IRCloggy #git 2011-12-10

Logs Search ←Prev date Next date→ Channels Documentation

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

2011-12-10

← diogogmt left00:00
← yshh left00:00
→ diogogmt joined00:00
ChanServ set mode: +v00:01
diogogmt hey guys00:01
I'm having a problem00:01
SDr diogogmt, welcome to the club00:01
diogogmt lol00:01
:)00:01
SDr and you're ircing: now you have 2 problems00:01
let's solve one with the other!00:02
;)00:02
← ph^ left00:02
← petteyg left00:02
diogogmt this is what happened : I have a github repo, I commited a lot of things there, but now I want them removed, I've tried using git rm but it ketp track of the logs, I want to removed everything and have a fresh repo00:03
maybe reset or rebase?00:03
jlk what the hell? could there be a bug in ls-files? git ls-files -d HEAD shows nothing, but the commit was removal of files...00:03
iheffner diogogmt: any forks or other clones of that repo?00:04
diogogmt yes00:04
→ gusnan joined00:04
← kermit left00:04
← heppy left00:04
← johnanderson left00:04
→ heppy joined00:05
→ kermit joined00:05
← glcrazy left00:05
diogogmt iheffner: if the repo has forks does it matters when reseting it?00:05
← sirclicksalot left00:05
jlk oh n/m, I'm a dork00:05
iheffner diogogmt: then consider those commits permanent. You can perhaps delete the repo and start a new one, but you have to scrub your history first.00:05
diogogmt: I'm not sure how github handles forks when the parent repo gets deleted.00:06
← TeckniX left00:06
cmn I'm fairly sure the forks won't get deleted00:07
← mxweas_ left00:07
iheffner diogogmt: but if you look at the world outside of github, if you put another repo at the same url, git isn't going to know that it isn't the same. It may conflict all over itself. or if one is a modified copy of the older one....00:07
diogogmt: ... it may just merge the histories together.00:07
← Vile left00:07
← cschneid left00:07
diogogmt but what if I just want to delete the history for the repo, how would I be able to do that?00:08
the problem is that I don't have Admin privileges on github, so I can't delete from their UI00:08
cmn if you don't have admin privs on the repo, are you sure deleting history is the right move?"00:09
→ heppy_ joined00:09
← heppy_ left00:09
nevyn you're likely to just piss of whoever is the admin00:09
cmn to change the history, you'll have to either delete branch or force-push a different version of the branch00:10
nevyn who'll revert whatever you've done by pushing their copy back up to github00:10
reset and force push00:10
cmn but doing that on a repo with forks is probably exactly the wrong move to take00:10
iheffner does github even allow forced pushes?00:10
cmn of course00:11
→ unnamedwill joined00:11
EugeneKay Yes, of course.00:11
iheffner well, there's the "delete / push" path, but I'd expect it balk at attempts to rewrite history.00:12
cmn there's a config option to stop you from deleting braches, but I'm not sure if the github interface lets you control that00:12
← heppy left00:13
→ ehsan joined00:13
← hobodave left00:13
SDr iheffner, okay, here's a good one: .git/info/exclude00:14
create an initial checkout script, whereby I do a full pull, then selectively add the files to the exclude list00:14
iheffner, what fail cases can you spot with this method?00:15
← ldlework left00:15
kevlarman SDr: you can't ignore files under version control00:15
SDr (.git/info/exclude is local to each repo)00:15
→ nvn joined00:16
cmn SDr: exclude and gitignore is to ignore untracked files00:16
it doesn't delete existing tracked files00:16
that would in fact break git00:16
kevlarman cmn: maybe something with git's sparse checkouts?00:18
cmn maybe... can you push with those?00:19
SDr cmn, each module is a single file00:19
kevlarman cmn: i would assume so, it's not shallow, all the objects are available locally00:19
SDr if I understand correctly, sparse checkout does directories00:20
which would mean 1 directory for each file00:20
kevlarman SDr: looks like it uses the same rules as .gitignore00:21
SDr: see man git-read-tree00:21
gitinfo the git-read-tree manpage is available at http://jk.gs/git-read-tree.html00:21
→ astrostl joined00:22
Caelum I ran a git rebase --skip during a conflict and it seems to have eaten all my code, the whole commit is gone, any way to get it back?00:22
kevlarman Caelum: see man git-reflog00:22
gitinfo the git-reflog manpage is available at http://jk.gs/git-reflog.html00:22
kevlarman you can find what your branch pointed to before the rebase and check that out or reset to it00:23
→ sirclicksalot joined00:24
cmn SDr: I might have missed this info, but how much of the development do you have to do at the client?00:24
Caelum kevlarman: whew, got the code back, thank you very much!00:24
← neurodrone left00:25
Caelum kevlarman: so git rebase --skip skips the whole commit?00:25
← Know1edge left00:25
kevlarman Caelum: yup00:25
Caelum kevlarman: I see, thanks00:26
← kennethreitz left00:26
← baburdick left00:27
→ dv310p3r joined00:27
→ baburdick joined00:28
→ sorin joined00:28
← cccaldas left00:29
→ subbyyy joined00:30
→ cccaldas joined00:30
← jlk left00:30
← cmyk_ left00:30
← g0bl1n left00:31
← giallu left00:32
→ lightcap joined00:32
← Chillance left00:32
← dv310p3r left00:33
← iocor left00:34
← sirclicksalot left00:34
← yeryry left00:34
→ yeryry4 joined00:34
→ traviscj joined00:34
→ nrichards joined00:35
→ oriba_ joined00:35
→ dv310p3r joined00:35
→ sirclicksalot joined00:36
← tvw left00:37
← oriba left00:38
← nrichards left00:38
→ kennethreitz joined00:39
→ jasonbray joined00:39
← traviscj left00:41
← ehsan left00:43
← j416 left00:44
← Destos left00:45
← unnamedwill left00:46
← ddkilzer left00:46
→ unnamedwill joined00:46
← SamB left00:46
→ Destos joined00:47
← jimmy1980 left00:47
← fridim_ left00:48
→ jimmy1980 joined00:48
← cakehero left00:49
← brian_g left00:50
→ rubyruy joined00:50
← mandric left00:51
← lightcap left00:52
← cjuner left00:52
← barcoder left00:53
→ __sorin__ joined00:53
→ jlebar joined00:56
← __sorin__ left00:57
← baburdick left00:57
→ __sorin__ joined00:57
→ SamB joined00:58
← TStubbs left00:58
→ baburdick joined00:59
mattalexx It's policy at my new job to squash my recent commits into a single commit with a detailed explanation and bug ids, then do a pull request. How can I do that but keep my smaller commits?01:01
frogonwheels mattalexx: separate branch01:01
← sirclicksalot left01:01
← cannonball left01:02
← dv310p3r left01:02
← SamB left01:02
← kvanderw1 left01:02
mattalexx frogonwheels, So create a new branch for each pull request?01:02
→ sirclicksalot joined01:03
← w09x left01:03
← tjholowaychuk left01:04
frogonwheels mattalexx: or have a public version of your working branch with the squished versions and lots of merges from the main one.01:04
← sorin left01:04
mattalexx frogonwheels, ... or develop in my own branch, then use another branch as a buffer between my branch and the mainline?01:04
frogonwheels mattalexx: that depends on how they want it.01:04
← nadavoid left01:04
→ sorin joined01:04
frogonwheels mattalexx: yep. you can always delete branches that have been accepted.01:05
← sorin left01:05
frogonwheels mattalexx: so many choices :)01:05
← __sorin__ left01:05
→ sorin joined01:05
← jaequery left01:07
← dankest left01:09
→ mxweas_ joined01:11
← tgunr left01:12
→ yairgo joined01:12
← subbyyy left01:13
← ProperNoun left01:14
yairgo is there an environment variable that I can set to make git deamon verbose when running over apache and http01:14
→ WhiskerBiscuit joined01:15
→ ProperNoun joined01:15
bremner yairgo: git daemon does not run over apache01:16
cmn git daemon doesn't run over apache and http, you're thinking of git-http-backend and from the mapage it does't look like there are specific flags for it01:16
bremner yairgo: man git-daemon01:16
gitinfo yairgo: the git-daemon manpage is available at http://jk.gs/git-daemon.html01:16
cmn but GIT_TRACE may work01:16
→ d2dchat joined01:16
← jkyle left01:17
← WhiskerBiscuit left01:17
yairgo bremner, I'm getting a "Service not enabled: 'receive-pack'" in my apache log and trying to diagnose how to get more information on the error. the only place that error exists in the git codebase is in the git-deamon area01:18
→ WhiskerBiscuit joined01:18
→ subbyyy joined01:18
← rubyruy left01:18
bremner yairgo: you should believe us, or find some other help ;)01:18
ChanServ set mode: +v01:18
WhiskerBiscuit Hello... I've been using mercurial for the past 8 motnhs and really love it (well DVCS at least). I keep hearing about git as well, and was wondering if there is a primer that would be good for someone famailar with Hg01:19
→ kenichi joined01:20
yairgo bremner, I didn't say I don't believe you, i was fully under the impression that the git deamon doesn't run over apache just can't find where else this error is coming from :(01:20
bremner everything I can find online relating to that error says that you need to run git daemon with --enable=receive-pack, I can't find how to set that for apache, any ideas?01:22
Mikachu funny, if you google for switching from mercurial to git, all the results are about switching from git to mercurial ;)01:22
bremner yairgo: does man git-http-backend help01:23
gitinfo yairgo: the git-http-backend manpage is available at http://jk.gs/git-http-backend.html01:23
← maletor left01:24
WhiskerBiscuit Mikachu: Not switching, just looking at it cos it seems git is used more01:25
← sirclicksalot left01:25
Mikachu well, the information you would need is the same01:25
WhiskerBiscuit does git use the changeset concept?01:25
yairgo bremner thank you01:25
WhiskerBiscuit hash codes for each commit?01:25
jast yes, though they're not stored internally as changesets. they're more like snapshots.01:26
Mikachu there's http://mercurial.selenic.com/wiki/GitConcepts01:26
it's the opposite, but i guess it's helpful01:26
WhiskerBiscuit do the installation packages for git include visual merge uilities like Kdiff?01:27
Mikachu that depends on who made the package01:27
git mergetool supports a bunch of different programs, and a bunch of different merge programs support git too01:27
← baburdick left01:27
bremner if you count linux distros as "installation packages for git", then yes01:27
← Behi left01:28
jast normally git mergetool uses external programs for that. git has built-in knowledge of how a whole bunch of them should be called.01:28
→ sirclicksalot joined01:28
Mikachu i always get confused when looking at anything more complicated than regular conflict markers01:28
WhiskerBiscuit i live in a windows world unfortunately01:28
← btanaka left01:28
jast yeah, personally I always just edit the files with conflict markers01:28
WhiskerBiscuit its definitely not worth it for me to switch my team to git01:28
→ Behi joined01:28
bremner how can you possibly know that?01:29
← dreiss left01:29
WhiskerBiscuit Thye had a shitfit going from svn to Hg01:29
← warthog9 left01:29
WhiskerBiscuit thought it would ruin their world01:29
bremner oic, fair enough.01:29
