IRCloggy #git 2012-02-20

Logs Search ←Prev date Next date→ Channels Documentation

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

2012-02-20

← jrnieder left00:01
→ sattu94 joined00:01
→ johnkpaul joined00:02
someprimetime so i'm doing git clone and giving it the url of where i set my repo up on my server and it's asking me for a username and password.. is it the server username/pass or is it a git username/pass (which i haven't set up if this is the case)?00:02
← wuj left00:02
someprimetime the server un/pw combo won't work so i'm assuming i may have no setup a un/pw for git00:03
posciak hi, I'm doing some work on top of a huge project. I'd like to backup that work periodically, but I wouldn't like to backup the whole repo and its history of that project, because it's huge and well backed up already. So I've been thinking that instead of creating a copy of the whole repo, I could use git bundle create foo master ^origin/master and copy the file over to backup storage regularly. Is this a00:03
good idea? Any better ways to do it?00:03
← satyanash left00:04
m1sc posciak: why not just pushing to backup-repo?00:04
→ ajt joined00:04
posciak m1sc: because the backup-repo would be very big00:05
there is no need to store the entire history00:05
just a few commits that I've made00:05
m1sc: well, I could also use a shallow backup-repo00:05
is that better?00:06
← dnjaramba left00:06
← mastro left00:07
sitaram bundle is pretty much the only sane way to backup a "top layer of commits" if the older stuff is already well protected from loss00:07
← [M]ax left00:07
← ixti left00:07
posciak ok, sounds great :)00:08
→ dnjaramba joined00:08
→ quazimodo joined00:09
→ [M]ax joined00:10
someprimetime what00:10
what's the benefit of something like gitosis00:10
← shibayu36 left00:11
cmn currently, none !gitosis00:11
gitinfo gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite00:11
cmn !gitolite00:12
gitinfo Want to host as many git repos (and users!) as you like, on your own server, with fine-grained access control? You want gitolite: https://github.com/sitaramc/gitolite00:12
cmn as the bot says, access control based on user, branch, file00:12
← johnhamelink left00:12
→ Biopandemic joined00:13
→ dotnull joined00:13
someprimetime thanks cmn00:13
→ Danielpk|rdlBNC joined00:14
→ satyanash joined00:14
← [M]ax left00:16
← sattu94 left00:17
← Biopandemic left00:17
→ Biopandemic joined00:17
→ [M]ax joined00:17
→ corenet joined00:18
corenet Hello everyone.00:18
← ph^ left00:19
← snearch left00:19
corenet New to git and have a (hopefully) simple question...00:19
FauxFaux corenet: !hi00:20
gitinfo corenet: [!welcome] Welcome to #git, a place full of helpful gits. If you have a question, feel free to just go ahead and ask—somebody should answer shortly. For more info on this channel, see http://jk.gs/git/ Take backups (type !backup to learn how) before taking advice.00:20
corenet If I have a bare git repository that has no working folder and I clone it, how do I get the files for the project?00:20
FauxFaux corenet: You have them. That's what clone does.00:20
corenet so if I set up a bare repository and I clone it, I have a full working copy on the new machine with all the files AND the .git folder?00:20
cmn with a normal clone, you get them automatically00:20
← dnjaramba left00:20
cmn you get them anyway, but a clone w/o --bare will checkout the files as well00:21
corenet But the bare repository only has the .git files, correct? Where would the local machine cloning get the working files from?00:21
Thanks for entertaining a noob :)00:21
cmn .git is where all the information is00:22
the working directory is for you to modify files; git has its own copy00:22
← Danielpk|rdlBNC left00:22
corenet So the .git directory actually has the files I am working with in addition to the local folder.00:22
cmn the .git directory is where everything is00:23
corenet Just trying to wrap my head around the notion of a bare repository for push/pull00:23
cmn the working directory is for humans00:23
what you transmit on fetch/push is the history information00:24
adding a working directory complicates things, it's not the other way around00:24
→ dnjaramba joined00:24
← tga left00:24
→ marvin_ joined00:24
← alexgordon left00:24
→ linusoleander joined00:25
posciak corenet: bare repository has all the files that you've commited. If you ever lose your working directory but not .git directory, you will only lose the changes that you've made since your last commit.00:26
i.e. diffs, not files00:26
← arietis left00:26
← dotnull left00:27
cmn git doesn't store diffs, it stores snapshots00:27
posciak yes00:27
I said he'd lose diffs00:27
note store :)00:27
→ tga joined00:27
posciak but good you've clarified this actually00:27
one of the common misconceptions :)00:27
cmn oh, I thought you meant to replace the "has all the files" with "all the diffs"00:28
→ arietis joined00:28
posciak ohh right, I didn't realize, sorry00:29
← dnjaramba left00:29
posciak I would've interpreted it this way too ;)00:29
← caseymcg left00:30
→ chrisdev joined00:30
→ rchavik joined00:30
← johnkpaul left00:30
→ cybersphinx joined00:32
← bitkiller left00:36
← the_cybersphinx left00:36
← crisp left00:36
← DrHouse|TARDIS left00:37
→ dnjaramba joined00:37
← cooldude left00:40
← dnjaramba left00:41
→ dnjaramba joined00:41
← Sonderblade left00:42
→ cooldude joined00:42
← cooldude left00:42
→ cooldude joined00:42
← cooldude left00:42
→ Archalia joined00:44
→ shibayu36 joined00:44
corenet Thanks for the answers everyone - had to walk away briefly.00:45
posciak corenet: you are very welcome00:45
← legnaleurc left00:45
→ cooldude joined00:45
← Archalia left00:46
← airborn left00:46
← corenet left00:47
← d0k left00:47
← marvin_ left00:47
→ g0th joined00:48
g0th hi00:48
I am about to commit a really huge change00:48
← brian_g left00:48
g0th I thought it might be useful to somehow give this point in the history a name, I guess that would go under "tag"?00:49
← soulcake left00:49
→ marvin_ joined00:49
← iocor left00:49
g0th so do I tag now "before" the big commit?00:49
or after or both?00:49
or how does that work usually?00:49
cmn it doesn't matter when00:49
a tag is a pointer to a commit, you can create one whenver00:49
g0th when I type tag ...00:50
git tag -a ....00:50
which commit does it tag?00:50
cmn it should default to HEAD00:50
g0th and that was the last one?00:50
→ alberto56 joined00:51
g0th let me put it differently, how do I get the situation back right before the big commit?00:51
how after and how do I get the diff later?00:51
cmn that's called HEAD~ in this case00:51
if you haven't committed, then you're interested in HEAD00:51
→ brian_g joined00:51
← alberto56 left00:51
cmn or you can check the name and go that way00:52
g0th I still have no clue what I should do now00:52
cmn git rev-parse HEAD will give you the commit's name00:52
g0th I never heard of that command00:52
cmn it's low level00:52
→ mikepack joined00:52
posciak g0th: tag the big commit and you can then refer to situation after by tagname and situation before by tagname~00:52
g0th I thought tags were something simple00:52
cmn they are00:52
g0th posciak: ok, that helps :)00:52
cmn you tag whatever you want00:52
posciak g0th: but as cmn says, it doesn't matter00:53
cmn if you omit what, it defaults to HEAD00:53
g0th what is the most common practise?00:53
cmn to tag commits00:53
g0th so I first commit, then git tag -a ... ?00:53
Z3Rt → Z3Rt|away00:53
→ legnaleurc joined00:53
g0th because then it tags HEAD commit which was the last one?00:53
cmn the order doesn't matter00:53
HEAD is the last one you made00:54
g0th the order matters very much00:54
if I first tag it will tag the commit before the big commit00:54
→ ploop joined00:54
cmn you can tell tag to tag whatever you want00:54
posciak g0th: you can tag both if that gives you peace of mind :)00:54
cmn it doesn't matter if you do it now or later00:54
g0th yes I get that00:54
cmn tell it to tag HEAD now or HEAD~ alter00:54
g0th but I want to do it now00:54
← zatan_ left00:54
cmn then do it now00:54
g0th ok, so the most common practise is to tag the big commit right? so I do the following _now_:00:55
cmn common practice is to tag important commits, usually releases00:55
g0th git commit... git tag ... and later how do I recover before/after/diff?00:55
cmn recover?00:55
g0th not to recover00:56
hmm00:56
cmn man git show if you want to see the changes a commit introduced00:56
gitinfo the git-show manpage is available at http://jk.gs/git-show.html00:56
← dnjaramba left00:56
posciak g0th: you can always checkout your tag, you can reset your branch to it, or do whatever you wish00:56
g0th to "check back" how it was before00:56
cmn what do you want to do?00:56
g0th ah checkout, thx00:56
cmn what do you mean by check?00:56
g0th maybe recover as you mentioned00:56
or to just check something which was lost in the big commit00:56
→ dnjaramba joined00:56
cmn it's not lost00:57
g0th exactly00:57
posciak g0th: nothing gets lost, even if you don't tag00:57
g0th that's why I want to know how to get it back using the tag name00:57
yeah00:57
but then I need to figure out exactly where etc00:57
so much work00:57
that's exactly why I want to tag it00:57
posciak g0th: normally, you should make your commit message informative enough to be able to find the big commit00:57
← libertyprime left00:57
g0th hmm, tag reference still seems simpler00:57
posciak yes, probably00:58
g0th I have no clue how to find a commit by message content00:58
cmn if there is so much in that commit that you can't figure out what it changes, you should split it into several00:58
posciak so I'm not saying don't do it00:58
g0th git checkout "tagname"?00:58
cmn yeah, just like any other ref00:58
g0th how would I show the commit diff?00:59
← cooldude left00:59
cmn 01:56:01 <cmn> man git show if you want to see the changes a commit introduced00:59
gitinfo the git-show manpage is available at http://jk.gs/git-show.html00:59
g0th thanks00:59
I know I ask too many questions, sorry ^^00:59
I abuse you guys a bit and you are right to "rtfm" me a bit01:00
← [M]ax left01:00
g0th it's just that I wanted to get this right01:00
← Kre10s left01:00
→ cooldude joined01:00
→ jameslordhz joined01:01
← ceej left01:01
← shibayu36 left01:01
jameslordhz hi01:01
→ [M]ax joined01:02
posciak g0th: we are friendly and answer happily. Sometimes it's just easier to point to manual than to retype it here01:02
g0th: but ask away if in doubt01:02
← pdtpatr1ck left01:03
g0th "For tags, it shows the tag message and the referenced objects."01:03
so git show "mytag" doesn't show the diff?01:03
← arietis left01:03
SethRobertson For annotated tags01:03
cmn that means for real tags, the ones that have a message01:03
"and the referenced objects"01:04
g0th ah01:04
SethRobertson But yes, it shows the diff, which is the referenced object01:04
g0th that would be the commit01:04
posciak g0th: it does show the diff too01:04
g0th the referenced object?01:04
cmn it shows whatever it would show for the object01:04
→ mikepack_ joined01:04
→ shibayu36 joined01:04
g0th referenced object===commit?01:04
posciak g0th: yes01:04
g0th ok thanks01:04
cmn in this case yes01:04
← mikepack left01:04
SethRobertson I wonder what would happen if you made an annotated tag on an annotated tag01:04
← edude03 left01:04
esc SethRobertson: the tag will point to the other tag01:05
cmn hm, interesting01:05
SethRobertson Yes yes. But what does `git show tag2` say?01:05
esc using '^{}' you can deref a tag IIRC01:05
← tga left01:05
esc SethRobertson: it will show the tag, the tag, and then whatever the last tag points to01:05
tree, commit, or blob01:06
;)01:06
at least it did that last time i checked, which is like > 12 months ago, so don't trust me on it ;)01:07
← mikepack_ left01:07
SethRobertson I tested it, and it does show all three objects. I find myself impressed that the dev cared enough to allow that to work01:07
← dnjaramba left01:07
SethRobertson I might not if I were writing it, unless recursion made sense for other reasons01:07
→ arietis joined01:08
→ txomon|home joined01:08
← txomon|home left01:08
→ txomon|home joined01:08
esc SethRobertson: well, it does just deref the sha of whatever if pointed to, until nothing is pointed to anymore01:08
← subbyyy left01:08
→ GeertB joined01:08
cmn I wouldn't be surprised if it did if (is_tag) { start_again(*tag); }01:08
it makes perfect sense to me to make it recursive01:09
esc SethRobertson: no wait, it shows the diff, right?01:09
→ Redbluegreen joined01:09
SethRobertson Yes, two tags and a commit01:09
← Redbluegreen left01:10
esc SethRobertson: hmmm, interesting01:10
→ anjen joined01:10
esc SethRobertson: so it probably needs some additional logic, beyond the text-book recursion...01:11
SethRobertson Not really.01:11
← fir31ight left01:11
→ Esa_S joined01:12
→ dv310p3r joined01:12
→ apok joined01:12
→ sudoquai_ joined01:12
jameslordhz just now i check out a git project from github, and make a git repo on my vps using git init --bare, then i import this project to eclipse, git status and find some changs, so i commit it to my vps repo using git remote add vps_repo, and git push vps_repo master , after that i find it is better to make modification on my own branch, but i have made changes to master branch and pushed it to repo git repo, so how to make a new branch based on the formmer commit01:12
→ dnjaramba joined01:13
→ fir31ight joined01:13
cmn !fixup01:13
gitinfo So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions. For hints type !fixup_hints in IRC. Remember: if you have pushed already, there are only a few things you can do without !rewriting_public_history (type that for more info)01:13
cmn the short answer would be git branch newbranch master01:13
→ mikepack joined01:14
SethRobertson Which would take care of making the new branch, but not fixing master01:14
Z3Rt|away → Z3Rt01:14
← Z3Rt left01:14
← Pugnax left01:14
← sudoquai left01:14
→ Z3Rt joined01:14
← anjen left01:15
jameslordhz cmn dude, i have already made change to master branch:(01:15
SethRobertson Yes. If only we had discussed that issue.01:15
cmn I should add a !dude hook to gitinfo01:15
SethRobertson Step one: make the new branch. Step two: get rid of the unwanted commit (see the fixup link)01:16
jameslordhz cmn so it is better to create a branch based on HEAD~1, but not HEAD01:16
cmn have you considered the answers?01:16
← hyperair left01:16
← Esa_S left01:16
cmn you said you wanted that commit in the new branch, why do you want to base that branch on an earlier commit?01:17
← dnjaramba left01:17
→ igotnolegs joined01:17
← chrisdev left01:17
→ lyddonb joined01:18
← posciak left01:18
← bgerber left01:18
→ dnjaramba joined01:18
jameslordhz cmn master branch is used to track project on github, and i have no priviledge to commit to that repo, and it's better that i do not pollute master branch, agree?01:18
→ herdingcat joined01:18
→ Esa_S joined01:18
cmn how is that different from anything we've said?01:18
furthermore, branches aren't tied to repos01:19
SethRobertson And it is *your* master branch. You can do with it whatever you want.01:19
cmn you can push to master on your repo01:19
jameslordhz SethRobertson dude, yes it's my branch, i can do anything, but i don't want conflict when i use git pull on master branch, that branch is reserved just to track updates from github, then i can get updates to my modification branch:)01:20
→ aspotashev joined01:21
cmn then do that01:21
jameslordhz cmn dude, i do not want conflicts on master branch:(01:21
cmn we've been telling you how to do that01:21
SethRobertson You don't need to reserve master for that, you would have upstream/master for that purpose. But whatever you want.01:21
jameslordhz cmn dude, you are not:(01:21
SethRobertson (20:16:17) SethRobertson: Step one: make the new branch. Step two: get rid of the unwanted commit (see the fixup link)01:22
(20:13:56) cmn: the short answer would be git branch newbranch master01:22
→ hyperair joined01:22
SethRobertson (20:13:16) cmn: !fixup01:22
gitinfo So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions. For hints type !fixup_hints in IRC. Remember: if you have pushed already, there are only a few things you can do without !rewriting_public_history (type that for more info)01:22
→ edude03 joined01:22
jameslordhz SethRobertson where is upstream/master branch? i can not find such branch :(01:22
cmn jameslordhz: have you looked at and tried to undestand what those do?01:22
← Esa_S left01:23
cmn step one is the first thing you pasted from me01:23
and step two is the fixup01:23
SethRobertson jameslordhz: In order to get updates from the true upstream, the place you do not have commits from, you need to add it as a remote `git remote add upstream git://public/url/for/project01:23
jameslordhz SethRobertson dude, i have make commit to master, and pushed it the vps_repo, so the master branch now is not the one before i made commits, how can git branch newbranch master works in the way i want:(01:23
→ sattu94 joined01:24
→ Esa_S joined01:24
cmn do you know what "git branch newbranch master" does?01:24
SethRobertson (Hint, the answer is not. See man git-branch)01:24
gitinfo the git-branch manpage is available at http://jk.gs/git-branch.html01:24
jameslordhz cmn dude, create a branch based on master's current commit point, right?01:25
→ ajaykools joined01:25
← satyanash left01:25
cmn and how is that not exactly what you asked for?01:26
and for the nth time, don't call me dude01:26
← dnjaramba left01:27
jameslordhz why git branch -r is different from -a?01:27
→ dnjaramba joined01:27
cmn because they're different options01:27
each shows you different things01:28
→ cooloney joined01:28
jameslordhz what is the difference then?01:28
cmn man git branch01:28
gitinfo the git-branch manpage is available at http://jk.gs/git-branch.html01:28
→ fisted_ joined01:28
g0th aaarg01:29
now I did the commit command without adding any files ^^01:29
lol01:29
← fisted left01:29
← [M]ax left01:29
← drizzd left01:30
g0th does it do anything if I do commit without any changes?01:30
→ [M]ax joined01:30
jameslordhz is it need to reserve master branch to track remote changes?01:30
→ forsto joined01:30
→ Aaaarg joined01:30
SethRobertson g0th: It prints nothing to commit (working directory clean)01:30
jameslordhz: No. You may do it, but it is not required01:31
cmn g0th: if there are no changes in the index, it will complain01:31
SethRobertson jameslordhz: You could give it another name, you could just depend on merges/rebases. You have the power.01:31
jameslordhz SethRobertson dude, is it a best practise to reserve master branch to track remote repo which you have no rights to write to?01:32
SethRobertson It is not a best practice. As I said, if you are tracking the true upstream there is already a branch reserved for that.01:32
bremner is it unreasonable to want to /kick someone for overusing the word dude?01:32
← vandal- left01:32
SethRobertson No!01:32
cmn not at all01:32
jameslordhz: branch names are just names; master is the default so many people use it, but there is no need for it01:33
→ bgerber joined01:33
SethRobertson But even when using master, reserving it to track another repo's master isn't exactly common. Some people do, some people do not01:35
← hyperair left01:35
jameslordhz cmn dude, the best solution is just as said, that is git branch newbranch master, right? then i could reset commits to Head~1 on master branch, that is what i want?01:35
ChanServ set mode: +o01:35
← ryanisnan left01:36
cmn since you clearly can't understand simple instructions and don't bother reading any documentation, I'm going to stop here01:36
jameslordhz SethRobertson so how to reserve a clean branch to track remote changes, and only make modification on my own branch?01:36
→ hyperair joined01:37
← herdingcat left01:37
ko1 bremner, chill, dude01:37
bremner looks innocent01:37
← dnjaramba left01:37
→ dnjaramba joined01:38
SethRobertson I agree. We have told you N times how to do what you want jameslordhz and I have already told you that you already have reserved a clean branch. Do what we tell you to and come back if it does not work.01:38
← ploop left01:39
jameslordhz SethRobertson you mean a branch named remotes/master, right?01:39
g0th how can I check if it commited?01:39
→ boombatower joined01:40
SethRobertson jameslordhz: For values of "remotes" equal to the name of a non-origin remote, yhes01:40
cmn 'git commit' will have told you, git show shows HEAD by default01:40
g0th can I see a history of commits?01:40
cmn git log01:40
→ DofLeFou joined01:40
cmn you might benefit from the !book01:40
gitinfo There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !vcbe and !parable01:40
g0th in git log I don't see my message01:41
does it mean it was not commited?01:41
(which would be perfect)01:41
→ dfr|mac joined01:41
SethRobertson g0th: Run: gitk --all --date-order01:41
Or `git log --all --date-order`01:42
In case it is on a different branch01:42
g0th I still don't see it01:42
ko1 git diff?01:42
jameslordhz SethRobertson you mean there is already some branches created by git to track remote updates, and these branches cannot be polluted by my modification, right?01:42
ko1 git status01:42
cmn what did git say when you told it to commit?01:42
g0th git diff is huuge01:42
← dnjaramba left01:43
g0th it printed something like a status01:43
← Heimidal left01:43
g0th forgot exactly01:43
→ dnjaramba joined01:43
→ dfr|mac_ joined01:43
g0th I don't think it printed "commited ...."01:43
SethRobertson jameslordhz: Yes01:43
cmn read it next time01:43
g0th is that normal if I didn't add _anything_01:43
SethRobertson g0th: yes01:43
cmn yeah01:43
g0th so it most probably didn't really commit?01:44
nice! :-)01:44
bremner set mode: -o01:44
← sekati left01:44
g0th ok, now it showed a completely different message, thanks a lot guys01:46
← cjuner left01:47
← dfr|mac left01:47
→ wuj joined01:47
← dnjaramba left01:48
jameslordhz say SethRobertson run git branch remotes/origin/master, what will happen?01:48
→ dnjaramba joined01:48
jameslordhz SethRobertson after that remotes/origin/master can no longer track remote changes, yes?01:49
SethRobertson After that you have a mess, but *a* remotes/origin/master will be tracking the right stuff.01:50
Run `git show-ref` and you will see the full name01:50
← sudoquai_ left01:50
→ nutron|w joined01:51
SethRobertson However, the basic answer is do NOT create a branch with a reserved word directory component or a remote's name directory component. It may work, but it is not a good idea01:51
→ jason237 joined01:52
← ajaykools left01:53
→ ceej joined01:53
← ceej left01:54
← ajt left01:55
← Esa_S left01:55
→ ajt joined01:56
jameslordhz SethRobertson so there is no need to reserve master branch to track remote updates, right? since there is a buildin branch to do work automatically, right?01:57
SethRobertson yes01:57
