IRCloggy #git 2010-11-03

Logs Search ←Prev date Next date→ Channels Documentation

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

2010-11-03

← Xeross left00:00
← dpino left00:00
← thierryp left00:01
← wsclock left00:01
fr0sty "source foo" plays the file 'foo' in your current shell (so the exports persist) try it with foo = "export MESSAGE=HELLO" && echo $MESSAGE00:01
wagle: ^00:01
assuming the heroku repository is added as a remote.00:02
← chrischambers left00:02
fr0sty JohnDav: 'git push heroku 1.0'00:02
← f0rkit left00:02
← goshawk left00:03
→ Tuller joined00:05
→ jrmuizel joined00:05
→ Blackguard joined00:05
← unhygienix left00:05
← amerine left00:06
← ijuma_ left00:06
← mtkd_ left00:07
← eletuchy left00:07
← LiamH left00:07
→ mtkd_ joined00:07
→ ijuma_ joined00:09
→ chrischambers joined00:09
→ hobodave joined00:10
→ gagzilla joined00:10
← hakunin left00:10
← gagzilla left00:10
→ hakunin joined00:11
← kenichi left00:12
← bdiu left00:12
→ Guest70226 joined00:13
← tvw left00:13
← Guest70226 left00:14
→ thierryp joined00:15
← ijuma_ left00:17
→ ijuma_ joined00:18
← brntbeer_ left00:19
→ rasztasd joined00:20
← clcto left00:20
← rasztasd left00:21
← MikhX left00:21
→ MikhX joined00:21
← jepler left00:21
→ abetaha joined00:22
→ dylanb joined00:23
← mtkd_ left00:23
← pedrobelo left00:23
→ mtkd_ joined00:23
→ pedrobelo joined00:24
→ jepler joined00:24
← coderdad left00:25
→ Heimidal joined00:26
→ thiago joined00:26
→ dreiss joined00:28
← mtkd_ left00:29
→ unhygienix joined00:30
← aah left00:30
→ bdiu joined00:32
← keystr0k left00:32
← wuj left00:33
← jrmuizel left00:33
→ juliohm joined00:34
← hasrb left00:37
→ kingfishr_ joined00:40
→ dduncan joined00:41
← odin_ left00:41
← dduncan left00:41
→ ki__ joined00:42
← mlq left00:43
← TeckniX left00:44
← Karen_m left00:44
← Adaptee left00:45
→ scottwd9 joined00:46
dylanb Evening... any Git gurus about?00:46
→ jmil joined00:46
thiago don't ask for someone, just ask your question00:46
dylanb There are two versions of a work-in-progress cheat-sheet at http://www.dylanbeattie.net/cheatsheets/git/00:47
But since I've put this together as I've learned my way around it00:47
I don't know whether it's (a) helpful and (b) correct00:47
and would appreciate some feedback.00:47
One version has speech-balloon annotations, the other doesn't - I don't know if they help or hinder.00:48
Any feedback or comments would be gratefully recevied :)00:48
← Marquel left00:48
← bhosie left00:49
jast I think it's great to have both00:49
→ MikhX_ joined00:50
← s0enke left00:50
→ ki___ joined00:50
jast might help to label the leftmost column (working tree/index, and by the way, you don't distinguish those two)00:51
← rendar left00:51
← ki__ left00:51
dylanb You mean - I fail to do so on my chart, or one customarily does not, distinguish those two?00:51
column labels are right at the bottom00:52
have they got lost down there, or do they need to be up top?00:52
fr0sty yeah, headers, not footers...00:53
jast only in the speech bubbles, though, aren't they?00:53
you don't distinguish. we generally do.00:53
← MikhX left00:53
MikhX_ → MikhX00:53
Gecko1986 could anyone explain me what this guy means with "leave a branch pointing at the most recent commit, so i know which commits to move"00:53
jast it's not *necessary* to care about the index, but it makes certain things much easier if you know that it's there00:53
Gecko1986 ? P00:53
← ki___ left00:53
jast Gecko1986: a bit more context perhaps?00:54
dylanb got a link to docs/something that'll explain the index to a newbie?00:54
→ ki__ joined00:54
fr0sty dylanb: also, mention the word 'merge' in conjunction with the 'pull' commands.00:54
or maybe have one of the pulls be an explicit fetch+merge00:54
dylanb A pull is effectively a one-step fetch-then-merge, right?00:54
Second pull is an explicit fetch-merge00:54
left hand side, after the pull request.00:55
← plediii left00:55
dylanb You think I should switch them - fetch/merge first, and pull later?00:55
fr0sty ah, right. glossed over that.00:55
→ Marquel joined00:55
fr0sty not necessary. and definitely keep the speech bubbles.00:55
Gecko1986: have a branch in your repository that points to the most recent commits. I'm not sure why you _wouldn't_ already have a branch there, though.00:57
← juliohm left00:57
← _nil left00:57
→ s0enke joined00:58
← chrischambers left00:58
→ QinGW joined00:59
← bdiu left00:59
frogonwheels dylanb: cute concept. don't shade out the column headers of the repo!00:59
dylanb: the bubbles are great!00:59
dylanb OK - label columns at the top (headers) instead of at the bottom; read up on working tree/index and show it on the chart...00:59
← orafu left01:00
→ plediii joined01:00
dylanb keep the speech bubbles...01:00
any more ?01:00
→ daveb joined01:00
daveb is there a way to tell the URL of a repos from a checkout/clone?01:00
frogonwheels git remove -v ??01:01
lol s/remove/remote01:01
→ noktoborus_ joined01:03
daveb that's it!01:04
SethRobertson shivan, a question about gitslave?01:04
daveb i found it by reading the docs. thanks01:04
→ brntbeer_ joined01:04
→ phantomcircuit_ joined01:05
← Amorphous left01:05
→ bdiu joined01:06
← hobodave left01:07
← brntbeer left01:08
← brntbeer_ left01:09
→ LiamH joined01:09
→ fr0sty___ joined01:09
→ brntbeer joined01:09
← daveb left01:09
← Marquel left01:10
← comawhite left01:10
→ stliu joined01:11
← boombatower left01:11
→ rayners joined01:12
dylanb ok - so the git index is where the un-committed commits live until they're committed ?01:12
so when you git -a somefile, that file is now in your working copy and in the index01:12
but won't be added to the actual local repo until you do a commit ?01:12
frogonwheels git add somefile .. but yes01:13
dylanb ah, yeah.01:13
great.01:13
Gecko1986 he meant to create a new branch basing it of the current one :P01:13
frogonwheels dylanb: when you only stage part of a change, that's where having the whole file in the index really comes into its own01:13
← SkramX left01:14
→ zackattack joined01:14
frogonwheels Gecko1986: so git checkout -b <branch> effectively :)01:14
zackattack once i git stash; how can i apply my changes?01:14
is it "git stash apply" ??01:14
dylanb frogonwheels: 'only stage part of a change'... where can I read more about that?01:14
frogonwheels zackattack: if you want to keep the stash around yes.. otherwise git stash pop01:14
→ Marquel joined01:14
→ Yuffster joined01:14
dylanb it's a new concept to me...01:14
frogonwheels dylanb: the best way of seeing it is to use git gui01:15
→ comawhite joined01:15
dylanb ok, cool. I'll check it out.01:15
frogonwheels dylanb: click on a file with multiple changes and right-click on a hunk and 'stake hunk'01:15
→ odin_ joined01:15
← andyburke left01:15
dylanb stage/stage ?01:16
I mean, stake/stage ?01:16
even ?01:16
frogonwheels ahh yes.01:16
→ simplechat joined01:16
dylanb Ok, I'll check it out.01:16
frogonwheels dylanb: that sounds like a line from a twilight movie.. :) 'stake hunk'01:16
dylanb Thanks so much for your time & assistance :)01:16
← ijuma_ left01:17
dylanb It is much appreciated.01:17
frogonwheels yw01:17
arghh.. I have a 'compile' batch file that I use to compile my projects.. and I keep typing 'git compile <options> '01:18
→ mithro joined01:18
Guest10739 → gdoteof_home01:18
→ ijuma_ joined01:20
← thiago left01:20
← kukks left01:20
← alberto56 left01:21
→ tohava joined01:22
→ Amorphous joined01:22
← jim left01:22
← ki__ left01:22
tohava Question, I have a repository in github, I pulled it and am working on it locally (on a branch). I wish to pull the changes done in the master branch on the remote repository into my branch as they add a feature I need, how do I do it?01:23
← acts_as left01:23
← airborn left01:23
→ jds_ joined01:24
← thierryp left01:24
frogonwheels tohava: git fetch && git merge origin/master (presuming it's on master)01:25
tohava: (and that you have your feature branch checked out)01:25
→ alberto56 joined01:26
frogonwheels tohava: or you could rebase your branch on top of the new master ... in which case git rebase origin/master01:26
tohava: (that'd be what I'd do - bt it depends on the case)01:26
tohava frogonwheels, if i do not rebase, does that mean that when I commit the branch there will be conflicts between the changes done in the master and the changes done in the master as were copied to my branch?01:27
i.e. does git understand that there is no need to apply the changes that I already downloaded again?01:27
→ kanru joined01:28
frogonwheels tohava: merge vs rebase .. is just whether you want your history to be linear or to include the fact that it was a branch an merge.01:28
tohava: merge being the branch/merge thing .. and rebase being linear01:28
→ xiangfu joined01:29
tohava frogonwheels, ok, in this case I do not mind a normal merge01:29
thanks01:29
← jds_ left01:29
← lemonchicken left01:30
→ Rogi_ joined01:31
← JEEB left01:32
← stliu left01:33
← Rogi left01:33
→ bhosie joined01:34
← Rogi_ left01:35
← hakunin left01:35
→ hakunin joined01:36
→ brahmana joined01:38
← MikhX left01:39
→ lemonchicken joined01:39
→ sergio_ joined01:39
jhelwig_ → jhelwig01:40
← brahmana left01:42
→ JEEB joined01:43
→ dissidentss joined01:44
← rayners left01:44
→ thierryp joined01:45
→ rayners joined01:46
← ignatp left01:47
→ ignatp joined01:47
← noktoborus_ left01:48
→ q0rban joined01:49
← pasky left01:49
← thierryp left01:50
→ noktoborus_ joined01:50
← zackattack left01:51
← kingfishr_ left01:54
krang hey all, if I realise that I'm working on the wrong branch but have only just created it, what's the best way to change branch without losing the changes in my working directory?01:54
frogonwheels krang: git checkout <right branch>01:55
krang frogonwheels: Won't that lose my changes in the working dir?01:55
frogonwheels krang: if that would clobber files it will refuse to do it01:55
krang: if the files are the same in both branches, you'll be in the new branch with the changes still there01:55
krang: if it refuses.. git stash && git checkout <right branch> && git stash pop01:56
→ pasky joined01:56
krang Oh yeah, wow. Thanks!01:56
← Marquel left01:56
← scottwd9 left01:57
← dylanb left01:57
← s0enke left01:58
→ kingfishr_ joined01:58
→ Buglouse joined01:58
← dr_win left02:00
→ thierryp joined02:00
→ whaley_ joined02:02
← whaley_ left02:02
→ dr_win joined02:03
← kingfishr_ left02:03
→ l2trace99 joined02:05
→ Marquel joined02:05
→ s0enke joined02:05
← q0rban left02:06
→ waterbourne joined02:06
→ metoikos joined02:07
→ kingfishr_ joined02:08
→ gretch joined02:08
→ MikhX joined02:09
frogonwheels krang: git doesn't like doing things with a dirty working directory unless 1) No merging is involved 2) YOu force it (reset --hard, checkout -f &c)02:09
→ ian_mac joined02:10
krang frogonwheels: So why did it let me checkout a different branch this time, even though the working dir was dirty?02:10
→ mlq joined02:10
frogonwheels krang: no merges02:11
krang: the files that you had modified were the same in both branches.02:11
krang frogonwheels: Oh, I see, so it viewed it as working from the last commit in either branch02:12
frogonwheels krang: Ithink it's more that it finds all the files it needs to change between the two branches - and makes sure those files haven't been modified02:12
krang frogonwheels: i was about to stash, checkout, pop. thanks for confirming that would be the right thing to do under different circumstances. It's also nice to know that I can't easily ruin my working directory with a checkout02:13
frogonwheels krang: :)02:13
← _Vi left02:14
→ davetoo joined02:15
→ acts_as joined02:15
← abetaha left02:16
← bdiu left02:18
← programble left02:19
→ Vampire0_ joined02:20
← phantomcircuit_ left02:20
← Vampire0 left02:20
← dreiss left02:21
→ Slix` joined02:24
Eridius does anyone know if I can assume printf is a shell built-in on all platforms git supports?02:26
Eridius isn't sure if POSIX declares it to be a builtin02:26
→ AndIrc_ joined02:27
← jschoolcraft left02:28
← AndIrc_ left02:28
← justin-george left02:28
frogonwheels Eridius: it's an exe in msys02:29
Eridius ok thanks02:29
← kingfishr_ left02:31
← cantoma left02:31
→ q0rban joined02:35
frogonwheels Eridius: why does it need to be built-in for you to use it ?02:36
← patthoyts left02:36
→ ki__ joined02:36
← alberto56 left02:38
Eridius frogonwheels: it doesn't, it's just less performant to spawn subprocesses02:39
I was writing some shell code and I decided to replace the use of sed with shell prefix/suffix variable manipulations, but I was calling printf several times. If printf isn't a builtin, that kinda defeats the optimization02:40
frogonwheels Eridius: ah. yes. presumably you know how to do what you want to do a different way :)02:40
← q0rban left02:40
Eridius but I made a change so I only call printf once, so it's good02:40
frogonwheels cool - makes sense.02:40
Eridius in my test, calling /usr/bin/printf once after doing the variable manipulation is over twice as fast as using sed02:41
→ brahmana joined02:41
Eridius and if printf is a builtin, it's roughly 50 times faster02:41
← perlygatekeeper left02:41
frogonwheels Eridius: impressive02:41
Eridius: presumably this gets called a lot in a particular process then02:41
← merlin83 left02:41
brahmana Hi,02:41
frogonwheels Eridius: what are you doing that needs printf though?02:41
brahmana I am new to git and have been reading the Pro Git book to understand git and its usage.02:41
Eridius probably not enough to warrant the optimization ;) but it's going to be called on every argument passed to the command02:41
frogonwheels: mostly because I don't know if echo supports -n on all platforms02:42
brahmana I have written down the workflow for a typical development cycle here : http://pastebin.com/N1MgTJBU02:42
frogonwheels ahh02:42
Eridius and we already have examples of using printf from shell in the codebase, but no uses of echo -n02:42
SethRobertson echo will support -n more than printf will be available02:42
brahmana Can someone please take a look at it and see if I have understood it correctly?02:42
Eridius SethRobertson: we already depend on printf being available, whether it's a shell builtin or just in the PATH. We don't depend on echo supporting -n at the moment02:42
frogonwheels Eridius: -n is fine in msysgit if it helps :)02:43
brahmana Also, at the end, on line 78, I have put in a question. Any pointer related to that will be very helpful.02:43
← rpg left02:44
frogonwheels Eridius: couldn't you do some kind of check for -n support in echo and set ECHON or something depending02:44
→ hasrb joined02:44
Eridius frogonwheels: as we already depend on printf being available, I see no reason not to just use that02:44
SethRobertson brahmana: Assuming the dev has committed, he should next fetch and create his tracking branch for alpha202:44
→ sako joined02:44
→ cantoma joined02:45
→ AAA_awright_ joined02:45
SethRobertson brahmana: Then the dev should merge his alpha1 into alpha2. Then he is done02:45
brahmana SethRobertson: You mean committed to alpah1, right?02:45
SethRobertson yes02:45
brahmana alpha1 being deleted doesn't have any influence on his local branch?02:45
SethRobertson I'm not entirely sure why you need BRANCHES for your alpha milestone work, but we can assume you do02:45
brahmana no. Deleting a branch does not get rid of anything important, it just deletes the name02:46
← Gunni left02:46
SethRobertson I actaully expect that all of you devs, if they did a `git branch -a` would still see that alpha1 branch unless they did something special02:46
``does not get rid of anything important'' if you have merged, I mean.02:47
← AAA_awright left02:47
brahmana SethRobertson: ok02:48
SethRobertson The dev could also rebase his work onto alpha2, but that would be slightly more tricky02:48
→ naderman_ joined02:48
← naderman left02:48
SethRobertson Or cherry pick, or....02:48
→ stliu joined02:49
← stliu left02:49
→ stliu joined02:49
