IRCloggy #git 2008-02-17

Provider of IRC logs since 2005.

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

	2008-02-17
	→ fultilt joined	00:05
aazar	hey guys, newbie question. Why does git tell me that it "cannot rebase: your index is not up-to-date". How do I make my index up to date?	00:08
	← madewokherd left	00:09
	← jdempsey left	00:09
	→ jdempsey joined	00:10
tpope	aazar: commit your changes, or stash them	00:11
aazar	tpope: so git stash, git rebase, then git stash --apply?	00:11
	← tjafk2 left	00:11
tpope	git stash apply, yeah	00:12
aazar	tpope: thanks. Will git overwrite the new versions of the files when I do git stash apply? or merge?	00:13
tpope	merge	00:13
aazar	thanks, appreciate the help	00:13
	tpope: actually another question, why doesn't git rebase do that automatically?	00:14
tpope	because nobody's implemented it?	00:15
aazar	ok. I was wondering if there was any more profound reason behind it.	00:15
tpope	I don't know	00:16
robinr	it maybe hard for stash to apply (hard conflicts) after rebase so the user better be alerted there is a potential problem	00:18
	→ mtcx joined	00:18
aazar	robinr: why would it be any harder to rebase what's in the stash than any other commit?	00:20
	← glommer left	00:23
	← mtcx left	00:23
	← trapni_ left	00:29
	← \ask left	00:31
	→ jdempsey_ joined	00:38
	← jdempsey left	00:38
	→ antirobotrobot joined	00:39
	→ pete__c joined	00:39
antirobotrobot	hi, i just made some changes to my master branch, but they're getting a little intense. i'd like to commit my change instead to a new experiemental branch. i've already created one - how do i switch to it without losing my changes :/	00:40
robinr	if the new branch is on top of your head just checkout it	00:47
antirobotrobot	i haven't commited yet	00:47
robinr	if not checkout -m will perform a threeway merge that will hopefully succeed	00:47
antirobotrobot	if i checkout it will switch to my previous version	00:48
robinr	it won't destroy your changes unless you force it to	00:49
	just git checkout otherbranch to start with	00:50
antirobotrobot	i'm pretty sure this overwrites all of my files	00:50
	oh, no okay	00:51
	← SuttoL left	00:52
robinr	checkout safe, checkout -m mostly safe but merge may create conflicts, checkout -f destroys uncommitted changes	00:53
antirobotrobot	okay - thanks	00:53
xyzzy42	is there a way to grep across revisions? For instance tell me what in what commit a given regex first appeared in?	00:55
	← jdempsey_ left	00:57
context	log supports that i think	00:58
	→ langenberg joined	00:59
context	yes	00:59
	xyzzy42: you can do that with git-log	00:59
robinr	sort of -S detects when the number of occurrences of a fixed string changes	00:59
	a.k.a pick-axe	00:59
context	man git-log read the very first couple lines of text, then read man git-rev-list	01:00
	← Yuuhi left	01:02
	→ jdempsey joined	01:02
	← Leonidas left	01:03
	← toxx_ left	01:07
	← jdempsey left	01:09
antirobotrobot	i want to get one file from a branch into my master	01:16
	it exists in master but has no local changes	01:16
	→ janm_ joined	01:17
	→ samgranieri joined	01:18
	← kelvie left	01:19
	→ thorat joined	01:19
	← AtharHameed left	01:20
	→ CareBear\ joined	01:20
CareBear\	hello! How can I tell which commit contains a given source line?	01:21
capisce	git annotate	01:22
CareBear\	Nice! Thanks! Is there a gitweb trick as well?	01:23
	guess not - now to git clone Linux. :p	01:25
	← janm left	01:28
	← langenberg left	01:29
	← fdr- left	01:30
	← reval left	01:31
	→ jdempsey joined	01:31
	← jdempsey left	01:31
	→ jdempsey joined	01:32
	→ robfitz_ joined	01:32
	→ bryanl joined	01:33
	→ fdr- joined	01:35
	→ Ramblurr joined	01:38
	→ kelvie joined	01:39
	← moya left	01:40
	← robfitz left	01:49
	← robinr left	01:53
bobesponja	hi	01:56
	can I create hooks in perl or ruby? and how do I tell git if they are post or pre (etc) commit hooks?	01:57
gitster	They have defined names and semantics. See Documentation/hooks.txt. Any executable would be usable as a hook --- you could even do one in C.	01:59
bobesponja	gitster: ok thanks, yes I read hooks.txt but what's the semantic? post-commit-<my hook>?	02:01
	or do I put all my post commit hooks in post-commit? I guess that must be it	02:02
	← kukks left	02:09
	← Ademan left	02:21
	→ ry_ joined	02:30
	→ Ademan joined	02:36
	→ lord_sauron joined	02:41
	← jdempsey left	02:42
	← antirobotrobot left	02:46
	← pete__c left	02:51
	← deitarion left	02:52
	→ moya joined	02:57
	→ sNov joined	02:57
	← sNov left	02:58
	→ pete__c joined	03:02
	→ FunkeeMonk joined	03:07
	← bobesponja left	03:36
	← moya left	03:39
	← zed_ left	03:40
	→ patnakajima joined	03:45
	← patnakajima left	03:46
	→ patnakajima joined	03:47
	→ bobwhoops joined	03:47
	← patnakajima left	03:49
	← thorat left	03:50
	← lord_sauron left	03:51
	← samgranieri left	03:56
	→ jdempsey joined	03:58
	→ jbms joined	03:59
	→ moya joined	04:03
	← lamont left	04:10
	← FunkeeMonk left	04:33
	→ doener joined	04:37
	← mithro left	04:49
	← doener_ left	04:55
	→ lamont joined	05:00
	→ srid joined	05:01
	← eternaleye left	05:05
	← jdempsey left	05:18
	← xyzzy42 left	05:23
	→ unreal_ joined	05:33
	← unreal left	05:33
	unreal_ → unreal	05:34
	→ eternaleye joined	05:36
	← Ingmar left	05:48
	→ Ingmar joined	05:49
CareBear\	thanks for the help! bye.	05:51
	← CareBear\ left	05:51
	→ CareBear\ joined	06:09
CareBear\	back for more :)	06:09
	Can anyone explain that a commit to linux-2.6.git dated before the Linux v2.6.15 commit isn't included in the -2.6.15 tarball?	06:09
DrNick	it was merged after v2.6.15	06:10
CareBear\	merge != commit ?	06:11
	Hm. I see.	06:11
DrNick	well, no, merges are just commits with multiple parents	06:11
CareBear\	The commit date is from when it was committed the first time - not when it was pulled into Linus' tree?	06:12
DrNick	the point I was making is that this predated commit was actually committed at that date prior to the v2.6.15 tag	06:12
	however it wasn't merged into linux.git's master until after the v2.6.15 tag	06:12
CareBear\	nods	06:12
	How can I find out when it was merged?	06:12
	(Ideally which was the next release, but I'll settle for a date. :)	06:13
paakku	possibly you could use git describe --contains	06:14
CareBear\	Excellent.	06:15
	2.6.16-rc1	06:15
	(followed by noise?)	06:15
	v2.6.16-rc1~169^2~16^2~25^2~6	06:16
	→ FunkeeMonk joined	06:16
paakku	it shows the ancestry path from v2.6.16-rc1 to the commit you asked for	06:17
doener	CareBear\: so you're looking for platform_device_del()? ;-)	06:17
CareBear\	doener : Correct. :)	06:17
paakku	the syntax is documented at git-rev-parse	06:17
CareBear\	paakku : Ok!	06:18
loops	CareBear\, the ancestry info actually specifies the exact commit.. so if you say git show "v2.6.16-rc1~169^2~16^2~25^2~6", you will see the [PATCH] Driver Core: Add platform_device_del() commit.	06:21
DrNick	v2.6.16-rc1~169 is 169 commits before v2.6.16-rc1, and v2.6.16-rc1~169^2 is the second parent of that commit (i.e. 2.6.16-rc1~169 is a merge)	06:22
	and then it goes from there	06:22
loops	and even though that example is rather ugly.. sometimes it's a friendly way to talk about commits in email etc..	06:23
DrNick	93ce3061be212f6280e7ccafa9a7f698a95c6d75 (or 93ce306) is much easier to talk about	06:24
loops	in this case there's no doubt bout that DrNick	06:25
CareBear\	I knew the id, just new to this distributed thing. :)	06:26
	I really like git though.	06:26
	→ patnakajima joined	06:33
	→ spearce joined	06:33
	← patnakajima left	06:34
	→ patnakajima joined	06:34
	→ [RIT]Rawn027 joined	06:40
	← patnakajima left	06:41
	→ patnakajima joined	06:41
	← ry_ left	06:43
	← cwillu left	06:45
gitster	I wonder how you would pronounce ~169^2~16^2~25^2~6 in English ;-)	06:48
spearce	"damn old"	06:49
gitster	Heh.	06:49
	I am committed to rewind and rebase "next", by the way.	06:49
spearce	yea, i saw that. go for it.	06:49
gitster	I need to update MaintNotes, as 1.5.4.2 builds...	06:49
	← [RIT]Rawn027 left	06:51
DrNick	sixth ancestor of the second parent of the 25th ancestor of the second parent of the 16th ancestor of the second parent of the 169th ancestor of v2.6.15	07:00
	← patnakajima left	07:00
	→ patnakajima joined	07:00
spearce	aka "damn old"	07:00
gitster	Gaah. Can you recite it with your eyes closed?	07:00
DrNick	actually, ^2~ is probably a common enough idiom that it might deserve it's own syntax	07:00
loops	yeah.. replace ^2~ with "m" for merge	07:01
spearce	please don't	07:01
loops	heh	07:01
	okay.. damn old it is	07:01
gitster	v2.6.12~4mmmmmmmmmm4?	07:01
spearce	as it is we have a line-noise revision syntax	07:01
loops	gitster, no.. you'd only replace ^2~	07:02
DrNick	should ask the Perl people for advice	07:02
spearce	yea, then people would want v2.6.12~4m{32}~4 or something. :)	07:02
DrNick	v2.6.15~169m16m25m6	07:02
loops	DrNick, right.. that's what i was thinking	07:02
spearce	then we'd need 169m = 2h. :)	07:02
DrNick	ITYM 120m	07:02
	→ tarbo_ joined	07:03
spearce	sorry, 169m = 2h49m.	07:03
	yet another long day at day-job.	07:03
	← Cerebon left	07:06
Ilari	spearce: I have seen git-gui crash on startup with some (sane-looking) LC_ALL values.	07:06
spearce	Ilari: that's never good.	07:07
	i basically run with no LC_* stuff set, so i'm picking up C/stupidamerican locale. :)	07:07
