IRCloggy #git 2010-06-29

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.

2010-06-29

ReinH Hi folks, where does git grab the executables in .git/hooks to run?00:00
hyperair joined00:00
ReinH I'm wondering if it uses find, and if so how it deals with the differences between GNU find and BSD find00:00
ntoll_ joined00:00
ntoll left00:00
ntoll_ntoll00:00
ShooterMG left00:00
ReinH which are currently making me sad in a non-git-related script that needs to be portable00:01
anjen joined00:02
ReinH oh, it just does a text -x for each hook individually00:03
Basstard` joined00:03
Paraselene_ left00:03
geckosenator left00:03
zx2c4 left00:05
[Po]lentino left00:05
ignatp left00:07
ignatp joined00:07
neunon joined00:07
icwiener joined00:07
nevyn_nevyn00:08
Paraselene_ joined00:09
geckosenator joined00:10
EricInBNE joined00:14
jrmuizel left00:14
jrmuizel joined00:15
hyperair left00:16
Titosemi left00:16
dreiss left00:19
dj_tjerk left00:20
geckosenator left00:20
brothers_ joined00:20
geckosenator joined00:22
harinath left00:22
gsan1 joined00:24
airborn left00:26
lispy left00:27
ExtraSpice left00:27
justin-george left00:27
rado1 left00:28
ntoll left00:28
rdeweerd joined00:29
Paraselene_ left00:30
Chaser__ joined00:30
Chaser__ left00:32
kenichi left00:32
Chaser left00:33
gnuvince left00:35
harinath joined00:35
eletuchy_ joined00:36
eletuchy left00:39
engla|ger left00:40
Goneri left00:41
dramsay left00:41
rado1 joined00:43
Goneri joined00:45
arthurmaciel see you!00:45
arthurmaciel left00:45
andrewl left00:47
jspiros left00:47
rryouumaa joined00:47
jtbandes joined00:48
geckosenator left00:48
jtbandes Any opinons on Mac clients: GitX, Gity, something else? (have we reached GitZ now?)00:48
MicahElliott joined00:51
jspiros joined00:52
geckosenator joined00:53
Heimidal joined00:53
bc0906524 joined00:53
hakunin left00:53
rdeweerd left00:54
Ambiguity left00:54
jtaby left00:55
Rodya left00:56
MicahElliott left00:57
andrewl joined00:59
sftp left01:01
yaotti_awyaotti01:12
Vampire0_ joined01:14
yaottiyaotti_aw01:14
yaotti_awyaotti01:15
gsan1 left01:15
dsdeiz joined01:16
dsdeiz heya, how do i check if there are new files that were not added to the repo? i did git status although it didn't show the new files :D01:17
berserker joined01:17
davr git status will show new files not added to the repo, unless they are ignored by .gitignore or .git/info/exclude01:17
kpreid_ joined01:18
dsdeiz even images right?01:18
oh, nvm, i get it now.. tnx!01:18
davr file type doesn't matter to git01:18
Ahri left01:18
Vampire0 left01:18
dsdeiz gotcha, thanks01:19
Ahri joined01:19
hobodave left01:20
dsdeiz left01:21
alezandro joined01:21
geckosenator left01:23
jtibu joined01:24
parasti left01:24
MetaCosm joined01:25
naderman_ joined01:26
offby1 jtbandes: some coworkers use GitX and seem to like it. It has a few rough edges though01:27
Joel_ joined01:27
_VVSiz_ joined01:28
tstclair left01:28
jtbandes left01:28
chrislerum joined01:29
mabrand left01:29
naderman_naderman01:30
naderman left01:30
naderman joined01:30
jtibu left01:30
VVSiz_ left01:31
pennyfx joined01:36
geckosenator joined01:37
yaottiyaotti_aw01:37
yaotti_awyaotti01:37
Ambiguity joined01:38
tychoish left01:39
alester left01:41
przemoc left01:43
kefah joined01:44
tychoish joined01:44
kefah hello, is this the right place to ask about gitolite?01:45
bremner yes01:47
well, there is a mailing list also.01:48
hyperair joined01:49
kefah bremner: thanks. I thought i could get some quick help: here is my problem: I setup gitolite, admin stuff seems is working properly. I was able to clone, modify both conf and keydir to add a new repo and new user.01:49
bremner: the new user is another account. all are on mylaptop.01:50
bremner: I can check the newuser via ssh [email@hidden.address] info and it shows that it has the right access to the wanted repo.01:51
pedrobelo joined01:51
kefah bremner: i clone that repo (still empty) successfully, but I fail to push any new files.01:51
rdeweerd joined01:51
bremner what error message do you get?01:52
kefah bremner: I read the ssh-troubleshooting many times, but couldn't figure it out.01:52
spearce left01:52
kefah bremner: $ git push01:52
No refs in common and none specified; doing nothing.01:52
Perhaps you should specify a branch such as 'master'.01:52
fatal: The remote end hung up unexpectedly01:52
error: failed to push some refs to 'gitolite@127.0.0.1:kee.git'01:52
bremner kefah: it sounds like you should follow the suggestion in the error message. This is a typical git message, not a gitolite problem afaikt01:53
sitaram bremner++01:53
kefah: man git-push will tell you what the default behaviour is when you "git push" instead of "git push remotename branchname"01:54
kefah: read that, and then the error message you got will make sense01:54
kefah bremner: thanks. I'll read that right a way.01:54
rryouumaa git log produces diffs. so what does git diff do?01:56
ereslibre_laptop left01:56
rryouumaa do most people use only log?01:56
geckosenator they have different functions though01:57
hyperair left01:58
hobodave joined01:58
fivetwentysix joined01:59
pedrobelo left02:01
rdeweerd left02:02
kefah sitaram and bremner: it worked with me when I typed "git push origin master". but i encountered something confusing:02:02
tildeequals left02:02
kefah when I did the same steps on the first user (the admin of gitolite, not the gitolite user) but using only "git push" it worked without a problem.02:03
also, when I repeated modifying something on the second user (normal user) it didn't require me to mention origin master.02:04
antlong joined02:04
yaottiyaotti_aw02:04
f00li5h does it make sense for me to ask for an "external" (like those in svn) with git?02:05
yaotti_awyaotti02:05
antlong left02:05
f00li5h so I can have stuff common over a couple of repositories... or are they just more remotes?02:05
kefah I know that I might sound too noob02:06
pragma_ barks verociously at f00li5h.02:07
pragma_ idly wonders if verociously is indeed a word, after noticing he meant to use an 'f'.02:07
[1]intripoon joined02:08
yaottiyaotti_aw02:08
przemoc joined02:08
yaotti_awyaotti02:08
rryouumaa ferociously voracious02:09
of all the zillions of git docs out there, which would explain at an executive level the difference between git diff and git log with examples?02:10
tstclair joined02:10
bremner kefah: part of the answer is in that first error message: after you push master, the two repos have a common ref, namely master.02:11
consolers joined02:11
bremner rryouumaa: tell me what you want to do, and I'll tell you which one to use.02:11
consolers use the red one02:12
intripoon left02:12
[1]intripoonintripoon02:12
rryouumaa well, i was trying to self-teach a man to fish, but the triggering problem, which i already kludged manually with magit, was diffing from a certain date to another date a subset of files. i'm guessing git log -p in that case?02:12
przemoc left02:13
bremner do you want to know the diff in terms of commits, or just the total effect?02:13
rryouumaa i think the total effect02:13
Heimidal left02:13
bremner then git diff might be better.02:14
rryouumaa ok so diff is total effect and log is each?02:14
kefah bremner: so more or less, by typing "git push orign master" for one time, it fixed the common ref, so there is no need to do that again every time. I think I got it. Many thanks for the help.02:14
bremner rryouumaa: well, log is a real monster with many options, but basically yes.02:15
sh1m left02:15
icwiener left02:15
rryouumaa yeah i kind of got lost with it02:15
thus the idea of examples02:15
hence*02:15
bremner rryouumaa: log is about examining history; diff is about comparing two commits.02:15
harinath left02:15
rryouumaa ok02:16
bremner There are several examples in the man page02:16
rryouumaa but diff does not have the since and until options02:16
harinath joined02:16
Alagar left02:16
rryouumaa in my version anyway02:17
hyperair joined02:18
bremner rryouumaa: it's a bit tricky to find, but see the man page of git-rev-parse, look for " A ref followed by the suffix @ with a date specification"02:19
kefah bremner: I just read that this problem might be specifically related to empty repo (like was my case).02:19
przemoc joined02:20
bremner kefah: well, it can arise in other cases too, but not in the simplest case of cloning a repo and pushing back the default branch.02:20
cytrinox_ joined02:21
kefah bremner: it seems I have to do more reading on the subject. appreciating your help. bye.02:21
bremner bye02:21
bremner sneaks off before getting sucked in more :)02:21
rryouumaa bremner: wait, so git diff takes --since and --until anyway? or i do master{yesterday}?02:22
adymitruk joined02:22
kefah left02:22
bremner somethink like master@{yesterday}02:22
hyperair left02:22
Elfe_ joined02:24
rryouumaa ok, thanks. what about the since thing (not that it is necessary given the existence of the @ thing)?02:24
Elfe left02:24
rryouumaa what an intricate system of commands. how do you guys remember all that stuff from 3 months back when you last needed it?02:24
yaotti left02:24
cytrinox left02:24
cytrinox_cytrinox02:24
rryouumaa i must say, git is powerful02:25
parndt word02:28
tstclair left02:31
hyperair joined02:33
flaguy48 left02:33
andrewl left02:36
andrewl joined02:36
zenazn_ left02:40
frogonwheels rryouumaa: knowing which man page to look at is 1/2 the battle02:41
LiamH left02:41
tildeequals joined02:41
rryouumaa yeah true, a little overwhelming sometimes though02:45
harinath left02:47
parndt understanding the man page is the second 1/202:47
timj__ joined02:52
eletuchy_ left02:54
chrislerum left02:55
zenazn joined02:55
timj_ left02:56
coderdad joined03:00
mhagger joined03:00
consolers git man pages are nothing compared to say rsync03:02
its a clarion of perspecuity03:02
perspicuity03:02
harinath joined03:03
warlock_mza joined03:08
saiyr9000 left03:09
saiyr9000 joined03:09
Joeconyers left03:10
justin-george joined03:10
pennyfx left03:11
saiyr joined03:12
warlock_mza left03:12
eletuchy joined03:12
hyperair left03:14
[J]stringo0 joined03:14
[J]stringo0 left03:14
saiyr9000 left03:15
stringo0 left03:16
zenazn left03:16
Joeconyers joined03:17
mhagger I'm confused about the --mirror options on the "git remote add", "git clone", and "git push" commands and the remote.<name>.mirror config option...03:18
(1) Does remote.<name>.mirror only affect pushing, or also fetching?03:18
(2) Does "git clone --mirror" set this config option?03:19
rado1 left03:20
mhagger Empirically, "git clone --mirror" *does* set this config option03:21
PhilipYartoo joined03:21
qingshan joined03:22
mhagger I guess my main concern is: if remote.<name>.mirror affects both push and fetch, then it seems that one is a single command away from disaster if one accidentally pulls from a "backup" repository.03:22
qingshan left03:23
David joined03:23
DavidGuest6191103:23
PhilipYartoo Hello. Anyone to help me on this issue? For example: Tom worked on branch t2444, then merge it into integration, then the integration branch goes on03:23
while later #2444 is bounced, so Tom has to work on #2444 again, while when he tries to merge integration into t2444, there are a lot of conflicts because there are a lot of changes after his last commit for t244403:23
Some conflicts are even not real conflicts like new codes, but git thinks they are conflicts, and the most difficult part is some configure files, what git displays is not whole configure files, but some codes splits which makes Tom thinks to deal with the merge conflicts is time consuming and not easy.03:24
BTW, integration is another branch.03:24
adamv joined03:25
PhilipYartoo I'm not sure whether Tom is using feature branching right or not.03:26
Guest61911DavidChen03:26
ralf_e left03:26
whaley left03:27
bousquet joined03:30
saiyr left03:31
PhilipYartoo My questions are 1.How to deal with these git merge conflicts? 2. Whether we are using git's feature branch on the right way? 3. If not, how?03:31
saiyr9000 joined03:31
avar how are they not "real conflicts" ?03:33
PhilipYartoo For example, in integration branch, other developers added some new functionalities03:35
which may not be relative to the codes in t244403:35
Joel_ left03:36
avar yes, new functionality that obviously conflicted with the topic branch since you're getting a merge conflict.03:40
keyvan left03:41
hohoho joined03:43
Zycon joined03:46
AAA_awright left03:47
neunon_ joined03:47
mw left03:47
PhilipYartoo How do you deal with the shared codes?03:49
neunon left03:49
neunon_neunon03:49
Goneri left03:51
Goneri joined03:52
johnw left03:53
eletuchy left03:53
rryouumaa left03:53
sftp joined03:54
AAA_awright joined03:54
Zycon ok.. so more specifically how do you deal with code that is shared between individual tickets (i.e. jobs) if you are feature branching... i.e. ticket 1 uses library C and ticket 2 uses library C and both tickets require minor updates to library C03:55
mhagger left03:55
hasrb_ joined03:55
Zycon as far as I know there's very little you can do about this besides just keep the tickets in separate feature branches still and then just merge the conflicts in C after03:55
anyone else have any good ideas about what to do in this situation?03:56
Bass10 left03:56
maleknet joined03:56
