:37
cbreak	guess that's as close as it gets	11:37
j0rd	i feel there was a mild flame in ther :)	11:37
cbreak	just a bit... :)	11:37
	→ airborn joined	11:38
	← radarek left	11:39
	→ Luke__ joined	11:41
	← happypolla left	11:44
	→ apsysoev joined	11:48
j0rd	so i just deleted my root/htdocs/.git child folder i was trying to get rid of, expecting that my root/.git would not all those files committed into the child...but the root seems to have them all. I dunno wtf	11:49
	← Bennid left	11:50
j0rd	i guess i just lost the history	11:50
	← airborn left	11:50
j0rd	what ever, just going to move on and hope my repo isn't messe	11:51
	→ Zephyrus joined	11:52
	→ Chris64 joined	11:53
j0rd	applying patches I made with `format-patch` wouldn't work as the files are already technically patched. I guess this is why i needed to "filter-branch" to revert back my changes in the child branch and then re-apply them	11:53
	all really confuzing.	11:53
	i'm going to love trying to explain git to my designers. I'm sure this will all make tonnes of sense to them :D	11:54
	→ jasonn1234 joined	11:54
cbreak	designers?	11:55
	git's not that good at dealing with big binary files	11:56
j0rd	is any VCS?	11:56
gebi	perforce, svn	11:56
j0rd	mostly just store jpgs and gifs and pngs	11:56
cbreak	well, it's not like git will fail or something	11:57
	it'll just not be too efficient	11:57
j0rd	which are usually a couple K and don't get changed often	11:57
gebi	ah, should have said earlier ;)	11:57
j0rd	but if they edit html, i want to be able to work in the VCS with them for that	11:57
gebi	binaries files in GB are problematic with git	11:57
j0rd	i don't do anything that large. having 1GB jpgs is also problamatic for web browsers	11:58
gebi	heh	11:58
	← pcwick left	11:58
j0rd	i've been toying with the idea of looking at Bazaar, but if drupal decides to go with git, so will I	11:59
	native gui clients would help me actually get the people I work with to use the VCS instead of just emailing me, or uploading changes	12:00
	i think git will prove to be too difficult for them	12:00
ciaranm	do you really want people who can't grasp basic git to contribute?	12:01
gebi	depends on, if they are not svn infested a quick introduction may save you many hours of work, if they expect to work with git like svn you better use svn for them	12:01
	and sync with git from there	12:01
j0rd	ciaranm: yes, i do	12:02
cbreak	look at gitx	12:02
	it's a nice gui	12:02
j0rd	ciaranm: unfortunately very few good graphic designers are actually computer scientists or kernel programmers	12:02
	and unfortunately as a web programmer, I am destined to collaborate with them	12:03
ciaranm	i think you're overestimating the knowledge required to use git	12:03
j0rd	so i need a version control system people can actually figure out and use, instead of one that's efficient in storage and performance	12:03
cbreak	you could make them a few scripts that they can just start	12:03
	one that does git commit -a	12:03
	one that does git pull & git push	12:04
gebi	scripts that work on OSx and Windows?	12:04
cbreak	one for each	12:04
j0rd	ciaranm: i think you're over estimating the patients and skill of people i work with :)	12:04
	and my spelling	12:04
	cbreak: ideally a VCS would be easily enough, that i wouldn't have to create scripts for my users to use it	12:05
cbreak	it is :)	12:05
ciaranm	the problem with easy is that sooner or later you're going to need not easy	12:05
cbreak	they can type it themselves	12:05
j0rd	http://code.google.com/p/tortoisegit/ seems to be simple enough for people to figure out. at least witih svn	12:05
cbreak	but I found that at least my sister has major problems even remembering basic commands	12:06
	like "cd" to switch to the correct folder	12:06
j0rd	cbreak: number of users i've worked with at my past job who used a VCS on the command line were 1	12:06
	and that was me	12:06
	everyone else uses a gui	12:06
	because they are stupid and lazy	12:06
	mostly lazy	12:06
	← xiong left	12:06
ciaranm	the lazy thing to do is to learn how to use git	12:06
	because sooner or later it'll save you from having to redo loads of work	12:07
j0rd	ya, but that takes time and experience. most users will give up after a project if it's not easy to use	12:07
	granted git is pretty good for normal workflow, it's gets hard when you have problems	12:07
	but so do all VCSs	12:07
ciaranm	no, with most VCSs you're screwed when you have problems	12:07
j0rd	ive rarely been screwed...but granted a have a lot more experience with other past VCSs	12:08
	← bagoor left	12:08
j0rd	and i didn't do a lot of branching, which caused problems	12:08
	which is eventually why i switched to git (branching)	12:09
	again though, my concern is that it'll be too hard for the people i collaborate to figure out :)	12:09
	but so are most VCSs	12:09
	work on a couple projects with people who don't get VCSs and who are not programmers. when get back to me with your experience	12:10
	s/when/then/g	12:10
ciaranm	the solution is to teach them how to use a VCS	12:11
j0rd	i work with people remotely	12:11
	not in the same office	12:11
ciaranm	so?	12:11
j0rd	i don't have that luxery	12:11
ciaranm	presumably they know how to read	12:12
	← henkboom left	12:12
j0rd	kernel.org documents are confusing	12:12
	even for me	12:12
ciaranm	there are plenty of other git documents	12:13