	any backtraces from the crash? or is it just disappearing on you?	07:08
	any specific LC_ALL values that i can try to test here?	07:08
Ilari	spearce: I need to build the newest version...	07:08
Mikachu	is there anything to build?	07:11
spearce	sure, there are many open source programs available. :)	07:12
Mikachu	i meant for git-gui :)	07:12
	→ samgranieri joined	07:12
spearce	it does a short transform that it does to generate the tclIndex and embed its install location into the main script.	07:12
	on the mac it does a bit more as it creates a full .app bundle to get the icon in the menu bar and dock right.	07:12
Mikachu	ah	07:12
spearce	but yea, given that its pure tcl/tk there isn't much to "build"	07:13
	although future git-gui's may have more. i have a pet project that translates C to Tcl.	07:13
	← tarbo left	07:13
Mikachu	that sounds a little bit crazy	07:13
Ilari	spearce: "Error in startup script: Error: No word lists can be found for the language "fi_FI".".	07:15
spearce	aaah, aspell freaked	07:15
	and git-gui didn't handle it gracefully by disabling aspell.	07:15
	yup, got that failure here.	07:16
	i'll try to fix it before i go to bed this morning. thanks Ilari.	07:16
	workaround would be to set gui.spellingdictionary = none in your .git/config or ~/.gitconfig to disable aspell too. or set it to a dictionary that aspell does support on your system.	07:17
Ilari	If i replace Aspell by Tmispell-voikko, it crashes with 'error in startup script'.	07:18
Mikachu	is there anyone here who enjoys sending typo fix patches to the list? http://git.mika.l3ib.org/?p=git.git;a=commitdiff;h=0d9aa23	07:18
spearce	what is Tmispell-voikko ?	07:18
Mikachu	sounds finnish	07:19
	← adante left	07:19
Mikachu	ah, fi_FI, not surprising then :)	07:19
	Mikachu s/.//*	07:19
spearce	git-gui is right now assuming aspell, and nothing else, so it may be crashing because its passing aspell specific command line options or something retarded like that.	07:19
Ilari	spearce: Ah. Aspell --version prints: "@(#) International Ispell Version 3.1.20 (but really Aspell 0.60.5)".	07:20
spearce	i think i may be able to back off on the aspell specific stuff and try to disable them except for when it really is aspell. probably take me a bit longer than just not crashing. :)	07:21
	← patnakajima left	07:25
	→ tommorris_ joined	07:26
Ilari	spearce: Spell checking is such a mess of programs...	07:26
spearce	yea. but its already saved me a good bit of typos in my commit messages. then again, i only use english, which is rather well supported.	07:27
	→ tommorris__ joined	07:37
	← tommorris left	07:41
	→ jeff2 joined	07:43
	← tommorris_ left	07:44
jeff2	I ran git-instaweb from my repository directory, but when my browser opens it says "403 Forbidden - No projects found". I do have a project there, how can I make git-instaweb find it?	07:44
	← CareBear\ left	08:05
	← nkallen left	08:06
	→ nkallen joined	08:08
	→ pagenoare joined	08:18
pagenoare	hi	08:18
	i am new in git	08:18
	i register my project on repo.or.cz, add my user, and mod user, go to /home/user/projekt	08:19
	git init	08:19
	← tommorris__ left	08:19
pagenoare	git checkout adress	08:19
	git add .	08:19
	git commit -m 'test'	08:19
	annd what now ?	08:20
	→ aunes joined	08:21
	← Ademan left	08:22
spearce	`git remote add origin [email@hidden.address]	08:22
	that will tell your local repository about the one that you have on repo.or.cz.	08:22
	→ glommer joined	08:22
spearce	then you can publish your changes to repo.or.cz with git push: `git push origin master` would publish the current contents and history of your branch master to the repository you called origin, in its own branch called master.	08:23
pagenoare	hmm, access denied	08:24
	password is 100% correct	08:24
spearce	repo.or.cz doesn't use passwords, it uses only ssh keys.	08:24
pagenoare	so, i paste incorrect ssh key?	08:25
	can i change it?	08:25
Ilari	pagenoare: You need to authorize the user to push. Admin authorization does not imply authority to push.	08:25
spearce	or your ssh client doesn't know to use that ssh key with repo.or.cz. i had to add an entry to my ~/.ssh/config file for repo.or.cz to make it use the proper key file.	08:25
pagenoare	spearce: i havent ~/.ssh/config	08:26
	pagenoare @ macos lori ls ~/.ssh	08:26
	id_rsa id_rsa.pub known_hosts	08:26
spearce	hmm, only one key, so not too likely the problem. i have like 10, and that's 7 more than most servers will allow my client to attempt before tossing it off. :)	08:27
	gitster says "we are at 304 stacks"	08:27
pagenoare	spearce: i paste incorrect in registration ;(	08:28
Mikachu	gitster: i just read that git reset -- paths description, and i think it's a bit confusing about what will happen when the file already exists in the previous commit	08:28
	← samgranieri left	08:29
Mikachu	ie, it says "remove the file from the index" which implies the next commit will actually remove the file	08:29
Ilari	Wonder why repo.or.cz doesn't even try running the pasted keys though ssh-keygen to check some really blatant errors... Probably because of laziness...	08:29
gitster	yeah, that is called duct-tape ;-)	08:29
	← vbgunz left	08:30
Mikachu	you can run keys through ssh-keygen? /me reads manpage	08:30
gitster	but it is surprising that the site has gathered so many projects.	08:30
pagenoare	spearce: Permission denied, please try again.	08:30
Mikachu	that's not a helpful suggestion, it's not very likely to not be denied the next time :)	08:31
	→ samgranieri joined	08:31
spearce	that's like the message i got today from a micro$oft product: "administrator is busy. please contact the administrator and tell them they are busy."	08:31
pagenoare	eh, what now	08:32
spearce	open a new user account with the correct ssh key, or email pasky and ask him to fix your key, or both. :)	08:32
pagenoare	cat ~/.ssh/id_rsa.pub	08:33
	it's my key, yea?	08:33
spearce	right	08:33
	→ kumbayo joined	08:34
pagenoare	spearce: with sshrsa?	08:34
	on top i have ssh-rsa	08:34
spearce	yea, the ssh-rsa part is part of the key line.	08:35
Ilari	pagenoare: Paste the entiere thing. ssh-rsa or whatever included.	08:35
	→ AtharHameed joined	08:36
pagenoare	Ilari: ok, but i have other login in bash, and in repo.or.cz	08:37
	how to i must login?	08:37
Ilari	pagenoare: '<useraccount>@repo.or.cz' as host to connect to?	08:39
	pagenoare: Or make an hostalias (they can specify user to connect as too).	08:39
gitster	Or use "User me" in "Host repo.or.cz" section in .ssh/config	08:40
	→ Ademan joined	08:40
	← Sho_ left	08:40
pagenoare	yea, works, thanks all	08:42
gitster	too many ways to skin the cat...	08:44
	← spearce left	08:45
pagenoare	how to add the dictonary into git repo?	08:45
Ilari	pagenoare: 'dictonary'?	08:46
pagenoare	folder	08:47
	sorry, i don't know how to write in english	08:47
Ilari	pagenoare: Add some file in it (or in its subdirectory).	08:48
gitster	we do not track directories. if you have an interesting contents (files) in a directory, just add that file.	08:48
	The containing directory comes for free ;-)	08:48
	← dsaxena left	08:49
Mikachu	hm, can you manually add that empty tree sha1 to a tree if you wanted to? would git create the empty dir on checkout then? only asking out of academic interest of course :)	08:49
pagenoare	yea, sorry :D	08:49
gitster	No.	08:49
pagenoare	it works	08:50
Mikachu	kay	08:50
gitster	Yeah, standard answer to "empty directory" question is "to have .gitignore that covers everything, as obviously you do not want to have anything tracked in that directory" ;-)	08:50
	← jeff2 left	09:00
	← asdx left	09:05
	→ pombreda joined	09:07
	→ lukass joined	09:07
	← eternaleye left	09:08
	→ eternaleye joined	09:08
	→ russell^84 joined	09:20
russell^84	is it possible to push to an newly created git repository on a webserver(apache) via webdav? i keep getting "Could not LOCK" from apache, due to the /refs/heads/master file not existing	09:23
	→ bschindler joined	09:28
	← Ademan left	09:31
	robfitz_ → robfitz	09:33
Tv	russell^84: DAV push is probably the most unreliable push method right now, encountering silly bugs is just about expected	09:33
	russell^84: if you can, use ssh, that's the one most of us use	09:34
russell^84	ah ok	09:34
	→ neosun joined	09:34
russell^84	cool thanks	09:34
	← pagenoare left	09:34
Tv	russell^84: and, as usual, let me point to http://eagain.net/gitweb/?p=gitosis.git;a=blob;f=README.rst ;)	09:35
	← samgranieri left	09:35
	→ tommorris joined	09:35
russell^84	yeah, the only reason i wanted dav support was because most of my coworkers are used to svn. Hey i also noticed there is ftp support, how useful/stable is it?	09:38
	← pergesu left	09:39
Tv	russell^84: ssh is the thing, git-daemon uses same protocol so works well (but is anonymous)	09:39
	honestly, i always use svn+ssh with svn	09:39
	i've never actually tried ftp push	09:39
aazar	Hey guys, is there a way with git diff to get a diff between my version of a file and the one residing on my remote origin?	09:39
Tv	aazar: git fetch && git diff remotes/origin/master -- path/to/file	09:40
aazar	Tv: thanks, I'll try it out.	09:40
russell^84	Tv: thanks for the help	09:42
	→ robinr joined	09:45
Tv	ftp is one of those protocols i wish would die away fast	09:46
