IRCloggy #git 2011-05-08

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.

2011-05-08

→ saschagehlich joined00:00
← v0n left00:00
→ dotblank joined00:01
dotblank Anyone know very easy to understand git tutorials? Need to link instructions in a wiki.00:02
← sylr left00:02
→ Err404NotFound joined00:02
canton7 git pro is good00:03
← malumalu left00:03
→ kenneth_reitz joined00:09
← candybar left00:09
→ aquabu joined00:11
← unreal left00:12
← Prism left00:13
← TomyLobo left00:14
→ lukeman joined00:18
→ sbell joined00:18
→ unreal joined00:19
← jjuran left00:19
→ __sorin__ joined00:20
→ candybar joined00:21
← aruiz left00:21
→ Freakshow joined00:24
← canton7 left00:27
← coldboot|home left00:29
← rendar left00:30
→ v0n joined00:33
← Leemp left00:33
← candybar left00:35
→ jjuran joined00:38
← gigamonkey left00:40
→ candybar joined00:41
← SethRobertson left00:41
← tty2 left00:43
→ tty3 joined00:46
← NetHawk left00:49
→ NetHawk joined00:50
← pantsman left00:50
← aquabu left00:50
→ thews joined00:51
→ hackez_work joined00:54
← candybar left00:56
← Bauer left01:01
→ hvgotcodes joined01:02
← babilen left01:02
← deekayen left01:05
← saschagehlich left01:05
→ Swimming_bird joined01:06
→ goldins73 joined01:07
← fs-sylr left01:08
→ fs-sylr joined01:08
yaotti_aw → yaotti01:08
← elb0w left01:08
← Yuuhi left01:08
← Bucciarati left01:08
→ elb0w joined01:09
← RaceCondition left01:09
→ Bucciarati joined01:10
→ candybar joined01:13
← arunce left01:15
→ arunce joined01:15
← g0bl1n left01:17
← thews left01:17
← davidcoallier left01:18
→ davidcoallier joined01:18
→ ezkl joined01:21
→ SethRobertson joined01:21
← eletuchy left01:22
← davidcoallier left01:23
← goldins73 left01:24
→ davidcoallier joined01:24
← jophish left01:26
→ thews joined01:30
← superdmp left01:32
→ ianawilson joined01:38
← thews left01:40
← lolfrenz left01:40
→ Bauer joined01:42
→ thews joined01:43
→ darkthought joined01:44
→ mikepack joined01:44
→ seivan joined01:45
← BeWilled left01:45
← SimonP86_ left01:47
← MissionCritical left01:47
→ alexandernst joined01:47
← Vampire0_ left01:48
ianawilson anybody have experience with gitosis? I'm having a terrible time getting users added to it01:48
alexandernst Removing a file and then going "svn up" will get the latest file from the repo. What's the equivalent in git to that?01:48
← NetHawk left01:49
← paeva left01:50
→ rgr joined01:52
→ FylGood joined01:52
→ eletuchy joined01:52
← eletuchy left01:52
← FylGood left01:52
→ eletuchy joined01:52
← linusoleander left01:54
→ paeva joined01:55
→ linusoleander joined01:58
SavoxAway → Savoxis01:59
→ NetHawk joined01:59
← AAA_awright left02:01
→ AAA_awright_ joined02:01
→ Vampire0_ joined02:02
← Arafangion left02:04
← kenneth_reitz left02:04
cirwin alexandernst: git requires you to think more than that. Probably git fetch && git checkout origin/master <file>02:04
alexandernst cirwin: thanks. I was just reading man02:04
did it with checout <file>02:04
s/checout/checkout02:04
cirwin ok02:04
← davidcoallier left02:06
→ deekayen joined02:09
← kukks left02:10
← __sorin__ left02:11
ianawilson anybody have experience with gitosis?02:12
SethRobertson !gitosis02:13
jast [automatic message] 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/gitolite02:13
ianawilson o02:14
wow02:14
that would have saved a ton of time02:14
← sion2k left02:14
SethRobertson tell that to whomever recommended gitosis02:14
ianawilson the first two things i read about it (https://help.ubuntu.com/community/Git) and some git + redmine recommended gitosis02:14
i assumed that i could trust the help.ubuntu.com guide02:14
SethRobertson file a ticket against ubuntu then02:14
ianawilson huh02:14
← Destos left02:15
← jnylen left02:15
→ MissionCritical joined02:15
ianawilson do you have any suggestions for a guide to get me started with gitolite?02:15
← linusoleander left02:16
SethRobertson The URL above is probably the best palce to get started02:16
← deekayen left02:16
← sergio left02:17
ianawilson ok02:17
well thank you!02:17
seriously saved me from pounding my head against it and google02:18
rudi_s !gitolite02:18
Hm.02:18
ianawilson hah02:18
:)02:18
rudi_s Worth a try ;-)02:19
← MikhX left02:19
sitaram rudi_s: <bang>gitosis I think. Although it would be nice if !gitolite replied with http://sitaramc.github.com/gitolite/doc/overkill.html :-)02:21
→ n8o-mba joined02:21
rudi_s jast: ^02:21
Yeah, !gitosis works, I thought !gitolite would provide some more details.02:22
jast Yeah: [automatic message] 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/gitolite02:22
rudi_s Hehe.02:22
→ jnylen joined02:22
rudi_s jast: ^ - maybe only reply if the nick exists in the channel02:22
→ kukks joined02:23
← vanmik left02:24
← LekeFly left02:25
→ Leemp joined02:25
← cybernd left02:27
→ verto joined02:28
→ MikhX joined02:28
→ brienna joined02:28
← blorbx left02:30
← drev1 left02:31
→ DrDogwelder joined02:33
← brienna left02:33
← n8o-mba left02:33
← MikhX left02:33
→ g0bl1n joined02:35
→ Arafangion joined02:35
← lastk left02:35
→ MikhX joined02:36
→ shiba_yu36 joined02:36
→ gigamonkey joined02:37
→ Textmode joined02:37
→ __sorin__ joined02:38
← Heimidal left02:38
← Behi left02:39
→ intripoon joined02:42
→ iori joined02:46
← intripoon_ left02:46
→ Rotham joined02:46
← jjuran left02:46
→ zhengyang joined02:48
→ mando joined02:49
→ verto_ joined02:50
← verto left02:50
← Anubisss left02:51
→ davidcoallier joined02:53
← hobodave left02:54
→ FrancisHaibin joined02:55
→ t3i joined02:56
← DrDogwelder left03:03
← Chillance left03:04
← sraue left03:09
AAA_awright_ → AAA_awright03:11
→ sraue joined03:11
← tty1 left03:12
← tty3 left03:13
← v0n left03:14
→ dl9pf_ joined03:14
← FrancisHaibin left03:15
← dl9pf left03:15
← theoretical left03:15
→ FrancisHaibin joined03:20
← gigamonkey left03:30
← aef left03:34
→ blorbx joined03:36
→ arunce_ joined03:37
→ aef joined03:38
← arunce left03:39
← alip left03:41
→ v0n joined03:42
→ alip joined03:44
← pyroscope left03:45
→ Corren joined03:46
← tyteen4a03 left03:47
← Textmode left03:47
← Corren left03:49
→ Guest1729 joined03:49
← mikepack left03:53
← g0bl1n left03:54
→ ianmcorvidae joined03:57
← ianmcorvidae left03:57
→ ianmcorvidae joined03:57
→ nadirvardar joined03:59
← mando left04:01
→ xxoxx joined04:02
← bremner_ left04:02
← MikhX left04:03
← OOPMan left04:03
→ bremner_ joined04:03
→ warthog9 joined04:05
→ warthog19 joined04:07
← warthog9 left04:10
← nadirvardar left04:11
→ LongBeach joined04:11
← hvgotcodes left04:12
→ MikhX joined04:15
→ tty1 joined04:18
→ verto__ joined04:18
→ tty2 joined04:18
← verto_ left04:18
→ orbiting joined04:19
→ tty3 joined04:20
→ ajpiaNOU joined04:21
← iori left04:22
← tty2 left04:22
→ OOPMan joined04:22
← tty1 left04:23
→ tty2 joined04:23
→ tty1 joined04:29
← blorbx left04:31
← tty3 left04:32
← tty2 left04:32
→ tty2 joined04:33
← MikhX left04:33
← oc80z left04:35
← ymasory left04:36
→ oc80z joined04:36
← engla left04:37
→ drev1 joined04:37
→ dermoth_ joined04:39
→ tty3 joined04:40
← tty2 left04:42
← dermoth|home left04:42
← tty1 left04:42
→ tty1 joined04:42
← __sorin__ left04:43
← jfkw left04:45
→ blorbx joined04:47
→ sechrist joined04:48
→ f0i joined04:49
→ MikhX joined04:49
← tty3 left04:51
← tty1 left04:52
← ianawilson left04:57
→ caseymcg joined04:58
Escherial say i'm doing a merge and i just want to take any differences from the branch i'm merging when resolving conflicts04:58
how would i go about doing that? :\04:58
i'm already halfway through manually merging them and it's a bit tedious04:58
← drizzd left05:00
← kukks left05:00
← cirwin left05:04
→ cesario joined05:05
→ bindaas joined05:09
→ angelsl joined05:10
warthog19 → warthog905:14
← ajpiaNOU left05:16
← blorbx left05:22
← codeshah left05:23
→ pyroscope joined05:23
← falican left05:24
→ falican joined05:25
→ jjuran joined05:31
→ xiangfu joined05:33
← Guest1729 left05:33
→ dr_ick joined05:34
→ adaptee joined05:35
← adaptee left05:36
→ adaptee joined05:36
→ jablko joined05:37
← sechrist left05:41
← metcalfc left05:42
→ sechrist joined05:42
→ Corren joined05:42
← joel_falcou left05:44
← bauruine left05:47
← Corren left05:49
yaotti → yaotti_aw05:51
yaotti_aw → yaotti05:51
← falican left05:54
→ Guest1729 joined05:55
← Guest1729 left05:55
→ maleknet joined05:56
jmcantrell isn't there a way to echo the root directory of a repo or the directory that '.git' sits in?05:57
Escherial: i think you can do something like git merge -s theirs06:02
Escherial jmcantrell: thanks :) i'll remember that for next time06:02
for the time being, i just manually merged them06:02
→ blorbx joined06:06
yaotti → yaotti_aw06:08
← [intra]lanman left06:09
→ iAsterisk joined06:10
→ aminpy joined06:14
→ Behi joined06:14
→ cirwin joined06:16
← OOPMan left06:16
→ genbattle joined06:19
← drev1 left06:20
→ legumbre_ joined06:22
→ tvw joined06:23
→ namemon joined06:23
← legumbre left06:23
← jablko left06:25
← bgerber left06:25
← TML left06:25
→ sirvaliance joined06:26
← dreiss left06:26
← sirvaliance left06:27
← orbiting left06:27
→ tore_ joined06:36
→ drev1 joined06:36
← heinrich5991 left06:37
→ bgerber joined06:37
→ TML joined06:38
→ heinrich5991 joined06:38
→ ggregori joined06:39
→ tyfighter joined06:44
→ Heimidal joined06:44
← tore_ left06:45
→ Dreadnaught joined06:45
← Dreadnaught left06:45
→ Dreadnaught joined06:45
← Kulrak left06:45
→ tore_ joined06:46
← blorbx left06:47
← davidcoallier left06:48
→ d0k joined06:51
→ urkud joined06:55
← amirite left07:04
→ amirite joined07:05
← smolyn left07:05
→ Raging_Hog joined07:08
→ hubertchang joined07:09
→ robertdavid joined07:10
← up_the_irons left07:12
→ ph^ joined07:13
← cesario left07:13
→ SimonNa joined07:13
→ glcrazy joined07:15
→ cesario joined07:16
→ Spockz joined07:20
← Raging_Hog left07:23
→ Raging_Hog joined07:23
← caseymcg left07:23
→ sirvaliance joined07:24
→ caseymcg joined07:26
→ up_the_irons joined07:27
← cthuluh left07:27
← sirvaliance left07:28
← guardian1 left07:29
← dr_ick left07:29
→ guardian joined07:29
← hubertchang left07:30
← warthog9 left07:31
→ ThomasLocke joined07:32
→ tewecske joined07:33
← tyfighter left07:34
← Spockz left07:36
→ envi joined07:36
← tvw left07:43
← ESphynx left07:54
→ Wizek_ joined07:56
← ajt left07:56
← alberto56 left07:57
← t3i left07:57
→ alberto56 joined07:57
← coldboot left07:58
← elb0w left07:58
→ simplechat joined07:58
← simplechat left07:58
→ simplechat joined07:58
← dermoth_ left07:58
← Mpenz left07:58
→ elb0w joined07:59
← Tommy[D] left07:59
→ coldboot joined07:59
→ Mpenz joined07:59
→ dermoth_ joined07:59
→ Tommy[D] joined08:01
→ vdv joined08:01
← darkthought left08:02
→ workmad3 joined08:03
→ serverhorror_ joined08:04
→ ESphynx joined08:04
→ darkthought joined08:05
→ amirite_ joined08:07
← workmad3 left08:08
→ leo2007 joined08:09
→ freshmilk joined08:09
← amirite left08:10
← verto__ left08:11
← drev1 left08:12
adamholt → adamholt_away08:12
adamholt_away → adamholt08:12
← Tommy[D] left08:13
← coldboot left08:14
← dermoth_ left08:14
← ggregori left08:14
← elb0w left08:14
→ dermoth_ joined08:15
← larva left08:16
→ Tommy[D] joined08:16
→ coldboot joined08:16
→ elb0w joined08:17
→ ch077179 joined08:18
← Mek left08:19
→ larva joined08:20
→ workmad3 joined08:20
→ verto__ joined08:23
→ warthog9 joined08:27
→ Mek joined08:28
← FrancisHaibin left08:32
→ n1ck3 joined08:32
← giallu left08:33
← v0n left08:35
→ jablko joined08:38
← jablko left08:38
→ jablko joined08:38
← seivan left08:39
→ Dave^| joined08:39
← workmad3 left08:40
← jjuran left08:40
→ Spockz joined08:44
→ seivan joined08:46
← dpino left08:46
← Spockz left08:47
← mdel left08:49
→ jjuran joined08:51
← darkthought left08:53