j0rd	yep, i know	12:13
	and i find those instead	12:13
	but you assume people want to learn a VCS. most cases they dont. they are being forced to use them. if they run into problems. I'd be the one having to resolve their issues. So I'd usually just get them to zip up their files send em over and I'd do it myself	12:14
	teaching them the art of VCS would take much longer	12:14
	← jasonn1234 left	12:14
imarcusthis	fish vs fishing	12:15
j0rd	all i'm trying to say, is for me. usability is one of the top concerns with me using a VCS with the people I work with. Because then, they might actually use it	12:15
	→ henkboom joined	12:15
j0rd	imarcusthis: ya, but you can't really teach everyone to fish, especially if theyre affraid of water. you'd have to get over that first :)	12:15
ciaranm	if they're not prepared to learn git, fire them and hire people who are	12:16
j0rd	i don't always get to choose who i work with	12:16
imarcusthis	quit and start your own company. worked out well for me.	12:16
j0rd	and i wouldn't expect a designer to understand git	12:16
cbreak	or just encourage them	12:16
	"Hey, I liked the version of that style from 2 weeks ago too, can you show me how it looked back then?"	12:16
j0rd	i'm thrilled if they get it, and Im supprized when they actually do	12:17
	it's rare	12:17
	otherwise an sFTP account with a repo for them. email notifications upon changes and commit the stuff myself	12:18
	;)	12:18
	not trying to change the world. just get work done	12:18
jast	I continue to be astounded at how much people underestimate the human capacity for learning]	12:19
imarcusthis	I work with designers	12:19
	they use git	12:19
	← giskard left	12:19
jast	but I agree that much of the existing documentation is not very friendly	12:19
imarcusthis	they even manage to do branching, somewhat.	12:19
j0rd	imarcusthis: do did i. but they're not very good at it	12:19
	imarcusthis: and if it breaks, who do they talk to?	12:19
imarcusthis	j0rd: you? :)	12:20
cbreak	send them to #git, we'll make fun of them.	12:20
j0rd	:)	12:20
	→ darkredandyellow joined	12:23
	[reaVer] → reaVer	12:23
darkredandyellow	guys, small problem with "--squash": "git merge bar --squash" when I am in branch foo.	12:23
	the merge happens correctly	12:24
	but: git log doesnt show anything from "bar", meaning the last SHA I see is the last from foo	12:24
	← JdGordon left	12:25
darkredandyellow	so the merge was successful, but I have no indication that it was even merged by just looking at it	12:25
ciaranm	darkredandyellow: --squash doesn't preserve information from the other branch	12:25
	darkredandyellow: think of --squash as creating a new, unrelated commit that does the same as all of the changes you'd otherwise be merging	12:25
darkredandyellow	ciaranm, yes, but i thought at least the merge would show up in git log?	12:25
	in other words: what can i do that i still see that there was a merge?	12:26
ciaranm	darkredandyellow: do a 'git status'. did your squash get committed?	12:26
	→ jasonn1234 joined	12:26
ciaranm	merge --squash isn't a 'merge'. it's a commit that does the changes that a merge would do.	12:26
	→ str1ngs joined	12:27
darkredandyellow	ciaranm, yes, my squash got committed	12:27
ciaranm	darkredandyellow: then 'git log' will show you that commit	12:28
darkredandyellow	ciaranm, you're right.	12:28
	thanks for the clarification	12:28
engla\|swe	darkredandyellow: the commit message should make it clear that it's a merge	12:28
str1ngs	I'm having trouble deleting a remote branch. when I run. git push origin :foo. I get ! [remote rejected] foo (deletion of current branch prohibited) . how I can I fix this?	12:28
offby1	str1ngs: maybe there's a command line switch like "--force"	12:29
str1ngs	offby1: ok let me check	12:29
offby1	or maybe you can log in to the remote box, and check out some other branch	12:29
str1ngs	its on github	12:29
	ciaranm mutters something about faq non-bare for good measure	12:29
	→ ilogger2 joined	12:46
	→ rseifert joined	12:52
	→ rajeshsr joined	12:56
	→ ghost_olli joined	12:59
	→ ghost_olli_ joined	13:06
	→ yann joined	13:06
	→ bandu joined	13:08
	← ghost_olli left	13:09
	← ghost_olli_ left	13:12
	→ mw joined	13:12
	← mw left	13:12
	→ mw joined	13:12
	→ mw2 joined	13:13
	→ bentob0x joined	13:13
	→ hohoho joined	13:16
	→ rubydiamond joined	13:17
	← bentob0x left	13:25
	→ jrmuizel joined	13:27
	→ ijuma joined	13:28
	→ plediii_ joined	13:29
	→ jasonn1234 joined	13:29
	→ psoo joined	13:31
	← psoo left	13:32
	→ poseid joined	13:43
	← jasonn1234 left	13:44
	→ shade_ joined	13:47
	→ VVSiz joined	13:48
	→ fr0sty joined	13:48
	→ khelll joined	13:58
	← rubydiamond left	13:58
	→ LauJensen joined	14:00
LauJensen	I have a branch with several commits that I need to clear out - What is the command to merge a branch, picking out one commit at a time?	14:00
	(is there a magit thing to help me do this)	14:00
cbreak	you can use git rebase -i to remove the commits	14:01
_ikke_	LauJensen: Try looking at git rebate	14:01
	rebase*	14:01
LauJensen	Ah ok	14:01
jast	rebase -i is magic	14:01
LauJensen	How do I make the new branch then ? Since checkout will give me all of the old commits, I dont think that would be me the result I want	14:01
	I need an empty branch right?	14:02
jast	rebase -i is for reshaping a branch, so you don't actually need to start from an empty branch	14:03
