IRCloggy #git 2011-10-02

Logs Search ←Prev date Next date→ Channels Documentation

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

2011-10-02

nb left00:01
kukks left00:02
cyphase joined00:03
idlemind324 left00:03
tangledhelix joined00:03
the_metalgamer left00:03
kumbayo left00:04
nb joined00:04
kaimi joined00:04
canton7 left00:06
tiesmaster left00:07
ph^ left00:08
cjs joined00:09
CareBear\ irc tv slow :(00:13
so slow that I have 217 lines of javascript in front of me00:13
kvanderw left00:13
jelmer left00:13
JoeyA joined00:14
sikao_lfs joined00:15
EM03_ joined00:15
EM03_ left00:15
EM03_ joined00:15
EM03 left00:16
EM03_EM0300:16
sh_sh left00:16
cakehero left00:16
friskd joined00:16
JoeyA left00:18
friskd left00:19
boombatower joined00:19
xpot-mobile left00:20
kukks joined00:21
whyking joined00:23
digitalfiz left00:23
sikao_lfs left00:25
Francisc joined00:25
cyphase left00:25
j416 left00:25
Guest96140 left00:25
sorin joined00:26
ChanServ set mode: +v00:26
Francisc Evening.00:26
sorinGuest2839100:26
AndChat joined00:26
Francisc Is it possible to remove files from repo? For example I have a local SQLite DB. It might be good to have the last 5 versions for example, but I don't need more than that. It's only data changes, no structural ones, and old data is irelevant in this case.00:27
AndChat left00:27
slackstation joined00:27
Francisc After I make a new commit I want to remove older versions.00:28
Just the data.db file.00:28
wilmoore left00:28
AndChat joined00:28
brian_g left00:28
Guest28391 left00:29
kvanderw joined00:29
wilmoore joined00:31
gbacon Francisc: you may want to run git filter-branch00:32
Francisc Thanks, gbacon but that won't remove old files from repo.00:33
gbacon Francisc: but it's a pretty serious step00:33
Francisc The idea is that old db is useless.00:33
Even if I lose the current db file, the old one is of now help whatsoever.00:33
So there is no point in having it there.00:33
gbacon git filter-branch definitely will remove the old db00:33
Francisc Oh.00:33
slackstation left00:33
rendar left00:33
Francisc Sorry, I'll have to read the manual about filter-branch then as I know it to do something else.00:34
Thank you.00:34
gbacon Francisc: see http://help.github.com/remove-sensitive-data/00:34
Francisc Thanks agian.00:34
CareBear\ Francisc : suggest strongly to never add the database in the first place.00:34
gbacon even if you have to pretend your db is sensitive00:34
shiba_yu36 joined00:34
digitalfiz joined00:34
CareBear\ Francisc : have a schema file instead.00:34
gbacon Francisc: keep in mind this is a disruptive thing to do.. you'll need to coordinate with any collaborators00:35
digitalfiz left00:35
digitalfiz joined00:35
Francisc OK.00:35
CareBear\ Francisc : the database can be generated from the schema so it should never be committed00:35
always write schema by hand00:35
Francisc And keep the schma in the rep?00:35
CareBear\ yes00:35
Francisc Hmm...00:35
CareBear\ the schema is code like any other00:35
Francisc You really are a bear that cares.00:36
:)00:36
gretch left00:36
CareBear\ aw :)00:36
iorec left00:37
Aaaarg joined00:38
shiba_yu36 left00:38
Francisc left00:44
fpcfan left00:44
kurtzhong joined00:45
[intra]lanman left00:47
topeak left00:47
macmartine left00:48
fpcfan joined00:49
shiba_yu36 joined00:49
gretch joined00:50
macmartine joined00:53
cjs What's the canonical way to see what branch is being tracked by a branch? "cat .git/config"?00:53
sythe left00:54
CareBear\ cjs : git config --get branch.name.merge00:58
JoeyA joined00:59
joelpet joined00:59
cjs Thanks.00:59
CareBear\ you're welcome01:01
IanWizard-Cloud left01:01
Jon47 left01:02
slackstation joined01:04
AndChat left01:04
JoeyA Is it possible to checkout an object, but apply its changes to the index so I can look over them?01:06
sitaram cjs: also try 'git branch -v -v', esp if you have more than one remote01:06
JoeyA I suppose I could merge with the source branch, then git reset back to where I started.01:06
SethRobertson JoeyA: Wait, what? Do you want to merge or checkout?01:06
kurtzhong_ joined01:07
SethRobertson JoeyA: And why do you need for it to be in the index when you can `git diff` it in the working directory?01:07
ziggles joined01:07
cjs sitaram: Ah, that was what I was looking for. I'd known it, but forgotten.01:07
berserkr left01:08
sitaram cjs: yeah. I was never a fan of 'git config --get' for interactive use; you get more from 'less .git/config' otherwise01:09
JoeyA I used git rebase -i to remove select changes from my history (in a backup branch). Now I want to ensure that I didn't remove changes I wanted.01:09
dfr|mac joined01:09
JoeyA err, separate branch01:09
sitaram cjs: however, there doesn't seem to be a single command that would print 'remotename/branchname'; too bad01:09
glennpratt joined01:10
SethRobertson JoeyA: so you want to diff specific files between two branches?01:10
aef joined01:10
kurtzhong left01:10
kurtzhong_kurtzhong01:10
cjs sitaram: I seem to get the remote name, e.g.: * master 0a289c6 [origin/master] Inline listen_on function, and re01:10
JoeyA I want to diff the current branch with another branch, and apply select differences to the current branch.01:11
jedahan joined01:11
jedahan left01:11
jedahan joined01:11
jMCg joined01:11
johnkpaul joined01:12
milki git cherry-pick?01:12
sitaram cjs: oh yeah that command does, that's why I gave it to you when I saw the 'git config --get' offered up as a solution. What I meant was "without all the other cruft, and without all the other *branches* as well"01:12
JoeyA The changes aren't specific commits.01:12
SethRobertson JoeyA: `git checkout targetbranch; git merge --no-commit sourcebranch` Hopefully you will get a merge conflict and you can use mergetool to go through the changes. Otherwise you can use `git add -p` to do so (after evicting the changes from the index)01:12
cjs sitaram: Ah. Yes, this was for interactive use, not for scripting, so the -v was what I needed.01:13
JoeyA Thanks01:13
cjs (Well, extra -v; my standard "br" alias is 'git branch -v "$@"'.01:13
SethRobertson JoeyA: Alternately you can just blow the contents of sourcebranch over targetbranch (I can provide commands if necessary) and then use `git add -p` to selectively apply. Note this would not cause a merge record and would make future merges difficult01:14
JoeyA git checkout -p seems to be doing what I want.01:14
But I'll find out soon enough.01:15
aef left01:15
SethRobertson git-checkout -p will also prevent future merges from working, unless you create a fake merge record afterwards01:15
milki why is that?01:16
Praise left01:17
fpcfan left01:19
aef joined01:19
shiba_yu36 left01:20
boombatower left01:20
fpcfan joined01:20
kerrick joined01:21
gusnan left01:22
dreiss left01:22
JoeyA Nevermind, I'll just walk. (Spaceballs reference)01:23
Thanks, though.01:23
CareBear\ We're being jammed.01:25
kurtzhong_ joined01:28
iXeno joined01:29
Vampire0 joined01:30
kurtzhong left01:31
iXeno_ left01:31
kurtzhong_kurtzhong01:31
churp left01:32
Vampire0_ left01:33
kurtzhong_ joined01:35
sbell left01:36
kurtzhong left01:38
kurtzhong_kurtzhong01:38
sythe joined01:39
kurtzhong_ joined01:39
sythe_ joined01:40
shiba_yu36 joined01:41
sythe left01:41
sythe_ left01:41
sythe joined01:42
kurtzhong left01:42
kurtzhong_kurtzhong01:42
ziggles left01:43
shiba_yu36 left01:43
kurtzhong_ joined01:47
mcdebugger left01:49
kurtzhong left01:50
kurtzhong_kurtzhong01:50
r0b0ty joined01:54
Aaaarg_ joined01:55
Aaaarg left01:56
Aaaarg_Aaaarg01:56
shiba_yu36 joined01:56
bitkiller left01:58
krz left02:03
JoeyA left02:03
cher left02:04
r0b0ty left02:05
aef left02:06
kurtzhong_ joined02:08
aef joined02:09
kurtzhong left02:09
kurtzhong_kurtzhong02:09
macabre joined02:09
johnkpaul left02:10
ISF joined02:10
BiggFREE left02:12
mandric joined02:13
tjubaluba_ joined02:14
Ziggy` joined02:15
Ziggy` Anyone here got any experience with EGit + Eclipse Helios Plugin?02:15
kerrick left02:17
Ziggy` left02:19
mattalexx_ joined02:20
Yuuhi` joined02:20
mattalexx_ I can't merge and I can't push. hat to do now? http://pastie.org/262476702:21
apok joined02:21
Yuuhi left02:22
niloc132 mattalexx_: you can try forcing the push02:22
rudi_s mattalexx_: git merge origin/master02:22
shiba_yu36 left02:22
r0b0ty joined02:23
digitalfiz left02:23
rudi_s (Note the / - and normally you shouldn't force a push.)02:23
ziggles joined02:23
niloc132 ah, i misread, thanks rudi_s02:23
shiba_yu36 joined02:24
mandric left02:24
rudi_s np02:24
mattalexx_ Darn I just forced a push02:24
Before you said that02:24
iXeno left02:24
rudi_s mattalexx_: Is it a public repo?02:25
mattalexx_ No02:25
http://pastie.org/262477402:25
rudi_s mattalexx_: Then you "only" lost the remote changes which were not merged.02:25
iXeno joined02:25
rudi_s git status should tell you what's going on.02:26
tangledhelix left02:26
mattalexx_ rudi_s, http://pastie.org/262477802:26
sorin joined02:26
mattalexx_ I'm a relative beginner and I'02:27
m not sure what unmerged paths means02:27
rudi_s mattalexx_: I think that happend after the failed octopus merge (which means mergeing muliple branches at once - was just a typo, the missing /). It means that git tried to merge another branch but failed and it asks you to resolve the conflict by editing the file.02:28
As you've already forced the push there's no need for you to resolve the merge, just unstage the changes as git tells you: git reset HEAD README.md02:29
Forcing a push modifies the upstream repository in a way which (may) causes commits to be lost - because they were not merged.02:29
mattalexx_ That seemed to work, thanks02:30
rudi_s If you're new to git have a look at the git book (readable online):02:30
!book02:30
jast *automatic message* 'Pro Git' is a complete book about git that is available online for free: http://progit.org/book/02:30
mattalexx_ And thanks for the ex[planation02:30
rudi_s np02:30
mattalexx_ I have another question: Why is the refspec "origin/master" instead of "origin master"? I could have sworn I've seen the refspec written in that syntax before02:31
rudi_s mattalexx_: It's just a convention. It means the branch master in the remote origin (which is the default remote when you clone a repository). But there can't be spaces in the name. So it could be origin-master (if you like to - you should stick to the defaults though), but not origin master.02:32
digitalfiz joined02:32
mattalexx_ rudi_s, I see, thanks02:33
rudi_s If you push branches you'll use the other refspec: git push origin master02:33
man git-push for more information.02:34
Which means push master to remote origin.02:34
kurtzhong_ joined02:35
helpshae joined02:37
kurtzhong left02:38
kurtzhong_kurtzhong02:38
nimrod10 joined02:40
nimrod10 left02:40
kurtzhong_ joined02:41
whyking left02:42
kurtzhong left02:42
kurtzhong_kurtzhong02:42
r0b0ty left02:43
nimrod10 joined02:44
nimrod10 left02:44
nimrod10 joined02:44
kurtzhong_ joined02:44
nimrod10 left02:45
haseeb joined02:46
kurtzhong left02:47
nimrod10 joined02:47
kurtzhong_kurtzhong02:47
Octane left02:51
helpshae left02:53
Fandekasp joined02:54
dabd joined02:56
mandric joined02:57
kurtzhong__ joined02:58
kurtzhong left02:59
kurtzhong__kurtzhong02:59
f0i joined03:00
Wizek left03:00
kurtzhong_ joined03:01
kurtzhong__ joined03:04
kurtzhong left03:04
kurtzhong__kurtzhong03:04
tjubaluba_ left03:06
kurtzhong__ joined03:06
kurtzhong_ left03:07
kurtzhong_ joined03:08
xiong left03:08
axl_ joined03:08
d4rk1ink left03:08
kurtzhong left03:09
kurtzhong_kurtzhong03:09
axl_ left03:09
f0i left03:09
kurtzhong__ left03:11
tangledhelix joined03:11
haseeb_ joined03:11
whyking joined03:11
haseeb left03:12
haseeb_ left03:15
xp_prg joined03:15
xp_prg I am new to git but know subversion and cvs, trying to learn some git, having problem cloning a github repository, trying git clone https://github.com/boothead/stargate.git and it says Cannot get remote repository information.03:16
any ideas what I am doing wrong03:16
?03:16
DrNick works for me. try again?03:18
whyking left03:18
haseeb joined03:18
Davey_ joined03:20
Davey_ left03:20
Davey_ joined03:20
xiong joined03:21
xp_prg nope same thing for me03:22
envi joined03:22
xp_prg well i am trying to get the source code if you care to post it for me in a tarball somewhere03:23
niloc132 there should be a tarball link on the github page03:23
https://github.com/boothead/stargate/tarball/master03:23
aef left03:24
niloc132 click 'Downloads' to get the options03:24
aef joined03:26
LongBeach joined03:27
sikao_lfs joined03:28
sikao_lfs left03:28
xp_prg thanks, I am humbled how I can't find these things myself03:31
aef left03:31
luxurymode joined03:31
xp_prg forgive my existence03:31
niloc132 haha03:31
xp_prg: these things can be tricky ;)03:32
seawise left03:32
tangledhelix left03:33
vilhelm_s joined03:33
SpyroNapDragon03:36
Nomikos newbie here. Can I turn a regular checkout into a bare repository by just throwing everything out, then moving everything from the .git dir into it, and renaming it to have a suffix '.git' >03:36
*?03:36
niloc132 there is one other flag that has to be flipped too - but it is most often suggested that instead you make another bare repo, and push --mirror the contents instead03:36
aef joined03:37
niloc132 http://stackoverflow.com/questions/2199897/git-convert-normal-to-bare-repository03:37
jedahan left03:38
kennethreitz left03:39
vilhelm_s How do I undo a "commit -a"? I want to go back to only the working tree being changed.03:41
kevlarman vilhelm_s: git reset HEAD^03:42
niloc132 or just git reset should work...?03:42
z8000 joined03:43
vilhelm_s Thanks03:43
z8000 left03:43
mattalexx_ Git is saying that a file was renamed in a commit, even though the files are not related, even if they have similar contents. It makes more sense as a deletion and an addition, how can I fix that?03:45
Nomikos Can I create a repo on a server that doesn't support git software, by cloning a local repo to it?03:48
niloc132 left03:48
kevlarman mattalexx_: git doesn't keep track of renames or copies03:49
it will consider files renamed or copied when querying history if they are very similar03:49
kurtzhong_ joined03:52
dermoth_dermoth|home03:52
kurtzhong left03:54
kurtzhong_kurtzhong03:54
luxurymode left03:55
towski joined03:55
mattalexx_ kevlarman, Thanks03:56
kurtzhong_ joined03:56
digitalfiz left03:58
Nomikos can I use ssh or sftp to push to a server that doesn't have git software running, as long as it has a ["valid"] .git dir in the target directory?03:58
kurtzhong left04:00
kurtzhong_kurtzhong04:00
kevlarman Nomikos: yup04:00
Nomikos kevlarman: do I just copy my local (created with git init) .git to the remote server?04:01
adamm left04:01
kurtzhong_ joined04:03
antoniojasr joined04:04
kurtzhong__ joined04:06
kurtzhong left04:07
kurtzhong__kurtzhong04:07
d4rk1ink joined04:07
xiong left04:08
mattalexx_ I did :!grep . Now how can I open second file?04:09
kurtzhong_ left04:09
vilhelm_s left04:10
Bass10 left04:10
kurtzhong_ joined04:11
axolx_ left04:11
ymasory left04:14
kurtzhong__ joined04:14
kurtzhong left04:15
kurtzhong__kurtzhong04:15
ymasory joined04:16
kurtzhong_ left04:17
chrisf_ joined04:19
cytrinox_ joined04:19
kurtzhong_ joined04:20
kurtzhong left04:21
kurtzhong_kurtzhong04:21
kukks left04:22
xiong joined04:22
cytrinox left04:22
cytrinox_cytrinox04:22
offby1 mattalexx_: there's probably a #vim channel; ask there04:23
milki kevlarman: how?04:24
i would say no...04:24
kurtzhong_ joined04:26
kevlarman milki: i'm being tired and wrong >.<04:26
kurtzhong left04:28
kurtzhong_kurtzhong04:28
mattalexx_ offby1, Whoops!04:28
Liquid-Silence joined04:29
icwiener left04:29
haseeb left04:30
anzime joined04:30
chrisf_ left04:31
milki heh04:32
Nomikos: so...no...04:32
Nomikos I sorta gathered :-)04:32
milki but well04:33
