IRCloggy #git 2010-02-06

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.

2010-02-06

→ scarabx joined00:00
← Paraselene_ left00:03
← schlortoo left00:05
← ludde left00:07
← soenke left00:08
→ jrmuizel joined00:09
cbreak are you the jSite bombe?00:09
Bombe cbreak, yes.00:09
cbreak thanks for that tool! :)00:09
Bombe No problem. :)00:09
→ Paraselene_ joined00:09
← Paraselene_ left00:10
→ kpreid_ joined00:10
← khmarbaise left00:12
→ pheaver joined00:16
← Yuuhi left00:17
→ Paraselene_ joined00:19
← Paraselene_ left00:19
→ arooni joined00:20
arooni how do i do a diff cmopaerd to whats in remote master branch00:20
← tewk left00:20
→ schlortoo joined00:20
cbreak diff remote/master..00:21
if your remote is called remote00:21
← pheaver left00:22
← jepster left00:24
← jrmuizel left00:24
← schlortoo left00:26
→ schlortoo joined00:26
← corni left00:28
→ jrmuizel joined00:31
← amerine left00:31
← docelic_ left00:32
← mindworx_ left00:33
← jrmuizel left00:33
← schlortoo left00:34
← ksn left00:34
← catalis left00:35
→ qshan joined00:35
← matthewmccull left00:36
← nadavoid left00:36
→ catalis joined00:36
← til_ left00:36
→ Paraselene_ joined00:37
→ Kelly_zhou joined00:37
→ schlortoo joined00:37
← ndim left00:38
→ mindworx joined00:40
← catalis left00:40
→ JoeConyers joined00:41
→ tewk joined00:41
→ pheaver joined00:42
→ flaguy48 joined00:42
← rpg left00:43
→ docelic_ joined00:46
← schlortoo left00:47
→ schlortoo joined00:49
← seanmccann_ left00:50
→ christophsturm joined00:50
← alley_cat left00:51
← mssm left00:52
← pheaver left00:52
← schlortoo left00:55
→ pheaver joined00:59
→ chewbranca joined01:03
→ agliodbs joined01:05
→ mklappst1hl joined01:05
agliodbs I accidentally overwrote a local file on my local branch with other data. How do I use git to revert the file? There is a remote branch to synch with, if it's relevant01:06
← mklappstuhl left01:08
← chewbranca left01:09
→ skyl joined01:10
→ schlortoo joined01:10
crab agliodbs: what does git status say? changed but not updated? then just git checkout -- filename (which git status should tell you)01:10
skyl git checkout -b pure upstream/master # fatal: git checkout: updating paths is incompatible with switching branches; Did you intend to checkout 'upstream/master' which can not be resolved as commit?01:11
agliodbs crab: thanks, that worked. git revert was looking complicated01:11
crab agliodbs: git revert commits a change that undoes another committed change. it's not like svn revert which throws away changes to the working copy.01:12
skyl I simply want to create a new local branch that will track a remote.01:12
crab skyl: refs/remotes/upstream/master?01:13
locklace skyl: git branch --track localname upstream/remotename01:13
← priidu left01:13
← godsflaw left01:14
→ richcollins joined01:17
→ Lake_ joined01:18
← pheaver left01:18
← _Vi left01:18
skyl it's strange that I have ``remote upstream already exists.`` but ``git branch -a`` doesn't show remotes/upstream01:18
→ tedoc2000 joined01:19
skyl I see it in .git/config and can fetch it ...01:20
locklace skyl: git branch -r should show entries like upstream/<branch>01:20
skyl nevermind me; just bungling momentarily, I've got it now, thanks :-)01:20
→ _Vi joined01:20
← tuxbrother left01:21
← richcollins left01:22
→ seanmccann joined01:22
→ jcromartie joined01:25
jcromartie is there anyway to un-push a commit?01:25
to delete that commit from the remote repo?01:25
kmap jcromartie: I don't really think so.01:26
← tedoc2000 left01:26
jcromartie I put some information in there that I would rather not.01:27
kmap The dirty way would be to remove the public repository, and re-push cleanly…01:27
jcromartie yeah01:27
kmap Or, manually edit the remote repository in some way (I am unaware of the details, though)01:27
← dreiss left01:28
noyb maybe that would be a cool feature, but I think commit-undo would need some form of stage area with a timeout before it's too late. Then the repo will take things from the stage and commit them.01:28
← [1]intripoon left01:28
jcromartie PROTIP: To compromise security, put your application's private key and signing certificate in version control.01:28
← shruggar left01:28
→ richcollins joined01:29
← Alagar left01:32
cbreak :/01:33
locklace jcromartie: push --mirror might work01:34
cbreak you can reset it away locally01:36
and then force push01:36
but note that the commit will still exist on the other repo01:36
until git gc will eventually collect it01:36
in the far far future01:36
flaguy48 jcromartie: ooops01:37
doki_pen BBombe: thanks01:37
jcromartie cbreak: but the commit won't be visible to other users of the remote?01:37
locklace cbreak: seems like --mirror might be better in that regard01:38
← magcius left01:38
← sgh left01:38
→ magcius joined01:38
← lectrick left01:38
← Guest79313 left01:38
doki_pen Bombe: thanks01:39
← tvw left01:39
jcromartie cbreak: that worked like a charm01:41
now I'm equipped to wreak havoc on our histor01:41
history01:41
cbreak oh, by the way01:42
if anyone pulled before you did that01:42
jcromartie they didn't01:42
cbreak he'll push the commit back next time he pushes01:42
and if anyone accesses the repository01:42
he can restore the commit from the reflog01:42
jcromartie well it's a hosted repo, and it's not a critical security thing01:45
I think we're covered01:45
← Modius left01:45
jcromartie it's just an Android app :P01:45
← Paraselene_ left01:49
→ paltman joined01:49
jcromartie thanks for the help01:50
← jcromartie left01:50
→ dreiss joined01:54
→ Fissure1 joined01:54
← richcollins left01:54
← seanmccann left01:55
→ Guest79313 joined01:55
← Fissure left01:56
→ ispeak_ joined01:57
→ paltman_ joined01:58
← mason-l left02:00
→ TheNewAndy joined02:00
→ Gitzilla joined02:01
← paltman left02:01
paltman_ → paltman02:01
← ispeak_ left02:05
→ methanol joined02:05
← TheNewAndy left02:05
→ State joined02:06
State Hello - If I do apt-get install git-core, do I need to refresh something to make the git command available in the shell?02:06
I'm sure I'm missing something.02:06
→ s1kx- joined02:06
AAA_awright Not for me... you can try opening a new shell anyways, that has the same effect02:07
State Hm. That's what I figured.02:07
← methanol left02:08
State If I'm sshed in, how do I re-login without actually disconnecting?02:08
adam12 State: Maybe try 'bash'02:08
State That's it. Thanks.02:08
adam12 State: What distro?02:08
State debian lenny02:09
adam12 ls /usr/bin/git02:09
State that works, but it opens another bash of course02:09
← S1kx left02:09
State Hm. Just did it again: seemed to work. Thanks though.02:10
fryguy it depends on what shell you are using02:13
csh-style shells require you to run rehash or something similar02:13
all bourne and korn derivatives should automatically refresh PATH02:13
← kipras left02:13
→ mason-l joined02:16
← paltman left02:17
← JackWinter left02:18
→ JackWinter joined02:21
← dreiss left02:22
State OK: I have an ssh key pair that I use to log in to the server personally, should I generate a different pair to push git updates?02:22
I would guess yes, especially if others are going to contribute02:22
fryguy State: why?02:23
cbreak create a pair for each person who contributes02:24
or better02:24
let the person who contributes generate one02:24
cbreak is speaking out of theory, not practical application02:24
→ vbgunz joined02:24
State ok, that makes sense - you can add keys to gitosis any time i guess02:25
→ VVSiz_ joined02:26
← Guest79313 left02:26
← _Vi left02:29
← VVSiz left02:29
→ Gentoochild joined02:30
← parasti left02:33
→ Zeelot2k joined02:35
→ dreiss joined02:36
→ ATH500 joined02:37
→ paltman joined02:39
→ MJBoa joined02:39
MJBoa hey guys, i have two version of a project, the original untouched version and one i worked on before using git. I want to add version control so I want the initial import to be the untouched version and I want to commit a second version with the changes i made already02:43
cbreak git init the first directory, add the things you want, commit.02:44
move the second directory's content into the first directory, add everything again, commit02:44
you have two commits02:44
make backups if you want to be careful and don't have git experience.02:44
MJBoa ok so overwrite everything, leaving the .git folder02:44
cbreak yes.02:45
MJBoa that makes sense thanks02:45
cbreak I guess you can also just copy/move the .git folder over to the other dir02:45
Gentoochild ah, good to see people are awake :-) I always forget it's a global world.02:45
b/c I've got a question, too02:45
cbreak lives in europe02:46
cbreak but the world is global indeed02:46
Gentoochild If I delete a commit using --amend, will the objects the commit refers to be deleted as well?02:47
cbreak --amend doesn't delete commits02:48
← ivan left02:48
cbreak it changes your last commit02:48
Gentoochild perhaps I'm mixing something up, I only two days into got *g*02:48
→ bremner joined02:48
Gentoochild I only=I'm only02:49
basically I want to change a file in that commit02:49
→ ajpiano joined02:49
← matgnt left02:49
→ aresnick joined02:50
cbreak if you amend a commit, the old version of that commit is still available in the reflog, and it will not be collected/deleted for quite some time02:51
MJBoa im having a little trouble comprehending the decentralized nature of git, every single person editing a project has their own repository right? what do you do with multiple users instead of normally pushing changes to the server? how do multiple users know that another user has made changes to the branch and pull down those changes?02:56
bob2 in practice you often have a central repository02:56
that everyone pushes and pulls from02:56
cbreak or everyone has a public repository where only he pushes to02:57
but everyone else can pull from02:57
to get that person's changes02:57
→ ivan joined02:57
cbreak I guess in practice, you have a mix of both02:58
MJBoa but how do you even push to a repository?02:58
Gentoochild well, people at Amarok do it that way: they have core devs with push rights to the central repo.02:59
cbreak with git push02:59
Gentoochild everyone else uses a clone and then requests merges into the main repo02:59
→ ideogon joined02:59
← rgr left02:59
← pkw left02:59
← agliodbs left03:00
→ sagsousuke joined03:01
→ Chani joined03:01
Chani soso... I fucked up03:02
cbreak backups?03:02
Chani I meant to type "git rebase master branch" and I typed "git rebase branch master"03:02
how do I undo that?03:02
cbreak git reset --hard HEAD@{1}03:03
probably03:03
Chani probably?03:03
cbreak no03:03
fryguy gonna need more than that03:03
cbreak rebase uses more than one03:03
fryguy gonna need to look at the log03:03
cbreak git reflog03:03
fryguy and find the commit you want to reset to03:03
cbreak then look which one is from before the rebase03:03
bob2 stop, cp -a the whole thing right now03:03
before fiddling03:03
cbreak then reset to that03:03
fryguy don't even need reflog, probably easier to just look at the normal log03:03
cbreak but as I said at the start, make backups :)03:03
rebase screws up logs fryguy03:04
it rewrites history03:04
Chani the good news is, it's a gitsvn repo and most of hte branch's code is also on reviewboard03:04