jast setting merge.conflictstyle to diff3 was kind of nice though01:29
Mikachu there's tools to use git with a hg repo and vice versa01:29
and with an svn repo for that matter01:29
i usually git co --conflict=diff3 when i need it01:30
→ baburdick joined01:30
Mikachu in the common case having the base code in the middle is just in the way01:30
WhiskerBiscuit ever since going to DCVS, it's changed the way we build stuff01:30
Mikachu (for me)01:30
← Behi left01:30
WhiskerBiscuit I keep telling them, "when in doubt, put it in a seperate branch"01:31
go design whatever crazy shit you want to, but dont mess with a production or beta branch01:32
jast DVCS tend to change things, yeah... :)01:33
→ garrettrayj joined01:34
WhiskerBiscuit only wish we used it a few years ago01:34
→ w09x joined01:34
← ProperNoun left01:34
→ ProperNoun joined01:34
→ jpieper joined01:34
← jasonbray left01:34
WhiskerBiscuit some dumbfuck of a project manager let the developers take a database out of 3NF because they were having performance issues and started storing everything in blobs01:35
Now we suffer for it until we can convert back...... There's a branch for that01:36
← mxweas_ left01:37
→ buntstift joined01:37
→ mdel joined01:37
← pantsman left01:37
→ gavin_huang joined01:38
→ Element9 joined01:38
bremner anyone know how git-update-index copes with concurrent writes? Is there locking?01:40
jast that's what index.lock is for, right?01:41
bremner sounds plausible ;)01:41
cmn yeah, it deals with them by not doing them01:42
jast we should totally change that01:43
because this way writes can starve. suppose you have a billion clients all trying to update the index at the same time...01:43
cmn isn't the problem somwhere else in that case?01:44
jast we must do our best to make this use case work!01:44
→ muddo joined01:44
jast doubtlessly we need to replace the index with an oracle database01:44
cmn repeat after me: git is not mysql01:44
← sorin left01:44
cmn oh, oracle would be eve better, yeah01:45
jast well yeah. git keeps data.01:45
or to quote someone from twitter: it loses lots of data but damn, is it fast01:45
mattalexx I just reset two (already pushed to remote) commits, commited them together as one commit, did a `git push origin HEAD --force` and now I see that commit in my commit history screen on github but not my source view. Has anyone seen this before?01:45
jast 'source view'? is that some github thing?01:45
cmn do one thing and do it well :)01:46
← berserkr left01:46
mattalexx jast, I don't know, it's the view where you can read the code.01:46
jast right01:46
cmn on github or where?01:46
mattalexx github01:46
jast I suspect it's a caching thing01:46
→ valaris joined01:46
mattalexx Darn I have to do a pull request before bed. How long does it usually take in your experience? jason23701:47
^ jast01:47
jast no idea. I don't use github a lot.01:47
cmn creating a pull request takes a few seconds01:47
→ berserkr joined01:47
bremner mattalexx: you could try #github01:47
jast iirc you can view files per revision anyway01:47
cmn how fast it gets merged depends on the maintainer01:47
mattalexx bremner, Hm, good idea.01:47
← infogulch left01:48
→ warthog9 joined01:48
→ warthog19 joined01:48
← warthog19 left01:48
← sirclicksalot left01:52
→ jaequery joined01:52
→ mirage335 joined01:53
→ sirclicksalot joined01:54
← jimmy1980 left01:55
→ jimmy1980 joined01:56
← mattalexx left01:57
← verto left01:57
← baburdick left01:58
← cccaldas left01:59
→ baburdick joined02:00
← mdel left02:00
→ jasonbray joined02:07
← w09x left02:07
→ w09x joined02:08
← dnjaramba left02:08
→ dnjaramba joined02:08
← jaequery left02:10
← valaris left02:12
→ mandric joined02:12
→ w09x_ joined02:12
← w09x left02:12
→ Chib joined02:13
→ Chillance joined02:14
← sirclicksalot left02:15
← kenperkins left02:16
→ kenperkins joined02:17
← kenperkins left02:17
→ sirclicksalot joined02:17
← msponge left02:17
→ msponge joined02:18
→ ognevsky joined02:18
← buntstift left02:19
rns If there is a subversion package I want and I have to get it using svn, can I turn it into a git repo?02:19
cmn that's what git-svn does02:20
→ Martin_ joined02:21
← Araxia_ left02:21
→ EricInBNE joined02:22
← DarkAR left02:22
← w09x_ left02:23
← towski left02:23
→ mxweas_ joined02:23
→ w09x joined02:23
→ heppy joined02:24
← orafu left02:25
← w09x left02:28
← baburdick left02:28
← warthog9 left02:29
→ orafu joined02:29
→ baburdick joined02:31
← ognevsky left02:31
← apok left02:32
← diogogmt left02:33
→ w09x joined02:33
→ jaequery joined02:34
← jasonbray left02:34
→ Yuuhi`` joined02:35
← mxweas_ left02:36
← Yuuhi` left02:37
← MANCHUCK left02:37
← ssaasen left02:38
← Yuuhi`` left02:38
→ Yuuhi`` joined02:38
← jimmy1980 left02:38
← Chillance left02:39
← msponge left02:39
→ jimmy1980 joined02:40
→ VladGh joined02:40
→ ponto joined02:40
← _ponto_ left02:41
→ dagda joined02:42
← gumnos left02:43
← baburdick left02:43
← sirclicksalot left02:43
← ColKurtz left02:43
→ SRabbelier|Lappy joined02:45
→ sirclicksalot joined02:46
← tystr left02:47
→ mxweas_ joined02:48
→ towski joined02:48
← towski left02:50
→ caseymcg joined02:52
← w09x left02:52
yairgo Does anyone know what the difference is between setting GIT_PROJECT_ROOT inside of a Location vs a LocationMatch directive on git with apache02:52
→ w09x joined02:52
← d2dchat left02:53
→ towski joined02:53
→ valaris joined02:53
← towski left02:54
← Vampire0 left02:54
← jimmy1980 left02:56
→ sirclicksalot1 joined02:56
← sirclicksalot left02:56
← mxweas_ left02:57
→ mgpcoe joined02:59
→ kenperkins joined02:59
→ towski joined02:59
→ mxweas_ joined03:01
→ jimmy1980 joined03:02
← dvide left03:03
← rosco_y left03:03
← kenperkins left03:03
→ shiba_yu36 joined03:05
→ zivester joined03:06
→ ericbarnes joined03:08
→ Vampire0 joined03:08
→ kvanderw joined03:09
→ vptr joined03:09
→ Vash joined03:10
← oriba_ left03:12
← sirclicksalot1 left03:12
→ jpalmer joined03:12
← jimmy1980 left03:12
→ igotnolegs joined03:13
→ SamB joined03:13
← jpalmer left03:13
→ jpalmer joined03:14
← ilteris_ left03:17
→ jimmy1980 joined03:18
← gusnan left03:20
← shiba_yu36 left03:21
→ eightyei1ht joined03:24
→ diogogmt joined03:24
← kvanderw left03:28
← diogogmt left03:29
→ sirclicksalot joined03:30
→ glennpratt joined03:30
← heppy left03:34
→ dotnull joined03:35
← dotnull left03:35
← towski left03:35
← sirclicksalot left03:35
→ dotnull joined03:36
→ sirclicksalot joined03:37
← JackWinter left03:39
← hyperair left03:41
→ CRCinAU joined03:43
→ JackWinter joined03:43
CRCinAU i have a stupid question.03:43
how do I check out this? http://git.kernel.org/?p=linux/kernel/git/jeremy/xen.git;a=shortlog;h=refs/heads/xen/next-2.6.3203:44
→ dr0id joined03:44
CRCinAU its the next-2.6.32 branch that I've had a mental snafu and can't quite grok it.03:45
← rockfruit left03:45
CRCinAU is this correct? git clone --depth 1 git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen.git next-2.6.3203:46
→ rockfruit joined03:46
→ herdingcat joined03:47
→ accel joined03:47
accel is there something like git; but for snapshots of datastructures in memory; rather than for files on disk?03:47
i.e. I have a set of data structures in the memory of my program; I want to be able to take snapshots of them, commits of them, then see it as a tree, and go back / forht along the logs03:48
it soudns like I'm starting ot describe a database03:48
→ dys` joined03:49
→ LongBeach joined03:50
← LongBeach left03:50
→ fisted_ joined03:50
→ munro joined03:50
← fisted left03:51
← nevyn left03:51
← WhiskerBiscuit left03:52
← z0ned left03:52
→ z0ned joined03:52
← dys left03:52
→ WhiskerBiscuit joined03:53
→ ijuma_ joined03:53
→ nevyn joined03:53
→ tchan joined03:53
← CRCinAU left03:54
→ heppy joined03:55
→ apok joined03:56
← fukenA left03:56
→ Nibato joined04:02
→ kenperkins joined04:02
← mxweas_ left04:03
← wilmoore left04:04
→ daveb_ joined04:04
← croepha left04:05
← sirclicksalot left04:05
→ johnanderson joined04:06
→ adrien__ joined04:06
→ sirclicksalot joined04:07
← valaris left04:08
← WhiskerBiscuit left04:08
← SDr left04:09
← jimmy1980 left04:13
→ mxweas_ joined04:14
← berserkr left04:16
→ Targen joined04:16
→ jlebar_ joined04:19
← jlebar left04:19
jlebar_ → jlebar04:19
→ jimmy1980 joined04:20
← caseymcg left04:20
→ otakutomo joined04:24
← jimmy1980 left04:24
spyro → napdragon04:26
← accel left04:27
← mxweas_ left04:29
→ jimmy1980 joined04:31
→ identify joined04:32
← sirclicksalot left04:32
← arvind_khadri left04:33
→ sirclicksalot joined04:34
← yairgo left04:38
→ namidark_ joined04:41
→ yshh joined04:41
→ mxweas_ joined04:45
← kenichi left04:51
→ wbednarski joined04:51
← wbednarski left04:52
→ disappea_ joined04:54
← raythecat left04:55
→ wbednarski joined04:57
← identify left05:00
→ Niedar joined05:00
← sirclicksalot left05:01
→ sirclicksalot joined05:03
→ flyz joined05:05
← mikepack left05:06
← Fretta left05:07
← flyz_ left05:08
→ errr joined05:09
→ jjore joined05:09
→ felher joined05:09
→ fr joined05:09
← yshh left05:10
← wbednarski left05:11
→ bindaasomatic joined05:12
← codeshah left05:14
→ frogonwheels joined05:14
← mxweas_ left05:16
→ mikepack joined05:17
→ OOPMan joined05:20
→ mxweas_ joined05:21
→ media-948 joined05:23
← media-948 left05:25
← media-488 left05:26
→ media-584 joined05:26
← ISF left05:30
→ msponge joined05:32
→ nrichards joined05:32
← mgpcoe left05:33
← Niedar left05:33
← jimmy1980 left05:35
← herdingcat left05:35
→ jimmy1980 joined05:36
← media-584 left05:36
→ herdingcat joined05:41
→ SegFaultAX joined05:42
← Xeross left05:42
← mandric left05:44
← muddo left05:44
→ muddo joined05:45
→ mandric joined05:47
→ raythecat joined05:48
→ Xeross joined05:48
→ jlebar_ joined05:51
← jlebar left05:51
jlebar_ → jlebar05:51
← ericbarnes left05:52
← Xeross left05:53
← herdingcat left05:55
→ CpuID2 joined05:55
→ media-886 joined05:57
← sirclicksalot left05:57
← gerard0 left05:58
→ gerard0 joined05:58
→ sirclicksalot joined06:00
→ Xeross joined06:01
← mxweas_ left06:02
← mandric left06:04
→ aroop joined06:07
→ the-fonz joined06:08
← marcuy left06:08
← heppy left06:10
← garrettrayj left06:10
← MacGyverNL left06:10
→ offbyone joined06:12
→ baburdick joined06:14
→ kadoban_ joined06:18
→ mxweas_ joined06:18
← glennpratt left06:19
→ Dave^| joined06:19
← subbyyy left06:20
← kadoban left06:21
← media-886 left06:22
← kukks left06:22
← astrostl left06:24
← mxweas_ left06:26
→ Tankado joined06:31
→ thingee joined06:32
← thingee left06:36
← vptr left06:40
→ rendar joined06:43
→ Josh_ joined06:46
← dirkle left06:47
← Josh_ left06:47
← metcalfc left06:49
← niloc132 left06:50
← [M]ax left06:51
← sirclicksalot left06:51
→ aspotashev joined06:51
→ ramusara joined06:52
→ heppy joined06:52
→ [M]ax joined06:53
→ PovAddict joined06:54
PovAddict what git version added .gitattribtes, and textconv diff stuff?06:55
aroop → aroop|away06:57
← Vash left07:03
→ yshh joined07:03
→ TrineLocke joined07:04
← nevyn left07:06
→ mxweas_ joined07:07
← mxweas_ left07:07
→ sirclicksalot joined07:08
← dotnull left07:11
→ nevyn joined07:12
← sirclicksalot left07:15
→ Vash joined07:16
→ sirclicksalot joined07:17
