IRCloggy #git 2012-05-12

Logs Search ←Prev date Next date→ Channels Documentation

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

2012-05-12

→ v0n joined00:01
Vinnie_win If the remote master has new commits, how do I bring my local master forward?00:03
(and preserve my new local commits)00:03
← anth0ny_ left00:03
← madewokherd left00:03
← amalloy left00:03
→ madewokherd joined00:04
sayd git pull --rebase00:04
→ kukks joined00:04
Vinnie_win cool! Worked00:04
→ warthog9 joined00:04
Vinnie_win Absolutely amazing, not even a merge conflict00:04
sayd cool, im sure there are reasons not to do it that way00:04
but i always forget why00:05
→ imMute joined00:05
← imMute left00:05
→ imMute joined00:05
← LazyTown left00:05
← tyatpi left00:05
→ jfischoff joined00:05
→ ehsan joined00:06
← txomon|nqas left00:06
← yitz_ left00:08
→ txomon|nqas joined00:08
← txomon|nqas left00:08
→ txomon|nqas joined00:08
sayd another way, if you just want all your changes in one big commit. you can make sure youre work is in a branch, checkout master, then git merge --squash branch00:08
diegoviola how do i send a pull request via git? without the github stuff00:09
→ williamcotton joined00:11
← Goplat left00:12
← shruggar left00:13
→ Zenith77|2 joined00:13
→ yhager joined00:14
← tharkun left00:14
← aselder left00:14
→ Destos joined00:15
← kenperkins left00:15
← macmartine left00:15
← tewecske left00:15
← yoklov left00:16
← Zenith77 left00:16
→ tharkun joined00:16
→ irqq joined00:16
← jaseemabid left00:17
← eletuchy left00:18
← tharkun left00:20
← yuriks left00:22
→ yuriks joined00:22
← jfischoff left00:22
yuriks diegoviola: send an email to the other guy?00:22
→ francisfish joined00:22
← fuzzybunny69y left00:23
diegoviola yuriks: will he have to add my repo as a remote?00:23
in order to pull00:23
→ fuzzybunny69y joined00:23
← psoo left00:24
← ehsan left00:24
← thelazydeveloper left00:26
→ gavin_huang joined00:26
→ tharkun joined00:26
← maletor left00:26
← gerard0 left00:27
→ gerard0 joined00:27
yuriks you can pull directly from url, iirc00:27
but it makes things a bit easier, I guess00:27
← irqq left00:28
diegoviola cool, thanks00:28
if i can pull directly from a url that's even easier00:29
ty00:29
← wilmoore left00:30
← soulcake_ left00:32
← JohnyBaby left00:32
→ Lord_DeathMatch joined00:36
← Xunie left00:36
→ `Assassin joined00:36
→ mritz joined00:37
← tharkun left00:37
→ HyperNerdV2 joined00:40
→ chihjen joined00:41
→ pdtpatrick joined00:41
→ tharkun joined00:43
← joaoh82 left00:43
← Araxia left00:47
← tharkun left00:49
→ Per4u3e joined00:51
→ QaDeS_ joined00:52
→ Evanlec joined00:52
→ tharkun joined00:53
← robotmay left00:53
→ cilly joined00:54
← QaDeS left00:55
→ soulcake_ joined00:56
← fuzzybunny69y left00:57
← soulcake_ left00:57
→ soulcake_ joined00:58
← soulcake_ left00:58
→ fuzzybunny69y joined00:58
← martinjlowm left00:58
← tbird72 left00:59
→ tmcmahon joined01:01
← fwg left01:01
→ carlosgaldino joined01:03
→ xiangfu joined01:04
mikepack Can someone define "upstream" and "downstream" concisely?01:04
→ elliottcable joined01:04
← fuzzybunny69y left01:05
← thearthur left01:06
→ krzkrzkrz joined01:06
CareBear\ mikepack : it depends a bit on context01:06
mikepack CareBear\: Not a good start :)01:06
CareBear\ mikepack : in general the terms are used to describe a repository you clone, and a repository which has been cloned from yours01:06
→ thiago joined01:06
CareBear\ does that make sense?01:07
← ColKurtz left01:08
← krz left01:08
krzkrzkrz → krz01:08
← tharkun left01:08
mikepack CareBear\: Not really :-/01:08
→ tharkun joined01:08
mikepack CareBear\: I thought there was more coming...01:08
CareBear\ let's talk specifics in stead01:09
what do you have and why are you asking? :)01:09
mikepack maybe "concisely" is too much of a restriction01:09
→ chihjen_ joined01:09
sayd downstream == local; upstream == remote?01:09
CareBear\ no01:09
no good01:09
mikepack : talk about your situation01:09
← chihjen left01:10
mikepack CareBear\: I have no specific situation. I also mostly understand the terms. My understanding is a newly cloned repo or newly created branch are "downstream" because changes made to it are created after the origin/master.01:11
CareBear\ there are very few rules01:11
which means that git allows you to mess up badly if you tell it to01:11
this is a feature01:11
thiago → thiago_LAX01:12
CareBear\ because it allows to do awesome stuff once you can tell it to :)01:12
← pdtpatrick left01:12
mikepack CareBear\: You're being too elusive01:12
→ intripoon_ joined01:12
CareBear\ you are being too unspecific :\01:13
→ Vampire0_ joined01:13
→ Zenith77 joined01:13
mikepack This isn't specific? "Can someone define "upstream" and "downstream" concisely?"01:13
CareBear\ no01:13
mikepack If not, what can make it more specific?01:13
CareBear\ context01:13
mikepack Ok, what possible contexts are there?01:13
CareBear\ lots01:13
mikepack So the terms differ based on context?01:14
CareBear\ it depends on how anal you are about meanings01:14
mikepack Well, I'm more of less looking for a good "rule of thumb"01:14
CareBear\ if you have a fuzzy matcher and can extrapolate then you can get a good feel for what is intended01:15
sometimes the meaning of upstream is more specific than other times01:15
← avinashhm left01:15
← intripoon left01:15
← Zenith77|2 left01:16
CareBear\ the second hit on google "git upstream" looks relevant too.. http://stackoverflow.com/questions/2739376/definition-of-downstream-and-upstream01:16
oh right01:16
GitHub uses "upstream" as a recommended standard name of a remote in their tutorials01:16
← Vampire0 left01:17
→ nrichards joined01:17
CareBear\ so in the context of those github instructions it has a quite specific meaning01:17
mikepack That's what I do as well. If I fork, I call the primary repo "upstream"01:17
CareBear\ "fork" is a bad word imo01:17
in general I'm rather adverse to github01:17
mikepack Github calls it fork...that's why I used the term.01:18
CareBear\ yes01:18
fork is however not a git term01:18
fork is rather a social term01:18
mikepack What's the purists analog?01:18
CareBear\ it depends on what you do01:19
mikepack I'm actually never really "forked" a repo locally01:19
CareBear\ there are several ways to do the analogous of github fork on the git commandline01:19
mikepack such as?01:19
CareBear\ one way is simply git clone01:19
← Lgb left01:20
CareBear\ that assumes you don't have any of the repository already01:20
if you *do* have some of the repository already then you will instead add a remote, fetch commits from there, and do something with those commits. possibly first create a branch, then merge or rebase or cherry-pick commits in one or the other direction01:21
← ymasory left01:21
CareBear\ this is what I do for my Linux kernel01:21
ScottNYC → ScottNYC|away01:21
← sayd left01:21
CareBear\ I have several different Linux kernel trees as remotes in my worktree01:21
and I mix them together in order to get the kernel code I want to run01:21
imMute CareBear\: I like to clone from another local repo, but setup the new repo's origin to be the real origin01:22
← h0h0h0 left01:22
CareBear\ imMute : no problem01:22
→ Lgb joined01:22
CareBear\ imMute : cd /tmp && git clone /var/lib/git/anotherlocal.git && cd anotherlocal && git remote origin set-url ssh://another.org/therepo.git01:22
imMute yeah, that01:23
CareBear\ or you can even edit .git/config manually01:23
→ elijah joined01:23
→ Ecesis joined01:24
← Ecesis left01:24
mikepack Ok, so if I clone a repo, I'm now downstream and origin is upstream. Does my clone always remain "downstream"?01:24
← tmcmahon left01:24
thiago_LAX mikepack: no01:25
mikepack: there's no downstream or upstream, aside from social convention01:25
mikepack: the repositories are equal as far as Git is concerned. The label is something you apply. As such, you can change at any time.01:25
CareBear\ mikepack : git by default remembers where you cloned from but you can change that as you like at any time01:25
mikepack thiago_LAX: I understand it's a social convention. But if I were discussing the current status of something, I would likely need to be using either "upstream" or "downstream"01:26
CareBear\ not neccessarily01:26
but let's keep going here01:26
mikepack Let me ask this01:26
thiago_LAX mikepack: if you want to use it, you can.01:27
CareBear\ yes! :)01:27
mikepack Given I'm "downstream" after a clone, when would I no longer be considered "downstream"01:27
← elijah left01:27
mikepack ?01:27
:)01:27
CareBear\ you aren't neccessarily downstream after a clone01:27
mikepack When do I become downstream?01:28
CareBear\ again, clone is a git term, downstream is a social term01:28
mikepack FWIW, the stack overflow question you linked me to describes it as such :)01:28
So, when I clone, I'm just "stream"01:28
:)01:28
CareBear\ when you clone you have a clone01:28
mikepack lol01:29
TOOOO ELUSIVE!01:29
:)01:29
CareBear\ have you used other version control?01:29
mikepack svn01:29
CareBear\ nod01:29
do you know crypto?01:29
mikepack What I'm gathering is that there really is no good definition for the terms01:29
Which is why I don't fully understand them01:29
and the reason you're being elusive is because the terms themselves are elusive01:30
CareBear\ "clone" is not at all elusive01:30
imMute mikepack: when you clone a repository, your clone is *exactly* the same as the original. one of them being "upstream" of the other is purely a management/social construct01:30
→ makeawish joined01:30
imMute thinks English isn't mikepack's native language01:30
CareBear\ mikepack : I'll add to what imMute just said that it's exactly the same content-wise01:30
mikepack Guys, I understand what a clone is, I use it quite frequently (given I did just learn about "forks")01:31
CareBear\ mikepack : but settings can and will be different, including the URL of the origin remote, which is created when git creates a repository by clone as opposed to init01:31
mikepack I understand "clone" is a command while "upstream" and "downstream" are social terms01:31
imMute okay, lets put it this way. you've got 3 people, all working on the same project. guyA starts it and makes his git repo publicly available. guyB clones it and starts helping. guyA could be considered "upstream" of guyB. guyC then clones guyA's as well. guyC (and guyB) are downstream of guyA.01:33
mikepack imMute: That's my current understanding.01:33
imMute now, suppose guyA goes off to college and stops working on the project and guyB 'takes over' managing the project. guyC would then pull-from/push-to guyB and guyB would be "upstream" of guyC01:33
OR01:34
guyB clones from guyA and guyC clones from guyB. guyB is downstream of A but upstream of C (assuming C doesn't interact with A)01:34
given that you can pull from and push to just about anywhere you damn well please, it really only makes sense to relate it to how code flows into a project01:35
→ Ecesis1 joined01:35
imMute for instance, Linus is upstream of *everybody* because it's generally accepted that his git.kernel.org repo is the Golden Master that everyone bases their work off of01:35
→ nOStahl joined01:36
mikepack imMute: Understood.01:36
imMute another way to think of it is projects that have dependancies (such as shared libraries). "upstream" would be projects that you base yours on, and "downstream" would be projects that use your library01:36
← Lord_DeathMatch left01:36
← Dave^| left01:37
mikepack imMute: Interesting, "downstream" and "upstream" can be used a means of describing dependencies, no?01:37
← tharkun left01:38
→ tharkun joined01:38
imMute sure. they're not concretely defined words01:38
→ Goplat joined01:38
mikepack If I branch off master, I'm not dependent on master and considered "downstream" of master01:38
← Ecesis1 left01:38
mikepack not dependent on master...not sure that's the best verbiage01:39
imMute if master is where "all the real work happens" then yes you'd probably be considered downstream of work happening on master01:39
why are you dwelling so much on these two words anyway?01:39
mikepack imMute: Does it seem like I'm dwelling or just being inquisitive? :)01:39
imMute inquisitive would be "I've never heard these terms before, lets go look them up on wiktionary or google.", what you're doing borders on OCD.01:40
SethRobertson upstream/downstream are normally used in the context of repositories, not branches01:40
mikepack imMute: I must be considered always dwelling by my peers then :)01:40
thiago_LAX especially since we told you that it's a social convention. They mean what you want them to mean.01:40
if you want them to change, they change.01:40
mikepack but it's a convention none-the-less, and conventions shouldn't just magically change01:41
CareBear\ ?01:42
conventions change every day01:42
imMute mikepack: move on with your life. seriously.01:42
← jstout24 left01:42
CareBear\ +101:42
mikepack imMute: this is clearly not the place for intellectual questions, apparently...01:42
→ Ecesis joined01:43
imMute uh, this is #git not ##philosophy...01:43
→ josefig joined01:43
← josefig left01:43
→ josefig joined01:43
mikepack So I should only really pose dry, easy to answer questions with straightforward answers?01:43
This conversation went south for no reason...01:44
← Ecesis left01:44
mikepack Good day gents, I appreciate you at least humoring me...honestly01:44
CareBear\ mikepack : please ask about concrete problems, be it with understanding models, technical terminology, or actual commands not working01:45
mikepack CareBear\: my question doesn't fall into "technical terminology"?01:46
CareBear\ mikepack : no, it doesn't01:46
mikepack : again, upstream and downstream are not neccessarily technical terms, only in some contexts, and you did not really have a particular context in mind01:47
mikepack CareBear\: Why ask only concrete questions? Seriously, how boring is that?01:47
CareBear\ it's a very efficient way to learn01:47
it's also a very efficient way to teach01:48
neersighted → neersighted|AFK01:48
← Chillance left01:48
mikepack Well, I guess I just prefer being inefficient.01:48
← notVert left01:49
← francisfish left01:49
← SegFaultAX left01:49
infogulch → infogulch|afk01:50
← Macca- left01:51
← r126t left01:53
← hachque left01:53
← tharkun left01:54
→ luke-jr joined01:57
luke-jr is there a way to grep through all git's objects, even unreachable from refs?01:58
thiago_LAX not easily01:58
← chihjen_ left01:58
luke-jr how about list all objects? (then I can do a loop of git-show..)01:59
thiago_LAX git fsck, maybe01:59
→ tharkun joined01:59
→ iamjarvo joined02:00
→ apok_ joined02:02
← apok_ left02:02
← tharkun left02:03
yuriks can you just ls the object store?02:03
imMute except for packed objects.02:04
→ lonewulf` joined02:04
yuriks oh02:04
← apok left02:05
← alanp left02:06
→ rudrab joined02:08
→ alanp joined02:08
→ kenperkins joined02:09
