IRCloggy #git 2010-08-20

Provider of IRC logs since 2005.

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

	2010-08-20
	← ajpiano left	00:01
	→ jmcantrell joined	00:01
	← Nemurenai left	00:02
	← yngress left	00:03
	← Nonpython left	00:04
	← Hanumaan left	00:05
elfMobile	So I'm doing development on a branch called featurex. I'm not ready to push yet. Someone checks stuff into master. I want to pull that new stuff before I push my new feature. Should I do a pull from master or from featurex?	00:06
frogonwheels	elfMobile: featurex is on the server as well?	00:07
JdGord	The date at the top of the first patch is Sep 2001! And it has what looks like a git hash?	00:07
frogonwheels	elfMobile: if not.. I would probably: git checkout master ; git pull ; git checkout featurex; git rebase master	00:10
	.. I think there might be a shorter way.. but that would work	00:10
	→ splnet joined	00:13
	← DrNick left	00:13
	→ DrNick joined	00:14
engla	JdGord: what can help you are the hashes on the lines starting with "index ", those record the blob ids that the patches apply to	00:15
	JdGord: however, a blob does not uniquely identify a commit (at all), unfortunately	00:16
elfMobile	frogonwheels, sorry... was afk. No featurex is not on the server.	00:16
splnet	how do you do a diff between 2 commits on a single file?	00:16
basicxman	splnet: git diff <rev>:file <otherrev>	00:17
elfMobile	frogonwheels, you say "git checkout master ; git pull ; git checkout featurex ; git rebase master".... but step 1 was failing for me because I had uncommited stuff it wouldn't let me checkout master	00:17
basicxman	splnet: <rev>:file <otherrev>:file sorry	00:18
frogonwheels	ok commit it or stash it then.	00:18
	elfMobile: ^^	00:18
	→ apol_ joined	00:18
frogonwheels	elfMobile: git stash; git checkout master ; git pull ; git checkout featurex ; git rebase master ; git stash pop	00:18
elfMobile	frogonwheels, never heard of stash... is it like a commit?	00:20
splnet	basicxman: hmm fatal: ambiguous argument '5347bd4cda2b6afc18f8acab48e52131f35ed13c:WifiStateTracker.java': unknown revision or path not in the working tree	00:20
	← apol left	00:21
frogonwheels	elfMobile: it puts your index away for future retrieval. look at man git-stash	00:21
Gitbot	elfMobile: the git-stash manpage can be found at http://git.or.cz/man/git-stash	00:21
	← quotemstr1 left	00:21
	← enherit left	00:22
frogonwheels	elfMobile: this is pretty much a text-box use of it- useful when you don't want to commit to anything but need a clean working tree	00:22
	s/box/book/	00:22
	← GoGi left	00:22
JdGord	engla: hmm, nuts	00:22
	→ quotemstr1 joined	00:22
frogonwheels	bb in 30	00:22
	→ mikeric joined	00:23
	← sijis left	00:25
	→ fod joined	00:27
	← JdGord left	00:27
	← aah left	00:27
	→ akahn joined	00:27
	→ CoinOpeBoy joined	00:28
	→ richardbronosky joined	00:28
basicxman	splnet: Can you paste your full command to http://pastie.org? Basically it's saying that revision doesn't exist, or that file doesn't exist. Make sure to specify the full file path.	00:28
	← Aaaarg left	00:29
	→ Aaaarg joined	00:29
	← Aaaarg left	00:29
splnet	basicxman: http://pastie.org/1103271	00:30
	I didn't use the full path. './' isn't sufficient?	00:31
	← jrmuizel left	00:31
splnet	hmm.. $PWD/file didn't work either	00:32
basicxman	You don't need ./ or $PWD	00:32
	That's probably your problem.	00:32
	← akahn left	00:32
	→ posciak joined	00:32
basicxman	You need to specify the path relative to the repo root	00:32
splnet	ah ok	00:32
	← zokier left	00:33
	← engla left	00:35
	← kalusn left	00:35
	← ExtraSpice left	00:35
splnet	basicxman: ok got it working thanks	00:35
	← dget left	00:36
	→ dget joined	00:36
	← zackattack left	00:39
	→ JdGordon\| joined	00:40
JdGordon\|	ah ha :) how do I find out the iniital branch point?	00:40
	← dget left	00:40
	← sh1mmer left	00:41
	→ Radar joined	00:43
	→ MrGando joined	00:44
MrGando	Hello guys, could anyone suggest me the best way to embed doxygen to a git repository ?	00:44
	← amerine left	00:45
	→ k0nichiwa joined	00:45
	← Yuuhi left	00:45
basicxman	splnet: np	00:46
	→ wezyde joined	00:47
	← jmcantrell left	00:48
	← SimonP86 left	00:48
	→ juanjoc joined	00:49
	→ sh1mmer joined	00:51
	← dr_win left	00:53
	→ scarabx joined	00:53
	← pigeon left	00:53
	→ pigeon joined	00:53
	← sh1mmer left	00:56
	→ Aaaarg joined	00:57
	← Aaaarg left	00:57
	← JdGordon\| left	00:58
elfMobile	can I do micro commits in my local repo that nobody copies from and then combine all of them into a single commit for when I push?	00:58
theoros_	you can edit, reword, reorder, combine, split commits however you like as long as you haven't pushed (or similar'ed) them	00:59
	go wild with rebase -i	00:59
	apol_ → apol	00:59
	→ guilhermeblanco joined	01:01
	→ Aaaarg joined	01:02
	← Shurakai left	01:03
	→ Yuffster joined	01:04
	→ ajpiano joined	01:04
	← markbao_ left	01:06
hachi	is there a fast way to find whether one commit compared to another is a child, grandchild, etc.	01:06
	or a parent, grandparent, etc.	01:07
	or neither	01:07
	← sivy left	01:07
	← ajpiano left	01:07
hachi	I'm thinking you can ask about the merge base between the two, and whichever one happens to be more 'parentish' will be returned as the merge base	01:07
	← posciak left	01:08
theoros_	hachi: you could also git rev-list A \| grep $(git rev-parse B)	01:14
	if you're feeling particularly mediocre.	01:14
	← lucasvo_2 left	01:14
	→ shiba_yu36 joined	01:15
	← g0bl1n left	01:15
	← wezyde left	01:16
	← friskd left	01:16
	← pdelgallego left	01:19
	→ sivy joined	01:20
	→ lucasvo_2 joined	01:20
	← mbroeker left	01:21
	jexmex → jexmex_	01:23
elfMobile	in general, should you ever work on your master branch? What if you do work on your master branch and the project doesn't accept your changes and advances without you..... then your master is out of sync with remote/master?	01:25
	← nothingHappens left	01:26
	← theoros_ left	01:26
	← jexmex_ left	01:27
SeveredCross	elfMobile: Git is probably smart enough to merge later changes.	01:27
elfMobile	i'm not saying its the end of the world... but in general do you work directly on master?	01:28
hachi	depends on what I'm doing, but yes, I do	01:28
MrGando	It's always recommended to work outside master and then merge	01:29
	→ Vampire0 joined	01:29
	→ jexmex_ joined	01:30
frogonwheels	elfMobile: when you pull, you can use -r to have your working changes rebased onto the origin/master	01:30
SeveredCross	I generally old do major work in a branch.	01:31
elfMobile	right... but then what would you do after you "git checkout master ; git merge some_feature" and tell someone to pull from you and they never do?	01:31
SeveredCross	You can do pulls from a remote branch.	01:31
	So you don't have to merge.	01:31
	← ObsidianX left	01:31
frogonwheels	elfMobile: or you can merge (which makes the tree more complicated, but isn't a bad choice either)	01:31
	← Vampire0_ left	01:31
frogonwheels	elfMobile: .. but for larger changes, I tend to make a branch... that way if I need to do a fire-fighting fix, I can do it directly to the real master	01:32
SeveredCross	Pretty much exactly what I do.	01:32
elfMobile	so when you petition someone to pull your changes would you ask them to pull from a branch and not your master?	01:32
	← shiba_yu36 left	01:32
	← jexmex_ left	01:32
	→ ajpiano joined	01:33
elfMobile	maybe I need to get back to reading	01:33
	→ shiba_yu36 joined	01:34
	→ mithro joined	01:34
	← CoinOpeBoy left	01:35
	← krawek left	01:40
	→ krawek joined	01:40
	→ yoh joined	01:40
yoh	quick question: config's merge.log=true is great but it limits number of listed commits -- is there way to raise the limit or remove it? i.e. how many commits to include in the merge commit message?	01:41
	← hebz0rl left	01:44
	← bauruine left	01:46
	→ bauruine joined	01:46
	→ rmbwebs joined	01:46
	← mikeric left	01:47
	→ beatak joined	01:47
	→ markbao_ joined	01:48
	← beatak left	01:48
	← nuoHep left	01:48
rmbwebs	quick question. was working on master branch, I had some local changes and wanted to commit them on a new branch (leaving master alone). I created the branch but forgot to check it out and wound up committing the changes to master. So now master and the new branch are basically reversed from what I want. . . how to correct?	01:49
	I figure I can checkout the new branch and merge master. that will make the new branch correct. How to make "master" point to the previous commit?	01:49
	→ nothingHappens joined	01:50
Chaser	git --reset hard ?	01:50
yoh	looking at the code of do_fmt_merge_msg -- there seems to be hardcoded limit of 20	01:50
	← Vampire0 left	01:51
	→ friskd joined	01:52
	← friskd left	01:52
rmbwebs	Chaser: was thart directed at me?	01:53
Chaser	rmbwebs, yes	01:54
frogonwheels	rmbwebs: git checkout master ; git reset --hard HEAD~1	01:55
	HEAD~1 == HEAD^	01:56
	→ eggcaker joined	01:56
	← eggcaker left	01:56
	→ hebz0rl joined	01:56
frogonwheels	rmbwebs: get used to doing git checkout -b newbranch (rather than git branch newbranch; git checkout newbranch)	01:56
	→ corni_ joined	01:57
rmbwebs	that worked great, thanks guys. If I hadn't merged the master into the new branch first, would I have lost the changes or would that commit still be floating around?	01:57
	frogonwheels: good advice. I usually do that	01:57
frogonwheels	floating round	01:57
	← juanjoc left	01:58
frogonwheels	rmbwebs: it would be in the reflog, and even git gc won't touch it until it is a certain age	01:58
rmbwebs	frogonwheels: is suspected. thanks. I'm starting to learn that branch names are really like pointers to various commits and reset is what you use to move the pointers around	01:59
frogonwheels	rmbwebs: exactly.. well branch neams are floating pointers - when HEAD refers to a branch (rather than a commit - which is called a detached head), then the branch advances with the commit	02:00
	→ stringoO joined	02:00
	← corni left	02:00
	→ wezyde joined	02:01
frogonwheels	rmbwebs: a tag is a pointer to a commit and is usually more permanent.	02:01
	← armenb_ left	02:01
rmbwebs	frogonwheels: right. it doesn't move with the commit.	02:02
frogonwheels	exactly.	02:02
	← canavar left	02:03
	→ ProLoser\|Work_ joined	02:03
frogonwheels	rmbwebs: cat .git/HEAD or ls -l ./git/HEAD	02:04
	err .git/HEAD	02:04
	→ gnufied joined	02:05
	← kylehayes left	02:06
rmbwebs	ref: refs/heads/newbranchname	02:07
	← SvenDowideit left	02:07
	← ProLoser\|Work left	02:07
frogonwheels	rmbwebs: get the idea?	02:08
	→ SvenDowideit joined	02:08
	→ drubiess joined	02:09
rmbwebs	yeah I think so. if I checkout some random commit, the contents of .git/HEAD is just the commit ID	02:09
	← ProLoser\|Work_ left	02:09
	← drubiess left	02:09
	→ frakturfreak_ joined	02:09
rmbwebs	so HEAD just refers to whatever you have checked out	02:09
frogonwheels	rmbwebs: that's precicely what HEAD is	02:11
	← sagsousuke left	02:11
	← aziz_ left	02:12
	← Radar left	02:13
	→ LongBeach joined	02:13
	→ Radar joined	02:14
	← frakturfreak left	02:14
	← wezyde left	02:18
	← kukks left	02:19
	← yoh left	02:21
anjen	Hi, durnig development over the past few months, I've been asked to add things to public (flash stuff specifically) that will bne migratnig to a more "proper treatment using S3 bukets to hold the content. Unfortunately, my code base is getting unweildingly large, and the git direcotry of my rails codebase represents a good 2/3 of that bulk. While I realize that the whole point of git is vc, I'm wondering if there's a *sound method	02:23
	for prasing out some of the crud that has accumulated due to the temproary storgae in public of these flash files...From perusing the git directories, it looks like they're the big ugly.	02:23
	*parsing, not prasing	02:23
	sorry for the typos. I really should have spell checked first	02:24
	← cytrinox left	02:24
	→ cytrinox joined	02:26
	→ zackattack joined	02:26
	← ajpiano left	02:27
	→ amerine joined	02:27
	← sftp left	02:27
	← spaceonline left	02:28
	→ akahn joined	02:28
	← txdv left	02:28
	← richardbronosky left	02:30
	← Blackguard left	02:30
	→ ignatp joined	02:31
	frakturfreak_ → frakturfreak	02:31
	← akahn left	02:32
	← Radar left	02:32
	← mattikus left	02:33
	→ Radar joined	02:33
	→ mattikus joined	02:34
elfMobile	can I interact with an hg server using git?	02:34
blueshift	anjen: if they're in your history, they're going to be in your repo..	02:35
	← welly left	02:36
blueshift	are you running out of disk space or something?	02:36
	→ frakturfreak_ joined	02:36
blueshift	you could have them in the tree without having them in git, if that works	02:37
	→ posciak joined	02:38
	→ jrmuizel joined	02:38
	→ welly joined	02:38
anjen	blueshift: yeah, they look like they've made it into my history. As to wehre they'll be in the future, they're all going to S3 buckets for more permanent storage..access will eb trhough "normal" methods, rather than via public dir calls. upload will bypass git altogether, jsut the handlers will be present in the code	02:39
	and btw, thanks for the reply. I was afraid it went blackhole	02:39
blueshift	this is quiet time in this channel	02:40
	← frakturfreak left	02:40
anjen	blueshift: So then my history is permanently bloated?	02:40
	and understood on quiet time	02:40
blueshift	unless you want to rewrite history, yes	02:40
elfMobile	If you have changes to a file staged and changes to that file also unstaged you can use diff to see between the two, and diff --staged to see between staged and the repo.... but is there a way to see diff between working directory and repo?	02:40
anjen	ah...well that's a bit beyond my present capabilities, so I guess I'll have to live with it a while longer	02:41
	→ Arafangion joined	02:41
blueshift	if you want to remove them it may be more painful the longer you leave it	02:41
	do other people have the history containing the flash files? or just you?	02:41
	do you push and pull to other repos?	02:41
anjen	blueshift...only a few days before I can start handling it. they've been present for months...	02:42
	jsut me	02:42
	→ frakturfreak__ joined	02:42
anjen	solo git repo, no one else active	02:42
blueshift	if you have only one isolated repo then rewriting history is not a big deal	02:42
anjen	hmm...pointers to the "how"?	02:42
blueshift	git help filter-branch is probably what you want	02:43
	← gnufied left	02:43
	← frakturfreak_ left	02:43
blueshift	but of course at the moment if you checkout an old rev, it will have flash files and work. if you remove those from history, checking out an old version won't have the flash files any more	02:43
	if that is ok, then fine.	02:44
anjen	thank you blueshift I'll study up in the API... now that I know what to look for that is :)	02:44
blueshift	also once you've rewritten history you will need to clear out the junk in the git repo	02:44
	all the old stuff will stay in there until you clean it	02:44
	that bit I'm hazy on, haven't done myself, but have a look at git prune, or fsck, or clean.. something like that	02:45
	or git gc maybe it is	02:45
anjen	blueshift: understood.	02:46
	→ gnufied joined	02:46
anjen	thanks for the advise :)	02:46
blueshift	np	02:48
	← mattikus left	02:50
	Mage_Dude_ → Mage_Dude	02:50
	← gnufied left	02:51
blueshift	anjen: you may end up with a commandline like: git filter-branch --tree-filter 'find . -path ./.git -prune -o -name "*.swf" -delete' HEAD	02:51
	that's a guess.. may need some fiddling	02:51
	→ datakid joined	02:52
	→ timj_ joined	02:52
	→ mattikus joined	02:52
datakid	if I check out the master, make changes, then realise I should have made a branch before making the changes, what is the recommended method for doing this?	02:53
	ie, making a branch with all my changes, leaving master as it was	02:53
	is it to get a git diff, roll back to master, branch, then re apply diff?	02:54
	← spearce left	02:54
datakid	sorry, relative newbie	02:54
blueshift	nah	02:55
	git branch newbranchname	02:55
	git reset --hard origin/master	02:56
	← timj left	02:56
blueshift	git checkout newbranchname	02:56
	beware that will nuke any uncommitted changes you have..	02:56
	← zenazn left	02:56
dominikh	wouldn't it be easier/safer to just use git-stash, create a new branch and then apply the stash?	02:57
anjen	just noticed your last suggestion, thanks blueshift	02:57
	← anjen left	02:57
blueshift	dominikh: maybe, never used stash.	02:58
	I should look at it	02:58
dominikh	blueshift: also won't your reset "nuke" all commits that have not been pushed to origin?	02:58
	← frakturfreak__ left	02:58
blueshift	they will be on newbranchname	02:58
	← Eaven left	02:59
	← k0nichiwa left	02:59
dominikh	well yeah, but he only wants current changes in a new branch, not all commits he has done since he last pulled	02:59
blueshift	oh, maybe. wasn't clear to me	02:59
	← steph021_mac left	03:00
dominikh	datakid: did you commit the changes yet?	03:00
	← corni_ left	03:00
	→ Raimondi joined	03:01
	← mdel left	03:02
datakid	dominikh, no, not yet	03:02
	I was afraid of soiling master	03:03
dominikh	datakid: git stash; git checkout -b newbranch; git stash apply	03:03
	← zackattack left	03:03
datakid	but I guess I could commit, make a new branch, then reset master to the original - is that how it works	03:03
	→ mdel joined	03:03
datakid	dominikh, ok, thanks	03:03
dominikh	that should do the trick	03:03
blueshift	datakid: yeah, don't listen to me. I think I had the wrong idea about what you were doing	03:03
	← Radar left	03:04
	→ Radar joined	03:05
datakid	sorry to be daft - does stash recursively go down filesystem, or do I need to stash each file that's changed?	03:05
dominikh	datakid: git stash takes all unstaged changes and puts them away	03:06
datakid	dominikh, ok thanks	03:06
	→ spearce joined	03:06
dominikh	man git-stash	03:06
Gitbot	dominikh: the git-stash manpage can be found at http://git.or.cz/man/git-stash	03:06
	→ spearce` joined	03:09
	→ mikeric joined	03:09
	→ gnufied joined	03:10
	→ dmlloyd joined	03:12
dominikh	gtg	03:12
	← spearce left	03:13
	spearce` → spearce	03:13
	← Radar left	03:14
	← Sigma left	03:14
	→ Radar joined	03:14
	← stringoO left	03:15
datakid	that worked all, thanks!	03:15
	← datakid left	03:15
	← vu3rdd left	03:15
	→ k0nichiwa joined	03:19
	← Radar left	03:24
	→ Radar joined	03:25
	← comawhite left	03:25
	→ kreitz joined	03:27
	→ comawhite joined	03:27
	→ shiba_yu_ joined	03:28
	← shiba_yu36 left	03:30
	→ sh1mmer joined	03:37
	→ zackattack joined	03:41
	← spearce left	03:41
	← Radar left	03:42
	→ Radar joined	03:43
	← codysoyland left	03:49
	→ codysoyland joined	03:49
	← nothingHappens left	03:50
	ippa → ip^sleepz	03:52
	→ terinjokes joined	03:52
	← terinjokes left	03:52
	→ terinjokes joined	03:53
terinjokes	I'm trying to use git-svn with a public repo, but every time I try cloning I get an error at the same spot "error: failed to run repack	03:53
offby1	:-(	03:53
terinjokes	yes, frownie face is so	03:54
	← lucasvo_2 left	03:54
terinjokes	it also says "object <blah> is corrupt"	03:55
offby1	I assume those are related ... but I really don't know.	03:58
	I've had "git-svn clone" fail when the repo was pretty big	03:58
	and it's awfully slow, even when it does work.	03:58
	→ fivetwentysix joined	03:58
terinjokes	15584 revisions	03:59
	→ chris_n` joined	03:59
terinjokes	fails at r10474	03:59
	(during the auto repacking part)	03:59
	← chris_n left	04:01
	→ lucasvo_2 joined	04:01
	← Slix` left	04:03
	← Raimondi left	04:04
	→ Raimondi joined	04:05
	← Radar left	04:05
	← scarabx left	04:06
	→ Radar joined	04:06
terinjokes	offby1: and scrolling through the backlog, i don't see a revision with an ID with the same as the object that is corrupt	04:08
	← Mage left	04:09
offby1	shrug	04:09
	no ideas, sorry	04:09
	→ oriba joined	04:13
	→ eydaimon joined	04:16
eydaimon	can I specify a branch to checkout when I've got a git repo? What's the syntax?	04:17
	← jfkw left	04:17
	← gnufied left	04:18
	← phlowx left	04:18
	← Radar left	04:19
offby1	git checkout branchname	04:19
	you really should read at least _one_ git doc :-\|	04:19
terinjokes	offby1: what does history say about repeating something and expecting different results?	04:20
	→ Radar joined	04:20
	→ Slix` joined	04:20
offby1	terinjokes: I gotcha	04:20
	← apol left	04:20
	← nairb774 left	04:22
	← sivy left	04:23
terinjokes	offby1: this time i'm running it on another box, so i guess I could expect different results...	04:24
	→ steph021_mac joined	04:27
	→ reflectivedev joined	04:28
	→ akahn joined	04:28
	→ chris_n joined	04:31
	→ nanotube joined	04:32
	← magcius left	04:32
	← zimbres left	04:32
	← akahn left	04:32
nanotube	so... i'm trying to create a pre-commit hook that automatically updates a version string with a datetime stamp... however, file which is edited by the pre-commit hook does not get committed, even when i call 'git commit -a'...	04:32
	so my question is... is there a way to get it in there? or maybe is there a better way to tag the version string with a datetime automatically for each commit?	04:33
	→ pdignan joined	04:33
blueshift	nanotube: why not do it the other way round, have your build mecha generate a version from git. various projects do it this way inc linux kernel	04:34
nanotube	blueshift: er well... i don't have builds... i just tell people to pull a snapshot out of git.	04:34
	→ magcius joined	04:34
nanotube	but then... i want to be able to quickly ask people what version they're running, so i can tell what code they have	04:34
blueshift	hmm, I see	04:35
	← chris_n` left	04:35
	← phantomcircuit left	04:35
nanotube	so any suggestions? :)	04:35
blueshift	you could commit with a script instead.. but there may be a better way I don't know	04:35
nanotube	hmm, so have a script that does changeversion; git commit -a, basically?	04:36
	guess that could work...	04:36
blueshift	if you always simply commit -a	04:36