→ topi joined07:17
→ sattu94 joined07:18
← nevyn left07:22
→ melvinram joined07:22
melvinram Is there a way to push all branches to a specified remote repo? I've googled but haven't come up with a good solution07:23
→ nevyn joined07:24
tango_ --all or --mirror, depending on the situation07:24
if you want to make one remote the exact mirror of a repo, look into --mirror07:25
→ mxweas_ joined07:25
← philip left07:25
melvinram tango_: thanks. i'll look into mirror07:26
← madewokherd left07:27
→ ilteris_ joined07:32
→ marthinal joined07:33
→ toabctl joined07:35
← mxweas_ left07:35
rendar what is the flag for git commit to make a commit from a .txt file?07:38
ok, -F07:39
← mikepack left07:39
→ mikepack joined07:40
→ akosikeno joined07:40
← mikepack left07:40
← Destos left07:40
→ xiangfu joined07:41
← toabctl left07:44
→ tewecske joined07:44
← sirclicksalot left07:45
→ MacGyverNL joined07:45
← MacGyverNL left07:45
→ MacGyverNL joined07:45
→ sirclicksalot joined07:47
→ jlebar_ joined07:47
← jlebar left07:47
jlebar_ → jlebar07:47
→ cybersphinx joined07:47
← drev1 left07:51
← the_cybersphinx left07:51
→ iocor joined07:54
→ mikepack joined07:54
← xsbeats left07:57
→ h0bbit joined08:00
→ drizzd joined08:02
← aspotashev left08:04
← mikepack left08:05
→ mikepack joined08:05
← iocor left08:06
→ iocor joined08:07
← iocor left08:07
← SegFaultAX left08:09
← mikepack left08:09
→ offby1 joined08:11
→ xangelo joined08:14
→ womanizer_ joined08:17
← xangelo left08:18
← TrineLocke left08:20
→ SeySayux joined08:21
SeySayux Hey, does anybody have experience with a bug tracking tool that 1) does not require me to molest my web server; and 2) has good git support?08:22
→ Dave^|| joined08:22
← offby1 left08:23
← otakutomo left08:24
← Dave^| left08:25
← Bass10 left08:26
← dharriga1 left08:28
→ dharrigan joined08:28
← dharrigan left08:28
→ dharrigan joined08:28
→ leo2007 joined08:31
leo2007 folks, any idea whether it is possible to sync a forked repo from its origin using the github website?08:32
→ the_cybersphinx joined08:32
ChanServ set mode: +v08:32
loocorez SeySayux: FogBugz? Not hosted, though, and it's paid. At my job we use Redmine, and it's okay. Feels cluttered.08:33
tango_ SeySayux: I don't know of any, but I recently came across http://searchco.de/blog/view/batf-big-arse-text-file08:33
→ maletor joined08:34
SeySayux loocorez: I'm using Redmine at the moment, and it seems a little bit low on features.08:34
← cybersphinx left08:34
loocorez SeySayux: Yeah, I'm not a big fan either. I have to find an alternative by next fall :P08:34
Let me know if you find a super awesome one08:34
SeySayux It starts to look like s/find/write/ ...08:35
→ warthog9 joined08:35
→ ph^ joined08:35
SeySayux tango_: Unfortunately, I do need a full-featured bug tracker. Everything BATF-compatible is already stored in Evernote.08:35
← aroop|away left08:36
← ponto left08:39
selckin jira the one and only </singing voice>08:40
← Chib left08:41
← leo2007 left08:42
CareBear\ hah08:43
← Tommy[D] left08:43
SeySayux Hmm, I might try Jira... Not that an open source project has such a budget, but hey, tpb :P08:45
tango_ tpb = ?08:45
→ Tommy[D] joined08:45
← Druid_ left08:45
← stringoO left08:46
SeySayux tango_: A website that is specialised in sharing paid software for a lower (ahem) price.08:46
tango_ rotfl08:46
→ eean joined08:48
← dharrigan left08:48
→ dharrigan joined08:50
→ Druid_ joined08:51
→ drev1 joined08:56
selckin pritty sure it has free tiers for open source08:57
→ cybersphinx joined08:59
SeySayux Neat :)08:59
rue SeySayux: I'd personally stay far from JIRA. There's pivotal and lighthouse09:00
← dharrigan left09:01
SeySayux rue: I'll add them to my to-test list09:01
→ dreiss joined09:01
→ psakrii joined09:01
SeySayux rue: Any founded criticism on Jira?09:01
rue SeySayux: Personal opinion: it's slow, unwieldy and ultra-enterprisey. A fair amount of people like it, but lots really don't09:02
← nevyn left09:03
← VladGh left09:03
SeySayux I'm kind of going for somewhat enterprisey here.09:03
← the_cybersphinx left09:03
← sirclicksalot left09:04
→ nevyn joined09:05
→ sirclicksalot joined09:05
→ bauruine joined09:07
← Tankado left09:08
→ dharrigan joined09:09
← dharrigan left09:09
→ dharrigan joined09:09
→ caseymcg joined09:12
→ ianmcorvidae joined09:13
← ianmcorvidae left09:13
→ ianmcorvidae joined09:13
← heppy left09:13
→ dharriga1 joined09:14
← dharrigan left09:15
→ airborn joined09:15
← disappea_ left09:16
→ Shusshu joined09:16
→ otakutomo joined09:17
→ disappe__ joined09:18
← womanizer_ left09:18
← akosikeno left09:19
← igotnolegs left09:20
→ ognevsky joined09:21
← raythecat left09:21
nevyn yawns09:22
← dharriga1 left09:26
→ ReekenX joined09:26
→ dharrigan joined09:27
→ kpreid joined09:30
→ d0k joined09:31
← sirclicksalot left09:31
→ sirclicksalot joined09:33
→ MissionCritical joined09:33
→ arvind_khadri joined09:34
← dharrigan left09:35
→ dharrigan joined09:36
← dharrigan left09:36
→ dharrigan joined09:36
→ glcrazy joined09:39
→ dzonder joined09:40
← joshsmith left09:42
→ fermo joined09:42
← drev1 left09:42
→ identify joined09:43
→ node86__ joined09:45
← dzonder left09:48
← Vash left09:49
→ the_cybersphinx joined09:50
← cybersphinx left09:52
→ honchu joined09:53
← sirclicksalot left09:55
→ proqesi` joined09:56
← proqesi` left09:56
→ proqesi` joined09:56
← honchu_ left09:56
→ MarcoSousa joined10:00
→ amirwebdev2 joined10:03
amirwebdev2 what is git ref?10:03
→ bartek joined10:03
← ognevsky left10:07
← nevyn left10:07
CareBear\ amirwebdev2 : I guess it refers to a commit reference, which can be expressed in lots of ways, see man git rev-parse10:08
gitinfo the git-rev-parse manpage is available at http://jk.gs/git-rev-parse.html10:08
→ ognevsky joined10:08
← lapistano left10:08
→ lapistano joined10:08
→ nevyn joined10:09
← node86__ left10:10
→ Suhana joined10:10
→ sirclicksalot joined10:12
← dharrigan left10:19
→ flavius joined10:20
→ dharrigan joined10:20
→ canton7 joined10:20
← sirclicksalot left10:23
← jimmy1980 left10:23
→ thiago joined10:23
→ cmyk_ joined10:24
← caseymcg left10:24
→ sirclicksalot joined10:25
→ fridim_ joined10:28
→ jimmy1980 joined10:30
→ kriegerod joined10:33
← yshh left10:33
← setmeaway2 left10:35
→ setmeaway joined10:36
← w09x left10:40
→ ipalaus_ joined10:40
→ yshh joined10:41
→ jasonbray joined10:41
← tchan left10:41
→ j416 joined10:42
→ aspotashev joined10:46
→ Guest29733 joined10:46
→ tchan joined10:47
← sirclicksalot left10:49
→ jrydberg_ joined10:49
→ khmarbaise joined10:50
→ pantsman joined10:50
← pantsman left10:50
→ pantsman joined10:50
→ sirclicksalot joined10:51
→ harshpb joined10:53
← dnjaramba left10:54
← jasonbray left10:54
→ cybersphinx joined10:55
→ Cromulent joined10:55
← kennethreitz left10:56
← the_cybersphinx left10:56
← Suhana left10:57
→ jasonbray joined11:01
→ ManDay joined11:04
← maletor left11:06
→ Chillance joined11:08
→ brian_g joined11:11
→ Suhana joined11:11
← sirclicksalot left11:13
← flavius left11:14
→ sirclicksalot joined11:14
← OOPMan left11:15
← dagda left11:17
→ barcoder joined11:17
← nevyn left11:17
→ nevyn joined11:19
← jimmy1980 left11:19
→ MarcoSousa1 joined11:22
← MarcoSousa left11:23
← Cromulent left11:25
→ jimmy1980 joined11:26
→ knifed joined11:29
→ TheDark12 joined11:29
→ raythecat joined11:31
← ognevsky left11:31
→ Nebukadneza joined11:32
Nebukadneza heho11:32
← j416 left11:32
← TheDark12 left11:32
→ ognevsky joined11:33
Nebukadneza i don't know if this is the right channel, but i'm trying to set up girocco and as the documentation is very sparse and there seem no dedicated girocco channel i thought i'd ask here11:34
so well, i think i've got most stuff running, made install, let it build its chroot, added the cronjobs11:34
i can create repos and users11:34
just the strange ssh setup it wants - how exactly does girocco want its own sshd started Oo?11:35
← jimmy1980 left11:35
→ jimmy1980 joined11:36
← Suhana left11:40
→ na3r joined11:40
→ _iron joined11:41
→ albel727 joined11:41
← dharrigan left11:43
→ psoo joined11:44
canton7 looks like jast is one of the main committers11:48
→ dharrigan joined11:48
← dharrigan left11:48
→ dharrigan joined11:48
→ dpino joined11:50
← jasonbray left11:50
← sattu94 left11:51
→ Suhana joined11:52
Nebukadneza mhhh, kay canton711:53
so then ... jast: ping! :)11:53
→ jasonbray joined11:55
← psoo left11:56
← ManDay left11:57
← identify left11:59
← raythecat left12:01
← psakrii left12:01
← sirclicksalot left12:01
→ flavius joined12:02
← Suhana left12:03
_ikke_ Jan Krüger is jast?12:08
→ Vile joined12:08
← flavius left12:09
→ ManDay joined12:09
cbreak _ikke_: jast is half man, half machine, answering man git requests in his spare time in between saving the world12:10
gitinfo _ikke_: the git manpage is available at http://jk.gs/git.html12:10
cbreak seems he's in-cognito at the moment :)12:10
← marthinal left12:11
frogonwheels man git12:13
gitinfo the git manpage is available at http://jk.gs/git.html12:13
← Element9 left12:13
frogonwheels cbreak: jast has created gitinfo now to do that funciton12:13
→ hc96 joined12:13
hc96 Hi! How can I find out, wether a file is under version control or ignored, without looking into the .gitignore in the whole tree?12:14
drizzd hc96: git ls-files -v -o -c <file>12:16
← kriegerod left12:16
→ Suhana joined12:16
drizzd (and combine with -i --exclude-standard)12:17
→ dharriga1 joined12:17
← dharrigan left12:17
frogonwheels hc96: of course a file can be both under version control and match the .gitignore filter12:18
← albel727 left12:18
→ albel727 joined12:18
hc96 drizzd: thank you!12:19
→ andrelopes joined12:19
iheffner hc96: you could git log the file12:19
→ LongBeach joined12:19
hc96 frogonwheels: oh yes, you're right. So just looking for the .gitignore wouldnt even be enough12:19
iheffner: nice :-)12:20
drizzd hc96: although git status --porcelain --ignored would be a more modern way to do it12:20
→ sirclicksalot joined12:20
EugeneKay _ikke_ - /whois jast ;-)12:21
← jlebar left12:23
→ dharrigan joined12:24
← dharriga1 left12:24
← andrelopes left12:25
→ alnewkirk joined12:25
← sirclicksalot left12:26
→ Vash joined12:31
← f0i left12:33
← cmyk_ left12:34
→ f0i joined12:34
← Suhana left12:35
← Guest29733 left12:36
→ j416 joined12:36
← Vash left12:37
→ codeshah joined12:38
→ sirclicksalot joined12:44
← rockfruit left12:48
← amirwebdev2 left12:50
→ Vash joined12:50
→ riper joined12:53
← sirclicksalot left12:54
→ sirclicksalot joined12:56
← yshh left12:58
← ManDay left12:59
← Vash left12:59
→ yshh joined13:00
→ replore joined13:05
→ dzonder joined13:08
→ psoo joined13:11
← harshpb left13:11
→ hc96` joined13:14
← hc96 left13:16
← babilen left13:16
← psoo left13:19
→ babilen joined13:20
← msponge left13:25
→ marcuy joined13:26
→ ZyX-I joined13:28
→ w0bni joined13:28
ZyX-I Hello. Where is the format of the last lines of “git log --raw” output described?13:31
← jimmy1980 left13:33
cbreak ZyX-I: it's from man git-diff if that helps13:34
gitinfo the git-diff manpage is available at http://jk.gs/git-diff.html13:34
← otakutomo left13:35
ZyX-I cbreak: Thanks.13:35