Nomikos But git can push over (for example) ftp, right?04:33
milki i wonder...04:33
milki tries04:33
milki maybe?04:33
>.>04:33
ive only used over ssh04:33
now what machine do i have that doesnt have git...04:33
kurtzhong_ joined04:33
milki i...dont have one...04:33
>.>04:33
Nomikos heh04:35
kurtzhong left04:36
kurtzhong_kurtzhong04:36
milki i think its entirely possible for git to do everything from the client side though...04:36
i just havent read about it04:36
Nomikos Various docs talk about how git understands different protocols, down to http - or are they only referring to pulling there?04:36
kurtzhong left04:36
milki ive only heard git, ssh, dumb http, and smart http04:36
for protocols04:36
they are transport protocols04:37
jedahan joined04:38
milki so its for communicating with a remote repo04:38
jedahan left04:38
sythe left04:38
Nomikos still trying to wrap my head around all the new concepts..04:38
milki Git natively supports ssh, git, http, https, ftp, ftps, and rsync protocols04:38
cool04:38
so i guess it does04:38
Nomikos I couldn't get it to clone *to* ssh//whatever though04:39
milki you cant clone to04:39
that is true04:39
i usually rsync it over if i cant clone from04:39
Nomikos maybe that falls under "deployment" rather, and I'm expecting it to do things it wasn't made for04:40
anzime left04:40
Nomikos ah ok04:40
milki you can use an access control layer that can auto-create new repos or somethign like that04:40
but git itself wont create new remote repos04:41
with clone at least...04:41
Nomikos "access control layer" ?04:41
dreiss joined04:41
milki like gitolite04:41
*cough cough*04:41
Nomikos googles04:41
cypha joined04:42
cypha i just did "git add ." and "git commit -m "mymsg""04:42
and I just realized I included all the user files04:43
how can I undo what i've done so far?04:43
milki do you want to keep the changes?04:43
cypha changes within files?04:44
milki yes, from before the last commit04:44
cypha oh, yes04:44
I just don't wanna push all these image files that were uploaded04:44
milki you can do a soft reset with git reset HEAD^04:44
or you can git rm those image files04:44
and then git commit --amend04:44
cypha how can I specify all the image files?04:44
there's like 5004:44
milki bash magic04:45
macmartine left04:45
cypha eh, I suck at bash04:45
so git reset HEAD^ will just undo the "git add ." and "git commit" ?04:45
but won't touch the contents of my files?04:45
milki i think so...04:45
:P04:45
pharvey joined04:45
milki it doesnt under the commit, but rather revert your HEAD04:46
when you make a new commit, your new head points to your new commit04:46
and your mistaken commit is just a dangling reference04:46
cypha i'm confused about all that you just said!04:46
lol04:46
milki which might disappear after a git gc04:46
cypha ok, i'm just doing it04:46
milki git works well with pictures04:47
but cant draw picturesi n irc TT04:47
so uh04:47
a - b - c04:47
where c is your mistaken commit right04:47
cypha so how do I not include a directory called "user_files"?04:47
milki git reset HEAD^ will go back to b04:47
and a new commit will then be c'04:47
you can use a .gitignore file04:47
which does simple wildcard matching04:47
cypha and it'll ignore that directory?04:47
as long as I put a .gitignore empty file in there?04:48
milki in the file, you want user_files*04:48
cypha oh04:48
milki git help ignore04:48
!04:48
you can put it at the parent directory of user_files04:48
or at the root, in which you want *user_files*04:48
or something04:48
you will know its working04:49
if you git status04:49
macmartine joined04:49
milki and user_files doesnt show up anymore04:49
texel left04:49
cypha :)04:49
milki tries out the git reset04:49
cypha so don't put it in .git directory?04:49
milki it doesnt belong in a bare repo04:50
it belongs in the repo itself04:50
as a regular committed file04:50
Nomikos the files, or the .gitignore file?04:50
milki the .gitignore04:50
cypha .gitignore04:50
etcetera has anyone tried assembla's free git hosting?04:50
milki read more04:50
mandric left04:50
Nomikos I have more reading to do ...04:50
milki git help ignore04:50
:P04:50
git help is your friend04:50
etcetera looking for free private repos...04:50
milki lol04:51
its pretty easy04:51
but sincei t requires like04:51
hd space04:51
minijupe left04:51
milki most people would not do it for strangers for free04:51
cypha milki, nothing about gitignore in there04:51
milki for "private" ones04:51
eh04:51
mine does -.-04:52
DESCRIPTION A gitignore file specifies intentionally04:52
...04:52
EugeneKay etcetera - free and private are mutually exclusive. You can run gitolite someplace you have a shell account(such as a VPS).04:53
milki but then you are paying for the VPS04:53
cypha i dunno, it's not there in mine04:53
milki :P04:53
etcetera EugeneKay: good point.04:53
Chillance left04:53
etcetera free, managed, git hosting with a web front end provided.04:53
and private.04:53
quadz joined04:53
milki aka github for poor people04:53
etcetera like assembla, unfuddle, or github. except free or trial.04:53
EugeneKay "free" is a difficult business model04:53
milki cypha: dunno...but you can still google it?04:54
Nomikos cypha: try http://help.github.com/ignore-files/04:54
cypha milki, so what'll I put in .gitignore if I want to ignore ./mysite/user_files/* ? (so I want user_files to show up as an empty directory)04:54
etcetera I need to sell a CTO on git. Trial will suffice.04:54
EugeneKay 1 month of a Github account is what, $5?04:54
milki you can put a .gitignore file in . with contents *mysite/user_files*04:54
or one in ./mysite with content user_files*04:55
selckin any server you allready have.04:55
milki git will not track the direcotry at all04:55
EugeneKay etcetera - I think you'd get farther with convincing the boss with SmartGit than GitHub - you probably already have all the fancy stuff github does, like issue tracker and docs.04:55
SmartGit, on the other hand, is hands-down the best git gui I have ever come across.04:55
cypha awesome, thanks milki :)04:55
Nomikos suddenly busy here :-p04:55
EugeneKay WAY more intuitive for n00bs, imo04:55
spatulasnout left04:55
offby1 EugeneKay: I dimly recall playing with that and being bitterly disappointed. Perhaps it's improved04:56
antoniojasr left04:56
EugeneKay offby1 - you have to do some fiddling-about to get the UI "how you like it", but once you do, it's verah nice.04:56
joshteam joined04:56
milki etcetera: http://fi.github.com/trial.html04:57
45 day trial?04:58
EugeneKay Github:FI is overpriced for small business use, IMO04:58
Davey_ left04:58
milki i think so too04:58
but hey, free trial still04:58
!04:58
you can then switch over to private repos :P04:58
EugeneKay Meh. gitolite <304:58
milki yup yup04:59
Davey_ joined04:59
cypha i'm getting a weird error when pushing http://pastebin.com/tmDAAbBX04:59
glennpratt left04:59
milki but etcetera said he needs a good web frontend too04:59
so gitolote+gitweb?04:59
o right04:59
so did you push the bad commit already?04:59
cypha milki, no04:59
milki huh05:00
then your branch is out of sync with remote05:00
cypha well, I want what I'm pushing to win05:00
milki then force push05:00
use the force!05:00
Nomikos push harder! :-p05:00
cypha merge remote changes first?05:00
what remote changes are they even talking about05:01
milki if you dont want to lose stuff05:01
then that would be advisable05:01
dunno05:01
you gotta look at your commit history05:01
of local and remote05:01
Nomikos Do you guys mostly use GUIs or CLI for gitting?05:01
milki and see whats different05:01
cli!05:01
EugeneKay Nomikos - SmartGit <305:01
I occasionally have to open a prompt to do something "funky", but such things would be difficult-to-impossible to GUI anyway05:02
joshteam left05:03
cypha ok, so how do I force my local files to push onto github?05:03
I don't want my local files changed05:03
milki git push -f05:03
you are overwriting history now though05:04
so make sure its what you actually want05:04
Nomikos Hmkay. I'm trying to "learn the basics" using CLI, but wonder if a GUI would not show me them more intuitively05:04
milki gui might have pretty pictures05:04
cypha milki, yeah, that's fine05:04
Nomikos then again, still reading lots..05:04
etcetera oh bummer. I suppose I should say can do push / pull over http(s)05:04
milki i would agree that pretty pictures are nice05:04
cypha i was just nervous i'd lose all my current work05:04
etcetera managing ssh keys is easily the greatest PITA evar.05:04
milki cypha: ah05:04
you will not lose local work05:04
cypha how can I remove all these .pyc files i pushed a while back?05:04
(i've added *.pyc to .gitignore now)05:05
haydenmuhl joined05:05
milki but you are rewriting remote history, so you may lose remote work05:05
EugeneKay Nomikos - IMO, the best way is to read the "for designers" a few times, play around in SmartGit, and then read Pro Git to learn what's going on under the hood.05:05
milki you would need to learn how to use filter-branch05:05
EugeneKay http://hoth.entp.com/output/git_for_designers.html05:05
cypha i just wanted to update really, but whatever, it's done now05:05
milki @cypha05:05
EugeneKay Note: SmartGit is not FOSS, if that matters to you05:05
Nomikos nods05:05
cypha milki, so no easy way to just delete files?05:05
milki nope05:05
well05:05
cypha hm05:06
milki filter-branch is the command05:06
but its not easy05:06
:P05:06
uh05:06
github has something on it05:06
cypha nothing on git is [for me]05:06
Nomikos EugeneKay: I've read the for-designers, that was helpful05:06
EugeneKay Just playing-about on a "test" repo does wonders05:06
milki http://help.github.com/remove-sensitive-data/05:06
cypha: you can try following that05:06
should be easy enough to follow05:07
sorin left05:07
minijupe joined05:07
milki i havent read any git books05:07
maybe i should...05:07
offby1 high-fives EugeneKay05:07
Nomikos ackpfft. I can mount the sftp thinger locally and push to that..05:07
EugeneKay hacks up a lung on offby105:08
cypha milki, not as easy as you think!05:08
offby1 oh, you too?05:08
EugeneKay Bronchitis :-(05:08
milki but github makes...o right05:08
you are using github05:08
:P05:08
Nomikos is it a really bad idea to have a .git dir in your webroot ?05:09
cypha http://help.github.com/remove-sensitive-data/ like i don't even understand what ~/tmp master* is supposed to be05:09
what's master*?05:09
milki i would say its not something you want to do05:09
thats the persons shell prompt05:09
offby1 Nomikos: I don't know. I'd think if your web server refused to serve it, you'd be OK05:10
cypha oh, so master* is just the command?05:10
milki nope05:10
you want to look for the $ on the next line05:10
cypha oh05:10
Nomikos I guess I can just put it one dir up, and live with the other random stuff being included, it's not that much05:10
milki Nomikos: or you .gitignore!05:10
er05:10
use05:10
offby1 cypha: yeah, it's a shell prompt. A particularly unusual one; github should rewrite that page to make it less confusing.05:10
milki indeed05:11
but actually it seems like a pretty useful one05:11
Nomikos hmmm.. well, - oh, yes, of course, I first edit .gitignore /then/ commit the stuff05:11
milki since it tells you your current branch too05:11
offby1 milki: sure05:11
yep05:11
milki didnt think about that one05:11
maybe i should make a git aware prompt05:11
wilmoore left05:11
offby1 Nomikos: I'm going out on a limb here, but: you might find this useful -- http://toroid.org/ams/git-website-howto05:12
cypha ah05:12
Nomikos reads05:12
offby1 milki: there's a "bash-completion" script that comes with git that does 95% of what you need05:12
all you gotta do is hack your .bashrc a little to make use of it05:12
milki ive seen it05:13
cypha ok, i have no idea where I'm actually deleting the files on git05:13
offby1 mua ha ha05:13
milki but i rarely use completion for git commands anyways05:13
its all aliases05:13
Nomikos offby1: I have sftp but no direct ssh access05:13
cypha why can't I just click and delete files!?05:13
milki cypha: ??05:13
click!?!05:13
what is this, windows?05:13
cypha i'm following http://help.github.com/remove-sensitive-data/05:13
milki the filter-branch command will delete the files from history05:14
cypha i don't see where I'm actually deleting *.pyc on github05:14
milki theresa git rm command within the filter branch command05:14
cypha where is "filter-branch" written in that example?05:15
milki last command of the first chunk of shell code05:15
cypha oh05:16
dude, I have no idea what's going on with that05:16
I have half a mind to just delete my current project05:16
and create a new one05:16
shouldn't be this hard to delete files05:16
milki replace Rakefile with *pyc05:16
o05:17
you dont care about history05:17
you just wnat to delete them1?05:17
git rm -.-05:17
cypha git filter-branch --index-filter 'git rm --cached --ignore-unmatch *.pyc' HEAD05:17
Cannot rewrite branch(es) with a dirty working directory.05:17
how dare it call my working directory dirty05:17
milki you can just do git rm *pyc05:17
HAHAHAHAH05:17
Aaaarg left05:18
cypha ok, removed, and then?05:18
git commit -m "updated"?05:18
milki ya05:18
xiong left05:18
cypha beautiful!05:19
so all that other stuff was to preserve history05:19
milki ya05:19
>.>05:19
cypha phew, luckily i'm not doing anything important05:19
thanks for your help milki :)05:19
really appreciate it05:19
milki np05:19
i like helping05:19
Aaaarg joined05:19
milki now if only there were more people to help...05:20
pharvey git subtree seems even more unwieldy than submodules05:20
milki i wanna check out sparse clones05:21
hyperair left05:22
cypha what language(s) r u developing in milki?05:22
milki i mainly do perl05:22
hyperair joined05:22
milki but im currently working on a python project05:22
cypha cool05:22
milki, i've got a good irssi script idea :)05:23
milki not another irssi script05:23
:P05:23
cypha lol05:23
milki and what will it do -.-05:23
cypha ur already falling asleep05:24
:)05:24
jedahan joined05:24
vinod joined05:24
dudicus left05:24
cypha ok, so using a buddy list/notify list, in a channel you can see who from your list is in the channel05:24
milki eh, thats wasting screen real estate -.-05:25
cypha bound to a key to show the list up top or whatever05:25
ohh, not permanent05:25
milki oo05:25
Nomikos hrm, git on ssh-mounted drives fails, "fatal: error when closing sha1 file", seems to be a common error05:25
milki wont most people just try to tab complete their nicks?05:25
and if it doesnt complete, then they arent there...05:25
cypha alllll the nicks in my list?05:25
etcetera ok, so I reinitialized a git repository in a higher dir.05:26
cypha if it just recursively goes through the names list?05:26
etcetera but it's got the *original* repo still in there.05:26
how do I remove the old directory?05:26
cypha i thought it would be pretty easy (Except I don't know perl)05:26
vinod left05:26
milki cypha: should be05:26
you can get a list of nicks in the current channel05:26
cypha yup05:26
milki and then you recurse check against the list05:26
cypha yup05:27
:)05:27
milki etcetera: cant really visualize what the new repo and original repo look like atm05:27
-.-05:27
etcetera ?05:27
milki cypha: i dunno how to temp display it though05:27
so you git inited in a higher dir05:27
etcetera milki: correct.05:27
created a directory, moved the old repo there. did git init and it reinitialized.05:28
milki ah, you wouldnt need to reinit05:28
you can just move the old repo contents to the new dir05:28
is this a bare repo or a regular repo?05:28
kaimi left05:28
etcetera bare?05:28
madewokherd joined05:29
milki uh05:29
git init --bare or git init05:29
originally05:29
etcetera git init.05:29