← iamjarvo left02:09
← btanaka left02:09
→ hachque joined02:09
→ tharkun joined02:09
rudrab Dear friends, as svn can store all my different programs vc in one place , say repo02:10
is there any such thing in git? or it creats different repo for different program?02:10
← kenperkins left02:11
imMute rudrab: it's entirely up to you. most people would prefer to have a separate repository for different projects02:12
← hwrdprkns left02:12
← p3rror left02:12
→ hwrdprkns joined02:12
→ Zenith77|2 joined02:13
rudrab imMute: actually as it is there in http://git-scm.com/book/en/Git-Basics-Getting-a-Git-Repository I cannot figure out how to make a central repo02:14
imMute you make a git repository with 'git init'02:14
ScottNYC|away → ScottNYC02:15
rudrab say, I want all my projects A,B,C etc git in a central dir02:15
imMute, here it says " you need to go to the project’s directory and type $ git init "02:15
← ScottNYC left02:16
imMute 'git init' creates a git repository. there is no "central/master" repository like in svn02:16
rudrab so the git repo will be created in the same dir as my source dir. is it correct?02:16
← Zenith77 left02:16
rudrab or I may make the repo outside the sourcedir?02:17
imMute yes, similar to the .svn folders in svn (except there's only one at the root)02:17
you might want to read http://git.or.cz/course/svn.html02:17
← gerard0 left02:18
→ igotnolegs joined02:18
→ davidbanham joined02:18
← osmosis left02:20
rudrab imMute, I found that the .git file is placed in the working directory itself.02:24
yuriks yes, that's the repo02:24
rudrab can you please let me know your strategy if you accidentally delete the source dir?02:25
CareBear\ mikepack : I didn't say you are inefficient. I think it's more a question of what you want to learn. the upstream/downstream stuff that can be learned in #git should be the technical ones in order to reach success02:25
imMute rudrab: have it mirrored somewhere else02:25
or keep it backed up like you should have your $HOME anyway02:25
rudrab imMute, may be i am not explaining correctly02:26
I have a $HOME/ABC project02:26
when i do a git init , the repo is created inside the $HOME/ABC , correct?02:27
imMute yes02:27
rudrab as $HOME/ABc/.git02:27
→ Macca- joined02:27
→ p3rror joined02:27
← Goplat left02:27
rudrab so you are advicing to "push" the mirror at some other place as well, right?02:28
imMute correct02:28
rudrab ok02:28
I got it02:28
imMute if you really only need the project on that one machine, it's just a really simplified backup strategy02:29
rudrab that is "clone" in git i guess02:30
ok02:30
→ BiggFREE joined02:30
imMute clone is more like pull than push02:31
→ TheShagg joined02:31
gitinfo set mode: +v02:31
TheShagg *sigh* how does one make sense of a git rebase 3-way-diff02:31
← orafu left02:32
TheShagg actually, maybe I know my problem02:32
rudrab ok....02:32
→ orafu joined02:32
TheShagg for some background, I went back in time, and split a commit into about 6 commits, and now trying to rebase the old head onto the head of my new commits02:33
normally this works fine, but thats because I split off one extra commit, not 502:33
i think its choking on the 502:33
i need to cherry pick02:33
rudrab imMute, thanks for your clarification.02:33
SethRobertson TheShagg: The number of commits doesn't matter. You'd have to use the three argument version of rebase for your usage model02:34
cherry-pick is fine too02:34
TheShagg cherry-pick vs rebase, in this case, is like a push vs pull sorta02:35
?02:35
← vervic left02:35
SethRobertson But honestly, your probably is probably that you reset to put yourself back in time instead of using `git rebase -i` (or you have merge/branches, in which case you are hosed)02:35
No, the three argument version of rebase and cherry-pick are essentially implemented in terms of each other02:36
← berserkr left02:36
SethRobertson It doesn't matter which you use02:36
ignoring merges of course02:36
← Eridius left02:36
← machty left02:37
TheShagg i am familiar with using rebase -i to squash/fixup, but can you use it to split a commit?02:37
→ vervic joined02:37
TheShagg what I have been doing is reset -> reset --mixed -> git add -p -> git add -p -> ....02:37
→ gerard0 joined02:39
← klj613 left02:40
← Destos left02:40
→ JohnyBaby joined02:40
← williamcotton left02:40
CareBear\ you can combine rebase -i with reset && add -p02:41
← txomon|nqas left02:42
→ Chib joined02:42
→ txomon|nqas joined02:43
← txomon|nqas left02:43
→ txomon|nqas joined02:43
← Zyrkon_ left02:44
← rudrab left02:45
→ Zyrkon joined02:45
→ kenperkins joined02:45
TheShagg CareBear\: can you elaborate a touch02:46
→ iamjarvo joined02:47
← alvaro_o left02:47
← carlosgaldino left02:53
TheShagg SethRobertson: BOOM, 3 argument rebase seems to be the ticket02:53
→ carlosgaldino joined02:53
← _huoxito left02:53
SethRobertson Using it with -p is probably a best practice02:53
→ dcorbin joined02:54
SethRobertson TheShagg: See also !perfect for how to use rebase-i to split commits02:54
gitinfo TheShagg: [!postproduction] So, you want to make your commit history look pretty before pushing? http://sethrobertson.github.com/GitPostProduction talks you through how to use 'rebase -i' to do this.02:54
→ ScottNYC joined02:55
← kenperkins left02:55
← SkAzZ left02:56
→ Yuuhi` joined02:59
← Yuuhi left03:01
TheShagg SethRobertson: thanks, this looks tasty, hopefully I will read it sometime tonight03:01
→ _huoxito joined03:02
← ScottNYC left03:03
← davidbanham left03:03
→ benbergman joined03:03
benbergman hello all; i just started playing with hooks and i noticed that when i push to origin and then clone, my hooks are not cloned03:04
is there a way to keep this synchronized?03:05
→ tiglog joined03:06
imMute benbergman: not automatically03:07
→ ScottNYC joined03:07
→ dr0id joined03:08
benbergman imMute: hmm... ok03:09
seems to me it could be useful to optionally sync scripts (i can see a desire to not always sync them)03:09
oh well, not a deal breaker03:09
imMute it's a security hole to have them automatically installed03:10
← v0n left03:10
← ben225 left03:11
benbergman fair enough03:12
→ quazimodo joined03:12
→ ben225 joined03:13
→ Zenith77 joined03:13
← dcorbin left03:13
← madewokherd left03:14
thiago_LAX → thiago03:14
← txomon|nqas left03:15
← nOStahl left03:16
← thiago left03:16
← Zenith77|2 left03:17
→ txomon|nqas joined03:17
← txomon|nqas left03:17
→ txomon|nqas joined03:17
← tiglog left03:23
← hachque left03:24
→ ymasory joined03:26
← marhaban1 left03:26
→ s0ber joined03:28
← Orphis left03:32
→ Orphis joined03:32
← drizzd left03:33
← QaDeS_ left03:33
→ pdtpatrick joined03:33
← pdtpatrick left03:34
← edude03 left03:35
→ jfischoff joined03:36
quentusrex_ is there a way to have git show changed files on a git pull?03:36
→ project2501a joined03:36
← joeytwiddle left03:37
nb → i03:37
← leonardotartari left03:38
→ khmarbaise_ joined03:39
→ pdtpatrick joined03:39
← ScottNYC left03:39
quentusrex_ I'm trying to configure my local git repo to show stats and a summary of changed files on a merge?03:40
s/?/\./03:40
i → nb03:40
← BiggFREE left03:40
← khmarbaise left03:41
khmarbaise_ → khmarbaise03:41
← pdtpatrick left03:42
quentusrex_ Found it: git config --global merge.stat true03:42
← btree_ left03:42
← redShadow left03:42
→ btree_ joined03:43
← btree_ left03:43
→ btree_ joined03:43
← krz left03:44
← kukks left03:47
→ pdtpatrick joined03:47
→ SegFaultAX joined03:49
→ Vooloo joined03:50
Vooloo I cant add new code to my new repo: error: src refspec master does not match any.03:50
← TonkaTruck left03:50
→ joeytwiddle joined03:50
← pdtpatrick left03:50
→ munichlinux joined03:51
→ pdtpatrick joined03:51
← lonewulftp left03:52
← az left03:54
← SegFaultAX left03:55
→ az joined03:56
← fisted left03:58
← mikepack left03:58
→ mikepack joined03:59
→ angelsl joined04:01
← _huoxito left04:01
→ SegFaultAX joined04:01
← munichlinux left04:03
← mikepack left04:03
← dzonder left04:03
→ xsbeats joined04:04
→ dv310p3r joined04:05
→ pjn_oz joined04:08
← jpalmer left04:08
← JohnyBaby left04:08
→ jpalmer joined04:08
→ kukks joined04:09
← cilly left04:11
→ cilly joined04:13
→ Zenith77|2 joined04:13
→ Goplat joined04:13
→ davidbanham joined04:14
← Zenith77 left04:17
← unreal left04:17
← pjn_oz left04:17
← tharkun left04:21
← HyperNerdV2 left04:21
→ pjn_oz joined04:22
→ chihjen joined04:22
→ tharkun joined04:27
← Per4u3e left04:27
→ maleknet joined04:28
→ davidban_ joined04:30
← davidban_ left04:30
← davidbanham left04:31
← SeySayux left04:35
jameslord why git revert failed:(04:35
http://iyanwu.com/pastebin/showthread.php?tid=5204:36
← tharkun left04:37
→ tharkun joined04:37
← Goplat left04:38
→ jceb joined04:38
← PerlJam left04:39
sp3ctum jameslord, it says you have unmerged files. do they show up on git status?04:39
→ PerlJam joined04:40
→ fisted joined04:43
jameslord sp3ctum yes, there are huge number of conflictions, how to fix it?04:47
← brokenshakles left04:48
← kukks left04:49
→ brokenshakles joined04:50
→ SeySayux joined04:50
← tharkun left04:50
sp3ctum jameslord, do you know where they came from? if yes, you can fix them with e.g. git mergetool04:54
→ QF-MichaelK joined04:55
→ tharkun joined04:56
← QF-MichaelK left04:58
← EyesIsMine left04:59
← tharkun left04:59
bryanmv`away → bryanmv05:06
→ wilmoore joined05:07
← hexa2k9 left05:07
→ hexa2k9 joined05:07
→ Prime1309 joined05:11
← mritz left05:12
→ Zenith77 joined05:13
← kadoban left05:14
→ tharkun_ joined05:16
← Zenith77|2 left05:17
jameslord sp3ctum it is imposibble to resolve such conflictions, since this merge is done between two branches from git svn05:20
→ macmartine joined05:20
→ teotwaki_ joined05:20
→ mrsrikanth joined05:22
← btree_ left05:26
← TheShagg left05:27
← ThunderChicken left05:30
→ ThunderChicken joined05:30
← rockband left05:33
→ xunie-laptop joined05:34
→ btree joined05:34
← OOPMan left05:34
← xunie-laptop left05:39
→ xunie-laptop joined05:39
xunie-laptop So, I have made 2 changes to one file and I'd like to commit one change but not the other.05:39
→ ScottNYC joined05:40
xunie-laptop When I stash, git automatically resets. Is stashing the way to go?05:40
← subbyyy left05:40
← txomon|nqas left05:40
jameslord how to cancel git merge,05:41
git revert failed since conflict happened05:41
CareBear\ xunie-laptop : no. use git add -p file05:42
→ txomon|nqas joined05:43
← txomon|nqas left05:43
→ txomon|nqas joined05:43
CareBear\ jameslord : look at git reset, but be mindful that it will delete your work if you tell it to05:43
nibalizer xiangfu: i mean, it kinda worksm but really cp /tmp is awesome until you're a git pro05:43
(i am not a git pro yet)05:44
← tharkun_ left05:49
→ tharkun_ joined05:49
→ michaelgamble joined05:50
michaelgamble heyo05:50
lets say i have a repo with 25 checkins, how do i revert the repository back to the 20th checkin eliminating 21 - 25 (reverting)05:51
ivan man git-reset05:51
gitinfo the git-reset manpage is available at http://jk.gs/git-reset.html05:51
← mrsrikanth left05:51
ivan if you want a new commit that undoes 21-25, git-revert05:52
→ Sigma joined05:52
michaelgamble no i want to back track eliminating the commits from the history05:53
← txomon|nqas left05:54
xunie-laptop So, I'm an idiot for thinking I should add .o's to git, and thankfully I haven't uploaded my git repo. Any great ways to fix this?05:55
ivan http://sethrobertson.github.com/GitFixUm/fixup.html05:57
might be more complicated if you added .o's in many commits05:57
xunie-laptop Thanks, you people are like so awesome.05:58
← xiangfu left05:59
michaelgamble man this stuff is confusing when your new to the terminology06:02
← stkrzysiak left06:02
← smw left06:05
→ smw joined06:05
→ gebbione joined06:09
← tharkun_ left06:10
→ tharkun_ joined06:10
→ solofight joined06:11
← macmartine left06:12
→ Zenith77|2 joined06:13
← solofight left06:14
← Zenith77 left06:17
→ OOPMan joined06:17
→ giallu joined06:19
← neurodrone left06:21
michaelgamble ok got another question06:23
← angelsl left06:23
bob2 michaelgamble, did you push your branch anywhere?06:23
← xunie-laptop left06:23
michaelgamble yeah i got my master reset to the revision number06:23
sh thingy06:24
bob2 if oyu pushed, reset is of no use06:24
jameslord git stash failed, why?06:24
http://iyanwu.com/pastebin/showthread.php?tid=5306:24
michaelgamble i did a git push --force06:24
now i look at the history and its where i want it to be06:24
bob2 jameslord, why are you in the middle of a merge06:24
→ Araxia joined06:24
jameslord bob2 i want to cancel that erroneous merge:(06:25
michaelgamble ok so new question.. have repoA all the old history of the past 3 years of a project, and repoB is every checkin from where repoA ends and forward06:25
how do i merge repoB into repoA so that i have one repo with all history to date06:25
bob2 jameslord, 'git checkout -f' iirc, throws away all local changes06:26
michaelgamble, depends06:26
a) don't do silly things like that in the future06:26
b) do you want a hacky splice, or do you want to rebase the B branch on top of A's branch?06:26
michaelgamble a) migrating svn repos that have history prior to trunk branch structure and then part way through switched to a trunk branch structure06:27
b) I'm not sure which i want06:28
A is nothing but trunk06:28
bob2 not sure what a means or why it matters06:28
michaelgamble B is trunk with a new branch06:28
bob2 well, think about whether you want to invalidate all existing branches based on B or not06:28
→ nowhere_man joined06:29
jameslord bob2 git checkout -f not works:(06:29
bob2 jameslord, it's useless to say that on irc, you need to bpaste.net what you ran and what the output was06:29
michaelgamble i want to end up with B's master and branch06:29
bob2 I'd suggest rebasing all of B06:29
michaelgamble but with A's master back history merged into the B's master06:29
bob2 since that leads to sanity in the future06:29
← Vooloo left06:30
← MestreLion left06:30
michaelgamble hmm06:30
bob2 it just replays every commit from B/whocares on top of A/whatever06:31
michaelgamble hmm06:32
i guess i don't understand why i can't just plunk B into A06:32
bob2 sure you do06:32
because every commit's ID is a hash, and part of the hashed data is the ID of the previous commit06:33
so obviously you can't just append B onto A, since it means changing the id of /every single past commit/06:33
michaelgamble damn06:33
bob2 rebasing means making new commits, that just happen to have identical content to the ones in B06:34
except the descend from commit(s) in A06:34
michaelgamble know where i might find a walkthrough on doing that?06:35
bob2 no06:35
michaelgamble got a git command i can look up?06:35
bob2 rebase06:35
michaelgamble so just to be clear, I'm rebasing B's repo with A's content?06:35
bob2 no06:36
→ Dave^| joined06:36
bob2 you're rebasing B/whocares on top of A/whatever06:36
you need to be more careful about "branches" vs "repositories" - repositories are red herring here06:36
