IRCloggy #git 2012-10-03

Logs Search ←Prev date Next date→ Channels Documentation

Provider of IRC logs since 2005.
WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

2012-10-03

← daniel_- left00:00
← Vile left00:00
← gtrak left00:01
→ Vile joined00:01
← dogarrhea2 left00:02
→ spaceonline joined00:02
→ skorgon joined00:03
→ jperry2_ joined00:04
← kvanderw left00:06
→ cybersphinx joined00:06
← nelse_ left00:08
← jperry2_ left00:09
→ davidban_ joined00:09
← the_cybersphinx left00:10
← aircan left00:10
← donri left00:12
← Hasbro left00:14
→ JoeQuery joined00:14
→ Eridius_ joined00:18
← jesseFromYVR left00:18
CareBear\ retro|cz : that does not make any sense.00:20
← jokajak left00:20
retro|cz CareBear\, why?00:20
That format looks better for me.00:20
CareBear\ retro|cz : a non-git diff does not have any git information so you might as well apply it to your workdir using patch00:20
what format?00:20
retro|cz I'm diffing two outputs of similar files.00:20
→ jperry2_ joined00:20
milki retro|cz: you probably just want to know what the default diff flags git uses00:20
CareBear\ and you want to know how to make diff output unified diffs?00:21
→ jokajak` joined00:21
retro|cz milki, probably00:21
CareBear\ try -u00:21
milki always uses -rub -.-00:21
← Eridius left00:21
→ oxidative joined00:21
retro|cz CareBear\, ok, give me a sec00:21
← ikolev left00:22
← Bombe left00:23
← les_sylvains left00:23
→ maxandersen1 joined00:24
← TeckniX_AFK left00:24
CareBear\ three minutes later00:24
milki LOL00:24
← simesy_work left00:24
retro|cz CareBear\, :-*00:25
→ simesy_work joined00:25
retro|cz I'm playing poker, sry :(00:25
milki wat00:25
← simesy_work left00:25
→ simesy_work joined00:25
retro|cz -u is what I was looking for00:25
milki tis called a unified diff00:26
retro|cz milki, and -rub is much better00:26
ok, thanks00:26
← cakehero left00:26
← breck left00:26
retro|cz thank you CareBear\ and milki00:27
← davidban_ left00:27
← _Vi left00:27
→ cakehero joined00:27
← maxandersen left00:27
→ _Vi joined00:28
→ Bombe joined00:29
→ breck joined00:29
→ ISF joined00:33
← rakm left00:33
← cheriv left00:33
← dpk left00:34
← kenperkins left00:35
← oxidative left00:35
→ macabre joined00:37
← JoeQuery left00:38
← jdunck left00:38
← msmithng left00:39
→ msmithng joined00:39
← msmithng left00:39
→ legumbre_ joined00:40
→ Hasbro joined00:42
→ xiangfu joined00:43
← hyperair left00:44
← spaceonline left00:44
→ Milossh joined00:44
← skorgon left00:45
→ Bosox20051 joined00:46
← dsirijus left00:48
← Eridius_ left00:49
← madsy left00:49
← gavin_huang left00:49
→ mdpatrick joined00:50
← metcalfc left00:50
← ISF left00:52
→ tinti joined00:53
← cyphase left00:54
← hwrd|work left00:54
← breck left00:55
← diegoviola left00:56
→ Eridius joined00:57
← Bosox20051 left00:57
← Eridius left00:57
← ehsan left00:57
→ Eridius joined00:58
→ cyphase joined00:58
→ boombatower joined01:00
→ tmcmahon joined01:01
← pheaver left01:02
→ flippo joined01:03
→ goshawk joined01:04
← pdtpatrick left01:06
← legumbre_ left01:06
← reuf left01:07
← jcao219 left01:08
← boombatower left01:09
rohan32|afk → rohan3201:11
→ macrover joined01:12
← Mocramis left01:13
→ Mocramis joined01:14
← isomorphic left01:16
→ intripoon joined01:16
→ isomorphic joined01:18
← etcetera left01:18
← Hasbro left01:19
← Lgb left01:20
← intripoon_ left01:20
← mmc1 left01:20
→ hyperair joined01:21
← antix_ left01:21
→ Lgb joined01:21
→ antix_ joined01:22
← Milossh left01:24
→ Milossh joined01:25
→ kvanderw joined01:25
← macrover left01:26
← _Vi left01:27
→ Vampire0_ joined01:28
← DancingBear left01:28
→ _Vi joined01:28
→ fedesilva joined01:29
← kvanderw left01:31
← Vampire0 left01:31
→ Targen joined01:34
→ WillMarshall joined01:36
← tinti left01:37
← itatitat left01:38
← Shadeness left01:38
← goshawk left01:42
← takus left01:43
→ takus joined01:44
→ asteve joined01:46
← asteve left01:46
→ asteve joined01:46
→ pheaver joined01:46
← Akufen left01:46
← fedesilva left01:47
→ fedesilva joined01:47
→ steini joined01:49
→ tinti joined01:50
→ kukks joined01:51
← kukks left01:51
→ kukks joined01:51
→ ISF joined01:53
→ hwrd|work joined01:53
→ jalama joined01:54
← hwrd|work left01:54
← Moussekateer left01:54
→ reshtnk7 joined01:55
→ thiago joined01:56
→ les_sylvains joined01:57
← Eridius left01:57
← tinti left01:58
→ jonathan__ joined01:59
← pheaver left02:03
→ metcalfc joined02:07
→ jcao219 joined02:08
← cakehero left02:08
← jonathan__ left02:09
← glennpratt left02:10
→ cakehero joined02:11
→ Eridius joined02:11
→ Eridius_ joined02:11
← tmcmahon left02:12
→ tchan joined02:13
→ intripoon_ joined02:13
← oogatta left02:14
← tavisto left02:14
← bartek left02:15
← intripoon left02:16
→ adamben joined02:16
← Eridius_ left02:17
← SkAzZ1 left02:18
→ ehsan joined02:18
← funnyfingers left02:19
→ zastern joined02:19
← steini left02:21
→ sponkles joined02:22
→ apok_ joined02:22
→ Turicas joined02:23
→ oogatta joined02:24
→ pheaver joined02:24
← jalama left02:25
← apok left02:25
apok_ → apok02:25
← look2thelight left02:28
← _Vi left02:28
→ _Vi joined02:29
← kpreid left02:30
→ kpreid joined02:30
← beneggett left02:30
→ jalama joined02:30
← jalama left02:31
→ kenperkins joined02:32
← cakehero left02:33
← Davey left02:34
← mlb- left02:34
← ehsan left02:35
→ mlb- joined02:35
→ ehsan joined02:35
← felipec left02:37
← kenperkins left02:38
← les_sylvains left02:40
→ williamherry joined02:42
← brentratliff left02:44
williamherry hi, I am reading a book about git, it says that within a git repo, directory that is gited(git init...) will be add as submodule, but what I see is just as normal directoy, can any one know this things?02:44
← orafu left02:45
← zastern left02:46
→ orafu joined02:47
rking williamherry: I don't follow.02:48
← Targen left02:48
rking When you "git init" it creates .git/ as a directory with a bunch of stuff in it. Not sure where submodules would come into play on this one.02:48
williamherry rking: with in a git directory, mkdir test, cd test, git init, git commit --allow-empty init, cd .., git add test02:49
→ DoNotDisturb joined02:49
← DoNotDisturb left02:49
→ DoNotDisturb joined02:49
← DoNotDisturb left02:49
→ DoNotDisturb joined02:49
← DoNotDisturb left02:49
→ DoNotDisturb joined02:49
← DoNotDisturb left02:49
← Samual left02:50
williamherry this will only add test directory,02:50
→ Samual joined02:50
← Samual left02:50
→ Samual joined02:50
← Samual left02:50
williamherry any file within test will not add02:50
rking Yeah, I don't think that's typical02:50
→ Samual joined02:50
← Samual left02:50
→ Samual joined02:50
← Samual left02:50
→ Samual joined02:50
← Samual left02:50
→ Samual joined02:50
→ assem joined02:50
← Samual left02:50
→ Samual joined02:50
← Samual left02:50
rking Not for me anyway; I'm used to submodules living elsewhere and being referenced via "git submodule add <url>"02:50
→ Samual joined02:50
← Samual left02:50
→ Samual joined02:50
← Samual left02:50
→ P1RATEZ joined02:51
→ Samual joined02:51
← Samual left02:51
→ Samual joined02:51
← Samual left02:51
→ Samual joined02:51
← Samual left02:51
→ Samual joined02:51
williamherry rking: like this http://pastie.org/490041802:52
→ DoNotDisturb joined02:52
← DoNotDisturb left02:52
→ DoNotDisturb joined02:52
← DoNotDisturb left02:52
→ DoNotDisturb joined02:53
← DoNotDisturb left02:53
→ DoNotDisturb joined02:53
← DoNotDisturb left02:53
→ DoNotDisturb joined02:53
← DoNotDisturb left02:53
→ DoNotDisturb joined02:53
← DoNotDisturb left02:53
→ DoNotDisturb joined02:53
rking williamherry: Are you opposed to creating the submodule somewhere else first?02:54
← jperry2_ left02:54
← Synthead left02:55
← kpreid left02:55
williamherry rking: I think I got this, `git add others` and `git add others/` are diff, that's what I confuse, thanks02:55
→ beneggett joined02:55
rking Aha, wow.02:55
← Samual left02:55
→ kpreid joined02:55
→ Synthead joined02:56
assem git clone'ing a large repo (emacs and efl) through a proxy finishes receiving objects then does fatal: The remote end hung up unexpectedly, early EOF, and index-pack failed. any hints?02:59
← P1RATEZ left02:59
assem from google'ing i tried git config --global http.postBuffer 3524288000 (large # chosen by me) but same result02:59
EugeneKay Don't use HTTPS03:00
!ssh03:00
gitinfo Please use SSH to talk to remote repos. http:// and git:// are completely unsecured. You can try to do "smart" https://, but it is troublesome to configure(web server/cgi authentication mechanisms, ACLs, etc) and use(client CA certificates, firewalling, protocol inefficiency)03:00
← ajw0100 left03:00
assem EugeneKay: i have no choice at work03:01
→ noyb joined03:02
CareBear\ change work03:02
assem *sadness*03:03
CareBear\ seems another good reason to change work! :)03:03
assem i just thought it seemed fishy that it finishes receiving all objs and blows up (or if not all objs, then like 90%+)03:04
is there any way to do an incremental clone? pick up where it left off?03:04
→ cakehero joined03:05
← beneggett left03:05
← Eridius left03:06
← guampa left03:06
← jschall_ left03:08
← linguini left03:10
→ beneggett joined03:10
← rchavik left03:11
→ Raging_Hog joined03:11
EugeneKay !resume03:11
gitinfo Git does not yet support resuming, torrenting, or parallel fetching of clones or other network traffic. You can make a single file out of a repo using 'git bundle' then use normal resumable download methods to get that file. See https://bundler.caurea.org. Also !gitolite can be told to maintain and send bundles using rsync; see https://github.com/sitaramc/gitolite/blob/master/src/commands/rsync03:11
→ ReekenX joined03:12
← cybersphinx left03:12
→ guampa joined03:12
→ Eridius joined03:14
→ jesseFromYVR joined03:15
rohan32 → rohan32|sleep03:24
← rohan32|sleep left03:25
← fedesilva left03:27
← _Vi left03:28
→ _Vi joined03:29
→ dv310p3r joined03:30
← witul left03:30
← juvenal left03:31
→ juvenal joined03:33
→ rchavik joined03:34
→ jschall joined03:34
← rchavik left03:34
→ rchavik joined03:34
← LucasTizma left03:35
→ BiggFREE joined03:36
→ superthebobfredm joined03:36
← BiggFREE left03:36
→ BiggFREE joined03:37
← jschall left03:38
← BSaboia left03:40
← asakura left03:43
→ asakura joined03:45
← apok left03:45
← maxandersen1 left03:46
← frogonwheels left03:46
→ maxandersen joined03:47
← maxandersen left03:47
→ maxandersen joined03:47
→ maxandersen1 joined03:47
← maxandersen left03:47
→ irqq joined03:48
← kcj left03:49
← fletch_ left03:49
→ kcj joined03:50
← Raging_Hog left03:52
→ jschall joined03:52
→ yoklov joined03:53
→ jdunck joined03:54
→ fedesilva joined03:54
← jwmann left03:56
→ OOPMan joined03:56
DoNotDisturb → Samual03:56
→ quazimodo joined03:56
→ aantix joined04:00
→ jwmann joined04:01
← banghouse left04:01
← sponkles left04:02
← OOPMan left04:02
→ diegoviola joined04:04
← superthebobfredm left04:04
← notfunk left04:04
← armenianeagle left04:05
→ notfunk joined04:05
← pheaver left04:08
→ pheaver joined04:09
← pheaver left04:10
← irqq left04:14
← xiangfu left04:18
→ hwrd|work joined04:19
→ OOPMan joined04:19
← _huoxito left04:21
→ pretty_function joined04:21
→ sangi joined04:22
→ CannedCorn joined04:24
← todd_dsm left04:25
← beneggett left04:27
← asteve left04:27
← _Vi left04:28
→ _Vi joined04:29
← adamben left04:29
← jcao219 left04:31
→ MestreLion joined04:31
→ xiangfu joined04:31
MestreLion is there any way to configure git add (or git commit, or both) to ignore commit hooks for a particular set of files?04:32
← juvenal left04:32
→ beneggett joined04:33
MestreLion use case: i'm setting up a git for a debian source package... and I don't want my hook to fix whitespaces and etc from upstream source code, only for my edited files04:35
← gregm_ left04:36
← kukks left04:37
← tomaw left04:37
→ todd_dsm joined04:38
ChanServ set mode: +v04:40
EugeneKay MestreLion - your hook would have to do that.04:40
It should be a simple matter of abusing diff to find out the changed files04:41
→ frogonwheels joined04:41
MestreLion I could customize my hook to apply changes only to debian/* files04:41
← jesseFromYVR left04:41
MestreLion i guess04:41
→ intransit joined04:44
← CannedCorn left04:47
← runvnc left04:48
MestreLion EugeneKay: the pre-commit hook relevant snippet starts with git diff-index --check --cached "$against" -- | ...04:50
EugeneKay Don't make me bash-fu tonight04:50
MestreLion so if I change that to git diff-index --check --cached "$against" -- Makefile debian/* | , would it get all files from debian *and its subdirs* ?04:50
→ jesseFromYVR joined04:50
EugeneKay !tryit04:51
gitinfo [!tias] Try it and see™. You learn much more by experimentation than by asking without having even tried. If in doubt, make backups before you experiment (see !backup). http://sitaramc.github.com/1-basic-usage/tias.html may help with git-specific TIAS.04:51
MestreLion ok, I promise this won't be bash-fu... I just don't know how git handles dirs when passed as parameters04:51
EugeneKay It considers all files that start with that path04:51
MestreLion nice04:52
so dropping the * and using only debian/ might be all I need... thanks :)04:52
EugeneKay man gitignore has a section on pattern atching as done by git04:52
gitinfo the gitignore manpage is available at http://git-scm.com/docs/gitignore.html04:52
← wonder95 left04:52
EugeneKay debian/* would be globbed by your shell04:52
MestreLion yeah, that's why i think i need to drop the *04:53
so just debian/ is enough for git to recurse all files and subdirs inside?04:53
EugeneKay Yup04:54
MestreLion thanks EugeneKay04:54
EugeneKay Though git doesn't care about dirs at all04:54
MestreLion I know04:54
→ StuckMojo joined04:55
MestreLion I was considering using a gitignore... but I may eventually change the source code.. I just don't want my pre-commit hook to "auto-fix" the upstream files, just my own work04:55
StuckMojo is there something like git-reset --hard that will even get rid of untracked files?04:55
i.e. get me back to pristine working dir state?04:56
MestreLion StuckMojo: rm -rf && git checkout ::P04:56
EugeneKay man git-clean04:56
gitinfo the git-clean manpage is available at http://git-scm.com/docs/git-clean.html04:56
MestreLion wow, there IS a git-clean now??? this is new stuff for me :D04:56
StuckMojo ah thanks04:56
i knew there had to be04:56
EugeneKay 'git reset --hard HEAD; git clean -fdx' should do the jbo04:57
MestreLion IIRC, it didn't until recently04:57
← jesseFromYVR left04:57
← quazimodo left04:58
→ jbub joined05:01
← dv310p3r left05:02
← macabre left05:05
→ private joined05:06
→ nickpresta joined05:06
SeoZ-work[AWAY] → SeoZ05:08
← thiago left05:14
→ AMfish joined05:15
← simesy_work left05:15
→ shruggar joined05:16
← warthog9 left05:16
← Bass10 left05:17
→ simesy_work joined05:18
→ jesseFromYVR joined05:20
← AMfish left05:20
← beneggett left05:21
← nickpresta left05:26
→ thirdknife joined05:26
→ ehsan_ joined05:28
← ehsan left05:29
← _Vi left05:29
→ pantsman joined05:29
← pantsman left05:29
→ pantsman joined05:29
→ beneggett joined05:30
→ _Vi joined05:30
→ j0ran joined05:32
→ quazimodo joined05:32
→ warthog9 joined05:33
→ mrgenixus joined05:33
mrgenixus why is my 'GIT_WORK_TREE' enviroment variable set? and can I just unset it?05:34
← shruggar left05:34
→ imachuchu joined05:34
→ patrick85 joined05:36
_ikke_ mrgenixus: Well, I don't know why it's set05:36
mrgenixus: But you can unset it05:36
→ dc5ala joined05:37
← jschall left05:37
→ Taichou joined05:39
→ jschall joined05:40
mrgenixus ok, wierd stuff05:40
that seems to have fixed me up05:40
← mrgenixus left05:40
← smw left05:41
→ skul joined05:45
→ jceb joined05:45
← akiress left05:46
← aristidesfl left05:47
→ fisted_ joined05:48
→ steveoliver joined05:50
← assem left05:50
← fisted left05:51
→ beautifulmind joined05:51
← warthog9 left05:53
→ smw joined05:54
← AlbireoX`Laptop left05:54
→ devsharpen joined05:54
← cakehero left05:57
→ ThomasLocke joined05:57
← steveoliver left05:57
→ AdrienBrault joined05:58
← hakunin left05:58
→ hakunin joined05:59
← mgorbach left06:00
→ cgs_bob joined06:00
→ cakehero joined06:00
→ maleknet joined06:00
← beneggett left06:00
→ mgorbach joined06:01
→ davidban_ joined06:01
cgs_bob hello all. git newbie here. how can I find the last revision for a file? currently master no longer has this file.06:02
imachuchu cgs_bob: so what you are looking for is the most recent commit that contains the file, correct?06:02
cgs_bob imachuchu: that's right06:03
← AdrienBrault left06:03
→ soee joined06:04
← hakunin left06:04
→ warthog9 joined06:07
← flippo left06:10
→ arturaz joined06:11
_ikke_ git log -- path/to/file06:11
→ koltroll joined06:12
← chaiz left06:12
← jbub left06:15
← beautifulmind left06:17
→ marthinal joined06:17
← Sonderblade left06:19
← yoklov left06:21
→ shruggar joined06:22
imachuchu cgs_bob: did _ikke_'s suggestion work for you (sorry, got busy)?06:22
← StuckMojo left06:23
→ mikepack joined06:23
cgs_bob imachuchu: I just tried _ikke_ suggestion. worked like a charm06:25
_ikke_: thanks a bunch06:25
→ papegaaij joined06:25
← const left06:25
cgs_bob cya later06:26
← cgs_bob left06:26
→ dpino joined06:27
← cakehero left06:27
→ friskd joined06:27
← unreal left06:29
← _Vi left06:29
← noyb left06:30
→ _Vi joined06:31
← aantix left06:31
→ drougge joined06:32
→ unreal joined06:32
→ apok joined06:34
→ jdmax joined06:37
← jdmax left06:37
→ jdmax joined06:37
← shruggar left06:37
→ thierryp joined06:38
→ senny joined06:39
→ subhojit777 joined06:39
← WillMarshall left06:41
← maleknet left06:42
gitinfo set mode: +v06:42
subhojit777 How the commit id is generated? Actually, I want to know whether the commit id is always unique?06:42
Sajbar it's always unique.06:44
→ flijten joined06:44
frogonwheels subhojit777: it uses the sha1 algorithm.06:44
subhojit777: the chances of it not being unique are vanishingly small.06:45
subhojit777 Where the algorithm is applied?06:45
frogonwheels subhojit777: on the contens of the commit object06:47
subhojit777 frogonwheels, right.. thanks06:47
→ harshpb joined06:48
→ shruggar joined06:48
→ raatiniemi joined06:48
→ maleknet joined06:49
→ aisbaa joined06:49
← metcalfc left06:50
→ WillMarshall joined06:51
← arturaz left06:54
← WillMarshall left06:54
← fedesilva left06:54
← shruggar left06:55
← project2501b left06:55
→ soiledpants joined06:56
→ replore_ joined06:57
→ caseymcg joined06:58
→ robustus joined06:59
→ shruggar joined06:59
← ph^ left07:00
← subhojit777 left07:00
← OOPMan left07:00
→ Sky[x] joined07:00
→ project2501b joined07:01
tomaw_ → tomaw07:05
→ suy_ joined07:06
→ adamben joined07:06
← suy_ left07:06
→ giallu joined07:07
← giallu left07:07
→ giallu joined07:07
← shruggar left07:07
→ suy joined07:07
→ afuentes joined07:10
→ angelsl joined07:11
← bigkm left07:13
← hwrd|work left07:13
→ osxorgate joined07:14
→ shruggar joined07:17
_ikke_ see man git hash-object07:18
gitinfo the git-hash-object manpage is available at http://git-scm.com/docs/git-hash-object.html07:18
← shruggar left07:18
→ subhojit777 joined07:18
→ dhruvasagar joined07:19
← Lgb left07:20
← maleknet left07:21
→ Lgb joined07:22
→ witul joined07:22
← ReekenX left07:22
← acrocity left07:24
→ maleknet joined07:25
→ svetlyak40wt joined07:26
→ acrocity joined07:26
← jdmax left07:26
