IRCloggy #git 2011-03-28

Logs Search ←Prev date Next date→ Channels Documentation

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

2011-03-28

steven cirwin: so you're saying, implementing `git merge` (even a basic, rudimentary version which deals with trivial-cases) is non-trivial and difficult?00:01
cirwin it's certainly non-trivial. Given that it's been solved so many times, a basic version probably isn't difficult, but solving it right certainly is.00:03
← Chillance left00:03
→ SirFunk joined00:03
cirwin you could try playing with operational transforms, they're a fun way to start00:04
steven i was hoping i could just shell out to `merge` :)00:04
cirwin zsh: command not found: merge :)00:04
if it exists, sure00:04
steven it certainly exists, but i cant figure out where it comes from00:05
comes default on os x i guess. or maybe it was installed with dev tools00:05
dunno00:05
cirwin writing merge drivers is probably more fun than re-implementing git00:06
→ tgunr joined00:08
→ g0bl1n joined00:10
← tjholowaychuk left00:10
→ __sorin_1 joined00:11
← TiBook left00:14
Fuu` → Fuu00:23
→ TiBook joined00:23
← oriba left00:25
→ cutmail joined00:27
→ mattyb joined00:27
→ dAnjou_ joined00:34
← dAnjou left00:37
→ jtshs256 joined00:37
dAnjou_ → dAnjou00:37
← __sorin_1 left00:38
→ __sorin_1 joined00:38
← ping-pong left00:39
← jtshs256 left00:39
← __sorin_1 left00:41
→ __sorin_1 joined00:42
→ shepherd joined00:47
← shepherd left00:47
→ shepherd joined00:47
→ ping-pong joined00:47
← __sorin_1 left00:47
→ bjtitus joined00:51
← erichammond left00:53
→ shepherd_ joined00:54
← hyperair left00:55
← shepherd left00:57
→ xiangfu joined00:57
→ msponge joined00:59
← msponge left00:59
← cutmail left01:00
→ marcamilly joined01:00
→ nazgul101 joined01:01
→ snover joined01:02
→ __sorin_1 joined01:02
→ alkavan_ joined01:03
snover Is there a git merge flag that allows you to selectively choose changes to merge in, like git add -p? I’ve read the man page for merge and didn’t see anything like that immediately.01:03
→ Remixman joined01:03
bob2 merge merges whole branches01:03
← nazgul102 left01:04
→ zorzar_ joined01:04
→ cutmail joined01:04
snover so…is that a no?01:05
_null github lists the command "git remote add origin [email@hidden.address] - but when i enter that command into my shell it says that the -u switch doesn't exist01:06
← zorzar left01:06
_null oh my bad, i meant git push -u origin master01:07
snover _null: get a newer version of git :)01:07
bob2 -u is reasonably new01:07
_null *sigh* shared hosts!01:07
← mattyb left01:07
_null i'll send in a ticket to my host and ask them if they'd consider upgrading their fleet. if not then it's not an essential feature i suppose01:08
snover it’s mostly just a convenience01:08
bob2 heh01:08
_null i'm stilly fairly new to git so i'm still learning the basics01:08
← theoros left01:08
← __sorin_1 left01:09
→ ymasory joined01:09
→ theoros joined01:09
→ lee__ joined01:09
snover _null: -u just adds upstream tracking references automatically; you should be able to add those with git remote01:10
← ymasory left01:10
→ ymasory joined01:10
→ cpugeniusmv joined01:10
_null okay :)01:10
← shepherd_ left01:11
← ymasory left01:11
→ ymasory joined01:11
← ymasory left01:11
→ TESTBROTEST joined01:12
← lee__ left01:12
← TESTBROTEST left01:12
→ lee__ joined01:12
← lee__ left01:13
→ erichammond joined01:18
← juvenal left01:18
→ juvenal joined01:18
← alkavan_ left01:19
← gretch left01:23
→ ptl joined01:24
← ptl left01:24
→ ptl joined01:24
→ smokex joined01:26
← adaptee left01:28
← larsl left01:29
→ rayners joined01:29
→ msponge joined01:32
→ boombatower joined01:33
← boombatower left01:33
→ boombatower joined01:33
steven cirwin: re-implementing (the basics of) git is fun in the same way writing a linked-list for the first time in C is fun01:33
<301:33
← snover left01:33
← rayners left01:36
avdg is tending to agree01:37
cirwin bleargh the basics of git is far too big to be able to do properly, that would just annoy me :p01:38
I have implemented a few linked lists, 'tis true01:38
avdg just the sorting of the tree is currently something I need more information though01:38
TiBook linked-lists in C *are* fun. And easy.01:38
avdg I'm not programming in c :p01:39
← sftp left01:40
→ rayners joined01:42
← urkud left01:43
→ jason237 joined01:44
→ panike joined01:47
← panike left01:49
→ panike joined01:50
→ michael joined01:52
michael → Guest1551801:52
→ michael_lf joined01:52
→ __sorin_1 joined01:53
← g0bl1n left01:54
← gebi left01:55
→ vitor-br_ joined01:55
← vitor-br left01:55
← panike left01:56
← __sorin_1 left01:57
→ __sorin_1 joined01:58
← cutmail left01:58
→ panike joined02:00
→ cutmail joined02:01
← __sorin_1 left02:01
→ gebi joined02:01
← nchaimov left02:02
→ alester joined02:03
→ smolyn joined02:03
← engla left02:03
→ gretch joined02:04
→ nchaimov joined02:04
← panike left02:04
← rustico left02:05
← smolyn left02:06
← flaguy48 left02:07
← cutmail left02:08
← Bass10 left02:13
→ cutmail joined02:14
← cutmail left02:14
→ cutmail joined02:15
← owst left02:15
→ dvaske joined02:15
→ tyteen4a03 joined02:15
← kukks left02:15
→ tanuki joined02:16
tanuki Is there any way to take the last N commits in a given branch and move them to a new branch?02:17
← wilmoore left02:17
tanuki Basically, I've realized that the last commit I made, and the changes I've made since that last commit, might not be a wise idea and probably should be in their own branch.02:17
→ wilmoore joined02:17
avdg tanuki: cherry-pick?02:18
cirwin tanuki git branch <new branch name>; git reset <where you want this one to be>; git checkout <new branch name>02:18
← guireisc left02:19
← avdg left02:19
← dvaske left02:19
→ bindaas joined02:20
→ Schmallon joined02:21
→ blaenk joined02:21
→ intripoon joined02:23
→ Yuuhi``` joined02:24
← Yuuhi`` left02:26
← intripoon_ left02:26
← sagsousuke left02:29
→ WakiMiko joined02:30
← Elfe left02:31
→ Elfe_ joined02:31
→ dvaske joined02:32
← jungejason left02:33
→ hyperair joined02:34
← WakiMiko_ left02:34
← javaanse_jongens left02:34
← TiBook left02:36
→ G__81 joined02:38
→ javaanse_jongens joined02:39
G__81 i have made a change in my linux kernel code and have committed the same in my git branch and have taken a patch too now i need to delete that and recreate a patch how do i delete the changes that i committed into my repo02:40
is that possible ?02:41
→ bdiu joined02:41
→ dr_ick joined02:43
SethRobertson G__81: have you pushed? (seems unlikely)02:44
→ smolyn joined02:44
← blaenk left02:44
→ pgib joined02:44
pgib is there a way to get diff to show a diff of /all/ changes; I mean, staged, not updated, and untracked?02:45
← Remixman left02:45
G__81 SethRobertson, No i have pulled the kernel repo and i have just committed the changes and i used git format-patch and git send-email and sent the patches too.Now i need to revert those changes and recreate an other patch02:45
SethRobertson, Since i am gonna create an other patch i don't want the older stuff in my branch so i need to create a new patch.02:46
I did the following git reset HEAD <File_Name>02:47
but i still see the commit ?02:47
cirwin pgib: git diff HEAD02:47
→ makubi_ joined02:47
SethRobertson G__81: git reset HEAD SHA02:47
cirwin of git diff --cached02:47
not sure exactly what you're asking for, man git-diff :)02:47
jast the 'git-diff' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-diff.html [automatic message]02:47
SethRobertson G__81: If you use gitk, right mouse on a commit and you can see the a reset branch to here option02:47
G__81: Read the man git-reset option to see what --soft --hard --mixed etc option you might want02:48
jast G__81: the 'git-reset' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-reset.html [automatic message]02:48
SethRobertson probably --soft02:48
→ jsick3 joined02:48
pgib yeah I did man git-diff ;) basically, if you do 'git status' it shows changes in 3 categories: staged, changed by not updated, and untracked. I was looking for a diff of all that data.02:48
jast the 'git-diff' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-diff.html [automatic message]02:48
pgib I guess I can just commit everything in 3 commits, take the diff across that and then roll them back out02:49
SethRobertson git diff HEAD02:49
Well, not untracked02:49
pgib see, I started working on something and realized I'm doing it wrong, and it is easier for me just to rewrite the feature while looking at the old diffs. I have stuff I haven't committed yet, and stuff that is untracked. I was avoiding clumping that together, but I can just make a junk-commit just to please the diff02:51
→ jutaro1 joined02:52
pgib I suppose i can also just look at those files, no real reason for a diff now that I think about it ;)02:52
← jutaro left02:53
G__81 SethRobertson, I did the following and it solved the problem git reset --hard HEAD~202:54
→ msponge_ joined02:54
G__81 02:54
SethRobertson, My Repo is now set to the old Head. Is that right ?02:54
I had 2 commits and i wanted to remove them so that now if i make a git format-patch i wanted it to be out of the base changes02:55
SethRobertson It is set to *a* old head. Only you can judge if it is the correct one02:55
← bdiu left02:55
SethRobertson Yes, if you added two that removed two02:55
G__81 Yeah coz now i need to resend the patch and i need to make a diff from the original version02:56
SethRobertson If you plan on doing this a lot, you might want to investigate stacked git and/or quilt or one of the other patch systems on top of git02:56
← msponge left02:56
msponge_ → msponge02:56
G__81 SethRobertson, Ok thanks will do that too but now since my commits are removed its just nothing but the original version of the code so i could still use git pull to fetch the latest changes from the repo and that would still continue to work right ?02:57
SethRobertson You could do that without resetting.02:58
→ ymasory joined02:58
G__81 yeah coz now i did a git pull it says already up to date i am not sure whether there are any changes that got in02:59
← p4tux left02:59
SethRobertson You can look at the reflog. It would tell you.02:59
If you did a `git pull --rebase` it would update the branch and leave your changes at the end, as if you reset, pulled, and the reapplied them.03:00
← dkasak left03:00
G__81 oh ok thanks03:01
→ TheUni joined03:02
← erichammond left03:02
→ mikepack joined03:02
Davey Has anyone seen an expansion of the NVIE branching model that accounts for multiple concurrent stable versions?03:04
← msponge left03:04
G__81 SethRobertson, I am sorry to bother you on this how do i make sure that my repo is ok coz i have a small doubt. Though its pointing to the right commit which was the original but yesterday's changes are not though i see some mails saying "Changes Applied" in the ML i am not sure whether those were pushed in by the kernel guys and whether my repo is unable to pull it03:07
is there some way where i can cross verify them ?03:07
SethRobertson dif against @{upstream}03:08
Davey if you have commit shas, what about git branch --contains sha03:08
?03:08
SethRobertson `git fetch` first if you are paranoid. `git remote update` as well if you are ultra paranoid03:09
tbranyen i am paranoid about being paranoid03:09
← sterlo left03:09
→ sdboyer joined03:10
G__81 SethRobertson, its still the same tried git getch and git remote update the git log still shows the original commit i safely assume that there are no changes pushed in hopefully03:10
SethRobertson BTW, just because someone committed your changes doesn't mean a maintainer did, and a maintainer committed your changes doesn't mean that Linus did. It is a process03:10
G__81 I was talking about git diff. Run git-log against origin/<branchname>03:11
Or use `gitk --all --date-order`03:11
← dstanek left03:12
→ babbos joined03:12
babbos i need to do a 'pull' from a python script... do i have to change my working directory to the git repo i want to pull to?03:13
G__81 SethRobertson, yeah it still shows the same thing so i feel it should be safe and right03:14
babbos is there a way i can have the path to the repo i want to pull to?03:14
SethRobertson babbos: man git. Among other things, $GIT_DIR03:14
jast babbos: the 'git' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git.html [automatic message]03:14
→ rustico joined03:14
← sdboyer left03:14
→ blaenk joined03:17
babbos trying it on a directory back from the git repo: git --git-dir=./myrepo pull ~/git/myrepo.git/ master03:17
blaenk I forget, is there a way to 'export' a repository?03:17
babbos won't work: fatal: not a git repository './myrepo'03:17
SethRobertson git archive03:17
blaenk03:18
blaenk that is 'clone' a repository without it being under git versioning (I guess no .git/ )03:18
oh thanks SethRobertson03:18
bob2 babbos: what is that meant to do03:18
babbos blaenk http://stackoverflow.com/questions/160608/how-to-do-a-git-export-like-svn-export03:18
blaenk thank you babbos03:18
babbos as i said, i am working through a python script that runs in an irelevant to my repo directory... from there i want to do a 'pull' command03:18
so i need to 'pull' without my working directory being the repo i want to pull to...03:19
SethRobertson babbos: You might need to set --work-tree as well03:19
← kd4joa left03:19
babbos just tried that indeed03:19
nope03:19
SethRobertson Though `cd foo && git bar` seems to be pretty easy03:20
babbos check it for yourself03:20
frogonwheels babbos: without looking too closely, possibly look at man git-fetch and man git-archive03:20
jast babbos: the 'git-fetch' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-fetch.html [automatic message]03:20
bob2 subprocess.Popen takes a cwd arg anyway03:20
so no need to chdir03:20
← freedrull left03:20
SethRobertson babbos: I just tried with both git-dir and work-tree set and it worked03:21
Remember git-dir needs to be ..../.git03:21
babbos seth thanks, i must be doing something wrong then... rechecking03:21
← dv310p3r left03:21
babbos ohhh03:21
that was it03:21
:)03:21
didn't catch the /.git detail, now it works! thank you Seth!!03:22
← vitor-br_ left03:23
→ seivan joined03:23
← gebi left03:24
SethRobertson Why don't they put spam checking on the git mailing list. It isn't that challenging03:24
→ Boggle joined03:25
→ noahcampbell joined03:26
← smolyn left03:26
→ leo2007 joined03:27
→ amondracorp joined03:29
→ cutmail_ joined03:31
← cutmail left03:32
→ helgikrs joined03:37
← ajt left03:37
→ deekayen joined03:38
→ warthog9 joined03:40
→ Error404NotFound joined03:40
← Error404NotFound left03:40
→ Error404NotFound joined03:40
→ jungejason joined03:41
← deekayen left03:41
← arunce left03:45
← maestrojed left03:47
← babbos left03:47
← Modius left03:47
← rustico left03:49
← unreal left03:53
→ unreal joined03:53
→ RichardBronosky joined03:58
→ towski joined03:59
→ Bonkers joined03:59
Bonkers is there a shorter way to do "git pull <remote> <branch>:<branch>" to pull down a remote branch that doesn't yet exist locally?03:59
→ bdiu joined04:00
← dr_ick left04:01
← cutmail_ left04:02
← Twey left04:02
cirwin Bonkers: git fetch gets all branches, then git checkout <branch> will do the right thing (on modern git)04:03
→ ratmice_ joined04:04
← Laggg left04:04
← cpugeniusmv left04:05
Bonkers is 1.7.4.1 modern? what exactly are you checkout out? refs/remotes/<remote>/<branch> leaves a detached HEAD, and git checkout <branch> just says no such branch04:05
← noahcampbell left04:05
Bonkers fetch only gets the tracking branches04:05
← OOPMan left04:05
← sunblush left04:06
← rayners left04:06
→ cutmail joined04:07
→ Modius joined04:07
→ koo6 joined04:10
→ Theravadan joined04:10
← javaanse_jongens left04:10
← unreal left04:11
cirwin git fetch with no arguments gets all the remote branches for me. git checkout <branch> detects the presence of a remote branch with the same name and checks it out with a new local branch04:11
→ djbpython1 joined04:12
cirwin 1.7.4.1 is the same version I'm using04:12
maybe it only works for origin04:12
Bonkers oh, it's likely because I have two remote branches by the given name04:13
I wonder by "git pull <remote> <branch>" doesn't automatically create a local branch04:14
djbpython1 Hi all, I need some hand holding, or a link to a guide. I have a local git repo, I'd like to clone it on my shared hosting server (which is already set up for git), what do I need to do that?04:14
← cutmail left04:14
Bonkers git clone <ssh_path>04:14
cirwin djbpython1: you need a git server that your shared hosting server can access — either set up dynamic DNS so it can access your local computer, or push via github04:14
→ cutmail joined04:15
→ unreal joined04:15
djbpython1 cirwin: yes, it is set up to be a git server, I think I need the push option04:15
cirwin what does "be a git server" mean?04:15
djbpython1 but i can't just say 'git push my.server.com'04:16
cirwin you can do that too, but first you need to do "git init" on my.server.com04:16
Bonkers djbpython1, you can, but you probably do'nt want to04:16
→ TiBook joined04:16
djbpython1 cirwin: I am supposedly able to push to it via ssh/http04:16
Bonkers are you just trying to deploy it?04:16
← G__81 left04:17
Bonkers what's your goal here?04:17
djbpython1 nope, I'm trying to have a place to put my code so my other computer can clone the repo04:17
so the two of them can push and pull between my server04:17
← dvaske left04:17
Bonkers central repo? you'll want to create a bare repo on the server04:17
then push to it04:17
→ tgunr_ joined04:18
Bonkers then add the server as a remote to both local repos04:18
you'll likely want to read up and google this stuff since you'll be much happier if you actually understand all of this04:19
djbpython1 ok, ill have to google this more, but nothing solid was coming up04:19
Bonkers it's not really all that complicated04:19
hopefully 'bare repository' will lead you in the right direction04:19
← tgunr left04:19
Bonkers and possibly consider using github since it's pretty much pure awesomeness04:19
djbpython1 like i'd expect this guide to point me in the right direction -> http://www.kernel.org/pub/software/scm/git/docs/everyday.html04:20
but it starts talking about origins and motherships and kind of confuses me04:21
→ dr_ick joined04:22
→ OOPMan joined04:23
Bonkers that's likely not the most helpful of guides...04:24
pretty terse04:24
I actually learned from the oreilly git book04:24
← xxneolithicxx left04:24
Bonkers totally worth it in my opinion, but you can likely find a middle ground between that kernel.org guide and the oreilly book04:24
to get you going, on the server do:04:25
mkdir foo.git; cd foo.git; git init --bare04:25
then in the local repo:04:25
git remote add origin <ssh_path_to_foo.git>04:25
djbpython1 what does origin mean?04:26
Bonkers origin can be anything04:26
just a name04:26
but normally if there's a sole remote in a repo, it's called origin04:26
djbpython1 ah04:26
Bonkers now I'm not quite sure how to push all local branches to the server, you can certainly push individual ones with04:27
git push origin <branch>:<branch>04:27
just 'git push origin' may push all local ones, but I'm not sure04:27
then to get the repo on the other machine, it's much easier, just: git clone <ssh_path>04:27
that will automatically pull down everythign and setup the remote04:28
← wilmoore left04:28
→ shiba_yu36 joined04:28
→ zoo-zed joined04:28
← zoo-zed left04:29
→ kushal joined04:29
→ wilmoore joined04:30
Bonkers did that get you somewhere?04:30
→ __sorin__ joined04:30
← ratmice_ left04:30
kushal I have few patches created using "git format-patch" command , now I want to apply those patches and also want to make sure that the original author's name and commit message should stay the same, how can I do that ?04:31
→ aalex joined04:31
frogonwheels kushal: man git-am04:31
jast kushal: the 'git-am' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-am.html [automatic message]04:31
→ __sorin_1 joined04:32
frogonwheels kushal: isn't it obvious from the name? :) not.04:32
kushal frogonwheels, jast I already read that, I am confused as I am having the patches as in files not in a mailbox04:32
Bonkers heh, well the git format-patch page does mention git-am04:32
frogonwheels kushal: just specify the files. it works.04:33
kushal: for eg: git am *.patch04:33
← __sorin_1 left04:33
kushal frogonwheels, ok, trying again04:33
djbpython1 Bonkers: think im just going to set up a repo on unfuddle04:33
frogonwheels kushal: the default format for format-patch is one that is accepted by am basically.04:33
s/am/'am'/04:33
Bonkers djbpython1, won't be any easier, but unfuddle isn't a bad choice, github has some much cooler features though04:34
← dAnjou left04:35
← amondracorp left04:35
→ dAnjou joined04:35
djbpython1 cant beat it on price04:36
Bonkers ya, definitely04:36
github is free for public repos though04:36
→ EugeneKay joined04:36
djbpython1 indeed04:36
→ __sorin_1 joined04:36
EugeneKay Out of a perverse curiousity, what is the largest known git repository, in GB / files tracked?04:37
→ harshu joined04:37
kushal frogonwheels, worked this time, thanks :)04:38
→ jsquared joined04:40
nevyn EugeneKay: kde4?04:41
EugeneKay How big? :-p04:41
← xiangfu left04:42
EugeneKay I'm closing in on an hour for "git add ." on my eBooks dir, just wondering if I'm setting a record here.04:43
→ sdboyer joined04:45
← __sorin__ left04:45
← __sorin_1 left04:45
→ iszak joined04:45
iszak So I want to host a private repository which isn't accessed via SSH, I want credential authentication, what are my options?04:46
bob2 why not ssh04:46
iszak security04:46
EugeneKay Not using ssh for security reasons is...... backwards.04:47
bob2 hahaha04:48
iszak not using git over ssh for security reasons04:49
I have nothing against using SSH in general, just not using it with git04:49
EugeneKay What are you trying to accomplish?04:50
bob2 what security issues do you forsee?04:50
iszak bob2, rm -fr /bin04:50
I want them to push to a server all their code changes.04:50
EugeneKay Not to be rude, but lrn2sysadmin04:50