jameslordhz SethRobertson dude, so if i want to read code of the not modified version, that is it only has updates from remote repo, how to do this?01:59
→ Turicas joined01:59
SethRobertson git remote update upstream; git checkout upstream/master; Assuming the remote is named upstream of course02:00
You will be on a detached head, but don't let that disturb you. Just don't make commits02:00
→ ploop joined02:00
→ BIRCOPI986 joined02:01
← hyperair left02:01
← Turicas left02:01
← BIRCOPI986 left02:01
→ ChaoWang joined02:03
→ Turicas joined02:03
jameslordhz SethRobertson dude, git checkout remotes/origin/master will create a branch named remotes/origin/master that is /refs/heads/remotes/origin/master, not a detached branch:(02:03
← dnjaramba left02:03
→ dnjaramba joined02:03
SethRobertson Why are you running different commands than the ones I told you to run?02:03
jameslordhz git branch remotes/origin/master, dude, i run this command02:04
after the command has been runned, i am not on a detached branch:(02:05
SethRobertson Yes. I didn't tell you to do that and bad things happened when you did. Please delete that branch and run the command I told you to02:06
jameslordhz git remote update origin?02:06
← matthiaskrgr left02:07
→ matthiaskrgr joined02:07
SethRobertson Have you deleted the branch? Does `git branch` print anything about remotes or anything with a / in it?02:08
→ pdtpatr1ck joined02:08
← marvin_ left02:09
← Chillance left02:09
jameslordhz http://paste.ubuntu.com/849446/02:10
yes, dude02:10
SethRobertson Good. Step one, create the remote like I told you to. Search for `git remote add` in your channel history02:11
Or read the man page02:11
← brian_g left02:11
jameslordhz SethRobertson dude, what do you want me to do? i have already add vps_repo using git remote add:)02:12
← dnjaramba left02:12
SethRobertson Then do `git remote update`02:12
→ dnjaramba joined02:12
jameslordhz SethRobertson http://paste.ubuntu.com/849448/ i am still on master branch, dude , not a detached branch:(02:13
→ kurtzhong joined02:13
SethRobertson Yes, but I think you will find I told you to run a command after that.02:14
(21:00:04) SethRobertson: git remote update upstream; git checkout upstream/master; Assuming the remote is named upstream of course02:14
Oh, and I don't recommend naming the remote "notes" either. You might confuse people into thinking it has something to do with the notes functionality in git02:14
← Yuuhi` left02:15
← tewecske left02:16
→ Jay_Levitt joined02:16
jameslordhz SethRobertson why i should use git checkout origin/master, but git checkout remotes/orgin/master, the only difference is that the latter creates a branch with a branch name, the formmer is a detached branch:)02:16
← stringoO left02:16
← Turicas left02:17
→ BiggFREE joined02:17
SethRobertson git has rules about branch names and trying to DWIM. Just do it the way that works. the way that I told you.02:17
← mklappstuhl left02:17
→ Goplat joined02:17
→ zdk joined02:18
jameslordhz SethRobertson thank you: ) but what is DWIM?02:19
FauxFaux DO WHAT I MEAN02:19
← dr_win left02:20
→ hyperair joined02:21
← Destos left02:21
→ Windwaker joined02:22
lahwran Windwaker: hi02:22
← linusoleander left02:22
gitinfo set mode: +v02:22
→ subbyyy joined02:22
← jameslordhz left02:22
Windwaker oh hai lahwran :p02:22
→ Vampire0_ joined02:23
Windwaker anyone know if you can un-squash commits in git?02:23
← arietis left02:23
FauxFaux Windwaker: Sure, rebase -i, edit them, then add -p the bits you want, and re-commit the part. That, or recover your deleted history from the reflog. !gka02:26
gitinfo Windwaker: For a better way to view the reflog, try: gka() { gitk --all $(git log -g --format="%h" -50) "$@"; }; gka02:26
← ploop left02:26
Windwaker FauxFaux: But I now have 1 commit (which were 2) and I want to un-squash them02:26
or am I missing something02:26
← Vampire0 left02:26
→ n8o-mba joined02:27
SethRobertson Windwaker: See !fixup. You can do it in two ways. You can find the commit before the rebase or you can split the commit in another rebase02:27
gitinfo Windwaker: So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions. For hints type !fixup_hints in IRC. Remember: if you have pushed already, there are only a few things you can do without !rewriting_public_history (type that for more info)02:27
→ xiangfu joined02:29
Windwaker It's probably just less trouble to delete the forked repo at this point actually :p02:29
then re-fork02:30
→ posciak1 joined02:30
→ herdingcat joined02:30
SethRobertson Not necessary, but that provides a third and fourth method. I'd actually recommend doing it the hard way to gain experience in a setting where it doesn't make a difference02:30
(The fourth way is to add the upstream repo as a remote and then reset master to upstream/master, BTW)02:31
→ simplechat joined02:33
← simplechat left02:33
→ simplechat joined02:33
→ TeckniX_ joined02:34
← aspotashev left02:34
→ n8o-mba_ joined02:35
← n8o-mba left02:35
← hyperair left02:35
← dnjaramba left02:35
→ dnjaramba joined02:36
← mikepack left02:37
→ weiyang joined02:37
← TeckniX left02:37
TeckniX_ → TeckniX02:37
← flavius left02:39
→ vervic joined02:39
← jds left02:40
→ jds joined02:41
→ TTSDA joined02:41
→ hyperair joined02:42
← ajt left02:45
← TeckniX left02:46
← hyperair left02:46
→ dotnull joined02:47
→ posciak2 joined02:49
← iamjarvo left02:49
→ mklappstuhl joined02:50
weiyang hi, all, I did clone a repo and do some changes and commit, after a while I want to update the repo. So I use git fetch. Currently it works fine. But I am afraid one day I may face the conflict. So normally, what should we do if we want to do some change? create a new local branch?02:50
← posciak1 left02:51
← forsto left02:51
← dnjaramba left02:52
→ forsto joined02:53
← maek left02:53
← dotnull left02:54
→ zulax joined02:55
→ wroathe joined02:55
← groovehunter left02:56
wroathe Are there any environment variables that git sets to get the current branch?02:56
like $GIT_BRANCH or something02:56
← Windwaker left02:56
FauxFaux git symbolic-ref HEAD02:56
← apok left02:56
→ hyperair joined02:57
→ resmo__ joined02:57
← diverdude left02:58
→ Justasic joined02:58
ChanServ set mode: +v02:59
→ caseymcg joined02:59
→ Nathandim joined03:00
← pdtpatr1ck left03:00
→ rurufufuss joined03:00
← resmo left03:01
← wuj left03:01
→ tga joined03:02
→ gumnos joined03:04
← t0rc left03:05
gumnos I've got a custom log format that colorizes, but when I pipe it, it dumps the escape characters into the piped output. Is there a way to get git to recognize that stdout isn't a tty and suppress the escape characters?03:05
→ shergill joined03:05
gumnos log --date-order --format='format:%Cblue%ci %Cred%h %Creset%s'03:06
→ intripoon_ joined03:06
← twisted` left03:07
→ twisted` joined03:07
→ Balrog joined03:08
→ groovehunter joined03:09
→ dnjaramba joined03:09
Justasic does anyone in here use cgit with syntax highlighting?03:09
→ mikepack joined03:10
← intripoon left03:10
SethRobertson wroathe: no03:10
← arvind_khadri left03:11
← th5th left03:11
← Biopandemic left03:12
SethRobertson gumnos: No, because you are telling git what to do. What you could do is write a wrapper which tests the tty and calls with different arguments03:12
← twisted` left03:12
→ apok joined03:12
→ mxweas_ joined03:13
gumnos I guess I was hoping it would be like some *nix tools have a "--color=auto" option. Ah well.03:13
Since it's an alias, I can just duplicate it, making one a colorized version, the other a pipeable versino03:14
← mikepack left03:14
← dnjaramba left03:14
→ uu joined03:15
← cakehero left03:15
→ wuj joined03:15
→ dnjaramba joined03:16
← zulax left03:20
weiyang how git diff to show change of a particular commit? I have to specify the two commit number?03:20
← tga left03:20
→ tga joined03:21
tga greetings03:21
SethRobertson weiyang: That is one of your many optinos03:22
← dnjaramba left03:22
weiyang SethRobertson, yes i think so , but i found it not convenient. I have to find the parent commit of which I really want to see03:22
→ dnjaramba joined03:23
SethRobertson gumnos: It does. git's self-colorization is enabled/disabled as necessary. You are not using that.03:23
weiyang SethRobertson, so i have to search the git log03:23
← mklappstuhl left03:23
SethRobertson weiyang: If you want to see a specific commit's diff, use `git show SHA`. You could also say `git diff SHA^...SHA`03:23
weiyang SethRobertson, oh, i could use SHA^03:24
← n8o-mba_ left03:24
weiyang SethRobertson, i know the git show, sine I want to use git difftool, so...03:25
SethRobertson, your SHA^...SHA works fine, thanks03:25
SethRobertson !thanks03:26
gitinfo Feeling thankful? Type "ExampleUser++", and ExampleUser will score karma points at http://carmivore.com (our preferred way to objectify self-worth). There's really no point to thanking me – instead, why not thank the person who made me help you?03:26
← metcalfc left03:26
→ mklappstuhl joined03:26
← p3rror left03:27
→ rtconner joined03:27
tga simple one: if I merged with another branch and got a conflict (binary file), and I want to ignore the other branch's version and keep mine03:27
← mxweas_ left03:27
tga do I git add my current file or what?03:27
← Z3Rt left03:29
← orafu left03:29
→ orafu joined03:29
→ n8o-mba joined03:29
← rcu left03:30
← wuj left03:30
← neunon left03:31
→ rcu joined03:31
→ smw_ joined03:31
gumnos SethRobertson: Is there some option/config that I missed to get auto-coloration? When I run the above log command with the color format codes and pipe it to less (or vim/xxd/od), the ANSI escape sequences are present in the output. My config for color.{diff,status,branch} are all set to auto with no other color settings specified03:32
→ neunon joined03:32
gumnos or is it that, by including color formatting codes in the format-string, it overrides these, rather than behaving like the other colorizing areas of git?03:32
→ glennpratt joined03:33
← caseymcg left03:33
→ Bucciarati joined03:34
← dnjaramba left03:34
→ dnjaramba joined03:34
SethRobertson Search for color in man git-config03:37
gitinfo the git-config manpage is available at http://jk.gs/git-config.html03:37
→ preflex joined03:37
← dnjaramba left03:38
→ dnjaramba joined03:39
← cccaldas left03:40
SethRobertson tga: Yes. You can use `git mergetool` to help you do the right thing, but you essentially cause the correct file to be placed into the index03:40
→ pdtpatr1ck joined03:45
← legnaleurc left03:46
← pdtpatr1ck left03:46
→ Jedateach joined03:47
→ macabre^ joined03:47
gitinfo set mode: +v03:47
Jedateach is it possible to add a sub module to a directory that is already tracked by git?03:47
← macabre left03:47
SethRobertson git does not track directories03:48
→ VladGh joined03:48
SethRobertson So you may add a submodule in a directory which has other files tracked by git03:48
Jedateach I'm trying to add the contents of a git repo to myproject/themes/ , where myproject is git enabled03:49
but I get the error 'themes' already exists in the index03:49
→ legnaleurc joined03:49
← mklappstuhl left03:50
SethRobertson If you have files under themes (or themes is a file) known to git, you may not hide those files with a subproject.03:50
→ drpaul joined03:50
← hyperair left03:50
Jedateach SethRobertson: so the directory needs to be empty, is that what you mean?03:51
SethRobertson empty or non-existant03:51
Jedateach Ok, thanks, make sense03:51
SethRobertson++03:52
← Jedateach left03:53
→ beautifulmind joined03:54
→ hyperair joined03:54
→ legnaleurc_ joined03:54
← legnaleurc left03:54
→ dnjaramba_ joined03:57
← [M]ax left03:57
← dnjaramba left03:57
← diegoviola left04:00
→ [M]ax joined04:00
← dnjaramba_ left04:02
← n8o-mba left04:03
→ iband joined04:04
← lyddonb left04:04
← OOPMan left04:05
→ Fandekasp joined04:05
→ arvind_khadri joined04:06
← arvind_khadri left04:06
→ arvind_khadri joined04:06
→ dnjaramba joined04:06
← legnaleurc_ left04:07
← BiggFREE left04:07
→ theotherguy joined04:08
→ legnaleurc joined04:08
← theotherguy left04:08
→ linoj joined04:08
→ Blaster joined04:10
Blaster hey how would I browse the entire contents of a file from a past commit04:10
← plasticle left04:10
wereHamster Blaster: git show <commit>:<file>04:11
Blaster thanks04:11
← weiyang left04:12
Blaster so I have to use the MD5 hash for the <commit> of that command?04:12
in windows git bash I can't copy from the console04:12
wereHamster it's a sha1 hash. and take your windows problem to microsoft.04:13
chrisf Blaster: if you have some name for the commit, relative to a branch, perhaps -- you can use that04:14
← arvind_khadri left04:14
chrisf you actually *can* copy from the windows console04:14
Blaster I wish I could use Linux for everything!04:14
chrisf: how?04:14
my window says MINGW32 at the top04:14
← legnaleurc left04:14
chrisf turn on quick edit mode, and it's even not a huge pain in the ass04:15
← dnjaramba left04:16
Blaster oh ok thanks I think I got it now04:16
→ legnaleurc joined04:16
→ mikepack joined04:19
← hyperair left04:20
← legnaleurc left04:21
← kukks left04:23
→ legnaleurc joined04:24
→ weiyang joined04:27
→ ajt- joined04:28
← Blaster left04:28
→ Blaster joined04:28
← rtconner left04:29
← legnaleurc left04:29
← notola left04:29
← linoj left04:29
← [M]ax left04:29
→ [M]ax joined04:30
→ theotherguy joined04:31
→ gavin_huang joined04:31
→ legnaleurc joined04:31
← legnaleurc left04:36
→ legnaleurc_ joined04:37
← bgerber left04:38
→ h0bbit joined04:39
→ bgerber joined04:43
← legnaleurc_ left04:44
→ arvind_khadri joined04:45
← arvind_khadri left04:45
→ arvind_khadri joined04:45
→ hyperair joined04:46
→ legnaleurc joined04:52
→ notola joined04:52
→ ajaykools joined04:52
← notola left04:53
→ Z3Rt joined04:53
← simplechat left04:54
→ lyddonb joined04:54
CareBear\ click system icon, edit->copy04:55
mark by dragging while holding left mouse button depressed04:55
press enter to make copy and exit copy mode04:56
← Z3Rt left04:59
← neunon left04:59
→ harshpb joined05:02
→ jony joined05:02
← lyddonb left05:03
← eletuchy left05:05
← arvind_khadri left05:06
→ eletuchy joined05:08
→ dc5ala joined05:10
← [M]ax left05:10
← Fandekasp left05:10
→ legnaleurc_ joined05:10
← legnaleurc left05:10
→ fbsddom0 joined05:11
→ mxweas_ joined05:12
← ajaykools left05:13
iband → iband|away05:13
iband|away → iband05:14
→ psakrii joined05:14
← orospakr left05:15
← infogulch left05:16
→ legnaleurc joined05:17
← legnaleurc_ left05:17
→ theotherguy_ joined05:17
← mikepack left05:17
→ arvind_khadri joined05:18
← arvind_khadri left05:18
→ arvind_khadri joined05:18
→ [M]ax joined05:18
→ ploop joined05:18
← xiangfu left05:18
→ krz joined05:18
→ vogxn joined05:19
← theotherguy left05:19
theotherguy_ → theotherguy05:19
→ xiangfu_ joined05:20
→ advatar joined05:21
← edude03 left05:22
← eletuchy left05:22
← legnaleurc left05:23
→ ryanisnan joined05:23
← joshsmith left05:23
→ xsbeats joined05:24
→ legnaleurc joined05:25
← h0bbit left05:28
→ mikepack joined05:32
← Enoria left05:32
← harshpb left05:34
← legnaleurc left05:35
→ stkrzysiak joined05:37
← diogogmt left05:40
→ krzkrzkrz joined05:41
← krzkrzkrz left05:41
→ krzkrzkrz joined05:41
→ legnaleurc joined05:41
← nevyn_ left05:41
→ nevyn joined05:41
← lonewulf` left05:42
→ zweb joined05:43
← krz left05:43
masterhumper → sejo05:43
→ krz joined05:44
← krz left05:44
→ krz joined05:44
← wroathe left05:44
← krzkrzkrz left05:46
→ lonewulf` joined05:46
→ theotherguy_ joined05:48
iband → iband|away05:48
← legnaleurc left05:48
← stkrzysiak left05:49
→ legnaleurc joined05:50
← mxweas_ left05:50
← theotherguy left05:51
theotherguy_ → theotherguy05:51
→ Fandekasp joined05:52
← Bass10 left05:52
→ JohnBat26 joined05:52
← theotherguy left05:53
→ theotherguy joined05:54
← aristidesfl left05:55
jmcantrell are there any env vars that are set when running a script as a git extension?05:57
← legnaleurc left05:57
→ kadoban joined05:57
→ dorkmafia joined05:58
→ legnaleurc joined05:59
dorkmafia hi05:59
i just did a git pull origin and it says "Your branch is ahead of 'origin/pod' by 1950 commits"05:59
nevyn you've been busy?05:59
dorkmafia i haven't made any commits to this branch …05:59
nope05:59
how do i reset my branch to origin?06:00
→ theotherguy_ joined06:00
dorkmafia or see the out going commits06:00
so weird06:00
nevyn git reset --hard origin/pod06:00
but this is06:00
distructive06:00
destructive even06:00
→ Heimidal joined06:00
← bgerber left06:00
dorkmafia i haven't made any commits though06:01
← mikepack left06:02
nevyn so I'd make a reference branch of where your branch is now.. just in case06:02
→ RajeevGEC joined06:03
← gchristensen left06:03
→ bgerber joined06:03
← legnaleurc left06:03
nevyn but then I'm paranoid.06:03
← theotherguy left06:04
theotherguy_ → theotherguy06:04
← jony left06:04
wereHamster win 1906:04
dorkmafia well it doesn't seem to make sense to me cause i git reset —hard origin/pod and it resets me then i git pull origin red-pod and it downloads a bunch of stuff06:04
and says i'm ahead of the branch by 1950 commits06:04
→ legnaleurc joined06:04
→ gchristensen joined06:04
weiyang after I run git fetch, the file I have changed seem not be updated....06:05
→ mikepack joined06:05
nevyn dorkmafia: how many remotes do you have?06:05
weiyang: fetch doesn't touch the working directory06:05
CareBear\ dorkmafia : did you ever create a commit?06:05
nevyn weiyang: you may want pull (fetch & merge) or to use rebase depending on your workflow06:06
CareBear\ I like git pull --rebase06:06
dorkmafia carebear no i didn't06:06
→ stkrzysiak joined06:06
nevyn dorkmafia: did upstream get rebased?06:06
dorkmafia i have no idea how could i tell?06:07
← RajeevGEC left06:07
dorkmafia on github?06:07
← hyperair left06:07
CareBear\ seems the only explanation06:07
nevyn dorkmafia: who's github? yours or someone elses?06:07
CareBear\ nevyn : "whose"06:07
← mikepack left06:07
CareBear\ sorry, I'm off now06:07
weiyang nevyn, hmm.... ok, I have done the git fetch, than i need to git rebase? let me view the help of git rebase :)06:08
dorkmafia mine06:08
← Fretta left06:08
→ cehteh joined06:08
nevyn weiyang: it depends on your workflow06:08
dorkmafia i was asking how i could tell if there had been an upstream rebase06:09
nevyn dorkmafia: then that seems unlikely if you own the github.06:09
you'd know if you'd done a force update rebase of the github06:09
dorkmafia oh there are a lot of people on the team06:09
→ nixmaniack joined06:09
← gchristensen left06:10
weiyang nevyn, i am relatively new. I get the repo before, on the master branch, i did some commit, now i want to update the master branch06:10
dorkmafia so when i do a git reset—hard origin/pod it takes me back to feb 4 but when i do a git pull origin pod it takes me yesterday06:10
nevyn dorkmafia: don't use pull it's going to mess you up if you're working with many remotes. try git remote update06:11
← sattu94 left06:11
nevyn weiyang: it depends how you want history to look06:11
dorkmafia aight06:11
will that update all of my remotes?06:11
nevyn weiyang: if you want masterA => masterB => masterC => yourworkhere => (masterD and E merged) => HEAD then use merge06:12
→ hyperair joined06:13
nevyn if you want masterA => B => C => D => E => yourworkhere then use rebase06:13
dorkmafia: yep06:13
dorkmafia nice06:13
weiyang nevyn, oh... thanks :)06:13
dorkmafia is there a way to do just one?06:13
nevyn fetch06:13
dorkmafia: why would you want out of date remotes?06:14
← jds left06:14
dorkmafia yah true06:15
nevyn weiyang: be aware... this will rewrite the yourworkhere commit.06:15
→ jds joined06:15
← ajt- left06:15
nevyn because it's parent commit is now E not C06:15
dorkmafia just want to make sure that this branch gets updated06:15
weiyang nevyn, ok, you mean the rebase method will recommit it. while the merge method will not?06:16
→ caseymcg joined06:16
nevyn the merge method is more "true" but often less useful06:16
iband|away → iband06:17
← hyperair left06:17
← xiangfu_ left06:17
weiyang nevyn, ok, yes, you can not change the order of the upstream06:17
nevyn weiyang: well it makes commiting your work upstream harder.06:17
the rebase mechinism always has yourwork ontop of upstream06:17
dorkmafia so now that brings me to the next fun question i'm pretty sure i have an idea of how to rebase a series of commits but I was wondering if anyone had a good explanation of what's actually going on..06:18
weiyang nevyn, oh, got it.06:19
nevyn, so rebase is a better choice06:19
nevyn weiyang: sometimes06:19
← zweb left06:19
← legnaleurc left06:19
nevyn often even06:19
→ harshpb joined06:19
weiyang :)06:20
← glennpratt left06:20
nevyn and by often I mean frequently not a child with no parents06:21
→ ThomasLocke joined06:22
← igotnolegs left06:22
→ dorisabayon joined06:22
→ jony joined06:22
→ dr_win joined06:23
← Fandekasp left06:25
→ sattu94 joined06:25
← JohnBat26 left06:26
→ JohnBat26 joined06:26
→ airborn joined06:29
→ Xizor joined06:30
→ notjohn joined06:30
vogxn hello all - how do i tell which branch a certain SHA-1 belongs to? i did git show-branch c98721w389 but that showed me c98721w389 which doesn't makes sense because i don't have a branch with that name06:31
→ legnaleurc joined06:32
→ _iron joined06:32
wereHamster vogxn: git branch --contains c98721w38906:32
← beautifulmind left06:33
vogxn wereHamster: thanks, that returned nothingg. now i am further confused because I can't checkin because that commit didn't have a reviewer :)06:33
wereHamster git doesn't require reviewers06:34
bob2 and the first thing does make sense, since both c98721w389 and whocares/somebranch both identify branches06:34
vogxn wereHamster: that's right. but our repo has to contain a reviewed-by in its comment message without which it rejects pushes06:34
→ theotherguy_ joined06:34
wereHamster and it never refuses to commit unless there is nothing to commit or the pre-commit hook refuses06:34
dorkmafia let's say i have a branch experimental with commits A,B,C,D and I want to rebase commits B,C,D over to my-branch what's the best way to go about this?06:35
vogxn wereHamster: it must be a pre-commit hook in the remote repo. thanks again. git branch --contains was what i was looking for06:35
wereHamster there is no pre-commit hook in th remote06:35
there is pre-receive though06:35
commits are local06:36
← theotherguy left06:36
theotherguy_ → theotherguy06:36
vogxn wereHamster: i see, that makes sense then. thanks06:36
← sattu94 left06:38
← madewokherd left06:38
→ TrineLocke joined06:39
→ pfFredd joined06:40
← vervic left06:41
→ legnaleurc_ joined06:42
← legnaleurc left06:43
← dfr|mac_ left06:46
← pfFredd left06:46
dorkmafia so i just rebased but it looks like it didn't include my last commit :(06:47
or rather it didn't include the first commit i made in the series06:47
i guess i need to change it to the commit before my last change06:48
→ xiangfu joined06:49
→ legnaleurc joined06:49
dorkmafia is ~1 appropriate to use in this case?06:49
or is it risky?06:49
← legnaleurc_ left06:50