nanotube	well for the most part. but i guess i'd make it take args, and pass those args on to git commit, to be more generic about it. :)	04:37
	thanks for the suggestion, blueshift... i'll stick around, and if nobody else comes up with anything better, that's what i'll do then. :)	04:38
blueshift	or you could work on the basis that whatever is in your current version file is the one to use, and bump it on commit for the next version	04:39
	so it says v1, you do a bunch of work, commit: it contains v1 and your work, and bumps the file to v2 ready for your next work	04:40
	or it could bump a minor number, and you edit by hand 1.1 -> 2.0 before commit if you feel the need	04:40
	→ catalis joined	04:40
	→ consolers joined	04:42
	→ Manganeez joined	04:43
	→ chris_n` joined	04:44
	→ NET\|\|abuse joined	04:44
Manganeez	Hello. I'm sure many people comes here and says the same thing: "I'm a complete and utter git newb." Well, I'm a complete and utter git newb.	04:45
terinjokes	offby1: they also say, 3rd times a charm... all checked out	04:45
blueshift	Manganeez: congrats. have a badge.	04:46
	→ chris_n`` joined	04:46
	← chris_n left	04:47
	← catalis left	04:47
	← jrmuizel left	04:49
	← chris_n` left	04:49
Manganeez	Here's what I'm trying to figure out how to do: I want to use git-svn to mirror wordpress's svn repo into my own bare git repo. I then want to branch from that to make a slightly customized version (a few of my "standard" config tweeks, plugins, etc.). I then want to branch from that for each of my wordpress sites so I can track per-site customizations. We are three devs, so I want all of this to be shared via a bare repo, and ideally	04:49
	, I'd like the upgrade to be as easy as switching some reference for that original vendor-tracking branch then pulling through the downstream levels. Does this make sense? Am I thinking wrong?	04:49
	← chris_n`` left	04:50
	← Spines left	04:51
	→ bcardarella joined	04:51
	← bcardarella left	04:52
offby1	terinjokes: weird.	04:52
	Manganeez: I honestly don't know. It sounds like it's worth a try.	04:53
nanotube	blueshift: hrm, none of those other alternatives strike me as being what i'd really want to do (which is for the timestamp to be equal to the datetime of the commit, and to avoid any manual editing)	04:53
Manganeez	hehe - I've been trying. Can't figure out details. Like how to make a branch track another branch in such a way that when I push it to the bare repo, the tracking association is maintained when another dev updates and does a checkout.	04:54
	← NET\|\|abuse left	04:54
blueshift	nanotube: do you actually need a timestamp? can you tag in the commit hook I wonder? you could tag based on the contents of the version file	04:55
	and if it was a simple single version number, you wouldn't need to manually edit	04:56
	I tend to use major.minor.build form where build gets auto bumped	04:56
Mage_Dude	nanotube: I missed the need for the timestamp. Errr why do you need the commit time re-recorded as a timestamp?	04:59
	← fivetwentysix left	04:59
	→ freimatz joined	05:01
	→ typemore joined	05:01
typemore	is there an open source tool (perferably runs on linux) that provides a nice wiki + bug tracker integrated with git?	05:01
	we have a largeish project, and i'd prefer to keep all documentation (in wiki format) + todo lists / trackers together with the git repo	05:02
	← pdignan left	05:02
	→ pdignan joined	05:05
Mage_Dude	typemore: No but if you write one that would be nifty. ;)	05:05
	→ Einstein_ joined	05:06
	← mdel left	05:07
	→ girishr joined	05:07
	→ jony joined	05:08
	← girishr left	05:08
nanotube	Mage_Dude: the thing is that i have no builds, i just tell people to pull snapshots out of git.	05:09
	Mage_Dude: and then i need a way to tell which code people are using... so i want to have each git commit stick in a timestamp into the version number. then i ask people "what's your version" and i know.	05:09
	← Einstein_ left	05:09
nanotube	Mage_Dude: that's the idea	05:09
Mage_Dude	Are the people capable of running a system command to check it?	05:10
	← apsysoev1 left	05:10
blueshift	nanotube: that would work with the version file, auto tag and bump on commit thing I was suggesting	05:11
nanotube	Mage_Dude: what do you mean? it is easy to see what version you're running... but not easy to figure out which commit it came from, unless version changes on every commit.	05:11
	blueshift: that's pretty much what i'm doing... or what do you mean by 'autotag' exactly? create git tags for every commit?	05:11
blueshift	yeah	05:12
Mage_Dude	nanotube: Are they also adding changes to tho code, but not in a different branch?	05:12
nanotube	so there would be two commits for every actual commit? and there would be a crapload of tags? and the tags would still not enable people to figure out ex-post which commit they pulled?	05:13
	→ MattDiPa_ joined	05:13
blueshift	tags aren't commits, there would be a crapload of tags, they would exist to make it easy for you to refer a version back to a commit	05:13
nanotube	Mage_Dude: no, people are just running the code. but if they come to me and say "such and such bug bla bla" i say "what version are you running" and i can then see if the bug could possibly have been fixed in subsequent commits.	05:13
blueshift	or you could skip the tags and do it by searching history for a version	05:14
Mage_Dude	So they can't run 'git log' to see the last commit?	05:14
	← posciak left	05:14
nanotube	blueshift: ah well yes not commits... but still, person goes and grabs a snapshot and installs... then how would he then figure out later which tag?	05:14
blueshift	nanotube: or, I think there's a hook that lets you rewrite commit messages, or preload them with stuff before you commit. you could read the version file and put it in the commit message	05:15
nanotube	Mage_Dude: well, i'd not tell them to 'go upgrade to the latest' if it won't help... so it would help to know which version they're running.	05:15
	→ svaksha joined	05:15
blueshift	nanotube: because the version is stored in a file. the commit hook reads that version and creates a tag (or commit message) based on it	05:15
	← MattDiPasquale left	05:15
nanotube	blueshift: Mage_Dude ok maybe i'm not explaining myself properly... let me try to describe from the beginning.	05:16
blueshift	version file is for the user, tag or commit message is for you to refer back to it	05:16
	← Radar left	05:16
	← typemore left	05:16
blueshift	or git log -S and just use the version file directly.. bit slower but maybe ok for you, and cleaner than lots of tags	05:16
	← pdignan left	05:16
	← tedoc2000 left	05:17
	→ gnufied joined	05:17
nanotube	i tell people to get and install code by grabbing snapshots from the git repo. they go to the gitweb page, click the 'snapshot' link, and get a tgz. they then install it and do whatever for $sometime.	05:17
Mage_Dude	Got it. They aren't running git...	05:17
	→ vu3rdd joined	05:18
nanotube	some time later, one of them may come by and report some undesirable behavior they're seeing. i want to be able to know what commit they are running, so i can tell whether i have made any relevant changes since that.	05:18
	Mage_Dude: right, they are not running git.	05:18
	← ignatp left	05:18
nanotube	so tags and stuff is not going to help them.	05:18
	only a changed version string, which they can easily get and tell me, will be of help.	05:18
	→ bobbyt joined	05:19
nanotube	i can't expect them to remember the sha hash of the commit that they pulled the snapshot from, or the tag name, or anything.	05:20
Mage_Dude	But of course you'd like to generate a file before committing the changes to the snapshot so that a user can figure it out.	05:20
	→ soreau joined	05:21
soreau	$ git clean -fdx	05:21
	Removing build/	05:21
	warning: failed to remove 'build/'	05:21
	What is the deal with this?	05:21
	← bobbyt left	05:22
	→ pdignan joined	05:23
	→ killerchicken_ joined	05:23
killerchicken_	hey, how do I edit a staged hunk?	05:24
	← kreitz left	05:24
nanotube	Mage_Dude: right. i tried it with a pre-commit hook, but that doesn't make it into the commit. blueshift suggested to just use a script, which first makes the change to version, then runs commit. that works just fine. but i was just wondering if there wasn't a "better way (tm)" :)	05:24
madewokherd	killerchicken_: modify the file, then git add it?	05:24
frogonwheels	killerchicken_: vim + fugitive allows you to	05:25
killerchicken_	madewokherd: I have lots of changes in the file that I don't want to stage	05:25
frogonwheels	killerchicken_: what _exactly_ are you trying to do	05:25
	killerchicken_: so modify it in the working dir, then stage the lines you wanted staged	05:25
killerchicken_	well, I did that	05:25
soreau	I want it to tell why it failed to remove the build directory, not just a warning with no explanation and keep on kicking	05:25
killerchicken_	except that patch -p doesn't show me those lines.	05:25
	and for each new try I need to go through all the changes again	05:26
soreau	Can someone please tell me why this fails? $ git clean -fdx \n Removing build/ \n warning: failed to remove 'build/'	05:26
Mage_Dude	nanotube: Nothing else is coming to mind at the moment.	05:26
frogonwheels	killerchicken_: so either use git gui or git cola or vim+fugitive	05:26
	blueshift sighs	05:26
	← flaguy48 left	05:27
madewokherd	or diff your working version of the file from the staged version, checkout the staged version, modify that, git add it, and reapply your diff	05:27
	← guilhermeblanco left	05:27
	← Manganeez left	05:27
nanotube	Mage_Dude: blueshift: well thanks :)	05:28
soreau	Hello folks?	05:29
	soreau waves hands frantically	05:29
soreau	Should be a simple issue......	05:29
	git clean -fdx doesnntwork. It says Removing build/ \n warning: failed to remove 'build/'	05:29
	wtf is this supposed to mean?	05:29
	no --verbose?	05:29
	It's permissions issue	05:31
	should have said Permission Denied	05:31
	Useful error messages = less headaches	05:31
	← soreau left	05:31
	← MattDiPa_ left	05:33
	→ MattDiPasquale joined	05:33
	← reflectivedev left	05:34
	→ sivy joined	05:37
	← zjason left	05:38
	→ roop joined	05:38
	→ thiago_home joined	05:40
	← MrGando left	05:42
	→ CrypTom joined	05:43
	→ marc_kdab_ joined	05:43
	← sivy left	05:44
	← consolers left	05:44
	→ timotei joined	05:44
timotei	anyone knows why I got this?: Checksum mismatch: trunk/data/core/units/undead/Corpse_Ghast.cfg e6810d74b525c653e75f0f2f8eeed60c46fe6980	05:44
	and how can I fix it?	05:45
	← mikeric left	05:46
frogonwheels	timotei: git-svn ?	05:46
timotei	yes	05:46
	:(	05:46
	frogonwheels shrugs.	05:46
timotei	why?:P	05:46
	I've got that once on monday,	05:46
frogonwheels	google "git checksum mismatch" you get a page of stuff relating to git-svn	05:47
timotei	redownloaded the git tarbal, reseted, and it worked till yesterday	05:47
	I've already googled	05:47
	hmm	05:47
	ok	05:47
frogonwheels	not really up to speed on problems in git-svn	05:47
	← Tuomas left	05:48
	← fod left	05:48
	← mquin left	05:48
	→ mikeric joined	05:48
CrypTom	Hi all, I'm new to git and trying to use git for my homedir. When I try to push to my QNAP NAS, git always crashes with the following error:	05:50
	Counting objects: 30639, done.	05:50
	Delta compression using up to 2 threads.	05:50
	Compressing objects: 100% (28627/28627), done.	05:50
	fatal: Out of memory, malloc failed, 1.12 GiB \| 135 KiB/s	05:50
	error: pack-objects died of signal 13	05:51
	error: pack-objects died with strange error	05:51
	error: failed to push some refs to 'ssh://.......'	05:51
	I've tried to repack with low depth and window values, without success	05:52
	← MattDiPasquale left	05:52
frogonwheels	CrypTom: did you do a git fsck?	05:52
CrypTom	frogonwheels: no, not yet	05:52
	→ MattDiPasquale joined	05:52
	← mutex_ left	05:52
CrypTom	do I just run "git fsck" or do I need to specify options?	05:53
frogonwheels	CrypTom: no options afiak.	05:53
CrypTom	frogonwheels: ok, fsck is running	05:53
	← adam__ left	05:53
	→ mutex_ joined	05:54
frogonwheels	CrypTom: repacking won't help also (afaik) - I believe the packing for push is independent of the local pack state.	05:54
	CrypTom: you could try a git bundle - but I expect it to fail similarly.	05:54
timotei	hmm. frogonwheels the solution was hidden :D	05:54
	no wonder didn't saw it	05:54
CrypTom	frogonwheels: ah, I see	05:54
timotei	I used: git svn rebase --ignore-paths data/core	05:54
	and it worked somehow:)	05:54
	thanks anyway	05:54
frogonwheels	np timotei	05:55
	→ fivetwentysix joined	05:55
thiago_home	is this an initial push? i.e., other side empty	05:55
	if it's not, then push will need to create its own pack	05:55
	→ mquin joined	05:56
	→ Dave^\| joined	05:58
	← timotei left	06:00
	← Textmode left	06:02
	→ girishr joined	06:03
	→ paja joined	06:03
	← girishr left	06:05
	→ Jippi_mac joined	06:05
	→ Radar joined	06:06
	→ sivy joined	06:08
	← quotemstr1 left	06:08
	→ Tuomas joined	06:09
	← krawek left	06:09
	← thiago_home left	06:09
CrypTom	frogonwheels: fsck showed about 50 dangling blobs, what does that mean?	06:10
	← MattDiPasquale left	06:12
	→ MattDiPasquale joined	06:13
	← svaksha left	06:13
	← madewokherd left	06:14
	→ svaksha joined	06:14
	→ Eaven joined	06:14
frogonwheels	CrypTom: I believe it means that some things are missing :\|	06:14
	← plediii left	06:14
frogonwheels	oh.. it says they aren't a problem.	06:15
	→ davidw joined	06:15
	davidw → Guest6932	06:16
	← sivy left	06:16
	← justin-george left	06:17
	← oriba left	06:18
	→ ph^ joined	06:19
	← paja left	06:21
frogonwheels	CrypTom: read the manapge - they're just things left over from operations that would get garbage collected eventually	06:21
	→ kernelzilla joined	06:24
	→ rchavik joined	06:24
	→ quotemstr__ joined	06:26
	← Solow left	06:26
	→ priidu joined	06:27
	← pdignan left	06:27
CrypTom	frogonwheels: ok, thanks, so I'm trying to push again	06:27
	← quotemstr left	06:28
	quotemstr__ → quotemstr	06:28
	→ akahn joined	06:28
frogonwheels	CrypTom: you might think about what version of git you're using?	06:29
	coyO → coyo	06:30
	← unhygienix left	06:32
	← MattDiPasquale left	06:32
	→ MattDiPasquale joined	06:33
	→ apsysoev joined	06:33
	← akahn left	06:33
	← hebz0rl left	06:34
	← lucasvo_2 left	06:35
	← patthoyts left	06:37
	→ lucasvo_2 joined	06:41
	→ doener joined	06:42
	← programble left	06:42
	→ ExtraSpice joined	06:44
	← Jippi_mac left	06:44
	→ apsysoev1 joined	06:51
	← MattDiPasquale left	06:52
	← apsysoev left	06:52
	→ dpino joined	06:53
	→ MattDiPasquale joined	06:53
	→ justin-george joined	06:54
	← Morphous left	06:55
CrypTom	frogonwheels: git version 1.7.0.4 (Ubuntu 10.04)	06:55
	frogonwheels: git push is running and almost at the point, where it always crashed	06:56
	→ galderz joined	06:58
	← galderz left	06:58
	→ galderz joined	06:58
CrypTom	frogonwheels: it crashed again at the same place as always	06:59
	→ trivol joined	07:00
	← LongBeach left	07:01
	← quotemstr left	07:02
	← brntbeer left	07:02
	← galderz left	07:03
	→ unreal_ joined	07:04
	← unreal left	07:07
	unreal_ → unreal	07:07
	→ yngress joined	07:07
	→ frakturfreak__ joined	07:08
	→ unreal_ joined	07:08
	→ Morphous joined	07:09
	← Yuffster left	07:10
	→ Schmallon joined	07:11
	← unreal left	07:12
	unreal_ → unreal	07:12
	← MattDiPasquale left	07:12
	→ MattDiPasquale joined	07:13
	→ bsodmike joined	07:13
	← hagabaka left	07:13
	→ StrongOrder joined	07:13
	→ hebz0rl joined	07:13
	→ Deesl joined	07:13
	← frakturfreak__ left	07:16
	Guest6932 → davidw	07:16
	← davidw left	07:16
	→ davidw joined	07:16
	→ ntoll joined	07:19
CrypTom	are there any other options to prevent "fatal: Out of memory, malloc failed" when using git push? Is this an issue on the remote maching (QNAP NAS) or local?	07:20
	← killerchicken_ left	07:20
	→ killerchicken_ joined	07:21
	← k0nichiwa left	07:21
	→ pdelgallego joined	07:23
ybit_	w/c	07:23
	→ nevyn^ joined	07:23
	← ybit_ left	07:23
	→ nuoHep joined	07:23
	→ leo2007 joined	07:24
	← jjuran left	07:24
	← Radar left	07:24
	→ jjuran joined	07:24
cbreak_work	CrypTom: it means your server has too little memory	07:25
leo2007	1772 has come out.	07:25
	← bsodmike left	07:26
leo2007	oh no	07:26
	1722	07:26
	→ bsodmike joined	07:26
	← leo2007 left	07:28
kernelzilla	CrypTom, this may related to the issue: http://git.zen-kernel.org/?p=kernel/zen-stable.git;a=commit;h=274e8f400363a453c0501fe964fdf3ebcc2a62f1	07:33
	← yashi left	07:34
	→ pdignan joined	07:36
CrypTom	cbreak_work: my NAS has 256MB RAM and while receiving the push, the free memory varies from 20-50MB, but never drops below 20MB?!	07:38
	→ tobiassjosten joined	07:39
	← Schmallon left	07:41
	← zackattack left	07:41
	→ voxpelli joined	07:42
	← voxpelli left	07:43
	→ voxpelli joined	07:43
CrypTom	cbreak_work: and the swap partition (512MB) shows about 180MB used	07:43
	so, there should be enough memory	07:44
	← fivetwentysix left	07:45
	→ tvw joined	07:45
	→ rraasch joined	07:46
	→ NET\|\|abuse joined	07:47
	→ airborn joined	07:48
	← bsodmike left	07:48
	← pdignan left	07:49
	→ engla joined	07:49
	← engla left	07:49
	→ engla joined	07:49
	← mikeric left	07:50
	→ nicoulaj joined	07:50
	→ pdignan joined	07:50
	→ bsodmike joined	07:50
	→ bsodmike_ joined	07:51
	← MattDiPasquale left	07:53
	→ MattDiPasquale joined	07:54
	← fahadsadah left	07:54
	→ whitman joined	07:54
	← bsodmike left	07:55
	bsodmike_ → bsodmike	07:55
	→ cash joined	07:56
	→ psoo joined	07:56
	→ parasti joined	07:57
	← gnufied left	07:57
	← parasti left	07:57
	→ parasti joined	07:58
	→ Radar joined	07:58
	→ fahadsadah joined	08:00
	→ girishr joined	08:02
	← steph021_mac left	08:02
	← parasti left	08:06
	← yann left	08:08
	← cash left	08:10
	→ bsodmike_ joined	08:11
	ChanServ set mode: +o	08:12
	wereHamster changed the topic to: 1.7.2.2 \| Homepage: http://git-scm.com \| Everyone asleep or clueless? Try [email@hidden.address] \| Channel log http://tinyurl.com/gitlog \| Mailing list archives: http://tinyurl.com/gitml \| Gits on git: http://tinyurl.com/gittalks \| Pastebin: http://gist.github.com/	08:12
	ChanServ set mode: -o	08:12
	→ parasti joined	08:12
	← MattDiPasquale left	08:13
	← bsodmike left	08:13
	bsodmike_ → bsodmike	08:13
	→ MattDiPasquale joined	08:13
	← justin-george left	08:15
	← girishr left	08:17
	→ [Po]lentino joined	08:17
rchavik	hi.. is it possible to undelete a local branch?	08:23
Rhonda	If you know the sha1 of its head, yes.	08:24
	Potential you can look that up in git reflog when you had it checked out recently.	08:24
	→ psankar joined	08:25
blueshift	reflog is my friend.	08:26
Rhonda	Isn't it everyone's? :)	08:26
	→ fivetwentysix joined	08:28
	← amerine left	08:28
	→ akahn joined	08:28
	← mattikus left	08:31
rchavik	ok.. i think i found it.. then what do i do with the sha1?	08:31
	← akahn left	08:33
	← MattDiPasquale left	08:33
wereHamster	rchavik: create the branch	08:33
	→ MattDiPasquale joined	08:33
rchavik	this way: git co -b <lostbrachn> <sha1> ?	08:33
	→ tonnzor joined	08:33
wereHamster	co is not a git command	08:34
rchavik	sorry.. checkout.. co is my alias :)	08:34
wereHamster	that's one way to create a branch	08:34
	← airborn left	08:36
	→ thedjpetersen joined	08:36
	→ tatsuya joined	08:36
	→ airborn joined	08:37
	→ dr_win joined	08:37
	← tvw left	08:37
	→ mikeric joined	08:37
	← airborn left	08:37
	→ airborn joined	08:38
	← voxpelli left	08:38
	→ voxpelli joined	08:38
	← Dave^\| left	08:38
	← Slix` left	08:38
	→ workmad3 joined	08:39
	→ tvw joined	08:40
tonnzor	hi. I want to generate changelog between old deployment version and the new one. I tried git log REV1..REV@, but it shows be ALL changes from repository including other branches	08:40
	→ sid3k` joined	08:41
tonnzor	git diff generates DIFF, not commits log	08:41
	← ribasushi left	08:42
	← bsodmike left	08:42
	← lemonchicken left	08:42
	← sigmonsays left	08:42
	← tehbaut left	08:42
	→ sigmonsay joined	08:42
	→ ribasushi joined	08:42
	← seanius left	08:43
	→ bsodmike joined	08:43
tonnzor	any ideas on generating changelog?	08:43
Rhonda	Why the @ at the end? git log rev1..rev2 works here.	08:43
	→ masterkorp joined	08:43
tonnzor	sorry, just a typo	08:44
	→ codeshepherd joined	08:44
	→ seanius joined	08:44
	→ tehbaut joined	08:44
	→ Titosemi joined	08:44
Rhonda	And be sure it's oldversion..newversion, not the other way round.	08:44
tonnzor	git log rev1..rev2 shows all changes from other branches during that time	08:45
Rhonda	When those got merged in, right?	08:45
wereHamster	tonnzor: in git a commit doesn't belong to any branch in particular.	08:45
codeshepherd	when i say "git pull origin mybranch:mybranch" .. it refuses to pull saying non--fast-forward .... however when i say "git fetch origin mybranch" and then do "git merge FETCH_HEAD" it works fine.. why is this? is there way to do fetch and merge at one go?	08:45
wereHamster	codeshepherd: git pull origin mybrach	08:45
codeshepherd	wereHamster: what is the difference between mybranch:mybranch and just mybranch?	08:46
doener	codeshepherd: "git pull origin mybranch:mybranch" does "git fetch origin mybranch:mybranch", which means "overwrite mybranch with origin's mybranch", which gets refused	08:46
	→ Schmallon joined	08:46
doener	codeshepherd: while "git pull origin mybranch" does "git fetch origin mybranch" which stores origin's mybranch in FETCH_HEAD only (as you already know)	08:47
	← bsodmike left	08:47
doener	codeshepherd: the remote and refspec arguments to pull are only for fetch. The target for the merge is always the checked out branch	08:47
	→ bsodmike joined	08:48
	← bsodmike left	08:48
codeshepherd	thanks doener	08:48
	→ fluter joined	08:49
rchavik	why git does not record branch delete operation in reflog?	08:49
fluter	hi, buddies, how to get the modified file list using git log?	08:49
	→ justin_ joined	08:50
wereHamster	rchavik: in which reflog should it record it?	08:50