→ KKA joined04:50
bob2 iszak: uh, no04:50
iszak: obviously you don't have people comitting as root04:50
tbranyen could probably easily chroot jail an ssh user04:50
and only allow access to the git binary04:51
iszak bob2, right I know, it's a pull system, but the user won't always be online so I can pull their code04:51
bob2 so what on earth are you talking about04:51
tbranyen wait did i just say easily xD04:51
bob2 in what way do you think git over ssh is exploitable04:51
EugeneKay lol chroot easy04:51
iszak yeah I looked into chroot, seemed difficult.04:51
EugeneKay I have yet to figure out why there is no easy chroot environment create script04:51
iszak then I looked into rssh, but it doesn't support git commands, there's a patch tho04:51
EugeneKay (for x distros)04:51
tbranyen iszak: i've gotten it working a couple of times, but i've had to use a guide to get it just right every time04:52
EugeneKay: Arch Linux has a few decent guides04:52
iszak tbranyen, link to guide?04:52
bob2 right, so you're imagining issues04:52
you probably want git and gitolite04:52
tbranyen iszak: I haven't done it since last year, but all I did was Google for it04:52
← ymasory left04:52
EugeneKay tbranyen - guides != "yum install user-chroot && user-chroot eugenekay"04:52
tbranyen I should probably bookmark stuff04:52
EugeneKay ;-)04:52
tbranyen EugeneKay: welp, sometimes getting your hands dirty is a good thing04:53
iszak I don't use yum anyway ;)04:53
EugeneKay tbranyen - I know that - but it would be a good gesture. I mean, it's 2011.04:53
iszak gitosis seems to be a solution.04:54
tbranyen Agreed, sudoing a user takes minimal effort04:54
why not jailing04:54
iszak despite being SSH, it seems secure04:54
bob2 lol04:55
EugeneKay iszak - what about ssh is insecure? User having a shell on the target box?04:55
Don't blame the protocol, blame the environment. ;-)04:55
clever ssh can be setup to run git and only git04:55
→ khmarbaise joined04:56
hyperair sb levelclear -level clientcrap,crap,joins,parts,quits,nicks,clientnotice04:57
← jsick3 left04:57
bob2 heh04:57
← harv left04:57
bob2 hyperair: /04:57
hyperair bob2: yeah =(04:58
← RichardBronosky left04:58
→ ThomasLocke joined04:59
← dAnjou left04:59
→ dAnjou joined04:59
iszak EugeneKay, it seems to be satisfied with gitosis05:00
← boombatower left05:00
← jfkw left05:00
← leo2007 left05:01
→ urkud joined05:02
→ dvaske joined05:04
← khmarbaise left05:04
→ txh joined05:05
→ dAnjou_ joined05:06
← Tv left05:06
← dAnjou left05:09
dAnjou_ → dAnjou05:09
→ roop joined05:11
→ liorkesos joined05:11
← Schmallon left05:13
→ excess^ joined05:15
← shiba_yu36 left05:17
→ HG` joined05:18
← fairuz left05:18
→ Engin joined05:18
← monokrome left05:19
→ shiba_yu36 joined05:20
← HG` left05:20
→ ExtraSpice joined05:20
→ aminpy joined05:20
→ ajt joined05:21
→ paja joined05:21
→ EricF joined05:28
→ Bombe joined05:28
← robotarmy left05:29
← urkud left05:31
← liorkesos left05:32
← Bombe left05:32
← cirwin left05:33
→ shine-neko joined05:34
→ doener joined05:35
← towski left05:36
← bdiu left05:38
→ j0ran joined05:38
→ Soliton joined05:39
← Yuffster left05:40
← Bonkers left05:42
→ dkannan joined05:42
EugeneKay fatal: confused by unstable object source data for 8e70af58cdbb989ad3c0170f23b1474f3478a46505:44
:-/05:44
← iszak left05:44
→ towski joined05:44
← Engin left05:45
→ javaanse_jongens joined05:46
← MattDiPasquale left05:46
→ Engin joined05:46
karstensrage do you guys do anything with WIP code ?05:48
like a sandbox repo or something05:49
← TiBook left05:50
doener karstensrage: like anything else, it gets its branch. No need for a separate repo05:50
→ milli joined05:51
→ LongBeach joined05:51
karstensrage fr'doener, no i mean like code that not even in a repo, like testing stuff out or investigating some new thing05:52
→ MattDiPasquale joined05:52
doener karstensrage: a whole new project then? If I work for more than, say, 15 minutes on it, it gets a repo05:53
→ Bombe joined05:53
→ TrineLocke joined05:53
doener karstensrage: there are basically no downsides to running "git init" ;-)05:54
karstensrage hmmm fr'doener ok, maybe thats the ticket05:54
→ Laggg joined05:55
← towski left05:56
karstensrage git init sandbox it is :)05:56
← slide left05:56
Laggg what are some cool git repos?05:57
karstensrage define 'cool'05:57
→ xiangfu joined05:58
← bauruine left05:59
← ISF left06:00
→ Twey joined06:01
→ airborn joined06:02
→ bauruine joined06:02
← bauruine left06:02
← xiangfu left06:03
→ slide joined06:03
← djbpython1 left06:03
← Laggg left06:04
→ _ollie joined06:05
← dAnjou left06:05
← Bombe left06:06
→ liorkesos joined06:07
← ericdc left06:07
← _ollie left06:07
→ dAnjou joined06:07
→ evaryont joined06:07
→ ericdc joined06:07
evaryont anyone know what is going on? When I execute git from the command line, I get the following error: "git: exec format error: hub" (with return code 126)06:08
→ Tasser joined06:09
→ elb0w joined06:09
← Octalot left06:09
evaryont oh, my bad, my bad... I was stupid. my git wrapper script was moved to a different folder. silly me. (eeps)06:10
→ lxsameer joined06:10
lxsameer i have a remotes/origin/HEAD, what does it means?06:10
← doener left06:12
→ thiago_home joined06:12
→ joel_falcou joined06:13
← liorkesos left06:14
→ cbreak_work joined06:14
→ fr0sty joined06:15
← troyt left06:16
→ troyt joined06:16
→ xiangfu joined06:16
← tgunr_ left06:17
→ Laggg joined06:19
← TrineLocke left06:21
→ ice799 joined06:21
→ TrineLocke joined06:21
← EugeneKay left06:22
→ kronos joined06:23
← phantomcircuit left06:26
→ bauruine joined06:27
→ urkud joined06:29
→ dannas joined06:29
← dr_ick left06:29
← LouisJB left06:31
← excess^ left06:31
→ hynek joined06:33
→ guardian joined06:34
→ EnginA joined06:34
→ shruggar joined06:35
← etnt left06:35
← project2501b left06:35
← roop left06:36
← Engin left06:36
→ nicoulaj joined06:37
→ matjas joined06:37
→ ph^ joined06:38
← dAnjou left06:38
→ dAnjou joined06:38
← slide left06:39
→ Bombe joined06:41
← Zenith77 left06:42
→ mklappstuhl joined06:42
← EricF left06:42
← ormaaj left06:43
→ tvw joined06:43
→ larsmoil joined06:45
→ giallu joined06:46
← madewokherd left06:47
→ fairuz joined06:47
→ mastroDani joined06:48
→ ormaaj joined06:48
→ rue joined06:50
→ slide joined06:52
→ jhoepken joined06:52
→ Engin joined06:52
← larsmoil left06:53
← dAnjou left06:53
→ larsmoil joined06:53
← mosno left06:53
← airborn left06:53
← dvaske left06:53
→ dAnjou joined06:54
← EnginA left06:55
← kronos left06:55
→ skoop joined06:56
← infid left06:56
← tvw left06:57
→ resmo_ joined06:58
→ infid joined06:58
→ radarek joined06:58
→ henry_ept joined06:59
← paja left06:59
← bjtitus left07:01
→ roop joined07:01
← dAnjou left07:02
→ dAnjou joined07:04
← xerora left07:05
→ teweWork joined07:05
← jungejason left07:06
← shennyg left07:08
← teweWork left07:08
← dreiss left07:10
→ teweWork joined07:11
← qwester left07:14
← dannas left07:14
→ shennyg joined07:14
← OOPMan left07:16
→ paja joined07:17
→ dvaske joined07:17
→ workmad3 joined07:18
fairuz hi07:18
how to reset to previous commit?07:18
→ crashanddie joined07:18
← crashanddie left07:18
→ crashanddie joined07:18
thiago_home by reset, do you really mean what git reset does?07:18
→ OOPMan joined07:20
fairuz thiago_home: i mean how to reset to previous commit, btw i found it.. git reset --hard HEAD .. thanks07:20
my other question..we have git diff to make patches..do we have git patch to apply the patches or we can just use normal patch command07:21
thiago_home git apply07:22
to make patches, use git format-patch07:22
to apply, use git am07:22
ChanServ set mode: +v07:22
henry_ept Hello.07:22
I have questions07:23
thiago_home we might have answers07:23
henry_ept I tried to compile git 1.7.4.2 myself on Mac OS X 10.6.7, but I failed.07:23
I think config.mak.autogen is not included in Makefile.07:23
← mastroDani left07:23
fairuz thiago_home: thanks. what are difference between git diff and git format-patch?07:23
henry_ept What's wrong?07:23
configure --> http://pastebin.com/VnBGPgB507:23
← evaryont left07:23
henry_ept config.mak.autogen --> http://pastebin.com/0SH9ZLTV07:23
I fixed Makefile manually, L309 'CC = gcc' --> 'CC = /usr/bin/gcc'07:23
then compiled successfully07:23
thiago_home fairuz: git diff just produces a diff of any two things07:23
fairuz: git format-patch creates a patch out of a commit07:24
crashanddie henry_ept, you don't have gcc in your path on OSX?07:24
← Bombe left07:24
thiago_home fisted_: it includes the commit message07:24
→ jds_ joined07:24
fairuz thiago_home: but git diff HEAD -- > my.patch will also create a patch ?07:24
thiago_home fairuz: yes07:25
marcamilly ok when I am looking at a diff on the command line, how do i read this now ? I am actually looking at the diff of a git flow07:25
thiago_home fairuz: but it won't include the commit message (because there is no commit)07:25
→ drake01 joined07:25
henry_ept crashanddie: I use gcc compiled myself.07:25
→ empity joined07:25
fairuz thiago_home: ah ok...i got it now..thanks07:25
thiago_home marcamilly: explain07:25
crashanddie henry_ept, why, exactly? Probably off topic, but still.07:25
henry_ept but when compile git, i want(need) to use /usr/bin/gcc07:25
marcamilly when I do git flow feature diff edit_comments, and I am on the develop branch, am I looking at a diff between the develop & edit_comments branch thiago_home ?07:26
thiago_home marcamilly: I have no idea what git flow is07:26
henry_ept apple's gcc very old, so i use gcc 4.6.007:26
thiago_home no such command on 1.7.3.207:26
marcamilly ooh thiago_home https://github.com/nvie/gitflow07:26
→ mishok13 joined07:27
marcamilly it's pretty kewl for managing your workflow on a project07:27
→ opalka joined07:27
← dvaske left07:27
marcamilly thiago_home: it was modeled off of this post: http://nvie.com/posts/a-successful-git-branching-model/07:27
→ robert_david joined07:28
→ Bombe joined07:28
marcamilly so my question might be moot then ?07:28
henry_ept However, When I use gcc 4.6.0 compiled myself, I failed to compile git. I need to use /usr/bin/gcc07:28
thiago_home henry_ept: and /usr/bin isn't on your $PATH?07:29
marcamilly: no, it's just that I have no clue what your question means.07:29
← smokex left07:29
thiago_home marcamilly: so I can't answer.07:29
marcamilly ok...hrmm07:29
anyone else in here familiar with git flow ?07:29
henry_ept So i set 'CC=/usr/bin/gcc ./configure ....'07:29
→ psoo joined07:30
henry_ept thiago_home: /usr/bin is in my PATH07:30
thiago_home henry_ept: then why do you need to set the compiler's absolute path?07:30
henry_ept: do you have another gcc?07:30
← Theravadan left07:30
henry_ept thiago_home: yes. i usaully use $HOME/local/gcc46/bin/gcc07:31
marcamilly thiago_home: this is what my diff looks like: https://gist.github.com/89011207:32
and am trying to figure out which is why07:32
→ dreiss joined07:32
marcamilly how do I know which branch is which ?07:32
or is this a diff of changes on one branch ?07:32
henry_ept gcc --> http://pastebin.com/2KEhWu4a07:33
← mklappstuhl left07:34
→ lhunath joined07:34
→ leo2007 joined07:35
lhunath I've got a patch file generated by git format-patch -M which contains a rename like, --- a/file1 b/file2, but git-am fails to apply this saying: fatal: git apply: bad git-diff - inconsistent old filename on line 10907:35
← urkud left07:36
thiago_home marcamilly: a is the first branch, b is the second branch07:36
→ henrikbjorn joined07:36
→ troessner joined07:37
marcamilly thiago_home: hrmm....ok.....so if I am on branch 'develop' and I execute a git command to basically show me the diff and I specify another branch...does that mean that the branch I am currently on becomes 'a', and the one I specified becomes 'b' ?07:37
← MattDiPasquale left07:38
thiago_home marcamilly: git diff => diffs the working tree to the index07:38
marcamilly: git diff --cached => diffs the index to HEAD07:38
marcamilly: git diff $commit => diffs the working tree to $commit (a is the working tree, b is $commit)07:38
marcamilly: git diff $commit1 $commit2 => diffs one commit to the other07:39
marcamilly hrmm....interesting07:39
well what i did was git flow feature <name> diff07:39
but you aren't familiar with git flow...sooo...07:39
:(07:40
henry_ept Makefile L1196 '-include config.mak.autogen' but 'CC = gcc' yet.07:41
← fairuz left07:41
← dAnjou left07:41
→ dAnjou joined07:42
→ liorkesos joined07:42
→ meandi2 joined07:43
henry_ept If true, CC should be /usr/bin/gcc(i will set) here?07:43
← helvete left07:43
← meandi left07:43
thiago_home henry_ept: why don't you just set the PATH to contain only one gcc, the one you want?07:43
→ MattDiPasquale joined07:44
→ L42y joined07:44
→ dvaske joined07:44
← MattDiPasquale left07:45
→ mastroDani joined07:46
← aminpy left07:47
→ MattDiPasquale joined07:47
← dkannan left07:47
→ dkannan joined07:48
← blueyed left07:48
→ davzie joined07:49
← mastroDani left07:50
henry_ept thiago_home: It's messy to rewrite $PATH setting only for compiling git :p07:50
And if succeed to compile, make install to $HOME because config.mak,autogen is not included, in which prefix = $HOME/local07:50
← notbrent left07:50
→ mabrand joined07:50
→ eijk joined07:51
← thiago_home left07:51
→ mastroDani joined07:51
← meandi2 left07:52
→ meandi joined07:52
← shruggar left07:55
← drake01 left07:57
→ tatsuya joined07:58
← tyteen4a03 left07:59
← Laggg left07:59
→ mklappstuhl joined08:00
→ dAnjou_ joined08:00
→ dpino joined08:02
← dAnjou left08:03
← theoros left08:04
→ nonnooo joined08:04
→ fairuz joined08:05
→ savetheinternet joined08:06
dAnjou_ → dAnjou08:06
← dvaske left08:06
savetheinternet i'm a bit of a git noob, and need some help making a download with the git tagging thing08:06
anyone here?08:06
← Error404NotFound left08:06
savetheinternet i created the 'tag' thing. http://pastie.org/private/ars4n7as26q25yqd8xpg but then what do I do?08:06
fairuz Hi, if I replace a file with other file with a same name. Do git detect it as delete and add or just content modify?08:06
savetheinternet i was reading http://gitref.org/branching/#tag08:06
do i need to make a commit?08:06
fairuz savetheinternet: afaik, a tag is already a commit08:07
just it has name08:07
savetheinternet how do I "push" it then?08:07
i'm trying to use the github download tagging thing08:07
>There aren't any downloads yet. You can create some by uploading files using form above or tagging a point in your repository.08:07
thiago savetheinternet: what are you trying to do?08:07
→ tyteen4a03 joined08:07
thiago you want to push your tag?08:07
savetheinternet yes08:07
i want to make downloads like this: https://github.com/openmelody/melody/downloads08:08
thiago git push remotename tagname08:08
→ PigeonFriend joined08:08
fairuz savetheinternet: something like git push origin mytag08:08
→ dc5ala joined08:08
savetheinternet much love.08:09
thank you08:09
fairuz thiago: any idea about my question? :D08:09
← roop left08:10
thiago fairuz: git does content detection08:10
fairuz thiago: ok08:10
marcamilly is there anyway to manually go through the differences between two branches and merge some of the changes from one branch into the other ?08:11
thiago fairuz: if you delete a file and re-add with different content, Git may or may not detect a rewrite08:11
it can detect a rewrite even without a delete+add08:11
→ MikeChelen joined08:11
marcamilly i would imagine that a GUI would make this process easier...is there a gui for OS X that would make it easier ?08:11
← L42y left08:11
← dAnjou left08:11
MikeChelen how do you push changes to .git/info/exclude to the remote repo?08:12
→ dAnjou joined08:12
marcamilly btw thiago do the - and + mean anything in the diff view ? or does it just show the lines for a & b respectively ? or does it mean lines deleted and added ?08:13
→ aminpy joined08:14
fairuz marcamilly: + means the lines are addes - lines are removed08:15
marcamilly ok fairuz thanks08:15
thiago marcamilly: no, just standard diff meaning08:16
← mastroDani left08:18
← dAnjou left08:19
→ Error404NotFound joined08:19
→ dAnjou joined08:20
jaeckel MikeChelen: info/exclude is only for the local repository! you need to add a .gitignore, c.f. man git-ignore08:22
fairuz thiago: i suppose git format-patch -k -1 will do diff between master and a commit before that?08:23
thiago fairuz: provided you have master checked out08:23
fairuz thiago: yup, thanks08:23
→ TrahDivad joined08:25
TrahDivad Hi ppl08:25
← dAnjou left08:26
TrahDivad How can I local changes I have made that haven't even been staged?08:26
How how can I delete08:26
thiago TrahDivad: your sentence is missing a verb08:26
TrahDivad I meant08:26
thiago ah08:26
bob2 git checkout filename # obliterate local changes to filename08:26
henry_ept thiago: btw, does configure and Makefile work well in your enviroment?08:26
thiago henry_ept: never tried08:26
TrahDivad bob2: How would I do it for all local changes?08:27
for all files even08:27
thiago henry_ept: I don't use configure08:27
bob2 don't provide a filename08:27
maybe also -f08:27
TrahDivad bob2 Thanks :) I'll give it a try08:27
→ dAnjou joined08:27
henry_ept thiago: ok :)08:27
thiago TrahDivad: git status. It will tell you.08:27
← savetheinternet left08:28
← robert_david left08:28
TrahDivad thiago: You're right, git status does indeed tell me. :S08:28
marcamilly thiago: there anyway to selectfully accept some changes commited to a branch, via a merge to another branch ?08:29
i.e. branch b has say 10 changes that I want in branch a08:29
thiago marcamilly: no08:29
marcamilly but it also has 15 changes that I dont want08:29
thiago marcamilly: merge merges everything08:29
marcamilly: it's all-or-nothing08:29
→ erichammond joined08:30
marcamilly so no way at all to get them individually ?08:30
thiago marcamilly: if you didn't want those changes, they should have been in separate branches08:30
marcamilly :|08:30
→ robert_david joined08:30
thiago marcamilly: you can cherry-pick each individually08:30
it's not the same thing08:30
marcamilly ok, will do that08:30
thiago we use selective cherry-picking for release branches08:31
small, hotfixes before a release08:31
we use merges for everything else08:31
← dAnjou left08:32
→ dAnjou joined08:32
→ mjt joined08:33
← Nonpython left08:34
→ bentob0x joined08:34
→ spuz joined08:34
← jds left08:34
→ Nonpython joined08:35
lxsameer i my remote repo has a HEAD, what does that mean ? what is wrong ?08:37
thiago nothing is wrong08:37
when you clone a repository, you get the HEAD checked out08:37
→ LouisJB joined08:38
→ saschagehlich joined08:39
→ achadwick joined08:40
← spuz left08:41
← dreiss left08:41
→ dannas joined08:41
fairuz thiago: i got an error while trying to do git am. I suppose I have to resolve it manually?08:42
or is there any tool to help? :D08:42
thiago fairuz: do you mean a conflict?08:42
→ mastroDani joined08:42
fairuz thiago: i think so08:43
error: patch failed: arch/arm/include/asm/pmu.h:2208:43
error: arch/arm/include/asm/pmu.h: patch does not apply08:43
Guest26955 hey all08:43
Guest26955 → cbx3308:43
cbx33 Mornin08:43
so guys - a question - how does git reset --merge differ from git merge --abort08:44
← dAnjou left08:44
→ dAnjou joined08:44
fairuz thiago: it even dont show me which line the conflict occured08:44
:(08:45
thiago fairuz: it doesn't. Use the standard patching tools to fix it08:45
fairuz thiago: ok08:45
← Spockz left08:48
→ JohnBat26 joined08:48
→ tvw joined08:50
→ rumith_ joined08:51
→ Saur joined08:51
← dAnjou left08:52
← Saur left08:52
→ dAnjou joined08:53
← AAA_awright left08:54
← rumith_ left08:54
← lhunath left08:55
→ AAA_awright joined08:55
← synisma left08:56
→ MacGyverNL_ joined08:57
← MacGyverNL_ left08:57
← dAnjou left08:59
→ dAnjou joined09:00
← tsbtmn left09:00
→ mastroWork joined09:01
→ tsbtmn joined09:02
← tsbtmn left09:02
→ tsbtmn joined09:02
Boggle yesterday I created a GitHub account and successfully loaded up an Xcode project into a new repository... since then I have modified the Xcode project. How do I upload the new project? Exactly the same way as I did in the first place? ie here ... http://help.github.com/create-a-repo/09:03
?09:03
how about if I am happy to just erase everything and copy the new one over?09:04
I don't need an old version and a new version, ...09:04
→ logankoester joined09:04
← mastroDani left09:05
← dAnjou left09:06
→ dAnjou joined09:06
← JohnBat26 left09:07
→ Spockz joined09:09
← st-5426 left09:10
← dAnjou left09:10
→ dvaske joined09:11
f00li5h Boggle: you'll want to commit your changes, then push the repo09:11
→ koltroll joined09:12
→ CareBear\ joined09:12
CareBear\ hi all!09:12
I'm writing a pre-receive hook that should detect rebase.09:13
→ mikelangelo joined09:13
CareBear\ what's an easy/efficient way to determine if the old commit exists in the history of the new commit?09:13
→ dAnjou joined09:14
thiago git log -1 $newcommit..$oldcommit. If you get no output, it exists in the history. If you get any output, it doesn't exist in the history.09:15
← dkannan left09:15
CareBear\ thiago : of course! thanks for the hint. :)09:16
oh no I misread that - not what I first thought it did09:17
but works perfectly09:17
thanks!09:17
← blaenk left09:18
← dAnjou left09:18
→ eschnou joined09:18
← opalka left09:18
→ opalka joined09:18
← opalka left09:18
→ opalka joined09:18
fairuz Hi, how to delete a remote branch? deleting in local and pushing to remote doesn't solve it09:19
tango_ fairuz: push remotename :branchname09:19
← davzie left09:20
→ dAnjou joined09:20
fairuz tango_: a weird command to delete a branch :D thanks09:21
→ dkannan joined09:21
jaeckel use git extras' "git delete-branch" :>09:24
← dAnjou left09:25
→ dAnjou joined09:25
← frerich left09:28
← marcamilly left09:28
→ harv joined09:28
← dvaske left09:30
← siprbaum left09:30
→ siprbaum joined09:30
→ mastro joined09:30
→ frerich joined09:31
← frerich left09:31
→ frerich joined09:31
→ nevyn^ joined09:31
← mikepack left09:31
→ thierryp joined09:34
← mastroWork left09:35
→ xiong joined09:36
← mabrand left09:39
← dc5ala left09:39
← mklappstuhl left09:39
← dAnjou left09:41
→ dAnjou joined09:42