← cirwin left08:54
→ babilen joined08:55
← jepler left08:55
→ rendar joined08:55
→ TomyLobo joined08:56
→ cybernd joined08:57
→ superdmp joined08:57
← n1ck3 left08:59
← adaptee left08:59
← genbattle left08:59
→ airborn joined09:06
→ jepler joined09:07
← caseymcg left09:11
adamholt → adamholt_away09:11
adamholt_away → adamholt09:12
← cesario left09:13
→ drizzd joined09:14
← stamina left09:20
→ psoo joined09:25
← namemon left09:29
← leo2007 left09:29
setmeaway i have question. how can i remove invalid refspec from git repository?09:31
← Heimidal left09:31
→ caseymcg joined09:31
→ aruiz joined09:31
← rendar left09:31
→ pantsman joined09:33
← pantsman left09:33
→ pantsman joined09:33
setmeaway solved. git push origin :refs/for/master09:35
→ leo2007 joined09:36
← Derander left09:36
→ Derander joined09:37
← dermoth_ left09:37
← madewokherd left09:37
← Buglouse left09:38
→ dermoth_ joined09:39
← caseymcg left09:40
→ setmeaway2 joined09:40
← marc_kdab left09:40
← setmeaway left09:43
← shiba_yu36 left09:44
→ shiba_yu36 joined09:45
→ hubertchang joined09:45
→ saschagehlich joined09:46
← hubertchang left09:49
← freshmilk left09:51
← aruiz left09:51
→ bauruine joined09:52
→ tansell-laptop joined09:52
→ freshmilk joined09:54
→ workmad3 joined09:54
→ amerine joined09:56
← workmad3 left09:56
→ TrineLocke joined09:59
← TrineLocke left09:59
→ TrineLocke joined09:59
→ helvete joined09:59
← Savoxis left10:00
← TrineLocke left10:00
→ Savoxis joined10:00
→ dpy joined10:02
← jayne left10:05
→ RaceCondition joined10:07
→ saschpe joined10:07
→ Fullmoon joined10:09
← Fullmoon left10:09
→ jstemmer joined10:11
→ carl-eric joined10:12
→ radarek joined10:14
carl-eric Hi - is there a way for a hook (or something similar) to intercept a merge and possibly to prevent the merge?10:14
← freshmilk left10:14
← pantsman left10:16
→ pantsman joined10:16
← pantsman left10:16
→ pantsman joined10:16
bremner_ intercept where? in your local repo?10:17
carl-eric bremner_: yes.10:17
We have a number of similarly named branches (organization and hysterical reasons, don't ask), and we have a defined merge direction between the various branches10:18
bremner_ you know you can just undo it, right?10:18
nevyn carl-eric: yes.10:18
carl-eric I'd like to recognize when I merge in the "wrong" direction by accident10:18
→ jayne joined10:18
← saschpe left10:18
carl-eric bremner, yes, I know it's undoable, but we've had it happen that it wasn't caught until a few days later, making it a bit more of a hassle10:18
nevyn ooh that'd be useful for me actually.10:19
carl-eric I couldn't find a hook that would let me do that (there is no pre-merge hook, for reasons unknown to me)10:19
nevyn you can hook incomming tho10:19
you mean in your repo.10:20
carl-eric Yes.10:20
bremner_ carl-eric: a heuristic would be to check in pre-commit10:20
← Octalot left10:20
carl-eric I'd like to catch it when I do "git merge <wrong-branch>"10:20
Let me have a look at pre-commit10:20
bremner_ oh, and use --no-fast-forward, so you always get a commit10:20
not ideal perhaps10:20
→ saschpe joined10:21
carl-eric Ah, yes this only works when there are no fast-forwards. Those are nice to have though10:21
bremner_ hmm. No clever ideas. You could replace git-merge with a wrapper...10:22
carl-eric I tried that - but I can't alias it (existing commands can't be shadowed by a git alias)10:22
nevyn carl-eric: read throught the pre-rebase.sample.10:22
might have some ideas.10:23
bremner_ carl-eric: well, you can always put a script git-merge in your path10:24
carl-eric Hm, I might try *post*-merge and just reset back to before the merge.10:24
nevyn carl-eric: rollback if it now has commits that are not in the branch it should have come from ?10:24
carl-eric nevyn: Something along those lines, yes.10:25
I'm going to try something and then post it here if it works.10:25
bremner_ personally I'd just report an error, rather than trying to reset from a hook10:26
carl-eric Well yes, I'm not going to silently reset ;-) Ideally I would like it to recognize where the merge is coming from, and if this particular merge hasn't been done before, ask whether to proceed.10:27
→ freshmilk joined10:28
→ TrineLocke joined10:28
← tansell-laptop left10:29
→ seivan_ joined10:29
→ genbattle joined10:30
carl-eric bleh, post-merge is not called when there was a conflict.10:31
← freshmilk left10:31
← seivan left10:31
→ _iron joined10:33
→ io2 joined10:34
← TrineLocke left10:35
albel727 hmm, isn't that kind of a bug? I guess, git should invoke post-merge on commit with several heads.10:36
nevyn albel727: it'd be nice if there were a pre-merge hook10:37
→ plause joined10:37
→ freshmilk joined10:38
→ awilkins joined10:38
carl-eric from the githooks manpage, about post-merge: "This hook cannot affect the outcome of git merge and is not executed, if the merge failed due to conflicts."10:38
Yes, a pre-merge hook would be very nice.10:38
← plause left10:38
albel727 nevyn: well, there is. kind of. prepare-commit-msg.10:39
← RaceCondition left10:39
albel727 it receives, what kind of commit is pending. regular, merge etc.10:39
and can abort committing.10:40
carl-eric That looks interesting. Maybe prepare-commit-msg together with post-merge can do it.10:40
→ mbroeker joined10:40
carl-eric Still feels like a kludge, but not having pre-merge, I'll give it a shot.10:41
→ ningia joined10:41
← babilen left10:44
albel727 well, a more honest approach might be pre-commit hook with checking, whether .git/MERGE_HEAD exists.10:44
← freshmilk left10:44
→ babilen joined10:45
→ RaceCondition joined10:46
carl-eric albel727: yes, but that won't fire if it's a fast-forward merge10:46
← tro left10:46
→ Anubisss joined10:46
carl-eric that's where I'd fall back to post-merge10:47
albel727 oh. well, I'm pretty sure prepare-commit-msg will also not detect fast-forwards.10:47
you can force --non-ff merges.10:47
← rgr left10:48
albel727 (btw)10:48
carl-eric I know, but I'd like to have ff merges.10:49
I'm going to try and see whether post-merge fires for ff.10:49
cbreak fast forward merges are not real merges10:49
→ Spockz joined10:51
→ tro joined10:51
superdmp when I see a list of deleted files, like: http://dpaste.com/540020/ how can I delete them all at once using git rm?10:52
→ brodo joined10:52
carl-eric superdmp: git ls-files --deleted | xargs -I file git rm "file"10:53
superdmp carl-eric: wow, that's one mouthful of syntax!10:53
carl-eric That's why I put it in a script as "git-rma" in my path10:54
Now I can just say: git rma10:54
albel727 superdmp: you man git add -u10:54
jast superdmp: [automatic message] the 'git-add' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-add.html10:54
superdmp i can imagine10:54
albel727 superdmp: or better yet, you learn of git add --all.10:55
superdmp actually, git add -u seems not to touch them10:56
carl-eric albel727: the git add options all assume that you just want everything added, on top of removing the deleted files.10:56
superdmp (I'm using an old git, btw)10:56
albel727 superdmp: that's because you forgot to add pathspec. do "git add -u ." (note the dot) in top level dir.10:56
carl-eric with my rma script I can just have it rm everything I deleted, but still pick what I want added10:56
superdmp albel727: no, I did git add -u .10:57
oh10:57
ahem10:57
I can't read10:57
albel727 superdmp: they're changes to be committed already.10:57
superdmp albel727: yes10:58
thanks guys10:58
→ Fullmoon joined11:01
→ hubertchang joined11:01
← babilen left11:01
← RaceCondition left11:03
→ RaceCondition joined11:06
← s0undt3ch left11:08
→ s0undt3ch joined11:10
→ tansell-laptop joined11:12
← ThomasLocke left11:12
← aminpy left11:15
← carl-eric left11:15
→ aminpy joined11:16
← tansell-laptop left11:18
→ mkramer1 joined11:18
← hubertchang left11:18
← Escherial left11:19
→ engla joined11:19
→ lolfrenz joined11:19
→ canton7 joined11:20
← mkramer left11:20
→ Prism joined11:21
→ mkramer joined11:21
→ babilen joined11:21
← pantsman left11:21
→ freshmilk joined11:22
← mkramer1 left11:23
← babilen left11:23
→ babilen_ joined11:23
← babilen_ left11:26
→ babilen joined11:27
← babilen left11:27
← jjuran left11:27
← dpy left11:31
→ berserkr joined11:32
→ LekeFly joined11:33
← f0i left11:35
→ babilen joined11:35
← Fullmoon left11:35
← mkramer left11:40
← Rotham left11:40
→ mkramer joined11:40
← brodo left11:40
→ Rotham joined11:40
→ Yuuhi joined11:42
← ph^ left11:44
→ ph^ joined11:45
→ f0i joined11:48
→ jophish joined11:49
← psoo left11:49
← engla left11:54
→ engla joined11:54
→ FrancisHaibin joined11:55
→ Juo joined12:01
← leo2007 left12:02
→ hvoigt joined12:04
→ brodo joined12:07
← saschpe left12:09
← Yuuhi left12:10
→ theoretical joined12:11
→ jfkw joined12:11
← MikhX left12:14
← f0i left12:16
← genbattle left12:17
→ sion2k joined12:21
→ gebbione joined12:24
gebbione hi all, beginners question ... i have a bare repo somewhere and i want to fetch it so that i can work on the proj locally12:25
do i need to create the repo locally first and then fetch from the remote bare repo?12:25
or am i supposed to just fetch it and git will recreate the local copy12:26
→ deuterium joined12:27
→ rgr joined12:29
← FrancisHaibin left12:29
daethorian gebbione: Just clone it and git will create it for you.12:30
← brodo left12:33
→ reliability joined12:33
reliability hi12:33
gebbione daethorian thanks12:33
→ brodo joined12:34
← serverhorror_ left12:35
← urkud left12:37
→ f0i joined12:37
← tango_ left12:39
→ tango_ joined12:40
→ namemon joined12:41
← hyperair left12:41
→ hyperair joined12:42
Teo`` → Teo`12:43
← sraue left12:43
→ urkud joined12:44
→ jjuran joined12:46
→ _Qman joined12:49
→ serverhorror_ joined12:54
→ berserkr1 joined12:55
← aminpy left12:56
← LongBeach left12:56
Aleric I just did a commit, but I wanted to do an amend... how can I correct that? I also have local changes in my working directory that have to be preserved but which weren't staged.12:56
← berserkr left12:57
Aleric What I'd like it to just undo the last commit, and then commit --amend it .12:57
rudi_s Aleric: git reset HEAD~ and then git add ... git commit --amend12:57
Aleric I have to do the add again?12:57
Can I reset it so it ends up staged?12:58
rudi_s Aleric: I think so - if you have uncommitted changes in your repository do a git stash first.12:58
→ leo2007 joined12:58
rudi_s Ah, I think it's git reset --soft - but I'm not sure.12:59
Aleric What about, ... yeah12:59
rudi_s Yeah, it's --soft.12:59
berserkr1 → berserkr12:59
rudi_s Man page mentions it.12:59
Aleric So, git reset --soft HEAD~12:59
?12:59
rudi_s Yes.12:59
lolfrenz that's ironic12:59
fucking up when trying to fix previous fuckup12:59
Aleric It's not like that.. I thought I already pushed the previous commit - but then after my intentional commit I found out it wasn't, so I changed my mind.13:00
→ SimonP86 joined13:02
→ rolfb joined13:02
← _iron left13:03
→ whaley joined13:03
← berserkr left13:04
→ berserkr joined13:04
← brodo left13:04
→ brodo joined13:05
← maleknet left13:05
→ RichardBronosky_ joined13:07
→ sraue joined13:07
← RichardBronosky left13:07
→ Yuuhi joined13:11
← sion2k left13:12
← Dreamer3 left13:12
← urkud left13:14
← bidi left13:14
← berserkr left13:14
→ Yuuhi` joined13:16
→ Chillance joined13:17
→ berserkr joined13:19
→ oriba joined13:20
→ crakdmirror joined13:20
← brodo left13:20
← Yuuhi left13:20
← hvoigt left13:21
← oriba left13:21
→ __name__ joined13:22
← __name__ left13:22
→ __name__ joined13:22
→ berserkr1 joined13:22
← berserkr1 left13:22
→ khmarbaise joined13:26
← verto__ left13:33
← namemon left13:36
← seivan_ left13:37
