IRCloggy #git 2012-04-12

Logs Search ←Prev date Next date→ Channels Documentation

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

2012-04-12

maestrojed CareBear\ you rock, got me fixed up. Thanks00:00
← maestrojed left00:01
→ sattu94 joined00:01
← mgpcoe left00:02
→ jalama joined00:02
→ acl_ joined00:03
← satyanash left00:04
← master-lincoln left00:05
CareBear\ maestrojed : you're welcome00:05
← tqmz left00:05
← subh left00:06
→ mischat joined00:07
kaen → Farva00:07
Farva → kaen00:07
→ blackmatrix_ny joined00:08
← kevwil left00:08
← rrn left00:09
← dreiss left00:10
→ rrn joined00:10
← robotmay left00:10
tizzo-afk → tizzo00:12
→ zivester joined00:12
→ slide23 joined00:12
→ kevwil joined00:13
← slide left00:14
← kevwil left00:15
← tvw left00:15
→ nif joined00:20
→ TeckniX joined00:20
← flavius left00:20
→ Bartzy joined00:20
← Chaser left00:25
→ aca20031 joined00:25
aca20031 If I have one file with local changes and for just that one file i want to get the remote versions of it, how can I do that? Will git rm do it or will it actually delete the file from the repository?00:25
← SimonNa left00:26
CareBear\ aca20031 : define remote versions00:27
aca20031 : you can use git checkout commit filename to checkout a file from a previous commit00:27
aca20031 local repository has changes to a file that i dont want, theres also a remote repository called "remote" that i added, i want to replace the file X.java with the same exact file in the remote00:27
← rrn left00:28
→ gchristensen joined00:28
aca20031 i see, git checkout name/of/file.java did it00:29
← gchristensen left00:29
← harish left00:29
→ rrn joined00:29
→ gerard0 joined00:30
aca20031 i get confused as to gits distinction between "push/pull" and "commit/checkout"00:31
kevlarman aca20031: sorry confuse you even more, but pull is not the inverse of push00:32
fetch is00:32
→ macmartine joined00:33
← Bartzy left00:33
aca20031 i hate you a little00:33
lol00:33
← the_drow left00:33
→ Goplat joined00:36
← beatak left00:36
→ macrover joined00:36
← machine2 left00:36
CareBear\ aca20031 : I recommend !book00:37
gitinfo There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !gitt !vcbe and !parable00:37
← bolt left00:37
→ pfFredd joined00:38
← ryden left00:38
← rrn left00:40
aca20031 I should, but im just using it for pet projects not professionally atm so im curerently under motivated to learn it as well as i should00:40
→ cooloney joined00:40
CareBear\ either you want to learn to use the tool or you do not00:41
→ machine2 joined00:41
→ alberto56 joined00:42
CareBear\ if you don't want to learn ("meh" included) then better not use the tool at all00:42
if you do want to learn then do it right, or you're just wasting your time (and ours to some degree)00:42
→ rrn joined00:42
aca20031 yeah yeah00:42
sorry, but thank you.00:42
CareBear\ I'm happy to help, but maybe less so if someone isn't actually interested in learning :)00:42
but it *is* a fair bit to learn!00:43
the best is to just use it00:43
as you are00:43
and read a bit on how to do things you want00:43
aca20031 im interested in learning, just not -everything- at the moment considering my limited usage00:43
CareBear\ sure00:43
aca20031 hard to pick and choose pieces from a book, if i do that im going to learn it in much depth00:43
CareBear\ don't have to read the whole book right away :)00:43
the book is not very heavy00:43
aca20031 if it helps, i googled pull vs fetch00:43
makes sense :)00:43
→ MrJones98 joined00:43
CareBear\ I also recommend man pages to learn about what the different commands do, as well as the details of each come time to go in depth00:44
SethRobertson We even have a trigger on that: !pull00:44
gitinfo pull=fetch+merge (or with flags/config also fetch+rebase). It is thus *not* the opposite of push in any sense. A good article that explains the difference between fetch and pull: http://longair.net/blog/2009/04/16/git-fetch-and-merge/00:44
aca20031 im actually using git bash in windows00:44
no man pages :(00:44
→ fission6 joined00:44
CareBear\ man git00:44
gitinfo the git manpage is available at http://jk.gs/git.html00:44
aca20031 oh nice00:44
bookmarked00:44
SethRobertson git help; git pull help, etc00:44
built in man pages00:44
← ocharles left00:45
SethRobertson I assume even for windows00:45
CareBear\ ass-u-me!00:45
SethRobertson Am I wrong?00:45
→ hobbyist joined00:45
CareBear\ there's not even a termianl00:45
terminal00:45
aca20031 it opens a web brower, so no youre not wrong00:45
← lonewulf` left00:45
SethRobertson git bash *is* a terminal.00:45
aca20031 git pull help opens file:///C:/Program%20Files%20(x86)/Git/doc/git/html/git-pull.html00:45
← rrn left00:46
CareBear\ SethRobertson : which one?00:46
gitinfo set mode: +v00:46
MrJones98 so i am having a really difficult time deleting some files from my history... these were deleted many commits ago and filter-branch --index-filter 'git rm -rf --cached --ignore-unmatch' doesn't seem to get rid of them00:46
→ subh joined00:46
SethRobertson Dunno. I can type commands. I have command line editing. I have perl. Seems terminalish00:46
CareBear\ MrJones98 : it seems like filter-branch should do what you need00:46
MrJones98 CareBear\: I thought so also00:47
CareBear\ SethRobertson : is it running in cmd?00:47
MrJones98 CareBear\: but i've tried it with tree-filter and index-filter and the files are still in my history00:47
→ gchristensen joined00:47
→ rrn joined00:47
← gchristensen left00:48
→ gchristensen joined00:48
SethRobertson CareBear\: Huh? When you ask for git bash you get a bash shell.00:48
MrJones98: How do you know? !repro00:48
gitinfo MrJones98: Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.00:48
MrJones98 SethRobertson: let me clean a session up and paste it for you00:49
CareBear\ SethRobertson : sorry yes, I mean the cmd command interpreter window00:49
← Guest28280 left00:49
SethRobertson MrJones98: Did you use the --all flag to do it to all branches, if you have multiple branches? Did you use --tag-filter cat if you have tags?00:49
MrJones98 SethRobertson: yes, i have used --all00:49
SethRobertson CareBear\: No, typically you have a new option in your explorer shell right button menu which brings up a window which might make you believe for a few blissful moments is on a sane OS00:50
← macrover left00:51
SethRobertson MrJones98: Yes, please pastebin an example command line and output and the method you determined that the file was still in history.00:51
→ dbgster joined00:51
MrJones98 SethRobertson: here's a link https://gist.github.com/236385500:51
SethRobertson: i will work from a new copy of the repos so you can see the output also00:52
→ lonewulf` joined00:52
→ jsutri joined00:52
SethRobertson MrJones98: You don't need to be a new copy. Doing it on the old copy is just fine. For an example SHA1 returned by git-log, please run `git branch --contains SHA1` and `git tag --contains SHA1`00:53
← joaoh82 left00:53
→ joaoh82 joined00:54
→ bolt joined00:54
→ rgwu joined00:54
→ ferdna joined00:55
→ subbyyy joined00:55
MrJones98 SethRobertson: hmm, perhaps it is deleted then since the command doesn't return anything00:55
→ Qoidnef joined00:55
← subh left00:56
SethRobertson MrJones98: Also, running `git checkout SHA1; git rm -rf --cached --ignore-unmatch env/; git status` might be interesting as well.00:56
→ CannedCorn joined00:56
SethRobertson Declare victory and press forward00:56
← pfFredd left00:56
← Mowsh left00:56
← AaronMT left00:57
→ Mowsh joined00:57
→ satyanash joined00:58
← w0ng left00:58
← joaoh82 left00:59
← usmanity left00:59
→ troulouliou_dev joined00:59
MrJones98 SethRobertson: my repository is still extremely large and i found a script online that prints out the largest blobs and the files are still there00:59
CareBear\ man git gc00:59
gitinfo the git-gc manpage is available at http://jk.gs/git-gc.html00:59
→ w0ng joined00:59
troulouliou_dev hi is it normal that after cloning and adding the upstream for github , i type "git fetch upstream master"00:59
← ziggles left00:59
SethRobertson MrJones98: Ah, now that is completely different. Step one: `git clone file:/path/to/repo othername00:59
CareBear\ MrJones98 : try cloning it into a new directory. see if they come along.00:59
troulouliou_dev but git branch -r do not show me the upstram/master branch00:59
← Guest87934 left01:00
CareBear\ troulouliou_dev : git branch -a should definitely show it01:00
SethRobertson MrJones98: That will tell us whether the repo can be shrunk to the size you expect. You could use that repo, but more likely you will want to fix your current one. See the bottom of the man git-filter-branch man page which tells you how to shrink...assuming the git-clone shrinks01:00
gitinfo MrJones98: the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html01:00
← KyleXY left01:00
← sattu94 left01:01
SethRobertson troulouliou_dev: Did you `git remote update upstram`?01:01
→ mgpcoe joined01:01
← mgpcoe left01:01
troulouliou_dev CareBear\, didn't too , i dont want to do "git fetch upstream"" without master caus ethere are tons of branch that i don't want01:01
SethRobertson, why ?01:01
← ColKurtz left01:02
troulouliou_dev when i type git fetch upstream master it says01:02
branch master -> FETCH_HEAD01:02
SethRobertson troulouliou_dev: You cannot really get only one branch. Certainly not under upstam/branchname01:02
Yes, don't use the four argument version of git-fetch or git-pull. It only ends in tears. You *can* use FETCH_HEAD instead of upstram/master01:03
But it will end in tears. It always does.01:03
← Qoidnef left01:04
→ brngardner joined01:04
troulouliou_dev SethRobertson, i have to fetch all the branches ?01:04
→ Qoidnef joined01:04
SethRobertson I recommend it, to avoid tears you know. Are you afraid of namespace pollution?01:05
← MrJones98 left01:05
SethRobertson You could try changing remote.upstream.fetch refspec01:06
You would probably want to also disable tag fetching for that remote as well.01:06
→ ryden joined01:07
CareBear\ ooh I want that01:07
!01:07
← berserkr left01:08
CareBear\ This? remote.<name>.tagopt Setting this value to --no-tags disables automatic tag following when fetching from remote <name>01:08
SethRobertson remote.<name>.tagopt01:08
CareBear\ lovely01:08
SethRobertson I'd like for the git-devs to get tag remotes namespace working. They were toying with it for a while01:08
CareBear\ can I get rid of them after the fact? simply tag -d ?01:09
SethRobertson Sure, if you can distinguish the bad'uns from the good'uns01:09
CareBear\ sure01:09
← dbgster left01:09
CareBear\ the bad'uns are all called r[0-9]+01:09
SethRobertson Actually what would be nice is a refspec like thing for which tags you want to accept01:10
→ barrbrain joined01:10
SethRobertson Or even better, glob or REs. for the fetch spec as well. Dreaming, I know01:10
← Vampire0 left01:11
→ KyleXY_ joined01:11
→ MrJones98 joined01:11
CareBear\ I did git tag|egrep|xargs git tag -d01:11
gitinfo set mode: +v01:11
CareBear\ awesome01:12
thanks01:12
← SegFaultAX|work left01:12
MrJones98 SethRobertson: looks like it might've shrunk01:12
trying it on a clean veersion since i think i might've mucked it up01:12
← whyking_ left01:12
← dhruvasagar left01:13
KyleXY_ → KyleXY01:13
→ dhruvasagar joined01:14
← apok left01:15
→ Guest87934 joined01:15
→ apok joined01:15
→ mwrosen joined01:17
← timoxley left01:18
→ cccaldas joined01:18
→ Vert joined01:22
→ xiangfu joined01:22
→ harish joined01:23
→ NickPresta joined01:24
NickPresta I have a project that is itself, a git repository. It depends on code that is checked into another repository (a library). Does it make sense to manage this dependency using git submodules (the library is a submodule for the main project)?01:24
→ timoxley joined01:24
← borodon left01:24
← lottadot left01:24
→ Vampire0 joined01:25
← fisted_ left01:26
← lonewulf` left01:26
→ Fandekasp joined01:26
→ fisted joined01:26
→ Paradox2000 joined01:27
← troulouliou_dev left01:28
→ eletuchy joined01:29
← gchristensen left01:29
→ brar joined01:29
← dstorrs left01:30
← TeckniX left01:30
← eletuchy_ left01:31
→ maestrojed joined01:32
→ robbyoconnor joined01:32
← illusi left01:37
→ mritz joined01:38
← MrJones98 left01:41
→ setmeaway2 joined01:42
← acl_ left01:42
→ ehks` joined01:43
→ iori joined01:44
← setmeaway left01:45
→ Davey joined01:46
← DrNick left01:46
← el2ro left01:47
← ehks left01:47
← setmeaway2 left01:48
← Davey left01:48
→ joaoh82 joined01:50
→ gchristensen joined01:51
→ Davey joined01:52
← kaen left01:52
← jalama left01:53
← alberto56 left01:53
← zivester left01:54
← CannedCorn left01:55
← ghvandoorn left01:56
← oriba left01:57
→ neurodrone joined01:57
→ setmeaway joined01:58
→ distg joined01:59
→ PigeonFriend joined02:00
← agile left02:01
→ ajaykools joined02:02
→ thursday joined02:02
SethRobertson NickPresta: That is one of your options. See also !subprojects02:03
gitinfo NickPresta: So, you want to add git repositories inside of other git repositories? Well, you have four main options. First is to just do it, add the repo to the outer project's .gitignore, and treat them entirely separately. Best if they are entirely seperate. Otherwise your best options are "!submodule" "!gitslave" and "!subtree" Try typing those commands into this IRC channel.02:03
→ chrisshattuck joined02:03
→ dbgster joined02:03
NickPresta SethRobertson: thank you. I ended up just installing the library as a Python package.02:03
← bielewelt left02:03
→ agile joined02:04
→ bielewelt joined02:08
→ d2dchat joined02:11
→ ChaoWang joined02:12
← cooloney left02:12
→ cooloney joined02:12
→ fling joined02:12
→ zivester joined02:13
← NickPresta left02:13
→ intripoon joined02:15
← orafu left02:16
→ orafu joined02:16
← Nibato left02:16
→ el2ro joined02:18
← intripoon_ left02:18
← zivester left02:19
← bartman left02:19
← jsutri left02:19
← Paradox2000 left02:19
→ Paradox2000 joined02:19
← PigeonFriend left02:19
← ChaoWang left02:20
← cooloney left02:20
→ duckxx joined02:21
← Paradox2000 left02:22
→ Nibato joined02:23
← williamcotton left02:23
→ Paradox2000 joined02:23
← Paradox2000 left02:23
→ Paradox2000 joined02:23
→ webflo1 joined02:24
← webflo left02:24
← LimeBlast left02:25
→ kingx joined02:26
→ bartman joined02:26
kingx hello02:26
gitinfo kingx: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.02:26
← thelazydeveloper left02:26
→ cooloney joined02:27
kingx first time working with version control, just wanted to get something simple figured out. I have a small project (simple html files), that I want to maintain with git. So I can edit the files on one computer and commit them to run on another.02:27
SethRobertson Read !book02:27
gitinfo There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !gitt !vcbe and !parable02:27
← mwrosen left02:28
CareBear\ kingx : moving files around (deployment) is somehow orthogonal to version control02:28
SethRobertson There is !web but...02:28
gitinfo [!website] Git is not a website deployment tool, but can sometimes play one in sufficiently simple/lax environments with a little help. One example of help is: http://toroid.org/ams/git-website-howto and another approach: https://gist.github.com/171423502:28
kingx well what i was thinking is (maybe i am wrong), I simply dont want to move files around every time I make a change.02:28
CareBear\ kingx : I like http://joemaller.com/990 for using git with web02:28
kingx so maybe I can commit to a git repo and then clone the git where i want the files running02:29
CareBear\ kingx : the term for what you want is automated deployment02:29
kingx or that makes no sense lol02:29
CareBear\ kingx : but it's important that you keep in mind that deployment is orthogonal to version control02:29
kingx : that said, there are a couple of popular ways to script around git, so that you *can* in fact deploy also using git commnads02:29
commands02:30
kingx ah i see02:30
now that makes sense02:30
CareBear\ recommend check out the links and simply try it out02:30
→ zivester joined02:31
← Araxia left02:31
CareBear\ I like the joemaller.com/990 solution although it's slightly more complicated, because it allows to work also on the server should that be neccessary02:31
← MestreLion left02:31
CareBear\ but you may want to try out the simpler solutions first02:31
→ Slix` joined02:31
CareBear\ I recommend experimenting with a dummy site02:31
→ gavin_huang joined02:31
CareBear\ just to quickly try out the different methods02:31
← mxweas_ left02:32
CareBear\ then decide on which one you like, and implement it for your real files02:32
→ jalama joined02:32
kingx yeh thats what i have going on, this web-focused git workflow seems like what i need02:32
i'll start reading the links you mentioned, thanks for pointing em out :)02:33
→ apok_ joined02:33
CareBear\ you're welcome, SethRobertson also helped02:33
later!02:33
← hobbyist left02:35
← mister_m left02:36
→ pastime joined02:36
← apok_ left02:37
← apok left02:38
→ pantsman joined02:39
← dhruvasagar left02:39
← legnaleurc left02:41
→ stkrzysiak joined02:41
← zivester left02:42
← d2dchat left02:42
→ dhruvasagar joined02:42
→ maletor joined02:43
→ mwrosen joined02:44
← timoxley left02:44
← marass_ left02:44
→ marass joined02:44
← dhruvasagar left02:44
← nif left02:45
← subbyyy left02:46
→ dhruvasagar joined02:46
← jim0203 left02:48
← pantsman left02:48
← jalama left02:49
← fukushima left02:51
→ Nightfire034 joined02:52
bigkitty HI02:53
can not use git clone right now:(02:53
what's wrong?02:53
SethRobertson !repro02:53
gitinfo Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.02:53
imMute nah, it's the moon phase messing with things. try again in 13.3 minutes02:54
→ Yuuhi`` joined02:54
← btanaka left02:55
→ zivester joined02:55
← Yuuhi` left02:56
← gavin_huang left02:57
← maestrojed left02:57
bigkitty http://iyanwu.com/pastebin/showthread.php?tid=3002:59
→ fukushima joined03:00
← ybit left03:02
SethRobertson Ah, well it looks like you have upgraded perl but not git03:02
What OS are you running?03:02
distrobution03:02
→ gavin_huang joined03:03
→ pjn_oz_ joined03:04
← cccaldas left03:04
→ scottrigby joined03:04
→ DrNick joined03:04
→ cjs joined03:04
→ mister_m joined03:05
← marowit left03:05
→ williamcotton joined03:05
bremner bigkitty: also, please be precise about what command you typed. It looks much more like "git svn clone" than "git clone"03:05
← Paradox2000 left03:06
bigkitty bremner i am running git svn clone:)03:06
thiago bigkitty: install your Perl SVN libraries03:07
bigkitty thiago how to install it?03:09
← subrosa left03:09
bremner umm. it depends.03:10
→ ybit joined03:10
bremner ask your OS support buddy?03:10
← williamcotton left03:10
→ marowit_ joined03:11
← travisfischer left03:12
→ jrots_ joined03:13
→ pjn_oz joined03:13
← dbgster left03:13
→ travisfischer joined03:13
← gavin_huang left03:14
→ mxweas_ joined03:14
← Contex left03:14
bigkitty bremner i ask u now:)03:14
→ mattsa joined03:14
← pjn_oz_ left03:14
bremner bigkitty: I don't even know what OS you are running. It looks kindof old and crufty; perl 5.8 is pretty obsolete.03:16
thiago bigkitty: this is #git, not #perl. Please check with your OS or Perl vendor how to install those.03:16
bigkitty: also, it's about the SVN libraries. Make sure you installed them from wherever you get Subversion.03:16
→ gavin_huang joined03:20
← mattsa left03:20
← jrots_ left03:21
→ jrots joined03:21
← gavin_huang left03:21
← mxweas_ left03:24
bigkitty thiago i am using centos 5.3 now:)03:25
thiago bigkitty: try asking in #centos03:25
→ mxweas_ joined03:25
← ferdna left03:26
jpalmer bigkitty: hint: the first thing #centos will do, is make fun of you for being on an ancient version.03:26
You'd be better off in #RHEL, #centos isn't a very pleasant channel.03:27
← jrots left03:27
← mxweas_ left03:28
→ gavin_huang joined03:28
← Nightfire034 left03:28
→ philoserf joined03:30
→ jrots joined03:30
← scottrigby left03:30
← philoserf left03:34
→ philoserf joined03:34
→ timoxley joined03:34
→ michaelmartinez joined03:35
← jrots left03:36
→ jrots joined03:36
→ maestrojed joined03:37
← maestrojed left03:38
← skoop left03:39
→ glennpratt joined03:41
← gavin_huang left03:42
→ maestrojed joined03:42
← jameswarren left03:43
→ literal_username joined03:44
← jrots left03:45
→ ericmuyser joined03:45
→ gavin_huang joined03:46
← eletuchy left03:47
literal_username Hi all, I have a file with 0400 permissions that belongs to the system user that is used to run the application in my repo. I want to include the file with the repo, but git can't stat it under my account. So my options seem to be to use git as my app's system user, which seems wrong, or to just not version the file at all, even though it belongs in there, or... any ideas?03:47
→ dr0id joined03:48
→ yasar joined03:48
literal_username I could give it 0440 permissions, but that's a minor security concern.03:48
→ Fullmoon joined03:49
← brngardner left03:49
yasar how do I see diff between staged file and last commited version of that file?03:49
jpalmer depending on the OS, you could also use acl's to gran your git user read perms in addition to the system users read perms.03:49
literal_username jpalmer: It's Linux. That sounds plausible.03:50
jpalmer literal_username: man setfacl, man getfacl03:52
rgwu yasar: git diff HEAD03:53
literal_username jpalmer: Sweet. This is new to me, so I really appreciate the knowledge. Thanks!03:53
← elnn left03:53
rgwu yasar: or git diff HEAD filename03:54
jpalmer literal_username: keep in mind, I don't know your environment, so this may not be the *best* answer. it only answers your immediate question. pesonally, I'd look for a workflow where you can utilize the traditional "ugo" permissions model.03:54
→ elnn joined03:54
thiago literal_username: note that git might decide to erase and recreate that file03:54
literal_username: it will reset to regular permissions (no ACL, 0666 | ~umask)03:54
→ BiggFREE joined03:54
thiago 0666 & ~umask, I mean03:55
jpalmer ^^ what thiago said03:55
literal_username Oh.03:55
→ ketas joined03:55
thiago literal_username: if you don't want git to touch the file, don't track it03:56
→ archdori joined03:56
literal_username I don't really mind if git touches it, but changing its permissions is undesirable.03:56