weiyang nevyn, hm... i run "git rebase master", but still not see the changes...06:51
nevyn, which step i may wrong?06:51
→ sattu94 joined06:51
weiyang richard@richard:~/git/qemu$ git rebase master06:51
Current branch master is up to date.06:51
→ legnaleurc_ joined06:53
← legnaleurc left06:54
← [M]ax left06:54
← acl_ left06:55
→ Spockz` joined06:57
← uu left06:57
→ marthinal joined06:57
← rcu left06:57
→ rickmasta joined06:57
→ jargon- joined06:58
← subbyyy left06:59
← advatar left06:59
→ [M]ax joined07:00
← [M]ax left07:00
← rickmast left07:00
→ legnaleurc joined07:00
→ rryter joined07:01
← legnaleurc_ left07:01
→ [M]ax joined07:02
← stkrzysiak left07:03
→ legnaleurc_ joined07:04
← legnaleurc left07:04
← kurtzhong left07:05
→ jutaro joined07:06
→ gmarselis joined07:08
← project2501b left07:09
← jutaro left07:09
← Spockz` left07:09
← bgerber left07:10
→ hyperair joined07:10
→ kurtzhong joined07:11
← ryanisnan left07:12
← theotherguy left07:12
← codewrangler left07:12
← Dreamer3 left07:13
→ ipalaus_ joined07:13
→ vdv joined07:15
→ cbreak-work joined07:16
← gmarselis left07:17
→ Dragooon joined07:17
← legnaleurc_ left07:17
→ pass joined07:17
← pass left07:18
→ papegaaij joined07:18
→ w09x joined07:19
← krz left07:19
→ legnaleurc joined07:19
→ jceb joined07:19
→ theotherguy joined07:21
← papegaaij left07:21
→ jakob joined07:21
→ papegaaij joined07:22
← theotherguy left07:23
→ flijten joined07:25
→ theotherguy joined07:25
→ pass joined07:26
← pass left07:26
← legnaleurc left07:26
→ legnaleurc joined07:28
→ skoop joined07:28
→ ryanisnan joined07:29
→ ajaykools joined07:32
← arvind_khadri left07:33
→ knilz joined07:33
→ Shusshu joined07:34
← Shusshu left07:34
← w09x left07:36
→ legnaleurc_ joined07:37
← legnaleurc left07:37
← sattu94 left07:38
→ Sitethief joined07:39
← TTSDA left07:39
→ netiru joined07:39
→ giallu joined07:41
→ raatiniemi joined07:41
← legnaleurc_ left07:41
→ legnaleurc joined07:42
→ TTSDA joined07:43
→ thiago joined07:45
← ipalaus_ left07:45
← adamm left07:45
→ glcrazy joined07:47
→ ph^ joined07:47
← ph^ left07:47
← knilz left07:48
← legnaleurc left07:48
← mishok13 left07:48
← glcrazy left07:49
→ teweWORK joined07:49
→ ph^ joined07:49
→ w09x joined07:50
→ sattu94 joined07:51
→ theotherguy_ joined07:51
← airborn left07:52
← tatsuya_o left07:54
← theotherguy left07:55
theotherguy_ → theotherguy07:55
← w09x left07:56
→ h0bbit joined07:56
→ theotherguy_ joined07:58
dorkmafia I have some commits that I cherrypicked and I accidentally picked some in the wrong order then pushed… is it possible to still rebase this branch to reorder the commits?07:58
← cooloney left07:59
← theotherguy left08:00
theotherguy_ → theotherguy08:00
→ rendar joined08:00
← lapistano left08:01
→ arvind_khadri joined08:01
← arvind_khadri left08:01
→ arvind_khadri joined08:01
mst yes, but you'll have to force push08:01
but that's ok provided nobody else pulled yet08:02
→ lapistano joined08:02
← Goplat left08:05
→ noyb joined08:06
→ mishok13 joined08:08
← ryanisnan left08:10
← vdv left08:10
→ krz joined08:10
← weiyang left08:11
→ gulden joined08:11
→ GeorgeSebastian joined08:11
← GeorgeSebastian left08:11
→ GeorgeSebastian joined08:11
dorkmafia ok08:12
what if they have?08:12
← justfielding left08:12
dorkmafia tell them to pull again?08:12
→ gnufied joined08:12
← gnufied left08:14
→ gnufied joined08:14
← herdingcat left08:14
← gnufied left08:15
→ psoo joined08:15
→ pro_metedor joined08:15
thiago no08:16
→ OOPMan joined08:16
thiago tell them to erase and clone again08:16
← js__ left08:18
→ brokenshakles joined08:18
→ tqmz joined08:18
dorkmafia ouch08:20
i can just revert and push08:20
right?08:20
thiago yes, git revert is fine08:21
← shergill left08:21
→ maleknet joined08:22
← tga left08:22
→ tga joined08:23
← tga left08:23
→ tga joined08:23
→ Dave^| joined08:24
→ advatar joined08:29
← JohnBat26 left08:30
← harshpb left08:32
→ chepre joined08:32
→ DrNick joined08:32
→ marvin_ joined08:33
← advatar left08:34
→ lb joined08:34
→ advatar joined08:34
→ gusnan joined08:34
→ JohnBat26 joined08:34
← advatar left08:35
→ johnhamelink joined08:37
→ codeshah joined08:37
→ harshpb joined08:37
→ milli joined08:38
→ stodan joined08:39
← notfunk left08:40
→ notfunk joined08:41
dorkmafia thiago: so can i revert a range of commits at once?08:43
→ ipalaus_ joined08:43
← QaDeS_ left08:43
thiago I think revert now accepts a range08:43
but if it doesn't, you can easily revert one at a time08:44
← kaimi left08:44
← sh4rm4 left08:45
dorkmafia git revert —hard SHA..SHA08:45
like that? but it gives me an error08:45
← lb left08:47
thiago no, not like that08:49
remove the --hard08:49
this is revert, not reset08:49
→ Salonce joined08:51
→ kaimi joined08:51
→ lb joined08:54
← olinux left08:55
→ shruggar joined08:55
pro_metedor how to return to stashed changes ?08:58
← [M]ax left08:58
shruggar pro_metedor: usually "git stash pop"08:59
pro_metedor thanks :)08:59
→ cooloney joined09:00
→ herdingcat joined09:01
← harshpb left09:01
← Heimidal left09:01
→ _Qman joined09:02
→ osa1 joined09:02
→ [M]ax joined09:02
osa1 is there a way to see diff of two commits? I have commit hashes09:02
→ Dreamer3 joined09:02
someprimetime i want to clone a repo from my server on my local machine, but i ssh with a port number generally how can i do this with git clone?09:03
git clone ssh://me@domain.com:PORT# ?09:03
← BlindOracle left09:04
← osa1 left09:04
← vogxn left09:04
someprimetime nvm yes09:05
thiago OOPMan: git show SHA109:07
OOPMan thiago: Wrong person09:07
thiago yes, sorry09:07
→ jatt joined09:07
thiago the person who asked has left09:07
← fayimora left09:08
← rue left09:08
→ fayimora joined09:09
→ beautifulmind joined09:09
→ Spockz joined09:09
→ eijk joined09:10
→ wadkar joined09:10
→ johnathanb joined09:10
→ FernandoBasso joined09:10
wadkar how do I see diff between two branches for a given file ?09:10
someprimetime so i just cloned my directory from my server into a folder on my local client machine which seemed to work fine (pulled all the contents from that repo), but when i go git branch i get: fatal: Not a git repository (or any of the parent directories): .git09:10
do i need to git init within the cloned directory?09:10
charon wadkar: git diff branch1 branch2 -- filename09:10
wadkar charon: sweet ! thanks09:11
→ whitman joined09:11
charon wadkar: or perhaps more usefully, gitk branch1...branch2 -- filename09:11
wadkar whats the difference between spaces and three dots ?09:11
charon (yes, that's something entirely different, i just wanted to point it out)09:11
thiago someprimetime: you need to cd into the directory that clone created09:11
wadkar: the space in git diff is equivalent to two dots09:11
someprimetime thiago: in it now09:11
thiago wadkar: A..B09:11
charon wadkar: the second one graphically looks at the difference in history between the branches, as pertaining to changes to filename09:11
thiago wadkar: A...B is A B --not $(git merge-base A B)09:11
someprimetime thiago: woops09:11
no i wasn't09:12
thanks09:12
← VladGh left09:12
someprimetime cool09:12
→ theotherguy_ joined09:12
charon thiago: except sadly in git-diff it's special-cased anyway ;)09:12
thiago charon: right09:12
someprimetime: ^^^^09:12
wadkar thiago: ohhk, thanks. I may not understand it completely but I think I get what you are trying to say09:12
→ LimeBlast joined09:12
wadkar I will go with gitk version09:13
← maletor left09:13
← theotherguy left09:14
theotherguy_ → theotherguy09:14
wadkar errm, does it matter which branch i name first ? i mean git diff 0.1 0.2 -- foo.sql09:14
charon wadkar: they're really special-cased. 'diff A..B' is the same as 'diff A B'. 'diff A...B' is the same as 'diff M B', where M is the "fork point" where B started diverging from A (= one merge-base of)09:14
someprimetime thiago: ok i did my first commit/push from my local machine on the repo that i cloned. so i just pushed it to my server.. now how do i know that my server received my changes?09:15
charon wadkar: however, in all other commands (log, gitk, etc) A..B means "everything contained in B that is not contained in A", whereas A...B means "everything contained in A or B but not in both"09:15
someprimetime sorry didnt' mean to ask you specifically this question in here09:15
i was gonna say thanks then just started adding to the message09:15
thiago someprimetime: well, git log origin/branchname should show them09:15
→ PigeonFriend joined09:15
thiago someprimetime: if git push said it was accepted without error, then it was accepted09:15
→ truzie joined09:16
charon wadkar: and yes, 'diff A B' gives the "opposite" diff of 'diff B A' :)09:16
thiago someprimetime: if you REALLY want to be sure, git ls-remote origin, then check that the SHA-1 you wanted to push shows up.09:16
someprimetime ah ok that'll do i think thanks :)09:16
this is cool.09:16
thiago someprimetime: note this only works if no one else submits another change right afterwards09:16
wadkar charon: aha ! understood, thanks for the simple explanation09:16
→ BlindOracle joined09:17
→ CombatCow joined09:17
wadkar charon: wow, I never knew git is so specific about what a simple extra dot means, the order of arguments .. phew , but after reading your explanation for 3rd time (and trying out few git diff/log commands) i get it now09:19
thanks charon , thiago for your quick help ! I will get back to work now09:20
later :)09:20
charon wadkar: glad i could help. i also made !dots for the next unsuspecting user09:20
gitinfo wadkar: In the log family (git-log, gitk, etc.) A..B means "everything in B but not in A" [formally: ^A B] and A...B means "everything in A or B but not in both" [formally: A B --not $(git merge-base A B)]. An empty "side" of the dots implies HEAD, so 'git log master..' is very different from 'git log master'!09:20
← milli left09:20
charon ah, --all was missing from the merge-base. fixed09:21
→ overbythere joined09:21
wadkar lol, thats handy !dots to remind myself of the simple explanation ... sweet !09:21
gitinfo In the log family (git-log, gitk, etc.) A..B means "everything in B but not in A" [formally: ^A B] and A...B means "everything in A or B but not in both" [formally: A B --not $(git merge-base --all A B)]. An empty "side" of the dots implies HEAD, so 'git log master..' is very different from 'git log master'!09:21
→ nax joined09:23
← nax left09:23
→ dpino joined09:24
← caseymcg left09:24
← thiago left09:24
→ milli joined09:25
→ Pugnax_ joined09:27
→ Tommy[D] joined09:28
→ Element9_work joined09:28
→ kpreid joined09:30
→ soulcake_ joined09:31
← Balrog left09:31
→ drizzd joined09:31
→ Pugnax__ joined09:32
→ flavius joined09:32
← Pugnax__ left09:32
→ toabctl joined09:32
→ Pugnax__ joined09:33
← Pugnax_ left09:35
→ achadwick joined09:37
→ akiress joined09:42
→ d3v1an7 joined09:42
← d3v1an7 left09:42
← noyb left09:43
→ ben1066 joined09:43
← DofLeFou left09:44
→ tatsuya_o joined09:44
← tatsuya_o left09:44
→ tatsuya_o joined09:45
→ dr_lepper joined09:45
← ben1066_ left09:46
→ eijk_ joined09:47
← eijk left09:48
someprimetime so it looks like i'm successfully able to create files from my server's repo and pull them onto my cloned local branch (master), but how do i do it the other way around?09:49
i created a new file locally in my master branch, committed it, then when i `git push` and then go to my server and do a `git pull` i get: fatal: 'origin' does not appear to be a git repository09:50
bob2 why would you run pull there?09:51
← shruggar left09:51
→ theotherguy_ joined09:52
bob2 and i suspect you don't understand why pushing to non-bare repos is discouraged09:52
→ Cromulent joined09:52
← tatsuya_o left09:53
someprimetime bob2: not quite sure total noob just trying to figure out how i can start editing files locally and merging them into master09:53
→ tatsuya_o joined09:53
bob2 I don't think that's what you're asking for at all09:53
← oc80z left09:53
→ oc80z joined09:53
← oc80z left09:53
bob2 short answer is if you want a checked out copy of your code on a server, you'll want to push to a bare repo09:54
and have that bare repo, in its' post-receive hook, update the checkec-out non-bare repo09:54
e.g. bu rynning 'cd /somedir ; git fetch ; git reset --hard origin/master'09:54
← theotherguy left09:54
theotherguy_ → theotherguy09:54
→ oc80z joined09:55
someprimetime oh, so will this allow me to have final say (on the server) of whatever i push locally can be eventually merged in with master?09:55
or am i just confused (i think the latter)09:55
bob2 very confused09:55
someprimetime :\09:55
what i don't understand is this: i make have my own local branch, i make changes, commit, push ... where the heck does that go? like how can i find out if it's been received on the server?09:56
bob2 your goal is simply to edit code on your desktop, and have things works so that when you push that code, it ends up updating some files on disk on a server?09:56
you're imaginging that things are much much more complicated than they are09:57
someprimetime bob2: my goal is to be able to locally create a bunch of different branches, push that code to my server.. be able to check if i want that code that local code is good enough to go live on my site09:57
bob2 when you have 'master' checked out locally, 'git push', by default, will push changes from /your local branch called 'master'/ to the remote branch called 'master'09:57
that's it09:57
it doesn't update files on disk09:57
it doesn't do any magic09:57
→ shruggar joined09:57
→ crooksey joined09:58
bob2 (er it obviously updates files in the git repo, but it doesn't update your source or whatever)09:58
someprimetime oh shit really?09:58
ah ok09:58
bob2 but you definitely didn't read the faq or a book09:58
since they all tell you to not push to a non-bare repo09:58
since it will confuse you09:58
someprimetime ok i'll do that now09:58
bob2 anyway09:59
delete code from server09:59
make a bare repo on server09:59
← hyperair left09:59
someprimetime does the directory of the bare repo matter? or just put it anywhere09:59
bob2 somewhere you can write to that isn't toally idiotic09:59
someprimetime ok09:59
bob2 /srv/git/whocares.git is fine09:59
or ~/repos/whoares.git09:59
etc09:59
you will find things far less confusing if give bare repos a .git suffix10:00
someprimetime ok10:00
bob2 then get rid of all the other stuff on the server10:00
then do 'git clone /srv/git/whocares.git/' somewhere on the server10:01
→ thiago joined10:01
→ Davey_ joined10:01
→ mulletro1 joined10:04
→ zdk_ joined10:04
→ PerlPilot joined10:04
→ tychoish joined10:04
← thiago left10:04
← oc80z left10:04
← PigeonFriend left10:04
← herdingcat left10:04
← gusnan left10:04
← tqmz left10:04
← jceb left10:04
← Xizor left10:04
← dorisabayon left10:04
← orafu left10:04
← zdk left10:04
← hwrdprkns left10:04
← tazle_ left10:04
← MissionCritical left10:04
← Noxville left10:04
← tychoish_ left10:04
← eli left10:04
← zorzar left10:04
← PerlJam left10:04
← jimi_hendrix left10:04
→ ypcat joined10:04
→ tazle__ joined10:04
zdk_ → zdk10:04
→ shennyg_ joined10:04
→ Fandekasp joined10:04
→ ouah joined10:04
← rryter left10:04
→ hyperboreean joined10:04
→ hwrdprkns joined10:04
→ nomadtacos joined10:05
→ unnamedwill joined10:05
→ zorzar joined10:05
← dorkmafia left10:05
ouah with new git version, git difftool run vimdiff in read only mode10:05
([diff] is set to "tool = vimdiff" in .gitconfig)10:05
→ PigeonFriend joined10:05
→ sattu941 joined10:05
ouah how can I change that so my files are writeable?10:05
← sattu941 left10:06
← johnathanb left10:06
→ mastro joined10:06
→ DangerousDave joined10:06
→ retoryter joined10:06
→ crash\ joined10:07
→ bgerber joined10:07
← retoryter left10:08
→ johnathanb joined10:08
→ orafu joined10:09
→ tqmz joined10:09
→ speakman joined10:09
→ przemoc joined10:09
→ NickASloan joined10:09
← xiangfu left10:09
→ Noxville joined10:09
← drpaul left10:09
→ oc80z joined10:09
← sawjig left10:09
→ thiago joined10:10
← LimeBlast left10:10
→ albel727 joined10:11
→ kurtzhong_ joined10:11
→ retoryter joined10:11
→ sawjig joined10:11
→ Xizor joined10:12
← kurtzhong left10:12
kurtzhong_ → kurtzhong10:12
→ gusnan joined10:12
→ jimi_hendrix joined10:13
→ Fissure joined10:13
→ heinrich5991 joined10:13
→ eli joined10:13
← netiru left10:15
← retoryter left10:16
→ netiru joined10:16
→ jave joined10:16
→ Matrixiumn joined10:17
← boombatower left10:17
→ senny joined10:17
→ pfig joined10:17
→ arietis joined10:17
→ retoryter joined10:17
→ Mission-Critical joined10:17
→ jceb joined10:18
→ jstemmer joined10:18
← jstemmer left10:18
Mission-Critical → MissionCritical10:20
← codeshah left10:20
→ herdingcat joined10:21
→ iocor joined10:21
← DrNick left10:23
→ jstemmer joined10:24
→ valtih joined10:25
← herdingcat left10:26
valtih why typing .classpath into the .gitignore does not have any effect? How do I stop the file being managed by git?10:27
FauxFaux valtih: .gitignore doesn't affect tracked files. See man git rm --cached.10:27
gitinfo valtih: the git-rm manpage is available at http://jk.gs/git-rm.html10:27
← theotherguy left10:28
→ theotherguy joined10:28
→ EyesIsServer joined10:29
valtih Actually, I do not need to remove the file from either workspace or repository. I just want to ignore it. Why gitignore works for one file but not the other?10:29
FauxFaux Perhaps you should address my statement instead of re-stating yours.10:30
qwertz_ valtih, did you add and commit .gitignore?10:31
cmn qwertz_: that's not the issue10:31
qwertz_ oh, i see. FauxFaux is right10:31
← Fandekasp left10:32
qwertz_ valtih, git rm --cached fileName removes the file from git tracking, however, it leaves a copy in your working directory10:32
→ cortana2 joined10:33
valtih ok, I'll try it10:33
Anyway I would like to know why not-tracked files can be gitignored but tracked can not10:34
cmn because that's the saner variant10:34
you either want to track files or you dont'10:34
← mastro left10:35
← CombatCow left10:35
valtih I see that git rm --cached is going to commit a file delete. I do not want to delete it from reporsitory.10:36
bob2 if you think about it for a bit, you'll probably agree that having it operate on anything other than untracked files would be super confusing10:36
← bgerber left10:36
qwertz_ valtih, what do you want then?10:36
→ na3r joined10:37
valtih I want to have a copy of file different from repository and git do not affect it10:37
← GeorgeSebastian left10:38
← Pugnax__ left10:38
→ khmarbaise joined10:38
cmn !config10:38
gitinfo [!configfiles] the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.10:38
cmn or just overwrite the file and don't commit it10:38
→ tcurdt joined10:39
valtih How !config is different from gitignore and how can I keep file not commited if this prefents any git operations (like checking out different branches)?10:40
gitinfo [!configfiles] the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.10:40
→ ajt joined10:40
qwertz_ valtih, just stash the change before checking out a new branch - then apply the stash again10:41
valtih that is interesting, thanks.10:41
→ bgerber joined10:42
valtih Though, I would like to know a less workaround method10:42
cmn what do you actually want to achieve?10:42
← ploop left10:42
FauxFaux The thing we've explicitly stated isn't possible without a workaround, duh.10:43
← shibayu36 left10:43
← Cromulent left10:46
→ Pugnax_ joined10:46
→ d0k joined10:46
← cooloney left10:50
→ xjonc joined10:51
→ sh4rm4 joined10:52
← _Vi left10:52
← [M]ax left10:55
← toabctl left10:55
← johnhamelink left10:55
→ [M]ax joined10:58
→ johnhamelink joined10:58
→ mneorr joined10:59
cmn anything other than "track file: y/n" is going to be a workaround11:00
→ dr_win_ joined11:02
as3llus_ → as3llus11:02
qwertz_ valith, maybe you can create your branch (that you don't publish) and merge the branch you want to track. if you want to publish your changes, you can rebase your local branch onto the master and leave out the commit that changes your file11:03
← zdk left11:03
Nickeeh I killed a git checkout, and now git bisect reset nor checkout wants to complete anymore since I have additional untracked files which weren't untracked. :/11:04
→ ASIRINeiluj joined11:05
Nickeeh (and thus will be overwritten)11:05
← tga left11:05
← Tureos left11:05
← dr_win left11:05
cmn what do you mean killed and what would you like to do?11:06
Nickeeh By killed I mean I terminated the process while it was busy11:07
and now I'm left with "untracked files" which weren't untracked before.11:07
← someprimetime left11:08
cmn what process?11:09
→ prudnikov joined11:09
Nickeeh cmn, git checkout xD11:10
cmn it sounds like you stopped the index from getting updated, so HEAD and the index disagree11:10
← rickmasta left11:10
← posciak2 left11:10
Nickeeh mmmokay11:11
→ toabctl joined11:11
Nickeeh How can I fix it?11:12
cmn that depends on what the fix is11:12
what would you like to do? how should the worktree/index/branch look like?11:12
Nickeeh Like some other commit I'd like to hceckout to11:12
cmn then you can use -f11:12
Nickeeh git checkout -f ?11:13
cmn to tell git to overwrite those untracked files, since they're not really untracked11:13
Nickeeh aaah11:13
great, thanks :)11:13
cmn but do be careful when using -f with anything11:13
Nickeeh of course11:13
→ boomshanka joined11:15
→ hyperair joined11:17
← boomshanka left11:17
→ chuck^norris joined11:18
← fruight left11:18
chuck^norris hello everyone, and please excuse my noobishness :) i'm playing around with git, i cloned a remote repository, did some local changes, and pushed back. git status on the remote site shows the changes, but how'd i actually coommit them please?11:18
← Nathandim left11:18
→ Tureos joined11:19
→ DarkAR joined11:19
→ fruight joined11:19
← jony left11:20
cmn you commit with git commit11:21
however, it sounds like you want to push to a non-bare repo, which isn't always that good an iea11:21
have you read the !book?11:21
gitinfo There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !vcbe and !parable11:21
chuck^norris cmn: i have to admit i haven't, reading now, thanks11:22
→ posciak2 joined11:23
→ twisted` joined11:23
→ legnaleurc joined11:26
← chuck^norris left11:26
← fruight left11:27
→ mklappstuhl joined11:30
← legnaleurc left11:30
← kurtzhong left11:32
→ kurtzhong joined11:32
← wadkar left11:32
← rchavik left11:32
← Pugnax_ left11:33
← [M]ax left11:34
← GeertB left11:34
→ [M]ax joined11:35
→ Pugnax_ joined11:35
→ kurtzhong_ joined11:35
← kurtzhong left11:36
kurtzhong_ → kurtzhong11:36
→ tga joined11:37
← tga left11:37
→ tga joined11:37
→ angelsl joined11:37
← ajaykools left11:38
← OOPMan left11:38
← arvind_khadri left11:38
→ Talvino joined11:38
→ na3r_ joined11:39
→ Black joined11:41
← gavin_huang left11:42
→ gavin_huang joined11:42
pom Can you help me with github here?11:42
EugeneKay !ask11:42
gitinfo yes, it's okay to ask questions here... in fact, you just asked one! ;) Pretty much any question is fine. We're not terribly picky.11:42
← na3r left11:42
→ Lgb joined11:43
EugeneKay Though if we don't know we'll point you to #github11:43
