IRCloggy #git 2012-02-09

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.

2012-02-09

← mandric left00:00
→ satyanash joined00:01
→ twisted`_ joined00:02
← twisted`_ left00:02
→ boombatower joined00:03
samuelj is that good?00:03
← airborn left00:03
← crisp left00:03
← sattu94 left00:04
→ Turicas joined00:04
SethRobertson Yes, but it exhausts my knowledge.00:04
← twisted` left00:04
→ rchavik joined00:04
SethRobertson You could check in .git/modules and .git/config for problems. I guess the former should be empty and the latter should not mention submodules00:05
← chrisgeorge left00:05
→ zulax joined00:06
← eijk__ left00:06
← Rogi left00:07
→ huntca joined00:07
← Yuuhi left00:07
→ Rogi joined00:07
samuelj thanks for the help!00:07
→ thebaddies joined00:08
thebaddies hi00:08
← rtconner_ left00:08
thebaddies why what do I have to type to rebase this on master? http://codepad.org/Z80ScohC00:09
SethRobertson Well, I'm going to wildly guess that you need to commit or stash your changes.00:09
thebaddies I did that.00:09
← jceb left00:09
SethRobertson Yet somehow you still have changes which have not been committed.00:09
frogonwheels thebaddies: no, you have some deleted files that you haven't comitted00:10
thebaddies why would I commit deleted files ?00:10
how do you commit a deleted file ?00:10
SethRobertson Because you did not delete them otherwise00:10
frogonwheels thebaddies: you did a git rm <filename> ?00:10
← Russell^^ left00:10
frogonwheels you need to commit the rm, if that's what you did.00:11
thebaddies yes and it says: fatal: pathspec 'wireless-ap/v1.1.bin' did not match any files00:11
frogonwheels thebaddies: no, I mean presumably you already did that?00:11
SethRobertson Just commit00:12
← iamjarvo left00:12
thebaddies thank you Seth00:12
SethRobertson !thanks00:13
gitinfo Feeling thankful? Type "ExampleUser++", and ExampleUser will score karma points at http://carmivore.com (our preferred way to objectify self-worth). There's really no point to thanking me – instead, why not thank the person who made me help you?00:13
milki i think seth wants karam00:13
SethRobertson I have the biggest ego around. cmn (I think) certified it00:13
→ glennpratt joined00:13
milki lol00:13
thebaddies SethRobertson++00:14
I'll write a karmbot for seth00:14
← nOStahl left00:15
→ kerrick joined00:15
→ Anubisss_ joined00:15
→ Yuuhi joined00:16
← spaceonline left00:17
→ p3rror joined00:18
← Anubisss left00:19
← adammokan left00:19
→ Gitzilla joined00:20
← aspotashev left00:22
→ jaisoares joined00:23
← jaisoares left00:24
→ rtconner_ joined00:24
← rtconner_ left00:25
Wombert if I have files that, from their original source, contain CRLFs00:26
is it still a good idea to convert them to LFs?00:26
(it's an existing repos converted from SVN, and I'm about to run git filter-branch with dos2unix on it00:26
SethRobertson It can make blame more complicated and if someone else is still using CRLF with them you should not.00:26
On the other hand, I would and do.00:26
Wombert the repos hasn't been pushed yet or anything00:26
SethRobertson Well, filter-branch will avoid the blame problem00:27
Wombert so the idea is to fix it before that happens with filter-branch00:27
yes00:27
defo don't want to run dos2unix and then commit everything :)00:27
SethRobertson As long as you have killed the DOS people, that is fine. Make sure to run the filter-branch in a tmpfs mount00:27
Wombert I'll trust my SSD00:27
:p00:27
SethRobertson tmpfs is still faster00:28
← totzeit left00:28
Wombert so even if the original files (vendor stuff I bundle from an external source) contain CRLFs, you'd advise me to do the conversion, right?00:28
MacGyverNL I usually do.00:28
SethRobertson If you might need to import more recent version of the vendor stuff, perhaps not. You would have to stay awake when doing so at least.00:29
MacGyverNL And I also instruct anyone I pull from to use LF-line endings as well.00:29
imMute SethRobertson: if there were still DOS people, couldn't they just use autocrlf or something?00:29
Wombert so, core.autocrlf input is a reasonable setting, and core.autocrlf input or true for Windows folks?00:29
MacGyverNL Well, "instruct" is a big word.00:29
"Suggest".00:29
Forcefully.00:29
SethRobertson Many people have come here with EOL problems. Whether they have poor practice I could not say00:30
MacGyverNL And on my windows machines I always put my core.autocrlf on input.00:30
My editors are allowed to mangle them all they want.00:30
(Not all editors can be set to use only LF endings, you see.00:30
)00:30
But all files going into the repo are unmangled.00:31
And that's why I prefer other people to do the same.00:31
Wombert that's what "input" does, right?00:31
and when you pull again it's LF00:31
MacGyverNL Yes - the files come out "as-is", they go in with CRLF->LF conversion.00:31
Wombert while with "true", it's CRLF out, and conversion to LF when committing00:32
MacGyverNL So that basically means that I also automatically scrub the repo from CRLF endings that strayed in there every commit.00:32
Wombert yea00:32
Wombert ponders00:32
Wombert so I'll do the filter-branch stuff00:32
MacGyverNL (Honestly, can't we all grow up and standardize on LF by now? MS? *nix? Mac?)00:32
Wombert well, CRLF is *actually* more logical00:32
imMute Wombert: howso?00:33
Wombert at least if you are a typewriter :p00:33
← huntca left00:33
imMute Wombert: well, derp :)00:33
Wombert imMute: go to next line, return to beginning, like a typewriter would :p00:33
← Xizor left00:33
MacGyverNL Okay, maybe not the discussion to have here.00:33
Wombert indeed not00:33
:p00:33
→ tty234_ joined00:34
Wombert so I'll do dos2unix with filter-branch, and in the future I just high five anyone who commits CRLFs00:34
with a chair00:34
in the face00:34
:p00:34
forrrrrcefully00:34
→ mxweas joined00:34
SethRobertson Write a receive hook to reject it00:34
Wombert well it'll be on GH00:34
:<00:34
← Anubisss_ left00:35
→ Anubisss joined00:35
Wombert also, why does Git not like trailing whitespace?00:35
← Anubisss left00:35
Wombert like indented newlines, my editor always fills up the blank with spaces/tabs00:35
but that seems bad practice somehow, at least git add -p complains and stuff00:36
imMute Wombert: I like that idea.00:36
Wombert imMute: which one :>00:36
← berserkr left00:36
← Destos left00:37
MacGyverNL Wombert: I personally don't use github, but can't you still use receive-hooks there?00:37
← anjen left00:37
← kenperkins left00:37
← Kicer86 left00:37
Wombert nah just post-receive hooks, MacGyverNL00:37
MacGyverNL I now envision something like a post-receive hook checking for CRLF endings and resetting HEAD to previous if it contains them ^_^00:38
MacGyverNL hides00:38
← tty234_ left00:38
Wombert harr00:39
:>00:39
← oc80z left00:40
← guampa left00:40
samuelj SethRobertson: Fixed! For future ref: in .git/config, in the submodules section, url was set but there were no options for fetch, so it just wasn't doing anything00:41
sorted now!00:41
→ oc80z joined00:41
→ cooloney joined00:42
→ cakehero joined00:43
→ aristidesfl joined00:43
← tatsuya_o left00:43
→ Chib joined00:43
→ tatsuya_o joined00:44
← glennpratt left00:45
→ s_segawa joined00:46
← samuelj left00:47
→ fr0stbyte joined00:48
← kerrick left00:48
→ sattu94__ joined00:49
← griotspeak left00:50
→ Kicer86 joined00:51
← jfelchner left00:51
← satyanash left00:52
→ ihodes joined00:53
← shysoft left00:54
→ guampa joined00:54
→ xiangfu joined00:55
← linusoleander left00:56
← AaronMT left00:57
← davidrussell left00:58
→ griotspeak joined00:58
→ davidrussell joined00:58
← iocor left00:59
← mabrand left00:59
→ vikarti joined01:00
← stuntmachine left01:00
← pandeiro left01:01
← macmartine left01:01
→ sythe_ joined01:02
← fr0stbyte left01:02
← beatak left01:03
→ linusoleander joined01:05
← linusoleander left01:06
← rbrooks left01:06
→ iamjarvo joined01:06
→ Heimidal joined01:07
← andres left01:08
→ forsto joined01:08
← alberto56 left01:10
wald0 what means -- in something like: git checkout -- * ?01:11
→ weiyang joined01:11
← Demego left01:13
Bearcat http://sitaramc.github.com/gitolite/sskm.html playing with the commands on this page..where is the sskm command? I can't find it in any package?01:14
→ jamescarr joined01:15
← fly91 left01:15
→ fly91 joined01:15
jamescarr is there a way to completely remove a file from history?01:18
someone committed a ton of quicktime videos by mistake01:19
SethRobertson git-filter-branch, but this is !rewriting_public_history01:19
gitinfo git-filter-branch: 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)01:19
jamescarr like, 3009 commits ago01:19
hmmm01:19
SethRobertson, thanks. It sucks because it takes eons for new people cloning the repo01:19
SethRobertson !thanks01:19
gitinfo Feeling thankful? Type "ExampleUser++", and ExampleUser will score karma points at http://carmivore.com (our preferred way to objectify self-worth). There's really no point to thanking me – instead, why not thank the person who made me help you?01:19
← mxweas left01:19
← diegoviola left01:21
← _obiJuan left01:22
← davidrussell left01:23
→ andres joined01:23
← oc80z left01:24
→ duckxx joined01:25
→ oc80z joined01:26
milki sethego++01:26
← ehsan left01:26
← oriba left01:26
→ mxweas joined01:26
← kpreid left01:27
← vandal- left01:28
← ScottE left01:28
→ macmartine joined01:29
← ajt left01:29
→ stringoO joined01:29
→ ajt joined01:30
Bearcat SethRobertson++01:30
hmm01:30
"SethRobertson++"01:30
double hmm01:30
SethRobertson ?01:31
Bearcat i was trying to ++ you as listed above01:31
bremner ++++01:31
SethRobertson And carmivore is not accepting? Meh01:31
Bearcat do i need to direct that at the gitinfo?01:32
SethRobertson now it is up, by two01:32
gitinfo is not involved01:32
Bearcat oh, i was expecting some sort of responce like "karama added" or something.01:32
SethRobertson silent but deadly01:33
Bearcat ha01:33
canton7++01:33
← mxweas left01:33
← sythe_ left01:34
Bearcat so ..uh...where is this sskm bin supposed to be located. Is it part of the gitolite package?01:34
→ sythe_ joined01:34
→ MarconM joined01:35
→ mxweas joined01:35
→ nemo_nihil joined01:35
→ antihero joined01:36
→ nomadtacos joined01:36
antihero how do I make a patch file for a specific commit?01:37
← daed left01:37
→ SpencerDailey_ joined01:38
SethRobertson git show COMMITSHA; git format-patch; etc01:39
→ sythe__ joined01:39
← sythe_ left01:39
→ herdingcat joined01:39
← Chowzzf left01:40
← Heimidal left01:41
→ cooldude joined01:43
← sythe__ left01:44
→ daed joined01:45
← nomadtacos left01:47
→ sythe_ joined01:47
← tty234 left01:48
← dr_win left01:48
nemo_nihil i have initialized a git repo and i get "Initialized empty Git repository in /home/git/myrepo.git/" and then on my local machine I run "git remote add origin [email@hidden.address] but then i gives me "fatal: Not a git repository (or any of the parent directories): .git" what am i doing wrong?01:48
bob2 you did 'git init --bare' right01:48
nemo_nihil yeah01:48
inside of the myrepo.git folder01:49
→ Chowzzf joined01:52
→ squiggy_ joined01:53
← mxweas left01:53
frogonwheels nemo_nihil: on your local machine are you in a git repo?01:54
→ ehsan joined01:55
frogonwheels nemo_nihil: I suspect you need to git init inside your local folder (presumably this is a first commit?)01:55
nemo_nihil no, i am trying to create a git repo on a cloud server and then copy it down01:55
→ mindlogger joined01:55
← acl__ left01:55
mindlogger guys I just accidentally removed my local repo how can I get my repo back from github?01:55
nemo_nihil ok that fixed it, am i using the right architecture for this tho?01:55
frogonwheels nemo_nihil: how are you providing content to the cloud server repo? or is that to come?01:55
nemo_nihil basically i am running a website and i want to use git as my revisiting system. it is all in the /var/www/ directory01:56
frogonwheels nemo_nihil: if this is your 'first commit' to the cloud server, then this is the right approach01:56
mindlogger: clone it01:56
→ acl_ joined01:57
frogonwheels mindlogger: wait, you removed .git ? or the whole tree?01:57
mindlogger frogonwheels: will I get all the detials for commits?01:57
acl_ does running "git gc --aggressive" remove dangling blobs/commits/trees from a repo? or is it not useful to remove these dangling things?01:57
mindlogger frogonwheels: I removed all .* files in that directory and the whole directory01:57
nemo_nihil ok so i would commit it from my local machine and then clone it to the /www/ directory?01:57
frogonwheels mindlogger: hopefully you didn't have anything you hadn't pushed up?01:57
mindlogger frogonwheels: no I had everything pushed up01:58
I just need to get them back to my local machine01:58
frogonwheels mindlogger: cool. just git init && git remote add <url to github>01:58
→ BiggFREE joined01:58
frogonwheels mindlogger: cool. just git init && git remote add origin <url to github>01:58
sorry01:58
mindlogger great, ty frogonwheels01:58
frogonwheels mindlogger: then git fetch01:58
← Slush- left01:58
frogonwheels mindlogger: anything uncomitted?01:58
SethRobertson acl_: Do not worry about dangling bits. They will go away ina few weeks. `git gc --aggressive` will get rid of those which are ready to be gotten rid of01:59
← mikeric left01:59
mindlogger frogonwheels: not really, everything was committed and pushed to the remote01:59
acl_ SethRobertson++01:59
SethRobertson: thanks - i thought it might be otherwise02:00
→ mikeric joined02:00
SethRobertson No. They are a valuable safety net02:00
← zulax left02:00
acl_ i'll keep that in mind02:00
← fly91 left02:01
→ gchristensen joined02:02
mindlogger frogonwheels: to push to github should I define my account with user and pass in the git config file?02:02
← Tureos left02:02
→ zdk joined02:03
mindlogger frogonwheels: I did what you said and it showed my branches and stuff but it didn't download the actual files!02:03
frogonwheels mindlogger: git fetch02:03
oh.. yeah02:03
→ twisted` joined02:03
mindlogger I did and I got remote: Counting objects: 1736, done.....02:03
← weiyang left02:04
frogonwheels mindlogger: yeah. almost there.. just let me think02:04
mindlogger: possibly git checkout -f <branch>02:05
mindlogger I can't see my branches02:05
frogonwheels mindlogger: there's nothing you want in your wd?02:05
mindlogger wd?02:05
→ dfr|mac_ joined02:05
frogonwheels mindlogger: working dir02:05
mindlogger: no, they're all 'remotes' still. git branch -a02:05
→ fr0stbyte joined02:05
→ replore_ joined02:05
mindlogger I get all remote branches02:06
no local02:06
frogonwheels if you have a branch 'origin/master' then git checkout master will check it out.02:06
mindlogger: it will create the local one and set up the tracking.02:06
just as if it were there.02:06
you might have to -f just to force the issue.02:06
← nemo_nihil left02:06
frogonwheels we probably should have just blown away the old dir and clone afresh..02:06
mindlogger frogonwheels: I didn't have the old dir, I just removed it02:08
how do I make a fresh clone ?02:08
frogonwheels mindlogger: git clone <url>02:08
mindlogger does it make a new dir or should I do that inside a dir ?02:09
frogonwheels mindlogger: that will make a new dir02:09
git clone <url> <dir>02:09
MrKeuner I have a git repository, if I tag versions of code as stable or staging, is it possible to pull the latest stable from the production server, and latest staging from the staginf server?02:11
mindlogger frogonwheels: thanks a lot [Solved]02:11
← Vampire0 left02:12
frogonwheels mindlogger: once again, remember the trick with branches - if you have remotes/origin/mybranch then git checkout mybranch will checokut mybranch and set up tracking02:12
imMute MrKeuner: tags dont change what they tag. once you make a tag pointing to a commit, its set in stone02:12
→ Goplat joined02:12
frogonwheels MrKeuner: tags shouldn't move around really. branches are the things that float02:13
mindlogger frogonwheels: right gochya!02:13
MrKeuner frogonwheels, have a stable, staging and a dev branch and pull only relevant branch from each server?02:13
RandalSchwartz not pull, checkout02:14
unless you also need to make local commits02:14
frogonwheels MrKeuner: ^^ what he said02:14
RandalSchwartz "pull" doesn't mean in git what it means most places02:14
pull = fetch + merge, for when you have local mods that you want to merge with upstream02:14
frogonwheels MrKeuner: !pull02:14
gitinfo MrKeuner: 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/02:14
MrKeuner checkout means the copy serving on web server folder is not actually a git repository but a copy of the latest commit?02:14
RandalSchwartz no - it forces the git workdir to be the same as a commit02:15
git checkout stable02:15
(workdir becomes "stable" commit)02:15
← Sigma left02:15
frogonwheels MrKeuner: you can use man git-export to grab a copy02:15
err.02:15
RandalSchwartz !man git-archive02:15
gitinfo the git-archive manpage is available at http://jk.gs/git-archive.html02:15
→ Chillance joined02:15
frogonwheels bleh02:15
RandalSchwartz: thanks02:15
that's what I meant02:16
RandalSchwartz I've made that mistake before :)02:16
← MarconM left02:16
← DarkAR left02:18
MrKeuner post-receive hook does a GIT_WORK_TREE=/var/websites/stable git checkout -f, right? this is what you suggest02:18
→ EricInBNE joined02:18
← BSaboia left02:18
tizzo-afk → tizzo02:19
frogonwheels MrKeuner: !website02:19
gitinfo MrKeuner: 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-howto02:19
← leo left02:19
← lqi left02:19
RandalSchwartz there's also https://gist.github.com/171423502:19
git-archive piped into tar then rsync02:19
if you need to meld with existing files, that's the model02:19
← Lgb left02:20
MrKeuner thank you. I'll read these02:20
→ D-Chymera1 joined02:21
nevyn so...02:21
← D-Chymera left02:21
frogonwheels RandalSchwartz: !website02:22
gitinfo RandalSchwartz: 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/171423502:22
frogonwheels :)02:22
← hyperair left02:22
RandalSchwartz thx02:22
nevyn so what if your production environment does something useful with branches?02:22
RandalSchwartz "does something useful"?02:23
nevyn sure.02:23
so vhost per branch02:23
→ Lgb joined02:23
→ leo joined02:23
← fr0stbyte left02:23
→ logankoester joined02:23
RandalSchwartz sure - then tie each one to a dir02:23
nevyn or customer per branch and when it's on the server the requests are fetched from the local git repo. not a checkout02:23
RandalSchwartz you can even just loop it02:24
nevyn so I'm not actually talking about websites02:24
RandalSchwartz "fetched from local git repo"?02:24
nevyn I'm talking about puppet02:24
where there are patterns for using git branches for environments...02:24
http://hunnur.com/blog/2010/10/dynamic-git-branch-puppet-environments/02:25
like this ^^^02:25
→ tqmz__ joined02:25
→ lqi joined02:25
nevyn so to create a new environment you just make a new branch and push it to the server. then test your client against that environment02:26
← tqmz__ left02:26
→ Vampire0 joined02:26
→ hyperair joined02:26
← abetaha left02:27
← cccaldas left02:28
← Chib left02:28
→ solidus-river joined02:28
solidus-river how do i revert the most recent merge to a branch?02:28
this is driving me insane02:28
i see02:28
git revert HEAD02:28
← neurodrone left02:28
solidus-river but it complains someting about merge and an -m option02:28
i've done no work since then02:29
← tqmz left02:29
→ lyddonb joined02:29
→ CAAEF5Y joined02:29
→ pilotbub joined02:29
→ zulax joined02:29
solidus-river there shoudl be a simply worded how-to on this02:30
nevyn do you want to revert? or just undo it?02:30
SethRobertson solidus-river: !fixit02:30
gitinfo solidus-river: [!fixup] 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)02:30
← griotspeak left02:33
← thebignoise left02:33
solidus-river how do you find out parent numbers?02:34
imMute git log02:35
← mtytel left02:36
← mcluskydodallas left02:38
→ abetaha joined02:39
← ehsan left02:41
→ diogogmt joined02:41
← hyperair left02:42
→ davidrussell joined02:42
→ TL joined02:43
→ weiyang joined02:44
→ ehsan joined02:44
→ Yuuhi` joined02:45
→ hyperair joined02:45
→ tokiya joined02:46
← diogogmt left02:46
← Yuuhi left02:46
← BiggFREE left02:46
→ BiggFREE joined02:47
→ diogogmt joined02:47
← pilotbub left02:47
→ w09x joined02:48
← w09x left02:49
→ w09x joined02:50
→ ASIRINeiluj joined02:50
← rickmasta left02:51
← diogogmt left02:52
→ diogogmt joined02:53
← mikeric left02:53
← davidrussell left02:55
→ dv310p3r joined02:56
← apok left02:57
→ gavin_huang joined02:57
← solidus-river left02:57
← zendeavor left02:58
zdk my history is A--B--C--D, I would like it to be A--C--D. any suggestions for the git command to do this?02:58
← w09x left02:58
SethRobertson zdk: !fixup02:58
gitinfo zdk: 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)02:58
← Bearcat left02:58
→ krz joined02:59
zdk SethRobertson: nice, thanks!02:59
← jason237 left03:00
→ rurufufuss joined03:00
← Chibby_ left03:01
→ Chibby_ joined03:01
← BiggFREE left03:02
→ xyloph joined03:03
→ zendeavor joined03:03
← diogogmt left03:04
← xyloph left03:05
→ diogogmt joined03:05
← krz left03:05
→ jason237 joined03:07
→ Destos joined03:09
← cakehero left03:10
→ wuj joined03:11
→ neurodrone joined03:11
← D-Chymera1 left03:12
→ D-Chymera joined03:13
← ipalaus left03:13
→ xyloph_ joined03:15
→ luoluoluo joined03:15
← diogogmt left03:17
← xyloph_ left03:18
← Kichael left03:18
wald0 if i have feature A saved in a commit, and i want to give it to the original author (remote repository), what's the normal procedure ? and how he can get updates about such feature ?03:18
← xiangfu left03:18
→ kermit joined03:19
SethRobertson Is this on github03:19
→ diogogmt joined03:21
← TL left03:21
← diogogmt left03:23
← zulax left03:23
→ diogogmt joined03:23
→ kennethr_ joined03:24
← merlin83_ left03:25
← alexgordon left03:25
→ xiangfu joined03:25
→ alberto56 joined03:25
→ fr0stbyte joined03:26
← kadoban left03:26
→ jaisoares joined03:26
→ wbednarski joined03:27
← wuj left03:27
← kennethreitz left03:28
← diogogmt left03:28
← philllllijw left03:29
→ diogogmt joined03:29
← f0i left03:31
sigs I'm guessing github has some system about keeping track of forks?03:31
← akiress left03:31
← acl_ left03:31
← mattalexx left03:31
← SpencerDailey_ left03:32
→ SpencerDailey_ joined03:33
← gchristensen left03:33
← diogogmt left03:36
← maletor left03:37
→ diogogmt joined03:37
← orospakr left03:38
← ihodes left03:39
← squiggy_ left03:40
→ maletor joined03:40
EugeneKay Yes, they track parents explicitly when you push the Fork button03:41
It's antique and inflexible and very un-git-like03:41
← orafu left03:41
→ mandric joined03:41
→ orafu joined03:42
milki but it works!03:42
← diogogmt left03:42
→ merlin83_ joined03:43
→ diogogmt joined03:43
← diogogmt left03:45
→ luxurymode joined03:45
→ diogogmt joined03:46
→ ormris joined03:46
← dv310p3r left03:46
gitinfo set mode: +v03:46
← maletor left03:50
← SpencerDailey_ left03:52
← kennethr_ left03:53
→ kennethreitz joined03:53
→ Heimidal joined03:54
← alberto56 left03:54
← sattu94__ left03:54
→ satyanash joined03:54
→ SpencerDailey_ joined03:55
← guampa left03:56
→ akiress joined03:57
← ko1 left03:57
← k1o left03:57
← kennethreitz left03:58
← SpencerDailey_ left03:58
← diogogmt left04:00
→ diogogmt joined04:00
→ ihodes joined04:03
← fr0stbyte left04:07
→ fr0stbyte_ joined04:07
← diogogmt left04:07
→ diogogmt joined04:08
→ tiglog joined04:08
← hyperair left04:08
← tiglog left04:09
→ tiglog joined04:09
→ tylere joined04:10
tylere Need some help. Changes files, cloned to a new branch with checkout -b, did a git add ., and then a commit.04:10
It looks like it logged it as a commit from "nothing", rather than as a branch from master04:11
so it logged everything as an addition04:11
even the unchanged files04:11
← diogogmt left04:13
→ diogogmt joined04:13
← ColKurtz left04:16
← cooldude left04:16
← diogogmt left04:16
→ diogogmt joined04:17
← fr0stbyte_ left04:18
← tystr left04:18
← Turicas left04:20
→ davidrussell joined04:20
← Heimidal left04:21
← diogogmt left04:21
→ diogogmt joined04:22
→ cooldude joined04:25