→ _sundar_ joined06:37
→ leavittx joined06:37
→ jceb_ joined06:38
→ angelsl joined06:40
← jceb left06:41
michaelgamble my problem started with not having a good way to migrate old svn repos over to git if they had a history that preceded a trunk branch structure06:41
← ben225 left06:42
bob2 suspect you could have done it, but that does sound terrible06:43
svn encourages people to fuck up06:43
michaelgamble yeah lol now I'm left cleaning it up06:43
bob2 since it's just a versioned file system, branches and tags are just policy06:43
← jceb_ left06:43
→ n2diy joined06:43
michaelgamble i may just say too bad on the history prior to the proper structure06:43
bob2 well, you can do the above06:43
it's one command to fix it in git06:43
michaelgamble to be honest i wish i could just hire someone to do it :)06:44
→ fukushima joined06:45
michaelgamble the other problem I'm having is my git server is only doing pull and push's at 1mb/s on a gigabit lan06:45
jameslord git checkout -f , but there is still lots of files unmerged, why?06:46
http://iyanwu.com/pastebin/showthread.php?tid=5406:46
bob2 why on earth are you using a git server for this06:46
you do all this locally06:46
michaelgamble you talking to me or james?06:47
bob2 michaelgamble, you06:47
jameslord bob2 yes, i do all this locally:)06:47
→ robotmay joined06:48
← anildigital_ left06:48
michaelgamble software development company with 10 developers working out of svn repos, sometimes from home06:48
want to switch the office over to git06:48
← Araxia left06:48
→ anildigital joined06:48
← anildigital left06:48
→ anildigital joined06:48
← madduck left06:48
→ Araxia joined06:49
michaelgamble not certain if i really need the back history migrated, but id rather not have to maintain our crappy svn server anymore06:49
bob2 I'm lost06:49
anyway06:49
clone A06:50
fetch the branches from B into A06:50
rebase all the branches you care about in B on top of A06:50
done06:50
only the clone and fetch involve a network at all06:50
→ ThomasLocke joined06:50
michaelgamble oh yeah sorry06:51
i got more reading to do :)06:52
git is proving to be quiet difficult to wrap my head around06:52
← Slix` left06:52
← brokenshakles left06:53
← tharkun_ left06:53
michaelgamble doesn't help that its 3 in the morning06:53
→ tharkun_ joined06:53
← ScottNYC left06:54
michaelgamble hey do these instructions seem correct to you?06:54
http://stackoverflow.com/questions/2428137/how-to-rebase-one-git-repository-onto-another-one06:54
bob2 remember that the difference between 'repository' and 'branch' is critical06:55
← Textmode left06:55
bob2 but yes, that first answer is the gist of what you want06:55
michaelgamble i think I'm clear on it06:55
I'm thinking i need to follow the instructions of "A and B are not the same repo"06:56
because i don't think my SHA1 would be common?06:56
bob2 nothing to do with that06:56
michaelgamble oh ok ill try the first instructions :)06:56
bob2 the important thing to realise is that the repository crap is unrelated, except insomuch as you need everything in one repo to rebase06:56
so it's a huge red herring, and you just need to use 'git remote add' and 'git fetch' before starting the actual work06:57
→ HyperNerdV2 joined06:58
→ brokenshakles joined06:58
→ project2501b joined06:59
← tharkun_ left06:59
← igotnolegs left06:59
← jfischoff left07:01
→ madduck joined07:02
← khmarbaise left07:03
→ tharkun_ joined07:04
← cilly left07:05
→ fanno joined07:06
→ cilly joined07:07
← Radium left07:08
← tharkun_ left07:08
← gavin_huang left07:11
→ Radium joined07:12
← cilly left07:12
→ Zenith77 joined07:14
→ thelazydeveloper joined07:14
→ mcluskydodallas joined07:14
→ tharkun_ joined07:14
← pjn_oz left07:17
← Zenith77|2 left07:17
← Lgb left07:20
→ kvda joined07:20
← Zenith77 left07:20
kvda Is there any reason why my .gitmodules submodule url is getting ignored?07:21
→ Lgb joined07:22
kvda ie git submodule update --init . fails as the url is pointing to git instead of https on github07:22
→ ToBeFree joined07:23
→ whitman joined07:23
← michaelgamble left07:23
kvda ah you have to use 'git module sync'07:26
← HyperNerdV2 left07:27
← joeytwiddle left07:27
→ Jackneill joined07:28
← Jackneill left07:28
→ Jackneill joined07:28
← hwrdprkns left07:31
← mcluskydodallas left07:35
→ francisfish joined07:37
← tharkun_ left07:37
→ tharkun_ joined07:37
→ HyperNerdV2 joined07:39
→ w09x joined07:39
→ matzbitz joined07:40
← matzbitz left07:40
← tharkun_ left07:42
← kvda left07:44
← Karmaon left07:44
→ tharkun_ joined07:48
→ fanno1 joined07:48
← fanno left07:51
→ joeytwiddle joined07:51
← josefig left07:52
→ josefig joined07:53
→ ill1cit joined07:53
← josefig left07:53
→ josefig joined07:53
→ blacktulip joined07:53
→ d0k joined07:54
← Silowyi left07:55
→ arvind_khadri joined07:57
→ gavin_huang joined07:58
← robotmay left07:59
← quazimodo left08:00
← wilmoore left08:00
→ sq-one joined08:01
← tharkun_ left08:01
← wbednarski left08:02
→ mishok13 joined08:03
→ tqmz joined08:03
→ gasbakid__ joined08:05
← gasbakid left08:06
→ tharkun_ joined08:07
→ rchavik joined08:08
→ sshres15 joined08:08
← rchavik left08:08
→ rchavik joined08:08
→ _md joined08:09
→ stkrzysiak joined08:10
jameslord hi08:10
← _md left08:11
→ robbyoconnor joined08:13
→ __CA__ joined08:14
→ r0bby_ joined08:14
← robbyoconnor left08:18
→ eijk joined08:20
→ martinjlowm joined08:21
← gasbakid__ left08:21
→ gasbakid joined08:21
← p3rror left08:23
→ kmels joined08:23
← ThunderChicken left08:23
→ ThunderChicken joined08:24
← ThunderChicken left08:24
→ ThunderChicken joined08:24
← tharkun_ left08:24
→ tharkun_ joined08:24
← Vile left08:24
kmels hello, I'm trying to serve a new git repo, I've one "initial commit" but when I do `git clone` I get a warning: cloned an empty repository. Is there a step I'm missing?08:24
→ demi` joined08:25
muep did you expect the repo you cloned to not be empty?08:25
← diegoviola left08:26
kmels yes.08:28
← blacktulip left08:28
kmels I expected it to have the files i had added08:28
muep I guess you did not push the commits where you added the files08:29
git commit will only modify the local repository08:29
← Cros79 left08:29
← sq-one left08:29
kmels where should I push them? I'd like to pull from there08:30
→ kpreid joined08:30
r0bby_ → robbyoconnor08:30
← giallu left08:30
muep kmels: you should have pushed (or directly created) some commits into the repo you cloned from08:32
← tharkun_ left08:32
muep kmels: or did you just add, but not commit some files?08:32
← robbyoconnor left08:32
kmels muep, it is a new repo, I'm trying to serve from it08:32
→ robbyoconnor joined08:33
kmels i.e. git init , add some files, git add *, git commit -a -m "initial commit". Then, tried to clone it, but.. I get it empty08:33
← eijk left08:33
← stkrzysiak left08:34
→ j416 joined08:34
→ resmo joined08:35
← chihjen left08:35
← __CA__ left08:35
→ superdmp joined08:36
muep kmels: how did you clone it?08:36
→ __CA__ joined08:36
kmels I didn't clone it :) I did a new dir and initialized a new repository with "git init"08:36
→ tharkun_ joined08:38
muep kmels: I mean, how did you "...tried to clone it, but.. I get it empty"08:38
kmels muep, ah, sorry! I did `git clone <url>`08:39
← arvind_khadri left08:39
kmels it might be something with the apache setup, i'm re checking08:39
→ arvind_khadri joined08:41
← __CA__ left08:41
jameslord hi all08:42
how to commit to svn using git svn?08:42
what is the command to do this?08:42
→ cesc joined08:43
← osX64 left08:44
muep jameslord: dcommit08:44
← josefig left08:44
mastro jameslord, http://bit.ly/JIl32U08:44
jameslord muep git svn dcommit, right?08:44
muep jameslord: I guess yes, but I do not remember exactly because it has been a while since I last used git-svn to work on something08:45
← ill1cit left08:46
→ sirdancealot2 joined08:46
jameslord muep should i git commit first before i use git svn dcommit?08:48
→ Lord_DeathMatch joined08:48
← iheffner left08:48
muep jameslord: afaik yes08:49
← mishok13 left08:50
→ Evanescence joined08:52
← kcj left08:53
→ kcj joined08:54
Evanescence how to "only" add and commit deleted files changes ? here is a screenshot of my git repo status: http://ompldr.org/vZHBxbw . use git add file... one by one is very slow. I want to know is there some tricks ?08:54
← tharkun_ left08:54
Evanescence And I can't now use Zsh to complete those deleted files.08:54
→ tharkun_ joined08:54
→ _md joined08:55
→ resmo__ joined08:55
→ bulletmark joined08:57
← Mowsh left08:57
← resmo left08:58
→ osX64 joined08:58
wizbit how can i revert back a file to a previous version?08:59
i cannot see any commands like git revert08:59
→ ph^ joined09:00
→ MestreLion joined09:00
raek wizbit: that depends on what you mean by revert? do you want to undo a commit or reset some files to the state of another commit?09:00
wizbit from git log09:00
i can see: commit c708f602d etc do i use those numbers to revert back09:00
MestreLion I'm getting the following error when trying git push --mirror upstream: error: unable to push to unqualified destination: HEAD09:00
The destination refspec neither matches an existing ref on the remote nor... etc etc09:01
wizbit raek: i would like to be able to go back in time, commit by commit until i finish at the original file09:01
raek wizbit: a single file or all files?09:01
wizbit single09:01
MestreLion ordinary pushes work fine. I've used push --mirror before on other remotes. Any idea what might be going on?09:01
raek wizbit: use "git checkout <commit> -- <filepath>"09:01
wizbit ohh checkout09:02
raek generally you use checkout or reset for these activities. it's not always very intuitive which one to use09:02
wizbit <commit> = 951a52343243243 blah ?09:02
raek exactly09:02
wizbit the special code09:02
ok09:02
excellent09:03
wizbit tets09:03
j416 the special code.. is a hash.09:03
raek wizbit: this is explains this pretty well: http://git-scm.com/2011/07/11/reset.html09:03
→ simplechat joined09:03
← simplechat left09:03
→ simplechat joined09:03
→ josefig joined09:03
← josefig left09:03
→ josefig joined09:03
j416 the git main website has been updated I realised the other day09:04
very nice!09:04
good job whoever made it! :)09:04
← Evanescence left09:04
wizbit IT WORKS!!!!09:04
raek Evanescence: one way is to add the directory that contains the files: "git add .weechat/python"09:04
oh, he left09:05
wizbit there are 2 more basic things i need to learn then ill read the whole docs09:05
reset _all_ files back to a previous state09:06
and creating a new branch, ie v2 of the whole project09:06
raek wizbit: as you probably have noticed there is acutally a "git revert" command, but it does not do the same thing as "svn revert"09:06
wizbit git branch shows me the version im on09:06
oh09:06
raek (it makes a new commit that undoes the diff of a previous commit)09:07
wizbit right o09:07
k09:07
git is powerful09:07
raek indeed!09:07
← josefig left09:08
raek so, to wipe the state of the working tree09:08
→ josefig joined09:08
← superdmp left09:09
raek you can use "git reset --hard HEAD" to throw away any changes and end up in the state after the last commit09:09
wizbit interesting09:09
so that effects all files09:09
raek if you have a clean working tree, you can set it to another commit by using "git checkout <commit>"09:10
yes09:10
note that checkout and reset behaves slightly different when given a file path09:10
wizbit tests09:10
raek if you do "git reset --hard <commit>" you also "checkout" that commit, but in addition you also set the head of the current branch you are on09:11
so be careful with that one09:12
wizbit k09:12
j416 wizbit: one rule to remember: anything that is _not_ committed, can be lost if you are not careful. If you have it committed, it can in most cases be recovered.09:13
← w09x left09:13
wizbit commits himself to a mental hospital09:13
wizbit ok, so imagine i want to create project-v2.009:14
← tharkun_ left09:14
→ tharkun_ joined09:14
j416 wizbit: (meaning, of course committed files are stored, but if you screw things up (maybe you lost a commit one way or the other), commits are usually not hard to get back. Plain files are.)09:14
raek wizbit: creating a new branch is super easy: do either "git branch <name> <commit>" or "git checkout -b <name> <commit>"09:14
wizbit is that all?09:14
jeeeze09:14
and a new branch will be stored in the same head repo09:15
*hub09:15
raek the first one just creates a new branch. in the second one you check it out too.09:15
j416 what's a hub repo09:15
raek wizbit: branches are just labels for "the special codes"09:15
mac- hey09:15
← simplechat left09:15
mac- I wish to use --since 6 days 1 hour 10 minues09:16
wizbit raek: so when mythtv moved from .24 to .25 they did a: git branch 0.25 2385723758532509:16
mac- and it works when I provied 6days1hour09:16
but not in case: 6days1hour10minutes09:16
gitinfo set mode: +v09:16
j416 wizbit: you want to read the docs and understand git from the ground up. You'll be happy you did.09:16
fanno1 any sugestion's on how to structure a repo i know if is different that svn, but i cant get my mind around doing it the "best" way i am thinking /master/..... (active development) then when we create a new release we make a new branch ? eg /1.8.x/.... (only maintenance) i know this is the aprotch we would use in svn (or atleast it is the one we use) what is the best aprotch..09:16
wizbit j416: ok i will do09:17
j416 wizbit: versions are usually tagged, using 'git tag'09:17
raek wizbit: sure. but this highly depends on the workflow09:17
wizbit git clone -b fixes/0.25 git://github.com/MythTV/mythtv.git09:17
fixes is like a folder for the branches09:17
raek it's common to create a tag when you release a new version09:17
wizbit oh09:18
j416 wizbit: you've got the terminology mixed up :) read the docs.09:18
wizbit ok i will09:18
fanno1 raek: was that for me ?09:18
raek (tags are like commits, only that they should always point to the same commit and not "move around" like branch heads)09:18
fanno1 raek: (with tags)09:18
j416 sorry for telling you to read the docs but really, you _will_ be glad you did :) it will be very logical and easy to understand most things once you grok the basic concept.09:18
@ wizbit09:19
fanno1 ok09:19
raek fanno1: heh, no. but I guess my answer applies to you question as well :-)09:19
wizbit j416: the website docs are nice to read09:19
j416 wizbit: so don't skip the introduction and info on git internals, it is very useful information.09:19
← camdenx left09:19
jameslord ls09:19
j416 good good09:19
wizbit git has a nice clean looking site, i wonder if that is a cms09:19
→ xiangfu joined09:19
→ camdenx joined09:19
raek wizbit: for me everything got easier to understand after I learned how git works internally09:20
fanno1 raek: but how do you keep track of making stable builds of changes ? made to a "old" tag ?09:20