Heimida__ joined03:57
AAA_awright left03:58
dreiss joined03:58
hasrb left03:58
jfkw left03:58
dr_win left03:58
sftp left03:59
welly joined04:00
AAA_awright joined04:01
leitao left04:03
alezandro left04:04
Pewpewarrows left04:05
hakunin joined04:05
DavidChen left04:05
anathematic joined04:06
PhilipYartoo left04:06
AAA_awright left04:08
eletuchy joined04:08
AAA_awright joined04:08
codeshepherd joined04:10
eletuchy left04:10
hasrb_ left04:12
hasrb joined04:12
jmcantrell left04:14
aresnick left04:14
leitao joined04:15
harinath left04:15
avar Explaining the problem better would be a good start. Like what sort of merge conflicts are these.04:15
Zycon avar: thanks.. i'll get some more details04:20
ignatp left04:22
brosner left04:22
scarabx left04:25
leitao left04:27
radioman-lt joined04:28
alezandro joined04:28
harinath joined04:30
fishey left04:30
consolers left04:31
ignatp joined04:32
marc_kdab left04:32
justin-george left04:35
rchavik joined04:36
Joeconyers left04:37
radioman-lt left04:40
jony joined04:40
MattDiPasquale joined04:40
MicahElliott joined04:42
brosner joined04:44
brosner left04:44
brosner joined04:44
brosner left04:44
emmanuel_oga left04:44
apsysoev left04:45
AAA_awright left04:47
anjen left04:48
coderdad left04:48
adamv left04:48
ijuma_ left04:50
rado1 joined04:50
hasrb left04:51
ijuma_ joined04:52
intripoon left04:56
AAA_awright joined04:56
dytrivedi joined04:57
harinath left04:58
vu3rdd joined04:59
girishr joined04:59
alezandro left05:00
brosner joined05:01
brosner left05:01
psankar joined05:01
jonmasters left05:01
brosner joined05:02
brosner left05:02
baegle joined05:02
baegle Ok, a bit lost reading the manuals for this particular use case. I have files that are going to be added on my next commit that I don't want added yet. I want to keep the local changes, but I don't want them to be committed when I do a git-commit. How do I remove them from the staged commit but keep the local modifications intact?05:03
freimatz joined05:04
FauxFaux You mean, you want to un-stage a file? That's a reasonably frequently asked question.05:05
baegle I was using the verb "remove", didn't think to search "unstage"05:05
AAA_awright git reset I think?05:06
baegle git-reset, but the manual entry strikes fear into me that it will remove my changes if I don't specify the right options05:06
AAA_awright left05:09
AAA_awright joined05:10
brosner joined05:11
brosner left05:11
AAA_awright left05:11
brosner joined05:11
brosner left05:11
AAA_awright joined05:11
harinath joined05:11
Yuffster left05:12
jonmasters joined05:14
brosner joined05:14
brosner left05:14
hagabaka joined05:15
brosner joined05:16
brosner left05:16
ignatp left05:17
walken joined05:18
walken hi...05:18
LongBeach joined05:18
priidu joined05:18
walken I am wondering, is there a git command that will set a non-branch ref to a known sha1 ?05:18
I thought git branch -f would do it, but it'll only work on refs/heads/* refs05:18
'git push . +sha1:ref' does it, but there's got to be a better way ???05:19
ignatp joined05:19
walken (I'm trying to set a ref in refs/notes/* space)05:19
p3rg2z joined05:20
blueshift joined05:22
brosner joined05:22
brosner left05:22
Heimida__ left05:23
MicahElliott left05:24
doener joined05:24
aliceinwire joined05:24
marc_kdab joined05:24
nairb774 left05:24
solofight joined05:25
fmardini joined05:25
solofight help | i had commited my changes in a no branch without knowing it, and when somebody noticed it they merged the no branch to master, and found that the commits were lost05:26
is this possible ?05:26
ExtraSpice joined05:26
solofight if yes, is there a way i can retrive my commits without hav to recode everything ?05:26
MattDiPasquale left05:26
blueshift solofight: can you find them in the reflog?05:27
alezandro joined05:28
nairb774 joined05:28
solofight blueshift, wait let me read about reflog and reply to your question05:28
fivetwentysix left05:28
fivetwentysix joined05:29
Belna joined05:29
rayners joined05:31
solofight blueshift, but is it possible that the commits made in no branch when merged to master get disappeared ?05:31
dgarr left05:31
blueshift solofight: don't know how that would work, sorry05:31
but if you do "git reflog" you may be able to find your commits and check them out or get their patches05:32
jmcantrell joined05:32
blueshift I don't *think* they should disappear unless they were deliberately dropped during the merge.. but don't know enough about that stuff, sorry05:33
p3rg2z left05:33
solofight blueshift, ok05:34
will try that05:34
and get back here if anything pops up05:34
thank you05:34
blueshift np, good luck05:34
dgrift left05:35
eighty4 joined05:35
apsysoev joined05:36
parndt left05:36
parndt joined05:37
brizly left05:38
Mage left05:39
AAA_awright left05:41
AAA_awright joined05:41
AAA_awright left05:42
dgrift joined05:43
roop joined05:43
Aaaarg_ joined05:43
kar8nga joined05:43
RobertLaptop left05:44
geckosenator left05:44
geckosenator joined05:45
Aaaarg left05:46
Aaaarg_Aaaarg05:46
AAA_awright joined05:46
StrongOrder left05:47
jmcantrell left05:48
mlangenberg joined05:49
eighty4 left05:50
eighty4 joined05:51
Heimidal joined05:51
PhilipYartoo joined05:51
Bennid joined05:55
dgrift left05:55
khmarbaise joined05:55
dgrift joined05:56
GodEater joined05:57
GodEater left05:57
GodEater joined05:57
programble left05:58
welly left05:58
MrGando joined05:58
codejunky left05:59
codejunky joined05:59
MrGando Hello! , is there a way to perform a "git fetch" while being outside a repository folder ??? ( this is to automate fetching with a script on my build server )05:59
baegle git clone?05:59
parndt chdir to the directory and run the command and then chdir back ?05:59
eighty4 left05:59
blueshift MrGando: git --git-dir=/where/it/is fetch06:00
or do what I did: "man git"06:01
MrGando Sorry , I did git --help , stupid me didin't think about man git . Thanks for the help :)06:02
radioman-lt joined06:02
MrGando blueshift: wow, this is interesting , I'm getting : fatal: Not a git repository: '/Path/to/repo' ... but if I do cd /Path/to/repo and perform the fetch, it goes fine06:04
Aaaarg_ joined06:04
mlangenberg left06:05
Morasique left06:06
blueshift MrGando: seems to want /.git on the end06:06
MrGando let's see06:06
nekoget joined06:06
MrGando You are correct...06:06
weird though06:06
blueshift hmm but things like git log don't work right, then06:06
Aaaarg left06:08
Aaaarg_Aaaarg06:08
blueshift hm it works if you specify both --work-tree and --git-dir06:08
--git-dir needing /.git on the end06:08
radioman-lt left06:08
blueshift anyway. you can use man as well as me.06:09
I should do work.06:09
aliceinwire left06:09
poseid joined06:10
Morasique joined06:10
thiago_home joined06:11
ige joined06:12
[Po]lentino joined06:14
tedoc2000 left06:15
cbreak_work joined06:15
bambule left06:17
khmarbaise left06:19
qingshan joined06:20
ignatp left06:21
level09_ joined06:21
jpic left06:21
timyhac joined06:21
SplinterOfChaos joined06:21
Titosemi joined06:21
dreiss left06:21
timyhac left06:21
ph^ joined06:22
eighty4 joined06:22
timyhac joined06:22
sagsousuke left06:24
sgh joined06:24
berserker left06:26
solofight left06:26
rchavik left06:26
eletuchy joined06:27
berserker joined06:28
geckosenator left06:29
roop left06:31
SplinterOfChaos left06:32
jensn joined06:33
jensn left06:34
MrGando left06:34
MrGando joined06:34
rox` joined06:36
hyperair joined06:37
thunksmoking|thunk06:37
dreiss joined06:38
roop joined06:39
thiago_home left06:40
qingshan left06:40
Jaxan joined06:41
dytrivedi left06:42
BlindOracle left06:42
avar left06:43
avar joined06:45
mtkd joined06:45
BlindOracle joined06:47
ExtraSpice left06:48
dwmw2_gonedwmw206:49
hyperair left06:49
smoking|thunkthunk06:49
Tass_ joined06:50
Tass_ is there a way to stage specific changes in a file only?06:51
hyperair joined06:52
project2501a left06:53
_rane Tass_: git add --patch06:53
shame you haven't known about it until now06:53
walken I am wondering, is there a git command that will set a non-branch ref to a known sha1 ?06:54
I thought git branch -f would do it, but it'll only work on refs/heads/* refs06:54
'git push . +sha1:ref' does it, but there's got to be a better way ???06:54
(I'm trying to set a ref in refs/notes/* space)06:54
Tass_ _rane, duh!06:54
Jaxan left06:56
hyperair left06:57
Tass_ how to go git add -p for all unstaged files?06:59
_rane `git add -p`06:59
ede joined06:59
MrGando left06:59
Tass_ ohhh06:59
taptapir joined07:00
jony left07:00
taptapir left07:00
oleide joined07:01
jreiners joined07:01
lhz joined07:03
masterkorp joined07:04
jony joined07:04
scatterp_ joined07:05
scatterp left07:05
jrmuizel left07:06
oleide git <307:09
disappearedng left07:10
rado1 left07:11
trivol joined07:14
skoop joined07:15
brundle joined07:15
Lgb left07:16
Gullik left07:16
giskard joined07:23
disappearedng joined07:24
timyhac left07:24
barismetin joined07:25
kar8nga left07:25
jony left07:25
pantsman joined07:26
Jochen_Hayek joined07:26
poseid left07:27
Gullik joined07:28
Lgb joined07:28
raviepic3 joined07:29
raviepic3 blueshift, hurray07:29
the commits are there when i see the git reflog07:29
so now what do i do ?07:30
Aaaarg left07:30
raviepic3 i have to take that commit and put it in main branch07:30
and make the effective along with other commits made by others after this one07:30
tatsuya joined07:33
lucsky joined07:33
NetHawk left07:33
berserker left07:37
sondreb joined07:38
timyhac joined07:38
Leemp left07:39
jony joined07:40
wereHamster raviepic3: just one commit? Or several consecutive ones?07:40
tatsuya left07:40
hakunin left07:40
tatsuya joined07:40
tatsuya left07:41
berserker joined07:41
airborn joined07:41
tatsuya joined07:42
nevyn^_ joined07:43
GodEater left07:43
tatsuya left07:44
raviepic3 wereHamster, i think 2 commits07:44
yeha07:44
2 commits07:44
Zycon left07:44
tatsuya joined07:44
raviepic3 and after that it displays a merge commit07:44
merge master07:44
so two commits are my custom code07:44
wereHamster, how do i get these commits and apply to master ?07:45
Louie` joined07:45
Louie` left07:45
GodEater joined07:46
GodEater left07:46
GodEater joined07:46
wereHamster raviepic3: cherry-pick07:46
krawek left07:47
raviepic3 wereHamster, googling07:48
will get back to you after using it07:48
i mean here07:48
Zycon joined07:49
bousquet left07:49
stamina_ left07:50
alphabeat joined07:50
stamina_ joined07:50
dahankzter joined07:50
Aaaarg joined07:50
drizzd left07:51
alphabeat hey all. i stupidly pushed a rebased branch to a remote bare repo, that only i was using, and then did a pull, resulting in a bunch of dupes. what's the best way to revert?07:52
drizzd joined07:52
[Po]lentino left07:52
raviepic3 wereHamster, i need to do a git cherry-pick commitid right ?07:53
but how do i get the sequence of commits ?07:53
should i be doing like commitid1..commitid_n ?07:53
level09_ left07:53
timyhac left07:54
j0rd_ left07:54
[Po]lentino joined07:55
giskard left07:55
NetHawk joined07:55
raviepic3 wereHamster, when i do a git cherry-pick commitid107:56
it says 1 commit cherry picked07:56
but when i do a git log07:56
the commit is not there07:56
dreiss left07:57
masterkorp left07:57
psoo joined07:57
Leemp joined07:58
ziro` joined07:58
waveclaw joined07:58
giallu joined07:58
Soliton left08:00
girishr left08:02
shruggar joined08:02
Soliton joined08:02
wereHamster raviepic3: you said it's only two commits, so do git cherry-pick <commit1>; git cherry-pick <commit2>08:03
raviepic3 wereHamster, when i do a git cherry-pick commit108:03
it says 1 commit cherry picked08:03
but when i see git log08:03
the commit which has been cherry picked is not there08:04
wereHamster, ^08:04
pantsman left08:04
waveclaw left08:04
wereHamster was the cherry-pick successful? No conflicts?08:05
raviepic3 yeah08:05
no conflicts08:05
ntoll joined08:05
wereHamster and are you using 'git log <bracnh>' or jsut 'git log'?08:06
raviepic3 git log08:06
wereHamster maybe the commit already is part of your branch history08:07
is 'git log ..<commit>' empty?08:07
raviepic3 if thats the case it should have shown the commit message in the git log right08:07
it didnt08:07
1 min08:07
ChUrP joined08:08
raviepic3 wereHamster, cool08:08
it worked08:08
when i do a git log08:09
the cherry picked commit showed now08:09
radioman-lt joined08:09
raviepic3 but how do i see the changes made in the commit ?08:09
sftp joined08:09
raviepic3 i did a git diff d53019617a0e594f701c9c9d4e8efcaaa4d62deb08:09
and nothing appeared08:09
wereHamster git show <commit>08:09
sgronblo_sgronblo08:11
[Po]lentino left08:11
masterkorp joined08:11
[Po]lentino joined08:13
emmanuel_oga joined08:14
hyperair joined08:15
girishr joined08:16
raviepic3 wereHamster, thank you very much08:16
wereHamster++08:16
raviepic3 left08:16
ExtraSpice joined08:17
jatt joined08:17
_rane wereHamster--; hah-ha!08:19
just kidding, you're the best!08:20
baegle left08:20
nevyn left08:20
nevyn joined08:22
ntoll left08:24
hyperair left08:24
ntoll joined08:25
timyhac joined08:26
Paraselene_ joined08:28
berserker left08:29
bauruine left08:29
hyperair joined08:30
berserker joined08:31
anathematic left08:31
walken left08:31
dpino joined08:32
project2501a joined08:34
warthog9 left08:36
gmarselis__ joined08:36
AlexC_ joined08:38
AlexC_ morning08:38
tvw joined08:38