wald0 EugeneKay: why you mean un-git-like ?04:27
EugeneKay It's strongly "centralized" to Github. You can't clone a random git:// link from the web, or change the "parent" of your clone, etc04:28
wald0 so is there no such feature out of github ?04:28
← Destos left04:29
wald0 what about i create feature A in a different branch, and i want to give my changes to the author of my git fork, he will be "track" my branch looking for new commits ?04:30
EugeneKay GitHub does it by pull requests, usually.04:33
← diogogmt left04:33
→ diogogmt joined04:34
wald0 yes but outside github (the author of my fork doesnt use it)04:34
→ ChaoWang joined04:35
← Wombert left04:37
← diogogmt left04:38
→ diogogmt joined04:39
← eletuchy left04:39
EugeneKay Provide a public git:// URL someplace he can pull, or email him a patch04:39
wald0 right, that's the idea, but how he will be noticed/updated about future changes on feature A from my repo ?04:42
EugeneKay You send him a fresh email?04:42
→ jfelchner joined04:42
wald0 no, im searching for an automated way so that i dont need to do that04:43
or think/remember about that04:43
→ kennethreitz joined04:44
EugeneKay post-commit hook involving 'mailx'04:44
← dreiss left04:44
← kennethreitz left04:45
← cooldude left04:45
← ormris left04:47
← Dmole_ left04:47
→ krz joined04:47
→ glennpratt joined04:49
→ caseymcg joined04:55
← mindlogger left04:56
→ kennethreitz joined04:59
→ beautifulmind joined05:00
← diogogmt left05:02
→ diogogmt joined05:03
← glennpratt left05:03
← mikepack left05:04
→ mikepack joined05:05
← krz left05:05
→ krz joined05:05
← diogogmt left05:08
→ diogogmt joined05:08
← sythe_ left05:08
← mikepack left05:09
← soidexe left05:09
← diogogmt left05:13
→ diogogmt joined05:14
→ cnorman joined05:14
→ mikepack joined05:14
← conradev left05:15
← diogogmt left05:15
→ diogogmt joined05:16
→ fayimora joined05:17
← legnaleurc left05:17
← jamescarr left05:18
→ fr0stbyte joined05:18
← rchavik left05:19
→ legnaleurc joined05:19
← wald0 left05:21
→ wald0 joined05:23
← mikepack left05:24
← kukks left05:24
← cnorman left05:24
← legnaleurc left05:24
→ LongBeach joined05:24
→ mikepack joined05:24
← mikepack left05:25
→ legnaleurc joined05:25
→ mikepack joined05:25
← diogogmt left05:28
← legnaleurc left05:29
→ legnaleurc joined05:30
→ diogogmt joined05:30
← Bass10 left05:30
→ dnjaramba_ joined05:33
wald0 how i can see the changes that was introduced in a branch since its creation ?05:33
← legnaleurc left05:34
bob2 that's hard05:34
← dnjaramba left05:34
bob2 but 'git log -p somebranch' is easy05:34
← diogogmt left05:34
→ legnaleurc joined05:35
bob2 hm maybe merge-base is plenty05:35
→ dnjaramba joined05:35
→ diogogmt joined05:35
wald0 bob2: git log -p branch seems like to show an output of all the changes from my master, to that old branch05:36
← codeshah left05:36
→ kadoban joined05:37
bob2 it shows the whole history of the branch yes05:37
← matthiaskrgr left05:37
→ pilotbub joined05:38
← dnjaramba_ left05:38
wald0 supposed that i create now a branch, and i do 3 different commits, how i can see -only- this history ?05:38
bob2 like i said, try merge-base05:38
← lyddonb left05:39
→ matthiaskrgr joined05:39
← replore_ left05:40
← diogogmt left05:40
→ Heimidal joined05:41
→ diogogmt joined05:41
→ Bass10 joined05:41
→ legnaleurc_ joined05:41
← legnaleurc left05:41
→ smw joined05:41
← mandric left05:45
← wbednarski left05:45
sente I have a simple question: http://c.sente.cc/0zSh/git -- how can I compare the remote repo to the current files without actually pulling? i'm very to git05:45
→ jony joined05:46
→ hyperair joined05:46
← Bass10 left05:46
← diogogmt left05:46
→ dreiss joined05:46
→ diogogmt joined05:46
milki do a git fetch05:48
instead05:48
← legnaleurc_ left05:48
milki and then compare it locally05:48
← kryl99 left05:49
milki you would be comparing origin/master to master05:49
← codewrangler left05:51
← pilotbub left05:51
← diogogmt left05:51
← macmartine left05:52
milki git pull does a fetch+merge05:52
git fetch will not change your working dir05:52
→ freimatz joined05:53
→ toabctl joined05:53
sente okay, thanks05:53
→ legnaleurc joined05:54
← fr0stbyte left05:54
wald0 maybe: git log HEAD..origin/master , sente, im just reading about it05:56
← hyperair left05:57
← aristidesfl left05:57
← ihodes left05:57
→ acl_ joined05:57
← krz left05:57
← dreiss left05:57
→ talios joined05:58
sente thanks06:00
→ macmartine joined06:02
→ legnaleurc_ joined06:03
← legnaleurc left06:04
→ ajaykools joined06:04
MrKeuner http://toroid.org/ams/git-website-howto on this website, the author creates a local repository and a remote bare repository and add bare repository as remote for local repository. Isn't that the same as creating a remote bare repository and git clone it to local ?06:06
bob2 well, i don't htink you can clone an empty bare repo06:09
← ajaykools left06:09
→ ajaykools joined06:10
← legnaleurc_ left06:10
→ legnaleurc joined06:12
← ceej left06:12
← caseymcg left06:13
milki ya you can06:14
itll just say you cloned an empty repo06:14
→ Slush- joined06:14
milki MrKeuner: yes its the same06:14
its how github repos are linked06:15
→ ThomasLocke joined06:15
→ fr0stbyte joined06:16
← alexim left06:16
wereHamster milki: eh? Github?06:17
milki github's new repo page?06:17
says to add remote06:17
MrKeuner I see, though when you clone your remote is called origin by default06:18
milki yes06:18
MrKeuner I guess you would have to remove that rmeote and add same remote with name web06:18
to comply with example in the text06:18
milki ya, but any programmer can handle variable changes like that06:18
MrKeuner just trying to understand why he did it that way06:19
or she06:19
→ harshpb joined06:19
← davidrussell left06:20
→ ihodes joined06:21
← legnaleurc left06:21
→ legnaleurc joined06:22
MrKeuner I don;t get this :) "I'm an egotistical bastard, and I name all my projects after myself. First Linux, now git." ~ Linus Torvalds06:22
so he doesn't name after himself?06:22
os is there a relation06:22
wald0 bob2: aparently, showing the work of a branch -only- can be done with "git diff HEAD..branch"06:22
milki somewhere explains what git means06:23
its a colloquial term06:23
sigs MrKeuner: http://www.etymonline.com/index.php?search=git06:23
bob2 wald0, nah06:23
→ martinjlowm-lptp joined06:25
→ martinjlowm joined06:25
← Targen left06:27
MrKeuner hehe06:28
← luoluoluo left06:30
→ LouisJB joined06:30
→ mabrand joined06:30
→ LeMike joined06:31
← _iron left06:32
→ Dave^| joined06:33
MrKeuner kind of recursive naming06:33
→ caseymcg joined06:35
← Goplat left06:36
→ artib joined06:37
→ disappearedng joined06:38
→ hyperair joined06:39
→ TrineLocke joined06:39
→ pidus joined06:40
← legnaleurc left06:41
← mikepack left06:42
→ marthinal joined06:42
→ legnaleurc joined06:43
MrKeuner I seek your wisdom: If I use my laptop as dev repo, my home desktop as stage repo, and website as production repo. and push changes to either dev, stage or web (and each repo will publish to a detached work tree on that machine) would this setup cause any problems?06:43
→ JohnBat26 joined06:45
← smw left06:45
MrKeuner how about this: I use my website for single repo which will have three branches dev, stage and prod, and post-receive will publish to etiher of three detached work tree /var/websites/dev or /var/websites/stage or /var/websites/prod06:46
← Chappo left06:47
→ mxweas joined06:47
← CEnnis91 left06:47
← r126t left06:47
← forrestv left06:47
← nru left06:47
← kokx left06:47
← pdobrogost left06:47
← koivula left06:47
← roland left06:47
← mfischer left06:47
← xpot-mobile left06:47
← hays|afk left06:47
← Ferris left06:47
← kevlarman left06:47
← mpenz left06:47
← Fudge left06:47
← crash\ left06:47
← totte left06:47
← FauxFaux left06:47
← surfdue left06:47
← mjt left06:47
← Pieter left06:47
← nevyn left06:47
← cmn left06:47
← zackiv31 left06:47
← milk left06:47
← GrowLinux left06:47
← blueyed left06:47
← xiaotian_ left06:47
← elliottcable left06:47
milki sounds fine...06:47
→ tqmz joined06:47
→ elliottcable joined06:47
← hyperair left06:47
→ antijava joined06:47
← legnaleurc left06:49
antijava Any idea why git archive barfs when I pass in a known good commit id? http://pastebin.com/8bMTXkD306:49
MrKeuner milki, both?06:49
milki ya06:50
1st on has different remotes basically06:50
and 2nd one has different branches06:50
→ legnaleurc joined06:50
→ jacekmigacz joined06:50
→ break57382 joined06:51
→ nru joined06:51
→ CEnnis91 joined06:51
→ r126t joined06:51
→ forrestv joined06:51
→ kokx joined06:51
→ pdobrogost joined06:51
→ mfischer joined06:51
→ koivula joined06:51
→ roland joined06:51
→ xpot-mobile joined06:51
→ hays|afk joined06:51
→ Ferris joined06:51
→ kevlarman joined06:51
→ mpenz joined06:51
→ Fudge joined06:51
→ crash\ joined06:51
→ totte joined06:51
→ mjt joined06:51
→ FauxFaux joined06:51
→ surfdue joined06:51
→ nevyn joined06:51
→ Pieter joined06:51
→ cmn joined06:51
→ zackiv31 joined06:51
→ milk joined06:51
→ GrowLinux joined06:51
→ blueyed joined06:51
→ xiaotian_ joined06:51
wolfe.freenode.net set mode: +v06:51
← toabctl left06:51
MrKeuner milki, I am not sure if I can run different post-receive hooks depending on which branch is updated. Is that possible?06:52
milki you wouldnt, you would detect what branch was committed to06:53
→ hyperair joined06:54
milki antijava: nope, i havent tried remote before...06:55
→ kerrick joined06:55
antijava If I just use master it works. But a raw hex id doesn't...06:56
The problem is that it is two separate command executions, so theoretically someone could have committed between the two operations.06:56
I want to get the id of master in the first call, and then archive that known point in time in the 2nd06:56
→ smw joined06:56
→ macrover joined06:57
MrKeuner milki, you mean I have to detect using git log | grep etc.?06:59
which I could06:59
← legnaleurc left06:59
milki doesnt post-receive have parameters?06:59
← artib left06:59
MrKeuner I have no idea06:59
milki find out?06:59
← acl_ left07:00
→ legnaleurc joined07:00
→ psakrii joined07:00
MrKeuner should I be reading git hooks?07:00
milki yes...07:00
← mxweas left07:00
← xpot-mobile left07:01
→ artib joined07:01
← antijava left07:02
→ xpot-mobile joined07:02
wereHamster MrKeuner: no, man githooks should be enough07:02
gitinfo MrKeuner: the githooks manpage is available at http://jk.gs/githooks.html07:02
milki isnt that the same thign?07:02
→ mxweas joined07:03
wereHamster oh, I misread him as 'should I be reading git books' -.-07:03
← daed left07:03
← LouisJB left07:03
→ dreiss joined07:03
MrKeuner hehe, thanks anyways07:03
I'll go read hooks07:03
hukbuks07:03
← mbrevda left07:04
→ mishok13 joined07:04
→ h0bbit joined07:05
← legnaleurc left07:07
← VladGh left07:07
→ c0w joined07:09
→ Karmaon joined07:10
→ legnaleurc joined07:10
→ scotters joined07:11
ChanServ set mode: +v07:11
scotters 'evening everyone!07:12
← c0w left07:12
← mxweas left07:13
→ c0w joined07:13
scotters quick q: Git-scm.com lists the latest stable release as 1.7.6.6, but the Mac binaries are at 1.7.7 and above. I'd like to upgrade from 1.7.1 on Snow Leopard and build from source. --- so, which release should I use?07:13
→ cbreak-work joined07:13
← c0w left07:14
← elnn left07:14
→ elnn joined07:14
← JohnBat26 left07:14
← duckxx left07:15
wereHamster scotters: read the topic07:16
milki http://code.google.com/p/git-osx-installer/wiki/BuildYourOwn07:17
scotters :) okay! That's the 5th different potential version I've encountered. It's hard to know what's currently canonical.07:17
milki o07:17
what07:17
→ r3wn joined07:17
r3wn hi07:17
milki is the latest version different from stable?07:17
scotters according to git-scm.com, yes.07:18
r3wn whats the best way to setup a private git repo? create a dir locked down with ssh access with private keys for all users?07:18
milki huh07:18
← xiangfu left07:18
milki r3wn: depends on number of users07:18
← KyleXY left07:18
r3wn 5-10 at most07:18
milki and how much control over access you want07:18
for a small number, ssh access is fine07:19
anymore, you might want to look at gitolite07:19
wereHamster scotters: the one in the topic is canonical07:19
→ KyleXY joined07:19
→ Sitethief joined07:19
wereHamster scotters: or also the latest tag in the git.git repository07:19
r3wn whats a typical setup with ssh?07:19
← legnaleurc left07:19
r3wn using keys?07:19
scotters wereHamster: thanks. That's why I stopped in. Much appreciated.07:19
r3wn 1 main user with keys for all other users?07:19
→ JohnBat26 joined07:20
MacGyverNL uses gitolite for his centralized repo's, even though most are accessed by a single person and none by more than four.07:20
milki yes, one main user07:20
scotters so why is the git-scm.com page listing 1.7.6.6?07:20
→ airborn joined07:20
MacGyverNL It's extremely easy to setup, and much safer than trusting those people with any other kind of push method which somehow involves me restricting their shell access but still enables them using git.07:20
And by "extremely easy" I mean: You can have it up and running in less than 30 minutes if you already know how git and SSH with public key crypto works.07:21
← boombatower left07:22
tango_ is setting up his first gitolite repository thingie, and has just found out that machine has no ssh open to the outside 8-/07:26
← macrover left07:27
→ senny joined07:28
scotters → scottley07:29
MacGyverNL tango_: But that means you're spending time setting up SSH in general - not gitolite ;)07:30
→ legnaleurc_ joined07:30
← scottley left07:31
→ digitteknohippie joined07:31
milki >.<07:31
no ssh1?07:31
bah07:31
!07:31
→ Shusshu joined07:31
digitteknohippie finally trying to get to grips with git. github's confusing me though... when it says "git remote add origin [email@hidden.address] do i just change the email address to my own, and leave the rest as it? or do i leave all like that, or...?07:33
tango_ MacGyverNL: yeah, actually the ssh is set up already and I'm using it, but the machine is only ssh-accessible from the intranet07:33
→ maletor joined07:33
milki digitteknohippie: no, thats not an email address07:33
→ csherin joined07:34
milki digitteknohippie: leave it as it is07:34
digitteknohippie ah, i see now. it's much like an address if you were ssh'ing.07:34
milki yes07:34
digitteknohippie thnx milki07:34
milki thats why you needed to upload your public key07:34
tango_ it's exactly that07:34
← fayimora left07:35
r3wn can you use periods in repo names?07:35
← legnaleurc_ left07:35
milki if it doesnt complain i guess?07:36
← cenuij left07:36
← beautifulmind left07:37
→ legnaleurc joined07:37
→ xiangfu joined07:37
→ aspotashev joined07:37
→ giallu joined07:39
← kerrick left07:39
→ gusnan joined07:40
→ teweWORK joined07:40
← chrisdev left07:41
→ papegaaij joined07:43
← Heimidal left07:44
← aspotashev left07:45
← fr0stbyte left07:46
→ Wombert joined07:47
← mishok13 left07:47
← legnaleurc left07:47
→ legnaleurc joined07:48
→ nixmaniack joined07:48
MrKeuner I have been reading this and other documents but don't have much luck yet, any hints as to how to detect different branch commits and run checkouts on different GIT_WORK_TREEs07:50
← twisted` left07:50
MrKeuner What I am looking for is http://toroid.org/ams/git-website-howto except for multiple branches and multiple detached trees instead of one07:52
→ legnaleurc_ joined07:54
← break57382 left07:54
← legnaleurc left07:54
sitaram MacGyverNL: 30 minutes? too much! in many cases gitolite install can be as simple as "yum install gitolite", copy your ~/.ssh/id_rsa.pub from your workstation as /tmp/YourName.pub on server, then "gl-setup -q /tmp/YourName.pub" as gitolite user07:55
MacGyverNL: then maybe count time to clone the admin repo to your workstation I guess but after that it's routine maint -- adding keys, repos etc07:56
→ cenuij joined07:57
MacGyverNL sitaram: I know - I included the time it took me to read the first few parts of the docs.07:59
← airborn left07:59
sitaram MacGyverNL: aah ok; that's fair then :-)08:01
→ k_sze joined08:01
sitaram I *have* sometimes been told "you wrote too much docs!"08:01
k_sze Hello.08:01
tango_ there is no such thing as 'too much docs'08:02
k_sze How do I peek at the content of a file at a specific tag?08:02
← legnaleurc_ left08:02
jast EugeneKay: nope. it just includes that line.08:02
MacGyverNL sitaram: People who say that are usually people who also think "docs? Why? You've got the code haven't you?" is an acceptable answer ;)08:02
tango_ k_sze: git cat-file tag:/path/to/file08:02
→ batlock666 joined08:02
EugeneKay jast - I've forgoten the context in my drankennas08:03
+s08:03
jast you asked about commit -s08:03
← dfr|mac_ left08:03
→ dfr|mac__ joined08:03
EugeneKay Ahhh yes08:03
sitaram MacGyverNL: or my favourite: "hop onto #git and hope for the best" ;-)08:03
jast also I've just merged your patch08:03
EugeneKay So my email says08:03
jast and "deployed" it08:03
EugeneKay mumbles something about git is not a deployment tool08:04
← artib left08:04
jast quite so! I'm a repeat offender in UGFWIINI, though, so it's okay08:04
MacGyverNL sitaram: By the way, I was telling somebody about gitolite a few days ago, he was using a self-compiled gitosis while his repo's provide gitolite - and I was unsure about one thing: did you also develop gitosis and just start anew with gitolite, or was that built by another dev / group?08:04
→ dfr|mac_ joined08:05
sitaram MacGyverNL: gitosis was by someone else. An author who, according to someone on this channel once, does not even reply to emails from his CEO, much less requests for support or feature requests ;-)08:05
MacGyverNL Ah, right, thanks.08:05
sitaram MacGyverNL: so I wrote my own08:05
← dfr|mac__ left08:05
→ dfr|mac joined08:06
tango_ did you fork gitosis or just wrote something new from scratch08:06
← xiangfu left08:06
jast gitosis was written by Tommi Virtanen, over at eagain.net08:06
gitolite is written from scratch08:06
different programming language and all08:06
sitaram tango_: stole the basic idea (ssh, command=... in authkeys file), kept the filename conventions, and changed everything else08:07
→ flijten joined08:07
sitaram tango_: although I believe the basic "ssh, authkeys..." idea is older than gitosis; it's standard openssh after all08:07
→ OOPMan joined08:08
tango_ does gitosis predate stuff like github or gitorious?08:08
→ lars___ joined08:08
sitaram no idea08:08
← macmartine left08:08
sitaram about the same I suppose but it should be easy to check08:08
jast I knew about gitosis before I knew about github, I believe08:08
but that doesn't say much08:09
k_sze tango_: $ git cat-file tag 1.16:mHHT/Properties/AssemblyInfo.cs08:09
fatal: git cat-file 1.16:mHHT/Properties/AssemblyInfo.cs: bad file08:09
← Chaser left08:10
jast gitosis development started in june 200708:10
sitaram tango_: gitosis first commit 2007-05-30, github launched 2008-04 (code may be older)08:10
← dfr|mac_ left08:10
tango_ k_sze: git cat-file -p 1.16:mHHTetc08:10
MacGyverNL Thing with github is we don't *know* holw old the code is.08:10
sitaram tango_: jast may be right in terms of when gitosis was made public; the commit time does not mean it was published08:10
← Wombert left08:10
MacGyverNL (Which is at once my main gripe with the service)08:10
jast well I'm going from the date of the first commit in the official repository08:10
→ Chaser joined08:11
→ legnaleurc joined08:11
jast MacGyverNL: how so?08:11
sitaram MacGyverNL: why is that a gripe? just another "SaaS" no?08:11
→ BiggFREE joined08:11
jast ah well, bbl08:11
MacGyverNL jast: I have a sort of problem with hosting one of the world's biggest repositories of open source software on a closed source platform.08:12
→ vdv joined08:12
← Khisanth left08:12
MacGyverNL But that's my irrational opinion, ignore me by all means - github is awesome, I just have a difference of opinion.08:12
A rather hypocritical one at that, since I use Windows as one of my main OS's.08:12
sitaram MacGyverNL: well they have to make money somehow; I guess I'm not a complete RMS anyway08:13
oh... bad!08:13
sitaram has no windows at home or work08:13
sitaram (well not *my* work desktop / laptop anyway)08:13
MacGyverNL Yes, yes, they have to make money, true - but with a model like github, money and publishing source doesn't have to be mutually exclusive...08:13
But it's a choice up to them, really.08:14
It's just that, if gitorious was as feature-rich as github, I'd vastly prefer it because that one is open source.08:14
→ soidexe joined08:14
sitaram MacGyverNL: ever talk to a VC (I'm not saying that is the case here, but you can get a lot more funding if you can show that you are trying to maintain a differentiator)08:14
→ raatiniemi joined08:15
← soidexe left08:15
← legnaleurc left08:15
→ psoo joined08:15
→ macrover joined08:16
→ legnaleurc joined08:16
MacGyverNL Understanbadable.08:16
Meh. SSH fail.08:16
Understandable*08:16
← mdel left08:16
k_sze tango: thanks. :D08:17
→ dpino joined08:17
→ LucasTizma joined08:17
MacGyverNL But I should also add I'm quite paranoid and have a security background - in my book, something that's not open sourced has automatic points against it for not being inspectable. And yes, that's unreasonable, and yes, I still use those services if they're good and there's no viable alternative.08:18
But it's unfair to apply that to github. I know.08:18
I can't make a real rational argument for this without at some point introducing philosophical differences.08:19
→ xiangfu joined08:19
MacGyverNL That's why I said you should ignore me.08:19
← Lgb left08:20
→ stodan joined08:20
← legnaleurc left08:20
MacGyverNL (To illustrate: LastPass is a very nice password service. But I have no way to inspect it and therefore I am not going to give my full database of passwords to them - even though many thousands of other people trust them, and even though they (claim to) use AES256 for everything that gets sent to them.08:21
Instead, I use KeePass, which is somewhat less easy to use but I have checked the code.)08:21
→ dc5ala joined08:21
← herdingcat left08:21
→ sameer_ joined08:21
← mva left08:21
→ yshh joined08:21
→ Wombert joined08:22
← JackWinter left08:22
→ JackWinter joined08:22
→ mva joined08:22
← elnn left08:23
→ ipalaus joined08:23
→ Lgb joined08:23
← MrKeuner left08:23
→ snearch joined08:23
→ paradoxgo joined08:23
→ elnn joined08:23
← hyperair left08:24
← paradoxgo left08:24
→ jceb joined08:25
gitinfo set mode: +v08:26
sameer_ Hi, I have done a commit with wrong message and pushed to remote repo(I realised at this point and did a ctrc+c, looked at UI of bitbucket dashboard, didnt show my pushed commit). So I locally did a `commit --amend` and tried to push then it was rejected(apparantly it was pushed!). I know that noone has yet pulled the latest comit, but might have pulled this branch earlier. Is it okay to force push?08:26
jast MacGyverNL: if you don't mind a complete lack of features, you can give repo.or.cz a try. if nothing else, I know it has decent support. [I'm one of the guys who does that...]08:26
← tiglog left08:26
→ mishok13 joined08:26
← Araxia left08:27
jast in fact, some of the limitations are in place because we have yet to figure out a way to get rid of them without introducing mostly unrealistic security issues...08:27
→ SPYGAME joined08:27
→ Khisanth joined08:27
jast sameer_: if you're quite certain that nobody has the old version of the commit, force push is okay08:27
sameer_ what does it mean by 'force push changes the history.' It will change only the diverged i.e. the amended commits history, right?08:27
jast sameer_: yes.08:28
← roland left08:28
MacGyverNL jast: Well, I now have gitolite which is enough for my purposes - I'm not involved in any large open source projects.08:29
sameer_ jast: thanks :). so the history of the older commits remains unchanged, right?08:29
MacGyverNL But it sounds interesting.08:29
jast sameer_: yes08:29
→ d3vic3 joined08:29
sameer_ <jast>: thanks :)08:30
← mva left08:30
→ roland joined08:30
→ thiago joined08:31
→ gnufied joined08:32
→ Araxia joined08:34
→ retoryter joined08:35
← tatsuya_o left08:36
→ gldn joined08:36
← LucasTizma left08:38
→ eijk joined08:39
