IRCloggy #git 2012-01-21

Provider of IRC logs since 2005.

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

	2012-01-21
FauxFaux	Oh, @git_base_url_list is a different variable, how useful.	00:00
	→ pdtpatrick_ joined	00:00
	variable is the sam	00:00
	← pdtpatrick_ left	00:00
Danny_Joris	cmn: that was it. I merged with my local 6.x branch and then it worked	00:00
	← tystr left	00:01
	→ lonewulf` joined	00:01
cmn	hm? that doesn't sound like it worked, it sounds like you worked around telling git what to push	00:01
	→ tommyvyo joined	00:01
	→ p3rror joined	00:02
	→ soulcake joined	00:03
	← matthiasgorgens left	00:06
	→ osmosis joined	00:07
	→ kenichi joined	00:08
ripero	FauxFaux: I couldn't see how git_base_url_list would solve the problem...	00:08
	→ matthiasgorgens joined	00:09
FauxFaux	It's entirely unrelated.	00:09
	← fly9 left	00:09
ripero	FauxFaux: ok	00:09
	← Guest5662 left	00:09
	← Vile left	00:10
	→ ricky_ joined	00:10
ripero	FauxFaux: I looked at the examples of the documentation, and they seem to solve this by defining a DocumentRoot, which I think I can't do since I am not running gitweb as a virtual server	00:10
FauxFaux	I run it both on the root of a domain, and on lighttpd, which has rewrite rules that actually make sense, so I can't help any further. :)	00:11
	ricky_ → ricky	00:12
	← beatak_m left	00:12
ripero	ok, sorry for insisting... I think I found my way through it, I just wanted to try that future users won't have the same problem...	00:13
	← mattly left	00:15
ripero	FauxFaux: ^ thanks a lot for all	00:15
	:)	00:15
	← scelis left	00:16
	→ techhelp joined	00:17
	← fr0stbyte left	00:19
	← kmc left	00:20
	← pantsman left	00:20
	← Carmivore left	00:20
	← giallu left	00:21
	← kenichi left	00:21
	← matthiasgorgens left	00:23
	← tommyvyo left	00:23
	→ Carmivore joined	00:23
	← ipalaus left	00:25
	← QaDeS left	00:25
	→ matthiasgorgens joined	00:26
	← jceb left	00:26
	← Danny_Joris left	00:26
	→ hal joined	00:30
hal	I have pulled someone else's branch into my branch by performing git pull theirrepo theirbranch	00:30
	→ notjohn joined	00:30
	← neurodrone left	00:31
hal	in my log, I have one commit, where the branch is merged together, which is now the HEAD	00:31
	← macUzer left	00:31
hal	could someone tell me the best way to make the commit before last the head, please?	00:31
FauxFaux	hal: !fixup	00:33
gitinfo	hal: So you made a bad commit and want to remove/fix it? Look at https://gist.github.com/1612395 for full instructions. Hints are: (1) NOT PUSHED/PUBLISHED: `git rebase -i $COMMIT^` or perhaps `git commit --amend` (or `git reset HEAD^`). (2) OTHERWISE, `git revert $COMMIT` to make a reverse commit. (3) If you have pushed and MUST remove it, use rebase or filter-branch and type !rewriting_public_history in IRC.	00:33
	← YesJustWolf left	00:33
	→ t-mart joined	00:34
	← baburdick left	00:35
hal	FauxFaux: wow, this gist is bloody brilliant!!! Thank you! :)	00:36
FauxFaux	/nick SethRobertson2	00:36
SethRobertson	heh	00:36
	My minions are assembling	00:37
	→ baburdick joined	00:37
t-mart	im writing my resume in LaTex and keeping versions with git. I want to push this repo to github. Generally, I don't want to push the resume.pdf to github because its binary/takes up too much space/not good for version control/etc. but I would like to tag some important releases that get pushed for linking from my github pages site. how should i manage this?	00:37
SethRobertson	git tag?	00:37
FauxFaux	t-mart: Technically you can tag files, but people don't really understand. Ignore it, and add -f it for specific commits?	00:38
kevlarman	FauxFaux: once you add it, it's no longer ignored	00:38
	→ scelis joined	00:39
SethRobertson	You could form a !superproject with some bits being sent to github and others not? It isn't exactly clear what you want yet	00:39
	!subproject?	00:39
gitinfo	[!subprojects] So, you want to add git repositories inside of other git repositories? Well, you have four main options. First is to just do it, add the repo to the outer project's .gitignore, and treat them entirely separately. Best if they are entirely seperate. Otherwise your best options are "!submodule" "!gitslave" and "!subtree" Try typing those commands into this IRC channel.	00:39
FauxFaux	Yeah. It's a shame there's no add --cached, or something.	00:39
	← psoo left	00:39
t-mart	ok, i've also .gitignore-d .pdf files...will i have to un-ignore that pdf for the tagged version, and then re-ignore it?	00:40
SethRobertson	That will push it out to github	00:40
	Once you add, it is there forever	00:40
	However, you could add it under a different name than the one your produce for testing	00:40
	ignore *.pdf and have make generate resume-working.pdf but `git add -f resume.pdf`	00:41
	← robotmay left	00:41
	← CannedCorn left	00:42
	← dankest left	00:42
	→ dankest joined	00:42
	← ramsey left	00:43
t-mart	thanks guys. im going to mull this over.	00:44
	→ mgpcoe1 joined	00:47
t-mart	ah, i think i'll just use github's downloads. more manual, yes, but also pretty uncomplicated.	00:47
	→ glennpratt joined	00:49
SethRobertson	Um, how does that solve the problem?	00:49
t-mart	its online storage that's not tracked	00:50
	← mgpcoe left	00:50
SethRobertson	OK, have fun	00:50
	← tvw left	00:50
t-mart	thanks =)	00:51
	← centipedefarmer_ left	00:52
	← Goplat left	00:52
	← joshteam left	00:53
	← bitkiller left	00:55
	← Karmaon left	00:55
	← ap3mantus left	00:56
	→ macabre joined	00:56
	→ Grafica joined	00:56
	← glennpratt left	00:56
	← yhager left	00:56
	← steffo left	00:57
	← duckxx left	00:58
	→ Danny_Joris joined	01:00
	→ fayimora joined	01:00
	→ wroathe joined	01:02
	→ apok_ joined	01:03
	← wroathe left	01:03
	← nicxvan left	01:03
	← mattcen left	01:03
	← dankest left	01:03
	← apok_ left	01:04
	← apok left	01:04
	→ apok_ joined	01:04
hal	am I allowed to ask github questions in here?	01:05
FauxFaux	You can try. There is a #github.	01:05
	← baburdick left	01:05
	← gusnan left	01:06
	← techhelp left	01:06
	← oriba left	01:06
	→ baburdick joined	01:07
hal	oh I thought I'd tried that and there was noone there. I will try again. but just a quickie - does anyone know the quickest way from the pull request page to find the branch name that the commit comes from of the contributor's repo?	01:07
	→ tatsuya_o joined	01:07
hal	I hope that made sense! ;)	01:07
FauxFaux	man git branch --contains= one of their commits?	01:08
gitinfo	the git-branch manpage is available at http://jk.gs/git-branch.html	01:08
	→ iamjarvo joined	01:08
hal	hmm, yeah I never thought of actually using git for that! ;)	01:08
	thank you FauxFaux :)	01:08
FauxFaux	It don't believe it necessarily hasn't have to have come from a current branch.	01:08
	+English.	01:08
	→ borodon joined	01:08
hal	what do you mean by that FauxFaux ?	01:09
	→ jonesy joined	01:10
	→ yhager joined	01:10
	← airborn left	01:10
	← apok_ left	01:10
	← baburdick left	01:11
	→ baburdick joined	01:11
hal	FauxFaux: aha, yes, thank you, it's git branch -r --contains=	01:11
	→ neurodrone joined	01:13
	→ JackWinter2 joined	01:14
	← JackWinter left	01:16
	← ehsan left	01:18
	← Chillance left	01:19
	→ oriba joined	01:20
	→ traviscj joined	01:21
	← pl0sh left	01:22
	→ Vile joined	01:23
	← mgpcoe1 left	01:25
	← ScottE left	01:27
	→ t0rc joined	01:27
	→ Tens[a]i joined	01:27
	← Tens[a]i left	01:28
	← maletor left	01:29
	→ apok joined	01:29
	→ Green1 joined	01:30
	← mandric left	01:30
	← RaptorX left	01:30
	→ Goplat joined	01:32
	← friskd left	01:32
	← afief left	01:33
	→ vervic joined	01:34
	→ gchristensen joined	01:34
	← baburdick left	01:36
	← Green1 left	01:36
	→ baburdick joined	01:36
	→ Green1 joined	01:37
	ChanServ set mode: +v	01:38
Grafica	I am just learning Git. The other day, I added files, but did not make any changes to them. Now I want to start working on them, but received a fatal error (Not a Git repository or any of the parent directories) when I checked the status. Then, I attempted to add files and received the same error.	01:38
offby1	Grafica: sounds like your current directory is wrong	01:39
	Grafica: a transcript would be helpful	01:39
	← Danny_Joris left	01:40
Grafica	How do I generate a transcript?	01:40
	→ ehsan joined	01:41
	→ glennpratt joined	01:41
Grafica	I added the files the other day.	01:42
	← mabrand left	01:42
mjago	Grafica: what is the error message?	01:43
	← icwiener left	01:43
	→ icwiener joined	01:43
	→ aidenhong joined	01:44
Grafica	Fatal: Not a Git repository or any of the parent directories: .git	01:44
offby1	Grafica: if you can cause the problem to happen again, then you can copy your terminal session and paste it to a paste site, so we can see it.	01:45
	Then we say "aha!"	01:45
	you did X when you should have done Y.	01:45
	otherwise we're just guessing.	01:45
Grafica	How do I copy it, do you mean make a screen shot?	01:47
	← martinjlowm left	01:47
FauxFaux	Grafica: Copy-paste it into http://pastie.org/	01:47
	← tatsuya_o left	01:48
	← ehsan left	01:48
Grafica	How do I copy it? The words don't select.	01:48
	→ tatsuya_o joined	01:49
FauxFaux	Git bash on Windows? RIght-click, edit.	01:49
mjago	Grafica: someone already said it - CD into root directory of the project and try again	01:49
sitaram	mjago: I think "root directory" might be hard for someone who doesn't know how to copy-paste... (I dont intend any offense)	01:49
mjago	sitaram: no offense taken	01:49
SamB	mjago: I don't think you were the one he didn't intend to offend ;-P	01:50
sitaram	mjago: I meant "I dont intend any offense to the person who can't copy-paste by mentioning it as if it were a really basic skill, which actually it is" ;-)	01:50
	← johnhamelink left	01:50
SamB	copying and pasting from Windows consoles is a little tricky, to be sure	01:50
offby1	SamB: it's easier if you fiddle cmd.exe to turn on "quickedit"	01:51
mjago	sitaram: SamB: we all start somewhere that's for sure	01:51
offby1	Grafica: anyway, a screen shot would be fine for now.	01:51
	Grafica: learn to select, copy, and paste from that console though; it's a useful skill anyway	01:51
mjago	i have to use windows when I do embedded stuff	01:52
sitaram	SamB: you may be right. I thought it could be done entirely with the mouse, but maybe I am mistaken	01:52
	← Green1 left	01:52
SamB	sitaram: I think it can be, but that's not the way most Windows users are used to doing it, as I understand it	01:53
	my understanding is that the expected method is (1) drag over text to select it (2) type Ctrl-C to copy it	01:53
mjago	SamB: Not on the command line?	01:54
FauxFaux	Return to copy! Ctrl+c would be too easy.	01:54
SamB	mjago: I mean the method users would expect to work	01:54
sitaram	SamB: I thought you right click, select "mark" etc... but again, it's been a while since I saw Windows (even at work!)	01:54
SamB	sitaram: that's the way you actually do it on the console, yes	01:54
	→ ap3mantus joined	01:55
	← LekeFly left	01:55
	← ap3mantus left	01:55
Grafica	offby1: How do you paste a screen shot into pastie?	01:56
	Or how do I cut, copy, and paste from the console?	01:56
	SamB tends to forget step 2, since half the time he's copying stuff from PuTTY sessions	01:56
SamB	Grafica: just copy and paste, actually; cut is not an option ;-)	01:57
	→ herdingcat joined	01:57
Grafica	I'm not able to copy anything from Bash.	01:57
	← dreiss_ left	01:58
SamB	Grafica: like sitaram suggests, try right clicking and choosing the "mark ..." option	01:58
sitaram	Grafica: http://www.google.co.in/search?sclient=psy-ab&hl=en&client=firefox-a&hs=BO5&rls=org.mozilla%3Aen-US%3Aunofficial&source=hp&q=copy%20paste%20from%20windows%20console	01:58
	→ Enchilada joined	01:58
Enchilada	Oh, shit. I need some urgent help here...	01:58
SamB	sitaram: geese, couldn't you have trimmed that URL down ?	01:58
Grafica	I did, I did. Nothing selects. The menu doesn't come up when I right click.	01:59
sitaram	SamB: my apologies... I meant to paste http://www.howtogeek.com/howto/windows-vista/copy-to-the-clipboard-from-the-windows-command-prompt/ (the first hit on that)	01:59
SamB	Grafica: oh, are you using rxvt or something?	01:59
Enchilada	I did a "git reset --hard hasOfSomeOldCommit" and then I did "git branch someBranchName"	01:59
Grafica	No	01:59
Enchilada	without first creating a branch for where I was most recently	01:59
	have I lost all my data??!!!	01:59
SamB	sitaram: oh, you ran up against their stat-gathering redirect, eh?	01:59
	or you really accidently pasted the search?	02:00
	SamB guesses the latter; there isn't enough stuff in there to encode that URL, on further inspection	02:00
sitaram	SamB: no; it was my error. I used 'y' instead of ';y' then select the number (I use vimperator; long story...)	02:00
SamB	ah	02:01
FauxFaux	SamB: Actually, there are much better ways to do stat gathering than rewriting the URL; they actually rewrite the url in search results so the target site recieves accurate referrer information. :)	02:01
SamB	FauxFaux: oh, is that so?	02:01
	interesting!	02:01
	→ Jarred joined	02:02
SamB	how in the world do they manage that?	02:02
Jarred	hi all	02:02
Enchilada	Someone? How can I recover from my disastrous loss?	02:02
Jarred	I want to ignore a directory in my git repository without deleting it?	02:02
	how do I do it	02:02
FauxFaux	Enchilada: !gka	02:02
gitinfo	Enchilada: For a better way to view the reflog, try: gka() { gitk --all $(git log -g --format="%h" -50) "$@"; }; gka	02:02
	→ lightcap joined	02:02
	← jasonbray left	02:03
offby1	Grafica: you have to do Alt+Space to bring up the menu, then E for "Edit", then I think k for Mark	02:05
	then you drag with the mouse	02:05
	then Alt+Space E C to copy.	02:05
	I think; this is from memory.	02:05
	→ Targen joined	02:05
	← baburdick left	02:06
	← lightcap left	02:06
Enchilada	FauxFaux: Ah, I didn't even know about git reflag. Okay, so I see the latest commit in the reflog output (i..e the commit that I thought was destroyed).	02:07
Grafica	Welcome to Git (version 1.7.6-preview20110708)	02:07
	Run 'git help git' to display the help index.	02:07
	Run 'git help <command>' to display help for specific commands.	02:07
	Administrator@D1R2NVG1 ~	02:07
	$ git status	02:07
	fatal: Not a git repository (or any of the parent directories): .git	02:07
	Administrator@D1R2NVG1 ~	02:07
	$ git add index.php	02:07
	fatal: Not a git repository (or any of the parent directories): .git	02:07
	Administrator@D1R2NVG1 ~	02:07
	$	02:07
	→ dankest joined	02:07
offby1	Grafica: ow. Not in the channel next time	02:07
	← abetaha left	02:07
FauxFaux	Grafica: IN A PASTEBIN	02:07
Enchilada	FauxFaux: so all I need to do now is do "git reset --hard theHashOfThatCommitIThoughtWasLost"?	02:07
EugeneKay	man git-rev-parse	02:07
gitinfo	the git-rev-parse manpage is available at http://jk.gs/git-rev-parse.html	02:07
FauxFaux	Enchilada: git checkout -bnewbranch $commitid; or branch/tag it from gitgui.	02:08
SamB	EugeneKay: this is not your terminal ;-;	02:08
FauxFaux	*gitk.	02:08
offby1	Grafica: yeah, anyway: your current directory isn't right, like I said earlier. You need to "cd /path/to/wherever/you/put/your/repository"	02:08
	→ baburdick joined	02:08
EugeneKay	SamB - no, it's my browser window	02:08
Grafica	OK, thanks.	02:08
SamB	EugeneKay: and here I was thinking it was your IRC client	02:08
Grafica	Sorry, command line is new to me too.	02:08
Enchilada	FauxFaux: is it safe to rely on only the first 7 characters of the hash?	02:08
FauxFaux	Enchilada: I don't know, does it work?	02:09
EugeneKay	SamB - http://youtu.be/fMrXMjxB6wk	02:09
Enchilada	i'll see	02:09
sitaram	Enchilada: for most repos yes. Some repos could get too big though	02:09
SamB	sitaram: do you recall if git checks for this at all?	02:09
Enchilada	sitaram: if there are two with the same, the git will just warn me i guess?	02:09
FauxFaux	Or someone could've messed with their repo such that the commit ids were full of duplicates. ¬_¬ https://github.com/FauxFaux/seqed/commits/master	02:10
	→ boombatower joined	02:10
SamB	FauxFaux: why did you call it "seqed"?	02:11
sitaram	SamB: most git commands that accept a SHA will complain about ambiguous ref or something but only if the repo gets too large	02:11
	→ coolstar-pc joined	02:11
FauxFaux	SamB: It's been seq[entialis]ed.	02:11
	gitinfo set mode: +v	02:11
Grafica	OK, that worked.	02:11
sitaram	Enchilada: yes it will warn you. In scripts try to use 10 just in case, but for interactive use its ok	02:11
offby1	\o/	02:11
Enchilada	thanks	02:12
	← ProLoser\|Work left	02:12
SamB	,dict seqential	02:12
	← dankest left	02:12
SamB	er, wait, this is not #emacs	02:12
Grafica	I thought the \o/ was for me.	02:12
	A head with arms up cheering.	02:13
sitaram	FauxFaux: how the hell did you manage that sequence!!!	02:13
FauxFaux	sitaram: http://blog.prelode.com/2010/12/git-set-commit-id/	02:13
Enchilada	FauxFaux: your "git checkout -b" statement would just be equivalent to first doing git --reset hard commitId and then creating a branch, right?	02:14
	← LondonGuy left	02:14
	→ LLStarks joined	02:14
FauxFaux	Enchilada: Apart from the fact taht mine is infinitely safer, more convenient, and makes it much harder to get into the situation you're in, sure.	02:15
Enchilada	heh ok	02:15
LLStarks	hi, is possible to search a git tree for a file that existed in previous commits?	02:15
	or current commits	02:15
FauxFaux	LLStarks: git log --all -- path/to/file	02:15
	→ jaisoares joined	02:15
	← oriba left	02:15
LLStarks	does it matter if i can't remember the path?	02:15