→ hubertchang joined13:38
→ barf00 joined13:45
ChanServ set mode: +v13:46
barf00 hi, how can I revert something that has been committed to master but not yet pushed?13:46
→ sion2k joined13:48
mastro barf00, if it is the last commit and you are on master you can change it and to a commit --amend13:48
canton7 barf00, you've got two options. You can either remove the thing permanently from history, or create a commit that undoes the changes you want to undo13:49
mastro if you want to get rid of the last commit (all) you can reset: git reset --hard HEAD~13:49
→ Saur joined13:49
mastro if it is a commit back in the log you can use git rebase interactive13:49
canton7 if you want to create a commit that undoes the changes, look up git-revert13:49
→ workmad3 joined13:51
→ thm joined13:52
barf00 thanks and bye :)13:52
← barf00 left13:52
→ Adaptee joined13:58
← radarek left14:01
→ aminpy joined14:02
→ __sorin__ joined14:02
← workmad3 left14:04
← freshmilk left14:05
→ ianawilson joined14:05
← ianawilson left14:05
zz_rickbradley → rickbradley14:07
← hubertchang left14:08
→ tyteen4a03 joined14:11
← rickbradley left14:11
→ n1ck3 joined14:15
← dim_ left14:23
→ dim_ joined14:25
→ the_sniff joined14:26
the_sniff i made a git bare (hub) and a git regular repo's. added a post-update script between them14:27
in one of this "dev" env i have a "Merge branch 'master' of ../hub"14:27
everytime the script ran14:27
in another one i dont14:27
why is this?14:27
← superdmp left14:27
← jjuran left14:29
thiago_home I guess the other one did a fast-forward14:29
→ gusnan joined14:32
→ seivan joined14:32
→ jshanab_wcw joined14:32
jshanab_wcw I have msysgit installed on my win7 64bit machine and am in the gitbash shell. I can't run a script because it does not have the execute flag set, so I try to chmod it and it won't change. No error, just no effect. Any suggestions?14:34
the_sniff thiago_home where is that defined? i do a regular pull. where is the default in it?14:34
← RobertLaptop left14:35
← kermit left14:35
canton7 jshanab_wcw, windows file associations?14:36
right-click, open with, select your preferred intepreter and make sure "always use this..." is checked14:37
thiago_home the_sniff: the default is to do a fast-forward if it can be done14:37
jshanab_wcw I am at the command prompt typing chmod +x ./filename, how do associations come into play?14:37
the_sniff thiago_home i never touched that definition so why is one doing a fast-forward and the other is doing recursive?14:37
← gebbione left14:37
canton7 windows doesn't really support chmod -- file permissions aren't done in the same way at all14:37
thiago_home the_sniff: because one can do a fast-forward and the other can't14:37
← d0k left14:37
canton7 there's no way to make a file as "executable"14:38
you just tell windows what program you'd like to open that sort of file with14:38
and it'll use that program when you next try and open the file14:38
the_sniff thiago_home why wouldnt it if its the exact same files in both sides?14:38
thiago_home the_sniff: probably because one the one that didn't do a fast-forward made the fast-forward possible14:38
the_sniff: are you starting from the same commit?14:38
the_sniff: merging the same commit?14:38
the_sniff yes its a live dev env14:38
yes14:38
always14:38
→ Corren joined14:38
thiago_home are you sure?14:39
let me repeat: are you merging the same commit in both cases, starting from the same commit?14:39
jshanab_wcw canton7 other files have the execute bit set and work fine :-) Newer windblows file systems have equivilent, does gitbash not yet support this? How do they instructions for chmoding the ssh file get followed?14:39
the_sniff you mean merging the sdame comming in both all 4 repos? then no14:40
canton7 jshanab_wcw, not come across that ever working, so afraid I can't help in that case14:40
thiago_home the_sniff: then that's the difference14:40
the_sniff: if a fast-forward is possible, a fast-forward will be made.14:40
jshanab_wcw Too bad. I was really enjoying having a decent shell in windows for a change ;-)14:41
the_sniff thiago_home why would a fast-forward not be possible14:42
→ smolyn joined14:42
the_sniff is the merge is always for the same versions?14:42
jshanab_wcw canton7. But, duh, it is a cmd script, LOL so hwo do I run that from gitbash LOL. (more coffee is indicated)14:42
the_sniff i mean when i do A -> B. B was the same as A was before14:42
same goes for the other case, i do X -> Y. Y was the same as X was before14:43
thiago_home the_sniff: a fast-forward is not possible if the current commit is not an ancestor commit of what you're trying to merge14:43
→ joel_falcou joined14:43
canton7 jshanab_wcw, if you've got testy.cmd, just type "testy.cmd" and hit enter ;014:43
the_sniff thiago_home so if i dont start from a "git clone" then a fast-forward is not possible?14:44
jshanab_wcw canton7. That won't work, bash doesnot know what to make, I have to switch to a command prompt. I was in GIT-BASH14:44
thiago_home the_sniff: it depends on the commit you're in, not how you got there14:45
the_sniff well the thing is it "should" be an 'anscetral' commit14:45
canton7 jshanab_wcw, I just tried running a .cmd file from git bash, and it was reasonable happy..... although, thinking about it, there won't have been any batchfile-specific stuff in there14:45
the_sniff in the above example14:45
B was git init'd and git commit -a'd14:45
and A pulled from it14:46
then A was changes and B started pulling from it14:46
nothig else14:46
just A changed14:46
jshanab_wcw canton7. right, I do not know what I was thinking. It would be an interesting tol for the bash shell, to run .bat commands.14:46
canton7 jshanab_wcw, true :)14:47
← glcrazy left14:48
jshanab_wcw I am using a framework that uses cmake to generate build files on linux,mac and windows, but the prep script is in each OS's shell language.14:48
thiago_home the_sniff: that's why14:48
the_sniff im missing the point14:49
→ glcrazy joined14:49
thiago_home the_sniff: a fast-forward is possible if the current commit is an ancestor of the commit being merged14:49
the_sniff: you've just described that case14:49
when A pulled B, B made a recursive merge14:50
now that repository is in commit C14:50
A is an ancestor of C14:50
so when you pull it into A, it's a fast-forward14:50
the_sniff so is there a way to flip it around>14:51
?14:51
← Err404NotFound left14:51
thiago_home yes, don't do what you're doing14:51
invert what you're doing14:51
the_sniff but A is a bare repo14:51
← jshanab_wcw left14:51
the_sniff how would i insert the files locally into it?14:52
thiago_home then use a non-bare14:52
the_sniff B is non bare14:52
A is bare so that everyone could commit into it14:52
thiago_home you cannot commit in a bare14:52
you can push to it14:52
the_sniff yeah i meant push my bad14:52
thiago_home why do you want to invert?14:52
the_sniff this is the type of git im using http://joemaller.com/990/a-web-focused-git-workflow/14:53
← mtx_init left14:53
the_sniff i think this would explain better then me14:53
→ mtx_init joined14:53
thiago_home irrelevant14:54
the_sniff why? i want to invert because A is the 'source' B is just the way apache see's it14:55
thiago_home why?14:55
why does it matter?14:55
the_sniff i get a lot of logs for 'git merge of X...' and i dont want them14:56
→ brian_g joined14:56
→ radarek joined14:57
thiago_home then use rebase14:57
the_sniff after i push i should do a rebase on A to make it the same as B?14:57
thiago_home don't pull; rebase14:58
or pull --rebase14:59
the_sniff do a pull --rebase in B everytime i want to get from A?15:00
→ Fuwex joined15:01
thiago_home yes15:01
the_sniff ok thank you i will give this a try15:01
← Prism left15:04
canton7 the_sniff, you may like the config key branch.autosetuprebase15:06
← awilkins left15:07
the_sniff thiago_home lol pull gives me error: unknown option `rebase' :S15:07
thiago_home which git version are you using?15:08
the_sniff git version 1.7.2.515:08
thiago_home git pull --rebase should work in that version15:08
it does a fetch followed by a rebase15:09
the_sniff http://pastebin.com/qrhBwTnP15:10
wereHamster the_sniff: git pull --rebase hub master15:10
the_sniff haha15:11
ok thanks15:11
btw should git pull --ff hub master do the same?15:11
→ mdel joined15:11
wereHamster no15:11
thiago_home --ff is only to negate a previous --no-ff15:11
it doesn't make a fast-forward possible where one wasn't before15:11
→ RobertLaptop joined15:12
the_sniff ok15:12
thanks15:12
→ workmad3 joined15:12
← dr_win left15:15
→ d0k joined15:22
← mtx_init left15:23
→ mtx_init joined15:23
→ n3m joined15:25
← n3m left15:26
→ kayue joined15:30
← ph^ left15:31
← the_sniff left15:31
→ ph^ joined15:33
→ coldboot|home joined15:34
← kayue left15:34
→ kukks joined15:36
→ psoo joined15:36
→ iori joined15:37
← jfelchner left15:37
← jophish left15:38
→ jophish_ joined15:38
← Adaptee left15:38
→ pantsman joined15:39
← pantsman left15:39
→ pantsman joined15:39
← __name__ left15:39
→ dannas joined15:40
← xiangfu left15:41
→ jorfan joined15:42
→ kermit joined15:43
← psoo left15:45
ChanServ set mode: +v15:46
jorfan Hello everyone !15:46
← dannas left15:46
← Corren left15:47
CareBear\ yasso15:47
→ Buglouse joined15:48
← selckin left15:49
← jablko left15:51
← flaguy48 left15:52
← f0i left15:53
→ JohnBat26 joined15:55
← workmad3 left15:55
jorfan guys can some1 help me with 1 modification...15:56
canton7 can try...15:56
jorfan im trying to figure out how to make a custom build of the GAIA MANAGER FOR PS315:57
i just want to remove a button thats all :(15:57
→ davidcoallier joined15:57
→ dannas joined15:57
canton7 jorfan, sounds like you might have more luck in a more related channel15:58
→ dreiss joined15:58
jorfan :(15:58
← jophish_ left15:58
CareBear\ jorfan : how on earth did you end up in #git ? :)15:59
jorfan : sorry, we have no idea about that stuff15:59
→ aquabu joined15:59
jorfan well, i was reading on the authors notices.. and in some part it said that he uses #git to update some things.16:00
so i downloadeed git. opened his GPL project16:00
and stuck on here16:00
→ Corren joined16:00
→ workmad3 joined16:01
CareBear\ git is the source code management tool16:01
canton7 jorfan, git is the system he uses to version his files. it has nothing to do with the content of those files16:01
jorfan sorry guys. now i got in16:02
canton7 it's a bit like asking microsoft, because he happened to write his project in notepad16:02
:P16:02
jorfan yeap. sorry obout it16:02
:P16:02
← jorfan left16:02
canton7 no worries :)16:02
← dreiss left16:03
← paeva left16:03
← serverhorror_ left16:03
← Pilif12p left16:03
CareBear\ canton7 : so you are like, microsoft?16:04
CareBear\ never knew16:04
canton7 I was stuck for a tool which he both knows, and knows the maker of :P16:04
canton7 is quite glad this isn't #microsoft16:04
→ whyking joined16:05
→ flaguy48 joined16:05
CareBear\ canton7 : would you consider it a bug in git if git am gets confused by arbitrary commit messages?16:06
← rgr left16:06
→ ISF joined16:07
canton7 random question? depends how arbitrary16:07
← seivan left16:08
← workmad3 left16:08
→ urkud joined16:08
→ madewokherd joined16:09
CareBear\ canton7 : should it?16:09
← Corren left16:09
→ rendar joined16:10
→ dv310p3r joined16:10
CareBear\ canton7 : the problem I had was with commit messages containing what looks like a patch, which trips git am up16:11
canton7 if the commit message screws with the markers that git am uses, then it's excusable i would have though? Not used git am / format-patch much though16:11