fairuz if I have c1 -> c2 -> c3 -> c4 , where i'm now at c4..Can i go back to c2 and reapply c4?09:42
Actually I just want to get rid of modification made in c309:43
tango_ fairuz: did you push your history already?09:43
if not, git rebase -i c1 and delete c3 from the chain of reapplied commits09:43
→ dvaske joined09:44
fairuz tango_: what do you mean by push history? push to remote?09:44
tango_ yeah09:44
fairuz tango_: it's only a local repo, so no remote09:44
tango_ excellent09:44
then git rebase is what you want, look up the man for further info09:44
fairuz tango_: ok thanks09:45
tango_ np09:45
fairuz tango_: but if we delete c3, do git smart enough to chain c1 with c4?09:45
or i understand it wrong09:45
tango_ if you only delete c3, you will have c1 -> c2 -> c409:45
without the c3 changes09:45
→ d0k_ joined09:47
← txh left09:47
fairuz tango_: so the rebase will make c1 -> c2 -> (2 branch here c4 and c3)..so i just delete the c3 to get only c1 -> c2 -> c4?09:48
tango_ no, the rebase will just plot c4 on top of c209:48
plop09:49
fairuz: read the rebase manpage09:49
fairuz: man git-rebase09:49
jast fairuz: the 'git-rebase' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-rebase.html [automatic message]09:49
fairuz tango_: ok thanks09:49
→ hellovetica joined09:52
→ ereslibre joined09:53
← ereslibre left09:53
→ ereslibre joined09:53
← dAnjou left09:54
→ dAnjou joined09:54
← hellovetica left09:55
→ shruggar joined09:57
← aminpy left09:57
shruggar how can I find the original id of the "current" commit during git-filter-branch?09:57
← mastro left09:58
→ aminpy joined09:58
shruggar I am trying to remove all commits which match a list of IDs (as git rebase -i does when you delete a line)09:59
→ st-8616 joined09:59
← nd__ left10:00
← tsbtmn left10:00
← st-8616 left10:00
← cutmail left10:01
→ st-8670 joined10:02
→ mastro joined10:03
← st-8670 left10:03
→ rendar joined10:03
← wilmoore left10:04
→ nd__ joined10:04
→ xudifsd joined10:05
→ quazi_farhan joined10:06
← genjin left10:07
→ tsbtmn joined10:07
← tsbtmn left10:07
→ tsbtmn joined10:07
→ larsl joined10:08
← tatsuya left10:08
← dannas left10:10
xudifsd Hi,there, GSOC's homepage says that mentor org will may have a template for student's application, but i can't find it in git, where is it?10:10
← tsbtmn left10:12
thiago orgs don't have to have a template10:12
← dAnjou left10:13
xudifsd oh, this means git don't have it right?10:13
thiago no10:14
I don't know if git has it or not10:14
→ spaceonline joined10:15
xudifsd eh, thanks anyway.10:17
← henrikbjorn left10:18
→ tsbtmn joined10:18
← tsbtmn left10:18
→ tsbtmn joined10:18
→ davzie joined10:19
← thierryp left10:19
xudifsd btw, should application submit to org or google?10:19
thiago you submit the application on the web interface10:20
there's only one interface to do it10:20
← teweWork left10:20
thiago but you should contact the organisation and talk to them10:20
← resmo_ left10:20
→ resmo_ joined10:21
← ph^ left10:22
xudifsd i see10:22
→ ph^ joined10:22
← workmad3 left10:24
→ dAnjou joined10:26
← hyperair left10:27
← Engin left10:28
→ mosno joined10:29
← mastro left10:29
→ st-9366 joined10:29
→ avdg joined10:29
→ vitor-br joined10:30
← st-9366 left10:30
→ sagsousuke joined10:32
→ minkben joined10:32
minkben if I already have a project directory with files, how do I setup a git repo locally and remotely?10:33
→ mastro joined10:34
→ gebi joined10:34
minkben or, how do I init a directory with file in it already and also tie it to a remote repo10:35
JPT hmm.. just initialize your repository, put the files in there, commit them and then clone your repository elsewhere.10:36
→ st-9471 joined10:37
minkben how?10:37
← st-9471 left10:37
→ st-9486 joined10:37
← st-9486 left10:38
→ dAnjou_ joined10:38
← Spockz left10:39
minkben do I have to start with an empty directory?10:39
avdg minkben: nope10:39
just go to the root of your project10:39
and 'git init'10:39
minkben but I want it on my server10:39
I did git init --bare on the server10:39
avdg and where is your project files?10:39
minkben locally10:40
avdg and the local files are already in a repo?10:40
minkben no10:40
I haven't made any repo for them yet10:40
avdg its better to create a local one10:40
minkben I don't know how to proceed10:40
avdg makes life easier10:40
→ st-9523 joined10:40
minkben I want to sync between school and home10:40
so I need a remote repo10:40
avdg just create a local repo, then push it to a remote repo10:40
← jophish left10:41
minkben how do I push it to the remote repo10:41
← st-9523 left10:41
← dAnjou left10:41
avdg well, just start with the local repo ;-)10:42
minkben local repo done10:43
avdg how do you get access to the server ssh?10:43
minkben yeah, ssh10:43
→ st-9554 joined10:43
avdg if you know the location, you could add that remote10:43
minkben how do I do that10:44
→ simplechat joined10:44
← simplechat left10:44
→ simplechat joined10:44
avdg first by defining your remote10:44
http://www.kernel.org/pub/software/scm/git/docs/git-remote.html <- git remote add10:44
← Sonderblade left10:45
avdg git remote add <name> <url>10:45
vdv how can i comment line in a .git/config? with #?10:45
avdg vdv: yes10:45
→ Sonderblade joined10:45
← st-9554 left10:45
→ Sonder joined10:45
→ dAnjou joined10:46
← dl9pf_ left10:47
← ereslibre left10:48
→ ereslibre joined10:48
→ dl9pf joined10:49
← dl9pf left10:49
→ dl9pf joined10:49
← dAnjou_ left10:50
→ wingo-pi joined10:50
→ evildmp joined10:50
minkben avdg, got it up and running, thanks dude!10:50
evildmp how do I abandon a commit that I haven't yet pushed?10:50
I don't want to abandon the changes to the files in my working tree10:50
avdg minkben: np10:50
evildmp I just don't want to commit the changes yet10:51
avdg don't forget to push ;-)10:51
minkben g2g10:51
← minkben left10:51
vdv how can i remove remotes that i don't need anymore?10:51
evildmp is: git revert HEAD what I need to do?10:51
avdg vdv: try git remote -d10:51
evildmp: try "git rebase -i"10:52
→ st-9779 joined10:52
← st-9779 left10:52
avdg *git rebase -i <commit where on top you want to rebase>10:52
→ urkud joined10:54
→ dr_win joined10:55
→ sweetpi joined10:56
→ st-9828 joined10:57
← st-9828 left11:01
→ Vampire0 joined11:01
← Vampire0 left11:01
→ Vampire0 joined11:01
→ st-9894 joined11:02
← st-9894 left11:02
← dvaske left11:03
→ st-9915 joined11:03
← dAnjou left11:03
→ dAnjou joined11:03
← st-9915 left11:04
→ st-9929 joined11:04
← st-9929 left11:05
← dAnjou left11:07
← urkud left11:07
→ st-10005 joined11:08
→ Spockz joined11:09
← st-10005 left11:13
→ dAnjou joined11:13
→ t3rm1n4l joined11:13
← EricInBNE left11:14
→ sebersole joined11:14
→ st-10046 joined11:15
← st-10046 left11:15
← muep_ left11:16
→ st-10058 joined11:16
← st-10058 left11:16
← bottiger left11:16
← leo2007 left11:16
→ urkud joined11:17
→ elyob joined11:18
← mastro left11:19
→ st-10104 joined11:20
→ muep_ joined11:21
→ roop joined11:23
→ flaguy48 joined11:23
→ owst joined11:25
← ptl left11:25
wingo-pi http://wingolog.org/archives/2011/03/28/git-brunch111:26
→ nicomen joined11:26
→ ptl joined11:26
nicomen I'm setting up gitweb as a subdir on my apache server, but I keep getting an error that HEAD doesn't point to a valid ref when doing a checkout. the rewrites on the server as supposedly invoking gitweb.cgi as gitweb.cgi/<project>/HEAD but it outputs an html page -- what am I doing wrong?11:27
→ g0bl1n joined11:27
thiago nicomen: nothing11:27
are you sure it's an error? It should say "warning"11:28
nicomen warning: remote HEAD refers to nonexistent ref, unable to checkout.11:28
thiago so it's not an error11:28
nicomen the HEAD file contains: ref: refs/heads/master11:29
is gitweb supposed to passthru the access to that file without invoking the cgi script?11:29
thiago gitweb has nothing to do with cloning11:30
are there any commits in branch master?11:30
← IceW left11:30
nicomen how do I check that?11:30
→ Gulug joined11:30
← michael_lf left11:31
nicomen (what I'm trying to do is let a repo be accesible thru http for git command line btw.)11:31
← Guest15518 left11:31
nicomen it works june fine by using ssh:// as the protocl and path to the repository on the box11:31
→ cutmail joined11:32
← roop left11:33
← dkannan left11:35
← st-10104 left11:36
→ q0rban joined11:36
→ etnt joined11:37
← sweetpi left11:38
→ dkannan joined11:39
→ chernjie joined11:41
← chernjie left11:42
→ mintux1 joined11:43
← dAnjou left11:43
→ dAnjou joined11:44
→ foocraft joined11:45
← lxsameer left11:47
← javaanse_jongens left11:48
→ hbf joined11:49
← gebi left11:50
→ uncaught_excepti joined11:50
← dyer left11:50
← shiba_yu36 left11:51
→ dyer joined11:52
← dyer left11:53
← dAnjou left11:54
→ mklappstuhl joined11:54
→ artefon joined11:54
→ dAnjou joined11:54
hbf After repacking tightly, should i do git-send-pack explicitly to the remote repo instead of push? With one repo packed/compressed better than the other, how does git push/fetch decide which repo's packing to (mostly) retain?11:56
thiago it retains the current pack11:57
jast you can't replace existing objects on the wire11:57
if you want to have things repacked on the remote, repack them on the remote11:58
hbf ok. so repack before push/send-pack merely affects the amount of network traffic?11:58
jast it might11:58
→ simenbrekken joined11:58
hbf ok. thanks.11:59
jast but since push generates a thin pack on the fly anyway... chances are it makes no difference11:59
← mr_daniel left12:00
hbf by default anyway, i notice push has default --thin and send-pack not.12:01
→ mr_daniel joined12:01
→ ph^_ joined12:02
jast without --thin it just means that the pack generated on the fly won't use deltas12:02
which is pretty much always a bad idea12:02
← ph^ left12:03
quazi_farhan let us just say, i cloned a repo at my local drive and there has been some update at the server. how can i update my local version? i mean which command?12:03
→ workmad3 joined12:04
jast quazi_farhan: generally, git pull (which will merge the upstream changes with your locally committed changes, if any)12:04
quazi_farhan jast, thank you...12:05
← Error404NotFound left12:05
← Spockz left12:05
hbf jast, thanks again.12:06
← mosno left12:06
← dAnjou left12:07
→ engla joined12:08
← engla left12:08
→ engla joined12:08
→ blueyed joined12:09
→ dAnjou joined12:10
→ Error404NotFound joined12:11
→ rseifert joined12:12
→ mastroDani joined12:13
→ leandrosansilva_ joined12:13
→ leandrosansilva joined12:13
← dAnjou left12:13
← leandrosansilva left12:13
→ gebi joined12:14
→ dAnjou joined12:15
→ leo2007 joined12:15
→ Err404NotFound joined12:16
← Error404NotFound left12:16
← leandrosansilva_ left12:16
← urkud left12:17
→ leandrosansilva joined12:17
← leandrosansilva left12:17
← rseifert left12:18
← SirFunk left12:18
← hobodave left12:18
→ rseifert joined12:18
← jds_ left12:18
→ mrevd joined12:19
← dkannan left12:19
→ nona joined12:19
← uncaught_excepti left12:20
→ johnpipi joined12:20
johnpipi how can i push a local tag i made to the remote bare repo's master?12:21
is that possible?12:21
→ urkud joined12:21
ChanServ set mode: +v12:22
johnpipi how can i push a local tag i made to the remote bare repo's master?12:22
is that possible?12:22
← kushal left12:22
→ durarara joined12:23
← durarara left12:23
hbf i'm new at this, but i suppose push <remote name> refs/tags/<tag name>12:23
→ thrope joined12:23
johnpipi let me try that12:23
what i am trying to do is deploy to multiple servers with git12:23
thrope I wonder if someone could help me - I am trying to apply a commit from one repository to another.12:24
I am trying to merge too different lines of development done outside git12:24
on the first one I have 3 commits, on the second I have the same first commit and then a different second commit12:24
I have branched the first repo to where they were in common, and now I would liek to somehow apply the same changes12:24
I tried git-format-patch and then git-am but it fails saying patch won't apply12:25
is there any other way to do it?12:25
→ bottiger joined12:26
→ dyer joined12:26
← dyer left12:26
→ dyer joined12:26
hbf add one repo as a 'remote' to the other with 'git remote', and pull the devel branch from there. then use git merge or rebase to join the development histories into a single branch.12:26
← nona left12:27
thrope previous rebase directory /home/robince/src/MatlabAPI_20091219/.git/rebase-apply still exists but mbox given.12:27
→ tstclair joined12:27
bottiger How do I move a submodule into another folder. Just doing "git mv submodule new_folder/" gives me a "fatal: source directory is empty ..." error12:27
thrope how do I use git remote?12:28
I tried git remote add largerem ../MatlabAPI_temp12:28
hbf thrope: don't know what that mbox given means mut maybe you tried rebase and failed, and need to clean up with git rebase --abort?12:28
← t3rm1n4l left12:28
thrope it says No rebase in progress12:28
I definitely havent run a rebase myself12:28
← mintux1 left12:28
thrope but I am on a (no branch) becuase I checked out an earlier commit12:29
avdg try git rebase —abort12:29
hbf huh. i don't know what that error message means then, i'm new at git myself. for remote, see example in 'man git-remote'. Or gittutorial & gitworkflows, if you havent' read those yet12:29
jast the 'git-remote' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-remote.html [automatic message]12:29
thrope it gives that error - "no rebase in progress"12:29
→ sunblush joined12:32
← meandi left12:33
→ sftp joined12:33
← dAnjou left12:33
→ Frtecek joined12:34
→ meandi joined12:34
→ dAnjou joined12:34
avdg thrope: what does "git rebase status" say?12:34
oh nvm12:35
avdg confused12:35
thrope it says "fatal: Needed a single revision"12:35
invalid upstream status12:35
← rseifert left12:35
thrope anyway I am trying to add it as a remote now12:36
avdg remote?12:36
so you push it to an other repo right?12:36
thrope now I have the two things I want to merge in different branches on the same repo12:37
so things are looking up i think12:37
avdg well, merge it locally and push the changes to the repo you want12:37
→ DiGitalX joined12:37
← sagsousuke left12:38
→ teweWork joined12:38
thrope yeah it worked now12:38
thanks for the help12:38
→ tatsuya joined12:39
johnpipi i was trying to follow this http://toroid.org/ams/git-website-howto12:40
→ drake01 joined12:40
→ SimonP86_ joined12:40
johnpipi but instead of "git push mysite" to push the master i want to push a tag to the remote and via the post-receive checkout into right dir12:41
→ SimonP86__ joined12:42
→ mastroWork joined12:43
← SimonP86 left12:44
→ nona joined12:44
← foocraft left12:44
→ nigewhite joined12:45
← pgib left12:45
← SimonP86_ left12:45
← Boggle left12:46
nigewhite If I accidentally make some edits on my master branch how can I "transfer" those changes to a new branch that I create?12:46
← ph^_ left12:46
→ unvs joined12:47
← mastroDani left12:47
unvs noobie question: what does it mean to publish a tag? does it push my local tag to the remote?12:47
→ ph^ joined12:48
rudi_s nigewhite: git branch newbranch and then git reset to remove them from the master branch (unless you've already pushed the changes!).12:49
← dyer left12:49
rudi_s unvs: AFAIK Yes.12:49
nigewhite I haven't committed them, just edited the files on disk.12:49
rudi_s nigewhite: Oh. Then just git checkout -b newbranch12:50
→ dyer joined12:50
← dyer left12:50
→ dyer joined12:50
nigewhite OK, I'll try. It's still scary to me though because I don't know what each command physically does, so I'm petrified of losing work.12:50
→ dannas joined12:51
rudi_s nigewhite: git stash save backup; git stash apply (after git adding new files).12:51
That way you at least have the changes stored somewhere (git stash list to see your stashed changes).12:51
Or just add it to the current branch and "move" it later.12:51
← larsmoil left12:52
nigewhite git stash save backup.12:52
→ Boggle joined12:52
nigewhite Is the "backup" bit significant?12:52
rudi_s nigewhite: No, it's just a name.12:52
man git-stash for more information.12:52
jast the 'git-stash' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-stash.html [automatic message]12:52
jsquared nigewhite: everything after "save" is a name you give to the stash.12:52
→ Vampire0_ joined12:53
← Vampire0 left12:53
nigewhite Thanks. The commands are so wordy that I never know with examples given which bit is just an extra bit of command and which is some identifier used by the command.12:53
Ah, the manpage. I tried that before I came here.12:53
The sentences do not mean anything in english.12:54
nicomen hm, it seems I end up with shortlog as action (and HEAD ends up as refname in some regexp) so it doesn't actually output the file when going to /git/project/HEAD12:54
thiago: ^12:54
← SimonP86__ left12:54
rudi_s nigewhite: It takes a while to get used to man pages - but there are good manuals for git, progit for example.12:54
nigewhite THanks, I'll google that.12:55
rudi_s np12:55
→ SimonP86 joined12:56
→ ericbarnes joined12:57
johnpipi is it possible to push a local tag like v4.0.0 to a remote bare repo master?12:57
jaeckel you want to push only this single tag?12:58
johnpipi yeah12:58
jaeckel so not git push --tags12:58
johnpipi overwrite master on remote12:58
jaeckel :)12:58
johnpipi i just want to push out the tag then post-receive will move files to right location12:59
on the server12:59
to deploy basically12:59
nicomen ok, seems like it is indeed supposed to serve files without going thru gitweb12:59
jaeckel then do git push --tags12:59
this should push the tag to the remote12:59
johnpipi my post-receive has this13:00
GIT_WORK_TREE=/home/mysite/public_html git checkout -f13:00
→ bartek joined13:00
johnpipi so if i push to the remote server it will update the live files13:00
but i want to be able to push over any tag13:01
not just master13:01
← cutmail left13:01
← bartek left13:01
rudi_s johnpipi: Why not just use rsync? AFAIK git shouldn't be used as deployment tool.13:02
→ bartek joined13:02
jaeckel johnpipi: you're following http://toroid.org/ams/git-website-howto ?13:02
crashanddie erhm... "shouldn't"?13:02
johnpipi yes13:02
crashanddie says whom?13:02
johnpipi it works good and can deploy to multiple servers13:03
i guess i could make a bash script to rsync to all servers13:03
4 webservers13:03
rudi_s crashanddie: Many people in this channel for example.13:03
→ IslandUsurper joined13:04
→ dv310p3r joined13:04
crashanddie and for what reasons, exactly?13:04
rudi_s rsync is simpler and faster than git (and you don't push the whole history to the server which might cause problems).13:04
crashanddie: For example if you're not careful you give everybody access to your complete history.13:04
And git wasn't designed for it.13:04
crashanddie wait, what?13:04
rudi_s crashanddie: ?13:05
crashanddie "Git wasn't designed to give everybody access to your complete history?"13:05
rudi_s crashanddie: Git wasn't designed to be used as deployment tool.13:05
crashanddie I thought that was *exactly* what Git was designed for.13:05
johnpipi well my git isn't push to public folders13:05
thats what the checkout is for13:05
no .git stuff in the live13:05
jaeckel rudi_s: that's why he uses post-receive hooks as described in http://toroid.org/ams/git-website-howto13:05
rudi_s crashanddie: It is - but not if you have a web server and push the history (including passwords in configuration files for example) to the web server and everybody can read the .git directory.13:05
johnpipi it works but only could make it do master13:05
but im a git noob13:05
rudi_s jaeckel: Yeah, but why use git for that? rsync is just a single line, and faster, and simpler.13:06
jaeckel because you can?13:06
johnpipi one reason i can do13:06
git push all-servers13:06
done13:06
rudi_s jaeckel: Hehe, good point.13:06
jaeckel :)13:07
→ liorkesos_ joined13:07
johnpipi rsync i have to write a for loop 3 lines of code13:07
ha13:07
crashanddie I'd also argue that if all the tools that are available to a linux dev had always been "used as designed", we wouldn't be very far...13:07
johnpipi or one i guess u could13:07
CareBear\ johnpipi : so you want the latest received ref to be published?13:08
← liorkesos left13:08
crashanddie not at all, he just wants to publish other branches than master13:09
← mklappstuhl left13:09
→ henrikbjorn joined13:09
johnpipi well i guess in most cases yes, but what i wanted to be able to do is this13:09
CareBear\ well, that's what I'm asking - what should be published when?13:09
johnpipi git push all-servers v4.0:master13:09
or whatever the syntax is13:09
← liorkesos_ left13:09
CareBear\ you want to be able to publish a given tag or branch13:10
right?13:10
johnpipi so tag v4.0 is pushed to the remote master and then post-receive is ran13:10
CareBear\ well13:10
johnpipi yeah13:10
← LouisJB left13:10
← urkud left13:10
johnpipi remote repo is a bare init repo13:10
CareBear\ yep13:10
johnpipi that is never used except to push stuff and make live13:10
CareBear\ so here's what I'd suggest13:10
ok13:10
do you have changes in live btw?13:10
← tomprince left13:10
CareBear\ that need to go back the same way?13:10
johnpipi nope13:10
CareBear\ (but in reverse)13:10
johnpipi just one way13:11
CareBear\ ok, that helps13:11
johnpipi out to live13:11
CareBear\ I'd suggest:13:11
johnpipi thats it13:11
to deploy13:11
CareBear\ you keep the bare repo13:11
but you let it have all the same tags and branches that work repos have13:11
→ thierryp joined13:11
CareBear\ overwriting master every time you push is not problem free13:11
and not so clean13:11
do you ever want to push something that does *not* get published right then and there?13:12
johnpipi no everytime i push will be to make something live13:12
CareBear\ perfect13:12
johnpipi is it better to nuke the remote first then do it or something13:13
Pieter yeah, nuke it from orbit13:13
CareBear\ then just push branches, and in the post-receive hook the last parameter is the branch name that was updated13:13
Pieter it's the only way to be sure13:13
CareBear\ disagree13:13
so there are two steps to this;13:13