→ glennpratt joined13:36
← knifed left13:36
→ harshpb joined13:37
→ Correcto joined13:38
→ qian joined13:39
canton7 whoa! when did gitinfo appear?13:39
!gitinfo13:40
!jast13:40
_ikke_ 13:15:40* frogonwheels | cbreak: jast has created gitinfo now to do that funciton13:40
→ jimmy1980 joined13:40
canton7 ye, just saw that13:41
And the old gitbot-standin.pl source is no longer available13:41
→ joshsmith joined13:42
← j416 left13:43
→ coredumb joined13:44
→ bahamas joined13:44
bahamas hello. i've created a new branch, but haven13:44
bassliner 't accidently hit the enter key while trying to type "'"?13:44
bahamas yes :)13:45
canton7 I have that trouble on my 1st gen eee -- the ' key is absolutely minute13:45
bahamas haven't committed anything in it. now i want to rename it. should i delete it and create a new one, or there is a different way?13:46
→ psakrii joined13:46
canton7 bahamas, git branch -m13:46
← ReekenX left13:47
bahamas canton7: great. thanks13:47
← coredumb left13:47
← khmarbaise left13:48
→ coredumb joined13:49
→ knifed joined13:49
← sirclicksalot left13:49
fisted_ → fisted13:51
→ sirclicksalot joined13:51
← coredumb left13:54
→ koo5 joined13:54
→ msponge joined13:55
→ coredumb joined13:56
→ jrm2k6 joined13:58
← gavin_huang left14:00
→ urbanmonk joined14:00
→ otakutomo joined14:01
← urbanmonk left14:02
← xiangfu left14:02
→ herdingcat joined14:04
← herdingcat left14:05
→ urbanmonk joined14:05
← Correcto left14:06
→ herdingcat joined14:06
← herdingcat left14:08
← knifed left14:08
→ herdingcat joined14:10
← jasonbray left14:11
← harshpb left14:12
→ kvanderw joined14:12
← qian left14:13
← urbanmonk left14:13
← sirclicksalot left14:15
→ sirclicksalot joined14:16
← aspotashev left14:23
← hc96` left14:23
← Vile left14:26
→ cmihai joined14:28
→ jasonbray joined14:30
→ cmyk_ joined14:30
→ Bass10 joined14:30
→ dvide joined14:31
← fisted left14:31
→ fisted joined14:33
← jimmy1980 left14:35
← yeryry4 left14:36
← fisted left14:36
→ yeryry joined14:36
→ jimmy1980 joined14:40
← replore left14:40
← fridim_ left14:42
← sirclicksalot left14:42
← glennpratt left14:43
→ replore joined14:43
→ sirclicksalot joined14:43
→ Vile joined14:46
→ ISF joined14:47
← alnewkirk left14:49
← kvanderw left14:50
← bindaasomatic left14:51
→ hyperair joined14:51
→ Cromulent joined14:52
→ skim1776 joined14:52
→ glennpratt joined14:52
skim1776 what if I commited without adding new files(forgot about add ., just made commit -a "some message") and switched to another branch - I've lost all new files, can I recover them?14:53
canton7 You won't have lost the new files -- they'll have "floated" across to the new branch14:53
just switch back to the old branch, add then, commit --amend14:54
(assuming you haven't pushed)14:54
→ xsbeats joined14:55
← jrm2k6 left14:55
skim1776 thank you! canton7, you're my hero!14:56
canton7 git will try very hard not to delete untracked content. reset --hard and checkout -f can cause accidental loss, but not much else14:56
skim1776 actually, when using IDE - does it worth to use git bash?14:56
canton7 untracked *files*, as opposed to just untracked content in existing files, are very hard to lose acceidentally14:57
I've yet to come across a git gui I like, although some of the mac ones look nice :P14:57
← dr0id left14:57
canton7 You've definitely (imo) got a lot more power on the command-line14:57
→ iocor joined14:58
← glennpratt left14:59
skim1776 yeah, command line is powerful if you know what to do15:01
and if you are learning it's a bit scary15:01
canton7 depends what you're learning from -- all the good resources are writtenf or the command-line15:02
*written for15:02
bremner and both IRC and IRC-inhabitants are text oriented15:02
→ niloc132 joined15:03
bremner can't really understand how to remotely support gui's. Just old fashioned maybe.15:03
→ matjas joined15:03
matjas is there a way to `git pull` in a specific directory without using `cd`?15:03
bremner matjas: you mean in a specific git repo?15:04
Mikachu GIT_DIR=foo/.git GIT_WORK_TREE=foo git pull, might work15:04
matjas bremner: yeah, i have a limited shell and can only use `git` commands15:04
→ eijk joined15:05
bremner then you can't really merge, if there are conflicts15:05
matjas Mikachu: thanks; unfortunately that doesn’t work15:06
← h0bbit left15:06
matjas /usr/libexec/git-core/git-sh-setup: line 71: basename: command not found15:06
//usr/libexec/git-core/git-sh-setup: line 71: sed: command not found15:06
fatal: Not a git repository (or any of the parent directories): .git15:06
i suppose this shell is TOO restrictive… can’t get anything done15:06
Mikachu well, i didn't know then that you weren't in a real shell15:06
matjas thanks guys!15:06
bremner yeah, I'm not sure git itself will work15:06
matjas i just cloned a repo; that worked fine15:06
→ khmarbaise joined15:06
canton7 a lot of git commands rely on shell scripting15:07
← w0bni left15:07
← herdingcat left15:08
← nevyn left15:08
← barcoder left15:08
drizzd canton7: it's not that many, but git pull is one of them15:08
cmn the complex commands do rely on shell, like pull and rebase15:09
→ nevyn joined15:09
→ mandric joined15:09
← msponge left15:09
canton7 I count about 20 fairly common ones15:10
→ msponge joined15:10
→ asoltys joined15:10
→ dr_lepper joined15:10
cmn complex as in composite, I mean15:10
← pantsman left15:11
→ afief joined15:12
← sirclicksalot left15:12
cmn then again, even a push calls an external command, doesn't it? (at least for http)15:12
→ Sonderblade joined15:13
← Sonderblade left15:13
cbreak any smart push involves at least two processes, one on your system, one on the destination :)15:13
→ Sonderblade joined15:14
→ sirclicksalot joined15:15
→ gbacon joined15:17
→ jrm2k6 joined15:18
cmn a dumb push would also rely on something on the remote system, wouldn't it?15:18
canton7 doesn't have to be a git process, though15:18
cbreak not sure if git supports dumb pushing15:18
canton7 dumb http doesn't need a git process running on the remote system15:19
cbreak yes, but it also is read only :)15:19
cmn if all else fails you could sftp the objects over, I guess15:20
→ subbyyy joined15:20
→ aspotashev joined15:20
← subbyyy left15:20
canton7 iirc, jgit supports s3 buckets?15:20
→ subbyyy joined15:21
cmn s3 as in the amazon key-value storage?15:21
← subbyyy left15:21
canton7 yeah15:21
cmn cool15:21
→ subbyyy joined15:21
→ ReekenX joined15:21
cmn it also supports using GFS directly (instead of as a regular fs), as the filesystem actually looks a lot like git underneath15:22
→ dv310p3r joined15:22
← Octalot left15:22
cmn but I guess not many people outside of Google actually care about that15:22
→ oriba joined15:23
→ kukks joined15:23
→ tgunr joined15:28
← yshh left15:29
← dv310p3r left15:34
← sirclicksalot left15:36
→ fukenA joined15:37
→ ehsan joined15:37
→ boblet joined15:38
→ sirclicksalot joined15:39
boblet hey all. I renamed a branch locally, and now I can’t push it to the remote branch. It doesn’t appear in “Local ref configured for 'git push'” for $ git remote show origin. How do I re-establish the link?15:40
← msponge left15:41
cbreak boblet: git push doesn't normallyuse a link, it uses matching names15:41
take a look at man git-defaults push.default15:41
boblet cbreak: so my local branch has to have the same name as the remote one?15:41
cbreak no15:42
but if they have different names, normal git push won't notice15:42
you have to git push remotename localbranch:remotebranch15:42
or set push.defaults to upstream and set upstream branch with git push -u or similar15:43
boblet cbreak: aaah, that explains it. ok I’ll read up on that. thanks15:43
Nimesh → NimeshNeema15:44
← dr_lepper left15:44
← iocor left15:45
→ dr_lepper joined15:47
← dr_lepper left15:47
→ dr_lepper joined15:47
← afief left15:47
→ jpmelos joined15:47
→ afief joined15:48
← nrichards left15:50
← mandric left15:50
← na3r left15:50
← disappe__ left15:50
→ msponge joined15:50
← lapistano left15:51
→ lapistano joined15:52
← EricInBNE left15:53
→ kmc joined15:54
← jpmelos left15:54
→ jpmelos joined15:55
→ berserkr joined15:57
→ iocor joined16:00
← jpmelos left16:01
← Cromulent left16:01
← sirclicksalot left16:02
→ jpmelos joined16:02
→ sirclicksalot joined16:03
← Vampire0 left16:04
← boblet left16:05
→ Vampire0 joined16:05
→ Textmode joined16:09
← jasonbray left16:10
→ cccaldas joined16:10
→ drewis joined16:11
← drewis left16:11
← iocor left16:11
← rendar left16:11
→ vptr joined16:12
→ iocor joined16:12
→ drewis joined16:12
← drewis left16:12
← brian_g left16:14
→ Anonymous9432563 joined16:15
→ brian_g joined16:15
← ipalaus_ left16:15
→ drewis joined16:15
← Anonymous9432563 left16:17
→ nicoulaj joined16:17
→ drizzd_ joined16:18
← bahamas left16:20
← drizzd left16:21
← jpmelos left16:21
→ jpmelos joined16:22
→ Cromulent joined16:23
→ marcino joined16:23
← jimmy1980 left16:26
→ w09x joined16:26
← msponge left16:27
← iocor left16:28
← sirclicksalot left16:28
← jpmelos left16:29
→ jpmelos joined16:30
← airborn left16:30
→ iocor joined16:30
← gbacon left16:30
→ pantsman joined16:31
← pantsman left16:31
→ pantsman joined16:31
→ sirclicksalot joined16:31
→ jimmy1980 joined16:32
← apok left16:34
← lnostdal left16:34
→ kvanderw joined16:34
→ glennpratt joined16:35
← cccaldas left16:36
← Cromulent left16:37
→ macRover joined16:37
→ raythecat joined16:39
← SRabbelier|Lappy left16:39
← oc80z left16:41
→ oc80z joined16:41
→ adamkochanowicz joined16:44
← raythecat left16:45
ChanServ set mode: +v16:46
← niloc132 left16:46
adamkochanowicz I initialized a remote --bare repo, then, after cloning it into a local non-bare repo, it was successful but there were no files16:46
Do I need to add and commit first?16:46
ZyX-I Is it possible to tell git that it should never expand globs?16:46
canton7 ZyX-I, it's probably your shell that's expending the globs16:46
ZyX-I canton7: No. I don't actually observe glob expanding, just noted that it is stated in documentation.16:47
cbreak adamkochanowicz: how do you think git knows anything about files?16:47
ZyX-I canton7: Like git help add16:47
canton7 ZyX-I, I think that refers to gitignores16:48
adamkochanowicz I would think you would have to commit and add. So that's what I did, but it said the operation must be run in a working tree.16:48
canton7 well, the stuff in the description does16:48
cbreak adamkochanowicz: it's sayint ghat for a reason :)16:49
ZyX-I canton7: No. See git help add, it is stated clearly there there are fileglobs interpreted by git.16:49
canton7 the globs stuff in <filepattern> I was fairly sure was your shell doing the heavy lifting16:49
cbreak you do indeed need a working tree if you want to commit anything16:49
adamkochanowicz cbreak: so now that I have all my files under my bare repo, how do I pull from that?16:49
ZyX-I cbreak: Just do “git rm '*'” in any git repo. '*' is not expanded by shell obviously, but all files still get removed.16:50
cbreak adamkochanowicz: bare repositories do not have files16:50
adamkochanowicz: they are bare.16:50
noworking directory.16:50