	it's a pain for firewalls and nats	09:46
	for just about everything i've ever used it for, http or sftp are better	09:46
	→ trochala joined	09:47
Tv	or rsync	09:47
	→ Eludias joined	09:48
russell^84	yeah, i know how you feel. But GSIFtp is really popular in the grid(globus) computing world and one of the reasons i'm playing around with git is to see if it can act as a scientific/research data store	09:52
	→ dash__ joined	09:58
	← tommorris left	09:59
	← FunkeeMonk left	10:01
	→ Leonidas joined	10:02
Tv	russell^84: unlikely for any real data	10:03
	russell^84: i put my digicam pics in git just to seek, and push/pull/pack are horribly slow with large files	10:03
	s/seek/see/	10:04
	→ janm joined	10:09
up_the_irons	Tv: oh hey Tv, how's it hangin	10:10
Tv	up_the_irons: the usual	10:10
	thinking if i spend the sunday cleaning my desk, i might see more than 30% of it on monday	10:10
up_the_irons	haha	10:11
	→ TuxCM- joined	10:12
up_the_irons	Tv: i'm shopping for a thinkpad. thinking i'll make my budget 1.5K and get what I can with that	10:12
Tv	mmm thinkpads	10:13
	i wish they still make them (as well) when i wanna switch to something smaller again	10:13
up_the_irons	Tv: so have you been pretty happy with your T61?	10:14
Tv	yeah	10:14
	i mean, it's big and clumsy	10:14
	but the kbd is the usual thinkpad quality, the screen is big & bright, and there's plenty of cpu power	10:14
up_the_irons	yeah, that's pretty much what i'm after	10:15
	→ tommorris joined	10:15
Tv	and the 2GB disk cache certainly helps	10:15
	;)	10:15
up_the_irons	i don't like the weight (little heavier than my MBP), but i'll just have to reason i'm not lifting it much, usually on the desk	10:16
Tv	yeah i carry it between desk and sofa, mostly	10:16
up_the_irons	yeah i bet that disk cache is nice ;)	10:16
Tv	the 12" was so much nicer to take to events	10:16
up_the_irons	yeah i bet	10:16
Tv	but the screen on this thing is big enough that i don't even want dual screen	10:17
	(because i'd never have the dual screen on the sofa anyway)	10:17
up_the_irons	yup	10:19
	ChanServ set mode: +o	10:19
up_the_irons	i feel that way about my mbp, at 15", i don't need another monitor. I use profont anyway	10:19
	Mikachu changed the topic to: 1.5.4.2 \| Homepage http://git.or.cz/ \| Everyone asleep or clueless? Try [email@hidden.address] \| Channel log http://colabti.org/irclogger/irclogger_log/git \| Mailing list archives: http://marc.info/?l=git \| Gits on git: http://tinyurl.com/2xq3ke \| You want $ID?: http://tinyurl.com/yqpgv9 \| Need to paste something?: http://git.pastebin.com	10:19
	Mikachu set mode: -o	10:19
up_the_irons	what the, the X61 now is energy start 4.0 compliant	10:19
	it wasn't before...	10:19
	mmm..	10:19
	Tv: is your screen type the LED backlit kind?	10:21
Tv	don't think so	10:21
	generation too old	10:21
	← janm_ left	10:21
Tv	it's brighter than any earlier thinkpad i've seen	10:21
	every time i've switched to a new one, the difference has been huge	10:21
up_the_irons	gotcha	10:23
Tv	haven't used it outside much yet, though, because it's so clumsy	10:23
up_the_irons	mm.. the x61s goes up to 1.8GHz only	10:23
Tv	yeah, power savings..	10:23
up_the_irons	yeah	10:23
Tv	then again, fast disk + plenty of ram >> cpu	10:23
up_the_irons	i don't think i've ever used my mbp outside either.. not much point	10:23
	yeah, for sure	10:24
Tv	i think my 12" was 1.8GHz	10:24
	it felt fast enough to do compiles etc	10:24
	difference between 1.8 and 2.2 is way less in reality than on paper	10:24
up_the_irons	yeah	10:25
	i'd agree	10:26
	→ TGEN_ joined	10:26
Tv	and for compiles etc, plenty of ram helps a lot more	10:26
up_the_irons	mm.. the T61 does have a 14" model.. perhaps a compromise in size/weight	10:26
Tv	don't want to be disk bound	10:26
	← aazar left	10:26
up_the_irons	right	10:26
Tv	up_the_irons: yeah, but check the pixels	10:26
up_the_irons	uhoh, pixels	10:27
Tv	up_the_irons: at least when i looked at that, it had horrible dpi	10:27
context	when it comes to laptops now	10:27
Tv	well ppi	10:27
context	its all about desktop realestate	10:27
	← tommorris left	10:27
up_the_irons	the high end has 200 nit	10:27
Tv	context: i went from 12" to 17", and will likely go back next ;)	10:27
context	you realize 12-17 is just about the entire laptop range	10:27
Tv	nit is brightness	10:27
	context: there's plenty of things below 12	10:28
context	...	10:28
Tv	oh damn not 17", i'm on crack	10:28
	15.4"	10:28
context	i bought a 15" mbp	10:28
	← TuxCM left	10:28
context	it replaced my desktop	10:28
	13" is to small	10:28
	17" is to big to carry around	10:29
up_the_irons	Tv: ok, then for some reason the pdf doesn't have the ppi	10:29
context	15" is the happy medium	10:29
Tv	12" with thinkpad kbd is just lovely	10:29
	just wanted more pixels than they could manufacture	10:29
	up_the_irons: oh they never do.. calculate it	10:29
up_the_irons	Tv: lol	10:29
	up_the_irons grabs a calculator	10:29
context	tv: call and pay them enough and you can get anything ;)	10:29
Tv	up_the_irons: i'm just saying the 14" i saw was like 1024x768	10:30
context	from my understanding thinkpads are garbage since ibm sold it	10:30
Tv	context: oh in this case not really, not without waiting	10:30
context	which was honestly the only pc laptop model i ever cared about	10:30
Tv	context: the high ppi display panels aren't really available in 12" sizes	10:30
	context: it's on the product list but you not available	10:30
up_the_irons	Tv: oh ok, the WXGA+ on 14.1" model does 1440x900, which isn't bad i don't think	10:30
context	12" is tiny anyway	10:30
Tv	context: 12" is physically enough if it was 1400x1050	10:31
context	asside from being completely portable	10:31
Tv	up_the_irons: that sounds workable	10:31
context	its just to small for me	10:31
Tv	up_the_irons: i think that just wasn't available when i was buying, and i couldn't wait	10:31
	context: full size thinkpad kbd etc	10:31
up_the_irons	Tv: right	10:31
Tv	though i do have good eyes, i'm sure if you don't you'll disagree about 12" screens being physically big enough	10:32
	other than that, the 12" has identical feel for use as any other thinkpad	10:33
up_the_irons	Tv: do you have the Intel 4965AGN card?	10:34
Tv	03:00.0 Network controller: Intel Corporation PRO/Wireless 4965 AG or AGN Network Connection (rev 61)	10:34
	up_the_irons: intel really did fine work with that one	10:35
	up_the_irons: open source goodness	10:35
up_the_irons	Tv: cool	10:35
	Tv: well, a nicely configured x61s is $1644; oh the hell with budgets. ;) Now let's see what that'll get me in a t61	10:36
	→ janm_ joined	10:39
up_the_irons	Tv: would you think the difference between a 3MB and 4MB proc L2 cache would be noticable?	10:43
Tv	not as badly as 2MB would be	10:43
	but depending on what you do with it, it will definitely have some effect	10:44
	maybe even almost as much as the 1.8GHz vs 2.2GHz thing	10:44
	that's more a question of are you pushing 7200rpm disks yadda yadda	10:44
	oh let me never again suffer from 5400rpm laptop disks..	10:46
up_the_irons	lol	10:46
	sure wish i didn't have to pay for a winblows license with this purchase	10:47
cehteh	use a SSD :)	10:47
Tv	up_the_irons: right now you sort of need windows to set up the bios fingerprint protection :(	10:47
	up_the_irons: though honestly i haven't played with the reader any more than that	10:48
up_the_irons	Tv: i wasn't really thinking of using the reader	10:48
	Tv: just more candy	10:48
Tv	up_the_irons: i'm not using it really except on poweron	10:48
	don't have it set up in linux	10:48
	→ mithro joined	10:48
up_the_irons	Tv: oh man, there's a 6MB L2 cache proc in the list. 6MB!!	10:48
Tv	but i kinda like my non-server computers not to start up without me	10:48
	hehe	10:48
	← russell^84 left	10:49
up_the_irons	Tv: yeah that's true	10:49
	← janm left	10:51
	→ drizzd joined	10:52
up_the_irons	Tv: interestingly, i'm using 1440x900 on my 15" MBP. That means, if I got a 14" T61, I'd have greater ppi. sweet	10:52
	→ janm joined	10:56
	→ Ademan joined	10:57
	← moya left	10:58
	→ moya joined	10:59
	→ adante joined	11:07
	← adante left	11:08
	← janm_ left	11:10
	→ ferdy joined	11:17
	→ Leonidas_ joined	11:21
	← Leonidas left	11:21
	Leonidas_ → Leonidas	11:21
drizzd	If I do git push origin :test I get "error: remote 'refs/heads/test' is not an ancestor of local '(delete)'"	11:26
	What's going on?	11:26
	→ HG` joined	11:26
	→ SuttoL joined	11:31
drizzd	Hmm, appears to be yet another problem with git-http-push.	11:31
	→ Yuuhi joined	11:32
	→ Beket joined	11:37
	← fultilt left	11:38
	← capisce left	11:41
	→ capisce joined	11:41
context	push over ssh, its more secure ...	11:41
	← reuss left	11:47
	→ Sho_ joined	12:04
	→ thresh_ joined	12:05
	← thresh left	12:06
Tv	and is the one 99.99% of the community uses, so has less bugs	12:09
paakku	how large is the community?	12:09
drizzd	context: I would if I could	12:11
Tv	at least thousands of people, but nobody knows	12:11
	not like they'd be buying licenses from a single corporation	12:12
drizzd	in any case, -f should not be required, even for push over ssh. right?	12:13
paakku	are there statistics on people visiting #git or posting to the mailing list?	12:14
Tv	paakku: archives are public, feel free to make (up) some	12:14
	drizzd: to delete branches? no	12:14