johnpipi so what is the syntax on my dev box to push it over13:14
→ durarara joined13:14
durarara hello13:14
johnpipi a tag13:14
CareBear\ 1. getting the stuff you want to publish into the bare repo13:14
johnpipi and overwrite master13:14
with the tag13:14
CareBear\ 2. getting it out to the live environment again13:14
johnpipi : git push v4.013:14
for example13:14
which is equivalent to git push v4.0:v4.013:14
fr0sty_ johnpipi: git push <remote> t<agname>:master13:14
CareBear\ ie. push the local branch called v4.0 to a branch with the same name in the bare repo13:14
fr0sty_ err, <tagname>:<master>13:15
→ Fullmoon joined13:15
johnpipi let me try it out13:15
CareBear\ I would suggest not using master as a special key word. it doesn't have to be13:15
johnpipi : for my step 1. above I think keeping 1:1 branch names is a good idea13:15
johnpipi : how are you currently doing my step 2.? git export in the post-receive hook?13:16
johnpipi my post-receive looks like this13:16
GIT_WORK_TREE=/home/mysite/public_html git checkout -f13:17
→ drubies|mbp joined13:17
CareBear\ johnpipi : okey. so public_html is also a repo, but one with a worktree?13:18
johnpipi nope13:18
← juvenal left13:18
johnpipi the repo is not in a public dir13:18
CareBear\ ahh no just checked out there sorry yes13:18
hm13:18
→ juvenal joined13:18
johnpipi /home/git/mysite.git13:18
← ormaaj left13:18
CareBear\ if this works it's a really good approach - but I would just add the branch name on the end13:18
mysite.git is what you push into, right?13:19
johnpipi is it better to make a branch instead of a tag13:19
then try to push that over13:19
or is that the same thing13:19
CareBear\ it's not the same thing, both should work equally well to push over, which to use depends on your project work methodology more than anything else13:20
→ jacobat joined13:20
CareBear\ let me test one thing13:20
jacobat Howdy, I'm trying to run git instaweb on a server of mine, but when I hit the server I get a 404 back - any ideas why?13:20
I'm using webrick as httpd13:21
bremner jacobat: check the default port in man git-instaweb13:21
jast jacobat: the 'git-instaweb' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-instaweb.html [automatic message]13:21
→ StephenGWills joined13:22
jacobat bremner: The default port is 1234, that's the port I'm hitting13:22
bremner oh,ok13:22
jacobat bremner: I can see the request in the acces.log file13:22
← xudifsd left13:23
CareBear\ johnpipi : I think you just need to adjust your post-receive a little, and things should work well. try this: GIT_WORK_TREE=/tmp/y git checkout -f "${3}"13:23
→ alberto56 joined13:23
jacobat Here's the relevant part of the log http://pastie.org/172627913:23
johnpipi what does the $3 do13:24
→ Ryback_ joined13:24
← Ryback_ left13:24
→ Ryback_ joined13:24
johnpipi what value will be in there13:24
CareBear\ johnpipi : "ref-name", so it will be e.g. refs/heads/master if you push master, or refs/heads/v4.0 if you push a v4.0 branch, or ref/tags/mar28 if you push a tag13:24
← sunblush left13:25
johnpipi sweet!13:25
→ el_bucket joined13:25
johnpipi thanks let me tinker with it for a few be back and report13:25
CareBear\ sure! note it's the ref-name on the *bare* repo, not the ref name of the repo you pushed from13:25
this is why I think it's a good idea to keep things 1:1 between work repo and the bare repo13:26
johnpipi k good point thanks13:26
bremner jacobat: can you try with a different http server?13:26
→ mitzi_ joined13:26
jacobat bremner: I haven't been able to get apache to serve it, I can give it another try13:26
→ ormaaj joined13:27
← sftp left13:28
→ larsmoil joined13:28
← tatsuya left13:29
← mitzi_ left13:29
← paja left13:29
← gebi left13:30
hbf How do I compare two repos to verify that their contents are identical?13:30
→ mitzi_ joined13:30
← mitzi_ left13:30
← bentob0x left13:30
CareBear\ hbf : that's a complicated question.. :) what do you want to compare more precisely?13:31
← Frtecek left13:31
→ rseifert joined13:31
← dAnjou left13:31
← rseifert left13:31
→ dAnjou joined13:31
jacobat bremner: I think I found it, I'm missing the gitweb installation... doh... thanks13:32
→ sftp joined13:32
→ rseifert joined13:32
→ Bass10 joined13:32
bremner ok ;)13:32
← ph^ left13:32
→ wtffff joined13:32
rudi_s hbf: Run git fsck --full and then compare the sha1 of master, if it's the same, the repos (should be) the same.13:32
jacobat will go and bang head against wall13:33
hbf I want to (a) run a test filter-branch and check that it doesn't change anything before doing the real thing, and (b) after filereing for real, check that the repo I cloned hasn't changed before I push the rewritten repo13:33
rudi_s, thanks13:33
← wtffff left13:33
rudi_s hbf: But that won't work if you rewrite the history.13:33
It only checks if the repos are really the same.13:33
fr0sty_ hbf: filter branch changes things (that's kind of the point)...13:33
→ Vampire0 joined13:33
hbf yes, that's why I want to run a no-op filter first just to see that I haven't messed up13:33
→ mitziyahu joined13:34
mitziyahu hello13:34
fr0sty_ hbf: then just compare the SHAs from the branch tips.13:34
→ EricInBNE joined13:34
mitziyahu oh finally13:34
← Vampire0_ left13:34
hbf right13:34
← mastroWork left13:34
mitziyahu im setting up a new repo, and cant figure out how to add files to it from a local dir.... all manuals i read had pull from some github or w/e13:35
CareBear\ copy files into repo, git add, git commit13:35
rudi_s mitziyahu: Also try the progit book, it's really good.13:36
mitziyahu im too lazy to read a whole book :o13:36
rudi_s mitziyahu: Then don't.13:36
mitziyahu thank you for allowing me ;p13:36
CareBear\ mitziyahu : then stay ignorant, and expect not to get much help here either.13:36
consider hiring a consultant to help you..13:37
← harshu left13:37
→ _Vi joined13:38
→ Vampire0_ joined13:38
mitziyahu lol13:39
jaeckel CareBear\: +113:39
← Vampire0 left13:39
cbreak_work mitziyahu: man gittutorial13:39
jast mitziyahu: the 'gittutorial' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/gittutorial.html [automatic message]13:39
Vampire0_ → Vampire013:39
→ mklappstuhl joined13:39
→ nothingHappens_ joined13:39
→ mastroWork joined13:39
mitziyahu thanks13:40
→ LiamH joined13:40
cbreak_work http://git-scm.com/documentation13:40
→ jimmy1980 joined13:41
d0k_ → d0k13:41
CareBear\ johnpipi : does it seem to work as you would like?13:41
← wingo-pi left13:42
→ v0n joined13:42
johnpipi CareBear\ : was just eating dinner going to test it now13:42
:)13:42
CareBear\ aha!13:42
sorry to disturb the meal :)13:43
johnpipi its cool all done ;)13:43
← Swimming_bird left13:43
→ petrux joined13:45
← aminpy left13:45
johnpipi when i do the push do i want to do git push all-sites <tagname>:master or <tagname>:<tagname>13:46
→ panike joined13:46
johnnywengluu i cloned a git repo from github but i can only see the master branch (git branch)13:46
johnpipi with the $3 will it checkout the tag im pushing13:46
johnnywengluu why arent the other branches fetched?13:46
CareBear\ johnpipi : yes13:46
→ dkannan joined13:46
CareBear\ johnpipi : push all-servers tag13:47
johnnywengluu anyone?13:47
CareBear\ johnpipi : which is equivalent to push all-servers tag:tag13:47
johnnywengluu : I think you have to explicitly say that you want to fetch further branches. have a look at git clone --help13:47
johnpipi said this13:47
remote: fatal: You are on a branch yet to be born13:47
To ssh://10.0.0.76/home/git/mysite13:47
* [new tag] v4.0.0 -> v4.0.013:47
CareBear\ johnpipi : have you already pushed the commits that the tag refers to?13:48
johnpipi : that would be done by pushing some branch13:48
johnpipi let me nuke the repo and try fresh13:48
CareBear\ johnpipi : tags are nothing but names for a commit13:48
johnpipi the bare one13:48
CareBear\ ok, but then you will need to also push some branch into it13:49
→ jsick3 joined13:49
johnpipi ok so push the master first13:49
jaeckel CareBear\: nah, when you push the tag it also pushes the commits13:49
CareBear\ jaeckel : it does? ok!13:49
← el_bucket left13:49
CareBear\ that's good! :)13:49
jaeckel CareBear\: yeah, it was a bit surprising for me as well to have sth like a detached HEAD that refers only to a tag :D13:50
johnnywengluu CareBear\: cant find anything about cloning the others13:50
← EricInBNE left13:50
→ sunblush joined13:50
→ kiorky joined13:50
CareBear\ johnpipi : supposedly no need to push master then, but I'm a little surprised actually :) give it a go13:50
johnpipi git push all-sites +master:refs/heads/master ?13:51
?13:51
CareBear\ johnnywengluu : am reading now; "creates remote-tracking branches for each branch in the cloned repository (visible using git branch -r)" so you should have tracking branches13:51
johnpipi : simply git push all-sites master13:51
is fine13:51
johnpipi k13:51
→ hyperair joined13:51
← mklappstuhl left13:51
kiorky Hi, i'm trying to build git on a snowleopard box, im getting this compilation error : http://pastebin.com/FyT1Wevd. Does anyone has an idea ?13:52
johnpipi says Everything up-to-date13:52
when i do the push v4.0.013:52
but that one is diff then master13:53
← watsoncj left13:53
CareBear\ kiorky : the build is using the wrong compiler. make sure to use the correct --host if calling ./configure13:53
← MattDiPasquale left13:53
CareBear\ johnpipi : --tags ?13:53
johnpipi use tags13:54
k13:54
CareBear\ git push --tags all-sites v4.0.013:54
was what you ran?13:54
→ icwiener joined13:54
johnnywengluu CareBear\: thanks13:55
kiorky CareBear\: This may be more complicated, look at the compilation flags, its already explicitly set to x86_6413:55
CareBear\: (at first i ve tried without anything with the same result)13:55
fr0sty_ CareBear\: that pushes all tags, which is probably not what he wants...13:55
← OOPMan left13:55
→ LouisJB joined13:55
→ lxsameer joined13:55
CareBear\ oh13:56
bleh. my bad13:56
fr0sty_ and a repository doesn't need to have branches...13:56
← joel_falcou left13:56
fr0sty_ (it usually does, but doesn't have to...)13:56
CareBear\ no, but git likes it when they do :)13:56
kiorky CareBear\: i thought of mixed arch dependency libraries but donno how to tell the linker where it links to 32b ones13:56
fr0sty_ true...13:56
→ cjk joined13:56
→ DaveIngram joined13:56
→ tatsuya joined13:56
SethRobertson -L/path/to/32b/ones?13:57
→ roop joined13:57
kiorky CareBear\: and i m not an osx expert, just porting a package manager to it ...13:57
← tatsuya left13:57
CareBear\ kiorky : I'd try to get it to use another compiler than "cc"13:57
→ tatsuya joined13:57
CareBear\ johnpipi : I'll test this too..13:57
lxsameer http://dpaste.com/526297/ , what is the third entry ?13:57
johnpipi i nuked remote bare13:57
kiorky CareBear\: which one for example ?13:57
CareBear\: cc links to gcc42 the one shipped with xcode13:58
johnpipi on local did git push all-sites master13:58
it puts the files13:58
then did git push all-sites v4.0.013:58
← cjk left13:58
johnpipi i pushed stuff over but files from v4 are there13:59
← xiong left13:59
CareBear\ are, or are not?13:59
is v4 in master, or in another branch?13:59
johnpipi then if i run the push 4.0 command again says up-to-date13:59
i just did13:59
git tag -a v.4.0.014:00
v4.0.014:00
to make the tag14:00
is that wrong?14:00
CareBear\ with master checked out?14:00
→ mastro joined14:00
johnpipi yes14:00
fr0sty_ johnpipi: what did you expect to happen?14:00
CareBear\ then v4.0.0 refers to the same thing as master, so there *is* no difference ;)14:00
→ SirFunk joined14:01
johnpipi i just put one html file in master test.html then made that tag then deleted test.html14:01
← dkannan left14:01
johnpipi then commited master so v4 has test.html and master doesn't14:01
CareBear\ ok!14:01
kiorky CareBear\: ?14:01
← evildmp left14:02
johnpipi mmmm14:02
CareBear\ kiorky : sorry, dunno, also no expert with macos14:02
johnpipi losing hair lol14:02
kiorky CareBear\: k14:02
CareBear\ johnpipi : let me test a bit.14:02
johnpipi k thanks man14:02
→ cjk__ joined14:02
johnpipi if i do this though git push all-sites v4.0.0:master14:03
CareBear\ mh yeah no14:03
johnpipi i get this14:03
! [rejected] v4.0.0 -> master (non-fast-forward)14:03
CareBear\ yeah14:03
it can work, but you will be overwriting master everytime, which isn't .. great14:03
johnpipi eyah14:04
yeah14:04
← cjk__ left14:04
johnpipi not best solution14:04
CareBear\ oh, hold on, maybe I made a boo-boo in that hook syntax14:04
let's see14:04
← mastroWork left14:04
CareBear\ yeah, I did14:05
ah, hm.14:06
unless there are actually commits to push, the hook will not run14:06
another thing is I got the hook syntax wrong.14:07
→ evildmp joined14:07
CareBear\ make it: #!/bin/sh read old new ref; GIT_WORK_TREE=/dir git checkout -f "${ref}"14:07
→ guireisc joined14:07
CareBear\ but doing this you'll basically not be able to switch back to something that already exists on the server14:08
:\14:08
→ koliz joined14:08
← gretch left14:08
CareBear\ hm14:08
johnpipi mmmm14:08
CareBear\ ah! yes - if you create a new tag.14:08
that doesn't exist before14:09
then it will go back14:09
→ Vampire0_ joined14:09
→ mando joined14:09
johnpipi what is the syntax to remove a tag on remote?14:09
can u do that14:09
CareBear\ yes, you can14:09
johnpipi but leave live14:09
CareBear\ git push all-sites :tagname14:09
fr0sty CareBear\: any ref update (creating a new one, or moving an old one) will fire the hook.14:09
CareBear\ no the other way around:14:09
git push all-sites tag:14:09
fr0sty Including tag deletion, btw14:09
CareBear\ eh14:09
no14:09
right the first time! :)14:10
← koltroll left14:10
johnpipi k let me update my post-receive and try it14:10
→ hobodave joined14:10
CareBear\ fr0sty : yes! but if the tag already has been pushed, then a new tag pushed, then the first tag can't be pushed to get the hook to run14:10
→ stepnem joined14:10
CareBear\ all right, we also need to cover deletion of stuff, let's see14:10
← Vampire0 left14:10
→ Yuffster joined14:10
CareBear\ yeah, the hook doesn't deal with deleted tags14:11
← meandi left14:12
→ Vampire0 joined14:12
→ sftp_ joined14:12
→ cannonball joined14:12
→ Raging_Hog joined14:12
← ericbarnes left14:12
→ meandi joined14:12
← sftp left14:13
CareBear\ hmm..14:13
fr0sty_ CareBear\: which is why he should push to some common name (e.g. 'git push remote mytag:deploy)14:13
← teweWork left14:13
johnpipi woot it works!14:14
CareBear\ mh, I dunno14:14
johnpipi thanks carebear14:14
→ Swimming_bird joined14:14
CareBear\ johnpipi : yeah, my testing works too, but add one thing:14:14
test "$new" = "0000000000000000000000000000000000000000" && exit 014:14
johnpipi just have to delete the remote first if i want to update if its the same14:14
CareBear\ after the read, before the checkout14:14
fr0sty_ I know you dont like the X:1 relationship but it's helpful sometimes...14:14
johnpipi what does that do?14:14
← Vampire0_ left14:14
CareBear\ johnpipi : it avoids trying to checkout when you delete a tag14:14
johnpipi : so that there is no error message, because you can't checkout a tag that has been deleted :)14:15
johnpipi yeah14:15
RichiH how well supported are binary diffs, these days? i seem to remember that git used to be quite inefficient when it came to binary data and i wonder if that's been improved14:15
CareBear\ johnpipi : should work well14:16
johnpipi yeah its great one command push to all servers14:16
nice14:16
→ Spockz joined14:16
johnpipi thanks a million14:16
CareBear\ you're welcome14:16
I'm also writing hooks :)14:16
← roop left14:17
→ bitkiller joined14:17
Boggle I just did:14:18
PiBook:NoXIB_iAd TEMPLATE sunfish7$ git commit -a -m 'ffs cant click adBanner'14:18
[master 1b3bbda] ffs cant click adBanner14:18
5 files changed, 4084 insertions(+), 3986 deletions(-)14:18
fairuz Hi, how to revert all recent changes (that not commited and not added yet)14:18
CareBear\ fairuz : git checkout14:18
Boggle Does this mean it sent it to my online GH repository?14:18
fairuz CareBear\: ty14:18
fr0sty_ Boggle: no14:18
Boggle or is it still local to my machine?14:18
→ TiBook joined14:18
← Spockz left14:18
karstensrage http://cmurphycode.posterous.com/git-10214:18
fr0sty_ commit is a local operation in git.14:19
karstensrage can someone explain to me how doing 'git reset --mixed HEAD~1' helps split one commit into two?14:19
← sunblush left14:19
→ Spockz joined14:19
Boggle so ' pushing ' is what gets it onto the server yes?14:19
fr0sty_ only 'git push' modifies remote repositories.14:19
CareBear\ Boggle : yep14:19
→ el_bucket joined14:19
→ jfkw joined14:19
Boggle The first time, I did:14:19
62 git remote add origin [email@hidden.address]14:19
63 git push origin master14:20
← drubies|mbp left14:20
Boggle so do I just repeat line 63?14:20
and that will push my second revision onto GH servers?14:20
→ drubies|mbp joined14:20
← el_bucket left14:20
fr0sty_ karstensrage: you can use 'git add' (maybe 'git add -p') to add only part of your changes, commit that, and repeat14:20
Boggle: yes.14:21
Boggle thx14:21
fr0sty_ assuming you are on branch 'master' locally.14:21
karstensrage fr'fr0sty, how do you add "part" of your changes?14:21
fr0sty_ karstensrage: look up: 'git add -p'14:21
jacobat It's awesome14:21
fr0sty_ allows you to add changes hunk by hunk.14:22
← TiBook left14:24
← henrikbjorn left14:24
→ teweWork joined14:25
→ tgunr joined14:26
→ sunblush joined14:26
karstensrage wow14:27
→ spearce joined14:27
karstensrage thats cool14:27
karstensrage groks the power14:27
→ kukks joined14:28
reuss what's the proper way of obtaining all references to a single commit? .. just like --decorate shows with log ...14:29
→ bkosborne joined14:29
thiago reuss: the SHA-1 is enough14:30
reuss thiago: i want to test if a given sha-1 also is referenced in other ways ..14:30
→ ericbarnes joined14:30
karstensrage thanks fr'fr0sty14:32
reuss say i pushed a ref refs/my_namespace/foobar -- how do i test if $commit is referenced as refs/my_namespace/foobar ? .. (this is in the bare repository, so no need to worry about fetching the refs in the namespace)14:32
← DaveIngram left14:33
← jfelchner left14:33
← matjas left14:33
thiago reuss: git rev-parse refs/my_namespace/foobar14:33
reuss: if it's the same SHA-1...14:33
panike window close 414:34
→ PabloM joined14:34
→ booginga joined14:34
→ bjtitus joined14:34
reuss thiago: oh yeah - that might work14:34
CareBear\ yawn14:34
→ OOPMan joined14:36
→ ereslibre_laptop joined14:36
← ereslibre_laptop left14:36
→ ereslibre_laptop joined14:36
← sunblush left14:37
← shame left14:37
→ mastroWork joined14:37
→ agile joined14:37
← ereslibre left14:37
→ el_bucket joined14:38
→ DaveIngram joined14:38
← Yuffster left14:38
← drake01 left14:38
→ foocraft joined14:39
← panike left14:39
← mastro left14:40
← rseifert left14:40
elb0w git-svn, does this just allow the use of git commands on a svn repo?14:40
CareBear\ in a way..14:41
elb0w for the most part anyway14:41
CareBear\ it allows a local git repo to interact with a remote svn repo14:41
elb0w ok14:41
jacobat Or converting a svn repo to a git repo14:41
← sdboyer left14:42
← nigewhite left14:43
← el_bucket left14:43
← DaveIngram left14:43
→ panike joined14:44
→ ceej joined14:44
→ deuterium joined14:45
→ krosenvold joined14:45
← opalka left14:46
steven does the index start off with whatever is in HEAD (assuming the user hasnt done 'git add' etc)?14:46
→ dr_ick joined14:46
→ oriba joined14:46
→ wolfe joined14:46
← giallu left14:47
→ DaveIngram joined14:47
← mastroWork left14:47
fr0sty_ steven: that's my understanding.14:48
← johnpipi left14:48
→ johnpipi joined14:48
→ dingomanatee joined14:49
← dAnjou left14:49
→ dAnjou joined14:50
→ sdboyer joined14:51
← thierryp left14:51
→ excess^ joined14:51
← excess^ left14:51
→ wilmoore joined14:51
→ ^Simon^ joined14:52
← sdboyer left14:52
← wilmoore left14:53
→ mastroWork joined14:53
Caelum what was the git branch --track syntax to make a branch tracking a remote branch14:53
→ rayners joined14:54
→ alex__c2022 joined14:54
Caelum when I try to do git branch --track remote local it says fatal: Not a valid object name14:54
panike Caelum: That should be reversed---i.e., git branch --track local remote14:55
→ beatak joined14:56
Caelum panike: aha, thank you very much14:56
← koliz left14:57
→ dkasak joined14:57
← PigeonFriend left14:58
esc Caelum: if the remote-tracking-branch exists already, say origin/pu, then doing 'git checkout pu' will automatically create a local branch, and set the upstream realationship14:59
← bauruine left14:59
Caelum esc: that's what I usually do, but I wanted to find out the older syntax for an email in case the person has an older git15:00
→ gebi joined15:00
→ airborn joined15:00
→ gretch joined15:00
← mitziyahu left15:01
← PabloM left15:04
→ excess^ joined15:06
→ PabloM joined15:06
← angelsl left15:06
→ roop_ joined15:07
← Bass10 left15:07
→ _psychic_ joined15:07
← mastroWork left15:08
← nothingHappens_ left15:08
← teweWork left15:09
← simplechat left15:11
→ Robin_Watts joined15:12
→ mastroWork joined15:12
→ dim_ joined15:13
← ^Simon^ left15:13
← radarek left15:13
← cbreak_work left15:13
← foocraft left15:14
ChanServ set mode: +v15:15
Robin_Watts Hi all. I hope this isn't an FAQ - I've done a web search, but couldn't find an answer...15:15
→ soc42 joined15:15
Robin_Watts We have a larg(ish) open source/commercial project that has until now been held in svn. We want to move over to using git for it.15:15
CrazyGir how would I roll a repo back to a specific point in time/commit15:17
nDuff presumes Robin_Watts has read the manual for git-svn in full?15:17
Robin_Watts The only real snag we can see is the loss of 'monotonically increasing easily human readable revision numbers'.15:17
nDuff CrazyGir, git reset15:17
fr0sty_ Robin_Watts: that's no loss...15:17
CrazyGir awesome, I'll go read!15:17
Robin_Watts nDuff: Yes. We are going to full blown git, not git svn.15:17
jaeckel fr0sty_: +115:17
shruggar Robin_Watts, do you know about "git describe" ?15:17
nDuff Robin_Watts, ...ahh, yup. If that's a showstopper for you, there's Bazaar.15:17