jpalmer literal_username: also, depending on which linux (saying "Linux" is kinda vague in this regard) you may have to enable kernel modules, or remount your filesystems to utilize the extended permissions/attributes.03:57
literal_username Arch, ext4 fwiw03:57
thiago literal_username: "touch" means rewrite03:57
literal_username: Git never writes to an existing file. It only creates a new one and replaces the old.03:58
jpalmer isn't familiar with selinux on arch, so I'll defer you to the docs ;)03:58
← enmand left03:58
← chrisshattuck left03:59
literal_username thiago: I get that. I don't follow the need to change permissions from the original on re-create, though.03:59
← madewokherd left03:59
thiago literal_username: because it's a new file03:59
literal_username: it gets the default permissions03:59
literal_username So even if I give this file 0440, and it gets updated, then git gives it 0666 & ~umask04:00
thiago no04:00
your operating system does04:00
← Fandekasp left04:00
thiago git doesn't care. It creates a new file.04:00
→ acl_ joined04:00
literal_username So user and group are not retained either.04:00
thiago it's the OS that chooses what the permissions, owner and group will be.04:00
correct04:00
← gchristensen left04:00
→ hagabaka joined04:01
← hagabaka left04:01
→ hagabaka joined04:01
literal_username Sure. I follow the logic, it just seems sorta annoying.04:01
thiago git can be configured to force the permissions04:01
but not user or group04:01
especially not user. chown requires root.04:01
literal_username Why do files get replaced instead of updated?04:03
← hyperair left04:03
← aca20031 left04:04
→ skylamer` joined04:04
→ legnaleurc joined04:04
→ tonesfrommars joined04:08
← Internet13 left04:11
← superdmp left04:11
→ superdmp joined04:11
← ajaykools left04:13
← ryden left04:13
← coyotebush left04:13
← b1lly left04:16
→ Internet13 joined04:16
← mwrosen left04:16
→ gchristensen joined04:19
← krz left04:19
← rgwu left04:19
← gchristensen left04:19
→ dstorrs joined04:21
← iamjarvo left04:23
→ brando753 joined04:25
→ ehsan joined04:29
← timoxley left04:31
→ richo joined04:31
← richo left04:31
→ richo joined04:31
richo I'm having a brainfart, for a given commit, how do I extract it's commit message?04:32
→ envi_ joined04:32
offby1 "git show"04:32
richo it seems insistent to show me the diff too04:33
all I want is the commit message04:33
cehteh git log -1 and some pretty format works for sure04:35
→ timoxley joined04:35
cehteh git cat-file commit $hash and some mangling may work too (but thats plumbing)04:35
→ zamN joined04:36
cehteh as well git-show with pretty format works04:36
richo plumbing is probably ok04:36
i tried git show --pretty=format:"%s" HEAD04:37
but that still shows me the diff04:37
→ FreeArtMan joined04:37
cehteh try git log -1 then04:38
→ beautifulmind joined04:38
cehteh is the full body of the message04:38
← Internet13 left04:38
richo git log -1 works great, thanks :)04:39
cehteh (did you see that percentB .. my client takes that as bold formatting, stupid)04:39
← timoxley left04:40
→ dvide_ joined04:41
← estevocastro left04:41
← yasar left04:41
← dvide left04:43
← lyddonb left04:43
← pastime left04:43
→ jameswarren joined04:44
→ Internet13 joined04:44
brando753 I followed the documentation for getting git to publish to my webs document root, though the files publish to the git repo they dont seem to update the documentation root, Im using gitolite as the git server04:44
← gavin_huang left04:44
brando753 http://toroid.org/ams/git-website-howto04:44
→ eletuchy joined04:45
richo cehteh: Yup, all sorted thankyou :)04:47
→ Error404NotFound joined04:48
→ timoxley joined04:48
← jameswarren left04:48
← Error404NotFound left04:49
→ gavin_huang joined04:51
← dstorrs left04:52
→ cime joined04:52
→ mbrochh joined04:52
mbrochh hey all. do I understand it correctly that "git push" tries to push all my branches?04:52
I often step into pifalls because of this, wouldn't it be much more intuitive if "git push" would only push the active branch? "git push -all" should push all branches.04:53
→ dvide joined04:53
← dvide_ left04:56
← maestrojed left04:57
→ apok joined04:57
→ thirdknife joined05:00
wereHamster mbrochh: man git-config -> push.default05:01
gitinfo mbrochh: the git-config manpage is available at http://jk.gs/git-config.html05:01
← diegoviola left05:02
mbrochh wereHamster: many thanks! super helpful!!05:02
→ _sundar_ joined05:03
→ maestrojed joined05:05
→ akiress joined05:06
← BiggFREE left05:09
← jhford left05:09
← yoklov left05:10
← electrogeek left05:10
← jrockway left05:10
← staticsafe left05:10
← magcius left05:10
→ electrogeek joined05:10
← electrogeek left05:10
→ electrogeek joined05:10
→ jrockway joined05:12
← ceej left05:14
→ magcius joined05:15
← Qoidnef left05:15
← thiago left05:16
← philoserf left05:17
← hotwings left05:17
← duckxx left05:18
→ harshpb joined05:18
→ rurufufuss joined05:18
← macmartine left05:18
→ philoserf joined05:19
← philoserf left05:21
→ jhford joined05:21
→ staticsafe joined05:22
← staticsafe left05:22
→ staticsafe joined05:22
← gavin_huang left05:22
→ macmartine joined05:23
← ology left05:24
← kukks left05:24
← tatsuya_o left05:26
← pom left05:26
→ tatsuya_o joined05:26
→ philoserf joined05:26
← flippo left05:26
→ replore_ joined05:27
→ gavin_huang joined05:27
← xiangfu left05:28
→ xiangfu joined05:29
← project2501b left05:29
← thirdknife left05:30
→ tewecske joined05:31
← tewecske left05:31
← envi_ left05:34
→ dreiss joined05:34
→ thirdknife joined05:34
→ Sigma joined05:34
→ caseymcg joined05:35
→ JohnBat26 joined05:36
→ notola joined05:36
sitaram brando753: try adding 'unset GIT_DIR' at the top of the post-receive hook; without digging much deeper into details that's the only thing I can think of05:38
← kingx left05:38
← notola left05:40
→ ajaykools joined05:40
→ notola joined05:41
← brokenshakles left05:42
→ tewecske joined05:43
← tewecske left05:43
← xiangfu left05:43
→ sandstrom joined05:43
← sandstrom left05:43
→ sandstrom joined05:43
→ DofLeFou joined05:44
milki brando753: !website05:44
gitinfo brando753: Git is not a website deployment tool, but can sometimes play one in sufficiently simple/lax environments with a little help. One example of help is: http://toroid.org/ams/git-website-howto and another approach: https://gist.github.com/171423505:44
milki brando753: people suggest the second method05:44
→ brokenshakles joined05:46
← notola left05:46
← richo left05:46
→ pfFredd joined05:46
← tatsuya_o left05:47
← macmartine left05:47
← beautifulmind left05:47
→ tatsuya_o joined05:47
→ subbyyy joined05:48
→ mishok13 joined05:48
← luke-jr left05:50
→ luke-jr joined05:50
← Goplat left05:51
← neurodrone left05:52
← mister_m left05:52
→ hyperair joined05:53
← tatsuya_o left05:54
→ tatsuya_o joined05:54
← Davey left05:54
→ psoo joined05:55
→ akraut joined05:57
→ mxweas_ joined05:58
→ airborn joined05:59
← OOPMan left06:00
→ fayaz joined06:02
← ben225 left06:02
→ flijten joined06:03
← smw left06:03
→ macmartine joined06:03
← RobertLaptop left06:03
← dhruvasagar left06:04
→ patrakov joined06:04
→ beautifulmind joined06:05
→ dhruvasagar joined06:06
← macmartine left06:06
patrakov Hi. I made a commit, then was asked to revert it (with the motivation "not now, and this piece of your code can be simplified"), reverted.06:06
← The-Compiler left06:07
patrakov Now I want to work on this again. How to revert the revert so that the result is not a commit but rather something that I can further edit and then commit?06:07
→ awkward joined06:07
← thierryp left06:07
← fayaz left06:07
← ericmuyser left06:08
← philoserf left06:10
→ ThomasLocke joined06:10
← ormaaj left06:11
→ Dave^| joined06:12
← sandstrom left06:12
→ cbreak-work joined06:13
→ tsdh joined06:14
brando753 milki, how do I get that second one to copy the files to the web root?06:15
it wasnt commented06:15
→ sacback joined06:15
patrakov ok, "git revert --no-commit e5979d184c" did almost what I wanted. The diff is in "git diff --cached"06:15
→ dpino joined06:15
→ Shusshu joined06:16
← agile left06:16
→ OOPMan joined06:17
← ajaykools left06:17
→ agile joined06:17
→ richo joined06:18
richo I'm trying to fetch just the content of an annotated tag, ie git [something] tag_name, I'm assuming I'll need git log -1 magic again?06:19
→ glcrazy joined06:19
milki brando753: you modify the rsync command06:19
← acl_ left06:20
→ RobertLaptop joined06:20
→ The-Compiler joined06:20
→ iocor joined06:20
→ lonewulf` joined06:23
← hyperair left06:23
→ Qoidnef joined06:25
→ slide_ joined06:26
→ ormaaj joined06:27
← gdoteof left06:27
← michaelmartinez left06:28
brando753 milki, I changed the line "rsync --delete -rlpgocv lib/ /web/geekcruises-secure/lib-ng/" to "rsync --delete -rlpgocv lib/ /srv/www/mywebsite/" but it still does not show up there06:29
← maestrojed left06:29
← slide23 left06:29
← maletor left06:30
← mishok13 left06:31
→ chayin_ joined06:32
milki you need to modify lib/ to be the dir you want to copy over as well06:33
man rsync06:33
→ giallu joined06:34
→ hyperair joined06:34
→ chayin__ joined06:35
→ mastro joined06:35
← k0rnel left06:35
← chayin left06:36
→ Sitethief joined06:36
→ magn3ts joined06:36
magn3ts How can I move my working tree 'back in time'06:36
I want to equivalent of "git checkout 1532a3d4a1a70f594c64d1196b2252ba4b302844"06:37
can I do something like that?06:37
EugeneKay magn3ts - of course. You probably want to create a branch when you checkout.06:37
magn3ts EugeneKay: you're just every fucking where.06:38
EugeneKay eg, git checkout -b goingback 1532a3d406:38
magn3ts EugeneKay: I explicitly do plan on modifying the code at all.06:38
EugeneKay More places than you know ;-)06:38
magn3ts so I don't want to branch or rebase I don't think.06:38
do not***06:38
magn3ts dammit06:38
← ipalaus_ left06:38
EugeneKay What are you trying to do? Android stuff?06:38
← chayin_ left06:38
magn3ts No, we introduced a bug over the last days in a project I'm on and I need to track down which commit caused it06:39
EugeneKay Ah, OK.06:39
Is it something you can test at various points?06:39
→ orkz joined06:39
magn3ts yes but only if the code's checked out06:40
EugeneKay Right06:40
← awkward left06:40
EugeneKay Create a new branch, call it 'test', from wherever you are now.06:40
Then use reset --hard $ID to move that branch around(and update your working-tree to match) to each suspected point06:41
← joaoh82 left06:41
EugeneKay You can use !lol to get the various SHAs you want to try06:41
gitinfo git config --global alias.lol "log --oneline --graph --decorate"06:41
magn3ts after switching to the branch of course :)06:41
← yrlnry left06:41
EugeneKay 'git checkout -b Foo HEAD' will create Foo from wherever you are now, and check it out.06:41
→ pantsman joined06:42
← pantsman left06:42
→ pantsman joined06:42
EugeneKay Shortcuts ftw06:42
magn3ts oh, I see.06:42
← FabTG left06:42
magn3ts I'll work on that then, thanks sir.06:42
EugeneKay As you'll be reset --hard'ing it a bunch it probably isn't important that your stuff is pushable.06:42
Just so you can find the problem06:42
→ sandstrom joined06:42
← sandstrom left06:43
→ sandstrom joined06:43
→ ph^ joined06:43
→ deebo joined06:43
deebo how do i reset the current index and checkout from .git directory?06:43
← zamN left06:44
deebo i moved the .git directory into a blank directory to get a "blank checkout" easily, but it wasnt that easy after all :P06:44
→ maletor joined06:44
EugeneKay deebo - !fixup, or man git-reset06:44
gitinfo deebo: the git-reset manpage is available at http://jk.gs/git-reset.html06:44
← OPK left06:44
EugeneKay !fixup06:44
gitinfo So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions. For hints type !fixup_hints in IRC. Remember: if you have pushed already, there are only a few things you can do without !rewriting_public_history (type that for more info)06:44
EugeneKay Stupid bot.06:44
← airborn left06:45
deebo ah yeah, reset to HEAD did it, thanks06:45
→ ben225 joined06:45
→ dc5ala joined06:46
→ wilmoore joined06:46
← wilmoore left06:47
→ rendar joined06:47
← barrbrain left06:49
→ airborn joined06:49
magn3ts EugeneKay: So if I delete the branch before doing a commit, it won't be pushed ever though right, or does doing a hard reset create a new "history point"?06:49
EugeneKay magn3ts - the branch shouldn't be pushed anyway, as it isn't a "tracked" branch.06:49
magn3ts EugeneKay: oh. does a first commit on a new branch cause it to be tracked then I assume?06:50
EugeneKay reset --hard force-moves the branch pointer to the commit you specify and checks it out06:50
→ marthinal joined06:50
EugeneKay No, --track(when checking out a remotes/foo/branch) or set-upstream cause it to be tracked.06:50
← Mowsh left06:51
magn3ts So if I didn't do that when I started... would I have to just do `git checkout --track branch` again?06:51
→ arvind_k joined06:51
magn3ts (theoretically, I don't want to do it here obviously)06:51
← giallu left06:51
EugeneKay I would use set-upstream06:51
Which is a branch subcommand06:51
magn3ts okay. I see. thanks06:51
EugeneKay WHY it doesn't have --, I don't know. Git's stupid porcelain needs a reworking.06:52
→ maestrojed joined06:52
← ben225 left06:52
← bartman left06:52
← tonesfrommars left06:53
← thirdknife left06:53
→ resmo_ joined06:53
→ jceb joined06:53
arvind_k → arvind_khadri06:54
← arvind_khadri left06:54
→ arvind_khadri joined06:54
milki maybe its not using the same library to parse options06:55
← Fullmoon left06:55
EugeneKay No, it's a case of unintelligent design.06:55
Actually, set-upstream DOES use --. I'm thinking of "set-url", which is git-remote.06:55
← mritz left06:56
← sacback left06:56
EugeneKay rambles for a bit about things that are terrible, but offers no comprehensive solutions to improving them06:56
→ Fandekasp joined06:57
→ netiru joined06:58
→ bartman joined06:59
magn3ts I almost want to create a script that will do a binary search for a bug. I can tell it two points, one where there was a bug, and one before it that didn't have the bug. It will do checkouts in a binary fashion and allow me to say if the bug was present at each checkout...06:59
EugeneKay man git-bisest06:59
IcePic bisect*06:59
EugeneKay man git-bisect06:59
gitinfo the git-bisect manpage is available at http://jk.gs/git-bisect.html06:59
→ syphar joined06:59
magn3ts :{06:59
awesome. I'm enjoying my gitfu expanding07:00
← jweinberg left07:01
EugeneKay I've never used it, it kinda scares me.07:01
← archdori left07:01
→ Element9 joined07:01
← WillMarshall left07:02
← machine2 left07:02
← blackmatrix_ny left07:04
→ giallu joined07:05
→ AlexeiKozlov joined07:05
← welly left07:05
→ sq-one joined07:05
← airborn left07:05
→ tewecske joined07:06
← FreeArtMan left07:06
brando753 milki, No what I dont understand is what lib/ is supposed to be to copy the file? I want it to post the entire repo into the web root, I understand rsync is syncing the two repos, however what do I identify as the whole directory for git? Is it just / ?07:07
← AlexeiKozlov left07:07
EugeneKay brando753 - are you trying to !deploy ?07:08
gitinfo brando753: [!website] Git is not a website deployment tool, but can sometimes play one in sufficiently simple/lax environments with a little help. One example of help is: http://toroid.org/ams/git-website-howto and another approach: https://gist.github.com/171423507:08
← iocor left07:08
brando753 EugeneKay, yes, and Im having trouble getting it to sync to the web root07:08
→ FabTG joined07:08
← webflo1 left07:09
EugeneKay I would recommend using the approach outlined in that gist, rather than the --work-tree approach you typically see in !blog posts.07:09
gitinfo Blog posts, while helpful and informative, are quite often outdated, give bad advice, or are just plain wrong. Please don't rely solely upon them, or treat them as authoritative.07:09
← dreiss left07:10
brando753 EugeneKay, thats the one Im using now, though I am not sure what to change the rsync command to, it has no comments07:10
EugeneKay man rsync ;-)07:10
← Slix` left07:11
→ yrlnry joined07:11
→ AlexeiKozlov joined07:11
gitinfo set mode: +v07:11
← maestrojed left07:11
brando753 EugeneKay, again Im unsure on how to reference the root git directory07:11
AlexeiKozlov hi every1. i have a question about installing gitolite. i got this mesasge from gl-system-install ->git describe failed -- your git is probably too old . what i suppose to do?07:12
→ slide23 joined07:12
EugeneKay In that example, git-archive is being used to "check out" the root of the git dir into a tmp directory, and then it is being CDed into.07:12
If you wanted to rsync the whole directory you would use ./07:12
s/directory/repo/07:12
If I was to build such a hook I would lay it out differently07:13
→ especially-corn joined07:15
especially-corn hello07:15
gitinfo especially-corn: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.07:15
← slide_ left07:15
brando753 EugeneKay, ya, its still not working, here is my rsync command in that hook " rsync --delete -rlpgocv ./ /srv/www/mysite/" and the hook is just pre-receive07:15
especially-corn i just did a repo init thing and its receiving from the repo, shouldn't i see the files it receives?07:16
EugeneKay brando753 - like I said, I'd build a radically different hook; it's provided mostly as an example. ;-)07:16
Scripting your own hook for this sort of thing is a good learn-by-fumbling experience. I advise you to just go for it.07:17
AlexeiKozlov i have fail in record_version sub in gl-system-install. it says "git describe failed -- your git is probably too old" . help me please.07:17
← mxweas_ left07:17
→ pro_metedor joined07:17
← brokenshakles left07:17
→ jweinberg_ joined07:18
brando753 im just trying to get it to work at this point :( I also tried the earlier example of "GIT_WORK_TREE=/srv/www/mysite git checkout -f" but that didnt work either ... Im trying to find if there is another reason both these methods fail to work ...07:18
→ [[thufir]] joined07:19
← jweinberg_ left07:20
[[thufir]] can I make n commits, then one push to github, and have each of those commits show up individually? Or, does it just show as one push?07:20
→ jweinberg joined07:20
→ rtuin joined07:21
→ SimonNa joined07:21
sq-one each show up individually07:23
EugeneKay [[thufir]] - github would show it as 1 push of N commits, in the dashboard thing. You can push them individually if you want, but there's no real gain except lots of spam.07:23
AlexeiKozlov someone can help me in that message "git describe failed -- your git is probably too old" ?07:23
EugeneKay AlexeiKozlov - git --version07:23
[[thufir]] and if I want to rollback (?) to commit x, it's just as easy either way?07:24
AlexeiKozlov it says 1.7.5.107:24
EugeneKay Depends how you want to rollback. !revert07:24
gitinfo "Revert" is a heavily overloaded term. Do you mean: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard the uncommitted changes in the working tree [git checkout -- .]; c) undo committing [git reset --soft HEAD^, but type !rewriting_public_history] ?07:24
EugeneKay Do you mean to move backwards in time, or to make your repo as it was in the past?07:25
← meejah left07:25
→ vdv joined07:25
→ meejah joined07:25
AlexeiKozlov EugeneKay it says 1.7.5.107:25
[[thufir]] as it was in the past.07:25
EugeneKay [[thufir]] - then you're making a new commit which just happens to have the same tree as an older one. It still coutns as going forward, so it'd show as aregular commit.07:26
AlexeiKozlov -- well, that's interesting. What's your OS?07:26
AlexeiKozlov EugeneKay my OS windows 2008 R2 and cygwin07:26
EugeneKay Oh07:27
AlexeiKozlov )07:27
EugeneKay cygwin git is known to be..... well, terrible.07:27
AlexeiKozlov it works meanwhile07:27
but not latest version07:27
→ teotwaki joined07:27
EugeneKay I would be really surprised if you manage to get gitolite running on it. I can't say I recommend it, at all.07:27
AlexeiKozlov really?07:27
i did this but07:27
not latest version. git works fine in cygwin. no bugs.07:28
← beautifulmind left07:28
EugeneKay My advice is to use "git for windows"(formerly msysgit), and to run gitolite on a proper *nix box.07:28
Inside of a VM is just fine.07:28
AlexeiKozlov when v2.1 comes out it makes install very hard :( before 2.1 i do every install wery well.07:29
← thursday left07:30
AlexeiKozlov EugeneKay can you help about gl-system-install.pm ?07:30
this code fails sub record_version { # this is really much easier in plain shell :( if ( system("git rev-parse --is-inside-work-tree >/dev/null 2>&1") ) { # for system() calls, perl 'true/success' is shell 'false/fail', which # means the command failed; we are not in a git work tree -f "conf/VERSION" or system("echo '(unknown)' > conf/VERSION"); } else { system("git describe --tags --long07:30
EugeneKay I cannot. I use a packaged version of gitolite which has the nice property of 'just working'.07:31
← maletor left07:31
AlexeiKozlov ) lucku07:31
y07:31
EugeneKay sitaram might be able to give you some pointers with some proper error messages, but that's a pretty big "maybe" - Windows is NOT a supported OS ;-)07:31
→ mishok13 joined07:31
AlexeiKozlov is anybody knows perl ? i need help with gl-system-intall.pm07:31
EugeneKay i wrote him some emails) he doesnt want to help after word "cygwin" :(07:32
← Sitethief left07:32
AlexeiKozlov and word "windows"07:33
EugeneKay I am completely unsurprised.07:33
→ maletor joined07:33
→ Sitethief joined07:33
EugeneKay It's a *nix app, and, well, Windows.07:34
AlexeiKozlov )07:34
its perl case) not a windows or cygwin07:34
→ brokenshakles joined07:34
deebo cygwin + git on windows is so much more usable than msysgit (or whatever its called now)07:35
AlexeiKozlov deebo ur right) i m noot using msysgit.07:35
i have gitolite(v.2.0) +git+cygwin works well on windows 2008 r2.07:35
→ blacktulip joined07:36
AlexeiKozlov but someone can help in perl for me?07:36
this if ( system("git rev-parse --is-inside-work-tree >/dev/null 2>&1") ) works bugly07:36
← subbyyy left07:36
→ williamcotton joined07:36
AlexeiKozlov if i run this command myself i get "git@NP4 /tmp/share/gitolite $ git rev-parse --is-inside-work-tree true"07:37
← richo left07:37
→ skoop joined07:37
← skoop left07:37
AlexeiKozlov there is true! but it doesnt recognized by perl07:37
← pantsman left07:38