fryguy cbreak: not in this case03:04
cbreak: interactive rebase rewrites history, normal rebase does not03:04
it just appends03:04
cbreak all rebase rewrites history. that's it's whole purpose.03:04
fryguy cbreak: no, it's not03:05
Chani looks in the log03:05
cbreak your own commits get put on top of someone elses03:05
so you insert history before your own commits03:05
fryguy cbreak: it's purpose, when not in interactive mode, is to move commits from one basepoint to another03:05
Chani okay, so the branch's stuff, which spans december-january, has been inserted in between february 2nd and 3rd in the svn commits.03:05
cbreak -> that's changing history03:05
Chani so maybe I can reset to february second then ask git-svn to update stuff again.03:06
cbreak Chani: take a look at git reflog.03:06
it is usually a good place to find what you want03:07
Chani hmmmm.03:07
cbreak maybe you find git log -g more comfortable to read03:08
"Reflog message:" might be helpful too03:08
← batok left03:08
← hipe left03:09
← erikvold left03:09
→ hipe joined03:09
← Adaptee left03:09
Gentoochild oh well, 4AM. Gotto go to bed. g'night03:10
→ Davey joined03:10
← Gentoochild left03:10
← mklappst1hl left03:10
Chani reflog's a bit confusing03:11
← Lake left03:11
← Lake_ left03:11
Chani svn rebase seems to create two entries in the reflog but I'm not sure why03:12
← Fullmoon left03:12
cbreak rebase removes your commits03:12
and then puts them on top of the place where you want them to be put on03:12
I'd have expected more than two to be honest03:13
but ah well03:13
I've never done it myself, and two sounds plausible too :)03:13
← sh1mmer left03:13
Chani cbreak: I did git reset --hard HEAD@{14} but I think I should have used 15. do I now use HEAD@{1} ?03:14
cbreak if you change your head by any means03:15
Chani or.. 16?03:15
cbreak the new head gets added to the list as 003:15
so 1603:15
you can look again :)03:15
the reflog is a history03:15
a log of the states of your HEAD03:16
it is totally unrelated to the commit history03:16
Chani yay! my history looks sane :)03:16
→ Lake joined03:16
cbreak the reflog supports things like HEAD@{1.hour.ago}03:16
Chani and now the branch is rebased :)03:16
cbreak ... I read03:16
never tried03:16
Chani wow, neat03:16
→ Lake_ joined03:17
Chani loves how everything can be fixed in git03:17
cbreak make backups first anyway :)03:17
→ rgr joined03:18
← jafl left03:18
← ArdRigh left03:19
→ j416 joined03:19
→ sh1mmer joined03:19
→ cytrinox_ joined03:20
← cytrinox left03:21
cytrinox_ → cytrinox03:21
← RomD` left03:24
← rgr left03:24
cj is there a way to get all of the diffs since a certain revision?03:24
cj guesses git diff has an arg that does this03:25
bob2 as in what? a list of diffs between successive revisions, or the cumulative one?03:25
fryguy git log -p for the first case, and git diff for the second case03:26
bob2 git log -p, git diff <c1> <c2>03:26
cbreak there's a lot of reading material in man git-diff03:34
Gitbot cbreak: the git-diff manpage can be found at http://git.or.cz/man/git-diff03:34
cbreak you have quite some freedom specifying what you want03:34
→ cori[s] joined03:35
← qshan left03:40
MJBoa whats the command to retrieve a copy of a branch without .git info? like svn export03:41
cbreak archive03:41
gives an archive03:41
like zip03:41
MJBoa: man git-archive has some pointers03:42
Gitbot MJBoa: the git-archive manpage can be found at http://git.or.cz/man/git-archive03:42
cbreak use it with > outputfile.zip for example03:42
→ Adaptee joined03:42
→ sumangurung joined03:42
← vbgunz left03:42
→ rgr joined03:43
← sumangurung left03:45
→ Textmode joined03:45
Textmode morning all03:46
← engla left03:46
← jschoolcraft left03:47
→ vbgunz joined03:47
adam12 morning03:48
AAA_awright In where, Europe? Good night from the west coast :)03:48
adam12 Your right, it's only 11pm here03:49
Oh well :(03:49
:)03:49
→ ArdRigh joined03:50
→ pheaver joined03:51
→ Rolenun joined03:53
← paltman left03:54
← MJBoa left03:55
→ MJBoa joined03:56
← schlortoo left03:57
→ remi joined03:57
← NetHawk left04:04
← remi left04:04
→ Pistos joined04:12
Pistos I am trying to set up a shared bare repo. I'm concerned, however, that when one user does operations on it, it will change perms (or create files with perms) such that other users won't be able to perform git operations on it any more.04:13
If my fears are well-founded; how can I set up UNIX permissions in the repo so that this won't happen?04:13
i.e. toying with umask, sticky bits, or whatnot.04:13
→ mdel joined04:16
← dreiss left04:25
sitaram Pistos: git init --bare --shared sets up perms the right way, then you have to add every potential user to the unix group that you did that 'git init' with04:26
Pistos: in other words, everyone has to be members of the group that was used to create the repo, and then yes it is the sticky bits; --shared will do that right04:27
Pistos sitaram: Perfect; thanks. Let me try with --shared.04:29
Somehow, I have vague memories of being told this in the past. :)04:29
→ rgr_ joined04:30
← rgr left04:30
→ NetHawk joined04:30
← jasonn1234 left04:32
→ ebneter joined04:32
← alfredlysebraate left04:32
→ Determinist joined04:33
← rgr_ left04:34
← aziz left04:40
← gnufied left04:41
→ rgr joined04:42
→ pinnacle joined04:42
← Lake_ left04:44
← Lake left04:45
→ schlortoo joined04:46
← tetha left04:47
→ tetha joined04:48
← christophsturm left04:52
← mentor left04:55
← justin-george left04:57
← aresnick left05:01
→ offby1` joined05:01
offby1` → offby105:01
← offby1 left05:01
→ offby1 joined05:01
→ pablo_ joined05:02
← ProperNoun left05:02
pablo_ how do u tell git wich ssh key are u going to use for authenticating with a remote repo???05:02
j416 pablo_: tell ssh, not git05:03
sitaram pablo_: you can do it using a "host" paragraph in ~/.ssh/config05:03
pablo_ sitaram: u have to put a reference to your private key there?05:03
← Pistos left05:03
sitaram pablo_: one second; pastebin something for you...05:04
pablo_ ok, thx05:04
→ paltman joined05:04
← Chillance left05:04
← Davey left05:05
sitaram pablo_: http://pastebin.com/m200d49a905:05
← vbgunz left05:05
→ godsflaw joined05:06
sitaram pablo_: then you use "git clone git95:reponame"05:06
MJBoa hey sitaram05:06
pablo_ sitaram: ok, gonna try that, thx a lot05:06
MJBoa if i add a host paragraph can i just say "ssh git95"?05:06
sitaram MJBoa: sure05:07
MJBoa cool didnt know that05:07
sitaram MJBoa: although it was written for gitolite, there's a lot of ssh "fundas" at http://sitaramc.github.com/0-installing/9-gitolite-basics.html and http://github.com/sitaramc/gitolite/blob/pu/doc/6-ssh-troubleshooting.mkd05:08
especially the first link has a "basics of ssh" or some such section; but it's more than basics of course05:08
→ ProperNoun joined05:09
← univate left05:11
← offby1 left05:11
sitaram and I like to believe I've written it for easy reading :)05:12
← Pewpewarrows left05:13
← pheaver left05:14
→ offby1 joined05:18
← Aaaarg left05:18
→ Aaaarg joined05:19
← DrNick left05:23
→ __iron joined05:24
→ pkw joined05:25
← _iron left05:28
→ sumangurung_ joined05:32
← sumangurung_ left05:37
hipe → hipe|din-din05:40
← mefesto left05:40
← MJBoa left05:41
← dominikh left05:42
→ gnufied joined05:43
← kingfishr left05:47
← camonz left05:53
← project2501a left05:54
→ Pewpewarrows joined05:57
← DavidKlein left05:58
← jfkw left06:05
→ dduncan joined06:05
← godsflaw left06:05
→ godsflaw joined06:06
← kukks left06:07
← crankharder left06:18
→ jrmuizel joined06:18
→ TheNewAndy joined06:18
→ bob2 joined06:19
Chani /wc06:19
doh.06:19
← Chani left06:19
→ hipe joined06:20
→ gitster joined06:20
← godsflaw left06:21
← hipe|din-din left06:21
→ seamushc joined06:22
seamushc Hey guys, getting a weird error from a git fsck: warning in tree 1ed767b9e5b028329e8a9b132471b7174c69d7d6: contains zero-padded file modes, I dont think it is causing any problems, but I am curious about what is causing it, and ideas?06:23
gitster what project is this?06:23
seamushc Its just a repo I am using for school, containing mostly .tex files06:24
a few binaries though06:24
gitster the reason I ask is that ancient git used by early days of kernel used to create such.06:24
these days, in modern implementations and re-implementations, it is considered a bug.06:24
whose git implementation are you using? C-git? EGit? Something else?06:25
seamushc the repo was made with git 1.4, because that is what came with cygwin (I recompiled my own after I found that out)06:25
→ DrNick joined06:25
gitster 1.4 is new enough not to suffer from "zero-padded modes".06:25
← TheNewAndy left06:25
seamushc strange...06:26
→ ankit9 joined06:26
seamushc Any ideas for a fix?06:26
gitster indeed it is.06:26
It is not a fix, but you might want to try running "git archive 1ed767b9..." to see what's in there.06:27
it will create a tarball (or perhaps zip).06:27
→ segher joined06:27
→ TheNewAndy joined06:28
← hadeees left06:30
→ justin-george joined06:31
← TheNewAndy left06:33
seamushc I will try that, thanks06:36
← MetaCosm left06:37
← rtyler left06:37
→ TheNewAndy joined06:38
seamushc it contained a latex file06:38
and nothing else06:39
← kinabalu left06:39
→ bdimcheff joined06:41
← ideogon left06:42
→ Pistahh joined06:42
→ hyperair joined06:43
← ATH500 left06:44
→ Lake_ joined06:44
→ Lake joined06:44
← scarabx left06:48
← Lake_ left06:48
← Lake left06:48
→ MetaCosm joined06:50
← Rolenun left06:51
← TheNewAndy left06:51
→ rtyler joined06:52
aspotashev_ → aspotashev06:54
← sh1mmer left06:55
← j416 left06:55
→ TheNewAndy joined06:57
← __iron left06:57
→ sh1mmer joined06:58
Ilari seamushc: Also earlier versions of github editor (whatever that was) did those.06:58
seamushc: Considering nature of the issue, I would leave it alone. It isn't serious.07:00
← dduncan left07:01
gitster As long as "archive" produces a good copy, you should be able to extract and use the contents fine. It's a different story if there are _other_ bugs in whatever created such an object, though ;-)07:02
→ mithro joined07:04
→ bdon-csua joined07:05
← bdon-csua left07:05
← sh1mmer left07:05
→ project2501a joined07:08
← aspotashev left07:09
→ reto` joined07:09
← adamm left07:10
← TheNewAndy left07:10
→ sde_antrix1 joined07:10
← gnufied left07:12
← pinnacle left07:13
← mithro left07:14
← pkw left07:16
→ TheNewAndy joined07:19
→ dreiss joined07:19
← paltman left07:19
→ gnufied joined07:20
← gnufied left07:20
→ brizly joined07:25
→ mithro joined07:26
← bdimcheff left07:26
← sagsousuke left07:26
→ stringo0 joined07:32
→ AkumaStreak joined07:32
← jrmuizel left07:35
← TheNewAndy left07:36
→ TheNewAndy joined07:37
→ bambule joined07:39
← TheNewAndy left07:42
← project2501a left07:43
→ TheNewAndy joined07:44
← RurouniJones left07:45
→ godsflaw joined07:47
→ pkw joined07:47
→ RurouniJones joined07:50
→ project2501a joined07:50
← TheNewAndy left07:52
→ pencilo joined07:57
seamushc Thanks guys07:57
quit07:57
← seamushc left07:57
← project2501a left07:58
← godsflaw left08:00
← pkw left08:02
← Bass10 left08:03
→ project2501a joined08:03
→ nitrate joined08:06
← project2501a left08:06
← ArdRigh left08:08
→ _Vi joined08:09