Robin_Watts Yes, thanks, I know about git describe. Let me finish the question...15:18
nDuff Robin_Watts, (which does a left-sort on the tree to assign revision numbers -- something that certainly could be done with git as well)15:18
→ joel_falcou joined15:18
← rando left15:18
→ tjholowaychuk joined15:18
fr0sty_ having authoritative revision numbers in a distributed system is not practical, but do finish your question...15:18
← johnpipi left15:19
nDuff fr0sty, ...well, it's practical if you have a defined central-repository from which they're measured...15:19
→ alvaro_o joined15:19
← seivan left15:19
fr0sty_ nDuff: where would you store the info? how would it be useful?15:19
nDuff fr0sty, ...if I say revno 4321 of Bazaar, everyone knows what I'm talking about15:19
fr0sty_ refs/fake_revs/***15:19
Robin_Watts In interactions with customers, many of them get snapshots from us. For us, (the developers) it's easy enough to look to see if we have a given commit included in our checkouts. For customers, it's a harder job.15:19
→ mastro joined15:20
→ bdiu joined15:20
← alex__c2022 left15:20
shruggar Robin_Watts: you should create an annotated tag for anything you actually release to customers15:20
nDuff fr0sty, ...and likewise, if I say 4321.3 (that's saying 4321 was a merge commit, and then I want the 3rd changeset which came in through it)15:20
← alvaro_o left15:20
Robin_Watts Given that we have a central 'public' repository to which only authorised developers can push, we do effectively have one 'special' repo.15:20
fr0sty_ nDuff: if I create a hotfix branch from last years release (213) and make three revisions, what numbers to they get?15:21
Robin_Watts I was wondering if there was a script anywhere that we could run as a hook on that repository to tag each new commit with a revision number.15:21
r1000, r1001, r1002 etc.15:21
fr0sty_ Robin_Watts: just roll your own if it is important.15:21
nDuff fr0sty, the numbers are where they're merged, not where they're branched from, and IIRC are branch-local.15:22
Robin_Watts Tags are hugely lightweight in git, right? so it wouldn't be a huge burden.15:22
nDuff RobertLaptop, yes, there are hooks for that. Personally, I think it's silly.15:22
CrazyGir thanks nDuff!15:22
fr0sty_ nDuff: so XXX only has context with a branchname?15:22
← soc42 left15:22
→ Bass10 joined15:22
fr0sty_ is confusedd, and undercaffeinated (as usual)...15:23
← Fullmoon left15:23
nDuff fr0sty, if my memory (which is increasingly vague) is accurate, yes. I can say "4321" to a Bazaar developer and they assume it's upstream mainline unless specified otherwise.15:23
← mastroWork left15:23
jaeckel Robin_Watts: either you write a script (like we did) that handles the former svn revision IDs and so on, or what I've found some days ago http://www.kernel.org/pub/software/scm/git/docs/gitattributes.html#_tt_ident_tt15:23
Robin_Watts fr0sty_: I was hoping that someone else might have already written the script (because 1) I'm lazy, and 2) they could probably arrange for merges to mention the revision numbers of parents too etc)15:24
But the main reason I was asking here was to check that there wasn't some fundamental reason why this was frowned upon.15:25
→ goshawk joined15:25
shruggar well, there are a couple of fundamental reasons why it's frowned upon :)15:25
Robin_Watts jaeckel: The guy converting from svn to git has taken care of everything (sorting out rather wierd repo format out w.r.t tags and branches). This is for future work.15:26
shruggar: Such as?15:26
jaeckel Robin_Watts: nah, with SVN revision IDs I wanted to say to handle $Revision$ and so on and fill in the values15:27
← eschnou left15:27
→ liquidproof joined15:28
→ mastroWork joined15:28
shruggar those already mentioned: an incrementing number is by nature non-distributed, while git is distributed. Using a strictly non-distributed method of referring to commits is likely to prevent you from thinking about things in a git-like manner. Also, if you auto-tag everything, you 1) can't roll back, since it's tagged. 2) will always see it in "git tag" output- most people don't like 10000 tags showing up when they run "git tag" :)15:28
Robin_Watts I can see that if were were working in a truly distributed sense, it wouldn't work - but while we will all have our own separate repos, the 'public' one is special.15:28
nDuff right, but those tags get pulled down15:28
← jhoepken left15:29
Robin_Watts shruggar: The public one is just that, public. Once stuff goes into that, we can't roll back anyway.15:29
nDuff so they'll be an annoyance to people who try to get a tag list15:29
which is usually something with more useful data, ie. released versions15:29
Robin_Watts But yes, I take your point about the tag list.15:29
avdg shruggar: what about hg, they have incrementing numbers too15:29
→ bauruine joined15:30
Robin_Watts marvels that git can't have tags that don't show up in tag lists by default. It seems more configurable than a configurable thing.15:30
← mastro left15:31
← DaveIngram left15:32
← psoo left15:32
← g0bl1n left15:32
fr0sty_ Robin_Watts: refs/not_Tags/r1234515:32
refs/revs/12345, maybe15:32
'then git checkout rev/12345'15:33
← tyteen4a03 left15:33
← dAnjou left15:33
→ ryanakca joined15:34
← oriba left15:34
Robin_Watts fr0sty_: And that won't show up in a git tag list? Sounds ideal.15:34
fr0sty_ they will not be easily searchable, but they will be there.15:34
→ dAnjou joined15:35
→ DaveIngram joined15:35
fr0sty_ again, I think the whole project is lunacy, but you are free to do as you wish...15:35
← xiangfu left15:36
← mikelangelo left15:38
→ Anubisss joined15:38
← bkosborne left15:40
→ nadavoid joined15:42
→ bkosborne joined15:44
→ xiong joined15:45
← evildmp left15:45
→ paja joined15:45
← vitor-br left15:46
← DaveIngram left15:47
← dAnjou left15:47
← cantoma left15:47
→ dAnjou joined15:47
← bkosborne left15:48
→ hobodave_ joined15:49
→ rando joined15:49
→ smolyn joined15:49
→ cantoma joined15:49
← hobodave left15:50
→ sagsousuke joined15:51
→ vitor-br joined15:51
← petrux left15:52
← nibblebot left15:53
→ DaveIngram joined15:53
→ mastro joined15:56
→ kumbayo joined15:57
← mastro left15:58
→ evanmeagher joined15:58
← mastroWork left15:59
→ doener joined15:59
→ jophish joined16:00
→ __name__ joined16:00
← __name__ left16:00
→ __name__ joined16:00
← leo2007 left16:01
← roop_ left16:01
← elyob left16:02
→ mehwork_ joined16:02
← ereslibre_laptop left16:03
→ robotarmy joined16:03
← mehwork left16:05
→ rgr joined16:07
→ saysjonathan joined16:07
← ping-pong left16:07
→ vbar joined16:08
ChanServ set mode: +v16:09
vbar hello?16:09
CareBear\ Hi Lennart. Go ahead and ask your question. :)16:10
← nevyn^ left16:10
→ nevyn^ joined16:10
vbar well16:10
← fr0sty left16:11
vbar I have done sth to my repos such that git merge will tell Already up to date. but git pull gets the changes16:11
← guireisc left16:11
← Bass10 left16:11
← xiong left16:12
→ ping-pong joined16:12
→ nevyn^_ joined16:13
vbar i am using bare remote repo vbhome, for example16:13
→ thm joined16:13
← sagsousuke left16:13
→ kenichi joined16:13
vbar and in my working repo I normally do a git fetch vbhome and git merge vbhome/master16:13
thm can I push branches to another location so that they are remote branches there?16:13
→ _ollie joined16:14
← _ollie left16:14
CareBear\ vbar : unless configured otherwise, pull = fetch + merge16:15
thm : no16:15
thm : use ssh :)16:15
thm CareBear\: how? I can't connect back and pull.16:15
CareBear\ ssh host git remote add ..16:16
→ drake01 joined16:16
thm and then?16:16
← nevyn^ left16:16
nevyn^_ → nevyn^16:16
CareBear\ thm : then you have added the remote tracking branches16:16
thm : in the remote repo16:16
→ noahcampbell joined16:17
thm and a push will use them?16:17
← nevyn^ left16:17
→ gmclelland joined16:17
thm I mean a push to host16:17
CareBear\ ?16:17
no16:17
let's try another way. please explain what you would like to achieve workflow wise?16:17
← evanmeagher left16:17
thm I have localhost and remotehost, and want to push from localhost to remotehost in such a way as if I had pull on remotehost from localhost.16:18
wereHamster thm: use git push16:19
CareBear\ thm : push from local is equivalent to fetch on remote16:19
thm : merge or rebase you will have to implement in a hook16:19
doener or fetch from the remote, merge locally and push the result16:20
→ dreiss joined16:20
thm but on remotehost it is not a bare repo. it is a repo with a workspace. so I want to avoid pushing master:master, because that would screw the workspace16:20
can I push master:remote/foo/master ?16:20
← TrahDivad left16:20
CareBear\ thm : not sure you can push into a repo with worktree at all16:20
→ HG` joined16:20
fr0sty_ CareBear\: faq non-bare16:21
jast CareBear\: Pushing to non-bare repositories is discouraged: please see https://git.wiki.kernel.org/index.php/GitFaq#non-bare [automatic message]16:21
CareBear\ thm : what you do is set up a bare hub repo in between16:21
fr0sty_ you _can_, but shouldn't. (generally)16:21
→ EmperorGonzo joined16:21
← EmperorGonzo left16:21
thm aha. so the bare hub repo would also reside on remotehost, in my case.16:22
CareBear\ right16:22
→ jungejason joined16:22
CareBear\ http://joemaller.com/990/a-web-focused-git-workflow/16:22
← rgr left16:22
thm I thought I could push to some branch and then on remotehost do a merge.16:23
CareBear\ what if there's a conflict16:23
you probably want to be resolving them locally16:23
ChanServ set mode: +v16:23
thm how do I populate the hub?16:24
wereHamster what is the hub?16:24
CareBear\ it is explained later in the article16:24
wereHamster : a bare central repo16:24
thm : you push into hub16:24
← SirFunk left16:24
gmclelland After forking a project(compass) in github how do I get the updates to compass into my fork? Do I then add the compass project branches as remotes and then pull down all the branches?16:24
wereHamster gmclelland: github has a kb article o that16:25
vbar I am not sure, but maybe my problem is similar to http://stackoverflow.com/questions/4176195/git-merge-x-reports-up-to-date16:26
→ towski joined16:26
vbar what are the steps to find out what the problem really is?16:26
gmclelland <wereHamster>Thanks - it looks like I can git fetch upstream, git merge upstream master16:27
vbar I cloned my bare repo into an empty folder and all changes seem to be there.16:27
← achadwick left16:28
vbar But if I fetch the bare repo, the changes are not even in the remote branch16:28
→ elyob joined16:28
← dAnjou left16:29
← empity left16:30
→ amerine joined16:30
← jsquared left16:30
→ rgr joined16:31
→ slonopotamus joined16:31
→ dAnjou joined16:31
vbar anyone has an idea ?16:32
wereHamster vbar: transcript please16:33
vbar transcript?16:33
→ mitziyahu joined16:34
vbar what do you mean exactly?16:34
← hbf left16:34
→ ereslibre joined16:34
← ereslibre left16:34
→ ereslibre joined16:34
mitziyahu i need a git tree that is also a working copy that many developers would commit to it... im looking on google for hours jumping through books. any pointers on what to search on google?16:34
CareBear\ mitziyahu : why?16:35
mitziyahu cause i want all the developers to see their results in one place (apache site) after they commit16:35
i think im failing at seeing the idea behind version control :16:36
CareBear\ no. each developer should probably have their own repo instead.16:36
see the link I just posted16:36
mitziyahu i wasnt logged in16:36
what was it?16:36
CareBear\ http://joemaller.com/990/a-web-focused-git-workflow/16:36
gmclelland <wereHamster>On http://help.github.com/fork-a-repo/ it say's to git fetch upstream, and git merge upstream/master. When I try that, it tells me upstream isn't a repository.16:36
mitziyahu and each developer has their own repo, the problem is i cant put it all to a working directory in the end16:36
thanks16:36
CareBear\ mitziyahu : yes, you can.16:36
mitziyahu : putting it all together in the end is something git is really really good at, actually16:37
mitziyahu i know16:37
im just failing to do it16:37
CareBear\ :)16:37
mitziyahu i think im too tired16:37
let me see if i get this straigh16:37
CareBear\ see that web page, it has a good explanation16:37
mitziyahu ok16:37
thanks16:37
← robotarmy left16:37
mitziyahu i dont know why but my dev machine (which is a vm machine) keeps dying when doing an rsync / push lol16:39
← kpreid left16:39
→ robotarmy joined16:39
← drake01 left16:40
quentusrex why is it that a "git remote add --tags upstream git://the.remote.server.com/repo.git" only checks out the tags? and not the branches and the tags?16:40
← dAnjou left16:40
quentusrex Why does a separate remote have to be added if I want both branches and tags from a remote repo?16:40
CareBear\ quentusrex : adding a remote has nothing to do with tags16:41
quentusrex CareBear\, when I add the remote with that line, then do a git fetch, only the tags are imported.16:41
CareBear\ so don't use --tags16:41
→ dAnjou joined16:41
CareBear\ --tags isn't a valid option to git remote add16:42
← doener left16:42
CareBear\ in 1.6.3.3 at least16:42
though that's a bit old now16:42
quentusrex CareBear\, http://www.kernel.org/pub/software/scm/git/docs/git-remote.html16:42
git remote add [-t <branch>] [-m <master>] [-f] [--tags|--no-tags] [--mirror] <name> <url>16:42
→ yairgo joined16:42
CareBear\ ok16:43
and the page explains what --tags means16:43
mitziyahu CareBear\ i wish he would do ls -l in his dir after the directories were created lol i cant focus enough16:43
← btipling_ left16:43
← Twey left16:43
CareBear\ mitziyahu : then you need to return to solving this problem when you can concentrate a bit more - or buy consulting from someone who can solve the problem for you so that you do not have to care16:44
quentusrex CareBear\, it explains that it imports every tag from the remote repo, but it does not say that it will import only the tags. And ignore any branches specified with "-t branch-name"16:44
CareBear\ quentusrex : did you tell git to fetch something other than tags?16:44
← airborn left16:44
← McMAGIC-- left16:44
→ abhaythehero joined16:45
← kassah left16:45
quentusrex without the -t flag no branches are fetched, even with the -t flag specifying a branch, no branches are ever fetched.16:45
→ Vampire0_ joined16:45
quentusrex even though in 'git remote show upstream' it specifically states the branches will be pulled in, they never are pulled.16:46
→ dAnjou_ joined16:46
mitziyahu CareBear\ haha i did it :p i just cant copy his hook scripts16:46
← Vampire0 left16:46
← dAnjou left16:46
dAnjou_ → dAnjou16:46
← LouisJB left16:47
← gmclelland left16:47
→ giallu joined16:48
→ Tv joined16:48
← iAsterisk left16:48
← Vampire0_ left16:48
→ airborn joined16:49
→ cjz joined16:49
→ adurity joined16:49
→ empity joined16:50
→ jajmon joined16:51
jajmon hello good sirs!16:51
← artefon left16:51
jajmon does git have anything against spaces in file names?16:52
crashanddie Nope16:53
← Sonderblade left16:54
→ alesan joined16:54
alesan hi!16:54
I have made some modifications to files locally. I want to pull and I know there is going to be some changes there.16:55
should I commit first or pull then commit then push?16:55
CareBear\ stash16:55
alesan mh why stash?16:55
← shruggar left16:55
CareBear\ or commit, fetch, rebase16:55
→ mikepack joined16:55
alesan fetch is...?16:55
CareBear\ half of pull16:56
wereHamster git fetch16:56
jajmon so any idea why my git diff shows changes as if the entire file was added?16:56
wereHamster and commit is git commit16:56
and rebase is git rebase16:56
CareBear\ jajmon : line endings16:56
fetch+rebase is equivalent to pull --rebase16:56
alesan so, why not simply commit and pull?16:56
slonopotamus CareBear\: your backslash is really confusing :)16:56
CareBear\ because pull defaults to merge16:56
jajmon CareBear\, could you elaborate a little?16:57
alesan well I want to merge... maybe there is something that I don't get16:57
CareBear\ slonopotamus : part of the reason it's there, but not so much for humans as for software. had nick since 1995. applications still can not handle it.16:57
wereHamster alesan: commit first16:57
CareBear\ alesan : if you want to merge (instead of rebase) then commit + pull is perfect16:57
→ ph^ joined16:57
← guardian left16:57
← davzie left16:57
alesan I need to read the difference between merge and rebase16:58
CareBear\ jajmon : if line endings in the file change then git will consider every line to have changed16:58
wereHamster alesan: yu said you have local modifications. Are they final? Is the feature implemented completely, the bug fixed?16:58
→ guardian joined16:58
wereHamster alesan: or are those changes which are half-done and you need to work o them late?16:58
jajmon i c16:58
must look into my vim settings, hmpf16:58
alesan wereHamster, which bug? in any case let;s say "yes"16:58
CareBear\ jajmon : because, in fact, every line *has* changed16:58
alesan no they are done16:58
wereHamster alesan: then why odn't you want to commit16:58
?16:58
← rando left17:00
→ radarek joined17:01
→ dajhorn joined17:02
→ McMAGIC-- joined17:02
← rgr left17:02
→ kassah joined17:02
← guardian left17:02
→ rgr joined17:03
→ artefon joined17:03
→ Vampire0_ joined17:04
← workmad3 left17:05
→ rando joined17:06
→ T-Pup joined17:06
→ antant joined17:09
antant Hey guys17:10
← Boggle left17:10
antant Hoping someone can help me sort out this 'fatal: Not a git repository (or any of the parent directories): .git' error17:10
← LekeFly left17:10
→ p3rg2z joined17:11
← p3rg2z left17:11
← thm left17:11
← resmo_ left17:11
vbar are you in the right directory?17:11
antant after I cloned it initially I renamed the folder for ease of access, but I've renamed it back now. Would that cause the error? If so how do I fix it17:11
→ superdmp joined17:12
← slonopotamus left17:12
← dreiss left17:12
superdmp i am puzzled - I have cloned https://github.com/SmileyChris/easy-thumbnails/17:12
on one of my machines I can git checkout simpler-get_thumbnail17:12
← rgr left17:12
superdmp on the other, I can't17:12
→ saschpe_ joined17:12
→ SirFunk joined17:13
superdmp why might this be?17:13
antant vbar: both pull and fetch failed with the same error.17:13
wereHamster superdmp: git --version17:13
← nothingHappens left17:13
wereHamster antant: did you cd into the cloned directory?17:13
antant nope17:13
superdmp wereHamster: the one that works is 1.7, the one that doesn't 1.517:13
→ jstemmer joined17:14
wereHamster superdmp: ...17:14
superdmp I guess that is my answer17:14
antant ~/Desktop/OpenELEC.tv is the cloed directory of https://github.com/OpenELEC/OpenELEC.tv.git17:14
I'm in ~/Desktop17:14
superdmp apt-get upgrade17:15
oops17:15
→ shruggar joined17:15
wereHamster antant: cd ~/Desktop/OpenELEC.tv17:15
→ variable joined17:15
→ rgr joined17:16
→ g0bl1n joined17:16
→ Zephyrus joined17:16
variable is there a way to normalize times of edits before I push them? (ie I want it to look like all my edits happened on a certain hour or whatever) (just for interest - no real reason)17:16
antant ah17:16
superdmp ok, a newer version of git is not available through apt-get, so that's a nuisance17:16
wereHamster variable: not without rewriting the commit17:16
alesan wereHamster, sorry, I had to eat a wafer17:16
→ LekeFly joined17:16
→ Dave^| joined17:16
alesan wereHamster, sure I want to commit, I was just asking what's the best procedure17:16
antant Cheers wereHamster. Didn't know you have to do it from inside the dir17:16
variable wereHamster: what do you mean?17:17
superdmp is there some way I can pull down the branch I need from github without upgrading git?17:17
→ dreiss joined17:17
wereHamster alesan: git commit, presumably17:17
← liquidproof left17:17
← juvenal left17:17
→ kipras joined17:17
wereHamster superdmp: fetch + checkout17:17
or + merge17:17
antant It's now asking me to specify a branch.17:18
→ DR01D5 joined17:18
variable wereHamster: I'm fine with changing history - but I'm curious if there is some kind of git command to do so or if I need to go "lower level"17:18
→ systemaddict joined17:18
wereHamster variable: git rebase or filter-branch17:18
→ bran joined17:19
→ matjas joined17:20
← rgr left17:20
→ rgr joined17:21
→ henrikbjorn joined17:21
superdmp sorry wereHamster , I can't work out the sequence of operations to do that17:22
the branch I want is https://github.com/SmileyChris/easy-thumbnails/tree/simpler-get_thumbnail17:22
→ MikhX joined17:23
→ nothingHappens joined17:23
superdmp git fetch just gets the latest updates, no?17:23
antant You would hope it would be that easy17:23
→ friskd joined17:24
← rgr left17:25
→ \yrlnry joined17:25
→ Titosemi joined17:25
CareBear\ superdmp : yes, fetch only downloads commits that are in remote repos to your "remote tracking branches"17:25
→ rgr joined17:26
← saysjonathan left17:26
superdmp CareBear\: ok - but I am trying to understand the suggestion above that fetch + checkout will let me get the branch I need17:26
← tatsuya left17:27
→ saysjonathan joined17:27
→ texel joined17:27
→ phrearch joined17:27
phrearch hello17:27
texel → Guest4122017:27
→ jrmuizel joined17:27
phrearch i forked a repo on github like http://help.github.com/fork-a-repo/ , but i was wondering if i can switch to an upstream branch now17:27
← deadlyicon left17:27
→ LouisJB joined17:28
→ squiggy joined17:28
alesan so how was to commit only part of a file?17:30
→ iAsterisk joined17:30
← amerine left17:30
→ phantomcircuit joined17:31
→ slonopotamus joined17:31
→ varioust joined17:32
← whaley left17:32
wereHamster alesan: man git add -> -p17:33
jast alesan: the 'git' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git.html [automatic message]17:33
← CarlFK left17:33
← DarthGandalf left17:33
→ koltroll joined17:33
→ sylr joined17:35
← sylr left17:37
← radarek left17:37
→ _ollie joined17:38
→ CarlFK joined17:38
→ t3rm1n4l joined17:38
steven hello17:38
→ justin-george joined17:39
→ beyondcreed joined17:39
← dreiss left17:39
steven is there a way to change the message (or author) of a commit thats sevral commits ago?17:39