brahmana SethRobertson: About the branches for alpha milestones, the idea is that at any point in time the 1.3 will have the previous milestones code, i.e. stable or semi-stable.. Is there a better way?02:49
SethRobertson It depends on whether you would ever release a patch to an older alpha release. I might be very inclined to tag the milestones when they are met instead of branching all over the place. But that is more personal preference02:50
← sako left02:50
brahmana Ok.. so when we reach the milestone, close the tree for any check-ins, create a tag and then re-open it.. that way?02:51
SethRobertson You don't need to "close the tree" whatever that means. Just pull it to your QA system, do final test of whatever that pull is, and then tag on the QA system and `git push --tags`02:52
→ scottwd9 joined02:52
← Mage left02:52
brahmana Ok, when I said close the tree I meant, lock the codelines from any further check-ins until the QA work is complete.02:53
But I guess I don't need that.02:53
SethRobertson: When I push the tag from the QA machine, tag will be applied to the same point in the commit history on the remote branch also, right? (Even if there were pushes after that point)02:56
bob2 tags mark revids02:56
so yes02:56
← thierryp left02:56
brahmana ok..02:56
← scarabx left02:56
brahmana thank you SethRobertson and bob202:57
← steph021 left02:57
→ Mage joined02:57
→ steph021 joined02:57
← steph021 left02:57
→ steph021 joined02:57
← gdoteof_home left02:58
→ thierryp joined03:00
→ perlygatekeeper joined03:01
← rayners left03:01
← steph021 left03:02
← bhosie left03:03
← petercoulton left03:06
← LiamH left03:06
→ blaines joined03:06
← thierryp left03:07
→ dfr|mac joined03:07
← blaines left03:07
→ [1]dukebag joined03:08
→ alnewkirk joined03:08
← brahmana left03:08
→ blaines joined03:09
[1]dukebag → dukebag03:09
→ Bass10 joined03:09
← blaines left03:10
← Tuller left03:10
← davetoo left03:11
→ blaines joined03:11
→ shiba_yu36 joined03:11
← Buglouse left03:11
← alester left03:12
→ bdiu joined03:12
← blaines left03:12
→ blaines joined03:15
→ thierryp joined03:16
← blaines left03:16
→ rayners joined03:17
→ blaines joined03:18
← dukebag left03:18
← blaines left03:19
→ Elfe joined03:19
← DrNick left03:21
← Yuffster left03:21
→ Yuffster joined03:22
← Heimidal left03:22
→ Heimidal joined03:22
← Elfe_ left03:23
→ Heimidal_ joined03:24
← Heimidal left03:25
← Bass10 left03:27
→ blaines joined03:32
← fr0sty___ left03:33
→ fr0sty___ joined03:33
← blaines left03:34
→ fr0sty__ joined03:34
← noktoborus_ left03:37
→ anirudhr joined03:37
→ ajpiano joined03:38
← dfr|mac left03:41
← adamm left03:41
← Mage_Dude left03:44
← ajpiano left03:46
→ dreiss joined03:46
dgtlmoon how do i make a 'patch' compatible patch when i run git-diff ?03:47
bob2 they already are patch compatible03:48
dgtlmoon nein03:48
bob2 wfm03:49
kadoban dgtlmoon: there's --patch, but that's the default03:49
dgtlmoon $ git diff |patch -p003:49
patch: **** Only garbage was found in the patch input.03:49
bob2 0 is wrong03:49
103:49
← scottwd9 left03:49
dgtlmoon $ git diff |patch -p103:50
patch: **** Only garbage was found in the patch input.03:50
03:50
fffffffffffffffffffffffffu ubuntu is adding ANSI color sequences03:50
bob2 'diff = auto' in the [color] block of ~/.gitconfig03:51
offby1 I just did "git add remote work; git remote update work" ... then "git checkout work/auth". I expected to see a new local branch set up to track work/auth, but instead got the usual "you are in detached HEAD state". I thought this was supposed to create a local branch for me ... am I misunderstanding something?03:51
bob2 presumably turns off if stdout isn't a tty03:51
kadoban dgtlmoon: --no-color? i dunno, mine doesn't do that on ubuntu03:51
→ Mage_Dude joined03:51
dgtlmoon xubuntu here03:51
Eridius dgtlmoon: your color options must be set to always show them03:51
→ intripoon joined03:51
Eridius dgtlmoon: it should be set to "auto", not "always"03:51
bob2 I bet xubuntu does not change tahat in any way03:51
check your ~/.gitconfig03:51
offby1: hm, maybe checkout -t ?03:52
dgtlmoon bob2, beeeeutiful, thanks heaps :)03:52
bob2 dgtlmoon: was it set to always in ~/.gitconfig?03:52
dgtlmoon bob2, no it was not there, thanks!03:53
← dr_win left03:53
bob2 weird03:53
→ blaines joined03:54
→ DrNick joined03:54
Eridius dgtlmoon: do you have any color options set to "always"?03:54
`git config -l | grep always`03:55
← blaines left03:55
dgtlmoon i think i had an old config03:55
← ignatp left03:57
← zombor left03:57
→ scarabx joined03:58
→ mikemiller joined04:00
→ ajpiano joined04:00
mikemiller I have a commit history shared with one other person. Does git provide a way to, between two commits, get a diff between HEAD and my first commit, without including the other person's commit?04:01
*other person's commitS04:01
→ justin-george joined04:02
mikemiller Does my question make sense?04:04
frogonwheels mikemiller: not really - not sure of the strcture the commits , and if you mean HEAD , the working directory or the index.04:05
mikemiller I think I can just do git diff --author and that will resolve most issues, --rej will solve the rest.04:06
→ mdel_ joined04:10
← thierryp left04:10
Gecko1986 is it possible to just apply _1_ commit located at some git repo?04:11
→ Dunearhp joined04:11
bob2 cherry-pick04:11
mikemiller hrm, that doesn't seem to work :/04:12
← ian_mac left04:14
Dunearhp Can anyone recommend a tool for visualising vim diffs that may make sense to non-programmers?04:15
s/vim/git/04:15
vim on the brain04:15
→ MmikeMRMA joined04:15
frogonwheels Dunearhp: what kind of diffs? Use Kompare or beyond compare as a difftool ?04:16
→ scottwd9 joined04:17
mikemiller hmm, does git diff even take an author?04:17
→ thierryp joined04:17
mikemiller not mentioned in the man page :(04:18
Eridius how would diff take an author?04:18
mikemiller my situation is this: I have a series of commits by myself and someone else. I only want to include my changes in the diff.04:19
(and not the other person)04:19
← MmikeDOMA left04:19
Dunearhp Kompare looks like it might do (from the google images) Their website is down/missing04:19
→ abetaha joined04:20
← thierryp left04:20
sitaram mikemiller: a diff is, by definition, between 2 specific points. If your commits are *interspersed* with his, you will need to create a temp branch, revert all his commits (on this temp branch) and then do the diff. Then you throwaway the branch and get back to your old one04:20
Dunearhp: kdiff3 is real nice too...04:20
mikemiller sitaram: okay, I'm fine with that...04:21
so:04:21
`git branch tempbranch`04:21
`git checkout tempbranch`04:21
(checking out the first commit I'm interested in before any of thise)04:22
then how do I merge in _just_ my commits?04:22
sitaram mikemiller: that would be a cherry-pick, or a series of them... in the correct order04:22
← brntbeer left04:23
mikemiller can I do a cherry pick selecting all of the commits at once?04:23
→ tty1 joined04:23
mikemiller I guess cherry picking these manually wouldn't be too bad...04:24
→ andyl joined04:27
SethRobertson You can cherry-pick a range04:28
If you can express a search in git log or whatever, you can cherry pick that search04:28
(through the magic of xargs)04:28
mikemiller ruby script solves the problem and I know the syntax :)04:29
→ devanjan joined04:29
← bdiu left04:32
← OOPMan left04:34
→ phantomcircuit joined04:34
← rayners left04:36
← devanjan left04:37
← dreiss left04:43
← sraue left04:44
→ amerine joined04:44
← shiba_yu36 left04:48
← Slix` left04:50
→ sraue joined04:50
→ shiba_yu36 joined04:51
→ devanjan joined04:52
← jmil left04:54
→ OOPMan joined04:54
→ steph021 joined04:54
← steph021 left04:54
→ steph021 joined04:54
← steph021 left04:55
→ steph021 joined04:55
← justin-george left04:56
← acts_as left04:56
← anirudhr left04:57
→ anirudhr joined04:57
← scottwd9 left04:58
← OOPMan left05:03
→ jtaby_ joined05:03
← pedrobelo left05:03
→ pedrobelo joined05:04
← pedrobelo left05:05
→ dreiss joined05:06
→ noktoborus_ joined05:12
→ merlin83 joined05:19
→ kphanee_ joined05:20
→ roop joined05:24
← Yuffster left05:24
← foocraft left05:26
kphanee_ hi... i have a question ...05:28
when cloning from a remote repo can we ask git to exclude some dir while cloning?05:28
→ acts_as joined05:30
bob2 no05:31
frogonwheels kphanee_: you can exclude a directory on checkout - but it will be still in the repo05:31
kphanee_: look at sparse checkout05:32
kphanee_ frogonwheels: thanks :) ... i will check on that...05:33
frogonwheels kphanee_: also there's a no checkout option for clone - which you would want to use before you set up sparse configs and do your chekcout05:34
kphanee_ ok...05:35
→ eletuchy joined05:35
← mikemiller left05:38
→ disappearedng joined05:40
← disappearedng left05:40
→ disappearedng joined05:40
← disappearedng left05:41
→ AAA_awright joined05:42
← AAA_awright_ left05:44
← ki__ left05:44
→ ki__ joined05:45
→ Xeross joined05:45
← Xeross left05:45
→ Xeross joined05:45
→ jmil joined05:48
← ki__ left05:49
→ boombatower joined05:50
→ Helios joined05:50
→ OOPMan joined05:50
→ doener joined05:54
→ thiago joined05:55
← scarabx left05:55
→ Hanumaan joined05:56
← bob2 left05:57
→ disappearedng joined05:59
qDot|Away → qDot_05:59
→ AAA_awright_ joined05:59
← AAA_awright left06:02
← madewokherd left06:09
→ _ollie joined06:10
→ fr0sty_ joined06:11
← _ollie left06:13
→ ki__ joined06:13
← cantoma left06:14
← hasrb left06:17
→ goshawk joined06:17
← MikhX left06:18
← kenyon left06:18
→ kenyon joined06:18
→ MikhX joined06:18
← jsquared left06:20
MmikeMRMA → MmikeDOMA06:24
→ cantoma joined06:27
→ tty2 joined06:29
← cantoma left06:32
← tty1 left06:33
→ _ollie joined06:33
← anirudhr left06:35
← amerine left06:36
qDot_ → qDot|Away06:36
← MikhX left06:36
← tty2 left06:36
→ MikhX joined06:37
← Blackguard left06:37
← jmil left06:37
→ tty1 joined06:37
→ anirudhr joined06:38
← metoikos left06:38
← _ollie left06:44
→ parasti joined06:48
→ apsysoev joined06:48
AAA_awright_ → AAA_awright06:48
← goshawk left06:56
→ radioman-lt joined06:57
→ apsysoev1 joined07:00
← apsysoev left07:00
→ nd joined07:01
nd hi there07:03
→ Torsten joined07:03
nd why 'git gc --auto' write message "Auto packing the repository for optimum performance" to stderr and not to stdout?07:04
→ mtkd_ joined07:08
→ opalka joined07:09
Eridius probably so the user sees it if stdout gets redirected07:09
← ReleaseX left07:10
→ xiong joined07:10
→ przemoc joined07:10
→ ReleaseX joined07:11
← apsysoev1 left07:11
← mtkd_ left07:14
← MikhX left07:16
→ MikhX joined07:16
← thiago left07:18
→ dkannan joined07:18
→ frerich joined07:19
← frerich left07:19
→ frerich joined07:19
→ cbreak_work joined07:19
→ nevyn^_ joined07:19
→ jds_ joined07:20
← jds_ left07:26
← Eridius left07:27
→ priidu joined07:30
← frerich left07:33
← parasti left07:35
← devanjan left07:35
→ saschpe joined07:36
← steph021 left07:38
→ steph021 joined07:38
→ kukks joined07:39
→ jds_ joined07:40
→ foocraft joined07:40
→ FabianB_ joined07:41
← FabianB left07:42
← Mage_Dude left07:46
← Heimidal_ left07:46
→ Heimidal joined07:46
← Torsten left07:47
→ pantsman joined07:48
→ Torsten joined07:49
← carpet_the_walls left07:49
← Heimidal left07:51
→ Mage_Dude joined07:52
→ ph^ joined07:52
← waterbourne left07:54
→ valerio joined07:56
← vpilo left07:57
← unhygienix left07:58
→ galderz joined07:59
← ph^ left07:59
← gretch left08:01
valerio → vpilo08:01
→ gretch joined08:02
← roop left08:03
→ plediii_ joined08:04
→ ph^ joined08:04
→ ntoll joined08:05
→ ExtraSpice joined08:06
→ ch077179 joined08:07
← ki__ left08:08
→ ignatp joined08:08
← plediii left08:08
plediii_ → plediii08:08
← Zenopus left08:08
→ metoikos joined08:09
→ |kukks| joined08:12
→ cimenta joined08:13
→ apsysoev joined08:13
→ walf443 joined08:14
→ TomyLobo joined08:14
→ Zenopus joined08:15
← golodhrim left08:15
← kukks left08:16
→ k_sze joined08:16
→ skoop joined08:17
→ psankar joined08:17
→ psoo joined08:18
elmargol A bug deleted files in HEAD. how do I revert HEAD to 5d67550?08:19
frogonwheels elmargol: you mean your master branch?08:20
elmargol somehow a merge or something failed and I lost almost all files08:20
5d67550 seems to be the last sane commit08:20
frogonwheels elmargol: or do you want to go back and branch from 5d6755008:20
selckin git reset --hard 5d08:20
frogonwheels elmargol: that will get rid of all commits from then on.08:20
elmargol: if you want you could put a branch on your current commit so that you can still see and cherry-pick changes08:21
cimenta hi, could somebody suggest any tutorial how to set up git repository on my linux server for my developer. I do not have any code localy but I want to make it ready for him and then I will check out the code to my comp. Never used git before.08:23
→ hobodave joined08:24
← AlexP left08:25
frogonwheels cimenta: describe more how you want to use it - is your developer remote - do you want a server to keep the finished changes &c08:27
→ quesada3 joined08:28
← boombatower left08:29
→ ayanes joined08:30
← jds_ left08:30
← Mage left08:30
→ jds_ joined08:30
cimenta frogonwheels: my server is overseas, I am in Australia and my developer is somewhere else. not sure what you mean if I want server to keep finished changes. I want to have one place where the current and all versions of code will be.08:31
naderman_ → naderman08:32
← naderman left08:32
→ naderman joined08:32
frogonwheels cimenta: git is a distributed source control. so your developer will have a clone of the repository which he will do his work on, commit to it and then push it (presumably) back to your sever.08:32
← priidu left08:32
→ elmob joined08:32
frogonwheels cimenta: so you will probably want to set up something like gitolite which handles access permissions and is secure (as ssh).08:33
cimenta: (push the changes that is).08:33
← pantsman left08:33
frogonwheels cimenta: The other option is to pay to host your project on github.08:34
cimenta: (assuming a closed project) .. you'll get all sorts of great web tools to manage changes and bug requests & stuff.08:35
cimenta: and the great thing about git, is that every clone of your repo is a backup.08:35
cimenta frogonwheels: ok it looks like gitolite is the way08:36
frogonwheels cimenta: it's hosted on github - it's got great doco08:37
→ mcnellis joined08:37
→ Schmallon joined08:37
mcnellis how can I merge just one file from a branch into my master branch08:38
cimenta frogonwheels: thank you08:39
→ resmo joined08:40
AAA_awright mcnellis: Maybe git cherry-pick is what you want, otherwise just copy the file and note in the commit log. A merge might not be what you want.08:41
A merge would imply all the changes made in each line of commits now appear in the same merge commit, which isn't the case.08:42
→ brahmana joined08:43
brahmana Hi all, a quick question :08:43
→ dr_lepper joined08:44
← dr_lepper left08:44
→ dr_lepper joined08:44
AAA_awright Question! What's the question?08:44
brahmana Git branches are for the whole repo right? i.e. I cannot create a new branch for a folder?08:44
AAA_awright Correct08:44
brahmana Thank you.08:44
AAA_awright brahmana: A branch is just a pointer to a particular commit, and a commit only lists a single root tree08:44
tango_ brahmana: otoh on that new branch you can only make changes in a particular folder08:44
AAA_awright You can use submodules if you want to track it seperately, I guess08:45
brahmana tango_: yeah... got it..08:45
AAA_awright: oh no.. nothing like that. Just that there were a few non-code folders which do not need any versioning per say. So was wondering if they can be kep there without the versioning08:46
(Effect of having used Perforce for quite some time, I guess)08:46
← cimenta left08:47
AAA_awright Individual files aren't revisioned so it doesn't make much sense anyways, it's just a series (well, a graph) of snapshots of an entire tree (directory structure).08:49
→ devanjan joined08:49
→ PointMan joined08:49
PointMan is it08:49
is it possible to completely remove a file from the repository?08:50