AlexC_ if I commit + push, and people then pull in my changes. What happens if I remove this commit and force a push? When others pull in my changes will that commit be removed for others to?08:38
brundle left08:38
parndt yes but only if you did it before anyone pulled08:39
AlexC_ so you mean no :P08:39
parndt ;) sure do08:39
AlexC_ ok, thanks08:40
parndt it's generally considered _very_ bad practice to remove commits08:40
you can revert a commit instead08:40
AlexC_ indeed, I know - just commited on wrong branch and only noticed today08:40
project2501a left08:40
parndt oops ;)08:40
project2501a joined08:40
AlexC_ yeah =\08:40
parndt done that before08:40
roop left08:41
gmarselis__ left08:41
AlexC_ mmm, I did a 'git diff 2.0.4 2.0.5' this morning and was expecting far more changes than I saw08:41
rayners left08:43
freshtonic left08:46
cilly joined08:46
Zycon left08:47
MrSparkle joined08:47
poseid joined08:47
MrSparkle left08:47
thunkthunk-asleep08:48
sunil joined08:48
TheNewAndy joined08:49
gmarselis__ joined08:49
_Vi joined08:50
harinath left08:50
bluenovember joined08:50
emmanueloga joined08:50
lucasvo_2 joined08:51
lucasvo_2 I have cloned a git repository locally and now push to the local "remote" regularly. Now when `cd` to the remote, I see uncommitted files but a `git diff` doesn't return anything. when I then do git pull from the remote's remote, the last commit that I pushed to the local remote is being discarded. what am I doing wrong?08:51
mbroeker joined08:52
project2501a left08:53
emmanuel_oga left08:53
ptor lucasvo_2: Maybe your 'remote' repo isn't a 'bare' repo? You should only really push to bare repos.08:54
Titosemi_ joined08:56
Titosemi left08:57
Titosemi_Titosemi08:57
lucasvo_2 ptor: ok, but how can I push the remote repo to my real remote (on github)?09:00
erikvold left09:01
erikvold joined09:03
dj_tjerk joined09:04
rajeshsr_ joined09:04
froschi joined09:05
harinath joined09:05
werdan7 left09:06
gmarselis__ left09:08
curtana joined09:08
ptor github remotes are bare repos. No push problems. If you want to sync local closes between them it's better to git pull between them. And push from anyone to github.09:10
s/closes/clones/09:10
erikvold left09:11
jamesmacleod joined09:12
jamesmacleod Hey guys, this is a really daft question, but I have a remote repository that I use with capistrano, but when I look at the folder it appears empty. what is the command to get it up-to-date with the latest commits from my local machine?09:13
curtana left09:14
wereHamster jamesmacleod: git push09:14
werdan7 joined09:14
jamesmacleod sorry I don't think I am explaining this very well09:15
ralf_e joined09:15
jamesmacleod I have pushed all the changes up to my remote repository, but when I ssh to the server and look at the server the directory appears empty apart from the git folders. how do I get that to show the files I have just pushed09:15
wereHamster jamesmacleod: faq non-bare09:16
jamesmacleod: https://git.wiki.kernel.org/index.php/GitFaq#non-bare09:16
wereHamster is wondering when the gitbot comes back09:16
cilly left09:16
wereHamster Pieter: ^^^09:17
curtana joined09:17
masterkorp left09:17
jamesmacleod brilliant thanks, I knew the logic behind what was going on, but I just didn't know the command. Thanks for your help09:17
cilly joined09:17
eletuchy left09:18
oriba joined09:19
babafds joined09:20
ziro` left09:20
d0k joined09:21
bababfds left09:21
curtana left09:22
curtana joined09:22
sebi_` joined09:24
hyperair left09:24
roop joined09:24
[Po]lentino left09:24
henkboom left09:28
gebi joined09:29
henkboom joined09:31
masterkorp joined09:31
ijuma joined09:31
harinath left09:31
rajeshsr_ what does a plain: git checkout when am in some branch?09:32
malumalu joined09:33
Bombe rajeshsr_, depends on the parameters.09:33
jds joined09:33
_rane nothing09:33
hukl joined09:33
_rane or does it?09:33
parasti joined09:33
jony left09:33
rajeshsr_ Bombe, no params!09:33
well, waht happened is i have made 2 set of changes now to my branch. i want to discard 1 change and commit another..09:34
Bombe Well, I just did a “git checkout” in a working tree, that is something that everybody can easily do for himself.09:34
And it appears to have done nothing, indeed.09:34
rajeshsr_ what can be done in this case?09:35
_rane have you staged either change?09:35
git add --patch; do 'y' for the hunk that you want to commit09:35
rajeshsr_ nope, they are pending changes..09:35
Bombe rajeshsr_, I would use “git gui” or “git add -p” to add only the parts you want to commit.09:35
berserker left09:35
_rane then git reset --hard when you only have changes left you want to get rid of09:35
rajeshsr_ how do you discard changes? like i created new files and all..09:36
oh, thanks!09:36
_rane git clean -f removes untracked files from teh repo09:36
PhilipYartoo left09:36
_rane or working tree09:36
berserker joined09:37
rajeshsr_ BTW, now i face a new problem. I had overridden a file with some content, but i actually wanna append it? If I revert i lose its contents, right? so what can be done?09:38
just cp somewhere and revert and append it?09:38
_rane that would be easiest, yes09:38
git checkout -- path/to/file to cancel changes in a file09:39
or git checkout HEAD -- path/to/file09:39
hyperair joined09:40
vu3rdd left09:40
project2501a joined09:42
Nebukadneza left09:42
rajeshsr_ _rane, ok, thanks!09:43
villemv joined09:43
rajeshsr_ is it possible to get a single file from a particular commit?09:44
teuf git show sha1:filename09:44
_rane I think he wants checkout09:44
teuf or git checkout sha1 filename09:44
_rane: yep, thought about it afterwards: )09:44
HG` joined09:45
harinath joined09:45
rajeshsr_ with a different name? that will override the present contents, right?09:46
_rane right09:46
but you could do git show sha1:filename > new_file09:47
hyperair left09:47
wereHamster show > file and checkout -- file is not necessarily equal09:48
Mike_MnD joined09:50
teuf wereHamster: ah?09:51
anathematic joined09:51
bentob0x joined09:52
wereHamster I'm not 100% certain, but IIRC checkout is running the filters on the file and show isn't09:52
jony joined09:53
wereHamster so show is more like cat-file -p09:53
oriba left09:53
Unode|NAUnode09:54
tomoyuki28jp joined09:55
Mike_MnD why "git help clone" says "No manual entry for git-clone" ??09:56
tomoyuki28jp Is there a way to create .gitignore which is valid only on my local machine? (My repository already has committed .gitignore file.)09:57
wereHamster Mike_MnD: maybe because you didn't install the git man pages.. ?09:57
curtana tomoyuki28jp, .git/info/exclude09:57
parndt tomoyuki28jp: http://abizern.org/2010/02/07/three-ways-of-excluding-files-from-git/09:57
tomoyuki28jp curtana: parndt: thanks a lot!09:58
Mike_MnD wereHamster: well that's is logical :P bu problem is this is on my webhosting and they installed it09:58
fivetwentysix left09:58
Mike_MnD well i'll have to live with it :D09:58
wereHamster tomoyuki28jp: or if you want a gitignore for all your lcoal repos, there's a way to do that throught core.excludesfile09:58
tomoyuki28jp wereHamster: I see, thanks!09:58
tizbac joined09:59
rajeshsr_ _rane, git show does what i wanted to do well! I like it better than reverting and putting it back. Although both seems to require almost same efforts.09:59
Thanks09:59
fivetwentysix joined10:00
rgr joined10:00
gebi left10:01
bauruine joined10:02
kar8nga joined10:02
reuss_reuss10:02
felipec joined10:03
psankarpsankar_brb10:04
henkboom left10:04
tomoyuki28jp curtana: I have modified committed files and want to ignore the changes. .git/info/exclude is not the one for me? (seems like doesn't work in my case)10:05
parndtparndt_away10:06
girishr left10:06
Ilari tomoyuki28jp: Tracked files can't be ignored.10:06
tomoyuki28jp Ilari: uhm. I see..10:06
rajeshsr_ git commit -a doesn't work with newly created files?10:06
cilly left10:06
tomoyuki28jp Ilari: thanks!10:06
Ilari rajeshsr_: Yup.10:06
henkboom joined10:07
rajeshsr_ Ilari, so, it is mainly for modification and deletion of file?10:07
mainly => only10:07
tomoyuki28jp In my project, some files are to set up dev environment so I need to change it for my machine, but don't wanna commit it.10:07
Ilari tomoyuki28jp: Commit template of that file and leave the real file untracked.10:08
cilly joined10:08
curtana maybe you could do something clever with a munging/unmunging stcript10:08
jony left10:08
Ilari rajeshsr_: It only picks modifications and deletions. But I don't even usually use commit -a (only when everything is tracked).10:08
tomoyuki28jp Ilari: I understand. thanks a lot for your help.10:08
jensn joined10:08
tomoyuki28jp left10:09
kar8nga left10:09
hyperair joined10:11
froschi left10:11
pragma_pragma|t10:12
poseid left10:12
pragma|tpragma_10:14
pragma_pragma|t10:14
keyvan joined10:17
aziz joined10:17
Chaser joined10:18
pragma|tpragma_10:18
pragma_pragma|t10:19
fod left10:20
Jochen_Hayek left10:20
pragma|tpragma_10:21
bentob0x left10:22
bentob0x joined10:22
jony joined10:22
dj_tjerk left10:23
skoop left10:26
lpolzer joined10:26
Jochen_Hayek joined10:27
lpolzer hey there10:27
lucasvo_2 left10:27
lpolzer quick question: how do I push a local branch to a specific remote branch?10:27
I tried "git push origin REMOTE-BRANCHNAME" but it didn't work...10:27
curtana git push remote localbranch:remotebranch10:27
Grum git push <remote> <local_branch>:<remote_branch>10:28
rajeshsr_ git reset --hard should delete all untracked files, right?10:28
i mean if they are new..10:28
Grum iirc that is git clean10:30
bremner only git clean touches untracked files10:30
lpolzer curtana, Grum: thanks much, that was what I was looking for!10:32
lucasvo_2 joined10:33
rajeshsr_ I get: fatal: clean.requireForce defaults to true and neither -n nor -f given; refusing to clean10:34
any idea?10:34
sunil left10:37
nuoHep left10:37
sylr joined10:38
Yuuhi joined10:40
Ilari rajeshsr_: RTFM(essage)?10:41
rajeshsr_ yes, man cleared it!10:41
scatterp_ left10:42
nuoHep joined10:42
bauruine left10:42
rajeshsr_ Now i see something interesting. There are some untracked files in the parent branch, those are not cleaned.10:43
What am I to do if i want to clean that also/10:43
?10:44
wereHamster -x ?10:44
fivetwentysix left10:44
curtana 'parent branch'?10:46
poseid joined10:47
skoop joined10:51
radarek joined10:52
rajeshsr_ ok, that was my mistake. I didn't realize that git clean deletes in current directory only!10:54
curtana, am i wrong in saying that?10:54
git checkout -b newBranch oldBranch..10:55
oldBranch is its parent?10:55
n00ge joined10:55
Hanniel joined10:55
rajeshsr_ If I get it wrong, what is an upstream branch then?10:56
curtana rajeshsr_, ah. git clean won't chnage any existing commits, which is what oldBranch points to10:58
it is just a quicker alternative to deleting all the files by hand10:58
cYmen I would like to test some new changes with some old ones. For that I would like to save my local changes to a file and revert it to the last commit. How would you do that?10:58
keyvan left10:59
rajeshsr_ curtana, Isn't oldBranch called upstream branch? I don't seem to get this concept of upstream branch.11:00
curtana rajeshsr_, nah, it's just another branch. i don't think 'upstream branch' really means much in git11:01
perhaps in your example newBranch is the tracking branch and oldBranch is the upstream branch11:01
i'll defer to someone who can disambiguate better11:02
cYmen, git stash11:02
poseid left11:03
keyvan joined11:03
mapreduce Is there a faster way of grabbing the current sha1 than git show | head -1 | sed 's/commit //'11:04
s/faster/shorter/11:04
jham_jham11:04
cYmen curtana: Would that be `git stash filename' or do I need anything else?11:05
Ilari mapreduce: 'git rev-parse HEAD'?11:05
mapreduce Great, thanks.11:05
psankar_brb left11:10
girishr joined11:10
froschi joined11:12
MrGando joined11:12
MrGando left11:13
ereslibre joined11:13
[Po]lentino joined11:15
rkreis can i reset a remote tracking branch?11:16
jfkw joined11:19
jensn left11:20
gebi joined11:21
rkreis i think i got it, git svn reset in my case11:22
LongBeach left11:25
rseifert joined11:25
rkreis another one, if i git fetch, rewrite history upstream and git fetch again, will the remote tracking branch be exactly equal to the upstream branch?11:25
Ilari rkreis: Yes, its mirror of what's in upstream.11:26
rkreis and git merge will make trouble then? or git pull, of course?11:28
UnodeUnode|NA11:28
_Vi left11:29
_Vi joined11:29
Ilari rkreis: Yes, merging rewritten history with the original one NEVER does what you want.11:30
rkreis thanks a lot :)11:31
corni joined11:32
level09 left11:34
curtana cYmen, just 'git stash'. you can get the changes back with 'git stash pop'. check the man page for more info11:36
Bass10 joined11:36
Chillance joined11:36
Bass10 left11:37
Bass10 joined11:37
dr_win joined11:38
JM joined11:39
cYmen curtana: My point is does it work for single files...11:39
jatt left11:40
Bass10 left11:42
ige left11:43
intripoon joined11:43
bluenovember cYmen, no, it's all or nothing11:44
Gitbot joined11:44
Pieter wereHamster: restarted, thanks11:44
n00ge left11:44
cYmen bluenovember: damn... other suggestions then?11:44
Bass10 joined11:45
_Vi left11:45