shruggar steven, does anyone else have a copy of that commit?17:39
← _ollie left17:39
← koltroll left17:40
← simenbrekken left17:40
← CarlFK left17:42
steven nope17:43
→ koltroll joined17:43
steven (sorry for the late reply shruggar)17:43
→ Twey joined17:43
→ cruzton joined17:44
shruggar steven: git rebase -i HEAD~(some number one or two larger than the number "ago" it was) select "edit". rebase will drop you into the history at a point where you can git commit --amend the wrong commit17:44
← starcoder2 left17:45
→ guardian joined17:45
ChanServ set mode: +v17:45
cruzton shruggar: woah, i have a question relating to that very topic17:46
shruggar: how can i get "reword" to also edit the commit times and author/commiter emails?17:46
← iAsterisk left17:46
→ Fretta joined17:46
← Bauer left17:47
→ amerine joined17:48
← jrmuizel left17:48
wereHamster cruzton: man git-commit-tree -> GIT_AUTHOR_DATE etc17:49
jast cruzton: the 'git-commit-tree' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-commit-tree.html [automatic message]17:49
→ WAAQLDW joined17:49
wereHamster but I still don't see the point of editing the commit times17:49
→ starcoder2 joined17:49
cruzton wereHamster: so people don't know when stuff was done? :) ill look into commit-tree, thanks17:50
← nicoulaj left17:50
→ mattly joined17:51
← Magix left17:51
← varioust left17:52
← bjtitus left17:52
→ sylr joined17:53
→ iAsterisk joined17:53
→ rcreasey joined17:55
→ Laggg joined17:56
→ Bauer joined17:56
rcreasey If I have two branches on my repo, 'master' and 'develop', and there's a symbolic-ref for HEAD pointing to master; is there a way to create another symbolic-ref to point to the 'most recent revision on the develop branch'?17:57
and then tell a remote about the new symbolic-ref?17:57
← sylr left17:57
→ jsquared joined17:57
rcreasey I basically want to have a HEAD-like reference to the develop branch, but still keep HEAD pointing at master.17:57
← nonnooo left17:59
→ lessthanzero joined17:59
← Titosemi left17:59
← tvw left17:59
← airborn left18:00
alesan wereHamster, so I did commit18:01
→ radarek joined18:01
alesan and now I did git fetch18:01
I should probably rebase now yes?18:01
← gebi left18:01
→ airborn joined18:01
← jajmon left18:02
← Barret1 left18:02
→ kingfishr joined18:03
← rgr left18:03
← vbar left18:03
← Laggg left18:04
← elyob left18:04
← t3rm1n4l left18:05
→ Laggg joined18:05
steven shruggar: will that also allow me to change the commit date?18:07
← lxsameer left18:08
→ lanthan joined18:08
shruggar steven, anything you can normally do with commit --amend18:09
steven oh, i dont think yo ucan chan ge the date with commit --amend18:09
← phantomcircuit left18:10
← mitziyahu left18:10
→ phantomcircuit joined18:10
→ maestrojed joined18:10
← Laggg left18:11
→ Laggg joined18:11
steven crap.18:11
ooh, --date18:12
← artefon left18:13
← erichammond left18:14
← thrope left18:15
alesan so git pull detected a "conflict" while rebasing18:15
it's a git pull --rebase18:15
I just want the remote server version as I have not touched those files; how do I do that?18:15
→ opalka joined18:16
← opalka left18:16
→ opalka joined18:16
← MikhX left18:16
→ MikhX joined18:16
→ dreiss joined18:17
→ Mannequin joined18:17
→ bentob0x joined18:18
← _Vi left18:18
→ Evious joined18:18
← radarek left18:18
← Laggg left18:20
→ Chillance joined18:21
→ Titosemi joined18:21
→ Laggg joined18:21
← Laggg left18:22
→ Laggg joined18:22
→ mitziyahu joined18:22
→ LouisJB_ joined18:22
← LouisJB left18:23
LouisJB_ → LouisJB18:24
→ eletuchy joined18:25
SethRobertson Is the best way to discover if a branch is fully merged to do a `git merge-base br1 br2` and the compare the result to the head of br1 and br2, and only complain if one or the other is not the merge base? `git cherry br2` returns SHA even though the branch is merged.18:25
→ soreau joined18:26
→ drake01 joined18:28
mrevd sethrobertson, how about git branch -a —contains featureBranch18:28
soreau Hey guys, I'm trying to figure out what commit is responsible for a certain function's args changed but I'm having some trouble. I tried git log -p -- the/file.c and / searched for the function name, but it didn't point to the correct commit. It just said ...skipping... and never showed the commit id pointing to the actual change18:28
systemaddict i just made a commit to my local repo but haven't pushed it yet - is there some way i can delete the commit, making it as though it never happened?18:29
← LouisJB left18:29
soreau So I can find the change in git log, but not the commit message or id to say which commit this actually happened in18:29
→ urkud joined18:29
CareBear\ systemaddict : well, yes, but it's kinda messy.. can you replace it with something else instead?18:29
systemaddict : then make the changes you do want, git add as usual, and then git commit --amend18:29
systemaddict that works thanks CareBear\18:30
CareBear\ soreau : git blame18:30
BinGOs git log -S perhaps18:30
← deuterium left18:31
← eletuchy left18:31
SethRobertson mrevd: Great idea. requires a grep, but still easier than the other method18:33
→ thrope_ joined18:33
← shine-neko left18:33
← koltroll left18:34
← jason237 left18:34
steven i totally just rebased successfully for the first time ever!18:34
<318:34
soreau CareBear\: I don't see how git blame works18:34
thiago soreau: the syntax or how it works internally?18:34
soreau thiago: The syntax of course18:35
thiago git blame filename18:35
→ jason237 joined18:35
← jason237 left18:35
← cruzton left18:35
steven git help blame :)18:36
→ artefon joined18:36
soreau bah18:37
← phrearch left18:37
soreau git blame is not helping me18:37
Even with -p, it doesnt show full patches18:37
thiago it's not supposed to18:37
soreau I need to know, when a certain function changed18:37
→ Bass10 joined18:38
thiago it shows when each line was changed18:38
CareBear\ and the commit that changed it18:38
→ varioust joined18:38
soreau I don't understand how I'm supposed to use this information to determine when this particular function change happened18:39
→ letas joined18:39
soreau I need to know the commit id and message18:39
CareBear\ it's right there18:39
git blame gives the commit id18:39
← SirFunk left18:40
CareBear\ that changed each line of the file18:40
last18:40
soreau It makes no sense whatsoever18:40
→ thiago__ joined18:40
CareBear\ did you try to run it?18:40
← mitziyahu left18:40
Laggg → Laggg|dishes18:40
CareBear\ if yes - show the line that you are interested in?18:40
letas hello folks - Stupid question: how can I push (only) commits that are meaningful to systems like github or beanstalkapp? I am kind of tired of the ever growing merges showing up18:40
soreau Yes it just gives a bunch of meaningless information18:41
CareBear\: It doesn't show anything interesting at all18:41
It's a bunch of jargon18:41
CareBear\ pastebin the output then?18:41
soreau and it doesn't stick to the file I passes it either18:41
← dannas left18:42
→ dannas joined18:42
← artefon left18:42
CareBear\ (of course also include the exact command you used, as always)18:42
steven soreau: git log -p file18:42
← Evious left18:43
soreau CareBear\: Look it: http://pastebin.com/Wz6t2JqV I need to know when notify_notification_new changed and broke API18:43
steven btw guys, is there a fork of git that stores timestamps of all the tree/blob info per commit, and then `touch`s the files when checking out (or cloning) to match the timestamps the files had when they were committed?18:43
rcreasey If I have two branches on my repo, 'master' and 'develop', and there's a symbolic-ref for HEAD pointing to master; is there a way to create another symbolic-ref to point to the 'most recent revision on the develop branch'?18:43
steven soreau: try the command i just gave you, i think its what you're looking for18:44
→ Evious joined18:44
steven rcreasey: 'develop' already is that.18:44
PabloM rcreasey, why do you want that?18:44
soreau steven: Yea I do that, then hit / to search for notify_notification_new and it goes to the change, but instead of showing the commit message or id, it shows ...skipping...18:44
CareBear\ soreau : and which line of the file are you interested in?18:44
steven soreau: u broke it18:44
soreau ??????????????????/18:44
steven soreau: that works for me. it shows commit info as well as diff18:45
soreau wtf18:45
steven oooooh18:45
i know why.18:45
you're drunk. go sober up and come back and try again :)18:45
→ maestroj_ joined18:45
soreau Since when?18:45
steven typical case of the mondays.. still hung over from sunday, amirite or amirite!18:45
CareBear\ soreau : which line in the file are you interested in?18:45
soreau CareBear\: I'm trying to figure out how to answer that question18:46
← troessner left18:46
← maestrojed left18:46
CareBear\ man..18:46
open the file18:46
← thiago__ left18:46
soreau CareBear\: Because the line number in master may have changed by now18:46
→ tizzo joined18:46
→ bytemask joined18:46
← harv left18:46
CareBear\ you want the current line number18:46
← maestroj_ left18:46
soreau steven: How dare you. That is very unprofessional18:46
CareBear\ this is not a professional support channel18:46
this is IRC18:46
:)18:46
soreau I didn't say it was18:47
I just said he's being unprofessional18:47
CareBear\ I'm just answering how people dare.18:47
steven soreau: hehe18:47
→ harv joined18:47
steven its true. im not being very professinoal right now.18:47
he's got a point.18:47
CareBear\ I agree18:47
soreau /ignore steven18:47
oops18:47
bremner heh18:47
steven try harder soreau18:47
← Zephyrus left18:47
← Nickeeh left18:47
steven you gotta press down real hard on the slash key for it to work18:47
→ Nickeeh joined18:48
soreau CareBear\: It would be libnotify/notification.h:11618:48
the closest I can guess18:48
CareBear\ man18:48
you should not guess this18:48
you say you want to find what caused a change18:48
then you really must know *what* that change is18:48
soreau CareBear\: But if the file has changed, how can you track a previous change by line number?18:48
CareBear\ to be effective18:48
I said the *current* line number18:49
soreau Well they omitted this line18:49
CareBear\ omitted?18:49
soreau so it really has no line number18:49
→ TESTBROTEST joined18:49
CareBear\ you mean deleted?18:49
→ pyrony joined18:49
soreau Yes, they omitted it ie. removed it from the file18:49
TESTBROTEST → theoros18:49
→ maestrojed joined18:49
soreau No, omitted as in nuked the line in question18:49
CareBear\ omit means something else - but ok!18:49
I see18:49
← thrope_ left18:49
soreau I'm trying to figure out when this happened18:49
CareBear\ understand18:49
is there some other part of the same commit that is still around?18:50
soreau So it's kinda difficult to give a line number to something that no longer exists :P18:50
CareBear\ of course18:50
you said it changed18:50
soreau CareBear\: I can't find the commit18:50
so I do not know18:50
CareBear\ you did not say that it was removed18:50
BinGOs git log -Sfunctionname #perhaps18:50
soreau Well it's an API change18:50
CareBear\ change and remove are fundamentally different operations, and I was giving advice for the case of a change18:50
yes18:50
→ marcamilly joined18:50
CareBear\ one way would be bisect18:51
← maestrojed left18:51
CareBear\ and let it grep18:51
soreau how cumbersome18:51
surely there is a better way18:52
CareBear\ might not be very cumbersome18:52
soreau BinGOs: http://pastebin.com/H0beLF9i18:52
CareBear\ soreau : what are the known good and known bad commits?18:53
known bad could just be master18:53
soreau CareBear\: That is the very thing I'm trying to find out18:53
CareBear\ soreau : any known good is fine, it can be old18:53
soreau Well yea, i could do bad master, good 0.5.x18:53
CareBear\ great18:53
soreau but that's a lot of commits to trudge through18:53
I could find this out manually if I wanted already18:53
I just figured there is a much easier way18:53
CareBear\ no, this will be automated18:53
soreau Well I guess if you're not building the code it isn't so bad18:54
steven git is just brilliant. i really like it.18:54
<318:54
→ mastroDani joined18:54
steven i misunderstood how git revert worked earlier and it didnt do what i wanted.. so instead i did `git checkout <sha> -- .` and committed :D18:55
soreau CareBear\: So you're suggesting 'git checkout HEAD~1 && if grep some_string some/file; then echo bingo. && exit 0; fi' or something?18:55
steven (which is what i thought git revert would do)18:55
ooooh no. git revert removes that change. i get it now!18:55
CareBear\ soreau : git bisect start master 0.5.x --; git bisect run grep -vq searchstring libnotify/notification.h18:55
steven i was thinking of it as "git revert-to" which is just silly of me18:55
soreau CareBear\: Can you walk me through this? Never used git bisect with grep before18:56
CareBear\ soreau : git runs grep for each commit and exit status determines if commit is good or bad18:56
soreau CareBear\: 0e04cf94cbd823118e937ec9fab285a76c2e6e29 is good and master is bad18:57
So how can I do it?18:57
← Gulug left18:57
CareBear\ soreau : git bisect start master 0e04cf94c --; git bisect run grep -vq searchstring libnotify/notification.h18:57
soreau tries18:57
CareBear\ replace searchstring with the unique function name18:57
it must not match anything but the deleted line18:57
→ thiago__ joined18:58
BinGOs soreau: I said 'git log -Sfunctionname' not 'git log -S functionname'18:58
CareBear\ oh, sorry, remove the -v18:58
and you can remove -q as well if you like18:59
so the second command becomes git run grep searchstring ...18:59
soreau BinGOs: ok18:59
← _psychic_ left18:59
CareBear\ if you already started the bisect I guess you'll have to restart it18:59
soreau CareBear\: Didn't work., Pointed to a version bump19:00
← Sonder left19:00
CareBear\ the first, or without -v ?19:00
→ Sonderblade joined19:00
→ _psychic_ joined19:00
soreau CareBear\: I ran this: git bisect start master 0e04cf94c --; git bisect run grep -vq "GtkWidget *attach);" libnotify/notification.h19:00
where "GtkWidget *attach);" is the string that was removed19:00
CareBear\ drop -v sorry about that19:00
soreau the line, even19:00
CareBear\ I confused what good and bad exit codes should be19:01
soreau BinGOs: No output19:01
CareBear\: Ok, so what's correct?19:01
→ lstdgtfp joined19:01
CareBear\ just -q or no option at all19:02
otherwise it looked good19:02
well..19:02
you may need to escape the *19:02
it's a regex after al19:02
→ erichammond joined19:02
soreau It's in quotes..19:02
CareBear\ * has special meaning in regexes19:03
not escape from shell19:03
→ thiago_home joined19:03
CareBear\ escape in regex19:03
← vitor-br left19:03
soreau Well crap, now I broke it19:03
CareBear\ git bisect start master 0e04cf94c --; git bisect run grep 'GtkWidget.*\*attach);'19:03
soreau git log -p <commit id> just returns19:03
even though I git bisect reset19:04
← urkud left19:04
→ boombatower joined19:04
soreau It just hangs now19:04
Bisecting: 32 revisions left to test after this (roughly 5 steps) \n [4026faf46b093d2f4207732f4019a595e08e8ada] Add tests for transient and resident \n running grep GtkWidget.*\*attach);19:05
CareBear\ CPU busy?19:05
soreau and it's stuck19:05
nope19:05
CareBear\ may have to double escape, in case git runs the command via sh19:05
which I think it does19:05
git bisect start master 0e04cf94c --; git bisect run grep '"GtkWidget.*\*attach);"'19:05
BinGOs I'm doubting this thing ever existed >:)19:05
soreau That is not the problem19:05
The repo or git is broken now19:06
→ thrope joined19:06
← mastroDani left19:06
soreau It wont run19:06
→ vitor-br joined19:06
CareBear\ is this a repo available online?19:06
soreau yes19:06
CareBear\ url?19:06
soreau http://git.gnome.org/browse/libnotify19:06
CareBear\ let's see19:07
soreau sighs19:07
soreau This should be *wai* easier19:07
CareBear\ doh, sorry, the filename needs to go on the end of course :)19:07
soreau Well nofuckingwonder :P19:08
← Nickeeh left19:08
Laggg|dishes → Laggg19:08
CareBear\ the search string matches two lines19:08
→ Nickeeh joined19:08
CareBear\ that does not work19:08
it must be unique19:08
what function is it?19:09
← jsick3 left19:09
→ alberto joined19:10
→ xerora joined19:10
→ radarek joined19:10
← alberto left19:10
← alberto56 left19:10
→ alberto joined19:10
CareBear\ soreau : ?19:11
soreau CareBear\: Since you're looking into this (thanks btw) the specifics are Good http://git.gnome.org/browse/libnotify/tree/libnotify/notification.h?id=c56f4855b1038a8b74b1d44e573eb380718a5892#n79 Bad http://git.gnome.org/browse/libnotify/tree/libnotify/notification.h#n11419:11
CareBear\: It's notify_notification_new19:11
→ CarlFK joined19:12
→ jfelchner joined19:12
CareBear\ ok19:14
let's see then19:14
the regex becomes a little tricky since all parameters are on separate lines (silly)19:14
→ amh345 joined19:15
amh345 i've just mangled git. eek19:15
CareBear\ soreau : git bisect start master c56f485 --; git bisect run sh -c 'sed -e "/notify_notification_new /,/);/!d" libnotify/notification.h |grep attach'19:16
amh345 git push origin master ==> " ! [rejected] master -> master (non-fast-forward)"19:16
soreau CareBear\: geez :P19:16
CareBear\ 27e05d0f9562a26163493d6cc1d5924b9a4ebf68 is the first bad commit19:16
soreau looks19:17
→ irc2samus joined19:17
→ jsick3 joined19:17
← thrope left19:17
→ mattalexx joined19:17
CareBear\ git log|grep notification had actually worked too, since you have good commit messages19:17
→ liquidproof joined19:17
irc2samus hello guys, can you tell me how to checkout a specific revision of a file without afecting the HEAD?19:17
I'd like to run tests with previous versions of a specific file19:18
mattalexx I'm an absolute beginner at git, but I have some experience with Subversion. Anyway, I'm following this tutorial: http://squaremasher.blogspot.com/2009/12/managing-dotfiles-with-git.html19:18
nDuff irc2samus, ...uhh, HEAD _is_ "whatever you have checked out"19:18
CareBear\ irc2samus : I think HEAD follows what you check out19:18
mattalexx At no point does he add the files to the commit, right? Isn't he missing something?19:18
← ptl left19:18
irc2samus I mean, making the file modified but not affecting the last commit19:18
amh345 git merge master says it's up to date. yes im getting this " ! [rejected] master -> master (non-fast-forward)" . can anyone help with this? im completely lost.19:18
mattalexx I'm running git commit -a -m 'Bla' and getting "nothing added to commit but untracked files present (use "git add" to track)"19:18
soreau CareBear\: Ah, this is it. One of the bisects I had came up with this commit but it was so big, I didn't realize the change was there19:19
CareBear\ irc2samus : git checkout never modifies commits19:19
nDuff irc2samus, ...oh, just one file, got it.19:19
fr0sty_ amh345: git fetch --all then merge again19:19
→ cocoadaemon joined19:19
soreau CareBear\: ie. this works too: git bisect start master 0e04cf94c --; git bisect run grep 'GtkWidget' libnotify/notification.h19:19
CareBear\ soreau : and the log message did not convince you? wow..19:19
soreau CareBear\: Thanks for your help19:19
→ ilteris___ joined19:19
soreau CareBear\: Nope, I didn't read it :)19:19
→ jhoepken joined19:19
nDuff irc2samus, ...so, if you have the hash of the file, you can retrieve it with git show...19:20
amh345 fr0sty_: git fetch --all say's fetching origin19:20
CareBear\ soreau : cool! I didn't look at the file contents, so I just went with a very very reliable test19:20
amh345 fr0sty_: then git merge master says already up-to-date19:20
CareBear\ although a bit more complicated19:20
irc2samus nDuff: hmm and then git apply right? sounds good I'll try19:20
ilteris___ hey guys, does anyone know how can I add a timestamp to my git in my terminal so it shows me how much it has been since the last commit?19:20
soreau CareBear\: I still don't completely understand what git run is doing, but I guess you could put any command there and it checks return value?19:20
CareBear\ soreau : correct. there are good explanations in git bisect --help19:21
fr0sty_ amh345: where are you trying to push?19:21
nDuff irc2samus, ...no need to apply if you're retrieving the blob with the file...19:21
amh345 fr0sty_: im trying to push my latest changes to a remote server. but somewhere things got out of sync19:21
soreau Good ghad, not the man pages19:21
fr0sty_ amh345: what command are you using to push ?19:22
przemoc86 → przemoc19:22
CareBear\ soreau : they are excellent19:22
amh345 fr0sty_: totally my fault too. i accidentally recreated .git and pushed19:22
BinGOs soreau: http://git.gnome.org/browse/libnotify/commit/?id=cc13b1892f79b7d33ab8d587a14db5d848e52fa419:22
cocoadaemon Ola. I am starting with Git, which looks awesome. I am seeing on a tutorial --bare should be used for projects using ssh to push / pull. But I don't understand how you could have that to work : bare only takes the .git dir, right ?19:22
soreau CareBear\: not if you have reading comprehension problems :P19:22
amh345 fr0sty_: git push origin master19:22
nDuff irc2samus, ...you should be able to use git ls-tree to get that blob's hash.19:22
fr0sty_ amh345: oh, that is important information...19:22
soreau CareBear\: Or, if you're otherwise allergic to man pages19:22
fr0sty_ (the recreating .git bit)19:22
irc2samus nDuff: but git show outputs the diff from the parent commit to the specified one right?19:22
fr0sty_ try 'git push -f origin master'19:22
amh345 fr0sty_: if possible. i could just start from sratch with my remote and local repos19:22
fr0sty_: everything that was up there was pretty much trash anywats19:23
← cocoadaemon left19:23
nDuff irc2samus, git show will show a _lot_ of different things, depending on what you ask it. If you give it the hash of the file in the state that you want, it gives you THAT FILE, not a diff of any kind.19:23
← slonopotamus left19:23
fr0sty_ amh345: then just recreate the remote repository if you can...19:23
soreau BinGOs: Nope, that's not it19:23