→ skoop joined07:39
AlexeiKozlov there says "# for system() calls, perl 'true/success' is shell 'false/fail', which # means the command failed; we are not in a git work tree"07:39
→ Lujeni joined07:40
AlexeiKozlov if i run "git rev-parse --is-inside-work-tree" from shell it returns "true"07:40
→ mastroWork joined07:42
AlexeiKozlov hm. perl says that : system(..) returns true.07:42
← tatsuya_o left07:44
← eeadc left07:44
→ weiyang joined07:44
→ eeadc joined07:44
← mastro left07:45
← comcor left07:45
→ comcor joined07:46
← williamcotton left07:48
← especially-corn left07:48
← matthiaskrgr left07:48
AlexeiKozlov help please. how i can change permission on everything in folder/07:49
?07:49
← Cope left07:50
→ Chaser joined07:50
→ arietis joined07:52
← legnaleurc left07:52
← maletor left07:54
→ Desproges joined07:55
sitaram AlexeiKozlov: you can remove the code that generates that message. For the rest, ask frogonwheels if he is here; he's the only guy I know who has managed to install gitolite on cygwin. I personally do not want to encourage such foolishness :-)07:56
← Nugget left07:56
sitaram deebo: cygwin+git may be *usable* but installing cygwin is a pain the rear-end. And for a *client*, msysgit is quite good; I'd argue it's better than cygwin07:57
AlexeiKozlov oo nice07:57
← hagabaka left07:57
→ shruggar joined07:57
AlexeiKozlov suprised)07:57
sitaram AlexeiKozlov: or you can remove the "--dirty=dt" or whatever part of that command07:57
AlexeiKozlov thank i try it now.07:57
sitaram AlexeiKozlov: that was added so that you can know if you're running a "dirty" (i.e., locally modified) gitolite07:57
AlexeiKozlov oh! ok i got thanks for explanation)07:58
← netiru left07:58
→ eijk joined07:58
sitaram AlexeiKozlov: you wrote me some emails? when?07:58
AlexeiKozlov yeah i wrote)07:58
wait 1 min07:58
← brokenshakles left07:59
sitaram but yeah if the subject line is cygwin I probably don't want to spend time. It's usually a waste07:59
(or windows)07:59
AlexeiKozlov Hi. i got gitolite 2.2 version and have one question about customized setup07:59
sitaram AlexeiKozlov: the problem is this: whatever problems you have are likely to be specific to your env, and since I dont have that (and never will), there's no point07:59
AlexeiKozlov this subject)07:59
→ Bartzy joined07:59
AlexeiKozlov ok ok) but its specific to perl . not cygwin i think.08:00
Bartzy How can I print the hash that git gave to a blob object (file) ?08:00
sitaram AlexeiKozlov: same thing; it works fine in a proper unix08:00
→ pmatthews joined08:00
AlexeiKozlov how to prefix nickname08:00
sitaram Bartzy: man git hash-object08:00
gitinfo Bartzy: the git-hash-object manpage is available at http://jk.gs/git-hash-object.html08:00
AlexeiKozlov ?08:00
sitaram AlexeiKozlov: "prefix nickname"? I have no idea what you are talking about08:00
→ brokenshakles joined08:01
AlexeiKozlov in irc chat )08:01
→ ben225 joined08:01
→ niro joined08:01
sitaram AlexeiKozlov: usually, type the first 2-3 characters, hit tab08:01
Bartzy Thank you.08:01
→ Nugget joined08:01
Bartzy why git hash-object and sha1sum gives me different results ? What Git is using for the hash besides the file itself ?08:02
→ netiru joined08:02
sitaram AlexeiKozlov: the other reason I won't *directly* help is that I am seriously prejudiced against both the proprietary OSs in popular use08:02
niro git status08:02
excuse me08:02
→ jlebrech joined08:03
→ francisfish joined08:03
AlexeiKozlov sitaram )) i only need help for perl) i do install now. report feedback later in 10 minutes08:03
niro I get the error, git-am is in progress. cannot rebase. What do I need to do to rebase?08:03
← glcrazy left08:03
sitaram AlexeiKozlov: if you need help for perl to use *gitolite*, that is *proof* that you are using a crappy OS08:03
→ TomyLobo joined08:03
sitaram AlexeiKozlov: no one needs to know perl to use gitolite, except some bare minimum syntax to add entries to ~/.gitolite.rc08:04
s/crappy OS/crappy git setup, possibly due to crappy OS/08:04
← Shusshu left08:05
→ alanp_ joined08:05
← alanp left08:05
EugeneKay Bartzy - man git-hash-object08:07
gitinfo Bartzy: the git-hash-object manpage is available at http://jk.gs/git-hash-object.html08:07
→ subbyyy joined08:07
AlexeiKozlov sitaram yeah) but i need some help) it not crappy OS. ots crappy cygwin env. )08:07
← netiru left08:07
sitaram AlexeiKozlov: <sigh> think about what is the *root* cause of your troubles08:08
→ thierryp joined08:08
EugeneKay AlexeiKozlov - And we've given you about all the advice we're capable of / willing to give. You are advised to use gitolite in a *nix environment or figure it out for yourself. If you have a problem with this, well, go away.08:08
AlexeiKozlov sitaram not a windows) belive me))08:08
→ elijah joined08:08
sitaram AlexeiKozlov: prove that it happens on Linux or BSD and I will help :)08:08
AlexeiKozlov EugeneKay why? i only need to know about recordversion sub in gl-system-install.08:08
← brokenshakles left08:09
AlexeiKozlov i dont want u help to setup it in cygwin ) i want only knowledge about gl-system-install08:09
EugeneKay AlexeiKozlov - it's self-documenting. Read the source / learn perl.08:09
AlexeiKozlov i dont want to learn perl ) its crappy language )08:09
sitaram AlexeiKozlov: and you expect help from me after saying perl is a crappy language?08:10
AlexeiKozlov what you for i you adivce me to learn everything by myself?08:10
EugeneKay Learning is important. Please go do some.08:10
AlexeiKozlov ahaha))) sitarem08:10
→ netiru joined08:10
AlexeiKozlov but what for this channel when?08:10
← weiyang left08:10
AlexeiKozlov send someone away?08:11
sitaram AlexeiKozlov: I already told you how to get past that error by removing the "--dirty=-dt" bit in the git describe.08:11
AlexeiKozlov to learn, yeahh?))08:11
sitaram yeah ! and i say thank you and now i go next step very well08:12
← ben225 left08:13
AlexeiKozlov sitaram and about perl. if its not crappy than ist not true cross-platform lang. right? ;)08:14
than=then08:14
neersighted → neersighted|AFK08:14
→ robotmay joined08:14
← Matrixium left08:15
sitaram AlexeiKozlov: I don't take kindly to people who say negative things about perl while they use *Windows* as their OS. You, sir, are a ...<never mind>...08:15
← caseymcg left08:15
EugeneKay hands sitaram a $INTOXICATING_BEVERAGE08:15
AlexeiKozlov ^)08:15
tango_ lol08:16
AlexeiKozlov )))08:16
cbreak-work cross platform?08:16
perl works on OS X, kind of.08:16
and the other unixes :)08:16
AlexeiKozlov or yeah ) but percent of mac os worldwide how much is ? :)08:16
cbreak-work 100% of the mac os worldwide08:17
AlexeiKozlov ))08:17
cbreak-work it comes preinstalled.08:17
← superdmp left08:17
← netiru left08:17
AlexeiKozlov beside, i m from russia, hello every1 )08:17
in russia mac os its nothing :)08:18
→ xiangfu joined08:18
AlexeiKozlov is nothing08:18
cbreak-work must be a coold and unfriendly place.08:18
→ Stava joined08:18
rue In Soviet Russia, you own Apples?08:19
cbreak-work evidently not.08:19
AlexeiKozlov yeah ) in soviet russia)08:19
in our country08:19
Stava git pull informs me that it would overwrite one file and it aborts because of that. Little does it know that I actually want to overwrite this file. Can I force the pull?08:19
AlexeiKozlov apples not a computer)))08:19
← francisfish left08:20
AlexeiKozlov biside Russia now is only Russia. Not a Soviet Russia)08:20
cbreak-work Stava: just check out the file if it's tracked08:20
or delete it if it's untracked08:20
tango_ Stava: remove the file08:20
← SunilThaha left08:20
Stava It is untracked08:20
tango_, cbreak-work that worked, thanks08:21
→ LimeBlast joined08:22
rue AlexeiKozlov: I know, it's just a Slashdot meme08:22
← harshpb left08:23
AlexeiKozlov rue doent know about Slashdot08:24
→ overbythere joined08:25
→ tqmz joined08:25
AlexeiKozlov pls help me about bash08:25
it my bin gitolite folder home\git\bin how to write PATH in bashrc to make this folder accessiable/08:26
?08:26
cbreak-work accessible?08:26
← __main__ left08:26
AlexeiKozlov yeah PATH to access files from home\git\bin folder08:27
→ SunilThaha joined08:27
→ netiru joined08:29
→ SkAzZ joined08:29
→ kpreid joined08:30
→ thirdknife joined08:31
AlexeiKozlov nevermind , i done .thanks08:31
→ marcogmonteiro joined08:31
→ superdmp joined08:36
← lausi left08:39
→ brokenshakles joined08:39
← mattalexx left08:39
→ lausi joined08:39
← sq-one left08:39
← arvind_khadri left08:40
→ tatsuya_o joined08:40
→ ipalaus_ joined08:41
→ __main__ joined08:42
← dhruvasagar left08:43
→ dhruvasagar joined08:45
← mastroWork left08:45
AlexeiKozlov and one more time)08:47
system("mkdir", "-p", "$repo.git") and die "$ABRT mkdir $repo.git failed: $!\n";08:47
→ tcurdt joined08:48
AlexeiKozlov can somebody explain?08:48
← tcurdt left08:48
← pfFredd left08:48
AlexeiKozlov if first fail when do die?08:49
wereHamster AlexeiKozlov: do you know perl?08:50
→ Advocation joined08:50
→ pfFredd joined08:50
wereHamster and do you know unix?08:51
→ Chib joined08:51
← jargon- left08:52
AlexeiKozlov no i dont. i run some script . and it gets me errors. i try to explain08:52
→ noisebleed joined08:52
← mischat left08:53
AlexeiKozlov try to solve)08:53
→ williamcotton joined08:54
→ jrots joined08:54
wereHamster haha, just read sitarams last message. Good luck.08:54
→ ryden joined08:54
AlexeiKozlov hm08:56
sad08:56
what for this channel?08:56
sad sad08:56
wereHamster for people who respect other peoples choices08:56
AlexeiKozlov really?08:57
← tokiya_ left08:57
AlexeiKozlov why sitaram doesnt respects my choice?08:57
he did it first08:57
so?08:58
← harish left08:58
AlexeiKozlov dont belive me?08:58
→ mastroWork joined08:58
AlexeiKozlov ask sitaram )08:58
or scroll chat08:58
← Guest83854 left08:59
← elijah left09:00
← JEEB left09:00
BinGOs why is that shelling out to make a directory ?09:00
→ elijah joined09:00
zendeavr → zendeavor09:00
← elijah left09:00
→ JEEB joined09:00
← JEEB left09:01
sitaram BinGOs: because in that version I did not want to use File::Path09:01
BinGOs oh right.09:01
→ elijah joined09:01
→ JEEB joined09:01
→ kronos joined09:02
BinGOs for the record, I got gitolite installed on cygwin fine.09:02
← elijah left09:02
BinGOs I didn't proceed to add any repos though.09:02
AlexeiKozlov oo hince09:02
can u help me/09:02
← babilen left09:02
sitaram BinGOs: then you can help this ... person (for lack of a better word)09:02
BinGOs I have paid $work to do right now.09:03
AlexeiKozlov i too got gitolite isntalled but v2.0 not 2.1 where setup process has changed09:03
BinGOs I just cloned the gitolite repo and ran the instructions from the Quick-Install.09:03
charon isn't that the version that accidentally *cough* doesn't work on windows at all ;)09:04
BinGOs I have removed everything now, as I didn't have a dedicated account and was using my own account.09:04
AlexeiKozlov charon dont get u09:04
BinGOs so u cant help me?09:04
BinGOs I have paid work to do. A load of powershell scripts to write :(09:05
AlexeiKozlov ?09:05
→ matthiaskrgr joined09:05
← apok left09:06
← syphar left09:06
→ mcepl joined09:06
BinGOs When I have some spare tuits I'll have a go at setting up gitolite properly. But it won't be today.09:07
→ syphar joined09:08
charon AlexeiKozlov: just poking fun at sitaram, he posted something on G+ about accidentally making it impossible to run gitolite on windows in one version (i guess it's fixed now)09:08
mcepl I know this is a bit off-topic but could somebody tell me the right command for ReviewBoard's post-review? When I try09:08
post-review --parent master --guess-summary --guess-description --tracking-branch giteng/master --username=something --password=xxxx09:08
I get error diff is empty (but when I run --output-diff I get a nice long diff).09:08
sitaram sadly yes09:08
BinGOs sitaram: I'll feed back to you if I encounter anything out of the ordinary.09:08
AlexeiKozlov charon dotn know about G+09:09
BinGOs &09:09
→ mintux joined09:09
sitaram BinGOs: please join the mailing list and send feedback there; someone who cares might read it ;-)09:09
AlexeiKozlov ))09:09
sitaram charon: the "master" branch, however, is not of that "accidentally not working on Windows" version ;-)09:09
BinGOs I'll just fork and fix >:)09:10
BinGOs & # really this time.09:10
AlexeiKozlov i installed v2.1 and above at home. here cant. some errors in perl code in system command.09:10
→ ignazioc joined09:10
mintux how can I define remote (pull)? when I run git remote -v I saw> web ssh://<ip>/<name.git> (push)09:10
jast mintux: and you just want to change the URL?09:11
→ jargon- joined09:11
charon sitaram: i have a perpetual inverse "problem" here: our windows using students tend to submit programs with system("pause") in them to the automated tester/grader. the security module promptly kills them for attempting to fork()09:11
→ marhaban joined09:11
← subbyyy left09:11
mintux jast:no in eclipse when I click on pull it said to me fatal: No remote repository specified. .. but push worked09:11
← ipalaus_ left09:11
→ ipalaus_ joined09:12
← marhaban left09:12
→ marhaban1 joined09:12
→ berserkr joined09:12
sitaram charon: there's a "pause" command? (or is that in Windows?)09:12
jast sitaram: windows09:12
← dr_win left09:12
→ BiggFREE joined09:12
charon sitaram: windows users apparently need it to hold open the 'cmd' window so they can see the output09:13
← marhaban1 left09:13
→ marhaban joined09:13
jast mintux: do you have tracking info configured for the current branch? (does git branch -vv show something like [web/yourbranch] for your branch?)09:13
→ dr_win joined09:13
FauxFaux charon: Train them to use getchar() / System.in.read() / whatever?09:13
jast charon: yeah, iff they don't run it from the console in the first place :)09:13
it'd be kind of the same on any other system09:13
← marhaban left09:13
→ marhaban joined09:13
← brando753 left09:14
mintux jast:it's my output > * master 315db99 ABout group completed09:14
→ marhaban1 joined09:14
← marhaban left09:14
→ Araxia joined09:14
gitinfo set mode: +v09:15
FauxFaux mintux: It's not interested in the /url/ of the remote, it's interested in which branch you want from the remote, which you need to specify. On the command line, you just add it to the end. Nobody knows how it works in Eclipse.09:15
← marhaban1 left09:15
→ marhaban joined09:15
gitinfo set mode: +v09:15
← wbednarski left09:16
← marhaban left09:16
mintux I use egit in eclipse . in command line it works ... but I don't know why it said you don't have remote09:16
→ marhaban joined09:16
magnus_ hey. while doing an interactive rebase of my branches, it says "[detached HEAD <sha>] as it applies one of the commits. whats wrong?09:16
← marhaban left09:16
→ WillMarshall joined09:16
→ marhaban joined09:17
← blacktulip left09:17
← marhaban left09:17
→ Qoidnef_ joined09:17
sitaram charon: oh ok09:17
← Qoidnef left09:18
→ _iron joined09:18
→ iocor joined09:20
→ arvind_khadri joined09:21
← arvind_khadri left09:21
→ arvind_khadri joined09:21
→ papegaaij joined09:21
AlexeiKozlov i got! something! system("mkdir", "-p", "$repo.git") this has "permission denied" because it wants create folder in \ not in ~ . i wrote system("mkdir", "-p", "/home/git/$repo.git") and it works now. how can i fix it globally? may be in perl some switch ?09:21
→ babilen joined09:23
AlexeiKozlov sitaram any ideas? u should know i suppose :)09:23
← hyperair left09:24
→ f0i joined09:25
← WillMarshall left09:25
jast mintux: oh, actually, I missed part of what you said earlier... "git remove -v" shows *only* the push line?09:25
ignazioc hi09:25
gitinfo ignazioc: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.09:25
mintux jast: yes09:25
ignazioc hi everyone, i need an help. I have a git project, i made a branch for some test (git checkout -b newbranch) today i want to discard this test so i write git checkout master but nothing happened in my working copy…the filesystem remain the same…how can i revert my working copy to the "master" version?09:26
jast mintux: okay. and you want to pull from the same URL?09:26
mintux jast: (push) and (fetch)09:26
jast: yes09:26
jast mintux: oh. okay. the "fetch" one applies to pull, too (pull is fetch+merge)09:26
← joeytwiddle left09:27
mintux jast: is it possible to define pull only not alone fetch ? because EGIT doesn't understand it and said you don't have remote error: fatal: No remote repository specified. Please, specify either a URL or a09:27
remote name from which new revisions should be fetched.09:27
jast mintux: git only has "fetch" and "push" specifications, and any git software that isn't completely stupid and braindead knows that09:28
mintux hmm09:28
→ hyperair joined09:28
jast mintux: my guess is that egit uses the remote "origin" by default... but you don't have that, yours is called "web"09:28
mintux: you can try explicitly telling it what to use, like this: git branch --set-upstream master web/master09:28
← xiangfu left09:29
ignazioc git checkout modify my working copy?09:29
jast ignazioc: the stuff that didn't change, is that uncommitted stuff?09:30
ignazioc committed in the other branch09:30
jast well in that case checkout should switch the affected files over to the version in master09:31
does "git status" say anything about the files in question?09:31
ignazioc when i switch branch so my working copy remain unchanged?09:31
jast well09:31
'git checkout', when switching branches, sets all files to the committed version from the target branch, except if uncommitted changes are involved09:32
rvsjoen_ → rvsjoen09:32
jast those are preserved if possible, otherwise checkout aborts before it does anything09:32
← Qoidnef_ left09:32
jast in that case it gives you an error message09:32
→ Qoidnef joined09:32
ignazioc and for new files?09:33
checkout "remove" them?09:33
← FACEFOX left09:34
← kpreid left09:35
→ kpreid joined09:35
AlexeiKozlov i got solve my problem!!!! o my god! it so simple! then it command executes it tries to access some path that git user have no access!09:35
→ arti_t joined09:36
← ryden left09:36
jast ignazioc: if they were committed to the other branch, then yes09:36
if they weren't... nope :)09:36
checkout will only overwrite/delete files if no data will be lost09:36
→ kanzie joined09:36
AlexeiKozlov this folder is out of cygwin :) that why it cant09:37
jast in the standard form, anyway (checkout can also be made to do other, more destructive things)09:37
→ cortana2 joined09:37
→ FACEFOX joined09:37
← FACEFOX left09:37
→ Qoidnef_ joined09:37
→ joeytwiddle joined09:37
→ batmanian joined09:38
← teknoid left09:38
→ teknoid joined09:39
→ FACEFOX joined09:39
→ tvw joined09:39
← FACEFOX left09:39
← Qoidnef left09:39
→ JPohlman1 joined09:39
AlexeiKozlov sitaram !! it works very well!!))) i got installed gitolite ))09:39
on win 2008 R2 x64 cygwin09:40
ignazioc i tried those commands:09:40
git init09:40
echo "one" > fileone.txt09:40
git add .09:40
git commit -m "first commit"09:40
git checkout -b newbranch09:40
echo "two" > filetwo.txt09:40
git checkout master09:40
← mpenz left09:40
← JPohlman1 left09:40
→ mpenz joined09:40
jast ignazioc: yeah, in that case the file won't be changed back because you didn't commit the change09:40
in fact the file is completely untracked in both branches09:41
ignazioc but after checkout master there are two files fileone.txt and filetwo.txt this is the standard behavior09:41
?09:41
→ FACEFOX joined09:41
← FACEFOX left09:41
jast so git will leave it well alone09:41
← LucasTizma left09:41
jast you never even told git to care about the file. why should it delete it?09:41
→ mastroDani joined09:42
ignazioc because in "master" branch this file wasn't.09:42
ok ok09:42
i understood09:43
→ FACEFOX joined09:43
kanzie If I have a local master and I want to add new remote branch. How can I set this up?09:43
ignazioc before i revert on my branch i need to commit09:43
→ LucasTizma joined09:43
← coredumb left09:44
← iocor left09:44
← mastroWork left09:45
← orkz left09:46
← AlexeiKozlov left09:47
← Qoidnef_ left09:47
→ Qoidnef joined09:47
jast kanzie: create new local branch, do whatever you like in it, git push origin yourbranch09:47
→ coredumb joined09:48
→ Qoidnef_ joined09:48
→ legnaleurc joined09:48
← ignazioc left09:49
← replore_ left09:49
kanzie jast: trying to figure out how to add new remote branch09:49
→ iocor joined09:50
← Qoidnef left09:52
→ thelazydeveloper joined09:52
→ replore_ joined09:52
← Lujeni left09:53
← replore_ left09:53
→ Qoidnef joined09:54
← Qoidnef_ left09:54
wereHamster kanzie: add a remote? git remote add ...09:55
← Qoidnef left09:55
kanzie but how do I specify a remote branch to pull from09:56
wereHamster kanzie: through tracking info09:56
kanzie: man git-branch -> --set-upstream09:56
gitinfo kanzie: the git-branch manpage is available at http://jk.gs/git-branch.html09:56
→ ocharles joined09:56
→ Qoidnef joined09:56
← timoxley left09:57
← arti_t left09:57
→ archdori joined09:58
← netiru left10:00
→ jds joined10:01
← Fandekasp left10:02
→ harish joined10:02
← williamcotton left10:04
← kadoban left10:05
→ mabrand joined10:06
← dr_win left10:06
→ alamaiso` joined10:09
→ flavius joined10:10
← alamaison left10:11
→ jatt joined10:11
→ netiru joined10:12
→ arti_t joined10:14
← alamaiso` left10:14
→ sdeboey joined10:14
→ pastime joined10:16
jast kanzie: sorry, to me your question meant: how do I create a new branch in the remote repository :)10:16
→ williamcotton joined10:16
kanzie np10:18
I got it now10:18
← hyperair left10:18
→ angelsl joined10:19
nimesh_accenture git ls-remote shows me one of the branches as "refs/heads/master" , so if i want to clone that particular branch is this the right command : "git clone http://git.gitorious.or.bla.bla -b master" ?10:21
← stroll2 left10:22
→ achadwick joined10:22
FauxFaux nimesh_accenture: Yes?10:22
→ falena joined10:22
→ mischat joined10:22
nimesh_accenture Fauxfaus: so you mean to say that is the right command?10:22
jast nimesh_accenture: you know it clones all branches by default, right?10:22
nimesh_accenture nopes ... i dodnt know that !10:23
jast so now you do :)10:23
nimesh_accenture yup!10:23
so is there a way to just clone a particular branch that i want?10:23