← mikepack left07:29
← Taichou left07:30
← _Vi left07:30
→ mmc1 joined07:30
→ BadDesign joined07:30
← BadDesign left07:30
→ BadDesign joined07:30
→ _Vi joined07:31
← friskd left07:33
← marthinal left07:34
→ ReekenX joined07:34
→ LucasTizma joined07:34
← osxorgate left07:36
→ xxc_michael joined07:36
→ maxagaz joined07:36
maxagaz hi07:36
→ osxorgate joined07:36
maxagaz my friend removed some files in his project and did "git add . / git commit -a / git push", but I still get these files07:37
← Eridius left07:37
← torbjorn left07:37
maxagaz after doing "git pull"07:37
→ Eridius joined07:37
maxagaz is there some cleaning to do somewhere ?07:38
← caseymcg left07:38
→ dangerousdave joined07:38
_ikke_ maxagaz: git add . afaik only adds modified files, it doesn't remove deleted files07:39
modified / new07:39
← guampa left07:39
← derrekito left07:41
esc use more bandwidth07:42
→ tapout joined07:43
→ marhaban joined07:43
← jceb left07:44
← LucasTizma left07:44
→ skorgon joined07:44
cmn maxagaz: you need to tell git that you don't want the files anymore, ask status07:45
maxagaz cmn: how to tell it ?07:45
← Vile left07:45
_ikke_ maxagaz: git status tells you07:46
← eletuchy left07:47
→ friskd joined07:47
← j0ran left07:47
← tapout left07:47
→ j0ran joined07:47
→ tapout joined07:50
← mmc1 left07:50
← tapout left07:50
→ tapout joined07:50
→ WillMarshall joined07:50
→ timj1 joined07:51
← pretty_function left07:51
← Learath2 left07:52
→ exhaze joined07:52
← WillMarshall left07:53
← timj left07:53
→ Learath2 joined07:53
williamherry what git server software github use, gitosis or gitolite, or others?07:53
_ikke_ williamherry: custom made07:54
non open-source07:54
→ shruggar joined07:54
williamherry _ikke_: oh, thanks, why there are two git server package, which one better?07:55
exhaze there's a talk about their stack, should be easily googlable07:55
_ikke_ !gitosis07:55
gitinfo gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite -- if you're already stuck with gitosis we'll try to help, but no promises!07:55
→ stodan joined07:56
williamherry thanks07:56
cmn they use github, hence the name07:56
→ LimeBlast joined07:56
cmn it's basically just some glue code between ssh, a mysql database and mostly-vanilla git07:56
← apok left07:58
← suy left07:59
← xiangfu left07:59
→ WillMarshall joined08:01
→ suy joined08:02
← marhaban left08:02
← osxorgate left08:04
→ eijk joined08:04
→ osxorgate joined08:04
→ mikecmpbll joined08:06
teotwaki that info is wrong, should "no longer maintained nor supported by the author".08:07
should be*08:07
→ sebrock joined08:07
← pantsman left08:07
→ arvind_khadri joined08:07
→ pretty_function joined08:07
cmn being gramatically imperfect doesn't mean that the information is wrong08:08
← WillMarshall left08:12
esc exhaze: you have more details on that talk?08:12
← patrick85 left08:13
→ xiangfu joined08:13
← simesy_work left08:14
exhaze esc: http://zachholman.com/talk/how-to-build-a-github08:14
→ simesy_work joined08:15
esc exhaze: thanks a bunch08:15
← Goplat left08:15
→ Element9 joined08:16
← subhojit777 left08:19
← suy left08:20
albel727 wonders, is there ever going to be a horizontal scroll in commit tree view of gitk08:21
→ Pasdefracse joined08:22
→ daniel_- joined08:23
→ shtrb joined08:24
shtrb how can I get the the last place I did git clone from ? (I did git svn clone url but lost the url)08:25
→ jargon- joined08:25
dr_lepper shtrb: look in .git/config08:25
shtrb dr_lepper, thnx08:25
→ svetlyak40wt_ joined08:26
→ Chillance joined08:26
← imachuchu left08:26
→ dzonder joined08:27
← svetlyak40wt left08:28
→ OOPMan joined08:28
← _Vi left08:30
→ crabmanX joined08:30
→ jkremser joined08:31
→ _Vi joined08:31
← LeMike1 left08:32
→ LeMike1 joined08:32
→ irqq joined08:32
→ subhojit777 joined08:33
→ psoo joined08:34
→ eletuchy joined08:35
SeoZ → SeoZ-work[AWAY]08:36
← acrocity left08:37
← davidban_ left08:39
→ sq-one joined08:40
← xiangfu left08:41
→ e66 joined08:43
→ acrocity joined08:43
→ arietis joined08:43
→ mtrd`w joined08:44
→ vdedik joined08:44
← harshpb left08:47
← diegoviola left08:47
→ harshpb joined08:47
← jargon- left08:48
→ jargon- joined08:48
← MacGyver left08:48
→ Desproges joined08:49
← crabmanX left08:50
→ nelse_ joined08:50
alexander__b when should you use a "real tag" and when should you use a "lightweight tag"?08:50
I don't see the point in signing and writing a tag msg for my use scenario, which at the moment is just to have a pointer to different release versions.08:50
_ikke_ alexander__b: Well, you seem to give the answer yourself08:51
alexander__b so is just doing "git tag -a alpha-1 commithashhere" the correct way to do this?08:52
_ikke_ alexander__b: Just leave out the -a08:52
alexander__b oh OK so just git tag versnr commit08:52
_ikke_ alexander__b: and commit can be left out if you want to tag the current commit08:54
← maleknet left08:55
← babilen left08:56
← wolog left08:57
alexander__b _ikke_: yes, I got that. thanks.08:57
→ blami_ joined08:57
alexander__b unrelated: how do I checkout a file based on commit hash? I want to look at a file long since gone from a repository.08:58
← acrocity left08:58
→ heke80 joined08:58
→ oxidative joined08:58
→ jas4711 joined08:58
_ikke_ alexander__b: git checkout <hash> -- file08:58
→ donri joined08:59
alexander__b ah! thanks08:59
← raatiniemi left08:59
→ marhaban joined08:59
heke80 hi , I have download cakephp zip package and would like to update now with git , git://github.com/cakephp/cakephp.git , how to do it?09:00
→ acrocity joined09:00
→ MacGyver joined09:02
← MacGyver left09:02
→ MacGyver joined09:02
alexander__b _ikke_: uhm, how do I "untag"?09:03
→ PigeonFriend joined09:03
alexander__b _ikke_: oh found it by searching the man for "delete" instead of "remove", sorry.09:03
← xxc_michael left09:03
→ leo2007 joined09:03
→ xxc_michael joined09:03
← vdedik left09:03
→ wolog joined09:04
← BiggFREE left09:05
leo2007 hello, in git-push.txt there is OPTIONS[[OPTIONS]], is that a typo?09:05
alexander__b _ikke_: OK, so my last thing to do now is to push the tags to origin. is git push --tags the correct thing to do? does this just make the tags available on upstream for everyone to pull? because I don't want to push any commits/changes, I only want to share these new tags.09:06
← fireh left09:06
← sq-one left09:07
→ bwidmer joined09:07
→ Vile joined09:08
→ dwijnand joined09:08
→ notVert joined09:09
← jesseFromYVR left09:09
dwijnand How can I make it less painful to work (diff, build commit, commit) while having some local-only changes in the working directory (to circumvent login and other local dev changes)09:11
FauxFaux !config09:11
gitinfo [!configfiles] It is recommended to store local configuration data in a file which is not tracked by git, but certain deployment scenarios(such as Heroku) may require otherwise. See https://gist.github.com/1423106 for some ideas09:11
_ikke_ alexander__b: Well, it also pushes everything reachable by the taggs09:11
tags09:11
→ fornext joined09:12
← hpa left09:12
→ berserkr joined09:13
dwijnand FauxFaux: thanks, I'll see if I find a solution for myself09:14
alexander__b _ikke_: what do you mean by reachable? the commits that are tagged? because they are long since pushed.09:14
→ the_cybersphinx joined09:14
_ikke_ alexander__b: Then they don't have to be sent again of course09:14
→ btree__ joined09:14
alexander__b _ikke_: OK cool. thanks!09:15
_ikke_: so now, when someone pulls, they will get the tags? also, I think github should show this as a notification somewhere, heh.09:16
→ samphippen joined09:16
← PigeonFriend left09:16
→ hpa joined09:16
→ PigeonFriend joined09:17
→ klj613 joined09:18
← btree_ left09:18
klj613 how do i search for all commits (latest to oldest preferably) for the same commit message?09:18
← leo2007 left09:18
esc klj613: git log --oneline | grep $MESSAGE09:19
_ikke_ klj613: what do you mean with the same commit message?09:19
esc is certainly one may of doing it09:19
← angelsl left09:19
_ikke_ esc: git log has the --grep option09:19
esc then it becomes09:19
git log --grep $MESSAGE09:19
i suppose09:19
_ikke_ But depends on what he means with the same commit message09:20
← ReekenX left09:20
→ leeN joined09:21
← psoo left09:21
← eletuchy left09:21
← marhaban left09:24
← unnamedwill left09:30
→ jimubao_ joined09:30
→ jimubao joined09:30
← _Vi left09:30
→ _Vi joined09:31
← Element9 left09:32
jimubao ?09:33
→ babilen joined09:34
_ikke_ ??09:34
→ und3f joined09:36
→ unnamedwill joined09:36
klj613 i meant the same commit message, as i've done rebases i wanted to get list of the multiple tries ive done..09:39
does merge do a different algorithm than rebase? when it _tries_ to solve conflicts09:39
→ FrenkyNet joined09:40
cbreak klj613: kind of09:41
_ikke_ klj613: There are better ways09:41
klj613: git cherry looks at the patch id09:41
← Eridius left09:42
cbreak what does he want to do?09:42
_ikke_ "i wanted to get list of the multiple tries ive done.."09:42
← MestreLion left09:43
klj613 when i merge everything okay, when i rebase it seems to be different results (application logic error, not getting a collection from doctrine)09:43
cbreak git reflog?09:43
klj613 and its not the conflicts i solve.. so must be the auto-solved conflicts. which i'd have to check it all lol09:43
cbreak what's a doctrine?09:43
klj613 doctrien, ORM09:43
cbreak no idea what you are talking about09:43
← jdunck left09:43
cbreak normally, git rebase uses the same algorithm as merging09:43
but of course, with different input09:43
klj613 im doing $entity->getFriends(); type of thing and getting null. (if i use rebase). but if i use merge it works as expected. so rebase/merge doing different conflict solving algorithms i assume09:44
cbreak it's into the other direction after all, and on all individual steps09:44
klj613 well. not sure whats going on then09:44
cbreak klj613: that's unrelated with git09:44
_ikke_ klj613: git doesn't have any conflict resolving algorithms09:44
cbreak $entity looks like bash script or... Yuck ... php09:44
klj613 cbreak, im doing a rebase, which ends up with different results than a merge09:44
cbreak or maybe perl09:44
klj613 = its related to git09:44
→ marthinal joined09:44
_ikke_ doctrine is php09:44
cbreak klj613: look at the results then09:44
klj613: man git-bisect09:44
gitinfo klj613: the git-bisect manpage is available at http://git-scm.com/docs/git-bisect.html09:44
cbreak with rebase that should be straight forward09:44
klj613 okay09:45
→ TheJH joined09:45
cbreak with git bisect, you can find the commit that introduces a bug09:46
then you can look at what that commit does09:46
klj613 okay thanks09:46
→ awilkins joined09:46
cbreak of course, that's not needed if you can look at the code (with a debugger or so) and see what is wrong09:46
← intripoon_ left09:46
← thirdknife left09:46
cbreak but I have no idea if php has anything advanced like a debugger09:46
_ikke_ But that requires the commits to be runnable09:47
awilkins Silly question ; is there a way to clone a repository such that it is an *identical* copy, e.g. - the new repo does not have the old repo as it's origin, instead, it has all the remotes and origin of the first repo09:47
→ intripoon joined09:47
cbreak awilkins: not directly09:48
you can just use rsync though09:48
awilkins This is from the POV of someone who is using git-svn and doesn't want to have to fiddle about setting up the SVN remotes again09:48
cbreak the git protocol does not expose configuration like remote URLs09:48
oh, git svn you can't clone anyway09:48
you will have to copy09:48
→ madsy joined09:49
← madsy left09:49
→ madsy joined09:49
→ beautifulmind joined09:49
← jimubao left09:49
→ jimubao joined09:49
→ thirdknife joined09:49
awilkins cbreak, This is a local clone of an SVN repo, so no need for protocols... it's a fairly large tree, hence I was interested in the hardlinking09:50
→ Dno joined09:50
awilkins It's a question that's been asked before but I couldn't find a simple answer09:50
→ einonm joined09:51
→ i42n joined09:52
→ Mode-M joined09:52
→ chaiz joined09:53
← dwijnand left09:53
cbreak awilkins: so you want to copy it on the same filesystem?09:53
awilkins cbreak, Yes, I know, cp -r seems easiest :-)09:54
cbreak or rsync, yes09:54
← fluxdude left09:54
awilkins Does rsync use hardlinks where available?09:54
jMCg awilkins: if you tell it to. -H09:54
awilkins Not terribly familiar with it.. it's on my list of things to learn09:54
cbreak I would not use rsync's hard linking09:55
that'd hard link the config and branches too I bet09:55
-> super confusing09:55
awilkins Has been pushed to the back of the list by the likes of Dropbox :-)09:55
cbreak dropbox, careful09:55
they are morons09:55
awilkins I don't trust it for anything important or secure09:55
But it's very handy for general managerial nonsenese09:56
And it works better than Ubuntu One09:56
← OOPMan left09:56
awilkins Which is a shame. Neither of them do proxies right (don't understand PAC scripts, despite libraries being easily available)09:57
← adhawkins-away left10:01
cbreak either way, if you want a local copy of a repo, just copy it without hard links10:02
← SeySayux left10:03
→ amcsi_work joined10:03
→ adhawkins-away joined10:03
adhawkins-away → Guest2403310:04
← Villadelfia left10:04
← iori left10:04
→ OOPMan joined10:04
awilkins cbreak10:05
I agree copying is probably easier, came up with a long-winded way of doing it thoguh10:05
→ Villadelfia joined10:05
→ Akufen joined10:05
← mikecmpbll left10:05
awilkins mkdir new-repo ; git clone --mirror old-repo new repo/.git ; cd new-repo ; git config --local --bool core.bare false ; git checkout10:06
→ SeySayux joined10:06
→ mikecmpbll joined10:06
→ xonix joined10:06
← blami_ left10:06
awilkins .. but doesn't clone the SVN stuff. So back to copying :-)10:06
I really do prefer the bzr-svn design to the git-svn design, don't see why it couldn't work for Git TBH.10:07
bzr-svn stores metadata in the SVN repo using it's arbitrary properties feature10:08
FauxFaux awilkins: You can possibly rsync the .git directory over the hardlinked clone to pick up any extra config? I can't think why that wouldn't work, unless you're using alternatives..10:08
← Villadelfia left10:08
← SeySayux left10:08
← jimubao left10:09
awilkins Wheras Git stores metadata in commit descriptions, which rewrites all the revisions you push to SVN, which breaks history10:09
→ jimubao joined10:09
→ SeySayux joined10:09
← jimubao left10:09
→ Villadelfia joined10:09
cbreak awilkins: git stores metadata in additional files10:09
that's why you can't clone it10:10
you can't store the meta data in the svn repo10:10
since that obviously requires write access to the svn repo10:10
awilkins cbreak, The data I'm thinking about only gets generated when you dcommit to SVN, so you need write-access for that10:11
It only gets stored in the local git repo though - the data doesn't end up in the commit comments on SVN, and it rewrites the commit to store the data, which of course, breaks the history because it changes the SHA1 of that revision10:12
cbreak that's not sufficient10:12
you don't have to store any additional data when committing10:12
since once you have committed, the commit is a normal svn commit10:12
and gets svn fetched as normal10:12
← jimubao_ left10:12
cbreak no difference from read only access10:12
git doesn't store anything additional when dcommitting10:13
awilkins It's harder to interoperate with other people using Git though, whereas AFAICT Bazaar has an advantage in that you can share around Bazaar branches that are pulls of SVN branches, and they can all interoperate10:13
← samphippen left10:13
awilkins cbreak, When dcommitting, Git rewrites revisions to store a git-svn commit ID10:14
cbreak in that case, bazaar must store svn meta data in bazaar history10:14
which git doesn't do10:14
(not completely)10:14
awilkins: no10:14
awilkins cbreak, Bazaar stores bzr metadata in svn properties10:14
cbreak it just svn fetches10:14
and then nukes the local commits away10:14
awilkins And it also indexes metadata in a local sqlite db10:14
cbreak and replaces them with what it got from svn10:14
awilkins: as I said: you need write access to the svn repo that way10:14
awilkins cbreak, the commits on svn dont' have the extra data in their commit description that I can see10:15
cbreak you can't store anything in svn properties without write access10:15
awilkins cbreak, Yeah, but you can't dcommit either10:15
cbreak so how would you share around bazaar svn clone branches10:15
if you can't modify the svn repo?10:15
git can do it just fine10:15
you can share around git svn cloned repositories as you would with a normal git repo10:16
the only thing that won't work is communicating with the svn remote10:16
because THAT needs additional meta data10:16
awilkins You can still share around the bazaar branches ; I think my point is that if you push from a bazaar branch to SVN your history remains interoperable with the other peoples bazaar branches10:16
But if you dcommit from git, you rewrite your history so everyone elses Git clone history is now disconnected from the SVN10:16
→ suy joined10:16
cbreak that's normal10:17
awilkins And you can just pull from either bzr or svn if you know the URLs10:17
cbreak and it's unavoidable10:17
since once you dcommit something, it is linearized10:17
it belongs to svn history10:17
awilkins I think Bazaar avoids that rewriting because rather than storing SVN metadata in it's own branches, it stores bazaar metadata in the SVN, and indexes SVN metadata in a temporary store that can be reconstituted from read access to the SVN repo10:20
When you start off pulling an SVN branch from Bazaar it churns away at the metadata for a while (you could share the sqlite around to speed this up, but you don't have to )10:20
→ knowj joined10:20
cbreak so bazaar's svn integration is completely worthless10:21
unless you actually have write access10:21
awilkins Why is it worthless?10:21
cbreak because it can't work10:21
awilkins I can confirm that it does work10:22
cbreak without you modifying the svn repo10:22
so, try it.10:22
awilkins You don't have to modify it to pull10:22
cbreak how do you store the meta data then?10:22
how do you store which svn subdirectory was cloned?10:22
how do you store which branch and tag directory was cloned?10:22
awilkins Which metadata, the metadata of the revisions you pushed to the SVN repo, when you can't because you don't have write access?10:22
cbreak how do you store the clone base?10:22
knowj I've made 2 branches for some magento updates I'm doing 1: update magento from 1.3->1.7.0.2 2: install a new payment system. Do I need to merge the 1: update 2: payment module install with the master or can I just merge the 2nd branch to merge all changes?10:23
cbreak committing to svn is not that important10:23
awilkins You just instruct Bazaar to clone an SVN url10:23
cbreak the most important part is being able to read it in the first place10:23
awilkins It reads the remote repo and constructs a local cache of metadata about it10:23
← sh4rm4 left10:23