Enchilada	FauxFaux: i don't want to memorize and know about all these options such as -b and whatnots. It's getting confusing enough already. Just wanna have a handful of commands and options at my fingertips	02:16
	→ lemon-tree joined	02:16
Enchilada	at least for starters	02:16
FauxFaux	Well, remove reset, and never use it again; making space for some safer commands.	02:16
LLStarks	thx	02:16
sitaram	FauxFaux: you blew my mind... can I assume there's some serious brute forcing going on here (no energy to look at the code; haven't had breakfast yet!)	02:16
offby1	Grafica: it was!	02:16
	\o/	02:16
	← GodEater left	02:16
offby1	the opposite, when things go badly, is /o\	02:17
	(head under arms)	02:17
mjago	Grafica: glad you sorted it	02:17
	\|o\|	02:17
FauxFaux	sitaram: Yeah; it just appends random whitespace to the end of the commit until it gets what it wants (git trims it just before my code); 5 digits is only 2bits, which is a tiny number compared to cpus (thinking 10 seconds here, with no parallelisation or optimisation).	02:17
mjago	~o~	02:18
	→ igotnolegs joined	02:19
	← Araxia_ left	02:19
	← aidenhong left	02:19
	→ aidenhong joined	02:20
	← Textmode left	02:20
Enchilada	FauxFaux: would "git branch branchName hashOfOldCommit" and then "git checkout hashOfOldCommit" have worked? (Sorry, but I'm trying to get it down to be as simple as possible.)	02:20
	→ abetaha joined	02:21
FauxFaux	Yes, that's what checkout -b does (probably literally?).	02:21
Enchilada	ok nice	02:21
	→ GodEater joined	02:21
	← GodEater left	02:21
	→ GodEater joined	02:21
	→ mvrilo joined	02:22
Enchilada	FauxFaux: and checkout doesn't actually move the master branch does it	02:22
FauxFaux	Enchilada: Wait, no. checkout [hash] will detach your head, so doesn't do the same thing.	02:22
	→ dorisabayon joined	02:22
Enchilada	hmm	02:23
	→ tommyvyo joined	02:23
	→ Davey_ joined	02:24
	← scelis left	02:25
	← hyperair left	02:25
	← Fandekasp left	02:26
	← ripero left	02:28
	→ kryl99 joined	02:28
	← shiba_yu36 left	02:29
	→ shiba_yu36 joined	02:30
	← glennpratt left	02:30
	← mvrilo left	02:31
	→ hyperair joined	02:31
	→ drale2k joined	02:31
	← s0undt3ch left	02:31
	→ jasonbray joined	02:32
	← metcalfc left	02:33
	→ mvrilo_ joined	02:36
	← hyperair left	02:36
EugeneKay	Man, the things I do for laziness.	02:37
	← baburdick left	02:37
	← apok left	02:37
EugeneKay	Rather than git out of bed and use my desktop, I just did naughty things to gitolite's .ssh/authorized_keys file to authorize my laptop's new pubkey.	02:37
	← mvrilo_ left	02:37
	← Q\| left	02:38
	→ baburdick joined	02:38
	← borodon left	02:40
	→ Yuuhi` joined	02:42
SamB	FauxFaux: so ... how is this:	02:42
	2012-01-20 21:23:00-0500 [HTTPChannel,173318,192.168.0.31] 192.168.0.31 - - [21/Jan/2012:02:23:00 +0000] "GET /psyq/man/ucode/gspS2DEX.html HTTP/1.1" 200 23205 "http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CCAQFjAA&url=http%3A%2F%2Fnaesten.dyndns.org%3A8080%2Fpsyq%2Fman%2Fucode%2FgspS2DEX.html&ei=DB8aT9_XM6fs0gH_tKnKCw&usg=AFQjCNFuhtkbqHjKr77EsatY2cB_R9giVA&sig2=CYLayfsawqpRJL2KlYIL8g" "Mozilla/5.0 (Wi	02:42
	ndows NT 5.1) AppleWebKit/535.11 (KHTML, like Gecko) Chrome/17.0.963.33 Safari/535.11"	02:42
	→ mandric joined	02:42
SamB	better than this:	02:42
	2012-01-20 21:05:21-0500 [HTTPChannel,173305,207.172.114.83] 207.172.114.83 - - [21/Jan/2012:02:05:2	02:42
	0 +0000] "GET /psyq/man/ucode/gspS2DEX.html HTTP/1.1" 200 23205 "http://www.google.com/search?q=site	02:42
	%3anaesten.dyndns.org&btnG=Google+Search" "ELinks/0.12~pre3-2 (textmode; Debian; Linux 2.6.30-1-686	02:42
	i686; 188x39-3)"	02:42
	← Yuuhi left	02:43
	← dorisabayon left	02:44
SamB	(aside from the newlines evidently caused by a bad interaction among less, screen, and putty when you scroll up, I mean)	02:44
Grafica	What is the command to add a directory?	02:45
	← Davey_ left	02:45
	← notjohn left	02:45
	→ notjohn joined	02:46
offby1	mkdir	02:47
	← thiago left	02:47
	← GodEater left	02:47
Grafica	The directory is already made. I just want to add it.	02:47
Enchilada	I modified some files. But I don't want to commit them into the existing branch. I want to create a new branch for them. I should have done this before modifying the files, but I didn't. What can I do?	02:49
EugeneKay	Not to make it sound obvious, but just create the new branch and commit them there.	02:49
	← ColKurtz left	02:50
Enchilada	don't I have to checkout first to move the HEAD?	02:50
	← coolstar-pc left	02:51
EugeneKay	Sure. But it won't stomp on your changes.	02:51
	← cakehero left	02:53
Grafica	offby1: The directory is already made. I just want to add it.	02:53
	→ GodEater joined	02:55
	← GodEater left	02:55
	→ GodEater joined	02:55
	← SamB left	02:56
	→ SamB joined	02:57
	← jonesy left	02:58
	← eikke left	03:01
	← igotnolegs left	03:01
	→ hyperair joined	03:02
	→ igotnolegs joined	03:02
	→ coder4711 joined	03:03
	← icwiener left	03:03
	← arvind_khadri left	03:03
	← boombatower left	03:04
Grafica	I am just learning Git. I have existing files and directories, and I added some files to the repository, but how do I add an existing directory to the repository?	03:04
EugeneKay	Git tracks files, not directories.	03:07
	When you add a directory, you're really adding all the files in that directory	03:07
	← neurodrone left	03:07
	← baburdick left	03:07
	→ baburdick joined	03:08
	→ mgpcoe joined	03:08
	→ dreiss_ joined	03:09
phillijw	you need to put a file into the direcetory	03:09
	← gchristensen left	03:11
Grafica	There ARE files in the directory (not yet added to Git).	03:11
	→ kenperkins joined	03:12
Grafica	EugeneKay: Yes, I want to add the directory and all the files in it.	03:13
	← kenperkins left	03:13
SethRobertson	git add directoryname	03:13
	I recommend reading !book	03:13
gitinfo	'Pro Git' is a complete book about git that is available online for free (there's a print version, too): http://progit.org/book/	03:13
	→ angelsl joined	03:13
	→ techhelp joined	03:14
Grafica	Thanks.	03:14
	← acrocity left	03:14
	← cmyk_ left	03:15
	← Sigma left	03:16
	← coder4711 left	03:16
	← Goplat left	03:17
	← pen left	03:17
	← hyperair left	03:17
	→ skylord5816 joined	03:18
	← skylord5816 left	03:18
	→ skylord5816 joined	03:18
	← skylord5816 left	03:18
	→ skylord5816 joined	03:18
	← jasonbray left	03:20
	→ jasonbray joined	03:21
	← MissionCritical left	03:21
	← Enchilada left	03:24
	← soulcake left	03:24
	← iamjarvo left	03:26
	→ intripoon_ joined	03:26
	→ ehsan joined	03:29
	← vervic left	03:29
	← intripoon left	03:29
	→ dankest joined	03:29
	→ vervic joined	03:30
	← _inc left	03:30
	→ _inc joined	03:30
	→ mvrilo joined	03:31
	← orafu left	03:32
	← _inc left	03:32
	→ _inc joined	03:33
	← kermit left	03:33
	← traviscj left	03:34
	→ orafu joined	03:34
	→ kermit joined	03:34
	← shiba_yu36 left	03:36
CareBear\	Grafica : how's that restructuring coming along?	03:36
	← vervic left	03:37
	→ MissionCritical joined	03:39
	← dreiss_ left	03:41
	← ehsan left	03:41
	← _inc left	03:41
	→ _inc joined	03:41
	← _inc left	03:43
	← demi` left	03:43
	→ _inc joined	03:43
	← ProperNoun left	03:45
	← Danielpk left	03:45
	→ fr0stbyte joined	03:46
	→ amigojapan joined	03:48
	← shruggar left	03:48
	← kryl99 left	03:49
	→ alberto56 joined	03:50
	← herdingcat left	03:51
	→ johnkpaul joined	03:51
	→ munichlinux joined	03:52
	→ newbie999 joined	03:52
newbie999	Hi. How I can get DATE of last commit in git for script ?	03:53
	← alberto56 left	03:53
offby1	newbie999: "git show" shows the date.	03:54
	It probably takes a "format" argument so that you don't have to do fancy parsing.	03:54
	← f0i left	03:54
SethRobertson	If it doesn't, `git log -n 1` will	03:54
offby1	If may baby won't love me, I know her sister will	03:55
	→ kenperkins joined	03:57
	← LLStarks left	03:58
newbie999	I think both commands are the same. Thx.	03:58
	← codewrangler left	04:00
	← OOPMan left	04:02
	← pinky left	04:05
offby1	newbie999: for your purposes, yes	04:06
	→ pinky joined	04:07
newbie999	offby1: Yes, with "--date=short" it exactly what I need.	04:07
	← jasonbray left	04:08
	← baburdick left	04:08
	→ baburdick joined	04:10
	→ OOPMan joined	04:10
	← mattalexx left	04:11
	← kenperkins left	04:12
	← dankest left	04:13
	← mgpcoe left	04:14
	← drale2k left	04:16
	← johnkpaul left	04:19
Grafica	CareBear\: Are you still there? I was outside cleaning snow off of my car.	04:20
	→ acrocity joined	04:21
	← mvrilo left	04:21
	← acrocity left	04:21
	→ Danielpk joined	04:23
	← Grafica left	04:23
	→ napster joined	04:24
	→ acrocity joined	04:25
napster	I'm new to git. Is it okay to use git commit -am "msg" always if I need to commit all changes in tracked files?	04:26
	← eletuchy left	04:28
	← candybar left	04:29
SethRobertson	k	04:29
	yes	04:30
	→ dreiss_ joined	04:31
	→ dankest joined	04:33
	→ ProperNoun joined	04:34
	← hal left	04:35
	tizzo-afk → tizzo	04:35
	← dankest left	04:38
	← baburdick left	04:39
	→ baburdick joined	04:39
	→ Bucciarati joined	04:39
	← munichlinux left	04:40
	← napster left	04:40
	→ Chibby_ joined	04:42
	← angelsl left	04:44
	← monk12 left	04:44
	→ monk12 joined	04:45
	→ drale2k joined	04:45
	← jaisoares left	04:46
	→ munichlinux joined	04:46
	← munichlinux left	04:46
	tizzo → tizzo-afk	04:46
	→ angelsl joined	04:46
	← monk12 left	04:47
	→ monk12 joined	04:47
	→ savage- joined	04:49
	→ monk13 joined	04:49
	← monk12 left	04:49
	← monk13 left	04:50
	→ monk13 joined	04:51
	→ demi` joined	04:52
	→ napster joined	04:53
	← monk13 left	04:53
	→ monk13 joined	04:54
	← diogogmt left	04:55
	← t0rc left	04:55
	← amigojapan left	04:58
	← dorkmafia left	04:58
	← boblet left	04:59
	← OOPMan left	05:00
	← baburdick left	05:09
	→ baburdick joined	05:09
	← fr0stbyte left	05:10
	← napster left	05:12
	→ rurufufuss joined	05:13
	→ LLStarks joined	05:17
	← Orbitrix left	05:17
	← madewokherd left	05:20
	→ neurodrone joined	05:20
	→ OOPMan joined	05:20
	→ Orbitrix joined	05:23
	→ Fandekasp joined	05:23
	← berserkr left	05:24
	← dreiss_ left	05:27
	→ aspotashev joined	05:27
	→ weiyang joined	05:27
	← baburdick left	05:30
	→ johnkpaul joined	05:30
	→ fr0stbyte joined	05:31
	→ mattly joined	05:31
	← Chibby_ left	05:31
	→ uu joined	05:32
	→ Goplat joined	05:33
EugeneKay	Yes, it will work, but it's not a particularly good habit to be in.	05:33
	Oh, he's buggered of.	05:33
	→ Kulrak joined	05:33
	← Dreadnaught left	05:34
	← johnkpaul left	05:34
SethRobertson	Meh, if you are not hiding the sausage making it doesn't make a lot of difference	05:35
	→ dankest joined	05:35
	offby1 hides the salami making	05:36
	offby1 glances around nervously.	05:36
	← acrocity left	05:37
	← dankest left	05:37
	← mandric left	05:38
	→ T-Gunn joined	05:38
	ChanServ set mode: +v	05:39
	← fr0stbyte left	05:40
	← monk13 left	05:40
T-Gunn	I'm trying to setup gitosis... im following the tutorial at http://www.redconfetti.com/2012/01/rails-3-on-whm-cpanel-vps-server/ and am at the part where i clone the repository to my local machine to be able to easily configure gitosis on the server and i get an error "fatal: protocol error: bad line length character: Shel"... any ideas?	05:40
	← savage- left	05:41
offby1	first piece of advice:	05:41
	faq gitosis	05:41
	sigh	05:41
	!gitosis	05:41
gitinfo	gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite	05:41
	→ monk13 joined	05:41
T-Gunn	ok thanks	05:41
	→ frerich2 joined	05:42
	→ acrocity joined	05:43
osmosis	what is hg ?	05:43
EugeneKay	Second piece of advice: !blog	05:43
gitinfo	Blog posts, while helpful and informative, are quite often outdated, give bad advice, or are just plain wrong. Please don't rely solely upon them, or treat them as authoritative.	05:43
	← frerich left	05:43
	→ fr0stbyte joined	05:44
	← pdobrogost left	05:44
SethRobertson	osmosis: Hg is one of the many other inferior revision control systems that exist in the world	05:47
	← OOPMan left	05:53
	→ psakrii joined	05:53
offby1	probably the best of 'em though :)	05:55
	→ gavin_huang joined	05:56
	→ nownot joined	05:56
	← nownot left	05:56
	→ nownot joined	05:56
	← Fandekasp left	05:56
nownot	if i want to create a new branch off of an old commit how do I do that?	05:56
	git checkout 5ea45e79b8089410564d3a1336b5f69da25fb6ff but it created a branch called (no branch)	05:57
	→ Dave^\| joined	05:58
offby1	no need for that.	05:58
	"git branch whatever-you-wanna-call-it 5ea45e79b8089410564d3a1336b5f69da25fb6ff"	05:59
nownot	wow, thanks	05:59
offby1	actually it hasn't created a branch called (no branch); that's git's way of telling you ... well ... that you're not "on" any branch :)	05:59
	→ xil joined	05:59
	→ dnjaramba joined	05:59
offby1	general-purpose git advice: 1) Read "Git For Computer Scientists" (http://eagain.net/articles/git-for-computer-scientists/); 2) Run "gitk --all"; 3) Learn about "git reflog"	05:59
	now do "git checkout master" or whatever to get back to sanity	06:00
	→ amigojapan joined	06:00
	← edude03 left	06:00
	← LLStarks left	06:02
	← fr0stbyte left	06:02
nownot	ok, thanks	06:04
	← nownot left	06:04
	→ dankest joined	06:05
	→ kenperkins joined	06:06
	← dankest left	06:07
	→ sluggo206 joined	06:08
	→ Fandekasp joined	06:08
	→ akosikeno joined	06:09
sluggo206	Can anyone tell me how to list the commits that are in one branch but not in another, without modifying either branch?	06:11
	← monk13 left	06:11
	→ fr0stbyte joined	06:12
	→ dorkmafia joined	06:12
sluggo206	I want to abandon a dev branch, but first I want to make sure there's nothing important in it. In the meantime I've merged dev into master and there are later third-party commits in master, so I can't tell by comparing the "git log"'s whether I'm missing any important changes.	06:12
SethRobertson	git log dev..master or git log dev...master	06:12
	→ monk12 joined	06:12
SethRobertson	git-cherry might also be interesting	06:13
sluggo206	I don't understand what cherry-pick does, or how it's different from a merge	06:13
	← [M]ax left	06:13
	→ Cam joined	06:13
sluggo206	I read the help page a few times and still don't understand what it;s for	06:14
SethRobertson	cherry-pick makes a copy of the commit. merge just tells git to look at both branches of history	06:14
	Which is of course different from `git cherry`	06:14
	→ [M]ax joined	06:15
CareBear\	sluggo206 : do you understand the difference between a commit and a branch?	06:16
	← Goplat left	06:16
sluggo206	A branch is a chain of commits	06:16
SethRobertson	!branch	06:16
gitinfo	A branch and a tag are just convenient ways of spelling the name of a particular commit. A commit represents a specific set of files and the history of all commits which came before it, and the SHA-1 hash tag official name provides cryptographic assurance of the lineage of a particular commit (and thus branch or tag). A branch's reference may change. A tag usually doesn't.	06:16
	Xamual → Samual	06:16
CareBear\	sluggo206 : your explanation is a bit of an oversimplification, but it's good enough	06:17
	← monk12 left	06:17
	← techhelp left	06:18
sluggo206	Anyway, I looked at "git help cherry" but I don't understand it either. When would you want to use cherry or cherry-pick?	06:18
CareBear\	sluggo206 : so consider that cherry-pick operates on changes in individual commits only, while merge operates on branches ie. commits including all their history	06:18
	← infid left	06:18
sluggo206	So how is cherry-pick different from merge --squash?	06:18
CareBear\	they don't compare at all	06:18
	again:	06:19
	cherry-pick operates on a single commit	06:19
	→ infid joined	06:19
CareBear\	pulls it out from the context of any branches where it exists, and tries to re-create the identical changes in your current branch	06:19
	without any connection to where the commit was	06:20
sluggo206	So it's like applying a patch?	06:20
CareBear\	that's not a bad comparison	06:20
sluggo206	OK.	06:20
	← dnjaramba left	06:20
sluggo206	So when would I want to use "git cherry"?	06:21
	→ dnjaramba joined	06:21
CareBear\	to find out what commits you might want to cherry-pick	06:21
	XaV`S → zz_XaV`S	06:21
CareBear\	I've used Git for years now and never used git cherr	06:21
	+yu	06:21
	-u	06:22
	it depends on workflow of course, I haven't had a need	06:22
sluggo206	It says "git cherry <upstream> <head>". But in this case I have two local branches, master and dev, and I don't care about the upstream. So if I'm in master, would that be <head> and the dev branch would be <upstream>?	06:24
CareBear\	that depends on how the branches relate to each other	06:25
sluggo206	Dev is theoretically more recent, but there have been changes to master in the meantime. I want to find out if there's anything in dev that's not in master that I care about.	06:26
CareBear\	it seems to me that git cherry would do just that	06:26
	→ tic joined	06:26
sluggo206	So if I'm in master, "git cherry dev" would be the right syntax?	06:26
CareBear\	yes, and it seems the other way around would also work	06:27
sluggo206	OK, it listed two commit IDs.	06:27
CareBear\	what prefixes?	06:27
sluggo206	"+ "	06:28
CareBear\	those that only exist in the <head> branch are prefixed with a plus (+) symbol	06:28
sluggo206	That makes sense, because those two changes are the ones only in master.	06:29
	So if it doesn't list anything besides those, it means there are no unique commits in dev?	06:29
	From the help it looks like, if there are no lines starting with "-", there are no unique commits in dev.	06:31
	← paul_irish left	06:32
	← Cam left	06:32
	← detaer left	06:33
	→ hh__ joined	06:34
	← queequeg1 left	06:34
	→ queequeg1 joined	06:35
	gitinfo set mode: +v	06:35
CareBear\	I guess I would run it the other way around	06:35
hh__	i did git reset --soft HEAD^, now my remote repo isn't in sync with my local repo, how do i tell the remote repo it should accept changes?	06:35