doener	rchavik: because the reflog gets deleted with the branch head?	08:50
	fluter: --name-only or --name-status	08:51
	→ bsodmike joined	08:51
	→ ciampix joined	08:52
fluter	doener: how to omit the message body in git log --name-only	08:52
wereHamster	fluter: use git idff	08:52
fluter	I only want to see which files are modified	08:52
wereHamster	diff even	08:52
	← MattDiPasquale left	08:52
	→ MattDiPasquale joined	08:53
rchavik	wereHamster, doener, ok i think i understand.. i kinda though that reflog is recorded in a logfile or something	08:54
fluter	git diff will output the whole patch	08:55
wereHamster	fluter: you need to use --name-only/--name-status of course	08:56
	→ bentob0x joined	08:56
fluter	wereHamster: I see, --name-only is what I want , thanks	08:56
	← Radar left	08:56
	→ bushwakko joined	08:57
	→ bobpitbull joined	08:57
bobpitbull	hi, any SmartGit users in here?	08:57
	← tuxillo left	08:58
	→ curtana joined	08:58
	→ Radar joined	08:58
	← roop left	08:58
	→ jatt joined	08:59
wereHamster	would it make a difference it there were?	08:59
jatt	Is it possible to activate syntax highlighting in gitweb's blob view?	08:59
	→ lemonchicken joined	09:00
bobpitbull	i'm just struggling with the whole Merge thing... i'm a die-hard Perforce user... just been using Git/SmartGit/GitHub for a few days now... our whole team is struggling to figure out how to use Merge.	09:00
workmad3	bobpitbull: you do realise you can set up git to use p4merge as a merge tool, right?	09:01
	← tonnzor left	09:01
	← psankar left	09:01
	→ bluenovember joined	09:03
	→ cash joined	09:04
	→ AlexC_ joined	09:04
AlexC_	morning	09:04
	→ Digitalos joined	09:05
AlexC_	how long does a git stash last for? I have some changes in 'master' that I am not ready to commit, so I just did 'git branch foobar' to branch off, then I did 'git checkout master' - but of course this kept the changes on master	09:05
	if I switch to 'foobar' and do 'git stash', then go back to master, will the stash still be there?	09:05
wereHamster	AlexC_: until you delete the stash	09:05
cxreg	AlexC_: not only will it still be there, but you can apply the stash to any branch you want	09:07
AlexC_	most excellent, thanks :)	09:07
	→ intripoon joined	09:07
	→ picardo12 joined	09:10
	← Gunni left	09:11
	← mithro left	09:12
	← bsodmike left	09:13
	→ bsodmike joined	09:15
	← lucasvo_2 left	09:15
bobpitbull	workmad: sorry for not responding... we don't have Perforce here unfortunately :-(	09:17
jaeckel	AlexC_: why not creating a development branch, checking in the changes and switching back to master? This would allow you to backup the changes you made that are not ready to be integrated into master.	09:17
	→ ixti joined	09:17
Radar	+1 jaeckel	09:18
wereHamster	bobpitbull: what's your problem? You don't know how to merge (with smartgit)?	09:18
AlexC_	jaeckel: it would still involve comitting half-done code with errors	09:19
wereHamster	so?	09:19
jaeckel	AlexC_: and where's the problem?	09:19
AlexC_	seems like a very messy way of doing it to me	09:19
bobpitbull	wereHamster: yeah... never used Git before. Most of our files are data files and, usually, it should just be a matter of accepting either our Local file or the Remote one. The whole "Accept Yours"/"Accept Theirs" thing in Perforce?	09:20
wereHamster	AlexC_: in git commits are not final, you can edit them as much as you want	09:20
AlexC_	true	09:20
wereHamster	bobpitbull: no idea how to do it with smartgit (or if it's even possible). Commandline ftw!	09:21
bobpitbull	wereHamster: i know that we're not recommended to use Data files in Git (at least that's what tekkub told me) - but most of our files are Data ones... we're being very careful not to work on the same data files at the same time... but, still, SmartGit constantly wants us to do merges	09:21
ptor	AlexC_: Keep the master branch clear (builds cleanly, lets you use git-bisect and the like). A feature/test branch doesn't have to, and before you merge to master you can squash commits or whatever.	09:21
jaeckel	AlexC_: and checking in half-finished things isn't something bad... noone is forced to merge your branch and that would be the only way how his build would break	09:21
bobpitbull	wereHamster: we have artists working here though... we can't expect them to learn the commandline options etc... it's hard enough getting them to do simple things in JavaScript and so on...	09:22
AlexC_	I may start doing that, though this situation very rarely appears with us - but thanks for the suggestions, it is a possibility	09:22
	→ lucasvo_2 joined	09:22
	← tvw left	09:22
ptor	AlexC_: To handle a very simple way of managing half-done work you just check in what you have (on your development branch), switch back to master, switch back to dev. branch, and then you just keep editing and 'git commit --amend' your last one. That way you can finish each commit piecewise.	09:24
wereHamster	bobpitbull: why are you even using git (for binary files)?	09:24
AlexC_	ptor: sure does sound simple	09:25
ptor	AlexC_: What I described there is of course a lot like what git stash can do, but I've found that if I have to stash all the time there's a risk of me doing something (e.g. a git reset for another reason) and I lose the work anyway.	09:26
	→ ivan-kanis joined	09:26
ivan-kanis	is there a way to clone just one file from a repo?	09:27
wereHamster	no	09:27
ptor	ivan-kanis: You can 'git archive' with a path (afaik) though.	09:28
	(not certain though)	09:28
ivan-kanis	hmm it looks promising	09:28
ptor	Now certain, did a 'man git-archive' :-) You can git archive with a path.	09:28
Gitbot	ptor: the git-archive manpage can be found at http://git.or.cz/man/git-archive	09:28
	→ jumoit joined	09:28
jaeckel	ptor: but you must do a clone of the whole archive before...	09:29
	s/archive/repository	09:29
ptor	ivan-kanis: You don't get a git repo out of that of course, so no 'git log' or anything (but it wouldn't make much sense for a single file).	09:29
wereHamster	not with archive --remote\	09:29
ptor	jaeckel: No, you can use the command against a remote git url	09:30
wereHamster	jaeckel: ^^	09:30
jaeckel	oh :)	09:30
	cool	09:30
ivan-kanis	too bad i'll have to untar it...	09:30
jaeckel	didn't know	09:30
ptor	ivan-kanis: Or unzip it.. in any case it can archive to stdout, so you just pipe it through tar, no .tar file will land on your disk.	09:31
ivan-kanis	ptor: thanks it will do the trick	09:31
	→ Saur joined	09:33
	← MattDiPasquale left	09:33
	→ MattDiPasquale joined	09:34
	← Arafangion left	09:34
curtana	could you not grab the commit that HEAD points to, then grab the trees it contains, and look for the file you want, and grab that file?	09:34
	→ Arafangion joined	09:34
curtana	with an appropriate amount of low-level commands	09:35
wereHamster	curtana: how do you 'grag the commit that HEAD ponits to'?	09:35
jumoit	any ideas on how to manage new repositories by gitosis.conf? for example, what if i want to create a new repo, called tools/test/project0, in the /home/git/repositories/?	09:36
curtana	i duno. i assumed you could do the same stuff that git-clone eventually does	09:36
wereHamster	jumoit: use gitolite	09:36
	→ arialdo joined	09:36
jumoit	wereHamster: i'm nothing else than novice in git. so,	09:37
ptor	curtana: 'git archive' does whatever low-level commands are needed, so that you won't have to.. :-)	09:37
	→ fero joined	09:38
curtana	it will avoid transferring all the other files in the repository?	09:38
	→ madewokherd joined	09:39
ptor	curtana: I don't know exactly what it does (i.e. what needs to be transfered from a remote repo). I guess it's just to try to get a single file from a huge remote repo and then it'll be pretty clear how much data goes over (e.g. monitor it with pload or whatever)	09:40
	→ khmarbaise joined	09:40
	→ alsuren_ joined	09:42
	→ \|albel727\| joined	09:42
	← bobpitbull left	09:43
	→ erlnoob joined	09:43
	← ExtraSpice left	09:43
	→ bilkulbekar joined	09:44
	← albel727 left	09:45
	← bauruine left	09:50
	← Juri_Agarin left	09:51
	→ Juri_Agarin joined	09:51
	→ bsodmike_ joined	09:51
	← bsodmike_ left	09:51
	→ Shurakai joined	09:52
	← fluter left	09:52
	← MattDiPasquale left	09:53
	← bsodmike left	09:53
	→ MattDiPasquale joined	09:53
	← AlexC_ left	09:54
	← fero left	09:55
	← ntoll left	09:55
	← fivetwentysix left	09:59
	→ galderz joined	10:00
	← galderz left	10:00
	→ galderz joined	10:00
	← codeshepherd left	10:01
	→ Tobias\| joined	10:02
	← jony left	10:02
	← benjoldersma_ left	10:04
Ilari	Ugh... "fatal: remote error: fatal: 'repositories/pub/gtd.git' does not appear to be a git repository"... Wonder what the heck did go wrong (it somehow passed control to git without creating the repo directory).	10:04
	→ Rasherz joined	10:06
ivan-kanis	is is safe to push onto git while the remote repository is being repacked?	10:06
Ilari	ivan-kanis: Yes.	10:07
	→ villev joined	10:09
Rasherz	Can anyone fill me in on why I keep getting a "remote end hung up unexpectedly" please?	10:11
	I'm just trying to clone from a VPS to local machine	10:11
	→ Paraselene_ joined	10:12
Rasherz	git clone git+ssh://root@myserver/folder/project.git	10:12
	asks for a password	10:12
	← justin_ left	10:12
Rasherz	enter that and it hangs up on me	10:12
	← MattDiPasquale left	10:13
	→ MattDiPasquale joined	10:13
Rasherz	been at this for a while but not moving anywhere	10:13
wereHamster	Rasherz: is git installed on the server?	10:15
	also, use just just ssh://, not git+ssh	10:15
Rasherz	i can run git --version through putty	10:15
	← NET\|\|abuse left	10:15
	→ benjoldersma_ joined	10:16
wereHamster	Rasherz: does the git repo on the server exist?	10:16
Rasherz	yeap	10:16
	ran the bare init	10:16
wereHamster	Rasherz: does 'ssh root@myserver git --version' work?	10:17
Rasherz	nope, getting command not found using that method	10:18
	so works through root access on putty but not through ssh. Do I have access issues?	10:18
wereHamster	no, it's simply that git is not in your PATH	10:19
	also, create a separate user account, don't use root	10:19
Rasherz	cool, will get right on that	10:19
	← Tobias\| left	10:20
Rasherz	cheers Hamster, was running around in circles there for a while	10:20
	← tapvt left	10:21
	→ jony joined	10:21
	← ciampix left	10:23
hachi	I've got a contrived test here, branch 'master', which has a couple extra commits on it on the remote side	10:24
	haven't pulled them on purpose, trying to get this to work	10:25
	git checkout master; git merge --no-ff release	10:25
	now I have a merge commit, and my history of master has diverged from that of upstream	10:25
	git fetch origin; git rebase -p origin/master	10:25
	→ sgh joined	10:26
sgh	Hi !	10:26
hachi	now my merge commit has been dropped, and the commit in 'release' has been played on master as a.. cherry pick or rebase	10:26
Ilari	That error message was apparently caused by some sort of error in configuration (that the configuration checker doesn't detect).	10:26
hachi	isn't -p supposed to preserve the merge commit, as a merge commit?	10:26
doener	old git version? -p needs -i, which is only implied in semi-recent git versions	10:28
hachi	1.7.0.3 sufficiently old?	10:28
	also, how does -i help?	10:28
	← hipe left	10:28
doener	no, -i is implied by -p since 1.6.1	10:28
	→ akahn joined	10:29
hachi	the rebase was in no way interactive when I ran it	10:29
doener	hachi: "rebase -i" is basically a different program, and that one supports -p, the plain rebase command doesn't	10:29
	→ aspotashev joined	10:29
sgh	If following the procedures described in "man gitworkflows" what would then be best practice to use when having to make an ugly hack in 'maint' but a better implementation shoould be done in 'master'. Merging 'maint' into 'master' would merge the ugly hack. Not merging would prevent furhter changen in 'maint' to be merged into 'master'. Any ideas?	10:29
Gitbot	sgh: the gitworkflows manpage can be found at http://git.or.cz/man/gitworkflows	10:29
hachi	ok, but assuming -i was called... it didn't pop up an editor at all	10:29
doener	hachi: if you use only -p, the editor is set to a no-op	10:30
hachi	ahh	10:30
	so then the thing I'm trying to figure out, where did my merge commit go	10:30
	← MattDiPasquale left	10:32
wereHamster	hachi: why not pull origin/master and then merge release?	10:33
	→ MattDiPasquale joined	10:34
	← akahn left	10:34
hachi	the situation is that other people are updating origin/master too	10:34
	so in reality I try to push, failed push	10:34
	I need to 'pull' again	10:34
	but we use rebase=true	10:35
	which I sorta like, I would like to rebase manually with -p and have it preserve the merge	10:35
	← hebz0rl left	10:35
hachi	so... I'm trying to do exactly what I think the manpage claims it can do	10:35
	← Eaven left	10:35
doener	hachi: after that rebase, does "git rev-list master..release" output anything?	10:38
	← blueshift left	10:39
	→ NET\|\|abuse joined	10:39
hachi	doener: yes, but it happens to be the same commit on both sides	10:40
	er	10:40
	→ froschi joined	10:40
	→ blueshift joined	10:40
hachi	it only lists one, like you're asking about	10:40
	but it's the same commit on both sides of the merge base, essentially cherry-picked	10:40
	er, same change on both sides	10:40
doener	hachi: ok, so "git rev-list --left-right --cherry-pick master...release \| grep '>'" is empty, right?	10:41
	→ Torsten joined	10:42
hachi	correct	10:42
doener	hachi: and "git rev-list --left-right --cherry-pick origin/master...release \| grep '>'"?	10:42
hachi	shows me the original commit I made on the release branch	10:44
	→ shruggar joined	10:44
	← jumoit left	10:44
	← andre_pl left	10:45
	→ Vampire0 joined	10:45
hachi	checkout master; branch temp1; reset --hard HEAD^^ (2 is just arbitrary); checkout -b release; $EDITOR something; commit -a; checkout master; merge --no-ff release; rebase -p temp1	10:47
	that should basically boil down exactly what I'm seeing	10:47
	→ fourcolors joined	10:48
fourcolors	hi, I have a git repository and I want to remove one of my remote branches. how can I do this?	10:50
wereHamster	fourcolors: branch -r -d ...	10:51
fourcolors	wereHamster: oh cool thansk	10:51
	thanks*	10:51
	← fourcolors left	10:52
	→ zaas joined	10:52
zaas	hi	10:52
	← MattDiPasquale left	10:53
	→ MattDiPasquale joined	10:53
zaas	is it common practise to push git to a test environment and from there into production environment, meaning that the git dirs are accessible over the web?	10:53
hachi	I just tested, I'm seeing the same behavior on OSX with 1.7.2.2	10:53
	← NET\|\|abuse left	10:54
doener	yeah, can reproduce here	10:54
	→ canavar joined	10:54
wereHamster	zaas: it's possible, but I don't recommend it. Git is a SCM, not a deployment tool.	10:55
selckin	zaas: it's a bad idea.	10:55
	→ emacs_noob joined	10:56
	emacs_noob → rails_noob	10:56
zaas	i see. so you only checkout into your local webroot and deployment goes without git. (any suggestion on deployment tools other then sftp?)	10:56
	→ Dave^\| joined	10:57
wereHamster	zaas: depends on the language/web framework	10:57
zaas	zend php	10:57
wereHamster	zaas: http://stackoverflow.com/questions/321502/what-tools-languages-do-you-use-for-php-web-application-deployment	10:57
	→ rolfb joined	10:58
	← khmarbaise left	10:59
	→ jinkochino joined	11:00
	rails_noob → ruby_noob	11:00
	ruby_noob → rails_noob	11:00
hachi	doener: do you have any advice? should I report this as a possible bug or anything?	11:00
zaas	thanks, will look into that. this should ease working on new features to test incremental building. kinda new to scm and want to be able to test features, roll back and deploy to live when happy	11:01
doener	hachi: yeah, at least report it, providing a testcase for the testsuite would be cool	11:01
zaas	i guess git + phing might to the job	11:01
	→ _ramo joined	11:01
	← _ramo left	11:01
	← Digitalos left	11:01
	→ mbroeker joined	11:02
doener	hachi: doesn't happen when the merge doesn't need to be forced	11:05
	← aspotashev left	11:05
hachi	yeah, it's kinda weird	11:05
	I feel like I'm gonna get told that it does this on purpose	11:05
doener	hachi: ok, I see what's going on	11:06
hachi	oh good, I haven't sent an email yet, so that's good at least :)	11:06
doener	hachi: rebase calls "git merge" internally, and of course it doesn't pass --no-ff	11:07
	→ Marquel joined	11:07
Marquel	morning.	11:07
doener	hachi: so you get a fast-forward	11:07
hachi	this would sorta be okay, except I'm placing a tag on the release branch	11:07
Marquel	how good is git pull (or git clone) on sloppy network connections with repositories about 1.5GB in size?	11:08
hachi	Marquel: about as good as TCP is	11:08
	→ elnur joined	11:09
Marquel	hachi: my problem is - one of our subversion clients is constantly failing now on checkout of parts of the tree and i can't figure out why.	11:09
hachi	and when I do this merge, it's because I want the tag to be reachable from master, as well as the whole 'branch -d' thing, where it doesn't like you deleting branches because they aren't "fully merged"	11:09
	doener: given this, does it seem like a gap in coverage to you as well?	11:09
	→ iravid joined	11:09
doener	hachi: I currently wonder why the commits on "release" get rewritten at all. I thought that was fixed a few releases ago	11:10
hachi	doener: oh?	11:10
	← Saur left	11:11
hachi	Marquel: when the network connection fails you may need to cd into the new repo dir and pick up again with 'git fetch'	11:11
	→ Saur joined	11:11
	← stliu left	11:11
hachi	the whole thing is more resiliant than SVN was, because checkouts happen from the local repo after the updates have happened	11:11
	so bascially, nothing you actually touch changes until all the network traffic is done	11:12
	SVN is	11:12
	(can't put SVN in past tense in general, very unfair of me :)	11:12
	← bremner left	11:12
	← MattDiPasquale left	11:12
	→ razvand joined	11:13
	→ MattDiPasquale joined	11:13
Marquel	hachi: well, this is an automated build box, which is not touched by anybody unless something goes wrong.	11:13
	← Rasherz left	11:13
hachi	doener: if I may inquire, what do you mean by you thought this was fixed... did they rewrite part of the merging or rebasing that you thought would take care of this?	11:13
	→ bremner joined	11:13
Marquel	hachi: so its task will be to do a git pull every night, run some build-tasks and be done with it.	11:13
	← rails_noob left	11:13
hachi	pull is probably easier	11:14
	I bet if it fails it just returns status != 0	11:14
	← jinkochino left	11:14
hachi	so you can just retry a few times on failure	11:14
Marquel	hachi: now it does a svn checkout and is stuck.	11:14
doener	hachi: someone took the time to fixup -p (or at least that was my impression, I used to avoid -p)	11:14
	→ Dave^\|\| joined	11:15
hachi	Marquel: are you talking about native git, or git-svn ?	11:15
	← Dave^\| left	11:15
Marquel	hachi: if i do a switch it will be native git. - if i can persuade my boss to do the switch	11:15
hachi	in both cases, I'm pretty sure you could just retry till status 0, with some reasonable backoff	11:15
Marquel	hachi: i don't know if cruise control .net will support that.	11:16
	→ daethorian joined	11:16
Marquel	me best hope is that git pull does not fail that fast...	11:16
hachi	then to be honest, your problems lie in a different realm from your version control	11:16
	maybe you should convice your boss to fix the network	11:17
Marquel	hachi: interestingly it worked flawlessly a few days ago...	11:17
hachi	doener: I like the idea of -p... I could possibly just emulate it too anyways	11:18
	the liklihood of merge conflicts on this is so low that I could just do	11:18
	pull, reset --hard origin/master, merge release, push	11:19
doener	s/pull/fetch/	11:19
hachi	indeed	11:19
	wait... no	11:19
	→ ExtraSpice joined	11:19
hachi	that would double merge	11:19
	← Saur left	11:19
hachi	but whatever, implementation details	11:19
	→ ForceBlast joined	11:20
doener	hachi: err, fetch doesn't merge, pull does, and reset --hard just kills that merge again	11:20
	← zaas left	11:20
hachi	pull destroys the old merge	11:20
	← jatt left	11:20
	← galderz left	11:20
wereHamster	pull doesn't destroy anything	11:20
hachi	if I fetch, the old merge wouldn't go away... oh wait, the reset would	11:20
	yeah	11:20
doener	hachi: ah, rebase pull...	11:20
	→ Saur joined	11:21
hachi	should	11:21
	git pull -p	11:21
	with rebase=true	11:21
	do the right thing, or am I being a little pushy in my expectations :D	11:21
doener	pushy	11:21
hachi	keke	11:21
	I saw that you can pass options to git-merge via pull, but it didn't say anything about rebase options with you pass --rebase	11:22
doener	but if that's the whole workflow, I'd just skip the "master" branch... git fetch origin; git checkout origin/master; git merge release; git push origin HEAD:master	11:22
hachi	I am boiling it down further than the actual use case	11:22
doener	hachi: there used to be a, uhm, quirk that allowed "git pull -S -p" to work, but... waves hand I never told you about any quirks	11:23
hachi	when you get down to it, I'm worried about this problem because people who are not me need to be able to do it	11:23
	→ khmarbaise joined	11:23
hachi	since I understand the problem it's a bit easier to work around... it's really hard for me to teach someone how to work around a problem until they encounter it	11:23
doener	-p is simply hard to get right	11:24
	← davidw left	11:24
hachi	my favorite people to train are the ones who lie about their being a problem... one of my coworkers blows away his git checkouts when something goes wrong	11:24
	like... a botched merge	11:24
	→ galderz joined	11:24
doener	hachi: btw: http://www.kerneltrap.com/mailarchive/git/2007/10/31/371510	11:25
	hachi: that might help in your case. But the whole --first-parent thing isn't _that_ meaningful (it depends on your working habits to "fit"), and I wasn't scratching my own itch, so I never got back to that patch	11:27