pom I forked someones project and now it's pretty much a new project based on theirs. It's not an in-place replacement. Can I rename it or should I start a new repo? I though it'd be nice to have the history left.11:44
Sorry, I'm a retard apparently. I found it.11:44
EugeneKay The "fork" convention only deals with Github's Pull Request stuff.11:45
FauxFaux pom: Seems like it'd make sense to give it a new name, so long as you give credit to the old project?11:45
EugeneKay You can delete you repo + recreate it, pushing the same history back up11:46
FauxFaux Iirc github is fine with renaming repos as many times as you want.11:47
pom There was a rename feature actually. Very simple. It's so hard when you can't manpage it. ;)11:47
FauxFaux Iirc it was even fine with me fixing the case on one of my projects. I was pretty impressed.11:47
→ legnaleurc joined11:48
→ artib joined11:50
→ arvind_khadri joined11:50
← psakrii left11:51
→ ajaykools joined11:51
← artib left11:53
→ artib joined11:53
← legnaleurc left11:53
→ bgy joined11:54
← ajt left11:55
→ ajt joined11:56
← lapistano left11:58
→ lapistano joined11:58
← nixmaniack left11:59
→ Dave^|| joined11:59
→ legnaleurc joined12:01
→ na3r__ joined12:01
← Dave^| left12:02
← na3r_ left12:02
← Dragooon left12:03
→ brian_g joined12:04
← legnaleurc left12:05
← kurtzhong left12:05
→ ph^_ joined12:06
← krz left12:06
→ fruight joined12:07
← keller left12:07
→ akosikeno joined12:08
← artib left12:09
← ph^ left12:09
← stodan left12:09
→ artib joined12:10
→ stodan joined12:11
→ legnaleurc joined12:11
→ fayimora_ joined12:11
← fayimora_ left12:11
→ fayimora_ joined12:12
← Element9_work left12:12
→ artib1 joined12:13
← artib left12:13
→ Element9_work joined12:13
← fruight left12:13
→ amigojapan joined12:13
← fayimora left12:14
fayimora_ → fayimora12:14
→ keller joined12:16
← smw_ left12:16
← artib1 left12:21
→ artib joined12:21
→ ph^ joined12:23
→ jameslord joined12:24
jameslord anybody here using svn?12:24
cmn if you have problems with svn, you should ask #svn12:25
← ChaoWang left12:25
← lapistano left12:25
→ lapistano joined12:26
← ph^_ left12:27
← artib left12:27
→ erichynds joined12:27
→ martinjlowm joined12:28
→ stamina joined12:29
→ uu joined12:30
→ theotherguy_ joined12:32
jameslord cmn, dude, i just want to use svn and git both:)12:33
→ lucas joined12:33
jameslord cmn but i am not sure if they can cooperate well：)12:33
cmn do you have a question about git?12:33
check man git svn12:33
gitinfo the git-svn manpage is available at http://jk.gs/git-svn.html12:33
jameslord cmn dude, yes, i know git svn, the problem is that i want to use svn , git and eclipse at the same time:)12:34
→ OOPMan joined12:34
jameslord cmn and i may switch to different platforms oftenly, sometimes i may use vim:)12:35
← theotherguy left12:35
theotherguy_ → theotherguy12:35
← toabctl left12:35
→ Pugnax__ joined12:35
cmn you editor is irrelevant12:35
canton7 I don't know if EGit can handle git-svn. I suspect not. Just use command-line git and you'll be fine12:35
jameslord EGit is bull shit:(12:36
cmn you can use git-svn and svn if you're careful12:36
canton7 well, I guess you can use EGit for commits and command-line git for dcommits, but hey12:36
jameslord i choose svn because subeclipse is nice tool:)12:36
← tga left12:36
cmn then use svn12:36
← ph^ left12:37
jameslord canton7 no, i just hate egit, it is bull shit:(12:37
canton7 then just use command-line git, jameslord12:37
jameslord cmn i want to have the power of git at the same time:)12:37
cmn then don't use svn12:37
← lucas left12:37
cmn it's your workflow, you get to decide12:38
→ ph^ joined12:38
cmn if you want the power of git, use git12:38
← Pugnax_ left12:38
jameslord cmn can i use svn and git at the same time12:38
cmn I've answered this already12:38
jameslord cmn for example, i can commit using svn or git12:38
→ advatar joined12:38
canton7 jameslord, the only way to tie git and svn together is to have a local git repo that talk to a remote svn repo12:39
→ lucas joined12:39
cmn or use subgit12:39
jameslord cmn dude, you told me git-svn, no svn:(12:39
→ bitkiller joined12:39
cmn exactly12:39
git-svn makes git interoperate with a svn repo12:39
and AFAIK subgit is the only thing that makes svn interoperate with a git repo, kinda12:40
jameslord cmn but i want to use subeclipse:(12:40
canton7 jameslord, you can not use a subversion tool on a git repo12:41
cmn you either want to use git or subeclipse12:41
jameslord cmn subeclipse is powerful:)12:41
cmn then use that, if it has the power you want12:41
jameslord cmn no, dude, i want to use git and subeclilpse at the same time:)12:41
cmn but you just implied it's not powerful enough for you12:41
canton7 jameslord, subeclipse talks to subversion. If you want to use subeclipse, you can't use git. If you want to use git, you can't use subeclipse12:41
cmn then use subgit, as I've said severl times12:41
← chepre left12:41
jameslord canton7 very sad to hear that:(12:42
← macabre^ left12:42
← mneorr left12:42
cmn each tool operates with itself, is that so hard to understand?12:43
take a look at subgit if you want to use svn and git to talk to the same repo12:43
canton7 jameslord, it's like... I can't use use a media player to read my email.12:43
jameslord cmn subgit is not so nice:( i must have a subgit server:(12:43
cmn jameslord: exactly12:43
← lucas left12:43
cmn for the reasons we've outlined above12:44
you need a http server to serve web pages, your mail server won't do it12:44
→ lucas joined12:45
← tatsuya_o left12:45
→ tatsuya_o joined12:47
→ iamjarvo joined12:47
→ Z3Rt joined12:49
jameslord hell ride:)12:50
nice12:50
← bitkiller left12:50
← ajt left12:51
← arietis left12:51
jameslord if you with git-svn, which style of usage will you choose12:52
git svn clone xxx/svn or git svn clone xxx/svn/trunk12:52
→ ajt joined12:52
jameslord which one is much attractive:)12:52
→ chepre joined12:52
canton7 jameslord, if you pass -t, -T and -b correctly, then it makes no difference12:53
jameslord and why?12:53
canton7 dude, what do you means?12:53
canton7 jameslord, read man git-svn. When you call 'git svn clone', you tell git-svn what folders you trunk, tags, and branches are in12:53
gitinfo jameslord: the git-svn manpage is available at http://jk.gs/git-svn.html12:53
← dhruvasagar left12:54
canton7 jameslord, git-svn is smart enough to notice when the URL you've given it ends in your trunk folder, and it removes that folder from the url12:54
→ dhruvasagar joined12:55
← iocor left12:56
→ bryanl joined12:56
→ Desproges joined12:57
→ theotherguy_ joined12:58
jameslord hell12:59
← Pugnax__ left12:59
→ toabctl joined12:59
jameslord i can init git repo using svn repo, that is git svn clone svn:// , can i init a svn repo using git repo? how to do this?13:00
→ twisted`_ joined13:00
cmn that's how you do it13:00
← hyperair left13:00
← theotherguy left13:00
theotherguy_ → theotherguy13:00
jameslord cmn dude, how to init a svn repo using git? svn git checkout git:// ?13:00
cmn you can't13:01
← twisted` left13:01
twisted`_ → twisted`13:01
← DangerousDave left13:01
cmn and you just said that's the opposite of what you wanted to do13:02
→ fayimora_ joined13:02
← fayimora_ left13:02
cmn and do read the manual pages, it's faster and more informative than asking how to use the tool every five minutes13:02
← ouah left13:03
jameslord cmn obviously i can init a git repo from svn repo:) but how to init a svn repo from git?13:03
cmn did you even read what I wrote?13:03
jameslord cmn yes, dude:)13:03
canton7 then which part of "you can't" didn't you understand?13:04
→ glcrazy joined13:04
cmn probably the same part of "don't call me dude" the last hundred times13:04
jameslord canton7 can i init a svn repo from git repo:)13:04
→ roland joined13:04
canton7 jameslord, no. no you can't. just as cmn said13:04
jameslord canon7 sorry to hear this, it is not fair:(13:05
← lucas left13:06
← theotherguy left13:06
canton7 jameslord, no-one has written such a tool, because no-one has a need to. If you need one, you should write it13:06
→ lucas joined13:06
← fayimora left13:06
← iband left13:07
cmn fair?13:07
if you're unhappy with the performance of svn, you should go tell them13:07
it has nothing to do with us13:08
← marvin_ left13:08
cbreak-work jameslord: you don't have to init svn repositories13:09
→ marvin_ joined13:09
cbreak-work the sysadmin/svnadmin will do that already13:09
← dv310p3r left13:10
← ajaykools left13:10
cmn if what you're trying to do is push your work back to the svn repo, that's what git-svn does13:11
shruggar if what you're trying to do is use an svn-like interface to operate on a git repository, I want you to seek your local priest for an exorcism :)13:12
→ chrisdev joined13:13
→ Know1edge joined13:13
← advatar left13:14
← iamjarvo left13:17
← JEEB left13:17
← jameslord left13:18
← maleknet left13:18
→ pandeiro joined13:18
→ maleknet joined13:19
→ dv310p3r joined13:19
→ jameslord joined13:19
jameslord why i cannot use git svn?13:20
http://paste.ubuntu.com/849942/13:20
→ alberto56 joined13:20
mika because you forgot to install git-svn?13:21
cmn you need to install it13:21
→ Pugnax_ joined13:21
← Black left13:24
← Juri_Agarin left13:25
→ Juri_Agarin joined13:25
cbreak-work jameslord: did you install the normal git?13:27
or some gimped precompiled binary?13:27
← jameslord left13:27
→ cakehero joined13:28
→ fayimora joined13:29
cmn the usual method of getting git is through the distribution13:30
→ amigojapan_ joined13:32
← amigojapan left13:33
→ Zet joined13:34
← Know1edge left13:34
ChanServ set mode: +v13:35
Zet I made some changes, haven't committed anything, but decided the changes are stupid. How do I get back to the latest commit?13:35
→ wroathe joined13:36
→ subbyyy joined13:36
cmn checkout -f13:36
← dv310p3r left13:36
← tatsuya_o left13:36
shruggar I recommend "git stash"13:37
cmn or checkout -- .13:37
→ tatsuya_o joined13:37
cmn stash would work13:37
→ Know1edge joined13:37
shruggar reversible tends to be a good option13:37
← Know1edge left13:37
→ Know1edge joined13:37
→ v0n joined13:37
shruggar git add . && git commit -nm 'Well, that was a stupid idea…' && git reset --keep HEAD^ :)13:38
← Know1edge left13:38
← chrisdev left13:39
shruggar the problem with "git stash" is, every now and then I type "git stash list" and I think "wtf? what the hell was I doing then?!" so, to me, it's like a reversible delete and nothing else :)13:39
→ Bass10 joined13:39
Zet I guess checkout -f is what I want13:40
the man page even says "This is used to throw away local changes."13:40
→ sdeboey joined13:40
→ QaDeS joined13:40
canton7 as usual, there are many ways of achieving the same end-result13:41
← subbyyy left13:41
← fayimora left13:42
shruggar There is more than one way to do it, though usually only one applies logically to your specific situation13:42
← tatsuya_o left13:44
→ glennpratt joined13:44
← marvin_ left13:44
→ fayimora joined13:45
→ tatsuya_o joined13:45
→ Jackneill joined13:45
← Jackneill left13:45
→ Jackneill joined13:45
→ Brammm joined13:45
→ alexgordon joined13:46
→ drupol joined13:47
Brammm Hi everyone, completely new to Git, but wanted to teach myself something new. I'm reading Pro git, but I just want to check if I'm still up to speed (just finished chapter 1): If I make a repository on my remote server, I do the git init thing there. Then if I want to work on that project locally, I do the git clone thing on my machine. Riught?13:47
→ cabaire joined13:48
→ rgsteele joined13:48
bremner correct13:49
→ Tommy[D]_ joined13:49
→ indstry joined13:50
← [M]ax left13:50
cmn that's one of the basic principles, not bound to get outdated13:50
← mulletro1 left13:50
drupol People using Gerrit around ?13:50
FauxFaux drupol: !ask13:51
gitinfo drupol: yes, it's okay to ask questions here... in fact, you just asked one! ;) Pretty much any question is fine. We're not terribly picky.13:51
→ Cromulent joined13:51
drupol Is it possible to create a group of users who can bypass the Gerrit review process and push directly to the branch ?13:51
→ [M]ax joined13:51
← cabaire left13:52
→ stuntmachine joined13:52
Brammm okay, thanks bremner and cmn :p I have some really outdated SVN knowledge, but that's it. Didn't find that in any other tutorials i've read so far. Liking the pro git book though13:52
→ cabaire joined13:52
← Tommy[D] left13:52
← OOPMan left13:54
→ advatar joined13:55
← EricInBNE left13:56
→ Dragooon joined13:56
→ cbeams joined13:56
→ macabre joined13:57
→ corenet_ joined13:58
← corenet_ left13:59
← Dragooon left14:00
← Jay_Levitt left14:00
→ crashanddie joined14:01
→ cooloney joined14:01
ChanServ set mode: +v14:02
cbeams I have a 'next' branch and a 'maint' branch, where directories have been renamed in 'next', but remain the same in 'maint'. My hope was to be able to perform backport cherry-picking from next to maint and then use `git cherry` to see which commits have not yet been backported, as people usually do. The trouble is, that `git cherry` relies on the output of `git patch-id` to determine whether an equivalent commit exists upstream. It app14:02
→ lyddonb joined14:02
→ johnkpaul joined14:03
→ Fandekasp joined14:03
shruggar cbeams: we lost you at "equivalent commit exists upstream. It app"14:04
cbeams oops. splitting up now.14:04
← Chib left14:04
→ krz joined14:05
cbeams I have a 'next' branch and a 'maint' branch, where directories have been renamed in 'next', but remain the same in 'maint'. My hope was to be able to perform backport cherry-picking from next to maint and then use `git cherry` to see which commits have not yet been backported, as people usually do. ...14:05
The trouble is, that `git cherry` relies on the output of `git patch-id` to determine whether an equivalent commit exists upstream. It appears that `git patch-id` does not consider renamed paths in any way. Per the manpage it says that patch-id calculates the SHA1 of a patch, ignoring whitespace and line numbers. ...14:05
→ codewrangler joined14:05
shruggar I'm guessing your problem is: the cherry-picked commits have merged into different files, so the patch-id is different?14:05
cbeams It seems that what I need is a switch to patch-id that ignores paths as well (or that more intelligently recognizes renamed paths). Anyone else ran into this? Any other approaches that I might consider other than `git cherry` for detecting which commits have not yet been backported from next to maint? Thanks, and I'm on 1.7.5.4, btw. EOM14:05
→ fisted joined14:05
cbeams shruggar: that's correct.14:05
→ Sigma joined14:05
cbeams particularly, into the same files within renamed directories14:06
so just renamed paths, but a rename is a rename all the same I suppose.14:06
← fisted_ left14:06
charon cbeams: any specific reasons why you are using a cherry-pick workflow instead of a topics-from-maint workflow?14:07
(or perhaps that decision is not even up for discussion)14:07
← indstry left14:07
← valtih left14:08
→ hyperair joined14:08
← hyperair left14:08
← groovehunter left14:09
→ olinux joined14:10
cbeams charon: we're basically working on a simplest-possible-approach-that-could-work workflow, where all development is happening on 'next' and as necessary, we backport fixes to 'maint' in order to roll maintenance releases. This allows committers to think only about 'next' and a single 'backporter' (project lead) to periodically perform backports batch-style. That's why it's necessary to be able to easily see what has and has not yet been14:10
← advatar left14:10
← amigojapan_ left14:10
charon has not yet been cut off because of line length restrictions? ;)14:10
cbeams where did I get cut off?14:10
charon "has not yet been"14:10
cbeams That's why it's necessary to be able to easily see what has and has not yet been backported. I'm open to a different workflow, but it needs to be as simple as possible. (some folks on the team including the project lead have weak git-fu). EOM14:11
charon i guess it depends on how long you expect the rename divergence to last (IOW, when maint will pick up the rename)14:11
→ amigojapan joined14:11
→ subh joined14:11
→ vandal- joined14:12
→ BiggFREE joined14:12
→ nikuyoshi joined14:12
→ dangerousdave joined14:13
cbeams 'maint' is actually a branch called '3.1.x'. 'next' is actually 'master', currently representing 3.2.x development. 3.1.x will never pick up the rename.14:13
When 3.2.x goes GA, a '3.2.x' branch will be spun off, at which point master will represent 3.3.x. At that point backports will not be a problem, but that will be about a year from now :(14:13
→ nixmaniack joined14:13
cbeams is it reasonable to expect here that `git cherry` (and therefore `patch-id`) actually pick up some awareness of renamed paths?14:14
← sdeboey left14:14
cbeams or at least provide the option to be path-ignorant completely?14:14
it feels like a hole in git's otherwise pretty amazing rename-awareness.14:14
→ herdingcat joined14:15
→ na3r_ joined14:16
charon a not-so-painful way might be to have a shadow copy of next which is rewritten (with filter-branch) to have the files in the old places14:16
→ ixti joined14:17
→ sdeboey joined14:17
→ tvw joined14:19
← na3r__ left14:19
→ JEEB joined14:19
→ simenbrekken joined14:19
shruggar that sounds like the best option, yeah14:19
← Lgb left14:20
← glcrazy left14:20
→ mastro joined14:20
simenbrekken I removed a submodule by simply rm -Rf'ing it along with .gitmodules but now I can't even do 'git status' because it's trying to read the non-existant submodule directory14:21
→ Lgb joined14:21
→ valtih joined14:21
charon simenbrekken: git rm --cached submoduledir14:22
← lucas left14:22
charon cbeams: also see !workflows for more ideas on how to organize your branching workflow. usually workflows with git-cherry are suboptimal because they don't exploit git's power in doing and handling merges14:23
→ lucas joined14:23
charon er, !workflow14:23
gitinfo Finding the right workflow for you is critical for the success of any SCM project. Git is very flexible with respect to workflow. See http://sethrobertson.github.com/GitBestPractices/#workflow for a list of references about choosing branching and distributed workflows.14:23
→ Black joined14:23
→ hyperair joined14:23
simenbrekken charon: says there's no matching files, is there any command to list the cache contents?14:23
← hyperair left14:24
charon simenbrekken: git ls-files -s14:24
← Brammm left14:24
shruggar yes, "…cherry-picking… " "…as people usually do…" sounds wrong to me :)14:24
← notjohn left14:24
simenbrekken charon: curiously it's not in that list14:24
→ fruight joined14:24
cbeams !workflow14:25
gitinfo Finding the right workflow for you is critical for the success of any SCM project. Git is very flexible with respect to workflow. See http://sethrobertson.github.com/GitBestPractices/#workflow for a list of references about choosing branching and distributed workflows.14:25
→ notjohn joined14:25
simenbrekken charon: I appear to be in some sort of limbo :(14:25
charon simenbrekken: in that case look for remnants of the submodule in .git/config14:25
simenbrekken: specifically, submodule.FOO.*14:25
simenbrekken charon: I did that and removed them from the config14:25
charon and it still wants to recurse there?14:25
simenbrekken yes14:25
← notjohn left14:26
charon huh. sounds like a bug. perhaps collect 'git ls-files -s', 'find . -name .gitmodules', 'git config -l' etc. and send a bug report to [email@hidden.address]14:26
simenbrekken charon: Ah, it was simply a .git ref file in a directory deep in some library folder14:27
→ dv310p3r joined14:28
charon i wish i had some clue as to how git-status detects what to recurse into14:28
→ VladGh joined14:29
← m1sc left14:30
→ na3r__ joined14:30
charon simenbrekken: btw i still think it's a bug that status became useless, it should never do that14:31
simenbrekken agreed, there's no reason why I shouldn't be able to recover from it or via verbose output get a clue as to why it's recursing14:31
← maleknet left14:32
← lucas left14:32
→ t4nkd joined14:32
→ lucas joined14:33
← na3r_ left14:34
→ Dragooon joined14:34
← subh left14:34
→ chrisdev joined14:34
→ RajeevGEC joined14:35
→ Raging_Hog joined14:35
RajeevGEC can i lock a git repo using a password?14:35
FauxFaux Not using the git protocol, but using ssh you get passwords for free.14:35
cbreak-work git doesn't do authentication14:36
← jonesy left14:36
← BiggFREE left14:37
RajeevGEC so if i have to create a local repo,then i would have to collect the ssh keys of each contributor,right?14:37
← lucas left14:37
← na3r__ left14:37
← gavin_huang left14:38
cbreak-work if you use gitolite, yes14:38
RajeevGEC i use gitosis14:38
cbreak-work :/14:38
→ hyperair joined14:38
cmn !gitosis14:38
gitinfo gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite14:38
→ juanfer joined14:38
cbreak-work then you don't need passwords14:38
← tatsuya_o left14:39
→ TeckniX joined14:39
→ alan_w joined14:40
→ Jay_Levitt joined14:40
→ schlumpi joined14:40
→ tatsuya_o joined14:40
RajeevGEC <cbreak-work>:sorry,,,i didnt get that...i am to have multiple repos,with access to each limited to a particular group of developers14:40
→ gchristensen joined14:40
cmn that's what gitolite is for14:41
→ m1sc joined14:41
RajeevGEC ok...i will try it right away:)...thanks:)14:41
← lmatteis left14:44
← RajeevGEC left14:44
cbreak-work raj, you don't need passwords because you have public/private keys14:45
→ valtih_ joined14:46
→ diogogmt joined14:46
→ DrNick joined14:47
Nickeeh Hey guys!14:47
→ aristidesfl joined14:48
Nickeeh Suppose just cloned a repo, then want to go back a commit or thousand and only apply commits touching a certain path, how would I do that?14:48
cmn you might be looking for git-subtree14:49
alternatively, git log -- file; store that in a file, and loop git cherry-pick14:49
← valtih left14:49
valtih_ → valtih14:49
← Cromulent left14:50
→ storrgie joined14:50
← alan_w left14:51
Nickeeh cmn, I'll look into both, thanks14:53
cbreak-work check out man git-filter-branch14:53
gitinfo the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html14:53
cbreak-work subdirectory filter can extract spatially partial history14:53
← olinux left14:53
Nickeeh cbreak-work, that sounds complicated. Would it be worth it if the merges are expected to be trivial and few?14:55
as in, 5 or less?14:55
shruggar I'm getting "git-shell died of signal 13" "fatal: the remote end hung up unexpectedly" other than that, everything seems just dandy. I expect one of my hooks is failing, but I don't know how to narrow it down (I've tried just adding debug statements everywhere,,, but no idea so far based on that)14:56
charon shruggar: 13 is SIGPIPE. perhaps one of your hooks doesn't read all of its input (it must)?14:56
shruggar ah, makes sense14:57
charon also see the recent "SIGPIPE means your hook is broken, so let's die() verbosely" thread between peff and jrnieder14:57
← nixmaniack left14:58
→ aldodelgado joined14:59
→ nixmaniack joined14:59
→ flyz joined15:00