CareBear\	hh__ : do you really want to rewrite your history?	06:36
	→ disappearedng joined	06:36
hh__	yea	06:36
	← disappearedng left	06:36
CareBear\	push -f	06:36
	→ disappearedng joined	06:36
hh__	thanks	06:37
	probably a no since i dont see it in --help switch, but would be nice to write shorthand commands like git c instead of git commit , etc.	06:38
nevyn	git alias	06:38
	← acrocity left	06:38
	→ giallu joined	06:38
xil	hi. I did a fetch/merge from an upstream repo that I wish I now hadn't. I'm reading about how to fix that but want some corroboration before I do anything. Some recommendations are to do a reset, but I don't understand why a reset would take me back to before the merge. Another was to do a rebase, but wouldn't I need to rebase off something older than the merge?	06:39
CareBear\	xil : indeed reset --hard will take you back to before the merge	06:39
	← kenperkins left	06:39
CareBear\	xil : reset --hard simply forces your current branch to point somewhere else than it currently does	06:39
xil	CareBear\: so what's the location I point it to? HEAD^?	06:40
CareBear\	xil : the commit before the merge commit	06:40
	← psakrii left	06:40
CareBear\	xil : if the merge commit is HEAD then HEAD^ would work	06:41
	→ acrocity joined	06:42
xil	CareBear\: what is HEAD exactly?	06:43
	I know it's a pointer, but to what precisely?	06:43
CareBear\	xil : the last commit on the branch	06:43
	→ bindaasomatic joined	06:43
	← Orbitrix left	06:44
sluggo206	Thanks for the help on "git cherry". I learned a new command.	06:45
xil	I see. Okay. So I just do 'git reset --hard <SHA>' where the <SHA> is the last commit then	06:45
	← fayimora left	06:46
xil	oh, but will that make it as if the merge I'm disregarding didn't ever happen, or is HEAD still at the merge commit?	06:46
	← rurufufuss left	06:46
nevyn	it'll make it as if it never happened	06:47
	← acrocity left	06:47
CareBear\	yes	06:47
xil	perfect. Thanks so much	06:47
	oh one last thing. Do I have to type out the whole hash, or is there a shortcut I can use? I'm only undoing the most recent commit	06:48
nevyn	is it a merge?	06:48
xil	yes	06:49
	← Fandekasp left	06:49
nevyn	xil: I'd copy it from git log	06:49
	but you only need enough of the hash to uniquely identify the commit	06:49
xil	yeah that's what I was trying to avoid, because for some reason I can't select from the log. Terminal just won't let me	06:50
	→ Orbitrix joined	06:50
	← EyesIsAsleep left	06:50
	→ leavittx joined	06:50
	→ acrocity joined	06:51
nevyn	what terminal program	06:51
xil	oh hmmm....this is interesting. So it says 'Merge x y'. Could I just use the x in that to get to the previous commit?	06:52
nevyn	no	06:52
xil	nevyn: I think the real problem is that I'm using elinks for man pages and such, I don't remember what the variable is called, READER?	06:52
	→ EyesIsAsleep joined	06:53
nevyn	PAGER	06:53
xil	that's it	06:53
CareBear\	xil : x and y are the two parents of the merge commit	06:53
	xil : you will want one of them, but you need to know which one	06:53
xil	I was using urxvt and couldn't scroll with my mouse so started using elinks as the pager. Now I use the Xfce terminal and don't need elinks, but haven't removed it out of laziness	06:54
CareBear\	"couldn't scroll with my mouse" also doesn't sound so good	06:54
	→ sikao_lfs joined	06:55
xil	CareBear\: so I did "git fetch upstream ; git checkout develop ; git merge upstream/develop ; git checkout feature/MyFeature ; git merge develop". One of the parents is clearly the merge from upstream/develop into develop, so should I choose the other parent?	06:55
	I want to undo the second merge, into my feature	06:56
	← acrocity left	06:56
CareBear\	xil : check the log	06:57
	and never make changes to remote repos	06:57
xil	oh okay, I found it. I was worried that the second parent was hidden way down in this list because the merge brought in a bunch of commits, but I found it. So I guess I still need to know, do I reset to the merge from upstream to develop, or to my last commit before all of this?	06:59
	also, what do you mean about changing remote repos?	06:59
	my only intention is to roll my feature branch back to before this merge	06:59
	the merge broke something because of another dev and in order to keep working I have to roll back because the dev isn't available to talk now	06:59
	→ macrover joined	07:00
CareBear\	so feature is not a remote?	07:00
	then it's fine	07:00
	→ Danny_Joris joined	07:00
xil	yeah feature is published with git flow, but I haven't pushed since before the merge, so this is all local	07:00
CareBear\	I have no idea what git flow means	07:00
xil	flow is some addon that helps with something, don't really know	07:00
	my supervisor told me to use it so I use it	07:01
CareBear\	you realize noone here neccessarily has any idea about it if it's not a part of standard git	07:01
xil	I do realize that	07:01
	but I thought I'd mention it in case you did	07:01
CareBear\	you need to check what commit hash you want	07:02
	we can't say	07:02
	→ acrocity joined	07:02
CareBear\	but it will be obvious from git log	07:02
xil	yeah I'm looking at the log now	07:02
CareBear\	if in doubt you can of course use git log --graph	07:02
xil	I'm just not sure which to pick because I have 2 options. One is literally the next to last commit, but that is the merge from upstream/develop -> develop. The other option is to go to the last commit I made, before all of these merges. Question is, if I do the latter will it affect the other branches, or only my branch	07:03
	my current branch*	07:03
CareBear\	reset only affects the current branch	07:04
xil	oooh okay. Now I understand why you were recommending git log --graph. Because I want to roll back to the last commit for this branch. But Right now I'm looking at the log in date order, so most of these commits don't have anything to do with my current branch	07:05
CareBear\	add --oneline to get more overview but less detail	07:06
xil	gitk	07:06
CareBear\	shrug whatever you prefer	07:07
	acrocity → Guest75108	07:07
xil	I mean, none of these options is convenient really, so I play around between them	07:07
	gitk isn't great, but it works for the graph for me	07:07
	← igotnolegs left	07:07
	← _inc left	07:08
xil	hmmm....I have a question about git stash. Hope I'm not getting on anyone's nerves with all these questions, I've just read a good amount of docs and am frankly still a bit lost. So I have tried using the stash a couple times and something seemed off. I'd just like some clarification	07:10
	when I run 'git stash' it only actually stores the stuff in my working tree and index correct?	07:11
	it doesn't care about the commit I'm at or anything like that?	07:11
	← neurodrone left	07:11
CareBear\	right	07:11
	it does store where you stashed	07:11
	but it's only informational	07:11
xil	so like, I made some changes since this last commit that I want to undo. I could 'git stash', then 'reset --hard <previous commit>' then 'git pop' and it would be the same as if I'd made those changes before the merge I'm trying to rid myself of?	07:12
CareBear\	so like, yeah	07:12
xil	lol	07:13
	I'm tired =P	07:13
	that's a lot for the help =]. I needed to get some stuff done and spent a lot of time trapped because of this broken code	07:13
	thanks*	07:13
CareBear\	you could always just have made a new branch	07:14
	from wherever you wanted to work	07:14
	git checkout -b newfoobla where_you_want_to_start	07:14
	← weiyang left	07:14
xil	ignorance. Didn't know I could do that	07:15
CareBear\	check out a !book	07:15
gitinfo	'Pro Git' is a complete book about git that is available online for free (there's a print version, too): http://progit.org/book/	07:15
skylord5816	CareBear\: D: your name is weirding me out	07:15
xil	I've read a good amount of that, but pieces here and there	07:15
CareBear\	skylord5816 : good	07:15
xil	it's not that I haven't seen 'git checkout -b newbranchname whereyouwanttostart' but it takes time to stick you know?	07:16
CareBear\	xil : don't think of commands, think of the underlying operations they represent	07:16
skylord5816	CareBear\: it's just the \ - it makes me get confused when skimming chat history	07:16
CareBear\	xil : don't learn syntax, learn expression	07:16
	← sluggo206 left	07:16
CareBear\	skylord5816 : perfect	07:16
	skylord5816 : I needed a unique character at the end of my nick in 1995 and picked \ because it was pretty uncommon and because maybe it would f up some software	07:17
	skylord5816 : it keeps delivering	07:17
skylord5816	lol :D	07:17
xil	CareBear\: I mean, the problem is really that I'm not learning git the way you might normally learn something, i.e., by studying it. Right now git is only a tool for me to get a job done, so I am learning with the short term in mind, so I don't focus on retaining the information	07:18
skylord5816	and this client doesn't even have an /alias D:	07:18
CareBear\	xil : so you are not really learning at all	07:18
xil	yeah	07:18
CareBear\	xil : instead you are just wasting both your time and mine	07:18
xil	well no, I wouldn't say that. I am remembering this because it was a problem you helped me overcome	07:19
	I retain that stuff well; I learn well from solving problems	07:19
CareBear\	sounds good	07:19
	← Guest75108 left	07:19
xil	but if I'm reading about how to do something and then I do it and it works, in a situation where there isn't a problem I just needed to learn some new command, then I don't retain as well, because it usually wasn't important	07:19
	not saying that's a good way to go about it, just what it is for me at this moment	07:20
	CareBear\: thanks again. The reset worked =D	07:24
CareBear\	xil : you're welcome	07:25
	→ stringoO joined	07:25
	← aspotashev left	07:25
xil	off to do the stuff that was necessary for. You're awesome, CareBear\ =D	07:26
	← xil left	07:26
	→ acrocity joined	07:29
	acrocity → Guest7614	07:29
	→ yshh joined	07:30
	← Guest7614 left	07:30
	← Danny_Joris left	07:32
	→ acrocity_ joined	07:40
	→ i42n joined	07:41
	acrocity_ → acrocity	07:43
	→ dfr\|mac joined	07:44
	→ Niya joined	07:47
	← skylord5816 left	07:48
Niya	I'm seeing a really strange thing... when I pull files from my repo, in this one file, I get these <82> characters (as shown in vim) - I'm using git in cygwin on Windows 7 64-bit... I know the files are correct on the server, so I'm really at a loss for what's happening.	07:49
	→ monk12 joined	07:49
EugeneKay	Niya - what editor are you using?	07:51
	vim?	07:51
Niya	Typically eclipse on OS X.	07:51
EugeneKay	My guess is something unicodey, and cygwin vim is barfing. Try Notepad++ in Unicode mode	07:51
Niya	It's really strange... the string, in vim, says "sBuilding<82>onstruction", but it should be a 'C'. Each of the cases are like that.	07:52
CareBear\	Niya : can you put the file online?	07:53
	no pasting, the file verbatim	07:53
Niya	I don't want to share the entire file, that's the thing.	07:54
CareBear\	then find someone you can trust who is competent to help	07:54
	→ baburdick joined	07:55
	← fr0stbyte left	07:55
Niya	Okay. Sorry about that.	07:56
	← hwrdprkns left	07:56
CareBear\	no need to be sorry	07:56
	→ hwrdprkns joined	07:56
CareBear\	just face that if you don't give useful information you can not be helped	07:56
	→ arif-ali_ joined	07:57
	← dfr\|mac left	07:57
Niya	Yeah. Thank you.	07:58
	→ dreiss_ joined	07:58
CareBear\	look at the file with a hex dump tool	07:58
	← arif-ali left	07:58
CareBear\	see what the bytes are	07:58
Niya	Oh, um. Actually... Is there a way to find a specific file inside the <project>.git repo?	07:58
CareBear\	perhaps the various ways the file is being decoded can be inferred	07:59
	define find?	07:59
Niya	I want to open it on the server, if possible.	07:59
CareBear\	define server?	07:59
	and open?	07:59
Niya	git repo	07:59
CareBear\	these are bad git terms	07:59
	I mean, these expressions don't map to git terms	07:59
Niya	okay, let me try again	07:59
	there is a place where I git cloned from.	08:00
CareBear\	start with explaining what you have	08:00
	okey	08:00
Niya	I can ssh to this computer.	08:00
CareBear\	no need	08:00
	your clone is equal	08:00
	did the remote repo have a worktree as well?	08:00
	or was it bare	08:00
	?	08:00
Niya	Yes, but I'm not convinced that something strange isn't happening as the bits are coming in. I know that's kind of paranoid ._.	08:01
CareBear\	nothing strange is happening	08:01
	as I said, your clone is equal	08:01
Niya	A worktree... like, a place where I cloned it there?	08:01
CareBear\	no	08:01
	a directory where the versioned files are stored	08:01
Niya	yes.	08:01
CareBear\	as opposed to only the git database	08:01
Niya	Hmm.. Maybe not	08:02
	→ churp joined	08:02
CareBear\	if you do ls in the directory you cloned from what do you see?	08:02
Niya	"branches config description HEAD hooks info objects packed-refs refs"	08:03
CareBear\	ok, so that is a bare repo	08:04
	ie. it does not have a worktree	08:04
Niya	Okay.	08:04
CareBear\	this is good and to be expected	08:05
	you can very much ignore it	08:05
	your clone is identical	08:05
	→ coder4711 joined	08:05
Niya	I see.	08:05
	Okay.	08:05
CareBear\	now	08:06
	besides all the identical stuff	08:06
	your clone also has a worktree	08:06
	the worktree contains a human-friendly representation of the git database	08:06
	ie. the files you work on, including the file with the character	08:06
Niya	ah, I see.	08:07
CareBear\	this representation is subject to various magic that git may do to help you	08:08
	mostly line ending conversion	08:09
Niya	Yes.	08:09
	→ ph^ joined	08:10
	→ LongBeach joined	08:10
	← LongBeach left	08:10
	→ candybar joined	08:11
	→ dirkle joined	08:13
	→ eletuchy joined	08:13
	→ munichlinux joined	08:15
CareBear\	git ls-tree will show you the blob hash for your files	08:15
	→ tewecske joined	08:16
CareBear\	git show hash will output the blob, you can pipe that to a hex dump program such as xxd	08:16
	← monk12 left	08:17
	→ dandaman1 joined	08:19
Niya	how am I supposed to use ls-tree? I don't understand "<tree-ish>"	08:20
	→ dnjaramba_ joined	08:20
	← dnjaramba left	08:20
kevlarman	CareBear\: revision:path is quicker	08:20
	→ caseymcg joined	08:21
	→ centipedefarmer_ joined	08:21
	→ Error404NotFound joined	08:22
	→ iocor joined	08:24
	← iocor left	08:24
	→ iocor joined	08:24
Niya	ah, I understand	08:24
CareBear\	kevlarman : yup. nice	08:24
	→ Phylock joined	08:25
CareBear\	kevlarman : I like hashes though	08:26
	← [M]ax left	08:27
Niya	I was thinking I'd be able to find the line with the file in question, look at the hash, and find that in objects/... but I am not correct?	08:27
	→ psakrii joined	08:28
CareBear\	line?	08:29
	no	08:29
	use git show	08:29
	→ [M]ax joined	08:30
	→ faber joined	08:30
Niya	I see. Okay. Thank you.	08:30
	← macrover left	08:31
	→ rendar joined	08:33
	→ queequeg1_ joined	08:36
	← resmo left	08:36
	→ pdtpatrick_ joined	08:37
Niya	Thank you again for your help.	08:37
CareBear\	so did you find the problem?	08:38
	← queequeg1 left	08:38
Niya	Well, not exactly? I found that, as you said, it is exactly like this in the repo, but for some reason Eclipse on OSX translates it to compile, and Eclipse on Windows doesn't.	08:39
	← michael_mbp left	08:39
	← centipedefarmer_ left	08:39
CareBear\	what bytes are there?	08:39
	git show \| xxd	08:39
	← dnjaramba_ left	08:40
Niya	it's 0x82	08:41
CareBear\	before and after	08:42
Niya	I think I understand now... yes. Basically, on OS X, the identifier is name<0x82>Ending and both in the declaration and use it uses this name	08:42
	In Windows, it gets converted, for some reason, to ","	08:43
	And that breaks the identifier.	08:43
CareBear\	git blame to find what moron committed the junk	08:44
Niya	Yes, I did that ^^;	08:44
CareBear\	ah	08:44
Niya	Again, thank you muchly	08:44
CareBear\	so you got what you deserved :p	08:44
Niya	Oh, I mean, I used git blame	08:44
CareBear\	aha	08:44
	ok	08:44
Niya	it wasn't my edit.	08:44
	Thank you again	08:45
CareBear\	you're welcome	08:45
Niya	And I should totally sleep	08:45
	Take care	08:45
	← Niya left	08:45
	← munichlinux left	08:45
CareBear\	sleep well	08:45
	← dirkle left	08:47
	→ resmo joined	08:49
	← Jarred left	08:50
	→ nikuyoshi joined	08:50
	← nikuyoshi left	08:50
	→ nikuyoshi joined	08:50
	← joshsmith left	08:51
	→ cesc joined	08:53
	← dreiss_ left	08:55
	→ arif-ali__ joined	08:55
	← gavin_huang left	08:56
	← arif-ali_ left	08:56
	← Sweetshark left	08:56
	← [M]ax left	09:00
	→ Sweetshark joined	09:00
	← Sweetshark left	09:00
	→ Sweetshark joined	09:00
	← iocor left	09:01
	→ koo4 joined	09:01
	→ [M]ax joined	09:02
	→ Textmode joined	09:06
	← amigojapan left	09:08
	← Vile left	09:09
	→ Vile joined	09:10
	→ khmarbaise joined	09:11
	→ LeMike joined	09:11
	→ herdingcat joined	09:13
	← LeMike left	09:17
	← cesc left	09:17
	→ cesc joined	09:18
	→ boombatower joined	09:19
	← justfielding left	09:19
	→ thejoecarroll joined	09:23
	→ tvw joined	09:24
	ChanServ set mode: +v	09:24
	→ FernandoBasso joined	09:27
	→ kmc joined	09:28
	→ cmyk_ joined	09:29
	← caseymcg left	09:30
	→ nazgul101 joined	09:31
	→ DarkAR joined	09:31
	→ jceb joined	09:31
	← disappearedng left	09:31
	→ roflin joined	09:33
	→ d0k joined	09:34
	→ disappearedng joined	09:36
	← Error404NotFound left	09:36
	→ martinjlowm joined	09:39
	→ iocor joined	09:40
	← Blaster left	09:40
	← iocor left	09:41
	→ Blaster joined	09:42
	→ martinjlowm-lptp joined	09:43
	→ iocor joined	09:44
	← iocor left	09:44