← stringo0 left08:10
← r11t left08:10
→ Zephyrus joined08:10
→ karpar joined08:10
→ solrize joined08:11
← AkumaStreak left08:12
→ TheNewAndy joined08:15
← Mage left08:19
← hipe left08:20
← TheNewAndy left08:20
← _Vi left08:22
→ anno^da joined08:22
← mindworx left08:23
← anno^da left08:24
→ hipe joined08:27
→ mindworx joined08:27
← mithro left08:29
← bcardarella left08:30
← rduplain left08:31
→ ixti joined08:32
→ RdeWilde_ joined08:33
← dreiss left08:33
← karpar left08:36
RdeWilde_ I can't checkout 'localAB' http://github.com/RdeWilde/rows-server/tree/localAB `git branch -l` does only show 'master' .. how can I fix this?08:45
thiago_home first of all, what did you clone?08:46
→ TheNewAndy joined08:46
← solrize left08:46
→ JayM joined08:48
← Aaaarg left08:48
→ Aaaarg joined08:48
→ vvsh joined08:48
RdeWilde_ thiago_home; `git clone [email@hidden.address]08:49
← TheNewAndy left08:51
RdeWilde_ thiago_home; I can checkout the remote branch, but it gives me kinda strange output; http://pastie.org/81218708:52
← spacebat left08:54
→ Aaaarg_ joined08:54
RdeWilde_ But it works for the files I think..08:54
← Aaaarg left08:55
Aaaarg_ → Aaaarg08:55
→ spacebat joined08:56
← krawek left08:56
drizzd_ RdeWilde_: why is the output strange? it says exactly what you're saying, that origin/localAB is a remote branch08:57
RdeWilde_ The last line, the name of the current branch.. it shows 7c83248234 or something08:58
That's normal?08:58
drizzd_ yes, since you don't have a local branch, you are basically sitting on the current commit that's currently the latest in origin/localAB08:59
if you make a commit, it will not be recorded in any branch08:59
you can't commit to a remote branch08:59
that's why it's verbosely suggesting that you create a local branch09:00
OTOH, if you're not planning to do commits there, it's perfectly fine09:00
→ ArdRigh joined09:01
→ pinnacle joined09:02
→ ATH500 joined09:02
→ ph^_ joined09:03
→ gebi joined09:04
← ph^ left09:04
← ph^_ left09:04
→ ph^ joined09:05
← MetaCosm left09:05
← ph^ left09:08
→ ph^ joined09:08
← ph^ left09:09
RdeWilde_ Ah, I understand. Thanks for your help. I understand now09:10
→ TheNewAndy joined09:11
→ priidu joined09:15
← dbatbold left09:15
→ dbatbold joined09:15
← TheNewAndy left09:16
→ nizuri joined09:17
← Gullik left09:19
→ meyering joined09:19
→ Gullik joined09:20
→ ph^ joined09:20
→ Sp4rKy joined09:25
Sp4rKy hi, sorry for this newbie question, but I would clone a branch of a git repository (on gitorious) on my computer09:26
what's the good syntax ?09:26
dennda git clone URL09:26
thiago_home clone the repository09:26
you get all branches09:26
Sp4rKy oh09:27
I have to clone the whole repo ?:'(09:27
there are a lot of branches, and so the whole repo will take so long :'(09:27
dennda that's not bad09:27
no09:27
git branches are cheap09:27
thiago_home usually, cloning all branches won't increase by more than 10% the transfer size09:27
dennda no data is duplicated09:27
Sp4rKy oh ok09:28
thanks for the tip :)09:28
← justin-george left09:32
→ TheNewAndy joined09:33
Sp4rKy gitorious is so slow :'(09:34
thiago_home tries09:38
thiago_home downloading at 1.9 MB/s09:38
Sp4rKy 0_o09:38
22KiB/s here09:39
for both tarball and git clone09:39
and the tarball was corrupted09:39
thiago_home it reached 2.01 MB/s09:40
average 1.9109:40
Sp4rKy :|09:40
I try for 2 days09:41
thiago_home Receiving objects: 100% (276637/276637), 199.63 MiB | 1.91 MiB/s, done.09:41
Sp4rKy always same09:41
:'(09:41
← Pewpewarrows left09:41
Sp4rKy tries from another location09:42
Sp4rKy mhhh, much faster from this other location09:44
strange since I have no network issue here09:44
→ ramshacklecpu joined09:46
← TheNewAndy left09:46
← vvsh left09:46
→ jessep joined09:46
→ Paraselene_ joined09:48
← segher left09:49
← doener left09:53
← pablo_ left09:54
→ fifthsally joined09:54
→ doener joined09:55
← ATH500 left09:55
→ project2501a joined09:57
→ pantsman joined09:58
← pantsman left09:58
→ pantsman joined09:58
← jessep left09:59
← project2501a left09:59
→ thisch joined10:00
→ project2501a joined10:00
← VVSiz_ left10:00
→ jessep joined10:00
→ VVSiz joined10:00
→ maleknet joined10:01
← jessep left10:02
→ catalis joined10:02
→ jessep joined10:03
← nizuri left10:04
← jessep left10:05
← priidu left10:09
→ robert_ joined10:10
→ sgh joined10:11
← fifthsally left10:13
→ UncleCJ joined10:15
→ alfredlysebraate joined10:16
← Kelly_zhou left10:16
→ TheNewAndy joined10:17
dsop aehm, shouldn't git --work-tree=c/git.git status just run fine ?10:19
← TheNewAndy left10:21
← niki left10:21
thiago_home sounds like you wanted --git-dir10:22
not --work-tree10:22
dsop no10:22
sorry git.git is not bare10:22
so let's say --work-tree=c/gitdir10:22
doesn't work for me on mac os x with 1.7-rc110:23
thiago_home you must set both10:24
← RobertLaptop left10:24
→ jinzhu joined10:26
→ malumalu joined10:27
← thisch left10:28
dsop thiago_home: thanks10:28
→ karasiov joined10:30
→ nizuri joined10:30
dsop thiago_home: git --work=tree=gitrepo --git-dir=gitrepo/.git diff-index HEAD doesn't work10:31
→ khmarbaise joined10:33
thiago_home it should10:34
though it did take me a while to get my settings right too10:35
dsop thiago_home: it doesn't, git status works fine, ls-tree works fine but diff-index doesn't10:35
→ RobertLaptop joined10:37
thiago_home diff-index is a plumbing command10:38
→ jasonn1234 joined10:38
dsop still I expect plumbing to respect --working-dir and --git-dir10:39
thiago_home: I need a plumbing command that shows me if the working dir is directy and I'm not in the working directory itself10:40
s,directy,dirty10:40
← gebi left10:45
→ d0ugal joined10:45
wereHamster → wereHamster|away10:46
thiago_home I don't know10:46
→ brizly1 joined10:48
← brizly left10:50
← ramshacklecpu left10:50
→ ph^_ joined10:52
→ jessep joined10:54
← sgh left10:56
← jessep left10:58
→ tvw joined11:00
RdeWilde_ Sorry, I'm lost. How can I get a remote branch local to edit it and push back? git branch myLocal remoteOne ?11:02
→ Octalot joined11:03
→ christophsturm joined11:03
→ brizly joined11:03
← brizly1 left11:04
→ spaceonline joined11:05
cbreak RdeWilde_: git checkout -t remotename/branchname11:06
you can also do git checkout -b newname remotename/branchname if you don't want tracking info11:06
thiago_home RdeWilde_: or what you said yourself11:07
then check it out11:07
cbreak or git branch -t remotename/branchname newname11:07
RdeWilde_ cbreak; What does `git branch localOne origin/remoteOne` do? :/11:07
Ah11:07
cbreak it creates a new branch origin/remoteOne that branches off localOne11:07
→ TheNewAndy joined11:07
← Sp4rKy left11:07
cbreak hmm.11:08
or vice versa?11:08
thiago_home vice-versa11:08
cbreak anyway, I think the -t is quite useful :)11:08
thiago_home it creates a local branch called localOne that tracks origin/remoteOne11:08
→ parasti joined11:09
→ corni joined11:10
← UncleCJ left11:11
cbreak I personally like git checkout -t most11:11
← nizuri left11:11
→ vvsh joined11:12
← TheNewAndy left11:12
← ixti left11:13
thiago_home I never remember the syntax11:13
← hipe left11:13
→ tjc joined11:16
← khmarbaise left11:17
← docelic_ left11:18
→ khmarbaise joined11:18
→ sgh joined11:19
→ UncleCJ joined11:19
→ gebi joined11:20
← rgr left11:21
→ rgr joined11:21
→ engla joined11:23
← engla left11:23
→ engla joined11:23
← UncleCJ left11:24
→ rbq joined11:27
→ mbroeker joined11:28
dsop ^parren11:30
arg11:30
← riz2 left11:30
→ pheaver joined11:30
→ docelic_ joined11:31
→ fmardini joined11:32
← gebi left11:33
← tjc left11:37
→ karpar joined11:38
→ TheNewAndy joined11:39
→ DavidKlein joined11:40
← cjs left11:44
→ orodent joined11:44
orodent ?DCC SEND "diaper_jesus" 0 0 011:44
← reto` left11:44
← noyb left11:44
← hyperbolist left11:44
← ankit9 left11:44
← enthooz left11:44
← cbreak left11:44
← orodent left11:44
→ cbreak joined11:45
→ enthooz joined11:45
→ hyperbolist joined11:45
→ ankit9 joined11:45
→ cjs joined11:45
→ odwaak joined11:47
← TheNewAndy left11:50
← cjs left11:57
→ DENZEL_FOXX joined11:57
DENZEL_FOXX ?DCC SEND "ff???f?" 0 0 011:57
← cbreak left11:57
← enthooz left11:57
← ankit9 left11:57
← hyperbolist left11:57
→ cjs joined11:57
→ hyperbolist joined11:57
→ enthooz joined11:58
→ cbreak joined11:58
→ ankit9 joined11:58
Textmode ...failed spambot?11:58
thiago_home yeah11:58
locklace i think you mean "succeeded"11:58
thiago_home no, failed. They couldn't write a proper CTCP.11:59
Textmode locklace: not compared to what I think its trying to do.11:59
locklace thiago_home: on the contrary.11:59
Textmode looks like its trying to do file transfers...11:59
← cbreak left12:00
thiago_home yes, which is why it failed12:00
they couldn't write the proper CTCP command12:00
→ cbreak joined12:00
← DENZEL_FOXX left12:00
locklace http://forums.linksysbycisco.com/linksys/board/message?board.id=Wireless_Routers&thread.id=8875812:04
cbreak: you'll want to get that fixed12:04
cbreak My firwmare coding ability is not that developed12:06
locklace you're already running the latest?12:08
i can't remember whether that firmware bug is specific to port 6667, but if so connecting to one of freenode's other ports should help (8001, 6665, etc.)12:09
cbreak the version from 2008, yes12:09
→ _iron joined12:10
← sgh left12:11
← ArdRigh left12:13
→ sgh joined12:13
cbreak the only newer update would be openwrt or new hardware12:14
locklace try connecting to irc.freenode.net:8001 instead12:16
cbreak sounds like a terrible hack12:17
→ raz joined12:17
locklace up to you.12:17
cbreak next time It'll connect to that port12:18
ah well. Linksys is out. DLink is out since the last router12:18
wonder what's out next.12:18
raz can anyone assist with a very simple merge/cherrypick here. i'm wondering what the cleanest way would be to the last patch from the "andy' branch (green) into my branch (black): http://github.com/razmataz/redis-py/network .. ideally i would like git to recognize that a merge happened but i think a plain merge wouldn't work because all the intermediate commits conflict... so what's the best way, cherrypick?12:19
→ rseifert joined12:20
→ GamaFranco joined12:21
cbreak if you cherry pick, you only get the changes from that commit12:22
← GamaFranco left12:22
cbreak if you merge, you get all12:22
raz yea. guess i have no choice but cherrypick because the branches have diverted so much that the intermediate commits would conflict.12:23
→ GamaFranco joined12:23
raz will git recognize that my cherrypick is bad on his branch or does it just look like another commit by me12:23
cbreak (and flash is evil!)12:23
raz s/bad/based/12:23
cbreak like any other commit by him I think12:24
GamaFranco hi12:24
raz hmm ok then i'll just go head and stop making this harder for myself than it need be ;)12:24
i hoped i could somehow get a fancy line drawn from his commit to mine, to visualize that i was merging his work12:24
but guess that doesn't work here12:24
← ben_alman left12:25
← parasti left12:25
GamaFranco I have one quick question: does git support permissions on a folder base? (i.e. some users have permissions to certain folders while others not)?12:25
→ bentob0x joined12:25
→ ben_alman joined12:26
cbreak no12:26