irc2samus nDuff: I don't know how to pass "hashes of files" :( I have the hash of the commit19:24
soreau BinGOs: It was this http://git.gnome.org/browse/libnotify/commit/?id=27e05d0f9562a26163493d6cc1d5924b9a4ebf6819:24
nDuff irc2samus, use git ls-tree to get the file's hash, as I previously suggested.19:24
→ cocoadaemon joined19:24
soreau BinGOs: Thanks for your help though19:24
irc2samus ahh missed that, sorry :)19:24
amh345 fr0sty_: just because im terrified im going to mangle something further. on remote i would rm -rf my git dir with my file.git ?19:24
→ logickal joined19:25
cocoadaemon mmm sorry I CTRL-Ced19:25
BinGOs soreau: first commit I looked at from 'git log -Snotify_notification_new'19:25
oh well.19:25
← cocoadaemon left19:25
→ alban_ joined19:26
soreau BinGOs: Ah, that command never worked for me. It just returns with no output19:26
irc2samus perfect! thanks nDuff :D19:26
→ ptl joined19:27
→ LouisJB joined19:27
← alban_ left19:27
→ cocoadaemon joined19:27
← robinsmidsrod left19:27
← soreau left19:27
mattalexx I thought "git commit -a -m 'Bla'" will automatically add new things to the repo. When I run it, however, I get "nothing added to commit but untracked files present (use "git add" to track)". What am I doing wrong?19:27
← mediaslave left19:28
nDuff mattalexx, git commit -a does not add new files.19:28
→ robinsmidsrod joined19:28
nDuff mattalexx, (and it's bad practice anyhow -- leads to committing things you don't want by mistake)19:28
→ mediaslave joined19:28
← logickal left19:29
cocoadaemon So, --bare : Good Practice As Project Sharepoint ?19:29
fr0sty_ amh345: yes, wherever your .git (or name.git) directory is, remove it. and 'git init --bare [--shared=group] name.git'19:29
amh345 fr0sty_: here goes nothing :)19:29
thanks for the help!19:29
← mando left19:29
nDuff mattalexx, ...do the right thing and use git add to add things, and git diff --cached to see what you have added, and _then_ commit. Or don't do the right thing, and use git add only for new files where it's mandatory.19:29
← _psychic_ left19:31
cocoadaemon So no one to help me figure out why using a project cloned with --bare is better than working on the original when you're more than one ?19:33
jast cocoadaemon: bare repositories are used at hosting locations (e.g. central servers); nobody can work in the working tree so no confusion can possibly happen when someone pushes new stuff to it19:34
for your own local work, of course it doesn't make sense to use a bare repository19:35
→ maestrojed joined19:35
cocoadaemon so basically, it's up to the project admin to push from the bared clone to the "master" ?19:35
→ ymasory joined19:35
jast uh, what? :)19:36
cocoadaemon I mean, when a user A pushed a change, he pushes it to the bare clone X, ok ?19:37
← owst left19:37
mattalexx nDuff, I'm following a tutorial for adding my dotfiles to a repo. This guy isn't adding the files at all. Is he doing it wrong in your opinion? http://squaremasher.blogspot.com/2009/12/managing-dotfiles-with-git.html19:37
jast right19:37
← beatak left19:38
cocoadaemon if that change is ok, someone who's responsible for the project can review user A's changes, accept it and push it to the non-bare project ?19:38
irc2samus mattalexx: github?19:38
← j0ran left19:38
mattalexx irc2samus, What about it?19:38
← maestrojed left19:38
jast cocoadaemon: what non-bare project?19:39
irc2samus mattalexx: if you're managing your own repos maybe a tool like gitosis can help you19:39
but putting them on a free account on github is much more convenient19:39
jast please consider recommending gitolite rather than gitosis; the latter is no longer maintained and supported19:39
mattalexx irc2samus, I believe I'll be hosting them on github, yes.19:39
→ codyrobbins joined19:40
irc2samus jast: ok, didin't tried that one only gitsos for the moment but I'll check it out19:40
→ beatak joined19:40
irc2samus *only gitosis19:40
cocoadaemon jast: the non-bared project would be the "full" version of the project, the one used as a source for the --bare clone19:40
codyrobbins I did a git pull and got all sorts of weird merge conflicts. How do I just completely undo the attempted pull and get back to a blank slate?19:40
Doing a git checkout still leaves the files with the conflict markers in them19:41
jast cocoadaemon: so why do you need both the bare repository and that one?19:41
cocoadaemon: git reset --merge19:41
sorry, wrong person19:41
codyrobbins: git reset --merge19:41
have more easily discernable nicknames next time :(19:41
cocoadaemon this place is damned crowded :)19:42
ok so, I'm not understanding.19:42
codyrobbins jast: Thanks! :)19:42
I kept looking at that option, but the documentation for it is completely opaque to me.19:42
jast codyrobbins: yeah, I'll admit that it takes a while until these manpages make sense19:43
→ ShooterMG joined19:43
cocoadaemon I have my project in /P/ with /P/.git ( and all the file hierarchy under ) which I clone in /B. Users will push to /B right ? And if I accept their changes, I'll push back to /P no ?19:44
( the clone being B for --bare )19:44
jast P is your personal repository?19:45
nDuff mattalexx, non-source usage falls into the category of special cases.19:45
CareBear\ then yes19:45
cocoadaemon nope, let's say P is a web app directory jast19:45
→ j0ran joined19:46
jast oh, so this is a deployment thing19:46
CareBear\ cocoadaemon : read this: http://joemaller.com/990/a-web-focused-git-workflow/19:46
jast git doesn't really do deployment19:46
CareBear\ third time tonight..19:46
cocoadaemon yes, this might clarify19:46
jast anyhow, you'd pull from B to P19:46
CareBear\ jast : it can19:46
with a little help :919:46
cocoadaemon ah, thanks CareBear19:46
jast it *can*, but it's not a natural at it19:46
Arrowmaster cocoadaemon: no read this instead http://toroid.org/ams/git-website-howto19:46
CareBear\ true19:46
Arrowmaster : that's good for single developer, not so good for multiple imo19:47
→ ereslibre_laptop joined19:47
Arrowmaster CareBear\: dont link people to sites that recommend hacky stuff like running git pull in a hook19:47
← ereslibre left19:47
nDuff mattalexx, ...but given that git commit -a does not auto-add files, it's clearly "wrong" in the sense of "doesn't work"19:47
CareBear\ Arrowmaster : not hacky - it's a different repository being pulled into19:47
nDuff mattalexx, indeed, the git commit man page is explicit about this; describing --all: "Tell the command to automatically stage files that have been modified and deleted, but new files you have not told git about are not affected."19:48
Arrowmaster CareBear\: that is hacky and often problematic19:48
mattalexx nDuff, Interesting. I'll close that taqb then19:49
*tab19:49
CareBear\ Arrowmaster : did you actually read through the page? the setup is pretty foolproof.19:49
← DiGitalX left19:49
Arrowmaster CareBear\: and he can use the one i linked just fine in a multi user environment if he is the one controlling what goes into the production repo19:49
CareBear\ then there is in fact only a single user19:49
Arrowmaster CareBear\: yes i have, and i know of problems caused by setups exactly like that19:49
cocoadaemon what's the purpose of the hook ? Auto push ?19:49
CareBear\ cocoadaemon : in joemaller's variant there are two hooks IIRC19:50
cocoadaemon : but the purpose of both is basically to move changes around to the right places19:50
← saschagehlich left19:50
← Raging_Hog left19:51
cocoadaemon both articles contain hooks19:52
are these so important? Or just useful?19:52
→ Nemesis13 joined19:52
← panike left19:52
← OOPMan left19:53
CareBear\ important19:53
if you want automated things19:53
Arrowmaster in the one i linked, the hook causes a directory to be updated with the repository contents when you push to the repository19:53
kumbayo is it possible in pre-commit hook to detect a commit --amend (i would like to detect this because currently the hook uses "git rev-parse --verify HEAD" to detect a unborn branch, but that is not true anymore on amend)19:53
← ereslibre_laptop left19:53
Nemesis13 i'm using git for a long time now but i forget the commands again and again... how do i merge two commits (which were not commited in a line)? was it git rebase -i or something?19:53
→ PointMan joined19:53
cocoadaemon jast: basically, the graph in the former article is what I meant : many users who push to a central node which is used as a gate to the prod repo19:53
CareBear\ cocoadaemon : the graph doesn't fit that case unless all users have equal publishing rights19:54
cocoadaemon Arrowmaster: so basically, it's the "automatize the push when I commit" ?19:54
Arrowmaster cocoadaemon: no19:54
← guardian left19:55
Arrowmaster cocoadaemon: you understand the difference between bare and non-bare repositories right?19:55
→ guardian joined19:55
PointMan um. on my origin I only want a master. but I accidently did git push in a branch, so there is a feature branch in the origin..When I do git checkout master; git push then I get an error about the feature branch not being able to ff..but I just want git to forget about the remote branch. I tried deleting the remote branch but it still wont let me push from master to origin master19:55
← quazi_farhan left19:55
CareBear\ PointMan : git push remote :branchtodelete19:55
Arrowmaster or git push remote --delete branchtodelete19:56
← codyrobbins left19:56
← radarek left19:56
cocoadaemon probably not, it might be the problem. As i said, I understand it only contains the "abstract" and no files. What I wonder is : when you push to a bare repository, in which form does it recorded what's been pushed? Hash only ? file too?19:56
← Slush- left19:57
← letas left19:57
Arrowmaster cocoadaemon: a bare repo and the contents of the .git directory in a normal non-bare repo are the same, the bare repo just lacks the working copy19:57
PointMan CareBear, Arrowmaster thanks :)19:58
→ sylr joined19:58
→ t_b joined19:58
Arrowmaster cocoadaemon: since you arent supposed to push into a non-bare repo, the page i linked uses a hook in a bare repo to update a directory with what would normally be the working tree19:58
cocoadaemon Arrowmaster: got that. bare <-> copy ".git" and F... the rest right ?19:59
Arrowmaster well its a little bit more than just that but yeah19:59
cocoadaemon so yeah, that's the tough part. What happens when you push to a bare repo.20:00
← t_b left20:00
cocoadaemon I guess my "moderation process" is a remain of svn and has nothing to do with git20:01
→ qwester joined20:01
mattalexx How to I clear out a git init altogether?20:01
jast remove .git20:02
there, all traces of anything git has ever done are gone20:02
cocoadaemon what I don't understand in this case is "how is it safer to have people work on the bare repo if you insta push to the master ?"20:02
mattalexx jast, thx, that's great20:02
jast along with any history you might have20:02
← sylr left20:03
CareBear\ cocoadaemon : don't push to a repo with a work tree20:03
cocoadaemon : git may not even let you20:03
steven i just realized that the algorithm to generate a tree from the index isnt as trivial as i thought it would be20:04
considering the index stores files as a flat array with relative pathnames instead of as a tree20:04
cocoadaemon CareBear\: So it's necessary to have the "pushable" part of the project as a --bare repo ?20:05
CareBear\ right20:05
cocoadaemon So this bare concept is something one need to understand right indeed.20:05
← superdmp left20:06
→ eletuchy joined20:06
CareBear\ a bare repo is the .git dir of a repo with work tree20:06
and a config change20:06
steven config change?20:06
CareBear\\: what do you mean by that?20:06
CareBear\ core.bare=true20:06
steven oh. thats it?20:06
CareBear\ I believe so20:07
steven cool. sounds useful as remote repos20:07
(since theres no working tree)20:07
CareBear\ right20:07
→ OOPMan joined20:07
→ owst joined20:08
cocoadaemon and no files ? So you cannot say " okay, let's have this Hub act as a filter which I'll use to verify commits before pushing to Prime"20:08
steven is that how a repo knows if it is a "remote" repo or not? or can any repo (even a non-bare one) be a "remote" in a push/pull operation?20:08
→ nibblebot joined20:08
→ Dr4g joined20:08
CareBear\ cocoadaemon : correct, you can not20:09
cocoadaemon : if you need that gateway you will have to place it before hub20:09
← Dr4g left20:09
→ Dr4g joined20:09
CareBear\ steven : any can be remote20:09
steven ah. cool.20:09
CareBear\ steven : remote is a property of a repository reference in a repository20:09
Dr4g Trying to use github on FreeBSD. I've installed git via ports, but this page shows me errornous stuff20:09
http://help.github.com/linux-set-up-git/20:09
CareBear\ a repo is a repo is a repo :)20:09
cocoadaemon Somehow, if you want to "filter" commits you need 4 repos : Dev server ( 1. bare + 2. working tree ) + Prod server ( 3. bare + 4. working tree )20:09
so peons commit to 1, which passes files to 2. Manager then tests, validates, pushes to 3 which passes to 4 ?20:10
CareBear\ cocoadaemon : you can skip 4. if only one person will ever push to 3.20:10
cocoadaemon but you need a working tree at the end??!20:11
CareBear\ cocoadaemon : also, 2. is more likely to be on manager workstation than on dev server20:11
steven i kinda wonder what happens if i try to co origin/master and make changes.20:11
→ nurey joined20:11
CareBear\ cocoadaemon : no, you can do what the ams-page describes and only do a GIT_WORK_TREE=public_html git checkout -f in the post-receive hook in 3.20:11
← PointMan left20:12
CareBear\ steven : try it!20:12
← yairgo left20:12
CareBear\ :)20:12
bremner Dr4g: you'll either need to be more specific, or ask on #github20:12
→ letas joined20:12
→ linusoleander joined20:12
cocoadaemon Darn. You lost me.20:12
#8 Git is easy. My A//20:12
;)20:13
CareBear\ cocoadaemon : git is easy to use, but the workflows are much more powerful therefor more complicated20:13
+e20:13
cocoadaemon yeah, not blaming, git is great. As a user, not a manager :)20:13
steven ooooh i get it, git status is slow on large repos because it has to build an entire tree object out of the working tree, right?20:14
→ Slush- joined20:14
cocoadaemon ( kidding. Git is also a steaming pile of shit for users too )20:14
CareBear\ also manager; you are *able* to set up things that would be nowhere nearly as easy with other tools :)20:14
→ Zephyrus joined20:14
← qwester left20:14
cocoadaemon ( or isn't it what the doc says ? :=D )20:14
→ mrq1 joined20:14
Dr4g thanks bremner20:15
→ qwester joined20:15
CareBear\ cocoadaemon : peons have local repos where they commit. they push into a bare repo 1. on a development server. manager pulls from 1. into her local repo with work tree (could be considered your 2. above)20:15
cocoadaemon So you say one needs only these : {public access}( 1. bare + 2. working tree ) + {resticted access} ( 3. bare[with stuff I don't get yet] )20:16
CareBear\ cocoadaemon : when manager has approved and/or massaged commits in 2. they get pushed to a bare repo on the production server 3. where three is a post-receive hook with read old new ref; GIT_WORK_TREE=/var/www/server/htdocs exec git checkout -f "${ref}"20:16
→ artefon joined20:17
cocoadaemon mmm... I need to slowly digest that CareBear\ :)20:17
PabloM cocoadaemon, or you could just use Gerrit, which has access control built in20:18
← excess^ left20:18
← thiago__ left20:18
→ excess^ joined20:18
cocoadaemon PabloM: Gitosis too? Yeah I know, just trying to figure out how the manual gearbox works, you know ;)20:19
→ mmw joined20:19
← saschpe_ left20:19
cocoadaemon CareBear\: so when a peon pushes to bare repo 1. what is kept from their commission(?) : file ? or hash ?20:20
CareBear\ cocoadaemon : come again?20:20
cocoadaemon : they create a commit in their local repo.20:20
cocoadaemon : that commit then travels into 1. as-is20:21
cocoadaemon let's take a very simple example. peon creates a text file, adds it, commit it to 1.20:21
CareBear\ no20:21
→ rseifert joined20:21
CareBear\ they push into 1.20:21
cocoadaemon right, sorry20:21
CareBear\ they commit locally20:21
:920:21
:)20:21
cocoadaemon the commit locally, push to 1.20:21
→ kpreid joined20:21
CareBear\ yep!20:21
→ koltroll joined20:21
cbreak that's the only kind of commit that exists...20:21
cocoadaemon ^the^they20:21
cbreak local one :)20:21
CareBear\ cbreak : aye20:21
cocoadaemon exactly.20:22
CareBear\ cocoadaemon : the manager would then do git pull in her work tree20:22
cocoadaemon Let's take a very simple example. peon creates a text file, adds it, commit it, pushes the commit to 1.20:22
how is the textfile kept on 1. ?20:22
cbreak same as on the source20:22
cocoadaemon is it a file ? only a hash ?20:23
cbreak it's stored in git's database20:23
CareBear\ cbreak++20:23
mrq1 hi :) a question about "git add -p" patch mode. if the changed lines are something like more than 10 lines apart they are recognized as a different hunk. is there a config option for this number? i need every changed line as a different hunk.20:23
cbreak that means there is either a blob in .git/objects20:23
CareBear\ cocoadaemon : git stores the commit (ie. all added lines in the file) in it's database20:23
cbreak or there's a blob in a pack file20:23
git stores content of files20:23
NOT diffs20:23
NOT added lines, but the complete new files20:23
CareBear\ true20:23
cbreak http://eagain.net/articles/git-for-computer-scientists/20:24
CareBear\ the blobs20:24
cocoadaemon cbreak: that's what I understood so far. Which I assumed makes it better for binary storage ?20:24
cbreak when you push, you push over the blob, the tree that points to the blob, the commit that points to the tree20:24
→ eijk_ joined20:24
cbreak and the commits the commit points to20:24
cocoadaemon: binary storage?20:24
CareBear\ cbreak EMUCHDETAIL20:24
possibly not useful :920:25
:)20:25
cocoadaemon hey, actually it made sense.20:25
CareBear\ okey!20:25
cbreak just look at the link above20:25
it's easy to understand20:25
... if you're a computer scientist :D20:25
cocoadaemon binary : SVN was bad at storing non-text files20:25
CareBear\ though not completely neccessary to understand20:25
cbreak git's bad too20:25
cocoadaemon since it uses only diff20:25
bremner zlib's bad20:26
cbreak if you change one bit in a file, git stores the complete file again20:26
→ radarek joined20:26
cbreak that's kind of bad for big binary files20:26
repack might be able to deal with it20:26
... but it'll be slow and use a lot of memory20:26
you should not use git to version control big and often changing binary files20:26
cocoadaemon I see.20:26
each system comes with its own downsides.20:27
bremner s/use git to//20:27
CareBear\ cocoadaemon : back to the example?20:27
cocoadaemon yeah.20:27
cbreak bremner: I heard a talk from the pixar guys from the main office, they use perforce for their version control20:27
they store the scenes, the models, assets, and so on20:27
cocoadaemon so the answer to my question is : 1. contains a blob of the text file that the user sent.20:28
← eijk left20:28
cbreak not sure how good it is though20:28
but I bet it's better than not having backups :)20:28
bremner cbreak: yeah, I imaging pixar just buys more disk20:28
cocoadaemon and also, ultimately the .git directory might be 10 times larger than the working tree20:28
cbreak depends20:29
← blindingdawn left20:29
cocoadaemon so technically, a bare repo isn't as "lightweight" as I thought it was20:29
→ PointMan joined20:29
CareBear\ not at all20:29
cbreak it saves a working copy20:29
nothing more20:29
CareBear\ the only difference is that it has no work tree20:29
← Titosemi left20:30
cocoadaemon so if each file was changed 10 times, it saves approx. 10% (files size being equal)20:30
cbreak bremner: as far as I remember, in an other stage of the process, they use folders with versions of the models20:30
→ sagsousuke joined20:30
cbreak bremner: and a symlink pointing to the newest approved version20:30
← alberto left20:30
cbreak all models and so on have to be approved from some leader guy before it can be used downstream20:30
cocoadaemon cbreak: interesting20:31
cbreak so it's kind of a manual full copy scheme20:31
→ shelhamer joined20:31
cbreak I doubt storage is a limitation for them though :D20:31
→ linusoleander_ joined20:31
→ synisma joined20:31
← linusoleander left20:32
linusoleander_ → linusoleander20:32
cocoadaemon ok so the whole bare repo / working tree is starting to make sense for me thanks to you folks20:32
→ whaley joined20:32
← bindaas left20:32
cbreak bare repos are there so you can push to them20:32
cocoadaemon what I need to investigate is the final step explained by CareBear\20:32
cbreak you can not reasonably push to repositories with working directories, for various reasons20:33
they are for exchange20:33
cocoadaemon I don't understand why yet, but I will20:33
cbreak repositories with working dir are for working :)20:33
→ _psychic_ joined20:33
cocoadaemon let's say I keep that as a motto :)20:33
← sagsousuke left20:34
← dannas left20:35
→ avdg1 joined20:35
← jhoepken left20:36
→ jhoepken_ joined20:36
cocoadaemon CareBear\: GIT_WORK_TREE somehow means "Wherever you (bare repo) may be stored on the disk, use this location as a working dir"20:36
← T-Pup left20:36
← avdg left20:36
← Tv left20:36
← PointMan left20:38
→ sagsousuke joined20:38
CareBear\ cocoadaemon : correct!20:38
→ _nil joined20:39
← matjas left20:39
_nil hi, i wnat to write a plugin for git20:39
is this possible?20:40
thiago_home what would this plugin do?20:40
_nil thiago_home: it would upload code reviews to google rietveld servers20:40
← sebersole left20:40
thiago_home so an extra command?20:40
_nil well several20:40
thiago_home just do them20:41
→ mando joined20:41
thiago_home name them git-commad1, git-command2, etc.20:41
you can run as "git command1"20:41
→ sebersole joined20:41
_nil thiago_home: i'm not sure how it would integrate20:41
i've worked with mercurial plugins20:42
→ anjen joined20:42
_nil basically are their git plugins20:42