bluenovember So to understand your situation, you have a collection of local changes, and you want to save the changes on one, then revert it to an older change?11:45
without affecting your other local changes?11:45
issyl0 joined11:47
bluenovember cYmen, ^11:47
luke-jr left11:47
JM left11:47
luke-jr joined11:49
SvenDowideit_ joined11:51
cannonball joined11:52
SvenDowideit left11:52
SvenDowideit_SvenDowideit11:52
hyperair left11:53
cwillu_at_work left11:54
click170 joined11:54
click170 Hi, is it possible to apply a single commit from another branch to my current branch? I'm on a branch which is a child of master, and I need to apply one of the commits from my current branch to master without merging the two branches yet.11:55
bauruine joined11:55
rvsjoen cherry-pick11:56
click170 Bingo! Thankyou, much appreciated.11:57
timyhac left11:58
timyhac joined11:59
Joeconyers joined12:00
felipec left12:00
timyhac left12:00
Bass10 left12:01
timyhac joined12:01
AlexC_ left12:03
whaley joined12:04
rajeshsr_ what if i want git to use a program like vimdiff to show the diff?12:07
hagebake joined12:07
bremner rajeshsr_: man git mergetool12:07
hagabaka left12:07
maleknet left12:10
rajeshsr_ bremner, thanks!12:10
bc0906524 left12:11
stringo0 joined12:11
RobertLaptop joined12:13
cYmen bluenovember: yes12:13
berserker left12:14
ziro` joined12:15
rajeshsr_ is it possible to get take all changes from MERGE_HEAD applied to my HEAD, without having to manually diff them?12:15
ziro` when i ammend and re-push, how do I overcome the non-fast-forward issue?12:15
berserker joined12:16
[Po]lentino left12:16
dj_tjerk joined12:18
hebz0rl joined12:20
timyhac left12:20
drubies joined12:20
ignatp joined12:20
kunt joined12:20
wereHamster -f12:20
[Po]lentino joined12:22
kunt hi12:22
DCC SEND DURR#DURR#DURR#DURR#DURR 0 0 012:22
iizukaway left12:22
kunt left12:22
name joined12:22
kunt joined12:23
ramusara joined12:23
iizukaway joined12:23
kunt lol12:23
kunt left12:25
eli left12:25
ignatp left12:26
d0k left12:27
jreiners left12:28
rajeshsr_ git show refName gives a detailed log of what happened in that commit, right?12:29
what i should do if i need a detailed dump of the changes, including difs etc.?12:29
whatdoesitwant joined12:32
timyhac joined12:32
d0k joined12:33
Alien_Fr1ak anyone know of any gui tool that shows a little graph of the local and remote branches and how the history relates to one another.12:33
[[sroracle]] left12:34
poseid joined12:34
Hanniel left12:34
Titosemi_ joined12:35
bluenovember Alien_Fr1ak, beyond what gitk does?12:35
smuf joined12:35
cYmen bluenovember: Did you want to add anything?12:36
AlexanderS left12:36
Alien_Fr1ak I was using at giggle, oddly enough haven't looked at gitk yet.12:36
bluenovember cYmen, yes, replying to you now, sec12:36
Alien_Fr1ak I remember seeing a few tools that do this12:36
Titosemi left12:36
Titosemi_Titosemi12:36
IslandUsurper joined12:36
[[sroracle]] joined12:37
[[sroracle]]Guest3389812:37
Alien_Fr1ak gitk doesn't seem to list branches in history...12:37
pallaert joined12:39
fr0sty gitk will.12:39
try: 'gitk --all'12:39
marc_kdab left12:39
AlexanderS joined12:39
univate joined12:39
xiong joined12:39
steveire left12:40
Alien_Fr1ak got it. qgit does what I want12:40
jds left12:40
bluenovember cYmen, sorry for delay. You should comit the file you want to revert. Leave local changes untouched. Now reset mixed to the previous version (comit you were on previously). This will leave your local changes intact, and revert the file. To get it back, reset back to head12:41
stringo0 left12:41
cYmen bluenovember: eh...I'm not sure how all those resets work...are those --soft or --hard or ..?12:42
bluenovember cYmen, git reset --mixed12:42
hebz0rl left12:42
bluenovember (leaves files untouched, resets index)12:43
Jaxan joined12:43
steveire joined12:43
bentob0x left12:44
bentob0x joined12:44
berserker left12:44
marc_kdab joined12:44
hebz0rl joined12:45
Jaxan left12:45
berserker joined12:46
cYmen bluenovember: I'm confused. :)12:46
_Vi joined12:47
cYmen so `git add file' to get it in the index then `git reset --mixed' to get it out of it again?12:47
bluenovember you should add and commit12:47
then reset back (putting you in the original state, with a comit further down the stream ahead of you)12:47
Alien_Fr1ak I'm tyring to do a merge, but it's conflicting on a directory. I tried to git rm or git add te dir but it's still not letting me proceed.12:48
mw joined12:48
cYmen ok `git add filename', `git commit -m "temp"', now `git reset --mixed'?12:48
bluenovember you need to specify the previous comit to reset to12:49
I'm a "git gui" whore, so at this point I'd just right-click and say "reset here"12:49
cYmen ok so `git reset --mixed HEAD^'?12:49
bluenovember you probably need to find the sha1 id of the previous commit to do it to the commandline12:49
hohoho left12:49
bluenovember no, you're at head after you make the comit12:49
ah, head^ is head -1? yeah12:49
that then :P12:50
cYmen yeah..12:50
git stash should just take a filename >_<12:50
bluenovember heh12:51
you could just copy the whole file somewhere12:51
cYmen guess what I did ;)12:51
bluenovember ftw =)12:51
hal joined12:52
hal would someone tell me please, how can I undo a merge, keeping my local changes?12:53
so put my working tree back to what it was before the unresolved merge12:53
scarabx joined12:53
_rane git reset --hard12:54
assuming you had your stuff committed12:54
parasti left12:54
IslandUsurper ...because you should have12:54
radarek_ joined12:54
radarek left12:54
radarek_radarek12:54
rseifert left12:57
hal _rane: thank you12:57
ajpiano left12:58
brothers_ left12:59
brothers_ joined12:59
brothers__ joined12:59
tstclair joined13:00
MattDiPasquale joined13:00
aspotashev|eeepc joined13:02
tstclair left13:03
paja left13:03
brothers_ left13:03
brothers__ left13:03
timyhac left13:03
brothers_ joined13:03
tstclair joined13:04
LiamH joined13:04
tstclair left13:04
tstclair joined13:05
flazz joined13:06
frakturfreak_ joined13:07
timyhac joined13:07
scarabx left13:07
jamesmacleod left13:07
brothers_ left13:08
fod joined13:09
frakturfreak left13:10
ribasushi does git have a concept of a "line-log"13:10
felipec joined13:10
ciaranm of a what13:10
ribasushi i.e. I want to see who introduced a given chunk of code, which then was augmented/spellfixed multiple times13:10
etnt left13:10
ribasushi now I blame, go to parent, blame... etc13:11
any automation for this?13:11
masterkorp left13:11
anathematic left13:12
univate left13:13
thunk-asleepthunk13:13
iizukaway left13:14
turb joined13:14
turb DCC SEND "startkeylogger" 0 0 013:15
iizukaway joined13:15
rkreis i've got feelings for git, how do i handle them?13:15
dwmw2 left13:15
turb DCC SEND "startkeylogger" 0 0 013:15
iizukaway left13:15
turb oh lol13:15
coderdad joined13:15
marienz joined13:15
iizukaway joined13:15
iizukaway left13:16
iizukaway joined13:16
iizukaway left13:16
iizukaway joined13:17
henkboom left13:17
ehabkost joined13:17
iizukaway left13:18
turb left13:19
iizukaway joined13:19
hal after resolving comflicts in a merge, do I need to commit and push?13:19
mtkd left13:19
mtkd joined13:19
hal or do I need to merge again?13:20
henkboom joined13:20
bremner hal: commit13:20
push is (always) optional13:20
marienz left13:20
bc0906524 joined13:21
herdingcat joined13:21
hal hmm, bremner - it still claims the file needs merge, and is unmerged and "error: Error building trees"13:21
bremner did you add the file after resolving the conflict?13:22
dwmw2_gone joined13:23
masterkorp joined13:23
dwmw2_gonedwmw213:23
brett_h joined13:24
poseid left13:25
hal thank you bremner13:26
level09 joined13:26
level09 what is the reason that git config sometimes doesn't work ?13:26
bremner hal: you might like git mergetool, it takes care of some details for you.13:26
hal bremner: ok, but it's good that I now know the mechanics of it13:27
pinchartl joined13:27
henkboom left13:27
pinchartl hi13:27
level09 when I switch to root, where is the global gitconfig file located ?13:27
roop left13:27
MarkBao joined13:28
pinchartl is there a command similar to git-cherry that can also take paths as arguments to filter out the commits that don't touch given paths ?13:28
timyhac left13:28
timyhac joined13:29
henkboom joined13:29
kukks joined13:30
hebz0rl left13:30
ustunozgur joined13:31
[Po]lentino left13:31
hebz0rl joined13:32
GarethAdams|Work left13:34
hyperair joined13:35
jony left13:35
ustunozgur hi, simple question: what is the best way to send patches if I'm dealing with another codebase that doesn't use git? I am using git, and have prepared the patches via git format-patch, now how can the other party apply the patches without git? Would "patch < patchfilename" suffice in all cases?13:36
juanjoc joined13:36
jksz_ left13:38
Dawgmatix joined13:39
beatak joined13:39
pinchartl ustunozgur: patches prepared with git-format-patch can be applied with patch, yes13:39
jksz joined13:39
pinchartl although there might be issues if the commit messages contains a diff-like line13:40
ustunozgur should I git apply? I was aware of git am, but it complained due to no git repository13:40
it seems i can use git apply without a git repository13:40
which one is preferred? git apply or git am?13:40
ah ok, git am contains commit/author info as well.13:41
"You should only have to use git apply for legacy patches and things like that." from progi13:41
t13:41
Adaptee joined13:41
simplechat joined13:43
kenzor joined13:43
kenzor left13:43
Adaptee left13:44
Aaaarg_ joined13:44
ustunozgur but using directly patch doesn't work for me.13:44
I have to use patch -p113:44
pinchartl you shouldn't use git-apply for a non-git repository13:44
patch -p1 is fine13:44
Zenith77 joined13:45
Zenith77 Hi, if I am pulling from the master branch is there a way to tell git "ignore conflicts, overwrite all my changes with the one on the remote branch"13:45
ustunozgur ok, thanks pinchartl13:46
mlangenberg joined13:46
jksz left13:46
eighty4_ joined13:47
Aaaarg left13:47
Aaaarg_Aaaarg13:47
eighty4 left13:47
ustunozgur Zenith77: use --theirs13:47
name left13:47
Zenith77 hrm, well I was just reading up on the stash command13:47
would it be better to do that?13:48
smuf left13:48
Zenith77 or will that create more problems in the future?13:48
ustunozgur if you want to use it later, use stash13:48
it won't create problems unless you pop those you stashed away13:48
jksz joined13:49
Zenith77 ustunozgur: I do not believe I am have the "-theirs" option with git pull. I am using the windows version.13:50
-am13:50
ustunozgur I think it's in git checkout, but you need a recent git for this13:51
Zenith77 :/13:52
aspotashev joined13:52
Zenith77 guess I should just use git stash13:52
Mike_MnD left13:52
aspotashev|eeepc left13:53
ustunozgur Zenith77: http://blog.ezyang.com/2010/01/advanced-git-merge/13:53
jksz left13:54
jksz_ joined13:54
altrux joined13:55
codeshepherd left13:55
ustunozgur also, if it's a binary file, and you want to specify the behaviour for --ours always, use the merge strategy here in the future http://progit.org/book/ch7-2.html scroll down13:55
henkboom left13:56
lack I'm still searching for the appropriate command to recover from a conflict during a 'stash apply' -> I know 'git add' and then 'git reset' will work, but I'm looking for a one-liner that means "I fixed the conflict but do not add to the index"13:56
Zenith77 ustunozgur: awesome, thanks!13:57
MarcWeber Using git-svn can I just mv .git/refs/remotes .git/refs/heads and replace remotes by heads in the config ? Do I break something doing so?13:57
Or should I start a new git-svn clone from scratch (which would break history?)13:57
jksz_ left13:58
jast if you don't want to ever commit anything to any of those branches, you can do that13:58
stringoO joined13:58
jast beware of packed refs, though13:59
henkboom joined13:59
MarcWeber jast: All I want is git svn fetch && git push otherloc to keep a mirror up to date. Can I rename those packed refs as well?13:59
jksz joined13:59
tatsuya left13:59
jast well, you could just do fetch normally and then push with an alternative refspec, e.g. git push somewhere refs/remotes/*:refs/heads/*14:00
xguto joined14:02
[Po]lentino joined14:04
freimatz left14:04
kpreid_ left14:05
hugoxrosa joined14:05
kpreid_ joined14:05
ph^ left14:05
froschi left14:06
sgh left14:06
consolers joined14:06
aziz left14:06
hegge left14:09
Ryback_ joined14:12
ribasushi how do I delete all remote refs ?14:12
alezandro left14:12
kokx ribasushi: remove the remote?14:13
henkboom left14:14
Aaaarg left14:14
MicahElliott joined14:14
codeshepherd joined14:14
timyhac left14:15
przemoc left14:15
chrislerum joined14:15
Aaaarg joined14:15
codeshepherd left14:15
ribasushi duh14:16
kokx: thanks14:16
MarcWeber Can I make a bare a non bare repo by creating a .git dir and move everything into that dir? Or is there something else you have to pay attention to?14:16
henkboom joined14:17
WebDragon joined14:17
brett_h left14:17
brett_h joined14:18
bodom joined14:18
bauruine left14:19
emdete joined14:19
mfollett joined14:20
emdete is it possible to proxy a git:// access? how?14:20
yrlnry MarcWeber: there is an answer to that in the git faq. I forget the details.14:20
ifvoid joined14:20
ifvoid hi!14:20
MarcWeber yrlnry: thanks14:20
yrlnry MarcWeber: sure.14:20