git doesn't do permissions12:26
GamaFranco ok, thanks12:26
cbreak there's some server side stuff12:26
→ UncleCJ joined12:26
cbreak that can do some kind of permissions12:26
→ blizard joined12:26
cbreak gitosis or something like that12:26
→ sgh_ joined12:27
locklace GamaFranco: you want something like gitolite, i guess12:27
GamaFranco good. where can I get that info?12:27
thanks12:27
i'll google it then12:27
cbreak or gitolite, yes. http://github.com/sitaramc/gitolite12:27
but I think it's branch based12:27
and not directory based12:27
← sgh left12:28
GamaFranco ok, thanks cbreak12:28
I use directory based permissions in my SVN repo and I was thinking in migrating to GIT. thanks for helping me exploring this alternative12:30
blizard new hack tools exploits,vulnerabilities,videos,shellcode on www.team3d.3xforum.ro12:30
new hack tools exploits,vulnerabilities,videos,shellcode on www.team3d.3xforum.ro12:31
new hack tools exploits,vulnerabilities,videos,shellcode on www.team3d.3xforum.ro12:31
new hack tools exploits,vulnerabilities,videos,shellcode on www.team3d.3xforum.ro12:31
new hack tools exploits,vulnerabilities,videos,shellcode on www.team3d.3xforum.ro12:31
new hack tools exploits,vulnerabilities,videos,shellcode on www.team3d.3xforum.ro12:31
new hack tools exploits,vulnerabilities,videos,shellcode on www.team3d.3xforum.ro12:31
new hack tools exploits,vulnerabilities,videos,shellcode on www.team3d.3xforum.ro12:31
new hack tools exploits,vulnerabilities,videos,shellcode on www.team3d.3xforum.ro12:31
new hack tools exploits,vulnerabilities,videos,shellcode on www.team3d.3xforum.ro12:31
← blizard left12:31
relaxok mhm12:31
zomg Apparently they did not teach spelling in hacker-U12:31
→ jelmer joined12:32
→ parasti joined12:34
→ sgh joined12:35
← sgh_ left12:36
Textmode zomg: you mean scriptkiddy-U :P12:36
zomg Yeah =)12:37
← jinzhu left12:38
← GamaFranco left12:40
← sgh left12:41
→ sgh joined12:41
← jelmer left12:41
← sgh left12:46
→ munichlinux joined12:48
← bentob0x left12:50
→ mssm joined12:51
→ anno^da joined12:52
← orafu left12:54
→ orafu joined12:54
← mssm left12:58
→ mssm joined12:58
→ methanol joined12:58
← munichlinux left13:00
→ qfmcnx joined13:00
→ Lake_ joined13:01
→ Lake joined13:01
→ munichlinux joined13:01
← Lake_ left13:02
→ segher joined13:02
← odwaak left13:02
← d0ugal left13:04
← qfmcnx left13:05
← kpreid_ left13:06
→ fwdyvx joined13:07
→ Lake_ joined13:11
→ jxaejb joined13:12
→ omerj joined13:13
elliottcable → nicholaskennethc13:13
← omerj left13:13
nicholaskennethc → elliottcable13:13
← fwdyvx left13:15
→ jschoolcraft joined13:16
→ til_ joined13:18
← jxaejb left13:19
→ pdshdp joined13:20
→ remi joined13:22
→ archis joined13:23
→ Geralt joined13:24
← Geralt left13:24
→ Geralt joined13:24
← karpar left13:27
← til_ left13:33
→ kipras joined13:34
→ Tureba joined13:35
← pdshdp left13:36
→ sgh joined13:36
← archis left13:37
→ wacodx joined13:37
→ til_ joined13:39
← karasiov left13:39
→ consolers joined13:41
→ aspotashev joined13:42
← cesario_ left13:42
→ gaveen joined13:43
← catalis left13:43
← consolers left13:43
→ minefield joined13:43
minefield OT but what is the moral opposite of SHA? what is a hash for detecting "similarity" in strings13:43
← secoif left13:43
minefield where O(n^2) "distance" comparisons are out of the question13:44
← munichlinux left13:48
cbreak hamming distance?13:48
I don't think there's a hash for it.13:48
→ krunk- joined13:48
krunk- morning13:48
cbreak but maybe you can try to compress the source13:48
and compare the compressed strings instead13:49
← rudi_s_ left13:49
→ camonz joined13:49
← pinnacle left13:49
krunk- I host some of my lab's git repositories on our server under a git user and committers are given access by submitting a public key13:49
cbreak (I read that people theorized comparing authorship similarities of text or language of text by their compression properties)13:49
→ jelmer joined13:50
→ rudi_s joined13:50
RdeWilde_ Should I use CRLF or LF? I though LF but when I do `git add . ` it says "warning: LF will be replaced by CRLF in" .. confused..13:51
krunk- if a user works on a clones repository from two different places and set up their email or name as different (e.g. foo bar vs. foo c. bar) it show's up as two different committers. Is it possible to map these committers and their commits?13:51
cbreak I always use LF13:51
RdeWilde_ cbreak; I understood that was right, but then I don't understand why Git says it's replacing the LF to CRLF :/13:52
(I'm on windows)13:52
cbreak autocrlf?13:52
it's optional13:52
RdeWilde_ Yes, that's enabled, but should that convert the CRLF to LF when pushing/changing :S?13:52
→ munichlinux joined13:53
RdeWilde_ Because it seems to do the opposite of what I understand what it is saying? :S13:53
cbreak it makes stuff in your working dir CRLF13:53
and stuff inside your repo LF13:54
I think?13:54
cbreak stays away from windows13:54
RdeWilde_ I'm confused :S I set my editor to use LF :/ And now it's converting it back? :p lol13:54
cbreak: I have no choice.....13:54
pasky minefield: i think there's no hash for it, at least i don't know about any, but there are algorithms to find subsequences of one string in another etc.13:54
cbreak disable autocrlf if you don't want to have it13:54
pasky minefield: it depends a lot on what do you define by "similarity"13:54
RdeWilde_ I'm not sure, on github help I read I should want to have it :/13:55
cbreak I think diff works by hashing each line to find equal lines13:55
← sejo left13:56
→ sejo joined13:57
← doener left13:57
← jelmer left13:58
pasky minefield: see section 4.1 of http://pasky.or.cz/~pasky/bc/bcthesis-final.pdf for some possibly-related algorithms13:58
← sejo left13:58
pasky minefield: ("diff" is really a lot about detection similiarities ;)13:58
→ jfkw joined14:00
→ nonken joined14:01
→ jceb joined14:01
← malumalu left14:02
→ malumalu joined14:02
→ radarek joined14:02
→ sejo joined14:07
sejo → Guest5509714:08
← wacodx left14:09
minefield cbreak looks like theres nothing short of a full textmining solution (identifying extracting features and running those through a clustering algo)14:09
→ xknucp joined14:11
← nonken left14:11
← Guest55097 left14:11
→ nonken joined14:12
→ masterhu1per joined14:12
← masterhu1per left14:12
→ masterhu1per joined14:13
← nonken left14:14
masterhu1per → evilsejo14:16
→ yziquel joined14:17
→ svenfark joined14:17
→ dominikh joined14:18
→ killerchicken_ joined14:18
svenfark Hi, I've got a question about git-bisect14:18
According to http://lwn.net/Articles/317154/ : "This is most efficiently done by performing a binary search between the last known working commit and the first known broken commit in the commit history."14:19
What about scenarios when I don't know exactly which is the first broken commit?14:19
killerchicken_ well, the goal of git bisect is help you find the first broken commit.14:20
svenfark When I would, and commits are small, I wouldn't need bisect?14:20
killerchicken_ the key point here is "known broken"14:20
so what you do is find one commit that is definitely ok, and one that is definitely broken14:20
and git bisect does the rest14:20
svenfark Ah, ok14:20
That was what I hoped14:20
killerchicken_ bisect is awesome, btw14:21
svenfark The sentence was little bit unclear to me14:21
→ kpreid_ joined14:21
svenfark Yeah, I find it extremely interesting14:21
So s/first/a/ is how to interpret that sentence?14:22
→ aziz joined14:22
← minefield left14:22
→ munichpython joined14:24
← munichlinux left14:27
← killerchicken_ left14:29
→ killerchicken joined14:29
killerchicken sorry, had a disconnect there14:29
was your question answered?14:29
svenfark killerchicken: Almost, under the shower, a second question came to my mind ;)14:35
When the first broken commit is known, does bisect go down to patch-level?14:35
killerchicken hm?14:35
svenfark ...because commits can be quite huge sometimes, depending on the project14:36
killerchicken oh14:36
no14:36
it tells you the broken commit14:36
svenfark ok14:36
killerchicken there is no sane way for a program to break down a commit14:36
svenfark thats fair enough14:36
killerchicken if your commits are huge, you're doing it wrong, most likely ;)14:36
svenfark hm, you're probably right14:36
yeah ;)14:36
It's not my project, I think I found a serious bug in KVM and the devs give bonus points when I do bisect it ;)14:37
→ d0ugal joined14:37
killerchicken I'm sure they do14:37
→ Rolenun joined14:37
svenfark I think even with bisect it'll take quite some time as the bug relates to live migration, so at least two machines involved, starting VMs, migrating them, checking liveness...14:38
It'll be quite some run script14:38
tetha heh, so it is more like 'write shell script to do the test, start the process and look again after a week'?14:39
→ codeswing joined14:39
svenfark I don't know yet if automating that will work14:39
→ Guest14599 joined14:39
killerchicken tetha: depends on how long a single test takes. git bisect finds bad commits very quickly, usually14:40
svenfark A hope a single test won't take longer than 15min14:40
tetha indeed, logarithmic time and all such nice things :)14:40
← jjuran left14:40
← jsquared left14:41
← rgr left14:42
← munichpython left14:44
→ sveajobb joined14:46
sveajobb Hi, I'm fiddling around in the .git/hooks/post-update hook. I do the following cd /path/to/cloned/rep/ ; git pull -f ; make && make install14:47
but I end up with a "Fatal: Not a git repository: '.'"14:47
what could be wrong? :)14:48
yziquel hi. i'd like to have a branch that is completely independent of existing branches. how can i do that? git branch makes it start from the current head. i understood that there was a startpoint argument, but how do i specify that i want a completely new branch, unrelated to the current code?14:49
→ schierbeck joined14:49
locklace yziquel: http://book.git-scm.com/5_creating_new_empty_branches.html14:49
yziquel locklace: thanks14:49
← candybar left14:49
→ masterhumper joined14:49
← masterhumper left14:50
← kipras left14:53
→ gnufied joined14:54
← malumalu left14:54
→ project2501b joined14:55
yziquel locklace: mighty helpful14:57
← project2501a left14:58
← svenfark left15:02
← UncleCJ left15:02
→ alley_cat joined15:03
← ProperNoun left15:05
← remi left15:07
→ karpar joined15:09
← maleknet left15:11
→ ProperNoun joined15:11
← ProperNoun left15:11
→ ProperNoun joined15:11
← Guest14599 left15:12
→ coolcat joined15:13
← cori[s] left15:14
evilsejo → sejo15:16
→ Bass10 joined15:18
→ sjiep joined15:18
→ remi joined15:18
→ candybar joined15:19
→ adamm joined15:20
← sjiep left15:22
→ icwiener joined15:23
→ cesario_ joined15:24
← Lake left15:24
← Lake_ left15:24
→ Lake joined15:26
→ Lake_ joined15:26
← vvsh left15:30
→ malumalu joined15:30
Lake made a terrible mistake15:31
and need help asap15:31
any takers?15:31
thiago_home no15:31
Lake :D15:31
thiago_home we don't know what your mistake was15:31
Lake k15:31
cbreak it was using windows15:31
or using svn15:31
or ... BOTH!15:31
killerchicken Lake : The biggest mistake you've made is not asking your question15:32
:)15:32
Rolenun sorry Lake, we dont offer psychiatric therapy for free =D15:32