	so if you want to keep your old branch, just create a copy and work on that	14:03
LauJensen	and I want an empty branch or a new repo, how would I do that ?	14:04
	→ jasonn1234 joined	14:04
	→ notjohn joined	14:04
	→ cele joined	14:04
	← cele left	14:05
jast	well, a new repository is simple, right? just run git init in an empty directory	14:05
LauJensen	sure, and then the rebase to loot from the old repo ?	14:05
jast	creating an empty branch in an existing repository is a bit messy... there's an explanation on http://book.git-scm.com/	14:05
	do you actually want to keep the old version of the history?	14:05
LauJensen	For those commits that are interesting yes, everything else can stay where it is	14:06
	s/can/should/	14:06
jast	oh well	14:06
	just create a copy of the branch (git checkout -b newbranch) and run git rebase -i on that, with the oldest commit you want to keep as its argument	14:06
	→ madewokherd joined	14:07
jast	and see the description of rebase -i (section "interactive mode" in the manpage)	14:07
LauJensen	I know I was a bit unclear to begin with, but this is what I would prefer: Make a new repo, rebase from the old repo, making a new shiny repo with a perfect history. Is it possible to rebase from another repo like that ?	14:08
jast	just clone the repo and reshape the clone	14:08
LauJensen	k, lemme try :)	14:09
	← jasonn1234 left	14:12
LauJensen	Ok heres the thing. Cloning it is fine, but the rebase should end of having only the master branch in a prettified version, rebasing will put me in another branch right?	14:13
	→ jasonn1234 joined	14:14
	→ t0rc joined	14:19
LauJensen	Someone must have the answer to this. Take a branch from 1 repo, and save is in a cleaned up form in another...?	14:20
	→ ChUrP1 joined	14:20
	→ sagsousuke joined	14:21
	← t0rc left	14:22
avar	Interesting. I can commit empty dirs by making them a subproject. I wonder what abusing that would harm..	14:22
thiago_home	it's easier to just make them non-empty	14:22
	→ aspotashev\|eeepc joined	14:23
	← ChUrP1 left	14:25
	→ slakmagik joined	14:25
	→ coderdad joined	14:26
	→ MattDiPasquale joined	14:29
	← rseifert left	14:32
	→ xiong joined	14:34
	← coderdad left	14:41
	← khelll left	14:42
	→ d0k joined	14:45
	← fr0sty left	14:49
	→ fr0sty joined	14:50
	→ zubin71 joined	14:51
zubin71	hi is there any place i can read about the git protocol alone? its implementation and stuff like that?	14:52
offby1	the git source	14:55
	← poseid left	14:55
	→ alsuren joined	14:55
	→ Jaxan joined	14:58
zubin71	offby1, okie	15:01
	thnkx!	15:01
	→ starcoder\|Svr joined	15:03
offby1	Don't call me no okie!	15:03
	→ geirr joined	15:05
	← alsuren left	15:10
	→ harinath joined	15:15
	→ level09 joined	15:17
level09	why is there origin/head and origin/master	15:17
	→ bdimcheff joined	15:17
	→ raichoo joined	15:18
	→ hyperair joined	15:18
wereHamster	you mean origin/HEAD, right?	15:19
level09	yes	15:20
	→ tizbac joined	15:21
	← jasonn1234 left	15:21
level09	why is that considered another branch?	15:21
thiago_home	because HEAD exists on the server	15:21
level09	isn't HEAD refers to my current location	15:21
thiago_home	no	15:21
level09	in the commit tree ?	15:21
thiago_home	well, yes	15:21
	but origin/HEAD is not HEAD	15:21
offby1	HEAD is the currently-checked-out branch, no?	15:21
thiago_home	but it's there to help you: you can simply write "origin", meaning origin/HEAD	15:22
offby1	or a plain SHA1 if you're "detached"	15:22
level09	thiago_home: but isn't that identical to origin/master?	15:22
thiago_home	it could mean something else	15:22
	and that's 7 characters longer	15:22
	→ n2d joined	15:24
thiago_home	by default, origin/HEAD points to the branch that origin's HEAD points to	15:24
	but you can change it locally (no effect on remote)	15:24
	← bandu left	15:25
n2d	when writing a commit message what should be taken in to account	15:25
	→ dreiss joined	15:26
	← mw left	15:26
wereHamster	n2d: <HEAD:Documentation/SubmittingPatches>	15:26
	← mw2 left	15:27
wereHamster	n2d: http://git.kernel.org/?p=git/git.git;a=blob;f=Documentation/SubmittingPatches;h=eb53e0636e3c3bab06e88ce3371945f5602c5756;hb=HEAD	15:27
	(where's the bot?)	15:27
n2d	wereHamster: sorry, i didnt get it	15:27
	wereHamster: thanks	15:27
	→ jensn joined	15:32
	→ jfkw joined	15:32
	← zubin71 left	15:33
	→ airborn joined	15:40
level09	how do I replace my branch when pulling ?	15:44
thiago_home	explain	15:45
offby1	what do you mean by "replace"?	15:45
level09	I want to force my branche to be overwritten by the remote one	15:45
offby1	ah	15:45
	"git fetch ; git reset --hard FETCH_HEAD"	15:45
	→ programble joined	15:45
	→ AkumaStreak joined	15:45
level09	this will create a new commit that is identical to the remote branch ?	15:45
offby1	no.	15:45
	Try It And See™.	15:46
level09	alright	15:46
	:DD	15:46
offby1	1) Read "Git For Computer Scientists" (http://eagain.net/articles/git-for-computer-scientists/); 2) run "gitk --all"	15:46
level09	i cant run gitk	15:47
	its centos	15:47
	server	15:47
jast	tig is a pretty nice alternative	15:47
thiago_home	level09: so?	15:48
level09	let me check that	15:48
	→ khmarbaise joined	15:49
level09	nice	15:49
	→ neutrino joined	15:50
	neutrino → tar-	15:50
	← harinath left	15:51
	← sagsousuke left	15:53
	← plediii_ left	15:56
	→ henkboom joined	15:59
	← Jaxan left	16:00
	→ khelll joined	16:05
	← khelll left	16:05
	← n2d left	16:05