adamkochanowicz cbreak: what is their purpose then?16:51
→ Samual joined16:51
cbreak they are for pushing and fetching16:51
as data exchange/storage repository16:51
adamkochanowicz how can you fetch if they are bare?16:51
cbreak with git fetch16:51
adamkochanowicz cbreak: Yes, I understand that's the command you use, but how could you fetch from something that is empty?16:51
← sirclicksalot left16:51
ZyX-I So, can I tell git it should never perform globbing?16:52
cbreak adamkochanowicz: why do you care?16:52
it's no different than if it were not empty16:52
adamkochanowicz cbreak: Let me ask a different question...16:52
First of all, I did this set up because I was told in IRC to do it this way. But I have a server with all my files on it. I want to have a local repo that pushes to and pulls from the server16:53
I was told not to push into a non-bare repo, so now I'm using a bare one16:53
but this is wrong too? :)16:53
→ offby1 joined16:53
wereHamster adamkochanowicz: on the servre you have a bare repo, on your local computer you have a non-bare repo.16:53
→ sirclicksalot joined16:53
cbreak adamkochanowicz: commit the files on the client...16:53
← jpmelos left16:53
cbreak in a non-bare repository16:53
adamkochanowicz There are no files on the client.16:54
cbreak then copy them to the client16:54
wereHamster because you use the server only to exchange data, you don't need a working tree there. But locally you work on the file so you have a non-bare repo.16:54
cbreak or make a client repo on the server16:54
wereHamster adamkochanowicz: how did you create the local repo?16:54
cbreak you can delete the files from the server after that16:54
adamkochanowicz cbreak: okay, once I do that and push to the bare repo, will I be able to view the files on the server?16:54
cbreak no16:54
→ cccaldas joined16:54
→ jpmelos joined16:54
cbreak of course not16:54
it's a bare repository16:54
← offby1 left16:54
cbreak there are no visible files16:54
wereHamster adamkochanowicz: why do you need to view the files on the server?16:55
adamkochanowicz cbreak: Now you understand.16:55
ZyX-I Forget, I see that if there is file that has name just the same as a glob, git does not perform globbing. So it should not be a problem.16:55
cbreak adamkochanowicz: I've always understood :)16:55
adamkochanowicz wereHamster: Because that is the bare minimum of what a server should do.16:55
canton7 ZyX-I, why are you so convinced that it's not the shell doing globbing?16:55
cbreak just do what I told you above and it'll work16:55
wereHamster adamkochanowicz: use gitweb maybe?16:55
cbreak the only place where you need the actual files is on the client after all16:55
adamkochanowicz cbreak: You just said that it wouldn't work.16:55
cbreak so there's no point to be able to access the files in a server repository16:56
adamkochanowicz cbreak: that is not the case, the client is not the only place where I need the actual files.16:56
cbreak that's why git doesn't bother wasting space on that16:56
ZyX-I canton7: Because I *read it in the documentation*. Just thinking about possible problems.16:56
→ texel joined16:56
cbreak of course it is16:56
by definition16:56
adamkochanowicz cbreak: what do you mean by "client"?16:56
cbreak place where you want files -> client repo16:56
a non-bare16:56
adamkochanowicz cbreak: Is that a git term or meaning my local machine?16:56
texel → Guest7877716:56
canton7 ZyX-I, the documentation's telling you a way to use 'git add', rather than something that 'git add' specifically does16:56
cbreak place where you just want data exchange -> server repo, a bare16:56
adamkochanowicz: no, it's independent of machines16:57
← TomyLobo left16:57
← koo5 left16:57
adamkochanowicz cbreak: Let me start from zero, actually, because this solution solves something I don't intend to do.16:57
ZyX-I canton7: Can you try “git rm -n '*'” in any repository that does not have file named “*”?16:57
cbreak client is where you push from and fetch into, server is where you push into, and fetch from16:57
adamkochanowicz cbreak: I see.16:57
cbreak you can't push into a repository with files16:58
(without knowing what you do)16:58
ZyX-I canton7: And you will see that it is not a shell. It does not expand '*'.16:58
adamkochanowicz cbreak: 1. I have my macbook air and a rackspace server...16:58
2. I want to edit files on the macbook and have them show up on the rs server so my coworkers can view the site16:59
canton7 ZyX-I, it attempts to remove all files and dirs in the current dir, which is exactly what I'd expect if it was the shell16:59
CareBear\ I like http://joemaller.com/990/16:59
adamkochanowicz cbreak: somewhere in that operation I need version control with git.16:59
ZyX-I canton7: Shell does not expand “'*'”. It would do this only if you remove “'”.16:59
wereHamster adamkochanowicz: then you don't need a git repository on the server. Use a deployment tool to deploy the files.16:59
cbreak adamkochanowicz: git does not solve that problem16:59
wereHamster adamkochanowicz: or even rsync, scp, ftp, whatever17:00
adamkochanowicz wereHamster: can you recommend me a good deployment tool? Other than drag and drop :)17:00
→ jkyle joined17:00
adamkochanowicz cbreak: okay.17:00
cbreak you can use git in combination with other tools if you want17:00
wereHamster adamkochanowicz: or if you are into gui apps, use your favourite gui app for that17:00
adamkochanowicz cbreak: So it looks like rsync should do the trick.17:01
ZyX-I canton7: And, from “git rm”: “If you want git to expand file glob characters, you may need to shell-escape them.”17:01
adamkochanowicz wereHamster: ideally it would be something automated17:01
ZyX-I canton7: s/git rm/git help rm17:01
adamkochanowicz cbreak: in the future, I'll have myself and my friend both working simultaneously on a website. Would you recommend each of us have copies of the repo (non-bare) and then just deploy as needed?17:03
wereHamster adamkochanowicz: yes17:03
adamkochanowicz cbreak: Because I was counting on git to merge the changes from myself and my partner on the backspace server17:03
wereHamster: okay, so I'll look into rsync17:03
wereHamster adamkochanowicz: well, that's a different story then. You said your coworkers would only view the site, not work on it17:04
bremner adamkochanowicz: push never merges17:04
→ alnewkirk joined17:04
wereHamster adamkochanowicz: if you want your coworkers work on the site in parallel with you, you need a bare repo somewhere that you'll use to synchronize your work17:04
adamkochanowicz wereHamster: sorry, two different projects.17:04
wereHamster is confused17:05
adamkochanowicz wereHamster: I think I'm good on the first project.17:05
wereHamster: starting over. I have another site where it will be me and my friend both coding on the website.17:05
cbreak adamkochanowicz: merging and deploying are two different issues17:05
wereHamster in general, if multiple people work on the same project, you need a bare repo somewhere.17:05
if you are alone, you don't need such.17:05
adamkochanowicz cbreak: who said they are the same?17:05
← cccaldas left17:05
wereHamster where you create that bare repo is not important, as long as all developers have access to it17:06
→ disappea_ joined17:07
adamkochanowicz wereHamster: so we would each have non-bare repos, push to/pull from a bare repo, and then we could just view the site on our own machines?17:07
wereHamster adamkochanowicz: no, a single bare repo that all of you use to exchange your work17:07
→ lnostdal joined17:08
adamkochanowicz wereHamster: okay, cbreak just told me I cannot view the files on a bare repo.17:08
wereHamster adamkochanowicz: you push into it, the other developers pull from it. That's how you share your work17:08
who said you need to view the files?17:08
adamkochanowicz wereHamster: I asked "we could just view the site on our own machines" and you said "no, a single bare repo…"17:08
wereHamster: maybe I misunderstood.17:09
wereHamster it's just a way how to get your changes to the other developers. They can view the files in their own repo after they pull them17:09
canton7 ZyX-I, fair play. Prefix it with a bang, it looks like17:09
wereHamster [your repo on you local machine] <-> [a bare repo somewhere] <-> [repo on your coworkers' machine]17:09
← ReekenX left17:09
adamkochanowicz wereHamster: Yes, I understand that. And to actually see and use the website, we are using our local computers, right?17:10
wereHamster adamkochanowicz: whatever works. git doesn't care what you do with the files. It's just a SCM17:10
→ harshpb joined17:10
← disappea_ left17:10
wereHamster adamkochanowicz: you can view the site locally, or in a virtual machine, or on rackspace, or ...17:11
adamkochanowicz wereHamster: But do I understand correctly that once we have our site in the bare repo that we *cannot* view the site on that server--since you cannot view files on a bare repo?17:11
ZyX-I canton7: No, “git rm '!existing_file_path'” reports that “!existing_file_path” does not match any files. But I noticed that if there is file “*” in a repository, then “git rm \*” will remove only it, same for other commands. It resolves possible issues.17:12
wereHamster adamkochanowicz: the bare repo doesn't have to be on a public server. You could use github or any other git hoster.17:12
adamkochanowicz: how to share changes with your coworkers and how you view the website are two completely different issues.17:13
adamkochanowicz wereHamster: Okay, that's fine that it doesn't have to be on a public server, but I'm trying to understand how the bare repo works.17:13
← sirclicksalot left17:14
adamkochanowicz wereHamster: I have a site on my local repo, I push it to the bare repo, I go to the server on which the bare repo is located. The site appears: Yes/No.17:14
wereHamster adamkochanowicz: it works like this: you push your changes into it. Your coworkers fetch the changes from it.17:14
no17:14
adamkochanowicz wereHamster: yes, I understand that.17:14
ZyX-I Found another problem: “git rm '\*'” will remove both '\*' and '*' files, same for other commands. How could I prevent git from doing these kind of things?17:15
adamkochanowicz wereHamster: Thank you.17:15
wereHamster adamkochanowicz: consider the bare repo a blackbox. It's just a 'place' which you can use to share changes.17:15
← glcrazy left17:15
→ msponge joined17:15
wereHamster adamkochanowicz: you can not do anything else with the bare repo than push changes into it and fetch changes from it.17:15
adamkochanowicz wereHamster: Okay, so now that that's established, I want to view the site--where do I go?17:16
← PovAddict left17:16
wereHamster adamkochanowicz: you start a local web server and view the website there17:16
or open the html file directly from the local filesystem17:16
→ sirclicksalot joined17:17
adamkochanowicz wereHamster: Okay. That is the essence of my question. thank you.17:17
wereHamster: Wow, that clears up a lot. Thank you. I've been creating and deleting repos like crazy. I needed to understand the core concept of these different things.17:19
