IRCloggy #git 2012-06-26

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.

2012-06-26

← wagle left00:00
← mikepack_ left00:02
→ leilaM joined00:03
← Zespre left00:03
→ Zespre joined00:03
→ kukks joined00:04
← kukks left00:04
→ kukks joined00:04
→ aspotashev joined00:04
← ashphy left00:05
← cjz left00:06
→ yizumi joined00:06
→ p3rror joined00:07
→ oriba_ joined00:08
← twisted` left00:09
← leilaM left00:10
← dzonder left00:10
→ Dave^| joined00:10
← aselder left00:11
← teknopaul left00:11
→ _huoxito joined00:12
← oriba left00:12
→ mishok13 joined00:12
→ wagle joined00:13
→ twisted` joined00:13
← adammokan left00:14
← neurodrone left00:15
→ rchavik joined00:15
← rchavik left00:15
→ rchavik joined00:15
← ebneter left00:15
→ iwoj joined00:16
← LekeFly left00:17
← tewecske left00:17
→ jjbohn joined00:18
← teweWORK left00:18
→ LekeFly joined00:18
→ lnostdal__ joined00:19
← serj left00:19
← mishok13 left00:20
→ serj joined00:20
← raykyri left00:20
← serj left00:20
← p3rror left00:20
← TrahDivad left00:21
→ mritz joined00:21
← Sil4nc4 left00:21
→ serj joined00:21
← mklappstuhl left00:22
← lnostdal_ left00:22
→ ebneter joined00:22
→ ziggles joined00:23
← Destos left00:23
→ Targen joined00:25
→ zx joined00:25
oriba_ → oriba00:25
← andref_dias left00:25
→ raykyri joined00:27
→ jbohn joined00:28
→ apok joined00:29
← hyperair left00:30
← jjbohn left00:30
→ hyperair joined00:31
← cdash004 left00:32
← Cromulent left00:33
→ p3rror joined00:33
← iamjarvo left00:34
← tr-808 left00:35
→ tr-808 joined00:35
← zx left00:35
→ kephas joined00:36
→ leo2007 joined00:36
leo2007 switching branches usually causes massive recompilation, which can take a long time. Has anyone been annoyed by this problem?00:36
bremner yes00:37
← yizumi left00:37
← jbohn left00:37
→ Escherial joined00:38
bremner dunno anything sensible to do about it though.00:38
Escherial hey, is there a way when resolving merge conflicts to simply take one of the two sources?00:39
← toothr left00:39
→ toothrot joined00:39
← toothrot left00:39
→ toothrot joined00:39
← nowhereman left00:39
Escherial i've pulled a commit with conflicts to my base, but i know the remote's correct and i don't want to slog through manually merging it all00:39
→ jhardinee joined00:39
leo2007 Escherial: checkout git merge -s theirs00:41
avpx ^00:41
→ booginga joined00:42
→ BarryCrayon joined00:42
← aantix left00:43
← ehsan left00:43
← HyperNerdV2 left00:44
← etcetera left00:44
→ ehsan joined00:44
→ tigrang joined00:45
← Zespre left00:46
→ Zespre joined00:46
→ etcetera joined00:46
← dreiss_ left00:46
← ipalaus left00:46
← BarryCrayon left00:47
← SkullG4m3 left00:47
Escherial erm, is it still possible to merge now that i'm in a conflicted state?00:47
← ology left00:47
Escherial i'm not really sure what that means, by the way...how does git keep track of the fact that i'm conflicted?00:47
bremner Escherial: git status00:48
well, technically the index00:48
← ehsan left00:49
← mritz left00:49
← oriba left00:49
bremner Escherial: one option is to use the output from git status -s and run git checkout remote/branch file on the appropriate files00:49
a one-liner in a shell00:49
Escherial ah, if it's just a matter of checking them out from the remote branch, i think i can handle that00:50
← wagle left00:51
→ kenperkins joined00:52
Escherial sorry, have to go; thanks for the advice00:52
→ wagle joined00:52
bremner good luck00:52
Escherial thanks :)00:52
← Escherial left00:52
← SkAzZ left00:52
→ bigsky_ joined00:53
← r0bby left00:56
→ ehsan joined00:58
← ISF left01:01
← MrGlass left01:01
→ MrGlass joined01:02
→ martialcoder joined01:02
← andrewhl left01:04
→ Bass2 joined01:04
← tonsofpcs left01:05
← mtytel left01:05
infogulch → infogulch|afk01:05
← etcetera left01:05
← Bass2 left01:05
→ tonsofpcs joined01:06
→ xinming_ joined01:06
→ ashphy joined01:06
← Goplat left01:07
← Bass10 left01:08
← avpx left01:08
← martialcoder left01:09
← xinming left01:09
← hwrd|work left01:10
← erkules left01:10
→ erkules joined01:11
← akiress left01:11
→ Vampire0 joined01:12
← maletor left01:14
← elnn left01:14
adityamenon ah, beautiful. so with -s theirs, what happens is git chooses 'their' versions of files, ignoring our files completely01:14
I knew it was something on those lines, but did not the exact details01:14
words like *merge strategy* got me petrified, thinking theres some complicated superlogic inside...01:15
← Vampire0_ left01:15
← btanaka left01:15
→ elnn joined01:16
milki aha01:17
→ brentratliff joined01:18
→ andrewhl joined01:19
← palm_vue left01:20
← Lgb left01:20
→ keller joined01:21
→ HyperNerdV2 joined01:21
→ Joseph_ joined01:21
→ palm_vue joined01:21
→ Lgb joined01:21
nevyn adityamenon: merge strategy is actually one of my fav bits of git.01:21
in that they're pluggable01:21
so you could have a merge strategy for arbitary data formats.01:22
← kermit1 left01:22
→ kermit1 joined01:22
adityamenon does that mean you write new implementations of merge to suit new data formats?01:22
kermit1 → kermit01:22
← Joseph_ left01:23
← bgerber left01:23
→ Joseph_ joined01:23
← jhardinee left01:23
→ Goplat joined01:24
← mocramis left01:24
← duckxx left01:25
→ mocramis joined01:25
← cmullican left01:25
→ jhardinee joined01:26
→ dreiss_ joined01:26
adityamenon wat?01:27
http://www.kernel.org/pub/software/scm/git/docs/git-merge.html01:27
under "ours"01:27
This option forces conflicting hunks to be auto-resolved cleanly by favoring our version. Changes from the other tree that do not conflict with our side are reflected to the merge result.01:27
Then again01:27
"This should not be confused with the ours merge strategy, which does not even look at what the other tree contains at all. It discards everything the other tree did, declaring our history contains all that happened in it.01:27
"01:27
WHAT should not be confused with what?01:28
milki "ours" strategy01:28
and "theirs" strategy01:28
adityamenon the paragraph describes what "ours" does - then says don't confuse with itself?01:28
milki hmm?01:29
cmn there are two ours01:29
one is a merge strategy, the other one an option to recursive01:29
adityamenon two ours?01:29
milki o01:29
interesting..01:29
so theres "ours" and _ours_01:29
haha01:29
adityamenon indeed... the paragraph I extracted is an option to recursive01:29
← Joseph_ left01:29
cmn it's all in man merge01:29
→ Joseph_ joined01:30
cmn though recursive being in the middle can get confusing; use the indentation01:30
← imexil left01:30
← leo2007 left01:30
← suedoh left01:30
← jwmann left01:31
← ben_alman left01:32
← ashphy left01:32
→ jwmann joined01:32
→ adityamenon90 joined01:32
→ mikepack joined01:32
→ bgerber joined01:32
← lonewulf` left01:32
→ ashphy joined01:32
← tyatpi left01:33
→ diogogmt joined01:33
← jhardinee left01:33
← adityamenon left01:34
adityamenon90 man merge is actually more beginner friendly than many blog posts01:34
→ tyatpi joined01:35
adityamenon90 first time I've experienced that with any documentation01:35
→ lonewulf` joined01:35
adityamenon90 → adityamenon01:35
→ ben_alman joined01:35
MetaCosm Any tips on emulating the integer expansion of $Rev$ in SVN on the git side01:35
adityamenon I only saw this man in a command line, and since command line can make mans look horrid, I did not read much01:35
→ intripoon_ joined01:37
cmn integer expansion?01:37
← ashphy left01:37
frogonwheels MetaCosm: There are many reasons not to.. but you can .. SHA1 is the safest, but the problem is you have to take out the expansion to get the sha1 of the file01:38
→ vitor-br joined01:38
MetaCosm frogonwheels: this moronic system uses it as an upward moving int value01:38
based on the SVN $Rev$01:38
cmn you can use git-describe to get the version of the tree01:38
MetaCosm to do versioning between components01:38
cmn though it won't necessarily give you a number01:39
→ Syzygy_ joined01:39