up_the_irons	Tv: did you get your T61 using the cpp? I'm in it now and it gives 15% discount, pretty nice :)	12:19
Tv	up_the_irons: cpp?	12:19
up_the_irons	Tv: lenovo.com/cpp (contractor purchase program)	12:19
	thresh_ → thresh	12:19
Tv	#define THINKPAD "T61"	12:19
	up_the_irons: nope, i'm registered with them as a reseller but only in EMEA	12:20
up_the_irons	Tv: the passcode is 556655, if ya ever wanna buy some accessories :)	12:20
	Tv: gotcha	12:20
Tv	one secret?	12:20
	whoa, stoopid	12:21
up_the_irons	yup	12:21
Tv	and they don't check anything?	12:21
	that's just silly	12:21
up_the_irons	i hear they don't check	12:21
	ppl on slickdeals.net were able to purchase through it	12:21
	it's probably one of those things they don't want you to find, but kinda do	12:22
	Tv: you have 4GB RAM right?	12:22
Tv	heck most local b2b resellers won't let you even see their prices without logging in	12:22
	up_the_irons: yup, maxed out	12:22
up_the_irons	Tv: nice	12:22
	Tv: i'm gonna max this puppy out too	12:22
	Tv: now let me say that "Executive Class" thinkpad is silly	12:23
	Tv: waaaaay ugly	12:23
	← neosun left	12:23
Tv	yeah i hope lenovo doesn't screw up the whole series, one by one	12:23
up_the_irons	hehe yeah	12:25
	Tv: does the 4-in-1 media card slot read SD and stuff?	12:26
Tv	i think my scanner will work a lot better if i plug it in first	12:26
	up_the_irons: no linux drivers afaik	12:26
up_the_irons	roger	12:26
Tv	SDIO was somehow DRM-infested, or something	12:26
	http://www.linuxdevices.com/articles/AT6640645071.html	12:27
	"You really can't offer an SD or SDIO stack in open-source form and meet the licensing requirements currently in effect for these devices."	12:28
	aka "wait for the next gen hardware and try again"	12:29
up_the_irons	ouch	12:29
	← Fullmoon left	12:29
Tv	personally, i've never moved data on sd/mmc cards from one device to another	12:29
	not even when i'd be in an environment with working sd readers	12:29
	"Secure Digital (SD) is a proprietary flash memory card format, including Content Protection for Recordable Media (CPRM) encryption hardware, based on the MultiMediaCard (MMC) format."	12:32
	(at http://www.olpcnews.com/hardware/peripherals/secure_digital_isnt.html)	12:33
	that's the DRM i remember	12:33
	← DeepB left	12:33
up_the_irons	Tv: gotcha	12:34
Tv	mmm OLPC might be trying to create open-source-compatible SD hardware	12:34
Leonidas	up_the_irons: I think my SD reader does work	12:34
	Yeah, it most certainly works.	12:35
Tv	Leonidas: binary only proprietary kernel module much?	12:35
up_the_irons	Leonidas: neato, what model do u have?	12:35
Tv	https://opensource.motorola.com/sf/projects/sd-mmc	12:35
	motorola and olpc seem to be working on it	12:35
	note that that's "simplified spec"	12:35
Leonidas	Tv: nope, no special modules. Just plain gentoo-sources. I was surprised that it worked out of the box.	12:35
Tv	so don't expect all the sdio devices to work with it, either	12:35
	Leonidas: anyway, it's still patented, so selling software with support is risky	12:36
	clean room implementations don't get around patents	12:36
	→ octopod joined	12:37
Tv	http://en.wikipedia.org/wiki/Secure_Digital_card#Openness_of_standards	12:37
Leonidas	up_the_irons: hmm. Hard to say. lspci and lsusb do not say anything about it.	12:37
up_the_irons	Leonidas: i meant model of computer :)	12:38
	Tv: if you could have your laptop w/ 14" and 1440x900, would you do it?	12:39
Leonidas	up_the_irons: It's a Samsung R60plus notebook. But I cannot recommend it really, since I killed my Fn-keys via bios-update.	12:39
up_the_irons	Tv: oh, and do you know the nit of your screen?	12:39
Tv	up_the_irons: maybe	12:39
up_the_irons	Leonidas: lol, ok	12:39
Tv	up_the_irons: i'd have to dig up the nits via lenovo.com i guess	12:39
	up_the_irons: might be 220?	12:39
up_the_irons	Tv: ok don't worry about it then	12:40
	Tv: i can't find it in the pdf, that's why i ask	12:40
	Tv: i think i'll go w/ 14" unless the 15" has superior brightness, etc..	12:40
Tv	up_the_irons: you have the ubersekr1t sales pdf i found back then?	12:40
Leonidas	up_the_irons: there was a bug with MMCONFIG and I tried updating the BIOS which does only work unter windows. So I booted a windows live cd and it worked. Just my Fn-keys are dead now and I cannot make the display darker or anything.	12:40
up_the_irons	Tv: i only have the one off their site	12:40
Tv	up_the_irons: is it like at least 20 pages of actual tech specs?	12:41
	up_the_irons: each model listed separately, etc	12:41
up_the_irons	Leonidas: ouch	12:41
	Tv: no, like 4 pages	12:41
Tv	up_the_irons: then you don't have the ubersekr1t one	12:41
	up_the_irons looks around	12:41
Tv	up_the_irons: the public web site has furiating little on specs	12:41
up_the_irons	Tv: do u think u could email me the ubersekr1t one? ;)	12:42
Leonidas	up_the_irons: maybe they start working once I update the BIOS on a properly installed vista, but I'd need a harddisk for that - argh, everything sucks.	12:42
Tv	not sure if i have it anymore	12:42
	i google reverse engineered the url	12:42
up_the_irons	Leonidas: ouch	12:42
	reverse engineer!	12:42
	"Hi, yes, I'm a reverse engineer. I designed the reverse gear in transmissions"	12:43
Tv	it's basically only meant for their resellers	12:44
	but is way more useful than most of their website	12:44
	just not user friendly	12:44
	ftp://ftp.software.ibm.com/pc/pcinstitute/psref/tabook.pdf	12:45
	that's probably it	12:45
	http://www.thinkwiki.org/wiki/Hardware_Specifications	12:45
up_the_irons	Tv: oh jesus that's slick	12:46
	Tv: there's even linux models?!	12:46
Tv	heh	12:46
up_the_irons	man, where do i get one of those	12:46
Tv	they're not top of the line, though	12:47
	you can just call them and order anything in there	12:47
	even if the website doesn't show it	12:47
	if they have it in store	12:47
	s/e$/age/	12:47
up_the_irons	oh ok	12:47
	well if it's not top of the line...	12:48
	← moya left	12:48
Tv	at least at one point the linux models were of the R line	12:48
	and in general, in the "business value" spirit	12:48
	pick intel chips yourself and you're just about better off, though paying the windows tax	12:49
	and if you do call them, make damn sure you get the right wlan chips etc	12:49
up_the_irons	yeah, that makes me nervous about calling; at least online i can double check everything, albeit i pay the windows tax	12:50
	tip: 15" models have integrated 4-in-1 reader, on 14" it would take up the express card slot	12:50
Tv	i had to call due to payment details	12:50
	up_the_irons: that's just case size, i think	12:51
up_the_irons	yup i'm sure	12:51
Tv	though the 12" had a 4-in-1 reader and a cdma modem in the express card slot	12:52
	my current express card slot has an extra flash in it.. maybe some day it'll work nicely in linux	12:53
up_the_irons	oh the 15" has a smart card reader built in?	12:53
	the spec says so	12:53
Tv	huh	12:53
	i don't see anything that size on this 15.4" T61	12:54
	unless it's hidden with the pc card slot	12:54
up_the_irons	Tv: page 35 of the uber secret pdf	12:54
Tv	yeah just trying to figure if that changed	12:54
up_the_irons	ah	12:55
	or, maybe it means its a compatible add on	12:55
	oh, it's true, just diff. config	12:56
	"PC Card Slot & Smart Card Slot [add $21.25]"	12:56
	Tv: ^^	12:57
Tv	ah so it is hidden in that thing	12:57
	not sure if i got that one, though	12:57
up_the_irons	yeah it's not the default	12:57
	default is express card slot	12:57
Tv	anyway, that's sort of silly because you can get smart card readers as pc cards	12:58
	if it takes the slot space anyway	12:58
	might as well have it changeable	12:58
	yeah i have two identical big empty slots	12:58
	with dog hair in them	12:59
up_the_irons	lol	12:59
	yeah i agree, no need to go with "fixed" configuration on the smart card.. when the form factor changes of the SC, i'd be screwed	12:59
	Tv: did u get a big drive in yours? (160GB+)	13:00
Tv	that waswn't available then	13:00
	i have the biggest 7200rpm that was available then, 100GB	13:00
	and over a terabyte of external storage ;)	13:01
up_the_irons	hehe cool	13:01
	it's $80 extra for 160GB (from 100GB), wondering if it's worth it. For a little over $1.20 for each extra gig, i suppose it is	13:02
	Tv: did you get anything above the base warranty?	13:03
Tv	no	13:04
up_the_irons	cool	13:04
Tv	up_the_irons: i'm outside the US, having a US warranty wouldn't have done much good for me	13:04
up_the_irons	oh yeah, duh	13:04
	→ destruct_ joined	13:06
up_the_irons	Tv: decisions, decisions... identically configured (as close as possible): 15" - $1550, 14" - $1500	13:07
	so it's basically do i want 14 or 15 inches..	13:07
Tv	wide/non-wide?	13:07
up_the_irons	both wide	13:07
	→ kukks joined	13:09
	→ rubydiamond joined	13:12
	← destruct left	13:16
	destruct_ → destruct	13:18
	TGEN_ → TGEN	13:20
	→ bschindl joined	13:23
	← bschindler left	13:25
	→ bschindler joined	13:27
up_the_irons	Tv: interesting, the contrast ratio on 14" is 300:1, on 15" it's 500:1. I like my contrast...	13:35
Tv	up_the_irons: yeah, and the sucky thing is those details change fast	13:35
up_the_irons	Tv: yeah	13:36
	→ nessundorma joined	13:36
Tv	up_the_irons: display panel manufacturing is a tight business	13:36