wereHamster: So I could do this: <Macbook>--push--<bare repo>17:19
wereHamster adamkochanowicz: !book read it17:19
gitinfo adamkochanowicz: 'Pro Git' is a complete book about git that is available online for free (there's a print version, too): http://progit.org/book/17:19
← nicoulaj left17:19
adamkochanowicz wereHamster: and then <public server>---pull--<bare repo>17:19
wereHamster: I think I've looked at every piece of git documentation, the git book, youtube vids, lectures each about a million times.17:20
wereHamster: I think it's my preconceived notion of how I thought git works that is muddling my understanding of it.17:21
cbreak it's fairly easy, as I said above17:21
→ mandric joined17:21
cbreak there are two types of repositories17:21
wereHamster adamkochanowicz: what is <public server>17:21
cbreak bares and non-bares17:21
← msponge left17:21
cbreak the bares are those used for data exchange/storage17:21
adamkochanowicz cbreak: Yes, I think I'm past that now. Thank you for your help though.17:21
cbreak and non-bares are used to work with them17:21
→ jasonbray joined17:21
adamkochanowicz Okay, got to go. Thanks for your help, guys. This makes a lot more sense :)17:23
← adamkochanowicz left17:23
← psakrii left17:23
→ madewokherd joined17:23
→ gusnan joined17:25
← brian_g left17:26
→ rendar joined17:29
← jasonbray left17:30
→ brian_g joined17:30
→ maletor joined17:31
cmihai Anyone using gitlab? I setup the latest on RHEL 6.2 and I can't log in with [email@hidden.address] 5iveL!fe17:37
cbreak what's 5iveL!fe?17:39
cmihai Default password.17:41
Mikachu you didn't mean localhost then?17:41
→ airborn joined17:43
← sirclicksalot left17:43
cmihai Nope. It's in db/fixtures/production/001_admin.rb :email => "admin@local.host", :password => "5iveL!fe". Though I suppose I didn't see that get executed by rake.17:44
*Sigh*. I so hate this ruby crap.17:45
→ sirclicksalot joined17:47
cmihai Yeah, I only did a db:seutp and not a db:seed_fu.17:47
Works now.17:47
← macRover left17:48
← MissionCritical left17:49
→ macRover joined17:51
← brokenshakles left17:54
← JackWinter left17:54
→ dotnull joined17:54
→ JackWinter joined17:56
← mandric left17:56
→ raythecat joined17:57
← kvanderw left17:58
→ simplechat joined17:59
← skim1776 left18:01
← jaequery left18:03
→ metcalfc joined18:03
→ rockfruit joined18:04
→ ipalaus_ joined18:04
→ psoo joined18:05
← Liquid-Silence left18:06
← JackWinter left18:07
→ VladGh joined18:07
← dpino left18:09
→ MissionCritical joined18:10
→ JackWinter joined18:10
← sirclicksalot left18:10
→ Spockz joined18:10
→ apok joined18:11
→ sirclicksalot joined18:12
← simplechat left18:13
→ offby1 joined18:17
→ brokenshakles joined18:18
← raythecat left18:19
← ipalaus_ left18:20
→ pauld joined18:20
→ gchristensen joined18:21
← offby1 left18:22
→ glcrazy joined18:23
→ kvanderw joined18:24
→ friskd joined18:25
→ mikepack joined18:26
← canton7 left18:26
← subbyyy left18:27
→ subbyyy joined18:28
→ tshauck joined18:29
→ raythecat joined18:30
← pauld left18:34
← Bass10 left18:34
← macRover left18:34
← sirclicksalot left18:35
→ erichynds joined18:35
→ yeryry4 joined18:36
← yeryry left18:36
→ sirclicksalot joined18:36
→ kvanderw_ joined18:39
→ DarkAR joined18:41
→ towski joined18:42
→ identify joined18:44
→ jdav_gon` joined18:44
← kvanderw_ left18:45
→ kvanderw_ joined18:45
→ mandric joined18:45
→ bindaasomatic joined18:49
← towski left18:49
→ vmiklos joined18:49
→ ManDay joined18:52
← kvanderw left18:52
← kvanderw_ left18:53
← iocor left18:53
← raythecat left18:53
→ kvanderw joined18:53
→ kvanderw_ joined18:55
← jkyle left18:55
→ cccaldas joined18:56
← gchristensen left18:57
→ aspotashev| joined18:58
← vmiklos left18:58
→ airborn_ joined18:58
← airborn_ left18:58
← aspotashev left18:58
→ chromaticwt joined18:59
← harshpb left18:59
→ harshpb joined19:00
← ManDay left19:00
← sirclicksalot left19:01
← yhager_ left19:02
→ sirclicksalot joined19:03
→ fisted joined19:06
→ avinashhm joined19:06
→ vmiklos joined19:06
← kvanderw_ left19:06
avinashhm Hi , i am trying to use git log --grep option .. i see that to find word1 or word2 - --grep="word1 \| word2" works .. but to find word1 and word2 --grep="word1\&word2" doesn't .. can some please help in searching all commits for word1 and word219:07
← ProperNoun left19:10
→ jaequery joined19:11
← fermo left19:12
→ ehsan_ joined19:12
→ Bass10 joined19:13
→ cwang joined19:13
→ Danielpk joined19:14
← Bass10 left19:14
→ niloc132 joined19:14
→ Bass10 joined19:14
→ iocor joined19:16
→ tmz joined19:17
← mandric left19:18
→ aidenhong joined19:18
← aidenhong left19:19
→ rurufufuss joined19:20
Mikachu avinashhm: --grep=word1 --grep=word2 --all-match19:21
← friskd left19:21
avinashhm Mikachu, thanks very much .. this is exactly what i was looking for :-)19:23
thanks again19:23
→ cannonball joined19:24
← sirclicksalot left19:25
← ehsan_ left19:25
← BlindOracle left19:26
→ sirclicksalot joined19:27
← jaequery left19:28
→ BlindOracle joined19:29
← hyperair left19:31
→ hyperair joined19:32
→ yhager_ joined19:33
← psoo left19:34
→ ehsan_ joined19:35
← MarcoSousa1 left19:35
→ MarcoSousa joined19:36
← cannonball left19:36
→ demi` joined19:37
← tshauck left19:38
→ infogulch joined19:41
← hyperair left19:41
→ dcarper joined19:41
← kmc left19:42
→ kmc joined19:45
→ nardev joined19:45
← jrm2k6 left19:45
← tmz left19:45
→ raythecat joined19:45
→ tmz joined19:46
→ hyperair joined19:47
← sirclicksalot left19:50
← Spockz left19:51
bassliner so i have a remote repository (gitolite), that i have pushed stuff to. when checking that repo out from another machine, i have some empty directories, tho i did "git add .; git commit -a" before pushing those. any hint why?19:51
SethRobertson Git doesn't support empty directories. It will not create them. So if it created the directory there should be at least one file in them `ls -a`19:52
Does `git status` say anything?19:52
→ sirclicksalot joined19:52
SethRobertson Does `git ls-tree -r HEAD | grep directory` say thing?19:52
bassliner SethRobertson: that sees them, yes.19:54
SethRobertson Define that19:54
← diverdud` left19:54
SethRobertson (I gave two commands)19:55
← ehsan_ left19:55
→ ehsan_ joined19:56
← iocor left19:56
← yhager_ left19:57
→ meLon joined19:58
← meLon left19:58
→ meLon joined19:58
kvanderw → kvanderw|afk19:58
→ disappea_ joined19:58
bassliner SethRobertson: well when i do a fresh git clone of that repo, i get some empty directories there, but on the machine that i pushed my stuff to that repo, i have stuff inside that dirs.19:59
SethRobertson: git status says nothing to commit there.19:59
cbreak bassliner: you're using submodules?19:59
SethRobertson But `git ls-tree -r HEAD | grep directory` shows the filename on the newly cloned repo?20:00
cbreak if so, do a recursive clone, or do a git submodule update --init20:00
← ehsan_ left20:00
bassliner SethRobertson: that shows that directory.20:00
SethRobertson: but not contents20:00
SethRobertson How about on the original repo?20:01
→ igotnolegs joined20:01
→ jkyle joined20:02
← kmc left20:04
→ drev1 joined20:04
→ ProperNoun joined20:05
bassliner http://fpaste.org/D3dL/20:06
that's basically what i do here20:06
→ ehsan_ joined20:06
← ehsan_ left20:06
← marcuy left20:06
← harshpb left20:06
← jkyle left20:06
→ heppy joined20:07
→ marcuy joined20:07
SethRobertson If you do `git ls-files -o -i --exclude-standard` on the source repo, does it show index.theme in it?20:08
I can't see why it would have created an empty directory on the cloned side.20:08
bassliner no.20:08
"git ls-files -o -i --exclude-standard" on acid in .themes does not show anything.20:09
SethRobertson If you go in Clear* directory and say `git add index.theme` and the `git status` does it say anything?20:09
← ianmcorvidae left20:10
bassliner oh now i see my mistake:20:10
fatal: Path 'Clearsightz-Darky/index.theme' is in submodule 'Clearsightz-Darky'20:10
and i simply went into Clearsightz-Darky and rm'ed the .git dir20:10
which probably broke it20:10
SethRobertson +1 to cbreak20:10
→ tshauck joined20:10
SethRobertson You might need to edit .gitmodules in the git root and then `git rm --cached Clear*ky` commit that and then you can probably re-add20:11
cbreak you can do that in one commit20:12
→ bewartse joined20:12
cbreak but don't expect that to propagate well to existing repositories20:12
← alnewkirk left20:13
→ ehsan_ joined20:15
← sirclicksalot left20:15
← ehsan_ left20:16
← haydenmuhl left20:17
→ [haydenmuhl] joined20:17
→ sirclicksalot joined20:17
→ BiggFREE joined20:17
← rurufufuss left20:18
← demi` left20:18
← Danielpk left20:18
← ProperNoun left20:18
bassliner hmm. i get "fatal: pathspec 'Clearsightz-Darky' did not match any files" when doing git rm -rf Clearsightz-Darky here.20:18
← codeshah left20:19
← kadoban_ left20:19
→ __iron joined20:19
SethRobertson You will note I said --cached. Does that directory still exist and have content?20:19
bassliner uhm yes, it does.20:20
SethRobertson Try with --cached and then try `git add Clear*ky`20:20
bassliner even with --cached i get "did not match any files"20:20
SethRobertson Fine. Proceed.20:21
Did you have anything in .gitmodules?20:21
bassliner not sure at what step i lost .gitmodules tbh20:21
but it's not there anymore20:21
ok seems like i'm screwed here.20:21
SethRobertson how so?20:21
bassliner what you mean by "proceed"?20:21
SethRobertson git add Clear*ky20:22
→ gg0 joined20:24
← nevyn left20:25
bassliner hm, i don't get it - i even see the files in .git/index20:26
→ kmc joined20:26
SethRobertson What do you not get?20:27
What happened when you ran that command?20:27
→ nevyn joined20:27
cbreak morale: don't use submodules unless you know what they are good for :)20:27
Mikachu you need to quote the pattern if it won't match any actual files20:27
bassliner well why git push still tells me "Everything up-to-date" when doing a git push, but when checking that out, i still get empty Clearsightz-Darky directory.20:27
SethRobertson cbreak: Any use of submodules will harm your moral20:28
Mikachu or just use the actual pathname as you did first20:28
→ redheadphones joined20:28
cbreak bassliner: did you commit?20:28
bassliner ah that's it. thank you.20:28
*punches self*20:28
← heppy left20:29
gg0 hi, I did "git remote prune origin" and seems I lost everything. any way to undo that?20:30
SethRobertson Define "lost everything"20:30
cbreak gg0: git remote prune origin will just delete remote tracking branches for which the remote does not have a branch anymore20:31
gg0 I had "master" and "upstream" branches which were "stale (use 'git remote prune' to remove)"20:31
and I had the bad idea to prune them20:32
cbreak well, after pruning they are gone.20:32
Mikachu if you remember their hash, you can create them again20:32
SethRobertson gg0: What does `git branch -a` say?20:32
gg0 Sebastian_: nothing20:32
← glennpratt left20:32
gg0 Sebastian_: sorry20:32
SethRobertson: nothing20:33
SethRobertson Impressive. Do you *have* a remote? `git remote`20:33
→ yhager_ joined20:34
← __iron left20:34