frogonwheels MetaCosm: what I have done is to use git describe to give me a version that I compile into a resource & include as s define.. so it gives me a compile-time version01:39
→ leilaM joined01:39
frogonwheels the point being that the file generated with the version is not in source-control01:39
cmn and if you need to describe dependencies, version numbers are much better than relying on revisions01:40
MetaCosm ... ugh, they got it expanded in like 10 places01:40
← intripoon left01:40
MetaCosm cmn: not arguing for correctness, just trying to get them off SVN01:40
→ alvaro_o joined01:40
MetaCosm so I Can start fixing stuff01:40
this is the last stumbling block01:40
cmn then you need a sensible system01:40
frogonwheels MetaCosm: Is this a run-time version or source version that it's doing?01:40
MetaCosm cmn: no argument, hard to build a sensible system on svn01:41
hence my new employment and moving to git01:41
cmn no harder than on anything else01:41
MetaCosm no need to beat the dead horse, I entirely agree!01:41
Just trying to come up with a mid-step01:41
frogonwheels MetaCosm: I had to move my team from Beyond Compare. You got it relatively easy ;P01:41
arg not beyond compare. I mean Team Coherence01:41
bleah01:41
brain missfire01:41
MetaCosm ... BC is a compare tool and awesome01:41
don't you badmouth it!01:41
frogonwheels MetaCosm: BC is a compare tool and awesome01:42
cmn if you want a monotonically increasing number, define a tag and convince git-describe to use that one as the base01:42
and take the middle par01:42
frogonwheels MetaCosm: sorry I totally agree, I use it, don't know why my brain misfired on that one.01:42
cmn that way you get pretty much the same as svn01:42
but fixing the system might be easier than jumping though all these hoops01:43
← osmosis left01:43
frogonwheels MetaCosm: can you convince the team to #include (or whatever) a version file that is always generated from git describe?01:43
← Joseph_ left01:43
MetaCosm this is so incredibly dumb... it is freaking javascript01:44
← dsantiago left01:44
cmn what do you actually need then?01:44
MetaCosm I am just going to kludge it till I can fix it01:44
cmn something that sorts correctly or really a number?01:44
MetaCosm time since epoch - X in that slot as a post commit hook01:45
→ murr4y joined01:45
MetaCosm as the same amount of value as a goddamn svn revision01:45
frogonwheels MetaCosm: there is a man gitattributes method that allows you to expand macros :|01:45
gitinfo MetaCosm: the gitattributes manpage is available at http://git-scm.com/docs/gitattributes.html01:45
→ neurodrone joined01:45
← neurodrone left01:45
→ neurodrone joined01:45
frogonwheels I'm pretty sure it's gitattributes anyway.01:45
← diegoviola left01:46
MetaCosm ahhh01:46
that could work in the short-term01:46
till we get to something that makes sense01:46
← AaronMT left01:46
frogonwheels I think it's use has the caveat *this is dumb, but if you gotta do it,01:46
→ weiyang joined01:46
← mwrosen left01:46
frogonwheels MetaCosm: and anwyay, the sha1 is calculated post crlf -> lf conversion anyway.. so expanding some inline text.. well.. not so nasty01:47
← bitkiller left01:48
adityamenon even if I say git merge --ff01:49
it says merge made by recursive01:49
can I force it to just forward in some places?01:49
cmn only if it can fast-forward, it sounds like it's not possible in that case01:49
adityamenon well, I don't know what I did... I think I made a global configuration or some project specific config01:50
← alvaro_o left01:50
adityamenon to *always* merge by recursive01:50
but actually, develop is just a straight forward branching of master01:51
made a few changes like adding a couple of lines01:51
then went back to master and demanded a fast-forward01:51
still does a recursive01:51
cmn recursive isn't the oppsite of fast-forward01:52
recursive is a side-effect of doing a real merge01:52
!lol to check if it could have actually been a ff01:52
gitinfo git config --global alias.lol "log --oneline --graph --decorate"01:52
← jargon- left01:55
← Zespre left02:00
→ Zespre joined02:00
adityamenon cmn, what does the command do?02:01
where can I find the global alias.lol?02:01
← Syzygy_ left02:01
adityamenon .lol? hard time believing that's a real extension, lol02:02
cmn the command there sets up an alias02:02
→ zdk joined02:02
adityamenon okay..02:02
cmn and what the alias executes gives you a graph of the current branch02:02
.lol is a real extension if you define it02:02
adityamenon well, of course :)02:03
cmn and it's not even an extension, it's the local part of the key02:03
adityamenon someone did define it though didn't they02:03
→ grantbdev joined02:03
cmn huh?02:03
adityamenon let's forget it02:04
I don't mean to criticize it :)02:04
← musee left02:05
→ jackh joined02:05
adityamenon C:\Program Files (x86)\Git\etc\gitconfig02:05
is that the global config file?02:05
← dr_diamond left02:06
cmn no, that's the system config02:06
← Lulzon left02:07
→ edude03 joined02:08
adityamenon okay, I also found C:\Users\username\.gitconfig02:08
← diogogmt left02:08
adityamenon apparently, I've not anything to mess with merge options automatically02:08
→ leeight_ joined02:08
adityamenon there's only a 'rebase autosquash true'02:09
that doesn't relate to my current problem...02:09
→ diogogmt joined02:09
← mdhopkins left02:10
cmn and what happened to checking whether it could really be a fast-forward?02:10
adityamenon sorry, I had to quench my curiosity, I just looked at the config file for the current project in .git, it also does not have any weird rules02:11
what is the command to make the alias local?02:12
git config --local alias.lol "log --oneline --graph --decorate"02:12
correct?02:12
cmn don't tell git to make it global02:12
local is the default02:12
→ igotnolegs joined02:12
adityamenon very good, I just ran git config alias.lol "log --oneline --graph --decorate"02:12
← aruiz left02:12
adityamenon now... should I try the command log?02:12
cmn the command log? you can use git lol, since you've configured the alias02:13
or just use the commadn that's in it02:13
adityamenon oh, I'm so dumb, thanks... just checked the local file again and see a new alias called 'lol'02:14
infogulch|afk → infogulch02:14
adityamenon now I see what you were saying before, too02:14
← diogogmt left02:14
cmn yes, that's what the command says to do02:14
adityamenon I was thinking alias.lol would create a new file02:14
called alias.lol02:14
facepalm @ me02:14
okay, just ran git lol...02:15
while it's a sexy graph in the command line, the equivalent of gitk02:15
cmn yes02:15
adityamenon how can I find out if it could have been an FF?02:15
→ juvenal joined02:15
cmn see if there are any commits on the destination branch after the other branch starts off02:16
→ diogogmt joined02:16
adityamenon no, there are none02:16
→ rking joined02:16
adityamenon develop: "updated readme"02:16
then it's directly merging into master02:17
there's no commits in master before the merge02:17
there's simply a merge commit after that02:17
master: Merge branch 'develop'02:17
is the command git merge --ff wrong?02:17
← cakehero left02:17
→ eephillip joined02:18
adityamenon I think I should try rebase02:18
jrdnn adityamenon If it says "Merge branch 'branch' into currentBranch" it wasn't a fast forward02:18
→ ISF joined02:19
adityamenon so if I want to force a fast-forward I should just rebase?02:19
in the past, I didn't have to02:19
if it could do a fast-forward it would02:19
I don't see a reason why it can't now02:19
I have no problems using rebase, of course02:19
← squillow left02:20
adityamenon none of the history is published yet, anyway02:20
I'm simply curious why it's like this, now - when it was different before02:20
← Nibato left02:20
cmn git still defaults to a fast-forward02:20
→ hwrd|work joined02:20