thejoecarroll	good morning. i'm wondering if anyone knows a way to perform git commits over ftp? i've tried to do so by mounting an ftp share with curlftpfs, but i get errors. i want to back up a mysql database from a web server with versioning, but the backup service the hosting provider offers only allows ftp connections, no ssh.	09:45
	→ darwinbits joined	09:46
	← aalex left	09:46
mjago	thejoecarroll: git bundle?	09:47
kevlarman	that requires running git bundle	09:47
	so same issue	09:47
	your next best bet is probably smart http	09:47
thejoecarroll	yeah, the problem is that i don't want the git repo on the server itself, taking up space, i want the repo with all its commits to live on a remote backup server, if that's possible	09:48
	← sikao_lfs left	09:48
thejoecarroll	and having no shell access to that backup server is really limiting	09:49
	@kevlarman : could you explain more please?	09:49
	→ Error404NotFound joined	09:50
kevlarman	i've always used ssh://	09:50
	→ gusnan joined	09:50
kevlarman	but see man git-http-backend for the gory details	09:50
gitinfo	the git-http-backend manpage is available at http://jk.gs/git-http-backend.html	09:50
	→ detaer joined	09:51
thejoecarroll	the problem here is that there's no possibility of using ssh (so other options like rsync etc. are also ruled out). the only write access i have to the backup storage is over standard ftp and i can't run any backend or clone/pull commands on it. essentially what i need is a way to get git to initialize, maintain and manipulate a repo on a remote filesystem over ftp. curlftpfs seemed perfect, but hasn't worked for me.	09:55
	→ sheldonels joined	09:56
	→ vdv joined	09:57
	ChanServ set mode: +v	09:57
sheldonels	Hi, can anyone help me find an explanation of how git gui's (and some other tools') stage line(s) code works? I'd like to implement the same feature in a text editor extension for git, but I'm not too sure how it manages to split up diffs.	09:59
	← [M]ax left	10:00
mjago	thejoecarroll: you might check http://bit.ly/yIE3cK out	10:00
	→ munichlinux joined	10:01
	← munichlinux left	10:01
	→ [M]ax joined	10:01
thejoecarroll	mjago: that link is broken	10:01
	← tvw left	10:03
	→ iocor joined	10:03
mjago	thejoecarroll: sorry, try http://bit.ly/ylE3cK	10:05
	← darwinbits left	10:06
	→ QaDeS joined	10:06
	← dorkmafia left	10:07
	→ arvind_khadri joined	10:08
	→ elenril joined	10:09
elenril	is there a way to checkout a given commit somewhere outside of the working tree?	10:09
nevyn	you can make a thin clone	10:10
	→ _iron joined	10:10
nevyn	but you can't commit to it	10:10
thejoecarroll	thanks, mjago. i had a look on stack exchange already but somehow missed this one. it seems that the OP was trying to do something very similar to me, and he seems to have had more luck with the approach i've tried. he mentions in post from dec 2010 that ftp support has been added to git itself for pull, but not for push at that time. has this changed?	10:10
elenril	hm, right, that could do the trick	10:11
	thanks	10:11
nevyn	elenril: just make a clone	10:11
	and use the hardlink stuff	10:11
	← tatsuya_o left	10:12
	→ tatsuya_o joined	10:12
mjago	thejoecarroll: that I couldn't tell you	10:12
thejoecarroll	ok, thanks anyway. time for me to go rtfm :-D	10:13
elenril	but can a shallow clone start somewhere else than HEAD?	10:13
nevyn	?	10:14
	when you clone you get the branch pointed to by HEAD	10:15
elenril	well let's say i want to checkout a five year old tree somewhere without affecting my current working tree	10:15
nevyn	is there uncommited stuff in your working tree?	10:16
elenril	no, why does that matter	10:16
nevyn	elenril: clone doesn't clone uncommitted stuff.	10:17
	elenril: so clone it.	10:17
	I don't understand what's difficult here.	10:17
	→ Sigma joined	10:17
	← Sigma left	10:17
	→ Sigma joined	10:17
nevyn	~/somegitthing$	10:17
	~/somegitthing$ mkdir ../someothergitthing	10:17
	actually	10:17
	~/somegitthing$ cd .. && git clone somegitthing someothergitthing	10:18
elenril	hmm, right, now that i think about it this will be very cheap on the same device	10:18
	→ sdoutob joined	10:19
sheldonels	elenril: it will use hard links by default (I think), which uses no extra space at all for the history	10:19
nevyn	elenril: with -l	10:19
	for local	10:19
	which uses hardlinks for objects	10:19
	→ airborn joined	10:20
	← [M]ax left	10:20
sheldonels	-l for clone is default for local folders, from the man page: "This is now the default when the source repository is specified with /path/to/repo syntax"	10:20
	→ [M]ax joined	10:21
sheldonels	so just the clone /path/to/repo /new/path will be enough :)	10:21
	← sdoutob left	10:21
nevyn	right.	10:21
	← cmyk_ left	10:21
	← fisted left	10:23
	→ rzec joined	10:24
	→ fisted joined	10:25
	→ s0undt3ch joined	10:27
	→ vervic joined	10:27
	→ kryl99 joined	10:27
	← koo4 left	10:28
	← Blaster left	10:30
	→ mlukashov joined	10:32
	← vervic left	10:32
rzec	is there a threshold of the number of tags that git can hold at which performance start to degrade?	10:33
sheldonels	they are stored as files so I think the degradation is dependant on the number of files on your filesystem	10:34
	→ pantsman joined	10:35
	← pantsman left	10:35
	→ pantsman joined	10:35
	→ drizzd joined	10:36
	← nowhere_man left	10:37
	← sheldonels left	10:37
	→ n2diy joined	10:38
cmn	up to the last few releases, a few hundred tags would make git slow down, as it was processing them quite inefficiently, but they fixed that	10:38
	EyesIsAsleep → EyesIsServer	10:39
	← stringoO left	10:41
	→ koo4 joined	10:41
	ChanServ set mode: +v	10:41
n2diy	I'm starting out with GIt, and need to setup a diff tool. I'm running Xubuntu 11.04, and have Jedit installed, suggestions?	10:42
	→ stringoO joined	10:42
	→ pdobrogost joined	10:43
	← vdv left	10:45
	← Textmode left	10:47
	→ johnhamelink joined	10:47
rzec	where would I notice the performance impact with have thousands of tags?	10:50
martinjlowm	n2diy: aren't you happy with 'git diff'?	10:50
	← Sigma left	10:51
n2diy	martinjlowm, I don't know? I don't even know if I have it? I just installed git 15 minutes ago.	10:51
	← aro left	10:52
n2diy	martinjlowm, Well, I have git diff, so I'm happy about that.	10:52
martinjlowm	oh okay, great :)	10:52
n2diy	martinjlowm, Thanks, that was easy enough.	10:53
martinjlowm	my pleasure	10:53
	← drale2k left	10:56
mjago	n2diy: from jedit-users list: http://bit.ly/wfE34J	10:57
	→ drale2k joined	10:57
n2diy	mjago, thanks.	10:58
	→ psoo joined	11:01
	→ Textmode joined	11:01
	← Textmode left	11:01
	← coder4711 left	11:04
	← xerora left	11:04
	→ iamjarvo joined	11:05
	← jceb left	11:05
	← akosikeno left	11:06
	← lapistano left	11:07
	→ lapistano joined	11:08
	← rzec left	11:10
	→ shiba_yu36 joined	11:13
	← stringoO left	11:15
	← iamjarvo left	11:16
	← Targen left	11:18
	→ Targen joined	11:18
	→ ipalaus joined	11:19
	→ akosikeno joined	11:20
	→ Textmode joined	11:21
	→ iamjarvo joined	11:23
	→ Sigma joined	11:25
	→ frozzenfire joined	11:25
	→ ravenzz joined	11:25
	→ jargon- joined	11:25
	→ Sonderblade joined	11:26
	← j0ran left	11:26
	← Sonderblade left	11:26
	→ Sonderblade joined	11:27
	→ RaptorX joined	11:28
	→ robotmay joined	11:30
	→ nowhere_man joined	11:30
	→ Cromulent joined	11:30
	→ j416 joined	11:30
ravenzz	hi I am new to git, I would like to use it for a private project I am going to work (a wp website). Could this article be good to start http://goo.gl/Lqn9C ?	11:31
	→ webrover joined	11:31
FauxFaux	Looks rather over-complicated to me.	11:31
cmn	that mixes setting up the repo on the server with actually starting	11:32
ravenzz	use git in a wp development workflow?	11:33
cmn	ravenzz: you'd be better off reading the !book and then trying to apply your knowledge to this particular project	11:33
gitinfo	ravenzz: 'Pro Git' is a complete book about git that is available online for free (there's a print version, too): http://progit.org/book/	11:33
ravenzz	oh I see	11:33
	→ afief joined	11:33
cmn	developing a wp plugin might have its own nuances, but the general git workflow should still apply	11:35
ravenzz	Before to start I would be sure that I can work just in the production server (locally). And I can setup the server and deploy(?) it when it will be finished	11:35
	eh indeed i am going to work on an entire website, it will include some plugin development as well	11:36
cmn	setting up the deployment later is no problem at all	11:36
ravenzz	ok than you. last question do you know any free service (like github) that offer free private repositories?	11:39
cmn	bitbucket	11:39
	at least they used to; I haven't checked in a while	11:40
ravenzz	cool	11:40
	Store all of your Git and Mercurial source code in one place with unlimited private repositories. <-	11:40
	→ stringoO joined	11:40
ravenzz	I owe you a beer	11:40
	← Textmode left	11:44
	→ Laurenceb_ joined	11:45
Laurenceb_	hi, im trying to upload to github over ssh	11:45
	← psoo left	11:46
Laurenceb_	Permission denied (publickey).	11:46
	fatal: The remote end hung up unexpectedly	11:46
	when i try git push on the remote machine	11:46
	how can i fix this?	11:46
cmn	your ssh isn't configured properly	11:46
	← Zenopus left	11:47
Laurenceb_	what is wrong with it?	11:47
cmn	it doesn't know which key to use to authenticate against github	11:48
Laurenceb_	oh right	11:48
FauxFaux	Laurenceb_: !ssh	11:48
gitinfo	Laurenceb_: [!gitolite_ssh] See http://sitaramc.github.com/gitolite/sts.html for steps to troubleshoot ssh/gitolite	11:48
	← detaer left	11:48
Laurenceb_	ok thanks	11:48
	bbl	11:49
	← eletuchy left	11:50
	→ eletuchy joined	11:50
	← Cromulent left	11:52
	← boombatower left	11:52
Laurenceb_	i dont see my error descriped there	11:54
FauxFaux	Laurenceb_: Being prompted for a password, as that document puts it, is the same as "permission denied (publickey)", your server config is jstdifferent.	11:55
	→ na3r joined	11:55
Laurenceb_	so how do i fix this?	11:55
	ive never done this before	11:56
FauxFaux	By diagnosing it using the document.	11:56
Laurenceb_	only ever uploaded directly from the machine	11:56
FauxFaux	You've never diagnosed a problem? You're in for a fun suprise!	11:56
	→ Textmode joined	11:58
	→ Dave^\|\| joined	11:59
	→ shruggar joined	11:59
	→ dw joined	11:59
	→ pidus joined	12:00
erickr	I'm thinking about using submodules for a code tree setup, but I am a bit uncertain its the best practice in this case. I have a main project-tree which have 10 deployments, the deployment consists of a config-file, some local settings, plugins and the main project tree. With subversion i used svn externals to always have the latest main project, but is setting the main project as a submodule of the deploymentrepo the best w	12:01
	← Dave^\| left	12:02
	← pidus left	12:03
FauxFaux	erickr: a) you got cut off after "repo the best". b) !subproje	12:03
gitinfo	erickr: [!subprojects] So, you want to add git repositories inside of other git repositories? Well, you have four main options. First is to just do it, add the repo to the outer project's .gitignore, and treat them entirely separately. Best if they are entirely seperate. Otherwise your best options are "!submodule" "!gitslave" and "!subtree" Try typing those commands into this IRC channel.	12:03
j416	yuck add it to gitignore o_O who recommended that..	12:04
	oh well.	12:04
erickr	FauxFaux: wasnt much after that, last sentence: "With subversion i used svn externals to always have the latest main project, but is setting the main project as a submodule of the deploymentrepo the best way to go? "	12:04
jast	j416: ask the edit history :)	12:05
erickr	!gitslave	12:05
gitinfo	gitslave (http://gitslave.sf.net) is useful to add subsidiary git repositories to a git superproject when you control and develop on the subprojects at more or less the same time as the superproject, and furthermore when you typically want to tag, branch, push, pull, etc. all repositories at the same time.	12:05
j416	jast: it was more of a rhetorical question. :)	12:05
erickr	!subtree	12:05
gitinfo	The git subtree merge method is ideal to incorporate a subsidiary git repositories directly in to single git repository with "unified" git history, where you only need to pull changes in from external sources not contribute your own changes back (which if technically possible is at least difficult). See http://progit.org/book/ch6-7.html Type "!subtree_alt" for more options	12:05
j416	ty	12:05
jast	:)	12:05
	← ravenzz left	12:06
	→ steffo joined	12:06
FauxFaux	j416: I have a project somewhere that'll pick up a dependency either from the system (common case), or, if it exists, a subdirectory. For that, the most common case is to just clone another module directly into the subdirectory. Normally development is independent etc.	12:06
j416	FauxFaux: I have a habit of cleaning out stuff with 'git clean -dfx' -- that would destroy an ignored nested repo.	12:07
	ignored files, to me, are files I don't care about	12:07
	→ sattu94 joined	12:08
jast	I never use git clean	12:08
	→ bk912131 joined	12:08
j416	it's very handy, particularly if you are forced to use systems other than git that pollute your directories with stuff	12:08
jast	there's usually a target in the build system for cleaning away most files I don't want	12:08
FauxFaux	j416: True. I use git clean -fdx when some state in my build system (or ide) is being pissy.	12:09
	← giallu left	12:09
erickr	FauxFaux: seems gitslave is good, but might be to complicated for this..	12:09
	→ pantsman- joined	12:09
j416	it would be cool if submodules could just be.. nested repos and treated as such, no config, no anything	12:10
erickr	FauxFaux: the gitignore, however ugly trick might be easiest in this case since the deploymentpart is really small...	12:10
j416	but I guess it's not that simple	12:10
erickr	j416: yeah, i like the way svn externals worked for this, i just svn up at the root and it more or less worked..	12:10
j416	I prefer symlinks to nesting repos.	12:10
FauxFaux	erickr: I really don't know anything about your usecase as I'd ignore it like the plague, as, I believe, would most people here. The aliases are all we have.	12:10
erickr	probably not.. :)	12:10
rue	j416: You have to configure it somewhat, otherwise you may as well just pull the code into the tree and commit it	12:11
	← arvind_khadri left	12:11
j416	rue: I guess.	12:11
	rue: haven't thought much about i.t	12:11
rue	Or, rather, if you don't need to configure it, you might as well	12:11
j416	it*	12:11
	→ eijk joined	12:11
erickr	FauxFaux: yes, thanks anyway for the pointers.. :)	12:11
FauxFaux	Yeah, someone just get on with it and implement subtree clones, and people will be much more willing to just bloat their repo. ¬_¬	12:12
j416	good point	12:12
	still tricky though	12:12
	shared libraries and that	12:12
	← pantsman left	12:12
	← pantsman- left	12:14
	→ bitkiller joined	12:15
	← airborn left	12:15
shruggar	the real problem with dependencies in code is there are two entirely separate cases with contradictory needs, both of which are required for sane development. There is the "I want a version which works with external resources /now/, regardless of what version the rest of the code saw at that point in history, so I can test other things" case, and there is the "I want the exact version that existed at this point, so I can test the interaction between thin	12:16
	→ Zenopus joined	12:17
	← iamjarvo left	12:18
erickr	shruggar: sounds like the easiest way then is to handle them as separate repos.. then you can branch back and forth any way you like..	12:18
	especially in development.	12:19
	deploying to production should be something like fetch version x and version y which we know works together..	12:19
cbreak	submodules? :)	12:20
erickr	is that the same?	12:20
	:)	12:20
	← herdingcat left	12:20
j416	the cleanest solution I have come up with so far is 1) if the projects are not independently versioned, then they are the same project = the same repo. 2) if they are independently versioned, put them in separate unrelated repos and implement controls in your software to check that the right version is being used	12:20
shruggar	submodules killed my dog and forgot to feed my hamster	12:20
erickr	i tried reading the submodule page but I must have missed the important steps..	12:20
cbreak	submodules are a very good solution to the problem that you want to bind each parent commit to a submodule commit	12:21
	and for development, they allow you to be at a branch tip as well	12:21
j416	of course if you have zillions of dependencies, (2) could be a pain	12:21
cbreak	(You just have to not use submodule update)	12:21
erickr	j416: good separation, in regards to 1, can you in some way exclude some folders when cloneing a repo?	12:22
j416	erickr: no	12:22
	→ Cromulent joined	12:22
cbreak	that makes no sense	12:22
erickr	haha.. :)	12:22
shruggar	at the very least, right now I need to make repos for everything so one doesn't need to run 10 entirely different commands in order to get a working checkout. But then, at the end of the day, anything which /can/ be gotten through apt-get, we get through apt-get	12:22
cbreak	part of a repo is useless	12:22
	every snapshot is bound together by a single tree hash	12:23
shruggar	eg: Imagemagick. I'm not planning on hosting that	12:23
erickr	i think my current problem with this is betweeen keyboard and chair at the moment..	12:23
	← bk912131 left	12:23
erickr	cbreak: if I create a releasebranch, can I include only 2 of the repos mainfolders in that?	12:24
	← whyking left	12:25
j416	erickr: why would you want to release half a project?	12:25