	→ harinath joined	16:06
	→ jmcantrell joined	16:07
	→ vu3rdd joined	16:07
	→ scarabx joined	16:08
	→ amitsamtani joined	16:11
amitsamtani	is anybody in here?	16:13
aep	no. just 608 other people	16:13
rkreis	607	16:14
	→ Mycr0ft joined	16:15
	← xiong left	16:15
	← harinath left	16:15
	← AkumaStreak left	16:15
	→ pcwick joined	16:16
	→ hal joined	16:17
	→ Titosemi joined	16:17
	→ icefox joined	16:17
	← Mycr0ft left	16:21
	→ ziro` joined	16:21
	← dreiss left	16:22
gebi	hm... error: git-svn died of signal 11	16:23
thiago_home	11 = SEGV	16:24
	→ dreiss joined	16:25
gebi	yea i know ;), though perl shouldn't segfault	16:25
	hm... nevermind, repo is good, git-svn seems to have done it's work	16:26
amitsamtani	thanks aep :)	16:26
	← jmcantrell left	16:26
	→ jmcantrell joined	16:27
amitsamtani	im really new to git, git push is supposed to push everything to the remote repository right? all the changes that i commit does not get pushed... it says "everything is up to date" anyone know why?	16:27
	← level09 left	16:28
	→ harinath joined	16:32
thiago_home	no	16:32
	git push pushes all like-named branches	16:32
	→ hannes joined	16:33
hannes	I'm going to import an subversion repository. first trunk, then I want to import branches, but these have not branched the complete trunk; somehow I wasn't able to find documentation how to do that properly (http://ivanz.com/2009/01/15/selective-import-of-svn-branches-into-a-gitgit-svn-repository/ was the best I could find)	16:34
	→ Mycr0ft joined	16:35
hal	How can a remote tracking branch be created in the fewest steps using the local command line, please?	16:35
	I always seem to have to create a testfile to commit and push before it will be created, and also run config commands so that it knows which is the correct remote	16:35
	isn't there a quicker way of doing this?	16:35
	→ corni joined	16:36
thiago_home	hal: git branch localbranch remotename/remotebranch	16:37
hal	thiago_home: hi and thank you....	16:38
	→ Yuuhi joined	16:38
hal	thiago_home: should the remotebranch be an existing branch, or what the new remote branch name will be?	16:38
thiago_home	it should be an existing branch	16:38
	→ ChUrP joined	16:39
hal	thiago_home: ok, and then what are the next steps, so that the new branch (currently local) can be seen by others?	16:39
	→ ThiefMaster` joined	16:39
	← ThiefMaster` left	16:39
	→ ThiefMaster` joined	16:39
	ThiefMaster` → ThiefMaster	16:39
amitsamtani	thaigo_home: so, if i create a new branch, create a new file, commit, then push... it will not push? how do i get my new branch to the repo?	16:40
hal	amitsamtani: err, are you hijacking our conversation?! ;-p	16:41
amitsamtani	hal: sorry about that	16:41
hal	amitsamtani: it seems that you have the exact same problem as me	16:43
amitsamtani	hal: ah, i just read ur messages, yea i think we do have the same problem.	16:44
	→ Tuller joined	16:44
amitsamtani	hal: i thought if i create a branch and push it, it would be pushed into the repo... but even when i create a file... its not getting pushed to the repo	16:44
	→ zimbres joined	16:45
hal	amitsamtani: yes, well, you could manually adjust the config files, by running the correct command, but I am sure there is a better way	16:45
	← dreiss left	16:45
	← hannes left	16:48
amitsamtani	hal: would you know why this would not work? git branch newbranch git checkout newbranch git touch newfile git commit -a -m "blah" git push .... am i missing something?	16:48
hal	thiago_home: can you help any further with this?	16:48
	→ radarek joined	16:48
hal	amitsamtani: you aren't specifying the remote branch that it is tracking	16:49
	→ Yuffster joined	16:49
hal	amitsamtani: as thiago_home said, git branch localbranch remotename/remotebranch	16:49
	or alternatively, if you are in master...	16:49
	amitsamtani: git checkout -b localbranch remotename/remotebranch	16:50
	which will switch to your new branch too	16:50
amitsamtani	right now i think i only have one branch... master... how would i create a new branch in the repo?	16:50
thiago_home	hal: sorry, was busy	16:51
	hal: you have a local branch and you want to share with other people?	16:51
	hal: push it	16:51
hal	thiago_home: it gives an error if I push it...	16:51
thiago_home	what error?	16:51
hal	thiago_home: sorry, not an error - it says, everything is up to date	16:52
	and even after running this command, the branch is not yet available to others	16:52
	on the remote	16:52
thiago_home	hal: "git push" only pushes like-named branches	16:53
	hal: if you have just created this branch locally, then by construction it doesn't exist in the remote	16:54
	so "git push" won't push it	16:54
amitsamtani	thaigo_home: 1, how do you create a branch remotely and 2, if you messed up and created a branch locally, how can we fix it and create it remotely as well now?	16:55
thiago_home	you need to push it to the remote	16:56
	you need to supply two more arguments to git push	16:56
amitsamtani	"git branch localbranch remotename/remotebranch"	16:57
thiago_home	2) your question doesn't make sense	16:57
amitsamtani	and i can specify whatever i want to remotename/remotebranch?	16:57
thiago_home	that creates a local branch called "localbranch", tracking the remote branch "remotebranch" from remote "remotename"	16:57
	yes, you can	16:57
amitsamtani	by doing this: git branch acl_aro_aco develop/acl_aro_aco i get "fatal: A branch named 'acl_aro_aco' already exists."	17:00