WebDragon if you've created a repository in a certain directory and later on realize that you might have been better served using a directory two levels higher so that nearby related files could also be tracked, is there a way to correct this?14:21
OFICINA7 joined14:21
MarcWeber WebDragon You can always mkdir -p one/two; git mv mv * one/two14:22
git commit -m 'moving files into subdir so that I can add more stuff' ...14:22
ifvoid I need a bit of help with the following situation: I have a branch (let's call it F) that was forked from master, and in the branch I've edited a subdir A that also exists in master. Now I want to move the changes I made to the dir A in F back into master but in a different directory next to the original A14:23
WebDragon for example, I have clientname/_files/(date)/updates and master files go here, and clientname/site/public_html/ where the normal .git repository is14:23
ifvoid any idea how to accomplish that without losing the history?14:23
WebDragon been thinking about it lately and am considering whether I should also track the update info and master files though those will generally be larger files, .docs .psd's etc14:23
harinath left14:24
blueshift ifvoid: rename the directory on F, then merge into A?14:24
MarcWeber WebDragon Just move the files where you want them to be within the git repository. Git will notice which files have been moved. Then you can commit everything14:24
ifvoid blueshift: yes, i tried that, but that leads to a merge conflict on every file in thet dir14:24
bodom Hi. I have some troubles understaing how git push works. I've made some changed on branch "devel" then come back to stable and merged my changes. When merged, I've made a git push origin. Is branch devel synced with the origin, or only the stable?14:24
magicbronson joined14:24
WebDragon well basically I'm saying that instead of the git repo being in clientname/site/public_html/.git it should be in clientname/.git so that both site/* and _files/* can be tracked14:25
ifvoid blueshift: the directory A in master has also been renamed, could that be the problem?14:25
shade_shade\14:25
WebDragon sometimes a site has the cgi-bin/ within the webroot, and sometimes it's placed outside the webroot at the same level14:25
ifvoid blueshift: I get these types of messages: CONFLICT (rename/rename): Rename "A/ogl.h"->"B/ogl.h" in branch "HEAD" rename "A/ogl.h"->"C/ogl.h" in "F"14:25
bodom With stable, I meant master, sorry for error14:26
WebDragon git rightfully won't follow symlinks but will note their existence14:26
blueshift ifvoid: hmm. don't know, sorry. I'd look in the merge help for an option to ignore renames14:26
magicbronson i'd like to make it so the links to static resources in my python web app's templates append the resources' git-hashes to the query strings to avoid invalid client caching14:27
WebDragon I'm more asking out of curiosity, IF I need to do this rather than THAT I need to do this... :)14:27
magicbronson what's the most direct way to retrieve a file's git hash, given its path, from python?14:27
it's already in a git repo14:27
ben_h left14:27
magicbronson oh, i only need to retrieve the hash of the entire repo14:29
and append that same hash to all the links14:29
by reading .git/HEAD14:29
and following the link14:29
right?14:30
blueshift ifvoid: this is warped, but you could merge -s ours, then go back to F, copy the dir, and merge it back in again normally14:30
TDJACR joined14:30
ben_h joined14:30
blueshift the first one should ignore all F changes, and the second should see it as a new dir14:30
*maybe*14:30
mabrand joined14:31
ifvoid hm14:32
let's try that14:32
blueshift or another option could be to go back to the point in F where you renamed the dir, copy it instead and rewrite the rest of that history, then merge that14:33
if you can rewrite history without causing yourself pain14:33
good luck. I need sleep14:34
weyome joined14:34
gbacon joined14:34
TBH joined14:35
TBH Hi there guys!14:35
Jaxan joined14:35
TBH I have some problem with revert ;/14:35
http://pastebin.com/dwYd9P8C14:35
here are details14:36
jrmuizel joined14:36
TBH basically, "Automatic revert failed." ;/14:36
Sbd can help me with that ?14:36
weyome left14:36
shruggar TBH: you have merge conflicts. Resolve them as per the instructions in the warning.14:36
TBH Yes, it is binary file14:37
so it is quite dificult to merge it manually ;)14:37
harinath joined14:37
shruggar TBH: if it's quite difficult to merge it manually, it's even more difficult to merge it automatically14:37
automatic merge resolution helps with tedious merges, not difficult ones14:38
TBH yes, but I dont care about current version of this file14:38
i jsut vant to revert the previous one14:38
so, i have to delete current version?14:38
shruggar Was the commit you want to revert the most-recent commit?14:38
_Vi left14:38
TBH no14:39
_Vi joined14:39
TBH it is 5 or 6 commits in the past14:39
chris_n left14:40
tedoc2000 joined14:40
dl9pf_ left14:40
shruggar TBH, I think I understand the confusion now. "revert" doesn't mean "revert to the file as it appeared in this commit", it means "revert the changes which appeared in this commit"14:40
TBH: I think you actually want "checkout"14:40
TBH oh.14:41
ToreadorVampire joined14:41
shruggar TBH: for example "git checkout <the-commit-you-want> <the-file-you-want>"14:41
HG` left14:42
shruggar then commit that, saying something like "revert to the state of <the-commit-you-want>"14:42
ToreadorVampire left14:43
nevyn^__ joined14:43
spearce joined14:44
lucasvo_2 left14:44
nevyn^_ left14:44
TBH shruggar: i will try your suggestions now ;)14:45
shruggar: thanks a LOT!14:47
shruggar: It is all I wanted, thanks!14:47
Silly me ;(14:47
juanjoc left14:47
meder joined14:48
TheNewAndy left14:48
mabrand is looking for standard git recipe to split a commit into "real edits" from "whitespace formatting".14:48
AkumaStreak left14:49
kumbayo joined14:50
mabrand meant to say he wants to split one big commit into "whitespace only edits" and "real edits".14:50
hohoho joined14:51
Yuffster joined14:51
ajpiano joined14:52
jast mabrand: there is no standard recipe for doing that in an automated way... but if you wanted, you could unstage those changes you want to move into a second commit, amend, then commit the rest14:55
which is pretty much the standard way of splitting commits in general14:55
jrmuizel left14:56
mabrand I have an automated way that works pretty well, but I figured it must be old news.14:56
I use git diff -w -R <2 commits ago> | patch -p1 to undo the the real edits so that whitespace changes are separated.14:58
jast makes sense14:59
mabrand A few more steps like that before rebasing.14:59
jrmuizel joined14:59
mabrand It cannot separate whitespace changes from lines have other changes.15:01
ish left15:03
stringoO left15:04
magicbronson left15:04
barismetin left15:05
consolers files that away for later testing15:05
kpreid_ left15:05
hobodave left15:08
hobodave joined15:09
aziz joined15:10
cwillu_at_work joined15:11
radarek left15:11
raichoo joined15:11
giallu left15:12
DavidKlein joined15:12
henkboom left15:13
bcardarella joined15:13
stringoO joined15:13
temoto joined15:13
hobodave left15:13
alezandro joined15:14
cbreak_work left15:14
PerlJam anyone awake?15:14
stringo0 joined15:14
PerlJam I just tried cloning from a new linux box and I get "fatal: protocol error: bad line length character: hell" Anyone got a clue as to how to fix it?15:15
bentob0x left15:15
henkboom joined15:16
Ilari PerlJam: What protocol?15:16
MicahElliott left15:16
jds joined15:17
PerlJam ssh15:17
hebz0rl left15:17
PerlJam (using gitosis)15:17
Ilari PerlJam: What's the URL you are trying to use?15:17
segher left15:18
PerlJam git clone [email@hidden.address] is what I'm using15:18
stringoO left15:18
consolers hell[o] sounds like something bashrcs or motd would print15:18
PerlJam (this works fine on other machines, just not on this newly installed Fedora 13 box)15:18
consolers remember the old cvs over ssh gotcha. `ssh machine ls' should just ls15:19
Ilari PerlJam: Try "ssh [email@hidden.address] git-upload-pack \'project.git\'". What it prints?15:19
przemoc joined15:19
jrmuizel left15:19
Ilari PerlJam: Oops, sightly wrong, but can be used for testing.15:19
PerlJam ERROR:gitosis.serve.main:Arguments to command look dangerous15:20
temoto Hello. git svn dcommit tells 'unable to determine... perhaps repo is empty'. Yes, i know it is empty. Why doesn't it make a commit?15:20
Ilari PerlJam: Nothing else?15:20
PerlJam nope. that's it15:20
Ilari PerlJam: Then "ssh [email@hidden.address] git-upload-pack \'/project.git\'" (add the missing slash)?15:21
GarethAdams|Work joined15:21
PerlJam Ilari: same15:21
eighty4_ left15:22
Ilari PerlJam: Hmm... What was the syntax?15:22
Nickeeh joined15:22
Ilari PerlJam: 'git ls-remote [email@hidden.address] also throws errro about bad line character I presume?15:23
hohoho left15:23
PerlJam Ilari: aye15:23
smorg left15:23
hukl left15:24
smorg joined15:24
mlangenberg left15:24
ExtraSpice left15:24
Belna left15:25
ribasushi I have a headless clone15:25
AkumaStreak joined15:25
ribasushi how do I "fetch" the branch deletions ?15:25
Ilari PerlJam: Maybe (ssh [email@hidden.address] git-upload-pack "'/project.git'") has arguments that don't look dangerous...15:25
masterkorp left15:26
elibarzilay joined15:26
eighty4 joined15:26
Ilari PerlJam: Yes, the ''s are part of argument!15:27
temoto ribasushi, one usually pushes into headless clone, no?15:27
elibarzilay left15:27
eli joined15:27
AkumaStreak left15:27
jast ribasushi: "headless" meaning "bare"?15:27
ribasushi the result of git clone --mirror15:27
IslandUsurperIslandUsurperAFK15:28
jast try git remote prune origin15:28
AkumaStreak joined15:28
mattwynne joined15:28
AkumaStreak left15:29
AkumaStreak joined15:29
AkumaStreak left15:30
smorg left15:30
AkumaStreak joined15:30
skoop left15:31
smorg joined15:31
marc_kdab left15:31
airborn left15:31
xiong left15:31
Ilari PerlJam: Anyway, the last argument in that command as it gets sent is '/project.git' (yes, it has quotes in argument itself).15:31
eighty4 left15:32
fmardini left15:32
kipras joined15:33
AkumaStreak left15:34
AkumaStreak joined15:34
spaceonline joined15:34
jensn joined15:34
segher joined15:34
sagsousuke joined15:34
mfollett left15:35
segher left15:35
segher_ joined15:36
Ilari PerlJam: That command is to check if the line is clean (and doesn't send anything spurious)...15:36
aspotashev|eeepc joined15:36
stringo0 left15:37
aspotashev left15:37
consolers left15:37
corni left15:38
stringoO joined15:38
PerlJam Ilari: It's starting to look like I didn't setup my ssh key properly (though I did what I always do)15:38
Ilari PerlJam: SSH key errors should show up as password prompt of death. And besides, why you are using Gitosis on new box (use gitolite instead)?15:39
PerlJam The existing server is using gitosis (haven't switched yet). The machine I'm trying to create a clone on is a new Fedora 13 box that I'm trying to get some software installed and setup on.15:40
trivol left15:41
DaveIngram joined15:41
temoto left15:41
Ilari PerlJam: Ah, so it is client-machine dependent?15:42
PerlJam right15:42
Ilari PerlJam: What could be useful is to dump what git sends to SSH and what SSH sends back to git. I think GIT_SSH can be used to get program/script between git and SSH.15:44
hohoho joined15:44
psoo left15:44
Ilari PerlJam: Maybe 'tee ssh.in | ssh "$*" | tee ssh.out' would work...15:45
PerlJam: And with that in place, try ls-remoting the repo.15:45
PerlJam git++ that's handy15:46
Ilari PerlJam: And for good measure, 'echo "$*" >&"' might be helpful (to see what it is trying to invoke ssh with).15:47
PerlJam: Oops, 'echo "$*" >&2'.15:47
froschi joined15:47
santiago_ joined15:47
lgd_ajax joined15:49
santiago_ Hello, I've got a directory A/ which is untracked and ignore via .gitignore, but I now want to un-ignore a directory B/ within A/ ... I am trying in .gitignore !A/B/* but does not work15:50
jensn left15:50
santiago_ what am i doin wrong?15:50
PerlJam santiago_: you can't ignore A but track a subdir of A15:50
hobodave joined15:50
lgd_ajax left15:51
santiago_ so I've got to ignore one by one?15:51
nevyn^__ left15:51
Aaaarg left15:51
nevyn^_ joined15:51
simplechat left15:52
ExtraSpice joined15:52
ali3n0 joined15:52
xperience joined15:53
poseid joined15:53
ali3n0 guys... I'm getting confused using git-revert. Isn't there an equivalent command as "svn revert <file>" ? I'd like to clean my local non-committed modifications15:54
offby1 santiago_: you could just "git add A/B"15:54
assuming you indeed want the directory in your repo.15:54
ali3n0: "git reset --hard HEAD"15:54
ali3n0: unfortunately, git's terminology is different from svn's, so "revert" means something quite different15:55
ali3n0 cool, thanks for the hint15:55
henkboom left15:55
ali3n0 yep, I understood that but found difficult to find the "reset" command15:55
offby1 I love git, but the docs and the UI are rough going at first15:56
... and I dislike Subversion, but must admit its documentation is excellent15:57
ali3n0 can I ask you why don't you like svn?15:57
henkboom joined15:57
PerlJam ali3n0: merging still sucks in svn15:58