← tolecnal left08:50
→ _ollie joined08:50
jaeckel git rm?08:50
← _ollie left08:50
jaeckel and if you want to remove it from the repository but keep the file itself try git rm --cached08:51
→ _ollie joined08:51
PointMan isnt it possible to revert a git rm?08:51
jaeckel for sure08:52
PointMan well I want to remove the file completely, as in no turning back08:52
jaeckel if you want to remove the file from the whole history you have to rewrite the tree I think08:52
← _ollie left08:52
jaeckel but not sure08:52
← avinashhm left08:52
PointMan well thats what I am asking :-)08:52
← theoros left08:53
← dreiss left08:53
← marc_kdab left08:54
→ tolecnal joined08:54
jaeckel if the file has been always commited alone, without any other file, you can do a rebase08:54
_ikke_ http://help.github.com/removing-sensitive-data/#purge_the_file_from_your_repo08:54
jaeckel or wait for someone using google like _ikke_ ;)08:55
→ mtkd_ joined08:57
← jds_ left08:57
_ikke_ This one even goes further: http://dound.com/2009/04/git-forever-remove-files-or-folders-from-history/08:59
→ rasztasd joined09:00
← mnemoc left09:02
← elmob left09:05
divVerent I have some clean/smudge filters, and autocrlf...09:05
when I merge a branch that for some reason does not have these filters applied09:05
→ davzie joined09:06
divVerent is it safe to simply amend the merge commit with the result of the filters?09:06
sounds fine to ME :P09:06
as these changes do not do anything about behaviour09:06
→ f0rkit joined09:06
→ airborn joined09:06
→ dels joined09:08
← _Unode left09:08
→ _Unode joined09:08
← resmo left09:08
_Unode → Unode09:08
Unode can anyone point me directions about how to rewind a git bare repository 1 commit09:10
divVerent which branch?09:10
Unode master09:10
divVerent you know that rewinding is bad once someone has pulled?09:10
Unode yes09:10
divVerent if you know that and acknowledge the problems it can cause:09:10
git update-ref refs/heads/master <currentcommit> <previouscommit>09:10
Unode thanks09:11
divVerent wait, wrong way round09:11
git update-ref refs/heads/master <previouscommit> <currentcommit>09:11
Unode ok09:11
that run on the bare repository, not on the cloned one right?09:11
divVerent yes09:11
update-ref works on a bare repo too09:11
→ roop joined09:11
Unode ok working on changing HEAD to HEAD^09:15
thanks09:15
→ petercoulton joined09:15
→ cantoma joined09:16
→ dpino joined09:19
→ CaMason joined09:19
→ brothers joined09:20
← dissidentss left09:23
→ resmo joined09:23
cbreak_work git's like a Hydra, if you reset HEAD to HEAD^, it instantly has a new HEAD :)09:24
sgronblo http://progit.org/2010/03/08/rerere.html I'm checking this out to learn about rerere... but to me it seems a bit weird that git outputs this:09:25
Resolved 'hello.rb' using previous resolution.09:25
Failed to merge in the changes.09:25
→ frerich joined09:25
← brothers left09:28
← PointMan left09:29
← hobodave left09:30
← sergio_ left09:31
← foocraft left09:32
→ tatsuya joined09:33
→ foocraft joined09:34
→ chrischambers joined09:39
← roger_padactor left09:40
→ AlexP joined09:41
← ch077179 left09:43
→ frsk joined09:45
→ HG` joined09:45
← OOPMan left09:47
← Hanumaan left09:48
→ akamike joined09:49
← MikhX left09:49
→ MikhX joined09:50
→ workmad3 joined09:52
← MikhX left09:54
→ artefon joined09:57
→ curtana joined09:58
← Schmallon left09:58
→ Schmallon joined09:58
→ icwiener joined10:00
← foocraft left10:00
→ bluenovember joined10:00
← Schmallon left10:03
→ Anti-X joined10:04
← stliu left10:05
← petercoulton left10:05
→ marc_kdab joined10:07
← merlin83 left10:08
→ thm joined10:09
← QinGW left10:09
← k_sze left10:10
thm if there is more than one tag describing a commit, which one will git describe prefer? the latest? or the shortest? annotated over lightweight?10:10
reto_ i bet its undefined10:11
→ MikhX joined10:11
thm can't find anything in the manpage10:11
← mithro left10:13
reto_ tolecnal: /* Multiple annotated tags point to the same commit. * Select one to keep based upon their tagger date. */10:13
ehm10:13
thm I mean10:13
← ntoll left10:13
reto_ thm: google for 03e8b541b38d95d1cd7b287963c82617a0469f8010:14
→ mattions joined10:14
← jtaby_ left10:16
→ lucsky joined10:17
← krosenvold_ left10:17
← icwiener left10:18
→ nevyn^__ joined10:24
← nevyn^__ left10:26
→ Chaser joined10:26
→ nevyn^__ joined10:27
← nevyn^_ left10:27
← kpreid left10:27
← hyperair left10:29
← plediii left10:30
→ g0bl1n joined10:32
← |kukks| left10:33
thm reto_: ok. and I guess annotated is preferred over lightweight10:33
reto_ thm: no idea sorry, just had the soruce around :)10:34
btw. this commit is available starting from v1.7.1.110:35
← eletuchy left10:37
→ eletuchy joined10:37
thm trying in a toy repo it seems that guess was right.10:41
→ icwiener joined10:41
← eletuchy left10:41
← dkannan left10:42
← yrlnry left10:42
→ lanthan_afh joined10:43
← Xeross left10:43
thm is gitosis still an active project?10:44
→ hyperair joined10:44
→ OOPMan joined10:44
→ jschoolcraft joined10:44
→ poisonbit joined10:45
poisonbit Hi starting with git here... I've make a branch and my changes on a branch... how I "apply" that branch changes to master ?10:46
→ petercoulton joined10:47
thm git checkout master ; git merge yourbranch10:47
→ zombor joined10:48
→ yrlnry joined10:48
poisonbit thanks, google was pointing me to import svn things with my keywords10:48
→ eletuchy joined10:48
→ silence joined10:50
→ ashirus joined10:50
← zombor left10:51
silence → Guest6505010:51
Juri_Agarin When tried to do a rebase to a branch, it complained about uncommitted local changes and failed to rebase some commits. However, before doing a rebase "git status" showed no uncommitted changes whatsoever. Any idea what could be causing this kind of behaviour?10:51
thm how do I clone a repo that has multiple remotes? I want to keep three or more repos in sync10:51
← ajpiano left10:54
← CaMason left10:54
← apsysoev left10:54
← xiangfu left10:55
reto_ Juri_Agarin: it doesnt even start if you have uncommited changes10:55
Juri_Agarin: ive read that git status might be out of date10:55
Juri_Agarin: well, the internal git status state10:55
Juri_Agarin: but a clal to git status should clear that 'outdated' state10:56
Juri_Agarin: and right after that git rebase shoudl work, but I'm not that sure if thats really the case :)10:56
→ btk joined10:56
→ simosx joined10:58
← simosx left10:58
→ simosx joined10:58
← ph^ left10:59
→ bentob0x joined11:00
→ dr_win joined11:01
← xiong left11:01
→ ph^ joined11:01
Juri_Agarin reto_: what is a clal?11:01
← nevyn^__ left11:02
← phantomcircuit left11:03
→ nevyn^_ joined11:03
← psankar left11:03
← Anti-X left11:03
thm call11:04
FauxFaux Grr, I wish there was no way to commit merge conflict markers.11:04
Pre-commit hook time, perhaps.11:04
← nevyn^_ left11:04
← noktoborus_ left11:06
← adys left11:06
← darkskiez left11:06
→ darksk1ez joined11:06
f00li5h kithub!11:07
→ spaceonline joined11:07
← gretch left11:07
← btk left11:08
→ rolfb joined11:08
→ Anti-X joined11:09
← zorzar left11:14
→ zorzar_ joined11:14
→ converge joined11:15
converge there is a way to see the differences of my local file and the remote one?11:15
→ parasti joined11:16
frogonwheels converge: you're best off doing a fetch (not merge) and then diffing your file against the origin/master veresion11:16
rolfb git diff name_of_remote -- path/to/file11:16
shouldn't that work?11:16
you might have to fetch first as frogonwheels said11:17
← MikhX left11:17
→ MikhX joined11:17
converge frogonwheels, rolfb worked, tks11:19
← erez left11:19
→ gzy joined11:19
→ Chillance joined11:19
gzy hi, can some one help me with a --pretty format for git log so it also shows tags?11:20
→ marcob joined11:20
Unode → _Unode11:21
← MikhX left11:22
← CoinOpeBoy left11:22
→ FsTluxe joined11:23
← bauruine left11:26
→ lucsky_ joined11:27
→ Yuuhi joined11:29
← lucsky left11:29
→ Ferris joined11:30
← roop left11:30
← jcdny left11:30
→ jcdny joined11:31
→ bitkiller joined11:31
← jcdny left11:33
→ jcdny joined11:34
← jcdny left11:35
→ jcdny joined11:35
→ koltroll joined11:35
→ apsysoev joined11:35
← anirudhr left11:36
→ erez joined11:43
albel727 gzy: please clarify. which formats did you try already, and what tags they were supposed to show you, but didn't?11:43
→ ih8mym8xd joined11:43
→ MetaCosm_ joined11:45
→ roger_padactor joined11:45
→ ehabkost joined11:45
← ih8mym8xd left11:45
→ Ryback_ joined11:45
← Ryback_ left11:45
→ Ryback_ joined11:45
gzy albel: i googled it.. i used the format:<string> and some magic formatting. I just wish it was documented in git help log.11:45
→ foocraft joined11:45
← jkp_ left11:45
albel727 gzy: so I guess it was something like "git log --format=%d --simplify-by-decoration". ok.11:46
→ R9dovoj joined11:47
gzy --pretty=format:"different_stuff i wish was documented"11:47
albel727 gzy: do you care to share the knowledge? =)11:47
cbreak_work gzy: man git-show11:48
→ JonasF joined11:48
albel727 cbreak_work: he's done already. =)11:48
gzy i ended up using:--pretty=format:"%Cred%h%Creset -%C(yellow)%d%Creset %s %Cgreen(%cr) %C(bold blue)<%an>%Creset"11:48
(modulo thecolors it leaves %h, %d, %an %cr...11:48
albel727 oh my11:49
→ engla joined11:49
cbreak_work albel727: he wanted documentation on formats, they are in there11:49
albel727 went to read the documentation11:49
albel727 also wonders, where is gitbot...11:50
gzy cbreak_work: indeed they are, thanks (though i still think it should be mentioned next to the pretty/format option of log help)11:50
← dels left11:51
gzy thanks!11:51
← gzy left11:51
albel727 I see it in man git-log too =/11:51
(section "PRETTY FORMATS")11:51
faq non-ff11:52
damn, gitbot is down again =/11:52
→ jkp joined11:53
← jkp left11:53
→ jkp joined11:53
→ golodhrim joined11:53
← tohava left11:54
← petercoulton left12:04
→ priidu joined12:04
← nd left12:04
→ nonno joined12:06
← icwiener left12:07
→ icwiener joined12:07
← jschoolcraft left12:08
← metoikos left12:08
→ metoikos joined12:09
← rgr left12:13
→ jaql joined12:15
→ goshawk joined12:16
→ paravastu joined12:17
← paravastu left12:18
← engla left12:19
← TomyLobo left12:20
→ dakota joined12:20
dakota I'm using git-http-backend for a read only service. How will I set http.receivepack to false for all repositories?12:22
→ jmil joined12:22
→ rgr joined12:23
→ _pingu joined12:24
→ roop joined12:24
→ dermoth joined12:24
← dermoth left12:25
← robgleeson left12:25
sitaram dakota: have you tried simply "git config --global http.receivepack false"12:26
→ robgleeson joined12:26
→ dermoth joined12:27
dakota sitaram: yup. But git-http-backend isn't running under my user (since it's running via apache).12:27
sitaram: I have also tried creating a .gitconfig file in the apache user's "home" with no luck12:27
→ nevyn^_ joined12:28
dermoth → dermoth|work12:28
dakota Basically I want authenticated read-only. The authentication is working fine, but the read-only part isn't. git-http-backend automatically enables the write service if the user is authenticated, this is the behaviour I want to stop12:28
← frerich left12:29
sitaram dakota: just for a test (be sure to erase it later), try "git config --system ..." (note, "system", not "global") and see if that works12:30
dakota: failing that, you have to go put that into every single repo using a "find" or something12:31
dakota sitaram: --system works12:31
sitaram dakota: and --global doesnt work because $HOME is not set for an httpd process; fair enough. You can either "setenv" it to some dir in your apache conf and then use --global and put it there, or you can just use --system12:32
dakota sitaram: working perfectly. SSH push/pull works. HTTP pull only12:32
sitaram ok12:32
dakota I'll just use --system. It does exactly what I want (Prevent any form of write access over http)12:33
thanks12:33
sitaram you're welcome!12:33
dakota didn't even know you could set system wide git configs12:33
← erez left12:33
→ icwiener_ joined12:34
→ erez joined12:34
sitaram oh yeah but it's rarely used. No one wants to be that sort of a dictator on a multi-user system I guess :) and not having $HOME is rare so most everything else is covered y --global12:34
s/y/by12:34
→ smuf joined12:35
→ Dave^| joined12:35
→ IslandUsurper joined12:36
dakota Agree. In this case it's for our central repo server (Setup somewhat like github - running redmine). We need the http service because one of our clients don't allow any form of ssh access to or from their server and we obviously don't want them to have any form of write access.12:36
→ petercoulton joined12:37
← _pingu left12:37
← icwiener left12:38
← FsTluxe left12:41
→ bauruine joined12:42
→ _Vi joined12:42
dj_tjerk different time of day.. trying again one last time:12:45
to make gitolite+gitweb+auth work, access to what folder needs to be authenticated? Also, am I supposed to make the projects.list file myself?12:45
← brahmana left12:47
← jherdman left12:47
← HG` left12:49
→ coderdad joined12:49
← jmil left12:51
→ Athox joined12:52
→ project2501b joined12:52
← Anti-X left12:54
← coderdad left12:56
← simosx left12:57
← albel727 left12:58
→ albel727 joined12:58
← foocraft left12:58
→ Schmallon joined12:58
← devanjan left12:59
→ SkramX joined12:59
← albel727 left13:00
→ albel727 joined13:00
→ engla joined13:00
← engla left13:00
→ engla joined13:00
jefferai-away → jefferai13:01
→ dels joined13:02
→ giles joined13:02
← giles left13:03
← Athox left13:03
→ flazz joined13:03
← shiba_yu36 left13:04
← warthog9 left13:06
→ orafu joined13:06
→ shiba_yu36 joined13:06
← priidu left13:08
← makubi left13:09
→ xiangfu joined13:10
→ thierryp joined13:10
→ dukebag joined13:11
← davzie left13:13
← roop left13:14
→ zombor joined13:14
→ teweWork joined13:14
← kanru left13:14
→ GarethAdams|Work joined13:15
teweWork does anyone have experience with git-stitch-repo?13:15
→ LiamH joined13:15
→ sebersole joined13:15
← Mpenz left13:15
→ Anti-X joined13:16
← mattions left13:17
→ davzie joined13:17
→ Pewpewarrows joined13:17
← yatiohi left13:18
← albel727 left13:19
→ albel727 joined13:19
→ foocraft joined13:21
→ fowlduck joined13:25
→ merlin83 joined13:25
→ mcchots joined13:26
→ rraasch joined13:26
→ sftp joined13:26
→ dfr|mac joined13:26
→ dvaske joined13:27
knittl how can i reset a branch without resetting it? xD13:30
say i have a WIP branch, and master13:30
i want WIP's tree to be HEAD's tree13:30
tango_ knittl: HEAD is just a symbolic ref to "the currently checked out commit"13:31
fr0sty knittle? 'git branch wip; git reset --hard <othercommit>'13:31
tango_ so you cannot move WIP to match HEAD unless you reset it13:31
knittl hm no13:31
tango_ knittl: if otoh you want HEAD to be WIP then you just git checkout WIP13:32
fr0sty unless I misunderstand.13:32
knittl i want something like reset --soft, but without actually moving branch refs13:32
→ herdingcat joined13:32
tango_ knittl: that doesn't make sense. could you be more specific?13:32
→ chrisred joined13:32
doener knittl: don't you rather wnat --hard without the move?13:32
knittl let me try13:32
cbreak_work what do you want to have in the end?13:32
doener knittl: --soft is just the move, so ...13:32
jast reset --soft does nothing *but* changing the ref13:32
knittl um13:32
ok13:32
cbreak_work git nop?13:32
knittl my mistake13:32
no, i have a master and a wip branch13:32
wip branch is ahead by 3 commits13:32
i'm on master13:33
doener knittl: I guess you want "add a new commit on WIP with the tree being the same as the tree of the tip of master"?13:33
knittl and want to read the WIP tree13:33
tango_ wait, let him finish13:33
knittl but still be on master13:33
tango_ knittl: git checkout wip13:33
oh13:33
knittl: git new-workdir13:33
knittl no, i want to be on master, i don't want to checkout wip13:33
jast git read-tree13:33
doener knittl: ok, so ther other way around, and what jast said13:33
fr0sty what do you want to happen to the three commits in between?13:33
jast that updates the index only, though13:33
albel727 yeah, + git checkout-tree13:33