milki ok, then you just want to move the entire directory to where you want it05:29
etcetera there are no old repo contents in the new dir. they are the original contents.05:29
milki a regular repo also has a hidden .git folder you need to copy over if you didnt05:30
xiong joined05:31
etcetera dunno what you're talking about.05:31
milki me too05:32
uh..05:32
so...you wanted to move the git repo to a higher directory05:32
etcetera I just did mkdir newdir; cd newdir; mv ../olddir/ newdir/olddir05:32
milki how did you "move the old repo"05:32
cypha milki, wget http://www.mimir.ch/mimir/irssi/files/aspell_complete.pl05:32
etcetera git init;05:32
milki hm05:32
cypha useful script05:32
milki where did you do the init05:33
cypha plus it can show/hide :)05:33
etcetera milki: in newdir.05:33
milki cypha: meh, i like my spelling mistakes05:33
cypha http://www.eckrall.co.uk/?page_id=5505:33
haha05:33
etcetera why is git this complicated. why doesnt it know to figure this stuff out for me or at least notify me of ill behavior.05:33
stupid git.05:33
milki what05:33
its expected behaviour05:33
in newdir05:33
you have a untracked oldir and its contants05:33
etcetera in newdir there is a .git05:33
milki right05:33
and in oldir you have .git05:33
etcetera no I don't.05:33
milki because you moved it there05:33
o.O05:34
well, lets start with. whats the problem again?05:34
etcetera when I pushed the repo to remote. the it contains olddir and newdir/olddir in the root.05:35
nore joined05:36
milki having newdir/olddir doesnt make sense to me05:36
newdir isnt in the repo itself05:36
nore hi everybody!05:36
milki moin05:36
what did you do after git init?05:37
etcetera git remote origin master [repo]05:37
git push [repo]05:38
and of course git add .; git commit05:38
in the normal order.05:38
milki that git add shouldnt have added anythign about newdir o.O05:39
nore can I use git with a project divided in two locations? I have a Django project where django files are stored in one location, and static files like photos, javascript, css are stored in other location. Can Git track all my files in different locations as a single project? thanx!05:39
milki nore: you can use submodules05:39
read up on them cuz i dont use them05:40
ymasory left05:40
milki kevlarman: now that i think about it, it doesnt make sense for the remote server not to have git05:40
nore milki, thanx!05:41
gbacon nore: do you mean your working copy is split into possibly different filesystems?05:41
milki kevlarman: since its stupid to reimplement git with ssh commands and stuff05:41
paparent left05:41
milki kevlarman: its all about git packs05:41
nore: o, what do you mean by "locations"05:41
-.-05:41
submodules is good if they are on different servers >.>05:42
etcetera milki: ahh there are some deleted files that are being pushed.05:42
milki eh, where did these deleted files come from o.O05:42
EugeneKay Space.05:42
milki omg05:42
not those files05:43
etcetera dunno, this is my first 10 minuets of git05:43
milki hmm05:43
jedahan left05:43
etcetera it's a bloody black hole wrapped in a black box.05:43
milki i suggest you start from scratch with your local and remote repo05:43
etcetera milki: hah. that's what I did.05:43
milki something wierd happened05:43
but you pushed05:43
dstufft left05:43
milki if you pushed, that means your remote wasnt from scratch05:44
!05:44
kaimi joined05:44
etcetera changes not staged for commit05:44
ReekenX joined05:44
nore gbacon, I mean for websites, you may have different files stored in different places05:45
milki places = folders?05:45
directories?05:45
servers?05:45
saurb joined05:46
milki git can "track them", but I think you are worried about how to serve files from a git repo05:46
nore milki in different directories, but they don't have the same parent directory05:46
etcetera this is ridiculous.05:47
these issues + assemblas crappy git hosting.05:47
gbacon for example, track content in /home/nore/src/project and /var/www/project/html all in the same repo?05:47
etcetera no repo. remote end hung up. ffs.05:47
milki hasnt header of assemblas05:47
milki s/header/heard05:47
nore gbacon, yes :)05:48
milki nore: it wont be pretty for a git repo05:48
nore and they are ONE single project05:48
milki git wants everything in one place05:48
spatulasnout joined05:48
milki this is where you need to split your idea of your development repository and production deployment05:48
BiggFREE joined05:49
milki your repo would have one hieararchy, but your production env has a different one05:49
and you have a way to deploy the files05:49
nore milki, it is not about development and production05:49
milki from repo to production05:49
quadz left05:49
Aaaarg left05:50
milki ok..?05:50
nore gbacon, milki: thanx guys. I just wanted to know if it was possible. I guess the only option to maintain separate repo for each directory05:51
etcetera error: failed to push some refs to [repo]05:51
w--t---f05:51
milki saaa05:51
etcetera gotta love the verbose error messages. love git.05:51
gbacon nore: you could have a symlink that points into the repo05:51
etcetera "just works"05:51
gbacon nore: e.g., ln -s /var/www/project/html /home/nore/src/project/html05:51
but that's hackish05:52
EugeneKay If you don't complain about the errors and actually learn what hey mean, it makes sense.05:52
etcetera E, [2011-10-02T05:51:55.177673 #17970] ERROR -- : no repository found05:52
gbacon better to have a proper deployment solution05:52
EugeneKay But yeah, whining is easier.05:52
milki dunno about the remote service etcetera is using05:52
why not use github!?05:52
tewecske joined05:52
dabd left05:52
nore gbacon: so if I create a symlink in a git-tracked directory, git will track all the files in that symlink?05:53
gbacon no, the other way around05:53
symlink on the deploy side into your repo05:53
git will track symlinks05:54
etcetera EugeneKay: never had these issues with hg.05:54
milki gbacon: track them as regular files?05:54
etcetera <- complainer monkey.05:54
RaceCondition joined05:54
milki eh05:54
it was easy for me to move from hg to git o.O05:54
gbacon but there's also a core.symlinks setting (see git config --help) that you could experiment with05:55
etcetera yup, it's assembla.05:55
gbacon milki: no, track them as symlinks05:55
etcetera fml.05:55
milki symlinsk are regular files >.>05:55
well05:55
i consider them to be files -.-05:55
can it track changes in the dir the symlinks point to?05:56
i think nore just wanted to have an easy way to do work from one dir, but albe to manage files in those other dirs05:56
and the answer to that would be no05:57
since the parents are differently05:57
s/differently/different05:57
etcetera when in doubt, remove .git and start over. can't imagine that being viable after I've done serious work.05:58
#winning.05:59
milki lolz05:59
when in doubt05:59
MimiEA joined05:59
milki git clone again05:59
etcetera yea git doesnt have the notion of indelibility. go figure.05:59
:-P05:59
milki nah, you just need to know how to mess with history06:00
i think commit history was the hardest thing to wrap my head around06:00
etcetera heh, when in doubt. don't worry your history isnt messy if you just clone.06:00
cjs_ joined06:00
milki but now i love to mess with history06:00
saurb left06:02
milki well06:03
ive spent 1.5hrs of my life here...06:03
i think i should go...06:03
cjs left06:03
draaaaaak left06:04
Nomikos we'll miss you!06:04
thanks for the help :)06:04
milki np06:04
dSebastien joined06:04
milki etcetera: good luck?06:04
etcetera milki: cheers.06:04
thanks for lending your ear.06:04
rchavik joined06:08
diverdude left06:09
gbacon left06:10
gretch left06:10
macabre left06:16
RaceCondition left06:16
saurb joined06:18
madewokherd left06:18
quadz joined06:18
mindmorass joined06:20
bartman left06:20
spatulasnout left06:21
cypha left06:21
m4dc0d3r joined06:23
flavius joined06:23
brack9 joined06:23
brack9 Hey. How smart is repo sync?06:24
mindmorass left06:24
brack9 is this even the right place to ask that question?06:24
Raging_Hog joined06:25
DrNick brack9: no, sorry06:26
seawise joined06:27
EM03 left06:31
xiong left06:34
OOPMan joined06:36
aceofspades19 left06:37
bsajsb joined06:37
iband joined06:38
Gitzilla left06:39
squeese left06:40
dudicus joined06:40
mindmorass joined06:41
mattalexx_ left06:42
Gitzilla joined06:42
bartman joined06:45
mindmorass left06:46
drev1 left06:46
ISF left06:46
xiong joined06:48
selckin left06:49
OOPMan left06:51
fisted left06:52
xiong_ joined06:53
Nomikos Woo, I think I have git working on an sshfs mount06:53
I had sftp access but not ssh access. http://dragffy.com/blog/posts/using-sshfs-with-bazaar-bzr-or-git06:54
xiong left06:54
iband left06:54
iband joined06:54
xiong_ left06:58
iband left06:59
minijupe left06:59
fisted joined06:59
xiong joined06:59
iband joined06:59
hakunin joined07:02
talios joined07:06
spatulasnout joined07:06
dexter_e joined07:07
quadz left07:08
macmartine left07:09
ph^ joined07:12
airborn joined07:13
ibandiband|away07:15
towski left07:22
iband|awayiband07:22
EugeneKay Git should work just fine over sshfs. It'll behave just like any other filesystem, just with lots of lag.07:26
mcdebugger joined07:26
balau joined07:26
mitsuhiko left07:29
iband left07:29
iband joined07:30
Nomikos EugeneKay: it worked partially, `git init` ran fine, just not git add foo07:31
adding the workaround=rename parameter to sshfs fixed it07:32
(... I /think/.. haven't tested anything else yet)07:32
nd__ joined07:39
thiago joined07:41
nd___ left07:42
Dave^| joined07:44
flijten joined07:46
Wizek joined07:46
gebi joined07:47
MimiEA left07:49
flavius left07:53
Chib joined07:53
flijten left07:55
NapDragonSpyro07:56
ziggles left07:59
jelmer joined08:01
i42n joined08:03
i42n left08:04
subbyyy left08:05
Joefish joined08:06
flavius joined08:06
na3r joined08:07
krisoijn left08:10
krisoijn joined08:11
krz joined08:11
workmad3 joined08:13
chrisf_ joined08:14
robotmay joined08:18
ibandiband|away08:18
cjs_cjs08:21
pantsman joined08:29
pantsman left08:29
pantsman joined08:29
mitsuhiko joined08:30
mitsuhiko left08:30
RevoOf1 left08:30
Jackneill joined08:36
iband|awayiband08:36
WeirdCat joined08:36
Wizek left08:40
cybersphinx joined08:40
the_cybersphinx left08:44
ReekenX left08:45
ch077179 joined08:45
Liquid-Work joined08:53
Liquid-Silence left08:53
Liquid-WorkLiquid-Silence08:53
solofight joined08:54
BlaDe__ left08:54
Martin_ joined08:54
tty2 left08:55
tty2 joined08:55
Liquid-Silence left08:55
solofight people i understand that git takes snapshot and not just a diff like few other version control systems. Now whats the advantage in this ?08:55
i mean how does it make a difference for an end user ?08:56
does this again come down to atomicity08:56
?08:56
m4dc0d3r_ joined08:57
m4dc0d3r left08:57
chrizz left08:58
ehsan joined08:59
Liquid-Silence joined08:59
mitsuhiko_ joined08:59
thiago solofight: it doesn't make a difference to the end user if the system is working09:03
solofight: a VCS is supposed to reliably reconstruct any point in time09:03
LiohAu joined09:03
thiago how it does that is irrelevant09:03
na3r left09:03
ndim It is conceptually elegant and simple and easy to implement to just take the data, calculate a hash over it, and then be done.09:05
BiggFREE left09:05
solofight thiago: why git avoided makign diff's and took path of taking snapshot ?09:09
talios left09:09
solofight thiago: so its also possible to use other methods like diff and still be able to reconstruct at any point in time ?09:10
mitsuhiko_ left09:10
mitsuhiko_ joined09:10
mitsuhiko_mitsuhiko09:10
solofight am confused ! :( why would somebody do a new thing like taking a snapshot when taking a diff was the usual practice for version control untill there was a big difference09:11
wbednarski joined09:12
haydenmuhl soloflight: I think using snapshots was part of a performance consideration.09:17
jceb joined09:20
wli joined09:20
solofight haydenmuhl: what would be the difference ? in terms of numbers09:21
Joefish left09:21
solofight haydenmuhl: at the end of the day GIT is performning a diff for its operations right ? so why not straight away use diff for storing versions ? why snapshot and then a reference to a snapshot if the file has not been changed .....etc ?09:22
sounds like a round about ! how will it aid performance ! ?09:23
cirwin left09:23
haydenmuhl solofight, storing a snapshot makes checking out a revision much faster.09:23
To recover the contents of a file from a specific revision, you just copy the snapshot. If you store diffs you might have to apply tens or hundreds of diffs to recover the version you want.09:24
diffs versus snapshots have different trade offs, and the git folks decided to go with snapshots.09:25
ktnarf left09:25
solofight haydenmuhl: oh, yes. But what about merging or any other operations which involves multiple commits09:25
haydenmuhl: oh so there is a trade off09:25
shruggar joined09:26
solofight haydenmuhl: where can i get a list of it ? to get a better understanding09:26
?09:26
haydenmuhl A list of what?09:26
wli left09:26
blaenk joined09:26
solofight haydenmuhl: list of factors involved in the tradeoff09:26
haydenmuhl I don't kow.09:26
know09:26
solofight oh09:26
nicoulaj joined09:27
orafu left09:28
orafu joined09:28
seawise left09:29
the_cybersphinx joined09:29
brian_g joined09:29
sitaram solofight: question for you. What should git do if I make 10 commits, alternating between some large content "A" and some other large content "B" for a file "foo". How much data should be stored09:31
cybersphinx left09:32
DarkAR joined09:32
seawise joined09:32
sitaram rather, how much should the repo size increase between commit 2 ("B" stored) and commit 1009:32
rupertgiles joined09:33
sitaram solofight: second, have you ever read the finding that a 'git checkout' is apparently faster than a 'cp -a'? Can you guess why?09:33
solofight: third: do you know any other VCS that will take advantage of similar content files in some unrelated part of the repo (i.e., no connetcion in terms of filenames, but content happens to be very similar). Git will happily make a delta with that unrelated file09:34
solofight assuming the changeset is either going to be in content A or B. The size of repository could be size of A + size of B - size of common items09:34
rupertgiles left09:35
solofight sitaram: as haydenmuhl said because its using snapshot retrival is easy than putting up several patches and reconstructing the file09:35
sitaram solofight: you didn't get my question. I commited file "foo" with large content "A" ("A" is just a symbolic name for that content). Then next commit I change file "foo" to now ocntain large content "B". Then I made 8 more commits, each alternating the *content* of file "foo" between "A" and "B", sort of toggling09:36
solofight sitaram: i am a person new to the concept and trying to learn it better with good understanding. so experimenting with the facts that i learn online09:36
sitaram solofight: so guess...09:37
solofight answer for third question^09:37
dexter_e left09:37
sitaram first one09:38
guess how much extra space repo should take between commit 2 and commit 10 (commit 1, foo="A", commit 2, foo="B"; commit 3, back to "A", etc)09:39
solofight sitaram: for your last question my guess would be it should identifies the similarities between the content of file across commits09:39
sitaram solofight: yes but let's deal with this A/B/A/B...toggling first09:39
psoo joined09:39
Tommy[D]_ left09:40
sitaram should the extra space be (1) almost 0 (2) 4*A + 4*B,(3) 4*(diff(A,B)) + 4*(diff(B,A))09:40
ThiefMaster left09:42
zkandor joined09:42
cher joined09:44