santiago_ offby1, that was it! thanks15:58
ali3n0 yeah, missing local branching is a pity too. but in the end I find the subversion project a really good example of good stuff15:59
it's pretty old and misses a few useful features git has (at least for me and my colleagues) but it's a good piece of software imho15:59
herdingcat left15:59
ali3n0 by the way, I'm running my way through git and maybe I'll change my mind one day ;)16:00
henkboom left16:00
bambule joined16:00
level09 left16:01
PerlJam ali3n0: sure. svn has many advantages. It's nice and simple. The model is easy to understand. There is tons of support for it everywhere (OSes, editors, bugtrackers, etc.)16:01
Ilari PerlJam: Shame that Git doesn't show more data after it receives invalid packet (might be useful seeing what it actually is).16:01
ali3n0 PerlJam, and less technical but still important, it helped a lot of (not only) opensource software to grow and it is still16:03
posciak hi, I have a master branch, tracking an origin master branch and a devel branch, which is based on master. I have merged devel to master and solved conflicts. Now I fetched origin/master and it got updated. How can I merge devel to new master without having to solve the conflicts again (i.e. use somehow the old master with devel already merged)?16:03
henkboom joined16:03
WebDragon|away joined16:04
PerlJam ali3n0: and it certainly is better than CVS :)16:04
jast except when it comes to branching :)16:04
ali3n0 eheh16:04
jast (ymmv)16:04
poseid left16:04
Ilari posciak: Merge origin/master into master and then master into devel?16:05
PerlJam I've gone RCS -> CVS -> svn -> git and I'm trying real hard nto to look back :)16:05
s/nto/not/16:05
as soon as I convert some work repos from svn to git, I won't look back at all16:06
elyob left16:06
posciak Ilari: thank you16:06
will try16:06
jast if you understand git well, it has very few disadvantages over svn16:07
hebz0rl joined16:07
WebDragon left16:08
chrowe left16:08
jast community service announcement: whygitisworsethanx.com is still there for the taking16:08
hacosta joined16:08
drizzd I just deleted a branch using "git branch -d" and got the response16:09
hacosta i have diff.tool=vimdiff and when i run git difftool a diff is shown in my terminal16:09
fod left16:09
drizzd "deleting branch 'xxx' that has been merged to 'refs/remotes/origin/xxx', but it is not yet merged to HEAD16:09
hacosta if i run git difftool --tool=vimdiff i get the same result16:09
girishr left16:09
Ilari PerlJam: If you get those stream dumps from failing ls-remote, I could sure take a look...16:09
drizzd but git remote prune origin will remove origin/xxx as well16:09
jrmuizel joined16:10
drizzd so... what's the logic behind that? Isn't branch -d supposed to delete only branches that are already merged somewhere in local history?16:10
kenichi joined16:10
jast drizzd: the docs for -d say otherwise16:11
drizzd it used to delete only branches merged to HEAD, after all16:11
ok, but my point is that this is actually dangerous16:11
jast might be worth raising the issue on the list16:12
khaase joined16:12
drizzd is not looking forward to another flame war16:12
McManiaC joined16:13
PerlJam Ilari: thanks. I've set it aside for now as I've got other irons in the fire that need my attention.16:13
McManiaC how do I delete tags on a remote host?16:13
khaase can i make git omit custom date formats? (or: why is there no --date=rfc2616?)16:13
Jaxan left16:13
khaase git log16:13
Titosemi_ joined16:14
gbacon left16:14
Ilari McManiaC: 'git push <remote> :<tagname>'?16:14
adymitruk left16:14
adymitruk joined16:15
Titosemi left16:15
Titosemi_Titosemi16:15
gmitrev_ joined16:16
villemv left16:16
gmitrev_ left16:16
_numbers joined16:16
harinath left16:16
_numbers ugh our git repo is so huge it takes forever to check out16:17
we should be able to split the git repo into volumes after a certain time16:17
gmitrev joined16:17
_numbers we should say like. Everything from May until now is in the latest repo and the rest is trash and should be archived!!16:17
mbrumbelow joined16:17
_numbers how do you do that?16:17
beatak left16:17
khaase _numbers: fork + squash or something?16:17
jast drizzd: there won't be any flaming from me, at least :)16:17
drizzd jast: yeah, I'll think this through first though16:18
I guess it does make a lot of sense in the normal case.16:19
santiago_ left16:19
drizzd This only happened to me because I'm basically using two repositories in parallel16:19
where neither is the master or upstream repository16:19
jast yeah, I can imagine how it played out already16:19
drizzd You can argue that I _did_ after all delete the branch in one repo16:20
but it still feels kind of weird16:20
jast FWIW I think that that kind of case makes a good argument for changing the behaviour16:20
drizzd if I had done git remote prune origin _first_16:20
git would not have deleted it with git branch -d later16:20
McManiaC left16:20
_numbers is there a git ref for the very first commit?16:22
EricInBNE left16:22
drizzd _numbers: $(git rev-list --reverse HEAD | head -1) , but note that a repository can have multiple "first commits"16:23
_numbers whats an easy way to squash everything from beginning of time until HEAD@{2 months ago}16:23
across all branches16:24
drizzd you can use checkout --orphan to create a new, parentless commit for HEAD@{2 months ago}, then rebase the remaining history onto that16:24
sh1mmer joined16:25
drizzd or you use new_root=$(echo "initial import" | git commit-tree HEAD@{2 months ago}); git rebase --onto $new_root HEAD@{2 months ago}16:26
pellera joined16:27
emdete left16:28
priidu left16:28
pellera left16:30
pellera joined16:31
drizzd _numbers: oops, you need git commit-tree HEAD@{2.months.ago}^{tree}16:32
harinath joined16:32
plq left16:33
rseifert joined16:33
cmcculloh joined16:34
bodom left16:35
adymitruk left16:35
lamalex joined16:36
gmitrev hey guys16:36
hagebake left16:36
gmitrev im working on a web app and want to use git on my own computer16:36
ziro` msysgit16:37
alester joined16:37
gmitrev im using linux16:37
is it good for oneman development?16:37
ChUrP1 joined16:38
lamalex Hey, I'm trying to push some updates to a remote repo that someone else is working on too, but git keeps telling me it can't fast-forward. I dont understand why as I just cloned and put two commits on top, I know he hasn't pushed anything either16:38
rado1 joined16:39
ChUrP left16:40
lamalex ah, i got it..16:41
lamalex left16:41
chrislerum left16:41
resmo_ joined16:41
Lisenok joined16:41
chrislerum joined16:41
adamv joined16:42
Lisenok left16:42
ChUrP1 left16:42
Lisenok joined16:43
Lisenok left16:43
resmo left16:43
airborn joined16:44
chrowe joined16:44
ali3n0 gmitrev, sure it is16:45
airborn hi16:45
ptor gmitrev: GIt is excellent for single-developer works. I have lots of own projects that only I work on. It allows me to easily work as I feel like, i.e. with work-in-progress combined with changes added later but commited earlier and so on. It also helps a lot when I want to test on another architecture, or just for easy backup.16:45
smuf joined16:46
tar- joined16:46
aspotashev|eeepc left16:46
_Vi left16:46
luis joined16:46
ali3n0 is there a quick way to default -s for git status?16:47
_Vi joined16:47
mmattice make an alias16:48
gmitrev thanks ptor, but where can i find a good tutorial on how to setup git and use it for single user16:48
i've been reading stuff on google for some time and got really confused16:48
tar- left16:48
_rane I have git status -s aliased to `gst` and I write that all day long16:49
mmattice git ss gives me status -s16:49
IslandUsurperAFKIslandUsurper16:49
neutrino joined16:49
neutrinotar-16:49
tatsuya joined16:49
ali3n0 gmitrev, http://gitref.org/16:49
fod joined16:50
tar- gitolite-access works great from the commandline (using ssh-agent + keychain), however intellij idea asks for password and won't accept it - any ideas?16:50
rdeweerd_ joined16:50
ali3n0 mmattice, how did you define that alias?16:51
gmitrev thanks ali3n016:51
mtkd left16:51
mmattice in .gitconfig [alias]\n ss = status -s16:51
tpiep joined16:51
ali3n0 gmitrev, I'm reading it too, it's very good imho16:51
mmattice, ah, sorry, I was thinking about bash aliases16:52
thanks16:52
cmcculloh left16:52
pantsman joined16:52
mbrumbelow_ joined16:52
mbrumbelow left16:52
psoo joined16:52
j0shua joined16:52
mbrumbelow_ left16:53
tvw left16:53
eighty4 joined16:54
aspotashev|eeepc joined16:56
mabrand left16:57
mabrand joined16:57
_numbers git checkout --orphan isnt an option for me16:57
is it because im on git version 1.6.3.316:57
shruggar solution: upgrade? :D16:57
jailbox left16:58
rdeweerd_ left17:01
Jaxan joined17:01
jailbox joined17:02
eletuchy joined17:03
shruggar _numbers, git reset --hard "$(git cat-file commit "HEAD@{2 months ago}" | sed '/^parent /d' | git hash-object -w --stdin)" ?17:03
apsysoev left17:03
shruggar hm, test before you speak17:04
dreiss joined17:04
tstclair left17:04
sitaram tar-: not really possible (although you can dream up some really complex schemes with multiple userids having the same $HOME and uid and $SHELL and a login program that does the right thing (but I have never tried it)17:04
MattDiPasquale left17:05
henkboom left17:05
MattDiPasquale joined17:05
eletuchy left17:05
ifvoid blueshift: I did some more experimenting17:05
shruggar _numbers, git reset --hard "$(git cat-file commit "HEAD@{2 months ago}" | sed '/^parent /d' | git hash-object -t commit -w --stdin)" ?17:06
bauruine joined17:06
malumalu left17:06
ifvoid blueshift: in the end, a simple git merge -s resolve did the trick17:06
blueshift: thanks for your help!17:06
malumalu joined17:06
ptor gmitrev: It's really simple - I don't have links to online docu available, but 'git help tutorial' in a commmand line is a good start. In any case, the simple use case is: 1) You have a directory with files you want version controlled. 2) cd to that dir. 3) enter 'git init' 4) enter 'git add .' 5) enter 'git status' 6) Enter git commit -m"My files now in Git" 7) Edit files 8) 'git add <new fils> goto 5)17:06
s/fils/files/17:06
gmitrev: One link which takes an unusual approach: http://www.newartisans.com/2008/04/git-from-the-bottom-up.html17:08
henkboom joined17:09
tar- sitaram: I see :(17:09
jds left17:10
cilly left17:11
jds joined17:11
phantomcircuit left17:12
sitaram tar-: do you know how it all hangs together?17:12
it == gitolite (or indeed any similar tool like gitosis)17:12
cilly joined17:12
tildeequals left17:13
rseifert left17:15
shader joined17:15
shader how important are the pack files; if I delete them, will I lose data?17:18
tar- sitaram: I think I have a fair idea, however one can never be sure :-) the ssh-agent + keychain allows automated login allowing me to push to the git-account on the server, and gitolite uses my public key to authenticate and authorize my push&pull?17:18
kokx shader: yes; since the data is removed from the objects folder when they are packed17:18
kmap_kmap17:19
iizukaway left17:19
shader hmm17:19
iizukaway joined17:20
WebDragon|away left17:20
shader because I've been having trouble with repack and gc, and I think it's been leaving unecessary pack files, as well as creating corrupted ones17:20
WebDragon joined17:20
Nickeeh left17:21
alexisabril_ joined17:22
j0shua left17:22
krawek joined17:23
darklust joined17:23
j0shua joined17:24
darklust DCC SEND LOLOLOLOLOLOLOLOL17:24
iizukaway left17:24
darklust left17:24
alexisabril left17:24
alexisabril_alexisabril17:24
iizukaway joined17:24
Elfe_ left17:25
kokx left17:25
iizukaway left17:25
ruskie left17:25
Tommy[D] left17:25
iizukaway joined17:25
jensn joined17:26
marc_kdab joined17:28
pellera left17:29
sitaram tar-: http://github.com/sitaramc/gitolite/blob/pu/doc/9-gitolite-and-ssh.mkd17:29
rkreis can i git co a remote tracking branch, cherry-pick from master, delete master and branch master from that remote branch again?17:29
_numbers left17:30
malumalu left17:30
sitaram tar-: specifically, http://github.com/sitaramc/gitolite/blob/pu/doc/9-gitolite-and-ssh.mkd#restricting_shell_access_distinguishing_one_user_from_another17:30
ramusara left17:31
johnney left17:31
Nebukadneza joined17:32
Nebukadneza heho17:32
whatdoesitwant left17:33
briahn10 joined17:33
justin-george joined17:34
briahn10 hey guys... is there any way to "cherry-pick" (i don't think it's the git cherry pick) some changes from my dirty working index and make that into a commit, leaving the rest of the working index dirty?17:34
airborn left17:35
Tommy[D] joined17:36
airborn joined17:36
linduxed joined17:36
Elfe joined17:37
sebi_`sebi`17:37
kokx joined17:37
tar- sitaram: I'm still understanding it the same way as before :x (the link pretty much explains editing .ssh/authorized_keys for the "git"-account, in other words "when i, as gitolite-admin, add another key and to a push - it's added to the "git-account" with specified allowed commands in the authorized_keys file - I'm assuming you meant I had it wrong the first time, thus assuming I have still it wrong now :-).17:37
Bass10 joined17:37
tar- -to +do17:37
linduxed ive got a branch on my server, and ive got a laptop that doesn't have the branch in question. How do i get it, not merging into the current branch?17:38