CareBear\ what are those markers?16:12
canton7 yeah, can see how that would trip it up. The man page says that "---\n", "diff -" and "Index: " are markers of the start of the patch16:12
CareBear\ is it the --- ?16:12
→ f0i joined16:12
CareBear\ ah, either of the three16:13
canton7 i would have thought that git format-patch would escape them, but that's a complete assumption16:13
CareBear\ I would have expected the same16:13
or.. not neccessarily escaped, because that is ugly16:13
→ Eaven joined16:13
canton7 at least a warning16:13
CareBear\ but I would expect that output from format-patch would always work with am16:13
canton7 goes in search of answers16:13
→ workmad3 joined16:14
CareBear\ awesome :)16:14
thank you so much!16:14
→ freshmilk joined16:14
← robertdavid left16:15
← apassi left16:15
→ apassi joined16:17
→ Gunni joined16:19
→ seivan joined16:19
← empity left16:19
← f0i left16:20
→ SimonP86_ joined16:20
← Ramblurr left16:20
← leo2007 left16:20
→ Ramblurr joined16:21
→ SimonP86__ joined16:21
← SimonP86 left16:23
← __sorin__ left16:23
→ f0i joined16:24
→ Corren joined16:24
→ VaL_ joined16:25
← SimonP86_ left16:25
← angelsl left16:26
← aminpy left16:28
canton7 CareBear\, hmmm. I can confirm that git format-patch does indeed leave "---" intact, with no warning. There doesn't seem to be any vaguely related flags, and I'm not familiar enough with the git source to find where emails are put together16:28
← Corren left16:29
CareBear\ canton7 : mm16:29
git grep Subject: ? :)16:29
← whaley left16:32
→ OOPMan joined16:34
canton7 CareBear\, well it looks like it's builtin/log.c, somewhere around line 80016:35
→ nevyn^_ joined16:35
→ noonoooo joined16:35
← jfkw left16:40
→ mando joined16:40
← nevyn^_ left16:43
← madewokherd left16:44
CareBear\ canton7 : question is if the issue is best handled at input or output16:44
canton7 : to handle on input requires either seeking or buffering the full patch16:44
this might be new16:44
and is undesirable of course16:45
canton7 CareBear\, well it's more complex... can also refuse to commit messages which might confused format-patch, but i don't like this16:45
→ leo2007 joined16:45
canton7 i'd be tempted to deal with it in format-patch. the patch emails follow a set format, and format-patch allows you to violate this with no warnings16:46
→ aminpy joined16:48
→ brodo joined16:50
jmcantrell isn't there a way to echo the root directory of a repo or the directory that '.git' sits in?16:50
← seivan left16:50
← workmad3 left16:51
CareBear\ canton7 : OTOH patches can come from other sources too, right?16:51
canton7 jmcantrell, I've got a ruby snippet -- might be useful in some way16:52
begin16:52
break if Dir.entries(".").include?(".git")16:52
cwd = Dir.pwd16:52
Dir.chdir("..")16:52
@path_to_root << "../"16:52
end while cwd != Dir.pwd16:52
← babilen left16:52
drizzd jmcantrell: git rev-parse --git-dir16:52
CareBear\ canton7 : at the very least it's a nice side-effect that git am accepts an email with an inline patch, and just works<tm>16:52
← urkud left16:52
drizzd or git rev-parse --show-cdup16:52
jmcantrell drizzd: very nice. thank you16:53
← kpreid left16:54
← VaL_ left16:54
→ urkud joined16:54
canton7 CareBear\, am I right in thinking that git-format-patch is the only tool that creates an email? and that the issue only appears when parsing patches in emails?16:55
CareBear\ canton7 : lots of tools can create emails with inline patches16:56
canton7 : but within git there's also rebase, which might or might not use format-patch as-is16:57
← radarek left16:57
→ Dreamer3 joined16:57
← Dreamer3 left16:57
CareBear\ canton7 : and there's also send-email16:57
→ nevyn^_ joined16:58
canton7 CareBear\, I thought git-send-email took patches generated by format patch, and just sent them?16:58
← gusnan left16:59
CareBear\ canton7 : send-email does not need patches as input16:59
→ workmad3 joined16:59
→ Dreamer3 joined16:59
← Juo left17:00
→ Juo joined17:00
→ serverhorror_ joined17:01
→ robertdavid joined17:03
→ babilen joined17:03
→ mishok13 joined17:03
← workmad3 left17:04
→ radarek joined17:04
→ codeshah joined17:05
← zhengyang left17:07
canton7 CareBear\, it looks like send-email uses format-patch internally17:07
→ zhengyang joined17:07
canton7 git-send-email.perl line 53717:07
← madsy left17:08
CareBear\ ok17:08
makes sense17:09
I just was not sure17:09
→ madsy joined17:09
← madsy left17:09
→ madsy joined17:09
→ unhygienix joined17:10
→ dr_win joined17:10
→ gusnan joined17:10
← gusnan left17:13
← phaedrix left17:13
canton7 CareBear\, not sure what rebase uses (still need to learn the ins and outs of the plumbling), but I'd be surprised if it used the format-patch syntax17:13
→ phaedrix joined17:13
wereHamster jmcantrell: git rev-parse --show-cdup17:14
→ jfkw joined17:14
← phaedrix left17:14
← madsy left17:15
→ zhengyang_ joined17:15
→ madsy joined17:15
← madsy left17:15
→ madsy joined17:15
← Saur left17:15
→ phaedrix joined17:16
← shiba_yu36 left17:16
← Dave^| left17:16
← KKA left17:17
← zhengyang left17:17
zhengyang_ → zhengyang17:17
← sion2k left17:18
→ KKA joined17:18
→ Dave^| joined17:19
← nevyn^_ left17:19
→ ereslibre joined17:20
← zhengyang left17:20
← radarek left17:20
← davidcoallier left17:22
→ cesario joined17:24
→ Phabtar joined17:26
ChanServ set mode: +v17:27
Phabtar Hi!17:27
→ Corren joined17:27
← babilen left17:28
→ davidcoallier joined17:29
→ lut4rp joined17:29
Phabtar I'm new to git, can anyone help17:29
← lut4rp left17:29
→ lut4rp joined17:29
← mando left17:30
wereHamster Phabtar: !doc17:30
jast Phabtar: [automatic message] a list of useful documentation of all kinds is here: http://git-scm.com/documentation -- or try the wiki at http://git.wiki.kernel.org/17:30
Phabtar I'm using msysgit in windows17:31
→ cthuluh joined17:31
Phabtar and also tried tortoisegit17:31
but couldn't clone a repository in gitorious.org17:31
it gets stuck halfway17:31
Also, I have a working copy of the corresponding SVN repository of the same project17:32
← dr_win left17:33
→ metcalfc joined17:34
→ engla_ joined17:34
→ ymasory joined17:36
← robertdavid left17:36
Phabtar But I want to make a shared branch of the project on gitorious and have a git working copy,17:36
← engla left17:37
Phabtar I'm afraid there is no solution other than cloning the git repo, is there jast or wereHamster?17:37
← JohnBat26 left17:39
← davidcoallier left17:39
← DrNick left17:40
← Rotham left17:42
→ Rotham joined17:42
→ babilen joined17:43
→ whaley joined17:47
→ workmad3 joined17:47
→ drev1 joined17:47
→ d03boy joined17:48
← brodo left17:48
d03boy I initialized a repo but I want to delete it and start over. How do I do that (I haven't committed anything to server yet)17:48
→ radarek joined17:49
→ brodo joined17:49
cbreak just delete the .git folder17:49
all of git is inside that17:49
d03boy ok. that's what I figured17:50
← envi left17:51
← lolfrenz left17:51
→ ThomasLocke joined17:52
→ jophish joined17:52
→ [intra]lanman joined17:53
← noonoooo left17:58
← berserkr left17:58
→ robertdavid joined17:59
← Freakshow left18:01
← brodo left18:02
→ brodo joined18:04
← workmad3 left18:05
← jophish left18:06
→ rgr joined18:08
→ workmad3 joined18:10
← n1ck3 left18:15
← khmarbaise left18:17
← thews left18:17
← mastertheknife left18:18
→ mastertheknife joined18:19
→ khmarbaise joined18:21
← metcalfc left18:21
→ qfyre joined18:22
← coldboot|home left18:23
→ mastorna joined18:24
← legumbre_ left18:24
→ metcalfc joined18:25
→ paullik joined18:25
← qfyre left18:25
→ qfyre joined18:26
ChanServ set mode: +v18:26
mastorna hey, when installing git, do you the dmg installs it in /usr/local/git and not /usr/local/bin - is that correct?18:26
← serverhorror_ left18:27
→ serverhorror_ joined18:27
→ kipras joined18:28
paullik Hi. I'm wondering what should i do when i have like 10 commits and no push and some one in my team pushed while i was commited, i cannot push...what shoud i do then?18:28
canton7 paullik, clarify? "you cannot push" - what do you try, and what does git say?18:29
paullik canton7, its a hypothetical situation, like: what happens if...18:33
→ ciupicri joined18:33
ciupicri how can I see the patch/diff for a single file from a single commit and then apply that to another file?18:33
→ DrNick joined18:34
canton7 you can either merge their commits into your branch, or rebase your commits on top of theirs18:34
ciupicri canton7, the problem is that they're different files18:34
canton7 sorry, was replying to paullik18:34
paullik ciupicri, yeah..18:34
ciupicri canton7, a sample file was modified and I want to update my work which of course is based on that18:35
ok18:35
← glcrazy left18:36
paullik canton7, yeah but if i merge their commits won't they change my modifications?18:37
← airborn left18:37
canton7 paullik, git is pretty smart here. if you've both changed the same line, that'll cause a merge conflict, which you'll have to resolve by hand. However, if you haven't stepped on each other's toes (haven't edited the same lines), git can merge them automagically18:38
that's a bit simplified, but it's basically it...18:38
paullik canton7, ok thanks18:39
canton7 paullik, I believe you'll get the exact same conflicts if you rebase, so it's up to how you want your history to look18:40
→ airborn joined18:40
paullik canton7, pff preety complicated...18:41
→ mikepack joined18:41
← RobertLaptop left18:41
← mikepack left18:42
← metcalfc left18:42
canton7 paullik, if two people are editing the same files, and both change the same line in different ways, you've got a situation where someone's got to decide which of those two people to listen to.18:42
→ sskceg joined18:42
canton7 in instances where the changes are compatible, (you've edited separate functions, for example), git is smart, and can merge both sets of changes18:43
you can't really avoid that18:43
← sskceg left18:43
→ glcrazy joined18:43
→ flavius joined18:44
paullik canton7, ok, thanks18:44
flavius Hi. If I have a remote, how do I create a branch attached to it so I can then do the diff?18:44
→ sskceg joined18:44
canton7 paullik, put it this way - it's better than locking files ;)18:45
→ starlays joined18:45
← workmad3 left18:45
canton7 flavius, what diff? :P18:45
paullik canton7, i see18:45
← aminpy left18:45
← robertdavid left18:45
flavius canton7: between the two branches :D18:45
canton7 flavius, so you want to diff a local branch against a remote one?18:46
← fr0stbyte left18:47
flavius canton7: I have a forked project from mine and my project, and I want to pick commits from the fork on an individual basis and merge them in18:47
I've set up the fork as a remote18:48
fetched it18:48
canton7 flavius, you can diff just by git diff local_branch remote/remote_branch18:48
if you're picking commits, look at git-cherry-pick18:48
use gitk / git log to find the sha1's of the commits you want to pick18:48
you don't need to create a local branch of the project you forked, though of course you could18:49
flavius and if I wanted, how would I do that?18:50
canton7 git checkout -t local_name remote/remote_name if you wanted it to track the remote branch18:51
and git checkout -b local_name remote/remote_name if you didn't18:51
flavius thanks18:51
canton7 np18:51
→ Heimidal joined18:51
← airborn left18:52
← xxoxx left18:53
→ legumbre joined18:54
→ AAA_awright_ joined18:55
→ hvgotcodes joined18:55
← ch077179 left18:56
→ Phabtar_ joined18:57
← Phabtar left18:57
→ xxoxx joined18:57
← AAA_awright left18:58
← vdv left18:59
→ AAA_awright joined18:59
← Phabtar_ left19:00
→ jatt joined19:01
→ serverhorror__ joined19:01
→ DTrejo joined19:02