awilkins If other people with write access have used bazaar to push revisions there, that metadata will be present in the svn: properties10:23
cbreak knowj: merging gives you all history10:24
look at the ancestor graph or so10:24
knowj cbreak: So if I merge the payment module branch it will also merge the changes for the 1.3 -> 1.7 update?10:24
→ sh4rm4 joined10:24
← Vile left10:24
cbreak depends10:24
is that change in the history?10:24
awilkins Bazaar only stores metadata in the SVN if you push revisions to it - which needs write access, but hey, so does dcommit10:25
cbreak no10:25
knowj cbreak: The payment update branch changes files but not all the same files as the 1.3->1.7 update10:26
cbreak dcommit is not required10:26
knowj: irrelevant10:26
awilkins dcommit is required if you want to store revisions in the SVN repo? Or is there another way?10:26
cbreak file changes don't matter10:26
← sluther left10:26
cbreak awilkins: normal svn commit10:26
knowj cbreak: Sorry I'm very new to git. I've mainly been using it badly as a quick rollback and a safeguard for updates10:26
awilkins cbreak, Well, yes, but you need write access for that too10:26
cbreak awilkins: someone does.10:27
other users can commit too10:27
they won't create any meta data10:27
can bazaar handle that?10:27
then it doesn't need that meta data10:27
→ avjpl joined10:27
← avjpl left10:27
awilkins It's the extra metadata about the interaction of bzr with svn that is the interesting bit10:27
cbreak knowj: if you made a branch, that branch was branched off something10:28
→ samphippen joined10:28
awilkins So other people using bzr as their commit client will generate metadata10:28
← e66 left10:28
cbreak awilkins: that's hardly relevant10:28
awilkins Why not?10:28
cbreak if multiple people use something modern, why waste time with a crappy svn server?10:28
← awilkins left10:28
→ awilkins joined10:29
awilkins sorry, thought ctrl-w was delete word10:29
Apparently in this client it's "leave"10:29
cbreak "Close Window" probably10:29
awilkins People use crappy SVN because they are being forced to by their corporate overlords10:29
All sorts of reasons10:29
cbreak knowj: you can do a git log branchname10:29
that will show you all commits in a branch10:30
do that on both your branches10:30
awilkins The project they are contributing to is stuck in the past, etc10:30
cbreak you can also do a git log --graph --oneline --decorate --all10:30
awilkins: exactly. So the client has to accomodate svn.10:30
awilkins It's not an argument to say "just use a better VCS" when discussing modules which permit VCS interop - obviously, if the newer VCS were the better option you'd be using it10:30
cbreak if you make a branch in svn, what happens?10:30
a copy into a branch directory happens10:31
← _Vi left10:31
cbreak git svn handles that of course10:31
awilkins cbreak, You can pull that if you want10:31
cbreak what happens if you branch in git?10:31
just a new ref to a commit10:31
dcommit does NOT preserve git branches.10:31
but git svn CAN make svn branches10:31
why is that so?10:31
that is so svn users also see the branches10:31
→ _Vi joined10:32
cbreak the fact that a branch was created is stored the svn way10:32
← shtrb left10:32
→ eletuchy joined10:32
cbreak git svn can extract branch creation from svn history without additional meta data10:32
and it stores it in svn without additional meta data10:32
which gives svn users maximal fidelity10:32
→ xiangfu joined10:32
← xiangfu left10:32
cbreak of course, you could just store some git specific meta in there10:32
but that'd be worthless for svn users10:32
awilkins We're not talking about branch creation ; we're talking about rewriting the history in the Git branch10:33
cbreak git branches don't rewrite history10:33
awilkins dcommit does10:33
cbreak as I said, FETCHING svn history does10:33
nope10:33
the second part of dcommit does10:33
getting the new svn history you just committed10:33
awilkins Ok... sorry, the result of doing a dcommit is that you end up with your branch at the tip of the SVN branch, not the branch you were on before10:33
irqq I have a ./foo file and ./bar directory. I want to mv foo bar, so I have bar/foo, but also I'd like to modify my repository as if it always was bar/foo. How?10:33
← oxidative left10:33
awilkins Which means that your tip is no longer the tip of anyone who for example, pulled from your branch10:34
knowj cbreak: It looks like payment-module was branched from magento-update. I merged payment module into magento-update then magento-update into master10:34
cbreak irqq: man git-filter-branch10:34
gitinfo irqq: the git-filter-branch manpage is available at http://git-scm.com/docs/git-filter-branch.html10:34
cbreak irqq: probably --index or --tree filter10:34
→ replore__ joined10:34
cmn consider it's !rewriting10:34
gitinfo [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to jump through hoops (and you have to tell them to), so it's infinitely better to just move on without rewriting. That said, you can use `git push -f` to force your new history through (may get rejected, though). See http://git-scm.com/docs/git-rebase.html#_recovering_from_upstream_rebase10:34
← rubious left10:34
awilkins And that person who pulled from you now has a totally different history back to the point where your SVN branch originated from10:34
Which Bazaar doesn't do10:35
cbreak awilkins: think about an svn user10:35
what does he see?10:35
awilkins A normal svn branch10:35
cbreak it should be obvious that a git/bazaar user has to see the same thing as an svn user10:35
awilkins He doesn't care. If he looks real hard he'll see some extra revision properties10:35
cbreak so you can not by principle store additional meta data that is required10:35
← replore_ left10:36
cbreak because svn users won't see that.10:36
awilkins That mean nothing to his VCS, they're just key / value paris10:36
cbreak and they won't create it if they do the same thing10:36
← adamben left10:36
← replore__ left10:36
awilkins The data isn't there for the SVN user - it's there to make the Bazaar users life easier10:36
cbreak the result of creating a branch in svn or git is the same10:36
awilkins You don't care about revisions the SVN users generated, because you can read the metadata from them already10:36
cbreak why is it the same?10:36
because that's the purpose of an scm10:37
awilkins But doing stuff in Bazaar has extra metadata which you would like to see10:37
← eletuchy left10:37
cbreak there's no point if you have different views on history10:37
awilkins But if you exchange revisions via SVN you won't see it10:37
Unless you put a copy there10:37
→ sobczyk joined10:37
sobczyk is it possible do do a shallow submodule clone?10:37
cbreak see, git is much better10:37
svn users get everything10:37
sobczyk: hmm?10:37
sobczyk I'd be helpful for my automated build machine10:38
awilkins cbreak, Well, that ain't true - you already said as much by saying that SVN won't see your GIt branches10:38
Unless you use git to push them there10:38
sobczyk I have a lot of history in my submodules and a full history is not needed on a CI machine10:38
← harshpb left10:40
cbreak sobczyk: you mean a temporally shallow clone?10:40
sobczyk yes10:40
cbreak you can clone them manually instead of git submodule update --init I guess10:40
no guarantees that that works10:40
sobczyk replace git submodule update --init10:40
→ harshpb joined10:40
cbreak and you would not gain much of that10:40
awilkins cbreak, Now, if I'm using Git, and I clone my co-workers clone of an SVN branch, including revisions he's not dcommitted, and he then dcommits them, any revisions I have committed are now orphans. Whereas with Bazaar, I can pull from his repo, or the SVN repo, it doesn't matter, and it just works.10:40
cbreak history is rather small10:40
sobczyk though the problem is when submodule does not point to the default head10:41
cbreak awilkins: not orphans10:41
awilkins cbreak, Revisions with a parent that is no longer a part of the history on SVN10:41
cbreak they were not in SVN's history anyway10:41
→ angelsl joined10:42
awilkins Because the revisions you get from doing a fetch are no longer the same revisions that are the ancestors of mine10:42
cbreak you will get the dcommitted commits from svn10:42
awilkins Whereas with bzr they are the same revisions10:42
→ noisebleed joined10:42
← noisebleed left10:42
→ noisebleed joined10:42
← skorgon left10:42
cbreak awilkins: it is impossible that they are the same10:42
after all, if you commit in svn, you create a new commit at the end10:42
what happens if someone else committed in between?10:42
with git: no problem10:43
the dcommitted commit is not the same as the one you used as source for dcommitting10:43
it has a different history10:43
awilkins With git you have to do the same as bzr which is rebase10:43
So no different10:43
cbreak (because it has a different history in svn)10:43
you don't need to do a rebase10:43
dcommit does that as part of its operations10:44
awilkins That's just semantics10:44
If dcommit does a rebase because it has to, then you have to rebase10:44
Just because it happens automatically doesn't mean it doesn't happen10:44
cbreak it doesn't do a rebase10:45
it does something that looks like one10:45
what it does is svn commit the new stuff10:45
awilkins Quack < sounds like a duck10:45
cbreak and then svn fetch the new history from svn10:45
as you can see, no rebase happened10:46
awilkins It can't commit the new stuff - someone else committed a revision in the meantime10:46
cbreak sure it can10:46
awilkins Now, OK, SVn lets you do that10:46
But git doesn't10:46
cbreak git svn does10:46
I've done it a few times10:46
if the svn commits don't conflict it's ok10:46
if they do conflict, then you have to do a git svn rebase10:46
which is a real git rebase10:47
awilkins The end result is as if you did a rebase anyway10:47
cbreak no10:47
it doesn't do git's advanced rebase resolution10:47
awilkins So who cares about how you get there - just like git doesn't care if you delete a whole tree and replace it (which SVN does)10:47
cbreak it's just svn's normal committing10:47
it is, as I said above, exactly what an svn user would do10:48
awilkins Incidentally, I greatly prefer Git there .. bazaar will slavishly follow the SVN instructions to delete all the files which breaks their history10:48
But Git doesn't care because it versions content10:48
← wolog left10:48
cbreak git history is a chain of snapshots10:49
awilkins Yes, I know that10:49
cbreak svn history is per-file10:49
it's quite an impedance missmatch10:49
awilkins So is Bazaars10:49
→ wolog joined10:49
→ replore joined10:49
cbreak but since git can reproduce most meta data from the content10:49
→ whitman joined10:49
cbreak it needs a lot less meta data than svn10:49
awilkins In this case, the mismatch is beneficial10:49
cbreak so it can reproduce sane history from weird stuff10:49
like bunch-of-tar-files10:49
or as in this case, chain-of-svn-commits10:49
awilkins Because we have several people on the project who can't use a VCS properly and just delete whole trees and replace them with the one they've been patching in another folder10:50
cbreak sounds like an svn merge.10:50
awilkins No, it's just lame user10:50
One revision where every file is deleted and then mostly replaced with an identical copy is someone copying a tree from one folder to another, probably in an IDE that automates SVN deletes10:51
This is why I stopped using bzr-svn and started using git svn10:51
Because git just thinks "oh, it's that tree again"10:51
Whereas bzr believes svn when it says all the files are new and breaks the history10:52
cbreak there are sometimes people who complain about git not storing moves/deletes10:52
git stores almost no meta data at all. File name, executability, that's it.10:53
→ zloy joined10:53
cbreak probably makes it more robust than svn :)10:53
awilkins So I prefer Git because it copes with the stupids better... and it's fast, and good, etc. But I like bzr-svn better (having had experience of both) because it doesn't change the tip of your local branch to something else when you push revisions to SVN10:54
← j0ran left10:54
awilkins I think the "porcelain" on Bazaar is probably marginally better too (when it comes to ease-of-use and graphical utils)10:55
cbreak more meta data means more fragility10:55
awilkins I can cope with that but my average users can't10:55
Although my opinion was formed first a few years ago when Git on Windows was relatively immature (I use Linux but I have to produce tools for Windows users)10:56
cbreak stay away from tortoise git10:56
awilkins I do, I'm a "CLI + GUI logging" kinda guy10:56
Invoke git-gui and gitk from the command line10:57
Same with bzr but qbzr log etc10:57
→ j0ran joined10:57
→ shtrb joined10:57
← hyperair left10:58
awilkins Also, Bazaar was more accessible to me for hacking because it's Python and my C is weak10:58
My Python was weak too, but easier to learn10:58
And it worked the best on Windows at the time (Mercurial - didn't like capital letters in paths, Git - too flaky on Windows, too hard for users)10:59
Mode-M imo theres no need to hack git ;)10:59
→ maleknet joined10:59
cmn there is if you want it to run well on Windows11:00
awilkins If Git on Windows had GUI tools as mature as TortoiseSVN it would be great11:00
But I'm not sure some of the visual metaphors are there for the DVCS stuff yet11:00
Mode-M mmh, I'm fine with msysgit on windows11:01
_ikke_ A TortoiseSVN kind of interface doesn't really make sense for git11:01
awilkins _ikke_ Hence we need newer visual metaphors for DVCS11:01
_ikke_ I found things like GitExtensions quite good on windows11:01
awilkins I don't mean slavishly copy all the dialogs from TSVN, but some are still relevant11:02
And need newer ones for branch / remote reference management / pull etc11:02
awilkins looks at GitExtensions11:02
awilkins Not really available when I started evaluating Git on Windows and I mostly use Linux so not had a good look at this11:03
← exhaze left11:04
← Mode-M left11:05
← danielfi_ left11:06
→ Ecesis1 joined11:06
← jargon- left11:08
← Ecesis left11:09
→ iguanna joined11:09
iguanna hi all, if I want to create my own repository i can do: git remote add myname git@gitosis:myname.git right?11:10
cmn maybe, if gitosis creates them automatically, also !gitosis11:11
gitinfo gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite -- if you're already stuck with gitosis we'll try to help, but no promises!11:11
→ SkAzZ joined11:13
→ torbjorn joined11:14
→ jalama joined11:18
← jalama left11:18
→ gusnan joined11:19
→ blami_ joined11:21
← heke80 left11:23
← maleknet left11:26
← devsharpen left11:29
→ RaceCondition joined11:31
← _Vi left11:31
→ sgambino joined11:32
→ _Vi joined11:32
← reshtnk7 left11:32
← Dno left11:37
→ _huoxito joined11:38
← svetlyak40wt_ left11:38
→ asteve joined11:38
→ svetlyak40wt joined11:38
→ maleknet joined11:39
cbreak iguanna: git remote add does not create a repository11:39
← btree__ left11:39
cbreak if you want to create a git repository, use git init (or indirectly via git clone)11:40
iguanna ok, I think to real about git, it cann't get the picture11:40
*read11:40
← maxagaz left11:43
→ Xethron joined11:46
Xethron heya, I was wondering, is there a way to export the "diff" view to like pdf or html?11:47
→ Dno joined11:47
Xethron I would like to show a friend not using git what changes I made to his code11:47
← C8H10N4O2 left11:48
wereHamster export it as plaintext11:48
Xethron meh11:48
wereHamster then send your friend an email and paste the diff there.11:48
Xethron I like the colors11:48
:P11:48
hehe11:48
ok11:48
→ amcsi_work_ joined11:50
← amcsi_work left11:50
→ Davey joined11:50
← harshpb left11:51
← amcsi_work_ left11:51
← _Vi left11:52
→ adamben joined11:52
← OOPMan left11:52
→ devsharpen joined11:52
→ erichynds joined11:54
→ jtri joined11:55
jtri hello, i finished my branch code, and did a 'git merge master'11:55
that went fine11:55
→ ReekenX joined11:55
jtri then i did 'git checkout master'11:55
and 'git merge mybranch'11:56
the second merge didn't seem to do anything11:56
then i did a checkout of the branch again, and did a push11:56
and did a checkout of the master and did a push11:56
then i logged into my server and tried a pull11:56
but the server is not getting the update?11:57
any ideas/advice/debug help is appreciated11:57
← mikecmpbll left11:59
→ C8H10N4O2 joined11:59
awilkins Xethron, Get your friend to use an editor that colourises universal diff?11:59
← dc5ala left11:59
→ mikecmpbll joined12:00
cmn jtri: that second merge most likely did a fast-forward, where it didn't create a new commit but pointed the branch to the same commit12:01
← lu_zero left12:01
cmn push isn't branch-dependent per default, see push.default in man git config12:01
gitinfo the git-config manpage is available at http://git-scm.com/docs/git-config.html12:01
FauxFaux Xethron: There are tools named things like "ansifilter" that'll convert terminal colour codes into html for you.12:01
→ lu_zero joined12:01
← lu_zero left12:01
→ lu_zero joined12:01
← sangi left12:02
jtri cmn: hmm... so what should i do?12:02
should i checkout the branch on the server?12:03
cmn we don't know what you want to do12:03
the pull probably indicates you're trying to do !deploy12:03
gitinfo Git is not a deployment tool. You can build one around it for simple environments. http://sitaramc.github.com/the-list-and-irc/deploy.html12:03
jtri oh, i finished the branch, and wanted to merge it back to master12:03
← maleknet left12:03
cmn pull is precisely the wrong tool12:03
jtri but made a mistake i guess ?12:03
cmn for that, you checkout master and do git merge other-branch12:03
→ Element9 joined12:04
→ xiangfu joined12:04
← Davey left12:05
→ harshpb joined12:06
→ mmc1 joined12:06
→ geekbri joined12:08
→ juvenal joined12:08
→ maleknet joined12:08
← ReekenX left12:09
jtri here are the results of 'git branch -a': http://pastie.org/490209512:09
→ cannonball joined12:10
jtri shoot, i think i made a mistake about which branch i merged!12:10
← retro|cz left12:11
→ psoo joined12:12
→ alvesjnr joined12:13
alvesjnr hi all. I changed a file (didn't commit) and want to revert it12:13
how can I do this?12:13
cbreak alvesjnr: git checkout -p filename12:15
→ e66 joined12:15
→ retro|cz joined12:15
← irqq left12:15
williamherry git rm --cached filename12:15
cmn williamherry: and that's meant to be what?12:16
→ hyperair joined12:16
→ irqq joined12:16
williamherry remove from index area12:17
cmn: am I wrong?12:17
→ ajw0100 joined12:17
cbreak williamcotton: you are wrong insofar that it does not answer the asked question12:17
→ pmorris joined12:17
alvesjnr tk12:17
→ jargon- joined12:18
williamherry oops, sorry12:18
← subhojit777 left12:18
gitinfo set mode: +v12:18
pmorris Where does git store the temporary commit message?12:18
→ bapa joined12:18
cmn .git/COMMIT_EDITMSG12:18
→ jperry2_ joined12:20
Xethron Thanks guys :D12:20
→ skyf joined12:23
→ hwrd|work joined12:23
← alunduil left12:23
→ tinti joined12:23
← harshpb left12:25
← TheFuzzball left12:27
→ thiago joined12:28
← dhruvasagar left12:28