knittl hmm. read-tree … should've thought of that myself13:33
doener ... add -u --reset to make it update the working tree as well13:34
jast if you want the working tree to match:13:34
knittl fr0sty: i don't care about them13:34
jast git checkout wip13:34
git symbolic-ref HEAD refs/heads/master13:34
foocraft → ewanas13:34
albel727 (...or was it git checkout-index...)13:34
jast saves you from mucking with clean etc. :)13:34
→ shiba_yu_ joined13:34
doener jast: uhm, just "git read-tree -u --reset wip" is a lot less trouble13:34
jast yeah, I saw that too late :)13:34
knittl too much information in a minute xD13:34
← dvaske left13:35
→ erikc joined13:35
fr0sty knittl: that is a good problem to have, I think...13:35
← erikc left13:35
fr0sty knittl: what are you trying to accomplish?13:35
knittl fr0sty: i'm hacking away on WIP branch13:36
because i'm still defining features13:36
then i checkout master branch again (clean working tree)13:36
← shiba_yu36 left13:36
knittl and want to have the state of WIP as 'unstaged'13:36
does that make sense?13:36
← ExtraSpice left13:37
cbreak_work git reset --hard WIPBRANCH13:37
git reset --soft HEAD@{1}13:37
why?13:37
← rolfb left13:37
knittl yeah, something like that13:37
why what?13:37
← hyperair left13:37
cbreak_work what is it good for?13:38
IslandUsurper why not just merge?13:38
cbreak_work you want to do some evil svn style pseudo merge?13:38
← mcchots left13:38
knittl no, but WIP branch is really messy13:38
IslandUsurper (once you're done hacking anyway)13:38
→ GeertB joined13:38
knittl messier than git.git/pu13:38
cbreak_work just rewrite it :)13:38
→ rolfb joined13:38
cbreak_work (before merging)13:38
jast merge --squash, then?13:38
or that13:38
knittl it's just adding a bunch of files, so i can push them for others to see13:38
→ erikc joined13:38
fr0sty and why is having three commits evil?13:39
knittl i then would only add some of the files, change others13:39
cbreak_work merge --squash is evil subversionism13:39
fr0sty (or pushing a WIP branch, for that matter)13:39
yeah, don't merge squash13:39
jast it has its place13:39
knittl …13:39
jast not often, admittedly, but sometimes it does13:39
engla merge --squash: This topic branch should just be one commit, I don't want to see that guys three correction commits.13:40
knittl WIP was just a quick way to share broken/unfinished/incomplete code with my colleagues13:40
engla: WIP is not even finished yet13:40
fr0sty engla: at least have the courtesy to rebase --squash and start your work from the rebased commit...13:40
→ apassi joined13:41
knittl i'll read about read-tree and probably just use the double reset (--hard + --soft) thingy13:41
apassi Hi, is there some pre- hook for git-reset --hard ?13:41
fr0sty 'rebase doesn't take --squash, I know' rebase -i + "edit to squash all but one commit" is what I meant.13:41
engla fr0sty: of course, I understood13:42
jast apassi: nope13:42
fr0sty apassi: no.13:42
erikc how would i get a list of the created, removed and changed files in a commit13:43
(im trying to merge into another source control system)13:44
fr0sty erikc: which one?13:44
knittl erikc: probably git whatchanged13:44
cbreak_work erikc: man git-log13:44
→ SimonP86 joined13:44
apassi =(13:44
cbreak_work there are various flags like --stat, --numstat and so on13:44
fr0sty erikc: the git-log manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-log.html13:45
apassi: what are you trying to do?13:45
cbreak_work it's probably evil13:45
apassi I think there should be some pre/post hooks for reset, because in my situation, there is only read acces for user to some files, and git-reset shows some errors, i just like to change the user rights for these files, until reset is done.13:45
erikc fr0sty: CMVC, ibm's quite old internal source control system13:46
fr0sty erikc: I see... which way are you trying to go? into our out of?13:46
→ rpg joined13:47
erikc both, i have cmvc->git working, working on the opposite way now13:47
→ jherdman joined13:47
← OOPMan left13:49
→ brntbeer joined13:49
→ dvaske joined13:49
fr0sty apassi: git doesn't manage read/write permissions.13:50
← Sigma left13:50
fr0sty btw, why do you want users to be able to modify files they are not allowed to modify?13:50
jast I've got a wrapper script around git for hooking into commands13:50
→ kenneth_reitz joined13:51
apassi fr0sty: git-reset --hard shows some errors, if current user dont have write permissions to file. I am developing in drupal environment, and it shows some annoying messages if settings files are not write protected.13:52
fr0sty: and in hook i could add write permission until git-reset is done, and after that remove permission.13:52
← mtkd_ left13:54
→ mtkd_ joined13:55
← artefon left13:57
← Strogg left13:57
→ TeckniX joined13:57
← skoop left13:57
← goshawk left13:58
→ jstemmer joined13:59
→ OOPMan joined13:59
← project2501b left13:59
← brntbeer left13:59
→ hendry joined14:00
hendry how do I `git clone` a certain branch in one step? http://review.webmproject.org/gitweb?p=libvpx.git;a=shortlog;h=refs/heads/aylesbury14:01
→ majd joined14:01
→ kanru joined14:01
fr0sty apassi: do you have multiple users sharing a working tree?14:01
→ henk joined14:02
henk hi14:02
→ akahn joined14:02
← thierryp left14:02
→ rossand joined14:02
← rossand left14:03
fr0sty hendry: 'git clone --no-checkout http://... && git checkout aylesbury'14:03
→ mtkd__ joined14:03
← workmad3 left14:03
henk i did a few changes to my repository but don't want to commit them all together but rather one by one. is it still possible to seperate the changes to several commits?14:03
dwmw2_gone → dwmw2_BOS14:03
← dwmw2_BOS left14:03
→ dwmw2_BOS joined14:03
← galderz left14:04
parasti henk: have a look at git add -p and git rebase -i14:04
→ petran joined14:04
parasti henk: interactive rebase is for already-commited work, though14:05
jzacsh → jzacsh_14:05
henk parasti: thanks!14:05
← mtkd_ left14:07
← henk left14:07
← robgleeson left14:07
← engla left14:09
→ robgleeson joined14:09
rasztasd hi, im mergin two files, how i decide if i want to resolve the conflict by selecting the remote or the local version?14:09
← robgleeson left14:09
→ robgleeson joined14:09
← SkramX left14:09
→ engla joined14:10
cbreak_work rasztasd: look at the code14:10
the developer is the only one who ca ndecide that14:10
← JohnDav left14:10
rasztasd i can deceide to keep the remote or local entirely14:10
→ chronos joined14:11
rasztasd but how, i dont want to use a difftool and click the click change button for every line14:11
cbreak_work just git checkout14:11
read the man page, there's --ours and --theirs14:11
→ pigdude joined14:12
← OOPMan left14:12
← SimonP86 left14:12
← walf443 left14:12
← teweWork left14:12
← Mage_Dude left14:12
← Torsten left14:12
← dakota left14:12
← kenyon left14:12
← mdel_ left14:12
← fr0sty left14:12
← marienz left14:12
← jhelwig left14:12
← zenwryly left14:12
← cannonball left14:12
← jammi left14:12
← jspiros left14:12
← GodEater left14:12
← unixtippse left14:12
← Fissure left14:12
← sivy left14:12
← TDJACR left14:12
← optixx left14:12
← jonmasters left14:12
← ilteris left14:12
← Tac left14:12
← zain left14:12
← LetoThe2nd left14:12
← LeLutin left14:12
← frakturfreak left14:12
← ndim left14:12
← vmiklos left14:12
← gcl left14:12
← knielsen left14:12
← nanotube left14:12
← kupsch left14:12
← Nickeeh left14:12
← bmalee left14:12
← BlindOracle left14:12
← frerich6 left14:12
← kandjar left14:12
← cxreg left14:12
← rickmasta left14:12
← mfoemmel left14:12
← Moo^_^ left14:12
← candybar left14:12
← SethRobertson left14:12
← wooli left14:12
← mika left14:12
← tehbaut left14:12
← ribasushi left14:12
← sbob left14:12
← Stx left14:12
← LucasWilcox left14:12
← ahf left14:12
← kloeri left14:12
← siyb left14:12
← eregon left14:12
← sitaktif_ left14:12
→ q0rban joined14:13
← engla left14:14
→ Gunni joined14:15
→ exiton joined14:16
← erikc left14:18
→ sivy joined14:18
apassi fr0sty_: yes and no, the group is www-data so apache can read it.14:18
← cballou left14:19
apassi fr0sty_: i could just add all write permissions, but i like to develop as close as i can to production environment, so i like to keep the rights as they are going to be in prod.14:19
rasztasd apassi: thx14:20
→ LRN joined14:20
→ coreyballou joined14:21
→ galderz joined14:21
← majd left14:22
→ MattDiPasquale joined14:22
→ boombatower joined14:24
→ nicoulaj joined14:24
← boombatower left14:24
→ boombatower joined14:24
→ bmalee joined14:24
→ marienz joined14:24
→ vadim joined14:24
→ fr0sty joined14:25
← dvaske left14:26
← koltroll left14:26
→ engla joined14:26
← engla left14:26
→ engla joined14:26
← fr0sty left14:27
← tty1 left14:27
vadim Hi, is it possible to checkout a subdirectory? I'm fine with cloning the entire repository, I just want to see a subdirectory on it. I have a repository that follows a FS layout (eg, var/www/foo, etc/apache2), and want to make it so that I can get the etc/apache2 files in /etc/apache2 on a box14:27
← elmargol left14:27
← jefferai left14:28
← sonnym left14:28
→ jefferai joined14:28
← steph021 left14:29
joevano The short answer is no. Git sees all files as a single unit.14:29
→ nanotube joined14:29
← radioman-lt left14:30
vadim joevano: ok, thanks14:31
→ JonMarkGo joined14:31
JonMarkGo I'm trying to git pull one of my branches and I keep getting "fatal: fsync error on '.git/objects/pack/tmp_pack_p2CbOe': Permission denied"14:31
→ Hackerdudecom joined14:33
vadim → 14WAAAFKI14:33
→ OOPMan joined14:33
→ SimonP86 joined14:33
→ teweWork joined14:33
→ dakota joined14:33
→ walf443 joined14:33
→ Mage_Dude joined14:33
→ Torsten joined14:33
→ kenyon joined14:33
→ mdel_ joined14:33
→ LeLutin joined14:33
→ jhelwig joined14:33
→ zenwryly joined14:33
→ cannonball joined14:33
→ Fissure joined14:33
→ jammi joined14:33
→ jspiros joined14:33
→ unixtippse joined14:33
→ GodEater joined14:33
→ TDJACR joined14:33
→ optixx joined14:33
→ knielsen joined14:33
→ jonmasters joined14:33
→ VABQ1HF joined14:33
→ Tac joined14:33
→ zain joined14:33
→ LetoThe2nd joined14:33
→ frakturfreak joined14:33
→ ndim joined14:33
→ vmiklos joined14:33
→ gcl joined14:33
→ kupsch joined14:33
→ Nickeeh joined14:33
→ BlindOracle joined14:33
→ frerich6 joined14:33
→ kandjar joined14:33
→ cxreg joined14:33
→ rickmasta joined14:33
→ mfoemmel joined14:33
→ Moo^_^ joined14:33
→ candybar joined14:33
→ Stx joined14:33
→ SethRobertson joined14:33
→ wooli joined14:33
→ sitaktif_ joined14:33
→ mika joined14:33
→ ribasushi joined14:33
→ tehbaut joined14:33
→ sbob joined14:33
→ LucasWilcox joined14:33
→ ahf joined14:33
→ kloeri joined14:33
→ siyb joined14:33
→ eregon joined14:33
→ crab joined14:33
→ tizbac joined14:34
JonMarkGo I'm trying to git pull one of my branches and I keep getting "fatal: fsync error on '.git/objects/pack/tmp_pack_p2CbOe': Permission denied"14:36
← marienz left14:36
→ fr0sty joined14:37
Hackerdudecom JonMarkGo: chown -R yourusername .git ?14:37
→ hyperair joined14:38
→ verto joined14:38
fr0sty vadim / 14WAAAFKI : look at sparse-checkout http://vmiklos.hu/blog/sparse-checkout-example-in-git-1-714:38
JonMarkGo Now I'm gettin a bunch of error: refs/remotes/unfuddle/features does not point to a valid object!14:39
Hackerdudecom I'd also make sure to use ls -l .git/objects/pack/* and see the permissions, make sure there's at least rw——— (owner can read+write)14:39
→ brntbeer joined14:39
← herdingcat left14:39
→ Number5 joined14:40
→ slyphon joined14:40
→ herdingcat joined14:40
→ tty1 joined14:40
← tty1 left14:40
→ tty1 joined14:40
← ewanas left14:41
← zombor left14:41
Hackerdudecom Do you do the management of the unfuddle remote?14:41
→ zombor joined14:42
Number5 A beginner's question. If I copy a directory with the git directory in it, to another location or pc, it's like cloning that directory? I mean I can assume everything stays like in the original directory. Just to be sure.14:42
→ noktoborus_ joined14:42
← mtkd__ left14:42
← dels left14:42
bremner Number5: it is not exactly like cloning, but yes, everything stays the same, unless you copy to some brain-damaged file system that doesn't know the difference between upper and lower case.14:43
albel727 (...like NTFS)14:43
→ shruggar joined14:43
→ alberto56 joined14:43
Number5 bremner: no, all my projects will run on Linux, at leas that's my goal. But suppose I want it on windose, than the best thing to do is clone?14:43
14WAAAFKI fr0sty: aha, that looks like just what I need, thanks a lot :-)14:43
→ scarabx joined14:44
14WAAAFKI darn14:44
14WAAAFKI → vadim314:44
→ ExtraSpice joined14:44
JonMarkGo Hackerdudecom - I'm still getting permission denied, everything is owned by root but its a VM where I'm a sudoer14:44
bremner Number5: probably. But I never use windows, so...14:44
JonMarkGo And I can pull other branches14:44
Number5 bremner: I totally understand you never us it :D14:44
shruggar what must I pass to "git log --follow" to tell it "Follow renames on file Foo, but when showing me stats, --name-only, etc, output all paths which the commit touched"?14:44
← golodhrim left14:45
engla shruggar: possibly --full-diff14:45
IslandUsurper Number5, the difference between copying and cloning is what the remotes point to14:45
shruggar --full-diff complains "usage: git logs can only follow renames on one pathname at a time"14:45
bremner IslandUsurper: to be precise, config information is lost by cloning, but not by copying.14:46
IslandUsurper right14:46
→ yairgo joined14:46
engla shruggar: --follow has a couple of limitations14:46
→ tj joined14:47
→ didi joined14:48
Number5 IslandUsurper: aah I understand, but you can always modify the remote, so that wouldn't be a problem (so does the config), right?14:49
Hackerdudecom JonMarkGo: It could be a corrupted reflog but it's more likely just a permissions problem. I'm assuming (never seen that error before myself) that ref/remotes/unfuddle/features can't get to the SHA file that it's pointing to. If you are on the remote, can git diff features..features^ give you a proper diff? If so, you're not corrupted on the server side and "everything will be fine" :-)14:49
fr0sty shruggar: do do passes: 'log --follow ... |git show --format="..."'14:49
one gets you the commits the other shows you what changed...14:49
hmm, probably need xargs in there, eh?14:50
→ dels joined14:50
shruggar yeah, I was trying to avoid xargs14:50
JonMarkGo Server side is fine14:50
shruggar (or similar)14:50
→ warthog9 joined14:50
Hackerdudecom Good. Are you accessing via git-ssh?14:50
← scarabx left14:50
shruggar git log --stdin, on the other hand, shows all the parents of the listed commits, too14:50
fr0sty you allergic or something?14:50
shruggar yeah :)14:50
fr0sty try benadryl.14:51
IslandUsurper Number5, sure. You can always chang the config file. It's just going to be different based on the command you use.14:51
didi I am a little confused about the way one is supposed to create patches. I read that projects like the Linux Kernel likes the `format-patch' way. But it seems to create a file that resembles an email. One should insert a `To:' header to it and send the e-mail as is or it should be attached to an actual e-mail?14:51
Number5 IslandUsurper, bremner thank you for your help.14:52
← dfr|mac left14:52
← hebz0rl left14:53
← tizbac left14:53
→ cirwin joined14:53
engla didi: how you transfer the file is not important, you can send it as attachment, or attach the file to a bug report. the important thing is that the whole patch is there (subject, commit log and diff)14:53
→ tizbac joined14:53
engla is irritated when someone sends him a "patch" and it's only a diff14:54
bremner hands engla the patch utility14:55
→ kenleycapps joined14:55
engla now I need to do a three-way merge, patch doesn't apply?14:55
Hackerdudecom JonMarkGo: My ls -lah of .git/objects on server side gives me permissions like −4–4–4— ownerfrommyteam ourgroup filename (and owner+group "x" for directories). Those are proper permissions if you're using git ssh. Check the sharedRepository and umask settings on gitconfig and set them to something sensible on the server.14:56