← vandal- left15:00
→ babilen joined15:00
← fayimora left15:01
→ RaceCondition joined15:01
← flyz_ left15:01
← DrNick left15:01
shruggar yeah, that got it. I'd had a post-receive hook which was only bothering to do anything if certain pre-conditions were met. Otherwise it would just exit (without reading)15:02
← JohnBat26 left15:02
← chepre left15:03
→ shibayu36 joined15:04
→ soc42 joined15:05
← beautifulmind left15:05
← valtih left15:08
← diogogmt left15:08
→ diogogmt joined15:09
→ shtrb joined15:09
← kpreid left15:10
kini I notice that when I define aliases with `git config alias.foo '!some shell command'`, the shell command does not see git environment variables like $GIT_AUTHOR_NAME, like it would if I were using git filter-branch. Why?15:11
thiago did you define those variables in the environment before calling git foo?15:11
cmn you might need to export then if git starts a subshell15:12
kini thiago: of course not :)15:12
thiago kini: then that's why15:13
kini: those variables aren't set unless you set them15:13
kini: git filter-branch is a special command15:13
kini these are environment variables that git itself sets in the subshell in which it calls shell functions defined in invocation of filter-branch15:13
→ airborn joined15:13
kini hmm15:13
cmn are you sure?15:13
kini so then how is my alias shell script supposed to know things like the author name, etc? :)15:13
cmn you tell your app15:14
or you query the config15:14
thiago filter-branch sets it15:14
not other commands. Just filter-branch.15:14
kini oh, querying the config sounds like a plan15:14
← ph^ left15:14
kini thiago: OK, I see15:14
← Fandekasp left15:15
← airborn left15:16
← uu left15:16
mastro I remember some years ago when I introduced some windows collegue to git I made him install a substitute shell for use with msysgit... I don't remember which one, can you suggest a good windows shell that can be used with msysgit?15:17
cmn msysgit brings bash15:17
it's all in there15:17
→ valtih joined15:17
cmn or you might want to use it with powershell15:17
→ stkrzysiak joined15:18
← stkrzysiak left15:18
cbreak-work Console215:18
that's the one my windows-using coworkers use15:18
→ stkrzysiak joined15:18
cbreak-work with git's bash as shell15:19
→ LimeBlast joined15:19
← rurufufuss left15:22
← toabctl left15:22
→ DrNick joined15:23
→ marc_kdab joined15:23
→ Pugnax__ joined15:23
← diogogmt left15:23
mastro cmn, I'm talking about the virtual terminal15:24
→ axl_ joined15:24
→ diogogmt joined15:24
mastro cmn, like gnome-terminal or kterm or whatever15:24
← axl_ left15:24
cmn then use what cbreak said, as it's apparently a terminal emulator15:24
or use the one that's already there than you can use by doulbe-clicking on the msysgit icon15:25
→ digiv joined15:25
→ axl_ joined15:26
→ iamjarvo joined15:26
mastro cbreak-work, thanks15:26
← Pugnax_ left15:26
mastro cmn, ah thanks I did not see his message :)15:26
→ alan_w joined15:27
← Pugnax__ left15:27
cbreak-work that happens me a lot15:29
← herdingcat left15:29
cbreak-work must have ninja skills above that of a normal mortal15:29
← arvind_khadri left15:30
→ arvind_khadri joined15:31
← arvind_khadri left15:31
→ arvind_khadri joined15:31
→ olinux joined15:31
→ fayimora joined15:32
→ jonesy joined15:33
→ Destos joined15:34
← tatsuya_o left15:36
→ skitrees joined15:37
← nixmaniack left15:38
→ mjd_ joined15:39
← brian_g left15:40
kini can someone explain this? http://pastebin.com/LmeAbyXW15:41
→ nixmaniack joined15:41
cmn the commands gets the arguments appended15:42
→ ph^ joined15:42
→ CannedCorn joined15:43
→ diegoviola joined15:43
cmn change it to !echo and see what happens, or run it with several arguments15:44
← arvind_khadri left15:44
→ ceej joined15:44
shruggar also note that if you need to process arguments (and do anything other than the most basic of "alias"ing) it usually makes more sense to use a script names git-yourcommand15:45
cmn indeed15:45
← LimeBlast left15:45
→ edude03 joined15:46
→ tatsuya_o joined15:46
shtrb how can I get file that I used git rm back ?15:47
← shibayu36 left15:47
shruggar shtrb: when did you use "git rm", what have you done since?15:48
shtrb another few commits15:48
→ brian_g joined15:48
shtrb shruggar , I have the release number (using git svn)15:48
cmn release number? you mean commit id?15:49
shtrb yep15:49
shruggar git show <commit where it last existed>:<path to file> > file15:49
I assume "git checkout <commit> -- <file>" would also work15:49
skitrees how do I get a diff between two branches15:49
as if I had run git diff with unstaged changes15:49
cmn you give the branches to diff15:50
← xjonc left15:50
skitrees cmn, ok... syntax?15:50
cmn man git diff15:50
gitinfo the git-diff manpage is available at http://jk.gs/git-diff.html15:50
← raatiniemi left15:50
cmn third syntax15:51
← kermit left15:51
→ garpenlov joined15:52
→ iocor joined15:52
← flijten left15:52
← bgerber left15:52
→ Brammm joined15:53
kini cmn: aha, of course :)15:53
→ bgerber joined15:53
← macmartine left15:53
→ vdv joined15:53
← [M]ax left15:55
→ glcrazy joined15:55
Brammm Hi guys, I have succesfully set up a bare repository (it's a basic web app) on my server, I've managed to clone it on my machine and am able to push to the remote master using SSH. Now I also want to be able to view the project online. I'm guessing I have to set up a post-receive hook. But what command should it execute?15:55
Should I clone the repository again on the machine and update from the bare one on push?15:56
← crooksey left15:56
→ flaguy48 joined15:57
← mastro left15:57
wereHamster Brammm: rsync the website to the server15:57
Brammm wereHamster: okay, i'll google that, I'm learning as I go :p15:58
→ tuyga joined15:58
→ mastro joined15:58
→ [M]ax joined15:58
→ mikepack joined15:59
← lb left15:59
← tqmz left15:59
← Desproges left16:00
→ harshpb joined16:00
→ beaufour joined16:00
← simenbrekken left16:01
← fruight left16:01
← harshpb left16:01
→ indstry joined16:02
sh4rm4 hmm since when is git no longer hosted at kernel.org ?16:02
i can't seem to find any release tarballs of newer versions...16:03
← brian_g left16:04
Brammm wereHamster: When I google around, I find a lot of people suggesting doing a checkout -f, is that also a solution?16:04
sh4rm4 ah... thats under "older releases"...16:05
hmm, y u no bz2 ?16:05
cbreak-work sh4rm4: you should be able to find a gzip tool for your system16:06
→ g0bl1n joined16:06
← g0bl1n left16:06
→ g0bl1n joined16:06
sh4rm4 its more about smaller download and disk storage16:07
← gchristensen left16:08
← diogogmt left16:08
bremner the management will be notified of your dissatisfaction.16:08
→ diogogmt joined16:08
→ joshsmith joined16:09
cmn sh4rm4: the tarballs are on code.google.com/p/git-core16:10
→ danielrmz joined16:10
→ toni_ joined16:10
→ kukks joined16:10
cmn Brammm: so you want to use git as a deployment system, then. Don't16:10
← fayimora left16:10
→ QiBala joined16:10
→ betovar joined16:11
← dc5ala left16:11
ChanServ set mode: +v16:11
Brammm cmn: okay, what would be the best approach then? It's only for testing and learning purposes. Use the rsync?16:12
cmn rsync is good16:12
it depends on what you need16:12
!website16:12
gitinfo Git is not a website deployment tool, but can sometimes play one in sufficiently simple/lax environments with a little help. One example of help is: http://toroid.org/ams/git-website-howto and another approach: https://gist.github.com/171423516:12
cmn the first link is that checkout -f but it's very limited in what it can deal with16:12
the second link is much better16:12
→ Slydder joined16:13
← diogogmt left16:13
Slydder hey all. am wanting to start developing wordpress plugins and am looking for a howto on how best to setup Git so that only my plugin code is included in the repository. does anyone here maybe have a link or could point me in the right direction?16:13
← QiBala left16:13
→ QiBala joined16:13
toni_ Hi! I'm trying to import a bazaar repo into git, but fail due to illegal characters in the bzr tags (in my case: version names with an epoch). trying bzr fast-export --plain didn't solve the problem, hand-editing the version numbers didn't help anything, too.16:13
← cabaire left16:14
→ diogogmt joined16:14
toni_ epoch: package version 1:2.3.4-016:14
Brammm Right cmn that looks complicated :p At the moment I don't need much: I just want an up to date copy of my master branch viewable for anyone, don't care about bugs :p16:14
cmn that's what the second link does16:14
a checkout -f might seem to work16:14
← cbreak-work left16:15
toni_ using vi on the fast-export file yields in various other crashes related to "unknown command"s16:15
s/yields in/resulted in/16:15
thiago fast-export produces a binary output16:15
Brammm yeah, at the moment the first link doesn't even work, however, I've not done the chmod thing, testing again.16:15
thiago make sure your text editor doesn't corrupt it16:15
toni_ thiago: looks like a text-file, though16:16
thiago toni_: it isn't16:16
toni_: not if you have a binary file in your repo16:16
toni_: each file's *exact* byte content is present in the fast-export16:16
toni_ thiago: it's bazaar's fast-export file, and it's only python code that i want to export, so...16:16
← Sitethief left16:17
toni_ (ok, some shell, makefile and similar stuff, too)16:17
→ jmoreira81_ joined16:17
→ harshpb joined16:17
toni_ vi usually doesn't corrupt files, afaik16:17
thiago toni_: still, your text editor might have corrupted it16:17
toni_: do a diff -u original new and check what it changed16:18
Brammm cmn: I'm getting the feeling the post-receive hook is not being executed, what could cause that?16:18
cmn wrong permissions, wrong repo16:18
Brammm i only have one repo16:18
→ ehsan joined16:18
← jmoreira81 left16:18
jmoreira81_ → jmoreira8116:18
cmn no16:18
Brammm and the push itself is succesful16:18
cmn if you can push, you have two repos16:18
→ crisp joined16:18
Brammm ah16:18
← soc42 left16:19
← diogogmt left16:19
Brammm I have a repo on my remote server and on my machine. I added the post-receive hook on the remote, do I have to add it locally? that doesn't make sense to me :/16:19
→ ohcibi joined16:19
cmn you have to put it in the remote where you want it to be executed16:20
← Spockz left16:20
→ diogogmt joined16:20
Brammm I also changed the dir to the dir of my webpage16:20
→ witti joined16:21
cmn that seems sensible16:22
← tuyga left16:22
← TrineLocke left16:22
→ alansaul joined16:22
← betovar left16:23
← prudnikov left16:23
toni_ thiago: umm... managed to import, but with now severely broken version numbers :/16:23
cmn version numbers meaning what?16:25
Brammm cmn: If I just put echo 'test' in the post-receive hook, I should see that output in the git gui window when pushing, no?16:25
project2501a hey guys, how do i transform a tracking repository to a working repo?16:25
cmn any echo from the hook will the passed to the git doing the push16:25
→ macmartine joined16:25
cmn I've no idea how git-gui presents it16:25
shruggar project2501a: neither of those are git terms. Can you clarify?16:25
Brammm aight, do you see it when doing it thorugh a terminal?16:26
→ GeorgeSebastian joined16:26
→ NetRoY joined16:26
gitinfo set mode: +v16:26
mjd_ project2501a: I'm not sure just what you mean, but if /path/to/repo is a path to the repository, you can make a working copy of it with "git clone /path/to/repo".16:26
project2501a shruggar: hm. funny, i am reading the o'reilly git book and they are taking about "tracking repository" and "working repository"16:26
cmn Brammm: yes, command-line git shows it16:26
project2501a shruggar: let me go back to the book and read it one more time16:27
→ kennethreitz joined16:27
← diegoviola left16:27
cmn project2501a: working repo is not a git term, but it might refer to the repo where you work16:27
← storrgie left16:27
project2501a cmn: "working directory"?16:27
cmn tracking repo seems wrong and/or misleading16:27
project2501a: that's a different thing16:27
shruggar project2501a: the nearest terms I can think of are "remote-tracking branch" and "working directory"16:27
Brammm not seeing the test, god damn, I feel I'm making some stupid mistake16:28
project2501a shruggar: those are probably it. i better go back and read and get my terminology right.16:28
→ dougsland joined16:28
← dougsland left16:28
→ dougsland joined16:28
project2501a sorry guys!16:28
← _Qman left16:28
shruggar (and there is no way to convert a remote-tracking branch into a working directory, as they are as different as pictures of apples and descriptions of someone else's apples)16:29
← diogogmt left16:30
shtrb shruggar, thank you it worked (at last)16:30
→ mxweas_ joined16:30
project2501a shruggar: *nod* what i really want to do is establish a chain between three points: developer->maintainer->production16:30
but as i said, i think i need more rtfm16:30
← jatt left16:31
→ mandric joined16:31
→ diogogmt joined16:31
← mxweas_ left16:32
→ wuj joined16:33
← GeorgeSebastian left16:34
→ Shakyj joined16:34
→ gchristensen joined16:34
→ zulax joined16:34
→ vandal- joined16:34
Shakyj Hey, anyone got a decent guide for setting git with a server? All the guides google gives me seem poor16:35
← diogogmt left16:36
cmn what do you mean "with a server"? do you want to work with others on some project?16:37
Shakyj cmn, Yeah, but not git-hub. So need a guide that does the full setup :)16:37
→ diogogmt joined16:37
cmn !gitolite16:37
gitinfo Want to host as many git repos (and users!) as you like, on your own server, with fine-grained access control? You want gitolite: https://github.com/sitaramc/gitolite16:37
Shakyj cmn, Cheers, I will try that :)16:38
cmn "not github" covers many things, including bitbucket and gitorious16:38
→ mxweas_ joined16:38
→ KayGridley joined16:38
Shakyj cmn, self hosted. not hosted elsewhere is what I was trying to cover by not git-hub. But this looks what I was after16:39
cmn so say self-hosted then16:39
← amigojapan left16:39
← skoop left16:39
→ jweinberg joined16:40
Brammm Is a push executed when there have been no commits?16:40
← papegaaij left16:40
← dhruvasagar left16:40
→ pfFredd joined16:41
cmn a push is executed when you say to16:41
if there are no differences, there is nothing to update, so it won't do that, and the update and post-receive hooks won't run16:42
← witti left16:42
← giallu left16:42
mjd_ You could of course be pushing a new ref to a repo that already has all of that branch's commits.16:42
So there might be something to do even if there are no new commits.16:43
← DarkAR left16:43
mjd_ → yrlnry16:43
→ zagra joined16:43
→ subh joined16:43
← mxweas_ left16:44
→ dhruvasagar joined16:44
cmn sure, but that's covered under difference16:44
yrlnry Someone asking bramm's question has probably not thought to consider this case specifically.16:45
I was answering the question that was asked,16:45
.16:45
← Black left16:45
← alexgordon left16:47
→ arvind_khadri joined16:47
→ _marvin joined16:47
← dr_win_ left16:47
→ dr_win joined16:47
→ tokiya joined16:48
→ alexgordon joined16:48
→ brian_g joined16:48
← h0bbit left16:49
Brammm god damn16:49
I've been testing so many times16:49
← jweinberg left16:49
Brammm suddenly, I see a copy of my working repo in my htdocs folder16:49
and now I don't know how it happened16:49
*bangs head*16:49
→ GeorgeSebastian joined16:49
→ arietis joined16:50
← Lisimba2 left16:51
→ Lisimba joined16:52
← RaceCondition left16:53
Brammm Ack, getting frustrated. How the hell can you even test this if I can't seem to get the hook working...16:53
cmn test locally first16:53
see if the script even does what you want it to16:54
→ Pugnax joined16:54
Brammm it just says 'echo "test"'16:54
and I'm not getting it in the bash window16:54
cmn then make sure it can be run on the remote system16:54
← arvind_khadri left16:55
cmn ssh in and try to run it16:55
Brammm I assume since I'm executing over ssh my account "bram" is also executing the script? So I have rights?16:55
>_< "bash: permission denied16:55
cmn if you're pushing as 'bram', then that's the user that needs to be able to execute the script16:55
Brammm well, that explains a lot :p16:55
cmn it's also the first thing I told you to check16:55
← NetRoY left16:56
Brammm Yeah, I probably didn't get it because of my ignorance16:56
god damn :p16:56
sitaram Brammm: and in general, use 'echo >&2' insteas of 'echo'16:56
→ Black joined16:56
sitaram instead*16:56
cmn huh, really? I've been using echo and haven't seen anything wrong; what could be the issues?16:57
← disappearedng left16:57
→ Heimidal joined16:57
→ mastroWork joined16:58
Brammm Okay, stupid as hell question, but how do I make sure I can execute the file? Google tells me chmod, but I don't understand the options16:58
(Git AND linux noob)16:58
cmn you give yourself execute permission16:58
chmod u+x script.sh16:58
sitaram cmn: depends on what is ttriggering the hooks. Sometimes the client side is still talking the git protocol and gets confused16:58
→ friskd joined16:59
Brammm yeah okay16:59
i'm owner now16:59
cmn oh, I see, in pre-receive that would be handy16:59
Brammm and the command worked16:59
← dr_win left16:59
Brammm okay.... sigh16:59
let's try this one more time :p16:59
← mastroWork left17:00
Brammm Okay17:00
← Heimidal left17:00
← Zenopus left17:01
→ Heimidal joined17:01
Brammm That's one hour of frustration I could've spent on something else17:01
:p17:01
thanks cmn for putting up with me, everything works now.17:01
← mastro left17:01
← stodan left17:01
Brammm sitaram: so if I want to output something like "Copied to staging server" I use "echo >&2 'copied...' "17:02
?17:02
← diogogmt left17:03
← whitman left17:03
← _marvin left17:04
← netiru left17:04
→ diogogmt joined17:04
→ dfr|mac joined17:05
← indstry left17:05
← vdv left17:05
← psoo left17:06
Brammm sitaram: thanks, it works ^^ I'm a happy nerd again :p17:06
thanks for the help guys, cya17:06
← iamjarvo left17:06
← Brammm left17:06
→ Bass2 joined17:09
→ Psylock524 joined17:09
← drupol left17:09
← jceb left17:09
→ Spockz joined17:10
← Bass2 left17:10
← soulcake_ left17:10
→ Bass2 joined17:10
← senny left17:11
→ Rogi joined17:11
← akiress left17:11
← diogogmt left17:11
→ mxweas_ joined17:12
→ arvind_khadri joined17:12
→ soulcake_ joined17:13
→ diogogmt joined17:13
→ Nikoli joined17:14
← csgeek left17:14
→ Ksugzulut joined17:15
← Slydder left17:15
← teweWORK left17:16
→ dfr|mac_ joined17:17
→ someprimetime joined17:17
← arvind_khadri left17:18
← joshsmith left17:18
→ joshsmith joined17:18
→ matthiasgorgens joined17:19
→ csgeek joined17:19
← bgy left17:20
← dfr|mac left17:20
→ Wafflenator joined17:21
gitinfo set mode: +v17:21
Wafflenator Hello, can someone tell me how to create a local remote or shortcut that creates an alias for a repo on my machine?17:21
Every time I startup git, I have to cd to a long path and its annoying. Can't find this through google searches.17:22
thiago git remote add17:22
huh?17:22
Wafflenator I tried that, it says its not a git repo (.git)17:22
thiago sorry, I didn't get it17:22
can you rephrase what you meant?17:22
harshpb → harsh_dinner17:22
Wafflenator git remote add gym "C:\Users\Waffles\Documents\Web Projects\Justin\Site"17:22
thiago and that doesn't work?17:23
Wafflenator I would like to create an alias for a local repository.17:23
No.17:23
fatal: Not a git repository (or any of the parent directories): .git17:23
It is a repo.17:23
thiago git log17:23
do you see anything?17:23
← gulden left17:24
← mishok13 left17:25
Wafflenator If I cd to it? Yes.17:25
thiago then cd to it17:25
then you can run git commands17:25
cmn put in a symlink somewhere on your filesystem17:25
or a shell alias17:25
thiago cmn: note that it's windows17:25
cmn then a shell alias17:26
Wafflenator How do I create a system link? I can't right?17:26
cmn NTFS doesn't support symlinks17:26
Wafflenator The idea is to not have to cd to it every time it is a long path.17:27
cmn they're symbolic links, not system links17:27
→ diegoviola joined17:27
cmn then use an alias; what shell do you use?17:27
Wafflenator How would I know?17:27
thiago Wafflenator: this is #git, we can help you with Git.17:27
cmn it tells you17:27
that too17:27
thiago Wafflenator: we can't really help you with how to use your system's shell.17:27
Wafflenator Git says its version 1.7.717:28
cmn that's git, not your shell17:28
Wafflenator thiago: then your sure git doesn't have the ability to do this?17:28
Yeah, how do I check my shell version?17:28
cmn git works on the local directory17:28
thiago Wafflenator: yes17:28
cmn how you check for the version depends on the whell17:29
thiago Wafflenator: git starts working once you're inside a repository17:29
cmn shell*17:29
thiago Wafflenator: except for git clone and git init17:29
Wafflenator Okay, nm then thanks guys. I can cd on init using a batch script I've done that before17:29
→ netiru joined17:29
← Wafflenator left17:30
← babilen left17:30
→ octoquad joined17:30
→ kenperkins joined17:31
flavius How could I split a hunk further when using git add -p file? I get this http://i.imgur.com/uuh2k.png and it doesn't let me split it down - I would like to stage only the deletion of that one line17:32
thiago press e and edit the hunk17:32
→ pdtpatr1ck joined17:32
thiago and please pastebin text, not images17:32
← subh left17:33
← johnathanb left17:33
flavius it's colored, it's easier to read, it features powerful git colors!17:34
:D17:34
← mklappstuhl left17:34
cmn a hung of picture isn't easier to read than text17:34
specially with that horrible font17:34
flavius cmn: what font would you like? I can set it :D17:34
→ Altreus joined17:35
← angelsl left17:35
cmn I want text17:35
Altreus I get this -- fatal: Unable to create temporary file: Permission denied -- but I'm not sure where it's trying to write the temporary file on the remote end. Can anyone help?17:35
er, when trying to push17:35
← alan_w left17:35
Altreus error: unpack failed: index-pack abnormal exit17:35
cmn ssh into the system and check the permissions in the objects dir17:36
← soulcake_ left17:36
Altreus ah, it's probably doing it by group -- let me try17:36
cmn: thanks that was it :)17:37
cmn you should try to figure out why that happened17:38
permissions don't change automatically17:38
← iocor left17:38
→ srcerer joined17:39
← tcurdt left17:39
← bgerber left17:39
→ griotspeak joined17:40
flavius using images is much more convenient for small things which don't need to be line-numbered - usually for visual things, not source code17:40
sh4rm4 i created a tag with -a - how can i push it upstream ?17:40
← mxweas_ left17:40
← Spockz left17:40
flavius and I'm using a script to take the screenshot, post to imgur, and copy the URL to my clipboard. what could be more convenient?17:40
cmn to let people read text however they wish17:41
→ multi_io_ joined17:41
cmn sh4rm4: git push tag tagname17:41
← thiago left17:41
cmn with the remote in there17:41
← netiru left17:41
← SegFaultAX left17:42