hachi	mm	11:27
doener	hachi: "git rebase --first-parent -p" would never cherry-pick the commits on "release", thus the merge wouldn't be a fast-forward	11:28
	→ psankar joined	11:28
hachi	I assume --first-parent is saying, only rewrite the first parent on these commits....	11:29
	cause they don't show in the example which is the first parent, and which is the second....	11:29
doener	hachi: "they" would be me	11:29
hachi	O_O	11:30
	ok	11:30
	→ fedesilva joined	11:30
	→ igstan joined	11:30
hachi	well, thanks for having written this then	11:30
	I saw the name, and assumed it wasn't you because of your handle, sorry bout that	11:30
doener	hachi: well, that was in 2007, and I never pushed it far enough to get into git.git	11:30
hachi	oh, well.... heh	11:31
doener	hachi: no problem, my ego is happy as is ;-)	11:31
	→ bauruine joined	11:31
doener	hachi: if you do "git rebase -i -p temp1" in your example, you'll two commits in the todo list, the one on "release" and the merge commit	11:31
	hachi: so rebase cherry-picks the commit from release, and then merges it (no-op, obviously)	11:32
	→ galderz_ joined	11:32
	← galderz_ left	11:32
	→ galderz_ joined	11:32
hachi	ahh	11:32
doener	hachi: err, fast-forward actually, the branch is picked on its own, anyway, you get the idea	11:32
	← galderz left	11:33
	galderz_ → galderz	11:33
doener	hachi: with the --first-parent thing, rebase would only put the merge commit into the todo first (it only walks the first-parent ancestry)	11:33
	hachi: so the merge would no longer try to merge the cherry-picked version of the commit on release, but the original commit	11:34
	hachi: basically, "rebase -p" doesn't recreate merges with the same second parent, but rebases the whole "subgraph", so the result just has the same structure	11:35
	doener hopes to make some sense, is ill, brain is slow	11:35
hachi	you are making sense	11:35
	I'm pretty sure I fully understand what you're saying	11:36
doener	good :-)	11:36
hachi	the patch for that is missing though	11:36
	alin-gym → alin-lunch	11:37
	← bilkulbekar left	11:37
doener	the --first-parent thing would make it rebase only the history it finds by walking the first parent ancestry (i.e. not the whole subgraph), and merges are recreated with the original secondary parent	11:37
hachi	I like it	11:38
	it took me a moment to realize that was the implementation, that's what I was saying before	11:38
doener	hachi: http://www.kerneltrap.com/mailarchive/git/2007/10/31/371897 -- there's some more description stuff in there	11:38
hachi	'only rewrite the first parent'	11:38
doener	hachi: "the first parent" sounded like a single commit to me ;-)	11:39
hachi	doener: I'm looking for the patch itself though	11:39
	I mean, I assume it exists	11:39
doener	hachi: yeah, there was some encoding problem that broke the archive it seems	11:39
	hachi: I still have it here	11:39
	← Radar left	11:39
hachi	oop, I found it on list-archives.org	11:39
	elegant and sweet	11:39
doener	ok, trying to rebase that gives ugly conflicts...	11:40
	→ davidw joined	11:41
	davidw → Guest58902	11:42
	← whitman left	11:43
hachi	so this patch never got merged in, but it spawned an enormous conversation about encoding issues	11:43
doener	I'd attribute that to my name, it happened again with other patches ;)	11:44
	(though those discussions weren't as big)	11:44
hachi	heh	11:44
	→ simplechat_ joined	11:45
	→ artefon joined	11:45
	→ galderz_ joined	11:47
	← galderz_ left	11:47
	→ galderz_ joined	11:47
	← galderz left	11:47
	galderz_ → galderz	11:47
	Guest58902 → davidw	11:47
	← davidw left	11:47
	→ davidw joined	11:47
	← psankar left	11:48
	→ roop joined	11:48
	← galderz left	11:51
	← voxpelli left	11:53
	→ kassens joined	11:54
	← ForceBlast left	11:54
	zz_rubydiamond → rubydiamond	11:54
	→ rocket joined	11:56
	→ cannonball joined	11:57
rocket	anyone else here experience issues with git-svn taking a huge amount of ram? seems there is a read-tree operations consuming tons .. anything I can do to limit the amount of ram its using?	11:57
doener	hachi: I guess the patch won't work anymore, at least not in the compact form it used to be in. Seems that -p now makes more assumptions about the commits being present in the todo list	11:58
	rocket: git version?	11:58
rocket	1.7.0.3	11:58
	→ d0k joined	12:00
	← lucasvo_2 left	12:01
	→ kpreid joined	12:01
rocket	doener: it seems to consume almost all the ram on this box .. eventually it will get some better	12:01
hachi	doener: ugh, so now I'm so far down with understanding the problem, I've forgotten the condition the -p flag works fine in	12:01
doener	hachi: AFAIK, when all the original merge commits happen to have their second (third, fourth, ...) parent being a descendant of the <newbase> (which defaults to <upstream>)	12:03
	→ aziz_ joined	12:03
doener	s/AFAIK/AFAICT/	12:03
hachi	yeah, I'm basically trying to replay my original thought process again now :)	12:04
	← canavar left	12:04
	→ corni joined	12:04
doener	hachi: ok, my head doesn't work well enough to get this sorted out right now.	12:05
hachi	thanks a ton	12:06
	honest	12:06
doener	hachi: willing to dive into the rebase--interactive script?	12:06
	hachi: I could give you some pointers, but no solution...	12:06
	→ lucasvo_2 joined	12:07
hachi	I agree with junio that it's odd to give special treatment to first parent... or are there precedents for that these days?	12:07
	it's been a while since he had that opinion	12:07
	→ Yuuhi joined	12:07
hachi	defacto use gives it somewhat special meaning	12:07
doener	I don't think so, except for the --first-parent option of rev-list, which inspired my patch	12:07
	but as I said, it depends on the workflow to "fit in"	12:08
	any criss-cross merging and it won't work anymore	12:08
	the original patch was made with git-svn in mind, and that even happens to use --first-parent internally to find its upstream (already did so back then)	12:09
	→ hohoho joined	12:09
doener	but in general, I also agree with Junio, it's kinda weird	12:09
	hachi: ok, so for a more general (but manual) solution:	12:10
	→ rasztasd joined	12:10
doener	hachi: around line 922 there's "Watch for commits that been dropped by --cherry-pick", that writes into the $DROPPED directory	12:10
	hachi: that would need to be adjusted to record all commits dropped from the todo list	12:11
	→ voxpelli joined	12:11
hachi	I actually just attempted the broken behavior using rebase -p -i origin/master	12:11
	sadly that doesn't do the right thing, I thought there was a slim chance it might	12:12
	doener: reading that section now	12:12
	I keep writing git hacks in shell, I didn't realize how useful that habit of mine would be in hacking on git	12:13
	→ aresnick joined	12:14
doener	hachi: hm, actually, no, I think you just need an extra pass that removes the files for commits not in the todo list from the $REWRITTEN directory	12:15
	hachi: I misread the code that uses the entries in those directories...	12:15
	← kpreid left	12:16
	→ xiong joined	12:16
doener	hachi: if you do that, you should be able to keep only the merge commit in the todo list (manually dropping the commit from "release")	12:16
hachi	you mean firing it up in -pi, and 'manually dropping' by deleting the line in the editor?	12:17
doener	hachi: yup. That's step one (and it works, just tried manually). Step two would be to add an option that can automatically drop the merged-in commits (i.e. basically --first-parent)	12:18
	hachi: http://git.pastebin.com/tanuZueg	12:19
	→ tvw joined	12:20
hachi	you tested with the inverse of the motion I'm trying to make, not that this matters I guess	12:21
	← rasztasd left	12:21
	→ byang joined	12:21
doener	hachi: hm? I just added more commits	12:21
	→ oriba joined	12:22
hachi	oh, I'm forgetting that temp1 isn't necesarally earlier in the history, because this is not cronological order	12:22
	still, my earlier sentiment of 'not that this matters' makes a lot of sense	12:23
	rebasing in any direction you want works fine, why shouldn't rebasing merges work too	12:23
	→ galderz joined	12:23
doener	hachi: http://git.pastebin.com/XGkQCk4J	12:24
	→ jetienne joined	12:24
	← artefon left	12:24
jetienne	q. silly question, how to create a tag which gonna be pushed in a remote dir. i tried 'git tag -a supertag' then git push, but it doesnt push anything	12:25
	→ apol joined	12:25
doener	hachi: that should be exactly the example you gave (even including the --no-ff requirement, which happens to work here, too, but there might be some more corner cases for that...)	12:25
	jetienne: git push <remote> supertag	12:25
	jetienne: push doesn't push tags unless you explicitly ask it to	12:25
jetienne	doener: ok thanks	12:25
	← corni left	12:26
	→ HG` joined	12:27
	← sonnym left	12:28
	← shiba_yu_ left	12:28
	→ akahn joined	12:29
	→ flazz joined	12:29
	← thedjpetersen left	12:30
	→ IslandUsurper joined	12:32
	← voxpelli left	12:33
	→ voxpelli joined	12:33
	→ ciampix joined	12:33
	← akahn left	12:34
	← doener left	12:40
	← voxpelli left	12:41
	→ voxpelli joined	12:42
	← kassens left	12:42
	→ beni- joined	12:44
	← s0enke left	12:46
	→ s0enke joined	12:46
	← voxpelli left	12:49
	→ voxpelli joined	12:49
	→ richardbronosky joined	12:50
	→ rayners joined	12:50
	→ davidw_ joined	12:50
	→ corni joined	12:51
	← davidw left	12:51
	← xiong left	12:51
	← picardo12 left	12:51
	← rgr left	12:54
	← apsysoev1 left	12:54
	→ rgr joined	12:54
	→ fod joined	12:54
	→ sivy joined	12:54
	← airborn left	12:55
	→ jatt joined	12:57
	← Shurakai left	12:57
	← freimatz left	12:58
	→ rgr_ joined	12:59
	→ Shurakai joined	12:59
	→ skoop joined	13:00
	→ scarabx joined	13:01
	← TTilus left	13:01
	davidw_ → davidw	13:01
	← davidw left	13:01
	→ davidw joined	13:01
	→ copelco joined	13:02
	→ kukks joined	13:03
	→ redbaritone joined	13:03
	← cash left	13:03
	← mbroeker left	13:05
	→ hedgehog83 joined	13:05
hedgehog83	hi, git send-email just hangs and doesn't do anything. how do i get it to be more verbose so i learn what's going wrong?	13:06
jaeckel	GIT_TRACE=1 git send-email....	13:06
	← yrlnry left	13:07
	← orafu left	13:08
	→ orafu joined	13:09
	← nvie left	13:09
	→ dpratt71 joined	13:09
	→ des joined	13:09
	← segher left	13:10
	← dpino left	13:10
rolfb	ooooh	13:10
	GIT_TRACE	13:10
	← rgr_ left	13:10
rolfb	didn't know	13:10
	thanks	13:10
	→ Danielpk joined	13:11
Danielpk	A noob question...	13:11
	→ segher joined	13:11
Danielpk	what is diff between branch and tag?	13:11
des	hi, i'd like to know how to reference a repository after a givven commit (something similar to svn's repo:foo rev:rXXX) which hash should I use? commit? tree? parent? all of them?	13:11
wereHamster	Danielpk: when you checkout a tag git detaches HEAD	13:11
	→ nvie joined	13:12
	→ kylehayes joined	13:12
wereHamster	des: you reference a repository by its url	13:12
des	I mean in a phrase like 'I imported this feature from git repo X at <whatIwantHere>'	13:12
	→ mkiwala joined	13:12
wereHamster	hash, ref, whatever you want	13:13
	something that identifies what you imported	13:13
Danielpk	wereHamster, humm in real world, what it does means?	13:13
	← froschi left	13:13
	→ hugoxrosa joined	13:13
	← sgh left	13:14
wereHamster	Danielpk: see tags as unmovable branches. A tag you create and it will always point to the same object. Branches move along as you make commits on them	13:14
	← rgr left	13:14
	→ rgr joined	13:15
dpratt71	cloned a project the other day; made some trivial changes; tried to discard my changes (via http://is.gd/errRl)...	13:15
	SvenDowideit → SvenDowide	13:15
des	wereHamster: I imported some code made by various commits in the git repo, but I imported it today, I want to be able to come back to this repo as I see it today. I hope that makes my question more clear	13:15
	SvenDowide → SvenDowideit	13:15
dpratt71	didn't seem to work	13:15
	← SvenDowideit left	13:15
	→ SvenDowideit joined	13:15
dpratt71	first question, how do I discard 'unstaged changes'?	13:15
wereHamster	des: make a tag ?	13:15
des	wereHamster: thanks for answering, btw	13:15
rgr	checkout	13:15
wereHamster	dpratt71: unstagedh changes: checkout -- .	13:16
hedgehog83	git send-email still hangs and the trace wasn't helpful. Now what?!	13:16
Danielpk	wereHamster, tags dont come with old commits and logs? is it deploy only?	13:16
dpratt71	ok, that may lead to my second question:	13:16
des	wereHamster: the repo is not mine	13:16
wereHamster	des: but you cloned it, right?	13:16
	you have it on your local computer	13:16
	→ yrlnry joined	13:16
wereHamster	Danielpk: how dyou mean 'deploy only'?	13:17
dpratt71	no...actually, it doesn't...I did 'git checkout .' but there still appeared to be unstaged changes	13:17
wereHamster	dpratt71: git status	13:17
des	wereHamster: no, actually I imported the code into a svn tree, and I just want to reference in the commit message where that code come from	13:17
	ip^sleepz → ippa	13:17
dpratt71	wereHamster: lists a bunch of files	13:18
wereHamster	des: you imported code from git into a svn repo, correct?	13:18
	← nvie left	13:18
des	wereHamster: correct	13:18
wereHamster	des: then put the commit/tree which you imported into the svn commit message	13:18
mkiwala	dpratt71: does your git status list files that have been checked into git previously? Or are these new files?	13:19
wereHamster	dpratt71: under which section? Changed but not updated or Changes to be committed ?	13:19
dpratt71	mkiwala: status of modified, so the former, I think	13:19
	← ippa left	13:20
	→ tryrty joined	13:20
dpratt71	wereHamster: changed but not updated	13:20
wereHamster	dpratt71: windows?	13:20
dpratt71	wereHamster: heh, yes	13:20
hedgehog83	so, git-send-email asks me plenty questions, and then it hangs after the last one, rather than send the email. help!	13:20
	← tryrty left	13:20
dpratt71	wereHamster: known issue?	13:20
des	wereHamster: oh, I see, the commit/tree hash is the same. I messed myself up, thanks!	13:21
wereHamster	dpratt71: most likely line endings problems	13:21
dpratt71	wereHamster: ok; fixable?	13:21
wereHamster	des: commit and tree hashes can not be the same	13:21
	dpratt71: I think so. But I'm not using windows	13:22
dpratt71	wereHamster: ok, I'll research, thanks	13:22
	→ nvie joined	13:23
	← nvie left	13:23
des	wereHamster: then maybe is github confusing me because the 'tree' hash liks links to /tree/{commit hash}, although the text for the <a> tag is different as you say	13:24
jaeckel	dpratt71: you've enabled autocrlf?	13:24
	→ Ryback_ joined	13:24
wereHamster	des: each commit contains the tree it references	13:26
	so given a commit you can always extract the tree	13:26
	→ JeffJohnson joined	13:27
JeffJohnson	howdy	13:27
	← ryanakca left	13:27
	→ jfkw joined	13:27
JeffJohnson	how can I refetch the remote branches when using git-svn? I deleted an remote branch locally but can't readd it to the git branch list with git svn fetch	13:28
	← apol left	13:28
	← rgr left	13:28
	← des left	13:28
	→ des joined	13:29
	→ tizbac joined	13:31
	← rchavik left	13:32
dpratt71	jaeckel: I'm assuming that the setting is the default value (I don't recall changing it); how do I check?	13:33
	→ kpreid joined	13:36
	← kpreid left	13:36
	→ RORgasm joined	13:36
	← des left	13:37
dpratt71	$ git config --get core.autocrlf	13:38
	true	13:38
	is that as it should be?	13:38
	→ frakturfreak__ joined	13:38
	frakturfreak__ → frakturfreak	13:39
	→ nvie joined	13:40
\|albel727\|	JeffJohnson: just out of curiosity, what was the method you used to delete remote branch?	13:40
	→ whaley joined	13:41
	→ unreal_ joined	13:41
	→ nothingHappens joined	13:41
rocket	doener: it seems to consume almost all the ram on this box .. eventually it will get some better	13:41
	anyone else here experience issues with git-svn taking a huge amount of ram? seems there is a read-tree operations consuming tons .. anything I can do to limit the amount of ram its using?	13:42
JeffJohnson	\|albel727\|: git branch -D -r xyz	13:42
	← scarabx left	13:42
JeffJohnson	\|albel727\|: and its still in the svn branches dir :)	13:42
	→ rgr joined	13:42
	→ SimonP86 joined	13:42
	← ciampix left	13:43
	← jony left	13:43
\|albel727\|	oh, the -r option too... that's gonna be a hard case. =) let me think for a while...	13:43
	← unreal_ left	13:44
JeffJohnson	\|albel727\|: im allready reclone it, cause i know no other solution, but would be nice if there is another way..:)	13:44
	← unreal left	13:44
jaeckel	dpratt71: that depends on what you want :)	13:45
	I'm always using the install setting 'commit as is/checkout as is'	13:45
\|albel727\|	Well, it's good to know that this is not a pressing matter. =) I would have sought of using some form of "git svn reset" or "git remote add".	13:46
jaeckel	otherwise there are magic conversions of line endings that are sometimes really confusing	13:46
	← HG` left	13:46
	→ unreal joined	13:46
jaeckel	and I haven't understood yet who needs that feature...	13:47
	← erlnoob left	13:47
	→ ryanakca joined	13:47
	→ unreal_ joined	13:47
	→ pfig joined	13:48
JeffJohnson	\|albel727\|: reset dont helps, remote add i never used before :)	13:48
\|albel727\|	JeffJohnson: yeah, I had my doubts about "svn reset" too. =) Now I'm looking at what can be done with remote add...	13:49
	→ SlickT10 joined	13:49
nanotube	\|albel727\|: thanks for the info on keyword expansion - looks like it's just the ticket. i'll give it a try :)	13:50
	← unreal left	13:50
	← copelco left	13:50
\|albel727\|	nanotube: np =)	13:51
	→ copelco joined	13:51
	→ fowlduck joined	13:51
nanotube	\|albel727\|: i'll be coming to you if i run into questions... so watch out! :D	13:51
dpratt71	jaeckel: well, I turned it off for this repository, started over, and the bogus changes are gone	13:51
	thanks for the help, all	13:52
\|albel727\|	oh no! I better run for my dear free time!! =))	13:52
nanotube	haha indeed ;)	13:52
	→ montylounge joined	13:53
	← SlickT10 left	13:53
	← Torsten left	13:53
	→ mfischer joined	13:53
	← madduck left	13:53
\|albel727\|	JeffJohnson: just out of curiosity, do you still have your broken repository?	13:54
	→ madduck joined	13:54
	← madduck left	13:54
	→ madduck joined	13:54
hedgehog83	for me, git send-email is still broken, and i don't know how to debug it. —smtp-debug doesn't do anything.	13:54
	→ bushwakko_ joined	13:55
JeffJohnson	\|albel727\|: no, but i could remove an branch that i dont use and try to refetch it:)	13:56
	→ artefon joined	13:57
\|albel727\|	JeffJohnson: that would be a risky affair, so I guess unless you're very enthusiastic... =)	13:58
	JeffJohnson: as for now, I'm just interested about, whether git reflog would have returned you the latest commit on your missing branch.	13:58
wereHamster	\|albel727\|: no, reflog is deleted along with the branch	13:59
JeffJohnson	\|albel727\|: problem is that after removing an branch in the svn, git-svn still will add it to the svn branch list, so i try to delete a few old ones everytime, but sometimes I inadvertently delete a wrong branch :D	13:59
	← bushwakko left	13:59
	← hedgehog83 left	14:00
	← bushwakko_ left	14:00
	← roop left	14:00
JeffJohnson	\|albel727\|: how i display reflog for the removed branch?	14:01
\|albel727\|	wereHamster: really? damn, that's just the worst possible behaviour imaginable. just out of curiosity, where a reflog for a branch is usually physically kept?	14:01
wereHamster	JeffJohnson: there is no reflog, it was removed along with the branch	14:01
	\|albel727\|: .git/logs	14:01
	← trivol left	14:02
JeffJohnson	ah thx wereHamster :)	14:02
	→ akahn joined	14:02
mfischer	is there a convenient way to specifiy (in .git/config) different "pull" vs "push" remotes?	14:03
	→ pigdude joined	14:03
	→ mar77i joined	14:03
wereHamster	mfischer: remote.<name>.pushurl	14:03
	← corni left	14:04
	← khmarbaise left	14:04
jaeckel	dpratt71: I recommend you to turn it off globally	14:05
	and if you really need it you can still turn it on in the repositories that need it	14:06
yrlnry	Is there an easy way to get git to write one patch file for each commit on a branch between A and B?	14:06
mfischer	wereHamster, thanks, seems to work...	14:06
\|albel727\|	JeffJohnson: yeah, it all looks just like wereHamster have said. let me think for a while of another ways to restore needed commit sha...	14:06
	← StrongOrder left	14:06
wereHamster	yrlnry: git format-patch	14:06
jaeckel	dpratt71: but do it by a 'reinstall' of msysgit, since there is still some magic done in the installer...	14:07
yrlnry	thanks.	14:07
wereHamster	\|albel727\|: when you delete a branch, git will prin tthe sha of the branch you just deleted. So if you just made the mistake it's easy to restore the branch	14:07
yrlnry	I did for h in `git log --format="%H" 1.001004..HEAD`; do git show $h > patch.$h; done	14:07
	→ icwiener joined	14:08
yrlnry	But git-format-patch seems like just the thing. Thanks again.	14:08
wereHamster	yrlnry: congratulatinos, you just (poorly) reinvented format-patch	14:08
	→ frakturfreak_ joined	14:09
yrlnry	Thanks!	14:09
\|albel727\|	wereHamster: Wow, cool. I've never paid attention to that before. Just to be sure, you're not aware of any other way to recover the sha in such situation?	14:09
yrlnry	My poor solution has a serious problem: It doesn't keep the order of the patches.	14:09
	← frakturfreak_ left	14:10
	← igstan left	14:10
	→ ikrabbe joined	14:10
wereHamster	\|albel727\|: no	14:11
\|albel727\|	thought so. real pity =/	14:11
mfischer	yrlnry, long time no see	14:11
yrlnry	Hi! It sure is.	14:11
	What have you been doing with yourself since 2001?	14:12
ikrabbe	Hi, maybe someone can help, as I don't use windows that often: I installed git on cygwin (1.7.1) today and checked out a repository. All files seem to have wrong line endings and some have changed permissions 00644->00755 after clone. so that I cannot use that working copy to work with the origin again.	14:13