	heavy competition but almost everything bought off them immediately	13:37
	that's why i cursed that much when i was trying to find that 12" with a good ppi	13:37
up_the_irons	Tv: hehe, i c	13:37
Tv	displays with that ppi are apparently better business at slightly bigger sizes	13:37
up_the_irons	ah	13:38
	← bschindl left	13:42
	→ reuss joined	13:45
up_the_irons	Tv: well, your secret pdf does say Jan 08, so it's probably pretty up-to-date	13:46
Tv	not for availability but yeah	13:46
	availability can change almost daily	13:46
up_the_irons	Tv: i think i gotta assume 300:1 CR on the 14". same CR as the 15" R series; so now i'm thinkin the 14" screen is a little "cheap"	13:46
	Tv: true true	13:47
Tv	iirc the guy on the phone said two weeks is the normal time they can stock things, at most	13:47
up_the_irons	yup	13:47
Tv	14" screen being cheap sounds perfectly believable	13:47
up_the_irons	Tv: oh crap, ya know what.. it's 300:1 on the 15" too if it's GMA. they only give 500:1 to NVIDIA models. ggggggrrrrrrrrrrrr	13:48
	up_the_irons throws a brick	13:48
Tv	oh huh	13:49
	different panels for different controllers, sounds weird	13:49
up_the_irons	very	13:49
Tv	i understood the connector is standard etc	13:49
	maybe its more about manufacturing just a few combinations	13:49
up_the_irons	or they figure the nvidia customers are gamers and want better colors	13:50
Tv	which is pretty much the same thing i said ;)	13:50
	not wanting to provide the extra combinations	13:50
up_the_irons	yeah	13:50
	:)	13:50
	Tv: think i figured it out. The WSXGA+ (plus is the key) is the 500:1 CR. The pdf doesn't show these w/o NVIDIA, but obviously you can get them w/ GMA	13:56
	Tv: so there's hope	13:56
Tv	heh	13:56
	← AtharHameed left	13:56
Tv	remember my rant about that "+"	13:56
up_the_irons	i do in fact	13:56
	actually, the "S" is important too	13:56
	← IRSeekBot left	13:57
Tv	it explicitly says "more than", but two "FOOGA+"s don't need to match	13:57
up_the_irons	WXGA+ on 14" is 300:1, so gotta have the "S" too	13:57
Tv	oh the S adds a lot of pixels	13:57
	XGA wasn't that much, IIRC, so more than XGA ain't much either	13:57
up_the_irons	yup	13:57
Tv	heck, XGA means 800x600@16bit or 1024x768@8bit	13:58
up_the_irons	lol	13:58
doener	for my T43, I had to talk to the dealer in terms of the 3-letter codes...	13:59
Tv	WCGA+	13:59
	doener: 3-letter?	13:59
	the product codes were 5 number and/or digits iirc	13:59
doener	do I want the 2668-WUM or maybe the 2668-WAM? or 2668-FOO?!	13:59
Tv	oh yeah, 7 letter	13:59
	2668WUM is the product code, then	13:59
	the prefix ain't always constant either	14:00
doener	2668 being the series IIRC	14:00
Tv	2668WUM can be almost identical to FE68XAM	14:00
doener	and the last three the specific components	14:00
	→ AtharHameed joined	14:00
Tv	i went through the parts list chip by chip on the phone when i ordered, just to make sure contents & prices match	14:01
	→ IRSeekBot joined	14:01
doener	I had a misunderstanding in one email, which made me order the wrong one.	14:02
	Fortunately, I could stil cancel it. The shop had offered me an almost identical model, just 10% cheaper and with a different product code	14:02
	I just lost track of which code was what stuff	14:03
Ilari	Yeah. Sometimes some clowns even change chipsets wthout changing model names... Buying craputers is easy, but buying good stuff is not trivial...	14:11
	→ shafty joined	14:13
simmel	That's why you make so you only have one choice and let the company choose for you. Go Mac! ; P	14:13
	← SuttoL left	14:14
simmel	</flame'n'rant>	14:14
Tv	one size fits steve jobs fanboys?	14:15
simmel	(Y)	14:16
	← kumbayo left	14:17
	← ferdy left	14:22
	→ priidu joined	14:25
	← tokkee left	14:27
up_the_irons	Tv: do u know what kind of batter life you get w/ your 9 cell?	14:42
Tv	up_the_irons: haven't really measured, but ...	14:42
	wlan on full brightness etc acpi estimate is 3h45min	14:42
	shutting down drivers is supposed to get at least an hour extra	14:43
up_the_irons	Tv: ok cool, tnx	14:43
	→ kumbayo joined	14:43
	→ langenberg joined	14:46
up_the_irons	k	14:46
	→ zhangkai joined	14:50
	← langenberg left	15:04
	→ tommorris joined	15:05
	← rubydiamond left	15:08
	← aunes left	15:10
	→ SuttoL joined	15:17
	→ zapnap joined	15:20
	→ aunes joined	15:20
	→ chris2 joined	15:20
	→ G_SabinoMullane joined	15:27
raggi	Tv / up_the_irons : not sure exactly what model you were talking about, but i've had 6 hours out of a macbook whilst on a plane (everything external disabled, and screen down on a night flight)	15:29
	← G_SabinoMullane left	15:29
raggi	which is enough for a trans-atlantic :)	15:29
	← priidu left	15:29
	→ priidu joined	15:32
	← tommorris left	15:34
	→ tommorris joined	15:34
	← harinath left	15:39
	→ harinath joined	15:40
	→ FunkeeMonk joined	15:42
	← sverrej left	15:56
	→ ferdy joined	16:01
	← FunkeeMonk left	16:07
	← srid left	16:13
	← ghtdak left	16:13
	← Beket left	16:17
	→ Beket joined	16:18
	→ vbgunz joined	16:19
Beket	Can git expand sccs keywords?	16:23
thiago_home	not the way you want	16:23
Beket	Thank you thiago_home	16:23
	→ Morphous joined	16:24
thiago_home	see http://git.or.cz/gitwiki/GitFaq#head-4a0afe71a2bb7734777a8b0b345e0308aefdbd40	16:24
	← jaalto left	16:27
	← Amorphous left	16:40
	→ bobesponja joined	16:42
	← bschindler left	16:43
	← Beket left	16:46
	→ Beket joined	16:50
	→ sverrej joined	16:51
	← priidu left	16:51
	← drizzd left	16:55
	← nessundorma left	17:01
	← IRSeekBot left	17:02
	→ IRSeekBot joined	17:05
	Morphous → Amorphous	17:10
	→ shaftyy joined	17:17
	← Beket left	17:22
	→ Beket joined	17:27
	→ sinkorswim joined	17:35
	← shafty left	17:35
	→ shafty joined	17:35
sinkorswim	http://uszla.me.uk/space/blog/2007/09/20 - on that page, the author uses git pull to merge a branch, can someone explain the difference between merge and pull pls	17:35
Mikachu	pull is fetch+merge	17:36
	pull . == merge	17:36
Ilari	As sidenote 'pull .' is at least informally deprecated.	17:37
sinkorswim	thanks	17:38
Tv	learning about git from people who are scrambling to learn it themselves doesn't always get you the best results ;)	17:38
	← harinath left	17:38
sinkorswim	learning about git from the git manual alone gets you unintended results :)	17:40
	→ kevwil joined	17:40
	← zapnap left	17:42
	← zhangkai left	17:42
	→ a-priori joined	17:43
	← a-priori left	17:46
	← eikonos left	17:46
	→ eikonos joined	17:47
	→ a-priori joined	17:48
	← shaftyy left	17:52
	→ pietia joined	17:56
	→ dsaxena joined	17:57
	→ toxx_ joined	17:58
	← tommorris left	17:58
	→ pengwn joined	18:03
	← Beket left	18:05
	← pengwn left	18:05
	→ pengwn joined	18:06
	→ langenberg joined	18:10
	→ shaftyy joined	18:11
	→ alley_cat joined	18:17
Ramblurr	i'm doing 'git diff' on a directory and my terminal.. goes blank	18:19
	until I C-z it	18:19
	→ tvachon joined	18:21
loops	Ramblurr, sounds like a pager issue.. try git --no-pager diff	18:24
Ramblurr	hm yea that did it	18:25
loops	you can choose a different pager with GIT_PAGER environment variable, or setting core.pager in your ~/.gitconfig	18:26
Ramblurr	loops: Oh. Haha.	18:26
	It was piping the output into less	18:26
	but there were no changes	18:26
	So less was blank.	18:27
loops	Ramblurr, ah, i think if you set your environment variable LESS=-RSX that situation ends up looking a bit better (not sure)	18:27
	export LESS=-RSX that is	18:28
Ramblurr	thanks	18:28
	that did it	18:28
Tv	Ramblurr: by "blank" you mean only had text on the last line, right?	18:29
loops	Tv, without the environment variable set, less will clear the screen before displaying anything on some terminals.	18:29
	← shafty left	18:30
Tv	loops: sure but it should still show the status line at the bottom	18:30
loops	it's the -X that helps	18:30
	→ samgranieri joined	18:30
	→ gitte joined	18:31
loops	Tv, yeah, i'm surprised less wouldn't at least show "(END)" or something like it	18:32
Tv	i'd claim it probably did, but that was missed	18:32
Ramblurr	nope	18:34
	the status line wasn't there	18:34
	LESS was = -R -M --shift 5	18:35
	gitte thinks that those git-svn tests are really slooow	18:35
tpope	wow what's with the wave of people who set $LESS lately	18:36
gitte	Doesn't OpenSUSE set it by default?	18:36
loops	some distros have it set by default	18:36
gitte	(Forgetting -X...)	18:36
tpope	yeah, I just mean this is like the third time in 24 hours	18:36
	and I've hardly watched the channel the full 24 hours	18:36
gitte	Heh.	18:37
	Problems tend to have that "clustering phenomenon"	18:37
	I also remember that this stupid "I want my files to have the date of the last commit touching them" issue came in clusters.	18:37
Mikachu	gentoo seems to set what Ramblurr has by default	18:40
Ramblurr	yup. that would be it	18:43
	→ aazar joined	18:47
	← kevwil left	18:50
	← janm left	18:51
	→ cypromis joined	18:56
cypromis	is there any solution to do something like svn externals in git ?	18:56
	ad	18:56
	is there a netbeans plugin for git ?	18:57
Mikachu	look at git-submodule	18:57
	(but don't ask me about it :)	18:57
gitte	There's a nice howto on the wiki AFAIR	18:57
cypromis	I am curious if I can take a svn repo we use	18:57
gitte	Netbeans plugin: no.	18:57
cypromis	and let people use git & svn in parallel	18:57
	ok will check that	18:57