← ihodes left08:40
→ sindresorhus joined08:40
→ rendar joined08:42
→ leavittx joined08:42
→ twisted` joined08:43
← lapistano left08:43
sitaram MacGyverNL: weirdly enough, gitolite *needs* github for one specific purpose: since I don't have a 'web site' anywhere else, I need a place to put the rendered HTML docs! Github does that great...08:43
→ lapistano joined08:43
→ legnaleurc joined08:45
→ shergill1 joined08:46
← shergill left08:46
MacGyverNL And gitorious most likely doesn't, right?08:47
sitaram MacGyverNL: no idea.08:48
→ BSaboia joined08:48
MacGyverNL (I looked at self-hosted gitorious before deciding on gitolite, but it was too much trouble for features I was never going to use, and features I expected weren't in there, so yeah...)08:48
← dnjaramba left08:50
sitaram MacGyverNL: also see https://github.com/sitaramc/hap (documentation is at the bottom of the source code). This is what I use and I highly recommend it. No passwords to store...08:50
many people don't like the concept, though. I find it perfectly acceptable08:50
MacGyverNL Speaking of "github" and "great docs" - I've recently encountered some 404's on links to your docs from other sites - did the structure change?08:50
→ dnjaramba joined08:50
← legnaleurc left08:50
→ legnaleurc joined08:51
← iamjarvo left08:51
← leavittx left08:51
← sindresorhus left08:52
sitaram MacGyverNL: yes they did; see the commit message for 6e2936508:52
→ arvind_khadri joined08:52
→ dr_lepper joined08:55
← dr_lepper left08:55
→ dr_lepper joined08:55
→ legnaleurc_ joined08:55
→ break57382 joined08:56
← legnaleurc left08:56
→ whitman joined08:56
→ Cyanure joined08:59
→ shruggar joined09:00
gitinfo set mode: +v09:00
Cyanure Hi09:00
MacGyverNL Wow - that *is* a huge change. But you have pictures now! ^_^09:00
Cyanure How can I specify that my local branch "experimental" always pull from the remote origin branch "experimental" ? And can I do that with commandline without editing any configuration file09:01
→ pilotbub joined09:01
← legnaleurc_ left09:01
kini is there an environment variable I can set which will make git ignore any config variables which change its output from the default?09:03
something like Mercurial's HGPLAIN variable, basically09:04
→ legnaleurc joined09:04
→ hyperair joined09:06
cmn what are you trying to do?09:07
if you need predictable output, use the plumbing commands09:07
← hyperair left09:07
→ PigeonFriend joined09:07
← PigeonFriend left09:09
← LeMike left09:10
→ dorkmafia joined09:10
→ g105b joined09:10
→ Blaster joined09:10
← roland left09:10
Blaster hey when you install Git on windows do you use ssh.exe (OpenSSH) or PuTTY?09:10
→ robotmay joined09:11
← legnaleurc left09:12
→ roland joined09:12
← senny left09:13
cbreak-work I heard both work09:14
← snearch left09:14
cmn you can get git to work either of them so you can continue using what you're confortable with09:15
← g105b left09:15
Blaster any way to store the password for my repository? it's like [email@hidden.address]09:15
→ senny joined09:15
Blaster Dont want to have to type the password each time I do a git pull or git push09:15
cmn Cyanure: git branch --set-upstream exp origin/exp09:15
Blaster: that's not a password for your repo, that's your password on that machine09:16
no idea if you can tell ssh to cache passwords, but you probably want to use keys09:16
Cyanure cmn, I googled it and find the combo (git config branch.experimental.remote origin) and (git config branch.experimental.merge experimental)09:17
Blaster Where should I start?09:17
Cyanure cmn, is there a difference ?09:17
sitaram MacGyverNL: yes I have pictures, and you should see the source for those pictures -- all plain text!09:17
Blaster What would be a good google search?09:17
MacGyverNL graphviz?09:17
cmn Cyanure: one you're doing it, the other one you're telling git to do it09:17
sitaram MacGyverNL: graphviz for a couple of them, "ditaa" for the others09:17
→ PigeonFriend joined09:18
cehteh Blaster: ssh-agent09:18
→ LimeBlast joined09:19
Cyanure cmn, so it does the exact same thing ?09:20
kini cmn: just trying to run the test suite of hg-git (which I believe is based on mercurial's own test suite)09:20
outputs don't match because the test suite calls git, and I have config settings that make git log output branch names, etc. etc.09:20
cmn Cyanure: yes, that's what I said09:20
← Retistic left09:20
kini so I wondered if I could pass some environment variable to make git produce output at default settings09:21
cmn kini: then the tests are buggy09:21
kini I guess so, lol09:21
it'd still be nice to have such a feature, though :)09:21
cmn depending on the output of a porcelain command is a bug09:21
→ legnaleurc joined09:22
← pilotbub left09:23
cmn porcelain may start doing things differently as well, so then you have to deal with different behaviours as well as different outputs09:23
as I said, if you need stable output (and in any script, anyway) use plumbing commands09:23
← dorkmafia left09:25
← Vile left09:26
← legnaleurc left09:26
Nickeeh Can I view the patch I made in my commit-message-editor when I enter "git commit" ?09:27
nru Nickeeh: git commit -v09:28
← fisted left09:28
→ daed joined09:28
Nickeeh nru, thanks! :)09:29
→ fisted joined09:29
→ i42n joined09:30
→ glcrazy joined09:30
→ amigojapan joined09:31
← xiangfu left09:31
← neurodrone left09:32
→ eugenmayer joined09:32
eugenmayer guys, when writing aliases using sh -c .. how to preserve the pwd?09:33
← NickASloan left09:33
→ Dreamer3__ joined09:33
→ QaDeS joined09:33
← moqq left09:33
← Dreamer3 left09:34
→ legnaleurc joined09:34
← lottadot left09:34
← mtkd left09:34
→ moqq joined09:34
← JEEB left09:34
← ianmcorvidae left09:34
→ ianmcorvidae joined09:35
← ianmcorvidae left09:35
→ ianmcorvidae joined09:35
← madsy left09:35
← daethorian left09:35
→ mtkd joined09:35
← avium left09:35
← unreal left09:35
← jjasinski left09:35
← hsb left09:35
← leo left09:35
→ hsb joined09:35
→ jjasinski joined09:35
→ overbythere joined09:35
→ madsy joined09:35
← madsy left09:35
→ madsy joined09:35
← jargon- left09:35
→ leo joined09:35
→ lottadot joined09:35
→ unreal joined09:36
→ xiangfu_ joined09:36
→ jargon- joined09:36
→ daethorian joined09:36
← xiangfu_ left09:36
→ NickASloan joined09:37
← macrover left09:37
→ achadwick joined09:38
→ Kingsy joined09:38
→ johnathanb joined09:39
← ianmcorvidae left09:39
→ ianmcorvidae joined09:39
← caseymcg left09:40
→ avium joined09:40
← shruggar left09:40
→ shruggar1 joined09:40
→ hyperair joined09:40
cmn preserve across what? what commands are you running?09:41
→ eijk_ joined09:41
→ xiangfu joined09:41
→ jamesaxl joined09:42
→ neurodrone joined09:42
← weiyang left09:44
eugenmayer cmn: imagin an alias09:44
cl = ! sh -c "git clone host:path/to/$0"09:44
the issue with this is, that pwd is not preserved, so the clone does not end up to be in the cwd09:45
→ tatsuya_o joined09:45
← eijk left09:45
cmn two or three issues with that09:45
eugenmayer and [url ] is not the same here09:45
← legnaleurc left09:45
cmn such an alias would be better as a insteadOf09:45
→ f0i joined09:45
eugenmayer cmd well, more to write09:46
cmn and you're starting a brand new shell for no good reason09:46
eugenmayer git clone insteadofalias:name09:46
→ noisebleed joined09:46
← noisebleed left09:46
→ noisebleed joined09:46
eugenmayer cmn: wel, afaik thats how you do $0 .. $1 stuff with git aliases09:47
eg09:47
tde = ! sh -c 'git tag 6.x-$0-$1'09:47
→ legnaleurc joined09:47
← glcrazy left09:47
eugenmayer i would love to avoid the new shel, but AFAIK there is no way to work with the args then09:47
cmn you use a function09:47
← xiangfu left09:47
cmn that's how I've always seen it in the docs and the list09:48
eugenmayer function? what exactly is that in git terminolgy?09:48
cmn nothing09:48
what you're doing is shell scripting09:48
eugenmayer basically, yes09:49
cmn not basically, this is shell programming, little to do with git09:49
→ cabaire joined09:49
wereHamster eugenmayer: I believe git exports an env var with the original pwd09:49
cmn allias.cl = f() { git clone whatever/$0} !f()09:49
gitinfo [!ff] A fast-forward is a special type of merge where you have a revision and you are "merging" another branch's changes that happen to be a descendant of what you have. In such these cases, you do not make a new merge commit but instead just update to his revision. This will happen frequently on a remote-tracking branch of a remote repository. (From man gitglossary) See also:09:49
→ xiangfu joined09:49
← talios left09:49
cmn ignore gitinfo09:50
I can't remember the exact syntax, but that's the way people who know write these things09:50
← neurodrone left09:51
← legnaleurc left09:51
← dreiss left09:52
cmn allias.cl = "!f() { git clone $0;}; `f $0`" seems to be the way to go09:52
→ legnaleurc joined09:52
cmn but really, use a insteadOf09:53
eugenmayer http://osiux.com/git-alias#id809:53
there are a lot of those examples09:53
cmn there's lots of examples for all kinds of things; I trust the git hackers more09:54
eugenmayer cmn: the last one does not work out09:54
cmn last what?09:55
heoa How can I install submodules specified by the submodule file?09:55
eugenmayer cl = "!f() { git clone dwdev:dw/$1;}; `f $1`"09:56
cmd ^^09:57
→ jatt joined09:58
eugenmayer the args are not passed, $0 is "sh"09:58
cmn so adjust for the real values09:58
← arvind_khadri left09:59
eugenmayer cmd, tried e.g09:59
cl = "!f() { git clone dwdev:dw/$1;}; `f $0 $1`"09:59
but even then, my string is not passed, when i call09:59
git cl somename10:00
← ajaykools left10:00
eugenmayer $1 is empty10:00
→ JEEB joined10:00
← BiggFREE left10:01
→ Vile joined10:01
← eno left10:02
← senny left10:03
→ eno joined10:04
→ senny joined10:05
wereHamster eugenmayer: cl = "!f() { git clone dwdev:dw/$1;} f"10:05
eugenmayer wereHamster: sh: -c: line 0: syntax error near unexpected token `f'10:06
sh: -c: line 0: `f() { git clone dwdev:dw/$1;} f 'wysiwyg''10:06
wereHamster why did you add the argument to f?10:06
eugenmayer well i called10:06
git cl wysiwyg10:06
(so in the cli, after adjusting the alias your way)10:07
wereHamster eugenmayer: git config --global alias.test '!f() { echo $1 $2 $3 $4; }; f'10:07
then run `git test arg1 arg2 arg4`10:07
see what it does?10:07
eugenmayer git test t1 t2 t310:07
t1 t2 t310:07
so, w orks10:07
wereHamster bravo!10:07
so, change echo to git clone or whatever you want10:08
eugenmayer wereHamster: you forgot the ; in the command.10:08
Kingsy guys, if I have a blank repo and I fetch from a --bare repo, it gets the objects but git log says fatal: bad default revision 'HEAD' and error: pathspec 'master' did not match any file(s) known to git10:08
wereHamster no I did not10:08
eugenmayer wereHamster: wait, no offence, was just FYI10:08
Kingsy is what it says when I try to checkout the master branch..10:08
eugenmayer wereHamster: eugenmayer: cl = "!f() { git clone dwdev:dw/$1;} f"10:08
→ tcurdt joined10:08
eugenmayer cl = "!f() { git clone dwdev:dw/$1;}; f"10:08
wereHamster ah, yes there10:09
eugenmayer wereHamster: should it be that way?10:09
ok, fine. Just bein sure to not miss something.10:09
Kingsy haha probably because git branch and git branch -av displays nothing, so I am assuming there isnt even a branch made.. which is confusing.. where did it sae the objects?10:09
eugenmayer wereHamster: cmn Thank you a lot!10:09
git is awesome.10:09
→ iocor joined10:10
FauxFaux Kingsy: Please slow down, learn something about git, and stop trying to do crazy things, then trying to get us to help you dig yourself out of whatever hole you've ended up in, it's saddening.10:10
Kingsy FauxFaux: oh ok, sorry I assumed what I was trying to do was pretty simple...10:10
eugenmayer wereHamster: :/ even then, the command seems ot fail for pwd. But FYI..iam in a "repo" were a folder is ignore, trying to checkout a repo there…its ends up to be in the "parents" root10:11
FauxFaux You haven't explained what you're trying to do or how you got into the situation you're in, but it's certainly not simple or the recommended way. Again: Return to start, read documentation, explain problem.10:11
wereHamster eugenmayer: all aliases are executed in the top level directory of the repo10:11
eugenmayer i see, no i get it10:11
So well, my real issue here is, not using submodules then10:12
→ arvind_khadri joined10:12
← arvind_khadri left10:12
→ arvind_khadri joined10:12
Kingsy FauxFaux: all I am trying to do is fetch from a bare repo into a blank repo10:12
→ ajaykools joined10:12
Kingsy thats it10:12
eugenmayer I avoided submodules since i had a bad experience with them 1.5 years ago…implemented a fork of braid with gitignore then..using this approach for a while10:12
FauxFaux Kingsy: You can't fetch an empty bare repo. It doesn't make any sense. There's nothing to fetch.10:12
Kingsy FauxFaux: lol no, the bare repo has stuff in it, I am just saying when I created it, I used --bare10:13
eugenmayer are submodules considered good practise for now, with nested repos? I general, its not what i would call "external" repo.. i dont need ot let the parent now the "submodules" at all. its basically a "Framework" as a parent..the "modules ( extenstions" ) are in subfolder10:13
← fruight left10:13
→ kombi joined10:13
→ d0k joined10:14
eugenmayer Kingsy: http://progit.org/book/10:14
kombi scared newbee: just rolled back to previous sha, how do I keep that version for real and get rid of "(no branch)"10:14
eugenmayer Kingsy: before you upset people in here even more, not getting help with the "tricky" stuff later anymore…you start there10:15
← psoo left10:15
Kingsy eugenmayer: ok np.. thanks10:15
wereHamster eugenmayer: git --version ?10:15
eugenmayer Kingsy: i would state, that this docs are by far the best written ones i have ever seen for a OS project in general10:15
Kingsy: its a book, a really good one.10:15
wereHamster: 1.7.4.4 / 1.7.5 on the severs10:16
(cli is OSX)10:16
wereHamster eugenmayer: update to 1.7.6+ and you'll have GIT_PREFIX10:16
eugenmayer basically, the servers are the limit, but i can upgrade to what ever needed10:16
wereHamster eugenmayer: read the 1.7.6 release notes10:16
kombi whoops, don't mean to upset anyone with a trivial task, it's just that it's a production website one day before launch..10:16
eugenmayer wereHamster: you are a hero.10:17
eugenmayer going to read10:17
← xerxas left10:18
Kingsy eugenmayer: point taken, thanks10:18
eugenmayer wereHamster: are you referencing to10:18
Processes spawned by “[alias] <name> = !process” in the configuration10:18
can inspect GIT_PREFIX environment variable to learn where in the10:18
working tree the original command was invoked.10:18
?10:18
← smw left10:19
wereHamster eugenmayer: yes, that's what I mean by GIT_PREFIX10:19
or do you see any other reference to GIT_PREFIX?10:19
→ psoo joined10:20
eugenmayer that would let my alias work i guess. Any comments on that "submodules or not" ?10:20
wereHamster I didn't follow that thread10:20
eugenmayer Because actually, my appraoch breaks any IDE integration of git yet ( PHPstorm for me here )10:20
wereHamster: should i rephrase in one sentence, or is you bandwith already limited? :)10:21
← ethanol left10:21
→ Sigma joined10:21
eugenmayer notices being pretty noisy in the channel. sorry10:22
→ Element9_work joined10:22
→ Ryback_ joined10:24
← xerora left10:24
wereHamster I just didn't pay attention to your other problems, that's all10:24
← Sitethief left10:26
glguy_ → glguy10:26
kombi how do I turn "no branch" into the new master (rather than merge it)?10:26
→ shruggar joined10:27
eugenmayer wereHamster: 2 or more years ago i tried submodules and didnt like the result at all. I have seen a lot of development on that field again and again, but did not retry it. Yet, my usecase might be even non-fitting. I dont need the parent project to actually no its submodules for "all clones", rather, if at all, only for the current clone ( localy ). The reason is, that the parent is more a "Framework", while the submodules are "modules"10:27
→ Nathandim joined10:27
→ fruight joined10:27
eugenmayer wereHamster: for the meantime, i forked braid and implemented a "managed .gitignore" with well, the same results you get with submodules, execpt some fixes ( compared to the old releases) and some extra stuff like email-report ( which were back then rather maintainance infos…if you try to pull all repos and you have a merge-request and fail..and things like that)10:28
iam not any longer using git for deployment, but of course for development, in the case stated above10:28
← hyperair left10:28
eugenmayer hopes to not be to verbosed10:28
← shruggar1 left10:29
← gxk left10:29
→ shruggar1 joined10:29
← shruggar left10:30
wereHamster eugenmayer: there has not been any significant development in the submodules10:30
jast eugenmayer: there are several approaches to !subprojects (see below). perhaps another will appeal to you more...10:30
gitinfo eugenmayer: 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.10:30
wereHamster maybe you want one of the alternatives !submodule10:30
gitinfo git-submodule is ideal to add subsidiary git repositories to a git superproject when you do not control the subprojects or more specifically wish to fix the subproject at a specific revision even as the subproject changes upstream. See http://book.git-scm.com/5_submodules.html10:30
wereHamster ah, what jast said10:30
→ gxk joined10:31
eugenmayer 1. subtree is no choice at all 2. submodule seems to not fit the case 3. gitignore breaks like any IDE / UI integreation for the subprepos ( while submodule seems not )10:32
so my last be are git-slaves10:32
→ Demego joined10:32
eugenmayer !gitslave10:32
gitinfo gitslave (http://gitslave.sf.net) is useful to add subsidiary git repositories to a git superproject when you control and develop on the subprojects at more or less the same time as the superproject, and furthermore when you typically want to tag, branch, push, pull, etc. all repositories at the same time.10:32
kombi I don't mean to muscle in on one of you but how do I get rid of that darned 'no branch'?10:34
eugenmayer kombi: i simply have no clue :)10:34
kombi ;)10:34
nru !subtree10:35
gitinfo The git subtree merge method is ideal to incorporate a subsidiary git repositories directly in to single git repository with "unified" git history, where you only need to pull changes in from external sources not contribute your own changes back (which if technically possible is at least difficult). See http://progit.org/book/ch6-7.html Type "!subtree_alt" for more options10:35
eugenmayer well, eventhough git-slaves looks interesting, i guess it will need a test how the IDEs play with it10:35
← dc5ala left10:35
eugenmayer nru: subtree is even a more tight "relation" from subrepo to parentrepo…in my case, that is even worse then submodules then10:36
→ beautifulmind joined10:36
→ pro_metedor joined10:36
eugenmayer nru: basically, parent and submodules are _completly_ seperated and no "history" of what version of sub1 has been when parent was in rev 12310:36
basically, all i need is . "hey could you pleae play nice when nested - you do not need to interact with eachother"10:37
kombi eugenmayer: a priceless answer the usual if(clueless){'why do you even want to do that?'}s10:37
eugenmayer gitignore kind of works..but well, with limitations stated above10:37
kombi *given10:37
nru eugenmayer: yeah, I just wanted to read the intro text for myself10:38
eugenmayer kombi: he, i hate those kind of questions in general10:38
nru: oh - sorry10:38
( i though you were referencing )10:38
kombi eugenmayer: same here;)10:38
→ plainas joined10:38
plainas hey all10:38
eugenmayer Heyho10:39
→ angelsl joined10:39
plainas I've installed git on windows, I'm assuming it installs ssh too?10:39
← maletor left10:39
jast kombi: let me guess, you checked out a tag or a commit directly at some point?10:40
eugenmayer plainas: yes, and then replaces your OS with a proper OS. Sure10:40
kombi jast: yip...10:40
jast: now I got (no branch) and master and would like to get rid of the former10:40
jast kombi: just checkout one of your branches. be aware that any commits you made on top of "no branch" will stop being referenced by anything if you do that. if you wanted to keep them, create a branch from the current state first.10:40
plainas where are my private keys? I need to transport them to other computer10:41
jast if you need to, you can merge or cherry-pick the commits to master10:41
EugeneKay plainas - msysgit includes a ssh-alike, but I recommend usage of plink in combination with the most-excellent Pageant and PuTTY10:41
jast plainas: ~/.ssh10:41
nru plainas: I recommend just installing Cygwin10:41
jast no idea what ~ is on msysgit, but "echo $HOME" in the git shell should tell you10:41
EugeneKay I replace msysgit's ssh.exe with plink.exe, and it 'just works'10:41
eugenmayer EugeneKay: well, it does not work as similar to what you want to use with former ssh git integrations like ssh aliases and all that10:41
nru so you don't have to bother about windows anymore10:41
jast eww10:42
using cygwin git with windows apps is just wrong :}10:42
EugeneKay jast - ~ on Windows correpsonds to %userprofile%, which is c:\users\USER\ nowadays10:42
plainas this is more related to ssh than git, but how do I use more than one key on the same computer?10:42
EugeneKay cygwin is a crime and should be burned10:42
nru jast: why?10:42
jast EugeneKay: yeah well, ISTR that (for example) cygwin does it differently10:42
EugeneKay plainas - magic :-p10:42
jast - yes, in many horrible and nasty ways10:42
eugenmayer EugeneKay: here you go..thats what i ment.10:42
MacGyverNL I see git for windows problems.10:42
jast nru: for one thing, cygwin git is built to assume that the native line ending is LF10:42
plainas eugenmayer: i can't?10:42
EugeneKay plainas - I actually dunno how to do multi-key with Pageant... I've never tried :v10:42
I know theoretically how-to with OpenSSH, but never done it in practice10:43
MacGyverNL plainas: Git for windows uses its own ssh executable, the ssh keys are in your home folder (C:\Users\plainas) in the .ssh subfolder, same structure as on linux.10:43
nru (sorry for the off-topic) 11:42:28 <+EugeneKay> cygwin is a crime and should be burned <- any pointers to explainations about this statement?10:43
eugenmayer there is a way, it feels like "just hack here, hack there, cross fingers, pray, hack here, pray" might work for 5%10:43
plainas I actually want to do that on linux, this is not completely relation to my other question10:43
MacGyverNL plainas: If you use putty SSH keys, however, you will need to convert them first.10:43
eugenmayer nru: iam with you. For anything you really wont to do under winos, while rally needing *nix libs, you have to use cygwin10:44
plainas how do I use multiple keys on a linux system10:44
EugeneKay nru - the package manager is bad to terrible, the posix emulation sucks, it breaks lots of standard windows stuff, the git client fails hard for a lot of things(usually UTF-related)10:44
I could go on, but I'm out of booze10:44
eugenmayer plainas: this is #git, nnot #ssh10:44
plainas: your buz word would be "ssh-alias" anyhow10:45
EugeneKay plainas - in a (very tiny) nutshell, as regards git: http://superuser.com/questions/272465/using-multiple-ssh-public-keys10:45
For more info, you really should `man ssh`10:45
kombi jast: thanks! Now I've got 'newBranch' and 'master'. When merging, how would I make sure 'newBranch' completely wins?10:46
← eugenmayer left10:46
→ eugenmayer joined10:46
EugeneKay kombi - man git-merge, look for "ours"10:46
gitinfo the git-merge manpage is available at http://jk.gs/git-merge.html10:46
← shruggar1 left10:46
kombi thx!10:46
← zraalb left10:46
eugenmayer guys, i want 1.5 git to be blacklisted! :)10:46
it drives me crazy on debian5… git checkou —track *sighs*10:47
→ artib joined10:47
EugeneKay Something something "unstable"10:48
eugenmayer but on the other side, its outstanding how much work goes into git10:49
it has developed to a "i can do anything" tool. It is used for so many things not being SCM related..like deployment, storage even backups..because its good in so many way.10:49