raek this explains it well: http://git-scm.com/book/en/Git-Internals09:20
wizbit would it make logical sense for a git repo to compress data?09:20
raek wizbit: they do09:21
wizbit to save space when lots of changes & bins are used09:21
ohhhhh09:21
lol09:21
→ Shusshu joined09:23
fanno1 raek: how do i seperate development/working stuff when working in same branch ? or should we branch development ?09:23
← xiangfu left09:24
← camdenx left09:24
← iamjarvo left09:24
fanno1 raek: i am having hard time wraping my head around how to use git propearly when it comes to "structure the repo"09:24
← josefig left09:25
fanno1 raek: today we have stable build in the root http://code.google.com/p/jfusion/source/browse/#svn%2Fbranches then in sub folders we have each version where atm 1.8.x is dev and 1.7.x is stable ?09:25
→ josefig joined09:26
← josefig left09:26
j416 fanno1: you would probably be interested in reading this http://nvie.com/posts/a-successful-git-branching-model/09:27
→ EyesIsMine joined09:28
→ khmarbaise joined09:30
← rchavik left09:31
← HyperNerdV2 left09:32
← j416 left09:32
→ harshpb joined09:33
fanno1 that was a good link =) he left tho before i could say thanks =)09:33
raek fanno1: the repo is just a graph of commits. you use as many branches as you need to "knit-on" new commits a the end09:34
→ xiangfu joined09:36
← harshpb left09:36
raek tags are great for pointing out important historical commits, like releases09:36
← tharkun_ left09:37
→ harshpb joined09:37
→ j416 joined09:38
→ stodan joined09:40
wizbit linus torvalds does not like github :-S09:40
cbreak ok. And? :)09:41
→ quazimodo joined09:42
→ tharkun_ joined09:42
wizbit cbreak: i treat him as king, that means i dont like github either09:43
cbreak He doesn't like OpenSUSE: https://plus.google.com/102150693225130002912/posts/1vyfmNCYpi509:43
wizbit yep09:43
or gnome09:43
← bulletmark left09:43
→ ipalaus_ joined09:49
cbreak hmm... he doesn't like C++ ... :(09:52
C++ is the best and most monstrous programming language these days!09:52
sitaram wizbit: make up your own mind; he's good and I do agree with a lot of stuff he says but the reason should not be "because he said it". This is technology, you can be rational!09:53
→ unreal joined09:53
→ _iron joined09:53
sitaram incidentally why doesn't he like github (or rather, what aspect of github doesn't he like?)09:54
(I suspect it may be stupid pull requests feature which needlessly adds an extra merge commit if you accept one)09:54
wizbit sitaram: https://github.com/torvalds/linux/pull/1709:55
sitaram bloody thing requires javascript to show the discussion09:58
cbreak I like gitolite more than github09:58
← localhost left09:58
sitaram huh? if you disable css you can actually see the whole discussion, even without JS...!09:58
cbreak sitaram: if you look at the source code, then probably as well :)09:59
→ shruggar joined09:59
sitaram yup09:59
→ pantsman joined09:59
← pantsman left09:59
→ pantsman joined09:59
cbreak stupid new "web designers" have no clue how to make html web pages these days09:59
→ joaoh82 joined09:59
sitaram wizbit: I've had this up for a long while now: https://github.com/torvalds/linux/pull/17 (see 2nd and 3rd paras)09:59
← osX64 left10:00
→ tewecske joined10:00
→ localhost joined10:01
sitaram wizbit: Linux actually went much further than I would. I don't even log into github to check for pull requests now, and certainly do not "discuss" it. One day I logged in and found a dozen or so pull requests and just deleted all of them10:04
only 2 I looked at, because the sender was someone I know; contacted him offline and dealt with it10:04
cbreak why is there a "R.I.P. Tree-Eating Monster" in the motd?10:06
→ w09x joined10:08
sitaram because they changed the logo10:10
→ simplechat joined10:10
← simplechat left10:10
→ simplechat joined10:10
sitaram cbreak: my wife (completely non techie) took one look at the new site over my shoulder and said "that does NOT look like an open source project's site"10:11
→ Superlokkus joined10:12
→ teweWORK joined10:13
→ osX64 joined10:13
← harshpb left10:14
← kcj left10:14
→ Tolnaiz joined10:14
→ kcj joined10:14
→ harshpb joined10:15
→ dzonder joined10:15
→ pfFredd joined10:16
← tharkun_ left10:18
→ tharkun_ joined10:18
← joaoh82 left10:18
→ arietis joined10:18
mac- how should I format --since string ?10:20
I wish to see all changes from 6 days 2 hours and 10 minutes10:20
ago10:21
← simplechat left10:21
← harshpb left10:22
→ harshpb joined10:23
← tharkun_ left10:23
cbreak oh, new site.10:25
← ThomasLocke left10:25
cbreak reminds me of some business page too...10:25
cmn it's not the extra merge he doesn't like (he needs to create them himself anyway), it's that a github pull request e-mail (which is what's a pull request for him) doesn't fit with the way a pull request works outside of github10:25
which isn't surprising barely noteworthy10:25
cbreak at least it's not as bad as arstechnica has become10:25
cmn: github could send emails with what ever sender address they wanted, if that's the reason for not liking it10:27
← RMapZero left10:27
cbreak technically10:27
they probably don't try10:27
→ h0bbit joined10:27
cmn the sender's address is set up so the system can post the response in the same thread10:27
mastro jameslord, is that hard to google and read^10:28
?10:28
→ tharkun_ joined10:28
→ skylamer` joined10:28
cmn github's pull requests are github-centric, which is natural10:28
← w09x left10:28
cmn but for the kernel, the actually use git in a distributed way, so a pull request doesn't do to the github repo, it goes into Linus' which he then pushes out10:29
→ w09x joined10:29
cmn it also doesn't help that there isn't even a real name associated with that commit, or an explanation10:30
→ dcorbin joined10:31
sitaram cmn: I'm not saying my reasons are the same as his; both are valid. His don't apply to me because mine is a much smaller project I suppose10:32
→ ronniy joined10:32
→ manorie joined10:32
sitaram cbreak: emails: (quoting from my own wiki page) The issues system does have an email interface, but it is not a substitute for email. I can't cc anyone else when I want to, for instance. Well I can, but any response the original requester then makes using the website will not get cc-d to the person I cc-d10:33
cmn sure, neither of your workflows agree with github, that's perfectly fine10:34
although, do you know about the refs/pull/ namespace?10:34
← tharkun_ left10:34
→ Shogoot joined10:35
← f0i left10:36
→ borior joined10:37
← harshpb left10:38
← vervic left10:38
→ nioncode joined10:38
sitaram cmn: yes; I found out one day when I did a fetch or ls-remote or something, but honestly, asking for the other guy's URL and branch name is much nicer and feels more natural10:38
cmn: one of the commentators said it best there: github is geared towards bringing in people at all technical levels and getting them coding, which is a laudable aim. As long as they don't send *me* patches :)10:39
cmn hehe10:39
sitaram there => on that linus/pull/17 thread10:39
← angelsl left10:39
→ Tinti joined10:40
sitaram honestly, *submitting* a patch via the web interface? That's like using notepad for programming10:40
cbreak they could just disable pull requests for such projects.10:40
sitaram (I didn't even know you could do that until I saw Linus ranting about it)10:40
→ tharkun_ joined10:40
cmn the initial problem is the guy who assumed github was the right place to send changes through, instead of reading the instructions10:40
the web interface is meant for simple things, not large hacking sprees10:41
← legnaleurc left10:41
→ legnaleurc joined10:42
← legnaleurc left10:42
→ f0i joined10:43
sitaram cmn: you could argue the same with notepad, but I would still fire someone who opened it up10:45
pasky omg new git homepage design10:45
that's pretty awful, isn't it?10:45
pasky wonders about resurrecting http://git.or.cz/index.html ;)10:45
cmn the github editor doesn't get confused if there's to CRs, so it's already better10:46
s/to/no/10:46
I mean, I wouldn't use it, given a choice10:47
← shruggar left10:47
cmn but if I wanted to change part of a README or some explanation in a text file for a large project... I'd be tempted, just so I didn't have to clone the repo10:48
WinterMute linus ranted about github pull requests?10:48
is there a link?10:48
raek cbreak: I don't think you can disable pull requests10:48
sitaram WinterMute: check backlog10:48
https://github.com/torvalds/linux/pull/1710:48
pasky hmm, actually, i got used to the new design pretty fast... i'm glad the tree-eating beast is gone, i was never a fan10:49
cmn WinterMute: all over the backlog, the twitterverse and you can just look at the torvals/linux repo10:49
→ giallu joined10:49
pasky the only thing really irking me is the ugly orange and very ugly logo... what's wrong with +++--- ?!10:49
← w09x left10:49
cmn text is scary10:49
or something10:49
pasky (and the LCD around current version number is maybe a bit silly, but better than the tree-eating beast ;)10:49
pasky tries to get the logo out of his mind again10:50
→ Alocado joined10:51
→ mishok13 joined10:51
Alocado hi... after upgrading my linux i always get a 401 error when i try to clone one of my repositories over https... any solution?10:53
cmn give it a password and username, as it's asking you to10:53
cbreak does your git have support for ssl?10:53
Alocado cbreak, no idea... do i have to enable it? is it not enabled by default?10:54
→ marhaban joined10:54
cbreak did you compile git?10:55
with a curl that has ssl support?10:55
j416 pasky: I think it looks much better than before, not to mention it being much more informative10:55
Alocado git version 1.7.9.5 (ubuntu 12.04 delivered packet) doesn't work.. tarball from github, selfcompiled, doesn't work10:55
j416 pasky: lots of more information -- if that information was there before (I don't know), it is now in a much more accessible format10:55
→ Zenith77 joined10:56
Alocado in the server logs i can see that git tries to request (GET) some files, but after the first ~3 GET requests there's no authentication delivered to the requests..10:56
→ angelsl joined10:57
cmn Alocado: how are you calling git? do you have .netrc set up?10:57
cbreak and that's why http is a stupid protocol for something like git.10:57
cmn is it asking you for a password?10:57
Alocado cmn, yes10:58
cmn, and no .netrc10:58
cmn which one?10:58
Alocado ... :P10:58
→ sdaone joined10:58
Alocado 123456 ;)10:58
→ w09x joined10:58
cmn so you're anwering "how are you calling git?" with "yes"10:58
Alocado <cmn> is it asking you for a password? <-- yes10:59
pasky j416: yes, that is true... as i said, after the original impulsive reaction, it's actually pretty good... except the logo ;)10:59
→ QaDeS joined10:59
j416 heh10:59
the logo is better than the other +++--- one imo10:59
Alocado calling git: git clone https://user@domain/path/10:59
j416 pasky: but, yes, if I had to complain about something, it would be that logo11:00
cmn then the password you're giving it is wrong11:00
j416 pasky: on the other hand, it is rather representative11:00
cmn if the server is behind a proxy, 1.7.10.2 just shipped a fix for proxies asking you to auth several times11:01
← sirdancealot2 left11:01
ChanServ set mode: +o11:03
canton7 changed the topic to: Latest version: 1.7.10.2 | http://git-scm.com | Central index of resources about git and this channel: http://jk.gs/git/ | R.I.P. Tree-Eating Monster | Getting "Cannot send to channel"? /msg gitinfo .voice11:03
ChanServ set mode: -o11:03
fanno1 from looking at http://nvie.com/posts/a-successful-git-branching-model/ do this look correct ? https://github.com/fanno/org.jfusion.jfusion we will be moving from svn to git i started master with out master development branch from svn (no tag created yet for that reason)11:03
Alocado cmn, i think my password is right because accessing the repository with a browser works11:04
cbreak Alocado: what do you see if you access it with a browser?11:05
Alocado because of directory index is disabled: 403. with url/config i can have a look to the config file...11:06
cmn the server disagrees with that assessment11:06
Alocado head does work.11:06
HEAD11:06
cmn what does that mean?11:06
works with what?11:06
cbreak so you say you can access it with a browser11:06
but you can NOT access it with a browser11:06
I don't get it11:06
← tharkun_ left11:07
→ tharkun_ joined11:07
Alocado cbreak, cmn assumed that my password is wrong11:08
cmn because that's what the server is telling you11:08
Alocado but with a browser i can view every existing file in the git repository11:08
cmn 401 means auth is still needed11:08
then maybe the username is wrong11:08
cbreak Alocado: so, you can show the HEAD file?11:08
Alocado cmn, i'm not an idiot...11:08
cmn but that's less likely to be the case11:08
Alocado cbreak, yes11:08
cbreak and the config file?11:08
Alocado yes11:08
cbreak and you can list the contents of the objects folder?11:09
→ cesc_ joined11:09
Alocado no (directory indexing disabled) but the git client never asks for that11:09
cmn Alocado: that's fine, I never said you were, but whatever11:09
← tharkun_ left11:10
← fanno1 left11:10
j416 fanno1: hint: if you don't know, maybe you should read the article again :)11:10
cmn and when you try to get to what git is trying to get to, namely info/refs?11:11
j416 and he left11:11
← cagedwisdom left11:11
→ solofight joined11:11
← solofight left11:11
Alocado cmn, works...11:11
cmn is curl/git creating several paraelell requests?11:11
→ rchavik joined11:12
← rchavik left11:12
→ rchavik joined11:12
Alocado cmn, here's my apache log: pastebin.com/cUZuxNju11:12
i'm not sure if it's parallel.. but there are multiple requests... but answered with 401... no idea why11:13
← cesc left11:13
cmn that doesn't show whether it is, but it looks like the bug I'm talking about11:13
upgrade11:13
← pantsman left11:14
→ tasyrkin joined11:14
cmn or stop using the dumb protocol11:15
← click left11:15
Alocado one moment please... compiling 1.7.10.211:15
← tasyrkin left11:17
← cesc_ left11:17
→ sirdancealot2 joined11:17
→ tharkun_ joined11:17
→ linusoleander joined11:19
Alocado "fatal: Authentication failed"11:19
no. stop ;) it's working now11:20
← tharkun_ left11:21
cmn not using the dumb protocol is a good idea anyway11:21
→ ColKurtz joined11:21
← nioncode left11:21
Alocado cmn, http is not a dump protocol only because git is too dump to authenticate11:22
bremner Alocado: man git-http-backend11:22
gitinfo Alocado: the git-http-backend manpage is available at http://jk.gs/git-http-backend.html11:22
bremner then you will see "dumb" and "smart" is used in a technical sense11:23
cmn Alocado: stop putting words in my mount11:23
s/mount/mouth/11:23
← khmarbaise left11:24
→ Vile joined11:25
cmn I never said to stop using http, or that it was dumb11:25
Alocado sry, i misunderstood this11:25