tpope	I saw someone speak on netbeans about a month ago and he noted that git support was currently absent in netbeans	18:58
gitte	I have no idea how much jgit is along... but if it is far enough, it should be relatively easy to make a Netbeans plugin, too.	18:59
paakku	is there some sort of git plugin for Apple's Xcode?	18:59
	gitte wonders if people will ask for a git plugin for their toaster, too.	19:00
paakku	hey I wanted it before but didn't remember to ask.	19:01
Mikachu	gitte: that would be great, start toasting in the kitchen, then clone it to the dining room and finish toasting it there	19:01
gitster	is toaster now a Porcelain?	19:01
	And when plumbing is not flushing, it backflows into your toaster? No thanks.	19:02
gitte	I saw some made from porcelain... does that count?	19:02
Mikachu	oh is that where "porcelain" comes from?	19:03
gitte	Yep.	19:03
	;-)	19:03
gitster	"git wrapper" was originally called "git potty". Eeek. go fiture.	19:03
	s/t/g/	19:04
gitte	IIRC it was Andreas' idea... maybe his child was in potty training, then?	19:04
gitster	Yeah, Linus began his original "intro/tutorial" with something like "this talks about plumbing, you may not usually be interested in how it works, because you would be using Porcelain, but when the plumbing is not flushing...". It was funny.	19:05
gitte	Yes, I like Linus' humour.	19:06
gitster	I am wondering what's the easiest way to help people who built on 'next'. I am thinking of pushing out a branch old-next that builds on the original 'next' before rewinding whose tree matches the rebased 'next'. And people can say "git rebase --onto origin/next old-next my-fork".	19:08
	Would that be sensible?	19:08
	s/rewinding whose/rewinding, whose/;	19:08
	→ Cerebon joined	19:08
gitte	I wonder if not most people switched to rebasing now.	19:08
Mikachu	you mean pull --rebase?	19:09
gitte	(I think I was one of the last resistance)	19:09
gitster	Mikachu: I mean fetch && rebase, but I presume they amount to the same thing.	19:09
gitte	Oh, I see!	19:09
	There is a subtle difference.	19:09
Mikachu	if you fetch first, you have to note the output or use the reflog	19:09
	pull --rebase automatically does the --onto bit for you with the old head as the base	19:10
gitte	Namely, if you use "git pull --rebase", you benefit from the recent fix, that it rebases --onto the new one, but using the old revision of the remote branch as base.	19:10
	Mikachu: yep, that's what I meant.	19:10
	gitte is being slow today, like git-svn unit tests.	19:10
gitster	I am wondering how well/badly it interacts with the totally rebased case and that is where I thought old-next might help.	19:11
	→ priidu joined	19:12
gitte	Sure.	19:12
	You might also tell people that you expect to delete that branch in the mid-term future.	19:13
	→ jkp joined	19:13
Mikachu	it's not that hard to look at the history and see when it stops having your name, and grab the next sha1, in a pinch	19:13
jkp	hello all...looking for some advice on running Git on top of a perforce repo?	19:13
gitster	Now 'next' is about 182 commits ahead of 1.5.4.2, instead of 1323.	19:14
jkp	we use Perforce at work, but i was hoping to experiment using a local Git repo to do my own experimental work on the code	19:14
	so..i see there are several scripts for doing this...does anyone have any thoughts on which is "best"?	19:14
	and also, just in general, is there a methodology / best practice to follow in terms of keeping the two in sync? should master always be rebased on the perforce repo? or do i rebase my branches?	19:15
	im assuming the same best practice applies to those using Git on top of svn or other SCM systems so it must be fairly well known by now	19:15
	(btw this is my first day looking at Git...just got through watching a couple of nice Google tech talks that do a very good job of selling it)	19:16
gitster	I do not work with P4 but generally if your project considers foreign SCM "the authoritative" copy, it is easier to make your git side follow them by rebasing, and not have any central copy on your git side.	19:16
gitte	Mikachu: yes, it is relatively easy, but I saw some people who did not have the idea to "git rebase --abort" and start over again... Better tell them beforehand.	19:16
Mikachu	do many people base their work on next?	19:16
gitte	Mikachu: I do.	19:17
Mikachu	but then gitster has to rebase it to merge into master?	19:17
Pieter	most do, I'd guess	19:17
gitte	Mikachu: but then, I _run_ next+patches. And AFAICT so do other people, like Randal, Linus, ...	19:17
Mikachu	curretly i run maint+patches, i'm not feeling adventurous	19:17
gitster	I had an impression that mostly they are master+patches.	19:17
jkp	gitster: ok, but do i have master rebased and use other branches for experimenting...then merge downstream to my branches from master to keep everythin in check?	19:17
gitte	jkp: I have _zero_ experience with Perforce, but we have "git-p4" in our contrib/. Maybe that is something for you?	19:18
jkp	gitte: there are a couple of options...aparantly thats not being maintained any more	19:18
	there is a python script which im leaning towards since i hate perl :p	19:18
gitte	jkp: it is updated in irregular intervals, mainly by Simon Hausmann.	19:18
jkp	which that script?	19:18
	→ TL_CLD joined	19:18
jkp	ahhhh, yes	19:19
gitte	jkp: git-p4	19:19
jkp	ok: http://git.or.cz/gitwiki/InterfacesFrontendsAndTools?highlight=%28perforce%29#head-bb56c0cc211da9af16917d005ac59179224373f3	19:19
	yeah, it does say that, sorry misread	19:19
	ok...ill look at that	19:19
	im just more trying to get a feel for the whole way of working	19:19
gitster	jkp: I do not do P4, but what I do is keep one branch (master) follow the "authoritative" foreign SCM, build any topics on top of it, and rebase topics on 'master' (after you make 'master' follow the foreign SCM which would likely to involve rebases).	19:20
jkp	the idea is this: i have a project in perforce which i want to make heavy changes to. so i want to keep pulling from perforce and keeping my master up to date....but it sems that Git will be very good since it tracks content not files	19:20
	gitster: right	19:20
	thats kinda what i thought	19:20
gitster	In a sense, I do not work on 'master' in git for that project.	19:20
paakku	In git-rev-list, do "foo..bar baz" and "foo..baz bar" always mean the same?	19:20
jkp	gitster: right	19:20
gitster	So updating 'master' from foreign SCM always fast forwards.	19:20
jkp	again, yes, this is what i thought	19:20
	then my topics would need to be rebased on master	19:21
paakku	the same set of commits, I mean, disregarding order	19:21
jkp	and eventually ill have things in a state where i can push back to master, and in turn hopefuly to perforce	19:21
gitster	Yup. And experiment inside git and when things are really good, feed that change to foreign SCM.	19:21
jkp	gitster: does that sound about right?	19:21
	right...good :)	19:21
	i like the sound of this a lot	19:22
gitster	The next "update from foreign SCM" will then kick your changes back to your 'master'.	19:22
jkp	sorry, you lost me there!	19:22
	if i update from the foreign SCM surely that just brings master in sync	19:22
	(to foreign SCM)	19:22
	my topics need to then take this into acount...right?	19:22
gitster	by "kick back", I didn't mean "rejecting".	19:23
jkp	and thats the one thing i didnt get form the talk i watched: when you rebase it changes the SHA1...how does this affect derived topics?	19:23
Mikachu	they are completely unaffected, which is not what you want usually :)	19:23
	ie, they will have the old pre-rebase branch in their history still	19:24
jkp	so what do you do to resolve this?	19:24
gitster	So when my changes I made in my topics mature, I feed them to foreign SCM, (I'd use git-cvsexportcommit if the foreign one is CVS, and perhaps git-svn dcommit if it is SVN). Then foreign side will have	19:24
Mikachu	usually you wouldn't rebase something with topics on top, but you can rebase them along too	19:24
gitster	change I made on my topic.	19:24
	The next update from foreign will have them in 'master'.	19:25
	Now, these changes will have different commit object name on 'master' and my topic.	19:25
	When you rebase my topic onto my updated 'master' that already has the changes, rebase notices them, not by comparing the commit object names, but by comparing the changes each commit introduces.	19:25
jkp	gitster: ok, so i thought it would work like this Perforce == authorative master - > my Git master - > my topics	19:26
gitster	Yes. That is correct.	19:26
	Then the changes are fed not to git master from your my topics directly.	19:26
jkp	so then i would be periodically rebasing master while my topics are incomplete	19:26
	← pietia left	19:26
jkp	(rebasing on the perforce repo)	19:26
	which is why i asked what do i then do to keep the topics in step?	19:26
	→ pietia joined	19:27
Mikachu	rebase them	19:27
jkp	assume for now im not pushing anything back to perforce, this is highly experimental work	19:27
gitster	rebase on master?	19:27
jkp	ok	19:27
	understood, and do i loose anything doing this?	19:27
	i was under the impression you loose things when you rebase	19:27
	(in terms of history)	19:27
gitster	But you are not publishing your git side history anywhere, so there is nothing lost.	19:28