← AAA_awright_ left19:02
← Corren left19:02
DTrejo hello, would anyone be able to help me figure out how a rebase, such as in this case: https://github.com/joyent/node/pull/102419:02
I don't have that commit on it's own branch, which makes things more complicated right?19:03
FauxFaux DTrejo: Single commit? Why not cherry-pick?19:04
DTrejo (I've never done a rebase or a cherry pick)19:04
← serverhorror_ left19:04
DTrejo FauxFaux: looking up cherry pick now19:05
→ airborn joined19:06
DTrejo FauxFaux: I'm really tempted to just save my changes somewhere, delete my fork and re-fork, check out v0.4, and then commit19:06
FauxFaux As a non-github user, it's really hard for me to work out wtf their interface means; someone else can probably tell you what you mean.19:07
→ davidcoallier joined19:07
DTrejo FauxFaux: basically: delete my whole repo, clone again, checkout the correct tag, and then commit again19:08
→ theoros_ joined19:08
DTrejo *correct branch19:08
← theoros_ left19:08
→ jophish joined19:10
→ RobertLaptop joined19:12
→ archis joined19:13
← sskceg left19:14
← archis left19:14
→ paeva joined19:15
← jophish left19:15
→ arand joined19:16
← Swimming_bird left19:16
arand I'm trying to add a git-svn clone as a new branch in an already existing git repository, how would I do this?19:17
← iAsterisk left19:17
→ fr0stbyte joined19:18
← dannas left19:20
← DTrejo left19:23
→ DTrejo joined19:23
DTrejo thank you for the advice FauxFaux :) have a good week19:23
← DTrejo left19:23
→ gunmuc joined19:23
← paeva left19:25
albel727 arand: try "git fetch <path to svn repo> <branchinsvnrepo>:<branchtocreateinexistingrepo>"19:25
arand albel727: Cheers19:26
→ dannas joined19:27
← dannas left19:28
→ dannas joined19:28
→ caseymcg joined19:28
← glcrazy left19:29
→ glcrazy joined19:30
albel727 arand: that said, if you're going to continuously update the new branch from git-svn repo, you'd be better off adding a git-remote and setting up a tracking branch.19:30
arand albel727: If I use just git I get Unable to find remote helper for 'svn' and If it use with git svn I get http://paste.debian.net/116397/ and nothing more happens19:30
← paullik left19:31
← airborn left19:31
arand albel727: How would I do that?19:31
Can I use git remote add with svn?19:31
albel727 wait a sec. I thought you have an already cloned git-svn repo.19:31
→ airborn joined19:32
arand Well I have an stand-alone clone via git svn clone, but I'd like to add this to an already existing git repo, is there any way to link it directly to the remote svn without going through my local clone?19:33
→ khmarbaise_ joined19:33
albel727 arand: well, you actually can pull your existing git repo into that git-svn clone, so you will have no intermediate repos.19:34
← electrogeek left19:34
← lamont left19:34
← drev1 left19:35
← khmarbaise left19:35
khmarbaise_ → khmarbaise19:35
→ drev1 joined19:35
← phantomcircuit left19:35
← Tianon left19:35
albel727 arand: that will require doing it all the other way round. i.e. "git remote add mylocalrepo <pathtoyourexistinggitrepo>" "git fetch mylocalrepo"19:36
after that you'll be able to "git checkout -t mylocalrepo/somebranch", and work on that.19:36
← khmarbaise left19:36
→ ch077179 joined19:36
← Neo left19:37
albel727 well, to avoid confusion you do that checkout -t only once, to create a local branch.19:37
you'll also need to change the url for mylocalrepo remote to point to your original repo at github or wherever.19:37
→ drev1_ joined19:37
← SJS left19:38
← Wizek_ left19:38
albel727 and for what it's worth, you can name your "mylocalrepo" remote "origin" instead.19:38
← drev1_ left19:38
→ fubada joined19:38
→ drev1_ joined19:38
arand Thanks!, I'm on it.19:38
albel727 so that it becomes more familiar.19:38
k19:39
→ Swimming_bird joined19:39
← jatt left19:39
← starlays left19:39
→ starlays joined19:39
starlays hello to all19:39
→ Neo joined19:39
fubada hi can someone please help im having trouble adding a new user. On my gitolite sevrer I "git clone /var/lib/gitolite/repositories/gitolite-admin.pub", I add my new user.pub to the keydir folder, try to commit but get an erro: http://pastie.org/187883819:40
→ ki__ joined19:40
starlays i need some help with this problem, i can`t make a pull request, i get this error: You asked to pull from the remote 'dotophp', but did not specify19:40
a branch. Because this is not the default configured remote19:40
for your current branch, you must specify a branch on the command line.19:40
→ davidrussell joined19:40
→ btf joined19:40
← drev1_ left19:40
fubada please help :P19:40
← lut4rp left19:41
albel727 sitaram: ^19:41
→ SJS joined19:41
← drev1 left19:41
starlays any help?19:41
albel727 starlays: what's pull request?19:41
CareBear\ no help19:41
albel727 just git pull?19:41
← davidcoallier left19:42
starlays yes19:42
thiago_home starlays: the error message you got is your help19:42
fubada albel727: you were really helpful the last time, any chance you can explain to me what im doing wrong19:42
starlays albel727: only git pull19:42
thiago_home starlays: you ran "git pull"19:42
← RaceCondition left19:42
thiago_home starlays: you need to tell it what to pull19:42
starlays thiago_home: don`t get it19:42
thiago_home: how?19:42
thiago_home starlays: by passing the extra arguments19:42
starlays: the remote name and the branch name19:43
CareBear\ fubada : that looks like a gitolite problem, nothing related to git19:43
albel727 fubada: I'm afraid, that's beyond my knowledge of gitolite. it looks like some internal error to me. I'm afraid you'll have to wait for sitaram.19:43
starlays thiago_home: how can i do that?19:43
thiago_home starlays: type more19:43
CareBear\ starlays : type. see man git pull19:43
jast [automatic message] the 'git-pull' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-pull.html19:43
← ciupicri left19:44
sitaram fubada: gitolite does not like being cloned via something other than ssh. If you really want to administer the gitolite-admin repo directly from the server (judging by your clone URL -- a local file path), it can be done but I would not advise it19:44
→ cirwin joined19:45
starlays 10x guys19:46
fubada sitearm I dont have my other pc with me that has the ssh key19:46
and i dont have any other means to add another key19:46
← tore_ left19:46
fubada what are my options19:46
→ giallu joined19:46
fubada i mean i only have one key in my gitolite-admin repo, and that key is not with me19:46
albel727 starlays: for future reference, if you're going to push back to that repo, use push's --set-upstream option. that will eliminate the need to specify branch in most cases.19:47
sitaram fubada: what version of gitolite is this?19:47
fubada ii gitolite 1.5.4-1ubuntu0.119:47
→ lut4rp joined19:48
→ balau joined19:48
→ jzellman joined19:48
sitaram fubada: that's far too old for me to remember how; lots of things have changed and you'll also be missing some of the convenience programs that gitolite now comes with19:49
fubada shit19:49
sitaram fubada: your best bet, if you *really*, *truly* understand ssh, is to do the following:19:49
← starlays left19:50
fubada i am open to upgrading19:51
do you have a ppa?19:51
sitaram generate a local keypair (if you dont already have one) using ssh-keygen, then manually add it to the ~/.ssh/authorized_keys file, as a new line, and then manually prefix it with the same long sequence (command="...", and those options) etc19:51
fubada: that assumes you have a decent editor, and you know how to not make it wrap lines etc :)19:51
fubada: what's a PPA?19:51
fubada launchpad ppa, ubuntu packages19:52
← jzellman left19:52
sitaram fubada: sorry; I don't do packaging. I'm what is called "upstream" I believe ;-)19:52
fubada i can add-apt-repository ppa:gitolite-admin and apt-get install your latest package19:52
sitaram Rhonda: ping19:52
fubada if you had that19:52
→ davidcoallier joined19:52
← airborn left19:53
fubada ah ok i see what youre saying with ssh keys19:53
ill do that19:53
→ airborn joined19:53
fubada that way i can add my key over ssh after19:53
← ki__ left19:53
sitaram fubada: there are probably saner methods but I would have to remember lots of forgotten details now. A *lot* has changed in the past year19:54
fubada im gonna give it a go now, thank you so much19:54
← davidrussell left19:54
sitaram fubada: make a backup of the file first, just in case (authkeys file)19:54
← OOPMan left19:55
← ISF left19:55
sitaram fubada: and it's 1:25am here, so I'm only going to turn up again in about 5 hours... sorry!19:55
need sleep19:55
← unhygienix left19:55
→ lamont joined19:55
sitaram good luck19:55
→ hungrygruffalo joined19:55
← hungrygruffalo left19:56
← gunmuc left19:56
← engla_ left19:57
→ engla_ joined19:57
→ hungrygruffalo joined19:57
→ verto joined19:59
← aesy left19:59
→ aesy joined19:59
← leo2007 left20:00
← mastorna left20:00
→ serverhorror_ joined20:00
→ drev1 joined20:01
→ phantomcircuit joined20:01
→ Tianon joined20:02
← Tianon left20:02
→ Tianon joined20:02
← serverhorror__ left20:02
→ Pilif12p joined20:03
→ xanonus joined20:03
→ electrogeek joined20:03
← electrogeek left20:03
→ electrogeek joined20:03
← dannas left20:03
fubada sitaram: i did all this. now im trying to remote clone it, but im gettinmg "object corrupt"20:04
← Juo left20:04
fubada and an md5 string20:04
any idea why?20:04
thiago_home probably a SHA-1 string20:04
fubada ?20:04
chrisf|work something's deeply wrong if a sha1-based system gives you md5 strings ;)20:04
fubada yea20:04
thats what i meant20:04
→ paeva joined20:05
← lut4rp left20:06
→ Phabtar joined20:06
← btf left20:06
← flavius left20:06
→ jfelchner joined20:06
→ btf joined20:07
← paeva left20:07
→ bsajsb joined20:08
→ OOPMan joined20:08
→ paeva joined20:08
fubada Cloning into gitolite-admin...20:09
fatal: object 8b2f014904d4fe0099ac9ea8875cdc2273f03e38 is corrupted20:09
fatal: The remote end hung up unexpectedly20:09
like that....20:09
→ paeva_ joined20:09
← ph^ left20:10
→ ph^ joined20:12
← paeva left20:13
albel727 now that's a real problem. you'll have to run git fsck and do all other stuff to repair it. do you per-chance still have that clone of admin repo, that you did locally?20:14
fubada you know i added my second key to the wrong place20:14
i should of added it to /var/lib/gitolite/.ssh20:14
not /home/myuser/sh20:14
.ssh20:14
← Phabtar left20:15
← airborn left20:15
albel727 private key should go to myuser.20:15
public key should go to gitolite/.ssh20:15
fubada right20:15
i didnt add the pubkey to the right place20:15
thanks20:15
→ airborn joined20:16
← guardian left20:16
→ vdv joined20:16
→ jzellman joined20:17
albel727 well, the cloning error message doesn't look like it was the problem with keys. let's hope it helps...20:17
→ muneeb joined20:18
→ RichiH joined20:19
← hungrygruffalo left20:19
→ hungrygruffalo joined20:19
RichiH is there a built-in way to fetch and then only merge to the newest signed & trusted tag?20:19
← airborn left20:19
→ airborn joined20:20
→ robertdavid joined20:20
albel727 sorry, I'm not sure I'm getting your problem. simple fetch + merge do not work for you?20:21
← hungrygruffalo left20:21
albel727 the tag, even signed, should behave just like every other tag, if I'm not much mistaken.20:22
→ unhygienix joined20:23
← simplechat left20:23
fubada albel727: i added the new key and commit the change20:24
→ gunmuc joined20:24
→ blorbx joined20:24
← paeva_ left20:24
fubada but now i get a corrupt messgae when i git clone gitolite@server/myrepo20:24
does the user.pub need to match the username on my workstation?20:25
← unhygienix left20:25