jast it just doesn't show you all of them after cloning (it creates local mirrors but not standard branches that you can work in), but you can still use "git checkout" to get 'em (which creates a standard local branch and sets things up correctly)10:24
nimesh_accenture and is above the right way?10:24
jast nimesh_accenture: you can do that, but why bother? the overhead due to extra branches is usually pretty low10:24
(and no, -b will just determine which one is checked out during cloning)10:24
nimesh_accenture ok... so the source code that i see will be the source code of that branch?10:25
jast yes10:25
nimesh_accenture ok.. cool10:26
thx jast10:26
jast master is typically the default, so chances are you don't even need to specify it10:26
the person maintaining the upstream repository specifies a default branch that you'll see if you don't use any extra arguments10:26
in 99% of cases it's master10:26
nimesh_accenture ya but there are other repositaries where i might want to check out one specific to my hardware10:27
→ OPK joined10:27
jast sure10:27
nimesh_accenture thx jast10:27
jast if you end up specifying the wrong one, you can always just checkout a different one after cloning10:27
nimesh_accenture ok...10:28
jast just to clarify: "git checkout foo" switches to branch foo, without fetching anything via the network10:28
nimesh_accenture ok!10:28
thx dude10:29
→ sbell joined10:30
→ whyking_ joined10:32
Advocation hey, I'm getting the message "-bash: git: command not found" when I type git into the command line10:32
have I broke it?!10:32
→ jalama joined10:34
→ Xizor joined10:34
→ dbgster joined10:34
→ oehman joined10:35
→ ryden joined10:36
→ enmand joined10:36
← enmand left10:36
m1sc Advocation: what says "which git"?10:40
← kyconquers left10:40
Advocation m1sc: sorry?10:41
→ mastroWork joined10:41
m1sc Advocation: probably git isn't installed or not in your PATH -- execute "which git" to check10:42
→ gjaldon joined10:42
Advocation "which git" returns nothing10:43
it was working, then I installed rbenv which is working fine, but git isn't now10:43
I wonder if my PATH is messed up10:43
← dhruvasagar left10:43
Advocation m1sc: echo $PATH returns - /Users/Advocation/.rbenv/shims:/Users/Advocation/.rbenv/bin:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/usr/X11/bin10:44
I'm guessing there should be a reference to Git in there...10:44
← kanzie left10:45
← mastroDani left10:45
← deebo left10:45
← cjs left10:45
→ Superlokkus joined10:45
→ Alexei joined10:45
gitinfo set mode: +v10:45
Alexei hi every1 again )10:46
→ dhruvasagar joined10:46
← ldlework left10:46
Alexei i setup gitolite well on windows 2008 R2 x64 in cygwin . works very well across all domain machines through SSH )10:46
so the question is " how i can add my repos into gitolite repo_base folder"?10:46
→ hyperair joined10:47
→ honschu_ joined10:47
→ beautifulmind joined10:47
Alexei i already have some repos and now i need to move them to gitolite repo_base folder10:47
→ felgru joined10:48
← sbell left10:48
→ sbell joined10:48
Alexei please?10:48
FauxFaux Alexei: Push them.10:49
m1sc Alexei: create those repos within gitolite-admin, add a new remote gitolite in each of the existing repos, push10:49
← sbell left10:49
→ ldlework joined10:49
← honschu left10:50
→ Anubisss joined10:51
→ dubd joined10:54
gitinfo set mode: +v10:55
Alexei oh really10:55
so simple)10:55
my history will be saved?10:55
→ Kaisyu joined10:56
bigkitty hi all10:56
how to enable git command autocompletation on centos?10:57
→ Kaisyu2 joined10:57
dubd bigkitty: im using zsh with oh-my-zsh and autocompletion works quite nice for git10:58
→ whitman joined10:58
→ mauror joined10:58
→ _Qman joined10:59
← ehsan left10:59
← f0i left10:59
bigkitty dubd but i want to use bash:(10:59
jast well you'll have to use the bash-completion then11:00
if centos doesn't install it for you, you can grab it from git's own repository, somewhere in contrib/11:00
← arvind_khadri left11:00
jast don't ask me how to set it up, though... probably all you'll have to do is source it in your profile11:01
← Kaisyu left11:01
Alexei one more Q: how i can refresh keys in authorized_keys from keydir ?11:02
m1sc you can just link the file from /etc/bash_completion.d/11:02
sitaram Alexei: are you AlexeiKozlov?11:02
Alexei sitaram yeah11:02
sitaram hmm ok11:02
Alexei lazy to type )11:02
sitaram, do u help me?11:02
sitaram no11:03
the gitolite master TOC at http://sitaramc.github.com/gitolite/master-toc.html11:03
Alexei why?11:03
sitaram ^11:03
← _Qman left11:03
Alexei )11:03
ok i see11:03
can u teel me that mmm11:04
where is that in TOC/11:04
&11:04
?11:05
hard to find there11:05
→ mcluskyd_ joined11:05
sitaram Alexei: (1) you changed your nick in a way that -- to me -- looks unnecessary (and gets around my /ignore). (2) you made code changes that should not have been necessary, to gitolite, so I expect other things to go wrong and you to keep coming back. But since I don't consider Windows usable as a server (despite success stories), I can't/won't help11:05
← mcluskydodallas left11:07
Alexei sitaram about (2) i revert changes back ) gitolite setup dont need this changes.11:07
it goes well w/o changes.11:07
hm . sad. u think that i get around your ignore? very sad. i change my nick because i access this channel through web.11:08
somebody please help in that Q: how i can refresh keys in authorized_keys from keydir ?11:09
← elnn left11:09
→ elnn joined11:10
→ Vortex34 joined11:10
Alexei i know how to do this by hands. i want auto11:11
and gl-setup-authkeys can do this. but working directory is unknown11:12
→ timoxley joined11:13
→ ph^_ joined11:14
→ williamcotton_ joined11:14
→ iamjarvo joined11:14
sitaram Alexei: http://sitaramc.github.com/gitolite/index.html#rtfm11:15
← _Vi left11:15
← williamcotton left11:15
williamcotton_ → williamcotton11:15
← patrakov left11:16
→ sbell joined11:17
← ph^ left11:17
sitaram and I already gave you a link to http://sitaramc.github.com/gitolite/master-toc.html -- did you at least *attempt* to find what you want in there?11:17
→ master-lincoln joined11:17
wereHamster IOW, Y U no read documentation?11:18
Alexei sitaram yeah i did. i find about gl-setup-authkeys11:20
and didnt found11:21
wereHamster find11:22
sitaram Alexei: gl-setup-authkeys is not mentioned anywhere in the docs except in the *developer notes* so stop bluffing. You went to the source code11:22
Alexei hm11:23
→ blackmatrix_ny joined11:23
← hyperair left11:23
sitaram Alexei: I'm sorry but I am at work and I do not have time for unproductive conversations with people who lie; you'll probably change your nick again but I'll burn that bridge when I come to it11:23
← sandstrom left11:23
Alexei omg sitaram)11:23
→ arvind_khadri joined11:23
← arvind_khadri left11:23
→ arvind_khadri joined11:23
Alexei i m too on work11:23
← mcepl left11:23
Alexei i asked u kindly11:24
if u dont want the do not11:24
← sdeboey left11:25
→ EtgarDizz joined11:25
→ setmeaway2 joined11:25
wereHamster then11:25
→ sdeboey joined11:25
Alexei yeah) mistype11:25
wereHamster you11:25
← archdori left11:25
Alexei aga11:26
→ f0i joined11:26
← setmeaway left11:27
Alexei sitaram read this book "you body cries for water" very useful book. its my advice.11:28
← pfFredd left11:29
← Sitethief left11:29
← master-lincoln left11:29
→ master-lincoln joined11:30
Bartzy How can I check if a local branch in my repo was pushed to a remote branch ?11:30
Alexei later i publish manual on russia about gitolite on cygwin and windows. it will be my second publish about that/11:30
Bartzy git remote show origin ?11:30
will only show for origin :|11:30
← _AlbireoX left11:31
← iocor left11:31
wereHamster Bartzy: what you can do is to check whether a local branch points to the same commit as a branch in a remote repo11:31
but git does not keep track of who pushed what where11:31
← jalama left11:33
→ enmand joined11:33
→ jalama joined11:33
← gjaldon left11:34
← OOPMan left11:34
→ sbell_ joined11:35
→ iocor joined11:36
← iocor left11:36
→ _AlbireoX joined11:37
← sbell left11:37
sbell_ → sbell11:37
→ hyperair joined11:38
Bartzy wereHamster: How can I check that ? rev-parse ?11:39
← sbell left11:39
Alexei bb every111:39
← Alexei left11:39
→ crippa joined11:40
crippa hi. question: I've cloned a repo, I edited a file, and run "git commit". If I pull the repo, the file is not changed. why?11:41
← f0i left11:41
wereHamster Bartzy: yes, rev-parse local branches, ls-remote to check the remote11:43
→ fayimora joined11:45
bremner crippa: are the changes in the local repo (git log)? where are you pulling from and to?11:45
→ ph^ joined11:45
← cortana2 left11:46
← oehman left11:46
crippa bremner: I have a local dir called "repo" and I pull from there to another dir.11:47
← zub left11:47
crippa if I edit a file, then run git commit, it says " Not currently on any branch." "no changes added to commit (use "git add" and/or "git commit -a")"11:48
← ph^_ left11:48
canton7 crippa, for a start, checkout a branch: git checkout branchname11:48
crippa and # modified: settings.xml11:48
I don't know if there is any branch actually11:49
canton7 crippa, 'git branch' ?11:49
crippa * (no branch) master11:49
← JohnBat26 left11:49
canton7 then you've got one branch, called master.11:49
crippa ok..11:49
→ JohnBat26 joined11:49
→ cccaldas joined11:49
canton7 you currently don't have that checked out, however, which means that any commits you make could be lost11:49
← juvenal left11:49
← arti_t left11:50
canton7 'git checkout master' to check it back out11:50
← babilen left11:50
crippa M settings.xml11:50
Switched to branch 'master'11:50
→ Superlokkus_ joined11:50
crippa this is what I have if I run the checkout11:51
so I think it's already checked out, right?11:51
and it recognize settings.xml as locally modified11:51
or not?11:51
canton7 hrm? It thinks you have uncommitted changes to settings.xml11:51
← Superlokkus left11:52
Superlokkus_ → Superlokkus11:52
crippa yes, I modified the file.11:52
I tried to commit these changes by typing git commit -m blablabla11:52
canton7 you need to stage it first. 'git add settings.xml'11:52
(or stage all uncommitted changes and commit in one go with 'git commit -a -m "Message"'11:52
crippa but I get the message: # Changes not staged for commit:11:53
# (use "git add <file>..." to update what will be committed)11:53
# (use "git checkout -- <file>..." to discard changes in working directory)11:53
#11:53
# modified: settings.xml11:53
#11:53
# Untracked files:11:53
# (use "git add <file>..." to include in what will be committed)11:53
#11:53
# settings.xml~11:53
no changes added to commit (use "git add" and/or "git commit -a")11:53
zokko use pastebin11:53
canton7 please use a pastebin for long pastes11:53
crippa sorry..11:53
canton7 crippa, as git explicitely tells you, settings.xml isn't staged for commit. Use 'git add <file>' to update what will be committed11:54
→ jsoa joined11:54
crippa ok that's the problem: if I run git add, and then commit, everything works.11:56
canton7 yup. Looks like you need to read up on git a bit. !book is good11:56
gitinfo There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !gitt !vcbe and !parable11:56
→ especially-corn joined11:57
especially-corn hello, what can i do if a download freezes?11:57
crippa But I'm trying to use jenkins to pull from the repo, and as I do that the local files are overwritten by the 'old' version in the repo...11:57
wereHamster especially-corn: what does that have to do with git?11:57
especially-corn it happened 3 times in a row Receiving objects: 14% (2787/19735), 295.14 MiB | 430 KiB/s11:57
← iori left11:57
especially-corn well i guess it does have something to do11:57
wereHamster especially-corn: ah, you're trying to *clone* a repo11:57
or *fetch*11:58
especially-corn not sure11:58
i just repo sync11:58
cbreak-work what's that?11:58
wereHamster git: 'repo' is not a git command. See 'git --help'.11:58
canton7 it's an android thing11:58
cbreak-work oh, that mobile thing11:58
especially-corn yes11:58
wereHamster crippa: I don't think it makes sense to have local changes in jenkins11:58
cbreak-work I would not try to clone/fetch hundreds of megabytes over celular networks11:58
especially-corn cbreak, neither would i11:59
← jsoa left11:59
crippa wereHamster: what do you mean?11:59
← ph^ left11:59
especially-corn it always freezes at about 14-16%11:59
→ Ryback_ joined11:59
← Ryback_ left11:59
→ Ryback_ joined11:59
wereHamster crippa: why do you have local changes?11:59
especially-corn anything i can do to keep whatever it downloaded?11:59
wereHamster no11:59
→ ph^ joined11:59
especially-corn because if i kill the process it goes back to (0/19735) and reattempts to download everything12:00
wereHamster yes12:00
cbreak-work especially-corn: git doesn't really have resumable communication12:00
you can try to fetch branches one by one in the hope to get less data at a time12:00
especially-corn so there is no pause, resume12:00
cbreak-work but that's it12:00
crippa wereHamster: I pulled from a repo, made changes and wrote back to the repo. If I then ask jenkins to 'checkout' the repo, I get the old (unmodified') version of the files12:00
→ erichynds joined12:00
wereHamster especially-corn: is that an initial clone or are you trying to update an existing repository?12:01
canton7 crippa, where is jenkins pulling from?12:01
especially-corn wereHamster, initial12:01
wereHamster !bundler12:01
gitinfo https://bundler.caurea.org/ will create a bundle of a repository that you can download using resumable transfer mechanisms, and then clone. This helps with slow or broken networks.12:01
gitinfo set mode: +v12:01
crippa the git repo12:01
canton7 crippa, which git repo? The one on your machine, or the one you cloned from?12:01
EtgarDizz having trouble fetching from remote, getting "error: unable to find 177......". anyone?12:01
wereHamster crippa: are you sure your changes made it to the repo that jenkins is pulling from?12:01
crippa wereHamster: I'm pretty sure12:02
canton7 EtgarDizz, please paste the command you ran and the full output into a pastebin12:02
crippa, do you understand what 'distributed' means in the context of git?12:02
wereHamster crippa: so what's preventing jenkins from updating its working tree?12:02
← master-lincoln left12:02
crippa canton7: no..12:02
→ xiangfu joined12:03
crippa wereHamster: I have no clue, it should work right?12:03
canton7 crippa, do you understand that if you clone from a repo, then commit, no changes are made to the repo you pulled from?12:03
wereHamster crippa: simple question: did you push after you committed the changes?12:03
EtgarDizz canton7, http://pastebin.com/bq2CGE1y12:03
wereHamster EtgarDizz: the remote repo is corrupted. restore from backup12:03
← tango_ left12:03
← especially-corn left12:03
crippa here we are.12:04
I did not, I thought commiting was enough12:04
thanks a lot12:04
wereHamster you thought. hm.12:04
crippa: go read !book12:04
gitinfo crippa: There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !gitt !vcbe and !parable12:04
→ tango_ joined12:04
crippa I should, yes.. thanks12:05
← dr_lepper left12:05
wereHamster but you won't. for some reason..12:05
← ldlework left12:05
→ cannonball joined12:05
→ dr_lepper joined12:07
← dr_lepper left12:07
→ dr_lepper joined12:07
EtgarDizz canton7, you mean there's no way to find the cause and fix it? I'm pretty sure I know which commit caused it, can that help?12:07
→ cortana2 joined12:07
← stkrzysiak left12:08
← thirdknife left12:08
canton7 EtgarDizz, what did that commit do?12:08
wereHamster the cause is irreleant. The repo is corrupted. You need to restore it.12:08
irrelevant even.12:08
EtgarDizz canton7, wereHamster: don't understand the question, it's just the last commit that was made, before it all was good. of course cause it irrelevant, just wondering if it's possible to pinpoint the problematic files and erase/restore them, instead of the whole repo12:09
canton7 EtgarDizz, have you done anything unusual in your local repo? Have you had any problems with the remote at all? tbh I'm not sure whether the problem is client- or server-side12:11
← dpino left12:11
→ griotspeak joined12:12
← nimesh_accenture left12:12
← pjn_oz left12:14
EtgarDizz canton7, I've had some trouble w/ permissions, not able to push stuff, so went to repo and did "chmod -R g+w ." and added appropriate users to the group12:14
→ FernandoBasso joined12:15
canton7 hrm, might a failed push cause this? I thought those things were atomic, but that might fall down when permission issues get involves12:16
→ apok joined12:17
← Qoidnef left12:17
← Advocation left12:17
→ nimesh_accenture joined12:18
EtgarDizz canton7, I thought it's atomic too... I also thought that if the pushing user is in group w/ write permission to the objects folder, it should be enough. wrong?12:19
→ jim0203 joined12:19
→ thiago_ joined12:20
← griotspeak left12:20
→ griotspeak joined12:20
→ thirdknife joined12:21
← xGambit left12:21
canton7 EtgarDizz, my guess is that a failed push *before* you resolved the permissions might have caused this12:21
← griotspeak left12:21
EtgarDizz canton7, is it possible to "re-push" the problematic commit from the local copy in which it was made?12:22
→ n1x joined12:22
← n1x left12:25
← thiago_ left12:25
→ thiago joined12:25
gitinfo set mode: +v12:25
← iamjarvo left12:26
crippa I pull, modify a file, "git add" it, "git status" says "# Changes to be committed: myfile", "git commit" says " 1 files changed", but when I push it says "Everything up-to-date12:26
why??12:27
cbreak-work you didn't commit?12:27
canton7 EtgarDizz, I honestly have no idea. Worth a try I guess. You'll need someone who's familiar with the internals of git's fetch/push mechanisms to answer that properly12:28
sajbar cbreak-work: if you just add a file it isn't commited.12:28
cbreak-work I know.12:28
crippa I did. after having run "git commit" it says "" 1 files changed",12:28
sajbar make that a reply to crippa12:28
crippa: then you can git push whereever12:28
cbreak-work crippa: and git log shows your commit?12:29
← SunilThaha left12:29
crippa and so I did, but it replies with a "everything is up to date"12:29
so teh file is not pushed12:29
cbreak-work what did you push? the branch you committed in?12:29
git doesn't push files anyway12:29
only pushes history12:29
sajbar crippa: git push whereever branchname12:29
→ SunilThaha joined12:29
→ ToBeFree joined12:30
crippa should I always specify the branchname? I only have 1 branch,the master12:30
gitinfo set mode: +v12:31
→ Advocation joined12:31
cbreak-work you don't need to in all cases.12:31
← mintux left12:32
crippa ok, so "git push" would be enough right?12:33
→ dr_leppe1 joined12:35
→ iocor joined12:35
← nimesh_accenture left12:35
← gavin_huang left12:35
← dr_lepper left12:37
← noisebleed left12:37
→ gavin_huang joined12:38
→ jsoa joined12:38
→ noisebleed joined12:38
← noisebleed left12:38
→ noisebleed joined12:38
→ d0k joined12:39
→ martinjlowm joined12:40
→ mastroDani joined12:41
→ t4nkd joined12:43
← erichynds left12:44
→ freeed joined12:44
← mastroWork left12:45
→ master-lincoln joined12:45
← apok left12:45
→ oriba joined12:47
→ chrisshattuck joined12:49
← blackmatrix_ny left12:50
→ subbyyy joined12:51
← subbyyy left12:51
→ subbyyy joined12:51
← subbyyy left12:51
gitinfo set mode: +v12:53
→ duckxx joined12:54
→ babilen joined12:54
← chrisshattuck left12:55
→ erichynds joined12:56
← dc5ala left12:57
→ TeckniX joined12:57
← iocor left12:57
→ juvenal joined12:58
→ gchristensen joined12:58
← gchristensen left12:58
→ macabre joined12:58
→ revagomes joined12:58
← Guest87934 left13:00
→ judaz joined13:00
← pastime left13:01
→ bartek joined13:02
judaz hello people, I need some help with git submodule. I've added a submodule locally and pushed it to the repo (github). At github I see that the .gitmodules has been updated, and it's exactly as mine locally. But, When I do a git clone --recursive the new submodule is not cloned AND when I check the .gitmodules files, my submodule does not exist. Any idea how to debug this or what can it be?13:02
→ AaronMT joined13:03
→ iamjarvo joined13:03
judaz also, in a git submodule status, the new module does not exist13:03
→ glcrazy joined13:03
judaz I see now that when I do a git submodule init, the submodule appears..13:03
but I need it to work on the clone --recursive13:04
← teratoma left13:05
← ph^ left13:05
→ Radium_ joined13:07
← Radium left13:08
← hyperair left13:08
PerlJam judaz: is your clone a mirror or bare?13:08
judaz PerlJam: you lost me there, I don't know the difference13:08
→ nimesh_accenture joined13:09
PerlJam judaz: then it's probably not relevant :)13:09
judaz PerlJam: hehe13:09
PerlJam judaz: If you "git clone --mirror" or "git clone --bare", even with --recursive, you won't get submodules initialized13:09
→ bauruine joined13:09
judaz PerlJam: oh!, no.. I'm not doing none of them13:10
PerlJam judaz: so you did "git clone --recursive URL" and that's it?13:10
judaz PerlJam: true13:10
PerlJam weird13:10
← nimesh_accenture left13:10
→ cholin joined13:10
judaz PerlJam: indeed13:10
oh wait.. I've tried doing a clone at my machine and worked13:11
→ hyperair joined13:11
judaz but the problem exists on the server (ec2)13:11
→ OOPMan joined13:11
← Chib left13:11
← dvide left13:11
← dhruvasagar left13:11
← rchavik left13:11
PerlJam judaz: maybe it's running a really old git client?13:12
← cooloney left13:12
→ dhruvasagar joined13:12
judaz PerlJam: ill check13:12
but the thing is13:12
the project has several submodules13:12
and it's not cloning only ONE of them13:12
I mean, the problem i've described happens only in one of them, but the other 5 works13:13
→ dpino joined13:13
PerlJam Hmm.13:13
judaz and the url of the submodule is public and over https:13:13
PerlJam Then ... what's different about that one? :)13:13
← d3vic3 left13:13
→ kyconquers joined13:14
judaz PerlJam: whats the difference between clone --recursive and git submodule init/update13:14
?13:14
cus, when I do an init and update13:14
the submodules is added13:14
but not during the clone13:14
tizzo → tizzo-afk13:15
→ d3vic3 joined13:15
tizzo-afk → tizzo13:15
→ martinjlowm-lptp joined13:15
→ mastroWork joined13:15
← w0ng left13:17
→ chrisshattuck joined13:18
→ w0ng joined13:18
→ neurodrone joined13:19
← mastroDani left13:19
→ v0n joined13:20