thiago_home	sounds self-explanatory	17:00
	do you want to change the branch that acl_aro_aco tracks, withtout deleting it?	17:00
hal	thiago_home: I am sorry, I don't understand. I have run git branch localbranch remotename/remotebranch. How do I make remotebranch exist on the remote?	17:00
thiago_home	hal: push it to the remote server	17:01
	hal: you need to specify the two arguments to git push	17:01
	hal: notice I'm not telling you what those arguments are. I expect you now to go read the manual.	17:01
hal	thiago_home: ok, I see, but could you be explicity about the git push command,?	17:01
	hmm	17:01
thiago_home	I could, but I won't	17:01
	← madewokherd left	17:04
	→ MicahElliott joined	17:05
hal	thiago_home: is it git push origin localbranch ?	17:05
thiago_home	yes	17:06
	→ mmw joined	17:07
amitsamtani	hal: thanks hal, thaigo_home	17:07
	→ khelll joined	17:08
amitsamtani	hal: just stumbled across this link too: http://www.gitready.com/beginner/2009/02/02/push-and-delete-branches.html	17:09
	→ ThiefMaster` joined	17:10
	← ThiefMaster` left	17:10
	→ ThiefMaster` joined	17:10
	← ThiefMaster left	17:10
	→ marlun joined	17:11
	→ rseifert joined	17:11
marlun	I'm trying to change branch to master but I am getting an error that I've got local changes. I thought I could just change branch and the changes will be left so that I can add them to the new branch?	17:12
thiago_home	that's almost correct	17:12
	you can't change branches if that would change a file with uncommitted changes	17:12
	you can use either the -m option to checkout, or you should stash	17:12
marlun	Is that because I added and commited part of the file in this branch?	17:13
	I used add --patch	17:13
thiago_home	you didn't commit	17:13
marlun	I did commit the part I added with add --patch.	17:13
	All of the changes I've got left is "Changed but not updated"	17:13
thiago_home	commit	17:14
	← radarek left	17:14
marlun	I have commited. The changes that is left I want to commit on the master branch.	17:14
thiago_home	it's what I said	17:15
	either use -m or stash	17:15
marlun	thiago_home: ok, thanks :)	17:16
hal	thank you for your help thiago_home - you were a great help! :)	17:16
	→ radarek joined	17:17
	← amitsamtani left	17:19
	→ amitsamtani joined	17:19
	← amitsamtani left	17:20
	→ amitsamtani joined	17:20
	← amitsamtani left	17:20
	← khmarbaise left	17:20
	→ OFICINA7 joined	17:21
	← radarek left	17:21
	← MicahElliott left	17:22
	→ froschi joined	17:25
	← yann left	17:28
	→ lowik joined	17:28
	→ amitsamtani joined	17:31
	→ comawhite joined	17:31
	→ liuchong joined	17:31
liuchong	some one can speak chinese?	17:32
	?	17:32
	← liuchong left	17:33
	← jmcantrell left	17:35
	← airborn left	17:35
	→ radioman-lt joined	17:35
	→ madewokherd joined	17:35
	→ airborn joined	17:35
	→ oriba joined	17:36
	→ Axius joined	17:36
	← oriba left	17:36
	→ przemoc joined	17:36
	← comawhite left	17:37
	← radioman-lt left	17:37
	→ comawhite joined	17:37
scatterp	[18:31:52] <liuchong> some one can speak chinese?	17:38
	[18:32:09] <liuchong> ?	17:38
	that was quick lol	17:38
zomg	Yeah why didn't you answer! You bastard!	17:39
	This is IRC	17:39
	You are supposed to react to every whim of the users, instantly!	17:39
	I does stand for Instant afterall	17:39
	;>	17:39
	→ radarek joined	17:46
	→ paja joined	17:47
	← hal left	17:50
tango_	be fair, he waited ONE WHOLE MINUTE before leaving	17:54
	← MattDiPasquale left	17:54
	→ lyosha joined	17:54
lyosha	Hi. Is there a way to tell git-svn to clone starting from a specific revision?	17:55
patrikf	lyosha: see man git-svn, --revision	17:56
	→ joelpet joined	17:56
	← radarek left	17:57
lyosha	patrikf: great, that worked, thanks!	17:59
	← icefox left	17:59
	→ psino joined	17:59
	→ mastro joined	17:59
	→ patmaddox joined	18:00
	← amitsamtani left	18:02
	← lyosha left	18:03
LauJensen	I have a branch with messy history. I want to walk each commit and cherry-pick a few to go into a new branch. Can somebody give me the exact command(s) to do that ?	18:07
	← vu3rdd left	18:07
	← hyperair left	18:08
	← tar- left	18:09
	→ neutrino joined	18:09
	neutrino → tar-	18:09
	→ hyperair joined	18:09
	→ Mage joined	18:09
rane_	so you want to completely get rid of some changes you've done?	18:10
LauJensen	rane_: Not really. I have one branch which contains 100 commits. 30 needs to go into 1 new branch, 20 into another, and the last 50 can just float around for historic purposes	18:10
rane_	well	18:11
	git checkout -b new-branch and interactive rebase	18:11
LauJensen	rane_: oddly enough, that just took 1 commit, put me in vi (eww) and then when I deleted the first line, simply stopped rebasing	18:12
rane_	I guess you'd have to give it something like git rebase -i HEAD@{100}	18:12
LauJensen	oh, is there a short notation for ALL instead of 100 ?	18:13
rane_	iirc there's no way to point to the "first" commit	18:13
	← hyperair left	18:13
LauJensen	k	18:14
	→ mw joined	18:15
	→ radarek joined	18:20
	← radarek left	18:21
	← jensn left	18:22
	← bdimcheff left	18:25
	← hohoho left	18:26
	← patmaddox left	18:27
	→ nsussman joined	18:28
	← nsussman left	18:28
	→ fundamental joined	18:30
	← khelll left	18:30
	→ apsysoev joined	18:36
	→ pennyfx joined	18:36
pennyfx	hi	18:37
	i'm a noob to git	18:37
	→ dj_tjerk joined	18:38
pennyfx	i have a few questions	18:38