← Yuuhi``` left02:21
jrdnn As long as only one branch has different commits it should fast forward (use --track when you make the branch and look in git status)02:21
cmn and --ff should take care of any other defaults02:21
← orafu left02:21
← thearthur left02:21
cmn care to pastebin the output?02:21
→ orafu joined02:21
adityamenon sure... just a second02:22
→ wrapids joined02:24
→ dsantiago joined02:24
← apok left02:25
← Mowsh left02:25
→ aruiz joined02:25
adityamenon http://pastebin.com/hT437FHk and the gitk looks like this: http://imgur.com/eDU2802:25
→ jjbohn joined02:26
adityamenon the 'updated readme' commit could have been an FF02:26
jrdnn --ff-only02:27
cmn I don't see how that would be a fast-forward02:27
→ Nibato joined02:27
→ SuperMiguel joined02:27
← ebneter left02:27
cmn the merge just before the one you created now makes it impossible02:27
← benjo2 left02:27
SuperMiguel So i have my rep folder in my computer, and every time i edit files, i have to do git add <file_name> because they are getting listed in git commit as not staged for commit..02:28
cmn that commit comes later than the merge base, and it's in the receiving branch02:28
SuperMiguel any idea why is not doing it automatically?02:28
adityamenon so if I do a merge --no-ff once along the lifetime, I am forever damned to use the same? :)02:28
cmn now, yo could have done a ff the other way around02:28
← raykyri left02:28
→ flippo joined02:28
cmn you can merge into the topic branch02:28
adityamenon SuperMiguel, "git add ."02:28
cmn then you can start doing fast-forwards02:28
← Araxia_ left02:29
adityamenon SuperMiguel, then "git commit -a"02:29
cmn SuperMiguel: you can use add -u or commit -a to stage the modified files02:29
jrdnn aditaymamenon You have to use non-ff until they get a common point in history, then if the one you are in when you run merge has no commits of the git merge --ff-only otherBranch doesn't have you can ff02:29
SuperMiguel adityamenon: cmn any of you use vim?02:29
jrdnn <-02:29
cmn what's that got to do with anything?02:29
← hwrd|work left02:30
jrdnn git is for code vim is for code :P both help programming a lot02:30
SuperMiguel cmn: jrdnn was just going to ask if you guys use any plugin for vim02:30
→ stardiviner joined02:30
← grantbdev left02:30
← eephillip left02:31
jrdnn SuperMiguel: Just .vimrc http://pastebin.com/52Cit2Aa02:31
→ adityamenon90 joined02:32
SuperMiguel jrdnn: not sure where git comes into play there?02:32
stardiviner I made a mistake, I committed a very secure file into repo, how to remove that from repo ? I use command: git rm -f --cache FILE, but after this, I still can see the commit history in repo about this file, how to remove it totally in repo instead of hard disk ?02:32
adityamenon90 jrdnn, thanks for --ff-only, I thought that was just --ff02:33
cmn, so I need to first merge master back into develop, then I can start doing fast forwards if I want to?02:33
→ maletor joined02:34
adityamenon90 I actually made a mistake, and committed changes to master :O just now02:34
then, I went to develop, asked to do git merge master --ff-only02:34
jrdnn adityamenon90 If you are in develop and run merge master, and you make some commits to develop, and not to master, you can checkout master merge develop and it will ff02:34
→ macabre joined02:34
cmn the way things are in that gitk view, you can merge into develop, and if you only do changes in develop, then you can do a ff02:34
adityamenon90 it worked, now I might be able to do FFs, not sure need to test..02:34
← adityamenon left02:35
adityamenon90 cmn, let's say I did not make the mistake I just wrote about02:35
If I were to do "git checkout develop", "git merge master" - wouldn't it say 'everything up to date' and quit?02:36
jrdnn aditayamenon90 If the branch you are currently in has no commits the branch you want to merge into it didn't have when the branch you want to merge into it was made, you can merge the other branch into it and it will be fast forward (just update pointers)02:36
adityamenon90 → adityamenon02:36
← aruiz left02:36
adityamenon jrdnn, that's not happening though02:37
master had no commits that develop did not have02:37
oh yeah02:37
there's the empty merge commit...02:37
those things can cause a problem?02:37
SethRobertson a commit is a commit02:37
← Kulrak left02:38
← BlindOracle left02:38
← md_5|away left02:38
→ hwrd|work joined02:38
SuperMiguel jrdnn: how does ur .vimrc helps with git?02:39
cmn it woudn't say it's up to date, because it's not02:39
master contains the merge02:39
adityamenon cmn, jrdnn I get it now02:40
thanks :)02:40
jrdnn SuperMiguel: Sorry, it doesn't. The empty merge commit is a seperate commit, made whenever a non-ff merge happens, and it counts as a seperate commit02:40
adityamenon You're welcome =)02:41
adityamenon jrdnn, I think you wanted to say that to me :D02:41
the one above addressed to SuperMiguel02:41
→ Davey joined02:41
jrdnn adityamenon: "Sorry, it doesn't" was about vimrc, the rest I meant to you, sorry02:41
adityamenon :)02:41
anyone else think those merge commits are super pesky?02:41
to look at, at least02:42
well, I guess it's because of my workflow02:42
← iwoj left02:42
jrdnn aditaymenon: They are good if you want to keep track of which branch a bunch of commits are from02:42
cmn it might just be that you have too many02:42
adityamenon master is not the 'holy branch' as recommended in gitflow02:42
I just keep pushing everything that works02:42
→ BarryCrayon joined02:43
adityamenon every *small* thing02:43
I should keep a new branch called 'testing' for the clients' purpose02:43
won't matter how ugly that looks02:43
→ Kulrak joined02:43
→ BlindOracle joined02:43
→ md_5|away joined02:43
adityamenon cmn, jrdnn you're both right02:44
← fruers left02:45
→ ShaunR joined02:46
ShaunR when trying to clone a repo i'm getting fatal: Unable to find remote helper for 'http'02:47
it's only on this machine though.02:47
no firewall02:47
← BarryCrayon left02:47
ShaunR any ideas what may be happening? it's a old centos 4.9 server and i had to build git from source02:47
frogonwheels ShaunR: don't you need something like curl to do the transport?02:48
ShaunR ya, was just reading somthing on that02:48
actually installing it now, and going to rebuild02:49
gitinfo set mode: +v02:49
← igotnolegs left02:50
← williamcotton left02:50
ShaunR frogonwheels: yep, curl was the issue, thanks!02:52
frogonwheels ShaunR: do one thing and to it well ;)02:52
*do02:52
→ FFForever joined02:54
← jjbohn left02:55
← edude03 left02:55
← marhaban left02:55
← diogogmt left02:55
→ diogogmt joined02:55
→ cooloney joined02:56
jrdnn frogonwheels: That's the UNIX philosophy +102:56
FFForever How come when compiling with make --prefix=/usr/local/ all it still compiles/installs in the same directory02:56
← hwrd|work left02:57
→ cccaldas joined02:57
→ edude03 joined02:57
→ jjbohn joined02:58
cmn because make isn't ./configure02:58
FFForever there is no ./configure02:58
cmn yes, there is, but that's not the point02:58
that's not how you pass options to make02:59
make PREFIX=02:59
FFForever ah, and I only have a configure.ac02:59
→ diegoviola joined03:00
cmn fine, but you should still pass the options to make properly03:00
FFForever fair enough.03:01
gitinfo set mode: +v03:01
cmn it will call autotools for you03:01
dpwright stardiviner, not sure if somebody already answered you (I did a quick scan down the log and couldn't see anything), but man git-filter-branch and look at the examples03:01
gitinfo stardiviner: the git-filter-branch manpage is available at http://git-scm.com/docs/git-filter-branch.html03:01
→ iwoj joined03:01
← jjbohn left03:02
← diogogmt left03:03
dpwright be warned that this entails modifying history, and if you've published it this may cause problems for people who've since pulled. Google for more information on the perils of modifying public history. But that should solve your problem.03:03