gg0 "git remote show origin" says origin is still there with just fetch and push urls20:34
← xsbeats left20:34
→ traviscj joined20:35
gg0 no branches20:35
→ heppy joined20:35
SethRobertson And if you `git fetch origin; git branch -a` you still have no branches?20:35
→ mandric joined20:35
← dr_lepper left20:35
← localhost left20:36
gg0 remotely branches are still there. problem is I had many changes to push20:36
SethRobertson So when you said "no branch" you mean "no local branches"?20:37
gg0 yep local ones20:37
← metcalfc left20:37
SethRobertson Please don't censor in the future. OK, next look at `git log -g | less` and see if you have a few example SHAs which might be of interest.20:38
→ localhost joined20:39
→ pidus joined20:39
← sirclicksalot left20:40
← traviscj left20:40
→ glennpratt joined20:40
→ iocor joined20:40
← glennpratt left20:40
gg0 fatal: bad default revision 'HEAD'20:41
SethRobertson That was the output of `git log -g` (NOT `git log`)?20:41
I guess you could say `git reflog` and see if that behaves differently20:42
→ dr_lepper joined20:42
← dr_lepper left20:42
→ dr_lepper joined20:42
gg0 same output even with git reflog20:42
→ sirclicksalot joined20:43
← dstorrs left20:43
SethRobertson OK. Next step. `git fsck` You will need to go through the unreferenced commits and find the last commit you made on each branch.20:43
gg0 though I found the last commit hash in the backlog20:43
SethRobertson You are confusing me again20:43
By "backlog" I assume you mean "reflog" which was the command you *just* said produced an error20:44
← redheadphones left20:45
→ redheadphones joined20:45
gg0 backlog = scrolling back terminal20:45
← nevyn left20:46
gg0 all git log, git log -g, git reflog give me bad default revision HEAD20:46
SethRobertson OK, then say `git checkout -b branchname SHA`20:47
And then `git branch branchname --set-upstream origin/branchname`20:47
gg0 git fsck http://paste.debian.net/plain/14883120:48
→ Spockz joined20:48
→ nevyn joined20:48
← Spockz left20:48
← cccaldas left20:49
SethRobertson Recreate the branch *first* then look at `git fsck` You will specifically want to examine the "dangling commits"20:49
← BiggFREE left20:53
← JackWinter left20:53
→ cannonball joined20:53
← aspotashev| left20:53
SethRobertson You can try, for example, saying `gitk --all --date-order SHA` where the SHA is from a dangling commit to try and visualize it20:54
const I accidentally branched from the wrong branch so my latest commit has a head which is wrong20:54
→ Danielpk joined20:54
SethRobertson const: So rebase or cherry-pick it20:54
→ airborn_ joined20:54
const I have HEAD, HEAD^ and HEAD^^. I'd like to move HEAD^ and HEAD^^ to a different branch20:54
← airborn_ left20:54
← yhager_ left20:55
SethRobertson const: git checkout goodbranch; git cherry-pic ${SHAofSHEAD^^} ${SHAofHEAD^}; git checkout badbranch; git rebase -i HEAD^^^20:56
const SethRobertson: and what do I use instead of 'pick'?20:56
SethRobertson ??20:58
Oh, in the rebase -i step? Just delete the lines20:58
const SethRobertson: thanks!21:00
→ kerrick joined21:00
→ felipec joined21:01
→ JackWinter joined21:01
→ ProperNoun joined21:03
← ProperNoun left21:03
napdragon → spyro21:03
← sirclicksalot left21:03
→ dstorrs joined21:05
→ csgeek joined21:05
→ sirclicksalot joined21:05
→ psoo joined21:06
→ jrm2k6 joined21:06
← berserkr left21:07
→ nowhereman joined21:08
const is it possible to see every branch with a certain commit ?21:08
Mikachu git branch --contains=somehash21:09
← rockfruit left21:10
const Mikachu: that isn't giving me what I want. I'm certain I committed commit X to a different branch but I don't see it21:10
← bartek left21:11
→ EyesIsServer joined21:11
← nowhere_man left21:11
← zub left21:11
→ zub joined21:11
SethRobertson Did you cherry-pick it?21:11
Can you find it with `git log -S`?21:12
Can you find it with `gitk --all --date-order`?21:12
const looks for it21:12
const SethRobertson: strange. I had X -> Y -> Z and it it seems that cherry-pick X _moved_ Y and Z along with X21:14
← rtyler left21:14
const just needs to fix my repo :-\21:14
← koo1 left21:14
SethRobertson Seems extremely unlikely21:14
cbreak that's not what cherry-pick does21:15
→ croby joined21:16
croby is there any way to prepend to the default commit message used when committing a squashed commit? i don't want to open the editor, just want to specify something on the command line21:17
← dstorrs left21:17
→ demi` joined21:18
SethRobertson Change your $GIT_EDITOR for that command to a command which prepends that message to the file listed on $1?21:18
croby oh interesting idea21:18
→ HG` joined21:21
→ caseymcg joined21:22
← niloc132 left21:23
→ niloc132 joined21:23
→ koo4 joined21:25
← riper left21:25
← cannonball left21:26
← iocor left21:26
← rendar left21:28
→ Octalot joined21:29
Sidnicious Quick poll: how do you all handle long lines in your commit messages — Do you have your editor set up to wrap at some number of characters, wrap by hand, write short commit messages, not worry about long lines, or something else entirely? How do you usually view your commit messages (git log, a GUI program, GitHub)?21:29
cbreak I just wrap by hand.21:29
→ alnewkirk joined21:30
thiago I have Emacs set to wrap automatically at 7221:30
and a commit hook that tells me if I'm past it21:31
jessopher yeah, i just wrap at around 75ish21:31
→ rockfruit joined21:31
thiago and the review system also reports it21:31
→ yshh joined21:31
← sirclicksalot left21:31
→ gchristensen joined21:32
Sidnicious All right, cool21:32
← bindaasomatic left21:34
← raythecat left21:35
Sidnicious Why do you do it (and why does the review system not like commits with long lines)?21:35
→ sirclicksalot joined21:35
thiago because they are meant to be 72 characters long21:35
← yshh left21:36
const I just accidentally deleted a branch which I didn't mean to21:37
is there any way to recover it?21:37
(it was git branch -D blahblah002)21:37
I typed nothing since then21:37
cbreak do you remember the sha?21:38
git branch -D might show it21:38
if so, just branch name sha21:38
→ offby1 joined21:38
← offby1 left21:38
→ offby1 joined21:38
← identify left21:38
← LongBeach left21:39
const cbreak: thanks!21:40
(more reasons to prefer 'closing' branches instead of deleting them)21:40
RandalSchwartz "closing" how?21:40
a branch is just a named commit21:41
const RandalSchwartz: mercurial concept. hg commit --close will make hg branch not show the branch unless you ask to show closed ones as well21:41
cbreak why not just tag it then delete it?21:42
const cbreak: its just simpler to to have it "out of the way". its essentially one extra bit attached to a branch21:42
cbreak useless extra bit :)21:42
const (yes, there are workarounds, but closing is easier)21:42
jessopher hah21:43
const cbreak: my point above was that when you start *deleting* things it makes it much easier to make a mistake21:43
cbreak one reason why git is so powerful compared to lesser alternatives is that it is extremely simple21:43
← DarkAR left21:43
const instead, do the "natural" thing and just say "I'm done with this branch, don't show it to me again"21:43
cbreak you can do the natural thing and say:21:43
ok, I never want to see this branch again21:43
let's delete it21:44
→ lemonchicken joined21:44
const cbreak: but thats *not* what I want21:44
cbreak or I want to keep this branch, let's tag it and then delete it21:44
const cbreak: also, *not* what I want to say21:44
jessopher the natural thing is just to snap the branch of and wait for it to grow back!21:44
RandalSchwartz it's so easy to recover a deleted branch though21:44
const cbreak: I want to say "lets keep this branch as-is on the off-chance I need to look at again, but lets put it out of mind unless I specifically ask for it"21:44
RandalSchwartz as long as you do it within 90 days21:44
and if you want it longer than that, yes, tag it21:45
cbreak so delete it and tag it const21:45
→ dstorrs joined21:45
const cbreak: understood, but that's a workaround for the lack of ability to say what I mean21:45
cbreak: I know *many* ways it can be done21:45
→ ProperNoun joined21:45
cbreak I have a whole bunch of deleted branches tagged in deac/workingthing1, dead/stupididea2 and so on21:45
const but my point was that the 'closing' concept is more natural in this case21:45
RandalSchwartz natural to you21:46
not to me21:46
cbreak closing makes no sense21:46
RandalSchwartz so it's an hg thing21:46
cbreak what if you want to continue?21:46
RandalSchwartz not a "natural" thing21:46
const RandalSchwartz: it happens to be that hg implements a natural thing21:46
RandalSchwartz You keep saying that like it's true for you, but it's not a universal truth21:47
const notes that I use hg, git, svn, p4, cvs, and fossil on a regular basis21:47
jessopher nature a vcs dont have much overlap in my experience21:47
cbreak I don't see how it is natural21:47
RandalSchwartz I am an example of someone for whom it is not natural.21:47
Mikachu if you want you can git branch -m thebranch closed/thebranch and just mentally ignore closed/*21:47
RandalSchwartz apparently, many in this channel agree21:47
cbreak I use namespaced tags for that21:47
const cbreak: RandalSchwartz does: "put this branch off to the side" mean "give this branch a different name and delete it"21:47
RandalSchwartz I'd say you're in the minority here21:47
← flyz left21:47
RandalSchwartz const - branches are fictional. just a name for a commit.21:47
← offby1 left21:48
RandalSchwartz when you get that a branch is a point, not a line, it becomes clearer.21:48
cbreak const: if you move it, you should not delete it unless you don't want to keep it21:48
→ offby1 joined21:48
const cbreak: but then git branch -a shows it21:48
cbreak renaming a branch renames a branch, it doesn't copy it21:48
RandalSchwartz I rapidly create and delete branchnames whenever I want a named commit21:48
cbreak I was talking about using tags21:48
← offby1 left21:48
→ offby1 joined21:48
cbreak not branches21:48
const cbreak: *sigh*21:48
RandalSchwartz I never think of them as lines. just points.21:48
cbreak because tags are not shown when you do git branch -a21:48
const meh21:49
const -> food();21:50
drizzd_ const: hmm, I use this script to hide branches: https://gist.github.com/d659b07efe14b9217abe21:50
← erichynds left21:50
const drizzd_: I have something like that21:50
drizzd_: there are many workarounds21:50
const -> food();21:50
cbreak no need for workarounds :)21:50
luckily, git supports everything natively21:50
with the same repository format as used years ago21:50
drizzd_ cbreak: how can I do this natively?21:51
cbreak that's the power of flexibility emerging from simple foundations and advanced analysis21:51
drizzd_: do what?21:51
← ehsan left21:51
drizzd_ never mind...21:51
jessopher i think some of the confusion of branch as a named commit, is it really isnt. A tag is, but a branch represents a series of commits, that once were, but now are not, named.21:51
← pidus left21:51
cbreak no21:52
a branch represents one commit21:52