sitaram solofight: there are 2 completely different things consider here. One: the git *object model*, which simply says "git stores entire blobs". If the sum total of sizes of *different* files in your repo's entire history is 4 GB, you can get 4 GB of blobs out of git using various commands.09:45
solofight sitaram: 309:45
sitaram solofight: in git it is (1) almost 009:45
solofight sitaram: actually if the content is same and toggled it should be option 109:45
sitaram solofight: would you have guessed that before this discussion started?09:46
na3r joined09:46
krz left09:46
rendar joined09:47
solofight sitaram: thinking about space conservation yes, thinking about progressive implementation no09:47
sitaram: but you definitely opened up new perspectives for me, thank you very much09:47
sitaram solofight: back to the 4 GB of blobs thing (long message a few lines up)...09:47
that is the object *model*. The *reality* is different09:48
solofight sitaram: i was about to ask09:48
solofight: there are 2 completely different things consider here. One: the git *object model*, which simply says "git stores entire blobs". If the sum total of sizes of *different* files in your repo's entire history is 4 GB, you can get 4 GB of blobs out of git using various commands.09:48
dont understand it sitaram. kindly explain the last part ?09:48
thiago if you have one file of size 10 bytes and another of size 15 bytes (and that's all in the repository), the total size of different files is 25 bytes09:49
solofight you mean to say no extra space is occupied apart from the files contents ? sitaram09:49
thiago "sum of total size of different files" doesn't include what isn't file09:49
sitaram solofight: what thiago said...09:49
thiago there is a little overhead of the trees and commits, but they are negligible. I'd estimate at around 1k per commit, in a very large repository.09:50
Tommy[D] joined09:50
thiago and that's a gross overestimation09:50
sitaram solofight: put another way, checkout each revision, run md5sum on each file, and count unique ones, and total their sizes09:50
marw joined09:50
solofight thiago: is that different in some other version control systems ? i mean foo file with 10kb and bar file with 15 kb will result in more than 25kb of content size ?09:50
marw i have deleted files waiting to be commited, but how do i push that without using -a?09:51
thiago solofight: by information theory, yes09:51
sitaram solofight: actually, what I am getting at is that in git it is likely to far less09:51
thiago if the VCS is working properly, it is capable of restoring any file and every commit cleanly09:51
solofight sitaram: thiago i found that its actually less because of the compression 10kb + 15KB sometimes resulting in less than 25kb09:51
thiago solofight: we're not talking about compressed size09:52
we're talking about information retrievable09:52
if a repository's entire history produces a total sum of different files of 4 GB, you have 4 GB of data09:52
*then* you apply compression09:52
Rvl joined09:53
sitaram actually, delta then compress, no?09:53
thiago I don't know the internals to be honest09:53
coffeejunk left09:53
Rejinka left09:53
ciaranm left09:53
sitaram solofight: I have to go now, but see http://keithp.com/blogs/Repository_Formats_Matter/ and look for "2.7GB"09:53
thiago: neither do I, TBH. I'm just guessing that compress then delta would not be very useful...09:54
psoo left09:55
haydenmuhl I believe git uses deltas when making pack files. Loose objects are just compressed. No?09:56
ThiefMaster joined09:56
Rejinka joined09:56
na3r left09:58
stodan joined09:58
dexter_e joined09:58
Yuuhi` left09:58
alip left09:58
tokkee left09:58
marw left09:59
tjubaluba_ joined09:59
haydenmuhl left09:59
Rejinka left09:59
radarek joined10:01
tokkee joined10:02
Rejinka joined10:02
Yuuhi` joined10:04
cenuij joined10:05
alip joined10:06
ciaranm joined10:06
coffeejunk joined10:08
shruggar left10:08
robotmay left10:10
mrniranjan joined10:10
ibandiband|away10:12
codebeaker joined10:15
jelmer left10:18
ketas-av left10:22
tjubaluba_ left10:22
hyperair left10:24
blaenk left10:24
tjubaluba_ joined10:25
hyperair joined10:25
Davey__ joined10:26
Davey_ left10:26
squentin left10:27
cybersphinx joined10:27
squentin joined10:28
the_metalgamer joined10:28
Shusshu joined10:31
the_cybersphinx left10:31
hyperair left10:32
dexter_e left10:33
codebeaker left10:33
soulcake joined10:35
wbednarski left10:36
caseymcg left10:36
jceb left10:38
Aaaarg joined10:39
soulcake left10:39
BlaDe__ joined10:40
Martin_ left10:40
zkandor left10:48
seivan joined10:51
Heimidal left10:51
krisoijn left10:52
tjubaluba_ left10:52
solofight left10:53
apok left10:58
apok joined10:59
draaaaaak joined11:01
dkannandkannan|away11:05
iband|awayiband11:07
replore joined11:07
Adaptee joined11:09
davegb3 joined11:13
__iron joined11:14
martinjlowm joined11:17
_iron left11:18
Kiall left11:18
cjuner joined11:18
Prasad joined11:19
Kiall joined11:22
hyperair joined11:22
mklappstuhl joined11:22
ibandiband|away11:23
stodan left11:24
soulcake joined11:25
ketas-av joined11:27
Aaaarg left11:27
Aaaarg joined11:28
iband|awayiband11:29
dSebastien left11:35
robotmay joined11:36
pantsman left11:37
dSebastien joined11:38
ottoshmidt joined11:38
yorick I had this idea of using git as an url shortener (with the truncated hashes). only problem is that hashes may become ambiguous in the future, breaking older links11:38
Shu joined11:38
Shu left11:39
ottoshmidt hey, git submodule status lists commits11:40
but how can I change one of those manually?11:40
Shusshu left11:41
cbreak ottoshmidt: cd into it, check out a branch11:42
or a commit directly if you so feel11:42
after that, in the parent, git add the submodule11:42
ottoshmidt cbreak, thanks11:43
cbreak yorick: what's the point? Why not just use a counter?11:44
bauruine joined11:45
w0bni joined11:45
ottoshmidt cbreak, but how do I delete it in the first place11:45
I mean the submodule11:45
cbreak you want to remove a submodule from the repository?11:45
yorick cbreak: so that identical uploads give the same id?11:45
cbreak yorick: hmm... you can do that with an internal full size hash table11:46
nixmaniack joined11:46
ottoshmidt cbreak, remove and than add again.. with a proper commit11:47
*then11:47
yorick cbreak: or I can have git do it for me11:48
Vile joined11:49
ixti joined11:49
cbreak ottoshmidt: why not just change it?11:49
yorick cbreak: also, I plan to extend it into a pastebin later :-)11:51
Rvl left11:52
solofight joined11:53
ottoshmidt cbreak, how?11:53
cbreak ottoshmidt: just like I told you above11:53
ottoshmidt: cd into it, check out a branch11:54
nyuszika7h left11:54
cbreak or a commit directly if you so feel11:54
shiba_yu36 left11:54
cbreak whops, wrong order11:54
check out the commit you want inside the submodule11:54
cd to parent repo, git add repo11:54
commit11:54
Shusshu joined11:54
solofight sitaram: thank you very much for the link http://www.keithp.com/blogs/Repository_Formats_Matter/11:55
sitaram++11:55
sitaram: ++11:55
thiago: ++11:55
nyuszika7h joined11:56
ottoshmidt cbreak, it won't initialize in the first mplace11:56
*place11:56
cbreak solofight: this is #git, not #c++11:57
ottoshmidt: so the folder is empty?11:57
solofight cbreak: so ?11:57
cbreak solofight: no operator++ implemented :)11:57
Shusshu left11:57
solofight :P nice one, how do i give credits to people here11:57
ibandiband|away11:57
ottoshmidt I changed remote url manually and... commits messed up11:57
cbreak usually by thanking them I think11:58
ottoshmidt: where did you change it? in .git/config?11:58
(of the parent)11:58
Prasad left11:58
matrixiumn joined11:58
ottoshmidt cbreak, yes11:58
matrixiumn hello11:59
hello11:59
oops11:59
iband|away left11:59
cbreak ottoshmidt: then you did git submodule init?11:59
and git submodule update?11:59
(for the repo)11:59
mrniranjan left11:59
shiba_yu36 joined11:59
cbreak it should start fetching and so on, so you should have a repository after update11:59
Shusshu joined12:00
DarkAR left12:00
ASIRINeiluj left12:00
ottoshmidt cbreak, yes12:01
on update I get an error12:01
cbreak which error?12:01
ottoshmidt Unable to checkout 'kdjf34o8u38u938ur etc12:01
cbreak hmm.12:01
sounds like the repository does not have the commit you want12:01
thiago ottoshmidt: that doesn't look like a valid anything12:01
cbreak so either you gave it a wrong url, or an url for a different repository12:02
thiago I suggest you type a valid commit instead of hitting the keyboard12:02
cbreak (which does not have the desired commit)12:02
_Qman joined12:02
thiago the SHA-1 is supposed to be unpredictable, but it's not rando12:02
cjs left12:04
asedeno_work joined12:04
DarkAR joined12:04
dabd joined12:07
ottoshmidt thiago, I see your point but isn't relevant here12:09
cbreak, I know that there is no such commit in the url specified by me12:09
that's why i want to indicate the commit myself12:10
cbreak ottoshmidt: that's why I told you to cd into the repo12:10
and git checkout something you want12:10
if you changed the URL, make sure to also change it in .gitmodules in the parent, if you want12:11
ottoshmidt cbreak, you got it vice versa12:11
I want to adapt my submodule to this branch12:11
without changing submodule...12:11
cbreak which part of my suggestion do you not want to follow?12:11
cggaurav joined12:11
ottoshmidt but telling parent that this is the commit you want12:11
cbreak and why?12:11
the parent records a commit in the submodule12:12
asedeno_work left12:12
ottoshmidt cbreak, where does it get that commit in the first place?12:12
cbreak ottoshmidt: you mean where it is recorded?12:12
ottoshmidt you know.. git submodule status12:12
lists them12:12
yes12:12
cbreak in the tree of the parent repository12:12
pulse00 joined12:13
ottoshmidt that's what I am talking about12:13
cbreak if the submodule is in subfolder/submodule12:13
ottoshmidt but I couldn't have found12:13
cbreak then git cat-file -p HEAD:subfolder lists it12:13
and git add subfolder/submodule adds a new commit of the submodule12:13
pulse00 hi there. when adding a submodule, git lists the submodule folder in the "changes to be committed" section as "new file". has anyone an idea why this would happen?12:13
ottoshmidt cbreak, is it in a text file?12:13
cbreak that's why I told you to git add the submodule12:13
ottoshmidt: no12:13
it's a tree object12:13
it's git internal meta data12:13
you can not change it directly12:13
robotmay left12:14
cbreak if you want to change it, I advise you to use git add12:14
ottoshmidt git submodule add?12:15
cbreak no12:15
git add submodulenamehere12:15
pharvey left12:15
BlastHardcheese joined12:15
cbreak for example, if your submodule is called hans/peter12:16
git add hans/peter12:16
cjs joined12:16
ottoshmidt and will it complain if there is already one?12:16
cbreak .... no?12:16
how long have you used git?12:16
ottoshmidt let's try12:16
cbreak git add adds new content12:16
ottoshmidt several months12:16
cbreak you use it all the time, every time you change a file12:16
ottoshmidt but submodules... less then a month12:16
cbreak you have to git add the file to commit the changes12:16
bassliner joined12:16
ottoshmidt yeah i know that12:17
cbreak ottoshmidt: I have to tell you something sad12:17
most of the time when people use submodules, they use them wrong12:17
ottoshmidt meaning?12:18
pharvey joined12:18
WeirdCat left12:18
bassliner hi, i get "error: could not revert 186301e... test hint: after resolving the conflicts, mark the corrected paths hint: with 'git add <paths>' or 'git rm <paths>' hint: and commit the result with 'git commit'" when trying to revert to an earlier revision. any hint what i'm missing? thanks.12:19
cbreak they are for refering to other projects' repositories12:19
not your own12:19
if you use them for your own, then they can become quite inconvenient12:19
ottoshmidt cbreak, that's right but.. I mean my clone of that other project12:19
by "my"12:19
cbreak well, submodules don't care. they will work either way12:19
just be aware that they might become a pain to manage :)12:20
ottoshmidt cbreak, I see12:20
c00kiemon5ter left12:20
canton7 joined12:20
nixmaniack left12:20
ottoshmidt btw in your example above did you mean hans is parent and /peter submodule?12:21
gebi_ joined12:21
cbreak bassliner: creating the revert commit probably gave a merge error12:21
sikao_lfs joined12:21
cbreak ottoshmidt: no12:21
ottoshmidt: I assumed that you are in the parent repository12:21
in its root12:21
pulse00 left12:21
cbreak hans is a folder, peter is the submodule name12:22
ottoshmidt they maybe the same?12:22
cbreak cd hans/peter would bring you into the submodule12:22
ottoshmidt I am in qt5 now which is parent, I guess12:22
qtbase is submodule12:22
which one is hans :)12:23
?12:23
sikao_lfs left12:23
cbreak type git submodule12:23
it should show lines like this: 7d0bbeb8ded56c2d4092b5e331e0cee4da7668a4 package/boost (heads/boost-snapshots)12:24
gebi left12:24
ottoshmidt git submodule12:24
no reaction12:25
cbreak then you don't have any.12:25
ottoshmidt sorry wasn't on parent dir12:26
bassliner cbreak: how would i solve that?12:26
ottoshmidt 3c54bd67dacc2a5d11246213377e804be364a387 qtbase (heads/master)12:26
cbreak bassliner: fix the conflicts, git add, git commit12:26
bassliner cbreak: sorry but i'm relatively new to vcs, so maybe i miss something.12:26
cbreak: thank you12:26
cbreak ottoshmidt: qtbase is the submodule12:26
cd into qtbase, check out the commit you want, cd into ..12:27
git add qtbase12:27
git status will show you what changed12:27
workmad3 left12:27
cbreak git commit -m "Changed Submodule qtbase to some different commit... because it's cooler that way"12:27
frsk joined12:27
Bass10 joined12:29
slonopotamus joined12:31
wbednarski joined12:31
cenuij left12:32
c00kiemon5ter joined12:32
cenuij joined12:33
cenuij left12:33
cenuij joined12:33
godlygeek left12:33
shirt joined12:33
bauruine left12:34
shirt i'm using gitolite with a single user. i lost my private key. i have root access to the server. how can i re-checkout the gitolite-admin repository so i can update the ssh keys?12:35
cbreak shirt: there are some special commands12:36
robotmay joined12:36
matrixiumn left12:36
robotmay left12:36
EugeneKay shirt - you can edit the gitolite users authorized_keys file12:37
cbreak http://sitaramc.github.com/gitolite/doc/ssh-troubleshooting.html#_losing_your_admin_key12:37
EugeneKay Or see that ^12:37
canton7 left12:37
shirt thanks12:38
ketas-li joined12:38
matrixiumn joined12:39
cbreak editing authorized_keys works too, but it's ... less... elegant :)12:40
and more risky12:40
and it will not stick12:40
EugeneKay It sticks long enough to let you clone, commit, push :-p12:40
sikao_lfs joined12:41
cjs left12:41
matrixiumn left12:41
matrixiumn joined12:41
akosikeno joined12:42
akosikeno left12:42
akosikeno joined12:44
davegb3 can i just confirm there's no way (msysGit) that git log <path> can ignore the case of <path> - core.ignorecase is true. I saw a patch in 2010, but doesn't look like it's in the official release12:45
draaaaaak left12:46
Jackneill left12:48
Jackneill joined12:49
EugeneKay I dunno if it matters what msysgit can/not do, Windows is notoriously bad at handling case12:50
Vortex35 left12:50
Vortex35 joined12:51
davegb3 :) that's as maybe, but I have a path that I can't be sure that the case is right, and need to know if it's been touched by a range of commits.12:52
iband joined12:52
stratwine__ joined12:53
nathggns joined12:55
akosikeno left12:56
akosikeno joined12:56
dfr|mac left12:56
pharvey left12:57
CannedCorn left12:57
khmarbaise left12:59
Cromulent joined12:59
khmarbaise joined13:00
Cromulent left13:00
sh_sh joined13:00
Praise joined13:02
canton7 joined13:02
canton7 left13:03
Davey__ left13:05
wbednarski left13:05
sorin joined13:06
matrixiumn can I change commit messages?13:06