→ tharkun_ joined11:27
cbreak http is dumb transport11:28
because there's no git on the other end to talk with11:28
→ simplechat joined11:28
← simplechat left11:28
→ simplechat joined11:28
cbreak just a web server that has zero clue what a git repository is11:28
so you will download the whole history every fetch11:28
(unless you can identify the packs you already have)11:28
cmn http is a silly thing to use anyway, because it's stateless11:29
→ georgets joined11:29
cbreak you need to update-index to create those indexes11:29
← simplechat left11:29
cmn and the git protocol is stateful11:29
cbreak http isn't made for such things, indeed11:29
imho it's one of the most abused protocols11:29
← arietis left11:29
cmn it's all Berners-Lee's fault11:29
cbreak http is fine as transport protocol for hyper text documents11:30
← tharkun_ left11:30
cmn OK, not quite, MS also had a lot to do with making something like SSH way to complicated11:30
seem way to complicated11:30
← ToBeFree left11:33
→ tharkun_ joined11:37
← catsup left11:42
← P-l left11:42
← quazimodo left11:42
← arvind_khadri left11:44
← tharkun_ left11:45
→ click joined11:46
← mishok13 left11:49
→ oriba joined11:49
→ tharkun_ joined11:51
← Tolnaiz left11:51
→ galaris joined11:52
→ linusoleander_ joined11:52
← linusoleander left11:52
linusoleander_ → linusoleander11:52
← georgets left11:54
mastro why git fetch log to standard error?11:54
→ quazimodo joined11:54
← QaDeS left11:55
→ txomon joined11:55
← txomon left11:55
→ txomon joined11:55
→ cissu joined11:55
← txomon left11:55
← cissu left11:56
cmn because those are random strings sent by the server11:56
by default it also contains a lot or \r and is pretty useless as any sort of output that you'd like to keep11:57
← rchavik left12:01
→ madewokherd joined12:01
← tharkun_ left12:02
→ GeorgeSebastian joined12:02
← GeorgeSebastian left12:02
→ GeorgeSebastian joined12:02
→ tharkun_ joined12:02
← maleknet left12:03
← tharkun_ left12:05
j416 I want to use filter-branch to change the commit message of one single commit12:05
is it possible to do this without it affecting the other commits as well, without explicitly checking for it in the script?12:06
I have:12:06
git filter-branch -f --msg-filter 'echo "New commit message"' d857067^..HEAD12:06
I only want this new message to affect commit d85706712:06
cmn this will affect any commit that comes after12:07
how old is that commit?12:08
j416 it's just a few commits back12:08
no merges or anything12:08
→ rchavik joined12:08
j416 I don't want to use rebase, since it'll change other things, like commit time12:08
I just want to change that one message12:09
cmn why is it that important to keep the committer time?12:09
→ thiago joined12:09
j416 I'll just check $GIT_COMMIT12:09
thanks anyway12:10
cmn: because I have OCD12:10
:P12:10
cmn fine, whatever12:10
→ tharkun_ joined12:12
→ enmand joined12:14
← oriba left12:14
→ bartek joined12:17
→ martinjlowm-lptp joined12:20
→ magic_al joined12:20
← rchavik left12:20
magic_al hey all i am kind of struggle configuring my git repo. can anybody help me?12:21
i have already initialized a repo on my computer and one on my server12:22
FauxFaux magic_al: Perhaps if you ask a real quesstoin.12:22
← Vile left12:22
magic_al well: where do i get the git url of my git folder on my server and how can i set up a branch?12:22
sorry i am new with linux as well as git12:23
→ rchavik joined12:23
→ catsup joined12:24
et if it's somewhere in your home directory, it's youruser@yourserver:path/to/your/repository12:24
→ arietis joined12:25
← tharkun_ left12:26
→ thiago_ joined12:29
← thiago left12:29
gitinfo set mode: +v12:29
thiago_ → thiago12:29
← GeorgeSebastian left12:30
← enmand left12:30
← galaris left12:30
→ tharkun_ joined12:32
→ BarryCrayon joined12:33
magic_al when i init my repo on my server i get the message 'Reinitialized existing Git repository in /home/alex_fechner/www/sensenman.com/htdocs/Documents/.git/12:36
' but if i try 'git push [email@hidden.address] existing Git repository in /home/alex_fechner/www/sensenman.com/htdocs/Documents/.git/12:36
' the bash returns 'fatal: Not a git repository (or any of the parent directories): .git'. whats wrong there?12:36
i mean i try 'git push [email@hidden.address]12:37
EugeneKay You need to be within a git repository locally as well12:37
Perhaps a !repro would help12:37
gitinfo Please paste (using https://gist.github.com/ or similar) a transcript (https://gist.github.com/2415442) of your terminal session, or at least explain exactly what you did that led up to the problem. This will help immensely with troubleshooting.12:37
← Superlokkus left12:38
← BarryCrayon left12:38
← tharkun_ left12:39
magic_al EugeneKay: thanks :)12:39
← n2diy left12:39
→ Chillance joined12:40
← j416 left12:40
← quazimodo left12:41
→ tharkun_ joined12:42
← ruskie left12:42
→ glennpratt joined12:43
→ harshpb joined12:45
← twisted` left12:45
← dvide left12:47
→ ruskie joined12:49
magic_al what means the error ' ! [remote rejected] master -> master (branch is currently checked out)' after pushing files into my git repo on a server?12:49
cmn !bare12:50
gitinfo an explanation of bare and non-bare repositories (and why pushing to a non-bare one causes problems) can be found here: http://bare-vs-nonbare.gitrecipes.de/12:50
cmn it means that updating the branch ref would leave the index and worktree completely out of date12:50
and you shouldn't be using non-bare repos on a server12:50
← bartek left12:50
← carlosgaldino left12:52
← ipalaus_ left12:53
→ legnaleurc joined12:53
← legnaleurc left12:54
← w09x left12:55
→ legnaleurc joined12:55
EugeneKay Are you trying to !deploy ?12:55
gitinfo Git is not a deployment tool, but you can build one around it for simple environments. Here is an example hook to get you started: https://github.com/EugeneKay/scripts/blob/master/bash/git-deploy-hook.sh12:55
← legnaleurc left12:56
→ legnaleurc joined12:57
← legnaleurc left12:57
← harshpb left12:57
magic_al cmn: i got the difference. but now i get the error: 'fatal: The remote end hung up unexpectedly'12:58
→ legnaleurc joined12:58
→ alberto56 joined12:58
← legnaleurc left12:59
← stodan left13:00
→ legnaleurc joined13:00
→ p3rror joined13:01
→ Superlokkus joined13:02
bryanmv → bryanmv`away13:02
→ subbyyy joined13:03
← subbyyy left13:03
← _md left13:04
→ stodan joined13:06
→ fuzzybunny69y joined13:06
← thiago left13:06
← Superlokkus left13:06
→ subbyyy joined13:07
← subbyyy left13:07
← francisfish left13:08
→ ipalaus_ joined13:11
magic_al i initalized a bare git on my server. if i triy to push files i get 'fatal: The remote end hung up unexpectedly'13:11
can anybody tell me that this error is about?13:12
← `Assassin left13:12
→ GeorgeSebastian joined13:12
← GeorgeSebastian left13:12
→ GeorgeSebastian joined13:12
→ hachque joined13:17
← tharkun_ left13:17
→ tharkun_ joined13:17
→ bartek joined13:17
→ dvaske joined13:17
← jefkin left13:18
→ subbyyy joined13:18
← subbyyy left13:18
← gavin_huang left13:18
axrfnu magic_al: remote properly set on your local repository? authentication properly working? proper permissions on remote host?13:18
→ alexukf joined13:19
magic_al axrfnu: yeah i have permissions on the server. how can i set remote on my repository?13:19
← Lgb left13:20
axrfnu git remote add ...13:20
!git-remote13:20
← cyphase left13:20
→ camdenx joined13:20
→ Lgb joined13:22
← slakmagik left13:22
← Tinti left13:24
→ slakmagik joined13:24
→ gavin_huang joined13:24
← camdenx left13:25
→ Per4u3e joined13:26
← jason237 left13:27
← cenuij left13:28
← gavin_huang left13:29
→ shreya joined13:29
EugeneKay man git-remote13:30
gitinfo the git-remote manpage is available at http://jk.gs/git-remote.html13:30
→ Goplat joined13:30
axrfnu EugeneKay: thank you. wasn't sure mine would work13:31
← kcj left13:32
← kcm1700_ left13:35
→ Vile joined13:36
← rurufufuss left13:36
→ psoo joined13:36
← Vile left13:36
→ objNoob joined13:37
→ subbyyy joined13:38
← rvsjoen_ left13:39
← tharkun_ left13:40
→ kcm1700 joined13:40
→ tharkun_ joined13:40
← alexukf left13:41
→ SkAzZ joined13:41
← arietis left13:42
← OOPMan left13:42
← kcm1700 left13:42
→ kcm1700 joined13:42
→ SkAzZ1 joined13:43
→ sdaone_ joined13:43
→ francisfish joined13:43
→ simplechat joined13:44
← simplechat left13:44
→ simplechat joined13:44
→ kenperkins joined13:44
magic_al i really stuck here: i have initalized a git repo (--bare) on my server. now is there anything else i have to do on the server side?13:45
← sdaone left13:45
canton7 depends hwo your planning on accessing it. if you're using ssh, just make sure that the user who you're going to ssh as has access to the bare repo13:46
→ robin joined13:46
← SkAzZ left13:46
→ kvda joined13:47
magic_al the user is me. so you meen i have to set chmod u=rw?!13:47
Zyrkon you wanne also have your local repo where you add files and commit13:48
← subbyyy left13:48
Zyrkon to that repo, you add the remote repo (the one on your server)13:49
gitinfo set mode: +v13:49
objNoob after installing git-core .. what's next?13:49
Zyrkon then you can just push and pull from it13:49
→ edude03 joined13:50
Zyrkon given that you have the permissions on the server (depends on how you establish the connection (i.e. ssh))13:50
← fuzzybunny69y left13:50
→ neurodrone joined13:50
objNoob i installed git on my vps - and i'd like to create a new repo to try things out13:50
offby1 objNoob: read the various tutorials on github.com; they're good13:51
Zyrkon reading a good git tutorial or handbook would be a good idea13:51
→ anEffingChamp joined13:51
offby1 short answer: "mkdir x; cd x; git init; touch some-file; git add some-file; git commit -m woohoo"13:51
Zyrkon the tutorials on github are nice13:51
objNoob i've been!13:51
Zyrkon but you probably wanne have a remote repo on the server and a normal, local repo on your pc13:52
you don't need to checkout a branch on the server13:52
← flaguy48 left13:52
← stodan left13:53
objNoob ok that was my next question, my dev machine needs git, and the server needs git13:53
← bartek left13:53
→ subbyyy joined13:53
magic_al now i initalized my local repo and added the remote and get the error 'fatal: '/home/alex_fechner/www/sensenman.com/htdocs/Documents' does not appear to be a git repository13:53
fatal: The remote end hung up unexpectedly'13:53
Zyrkon you need git on any machine you wanne work with git repositories, yes13:53
→ cenuij joined13:54
← cenuij left13:54
→ cenuij joined13:54
objNoob and then i link the local and remote repositories together?13:54
Zyrkon normally you have 1 or two machines. one is your pc and you work on the files, and one is your server/backup place you push to13:54
when working alone^^13:54
← JEEB left13:55
Zyrkon of coof course you can also push to github etc. to publish your code13:55
-coof13:55
← SkAzZ1 left13:55
Zyrkon ^^ obj: yes13:55
→ mishok13 joined13:55
objNoob thanks, that's where i'll go from here.13:56
→ alexukf joined13:56
Zyrkon on your local repository: git remote add <shortname> <url>13:56
(to the remote git repository ending in .git)13:56
canton7 magic_al, so if you ls /home/alex_fechner/www/sensenman.com/htdocs/Documents you get HEAD, hooks, objects, etc13:56
magic_al yes i do13:57
canton7 so if you 'git clone '/home/alex_fechner/www/sensenman.com/htdocs/Documents' ...13:57
← rchavik left13:58
← rudi_s left13:58
magic_al canton7: same error here13:58
do i have to care about branches if i use a --bare repo?!13:59
canton7 well, the repo will have branches, just as a non-bare one will. nothing out of the orginary13:59
*ordinary13:59
objNoob one thing i'm having a tough time understanding is where the url comes into play... do i need to make a repo in the web area?14:00
on the remote?14:00
canton7 and you're running 'git clone' from the same machine as the bare repo's on? (I know that's obvious, but you'd be surprised at some of the stuff people do)14:00
magic_al canton7: no i thought i am supposed to try to clone the server's machine to my local one14:01
canton7: the bare repo is on the server and a non-bare is on my local machine14:01
cmn then use a url that tells git to grab the repo from the other machine14:02
canton7 magic_al, right. I'm not quite sure how to put this... If you direct git at '/home/alex_fechner/www/sensenman.com/htdocs/Documents on the *local* machine, it can't automatically access that location on your *remote* machine. That would be something of a security hole14:02
cmn it also looks like you're trying to give the git repo to your webserver directly; don't14:02
canton7 you have to set up some kind of transport: http or ssh. ssh is better14:02
← shreya left14:03
← GeorgeSebastian left14:03
canton7 and then you tell git to use ssh (or http) to access the repo on your remote computer14:03
objNoob ok so typically this is bad git-update-server-info14:03
→ fuzzybunny69y joined14:03
canton7 update-server-info is used for dumb http -- that's the worst of the worst14:03
cmn objNoob: that is used for the dumb http protocol, which is a bad idea14:03
objNoob and better to use ssh for transpo14:03
canton7 smart http isn't *too* bad (but a bit of a pain to set up), but ssh is better14:04
→ GeorgeSebastian joined14:04
→ traviscj joined14:04
magic_al canton7: i direct to [email@hidden.address]14:04
← simplechat left14:04
→ enr joined14:05
→ theCole joined14:06
magic_al canton7: so whats wrong? i mean i have a repo on my server and want to push files to it. why does it say the folder wouldn't be a repo?!14:07
canton7 magic_al, so 'git clone [email@hidden.address] fails with 'can't find repo'?14:08
→ karpar joined14:08
magic_al fatal: '/home/alex_fechner/www/sensenman.com/htdocs/Documents' does not appear to be a git repository14:08
fatal: The remote end hung up unexpectedly14:08
← Radium left14:09
magic_al canton7: And the folder is no protected on my server14:11
canton7: cant even ls it14:11
→ hyperair joined14:11
canton7 what do you mean, 'can't even ls it'?14:11
bryanmv`away → bryanmv14:11
magic_al canton7: 'ls Documents' returns 'no permissons'14:12
canton7 OK, and how do you expect git to be able to read it?14:12
← theCole left14:12
cmn then that'll be why you still can't read it when running git14:12
← psoo left14:12
← agile left14:13
magic_al canton7: sorry for confusion: first i tried to clone the repo from the server to my local machine. it returned the error i mentioned. and now i tried to enter the repo folder with a normal ssh connection and it tells me 'no permisson'. i just mean the clone command must have had a connection to the folder14:14
canton7 'must have had a connection to the folder'? Set up your permissions so you can ssh to the folder, then try git again14:15
magic_al canton7: chmod u=rw, or?14:16
← Jackneill left14:16
canton7 depends on how your system's set up. probably14:16
← tqmz left14:17
→ mritz joined14:18
← traviscj left14:18
← skylamer` left14:18