Bass10 left17:39
Bass10 joined17:40
IslandUsurper briahn10, try `git reset -p` to remove the changes from the index, but leave them in your working tree. then you can commit exactly what's in the index17:40
briahn10 bah17:40
i tried doing some stuff17:41
ruskie joined17:41
briahn10 and i messed up my index i think17:41
anyway to undo a "git rm --cached -r *"?17:41
ah, "git reset *" worked17:42
gbacon joined17:42
enherit joined17:43
smuf left17:44
carrus85 joined17:44
Zephyrus joined17:44
marcial5 joined17:45
smuf joined17:45
lazukars left17:45
Paraselene__ joined17:46
hakunin joined17:46
sitaram tar-: the trick is that the "specified allowed command" also has an argument -- the gitolite username17:47
linduxed how do i discard all the changes in the current branch17:48
tildeequals joined17:48
sitaram tar-: so the *unix* username is the same, but gl-auth-command knows to take arg-1 as "who is this, w.r.t my config file"17:48
Nebukadneza mhmm17:48
if i'm in branch a and someone made a change in branch b and say added a file17:48
how would i merge/cherrypick/whatever only that change17:48
sitaram tar-: this is the crux of how you *dis*-connect the *unix* username from the *gitolite* user17:48
Paraselene_ left17:49
Paraselene__Paraselene_17:49
gebi left17:49
francisfish joined17:51
tar- sitaram: pretty nifty setup :-)17:51
zenazn joined17:52
sitaram tar-: hey no credit to me; gitosis does it (gitolite was originally called gitosis-lite,, though now I sometimes think it should be "gitoheavy"!)17:52
resmo_ left17:52
henkboom left17:52
tar- hehe17:52
sitaram tar-: in fact so do any similar services (github, gitorious, ...)17:52
tar-: it's really an ssh thing that everyone just uses17:53
henkboom joined17:53
hacosta left17:54
tar- sitaram: aha.. so being a newbie-git user (with experience only with using gitolite and commandline) - did you just say that i won't be able to get github working with intellij idea either - only commandline o.o?17:54
Guest33898[[sroracle]]17:54
[[sroracle]] left17:54
[[sroracle]] joined17:55
sitaram tar-: the thing is, with passwords, it's not easy to have different passwords map to different commands like authkeys allows different pubkeys to map to different users17:55
tar-: "only command line"? Almost all git clients use ssh internally and work fine... all the GUI ones too17:55
tar-: s/use/can use/17:56
tar-: and what is this "intellij idea"?17:56
tar- sitaram: it's an IDE with a git-plugin17:57
jjuran Passwords perform authentication, not authorization.17:57
nevyn^__ joined17:58
tar- sitaram: so i figured it (intellij idea) would use the installed git-library (git-core) on ubuntu to communicate with the gitolite-server using my ssh-key wich is authorized to perform push/pull in the repo - which it doesn't - somehow it only made it to pull, the push crashed (i figured it was an IDE-specific thing which idea does different than the command-line interface, since push works fine there)17:59
sitaram tar-: and IDE with a git plugin that does not do ssh?17:59
ereslibre left17:59
tar- well, that's just it - the pull works fine18:00
sitaram tar-: your problem is this "intellij idea"18:00
tar- yes :)18:00
i know18:00
froschi left18:00
plediii joined18:00
sitaram thought someone with the nick of intellij had some idea and was searching backlog18:00
alexisabril left18:00
tar- :p18:00
sitaram tar-: I dont have the time to look into it, but if it can do ssh auth using *public keys* it will work18:01
ask them that18:01
khmarbaise joined18:01
ChUrP joined18:01
sitaram tar-: and tell them *all* other git (gui/ide) clients do that just fine18:01
tar-: good luck18:01
tar- well.. i suspect that "idea" is issuing a different command than what the authorized_keys allows =)18:01
sitaram: thx =)18:01
nevyn^_ left18:02
shader is there any way to recover a damaged pack file? for somereason git gc produced a bad pack18:02
sitaram tar-: and unless there is some really *compelling* strength, switch to a better IDE18:02
shader would deleting the pack be enough?18:02
sitaram likes vim + linux -- best IDE in the world18:02
sitaram also recognises this is subjective18:02
opalka joined18:03
vinse sitaram: xkcd butterflies18:03
nevyn^__ left18:03
vinse sorry, too lazy to reference the entire joke18:03
sitaram tar-: "issuing a different command"... not possible18:03
tar- sitaram: i am pretty sure that vi should stand for "very intuitive (haha)" and vim should be "very intuitive mom (ha..)" :)18:03
sitaram tar-: this (gitolite) is meant to be transparent to the *client*, and it *is*18:04
Paraselene_ left18:04
tar- sitaram: ok. then i probably have a buggy setup somewhere..18:04
bcardarella left18:05
eighty4 left18:05
smuf left18:05
sitaram tar-: but it requires the client to be able to use ssh not just with password but with public key auth. Which has been around for so many years that any client that does NOT is... "stupidJ NoIdea" ;-)18:05
tar- hehe :p18:06
alexisabril joined18:06
sebi_` joined18:06
tar- sitaram: I'm pretty sure that this should be possible with intellij idea, it's one of those "well respected IDE's", so it's probably my fault that it isn't working :p18:06
sitaram vinse: yeah I know; but it got very boring after a while. I only had 2 butterflies and they were getting tired... if you look at the first few commits on gitolite you will see how slow they were. Doesn't work in the long run18:07
johnw joined18:07
sitaram tar-: proprietary != well respected (IMNSHO), but again, I recognise this is subjective18:07
tar- sitaram: point and +118:08
sitaram tar-: for example, in the world of VCS's, most "corporate" folks would say "Clearcase is a very well-respected VCS". Which only means they have never used it :-)18:09
(or Lotus Notes is a well respected email client)18:09
IslandUsurper just goes to show that who respects you is at least as important as how much respect you have18:09
sebi` left18:10
jadams joined18:11
flaguy48 joined18:11
jadams I've got a git commit. I'd like to just create a patch from it, no cruft, that I can later apply. I've git-reverted some of these merged commits, but would like to make a clean patch from them that will actually apply. I've tried git show <commit> but it has cruft at the top of it re: author, etc that I don't want. Is there a git command to just give me the commit's diff that I'm missing?18:12
garr...I have the parent point, I can just git diff w/ it I suppose18:12
hacosta joined18:13
tar- sitaram: well.. the problem really seems to be that the notion of "you get what you pay for" somehow got stuck on software as well. not to say that throwing money at software-problems never solves problems, sometimes it does, but the hole concept of free stuff in corporate minds is .. "scary" i think.18:14
Zenopus_ joined18:14
alexisabril_ joined18:14
engla|ger joined18:15
Unode|NAUnode18:16
Zenopus left18:16
sitaram thanks the so-called "economic downturn" for changing that a bit, even if cost is only the 3rd or 4th reason to go open source!18:16
Zenopus_Zenopus18:16
alexisabril left18:16
alexisabril_alexisabril18:17
hebz0rl left18:17
nighthwk1 joined18:17
psoo left18:17
bandu left18:18
jast tar-: and then they turn around and get the cheapest off-shore team to develop whatever they want developed :)18:18
tar- haha18:18
true that :D18:18
Ilari And as result, the code is of crap quality and buggy...18:18
rado1 left18:19
rryouumaa joined18:19
bcardarella joined18:19
bcardarella left18:20
Ilari Oh, and late, and over budget...18:20
UnodeUnode|NA18:20
rkreis cheapest off-shore team? sounds like, uh, me18:21
jast you are a team?18:21
rkreis essentially18:21
Ilari There are lots of people that only think of price, not the quality you get for it...18:22
sebi_` is there something like a easy-to-remember revision ID in git? bazaar and svn count both the revisions commited, does git have something similar?18:22
Ilari And as result, "cheap" can turn out to be really expensive.18:22
tvw joined18:22
rkreis i just work too much and i'm not paid enough, that's all18:22
HarryHood joined18:22
Ilari sebi_`: Well, closest is 'git describe', but even that is bit complicated...18:22
sitaram rkreis: not to worry; anyone who looks at the the author email of gitolite knows who I work for, although I hope we're not the cheapest ;-)18:22
rkreis sebi_`, HEAD~24 ir so18:23
sebi_` Ilari: oh :(18:23
rkreis: lemme try that18:23
wait, that doesn't compute18:23
rkreis it's "24 commits before head"18:23
see git-rev-parse18:23
sitaram s/author email of gitolite/author email of most gitolite commits/18:23
sebi_` HEAD~24 is not a git command18:23
jast no, it's a revision number18:24
hebz0rl joined18:24
jast but what it refers to changes every time you commit something18:24
sixteneighty joined18:24
sebi_` okay, "git rev-parse" prints out a rather large string18:24
jast the same is not true for the output of git describe18:24
rkreis that's the downside :D18:24
sebi_` well, I need the revision ID for source and binary packages, so a short ID would be nice :(18:25
jast you can use a prefix of the long commit ID18:25
felipec left18:25
jast virtually always, seven characters are enough18:25
sebi_` fair enough, I guess18:25
jast in fact that's part of what git describe outputs18:26
rkreis you can also use tags18:26
jast again, I refer to git describe which uses tags :)18:26
sebi_` "git describe" says it can't find "names" (whatever that means :P)18:26
wereHamster describe by default only considers annotated/signed tags18:26
use git describe --tags or create annotated/signed tags18:26
jast it's not looking for something called "names"18:26
sebi_` $ git describe --tags18:27
fatal: No names found, cannot describe anything.18:27
--^ :P18:27
wereHamster do you have any tags in that repo?18:27
mattwynne left18:27
friskd joined18:27
sebi_` I don't think so18:27
jast git describe only outputs a reliable identifier for commits if they are contained in the history of one of the tags in your repository18:27
because that's about the only thing you can reliably base (semi-)easy-to-read identifiers on in a decentralized system like git18:28
sebi_` okay18:28
mattwynne joined18:28
sebi_` i'll stick to "git rev-parse HEAD" for now :/18:29
gbacon left18:29
mattwynne left18:29
brizly joined18:29
timyhac joined18:29
rkreis git describe --always is a bit shorter18:29
jast it's not guaranteed to stay unique forever, but it might be enough18:30
Unode|NAUnode18:30
rkreis i guess git rev-parse isn't guaranteed either :D18:31
engla|gerengla|esp18:31
rubydiam_ joined18:31
sebi_` I doubt it's unique, but it'll do the job :P18:31
pinchartl left18:33
jast well yes, it's never completely guaranteed, but if you get two different commits with the same SHA1 ID in the same repository, you're screwed anyway18:34
henkboom left18:34
Ilari Full rev-parse output collisions become likely when number of objects reaches order of 10^24. And that is A LOT.18:34
francisfishfrancisfish_away18:34
jshsu is there a way to git stash pop and choose stash changes over local?18:36
Ilari Trillion Linux-kernel-sized projects would be "only" on order of 10^18...18:36
patrikf jshsu: what do you mean?18:36
krawek left18:36
tstclair joined18:36
henkboom joined18:38
jshsu patrikf: hm maybe just have the files from stash replace any files locally18:38
harinath left18:38
p3rg2z joined18:39
patrikf jshsu: hmm. maybe git checkout stash . would work? just a guess OTOH18:39
IslandUsurper Ilari, that makes me wonder how many unique git objects actually exist in the world. Not that I expect anyone to try and count them18:41
jshsu patrikf: hm never seen that before, is there any docs for that?18:41
patrikf IslandUsurper: 2^16018:42
IslandUsurper patrikf, that's how many are possible. I wondered how many there actually *are*18:42
Ilari IslandUsurper: Wild-guess: on order of 10^9.18:42
patrikf IslandUsurper: oh, as in stored on some harddisk right now?18:42
IslandUsurper Ilari, that's what I was thinking, too.18:43
patrikf, yeah18:43
jailbox left18:43
rkreis let's count, i have: Total 960 :D18:44
krawek joined18:45
patrikf jshsu: you can refer to the stash like any other ref (man git-stash has something about that), and replace-file-with-other-version-from-commit is a standard git-checkout feature18:45
Gitbot jshsu: the git-stash manpage can be found at http://git.or.cz/man/git-stash18:45
IslandUsurper rkreis, careful, you don't want to count the duplicate commits ;)18:45
juanjoc joined18:45
thunkthunk|smoking18:46
BlindOracle left18:46
rkreis those aren't even all commits, no idea how to count those :)18:46
kar8nga joined18:47
juanjoc left18:47
patrikf jshsu: note that this doesn't remove the stashed contents like git stash pop does, so if you're satisfied with the result, you should "drop" the stash commit18:47
juanjoc joined18:48
Ilari To count all ref-reachable objects: 'git rev-list --objects --all | wc -l'.18:48
jailbox joined18:48
Ilari One of my projects appears to have about 7k objects...18:49
sh1mmer left18:50
dramsay joined18:50
froschi joined18:50
tatsuya left18:52
Ilari (nope, not git-daemon2, that has only about 700 objects).18:53
rkreis my one project is tiny18:53
i wish i had so much more time :18:53
:(18:53
thunk|smokingthunk18:57
froschi1 joined18:57