lu_zero matrixiumn: git commit --amend13:06
matrixiumn thank you13:06
fortyfoxes joined13:08
fortyfoxes left13:08
whitman joined13:09
shirt cbreak: that gitolite technique won't work, i don't have a gl-admin-push command!13:12
cbreak it works if you have a somewhat recent gitolite13:12
if not, consider upgrading :)13:12
matrixiumn: this is history rewriting, so only do it if you have not yet pushed13:12
the_cybersphinx joined13:13
tangledhelix joined13:13
shirt cbreak: running gitolite 1.5.4 from ubuntu13:14
cbreak that's fairly ancient13:14
you've not updated in over a year13:14
FernandoBasso joined13:15
glennpratt joined13:16
shirt it's the latest version in ubuntu 10.10. ubuntu 11.04 only has gitolite 1.5.713:16
fortyfoxes joined13:17
cybersphinx left13:17
bremner seriously?13:17
debian stable has 1.7.x13:17
dsjkbfal joined13:17
shirt http://packages.ubuntu.com/natty/gitolite13:18
jstemmer joined13:18
shirt weird13:18
bremner oh, gitolite13:18
I was thinking of git, sorry13:18
I'd say gitolite is about as easy to install from source as packages (and I'm a debian developer)13:19
so just clone it and install it.13:19
drdave joined13:20
bsajsb left13:21
maleknet joined13:24
sphex joined13:24
cybersphinx joined13:25
toofar joined13:26
matrixiumnAmaria13:27
fortyfoxes left13:28
Amariamatrixiumn13:29
kumbayo joined13:29
sphex hello. hey, is it possible to have a branch based on a subdirectory of another branch? ex, say branch A has paths like "dir/1", "dir/2", "dir/3", could I have branch B with paths like "1", "2", "3", and still be able to merge easily from branch A's "dir1" files?13:29
the_cybersphinx left13:29
workmad3 joined13:30
fortyfoxes joined13:31
FauxFaux sphex: There's a subtree merge concept, but I've never used it.13:31
dsjkbfal left13:32
bremner sounds a bit backwards from subtree merge. That would be going from B to A, I think13:32
cbreak sphex: that doesn't really make sense13:33
gretch joined13:33
cbreak you could have branch B as subdir merge partner for A13:33
but vice versa... maybe not13:34
whyking joined13:34
fortyfoxes left13:34
fortyfoxes joined13:34
FernandoBasso Can someone please take a look at this error message: http://codepad.org/VYos0IWJ13:35
matrixiumn how can I totally delete a file from a repo also deleting all commits involved with it?13:35
bremner FernandoBasso: did you try #github?13:36
FernandoBasso bremner: Oh, I was not aware of the existence of this channel. Thanks.13:36
sphex cbreak: ok. yeah, that's a bit weird I guess. I'm thinking of doing that because there's a repo with a template for an application in a subdirectory. everything's designed so that this directory is the root of the application's files...13:36
fortyfoxes left13:37
CannedCorn joined13:37
fortyfoxes joined13:37
pharvey joined13:38
godlygeek joined13:39
madewokherd joined13:39
nathggns left13:42
fortyfoxes left13:42
nathggns joined13:42
sitaram shirt: you still here?13:43
sphex cbreak: do you think it would be worth it to try to make branch B with a commit containing A's subdirectory as its root files and A's top commit as the parent commit, and then see what happens with a "subtree" merge?13:43
hrm and I see there are two kinds of subtree merges :/ guess I should try both.13:44
and hopefully find a way to make this the default for merging from A13:44
maleknet left13:44
cbreak you can try what ever you want13:44
but I've always considered any kind of subtree merging messy13:45
fortyfoxes joined13:45
rane_ left13:45
sitaram shirt: in your version of gitolite there is a program called "gl-emergency-addkey"; should work.13:45
tangledhelix left13:45
sphex damn straight I can try whatever I want. but I'd rather not, I'm really lazy. that's why I'm here.. geeez.13:46
rane_ joined13:46
sphex meh :p anyway, thanks cbreak13:47
workmad3 left13:47
FernandoBasso left13:50
shirt sitaram: thanks :) i actually already manually set a bunch of the GL_ environment vars and pushed, and i think it worked13:51
nb left13:51
FernandoBasso joined13:51
stratwine__ left13:52
asedeno_work joined13:53
sitaram ok, but please consider upgrading. 1.5.4 is 14 months old, which is an eternity in gitolite time :)13:54
Guest13264 joined13:54
shirt sitaram: gonna upgrade server to ubuntu 11.04 soon, so i'll get gitolite 1.5.7 :D thanks for gitolite btw13:56
bremner 1.5.7 is not much newer13:56
CareBear\ can I override the sort order of the gitweb repository list?13:56
madewokherd left13:56
CareBear\ I tried creating a $projects_list file listing repos in desired order, but they seem to always be sorted alphabetically by gitweb13:57
gusnan joined13:57
CareBear\ in theory this is a good idea13:57
in practise I would very much like to control the order13:57
FauxFaux Betting that's a one-line patch. :)13:58
CareBear\ probably, but the question is if I actually need to patch, or can just configure13:59
slonopotamus alphabet is good13:59
CareBear\ well, no13:59
it seems good, but it fails14:00
see http://git.libusb.org/14:00
FauxFaux CareBear\: Well, find where the patch would be made, and see if there's a config variable around it.14:00
CareBear\ FauxFaux : indeed14:00
slonopotamus : libusb.git is the primary repo. it is stupid that it is listed last.14:00
$order ||= $default_projects_order;14:01
@projects = sort_projects_list(\@projects, $order);14:01
fortyfoxes left14:03
Guest13264 left14:03
fortyfoxes joined14:03
CareBear\ I basically do want alphabetical, except that - should collate after everything else14:03
here's what the sort function does: @projects = sort {$a->{$oi->{'key'}} cmp $b->{$oi->{'key'}}} @$projlist;14:04
bitkiller joined14:04
CareBear\ not so easy to get what I want I think14:04
dreiss left14:04
the_cybersphinx joined14:05
shruggar joined14:06
tlvb joined14:08
cybersphinx left14:08
tlvb is it possible to merge a specific list of files from another branch?14:08
CareBear\ merge operates on commits, not on files14:08
so I think no14:09
cbreak merge operates on histories14:10
not only individual commits14:10
sitaram tlvb: actually you can, but it's devious, and it's not a true merge in terms of future work14:10
CareBear\ sure but not *contents* of commits is the point14:10
sitaram tlvb: future merges will be much more troublesome14:11
CareBear\ yay trouble \o/14:11
tlvb heh14:11
nathggns left14:12
tlvb what I have is my-old branch created from the old master, and now I have created my-new from a later master, what I need is to get certain, but not all files from my-old into my-new (I guess redoing it with git checkout my-old; git checkout -b my-new; git merge master -s recursive - theirs14:13
bah, enter pressed too soon14:13
nb_ joined14:14
nb_ left14:14
tlvb git checkout my-old; git checkout -b my-new; git merge master -s recursive -X theirs; ...or something like that14:14
niloc132 joined14:16
tlvb meh, I should stop being a lazy bum and just google it... http://stackoverflow.com/questions/449541/how-do-you-merge-selective-files-with-git-merge14:16
airborn left14:16
cbreak well, the answer they give is wrong14:16
sitaram tlvb: if after this you will never use my-old (or develop on it), it's reasonable to do something like this14:16
cbreak the correct answer is: you can't14:16
instead, do what they write, but be aware that it is NOT merging14:17
ehsan left14:17
khmarbaise left14:18
tlvb I used the term merging since I want to bring (parts of) two branches together, the meaning is not neccesarily as strict as exactly what a git merge does14:18
fortyfoxes left14:18
radarek left14:19
radarek joined14:20
Guest20729 joined14:21
[intra]lanman joined14:22
[intra]lanman left14:22
[intra]lanman joined14:22
antoniojasr joined14:22
wilmoore joined14:22
jalama joined14:22
wbednarski joined14:22
tlvb how about "git checkout source_branch <paths>..." as proposed here?: http://goo.gl/rQRI14:22
slonopotamus left14:22
Chillance joined14:23
Guest20729 left14:23
jalama left14:24
FleetFox joined14:24
tlvb ...trying the thing of my own device I posted earlier actually seems to have done what I wanted14:24
sitaram tlvb: you want to use the merge machinery to get the best possible merge for certain files, and after that forget about that branch. I've done such things myself, it's not a big deal if you go slow and careful and explore the --ours and --theirs options of 'git checkout' to fix up files before committing14:24
radarek left14:24
sitaram tlvb: pedantically, you should not do such things, but if you're doing it with your eyes open, and know the future risks, why not...?14:25
saurb left14:25
sitaram tlvb: *key* is to forget about my-old for any future work after this14:25
dexter_e joined14:26
tlvb yeah, it was temporary anyway14:26
FleetFox Hello. I have a framework that i need to keep updated and app based on it. I can one git repo just contain other repos?14:26
cbreak submodules?14:26
or maybe git slave or similar if you don't need strong binding.14:26
SethRobertson gitslave, not git slave14:27
tlvb sitaram: because future merging from it will be difficult?14:27
sitaram tlvb: very14:27
SmartViking left14:27
FleetFox Thanks.14:27
rchavik left14:27
mastro joined14:28
FleetFox left14:28
sitaram tlvb: I would expect conflicts on the files that were *not* picked up from my-old as part of the merge. Can't be sure though, and a lot depends on the actual contents14:30
dSebastien left14:30
eijk joined14:31
tlvb I guess, well there should be no need to merge from it in the future, in fact it was not named my-old, but leo-tmp, which may hint at it's purpose :)14:31
DarkAR left14:32
whyking left14:33
tlvb Thankyou all for your help, it is much appreciated.14:33
wilmoore left14:33
psoo joined14:34
flaguy48 left14:34
nb joined14:35
khmarbaise joined14:35
nbGuest6615414:35
Dave^|| joined14:37
tangledhelix joined14:38
dec joined14:38
thrillERboy joined14:40
thrillERboy Hi, Is cloning from a working copy is okey? or should I always clone from a bare repo?14:41
cbreak you can clone from any repo14:41
but you should only push to a bare repo14:41
so if you clone from a non-bare, you can't push back14:41
sitaram if you never push back, sure. But it's bad for long term use to get into such bad habits14:41
cbreak (you can push to an other one though)14:41
sitaram (you can even push to a non-checked out branch on a non-bare, to be technically accurate)14:42
Dave^| left14:42
sitaram but it's a bad habit; I wouldn't do it "in production" so to speak14:42
fortyfoxes joined14:42
Guest66154 left14:42
Guest66154 joined14:42
shruggar left14:43
cbreak you can push to every branch if you know what you're doing :)14:43
pharvey left14:43
Guest66154nb14:43
cbreak but most people who don't know what they are doing think they are doing something that they are not actually doing14:43
frogonwheels left14:43
cbreak so they are disapointed and moan when their stuff breaks when pushing to non-bare14:43
slackstation left14:43
iband left14:44
sikao_lfs left14:47
cggaurav left14:47
macabre joined14:48
jalama joined14:49
khmarbaise left14:49
flowenol joined14:50
flowenol hi, i know this goes against version control ideas but, is it possible to erase all history from a git repository?14:51
thrillERboy When I clone from non-bare repo and do a commit, I always have my branch ahead of the number of commits I made in the cloned repo.14:51
thiago flowenol: rm -rf .git14:52
thrillERboy: as expected14:52
flowenol thiago and then i can redo init?14:52
thiago flowenol: yes14:52
thrillERboy But, cloning from bare repo doesn't cause this prob. So, I'll stick with cloning from bare repo from now.14:53
cmn flowenol: you can create a new branch w/o any history with git checkout --orphan somenewbranch14:53
the index is kept, so if you do a commit then, you'll get the same tree state, but a new root-commit14:53
thiago thrillERboy: the local clone behaves exactly the same way14:53
flowenol well im using git for a different than usual case, im using it to sync folders between servers14:53
thiago thrillERboy: if you commit, you're 1 commit more ahead than what you were based on14:53
flowenol: always pull14:53
flowenol from time to time im planning to kill the history14:53
matrixiumn how can I totally delete a file from a repo also deleting all commits involved with it?14:54
flowenol thiago yup thats what I do14:54
thiago matrixiumn: man git-filter-branch14:54
matrixiumn: note that everyone will basically have to re-clone14:54
thrillERboy somebody comitted pron :D14:54
sou_ joined14:54
johnkpaul joined14:54
flowenol thiago so if i remove .git, reinit it, on the other side can i rm .git and clone? or should i do something else14:55
matrixiumn how do filter a branch?14:55
thiago flowenol: on the other side, you should rm -rf the entire dir and then reclone14:55
cmn matrixiumn: the man pages has several examples14:55
thiago matrixiumn: man git-filter-branch14:55
berserkr joined14:55
flowenol thiago ok thanks14:55
fortyfoxes left14:56
fortyfoxes joined14:56
matrixiumn don14:56
*don't you mean main?14:56
thrillERboy thanks thiago cbreak sitaram14:57
thiago matrixiumn: who are you asking that question to?14:57
matrixiumn you14:57
kvanderw left14:57
mandric joined14:57
thiago matrixiumn: no, I didn't mean main14:57
whatever it is that you meant14:57
matrixiumn ok14:58
koo6 left14:58
satonio joined14:58
johnkpaul left14:59
satonio hi i did some changes to the code and then i tried git reset and then idk wht happened it seems it was commited and i don't know how can i get back removing all the code i done14:59
xiangfu joined14:59
thiago satonio: reset modifies the working dir14:59
xxthink left15:00
thiago satonio: it discards modifications and restores from either the index or from the commit you selected15:00
satonio hmm thats what i wanted to do15:00
simplechat left15:01
thiago and what's wrong?15:02
satonio but now if i try git checkout15:02
it says Your branch is ahead of 'origin/master' by 2 commits.15:02
thiago if you think you've lost data, stop running commands that modify even more15:02
satonio i want to lose data15:03
godlygeek left15:03
Daniel0108 left15:03
diegoviola joined15:03
thiago satonio: what data do you want to discard?15:03
satonio the changes i made15:03
thiago satonio: those 2 commits?15:03
satonio yes15:03
(i think so at least)15:03
thiago satonio: run git status15:03
satonio: any modification you still have will be discarded too15:04
satonio: are you sure you want to discard? Uncommitted changes cannot be brought back.15:04
satonio no problem15:04
radarek joined15:04
thiago git reset --hard origin/master15:04
unreal left15:04
Daniel0108 joined15:05
eijk_ joined15:05
whyking joined15:06
satonio ty15:06
unreal joined15:06
xiangfu left15:06
cakehero joined15:06
fortyfoxes left15:06
fortyfoxes joined15:07
shirt "Git is a jumble of random nearly-commentless code, full of globals and strange state and not at all clear control flows. On the other hand, it's also much more Unixy than the code I'm used to reading, doing all sorts of tricks like using mmap() instead of read() (because the latter just involves an extra copy, y'know?) and forking. I am simultaneously impressed and terrified of what's likely going on in my kernel."15:07
godlygeek joined15:07
thrillERboy left15:08
Chib left15:08
satonio excellent it seems ok now15:08
cmn it's not random code, but some of the codeflows can be very tricky15:08
eijk left15:09
jalama left15:09
jalama joined15:09
Daniel0108 left15:10
Daniel0108 joined15:10
sitaram shirt: where did that code come from? more important, when was it said?15:10
shirt SicoWork: 2008, http://evan-tech.livejournal.com/254793.html15:10
er, i mean sitaram15:10
canton7 joined15:10
sitaram thanks15:10
FernandoBasso left15:11
tangledhelix left15:11
jast mmap() is hardly a "trick"15:11