← fruight left10:49
← Skyx-mobile__ left10:50
EugeneKay git, at it's core, is a file history model. The porcelain VCS can be adapted to do deploy/backups, but it takes ork and is far from perfected10:51
→ Skyx-mobile__ joined10:51
← babilen left10:51
eugenmayer EugeneKay: well, i dont ment to replace Bacula with git, its just for specific tasks10:51
→ shruggar joined10:51
eugenmayer EugeneKay: people use it as a sftp / ftp replacement for customers to deploy code on dev-servers and all kind of things ( which works out pretty good and clever )10:52
EugeneKay people also use crack coaine :-p10:52
eugenmayer it just scales, its so good with hooks. I really like one10:52
EugeneKay: well, you have to go for "compromises" when you have "customers"10:52
"FTP or git…lemme think"10:52
← gjer984 left10:52
→ berserkr joined10:53
EugeneKay Customers are evil and should be burned at the stake10:53
eugenmayer I guess, there is nothing to add.10:53
EugeneKay With their unrealistic "uptime" expectations and demands for "features"10:53
→ Chappo joined10:54
eugenmayer Deadline? Well…what about yesterday, OR if no other way, today…and excalation..tomorrow10:54
"But you had 500k gazzilion of change requests".. "What are change requests..i cant remember" :)10:55
← Wombert left10:55
← forsto left10:56
→ lorian joined10:56
← legnaleurc left10:56
→ gjer984 joined10:58
→ flyz joined11:00
← kombi left11:00
pro_metedor Your branch is behind xxx by 1 commit, and can be fast-forwarded.11:01
fast-forwarded to what ?11:01
→ dr_win joined11:02
shruggar pro_metedor: fast-forwarded to be exactly xxx11:02
pro_metedor: it means "You based your branch on xxx. Since then, xxx has moved on. You can base your work on the current state of xxx without needing to merge, if you want"11:04
← Element9_work left11:04
pro_metedor so if I pull this, I will be up to date?11:05
shruggar you can pull with --ff-only11:05
pro_metedor what ff will do ?11:06
shruggar fast-forward. "fast forward" means "if I do a reset --keep to X, no commits will be left behind. So I'll do that instead of doing an actual merge"11:07
→ TTSDA joined11:08
shruggar --ff-only means "do a fast-forward if possible, otherwise do nothing"11:08
← zdk left11:08
→ fruight joined11:10
→ kombi joined11:12
← fruight left11:15
← artib left11:18
→ artib joined11:18
← amigojapan left11:20
→ isaak joined11:22
← eno left11:22
← jony left11:23
→ eno joined11:24
→ babilen joined11:24
→ hyperair joined11:25
→ cortana2 joined11:27
shruggar Is there a good way to programmatically determine whether a rebase was completed vs requiring more input? eg: my script says "git rebase master", afterwards I either want to continue with the next action, or drop into a shell to complete the rebase11:27
→ Xizor joined11:28
← OOPMan left11:28
cmn shruggar: if you look at how the bash __ps1 does it, you can figure out whether git is still in the middle of a rebase11:28
Mikachu doesn't the command return 1 if it aborts?11:29
← eno left11:30
→ eno joined11:30
eugenmayer is it possible to find out, which on which branch a tag was created?11:31
→ Element9_work joined11:31
→ Wombert joined11:31
← iocor left11:32
eugenmayer Iam in charge, this question is stupid, due git refs are refs on commits..but well, isnt it still a graph? and if a tag is not a ref a root of a branch, but rather a unmerged node (comparing to master), that tag "is on a specific branch", is it that waY?11:32
← thiago left11:32
→ akosikeno joined11:32
← i42n left11:33
shruggar eugenmayer: refs exist outside of the graph11:34
eugenmayer My case: iam using a make file to crreate a project, which basically consist out of several repos. While building, the tag-refs are used referenced in the make file..but later, when trying to go for changes, i need to somehow checkout the branch of that tag, which will be the most "possible" one…or..well i branch off11:34
→ cccaldas joined11:34
← artib left11:34
→ artib joined11:34
eugenmayer shruggar: they are pointers, i know11:34
cmn git branch --contains $(git rev-parse mytag)11:34
eugenmayer shruggar: but if we look at the graph as a directed timeline, a pointer is not a pointer to a "level of deepness", but rather one node, isnt it that way ?11:35
← Dreamer3__ left11:35
cmn a tag points a commit, it doesn't say which branch it's contained in because that changes with time11:35
eugenmayer cmn ➜ drupalwiki_user git branch --contains $(git rev-parse 6.x-5.0-beta3)11:36
* (no branch)11:36
not working11:36
cmn is it a lightweight tag?11:36
eugenmayer cmn yes, thats clear. What you suggested is what iam looking for11:36
its clear its not part of the "tag meta informations"11:37
cmn, never heard that term11:37
cmn git show $tag, what does it say?11:37
eugenmayer git show-ref | grep 6.x-5.0-beta311:37
fe457cdb804c636f0590a220a1051a57d030e6d0 refs/tags/6.x-5.0-beta311:37
cmn that's not what I said11:38
eugenmayer cmn it shows me a diff11:38
cmn does it just show the commit or does it include some text inside the tag?11:38
← dnjaramba left11:38
cmn a lightweight tag is a ref to a commit that lives in refs/tags11:39
→ simenbrekken joined11:39
eugenmayer cmn sorry, cant follow. Will not be able to pastebin you all the output. Do you need something of the n10 lines?11:39
cmn a real tag is a tag object11:39
cat-file -t fe457cdb804c636f0590a220a1051a57d030e6d011:39
what's that say?11:39
simenbrekken Hello, my 'git branch -r' shows origin/HEAD -> origin/refactor which doesn't exist, how do I change what origin/HEAD points to?11:39
← kombi left11:39
← ChaoWang left11:40
eugenmayer cmn: http://pastie.org/3347553 ..after that, the diff follows11:40
cmn then it's a lightweight tag11:40
→ g105b joined11:40
cmn so that command should have worked11:40
eugenmayer cmn: never heard of11:40
hm11:40
cmn man git tag surely tells you about them11:41
gitinfo the git-tag manpage is available at http://jk.gs/git-tag.html11:41
eugenmayer git version 1.7.4.411:41
← g105b left11:41
cmn I've never seen branch --contains return anything like * (no branch)11:41
cbreak-work detached HEAD? :)11:41
cmn yeah11:42
← tcurdt left11:42
→ tcurdt_ joined11:42
→ dnjaramba joined11:42
cmn that's what I thought11:42
but the star makes no sense11:42
eugenmayer cmn: sell sure it is detached11:42
EugeneKay The Legend of Sleepy Hollow strikes again11:42
eugenmayer git branch --contains fe457cdb804c636f0590a220a1051a57d030e6d011:42
* (no branch)11:42
(just tried it with ths SHA directly)11:42
cmn so that commit isn't in a branch11:42
eugenmayer it is :)11:43
cmn not one that branch sees11:43
add -a11:43
← csherin left11:43
eugenmayer --11:43
git show-ref | grep fe457cdb804c636f0590a220a1051a57d030e6d011:43
fe457cdb804c636f0590a220a1051a57d030e6d0 refs/remotes/origin/spaces11:43
fe457cdb804c636f0590a220a1051a57d030e6d0 refs/tags/6.x-5.0-beta311:43
cmn neither of those are local tags11:43
eugenmayer ok, i see, origin..11:43
cmn yes11:43
hence my "add -a"11:44
eugenmayer i see11:44
ok, so i might go the —mirror trick again, to get all refs11:45
cmn or use -a11:45
and detect whether it's a remote-tracking banch and hack off the beginning of the name, which you're going to do anyway11:45
eugenmayer (git clone —mirror path/.git && cd path && git config core.bare false && git checkout TAG)11:45
← dnjaramba left11:46
← tcurdt_ left11:46
→ tcurdt joined11:46
eugenmayer its the only way i know to "fetch all refs"11:46
cmn checking out a tag isn't going to help11:46
→ rendar_ joined11:46
cmn fetch && fetch --tags fetches all you need11:46
eugenmayer cmn: yes, at lest they will be local then ( i get our -a idea, its working ..but to have them local )11:46
cmn then do a checkout11:46
← twisted` left11:46
cmn why do they need to be local?11:47
→ twisted` joined11:47
← rendar left11:47
eugenmayer cmn: to simplify scrits ( the only reason)11:47
cmn that's not a need11:47
→ dnjaramba joined11:47
eugenmayer hot to get the "current" sha of the repo? ( current rev ) to be used with —contains?11:48
cmn define "current sha of the repo"11:48
eugenmayer ah. simple11:48
git branch -a --contains11:48
does the same for "the current commit of your wc"11:48
cmn: iam sure my terminology can confuse you - iam not mastering git to well yet, sorry11:49
cmn wc doesn't exist in git; were you asking to see which branch contains the current HEAD?11:49
→ csherin joined11:49
eugenmayer cmn: yes, sorry. Exactly11:49
cmn: or no, your question was the answer to my question above. I just simplified the case to the real example then, were the tag is already HEAD and i know want to change head to the branche of the "tag"11:50
cmn: so it is all my bad11:50
(tob be expected :) )11:50
→ brian_g joined11:50
eugenmayer _thank you_11:51
← no1home left11:51
sameer_ :)11:51
→ canton7 joined11:52
dwmw2_gone → dwmw211:52
← dnjaramba left11:52
→ krz joined11:53
← plainas left11:53
← eno left11:54
→ eno joined11:55
← dr_lepper left11:56
← Wombert left11:56
→ zavierlo joined11:57
← pro_metedor left11:58
→ Dave^|| joined11:59
← zavierlo left11:59
→ fruight joined12:03
← Dave^| left12:03
← psakrii left12:03
← eno left12:05
→ eno joined12:07
→ ph^ joined12:07
→ legnaleurc joined12:08
← artib left12:10
→ fayimora joined12:10
→ alberto56 joined12:16
→ johnkpaul joined12:16
→ legnaleurc_ joined12:16
← legnaleurc left12:17
→ robertstuttaford joined12:19
← eno left12:20
robertstuttaford context: so i'm using jenkins-ci and github. two jenkins jobs, staging and production. staging uses branch develop, production uses branch master. however, github pings both jobs when pushes are received to either branch.12:20
question: i want to know if i can tell _after_ the git fetch that jenkins does whether any actual changes came down12:20
← legnaleurc_ left12:20
robertstuttaford if so, how? i can use this info to make the deploy script stop working if there werent any changes.12:21
← johnkpaul left12:21
→ eno joined12:21
cmn jenkins should already be noticing that12:21
→ oriba joined12:21
robertstuttaford it does show 0 changes, yes. but i don't know how to give that information to my deployment script12:21
cmn don't call your deployment script if there aren't any changes12:22
robertstuttaford i'd love to know how to do that12:22
← harshpb left12:23
← heptadecagram left12:23
← juvenal left12:23
cmn it's avoiding doing something... just don't do it12:23
robertstuttaford i don't know how to configure the jenkins job to not call my deploy script when no changes are found12:24
cmn then you probably want to ask in a jenkins forum12:24
← Gitzilla left12:24
← netzapper left12:24
robertstuttaford i'll ask in the right channel12:24
purely for interest, is there a way to detect this with git?12:24
→ Chib joined12:25
→ legnaleurc joined12:25
cmn kinda12:26
you can store the old value somewhere and then compare against the new value12:26
robertstuttaford 'value' being the commit hash for HEAD?12:27
cmn for whatever branch you're interested in12:28
it's probably HEAD12:28
robertstuttaford gotcha12:28
→ linusoleander joined12:30
← Vortex34 left12:30
→ Fandekasp joined12:30
← s_segawa left12:31
← xiangfu left12:33
← eno left12:34
← break57382 left12:34
→ digitalfiz joined12:35
→ hotsyk joined12:36
→ eno joined12:36
← legnaleurc left12:36
→ aristidesfl joined12:36
→ Gitzilla joined12:36
← dougsland left12:37
← Skyx-mobile__ left12:37
← T_X left12:38
← robertstuttaford left12:39
→ xiangfu joined12:39
← arvind_khadri left12:39
magn0z_ can i display all local branches that has not yet been merged to my current branch (master)?12:39
← ajaykools left12:40
→ legnaleurc joined12:40
← papegaaij left12:41
→ DarkAR joined12:41
← jaisoares left12:42
→ tiglog joined12:42
→ dcarper joined12:42
→ Vortex34 joined12:43
← ehsan left12:44
→ i42n joined12:44
→ codewrangler joined12:44
→ silv3r_m00n joined12:45
silv3r_m00n I did a git add . & git commit -am "done" , is it wrong to do like that ? 2 commands at the same time12:45
→ T_X joined12:45
← T_X left12:45
→ T_X joined12:45
→ leavittx joined12:46
← stringoO left12:47
→ antgel_ joined12:47
← jamesaxl left12:47
→ jamesaxl joined12:47
← eno left12:48
← silv3r_m00n left12:49
cbreak-work silv3r_m00n: don't.12:49
that looks like it could provoke race conditions12:49
gitinfo set mode: +v12:49
antgel_ hi all. i've set up a git server on a past project, i used gitolite (or gitosis, can't remember), and it worked great. on my new project, most developers use windows, and there are no linux servers available. the easiest network resource for me to get hold of is a windows share. is there any way to set up a shared git repository with such limited access?12:50
gitinfo set mode: +v12:50
→ eno joined12:50
antgel_ i don't even have ability to set up a web server on the network12:50
FauxFaux antgel_: You can push and pull to a windows share, yes, but I'm not sure you can implement security in any sane way in that kind of environment.12:50
← teweWORK left12:50
→ teweWORK joined12:51
antgel_ FauxFaux: so i'd have problems with access control etc?12:51
FauxFaux Yes.12:51
cbreak-work antgel_: using a git repository on a windows share is asking for trouble12:51
antgel_ FauxFaux: okay, thanks. makes sense. having said that, we *can* enable and disable shared folder permissions. so i could put \\someserver\ourgroup\git under permissions whereby only our group could access it, at a windows network level12:52
cbreak-work: so i see :(12:52
cbreak-work you can just run a unix in a vm or so12:52
→ qwertz_ joined12:52
antgel_ this is awful. someone else is recommending that we use perforce12:52
→ erichynds joined12:52
cbreak-work why?12:52
does it offer security on windows shares? :)12:53
→ shiba_yu36 joined12:53
antgel_ cbreak-work: because he's used it before, it's "user-friendly" and he doesn't get it that version control can be metadata on his existing filesystem rather than some monolithic blob12:53
→ dougsland joined12:53
← dougsland left12:53
→ dougsland joined12:53
→ bitkiller joined12:54
FauxFaux antgel_: Theoretically it's possible to start cygwin ssh server and have things run under that, but it's really terrible and you'd probably rather gouge your eyes out (although I wouldn't go as far as using perforce); if that's an option?12:54
← dwmw2 left12:56
antgel_ i don't have any machines that i have access to that i can run as a server. i'm in a pretty locked-down massive windows-centric LAN environment12:56
i might see if i can get a machine here and just install linux on it12:57
cbreak-work antgel_: perforce doesn't need a server?12:57
antgel_ cbreak-work: provided by corporate IT dept *bangs head against wall*12:57
→ Dreamer3 joined12:57
cbreak-work and they can't provide a git server? Or a unix vm?12:57
antgel_ and my colleague thinks it's a *good thing* that they can "support" it12:57
cbreak-work well, it is a good thing12:57
but if the thing they support is terrible then their support might not be the thing to worry about12:58
jast note to self for future reference: making beef stock is quite a bit of work12:58
→ wbednarski joined12:58
shruggar can't you get a server these days for $50 that fits inside an RJ45 connector?12:58
FauxFaux I must say, our corporate-IT hosted SVN server is suprisingly good. That is, it's generally up, which is a miracle as far as I'm concerned. Certainly more often than the issue tracker or wiki are up.12:58
antgel_ cbreak-work: what if their support is terrible and the thing they support is terrible? ;)12:58
i'm just kidding of course12:59
→ dwmw2_gone joined12:59
MacGyverNL Whenever somebody mentions perforce I'm reminded of Torvalds' talk at google.12:59
"so you use... what dó you use here? Perforce? Perforce. I'm.... sorry."13:00
I can't help it, I'm forever spoiled.13:00
antgel_ i emailed the team with a great summary of why to use git and my team member ignored everything and said why not use perforce?13:00
← eno left13:00
shruggar one of the great things about git is that you can use it without a server13:01
;)13:01
→ rchavik joined13:01
← rchavik left13:01
→ rchavik joined13:01
→ eno joined13:02
→ dr_lepper joined13:03
→ sattu94 joined13:04
→ Targen joined13:04
→ _Qman joined13:04
→ tvw joined13:05
← dr_win left13:05
← dermoth|home left13:06
→ jonesy joined13:07
← _Qman left13:10
← Dreamer3 left13:13
→ johnkpaul joined13:13
← cabaire left13:13
← eno left13:13
← dwmw2_gone left13:14
→ eno joined13:15
→ fedesilva joined13:17
→ juvenal joined13:17
← ASIRINeiluj left13:17
→ dwmw2_gone joined13:18
← eugenmayer left13:19
← luxurymode left13:19
→ dermoth|home joined13:20
← Fudge left13:20
→ cabaire joined13:20
→ linkedinyou joined13:20
← berserkr left13:20
← eno left13:23
→ gchristensen joined13:23
→ alexgordon joined13:23
→ eno joined13:24
→ iamjarvo joined13:24
← johnkpaul left13:25
← nixmaniack left13:27
→ MrKeuner joined13:27
MrKeuner hello, any hints as to how to detect different branch commits in order to run checkouts on different GIT_WORK_TREEs?13:28
FauxFaux For what purpose?13:28
MrKeuner if branch dev is committed I'd like to checkout -f to mu dev.domain.name folder13:29
if prod branch is committed checkout -f to domain.name folder13:29
← isaak left13:30
FauxFaux So, by "committed", you mean "pushed"? And you're doing this from a post-recieve or update hook?13:30
MrKeuner sorry, yes pushed13:30
FauxFaux As in, you're using git as a deployment tool.13:30
MrKeuner yes13:30
FauxFaux edges away.13:30
MrKeuner :)13:30
what's wrong with it?13:31
← variable left13:31
← D-Chymera left13:31
← eno left13:32
← gchristensen left13:33
→ eno joined13:34
→ masterkorp joined13:37
masterkorp hello, is it possible to ignore a spefific line of a file ?13:37
FauxFaux Nope.13:37
← DarkAR left13:38
→ arvind_khadri joined13:39
← arvind_khadri left13:39
→ arvind_khadri joined13:39
shruggar is there a !config command for the bot? or !localconfig or something?13:39
gitinfo [!configfiles] the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.13:39
masterkorp FauxFaux: oh:(13:40
← alexgordon left13:40
← masterkorp left13:40
shruggar masterkorp: the above line about config-files was for you13:40
← fayimora left13:40
← eno left13:41
→ fayimora joined13:41
MrKeuner FauxFaux, I just wonder why you edged away13:41
shruggar MrKeuner: it's a common discussion around here, and got somewhat rambly yesterday :)13:42
→ eno joined13:43
MrKeuner :)13:43
anyone from the other side of the ramble around? :)13:43
← EricInBNE left13:43
→ variable joined13:44
→ armin_ joined13:44
→ JC_Yang joined13:44
→ thiago joined13:44
shruggar MrKeuner: "pushing to deploy" is a commonly-misunderstood concept used by those new to git. "using git as a deployment tool" is generally discouraged, though using git as part of a larger deployment solution is generally encouraged. There are differences of opinion about what constitutes a solution which is "good enough"13:44
← arvind_khadri left13:44
MrKeuner shruggar, could you suggest a larger deployment tool using git for versioning part?13:45
→ tcurdt_ joined13:46
JC_Yang hi, all. Is there anyone know why git-scm.com sometimes will put old version on the homepage? I've seen this twice. Now 1.7.6.6 is on it. I can't remember which version it was the first time I saw this situation.13:47
← tcurdt left13:48
tcurdt_ → tcurdt13:48
shruggar MrKeuner: I find them all unsatisfactory, and my own implementation isn't generic enough to share (and I'm too lazy to make it so)13:48
← jacekmigacz left13:49
nru JC_Yang: I guess they put the most recent release13:50
← demet8 left13:50
→ wbednars_ joined13:50
nru and not the last release in the last release branch13:50
← bolt left13:50
MrKeuner shruggar, is deficiency of using git as a deployment tool related to projects with DB portion as well as files? I was also thinking there must some other way to handle that kind of situation13:50
→ neurodrone joined13:50
→ guampa joined13:51
sitaram cbreak-work: what sort of trouble is running a git repo on a windows share?13:51
→ revagomes joined13:51
← eno left13:52
→ cburroughs joined13:52
→ vandal- joined13:52
← wbednarski left13:52
← GameGamer43 left13:52
→ GameGamer43 joined13:52
→ bolt joined13:52
shruggar MrKeuner: that's part of it, yes13:53
→ gchristensen joined13:53
→ eno joined13:54
shruggar MrKeuner: git also doesn't track permissions or directories. And releasing with pure git is not atomic (people can be running a "partially updated" site)13:54
← Targen left13:54
← bartman left13:54
→ ephelion joined13:55
→ capncodewash joined13:56
JC_Yang nru: Thanks. That might make sense. But IMHO they should have some clear statements about the different branches, or that would be misleading. For me, personally, will consider that there might be serious regression in the latest release so they rollback. The site is not enough informative at this point.13:57
→ arvind_khadri joined13:57
← digitalfiz left13:57
← guampa left13:57
→ dnjaramba joined13:58
nru JC_Yang: I also think that should be changed; but it appears to be the current behaviour.13:58
MrKeuner shruggar, any solution that comes closest to a proper deployment?13:58
← Milossh left13:59
← ipalaus left14:00
→ bartman joined14:00
→ dorisabayon joined14:01
← eno left14:01
sitaram jast: can your bot listen on #gitolite also and allow us the same (or different, I dont mind which) set of triggers?14:01
→ Fudge joined14:01
→ lyddonb joined14:01
mpenz sitaram: just replied to your emails14:02
→ digitalfiz joined14:02
→ AaronMT joined14:02
→ eno joined14:02
sitaram mpenz: ok thanks!14:02
will check14:02
mpenz sure no prob, btw congrats14:03
← GodEater_ left14:03
← linusoleander left14:03
knitt1 → knittl14:04
← Fandekasp left14:05
← hotsyk left14:05
→ GodEater joined14:05
← GodEater left14:05
→ GodEater joined14:05
→ crisp joined14:05
jepler hm. Should I expect this to print the commandline shown to invoke the mergetool (it doesn't with 1.7.8.3)? GIT_TRACE=2 git mergetool14:05
→ ephelion_ joined14:06
← ephelion left14:06
ephelion_ → ephelion14:06
← ephelion left14:06
tizzo → tizzo-afk14:08
← oriba left14:08
← ph^ left14:10
← eno left14:11
→ ph^ joined14:11
← teknoid left14:11
→ heptadecagram joined14:12
Kingsy how come when I type git clone git://github.com/paulboone/ticgit.git it just says Initialized empty Git repository in /home/chris/website.git/ticgit/.git/ and waits forever? I am trying to follow that book is all14:12
→ tcurdt_ joined14:12
→ eno joined14:12
← dnjaramba left14:13
→ dnjaramba joined14:14
→ intripoon joined14:14
← dorisabayon left14:14
← tcurdt left14:14
cmn have you tried other hosts?14:15
github is having connectivity issues atm14:15
← marthinal left14:17
← tcurdt_ left14:17
jast sitaram: sorry, the code currently assumes one channel only. but you can download the source and run one yourself. :}14:17
→ tcurdt joined14:17
Kingsy cmn: ah ok14:18
← joshsmith left14:18
→ ceej joined14:18
tizzo-afk → tizzo14:18
sitaram jast: thanks. It's not that bad yet so I will take a rain check on that :)14:18
← stic02 left14:19
→ stic02 joined14:19
→ stuntmachine joined14:20
← Lgb left14:20
← eno left14:20
← satyanash left14:21
→ Fandekasp joined14:21
→ eno joined14:21
→ satyanash joined14:21
→ josesaldana joined14:22
MrKeuner shruggar, I'll have to give git as deployment tool. For a beginner in git, would it be easier to handle two separate repos dev and prod; or two branches in one repo that will checkout to dev if dev branch and prod if prod branch is pushed14:22
← fedesilva left14:22
MrKeuner give git a chance14:22
← juvenal left14:22
gitinfo set mode: +v14:23
tizzo → tizzo-afk14:23
→ Lgb joined14:23
josesaldana Is there a way to do git sash pop stash@{2} for an specific file?14:23
stash*14:23
yrlnry You can git show stash@{2}:path14:24
or git checkout stash@{2} -- path14:24
← i42n left14:24
yrlnry I'm, not sure what "pop" would mean.14:24
→ boombatower joined14:24
← armin_ left14:25
→ guampa joined14:25
→ Wombert joined14:26
PerlJam I'm pretty sure you can't pop a partial stash.14:26
→ juvenal joined14:26
← frogonwheels left14:27
← xiangfu left14:27
← matthiaskrgr left14:28
← eno left14:28
jast that doesn't even make sense14:28