← ShaunR left03:04
frogonwheels stardiviner: !rewrite03:04
gitinfo stardiviner: [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)03:04
← phantomtiger left03:05
← iwoj left03:05
→ diogogmt joined03:05
← netj left03:06
gerard0 → gerardo03:06
← diogogmt left03:10
→ diogogmt joined03:10
← mcluskydodallas left03:12
← Goplat left03:13
← kukks left03:15
← diogogmt left03:15
← Zespre left03:16
→ diogogmt joined03:16
→ Zespre joined03:17
→ posciak joined03:19
← matt_taws left03:20
→ matt_taws joined03:20
→ Goplat joined03:20
→ avpx joined03:23
→ williamcotton joined03:25
stardiviner dpwright: frogonwheels thanks, I will check out manual, about filter-branch, and rewrite,03:28
← jstout24 left03:28
→ ipoval joined03:29
→ iwoj joined03:29
→ kcm1700_ joined03:29
← Per4u3e left03:29
← kcm1700 left03:30
← elnn left03:30
→ DofLeFou joined03:30
→ elnn joined03:30
← alansaul left03:31
→ netj joined03:32
→ rmccue joined03:34
→ iori joined03:35
→ hwrd|work joined03:36
adityamenon in my .gitignore file, I have03:36
application/config/database.php03:36
.htaccess03:36
rmccue Hi, I recently moved a repository (as in, `mv olddir newdir`) and my submodules are now broken, as they're pointing to the old path. How can I fix this?03:36
adityamenon in two lines03:36
but they are not being ignored03:36
they are being pulled into clones03:36
in a different repository, they were ignored03:37
I had to manually create them by renaming03:37
.htaccess.example03:37
← bgerber left03:37
adityamenon and application/config/database.php.example03:37
as it should be...03:37
but this time that's not working03:37
what may I be doing wrong?03:37
rmccue adityamenon: They've probably already been added to the repository with `git add -f`03:37
cmn if they're "pulled" into repos, that means they're being tracked03:37
you can't ignore tracked files03:37
avpx adityamenon: !enter03:38
gitinfo adityamenon: The enter key is not a punctuation mark. Please use fewer lines, with complete sentences on them.03:38
adityamenon avpx, sorry about that03:38
avpx No problem, just thought you should know.03:38
→ atomoul joined03:38
← atomoul left03:38
adityamenon I'm googling 'how to untrack files' now03:38
rmccue adityamenon: `git rm <filename>`03:39
adityamenon if I untrack these files, then later use "git add ."03:39
avpx The --cached option may be useful03:39
cmn !untrack03:39
gitinfo to stop git from tracking a file, without deleting the file, use "git rm --cached <file>". Please note that if you push/pull this untracking, the file will be actually be deleted on disk by the remote repo.03:39
rmccue Anyone know how to unhardcode my submodule paths?03:39
adityamenon will .gitignore work then?03:39
cmn rmccue: newer git versions use relative paths03:40
adityamenon in one sentence: if I untrack these files, then later use "git add .", will .gitignore work then?03:40
it's a simple matter to test of course :) sorry to be lazy03:40
cmn adityamenon: yes, because they won't be tracked then, so git will ignore them03:40
rmccue cmn: Ah, so I need to upgrade git? Is there any way to do it without that?03:40
EugeneKay !tias03:40
gitinfo Try it and see™. You learn much more by experimentation than by asking without having even tried. If in doubt, make backups before you experiment (see !backup). http://sitaramc.github.com/1-basic-usage/tias.html may help with git-specific TIAS.03:40
adityamenon cmn, rmccue thank you :)03:40
cmn rmccue: it won't update them; go into the submodule and update the path in .git03:41
rmccue cmn: I accidentally rm -r'd .git/modules/module-name so I can't do that.03:41
cmn then you'll have to re-init the submodule03:42
which may or may not work in this case03:42
rmccue It doesn't, unfortunately03:42
cmn hardly surprising03:43
rmccue So I'll have to upgrade then, I guess?03:43
cmn the easiest way might just be to re-clone03:44
→ atomoul joined03:44
rmccue Good idea03:45
← madewokherd left03:47
← codeM0nK3Y left03:47
→ akiress joined03:50
← tyatpi left03:51
rmccue cmn: Thanks for the help.03:51
← hwrd|work left03:53
→ ashphy joined03:54
→ bgerber joined03:55
← rmccue left03:55
← wrapids left03:57
→ netzmonster joined03:57
← netzmonster left03:57
→ Dave^|| joined03:58
← ashphy left03:59
adityamenon !backup03:59
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/#backups03:59
← pielgrzym left03:59
→ hwrd|work joined04:00
→ mcluskydodallas joined04:00
← simesy_work left04:01
→ caseymcg joined04:01
→ pielgrzym joined04:01
← Dave^| left04:01
← mcluskydodallas left04:02
← booginga left04:05
← Zespre left04:07
→ Zespre joined04:08
← tigrang left04:08
→ ology joined04:11
← cilly left04:12
→ cilly joined04:13
→ Jackneill joined04:15
← ology left04:17
← brentratliff left04:17
→ jargon- joined04:17
→ fuzzybunny69y joined04:18
← SuperMiguel left04:21
→ jhardinee joined04:22
← Arrowmaster left04:25
← cccaldas left04:25
← alta left04:25
→ BiggFREE joined04:26
→ Arrowmaster joined04:26
← OOPMan left04:26
→ beautifulmind joined04:27
→ alta joined04:28
← _huoxito left04:29
→ simesy_work joined04:29
→ ashphy joined04:30
← hwrd|work left04:31
← Zenith77|2 left04:32
→ naught101 joined04:34
→ dvide joined04:36
infogulch → infogulch|afk04:38
→ glosoli joined04:40
glosoli how can I do git pull with different user ?04:40
← Goplat left04:41
avpx man git-remote04:41
gitinfo the git-remote manpage is available at http://git-scm.com/docs/git-remote.html04:41
avpx glosoli: Assuming SSH, you just change the remote URL using git-remote per that man page04:42
→ gavin_huang joined04:42
glosoli well I don't want to change URL04:42
is there any way ?04:42
→ PRETTY_FUNCTION joined04:42
glosoli I just want to make pull once from different account04:42
→ BarryCrayon joined04:43
→ eedeep joined04:43
← BiggFREE left04:43
→ jceb joined04:43
eedeep hi there. the workflow I'm trying to achieve is to squash all my commits into a single commit on my feature branch before merging it in04:43
→ igotnolegs joined04:44
eedeep I'm doing this with git rebase -i HEAD~N on my feature branch, where N is the number of commits I want to squash down04:44
avpx glosoli: Can't you just git pull <repository> and be done?04:44
glosoli nah I can't :)04:44
avpx What happens?04:45
glosoli it asks for current user password04:45
afaik04:45
avpx What protocol are you using? ssh?04:45
eedeep I'm trying to understand the difference between doing git rebase -i HEAD~N on my feature branch and then doing git merge myfeature to merge the feature branch into say master and just doing git checkout master && git merge --squash myfeature04:45
→ Syzygy_ joined04:45
glosoli avpx: yes04:45
← diegoviola left04:46
eedeep they seem to be trying to achieve the same thing, more or less04:46
avpx glosoli: So you're using the URL git://user@remote.com/repo04:46
You have the user@ part, correct?04:46
← tonsofpcs left04:47
← BarryCrayon left04:47
→ tonsofpcs joined04:48
glosoli avpx: ah nevermind, my problem is I don't understand if some git catalog is added to the computer it knwos about other user who has access ?04:48
avpx glosoli: Could you rephrase that? I don't understand.04:49
→ d4rklit3 joined04:49
d4rklit3 hi04:50
gitinfo d4rklit3: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.04:50
d4rklit3 im kind of in a bind.04:50
i need to roll back a file one reviision back04:50
not sure how to find the revision hash04:50
avpx eedeep: I'm trying to answer your question thoughtfully (or at all)04:50
→ tyll joined04:50
← iwoj left04:50
eedeep avpx: thanks. does the question make sense to you?04:50
avpx eedeep: Really, it seems to me that the difference is that rebase is just playing the diffs in the order you've selected and creating new commits, while merge is actually performing a 3-way merge like usual, and then just not making a commit04:51