Lake i merged a branch into master that was never meant to be. the merge brought in tons of commits. and now they are pushed to the repo for other devs...15:32
→ bodisiw joined15:32
Lake can i reverse the merge without leaving history?15:32
cbreak with git reset15:32
BUT15:32
killerchicken yes, but you will make everyone sad who pulled from you already.15:32
Lake no one has pulled15:33
cbreak you have to reset on every single repository15:33
Lake it just happened15:33
killerchicken Never assume devs don't have a script that pulls every 3.14 seconds15:33
cbreak maybe git reset --hard HEAD@{1}15:33
but check your reflog first15:33
killerchicken reset the branch, the do git push <remote> +branchname15:33
Lake omfg, you are git gods15:35
thank you, thank you15:36
cbreak no problem :)15:36
killerchicken sure, you're welcome. bank acct # follows via pm...15:36
;P15:36
sounds like some pretty juicy history if nobody is allowed to see it :)15:37
Lake killerchicken: totally15:38
;)15:38
i'd tell you about it.. .but then i'd have to... well, you know15:38
killerchicken kill yourself? No problem.15:39
:-)15:39
Don't make a mess15:39
Lake :D...exactly.15:40
cbreak, killerchicken: i gotta get my oil changed now, but really. thanks so much.15:40
→ killerchicken_ joined15:40
killerchicken_ hehe, looks like someone didn't get fired or worse ;P15:41
→ eg3 joined15:42
cbreak good :)15:42
eg3 i have a problem with mit git-repo15:42
killerchicken_ oh noes!15:42
eg3 i checked in files under "no branch"15:43
than i checked out the master branch with git checkout master15:43
← s1kx- left15:43
eg3 an now, the new added files are gone15:43
killerchicken_ git reflog15:43
cbreak yes. that's normal15:43
killerchicken_ find the stuff that's missing15:43
cbreak indeed. git cherry-pick HEAD@{1}15:43
killerchicken_ git cherry-pick <id>15:43
eg3 how can i switch back to the last added files?15:44
killerchicken_ cbreak: you're always assuming they didn't try all kinds of shit already before coming here ;P15:44
cbreak yes.15:44
because he didn't say he did do any shit :)15:44
killerchicken_ pff15:44
:-)15:44
→ vbgunz joined15:44
← killerchicken left15:44
cbreak eg3: git log -g shows you the reflog15:44
you can get the changes with cherry-pick to where you are15:45
it is working as expected that you lose commits that are not commited to a branch15:45
→ krawek joined15:45
→ karasiov joined15:47
→ munichpython joined15:49
← pheaver left15:49
→ pheaver joined15:49
← camonz left15:50
← ankit9 left15:51
← Plouj left15:51
← infogulch left15:52
← EricInBNE left15:52
eg3 greate, thank you cbreak :)15:53
killerchicken_ sitaram: hey. Sorry for not responding to your email, I've never looked at wildrepos.15:53
→ infogulch joined15:53
← hyperair left15:53
bodisiw if ive cloned a repository on windows and the line endings have got things screwed (i.e., it says i have modified files but i havent, just the crlf dont align)... then if i set autocrlf to true and do a git reset --hard, it should get things back to stable state right?15:53
→ scarabx joined15:54
cbreak you changed the line endings?15:55
that's a modification15:55
← eg3 left15:55
bodisiw hmm15:56
cbreak, but shouldn't git reset --hard revert that modification?15:56
← Geralt left15:56
cbreak it'll remove every modification15:56
bodisiw it does not remove any modification!15:56
maybe i need to update my git15:57
→ dreiss joined15:57
killerchicken_ bodisiw: well, you can set different lineending modes for windows15:57
→ tstclair joined15:57
→ jrmuizel joined15:58
killerchicken_ git reset --hard would respect them, I think15:58
Lake killerchicken_: i notice that my files are still in git's object history. should i use filter-branch to remove traces?15:58
bodisiw killerchicken_, right, i was hoping so15:58
killerchicken_ Lake: Nice to see someone being paranoid15:58
← munichpython left15:58
killerchicken_ Lake: git gc --prune=now should kill them, I think15:59
bodisiw but for some reason it doesn't seem to work right... i _know_ the dirty working tree is due to line ending incompatibility, but reset does not seem to switch it up15:59
cbreak git gc should not remove anything that's still referenced15:59
Lake so, killerchicken_ gc and then git push master +branch ?15:59
← groovy77 left15:59
killerchicken_ Lake: You'd need to run it on the server side, too15:59
cbreak: but it isn't referenced anymore16:00
cbreak the reflog16:00
Lake hm, the server is hosted, not sure I can get at it16:00
killerchicken_ Lake: Looks like you lose16:00
→ groovy77 joined16:00
cbreak the reflog is also considered for git gc, it won't collect stuff referenced by it16:01
killerchicken_ Lake: maybe git push allows you to kill some objects by sha116:01
but I don't know16:01
→ ndim joined16:01
← jschoolcraft left16:02
← tvw left16:02
killerchicken_ anyone knowledgable about git-svn here? I'd like to get $Revision$ expansion working like it works in svn - I'm forced to work with an svn repo, and a build-script assumes the Revision stuff is expanded -.-16:03
→ eg3 joined16:03
→ jschoolcraft joined16:04
parasti killerchicken_: git-svn doesn't handle those, from what I can see. so your question probably has more to do with something like gitattributes than git-svn16:08
killerchicken_ parasti: hm16:09
grml16:09
← jrmuizel left16:09
killerchicken_ I guess I'll just use plain svn whenever I need to build the thing16:09
→ jelmer joined16:09
← JackWinter left16:11
← eg3 left16:11
→ JackWinter joined16:12
→ pkw joined16:15
← jelmer left16:15
bodisiw there has got to be a sane answer to this... a fresh clone should not tell me i have made changes!16:15
← ReinH left16:16
← Tureba left16:16
→ t0rc joined16:17
← Lake left16:18
← Lake_ left16:18
← xknucp left16:19
drizzd_ bodisiw: from the documentation of autocrlf I cannot think of a way that would mark your files as modified after clone, no matter how you set autocrlf.16:20
If you change autocrlf _after_ you clone, that is another matter16:20
bodisiw hmm then its buggy16:20
it does not matter if i change the setting or not16:20
thiago_home recheckout the files16:21
→ psoo joined16:21
bodisiw that doesnt work either, if you refer to git checkout -- filename16:21
thiago_home I refer to rm -rf *; git reset --hard16:22
drizzd_ bodisiw: you're not using jgit by any chance?16:22
bodisiw oh, i have to rm them first?16:22
okay, lemme try it16:22
i thought reset --hard would do that without deleting them first16:22
jast only if reset --hard is aware that the files changed16:23
bodisiw oh okey16:23
jast randomly changing autocrlf always makes these things a bit tricky16:23
bodisiw drizzd_, not jgit afaik, the git.cmd from git extensions (google code)16:23
drizzd_ if git status says the files are modified, why would git reset not know16:23
jast good point16:24
← raz left16:24
jast I didn't follow the discussion16:24
thiago_home because the feature is broken16:24
jast only just got here16:24
drizzd_ although I'm not sure how that works16:24
→ qshan joined16:24
← sde_antrix1 left16:24
cbreak wouldn't rm -rf * delete your .git folder?16:25
← ijuma_ left16:25
drizzd_ bodisiw: git extensions seemed rather broken last time I tried. Is ist usable?16:25
bodisiw drizzd_, well, you see how usable it seems for me :-P16:25
is there a preferable windows client?16:25
locklace cbreak: try typing "echo *" in one of your git working directories16:26
drizzd_ bodisiw: the official one16:26
← icwiener left16:26
bodisiw okay, last time i tried there it didnt seem to exist!16:26
i will check that... i hope it makes things easier :-)16:26
→ sjiep joined16:27
thiago_home cbreak: * doesn't matcn .*16:27
not by default anyway16:27
cbreak hmm. Interesting.16:27
drizzd_ FWIW, I just tried cloning a repo with core.autocrlf = true, and it also shows all files as modified16:27
cbreak are the files in the repository crlf?16:27
→ icwiener joined16:28
drizzd_ yes16:28
→ kukks joined16:28
bodisiw cbreak, that's kind of my suspicion as well at this point... i'm going to ask the maintainer16:28
cbreak then the files will get automatically converted to LF, so they changed from the repository?16:28
drizzd_ oops, make that a no. And I think you didn't ask me.16:28
bodisiw my understanding is that files in repository should always be LF to make things work nicely?16:28
→ Yuuhi joined16:30
sjiep I have a question (more a conformation0 on a basic concept of git:16:32
with svn you have a central repository where you do you commits and updates (thats why it's centralized version control), with git I have the idea that there can be hundreds of repositories (every computer managing there own) that can be linked together to do version control.16:32
Is this idea correct?16:32
drizzd_ bodisiw: right, now that cbreak asked I know what's going on. the reasons your files are shown as modified is this: they already have CRLF format in the repo, but autocrlf strips CR from the work tree files.16:32
→ offby1` joined16:32
bodisiw drizzd_, right on... so what is the cut-to-the-chase way to fix it?16:33
← offby1 left16:33
offby1` → offby116:33
← offby1 left16:33
→ offby1 joined16:33
drizzd_ bodisiw: git config --global autocrlf false16:33
sjiep anyone?16:33
drizzd_ bodisiw: git config --global core.autocrlf false16:33
cbreak maybe you can rewrite history16:33
bodisiw i am writing to the project maintainer right now to try to get him to pay attention and give consistent line endings in the repo16:33
cbreak sjiep: yes16:33
but you can also do it with just one repo16:33
→ nonken joined16:34
→ chewbranca joined16:35
drizzd_ I wonder why msysgit insists on having autocrlf = input by default. It's not like CRLF is such a big deal, is it?16:35
sjiep cbreak: thx, my situation is as follows, I have a NAS at home as a central storing point for al my files, all files on the NAS can be easily backup-up. So logically I would to store all my repositories on the NAS aswell, how would you approach this with git?16:35
cbreak you waste ONE BYTE per line16:35
drizzd_ cbreak: haha16:36
→ solofight joined16:36
cbreak sjiep: just make a bare repo on the nas16:36
and push to it16:36
bodisiw i think broad compatibility is the main issue... if you use vim, youre fine with whatever16:36
cbreak (for each repo you use)16:36
drizzd_ bodisiw: I think "notepad.exe" is one of the few editors that can't deal with it16:36
bodisiw but maybe if you use visual studio or some other wacky windows editor, it will break without 'proper'' crlf... not to mention it16:37
← ProperNoun left16:37
← nitrate left16:38
sjiep cbreak: don't you need a repo on the computer you are pushing from then as well?16:38
bodisiw drizzd_, well there is also the issue of new files created using whatever default eol16:38
cbreak sjiep: yeah, of course16:38
one repository with working directory to work with16:38
→ ijuma_ joined16:38
bodisiw perhaps i am overanalyzing all this... and should just set autocrlf to false on all platforms and situations16:38
cbreak one bare to push to on the nas16:38
maybe an other repo with working dir on the laptop where you can pull to16:38
Rolenun drizzd_: msysgit defaults to your settings during installation. (it asks you which line endings you want to use)16:38