→ satyanash joined14:19
→ tqmz joined14:20
← satyanash left14:20
← glennpratt left14:20
→ satyanash joined14:23
→ theCole joined14:24
← kvda left14:24
magic_al canton7: all right it works now. thanks a lot. you were right. it was the permission thing :)14:26
→ xjiujiu joined14:26
canton7 magic_al, cool!14:26
← brokenshakles left14:33
→ jalama joined14:34
→ btree_ joined14:34
oddraisent Is there a way to cherry-pick single file from commit?14:35
← btree left14:38
canton7 oddraisent, git checkout <commit> -- <file>14:39
hachque Is there a way to tag certain commits along master?14:39
oddraisent М14:39
canton7: right14:39
hachque Basically I want to have something that's "beta-stable"14:39
and every now and then I can bring "beta-stable" up to the current tip of master14:39
but I don't want it to be it's own branch, because I'm not making any distinct changes on it14:39
I'm just tagging a certain commit as "beta-stable"14:40
oddraisent Is there a way to pop changes to single file from stash?14:40
hachque is there a way to do that in Git?14:40
cmn tags in git should't change14:40
why do you have master and beta-stable with seems to be some subset of master?14:40
hachque because I want people to be able to checkout beta-stable and know that they're going to get something that builds14:41
cmn is master meant to be stable or not?14:41
hachque master is unstable14:41
it breaks all the time :P14:41
cmn then create a stable branch that you don't break14:41
hachque yeah but I don't want to add seperate commits to that branch14:41
cmn you don't have to add commits to a branch14:41
← Shogoot left14:42
hachque oh okay14:42
← WebDragon left14:42
cmn you're looking for a ref that changes every once in a while14:42
that's a branch14:42
hachque so I can associate an existing commit on master with a new branch beta-stable?14:42
cmn so you make a branch point to any commit you want14:42
oddraisent canton7: wait, it's no right14:42
→ yaroslav joined14:42
← _root_ left14:42
→ nOStahl joined14:42
oddraisent git co /= git cherry-pick14:42
cmn clearly, otherwise they'd have the same name14:42
oddraisent I need to apply commit diff to current file state14:43
hachque oh okay, that sounds like what I want then, how do I go about doing that?14:43
canton7 oddraisent, I know. You cherry-pick commits. You can pull a file from a given commit into the working copy, though, which is what checkout in file mode does14:43
hachque I know how to create a branch and checkout to it14:43
but I don't know how to directly modify where a branch points14:43
cmn man git branch14:43
gitinfo the git-branch manpage is available at http://jk.gs/git-branch.html14:43
← yaroslav left14:44
cmn you'll probably end up doing a lot of git branch -f and git push -f14:44
but you might want to make your workflow somewhat saner14:44
oddraisent canton7: what for there is "--"?14:44
hachque cmn: I won't be changing beta-stable that often14:44
so it shouldn't be an issue14:44
cmn it will be if people are basing work on it14:44
canton7 oddraisent, it's a disambiguator14:45
hachque people won't / shouldn't be14:45
people should base work on master14:45
because that's the development line14:45
cmn people should base work on the branch that's broken half the time?14:45
canton7 you can always have v1.0.1-beta and v1.0.2-beta branches, for example14:45
→ rmccue joined14:45
hachque it's not broken all the time, it's just that it can be broken14:45
and for people who want to checkout the source and go cmake . && make14:46
that's not too good14:46
it's usually not too seriously broken in the worst of cases14:46
canton7 hachque, have you come across gitflow, or the basics of it?14:46
rmccue Hi there! `git ls-tree origin/master` gives a list of files, including one called "screenshot-1.png", but `git ls-tree origin/master screenshot*` doesn't give anything. Any ideas?14:46
canton7 (master is always stable and is for releases only, development work happens on the develop branch, which is merged into master?)14:46
magic_al theoretical question: what if a user pushes a file to a git server he has worked with. another user works with the same file at the same time and pushes that to the same server. does git prevent such situations?14:46
cmn for people who want a tarball to build, you should provide a tarball14:46
hachque I will assume people who know how to program also know how to checkout an earlier revision and merge if they really need it :P14:46
GitHub does provide an automatic tarball14:46
cmn then you have no worries14:47
magic_al: git doesn't push files around14:47
canton7 rmccue, git ls-tree origin/master:screenshots14:47
→ twisted` joined14:47
cmn the second dev would get told that his branch has diverged14:47
canton7 rmccue, although, globs won't work14:47
rmccue canton7: "Not a valid object name origin/master:screenshots" - man page for ls-tree says I can do <treeish> [paths...] and use glob-like syntax14:48
→ sq-one joined14:48
canton7 rmccue, yeah I onyl glanced at your question -- I though there was a 'screenshots/' folder14:49
cmn hachque: FTR, that is a crazy way to develop software14:50
specially with git14:50
magic_al is there a way to get only those files from the git server which have been modified?14:50
cmn magic_al: modified from what?14:51
git doesn't deal with files at that level14:51
are you looking for fetch?14:51
← twisted` left14:51
rmccue canton7: Unfortunately not, and the glob matching is why I'm trying to do this. Looks like I may have to grep it myself14:51
← xiangfu left14:51
canton7 rmccue, use something like git ls-tree origin/master | grep screenshot*14:52
yeah14:52
cmn or 'screenshot*' to avoid the shell from expanding the glob14:53
canton7 good point14:53
→ pyhead joined14:53
← Gitzilla left14:53
← bartman` left14:54
← theCole left14:54
magic_al cmn: fetch sound like what i was looking for but dosnt have any effect on my folder14:54
cmn exactly14:55
fetch grabs the missing objects/information14:55
what you called files14:55
if you want to merge those changes use merge14:55
rmccue canton7: I'd prefer to avoid that if I can. I'd like to find out why ls-tree isn't accepting my wildcard first14:55
magic_al cmn: yeah but it dosn't dl the files i deleted for testing14:55
cmn or set up your branch to track the remote one and pull14:55
hachque cmn: I've created the local branch pointing to the correct commit, but I forget how to set it up for remote tracking14:55
cmn it does't download any files14:55
hachque given that it doesn't exist on GitHub yet14:55
if I do git push will it be created automatically?14:56
cmn no14:56
→ _root_ joined14:56
magic_al cmn: what else does it do?14:56
cmn if it's new and you haven't changed push.default, you have to push is explicitly14:56
magic_al: what else does what do?14:56
hachque okay thanks14:56
→ vervic joined14:56
magic_al cmn: what else does fetch do if not downlaoding missing files?14:57
cmn 16:55:16 <cmn> fetch grabs the missing objects/information14:57
it's not else14:57
git doesn't download files, ever14:57
← mritz left14:57
magic_al cmn: and how can i used the objects/information to get the files?14:57
cmn checkout, merge, show...14:57
do read the !book to know how to use git to work with remotes14:58
gitinfo There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !gitt !vcbe and !parable14:58
← edude03 left14:58
→ toabctl joined14:59
→ mritz joined14:59
→ tehbam joined14:59
← lapistano left15:00
→ bartman` joined15:00
→ lapistano joined15:00
rmccue In the mean time, I'll use `git ls-files --with-tree=origin/master screenshot*` but I don't really have enough information with that15:00
→ rootfield joined15:01
cmn do you have files that match that glob in the current dir?15:01
→ Codif joined15:02
← Prime1309 left15:02
← manorie left15:02
cmn nvm, ls-tree doesn't take a pathspec15:03
→ theCole joined15:04
→ JohnyBaby joined15:04
rmccue cmn: It takes a treeish and a path pattern (not a pathspec though)15:04
cmn it takes paths15:04
what do you call a path pattern if not a pathspec?15:05
rmccue "note that this isn't really raw pathnames, but rather a list of patterns to match"15:05
It's a glob pattern, which is a subset of pathspec (iirc)15:05
← xjiujiu left15:05
cmn so it doesn't take a pattern15:06
rmccue "a list of patterns" would indicate that it does15:06
cmn a pathspec is the one with the glob15:06
→ Gitzilla joined15:06
cmn gah, read the wrong15:06
rmccue :)15:07
← _root_ left15:07
cmn but it doesn't seem to take any pattern at all15:07
→ gusnan joined15:07
magic_al cmn: i really dont get how to merge folders together. is there no feature to just get all files which are more actual on the server than on the local folder?15:07
cmn you don't merge folders, whatever those are15:08
rmccue cmn: The docs say it takes a pattern, and "subdir/*" seems to work as intended15:08
cmn you merge histories15:08
rmccue: how are you calling it?15:08
rmccue cmn: Via bash at the moment15:08
cmn and what are you telling bash to run?15:08
rmccue I've tried quoting, so shell expansion doesn't seem to be the issue15:09
→ OOPMan joined15:09
rmccue Literally, `git ls-tree origin/master screenshot*`15:09
cmn does `echo screenshot*` print anything?15:09
rmccue (Where `which git` = /usr/bin/git, `git --version` = 1.7.0.4)15:09
cmn that is ancient15:09
rmccue It echos "screenshot*"15:10
I've also tried on newer15:10
cmn but that's the one that doesn't work, isn't it?15:10
rmccue The newer version does not work either15:10
cmn the command15:10
you saif one with subdir/* worked15:10
rmccue Yes, `git ls-tree origin/master subdir/*` gives the expected result15:11
cmn and do you have a subdir with the right stuff in it on your filesystem?15:11
← mritz left15:11
rmccue Hmm, good thinking, 'sec15:12
cbreak if you want to use globs in git, make sure your shell doesn't consume them15:12
rmccue cmn: Aha, that does appear to be shell expansion15:13
cbreak quote or escape them.15:13
rmccue cmn: So, that's a red herring then15:13
cbreak: I've tried the quoted version too.15:13
→ thiago_ joined15:13
→ edude03 joined15:14
← eightyeight left15:14
→ psoo joined15:15
→ _root_ joined15:15
→ rudi_s joined15:15
→ solussd_ joined15:15
← kenperkins left15:17
→ Superlokkus joined15:17
→ cjuner joined15:18
← gebbione left15:21
→ camdenx joined15:21
→ iamjarvo joined15:22
→ glennpratt joined15:23
← camdenx left15:25
← Superlokkus left15:26
→ DofLeFou joined15:26
← tharkun_ left15:26
→ tharkun_ joined15:26
tharkun_ → tharkun15:28
→ maleknet joined15:29
→ stkrzysiak joined15:30
gitinfo set mode: +v15:31
thiago_ → thiago15:31
→ v0n joined15:33
→ klj613 joined15:34
→ kenperkins joined15:34
→ Karmaon joined15:34
→ agile joined15:36
← karpar left15:37
← jesse left15:39
← xsbeats left15:40
← martinjlowm-lptp left15:42
← Gitzilla left15:42
→ the-fonz joined15:43
→ tshauck joined15:45
→ jesse joined15:45
← ipalaus_ left15:46
← kenperkins left15:46
→ kenperkins joined15:48
← rmccue left15:49
← yrrep left15:50
← DofLeFou left15:50
← Codif left15:50
→ Vert joined15:50
← maleknet left15:51
→ Gitzilla joined15:51
← psoo left15:53
→ felgru joined15:53
→ Textmode joined15:54
→ Lulzon joined15:54
← nrichards left15:55
← objNoob left15:59
← _root_ left15:59
← _sundar_ left16:01
← robin left16:02
← GeorgeSebastian left16:02
→ airborn joined16:02
← ThunderChicken left16:02
→ traviscj joined16:02
← fisted left16:02
→ macmartine joined16:05
→ psoo joined16:05
→ Anubisss_ joined16:05
→ _root_ joined16:06
→ fisted joined16:07
→ kinetic joined16:07
→ QaDeS joined16:08
← Anubisss left16:08
→ kadoban joined16:10
→ pandeiro joined16:11
← babilen left16:11
← kmels left16:12
← anEffingChamp left16:14
→ ipalaus_ joined16:15
← jalama left16:15
→ jalama joined16:15
→ harshpb joined16:15
← az left16:15
← psoo left16:16
→ psoo joined16:16
← eicto left16:16
→ kjg joined16:17
← harshpb left16:17
→ harshpb joined16:17
← _root_ left16:18
→ snearch joined16:18
← snearch left16:18
→ snearch joined16:18
→ __iron joined16:19
→ fanno1 joined16:19
→ jceb_ joined16:19
→ chunk1 joined16:20
← chunk1 left16:20
→ _root_ joined16:21
→ carlosgaldino joined16:22
thiago → thiago_MUC16:24
← kjg left16:24
→ kjg joined16:24
← harshpb left16:25
→ harshpb joined16:25
→ kukks joined16:27
← iamjarvo left16:27
→ eicto joined16:27
← cjuner left16:28
→ juvenal joined16:30
← eicto left16:32
BinGOs So is the github repo the official source for tarballed or can I get them from somewhere else where the links aren't broken16:32
← dr0id left16:32
→ blacktulip joined16:33
← psoo left16:33
BinGOs Aha, found the code.google.com url in my firefox history.16:34
← BLZbubba left16:34
→ BLZbubba joined16:34
← sq-one left16:35
→ popcorn joined16:35
cmn for what?16:36
← jalama left16:37
← JohnyBaby left16:37
→ berserkr joined16:38
→ babilen joined16:40
BinGOs git of course16:40
→ eicto joined16:40
→ jceb joined16:40
← jceb_ left16:41
→ flaguy48 joined16:41
cmn you'd be surprised what people ask16:41
and yeah, Junio primarily uploads his tarballs to google code16:42
← thiago_MUC left16:42
BinGOs well git-scm.com used to point to that.16:42
→ thiago_MUC joined16:42
→ stodan joined16:42
BinGOs but it's gone all .... 'designed'16:43
cmn it still pionts there16:43
oh, nvm16:44
BinGOs all the links under the television are to github.com/git16:44
cmn it links in /downloads under 'older releases'16:45
→ usiegj00 joined16:45
cmn had the idea that is was more prominent16:45
→ az joined16:46
usiegj00 heya--anyone know how to simulate git-blame by navigating trees/commits/blobs directly?16:46
cmn the github tarballs will do you as nicely, only the prebuilt doc tarballs would be missing16:46
BinGOs intuitive16:46
← sdaone_ left16:46
cmn you usually get git either from your distro or via the repo16:46
← Goplat left16:48
← glennpratt left16:50
← alexukf left16:52
← jceb left16:53
← madsy left16:54
← hachque left16:54
BinGOs cool I have 1.7.10.2 installed.16:54
→ solofight joined16:54
← toabctl left16:55
← Shusshu left16:55
← cbreak left16:55
← traviscj left16:55
← Dreamer3 left16:56
← solofight left16:57
← harshpb left16:57
→ cbreak joined16:57
← Lord_DeathMatch left17:00
→ Sky[x] joined17:00
→ madsy joined17:01
← madsy left17:01
→ madsy joined17:01
thiago_MUC → thiago17:01
← n3m left17:01
→ eijk joined17:01
← MestreLion left17:03
← thiago left17:03
← smw left17:03
← nutron left17:04
→ bondar joined17:04
→ UAA1N19 joined17:04
← bondar left17:04
← gusnan left17:05
→ nutron joined17:05
→ n3m joined17:06
← popcorn left17:06
→ bondar joined17:06
← bondar left17:06
← angelsl left17:06
→ bondar joined17:06
← bondar left17:06
madsy → Madsy17:07
← tharkun left17:07
→ khmarbaise joined17:08
← tshauck left17:10
→ angelsl joined17:10
→ nrichards joined17:10
→ smw joined17:11
← nrichards left17:12
→ eightyeight joined17:12
→ tharkun joined17:13
← Alocado left17:15
← kjg left17:15
→ Radium joined17:15
← _root_ left17:15
← solussd_ left17:16