→ beatak joined13:21
→ orkz joined13:23
→ jim0203_ joined13:25
judaz PerlJam: I see a difference at the last module13:26
it also has a submodule13:26
so is13:26
MAIN -> submodule -> submodule 213:26
although, it works here at my machine, but not at EC213:26
← teknoid left13:26
← BiggFREE left13:26
→ andrewvos joined13:27
← duckxx left13:27
andrewvos how do I `git checkout -b meh` and not fail if the branch *does* exist?13:27
→ adhawkins joined13:27
wereHamster andrewvos: if branch exists then checkout else checkout -b13:27
→ mattdipasquale joined13:27
← mattdipasquale left13:27
→ mattdipasquale joined13:28
andrewvos wereHamster: I'm doing this in code13:28
← tonyj_ left13:28
andrewvos Oh wait13:28
-B looks interesting13:28
wereHamster 'in code'?13:28
← jim0203 left13:28
jim0203_ → jim020313:28
← kermit left13:28
← jim0203 left13:28
→ tonyj joined13:29
andrewvos yay -B worked13:29
wereHamster: I'm writing a script.13:29
→ kermit joined13:30
wereHamster you can check the presence of a branch in a script..13:30
← bauruine left13:30
→ mrwalker joined13:31
→ gchristensen joined13:33
→ mantic0 joined13:33
→ jim0203 joined13:33
→ Vortex35 joined13:33
→ ceej joined13:34
→ tyatpi joined13:35
→ FreeArtMan joined13:35
← Vortex34 left13:36
← gchristensen left13:37
→ _Vi joined13:37
tizzo → tizzo-afk13:37
→ fisted_ joined13:37
← fisted left13:37
← tonyj left13:38
→ pandeiro joined13:38
→ Davey joined13:41
→ iamjarvo1 joined13:41
← iamjarvo left13:42
← iamjarvo1 left13:43
→ iamjarvo joined13:43
→ mastroDani joined13:43
← marthinal left13:44
← Advocation left13:44
← kpreid left13:45
→ codewrangler joined13:46
← w0ng left13:47
← mastroWork left13:47
→ cakehero joined13:47
→ w0ng joined13:47
→ tonyj joined13:48
← mattdipasquale left13:50
→ aristidesfl joined13:51
→ dv310p3r joined13:51
→ oehman joined13:52
→ mattdipasquale joined13:52
→ Nightfire034 joined13:52
→ eagles0513875 joined13:52
← oriba left13:52
← tonyj left13:53
eagles0513875 hey guys what does one need to do to merge a particular branch13:53
cbreak-work switch to the destination branch and type git merge sourcebranch13:53
← DofLeFou left13:53
← [[thufir]] left13:53
← andrewvos left13:53
→ dfr|mac joined13:53
→ marthinal joined13:54
← Maccer left13:54
eagles0513875 cbreak-work: destination branch in my case master then git merge BRANCH NAME13:55
of the branch you want to merge right13:55
→ ehsan joined13:55
digiv|afk → digiv13:56
← dv310p3r left13:56
→ chepre joined13:57
← swappoke left13:57
← mischat left13:57
→ dfr|mac_ joined13:57
→ nadavoid joined13:58
← glcrazy left13:58
→ juanfer joined13:58
→ gchristensen joined13:59
← robbyoconnor left13:59
→ wbednarski joined14:00
→ kpreid joined14:00
ironfroggy why cant i do "git branch -D origin/branch" ?14:00
→ Maccer joined14:00
← dfr|mac left14:00
SethRobertson I don't know. Why can't you? !repro14:01
gitinfo Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.14:01
cbreak-work ironfroggy: don't mess with remote tracking branches14:01
SethRobertson git remote prune origin14:01
ironfroggy i need to, its wrong. somehow a remote tracked branch is not what it really is on the remote.14:01
eagles0513875 ironfroggy: its just git branch -D BRANCH NAME14:01
bremner eagles0513875: no14:01
→ d2dchat joined14:02
ironfroggy when i checkout the branch, it includes changes from a local branch i never pushed14:02
← Kaisyu2 left14:02
cbreak-work ironfroggy: doesn't matter.14:02
→ mischat joined14:02
cbreak-work ironfroggy: git remote update first if you want to change it14:02
(to the current state of the remote)14:02
SethRobertson ironfroggy: git remote update -p origin; Then if you still think it is bad we can compare14:02
ironfroggy: git rev-parse origin/branch; git ls-remote origin | grep branch14:03
← pandeiro left14:03
← mischat left14:05
→ mischat joined14:06
← JohnBat26 left14:07
← dfr|mac_ left14:08
→ dfr|mac joined14:08
← chrisshattuck left14:08
ironfroggy SethRobertson: it does not LOOK like it has this other branch14:08
but when i check it out, inevitably, the same changes are always there14:08
← ehsan left14:09
adhawkins I created a branch locally, then pushed it up to a repository. Is this branch now bring tracked correctly or do I need to do something else to track it?14:09
SethRobertson You are not being clear. What was the result of the rev-parse/ls-remote commands I mentioned?14:09
adhawkins How can I tell which branches are being tracked and which aren't?14:09
SethRobertson adhawkins: git branch -vv Look in []14:09
→ lyddonb joined14:09
jast adhawkins: git branch -vv. chances are that didn't magically set tracking info, unless you used -u with push.14:09
SethRobertson adhawkins: If you did not push with -u, you didn't change anything14:09
ironfroggy SethRobertson: they both show the same hash, the head of the remote branch i'm trying to checkout14:10
adhawkins I did a git push origin branchname, which has created the branch at the remote, but it doesn't appear to be tracked14:10
andy@atom:~/software/libmusicbrainz/libmusicbrainz.git$ git branch -vv14:10
* bug-fixing-2012-04 ead2865 Update NEWS in preparation for new release14:10
master df0ce24 [origin/master] Remove README.txt14:10
ironfroggy it is the branch that i was on when i originally checked it out, with a clean working copy14:10
SethRobertson ironfroggy: So what makes you think they are not the same if the two hashes are identical? You just don't *think* that is what is on the branch?14:10
jast adhawkins: yeah, so use -u next time, or just branch --set-upstream14:10
→ macrover joined14:10
← Vert left14:11
→ brngardner joined14:11
adhawkins git branch --set-upstream bug-fixing-2012-04 origin/bug-fixing-2012-0414:11
Seems to have done it14:11
That correct?14:11
SethRobertson ironfroggy: Remember that master, origin/master, and origin's master are three different branches14:11
ironfroggy i know this14:12
SethRobertson So why do you think origin/master and origin's master are different?14:12
ironfroggy the problem is that when i do "git branch origin/master origin/branch" i see changes that i have personally made in another branch and never pushed to origin yet14:12
jast adhawkins: looks good14:12
I gotta go14:12
adhawkins Excellent, thanks14:13
jast see y'alls later14:13
SethRobertson ironfroggy: whoa whoa. If you are doing command like you said, you are making a major mistake. Do `git branch` Do you see anything like origin/ in the output? `git branch | grep origin/`14:13
→ tonyj joined14:13
SethRobertson The command you mention makes a local branch with a remote name. A major major problem.14:14
ironfroggy SethRobertson: typo!14:14
i meant diff, not branch14:14
SethRobertson Well, do that `git branch| grep origin/` anyway. It might explain things14:14
← mischat left14:14
ironfroggy no origin in there14:15
→ mischat joined14:15
← gavin_huang left14:15
eagles0513875 hey guys how will i know that a branch has been merged i did git branch and the branch i merged is still listed there14:15
once the branch has been merged can it be deleted?14:15
bremner eagles0513875: that is why git branch -d exists14:15
SethRobertson ironfroggy: So...perhaps you made a mistake and pushed it anyway? the default setting of push.default in current git will cause branches other than the one you are on to be pushed as well.14:15
eagles0513875: You can run `gitk --all --date-order` to see what happened too14:16
ironfroggy if i did i dont know how it would end up in some else's branch, but i'll look closer14:16
eagles0513875 bremner: ok14:16
jepler I cloned a repository with --shared, and later realized this was not desirable in my use case. What do I do to "un-share" the repository?14:17
← fisted_ left14:17
← mischat left14:18
→ mischat joined14:18
eagles0513875 hey guys how do i remove trackign on a remote branch14:21
remote server14:21
SethRobertson jepler: Just chmod the directory recursively and delete core.shared man git-config14:22
gitinfo jepler: the git-config manpage is available at http://jk.gs/git-config.html14:22
jepler git clone -s is the one that sets up objects/info/alternates, not the one that sets core.sharedRepository14:22
→ chrisshattuck joined14:22
jepler (that's exactly the mistake I made)14:22
SethRobertson eagles0513875: Can you restate your desire? Possibly `git remote remove` or possibly something else. I could not understand14:22
jepler: Oh, that shared. `git repack -Ad` I think14:23
→ envi_ joined14:23
→ chrisgeorge joined14:23
→ duckxx joined14:23
→ mikepack joined14:23
eagles0513875 SethRobertson: i have a repository i merged with master and deleted how do i remove said reposityr on the remote from being tracked14:23
jepler SethRobertson: ah yes14:24
SethRobertson You want to remove origin/foo? You cannot. You can remove origin's foo and then do a pruning remote update14:24
jepler now I can delete alternates and git fsck is happy14:24
thank you SethRobertson14:24
→ gavin_huang joined14:24
← Nightfire034 left14:25
eagles0513875 SethRobertson: git push origin :BRANCH NAME <--- then what does that do from what i looked up that removes the trackign on the remote14:25
adhawkins If I delete a remote branch (git push origin :branchname) presumably all its history is still maintained?14:25
← tqmz left14:26
→ Nightfire034 joined14:26
eagles0513875 git push origin username:BRANCH NAME <-- is that right assumption adhawkins14:26
adhawkins eagles0513875: Don't ask me questions, I'm still a newbie :)14:27
eagles0513875 wait14:27
this is the error i get SethRobertson error: src refspec jaquilina does not match any. error: failed to push some refs to 'jaquilina@git.eagleeyet.net:/home/jaquilina/intradev'14:27
does that mean the merge also merged the remote tracking for me into master?14:28
← gchristensen left14:28
SethRobertson eagles0513875: `git push --delete origin foo` will delete branch foo on origin. Then `git remote update -p origin` will update your copy of origin and delete stale branches like origin/foo14:28
→ airborn joined14:28
eagles0513875 SethRobertson: doesnt this do the same thing git push origin jaquilina:djangoAndSouthSetupScript14:28
→ mischat_ joined14:28
SethRobertson You can use `git ls-remote origin` to see what branches are still on origin.14:28
← mischat left14:29
mischat_ → mischat14:29
SethRobertson No, that does something completely different. It does not delete djangoAnd... it uploads jaquilina to the branch named djangoAnd... on origin14:29
eagles0513875 Branch ls-remote set up to track remote branch master from origin. SethRobertson when i run git ls-remote origin14:29
← envi_ left14:30
SethRobertson Um, you didn't run that command then. You ran some other command. !repro14:30
gitinfo Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.14:30
→ fayaz joined14:32
tizzo-afk → tizzo14:32
← rrn left14:32
← brentdanley_ left14:32
← chrisshattuck left14:32
adhawkins If I delete a remote branch (git push origin :branchname or git push --delete origin branchname) presumably all its history is still maintained?14:33
→ ericmuyser joined14:34
→ rrn joined14:34
→ bloony joined14:35
shruggar adhawkins: on the remote? not much, no14:35
adhawkins Even if I've been pushing the branch up to the remote?14:35
→ envi_ joined14:35
bloony whats the best way of looking at a previous file? eg. how the file looked 5 commits back14:35
→ pandeiro joined14:36
shruggar adhawkins: it removes the remote reference, which means the remote will consider it okay-to-delete next time gc runs14:36
SethRobertson adhawkins: For a while. Eventually, if no other ref points to those commits, the unreferenced history will go away14:36
eagles0513875 SethRobertson: :D problem git push --delete origin BRANCH NAME :D did the trick14:36
SethRobertson shruggar: After a few weeks14:36
← duckxx left14:36
→ duckxx joined14:37
adhawkins shruggar, SethRobertson: Ok, I assumed that it would still retain the history from when that branch was being worked upon.14:37
→ danielpunt joined14:37
adhawkins Or will that history be present in the merges back to the main branch (master in this case)?14:37
SethRobertson adhawkins: If you merged the branch, then there is still a reference to it, and the history will not be lost14:38
→ fisted joined14:38
adhawkins Ah ok, that makes sense.14:38
← mischat left14:38
SethRobertson Until such time as you delete the reference to the branch you merged it into14:38
→ mischat joined14:38
SethRobertson Unless you merged that. etc14:38
adhawkins SethRobertson: Well, the merge will go into 'master' which won't get deleted.14:38
→ psoo_ joined14:38
→ pastime joined14:38
← psoo left14:38
→ Radium joined14:39
→ chrisshattuck joined14:40
← macrover left14:40
← Radium_ left14:40
eagles0513875 thanks for your patience and help with me SethRobertson :D14:40
← psoo_ left14:40
→ ology joined14:41
→ alexisabril joined14:41
→ mastroWork joined14:42
← mischat left14:42
→ macrover joined14:43
→ mischat joined14:43
← duckxx left14:44
← beautifulmind left14:44
→ duckxx joined14:44
← SimonNa left14:45
← brokenshakles left14:45
← mastroDani left14:45
→ ceej_ joined14:46
← lyddonb left14:46
← ceej left14:46
ceej_ → ceej14:46
→ storrgie joined14:47
← duckxx left14:47
→ duckxx joined14:47
← w0ng left14:48
neersighted|AFK → neersighted14:48
← eagles0513875 left14:48
→ lyddonb joined14:48
← danielpunt left14:48
→ w0ng joined14:49
← Strogg left14:49
← ehks` left14:50
→ Strogg joined14:50
← mattdipasquale left14:52
→ brokenshakles joined14:52
harish → harish_afk14:53
→ lacrymology joined14:53
→ pfFredd joined14:54
← pandeiro left14:55
← storrgie left14:56
→ storrgie joined14:56
← xiangfu left14:57
→ Stone44 joined14:57
← __main__ left14:57
← giallu left14:58
→ __main__ joined14:59
← EricInBNE left14:59
← mikepack left14:59
→ saml joined14:59
saml how can I list the files taht's version controlled? do I export?15:00
← __main__ left15:00
→ joaoh82 joined15:00
→ stuntmachine joined15:01
→ __main__ joined15:02
→ gusnan joined15:03
← neurodrone left15:03
bremner saml: man git ls-files15:03
gitinfo saml: the git-ls-files manpage is available at http://jk.gs/git-ls-files.html15:03
bremner or, if doing something weird, see man git-ls-tree15:04
gitinfo the git-ls-tree manpage is available at http://jk.gs/git-ls-tree.html15:04
saml whoa didn'tk now... it's not under git help15:05
→ coyotebush joined15:05
← jatt left15:05
← ThomasLocke left15:05
→ rye_ joined15:06
→ Zenopus_ joined15:06
→ negev joined15:07
← brar left15:07
negev hi, is there a simple way to select which submodules you want to check out in a given repo? we were thinking of having a bunch of optional modules for our application as git submodules, but it seems that the submodule concept in git tends towards having /all/ the submodules checked out rather than just the ones you want15:08
→ zealiod joined15:08
← Zenopus left15:08
Zenopus_ → Zenopus15:08
zealiod is it possible just to allow a user only access to one directory of a repo?15:08
shruggar zealiod: no15:08
bremner not easily15:08
shruggar not at all15:08
bremner gitolite can fake it, iirc15:08
zealiod cool15:09
shruggar gitolite can fake it for writes, not reads, afaik15:09
bremner but probably, ask yourself why you are doing this?15:09
→ ehsan joined15:09
← nickcooley left15:09
→ broonie joined15:09
→ mikepack joined15:10
bremner since it is mostly like better and easier to seperate out things into different repos.15:10
← Stone44 left15:10
← broonie left15:10
→ spiette joined15:10
bremner shruggar: hmm. That makes sense. I only remember some trick involving temporary repos; but absent running filter-branch, it does sound hard to accomplish,15:11
← dfr|mac left15:12
shruggar it sounds like more of an explanation as to why it's not possible :)15:13
→ cgrage joined15:13
bremner well, you know, you could pay gnomes to type in some content from that branch ane email it to you, but I agree that might not count as git at that point.15:14
← wolog left15:14
← rye_ left15:14
shruggar everything can be solved with a large enough budget and mechanical turk, provided you're okay with high response time15:15
→ francisfish joined15:16
→ psakrii joined15:17
zealiod hmm15:17
i just want one of our team just to be able to read and write to part of the repo… mainly because I don't want him looking at all the other code in there15:17
→ iocor joined15:17
→ Carraway joined15:18
bremner make a submodule or use gitslave for that part15:18
← zealiod left15:19
bremner or not15:19
→ stkrzysiak joined15:19
→ wolog joined15:20
→ TheHaven joined15:20
← stkrzysiak left15:20
→ Vert joined15:20
→ stkrzysiak joined15:20
← flijten left15:21
→ aleray joined15:21
← resmo_ left15:22
← noisebleed left15:22
aleray hi, can I create a branch for which starting point is an old revision of master ?15:22
← tsdh left15:23
shruggar aleray: of course. git checkout -b <new name> <starting-point>15:23
← pfFredd left15:23
shruggar <starting-point> can be any commitish15:23
→ zamba joined15:23
aleray shruggar, nice, easy :)15:23
zamba is gitosis still used?15:23
shruggar !gitosis15:24
gitinfo gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite15:24
zamba thanks :)15:24
← negev left15:24
zamba but not a deb package :p15:24
cortana2 should be an easy backport15:25
in fact, 2.0.3 is in squeeze-backports :)15:25
zamba what i want is the following: two developers working on a web project.. php code and mysql database.. one of the developers are using windows and the other is using linux.. and then we have the production web server where the project should be running..15:26
gitolite good for this?15:26
cortana2 well... git is a component of such a setup, but you would need other stuff15:26
i will say this right now... if your project is going to be deployed on linux, then the devlopers must use linux too15:26
same goes for windows, etc15:26
zamba ok..15:26
but what other stuff are you talking about?15:27
cortana2 anything else will just drain your sanity in the long run15:27
→ marhaban joined15:27
cortana2 well, i was thinking of a system for deploying the code to the web server, etc... a lot of people manage to misuse git for that with some success :)15:27
← codewrangler left15:27
zamba what kind of system for deploying the code do you suggest then?15:27
i was thinking of just using git there directly15:28
git update or whatever15:28
canton7 zamba, !deploy15:28
gitinfo zamba: [!website] Git is not a website deployment tool, but can sometimes play one in sufficiently simple/lax environments with a little help. One example of help is: http://toroid.org/ams/git-website-howto and another approach: https://gist.github.com/171423515:28
cortana2 ah that factoid is more useful than i remember :)15:29
← juanfer left15:29
← pro_metedor left15:30
← infogulch_ left15:30
→ noisebleed joined15:30
← noisebleed left15:30
→ noisebleed joined15:30
zamba i like that first approach15:30
→ mritz joined15:31
zamba though i don't fully understand git remote15:31
or the distributed model here :p15:31
basically both developers need to commit/checkout from the same central repository15:31
← FACEFOX left15:32
cortana2 uh huh. that much is easy to set up15:33
see chapter 4 of the user guide, or chapter 3 of the git community book15:34
→ theCole joined15:34
→ kenperkins joined15:34
cortana2 http://schacon.github.com/git/user-manual.html and http://book.git-scm.com/ respectively15:34
← joaoh82 left15:34
→ t4nkd`workin joined15:35
← oehman left15:36
← pastime left15:36
← OOPMan left15:37
→ iqueeee joined15:37
← gavin_huang left15:38
← t4nkd left15:38
← iqueeee left15:39
← Vampire0 left15:39
→ juanfer joined15:39
→ giallu joined15:39
← juanfer left15:40
→ qos joined15:40
→ cccaldas_ joined15:40
→ neurodrone joined15:40
→ SimonNa joined15:41
← papegaaij left15:42
← _sundar_ left15:42
← arvind_khadri left15:42
← rtuin left15:43
← cccaldas left15:43
← t4nkd`workin left15:43
→ kukks joined15:44
→ t4nkd joined15:44
← angelsl left15:45
← batmanian left15:50
→ JohnBat26 joined15:51
← chepre left15:52
← Poapfel left15:53
→ Vampire0 joined15:53
→ Poapfel joined15:54
→ xou joined15:54
← Poapfel left15:54
→ Poapfel joined15:54
← mastroWork left15:56
→ ziggles joined15:57
← lyddonb left15:57
→ notola joined15:57
← brngardner left15:57
→ lyddonb joined15:59
→ ShadesofGraylin| joined16:00
→ n1x joined16:00
→ subh joined16:00
← kermit left16:00
→ juanfer joined16:01
→ hagabaka joined16:01
← hagabaka left16:01
→ hagabaka joined16:01
→ t4nkd`workin joined16:01
← n1x left16:01
← ShadesofGraylin| left16:02
← qos left16:02
← t4nkd left16:03
t4nkd`workin → t4nkd16:03
← shruggar left16:03
← jceb left16:03
→ CannedCorn joined16:04
← t4nkd left16:04
← fisted left16:04
← EtgarDizz left16:04
→ t4nkd joined16:05
← kenperkins left16:05
SethRobertson jast?16:05
← gusnan left16:05
← whitman left16:05
SethRobertson jast: .trigger_edit fetch4 Don't use the four argument version of `git fetch <repo> <branch>` or `git pull <repo> <branch>`. It will only end in tears. It updates FETCH_HEAD instead of <repo>/<branch> and always seems to utterly confuse anyone who is foolish enough to use it.16:05
← thirdknife left16:06
← mishok13 left16:07
SethRobertson jast: Also, it would be nice if someone said "Please see !last." if gitinfo would strip "." if it didn't match (or perhaps even better did de-suffixing as a second pass if that didn't match)16:08
→ shruggar joined16:08
← Sigma left16:09
← jrots left16:10
→ maestrojed joined16:10
← mritz left16:11
→ brngardner joined16:11
jast SethRobertson: what about the first one?16:11
← kpreid left16:11
← comcor left16:11
→ comcor joined16:11
jast (I'll do that . thing when I find time)16:11
← robotmay left16:11
SethRobertson jast: It doesn't allow [0-9]16:11
← TheHaven left16:12
canton7 hmm, I raised the "fetch4" issue on the git ML a little back, and Junio and peff are in favour, and peff's even got a patch. I just haven't had the time to figure out why it breaks the tests it does... If anyone's got time, please http://thread.gmane.org/gmane.comp.version-control.git/19225216:12
jast SethRobertson: oh, right. will take care of that, too, then.16:12
← flaguy48 left16:13
← thierryp left16:13
→ ph^ joined16:13
← dr0id left16:14
→ macmartine joined16:14
→ joaoh82 joined16:14
← matthiasgorgens left16:14
← noisebleed left16:15
→ SegFaultAX|work joined16:15