	← Axius left	18:38
pennyfx	i made some changes to some code in a local git repo.. i didn't check anything in.. i just want to revert back .	18:38
	how can i do that?	18:38
rane_	git reset --hard HEAD	18:39
	you lose everything you haven't committed	18:39
pennyfx	even new files?	18:39
	→ radioman-lt joined	18:39
rane_	ot those	18:39
	+n	18:39
pennyfx	k	18:39
rane_	my bad	18:39
	→ hyperair joined	18:40
pennyfx	thanks it worked	18:41
	← apsysoev left	18:42
	→ apsysoev joined	18:42
	→ RobertLaptop joined	18:42
	→ Heimidal joined	18:43
	← hyperair left	18:44
	→ Pewpewarrows joined	18:45
	→ tvw joined	18:45
	→ dl9pf joined	18:46
	← tar- left	18:48
	→ tar- joined	18:49
	→ dreiss joined	18:49
	→ MattDiPasquale joined	18:51
	→ yann joined	18:54
	← tvw left	18:54
	→ hyperair joined	18:55
	→ froschi1 joined	18:55
	→ giskard joined	18:57
	← froschi left	18:58
	→ francisfish_away joined	19:00
	← dreiss left	19:00
	← Tuller left	19:01
	→ alexyk joined	19:02
	→ Gitzilla joined	19:05
	→ seanius joined	19:05
	→ flaguy48 joined	19:06
	← alexyk left	19:13
	→ Louie` joined	19:14
	→ aresnick joined	19:16
	→ hohoho joined	19:21
	→ Jaxan joined	19:22
	→ amitsamtani joined	19:22
	→ __ed joined	19:23
	→ ajpi joined	19:26
LauJensen	Its amazing that this has to be explicitly configured, but how do I avoid getting thrown into vi every time I rebase ?	19:27
thiago_home	how would git guess which editor you prefer if you didn't configure?	19:27
LauJensen	thiago_home: Git should know that intelligent people would like something other than vi :)	19:28
	→ NfNitLoop joined	19:28
cbreak	it uses vim because of that	19:28
thiago_home	Git, like any other Unix tool, defaults to vi because that's what they all do	19:28
	it's the only editor that one can always expect to find, even in older and exotic Unix	19:28
cbreak	it uses your $EDITOR if you haven't configured core.editor or similar :)	19:28
	LauJensen: man git-config	19:29
	← flaguy48 left	19:29
	← scarabx left	19:29
	→ tpope joined	19:30
	→ kandinski joined	19:30
	← amitsamtani left	19:30
wereHamster	or $GIT_EDITOR or $VISUAL	19:31
LauJensen	export EDITOR worked just fine	19:31
	ajpi → ajpiano	19:32
	← NfNitLoop left	19:32
	→ khelll joined	19:34
	→ coderdad joined	19:34
	→ tulcod joined	19:37
	← MattDiPasquale left	19:37
	← mastro left	19:38
	→ lanthan joined	19:41
	→ mastro joined	19:41
	→ xiong joined	19:41
	→ sftp joined	19:42
	← RobertLaptop left	19:46
	→ Titosemi_ joined	19:47
	← Titosemi left	19:48
	Titosemi_ → Titosemi	19:48
	→ RobertLaptop joined	19:48
	← notjohn left	19:49
	→ amitsamtani joined	19:51
	← amitsamtani left	19:51
badboy_	l	19:51
fundamental	Is there some way to do a partial merge between branches (ie merge commits before x from another branch)?	19:53
cbreak	just merge the sha	19:55
	→ shruggar joined	19:55
	← ziro` left	19:57
fundamental	cbreak: thanks, that seems to do it	20:00
	→ p3rg2z joined	20:02
	← rseifert left	20:02
	→ little_owl joined	20:05
	← henkboom left	20:12
	← lowik left	20:13
	← little_owl left	20:15
	→ henkboom joined	20:16
	→ icefox joined	20:19
	← rajeshsr left	20:21
	← tulcod left	20:21
	→ eighty4 joined	20:21
	← raichoo left	20:22
	← harinath left	20:23
	→ mtkd joined	20:25
	← tar- left	20:26
	← pcwick left	20:26
	→ neutrino joined	20:27
	neutrino → tar-	20:27
	→ dreiss joined	20:27
	← comawhite left	20:30
	← d0k left	20:32
	← eighty4 left	20:33
	→ eighty4 joined	20:34
	← corni left	20:36
	→ amitsamtani joined	20:37
	← amitsamtani left	20:38
	→ harinath joined	20:39
	→ engla\|swe joined	20:40
	← jrmuizel left	20:43
	← Heimidal left	20:45
	→ oficina7_ joined	20:46
	→ chrislerum joined	20:47
	→ parndt_a_ joined	20:47
	← chrislerum left	20:48
	← OFICINA7 left	20:48
	← tar- left	20:49
	← p3rg2z left	20:49
	→ neutrino joined	20:49
	neutrino → tar-	20:49
	→ p3rg2z joined	20:50
	← hohoho left	20:51
	→ MattDiPasquale joined	20:55
	→ bx2 joined	20:55
	→ Aaaarg joined	20:58
	→ rryouumaa joined	20:58
	→ jasonn1234 joined	20:59
	→ thunk\|away joined	21:01
	← jasonn1234 left	21:02
	→ offby1` joined	21:02
	offby1` → offby1	21:02
	← offby1 left	21:02
	→ offby1 joined	21:02
	thunk\|away → thunk	21:02
	← mw left	21:03
	→ tvw joined	21:03
	← shruggar left	21:06
	→ masterkorp joined	21:07
	→ jrmuizel joined	21:08
	oficina7_ → OFICINA7	21:09
	← jrmuizel left	21:09
	← marlun left	21:10
	← ChUrP left	21:10
	→ jrmuizel joined	21:14
	→ Snow-Man joined	21:17
	→ Adaptee joined	21:18
	→ shadowmaster joined	21:20
	← Titosemi left	21:21
	→ andreaa joined	21:22
	→ marc_kdab joined	21:22
	→ pantsman- joined	21:23
	← pantsman- left	21:23
	→ chrislerum joined	21:25
	← LauJensen left	21:25
	→ bodom joined	21:26
	→ Tuller joined	21:28
	→ MattDiPa_ joined	21:29
	← Jaxan left	21:29
	← Tuller left	21:29
	← MattDiPasquale left	21:29
	→ neurodamage joined	21:33
neurodamage	with the git archive command is there anyway to specify a subtree of folders in the project instead of archiving the whole thing?	21:33
mastro	neurodamage, I don't think so since git do not work that way :)	21:34
patrikf	neurodamage: yes, read the manpage	21:34
	→ yairchu joined	21:36
	← Mycr0ft left	21:36
	← aresnick left	21:36
	→ erikvold joined	21:37
	→ flaguy48 joined	21:38
mastro	neurodamage, patrikf uh... well :) i was wrong good to know	21:40
	← yairchu left	21:42
	← bx2 left	21:44
	← radioman-lt left	21:45
	→ AAA_awright joined	21:45
	← AAA_awright left	21:46