→ khmarbaise_ joined12:30
→ guampa joined12:33
→ funnyfingers joined12:34
→ noscript joined12:39
← kcj left12:40
← daniel_- left12:41
→ prudnikov joined12:41
→ tmcmahon joined12:42
← asteve left12:42
→ asteve joined12:42
jokajak` → jokajak12:44
← shtrb left12:45
→ ReekenX joined12:46
mtrd`w Hello! I try to get some files from a distant git repository. I first use "git clone ----" and i get the "origin" version. I'd like to get a newer version in origin/name. How can i do that ? Thank you!12:48
_ikke_ mtrd`w: git clone gives you the latest version12:48
cbreak mtrd`w: git checkout that branch12:48
FauxFaux mtrd`w: 'git checkout name'12:48
_ikke_ Oh, you want a different branch12:49
mtrd`w yes :)12:49
okay, i'm on it!12:49
← skyf left12:49
→ dv310p3r joined12:49
→ leo2007 joined12:49
→ teplyakoff joined12:49
→ d3vic3 joined12:51
→ Tohuw joined12:51
mtrd`w yeah! thank you :) have a nice day12:51
Tohuw How can I view the contents of a directory from a previous commit?12:51
→ freeed joined12:51
_ikke_ Tohuw: View in what way?12:52
cbreak Tohuw: ls-tree maybe? :)12:52
← BadDesign left12:52
→ jbrechtel joined12:53
← Pasdefracse left12:53
Tohuw cbreak: Yes, that's what I need. Thanks!12:55
← whitman left12:55
← dangerousdave left12:55
→ banghouse joined12:56
→ skyf joined12:56
→ dangerousdave joined12:56
← C8H10N4O2 left12:57
→ dansan joined12:58
← adamm left12:58
← samphippen left12:58
→ bartek joined12:59
→ adamm joined12:59
→ guns joined12:59
→ jesseFromYVR joined13:00
dansan I'm having a bit of trouble with rebase --onto. On my branch 'mmotm' (based off of remotes/mmotm/master), I have 30-odd patches. I want to rebase those onto my 'next' branch (which is based off of remotes/linux-next/master), so essentially copy my patches over. I know I can just do format-patch and them git am them, but I'm trying to learn this better13:00
mmotm appears to be broken at the moment, so I want to test my patches on something that works :)13:01
→ chaiz_ joined13:03
← erichynds left13:04
← guns left13:04
→ erichynds joined13:04
→ cilly joined13:04
→ aristidesfl joined13:05
← chaiz left13:05
→ daniel_- joined13:07
→ Davey joined13:07
← arietis left13:07
→ hellopat joined13:08
← jtri left13:08
→ C8H10N4O2 joined13:09
→ cakehero joined13:09
pmorris My terminal hung up but git commit is still running and the editor is still running13:10
How can I recover my commit message from the running editor?13:10
cbreak if the editor runs13:11
just save13:11
and exit the editor13:11
that will make git commit continue13:11
pmorris But I can't take control of the editor since it belongs to a dead terminal13:11
← alexander__b left13:11
cbreak then copy .git/COMMIT_EDITMSG or so13:11
pmorris That's where it saves the message but the message wasn't saved13:12
The message only exists in memory of a process belonging to a terminal that has hung up13:12
_ikke_ pmorris: try ctrl+aq13:12
pmorris: try ctrl+q13:12
cbreak pmorris: then attach a debugger13:12
and analyze the RAM of the editor13:12
← chaiz_ left13:12
cbreak the message must be somewhere in there13:12
_ikke_ pmorris: If you happened to hit ctrl+s, then you froze the terminal13:12
pmorris I didn't hit anything, the terminal HUNG UP13:13
Xethron well13:13
→ alexander__b joined13:13
_ikke_ pmorris: well, that's the same ctrl+s does13:13
pmorris The network dies every 30 minutes on idle connections13:13
→ chaiz joined13:13
_ikke_ pmorris: Are you working over ssh?13:13
Xethron Is it a terminal emulator?13:13
pmorris Yes I am SSHing13:13
Xethron Like running inside a x11 enviroment?13:13
Can't you just select the text and copy it?13:14
cbreak I say: Debugger13:14
that's the only way to be sure13:14
pmorris Normall when the shell hangs up nano dumps its contents into *.save and exits13:14
But because git spawned it and git doesn't terminate on hangup it's still alive13:14
There's probably a signal that can be sent to the nano process to make it dump to *.save and exit but I don't know which one and if I get it wrong it will exit without dumping13:15
Actually I can probably just send HUP13:15
cbreak why nano?13:15
you should use vim13:15
pmorris No, fuck you13:16
_ikke_ !fsck13:16
gitinfo [!profanity] Hey! This is a family-safe channel, so keep it frakking clean you fierfekker!13:16
→ samphippen joined13:16
→ macabre joined13:16
← suy left13:17
→ suy joined13:17
FrenkyNet → FrenkyNet|work13:18
dansan cbreak: you should use ms notepad13:18
cbreak dansan: that thing doesn't work13:18
dansan so?13:18
← pretty_function left13:19
→ _Vi joined13:19
cbreak dansan: honestly, get a brain13:19
pmorris: honestly, get some manners, and then get a brain13:19
dansan cbreak: ahh, that's all the advice I needed, thanks so much! :)13:19
cbreak no problem.13:19
dansan I don't like nano either, but some people do13:19
← Lgb left13:20
cbreak so?13:20
dansan no reason to tell somebody to change just because it's not the editor you like13:20
← ehsan_ left13:20
cbreak if you want to remotely connect to an other editor instance13:20
then vim can do that13:20
dansan but good come back anyway! :)13:20
cbreak so, use vim if you want to do something vim can do13:20
got it?13:20
dansan what if you want to do something that some other text editor can do as well?13:20
cbreak http://stackoverflow.com/questions/2760014/how-can-i-configure-vim-so-that-when-i-send-the-process-a-usr1-signal-it-saves-a13:20
← dpino left13:21
dansan shouldn't be too hard to patch nano or emacs or some such to do the same13:21
→ Yuuhi joined13:21
cbreak do it13:21
and send pmorris the patch13:21
dansan I use vim, lol!!13:21
← kpreid left13:21
← e66 left13:22
→ kpreid_ joined13:22
→ v0n joined13:22
→ Lgb joined13:22
← samphippen left13:22
→ chuckharmston joined13:22
← tmcmahon left13:22
← psoo left13:22
→ psoo joined13:23
→ samphippen joined13:24
← chaz left13:24
← TomWij left13:24
→ TeckniX joined13:25
→ Gx4xx joined13:25
→ TomWij joined13:26
→ legumbre_ joined13:27
← RaceCondition left13:27
← Gx4 left13:28
← j0ran left13:28
→ j0ran1 joined13:28
← bartek left13:28
← prudnikov left13:29
→ bartek joined13:29
→ RaceCondition joined13:29
→ chaz joined13:31
→ dpk joined13:31
← madewokherd left13:32
→ les_sylvains joined13:34
→ achadwick joined13:34
→ nwest joined13:35
← RaceCondition left13:36
→ alunduil joined13:37
← adamben left13:37
→ crispus joined13:37
← ReekenX left13:38
← alunduil left13:39
→ Destos joined13:39
← dpk left13:39
→ alunduil joined13:39
→ alunduil_ joined13:40
→ Hasbro joined13:40
→ pfFredd joined13:41
→ jbrechte_ joined13:43
← cagedwisdom left13:45
← jbrechte_ left13:45
→ jds_ joined13:47
→ e66 joined13:47
← e66 left13:47
jds_ Heya13:47
← jbrechtel left13:47
jds_ Anyone familiar with libgit2? I'm trying to figure out why gitx is really slow on my repo...13:47
I see an awful lot of calls to git_path_dirload_with_stat on directories that are ignored by my gitignore13:48
was wondering if that was expected13:48
→ derpops joined13:48
→ msmithng joined13:50
← samphippen left13:52
→ lacrymology joined13:54
→ glennpratt joined13:54
← dansan left13:56
← Turicas left13:56
← psoo left13:57
→ isomorphic_ joined13:57
← lu_zero left13:58
→ logik-bomb joined13:58
logik-bomb how can I check who created a remote branch?13:58
_ikke_ logik-bomb: you can't13:58
logik-bomb :X13:58
← isomorphic left13:59
← beautifulmind left13:59
_ikke_ logik-bomb: a branch in git is just a pointer to a commit13:59
→ prudnikov joined13:59
_ikke_ logik-bomb: It's not recorded who created it13:59
← mikecmpbll left14:00
logik-bomb I see14:00
cmn fetch, look at the log and chances are whoever has the latest commits is the one who pushed it14:00
logik-bomb I'm just doing some clean up14:00
→ dhruvasagar joined14:00
logik-bomb and it would be usefull to know who owns what14:00
:)14:00
thanks cmn14:00
and _ikke_14:00
→ lu_zero joined14:00
← lu_zero left14:00
→ lu_zero joined14:00
← bartek left14:00
→ bartek joined14:01
cmn the most useful is to just ask your coworkers14:01
← juvenal left14:01
→ samphippen joined14:02
← SJr left14:02
← xxc_michael left14:02
← soee left14:04
← eijk left14:04
← samphippen left14:04
← bapa left14:05
← xiangfu left14:05
→ bapa joined14:05
→ samphippen joined14:05
→ ehsan joined14:07
← hyperair left14:08
→ Vile joined14:08
→ whitman joined14:08
→ RJ3000_ joined14:09
← RJ3000_ left14:09
← dr_lepper left14:09
→ RJ3000_ joined14:09
← Davey left14:10
→ juvenal joined14:10
← nelse_ left14:11
← skyf left14:11
→ talha joined14:12
→ trungqtran joined14:14
→ revagomes joined14:14
→ e66 joined14:16
C8H10N4O2 I have 7 repos on my server and so far the two other developers have had access to all of them because they are working on all of them. Now I am bringing someone in for one project only. How do I restrict their access to my remote repo to only that one repo14:16
server is ubuntu 12.0414:16
_ikke_ C8H10N4O2: How do people connect to those repostories?14:16
← skul left14:16
C8H10N4O2 I just created users added them to the git group and they just use username@ipaddress:/path/to/repo14:17
cbreak I'd just use gitolite14:17
← kermit left14:17
cbreak otherwise you'll have to mess around with fs access rights14:17
→ dr_lepper joined14:17
_ikke_ !gitolite14:17
gitinfo Want to host as many git repos (and users!) as you like, on your own server, with fine-grained access control? You want gitolite: https://github.com/sitaramc/gitolite - Documentation: http://sitaramc.github.com/gitolite/master-toc.html14:17
C8H10N4O2 yeah but if i use gitolite or gitosis wont i have to re do everything14:17
cbreak so?14:18
milki C8H10N4O2: you'll be happier14:18
trust us14:18
_ikke_ C8H10N4O2: It allows for importing existing repositories14:18
cbreak it'll take less time than doing it with fs access rights :)14:18
← mgorbach left14:18
→ mgorbach joined14:19
→ RaceCondition joined14:21
← stodan left14:21
→ kermit joined14:22
← RaceCondition left14:23
C8H10N4O2 yuck I don't want to have to set something else up.14:23
cbreak then do it yourself14:23
with file system permissions14:23
git doesn't do permission checking14:24
git doesn't do authentication or authorisation14:24
_ikke_ no ACL in general14:24
C8H10N4O2 no no no ill just use gitolite or gitosis. In each repo you should be able to provide public keys14:24
→ hyperair joined14:24
cbreak with gitolite, public keys identify users14:25
and users can have permissions to read and/or change repositories14:25
the change part is extremely granular14:25
C8H10N4O2 is it quick to install and set up?14:25
cbreak yeah14:25
C8H10N4O2 ok fine ill use that14:25
→ pfFred joined14:25
C8H10N4O2 now I have to change all my hooks for deployment too then right?14:26
cbreak http://sitaramc.github.com/gitolite/install.html14:26
gitolite uses its own hooks14:26
for update14:26
the rest it doesn't care about14:26
but using an update hook for deployment would be stupid anyway14:26
you probably used post-receive or similar14:26
C8H10N4O2 yeah14:26
anyway Ill try it out14:27
→ chrisgeorge joined14:28
← pfFredd left14:28
→ Spockz joined14:29
→ Zenopus joined14:29
→ Psy-Q joined14:30
→ AaronMT joined14:30
← maleknet left14:31
← thiago left14:32
sam____ if i have made several commits on a branch (which I've pushed to our 'central' repo), can I squash those commits up before I merge the branch with the master, and if so how?14:32
wereHamster merge --squash?14:32
or rebase and squash, or soft reset and commit14:33
→ Geen_Nuar joined14:33
← fornext left14:33
sam____ ok, I'll investigate those options. thanks!14:33
← Vile left14:33
← Spockz left14:33
→ alessand1o joined14:34
comps is it possible (somehow via --no-commit ?) to merge 4 heads into one merge commit, using the recursive strategy, merging just two heads at a time (and resolving conflits)?14:34
cmn no14:35
octopus merges are only possible without conflicts14:35
wereHamster comps: do your merges, one after the other, then take the resulting tree, and commit that as a commit with 5 parents14:35
comps in theory, it's possible to merge first two heads, then merge the resulting tree with a third head, using the resulting tree for the already existing merge commit14:35
← zloy left14:36
wereHamster comps: this might be useful as a guide: https://blog.caurea.org/2009/11/19/subtree-octopus-merge.html14:36
alessand1o → Spockz14:36
wereHamster see the script how I generate a merge commit and manually specify the heads that should be part of it14:36
comps okay, thanks14:37
→ rednaks joined14:38
← rednaks left14:38
→ rednaks joined14:38
comps the easiest solution that comes to my mind is to stack one merge on top of another, then take sha of the final tree, reset to the before-merge state and recreate the one merge commit, manually linking parents and using the final tree from the "stacked" merge14:38
→ Davey joined14:39
← Geen_Nuar left14:39
→ Geen_Nuar joined14:40
← Geen_Nuar left14:40
→ RaceCondition joined14:40
← v0n left14:41
→ apok joined14:42
→ MyTallest joined14:43
← papegaaij left14:44
← rchavik left14:45
→ ShadeTornado joined14:46
← Sky[x] left14:46
← MyTallest left14:47
← ShadeTornado left14:47
→ Vile joined14:47
xonix Hi, how can I bring my forked git repository to the latest state?14:47
→ ShadeTornado joined14:47
_ikke_ xonix: define latest state14:48
→ MyTallest joined14:48
← Xethron left14:48
xonix _ikke_: i forked a repository and the repository from which I have forked has new commites14:49
← MyTallest left14:49
→ tjadc joined14:49
→ amcsi_work joined14:50
→ flavius joined14:50
_ikke_ xonix: You need to clone your repo locally, add the original repository as a remote, fetch from that remote, and push to your fork14:50
→ dc5ala joined14:50
← banghouse left14:50
→ nelse_ joined14:51
← dhruvasagar left14:51
→ MyTallest joined14:53
→ dhruvasagar joined14:54
← hwrd|work left14:54
→ Loggie joined14:54
← ThomasLocke left14:55
→ NcA joined14:57
xonix _ikke_: should i see some changes after fetching from the remote? i took a look with git log but there are now changes14:58
← NcA left14:58
xonix _ikke_: no changes :)14:58
→ NcA^ joined14:58
_ikke_ xonix: That's right, because they haven't been merged yet14:58
← dc5ala left14:58
→ goshawk joined15:00
← _ivar_ left15:00
→ hwrd|work joined15:01
← goshawk left15:01
SirCmpwn → SirCmpwn_15:01
← teotwaki left15:02
→ Beoran_ joined15:02
← Beoran__ left15:03
→ goshawk joined15:03
→ duckxx joined15:04
→ dpino joined15:04
→ teotwaki joined15:05
xonix _ikke_: thanks works fine :)15:06
← arvind_khadri left15:07
← Akufen left15:08
→ dpk joined15:08
← Davey left15:08
← juvenal left15:11
_ikke_ xonix: your welcome15:12
you15:12
you're15:12
→ harshpb joined15:13
→ tplaner_ joined15:13
← goshawk left15:14
→ juvenal joined15:14
Vampire0_ → Vampire015:15
← bapa left15:15
→ hakunin joined15:16
← svetlyak40wt left15:16
← juvenal left15:16
← replore left15:17
← senny left15:20
→ DancingBear joined15:21
→ fornext joined15:21
→ duckx joined15:22
← nelk left15:22
→ Harzilein joined15:23
Harzilein hi15:23
what is the proper utility to check availability of a git:// repository?15:24
→ adamben joined15:24
← wave left15:24
← sebrock left15:24
← duckxx left15:24
→ Evanlec joined15:25
Evanlec Hey guys, I think I may have lost a bunch of code that I stashed15:25
I stashed my changes in my exp branch15:26
checked out master15:26
and did git stash pop15:26
lots of merge conflicts15:26
so i did git reset hard15:26
is my stashed changes gone forever now?15:26
or can I still bring them back15:26
cmn have you checked?15:26
Evanlec how?15:26
← harshpb left15:27
← amcsi_work left15:27
→ elbeardmorez joined15:27
→ thiago joined15:27
cmn show, list, probably others, see man git stash15:27
gitinfo the git-stash manpage is available at http://git-scm.com/docs/git-stash.html15:27
Evanlec cmn: git stash list doesnt seem to show the stash that i saved15:28
← teotwaki left15:29
→ wave joined15:29
Evanlec i dont know if i named the stash or not when i saved it15:30
but i dont think i did15:30
← trungqtran left15:31
← blami_ left15:31
→ goshawk joined15:31
← Chillance left15:32
cmn if you don't give it an explicit name, it gets the one from the commit when you stashed it15:32
Evanlec ah15:32
→ swiety joined15:34
← Loggie left15:34
→ trungqtran joined15:34
← samphippen left15:34
← d3vic3 left15:34
Evanlec cmn: using git fsck ---unreachable i see a bunch of blobs and commits , could it be one of those?15:35
→ OOPMan joined15:36
Evanlec cmn: the git-stash manpage says to try using: git fsck --unreachable | grep commit | cut -d\ -f3 | xargs git log --merges --no-walk --grep=WIP15:36
→ zastern joined15:36
→ kevwil joined15:36
milki Evanlec: how are we supposed to know which one it is. look at them and see15:36
Evanlec but that command fails complaining about the cut command delimiter must be a single character15:36
→ beasty joined15:36
Evanlec milki: how can I look at them?15:36
← beasty left15:36
→ beasty joined15:36
beasty hi guys15:36
gitinfo beasty: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.15:36
milki they have references15:37
you can git show them15:37
← tjadc left15:37
beasty i pondering when i do a 'git pull' on a production server15:37
milki or a sha15:37
Evanlec ok so `git show <sha1>' ?15:37
milki Evanlec: !tias15:37
gitinfo Evanlec: Try it and see™. You learn much more by experimentation than by asking without having even tried. If in doubt, make backups before you experiment (see !backup). http://sitaramc.github.com/1-basic-usage/tias.html may help with git-specific TIAS.15:37
beasty it takes up to 47 seconds before it's finished15:37
is this normal ?15:37
← Gx4xx left15:37
milki beasty: depends on a variety of things15:38
← Ecesis1 left15:38
→ YoMomma joined15:38
beasty milki: like ?15:38
it was just 1 file15:38
milki git doesnt pull individual files15:38
→ samphippen joined15:38
← osxorgate left15:38
← drougge left15:38
→ Mode-M joined15:38
beasty ok15:38
milki git syncs hisotry15:38
beasty but the machine is a p4 with 32gb ram15:39
← Zenopus left15:39
← hwrd|work left15:39
→ raatiniemi joined15:39
→ Loggie joined15:40
cmn it depends on the amount of references, how complex the history is, how packed or unpacked the objects in the repository are...15:40
→ Anders_J joined15:40
← OOPMan left15:40
milki beasty: http://git-scm.com/about/small-and-fast contains some information about what affects performance15:40
← ajw0100 left15:40
← daniel_- left15:42
← irqq left15:42
→ OOPMan joined15:44
→ harshpb joined15:44
Evanlec cmn: okay so my working tree of my exp branch still seems as it was before the git stash....just that the files are not tracked15:45
→ ajw0100 joined15:45
cmn branches don't have working trees15:46
Evanlec cmn: hmm, okay,15:46
→ irqq joined15:46
→ pandeiro joined15:47
Evanlec cmn: but the files seem to be as they were before i did my stashing silliness15:47
→ indel joined15:47
cmn have you actually determined that the stash was lost?15:47
Evanlec cmn: it seems to be...there is 2 saved stashes,15:47
can I look at a stash particularly?15:48
← aisbaa left15:48
cmn yes, take some time to familiarise yourself with your tools15:48
→ turion joined15:48
← cakehero left15:48
Evanlec git show stash@{0}15:48
→ Zenopus joined15:48