eedeep: That is, merge --squash is actually doing a merge but it's not creating a merge commit. But rebase -i with the "squash" command is basically just replaying patches in order.04:52
eedeep avpx: hmm ok right. I guess what I'm interested in really is just the end result...what I want is just to see a single commit on master which contains my new feature04:53
← jceb left04:54
eedeep avpx: so with given that --squash doesn't create a merge commit, perhaps that's not waht I'm after04:54
avpx d4rklit3: If you just want to pull that file out of a previous commit, try the git checkout command. For the previous commit, the command looks like: git checkout HEAD^ <file>04:54
← edude03 left04:54
avpx eedeep: That seems like a reasonable analysis. I'd say that rebase -i is what you are after.04:54
→ aruiz joined04:54
eedeep avpx: my problem with doing git rebase -i HEAD~N on my feature branches is that for some reason I seem to have to resolve tonnes of merge commits during htat process04:55
I have to resolve them, do git rebase --continue, get more merge conflicts....repeat04:55
avpx Now I'm a tad confused: where are the merge commits comming from?04:56
coming*04:56
eedeep avpx: you mean "where are the merge conflicts coming from?04:56
"04:56
?04:56
avpx You said commits earlier, that's what I thought you meant04:57
eedeep oh! sorry, I meant "merge conflicts"04:57
avpx I see. Well, I don't know what to tell you. squash shouldn't give you merge conflicts because essentially it's just dropping certain commits04:58
eedeep I'm rebasing my feature off of develop before initiating the git rebaes -i HEAD~1 ..... I dont know if that's the source of my conflicts perhaps04:58
→ tigrang joined04:58
eedeep avpx: when you say "squash" you mean the kind of squash you do when you do git rebase -i HEAD~N yeah?04:58
avpx yeah04:58
I can't see any reason why squash / fixup would produce conflicts04:59
Unless you were changing the order of the commits, or something04:59
← Zenmaker__ left05:00
← redii left05:00
avpx d4rklit3: Have you resolved the problem?05:00
Remember that as long as you committed something, it's probably still around.05:00
→ redii joined05:01
→ Zenmaker__ joined05:01
eedeep avpx: what I usually do during a git rebase -i HEAD~N is leave the first commit as pick and then change all those below it to squash ....is that the normal procedure?05:01
← mikepack left05:01
avpx Yeah05:01
Hmm, unfortunately I really don't know why you would get conflicts during that.05:01
eedeep yeah though so. oh well I'll try to get to the bottom of what's going on and report back if I get a magical insight.05:02
avpx Please do05:02
eedeep thanks for your help05:02
avpx Of course05:02
← ziggles left05:04
→ marc_kdab_ joined05:04
← fisted left05:05
← dhruvasagar left05:06
→ hwrd|work joined05:08
→ harshpb joined05:09
→ fisted joined05:09
← ISF left05:09
← harshpb left05:11
→ ISF joined05:12
→ harshpb joined05:12
← tehbaut left05:13
← Jon30 left05:14
← delinquentme left05:15
→ tehbaut joined05:15
→ boobah joined05:16
← diogogmt left05:17
constant is there a way I can say 'git edit commit log of abc13456' without rebase -i ?05:18
← Zespre left05:18
→ arvind_k joined05:18
→ Zespre joined05:18
← neurodrone left05:18
avpx constant: No, you have to rebase05:18
constant avpx: I don't mind rebasing or whatnot05:19
but I want to avoid the editor05:19
avpx Oh, I see...05:19
← harshpb left05:19
→ maleknet joined05:19
← aruiz left05:20
→ harshpb joined05:20
avpx constant: I suppose you could use git rebase -i with careful use of EDITOR05:20
constant haha true05:20
→ rmccue joined05:21
arvind_k → arvind_khadri05:21
avpx Hmm, you could also manually apply the changes in a separate branch05:21
← nutron left05:21
avpx Let's see now. Let's say you have A -- B -- C05:21
← arvind_khadri left05:21
→ arvind_khadri joined05:21
→ JeremyWei joined05:21
constant avpx: hm?05:21
avpx constant: So in this example master points at C. Create a new branch at B, check it out, and use git commit --amend to change the log message05:23
Now you'll have something like:05:23
A -- B -- C <-- master05:23
\05:23
B'05:23
constant yeah05:23
avpx Now you can rebase C onto B'05:23
constant I could do that05:23
but then it doesn't achieve my ultimate goal05:24
avpx What is the ultimate goal?05:24
constant of being able to edit past logs in 1 command, though it does avoid rebase -i :)_05:24
avpx Ah yeah. Well, you can cheat and use && to glue the commands together :p05:24
constant avpx: there is meta-data in commit logs which change after N commits, its annoying to go back every time with rebase -i and fix them05:24
avpx I still think that a good choice of EDITOR would allow this05:25
constant yeah05:26
I could play with 'sed' or something of the like05:26
avpx Just make a simple EDITOR that replaces the "pick" on the first line with "edit"05:26
constant :-)05:26
avpx Or something similar05:26
← rmccue left05:26
→ iwoj joined05:26
← maleknet left05:27
constant avpx: EDITOR="sed 1s/pick/reword/"05:27
avpx Have you tried that? Does it work?05:28
constant trying it now05:28
avpx I can't remember if sed needs an argument to edit in-place05:28
→ maleknet joined05:28
avpx Ah, yeah, I think you'll need sed -i05:29
constant avpx: problem is05:29
← thornofmight left05:29
constant it also affects the "edit the commit message" EDITOR :)05:29
→ arti_t joined05:29
avpx Ooh.05:30
→ guru joined05:30
quazimodo if i have *.conf in .gitignore, will that ignore *.conf in all subdirs too?05:30
→ thornofmight joined05:30
avpx constant: I don't know what to tell you, then :p05:31
frogonwheels quazimodo: yep05:31
← harshpb left05:31
constant avpx: though with some playing I could probably figure out what type of file I'm seeing and either run sed or vim05:31
but that would be ....05:31
silly05:31
:)05:31
→ boombatower joined05:31
avpx It's true.05:32
→ rmccue joined05:32
→ harshpb joined05:32
→ aruiz joined05:33
← leeight_ left05:33
frogonwheels constant: It's going to come up with an editor to change the commit message anyway.. or were you going to use -m ?05:33
← bgerber left05:34
→ ThomasLocke joined05:34
← stardiviner left05:34
constant frogonwheels: I'm fine with an editor to change the log message05:34
but I want to avoid the middle step, when I already know which hash I need to reword05:34
constant feeds frogonwheels flies05:34
dpwright constant: wouldn't defining an alias which cheats and uses && like avpx said be good enough?05:34
constant dpwright: likely05:35
I'll try at some point05:35
avpx I'm not sure if you could do it with alias, but a tiny shell script maybe05:35
constant just hoping there was something like git ammend <hash>05:35
frogonwheels constant: git checkout abc13456 && git commit --amend && git rebase --onto HEAD abc123456 branchname05:35
constant: something like that?05:35
avpx frogonwheels: That's what I was suggesting earlier05:35
← maleknet left05:35
constant avpx: frogonwheels yeah, I'll probably end up with that05:36
→ abstrusenick joined05:36
→ maleknet joined05:37
← macabre left05:40
← ashphy left05:41
→ ashphy joined05:41
→ ashphy_ joined05:43
← ashphy left05:43
← d3vic3 left05:45
→ leeight_ joined05:46
→ d3vic3 joined05:46
← maleknet left05:46
→ rbarraud_ joined05:47
→ maleknet joined05:47
→ twb joined05:51
→ j416 joined05:51
twb Can I /msg gitinfo with questions? I tried and it doesn't seem to work.05:53
cmn questions?05:53
→ marthinal joined05:53
cmn it responds to the same commands as in the channel05:53
→ airborn joined05:53
twb I must have picked a word he didn't have an entry for, then05:53
$coworker wants something prettier than log --graph05:54
I was hoping to pester gitinfo directly instead of the channel :-)05:54
cmn prettier as in what? there's gitk, --graph has --decorate05:55