→ mikepack joined17:16
← mikepack left17:16
→ mikepack joined17:16
← mikepack left17:17
→ mikepack joined17:17
→ upfront joined17:20
← eicto left17:20
← tharkun left17:20
→ _root_ joined17:20
→ eicto joined17:20
← v0n left17:21
← __iron left17:23
→ shruggar joined17:23
→ tharkun joined17:25
→ vespakoen joined17:25
← tqmz left17:26
→ raatiniemi joined17:27
vespakoen Hey guys, I have got a project on github, and i also have a server with gitolite on it, i would like to be able to push my changes to both repositories. but i dont know how to get them "synced" first, does anybody have pointers?17:27
← teweWORK left17:28
→ Silowyi joined17:28
canton7 vespakoen, push to both? then they'll both be in sync with your local repo17:28
gitolite also supports automatic mirroring iirc, if you prefer17:28
vespakoen ok, but my github project already has a history17:28
canton7 so, clone the github repo, push to the gitolite one17:29
vespakoen canton7, that sounds good17:29
← marhaban left17:29
vespakoen actually, maybe i dont need the history on the gitolite server, i only put it there for the post-commit hook17:29
oh yeh, that sounds pretty logical17:29
et vespakoen: if you push, you get all the history ...17:29
vespakoen ok sweet, that worked (git clone + remote add + push)17:31
that was pretty stupid of me.. =P17:31
anyways, thanks a lot17:31
canton7 heh, we all have our moments :P17:31
vespakoen yeh i guess hehe17:31
← pandeiro left17:35
← berserkr left17:37
→ HyperNerdV2 joined17:39
← _root_ left17:40
→ harshpb joined17:43
→ soreau joined17:43
soreau Hey guys, running F16 here.. git send-email seems to not exist. I have git --version 1.7.7.617:44
→ _root_ joined17:44
soreau I tried git-send-email as well, this command does not exist. git --help shows no send-email command17:45
What do I need to get the send-email command?17:45
→ ology joined17:46
milki the command isnt git-send-email, its git send-email17:46
--help is only a subset of available commands17:47
cbreak git send-email is part of git17:47
it comes with the standard git sources17:47
→ glennpratt joined17:47
soreau $ git send-email --help17:48
No manual entry for gitsend-email17:48
milki git help send-email17:48
cmn some distros make send-email a different package17:48
milki man git-send-email17:48
gitinfo the git-send-email manpage is available at http://jk.gs/git-send-email.html17:48
soreau git: 'send-email' is not a git command. See 'git --help'.17:48
cmn some distros make send-email a different package17:48
milki really o.O17:48
soreau cmn: I searched the repos for 'send-email', it returned nothing17:48
cmn then maybe your distro doesn't have it17:48
EugeneKay yum search git-17:48
milki whats the distro?17:48
soreau F1617:49
cmn in Debian it's git-email17:49
EugeneKay It's probably the git-email package, but I'm looking at SL6, not F16.17:49
So YMMV17:49
cmn there's also probably a git-all package17:49
EugeneKay Yup.17:49
soreau cool, found it17:49
git-email17:49
git all yes17:49
soreau installs17:50
→ bitkiller joined17:50
soreau thanks guys17:50
EugeneKay See also man yum, look at 'provides'17:50
soreau EugeneKay: yea that doesn't help17:50
→ josefig joined17:50
← josefig left17:50
→ josefig joined17:50
soreau unless there is a separate specific file for git-send-email17:50
EugeneKay Then you probably did it wrong because `yum provides */git-send-email` comes back with git-email for me.17:51
soreau wut17:51
ah, I only did yum search17:51
EugeneKay ;-)17:52
cbreak what'd cause a distro to butcher git into pieces? :/17:52
← Chib left17:52
EugeneKay People who don't want the dependencies.17:52
git-email requires a bunch of perl modules; git-svn requires svn stuff; git-gui requires X et al17:53
cbreak you can install them anyway17:53
it's a perl script17:53
EugeneKay `yum install git-all` comes back with a list of 31 dependences above and beyond what git-core requires. That's 29MB of packages I'd rather not have if I'm never gonna use them17:54
Especially cvs/svn.17:54
cbreak you don't need the dependencies17:54
just install git without them17:54
EugeneKay Tell that to yum.17:54
cbreak and then if you need them you can install the deps later17:54
← vespakoen left17:55
← mastro_ left17:55
cbreak that thing doesn't support optional dependencies?17:55
EugeneKay Nope. Hence the splitting.17:56
cbreak lame.17:56
EugeneKay "rah rah rah red hat is stupid" etc17:56
→ objNoob joined17:57
← fanno1 left17:58
soreau yea, it sucks :)17:59
thanks again all17:59
← soreau left17:59
← nutron left17:59
→ Diemuzi joined17:59
→ Mowsh joined17:59
→ jalama joined17:59
→ nutron joined17:59
← Per4u3e left18:00
← linusoleander left18:00
→ JEntrep joined18:00
← n3m left18:02
→ cirwin joined18:02
→ n3m joined18:02
→ marhaban joined18:03
→ mastro_ joined18:04
→ ConstantineXVI joined18:04
→ _md joined18:04
← JEntrep left18:05
→ poppiez joined18:05
poppiez stupid question: in my current repo I have a structure like this: /static/js/ and /static/css/. I'd like to add a submodule in my repo with both JS and CSS files. how do I make that work with my current structure?18:06
that submodule will include both JS and CSS so what dir should it be in?18:07
→ traviscj joined18:09
→ arietis joined18:10
← tharkun left18:11
cbreak poppiez: submodules are exactly one directory18:11
poppiez cbreak: yes, exactly my problem18:12
cbreak so maybe /module/static/js and /module/static/css18:12
then you have of course two js and two cs folders18:12
but that shouldn't matter too much18:12
your build system can resolve that18:12
← harshpb left18:13
← juvenal left18:13
→ harshpb joined18:13
→ wbednarski joined18:14
poppiez cbreak: right, ok. I guess that's fine. thx :)18:16
→ tharkun joined18:16
← _root_ left18:20
milki aw, the git osx installer requires 10.518:22
← neurodrone left18:22
← francisfish left18:23
← Nibato left18:23
cmn be a man and use homebrew18:23
← harshpb left18:23
milki has never used a mac before18:24
milki i have no idea what i need to do18:24
milki googles homebrew18:24
cmn it's gentoo for mac18:24
bremner ,twb-fix18:24
oh, wrong bot18:25
milki hm18:25
i need to figure out how to copy and paste in mac18:25
o18:25
right, theres a command key18:25
cmn replace ctrnl with the odd key they have next to space18:26
→ _root_ joined18:26
milki great, ssl certs are outdated as well18:26
← marhaban left18:26
milki wait18:26
the git hub page says requires 10.5+18:27
for homebrew18:27
this is ppc18:27
cmn oh... ppc, now that's old school18:27
cbreak you could install git from source18:27
I did that in the old days, on 10.518:27
milki ok18:28
cbreak but I don't know if current git works on that18:28
and I never tried it before 10.518:28
→ lonewulftp joined18:28
cbreak at least not as far as I can remember18:28
← asakura1 left18:28
← lonewulf` left18:29
← ska-fan left18:29
→ hellopat joined18:29
→ ska-fan joined18:29
cbreak imho, 10.6 is the best OS X so far. Unfortunately, apple decided to make it worse with 10.7 :(18:29
→ Nibato joined18:29
→ toabctl joined18:30
← satyanash left18:30
← toabctl left18:30
milki o.O18:30
→ satyanash joined18:30
milki where should i set my prefix?18:31
o, theres still a user local18:31
er18:31
/usr/local18:31
milki tries that18:31
→ gusnan joined18:31
et cbreak: what exactly do you find worse with 10.7?18:31
kevlarman et: mission control is a huge downgrade from expose18:32
milki unrecognized option -pthread18:32
o.O18:32
cbreak yeah. I didn't use expose much, but the new one is worse18:32
I am mainly anoyed by the screwed up Spaces18:32
milki pthread sounds useful...18:32
cbreak (apples multi virtual screen thing)18:32
→ ToBeFree joined18:33
← ToBeFree left18:33
→ ToBeFree joined18:33
→ shoerain joined18:33
→ lonewulf` joined18:34
← Diemuzi left18:34
← leavittx left18:34
← glennpratt left18:35
← sureaint left18:36
→ meoblast001 joined18:37
milki oo18:37
abort18:37
rquires asciidoc18:37
← subbyyy left18:37
meoblast001 hi, i have a question about how git does things internally. i wonder frequently what actually happens deep within git when i do certain operations18:37
milki to run doc proboably18:37
et hm, i don't feel that way ... expose was very cluttered if you had many windows so i didn't use it (but i don't use mission control much either)18:37
milki meoblast001: ask away18:37
cbreak meoblast001: !cs explains the general structure18:38
gitinfo meoblast001: 'Git for Computer Scientists' explains the basic structures used by git, which is very helpful for understanding its concepts. http://sitaramc.github.com/gcs/18:38
meoblast001 when i do a "git mv", does git handle it internally like a move? does it include the entire contents of the file being added and then removing from the old location?18:38
→ subbyyy joined18:38
← subbyyy left18:38
cbreak et: now it resizes all windows to the same size, which doesn't even leave size clues18:38
milki it probably does an filesystem move and then update the index entry for its new location18:38
cbreak meoblast001: git mv is like mv18:39
milki the rename detection maybe or may not work18:39
cbreak followed by staging the removal at the old location and the appearance at the new18:39
← vervic left18:39
meoblast001 so internally it's adding a whole new file and all its contents?18:39
cmn yes, that's how moving files work18:40
meoblast001 eep, hm, guess i should make sure i don't excessively move things then18:40
cbreak remember: git internally stores files by their contents18:40
cmn why?18:40
cbreak that means you can have thousands of copies of your file and they won't take up much more than one18:41
cmn moving introduces less change than editing a file18:41
meoblast001 i was always under the assumption that it would just keep track of that fact that a file left a location and entered a new location18:41
cbreak no18:41
cmn git doesn't do that for any file18:41
cbreak git doesn't track changes18:41
git doesn't track any kind of change18:41
cmn it will record the new file having contents so and so18:41
cbreak the only thing git DOES track is state. Like a snapshot.18:41
when you commit, the state in the index gets associated with a commit object and inserted into history18:41
meoblast001 aah, ok18:42
perhaps i should then read up more on the internals of git18:42
cirwin !parable18:42
gitinfo 'The git parable' provides some good reasoning behind git. http://tom.preston-werner.com/2009/05/19/the-git-parable.html18:42
← zivester left18:42
cbreak if you have a file that was never in git (because you changed one single character in it)18:42
then git will store that whole new file18:42
← Rafusek left18:42
cbreak if you don't change the file, it'll just use the one it already has.18:43
→ Rafusek joined18:43
→ zivester joined18:43
meoblast001 ah18:43
→ JEEB joined18:43
cbreak and sometimes git squeezes space out of the history by doing delta compression18:43
but that's just for making history compact, it doesn't change the conceptual representation of the snapshots18:44
← Zenith77 left18:45
milki do i need to install everything by source for ppc mac?18:46
← babilen left18:46
meoblast001 cbreak: ah, thanks18:47
← shruggar left18:47
← ology left18:48
→ caseymcg joined18:49
← zivester left18:52
→ zivester joined18:53
← lonewulftp left18:53
→ subbyyy joined18:53
← subbyyy left18:53
→ subbyyy joined18:53
→ asakura1 joined18:53
→ lonewulftp joined18:54
← stodan left18:55
→ alexukf joined18:55
← ipalaus_ left18:55
← Nibato left18:57
← zivester left18:58
→ gtrak joined18:59
→ zivester joined18:59
cbreak milki: I did back then, but maybe there are binary packages18:59
git 1.4 was current I think. Maybe even 1.318:59
← zivester left19:00
→ zivester joined19:01
← Gitzilla left19:02
← zivester left19:02
→ Nibato joined19:03
← dvaske left19:03
→ zivester joined19:03
→ Codif joined19:05
← mishok13 left19:05
← syphar left19:06
→ twisted` joined19:06
→ syphar joined19:06
← poppiez left19:06
← Codif left19:07
→ bartek joined19:07
← zivester left19:08
→ joaoh82 joined19:09
→ zivester joined19:09
← bartek left19:13
→ Gitzilla joined19:15
← tharkun left19:17
← _root_ left19:17
← stkrzysiak left19:17
← carlosgaldino left19:19
← Lgb left19:20
→ Xizor joined19:21
→ ology joined19:21
→ gumnos joined19:21
← Nibato left19:21
← mwrosen left19:22
→ shruggar joined19:22
→ Lgb joined19:22
→ _root_ joined19:22
→ Ky0sh1r0 joined19:23
← Ky0sh1r0 left19:23
→ Ky0sh1r0 joined19:23
→ Nibato joined19:28
← Sigma left19:29
→ FreeArtMan joined19:29
→ carlosgaldino joined19:31
← bgerber left19:31
← sirdancealot2 left19:32
→ bgerber joined19:33
→ teweWORK joined19:33
← lonewulftp left19:34
→ enmand joined19:34
→ sq-one joined19:35
→ lonewulftp joined19:35
neersighted|AFK → neersighted19:36
→ adamm joined19:36
← kinetic left19:41
→ hwrdprkns joined19:42
→ n1x joined19:44
← jalama left19:45
← kermit left19:46
← pfFredd left19:46
← _root_ left19:47
→ kermit joined19:47
← kermit left19:47
→ Slix` joined19:47
← lonewulftp left19:50
→ lonewulftp joined19:51
→ iamjarvo joined19:53
→ _root_ joined19:54
← ronniy left19:54
→ jason237 joined19:55
← angelsl left19:55
← mikepack left19:56
→ angelsl joined19:56
→ mikepack joined19:57
→ xsbeats joined19:57
→ hwrd|work joined19:58
← mikepack left20:01
→ vespakoen joined20:01
vespakoen Hey guys, I put gitolite on my debian machine, and i made a post-receive hook, in there i do "sudo -u server /path/to/updater.sh" in updater.sh I have "cd /home/server/myproject" and "git pull" it actually works fine, but the new files have got this permission "-rw-------" is this because of the sudo -u server?20:02
files that i pulled from git "the normal way" have "-rw-r--r--" btw20:03
← enmand left20:04
← xsbeats left20:04
← snearch left20:04
vespakoen or do you guys maybe have better ideas to get my repo updated with a post-receive hook?20:04
← zivester left20:04
← hwrd|work left20:04
→ kermit joined20:06
← teotwaki_ left20:07
bremner vespakoen: !deploy20:07
gitinfo vespakoen: Git is not a deployment tool, but you can build one around it for simple environments. Here is an example hook to get you started: https://github.com/EugeneKay/scripts/blob/master/bash/git-deploy-hook.sh20:08
← raatiniemi left20:08
josefig vespakoen: wait I did the same, hold.20:08
← joaoh82 left20:09
→ mritz joined20:09
→ thiago joined20:10
vespakoen wow, a !deploy command, that is awesome =D20:10
gitinfo Git is not a deployment tool, but you can build one around it for simple environments. Here is an example hook to get you started: https://github.com/EugeneKay/scripts/blob/master/bash/git-deploy-hook.sh20:10
vespakoen oops20:11
josefig vespakoen: https://gist.github.com/2668695 my post-update file, but I use it for staging server not for deployment, I agree with bremner, for deploy you should use git archive [branch] | tar -x -C /path/to/deploy20:11
vespakoen: your problem is maybe the umask 002220:11
vespakoen yep, that is for sure20:12
it is for a test server20:12