	→ Jaxan joined	21:46
	→ yacoob joined	21:46
yacoob	F1. I've cloned same repository out on two different hosts - and I see different histories for the same branch.	21:47
	← tizbac left	21:49
	→ hobodave joined	21:50
	← mmw left	21:50
	→ tizbac joined	21:50
	← tizbac left	21:50
	← engla\|swe left	21:51
	← jrmuizel left	21:55
	→ jrmuizel joined	21:56
	← jrmuizel left	21:57
rudi_s	yacoob: Maybe somebody pushed to that branch after you cloned the first repo.	21:57
	If you just copied the repository that can't happen.	21:57
yacoob	rudi_s, nope. I suspect the problem was outside git.	21:58
	(was accessing the repository via smb, and somehow got stuck at specific version of filesystem...)	21:58
	← chrislerum left	21:59
	→ patmaddox joined	22:01
yacoob	hm, could have forced inode cache to clear, instead of rebooting	22:02
	← Jaxan left	22:03
	→ eighty4_ joined	22:04
	← lanthan left	22:04
	← eighty4 left	22:06
	← tar- left	22:08
	← neurodamage left	22:09
	→ ales joined	22:12
	→ jrmuizel joined	22:14
	← harinath left	22:16
	← mastro left	22:17
	→ aliceinwire joined	22:18
	→ priidu joined	22:19
	→ notjohn joined	22:21
	← patmaddox left	22:23
	→ patmaddox joined	22:23
	→ engla\|swe joined	22:24
	← mtkd left	22:27
	← patmaddox left	22:28
	→ harinath joined	22:32
	← icefox left	22:32
	← giskard left	22:33
	→ comawhite joined	22:35
	→ krawek joined	22:35
	→ EricInBNE joined	22:35
	→ jensn joined	22:37
offby1	networked file systems scare me	22:37
	they're ok for editing a single file in my editor, but running any program like git across them ... just gives me the willies	22:38
	← ales left	22:41
	→ darkestkhan joined	22:43
drizzd	offby1: there are certain version control systems which are based on the idea of only using through a network file system	22:44
	because it's "fast"	22:44
offby1	only such I can think of is Visual Source Safe	22:44
patrikf	networked file systems aren't bad per se...	22:45
offby1	not bad; just willies-inducing	22:45
drizzd	offby1: I was talking about clearcase	22:45
offby1	well, presumably clearcase was _designed_ to run over one, so I assume it works fine.	22:46
	I mean things like git over SMB.	22:46
	→ jshsu joined	22:46
drizzd	no, clearcase has its own file system	22:46
patrikf	offby1: I don't see what's wrong with that	22:46
	offby1: unless you're suggesting mutliple people work within one working tree	22:46
drizzd	which you use as "working copy"	22:46
offby1	patrikf: well, I have no specific problems. But unless the docs clearly say "Yeah, we use this over SMB all the time; it works great", I'd be wary.	22:46
	→ homiziado joined	22:47
drizzd	this allows you to "check out" files in zero time	22:47
offby1	no, no, I wasn't talking about multiple people sharing one work tree. (That's madness!)	22:47
drizzd	neither am I	22:47
offby1	patrikf was	22:47
drizzd	oh	22:47
patrikf	no, I thought offby1 might be ;-)	22:47
	because that'd scare me too :D	22:48
offby1	anyway: if I tell git my repo is local (by cloning from a path like /foo/bar rather than host:/foo/bar), and yet it's actually an SMB-mounted file system ... _that's_ the scenario that worries me	22:48
	→ stringo0 joined	22:48
patrikf	I think git properly uses file locking.	22:50
offby1	I always assume stuff like file locking doesn't really work over SMB	22:51
	parndt_a_ → parndt	22:54
	← darkestkhan left	22:55
	← xiong left	22:58
	← aliceinwire left	22:58
	→ mw joined	23:00
	← mw left	23:00
	→ mw joined	23:00
	→ aliceinwire joined	23:01
	→ psoo joined	23:06
	→ FFForever joined	23:06
FFForever	Where would I place a hook to only affect 1 repository? (If it matters I am using gitolite)	23:07
wereHamster	in the repo's hook directory	23:08
offby1	would you have to log in to the server to fiddle that?	23:09
	→ icefox joined	23:09
wereHamster	yes	23:09
FFForever	does it matter what language a hook is written in?	23:10
wereHamster	no	23:11
	FFForever: man githooks	23:11
FFForever	No manual entry for githooks	23:12
	← masterkorp left	23:12
wereHamster	FFForever: install the git man pages. In the meantime, use http://www.kernel.org/pub/software/scm/git/docs/githooks.html	23:12
	← coderdad left	23:15
parndt	hey, I'm trying to merge in some commits from another fork but want to only merge up to a specific commit -- what's the best way to do this for a clean history (presumably avoiding cherry-pick?) - I want to pull around 5 commits.	23:16
andres	parndt: simply do git merge commit-id	23:17
	→ luke-jr joined	23:17
luke-jr	any way to find out the progress of a running git pack-objects?	23:18
	← paja left	23:18
luke-jr	eg, find out a % this 6 hours CPU time is	23:18
davr	6 hours? do you have like 10GB git repo or something?	23:20
parndt	andres: thanks :) that works..	23:22
	→ Bass10 joined	23:22
	← stringo0 left	23:22
	← airborn left	23:23
	→ stringo0 joined	23:24
	← Bass10 left	23:24
	→ Bass10 joined	23:25
	← p3rg2z left	23:25
	← Adaptee left	23:25
luke-jr	davr: far bigger I think	23:25
	→ alsuren joined	23:26
luke-jr	one is 9.6 GB	23:27
andres	Depends hugely on what window/depth parameter are used I would say...	23:27
parndt	andres: is there any way to give merge a start position to merge from too? like start at this commit end at this commit	23:28
davr	parndt: git merge doesn't work like that. you have to merge an entire branch, if you just want some of the commits then you'll have to do cherry-pick or rebase	23:28