flavius cmn: come on, it's monospaced, as for programmers. and I bet you don't set the fonts "as you wish" for pastebin sites, you just use the site's / browser's default font17:42
→ bgerber joined17:42
flavius be honest to yourself, what you say is pure religion17:42
cmn so you'd rather make it harder for people who are trying to help you?17:43
← retoryter left17:43
flavius was it harder? be honest17:43
sh4rm4 cmn++17:43
cmn are you completely sure that everybody in this channel has perfect sight?17:43
← multi_io_ left17:43
→ SegFaultAX joined17:43
cmn that you're not excluding people by posting something they can't read?17:43
→ mklappstuhl joined17:43
cmn and yes, it is harder17:43
→ retoryter joined17:44
flavius I would rather not put a burden on people with problems to help me, they have (understandably) enough to deal with17:44
cmn you have to adjust to the other person's environment and find the small piece of the picture that shows the important stuff17:44
→ multi_io_ joined17:44
multi_io_ I want to git push to a repo which I have access to only by moving files there via USB sticks.17:45
cmn use a bundle17:45
flavius multi_io_: use the file protocol17:45
multi_io_ how do I accomplish this?17:45
→ cenuij joined17:45
← cenuij left17:45
→ cenuij joined17:45
← Element9_work left17:45
← mklappstuhl left17:45
→ mklappstuhl joined17:45
shruggar if it's small enough to not need line numbers, it's small enough to paste in IRC (ie: it's one line long :))17:45
multi_io_ bundle, ah17:45
→ h0bbit joined17:46
multi_io_ flavius: you mean having a repo on the USB stick?17:46
flavius multi_io_: yes, you can just set the remote's url to a local path, and on the usb stick you'd have the bare repo17:47
multi_io_ yeah, I get that17:47
→ iamjarvo joined17:47
→ Araxia joined17:47
→ redheadphones joined17:47
multi_io_ actually I don't really have a USB stick right now, I prefer emailing files for now, so git bundle it is17:48
← gchristensen left17:48
→ orospakr joined17:48
multi_io_ :)17:48
→ alan_w joined17:48
← zulax left17:48
→ f0i joined17:48
flavius patches over e-mail... I think we need git for another 20 years around so we can get rid of the stone age :D17:49
← digiv left17:49
cmn huh? a bundle isn't a patch17:49
→ Sadin_ joined17:49
← retoryter left17:51
← johnhamelink left17:52
→ Biopandemic joined17:54
← bgerber left17:55
← garpenlov left17:55
→ igotnolegs joined17:57
SethRobertson Better than a patch17:57
← igotnolegs left17:57
← achadwick left17:57
← Biopandemic left17:57
← diogogmt left17:58
→ OOPMan joined17:58
→ mbroeker joined17:59
→ dc5ala joined17:59
→ diogogmt joined18:00
→ airborn joined18:00
← toni_ left18:00
→ lacrymology joined18:01
← sdeboey left18:01
→ garpenlov joined18:02
→ fruight joined18:03
→ maletor joined18:03
← pdtpatr1ck left18:04
← shruggar left18:04
→ sudoquai joined18:05
→ prudnikov joined18:06
← tatsuya_o left18:08
→ bgerber joined18:08
← alansaul left18:08
→ Error404NotFound joined18:09
→ disappearedng joined18:09
→ Cromulent joined18:10
← overbythere left18:10
→ j416 joined18:11
← friskd left18:11
← fairuz_ left18:11
→ Russell^^ joined18:12
→ fairuz_ joined18:13
← KayGridley left18:13
← JEEB left18:14
← alan_w left18:14
← Psylock524 left18:14
→ prateekagr joined18:15
→ w09x joined18:15
→ kermit joined18:15
← kermit left18:15
← Dragooon left18:16
→ mxweas_ joined18:16
Vampire0_ → Vampire018:17
→ ferdna joined18:17
→ sekati joined18:18
→ stockholm joined18:19
harsh_dinner → harsh18:19
gitinfo set mode: +v18:20
harsh → harshpb18:20
→ kermit joined18:20
← kermit left18:20
→ thiago joined18:20
← dharrigan left18:20
← apok left18:20
→ kermit joined18:20
← kermit left18:20
stockholm hi18:20
i have a github project18:21
with changes18:21
← tvw left18:21
→ Lujeni joined18:21
stockholm and i want to merge them into an other github project18:21
→ pdtpatr1ck joined18:21
stockholm let me look up the URLs...18:21
← w09x left18:21
→ kermit joined18:22
SethRobertson stockholm: Typically you submit a pull request via the github web ui18:22
stockholm i want to merge this https://github.com/stockh0lm/setup-storage/tree/andreas/setupstoragerefactor into https://github.com/faiproject/18:23
→ apok joined18:23
cmn do you have access to the faiproject?18:23
also note that the second URL for a user, not a repo18:23
stockholm SethRobertson: this is more advanced. i used svn2git to create my stuff, and worked on that, in a branch, and later the whole project moved to github18:24
← nomadtacos left18:24
← apok left18:24
cmn why did you use svn2git when the project is already in git?18:24
stockholm oh,https://github.com/faiproject/fai-project.org ?18:24
SethRobertson Well...add their project as a remote and see if the SHA line up. If not, you will be in for some work18:25
→ Chowzzf joined18:25
← harshpb left18:25
stockholm cmn: it wasnt at that time. or it was still in limbo and transition18:25
→ harshpb joined18:25
stockholm SethRobertson: how do i do that? add it as a remote?18:25
SethRobertson: and how do i see it the sha line up?18:26
SethRobertson git remote add faiproject git://url.that/github/profiels18:26
someprimetime http://toroid.org/ams/git-website-howto this is a good guide on writing content locally and then pushing it to the server, but basically i want to do it the other way around: e.g. initialize a repo on my server which is a folder of all the files I want to edit, then make branches locally and then push those to that repo. would this same guide apply?18:26
→ tewecske joined18:26
cmn someprimetime: that howto is for abusing git as a deployment tool18:26
→ fayimora joined18:26
someprimetime so it's a shitty one?18:26
SethRobertson someprimetime: Yes. Create the repo using your workflow and then use the toroid guide from then on18:27
cmn it's enough for some usecases18:27
→ DofLeFou joined18:27
cmn you still want to have a bare repo18:27
someprimetime it's just me going to be doing everything so i don't need gitolite or a way to manage users18:27
ok i'll go with the bare repo that's what i've been hearing18:27
→ smoge joined18:27
SethRobertson stockholm: Use `gitk --all --date-order` to see if commits before your local changes have appear on the same branch or there are two branches, one for you and one for them.18:27
cmn that has nothing to do with whether that is a good idea18:27
someprimetime: as you'll want to have a bare repo on the server anyway, init a repo where your files are, add them and push to the bare repo18:28
all of this on the server18:28
someprimetime ok sweet thx18:28
cmn then you can use whatever method for working with your files18:28
→ dharrigan joined18:29
← dharrigan left18:29
→ dharrigan joined18:29
→ felkor joined18:29
→ Spockz joined18:30
stockholm SethRobertson: what url is that in "git remote add faiproject git://url.that/github/profiels18:30
?18:30
SethRobertson provides18:30
← mxweas_ left18:30
stockholm i am an owner of the project, but i dont get github to display a proper write-url18:30
cmn it's one of the options at the top where it shows you the url18:31
but you should already have one; aren't you looking for the upsteream url?18:31
SethRobertson yes18:31
stockholm i look at this:18:32
https://github.com/faiproject/fai-project.org18:32
cmn then there's no way github is going to give you a write url18:32
← pdtpatrick left18:32
SethRobertson It will give you a git:// URL18:32
stockholm https://github.com/faiproject/fai-project.org18:32
err18:32
sorry for that18:32
cmn that's the other one18:32
→ katspaugh joined18:32
stockholm [email@hidden.address]18:32
but that is a read-only url, right?18:33
cmn that depends18:33
it's a ssh url18:33
stockholm ah, let me try, then18:33
cmn if you're allowed to, it's a read-write url, but that's irrelevant right now18:33
→ shruggar joined18:33
SethRobertson Actually, to tell you the truth what you are going to want to do eventually is create a new repo which is a fork of faiproject and then move your changes from your old repo to that repo, and then delete your previous repo18:33
stockholm SethRobertson: yes, that sounds right18:34
ddilinger i need some help understanding git diff. I do 'git checkout master', look at some file, 'git checkout origin/master' look at same file, they are the same. 'git checkout master; git diff master origin/master' says they are different, as if its not looking at current master, but something else18:34
stockholm SethRobertson: how do i do THAT? :-)18:34
← OOPMan left18:34
gitinfo set mode: +v18:34
katspaugh Hi! What's the proper way to add a third-party library to my project? The library has it's own repository, but I want to distribute the library along with my code.18:34
cmn !subrepos18:34
gitinfo [!subprojects] So, you want to add git repositories inside of other git repositories? Well, you have four main options. First is to just do it, add the repo to the outer project's .gitignore, and treat them entirely separately. Best if they are entirely seperate. Otherwise your best options are "!submodule" "!gitslave" and "!subtree" Try typing those commands into this IRC channel.18:34
felkor hi, why "GIT_WORK_TREE=../gitrepo.git git pull" works in shell and in a script it returns "/usr/lib/git-core/git-pull: 139: cd: can't cd to /home/user/gitrepo.git". is there a way to do it in a shell?18:34
katspaugh Thanks, cmn!18:34
thiago felkor: are you running that with the same user?18:35
felkor yes18:35
cmn felkor: that's not going to do what yo think it will18:35
katspaugh !submodule18:35
gitinfo git-submodule is ideal to add subsidiary git repositories to a git superproject when you do not control the subprojects or more specifically wish to fix the subproject at a specific revision even as the subproject changes upstream. See http://book.git-scm.com/5_submodules.html18:35
← eli left18:35
felkor cmn in the shell it does18:35
thiago felkor: are you sure that gitrepo.git contains a *checkout* of your repository?18:35
cmn why do you want to use a metadata dir for storing your files18:35
felkor thiago, no it doesnt, but it works in the shell18:36
← lacrymology left18:36
cmn what does work mean in this case?18:36
that command doesn't seem to make any sense18:36
felkor cmn, thiago I run in from the git checkout dir18:36
cmn if you run it from the checkout dir, why are you setting a different one?18:37
and telling it to use a .git dir?18:37
← smoge left18:37
felkor because of ikiwiki18:37
http://ikiwiki.info/tips/laptop_wiki_with_git/18:37
→ joshteam joined18:37
felkor this way i would update my wiki without having to add a git remote18:38
→ samnob_ joined18:38
felkor i was truing that and it works on the shell18:38
*trying18:38
someprimetime i just made a bare repo in ~/git/mydomain.com and did `git init` on the server... was i supposed to make a folder called mydomain.git within mydomain.com and then cd into that and `git init` within that (mydomain.git)?18:39
SethRobertson stockholm: Get the first repo in shape first. Then worry about forking and putting your changes on the fork18:39
felkor it updates the checkout dir without adding the original remote of the gitrepo.git, that was cloned with got clone --mirror18:39
*git18:39
thiago felkor: GIT_WORK_TREE should point to a checkout18:39
cmn felkor: so ../gitrepo.git is where you're trying to pull from?18:40
→ indstry joined18:40
felkor cmn, y, i think so18:40
stockholm SethRobertson: both repos are in shape, taken as they are.18:40
cmn then telling git to checkout files there is completely the wrong thing to do18:40
someprimetime nvm got it18:40
stockholm afaik18:41
felkor thiago, checkout tree is the result of "git clone gitrepo.git" without --mirror?18:41
someprimetime i didn't `git init` locally18:41
SethRobertson stockholm: Yes, I mean answering the question about whether the SHA are the same and getting your changes to be after all remote changes, etc.18:41
← ipalaus_ left18:41
thiago felkor: yes18:41
← nikuyoshi left18:41
→ JEEB joined18:41
← diogogmt left18:42
→ RaceCondition joined18:42
← SegFaultAX left18:42
→ vacho joined18:42
vacho guys..18:42
how do I go back to an earlier commit?18:43
→ infogulch joined18:43
cmn go back to do what?18:43
→ ipalaus_ joined18:43
SethRobertson vacho: To view or to change your repo?18:43
vacho I made some changes that I am not happy with, how do I go back to my last commit?18:43
SethRobertson vacho: !fixup18:43
gitinfo vacho: So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions. For hints type !fixup_hints in IRC. Remember: if you have pushed already, there are only a few things you can do without !rewriting_public_history (type that for more info)18:43
→ diogogmt joined18:43
→ SegFaultAX joined18:43
vacho i dont care about history..18:43
just want to go back to an earlier state18:44
SethRobertson Then follow the link and answer the questions18:44
cmn then follow thosesteps18:44
← skitrees left18:44
felkor what i did: git clone --mirror anothergit.git gitrepo.git; git clone gitrepo.git gitrepo ; cd gitrepo ; GIT_WORK_TREE=../gitrepo.git git pull <- it pulls from the original repo in the shell and gives error in script18:45
→ notjohn joined18:45
cmn it pulls from ../gitrepo.git because that's it's "origin" repo18:45
vacho that's an awesome website, wow!!18:45
→ skitrees joined18:46
someprimetime on my local machine i just created a folder, git init'd in it and added my bare repo from my server.. now i'm checking to see if it's added by going `git remote show origin` and it's asking me for a pw. i'm putting the pw to my server but it's telling me wrong pass18:46
← skitrees left18:46
someprimetime vacho ++ i totally just bookmarked that too haha18:47
SethRobertson someprimetime: `git remote show origin -vv` says what URL?18:47
vacho hehe18:47
← shruggar left18:47
someprimetime SethRobertson: error: unknown switch `v'18:47
← edude03 left18:47
SethRobertson git remote -vv show origin18:47
bah18:47
→ skitrees joined18:47
→ bitslip joined18:47
← diogogmt left18:48
someprimetime it's asking me for my server pw again18:48
i entered it, and it's saying permission denied18:48
Permission denied (publickey,gssapi-keyex,gssapi-with-mic,password).18:48
do i need to regenerate my keys?18:48
→ diogogmt joined18:48
SethRobertson Try `git remote -vv`18:48
← prateekagr left18:49
someprimetime origin ssh://someprimetime@mydomain.com:3000/~/git/mydomain.com (fetch)18:49
origin ssh://someprimetime@mydomain.com:3000/~/git/mydomain.com (push)18:49
→ alanp joined18:50
SethRobertson OK. Run `ssh [email@hidden.address] and try to log in. If that works, then it is git's fault. If it does not work, then it is not git's fault.18:50
← SegFaultAX left18:50
← Shakyj left18:51
someprimetime SethRobertson: ssh: Could not resolve hostname mydomain.com:3000: nodename nor servname provided, or not known18:51
but i usually prefix it with -P 3000 before i ssh to specify the port name18:51
err -p rather18:51
ssh -p 3000 [email@hidden.address] which works18:51
chrisf => git's fault.18:51
← bgerber left18:51
someprimetime so i should just restart?18:52
← Z3Rt left18:52
→ SegFaultAX joined18:52
→ tqmz joined18:52
← fairuz_ left18:52
SethRobertson someprimetime: Try `GIT_TRACE=1 git ls-remote origin`18:52
→ fairuz_ joined18:53
someprimetime SethRobertson: pardon the paste source, but pastebin is down: http://bin.cakephp.org/view/3023461618:53
→ bgerber joined18:53
→ w09x joined18:54
someprimetime so i think i need to change my port to 3000 in my ssh config, maybe?18:54
→ shergill joined18:54
→ Shakyj joined18:54
→ shruggar joined18:54
SethRobertson Um, that's odd. Try `ssh -p 3000 [email@hidden.address] ls ~/git/mydomain.com/config`18:55
← felkor left18:56
← GeorgeSebastian left18:57
→ mxweas_ joined18:58
→ filosofo joined18:59
← brian_g left18:59
ChanServ set mode: +v18:59
← posciak2 left18:59
→ posciak joined18:59
→ traviscj joined19:00
filosofo How can I get the author of a specific hash when my current working directory is not in the repository?19:00
samnob_ → samnob19:00
→ advatar joined19:01
selckin git_dir it19:01
← SegFaultAX left19:01
samnob Is this a good channel to ask about using "repo" code.google.com/p/git-repo/ ?19:01
→ SegFaultAX joined19:02
SethRobertson filosofo: You submitted a patch or someone lied.19:02
samnob: Not as such. Are you trying to use it for non-android things?19:02
samnob No it's android-ish.19:03
← Error404NotFound left19:03
SethRobertson You may ask, but I've never heard anyone profession great experience with it other than random users.19:03
someprimetime SethRobertson: i figured it out19:04
samnob I just want to use different projects (android forks [cyanogenmod and replicant]) without too much duplication19:04
someprimetime i created the directory as root but was trying to look for the directory that would be created as a user19:05
so the home directory paths were different19:05
← mxweas_ left19:05
→ gv joined19:05
someprimetime cause when i'd do cd ~/ as root, it'd be diff as someprimetime cd ~/19:05
← infogulch left19:05
someprimetime thanks for helping19:05
→ rickmasta joined19:06
SethRobertson samnob: If you do the cloning manually you can use --reference to try to share the second with the first. Good luck with that19:06
someprimetime: So use ~user or /home/user19:06
stockholm how can i export (?) my changes as patches+commit messages and apply them as patches to an other git project?19:06
SethRobertson stockholm: First step, add the other repo as a remote and then determine if the SHA line up.19:07
samnob Ugh, I think I'm in over my head. I may just blow off the upstream.19:07
kevlarman stockholm: git-format-patch and git-am19:07
stockholm: or git-send-email in place of the former19:07
filosofo selckin, thanks19:08
← filosofo left19:08
stockholm SethRobertson: i did add it as a remote, but i dont see any newer commits then mine. (do i need to pull?)19:08
chrisf you need to *fetch*.19:08
← DofLeFou left19:09
SethRobertson stockholm: You are not looking for newer commits, you are looking for older commits. Scroll (using `gitk --all --date-order`) to sometime BEFORE you made your commits. Do you see multiple lines of development with each commit appearing twice?19:09
← xsbeats left19:10
→ mishok13 joined19:10
SethRobertson stockholm: Assuming, yes, you have `git remote update`19:10
← Vampire0 left19:10
SethRobertson or fetched19:10
→ brian_g joined19:11
stockholm SethRobertson: that gives me: * [new branch] master -> faiproject/master19:11
is that good?19:11
SethRobertson yes19:11
not try the `gitk` thing19:11
stockholm ok19:11
→ skoop joined19:12
→ Z3Rt joined19:12
dfr|mac_ → dfr|mac19:12
→ ketas joined19:13
← diogogmt left19:14
→ OOPMan joined19:15
← SegFaultAX left19:15
stockholm SethRobertson: now i see tons of duplicate entries.19:15
SethRobertson: except my own, which are only in my branch19:15
→ diogogmt joined19:16
← cortana2 left19:16
stockholm SethRobertson: i guess that means that the shas dont match19:16
SethRobertson stockholm: OK. Now you want to make a local tracking branch off of the new $remote/master (or whatever branch you are interested in) . `git checkout -t othermaster $remote/master`19:17
Right, the SHA do not match. You will want to cherry-pick your changes over to a new branch19:17
stockholm ok19:17
and $remote is the git url that i get from github from the fai-project?19:18
← wuj left19:18
← truzie left19:18
SethRobertson No, $remote is the name in the `git remote add $remote $URL` command19:18
stockholm ah19:18
~/src/fai(master)$ git checkout -t othermaster faiproject/master19:19
fatal: Missing branch name; try -b19:19
→ nikuyoshi joined19:19
← octoquad left19:19
← Spockz left19:20
SethRobertson You will need to add a -b before or after the -t. The man page doesn't say which order, in a fit of stupidity19:20
stockholm SethRobertson: ok19:20
← diogogmt left19:20
thiago the order shouldn't matter19:20
SethRobertson It does19:21
← OOPMan left19:21
thiago huh?19:21
so -t takes an option?19:21
→ boombatower joined19:21
stockholm ~/src/fai(master)$ git checkout -t -b othermaster faiproject/master19:21
Branch othermaster set up to track remote branch master from faiproject.19:21
Switched to a new branch 'othermaster'19:21
after, i guess :-)19:21
→ diogogmt joined19:21
cmn -b gets an option, so it should be the last one19:21
← mishok13 left19:21
stockholm SethRobertson: and now?19:21
→ gchristensen joined19:22
SethRobertson stockholm: Before you mean. Well now you will need to find the SHA of the commits you made. Are they all in a row?19:22
stockholm yes19:22
they are19:22
thiago from the manual, it doesn't seem like -b takes an option19:22
← vacho left19:22
thiago it's just a switch19:22
→ A1kanaro joined19:22
→ Spockz joined19:22
SethRobertson stockholm: OK, then find the SHA of the first commit and the SHA of the last commit, and run `git cherry-pick FIRST...LAST`19:22
→ infogulch joined19:22
cmn oh, good point19:22
→ Vampire0 joined19:23
SethRobertson I would greatly support anyone who wants to submit a patch for this. Why, I might even give them a ++19:23
← martinjlowm left19:23
cmn in modern git, the -b isn't needed, is it?19:24
← nikuyoshi left19:24
→ martinjlowm-lptp joined19:24
SethRobertson Well, stockholm reported an error when I tried to not include it19:24
cmn with which version, though?19:25
SethRobertson `git checkout -t randomname nonorigin/master`19:25
cmn of git, I mean19:25
SethRobertson Yes, I know. That doesn't follow the autocreation rules.19:25
stockholm: `git -v`?19:25
→ smw_ joined19:26
SethRobertson Or even git --version19:26
← diogogmt left19:26
→ martinjlowm joined19:26
SethRobertson cmn: It fails in 1.7.8.419:27
(rm -rf z; git init z; cd z; echo A>A; git add A; git commit -m A; git remote add other ..; git checkout -t randomname other/master)19:27
← [M]ax left19:28
→ diogogmt joined19:28
cmn fails with 1.7.9.rc0 (no I idea why I have that version); maybe it works for origin19:28
→ [M]ax joined19:29
← lyddonb left19:29
← brian_g left19:29
cmn hm, other/master doesn't even exist in that example19:29
SethRobertson cmn: Adding a `git remote update` doesn't help19:31
→ mxweas_ joined19:31
cmn no, what does work is git checkout -t origin/master19:31
← valtih left19:31
cmn was miremembering then19:32
← diogogmt left19:32
SethRobertson How would that work if master already existed?19:32
fatal: A branch named 'master' already exists.19:32
You probably wouldn't need to even use -t in that circumstance.19:32
← advatar left19:32
→ SegFaultAX joined19:32
SethRobertson (assuming master didn't exist I mean)19:33
tango_ --set-upstream19:33
cmn I changed the line a bit19:33
tango_ (if it exists)19:33
cmn 'cos I'm testing in /tmp and .. isn't a repo19:33
SethRobertson No, it errors out19:33
→ lyddonb joined19:33
← ASIRINeiluj left19:33
← mxweas_ left19:34
cmn it works if master is unborn19:34
→ brian_g joined19:35
cmn but yeah, I was thinking of something different19:35
SethRobertson You could also say `git checkout master` and it would work (if master was unborn)19:35
ereslibre_laptop → ereslibre19:35