any modern operating system has the concept of mapped memory, backed by files or not15:11
DofLeFou joined15:12
xxthink joined15:13
sitaram shirt: a git clone in haskell? great -- take a system written by someone who obsesses over a few milliseconds of extra time on some obscure operation, with a language where they only start to worry if it crosses 20 minutes or so :-)15:13
adamm joined15:14
jast hey, yet another re-implementation of the absolute bare bones?15:14
obiwahn :)15:14
Daniel0108 left15:14
jast once someone re-implements packing and object walking, I'll start being interested15:15
psoo left15:15
sitaram jast: that was 2008. They're still compiling I guess ;)15:16
CareBear\ now be nice15:16
sitaram people forget that even GHC (the Glasgow Haskell Compiler) switched from the Haskell flagship VCS (Darcs?) to git15:16
CareBear\ C is not for everyone15:16
ISF joined15:16
CareBear\ who cares15:16
khmarbaise joined15:17
CareBear\ that's a compiler, not the language users15:17
fortyfoxes left15:17
CareBear\ it is OK to have a different opinion15:17
it is GOOD too!15:17
even15:17
s/too/even/ I mean :)15:17
sitaram cat hitting your enter key in between your words?15:17
thiago for someone who, like me, is used to writing OOP and thread-safe code in C++, git's codebase is surprising in the beginning15:18
EM03 joined15:18
EM03 left15:18
EM03 joined15:18
thiago the use of global variables, for example15:18
but that's just the way the developers working on Git do it15:18
DarkAR joined15:19
bremner sitaram : the main issue with haskell is lack of contributions15:20
Daniel0108 joined15:20
obiwahn C is not for everyone - C is not for every purpose15:20
tangledhelix joined15:21
sitaram obiwahn: I dont use C myself -- cant imagine writing C anymore -- but I dont criticise its use15:21
airborn joined15:21
obiwahn i think c is good as well - but sometimes it is ok to stick to other languages15:22
jast I don't like most C++ code, and I'm not a huge fan of C either15:22
but I hate all languages anyway15:22
thiago jast: even English? :-)15:22
obiwahn mrgrl mgrl15:22
sitaram bremner: sometimes it's just the "we're superior" that gets me. I see a few posts like that (over the years, from various languages, not just Haskell) and after that I start to generalise15:22
ottoshmidt left15:22
jast the problem with English is that it's hard to parse15:22
sitaram *nod*15:23
this is why my documentation changes so often, more than the code!15:23
jast not that I'm going to evangelize Lojban... that's worse, for other reasons15:23
macmartine joined15:23
jast wait until my own programming language hits the market15:26
fortyfoxes joined15:26
jast but don't stop eating until then, lest you starve15:26
I'm still figuring out one of the language's building blocks15:26
probably been going at it for four years now... :)15:27
toabctl joined15:27
drev1 joined15:28
obiwahn hard to parse:) its syntax is not even unambiguous worse it is not context free^^15:29
workmad3 joined15:29
fortyfoxes left15:30
fortyfoxes joined15:30
Liquid-Silence left15:33
cytrinox_ joined15:33
bremner sitaram: yeah, smugness without code is doubly irritating.15:34
sitaram *nod*15:35
jast even worse if it's paired with stupid insults, like this: http://teddziuba.com/2011/10/node-js-is-cancer.html15:35
well, I guess that's one guy you don't have to take seriously anyway15:35
sitaram that guy sounds like an el reg hack -- similar name at least15:35
minijupe joined15:35
ottoshmidt joined15:35
sitaram oh yeah -- same jerk15:37
cytrinox left15:37
cytrinox_cytrinox15:37
sitaram he's the reason I unsubscribed my el reg feed in my rss reader15:37
the rest wasn't compelling enough to compensate15:37
jast I don't even know who that is15:38
sitaram used to love their sense of humour and irreverence, but it's gone down quite a bit15:38
jast I only know Zed Shaw15:38
sitaram jast: "the register" is an online tech rag. Used to be nice...15:38
this guy was one of their correspondents15:38
jast oh, that15:39
sitaram probably still it15:39
jast haven't read anything of theirs in years15:39
fortyfoxes left15:39
Octalot left15:40
sitaram s/it/is15:40
cybersphinx joined15:43
toabctl left15:43
spaceonline joined15:43
godlygeek left15:44
flaguy48 joined15:46
godlygeek joined15:46
Fandekasp left15:46
the_cybersphinx left15:46
workmad3 left15:47
kumbayo left15:47
tangledhelix left15:50
Frankkkkkk joined15:50
dim__ left15:50
the_cybersphinx joined15:51
godlygeek left15:53
ottoshmidt left15:53
brian_g left15:53
OOPMan joined15:54
godlygeek joined15:54
cybersphinx left15:54
brian_g joined15:54
subbyyy joined15:55
macmartine left15:56
xpot-mobile joined15:57
Prasad joined15:57
j416_ joined15:58
j416_j41615:58
vmiklos left15:58
angelsl left16:02
dSebastien joined16:03
dSebastien left16:04
Frankkkkkk left16:04
Shusshu left16:04
ottoshmidt joined16:04
d4rk1ink left16:08
tangledhelix joined16:08
matrixiumn left16:12
solofight left16:12
vmiklos joined16:12
Jon47 joined16:12
solofight joined16:13
kukks joined16:14
matrixiumn joined16:14
matrixiumn left16:15
matrixiumn joined16:15
omf_ joined16:16
psoo joined16:19
ouah joined16:21
godlygeek left16:21
gebi_ left16:22
psoo left16:23
towski joined16:23
macmartine joined16:25
SpyroSpyraway16:25
asimpletest joined16:28
Deesl joined16:29
DarkAR left16:29
shruggar joined16:30
flavius I've put this line in .git/info/exclude but git-status still shows that file as "Changes not staged for commit": .gdbinit16:30
SpyrawaySpyro16:31
replore left16:31
jast flavius: rules for ignored/excluded files only apply to stuff that isn't already tracked16:31
shruggar left16:31
flavius is there any way to do it for this one file, even though it's tracked? I just don't want further modifications I (not others) do on it to be visible to me16:32
solofight left16:32
OOPMan left16:32
rane_ i suppose it's by design that there is no way ignore tracked files, even though you may find it convenient sometimes16:32
sitaram git rm --cached16:33
flavius I don't want to change that file, just to not see it16:33
ottoshmidt left16:33
sitaram check NOTES section in man gitignore16:34
flavius so this means it's not possible?16:34
godlygeek joined16:34
flavius oh, found it. thanks16:35
drizzd joined16:35
jast note that --assume-unchanged is pretty volatile16:35
for example, 'git reset' unsets it16:36
you can do a bunch of tricks involving sparse checkout, but that's a bit more work16:36
the good way to do it is to simply never stage changes to that file :)16:37
flavius I'm not staging it, but git-status still shows it in red under "Changes not staged for commit"16:37
Anyway I guess I'll just live with it16:38
ehsan joined16:38
flavius I can "trick" it anyway, by using gdb's source in this particular case16:39
solofight joined16:39
EricInBNE left16:43
mattalexx joined16:44
mattalexx How do I reverse a commit and then make the reversal a new commit with a new message?16:44
niloc132 mattalexx: git revert <commit ref>16:45
chrizz joined16:45
ottoshmidt joined16:46
mattalexx Is there any way to do that while adding a couple additional files to the commit?16:46
niloc132 err, you can squash it with another commit16:47
j416 then it is technically not a reversal16:47
it's a reversal and an add ..16:47
possible but confusing.16:47
mattalexx j416, Right, I forgot to take that into consideration in my original querstion16:47
traylenator joined16:47
j416 mattalexx: an easy way would be to revert, then amend whatever you like to that. It's a normal commit.16:47
_Qman left16:48
Vile left16:48
mattalexx NP, not worth the time. The commit's small. I just wanted to know if it was possible16:48
matrixiumn left16:49
tjubaluba_ joined16:49
Vile joined16:50
godlygeek left16:51
towski left16:53
replore_ joined16:53
Adaptee left16:53
wwalker joined16:54
nathggns joined16:54
wwalker How can I tell what code I last push'd upstream? git status; git log; count back X commits?16:55
ymasory joined16:55
FauxFaux git checkout -b my_last_push @{u}16:55
Spockz left16:56
Spockz joined16:56
_Qman joined16:57
shruggar joined17:00
mklappstuhl left17:00
satonio left17:01
arvind_khadri joined17:01
codebeaker joined17:01
arvind_khadri Hi, how can I throw away all the changes made and go to a previous state down the lane?17:02
Samoi left17:02
seivan left17:02
m1sc arvind_khadri: reset --hard17:03
arvind_khadri m1sc, thanks :)17:03
arunce_ joined17:03
Samoi joined17:04
Samoi left17:04
Samoi joined17:04
Prasad left17:05
fortyfoxes joined17:05
arunce left17:06
cytrinox_ joined17:06
arvind_khadri m1sc, after doing a reset, how do I solve the fast forward thing?17:07
d4rk1ink joined17:08
codebeaker left17:08
wwalker FauxFaux: Thank you. That's cool. However, I miscommunicated. I just wanted to find the revision hash of the last push so I can diff/review commit messages before I push.17:08
stodan joined17:08
localhost left17:08
FauxFaux wwalker: That's what @{u} is. :)17:08
shurane left17:09
cytrinox left17:10
cytrinox_cytrinox17:10
shirt how can i change the commit message of an old commit?17:10
localhost joined17:10
bremner shirt: assuming you understand the hazards of rewriting history, you can use "git rebase --interactive"17:11
FauxFaux rebase --interactive and reword.17:11
m1sc arvind_khadri: fast forward thing?17:11
fortyfoxes left17:12
arvind_khadri m1sc, I want to go back down the lane, and also want the main repository to go back too. So I reset and commit ?17:12
Pupeno left17:12
bremner arvind_khadri: you should probably not use reset for history you have shared17:12
mklappstuhl joined17:13
tjubaluba_ left17:13
arvind_khadri bremner, ohh, so should I be going with rebase?17:13
bremner arvind_khadri: no, with revert, unless you can force-push17:14
godlygeek joined17:14
wwalker FauxFaux: how can I get a git command to just output @{u}?17:14
FauxFaux rev-parse17:14
wwalker FauxFaux: thank you17:14
arvind_khadri bremner, and after revert how do I make the same available at the repository.17:16
bremner just push it.17:16
revert makes a new commit17:16
mandric left17:16
brian_g left17:18
Akari joined17:18
ajt left17:18
solofight left17:19
arvind_khadri bremner, I get this http://paste.pocoo.org/show/486111/17:19
wwalker so, "git rev-parse @{u}" shows me a git rev hash that is the 3rd one in the "git log" output, yet "git status" says "...Your branch is ahead of 'origin/master' by 6 commits..." These things don't correlate for me. What am I not understanding?17:19
cbeams joined17:19
wwalker @{u} and log say I'm 2 commits ahead. (a link to a page (web or man) describing what I'm misunderstanding would be great)17:20
ajt joined17:20
wereHamster wwalker: git log origin/master..17:21
that should show you six commits17:21
asimpletest left17:22
FauxFaux is confused.17:22
bremner arvind_khadri: and are there in fact changes on the remote that you should merge?17:22
wereHamster my guess is he has merges in the history and log shows a different order than he's expecting17:22
FauxFaux Oh, merges, olol.17:23
cybersphinx joined17:23
arvind_khadri bremner, I merged them and then did a reset17:23
bremner why a reset?17:23
Akari left17:23
brian_g joined17:24
ottoshmidt left17:24
Akari joined17:24
arvind_khadri bremner, aah I was told that if i want to go down the lane, i should use reset --hard17:25
bremner arvind_khadri: are you the only user of this repo?17:25
michael_mp joined17:25
michael_mp hello17:25
arvind_khadri bremner, no17:25
the_cybersphinx left17:26
icwiener joined17:26
bremner arvind_khadri: I guess you should start again. Use the reflog (man git-reflog) to find a state before you started rewriting history.17:26
michael_mp how do I format a patch to across multiple commits, git format-patch -1 sha1..sha2 ?17:26
workmad3 joined17:27
arvind_khadri bremner, and what should I be using? revert?17:27
bremner yes17:27
michael_mp hmmI should try that...17:27
bremner arvind_khadri: how much history do you need to undo?17:27
akosikeno left17:28
arvind_khadri around some 5 commits17:28
wwalker wereHamster: so "git log origin/master" shows me tons of commits, but starts (most recent) with @{u}. That makes sense. That (and "git log") make sense (I thought I was 2 commits ahead of origin/master). But still status says 6...17:28
bremner arvind_khadri: sounds not too bad to revert then.17:29
Akari left17:29
wwalker can I run "git push" so that it tells me what commits it _would_ push but doesn't push them?17:29
preflex_ joined17:31
bremner arvind_khadri: see in particular the --no-commit option17:31
preflex left17:32
arvind_khadri bremner, oh ok :)17:32
preflex_preflex17:32
wwalker ah. a diff of "git log" and "git log origin/master" shows 4 merges (as wereHamster predicted).17:32
voker57 joined17:32
LiohAu left17:32
voker57 can you create a new branch without parent?17:33
FauxFaux voker57: checkout --orphan17:34
goshakkk joined17:35
goshakkk left17:35
voker57 thanks17:35
jakobm joined17:35
ottoshmidt joined17:36
churp joined17:36
workmad3 left17:36
arvind_khadri bremner, just asking, why shouldnt I be going with git-checkout?17:37
bremner arvind_khadri: checkout is about the state of your working tree, not the repo.17:37
jakobm left17:38
arvind_khadri bremner, ohh ok :) I hope am not bugging you. Am new to git, so just asking17:38
goshakkk joined17:38
radarek left17:39
wwalker FauxFaux: wereHamster: Thank you. now I understand my current state.17:39
radarek joined17:39
goshakkk left17:40
toabctl joined17:40
holybit joined17:41
wereHamster wwalker: I asked you to run 'git log origin/master..'17:42
the two dots are very much relevant, not just ellipsis17:42
wwalker Of the 6 commits , I only want to push 2 upstream, is this the right way: "git push origin master:b11eed14701f5d3c7a947bacc1789018873153e6"17:42
wereHamster no17:43
it's <src>:<dst>17:43
churp left17:43
cbeams left17:43
arvind_khadri bremner, so how can I specify the number of reverts?17:44
radarek left17:44
godlygeek left17:44
obiwahn http://book.git-scm.com/4_git_treeishes.html17:45
SoCo_cpp joined17:45
wwalker wereHamster: ah, the .. is very useful, thank you17:46
linusoleander joined17:46
jakobm joined17:46
bremner arvind_khadri: just list the commits you want to revert on the command line. git log --oneline can tell you the hashes17:47
the_cybersphinx joined17:47
arvind_khadri bremner, obiwahn thanks..17:47
SoCo_cpp I have a workflow issue that I can't explain well enough to google it...17:48
... I have an existing repository with a few main directories art and docs. I want to clone or create a git so I can commit my /var/www/myproj as the directory www in my existing repository.17:48
arvind_khadri bremner, so I do it one by one, I could also specify a range, as the man page says?17:48
wwalker so, to push only 2 of the 6 commits that HEAD is ahead of origin/master, I need to create a branch from origin/master, and merge the 2 commits there, and then push?17:49
bremner arvind_khadri: I think a range should work, but I'm not 100% sure17:49
arvind_khadri bremner, oh ok, then I will go doing it one by one.17:49
ajt left17:50
SoCo_cpp ... so my problem is how to make /var/www/myproj be www in the repository. do I have to clone the repo at /var, then comit have /www/myproj in the repo, or can I make it just www.17:50
cybersphinx left17:51
Yuuhi`` joined17:52
radarek joined17:52
babbio joined17:53
arvind_khadri bremner, thanks a lot, for being so patient17:53
babbio left17:53
bremner arvind_khadri: did you get it to work?17:53
Joefish joined17:53
arvind_khadri bremner, yes :)17:53
bremner great.17:54
arvind_khadri is there a way I can just check if two commits are equal?17:54
like the contents of them are same?17:54
bremner git show?17:54
arvind_khadri will read the man pages :)17:55
SoCo_cpp is there a way to clone just a sub directory of a git repo ?17:56