teweWork can i rewrite a file in history which is originally encoded ISO-8859-2 to look like it was always UTF-8 ? :)14:56
fr0sty teweWork: man git-filter-branch14:56
Hackerdudecom JonMarkGo: As it seems to me that your git ssh is putting things as the wrong owner. You also need to chown + chgrp to set proper owners and permissions14:57
didi engla: Oh, OK. So attach the 0001-thingy. Thank you.14:57
→ thiago joined14:58
engla didi: for linux and git projects, you need to send it as an email. in other context it's up to you, maybe attaching it etc14:58
teweWork fr0sty: so it can be done, i haven't got to the "how to" yet :)14:58
→ err_ok joined14:59
← mkramer left15:00
→ mkramer joined15:00
fr0sty teweWork: it involves --tree-filter and a program which rewrites the file to have the correct encoding.15:00
→ mfacenet joined15:00
← mfacenet left15:00
→ mfacenet joined15:00
← dakota left15:00
didi engla: Hum, right. That's a problem then. I'll need to find a program that would accept the file as a prepared e-mail. Another thing, I need to be more verbose in my commit messages, ;).15:00
teweWork fr0sty: thx15:00
→ Yuffster joined15:01
← ashirus left15:01
→ jrmuizel joined15:02
didi types git commit --amend [...]15:02
→ marienz joined15:02
cirwin I often find myself doing a git rebase -i to swap the latest two commits on a branch, is there a way to use non-interactive rebase to do that, or a safe set of other commands I could run in an alias?15:03
← rraasch left15:03
Guest65050 → ajpiano15:05
→ hebz0rl joined15:05
esc cirwin: get the hashes, reset, and cherry pick in correct order?15:05
Bombe cirwin, sure. git checkout -b swap-last-two HEAD~2; git cherry-pick master; get cherry-pick master~; git checkout master; git reset --hard swap-last-two; git branch -d swap-last-two.15:05
fr0sty cirwin: you can do: GIT_EDITOR=some_script git rebase -i HEAD~215:05
Bombe Assuming you’re on master.15:05
cirwin ah, I like fr0sty's approach15:05
fr0sty and have somescript switch the two lines then exit15:06
cirwin the problem with just running multiple cherry picks is checking everythign ran successfully15:06
← _Vi left15:06
cirwin and aborting cleanly15:06
and saying what happened so I can fix it15:06
rebase already does that15:06
thanks :)15:06
← sambarino left15:06
→ yngress joined15:06
→ _Vi joined15:07
engla cirwin: git rebase -i with GIT_EDITOR=tac could do it15:08
cirwin engla: already there :)15:08
→ sarpanch joined15:09
cirwin hrm, maybe not15:09
needs to do in-place modify15:09
← _Vi left15:09
→ alex__c2022 joined15:10
engla something like tac $1 | sponge > $115:11
sponge from moreutils15:11
cirwin sed -i '1{x;d};2{p;x}'15:12
← Yuffster left15:13
→ sonnym joined15:13
← shennyg left15:13
→ Yuffster joined15:13
engla that works nicely15:13
I don't know sed15:13
→ adamcooke joined15:14
→ warbee joined15:14
→ karpar joined15:14
cirwin well, it took me four attempts to write that sed script :)15:15
→ _Vi joined15:15
→ ShooterMG joined15:15
→ eventualbuddha joined15:15
← marienz left15:16
← xiangfu left15:18
← rasztasd left15:19
→ rasztasd joined15:19
← OOPMan left15:20
IslandUsurper → IslandUsurperAFK15:20
← tizbac left15:21
→ tizbac_ joined15:21
→ workmad3 joined15:22
→ ymasory joined15:24
→ JoshS joined15:24
← icwiener_ left15:26
didi engla: Is it possible to write something at the `format-patch' e-mail in a way that it won't show up in the history?15:26
→ sarpanch_ joined15:26
← err_ok left15:27
engla didi: sure, I think it's commonly written after the "---" line before the diffstat15:27
→ Heimidal joined15:28
← sarpanch left15:29
→ marienz joined15:30
← sarpanch_ left15:32
→ sarpanch joined15:32
didi engla: By `diffstat' do you mean the part like "foo.c | 2+-"?15:33
← fr0sty__ left15:33
← dels left15:33
→ fr0sty____ joined15:33
← curtana left15:33
← fr0sty___ left15:34
→ fr0sty__ joined15:34
→ OOPMan joined15:34
sarpanch I have created a post-commit hook for gollum wiki to update remote git repository, however the wiki commits the repository but does not fire the hook. Manual commits fire the hook just fine. Any ideas what I might be missing?15:35
← opalka left15:35
→ curtana joined15:36
← JonMarkGo left15:36
frerich6 What's a good way to transplant a (non-consecutive) set of commits from on repository into a new one? We've been doing some work but now decided that this should really be in a repository of its own. Is there any way to keep the commits? All the work was done in one directory, and doing a 'git log .' in that directory basically shows all the commits I want to transplant. I looked at 'git format-patch' for whether15:37
it could generate a patch set which I could reapply in a new directory - but that doesn't seem to work.15:37
← Guest78925 left15:38
← psoo left15:39
engla didi: that's the diffstat yeah15:39
frerich6 Ah, I could filter-branch to mark everything but my desired directory for garbage collection.15:39
didi engla: Nice. Thank you.15:39
engla frerich6: you can use the 3rd party tool git-subtree to split out that subtree15:40
← JoshS left15:40
→ coderdad joined15:40
engla frerich6: I don't know if that tool will remove the "empty commits" in the resulting history though.15:40
fr0sty frerich6: git filter-branch --subtree-filter15:40
→ majd joined15:41
engla (git-subtree is just an interface to filter-branch and merge subtree etc, quite conveniently packaged. It's anyone's choice which method to use)15:41
fr0sty if you have empty commits you can get create a list of commits and drop them with rebase -i15:41
← thiago left15:42
← sarpanch left15:42
→ bcardarella joined15:42
engla I'd use git subtrees split15:42
*git subtree15:42
← ymasory left15:43
← marienz left15:44
← ph^ left15:45
← herdingcat left15:47
→ archis joined15:47
← dwmw2_BOS left15:47
→ jmil joined15:49
← hyperair left15:50
← etnt left15:50
← rolfb left15:51
← yngress left15:52
← simplechat left15:53
→ ochosi joined15:53
→ dwmw2_BOS joined15:53
ochosi hello everyone, i have a maybe trivial question, but i couldn't find out via google. can git be used with files that contain formatted text (e.g. rtf)?15:53
← f0rkit left15:54
fr0sty ochosi: man git15:54
'git - the stupid content tracker'15:54
← exiton left15:54
sitaram ochosi: it can, but binary files (which includes stuff like rtf, even if it seems like text) do not *merge* well, or at all. As long as you are aware of that, it'll work fine15:54
fr0sty git doesn't care.15:55
but like sitaram says: git can't help you with 'funky' files (merging or diffing).15:55
← majd left15:55
ochosi merges are noto that important to me, i'm working alone on a text-edition project15:55
but i want to document the changes15:55
← nuoHep left15:56
sitaram then go ahead15:56
ochosi so to reformulate my initial question: can git track the changes in a binary file in a human-readable way?15:56
or is there any other way you know of doing versioning with rtfs and the like15:57
engla it's impossible in the general case, if you have any standards for "human readable"15:57
→ mklappstuhl joined15:57
sitaram ochosi: track, yes. human-readable, probably not. I dont know *any* way in which it can track it which is readable by me15:57
mklappstuhl hey15:57
SethRobertson If `diff` does something useful, then the answer is yes.15:57
If diff does not, then you can write your own diff which does and git will use that15:57
curtana you can configure git to convert your RTF files to text for the purposes of diffing15:57
ochosi i can tell git to use a certain diff?15:57
SethRobertson yes15:57
ochosi ah, wow15:57
← adamcooke left15:58
ochosi that all sounds pretty cool15:58
curtana i think the feature is called textconv15:58
ochosi ok, i'll go ahead and google that :)15:58
→ nuoHep joined15:58
mklappstuhl when i have a repo with submodules in /repo1/ and then do "git init" in / can I add my /repo1/ as submodule afterwards?15:58
ochosi thanks everyone!15:58
curtana good luck15:58
← converge left15:58
SethRobertson mklappstuhl: Depending on what you are doing, you may find gitslave matches better than submodules for your workflow. http://gitslave.sf.net15:59
← _Vi left15:59
→ _Vi joined16:00
← hakunin left16:00
SethRobertson ochosi: There is also diff.external and difftool.* and the external diff driver (see gitattributes)16:00
and git-config16:01
dwmw2_BOS → dwmw2_gone16:01
mklappstuhl SethRobertson: looks interesting but i'd like to try onboard tools first.. also i dont need cross-repository branching for example..16:01
→ alin-gym joined16:02
SethRobertson Then submodules might be right. With your question, I don't think git-submodules (or gitslave) supports recursive subprojects especially well, but you can certainly move who controls the subprojects16:02
frerich6 I used 'git filter-branch --subdirectory-filter' to split a single directory into a new repository. However, there is still a huge pack file which surely contains mostly unneeded objects. How can I enforce that this is repacked? I tried 'git gc --aggressive' and 'git prune' and 'git reflog --expire=now --all' and 'git repack -ad' but nothing seems to touch the pack file.16:02
← archis left16:03
← Number5 left16:03
fr0sty frerich6: was this a clone of your other repository?16:04
→ ashirus joined16:04
frerich6 fr0sty: Yes. I cloned the other repository using --no-hardlinks (since I thought maybe those hardlinks cause the issue)16:04
SethRobertson Isn'tit `git repack -Ad`?16:04
engla clone it again16:04
or remove the backup refs16:04
fr0sty +1 engla16:04
mklappstuhl SethRobertson: haha then i'll be the brave who tries :D16:05
→ hyperair joined16:05
frerich6 engla: Hmm, cloning again didn't seem to make a difference. I must have some dangling refs... but there's nothing in .git/refs except .git/refs/heads/master16:05
→ justin-george joined16:05
→ kenleycapps_ joined16:06
frerich6 Actually, something else is odd: I can't seem to check out a particular version. My working tree is always empty.16:06
cbreak_work frerich6: how did you clone?16:06
← shiba_yu_ left16:06
← abetaha left16:06
frerich6 cbreak_work: I cloned locally using 'git clone --no-hardlinks foo bar'16:06
q0rban → q0rban|lunch16:07
cbreak_work check git log -p, if the diffs are strange, maybe filtering failed16:07
as for cloning, try using git clone file:///...16:07
→ sftp_ joined16:07
← kenleycapps left16:07
← sftp left16:07
kenleycapps_ → kenleycapps16:07
frerich6 I am on Windows right now, using msysgit. I wonder that affects this somehow.16:08
→ marienz joined16:10
← justin-george left16:10
→ alvaro_o_ joined16:11
→ dreiss joined16:12
← sftp_ left16:12
→ sftp_ joined16:13
→ yngress joined16:13
← cbreak_work left16:13
frerich6 Hm. FWIW, the repository is much smaller than the original already, but still way too large. It seems filtering the branch really went wrong - all the trees of the commits are empty.16:13
→ justin-george joined16:13
→ wsclock joined16:15
→ dels joined16:15
→ didi` joined16:16
→ elmob joined16:16
frerich6 Ah, interesting: after filtering the branch, pushing to a bare git repository actually just pushes the needed objects.16:16
engla git-subtree maybe16:17
← didi left16:18
← eno left16:18
→ ymasory joined16:19
← engla left16:19
→ eno joined16:20
→ ignatp_ joined16:20
→ slonopotamus joined16:20
← Torsten left16:21
← dreiss left16:22
← marienz left16:22
← _jesse_ left16:23
← AAA_awright left16:23
frerich6 engla: I'll give git-subtree a shot when I'm back home at my FreeBSD machine. git filter-branch *almost* works. I get the right commits, but all the trees of the commits are empty16:23
← _Vi left16:23
← warbee left16:24
→ warbee joined16:24
→ aah joined16:25
→ FsTluxe joined16:26
← tg left16:26
reto_ FsTluxe: werent you the one who asked yesterday about merging sub directories?16:26
→ tg joined16:27
reto_ negative..16:27
FsTluxe: http://stackoverflow.com/questions/277029/combining-multiple-git-repositories16:27
dont know if it helps anything16:27
→ icwiener joined16:28
← saschpe left16:28
→ aspotashev joined16:29
→ nevyn^__ joined16:31
IslandUsurperAFK → IslandUsurper16:31
← metoikos left16:31
→ Torsten joined16:31
→ irc2samus joined16:32
← nevyn^__ left16:32
← kenleycapps left16:34
← g0bl1n left16:34
← warthog9 left16:34
← jherdman left16:34
→ jherdman_ joined16:34
← jherdman_ left16:35
→ jherdman joined16:35
← nevyn^_ left16:35
← bauruine left16:36
← justin-george left16:37
didi` → didi16:39
→ justin-george joined16:39
→ thierryp joined16:39
← nonno left16:40
← kanru left16:41
→ abetaha joined16:43
→ whatdoesitwant joined16:44
← marc_kdab left16:44
→ rossand joined16:44
fr0sty → fr0sty___16:44
→ Phrogz_ joined16:45
→ psoo joined16:45
qDot|Away → qDot_16:47
← knielsen left16:47
← jherdman left16:48
← thierryp left16:48
← kenneth_reitz left16:49
→ thierryp joined16:49
→ kumbayo joined16:49
← Phrogz_ left16:51
← thierryp left16:51
→ ph^ joined16:51
← Schmallon left16:52
← rossand left16:52
→ jherdman joined16:54
← sftp_ left16:56
przemoc hi, can I import hg repo into git (via hg-git's gexport) with authors fixing (i.e. changing login to real name and adding mails) w/o commiting .mailmap?16:57
q0rban|lunch → q0rban16:57
→ spearce joined16:57
przemoc this problem is not covered at hg-git.github.com page16:57
→ kenichi joined16:58
fr0sty_ → fr0sty16:59
mklappstuhl when i have a gitignore with "*" in the first line: how can i whitelist an entire folder recursively? !folder doesnt work recursively16:59
← teweWork left17:00
→ sftp joined17:00
mklappstuhl and !folder/**/**/**/* does not really feel right :)17:00
← mlq left17:01
← q0rban left17:02
← ymasory left17:02
ashirus hi all17:03
got a weird problem with git --config diff.tool17:04
the settings don't seem to stick17:04
fr0sty mklappstuhl: why do you have that?17:04
"*" in .gitignore17:04
mklappstuhl fr0sty: dotfiles (dont know if i got the question though)17:04
→ kipras joined17:05
← aspotashev left17:05
fr0sty the question is: why are you, as a rule, ignoring every file in the tree?17:05
mklappstuhl fr0sty: ah, because maintaining a list of things that shouldnt be inside the repo would be harder17:05
ashirus currently I'm trying to set 'git config --global diff.tool p4merge'17:05
and then if I do 'git config difftool.p4merge.cmd=foo.sh' it doesn't stick17:05
fr0sty what 'things' _are_ in the repo, if I may dare to ask...17:05
mklappstuhl .vim .zsh .irssi etc. pp.17:06
ashirus it keeps on calling my old config for difftool17:06
← disappearedng left17:06
ashirus anyone experienced this?17:06
fr0sty mklappstuhl: so you have a ~/.git folder?17:06
mklappstuhl fr0sty: yup17:06
fr0sty ashirus: man git-config17:06
you need --set or --add17:06
mklappstuhl: that is a fantastically bad idea, generally...17:07
mklappstuhl fr0sty: okay :D why17:07
fr0sty http://pyverted.com/linux/setting-up-version-control-on-your-home-directory/2009/08/17:08
→ disappearedng joined17:08
← disappearedng left17:08
→ disappearedng joined17:08
ft mklappstuhl: so, you're not scared of accidentially running "git clean -xdf" in the wrong directory?17:08
fr0sty Bart's comment pretty much covers it.17:08
ashirus fr0sty: I know git-config but it's not using the difftool I'm settting17:08
fr0sty ashirus: what does 'git config -l' give you for diff.tool or difftool.* ?17:09
← smuf left17:09
→ necromancer joined17:09
→ kukks joined17:10
→ nullvariable joined17:10
← fr0sty left17:10
ashirus diff.tool=p4merge17:10
→ steph021 joined17:10
ashirus difftool.p4merge.cmd=sfgdfg.sh17:10
bartman fr0sty: a friend of mine lost of all $HOME once... same problem, but with .hg17:10
s/of all/all of/17:10
ashirus (sfgdfg.sh is nonsense put there just to verify it's not suing that config for some eason)17:10
→ fr0sty joined17:11
ashirus how could git be ignoring the difftool.p4merge.cmd setting?17:11
→ dreiss joined17:12
bartman fr0sty: oops, that's what I wrote in the blog commands17:12
mklappstuhl fr0sty: okay got it17:12