cmn but only for origin19:35
← zagra left19:35
SethRobertson Seems probable19:36
→ OOPMan joined19:36
cmn oh, apparently not19:36
it works just fine if the remote is called "other"19:36
← shruggar left19:37
← samnob left19:38
cmn it just needs to be the only remote19:38
← indstry left19:41
stockholm http://paste.debian.net/156942/19:41
SethRobertson: what does it want from me?19:41
SethRobertson stockholm: Huh?19:41
→ apok joined19:43
stockholm SethRobertson: do you understand the paste?19:43
→ johnhamelink joined19:44
SethRobertson Sorry, I missed the paste. Run `git mergetool`19:44
→ sdeboey joined19:44
SethRobertson Or decide whether you want to delete that file or keep it19:44
→ fedesilva joined19:44
cbreak or just read what status says and resolve it19:44
← Cromulent left19:45
wereHamster stockholm: line 21 to be precise19:45
→ shruggar joined19:47
→ cmyk_ joined19:49
→ _marvin joined19:50
→ Element9 joined19:50
→ mastro joined19:51
← traviscj left19:52
→ savant joined19:52
savant hi all19:52
I'm rewriting the history of a project19:53
(yes, very bad)19:53
to remove all files not in a specific whitelist.19:53
SethRobertson git-filter-branch19:53
savant is there a way, other than specifying the files I want to delete, to delete all files not in the whitelist?19:53
cbreak write a script19:54
→ zulax joined19:54
SethRobertson git ls-tree -r HEAD | egrep -v whitelist19:54
stockholm SethRobertson: now i decided to remove the file, and i think the merge is happy now19:54
cmn rm --cached *; git add <whitelist>19:54
cbreak or git reset HEAD instead of add19:54
SethRobertson cmn++19:54
stockholm SethRobertson: now i push?19:54
SethRobertson: how do i push?19:54
← garpenlov left19:54
SethRobertson stockholm: No, what you do now is fork the new upstream project branch19:54
stockholm ah19:56
→ traviscj joined19:56
stockholm hu19:56
SethRobertson: can you explain why? i dont get it19:57
where in the process are we just now?19:57
← PigeonFriend left19:58
SethRobertson stockholm: As I understand it, you now have a branch which consists of the true upstream with your patches after their commits19:58
stockholm: Now, you want to submit your patches to them and continue to track their repo while doing so19:58
stockholm: So, in order to use the github patch passing system, create a fork of their repo.19:59
stockholm SethRobertson: i have a local branch of their repo, with my patches applied, right? havent i created a fork already, in that way?19:59
→ kerrick joined20:00
stockholm i am an owner of the project, so i should be able to push, right?20:00
SethRobertson You have an git-only fork, which is not associated with the true upstream using github's methods. If you want to pass the patches via github (the expected way) you need to create a true github fork20:00
← katspaugh left20:00
SethRobertson You *can* push your changes to your repo, `git push -f origin mybranch:master` It just isn't going to get you where you want to go20:01
→ Jitse joined20:01
gitinfo set mode: +v20:01
→ xsbeats joined20:02
stockholm SethRobertson: ok, so how do i create that fork?20:02
← skitrees left20:02
Jitse hey guys, i've made a git user, gave him /usr/bin/git-shell, added two keys to the authorized_keys, added git to the AllowUsers option of sshd_config, restarted the ssh daemon; but now the server just tells me i can't ssh with the git user, as i have a wrong key20:02
SethRobertson stockholm: Go to the upstream github page, where you got the URL, and there should be a button on it to Fork20:03
cmn Jitse: does your ssh know to send the right key?20:03
← xsbeats left20:04
Jitse yes, when i do ssh -vT [email@hidden.address] he tells me he's using my default ~/.ssh/id_rsa.pub and id_rsa20:04
i'm on linux now ^^20:04
stockholm found it, i think20:04
← brian_g left20:04
cmn and it logs you in?20:04
Jitse no it does not, it tries the keys and then says no more authentication methods to try20:04
permission denied (publickey)20:04
i'll pastebin my console output20:05
SethRobertson stockholm: Once you fork, you will want to find the read-write URL for your new repo. You will then once again `git remote add myfork url-of-new-repo; git remote update`20:06
← alberto56 left20:06
cmn then that's not the right key20:06
SethRobertson stockholm: Then you can run `git log myfork/master...mybranch` and ensure that only the commits you just cherry-picked are listed as being ready to be sent. Also that `git log myfork/master...mybranch` returns nothing.20:07
stockholm SethRobertson: on it...20:07
→ brian_g joined20:08
← arosen left20:08
SethRobertson stockholm: Then you say: `git push myfork mybranch:master` which will upload your changes, then use the github web ui to submit a "pull request"20:08
← brian_g left20:08
→ arosen joined20:08
Jitse this is my console output: http://pastebin.com/RSSK7JQ020:08
← dc5ala left20:09
← krz left20:09
cmn that doesn't say anything other than the server refusing your key20:09
← ipalaus_ left20:09
Jitse *cries*20:09
i know20:09
→ Yuuhi joined20:09
cmn more levels of debug might help more20:10
depending on what the problem actually is20:10
stockholm SethRobertson: hu?20:10
~/src/fai/trunk/bin(othermaster)$ git push myfork othermaster:master20:10
Everything up-to-date20:10
Jitse how do i get more levels of debug? :p20:10
cmn make sure that key is indeed in authorised_keys20:10
Jitse i double-checked that20:10
← g0bl1n left20:10
Jitse it's just one line per key right?20:10
cmn you give it more v's20:10
a key is one line, yes20:10
SethRobertson stockholm: But `git log myfork/master...othermaster` returned stuff?20:10
cmn have you checked to see what the server's auth logs say?20:11
Jitse let me have a look20:11
stockholm http://paste.debian.net/156953/20:11
no, git log myfork/master...othermaster does return nothing20:12
Jitse oh crap >.< "git disallowed because account is blocked"20:12
how do i unlock it without giving it a password?20:12
locked*20:12
cmn don't know20:13
I let sitaram figure it out and use gitolite20:13
SethRobertson stockholm: `git branch` Look for the line with the *. What is that name?20:13
→ scelis joined20:13
stockholm * othermaster20:13
→ LordCrc joined20:14
stockholm SethRobertson: ^20:14
SethRobertson Use othermaster instead of mybranch20:14
→ rickroberts_ joined20:14
stockholm ah20:14
someprimetime i initialized a bare repo on my server (added the folder on the server that i want to commit as an origin), created a local repo and added that bare repo as an origin, now when i go into my local repo and go git pull i get `Your configuration specifies to merge with the ref 'refs/head/master' from the remote, but no such ref was fetched.` here is my .git/config on my local machine: http://pastebin.com/ahS3QJyW20:15
LordCrc hi, i just made a commit to a local repository, and want to undo it as i forgot some comments in one file... how do i undo the commit (ie ala hg rollback)?20:15
stockholm SethRobertson: you used mybranch and myfork20:15
someprimetime basically i want to pull everything i committed from the server to my local branch now20:15
Jitse whippitydoo20:15
stockholm SethRobertson: "git push myfork othermaster:master" ?20:15
SethRobertson stockholm: Yes, because I either didn't know or didn't remember the branch name you used20:15
stockholm: Yes, that20:15
stockholm that does nothing20:15
Jitse cmn: now it tells me interactive git shell is not enabled, git-shell-commands should exist and have read and execute access20:15
cmn Jitse: man git shell then20:16
gitinfo Jitse: the git-shell manpage is available at http://jk.gs/git-shell.html20:16
stockholm that is what i did20:16
http://paste.debian.net/156953/20:16
← zulax left20:16
SethRobertson Run the logs. `git log otherbranch...myfork/master` `git log myfork/master...otherbranch`20:17
← SegFaultAX left20:17
→ skitrees joined20:18
scelis sitaram: have you ever considered letting people create and edit shared hooks directly from the gitolite-admin repo?20:18
← harshpb left20:18
← rsaaguy left20:19
stockholm SethRobertson: i assume otherbranch == othermaster for me20:19
SethRobertson: and that gives nothing20:19
← Lgb left20:20
SethRobertson stockholm: Um... run `gitk --all --date-order` and look for those commits you cherry-picked20:20
→ DarkAR joined20:21
→ Lgb joined20:21
→ RCMatt joined20:22
→ akiress joined20:22
→ aspotashev joined20:23
← Targen left20:24
gv hey everyone, i just wanted to quickly announce a git subcommand i wrote20:25
it's for those days when you find yourself sending out a ton of cgit/gitweb/etc links ...20:25
→ digitalfiz joined20:25
← digitalfiz left20:25
→ digitalfiz joined20:25
gv ... and you just can't take the clicking any more20:25
https://github.com/gvalkov/git-link20:25
→ Targen joined20:26
gv </end-of-shameless-plug>20:26
← Gitzilla left20:26
← LordCrc left20:27
← Talvino left20:27
← DarkAR left20:27
→ ploop joined20:27
stockholm doh, i dont get it, the merging seems to have gone wrong...20:28
SethRobertson stockholm: You should not have merged yet, unless you were talking about cherry-picking20:29
→ sebrock joined20:29
stockholm yes, that is what i meant20:29
i thought that was a merge20:29
← Davey_ left20:30
→ xsbeats joined20:30
← griotspeak left20:30
SethRobertson No. You can only merge with common merge-bases, and since the SHA of the old commits were different, we did not merge20:30
CareBear\ hallå stockholm20:30
stockholm hej CareBear\20:31
SethRobertson stockholm: So, you cannot locate those commits you cherry-picked?20:31
→ rue joined20:31
→ _7mj joined20:31
stockholm SethRobertson: i see them, but they are in a seperate branch like before i cherrypicked20:31
SethRobertson Right. You see the original commits you made but not the ones created as a result of the cherry-pick20:32
?20:32
stockholm yes20:32
SethRobertson what does `git status` say?20:32
stockholm SethRobertson: nothing, all is fine.20:33
→ oriba joined20:33
← localhost left20:33
stockholm and the files that i worked on are not in the othermaster branch, but only in my own20:33
SethRobertson Well, I can think of nothing to do other than try the cherry-picks again.20:33
← sebrock left20:33
stockholm where i had them origianlly20:33
SethRobertson stockholm: Those files are not committed into git if you see them now20:33
The ones listed under "Untracked files"20:34
← lyddonb left20:34
← prudnikov left20:34
→ prudnikov joined20:35
SethRobertson If the files are not under untracked files, please paste the git-status20:35
→ localhost joined20:35
← __class__ left20:36
→ lyddonb joined20:36
→ psoo joined20:36
→ __class__ joined20:37
← xsbeats left20:37
→ pfFred joined20:37
← pfFredd left20:38
diegoviola how do i show changes that hasn't been commited yet?20:39
committed20:39
Enlik diegoviola: git status20:40
Jitse scp -r proj.git [email@hidden.address] -> fatal: unrecognized command 'scp -r -t -- /home/git/proj.git' lost connection20:40
shruggar "git diff" or "git diff --cached"20:40
Jitse what's wrong with that command guys?20:40
← dpino left20:40
shruggar depending on if they have been staged yet20:40
→ xsbeats joined20:41
shruggar Jitse: is that supposed to be a git-related question?20:41
if so, then there's plenty wrong with it :)20:41
Jitse kind of, i'm copying a git repo ;-)20:41
a site's telling me that's how you start an empty repo20:41
diegoviola thanks20:41
Jitse mm wait a sec, how could that work, git has the git-shell only20:42
alrighty that's fixed, ignore my dumb questions20:43
→ sroy2 joined20:44
cmn why not push?20:44
Jitse cuz cuz...20:44
ChanServ set mode: +v20:44
← rgsteele left20:44
→ jsick3 joined20:45
→ ipalaus_ joined20:45
Jitse i cloned it now, it tells me i cloned an empty repo, as in there haven't been any commits yet; now do i have to do anything to make it a proper working repo?20:46
sroy2 Hi, any advice when "git pull" returns "Already up-to-date." when you know for a fact that there are files in the repo that you don't have?20:46
Jitse for example do a first commit before it works, or create the master branch?20:46
← infogulch left20:46
cmn Jitse: before it works?20:46
→ notola joined20:46
← notola left20:46
cmn do you have a git repo already or not?20:47
Jitse yes, and i can clone it to my client, it then tells me it's an empty repo20:47
← PerlPilot left20:47
Jitse but it's been cloned20:47
cmn because it's empty20:47
Jitse yup20:47
however20:47
cmn but that's not what you want to do20:47
you want to push to that repo on the server20:47
Jitse oh!20:48
it works!20:48
i just pushed to the master20:48
→ PerlJam joined20:48
Jitse let's see if i can clone this on windows now20:48
wooptidooo20:49
→ LucasTizma joined20:49
Jitse why do people recommend gitolite or others nowadays? something about not having to use ssh directly?20:49
or is git-shell just as safe either way20:49
cmn you do use ssh20:50
gitolite simplifies user, repo and permission management20:50
Jitse right fair enough20:50
cmn and gives a unified hook set20:50
Jitse which i won't really need yet, as i'm the only user20:50
cmn and many more things20:50
→ HedgeMage joined20:50
Jitse thanks for the info!20:50
cmn it's very useful even for one user20:51
← Raging_Hog left20:51
Jitse why is that?20:51
if i won't really need anything else but fetching, committing, pushing, merging20:51
sroy2 meh nvm - had to reset the head back a ways in the repo to get it to work… :(20:52
→ Gitzilla joined20:52
cmn for the reasons I've listed20:52
Jitse fair enough20:53
cmn it also provides a way to set each repo's options20:53
← traviscj left20:53
PerlJam Jitse: if you're the only user, you can protect yourself from your own mistakes with gitolite and a nice config file.20:53
→ zackiv31 joined20:54
← xsbeats left20:55
→ stringoO joined20:56
← akiress left20:56
← flyz left20:56
HedgeMage So, if I have a git repo with a branch A with commits a1, a2, and a3, and a branch B that forks at a3 and has some commits of its own (b1, b2) what happens to branch B when I'm on HEAD of branch A (a3) and I do "git reset --hard HEAD^" ?20:56
→ mishok13 joined20:56
Jitse alright20:56
→ xsbeats joined20:56
→ akiress joined20:57
cmn HedgeMage: nothing20:57
you're modifying branch A20:57
← h0bbit left20:57
→ yeryry joined20:57
→ traviscj joined20:57
HedgeMage cmn: right, but I'm making a3 go away, right? so then does it only exist on branch B from that point, or did I make something explode in a way I don't expect?20:57
cmn you're not making a3 go away20:58
shruggar then it's as if you branched from a220:58
HedgeMage just wants to make sure she doesn't destroy anything important20:58
shruggar and one of the commits unique to B is a320:58
cmn you're setting branch A to a220:58
HedgeMage shruggar: thank you, that's what I hoped!20:58
someprimetime if i created a bare repo on my server, do i have to clone the directory that contains .git within that bare repo?20:58
← ipalaus_ left20:58
someprimetime the directory that contains .git would be where all my files live20:58
cmn a bare repo doesn't have a .git within20:58
someprimetime i know20:58
that's why im wondering if within my bare repo i clone the directory that has the .git in it20:59
← jonesy left20:59
cmn what do you mean within your bare repo?20:59
you clone a repo, not a directory20:59
what are you trying to achieve?20:59
→ EricInBNE joined21:00
← aristidesfl left21:00
someprimetime cmn: im almost setup. i want to edit files locally and push those to my bare repo on mym server.. i set up the bare repo in ~/git/ then locally i did the same thing and then i went into the folder on my server where all my files are that i want to work with and did a git init within it21:01
→ dcarper joined21:01
someprimetime i added the origin of the bare repo locally but when i do git pull it doesn't get anything21:01
→ reliability joined21:01
reliability hi21:01
someprimetime well, it just gives me this: `Your configuration specifies to merge with the ref 'master' from the remote, but no such ref was fetched.`21:01
← dcarper left21:02
someprimetime this is in my .git/config locally: http://pastebin.com/ahS3QJyW21:02
cmn are you trying to pull from the bare repo?21:02
→ dorkmafia joined21:02
someprimetime cmn: yes21:02
cmn that won't work21:02
by "from" I mean whilst in the bare repo21:02
someprimetime well i went into the directory that contains all my site files and did git init21:02
then git commit21:02
comitted all of them21:02
and then i went back into the bare repo and added that as an origin21:03
so i assume that the bare repo would now contain all those files i committed from the server21:03
cmn why?21:03
the bare repo is the one that serves as a sync point21:03
someprimetime well so i could pull them locally from that bare repo21:03
oh21:03
cmn so you need to push to it21:03
!pull21:03
gitinfo pull=fetch+merge (or with flags/config also fetch+rebase). It is thus *not* the opposite of push in any sense. A good article that explains the difference between fetch and pull: http://longair.net/blog/2009/04/16/git-fetch-and-merge/21:03
someprimetime oh what21:03
ok21:03
Jitse what are the correct permissions for .ssh with authorized_keys for the git user? 600 makes it go awry21:04
← thebignoise left21:04
cmn .ssh should probably be rw by the user21:05
the dir itself would be x as well, presumably21:05
and anything under would be rw21:05
someprimetime cmn: oh shit21:06
it works21:06
i just did git push origin master within my /app directory that contains all my site files21:06
and it seems to have pushed to the bare repo21:06
Jitse 700 seems to be ok21:06
← savant left21:06
someprimetime and now i did git pull locally and it retrieved the objects21:06
cmn if you set 600 to .ssh/ you probably can't list the files in it21:06
someprimetime: yes, doing things in the right order does help21:06
someprimetime lol21:06
hell yeah man21:06
FINALLY got it21:06
← rickroberts_ left21:07
← sroy2 left21:07
← marthinal left21:07
Jitse this is jizztastic21:07
← glcrazy left21:08
→ thebignoise joined21:08
Jitse cheers guys21:09
← Jitse left21:09
← xsbeats left21:10
→ xsbeats joined21:12
→ tarwich joined21:12
tarwich Will git clone + git commit work between two local folders?21:13
cmn commit is only local21:13
a clone will work in the same filesystem21:13
or any other local filesystem, for that matter21:14
→ tchan joined21:14
tarwich But can the clone (B) of A commit to A and A see the updates such that another clone (C ) could see them?21:15
cmn the clone of a commit? what do you mean?21:15
to move changes between repos you need to push or fetch21:15
you might be looking for the new-workdir script in contrib/21:15
tarwich Ahh...21:15
Thanks. I'm a recent SVN convert and thought commit pushed.21:16
cmn no, git is distributed21:16
← Jackneill left21:16
cmn committing is a different step from pushing out changes21:16
tarwich I see, so then I'd have to cd to A and git pull from B, then21:17
cmn that's one way21:17
tarwich LOL That's short for "But not a good one" :-)21:17
cmn but having many repos locally isn't usually a good workflow21:17
many repos of the same project, that is21:18
tarwich Yes… It's just temporary while I'm waiting for our IT dept to get the GIT server online.21:18
cmn that still doesn't explain why you want three repos21:18
tarwich Ha ha ha… You're right…. In part I was just familiarizing myself with GIT.21:19
But also, we're using a Synchronization service to push all files from one folder (A) to several computers on the network.21:19
cmn a synchronisation service? why?21:20
that sounds more like a deployment system21:20
tarwich You're right… It's not a good idea… But it's temporary.21:20
← Destos left21:20
tarwich Office policy is that all work has to be in DropBox.21:20
→ Destos joined21:21
← w09x left21:21
PerlJam wow21:21
tarwich However, we need GIT for development because Dropbox is not working right.21:21
PerlJam double wow21:21
← skoop left21:21
tarwich *shamed*21:21
cmn no wonder21:21
Dropbox isn't for development21:21
← gv left21:22
cmn you can use dropbox to put a few files for distribution, but I'm weary of any company policy that sepcifies that someone else must have all the data21:22
Valodim dropbox had some fun security desasters in its time21:22
PerlJam not to mention the security implications21:23
tarwich LOL… And don't forget security21:23
PerlJam okay, well Valodim mentioned them :)21:23
← xsbeats left21:23
cmn right, that's kinda part of the "someone else must have all the data"21:23
→ pandora17 joined21:23
PerlJam tarwich: how many people are supposed to follow this "office policy"?21:24
Valodim company policy to put files in dropbox == lack of a company policy to read up on how to properly use git before a couple of files were lost due to misuse of git21:24
PerlJam (I can kinda see how 2 or 3 people could come up with this and continue to use it)21:24
← lyddonb left21:24
tarwich Started as 3… Now we have 621:24
PerlJam ah, so you're just experiencing growing pains21:25
tarwich Guys thanks. I now see how it works… So we're definitely going to need that server up.21:25
But I get it now. PerlJam Valodim you rock21:26
SethRobertson Also, read !book21:26
gitinfo There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !vcbe and !parable21:26
→ lyddonb joined21:27
rudi_s !vcbe21:28
gitinfo 'Version Control By Example' gives a good overview of the different VCSes available. The author will even mail you a dead-tree copy for free. http://ericsink.com/vcbe/21:28
SethRobertson You can also /msg gitinfo those keywords to avoid spamming hte channel21:29
rudi_s Sorry, will do.21:29
← iamjarvo left21:31
← smw left21:33
← wroathe left21:34
→ Darren joined21:35
← RCMatt left21:35
→ Diniden joined21:36
← jsick3 left21:38
← shtrb left21:39
← mbroeker left21:40
← Darren left21:40
← olinux left21:41
someprimetime i added a new file in my local repo.. pushed it to master within the bare repo and set up a post-receive hook in the bare repo to deploy the changes to the website whenever the central bare repo is updated, but i'm not able to find the file blah.php anywhere on the server21:43
blah.php is the new file i made on my local repo fwiw21:43
← sdeboey left21:43
→ sdeboey joined21:44
PerlJam someprimetime: so ... time for you to debug your deployment script21:44
→ giallu joined21:44
→ NiklasFiekas1 joined21:44
← NiklasFiekas left21:45
→ murr4y joined21:45
→ linusoleander joined21:45
SethRobertson stockholm: You have been quiet21:46
← krisfremen left21:47
→ krisfremen joined21:47
← _iron left21:47
← NiklasFiekas1 left21:47
← stuntmachine left21:48
← Ksugzulut left21:48
← kerrick left21:49
reliability Suppose I have checked out a specific branch different from the master branch and now I want to merge a remote master branch into the master branch without checking out the master branch. How do I do that?21:49
← stringoO left21:49
← Black left21:50
Valodim is it a fast forward merge?21:50
← [M]ax left21:50
reliability yes21:50
Valodim you can just git fetch then, and update your master ref to orgin/master's commit21:51