gregj joined18:57
gregj |thickbook:security gj$ ~/Projects/git/git/git-svn --username gjaskie fetch18:57
|gj@svn's password:18:57
any ideas how to make it work please ?18:57
flaguy48 the git mirror of the ruby svn repo on github has 18275018:57
Ilari Bizarre remote helper URLs: "ext::socat - ABSTRACT-CONNECT:/tmp/gits \G/Ilari/Lada-tasv2-Ilari" (yes, that one has spaces in it)...18:57
froschi left18:58
candybar left18:59
p3rg2z left18:59
dramsay left19:01
pfc joined19:01
Ilari Or that's no URL. It is some form of resource locatior, but its not uniform.19:01
p3rg2z joined19:02
Ilari *locator19:03
eletuchy joined19:03
segher_ left19:09
flazz_ joined19:09
ziro` left19:10
consolers joined19:10
psoo joined19:11
segher joined19:11
rseifert joined19:11
flazz left19:11
pantsman left19:12
dramsay joined19:12
ExtraSpice left19:12
ntoll left19:13
LongBeach joined19:13
UnodeUnode|NA19:15
giskard joined19:15
seghersegher_19:15
segher_segher19:15
patrikf Ilari: looks more like an unformed resource locator19:16
dpino left19:16
GarethAdams|Work A uniform awesome locator?19:17
henkboom left19:17
MarkBao left19:19
henkboom joined19:20
Ilari That one talks git:// protocol to abstract namespace unix domain socket...19:21
consolers did you fix pearljams ssh problem? just curious19:22
Ilari consolers: No.19:22
consolers k19:23
magicbronson joined19:23
tatsuya joined19:23
bluenovember_ joined19:23
Paraselene_ joined19:23
cilly left19:24
consolers are unboxed guest additions on windows available somewere?19:25
gghfgd19:25
sebi_` left19:25
bluenovember left19:26
meder left19:27
nadavoid joined19:28
giskard left19:28
gregj left19:28
Jaxan left19:29
DavidKlein left19:30
dramsay left19:30
MarkBao joined19:30
eletuchy left19:31
khmarbaise left19:32
khmarbaise joined19:32
shruggar left19:33
mabrand comments welcome on this way to split whitespace changes into new commit: http://qt.pastebin.com/7qEcDQeG19:33
dramsay joined19:36
mabrand correction: http://pastebin.com/tMzFxpZr19:37
chrislerum left19:37
Unode|NAUnode19:39
curtana left19:40
ustunozgur left19:41
nevyn^_ joined19:41
warthog9 joined19:42
bluenovember_ left19:42
henkboom left19:45
chrislerum joined19:45
pallaert left19:46
sixteneighty left19:47
davr mabrand: I use this: http://stackoverflow.com/questions/1751612/git-tool-to-remove-lines-from-staging-if-they-consist-only-of-changes-in-whitespa/1752872#175287219:47
henkboom joined19:48
ph^ joined19:48
davr Not exactly the same thing, but similar19:48
timyhac left19:51
mabrand I don't see yet how that works with "-w" option.19:52
I mean I don't see yet how it works without -w option.19:52
Titosemi_ joined19:53
timyhac joined19:53
petercoulton joined19:53
Titosemi left19:54
Titosemi_Titosemi19:54
mabrand Now I see it. -b is a lot like -w.19:54
UnodeUnode|NA19:55
xguto left19:55
p3rg2z left19:55
helo_ what is this function commonly referred to: http://upload.wikimedia.org/math/8/a/b/8ab0fabd2fffb1fc999ce7a59d0ccbdd.png19:57
doh, wrong chan19:57
patmaddox joined19:57
p3rg2z joined19:57
radioman-lt left19:57
helo_ left19:58
helo_ joined19:58
IslandUsurper I feel like I should know that one...19:59
radioman-lt joined19:59
bremner entropy of some kind?20:01
pipegeek joined20:01
eletuchy joined20:01
pipegeek left20:01
IslandUsurper bremner, at first I thought it was physics-related, but now I'm thinking it might have to do with probabilities20:01
Unode|NAUnode20:02
henkboom left20:06
paja joined20:06
rubydiam_ left20:06
sie joined20:07
rryouumaa my guess is specifically shannon entropy20:08
henkboom joined20:08
mjf joined20:08
sylr left20:09
rryouumaa hey no fair #emacs just got closed. was there a flooder?20:09
sylr joined20:11
eighty4 joined20:12
WebDragon|away joined20:12
BlindOracle joined20:13
WebDragon left20:13
WebDragon|awayWebDragon20:13
nevyn^_ left20:15
Ilari Yup, Shannon entropy.20:16
Except that b is usually 2 then (rarely also b = e is used).20:17
khaase left20:17
francisfish_awayfrancisfish20:18
MarkBao left20:20
royalty joined20:21
royalty left20:21
apsysoev joined20:22
Paraselene__ joined20:22
bc0906524 left20:22
Paraselene_ left20:24
Paraselene__Paraselene_20:24
mabrand left20:24
redondos left20:26
redondos joined20:27
lispy joined20:27
benjo2 joined20:27
bambule left20:28
keyvan left20:28
timyhac left20:30
ehabkost left20:30
Jferro joined20:32
mattwynne joined20:33
phantomcircuit joined20:33
sie left20:33
rdeweerd joined20:34
timyhac joined20:35
consolers left20:37
skoop joined20:37
Zephyrus left20:38
ali3n0 left20:40
lucsky left20:40
khmarbaise left20:40
sinasalek joined20:43
sinasalek Hi, does anyhone knows whether it's possible to manage branches and tags via Git and the push them to CVS?20:44
justin-george left20:46
justin-george joined20:46
dgrift left20:47
skoop left20:47
justin_george joined20:48
xperience left20:48
Landon joined20:49
Landon does anyone here use gitosis? for some reasin .gitosis.conf wasn't updating even with post-update chmodded to 755, so I just added "cp gitosis.conf /home/git/.gitosis.conf" for now, but is there a better way to fix that?20:50
Ilari Landon: Gitosis is obsolete.20:51
Landon really? what's used now?20:51
rdeweerd left20:51
eletuchy left20:51
justin-george left20:51
eletuchy joined20:51
sinasalek Hi, does anyhone know whether it's possible to manage branches and tags via Git and the push them to CVS?20:52
opalka left20:52
dwmw2dwmw2_gone20:52
markbao_ joined20:53
xiong joined20:53
disappearedng left20:53
kumbayo left20:56
rseifert left20:56
timyhac left20:56
kar8nga left20:57
rox` left20:57
Landon Ilari: what is the replacement then?20:58
thiago_home joined20:58
thiago_home left20:58
thiago_home joined20:58
Ilari Landon: Gitolite20:59
thiago_home left20:59
thiago_home joined20:59
Landon thanks, I'll look into that20:59
mattwynne left21:00
mattwynne joined21:00
frakturfreak_frakturfreak21:01
francisfish left21:01
cannonball left21:02
Mukophage joined21:02
eletuchy_ joined21:02
nighthwk1 left21:03
kinectio1kinections21:04
rado1 joined21:04
mattwynne left21:05
eletuchy left21:05
mattwynne joined21:07
carrus85 left21:08
Bass10 left21:08
timyhac joined21:09
sinasalek left21:09
tpiep left21:09
disappearedng joined21:10
ChUrP left21:11
sagsousuke left21:11
linduxed left21:12
JEEBcz joined21:12
Gitzilla left21:13
Gitzilla joined21:13
JEEB left21:14
ajpiano left21:16
priidu joined21:16
Landon left21:17
timyhac left21:17
timyhac joined21:17
p3rg2z left21:18
eighty4 left21:19
Mukophage left21:20
anjen joined21:22
aresnick joined21:23
Eaven joined21:24
p3rg2z joined21:24
BlueDogArcholite joined21:25
BlueDogArcholite I am looking for some help with a gitosis setup21:25
its returning PTY allocation request failed on channel 021:25
ERROR:gitosis.serve.main:Need SSH_ORIGINAL_COMMAND in environment.21:25
I am not getting much traction on how to solve it21:26
chrislerum left21:26
WebDragon left21:26
wereHamster BlueDogArcholite: switch to gitolite21:26
BlueDogArcholite @wereHamster thanks, will try that21:27
phantomcircuit left21:29
_Vi left21:29
gbacon joined21:30
rryouumaa what's git diff's binary detection algorithm? can't seem to convince it that the file is plain text. using a front end.21:30
raichoo left21:31
thiago_home rryouumaa: you don't need to21:31
pantsman joined21:31
thiago_home rryouumaa: but take a look at man gitattributes21:32
Gitbot rryouumaa: the gitattributes manpage can be found at http://git.or.cz/man/gitattributes21:32
thiago_home rryouumaa: note the part about "diff"21:32
flazz_ left21:32
brosner joined21:32
brosner left21:32
IslandUsurper left21:33
radioman-lt left21:34
BlueDogArcholite left21:35
rryouumaa well, i do bec i want to see the diff. just for thius one file i want to fix it so that it gets recognized. i don't see any funny chars (though i could be mistaken).21:35
UnodeUnode|NA21:38
rryouumaa thanks though. could be useful to tell git about extensions.21:38
tstclair left21:40
Titosemi left21:40
pedrobelo joined21:40
redondos left21:41
redondos joined21:41
lpolzer_ joined21:42
eletuchy_ left21:43
aresnick left21:44
florianb joined21:44
lpolzer left21:46
j0shua left21:48
[Po]lentino left21:49
ph^ left21:49
parndt_awayparndt21:50
Bennid left21:51
mbroeker left21:51
ignatp joined21:52
altrux left21:58
jtaby joined21:59
Tuller joined21:59
drizzd left22:00
spb left22:00
giskard joined22:00
spb joined22:01
LongBeach left22:02
aresnick joined22:02
mbroeker joined22:04
DaveIngram left22:04
jds left22:06
thunkthunk|smoking22:06
mbroeker left22:08
magicbronson left22:09
_Vi joined22:11
DaveIngram joined22:12
berserker left22:12
florianb left22:12
DaveIngram left22:13
berserker joined22:14
LiamH left22:15
carrus85 joined22:15
Aaaarg joined22:15
candybar joined22:16
tizbac left22:20
dbatbold left22:23
hohoho left22:23
markbao_ left22:24
lhz left22:24
DaveIngram joined22:25
tar- left22:25
DaveIngram left22:25
candybar left22:26
dbatbold joined22:26
candybar joined22:26
jeffhung_ left22:28
jeffhung joined22:28
apsysoev left22:31
hasrb joined22:33
ralf_e left22:34
berserker left22:36
berserker joined22:38
dramsay left22:38
gmitrev left22:38
EricInBNE joined22:39
patrikf left22:39
MarkBao joined22:40
ajpiano joined22:40
martinciu joined22:43
coderdad left22:45
hebz0rl left22:46
coderdad joined22:46
p3rg2z left22:46
coderdad_ joined22:47
coderdad_ left22:48
coderdad_ joined22:48
engla|esp left22:48
Tuller_ joined22:50
Tuller left22:51
coderdad left22:51
Tuller_Tuller22:51
thiago_home left22:52
alexisabril left22:53
gbacon left22:53
coderdad_ left22:53
patrikf joined22:53
juanjoc left22:54
meder joined22:55
MattDiPasquale left22:56
meder goal: i want to clone my local git repo on my server. so far: i created a bare remote repo, mkdir ~/blah/git and then cd'd into it, and git --bare init. then on my local i did git remote add origin ssh://server.com/home/user/blah/git and then git push origin master22:57
am i on the right path, somewhat?22:57
giskard left22:58
meder when i invoked the git push origin master cmd it didnt push the changes to my remote, so i'm probably still missing some critical step.22:58
or doing things backwards, which i hope i'm not22:58
tychoish left23:01
pantsman left23:01
tychoish joined23:01
benjo2 left23:03
PerlJam left23:03
Mage joined23:04
jast meder: looks fine so far. did push complain in any way?23:05
meder no23:05
jast what did it say?23:05
meder but i'm not really seeing anything special in the remote dir23:05
that is..23:05
jast what special things did you expect?23:05
meder i'm not seeing those files which i have locally23:05
jast yes, it's a bare repository23:05
that means it has no working tree23:05
meder right23:05
i wanted to clone my local to it23:05
jast it only has the database stuff that stores the history23:05
meder ah.23:06
so would i checkout from remote?23:06
jast you could now clone other repositories from it23:06
meder what methods can i invoke todo that now?23:06
jast well, you can do something like git clone ssh://server.com/and/so/on from anywhere else and see pretty much what you've got locally right now23:06
eletuchy joined23:07
JEEBczJEEB23:08
JEEB left23:08
JEEB joined23:08
martinciu left23:10
PerlJam joined23:10
Yuuhi left23:12
MarkBao left23:12
meder left23:12
ralf_e joined23:13
peper left23:13
airborn left23:14
tsundsted joined23:15
brizly1 joined23:16
ralf_e left23:17
tvw left23:18
Jferro left23:18
brizly left23:18
yashi left23:19
Cesario left23:21
aliceinwire joined23:21
xiong left23:22
peper joined23:22
hobodave left23:23
psoo left23:23
Tuller left23:24
phantomcircuit joined23:25
lispy Does git have a verbose mode? I'm getting an error and I'm having a hard time figuring out why23:27
error: file write error (Input/output error)23:27
fatal: unable to write sha1 file23:27
kipras left23:27
lispy I can write to the disk just fine using other means23:27
valros joined23:28
priidu left23:30
ignatp left23:33
hobodave joined23:33
lispy Nevermind, I found that it's git pull -v not git -v pull, and that git --help is different than git pull --help23:33
hugoxrosa left23:33
lispy Futhermore, I tried my pull a 4th time and it "just worked"23:33
lispy boggle23:33
Cesario joined23:34
waveclaw joined23:35
thunk|smokingthunk23:38
przemoc left23:38
freshtonic joined23:39
Ryback_ left23:39
aziz left23:39
jrmuizel left23:40
Jochen_Hayek left23:40
alezandro left23:42
fcoury joined23:43
fcoury left23:43
berserker left23:45
gcristian_ left23:45
przemoc joined23:46
engla|esp joined23:47
Jochen_Hayek joined23:47
berserker joined23:47
warthog9 left23:50
briahn10 left23:51
ignatp joined23:51
py1hon joined23:51
marcial5 left23:51
JEEB left23:53
gbacon joined23:54
mjf left23:54
lispy left23:55
hacosta left23:56
bcardarella joined23:56
dgs joined23:56
johnw left23:57
dgs is it possible to simply revert a specific commit that impacts only one file? I've googled around, and it seems you can do (eg http://www.kernel.org/pub/software/scm/git/docs/howto/revert-branch-rebase.html) but that seems like a huge deal of work when I could just copy the old line from git diff and paste it into the current file....23:58
aliceinwire left23:58
jmcantrell joined23:58
dgs however, I know if I always just do it, I'll never learn to use git properly....23:58
s/it/that23:58

Logs Search ←Prev date Next date→ Channels Documentation