and there are some files i added on the server itself (database config etc) so git archive wont work for me20:12
vespakoen tried umask20:12
vespakoen tries*20:12
josefig vespakoen: why not? git archive works fine.20:13
← alexukf left20:13
vespakoen josefig, it will not delete stuff ?20:13
← sq-one left20:14
cbreak josefig: there are reasons not to tar to the target location20:14
instead, tar to a new location, then change a symlink or move20:14
josefig vespakoen: what you mean? it overrides only the files which have been updated I guess, It's how is working for me.20:14
vespakoen ah okay20:14
→ sq-one joined20:14
← gumnos left20:15
vespakoen thats nice, will try it later20:15
for now, umask did the trick =)20:15
josefig cbreak: but with tar -x -C /path/to/deploy only exports the files not a tar file, or maybe I didn't follow you.20:15
vespakoen: I'm glad that worked for you =)20:15
cbreak josefig: there are two problems with doing that20:15
the first is that you can never remove files that are no longer in the tree with your technique20:16
if you write to a new location, every file gets written, but only existing files of course20:16
deleted files vanish with the previous version20:16
vespakoen i dont really care about that, testing only, so im still safe at this point =)20:16
cbreak the second is that your tar is slow20:16
you might have a new file a.whatever20:16
and an old file z.whatever20:16
and if both are used at the same time, you get into trouble20:16
josefig cbreak: oh I see your point, is good to know that, thank you ! =)20:17
cbreak if you exchange "atomically", that will not happen, or only happen much rarer20:17
well, it's mainly important for those web app things20:17
EugeneKay josefig - git-deploy-hook does archive | tar, but with rsync tacked on at the end so you can use the magic that is the --delete and --exclude family ;-)20:17
It also has some git-config goodness baked in20:17
cbreak does it atomic exchange?20:17
I think rsync can't do that20:18
← Ky0sh1r0 left20:18
EugeneKay No, but you can code that in if you like.20:18
← kadoban left20:18
cbreak I am not a web dev... I don't care about such things :)20:18
josefig EugeneKay: Oh I see :D20:18
EugeneKay Then shush and let the "for simple environments" hook be simple ;-)20:18
cbreak our build server just moves ahead with git reset --hard to which ever tag it wants to test20:18
vespakoen rsync is awesome btw20:18
→ glennpratt joined20:18
EugeneKay cbreak - FWIW, one of the things on my to-maybe-do list is a deploy.$FOO.pre and .post command20:19
Which would support atomic switching20:19
josefig EugeneKay: your to-maybe-do? are you a git developer ?20:20
← lapistano left20:20
josefig I didn't catch that.20:20
→ lapistano joined20:20
← vespakoen left20:21
← HyperNerdV2 left20:22
→ enmand joined20:22
← robbyoconnor left20:23
← syphar left20:24
← wbednarski left20:24
← enmand left20:25
← caseymcg left20:25
← frikinz_ left20:25
→ t0rc joined20:26
← t0rc left20:26
→ bsajsb joined20:28
← SegFaultAX left20:28
← _root_ left20:29
→ jarlen joined20:31
← n1x left20:32
→ _root_ joined20:33
jarlen I've build a framework I want to use for a range of different projects. I want to be able create a folder for each project and include the framework as an external git repo (so I can work on the framework in a central place independent from all projects), each project-specific code is then contained in a subdir of the framework setup in each project dir.20:34
What should I be looking for to include the framework as an external git repo in each project. Is this git submodules, or something different?20:35
← glennpratt left20:36
→ smw_ joined20:36
→ rryoumaa joined20:36
← Vinnie_win left20:36
← FreeArtMan left20:36
→ juvenal joined20:38
← Nibato left20:38
← thiago left20:38
→ marhaban joined20:39
← grobe0ba left20:40
→ francisfish joined20:41
→ Nibato joined20:44
→ HyperNerdV2 joined20:45
→ rcu joined20:45
← francisfish left20:46
→ Ky0sh1r0 joined20:46
cbreak jarlen: git submodule is great for including repositories that are not concurrently developed20:46
← lonewulftp left20:46
← _root_ left20:46
← rcu left20:46
→ rcu joined20:47
→ lonewulftp joined20:47
jarlen cbreak: and if I'll develop on the external repository and want to be able to update all the projects?20:49
← fisted left20:49
→ mikepack joined20:49
cbreak then you have to go into the repository of each project20:49
tell it to use the new version20:50
then recompile and check if the new version works20:50
and then commit the new submodule state once you have validated that it works20:50
← khmarbaise left20:50
EugeneKay josefig - no, I'm a git. I just happen to be the guy who write the hook listed in ! deploy20:51
jarlen I wouldn't expect it to auto-update :-), but will it allow me to just go into each project and pull the updates to the framework?20:52
← angelsl left20:53
→ _root_ joined20:54
jarlen I just want to make sure that it allows me to update the framework from an external repository as well as updating/pushing each project, without having to set up the git repositories each time, even though the external project is in a parent directory of the actual project20:55
← rryoumaa left20:55
→ fisted joined20:55
cbreak jarlen: you have to go into the submodule of each repository20:56
then check out a branch20:56
pull20:56
then go out20:56
do the stuff I wrote above20:57
then git add submodule20:57
and git commit with a message indicating that you updated the submodule20:57
← lonewulftp left20:57
jarlen ok, I'll look a bit more into how git submodule works exactly, sounds like it is what I need20:59
→ lonewulftp joined20:59
jarlen thank you20:59
→ thiago joined21:00
← gtrak left21:00
→ Vinnie_win joined21:00
← tehbam left21:02
← lonewulftp left21:04
→ michaelmartinez joined21:05
→ Sigma joined21:06
r0csteady bu 3121:08
→ xsbeats joined21:09
← f0i left21:11
← michaelmartinez left21:12
← Ky0sh1r0 left21:12
→ lonewulftp joined21:13
← Sigma left21:13
→ glennpratt joined21:17
→ karoshi joined21:20
← vitor-br left21:20
← lonewulftp left21:21
← glennpratt left21:21
← karoshi left21:21
→ camdenx joined21:22
← alexim left21:23
← xsbeats left21:23
← camdenx left21:24
→ teotwaki_ joined21:26
← sophacle1 left21:27
→ xsbeats joined21:28
→ sophacles joined21:28
← _root_ left21:29
→ Daemon404 joined21:32
Daemon404 is there any way to re-add an svn remote to an existing git repo?21:32
in any sane way21:32
→ pantsman joined21:33
← pantsman left21:33
→ pantsman joined21:33
→ Ripp__ joined21:35
← Radium left21:36
← Dave^| left21:36
bryanmv → bryanmv`away21:36
→ bartek joined21:37
→ glennpratt joined21:38
← lonewulf` left21:41
← sq-one left21:44
→ _root_ joined21:45
→ psoo joined21:46
→ sirdancealot2 joined21:47
← psoo left21:49
← ology left21:49
→ iadicicco joined21:50
← Xizor left21:50
→ Radium_ joined21:50
← _md left21:52
← txomon|home left21:52
→ adiabatic joined21:54
→ FernandoBasso joined21:56
→ lonewulftp joined21:56
bryanmv`away → bryanmv21:58
← quentusrex_ left21:58
← cirwin left21:59
→ pfFredd joined21:59
← Element9 left22:03
→ Beelsebob joined22:07
Beelsebob So, what's the best way to approach using git on a FreeNAS 8.2 box?22:07
← Daemon404 left22:07
Beelsebob (i.e. using it as a git server)22:08
← glennpratt left22:08
← ConstantineXVI left22:10
← sophacles left22:11
← blacktulip left22:11
← lonewulftp left22:11
→ sdaone joined22:11
cmn install git on it22:12
→ sophacles joined22:12
← _root_ left22:12
cmn Beelsebob: and then take a look at !gitolite22:12
gitinfo Beelsebob: Want to host as many git repos (and users!) as you like, on your own server, with fine-grained access control? You want gitolite: https://github.com/sitaramc/gitolite - Documentation: http://sitaramc.github.com/gitolite/master-toc.html22:12
→ zivester joined22:12
Beelsebob wait… I fail22:12
please kindly ignore me22:12
I had git on the brain, and tried to join #freenas22:12
cmn okay...22:13
← matthiaskrgr left22:13
← Vert left22:14
→ lonewulftp joined22:16
← martinjlowm left22:17
← HyperNerdV2 left22:18
→ _root_ joined22:19
→ pfFred joined22:19
← _iron left22:19
← pfFred left22:19
← twisted` left22:20
← pfFredd left22:21
← lonewulftp left22:23
→ twisted` joined22:24
→ _huoxito joined22:26
mjessome_ → mjessome22:28
← whitman left22:29
← smw_ left22:31
← rcu left22:31
← d0k left22:32
← meoblast001 left22:33
← ToBeFree left22:34
← arietis left22:34
← fuzzybunny69y left22:36
→ cirwin joined22:37
← thiago left22:38
→ kcj joined22:39
→ rvsjoen joined22:40
← _root_ left22:45
→ cybersphinx joined22:47
← mikepack left22:49
→ mikepack joined22:50
← the_cybersphinx left22:51
→ _root_ joined22:51
← mikepack left22:54
EugeneKay !botsnack22:55
gitinfo Om nom nom22:55
← Beelsebob left22:55
→ lonewulf` joined22:56
← pyhead left22:56
← ph^ left22:58
→ lonewulftp joined23:00
→ fuzzybunny69y joined23:01
→ marhaban1 joined23:01
← fuzzybunny69y left23:01
← p3rror left23:02
← marhaban left23:04
← josefig left23:04
← gusnan left23:05
← Gitzilla left23:07
← mritz left23:08
→ igotnolegs joined23:09
← lonewulftp left23:10
→ linusoleander joined23:10
← iamjarvo left23:10
← Myhro left23:10
→ kinetic joined23:11
← traviscj left23:11
← enr left23:14
→ mxweas_ joined23:14
← mxweas_ left23:15
→ p3rror joined23:16
← _root_ left23:17
→ lonewulftp joined23:19
→ joaoh82 joined23:19
← eijk left23:23
→ camdenx joined23:23
→ stkrzysiak joined23:26
← subbyyy left23:26
→ tvw joined23:27
→ Goplat joined23:27
← camdenx left23:27
→ josefig joined23:29
→ _root_ joined23:29
TomyLobo i'm used to using "head" instead of "HEAD" since i usually work on windows23:31
cmn that's a bad habit23:31
lower-case head could mean any number of things23:31
TomyLobo however, on linux i *have* to use HEAD, which is harder to type and hard to get used to for me now. can i change that somehow?23:31
cmn you can't, that's its name23:32
TomyLobo so could upper-case HEAD, cmn, if i really mean to23:32
cmn HEAD in git is a well-defined term23:32
head less so, though it often means branch tip23:32
if you want to use the word "HEAD" to mean other things, that's your problem23:33
TomyLobo no i dont23:33
i want head to mean HEAD23:33
basically23:33
what HEAD does, i don't care. i dont use it23:33
i'd rather type out the branch name instead of using the shift key for HEAD23:33
cmn there's caps lock23:34
but head doesn't mean HEAD23:34
TomyLobo i'd rather type out the branch name instead of using the caps lock key for HEAD23:34
→ Gitzilla joined23:34
← stkrzysiak left23:34
TomyLobo head doesnt mean anything on linux23:35
except if i have a directory of that name23:35
cmn it does in git23:35
TomyLobo i'm talking about git23:35
← lonewulftp left23:35
TomyLobo what does it mean there?23:35
cmn the don't talk about its meaning on Linux23:35
TomyLobo ok let me rephrase23:35
→ lonewulftp joined23:36
TomyLobo lower-case head doesnt mean anything in git on linux, except if i have a directory/branch/tag/whatever of that name23:36
at least my git doesn't know it23:36
cmn lower-case head is often used instad of tip when talking about branches23:37
→ mritz joined23:37
TomyLobo yes when *talking* about it23:37
← zivester left23:37
cmn when else do you use words?23:37
TomyLobo i don't *talk* to git now, do i?23:37
cmn yes, you do23:37
TomyLobo i don't have a habit of talking to inanimate things23:37
→ zivester joined23:38
TomyLobo but we're drifting off into technicalities again23:38
→ txomon|home joined23:38
TomyLobo what do you mean by "often" and do you actually mean the keyword "head"?23:38
→ fuzzybunny69y joined23:38
TomyLobo on a command line that starts with "git"?23:38
SethRobertson In lower case no, in upper case yes23:39
TomyLobo that's what i'm saying. cmn tries to confuse me by claiming otherwise23:39
→ kvda joined23:39
SethRobertson no (19:37:01) cmn: lower-case head is often used instad of tip when talking about branches23:39
TomyLobo yes and then he went on and told me i'd be "talking" to git23:40
← borior left23:40
SethRobertson And so you do. In a very technical language23:40
cmn and now you're restricting the use of "talk" so you can claim my arguments are invalid23:41
you communicate with git23:41
otherwise, it won't do anything23:41
TomyLobo i dont claim your argument invalid. you havent made any23:41
← Gitzilla left23:42
cmn fine23:42
← sophacles left23:42
TomyLobo you havent even made a statement to discuss23:42
cmn fine23:42
→ sophacles joined23:43
→ pushkar joined23:44
TomyLobo is there any git command that will accept lower-case "head" in the same place where it'd accept an upper-case "HEAD", except for log messages and such?23:45
← shruggar left23:46
iadicicco well if you have another ref named "head." like a tag or branch or something.23:46
but then it's a different ref.23:47
probably.23:47
TomyLobo and other than that?23:47
iadicicco there will never be a situation where head and HEAD always mean the same thing23:47
TomyLobo if not, is there any other harmful effect to be expected from mapping head to HEAD?23:48
cirwin TomyLobo: you'll confuse yourself when using systems that don't have that23:48
TomyLobo i can cope with that :)23:48
cirwin good good23:49
iadicicco if you have a branch called "head"23:50
then you wouldn't be able to use it23:50
TomyLobo i wouldnt be able to use it on windows either23:51
← kvda left23:51
EugeneKay You'll need to have your head checked23:51
TomyLobo so i dont think having a "head" branch is a valid situation anyway23:51
← tvw left23:52
TomyLobo unless you want to specifically avoid having your branch checked out on windows23:52
← fuzzybunny69y left23:52
→ subbyyy joined23:52
← lonewulftp left23:53
→ fuzzybunny69y joined23:53
→ Gitzilla joined23:54
→ lonewulftp joined23:54
EugeneKay TomyLobo - FWIW, 'ln -s .git/HEAD .git/head'23:54
TomyLobo that's bound to break somethind :D23:55
← Chaser left23:55
EugeneKay `git rev-parse head` and `git rev-parse HEAD` come back the same. !weasels apply.23:55
gitinfo The consequences of this proposal are not well-defined. A band of furious weasels may infest your undergarments, or it might work just fine. You should !backup then !tryit and let us know what happens.23:55
TomyLobo -d+g23:55
TomyLobo gets weasel-proof pants and tries23:56
milki o.O23:57
→ iamjarvo joined23:57
milki is this weasel reference from a mailing list post or something?23:57
EugeneKay Old saying of mine.23:58
milki aha23:58
EugeneKay Undoubtedly some pop culture reference from my childhood which I have long forgotten23:58
→ pandeiro joined23:58
EugeneKay Ah, I know what it was - Top Gear.23:58
← lonewulftp left23:58
→ ConstantineXVI joined23:58
EugeneKay Damned if I can remember what episode, though.23:59
TomyLobo well that didnt work23:59
cmn well, no the ln command is wrong, it makes head point to .git/.git/HEAD23:59
TomyLobo that's because ln is kinda literal in its interpretation23:59

Logs Search ←Prev date Next date→ Channels Documentation