there*20:42
← joel_falcou left20:42
thiago_home it doesn't need integration20:42
they are just extra commands20:42
_nil ok i don't get the difference :)20:42
thiago_home has a script that pastes the named commits to pastebin.com20:42
_nil can you explain a bit?20:42
thiago_home not really20:43
you have a script that does stuff20:43
period. That's it.20:43
_nil that's not a part of git though20:43
that's an external wrapper20:43
← sagsousuke left20:43
IslandUsurper git commands are only related in that they know where to look for files (as I understand it)20:43
→ logickal joined20:43
thiago_home sure20:44
why do you need something to be part of Git?20:44
← amh345 left20:44
← Laggg left20:45
→ avdg joined20:45
→ panike joined20:46
_nil thiago_home: you're just asking me questions back20:46
thiago_home yes20:47
_nil i just want to know if git has a plugin infrastructure20:47
that's all20:47
← excess^ left20:47
thiago_home no, there is no plugin interface20:47
_nil i can't seem to find a reliable stource20:47
ok20:47
cbreak git is a plugin infrastructure20:47
thiago_home I'm asking you what kind of information you're expecting from the plugin20:47
cbreak it consists of dozens of tiny plumbing and porcelain tools20:47
systemaddict is also a plugin infrastructure20:47
→ excess^ joined20:47
← avdg1 left20:47
← excess^ left20:47
← linusoleander left20:47
_nil thiago_home: the plugin will enable me to interact with reitveld and then do git commands accordingly20:48
i.e.20:48
→ yairgo joined20:48
_nil git p <pending>20:48
← skoop left20:48
_nil will show me my open / pending changes20:48
git submit <change> will commit a change20:48
cbreak how will a plugin infrastructure enable that?20:48
← Guest41220 left20:48
← goshawk left20:48
_nil see mercurial codereview plugin20:48
i'm trying to replicate that on git20:48
though it seems not possible20:49
cbreak just put a new git-submit program into /usr/local/libexec/git-core/...20:49
thiago_home _nil: just name your program "git-p" and put it somewhere in $PATH20:49
that's enough20:49
cbreak the path I wrote above is used by git itself20:50
so it might or might not be a good idea to use it20:50
thiago_home has $HOME/bin/git-pastebin20:51
thiago_home I can run it with "git pastebin"20:51
_nil ok well i'll figure sth out20:51
thanks for your suggestions20:51
← abhaythehero left20:51
← nona left20:51
← sftp_ left20:52
← eletuchy left20:54
← drubies|mbp left20:54
→ irc2samus_ joined20:55
← guardian left20:55
→ tewecske joined20:55
→ guardian joined20:55
← Zephyrus left20:56
→ eletuchy joined20:56
← jhoepken_ left20:56
→ khmarbaise joined20:57
← irc2samus left20:57
irc2samus_ → irc2samus20:57
→ bronson joined20:59
→ linusoleander joined21:00
← khmarbaise left21:01
← LongBeach left21:02
← koltroll left21:05
← IslandUsurper left21:05
← tizzo left21:05
← lostincake2 left21:07
→ sftp joined21:08
→ EricInBNE joined21:09
← HG` left21:10
← jophish left21:11
steven is it possible for two different objects to accidentall have the same sha1 hash?21:11
← erichammond left21:11
← varioust left21:11
mmattice highly improbable, but sure21:11
cbreak if you find a collision, maybe you get a prize from linus21:12
← tewecske left21:12
steven has anyone found one yet?21:12
← eijk_ left21:12
← Nickeeh left21:14
CareBear\ http://eprint.iacr.org/2008/469.pdf21:14
steven sweet21:16
→ hakunin joined21:16
← icwiener left21:17
hakunin Anybody knows git?21:17
Just kidding21:17
My colleague used to forget to set his creds in gitconfig21:17
now he has like 10 different users in git history all being himself21:17
wondering if there's any way to consolidate21:17
→ Mark_Morcom joined21:17
PerlJam hakunin: sure ... it just requires that you rewrite history21:17
SethRobertson git filter-branch if he really has to, but it changes history and forces everyone else to be annoyed21:17
Assuming he has pushed21:17
→ sraue_ joined21:18
→ Nickeeh joined21:18
hakunin yeah rewriting history isn't good, not sure if this type of problem justifies it21:18
← sraue left21:19
hakunin PerlJam: SethRobertson: so if i have remote tracking branches with a few people participating, what happens when they attempt to pull after such a change?21:19
← Nickeeh left21:20
CareBear\ they must pull -f21:20
→ Nickeeh joined21:20
CareBear\ and they would need to rebase manually21:20
if they have local changes21:20
PerlJam hakunin: All those things CareBear\ is saying and probably more.21:21
hakunin: for instance, did they have any tags?21:21
SethRobertson They should `git pull --rebase` actually21:21
hakunin not locally, all tags are shared21:21
we normally pull --rebase around here, so that wouldn't be too new, but i see...21:21
SethRobertson If they have local tags they are messed up. Global tags should be updated during the filter-branch (common error)21:21
CareBear\ hakunin : the tags are all bogus too then21:21
PerlJam hakunin: also, once they've pulled forcefully, they'll have bunches of dangling commits in their repo that they'll need to clean.21:22
← kenichi left21:22
← Nickeeh left21:22
CareBear\ big fun yes!21:22
hakunin great21:22
← mmw left21:22
PerlJam hakunin: but it might be worth it to go through the pain *once*21:23
CareBear\ whatperljam said21:23
← liquidproof left21:23
→ jds joined21:23
homa_rano is there some way I could run git-diff-{index,tree} against an empty repository? as in show all existed files as added?21:23
→ k0ral joined21:23
hakunin well, i thought it'd be something like this. i really only have 3 people on this, one of which only comes in like once-twice a year. so maybe if we're side by side in office i could do this with the colleague.21:24
← yairgo left21:24
CareBear\ then hit messy guy over head21:24
with suitable inflatable hammer or so :)21:24
← kumbayo left21:24
→ eijk_ joined21:24
hakunin already did a bit more. created shared dotfiles and made him use my setup with his configs21:25
→ vitalius joined21:25
CareBear\ or install a gatekeeper21:25
→ Nickeeh joined21:25
CareBear\ ie have actual review ;)21:25
← rseifert left21:25
hakunin he's a frontend dev, so never focused on setting up local dotfiles/configs much21:25
vitalius why would gitosis ask for password when the id_rsa.pub was used? I heard it may be not random enough, is that true?21:25
PerlJam you could have a hook that verifies user information21:25
vitalius: 1) use gitolite instead of gitosis21:26
k0ral hello, if I'd like to share a subset of a git repository, not all files, what is the simpler way to do it ?21:26
CareBear\ vitalius : if your private ssh key was generated on a debian system with the openssl that had no randomness then yes ssh might reject it21:26
PerlJam vitalius: 2) it would ask for a password when the pub doesn't match up with what it thinks is valid.21:26
hakunin that's a nice idea, though since it's just 2 of us — a slip like this needs only to be fixed once, he now learned to care for that stuff21:26
← gretch left21:27
hakunin PerlJam: ^^21:27
CareBear\ hakunin : except you said it went wrong 10 times21:27
I would expect such a user to be incapable of reliably following protocol21:27
ever21:27
hakunin CareBear\: more like: he never had hard-configed creds, so he has 10 different names bcause he took laptop into multiple networks where it populated name@host for him automatically, with different host21:27
CareBear\ aha21:28
vitalius hmm i tried generating keys on mac, same story. How would I check if hook verifies user information?21:28
CareBear\ hakunin : then hopefully it'll stick! :)21:28
k0ral anyone noticed my question at least ? :)21:28
hakunin we'll hope so : )21:28
← Dave^| left21:28
CareBear\ k0ral : sure21:29
hakunin CareBear\: PerlJam: SethRobertson: thanks, really appreciate it21:29
steven hehehe, you can commit in the future with --date :D21:29
→ erichammond joined21:29
steven i wonder if github gets confused when it sees future-dated commits21:30
← eijk_ left21:30
→ eijk_ joined21:30
avdg I think github just wouldn't care about it21:32
← eletuchy left21:33
steven obviously git doesnt care either :)21:34
→ eletuchy joined21:34
← hobodave_ left21:35
← dkasak left21:36
← ericbarnes left21:36
→ dkasak joined21:36
← dajhorn left21:37
← ExtraSpice left21:37
→ Tv joined21:37
← jstemmer left21:37
jds I'm in the middle of a merge, but kinda made a mess of a file21:38
can I reset that specific file and start again?21:38
→ foocraft joined21:38
thiago_home reset to conflicted-merge state?21:38
← rayners left21:38
→ rorist joined21:39
← TrineLocke left21:41
← vitor-br left21:41
→ sdboyer joined21:42
← artefon left21:44
→ thiago__ joined21:44
← mrq1 left21:44
→ ereslibre joined21:44
← ereslibre left21:44
→ ereslibre joined21:44
k0ral not even an idea of how one could share only a subset of his repository ?21:44
→ verto joined21:44
mattly is there a way to retrieve a stash after I dropped it? They don't seem to stay in the reflog :/21:45
→ workmad3 joined21:46
← PabloM left21:47
→ vitor-br joined21:47
← mrevd left21:48
→ pedrobelo joined21:48
thiago_home mattly: stash uses the reflog21:48
mattly: once you drop it, it is removed from the reflog21:48
mattly thiago_home: gotcha, thanks21:48
← airborn left21:48
thiago_home mattly: there's no reflog of the reflog21:48
mattly :/21:48
amerine mattly: Yeah21:49
mattly I should probably spend an hour or so reading up on the stash docs before using it more heavily21:49
← ShooterMG left21:49
← booginga left21:51
← workmad3 left21:51
← ThomasLocke left21:53
← paja left21:54
→ xiong joined21:56
→ mrevd joined21:56
← bartek left21:56
← vitor-br left21:56
← cocoadaemon left21:56
jds thiago_home: Yeah, reset that 1 file to where it was immediately after 'git merge', but before I started screwing around with my mergetool21:57
→ paja joined21:57
→ warlock_mza joined21:57
→ uncaught_excepti joined21:57
→ drubies|mbp joined21:58
← nurey left21:58
→ sylr joined21:58
← uncaught_excepti left21:58
CarlFK I made about 50 changes accross 17 files. 40 of the changes I want to keep, the rest dump. (yeah, on the fly hacking to make it work)21:58
my thought is to git diff >foo, vim foo, remove the stuff I don't want, then patch a fresh checkout.21:58
→ uncaught_excepti joined21:59
CarlFK is there a more git foo way of doing this?21:59
← ereslibre left21:59
← robert_david left22:00
cbreak CarlFK: git checkout -p22:01
→ vitor-br joined22:01
cbreak answer yes if you want to remove the thing (and check out the staged version instead)22:01
CarlFK plan B: diff>foo, commit everything, use foo to track down the changes I didn't want comited and undo them by hand (those are trivial, like using a hardcoded value instead of reading from a .conf22:01
← k0ral left22:01
← DR01D5 left22:01
cbreak no if you want to keep it22:01
or git add -p, and only add to the staging area what you want to commit22:01
leave the rest uncommitted22:01
← uncaught_excepti left22:01
→ diegoviola joined22:02
CarlFK cbreak: is 'thing' a whole file or diff chunks ?22:02
cbreak the latter22:02
try it out22:02
CarlFK awesome22:02
← sylr left22:02
← q0rban left22:05
← vitor-br left22:06
cbreak isn't it cool? :)22:06
→ sketchyd joined22:06
← thiago__ left22:06
CarlFK beyond cool22:07
← guardian left22:07
cbreak add -p is a special case of add -i22:07
← paja left22:07
cbreak take a look at that too if you want22:07
CarlFK Discard this hunk from worktree [y,n,q,a,d,/,j,J,g,e,?]? too many options! :)22:08
← Mark_Morcom left22:08
cbreak try e22:08
← sebersole left22:08
cbreak for edit :)22:09
s is split22:09
CarlFK nuh uh...22:09
cbreak or just try ? and it'll tell you22:09
← Cap_J_L_Picard left22:09
CarlFK impossible!22:09
edit.. that's just too crazy useful22:09
← drubies|mbp left22:09
→ vitor-br joined22:09
cbreak if you use git add -p, don't forget to do a diff --cached and diff afterwards to check if what you wanted to happen really did :)22:10
→ judaz joined22:10
judaz hey... a question22:10
git clone -b [branch] [repo] [directory]22:10
does a clone of a branch of repo in directory?22:10
CarlFK cbreak: I can't handle the options :)22:11
→ ereslibre joined22:11
← ereslibre left22:11
→ ereslibre joined22:11
← alesan left22:11
CarlFK Split into 2 hunks. - omg, it did what I want.22:11
← empity left22:14
← f00li5h left22:14
← synisma left22:15
→ Cap_J_L_Picard joined22:15
← dpino left22:16
→ sylr joined22:16
← LiamH left22:16
→ rideh joined22:16
← xerora left22:18
← lessthanzero left22:18
← Swimming_bird left22:19
← SethRobertson left22:19
← rorist left22:19
→ SethRobertson joined22:20
← sylr left22:20
→ sylr joined22:21
← dv310p3r left22:21
→ f00li5h joined22:22
← bdiu left22:23
→ nurey joined22:23
→ sako joined22:25
→ marciodias joined22:25
← spaceonline left22:25
→ dajhorn joined22:25
→ Swimming_bird joined22:26
← sylr left22:26
sako hey guys, i know theres many ways to do this.. what is the best way to revert an entire git repo to a certain commit?22:26
still not so great with altering history :?22:26
:/22:26
teuf you can use git reset --hard but this alters the history22:27
sako i dont mind losing everything22:27
teuf you can use git checkout SHA1 -- .; git commit22:27
sako that came after this commit22:27
← Nemesis13 left22:27
teuf and this will create a new commit resetting the repo to the old state22:27
sako: are other people using your repository ?22:27
sako actually id prefer to lose everything that came after :P22:27
no22:27
teuf ok, so git reset --hard then22:28
← opalka left22:28
sako o just git reset --hard sha1/22:28
?22:28
teuf yep22:28
unless you have uncommitted changes in your working dire22:28
← ph^ left22:30
→ paja joined22:32
← irc2samus left22:32
→ PabloM joined22:33
← dyer left22:33
→ \monster\ joined22:33
\monster\ hey guys, whats the proper way to undo last commit, I commited the wrong file ..I currently have uncommited changes in current branch and don't want to lose anything, I just want to undo the last one22:34
← nurey left22:35
teuf \monster\: git reset HEAD^22:35
← nibblebot left22:35
CareBear\ \monster\ : or use git commit --amend to replace the previous commit with some new changes22:35
→ Zenith77 joined22:35
\monster\ hmm22:36
← jutaro1 left22:36
← rideh left22:36
\monster\ reset HEAD^ will only undo the last commit, it won't touch anything in current branch thats uncommited?22:37
teuf \monster\: yes22:37
\monster\: if you are worried, do a git diff HEAD^ >backup.diff22:37
\monster\ oki22:37
← sdboyer left22:37
sako teuf: after i do the reset --hard i have to force push it?22:38
teuf sako: you changed history using reset --hard, so you can't push unless you use --force22:38
sako i see ..22:38
k thanks :)22:38
← thiago_home left22:39
→ superdmp joined22:39
← logickal left22:40
← ahf left22:40
PuffTheMagic is there a way to revert multiple consecutive patches?22:40
s/patches/commits/22:41
→ sdboyer joined22:41
→ SirFunk joined22:42
engla PuffTheMagic: you can apply the reverse diff over the range of commits. or revert them one by one then squash the reverts together22:42
PuffTheMagic: I wonder if git revert doesn't take multiple commits nowadays though22:42
CareBear\ dragon!22:43
← orafu left22:44
→ tudor joined22:47
← Sonderblade left22:49
→ simplechat joined22:49
← simplechat left22:49
→ simplechat joined22:49
← tudor left22:50
→ project2501a joined22:50
→ angelsl joined22:50
neuro_damage is it safe to say that if there is another git project B which is a "module" to a larger project A, is it possible where project A can suck in portions or all of project B22:52
just curious about how to do that in particular with git22:52
I suppose that's sub-modules, I'm not too familiar with that part22:52
← systemaddict left22:53
karihre get familiar with it :)22:53
← sketchyd left22:53
→ orafu joined22:53
→ systemaddict joined22:53
→ ISF joined22:54
→ tudor joined22:55
bob2 is there a way to find all the heads in a repository?22:56
← tudor left22:56
← beatak left22:57
bob2 hm, fsck --lost-found maybe22:57
CareBear\ including detached ones?22:57
PuffTheMagic git reset --hard a; git reset --mixed d22:58
^^ seemed to do the trick22:58
← jds left22:59
bob2 CareBear\: yeah22:59
was fiddling too much and deleted a branch and can't find it in the reflog23:00
← unvs left23:00
→ patmaddox joined23:00
← cannonball left23:00
bob2 (the list of non-detached heads is just everything in refs/ right?)23:00
← dim_ left23:01
→ niblha joined23:01
steven guys,23:01
whats the trick that git uses to display things like "git log" or "git diff" using a less-like interface?23:01
← boombatower left23:01
steven how does it do that? does it just shell out to `less`, if its available?23:01
← phantomcircuit left23:02
→ boombatower joined23:02
← boombatower left23:02
→ boombatower joined23:02
bob2 PAGER23:03
or less23:03
← radarek left23:03
→ Buglouse joined23:03
steven oooh.23:03
← eijk_ left23:04
← rendar left23:05
→ sunblush joined23:05
← Yuuhi``` left23:06
→ Yuuhi``` joined23:06
← _psychic_ left23:06
→ Misafir112121030 joined23:06
← fisted_ left23:06
→ fisted joined23:07
← marciodias left23:07
→ RaceCondition joined23:08
← tgunr left23:10
→ tizzo-afk joined23:10
← MikeChelen left23:10
→ varioust joined23:10
← kipras left23:11
← sdboyer left23:12
→ sdboyer joined23:12
← Misafir112121030 left23:13
→ seivan joined23:14
→ bdiu joined23:15
← giallu left23:16
→ ahf joined23:16
← Amorphous left23:17
→ alberto56 joined23:17
jayne why is git cat-file blob <sha1> giving me a bunch of crap instead of the file contetns?23:21
^contents23:21
← xiong left23:21
→ icwiener joined23:23
← superdmp left23:24
← tizzo-afk left23:25
← shruggar left23:26
← wolfe left23:29
→ tizzo-afk joined23:29
ayust is the SHA you're giving it actually a blob?23:30
→ madewokherd joined23:30
← panike left23:30
ayust and if it is, is it a text file blob?23:30
jayne I think I must have typoed it... I got it to work now... which leads to the next question:23:30
why does git insist a file is modified, even though the working-dir contents match cat-file byte-for-byte23:31
← meandi left23:32
ayust git update-index --refresh23:32
→ Amorphous joined23:32
steven :)23:32
jayne it says b512: needs update23:32
Dr4g I got changes staged for commit23:32
but i want to take them off23:33
jayne (doesn't status do an update-index anyway?)23:33
Dr4g how can i do that23:33
jayne Dr4g: git reset23:33
ayust jayne: status doesn't --refresh23:33
Dr4g thanks jayne23:33
→ meandi joined23:33
jayne status still thinks b512 is modified though23:33
I did a git cat-file blob $sha1 > b512 (where $sha1 comes from master)23:34
ayust try just resetting it?23:34
git reset HEAD b51223:35
jayne "Unstaged changes after reset: b512"23:35
ayust k, what does git diff show then?23:35
steven are tree objects ever created when populating the index, or are they only created when creating a commit so the commit has somethign to attach to?23:36
jayne every single line in the file removed, then added back23:36
ayust jayne: line ending issues?23:36
jayne I would have thought, but cat-file should have bypassed any of that23:36
steven from what i understand, when populating the index, trees are never involved, only blobs keyed by pathnames23:36
is that correct?23:36
ayust jayne: check manually?23:37
jayne Just to prove it to myself, I've swapped the line endings in the file and it still isn't happy.23:37
ayust k. encoding?23:37
jayne If I use a difftool of vim -b, it shows the line endings match23:37
how would I check encoding?23:38
← tizzo-afk left23:38
ayust vim's statusline should have it i think23:38
steven nobody knows?23:38
ayust jayne: also, "file -i b512" should give you a charset23:38
← boombatower left23:39
→ dyer joined23:40
← dyer left23:40
→ dyer joined23:40
jayne text/plain; charset=us-ascii23:40
→ johnw joined23:40
ayust k23:41
try checkout then; git checkout HEAD -- b51223:42
← bdiu left23:42
→ phantomcircuit joined23:42
jayne it completes without errors, but still shows b512 modified in status23:43
← Chillance left23:43
→ _psychic_ joined23:43
ayust modified but unstaged?23:43
jayne yes23:43
ayust huh.23:44
→ rcu joined23:44
→ boombatower joined23:44
← boombatower left23:44
→ boombatower joined23:44
jayne normally that can happen with the repository has CRLF endings in it, but one of the things I've tried should have worked around that if it were the issue23:44
→ rchavik joined23:45
→ bdiu joined23:45
← larsl left23:45
ayust what's your eol config variable set to?23:45
core.eol23:45
← rchavik left23:46
→ rchavik joined23:47
jayne this is v1.6.5, so it doesn't have core.eol (but it has core.autocrlf)23:48
← Ryback_ left23:48
← rchavik left23:48
jayne I've tried setting it to false, but it still behaves the same23:48
← rcu left23:48
→ rchavik joined23:49
← smolyn left23:49
steven so sad23:50
maybe im on ignore..23:50
jayne steven: or simply, nobody here knows the answer and you should try the mailing list instead23:50
→ jason237 joined23:50
steven oh. but it seems like such a simple quesiton23:50
← linusoleander left23:50
steven hey Tv hows it going?23:50
jayne "why do magnets work?" seems like a simple question, too23:50
→ fuzzybunny69y joined23:50
steven thats totally not analogous ;)23:51
← ptl left23:51
jayne whether trees are part of staging is completely outside of my experience with git, so it's at least analogous in sense that how spin generates magnetism would be outside the experience of non-scientists.23:53
→ cordoval_ joined23:54
steven oh.23:55
but its not far-fetched to imagine someone here might know :)23:56
← sftp left23:56
jayne indeed, but perhaps you've hit an unlucky/sleepy time23:56
ayust yeah, a lot of the people who are knowledgeable about the internals don't seem to be active atm23:57
steven rats23:57
well thanks anyway guys :)23:57
← MikhX left23:57
steven although i should have checked the git book again lol - http://book.git-scm.com/7_the_git_index.html23:57
it kind of answers my question23:57
← d0k left23:57
← sako left23:58
→ Laggg joined23:58
← tjholowaychuk left23:58
← judaz left23:58
← bdiu left23:58
→ ptl joined23:59

Logs Search ←Prev date Next date→ Channels Documentation