Evanlec would that work?15:48
cmn does it?15:48
Evanlec it does15:49
sorry, i know, try it yourself15:49
i just dont want to lose anymore work!15:49
cmn the commands you say you did wouldn't get rid of any stashes15:49
thiago show can never destroy work15:49
← goshawk left15:50
Evanlec thiago: true15:50
cmn: git stash pop doesn't get rid of a stash?15:50
← Hasbro left15:50
cmn only if it applies cleanly15:50
Evanlec it didnt15:51
so i did a git reset HEAD --hard15:51
so the stash is still there?15:51
→ steveoliver joined15:51
← FrenkyNet|work left15:51
cbreak git stash list15:52
Evanlec its funny because now back on my exp branch where I saved the stash, if i stash pop it shows conflicts15:52
cbreak: i see two stashes with that...I think the latest might be the right one15:52
if i didnt name the stash when i saved, what would it be called?15:52
cbreak stash@{1} or so15:52
Evanlec well yes15:53
but it also says stash@{0}: On namespace: namespace refactor15:53
namespace being the branch iw as on15:53
and namespace refactor maybe is the stash name?15:53
→ john_anderson joined15:53
cbreak that's from the commit message you had checked out15:53
Evanlec its not a commit name15:53
thiago stashes don't have a commit message, so the command invents one for you15:53
it's the branch you were on and the commit you had checked o ut15:54
Evanlec git log doesnt show any commit named "namespace refactor"15:54
cbreak yeah, and?15:54
Evanlec if I did "git stash save namespace refactor" would that explain it?15:54
← kevwil left15:54
turion hi! I'm having a little convoluted problem: I worked on myrepo/dev_someissue and a friend cloned my repo as "hisrepo" and did a lot of new work. Now I want to have his branch hisrepo/dev_someissue as a new branch myrepo/dev_someissue_hiswork. Should I do something like "git fetch hisrepo/dev_someissue:dev_someissue_hiswork"?15:54
Evanlec cbreak: so where did it come up with that "namespace refactor" text associated with that stash?15:55
cbreak as I said: it takes the commit message of the commit you had checked out15:55
Evanlec cbreak: right, but i cant find any such commit message15:56
Tohuw turion: So he as performed work on top of your branch, and you want those changes as a separate branch?15:56
Evanlec cbreak: if i saved it with a specific name, it wold use that instead of the commit message right?15:56
cbreak git log stash@{0}15:56
turion Tohuw, yes15:56
cbreak you can pick one if you want15:56
but you said you didn't15:56
Evanlec cbreak: oh, there ti is15:57
→ psoo joined15:57
cbreak so it picked the message from your newest commit15:57
Evanlec cbreak: apparently it did15:57
cbreak: strange that regular git log doesnt show that15:57
cbreak regular git log is git log HEAD15:57
it only shows the current branch15:57
Evanlec oh15:57
cbreak NOT the one you happened to do git stash some time ago15:57
Evanlec i did the git stash today15:58
turion Tohuw, I'm also wondering how to set up the remote correctly. Do I have to do "git remote add -t dev_someissue hisrepo https://url"? Or is the -t switch not what I want?15:58
Tohuw turion: Furthermore, I assume you want to avoid having to do a local merge with your version of the branch (because otherwise you could simply merge in his changes, branch it, then on dev_someissue reset to an earlier commit)15:58
Evanlec the git log stash@{0} shows date oct 3 10:56 AM which sounds about right15:58
→ etcetera joined15:58
Evanlec of when i stashed15:58
turion Tohuw, yes I'd like to avoid that15:59
Tohuw turion: If you have the repo set remotely already, then you just need to push the branch.15:59
→ marshall joined15:59
marshall hey git15:59
gitinfo marshall: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.15:59
← samphippen left15:59
EugeneKay You're a git15:59
← logik-bomb left15:59
→ kevwil joined15:59
turion Tohuw: I haven't set his repo up yet16:00
→ juvenal joined16:00
Tohuw turion: Well, how about this: checkout dev_someissue at the commit before he made changes, then create a new branch from that level, then merge in his changes from his branch.16:00
← Dno left16:00
Tohuw turion: Why does he need an entirely separate repo?16:00
← psoo left16:01
→ DanyO83 joined16:01
turion Tohuw: Well, maybe he doesnt, but we started off a bit unorganised16:01
marshall i've updated the submodules in my dotfiles and pushed to github, but every time i pull from the origin, i get unstaged new commits for each submodule. what is going wrong?16:01
Tohuw turion: In general, a project should be one repo, and all collaborators work off that repo. There are obvious exceptions to that, but for a small project, it is usually painful to have complex multi-repo setups.16:02
→ mityaz joined16:02
← hyperair left16:03
Tohuw turion: May I suggest you give a read through this: http://nvie.com/posts/a-successful-git-branching-model/16:03
turion Tohuw: Ok, I'll work into that direction. So I should give him write-access and just let him push his branch under a different name?16:03
Tohuw turion: Yes, he could certainly do that. In general, when I collaborate on a repo, major changes are pushed as unique branches, because it's not too hard to merge things together when needed. It's not like you're paying by the branch. Branch away!16:04
→ bits8mybytes joined16:04
← bits8mybytes left16:04
→ bits8mybytes joined16:04
Tohuw IMO, the best habit you can develop when using Git is to "abuse branches".16:04
← notVert left16:04
→ aantix joined16:04
turion ok, that's a good advice16:05
Evanlec cbreak: thanks for your help, i popped the stash and it appears my work is restored16:05
cmn: thank you as well16:05
cbreak np16:05
Evanlec cbreak: i think in the future, I should commit changes on my feature/exp branch, instead of stashing16:06
cbreak marshall: chances are that you don't understand how submodules work16:06
turion Tohuw: but in that case he has to "git checkout dev_someissue", "git checkout -b "dev_someissue_mysolution" and then "git push dev_someissue_mysolution"16:06
← suy left16:06
→ AlbireoX joined16:06
Evanlec unless im making a very quick change on another branch16:06
turion and somehow resetting the position of his dev_someissue to my position16:06
← elbeardmorez left16:06
← Internet13 left16:07
cbreak marshall: if you have a submodule, then that submodule is a full self contained git repository16:08
the parent git repository decides on what to do with it, in particular which submodule commit it wants16:08
← jargon- left16:09
cbreak merging in the parent or the submodule doesn't affect the other16:09
← apok left16:09
thorbjornDX Hi, I have a branch with about a half dozen commits that I would like to put in another branch, and I would like to roll back the first branch to before those commits. What are my options? (none of these commits are public)16:09
Evanlec question: if git status says: changes not stashed for commit: deleted <some-file> , "use git add/rm <file> to update what will be committed", i should use git rm for deleted files right?16:09
→ SegFaultAX|work joined16:09
cbreak thorbjornDX: merge them into the other then reset them away from the first16:09
→ samphippen joined16:09
cbreak Evanlec: git add16:10
Evanlec or should I use git add, as in to "add" the deletion16:10
oh k16:10
thorbjornDX cbreak: would it work to just branch off of the main branch and then 'reset' that branch?16:10
cbreak you can use git rm, but add stages everything16:10
Evanlec what would git rm do?16:10
ok16:10
cbreak and if the file is already gone, you might have to use rm --cached if you opt to using rm16:10
← sobczyk left16:10
Evanlec ah right16:10
← flijten left16:10
→ imMute joined16:10
← imMute left16:10
→ imMute joined16:10
Evanlec i believe it is gone16:10
cbreak thorbjornDX: what?16:10
Evanlec so will i use git add or git rm --cached?16:11
cbreak your pick16:11
Evanlec doesnt matter?16:11
cbreak nope.16:11
Evanlec ok16:11
cool16:11
cbreak syntax is different. result is the same.16:11
← _huoxito left16:11
thorbjornDX cbreak: git checkout <branch w/ commits>; git branch <newbranch>; git reset HEAD^^^^^;16:12
cbreak sure.16:12
if the branch you want to move the commits into doesn't already exist16:12
Evanlec cbreak: oh git is so smart, i did git rm <file> and now it has it staged as a rename16:12
← jesseFromYVR left16:12
Evanlec cbreak: brecause really i just moved the file16:12
thorbjornDX cbreak: ah, yes. That's the case. Thanks :)16:12
→ Internet13 joined16:12
cbreak Evanlec: that's just rename detection. git blame is much cooler16:13
that thing can even detect if you move part of a file into an other file16:13
→ ReekenX joined16:13
← freeed left16:13
→ Hasbro joined16:13
cbreak thorbjornDX: also, you'd probably git reset --hard 123141216:13
--hard to also kill the working/index changes16:14
thorbjornDX cbreak: stashed already :)16:14
Evanlec ooh16:14
cbreak stashing is not the right thing to do with them16:14
Evanlec cbreak: i seem to recall rename detection not always being a feature in way older git versions16:15
thorbjornDX cbreak: well, I have unstaged changes that I don't want to deal with16:15
Evanlec i could be wrong16:15
cbreak Evanlec: it has always been a feature.16:15
not sure how good it was16:15
Evanlec ok16:15
cbreak or if it was implemented16:15
← ISF left16:15
← harshpb left16:15
Evanlec well I alwys try to do 'git mv'16:15
cbreak but it was accounted for :)16:15
← chaz left16:15
thorbjornDX cbreak: I'm adding another developer to this project, and I'd like to roll back a WIP branch that I had (and maybe use it later)16:15
Evanlec to avoid the added / deleted stuff16:15
← irqq left16:15
cbreak git mv is just a shorthand for mv && git add16:15
Evanlec Ahh16:15
cbreak: useful to know16:16
cbreak it does nothing special at all16:16
→ freeed joined16:16
cbreak rename detection is a feature of git diff/git log/git blame/...16:16
Evanlec and git rm is shortcut for rm ** rm --cached ?16:16
alexander__b cbreak: no it's short for mv && git rm [oldfile] && git add [newfile], right?16:16
Evanlec &&16:16
i mean16:16
→ jbrechtel joined16:16
cbreak alexander__b: same thing.16:16
→ alpjor joined16:16
cbreak alexander__b: if you move a file, git will tell you that you have a new untracked and an unstaged deletion16:17
you can use git add to resolve both16:17
or you can use git rm to resolve the unstaged deletion16:17
→ Vinnie_mac joined16:17
cbreak thorbjornDX: then staging is definitely NOT what you want16:18
stashing not either16:18
you'd ideally just make a new branch16:18
← friskd left16:19
thorbjornDX cbreak: I have two things I'm dealing with: 1) unstaged changes that I'd rather not deal with at this point (stashed), 2) un-pushed commits that I'd like to store in a non-public branch (branched + reset --hard)16:19
cbreak just commit the unstaged changes...16:20
put them away in some other branch16:20
stashes are way too easy to lose track off16:20
thorbjornDX cbreak: okay. Are they easy to lose just because of how they're pushed/popped?16:21
(and named)16:21
→ metcalfc joined16:21
cbreak mainly because people forget that they have them16:21
or where they are16:21
just make a branch...16:21
thorbjornDX cbreak: point taken. Making a branch now16:21
Evanlec cbreak: agreed about stashes16:21
after this scare16:21
thorbjornDX I mainly use stashes when I make changes in the wrong branch, and need to merge them with another one16:22
cbreak I don't use stashes for anything I want to keep for more than a few minutes16:22
→ Shusshu joined16:22
Evanlec thorbjornDX: yes thats hwat I do, but if the stash doesnt merge clearnly with the other branch, then you're in a mess16:22
← marshall left16:22
alexander__b cbreak: /msg aarolilja tippa d e siste16:22
err16:22
Evanlec like I am now16:22
thorbjornDX Evanlec: ah, I gotcha16:23
→ fly9 joined16:23
Evanlec i ended up doing git reset HEAD --hard16:23
but then didnt know where my stash went16:23
luckily it seemed to still be around16:23
← Shusshu left16:23
Evanlec in git stash list16:23
thorbjornDX 'git stash list' and 'git tag -l' always confuse me, haha16:23
cbreak stash pop only pops if the popping applies cleanly16:23
Evanlec but my tree i still messed up16:23
cbreak: yes, that is what saved me16:24
cbreak it won't mess up anything16:24
Evanlec it didnt apply cleanly so the stash stuck aroun16:24
cbreak it'll just apply like a cherry-pick16:24
and if it fails, you have to resolve merge conflicts16:24
← les_sylvains left16:24
Evanlec cbreak: well what i meant to do was stash pop, and then commit only one file16:24
that i wanted committed on master16:24
i didnt want to merge all the changes16:25
what would've been the correct course of action?16:25
cbreak stashes are everything16:25
Evanlec considering i made the change on the wrong branch, but wanted that file commit on master branch16:25
cbreak if you want to only part of a stash16:25
Evanlec yea16:25
cbreak then you'll have to manually work16:25
Evanlec hmm, well isnt there a better way thatn stash?16:25
cbreak just switch to the branch and commit16:26
Evanlec maybe just git co master, commit <blah>16:26
yea16:26
but it'd probably complain about overwriting working files or w/e still16:26
which I think is why I stashed in the first place16:26
← dpino left16:26
Evanlec so maybe I should've commited, switch to master, and merge that commit16:27
commited just that file I mean16:27
cbreak no16:27
← chaiz left16:27
Evanlec or make a new branch16:27
idk16:27
cbreak if you can't switch branches, then stash16:27
and at destination stash pop16:27
→ jargon- joined16:27
cbreak and then just add and commit the file you want16:27
Evanlec cbreak: but then i run into the issue i just had16:28
ok16:28
cbreak what issue?16:28
Evanlec but then what to do with the mess left over?16:28
cbreak no mess16:28
Evanlec when i stash popped i got all kindds of conflicts16:28
cbreak just resolve them16:28
Evanlec and things i didnt want16:28
i didnt want those16:28
i just wanted to commit that one file16:28
cbreak then git stash apply instead of pop16:28
and remove the rest with git checkout -p for example16:29
Evanlec or git reset hard?16:29
cbreak if you want to nuke everything16:29
Evanlec thts a dangerous command i know16:29
;p16:29
master was clean before the stash pop though16:29
→ FrenkyNet joined16:30
← Hasbro left16:30
→ xerxas joined16:30
Evanlec cbreak: so there's no way to just git add/merge/commit a file from one branch to another?16:30
EugeneKay You can fake it by doing a merge and then reset-ing all except the file(s) you wanted to merge16:31
Evanlec leaving the rest of the changes alone16:31
← LimeBlast left16:31
EugeneKay This is because git merges trees16:31
Evanlec EugeneKay: ah16:31
EugeneKay: seems kinda hacky16:31
EugeneKay That's because it is. The usual answer is "don't do that"16:31
Evanlec but I trust you know what you're tlaking about16:32
→ boombatower joined16:32
Evanlec its often the case though that I make a change on the wrong branch16:32
← klj613 left16:32
Evanlec and only want one change from that branch to be on a different brach16:32
not all th changes16:32
EugeneKay cherry-pick the commit?16:32
← afuentes left16:32
Evanlec well usually im on an exp branch so have no committed16:33
but thats anoter way16:33
EugeneKay !float16:33
gitinfo If you have made a change in your working directory and have NOT YET COMMITTED, you may "float" that change over to another (`git checkout oldbranch`) or new (`git checkout -b newbranch`) branch and commit it there. If the files you changed differ between branches, the checkout will fail. In that case, `git stash` then checkout, and `git stash apply` and go through normal conflict resolution.16:33
Evanlec exp branch doesnt matter if i make sily commits anyway16:33
→ ada1 joined16:33
→ kenperkins joined16:33
Evanlec EugeneKay: ah cool16:33
← Desproges left16:34
Evanlec EugeneKay: only thing is, git stash apply, i didnt want all the changes applied, just one file16:34
but we covered this16:34
EugeneKay git-add just the one file, make a commit, then cherry-pick it ;-)16:35
alexander__b is there a way to view diff with vim? i.e. not to open the affected files with vim, but to just view the diff file with vim.16:35
→ jonshea joined16:35
← ReekenX left16:35
Evanlec EugeneKay: yea, i think thats the best way16:35
turion Tohuw: Thanks a lot, it worked very well!16:35
EugeneKay git-stash does essentially this, but it doesn't use a normal branch - it stores it under refs/stashes/(I think... I don't use it because stashes are evil)16:35
Evanlec EugeneKay: it was made evne more complicated hwoever because i also moved the file, and mad changes to it16:35
→ skyf joined16:35
→ cheriv joined16:35
→ sorin joined16:36
← thierryp left16:36
EugeneKay You poor git16:36
Evanlec poor git16:37
← turion left16:37
← C8H10N4O2 left16:38
Evanlec git doesnt like the way i work16:38
but i like the way git works16:38
mostly16:38
if that makes any sense16:38
EugeneKay Have you read !concepts ?16:38
gitinfo "Git Concepts Simplified" explains the basic structures used by git, which is very helpful for understanding its concepts. http://sitaramc.github.com/gcs/16:38
Evanlec Ah, no i havent16:38
EugeneKay It helps16:39
Evanlec EugeneKay: I'll add that to our teams developer wiki16:39
we're all in process of migrating from svn to git16:39
→ Sky[x] joined16:39
EugeneKay You poor gits16:39
Evanlec im actually far ahead of the other guys with git16:39
It's scary to think that because I still get completely lost like earlier today16:39
scary to think of them trying to use git for anything non-trivial16:40
EugeneKay It bugs me jsut how user-unfriendly git can be16:40
Evanlec yes, myself included16:40
EugeneKay And amazes me just how well I've learned it >_>16:40
Evanlec I tried mercurial briefly because it seems to be more friendly, but16:40
EugeneKay The problem with Hg is that Linus didn't write it16:41
← sgambino left16:41
Evanlec then I was appauled (sp?) that it did not have lightweight branching16:41
→ sgambino_ joined16:41
EugeneKay appalled16:41
Evanlec i had to copy my entire tree to make a new branch16:41
EugeneKay Unless you're referring to PaulePanter16:41
Evanlec which was totally unacceptable16:41
one of my favorite features of git is lightweight branching, one directory for EVERYTHING16:41
← giallu left16:41
Evanlec i hate cluttering my fs16:41
← raatiniemi left16:42
EugeneKay Yeah, that comes from how branches exist in git16:42
← microcode left16:42
Evanlec they're awesome16:42
EugeneKay It's just a reference to a commit SHA16:42
Evanlec but they confuse the shit out of svn users16:42
EugeneKay There's some stuff on top of that like the !reflog16:42
gitinfo The git reflog (`git log -g`) records the SHAs of your HEADs for 2+ weeks. `git checkout -b myrestore OLDSHA` and `git reset --hard OLDSHA` will relink to that state via a new and current branch respectively, see http://sethrobertson.github.com/GitFixUm/ for full details. WARNING: reset --hard will trash any uncommitted changes! Visualize with: gitk --all --date-order `git log -g --pretty=%H`16:42
EugeneKay Make them !unlearn ;-)16:42