bartman: ^^17:12
bartman hehe17:12
yeah17:12
← FsTluxe left17:13
ashirus and before you ask there'\s no override in .git/config17:14
surely a call to git dofftool ought to be reporting an error as sfgdfg.sh doesn't exist17:14
→ freq9 joined17:15
fr0sty ashirus: what happens instead?17:15
ashirus it launches with the old difftool17:15
mklappstuhl bartman: fr0sty its interesting how much mistakes and bad practices you can make only managing you dotfiles :d17:15
ashirus (that I set using git config itself!)17:15
← karpar left17:16
freq9 i forked a repository a few weeks ago and now git tells me that my fork's master is 397 commits ahead of 'origin/master'. what do i need to do to be up-to-date with origin/master?17:17
ashirus basically the full story is that I set: git config --global difftool.p4merge.cmd p4merge '$LOCAL' '$REMOTE'17:17
→ kalusn joined17:17
ashirus and then I realised that as I'm on cygwin I need to 'cygpath'-ise the parameters, so I set:17:18
fr0sty freq9: git push origin master17:18
ashirus git config --global difftool.p4merge.cmd "p4merge \$(cygpath -w \$LOCAL) \$(cygpath -w \$REMOTE)"17:18
but this second change hasn't stuck17:18
← slonopotamus left17:18
ashirus and it's still calling the original difftool cmd17:18
freq9 fr0sty: wow :) thanks17:18
ashirus (which doesn't work)17:18
how is this possible?17:19
← didi left17:20
→ Rogi joined17:21
→ Deesl joined17:21
← sftp left17:23
→ sftp joined17:23
← sivy left17:23
fr0sty ashirus: change the second one from 'p4merge' to 'echo'17:24
see if the shell is doing what you expect.17:24
→ normanrichards joined17:24
→ rox` joined17:24
← airborn left17:25
→ sivy joined17:26
ashirus fr0sty: it launches vimdiff!17:26
← galderz left17:26
← ayanes left17:26
← akamike left17:27
ashirus fr0sty: specifically, it reports 'git config option diff.tool set to unknown tool: echo'17:27
→ priidu joined17:27
fr0sty no, change the p4merge.cmd = 'echo \$(cygpath...'17:29
← icwiener left17:29
← freq9 left17:31
← GarethAdams|Work left17:31
→ icwiener joined17:31
← spearce left17:31
ashirus fr0sty: done so, it ignored the echo settings and still tries to launch p4merge17:32
← shruggar left17:32
← kalusn left17:32
→ patabugen joined17:32
ashirus fr0sty: did you expect it to each to the shell? I only wish! It's using the old invalid p4merge config from somewhere17:32
← sftp left17:33
→ sftp_ joined17:33
→ Evious joined17:34
ashirus diff.tool is always an alias that points to diftool.XXX .cmd, right?17:35
fr0sty what command are you using to invoke the difftool?17:35
ashirus git difftool17:35
← vadim3 left17:36
fr0sty and 'git difftool -tp4merge' launches vimdiff?17:38
→ spearce joined17:38
patabugen Hey :) I'm getting "fatal: The remote end hung up unexpectedly" and can't figure for the life of me why it would say that. I can ssh no problem, but git clone ssh://.... fails (this is why Is started using Dropbox and had my problem of yesterday)17:39
← spearce left17:39
ashirus fr0sty: no, it launches p4merge but with invalid parameters - $LOCAL instead of $(cygpath $LOCAL etc.17:39
→ spearce joined17:40
ashirus so p4merge exists with an error17:40
rasztasd how to tell what is the last common commit of two commits or branches?17:40
ashirus (as it doesn't understand linux paths, cygpath converts them to windows paths that p4merge understands)17:40
doener rasztasd: git merge-base17:41
rasztasd doener: ty17:41
→ kenleycapps joined17:42
fr0sty ashirus: and if your difftool.p4merge.cmd="echo ..." it still executes p4merge?17:42
ashirus fr0sty: yes!17:42
fr0sty set diff.tool=foo difftool.foo.cmd=blah... and git difftool -t foo17:43
just for grins...17:43
Kobaz how do i revert a single file to what was in the last committed state17:43
fr0sty otherwise I'm out of ideas.17:43
FauxFaux Kobaz: git checkout HEAD file17:43
Kobaz k17:43
ashirus fr0sty: "Unknown merge tool foo"17:44
this is how I set it: git config --global diff.tool=foo17:45
is that wrong syntax or something?17:45
→ mastro joined17:45
FauxFaux What's foo?17:45
fr0sty ashirus: does git config -l show diff.too.=foo ?17:46
ashirus fr0sty: no17:46
fr0sty: it shows diff.tool=echo (from our experiment 2 mins ago)17:46
fr0sty git config --global --set diff.tool=foo17:46
ashirus fr0sty: but it should show diff.tool=foo17:46
fr0sty --add, not set17:46
→ warthog9 joined17:48
fr0sty ashirus: just edit ~/.gitconfig directly.17:49
← robgleeson left17:49
ashirus fr0sty: I have done! and it's still ignoring those settings! I'm going mad!17:50
forget it. i'm just going to reinstall cygwin, git, p4merge17:52
← whatdoesitwant left17:53
← davzie left17:53
← kukks left17:53
fr0sty a wise move.17:53
ashirus fr0sty: :) thanx17:53
fr0sty: for your help17:54
fr0sty reinstall the OS while you are at it.17:54
Ubuntu is nice ;-)17:54
→ foocraft joined17:54
ashirus lol17:54
→ sftp joined17:54
← sftp_ left17:54
← rasztasd left17:56
ashirus fr0sty: hmm, apprently it may have not been the settings that weren't getting read17:56
If I do:17:56
git difftool -x "p4merge \`cygpath -w \$LOCAL\` \`cygpath -w \$REMOTE\`"17:56
→ robgleeson joined17:56
← eletuchy left17:56
ashirus it still fails - the backticked parts don't seem to be executing properly17:56
→ hobodave joined17:57
← imMute left17:58
fr0sty not sure how backticks work in cygwin.17:59
why don't you write a wrapper-script and execute that?17:59
rather than trying the shell magic?17:59
ashirus fr0sty: well my problems started when I tried to invoke a wrapper script18:00
← cenuij left18:00
← dr_win left18:01
→ leadZERO joined18:01
ashirus fr0sty: I've finally got it to work via git difftool -x ./git-diff-bcomp-wrapper.sh18:01
leadZERO Anyone have issues using Git for Windows with Symantec products?18:01
ashirus fr0sty: so I'm going to alias that and forget the real settings18:01
thx18:01
← ignatp_ left18:02
adamholt → adamholt_away18:04
adamholt_away → adamholt18:04
→ mlq joined18:04
← mlq left18:04
→ mlq joined18:04
→ eletuchy joined18:04
→ dr_win joined18:05
← ShooterMG left18:05
→ thiago_home joined18:06
← dpino left18:06
← priidu left18:07
→ carpet_the_walls joined18:07
fr0sty no problem. glad you got a solution.18:07
← workmad3 left18:07
→ priidu joined18:09
→ LongBeach joined18:09
→ PauLej joined18:10
→ imMute joined18:10
← apsysoev left18:10
← rox` left18:10
→ remur_030_v2 joined18:10
remur_030_v2 can I ask questions about msysgit here too? My name contains an ö (o with umlaut), when I try commiting it complains about non-utf8 commit message18:11
→ q0rban joined18:12
remur_030_v2 I found a mail where modifying the name fixed it, so I replaced it with the international oe instead of ö, but utf-8 is perfectly capable of ö, how can I submit my 'real' name? =)18:12
← leadZERO left18:12
→ bauruine joined18:13
← chronos left18:13
→ rox` joined18:14
→ chronossc joined18:14
← chronossc left18:14
IslandUsurper remur_030_v2, can you tell what encoding the config file that has your name is in?18:14
I bet it's not UTF-818:14
remur_030_v2 IslandUsurper: I don't know about a config file? I just set it with 'edit -> options'18:15
it's a plain textbox18:15
← nicoulaj left18:15
← hyperair left18:15
← erez left18:15
IslandUsurper err, maybe I'm confused. I thought msysgit was mostly CLI, not GUI18:16
remur_030_v2 IslandUsurper: that's the msysgit bash, I have no idea how to set it there though18:17
I also can't find any configuration directory/file in my user directory18:17
though there propably is one somewhere =/18:18
IslandUsurper remur_030_v2, it should be called .gitconfig18:18
→ g0bl1n joined18:18
getxsick hi18:18
i cloned a repo, did some changes and wanted to push it back but i got this: http://paste.pocoo.org/show/285630/18:19
FauxFaux I'm trying to demo something and I'd like to EDITOR='grep -v foo' git rebase -i HEAD~~ #, or sed -i.. it seems that if the editor isn't interactive rebase does.. nothing? Apart from printing the (edited) rebase message to stdout.18:19
fr0sty getxsick: pushing to non-bare repositories is discouraged18:19
FauxFaux Oh utter penis. I'd done sed. That makes more sense.18:20
fr0sty FauxFaux: GIT_EDITOR you mean?18:20
FauxFaux No. It works fine with sed -i.18:21
getxsick fr0sty: 'git init' initializes non-bare repo, right?18:23
FauxFaux Yes.18:23
fr0sty yep: git init --bare, on the other hand, does not.18:23
getxsick so, in short summary "hub" repos should be bare ones, right?18:24
→ andrewwatts joined18:24
remur_030_v2 IslandUsurper: ah found it, well it obviously uses unix linebreaks, I opened it with vi and did :set enc=utf-818:24
fr0sty getxsick: yep.18:25
→ __name__ joined18:26
← __name__ left18:26
→ __name__ joined18:26
remur_030_v2 ah and saved, now it commits fine but on the 'git bash' it displays the ö as <C3><B6> which is proper utf818:26
code but not symbol?18:26
← abetaha left18:26
← petran left18:26
IslandUsurper remur_030_v2, might need to use a TrueType font18:27
→ abetaha joined18:27
→ erez joined18:28
remur_030_v2 IslandUsurper: ah i think you are right, the git gui displays it correctly in the log, thanks alot for the help!18:28
IslandUsurper no problem18:28
→ Heimidal_ joined18:28
← g0bl1n left18:30
← rox` left18:31
→ xiong joined18:31
→ mrenouf joined18:31
→ phantomcircuit joined18:32
→ hyperair joined18:32
← Heimidal left18:32
← Deesl left18:32
mrenouf If I only have a remote branch and not a local tracking branch, do I actually have all the revisions when I fetch, or just a pointer to the commit-ish of the remote branches?18:32
→ Deesl joined18:32
→ nadim joined18:32
fr0sty mrenouf: you have all of the history18:32
mrenouf fr0sty: ty18:33
wereHamster mrenouf: a pointer to a commit-ish implies all its history18:34
→ wonder95 joined18:34
wereHamster (unless you have a shallow clone)18:34
← kenleycapps left18:35
remur_030_v2 uh now that came up, so when I clone for example the kernel.git i pull all the history right? sounds like quite a large bulk for me18:35
→ rox` joined18:35
mrenouf remur_030_v2: smaller than you might expect18:35
though I don't have stats18:35
→ ymasory joined18:37
← chrisred left18:38
remur_030_v2 ok, but I figured stuff like binary firmware etc are hard to track in history as you can barely 'diff' them, so any binary patch will likely increase the clone by it's size right?18:38
wereHamster remur_030_v2: correct18:38
remur_030_v2: try to only put text files into git18:38
mrenouf remur_030_v2: doesn't git handle that efficiently?18:38
git only sees bits right? it stores deltas regardless.18:39
remur_030_v2 wereHamster: will keep it in mind then =P thanks for the infos18:39
← lanthan_afh left18:39
IslandUsurper mrenouf, git doesn't store deltas18:39
(much)18:39
← justin-george left18:39
mrenouf IslandUsurper: you mean, of binaries?18:40
← PauLej left18:40
IslandUsurper I mean it hardlinks blobs, and computes the diffs when it needs to18:40
mrenouf ok, I will like to look more closely some time. So far have steered clear of the "plumbing" ;-)18:42
← spaceonline left18:43
→ Fisherman joined18:44
Fisherman Hello18:44
→ sandman06 joined18:44
Fisherman How do I specify what ssh key to use when pushing?18:44
I can't start ssh-agent, for some reason18:44
→ andyburke joined18:45
→ FsTluxe joined18:45
cirwin Fisherman: edit your .ssh/config18:46
→ krosenvold_ joined18:46
← tweakism left18:46
← R9dovoj left18:46
→ goshawk joined18:49
→ jannau joined18:49
mrenouf When I use cherry-pick from a work tree to an another branch, and later merge that branch back into my work-tree, it's normal to see the same commits echoed right? Is there any cleaner way to handle that? I want to merge changes in but only push out selected commits18:49
Fisherman thank you cirwin18:49
jannau hi18:49
← bluenovember left18:50
mrenouf I'm using git-svn and I have lots of local specific-to-me changes made I want to track, but not push back to svn.18:50
so I've been using a work flow of merge-in cherrypick-out between a private branch and master18:50
jannau is there a way to limit :/"search for string in commit message" to commits reachable from the current branch18:51
→ MikhX joined18:51
← jdav left18:53
→ SianaGearz joined18:55
← tatsuya left18:56
← sandman06 left18:57
← rox` left18:57
→ jdav joined18:59
avar it should do that already I think, unless you have --all19:00
but git log :/foo branch I think19:00
← Heimidal_ left19:01
→ rox` joined19:02
← hobodave left19:03
→ plediii joined19:03
→ joel_falcou joined19:03
→ saschpe joined19:03
← saschpe left19:04
← psoo left19:04
→ saschpe joined19:04
← __name__ left19:04
← Fisherman left19:07
← curtana left19:07
patabugen Hey19:07
I'm having some troubles :(19:07
I have two copies of the same repos on the same branch but with changes which need merging19:07
One of them has a corrupt .git19:08
I'm stuck :( What can I do?19:08
FauxFaux Corrupt in what way? What did you run and what did it say?19:08
patabugen I was in here yesterday and the conclusion was that Drobox had broken it19:09
Hang on, let me get the error19:09
$ git status fatal: unable to read tree 792545082113cf2a182b698adafcf0019ed78b4919:10
← eletuchy left19:10
patabugen The same if I try to commit19:10
FauxFaux: This is the output from fsck --full http://pastebin.ca/198090919:11
FauxFaux Ouch. Have you tried copying the contents of working-repo/.git/objects into broken-repo/.git/objects?19:11
→ programble joined19:11
← programble left19:11
→ programble joined19:11
patabugen I haven't no19:11
Should I just write those which don't exist and ignore those which do?19:12
FauxFaux I'd copy everything. Of course, you've done full backups of both so it doesn't matter what happens.19:12
patabugen When you say copy everything, you mean replace those which were there already?19:12
→ s0n-of-a-gun joined19:12
→ sgh_ joined19:12
FauxFaux Yeah.19:13
doener git uses content addressable storage: same name => should have the same content, otherwise one version is corrupted19:13
→ s1akr joined19:13
s1akr NOTICE: FREENODE IS CHANGING HOW YOU CONNECT TO IT! At 10pm EST tonight Freenode will enforce SASL for ALL connections! If you don't fix your client to use SASL, YOU WILL BE UNABLE TO CONNECT TO FREENODE! PLEASE JOIN #FREENODE AND/OR MSG A STAFFER FOR DETAILS! s1akr sgh_ s0n-of-a-gun programble saschpe joel_falcou plediii rox` jdav SianaGearz MikhX jannau goshawk krosenvold_ FsTluxe andyburke ymasory wonder95 nadim Deesl hyperair phantomcircuit mrenouf xiong er19:13
NOTICE: FREENODE IS CHANGING HOW YOU CONNECT TO IT! At 10pm EST tonight Freenode will enforce SASL for ALL connections! If you don't fix your client to use SASL, YOU WILL BE UNABLE TO CONNECT TO FREENODE! PLEASE JOIN #FREENODE AND/OR MSG A STAFFER FOR DETAILS! disappearedng kipras kenichi jherdman ph^ kumbayo irc2samus Torsten tg aah warbee eno elmob dels wsclock yngress alvaro_o_ ashirus alin-gym nuoHep mklappstuhl dwmw2_gone ochosi jmil bcardarella coderdad O19:13
NOTICE: FREENODE IS CHANGING HOW YOU CONNECT TO IT! At 10pm EST tonight Freenode will enforce SASL for ALL connections! If you don't fix your client to use SASL, YOU WILL BE UNABLE TO CONNECT TO FREENODE! PLEASE JOIN #FREENODE AND/OR MSG A STAFFER FOR DETAILS! ribasushi mika sitaktif_ wooli SethRobertson Stx candybar Moo^_^ mfoemmel rickmasta cxreg kandjar frerich6 BlindOracle Nickeeh kupsch gcl vmiklos ndim frakturfreak LetoThe2nd zain Tac 18VABQ1HF jonmasters19:13
← s1akr left19:13
→ thierryp joined19:13
← rox` left19:13
→ tildeequals joined19:13
← saschpe left19:13
patabugen yay! That seems to have worked!19:13
remur_030_v2 → remur_03019:14
patabugen Well git status tells me something at least19:14
→ saschpe joined19:14
FauxFaux You probably want fsck --full to pass too.19:14
← goshawk left19:14
patabugen A dangling commit and some dangling blogs19:14
FauxFaux That's a pass.19:14
patabugen yay19:14
git gc will clean that up?19:15
FauxFaux Yep. git repack -adf # is probably a good idea, too.19:15
→ goshawk joined19:15
→ Goldfisch joined19:15