← balau left20:25
← mtx_init left20:25
→ mtx_init joined20:26
← d0k left20:26
→ guardian joined20:26
→ unhygienix joined20:26
← unhygienix left20:26
→ unhygienix joined20:27
albel727 fubada: what is "server", exactly? is that just a regular url, that you've snipped for your convenience, or is it a ssh virtual host, that gitolite have created for you during setup? or what?20:27
← unhygienix left20:27
albel727 fubada: and by the way, is that server/myrepo, or server:myrepo?20:28
← bindaas left20:29
fubada server:repo20:30
albel727 and my first question?20:30
← sechrist left20:30
fubada snipped20:30
for convinience20:30
← btf left20:30
← dwmw2_gone_ left20:31
albel727 ok. is that "workstation" the same user, where you've set up your gitolite from?20:31
or rather, what does ssh -v gitolite@server:myrepo say?20:31
fubada one sec im so confuse dheh20:31
appreciate the help20:31
:(20:31
→ RaceCondition joined20:32
← jfelchner left20:33
→ hungrygruffalo joined20:33
← Dreadnaught left20:35
→ jfelchner joined20:35
→ sechrist joined20:36
→ dr_ick joined20:38
→ jjuran joined20:39
← freshmilk left20:40
← eletuchy left20:41
← jstemmer left20:44
→ ISF joined20:44
fubada albel727: hi can you please help me figure this out20:48
http://pastie.org/private/rvmmaybksqcclt3olctv0g20:48
i added ateda.pub, pushed it20:48
i see the entry for it in /var/lib/gitolite/.ssh/authorized_keys20:48
seems like i can log in without a pw20:49
← bsajsb left20:49
fubada but i cant seem to checkout as that user20:49
← airborn left20:49
fubada i can do it no problem with another user20:49
← urkud left20:49
→ urkud_ joined20:49
→ airborn joined20:49
→ Prism joined20:50
→ freshmilk joined20:50
← RobertLaptop left20:50
← urkud_ left20:52
fubada someone ?20:52
http://pastie.org/private/rvmmaybksqcclt3olctv0g20:53
albel727 fubada: may you paste what does ssh [email@hidden.address] and git clone say for that another user?20:53
fubada k20:53
albel727 incidentally, what is the git --version on the server?20:55
and on your client, too.20:55
← SimonNa left20:55
← dr_ick left20:55
fubada albel727: yikes20:56
the other user cant clone that repo either20:56
problem must be repo side20:56
albel727 I see.20:57
fubada i can simply recreate the repo20:57
whats the best way to blow it away and re-add20:57
albel727 hmm, nothing short of commenting out the line from gitolite config, pushing that, going to the server, removing repo directory, uncommenting the line and pushing again, comes to mind.20:58
that would be cleanest of all, ofc. but you can also just create a new repo, add necessary remote and force-push that.21:00
→ jablko joined21:01
← jablko left21:01
→ jablko joined21:01
fubada error: insufficient permission for adding an object to repository database ./objects21:02
sigh21:02
constant problems it seems21:02
RichiH albel727: sorry for the late reply21:02
→ RobertLaptop joined21:02
RichiH what i want to do is that i only merge what's been signed with a trusted key21:03
albel727 fubada: that was from git push in gitolite-admin?21:03
fubada yes21:03
RichiH is managing his config files with git and wants to protect against injected configs21:03
fubada after changing the config to remove the repo im having toruble with21:03
← ThomasLocke left21:03
cbreak you can only merge whole lines of history21:04
you can not take selected commits out and not merge them unless you rewrite history21:04
albel727 fubada: oh well. I'd say you somehow screwed up your ssh config. or that old gitolite is buggy in some respect.21:04
fubada ok21:04
cbreak so either make sure that all who sign only do so if they want to sign the whole line of history21:04
fubada im not fond of gitolite21:04
is there a way to host my own git server and users21:04
without gitolite21:04
cbreak or rewrite history before merging21:04
sure21:04
gitolite only makes it easier21:04
it does NOT make it possible21:04
fubada obviously not :P21:04
cbreak obviously it does :)21:05
without it, you have to make system users21:05
git init --shared the repositories21:05
fubada thats...fine i gues21:05
system users21:05
cbreak create user accounts on the server for the people who you want to push/pull21:05
and put them into the group that you set for the git repo folder21:05
albel727 fubada: https://github.com/sitaramc/gitolite/blob/pu/doc/overkill.mkd21:05
cbreak (well, init needs --bare obviuosly)21:06
without gitolite you will not have all the nice features like branch level permissions, read only access, wildchar repos or similar21:06
but basic pushing/pulling will work21:06
fubada i have 3 users21:06
i think im gonan rid of gitolite21:06
← f0i left21:06
→ sechrist_ joined21:07
cbreak fubada: seems your repo is corroupted21:07
RichiH cbreak: what i want is to merge everything up to the most current signed tag21:08
there not be any branches, so "most current" makes sense in my repos21:08
cbreak RichiH: you could use git describe, it will use the most recent annotated tag as base21:08
you have to check yourself if they are signed and if the signature is valid21:09
and if you trust the signing key21:09
and if you trust the person that posesses the signing key21:09
RichiH ok, i had hoped there was a ready-made solution, but i'll just script it. thanks :)21:09
cbreak I am not aware of one21:09
albel727 and if you trust your trust to the person that possesses the signing key =)21:09
fubada hows gitorious?21:09
cbreak gitorious is a service21:10
it's comparable to github21:10
fubada k21:10
albel727 well, it does have an engine.21:10
which, theoretically, one can install.21:10
← karstensrage left21:10
bremner_ well, you can install it yourself too, but if you find gitolite confuzzling21:10
← reliability left21:10
bremner_ you will probably not like setting up gitorious21:10
← sechrist left21:10
cbreak yeah, gitolite is much simpler and more userfriendly :)21:10
→ karstensrage joined21:11
cbreak almost no dependencies, simple instructions21:11
fubada i think ununtu gitolite is just old21:11
buggy21:11
cbreak ubuntu? hahaha21:11
the whole thing is ancient :D21:11
fubada what is21:11
ubuntu is?21:11
cbreak yes21:11
prehistoric21:11
karstensrage → Guest8421221:11
fubada i think its the best distro to date21:11
bremner_ fubada: it's easy to install gitolite from source.21:11
cbreak some weeks ago someone came here with git 1.5.something21:11
fubada debian is even farther behind21:11
bremner_ fubada: that is neither true nor relevant21:11
albel727 remembers that moment )21:11
bremner_ at the moment21:12
cbreak I mean, I started with git 1.5.4 many years ago21:12
even apple's git is newer :)21:12
anyway, I recommend installing both git and gitolite from their respective git repositories21:12
albel727 ok, cbreak, just advise him a new distro =)21:12
cbreak that's what I did at work21:12
we use RHEL, but it's similarly prehistoric21:13
RichiH next question: I am using sed in a pre-commit hook to change a file i added. the change happens, but it's not actually commited21:13
cbreak only thing I used the stock stuff was to clone the new versions and compile from source21:13
RichiH: of course not21:13
RichiH: you change the file, right?21:13
git doesn't commit the files21:13
git commits the stuff that is staged21:14
albel727 it commits index21:14
cbreak I don't know if it's valid to change the index in that hook either21:14
← Fuwex left21:14
albel727 RichiH: you may want to look at clean/smudge filter functionality.21:15
← rolfb left21:15
albel727 what is that you use your sed for?21:15
← kinesis left21:15
cbreak I read somewhere that sed is turing complete21:16
so ... potentially anything :)21:16
albel727 lol21:16
cbreak but if he wants to change a commit behind the back of a user21:17
it's probably nefarious21:17
albel727 Well, it might be turing complete, but writing anything complex on it will be harder than even on assembler. I'd write a higher level language and call it Ced++ ))21:17
fubada so ok, without gitolite, just git, i need to add my pubkeys to /home/git/.ssh/authorized_keys?21:17
thats it21:17
?21:18
cbreak fubada: no21:18
RichiH albel727: i keep on hearing smudge filters... i really need to check them out :p21:18
cbreak you need to create one user per user21:18
← io2 left21:18
→ dajhorn joined21:18
→ brodo_ joined21:19
albel727 cbreak: I gave him gitolite-overkill link. it explains how to work with --shared repo.21:19
cbreak and put the public key of that user into the ~username/.ssh/authorized_keys21:19
RichiH if it git added all staged files after my sed, that would work as well, no?21:19
albel727: replacing passwords and/or gpg-encrypting the file21:19
cbreak RichiH: that sounds stupid21:19
albel727 RichiH: as cbreak said, it might work, or it might not. and even if it works, not sure if that's supposed to.21:19
RichiH *shrug*21:19
cbreak and I wouldn't rely on it to work21:19
it might change in other versions too21:20
RichiH i only want it for low-security passwords anyway21:20
cbreak how about not committing passwords? :)21:20
RichiH if i want to keep, e.g., my pianobar config in git, i need to commit the password _somehow_21:20
← brodo left21:21
brodo_ → brodo21:21
albel727 RichiH: well, you certainly can use clean/smudge filter for that. read http://progit.org/book/ch7-2.html "keyword expansion" section. and man gitattributes.21:21
jast RichiH: [automatic message] the 'gitattributes' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/gitattributes.html21:21
fubada ok21:22
thanks gjys21:22
guys21:22
← robertdavid left21:23
RichiH thanks, i'll have a look21:23
arand Is there a way to do this more quickly in a rebase conflict: "git show :3:debian/changelog > debian/changelog" "git add debian/changelog" If I already know which version I want to pick for all conflicts?21:24
RichiH whoah, the diffing of binary files is awesome21:25
→ f0i joined21:26
← candybar left21:26
→ cynic_spasm joined21:26
→ candybar joined21:27
jast arand: replace first step with, for example, git checkout --theirs <files>21:27
I think --theirs corresponds to stage 321:27
arand Ok, cheers21:27
→ ajt joined21:28
→ paeva_ joined21:29
→ dpy joined21:30
RichiH hmm, there is no merge strategy that basically says "don't even try, just dump a copy of each version in my working tree and let me figure out the rest"?21:30
at least the manpage does not seems to list anything i could make act that way21:31
wereHamster RichiH: dump that when? In case of a conflict? Or always?21:31
← sechrist_ left21:36
RichiH in case of conflict, yes21:36
arand I have three branches, one completely independent and two in a liner history. How can I rebase so that the two branches gets attached to the top of the one independent branch?21:37
albel727 RichiH: there is. it's called binary merge strategy.21:37
fubada ok21:38
i got it working with just git21:38
albel727 RichiH: well, actually it just leaves an unmodified copy of your file and signals conflict, but you can easily checkout all other versions of files, using checkout --theirs checkout --ours and checkout --base, iirc.21:38
fubada is it still possible to do a post-commit hook?21:38
→ workmad3 joined21:38
fubada i want at the time of commit, for the server to refresh stuff under htdocs/foo21:38
wereHamster RichiH: in case of a conflict, you can access the versions using <n>:<path>, e.g. :0:README, :README. man git-rev-parse21:38
jast RichiH: [automatic message] the 'git-rev-parse' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-rev-parse.html21:38
arand ...I'm trying to get something in order for a svn commit, but when I simply merge it squashes everything and only commits one thing with the merge as the commit message...21:39