cbreak	erickr: you can delete the other folders, but then the project will be incomplete	12:26
	← Mikachu left	12:26
cbreak	and the first commit in the branch would be "deleting half the project. Why did we even write that code?"	12:26
j416	erickr: if you have external things in your repo too, then that is part of your project whether you like it or not -- if you don't need it, don't add it to the repo. It's needed as much for the release as for development, is it not.	12:26
erickr	i have 11 folders, one is shared between the others but the 10 others only needs itself and the one global one..	12:26
	→ Mikachu joined	12:26
cbreak	then why are they in the same repository?	12:27
erickr	well, thats perhaps the problem..	12:27
j416	erickr: I would version the global one and create 10 more repos for the others.	12:27
cbreak	you should have 11 different ones.	12:27
erickr	good	12:27
	lets say I do that,	12:27
cbreak	and I'd use submodules in this situation	12:27
	but gitslave might also work	12:28
erickr	ahh..	12:28
cbreak	depending on how strong you want your binding to be	12:28
erickr	currently I just want the latest master of the global one and the local one, but later on i'll likely need to pin them both to a specific version when deploying..	12:29
	← DarkAR left	12:29
erickr	thanks for the brain-sparring, i'll probably split it up in multiple repos..	12:30
	→ berserkr joined	12:30
	→ mvrilo joined	12:30
	→ bouzbou joined	12:30
j416	the problem with submodules that I see..	12:31
	is that you have to remember to update them in your projects	12:31
	in case someone has done some update	12:31
SethRobertson	They are annoying.	12:31
	← Zenopus left	12:31
	→ fserb joined	12:31
	→ whyking joined	12:31
SethRobertson	gitslave fixes your different repos to each other at tag boundaries	12:31
j416	s/the problem/a problem/	12:32
SethRobertson	However, if you think you will need it fixed at every commit, you need the annoyance of submodules	12:32
erickr	j416: yeah, well, that could at least be fixed but forcing us all to remember. :)	12:32
j416	SethRobertson: interesting	12:32
	erickr: trying to force people to remember to do things.. won't work very well	12:32
cbreak	submodules aren't hard to forget	12:33
erickr	j416: well, then, documenteted then?	12:33
	:)	12:33
cbreak	if they don't match, they show up in every git status	12:33
j416	for people to do things regularly, it has to be dead simple and take no time	12:33
erickr	j416: can do a script..	12:33
SethRobertson	Or break things so utterly it is in your face	12:33
cbreak	git submodule update is rather easy :)	12:33
SethRobertson	!submodule_change	12:33
gitinfo	In order to change a submodule you must go into the submodule repository, check it out to the appropriate branch, make the needed change (possibly involving git pull), commit the change, cd .. (out of the submodule), git commit -m "Updated submodule" submodulepath	12:33
	← Cromulent left	12:33
j416	cbreak: 'git fetch' also fetches submodules?	12:33
	→ Zenopus joined	12:34
cbreak	update does	12:34
	→ psoo joined	12:34
cbreak	I think...	12:34
	fetches and resets them	12:34
j416	that's the thing.. you need to run update to check if there is a new version then	12:34
	→ bk912131 joined	12:34
	← bk912131 left	12:34
cbreak	no	12:34
	→ sacred_coder joined	12:34
SethRobertson	--recurse-submodules[=yes\|on-demand\|no]	12:34
j416	'git fetch' is enough	12:34
	?	12:34
cbreak	the parent decides wether there's a new version	12:34
SethRobertson	in git-fetch	12:34
j416	the parent decides?	12:34
cbreak	if you have a new parent, then git status will tell you that it wants a different child sha	12:34
	strong binding :)	12:35
j416	what do you mean by parent in this case?	12:35
cbreak	updating is the easy part in git submodules	12:35
SethRobertson	superproject	12:35
erickr	so each parent will include a pointer to a commit of the submodule?	12:35
cbreak	the parent is the git repository that contains the submodules	12:35
j416	ok	12:35
cbreak	each commit will contain the sha of the commits in the submodules it wants	12:35
j416	cbreak: and if one of the subprojects is updated, how will the superproject know?	12:36
erickr	that sounds perfect.	12:36
cbreak	if you want to change which commit the parent points at, you have to change that inside the submodule dir as seth already mentioned	12:36
	j416: the superproject doesn't care	12:36
j416	cbreak: that's what I mean	12:36
cbreak	it wants the same commit until you tell it that it wants a different one	12:36
SethRobertson	and then commit that sha in the superproject	12:36
	← disappearedng left	12:36
j416	cbreak: if you're working in the superproject you need to keep track of if there are changes in the subproject	12:36
cbreak	not really	12:37
j416	how so?	12:37
erickr	so, to deploy a change in all superprojects, I would need to commit in each superproject?	12:37
	→ disappearedng joined	12:37
j416	erickr: yeah.	12:37
cbreak	I only need to change the submodule commit if I want new things	12:37
	that's the whole point of submodules	12:37
SethRobertson	With gitslave you essentially treat the set of repos like one repo and use `gits <command>` instead of `git <command>`	12:37
cbreak	_strong_ binding	12:37
erickr	j416: that sucks in our case since most development is done in the submodule in this case,	12:37
SethRobertson	gitslave may be for you then	12:37
j416	SethRobertson: I'll look into that, thanks	12:38
cbreak	how does that handle shared subrepos SethRobertson ?	12:38
j416	the problem for us would be the risk of a superproject not having the latest submodules and being deployed	12:38
erickr	SethRobertson: using a lot of addons and extensions feels a bit like going to use mercurial.. :)	12:38
	j416: exactly.	12:38
cbreak	j416: for users of git-submodule, that's not a risk, that's a benefit :)	12:38
	because the latest version might break stuff	12:38
erickr	true though..	12:39
SethRobertson	cbreak: You can either develop on different branches or insist that commits pass multiple regression tests	12:39
cbreak	if you've ever had ffmpeg as submodule... you know what I mean	12:39
	gitinfo set mode: +v	12:39
j416	cbreak: then I guess submodules is not for our kind of projects	12:39
bouzbou	Hi, is there a way to save a ssh password somehow in git's configuration in order to be able to pull and push without entering it each time?	12:39
j416	bouzbou: private/public keys	12:39
cbreak	bouzbou: just use ssh public/private keys	12:39
SethRobertson	bouzbou: Use ssh-agent or pagaent	12:39
j416	bouzbou: not a git thing, an ssh thing	12:39
cbreak	maybe you can write your password into the UNENCRYPTED SSH CONFIG FILE	12:40
	...	12:40
	(do not!)	12:40
bouzbou	okay so it has nothing to do with git? it's a key that I have to generate and that's it?	12:40
	← Zenopus left	12:40
cbreak	you have to tell the server that that key is authorized of course	12:40
	google for .ssh/authorized_keys	12:41
	or what ever...	12:41
j416	googling for whatever likely won't help	12:41
	← noz left	12:41
	← disappearedng left	12:41
bouzbou	ok thank you I'm going to try to find some tutorial on it	12:42
j416	bouzbou: http://www.google.se/search?sourceid=chrome&ie=UTF-8&q=setting+up+private+public+keys+ssh maybe	12:42
erickr	bouzbou: http://www.csua.berkeley.edu/~ranga/notes/ssh_nopass.html	12:42
	→ txomon\|home joined	12:42
	← txomon\|home left	12:42
	→ txomon\|home joined	12:42
bouzbou	erickr: thanks for that link, I'll follow the instructions :)	12:43
drizzd	While I am running the test suite I am essentially blocked from doing any other changes. Any ideas how to fix that?	12:43
cmn	cp -r; or use a different workdir for other changes	12:45
erickr	j416: I'll just go and split up my repo now. :)	12:45
j416	erickr: :)	12:45
	← infid left	12:46
erickr	j416: can we conclude that subtrees suffer from the same problam as the submodules?	12:46
j416	erickr: I have used neither, so I cannot tell. I think I am allergic to every solution still.	12:46
	← bouzbou left	12:46
	→ canton7 joined	12:46
j416	erickr: if you find a good solution to your problem, please do tell	12:46
	→ Zenopus joined	12:46
j416	erickr: we're kind of in the same situation :)	12:47
	erickr: only, we haven't switched to git yet ($dayjob)	12:47
	→ munichlinux joined	12:49
	← fserb left	12:51
	← berserkr left	12:52
	→ Spockz` joined	12:52
	→ infid joined	12:53
	→ mishok13 joined	12:54
	→ berserkr joined	12:55
	→ LeMike joined	12:55
	← Spockz` left	12:55
	→ Spockz` joined	12:56
	← ipalaus left	12:57
erickr	j416: i havent switched either, checking all the angles before..	12:57
j416	erickr: sounds good. :)	12:57
erickr	you have weborojects or others?	12:58
j416	at $dayjob, windows dev	12:58
	→ ipalaus joined	12:58
j416	personal projects mostly anything	12:58
	→ crisp joined	13:00
	← Zenopus left	13:02
	→ pro_metedor joined	13:02
	→ arvind_khadri joined	13:03
	→ gianlucadv joined	13:04
	← FernandoBasso left	13:05
	← pro_metedor left	13:06
	← cesc left	13:07
	← nogginBasher left	13:07
	← resmo left	13:08
	← psoo left	13:09
	→ teddybeermaniac joined	13:11
	← teddybeermaniac left	13:15
	→ resmo joined	13:23
	← afief left	13:25
	← pdtpatrick_ left	13:26
	→ afief joined	13:27
	→ bouzbou joined	13:28
	→ markiv joined	13:28
	← arvind_khadri left	13:29
	← markiv left	13:31
	gitinfo set mode: +v	13:31
bouzbou	Hi, I've installed a public key on my server and generated a private key, I'm able to login to my server via Putty, now I'd like to configure my private key with git, I'm on Windows, how can I do that? thanks	13:31
	→ fayimora joined	13:31
	→ iamjarvo joined	13:32
	→ markiv joined	13:33
	← berserkr left	13:33
canton7	bouzbou, did you tell the msysgit installer to use plink over openssh when you installed? If so, you're good to go -- just load the key into pageant. If not, make sure you've got a non-putty private keyfile (if you're not sure, open puttygen, conversion -> import key, import your ppk, save private key. It needs to be called "id_rsa", and stashed in C:\Users\<name>\.ssh\	13:36
	bouzbou, erm, you might need to "load existing private key" over "conversions -> import key", it's been a while since I used puttygen	13:37
bouzbou	okay, I'm trying to do that	13:38
	thanks	13:38
	→ ypcat_ joined	13:39
erickr	any other swedish lamp-programmers in here looking for new challenges, of course with git involved. :)	13:39
	← ypcat_ left	13:39
	→ markiv_ joined	13:40
	← ypcat left	13:40
	→ ypcat joined	13:41
	→ Zenopus joined	13:41
	← markiv left	13:41
	markiv_ → markiv	13:42
	→ berserkr joined	13:43
	→ Zenopus_ joined	13:43
	bremner avoids programming swedish lamps	13:44
bouzbou	canton7: I don't really get what you mean by "plink over openssh", I've put my id_rsa and id_rsa.pub files in C:\Users\Bouzbou\.ssh and I'm still beign asked for my password when I do a git pull or push	13:45
	← Zenopus left	13:45
	← txomon\|home left	13:46
	← Zenopus_ left	13:48
	→ txomon\|home joined	13:48
	← txomon\|home left	13:48
	→ txomon\|home joined	13:48
bouzbou	I followed a tutorial and keys seems to be working, I can connect via Putty with my private key and I don't have to enter my password, that's great, now I guess I only have to tell git (in a config file or something like that) where that private key is? Am I wrong?	13:51
	← markiv left	13:52
	→ eikke joined	13:53
canton7	bouzbou, putty uses a different private key format to openssh. Thing is, I'm currently failing to remember how to tell putty to save your private key in openssh format...	13:54
erickr	bouzbou: no idea how it works on windows..	13:54
	← [M]ax left	13:55
canton7	bouzbou, when you install msysgit, you should get asked whether you want to use plinkk (part of the putty family, uses .ppk keys, pageant for key caching, etc), or openssh	13:55
erickr	bouzbou: i would say switch os, but it seems a bit radical. :)	13:55
	← tro left	13:57
bouzbou	canton7: okay, and I need openssh instead of plink?	13:57
	← nikuyoshi left	13:57
	→ [M]ax joined	13:57
bouzbou	erickr: Yes I'll think about that some day ;)	13:58
canton7	bouzbou, you can use either with git, but the msygit installer needs to be told to use plink if you want to use plink.	13:58
	bouzbou, Right now, I think your options are: 1. Generate a new key using openssh. 2. re-install msysgit, telling it to use plink	13:59
	bouzbou, ooh! Apparently if you load your putty private key into puttygen, you can save as an openssh key using the conversions menu	13:59
bouzbou	canton7: I think it's already openssh, PuttyGen displays : "Public key for pasting into OpenSSH authorized_keys file"	14:00
canton7	bouzbou, that doesn't mean what you think it means. puttygen provides the openssh public key as a courtesy, as most servers use openssh	14:00
	← afief left	14:00
	→ jasonbray joined	14:01
bouzbou	ok, so conversions, export openssh key	14:01
canton7	yup	14:01
bouzbou	ok I'm putting it in my .ssh folder in my user folder and I'm trying again	14:02
	→ Cromulent joined	14:02
	← jasonbray left	14:03
bouzbou	canton7: great thank you so much it's working! :)	14:04
canton7	bouzbou, sweet!	14:04
	← sven^ left	14:04
	→ afief joined	14:04
bouzbou	I have a last question! :) My git repo is hosted on the same server than my website	14:05
	← psakrii left	14:05
	← Spockz` left	14:06
bouzbou	I've created a shell script to deploy the website in production, with a few command and ther is a git pull in it, when I launch that script it's asking for the server password, do I need to put my private key on my server too?	14:06
cbreak	do NOT use git pull in a script	14:07
	it's dumb	14:07
canton7	!website	14:07
gitinfo	Git is not a website deployment tool, but can sometimes play one in sufficiently simple/lax environments with a little help. One example of help is: http://toroid.org/ams/git-website-howto	14:07
cbreak	if you want the server to be able to fetch, then just give it its own private key	14:07
	← dvide left	14:07
	← secoif left	14:09
	← fayimora left	14:09
bouzbou	okay, so it's wrong to use git pull to update my production website? Then how should i do?	14:09
cbreak	pulling is interactive	14:10
canton7	the main problem with "git pull" is that it does a merge, which can lead to conflicts	14:10
cbreak	you can NOT do it properly in a script, unattended	14:10
canton7	then you have a website which is showing "<<<<<<<<<<<" everywhere	14:10
cbreak	if you have to use git, use git archive or git reset --hard	14:10
bouzbou	okay but if I do NEVER update any single line directly in the production, there won't be any conflict right?	14:11
canton7	then why not just use reset --hard ?	14:11
Mikachu	it's also a race on update, in case the webserver tries to read a file at the moment git writes it, or if files depend on eachother it might read an old version of one file and a new version of another and get some weird error	14:12
	the best way is to update the whole tree atomically	14:12
	→ diegoviola joined	14:12
Mikachu	and make sure the webserver isn't reading anything while you do it	14:12
cbreak	I would use git archive	14:12
	← afief left	14:12
cbreak	you can write the whole tree into a new directory, and then change a symlink in the web root	14:13
	which should be relatively atomic	14:13
canton7	^^ that's what "proper" deployment systems do, that I've seen at least	14:13
	→ Heisenmink joined	14:13
Mikachu	yeah, except the server doesn't read the files atomically :)	14:13
cbreak	doesn't matter	14:13
Mikachu	the really paranoid way is to change the webroot in the config and do a graceful restart (in apache terms)	14:13
cbreak	since you never change files	14:13
	you just write new ones	14:13
Mikachu	cbreak: it might still read the old version of a file and a new version of another file, eg in an include() statement	14:14
	→ rzec joined	14:14
cbreak	no	14:14
	include are relative :)	14:14
	(unless you write bad code)	14:14
	maybe php is bad.	14:14
Mikachu	heh, i guess that depends on the scripting language too	14:14
	i don't know any in enough detail	14:14
	← gianlucadv left	14:15
bouzbou	cbreak: okay so a clean deployment should be, your production is only a folder with a symlink, and when you want to update, you generate a new folder, and then change the symlink to that new folder?	14:16
cbreak	that's what I would do	14:16
	→ tomislater joined	14:16
cbreak	but I am a real developer, not a web programmer, so I have little experience with that :D	14:16
cmn	you probably want to invalidate any caches you have	14:17
cbreak	yeah	14:17
	that makes sense	14:17
	→ _inc joined	14:17
cbreak	and also consider side effects	14:17
	like changes in dependent external data like a database	14:18
	← awx left	14:18
bouzbou	and in the deployment script git archive? to create the folder with the tree structure?	14:18
	because with git clone I get the .git folder too	14:18
cbreak	maybe you should ensure that all old scripts have stopped executing before changing the symlink	14:18
cmn	right, you don't want to kill apache whilst the script is updating stuff on the database	14:18
cbreak	bouzbou: you'd use git clone --bare	14:18
	and for deployment git archive \| tar	14:18
	→ awx joined	14:19
bouzbou	ok, thank you cbreak	14:19
	have a good day all :)	14:19
	← bouzbou left	14:19
	→ jasonbray joined	14:19
cbreak	maybe there's a way to prevent old instances of the script from being started	14:20
	← Gitzilla left	14:20
cbreak	without shutting down apache, or doing something else that causes errors	14:20
	→ dv310p3r joined	14:20
	→ edude03 joined	14:25
	→ Chillance joined	14:28
	← T-Gunn left	14:29
	→ psoo joined	14:29
	← shiba_yu36 left	14:29
	→ gavin_huang joined	14:32
	← jargon- left	14:32