→ herdingcat joined14:28
→ dv310p3r joined14:28
→ dr-evil joined14:29
→ OOPMan joined14:29
← stuntmachine left14:29
→ eno joined14:29
→ flojistik joined14:29
→ totte_ joined14:29
→ FauxFaux_ joined14:30
→ Ferris_ joined14:30
← pdobrogost left14:30
← milk left14:30
← FauxFaux left14:30
← blueyed left14:30
← koivula left14:30
← cmn left14:30
← Pieter left14:30
← mjt left14:30
← totte left14:30
← nru left14:30
← CEnnis91 left14:30
← Ferris left14:30
→ CEnnis91 joined14:30
← vdv left14:30
→ koivula joined14:30
→ Pieter joined14:30
→ matthiaskrgr joined14:30
gitinfo set mode: +v14:30
→ mjt joined14:30
→ cakehero joined14:30
→ astrostl joined14:30
→ blueyed joined14:30
dr-evil Hello all. I have a trouble with gitweb :( It doesn't show me 1 repo. I've checked a lot of stuff, but found no error. I don't know perl :(14:31
→ milk joined14:31
← mishok13 left14:31
→ cmn joined14:32
← cakehero left14:32
FauxFaux_ → FauxFaux14:32
→ iocor joined14:33
cmn you can probably apply part of a stash14:33
and then drop the rest14:34
→ ipalaus joined14:34
dr-evil cmn: hm?14:34
→ ISF joined14:34
cmn what part?14:34
→ jacekmigacz joined14:35
yrlnry I guess you could apply it, then use git chcekout -p or somehing to get rid of parts you didn't want..14:35
→ mishok13 joined14:35
→ v0n joined14:36
cmn yeah, I thought stash let you do that itself14:36
← freimatz left14:36
cmn you have to checkout -p14:36
→ BorX joined14:36
→ stuntmachine joined14:36
→ bartek joined14:36
← eno left14:37
← sattu94 left14:37
← arvind_khadri left14:38
ChanServ set mode: +v14:38
jast ahh. much better.14:38
→ eno joined14:38
→ TeckniX joined14:39
→ t4nkd joined14:40
MrKeuner I'll have to give git a chance as deployment tool. For a beginner in git, would it be easier to handle two separate repos dev and prod; or two branches in one repo that will checkout to dev if dev branch and prod if prod branch is pushed?14:40
cmn git is not a deployment tool14:40
← fayimora left14:41
shruggar MrKeuner: what you just said makes no sense at all. Are you saying that "dev" and "prod" are on the same machine??14:41
MrKeuner shruggar, yes14:42
shruggar in that case, may I ask: wtf?14:42
MrKeuner name.com and dev.name.com14:42
shruggar same machine, same directory?14:42
← flojistik left14:43
MrKeuner shruggar, same machine different directories14:43
cmn MrKeuner: you're thinking of branches in a way that doesn't correspond to git14:43
branches exist in repositories14:43
what you might want to do is deploy the dev and prod branches to different directories in the machine14:43
but that's the job of your deployment system14:44
shruggar how's this for something simple: You're new to git, so don't use it as a deployment tool until you're not new to git14:44
MrKeuner http://toroid.org/ams/git-website-howto I read this14:44
I'd like to implement it with multiple branches and multiple target directories14:44
→ chrisgeorge joined14:45
cmn then you need something much more complicated that that14:46
those scripts are for an extremely simple system14:46
← mfischer left14:46
cmn it doesn't deal with many things14:46
→ linusoleander joined14:46
← LekeFly left14:46
← GrowLinux left14:46
← kevlarman left14:46
← disappearedng left14:46
← linusoleander left14:46
← cabaire left14:46
→ LekeFly joined14:47
← mpenz left14:47
MrKeuner or wold it be easier to have two separate repositories dev and live and I will have two remotes on my local machine copy. and push to dev when needed and push to live when ready14:47
fir31igh1 → fire31ight14:47
← troyt left14:47
cmn that deals with one issue14:47
→ dr_win joined14:48
cmn there are still issues that those scripts don't deal with14:48
→ chrisdev joined14:48
MrKeuner I searched for web deployment tools with no chance. I see there is capistrano but it looks like it is primarily for RoR14:49
cmn !website14:49
gitinfo 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/171423514:49
← thiago left14:49
PerlJam MrKeuner: capistrano is general. it works as long as you're comfortable with Ruby14:49
Kingsy when you are pushing to a server using ssh say ssh://user@domain.com:port/address/to/git/ does that mean that it will be executing all commands as "user" so as long as the dir is writable to "user" your ok ?14:49
cmn look at the second link14:49
Kingsy: you're mostly ok14:49
it depends what you're doing afterward14:50
→ derekpa joined14:50
← tiglog left14:50
→ mpenz joined14:50
→ kevlarman joined14:50
→ GrowLinux joined14:50
→ dvaske joined14:50
Kingsy cmn: hmm my push is getting an error.. remote: error: refusing to update checked out branch: refs/heads/master :S14:50
cmn: I am just following that book see14:50
cmn I doubt that14:50
→ arvind_khadri joined14:50
← arvind_khadri left14:50
→ arvind_khadri joined14:50
Kingsy cmn: well I am following the book with my own example :)14:51
cmn and you forgot to initialise the repo as bare14:51
→ grekko joined14:51
Kingsy cmn: you cant push to a normal repo?14:51
→ nru joined14:51
cmn read the error message14:51
Kingsy its refusing to update a branch that has been checked out14:52
cmn the issue is that you're trying to move the active branch14:52
← eno left14:52
cmn yes14:52
→ hotsyk joined14:52
← Womkes left14:52
cmn because you're trying to move the worktree's base from under it14:52
← rurufufuss left14:52
cmn and git isn't letting you14:52
Kingsy cmn: yeah that make sense.. just for learning its helpful to have a git repo on my server and local machine that I can checkout (i.e it has a work tree)14:53
← gchristensen left14:53
→ eno joined14:53
→ Womkes joined14:53
cmn that sentence has too many possible meanings14:54
it's rarely helpful to have a worktree on the server14:54
canton7 Kingsy, have a bare repo on the server that you push to, then create a clone of that on your server, if you want to have a work tree on the server14:54
→ lacrymology joined14:54
Kingsy np, will do thanks14:54
cmn or even better don't have a worktree on the server, because that's going to cause confusion14:55
→ john2496 joined14:56
→ kennethr_ joined14:56
← kennethr_ left14:56
→ hotsyk_ joined14:56
→ cakehero joined14:57
→ Fangorn_ joined14:57
← kennethreitz left14:58
← iocor left14:58
← ipalaus left14:58
→ n8o-mba joined14:58
← kaste left14:59
← hotsyk left14:59
→ marthinal joined14:59
→ iocor joined14:59
→ oo_miguel joined15:00
← cooloney left15:00
→ fedesilva joined15:00
→ mfischer joined15:00
oo_miguel hi15:01
← Cyanure left15:01
← jeffhung_ left15:01
→ remote joined15:01
→ joshsmith joined15:02
← eno left15:05
→ spiette joined15:05
→ diogogmt joined15:05
→ Jackneill joined15:06
← Jackneill left15:06
→ Jackneill joined15:06
← dvaske left15:06
→ eno joined15:07
← jordanm left15:08
← JohnBat26 left15:08
← diogogmt left15:08
→ Skyx-mobile__ joined15:08
→ beatak joined15:10
→ tty234 joined15:10
← fedesilva left15:10
← beautifulmind left15:10
cmn !hi15:10
gitinfo [!welcome] Welcome to #git, a place full of helpful gits. If you have a question, feel free to just go ahead and ask—somebody should answer shortly. For more info on this channel, see http://jk.gs/git/ Take backups (type !backup to learn how) before taking advice.15:10
← crisp left15:11
shruggar pfft, backups before advice.. how un-git-like15:11
→ Dreamer3 joined15:11
shruggar doesn't the bot know about the reflog? :)15:11
→ crisp joined15:11
FauxFaux shruggar: I'm having a problem with git gc --expire=now!15:11
→ _obiJuan joined15:12
cmn is it keeping too much data?15:13
→ duckxx joined15:13
→ gchristensen joined15:13
shruggar git config gc.reflogexpire never ?:D15:13
→ fly91 joined15:13
thebaddies → Mubarak15:14
shruggar I don't add many large binaries that I later change my mind about15:14
→ tjaalton joined15:14
← kermit left15:14
→ jordanm joined15:14
← chrisgeorge left15:15
← eno left15:15
→ chrisgeorge joined15:15
← chrisgeorge left15:15
→ chrisgeorge joined15:16
→ eno joined15:16
Mikachu it's hard for git to recover from rm -rf .git15:16
→ troyt joined15:16
shruggar Mikachu: do you plan on giving rm -rf .git as advice?15:17
FauxFaux I've resolved problems by re-cloning many times.15:17
tjaalton I've a strange issue where git fails to pull tags from a remote15:17
well, a certain tag15:17
FauxFaux tjaalton: git ls-remote15:17
tjaalton FauxFaux: what about that?15:18
→ fayimora joined15:18
← cakehero left15:18
← binjured left15:19
→ berserkr joined15:19
→ jmcantrell joined15:19
FauxFaux It will allow you to provide a not completely useless question for us to have some hope of answering.15:19
nru 16:16:48 < Mikachu> it's hard for git to recover from rm -rf .git <- lol15:19
← satyanash left15:20
Mikachu "pull tags" is sort of nonsense too, pull doesn't update tags, it just merges a particular branch15:20
tjaalton FauxFaux: ok. how about this one; git clone git://git.debian.org/git/collab-maint/pkg-sssd.git; cd pkg-sssd; git remote add upstream git://git.fedorahosted.org/git/sssd.git; git remote update15:21
compared to 'git clone git://git.fedorahosted.org/git/sssd.git'15:21
cmn fetch --tags if you want to fetch tags15:21
(that aren't auto-follow)15:21
← iocor left15:21
tjaalton actually, it's one of the branches that isn't updated either15:22
sssd-1-815:22
→ mp_ joined15:22
cmn git branch -a?15:23
what's that say?15:23
← Skyx-mobile__ left15:23
→ satyanash joined15:23
tjaalton umm, it shows all the branches, looks ok15:23
cmn right then15:23
← eno left15:23
FauxFaux tjaalton: http://pastie.org/3348663 What's wrong?15:24
tjaalton FauxFaux: you don't have sssd-1_8_0_beta215:24
← rchavik left15:24
FauxFaux tjaalton: Interesting. As cmn said, though, fetch upstream --tags fixes it.15:25
→ eno joined15:25
→ rchavik joined15:25
tjaalton FauxFaux: ah.. cool, thanks15:25
FauxFaux What does auto-follow mean, cmn?15:25
← p3rror left15:25
cmn if you're downloading a tagged object, you'll also download the tag15:26
← gnufied left15:26
cmn that's the default behaviour for git15:26
→ duckx joined15:27
→ VladGh joined15:27
FauxFaux I don't follow why that would mean that you wouldn't fetch some tags, though?15:27
tjaalton hmm ok, I think I know why the tag wasn't fetched15:27
← duckxx left15:28
tjaalton it should be from the sssd-1-8 branch, but isn't15:28
cmn if the tag points to some object that isn't covered by the branches you download, you won't get it15:28
tjaalton yep15:28
makes sense now15:29
didn't notice that it doesn't match what's on the branch15:29
I'll let upstream explain :)15:29
FauxFaux Heh, I hadn't noticed that. I leave tagged bits of history lying around all the time (not even remotely on branches).15:30
→ Francisc joined15:30
← _ikke_ left15:31
← Chib left15:31
Francisc Say I have 3 commits, c1, c2 and c3, with HEAD pointing to c3. Running "git reset HEAD~2" will move HEAD to c1. What happens to c2 and c3. Are they removed?15:32
FauxFaux Francisc: They become unreachable and will eventually (after a few weeks) be garbage collectable, unless you make a branch or tag or etc. from them.15:32
Francisc I know the Working Directory does not get touched (unless using --hard).15:32
I see.15:33
← angelsl left15:33
Francisc Is it possible to checkout c3 for example if that time has not passed?15:33
← eno left15:33
shruggar Fracisc: of course.15:33
cmn checkout c315:33
← boombatower left15:33
Francisc Yeah, but what if I don't know the name?15:33
→ b0fh_ua joined15:33
cmn if you don't remember what c3 stands for, look at the reflog15:33
Francisc How can I list them.15:33
shruggar Francisc: git reflog15:33
FauxFaux !gka15:33
gitinfo For a better way to view the reflog, try: gka() { gitk --all $(git log -g --format="%h" -50) "$@"; }; gka15:33
Francisc Aha.15:33
OK. Thank you guys.15:34
→ dp_ joined15:34
b0fh_ua Hi all! I have several local branches, how can I see what are the latest commit messages in each of them?15:34
Francisc The removal of squashed commits is done automatically?15:34
→ orospakr joined15:34
b0fh_ua without checking out each of them15:34
Francisc I'm talking about a local (not remote) repository.15:34
→ iocor joined15:34
→ thiago joined15:34
← v0n left15:34
→ macmartine joined15:35
cmn discarded commits will be garbage-collected15:35
← dp_ left15:35
Francisc I got that, but how?15:35
cmn b0fh_ua: git branch -v15:35
Francisc When does GC run?15:35
← SPYGAME left15:35
→ eno joined15:35
cmn every once in a while15:35
and when you tell it to15:35
Francisc There's no system service that checks that. So it has to be triggered by something else.15:35
FauxFaux Francisc: At the end of some commands, sometimes. And it doesn't touch things in the reflog, which is automatically purged after a few weeks.15:35
Francisc Like when I commit next time?15:35
Aha.15:35
OK.15:35
Thank you all.15:35
→ airborn joined15:36
← chrisdev left15:36
→ _iron joined15:36
← mishok13 left15:37
← duckx left15:37
jast autogc does only a subset of gc operations15:37
→ chrisdev joined15:38
← fruight left15:38
← snow_usa left15:38
→ duckxx joined15:39
← tjaalton left15:39
→ mishok13 joined15:39
→ nrichards joined15:39
← eno left15:40
← neurodrone left15:40
b0fh_ua okay, how to find detached head?15:40
→ eno joined15:40
← pidus left15:40
b0fh_ua or some detached commits15:40
cmn what do you mean?15:40
FauxFaux b0fh_ua: !gka15:40
gitinfo b0fh_ua: For a better way to view the reflog, try: gka() { gitk --all $(git log -g --format="%h" -50) "$@"; }; gka15:40
← rchavik left15:41
b0fh_ua cmn: looks like we missed the commit somehow15:41
→ p3rror joined15:41
cmn missed?15:41
as in lost?15:41
b0fh_ua cmn: seems so15:41
cmn then do what FauxFaux says15:41
b0fh_ua guy did rebase after some amend of commit he was not owning15:41
so something is wrong here15:41
→ ehsan joined15:41
b0fh_ua and we don't have gitk - it's on windows15:42
→ armin_ joined15:42
canton7 msysgit does have gitk15:42
→ imgrey joined15:42
b0fh_ua and how to use gka?15:42
FauxFaux In git bash, or by hand.15:42
b0fh_ua okay what arguments should be given?15:42
FauxFaux To gka? None.15:43
imgrey good evening #git15:43
could somebody enlighten me on how to perform pull on remote server after I pushed changed to master ?15:43
it says: """15:43
You asked me to pull without telling me which branch you15:43
want to merge with, and 'branch.master.merge' in15:43
blah blah blah """15:43
FauxFaux imgrey: And?15:43
SethRobertson imgrey: What command did you try to run?15:43
Is the server a bare repo?15:44
imgrey SethRobertson, git pull15:44
SethRobertson, yes, it is bare repo15:44
→ touk joined15:44
SethRobertson So you are pushing to a server and then having it pull? Are there two servers?15:44
imgrey SethRobertson, my .git/config: http://pastie.org/334878815:44
→ gnufied joined15:45
imgrey SethRobertson, I commit on my home machine with git push and then Im trying to update server with git pull15:45
FauxFaux imgrey: Sounds about right. Well, you should probably read that git output which tells you what you need to do.15:45
SethRobertson How many repositories do you have?15:45
← herdingcat left15:45
imgrey SethRobertson, only one15:45
← eno left15:46
← arvind_khadri left15:46
→ _ikke_ joined15:46
imgrey FauxFaux, I know, but I tired to read docs about git. Now I want it to just apply my changes without questions15:46
canton7 imgrey, the argument-less form of "git pull" requires so-called "tracking information" to tell it what to fetch and merge into the current branch. You don't have tracking set up. Therefore either use the full form of git pull -- "git pull origin <branchname>", or set up tracking, see push -u and branch --set-upstream15:46
SethRobertson Let us try this again. You commit some code somewhere. That is a first repository (#1). Then you push somewhere. That is a second repository (#2) and is presumably bare. Then you are pulling. Where are you pulling from?15:46
→ eno joined15:46
→ cakehero joined15:47
← raatiniemi left15:47
imgrey SethRobertson, I pull from the same bare repository15:47
SethRobertson, that I pushed to15:47
candybar, ok, I do % git pull origin master15:47
error: Your local changes to the following files would be overwritten by merge:15:48
users/admin.py15:48
Please, commit your changes or stash them before you can merge.15:48
→ troydm joined15:48
FauxFaux READ THE FUCKING OUTPUT15:48
Ahem.15:48
imgrey FauxFaux, go **** yourself15:48
canton7 I could try and explain that message, but I really wouldn't be able to make it any clearer than the message already does15:49
← wbednars_ left15:49
troydm ok i have some branch lines in git log-tree (it's short alias for tree) that are colored out with different colors, i want to know which config variable controls color values15:49
→ _mpu joined15:49
imgrey candybar, what stash is ?15:49
← Wombert left15:49
FauxFaux imgrey: Trivial to google.15:49
canton7 1. my name isn't candybar, 2. man git-stash15:49
gitinfo the git-stash manpage is available at http://jk.gs/git-stash.html15:49
_mpu hi, sometimes git status informs me about the status of my repository compared remote ones, sometimes not.15:50
what setting governate this?15:50
which15:50
shruggar the "upstream" of a branch15:50
FauxFaux _mpu: It's based on the presence of tracking information that you see in git branch -vv15:50
canton7 _mpu, iirc, whether there's anything to report, and whether tracking infor is set up for the branch15:50
imgrey OMG, Conflict, conflict!15:50
CONFLICT (content): Merge conflict in users/admin.py15:50
what do I do ?15:50
→ Phoenixz joined15:51
canton7 imgrey, fix the conflict. Edit users/admin.py, find the merge markers, and sort the conflict out15:51
imgrey, or launch git mergetool15:51
imgrey, then "git add users/admin.py; git commit"15:51
FauxFaux troydm: https://github.com/gitster/git/blob/master/log-tree.c#L2515:52
_mpu canton7, FauxFaux: ok, so it is related to the branch command.15:52
Phoenixz I have project A with its git repo, and now I need to add a (jquery based) project B.. Project B has its own git repo already, but somehow I need to get it into the tree of A.. I could just copy the WT of B into A, and be done with it, but I'd like to know if there may be an option in GIT to include another repo somewhere into the current repository...15:52
← codewrangler left15:52
imgrey finally succeeded15:53
→ Guest4644 joined15:53
imgrey candybar, thank you very much15:53
SethRobertson, thank you too15:53
troydm FauxFaux: hmm thx15:53
SethRobertson !thanks15:53
gitinfo Feeling thankful? Type "ExampleUser++", and ExampleUser will score karma points at http://carmivore.com (our preferred way to objectify self-worth). There's really no point to thanking me – instead, why not thank the person who made me help you?15:53
imgrey FauxFaux, get a life15:53
← eno left15:53
canton7 please stop spamming candybar....15:53
FauxFaux imgrey: Learn to read? :)15:54
← linkedinyou left15:54
← JPT left15:55
→ linkedinyou joined15:55
→ eno joined15:55
→ JPT joined15:55
→ alex__c2022 joined15:56
← Guest4644 left15:56
RandalSchwartz Phoenixz - use a git subtree merge15:56
or get "git-subtree" (third party, but really should be part of the core)15:56
→ Wombert joined15:57
→ zatan joined15:57
cmn that's in progress15:57
← yshh left15:57
zatan Hi is there any command to see my git configurations ?15:57
cmn git config15:57
← mp_ left15:57
cmn -l shows everything15:57
if you're looking for something very specific, other tools might give output tailored to the specific use case15:58
→ v0n joined15:58
← iocor left15:58
zatan cmn, thanks ;]15:58
← linkedinyou left15:58
→ linkedinyou joined15:58
→ arvind_khadri joined15:58
→ jasonbray joined15:59
troydm FauxFaux: so regarding those colors it means they are hardcoded and there is no way to change them from config15:59
?15:59
FauxFaux That would be my guess, yes.15:59
troydm FauxFaux: ic, thx15:59
← cburroughs left15:59
← oo_miguel left16:00
← Francisc left16:00
← eno left16:00
← JC_Yang left16:01
← _mpu left16:01
→ duckx joined16:01
← cccaldas left16:02
→ eno joined16:02
→ fedesilva joined16:02
← duckxx left16:03
→ i42n joined16:05
→ kevwil joined16:05
→ ipalaus joined16:05
→ Cromulent joined16:06
→ ScottE joined16:06
→ PySloan joined16:06
← BorX left16:07
← Samoi left16:07
→ lost_eden joined16:07
→ Targen joined16:08
lost_eden I've just tried to setup Gitweb but an existing repository isn't showing up - is there something special I have to do to 'add' an existing repository?16:08
→ Samoi joined16:09
→ Sailias joined16:09
← b0fh_ua left16:09
→ bascht joined16:10
ChanServ set mode: +v16:10
← PySloan left16:11
← eno left16:14
→ rzec joined16:14
→ iocor joined16:14
→ mr_mustard joined16:14
gitinfo set mode: +v16:15
Sailias is there a preferred way to hide users commits? eg, i am working on a repo and want other developers to work on it, but i don't want them to commit to master, i want everything to look like it's coming from me16:15
→ PySloan joined16:15
PerlJam Sailias: sounds disingenuous16:15
← cbreak-work left16:15
← airborn left16:15
cmn don't allow them to commit on master16:16
rzec say I want everyone to use core.filemode false, it there a way to setup the bare repository everyone clone from so that there cloned repository has this configuration?16:16
← batlock666 left16:16
PerlJam Sailias: or do you really mean to say you'd only like to see your commits?16:16
cmn how you do that depends on the system you use16:16
rzec: configuration isn't cloned16:16
→ eno joined16:16
cmn you can include a default config in your repo as a file, but that's pretty much it16:17
← gldn left16:17
Sailias PerlJam, i'd like all commits to the repo to look like they are coming from me. So perhaps a fork that they work on, and i point that master to the original repo and rebase all their commits16:17
← arvind_khadri left16:18
cmn why?16:18
what possible use can that have?16:18
Sailias cmn, was that directed to me?16:18
cmn yeah16:18
→ Wombert_ joined16:19
← jceb left16:19
Phoenixz RandalSchwartz: Thanks, will look into it!16:19
Sailias cmn, a client wants me to work on the project, but i'm going to sub contract some of the extra work out, and i don't want them to know about it16:19
→ ASIRINeiluj joined16:19
cmn figures, it's about lying in the repo16:19
PerlJam like I said ... disingenuous16:20
cmn then grab their repo once they're done and filter-branch16:20
jpalmer git commits don't care about your ethical dilemmas ;)16:20
Sailias cmn, i never said i had an ethical question, just a technical one ;)16:20
cmn trying to solve a social problem with a technical one brings problems16:20
that's why reasons are important16:20
PerlJam Sailias: this is a technical question I won't expend any brain power on :)16:20
← Wombert left16:21
Wombert_ → Wombert16:21
MrKeuner cmn, you have suggested this pre-receive hook. what does old new branch stand for?16:21
cmn hm?16:21
MrKeuner https://gist.github.com/171423516:21
line 216:21
cmn oh that16:21
PerlJam MrKeuner: that's what gets passed to the hook16:22
cmn the script gets fed three variables per line on stdin16:22
MrKeuner PerlJam, env variables?16:22
cmn old and new are the old and new commits16:22
MrKeuner shouldn't they be $old $new $branch then16:22
cmn not there16:22
PerlJam MrKeuner: no, that's not how read works16:22
cmn the read is what's creating them16:22
MrKeuner ah ok16:22
← eno left16:23
→ dermoth_ joined16:23
nru hi16:23
is there a simpler equivalent to doing the following?16:24
git checkout B; git update-ref --no-deref HEAD A; git commit -a; git checkout B; git cherry-pick HEAD@{1};16:24
← stodan left16:24
cmn what's the problem you're trying to solve16:24
nru that is, commiting the state of branch B on top of branch A16:24
cmn you're presenting a solution16:24
→ BorX joined16:24
cmn rebase?16:24
→ eno joined16:24
nru cmn: hum, I don't see how rebase would be used to do that.16:24
cmn cherry-pick doesn't commit the state of anything, it applies a patch16:25
nru (that's why I was asking with a solution; that's easier to convey)16:25
cmn are you looking for pull --squash?16:25
merge --squash, rather16:25
nru no16:26
I don't want to merge16:26
cmn that doesn't merge16:26
← lost_eden left16:26
→ dermoth__ joined16:26
← dermoth|work left16:26
→ notjohn joined16:26