→ kermit joined16:17
→ Advocation joined16:17
→ eph3meral joined16:17
← RobertLaptop left16:18
← orkz left16:18
→ maletor joined16:19
← Desproges left16:19
→ mritz joined16:19
→ dr_win joined16:19
→ DarkAR joined16:20
→ dfr|mac joined16:21
→ dfr|mac_ joined16:21
← macrover left16:22
→ Sigma joined16:23
← Sigma left16:23
→ Sigma joined16:23
→ RobertLaptop joined16:23
← adhawkins left16:23
neersighted → neersighted|AFK16:24
← Advocation left16:24
→ noisebleed joined16:24
← noisebleed left16:24
→ noisebleed joined16:24
← pmatthews left16:24
→ michaelmartinez joined16:25
← michaelmartinez left16:25
← theCole left16:25
← dfr|mac left16:25
→ mxweas_ joined16:26
← niro left16:27
→ Mizoa joined16:28
→ johnkpaul joined16:28
← wbednarski left16:28
johnkpaul does anyone have a good resource for understanding why git push origin :branchname deletes the branch16:28
I vaguely know that the way you name a branch means something16:29
but what the colon does, I want to learn16:29
canton7 johnkpaul, http://progit.org/book/ch9-5.html16:29
johnkpaul great, thank you canton716:29
sitaram shruggar, bremner: gitolite can fake out read-restrictions using a "partial copy" repo but only for branches not for directories. http://sitaramc.github.com/gitolite/g3/non-core.html#partial-copy16:29
canton7 johnkpaul, also there's git push --delete, which is a nicer alias for the same16:29
SethRobertson johnkpaul: localname:remotename If the localname is empty, you are saying push nothing there, or delete it16:29
→ kenperkins joined16:29
shruggar syntax for git push: git push <where> <local-object>:<remote-ref>16:29
→ dreiss joined16:30
→ usmanity joined16:30
→ nbigaouette joined16:30
→ jehoshua02 joined16:30
johnkpaul ooh I see16:30
← whyking_ left16:30
johnkpaul awesome, that's simpler than I thought16:30
← mxweas_ left16:30
shruggar the "colon" means "I didn't specify anything to push to the remote ref"16:30
johnkpaul git push has a lot of different ways to use it16:30
jehoshua02 I need to restore a file back to where it was in a previous commit.16:30
canton7 jehoshua02, checkout in file mode: git checkout HEAD^ -- path/to/file16:31
(or HEAD, if you haven't committed yet)16:31
johnkpaul so when you say "remote ref" a ref is a particular branch?16:31
canton7 branches are refs. Some other things are refs, too (like tags)16:31
→ revagomes_ joined16:32
shruggar a ref is a ref. It's a pointer to an object. Usually this is a commit object, but it could be a tag or a blob16:32
johnkpaul oh ok, I'm looking at the progit info16:33
← dfr|mac_ left16:33
← revagomes left16:33
shruggar or a tree…. though I can't think of why one would do that16:33
→ oriba joined16:33
jehoshua02 Thanks.16:33
← jehoshua02 left16:33
→ ISF joined16:33
→ dfr|mac joined16:33
neersighted|AFK → neersighted16:33
← glennpratt left16:34
→ raatiniemi joined16:35
← oriba left16:35
johnkpaul thanks shruggar and canton7, definitely learning a lot16:36
← jargon- left16:36
→ oriba joined16:36
← jlebrech left16:37
nbigaouette Hi all, I'm using git for my code which outputs a lot of data to "output/". There's nothing in that folder managed by git except .gitignore (which contain only "*") and .kdev_ignore (which tells Kdevelop to ignore the content). The problem is that I want to keep my data centralized and organized. I thus remove the "output" folder and create a link instead. But then git cannot do some operations like rebase.16:37
Doing "git checkout -- output/.gitignore" deletes the symbolic link and re-create the folder, breaking all my running simulations...16:38
→ gerard0` joined16:38
nbigaouette Is it possible to have the "output" folder tracked by git (so a clone will create the folder with .gitignore * and .kdev_ignore) but then ignore everything done to it?16:39
← Mizoa left16:40
← gerard0 left16:40
← kenperkins left16:40
shruggar nbigaouette: why not have your "output" folder get created by your build process, including moving a gitignore template into place?16:41
← asakura left16:42
→ kpreid joined16:42
→ flaguy48 joined16:42
→ jrots joined16:42
eph3meral is there a way to share the results of rerere ?16:43
nbigaouette shruggar: Yes could be. That still leave the git question unanswered though... I wish I knew hwo to stash/rebase without breaking my running simulations...16:43
← Sigma left16:43
→ jargon- joined16:43
eph3meral as in, it appears that after running a merge with rerere enabled, I now have a .git/rr-cache folder (which I presume contains the results of my decisions and merging)16:43
→ asakura joined16:44
eph3meral but I have multiple people working with me on the same topic branch, and if one of them wants to rebase, we don't want them to have to re-make the same decisions that I made16:44
shruggar nbigaouette: it answers the question. It says: "Problem with tracking a file in git? Then don't track that file in git"16:44
thiago rebased commits are your own16:45
you don't share those commits, which means that other people won't rebase them16:45
accordingly, the rerere cache is your own too16:45
← daaku left16:45
← overbythere left16:46
→ mintux joined16:46
mintux when I want push my repo to the server on internet the speed is too low. i use ssh to connect to remote16:46
it's to slow16:47
← notola left16:47
← ipalaus_ left16:47
→ brentdanley_ joined16:47
bremner sitaram: ah, memory corruption on my part, not too surprising.16:47
SethRobertson mintux: So...get a faster network connection? Ship git-bundles via fedex?16:48
nbigaouette shruggar: Putting the "not using git" solution aside, I think the issue here is with the symbolic link. I remember having read some things about it but maybe I can get a clearer picture here. I understand git can have issue with tracking a symbolic link versus the content of the file, but what about links to folders?16:48
→ Araxia_ joined16:48
→ ipalaus_ joined16:48
SethRobertson nbigaouette: Git can use symlinks with whatever destination just fine. Not all OSen can use symlinks of course16:49
nbigaouette What's the reason for not following the link and keeping track of the folder pointed to?16:49
mintux SethRobertson: i only this connection here. my ftp upload speed is 60 KB/s but in git is 516:49
SethRobertson nbigaouette: Git tracks the symlink. If you want to track the destination you need to track that in the same or different repo. Note that git does not track directories, only files in directories16:49
← cakehero left16:49
mintux but here I have this connection16:49
← judaz left16:50
nbigaouette With symbolic links, I mean "track that folder, then change the folder with a symolic link to another." I though git should have tracked the files in the linked folder just fine.16:50
SethRobertson mintux: If you have a linux system as your ssh destination/git server, try `dd if=/dev/zero bs=8k count=8k | ssh destination time dd of=/dev/null` and see how fast that is.16:50
nbigaouette But then some operations (like checkout -- folder/file) will delete the link and put back the folder...16:50
mintux SethRobertson: run it on my local repo or server remote ?16:51
SethRobertson nbigaouette: Exactly so. Perhaps you should try using --bind mounts.16:51
← revagomes_ left16:51
→ revagomes_ joined16:52
tango_ stupid google, if I search for 'gitolite wildcard repositories' I want the sitaramc page, not everything else BUT16:52
SethRobertson mintux: Run that command locally. It measures the speed of ssh to the remote server, in case ssh is being traffic-shaped or something16:52
→ slide joined16:52
tango_ sheesh16:52
nbigaouette SethRobertson: bind mounts are a good idea. They require root though or setting sudoers...16:52
← revagomes_ left16:52
→ revagomes_ joined16:53
← tewecske left16:53
← revagomes_ left16:53
nbigaouette Again, if git tracks files and not folders, why is it deleting the symbolic link to replace it with a folder? Shouldn't it be able to just traverse the link and work with the file in the pointed-to folder?16:53
→ revagomes joined16:53
SethRobertson If you are treating the symbolic link like a directory, git is treating it like a directory and feels free to replace the link with a directory. Trying to fool git never ends well16:54
mintux SethRobertson: I run this : dd if=/dev/zero bs=8k count=8k | ssh 85.17.235.94 -p 2233 time dd of=/dev/null the cursor is blinking16:54
→ tewecske joined16:55
← tewecske left16:55
← slide23 left16:55
→ pastime joined16:56
mintux SethRobertson: what should I see in output ?16:56
SethRobertson: it doesn't finish16:56
→ JEntrep joined16:56
nbigaouette Is there a technical reason why git wouldn't be able to follow the links and work with the files there instead of going inside the folder?16:56
→ Sigma joined16:56
nbigaouette I'm just trying to understand how git "thinks" so I don't loose my data again16:56
→ slide23 joined16:57
SethRobertson Then lower "count=8k" to "count=1k" or something. You should get a duration and possibly a rate. If you don't get a rate you need to multiple 8k*1k (or whatever the count is) and then *8/(minutes*60+seconds) to get a bitrate16:57
→ Mowsh joined16:57
← dreiss left16:57
← cortana2 left16:58
→ apok joined16:58
SethRobertson nbigaouette: Git will track a symbolic link as an object to be tracked. Git will track a file as an object to be tracked. Git will track a subrepository as an weird object to be tracked. It also understands directories as path components. That is all16:58
→ whyking_ joined16:59
→ michaelmartinez joined16:59
→ orkz joined16:59
← slide left16:59
SethRobertson However, when someone clones the repo, they are going to get a directory, not a symbolic link to somewhere not necessarily present on their system. When you switch branches and the new working tree no longer has that object, what do you expect to happen?17:00
There are many technical reasons why what you want to do would be confusing to any VCS17:00
← fission6 left17:00
→ tewecske joined17:01
nbigaouette Ok, I think I get it. I tell git to track path/file, but then is sees symlink/file. Because it treats links as files, it will not follow "symlink" to track "file", it will think "symlink" is a new file to track. Thus, using "git checkout -- symlink/file", it sees "symlink" in the way of "file" instead of following it and replace the link with a folder.17:01
SethRobertson Not exactly. But the take-away is, "don't do that"17:02
→ glcrazy joined17:03
nbigaouette Except that git does not work with that now, why?17:03
← JEntrep left17:03
→ theCole joined17:04
nbigaouette I mean, I still think that technically it should be possible. Tell git to track folder/file. Replace "folder" with "symlink", "file" is present in "symlink" too. Git should just follow symlink to get to file.17:04
← d3vic3 left17:05
nbigaouette I understand that adding "symlink/file" to the repo the first time would be confusing. But this is not what I'm doing.17:05
→ lindenle_ joined17:06
gitinfo set mode: +v17:06
→ mishok13 joined17:06
→ d3vic3 joined17:06
SethRobertson It does. But it also feels free to remove the symlink because you didn't ask to track it.17:06
→ alvaro_o joined17:06
SethRobertson That is why I said "not exactly"17:06
← alvaro_o left17:07
→ slide joined17:07
← tewecske left17:08
→ leavittx joined17:08
→ chinopozo joined17:08
← lacrymology left17:09
← slide23 left17:09
mintux SethRobertson: I got some result so what should I do now?17:09
← tyatpi left17:09
SethRobertson pastebin them if you don't want to do what I said17:10
Along with the command17:10
← lasers left17:11
→ dstorrs joined17:11
→ tewecske joined17:11
nbigaouette Hum ok thanks for the clarifications.17:11
← sdeboey left17:12
← shruggar left17:13
mintux SethRobertson http://codepad.org/eNHrpUXZ17:13
← ajpiano left17:15
→ ajpiano joined17:16
← chinopozo left17:17
→ daaku joined17:17
→ slide23 joined17:17
→ qt-x joined17:17
SethRobertson OK, 46kb/s. Now you need to test git. On the server, `git init --bare /tmp/testrepo` and on the client `git clone server:/tmp/testrepo` then `cd testrepo; dd if=/dev/urandom count=1k bs=8k of=randomfile; git add randomfile; git commit -m randomfile; time git push -v --progress` You might want to run the "time git push" command under `script` so that you can more easily upload the output, but whatever.17:17
mintux SethRobertson the ssh connection is too slow . is it possible git protocol instead of it ?17:18
← jargon- left17:18
SethRobertson mintux: You can do the same test except have the "server" and "client" be on the same system and use "localhost" as the "server" name in the git-clone command. It should be blazingly fast unless your disks are stupid.17:19
← slide left17:19
← dfr|mac left17:20
→ dfr|mac joined17:20
mintux SethRobertson:on local network is too fast but I though the ISP do something on ssh speed so if I can use another protocol such as http or git my problem be solved17:21
← glcrazy left17:21
← DarkAR left17:22
SethRobertson mintux: We observed that your ssh speed is 46 kb/s. I thought you said ftp got 50 kb/s. Do you expect a faster speed?17:22
nbigaouette For the posterity, my "output" folder is now created from the makefile using "order-only prerequisite". See "Types of Prerequisites" @ http://www.gnu.org/software/make/manual/make.html#Prerequisite-Types17:22
mintux SethRobertson: no 50 KB/s17:22
SethRobertson For the posterity, you said that in the wrong channel.17:22
mintux SethRobertson: and download is 150 or 200 KB/s17:22
← marcogmonteiro left17:23
SethRobertson mintux: Ah, well dd said 46 kB/s anyway. Or 370kb/s17:23
In other words, the same speed17:24
→ mxweas_ joined17:24
← achadwick left17:24
→ dfr|mac_ joined17:25
← brngardner left17:25
← dfr|mac left17:26
→ dfr|mac__ joined17:26
→ Chillance joined17:26
→ dfr|mac joined17:26
→ TikaL joined17:26
→ arvind_khadri joined17:27
← JohnBat26 left17:27
→ whitman joined17:27
← mintux left17:27
← ipalaus_ left17:28
→ rzec joined17:29
→ slide joined17:29
← dfr|mac_ left17:29
→ krz joined17:29
← dfr|mac__ left17:30
→ ejb joined17:30
← slide23 left17:31
← ajpiano left17:31
→ frnkkpp joined17:32
→ ajpiano joined17:32
ejb Hi, can I have git repos within git repos? For example, if I have foo/* in a repo. But foo/bar/ be a different repo?17:32
→ glcrazy joined17:33
← pppaul left17:33
→ brngardner joined17:33
← leavittx left17:33
→ pppaul joined17:34
ejb Can I just put foo/bar/ in .gitignore and treat it as separate?17:34
eagles0513875_ → eagles051387517:34
offby1 ejb: that's called a "submodule"17:35
← pastime left17:35
FauxFaux It'll automatically add itself to the secret gitignore.17:35
m1sc ejb: !subrepos17:35
gitinfo ejb: [!subprojects] So, you want to add git repositories inside of other git repositories? Well, you have four main options. First is to just do it, add the repo to the outer project's .gitignore, and treat them entirely separately. Best if they are entirely seperate. Otherwise your best options are "!submodule" "!gitslave" and "!subtree" Try typing those commands into this IRC channel.17:35
→ pastime joined17:35
FauxFaux I thought.17:35
→ fission6 joined17:35
ejb great, thank you17:35
→ lacrymology joined17:37
← revagomes left17:37
→ revagomes joined17:37
← tatsuya_o left17:38
← Bartzy left17:39
→ rockets joined17:40
→ pfFredd joined17:40
→ knapper_tech joined17:40
→ dreiss joined17:40
← aleray left17:41
gitinfo set mode: +v17:41
knapper_tech how can i get old files out of the COMMIT_EDITMSG without committing the change to the rest of the version control?17:41
→ btanaka joined17:44
← hagabaka left17:44
← erichynds left17:45
m1sc knapper_tech: why do you want to touch this file and what problem do you try to solve?17:45
← alexisabril left17:46
← master-lincoln left17:47
→ Russell^^ joined17:48
→ kadoban joined17:48
→ spaceonline joined17:48
← zackiv31 left17:49
← dr_leppe1 left17:49
← orkz left17:50
→ dr_lepper joined17:50
← dr_lepper left17:50
→ dr_lepper joined17:50
← tvw left17:51
← nbigaouette left17:51
→ ThePendulum1 joined17:51
gitinfo set mode: +v17:51
ThePendulum1 Greetings17:51
eph3meral is there a better way to share merge conflict recordings other than sharing .git/rr-cache folders?17:51
← txomon|home left17:51
→ daaku_ joined17:51
ThePendulum1 I have been using git for quite a while now, but only personally. I now want to work with two people using Bitbucket. How do I allow him to push to Bitbucket?17:52
→ webflo joined17:52
→ pandeiro joined17:52
← cyphase left17:52
← daaku left17:53
daaku_ → daaku17:53
→ cakehero joined17:53
→ txomon|home joined17:53
← txomon|home left17:53
→ txomon|home joined17:53
charon checks if there is a #bitbucket17:53
→ leavittx joined17:54
eph3meral ThePendulum1, go to the admin tab of your repo (at the top)17:54
m1sc eph3meral: interesting question. afaik no (don't even use rerere), maybe you want to ask on the ML17:54
eph3meral ThePendulum1, then go to the "access management" menu item in the left menu17:54
← marhaban left17:54
← dstorrs left17:54
eph3meral ThePendulum1, then you can add users, there will be a search box and some buttons [read] [write] [admin] to give that level of access to said users17:54
← whyking_ left17:55
→ Advocation joined17:55
eph3meral ThePendulum1, if you want to give them write access, they must checkout the repo using SSH (not using HTTP)17:55
ThePendulum1 eph3meral: Thanks a lot! I feel quite retarded for not figuring that out myself.17:55
eph3meral m1sc, yeah, rerere is awesome :)17:55
ThePendulum1 eph3meral: What exactly is the difference between checkout and pull?17:55
→ the_cybersphinx joined17:56
→ aristidesfl1 joined17:56
← tewecske left17:56
→ erichynds joined17:56
→ whyking_ joined17:57
eph3meral ThePendulum1, sounds like you're probably coming from subversion, yes?17:57
← srcerer left17:57
← mischat left17:58
ThePendulum1 eph3meral: I've used subversion for about a week a few years ago. I just haven't used git for more than pulling in the morning, committing during the day, and pushing in the evening, lol17:58
eph3meral ThePendulum1, you could probably benefit from essentially throwing away everything you used to know about subversion, and reading this http://progit.org/book/17:58
← alex__c2022 left17:58
cbreak checkout is for branch switching, pull is for fetching+merging in one command17:58
eph3meral ThePendulum1, ok, well checkout is how you change branches17:58
← Nightfire034 left17:58
→ kenperkins joined17:58
← jfelchner left17:58
← cybersphinx left17:58
eph3meral ThePendulum1, pull is for retrieving other people's changes from a remote repo17:58
→ mischat joined17:59
eph3meral ThePendulum1, actually retrieving and merging said changes, because as cbreak just mentioned, pull is actually an alias for: git fetch; git merge;17:59
← aristidesfl left17:59
eph3meral ThePendulum1, honestly, you should really read http://progit.org/book/ ... I've been using git for a few years now, but hadn't ever read that book, and I wish I had read that book much earlier17:59
← daaku left17:59
eph3meral ThePendulum1, it's actually pretty short, and you'll actually understand what you're doing :)18:00
← theCole left18:00
→ mischat_ joined18:00
ThePendulum1 eph3meral: Thanks for the link, I've been looking for that book for a while :p18:00
→ daaku joined18:00
→ n1x joined18:01
eph3meral ThePendulum1, if you've been using git for a while, you can skip some of chapter one, i.e. the sections on installing and first time setup etc, but the history and basics sections are almost critical because git doesn't work like other repos, so understanding how git thinks about your files and changes will be very useful down the road18:01
ThePendulum1, really, if you read chapters 1-3 (which you can certainly do in under an hour I think), you'll be good to go, read the other ones later at your leisure18:02
knapper_tech m1sc: Every time I commit the number of "deleted" files is gigantic (30k). These files don't exist anywhere in the repo, they just have some history information18:02
ThePendulum1 eph3meral: I think 2, 3 and 5 are the most relevant now18:03
← mischat left18:03
mischat_ → mischat18:03
m1sc knapper_tech: meaning "I don't what to have all the old file in each commit"?18:03
← netiru left18:04
← pandeiro left18:04
← eicto left18:04
→ eicto joined18:05
← btanaka left18:05
→ tewecske joined18:06
m1sc ..want..files..18:06
FauxFaux s/files/penis/18:06
← tewecske left18:06
← ISF left18:07
ThePendulum1 Hmm... so if I continue working on a project that is hosted on Bitbucket, and someone may have made changes, what should be the first thing I do?18:07
← glcrazy left18:07
RichiH looks at FauxFaux18:07
← cenuij left18:07
ThePendulum1 Checkout? Pull?18:07
eph3meral ThePendulum1, you won't be able to push until you pull18:07
→ glcrazy joined18:07
eph3meral ThePendulum1, so, you won't really have to worry about that18:07
→ cenuij joined18:07
→ dstorrs joined18:07
eph3meral ThePendulum1, stop thinking in terms of subversion, checkout is a subversion thing18:07
RichiH ThePendulum1: pull18:08
← cenuij left18:08
→ cenuij joined18:08
eph3meral ThePendulum1, git checkout is very different from subversion checkout18:08
RichiH that will fetch & merge (if possible)18:08
then, you resolve conflicts, if any18:08
commit the resolution18:08
ThePendulum1 eph3meral: I don't even know why I started about checkout, I've never used it and I just thought I should do so18:08
RichiH hack hack hack18:08
→ cyphase joined18:08
RichiH and push18:08
ThePendulum1 So I shouldn't be using checkout?18:08
→ infogulch_ joined18:08
RichiH no18:08
ThePendulum1 Okay18:08
→ tewecske joined18:09
RichiH well, yes18:09
ThePendulum1 oh18:09
← tewecske left18:09
RichiH argh18:09
let's say "correct"18:09
eph3meral ThePendulum1, until you know what it is18:09
ThePendulum1, no, you don't need checkout18:09
RichiH "So I shouldn't be using checkout?" -> "that is correct"18:09
eph3meral ThePendulum1, you only need commit, push and pull... now please, just read chapters 2 and 3 of progit.org18:09
RichiH hates negated questions18:09
→ pandeiro joined18:09
ThePendulum1 If I am working on a project with a friend of mine, all we do is pull from bitbucket when we start working, commit every now and then, and push it to bitbucket as soon as it is worthy18:09
RichiH eph3meral: arguably, add will come in handy, as well18:10
→ estevocastro joined18:10