	In your project, P4 side is the authoritative one everybody else looks at.	19:28
jkp	right, but i dont loose any commit history? say i wanted to rollback something ive done, or go back in time in my local Git repo	19:28
Mikachu	rebase -i loses a bit of history depending on what you do, but not a normal rebase that just updates to a newer upstream	19:29
jkp	ok...i guess im just gonna have to try it :)	19:29
	im very excited about the idea of being able to do this though...perforce is a nightmare for me. i work offline a lot an i also do experimental work	19:29
	i hate the whole connectedness of perforce	19:30
	doing this means i should be able to work my way in safety wherever i am, then push back when im done	19:30
	← tvachon left	19:31
careo	I'm setting up a central shared repository for some coworkers as they move to git from svn. Am I correct that git-daemon, gitosis, and friends aren't really applicable if I only want 4 or 5 people to have read an commit access, with no other public/anonymous access of any sort? So just the plain vanilla setup is perfect?	19:32
Mikachu	they need a bit of ssh access, but i guess that's implied by saying gitosis	19:33
	oh right you said not gitosis	19:33
careo	Mikachu: right, but in this case everyone already has shell accounts. so gitosis doesn't seem particularly necessary	19:34
Mikachu	then they need ssh access for committing	19:34
	right right	19:34
	then it should be fine	19:34
	assuming you all cooperate :)	19:34
careo	ok. cool. just wanted to be sure I wasn't missing something obvious :)	19:34
	→ bdiego joined	19:35
Ilari	careo: You just need to set up the permissions so that everybody having commit access can write to that repo. There's 'shared flag' that causes git to do some permission manipulation to make things easier.	19:35
careo	Ilari: ok. so put everyone in the 'git' group, have the repo owned by that group too, and everything's cool assuming no one has a stupid umask?	19:36
Ilari	careo: Make all directories in repo setgid and set the shared flag (it will override group permissions for created files).	19:37
jkp	wow: git can mirror my entire perforce repo!!!	19:37
Mikachu	wasn't that the point?	19:37
careo	Ilari: okay. thanks.	19:37
paakku	jkp: how's the space consumption vs. perforce?	19:38
jkp	paakku: im about to find out :)	19:38
	→ tvachon joined	19:38
jkp	this is exciting stuff!	19:38
	→ schelcj joined	19:38
jkp	i love the idea of gettig Gits advantages for my whole repo...locally	19:38
gitte	Heh.	19:39
Mikachu	jkp: did you think you would just get the last commit?	19:39
Ilari	careo: Well, stupid umasks can still be somewhat of a problem, but now more in 'allows access that shouldn't be allowed' direction.	19:39
jkp	Mikachu: i didnt imagine id get the whole repo	19:39
gitte	I remember how excited I was the first time I imported a repository with git-cvsimport, and worked on it locally...	19:39
jkp	question though ebfore i go any further	19:39
Mikachu	heh	19:39
jkp	im looking at taking my whole main branch from Perforce	19:39
	what is considered a reasonable Git project size?	19:40
schelcj	could someone enlighten me as to how you setup a remote repos with the content and not just a bare repo?	19:40
jkp	i saw in the talk that it should be split up	19:40
careo	Ilari: that's ok in this case. the server it's on is just a development machine with nothing open but ssh and some http servers on occasion	19:40
jkp	now each component in our system might have say a couple of hundred files	19:40
Ilari	schelcj: Ususally not a good idea. If you must, setup hooks to update the checked out copy on push.	19:40
jkp	that doesnt seem massive...could i cope with say, 10000 files in one hit?	19:40
	jkp checks the size of the main branch	19:40
Mikachu	jkp: linux is just under 200MB for the .git dir, and quite a lot of files :)	19:41
schelcj	Ilari: ok, if i setup a bare remote repo, what is the prefered method for getting the contents of the repo to another dev?	19:41
	← jdq left	19:41
Ilari	schelcj: fetch it?	19:42
	→ jdq joined	19:42
jkp	the only thing is lots of the stuff in there isnt really relevant to me	19:42
gitte	gitster: if you saw what we have to cope with on Windows, you would not call git.git "our mess" ;-)	19:42
schelcj	Ilari: guess i am missing something basic here, but if the remote repo is bare, where does another dev get the contents?	19:42
jkp	its for other platforms...im going to be working on OSX and a load of the stuff is just windows related	19:42
schelcj	Ilari: do i just provide it via ftp/rsync/tarball?	19:43
jkp	25k files	19:43
	is that too many?	19:43
Ilari	schelcj: The recommended methods for fetching are dedicated git:// protocol for anonymous access and ssh for autheticated access.	19:43
	jkp wonders if he should do seperate repos for each component he is interested in	19:43
Mikachu	jkp: linux is 23771 files	19:43
	jkp: i think the splitting up argument is about things like the kde repo where every kde program/lib is in the same repo	19:44
jkp	it would be easiest for me to just grab it all...but as i say a shit load is not relevant	19:44
	for example we have a windows Python install in the repo so that people can checkout and it just builds correctly	19:44
schelcj	Ilari: i am still cornfused...	19:44
jkp	and a whole load of third party windows stuff in there	19:44
schelcj	Ilari: what i have read about setting up a git-deamon still means i do a git clone --bare	19:46
	Ilari: i am lost about the contents of the repo when you do a --bare	19:46
Ilari	schelcj: I would make the bare repo by 'git --bare init' and then pushing what one wants there.	19:46
schelcj	Ilari: oh, so start empty then push into it?	19:47
Ilari	schelcj: Yes.	19:47
schelcj	Ilari: ok, guess i can go that route	19:47
Ilari	schelcj: Watch out. It's 'git --bare init', not 'git init --bare'!	19:47
schelcj	Ilari: ok thanks	19:48
Ilari	schelcj: Also see git remote. It can be used to register nicknames for remote repos.	19:49
Mikachu	git has quite a lot of "historical artifacts" for its age :)	19:49
schelcj	Ilari: oh, i thought i had to edit a config for that	19:50
	Ilari: nice, thanks	19:50
	→ dash_ joined	19:53
	aggieben_ → aggieben	20:01
pengwn	has any one worked with collaborating with git and clearcase?	20:02
	would like some pointers links to that .	20:02
	← dash__ left	20:04
	→ johan-s_ joined	20:05
jkp	does this make any sense to anyone?	20:07
	"A useful setup may be that you have a periodically updated git repository	20:07
	somewhere that contains a complete import of a Perforce project. That git	20:07
	repository can be used to clone the working repository from and one would	20:07
	import from Perforce directly after cloning using git-p4"	20:07
	→ hipertracker joined	20:07
jkp	thats sounds to me like you use your initial full clone as a master for other clones	20:07
paakku	yes it does	20:08
jkp	so you keep your initial clone updated from the source, then you clone that to your own local Git repos	20:08
	i dont understand the wording though...maybe the author wasnt english	20:08
	"and one would	20:09
	import from Perforce directly after cloning using git-p"	20:09
paakku	import from Perforce (directly after cloning) using git-p	20:09
jkp	into thid master clone?	20:09
	← bdiego left	20:09
jkp	and then what, use that as the master for your real repos (the ones where you do the work)?	20:09
paakku	it seems to mean that you clone the master to a new repository, then pull from perforce to that new repository	20:10
jkp	that seems odd	20:10
Tv	oh wow	20:10
	$ git --version	20:10
	git version 1.0.0	20:10
jkp	by master are you referring to Perforce or the master Git repo paakku ?	20:10
paakku	git	20:11
Ilari	Tv: Just how old is that?	20:11
jkp	that seems odd	20:11
paakku	I don't know perforce though	20:11
Tv	-rwxr-xr-x 1 root root 7824 Dec 21 2005 /usr/bin/git	20:11
jkp	so i would do PErforcec - > Git Master clone - > other clonse	20:11
Ilari	Tv: Built on the day it was released...	20:11
jkp	and the from other clone use git-p4 directly?	20:11
paakku	I guess it would be faster than running git-p4 on an empty git repository	20:12
Tv	damn that thing has been hosting repos all this time	20:12
paakku	plus maybe it's more likely to generate the same SHA1s in all git clones	20:12
jkp	hrm	20:12
	this is what im doing now...im cloning the entire repo to a local Git repo	20:12
	so lets call this my Git master	20:13
	→ bdiego joined	20:13
jkp	according to that...what would i do next?	20:13
Ilari	Tv: Isn't that bit equivalent to those 'server physically lost' legends? :_)	20:15
	*:-)	20:15
	← loops left	20:15
Tv	Ilari: oh i have some war stories	20:15
paakku	you would clone your Git master to another Git repository, I suppose. But I don't know why the instructions are that way. Why not just use the Git master? Unless perhaps if the idea is that multiple people are cloning the same Git master.	20:15
Tv	Ilari: did an inside-company expert gig setting up a postfix cluster	20:15
	Ilari: they took the machines into the data center, and some months later had lost them	20:16
	Ilari: i made them beep until they found them	20:16
	Ilari: then they didn't know which one was which	20:16
	Ilari: so i ran find / on one and they looked at the hard drive leds	20:16