gitinfo Users of centralized VCSes will need to unlearn a number of habits in order to learn git. Many commands, such as "checkout" and "revert" bear little in common to those holding the same name in other systems.16:42
Evanlec lol16:42
true16:43
the reflog i still havent explored much yet16:43
one issue we're facing with migration, and maybe you have a good solution16:43
EugeneKay I've not used it in...... ages.16:43
← MyTallest left16:43
Evanlec is with svn we're using svn:externals, which allows us to do a partial checkout from our shared lib repository16:43
EugeneKay It depends how your SVN repo(s) is(are) set up.16:43
Evanlec into each application16:43
with git, you cannot do a partial checkout16:44
EugeneKay Sure you can. Man git-archive16:44
Evanlec okay16:44
EugeneKay man git-archive16:44
gitinfo the git-archive manpage is available at http://git-scm.com/docs/git-archive.html16:44
EugeneKay Stupid bot16:44
esc you can also clone with --depth, right16:44
Evanlec there was another article i saw called git subtree16:44
EugeneKay You can, but don't16:44
!subrepos16:44
gitinfo [!subprojects] So, you want to add git repositories inside of other git repositories? Well, you have four main options. First is to just do it, add the repo to the outer project's .gitignore, and treat them entirely separately. Best if they are entirely seperate. Otherwise your best options are "!submodule" "!gitslave" and "!subtree" Try typing those commands into this IRC channel.16:44
→ MyTallest joined16:45
EugeneKay It depends upon how you're !deploy ing16:45
gitinfo Git is not a deployment tool. You can build one around it for simple environments. http://sitaramc.github.com/the-list-and-irc/deploy.html16:45
Evanlec !subtree16:45
gitinfo The subtree merge method is great for incorporating a subsidiary git repo into your current one with "unified" history. Read http://git-scm.com/book/en/Git-Tools-Subtree-Merging for more info, or try one of the !subtree_alternatives16:45
→ Sigma joined16:45
EugeneKay Submodules are pretty common for purely library subrepos16:45
But they have..... issues.16:45
Evanlec EugeneKay: but often we dont want the entire library subrepo16:45
its not that its huge but16:45
we'd have to re-arrange paths and such16:45
EugeneKay Just a subdir of the library?16:45
Evanlec how?16:46
I mean16:46
EugeneKay I mean, what portion of the lib do you want?16:46
Is it just library/foo/ ?16:46
Evanlec well for example16:46
yes16:46
moben couldn't you split it up?16:46
Evanlec library/foo/16:46
i only want foo16:46
in my main repo16:46
EugeneKay Then split up your library into library-foo/, library-bar/, etc16:46
And submodule that16:46
Evanlec mmm...that sounds a bit ugly16:46
i mean16:46
there's 8 directories in the lib repo16:47
EugeneKay That's only 8 repos16:47
Evanlec ...16:47
imMute Git: I hope you have a big hard drive. sucks, but that's how git is (currently)16:47
Evanlec EugeneKay: wouldnt subtree or something be better solution?16:47
EugeneKay I have one project here with 37 submodules16:47
Evanlec EugeneKay: we have the library repo in git now btw16:47
EugeneKay subtree is for including16:47
← berndj left16:48
EugeneKay I think you could make it do that, but I dunno16:48
moben Evanlec: subtree is what you then use to get the split up libs into your repo16:48
Evanlec yea16:48
thats kinda what I thought16:48
EugeneKay If your lib is really 8 different logical parts then each part should be its own repo16:48
Evanlec like we just want libs/Session/LDAP/ in Myrepo/LDAP16:48
EugeneKay: okay fair enough but16:48
← jbrechtel left16:49
Evanlec EugeneKay: i dont know if thats the case16:49
ive not looked at it thoroughly16:49
EugeneKay Talk to your project manager ;-)16:49
Evanlec he quit16:49
i mean16:49
→ microcode joined16:49
EugeneKay HA16:49
Evanlec the guy who wrote the lib is long gone16:49
:\16:49
heh16:49
but yea moben i think we'll try git subtree16:50
for including bits of the shared-lib repo16:50
EugeneKay You've just given me a feature idea for my git-library project16:50
Evanlec !subtree16:50
gitinfo The subtree merge method is great for incorporating a subsidiary git repo into your current one with "unified" history. Read http://git-scm.com/book/en/Git-Tools-Subtree-Merging for more info, or try one of the !subtree_alternatives16:50
EugeneKay Yay feature creep16:50
Evanlec !subtree_alternatives16:50
gitinfo [!subtree_alt] The git subtree merge method is hard to export changes from.. https://github.com/apenwarr/git-subtree provides another method which appears to be easier to export changes from. Also as a no-change-exported method, see https://metacpan.org/module/git-stitch-repo which claims to generate a unified history instead of merged branches.16:50
← kevwil left16:51
Evanlec hmm, hard to export changes from16:51
sigmonsays_ Is there an easy command to see all changes from 2 weeks ago for each branch?16:51
EugeneKay Meaning for contributing back to your library from MyProject16:51
Evanlec I dont think we'll ever be modifying the subrepo files16:51
EugeneKay sigmonsays_ - !treeish16:51
gitinfo sigmonsays_: A tree-ish is something that looks like a tree. Read 'man gitrevisions' and http://git-scm.com/book/en/Git-Tools-Revision-Selection16:51
Evanlec EugeneKay: yea, i dont think we'll ever be doing that16:51
→ sq-one joined16:51
Evanlec EugeneKay: if anything we'd modify the library repo itself16:51
EugeneKay Evanlec - then no worries16:52
Evanlec since the repo is shared across many projects16:52
→ cakehero joined16:52
→ friskd joined16:52
→ hwrd|work joined16:52
Evanlec and thus making changes must be done very carefully16:52
EugeneKay: coolness, Thank you so much for your advice16:52
← friskd left16:52
EugeneKay Sure.16:52
Evanlec moben: ty as well16:52
← funnyfingers left16:53
→ funnyfingers joined16:54
→ jbrechtel joined16:54
← duckx left16:54
→ alien_talk joined16:56
← Vinnie_mac left16:57
alien_talk hey i'm using cent-os and when i'm trying to do a "git daemon..." it says git: daemon is not a git command16:57
any clues?16:57
→ ReekenX joined16:57
← SegFaultAX|work left16:57
cmn install it16:58
EugeneKay alien_talk - git-daemon is not installed in the git-core package.16:58
alien_talk - what are you trying to do, though?16:58
→ Hasbro joined16:58
→ apok joined16:58
← msmithng left16:59
alien_talk was just tyring to share a repo on a remote server16:59
thanks EugeneKay and cmn didn't know it wasn't part of the core17:00
EugeneKay Do it via ssh17:00
→ jesseFromYVR joined17:01
alien_talk was juts following this tutorial:http://www.davegardner.me.uk/blog/2010/01/29/setting-up-git-on-centos-5-server/17:02
← jbrechtel left17:03
← LeMike1 left17:03
EugeneKay !blog17:03
gitinfo Blog posts, while helpful and informative, are quite often outdated, give bad advice, or are just plain wrong. Please don't rely solely upon them, or treat them as authoritative.17:03
→ LeMike joined17:04
EugeneKay And that is a terrible post - it recommends using git-daemon, which uses git://. NOT what you want for security. !ssh17:04
gitinfo Please use SSH to talk to remote repos. http:// and git:// are completely unsecured. You can try to do "smart" https://, but it is troublesome to configure(web server/cgi authentication mechanisms, ACLs, etc) and use(client CA certificates, firewalling, protocol inefficiency)17:04
← mtrd`w left17:04
← angelsl left17:04
EugeneKay !gitolite would serve a lot better for 2+ people, or if it's just you then you cna use the simple user@host:/path/to/repo.git/ syntax for git-remote17:05
gitinfo Want to host as many git repos (and users!) as you like, on your own server, with fine-grained access control? You want gitolite: https://github.com/sitaramc/gitolite - Documentation: http://sitaramc.github.com/gitolite/master-toc.html17:05
EugeneKay Which runs over SSH17:05
alien_talk thanks17:06
EugeneKay Sure17:06
← glennpratt left17:06
→ glennpratt joined17:07
→ eijk joined17:07
→ kukks joined17:08
← kukks left17:08
→ kukks joined17:08
→ psoo joined17:08
← psoo left17:08
← apok left17:08
→ apok joined17:09
→ SegFaultAX|work joined17:10
← Hasbro left17:10
← shruggar left17:10
→ jaequery joined17:10
→ fletch_ joined17:12
← adamben left17:12
← einonm left17:14
→ goshawk joined17:14
→ giallu joined17:14
→ harshpb joined17:15
→ duckxx joined17:16
→ dustingetz joined17:16
← fornext left17:17
→ ncr100 joined17:17
→ armenb joined17:17
→ Raging_Hog joined17:17
← swiety left17:17
← berserkr left17:18
→ marshall joined17:18
← Vile left17:19
→ duckx joined17:19
→ SirCmpwn joined17:19
← PigeonFriend left17:19
→ kumbayo joined17:20
← duckxx left17:22
← awilkins left17:22
← etcetera left17:22
armenb I have a git-svn-related question. do any organizations use a single git-svn checkout to front for multiple people who want to use git amongst themselves?17:23
← Anders_J left17:24
← harshpb left17:27
← sorin left17:27
→ harshpb joined17:28
← tango_ left17:28
→ jez joined17:28
gitinfo set mode: +v17:28
jez guys, I sent a message to the git mailing list about fixing the git p4merge script and got no useful replies. Any ideas how I can get someone to care? :-)17:29
→ etcetera joined17:29
← duckx left17:30
→ shtrb joined17:30
→ tango_ joined17:30
→ martinjlowm joined17:30
← iguanna left17:30
milki more examples, motivation, test cases, and maybe an approach to a patch solution?17:31
← trungqtran left17:31
jez i've basically shown in my post what a patch would look like17:31
→ iguanna joined17:32
→ iguanna_ joined17:33
← shtrb left17:34
→ shtrb joined17:34
→ trungqtran joined17:34
shtrb I just learn the hard way that apperntly git reset --hard does not restore all info : I have some files (~20K) that are shared via a dropbox account and all files are backed up with git localy. I have a space problem so when the dropbox account was off I deleated the files then do a git --hard reset and tried to resync via the dropbox app, apperntly it had seen the delete and add operation so I just delated ~4Gb of data and started to upload that.17:34
→ les_sylvains joined17:35
← marthinal left17:36
← iguanna left17:36
iguanna_ → iguanna17:36
← mvensky left17:36
armenb is there a way to push commits between two bare repositories?17:37
kumbayo Is there a way to tell git diff certain lines are "boring" like whitespace, }, END_IF, end sub, ... Lines that contain no useful information, that should be just added to a hunk instead of staying there as context between 2 hunks17:37
← babilen left17:37
→ marthinal joined17:37
armenb e.g. if there are users on two sides of a firewall17:37
→ sorin joined17:38
← Synthead left17:38
milki kumbayo: man diff or provide your own diff program17:39
→ Synthead joined17:39
→ bc joined17:39
← harshpb left17:39
→ nelk joined17:39
milki armenb: make a hole in the wall17:39
→ harshpb joined17:40
armenb milki: thanks, but that's not an option.17:40
bc To revert only some changes between head and a particular commit, I would use `git reset -p a3f44c' right?17:40
milki the do it offline with physical media and bundles/patches17:41
bremner armenb: what do bare repos have to do with firewalls?17:41
milki bc: !fixup17:41
gitinfo bc: So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!17:41
→ jdunck joined17:41
armenb bremner: there are two sets of users, and they cannot access each others' networks. however, the repositories lie on networks that can access each other.17:42
imMute that doesn't make sense17:42
EugeneKay armenb - sure, a bare repo can have a remote the same as a normal user17:42
You can't do any work-tree operations(such as merge) however17:42
bremner armenb: can you you use hooks to mirror (essentially what EugeneKay suggests)17:42
armenb a simple solution to this seems to be to create a non-bare repo somewhere between the two bare repos and use it to just shuttle commit objects back and forth, even through a cron job17:43
I want to avoid that step altogether and push between the bare repos. is this possible?17:43
→ friskd joined17:43
imMute armenb: you can push between bare repos, just like regular repos17:43
armenb hmm, ok17:44
bc milki: the problem is that I want to revert only some specific portions of code. The GitFixUm link would have me do this: git reset --hard @{u}17:44
← Evanlec left17:44
← dhruvasagar left17:45
milki reset will not do portions of commits17:45
youll probably need to do this manually17:45
git diff rev > patch17:46
edit it and apply17:46
→ friskd_ joined17:47
← friskd left17:47
friskd_ → friskd17:47
bc milki: I see, thank you17:47
← LekeFly left17:48
← rvsjoen left17:48
→ C8H10N4O2 joined17:49
← harshpb left17:49
→ ncd|leeN joined17:49
← delt0r left17:50
→ rubious joined17:50
→ arturaz joined17:50
→ irqq joined17:50
← Tommy[D] left17:51
→ btree joined17:51
→ rvsjoen joined17:51
← rvsjoen left17:51
→ rvsjoen joined17:51
← indel left17:51
→ btanaka joined17:52
→ Tommy[D] joined17:52
← leeN left17:53
← mmc1 left17:54
→ pantsman joined17:55
→ duckxx joined17:57
→ Goplat joined17:59
→ carpeliam joined17:59
→ jbrechtel joined17:59
→ jalama joined17:59
carpeliam hey all, how do i find which commit a particular branch split off from another branch?17:59
milki carpeliam: man git-merge-base18:00
hmmm18:00
carpeliam: man git-mergebase18:00
hm....18:00
well itswell18:01
→ mvensky joined18:01
milki its merge-base -.-18:01
gitinfo carpeliam: the git-merge-base manpage is available at http://git-scm.com/docs/git-merge-base.html18:01
→ hyperair joined18:02
← mvensky left18:02
carpeliam milki: thanks. what if I'm not sure what branch a particular branch was split from?18:03
← erichynds left18:03
← btanaka left18:03
milki hmmm, visually with !lg maybe18:04
gitinfo git config --global 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"18:04
→ kpreid joined18:05
← nelse_ left18:05
← SegFaultAX|work left18:06
← blorbx left18:06
→ flijten joined18:07
→ TheFuzzball joined18:07
← kpreid_ left18:07
← jbrechtel left18:08
← marshall left18:09
← jalama left18:11
← codertux left18:12
→ taha joined18:13
← ada1 left18:13
← taha left18:14
← friskd left18:15
← tinti left18:15
→ friskd joined18:15
← Element9 left18:15
← friskd left18:16
→ frooh joined18:16
→ friskd joined18:16
← friskd left18:16
frooh what's the best way to list the files in a given directory in a given branch?18:16
→ babilen joined18:16
frooh I thought ls-files would work but I was mistaken18:16
milki was about to say that -.-18:16
hmmm18:16
→ tinti joined18:17
milki whats wrong with ls-files?18:17
frooh milki: I think ls-files is just the checkout18:17
no where does it take a branch18:17
← tinti left18:17
→ psoo joined18:18
frooh I think I may have used ls-tree in the past but it was pretty involved18:18
→ PapaSierra2 joined18:18
← LeMike left18:18
→ maletor joined18:18
milki hm18:18
ls-tee with --name-only18:18
→ tinti joined18:19
→ codertux joined18:19
milki er18:19
tree18:19
frooh milki: yeah, that seems to work18:19
git ls-tree $branch $dir/ --name-only18:19
milki++ # thanks18:19
milki oooo18:20
→ kerrick joined18:20
← Mode-M left18:23
→ iwoj joined18:23
gitinfo set mode: +v18:23
iwoj I moved a bunch of files manually instead of using "git mv"18:24
It's too complex for me to move them back now18:24
is there a way to reattach the files to their git histories?18:24
imMute iwoj: git rm the old names and git add the new ones.18:25
iwoj (instead of causing a mass of deletes and adds?)18:25
imMute iwoj: git mv is an "alias" for git rm && git add18:25
milki iwoj wants rename detection18:25
iwoj yeah18:25
move detection18:25
→ rtjure joined18:26
imMute the short of it is "you dont". git doesn't actually track renames, it just makes it look like it does when it shows history output18:26
milki nods. its misleading18:26
→ reshtnk7 joined18:27
→ Targen joined18:27
→ talexb joined18:27
← pantsman left18:28
→ Eridius joined18:28
→ pantsman joined18:28
← pantsman left18:28
→ pantsman joined18:28
→ tomasm- joined18:28
shtrb git reset ?18:28
← bartek left18:29
frooh I want to see a list of files that were created between a given branch and the current one, the obvious solution is git diff ^from to --name-only, but that includes modified (not just created)18:29
any tips?18:29
talexb is there some obvious way for me to find which commit deleted a directory, if I know which commit the directory existed in, and which commit they were gone?18:29
frooh, I usually use a GUI like gitk for that.18:30
← eMBee left18:30
frooh I know how to do it by hand, this is for a script.18:30
→ eMBee joined18:30
→ kvothe joined18:31
← kvothe left18:31
→ kvothe joined18:31
← i42n left18:31
talexb Hmm .. Looks like I'm going to try bisect on this.18:31
frooh this is ghetto, but it works:18:31
→ fisted joined18:32
frooh git diff ^from to --stat | grep -v '-'18:32
it would mistakenly skip binaries, but I don't care about them right now18:32
← Guest24033 left18:32
→ adhawkins joined18:32
adhawkins → Guest8585518:33
milki frooh: git log with --stat18:33
← kerrick left18:33
← Guest85855 left18:33
frooh milki: ... and that would be better than diff because?18:33
← tomasm- left18:33
→ adhawkins2 joined18:34
milki it tells you modify/delete/create with symbols i think18:34
← fisted_ left18:34
milki o18:34
you have --stat18:34
nvm18:34
→ variable joined18:35
← adhawkins2 left18:36
→ adhawkins joined18:36
← kvothe left18:37
→ uirono joined18:38
← aristidesfl left18:38
← thirdknife left18:38
milki o hey18:38
--name-status is the one18:38
with A, M, D18:39
← leo2007 left18:39
frooh aw awesome18:39
milki: thanks!18:39
milki :P18:40
→ mmc1 joined18:40
chrisgeorge Is there a way to show a range of commits for a specific file, between two dates?18:42
--after and --before don't work with a single file.18:42
→ bartek joined18:43
milki it doesnt?18:43
← etcetera left18:43
→ anabelle joined18:43
frooh milki: if interested this works perfectly: git diff HEAD ^origin/release --name-status -- var | grep -e '^A' | cut -b3-18:43
chrisgeorge it certainly isn't showing results whereas git log myfile does :\18:44
← eean left18:44
chrisgeorge (after that certain date)18:44
→ kerrick joined18:45
← jesseFromYVR left18:45
→ rohan32 joined18:46
← Eridius left18:46
← ReekenX left18:46
← steveoliver left18:47
← sontek left18:47
← iguanna left18:47
milki worksforme18:47
→ jesseFromYVR joined18:48
→ iguanna joined18:48
→ Eridius joined18:49
→ etcetera joined18:49
← hellopat left18:50
cmn check what commits git's actually looking at18:50
→ hellopat joined18:50
→ dcs_ joined18:52
→ kvothe joined18:52
← kvothe left18:52
→ kvothe joined18:52
← dcs_ left18:53
→ Vinnie_mac joined18:53
→ _iron joined18:58
→ daniel_- joined18:58
→ jperry2__ joined19:00
← jperry2__ left19:00
→ Dno joined19:00
← kerrick left19:01
← hwrd|work left19:01
→ jperry2__ joined19:03
← jperry2_ left19:04
← PapaSierra2 left19:04
→ hwrd|work joined19:04
→ PapaSierra2 joined19:04
→ jbrechtel joined19:05
← shtrb left19:05
← pfFred left19:05
← und3f left19:06
→ artnez joined19:07
← arturaz left19:07
→ shtrb joined19:07
← carpeliam left19:07
← JSharpe left19:08
← shtrb left19:09
→ JSharpe joined19:09
← derpops left19:09
← jbrechtel left19:09
→ v0n joined19:09
← iwoj left19:10
← alunduil_ left19:11
← alunduil left19:11