	← rraasch left	14:13
	→ chris_n`` joined	14:13
	chris_n`` → chris_n	14:13
ikrabbe	I haven't changed anything else but just installed a plain cygwin with git as it is packaged today	14:14
	← ixti left	14:14
	→ roop joined	14:14
ikrabbe	actually msysgit (some 1.7.0.. Version) does the right thing when choosing (leave files as-is).	14:15
	but I would better like using the cygwin version	14:16
	My old cygwin installation did it the right way, so I'm quite confused	14:16
jaeckel	core.autocrlf	14:18
	ikrabbe: ^^	14:18
\|albel727\|	ikrabbe: what does git config core.autocrlf...	14:18
	ikrabbe: yeah, jaeckel was faster =)	14:19
jaeckel	ikrabbe: and core.filemode	14:19
	→ ncopa joined	14:19
ncopa	hi	14:19
	Delta compression using up to 8 threads.	14:19
	error: pack-objects died of signal 11	14:19
	← alin-lunch left	14:19
ncopa	i got this with git-1.7.2.1	14:19
	upgraded to 1.7.2.2	14:19
jaeckel	ikrabbe: they should both be false	14:20
	← skoop left	14:20
ikrabbe	jaeckel and \|albel727\|, yes I though about these myselve and I will cross check again, but I haven't configured them at all, just use the plain installation... Problem is that it's not my PC... I will check in some minutes	14:20
jaeckel	ikrabbe: if it's not your PC you can still edit the settings of the repo	14:20
\|albel727\|	jaeckel: are you sure about that? I'd have thought about true for autocrlf, at least. where I'm wrong?	14:21
jaeckel	ikrabbe: if you won't mess up the users' config	14:21
	→ brntbeer joined	14:21
jaeckel	\|albel727\|: when you set autocrlf to true it will change all unix to win line endings	14:21
	zorzar_away → zorzar	14:22
\|albel727\|	jaeckel: yep. In the working copy. and turn them back to unix on commit.	14:22
	jaeckel: looks like we have to clarify, what does he consider a "wrong line ending"...	14:22
jaeckel	\|albel727\|: but that's his problem "... files seem to have wrong line endings	14:22
	and some have changed permissions 00644->00755 after clone"	14:22
	robinbowes → robinbowes_on_ho	14:23
	robinbowes_on_ho → robin_on_holiday	14:23
\|albel727\|	jaeckel: It's just that I'd have thought, that "wrong file ending" on windows means "unix file ending" =)	14:24
	*line	14:24
ikrabbe	jaeckel my main problem is that after checkout all files are marked changed	14:24
	after clone	14:24
jaeckel	\|albel727\|: even visual studio can handle unix file handling ;)	14:24
\|albel727\|	jaeckel: hmm, never actually tried =) what, automatically?	14:25
ikrabbe	\|albel727\|: wrong file ending is windows file ending, everywhere	14:25
\|albel727\|	ikrabbe: lol	14:25
	ndimitrij → ndim	14:25
jaeckel	ikrabbe: ok, then check these two settings git config --get core.filemode; git config --get core.autocrlf	14:25
	→ ajpiano joined	14:25
	→ spearce joined	14:26
jaeckel	\|albel727\|: yeah, for sure - it even notices if there are mixed file endings AND lets you choose to convert to either unix or windows =)	14:26
ikrabbe	eclipse manages even to mix file endings in single files, when you use mac, windows and linux you get a real mess of files	14:26
	try to run diff on them ;)	14:26
jaeckel	hrhr	14:26
	← mar77i left	14:26
	→ stringoO joined	14:26
ikrabbe	and it does that silently	14:27
\|albel727\|	ikrabbe: well, what can I say. I guess the times of me babysitting studio and always feeding it consistent endings are over =).	14:27
	→ steph021_mac joined	14:27
\|albel727\|	oh sorry	14:27
pigdude	I'm just beginning to use branches now...if I create a branch, what do I need to do to push it?	14:27
	← Mpenz left	14:27
\|albel727\|	i meant not ikrabbe, but jaeckel =)	14:27
	← steph021_mac left	14:27
	← vu3rdd left	14:28
jaeckel	\|albel727\|: I wished, mine too =).	14:28
	rubydiamond → zz_rubydiamond	14:28
jaeckel	\|albel727\|: but when you ever worked with Mentor Graphics Tools, you can say that VisualStudio is a real IDE ;)	14:29
	→ Mpenz joined	14:29
	← byang left	14:29
ikrabbe	the only ide worth thinking about is vim	14:29
jaeckel	and eclipse is the allmighty solution for everything...	14:29
ikrabbe	and ED is the UNIX Editor !	14:30
	→ alin-gym joined	14:30
	← Deesl left	14:32
yrlnry	pigdude: git push remote branch	14:33
	→ leo2007 joined	14:33
yrlnry	pigdude: "remote" is the name of the remote, typically "origin". "branch" is the name of your branch.	14:33
	← marc_kdab_ left	14:33
pigdude	yrlnry: so if I just did `git checkout -b some` then I'd do `git push remote some`?	14:33
yrlnry	yep.	14:33
pigdude	thanks yrlnry	14:33
yrlnry	you're welcome!	14:33
leo2007	hello, during rebase I have one commit skipped. How to find that commit again?	14:34
	→ g0bl1n joined	14:34
	← jatt left	14:34
leo2007	I only have the first line of the commit but couldn't find it in reflog.	14:34
yrlnry	leo2007: after the rebase, you can use ORIG_HEAD to look at the old, unrebased branch.	14:34
	leo2007: try git log ORIG_HEAD	14:34
\|albel727\|	jaeckel: zomg. Mentor Graphics Tools? The ones for circuit design? Never used them but screenshots I see look really horrible =)	14:35
yrlnry	Or if it is an ancestor of something you can find in the reflog, say at HEAD@{17}, then git log HEAD@{17} will show you the ancestor commits of that, including the one you want.	14:36
	→ chrislerum joined	14:36
leo2007	yrlnry: no, it is not there.	14:36
	← dr_win left	14:36
ikrabbe	ok, : git on cygwin... (mkdir test; cd test; git init; git config --get-all) show filemode=true and ignorecase=true, which both was a problem	14:36
chrislerum	http://gist.github.com/540444 educating needed - how can i get these commits back?	14:36
	→ Deesl joined	14:36
	→ eletuchy_ joined	14:36
	zz_rubydiamond → rubydiamond	14:37
	→ Torsten joined	14:37
	← montylounge left	14:37
	← [Po]lentino left	14:37
ikrabbe	I meant (git config -l) ofc	14:37
	← Deesl left	14:37
yrlnry	leo2007: did you look at the reflog for the rebased branch itself?	14:37
	→ dr_win joined	14:37
leo2007	git log ORIG_HEAD only shows upto the commit that is one ahead of omitted commit.	14:37
\|albel727\|	ikrabbe: and what about global settings?	14:37
yrlnry	say you rebased 'topic'; git reflog show topic	14:38
leo2007	yrlnry: I did it in magit.	14:38
ikrabbe	there where no global settings	14:38
leo2007	an emacs frontend for git.	14:38
\|albel727\|	ikrabbe: then what about system ones?	14:38
yrlnry	what do you mean "only shows upto...". It should show all the way back to the beginning of the repository.	14:38
	← eletuchy left	14:39
ikrabbe	hmm, might be worth checking	14:39
	→ bodelaire joined	14:39
wereHamster	chrislerum: which commits?	14:40
	← Marquel left	14:40
leo2007	yrlnry: this is the msg printed when conflict happens http://imagebin.org/110532	14:40
	→ tedoc2000 joined	14:40
chrislerum	wereHamster: all 4 - i'm reading about this on gitready.com, looks like i'll have my answer momentito ...	14:41
	→ kumbayo joined	14:41
wereHamster	chrislerum: you mean HEAD@{2-5}	14:41
chrislerum	yes, those 4 commits	14:42
leo2007	yrlnry: the output of git log ORIG_HEAD looks like this http://paste.pocoo.org/show/252435	14:42
wereHamster	chrislerum: and where do you want to have those commits?	14:42
ikrabbe	chrislerum, also check your "working" branch as "checkout: moving from master to master" is suspect	14:42
chrislerum	ikrabbe: ja that doesn't make sense to me	14:42
	← voxpelli left	14:42
	← lucasvo_2 left	14:43
chrislerum	wereHamster: where they were a second ago - on a branch called 'working'	14:43
wereHamster	chrislerum: git branch working 7bda753	14:43
	← eletuchy_ left	14:43
	→ eletuchy joined	14:43
	→ flaguy48 joined	14:44
chrislerum	wereHamster: you win the beer - thanks mucho. now to educate myself on the wherefores	14:45
leo2007	yrlnry: when the conflict happened I didn't fix it but git rebase --skip	14:45
	→ trtzrdb joined	14:45
wereHamster	leo2007: the old branch head is still in the reflog of the branch you rebased	14:45
	→ loincloth joined	14:46
	→ jinkochino joined	14:46
wereHamster	you can find that commit there	14:46
	← bodelaire left	14:46
	← trtzrdb left	14:47
	← jetienne left	14:47
leo2007	wereHamster: I can switch to branch@{1} and find the commit. but it is not in ORIG_HEAD.	14:48
	← ryanakca left	14:49
wereHamster	to?	14:49
	s/t/s/	14:49
	→ lucasvo_2 joined	14:49
	← psoo left	14:49
offby1	this sounds like one of those cases where "gitk --all" would be helpful	14:50
\|albel727\|	I concur	14:50
	← beni- left	14:50
	← krynnotaur left	14:51
leo2007	wereHamster: nothing I was just saying that because of yrlnry's earlier suggestion.	14:52
wereHamster	if branch@{1} != ORIG_HEAD, then it looks like you did something else after the rebase that changed ORIG_HEAD	14:53
	→ ryanakca joined	14:53
	← dpratt71 left	14:55
	rubydiamond → zz_rubydiamond	14:56
	→ hagabaka joined	14:56
	← hagabaka left	14:56
	→ hagabaka joined	14:56
	← vdv left	14:57
	→ Yuffster joined	14:57
	→ erlnoob joined	14:57
	← jamescarr_ left	14:59
	→ hipe joined	14:59
	→ durarara joined	15:00
	← iravid left	15:01
	→ lteo joined	15:01
	→ iravid joined	15:02
	← ajpiano left	15:04
	← Schmallon left	15:04
	→ TeckniX joined	15:05
	← bauruine left	15:06
	→ Schmallon joined	15:06
	→ apsysoev joined	15:06
	→ bhah joined	15:06
lteo	git experts.. if i committed files A, B, C.. and then realized that i should not have committed C, how do i remove C from that commit?	15:07
killerchicken_	is it the latest commit?	15:07
lteo	yes	15:08
cbreak_work	did you push?	15:08
lteo	not yet push. this is on a local branch	15:08
wereHamster	lteo: git reset HEAD^ -- C; git commit --amend	15:08
cbreak_work	then you can use commit --amend or reset or rebase -i (in increasing order of power/destructiveness/diffuculty)	15:09
	→ Martinp24 joined	15:10
Martinp24	THESE FLOODS ARE BEING DONE WITH PORT 80, 8080 AND 3128 L3 PROXIES! WHY DOES FREENODE PORTSCANNER SUCK SO MUCH? WILL THEY EVER FIX IT? DO THEY EVEN CARE? Martinp24 bhah apsysoev Schmallon TeckniX iravid lteo durarara hipe erlnoob Yuffster hagabaka ryanakca lucasvo_2 jinkochino loincloth flaguy48 eletuchy kumbayo tedoc2000 dr_win Torsten chrislerum g0bl1n leo2007 alin-gym Mpenz stringoO spearce brntbeer ncopa roop chris_n ikrabbe icwiener pigdude akahn artef	15:10
	THESE FLOODS ARE BEING DONE WITH PORT 80, 8080 AND 3128 L3 PROXIES! WHY DOES FREENODE PORTSCANNER SUCK SO MUCH? WILL THEY EVER FIX IT? DO THEY EVEN CARE? hugoxrosa mkiwala kylehayes segher Danielpk orafu redbaritone kukks Shurakai sivy fod davidw rayners richardbronosky s0enke IslandUsurper flazz galderz oriba tvw aresnick hohoho Yuuhi aziz_ d0k cannonball rocket simplechat_ fedesilva Saur ExtraSpice daethorian Dave^\|\| bremner MattDiPasquale razvand elnur r	15:10
	THESE FLOODS ARE BEING DONE WITH PORT 80, 8080 AND 3128 L3 PROXIES! WHY DOES FREENODE PORTSCANNER SUCK SO MUCH? WILL THEY EVER FIX IT? DO THEY EVEN CARE? bluenovember lemonchicken curtana bentob0x Titosemi tehbaut seanius masterkorp ribasushi sigmonsay sid3k` workmad3 mikeric tatsuya parasti fahadsadah pdignan nicoulaj engla tobiassjosten jjuran nuoHep nevyn^ pdelgallego killerchicken_ Morphous yngress priidu kernelzilla ph^ svaksha Tuomas mquin mutex_ Cryp	15:10
	THESE FLOODS ARE BEING DONE WITH PORT 80, 8080 AND 3128 L3 PROXIES! WHY DOES FREENODE PORTSCANNER SUCK SO MUCH? WILL THEY EVER FIX IT? DO THEY EVEN CARE? DrNick splnet redondos frogonwheels disappearedng GodEater Snow-Man coyo ianmcorvidae kenyon tro Getty EricInBNE bburhans crash\ dotCOMmie jeffryh knielsen mkramer1 Mage_Dude mastroDani Fullmoon b14ck xography vinse shade_ agile cbreak_work drizzd_ caseyw alester jtaby mrtazz zorzar Tommy[D] rektide alezan	15:10
	THESE FLOODS ARE BEING DONE WITH PORT 80, 8080 AND 3128 L3 PROXIES! WHY DOES FREENODE PORTSCANNER SUCK SO MUCH? WILL THEY EVER FIX IT? DO THEY EVEN CARE? jkp meuh koki mastro ThiefMaster pcc1 felipe StyleWarz majoh Hadaka guerrilla injekt przemoc patrikf stuffcorpse kmap grumpytoad jjore peper ekontsevoy ivan Scotto_ selckin teknotus tomaw joevano biz tgunr zomg jettero th1 stepnem metze benjamin jtauber reaVer kadoban unixtippse tg tmz kokx\| eplots_ Carn H	15:10
	← Martinp24 left	15:10
brntbeer	sigh	15:10
pigdude	pc	15:10
	← pigdude left	15:10
ThiefMaster	idiots..	15:10
curtana	whut	15:10
	← lucasvo_2 left	15:10
loincloth	who gives a shit	15:10
kylehayes	claps	15:10
lteo	wereHamster: thank you, that worked :)	15:10
cbreak_work	port scanner?	15:11
lteo	cbreak_work: thank you too, i still need to wrap my head around rebase	15:11
loincloth	SOMETHING ON THE INTERNET IS NOT 100% CORRECT	15:11
Getty	oh my god!	15:11
mkramer1	wah wah LOOK AT ME	15:11
Getty	is internet by apple?	15:11
brntbeer	/panic	15:11
stringoO	lol	15:11
	← chrislerum left	15:11
basicxman	loincloth: http://xkcd.com/386/ ;D	15:11
	→ psankar joined	15:11
	→ alex__c2022 joined	15:11
loincloth	hehe yeah	15:11
tedoc2000	sigh.. and I thought someone cared about me here	15:12
	wereHamster comforts tedoc2000	15:12
	→ lamont` joined	15:12
	tedoc2000 sobs	15:12
	stringoO git cares about your tedoc2000	15:12
tedoc2000	:)	15:12
	stringoO you*	15:12
tedoc2000	I care about you too git	15:12
	→ ajpiano joined	15:12
	← lamont left	15:12
	← ajpiano left	15:12
	→ jamescarr_ joined	15:13
	→ ajpiano joined	15:13
	← cbreak_work left	15:13
	← killerchicken_ left	15:14
	→ sagsousuke joined	15:15
	→ bterm joined	15:16
	lamont` → lamont	15:17
	← AZed left	15:18
	→ AZed joined	15:18
	→ eletuchy_ joined	15:18
	← alex__c2022 left	15:19
	→ sftp joined	15:20
	← eletuchy left	15:20
	→ alex__c2022 joined	15:20
	→ kipras joined	15:21
	← ikrabbe left	15:21
bterm	new to git: can someone point me towards how to add a existing repo to a new repo i have. example is a library that i want to add to a new project and be able to pull updates from the main libraries repo but also add local modifications to my new projects repo	15:21
	→ bcardarella joined	15:22
	→ skoop joined	15:22
broonie	bterm: git submodule is probably what you're looking for.	15:22
	← unreal_ left	15:23
	→ beatak joined	15:23
diamonds	is there a way to force push overwrite the remote?	15:26
	→ Garyy joined	15:26
Garyy	THESE FLOODS ARE BEING DONE WITH PORT 80, 8080 AND 3128 L3 PROXIES! WHY DOES FREENODE PORTSCANNER SUCK SO MUCH? WILL THEY EVER FIX IT? DO THEY EVEN CARE? Garyy beatak skoop bcardarella kipras alex__c2022 sftp eletuchy_ AZed bterm sagsousuke ajpiano jamescarr_ lamont psankar bhah apsysoev Schmallon TeckniX iravid lteo durarara hipe erlnoob Yuffster hagabaka ryanakca jinkochino loincloth flaguy48 kumbayo tedoc2000 dr_win Torsten g0bl1n leo2007 alin-gym Mpenz strin	15:26
	← Garyy left	15:26
	← leo2007 left	15:26
	→ niko joined	15:26
ajpiano	lol	15:26
skoop	WTF	15:26
tedoc2000	wow	15:26
bcardarella	interesting	15:26
tedoc2000	go go gadget script kiddie!	15:27
	→ zenwryly joined	15:27
	← eletuchy_ left	15:27
	→ Heimidal joined	15:27
	← Heimidal left	15:27
diamonds	I have this issue where I made some changes, commited and pushed only to find out it doesn't jibe with the original remote (not my fork)	15:27
	→ Heimidal joined	15:28
	← comawhite left	15:28
	→ eletuchy joined	15:28
diamonds	so I rolled stuff back locally and made a couple small changes, but now I want to push	15:28
	← g0bl1n left	15:28
NfNitLoop	diamonds: you need to merge first.	15:28
diamonds	is there a way to merge and say "always use my copy"	15:28
	?	15:28
	because I don't really want to pick thru each difference	15:28
	I want to overwrite the remote	15:29
NfNitLoop	oh, if you really just want to blow away history, you could push and overwrite the remote...	15:29
	but you'll lose history.	15:29
diamonds	ok...	15:29
NfNitLoop	I forget that option. git help push. :)	15:29
	→ saschae1986 joined	15:29
NfNitLoop	aah. --force	15:30
diamonds	fucky duck… I merged and it did the opposite of what I wanted...	15:30
	I want MY copy's changes to stay	15:30
	← saschae1986 left	15:31
	← tobiassjosten left	15:31
	→ herdingcat joined	15:31
	← engla left	15:32
	← herdingcat left	15:32
	→ saschae1986 joined	15:32
saschae1986	Hi	15:32
	I've got a problem with git svn	15:32
	→ ntoll joined	15:32
	→ herdingcat joined	15:32
saschae1986	everything works fine until now. i've got a branch local "development" wich points to the svn branch "dev" (remotes/dev)	15:33
	← alester left	15:33
saschae1986	now if i merge changes from the master branch (wich points to the svn trunk) the svn path of my local devlopment branch is changed to the svn trunk	15:34
	← tvw left	15:34
saschae1986	can anybody help me with this problem or give me some hints?	15:34
	← Torsten left	15:35
	→ mbroeker joined	15:36
	→ tomsdale joined	15:36
diamonds	jesus christ… I'm working on a repo with literally 2 files	15:37
	and I've f'd up my repo beyond what I can fix	15:37
	:'(	15:37
	I'm just learnig	15:37
	it's not critical	15:37
	→ mfacenet joined	15:38
	← elnur left	15:38
	→ ComaWhite joined	15:38
	← ivan-kanis left	15:39
Ilari	diamonds: As in, repo corrupt?	15:39
diamonds	no	15:39
	I just can't figure out how to go back	15:39
	revert	15:39
	← bcardarella left	15:39
diamonds	I'm looking at git help revert now	15:39
Ilari	diamonds: reset or checkout?	15:39
diamonds	but sadly I don't know how to access "See the revert-a-faulty-merge How-To[1] for more details."	15:39
	my man skills are limited	15:40
	→ hedgehog83 joined	15:40
	→ alester joined	15:41
hedgehog83	hi, git send-email hangs with this strange error message: http://pastebin.com/CLSExvMv	15:41
	→ webus joined	15:42
diamonds	Ilari: … I forked, checked out, did some work, pushed, realized the main repo (not my fork) had been updated, added fetched merged it, couldn't resolve conflicts, decided to roll back before my changes and merge in the main remote branch, did that, made a couple small changes, tried to push to my remote (the fork), and couldn't because it still reflected my FIRST set of changes, fetched it, merged it in, and it's changes overwrote what I	15:42
	ugh	15:42
	I think I'll just can it and start over :(	15:42
	Hunner_ → Hunner	15:42
	← erlnoob left	15:43
diamonds	i really wish I could work it out tho!	15:43
Ilari	diamonds: 'git log' shows what commits it has.	15:43
diamonds	I know	15:43
	i've tried to git revert <hash> repeatedly	15:43
	but I can't get up past a merge	15:44
	and I'm not sure how to specify the parent	15:44
Ilari	diamonds: Use reset instead?	15:44
diamonds	ok	15:44
	→ bcardarella joined	15:44
Ilari	diamonds: But don't reset past what's in main repo.	15:44
	← mbroeker left	15:44
	← lteo left	15:45
	→ Torsten joined	15:45
Ilari	diamonds: (because you presumably don't want to undo stuff from main repo).	15:45
diamonds	I reset but it's not being reflected in the files	15:45
	they still look… the same as when I messed them up by merging	15:46
Ilari	diamonds: --hard if you want to blow the uncommitted changes away.	15:46
	← nicoulaj left	15:47
diamonds	how do I get to this man page revert-a-faulty-merge How-To[1] ?	15:47
	← arialdo left	15:48
diamonds	and if I reset, how can I move back forward?	15:48
wereHamster	diamonds: http://www.kernel.org/pub/software/scm/git/docs/howto/revert-a-faulty-merge.txt	15:48
	→ mattikus joined	15:50
	→ jrmuizel joined	15:50
	→ slonopotamus joined	15:51
	← hipe left	15:52
diamonds	how do I move forward thru commits after resetting?	15:52
	I can see them in the history with gitk --all	15:52
	← artefon left	15:52
slonopotamus	diamonds: you see them because they're on a branch?	15:52
	→ unreal joined	15:52
diamonds	hmm… yes?	15:53
	remotes/origin/master	15:53
	and other remotes	15:53
slonopotamus	so... you can reset to them :)	15:54
diamonds	but I have uncommitted changes it says	15:54
	tho I didn't change anything	15:54
	it must be the result of merging and/or reverting	15:54
	← sftp left	15:54
slonopotamus	diamonds: what git status says?	15:55
diamonds	says I have 2 modified files	15:55
slonopotamus	well, sw doesn't lie :)	15:56
diamonds	I don't know why tho, I didn't change them after the messed up merge	15:56
	→ justin-george joined	15:56
diamonds	I don't know how they got "modified"	15:56
slonopotamus	diamonds: you can git reset --hard, that'll blow away all working copy changes WITHOUT A WAY TO RECOVER THEM	15:57
diamonds	ok	15:57
slonopotamus	you've been warned :)	15:57
diamonds	I just used git checkout — files	15:57
	and now I can fast forward, it tells me :)	15:57
slonopotamus	that doesn't reset index	15:58
	→ Nemurenai joined	15:58
	← beatak left	15:58
diamonds	so to fast forward:	15:59
	→ sftp joined	15:59
	← justin-george left	15:59
diamonds	git checkout <shah> right?	15:59
wereHamster	'git pull'	15:59
diamonds	I'll try it	15:59
wereHamster	diamonds: no, git checkout <sha1> will switch HEAD and your working tree tot hat commit	15:59
	→ justin-george joined	15:59
	→ friskd joined	15:59
diamonds	i'm really flummoxed. before, when I checked out a commit, my working copy reflected that	16:00
wereHamster	yes, it does	16:00
	diamonds: though you'll end up on a detached head (faq detached)	16:00
Gitbot	diamonds: You are on a detached HEAD and might lose commits.. See https://git.wiki.kernel.org/index.php/GitFaq#detached	16:00
	← psankar left	16:00
diamonds	I think I might have found the error	16:01
	← kumbayo left	16:01
diamonds	git tricked me into pointing my browser to the wrong local directory!!	16:01
	→ kipras_ joined	16:01
diamonds	that trickster… :p	16:01
	← kipras left	16:01
diamonds	ok I'm where I want to be, before the faulty merge	16:02
	← CrypTom left	16:03
brntbeer	if i accidentally apply a rebase to my branch, will a git reset hard^n go back to the way things were, where 'n' is the number of commits applied to my current branch	16:03
diamonds	why does gitk —all see things git log —all doesn't?	16:03
slonopotamus	diamonds: reflog maybe?	16:03
	(just a guess)	16:03
wereHamster	brntbeer: no	16:04
	brntbeer: reset --hard branch@{1}	16:04
	← ComaWhite left	16:04
brntbeer	thanks	16:04