cbreak and an other repo on some hosting service for public interaction16:39
ben_alman → ben_irish16:39
drizzd_ Rolenun: yes, but when it asks, there is a choice selected by default. Most people seem to have autocrlf set to true or input by default.16:39
bodisiw there is conflicting advice all over the web on what makes sense for the autocrlf settings16:40
Rolenun drizzd_: Would you prefer an option specifying None and forcing ppl to select it :D16:40
drizzd_ Rolenun: yes16:41
in fact, I think it would be better not to ask at all16:41
it can always be activated later16:41
sjiep cbreak: so you make an empty repo on the NAS, then you make a repo on your local computer where you start adding files to and you can then just push these to the empty on on the NAS (and indeed as you said if you have a (empty) repo on a laptop you can easily pull from the repo on the nas to get the project on your laptop)16:41
Rolenun drizzd_: it's just an option lol, it can always be set later ;)16:41
cbreak doesn't matter16:41
you can clone them afterwards too16:41
if you have one repo, you can make as many more as you want16:42
→ ph^__ joined16:42
sjiep mmm, ok I think I get the idea, just confused me at the start a bit because I was used to svn which has a slightly diffferent approach16:42
I guess I'll just have to start working with it and learn on the way ! :D16:43
cbreak: thx for the help!16:43
Rolenun sjiep: http://progit.org/book/16:43
cbreak http://eagain.net/articles/git-for-computer-scientists/ :)16:44
← qshan left16:44
sjiep Rolenun: I'll give them a read, thx!!16:44
Rolenun cbreak: oh brother, not that crackpot article :D16:44
→ offby1` joined16:44
cbreak it's nice.16:44
it has PICTURES!16:44
Rolenun sjiep: np :)16:44
lol16:44
→ ProperNoun joined16:45
← ProperNoun left16:45
→ ProperNoun joined16:45
sjiep haha lo, those picture will keep me going :P16:45
it reminds me of uni16:45
← offby1 left16:46
← JasonWoof left16:50
→ sumangurung joined16:50
→ nitrate joined16:50
← ProperNoun left16:51
← karpar left16:53
ryanakca Is it possible to merge only changes to a certain file from branch B in branch A?16:54
thiago_home no16:54
a merge merges all changes16:54
← nitrate left16:55
← sumangurung left16:56
bodisiw drizzd_, thank you so much for the help. i have wasted a ridiculous amount of time trying to get a straight answer on that eol question16:57
→ AkumaStreak joined17:00
drizzd_ yw17:01
← pantsman left17:04
← sshc left17:06
← kukks left17:08
ryanakca thiago_home: Thanks17:08
→ |Rolenun| joined17:09
← jceb left17:09
← JackWinter left17:09
→ JackWinter joined17:10
← Rolenun left17:12
← solofight left17:12
jschoolcraft is there a good way to use git-svn for a svn repo with 22,000 commits?17:12
not specifying revision takes days, makes the .git directory almost 1GB and it's still going17:13
using git-svn init, then git-svn fetch -r 22414 doesn't return what I expect it to (the repo at that revision)17:13
i'll RTFM, just point me to it.17:14
|Rolenun| → Rolenun17:14
← Rolenun left17:15
→ Rolenun joined17:15
← pheaver left17:15
← psoo left17:16
parasti jschoolcraft: what does it return? if the entire repo is too big and you don't care about older commits, git svn clone -r HEAD will DTRT (as should init+fetch+checkout)17:17
→ mmw joined17:18
← spaceonline left17:21
→ mutex0 joined17:22
mutex0 anyone know where the p4merge tool for mac went? it seems to be missing :P17:22
← gitster left17:23
← Mizar-jp left17:24
← mmw left17:26
← anno^da left17:26
→ mmw joined17:27
→ drizzd joined17:27
→ icwiener_ joined17:27
← dreiss left17:29
→ twirp joined17:29
← icwiener left17:30
→ hipe joined17:31
← drizzd_ left17:31
← DavidKlein left17:31
→ DavidKlein joined17:32
→ MJBoa joined17:33
← mutex0 left17:34
← ph^__ left17:34
← chewbranca left17:34
← nonken left17:34
→ nonken joined17:36
← MJBoa left17:36
→ ivenkys joined17:36
← yziquel left17:38
→ pheaver joined17:38
→ Bluehorn joined17:39
Bluehorn hoi17:39
jotik hey17:39
← bambule left17:40
Bluehorn Somebody here having experience with git bisect? It keeps telling me that I have 8 steps left.17:40
→ kukks joined17:40
thiago_home even after you tell it good/bad?17:40
Bluehorn thiago_home: I had to use git bisect skip two times and now even good/bad does not make it progress.17:41
thiago_home I don't know then17:41
Bluehorn thiago_home: Unfortunately, the documentation does not state exactly how skips are handled.17:41
thiago_home: too bad, thanks anyway17:41
It's only that I compiled the kernel some 20 times already, and I don't have the time to add another 50 on top of that.17:42
charon Bluehorn: well, since this is the kernel... perhaps you can post 'git bisect log' for investigation?17:42
(and/or reuse/modify it once you know what's wrong)17:43
→ Mizar-jp joined17:43
Bluehorn charon: I wanted to finish bisecting and report it to lkml.17:44
charon: I guess that skipping will reduce the performance of git bisect considerably, I am just wondering how bad it can get. And if I have any way to find out without reading the log and comparing it with git log of the tree.17:44
← nonken left17:46
charon Bluehorn: btw, i think there were some changes to bisect a few months back. what's your git version?17:46
Bluehorn charon: 1.6.6.117:47
→ bambule joined17:47
charon ah, so you already have those17:48
← jasonn1234 left17:48
charon (62d0b0d is in 1.6.4)17:49
Gitbot [git 62d0b0d]: http://tinyurl.com/yjarm4m -- bisect: when skipping, choose a commit away from a skipped commit17:49
← twirp left17:49
← engla left17:49
→ twirp joined17:49
→ nonken joined17:51
Bluehorn charon: hmm, makes sense, as one would assume that a series of commits is broken17:51
→ Geralt joined17:51
← Geralt left17:51
→ Geralt joined17:51
→ mtheoryx joined17:52
← hgimenez left17:52
Bluehorn torsten@pulsar:~/mirror/linux-xfs-bisect$ git bisect good17:52
Bisecting: 87 revisions left to test after this (roughly 7 steps)17:52
hmpf, it keeps saying that17:53
→ sumangurung_ joined17:53
← sumangurung_ left17:53
→ hgimenez joined17:53
offby1` → offby117:53
← offby1 left17:54
→ offby1 joined17:54
→ jasonn1234 joined17:55
← offby1 left17:55
← Geralt left17:56
← jasonn1234 left17:56
← bambule left17:56
→ arm2 joined17:58
→ bambule joined17:58
arm2 hack tools exploits,vulnerabilities,videos,shellcode,ubuntu on www.team3d.3xforum18:01
← arm2 left18:01
← mindworx left18:03
← mssm left18:04
← kopa left18:04
← gnufied left18:04
→ mssm joined18:04
→ gnufied joined18:06
← gnufied left18:06
← et left18:06
→ et- joined18:06
et- → et18:06
← twirp left18:07
← kiva left18:07
→ kiva joined18:07
→ nitrate joined18:10
→ tatsuya joined18:12
→ solofight joined18:13
← State left18:13
← solofight left18:13
→ solofight joined18:14
→ mindworx joined18:14
→ archis joined18:14
← solofight left18:16
→ _Vi joined18:16
← Nugget left18:19
→ TTilus joined18:20
Bluehorn okay, it is down to 12 revisions and 4 steps :-)18:20
← Determinist left18:21
ben_irish → ben_alman18:22
→ ATH500 joined18:22
← krawek left18:23
→ Nugget joined18:24
← nonken left18:29
→ sh1mmer joined18:32
← lemonchicken left18:33
← ajpiano left18:33
← adamm left18:35
← ph^_ left18:38
→ lemonchicken joined18:39
→ mithro joined18:39
→ ajpiano joined18:39
← Pupeno left18:44
← Yuuhi left18:47
karasiov olo18:47
← Textmode left18:49
← sgh left18:51
→ Yuuhi joined18:53
← sh1mmer left18:53
→ psoo joined18:53
→ |Rolenun| joined18:53
→ sh1mmer joined18:53
→ Chillance joined18:54
|Rolenun| → Rolenun_18:54
← Rolenun_ left18:55
→ Rolenun_ joined18:55
← Rolenun left18:55
Rolenun_ → Rolenun18:55
→ alt-dot-net-geek joined18:56
← project2501b left18:57
→ ph^_ joined18:58
← psoo left19:00
← kiva left19:01
→ kiva joined19:01
← gambler left19:02
→ MJBoa joined19:05
→ sshc joined19:06
→ bdimcheff joined19:09
→ jepster joined19:10
jepster hi, is it better to have a testing branch for a project with only one developer?19:10
→ Chaser_ joined19:10
Ilari jepster: I would say yes.19:11
→ r11t joined19:11
jepster can guit generate automatic version-numbers?19:11
git*19:11
→ ereslibre joined19:12
← ereslibre left19:12
→ ereslibre joined19:12
→ priidu joined19:13
Ilari jepster: 'git describe'?19:14
jepster: (but that needs at least one annotated tag in past).19:14
→ godsflaw joined19:15
← mmw left19:18
← ereslibre left19:22
→ jrmuizel joined19:23
→ project2501a joined19:29
← xinming left19:29
→ |Rolenun| joined19:29
→ reto` joined19:29
← jrmuizel left19:30
→ Raydiation joined19:31
← simNIX left19:31
Raydiation hi, i just switched my pc, how can i get access to my master repo and start commiting?19:31
→ justin-george joined19:32
Ilari Raydiation: Clone the repo?19:32
← Rolenun left19:32
Raydiation Ilari: did that anything else?19:32
set the global user and email via global conifg19:32
Ilari Raydiation: Yeah, that's a good idea.19:33
← codeswing left19:34
Ilari Raydiation: Well, that should be about it... Use 'git branch' and 'git branch -r' to show what you have.19:34
← aleray left19:34
Raydiation hm when i try to push i get a non fast forward error19:38
http://paste.pocoo.org/show/174676/19:39
→ jrmuizel joined19:39
Raydiation git pull repo doesnt solve it19:39
← JackWinter left19:40
Ilari Raydiation: 'git pull <repo> master'?19:40
Raydiation: (while on master)?19:40
Raydiation: <repo> is likely 'origin'.19:40
← justin-george left19:40
→ beni- joined19:40
Raydiation LiamH: yes19:40
Ilari: yes19:41
→ Lake_ joined19:42
→ Lake joined19:42
← karasiov left19:42
← pheaver left19:43
← jrmuizel left19:44
← jspiros left19:45
Raydiation hm just to get it right: Merge remote branch 'laudio/master'19:45
that means i have merged my branch with the master branch19:45
do i have to merge this everytime?19:45
→ JackWinter joined19:45
Raydiation or will git push laudio master be enough?19:45
Ilari Raydiation: You get that "not fast forward" error if remote side has commits you don't have.19:46
Raydiation hm i pushed everything i had when i left19:47
no one else has access to it19:47
so there shouldnt be changes19:47
Ilari Raydiation: You cloned this repo after that? Because if you did it before, this repo doesn't have your newly pushed changes.19:47
Raydiation: ... Unless you fetch and merge those changes.19:48
Raydiation yes i cloned it now19:49
→ gasull joined19:49
Raydiation i fetched and merged19:49
→ jrmuizel joined19:50
Ilari Raydiation: You could use 'gitk HEAD@{1}..HEAD' to check what's new.19:51
← radarek left19:51
→ gambler joined19:52
← nitrate left19:53
← christophsturm left19:53
gasull Hi all. what is the git-svn equivalent of "svn co"? Thank you.19:57
→ _Rolenun_ joined19:58
gasull git-svn co didn't work for me19:59
← warthog9 left19:59
← |Rolenun| left20:01
← jepster left20:03
→ paltman joined20:05
→ tim joined20:06
gasull OK. I found it: git svn clone20:06
→ paltman_ joined20:06
← gasull left20:06
← jhelwig left20:06
→ jhelwig joined20:07
tim hi, trying to use git send-email via my smtp server, it dies at line 984: $smtp->mail( $raw_from ) or die $smtp->message; .... any idea, what i am doing wrong?20:07
← paltman left20:10
paltman_ → paltman20:10
→ MikeJ joined20:10
← sh1mmer left20:10
→ sh1mmer joined20:10