reliability The thing is, someone made changes to a public repository and i want to merge that repositoy with my local one but while doing that I don't want to switch from the specific branch into the master branch.21:51
Valodim what's the problem with switching branches?21:52
→ [M]ax joined21:52
reliability It's php code that's being in use...21:52
← bgerber left21:52
Valodim ..git clone it somewhere else?21:53
canton7 you can git branch -f master origin/master, but of course you risk throwing away commits if you're not careful. Is there any reason why master has to be up to date?21:53
→ green- joined21:53
green- I'm a bit confused as to what happened to some changes in my working copy. I have a ton of changes, which are intact, but git no longer sees them as changes waiting to be added or committed.21:54
← dougsland left21:54
green- I ran a git pull to grab some changes another user made on some other files, and now my waiting changes no longer show via git diff or git status, etc.21:54
→ garpenlov joined21:55
green- The last 'git pull' resulted in a merge with the master, and I'm wondering if that has anything to do with it.21:55
What am I missing?21:55
oh, git log --name-status also does not show the changed files as having their changes committed at any time21:55
reliability locally, no changes have been made to neither master nor the specificbranch. So I could just merge origin/master with master and then checkout some specific files from master, right?21:55
← gchristensen left21:56
→ bgerber joined21:56
→ Galaxor joined21:56
Galaxor I'd like to see some abandoned commits... stuff that was on a branch that's been deleted. How do I do that?21:57
canton7 green-, git deon't let you merge (and hence pull) if you have uncommitted changes21:57
→ linoj joined21:58
canton7 Galaxor, git reflog or !gka21:58
gitinfo Galaxor: For a better way to view the reflog, try: gka() { gitk --all $(git log -g --format="%h" -50) "$@"; }; gka21:58
someprimetime PerlJam: thanks got it21:58
nice username btw ;D21:58
green- canton7: then i can't make sense of what's going on. the changes were absolutely not committed. i've been coding all day and haven't committed.21:58
canton7 green-, did you stash them?21:58
green- canton7: if i did, i have no idea how i did21:59
canton7 green-, did you run anything like "git checkout -- <anything>", "git reset --hard", "git checkout -f", etc?21:59
green- canton7: nope. and the changes are still in the files. nothing got wiped out. git just doesn't see them as changes.22:00
← joshsmith left22:00
← sdeboey left22:00
reliability so, isn't it possible to merge master and origin/master without touch the currently checkouted branch?22:00
← notjohn left22:00
canton7 green-, aah, if they're definitely present, then they were committed somewhere along the line. git's probably hiding the commit that introduced them because it thinks that commit is uninteresting -- a merge, for example22:00
green-, did you have any conflicting merges?22:01
green- canton7: yes22:01
canton7 green-, did you accidentally add your untracked files?22:01
green- they weren't untracked, they were tracked22:01
canton7 reliability, you can only merge into the currently-checked-out branch. You can forcefully update master to point at origin/master, as I said22:01
green- how do i get it to stop hiding the details in git log so I can see when and if it put them in somehow22:01
if it did, i can't imagine why it would commit them without giving me the opportunity to enter notes22:02
canton7 green-, run git glame / git gui blame on the file?22:02
← tarwich left22:02
← rendar left22:03
→ DarkAR joined22:03
← drizzd left22:03
green- not sure what i'm looking at22:03
← milli left22:03
canton7 green-, you're after oen of the lines which you don't think should be committed yet. The info on the left will tell you the commit in which that line was introduced22:04
← eijk_ left22:04
→ gchristensen joined22:04
← akosikeno left22:05
→ eijk_ joined22:05
→ alberto56 joined22:06
← nixmaniack left22:06
canton7 green-, http://thread.gmane.org/gmane.comp.version-control.git/189629 might also be relevant22:06
Galaxor canton7: I can't figure out what to put in git reflog --expire-unreachable=X22:07
green- canton7: it looks as if they may have gone in during the recent merge, as you suggested. but ... those changes were never added, so why would it commit them .. it also gave no opportunity to enter notes .. and why wouldn't it show them in 'git log --name-status' ?22:07
Galaxor canton7: the manpage says <time>, which defaults to 30 days. It didn't accept a number, like 90, and it didn't accept an isodate, like "2012-01-01 00:00:00"22:08
"unrecognized argument".22:08
canton7 green-, unless i'm missing something, the merge should never have happened if you have unstaged changes anyway... Maybe someone else has an answer to that22:08
Galaxor, what are you trying to do? Why are you trying to expire the reflog?22:08
green- canton7: i can see the changes via "git diff" if i do "git diff origin" ... but otherwise, they don't show. does that provide any insight into what might be going on?22:10
→ panthrodeb joined22:11
→ iocor joined22:11
Galaxor canton7: Oh. That's not what I was trying to do... I thought that meant that it was going to only show me unreachable commits newer than 30 days. I want to see older unreachable commits.22:11
← diegoviola left22:12
canton7 green-, does adding -c, --cc, -m, or --full-history to git log help?22:12
→ milli joined22:13
canton7 Galaxor, you'll just have to scroll back further I'm afraid. By default, reflog entries hang around for 90 days iirc22:13
green- canton7: yep, -m did. and it shows those files as merged/commited during that last log item22:13
← arosen left22:13
→ arosen joined22:13
green- canton7: but i don't get it, i didn't run git commit ... i ran git pull22:13
canton7 green-, did that merge conflict? I know you said one of them did22:14
green- canton7: one file conflicted22:14
(not one of the files at issue in this current problem)22:14
canton7 so, you'll have run "git commit" to resolve the conflict22:14
Galaxor canton7: Dang. I think the commit I want isn't in there, then. Maybe I imagined it to begin with.22:14
canton7 Galaxor, !gka might make the searching easier22:15
gitinfo Galaxor: For a better way to view the reflog, try: gka() { gitk --all $(git log -g --format="%h" -50) "$@"; }; gka22:15
← v0n left22:15
green- canton7: sure, for the conflicted file .. that doesn't explain why it committed all those other files when the changes weren't added ... it even deleted a file I never used git to remove ... just accidentally use a filesystem mv to rename it22:15
and it deleted that file22:15
i'm totally confused22:15
canton7 green-, same. I have no idea why git let you start the merge.22:16
green-, you definitely just 'git added' that one file, and didn't 'git add .' or 'git add -u'?22:17
anyone: does anyone know why git might let the user start a merge with uncommitted changes (to tracked files)22:17
green- canton7: i didn't git add anything22:17
← _7mj left22:18
canton7 green-, then how did you resolve the merge conflcit?22:18
green- the one file is still in conflict ... i ran 'git pull' and it notified that the one file was in conflict and committed un-added changes without my asking it to22:18
canton7 wait.. so you're in the middle of a merge at the moment?22:18
green- ok, scratch that22:19
ok so here's what happened22:19
scrolling back through buffer22:19
→ shibayu36 joined22:20
green- so i ran "git pull" and it complained that a file was in conflict and that it wouldn't let me pull until i fixed that22:20
canton7 what *exactly* did it say?22:20
green- so i didn't want changes to get lost ... so i ran (apparently) a git commit -a22:20
then it committed these changes without asking for notes and said22:20
master e25236d] Merge branch 'master' of file:///srv/git/njwxnet22:21
canton7 ah, right I know what happened, I think22:21
green- so instead of giving me a chance to enter details of the commit22:21
it just spit out that stuff about the merge22:21
← someprimetime left22:21
green- leaving the one file in conflict (where it remains)22:21
canton7 before you ran "git pull", you'd previously started another merge. that merge had conflicted, and you'd left it in a conflicted state and carried on working22:21
green- i guess, maybe22:22
← t4nkd left22:22
canton7 when you ran "git pull" a second time, git moaned that you were still in the middle of the previous merge, with conflicted files22:22
← gchristensen left22:22
canton7 so you ran "commit -a". Running commit in the middle of a conflicted merge resolves the conflicts and completes the mereg22:22
green- well the conflict still exists22:22
canton7 which is why it didn't ask you for a message, but rather finished the previous merge22:22
→ smoak joined22:23
green- heh, confusing22:23
← bgerber left22:23
green- here's how it went22:23
canton7 (adding your modifications into the merge commit, creating the evil merge you're currently having a problem with)22:23
← linduxed left22:23
gitinfo set mode: +v22:23
← Shakyj left22:23
→ bgerber joined22:24
green- tried a pull, it complained a file (import.xml) would be overwritten that was changed, so commit that file ... i commited that single file and then tried pull again ... it then tried merging import.xml and failed and said to fix conflicts ... then i ran git commit -a to get the other stuff in22:24
that is when this odd merge happened22:24
i think :)22:24
→ Shakyj joined22:24
green- import.xml remains in conflict22:24
← [M]ax left22:24
smoak im trying to squash 3 commits down into one commit using git rebase -i HEAD~3 but the commits are listed in reverse order (ie oldest first), is this normal and how its supposed to work?22:25
canton7 smoak, yup22:25
green- i'd like to roll back the delete it did so that I can git mv the one file properly so i don't lose change tracking22:25
can i do that?22:25
canton7 green-, well for one thing, git doesn't track moves, it works them out after-the-fact22:26
→ caseymcg joined22:26
canton7 that said, you'd do well to undo that dirty merge22:26
smoak canton7: but then if i try to squash the oldest (ie the top line) it says "Cannot 'squash' without a previous commit'22:26
canton7 smoak, yeah, you can't use squash on the top line. use git rebase -i HEAD~4 instead22:26
cespare smoak: change the second and third to squash if you want to squash all three together.22:27
→ [M]ax joined22:27
smoak ok thank you i will try that22:27
← maletor left22:28
canton7 green-, I'd identify the first commit you made prior to those messed-up merges, and git commit --mixed <commit> to take those merges out of history. Then sort your files out, and make some sensibe merges22:28
assuming you haven't pushed any of those merges, of course22:28
green- canton7: haven't pushed them22:29
canton7: can i just rollback that entire last commit?22:29
canton7 green-, yeah, reset --mixed (or --soft, DO NOT use --hard) to the commit before the merge commit22:29
→ maletor joined22:30
← lyddonb left22:30
→ gchristensen joined22:30
→ |Agent joined22:31
← LucasTizma left22:31
canton7 green-, of course, I'd recommend !backup if you haven't done history modification before22:31
gitinfo green-: Taking a backup of a git repository is always a good idea, especially when taking advice over IRC. Usually, the best way to TACTICALLY back up a git repo is `git clone --mirror`. However, some unusual maintenance might require `tar cf repo-backup.tar repodir`. Testing in a clone is also an excellent idea. See also http://sethrobertson.github.com/GitBestPractices/#backups22:31
→ Zenopus joined22:31
← juanfer left22:32
→ nixmaniack joined22:33
green- canton7: i'm still trying to make sense of what happened. things seem to have been split into 2 commits with merges involved. I'm clueless. Here's the relevant log entries: http://pastie.org/342348422:33
if it isn't obvious, I'm new to git :) coming from svn22:34
|Agent I merged a local branch into my local master then pushed my local master to origin/master (origin is a GitHub account). I didn't actually want to do this. I want my origin/master to come from upstream/master (upstream is the project's main repo).22:34
canton7 green-, yeah, those look like two evil merges (merges which introduce new changes)22:34
|Agent So I reverted the merge and pushed the revert up to origin/master. But now I can't fetch upstream/master and push it to origin/master to make my GitHub account up-to-date, because that loses history. How can I do this?22:34
green- canton7: why are they evil?22:34
canton7 green-, I'd need to see an exact scrollback (complete with the messages git gave you) to be abel to tell exactly what happened22:34
green-, an evil merge is a merge which introduces changes which weren't present in either of the branches being merged22:35
|Agent (er, not fetch, but rather rebase to upstream/master and push that to origin/master)22:35
canton7 green-, in your case cause by, I believe, you doing a pull, leaving it in a conflicted state, and continuing to work22:35
green- canton7: and what leads you to believe it has introduced those false changes?22:35
canton7: but only 1 file was conflicted, and that file remains in conflict22:35
canton7 green-, define "in conflict". How are you determining that it's in conflict?22:36
green- git says so, right in those log entries22:36
← digitalfiz left22:36
→ joshsmith joined22:36
canton7 green-, those were conflicts which happened during the merge, and which were resolved during the merge. I thought you said that the file was *still* in conflict?22:37
→ pfFreddd joined22:37
→ linduxed joined22:37
dorkmafia is there a way to run git with group permissions?22:37
→ ryanisnan joined22:37
dorkmafia i have two users using the same git directory22:38
← danmackay left22:38
green- canton7: i didn't know that indicated it was resolved22:39
← cbeams left22:39
green- i thought the log indicated it was in conflict so nothing was done22:39
canton7 green-, ah ok. Git doesn't let you finish the merge unless everything's resolved.22:39
green- i updated the paste with commands: http://pastie.org/342348422:39
← pfFred left22:39
green- if git introduced changes into import.xml, i can live with that, easily fixed22:39
i'm only concerned about the other files22:39
← linoj left22:40
← oc80z left22:41
→ oc80z joined22:41
← shibayu36 left22:41
canton7 green-, when abouts, relative to those 4 commands you pasted, were the modifications to import/datecode/mesowest, etc, made?22:41
green- i believe the other developer just committed those recently22:41
→ danmackay joined22:41
canton7 green-, oh! so wait... The files you had uncommitted changes in aren't listed in lines 29-37?22:42
← Sadin_ left22:42
dorkmafia anyone know if there is a git configuration to run the directory as a group instead of a user? :)22:42
← linduxed left22:43
canton7 dorkmafia, look at --shared for man git-init, and core.sharedRepository in man git-config. I can't remember exactly, but I *think* you have to re-run git init --shared (which doesn't remove any data)22:43
gitinfo dorkmafia: the git-init manpage is available at http://jk.gs/git-init.html22:43
→ TrineLocke joined22:43
dorkmafia tyty22:43
axl_ what is the command to get rid of local changes?22:44
← mishok13 left22:44
canton7 axl_, git status will tell you22:44
dorkmafia now to figure out how to get hudson/jenkins to do this by default :(22:44
← CannedCorn left22:44
← |Agent left22:45
← nixmaniack left22:45
← TrineLocke left22:46
← shruggar left22:46
← DarkAR left22:47
← [M]ax left22:48
→ mxweas_ joined22:49
→ zulax joined22:49
green- canton7: no, they are ... all the stuff in lib/perl mostly .. it did go in during that merge22:49
they're in .. so that's good ... there are 2 issues as far as i can see22:49
← alexgordon left22:50
canton7 green-, the fact that they're committed during a merge is *not* good -- it's very confusing, as you've found out22:50
→ [M]ax joined22:50
← ThomasLocke left22:50
canton7 green-, if the changes were made between line 9 and line 15, I can understand what happened perfectly. If not... I'm confused22:51
green- 1) import.xml is botched and looks like it has conflict queues in it (stuff like "HEAD" and commit IDs and 2) the file Importer.pm got deleted from tracking because i used a filesystem mv instead of a git mv22:51
← johnkpaul left22:51
axl_ is there a git command to get a list of all untracked files?22:51
canton7 green-, as I said, "git mv" === "mv file && git add newfile && git rm oldfile". Git doesn't track moves22:51
← cakehero left22:51
green- canton7: that was misspeaking on my part, they were not committed during a merge, or i don't think .. they went in w/ commit -a as i did run that22:51
but i can't tell why tehre are two log entries22:51
← psoo left22:51
→ alexgordon joined22:51
axl_ i was trying to use git ls-files but did not find an appropriate option22:51
← Chowzzf left22:52
green- one maybe from a pull merge and one from the commit? i don't know why some files went in the first time and some the second22:52
canton7 green-, I think the two log entries are as a result of one of the flags you passed to git log -- it's showing you the commit e252 with respect to each of its parents22:52
→ angelsl joined22:52
canton7 green-, notice that the two log entries are both for the same commit22:52
→ Chowzzf joined22:52
← Jay_Levitt left22:53
← angelsl left22:53
green- that was 'git log --name-status -m'22:53
← j416 left22:53
canton7 green-, probably the -m then22:53
→ cakehero joined22:53
axl_ ok i got it: git ls-files --other --exclude-standard22:53
← cakehero left22:53
green- there's mentions of merges, but i'm not clear why ... the only file that should have needed to be merged was import.xml22:53
canton7 green-, what I would do to fix this: First cp the repo to make a backup of it. Then git reset --mixed bb913af to undo the merge. Then git checkout -- import/config/import.xml to remove the merge markers from import/config/import.xml22:54
→ shibayu36 joined22:54
canton7 green-, ok, obviously the point hasn't quite gone across. When you do a merge which conflicts, git stops in the middle of the merge and tells you to sort it out. Normally what you'd do is to edit the conflicted file(s), add them, then run "git commit" to create the merge22:55
← macabre left22:55
green- am i misunderstanding what a merge is in git ... if files on the master and those in my working copy differ .. it is going to try to merge those files, correct?22:56
canton7 green-, what happened in your case was that git tried to merge, conflicted, stopped.... Then as well as adding the conflicted file, you added a load of other files *as well* (which I believe were modified between you running the 'git pull' which caused the conflict, and you running 'git commit -a'). You then ran 'git commit -a', whcih concluded the merge and created the merge conflict22:56
In this way, you created a merge commit that does more than just fix a conflicted file -- it also adds a load of new changes. This is termed an "evil merge", for fairly good reasons22:56
green-, yes, it merges files but also the histories of those files (unlike subversion, until recently)22:57
← dangerousdave left22:57
→ linoj joined22:57
green- crazy. hard to wrap your head around. but rolling back seems to be the thing to do. this time i'll reset those commits and commit my changed files before trying to pull22:57
→ johnkpaul joined22:57
green- but will definitely backup first22:57
unfortunately, my attempt will have to wait about an hour22:57
← airborn left22:58
green- canton7: thanks for the guidance, sorry for the insanity. hopefully i can clean this up relatively easily.22:58
canton7 green-, yeah like I said, reset --mixed back to that last good commit (bb913af), then get tid of the merge markers in import/config/import.xml with, be restoring it to how it was at bb913af22:58
← skitrees left22:58
green- will give that a try shortly22:58
thanks again, bbiab22:58
→ ipalaus_ joined22:59
canton7 green-, no worries. I've learnt a new thing that users can do to get evil merges :P I'd advise using __git_ps1 or something similar in your terminal prompt to remind you whether you're in the middle of a merge, what branch you're on, etc22:59
← gchristensen left22:59
green- canton7: one question as i'm on my way out: how were you able to tell that bb913af was the last good commit (which it is)22:59
← aspotashev left23:00
canton7 green-, from the info you posted, that's the commit which was made (on line 5-7) when you committed import/config/import.xml just before that fatal 'git pull'23:00
green- nm, i think i see23:00
→ linduxed joined23:00
canton7 green-, it's also the first parent of the merge commit, as you cans ee from the latter half of line 2023:00
← johnkpaul left23:01
green- forgot i pasted the command and responses, but did realize you could have gotten it from the parent of the merge commit23:01
← dzonder left23:01
canton7 :)23:01
← A1kanaro left23:01
← nutron|w left23:01
← ipalaus_ left23:02
→ ipalaus_ joined23:02
ipalaus_ → ipalaus23:02
← d0k left23:02
→ nikuyoshi joined23:02
← cmyk_ left23:02
→ skitrees joined23:03
← reliability left23:03
← stkrzysiak left23:03
← fedesilva left23:04
← aldodelgado left23:05
→ psoo joined23:05
← psoo left23:05
→ kerrick joined23:05
← pdtpatr1ck left23:05
→ pdtpatr1ck joined23:06
→ sasa90 joined23:06
← sasa90 left23:06
→ griotspeak joined23:07
→ gchristensen joined23:07
→ axl__ joined23:07
→ johnkpaul joined23:08
← kermit left23:10
← axl_ left23:11
→ CannedCorn joined23:14
← Diniden left23:14
→ pdtpatr1ck_ joined23:15
← johnkpaul left23:15
← sawjig left23:15
← pdtpatr1ck_ left23:16
← prudnikov left23:17
← pdtpatr1ck left23:17
→ dougsland joined23:17
← dougsland left23:17
→ dougsland joined23:17
← dv310p3r left23:17
← linduxed left23:18
→ sawjig joined23:20
← FernandoBasso left23:20
← rbrooks left23:22
← nikuyoshi left23:23
← fayimora left23:24
→ shibayu3_ joined23:26
← shibayu36 left23:26
← tewecske left23:28
→ Jay_Levitt joined23:30
→ kermit joined23:31
← maletor left23:31
← linoj left23:31
← iocor left23:32
→ maletor joined23:32
← scelis left23:32
← kaimi left23:32
→ kaimi joined23:33
← TeckniX left23:33
← macmartine left23:34
← martinjlowm-lptp left23:34
→ ska-fan joined23:34
← ph^ left23:36
← pfFreddd left23:37
→ fayimora joined23:37
← Destos left23:37
→ subbyyy joined23:37
→ bigkm joined23:38
→ linduxed joined23:38
→ stuntmachine joined23:39
→ rchavik joined23:40
← panthrodeb left23:41
← vandal- left23:42
← maletor left23:42
← sekati left23:46
← pandeiro left23:46
→ zweb joined23:47
→ maletor joined23:50
← swombat left23:51
→ iocor joined23:51
→ _obiJuan joined23:53
← ipalaus left23:55
← HedgeMage left23:56
→ swombat joined23:57
← iocor left23:58
← glennpratt left23:58

Logs Search ←Prev date Next date→ Channels Documentation