wereHamster	brntbeer: reset HEAD^n would remove all the commits you had on that branch before	16:05
	→ bcrouse joined	16:05
brntbeer	fking savior, thanks	16:05
	client meeting in an hour	16:05
wereHamster	'rebase otherbranch; reset --hard HEAD^n' is the same as 'reset --hard otherbranch'	16:05
diamonds	well lost a couple changes in my flailing but not so bad	16:05
wereHamster	diamonds: the most you lost are uncommitted changes. Everything else is available in one of the reflogs	16:06
	you just need to know where to look for it	16:06
brntbeer	wereHamster: it seems im reading the git-rebase info online wrong. if i wanted to commit upstream changes from master onto my current branch, is that git rebase master current_branch ?	16:06
	← yngress left	16:06
	← hagabaka left	16:06
diamonds	yeah I think i may have not committed	16:07
wereHamster	brntbeer: git merge master	16:07
	→ psankar joined	16:07
	→ ComaWhite joined	16:07
wereHamster	diamonds: now you know what we mean when we tell people to 'commit early, commit often'	16:07
	→ amerine joined	16:08
	← mkramer1 left	16:08
diamonds	:)	16:08
	→ hagabaka joined	16:08
	← hagabaka left	16:08
	→ hagabaka joined	16:08
diamonds	yeah that's why I'm messing around with it now before I start using it at work	16:09
	try to get as many mistakes out of the was as possible	16:09
	→ hedgehog831 joined	16:11
	← hedgehog831 left	16:11
	← hedgehog83 left	16:12
	→ mbroeker joined	16:12
	→ hedgehog83 joined	16:13
	→ Kipras__ joined	16:13
	← saschae1986 left	16:14
	IslandUsurper → IslandUsurperAFK	16:14
	← kipras_ left	16:15
	← pdignan left	16:19
	→ psoo joined	16:19
	→ erlnoob joined	16:19
	→ corni joined	16:20
	→ enherit joined	16:20
	→ spaceonline joined	16:20
	← JeffJohnson left	16:22
	→ artefon joined	16:22
	← priidu left	16:22
	→ tj_ joined	16:24
	← brntbeer left	16:25
	← asmega left	16:26
	→ vianaweb joined	16:27
	→ MattDiPa_ joined	16:27
	← eletuchy left	16:27
	← vianaweb left	16:27
	→ Theravadan joined	16:27
	← MattDiPasquale left	16:27
Theravadan	this doesnt seem to create a remote branch, anyone know why? git push origin origin:refs/release-candidate/2.06.7	16:28
fr0sty	why should it?	16:28
	← psoo left	16:29
Theravadan	do you know the correct syntax?	16:29
fr0sty	Theravadan: man git-push	16:29
Gitbot	Theravadan: the git-push manpage can be found at http://git.or.cz/man/git-push	16:29
	← Theravadan left	16:29
fr0sty	it is: 'git push <remote> <refspec>'	16:29
	<refspec> can be several things.	16:30
	<src>:<dest> is the usual format	16:30
	→ mrevd joined	16:31
	← alsuren_ left	16:31
	← mrevd left	16:31
fr0sty	what are you trying to push? what do you want the branch named on 'origin'?	16:31
	← EGreg left	16:31
	→ krawek joined	16:32
	→ EGreg joined	16:32
fr0sty	the refspec you gave tries to push a branch named 'origin' to a 'refs/release_candidate/2.06.7' on the remote named origin	16:32
	What are you trying to accomplish?	16:33
	(sorry for the wall of text)	16:33
	← galderz left	16:33
	→ slunk joined	16:33
tomsdale	I did a 'git cherry-pick' SHA on a previous commit. Now my head has ++<<<<<<< HEAD parts inserted in the files. What's the next step to pick the lines I want to keep from the old and the new file?	16:34
	← copelco left	16:35
	→ eletuchy joined	16:35
	← ntoll left	16:35
	← EricInBNE left	16:36
	→ zorzar_ joined	16:36
	→ inetic joined	16:36
	→ copelco joined	16:36
	← copelco left	16:36
	→ blueyed joined	16:37
blueyed	Can I undo a "git checkout ." to get back local uncommited changes?	16:37
	← nadim left	16:38
	← sagsousuke left	16:38
	← zorzar left	16:38
nothingHappens	i'm pretty sure no	16:39
	→ ciampix joined	16:40
slonopotamus	blueyed: no	16:41
	→ aah joined	16:41
	← ciampix left	16:41
blueyed	too bad then. Can you configure a "are you sure" prompt for such invasive actions?	16:41
jast	no. it is assumed that you know what you are doing. :}	16:42
	← bcardarella left	16:42
	← mikeric left	16:42
slonopotamus	blueyed: post a suggestion to git maillist (if there ain't one already)	16:43
jast	fwiw no other SCM system that I'm aware of has a prompt like that	16:43
blueyed	yes. It was just that I did a "distclean", noticed that it removed required files, which are in git and wanted to get them back. I've forgotten about the changes from weeks ago laying around uncommitted.	16:43
slonopotamus	jast: it is inconsistent with git clean that requires explicit -f/-d	16:44
blueyed	jast: FWIW bzr will move the original files to *.~$NR~	16:44
jast	git clean completely removes files. that's a bit of a difference.	16:44
	and "checkout" requires an explicit path argument for overwriting files, unlike clean	16:44
	→ r4tune joined	16:44
slonopotamus	blueyed: you don't want to have uncommited changes for long time	16:45
blueyed	well, a dot is typed easily, when you've got used to it.	16:45
	slonopotamus: I know, yes.	16:45
r4tune	is there a command to obliterate all trace of a file that should never have existed in a git repo, so it was like it was never there?	16:45
blueyed	It's where I left off.	16:45
jast	well, a prompt is easy to confirm, when you've got used to it.	16:45
	← psankar left	16:45
	→ vu3rdd joined	16:45
r4tune	it's had several commits made to it	16:45
	→ catalis joined	16:45
jast	r4tune: man git-filter-branch, and note that all commit IDs will change, thus making further merges with old versions of the repository very difficult.	16:45
Gitbot	r4tune: the git-filter-branch manpage can be found at http://git.or.cz/man/git-filter-branch	16:45
r4tune	thank you	16:46
jast	third example or so in that page	16:46
slonopotamus	jast: they both (+ git reset) do the same - throw away working copy changes	16:46
	→ ereslibre joined	16:46
jast	no, clean doesn't throw away changes, it throws away unrelated files	16:47
	and all of the three commands require an extra argument to actually clobber files; with clean it's -f, with checkout it's a path, with reset it's --hard	16:47
slonopotamus	btw, is git reset <path> && git checkout <path> doable in one command?	16:47
	← corni left	16:47
jast	git checkout HEAD <path>	16:48
slonopotamus	jast: you type paths from parent to child so accidental <ENTER> in the middle of the path may checkout more than you wanted	16:48
	clean/reset - agreed	16:48
jast	how often do you start checkout paths with '.'?	16:49
	I'd argue that it's very hard to, say, hit enter accidentally in the middle of typing '..'	16:49
slonopotamus	doesn't matter	16:49
	← jinkochino left	16:49
slonopotamus	git checkout dir<woops, enter>/dir/file	16:50
jast	at any rate, it's just as easy to accidentally confirm a prompt as it is to accidentally type the wrong path	16:50
	e.g. if you don't actually read the prompt... and I know from quite a bit of experience in this channel that many people completely ignore everything git outputs	16:50
	you wouldn't believe how often people say something like "rebase failed, how do I know which files are conflicted?"	16:51
	r4tune looks at the commit message logs	16:51
slonopotamus	jast: well, at least you type --hard/-f intentionally	16:51
r4tune	rm -rf .git	16:51
	there, file never existed :D	16:51
jast	r4tune: okay, if you don't have a lot of interesting history that's definitely the easiest way :)	16:51
r4tune	there was too many references to that which should not have been there in the log	16:52
slonopotamus	r4tune: i teached myself to never ever do rm when not in the direct parent dir of file i'm going to remove :)	16:52
jast	slonopotamus: personally I would do a lot of restructuring in git's commands anyway, but it would break backward compatibility big time	16:52
r4tune	and i fgrepd for it ./.git/COMMIT_EDITMSG has references to it	16:52
jast	I always check CWD and the command I entered before I hit enter on anything that's destructive :)	16:53
slonopotamus	rm -rf .git isn't very destructive since repo contents are pushed to other clases	16:53
wereHamster	jast: isn't cwd in your shell prompt?	16:53
jast	yes it is	16:53
	with particularly destructive commands I'll even put them in the cut buffer and do more elaborate checks before I actually use them	16:54
	← mattikus left	16:54
jast	for example, every time I do something on a disk device, even if I'm pretty sure that I've picked the right device path I still make sure once more before I run the command	16:54
r4tune	jast, i make sure then put it variable :D	16:55
	→ doener joined	16:55
jast	that would work if device names didn't change	16:55
	r4tune refers through the udev by-id to devs	16:56
ice799	Hi there. If i have a remote branch named X and a local branch named A, what do I need to change so that I can push my changes on local branch A to remote branch X ?	16:57
jast	strangely enough, not every USB storage device I use (and I use lots of them) has the same ID	16:57
ice799	i cant remember. i tried modifying the .git/config merge line, but it didnt seem to help.	16:57
r4tune	thats the point they're quite unique	16:57
	good reference	16:58
	← henryvps left	16:58
wereHamster	ice799: git push --set-upstream origin A:X	16:58
	← intripoon left	16:58
jast	yeah, so every time I do something on a USB device I have to create a new variable for its ID?	16:58
	ice799: note that by default, push will not reuse the mapping you create in that suggested command	16:58
	→ sonnym joined	16:58
r4tune	no, only when you're doing mega-destructive things :)	16:58
	← Torsten left	16:59
	← pfig left	16:59
	← mastroDani left	16:59
r4tune	it's pretty hard to dd syslinux onto your lvm setup via the by-id route	16:59
	(when you want it to boot a usb key)	16:59
ice799	wereHamster, jast: thanks	17:00
	→ f13o joined	17:01
	→ Torsten joined	17:02
	← f13o left	17:02
	→ EXABA7N3 joined	17:02
	← rgr left	17:02
blueyed	FWIW, Eclipse's local history was able to get me the files back from before "git co ."	17:04
	← ryanakca left	17:04
	← GodEater left	17:04
	→ jvasak joined	17:04
	← catalis left	17:05
	→ mkramer joined	17:07
	→ beni- joined	17:07
	→ canavar joined	17:07
bcrouse	I'm having some trouble establishing a functional workflow. anyone feel like taking a look? http://bit.ly/c1LbNj	17:08
dsop	someone ever used skip-worktree?	17:09
	← hohoho left	17:09
	→ ryanakca joined	17:10
wereHamster	bcrouse: your solution is not to rebase	17:10
bcrouse	wereHamster: you think the logs will still be reasonably usable in a few years then? I'm not experienced, but I was guessing it would get really tough to read	17:12
wereHamster	bcrouse: from what I understand the rebase causes conflicts because the developer did merge base/master into his branch prior to running the rebase, right?	17:12
dsop	is there a command to see the current content of the index?	17:12
	→ bauruine joined	17:12
wereHamster	dsop: ls-files ?	17:13
bcrouse	wereHamster: right. for example, I merge from base, push to origin, and then another developer runs the pull/rebase. that's what creates all the conflicts	17:13
	→ picardo12 joined	17:13
doener	bcrouse: client/master == master?	17:13
bcrouse	doener: not sure what you mean. by client/master I just mean to notate the master branch of the client repo	17:14
wereHamster	bcrouse: the question is, what did the other developer do in his repo? merge base/master? Or only do simply client-specific commit?	17:14
	← Schmallon left	17:14
bcrouse	wereHamster: only client-specific commits	17:14
doener	bcrouse: so "on client/master" actually means "in the 'client' repo, with 'master' checked out"?	17:14
bcrouse	wereHamster: I've been trying to control who is doing the merging from base into client	17:14
	doener: exactly	17:15
	← apsysoev left	17:15
wereHamster	bcrouse: than it's likely that a merge will also cause conflicts.	17:15
doener	bcrouse: ok. client/master looked like a remote tracking branch	17:15
	→ catalis joined	17:15
bcrouse	doener: ahhh sorry. should've chosen better language	17:15
doener	bcrouse: is "master" setup with origin's master as its upstream?	17:16
	bcrouse: what does this say in the client repo? git rev-parse --symbolic-full-name master@{u}	17:16
	← yotaff left	17:16
bcrouse	refs/remotes/origin/master	17:16
	← Torsten left	17:17
doener	bcrouse: OK, you're most likely being hit by a, uhm, corner case with "pull --rebase" trying to deal with upstream being rewritten	17:17
	→ girishr joined	17:17
	→ mikeric joined	17:18
	← ExtraSpice left	17:18
doener	bcrouse: basically, it's because you use "git pull --rebase origin master", which internally does "git fetch origin master", which does _not_ update origin/master (the remote tracking branch)	17:18
	bcrouse: and then "pull --rebase" fails to correctly detect which commits to actually rebase	17:18
	bcrouse: I'd bet that things start working as expected if you do just "git pull --rebase"	17:19
bcrouse	doener: that sounds correct, according to the symptoms I've seen. any advice?	17:19
	← Kipras__ left	17:19
bcrouse	doener: I'll give that a shot, thanks!	17:19
	← picardo12 left	17:20
	← tstclair left	17:20
	→ venky10 joined	17:20
doener	that will use the upstream configuration for the checked out branch and should then do The Right Thing	17:20
venky10	Hi all. I'm a newb and I'm stuck in something awkward that I'm not able to understand	17:20
	← shruggar left	17:20
	→ nepalaya joined	17:20
venky10	Can someone please help?	17:21
bcrouse	doener: just so I understand then, that's a workaround? or will that do something specifically different	17:21
	← markbao_ left	17:22
	→ itrebal joined	17:23
	→ thiago joined	17:23
fr0sty	venky10: just ask your question...	17:23
	→ priidu joined	17:24
	→ rseifert joined	17:24
	→ eno joined	17:24
venky10	fr0sty: I was learning how git push works and trying out a simple push from a cloned repository to the main repo. My main repo has only one branch "master" and so does my local repo. The local repo is updated to the latest changes in the main one.	17:25
	← tatsuya left	17:25
	← Soliton left	17:26
venky10	This was done through "git pull" on the local repo.	17:26
	Now, when I commit some changes in the local repo and do a "git push" I see the changes in the main repo only in the cached diffs	17:26
wereHamster	venky10: faq non-bare ?	17:26
Gitbot	venky10: Pushing to non-bare repositories is discouraged. See https://git.wiki.kernel.org/index.php/GitFaq#non-bare	17:26
doener	bcrouse: that's the way it's meant to be used	17:27
bcrouse	doener: oooh ok. thanks so much!	17:27
	IslandUsurperAFK → IslandUsurper	17:27
wereHamster	doener: maybe pull should fail when --rebase is given along with remote and refspec	17:27
	→ petercoulton joined	17:27
venky10	wereHamster,fr0sty:I wasn't aware of non-bare. I'll look into it. Thanks.	17:27
doener	wereHamster: that, or maybe the "detect the right onto/upstream parameters" thing should be turned off then	17:28
	→ khmarbaise joined	17:28
petercoulton	Is there a way to checkout a single line of a file, maybe by editing a patch, add -i style?	17:28
doener	wereHamster: I'm not sure what I'd prefer... But then again, I usually don't use pull	17:28
wereHamster	yeah, me neither	17:28
jast	petercoulton: checkout -p, of course :)	17:28
bcrouse	doener, wereHamster: if you don't use pull, what do you suggest?	17:29
wereHamster	bcrouse: fetch + merge or rebase	17:30
	bcrouse: pull is just a wrapper around those, and as you saw yourself it will not work correctly in all corner cases	17:30
bcrouse	wereHamster: I thought that was just the long way of doing what pull does?	17:30
petercoulton	jast, of course! Cheers.	17:30
	→ picardo12 joined	17:30
jast	in case you got a different impression: pull isn't really a wrong thing to do in general, the guys are just discussing possible changes to how it works	17:30
	(@ bcrouse)	17:31
bcrouse	ok, that makes sense	17:31
	← MattDiPa_ left	17:31
	→ Bass10 joined	17:32
venky10	Can someone please explain what a 'hook' is?	17:34
	← Bass10 left	17:34
	→ Soliton joined	17:34
	← ComaWhite left	17:34
	→ Bass10 joined	17:34
	→ p3rg2z joined	17:35
fr0sty	venky10: man git-hooks	17:35
	venky10: man githooks	17:36
Gitbot	venky10: the githooks manpage can be found at http://git.or.cz/man/githooks	17:36
	→ slonopotamus_ joined	17:36
	← slonopotamus left	17:36
	→ MattDiPasquale joined	17:36
fr0sty	basically, they are scripts that run at certain times which allow actions to be automatically taken (up to and including refusal to complete a requested action)	17:36
	← Fullmoon left	17:38
venky10	So, the wiki for git push to non-bare talks about post-update hook. Does this mean that I write a script to merge 'pushed' changes from a temp branch to master branch on remote repo?	17:38
	← nevyn^ left	17:39
cbreak	no	17:40
	merging can not be automated	17:40
	you can write one that just deletes what ever is there	17:40
patrikf	venky10: you don't normally push to non-bare repos.	17:40
	→ nicoulaj joined	17:41
cbreak	and replaces it with what you push	17:41
	← girishr left	17:42
venky10	Okay. So, can I have "dummy" branch set to working branch always and then all 'pushes' coming in can go into "master" right?	17:42
	adamholt → adamholt_away	17:43
	adamholt_away → adamholt	17:43
crab	venky10: what is it you're trying to do?	17:43
	→ GodEater joined	17:43
	← GodEater left	17:43
	→ GodEater joined	17:43
	← jvasak left	17:45
venky10	crab:I'm just trying to figure out how to update remote repository using "git push". I want to get the changes into "master" branch of the remote repository	17:45
	← curtana left	17:45
	→ Spines joined	17:45
crab	venky10: (1) push to a bare repository, (2) maybe see http://toroid.org/ams/git-central-repo-howto	17:45
	→ programble joined	17:46
venky10	crab:Does bare repository mean non-current and no-changes-pending repository?	17:47
	→ Chillance joined	17:47
	← bluenovember left	17:47
crab	venky10: no, bare means a repository initialised with git init --bare	17:48
	← slunk left	17:48
	→ roygbiv joined	17:48
crab	i.e. one without a working tree.	17:48
	← ncopa left	17:49
	→ kandjar joined	17:50
kandjar	hi there	17:50
	zorzar_ → zorzar	17:50
kandjar	I've been hosting my git repo on my linux box, but keep getting some permission issue when I push commits...	17:51
	every once in a while I had to refresh the perm on the object forder.	17:51
venky10	crab: So, its a repository created so that people can simply do "git push" to show up on that repository	17:51
crab	venky10: pretty much, yeah.	17:52
	→ sergelamoss joined	17:52
venky10	crab:Can I create a branch off of the bare repository master and start making changes on this new branch?	17:52
crab	venky10: you mean clone the bare repository and work on a local branch? yes, of course.	17:53
kandjar	"drwxrwsr-x 104 kandjar git-user 4096 2010-08-10 20:30 objects" this is the permission for my object folder, am I missing something?	17:53
crab	venky10: (look at that howto)	17:53
	← RORgasm left	17:53
	→ steph021_mac joined	17:53
	→ RORgasm joined	17:53
venky10	crab:Thanks!	17:53
	← iravid left	17:53
	← adymitruk left	17:54
	← parasti left	17:54
	← inetic left	17:56
	→ theoros joined	17:56
	→ ComaWhite joined	17:57
	→ parasti joined	17:58
	→ ciampix joined	17:59
	→ zenazn joined	17:59
	← catalis left	18:00
	← itrebal left	18:00
	← slonopotamus_ left	18:01
	← ereslibre left	18:01
	→ slonopotamus_ joined	18:01
kandjar	anyone?	18:01
ciampix	I ... but I'm not a GIT guru ... sorry...	18:02
	→ badluck2 joined	18:02
kandjar	:)	18:02
jacobat	Anyone know how to get git log output like this: http://skitch.com/jacobat/d16pn/git-inception	18:03
	→ hagebake joined	18:03
	← hagebake left	18:03
	→ hagebake joined	18:03
	← hagabaka left	18:04
basicxman	Yeah, git log --pretty="%h - %s - %cr"	18:05
	Rougly anyways	18:05
badluck2	Hi. i am using git as a client for svn. I tried to dcommit and now it looks like (master\|REBASE) and i dont know how to get it to become (master)	18:05
	← erlnoob left	18:05
jacobat	it complains about %h: preexec:printf: %h : invalid directive	18:07
tomsdale	I did a git cherry-pick (with gitk). It seems to have created .orig files and git status says 'both added:' What is the next command to get back to just 1 version?	18:07
	→ adymitruk joined	18:08
theoros	what does "get back to just 1 version" mean?	18:08
	→ itrebal joined	18:09
basicxman	jacobat: Works great for me, can you paste your command + full output to http://pastie.org ?	18:10
	Also: http://theroboticsuniverse.net/~andrew/gitnotes.txt for format modifier reference.	18:11
jacobat	basicxman: http://pastie.org/1104634	18:11
	← razvand left	18:12
tomsdale	theoros: git seems to be aware that it currently has 2 version of files in the working branch one with an additional .orig files. I just deleted the .orig files and commited. Seems to be fine now.	18:12
basicxman	jacobat: What version of git are you using?	18:12
jacobat	1.7.2.2	18:13
theoros	tomsdale: git sees it is a copy of content, i think	18:13
	tomsdale: treat it as any other file and git rm it	18:13
	→ hearit joined	18:13
theoros	tomsdale: also cherry-pick at the console, it's more fun that way :)	18:13
	plus you get to do things like line up a bunch of commit hashes and pipe them into git cherry-pick	18:14
	zz_rubydiamond → rubydiamond	18:15
	← alex__c2022 left	18:15
	→ erlnoob joined	18:15
	→ mattwynne joined	18:16
basicxman	jacobat: Not sure sorry, perhaps a bug.	18:16
	tj_ → tjholowaychuk	18:17
tomsdale	theoros: I tried that first but too but didn't know how to do the actual merge with the <<<<-HEAD parts in my files. What's the actual command for this?	18:17
jacobat	basicxman: Alright, thanks anyway :)	18:17
basicxman	jacobat: You could always ask it on the mailing list (git@vger.kernel.org)	18:18
jacobat	right	18:18
ciampix	basicxman: thanks for the git log tip... lurking here is really useful... :-)	18:19
basicxman	ciampix: np!	18:19
theoros	tomsdale: what do you mean by "do the actual merge" with those parts? you mean declare a merge conflict resolved?	18:20