	→ jonesy joined	14:33
	→ shiba_yu36 joined	14:35
	← n2diy left	14:37
	→ stefankolb_ joined	14:38
	→ n2diy joined	14:39
	stefankolb_ → stefankolb	14:39
	← stefankolb left	14:40
	→ Zenopus joined	14:41
	→ munichpython joined	14:43
	← mvrilo left	14:44
	← tatsuya_o left	14:44
	→ tatsuya_o joined	14:46
	← munichlinux left	14:46
	← drale2k left	14:46
	← Laurenceb_ left	14:46
	→ dreyescat joined	14:47
	← Cromulent left	14:47
	← dreyescat left	14:48
	→ dreyescat joined	14:49
	← Zenopus left	14:49
	→ Fandekasp joined	14:51
	→ munichlinux joined	14:51
	→ madewokherd joined	14:52
	← munichpython left	14:53
	← dreyescat left	14:53
	← flaguy48 left	14:53
	→ nikuyoshi joined	14:54
	→ bartman joined	14:54
	← tomislater left	14:55
	← uu left	14:55
	→ Goplat joined	14:56
	← bindaasomatic left	14:57
	→ Zenopus joined	14:59
rzec	will all errors that prevent a git command from continuing to run be prefixed with fatal: or error:?	15:01
	→ mvrilo joined	15:01
	← munichlinux left	15:01
	← jasonbray left	15:02
cmn	most of the code calls die() which prints "fatal"	15:03
	→ Destos joined	15:03
cmn	but the error code is what you should probably be using	15:03
	→ macrover joined	15:03
	← rebe left	15:04
	← eikke left	15:04
	tizzo-afk → tizzo	15:05
	→ munichlinux joined	15:08
	← tatsuya_o left	15:10
	→ tatsuya_o joined	15:10
	→ Gitzilla joined	15:11
	← iocor left	15:11
	→ codewrangler joined	15:13
	→ sandbags joined	15:13
rzec	cmn, well I am using python to create some scripts, can I get the error code using that/	15:15
shruggar	rzec; how are you calling git?	15:15
cmn	probably	15:15
rzec	shruggar: subprocess.Popen	15:16
shruggar	also, I assume python has a git library, why not use that?	15:16
	→ brian_g joined	15:16
rzec	from what I have read, most of the git library are subpar, a lot of people recommend just using subprocess	15:16
	→ iocor joined	15:17
	← awx left	15:18
jast	most libraries just call the git binary	15:18
shruggar	most libraries just call the git binary, but probably already have error detection worked out	15:19
rzec	how would I call the git binary from python?	15:20
shruggar	looks like the object given by subprocess.Popen has a "returncode" property after termination	15:20
	→ awx joined	15:20
shruggar	if that is nonzero, something went wrong	15:20
cbreak	or right	15:20
	some commands use the exit code for signaling information	15:20
	→ Raging_Hog joined	15:22
shruggar	let's burn those commands	15:23
	→ guide joined	15:23
	← sattu94 left	15:24
	tizzo → tizzo-afk	15:24
	→ neurodrone joined	15:24
cbreak	shruggar: I like it	15:25
	then you can do things like if git diff --exit-code HEAD; then	15:25
	← dv310p3r left	15:26
rzec	ok, I will assume a non-zero returncode is a git error	15:27
shruggar	cbreak: okay, I accept the utility when such a think is explicitly asked for	15:27
	→ flaguy48 joined	15:29
	← mishok13 left	15:29
shruggar	I will similarly not complain that "false" exits with a code of 1 :)	15:31
	→ yots joined	15:32
cbreak	git rev-parse --is-inside-work-tree	15:32
	I am not sure I like it	15:33
	it just errors when I am not in a work tree, unlike it is documented	15:33
	→ kenperkins joined	15:34
	→ warez joined	15:34
raz	gnah	15:35
	i wish git subprojects weren't such a royal pita	15:35
	← t-mart left	15:38
	→ icrazyhack joined	15:38
	→ Cromulent joined	15:39
	← Error404NotFound left	15:40
	→ FernandoBasso joined	15:40
bremner	I guess you mean submodules?	15:40
	← proq left	15:41
	→ proq joined	15:41
	→ jaisoares joined	15:43
	→ oriba joined	15:44
kini	I have two branches A and B whose common ancestor is pretty far back. I created a branch C at A, and commited some changes to C which only create new files and don't modify old ones. Is there a way to rebase A..C onto B (instead of C onto B)?	15:47
	→ mandric joined	15:47
kini	or does that not even make sense?	15:47
	→ centipedefarmer_ joined	15:47
	← kenperkins left	15:47
	← tatsuya_o left	15:48
	→ eMBee joined	15:48
	← innoying left	15:48
canton7	kini, do you mean the example in man git-rebase that starts with "Another example of --onto option is to rebase part of a branch" ?	15:48
gitinfo	kini: the git-rebase manpage is available at http://jk.gs/git-rebase.html	15:48
	→ tatsuya_o joined	15:48
luke-jr	110 GB git repo, but can't figure out why… biggest object according to some script bremner recommended was 10 KB :/	15:49
bremner	well, the script could be wrong ;)\	15:49
	luke-jr: did you look at the logs?	15:49
kini	canton7: wow, how did I miss that... thanks :)	15:49
Mikachu	luke-jr: what happens if you clone it?	15:49
	← icrazyhack left	15:50
luke-jr	bremner: no, what logs?	15:50
bremner	err, git log	15:50
luke-jr	Mikachu: don't have enough disk space to clone it…	15:50
	bremner: oh, what am I looking for there?	15:50
bremner	committing gigantic binaries? more than a million commits?	15:50
	← ethanol left	15:51
luke-jr	that script wouldn't show gigantic binaries?	15:51
	→ ethanol joined	15:51
	← ethanol left	15:51
	→ ethanol joined	15:51
luke-jr	counting commits… will probably take a while	15:51
bremner	all software has bugs. That goes double for scripts grabbed from blogs	15:51
luke-jr	it would not surprise me if there were over a million commits	15:51
bremner	all with binaries?	15:51
luke-jr	all text	15:52
	← martinjlowm left	15:52
canton7	kini, heh. That's the only bit of that manpage I keep going back to :P	15:52
kini	:)	15:52
luke-jr	164,110 commits	15:52
	→ Error404NotFound joined	15:52
Mikachu	luke-jr: what does git count-objects -v say?	15:52
luke-jr	count: 4990 size: 509380 in-pack: 1099229 packs: 2 size-pack: 2464119	15:53
	→ martinjlowm joined	15:53
Mikachu	that's only 3GB, what does du -hs .git say?	15:54
luke-jr	103G	15:54
Mikachu	cd .git; du -ms * \| sort -n?	15:54
luke-jr	105191 objects	15:54
	fwiw, the origin repo for this is 18G	15:55
Mikachu	and same in objects?	15:55
luke-jr	104692 pack	15:55
Mikachu	and in pack? ;)	15:55
luke-jr	a ton of tmp_* files	15:56
Mikachu	you can delete those	15:56
luke-jr	http://paste.pocoo.org/show/538279/	15:56
cmn	those would be from failed fetch attempts	15:56
luke-jr	>_<	15:56
Mikachu	git gc would also delete them, iirc	15:56
luke-jr	cmn: or push?	15:56
cmn	failed as in someone killed the process	15:56
luke-jr	Mikachu: it didn't	15:56
	cmn: oh, fun	15:56
cmn	I don't think push would create them	15:57
	→ johnkpaul joined	15:57
luke-jr	well, this repo is only pushed to…	15:57
jast	pushed to is a different story	15:57
cmn	that's like fetch	15:57
jast	that would probably create temporary packs, yes	15:57
luke-jr	ok	15:57
	so since it seems to make them often, any way to get git to delete them more proactively? ;)	15:57
	→ alx- joined	15:58
luke-jr	maybe a pushto hook?	15:58
	or is it unsafe there?	15:58
cmn	I would have thought git-receive-pack would delete them if it has a problem	15:58
	← elenril left	15:58
	→ kenperkins joined	15:58
luke-jr	2.9G after deleting all tmp_*	15:59
cmn	heh	15:59
jast	cmn: yeah, unless it got killed	15:59
cmn	right, in which case there isn't really anything you can do	15:59
luke-jr	I'll admit to killing git a few times, but nowhere near this often :P	15:59
	← johnkpaul left	15:59
jast	you could clean them up in a cronjob	15:59
Mikachu	make sure to only delete files not currently being written to though	15:59
jast	of course, who knows when these were created. perhaps it was some kind of exceptional case.	15:59
cmn	but you'd have to make sure you're not deleting a pack that's being written to	15:59
luke-jr	curiously, my other such repo has none	15:59
	jast: I should have looked at the timestamps before deleting…	16:00
	← thomas_b left	16:00
	→ alberto56 joined	16:00
jast	yes :)	16:00
	→ techhelp joined	16:00
	← martinjlowm-lptp left	16:00
	← RaptorX left	16:00
	→ RaptorX joined	16:00
	← martinjlowm left	16:00
luke-jr	hum, I guess if this repo is ♥G and the origin is 18G, time to `git gc' that too :P	16:00
	err	16:00
	< 3G	16:00
SamB	lol	16:00
Mikachu	git prune should delete them	16:00
	→ OOPMan joined	16:00
Mikachu	maybe gc doesn't	16:00
jast	♥G	16:01
	I'm going to make all my git repos that size from now on	16:01
	→ elenril joined	16:01
Mikachu	cute❣	16:01
luke-jr	:p	16:01
SamB	Mikachu: what in the world ?	16:01
luke-jr	if anyone else wants it, http://gitorious.org/gitbackup/gitbackup is the tool I use for these :P	16:01
Mikachu	you were not aware of the heart exclamation point in unicode?	16:02
cbreak	git gc --aggressive should clean up all garbage	16:02
SamB	indeed not!	16:02
Mikachu	what about pile of poo?	16:02
luke-jr	☠	16:02
SamB	do they have a heart-i and heart-j, too?	16:02
luke-jr	⚀ ⚁ ⚂ ⚃ ⚄ ⚅	16:02
Mikachu	i don't think so	16:02
jast	prune/gc don't clear out temporary pack files	16:03
SamB	but I've actually heard of those characters!	16:03
Mikachu	there's code in builtin/prune.c that deletes tmp_*	16:03
jast	only tmp_obj_	16:03
Mikachu	oh, i guess i didn't read carefully enough	16:03
jast	oh, hmm	16:03
	← Akari` left	16:03
jast	it has two cleanups	16:03
luke-jr	the joys of abusing git for maildir…	16:03
jast	one cleans tmp_obj_ only, the other appears to clean tmp_*	16:03
	← shiba_yu36 left	16:04
luke-jr	oh well, even if I can't automate it, at least I know where to look now ;)	16:04
	← Fandekasp left	16:05
luke-jr	so 3 GB for a git repo of my wife's maildir since 2009. pretty good compression :P	16:05
	→ Akari` joined	16:06
jast	files in maildirs don't tend to change much, so that makes sense	16:06
	plus a lot of headers are likely to be very similar	16:07
	→ frooh joined	16:07
	→ psakrii joined	16:08
	← frooh left	16:08
	→ frooh joined	16:08
frooh	is there a config option to default git pull (for all branches) to ff-only?	16:09
cbreak	yes	16:11
frooh	cbreak: what is it? I couldn't find it in the config docs	16:11
cbreak	unfortunately it is very ... overreaching	16:12
frooh	??	16:12
	like, it applies to all merges?	16:12
cbreak	yes	16:12
frooh	bah.	16:12
cbreak	frooh: man git-conf0g	16:12
	git-config of course	16:12
frooh	cbreak: got it open already	16:12
cbreak	look for merge.ff	16:12
frooh	huh, that's not there	16:12
jast	old git version, then :)	16:13
cbreak	and it can't even be overridden from the cli	16:13
frooh	1.7.8?	16:13
cbreak	frooh: man git-config, the online version should have it	16:13
frooh	ooh	16:13
gitinfo	frooh: the git-config manpage is available at http://jk.gs/git-config.html	16:13
frooh	lol	16:13
	docs are in the wrong folder	16:13
	I think it would be better for me to make an alias.	16:14
	cbreak: thanks	16:15
cbreak	I always use git pull --rebase	16:15
frooh	that's equally good.	16:16
	just not surprise merges :-)	16:16
	← gusnan left	16:16
	← Destos left	16:16
	← Akari` left	16:16
	→ sebrock joined	16:18
	← OOPMan left	16:20
	→ eikke joined	16:22
	→ amigojapan joined	16:23
	→ soulcake joined	16:23
	→ pantsman joined	16:24
	← pantsman left	16:24
	→ pantsman joined	16:24
	→ Akari` joined	16:26
	← macrover left	16:28
	← sacred_coder left	16:28
	→ sacred_coder joined	16:28
	← fisted left	16:30
	→ fisted joined	16:30
	tizzo-afk → tizzo	16:30
	→ stoffus joined	16:33
	→ sacred_coder_ joined	16:36
	← kenperkins left	16:41
	→ kenperkins joined	16:42
	→ pidus joined	16:43
diegoviola	is there a way to see the remote info other than "cat .git/config"?	16:46
Mikachu	git remote -v	16:46
diegoviola	ty	16:46
	git rocks, ty	16:46
	→ detaer joined	16:48
	← yots left	16:48
	← frooh left	16:48
	→ ehsan joined	16:49
	← johnanderson left	16:53
	← EricInBNE left	16:54
	← _inc left	16:56
	→ _inc joined	16:56
	→ gusnan joined	16:59
	→ mgpcoe joined	16:59
	→ Karmaon joined	17:01
	→ aalex joined	17:02
	→ muzone joined	17:03
muzone	Does anybody know why I'm getting this error? https://gist.github.com/1653291 - STATUS_ACCESS_VIOLATION, cannot fork(), Could not fetch origin	17:05
	← amigojapan left	17:05
	← pantsman left	17:06
	← jaisoares left	17:06
	← mgpcoe left	17:08
	→ rath joined	17:08
cmn	it looks like your server isn't allowed to execute programs; or something like that	17:10
	→ flor joined	17:10
cmn	it's a windows question, judging from the .exe extension	17:10
	← Heisenmink left	17:13
	← warez left	17:16
	← mvrilo left	17:17
	→ cjuner joined	17:17
muzone	cmn: it is indeed, im using cygwin on windows 7..	17:17
hh__	is it possible to clone a certain branch? and every time i pull i only pull changes from that branch?	17:18
cjuner	Something is wrong with git or my default pager. I have neither changed the settings for crlf handling nor have I set a custom pager. I see lots of "^[[m" characters in git diff's output (in every line, also the ++/-- lines) when there are no such characters in any of the files. If I run git diff \| less it's normal. Any ideas?	17:18
cmn	hh__: support for magically doing this is coming in clone, but for now you'll have to do it by hand; git init; git remote add -f --t branch -m branch origin <url>; git checkout -t branch	17:19
	cjuner: those look like terminal colour codes	17:20
	gitinfo set mode: +v	17:21
cmn	muzone: then I can't help you; other than suggest searching for allowing a web server to execute stuff or use a unixy OS	17:21
cjuner	If it helps, this is an example of what I get: "^[[1mdiff --git a/tests/unittests/Makefile.am b/tests/unittests/Makefile.am^[[m"	17:21
rath	cmn: is it coming in next release?	17:21
muzone	thanks cmn !	17:22
	← jdav_gone left	17:22
cmn	cjuner: right, looks like it's a colour code; either unset the colour options, or use a better pager/shell	17:22
cjuner	cmn, hm even with default settings in bash or dash I have the same problem.	17:22
	→ mvrilo joined	17:22
	→ malphaet joined	17:22
cmn	it could be your terminal emulator	17:22
	there are many layers	17:22
	rath: it probably won't come in 1.8.9	17:22
cjuner	cmn, it's gnome-terminal. Used to work fine. What pager with what options does git use by default? If I do git diff \| less everything's fine.	17:23
	← Cromulent left	17:23
cmn	git uses less by default	17:23
	by "everything's fine" do you mean that you see colours?	17:24
	← catsup left	17:24
	→ catsup joined	17:24
	← techhelp left	17:24
	← gavin_huang left	17:25
	→ toobluesc joined	17:25
	→ Destos joined	17:25
	← eikke left	17:26
cjuner	cmn, no, no colors.	17:26
	cmn, just no weird color codes either.	17:26
	But you're right, I configured git to use colors.	17:26
	git diff \| less -R or git diff \| less --raw also looks normal (but again without colors)	17:27
	xterm shows the same problem. So I guess I can rule out the terminal emulator as a source of the problem	17:27
cmn	right, so you told git to use colours when outputting to the terminal screen; so piping to less removes the colours	17:28
cjuner	Yes.	17:28
	But less -R should show them if they were input correctly.	17:28
	← stoffus left	17:29
cmn	it's odd that it would change; maybe you've changes your $LESS or whichever env var it reads the options from	17:29
cjuner	LESS or PAGER are not set to anything :S	17:29
	export PAGER=less makes it work fine	17:30
	including colors	17:30
	Maybe git tries to use some other pager by default (although I did not tell it to)?	17:30
cmn	it should try to use less by default	17:30
	maybe your $TERM has something wrong	17:31
	← iocor left	17:31
	→ dotnull joined	17:31
cjuner	TERM=xterm .. and I see the same effect in gnome-terminal and in the actual xterm	17:32
	→ eikke joined	17:32
Mikachu	cjuner: git uses less -FSRX by default	17:32
cmn	having xterm might cause some problems; shouldn't it be xterm-color?	17:33
	→ iocor joined	17:33
	← bgerber left	17:33
Mikachu	no	17:33
cjuner	"git diff \| less -FSRX" works but strips colors.	17:33
	export TERM=xterm-color does not change anything	17:33
Mikachu	if you explicitly pipe the output, color is disabled by default	17:34
	you'd have to say git diff --color=always	17:34
	← [M]ax left	17:34
cjuner	ah	17:34
	git diff --color=always \| less -FSRX works as well	17:35
	→ hseg joined	17:35
	→ t0rc joined	17:35
cjuner	Hm it seems ubuntu has git patched to use sensible-pager, which on my system appears to be something weird called "lv"	17:35
hseg	Hi. Is there a way to change my username?	17:35
Mikachu	in git or on irc?	17:35
	→ bgerber joined	17:35
hseg	git	17:35
cmn	hseg: are you looking for #github?	17:35
hseg	thx. sorry	17:36
	← hseg left	17:36
	→ [M]ax joined	17:36
Mikachu	nice save, thanks :)	17:36
cjuner	Ok, problem solved. Removed that weird lv pager.	17:36
	Thanks everyone.	17:36
	→ DofLeFou joined	17:38
	tizzo → tizzo-afk	17:39
	← sacred_coder left	17:41
	← msponge left	17:42
	→ valtih_ joined	17:44
	← valtih_ left	17:44
	→ valtih joined	17:45
LeMike	-- hi there. is it possible to put two repo together? one repo ends on 31/12/2011 and the other starts in 2012 with "the head of the first" but as a new bare repo.	17:46
	gitinfo set mode: +v	17:46
	→ jceb joined	17:46
cbreak	yes	17:46
	gitinfo set mode: +v	17:46
valtih	I am the only commiter to a remote repository. I developed my branch locally. Now, pushing, I'm getting conflicts. How it is possible? Why it just does not fast-forward?	17:46
cbreak	just git rebase the one on the other	17:46
	← DofLeFou left	17:47
cbreak	valtih: gitk origin/master master and find out	17:47
valtih	It shows me the conflicts with my own files	17:47
LeMike	okay. didn't found that. and can I search in whole repo (all revs) for a special folder or it's files via log?	17:47
	← _inc left	17:47
cbreak	LeMike: what are you talking about?	17:48
	← i42n left	17:48
valtih	why newer commits just do not take over the history?	17:48
cbreak	valtih: because that would suck?	17:48
	→ _inc joined	17:48
cbreak	just do what I told you	17:48