← DanyO83 left19:12
← Vinnie_mac left19:12
← marthinal left19:13
→ aspotashev joined19:13
→ eletuchy joined19:13
← kvothe left19:14
→ dogarrhea2 joined19:14
← dogarrhea2 left19:14
← anabelle left19:15
← _Vi left19:15
→ und3f joined19:15
→ dogarrhea2 joined19:16
→ jbrechtel joined19:16
← dogarrhea2 left19:16
→ dogarrhea joined19:16
← mityaz left19:18
→ wilsonj joined19:19
← Lgb left19:20
← maletor left19:20
→ Lgb joined19:21
← dogarrhea left19:22
→ Feh joined19:22
esc sitaram: any thoughts on scaling gitolite to 10,000 users?19:22
_ikke_ esc: Good luck handling that many ssh sessions ;)19:24
→ ISF joined19:24
PerlJam esc: Seems like the config file parsing would be your only slowdown (assuming you're mentioning all 10,000 users in it)19:25
→ dogarrhea joined19:25
dogarrhea what are the difference between gitlab and github19:25
→ dpino joined19:25
milki dogarrhea: gitlab is free!19:26
→ avjpl joined19:26
dogarrhea in regards to workflow/processes19:26
_ikke_ dogarrhea: You can't install github yourself19:26
← retro|cz left19:26
milki dogarrhea: none19:26
← avjpl left19:26
milki not that i know the features of gitlab...19:26
but it sorta looks the same >.>19:26
dogarrhea we were on github for a while. have this "upstream" stuff with a master repo and your "forked clone" which you push to. is this much different from gitlab?19:26
← OOPMan left19:26
→ sluther joined19:26
_ikke_ dogarrhea: No, that stays basically the same19:27
dogarrhea we were also having trouble with code review integration19:27
gitinfo set mode: +v19:27
wilsonj I'm having trouble getting gitolite mirroring to work properly. I have a master server named git_dev with all the repos, and a slave server named backup_dev. I also have gitlab implemented as well. It seems everything was configured properly, and when I push a change to git_dev, i do get this ====> remote: (26590&) git_dev ==== (gitolite-admin) ===> backup_dev, but nothing is being replicated on the backup_dev (most importantly the gitolite.conf from the19:27
gitolite-admin repo)19:27
← hwrd|work left19:28
→ SegFaultAX|work joined19:28
← JSharpe left19:29
→ JSharpe joined19:29
→ hwrd|work joined19:30
→ Industrial joined19:32
← jbrechtel left19:33
→ maletor joined19:34
← Milossh left19:35
Industrial I'm running a VPS and I'd like to use git as part of the deployment of some NodeJS webservers, but I'm not sure how to configure git and/or how the workflow works. What i'd like to do is have a central git repo/server that I can push to from home and also fetch and deploy tags from.19:35
wereHamster Industrial: !deploy19:36
gitinfo Industrial: Git is not a deployment tool. You can build one around it for simple environments. http://sitaramc.github.com/the-list-and-irc/deploy.html19:36
wereHamster and welcome to #git :)19:36
Industrial I know how to set up a git repo and I've used github before, so I know how the setting a remote and pushing works, but I haven't done the server side.19:36
hi ;)19:36
_ikke_ Industrial: Use something like capistrano19:36
gitinfo set mode: +v19:36
_ikke_ Industrial: A real deployment tool19:36
foraks anyone fancy some psychic debugging? http://pastebin.com/rv6akn1r19:36
_ikke_ Industrial: It's not very hard to learn19:36
foraks: echo $GIT_WORK_TREE19:37
foraks it's empty19:37
_ikke_ foraks: What is everyting else?19:37
← JSharpe left19:37
foraks all of git19:37
committing, merging, rebasing19:37
→ JSharpe joined19:37
foraks diffing with `git diff` works19:37
everything else works as it should except those commands (oh and add -i also doesn't work)19:37
← PapaSierra2 left19:38
_ikke_ foraks: git add <file> does work?19:38
foraks though I'd hazard a guess that the other commands taking --patch don't work19:38
_ikke_: yep19:38
← Dno left19:39
_ikke_ foraks: I'd take it to the mailing list19:39
foraks is GIT_WORK_TREE meant to have something?19:39
_ikke_ foraks: nope19:39
foraks GIT_DIR is also empty19:39
okay19:39
← NcA^ left19:39
_ikke_ git config --list --global?19:40
foraks: [email@hidden.address]19:40
← iguanna left19:40
→ OOPMan joined19:41
→ NcA^ joined19:41
foraks _ikke_: at the bottom http://pastebin.com/rv6akn1r19:41
_ikke_ foraks: What happens if you explicitly pass in --work-tree?19:41
→ svetlyak40wt joined19:42
→ \bMike\b joined19:42
_ikke_ git --work-tree path/to/repo19:42
and then the command19:42
← maxandersen1 left19:42
\bMike\b Is there a way to have `git add -p` skip whitespace changes like `git diff -b` does?19:42
→ maxandersen joined19:42
← maxandersen left19:42
→ maxandersen joined19:42
_ikke_ foraks: Your on windows19:42
?19:42
foraks yeah19:42
→ maxandersen1 joined19:43
← maxandersen left19:43
_ikke_ msysgit or cygwin?19:43
← longbeach left19:43
cmn you don't seem to have any repo-specific configuration19:43
→ derpops joined19:43
_ikke_ cmn: because of the --global19:43
foraks msysgit19:43
← JSharpe left19:43
cmn oh, nevermind me then19:43
foraks `git --work-tree . difftool --cached` worked19:43
→ JSharpe joined19:43
foraks oh wait19:43
← JSharpe left19:43
foraks ignore that19:43
"git difftool --cached" seems to work as well19:44
_ikke_ but not without the --cached?19:44
foraks seemingly19:44
just double checking19:44
→ TeckniX_ joined19:44
_ikke_ because --cached doesn't need the work-tree obviously19:44
foraks good point19:44
← TeckniX left19:45
TeckniX_ → TeckniX19:45
_ikke_ foraks: Have you tried reinstalling msysgit?19:45
foraks no, but I probably should.19:45
→ jceb joined19:45
foraks something's messed up somewhere anyway19:46
_ikke_ yeah19:46
And my knowledge of git internals is too limited to be of effective help with that19:46
← maxandersen1 left19:46
foraks thanks for the help. good to know i'm not missing something obvious19:47
cmn if rev-parse dosen't think it's a repo, reinstalling shouldn't help19:47
did the computer/disc crash?19:47
foraks nope19:47
cmn what does .git/HEAD have?19:47
foraks ref: refs/heads/master19:47
_ikke_ nothing wrong with that it seems19:48
→ macrover joined19:48
← jesseFromYVR left19:48
→ marhaban joined19:48
→ Moussekateer joined19:48
talexb Wow -- git bisect rocks. Love it.19:48
cmn and .git/ has objects/ and refs/ inside it, presumably?19:49
← flijten left19:49
foraks yep19:49
this happens with all repos, so I doubt the repo itself has anything wrong with it19:49
cmn oh, that's really odd19:49
foraks I can try freshly cloning something19:49
sec19:49
← psoo left19:49
foraks yeah just cloned something19:50
same thing19:50
$ git add -i19:50
fatal: Not a git repository: '.git'19:50
rev-parse --show-prefix: command returned error: 12819:50
$ ls .git19:51
HEAD config hooks info objects refs19:51
branches description index logs packed-refs19:51
_ikke_ seems something messed up with your git install19:51
← aspotashev left19:51
← armenb left19:51
→ aristidesfl joined19:51
foraks yup19:52
← patie1 left19:52
foraks are there any env variables git expects to be set19:52
_ikke_ foraks: nope19:52
foraks hm19:52
_ikke_ export | grep GIT doesn't return anything for me19:53
those env variables are for when for example working with bare repositories19:53
or for scripts19:53
thorbjornDX where's a good tutorial for someone who has no experience with rev control?19:53
_ikke_ !book has several19:54
gitinfo There are several good books available about git; 'Pro Git' is probably the best: http://git-scm.com/book but also look at !bottomup !cs !gcs !designers !gitt !vcbe and !parable19:54
→ patie joined19:54
_ikke_ !parable is a good one explaining the reasoning behind git19:54
gitinfo 'The git parable' provides some good reasoning behind git. http://tom.preston-werner.com/2009/05/19/the-git-parable.html19:54
foraks _ikke_: okay, thanks. i gotta go, but i'll try reinstall it later. if you think of anything else feel free msg me or highlight me in here19:55
_ikke_ foraks: right19:55
foraks: Otherwise you might want to hit the msysgit mailing list19:55
thorbjornDX _ikke_: thank you :)19:55
← guampa left19:56
← asteve left19:56
_ikke_ thorbjornDX: You're welcome. If you have any questions, feel free to ask19:56
← TeckniX left19:56
← john_anderson left19:57
→ cagedwisdom joined19:57
→ arietis joined19:57
thorbjornDX _ikke_: I personally have a reasonable amount of experience w/ git, but I am attempting to train someone who doesn't come from a software background, and who hasn't used anything in the past19:58
_ikke_ thorbjornDX: Ah ok19:58
thorbjornDX _ikke_: so I'm attempting to modify my vocabulary to make the transition less painful19:58
_ikke_ thorbjornDX: then parrable is very good. It explains version control from the very beginning19:58
→ asteve joined19:59
_ikke_ ie, it doesn't start with git, but it builts the concepts git uses up from the beginning19:59
→ srayner joined19:59
thorbjornDX _ikke_: also, I have his machine set up with the windows git gui, is this the best place to start? or should I try to get him working with git bash?19:59
← asteve left19:59
→ pfFredd joined20:00
→ guampa joined20:00
_ikke_ thorbjornDX: Well, I don't think that person is very comfortable with the command line yet20:00
thorbjornDX: Also look at GitExtensions20:00
← skyf left20:01
→ thillux joined20:01
thorbjornDX _ikke_: what does gitextensions provide over the default git gui? (from a beginner's perspective, mostly)20:02
_ikke_ thorbjornDX: Not sure if it has extra features or something20:02
thorbjornDX: Has been a while since I used it20:03
← cakehero left20:03
_ikke_ thorbjornDX: But I have a vague recalling finding it better than git gui20:03
thorbjornDX _ikke_: okay, thanks for the suggestions20:03
← thillux_ left20:04
← chrisgeorge left20:05
→ dmsuperman joined20:05
→ kcj joined20:05
dmsuperman git checkout -- app/submodule, git submodule update and git reset app/submodule will not reset my submodule20:05
which is marked dirty20:05
and git diff just says "dirty" so I can't see what's dirty about it20:05
thorbjornDX dmsuperman: can you cd to the submodule and run the checkout from there?20:05
_ikke_ dmsuperman: what does git status tell you20:06
← duckxx left20:06
← Raging_Hog left20:06
_ikke_ thorbjornDX: git checkout doesn't really do anything20:06
dmsuperman (modified content) _ikke_20:06
→ NcA^___ joined20:06
_ikke_ dmsuperman: Then you have to go to the submodule directory, and either commit the changes, or remove them20:06
dmsuperman: Before committing, make sure if you are actually on a branch20:06
dmsuperman Oh I see -- there's no way to do that without CDing into each submodule?20:07
← NcA^ left20:07
_ikke_ dmsuperman: Nope20:07
dmsuperman Why wouldn't git checkout -- app/submodule work but cd app/submodule git checkout -- . does?20:07
→ zaafouri_ joined20:07
_ikke_ dmsuperman: they're just each their own repository20:07
dmsuperman Doesn't make sense to me20:07
→ chrisgeorge joined20:07
dmsuperman Yeah but I can update them from the top level20:07
thorbjornDX dmsuperman: you can do 'git submodule foreach <somecmd>', but you can't target individual repositories with that20:07
dmsuperman and everything else20:07
_ikke_ dmsuperman: Because the parent repo doesn't really know about the contents of the submodule20:07
dmsuperman sure it does, git status tells me about it :P20:07
Enough to prevent me from rebasing in the parent20:08
if the submodule is dirty20:08
← jperry2__ left20:08
_ikke_ dmsuperman: Yeah, but that's all git knows20:08
← madsy left20:08
dmsuperman Right so if it knows that much, the ability to clean it up seems like a logical next step to me20:08
anyway that answers my question20:08
just pointing out the inconsistency20:08
→ abletony84 joined20:09
dmsuperman thanks20:09
milki dirty modules!20:09
→ jperry2_ joined20:09
← rednaks left20:09
abletony84 Hey anybody know what's wrong? (I just reinstalled Windows+Cygwin) % git clone https://github.com/myforum/myforum.git20:09
error: error setting certificate verify locations: CAfile: /usr/ssl/certs/ca-bundle.crt CApath: none while accessing https://github.com/myforum/myforum.git/info/refs - fatal: HTTP request failed20:09
cmn your curl doesn't have/know where to find the root CA certificates20:10
→ TeckniX joined20:11
→ JSharpe joined20:11
cmn how to fix that is curl/cygwin specific20:11
abletony84 thanks20:11
→ NiveusLuna joined20:12
NiveusLuna Let's say that I already have a git repo locally, and want to make a remote version of it that I can push to. What's the best way to go about this?20:13
→ armenb joined20:13
← kumbayo left20:13
cmn log in to the server and run git --bare init in the appropriate place20:13
if you're using gitolite, put it in the config, for other systems, use whatever they have20:13
NiveusLuna that's it?20:13
then push once it's configured, obviously20:14
cmn yeah20:14
thorbjornDX NiveusLuna: it helps to have a git user if you want multiple people to access the remote, and gitolite helps with per-repo permissions quite a bit20:14
NiveusLuna wow, that's nice and easy20:14
← dmsuperman left20:14
← e66 left20:14
← dzonder left20:15
← jason237 left20:15
talexb NiveusLuna, Or 'git remote add origin git@git:$yourRepoNameHere', then 'git push origin master' ..20:15
Where git@git is the appropriate way to get to your git server.20:16
cmn except for that you need the remote repository to exist or have gitolite wild repos20:16
← TeckniX left20:16
talexb gitolite++20:16
NiveusLuna talexb: like, say, "user@host:/path/to/reponame"20:16
_ikke_ talexb: thank sitaram for that20:16
talexb cmn Nope -- not the way I have mine set up .. pushing to the remote creates the repo for me.20:17
_ikke_, I have. :)20:17
← srayner left20:17
cmn what?20:17
you have your own system?20:17
talexb cmn My own git server? Yes. (And one at home, too.)20:18
← JSharpe left20:18
cmn system for managing repos20:18
→ JSharpe joined20:18
cmn since you say that you don't need gitolite wild repos20:18
talexb cmn Just git running on a server, along with gitolite for administration.20:18
NiveusLuna uhhh.... how do you modify the repo config with msysgit?20:19
Git Extensions is also acceptable. I have both.20:19
talexb cmn I may have mis-communicated .. I have gitolite on my git server.20:19
cmn no, it looks like you stopped reading mid-sentence20:20
talexb cmn But I still don't need to ssh onto the server and do a git init --bare .. I just push what I have locally to the server, and It Just Works(tm).20:20
cmn you never have to do that with gitolite20:20
← abletony84 left20:21
dangerousdave why is restart.txt not ignored? http://pastie.org/490448720:21
cmn because you haven't told git to ignore it20:21
the one rule you have mentioning tmp/ skips over that level20:22
and the '**' isn't doing anything20:22
dangerousdave cmn: oh20:22
cmn: ok, thanks20:23
→ asteve joined20:23
→ madsy joined20:23
← madsy left20:23
→ madsy joined20:23
dangerousdave cmn: whats the point of it? (its not my code)20:23
the ** i mean20:23
← prudnikov left20:23
→ hidekin joined20:23
← JSharpe left20:24
_ikke_ dangerousdave: there is no point20:24
→ JSharpe joined20:24
dangerousdave like a sphere?20:24
cmn it's a failed attempt at using rsync wildmatching20:25
dangerousdave or a sausage?20:25
oh20:25
_ikke_ Like the matrix20:25
dangerousdave :-)20:25
← flavius left20:25
_ikke_ It's very simple. * matches everything except for /20:26
dangerousdave that's fortunate, because i too, am very simple20:27
← madsy left20:27
wereHamster dangerousdave: gitignore doesn't support ** matching20:27
dangerousdave hmm20:27
→ imachuchu joined20:27
wereHamster dangerousdave: and there is no need to use ** in your case. Just put tmp/ into your .gitignore20:27
dangerousdave i am sure that was some sort of rails standard gitignore, bare with me20:27
wereHamster I doubt that.20:28
NiveusLuna Think I got it!20:28
argh, this test will take a while20:28
cmn there is work to include this wildmatching, but it probably won't be there for a while20:28
← JSharpe left20:29
→ john_anderson joined20:29
→ JSharpe joined20:29
dangerousdave wereHamster: _ikke_: cmn: http://railscasts.com/episodes/96-git-on-rails, http://stackoverflow.com/questions/2542353/what-is-a-good-gitignore-to-use-with-rails-on-heroku20:30
there's more20:30
cmn all those say you should use tmp/**? a perfect case for !blog20:31
gitinfo Blog posts, while helpful and informative, are quite often outdated, give bad advice, or are just plain wrong. Please don't rely solely upon them, or treat them as authoritative.20:31
wereHamster dangerousdave: https://github.com/rails/rails/blob/master/railties/lib/rails/generators/rails/app/templates/gitignore20:31
→ steveoliver joined20:31
→ duckxx joined20:31
→ _Vi joined20:31
wereHamster dangerousdave: somebody probably told him that ** doesn't work.20:31
dangerousdave: https://github.com/rails/rails/commit/9141c1fa6d54024d8bd73774633cf79a601d1dfa20:31
← njbair left20:31
← john_anderson left20:31
dangerousdave wereHamster: that seems like a better authority20:31
← jceb left20:32
cmn heh, minor fix?20:32
dangerousdave wereHamster: ha!20:32
thanks guys!20:32
→ madsy joined20:33
← madsy left20:33
→ madsy joined20:33
dangerousdave i knew you were right, just needed proof ;-)20:33
→ alunduil joined20:33
→ njbair joined20:33
cmn man gitignore is pretty authoritative20:33
gitinfo the gitignore manpage is available at http://git-scm.com/docs/gitignore.html20:33
← asteve left20:33
← _Vi left20:35
← SebastianFlyte left20:35
← daniel_- left20:35
← robustus left20:35
→ _Vi joined20:35
← hellopat left20:36
thorbjornDX man, gitignore is pretty cool20:37
rohan32 → rohan32|afk20:37
→ jesseFromYVR joined20:37
NiveusLuna how very interesting.20:37
it works great20:37
thank you guys!20:37
thorbjornDX NiveusLuna: do you know about ssh-agent?20:38
← _iron left20:38
thorbjornDX NiveusLuna: it saves you from typing in your ssh passphrase each time you ssh to a machine or do a git push20:38
→ daniel_- joined20:40
→ SebastianFlyte joined20:40
→ nelse_ joined20:40
← ncr100 left20:42
← geekbri left20:43
→ sontek joined20:43
→ btree_ joined20:44
→ BSaboia joined20:44
← chuckharmston left20:45
← zaafouri_ left20:45
→ chuckharmston joined20:46
← btree left20:47
← madsy left20:47
← kenperkins left20:47
→ kenperkins joined20:48
← koltroll left20:49
← jperry2_ left20:49
→ maxandersen joined20:51
← maxandersen left20:51
→ maxandersen joined20:51
← Loggie left20:52
→ manithree joined20:52
→ alunduil_ joined20:52
→ Shadeness joined20:52
← cannonball left20:54
→ john_anderson joined20:54
← alunduil left20:55
→ cannonball joined20:56
→ joaoh82 joined20:57
Industrial I made a clone over ssh, what is the correct way to push it back? https://gist.github.com/20b9ea6a8c98919d200220:57
← jesseFromYVR left20:58
← dogarrhea left20:58
→ jesseFromYVR joined20:58
← john_anderson left20:59
← tdanderson__ left20:59
← les_sylvains left20:59
← glennpratt left21:00
→ john_anderson joined21:00
Industrial I also made some local branches that I'd like to synchronize between the two (wrong thinking?), I'm not sure how :)21:00
→ asteve joined21:01
→ jcao219 joined21:02
← chrisgeorge left21:02
← joaoh82 left21:02
→ chrisgeorge joined21:03
← bartman`` left21:04
← C8H10N4O2 left21:04
→ jason237 joined21:05
← CrazyGir left21:05