← infid left21:40
→ MikhX joined21:40
→ Ziaix joined21:41
→ infid joined21:41
Ziaix Hey #git, I need a little help. I've messed up a branch, and want to revert to the last upstream commit. Can anyone tell me how to do this?21:42
wereHamster Ziaix: git reset --hard @{u}21:42
→ berserkr joined21:43
Ziaix I now get an odd message from git status: "# Your branch and 'origin/master' have diverged, and have 3 and 9 different commit(s) each, respectively." Any idea what I should do about that?21:44
wereHamster Ziaix: git reset --hard origin/master21:44
Ziaix Thanks :)21:45
wereHamster @{u} shoudl be the same as origin/master, I have no idea why the first command didn't work.21:45
but I'm too lazy to find out why21:45
← workmad3 left21:45
← smolyn left21:46
albel727 oh well, aren't we all?21:46
← kipras left21:46
Ziaix Heh. Not as important, but am I able to do that to the changes on the remote branch too?21:46
albel727 you mean you have pushed your screwup before?21:47
Ziaix yep. Some of my screwup got pushed before I noticed what I'd done.21:48
(I'm not used to working with an upstream repo, have only used a single remote previously.)21:48
albel727 oh, then it's a problem. if you're sure nobody have pulled that, you can fix it with git reset and a force-push. otherwise, use git-revert.21:48
← glcrazy left21:49
Ziaix Yeah, no-one will have pulled it. (I know for a fact that the other person is asleep)21:49
← berserkr left21:50
→ berserkr joined21:50
albel727 well then, use git reset --hard <commit> to reset your master to the point of history you want, and then git push --force origin master:master (or whatever is your remote and branches are)21:51
wait a sec. single remote? you mean you have origin and upstream?21:52
oh well, it's not important. unless you've somehow screwed up upstream.21:52
Ziaix No, my changes haven't been pulled into upstream.21:52
albel727 yeah, I thought that it's improbable.21:53
Ziaix Origin is a fork of the upstream repo, and I forgot to pull in changes before writing and committing stuff.21:53
albel727 that's not so much of an error.21:53
← aalex-home_ left21:53
← aalex left21:53
albel727 in itself.21:53
→ wafcio joined21:54
albel727 I'd say it's not an error at all. It's easily fixed by just pulling the upstream and pushing again.21:54
→ Corren joined21:54
Ziaix Not that bad, but the changes clashed with the upsteam commits, and I don't think I could safely merge them.21:54
albel727 why not? doing conflict resolution is usually easier than writing from scratch.21:55
→ aalex joined21:56
← Corren left21:56
Ziaix I'd made minor changes, but couldn’t work out how to merge properly.21:56
← wafcio left21:56
→ tansell-laptop joined21:57
← serverhorror_ left21:58
albel727 well, I guess it also happens from time to time.21:59
→ JohnnyL joined21:59
JohnnyL Any of you use fossil?21:59
← hvgotcodes left21:59
albel727 sounds familiar...21:59
→ Corren joined22:00
albel727 oh, it's a vcs too. how nice.22:00
Ziaix Why is it that after a revert I have untracked files?22:02
JohnnyL It's very nice. Just haven't figured out a way to go back to a previous version (aborting commit changes) then recomiting. You can undo an update, just not a commit. (afaik)22:02
albel727 Ziaix: you went with git-revert?22:03
JohnnyL I checked out a pervious version. Hm, perhaps I didn't go back 'twice'.22:03
Ziaix Yes.22:03
← kingfishr left22:05
albel727 no idea, then. it shouldn't have happened. maybe you've reverted .gitignore or something.22:05
Ziaix They aren't things that would be in .gitignore :S22:05
← Corren left22:05
Ziaix Gah, I seem to just be making more mess.22:05
Arrowmaster a revert only applies the opposite of a specific commit, if you had untracked files before you did the revert you will still have untracked files22:06
albel727 as long as history is fine - everything is fine.22:06
← Dave^| left22:06
albel727 untracked files can be gotten rid of with git clean.22:06
Ziaix albel727: my history is fine, I'm just having trouble getting it back.22:06
albel727 why not git-reset + force-push?22:06
← canton7 left22:07
Ziaix I tried that.22:07
albel727 and what happened?22:07
Ziaix I ended up with untracked stuff, I think.22:07
albel727 or rather, show me your git reset and push command.22:07
JohnnyL yeah this is no good, if you check out to a previous version in fossil it will include all changes even after the previous version.22:08
← ISF left22:09
→ ISF_ joined22:09
Ziaix How insensible would be removing the local stuff, and starting again by fetching the remote?22:09
albel727 very. there's nothing you can do with git clone, that you can't do with git-reset22:10
you just need to show me your commands.22:11
albel727 sounds like a child molester. =/22:12
← Corey left22:12
albel727 ...especially because my persuasion is not working.22:12
→ jophish joined22:13
Ziaix Heh. Well, I currently have a terminal with a long backlog of various attempts to fix things.22:13
albel727 pastebin it, why don't you?22:13
Ziaix http://pastebin.com/2EjXvpAy22:14
albel727 also give me the output of "git log --graph --oneline --decorate --all" and tell me where you want your master to be.22:14
→ sgeigerbot joined22:15
albel727 oh my, what a mess.22:15
nowhere_man I went throug git(1) but didn't find the command to remove commits for the object store, which one is it?22:15
albel727 well, we'll be fixin all that in a jiffy.22:15
Arrowmaster nowhere_man: gc or repack22:15
← ereslibre left22:16
FauxFaux needs to get the actual pasting bit of http://gitpad.goeswhere.com/ to work properly.22:16
Ziaix http://pastebin.com/H7HgNsm8 and trying to get back to: 0b793e822:16
And yeah, I've made a mess.22:17
← sgeigerbot left22:17
Ziaix I've gone the lead pipe route, in a somewhat silly attempt to make it work on my own.22:17
albel727 Ziaix: git checkout master; git reset --hard 0b793e8; (check that everything is in order); git push --force origin master:master22:18
← RobertLaptop left22:18
Ziaix Done git checkout master; git reset --hard 0b793e8; have an untracked folder.22:19
RichiH albel727: the manpage does not list "binary" as merge strategy22:19
nowhere_man Arrowmaster: gc or repack won't remove commits that are still reachable...22:19
← phaedrix left22:19
RichiH or should i simply mark the files as binary in .gitattributes ?22:19
Arrowmaster nowhere_man: you cant remove commits that are still reachable, unless you mean still reachable in the reflog but nowhere else22:20
→ smolyn joined22:21
nowhere_man what if I did a bunch of commits and just want to get rid of them?22:21
→ Engin joined22:21
nowhere_man like 'hg strip' with Mercurial?22:21
albel727 Ziaix: well, you did have it untracked at that point of history, then. you'll be able to read it if you want afterwards. you can do git merge upstream/master before the push. you may also remove all untracked files with git clean -fd, prior to merge.22:21
*able to add22:21
Engin there's no central repository, I zipped and send my repo to my friend, he did changes, I did changes. What is the proper way to go from now ?22:21
→ phaedrix joined22:22
albel727 Ziaix: thing is, I strongly suspect, that that folder came from upstream history. you're not going to lose it if you remove it anyway, and it will reappear after the merge.22:23
← drizzd left22:23
albel727 RichiH: yeah, that's what I meant by binary merge strategy, sorry if it confused you.22:24
spetrea can I put google analytics on my github account page ?22:25
Ziaix albel727: Okay, this all seems to have worked.22:25
bremner_ spetrea: sounds like a question for #github22:25
← Adys left22:25
→ diegoviola joined22:26
albel727 Ziaix: sure? the directory reappeared? the history and contents of master and origin/master look ok?22:26
spetrea bremner_: asked there, no answer22:26
← Guest84212 left22:26
← giallu left22:26
Ziaix Yep, everything seems to be there22:26
spetrea bremner_: so I decided to ask questions all over the place22:26
albel727 that's fine then.22:26
→ karstensrage joined22:26
dpy → dpy_zzz22:28
← vdv left22:29
RichiH albel727: ok, great, thanks22:30
→ madewokherd joined22:30
→ babbio joined22:31
← arand left22:31
RichiH albel727: when running git pull, the smudge filter is run first, then the clean filter is run. that's not how i understood the docs22:32
→ Adys joined22:32
← ningia left22:32
albel727 RichiH: merges right after adding smudge filter is an edge case. the main thing, is that the end result of merge is right. is it?22:34
← Ziaix left22:35
albel727 and btw, I think I might have misunderstood you about binary strategy. you meant setting "binary" attribute, or setting "-merge"/"merge=false" attribute?22:35
→ davidrussell joined22:35
albel727 and by the way, how do you check the order of filter runs?22:36
← freshmilk left22:36
← airborn left22:37
← radarek left22:38
→ Kuwaga joined22:39
← ch077179 left22:40
← qfyre left22:40
← smolyn left22:41
RichiH albel727: yes, the end result is correct22:41
albel727: i just want to make sure certain files are _never_ merged22:42
albel727: smudge gpg encrypts, clean gpg decrypts. if i need to enter the password once, it's clean. if i need to enter it once, it's the smudge filter.22:43
← babbio left22:43
RichiH also, when pulling, i can enter the wrong password in the smudge after pulling and it's still OK22:43
← davidcoallier left22:44
← candybar left22:44
albel727 encrypting in clean/smudge filter, huh? now that is not very good idea.22:44
at least unless you write a wrapper script which will ask for the correct password and call gpg as many times as necessary to produce correct output.22:45
→ rsewill joined22:46
RichiH albel727: i will use keys for production anyway22:46
← jablko left22:46
RichiH that way, gpg can simply exit non-zero22:46
albel727 well then, I'm not sure, how it ended up ok with entering wrong password. try debugging it all with redirecting all streams erros etc to some files.22:47
*errors22:47
RichiH atm, i am using symmetric encpryption and a passphrase22:47
← dim_ left22:48
albel727 then how do you distinguish clean from smudge?22:49
← saschagehlich left22:49
albel727 in any case, that doesn't affect in the slightest, how you should debug the filters.22:49
← dpy_zzz left22:50
← rsewill left22:52
albel727 RichiH: btw, did you read about clean/smudge invariants, that you should uphold? the clean+clean=clean identity and smudge+smudge+clean=clean identity?22:52
that would be relatively bothersome to implement for encryption, I'd say.22:53
RichiH albel727: smudge encrypts, clean decrypts22:53
→ ch077179 joined22:53
albel727 yeah. so smudge+smudge+clean = 2encrypts +1 decrypt = 1 encrypt22:53
RichiH albel727: hmm, i will need a wrapper for that22:53
→ workmad3 joined22:54
← kermit left22:54
albel727 note that in general, smudge+smudge+smudge+smudge+...+clean should be = clean. and clean+clean+...+clean should be = clean. don't expect git to run only clean+clean and smudge+smudge+clean combinations.22:55
are you really sure, that you need to keep encrypted files in your repo?22:56
← jophish left22:56
RichiH yah, i will simply parse the first line.. or just run file on it22:56
albel727 well, that's how I imagined it, indeed.22:57
← cesario left22:57
→ candybar joined22:59
← mdel left23:00
fubada albel727: i got most thjings working, but have a slight issue with pushing to another remote (production)23:01
← nag left23:01
fubada it says it cant allow me to push to another remote because..one sec23:01
→ jophish joined23:02