Goldfisch I just finished getting rid of some old, remote branches using git push origin :<branch_name>. Now, on another site, I want to fetch this update of deleted branches, but I can't seem to find the command to do this.19:16
← goshawk left19:16
← coderdad left19:16
→ Dominian` joined19:16
Dominian` NOTICE: FREENODE IS CHANGING HOW YOU CONNECT TO IT! At 10pm EST tonight Freenode will enforce SASL for ALL connections! If you don't fix your client to use SASL, YOU WILL BE UNABLE TO CONNECT TO FREENODE! PLEASE JOIN #FREENODE AND/OR MSG A STAFFER FOR DETAILS! Dominian` Goldfisch saschpe tildeequals thierryp sgh_ s0n-of-a-gun programble joel_falcou plediii jdav SianaGearz MikhX jannau krosenvold_ FsTluxe andyburke ymasory wonder95 nadim Deesl hyperair phant19:16
← Dominian` left19:16
→ coderdad joined19:16
← ashirus left19:16
← s0n-of-a-gun left19:17
→ goshawk joined19:18
wereHamster Goldfisch: git remote prune19:18
or delet the branches manually19:18
hyperair someone please +r this channel please19:18
i don't want to keep getting pinged by these morons19:18
→ ca1 joined19:18
ca1 NOTICE: FREENODE IS CHANGING HOW YOU CONNECT TO IT! At 10pm EST tonight Freenode will enforce SASL for ALL connections! If you don't fix your client to use SASL, YOU WILL BE UNABLE TO CONNECT TO FREENODE! PLEASE JOIN #FREENODE AND/OR MSG A STAFFER FOR DETAILS! ca1 goshawk coderdad Goldfisch saschpe tildeequals thierryp sgh_ programble joel_falcou plediii jdav SianaGearz MikhX jannau krosenvold_ FsTluxe andyburke ymasory wonder95 nadim Deesl hyperair phantomcircu19:18
← ca1 left19:18
ChanServ set mode: +o19:18
← programble left19:18
wereHamster hyperair: how do I do that?19:18
→ tvw joined19:18
→ gunbert` joined19:19
gunbert` NOTICE: FREENODE IS CHANGING HOW YOU CONNECT TO IT! At 10pm EST tonight Freenode will enforce SASL for ALL connections! If you don't fix your client to use SASL, YOU WILL BE UNABLE TO CONNECT TO FREENODE! PLEASE JOIN #FREENODE AND/OR MSG A STAFFER FOR DETAILS! gunbert` tvw goshawk coderdad Goldfisch saschpe tildeequals thierryp sgh_ joel_falcou plediii jdav SianaGearz MikhX jannau krosenvold_ FsTluxe andyburke ymasory wonder95 nadim Deesl hyperair phantomc19:19
NOTICE: FREENODE IS CHANGING HOW YOU CONNECT TO IT! At 10pm EST tonight Freenode will enforce SASL for ALL connections! If you don't fix your client to use SASL, YOU WILL BE UNABLE TO CONNECT TO FREENODE! PLEASE JOIN #FREENODE AND/OR MSG A STAFFER FOR DETAILS! nullvariable necromancer disappearedng kipras kenichi jherdman ph^ kumbayo irc2samus Torsten tg aah warbee eno elmob dels wsclock yngress alvaro_o_ alin-gym nuoHep mklappstuhl dwmw2_gone ochosi jmil19:19
← gunbert` left19:19
← andyburke left19:19
← wonder95 left19:19
SianaGearz wtf is this scam?19:19
hyperair wereHamster: /mode #git +r, i think19:19
tg fuk these spammers19:19
wereHamster mode #git +r19:19
wereHamster set mode: +r19:19
nullvariable wtf19:19
→ TomyLobo joined19:19
ChanServ set mode: -o19:19
nullvariable legit messages like that come from system notices...19:20
hyperair wereHamster: thanks.19:20
nullvariable sighs19:20
ymasory what's this scam about. how does joining #freenode make anyone money?19:20
hyperair nullvariable: i think real messages come via /notice19:20
→ CoinOpeBoy joined19:20
FauxFaux I suspect they're just trying to annoy the hell out of the staff.19:20
hyperair that's probably it, yeah19:21
FauxFaux Social engineering DDOS.19:21
ymasory i'll bet it's working19:21
nullvariable hyperair: right, from a system account or marked as a system wide message...not from someone join/parting a chan19:21
→ Adys joined19:21
← sgh_ left19:21
→ sgh_ joined19:21
hyperair nullvariable: ah sorry i misread your message.19:21
nullvariable FauxFaux: lol not a bad way to describe it actually... though aren't most people using IRC smart enough to notice such things?19:22
SianaGearz and then trailing junk to try to ward off filters.19:22
nullvariable hyperair: no worries :)19:22
FauxFaux nullvariable: < Dogbert> Imagine someone of average intellegence. Now, think. Half of the population are stupider than that.19:22
nullvariable I can't think it would be hard to write a script that booted them for saying more than a certain number of nicks in a single line at one time...19:23
SianaGearz oh it's not junk. it's uselist to trigger highlights19:23
*user19:23
remur_030 nullvariable: they get axed pretty quick ...19:23
FauxFaux It's even worked on us; see? OFF TOPIC KICK THE OFFTOPICERS.19:23
← tvw left19:23
nullvariable SianaGearz: yep :) that's why I'm here... was busy coding19:23
hyperair was busy digging into zram code19:23
SianaGearz hey so was i.19:23
nullvariable FauxFaux: lol19:23
hyperair the goddamn module refused to initialize them zram devices19:23
nullvariable so apparently we all needed an excuse to come hang out?19:24
hyperair apparently.19:24
SianaGearz hello there hyperali, i can see you on another network.19:24
nullvariable hyperair: I don't envy your pain19:24
SianaGearz ehr hyperair19:24
tab laziness fail19:24
hyperair nullvariable: i've figured it out ^_^19:24
and this concludes my first kernel hacking venture19:25
i think19:25
SianaGearz \o/ gratz19:25
nullvariable nice19:25
hyperair thanks =)19:25
well, first kernel hacking venture aside from solving merge conflicts and random bitrot19:25
→ workmad3 joined19:26
SianaGearz uhm i'm kinda scared to touch stuff there. call me crazy but i felt less uncomfortable tearing apart gcc. not that i got far there, i didn't.19:27
← thierryp left19:27
→ Fazer2 joined19:28
hyperair SianaGearz: actually so am i.19:28
but this is compcache19:28
it's a fairly simple thing that doesn't go really deep19:28
Fazer2 hey, any msysgit developers here?19:28
remur_030 Fazer2: just shoot the question19:29
Fazer2 can I have stable "Git for Windows" installed and test development Msysgit on the same system?19:30
jannau avar: no, doesn't work. 'git log :/foo $branch' shows the first matching commit on $other_branch and the top commit in $branch19:30
avar dunno, what do the docs say about what ;/ should do?19:31
jannau and man gitrevisions explicitly states that it will show the first commit reachable from any ref19:31
Fazer2 because I'm trying to find the cause of Git GUI freeze, I tried it on virtual machine, but can't reproduce it there19:31
jannau avar: it works as documented19:32
avar right19:32
← goshawk left19:32
← q0rban left19:34
→ MikhX_ joined19:36
← jherdman left19:36
→ jherdman joined19:37
← kandjar left19:39
← MikhX left19:39
MikhX_ → MikhX19:39
→ pantsman joined19:39
← Goldfisch left19:40
← krosenvold_ left19:42
← GeertB left19:43
← necromancer left19:44
→ justin-george joined19:44
← Chaser left19:44
← patabugen left19:46
→ kukks joined19:46
← workmad3 left19:47
→ alus joined19:47
alus I just committed something but I have not pushed it19:47
how do I undo that commit?19:47
dwmw2_gone git reset --hard HEAD^19:48
and it's really gone19:48
← thiago_home left19:48
← parasti left19:48
alus what do you mean "really gone"?19:48
remur_030 deleted19:48
dwmw2_gone the change you made are reverted.19:48
remur_030 left no trace19:48
alus are the files still modified on disk19:48
dwmw2_gone no19:48
remur_030 no19:48
alus well that doesn't undo the commit, that undoes the commit and all my work19:49
how do I just undo the commit?19:49
remur_030 revert and commit#19:49
dwmw2_gone CRAPCOMMIT=$(git rev-parse HEAD) ; git reset --hard HEAD^ ; git show $CRAPCOMMIT | patch -p119:49
remur_030 git revert bams19:49
dwmw2_gone there are probably better ways :)19:49
alus remur_030: that would add a second commit19:49
remur_030 ah right, leave no trace, dwmw2_gone suggestions looks good then =)19:50
otherwise git is written so you can't modify history19:50
→ airborn joined19:51
alus that doesn't make sense here. nothing was pushed19:51
← aah left19:51
jannau after looking at get_sha1_oneline() the only way make it do what I want is patching19:51
alus there are certainly some files on disk which could be modified to put things back the way they were19:52
doener alus: git reset --soft HEAD^19:52
alus: moves the branch head back, but doesn't touch the index or working tree19:52
alus doener: ok, but not "git diff" says nothing19:53
er, s/not/now/19:53
doener alus: sure, cause that diffs between the working tree and index, and they're still the same.19:53
alus so how do I undo that too19:53
remur_030 alus: how deep in is the history is the faulty commit? you'll loose the history after that19:53
doener alus: git reset19:54
remur_030 you sure you can't just revert the single commit? is a clean history that important?19:54
alus ok, "git reset --soft HEAD^ && git reset"19:54
doener alus: you just asked for undoing the commit, not the staging19:54
alus doener: I only typed one command19:54
doener: I want to undo that one command19:54
doener alus: that or just "git reset HEAD^" (but _not_ "git reset --soft HEAD^ && git reset HEAD^")19:54
alus ok19:54
doener alus: and didn't even tell us which command that was...19:55
alus doener: oh, true. by "I just committed something" I meant "git commit -a -m 'blah'"19:55
doener alus: that's "stage all modified files and commit", not just "commit something"19:56
alus doener: and "git commit foo.txt -m 'blah'"?19:57
is stage that one file and commit?19:57
doener alus: no, that's "create and use a temporary index, starting out equal to HEAD, stage foo.txt, commit, drop the temporary index"19:58
alus muh wha19:58
well, will 'git reset "HEAD^"' still undo that?19:59
→ yano joined19:59
doener alus: "git commit foo.txt" will commit only the changes to foo.txt, even if you already added e.g. "bar.txt" to the index19:59
alus right, ok19:59
doener alus: that and possibly more. The index will be clean, so anything you had added to the index is reverted, too19:59
alus yikes19:59
so is there a command I can use it either situation?19:59
doener no20:00
alus ugh20:00
so how would I undo that one?20:00
← pantsman left20:00
doener which one? "edit foo and bar; git add bar; git commit -m yadda foo;"?20:01
→ ChrisPitzer joined20:01
doener git reset HEAD^; git add bar20:01
alus hm20:01
but that's only if I used git add originally20:02
doener the index state is "destroyed" by the reset, so you have to reconstruct it, if you really want to undo _exactly_ the stuff that commit command did20:02
→ eletuchy joined20:02
alus I call that redo, not undo ;)20:02
ChrisPitzer hey. I need to roll back one commit. how do I do that?20:02
alus but I never do "git add" unless I'm adding a new file, so "edit foo and bar; git commit -m yadda foo; git reset HEAD^;" works fine to preserve changes in bar on disk?20:02
doener alus: sure, a --mixed reset (the default) won't touch the working tree20:03
alus ok, cool20:03
ChrisPitzer: good luck!20:03
doener: thanks20:03
← alus left20:04
← jrmuizel left20:04
← cannonball left20:04
→ ceez joined20:04
→ brendon_work joined20:04
remur_030 ChrisPitzer: you can do 'git revert'20:04
ChrisPitzer remur_030: how do I specify what commit i want to revert to?20:05
→ jrmuizel joined20:05
remur_030 ChrisPitzer: you checked the man page by chance? revert only reverts a single commit...20:05
← jmil left20:09
cbreak revert doesn't really revert a commit, it creates an undo commit20:09
the only way to properly change published history20:09
there are other methods such as git rebase -i and other history rewriting tools20:10
but those are as said for rewriting history. with all the consequences.20:10
← ph^ left20:13
← Torsten left20:14
→ ph^ joined20:15
→ schacon joined20:19
schacon is there any good crlf guide around?20:19
→ rendar joined20:21
rendar can i generate a myproject.tar.gz of the current snapshot of a branch?20:21
cbreak yes20:22
rendar how cbreak?20:22
cbreak now you're asking difficult questions :)20:22
rendar: man git-archive20:23
rendar :P20:23
thanks20:23
← ehabkost left20:25
→ golodhrim joined20:28
→ Eridius joined20:30
→ f0rkit joined20:32
← f0rkit left20:33
→ Sigma joined20:34
← marcob left20:34
← Deesl left20:40
← dl9pf_ left20:42
← irc2samus left20:43
← doener left20:47
bremner snapshot = "!g() { cdup=`git rev-parse --show-cdup`; repo=` cd $PWD/$cdup && pwd -P` ; sha=`git rev-parse --short HEAD`; git archive --format=tar HEAD | gzip -9 > $repo/../`basename $repo`-${sha}.tar.gz; } ; g"20:49
rendar: there is the alias I use ^20:49
→ kenneth_reitz joined20:52
← zombor left20:52
← kenneth_reitz left20:52
← abetaha left20:52
← schacon left20:52
→ kenneth_reitz joined20:53
cbreak I use this: zip = ! BN=$(basename -s .git $(git rev-parse --show-toplevel)) && git archive --format=zip --prefix=$BN/ HEAD > $BN-$(git id).zip20:54
id = describe --tags --long --abbrev=6 --dirty="*"20:55
→ workmad3 joined20:55
← jaql left20:56
← pmcnabb left20:57
← Pewpewarrows left20:57
← fr0sty left20:58
← workmad3 left20:58
← dr_lepper left20:59
→ ymasory_ joined20:59
← sebersole left21:00
← jherdman left21:01
→ fr0sty joined21:06
→ workmad3 joined21:06
← sonnym left21:07
→ sgh joined21:07
← sgh_ left21:08
← boombatower left21:08
← rgr left21:10
← Fazer2 left21:11
← Elfe left21:11
→ Elfe_ joined21:11
← LiamH left21:13
ChanServ set mode: +o21:13
wereHamster set mode: -r21:13
← alex__c2022 left21:13
ChanServ set mode: -o21:14
← workmad3 left21:16
← sgh left21:17
← dukebag left21:17
→ BiggFREE joined21:17
← joel_falcou left21:17
→ sgh joined21:18
← flazz left21:21
← priidu left21:22
← bcardarella left21:23
← elmob left21:24
→ krosenvold_ joined21:25
→ dukebag joined21:26
→ unhygienix joined21:28
→ ProperNoun joined21:28
← ProperNoun left21:28
→ ProperNoun joined21:28
← nuoHep left21:28
→ enherit joined21:29
← Evious left21:29
→ Evious joined21:30
← warthog9 left21:31
← bentob0x left21:31
← IslandUsurper left21:32
← ymasory_ left21:33
→ sgh_ joined21:34
← sgh left21:34
← dukebag left21:35
→ q0rban joined21:36
← nullvariable left21:37
← mfacenet left21:37
→ hefox joined21:39
→ warthog9 joined21:39
← hefox left21:39
foocraft hey all21:39
→ night_shade joined21:39
foocraft I was wondering if there's a workflow that supports "hacky" behavior :p21:40
← welly left21:40
foocraft like, take my code, do a lot of things for a few hours, and be able to revert back to the old version if things don't work out well, while letting the "current" version for others that pull from the repo be that version that I fell back to.21:40
→ sonnym joined21:41
wereHamster sure21:41
→ cannonball joined21:41
crash\ branches?21:41
wereHamster create a branch or not, use stash, or commit along the way.. your choice, really21:41
foocraft (that word scares me)21:41
PerlJam sounds more like branchy workflow to me21:41
foocraft branches is something I never understood completely, probably because it sounds unbelievable21:42
← cirwin left21:42
sgronblo foocraft: what you just described sounds exactly like the workflow you get when using branches21:42
foocraft so I "branch" so some work, and suppose I am successful in adding the new feature, how can I set my bare repo to contain this branch21:42
wereHamster foocraft: you push your branch into that repo21:43
sgronblo you can merge it back into your master branch and then push21:43
or you can push the branch to the repo first if you want others to have access to it21:43
← night_shade left21:43
foocraft so branch optimize_queue ; stuff ; git commit -a ; git push21:43
s/branch/git branch/21:44
sgronblo there must be some good guide online for this21:44
some details are missing from your suggested procedure21:44
→ ignatp_ joined21:45
foocraft sgronblo: wanna pastebin a sketch of a possible sequence of git commands?21:46
(that would be highly appreciated)21:46
sgronblo foocraft: well I guess I could, but I just think you could probably find much more detailed guides by googling for something like git topic branches21:46
foocraft I did, but they seem to be very...ugh21:47
I mean, I don't get branching for some reason21:47
it seems all magical heh21:47
sgronblo its quite simple once you get it21:47
← hyperair left21:47
→ thierryp joined21:47