there also a few GUIs, but the graph they show is pretty much the same05:55
twb Well as it happens I solved his immediate issue by telling him to use --color, but he wanted to know about GUI type alternatives05:55
← harish left05:55
→ Blaker0111 joined05:55
cmn setting color.ui to auto would make everything prettier05:56
git-scm.com has a list of the prettier ones05:57
though not a lot for Linux et al05:57
twb http://jk.gs/git/bot/faq.php looks pretty barren compared to what I thought, but oh well.05:57
cmn: thanks05:57
← maleknet left05:57
→ mishok13 joined05:57
Blaker0111 Hi. I have a git repository that I would like to rename. The repository is located in /path/repo-name.git my server. I have a working copy on my local pc in /path/repo-name . Do I just rename repo-name.git to new-repo-name.git and my local directory to new-repo-name/ ?05:57
← ashphy_ left05:57
twb Ah, they're in http://jk.gs/git/bot/trigger.php05:57
cmn well, that's the faq for the bot/channel, not for all of git05:58
twb yeah understood05:58
cmn ah, yes, the faq is a different thing from the ! triggers05:58
→ ashphy joined05:58
twb The "list of things the bot knows"05:58
→ maleknet joined05:58
← iwoj left05:59
j416 twb: look at git log --format05:59
→ jceb joined06:00
← glennpratt left06:00
avpx Blaker0111: Not exactly, no.06:00
→ herdingcat joined06:00
avpx Blaker0111: It's fine to rename the upstream repo (that is, the one on your server) but the downstream repo (the one on your PC) needs to have its remotes changed. This is done using the git remote command. See: man git-remote06:01
gitinfo Blaker0111: the git-remote manpage is available at http://git-scm.com/docs/git-remote.html06:01
j416 twb: this is what I mostly use, I have an alias for it: http://dl.dropbox.com/u/3207594/24h/gitlog_format.png06:01
→ jmd joined06:01
← vitor-br left06:01
Blaker0111 avpx: gotcha, ill talk a look -- thanks for the help :D06:02
← ashphy left06:02
avpx Good luck06:02
← kcj left06:02
← maleknet left06:04
← Fullmoon left06:05
← aruiz left06:06
→ toabctl joined06:07
← rmccue left06:07
→ afuentes joined06:07
← beautifulmind left06:07
→ musee joined06:08
→ sdeboey joined06:09
← glosoli left06:11
← Silex left06:12
← j416 left06:12
Blaker0111 avpx: could i also just rename the /path/repo-name.git folder to /path/new-repo-name.git on the remote server, delete my local working copy, and re-checkout the newly named repo?06:13
→ Silex joined06:14
avpx Blaker0111: The correct term would be "re-clone" because checkout means something different in git than CVS/SVN. But yes. Still, changing the remotes is easier.06:14
See also: !tias06:14
gitinfo Try it and see™. You learn much more by experimentation than by asking without having even tried. If in doubt, make backups before you experiment (see !backup). http://sitaramc.github.com/1-basic-usage/tias.html may help with git-specific TIAS.06:14
_ikke_ Blaker0111: with git remote set-url you can change the urls for remotes06:14
← eletuchy left06:14
→ cbreak-work joined06:15
Blaker0111 ah ok i was trying git remote rename, but git remote set-url was the ticket06:16
thanks both of you :)06:16
Silex is it just me or !tias says "Try it and seeM- You learn..." ?06:17
gitinfo Try it and see™. You learn much more by experimentation than by asking without having even tried. If in doubt, make backups before you experiment (see !backup). http://sitaramc.github.com/1-basic-usage/tias.html may help with git-specific TIAS.06:17
← xtagon left06:17
dpwright constant: Knocked something together based on frogonwheels' suggestion: https://gist.github.com/299370106:17
Silex someone should correct the text06:18
_ikke_ Silex: What should be changed?06:18
Blaker0111 SIlex: to me its showing up as a TM symbol after see06:18
dpwright tested it and it seems to work, but you need to be a bit careful with it06:18
Silex the M- part06:18
oh, maybe my client doens't render that symbol properly06:18
_ikke_ Yup06:18
Here it works06:18
Silex a'ight06:18
→ dc5ala joined06:20
→ hoschi joined06:21
← weiyang left06:23
← williamcotton left06:23
→ bgerber joined06:24
← Rhonda left06:25
→ Rhonda joined06:25
→ soee joined06:25
→ rtjure joined06:26
← andi left06:27
← BlindOracle left06:27
→ _Qman joined06:27
→ senny joined06:27
→ rath joined06:28
← _Qman left06:29
← Blaker0111 left06:29
→ _Qman joined06:30
← xreal left06:31
← madsy left06:32
→ tsdh joined06:32
← kkiero left06:32
→ boombatower_ joined06:32
← boombatower left06:32
→ kcj joined06:33
→ adityamenon90 joined06:33
boombatower_ → boombatower06:34
← adityamenon left06:35
→ gour joined06:37
← toabctl left06:38
→ aristidesfl joined06:38
← adityamenon90 left06:40
← ferdna left06:41
← mishok13 left06:41
→ Xizor joined06:42
→ jonnu joined06:42
→ BarryCrayon joined06:43
→ eletuchy joined06:44
← jceb left06:45
→ ph^ joined06:47
→ superdmp joined06:47
← Davey left06:48
→ weiyang joined06:48
→ ottoshmidt joined06:48
← BarryCrayon left06:48
← harshpb left06:48
→ harshpb joined06:48
← airborn left06:48
→ beaky joined06:49
→ dhruvasagar joined06:50
← francisfish left06:52
← keller left06:52
→ adityamenon joined06:53
→ ashphy joined06:53
← bgerber left06:53
← Zespre left06:56
md_5|away → md_506:56
← ashphy left06:57
→ Zespre joined06:57
→ ashphy joined06:57
← adityamenon left06:57
→ shruggar joined06:59
→ yizumi joined07:00
← Syzygy_ left07:00
→ thiago joined07:01
← Xizor left07:05
→ iwoj joined07:06
← avpx left07:06
← ashphy left07:06
→ harish joined07:07
→ pantsman joined07:07
← pantsman left07:07
→ pantsman joined07:07
→ ashphy joined07:07
→ madsy joined07:08
→ barmaglot joined07:10
← barmaglot left07:10
→ QaDeS joined07:10
→ barmaglot joined07:11
← ashphy left07:12
← aspotashev left07:12
← denysonique left07:12
→ aspotashev joined07:12
→ denysonique joined07:13
→ mishok13 joined07:15
← beaky left07:16
→ nisseni joined07:16
→ legumbre_ joined07:17
→ Huevo joined07:17
← hoschi left07:18
→ e-ndy- joined07:19
→ crash\ joined07:19
← rrs left07:19
← Orbitrix left07:19
← surfdue left07:19
← e-ndy left07:19
← malesca left07:19
← crash\\ left07:19
← CEnnis91 left07:19
→ surfsue joined07:19
surfsue → surfdue07:19
← dougsland left07:19
← jab416171|Cloud left07:19
→ rrs joined07:19
→ dougsland joined07:19
← palm_vue left07:20
← Lgb left07:20
→ CEnnis91 joined07:20
→ jab416171|Cloud joined07:20
→ malesca joined07:21
→ antgel joined07:21
← redii left07:21
← elnn left07:21
← naught101 left07:21
→ Lgb joined07:21
→ palm_vue joined07:21
← AlbireoX`Laptop left07:22
← andrewhl left07:22
→ elnn joined07:22
antgel hi all. i have a question about git mv. does it apply just to the branch i'm working on? i have a branch 'master' and a branch 'foo'. if i git mv some files on foo, then git checkout master (without committing my git mv changes), git status reports the moves. yet i don't want them to apply to master. what am i doing wrong?07:23
→ oo_miguel joined07:23
_ikke_ antgel: You first have to commit them07:23
← legumbre left07:23
← oo_miguel left07:23
→ oo_miguel joined07:24
_ikke_ antgel: Anything that is not committed does not belong to any branch07:24
→ tigrang_ joined07:24
oo_miguel Does git also remeber file permissions?07:24
_ikke_ oo_miguel: Only executable07:24
oo_miguel ok thank you07:24
← tigrang left07:24
→ psoo joined07:25
← boombatower left07:28
← ColKurtz left07:28
→ RickXy joined07:29
← agile left07:30
← rtjure left07:31
→ Aaeriele_ joined07:31
→ pureluck_ joined07:31
→ styx__ joined07:31
→ Ferris_ joined07:31
→ valberg_ joined07:31
→ sriehl_ joined07:31
← orafu left07:31
→ kdub3_ joined07:31
→ wereHams1er joined07:31
→ piotr__ joined07:31
→ Spockz_ joined07:32
→ zaiste_ joined07:32
→ telmich_ joined07:32
← telmich_ left07:32
→ telmich_ joined07:32
→ nipnop joined07:32
→ libertyp1ime joined07:32
← notfunk left07:32
→ orafu joined07:32
→ flijten joined07:32
← yizumi left07:32
→ cytrinox_ joined07:32
→ notfunk joined07:32