fubada http://pastie.org/private/vlwbfzk4ykmkjycpq8vmg23:03
can someone clarify why i cant do this23:03
jast what part of the message isn't clear?23:03
fubada im assuming its because theres no way to push the same change to origin at the same time23:03
← nazgul101 left23:03
fubada its clear but..how can I push to "origin" most times, and when im ready to "production"23:04
jast it has nothing to do with how many repositories you are pushing to23:04
it's because the target repository is a so-called non-bare repository, i.e. one that has a working tree (i.e. a place you can directly edit the files)23:04
albel727 fubada: I suppose, your production repo is actually a directory with a web homepage?23:04
jast origin is a bare repository, so it works there, but production apparently isn't, hence the detailed message explaining what's going on23:05
← thiago_home left23:05
fubada production is just a git clone23:05
of /home/git/repos/myrepo23:05
albel727 fubada: the point is, do you really have to have the files checked out in that dir?23:05
→ nazgul101 joined23:05
→ eydaimon joined23:05
eydaimon how can I check a remote if I'm ahead or behind?23:06
FauxFaux git fetch && gitk --all23:06
← brian_g left23:06
fubada albel727: im sorry i dont understand the question.23:06
i guess, yes23:06
albel727 eydaimon: or git fetch + git status. or git fetch + git log --graph --oneline --decorate --all23:06
fubada: did you install the post-receive hook I told you about?23:06
eydaimon albel727: thanks23:07
fubada my post-receive hook does a "cd /home/www/dev/ftns.co/bla & git pull"23:07
← caseymcg left23:07
eydaimon albel727: i swear I did some command that showed me that "remote repository is ahead by 1 checkin"23:08
fubada i want to be able to push to "production" remote which is a git clone of /home/git/repos/myrepo under /home/www/prod/htdocs/blah23:08
albel727 fubada: you should have not reinvented a wheel and used an already made hook, that covers all the cases. in any case, wait a sec, I always forget that git config option name...23:08
eydaimon: that's what git status usually tells, if you're indeed ahead or behind the remote.23:09
fubada albel727: i dont necessarily want to auto deploy to "prod" with every commit23:09
i do that for "dev"23:09
jast eydaimon: git checkout without args, or git branch -vv to see it for all branches (of course all of this is only up-to-date if you fetched recently)23:09
← hungrygruffalo left23:09
fubada i only want to change prod files if I push to remote "prod"23:09
eydaimon albel727: but it doesn't always check with the remote... sometimes I do git status, and it says nothing about remote even though it is ahead. in fact, if it had to check remote for every git status, then it wouldn't be good decentralized practice it would seem like23:09
← workmad3 left23:10
fubada thanks albel72723:10
← ch077179 left23:10
albel727 fubada: as long as you need those files in prod, there's no helping it. do "git config receive.denyCurrentBranch false" in prod remote.23:10
jast eydaimon: well, the only things that actually talk to the remote are "git fetch" and "git ls-remote" (which just lists all refs along with their current SHA1 IDs in the remote repository)23:11
for receiving information, that is. there are also a few obscure additional things that are not relevant to the current situation.23:11
fubada albel727: so if I push to remote 'production' what is the state of remote 'origin"?23:11
← Amorphous left23:11
fubada does it also get the changes?23:11
albel727 fubada: ofc not.23:11
fubada ok i dont think thatd ever be a case anyways23:12
→ PovAddict joined23:12
eydaimon jast: ok, thanks23:12
PovAddict is it possible to keep GIT_COMMITTER_* untouched after rebasing?23:12
rebase keeps the author, but sets the committer to my own data23:12
fubada albel727: your post-receive hook can handle dev and prod cases?23:12
if so, do you mind sharing it again23:12
← davidrussell left23:13
albel727 fubada: that post receive hook just does what your hook does, but it does it always correctly and with consideration to some edge-cases.23:13
!faq non-ff23:13
jast [automatic message] Your push would lose changes on the remote: please see https://git.wiki.kernel.org/index.php/GitFaq#non-ff23:13
fubada thanks23:13
albel727 you'll find a link to it there ^23:13
→ mikepack joined23:13
PovAddict guess I'll have to patch rebase23:14
albel727 thinks so too23:14
albel727 it's really a bug.23:14
don't hesitate to ping the git mailing list with your patch.23:15
PovAddict I tried a complete abomination involving a post-rewrite hook to store the committer info and a filter-branch script later to set it back in the new commits using the info stored by the hook23:15
albel727 I feel your pain =/23:16
PovAddict but it behaved wrongly if the rebase removed commits (I got the committer of the rewritten A set to the committer of the deleted A^)23:16
jast at least it's a very straightforward patch23:16
PovAddict well sort of23:16
jast note that you actually have to edit git-am.sh (which rebase uses)23:16
though I guess tracing back the previous information might be a bit tricky23:17
PovAddict rebase-i doesn't23:17
and for both rebase and rebase-i I'll have to add a new function like get_author_ident_from_commit but for committer info23:18
← crakdmirror left23:19
← xanonus left23:20
← tewecske left23:20
→ kermit joined23:21
→ Corren joined23:23
← TomyLobo left23:26
→ Amorphous joined23:27
← kukks left23:27
← tansell-laptop left23:28
→ bryanl joined23:30
→ tansell-laptop joined23:30
← pantsman left23:33
→ mdel joined23:34
← codeshah left23:35
← cynic_spasm left23:35
→ DrkShadow joined23:37
DrkShadow Trying to pop changes off the stash, what does it mean, "Cannot apply to a dirty working tree, please stage your changes", specifically "stage your changes" ?23:37
PovAddict 'git add' = staging23:38
DrkShadow ah. Thanks.. :-)23:38
AAA_awright cache=index=stage. I prefer stage and so do a lot of people but cache.h is where the code is and index is how the functions are named so go figure23:39
→ frogonwheels joined23:40
DrkShadow any way I can pop multiple stashes?..23:40
← Sonderblade left23:40
DrkShadow I have three items stashed, trying to pop them..23:41
→ moura joined23:41
← Rotham left23:41
frogonwheels DrkShadow: prob one at a time.23:41
→ Rotham joined23:42
fission6 i am trying to push to my origin and i am getting, error: failed to push some refs to23:42
frogonwheels ie you may need to create a (temporary) commit to squish all the pops into23:42
DrkShadow frogon: any way I can do it without commiting after each pop? :-/ I stash pop, make the changes according to the merge that's required, but it's still in the list.. if I pop it again, I have to make the merge corrections again.23:42
fission6 i tried to do git commit --amend and i think it messed something23:42
DrkShadow hmm. Amend..23:42
albel727 after a push...23:42
DrkShadow nope.23:42
PovAddict DrkShadow: if "pop" gets a conflict, it won't remove the stash from the stack23:42
fission6 git commit -aAborting commit due to empty commit message. got this too23:43
PovAddict DrkShadow: after solving the conflicts and all, if you want to get rid of it from the stash stack, use 'git stash drop'23:43
DrkShadow povaddict: got it; pop, merge, correct, remove. :-)23:43
Thanks!23:43
← whaley left23:43
fission6 " ! [rejected] v.1 -> v.1 (non-fast forward)"23:44
albel727 !faq non-ff23:44
jast [automatic message] Your push would lose changes on the remote: please see https://git.wiki.kernel.org/index.php/GitFaq#non-ff23:44
fission6 so what do i do23:44
frogonwheels fission6: ok, so I'd look at gitk which will probably explain what you've done.23:44
fission6: did you try and ammend a pushed commit?23:44
fubada albel727: if i have remote: origin, dev, and prod..is there any way to automatically push to origin when i push to either prod or dev23:44
← PovAddict left23:44
fission6 i think so frogonwheels23:44
fubada this whole decentralized thing is tough to grasp :P23:45
fission6 i know my code is fine on my side and i am the only person using it23:45
so i just need to get a nice commit on my side and push it and it will be fine23:45
→ PovAddict joined23:46
CareBear\ fubada : but oh so sweet once you do23:46
PovAddict augh23:46
frogonwheels fission6: you _can_ just force the push if that is the case.23:46
fission6 yea23:46
PovAddict turns out my patch to git-rebase--interactive to make it keep the committer info only works in uncommon cases such as squash23:46
not in the simple 'pick'23:46
fission6 i think thats fine, how do i do that git push --force23:46
PovAddict because 'pick' just delegates to cherry-pick23:46
and cherry-pick is C23:46
frogonwheels fission6: normally I'd avise against it unless you have a really good reason.. since it mucks up other developers.23:46
fission6 im the only developer, how else would i "properly" resolve23:47
← Rotham left23:47
frogonwheels fission6: just rebase your commits onto the remote branch, and edit the commit to reflect that you've done that.23:47
fission6: the ability to force push is useful when it is really needed though :)23:48
fission6 hmm ok23:49
thanks23:49
← ph^ left23:49
→ daviddoria joined23:49
fubada 19:44 < jast> [automatic message] Your push would lose changes on the remote: please see https://git.wiki.kernel.org/index.php/GitFaq#non-ff23:49
→ petar joined23:49
fubada sorry23:50
19:44 < jast> [automatic message] Your push would lose changes on the remote: please see https://git.wiki.kernel.org/index.php/GitFaq#non-ff23:50
petar is there a way to use sdiff with git-mergetool?23:50
← alip left23:50
albel727 fubada: no, it's not possible, unfortunately. unless you write a script.23:50
chrisf|work → ihptru23:50
← daviddoria left23:50
ihptru → chrisf|work23:51
fubada albel727: if I have people push mainly to "dev" remote, and they forget to push 'origin' what type of a bad situation can result?23:51
do i want to stress pushing to origin first23:51
← rendar left23:52
cmyers fubada: it is kinda not intuitive for people used to centralized, when all the sudden you have many different remote repos and branches to push to.23:52
fubada right...im coming from svn23:52
albel727 fubada: I'd say, make them push wherever you want and update the other remotes from the first one yourself.23:52
fubada gotcha23:52
cmyers in my experience, it is bad to have people learn "the workflow" and mindlessly parrot it23:52
fubada thanks23:52
cmyers in git, it is really tough to make a workflow that always works23:52
← petar left23:53
cmyers there are conflicts, non-FF pushes, etc. people have to look at the branches, understand the situation, then understand how to use the tools to resolve it.23:53
There are a few resources I can recommend though, lemme look 'em up23:53
fubada thanks23:53
cmyers http://marklodato.github.com/visual-git-guide/index-en.html <-- "23:53
"Visual Git" is very helpful to people I think23:53
shows with pictures what is going on23:54
← deuterium left23:54
cmyers http://nathaniel.themccallums.org/2010/10/18/using-git-fast-forward-merging-to-keep-branches-in-sync/ has a lot of example merges23:54
and http://www.ndpsoftware.com/git-cheatsheet.html is a good command reference for folks23:54
← micro`_ left23:54
← deadaim left23:54
→ alip joined23:54
→ micro`_ joined23:54
→ deadaim joined23:55
← kermit left23:55
fubada thank you albel727 and cmyers23:56
and all the others23:56
bye23:56
← fubada left23:56
← Chibby_ left23:58
← PovAddict left23:59

Logs Search ←Prev date Next date→ Channels Documentation