foocraft (a bit of a future question: are hooks per branch or per repo?)21:48
← thierryp left21:48
→ kar8nga joined21:49
→ thiago_home joined21:50
sgronblo http://pacific.mpi-cbg.de/wiki/index.php/Git_topic_branches here is one article that was in the first 10 hits on google21:50
if it scares you atm, i suggest you read more than one21:50
PerlJam foocraft: per repo, but the hook has enough information to distinguish branches and such.21:51
→ darkredandyellow joined21:51
foocraft sgronblo: thanks a lot21:52
→ s0ber_ joined21:52
sgronblo personally i create a new branch for nearly any changes i start making unless its a very simple repository or if im in some "prototyping stage" of a project21:53
→ dissidentss joined21:53
→ marienz joined21:54
darkredandyellow guys, I am getting a: "warning: remote HEAD refers to nonexistent ref, unable to checkout." when doing a "git clone....".21:54
Now, I know _why_ this is: "git branch -a" shows: remotes/origin/HEAD -> origin/master21:54
but as far as googling / looking at "git remote set-head" goes I am more confused than before21:55
← s0ber left21:55
darkredandyellow am I right: I can NOT set the HEAD on the repo.21:55
wereHamster darkredandyellow: is the remote repo empty?21:55
← sivy left21:55
darkredandyellow * remote repo21:55
wereHamster, no, there are some branches, but no master21:55
which is on purpose21:56
← icwiener left21:56
wereHamster darkredandyellow: then go into that repo and make sure HEAD points to one of those21:56
→ thierryp joined21:56
darkredandyellow wereHamster, I C. unfortunately it's on sourcerepo.com, so I'll have to drop them email.21:56
wereHamster, thanks for the clarification21:56
← ignatp_ left21:59
← thierryp left22:01
← carpet_the_walls left22:01
→ Atlantic777 joined22:01
← dreiss left22:03
→ hobodave joined22:05
→ Tuller joined22:06
→ abetaha joined22:06
← remur_030 left22:06
→ thierryp joined22:07
Atlantic777 How can I delete some commit messages or even better reset the whole repo to the start?! I did some testing commits and I now wan't to delete them and start seriously.22:08
← sftp left22:08
Eridius you can use `git reset --hard $SHA` to reset the branch back to $SHA, throwing away any commits between $SHA and the tip of the branch22:09
← kipras left22:09
Eridius if you want to just throw away the last 3 commits, you could use `git reset --hard HEAD~3`. Similarly if you want to throw away all local commits you can use `git reset --hard origin/master` (or in more recent gits, `git reset --hard @{u}`)22:09
→ sftp joined22:09
getxsick fr0sty: around?22:11
← ChrisPitzer left22:11
← thierryp left22:12
← fr0sty left22:12
fr0sty___ → fr0sty22:12
fr0sty yes?22:12
→ fr0sty_ joined22:12
→ blahdeblah joined22:12
Atlantic777 Eridius, I'm getting some strange messages about fast forwarding. :/22:12
Eridius Atlantic777: `git reset` doesn't provide any messages about fast-forwarding22:13
Atlantic777 Eridius, I get them when I try `git push` after hard reset.22:13
← saschpe left22:13
Eridius ok, it's telling you that you can't push because the push would not be a fast-forward22:13
which means you threw away commits that you had already pushed22:13
are you certain you want to throw them away on the remote? This changes history and causes problems for anyone else who has pulled from your remote22:13
Atlantic777 yap, that's the point :D22:13
They are just test1.txt, test2.txt etc.22:14
sgronblo the point is to cause problems? :)22:14
← noktoborus_ left22:14
blahdeblah Why does 'git status' return 1 on a clean working directory on v1.5.5.6, yet return 0 on a clean working directory on 1.7.0.4?22:14
fr0sty getxsick: you rang?22:14
Atlantic777 You can see the remote if you don't believe me. :)22:14
Eridius Atlantic777: if you're certain you want to throw them away on the server, you can use the -f flag to git push, or you can prefix the refspec wtih + (if you use -f make sure you provide exactly the refspecs you want, if you push all matching branches it may not do what you intended)22:15
getxsick fr0sty_: according to my problem (A and B companies), if developers from B's company would use one username, that would be easy, right? but is it possible to track somehow who did which changes? like adding some hooks? just to see it in company B, at least not by 'git log' something external22:15
fr0sty_: yeah, sorry, had to phonecall22:15
← lucsky_ left22:15
Eridius blahdeblah: my impression is that `git status` no longer has a useful exit code22:15
blahdeblah Eridius: :-( OK, thanks22:16
← abetaha left22:16
Eridius blahdeblah: usually you want some combination of flags to `git ls-files` to find out about dirty trees22:16
blahdeblah reads man page for git ls-files22:17
← kar8nga left22:17
Atlantic777 Eridius, ok I've removed some files, but there are still commit messages. ( http://git.ns-linux.org/openpeticija.git/ )22:17
→ f0rkit joined22:18
← brntbeer left22:18
Atlantic777 Eridius, and there is still the first file (frontend3.py) :/22:18
blahdeblah What is the difference between cached and committed for the purposes of git ls-files?22:19
← yairgo left22:19
→ sivy joined22:19
wereHamster blahdeblah: do you what the index is?22:20
← kumbayo left22:20
blahdeblah wereHamster: I presume that should be "do you know what..."22:21
wereHamster yes22:21
blahdeblah Well, the answer is yes, for very vague values of yes22:21
I'm still trying to learn git and haven't got the jargon down22:21
jzacsh_ → jzacsh22:22
wereHamster blahdeblah: why don't you start with http://progit.org?22:22
fr0sty getxsick: I still fail to understand any legitimate business reason for throwing away committer information.22:22
→ Blackguard joined22:22
fr0sty is even having a useful 'Author' name not possible?22:23
blahdeblah wereHamster: Because i don't have money to buy books and find mostly that i end up not using them because online resources are more up-to-date?22:23
wereHamster blahdeblah: why would you need to buy anything?22:23
blahdeblah saw the apress logo and assumed they weren't giving out books for free22:24
← warthog9 left22:24
blahdeblah checks now22:24
wereHamster blahdeblah: you can read the book online22:24
or even download its source!22:25
blahdeblah Not sure how that's going to help me interpret the git ls-files man page without reading the whole book first...22:25
← brosner left22:25
wereHamster you'll understand what the index is22:26
and then what --cached means in the context of just about any git command22:26
→ bcardarella joined22:26
fr0sty getxsick: an alternative would be to make up some bogus 'committer' names 'CompanyB-123', CompanyB-124, etc. and go with that.22:26
blahdeblah strange that the man pages for git-status and git-commit don't ever mention the word cache22:27
→ engla joined22:27
fr0sty getxsick: basically, there is no way to retains authorship information while at the same time discarding all authorship information22:27
gotta run...22:27
wereHamster blahdeblah: do they mention 'index'? or 'staging area'?22:27
foocraft I want to start going back in time and branching from there22:28
wereHamster foocraft: git branch <name of you rnew branch> <commit which you want to start that branch at>22:28
getxsick fr0sty: hmm, at least i could just add hooks for commits in B and make external DB to know who did what ;/22:28
foocraft wereHamster: thanks! :)22:28
blahdeblah wereHamster: yes, but i though cached and committed were the same thing22:29
→ Bass10 joined22:29
wereHamster blahdeblah: then you didn't get the concept of the 'index' at all22:29
← Bass10 left22:30
blahdeblah My laptop has man 7 gitglossary, even though the server (v1.5.5.6) doesn't, and it says that cache==index22:30
← hobodave left22:30
wereHamster that's true22:30
← normanrichards left22:30
→ Bass10 joined22:30
engla gitglossary doesn't even mention "stage" here :-(22:31
→ eaven joined22:31
← coderdad left22:31
blahdeblah engla: indeed!22:31
The plethora of terms is rather confusing for us newbies22:32
→ rhl joined22:32
engla ah, gitglossary is not written for newbies anyway. it's not like its explanation of index will enlighten anyone who doesn't already know git innards22:32
eaven → Eaven22:32
→ dreiss joined22:32
rhl Hi, can someone explain to me what this error means? http://fpaste.org/zAaW/22:32
wereHamster rhl: that you nede to commit your changes first or stash them22:32
engla blahdeblah: I can understand that. I say 'staging area' or 'index' for short. you never use 'cache'22:32
← Yuffster left22:33
→ jeff2 joined22:33
rhl wereHamster, I did a git commit -m "message here"22:33
← fowlduck left22:33
blahdeblah engla: OK - thanks22:33
rhl and it seemed to work out alright?22:33
← TeckniX left22:33
wereHamster rhl: what does 'git status' say?22:33
rhl http://fpaste.org/tLTp/22:33
→ alester joined22:34
→ Linc joined22:34
blahdeblah Maybe i would be better stating my problem in more practical terms: i'm trying to write a script that will run as a regular cron job to 1. add unstaged files, 2. commit all changes, 3. pull from remote, and i'm trying to work out how to do 1 in the most sensible manner.22:34
foocraft how can I create a branch locally, and let that branch be visible in the master repo?22:35
so I know how to branch and switch between branches22:35
wereHamster blahdeblah: committing or pulling from a cron script is not recomeded22:35
→ drizzd joined22:35
← pigdude left22:35
foocraft but not how to let the available set of branches visible across all repos22:35
rhl wereHamster, any clue?22:35
blahdeblah wereHamster: That may be, but it's required in this case, so i need to work out how to do it as best i can.22:36
wereHamster rhl: git rm --cached gen-graph.py && git commit -m 'remove that file'22:36
→ sgh joined22:36
wereHamster blahdeblah: so explain, what will you do when there's a conflict?22:36
rhl cool that worked, why did that work though?22:36
← sgh_ left22:36
wereHamster rhl: you had that one file locally modified. YOu needed to commit it22:37
rhl so I removed gen-graph.py so it needed to know that22:37
and I didn't tell it?22:37
wereHamster yes22:37
blahdeblah wereHamster: I'm happy to let it stay conflicted - conflicts will be rare in this repo because most of the time the files which are pulled from origin are not touched, and most of the files in the local repo are not even present in origin.22:37
frogonwheels rhl: if you use git rm <file> that does the file removal and adds it to the index22:38
← drizzd_ left22:38
rhl frogonwheels, aha I see.22:38
frogonwheels rhl: .. which is effectively (I guess) rm <file> && git rm --cached <file>22:38
rhl so I should alias rm to something fancy that checks if .git exists then does the git rm accordingly?22:38
frogonwheels rhl: or just get used to using git rm22:39
rhl frogonwheels, right....22:39
wereHamster rhl: you sholud know yourself whe you are in a git repo and use git rm instead of rm22:39
rhl sure, it's true.22:39
frogonwheels rhl: or alternatively use git gui which then allows you just to click on the removed file to 'stage the change'22:39
← jeff2 left22:39
wereHamster rhl: or use git status and git add -A ...22:39
frogonwheels rhl: many ways to skin that cat depending on how you like to work.22:39
rhl so i'm using git svn22:39
and I have two folders, say A, B22:40
Atlantic777 Ok... I have done somehow the hard reset, but can I even delete the first commit? To get completely clean remote?22:40
rhl folder A is a fork of code from another person, its a huge number of files, B is my new directory, which is just one file right now22:40
→ workmad3 joined22:40
wereHamster Atlantic777: git reset --hard origin/master22:40
rhl when I git svn dcommit, it seems to **always** push folder A, even though nothing changed22:40
isn't that wrong?22:41
← Linc left22:42
Atlantic777 wereHamster, but there's still the first commit. :/22:42
wereHamster Atlantic777: rm -rf .git ?22:42
← akahn left22:42
Atlantic777 wereHamster, maybe... :D22:42
→ warthog9 joined22:43
← f0rkit left22:45
← thiago_home left22:45
→ thiago_home joined22:45
→ _Vi joined22:45
→ marc_kdab joined22:45
← sgh left22:45
→ Schmallon joined22:46
→ hobodave joined22:47
→ amerine joined22:47
→ brothers joined22:48
→ noktoborus_ joined22:48
→ sgh joined22:48
← brothers left22:48
→ brosner joined22:48
→ zombor joined22:49
← zombor left22:49
→ zombor joined22:49
← blahdeblah left22:50
→ brothers joined22:50
→ psoo joined22:50
← dreiss left22:52
← sgh left22:53
← warbee left22:53
← hobodave left22:53
→ rgr joined22:53
→ theoros joined22:54
rgr I have just realised that I removed something from git a long while back using git rm (a dir that I forgot was useful). Can I get it back from git? How do I find it?22:55
← q0rban left22:55
← ph^ left22:56
frogonwheels rgr: you can use man git-log with a path to find out the commit, and then you can checkout the directory from before that (specifying a revision)22:57
→ abetaha joined22:58
← ymasory left22:58
→ ereslibre joined22:59
→ hobodave joined22:59
← hobodave left23:00
→ hobodave joined23:00
← workmad3 left23:02
← kenichi left23:03
← jannau left23:03
rgr thank you23:04
→ f0rkit joined23:04
← abetaha left23:04
→ dreiss joined23:04
← Atlantic777 left23:05
← psoo left23:05
← dels left23:06
→ awilkins joined23:06
→ hal joined23:06
hal what places are user.name setting kept? I am certain that I set this in a previous project but I can't find it anywhere23:07
← Schmallon left23:07
→ ProperN[out] joined23:07
hal in the project it's .gitconfig, isn't it? What if it is a global setting23:07
awilkins There's a global setting for it23:08
← BiggFREE left23:08
→ AAA_awright joined23:08
awilkins The current version of git I had actually prompted me to set it (after unhelpfully assuming it's value for the first commit)23:08
→ joeytwiddle joined23:09
← joeytwiddle left23:09
→ artefon joined23:09
hal awilkins: ah ok, I have found .gitconfig in my home dir23:09
← ProperNoun left23:09
hal thank you awilkins23:09
ProperN[out] → ProperNoun23:09
→ abetaha joined23:10
→ joeytwiddle joined23:10
← Dave^| left23:11
← tizbac_ left23:12
← Arrowmaster left23:12
← alester left23:14
→ marc_kdab_ joined23:14
→ Arrowmaster joined23:15
→ hakunin joined23:15
→ mmc joined23:15
→ jargan joined23:16
← siprbaum left23:16
→ siprbaum joined23:16
→ Heimidal_ joined23:16
→ scarabx joined23:17
← jast left23:17
← marc_kdab left23:18
← Heimidal_ left23:18
→ nullvariable joined23:18
→ checkers joined23:19
checkers I committed a few hacky fixes, now someone has made proper fixes for the problem. I want to commit an 'undo' of my hacky fixes. Is there a simple way to do this?23:19
s0ber_ → s0ber23:19
← mlq left23:19
← hebz0rl left23:20
→ tiglionabbit joined23:21
→ mlq joined23:21
← mlq left23:21
→ mlq joined23:21
tiglionabbit is it possible to get a similar behavior to git svn clone --no-minimize-url in git version 1.5.5.6?23:22
I'm trying to clone a part of an svn repository. The whole repository too large to clone the whole thing23:22
← jrmuizel left23:22
← Tama00 left23:23
→ jaql joined23:24
← sftp left23:25
← f0rkit left23:25
← abetaha left23:26
mklappstuhl bartman: hey.. i'd like to use your dot-files script but i dont get any output from any command23:26
← Ryback_ left23:27
→ sftp joined23:28
→ whatdoesitwant joined23:30
← awilkins left23:31
← kenneth_reitz left23:32
← ajpiano left23:32
← ereslibre left23:34
→ ereslibre joined23:35
← ereslibre left23:35
→ ereslibre joined23:35
→ Tama00 joined23:37
→ cirwin joined23:37
← ProperNoun left23:37
← ceez left23:38
← andrewwatts left23:38
→ jmil joined23:41
→ parasti joined23:42
→ andrewwatts joined23:44
← resmo left23:44
← erez left23:44
← whatdoesitwant left23:45
→ madewokherd joined23:49
← flaguy48 left23:50
→ brntbeer joined23:52
→ ki__ joined23:52
← frogonwheels left23:53
→ frogonwheels joined23:54
adamholt → adamholt_away23:56
adamholt_away → adamholt23:56
→ flaguy48 joined23:56
← Tama00 left23:56
← hobodave left23:57

Logs Search ←Prev date Next date→ Channels Documentation