	do you see two lines of history? one?	17:49
	← sacred_coder_ left	17:49
cbreak	you can use git log --graph --oneline --decorate instead of gitk too, if you have no GUI	17:49
LeMike	cbreak: having 100 revisions with folders that the head doesn't have anymore since perhaps 50 commits or more	17:49
cbreak	LeMike: what?	17:49
	→ dreiss_ joined	17:49
cbreak	you want to find... a folder?	17:50
	if you know the path, git log -- path	17:50
LeMike	thats what i ment. thanks again cbreak! :)	17:50
valtih	will I need to merge every commit I want to push with my local commits? what the hell?	17:51
	→ nixmaniack joined	17:51
cbreak	no	17:51
	→ mishok13 joined	17:51
cbreak	why do you think that?	17:51
	all you have to do is to make the history you want to push be a descendent of the history that is already there	17:51
	so it can be fast forwarded	17:51
	everything else would make you lose history obviously	17:52
	which would be really bad	17:52
	← mvrilo left	17:52
	← malphaet left	17:52
	→ v0n joined	17:52
	→ drizzd_ joined	17:53
	→ mvrilo joined	17:53
	← drizzd left	17:53
valtih	there were local/remote masters and my branch B, in one commit. Then, masters have developed separately from B. Local B was rebased on top of masters. Now, there is a diverged remote B in the middle of history line. How can I keep it with local B, without the need to merge all the time?	17:56
cbreak	you're not making sense	17:57
	if you have two divergent lines of history, it should be kind of obvious that they will stay divergent	17:58
	until you merge	17:58
	or rewrite one of the histories to not be divergent (with rebase for example)	17:58
	← dotnull left	17:58
valtih	I rebased local B on master. I do not want to repeat the process with remote B. I just want to push local B to synchronize with remote B	17:59
	← cehteh left	17:59
	← hh__ left	17:59
	← nacho_ left	17:59
cbreak	you can kill the remote history if you want	18:00
	by force pushing something else	18:00
valtih	good idea	18:00
	→ cehteh joined	18:00
valtih	but, what is the proper way to keep remote and local in sync?	18:01
	→ nacho joined	18:01
cbreak	there's no way	18:01
	either merge them	18:01
	nacho → Guest24222	18:02
cbreak	or rebase onto new remote history	18:02
	a git branch can only hold exactly one commit	18:02
	if you want more than that, use more branches	18:02
	→ adamm joined	18:03
	→ AlexP joined	18:04
	→ gianlucadv joined	18:05
	→ maletor joined	18:05
valtih	i need to keep a separate branch, not used by everybody. It must be on top of the master. Because master is also developed, I need constantly rebase my custom branch on top of master. Now, it turns out that I also need to constantly rebase its remote shadow. Right?	18:06
	→ warez joined	18:06
cbreak	no	18:06
	you can just force push to kill the old version	18:06
	offby1 nods gravely	18:06
cbreak	of course, that would delete any changes the remote branch had	18:06
offby1	it's sorta a pain but I've done it	18:06
cbreak	I do it all the time with feature branches	18:07
offby1	valtih: you could also regularly merge form master into your branch, but then your branch will be 80% merge commits :-\|	18:07
cbreak	but since no one else is supposed to commit to them, the history rewriting is ok	18:07
offby1	that's what I do with a fork of emacs -- my branch contains some of my work that hasn't been accepted upstream	18:07
	and I like having it on all my boxen so I have lots of copies of it	18:07
cbreak	I'd use rebase	18:08
valtih	thanks	18:08
	→ tcurdt joined	18:09
	← maletor left	18:09
rath	but use "git push -f [remote] [branch]" rather than "git push -f", the last one effect other branches too	18:10
	→ macmartine joined	18:10
	← warez left	18:12
	→ warez joined	18:12
cbreak	rath: man git-config	18:12
gitinfo	rath: the git-config manpage is available at http://jk.gs/git-config.html	18:12
cbreak	I can recommend push.default to upstream	18:13
tcurdt	I would like to get a git log with all the diffs for a single file	18:13
	I was playing with --full-diff on git-log but that does not seem to be it	18:13
	pointers?	18:13
cbreak	tcurdt: man git log -p --follow -- file	18:13
gitinfo	tcurdt: the git-log manpage is available at http://jk.gs/git-log.html	18:13
tcurdt	cbreak: thx	18:13
	tizzo-afk → tizzo	18:14
	→ jaisoares joined	18:14
	→ dvdrtrgn joined	18:15
rath	cbreak: thx for the hint	18:15
	← psakrii left	18:16
	→ psakrii joined	18:16
	← Karmaon left	18:17
	→ macrover joined	18:17
	← TomyLobo left	18:19
	→ TomyLobo2 joined	18:19
	TomyLobo2 → TomyLobo	18:19
rath	<cbreak> I can recommend push.default to upstream << this should be the default i think	18:19
cbreak	matching makes sense	18:19
	in many situations	18:20
	(when using git similar to a centralized scm)	18:20
cmn	but it's not what most people expect it to do	18:20
cbreak	people who read the docs expect it :D	18:20
cmn	exactly	18:20
rath	it make sense for maintainers ;)	18:21
cmn	there's a proposal for changing the default to upstream, but nobody has really stepped up to make the patches ready for inclusion	18:21
	for maintainers?	18:21
	of what?	18:21
rath	of the project	18:22
cmn	how?	18:23
	← jaisoares left	18:24
cmn	you're still a developer with several branches	18:24
	and you still don't want your private experiments to end up in the blessed repo	18:24
	running push with matching could push up random commits you forgot were there	18:25
	→ pdtpatrick_ joined	18:25
	→ joshteam joined	18:26
	← robotmay left	18:26
rath	if i have a branch i'm working on, it's my branch but it's pushed to a server, i rebase something, and now i want to push it and i'm stupid enough to use "git push -f"	18:28
cmn	then what? what's that got to do with being the maintainer?	18:30
	← sandbags left	18:30
	← munichlinux left	18:31
shruggar	I don't know why I don't trust rebase -p maybe it's because git rebase -i -p doesn't show a graph	18:31
rath	the default is actually to bring the remote in the exactly state of your local	18:31
	and that's the "maintainer-case" i think, but it's just my opinion	18:32
	→ jdav_gone joined	18:32
	→ _pingu joined	18:32
AlexMax	Is there a reason why hg-fast-export.sh doesn't seem to recognize my hg repository? Or is there something I should be using instead?	18:33
	→ Telugodu joined	18:33
	→ eean joined	18:34
cmn	how does asking in #git help with your hg tool?	18:35
	← RaYmAn left	18:35
bremner	it's actually a git-hg bridge	18:36
cbreak	fast-export is a companion to git-fast-import I think	18:36
	cmn: man git-fast-import	18:36
gitinfo	cmn: the git-fast-import manpage is available at http://jk.gs/git-fast-import.html	18:36
	← Telugodu left	18:36
cmn	rath: that depends on how you work; if you have a repo solely for maintiner stuff then yes; otherwise no	18:36
	→ Telugodu joined	18:36
	← mlukashov left	18:37
	← Goplat left	18:37
	← _pingu left	18:37
	← Telugodu left	18:37
	← dw left	18:37
	→ svalovic joined	18:37
cmn	I know what the fast-import format is; does hg-fast-export.sh actually depend on git?	18:37
AlexMax	cmn: It imports into git	18:40
	← faber left	18:40
	← ehsan left	18:41
AlexMax	presumably you're supposed to create a git repo, cd into it, then run the script while pointing to the old hg repo	18:41
rath	the current default is the case of a repo solely for maintainer	18:41
	→ jaisoares joined	18:41
rath	as a developer who working on one branch, why do i want to force a push on all branches i've tracked by default?	18:42
cmn	probably because you're meant to have your own repo as origin	18:43
rath	anyway	18:43
	zz_XaV`S → XaV`S	18:43
rath	in my case it's a central repo for many more developers	18:43
Mikachu	you can change what git push does with no options	18:43
	← jaisoares left	18:43
	← nixmaniack left	18:44
	← RobertLaptop left	18:44
	→ RobertLaptop joined	18:44
rath	and in the end i was wondered about this default, i personally havn't expect it	18:45
	→ d2dchat joined	18:45
cmn	then you're not using the method that was expected when it was first set	18:45
	← thejoecarroll left	18:45
	← LeMike left	18:46
	→ notola joined	18:46
rath	i would change the default for that, but anyway	18:48
	→ hyperair joined	18:48
	→ jaisoares joined	18:48
AlexMax	Is there a way to create extra commits using git filter-branch?	18:48
	→ pantsman joined	18:48
	← pantsman left	18:48
	→ pantsman joined	18:48
AlexMax	i.e. splice a commit in between commits A and B?	18:49
	I'd _really_ prefer not to have to rebase the commit, as it's proven to be a huge headache.	18:49
cbreak	AlexMax: use git rebase -i	18:49
	it's much easier than messing around with filter-branch	18:50
cmn	it sounds like the perfect way to use rebase	18:50
AlexMax	cbreak: I have found the opposite	18:50
cbreak	don't blame the tools :)	18:50
AlexMax	Especially because rebase doesn't care about things you've branched from.	18:50
cbreak	rebase doesn't care about branches anyway normally	18:50
Mikachu	everything is possible with filter-branch	18:50
AlexMax	Right, and git filter-branch does	18:51
	← iocor left	18:51
cbreak	not much more	18:51
	filter branch doesn't care about history	18:51
	it just preserves topology	18:51
AlexMax	Right well.	18:51
cbreak	rebase cares about history, so if you add a new file in the past with rebase, it will be there in the present	18:51
	→ joshsmith joined	18:52
cbreak	if you add a new file in the past with filter-branch, it will vanish in the next commit	18:52
	← mishok13 left	18:52
cbreak	that's why changing history interactively is best done with filter-branch	18:52
	I mean rebase...	18:52
	filter-branch is best if you want to use a script	18:52
AlexMax	Right	18:52
	← AlexP left	18:52
AlexMax	I've been doing filter-branches	18:53
	And I'm happy with the results	18:53
	But now I want to add an extra commit in some circumstances.	18:53
	→ disappearedng joined	18:53
	← disappearedng left	18:53
	→ disappearedng joined	18:53
cbreak	rebase -i is the way. I read it can preserve merges too, somehow	18:54
	← arif-ali__ left	18:54
cbreak	but don't ask me how it does that.	18:54
AlexMax	I think you have to manually rebase --onto afterwards	18:54
	← notola left	18:54
cbreak	that also kills branches (without -p)	18:55
	→ AlexP joined	18:56
Mikachu	rebase -p is probably a bit dangerous	18:56
AlexMax	I'm still working on my private git repo that is trying to go open source	18:56
Mikachu	how about using two grafts to splice in your extra commit?	18:56
cbreak	that could work	18:56
	or replace	18:57
AlexMax	cbreak: It's adjusting the copyright date on a file.	18:57
	or a bunch of files, actually	18:57
cbreak	copyright date?	18:57
AlexMax	Right	18:57
cbreak	does that matter?	18:57
	you think you care about it 70 years after you die? :)	18:57
AlexMax	Well, I've successfully managed to add a license to every commit in the past.	18:57
Mikachu	note that if you make a change in a commit while filtering, you need to perform the change in every commit after that as well	18:57
	since commits are just snapshots of the whole tree	18:58
	if you did that, you probably noted that already :)	18:58
AlexMax	However, I went the extra mile and made sure the copyright dates were logical	18:58
	← psoo left	18:58
AlexMax	i.e if you look at a commit made in 2010, it only says copyright X-2010	18:58
cbreak	that makes no sense	18:58
	→ bindaasomatic joined	18:58
AlexMax	cbreak: How does that not make any sense?	18:59
	If you make a commit here in the year 2012	18:59
cbreak	it should be copyright 2010-2100	18:59
	if you plan to die in 2030	18:59
cmn	cbreak: that's not what the years are for	18:59
Mikachu	that's not how copyright notices work, cbreak :)	18:59
AlexMax	you put in Copyright (C) XXXX-2012	18:59
	And then bump it up by one with every year.	18:59
	So I'm just making it logical	18:59
cbreak	what's that good for?	18:59
	I mean, the notices	19:00
AlexMax	Seeing a commit made in 2011 have a Copyright (C) XXXX-2012 would be very confusing	19:00
cmn	AlexMax: if you care about the years being right, you might want to know that (c) isn't actually recognised as a copyright symbol	19:00
Mikachu	for noticing people of when the copyright is from	19:00
cbreak	they could use git log	19:00
Mikachu	in many countries the notice isn't required at all	19:00
AlexMax	The question I'm being asked is "I noticed that when the year changed, the year change was put in the same commit that this other change was done"	19:00
cmn	most, in fact	19:00
AlexMax	So now I'm asking if it's possible to just stick the year change in its own commit	19:01
	retroactively	19:01
cbreak	probably is	19:01
cmn	AlexMax: you'd have to stick that change in every single commit for year X	19:01
	← na3r left	19:01
shruggar	at the end of the day, when a copyright dispute actually occurs, people are going to look for evidence of publication history. "git log" is better evidence than some arbitrary date in a text file	19:01
AlexMax	Well yeah, that's why filter-branch exists	19:01
shruggar	and of course, neither would hold up in court.	19:01
AlexMax	To make the same change in every single commit	19:02
	and in fact I've successfully done it	19:02
Mikachu	AlexMax: try something like this, check out the last commit of 2011, make a new commit changing everything to 2012 and tag this commit with something or write down the sha1, then add a graft for the first 2012 commit to have your new commit as its parent	19:02
cbreak	shruggar: considering the fact that he's rewriting his whole history	19:02
	→ timonia joined	19:02
cbreak	I doubt any of thta is useful evidence	19:02
shruggar	and of course, copyright law makes so little sense these days that it can be safely ignored ;)	19:02
	← krz left	19:02
cbreak	what Mikachu said also works with man git-replace	19:02
gitinfo	the git-replace manpage is available at http://jk.gs/git-replace.html	19:02
Mikachu	the date probably doesn't matter since they extend the duration of copyright by the rate of one year / year	19:02
AlexMax	Mikachu: Well, what would be super-easy for me	19:02
	would be to simply put in 2007-2012	19:03
	in every single commit	19:03
	going back to 2007	19:03
	← joshteam left	19:03
AlexMax	and then just bump the date up in the future	19:03
cbreak	or you could just only write the year the file was created	19:03
	which is what I do	19:03
cmn	you want to put a 2007-2012 copyright notice in a file change in 2007?	19:03
	→ ehsan joined	19:03
cbreak	because the .h file template of XCode does it for me	19:03
AlexMax	cmn: The revision wasn't public in 2007	19:04
	So yes	19:04
cmn	so?	19:04
	→ metcalfc_ joined	19:04
	← metcalfc_ left	19:04
AlexMax	The source tree was _released to the public_ in 2012	19:04
	→ metcalfc joined	19:04
cmn	that doesn't matter	19:04
	distribution doesn't extend the copyright	19:05
SethRobertson	I have a program which uses git (and CVS) to compute what years a file was created/modified in. Some management want companyfounded-today, some want filecreated-today, some want filecreated-filelastmodified, some want each,year,file,was,modified.	19:05
	→ acl_ joined	19:05
cmn	and it doesn't make sense to claim a 2012 copyright on a file in 2007	19:06
AlexMax	Mikachu: If I graft that onto a repository that already has that 'year change' modifiecation already done via git-filter	19:06
	← macrover left	19:06
SethRobertson	It does it all, in a very hacky and stupid way. It also tries to compute (c) vs ⓣ vs latin-1 ⓣ	19:06
cmn	I see a circled 't'	19:06
	how odd	19:06
	← ehsan left	19:07
AlexMax	Mikachu: So if I run a git-filter that uses the commit date to write the copyright dates to the appropriate files	19:07
SethRobertson	But yes, even with all of that stupidity, it is silly to back in history and retroactively claim the copyright in 2012 when the virtual time is 2007	19:07
AlexMax	then graft in a commit that actually makes those changes	19:07
	← queequeg1_ left	19:07
AlexMax	that will work?	19:07
	without seeing 20 million files change in the non-date-change commit?	19:07
	→ airborn joined	19:07
	→ queequeg1 joined	19:08
	→ jceb_ joined	19:08
	← dvdrtrgn left	19:08
	← joshsmith left	19:08
SethRobertson	It doesn't make sense. The filter-branch forces that the file was created with the future date	19:08
	→ dvdrtrgn joined	19:08
	→ joshsmith joined	19:08
Mikachu	AlexMax: yeah	19:08
	← disappearedng left	19:08
	← eikke left	19:09
	→ LeMike joined	19:09
Mikachu	AlexMax: since then the diff from that commit to the next will not contain the copyright date change anymore	19:09
rath	AlexMax: choose your license properly, so you don't need to update all of your files every year with the correct (c) year. ... the idea to do it in this way is just crazy	19:09
	and make no sense at all for me	19:10
	← jceb left	19:10
	→ Goplat joined	19:10
Mikachu	eh?	19:11
cmn	how does a license change the need to update copyrights?	19:12
	→ milito joined	19:12
cmn	...notices	19:12
	← milito left	19:12
rath	it doesn't	19:13
	gitinfo set mode: +v	19:13
dvdrtrgn	isn't the term of copyright like 80 years?	19:14
rath	why do you have to update your (c) comments every year?	19:15
AlexMax	Mikachu: So a revision in git is represented by the state of the tree?	19:15
	And not the diff between revisions?	19:15
	ChanServ set mode: +v	19:15
	← Destos left	19:15
	tizzo → tizzo-afk	19:16
Mikachu	AlexMax: exactly	19:16
	any diff you see is generated when you see it	19:16
AlexMax	ah	19:16
	What I could do	19:16
	is graft in the year change commits first.	19:16
Mikachu	a 'rebase' operation works in units of diffs though	19:16
AlexMax	and then let my existing filter-branch script run	19:16
Mikachu	but filter-commit doesn't	19:16
	tizzo-afk → tizzo	19:17