eph3meral RichiH, hence why he should read progit.org before asking further questions18:10
← bartek left18:10
ThePendulum1 That's what I was about to do anyway, I don't think I'm missing any information18:10
RichiH eph3meral: there's that18:10
eph3meral ThePendulum1, we are a manual supplement, not a manual replacement :)18:10
ThePendulum1 I just got confused because I thought someone told me I should use checkout, lol18:10
RichiH ThePendulum1: the short version is that git is hard to grasp at first and you just have to dive in headfirst18:10
eph3meral yeah so stop it, and just read the book, kthx bye18:10
ThePendulum1 I don't get what's going on now. I was already doing what I should be doing, so what's the point of reading the book once again?18:11
→ ISF joined18:11
← freeed left18:11
RichiH the longer version is that you need to at least skim the book, think about it, and then skim it again, slower this time18:11
ThePendulum1 I have read it multiple times18:11
RichiH ah18:11
→ bemson joined18:11
→ bartek joined18:12
← FreeArtMan left18:12
ThePendulum1 That's why I came here with a question about checkout, because I couldn't recall how it was relevant to what I was willing to do. Turns out I don't need to use if after all, so I'm back to where I was.18:12
← fmilo left18:12
→ tktiddle joined18:13
tktiddle How can i give a user permissions to push to a repo?18:14
_ikke_ tktiddle: !context18:14
gitinfo tktiddle: Without giving us more context, it's impossible to tell what's going on. Please provide details about what your situation is. (!repro)18:14
tktiddle I have a git repo and i want to add another user to it, so that can push to the repo, currently there is a perission denied error when they try to push18:15
→ adhawkins joined18:15
_ikke_ tktiddle: git itself doesn't handle authorization18:16
tktiddle: How do they access the repo?18:16
tktiddle i think its a gitolite server, i.e over ssh18:16
_ikke_ tktiddle: Ah18:16
→ wicchan joined18:16
_ikke_ tktiddle: Do you have admin access?18:16
tktiddle sorry didnt explain18:16
→ Bass10 joined18:17
→ rosascarlatta joined18:17
tktiddle yeah i can sudo on the server18:17
→ sq-one joined18:17
_ikke_ Do you have the gitolite-admin repository?18:17
← rosascarlatta left18:17
← Vert left18:18
← JEEB left18:18
tktiddle hmm, looks like not18:19
← mxweas_ left18:19
→ JEEB joined18:19
→ aristidesfl joined18:20
→ fmilo joined18:20
tktiddle Ive got some packages galled gitk and gitweb installed18:20
→ mdhopkins joined18:20
tktiddle dosn't look like a gitolite atall18:20
← mischat left18:20
_ikke_ tktiddle: Who set up gitolite?18:20
← Danielpk left18:21
← aristidesfl1 left18:21
← rurufufuss left18:21
→ marcogmonteiro joined18:21
← glcrazy left18:21
tktiddle commander18:21
on holiday18:21
_ikke_ http://sitaramc.github.com/gitolite/adminpush.html18:21
ThePendulum1 Thanks for your support, cheers!18:21
→ glcrazy joined18:21
← ThePendulum1 left18:21
_ikke_ tktiddle: gitolite is administered through the gitolite-admin repo18:21
← osX64 left18:22
tktiddle its not here, what other type of git server could it be?18:22
→ Vert joined18:23
← jefferai left18:23
_ikke_ tktiddle: git doesn't require a server18:23
→ glennpratt joined18:23
_ikke_ tktiddle: repositories are normally just accessed through ssh18:23
← cbreak left18:23
tktiddle cool18:23
→ cbreak joined18:24
tktiddle wish i could work it though18:24
_ikke_ tktiddle: Although I can recommend using gitolite18:24
← whyking_ left18:24
→ mastro joined18:25
_ikke_ tktiddle: But gitolite doesn't have to be installed through the package manager18:25
→ jefferai joined18:25
→ pantsman joined18:26
← pantsman left18:26
→ pantsman joined18:26
→ zackiv31 joined18:27
→ Ky0sh1r0 joined18:28
→ hagabaka joined18:29
← hagabaka left18:29
→ hagabaka joined18:29
adhawkins I'm sure this is an FAQ, but a quick google isn't giving much help.18:30
What's the closest equivalent to $Id$ in git?18:30
_ikke_ man gitattributes18:30
gitinfo the gitattributes manpage is available at http://jk.gs/gitattributes.html18:30
adhawkins I'm interested in the committer and commit date in particular18:30
What does ident provide?18:30
_ikke_ adhawkins: But what do you want to achieve?18:31
adhawkins To show when a file was last changed, and who did it18:31
In the file itself18:31
It's more a convention thing to be honest!18:31
← maestrojed left18:32
← jefferai left18:32
_ikke_ Do you really need that information in the file?18:33
with gith it's extremely easy to find out18:33
adhawkins I'm thinking more for people who just download the 'release' tarball.18:33
← raatiniemi left18:33
→ whyking_ joined18:34
_ikke_ Well, see that gitattributes page18:34
→ jefferai joined18:34
adhawkins But all ident does it put the hash in isn't it?18:35
← Davey left18:35
adhawkins That's the bit I'm least interested in :)18:35
_ikke_ Hmm18:35
I guess you'd have to create such a filter yourself then18:36
→ tqmz joined18:37
→ tatsuya_o joined18:37
← tatsuya_o left18:37
→ Ontolog joined18:37
→ tatsuya_o joined18:37
→ marhaban joined18:37
flavius is it possible to view all files under revision control, preferably as a tree? (as in "the tree command")18:38
Ontolog During a rebase sometimes i get mysterious conflicts for files that I'm only working on in the branch that is being rebased. How can I tell git to just accept the code in the branch that is being rebased for certain files?18:39
→ mxweas_ joined18:39
adhawkins _ikke_ Ok, thanks18:39
← TikaL left18:39
_ikke_ flavius: git ls-tree -r HEAD18:39
flavius _ikke_: thanks18:41
→ ColKurtz joined18:41
Ontolog For example, I have a branch 'foo' off of 'develop'. I run `git rebase develop` while foo is the current branch. I get a conflict for a while that I'm only working on in `foo`. How do I tell git to accept foo's version of the file?18:42
(Without manually editing the merge markers)18:42
← storrgie left18:43
→ master-lincoln joined18:43
← koo3 left18:43
→ sdeboey joined18:44
→ ThePendulum1 joined18:44
→ koo3 joined18:45
gitinfo set mode: +v18:46
ThePendulum1 What if someone committed a change and pushed it to Bitbucket the rest of the team can't agree on, can I just work from the previous version and still push it to Bb?18:46
surfdue Quick question I just merged the dev branch into master and now on my master testing if I try to pull I have many conflicts. What command can I run to retreieve the latest content and force merge using the remote repo (theirs) ignoring mine completely, but still ignoring all the files/folders in the gitignore.18:46
ThePendulum1 'the rest of the team can't agree on' should be after 'a change'18:46
← jrots left18:46
FauxFaux You can work from wherever you want and fix it later, sure. You can't push over their change without !rewriting , though.18:46
gitinfo [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)18:46
← spaceonline left18:46
← brentdanley_ left18:46
→ fisted joined18:47
_ikke_ ThePendulum1: You might better work via pull requests18:47
in the future18:47
ThePendulum1 _ikke_: That's my issue exactly. If I pull from Bitbucket, I will have his ... well, 'ruined' version?18:47
_ikke_ ThePendulum1: Does everybody commit to master?18:48
ThePendulum1 Yeah18:48
Well, not yet, though18:48
_ikke_ ThePendulum1: I could recommend githubs workflow18:48
ThePendulum1 I am just trying to solve some issues before they occur18:48
← Araxia_ left18:48
_ikke_ ThePendulum1: Everybody works on topic branches18:48
and makes pull requests to get them merged into master18:49
Only if someone else approves, it get's merged into master18:49
→ linguini joined18:49
← glcrazy left18:49
→ glcrazy joined18:49
→ Danielpk joined18:49
← Advocation left18:49
→ brentdanley_ joined18:50
→ Advocation joined18:50
ThePendulum1 _ikke_: So the other guy should just work on another branch, and I will merge it as soon as I like it?18:50
← Gentlecat left18:50
_ikke_ ThePendulum1: Yes18:50
← superdmp left18:50
ThePendulum1 _ikke_: How exactly am I supposed to do the latter? Pull from the topic branch, and then push it to the master one?18:50
surfdue FauxFaux: what do you think is the best way, I didnt intend to do what i did last night what i was trying to do is merge the two branches (same except the changes/adds/deletes), i was hoping to keep the log on master so basically i imagined it saying master -18:50
→ aleray joined18:50
surfdue commit "Pull dev into master" with all the changes18:51
→ gchristensen joined18:51
← Ontolog left18:51
FauxFaux surfdue: Not you.18:51
surfdue FauxFaux: and advice on how to properly merge the two branches?18:51
_ikke_ ThePendulum1: Bitbucker supports pull requests18:51
→ osX64 joined18:51
← brngardner left18:52
ThePendulum1 _ikke_: So the other developer shoots a pull request, and I approve, correct?18:52
_ikke_ ThePendulum1: Yes18:52
ThePendulum1 How do I prevent the other developer from pushing to the master branch?18:53
knapper_tech m1sc: yes. as it is, the list of modified files is useless18:53
surfdue ThePendulum1: you can't but dont create a hook.18:53
knapper_tech m1sc: this is bad becuase i usually read through it to remember changes18:53
surfdue you can reverse the commits and discipline them18:53
ThePendulum1: or if your using gitolite you can give access to certain branches18:53
← n1x left18:53
← adhawkins left18:54
→ drizzd joined18:54
ThePendulum1 surfdue: So I just tell the guy not to push to the master, and if he does, I can revert it?18:54
← tktiddle left18:54
_ikke_ ThePendulum1: yes18:54
ThePendulum1 Alright, thanks for the information :)18:55
← francisfish left18:55
surfdue and beat his ass18:56
ThePendulum1 So the other developer just cloned my project. What does he have to do to start working on his own branch now? He has already made some changes now, but he hasn't staged or committed anything yet18:56
← subh left18:57
→ umc joined18:57
surfdue So the question really is how do I properly merge the current changes in the dev branch into the master branch and then update the master git repository with the least amount of downtime.18:57
_ikke_ ThePendulum1: git checkout -b new_branch18:57
ThePendulum1: where new_branch is the name of his branch18:57
ThePendulum1 Okay, so that's where checkout comes in!18:57
I figured new_branch, but what exactly is -b?18:57
_ikke_ ThePendulum1: checkout is quite versatile18:57
ThePendulum1: -b creates a new branch in stead of trying to check out an existing one18:58
ThePendulum1 Ah, okay, makes sense18:58
← aristidesfl left18:58
→ aristidesfl joined18:58
_ikke_ ThePendulum1: normally you make a branch with git branch <new_branch>18:58
ThePendulum1 Does he have to do this before staging and comitting?18:58
_ikke_ ThePendulum1: Before committing, yes18:58
ThePendulum1 _ikke_: Okay, I'll remember that.18:58
_ikke_ ThePendulum1: Not before staging18:59
← williamcotton left18:59
ThePendulum1 Okay, thanks :)18:59
_ikke_ ThePendulum1: but git checkout -b is a convenient shortcut18:59
ThePendulum1 So he actually should've created a new branch before cloning the project? Or how was he supposed to get the files from bitbucket initially?18:59
_ikke_ ThePendulum1: No18:59
odin_ is there an easy way to check what a .gitignore/.gitattributes file-spec matches in my working tree ?19:00
← rzec left19:00
← thiago left19:00
ThePendulum1 _ikke_: Okay19:00
→ mister_m joined19:00
_ikke_ ThePendulum1: He clones the project, creates a new branch, and makes his changes, commits, and then pushes19:00
and then he makes a pull request19:00
ThePendulum1 _ikke_: Okay. Why should he make a pull-request after pushing?19:01
→ nightlycoder joined19:01
ThePendulum1 _ikke_: So as soon as he has created a branch using checkout, how does he properly push it to bitbucket?19:01
→ thierryp joined19:01
_ikke_ ThePendulum1: That way he communicates to you that his changes are finished and wants them merged in19:01
ThePendulum1: just git push origin branch_name19:02
ThePendulum1 _ikke_: Oh, indeed, I am sorry I read just pulling it, rather than making a pull-request. My bad, hehe!19:02
→ thiago joined19:02
ThePendulum1 _ikke_: It's pretty easy I see, thanks a lot for explaining!19:02
← martinjlowm-lptp left19:03
→ PElshoff joined19:04
→ krzkrzkrz joined19:06
← macmartine left19:06
→ Araxia_ joined19:07
← krz left19:08
krzkrzkrz → krz19:08
→ brngardner joined19:08
→ snearch joined19:09
← ThePendulum1 left19:10
← txomon left19:11
← txomon|home left19:11
← stuntmachine left19:12
→ axl___ joined19:13
← mdhopkins left19:13
→ txomon|home joined19:13
← txomon|home left19:13
→ txomon|home joined19:13
← txomon|home left19:14
← LucasTizma left19:15
← axl_ left19:16
axl___ → axl_19:16
→ subh joined19:16
← n3m left19:17
→ n3m joined19:17
→ francisfish joined19:18
→ project2501b joined19:19
← koo3 left19:20
← estevocastro left19:21
→ es joined19:22
→ SABH9VM joined19:22
→ estevocastro joined19:23
← francisfish left19:23
→ Laibsch joined19:24
→ williamcotton joined19:24
→ ceej_ joined19:25
Laibsch I have more than one remote branch locally. When I push tags to one of the remote branches, the remote repo is polluted from tags from the other remote repos. Is there a way to avoid this?19:25
← ceej_ left19:26
offby1 Laibsch: I suspect that you can be more specific. E.g., "git push remote tagname" should push just that one tag19:26
← es left19:27
Laibsch offby1: I'll try that. Any way to generally exclude certain tags?19:27
← _Vi left19:27
← EditThat left19:28
offby1 can't think of one19:28
→ Edit_ joined19:28
← ceej left19:28
→ lo0m joined19:29
→ shruggar joined19:29
→ _Vi joined19:29
canton7 one way we will have namespaced tags, and things will be nicer...19:30
← daaku left19:30
Laibsch OK, thanks. It doesn't quite seem right for git to push tags that point to a commit that is not available in that particular repo.19:30
← qt-x left19:30
Laibsch git does have the information to do the right thing on its own.19:30
→ martinjlowm-lptp joined19:31
→ kerrick joined19:33
← brngardner left19:33
← bartek left19:33
→ bartek joined19:33
← mcluskyd_ left19:34
← Araxia_ left19:35
→ Araxia_ joined19:36
← psakrii left19:36
← bemson left19:37
← lonewulf` left19:37
→ ceej joined19:38
← vdv left19:39
→ vdv joined19:39
← subh left19:39
→ lonewulf` joined19:40
← Ky0sh1r0 left19:43
← kpreid left19:44
→ krzkrzkrz joined19:44
← _Vi left19:44
← krz left19:46
krzkrzkrz → krz19:46
← glcrazy left19:46
→ ThePendulum1 joined19:46
→ TikaL joined19:46
→ |apriori| joined19:46
|apriori| hey guys19:47
does anyone know why two clients interoping with svn via git svn dont have the same SHA1 for e.g. HEAD?19:47
→ xsbeats joined19:47
_ikke_ |apriori|: Because each commit gets rewritten with the svn id in it19:48
← toothr left19:48
|apriori| _ikke_: and the svn id is? not the revision?19:48
or how should that differ across clients?19:48
→ aristidesfl1 joined19:48
_ikke_ |apriori|: It's a string written in the commit message19:49
|apriori|: git-svn-id: ...19:49
|apriori| _ikke_: yes, but that doesnt really answer my question..19:49
← martinjlowm-lptp left19:49
← falena left19:49
|apriori| that ID is the same everywhere..19:49
its pointing to the exact same repository19:50
and it includes a SHA1sum + repo url + repo rev19:50
_ikke_ Yes, but the commit date changes because the commit is rewritten19:50
|apriori| ah19:50
so two clients wanting to match the results of git svn rebase.. would actually have to do it at the exact same time :/19:50
← aristidesfl left19:50
canton7 it's generally a bad idea to have a git repo, which uses git-svn, talk to any other git repos19:51
→ whitey`` joined19:51
canton7 since things end up getting rewritten so often19:51
← mastro left19:51
|apriori| canton7: yes, because of the "issue" I mentioned19:52
_ikke_ |apriori|: because of the way SVN works19:52
|apriori| canton7: would also explain why pulling from such a repo would result in a full pull of all objects19:52
canton7 |apriori|, and other. Such as git having to rebase everything all the time to play by svn's rules19:52
*others19:52
_ikke_ |apriori|: SVN only knows lineair history, so everything would have to be rebased to create that19:53
← Dulak left19:53
← w0ng left19:53
|apriori| I know why I hate svn :P19:53
canton7 and every time you dcommit a commit, that commit has to get rewritten...19:53
→ Dulak joined19:53
→ w0ng joined19:54
→ macabre^ joined19:54
← sdeboey left19:55
→ mastro joined19:56
→ kpreid joined19:57
→ toothr joined19:58
← macabre left19:58
← Dulak left19:59
→ brngardner joined20:00
→ hashpuppy joined20:00
← w0ng left20:01
← whitey`` left20:01
← brngardner left20:01
→ Davey joined20:01
hashpuppy stupid question. let's say we develop on develop and merge to master. what happens when a user develops on master. doesn't merge the changes into develop. then another user develops on develop and merges those changes into master? we need to somehow catch this and manually merge master to develop?20:01
← skylamer` left20:02
hashpuppy or will we get a warning when trying to merge develop to master now?20:02
← umc left20:02
← marcogmonteiro left20:02
→ w0ng joined20:02
→ marcogmonteiro joined20:02
gitinfo set mode: +v20:02
→ kaen joined20:02
ThePendulum1 So another developer is working on a project, and he has his own branch. How do I pull from his branch so I can have a look?20:02
← kaen left20:03
→ kaen joined20:03
← thierryp left20:03
← vdv left20:05
← pppaul left20:05
_ikke_ ThePendulum1: He could push his branch to the shared repo20:05
ThePendulum1 That's what he did, but his branch is exactly the same for some reason20:05
lo0m hashpuppy: youll get a warning if theres a merge conflict..20:05
ThePendulum1 He did make changes, but they seem to have been disappeared20:05
hashpuppy ok, thanks :)20:06
→ tvw joined20:06
lo0m but youll have to merge master to develop nevertheless20:06
proq I'm trying to erase a commit that shouldn't have happened. when I do git reset --soft HEAD~1 and then git revert HEAD, I get the error: "fatal: Your local changes would be overwritten by revert. Please, commit your changes or stash them to proceed." but git diff shows no changes. how do I finish reverting the commit?20:06
→ pppaul joined20:06
|apriori| proq: e.g. by stashing20:06
toothr ThePendulum1, did he commit his changes?20:07
→ jrots joined20:07
ThePendulum1 I told him to do so, I guess he missed something20:07
proq |apriori|: aha, thanks.20:08
← jrots left20:08
→ Kobaz joined20:08
_ikke_ proq: Why do you do it that way?20:08
← webflo left20:08
hashpuppy i guess my question is, if you have a master and development, how do you guys keep an eye on development to make sure it doesn't get out of sync w/ master (for example, when someone commits directly to master)?20:08
|apriori| proq: its quite simple.. use git stash before trying the deletion, and git stash pop to get your stashed changes back after you're done20:08
Kobaz how would i list all the commit numbers in between two commits20:08
like, what's the log history between master and origin/master20:08
proq _ikke_: that was the first hit on google when I was looking for instructions (http://nakkaya.com/2009/09/24/git-delete-last-commit/)20:09
|apriori| hashpuppy: e.g. by enforcing fast-forward only20:09
_ikke_ proq: Have you shared them?20:09
proq _ikke_: what do you mean?20:09
lo0m when someone commits directly to master we chop his hands off :-)20:09
_ikke_ proq: Those are two alternatives20:09
git reset and git revert20:09
→ sdeboey joined20:09
proq _ikke_: oh, I see20:09
hashpuppy ok, this just happened to me. i cloned the repository and it defaults to master. i didn't think about it and wound up committing there20:10
_ikke_ Have you pushed that commit?20:10
← pastime left20:10
→ alberto56 joined20:10
hashpuppy me? yeah. not really worried about it cause we have a small team20:10
_ikke_ no proq20:10
← mauror left20:10
proq _ikke_: no. I'm only experimenting with pulling commits atm20:10
→ jrots joined20:10
→ khmarbaise joined20:10
← __class__ left20:11
→ m15k joined20:11
_ikke_ proq: git reset --hard (will remove any uncommited changes)20:11
proq where by pulling I mean erasing :)20:11
|apriori| btw. , is there any sane way one could setup a git interop server for a svn repository?20:12
or would that make things even much worse?20:12
→ __class__ joined20:12
_ikke_ |apriori|: You just cooperate through svn20:12
m15k I try to create a signed tag, but I got a gpg: signing failed: secret key not available20:12
← marthinal left20:12
proq _ikke_: thanks, I didn't know such a useful feature existed20:12
→ LucasTizma joined20:12
proq and potentially dangerous! hehe20:13
_ikke_ yes20:13
→ j416 joined20:13
→ pastime joined20:13
_ikke_ proq: normally you would do git reset --hard HEAD~120:13
→ mischat joined20:13
← pppaul left20:13
← fahi left20:13
→ fahi joined20:14
→ pppaul joined20:14
← mischat left20:14
→ mcluskydodallas joined20:14
|apriori| _ikke_: assume you got several devs working on a project who actually know git.. and would rather use that (e.g. use branches, push & pull from each other) and have e.g. autopush to svn from svn...20:15
→ lusers joined20:15
|apriori| _ikke_: is that any sane or would one just enter another hell with that?20:15
→ glcrazy joined20:16
|apriori| (and say its not an option to port the repo over to git)..20:16
ThePendulum1 Is it possible to remove a remote branch, when the local branch no longer exists?20:17
|apriori| ThePendulum1: didnt try it.. but create a local remote tracking branch.. and delete that with the remote one20:18
← pppaul left20:18
→ pppaul joined20:19
→ jargon- joined20:19
← enmand left20:20
← j416 left20:20
_ikke_ |apriori|: Any history that is not dcommitted to svn yet can just be used normally20:20
← lusers left20:20
← __class__ left20:20
→ __class__ joined20:21
← jsoa left20:21
← Davey left20:22
→ robinds joined20:22
→ btanaka joined20:22
← pastime left20:23
→ adhawkins joined20:23
← coyotebush left20:23
← Ryback_ left20:24
← arietis left20:24
← ThePendulum1 left20:24
← SkAzZ left20:24
← pppaul left20:24
→ EricInBNE joined20:24
→ j416 joined20:24
Laibsch Does Junio hang around here?20:24
← axl_ left20:24
_ikke_ Laibsch: afaik, not20:24
Laibsch Ok, thanks20:25
← j416 left20:25
→ pppaul joined20:25
→ axl_ joined20:25
→ j416 joined20:25
→ ipalaus_ joined20:26
→ Davey joined20:26