nru I want a commit on branch B that makes the working tree associated with branch B the same as the working tree associated with branch A16:26
→ alexgordon joined16:27
→ boombatower joined16:27
→ LetoTo joined16:27
→ pidus joined16:27
→ neurodrone joined16:27
nru (by "the working tree associated with branch A" I mean "the result of 'git checkout A'". Is there a better way to say that?)16:27
cmn the commit's tree16:27
nru yes16:27
Sailias cmn, PerlJam: i don't think what i'm asking is that unethical, it's about having a dev team that i can make commits look like they're coming from "My Dev Shop" and not John Doe. It also comes to not exposing the names of my employees.16:27
cmn, PerlJam: but ty for your time16:28
LetoTo should i be worried about pushing a new local branch to remote when it says: "emote: fatal: bad object 0000000000000000000000000000000000000000". It seemeed to have worked anyway16:28
← iocor left16:28
cmn git checkout B; git checkout A -- .; git commit -a would get close; not sure what would happen to deleted files16:28
Sailias: nobody but you mentioned it being unethical16:28
nru so my problem, hopefully restated in a better way: I want to commit the tree of A's head on top of B, whatever B is16:28
← Wombert left16:29
dermoth__ → dermoth|work16:29
cmn you can go low-level and create the commit yourself16:29
nru how would I do that?16:29
← d3vic3 left16:29
← jatt left16:29
cmn and make it point to A's tree16:29
← mr_mustard left16:29
cmn there's commit-tree, I haven't played much with that part16:29
→ CrazyWoods joined16:29
→ smw joined16:29
→ alvaro_o joined16:29
← dermoth_ left16:29
cmn man git commit-tree16:29
gitinfo the git-commit-tree manpage is available at http://jk.gs/git-commit-tree.html16:29
nru thanks16:30
cmn it looks like it does what you want to do16:30
← alvaro_o left16:30
cmn git checkout B; git commit-tree -m "fancy commiting" A^{tree}16:30
nru indeed! thanks16:30
yrlnry And then it prints out a new hash for the new commit and you make a ref to that hash.16:30
Don't forget that step!16:30
nru uh; how is A^{tree} different from A?16:31
← fayimora left16:31
→ d3vic3 joined16:31
cmn A is a commit16:31
→ arvind_khadri joined16:31
cmn A^{tree} prints the commit's tree16:31
yrlnry You can just say "A" instead of "A^{tree}" and it will do the right thing.16:31
CrazyWoods what shall i do if i had multi project which are like to use git. just create a single repository and store all my project in the repository or create multi repository for every project?16:31
cmn yrlnry: are you sure? it says <tree> and not <tree-ish>16:31
→ Tureos joined16:31
← eno left16:31
yrlnry cmn: I am mistaken.16:32
PerlJam CrazyWoods: if they are tightly coupled, one repo. If loosely-coupled (or should be), multiple repos.16:32
CrazyWoods: that's my personal rule-of-thumb16:32
nru 17:30:59 <+yrlnry> And then it prints out a new hash for the new commit and you make a ref to that hash. <- can't I just cherry-pick it?16:32
cmn I guess that would work16:33
yrlnry cherry-pick it to where?16:33
It's already set up with the parent you want.16:33
cmn git update-ref HEAD <that hash> would be more correct16:33
→ eno joined16:33
PerlJam CrazyWoods: (it's rare that I've used one repo for multiple projects ... I've only done it once that I can recall)16:33
Mikachu it's easier if you just checkout b; read-tree a; commit16:33
nru yrlnry cmn: OK, thanks very much16:34
CrazyWoods PerlJam, thank you16:34
nru Mikachu: oh! thanks16:34
cmn Mikachu raises a very good point16:34
Mikachu (read-tree won't update the working tree though, so add a reset too afterwards)16:34
→ diogogmt joined16:34
→ snow_usa joined16:35
← armin_ left16:35
← teweWORK left16:37
← webben left16:38
← lapistano left16:38
→ nixmaniack joined16:38
→ lapistano joined16:38
cmn I still think a squash-merge would solve this as well16:38
← d3vic3 left16:38
← shiba_yu36 left16:38
← leavittx left16:39
nru cmn: reading the man, I can't see how16:39
← eno left16:39
→ kennethreitz joined16:40
cmn it takes the difference between HEAD and the other branch and applies the diff on top16:40
unless you really want to loose all the stuff you've done on the current branch16:40
nru cmn: I really want to lose it16:40
otherwise I would have done a merge16:41
cmn a rename seems appropriate16:41
← chrisdev left16:41
→ eno joined16:41
cmn as now you have a bunch of commits and a big one w/o explanations16:41
nru also to me it looks like git merge --squash is really like git merge; git reset --soft HEAD@{1} so I wouldn't have bothered asking16:42
← Sailias left16:42
← flijten left16:42
nru cmn: in fact, I had already applied part of the commit which tree I wanted to transplant16:42
cmn git merge --squash is git merge --no-commit B; git reset --soft HEAD^; git commit16:42
← derekpa left16:43
→ sebbi joined16:43
nru wait, the doc says that --squash does not actually make a commit16:43
cmn so it does16:44
nru (In my case, I was in the process of cutting a commit into smaller parts; only I seem to have done so in an absurdly complicated way)16:44
← johnathanb left16:45
→ ajaykools joined16:47
mastro I'm having W: RA layer request failed: Server sent unexpected return value (405 METHOD_NOT_ALLOWED) in response to REPORT request for '/svn/docasu/!svn/bc/355' at /usr/lib/git-core/git-svn line 214416:47
while trying to svn clone a repository...16:47
it's a warning and I would ignore it if it wasn't that after saing: "This may take a while on large repositories" it stop without doing anything16:48
← antgel_ left16:48
← BSaboia left16:49
← brian_g left16:50
Kingsy curious, some functionality I don't quite understand... if I do git checkout -b fault; echo "this" >> somefile.txt; git checkout master; why does it say M somefile.txt when you change the branch? and somefile.txt seems to have the change from fault in the master branch16:50
→ brian_g joined16:50
← Shusshu left16:50
→ pfFredd joined16:51
← eno left16:51
Kingsy I mean I previously merged another branch with master.. but that has gone, I don't see why its passing the changes of a file that hasnt even been staged to another branch16:51
cmn Kingsy: the file isn't in any branch16:51
RandalSchwartz uncommitted files are dragged along when you do checkout, unless there'd be a conflict.16:51
Kingsy cmn: ahhh16:51
→ BSaboia joined16:51
Kingsy so you hvae to stage it16:52
cmn no16:52
Kingsy you have to commit it?16:52
cmn a branch is a pointer to a commit16:52
Kingsy ah ffs16:52
→ mikepack joined16:52
Kingsy I knew that16:52
cmn: thanks16:52
mattp_ is it possible to checkout only certain directories when specifying a commit?16:52
→ eno joined16:52
← eno left16:52
→ eno joined16:52
cmn I should probably pimp out !thanks more16:52
gitinfo Feeling thankful? Type "ExampleUser++", and ExampleUser will score karma points at http://carmivore.com (our preferred way to objectify self-worth). There's really no point to thanking me – instead, why not thank the person who made me help you?16:52
cmn Seth is getting way ahead16:52
→ dAnjou joined16:52
RandalSchwartz mattp_ - if you're not going to be making a new commit, sure16:53
cmn mattp_: you can do git checkout <commit> -- <files>16:53
mattp_ ie say i want to check out an old version of root/ where all my html templates are, but leave the web app files alone16:53
RandalSchwartz that just replaces <files> in your current commit with that commit16:53
cmn then do that16:53
but be careful if you want to commit afterwards16:53
← aristidesfl left16:54
Kingsy cmn++16:54
mattp_ sweet, that worked beautifully. git rules16:54
→ mandric joined16:54
mattp_ so when i want to get back to HEAD, `git reset --hard HEAD` will do it? i dont care about tossing any changes16:54
cmn "state of HEAD" technically, but yeah16:55
← smw left16:55
cmn or checkout -f16:55
← sameer_ left16:55
← h0bbit left16:58
→ sameer_ joined16:58
← retoryter left16:58
CareBear\ reset --hard will leave untracked files16:58
→ Milossh joined16:58
cmn well yeah16:59
untracked files don't get touched16:59
→ tewecske joined16:59
← Cromulent left17:00
Kingsy hmm not sure I agree with the way merging works.. unless I am misunderstanding, if you change the same line of the same file you need to pick between one or the other?17:00
when you merge I mean17:00
CareBear\ yes17:01
→ webben joined17:02
CareBear\ you get a conflict17:02
and have to resolve it17:02
→ dfr|mac_ joined17:02
cmn no17:02
← i42n left17:02
← whitman left17:02
cmn if you change the same line in two branches, git tells you you need to decide17:03
→ teknoid joined17:03
cmn and it presents you with both versions, you as a human decide what the real solution is17:03
← overbythere left17:04
← dfr|mac left17:04
Kingsy what if the scenario arises where for example.. you have a production copy.. you open a branch to work on a new version ( you add one line to the file ) something comes up, so you go back to the master and apply a "fix" to the file ( by adding a line to the master ) you then go back to the upgrade branch and finish editing your line, you switch back to master and merge.. its going to ask you to choose.. what if you wanted both?17:04
you still want the code to work (with the fix) but you want your new version too17:04
Mikachu you can resolve the merge however you want, one side, both sides, something completely different17:05
← arvind_khadri left17:05
Kingsy Mikachu: ah right, I didnt realise... sorry17:05
Mikachu np17:05
→ Heimidal joined17:05
Kingsy lol loads of typing for nothing then.. I assumed it was asking you to choose one or the other17:05
→ i42n joined17:05
cmn what made you think that?17:05
← eno left17:05
Kingsy cmn: just the way it was set out in the filew17:06
nevermind, I was just being dumb17:06
cmn that's the standard three-way merge conflict markers17:06
← Vampire0 left17:06
cmn it's showing you both versions17:06
→ eno joined17:06
← ajaykools left17:06
→ dreiss joined17:07
→ Vampire0 joined17:07
← robotmay left17:07
Kingsy cmn: yeah I see, and I guess you can address it with vimdiff or whatever and decide to have both if you want..17:07
thanks for the clarifaction17:07
cmn you can use a tool if it's a complex conflict17:08
often you can just do it by hand17:08
Kingsy mergetool ?17:08
→ airborn joined17:08
← jacekmigacz left17:08
→ dfr|mac joined17:08
cmn mergetool calls whatever tool you've configured17:08
→ Destos joined17:09
nru Kingsy: mergetool lets you launch a tool; kdiff3 and meld are both very good and intuitive17:09
Kingsy ah I see.. yeah as set in --config17:09
← zatan left17:09
← giallu left17:09
nru Kingsy: you can try out tool with git mergetool -t meld for example17:10
*tools17:10
← shergill1 left17:10
Kingsy ah ok, nru thanks17:10
→ shergill joined17:10
← capncodewash left17:11
→ dfr|mac__ joined17:11
← Phoenixz left17:11
← BorX left17:11
→ mr_mustard joined17:11
← dfr|mac_ left17:12
→ jacekmigacz joined17:12
mr_mustard I set a post receive hook to do the following: GIT_WORK_TREE=/home/user/public_html/www.site.com git checkout -f. But apparently, it only executes when I push the master branch. How do I set it to execute with another branch?17:13
nru damn, now I'm affraid I might have borked something.17:13
what's the standard way to cut a commit into multiple smaller commits17:13
→ fruight joined17:13
nru when the splitting is non-trivial (might need manual editing) ?17:13
cmn reset and selective adds17:13
← dfr|mac left17:13
nru cmn: what about when the splitting needs to manually edit stuff?17:13
cmn commit, stash, ammend17:14
nru !17:14
indeed17:14
thanks17:14
← sameer_ left17:15
cmn it's often better to commit too much17:15
→ eijk__ joined17:15
← brokenshakles left17:15
← eijk_ left17:15
← grekko left17:15
cmn and then squash the commits together; don't be commit-shy17:15
nru indeed; my problem ran deeper17:16
→ maletor joined17:16
→ dfr|mac joined17:16
← Nilbus_ left17:16
→ Nilbus joined17:17
→ dorisabayon joined17:17
→ fayimora joined17:18
→ chrisdev joined17:18
← eno left17:19
← tcurdt left17:20
→ norton- joined17:20
← chrisdev left17:20
← Fandekasp left17:20
← dfr|mac__ left17:20
→ sameer_ joined17:21
→ eno joined17:21
→ mxweas_ joined17:23
Mikachu nru: or if you're hardcore, add -e17:23
← psoo left17:25
nru Mikachu: \o/ I did not know this existed17:25
Mikachu (you don't need to manually adjust the hunk headers if you add/remove lines)17:25
← Kingsy left17:25
→ xsism joined17:26
cmn but you do need to write the patch format by hand17:26
I always forget the initial SP when writing context17:26
Mikachu if you're slightly less hardcore, add -p is also nice17:26
← LetoTo left17:26
Mikachu i think it also lets you pop up the editor for individual hunks?17:26
→ johnathanb joined17:26
cmn yeah17:27
with 'e'17:27
← LimeBlast left17:27
nru pretty cool17:27
← dreiss left17:27
→ Syco54645_work joined17:28
→ iocor joined17:28
Syco54645_work i am looking for a good graphical git client17:29
that is cross platform17:29
← mr_mustard left17:29
cmn gitk, git-gui17:29
both come with git and are cross-platform17:29
Syco54645_work ok, i tried gitk17:29
cmn they're good at what they do; gitk is ugly, but useful17:29
Syco54645_work yes, it was ugly but very very functional17:30
cmn it of course depends on what you mean by good17:30
→ kermit joined17:30
Syco54645_work easy to use and such, im fine with cli17:30
it is the coworkers who i am trying to help out17:30
← kermit left17:30
← johnathanb left17:30
→ BryanRuiz1 joined17:30
Syco54645_work you say git-gui comes with git?17:31
BryanRuiz1 can i unmerge a branch :)17:31
cmn yeah17:31
BryanRuiz1: maybe17:31
BryanRuiz1: http://www.opensource.apple.com/source/Git/Git-17/src/git-htmldocs/howto/revert-a-faulty-merge.txt17:31
Syco54645_work cmn: hmmm, doesnt appear to be on my system17:32
cmn 'git gui'17:32
the dashed version doesn't work for typing it17:32
→ kermit joined17:32
→ brokenshakles joined17:32
cmn people seem to like 'git gui blame <file>' a lot as well17:32
→ linusoleander joined17:33
← senny left17:33
→ snearch joined17:33
← norton- left17:33
Syco54645_work git: 'gui' is not a git command. See 'git --help'.17:33
→ norton- joined17:34
← kadoban left17:34
cmn your package manager might have split that out to a different package17:34
← OOPMan left17:34
cmn in Debian it's in the git-gui package17:35
→ eijk joined17:35
→ notola joined17:36
Syco54645_work cmn: ah that was it17:36
Syco54645_work is on mint, needed a change17:36
→ mcluskydodallas joined17:37
← srcerer left17:37
cmn mint is based on debian/ubuntu so pretty much the same17:37
Syco54645_work cmn: indeed17:37
→ anjen joined17:38
← eijk__ left17:38
Syco54645_work cmn: is there a way for them to rename and move files in git-gui?17:38
cmn I'm not sure, I don't use it myself17:38
← marthinal left17:39
← eno left17:39
→ OOPMan joined17:40
→ eno joined17:40
→ rtconner joined17:41
Syco54645_work cmn: well thanks for the help regardless, i assume the mac users here will use tower17:41
MrKeuner I have multiple remotes, I commit some change to one of them and doing git status I don't see the customary message that my copy is 3 commits ahead, etc17:42
am I doing something wrong?17:42
Syco54645_work i was just trying to find something that i could get familiar with before i introduced them to scm via git17:42
MrKeuner is that because git canot know which remote to compare with?17:42
→ apok joined17:43
cmn you tell git which remote to compare with17:43
MrKeuner git status remote2?17:44
cmn branch --set-upstream17:45
← gnufied left17:46
← OOPMan left17:46
← eno left17:46
→ tystr joined17:47
→ eno joined17:47
← CrazyWoods left17:47
← dpino left17:48
BryanRuiz1 "Your branch is ahead of 'origin/master' by 6 commits." How do I look at what these commits are? Just git log at the last 6?17:48
→ jbwiv joined17:48
SethRobertson Or `git log @{u}...`17:49
← apok left17:49
jbwiv guys, I have a series of commits which were committed using the wrong email address as the author. is there a way to fix this?17:49
← dAnjou left17:49
SethRobertson Have you pushed?17:49
thiago does it include any merges?17:49
jbwiv SethRobertson, yes17:49
thiago do other people have them?17:49
jbwiv thiago, yes, I believe so. some of the commits are from the very early days17:50
SethRobertson jbwiv: Then this is !rewriting_public_history and is not recommended. git can do it, configuration on the server allowing, of course.17:50
gitinfo jbwiv: 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)17:50
jbwiv thiago, only one (and github), and I can tell him to re-clone17:50
→ sec^nd joined17:50
→ apok joined17:50
SethRobertson git-filter-branch is probably the best solution for you17:50
← joshsmith left17:50
SethRobertson with the above caveats17:50
thiago agreed17:50
jbwiv ok, thanks guys17:51
sec^nd I have a git repo which is extremly slow when updating or git status'ing, is there a way to speed it up ? the git repo is here git://pkgbuild.com/aur-mirror.git17:51
MrKeuner branch --set-upstream dev master; git status tells me that I am ahead by 3 commits; when do git push dev master I get no change17:51
← shergill left17:51
cmn that's because pushing doesn't change the dev branch, only $remote/dev17:51
or master or whatever17:52
→ churp joined17:52
MrKeuner cmv I am in dev and pushing to dev master17:52
SethRobertson sec^nd: `git gc --aggressive` and buy faster disks. There was also a recent thread about this on the git mailing list17:52
MrKeuner cmv I am in branch dev and pushing to remote dev master17:52
→ airstream66 joined17:53
sec^nd SethRobertson: can you point me to the thread ?17:53
→ johnanderson joined17:53
→ DragonSlay joined17:53
← rzec left17:53
← Element9_work left17:54
SethRobertson MrKeuner: When you explictly name a remote branch in push it does not update the remotes/ Set push.default correctly (tracking) and skip the explicit branch name17:54
sec^nd SethRobertson: I did git gc --aggressive last night and woke up to a computer that was swapping out :/17:54
computer has 2G of ram and 3G of swap17:54
cmn MrKeuner: only one of those can be a remote name17:54
DragonSlay what email address should i give for git config? should i use my local email address?17:54
cmn you need to specify the complete name of the branch17:54
sec^nd git was using way to much ram17:54
cmn remote/branch17:54
sec^nd and wasn't done after running all night17:54
→ soc42 joined17:55
→ gltest joined17:55
SethRobertson sec^nd: See subject "large(25G) repository in git"17:55
sec^nd: Also see !large_files and more importantly, to you, see:17:55
gitinfo sec^nd: Git isn't yet great at large files (larger than RAM). See http://git-annex.branchable.com/ and https://github.com/schacon/git-media Find them (after gc) with: git verify-pack -v .git/objects/pack/pack-*.idx | grep blob | sort -k3nr | head | while read s x b x; do git rev-list --all --objects | grep $s | awk '{print "'"$b"'",$0;}'; done17:55
MrKeuner SethRobertson, does not update as a security precaution, right otherwise what I am trying to do makes sense, although dangerous?17:55
← Heimidal left17:55
SethRobertson sec^nd: !large_repo17:55
gitinfo sec^nd: [!large_repos] Git can be slow in the face of large repositories. There are git-config options which can help. pack.threads=1; pack.deltaCacheSize=1; pack.windowMemory=512m; core.packedGitWindowSize=16m; core.packedGitLimit=128m. Other likely ones exist.17:55
→ Heimidal joined17:56
← achadwick left17:56
SethRobertson MrKeuner: Security? I doubt it. Probably something to do with explicit naming meaning you are doing weird stuff and git shouldn't try to add too much convenience. Same with pulling/fetching with an explicitly named branch17:56
← gltest_ left17:56
← Heimidal left17:57
← mxweas_ left17:57
← kevlarman left17:58
→ kevlarman joined17:58
← jason237 left17:59
← abetaha left18:00
SethRobertson sec^nd: I gave you a wrong reference. The subject you should be looking for is "Git performance results on a large repository"18:00
→ vacho joined18:01
← LongBeach left18:01
vacho im adding this line to my .gitignore file but it does not listen to me :) "sites/default"18:01
MrKeuner SethRobertson, are you saying branch name = remote name is a bad idea?18:01
vacho if I make changes to files in that folder it still wants me to commit and push..why>18:01
sec^nd thanks18:01
→ magmatt joined18:01
SethRobertson MrKeuner: Just change the push.default setting in man git-config18:02
gitinfo MrKeuner: the git-config manpage is available at http://jk.gs/git-config.html18:02
magmatt I want to list the author for each branch in the output of `git branch --no-merged`18:02
MrKeuner ok18:02
SethRobertson magmatt: Branches do not have owners18:02
magmatt: !branch18:02
gitinfo magmatt: A branch and a tag are just convenient ways of spelling the name of a particular commit. A commit represents a specific set of files and the history of all commits which came before it, and the SHA-1 hash tag official name provides cryptographic assurance of the lineage of a particular commit (and thus branch or tag). A branch's reference may change. A tag usually doesn't.18:02
magmatt SethRobertson: commits do18:02
I just mean the head commit that the branch is pointing to18:03
→ jason237 joined18:03
vacho anyone please?18:03
← ceej left18:03
SethRobertson magmatt: Then post-process the output so you can feed the SHA/branch names into git-show, git-log, or similar18:03
← eno left18:03
→ marthinal joined18:04
SethRobertson vacho: !gitignore18:04
gitinfo vacho: [!gitignore_whitedir] A global/directory .gitignore blacklist with a (potentially) subdirectory whitelist is not easy to specify in git. However, something like `echo -e '/*\n!/a/\n/a/*\n!/a/b/\n/a/b/*\n!/a/b/c/' > .gitignore` (ignore everything but a/b/c directory) or `echo -e '*\n!*/\n!*.c' > .gitignore` (ignore everything but *.txt files) may do what you want.18:04
magmatt SethRobertson: dang, that's what I was already doing; k thanks18:04
← magmatt left18:04
→ Heimidal joined18:04
SethRobertson vacho: Sorry, wrong reference18:04
→ CannedCorn joined18:04
nru is there a way to edit a commit diff in a way similar to add -e ?18:04
SethRobertson vacho: !ignore_tracked18:04
gitinfo vacho: Git only applies ignore patterns to untracked files. You can't use ignore patterns to ignore changes to files that are already tracked by git. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.18:04
SethRobertson Unless I don't understand the problem, in which case !repro18:04
gitinfo Please paste (using https://gist.github.com/ or similar) a transcript of your terminal session. This will help immensely with troubleshooting.18:04
← shruggar left18:05
SethRobertson nru: Once you have committed, you can only rebase-i (or reset --soft for the most recent commit) and then redo the commit18:05
→ eno joined18:05
SethRobertson Using add -e or whatever18:05
nru ok; thanks18:06
← ASIRINeiluj left18:06
→ Heimidal_ joined18:06
← hotsyk_ left18:07
← t4nkd left18:08
→ t4nkd joined18:08
← MrKeuner left18:08
jaalto Does Git have some ocnfiguration variable that would stop it starting EDITOR on conflict merge?18:10
→ joe-mac2 joined18:10
→ ajaykools joined18:10
← Heimidal left18:11
gitinfo set mode: +v18:11
joe-mac2 is tehre a way besides writing a little script that with git pull I can not only pull in the changes to my local branch that's mapped to a remote bvranch, but do it for all branches that currently have a mapping to a remote branch?18:11
← eno left18:11
→ abetaha joined18:11
→ coachz joined18:12
jaalto joe-mac2: No git command, so you need to write a shell script18:12
coachz hi, is there any easy way to undelete a branch on my local box ?18:12
joe-mac2 jaalto: thought so, easy enough, thanks18:12
← joe-mac2 left18:12
→ eno joined18:12
→ Skyx-mobile__ joined18:12
→ Russell^^ joined18:12
← gchristensen left18:13
← flyz left18:13
jaalto coachz: If you delete the branch, Git shows the SHA. Just use the sha to branch again.18:13
→ gchristensen joined18:13
coachz how do i display the sha for the branch ?18:13
→ arvind_khadri joined18:14
← arvind_khadri left18:14
→ arvind_khadri joined18:14
→ kenperkins joined18:14
PerlJam coachz: git rev-parse <branchname>18:14