jkp	paakku: i think thats it	20:16
	to my mind it makes most sense to make a branch to work on in the same repo	20:16
	do stuff there, then periodically rebase master	20:16
	but i guess his way you can keep the mirrior repo set to master and do nothing except rebase that and pull and push to it	20:17
	just seems like an extra layer i maybe dont need though	20:17
	← sverrej left	20:21
	→ loops joined	20:22
	← hipertracker_ left	20:23
	← fujin left	20:25
	→ fujin joined	20:26
	→ moya joined	20:27
	→ jengelh joined	20:30
	← a-priori left	20:30
	← fujin left	20:31
careo	"error: bad graft data" is such a helpful message :/	20:32
gitte	Probably your graft data _is_ bad.	20:32
	How many grafts have you put in there?	20:33
careo	oh, I know it was bad. my script didn't actually flush the entire line to the file	20:33
	that was more of a mild venting than anything.	20:33
	the real culprit lay in ruby's puts method.	20:34
	gitte thinks that it's always ruby's fault when ruby's involved.	20:36
careo	heh	20:37
	most likely	20:37
	but at least this script worked(ish), unlike the shell script that was trying to do the same ;)	20:37
	(of course, that was more my fault than bash's)	20:38
	gitte likes bashing	20:38
careo	I have to admit, bash scripts are growing on me since I started using git	20:39
Tv	careo: write to temp files, rename into place?	20:40
careo	Tv: forcing it to flush the entire line worked. but man, that took forever to debug :/	20:41
	→ sverrej joined	20:42
	← amf left	20:42
	→ docelic joined	20:43
	→ nsdk joined	20:48
	→ fultilt joined	20:56
	gitster takes offence	20:58
gitte	At what?	20:58
	← nsdk left	20:58
gitster	somebody talking as if the scripts we ship are bash scripts.	20:58
gitte	Heh. No, no, we went through _pains_ to make sure that they work with other shells, too.	20:59
	gitte remembers that dash peculiarity.	20:59
	← pombreda left	20:59
	→ a-priori joined	21:00
gitte	BTW my theory is not correct, about that tag auto-fetching	21:00
	I just fetched next and master, and only had to fetch _one_ object for the v1.5.4.2 tag.	21:00
cehteh	mhm	21:00
	is there a function to get the tree sha1 from a commit object in simple way	21:02
gitte	I'd look into how ":" is resolved in get_sha1_<whatever>()	21:02
	Oh, no! slapsherhead	21:03
	I read your question again. Just parse the commit (parse_commit(commit)) to make sure that it is parsed, and then access commit->tree.	21:03
	(Or for the sha1: commit->tree.object.sha1)	21:04
cehteh	yes, just wondered if there is some ad-hoc function for that	21:04
gitte	Why a function?	21:04
cehteh	libgit is a bitch :P	21:04
gitte	You do not need a function when the data is right there.	21:04
cehteh	ok not because of that	21:04
gitte	s/data/datum/	21:05
cehteh	yep parsing commit	21:05
	← pengwn left	21:05
	← lukass left	21:09
	← HG` left	21:13
	← tvachon left	21:15
	→ tvachon joined	21:18
	← tvachon left	21:20
	← TL_CLD left	21:23
	→ fujin joined	21:23
	→ wvl_ joined	21:23
	→ DeepB joined	21:30
	← mithro left	21:32
	← bdiego left	21:34
	→ kampasky joined	21:34
	→ mithro joined	21:41
	→ Ryback_ joined	21:42
	← Eludias left	21:48
	→ SuperX joined	21:52
	→ robbyonrails joined	21:54
	← SuperX left	21:54
robbyonrails	hello all, i'm trying to import a svn repos into git and need to pass my svn username/password somehow	21:55
	am using git svnimport, is there a way to pass user/pass?	21:55
gitte	Note: svnimport is deprecated.	21:55
Eridius	use git-svn	21:56
robbyonrails	ah, i was following this: http://daniel.collectiveidea.com/blog/2008/2/9/on-git-github	21:56
Eridius	huh, wonder why he used git-svnimport	21:56
robbyonrails	its what shows up on the GitHub docs	21:56
Eridius	huh	21:57
robbyonrails	http://github.com/guides/import-from-subversion	21:57
	Eridius just heard about github today through tobias lütke's blog	21:57
robbyonrails	trying: git-svn clone http://path/to/svn/repos -A /path/to/author/file	22:00
	gitte gives up on Len's problem for today.	22:01
	← octopod left	22:04
	← a-priori left	22:09
fujin	I just initiated a bare, shared repository (for publishing) and am now trying to push stuff into it with git push --all git://localhost/puppet.git	22:13
	and it's giving me No refs in common and none specified; doing nothing.	22:14
Ilari	fujin: Try moving that --all into end of command?	22:14
Eridius	Ilari: docs actually show it first, as a flag	22:15
	fujin: you could try an initial git push git://localhost/puppet.git master	22:15
gitte	No, no, I think the problem is git://	22:15
Eridius	or you could try --mirror	22:15
Mikachu	fujin: git://localhost seems pretty stupid, just give the direct path to the repo	22:15
Eridius	since --mirror is explicitly documented as pushing new refs	22:15
fujin	uhm	22:17
	http://rafb.net/p/NLfEru78.html	22:17
	I don't have --mirror afaict	22:18
	what version has that?	22:18
Mikachu	fujin: did you try with the local path and --all?	22:18
	fujin: do you have local commits?	22:18
	← kumbayo left	22:18
Mikachu	are you sure the bare repo is inited correctly?	22:19
fujin	Mikachu: http://rafb.net/p/XnQa3v95.html	22:19
	I believe so, git --bare init --shared	22:19
Eridius	what's the (branch: master) in your prompt?	22:19
Mikachu	are you using a sane git version? ie 1.5 or newer?	22:19
fujin	oh god	22:19
	the branch: master in my prompt is __git_prompt	22:19
	Mikachu: yeah 1.5.2	22:19
	I'm an idiot	22:19
	no local commits	22:19
	forgot I nuked .git before	22:19
Mikachu	^_~	22:19
fujin	working now	22:20
Mikachu	it's good to have this list of default questions sometimes	22:20
Eridius	git prompt??	22:20
Mikachu	Eridius: is there an echo? :)	22:20
fujin	Eridius: yes, it's a git prompt	22:20
	shows current branch	22:20
	ls	22:20
	mt	22:20
Mikachu	Eridius: it comes with git	22:20
Eridius	where? i don't see a git-prompt, nor do I see anything in contrib	22:20
fujin	export PS1="\u@\h \$(__git_ps1 '(branch: %s) ')\w"'\[\033k\033\\\]\$ '	22:20
Mikachu	Eridius: ls completion	22:21
fujin	sorry - the \033k\033 is so that Screen shows my current running app as the app title in screen	22:21
	the important stuff is the \$(__git_ps1 stuff.	22:21
	you need the git bash completion stuff loaded	22:21
Eridius	you're still missing the problem here: I don't know of any bash support for git. Where is this stuff?	22:22
	Eridius runs, will be back in a bit	22:22
Mikachu	Eridius: cd contrib/completion; ls	22:22
fujin	In magical debian land	22:22
gitte	fujin: what does "git rev-parse --verify master" say?	22:22
fujin	oh, it actually comes with git? cool	22:22
Mikachu	gitte: we covered that already	22:22
	:)	22:22
fujin	gitte: it's working now, d21ac29be5d8519fef1619799809c8bf0dd4c324	22:22
gitte	Ah, okay.	22:23
	gitte is always late	22:23
Mikachu	gittardy	22:23
	→ tvachon joined	22:23
fujin	ugh, brb	22:25
	← fujin left	22:25
	← pietia left	22:26
	← alley_cat left	22:27
	→ bschindler joined	22:29
	→ fujin joined	22:29
	→ eternaleye_ joined	22:34
	← docelic left	22:42
	← johan-s left	22:44
	johan-s_ → johan-s	22:44
	← Leonidas left	22:45
	→ a-priori joined	22:47
	← langenberg left	22:50
bobesponja	hi	22:56
	← eternaleye left	22:56
	→ technoweenie joined	22:56
bobesponja	can I make "git log -1" only output the commit message of the latest commit?	22:56
thiago_home	git show	22:56
	git log HEAD^!	22:57
bobesponja	thiago_home: that also shows the commit hash, author name, email etc, I just want to see the commit message, I guess I need to do some parsing	22:59
loops	bobesponja, git log -1 ?	23:00
	oh..	23:00
bobesponja	:)	23:00
loops	bobesponja, git log -1 --pretty=format:%s%b is close	23:01
Pieter	git log --pretty=format:"%s%b" HEAD^!	23:02
	← technoweenie left	23:02
bobesponja	loops, Pieter thanks	23:02
loops	bobesponja, %s%n%b is better :o)	23:04
bobesponja	loops: I don't see the difference but I'm gonna trust you and use it :)	23:05
Pieter	newline between subject and body	23:05
	hmm	23:08
	what does "git checkout" without any arguments do?	23:08
loops	Pieter, it would seem, nothing useful ;o)	23:11
thiago_home	it prints the same output as git-diff --name-status, but slower	23:12
Mikachu	no	23:13
	if you git-add the changed files, git checkout still prints them, but git diff --name-status doesn't	23:14
	:)	23:14
thiago_home	ok, then it's the same as git-diff --name-status HEAD	23:14
	but also slower	23:16
Pieter	:)	23:16
thiago_home	git-diff --name-status is almost instantaneous for me	23:16
Pieter	it's a kinda weird feature	23:16
thiago_home	with the HEAD argument, there's a slight but noticeable delay	23:16
	git checkout has a longer delay	23:17
	← ferdy left	23:19
gitte	Are you talking about the builtin checkout, or the shell script one?	23:19
loops	Pieter, don't think checkout was ever meant to be used without parameters.. so its behavior is just coincidental	23:19
	← bschindler left	23:24
	ebzzry_ → ebzzry	23:27
	→ kevwil joined	23:29
	← sinkorswim left	23:29
	← gitte left	23:30
	← priidu left	23:32
Pieter	loops: yeah, might be.. just wanted to see what it does :)	23:37
	I'd thought maybe it'd do the same as git checkout . , which was what I wanted to do	23:37
	← shaftyy left	23:44
	→ tjafk2 joined	23:55
	← chris2 left	23:56
	← codeunderload left	23:59
	← trochala left	23:59

Logs Search ←Prev date Next date→

Channels Documentation