AlexMax	it will change the year on the year change commit	19:17
Mikachu	that would work too	19:17
	← steffo left	19:17
AlexMax	Thanks for your useful information Mikachu	19:17
Mikachu	there's a special option for commit that lets you commit something with no changes	19:17
AlexMax	Also perhaps one at an arbitrary date?	19:18
Mikachu	the filter-branch operation should also make the graft permanent	19:18
	there are ways to edit those too, yes	19:18
	should be in the manpage	19:18
AlexMax	it should	19:18
	hg commit allows a --date option	19:18
	git ought to as well	19:18
	→ aspotashev joined	19:19
	← matsebc left	19:19
AlexMax	Ah	19:20
	check out the commit that goes before every year change	19:20
	← rath left	19:20
AlexMax	do an empty commit with the next year	19:20
	as a custom date	19:20
	then let the filter-branch do the rest	19:20
	→ matsebc joined	19:20
	← dvdrtrgn left	19:21
AlexMax	after grafting the commit inbetween them	19:21
	→ dvdrtrgn joined	19:21
	← tgunr left	19:21
	→ techhelp joined	19:22
	→ tgunr joined	19:22
	← candybar left	19:23
	→ maletor joined	19:26
	← d2dchat left	19:28
	→ msponge joined	19:28
	→ dharrigan joined	19:30
	← dharrigan left	19:30
	→ dharrigan joined	19:30
	→ OOPMan joined	19:30
	← drizzd_ left	19:30
	← dvdrtrgn left	19:30
	→ nbigaouette joined	19:30
	→ dvdrtrgn joined	19:32
	← svalovic left	19:32
	← bartman left	19:33
	← psakrii left	19:33
	tizzo → tizzo-afk	19:34
AlexMax	damnit	19:34
	← t0rc left	19:34
pdtpatrick_	Question -- anyone used araxis or kaleidoscope ? http://www.kaleidoscopeapp.com/	19:35
	http://www.araxis.com/merge_mac/overview1.html	19:35
	→ VladimirZ joined	19:36
VladimirZ	after 'rebase' i again lost all commits in branch	19:36
	is there any good workaround for that?	19:36
SethRobertson	VladimirZ: See !fixup for recovery. Afterwards, come back for proper rebase usable	19:37
gitinfo	VladimirZ: So you made a bad commit and want to remove/fix it? Look at https://gist.github.com/1612395 for full instructions. Hints are: (1) NOT PUSHED/PUBLISHED: `git rebase -i $COMMIT^` or perhaps `git commit --amend` (or `git reset HEAD^`). (2) OTHERWISE, `git revert $COMMIT` to make a reverse commit. (3) If you have pushed and MUST remove it, use rebase or filter-branch and type !rewriting_public_history in IRC.	19:37
	← XaV`S left	19:37
	→ zz_XaV`S joined	19:38
	← cbreak left	19:39
	← techhelp left	19:39
VladimirZ	SethRobertson: I made a rebase not commit	19:40
SethRobertson	Look at the first link	19:40
VladimirZ	can not find it in reflog :/	19:40
	→ tbrock joined	19:40
SethRobertson	I assure you it is there	19:41
	← crisp left	19:41
	← FernandoBasso left	19:41
VladimirZ	found it	19:41
	:)	19:41
	← zz_XaV`S left	19:41
	→ bartman joined	19:41
VladimirZ	two days of work saved xD	19:42
	→ frode15243 joined	19:43
AlexMax	Hrm	19:44
	with the grafts in place, glog looks funny	19:44
SethRobertson	Not entirely surprising. If you remember, everyone recommended against it	19:44
AlexMax	however the log looks fine and as-expected	19:45
	← Chillance left	19:45
AlexMax	as long as I'm tracking single commits	19:45
	lets see what happns if i filter-branch and make the graft permanent	19:46
	probably log looks weird because it's sorting based on commiter date, not by commit date	19:47
	which i don't think --date= changes	19:47
	← jast left	19:49
	→ jast joined	19:49
	→ mgpcoe1 joined	19:51
	→ sbell joined	19:52
	← OOPMan left	19:54
	→ tomislater joined	19:54
	← erickr left	19:55
	→ fr0stbyte joined	19:57
	← fr0stbyte left	19:58
	← Sigma left	19:58
	→ nrichards joined	20:00
	→ cybersphinx joined	20:00
	→ PugMajere joined	20:01
	→ iocor joined	20:01
cmn	you can't change the committer date, but you can change the author date	20:02
	← the_cybersphinx left	20:03
	← arosen left	20:03
	← aidenhong left	20:04
	→ Chillance joined	20:06
	← angelsl left	20:06
	← tcurdt left	20:06
Mikachu	of course you can	20:07
FauxFaux	You can set the committer date with --env-filter..	20:07
Mikachu	and with git commit-tree	20:08
	→ OOPMan joined	20:08
FauxFaux	Literally the solution to all problems.	20:08
	→ Heimidal joined	20:09
	← resmo left	20:09
	← akosikeno left	20:09
	→ Destos joined	20:10
cmn	I can also set the clock on my computer back	20:10
	→ johnkpaul joined	20:10
	← LeMike left	20:10
	→ Cromulent joined	20:11
cmn	but that kinda wasn't the point; of course if you're grafting so some irrelevant text can be inserted, there's no sensible way to do this	20:11
	→ joshteam joined	20:12
	→ Sigma joined	20:12
	→ cbreak joined	20:13
	← iocor left	20:19
	→ resmo joined	20:21
	← Dave^\|\| left	20:21
	← johnkpaul left	20:22
	→ icwiener joined	20:24
	→ d2dchat joined	20:25
	→ erickr joined	20:28
	ipalaus → ipalaus\|eat	20:28
	→ iocor joined	20:29
	← JackWinter2 left	20:30
	← tgunr left	20:32
	→ tgunr__ joined	20:32
	→ cccaldas joined	20:33
	← frode15243 left	20:33
	← Sigma left	20:34
	→ JackWinter joined	20:34
	← cccaldas left	20:35
	→ cccaldas joined	20:35
	← cccaldas left	20:36
	→ cccaldas joined	20:36
	→ koo5 joined	20:36
	→ nicxvan joined	20:37
	→ vdv joined	20:37
	← tomislater left	20:39
	← rzec left	20:39
	→ _dzonder joined	20:39
	← dzonder left	20:40
	→ anj joined	20:40
	→ cyb3r3li0 joined	20:41
	← cyb3r3li0 left	20:41
	← oc80z left	20:42
	← bitkiller left	20:42
	← nbigaouette left	20:43
	→ oc80z joined	20:44
	→ cyb3r3li0g joined	20:44
	→ cesc joined	20:46
	→ LeMike joined	20:47
	← vdv left	20:48
	ipalaus\|eat → ipalaus	20:48
	tizzo-afk → tizzo	20:50
	← soulcake left	20:50
	→ beford joined	20:52
	gitinfo set mode: +v	20:52
	← gianlucadv left	20:54
	← tbrock left	20:55
	← cyb3r3li0g left	20:56
	← leavittx left	20:57
	→ cyb3r3li0 joined	20:58
	← txomon\|home left	20:58
	← bindaasomatic left	21:00
	← anj left	21:00
	← pikpik left	21:00
	→ pikpik joined	21:00
	← pikpik left	21:00
	→ pikpik joined	21:00
	→ ruizander joined	21:02
	← Goplat left	21:02
AlexMax	How do i show all commits with no parent commit?	21:03
	← mandric left	21:03
AlexMax	for some reason this repository has more than one	21:03
FauxFaux	AlexMax: They're called root commits; google for them, not unusual.	21:03
AlexMax	FauxFaux: Perfect, thanks	21:04
	← ruizander left	21:04
SethRobertson	AlexMax: You can `git log --pretty='%H %P --all \| grep -v ' '`	21:04
	Without testing, that should print all orphan commits	21:05
	closequote ' after %P	21:05
	egrep -v ' .'	21:06
	← _inc left	21:07
	→ _inc joined	21:07
	→ xil joined	21:08
AlexMax	I swear it's not my fault	21:09
	i think it's a git-svn artifact	21:09
	:)	21:09
xil	hey everyone. When I do "git status" I see it says I can "(use "git reset HEAD <file>..." to unstage)" but does that mean it will move the file(s) to the working tree, or remove them altogether?	21:10
FauxFaux	AlexMax: Why are you worried about it? It's perfectly normal. I frequently merge unrelated codebasese together.	21:10
	xil: It'll just remove your intention to commit the file this time.	21:10
xil	FauxFaux: so when I run git status again the file should be listed under "modified but unstaged"?	21:11
FauxFaux	Yes.	21:11
xil	awesome, thank you =]	21:11
	so if I did want to roll back that file to the last commit then I could reset HEAD the file, then check the file out again?	21:12
	tizzo → tizzo-afk	21:12
	← mgpcoe1 left	21:14
	← tgunr__ left	21:16
	→ vdv joined	21:18
	→ Wulf joined	21:19
Wulf	Hello	21:20
	→ dotnull joined	21:20
SethRobertson	xil: !fixup has full info	21:21
gitinfo	xil: So you made a bad commit and want to remove/fix it? Look at https://gist.github.com/1612395 for full instructions. Hints are: (1) NOT PUSHED/PUBLISHED: `git rebase -i $COMMIT^` or perhaps `git commit --amend` (or `git reset HEAD^`). (2) OTHERWISE, `git revert $COMMIT` to make a reverse commit. (3) If you have pushed and MUST remove it, use rebase or filter-branch and type !rewriting_public_history in IRC.	21:21
	← notjohn left	21:21
xil	SethRobertson: oh I'm just curious after the fact. My original question was just a quick curiosity because I see that message all the time, and the rest just sort of came out of that	21:22
	← vdv left	21:22
	← v0n left	21:23
	← nikuyoshi left	21:23
	→ nikuyoshi joined	21:24
	→ notjohn joined	21:26
	→ localheinz joined	21:27
	→ martinjlowm joined	21:28
	← khmarbaise left	21:28
	← localheinz left	21:28
	→ dv310p3r joined	21:28
	← nikuyoshi left	21:28
berndj	part	21:28
	← berndj left	21:28
	← dotnull left	21:29
	← toobluesc left	21:30
	← tewecske left	21:33
	← valtih left	21:34
milki	!rewriting_public_history	21:34
gitinfo	Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)	21:34
	→ dorkmafia joined	21:35
	→ iiu7 joined	21:37
	← phaedrix_ left	21:38
	← sebrock left	21:38
iiu7	I'm cloning and in the clone deleted a lot of branches I don't need. I want to be able to push back without the branches I deleted also gets pushed. What's the right way to deal with that?	21:38
	← Destos left	21:39
	← notjohn left	21:39
cmn	if you deleted the branches, they won't get pushed	21:39
canton7	iiu7, the only way to delete remote branches it though "git push --delete remote branchname" or "git push remote :branchname"	21:39
	*is through	21:39
Mikachu	how would they get pushed if you deleted them? :)	21:39
	→ pippo1 joined	21:40
iiu7	Mikachu :) I mean so that the "deletion" won't be repeated on the original.	21:40
Mikachu	it won't, it's the same as if someone else pushed a new branch that you don't have yet	21:41
	← detaer left	21:41
	→ vdv joined	21:41
iiu7	I see, alright, thanks!	21:41
	← bgerber left	21:42
	← martinjlowm left	21:42
	← iiu7 left	21:42
	→ zulax joined	21:42
	→ martinjlowm joined	21:43
	← nrichards left	21:44
	← LeMike left	21:50
	← pidus left	21:51
	← pippo1 left	21:54
	→ Destos joined	21:56
Wulf	How should I create a new repository which is used by several remote users? So far I found "git init --bare" and "git config receive.denyNonFastForwards true". Is there more?	21:57
	→ tigrang joined	21:58
	← iocor left	21:58
cmn	Wulf: at the very least, you need git init --bare --shared	22:00
	otherwise you'll have problems with permissions	22:00
canton7	cmn, remote users?	22:00
	→ candybar joined	22:00
cmn	Wulf: but you should take a look at !gitolite	22:01
gitinfo	Wulf: gitolite allows you to host any number of repositories with extensive access control options for any number of users, using just one system account: http://github.com/sitaramc/gitolite	22:01
	← p3rror left	22:01
cmn	canton7: ssh counts as remote, doesn't it?	22:01
canton7	ah, you're assuming different ssh users, of course. Quite right	22:01
Wulf	actually I'm trying to set up something similar to github. One ssh account for all users	22:02
	should be good enough to manage my keys in ~/.ssh/authorized_keys?	22:02
canton7	gitolite will be a lot nicer to use	22:03
	→ iocor joined	22:03
	→ bgerber joined	22:04
	← eean left	22:04
cmn	definitely gitolite	22:04
Wulf	I'll have a look at it.	22:04
	← iocor left	22:05
	→ LucasTizma joined	22:05
	← Matrixiumn left	22:05
	→ Matrixiumn joined	22:06
	→ Heisenmink joined	22:06
	← cccaldas left	22:07
	← zulax left	22:10
	→ nrichards joined	22:10
	→ Milossh joined	22:11
Milossh	if I add some files in .gitignore, remote git repo will not know about it, as in, it's client side only?	22:11
	← tatsuya_o left	22:14
	→ eean joined	22:15
	→ jeffisabelle joined	22:15
	→ techhelp joined	22:15
	→ p3rror joined	22:15
	→ disappearedng joined	22:15
	→ tatsuya_o joined	22:15
cmn	if you don't commit the file, it'll stay in your workdir only	22:17
	if want to share it, you can commit it like any other file	22:18
Milossh	you mean, I don't commit the .gitignore?	22:18
cmn	yes	22:19
Milossh	cool, thanks!	22:19
	← stringoO left	22:20
	← martinjlowm left	22:20
	→ rebe joined	22:20
	← yshh left	22:21
	→ yshh joined	22:21
	→ martinjlowm joined	22:21
	← martinjlowm left	22:22
	→ caseymcg joined	22:23
	← LucasTizma left	22:25
	→ elprado joined	22:25
	← elprado left	22:26
	← timonia left	22:28
	tizzo-afk → tizzo	22:35
	← vdv left	22:37
	→ Heisenmink_ joined	22:37
	→ kiero_ joined	22:38
	gitinfo set mode: +v	22:39
kiero_	re	22:39
	← fisted left	22:39
kiero_	i've got a question about doing a project with git. front-end as well as back-end work	22:40
	i set up a repository then i'm starting to creating html files of views and when I'm finished I want to set up a Ruby on Rails application	22:40
	← Heisenmink left	22:41
kiero_	what should I do? delete views? or set up another git repo?	22:41
	→ fwg joined	22:42
	tizzo → tizzo-afk	22:42
	← Kenextre1 left	22:43
cmn	that would depend on how the transition happens; if it's a new start, a new repo would make sense	22:43
	if it's a more gradual change, then keeping the history might be useful	22:44
kiero_	i mean that then i need to move the html files to ror views somehow	22:44
	tizzo-afk → tizzo	22:44
kiero_	and i'm confused	22:44
	→ fisted joined	22:45
kpreid	kiero_: is this something you do a little bit at a time? then keeping it in one rep lets you track/manage your progress	22:46
	you can always move existing stuff to a subdirectory	22:47
kiero_	what do you mean: is this something you do a little bit at a time?	22:48
	?	22:48
cmn	the changes	22:48
	whether you gradually move from one type of site to the other	22:49
	tizzo → tizzo-afk	22:49
j416	kiero_: put it all in a repo from the start, it will be fine.	22:49
	kiero_: you can move files as you like, later.	22:49
	→ infogulch joined	22:52
kiero_	yes it's gradually I think	22:52
	so, creating html files	22:52
	moving it to temporary folder	22:52
	then	22:53
	set up rails app	22:53
	moving html files to templates	22:53
	and delete temporary files with html files right?	22:53
cmn	so keep it in one repo	22:53
	← jaisoares left	22:54
	← webrover left	22:54
kiero_	ok, thx	22:55
	← joshteam left	22:56
	→ ehsan joined	22:57
	← gusnan left	22:58
	← caseymcg left	23:01
	→ t0rc joined	23:02
	→ nikuyoshi joined	23:03
	← rebe left	23:04
	→ Aleric joined	23:05
	→ stringoO joined	23:06
Aleric	hi, how can I create a new branch and fill it with branch master of some given repository (http://..../foo.git) ?	23:06
	or, I already have that repository... how can I list all repositories that I have (that would be updated when I do a 'git fetch -a')?	23:07
kpreid	those are called remotes	23:07
	'git remote' will list them, -v with urls	23:07
Aleric	haven't used git for a year... I forgot everything :(	23:08
	thxs	23:08
	→ Guest59942 joined	23:08
kpreid	if you already have it then git branch or git checkout can be used to create a local branch with the remote tracking branch as upstream	23:08
	← infogulch left	23:10
Aleric	My current working tree is a mess... I want to find out what I did (a year ago) by checking out the upstream master and then making a diff between my tree and that remote around the time that I abandoned it.	23:10
cmn	you don't need to checkout anything for a diff	23:10
	just use git diff otherbranch	23:10
Aleric	I am trying to create a new branch with just that remote master (origin/master)	23:11
	but I want to look at it :/	23:11
	need to find the right date and stuff	23:11
cmn	ok, then a log might help you more than a checkout]	23:11
SethRobertson	If you use `gitk --all --date-order` you can browse commits graphically and diff A-B	23:11
	← nikuyoshi left	23:12
	← _iron left	23:12
	← candybar left	23:13
	← warez left	23:14
	→ iocor joined	23:14
	→ candybar joined	23:14
	→ jaisoares joined	23:18
	← techhelp left	23:19
	tizzo-afk → tizzo	23:20
	← maletor left	23:21
	← nrichards left	23:21
	← jceb_ left	23:21
	← Raging_Hog left	23:21
	tizzo → tizzo-afk	23:22
	← jeffisabelle left	23:23
	← Destos left	23:24
	→ BiggFREE joined	23:24
	→ warez joined	23:24
	→ rickmasta joined	23:25
rickmasta	Hey guys, I just received "fatal: loose object 293085e6ecfa1ce96a207de9cee778121b62af48 (stored in .git/objects/29/3085e6ecfa1ce96a207de9cee778121b62af48) is corrupt", what does this mean?	23:27
	← JackWinter left	23:27
	→ innoying joined	23:27
	← warez left	23:27
cmn	what is says; that object is corrupt	23:27
	← jaisoares left	23:27
cmn	either a problem with the hardware, or the fs, probably	23:28
rickmasta	cmn: What can I do to fix this?	23:28
	← yshh left	23:29
cmn	get the right object from another repo or backup	23:29
	s/right/non-corrupt/	23:29
rickmasta	Did you just change your text?	23:30
	Sorry, but I'm not sure what that means.	23:30
cmn	it's a regex, replace "right" with "non-corrupt"	23:31
	→ xerora joined	23:32
	← acl_ left	23:33
	→ Davey_ joined	23:33
	→ JackWinter joined	23:34
	← d0k left	23:35
	→ macrover joined	23:36
	→ Chibby joined	23:36
	← frerich2 left	23:37
	→ caseymcg joined	23:38
	→ zibri joined	23:39
	→ JB3181 joined	23:42
	→ fayimora joined	23:43
	→ robotmay joined	23:43
	← [M]ax left	23:44
	← nicxvan left	23:45
	→ ISF joined	23:46
	→ frerich joined	23:46
	← frerich left	23:46
	→ frerich joined	23:46
	← Guest59942 left	23:46
	→ [M]ax joined	23:48
	← JackWinter left	23:48
	→ JackWinter joined	23:49
	→ blakefrost joined	23:49
	→ nikuyoshi joined	23:50
	← JackWinter left	23:51
	← Cromulent left	23:51
	→ soulcake joined	23:51
	→ vervic joined	23:53
	→ JackWinter joined	23:53
	→ keller joined	23:55
	← jjasinski left	23:56
rickmasta	cmn: I cannot find this specific object	23:56
	I've tried cloning my other repos	23:56
	Well, I only tried an older version of this same repo.	23:57
	→ jjasinski joined	23:57
cmn	if it's an older version it might not have it	23:57
	you should run 'git fsck' on the server to see if it has other problems	23:57
	→ fayimora_ joined	23:58
rickmasta	Should I try another repo?	23:58
	← diegoviola left	23:59

Logs Search ←Prev date Next date→

Channels Documentation