→ neunon joined18:14
← neunon left18:14
→ neunon joined18:14
coachz fatal: ambiguous argument 'jb_dash2': unknown revision or path not in the working tree.18:14
jaalto coachz: I mean, if you just deleted the branch, Git should have outputted the SHA with the last delete. But if you did it some time ago...18:14
← CannedCorn left18:15
coachz i have switched branches since then18:15
jaalto coachz: in that case, do some archeology with " git reflog show -COUNT"18:16
→ LeMike joined18:16
→ disappearedng joined18:16
coachz i think i see it18:16
so i can "git checkout SHA1" ?18:17
jaalto coachz: Use the SHA you see to make a new branch, and you're back18:17
coachz like this? "git checkout -b SHA1" ?18:17
jaalto coachz: You don't usually want "checkout" and you'd be put to a "no branch"18:17
CareBear\ git checkout -b foo sha118:17
← disappearedng left18:17
CareBear\ where foo is the name of the branch you want to create and sha1 is a refspec18:18
coachz thank you18:18
jaalto ...and.. => "or you'd be put..."18:18
CareBear\ -b is of course documented in man git-checkout too18:18
gitinfo the git-checkout manpage is available at http://jk.gs/git-checkout.html18:18
coachz yeah i use -b a lot :-) thanks very much18:18
jaalto CareBear: That backslash in your nick ruins my display :-)18:18
→ kukks joined18:18
CareBear\ jaalto : fix your shitty display18:18
← tatsuya_o left18:19
jaalto CareBear: It's lie red flag18:19
like18:19
CareBear\ jaalto : fix your display18:19
→ tatsuya_o joined18:19
Nugget /IGNORE CareBear\ is a valid fix. Just sayin'18:19
→ alx- joined18:19
→ dorkmafia joined18:19
→ flyz joined18:20
CareBear\ jaalto : I added a trailing backslash because I needed an extra character to avoid a nick collision and backslash was pretty unusual, and looked nice. also I thought it fun to see if it would cause any problem with various software that came across my nick, like bots and such18:20
jaalto : that was in 199518:20
← PigeonFriend left18:20
← iocor left18:21
CareBear\ jaalto : it really amazed me how many programs were unable to handle the character properly18:21
Nugget if you'd chosen a back-tick the internet might not have survived.18:21
CareBear\ jaalto : some time later it made me very angry when I saw some software having problems18:21
jaalto : these days I just laugh and spit at the shitty software18:22
jaalto CareBear: I'm from 1992, but I would never use unreadable character for conversation channel. No pun intended.18:22
← pfFredd left18:22
CareBear\ I can read it fine :)18:23
← airstream66 left18:23
Nugget I just wish it ended :\ instead of just \ so it looked more like a DOS prompt.18:23
← DragonSlay left18:23
CareBear\ Nugget : it's slightly less universal I think..18:23
Nugget CareBear:\>18:23
CareBear\ Nugget : hahaha, nice! :)18:24
SethRobertson ¿Unreadable characters‽18:24
CareBear\ is : allowed in nicks?18:24
← notola left18:24
Nugget no clue18:24
→ hydrozen joined18:24
CareBear\ I think not18:24
← tatsuya_o left18:24
CareBear\ it's used to say that a command has a parameter with spaces18:24
Nathandim → Nathan|dnd18:25
CareBear\ also, back then, the limit was 9 characters18:25
I guess it still is on those networks18:25
← jfelchner left18:26
← thiago left18:26
→ whitman joined18:27
→ rbrooks joined18:27
← diogogmt left18:28
→ OOPMan joined18:28
→ dreiss joined18:29
← eno left18:30
→ eno joined18:30
→ slicslak joined18:30
slicslak i want to master into my feature branch, but before i do that I want to see what changes that will bring. what's the best command to do that18:31
← cortana2 left18:31
← t4nkd left18:31
SethRobertson slicslak: git diff master...feature (or is it ..? Try both until you find the right set)18:31
.. definitely18:32
→ chisler joined18:32
← OOPMan left18:33
→ ceej joined18:34
→ preflex_ joined18:35
← preflex left18:35
preflex_ → preflex18:36
← vacho left18:36
→ thiago joined18:36
← oc80z left18:36
← tvw left18:37
Mikachu actually he wants ...18:37
→ khmarbaise joined18:37
→ flijten joined18:37
← alberto56 left18:37
→ oc80z joined18:38
→ rath joined18:41
gitinfo set mode: +v18:41
rath is there actually a way to use git's end-of-line normalization with CRLF instead of LF?18:42
cmn how do you mean?18:42
← Heimidal_ left18:42
cmn you can tell git which one you want, can't you?18:42
rath yes, but not for normalization of existing files18:43
you've to normalize them one time with LF18:43
or in other words...you have to normlize them one time...and it is LF18:44
but our files are mostly CRLF, so in the first normalization it would change the most files...18:44
← coachz left18:44
← oc80z left18:45
rath there is no way to change it (by configuration?) to CRLF, right?18:46
→ srcerer joined18:46
← Skyx-mobile__ left18:46
← eno left18:46
→ oc80z joined18:46
cmn if you want your files to be in CRLF, why not store them that way?18:47
← berserkr left18:48
← Chibby_ left18:48
→ eno joined18:48
rath we have stored it that way, but we want to ensure that any other commit doesn't change it18:49
→ OOPMan joined18:49
rath and happens many times because of different OS's18:49
→ g105b joined18:50
SethRobertson Enforce it in a receive hook on the server. That is your only sure bet18:50
cmn then tell git on the other OSs not to mess with line endings18:51
if people complain about line endings, they should come back from the 90s18:52
← cakehero left18:53
← Ryback_ left18:53
rath the problem is that the IDEs or build management systems likes to change it...18:54
cmn configure them not to18:54
← airborn left18:55
→ dfr|mac_ joined18:56
→ Invader joined18:56
Syco54645_work cmn: sublimetext 2 has a git integration plugin18:57
rath many windows tools can't be configured to do this, so git's normalization is actually the way, but it does LF18:57
→ elmasry joined18:57
← Milossh left18:57
cmn why are windows tools writing LF?18:58
rath is there a way to change it to CRLF or make it configurable?!18:58
git normalization does18:58
cmn then don't normalise it18:58
← duckx left18:58
→ duckxx joined18:59
← rtconner left18:59
← dfr|mac left18:59
← xsism left18:59
→ alberto56 joined19:00
← kermit left19:00
→ papegaaij joined19:01
rath one part of the tools does CR, another part does CRLF. so we have to normalize, but not LF19:01
thiago no one does CR-only19:01
that's Mac OS 919:01
← jasonbray left19:02
rath sorry, i meant LF19:02
← eno left19:03
→ xsism joined19:03
rath it would be nice if git gets a configuration for that19:04
→ eno joined19:04
→ zulax joined19:04
← babilen left19:05
→ jasonbray joined19:05
← bartek left19:05
← anjen left19:06
→ bartek joined19:07
← dnjaramba left19:09
→ astrostl_ joined19:10
← slicslak left19:11
← bitkiller left19:12
← astrostl left19:12
← fire31ight left19:12
→ LucasTizma joined19:12
→ dnjaramba joined19:12
← p3rror left19:12
→ fire31ight joined19:12
→ cybersphinx joined19:13
→ duckx joined19:13
← duckxx left19:13
rath SethRobertson: how can i set up a hook on the server?19:14
SethRobertson !any, but to answer your question, just dump it in .git/hooks as is described in man githooks19:15
gitinfo the githooks manpage is available at http://jk.gs/githooks.html19:15
SethRobertson !any as well19:15
gitinfo [!anyone] Usually, it does not help to ask for someone specific to answer your question. No one is the sole expert on anything, or even if they may be, there are plenty of others who know enough to help you19:15
→ Solow joined19:15
→ joshsmith joined19:15
Solow If I check something out inside of a checkout, will the checkout it's in ignore it?19:15
SethRobertson No19:16
← the_cybersphinx left19:16
Solow SethRobertson: Is it possible to make it ignore it? I don't want to pollute the "parent" checkout.19:16
Mikachu .gitignore or .git/info/exclude as usual19:16
rath SethRobertson: ah ok, i know it but saw it in a wrong situation, thx19:17
← ajaykools left19:17
Solow Not that usual for me, haven't worked with it that much yet. Thanks.19:17
SethRobertson Describe more what you are talking about, what command are you using to do the nested checkout19:17
← thiago left19:17
Solow SethRobertson: Just a simple git clone, cd into that directory, and do another git clone of something else.19:17
Mikachu well, it's not different from any other untracked file/folder you want to ignore19:17
SethRobertson Solow: !subprojects19:17
gitinfo Solow: 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.19:17
← eno left19:18
Mikachu shouldn't he suggest to /msg the commands instead? :)19:18
Solow SethRobertson: gitignore it is19:18
I'll google that one. Merci19:18
SethRobertson man gitignore19:18
gitinfo the gitignore manpage is available at http://jk.gs/gitignore.html19:18
yrlnry I thought that if you put repo A inside repo B, that commands in B would automatically ignore files in A because they would see the A/.git directory.19:19
← JStoker left19:19
→ Retistic joined19:19
SethRobertson Mikachu: Perhaps, but it sometimes helps to know whether someone did some more research. Perhaps syntax so that you can ask the question publicly but get the answer privately?19:19
Mikachu heh19:19
→ eno joined19:19
Mikachu yrlnry: yes, but A won't ignore B19:20
SethRobertson yrlnry: Mostly not. `git clean` does, I know, but submodules otherwise complicate that model19:20
→ JStoker joined19:20
← dnjaramba left19:20
SethRobertson I take that back. My most recent git is not ignoring during clean19:21
← zulax left19:21
← Syco54645_work left19:21
→ Milossh joined19:21
← zivester left19:21
→ zivester joined19:22
→ p3rror joined19:24
→ cakehero joined19:24
Solow Cool thanks. I used the exclude.19:25
→ dnjaramba joined19:27
← rath left19:27
→ Goplat joined19:27
→ iocor joined19:27
← Goplat left19:27
→ Goplat joined19:27
← iocor left19:27
→ iocor joined19:27
← khmarbaise left19:28
→ vacho joined19:29
vacho how do I tell git to untrack a file it's already tracking??19:29
→ shruggar joined19:30
← Vile left19:30
SethRobertson git rm [--cached]19:30
See also !config_files, which may or may not apply to your case19:30
gitinfo [!configfiles] the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.19:30
← astrostl_ left19:30
→ hobodave joined19:31
→ sythe joined19:31
← dnjaramba left19:31
← rbrooks left19:31
← shruggar left19:31
← eno left19:32
← abetaha left19:32
vacho SethRobertson: git rm --cached sites/default/settings.php19:32
fatal: Not a git repository (or any parent up to mount parent )19:32
→ michaelh joined19:32
← fayimora left19:32
SethRobertson Where is your git repo? In sites?19:33
vacho im not sure, how do I check?19:33
SethRobertson find . -name .git19:33
→ eno joined19:33
mjt heh19:34
vacho ./Sites/dccomics/.git19:34
→ shergill joined19:34
vacho Sites/sitename/.git **19:34
SethRobertson If that is the only output, then, err, sites/default/settings.php is not under version control19:34
→ abetaha joined19:35
vacho it is!19:35
SethRobertson If there is a sites/default/.git, then cd to that directory and run the git command19:35
vacho im looking at my local cloned git19:35
SethRobertson Then delete it in your local cloned git19:35
vacho that deletes it remotely too19:36
SethRobertson All too true.19:36
You can look at the URL for your local cloned git and find that directory on the server19:36
→ Wombert joined19:37
→ dnjaramba joined19:37
vacho sorry SethRobertson im not following 100%19:37
→ panpainter joined19:37
vacho if i want to untrack a file..do I need to do it from my local git or remotely19:37
RandalSchwartz you can only "untrack" locally19:38
← LucasTizma left19:38
SethRobertson You can do it in either location, but you can only save the file on the system you run the command in (--cached)19:38
RandalSchwartz then commit that, and push a new commit19:38
vacho ok..I want to untrack settings.php so everytime I make a change to it it does not get pushed remote19:38
SethRobertson !config_files19:38
gitinfo [!configfiles] the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.19:38
RandalSchwartz if you untrack it, it'll be deleted remotely19:38
if you push again19:39
basically, you're saying "don't look at this file at all"19:39
→ dfr|mac joined19:39
vacho I do not want it deleted remotely, then how do I do it?19:39
RandalSchwartz see !configfiles19:39
gitinfo the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.19:39
vacho I read that post, didn't help me.19:39
RandalSchwartz git mv config.php config.php.sample19:39
git commit -a -m 'renaming sample config'19:40
then have an installer that cp config.php.sample config.php19:40
so that it can be edited locally without hurting the upstream19:40
cmn you can try git update-index --assume-unchanged settings.php but that can be dangerous19:40
SethRobertson Very dangerous. You are lying to git. It is described on that link. Very very much not recommended19:40
cmn having a way to specify local configuration is a very useful thing to have at any rate19:40
vacho ok19:41
→ mxweas_ joined19:41
← dfr|mac_ left19:43
← dnjaramba left19:45
→ DarkAR joined19:45
→ infogulch joined19:46
← Destos left19:46
→ adiabatic joined19:47
→ MrKeuner joined19:47
← MrKeuner left19:47
→ MrKeuner joined19:47
MrKeuner hi, git branch --set-upstream dev master returns "Branch dev set up to track local branch master" I was intending to track remote dev's master19:48
→ linusoleander_ joined19:49
← linusoleander left19:49
linusoleander_ → linusoleander19:49
elmasry *Hi*, I'd like to use github, add my public ssh key, works fine, I have my local repository, but when I try to 'git push origin master' it gives: 'fatal: ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDegixhC39yszxoehyFx+JHQA2RypdNDtg8BIuK29EhPrxkazMNUu6+JKsSEljW87+vQ86B76CnBv6TAK0Hh8yZlm7sf5gn4Gui9oh+S9lw0IJJQJHnct4/SwSg1GMwFAYJRQvN3SKP7hzrg+G4M8MkF7lV11TZlM55054PhUHg1xZZmxSXM0LNmFa531qlOk8cq9zIRlpZnmAwiIUENpIBF74lOI0JRuUKlKfwMmrrSa0uHV/X8KbgnAz4Piw++Zp+ziBEC0PnKx2P4WP19:50
oops sorry19:50
→ di-five joined19:50
← eno left19:50
SethRobertson MrKeuner: Then say "dev dev/master"19:50
→ dnjaramba joined19:50
elmasry it gives me: 'fatal: the remote end hung up unexpectedly'19:50
→ delinquentme joined19:51
delinquentme hey all how do I edit the commit messages that I pull up when using " git log "19:51
SethRobertson elmasry: What does `git remote -vv; git branch -vv` say?19:51
delinquentme question mark.19:51
talexb elmasry, I'm guessing it's an authentication problem ..19:51
→ unnamedwill joined19:51
SethRobertson delinquentme: See !fixup19:51
gitinfo delinquentme: 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)19:51
→ eno joined19:52
→ Raging_Hog joined19:52
→ fayimora joined19:52
elmasry talexb: I tried: 'ssh -T [email@hidden.address] and worked fine19:52
← fayimora left19:52
← BlaDe^ left19:53
delinquentme so rebase works just fine .. however the commit messages in gitk and git log are different19:53
talexb elmasry, Don't know what else to suggest. Contact github maybe?19:54
SethRobertson delinquentme: Did you restart gitk?19:54
elmasry SethRobertson: origin git@github:emz/Algorithmic-Form-Generation.git (fetch)19:54
origin git@github:emz/Algorithmic-Form-Generation.git (push)19:54
delinquentme yeah19:54
← g105b left19:54
SethRobertson delinquentme: Then delete .git/gitk.cache19:54
delinquentme gitk looks good .. git log does not19:54
SethRobertson elmasry: I cannot help but notice the lack of ".com" in your URLs19:55
delinquentme git log contains all the commit messages of the commits i squashed with rebase19:55
SethRobertson delinquentme: Did you run that command and it still doesn't work?19:56
Oh, wait, I misread.19:56
delinquentme: What does `git branch` report? Are you on a * branch?19:56
delinquentme * master19:57
→ tatsuya_o joined19:57
SethRobertson And if you delete .git/gitk.cache and type `gitk master` you see something different than `git log master`?19:57
← Milossh left19:58
← zendeavor left19:58
← dnjaramba left19:58
← mxweas_ left19:58
← vacho left19:59
→ shruggar joined19:59
← Fangorn_ left19:59
delinquentme correct SethRobertson20:00
SethRobertson delinquentme: That is, err, disturbing. Is this a repo you can make public? Have you tried upgrading git?20:01
→ devth joined20:01
delinquentme here is the git log http://pastie.org/3350267 ... and here is gitk http://i.imgur.com/SAJbw.png20:02
no i havnt upgraded git20:02
devth is git-request-pull the only way to show a branch's "description"?20:03
MrKeuner SethRobertson, git branch --set-upstream dev dev/master returns "fatal: Not a valid object name: 'dev/master'"20:03
→ dnjaramba joined20:03
gitinfo set mode: +v20:03
SethRobertson MrKeuner: Please pastebin: git branch -avv20:04
→ robotmay joined20:04
touk what happens when you push changes to a repo that conflict with someone else's push?20:04
cmn devth: git config branch.$branch.description also will20:04
RandalSchwartz touk - generally, you can't20:04
if it's "fast forward only" by default, your push is rejected.20:05
SethRobertson delinquentme: I don't obviously see what the difference is20:05
cmn touk: conflict in what way?20:05
← Wombert left20:05
devth cmn: ty20:05
MrKeuner SethRobertson, http://pastebin.com/jiPznjwR20:05
→ pmenon joined20:05
cmn touk: the commits are made locally, when you push, you're asking the server to push the branch forward20:05
SethRobertson MrKeuner: Do you have a remote set up? Is `git remote` empty?20:06
cmn if what you're trying to push isn't a desdendant of what is there, git will complain20:06
→ slicslak joined20:06
cmn you can force the push, overwriting the history whoever else wrote20:06
MrKeuner SethRobertson, no there are three remotes push and pull for each20:06
fetch and push20:06
← hydrozen left20:06
→ BiggFREE joined20:07
delinquentme SethRobertson, in the git log commit message a the top20:07
→ zendeavor joined20:07
MrKeuner SethRobertson, 1.7.1 here, is it a problem20:07
delinquentme Id like the message to not have the more more and added readme formatting20:07
touk RandalSchwartz cmn - so if you push and you would be overwriting a change you didnt get that someone else pushed to the server, it will complain, and then you will have to do a pull, merge conflicts, then push?20:07
elmasry SethRobertson: OK thanks, problem solved.20:08
cmn delinquentme: not have it or not see it by default?20:08
→ kadoban joined20:08
cmn the first paragraph is the subject of the commit20:08
and that's what both gitk and git log --oneline show20:08
touk RandalSchwartz cmn - also, if someone pushes something they shouldnt have, can you revert their push?20:08
→ jlk joined20:08
← dnjaramba left20:08
MrKeuner SethRobertson, push.default is set to tracking20:09
jlk Ok, so I know git has a way of fairly easily telling what tags are reachable by which branches or commits, because you can fetch a specific branch and only get the associated tags20:09
→ dnjaramba joined20:09
SethRobertson delinquentme: I cannot see the full message in the gitk output, however the first line looks the same.20:09
jlk what I'm looking for is the ability to list out what tags are reachable from a given branch, but I'm not finding any good way of doing this.20:09
delinquentme so how do i edit commit 9bfb8d9bce96 .. so that the commit message is *ONLY* "formated readme file for github markdown"20:09
cmn git commit --amend20:09
← krz left20:10
cmn touk: right, your client will refuse to push non-ff history20:10
SethRobertson MrKeuner: Try `git remote update` and then re-run `git branch -a` You should see new branches show up20:10
delinquentme lol i SWEAR I tried that20:10
wald0 github has a limit of 300 mb of space only ?? i was thinking on made all my works from it... :(20:10
→ cccaldas joined20:10
delinquentme thanks guys :D20:10
touk cmn - i think i know the diff between a ff and a non ff merge, but whats non-ff history?20:11
cmn touk: so you should make it a descendant, either by rebase or merge or whatever20:11
pretty much the same thing20:11
fast-forward means one is a descendant of the other20:11
MrKeuner SethRobertson, I did git remote update but git branch -a still shows *dev and master20:11
cmn wald0: it's a soft limit20:11
touk cmn - ok, so in order to make it a descendent you just need to pull the changes down, merge the conflicts, then you should be able to push20:11
cmn to stop people from using it as file storage, if you really need more, you can discuss it with the support team20:12
touk: yes20:12
SethRobertson MrKeuner: Try `git ls-remote <remotename>` Do you see thinkgs like refs/heads/....20:12
MrKeuner: oh!20:12
touk cmn - great thanks, still trying to wrap my mind around all this20:12
MrKeuner no emtpty20:12
it would be my first push if that push had worked20:12
SethRobertson MrKeuner: Ah, that is the problem. `git push -u dev dev:master` and everything should start working.20:13
MrKeuner SethRobertson, git branch --set-upstream dev dev/master requires dev/master to exist which requires an initial push -u?20:14
wald0 cmn: so it is really not a limit of size if i really need more ? i say that because if is not the case i stop thinking my plans about put everything there (source codes, yes)20:14
MrKeuner SethRobertson, yes it works20:15
→ ASIRINeiluj joined20:15
cmn if you can convince them it's a good idea to give you more sure20:15
MrKeuner SethRobertson, thanks20:15
cmn I don't know the exact policy20:15
someone put up a webkit mirror there IIRC and that's definitely more than 300MB20:15
→ Sonderblade joined20:16
cmn most people don't need anywhere near that much20:16
→ djMax joined20:16
← Sonderblade left20:16
djMax I cloned a github repo and made some changes. I *then* forked it github to my own account. How do I push the changes to my new fork?20:16
→ Sonderblade joined20:17
→ danner joined20:17
→ Fandekasp joined20:17
cmn man git remote20:18
gitinfo the git-remote manpage is available at http://jk.gs/git-remote.html20:18
← di-five left20:18
→ anjen joined20:19
danner hello. i'm trying to generate a unified patch/diff from a sequence of commits, all by me, when a number of other commits have happened in the meantime by others. git log --author=danner was used, but i'm not sure what else i can do to help me - maybe cherrypick them all into a branch?20:19
djMax I want set-head or set-url?20:19
← Lgb left20:20
pom When I tabcomplete: http://pastebin.com/yxjztAAb20:20
← brian_g left20:20
cmn either add or set-url20:20
pom How does git know about "asenine"?20:20
cmn why would you want set-head?20:20
→ brian_g joined20:21
cmn pom: git doesn't20:21
it's the autocompletion20:21
← herriojr left20:21
← eno left20:21
← dorisabayon left20:21
← linusoleander left20:21
cmn git has nothing to do with that20:21
← gchristensen left20:21
← iocor left20:21
djMax set-url worked, thanks20:21
← lorian left20:22
danner or, perhaps a diff on a set of files from one commit to HEAD? other authors worked on other files, but I was the only one to work on my set of files.20:22
CareBear\ is there also git get-head ?20:22
pom cmn: Alright. So, any idea where this comes from? I'm currently on Fedora install of git, and think I installed a git-completion package.20:22
← papegaaij left20:22
→ Milossh joined20:22
pom And in my case, what is "asenine" here? An "origin"?20:23
cmn CareBear\: are you making a joke or asking for real?20:23
→ eno joined20:23
cmn pom: git-completion has a very high likelyhood of being where your completions come from yeah20:23
→ Lgb joined20:23
CareBear\ cmn : joke20:23
cmn asenine looks to be a rmote20:23
CareBear\: I thought so at first, but then I couldn't decide20:23
pom: origin is the default name for the remote you clone from20:24
→ Vile joined20:24
pom Okay, so "asenine" in this case is a remote.20:24
cmn right20:24
pom cmn: Okay, thanks. I'll sort it out now.20:25
← mishok13 left20:26
← Invader left20:26
→ herriojr1 joined20:26
jaalto What with github.com? Am I the only one that can't acess the site?20:27
← i8degrees left20:27
← sebbi left20:27
talexb jaalto, Works for me.20:28
cmn http://www.downforeveryoneorjustme.com/github.com20:28
← anjen left20:28
cmn also https://status.github.com/ and #github20:29
→ khmarbaise joined20:29
Nathan|dnd → Nathandim20:30
← lacrymology left20:30
→ LongBeach joined20:30
← LongBeach left20:31