← igotnolegs left07:33
← thursday left07:34
→ [W] joined07:34
→ Gizmokid2010 joined07:35
→ neersigh^ joined07:35
→ Zenopus_ joined07:35
← wereHams1er left07:36
← rrs left07:36
← shruggar left07:36
← harshpb left07:36
← dc5ala left07:36
← JeremyWei left07:36
← fisted left07:36
← iori left07:36
← libertyprime left07:36
← Tasser left07:36
← Khisanth left07:36
← felher left07:36
← project2501a left07:36
← snappy left07:36
← Zenopus left07:36
← pureluck left07:36
← piotr left07:36
← kdub3 left07:36
← sriehl left07:36
← wereHamster left07:36
← eighty4 left07:36
← Spockz left07:36
← zaiste left07:36
← SunilThaha left07:36
← valberg left07:36
← marienz left07:36
← styx_ left07:36
← lectrick left07:36
← neersight[d] left07:36
← Gizmokid2005 left07:36
← keltvek left07:36
← telmich left07:36
← werring left07:36
← cytrinox left07:36
← constant left07:36
← Ferris left07:36
← Aaeriele left07:36
cytrinox_ → cytrinox07:36
Gizmokid2010 → Gizmokid200507:36
telmich_ → telmich07:36
Zenopus_ → Zenopus07:36
→ wereHamster joined07:37
→ teknopaul joined07:37
→ keltvek joined07:38
→ eighty4_ joined07:38
→ fisted joined07:39
→ ccapndave joined07:41
→ agile joined07:41
e-ndy- → e-ndy07:42
← atomoul left07:42
← aspotashev left07:43
→ aspotashev joined07:44
→ _br_ joined07:46
→ Axsuul joined07:47
Axsuul is it good practice to keep around old branches (merge on commit)?07:47
→ jkremser joined07:47
→ simesy joined07:49
← dreiss_ left07:49
→ shruggar joined07:49
wereHamster Axsuul: do you need them?07:51
Axsuul Not really but it helps to see the history07:51
wereHamster does it?07:51
Axsuul I like seeing the history graph07:52
deebo i usually just have one called 'dev' i do everything in, then rebase it, merge to master and push changes to origin07:52
→ adityamenon joined07:52
wereHamster Axsuul: to see thi history graph you don't have to keep the refs around07:52
deebo unless i really need a branch for long term development07:52
→ maloik joined07:53
maloik I had to make some changes/deletions while I had started working on something else, normally I'd just add all the files I want in this commit and then commit, leaving out those files that belong to the next commit07:53
How can I add some deletions as well ?07:53
wereHamster git add <the deleted file>07:53
or run 'git status' and read07:54
maloik wereHamster: I thought so too but that doesnt seem to work07:54
→ adhawkins joined07:54
maloik when I do that, status doesnt return this deletion to the staged changes07:54
wereHamster maloik: can you pastebin the output from git status?07:55
→ whitman joined07:55
← shruggar left07:55
maloik wereHamster: http://pastie.org/415298807:55
→ bubbles10101 joined07:55
wereHamster maloik: hm. sorry. git rm --cached <the delteed file>07:56
maloik huh... i just tried that before and didnt get the impression it worked07:56
seems to do the trick now, thanks07:56
→ dr_diamond joined07:56
← tigrang_ left07:57
→ BarryCrayon joined08:00
← Zespre left08:01
→ Zespre joined08:01
→ LimeBlast joined08:02
← ipoval left08:02
← ill1cit left08:03
→ Khisanth joined08:04
→ constant joined08:04
→ project2501a joined08:04
→ SunilThaha joined08:04
→ TheHaven joined08:05
→ harshpb joined08:05
→ francisfish joined08:05
→ dc5ala joined08:05
← pantsman left08:05
→ marhaban joined08:06
→ ziro` joined08:07
← d4rklit3 left08:07
← tsdh left08:08
← adityamenon left08:08
← thiago left08:09
→ iori joined08:09
→ Tasser joined08:09
→ felher joined08:09
→ marienz joined08:09
antgel _ikke_: thank you08:10
_ikke_ antgel: your welcome08:10
gour morning08:11
→ uman joined08:11
→ yizumi joined08:11
jmd _ikke_: s/your/you're/08:11
← Vile left08:12
_ikke_ jmd: Ah yeah, autotypism08:13
→ blacktulip joined08:14
uman Hi, can someone give me a hint on the difference between "git push" and "git push -u" ? I didn't understand the man page's terse description08:14
_ikke_ uman: -u or --set-upstream sets some config options that lets git know which remote branch your branch is tracking08:15
← leeight_ left08:15
→ OOPMan joined08:15
gour i've started reading pro-git book in order to learn (some) git and wonder how to have treeless repo in one location and working directory in another (that's what bzr's lightweight checkout does)? i know there is option to do: git init --bare, but waht about the 2nd part?08:15
_ikke_ uman: (when git status says, your branch is behind .. and can be merged is a result of that)08:15
uman _ikke_: okay, so after I use it once to get the remote and local repositories aligned, I should only need to "git push" after that ?08:15
→ stodan joined08:15
_ikke_ uman: yes08:15
gour: You clone from the bare repository08:15
uman _ikke_: okay, thanks :)08:16
← weiyang left08:16
_ikke_ uman: also look at the push.default config option08:16
gour _ikke_: and i'll get just working repo? let me try it...08:17
← hwrd|work left08:17
_ikke_ gour: Yes08:18
→ marcogmonteiro joined08:18
→ jceb joined08:18
→ und3f joined08:18
gour _ikke_: cool...i (foolishly) thought git is too complex in comparison with other dvcs, but, although very powerful, it seems to be not so complicated...08:19
_ikke_ gour: !simple08:19
gitinfo gour: At it's heart, git is made up of many concepts that are individually simple. Getting the whole picture right is often tricky, and it's usually about breaking up the complex concept into its simple, individual parts and groking those.08:19
→ lectrick joined08:19
gour :-)08:19
uman should be "grokking"08:20
← eletuchy left08:20
_ikke_ The problem with git is that it's not afraid to inform the user about the inner workings, where that is shielded off from the user in most other SCM's08:20
uman: Yeah08:20
← marienz left08:21
_ikke_ .trigger_edit simple At it's heart, git is made up of many concepts that are individually simple. Getting the whole picture right is often tricky, and it's usually about breaking up the complex concept into its simple, individual parts and grokking those.08:21
gitinfo _ikke_: Okay.08:21
← jkremser left08:21
← yizumi left08:21
← iwoj left08:22
← sdeboey left08:22
← abstrusenick left08:22
→ tcurdt joined08:23
← tcurdt left08:23
→ yizumi_ joined08:23
← alanhollis left08:24
→ marienz joined08:24
→ shruggar joined08:25
← Pupeno_W left08:25
← aspotashev left08:25
← codeshah left08:25
→ Pupeno_W joined08:25
→ aspotashev joined08:25
← ocharles left08:26
← Mkaysi left08:26
← fuzzybunny69y left08:26
→ mklappstuhl joined08:27
→ puneetgoyal joined08:27
→ Lord_DeathMatch joined08:27
→ ocharles joined08:28
→ ill1cit joined08:28
→ Mkaysi_ joined08:29
← jim left08:29
→ kpreid joined08:30
luka12345 i would like to checkout folder (and all subfolders) to specifc commit08:30
anybody knows how to do that?08:30
but without touching all other folders08:30
jmd That doesn't make an awfull lot of sense.08:31
luka12345 well it does in my case08:31
jmd You can reset or checkout to a particular commit.08:31
You can list the changes of a particular commit.08:32
luka12345 and i want to do that but only on 1 folder08:32
jmd But what do you mean "folders to a specific commit"?#08:32
luka12345 doing it by hand is out of option08:33
→ tatsuya_o joined08:33
luka12345 i want to go back in history - checkout commit08:33
→ crippa joined08:33
jast you can't do a partial 'go back in history'08:33
you *can*, however, overwrite a set of files with an old version08:33
e.g.: git checkout <commit ID> <path>08:34
luka12345 let me try that one08:34
→ ipalaus joined08:34
crippa I have a file called ".gitignore" in my local folder and it contains "*.class". Nevertheless, my .class files are shown when I digit "git status". How to avoid it?08:34
→ durre joined08:35
jast crippa: are your .class files already added to git? because then the rules no longer apply to them...08:35
→ berserkr joined08:35
crippa that could be the case. How can I remove them?08:35