bremner not really17:56
Yuuhi` left17:56
dv310p3r left17:56
SoCo_cpp how can I have just part of a copy of a repository?17:56
_Qman left17:57
FauxFaux With 'cp'.17:57
jargon- left17:58
bremner or git-archive17:59
SoCo_cpp but I want to be able to continue working with just that part of the repository and be able to push changes.17:59
freerabit joined18:00
caseymcg joined18:00
SoCo_cpp I want to *checkout* just the www directory in my repo into my /var/www/myproj folder, and be able to work with that in place.18:00
freerabit left18:00
FauxFaux SoCo_cpp: I want a billion USD to implement that feature, hand them over, please.18:01
SoCo_cpp FauxFaux: So are you saying that is not within the ability of git?18:01
canton7 left18:02
SoCo_cpp For a billion USD, I'd turn around and produce that same workflow in SVN without modification18:02
ThiefMaster the idea of git is that you have everything available locally18:02
wwalker SoCo_cpp: clone the repo into /home somewhere and symbolic link /var/www/myproj to the location?18:02
bremner SoCo_cpp: yes, this is something that is easier in svn18:02
arvind_khadri left18:03
SoCo_cpp ok. thank you. I thought I was looking for the git way to do it. But I think the symbolic link approach should work fine.18:03
wwalker SoCo_cpp: in svn you then have svn poop all over your web tree, don't forget to tell apache to ignore all .svn directories18:04
bremner wwalker: we don't need to have a pissing contest about what is better18:04
we know git is better18:04
arvind_khadri joined18:05
arvind_khadri left18:05
arvind_khadri joined18:05
wwalker SoCo_cpp: with the symbolic link way though, apache knows that you are no longer under /var/www/myproj so you may need to clone a <Directory...> stanza18:05
SoCo_cpp ok, I was considering if this may lead to user/group issues too.18:06
wwalker bremner: ? I told him to use git...18:06
FauxFaux It's still a pissing concept.18:06
bremner SoCo_cpp: you may also want to investigate "sparse checkout"; I have not used it myself18:06
khmarbaise_ joined18:07
bremner http://vmiklos.hu/blog/sparse-checkout-example-in-git-1-718:07
SoCo_cpp ok. thanks bremner18:07
bremner note the difference between sparse _clone_ and sparse _checkout_18:08
wwalker I also told him how to make svn work. Don't know how I made it a pissing contest. Sorry.18:08
khmarbaise left18:09
khmarbaise_khmarbaise18:09
bremner wwalker: ok, my apologies. I thought we were having and "SVN sucks!" moment18:09
cybersphinx joined18:09
SoCo_cpp I need to devote some documentation time to some git workflow papers, im sure of this. I'm only fine as long as I keep things simple.18:10
wwalker bremner: no problem. That wasn't my intention. It's easy to write ambiguous stuff in text where there's no inflection.18:10
jargon- joined18:11
codebeaker joined18:12
SoCo_cpp some googling has lead me into reading about git Submodules...18:12
solofight joined18:13
wwalker SoCo_cpp: I've used both (more svn than git as I'm still learning to use git (and will be for another few years it seems - new cool stuff around every corner in git)). with either one, you have permissions and ownership issues and should hide the .git or .svn dirs from the web client18:13
the_cybersphinx left18:13
solofight people i heard git is a one way street once initiated. anything that you do is just an another addition in git's database is that true ?18:13
SoCo_cpp wwalker point taken about the .git dirs18:13
godlygeek joined18:13
solofight absolutely no way to skip or delete any content from git repository ?18:14
ReekenX joined18:14
j416 solofight: not true18:14
EugeneKay solofight - 99% of stuff is adding to git's history. Stuff like filter-branch can rewrite it, but that's dangerous and invalidates the entire commit tree, which must then be re-pushed.18:14
the_cybersphinx joined18:14
EugeneKay (and re-accepted by all people who have the repo)18:14
solofight oh18:14
thank you j416 EugeneKay18:14
j416 EugeneKay: 'dangerous'?18:14
EugeneKay j416 - comparatively.18:15
j416 I'd rather call it handy :)18:15
bremner j416: in the sense you might have to force-push18:15
j416 it changes history, but that is to be expected18:15
knifed joined18:15
bremner not to a complete newb it isn't18:15
EugeneKay It can also cause monkeys to fly out of your butt. So..... keep it away from the newbs.18:15
bremner scroll back about 30 minutes18:15
j416 w/e18:15
d3vic3 joined18:16
chrisf_ left18:16
Cromulent joined18:18
cybersphinx left18:18
mytec left18:19
khmarbaise left18:22
fortyfoxes joined18:22
voker57 Is it possible to create a branch in repo without any commits?18:22
fortyfoxes left18:23
ymasory left18:23
bremner asked and answered. as they say on court TV18:23
wwalker voker57: http://book.git-scm.com/5_creating_new_empty_branches.html18:23
macabre left18:23
mandric joined18:24
offby1 voker57: no.18:24
bremner oh, sorry, I misread.18:24
MADMADAWAR joined18:24
offby1 hm, if there's a book with a whole chapter on the subject, perhaps my "no" was premature18:25
Daniel0108 left18:25
voker57 yes, this magic works18:25
squeese joined18:25
offby1 I was technically right, but ...18:25
by definition, a branch is a reference to a commit18:26
so if you have no commits at all, you cannot have any branches.18:26
EugeneKay God, I hope people don't actually do it that way.18:26
offby1 we nerds tend to be sticklers for precision, even at the expense of clarity18:26
EugeneKay voker57 - git checkout --orphan <newbranch>18:26
macRover joined18:27
omf_ I find myself needing to convert some cvs repos to git. I want to keep as much historical data as possible. Is git cvsimport or cvs2git the way to go?18:27
EugeneKay Though you need 1.7.2+ for that18:27
macmartine I'm made 8 checkins to a fork I made on github. But I only want to make a pull request back to the main form for 3 of them. How do I do that? I'm in the 'change commits' page, but still can't figure it out/18:28
codebeaker left18:28
voker57 EugeneKay: doesn't work without any commits18:28
bremner voker57: so make an empty commit?18:28
offby1 omf_: dunno ... I suspect there's at least one easy and accurate way, but I don't know which it is18:29
I suppose you could try both :-|18:29
voker57 works18:29
still a workaround though18:30
Cromulent left18:30
starlays joined18:30
starlays left18:30
starlays joined18:30
towski joined18:30
bremner why do you want branches without commits?18:31
macRover left18:31
Daniel0108 joined18:32
drizzd left18:32
d3vic3 left18:32
linusoleander left18:32
the_cybersphinx macmartine: Split it into several branches, and only have the commits you want pulled on one.18:32
pantsman joined18:32
pantsman left18:32
pantsman joined18:32
linusoleander joined18:32
voker57 I'm writing a script to use git to maintain several completely separate histories18:32
macmartine got it. thanks18:32
slonopotamus joined18:34
the_cybersphinx macmartine: Depending on what your changes do and how picky your upstream is, do one branch per feature, and don't mix things together.18:34
macmartine the_cybersphinx: yeah, that's my next question, seems like you'd end up with tons of branches. do you remove them after?18:35
the_cybersphinx Once they're merged you can just delete them, yes.18:36
tangledhelix left18:36
workmad3 joined18:36
macmartine cool, thanks18:37
omf_ offby1, I am trying cvs2git now.18:37
friskd joined18:38
mandric left18:38
gretch left18:39
lanthan joined18:39
adamm left18:39
lanthan left18:40
slonopotamus left18:40
chrisf_ joined18:40
SoCo_cpp left18:41
lanthan joined18:41
macmartine left18:41
SamB joined18:42
ReekenX left18:42
MUILTFN joined18:43
SamB Is there a way to make "git gui" use the --patience diff algorithm? I looked in git-diff(1), git-config(1), and git-gui(1) but didn't find anything helpful :-(18:43
solofight left18:45
flowenol left18:46
berserkr left18:46
sythe joined18:49
sythe left18:49
sythe joined18:49
toabctl left18:49
mikeyutley joined18:50
bitkiller left18:50
seawise left18:50
macabre joined18:50
towski left18:51
arvind_khadri left18:56
mjt left18:57
AndChat| joined18:57
ottoshmidt left18:59
Vile left19:01
MADMADAWAR left19:04
SpyroSpyraway19:04
traylenator left19:05
dexter_e left19:05
MADMADAWAR joined19:05
chrisf_ left19:05
MADMADAWAR left19:05
xp_prg left19:06
slonopotamus joined19:07
kvanderw joined19:09
cirwin joined19:10
macmartine joined19:10
icwiener left19:12
cbeams joined19:13
frsk_ joined19:13
Joefish left19:14
__iron left19:14
frsk left19:14
frsk_frsk19:15
SpyrawaySpyro19:16
workmad3 left19:20
canton7 joined19:20
mandric joined19:23
Nickeeh How do I revert a commit, but store it's changes in my working directory?19:23
so I could change them, and reapply them..19:23
niloc132 Nickeeh: if it was the last commit, you can reset to the one before it19:24
if not, you'll need someone else around here to help :)19:24
fisted_ joined19:24
cmn do you want to change the last commit? the --amend option to commit can do that19:24
Nickeeh niloc132, i've already pushed it and it has been pulled.19:24
so it's been shared19:25
davegb3 git revert --no-commit ?19:25
i think.19:25
cmn are you trying to bisect a regression? there's a tool for that19:25
niloc132 or git revert, then reset to the commit prior to that19:25
Nickeeh cmn, nope, I pushed something wrongly, but it was a time ago19:26
antoniojasr left19:26
Nickeeh and I want to revert that, but not destroy my changes.19:26
davegb3, i'llcheck that out, thanks19:26
offby1 "git revert --no-commit" sure _sounds_ like the right thing19:26
Nickeeh yeah, but it's not19:26
it stashes the changes to *revert* the commit19:27
fisted left19:27
Nickeeh I want to.. simply grab the changes.19:27
slonopotamus left19:27
Nickeeh hmm19:27
ASIRINeiluj joined19:27
Nickeeh So say we have commits A-B-C19:27
B is the one I want to revert19:27
cmn git show some-commit | git apply, maybe?19:27
offby1 "git show SHA > yow.diff"19:27
maybe "git cherry-pick --no-commit"19:27
Nickeeh can't I do "git revert B", "git diff --patch A B"19:28
git apply patch19:28
something like that?19:28
offby1 or even "git cherry-pick ; git commit --amend"19:28
Nickeeh I haven't cherrypicked. :/19:28
niloc132 or git reset <before revert>19:28
offby1 Nickeeh: don't see why not; Try It And See™.19:28
niloc132 that'll keep you at HEAD, but stage up what it looked like before the revert19:28
minijupe left19:28
cmn niloc132: no, the commits have been made public19:28
Nickeeh okay, thanks19:28
niloc132 revert doesnt change history19:28
giallu joined19:28
Nickeeh niloc132, revert.. does..19:28
niloc132 it adds a new commit, that way you can publish that too19:28
Nickeeh well yeah19:29
but you said reset19:29
niloc132 yeah - revert, then reset19:29
diegoviola left19:29
niloc132 reset to get ready to make another commit19:29
Nickeeh hmm19:29
niloc132 A has the old, good code19:29
B has the newer, published, bad code19:29
git revert makes C, we publish that (same content as A)19:29
keelerm joined19:29
niloc132 then we git reset back to B, so we can make D, a variation on B19:30
Nickeeh hmm, will that work? :O19:30
I'll try19:30
fisted_ left19:30
shruggar left19:30
niloc132 A-B-C-workingdir, unttil you commit after the changes made after the reset19:30
lporras joined19:30
niloc132 reset, change until it is a good version of B (say B'), then commit as D19:30
fisted joined19:32
marienz_ joined19:32
grimwood joined19:34
DarkAR joined19:34
mikeyutley left19:35
AndChat| left19:36
stodan left19:36
macmartine left19:36
Liquid-Silence joined19:40
marienz left19:40
mikeyutley joined19:41
squeese left19:41
LongBeach left19:42
marienz_marienz19:43
squeese joined19:45
Davey_ joined19:47
Davey_ left19:47
Davey_ joined19:47
ouah left19:48
Davey_ left19:48
giallu left19:48
toabctl joined19:48
sorin left19:49
cybersphinx joined19:54
sh_sh left19:55
davegb3 left19:55
marienz left19:55
flavius left19:55
the_cybersphinx left19:57
obiwahn mh19:57
ph^ left19:58
wilmoore joined19:58
ph^ joined19:58
ph^ left19:58
ph^ joined19:59
ChanServ set mode: +v19:59
obiwahn would not to be too had to implement a partial read only checkout outside the working dir would it?19:59
sorin joined19:59
keelerm At work, we recently started using git, and we're maintaining multiple versions of the product in one repo. If I'm working on an issue for an older version that will be brought into a later version, I first create a feature branch from the older version. Once I'm ready to bring this issue into the later versions, what is the recommended way to handle this?20:00
obiwahn is that what soco asked for?20:00
Raging_Hog left20:03
marienz joined20:04
MUILTFN left20:05
tangledhelix joined20:06
giallu joined20:07
cbreak obiwahn: man git-archive20:08
jalama left20:09
macabre left20:11
cbreak keelerm: just merge?20:11
MUILTFN joined20:12
kermit left20:13
keelerm cbreak: If portions of the code base have diverged, won't this bring in potentially erroneous commits?20:13
s0ber left20:13
Tommy[D] left20:14
s0ber_ joined20:15
s0ber_s0ber20:15
mklappstuhl left20:16
jalama joined20:16
giallu left20:17
cbreak keelerm: no, only the changes on the branch get merged20:18
git will find the common merge base20:19
shruggar joined20:19
Tommy[D] joined20:19
keelerm cbreak: So to clarify, even if the two branches diverged a couple of weeks ago, and both have had 10 separate commits each, if I branch from v1.0, made commit A and merge into v2.0, it will only bring in commit A?20:21
cbreak it will bring in all commits that it doesn't have already20:22
if you want to make a feature branch for a bug fix that you can merge into two branches, base it on the merge base, or before20:23
so before they diverged20:23
maybe on the commit that introduced the bug20:23
or you could just duplicate the commit with cherry-pick20:23
but that will make history harder to read20:23
j416 left20:23
jimubao joined20:25
thews left20:25
mtwdmn joined20:25
lahwran I have a bunch of users that should have the same access to many repos in gitolite. what is a good way to only declare the access rules once?20:26
keelerm Okay. That makes sense. Thanks for the assistance.20:27
thews joined20:27
starlays left20:29
cbreak lahwran: you can make a list of repositories20:29
@somerepos = foo bar baz pizza20:29
macabre joined20:30
cbreak repo @somerepos \n RW = @theuserlist20:30
kermit joined20:30
dabd left20:30
envi left20:31
mikeyutley left20:33
mikeyutley joined20:33
keelerm left20:34
wbednarski left20:36
lahwran cbreak: ah, cool20:36
cbreak: is it still necessary to declare the repos separately?20:38
zodiak left20:38
Araxia__ joined20:40
cbreak no20:41
there's even an implicit @all group20:41
I use it to give myself all rights for all repositories :)20:41
lahwran so I can just go20:42
@myrepos = repo1 repo2 repo320:42
repo @myrepos \n access stuff20:42
balau left20:42
lahwran and that's all that needs to be in the file?20:42
assuming I want the same access for all repos20:42
cbreak yeah20:43
I only found out about that last week20:43
when I reread the manual :)20:43
it's a useful lecture20:43
wbednarski joined20:43
lahwran awesome20:44
Jackneill left20:44
cbreak apparently you can make groups of users, repositories, or regular expressions20:44
adamm joined20:44
lahwran wait20:44
groups must just expand to their values20:44
Heimidal joined20:44
lahwran so most likely if I did @test = repo asdf20:44
and then I had just @test on a line ...20:45
nicoulaj left20:46
Heimidal left20:50
Deesl left20:51
caseymcg left20:51
kpreid left20:51
sbell joined20:52
javalova joined20:52
Heimidal joined20:53
jakobm left20:53
Deesl joined20:55
towski joined20:56
towski left21:00
kpreid joined21:00
Zeelot left21:05
Zeelot joined21:05
the_cybersphinx joined21:06
macmaN left21:08
javalova left21:08
minijupe joined21:08