basicxman	jacobat: Also try %H instead of %h just to see what happens, %H is the full hash, %h is abbreviated like the picture you showed.	18:21
jacobat	same error	18:21
basicxman	Interesting, you are in a repo that has commits right? (sorry I have to ask the dumb question)	18:22
jacobat	Yeah, I am :)	18:22
	→ bcardarella joined	18:22
	slonopotamus_ → slonopotamus	18:22
	→ reflectivedev joined	18:22
jacobat	git log works	18:22
	← bauruine left	18:22
tomsdale	theoros: I did 'git cherry-pick #SHA'. It told me no automatic merge was possible and inserted sections of ++<<<<<<< HEAD parts into my files where there were conflicts. I didn't know how to proceed on the commandline from this point onwards.	18:23
basicxman	Ah hmm, interesting how only %[Hh] isn't working. If you take %h out all together does it do the other two?	18:23
theoros	tomsdale: it's called a merge conflict.	18:23
crab	tomsdale: edit the file, fix the conflict, then git add $file, and git commit -c $sha	18:23
theoros	tomsdale: git was unable to merge in a clean way the new changes	18:23
	← hedgehog83 left	18:23
theoros	tomsdale: and if you read the output of the conflict message, telling you there's a conflict, it will tell you exactly how to mark the conflict as resolved.	18:24
	← sh1mmer left	18:25
	← bentob0x left	18:25
tomsdale	thx, I will try it manually the next time.	18:27
	→ rutski joined	18:27
rutski	say I just did a "git rm foo.c" by mistake, followed by a commit & push	18:27
	how might I get foo.c back?	18:27
	rutski is still kinda new to git	18:28
	← quantumelixir left	18:28
jast	you can get an older version like this: git checkout <old commit> foo.c	18:28
	→ quantumelixir joined	18:28
jast	after just one commit, HEAD~ as the commit ID should do the trick	18:28
rutski	ah, rgiht; and git-log gives me the commit	18:28
	neat	18:28
	ah	18:28
	→ drizzd joined	18:28
jast	~ = parent :)	18:28
jacobat	basicxman: Haha, it works in bash, but not in zsh	18:29
rutski	wait... git-log doesn't show me which commit had the git-rm though	18:29
	(I know I can use ~, but I'm just curious)	18:29
	← khmarbaise left	18:29
rutski	how might I find the ID of the commit which had the rm?	18:29
jast	jacobat: you may have to escape the []. zsh has options that make it interpret, for example, {abc} as {a,b,c}	18:29
	rutski: if you use that commit ID, you won't get the file back because in that commit the file is already gone :)	18:29
jacobat	jast: This is related to format strings	18:30
basicxman	jacobat: Ah!	18:30
jast	jacobat: yes, and zsh interprets [aa] differently than [a] on the command line	18:30
basicxman	jast: [] isn't actually in the format string.	18:30
jacobat	jast: git log pretty="%h" works in bash, not in zsh	18:30
jast	oh. I see now.	18:30
	← venky10 left	18:30
	← kylehayes left	18:31
	← drizzd_ left	18:31
	→ g0bl1n joined	18:31
	← picardo12 left	18:31
	→ sh1mmer joined	18:32
jast	jacobat: well, in my zsh it works :)	18:32
	→ olaf joined	18:32
jacobat	Hmm... for bash, printf is /usr/bin/printf, in zsh it's a shell command	18:32
	→ meGenius joined	18:32
jacobat	jast: lucky you ;)	18:32
	jast, which git version?	18:32
meGenius	hi all!	18:32
jast	1.7.1.something	18:32
	I guess I could update while I go buy some food	18:33
jacobat	then it might not work anymore ;)	18:33
	→ Aaaarg_ joined	18:33
meGenius	i spent all the day trying to know how to work around this thing called git	18:33
	→ tstclair joined	18:34
meGenius	i want to set a new git bare repository, but, i don't know how to add files to it	18:34
	← bcardarella left	18:34
jast	meGenius: bare repositories are used only for pushing to them from other repositories	18:34
meGenius	i googled a lot but didn't find a reference for this	18:34
	→ Mage joined	18:34
jast	jacobat: still works :)	18:35
	→ ereslibre joined	18:35
jacobat	jast: which version of zsh?	18:35
meGenius	jast: in fact, the image is not clear	18:35
jast	I'm on git 1.7.2.2.108.gc1196 now	18:35
	and zsh 4.3.10	18:35
jacobat	mkay	18:35
	maybe my zsh is too old	18:35
	nope, not it	18:36
jast	what does it actually do, anyway? I mean, "doesn't work" is not very specific	18:36
	← Aaaarg left	18:37
	Aaaarg_ → Aaaarg	18:37
jast	meGenius: sorry, my crystal ball is full of smoke, so it can't tell me what part is unclear to you	18:37
	perhaps if you tell me... :}	18:37
	→ pflanze joined	18:38
jacobat	jast: http://pastie.org/1104634	18:38
	but interestingly it works if I make it an alias	18:38
	so apparently zsh is doing some shell expansion	18:38
meGenius	jast: i use a very good kind of tissues i should send you some :P	18:38
	however, what do you mean by pushing from other repos	18:39
jast	well, I assume that by "bare repository" you meant the "--bare" switch for git init. right?	18:39
meGenius	does that mean that i should have a local repo & remote repo & make the changes to the local & push to the remote??	18:39
	yup!	18:39
jast	okay	18:39
	bare repositories are mostly meant for hosting stuff	18:40
	→ unreal_ joined	18:40
jast	so they prevent directly committing inside them by not having a checked out copy of the files anywhere	18:40
	in any other case, people pushing changes to the repository could potentially mess up whatever someone else is currently working on in the checked out files	18:40
	clear so far?	18:41
	← unreal left	18:41
	unreal_ → unreal	18:41
	← olaf left	18:41
meGenius	i understood why there's bare repos	18:41
jast	(gotta step out for 20 minutes or so, but I'll get back to this topic when I return)	18:42
meGenius	i want to know - now - how can i add files to it & edit then letter	18:42
	← ereslibre left	18:42
meGenius	jast: tyt! i'll seek the help for the others ;)	18:42
	→ ludototo joined	18:42
jast	if you still have any questions left when I come back, I'll be happy to get back to this	18:42
meGenius	okay! guys, is there anyone who is ready to take jast's place??	18:42
	thnx, jast :)	18:43
	← ludototo left	18:43
	→ olaf joined	18:44
	← roygbiv left	18:44
	→ ludototo joined	18:45
	→ bcardarella joined	18:49
	→ Textmode joined	18:50
	← bcardarella left	18:51
	→ bcardarella joined	18:51
	→ drizzd_ joined	18:51
	← gaspar\|work left	18:51
	→ apsysoev joined	18:51
	→ Titosemi_ joined	18:52
	← apsysoev left	18:53
	→ kogent joined	18:53
	← Titosemi left	18:54
	Titosemi_ → Titosemi	18:54
	← ComaWhite left	18:54
	→ unreal_ joined	18:55
	← drizzd left	18:55
fr0sty	meGenius: '	18:55
	expand on: 'add files to it and edit them later'	18:56
	specifically the 'it'	18:56
	← ludototo left	18:56
	→ murz joined	18:57
	← unreal_ left	18:57
	→ nevyn^ joined	18:57
	→ comawhite joined	18:57
	→ unreal_ joined	18:58
	← unreal left	18:58
	unreal_ → unreal	18:58
	→ sgh joined	18:59
	← ciampix left	18:59
	← sftp left	18:59
	→ unreal_ joined	19:00
	→ bosie joined	19:00
	← nevyn^ left	19:00
	← unreal left	19:00
	→ chronos joined	19:00
	← badluck2 left	19:01
jast	meGenius: now, since we have already established what bare repositories are, is it also clear that they are supposed to be used as a central place for people (or just one person, if you feel like it) to collect changes from different places?	19:02
	unreal_ → unreal	19:02
	→ ntoll joined	19:02
	← bosie left	19:03
jast	jacobat: looks like something tells zsh that the string should be processed by printf. out of interest, does this work better: git log --pretty=%h (without quotes)	19:03
	← comawhite left	19:03
	→ comawhite joined	19:04
	→ Schmallon joined	19:04
meGenius	jast: sorry i was afk :)	19:05
jast	jacobat: I think some part of your shell environment defines a wonky preexec function	19:05
jacobat	jast: I found the issue - it was with a zsh config for setting the title of the terminal tab	19:05
jast	ah... fun	19:05
jacobat	sort of ;)	19:06
meGenius	fr0sty: the bare repo is a repo anyway, that means, i can add files to it, right??	19:06
jast	well :)	19:06
	technically yes, but you're not supposed to work inside a bare repo, so it's been deliberately made very hard to do that	19:06
fr0sty	meGenius: you can push commits to it.	19:06
jast	bare repositories are usually updated by pushing to them from other, non-bare, repositories	19:07
	→ Shurakai_ joined	19:07
jast	normal mode of operation for a collaborator is that he clones the central bare repository, changes stuff locally, commits it, then pushes back to the central repo	19:07
	← Shurakai left	19:07
meGenius	ah! i see	19:08
	← richardbronosky left	19:08
meGenius	so i should clone the repos locally!!	19:08
	meGenius is trying that...	19:08
	→ Eaven joined	19:08
	→ Radar joined	19:09
	← nepalaya left	19:10
	← durarara left	19:10
	← blueyed left	19:16
	← artefon left	19:16
	→ dgovinda joined	19:16
	→ nadim joined	19:16
	← petercoulton left	19:18
	→ Kipperoo joined	19:19
	← SimonP86 left	19:20
Kipperoo	Hello. Can anyone recommend a rich text document format that plays nice with git?	19:20
	→ catalis joined	19:21
	→ andre_pl joined	19:21
	→ bhearsum joined	19:22
meGenius	how to push to the bare repo?? i cloned it & added files & then commited the changes, git push gives me fatal: The remote end hung up unexpectedly error.	19:23
andre_pl	is it possible to have code in a git repo rooted at /dir1/ (ie. /dir1/.git/ contains the configs) and clone it into /dir2/ without creating dir1 within dir2?	19:23
bhearsum	is there a generally accepted way to mirror an hg repo into a git one? i've got upstream source that i develop against, but i want to do my development in git	19:23
	→ durarara joined	19:24
wereHamster	Kipperoo: tex	19:24
Kipperoo	Tex. What about something likely to be supported by 'doze?	19:24
bremner	what do you mean supported by windows?	19:25
	TeX can make PDF, btw	19:25
jast	define "plays nice with git"	19:25
wereHamster	I believe there's a tex distribution for windows	19:25
Kipperoo	Umm, easily accessible - editable - by users who are on windows machines.	19:25
theoros	wereHamster: clever answer	19:25
	miktex is decent	19:25
jast	git doesn't edit the files, users do :)	19:25
theoros	miktex + texniccenter	19:25
Kipperoo	Yes, I should have been a lot more specific, haha.	19:26
	→ richardbronosky joined	19:26
	← roop left	19:26
Kipperoo	I mean a format that a merge tool will be able to reasonably work with.	19:26
jast	that depends more on the merge tools you can find than on git :)	19:26
	I don't really know any merge tools for rich text formats	19:26
bremner	that is not the problem. The problem is you want wysiwyg	19:26
Kipperoo	Word itself has a merge tool	19:27
	← ntoll left	19:27
Kipperoo	But it doesn't need to be wysiwyg, just a text format would be fine... other than HTML :)	19:27
jast	yes, but can it use three different copies of a file in order to do the merging?	19:27
	→ archis joined	19:27
Kipperoo	No, certainly not, just two. And poorly. haha.	19:27
fr0sty	meGenius: how did you clone?	19:28
	(what command)	19:28
jast	well, there are tons of markup languages	19:28
	← archis left	19:28
jast	and all of them suck in at least two significant ways	19:28
meGenius	git clone [link to the repo], fr0sty	19:28
jast	meGenius: what protocol did you use for cloning in the link?	19:28
meGenius	jast: git://	19:29
jast	ah	19:29
theoros	you can't push over git:// iirc	19:29
jast	that protocol is meant for anonymous access	19:29
fr0sty	andre_pl: git doesn't know, or care, what ../ is.	19:29
theoros	meGenius: is it local to your computer?	19:29
	→ bcardarella_ joined	19:29
meGenius	theoros: the bare repo, nope!	19:29
jast	that means that you can enable anonymous push, so that anyone can push to the repository without security checks	19:29
	or you can switch to a different protocol, e.g. ssh://	19:29
meGenius	jast: the git hosting is not mine, it's provided by another group	19:30
jast	well, they probably have some other protocol available	19:30
meGenius	ssh:// is available too	19:30
jast	yeah, use that	19:31
	you can even change the URL after the fact now, without cloning again	19:31
	meGenius is trying ssh	19:31
fr0sty	andre_pl: and you can give a [<directory>] option to git clone to put it in whatever directory you want.	19:31
	← Kipperoo left	19:31
jast	just edit .git/config and put in the new URL where the old one is	19:31
meGenius	okay, jast	19:31
	hey, it works :D	19:32
	← richardbronosky left	19:32
	← mattwynne left	19:32
meGenius	git is more pwoerful, & a trouble-maker version system :P	19:32
	← g0bl1n left	19:32
	← bcardarella left	19:33
	bcardarella_ → bcardarella	19:33
	→ hipe joined	19:35
andre_pl	fr0sty: thanks.	19:35
	→ robhawkes_ joined	19:36
	← robhawkes_ left	19:37
	← fod left	19:37
	→ Jippi_mac joined	19:38
jacobat	So pretty: git log --pretty="%C(yellow)%ad %Cred%h%Cblue (%an)%Cgreen%d%Creset %s" --date=short	19:38
	:)	19:38
	→ robhawkes joined	19:38
	→ Cstep4 joined	19:39
	→ xiong joined	19:41
	→ LongBeach joined	19:41
	← catalis left	19:41
	← robhawkes left	19:42
	← rolfb left	19:43
	← Radar left	19:44
	← murz left	19:44
	→ patthoyts joined	19:45
	← RORgasm left	19:45
	← Cstep4 left	19:45
	→ RORgasm joined	19:45
	→ tvw joined	19:46
	→ marshall joined	19:46
	← benjoldersma_ left	19:47
SeveredCross	jacobat: That is pretty	19:48
	← simplechat_ left	19:48
jast	I feel pretty, oh so pretty, ... :}	19:48
	I would just use tig instead, but ymmv	19:49
	← bhearsum left	19:50
	← mfacenet left	19:51
	← erlnoob left	19:52
	← Schmallon left	19:53
	← vu3rdd left	19:53
	→ matty[pac] joined	19:55
	→ albel728 joined	19:55
matty[pac]	Is there a way in git to look for content that was committed but isn't there anymore. The way I did it was to go "git log -p \| grep -B10 -A10 'highlight'" to search for highlight. Is there an easier way?	19:55
	→ richardbronosky joined	19:55
whaley	jtauber: ping?	19:56
wereHamster	matty[pac]: maybe you want git log -Shighlight	19:56
matty[pac]	Awesome! Thanks were =)	19:57
jtauber	whaley: pong	19:57
	→ sftp joined	19:57
whaley	jtauber: /query ok?	19:57
jtauber	sure	19:57
	→ comawhite_ joined	20:00
	← disappearedng left	20:01
	← comawhite left	20:01
	→ thiago_home joined	20:01
	← oriba left	20:02
	← sonnym left	20:04
	← villev left	20:04
	→ oriba joined	20:04
	← bcardarella left	20:05
	← workmad3 left	20:06
	← RORgasm left	20:06
	→ RORgasm joined	20:06
	→ Radar joined	20:09
dominikh	jacobat: git log --graph --pretty=format:'%Cred%h%Creset -%C(yellow)%d%Creset %s %Cgreen(%cr) %C(bold blue)<%an>%Creset' --abbrev-commit --date=relative	20:14
	jacobat: that's what I use ;)	20:14
	→ AkumaStreak joined	20:15
albel728	dominikh: 0_0	20:15
	← dgovinda left	20:17
	← richardbronosky left	20:19
	→ bamccaig joined	20:21
	← albel728 left	20:21
	← \|albel727\| left	20:21
mkiwala	That's not a bad setup, but one would be better to create some configuration so that the whole command doesn't have to be typed each time.	20:21
	The guy who sits next to me does just that.	20:21
	→ \|albel727\| joined	20:21
	← sh1mmer left	20:21
dominikh	aka an alias :)	20:21
doener	mkiwala: like ""git config alias.lg log --graph --pretty=format:'%Cred%h%Creset -%C(yellow)%d%Creset %s %Cgreen(%cr) %C(bold blue)<%an>%Creset' --abbrev-commit --date=relative"?	20:21
bamccaig	Is there a good reason for the individual tools (e.g., git-add) to be kept out of the PATH? On Fedora, it seems those are installed in /usr/libexec/git-core/, and `git` itself is installed in /usr/bin/git. I find it nice to use bash's auto complete to find commands. :P	20:22
doener	hm, add --global to that call	20:22
	bamccaig: use the supplied bash completion script ;-)	20:22
bamccaig	doener: ?	20:22
doener	bamccaig: then you can also complete e.g. branch names	20:22
	<HEAD:contrib/completion/git-completion.bash>	20:23
Gitbot	[git git-completion.bash]: http://tinyurl.com/yg6y8cq [blob]	20:23
	→ yxa joined	20:23
bamccaig	doener: How do I use that? :\	20:23
doener	bamccaig: just source it	20:24
bamccaig	Ah, nevermind, it says so in the file. :)	20:24
doener	bamccaig: note that you'll want to use the completion script for your git version, so it doesn't complete stuff that doesn't make sense in your setup	20:25
	← marshall left	20:25
doener	bamccaig: I have no clue where fedora installs that thing, though	20:26
mkiwala	doener: very similar to that. I know he also has a --branch in there, but the format spec might be a little different...	20:27
	← andre_pl left	20:27
	→ rraasch joined	20:27
bamccaig	Good point. Apparently it's at /usr/share/doc/git-1.7.2.1/contrib/completion/git-completion.bash.	20:28
doener	mkiwala: that was just an example of how to setup an alias so one doesn't have to retype the whole thing all the time	20:28
bamccaig	Though that path will probably change when Git's version does...	20:28
	→ sh1mmer joined	20:28
bamccaig	I guess I could parse the version from git --version?	20:28
	→ cbeust joined	20:29
	→ richardbronosky joined	20:29
	rubydiamond → zz_rubydiamond	20:29
bamccaig	There's no git-completion found on my Gentoo box. :(	20:32
	← icwiener left	20:33
	→ pigdude joined	20:34
pigdude	how do I list remote tags? I pushed tag 0.1.1 but checkout says that the pathspec matches nothing	20:35
	→ iravid joined	20:35
pigdude	oh nevermind it was an issue in my script	20:35
bamccaig	Is it normal to have duplicates in `git log master..origin/master` ?	20:41
	← RORgasm left	20:41
	← reflectivedev left	20:41
	← r4tune left	20:41
	→ RORgasm joined	20:42
	→ montylounge joined	20:42
	→ reflectivedev joined	20:43
	← hearit left	20:43
	← iravid left	20:46
	← eydaimon left	20:46
	zorzar → zorzar_away	20:46
	← xiong left	20:47
	← rseifert left	20:49
	← olaf left	20:51
wereHamster	bamccaig: no, but it's possible	20:51
bamccaig	Maybe if I add the doc USE flag for git in Gentoo it might get me the git-completion.bash file...? :\	20:52
	wereHamster: Does it have a specific meaning or does it vary?	20:52
	→ brendon_work joined	20:52
	← brendon_work left	20:52
	→ brendon_work joined	20:53
wereHamster	bamccaig: did you use cherry-pick?	20:53
	← brendon_work left	20:53
	→ iravid joined	20:53
	→ brendon_work joined	20:53
bamccaig	No.	20:53
	← brendon_work left	20:53
wereHamster	sorry, upstream must have messed up, not you	20:53
	→ brendon_work joined	20:54
	← brendon_work left	20:54
jast	could've been a weird rebase, too	20:54
bamccaig	I just created a repository with existing code, cloned it, then edited the same file in both to see the result of a merge conflict (to compare to how Mercurial handled it).	20:54
	→ brendon_work joined	20:54
bamccaig	I then tried pulling from each.	20:54
jast	or a merge after history altering	20:54
wereHamster	in any case, you'll get duplicates if you apply the same commit to two different commits. You'll end up with two commits with the same diff and commit message, but different parents	20:54
	Morphous → Amorphous	20:55
	← fluzios left	20:55
	→ s0ber joined	20:55
	→ flugsio joined	20:56
	← durarara left	20:56
bamccaig	I see.	20:58
	I'm not really sure what happened... I think I pulled from A into B and from B into A, but haven't committed the merge (which conflicted for both, intentionally).	20:59
	Somehow, from A, I'm seeing A's latest commit twice (and not in chronological order) with git log..	20:59
	Somehow, from A, I'm seeing A's latest commit twice (and not in chronological order) with `git log master..origin/master` (where origin is B).	20:59
bremner	bamccaig: sometimes gitk is more helpful than git log	20:59
	→ jeff_ joined	21:00
	jeff_ → Guest27034	21:00
fr0sty	or: 'git log --graph --oneline'	21:00
	Guest27034 → marshall	21:00
bamccaig	bremner: Isn't that a GUI (i.e., TK) front end? NEVAR!	21:01
	← canavar left	21:01
	bamccaig installs gitk.	21:01
	← montylounge left	21:01
	hagebake → hagabaka	21:03
	← hugoxrosa left	21:03
	→ Slix` joined	21:04
	← masterkorp left	21:04
basicxman	bamccaig: No!! Terminal-fu FTW	21:04
	← meGenius left	21:04
	← henkboom left	21:04
	→ bauruine joined	21:06
	→ henkboom joined	21:07
	→ Blackguard joined	21:08
bamccaig	Yeah, I doubt I'll use that often. Graphical representations are helpful to visualize branching, etc... And make sense of it when you're lost. :P That's about it....	21:08
	Anyway, thanks for the help #git. I shall depart.	21:08
	← rayners left	21:08
	← emmanueloga left	21:09
	← marshall left	21:10
	← iravid left	21:10
	← bauruine left	21:11
	← ajpiano left	21:11
	→ iravid joined	21:11
	→ J3roen joined	21:12
	← cannonball left	21:12
	comawhite_ → comawhite	21:13
	← jjuran left	21:14
	← redbaritone left	21:14
	→ patrick99e99 joined	21:15
J3roen	Hi there! Question; I have a base CMS installation (sandbox) in Git with all my tweaks and stuff. I would like to base all other (new) client sites on this install and whenever I edit the sandbox, merge with the client website. What is the best way to implement this?	21:15
	→ level09 joined	21:16
patrick99e99	Howdy everyone... I have just realized that file names changes are not being added to my git status...... For example if I do, mv x y; .. and then git add .; when I do git status, I do not see this change appearing, and so I am finding when I push my code, that change is not happening on the server	21:16
	how can I correct this/	21:16
	?	21:16
selckin	stage the delete aswel	21:17
patrick99e99	selckin, was that for me?	21:17
selckin	no the guy	21:17
IslandUsurper	patrick99e99, is the new file getting added though?	21:17
	← unreal left	21:18
drizzd_	patrick99e99: git add -A	21:18
J3roen	patrick99e99: better to use git mv, and let git handle the moving.	21:18
patrick99e99	IslandUsurper, no... but actually what's happening is I had file names with upper case letters, that I changed to lower case... and they arent showing up	21:18
	← tstclair left	21:18
	← sgh left	21:18
patrick99e99	ah so I should move them back to have upper case names	21:18
	and then do git mv ?	21:18
J3roen	That's what I would do, but then again, I'm a git newbie :)	21:19
selckin	no, just stage the delete and stage the new new file	21:19
	→ xiong joined	21:19
mfischer	anyone know the way to get shortlog to not undo linebreaks in commit messages?	21:19