← _Vi left20:10
MikeJ hello #git, wondering if anyone has a clever way to do the equivalent of ignoring whitespace for diff, but for changes in whitespace that span multi line, like a \n added in the middle of a string20:11
at least visually --color-words='[a-z]|[^[:space:]]' brings me pretty close20:12
but really, I want to do a diff with just the hunks that the above colors20:12
← AkumaStreak left20:13
→ sumangurung joined20:13
← scarabx left20:14
← priidu left20:14
→ mmw joined20:18
ajpiano → bot-t20:19
bot-t → Guest7428420:19
→ nitrate joined20:19
Guest74284 → ajpiano20:20
← sumangurung left20:20
← beni- left20:20
→ NET||abuse joined20:21
→ AkumaStreak joined20:21
← kpreid_ left20:21
remi has anyone had recent problems with git + vim (or macvim) for commit messages?20:22
→ kpreid_ joined20:22
remi whenever I type git commit, I immediatly get "Aborting commit due to empty commit message." even before macvim's window is open20:22
and with vim, I can type a commit message, but when I save and close, I get this: "error: There was a problem with the editor 'vim'."20:23
tpope mvim forks into the background. you need mvim -f20:23
adam12 remi: vim -f20:23
→ scarabx joined20:24
tpope and the vim that ships with os x has been patched into retardedness by apple20:24
→ strongorder joined20:24
remi tpope + adam12: same thing happens with either "mvim -f" and "vim -f"20:25
← ph^_ left20:25
→ _Vi joined20:25
patrikf remi: how did you change the editor setting?20:25
remi patrikf: I tried $EDITOR first, then git's core.editor20:26
patrikf remi: $VISUAL?20:26
→ |Rolenun| joined20:26
tpope the issue with the apple-raped console vim is that it's patched to exit with nonzero status if ANYTHING generates an error. like if you accidentally type :W or if any of your plugins use a try/catch block20:26
→ Pewpewarrows joined20:27
remi tpope: oh, that sucks20:27
tpope you can symlink vim to mvim and it will work in the console20:28
or install another vim, e.g., macports20:28
← jrmuizel left20:28
→ webus joined20:30
← _Rolenun_ left20:30
remi tpope: I found this though: http://blog.martinemde.com/post/348706717/set-editor-to-macvim-for-git-commit-messages20:30
tpope looks okay though I don't know what that autocmd does20:32
← paltman left20:32
|Rolenun| → Rolenun20:32
← Rolenun left20:32
→ Rolenun joined20:32
tpope also, VISUAL is preferred to EDITOR. apparently purchasing a mac requires you to be completely ignorant to unix history20:32
← tatsuya left20:33
→ sgh joined20:34
→ ph^_ joined20:34
tpope can you tell I'm a reluctant bitter apple user?20:34
remi tpope: export EDITOR='mvim -f' finally worked (although it didn't seem to work the first time I tried it...)20:35
← sgh left20:37
→ agliodbs joined20:37
→ fmardini_ joined20:37
→ dreiss joined20:37
→ v0n joined20:38
← rbq left20:38
← _iron left20:39
→ _iron joined20:41
locklace tpope: ahaha, that's so mind-numbingly lame20:41
but on par with what i expect of apple20:41
→ sgh joined20:42
← _Vi left20:43
→ chrisf joined20:45
← Will left20:48
→ Will joined20:48
chrisf hi #git, i'm been trying to get a newbie on my team to rebase onto upstream before she pushes her changes, but for some reason `git pull --rebase` is gets confused and rebases commits that are already in the upstream too (applying the second copy on the top of the new HEAD). This does not happen on my machine, only hers, so I suspect something is broken in her config. What should I be looking for?20:49
→ magnus_ joined20:49
← webus left20:49
magnus_ → Guest1166220:49
Guest11662 hey! i've created a submodule in my git repo which points to another repo. the submodule repo has been updated with a new branch, which i would like to switch to in my "main" project, but "git pull" and then "git checkout <branch>" doesn't allow me (it says <branch> did not match any files know to git)20:50
so basically, im not sure if my pull actually did get the new branch.. but it mentioned it in the pull20:50
→ jasonn1234 joined20:51
chrisf Guest11662: `git branch -av` to see where everything is pointing.20:52
→ seh joined20:52
seh → Guest7441820:52
chrisf Guest11662: chances are you don't actually have a *local* branch for what you're trying to switch to.20:53
Guest11662 oh, ok. i see the problem now. the newly created branch doesn't seem to be a local branch20:53
do i always need to create any new branches created outside my repo, when i pull?20:53
chrisf ~ when you *fetch*.20:53
pull does not mean what you think it means.20:54
← dreiss left20:54
Guest11662 sorry, when i fetch20:54
Guest74418 I need some help figuring out how to abandon some local changes.20:54
I have a repository cloned from a public one hosted at GitHub. I never modify any of the files.20:55
chrisf yes, the fetch is just going to create remote branches. if you want to work with them locally, `git checkout $remote/$branch` then `git checkout -b $branch` to make a local branch to work with.20:55
locklace Guest74418: git reset --hard20:55
Guest74418 Yet today I ran 'git pull', and it failed with the "not uptodate. Cannot merge" message.20:55
I've been trying that, and nothing I do seems to abandon these changes.20:56
chrisf Guest74418: check the output of `git status`.20:56
Guest74418 I run "git reset --hard", it completes successfully, but afterward "git status" reports the same modifications present.20:56
chrisf do you have *new* files there?20:56
→ hegge joined20:57
Guest74418 Well, /I/ didn't put them there, but I don't know if they're new to the current HEAD at which branch "master" points.20:57
How could I tell?20:57
chrisf quick & dirty: `git add .; git reset --hard`20:57
Guest74418 I tried that. It didn't have any apparent effect.20:58
chrisf it doesn't know to reset *new* files that arent in the index.20:58
Guest74418 Can I just delete them?20:58
chrisf pastebin the output of `git status` kthx20:58
Guest74418 One moment.20:58
http://www.pastebin.org/8720120:59
← tim left20:59
chrisf odd, that should just work.21:00
← selckin left21:00
Guest74418 % git remote show origin21:00
* remote origin21:00
Fetch URL: git://github.com/richhickey/clojure-contrib.git21:00
Push URL: git://github.com/richhickey/clojure-contrib.git21:00
HEAD branch: master21:00
Remote branches:21:00
1.0.x tracked21:00
1.1.x tracked21:00
clojure-1.0-compatible tracked21:00
gh-pages tracked21:00
master tracked21:00
maven tracked21:00
new tracked21:00
Local branch configured for 'git pull':21:00
RandalSchwartz wipes the paste from his screen21:00
→ selckin joined21:00
Guest74418 Sorry about that.21:00
→ _Vi joined21:01
→ Axius joined21:02
chrisf try forcing the target of th: `git reset --hard master`21:02
Guest74418 Some of the changes appear to just be whitespace -- maybe CRLF conversion.21:02
locklace chrisf: does git log origin/master.. (or whatever) in her repo show the expected thing21:02
→ melter joined21:02
Guest74418 Running "git log origin/master" does give credible output, yes.21:02
→ resmo_ joined21:02
resmo_ hi21:02
resmo_ → resmo21:03
drizzd Guest74418: do you have autocrlf set to true?21:03
Guest74418 Yes.21:03
chrisf Guest74418: you're getting pwned by crlf conversion.21:03
Guest74418 I think I have that set globally.21:03
Ah.21:03
chrisf you, uh, don't want to do that ;)21:03
locklace: afaik, yes.21:04
Guest74418 I've always been confused on this front, because I run git within Cygwin on Windows, and all kinds of trouble ensues.21:04
drizzd Guest74418: you probably have CRLF line endings in the repo, so autocrlf will be confused.21:04
isn't git via cygwin slow?21:05
thiago_home yes21:05
Guest74418 Not that I can tell.21:05
chrisf Guest74418: is there a reason you don't use msysgit? works fine for me, and it's fast.21:05
→ denny joined21:05
Guest74418 I have Cygwin set up and working, and git comes as a package, and seems to work well enough so far.21:05
I can try msysgit, though. First, I'll try changing my global core.autocrlf setting.21:06
denny git diff gives me the difference what I've commited locally and what's currently in the files (I think) - what gives me the difference between what I've commited locally and what I've pushed to master?21:06
Guest74418 That fixed it. Thank you.21:07
chrisf denny: try something like `gitk origin/master...master`21:07
drizzd denny: if by pushed to master you mean "pushed to master on origin", it's git diff HEAD origin/master21:08
denny great, thanks21:08
← sh1mmer left21:10
→ sh1mmer joined21:10
locklace chrisf: strange one. be interesting to see her config and output of remote show origin, log origin/master.., pull --rebase -v, maybe compare what happens with a manual fetch and rebase21:11
← Guest74418 left21:11
← Lake_ left21:12
← adam12 left21:12
← sh1mmer left21:12
chrisf locklace: thanks. will have a look when she next surfaces, and i might be back :)21:13
tetha hm, I just had the problem of accidentially adding some really large files, resulting in endless pushes and other annoying things. is there a way to see the size of the treeish at a certain commit?21:14
cbreak every commit since you added the big file contains the big file21:16
→ froschi joined21:18
→ Mage joined21:18
→ catalis joined21:20
denny right, I tried to do a git pull and it threw a load of conflict errors. So I edited all those files and removed the conflicts, and tried to git commit (which is what it said to do) and it says 'needs merge' for a load of the files. What am I missing?21:22
cbreak you have to add the files you changed21:22
to resolve the conflicts21:22
denny thanks :)21:23
← alt-dot-net-geek left21:24
→ alt-dot-net-geek joined21:25
denny hooray, I think my machines are back in sync21:26
I wish I knew what I'd done to screw them up that badly, I didn't htink I was doing anything complicated21:26
cbreak, drizzd, thanks for the help21:26
→ triple_xxx joined21:27
triple_xxx trying to understand git the way it tracks updates to packages before merging??21:28
← bl4 left21:28
triple_xxx is there a resource I can read????21:28
locklace triple_xxx: not following your question, but there are plenty of resources; see http://git-scm.com/documentation21:29
cbreak packages?21:30
no context found21:30
triple_xxx git tracks project history correct...21:30
→ Pupeno joined21:30
← d0ugal left21:31
→ VeXocide joined21:31
VeXocide hi, I've got this situation, http://codepad.org/n79GxqLS and would like to copy the upstream/0.6-devel branch to origin, how would I do so?21:32
triple_xxx git more a developers tool ... then to manage package installs on a computer...21:33
cbreak packages?21:34
again, I have no idea what you mean21:34
locklace triple_xxx: git is not a package manager21:34
triple_xxx cbreak: I use pacman for arch linux was wondering if it tracks when I would apply new package and keep track of versions updated...??21:34
locklace VeXocide: something wrong with git push origin upstream/0.6-devel?21:35
← MikeJ left21:35
cbreak git doesn't care in the slightest what system you are on21:35
VeXocide locklace, no, just didn't know how to achieve it, let me give that a shot21:35
cbreak it doesn't care if you update your linux or if you install internet explorer21:35