→ pastime joined20:26
← TikaL left20:26
→ ThePendulum1 joined20:27
← ThePendulum1 left20:28
← ziggles left20:28
← sq-one left20:28
→ diegoviola joined20:28
→ Adys_ joined20:28
← juvenal left20:28
→ mdhopkins joined20:29
← lonewulf` left20:29
← Adys left20:29
← ericmuyser left20:29
→ Tasser joined20:30
← glcrazy left20:30
← pppaul left20:30
Tasser http://sprunge.us/OGhH ups?20:30
→ pppaul joined20:30
← whyking_ left20:31
← mdhopkins left20:31
Laibsch Can somebody please help me understand the case Junio gives in http://thread.gmane.org/gmane.comp.version-control.git/195360/focus=195361 ?20:31
← pastime left20:31
→ _class_ joined20:31
→ maestrojed joined20:31
Laibsch sounds like linear history to me and I can't come to the conclusion he gave20:31
← Soliton left20:32
← __class__ left20:32
Laibsch I think I understand better now, but to me it sounds like the information is there in git but just not in "git push"20:33
→ Soliton joined20:33
milki does git push --tags only push tags and not relevant objects as well?20:34
→ mastroWork joined20:34
→ marcogmonteiro_ joined20:34
Kobaz hmm20:34
so is there a way to get the log entries between revs or is diffing the only way20:34
_class_ → __class__20:35
→ jfelchner joined20:35
milki should work, see refspec20:35
← EricInBNE left20:35
→ cagedwisdom joined20:35
← pppaul left20:35
charon Kobaz: man gitrevisions, specifying ranges20:36
gitinfo Kobaz: the gitrevisions manpage is available at http://jk.gs/gitrevisions.html20:36
Kobaz k20:36
← robinds left20:36
→ pppaul joined20:36
Vinnie_win after I issue git reset --hard HEAD^ how do I push the rollback to the remote? git push says "Everything up-to-date"20:36
Kobaz oh cool20:37
nice20:37
← marcogmonteiro left20:38
marcogmonteiro_ → marcogmonteiro20:38
→ coyotebush joined20:38
→ robbyoconnor joined20:39
← iamjarvo left20:40
→ QiBala joined20:40
→ aristidesfl joined20:42
milki Vinnie_win: if its not a tracked branch, you need to specify it manually20:42
→ dv310p3r joined20:42
→ _Vi joined20:45
← aristidesfl1 left20:45
→ amcsi joined20:46
← jrots left20:46
← Tianon left20:46
proq how do I wipe out an entire branch with several months worth of fixes from another branch, without prompting or merge conflicts?20:47
← hagabaka left20:47
→ BW^- joined20:47
← mister_m left20:48
_ikke_ wipe out?20:48
← Sigma left20:48
→ boondog joined20:48
← __class__ left20:48
→ brngardner joined20:49
BW^- in a git repo, say that i have a bunch of big file commits that i want to get rid off completely, as not to take up unnecessary space neitehr on my harddrive, nor on github's, nor on anyone else's . Does git have features for listing all commits in total in it incl filenames, and, Does git have the feature to permanently erase a commit, and to propagate this permanent erasure also on pull/push, so that what I asked for now is fully accomplished?20:49
it's an app that generates C files needed for compilation, they're super big, but you only need the latest version of them, and the version from major app versions. all intermediary take hudnreds of megabytes, so iw ant to erase them completely.20:49
milki git log filename20:49
and you can use filter-branch to edit those commits20:50
BW^- ..from everyone's disks, it'd be completely unrealistic to have them around20:50
milki: how?20:50
→ _class_ joined20:50
milki man filter-branch20:50
BW^- milki: i'm asking this on behalf of someone else, so i need sth clear to say20:50
proq _ikke_: yes. wipe out as in replace one branch with all the changes from another20:50
milki BW^-: theres an example there20:51
BW^- aha20:51
ok20:51
milki you can also refer to github's remove sensitive files faq20:51
BW^- aha cool20:51
_ikke_ proq git checkout other; git branch -f one20:51
← PElshoff left20:51
BW^- milki: wait, git filter-branch you mean right?20:52
not the os app filter-branch?20:52
→ jrots joined20:52
BW^- _ikke_: that was to me?20:52
milki no clue what os app is20:52
BW^- OS as in operating system20:52
_ikke_ BW^-: No, to proq20:52
BW^- k20:52
← robbyoconnor left20:53
← erichynds left20:53
← pandeiro left20:53
← boondog left20:54
milki o20:54
right20:54
its git20:54
← _class_ left20:54
→ mister_m joined20:54
← pppaul left20:56
← ipalaus_ left20:56
→ aspotashev joined20:56
→ pppaul joined20:56
← p0t4t0 left20:57
← chrisshattuck left20:57
→ ThePendulum1 joined20:57
gitinfo set mode: +v20:58
ThePendulum1 I pulled from someone's branch, but now I can't pull from the master because it's "Already up-to-date" because the guy's branch is ahead of the master20:58
How do I pull from the master anyway?20:58
_ikke_ ThePendulum1: !pull20:58
gitinfo ThePendulum1: pull=fetch+merge (or with flags/config also fetch+rebase). It is thus *not* the opposite of push in any sense. A good article that explains the difference between fetch and pull: http://longair.net/blog/2009/04/16/git-fetch-and-merge/20:58
→ enmand joined20:58
→ lonewulf` joined20:59
← macabre^ left20:59
→ ipalaus_ joined20:59
→ macabre joined20:59
Laibsch I want to merge upstream branch into master, but there are conflicts. I want uptream to have the upper hand. So I did "git merge -s recursive -Xtheirs upstream" as detailed in "git help merge", but the conflicts still prevail. What am I doing wrong?21:00
← cannonball left21:01
← fahi left21:01
ThePendulum1 _ikke_: What should I be looking at to avoid the 'Already up-to-date'?21:01
→ wbednarski joined21:01
_ikke_ ThePendulum1: It means you already have all history locally21:01
ThePendulum1: But it may not be merged in21:01
ThePendulum1 I accidently pulled his topic branch into my master :|21:01
← brentdanley_ left21:02
→ felgru_ joined21:03
→ ObseLeTe joined21:03
ObseLeTe Hi21:04
gitinfo ObseLeTe: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.21:04
ObseLeTe Can i go to back before a git pull?21:04
← airborn left21:04
milki if you remember which commit it was yes21:04
reflog might help21:04
← pppaul left21:04
ObseLeTe i want back to before the first git pull21:05
→ pppaul joined21:05
→ PElshoff joined21:05
← felgru left21:06
→ wicchan_ joined21:06
← whitman left21:07
← snearch left21:07
→ snearch joined21:07
← bartek left21:07
→ Superlokkus_ joined21:07
→ mikepack_ joined21:08
→ PerlPilot joined21:08
→ ybit2 joined21:08
→ el2ro_ joined21:08
← sdeboey left21:08
← wicchan left21:08
← btanaka left21:08
← NimeshNeema left21:08
← el2ro left21:08
← cbreak-work left21:08
← andres left21:08
← merkosh left21:08
← PerlJam left21:08
← Superlokkus left21:08
← n2diy left21:08
← narthollis left21:08
← luka12345 left21:08
← dhoss left21:08
← mikepack left21:08
← ybit left21:08
Superlokkus_ → Superlokkus21:08
→ andres_f joined21:08
→ merkosh joined21:08
→ cbreak-work joined21:08
→ n2diy joined21:09
← marhaban left21:09
→ narthollis joined21:09
→ luka12345 joined21:09
→ NimeshNeema joined21:10
→ abetaha joined21:10
← alberto56 left21:10
→ dhoss joined21:10
← abetaha left21:10
← mantic0 left21:11
proq I'm trying to overwrite all files in one branch (master) with another branch (nu). when I use "git checkout master; git branch -f nu" it doesn't change anything in master. I tried the other way around and saw no results either.21:11
lo0m reset before the pull like "git reset --hard master@{12:15}" should work, ObseLeTe21:12
← rendar left21:12
→ MestreLion joined21:12
← tatsuya_o left21:12
|apriori| lo0m: whats that syntax? master@{12:15} ?21:13
ist that a commit number range?21:13
*is21:13
→ tatsuya_o joined21:13
lo0m |apriori|: it's time.. even master@{"2 hours ago"} should work21:13
← ThePendulum1 left21:13
→ alberto56 joined21:14
|apriori| lo0m: oh, ok21:14
← PerlPilot left21:14
lo0m man git-rev-parse21:14
gitinfo the git-rev-parse manpage is available at http://jk.gs/git-rev-parse.html21:14
MestreLion Ok, tricky one (for me): after some rebases/add -p/etc, I noticed I lost a block of code. I was able to find it in a commit in the reflog. Is there a way to show all diffs from all commits, including the reflog, so I can look if that code ended up elsewhere?21:14
← jrots left21:14
→ PerlJam joined21:14
→ jrots joined21:15
→ acl_ joined21:15
→ sdeboey joined21:15
→ jlebrech joined21:15
MestreLion Something like git magic --all-descendants-of <commit> --include-reflog | grep "my list code" ?21:15
lost*21:16
→ brentdanley_ joined21:16
→ ThePendulum1 joined21:17
→ __class__ joined21:17
gitinfo set mode: +v21:17
← kpreid left21:17
ThePendulum1 Is it possible to ignore an entire folder at once?21:17
MestreLion ThePendulum1: add dir/ to .gitignore21:17
← PElshoff left21:17
→ tewecske joined21:18
← cgrage left21:18
← tewecske left21:18
MestreLion assuming you dir/ and all of its files are currently untracked21:18
ThePendulum1 Meh. It indeed ignores the folder, but it shows itself as to be added21:18
→ martinjlowm-lptp joined21:18
← pppaul left21:18
← dv310p3r left21:18
MestreLion ThePendulum1: are they part of any commit, or just the staging?21:18
← Superlokkus left21:19
proq nevermind, I needed to git checkout master after my last step :P21:19
ThePendulum1 It's a subdirectory that shouldn't be part of git at all21:19
It's not staged yet21:19
→ pppaul joined21:19
ThePendulum1 (and shouldn't be)21:19
MestreLion ThePendulum1: was it ever part of any commit?21:19
ThePendulum1 MestreLion: It has never been part of anything regarding git, I just added the folder21:20
The folder seems to be ignored now21:20
But git shows .gitignore itself as an untracked file21:20
Should that one be added?21:20
← kerrick left21:20
MestreLion so the .gitignore trick should work... currently they are untracked files, correct? git add . or git show will not list them21:20
ThePendulum1 Correct21:21
But it does list the .gitignore file itself, so I'm wondering if I have to add that to git?21:21
eph3meral ThePendulum1, if you haven't set .gitignore to be tracked, then yes it will be listed as untracked, and you should track it by running: git add .gitignore21:21
MestreLion yes, I reccomend you to add the .gitinore to git21:21
ThePendulum1 eph3meral: Okay, I just wasn't sure I had to track it, hehe21:21
MestreLion you don't *have* to for it to work...21:21
→ smw joined21:21
ThePendulum1 I figured21:21
eph3meral ThePendulum1, strictly speaking, no you don't have to, but if you don't want cruft in your project, you should probably track it so other users don't commit crufty files21:22
ThePendulum1 ^21:22
MestreLion but since this .gitignore will be useful for anyone using the project, it makes sense to make it a part of the repo21:22
→ dv310p3r joined21:23
← mxweas_ left21:23
MestreLion ThePendulum1: classic examples are ignoring po/*.mo files, build/ , etc...21:23
so, guys... any help on my "magic command" ?21:24
→ s0ber_ joined21:24
MestreLion git magic --all-descendants-of <commit> --include-reflog | grep "my lost code" ?21:24
if need to find out if a piece of code I've found in the reflog ended up in any other commit21:25
← arvind_khadri left21:25
← Maccer left21:25
→ p3rror joined21:25
→ notola joined21:25
← notola left21:25
← s0ber left21:26
s0ber_ → s0ber21:26
→ LongBeach joined21:26
← snearch left21:26
← legnaleurc left21:26
← brentdanley_ left21:26
→ kpreid joined21:27
lo0m MestreLion: i guess you can use bisect to pinpoint the commit where it went missing21:28
← kpreid left21:28
← m15k left21:28
→ kpreid joined21:29
→ Tianon joined21:29
← Tianon left21:29
→ Tianon joined21:29
← __class__ left21:29
MestreLion lo0m: I don't need to know where it went missing.. I've already found it in the reflog. I did some "extreme makeover" with rebases, so I may have screwep up... no big deal since I've found it. I just want to know If it was erroneusly added to another commit/branch21:29
← pppaul left21:29
← kpreid left21:29
→ pastime joined21:30
→ kpreid joined21:30
← p3rror left21:30
← kpreid left21:31
→ pppaul joined21:31
→ kpreid joined21:31
lo0m MestreLion: oh, ok, my bad ;-)21:31
MestreLion lo0m: So I would like to "scan" all commits since a given ancestor and grep for my code. If It's not anywhere, good, It was deleted and I can cherry-pick it from reflog21:32
← spiette left21:32
← kpreid left21:32
← _iron left21:32
← saml left21:32
→ kpreid joined21:32
MestreLion I want want to avoid the embarrassment of cherry-picking the lost code from reflog and later find out that it was already present in a completely unrelated branch21:33
← shruggar left21:33
← dstorrs left21:33
← j416 left21:33
← Element9 left21:34
lo0m MestreLion: would love to know that, cant help...21:34
→ aalex joined21:34
MestreLion do you know how to show a diff from all commits since <commit> ?21:34
that would be a start21:34
→ Pupeno_W_ joined21:35
← pppaul left21:35
→ p3rror joined21:36
→ pppaul joined21:37
← Pupeno_W left21:38
Pupeno_W_ → Pupeno_W21:38
← ThePendulum1 left21:39
← sdeboey left21:39
← cbreak-work left21:40
← BW^- left21:40
→ dstorrs joined21:40
lo0m cant find anything useful21:40
MestreLion i'm trying git log --since=v0.30c --all21:41
→ __class__ joined21:41
→ marhaban joined21:41
← ipalaus_ left21:41
← pppaul left21:41
← yoh left21:41
→ pppaul joined21:43
← beatak left21:43
→ Maccer joined21:44
→ Bartzy joined21:45
lo0m you can add -S "taht code" .. or not?21:47
← pastime left21:49
MestreLion yessssss21:49
:D21:49
git log --all -S"mycode"21:49
andres_f → andres21:49
lo0m i love git :-)21:49
← rudi_s left21:50
← __class__ left21:50
← jrots left21:51
← timoxley left21:51
← pppaul left21:51
→ pppaul joined21:51
→ __class__ joined21:52
← macabre left21:52
← ObseLeTe left21:52
← marhaban left21:52
→ barr5790 joined21:53
← dv310p3r left21:53
← cime left21:54
→ williamcotton_ joined21:54
← knapper_tech left21:54
MestreLion hummm, not quite right21:54
← Dave^| left21:54
← pppaul left21:55
MestreLion --all only scans from branch tips21:55
← crash\ left21:55
→ ghickman joined21:55
→ pppaul joined21:55
MestreLion and -S only works for commits that *introduced* (or removed) that code21:55
→ crash\ joined21:55
MestreLion so if my code was inserted in a non-tip commit, it won't show21:55
← __class__ left21:56
FauxFaux MestreLion: "git diff sometag" will show you the diff with the tag. -S with --all will show all commits reachable from branches. Try --tags if your code is only in a tag?21:56
→ tonesfrommars joined21:56
← williamcotton left21:57
williamcotton_ → williamcotton21:57
← brentdanley left21:57
→ __class__ joined21:58
MestreLion FauxFaux: actually, I'm using -S because I don't know where the code is, IF it is somewhere21:59
→ Nightfire034 joined22:00
← Davey left22:00
FauxFaux Well, if you can't find it with -S with --all or --tags, you probably don't have it.22:00
← crash\ left22:00
MestreLion FauxFaux: I did a quick test with git log --all -S"my code", and it only found commits that were tips of a branch22:00
→ Francisc joined22:00
→ _class_ joined22:01
→ iamjarvo joined22:01
Francisc 1.7.10 is latest version, right?22:01
→ crash\ joined22:02
FauxFaux Francisc: Topic.22:02
Francisc Thank you.22:02
FauxFaux MestreLion: "git log --oneline --all -S. | wc -l" prints 14055 here. I'm reasonably sure I don't have 14055 branches.22:02
← t4nkd left22:02
← _class_ left22:02
← LongBeach left22:02
← __class__ left22:02
← kermit left22:02
← martinjlowm-lptp left22:03
← revagomes left22:03
→ kermit joined22:03
MestreLion git log --oneline --all -S"XDG_DATA_HOME" shows only 2 commits (both are tips)22:03
it fails to find a 3rd commit that I can see here22:04
→ __class__ joined22:04
← juanfer left22:04
← aristidesfl left22:04
→ mischat joined22:04
← lonewulf` left22:05
← duckxx left22:05
← lacrymology left22:05
← marcogmonteiro left22:06
← __class__ left22:06
→ aristidesfl joined22:06
→ brentdanley joined22:06
← Francisc left22:07
→ Nikoli_ joined22:07
→ eeanm joined22:07
← acl_ left22:07
→ mxweas_ joined22:07
← ISF left22:07
MestreLion FauxFaux: actually, no... out of (at least) 3 commits, it is finding 2, 1 is a tip, the other is not, and he missed one that is a tip22:08
← Nikoli left22:08
← DrNick left22:08
← aleray left22:08
tizzo → tizzo-afk22:09
→ _class_ joined22:09
← eean left22:09
lo0m and with -g option?22:10
← chrisgeorge left22:11
MestreLion oh... ignore whitespaces22:11
i guess that's why22:11
FauxFaux Hmm?22:11
_class_ → __class__22:13
MestreLion the commit log is not finding22:13
if you grep that commit, the particular like introduces only whitespace22:13
line*22:14
→ thatryan joined22:14
MestreLion so maybe that's it22:14
→ ericmuyser joined22:14
← lo0m left22:14
← Carraway left22:14
MestreLion care to test with my repo FauxFaux ? It's only 4MB22:14
then I could show how git log is missing at least one commit22:15
and this is really puzzling me22:15
git://github.com/MestreLion/singularity.git22:16
← fisted left22:16
← cccaldas_ left22:16
MestreLion Or I could paste a terminal session22:16
← ehsan left22:17
← mxweas_ left22:17
→ lonewulf` joined22:18
→ Eridius joined22:18
← barr5790 left22:18
variable I frequently do "git branch --contains <foo>" and then git checkout the one and only branch returned22:18
is there a way to do that in one step?22:18
MestreLion FauxFaux: http://sprunge.us/BZNR22:19
FauxFaux: first 2 lines are git log --oneline --all -S"XDG_DATA"22:20
→ williamcotton_ joined22:20
→ _class_ joined22:20
MestreLion FauxFaux: and the rest is git show f11d302b0e8122:20
→ sdeboey joined22:20
← iamjarvo left22:20
MestreLion you can clearly see that this 3rd commit has "XDG_DATA" code22:20
but still git log didn't list it... any ideas why?22:21
← sj` left22:21
→ mxweas_ joined22:21
→ DrNick joined22:21
← d0k left22:22
johnkpaul → johnkpaul-afk22:22
← williamcotton left22:22
williamcotton_ → williamcotton22:22
← __class__ left22:23
→ lo0m joined22:23
_class_ → __class__22:23
MestreLion Humm, I know why :)22:24
FauxFaux git rev-list --all | grep f11d302b0e8122:24
MestreLion looks like pickaxe is a bit picky about "introduces chances"... if file has the string both before and after the commit, it does not show it22:25
even if git diff mentions it22:25
→ jonash joined22:25
← TeckniX left22:26
→ localhost joined22:26
jonash git ls-tree <commit> shows type 'commit' for some <commit>. What's the reason for having references to a commit instead of a blob/tree?22:26
→ fisted joined22:27
→ srcerer joined22:28
jast jonash: submodules22:28
← brentdanley left22:28
← zeom left22:28
← FernandoBasso left22:30
→ spatulasnout joined22:30
jonash jast: could you elaborate a bit more on that? what does it mean if a tree entry references a commit?22:30
← lo0m left22:30
← dbgster left22:30
→ SkAzZ joined22:30
jast it means that there's a submodule in that place, and that that commit uses that commit of the submodule in that place22:30
← tqmz left22:31
jonash that makes sense22:31
what if I didn't use submodules in that particular repo?22:31
← Laibsch left22:32
→ Dulak joined22:32
← nightlycoder left22:32
→ ceej_ joined22:32
jast then something is funky22:33
→ erichynds joined22:33
← lyddonb left22:33
← stkrzysiak left22:33
jonash so I'm walking through a repo's history using dulwich (a Python binding). From dulwich's control flow I can say that that particular tree is a 'legacy' object22:34
← ceej left22:34
ceej_ → ceej22:34
jonash I don't know how many 'legacy' there though22:34
and if that is related somehow?22:34
→ stkrzysiak joined22:35
← Xizor left22:35
← Vinnie_win left22:35
← martinjlowm left22:37
← frnkkpp left22:37
→ Vinnie_win joined22:38
→ brentdanley joined22:38
→ ptw joined22:39
← MestreLion left22:40
← sdeboey left22:40
← mxweas_ left22:40
← mikepack_ left22:41
← SkAzZ left22:41
← alester left22:41
→ mikepack joined22:41
→ mxweas_ joined22:41
→ sdeboey joined22:41
← Spockz left22:42
→ duckxx joined22:43
→ subrosa joined22:44
← ghickman left22:45
← dstorrs left22:46
← mister_m left22:46
← ology left22:48
→ yoh joined22:48
← fayimora left22:48
→ macrover joined22:50
jonash jast: still here?22:50
→ imaginationac joined22:50
→ SkAzZ joined22:51
jast jonash: well, I can't tell you what happened because the only way a commit can legitimately be referenced from a tree is if it's for a submodule22:51
jonash jast: okay, thanks. I'll go and double-check I really didn't use any submodules for that repo.22:52
thanks a lot!22:52
← jonash left22:53
← eijk left22:53
← imaginationac left22:54
→ knapper_tech joined22:54
← estevocastro left22:54
→ ben225 joined22:55
← dr_win left22:55
← Russell^^ left22:55
→ blackmatrix_ny joined22:55
← crash\ left22:55
← ben_alman left22:56
→ crash\ joined22:56
→ chrisgeorge joined22:57
← ToBeFree left22:58
← Nightfire034 left22:58
→ zeom joined22:58
← felgru_ left22:58
→ guns joined23:01
→ timoxley joined23:03
← cakehero left23:03
→ ben_alman joined23:05
→ Spockz joined23:05
← mxweas_ left23:07
← nadavoid left23:08
→ whyking_ joined23:08
→ welly joined23:09
→ acl_ joined23:09
← berserkr left23:11
← toothr left23:12
→ tewecske joined23:12
→ berserkr joined23:12
← tewecske left23:12
← joaoh82 left23:13
← pantsman left23:13
→ pastime joined23:13
→ mister_m joined23:14
→ dstorrs joined23:14
→ joaoh82 joined23:15
→ toothr joined23:15
← kpreid left23:15
→ rgwu joined23:16
← mtkd_ left23:16
← gchristensen left23:17
← erichynds left23:17
← d2dchat left23:18
← enmand left23:19
← drizzd left23:19
← fission6 left23:20
← giallu left23:20
→ devth joined23:20
→ fission6 joined23:20
→ kpreid joined23:20
variable → const23:21
← acl_ left23:22
→ enmand joined23:23
→ iamjarvo joined23:23
→ cccaldas joined23:24
← enmand left23:24
← |apriori| left23:25
→ thedangler joined23:25
← fission6 left23:25
thedangler Anyone using sourcetree? if so how do you create a local branch that is the same as the remote one?23:26
← dpino left23:26
→ Nightfire034 joined23:27
thedangler got it. nvm23:27
← alberto56 left23:34
← TomyLobo left23:35
→ fayimora joined23:36
← michaelmartinez left23:36
← fayimora left23:36
← williamcotton left23:37
→ mailson joined23:38
← Chillance left23:43
← stkrzysiak left23:43
← pfFredd left23:44
→ rchavik joined23:44
← iocor left23:45
← duckxx left23:46
← sdeboey left23:47
← eph3meral left23:48
→ Per4u3e joined23:50
← mritz left23:50
→ madewokherd joined23:51
→ fission6 joined23:52
← madewokherd left23:56
→ madewokherd joined23:56
← madewokherd left23:58
← maestrojed left23:58
→ ISF joined23:59
→ madewokherd joined23:59

Logs Search ←Prev date Next date→ Channels Documentation