← ehsan left21:06
→ bartman`` joined21:06
→ ehsan joined21:06
→ CrazyGir joined21:06
← bgerber left21:06
← nwest left21:06
CrazyGir → Guest2543921:06
→ skorgon joined21:06
← skorgon left21:06
→ skorgon joined21:06
← txdv left21:07
gitinfo set mode: +v21:07
→ txdv_ joined21:07
← john_anderson left21:07
→ LeMike joined21:07
← LeMike left21:07
→ john_anderson joined21:08
→ LeMike joined21:08
m1sc Industrial: regarding your first question: !bare21:09
gitinfo Industrial: an explanation of bare and non-bare repositories (and why pushing to a non-bare one causes problems) can be found here: http://bare-vs-nonbare.gitrecipes.de/21:09
← LeMike left21:10
→ LeMike joined21:10
Industrial m1sc: I see21:11
m1sc Industrial: synchronize?21:11
imachuchu Industrial: to sync branches between two repos use push and pull/fetch (go read up the difference between the two)21:11
→ klj613 joined21:11
Industrial nvm, just pushing created the branches21:11
er, right.21:11
→ orkz joined21:11
→ ncr100 joined21:11
← goshawk left21:12
← irqq left21:13
← guampa left21:13
→ glennpratt joined21:13
← LeMike left21:14
← uirono left21:14
Industrial and git clone has --bare, yay my files are Saved(tm)!21:14
← btree_ left21:15
→ apok_ joined21:15
← bartek left21:15
← juvenal left21:15
→ cakehero joined21:16
← apok left21:17
apok_ → apok21:17
← txdv_ left21:18
→ uirono joined21:19
→ pdtpatrick joined21:20
← skorgon left21:21
→ Sonderblade joined21:21
Kronuz hello21:22
← sorin left21:22
Kronuz hey, I'm using git to hold some files which are utf-16 encoded21:22
→ Xethron joined21:22
← Guest25439 left21:22
Kronuz I added [diff "localizablestrings"] textconv = "iconv -f utf-16 -t utf-8" *and* [encoding "utf16"] textconv = "iconv -f utf-16 -t utf-8" to my .gitconfig21:22
and then I have "*.csv diff=localizablestrings" in a file named .gitattributes21:23
(*.csv are the utf-16 encoded files)21:23
→ diverdude joined21:23
Kronuz …now, the git diff works fine and finds the differences ok21:23
diverdude did anybody in here manage to integrate git with redmine?21:23
← legumbre_ left21:24
Kronuz …however, when I do `git add -p`, I don't get the differences listed there.. and neither when a merge occurs, it only says "conflicts" in binary file or something like that21:24
→ g_ joined21:24
Kronuz what am I missing?21:25
← talexb left21:25
Kronuz …unfortunately I couldn't find any resource anywhere explaining how to properly "fix" this issue21:25
the .gitattributes and .gitconfig as I have them is as far as I could get21:26
→ bgerber joined21:27
← g_ left21:28
EugeneKay Kronuz - probably support for UTF-16 in git-merge.21:28
→ guampa joined21:28
→ gdoteof2 joined21:28
← gdoteof2 left21:28
← tplaner_ left21:28
→ _g joined21:29
_g i needed to move two recent commits from master to a new branch. i did it, but now master is behind remote/master by two commits, and i don't know how to push it my github repo21:29
→ Sonder joined21:29
→ CrazyGir_ joined21:30
CrazyGir_ → Guest9998921:30
imachuchu _g: how did you move them? Were those two commits already pushed to remote/master?21:30
_g imachuchu: yes21:30
← revagomes left21:30
_g i basically did this:http://stackoverflow.com/questions/1628563/move-recent-commit-to-a-new-branch21:30
→ opus joined21:31
← john_anderson left21:31
→ revagomes joined21:31
Kronuz http://www.entropy.ch/blog/Developer/2010/04/15/Git-diff-for-Localizable-strings-Files.html21:31
imachuchu _g: well there is the problem. Your master says "there are two newer commits on the remote" which are the two that you moved to a seperate branch21:31
_g yes i am aware :)21:31
that is why i am in here21:31
Kronuz *that's what I'm using21:31
_g i am the admin of that repo21:31
→ john_anderson joined21:32
_g i just shouldn'thave made the commits there21:32
← devsharpen left21:32
_g i realized today21:32
← Xethron left21:32
_g having made them a couple days ago21:32
← asteve left21:32
moben and you want the commits gone from the remote?21:32
_g so i should just make a new commit that diffs against the new branch i guess21:32
imachuchu _g: if no-one else cares/has seen the remote repo then you can just force that master branch to point at the earlier commit, but that changes public history21:33
→ tomkralidis joined21:33
moben _g look at !rewrite21:33
gitinfo [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to jump through hoops (and you have to tell them to), so it's infinitely better to just move on without rewriting. That said, you can use `git push -f` to force your new history through (may get rejected, though). See http://git-scm.com/docs/git-rebase.html#_recovering_from_upstream_rebase21:33
imachuchu (sorry, didn't have the bot reference handy)21:33
→ ive joined21:33
_g ok21:33
→ txdv joined21:34
← SegFaultAX|work left21:34
← trungqtran left21:34
← jason237 left21:34
moben so it is best to either live with it or add them back to your master and revert them there21:34
_g so the "right" thing to do is somehow apply a diff between where i want the branch to be and where master is on the remote branch, and apply that to my local machine, and *then* push21:34
tomkralidis hi, I have a project (GitHub) where there were numerous changes made to master branch (e.g. from 10 September 2012 to today). I'd like to apply these to another branch in the repo? Any suggestions on how to do this?21:35
← _Vi left21:35
moben _g: add them back and 'git revert' them. that's what I'd do21:35
← Guest99989 left21:35
_g moben: ah revert ok thank you21:35
moben tomkralidis: another branch that you already have?21:36
tomkralidis moben: yes, the other branch exists, and was created on 10 September 2012 as part of a release.21:36
→ _Vi joined21:36
tomkralidis the truth is we neglected to apply these changes to the e.g. 1.4 branch in addition to master21:37
← wilsonj left21:37
moben so it is a selection of bugfixes that you want for 1.4.1?21:37
→ CrazyGir2 joined21:37
moben or do you want all of master in 1.4 too?21:37
← imachuchu left21:38
← DeviceZer0 left21:39
← svetlyak40wt left21:40
→ juvenal joined21:40
tomkralidis moben: preferably a selection of bugfixes (still sifting), but a wholesale push as well as an option21:41
→ iguanna joined21:41
moben you can `git cherry-pick` individual commits21:42
← pantsman left21:42
← martinjlowm left21:43
moben or just `git merge` your master branch into the other one21:43
not sure if there are other options21:43
tomkralidis ah, ok. that easy.21:43
tomkralidis is looking up git cherry-pick21:43
tomkralidis thanks moben21:43
← rohan32|afk left21:44
← AaronMT left21:44
→ rohan32 joined21:44
tomkralidis so I would run git cherry-pick from the e.g. 1.4 branch, and it would recognize the master branch SHAs?21:44
→ AaronMT joined21:45
→ ghartmann joined21:45
moben the commit hashes are not tied to a branch21:45
well, not really21:45
so you can just run `git cherry-pick 12313234`, yes21:46
→ shirish joined21:47
moben you may get to resolve conflicts21:48
← Sonder left21:48
← jesseFromYVR left21:48
tomkralidis thanks again21:48
← Sigma left21:49
← whitman left21:49
← timj1 left21:49
→ timj joined21:50
→ nowhere_man joined21:51
→ fir_ed joined21:52
← Sonderblade left21:53
fir_ed How do I download a copy of a repo?21:53
Vinnie_win fir_ed: git clone <url>21:53
fir_ed ah. clone.21:53
ty.21:53
→ Sonderblade joined21:54
← ajw0100 left21:55
← dv310p3r left21:55
← glennpratt left21:56
← john_anderson left21:58
← iguanna left21:58
→ gaius65 joined21:59
gaius65 Hey guys!21:59
Vinnie_win sup21:59
→ john_anderson joined21:59
gaius65 I have two branches: master, new-feature21:59
Vinnie_win congrats21:59
→ asteve joined21:59
klj613 lol21:59
gaius65 Vinnie_win: ups, I hit enter to early^^21:59
Vinnie_win You should end each transmission with "over" over.22:00
gaius65 After committing a few things to new-feature, I now commited a bug-fix to master. How can I now also get the bug fix into my new-feature branch?=22:00
← v0n left22:00
Vinnie_win copy that gaius65, use git merge master, over.22:00
moben gaius65: or cherry-pick22:00
← njbair left22:00
Vinnie_win gaius65: or git rebase22:00
klj613 gaius65, merge, rebase, cherry-pick. take your pick ;p22:00
moben but only use rebase if you did not push new-feature yet22:01
gaius65 klj613: what's the easiest way to get just all commits from master into new-feature?22:01
← ncr100 left22:01
gaius65 new-feature is already pushed22:01
tomkralidis gaius65: my question exactly a few minutes back! See convo22:01
→ P1RATEZ joined22:01
Vinnie_win moben: well even if you pushed new-feature to a remote, if no one else is looking at the branch then it wouldn't hurt...much22:01
klj613 easiest? to make sure feature has everything master has? merge22:01
← FrenkyNet left22:01
Vinnie_win I have --force on a hot key22:02
← manithree left22:02
→ Vile joined22:02
→ jamuraa joined22:03
→ jesseFromYVR joined22:05
→ chaiz joined22:05
← Destos left22:05
← dpino left22:06
gaius65 Vinnie_win, klj613: thanks, worked :)22:06
Vinnie_win Roger that22:06
← maxandersen left22:06
← derpops left22:07
← ncd|leeN left22:07
klj613 gaius65, if master had something you didnt want to merge in... then you'd use cherry-picks. (or rewrite the whole history to your will... if you know how :p)22:07
← shirish left22:07
Vinnie_win with a merge though, now the order of commits is different between master and feature-branch...right?22:07
jamuraa is the easiest way to make a new branch without a specific commit, just to branch from a common ancestor and cherry-pick the ones you want? or can I make a new branch, rebase -i against a common ancestor?22:08
klj613 Vinnie_win, open up a GUI tool and have a look22:08
→ StuckMojo joined22:08
Vinnie_win Yeah I think it is jlk22:08
← NiveusLuna left22:08
Vinnie_win *klj22:08
StuckMojo does git-push origin already use ssh/scp compression?22:08
gaius65 klj613: ok. I'm trying to keep my git usage simple, so this is good :)22:08
klj613 jamuraa, without a specific commit? i'd use rebase -i (cherry pick be.. easier to use). and that will be rewriting history so you'd have to force push22:09
→ DeviceZer0 joined22:09
rohan32 → rohan32|afk22:09
klj613 gaius65, make a dummy repo. spend a few hours destroying it whilst looking at a GUI. learn how git behaves :p22:09
← Sonderblade left22:09
← tomkralidis left22:10
sitaram esc: (and also _ikke_ and PerlJam) -- re scaling to 10,000 users. Let's decouple the "compile" step, which only happens once in a while, from the access check that happens for every "contact". The latter is O(1) now so it doesn't matter. Your real problem will be sshd's linear scan over the authkeys file, so if you use httpd (with mod_auth_ldap instead of htpasswd) it should work already22:10
← fir_ed left22:10
gaius65 klj613: I already did… but I use stuff like merge once every three, four month… so it's easy to forget a bit everythime...22:10
Vinnie_win: histories seem to be both be good22:10
Vinnie_win When will git-subtree become part of the official distro22:10
→ maxandersen joined22:10
← maxandersen left22:10
→ maxandersen joined22:10
klj613 :o i use merge all the time... rebase daily (sometimes more than merge!) lol22:10
milki Vinnie_win: yesterday22:11
Vinnie_win milki: really?22:11
milki Vinnie_win: or in two months22:11
← sq-one left22:11
moben I thought it was already22:11
Vinnie_win there's nothing in the release notes about it22:11
moben merged to contrib/ someone said iirc22:11
← dangerousdave left22:11
jamuraa klj613: I don't want to rewrite history, just have a new branch (branch-two) that only has 1,3,4,5,6 of the 6 commits that I have on branch-one. (both branch from master)22:11
klj613 why dont you merge back into master? or do your feature branches actually take months lol. im shocked you rarely use merge22:12
→ v0n joined22:12
← duckxx left22:13
→ Sonderblade joined22:13
→ SegFaultAX|work joined22:13
← rtjure left22:13
sitaram PerlJam: (re esc's question) fedora's config is already 11,000+ repos, and the conf file is 126,000 lines. The killer here is the POST_CREATE scripts actually, when I try it with v3 (or the eqvt in v2). Luckily fedora don't use all that.22:14
← ive left22:14
← kermit left22:15
→ storrgie joined22:16
← pandeiro left22:17
← TheJH left22:17
← phantomtiger left22:18
← ISF left22:18
gaius65 cya guys22:19
← lacrymology left22:22
← P1RATEZ left22:23
← jargon- left22:23
→ berserkr joined22:23
← RaceCondition left22:28
← juvenal left22:28
→ RaceCondition joined22:29
ghartmann → teste22:29
teste → ghartmann22:29
← Praise left22:30
→ ShadesofGraylin| joined22:30
→ Lulzon joined22:30
→ duckxx joined22:31
← chrisgeorge left22:31
← storrgie left22:31
→ Praise joined22:32
→ yoklov joined22:34
→ xaka joined22:35
← _Vi left22:35
← ghartmann left22:35
← klj613 left22:35
→ _Vi joined22:36
← MyTallest left22:37
→ arturaz joined22:37
xaka if i want to merge A to B and A is behind master repo, should i pull the changes to A first and then do merge or i can pull changes to B from A directly? (git checkout B; git pull origin A). Which one way would be better?22:38
adhawkins → adhawkins-away22:38
← und3f left22:39
thiago doesn't matter22:39
← arietis left22:40
rohan32|afk → rohan3222:41
← alunduil_ left22:41
→ pts\3 joined22:42
← gaius65 left22:42
← fly9 left22:43
pts\3 i cloned a repo, deleted a file, now i want to resync .. yet any combination of pulling, fetching, and merging i do just says "already up to date"22:44
← zastern left22:44
← asteve left22:44
→ _huoxito joined22:45
← eijk left22:45
\bMike\b pts\3: What does "resync" mean?22:45
→ juvenal joined22:45
→ Chib joined22:46
pts\3 i want to make my local directory match the repository , as it did when i cloned it22:46
← zorzar left22:47
→ cjz joined22:47
cjz hmm22:47
all my gitosis repos have file called description with the name in the repo22:47
all the new gitolite repos have no file called that?22:48
← hwrd|work left22:48
← v0n left22:48
pts\3 i tried every command on this page to no avail. https://help.github.com/articles/fork-a-repo22:48
milki correct, its not default anymore22:48
cjz the post-receive.email script im using seems to look for that file22:48
← RaceCondition left22:48
\bMike\b pts\3: git reset --hard HEAD?22:48
Vinnie_win I just added a new remote, how do I check out a local branch that trakcs a remote branch in that remote?22:49
pts\3 oh my, \bMike\b , that is very nice. thank you.22:49
→ Agis__ joined22:50
← juvenal left22:50
\bMike\b np22:50
→ storrgie joined22:50
Agis__ hey guys, I have 2 local branches and I want to checkout a whole directory and it's contents from branch B to branch A (which is my current)22:50
overriding the directories, any ideas?22:50
→ monokrome joined22:50
thiago git checkout B -- dirname22:51
monokrome Hi. Is it intended that after doing `git checkout --orphan somebranch`, the ,gitignore still exists in the new branch?22:51
← chuckharmston left22:51
monokrome seems odd22:51
Vinnie_win how do I check out a remote branch22:52
Agis__ thiago: so it's git checkout <branch-name> -- example/ ?22:52
→ chuckharmston joined22:52
← sgambino_ left22:52
EugeneKay monokrome - !orphan, see 2nd stop22:53
gitinfo monokrome: To create an orphaned branch(unrelated to existing history): `git checkout --orphan newbranchname`. This will leave your index/worktree as-is(use `rm .git/index; git clean -dfx` to delete EVERYTHING). If you get 'error: unknown option `orphan`' see !orphan_old. For an empty/null commit see !orphan_null.22:53
EugeneKay Vinnie_win - 'git checkout -b foo remote/foo'22:53
Vinnie_win EugeneKay: Thanks22:53
→ jargon- joined22:53
EugeneKay Agis__ - rm -r foo/; git checkout bar -- foo/22:53
Agis__ - the second part will do it, but will leave any files which exist in your current branch but not the other one22:54
Hence rm first22:54
→ juvenal joined22:54
→ C8H10N4O2 joined22:55
Agis__ i see22:55
← simesy_work left22:55
Agis__ and how I undo any changes I've made and not commited?22:55
cjz solved my own problem22:55
→ simesy_work joined22:55
cjz using $GL_REPO rather than looking for description22:55
EugeneKay Agis__ - man git-reset22:55
gitinfo Agis__: the git-reset manpage is available at http://git-scm.com/docs/git-reset.html22:55
← apok left22:58
→ apok joined22:59
← macabre left22:59
← revagomes left23:00
→ v0n joined23:00
← rohan32 left23:01
→ zonderH joined23:01
→ rohan32 joined23:01
← Sky[x] left23:04
← jargon- left23:04
moben !orphan_null23:05
gitinfo To create an orphan branch with the null/empty commit, use: `git checkout -b orphan $(echo "empty" | git commit-tree 4b825dc642cb6eb9a060e54bf8d69288fbee4904)`23:05
← siprbaum left23:07
→ siprbaum joined23:07
← v0n left23:07
← guide left23:07
→ hwrd|work joined23:08
→ madewokherd joined23:08
→ guide joined23:09
→ ajw0100 joined23:11
← SirCmpwn left23:12
→ Destos joined23:12
← giallu left23:13
monokrome EugeneKay: So, it's basically not any easier than just pushing an empty repository's master branch to a remote branch?23:13
← marhaban left23:14
← hidekin left23:16
← Destos left23:16
EugeneKay Uh, huh?23:17
frogonwheels monokrome: !xy23:17
gitinfo monokrome: This sounds like an "XY Problem" http://mywiki.wooledge.org/XyProblem So let's step back for a minute. What are you actually trying to achieve? Why are you doing it this way?23:17
EugeneKay I'm not sure what you're really trying to do23:17
monokrome I'm rewriting a project and want an empty branch23:18
EugeneKay checkout --orphan gives you a branch(and HEAD) that is a new initial commit. This is roughly analogous to the state that you get from 'git init'23:18
If it's a total project rewrite I would create a new repo and place for it on whatever centralized server you're using.23:18
← storrgie left23:18
rohan32 → rohan32|hw23:19
SirCmpwn_ → SirCmpwn23:19
→ akiress joined23:20
← SegFaultAX|work left23:22
← macrover left23:22
cbreak monokrome: an empty repository can not be used as source for a push23:24
it has no history for pushing23:24
→ duckx joined23:24
→ storrgie joined23:24
← Synthead left23:24
← arturaz left23:25
← simesy_work left23:25
← storrgie left23:26
→ simesy_work joined23:26
← ShadeTornado left23:26
← duckxx left23:27
→ duckxx joined23:28
← simesy_work left23:29
→ simesy_work joined23:29
→ Chillance joined23:29
→ SegFaultAX|work joined23:30
← jesseFromYVR left23:30
→ jesseFromYVR joined23:31
← duckx left23:31
← guampa left23:33
→ pfFred joined23:34
← babilen left23:35
← john_anderson left23:35
→ FernandoBasso joined23:35
← _Vi left23:36
→ guampa joined23:36
→ _Vi joined23:37
← pfFredd left23:37
← chuckharmston left23:38
← pfFred left23:39
→ msmithng joined23:40
→ babilen joined23:41
→ rchavik joined23:42
← rchavik left23:42
→ rchavik joined23:42
→ blorbx joined23:43
← mmc1 left23:44
SeoZ-work[AWAY] → SeoZ23:46
→ jperry2_ joined23:47
→ chrisf|walrus joined23:47
← Fandekasp left23:47
← gusnan left23:49
← SegFaultAX|work left23:54
→ Milossh joined23:54
← akiress left23:55
→ SegFaultAX|work joined23:56
← _g left23:57
← opus left23:57
→ adamben joined23:57
← hwrd|work left23:58
→ BSaboia__ joined23:58
← cjz left23:59
→ hwrd|work joined23:59
← BSaboia left23:59
← jperry2_ left23:59

Logs Search ←Prev date Next date→ Channels Documentation