locklace triple_xxx: not the right tool. try asking for help in #archlinux regarding what it is you're specifically trying to achieve21:35
VeXocide locklace, that didn't seem to do the trick :/21:36
cbreak get the branch to local21:37
with git checkout -t upstream/...21:37
then push it21:37
locklace VeXocide: maybe you need to create a local tracking branch for it first with git branch --track and push taht21:37
← godsflaw left21:37
triple_xxx locklace: just wondering if the tool could track package services across a network to see what is installed on most computer... git is one powerful tool wish it worked this way... :(21:38
→ godsflaw joined21:38
VeXocide locklace, so git branch --track upstream/0.6-devel; git push origin upstream/0.6-devel right?21:38
← godsflaw left21:38
locklace triple_xxx: you could force it into doing something like that but it's almost certainly not what you really want21:38
← denny left21:38
triple_xxx locklace: I see that , but it does have one power idea of management over system development.. that really makes sense :)21:39
power=powerful21:39
cbreak git couldn't care less about your system21:40
locklace triple_xxx: people have encountered and solved almost every sort of problem related to package deployment on networks of similar hosts; you should really ask for help in your os channel21:40
→ tlb joined21:40
← tchan left21:40
locklace VeXocide: git branch --track 0.6-devel upstream/0.6-devel, then git push origin 0.6-devel21:40
cbreak I used checkout for a reason :)21:41
you can use branch too21:41
→ bl4 joined21:41
cbreak idea is to try to create a local branch before pushing21:41
← Pewpewarrows left21:42
VeXocide locklace, brilliant, that seemed to do the trick21:42
triple_xxx locklace: not even close in this sense .... As IT personel there idea is probe the system and send over the package for installation nothing keeps this idea on what worked and what doesn't... Most databases would require manual entry and level alot to be desired...21:42
cbreak what package?21:43
git tracks file contents21:43
locklace triple_xxx: nevertheless, that's really an os deployment issue, and should be discussed in that context. if your os community isn't knowledgeable or helpful enough, consider using a different os21:44
← jasonn1234 left21:44
← aspotashev left21:44
← RdeWilde_ left21:44
triple_xxx locklace: ok will run along21:44
← triple_xxx left21:44
→ jasonn1234 joined21:45
→ jsonau joined21:47
jsonau ello guys, i need to reformat and i gitted a folder. what do i do to "export" this?21:48
← vbgunz left21:48
locklace jsonau: export how21:48
← fmardini_ left21:49
→ pklingem1 joined21:49
cbreak just copy the folder somewhere21:51
if you want to keep the history21:51
← remi left21:52
← Axius left21:53
→ kkszysiu joined21:53
kkszysiu hello21:54
→ johnw joined21:54
kkszysiu if I have fork of github branch. Can I split two forks into one?21:54
dennda that wouldn't be a split but a merge21:54
kkszysiu but Im talking about for example21:55
→ Pewpewarrows joined21:55
kkszysiu from http://github.com/kordless/zoto-server to http://github.com/kkszysiu/zoto-server21:55
he made changes and I21:55
cbreak you want to merge stuff?21:56
kkszysiu yeah21:56
cbreak kkszysiu: man git-merge might help21:57
Gitbot kkszysiu: the git-merge manpage can be found at http://git.or.cz/man/git-merge21:57
← JoeConyers left21:59
← strongorder left22:00
← jasonn1234 left22:01
← David-T left22:01
→ jmcantrell joined22:03
← schlortoo left22:04
→ charlenopires joined22:06
→ hyperair joined22:06
→ bauruine joined22:08
← jschoolcraft left22:09
← khmarbaise left22:10
aCiD2 Can I cherry-pick a set of revisions in sequence?22:12
← melter left22:12
← jmcantrell left22:14
|Mike| Hmz, who's using trac with git ?22:14
http://pastie.org/81278122:15
thiago_home aCiD2: you can cherry-pick any revision22:16
aCiD2: singular22:16
aCiD2 right, but I need to cherry pick a range22:17
→ Thell joined22:17
aCiD2 (We're migrating from svn to git, and the new git-svn is different from the old one as we used an authors file)22:17
thiago_home you can't pick a range22:17
aCiD2 barn22:17
darn*22:17
thiago_home you can pick only one commit, one at a time22:17
aCiD2 oh well, have to move them over one by one22:17
aCiD2 sets a script up22:18
→ rbq joined22:18
thiago_home you can do a rebase --onto22:18
cbreak you can merge whole branches22:18
thiago_home please read the manpage and understand the graphs22:18
aCiD2 that gives a ton of corflicts22:18
like I said, the commits are different22:18
so a rebase doesn't work22:18
there are /no/ common commits22:18
→ godsflaw joined22:19
→ snearch_ joined22:19
thiago_home the rebase has to work22:19
if it doesn't, cherry-pick won't work either22:19
rebasing and cherry-picking a range of commits is the same operation22:20
git format-patch | git am is also the same operation22:20
aCiD2 hrm, cherry-picking has worked but rebase always seems to conflict22:21
maybe the conflicts aren't me assuming rebase is failing but is a real conflict22:21
← sgh left22:21
→ aresnick joined22:25
← Helios left22:26
← nitrate left22:26
← catalis left22:27
→ Helios joined22:30
← mtheoryx left22:31
← til_ left22:31
→ David-T joined22:31
aCiD2 Also, we've just set up gitweb - does gitweb support doap files?22:31
David-T → Guest7935422:31
thiago_home doap?22:32
← Helios left22:32
→ MikeJ joined22:33
aCiD2 http://git.gnome.org/browse/nanny/tree/nanny.doap22:33
chris_n what sort of issues will 'git push --force foo' cause for those who are fetching branch foo?22:33
aCiD2 cgit supports it, wondered if gitweb did22:33
doesn't seem to pick my file up though, which is a shame22:33
cbreak chris_n: depends on what you push22:33
← snearch_ left22:33
cbreak if you push an ancestor of what was there before22:34
the people will not notice it22:34
or worst case push back what you removed22:34
if you push something completely different22:34
they will get trouble when trying to pull22:34
the nasty merging kind of trouble22:34
if you push a descendent, you wouldn't have to force it22:35
→ chewbranca joined22:36
chris_n cbreak: foo tracks a branch foo in one remote; I rebase, do work, commit, and then push to a second remote and have to use --force22:36
→ dreiss joined22:36
chris_n is there a way to avoid the use of --force here?22:37
→ Helios joined22:37
← project2501a left22:37
cbreak if you rebase22:37
you have to push with force22:37
because it falls under the "completely different"22:37
chris_n ahh22:37
cbreak people should NOT pull from such locations22:37
instead, they should probably just clone again each time you change something22:38
no one should base any work on it22:38
for obvious reasons22:38
they could fetch + reset maybe.22:38
← jsonau left22:39
chris_n cbreak: is there a better way for me to follow one remote branch, commit my work on top, and push to my own remote repo?22:39
← Guest79354 left22:40
→ tvw joined22:40
cbreak depends. you could instead of rebasing merge instead22:40
→ ntoll joined22:42
← corni left22:43
locklace chris_n: the fundamental rule of rebasing is: **do not rebase commits that you have pushed to a public repo**22:44
→ snearch_ joined22:44
→ riz2 joined22:45
→ mtheoryx joined22:45
chris_n so I really should have a tracking branch, rebase that branch; create a working branch, merge the tracking branch with the working branch, push the working branch to my remote repo?22:45
← snearch_ left22:45
→ EXAAB3AV joined22:45
← Kissaki left22:46
locklace chris_n: your options are to either merge instead of rebase, or push with -f/--mirror and require everyone to clone your repo or rebuild your branch from scratch (and all work based on it) whenever you push22:46
chris_n k, I think I understand, tnx22:47
→ Kissaki joined22:48
← Araneidae left22:48
→ Araneidae joined22:49
← charlenopires left22:51
← Zephyrus left22:52
← mdel left22:52
← dreiss left22:56
→ impulze joined22:57
impulze heya is it possible to set per-repository configuration stuff (send-email/format-patch) without putting it in .git/config in the repository (because it's shared by it)?22:58
cbreak why not?22:59
impulze how? :P22:59
cbreak I mean, why not into .git/config?22:59
it's per repo, so it should be in the repo22:59
impulze because the repos shares it22:59
cbreak it only applies to the repo22:59
impulze ?22:59
the repository has it in its tracked files22:59
so it always considers my config changed23:00
cbreak no23:00
impulze hm?23:00
→ dreiss joined23:00
cbreak everything inside .git is not tracked23:00
impulze oh ok :)23:00
ta23:00
← rseifert left23:00
→ sumangurung joined23:01
pasky Is there some good single-paragraph summary of the notes implementation?23:02
→ Digicrime joined23:03
→ nitrate joined23:03
← Digicrime left23:03
pasky there's git-notes.txt manpage but it's remarkably terse23:03
← godsflaw left23:04
pasky ah, core.notesRef is more verbose on that23:04
← sumangurung left23:06
→ psoo joined23:06
← Paraselene_ left23:07
→ Paraselene_ joined23:07
← rbq left23:10
← sshc left23:14
→ b14ck joined23:14
← nitrate left23:14
b14ck Hi all. I'm having issues creating a tag for my project. I'm running: git tag -a -v 1.0 But it won't accept that and keeps displaying usage. Anyone know what I'm doing wrong? It seems like I'm doing it right according to the man page.23:14
← brizly left23:15
→ alezandro joined23:15
← pklingem1 left23:16
rudi_s b14ck: Don't use -v, it's used to verify the tag later AFAIK.23:16
impulze is there a config option for find-copies-harder?23:18
← psoo left23:20
→ nonken joined23:21
← ntoll left23:21
← nonken left23:21
→ nonken joined23:22
← hyperair left23:23
→ pklingem1 joined23:23
← pklingem1 left23:24
→ ProperNoun joined23:25
← alfredlysebraate left23:25
← mbroeker left23:26
→ bcardarella joined23:27
→ nik0 joined23:29
← bremner left23:29
Thell → Thell|Away23:30
→ bremner joined23:30
niko → Guest7746223:32
nik0 → niko23:33
→ Gentoochild joined23:33
→ justin-george joined23:34
← Guest77462 left23:35
→ jspiros joined23:35
→ cloudhead joined23:36
← mmw left23:39
← ProperNoun left23:40
← malumalu left23:40
Gentoochild hi again. :-) Please help me not dying dump in understanding the following most basic question. *g*23:42
← b14ck left23:42
Gentoochild consider the following: http://141.24.172.20/pub/develop.txt23:42
I wanna contribute to a big project by developing a new feature23:43
so I create a new branch.23:43
→ nitrate joined23:43
← jspiros left23:43
← robert_ left23:43
Gentoochild I need a while to develop that feature, and in parallel I'm pulling the latest commits from the public repo to stay up to speed.23:44
→ jjuran joined23:44
← jjuran left23:44
Gentoochild what is the correct way now to incorporate the latest master commits into my branch without messing up the branch tree?23:45
→ ProperNoun joined23:45
→ jspiros joined23:46
→ pantsman joined23:46
locklace Gentoochild: git pull --rebase23:46
← tvw left23:48
→ d0ugal joined23:49
← bdimcheff left23:53
→ pheaver joined23:55
← NET||abuse left23:56
→ charlenopires joined23:56
→ Gentoo_child joined23:57
← Raydiation left23:57
← Gentoochild left23:57
← nitrate left23:58

Logs Search ←Prev date Next date→ Channels Documentation