parndt	right..	23:29
	by rebase do you mean undo the commits after merging?	23:29
luke-jr	parndt: AFAIK, only Darcs and Subversion really support cherry-picking	23:29
	hmm, /proc/19987/io says 8.8 GB done XD	23:29
davr	parndt: no, don't merge at all, only rebase	23:30
parndt	my problem with cherry picking is it messes up history, right?	23:30
luke-jr	davr: FWIW, my other repo is 17 GB	23:30
davr	luke-jr: stop putting huge binaries in git? :P	23:30
luke-jr	parndt: what you're asking is the definition of cherry picking	23:30
	davr: 100% text	23:30
andres	luke-jr: I wouldnt really consider svn's "support" as really supporting it	23:30
luke-jr	well, at least 99.9%	23:30
parndt	okay thanks I'll read up on rebase some more. bbl	23:31
	parndt → parndt_away	23:31
luke-jr	andres: recording it is for the most part support	23:31
	→ brizly joined	23:31
luke-jr	andres: if you have a better post-cherrypick merge algo than Subversion uses, go for it ;)	23:31
	cherry picking is a matter of recording; merging past one is a matter of merge algorithms	23:31
andres	luke-jr: well, its more their recording that sucks if my memory serves right (i.e. merge across more than one branch and it gets interesting)	23:32
	→ t0rc joined	23:32
	→ leavittx joined	23:32
luke-jr	davr: also, at least 99% under 4 KB :p	23:33
	← Louie` left	23:35
	→ nfrs joined	23:35
	→ Louie` joined	23:35
davr	sounds like you might be better off with a database or something	23:36
luke-jr	davr: unlikely. why?	23:37
	→ phantomcircuit joined	23:38
mugwump	11:29 < luke-jr> parndt: AFAIK, only Darcs and Subversion really support cherry-picking	23:38
davr	I mean, I'd hate to clone that repo, would take ages. It's definitely not what was thought of when git was designed. But I guess if it works for you, then great	23:38
mugwump	^^ fallacy of the day :D	23:38
luke-jr	davr: works better than anything else, so far XD	23:38
mugwump	I know of several others that support it well	23:38
luke-jr	davr: git seemed ideal because everything else scans the checkout for commits	23:38
	whereas git lets me specify the exact files	23:39
davr	I mean it doesn't sound like the kind of thing that belongs in source control in general (I'm assuming it's not 4.5 million source code files)	23:40
	← Louie` left	23:40
nfrs	I have two branches: "local", where all development occurs, and "testing", which contains small patches needed to run the program on another machine. "testing" should always get the same commits that go into "local". what is the best way to maintain such configuration? obviously, I want to avoid the need to merge each commit into testing by hand	23:40
	← icefox left	23:40
luke-jr	davr: nah, it doesn't need branching, but nobody makes a compressed history :p	23:40
davr	nfrs: I would probably rebase testing onto local, assuming the changes on testing are small and unlikely to conflict with anything in local	23:40
nfrs	also, I want to avoid the situation where I'll forget to merge the changes into testing and encounter an already fixed bug	23:40
	davr: I tried that, but it doesn't work well. probably because the testing branch is shared with another developre	23:41
	← alsuren left	23:41
luke-jr	davr: the 17 GB repo contains over 684,000 files	23:41
	the working copy is 8.1 GB	23:42
	strange, I seem to remember the wc being bigger...	23:42
davr	nfrs: then merging is probably the easiest. you could probably have a simple post-commit hook that attempts merging from local into testing	23:43
	something along the lines of if [ "$current_branch" == "local" ]; git checkout testing && git merge local && git checkout local	23:43
	if the merge fails, it'll still be on the testing branch waiting for you to manually resolve it, else it will automatically switch back to local	23:44
luke-jr	(in the meantime, all the timestamps are touched and your next build recompiles everything -.-)	23:44
nfrs	davr: I think that won't work well, either, since we rebase our local changes when pulling. so I'm uncertain what happens, when: a) main branch commit occurs; b) the commit gets pushed to the testing branch; c) pull --rebase rewrites history	23:44
davr	so you can't merge into testing, you can't rebase testing, you're not giving me many options here :P	23:46
nfrs	well, if there were an easy solution, I wouldn't be here, in the first place :)	23:46
davr	Maybe you just need to not push every single commit to testing one by one, and instead adopt a strategy where you only test on the separate machine every so often?	23:47
	then when you do it you can make sure that your local history wont get rebased in the future	23:47
	eg only update remote testing branch when you push to the remote development branch	23:48
	← mw left	23:49
nfrs	I'm not quite following	23:50
	"wont get rebased in the future" - ?	23:50
	→ AAA_awright joined	23:50
	→ chrislerum joined	23:50
	← engla\|swe left	23:52
nfrs	right now, the only solution I see is to pull changes from a post-push hook	23:53
	→ Tureba joined	23:54
nfrs	that will ensure that there are no local changes that will be rebased	23:54
	← chrislerum left	23:56
	shade_ → shade\	23:56
nfrs	but that would make the remote branch less stable, as I will be forced to push before testing	23:57
	and I will probably forget to push once in a while	23:58
	→ locklace joined	23:59

Logs Search ←Prev date Next date→

Channels Documentation