← jasonbray left20:31
← BryanRuiz1 left20:31
← asakura left20:31
→ asakura joined20:33
← erichynds left20:34
← cccaldas left20:34
→ Heimidal joined20:35
← Solow left20:35
← robotmay left20:35
→ robotmay joined20:35
← revagomes left20:36
→ no1home joined20:36
→ babonk joined20:37
babonk How can I find the most recent version of a file by name? The file is currently deleted20:38
I'd prefer not going through gitk to find the commit code20:38
→ acl_ joined20:39
cmn git log -- file20:40
→ vrillusions joined20:41
← marthinal left20:41
← ISF left20:41
← eno left20:41
→ lacrymology joined20:42
→ eno joined20:43
→ diegoviola joined20:43
← robotmay left20:44
← elmasry left20:44
→ disappearedng joined20:45
→ methods joined20:46
← sec^nd left20:46
→ ISF joined20:46
methods how would you see just changes from 1 revision20:46
PerlJam git show REV20:47
MrKeuner sometimes git status shows file as modified but git diff reports identical file, why would that be? permissions?20:50
← touk left20:51
devth MrKeuner: depends on if they've been staged or not. try git diff --staged20:51
→ linkedinu joined20:51
cmn MrKeuner: there are several levels of modification; what exactly does it say?20:52
← glguy left20:52
→ pfFredd joined20:52
tizzo-afk → tizzo20:52
MrKeuner Changed but not updated20:52
← babonk left20:52
methods PerlJam: thanks i was hung up using diff20:52
→ glguy joined20:52
← glguy left20:53
MrKeuner no changes added to commit20:53
← methods left20:53
← djMax left20:54
← linkedinyou left20:54
cmn so changes between worktree and index, then?20:54
if you have changed the permissions, that's a change20:54
and there are no textual changes, so diff won't show anything20:55
MrKeuner cmn, yes I have changed permissions, but I thought git did not care for permissions20:55
cmn for some20:56
it tracks the user execute bit20:56
MrKeuner some?20:56
oh ok20:56
← gavin_huang left20:56
cmn that's the only one that's relevant for anything20:56
← lacrymology left20:56
→ sec^nd joined20:57
EugeneKay If you want permissions tracked in git, do it in userspace. getfacl . > permissions.txt works well.20:57
MrKeuner I don't care about them for now, but thanks for the pointer20:58
→ mjs2600 joined20:58
→ lacrymology joined20:58
→ diogogmt joined20:59
→ duckxx joined20:59
→ fr0stbyte joined20:59
← disappearedng left20:59
← duckx left20:59
→ disappearedng joined21:00
→ disappea_ joined21:01
← disappearedng left21:02
← Jackneill left21:02
← lyddonb left21:02
→ MariusKarthaus joined21:02
← khmarbaise left21:03
→ zulax joined21:04
→ lyddonb joined21:05
mjs2600 Hi, I'm kind of new to git, but I having a really annoying problem. I want one git repository to backup all of my code, but git won't add a directory if it's part of another git project, which most of my code is. Is there anyway around this Or am I just approaching this the wrong wany and missing some terrible inefficiency in doing it this way?21:05
← oc80z left21:05
→ di-five joined21:05
→ aristidesfl joined21:06
cmn why do you have the outer repo?21:06
you could use subprojects, but I don't see how that helps backpus21:07
→ oc80z joined21:07
mjs2600 so a backup script can keep track of all of my work21:07
SethRobertson mjs2600: !backups21:08
!backup21:08
gitinfo Taking a backup of a git repository is always a good idea, especially when taking advice over IRC. Usually, the best way to TACTICALLY back up a git repo is `git clone --mirror`. However, some unusual maintenance might require `tar cf repo-backup.tar repodir`. Testing in a clone is also an excellent idea. See also http://sethrobertson.github.com/GitBestPractices/#backups21:08
SethRobertson The link provides a scriptlet which might help21:08
mjs2600 Thanks!21:08
→ tangledhelix joined21:08
→ glennpratt joined21:09
→ savid joined21:11
→ jasonbray joined21:11
savid How can I see what my remote is currently set to without actually querying the remote?21:11
SethRobertson Define "set to"21:11
savid It seems to just hang when I do "git remote show origin"21:11
SethRobertson git remote -vv?21:11
savid SethRobertson, ah, thanks that's what I neede :)21:12
← linkedinu left21:12
→ linkedinu joined21:12
← eno left21:15
← lacrymology left21:15
→ btanaka joined21:15
→ eno joined21:16
→ soidexe joined21:17
→ lacrymology joined21:17
← mjs2600 left21:18
← diogogmt left21:18
← merlin83_ left21:18
→ diogogmt joined21:19
→ mxweas_ joined21:19
← Raging_Hog left21:21
← heptadecagram left21:21
← norton- left21:21
→ xerora joined21:21
← dr_win left21:21
→ number_number joined21:21
← di-five left21:21
← eno left21:22
→ aspotashev joined21:23
→ retoryter joined21:23
← chisler left21:23
→ iocor joined21:24
→ Jeremy_ joined21:24
Jeremy_ → Guest5861521:24
→ eno joined21:24
→ joch_n joined21:24
← rendar_ left21:25
← brian_g left21:25
ChanServ set mode: +v21:25
← glennpratt left21:25
Guest58615 → GitCurious21:26
← Goplat left21:26
→ brian_g joined21:26
GitCurious hi21:26
SethRobertson !hi21:26
gitinfo [!welcome] Welcome to #git, a place full of helpful gits. If you have a question, feel free to just go ahead and ask—somebody should answer shortly. For more info on this channel, see http://jk.gs/git/ Take backups (type !backup to learn how) before taking advice.21:26
EugeneKay Who added that last bit? Do you doubt our good advice? ;-)21:27
← lacrymology left21:27
→ glennpratt joined21:27
cmn aren't gitinfo's triggers managed in git?21:27
← disappea_ left21:27
cmn if not, shame on us all21:27
GitCurious I'm looking to move our repository from svn to some decentralized system, at that point I'm hesitating between git and mercurial21:27
EugeneKay No, SQLite DB, I think.21:28
https://github.com/jast/gitinfo21:28
→ disappearedng joined21:28
← Dave^|| left21:28
GitCurious does somebody have some killer argument in favor of Git ? :)21:29
→ tgunr_ joined21:29
EugeneKay It turns you gitsexual21:29
SethRobertson !isgitbetter21:29
gitinfo we think so, but if you don't, please go away and use $YOURFAVOURITESYSTEM. See http://whygitisbetterthanx.com/ among many many other places.21:29
EugeneKay GitCurious: http://www.whygitisbetterthanx.com/21:29
→ disappea_ joined21:30
number_number the output of "git diff branch file" shows me the differences between my file on branchA and my current branchB.. I'd like to take one or two improvements from branchA and put them in branchB.. but not all.. I'd ideally like to open my difftool and copy over the bits I like..21:30
← disappea_ left21:30
→ ihodes joined21:30
← disappearedng left21:30
number_number but otherwise I'd like to keep to two branches completely seperate.. if that makes sense21:30
GitCurious I'm new to decentralized SCM, the only things that makes me prefer mercurial over git are the changeset + numbered revision21:30
cmn checkout -p branch file might do what you want21:30
→ disappearedng joined21:30
SethRobertson GitCurious: One company I know went with Mecurial since it reduced the amount of power their developers had. They didn't want developers to be able to fix up problems in old commits if they happened to discover them *before* pushing. Strange but true.21:30
GitCurious is being gitsexual some kind of goal?21:31
thx Eugene, checking that link21:31
→ linusoleander joined21:32
cmn the revision numbers are only ever local; git allows you to specify commits relative to each other21:32
so you can do much the same21:32
→ kergoth joined21:32
SethRobertson GitCurious: I'm pretty sure that any distributed SCM is not going to have one constant integer for a revision which cannot change. Specifically, the number is only fixed relative to certain other information.21:32
Sequential integer at least21:33
GitCurious @Seth right that is weird!21:33
number_number cmn: wow, that's absolutely perfect \o/21:33
← retoryter left21:33
wereHamster SethRobertson: '!gitisbetter' reads as 'not gitisbetter' =~ 'git is not better', feels weird. It should be !gitisntbetter, dontyathink?21:34
or !gitisworse21:34
adiabatic !gitisntbetter has too many negatives21:34
GitCurious yeah , I'm not saying revision number is anything revolutionnary ... it's just that I'm currently a svn user, which is already a big improvement as I used to work with Perforce before :/21:34
Mikachu that is !helpful21:34
SethRobertson There are a number of places in gitinfo where the ! trigger sometimes makes sense when thought of as "not" and where it does not21:34
← disappearedng left21:35
← mcluskydodallas left21:35
cmn GitCurious: svn revision numbers are global, hg's are local21:35
it's not the same thing at all21:35
→ mcluskydodallas joined21:35
SethRobertson GitCurious: You can compute a number with git if you want to, or just bite the bullet and use the globally unique SHA revision id. It is surprisingly easy to use after a while21:35
← delinquentme left21:35
cmn and you rarely need to do that anyway21:36
→ merlin83_ joined21:36
← jason237 left21:36
SethRobertson Any place where you are using a revision number for something important, you probably should be using a tag anyway, either that or the SHA would work just the same21:36
← matthiaskrgr left21:37
← glennpratt left21:37
GitCurious yeah that's right, I just like the fact that an integer is human friendly ... not much of a technical consideration21:37
← kevwil left21:38
← dr_lepper left21:38
← mxweas_ left21:38
vrillusions where I'm working there was a LOT of push-back. One of the ones being are db patch system was based off the increment revision id's of svn. Of course we redid things so it works21:38
→ matthiaskrgr joined21:39
vrillusions GitCurious: protip, you don't need to know the entire hash, only the first, say 5 or so21:39
so I'll often here the dialog "what's your revision at?" "ec829" "OK that's what mine is"21:40
git works the same way (without the yelling). `git lgo ec829` would work21:40
→ Bass10 joined21:40
wereHamster GitCurious: for what do you think you need the numbers? To tell your coworkers which commit you are talking about? Or when browsing the history locally?21:40
← pidus left21:40
vrillusions er, git log21:41
→ wilmoore joined21:41
→ Cromulent joined21:42
← fire31ight left21:43
→ fire31ight joined21:43
GitCurious we dont use them yet ... but I'm trying to think a bit further ... and it seems to me that referring to revisions numbers to settle some versioning scheme might be interesting. Also for those things you say actually wereHamster21:44
cmn if you need to version your schemes, use a version for them21:45
vrillusions GitCurious: probably if you get a need to reference a commit then you'd probably want to make a tag. Unlike svn tags in git just point to a commit, not a whole copy of the code base. So you can make as many tags as you want and it doesn't fill disk space (locally, server side svn uses hard links and stuff)21:46
→ dr_win joined21:46
→ rasmusth joined21:47
rasmusth error: unable to unlink old 'sites/default/default.settings.php' (Permission denied)21:47
does anybody know this error21:47
→ jasonb joined21:48
wereHamster rasmusth: what do *you* think it means?21:48
GitCurious that is a possibility as well right ... I'll give it some thought, thanks vrillusions.21:48
← dr_win left21:48
← fisted left21:49
← soc42 left21:49
cmn rasmusth: have you seen the error messages in the parentheses?21:50
← ISF left21:51
← asakura left21:52
← jonesy left21:53
GitCurious have to go, will probably hang around here those days ... thanks and see you!21:53
← GitCurious left21:53
→ fisted joined21:54
← Chillance left21:54
vrillusions rasmusth: `ls -l sites/default` will also give a clue21:54
→ heptadecagram joined21:55
→ t4nkd joined21:55
→ ko1 joined21:56
← oc80z left21:56
→ AbsLacKWare joined21:56
← shruggar left21:57
AbsLacKWare hi... someone from brazil ?21:57
→ glennpratt joined21:58
← lars___ left21:58
→ Aktom joined21:58
← Aktom left21:58
→ oc80z joined21:58
← jasonbray left21:58
← Cromulent left21:59
← joch_n left21:59
← AbsLacKWare left21:59
→ glguy joined22:01
wereHamster why would that be relevant?22:03
← Zenopus left22:04
← whitman left22:05
→ herriojr2 joined22:05
← ChoiZ left22:05
→ cp_clegg joined22:05
→ tzervo joined22:06
← tommyvyo left22:06
→ ChoiZ joined22:06
→ disappearedng joined22:07
← herriojr1 left22:08
→ giallu joined22:10
← diogogmt left22:11
← _iron left22:11
EugeneKay Something something language other than English22:11
→ nurey joined22:11
→ asakura joined22:11
→ preflex_ joined22:12
← canton7 left22:12
← preflex left22:12
← xsism left22:12
→ diogogmt joined22:12
preflex_ → preflex22:12
← joshsmith left22:12
cmn but any language they speak in Brazil they speak in many other places around the world22:14
← churp left22:14
yrlnry Surely there are all sorts of obscure Amazonian languages that are not widely spoken elsewhere.22:14
→ gchristensen joined22:15
cmn this is true22:16
though in that case "Brazil" would also be the wrong place to ask for22:16
← heptadecagram left22:17
← kergoth left22:17
← stuntmachine left22:18
← linkedinu left22:19
← BiggFREE left22:21
EugeneKay belches loudly22:21
MrKeuner :)22:21
→ linkedinyou joined22:21
← flijten left22:23
→ frogonwheels joined22:24
← gchristensen left22:25
← QaDeS left22:25
← TrineLocke left22:26
← diogogmt left22:29
→ diogogmt joined22:30
← jlk left22:32
← nurey left22:33
← digitalfiz left22:33
→ LucasTizma joined22:33
→ canton7 joined22:33
→ EricInBNE joined22:35
→ linkedinu joined22:35
← ThomasLocke left22:36
← glennpratt left22:36
← linkedinu left22:36
→ Francisc joined22:38
→ Inda5th joined22:39
← linkedinyou left22:39
Francisc If I git reset HEAD~, will I be allowed to push to remote origin, or is it the same as git commit --amend ?22:39
← nazgul101 left22:40
→ glennpratt joined22:40
chrisf Francisc: it's a history change.22:40
→ drizzd joined22:41
→ duckx joined22:41
← duckxx left22:41
← axl_ left22:41
← iocor left22:41
← spiette left22:42
← boombatower left22:42
Francisc Hey. I know that, but I'm still not sure how it affects push.22:43
bob2 push by default doesn't let you do anything other than /append commits/22:44
→ thiago joined22:44
→ axl_ joined22:44
gitinfo set mode: +v22:44
bob2 if you alter your branch such that it isn't just remote+some more commits, you can't push22:44
← axl_ left22:44
bob2 so reseting into the past is fine, but any commit after that means you cannot push22:44
→ erichynds joined22:45
← glguy left22:45
→ axl_ joined22:45
Francisc Got it. Thanks.22:45
← axl_ left22:45
→ axl_ joined22:46
← Francisc left22:47
← macUzer left22:47
→ nazgul101 joined22:47
← dreiss left22:49
→ cooldude joined22:49
→ dreiss joined22:49
→ gchristensen joined22:50
← n8o-mba left22:51
← snearch left22:52
← iamjarvo left22:52
→ tatsuya__ joined22:53
← tatsuya_o left22:53
bob2 hot22:53
← fr0stbyte left22:53
bob2 git's model is reasonably simple i think22:54
once it clicks22:54
→ eletuchy joined22:54
cmn it seems the hardest part is realise the lack of magic working in the background22:54
← disappearedng left22:55
ChanServ set mode: +v22:55
→ disappearedng joined22:55
← martinjlowm left22:55
→ joshsmith joined22:55
← ipalaus left22:56
← bgerber left22:56
← dcarper left22:56
← lyddonb left22:56
← pmenon left22:56
← akosikeno left22:57
bob2 heh22:57
← shurane left22:57
iheffner sufficiently advanced technology...?22:57
← martinjlowm-lptp left22:57
bob2 the other thing people seem to struggle with is that packs/deltas are an implementation detail and to you it's just that git snapshotted everything ever22:57
talexb .. thank you Arthur C Clarke.22:57
← disappearedng left23:00
→ iamjarvo joined23:00
→ berserkr joined23:01
← macmartine left23:01
→ MistuhKurtz joined23:02
← t4nkd left23:02
→ bgerber joined23:03
← Sonderblade left23:03
→ EspadaV8 joined23:03
EspadaV8 hi, has anyone tried using git svn over a CIFS and/or NFS share? I'm currently getting the following error - http://dpaste.org/frdOO/23:03
← LeMike left23:05
← tatsuya__ left23:05
→ tatsuya_o joined23:05
← john2496 left23:05
→ dr_win joined23:05
← gusnan left23:06
cmn it looks like git gives up too early23:07
or that NFS is too crappy an FS23:07
← iamjarvo left23:07
→ disappearedng joined23:08
← diogogmt left23:09
← ko1 left23:09
← jasonb left23:09
← tangledhelix left23:09
← kadoban left23:09
← Retistic left23:09
← mikepack left23:09
← orospakr left23:09
← Dreamer3 left23:09
← TeckniX left23:09
← T_X left23:09
← gjer984 left23:09
← gxk left23:09
← daethorian left23:09
← madsy left23:09
← mtkd left23:09
← JackWinter left23:09
← DrHouse|TARDIS left23:09
← jpalmer left23:09
→ gjer984 joined23:09
→ orospakr joined23:09
→ jasonb joined23:09
→ gxk joined23:09
→ jaisoares joined23:09
→ madsy joined23:09
← madsy left23:09
→ madsy joined23:09
→ JackWinter joined23:09
→ Retistic joined23:09
→ kadoban joined23:09
→ diogogmt joined23:09
→ Dreamer3 joined23:09
→ mikepack joined23:09
→ ko1 joined23:09
→ tangledhelix joined23:09
→ daethorian joined23:09
→ mtkd joined23:10
→ jasonbray joined23:10
→ shiba_yu36 joined23:10
→ bradly joined23:11
← MariusKarthaus left23:11
tizzo → tizzo-afk23:11
← amgarching left23:12
← crisp left23:13
→ TeckniX joined23:14
→ T_X joined23:14
← akiress left23:14
← T_X left23:14
→ T_X joined23:14
← tzervo left23:14
→ akiress joined23:14
← alx- left23:14
→ amgarching joined23:15
→ iocor joined23:16
wald0 how i can know the last commit id-revision (or date) from an external repository ? (without being myself in a clone of it)23:17
canton7 wald0, git ls-remote23:18
wald0 so, the equivalent of "svn info svn://domain..." from anywhere23:18
thx23:18
→ tcurdt joined23:18
→ dorisabayon joined23:18
← tcurdt left23:18
← tewecske left23:19
cmn maybe, depends on what svn info does23:19
ls-remote tells you the refs that the remote offers, nothing more23:19
well, you can guess the git version from the extensions23:20
← fedesilva left23:20
→ plusk joined23:21
← PySloan left23:21
← Fandekasp left23:21
→ glennpratt_ joined23:22
← MacGyverNL left23:24
← simenbrekken left23:25
← glennpratt left23:25
← pfFredd left23:26
→ jason237 joined23:27
→ mdel_ joined23:29
← ehsan left23:29
← i42n left23:29
← panpainter left23:29
→ mitsuhiko joined23:30
mitsuhiko hey everybody23:30
← Nathandim left23:30
mitsuhiko My git client creates "Merge branch 'master' of server" whereas the one of a collegue creates "Merge remote-tracking branch 'origin/master'". Anyone know why the difference?23:30
(default commit message)23:30
← Aaaarg left23:31
→ panpainter joined23:31
← panpainter left23:31
iheffner mitsuhiko: you are saying 'git merge master' and your collegue is saying 'git merge origin/master'23:31
kevlarman mitsuhiko: git --version for both?23:31
← aspotashev left23:31
→ panpainter joined23:31
mitsuhiko kevlarman: unfortunately i can't provide that right now since he's offline23:31
in my case it's 1.7.5 and as far as I know the "Merge branch 'master' of server" has been like that for a long time23:32
kevlarman actually iheffner's explanation sounds pretty plausible23:32
mitsuhiko iheffner: i think we both say git pull23:33
nevermind. iheffner is correct23:33
i say git pull, he merges by and23:33
thanks :)23:35
← hobodave left23:35
→ txomon|home joined23:35
← txomon|home left23:35
→ txomon|home joined23:35
← notjohn left23:35
← drizzd left23:37
← txomon|home left23:37
→ nOStahl joined23:38
nOStahl hey guys23:38
trying to setup my new macbook pro as admin machine23:38
bremner "admin machine"?23:39
nOStahl added my new .pub to gitolite keydir and added commited and pushed23:39
← jasonbray left23:39
nOStahl cant clone it down it asks for password23:39
cmn set up ssh so it sends the key to the ssh server23:39
nOStahl ah is osx not doing that ?23:39
my first time owning a mac heh23:40
← dr_win left23:40
RandalSchwartz same as on linux23:40
edit .ssh/config23:40
iheffner nOStahl: you need to make sure you are running an ssh agent and add you key to it23:40
RandalSchwartz make sure you're logging in as user "git" (or maybe "gitolite")23:40
nOStahl ah i have a config file23:40
cmn nOStahl: you're the one who has to tell ssh23:40
it makes no differece what OS23:41
RandalSchwartz Host gitolite.host.example.com23:41
User gitolite23:41
IdentityFile ~/.ssh/foo_rsa23:41
nOStahl aye i got that in my config23:41
RandalSchwartz because you aren't logging in as you... you're logging in as git.23:41
then try "ssh -v that.host.name"23:42
and see what key it offers23:42
nOStahl just tried testing the config with regular ssh23:42
and it asks for password there too23:42
→ rchavik joined23:42
RandalSchwartz something isn't in place then23:42
you edited gitolite-admin ?23:42
← rchavik left23:42
→ rchavik joined23:42
→ pekunz joined23:42
RandalSchwartz and pushed it back?23:42
and your user has access to this other repo?23:42
nOStahl aye I added my new macbooks key from my old dev laptop23:43
and pushed it23:43
RandalSchwartz as given in gitolite.conf23:43
iheffner nOStahl: check the file permissions; .ssh/ should be 070023:43
nOStahl I chmod 700 the .ssh file too already23:43
RandalSchwartz wait- you're already accessing gitolite-admin with some keypair right?23:43
iheffner nOStahl: .ssh/authorized_keys should be 060023:43
→ dr_win joined23:43
cmn so what does -v say?23:43
RandalSchwartz why are you adding a second keypair?23:43
→ stamina joined23:43
→ shurane joined23:44
cmn different computer23:44
→ Element9 joined23:44
RandalSchwartz and youre not using the same "username" for both keys, right?23:44
nOStahl debug1: Authentications that can continue: publickey,password23:44
debug1: Next authentication method: publickey23:44
debug1: Offering public key: /Users/nostahl/.ssh/id_rsa.pub23:44
debug1: Server accepts key: pkalg ssh-rsa blen 27723:44
debug1: PEM_read_PrivateKey failed23:44
debug1: read PEM private key done: type <unknown>23:44
different username23:44
RandalSchwartz weird23:44
← alex__c2022 left23:44
← DarkAR left23:44
← koo4 left23:44
nOStahl there's the tail of what it reports with -v23:45
← bartek left23:45
← canton7 left23:45
iheffner try more -v23:45
← tylere left23:45
nOStahl PEM_read_PrivateKey failed ?23:46
← chrisgeorge left23:46
→ bartek joined23:47
cmn is your key readable? what does -vvv say?23:47
and why on earth are you using a different username?23:47
nOStahl debug3: sign_and_send_pubkey23:47
debug1: PEM_read_PrivateKey failed23:47
debug1: read PEM private key done: type <unknown>23:47
debug2: no passphrase given, try next key23:47
debug2: we did not send a packet, disable method23:47
debug3: authmethod_lookup password23:47
debug3: remaining preferred: ,password23:47
← duckx left23:48
→ duckx joined23:48
iheffner nOStahl: looks like you have a passphrase on the key but you don't have an SSH_ASK_PASS(?) set23:49
cmn shouldn't ssh fall back to asking on the console?23:49
→ cvprog joined23:50
← brian_g left23:50
← sec^nd left23:50
→ brian_g joined23:51
RandalSchwartz walks away23:51
nOStahl hrmm23:51
← bartek left23:51
cmn does it work if you ssh-add your key beforehand?23:52
← EspadaV8 left23:52
→ mtytel joined23:53
← duckx left23:54
nOStahl what do you mean23:54
IM not familiar23:54
bob2 wish os x had ssh-copy-id23:54
cmn man ssh-add23:55
iheffner nOStahl: in Terminal: eval $(ssh-agent) ; ssh-add23:55
nOStahl did the trick!23:56
ty guys23:56
iheffner nOStahl: do you understand what was not working and why it is now?23:56
← BSaboia left23:57
kevlarman osx has an ssh agent running by default23:57
iheffner nOStahl: and if you do not take steps to make that a "permanent" setup on login, you will run into this not working again in the future.23:57
nOStahl not exactly, intuition tells me even though I had a passphrase set it was not registered23:57
cmn the gitolite docs probably have an explanation23:57
under "not a gitolite problem"23:58
kevlarman and it should pop up a window by default23:58
← pekunz left23:58
← k_sze left23:58
sitaram cmn: yes but that doc is too broad concerning ssh; it leads to http://sitaramc.github.com/gitolite/sts.html (ssh troubleshooting)23:58
iheffner I think what was happening is that even if OS X has an agent with the correct key added to it, if your shell in Terminal does not have the right ENV settings, ssh won't/can't use the key23:59

Logs Search ←Prev date Next date→ Channels Documentation