kvanderw left21:08
cybersphinx left21:10
Spockz left21:12
numbah left21:13
numbah joined21:14
mastro left21:17
JorgeRuiz left21:18
bitkiller joined21:19
macmaN joined21:20
Error404NotFound left21:22
luxurymode joined21:23
Hexxeh joined21:25
toabctl left21:26
ziggles joined21:26
ziggles left21:26
m4dc0d3r_ left21:27
javalova joined21:28
whitman left21:29
mastro joined21:30
arosen Anyone know why i'm getting this error when I try to push ? http://pastebin.com/0zzamZvG21:31
dim_ joined21:32
FauxFaux arosen: Read the output aloud, then, if it still doesn't make sense, ask again.21:33
caseymcg joined21:33
Dave^|| left21:35
arosen FauxFaux: sorry :/ I followed this http://thelucid.com/2008/12/02/git-setting-up-a-remote-repository-and-doing-an-initial-push/21:36
I'm getting error: src refspec master does not match any. now21:36
FauxFaux What command are you running?21:36
canton7 arosen, you omitted the --bare on the initial git init, by the looks of things21:37
(for the first problem, that is)21:37
arosen I just rm -fr the project on the server and it worked this some21:38
this time*21:38
sorry about that.21:38
thank21:38
s21:38
luxurymode left21:39
TheAnonymous joined21:40
rideh joined21:40
nardev joined21:40
richcollins joined21:41
richcollins When I pulled I got conflicts. I resolved the conflicts and committed. When I pull again after that -- I get the same conflicts again.21:42
How do I tell git that the conflicts are resolved?21:42
thiago are you using pull --rebase?21:42
ouah joined21:42
mcdebugger left21:43
richcollins thiago: No just git pull21:43
thiago richcollins: but you don't have a config setting to turn rebasing on?21:43
dfr|mac joined21:43
richcollins I didn't change any config settings21:44
For the repo or for my global git preferences?21:44
thiago then it's not exactly the same conflict21:44
it just looks like it21:44
you can turn rerere on to see if it helps21:44
richcollins thiago: No one pushed in the mean time21:44
I pulled, committed and immediately pulled again21:44
same lines in the same files21:44
thiago then it should have said "Already up to date"21:45
richcollins OK but that isn't what is happening -- any other suggestions?21:45
thiago something is going on that you're not telling us21:45
maybe you don't realise that it's important21:45
please pastebin the entire session: the commands you typed, the outputs from git21:46
I don't need to know how you fixed the conflict, I know it happened between git pull and git add21:46
macmartine joined21:47
the_metalgamer left21:47
sorin left21:47
shruggar left21:48
richcollins thiago: Worked when I did it from the command line. Must have been the github app tripping it up21:48
thiago: thanks for your help21:48
thiago you were doing a server-side merge?21:49
richcollins I did a pull from a server21:49
Fixed conflicts21:49
thiago using the "git pull" command-line tool?21:49
richcollins committed using the github app21:49
thiago: right: git pull origin master21:49
thiago ok21:49
then you fixed the conflicts21:50
richcollins Not sure what github app does when you hit "commit"21:50
thiago git add, git commit, git pull again?21:50
richcollins yeah that worked21:50
jargon- left21:50
thiago you were running a GUI front-end for Git?21:50
richcollins thiago: Yeah the Github Mac App21:50
JackWinter left21:50
dim_ left21:51
thiago ah21:51
never heard of it. Sounds like you need to report a bug.21:51
richcollins http://mac.github.com/21:51
seawise joined21:51
thiago I don't have a Mac21:51
richcollins looks like I'm out of date -- guess the app doesn't autoupdate21:51
FauxFaux Perhaps you should use a system with package management ZING.21:52
JackWinter joined21:52
SamB it also helps if the packages are reasonably cohesive21:53
e.g. MSI helps little...21:53
alnewkirk joined21:53
richcollins FauxFaux: os x has homebrew21:53
mikeyutley left21:53
FauxFaux Apparently it doesn't work properly.21:53
richcollins There isn't a package for the Github app21:54
FauxFaux Exactly.21:54
richcollins Oh is there one for your OS?21:55
FauxFaux Nope.21:56
richcollins Exactly.21:56
mattalexx left21:56
FauxFaux ZING.21:56
niloc132 ok, you've both established you have the moral high ground, perhaps we can move on...21:57
brack9 left21:57
DofLeFou left21:59
workmad3 joined21:59
mandric left21:59
madewokherd joined22:02
knifed left22:03
fahadsadah left22:07
eijk_ left22:07
fahadsadah joined22:08
shruggar joined22:09
Deesl left22:09
Uranellus joined22:09
sou_ left22:09
Uranellus left22:09
Uranellus joined22:09
ChanServ set mode: +v22:10
fahadsadah left22:10
lporras left22:14
fahadsadah joined22:14
adamm left22:16
ehsan left22:16
tangledhelix left22:16
ph^ left22:18
[intra]lanman left22:19
richcollins left22:19
jakob left22:20
mquin left22:20
tomaw left22:20
jhoepken left22:20
jpjacobs left22:20
RokSteady left22:20
malesca left22:20
Nibato left22:20
zeom left22:20
CareBear\ left22:20
SJS left22:20
awx left22:20
Azrathud left22:20
cehteh left22:20
cxreg left22:20
eboyjr left22:20
kandinski left22:20
mfb left22:20
ruskie left22:20
dcelix left22:20
k-man left22:20
elhadr left22:20
hsb left22:20
eightyeight left22:20
tharkun left22:20
phantomcircuit left22:20
luka12345|wiik left22:20
anildigital_work left22:20
matthiasgorgens1 left22:20
BinGOs left22:20
mattp_ left22:20
bassliner left22:20
nick_h left22:20
BeerSerc left22:20
foocraft left22:20
kd35a left22:20
spb left22:20
andres left22:20
roue left22:20
Marquel left22:20
Gate left22:20
felipe left22:20
Will| left22:20
Blackshark left22:20
guide left22:20
speakman left22:20
Soliton left22:20
Guest17735 left22:20
skyjumper left22:20
harleypig left22:20
cjackson left22:20
Bucciarati left22:20
seanius left22:20
pizza_ left22:20
tabakhase left22:20
twem2 left22:20
larva left22:20
egs left22:20
bjornbjorn left22:20
xrogaan left22:20
biz left22:20
nathggns left22:21
unreal left22:21
nore left22:21
aef left22:21
nimrod10 left22:21
sphex left22:21
jonanin left22:21
ianmcorvidae left22:21
SRabbelier left22:21
sonnym left22:21
Wooga left22:21
milki left22:21
mtx_init left22:21
shadowmaster left22:21
Kicer86 left22:21
alanp left22:21
odin_ left22:21
ouah left22:21
marienz left22:21
squeese left22:21
cirwin left22:21
lanthan left22:21
michael_mp left22:21
jstemmer left22:21
Aaaarg left22:21
ciaranm left22:21
alip left22:21
thiago left22:21
iXeno left22:21
bgerber left22:21
Pacman left22:21
emptyflask left22:21
cbreak left22:21
troyt left22:21
oc80z left22:21
pl0sh left22:21
siprbaum left22:21
fr| left22:21
jabular left22:21
dfamorato left22:21
zoran119 left22:21
jmcantrell left22:21
stamina left22:21
JEEB left22:21
muzone left22:21
Kulrak left22:21
helgikrs left22:21
zorzar left22:21
ew0 left22:21
balor left22:21
AstralStorm left22:21
architekt left22:21
jcarouth left22:21
Mizar-jp left22:21
saurik left22:21
piquadrat left22:21
davidboy left22:21
fujin left22:21
cesario left22:21
tfnico left22:21
eregon left22:21
piti left22:21
Stravy left22:21
majoh left22:21
rook2paw1 left22:21
KRF left22:21
rm-rf left22:21
charon left22:21
ndim left22:21
harv left22:21
theoros left22:21
Gizmokid2005 left22:21
neuro_damage left22:21
heph left22:21
mjsilva left22:21
arnoldB left22:21
rjbs left22:21
thoraxe left22:21
ttvd left22:21
chrizz left22:21
mattp_ joined22:21
friskd left22:21
brian_g left22:21
ouah joined22:22
ouah left22:22
ouah joined22:22
Uranellus hello, I22:22
shruggar left22:23
biz joined22:23
xrogaan joined22:23
bjornbjorn joined22:23
egs joined22:23
larva joined22:23
tabakhase joined22:23
pizza_ joined22:23
seanius joined22:23
Bucciarati joined22:23
cjackson joined22:23
harleypig joined22:23
Guest17735 joined22:23
twem2 joined22:23
Soliton joined22:23
speakman joined22:23
guide joined22:23
Blackshark joined22:23
Gate joined22:23
Will| joined22:23
felipe joined22:23
Marquel joined22:23
skyjumper joined22:23
andres joined22:23
spb joined22:23
roue joined22:23
foocraft joined22:23
BeerSerc joined22:23
nick_h joined22:23
BinGOs joined22:23
matthiasgorgens1 joined22:23
anildigital_work joined22:23
luka12345|wiik joined22:23
phantomcircuit joined22:23
tharkun joined22:23
eightyeight joined22:23
hsb joined22:23
elhadr joined22:23
k-man joined22:23
dcelix joined22:23
mfb joined22:23
kandinski joined22:23
cxreg joined22:23
eboyjr joined22:23
cehteh joined22:23
awx joined22:23
SJS joined22:23
ruskie joined22:23
CareBear\ joined22:23
zeom joined22:23
Nibato joined22:23
malesca joined22:23
RokSteady joined22:23
jhoepken joined22:23
tomaw joined22:23
Azrathud joined22:23
mquin joined22:23
jpjacobs joined22:23
jakob joined22:23
kd35a joined22:23
bassliner joined22:23
friskd joined22:23
marienz joined22:23
squeese joined22:23
cirwin joined22:23
lanthan joined22:23
michael_mp joined22:23
nathggns joined22:23
chrizz joined22:23
unreal joined22:23
sphex joined22:23
jstemmer joined22:23
Aaaarg joined22:23
ciaranm joined22:23
alip joined22:23
thiago joined22:23
nore joined22:23
aef joined22:23
nimrod10 joined22:23
iXeno joined22:23
bgerber joined22:23
Pacman joined22:23
jonanin joined22:23
ianmcorvidae joined22:23
emptyflask joined22:23
SRabbelier joined22:23
cbreak joined22:23
sonnym joined22:23
Wooga joined22:23
milki joined22:23
troyt joined22:23
oc80z joined22:23
pl0sh joined22:23
muzone joined22:23
jabular joined22:23
hubbard.freenode.net set mode: +v22:23
siprbaum joined22:23
fr| joined22:23
dfamorato joined22:23
mtx_init joined22:23
zoran119 joined22:23
shadowmaster joined22:23
odin_ joined22:23
alanp joined22:23
Kicer86 joined22:23
jmcantrell joined22:23
stamina joined22:23
JEEB joined22:23
ttvd joined22:23
rjbs joined22:23
arnoldB joined22:23
mjsilva joined22:23
neuro_damage joined22:23
Gizmokid2005 joined22:23
thoraxe joined22:23
theoros joined22:23
harv joined22:23
ndim joined22:23
charon joined22:23
rook2paw1 joined22:23
majoh joined22:23
Stravy joined22:23
piti joined22:23
eregon joined22:23
tfnico joined22:23
cesario joined22:23
fujin joined22:23
davidboy joined22:23
piquadrat joined22:23
KRF joined22:23
saurik joined22:23
Mizar-jp joined22:23
architekt joined22:23
AstralStorm joined22:23
balor joined22:23
jcarouth joined22:23
ew0 joined22:23
zorzar joined22:23
rm-rf joined22:23
helgikrs joined22:23
heph joined22:23
Kulrak joined22:23
brian_g joined22:23
grimwood left22:23
i42n joined22:23
soulcake left22:26
Uranellus hello, I'm trying to implement an automatic rollrout/update mechanism using git. in order to do so, I'm mirroring a github repo, and trying to create local clones of that local mirror .. I've tried lay that out using https://gist.github.com/125804522:26
nore left22:28
bburhans left22:28
Uranellus now my problem is, that if i run the update process, in step 1 it creates a new pack and fetches some stuff, but after all if i run git log in the local mirror, it does not show any of the new commits. if i try to pull of that local mirror in any of the copies, it doesnt do anything either. no changes are pulled.22:28
bburhans joined22:28
FauxFaux git branch -avv # in every repo and stare at the output.22:29
i42n left22:33
mandric joined22:33
Uranellus in the local mirror the remotes/origin/* are correct (as in the newest commits are shown), but the local branches, still point to an old commit. FauxFaux, a hint how to fix that?22:34
kennethreitz joined22:34
baburdick left22:34
khaije|minder joined22:34
EricInBNE joined22:34
ymasory joined22:35
baburdick joined22:35
FauxFaux No idea!22:35
canton7 Uranellus, you've pretty much answered your own question.... you can either merge the remote branches into your local ones (requires some checking-out magic), or do a bit of resetting methinks. or you might be able to do some refspec magic on the fetch22:36
niloc132 left22:36
workmad3 left22:36
Uranellus FauxFaux, thanks anyway, helped me narrow it down :)22:36
TheAnonymous left22:37
tjubaluba_ joined22:37
bigkm joined22:38
Amorphous left22:38
kennethreitz left22:38
Uranellus canton7, `git remote update' just did the trick .. guess that it does what `git fetch' already did, but also updated all refs..22:39
canton7 Uranellus, double-check that. git won't update the local branches without explicit instruction22:40
kennethreitz joined22:40
p3rror joined22:41
kvanderw joined22:41
canton7 wonders if git fetch origin refs/heads/*:refs/heads/* will do what you want, or even if it's sane...22:41
thiago left22:43
jimubao_ joined22:43
replore_ left22:44
jimubao left22:45
thews left22:45
thews joined22:46
Uranellus canton7, just did, but wouldnt mind if you had a short look .. https://gist.github.com/1258045#file_output22:46
pay special notice to the MOODLE_21_STABLE branch .. before: 4b85dc1..3973662 (local..remote) .. after both are at 397366222:47
knifed joined22:47
Tommy[D] left22:48
kennethreitz left22:49
simplechat joined22:50
angelsl joined22:50
jmah left22:51
tewecske left22:51
canton7 Uranellus, aah of course, forgot that --mirror did that. git fetch --all might do the same as that remote update22:51
tshauck joined22:52
Saturn28881 joined22:54
wilmoore left22:54
ouah left22:54
antoniojasr joined22:54
antoniojasr left22:54
Uranellus canton7, very well :)22:55
wilmoore joined22:55
drdave left22:55
berserkr joined22:56
Uranellus another related question: do I have to take any precaution for that '--local' clone?22:56
MimiEA joined22:56
Praise left22:57
Amorphous joined22:58
canton7 Uranellus, other than the fact that it's redundant (relative path implies --local), none that I know of.22:58
just be aware of the hardlinking that's going on22:58
Tommy[D] joined22:59
CannedCorn left22:59
nazgul101 joined23:00
caseymcg left23:01
nazgul102 left23:01
DarkAR left23:03
Uranellus alright, thanks very much and have a nice day :)23:03
nardev left23:04
nardev joined23:06
rideh left23:07
antoniojasr joined23:08
alberto56 joined23:09
toabctl joined23:10
frogonwheels joined23:10
airborn left23:10
canton7 left23:11
johnkpaul joined23:11
archis joined23:13
johnkpaul left23:13
DarkAR joined23:14
toabctl left23:14
kermit left23:15
DarkAR left23:15
niloc132 joined23:15
archis left23:16
canton7 joined23:17
pantsman left23:19
Jon47 left23:19
ixti left23:21
xpot-mobile left23:21
Davey_ joined23:21
toabctl joined23:21
phinze left23:21
jelmer joined23:22
tiffany joined23:24
canton7 left23:25
tjubaluba_ left23:26
tiffany I've been searching around about this problem I'm currently having and have found several threads asking but no replies... For some reason, whenever someone tries to clone a git repository from my server over http, it results in an empty git repository. Cloning through git daemon and ssh work, but http does not. I've run git update-server-info, and it didn't seem to do anything.23:27
phinze joined23:27
toabctl left23:28
Davey_ left23:32
cakehero left23:32
cbreak tiffany: where did you run it?23:33
tiffany where I'm pulling from?23:34
cbreak no23:34
where did you run update-server-info?23:34
toabctl joined23:34
tiffany oh, in the directory for the bare repository :s23:34
Saturn28881 left23:35
kermit joined23:35
cbreak do you have a folder info/refs inside it?23:35
javalova left23:36
cbreak file :)23:36
toabctl left23:39
javalova joined23:39
SpyroSpyraway23:40
tiffany # file info/refs23:41
info/refs: empty23:41
cbreak it's not empty here23:41
you sure update-server-info worked?23:42
seawise left23:42
w0bni left23:42
javalova left23:43
caseymcg joined23:43
macmartine left23:44
Uranellus left23:45
kennethreitz joined23:47
gretch joined23:47
cirwin left23:47
_codeman joined23:47
seawise joined23:48
nixmaniack joined23:49
tiffany I ran it again and it made it non-empty23:49
tshauck left23:49
tiffany huh, that worked :|23:50
lacrymology left23:51
intchanter left23:53
intchanter joined23:53
johnkpaul joined23:55
toabctl joined23:57
cirwin joined23:59

Logs Search ←Prev date Next date→ Channels Documentation