a branch and a tag are 100% the same thing21:52
they arejust in different subfolders inside .git21:52
jessopher until you start using them21:52
cbreak so they get treated differently21:52
SethRobertson Except 1) git provides an api so that you can trivially move a branch and you cannot trivially move a tag, and 2) they are treated different on fetch and push class operations21:54
drizzd_ const: a branch graveyard has been discussed repeatedly, but unfortunately there are some technical difficulties, and it is also difficult to find a consesus on what exactly it _should_ do21:54
jessopher to call them the same, and then beable to ascribe different properties and semantics to their usage is deceptive. They share the same concrete representation, but they are certainly not the same21:54
SethRobertson But under the covers they are exactly the same21:54
drizzd_ in particular, how long deleted branches should stay around21:54
cbreak not at all21:54
if I delete a branch then I want it gone21:55
I often create branches to easily "cherry pick" a series of commits by rebasing them all at once21:55
those get killed within one minute or two21:55
if you want to keep branches alive but hide them, it'd be more reasonable to implement some kind of ignore-ability for refs21:56
SethRobertson Which is what the script moving stuff to hidden does very well21:57
← psoo left21:58
← sirclicksalot left22:00
Sidnicious thiago: Didn't want to interrupt :). Why are they meant to be 72 characters long? I'm asking because I just stopped worrying about it a little while ago. I realized that, for me, the *only* reason to wrap commits was that git log and git show didn't know how to wrap them. Most GUIs do. Hell, my boss at work uses a Git UI to do all of his commits which uses a proportional font and doesn't provide any way to force line wrapping.22:00
← tmz left22:00
SethRobertson "They?" Commit messages?22:01
Sidnicious Right, sorry22:01
→ sirclicksalot joined22:01
SethRobertson They don't have to be, but it helps command output line up on a standard 80 column display. eg `git log --oneline`22:02
Remember this is only true of the first line (and the next line must be blank)22:02
Sidnicious I started working on a patch a little while ago which would let `git log` and the like do the wrapping when they display commit messages, which would seem to solve the problem completely, for me. But, I just wanted to get a feel for how other people work.22:03
SethRobertson It already does wrap.22:03
It just isn't at "neat" when displayed onscreen22:03
Sidnicious Yeah, I'm all in favor of making the first line short, for the same reason it's good to keep email subjects short. I'm mainly talking about the rest of the commit (after the blank line), which can get longer in some cases22:03
Hm?22:04
← ramusara left22:04
→ boombatower joined22:04
Sidnicious When does it wrap?22:04
SethRobertson Subsequent lines can be as long as you want, though 80 columns might be a good suggestion, again so you might not get poor line wrapping22:04
For example: `git log --oneline` to pick a random suggestion that I have not mentioned in at least two minutes.22:05
But `git log` as well, and `git show` and...well everything.22:05
Of course it doesn't perform word wrap, it performs character wrap.22:05
Sidnicious For me, none of those commands wrap at all… I think that's the default.22:06
i.e. I have to scroll left and right if a line is wider than my terminal22:06
SethRobertson That is a problem with your terminal, not with git. Other terminals behave differently.22:07
drizzd_ Sidnicious: that's because by default git calls less with -SFRX, where -S says "do not wrap"22:07
Sidnicious ^22:07
SethRobertson Ah, I have set GIT_PAGER=""22:08
→ metcalfc joined22:08
SethRobertson However, even if I unset it, it still works the same page22:09
sorry, same way. It still character wraps22:09
Sidnicious Well, what if git did word wrapping?22:09
→ fisted_ joined22:10
SethRobertson It certainly could not be allowed to do it if the output wasn't going through GIT_PAGER since there could be programic users needing exact input/output identity.22:10
← afief left22:11
← fisted left22:11
→ tmz joined22:11
Sidnicious It would, absolutely, be optional. Usually when git output is being piped to a program, you'd use a custom format string, right?22:12
SethRobertson Not at all.22:13
Sidnicious How do you usually use it?22:13
SethRobertson --oneline, as a unique example, is something I often use for speed. Also grep'ing the lines I care about, etc.22:13
→ neurodrone joined22:14
drizzd_ SethRobertson: funny, I can't get it to _do_ word wraps, even with GIT_PAGER='less' git log. But git log | less works as expected... IIRC git does something strange here...22:15
SethRobertson It does character wraps, not word wraps, for me. But someone was claiming earlier that git used default flags. Try ^Z and inspect the arguments (and $LESS environmental variable) of less22:16
ps axewww | grep less22:16
← brian_g left22:17
← chromaticwt left22:17
← tmz left22:17
Sidnicious My proposed patch would, right now, not affect --oneline. It would mainly be directed at git log, show and the like with --pretty=medium and up22:18
← nardev left22:18
→ tmz joined22:18
Sidnicious Optionally, of course, as a config item/command line flag22:18
drizzd_ SethRobertson: ah indeed, git sets LESS, so this works: LESS= GIT_PAGER='less -FRX' git log22:18
SethRobertson Or it seems you could just adjust the flags that git passes along.22:19
← tmz left22:20
drizzd_ a fix to this very problem seems to be documented in git-config(1): For example, to disable the S option in a backward compatible manner, set core.pager to less -+$LESS -FRX22:21
fisted_ → fisted22:22
Sidnicious The idea is to do word wrapping, respecting indentation, so that commits look good even if they weren't wrapped beforehand22:22
→ tmz joined22:23
→ aaronfeng joined22:23
Sidnicious Here's an example. With LESS= GIT_PAGER='less -FRX': http://cl.ly/1u0a1x0y3N1V1C1d3q3F/Screen%20Shot%202011-12-10%20at%205.23.38%20PM.png22:24
Patched git: http://f.cl.ly/items/2A2r1A3M1Y0R0b423z13/Screen%20Shot%202011-12-10%20at%205.23.41%20PM.png22:24
Mikachu Sidnicious: if you autowrap commit messages, you can't draw ascii art in them safely ;)22:25
const Mikachu: :)22:25
Sidnicious :)22:26
← jimmy1980 left22:26
← boombatower left22:26
← sirclicksalot left22:26
← subbyyy left22:27
Sidnicious Well, as it is now, the patch prints lines verbatim if they begin with whitespace, which seems to be the convention for that sort of thing. As you can see in that commit messages, it indents lists properly too22:27
er, as you can see in the screenshot22:27
→ subbyyy joined22:27
drizzd_ FWIW, there is also a format option to do something like that: git log --format='%w(72)%B'22:28
Sidnicious Mikachu: that help?22:28
→ sirclicksalot joined22:28
Sidnicious That's true, but there's no way to use it with --pretty22:28
→ jimmy1980 joined22:30
aaronfeng is there a way to force push a branch even if the branch is currently checked out? the checked out branch should never have any changes.22:30
→ nardev joined22:31
Mikachu if you're pushing to a non-bare repo, don't22:32
aaronfeng Mikachu: yes, I'm pushing to a non-bare repo. I have a vm setup for testing, I want to push directly to it. however, I can't when that branch is checked out.22:34
RandalSchwartz aaron - don't do that22:34
Mikachu no support for that22:34
RandalSchwartz push to bare repos22:34
Mikachu it will lead to pain, suffering and death22:34
RandalSchwartz and widespread adoption of suboptimal family values!22:35
aaronfeng I am pushing to a bare repo in the end, this is just my testing cycle22:35
RandalSchwartz move your testing to the end then :)22:35
Mikachu if you push to the checked out branch, the working tree and index will not be updated22:35
← subbyyy left22:35
→ felixsulla joined22:35
← glcrazy left22:36
SethRobertson aaronfeng: If you will never ever modify the target repo, you can use a hook to do it22:36
Mikachu you can check out no branch in the target repo i guess22:36
aaronfeng SethRobertson: yes, I will never modify that repo22:36
Mikachu (then why is it not bare?)22:36
RandalSchwartz I think the "push to live" works far better"22:36
SethRobertson aaronfeng: http://toroid.org/ams/git-website-howto22:37
RandalSchwartz I have my bare repo set up so that when I push to the "live" branch, it also git-archive | tar xvf22:37
works nicely22:37
git push origin master:live22:37
and that also ensures that I know what was pushed live last22:37
since it's the live history22:37
SethRobertson The toroid.org thing is `git push live`22:37
Well, the live history is an advantage22:38
RandalSchwartz ahh - that uses git checkout -f22:38
SethRobertson I guess the toroid would have it in the reflog22:38
RandalSchwartz I'm using git only to manage a portion of the site22:38
so I have to be careful what I add/delete22:38
← tshauck left22:47
→ zehrique joined22:49
← drizzd_ left22:50
← diminoten left22:53
→ ehsan joined22:54
← sirclicksalot left22:55
→ codeshah joined22:55
← HG` left22:57
→ sirclicksalot joined22:57
← mcluskydodallas left22:57
← marcino left22:58
← khmarbaise left22:59
← eijk left22:59
← the-fonz left23:01
→ Lacrymology joined23:02
→ diminoten joined23:02
← Lacrymology left23:02
gg0 great. restored all 3 branches by playing with checkout -b name SHA + dangling commits from fsck23:02
→ Lacrymology joined23:03
gg0 thanks for your hints SethRobertson23:03
← jrm2k6 left23:03
← gusnan left23:04
gg0 restored branches are stale again. I think I'll keep them stale this time ;)23:05
→ iocor joined23:06
← Lacrymology left23:07
→ Lacrymology joined23:08
→ Per_Vers2 joined23:09
← Per_Vers2 left23:09
← diminoten left23:09
→ diminoten joined23:09
← bewartse left23:10
← felipec left23:11
← Lacrymology left23:13
→ Lacrymology joined23:13
← Lacrymology left23:14
SethRobertson I have to think something else is going wrong. What version of git are you running?23:15
← ehsan left23:16
→ nixmaniack joined23:17
← Shusshu left23:17
← caseymcg left23:19
→ philadams joined23:19
→ raythecat joined23:21
→ ziggles joined23:22
→ brian_g joined23:22
← sirclicksalot left23:22
gg0 debian testing $ git --version23:22
git version 1.7.7.323:22
← kmc left23:23
← Morphous left23:24
→ cannonball joined23:24
gg0 http://paste.debian.net/plain/14886423:24
→ sirclicksalot joined23:25
← aaronfeng left23:27
← cmyk_ left23:27
→ canton7 joined23:28
→ berserkr joined23:28
← MarcoSousa left23:30
← cannonball left23:32
→ jasonbray joined23:33
→ Pilif12p joined23:34
→ felipec joined23:37
← jasonbray left23:38
→ Morphous joined23:39
→ jrm2k6 joined23:41
← dcarper left23:42
← metcalfc left23:45
→ chromaticwt joined23:46
→ aaronfeng joined23:47
← sirclicksalot left23:48
→ Chibby_ joined23:48
← frogonwheels left23:50
→ sirclicksalot joined23:51
← marcuy left23:53
→ subbyyy joined23:54
← jrm2k6 left23:54
← Chibby_ left23:57
← rockfruit left23:57

Logs Search ←Prev date Next date→ Channels Documentation