J3roen	selckin: Will that not destroy the history of the file?	21:19
	→ unreal joined	21:19
selckin	no.	21:19
patrick99e99	selckin, I am a newbie too.. I dont know what "stage the delete" means	21:20
J3roen	selckin: What's the point of git mv then?	21:20
selckin	tell git the file has been deleted, git rm, or use the git add -u/-A magic	21:20
	J3roen: lazyness	21:20
	← doener left	21:21
	→ canavar joined	21:22
patrick99e99	selckin, could you give me an example of what you'd type exactly if you had a file called Blah.txt that you wanted to be blah.txt ?	21:22
J3roen	ok.. I'd rather be lazy then tired tough :)	21:22
selckin	mv foo bar; git status; "hey look it says untracked bar, and foo seems to be gone"; git rm foo; git add bar; git status; "hey look it detected that i renamed foo to bar"; git commit	21:22
patrick99e99	ohh	21:22
	ok	21:22
	that is simple	21:22
	cool	21:22
	← Mage left	21:22
	→ Mage joined	21:23
	← richardbronosky left	21:23
J3roen	selckin: any idea how I would go about my question?	21:23
	selckin: In future I would still advice git mv btw :)	21:24
selckin	i couldn't care less	21:24
patrick99e99	selckin, well that was great except when I typed git rm Blah.txt	21:24
	it deleted blah.txt	21:24
	and now the file I want to add is gone	21:25
selckin	patrick99e99: you're on windows. go yell at microsoft	21:25
patrick99e99	I am on mac os x	21:25
selckin	they have case insensitive file systems too?	21:25
	go yell at them too	21:25
IslandUsurper	patrick99e99, then yell at whoever gave you a case-insensitive file system	21:25
patrick99e99	hmm	21:26
J3roen	lol..	21:26
	→ erlnoob joined	21:26
	→ ciampix joined	21:26
jast	seems like by default, OS X is case-insensitive, though it does remember case	21:26
	→ adam[pac] joined	21:26
J3roen	selckin: but you couldn't care less about what btw?	21:27
	→ sgh joined	21:27
selckin	J3roen: what you recommend	21:27
J3roen	selckin: well I don't also.. Just looks easier to me. It wasnt a bitchslap.	21:28
selckin	i just develop and rename stuff when i want, when i'm ready to commit something i go to my terminal and do it, i'm not gonna go git mv everytime when my IDE refactoring support will rename everything automaticly	21:28
	git is smart enough to figure it out, not a pain like svn	21:29
	← IslandUsurper left	21:29
J3roen	selckin: of course, I wouldn't also! But when you delete from the command line... That's what I meant...	21:29
	← pigdude left	21:30
selckin	in this case it might be the only way to rename due to using a bad filesystem	21:30
	but i'd fix the filesystem	21:30
	← skoop left	21:31
	→ airborn joined	21:31
	→ mattikus joined	21:33
	→ ben_alman joined	21:34
ben_alman	I stashed a file, made some changes to the original, and committed.. then i did a stash pop and there were conflicts, so i manually merged the file and i get this http://gyazo.com/a540074dd1e90c903937334b1393d4c4.png	21:36
	what is that diff telling me?	21:36
jast	probably that the file permissions are different	21:36
	commands that cause conflicts generally tell you more about the nature of the conflicts while they encounter them	21:37
	← Nemurenai left	21:37
ben_alman	it's got the same permissions as everything else	21:37
	→ canavar_ joined	21:37
selckin	ben_alman: try git diff --cached	21:37
	← ciampix left	21:37
ben_alman	"* Unmerged path jquery.ba-bbq.js"	21:38
	← canavar left	21:38
jast	oh, you've already fixed the file?	21:38
	then just git add it	21:38
	→ canavar__ joined	21:38
selckin	did you fix the conflicts allready?	21:38
ben_alman	yeah	21:39
	but it's not ready to commit, i still have further changes	21:39
selckin	git reset HEAD .	21:39
ben_alman	what does that do?	21:39
selckin	unstage it	21:39
jast	yeah, but so what? you can stage it anyway, can't you?	21:39
	it's not like you're forced to commit it straight away	21:39
ben_alman	i dunno	21:39
	i'm far from a git pro	21:39
selckin	normally when you pop a stash, it's not staged, but when it conflicts, all the changes from the stash will still be staged	21:40
ben_alman	all i know is that when i did a git stash pop and manually merged the changes, this is where i'm at	21:40
jast	git add just stages the current version for the next commit, but you can always restage a different version later	21:40
	← canavar_ left	21:40
	→ canavar joined	21:40
ben_alman	hmmn	21:40
	let me try	21:40
selckin	so now git reset HEAD .	21:40
ben_alman	ok	21:40
selckin	and all will be well	21:40
jast	reset HEAD . actually has the same effect by unstaging everything	21:40
	→ hohoho joined	21:40
ben_alman	hmmn....	21:41
	that's weird	21:41
	i did git reset HEAD .	21:41
	← RORgasm left	21:41
selckin	and?	21:42
jast	and then his head exploded	21:42
ben_alman	now when i do a git diff of the file, it says i'm adding stuff that was actually committed in my last commit	21:42
	i'm just double checking	21:42
	yeah, that's weird	21:43
	→ canavar_ joined	21:43
	→ Sigma joined	21:43
	← canavar left	21:43
ben_alman	it's diffing as if my last few commits never happened	21:43
	← canavar__ left	21:43
ben_alman	even though i see them on github	21:43
jast	do they still show up in git log?	21:44
ben_alman	yes	21:44
	was just checking that	21:44
jast	hm.	21:44
selckin	prove it	21:44
ben_alman	sec	21:44
	selckin puts 10$ on git diff -w	21:44
ben_alman	what's that	21:45
	http://gyazo.com/55bae2903149f3f166a145df1b70385a.png	21:45
	http://gyazo.com/5d2c2a5678b8ebc402e124a61e4c4ade.png	21:45
	http://github.com/cowboy/jquery-bbq/blob/master/jquery.ba-bbq.js	21:45
	← hohoho left	21:45
ben_alman	i can guarantee that i have no idea what i'm doing.	21:46
	← stringoO left	21:46
	← friskd left	21:46
selckin	well you stashed the file with 20/ ?	21:46
	so now you have it back like that	21:46
	← Danielpk left	21:46
ben_alman	what is 20/	21:46
selckin	the dates? like the only changes in your diff	21:47
ben_alman	er	21:47
	← Jippi_mac left	21:47
ben_alman	ok, let me backtrack	21:47
	this diff is between the file's current unstaged state and... ?	21:47
selckin	HEAD	21:47
ben_alman	ok. what is HEAD, the most recent commit?	21:48
	or how can i find out what HEAD is at	21:48
selckin	of the current branch	21:48
	usually the top thing in git log	21:48
	→ WaterRatj joined	21:48
WaterRatj	I installed git but doen't reconize the git command	21:48
ben_alman	ok so the top thing in git log is the github link i just sent	21:48
	→ richardbronosky joined	21:48
ben_alman	the github link i sent has a date of 8/20	21:49
	← stymo left	21:49
selckin	github is irrelevant	21:49
ben_alman	so why does the diff show that i'm adding 8/20 if it's already 8/20 at HEAD	21:49
drizzd_	WaterRatj: did you install from source?	21:49
selckin	because your commit is at 7/24	21:49
WaterRatj	drizzd_: I used apt-get install git	21:49
	→ t0rc joined	21:50
ben_alman	selckin this is the commit http://github.com/cowboy/jquery-bbq/blob/6c6bade0c415c2080ea59e195e10650039edda46/jquery.ba-bbq.js	21:50
	it's 8/20	21:50
drizzd_	WaterRatj: then you probably just need to type rehash or open a new terminal	21:50
selckin	ben_alman: thats not your head commit	21:50
ben_alman	i made it right before i did the git stash pop (with 1 commit of a separate file intervening)	21:50
	that is the top of my git log	21:51
jast	WaterRatj: in some versions of your distribution, the git package is actually called git-core	21:51
	← rutski left	21:51
	ben_alman is so confused	21:51
jast	and there's something different in the "git" package	21:52
selckin	ben_alman: whatever, too comfusing, and that is not the commit in the screenshot, diffrend sha1	21:52
ben_alman	it's the 2nd commit	21:52
	same sha1	21:52
	the 1st commit, like i said modified a different file	21:52
selckin	so, what is your question?	21:52
WaterRatj	jast: thanks, that worked	21:52
ben_alman	1 min	21:52
jast	ben_alman: what is the exact diff command you're using?	21:54
	→ stringoO joined	21:54
selckin	just git add -p the parts you want, git checkout -- all the date changes you don't	21:54
jast	selckin: the commit IDs are the same, actually	21:54
	looky here: http://github.com/cowboy/jquery-bbq/commit/aa12c92168fe92d051d24c9ab4249286108a692b	21:55
selckin	he said it was head commit, i only compared the first	21:55
jast	that's the same commit as at the top of the log screenshot	21:55
	and it definitely includes the date changes	21:55
	but I'm not going to speculate any further until my question is answered	21:56
selckin	how he got there is probably not usefull anymore now, and the solution involved git add -p file; nnnnnynnynynynynynyn; git checkout -- file; git reset head file	21:57
	selckin goes to sleep	21:57
jast	I'm not sure whether there is actually anything that needs to be solved	21:57
ben_alman	sorry, wife just called me.. 1 sec	21:57
	:P	21:57
jast	sure	21:57
	← AkumaStreak left	21:58
	← WaterRatj left	21:58
	← werdan7 left	21:59
	← bcrouse left	21:59
	→ marioxcc joined	21:59
marioxcc	hi	21:59
ben_alman	so yeah. i don't know what the head commit is, i assumed it was the one at the top of the git log because all i've done since that commit is git stash pop; edit file to manually merge; git reset HEAD; then the diff	21:59
	← loincloth left	21:59
marioxcc	why the blobs objects don't have names equivament to their SHA1?	21:59
ben_alman	so what i'd like to do is get the diff looking more like just the things that have changed since my latest commit	21:59
jast	marioxcc: they do.	21:59
	ben_alman: question repeated, what is the exact diff command that produces the confusing output?	22:00
ben_alman	git diff jquery.ba-bbq.js	22:00
	might it be a weird config setting?	22:00
	it would have to be, right?	22:01
	→ loincloth joined	22:01
jast	does git diff HEAD jquery.ba-bbq.js show any date changes?	22:01
marioxcc	jast: sorry, i mean the output of git cat-file blob	22:01
	← sgh left	22:01
ben_alman	jast nope	22:01
	looks as-expected	22:01
	so it's not diffing against the head?	22:02
jast	marioxcc: where does that output actually show any names?	22:02
ben_alman	wait.	22:02
jast	ben_alman: diff without further arguments diffs against the index	22:02
ben_alman	my git status shows staged files.	22:02
	i thought git reset HEAD . unstaged everything	22:02
marioxcc	jast: git cat-file blob <NAME>	22:02
	← tg left	22:02
marioxcc	that's the "name" of the object, isn't it?	22:02
jast	marioxcc: so what? name can be a sha1 name or any of the shorthands explained in http://git.or.cz/man/git-rev-parse	22:03
	→ tg joined	22:03
	→ Slix`` joined	22:03
	→ bauruine joined	22:03
ben_alman	should i do git reset HEAD . again?	22:03
marioxcc	jast: yeah, but the output of git cat-file blob SHA1 hash to SHA1 if piped to a SHA1 computer	22:03
ben_alman	i don't think i care about what's staged, i just want to work with what's unstaged vs the HEAD	22:04
	→ sgh joined	22:04
jast	marioxcc: can you rephrase that in a way that confuses me less? :)	22:04
	→ unreal_ joined	22:04
	← sh1mmer left	22:06
marioxcc	oh, sorry	22:06
	← Slix`` left	22:06
	← Slix` left	22:06
marioxcc	git cat-file blob 9488cbfe8cb920d0630f9aa077a0c1ca4a798a7f \| sha1sum	22:06
	866f2a4322b245245b9c0b5a74007b2c6be7fe14 -	22:06
jast	ben_alman: I dunno what happened there, but the surefire way to unstage absolutely everything is "git reset" with no additional arguments	22:06
	→ CoinOpeBoy joined	22:06
marioxcc	these are not equal, that's what I mean.	22:06
bamccaig	This is fucked. It's the middle of August and I could put he heat on....	22:06
ben_alman	ok jast and that won't change my files as they are on the hard drive	22:06
	that'll just change what's staged?	22:06
bamccaig	Whoops, wrong channel. :-[	22:06
jast	marioxcc: git hashes differently than the sha1sum binary does	22:06
	marioxcc: git cat-file blob 13245 \| git hash-object --stdin will give you the same sha1 name back	22:07
	→ werdan7 joined	22:07
	← unreal left	22:07
marioxcc	ok	22:07
	unreal_ → unreal	22:07
	→ bousquet joined	22:07
Scotto_	how would i commit only part of the changes in a file?	22:07
marioxcc	yes, it's strange	22:07
	← oriba left	22:07
	← mattikus left	22:07
	← flazz left	22:08
jast	ben_alman: yep. reset never changes working tree files unless you use the --hard switch... and then it kills all unstaged changes, hence the aggressive name for the switch :)	22:08
	Scotto_: git add -p and git commit (without -a), assuming you haven't staged the full file yet	22:08
ben_alman	thanks for the help jast and selckin :)	22:08
	← bauruine left	22:09
Scotto_	i staged it	22:09
	but i can do a reset	22:09
	adamholt → adamholt_away	22:09
	adamholt_away → adamholt	22:09
jast	yeah, probably easiest to unstage the changes in that file, then do something along the lines of what I suggested	22:10
	or you could use the completely interactive stuff in git gui :)	22:10
Scotto_	git gui sucks over samba	22:10
jast	where you can stage/unstage line by line (git add -i allows the same thing but it's slightly more unwieldy)	22:10
Scotto_	:D so im stuck with putty	22:10
	→ sh1mmer joined	22:11
	→ fero joined	22:11
	← mfischer left	22:11
	← Aaaarg left	22:11
	→ jjuran joined	22:12
	→ Gudy joined	22:13
	→ Aaaarg joined	22:14
Gudy	Hi, I have a setup where I have an origin coming from gnome, and I'm pushing to github, I had made some local changes, which were in conflict with my latest pull from the origin, I've commited the merge, however git won't let me push to github anymore	22:14
	with the error message : "! [rejected] master -> master (non-fast-forward)"	22:15
	would anyone know how to fix this ?	22:15
	← stringoO left	22:16
	← alester left	22:16
selckin	did you rebase things?	22:16
	→ cpr420 joined	22:16
Gudy	I guess not, as I don't know what rebase is	22:16
wereHamster	Gudy: faq non-ff	22:17
Gitbot	Gudy: Your push would lose changes on the remote. See https://git.wiki.kernel.org/index.php/GitFaq#non-ff	22:17
jast	did you do anything else between pulling, changing stuff, committing, and trying to push?	22:17
	→ SandCube joined	22:17
jast	chances are that you simply need to pull again, but some commands, e.g. git reset and git rebase, can cause other problems	22:17
Gudy	I've pulled from all	22:18
	the order was :	22:18
	pull from origin, make changes, push to github (weeks ago), then today : pulled from origin, got conflicts, edited the conflicts, did a commit, git detected a "merge commit" on its own, and now I cna't push to github	22:19
jast	the most likely case is that someone else pushed something new while you were busy fixing the conflicts	22:19
	in which case you can simply pull again and then push	22:19
Gudy	I've tried a git pull --all, and it says "up to date"	22:20
SandCube	anyone can point me to a good reading about how to setup and use a repo using git-http-backend (smart http)?	22:20
jast	okay. do a git fetch and look at the differences between upstream and local, e.g. gitk yourbranch origin/yourbranch	22:21
	→ ablemike joined	22:21
jast	perhaps you'll see something that helps	22:21
ablemike	git checkout staging	22:21
	Checking out files: 100% (68/68), done.	22:21
	Switched to branch 'staging'	22:21
	→ unreal_ joined	22:21
jast	SandCube: you mean the examples in its manpage? :P http://git.or.cz/man/git-http-backend	22:21
ablemike	then:	22:21
	.git/hooks/post-checkout:14: unexpected return (LocalJumpError)	22:21
	?	22:21
	wtf does that mean?	22:21
	→ KragenSitaker joined	22:22
jast	ablemike: it means that something is broken in that hook. that hook probably comes from a source other than git itself.	22:22
ablemike	gitosis	22:22
	crap	22:22
jast	gitosis is unmaintained and unsupported software, I suggest switching to gitolite if you can't figure out what's going on	22:23
	← sgh left	22:23
Gudy	jast: I'm not working on my own branch, maybe that's the problem ...	22:23
jast	Gudy: "yourbranch" could just as well stand for "master"	22:23
	ablemike: I'd help figuring out the issue... but I despise python. sorry. :)	22:23
	→ Slix` joined	22:24
	→ ciampix joined	22:24
	← unreal left	22:24
	unreal_ → unreal	22:24
ablemike	heh	22:24
	this is the first time gitosis has shat on me	22:24
	i too despise python	22:24
	→ picardo12 joined	22:25
jast	so far I've been putting off switching away from gitosis myself, so I feel for you	22:25
	← Dave^\|\| left	22:26
KragenSitaker	Hi. I have a git repository with one free-software directory and the rest proprietary (and confidential). I'd like to push the free-software version to github and cease developing it in the shared tree. Is git filter-branch the right tool to use to make the new repo?	22:26
	It looks like it can do the job but I'm not familiar with it. Is it the right thing or should I look elsewhere?	22:26
	← Heimidal left	22:27
	← adamholt left	22:27
jast	filter-branch is usable for creating a filtered version of an existing repository, yes	22:27
KragenSitaker	I'd like to preserve the history of this subdirectory but not the other dirs.	22:27
jast	filter-branch will automatically discard all commits that only contain files that have been filtered out	22:28
	← itrebal left	22:28
KragenSitaker	cool	22:28
jast	you'll want the --subdirectory-filter option	22:28
KragenSitaker	Thanks, that's exactly what I want! I missed it when I skimmed the man page the first time	22:29
SandCube	jast: I mean a walkthrough, a real life example, something more specific	22:29
	→ engla joined	22:29
SandCube	jast: sorry for late reply	22:29
	→ adamholt joined	22:29
	← engla left	22:30
	→ engla joined	22:30
jast	SandCube: dunno. I haven't really looked into setting it up yet. personally I could probably get by with the information in the manpage.	22:30
	→ tobiassjosten joined	22:30
	← rraasch left	22:33
	→ halfline2 joined	22:33
KragenSitaker	Nice, --subdirectory-filter even moved the contents of the subdirectory to the top level	22:33
	if I clone the resulting repo, will my clone contain only things that are referenced from the history of a head?	22:34
	→ _numbers joined	22:34
jast	yep. it's like someone anticipated exactly the situation you found yourself in! wow. the git developers must have superpowers.	22:34
	← _numbers left	22:34
KragenSitaker	Or will it still contain the old confidential blobs, just unreferenced?	22:34
	ezzieyguywuf → ezziey_drinking	22:34
jast	the filtered repository will, but if you push into a fresh repository, they will be gone there	22:34
KragenSitaker	I know. the superpowers of community never cease to amaze me	22:34
	← richardbronosky left	22:35
jast	or you can clone (take care to use a file:// URL if it's a local clone, otherwise git just reuses the pack files)	22:35
	or you can follow the elaborate list of steps to clean a repository post filter-branch, as explained in its manpage	22:35
	← adam[pac] left	22:36
	→ kelvie_ joined	22:37
	→ picardo12_ joined	22:38
	← picardo12 left	22:38
	picardo12_ → picardo12	22:38
Scotto_	wow ive never had save() add slashes for single quotes until now	22:38
	is something wrong? or is this somethin i have to disable?	22:39
	← parasti left	22:39
jast	I'm pretty sure I have no idea what you're talking about	22:39
Scotto_	ah n/m its because of Sanitize::clean	22:39
	dunno why i even used Sanitize::clean on $this->data	22:40
	is there a standard way of protecting data when saving to the db?	22:40
jast	what language are we talking about?	22:41
floppyears	hi guys, I know that with git some projects store their whole changelog in the tag's commit message	22:41
	what is the best way of using the commit message? Storing everything in one line or keeping the commit messages to 80 character lines?	22:41
jast	what is "everything"?	22:41
Scotto_	jast: lol	22:41
	jast: wrong chan, im sure youve been there before	22:42
jast	ah, that explains a thing or two :)	22:42
Scotto_	yeah	22:42
	→ scarabx joined	22:42
Scotto_	it probably means its time to go home	22:42
jast	it does sound more like Perl than PHP, I suppose	22:42
	← KragenSitaker left	22:42
	→ msieradzki joined	22:42
jast	at this point the PHP folks would probably have some kind of complex framework thing instead of a simple function call	22:43
	→ sgh joined	22:43
	← picardo12 left	22:43
	← brendon_work left	22:43
	← yxa left	22:43
	← pflanze left	22:44
	← davidw left	22:44
marioxcc	bye	22:44
	← marioxcc left	22:44
	→ EricInBNE joined	22:44
	← sergelamoss left	22:46
	→ bhosie joined	22:46
	→ whatdoesitwant joined	22:46
	← TeckniX left	22:47
	→ KragenSitaker joined	22:48
	← bousquet left	22:49
	← erlnoob left	22:52
	→ Nickeeh joined	22:54
	→ [M]ax joined	22:54
	→ phantomcircuit joined	22:55
	→ tomsdale_ joined	22:55
	← tomsdale left	22:56
	← nicoulaj left	22:56
	→ jason237 joined	22:58
	← bhosie left	22:59
hagabaka	is there a way to grep the history of a file for changes involving a line?	22:59
patrikf	hagabaka: git log -S	23:01
	← lamont left	23:01
	← Titosemi left	23:02
	→ lamont joined	23:04
hagabaka	thanks	23:05
	→ zenazn_ joined	23:05
	← zenazn left	23:05
	← jrmuizel left	23:05
	← devurandom left	23:07
	← frakturfreak left	23:07
	→ devurandom joined	23:08
	terinjokes is having a "why is this in mercurial" moment	23:08
	← herdingcat left	23:08
	← matty[pac] left	23:11
	← fowlduck left	23:13
	← halfline2 left	23:13
	← tizbac left	23:14
	← chronos left	23:25
	→ _numbers joined	23:26
	← _numbers left	23:26
	→ hal_5_ne2 joined	23:27
	← spaceonline left	23:29
	← Ryback_ left	23:30
	← phantomcircuit left	23:30
	← akahn left	23:31
	← jjuran left	23:32
	→ sshc_ joined	23:38
	← level09 left	23:38
	← kogent left	23:39
	→ Schmallon joined	23:40
	← sshc left	23:41
	← hal_5_ne2 left	23:43
	← mbroeker left	23:44
	→ hal_5_ne2 joined	23:44
	← airborn left	23:46
	→ mbroeker joined	23:46
	→ hebz0rl joined	23:47
	→ artefon joined	23:48
	← skeeziks left	23:48
	→ mdel joined	23:48
	← amerine left	23:49
	→ Theravadan joined	23:52
	← pdelgallego left	23:52
	→ rgr joined	23:52

Logs Search ←Prev date Next date→

Channels Documentation