jast git rm --cached *.class08:36
← elnn left08:36
luka12345 jast: that was what i needed08:36
jast that won't be recursive, of course08:36
← ccapndave left08:36
durre I just got back from vacation and found a lot of faulty commits in the repo. whats the best practice for reverting to a certain commit?08:36
luka12345 jast: thank you08:36
→ alanhollis joined08:36
jast !revert08:36
gitinfo "Revert" is a heavily overloaded term. Do you mean: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard the uncommitted changes in the working tree [git checkout -- .]; c) undo committing [git reset --soft HEAD^, but type !rewriting_public_history] ?08:36
crippa jast: thanks a lot, that made teh trick.08:37
→ elnn joined08:37
→ tasslehoff joined08:37
durre I think I mean a) :)08:38
→ jkremser joined08:38
→ Element9 joined08:39
→ shreya joined08:39
tasslehoff is there a nicer way to do "git diff master origin/master > foo.diff"? I want all the changes in my master as one patch.08:40
jast if you want it as a single patch, I don't think there is08:41
← ill1cit left08:42
→ ill1cit joined08:43
wereHamster tasslehoff: what is wrong with "git diff master origin/master > foo.diff" ?08:43
how does that not work for you?08:43
luka12345 jast: after git checkout i did also git reset <commit ID> <path>08:46
← Lord_DeathMatch left08:46
← elnn left08:46
→ Lord_DeathMatch joined08:46
→ elnn joined08:46
luka12345 that way i dont have to deal with uncommited files in that commit08:46
← harshpb left08:46
← uman left08:47
→ jatt joined08:48
→ Vile joined08:49
→ weiyang joined08:49
→ Joshua_MG joined08:51
→ dzonder joined08:51
→ PigeonFriend joined08:51
tasslehoff wereHamster: I just wondered if there was a format-patch curse to do it08:53
Joshua_MG hi. Dummy question about who is starting on git: if I have a poroject with files 'a' and 'b' on my master, and create a branch that contains only the file 'a', when I update the content on master I can just merge the master on my branch and only the file 'a' will be updated (I mean, the file 'b' will not be "recreated" on my branch)?08:53
wereHamster tasslehoff: no. format-patch creates one patch for each commit08:53
FauxFaux Joshua_MG: !tias08:53
gitinfo Joshua_MG: Try it and see™. You learn much more by experimentation than by asking without having even tried. If in doubt, make backups before you experiment (see !backup). http://sitaramc.github.com/1-basic-usage/tias.html may help with git-specific TIAS.08:53
jast Joshua_MG: no08:54
but there's a trick to make it happen like that08:54
→ tqmz joined08:54
jast branch off, create the commit where you delete the other file (so that your new branch has only 'a' left), then go back to master and do a git merge -s ours yourbranch08:54
it's not pretty... but it works08:55
Joshua_MG uhm, got it08:55
tks08:55
→ BlindOracle joined08:55
jast you do this '-s ours' thing only once, though08:55
it's just to tell git: don't actually merge these changes but pretend that you did, so that they won't be applied again during a future merge08:56
Spockz_ → Spockz08:57
← aspotashev left08:57
→ thiago joined08:57
← twb left08:58
← iori left08:59
← Tasser left08:59
← felher left08:59
→ harshpb joined08:59
→ felher joined09:00
→ iori joined09:00
→ Tasser joined09:00
tasslehoff When I try to apply that patch with "patch -p1 --dry-run < foo.diff", most files work, but some that should be deleted give "Reversed (or previously applied) patch detected!". Any idea what could cause that?09:01
→ kinetic joined09:03
← kinetic left09:03
→ ashphy joined09:03
→ andref_dias joined09:04
thiago tasslehoff: try apply the patch with git apply instead09:04
tasslehoff seems to be when it encounters binary files09:04
cbreak-work git am is for applying patches made by format-patch09:05
→ teonanacatl joined09:06
teonanacatl hello09:06
gitinfo teonanacatl: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.09:06
tasslehoff I haven't tried git am, but both patch and git apply struggles with the binary files09:06
teonanacatl i've got a submodule pointing to a specific tag. how do i set this to HEAD for example, or another tag?09:06
[W] → Werring09:06
→ adityamenon joined09:07
← eighty4_ left09:07
→ apok joined09:07
→ eighty4 joined09:07
gitinfo set mode: +v09:07
gitinfo set mode: +v09:07
cbreak-work teonanacatl: submodules don't point to a tag09:08
they point to a commit09:08
that's the only thing they can point to09:08
also, submodules are always on HEAD09:09
→ cortana2 joined09:09
← bubbles10101 left09:09
cbreak-work like any other git repository too, since HEAD is the current commit/branch/...09:09
if you want to change to which commit a submodule points to, check it out, then add and commit in the parent09:09
teonanacatl git submodule shows: 7460ef14105be800f0ec9c7489d106748633eaa3 foo/bar (my-tag) - so i thought it is pointing to the tag "my-tag"09:10
other submodules are showing heads/master09:10
→ adityamenon90 joined09:10
cbreak-work teonanacatl: that's just the repository that you have as submodule09:12
submodules themselves are only recorded by commit09:12
teonanacatl i cd's into the submodule, checked out master, cd'd to the parent-project and committed - but it's still saying "my-tag"09:12
← adityamenon left09:12
teonanacatl this is so confusing.09:13
cbreak-work what does git status show?09:14
teonanacatl nothing to commit (working directory clean)09:16
← Zespre left09:16
→ Zespre joined09:17
← ipalaus left09:17
cbreak-work then it's ok.09:18
← Zespre left09:19
← havocologe left09:20
← d3vic3 left09:23
→ d3vic3 joined09:23
← d3vic3 left09:24
← jkremser left09:26
→ havocologe joined09:27
← simesy left09:27
← dhruvasagar left09:27
→ d3vic3 joined09:28
→ dhruvasagar joined09:28
← d3vic3 left09:28
← simesy_work left09:29
→ d3vic3 joined09:30
→ sylr joined09:31
← jceb left09:33
→ bubbles10101 joined09:33
← ashphy left09:34
→ theCole joined09:36
→ dr_lepper joined09:36
← dr_lepper left09:36
→ dr_lepper joined09:36
← caseymcg left09:37
← GameGamer43 left09:37
→ ashphy joined09:37
← heroux left09:38
← d3vic3 left09:39
→ heroux joined09:39
→ GameGamer43 joined09:39
→ jkremser joined09:39
→ tvw joined09:39
→ d3vic3 joined09:40
→ Per4u3e joined09:41
→ gusnan joined09:42
→ argoneus joined09:44
→ mabrand joined09:44
→ ipalaus joined09:44
← argoneus left09:45
→ argoneus joined09:45
→ achadwick joined09:45
← aristidesfl left09:47
→ aristidesfl joined09:49
→ rtjure joined09:50
→ kkszysiu_work joined09:50
← rbarraud_ left09:51
→ arietis joined09:51
← weiyang left09:51
← GameGamer43 left09:52
← Per4u3e left09:52
← jackh left09:52
← aristidesfl left09:53
→ GameGamer43 joined09:53
← ISF left09:54
→ Per4u3e joined09:54
shruggar External company sends entire month of work as a single commit, then notices they can't push a tag "we find them useful for tracking releases"09:56
my mind boggles09:56
→ aristidesfl joined09:56
→ jceb joined09:56
shruggar is it really such a difficult concept that if your commit-message has bullet points listing out what changes were made, those changes ought to be separate commits?09:57
← puneetgoyal left09:57
→ leeight_ joined09:57
FauxFaux I was in charge of recieving zipped source trees from contractors, and asking them for the missing files necessary to make it build, at my last job.09:59
andres Poor you.09:59
← superdmp left10:00
→ simesy_work joined10:01
→ puneetgoyal joined10:01
← jmd left10:01
→ sh0ne joined10:01
→ adhawkins_ joined10:01
shruggar we've just starting working with a company in india, and I've finally been listened to regarding contractual stipulations of how code should be delivered (it must be sent in a form which is conducive to code review, and various bullet points about what that actually means)10:01
← sh0ne left10:02
→ sh0ne joined10:02
shruggar I still need to tell them at least once a week "No, you cannot have access to our live back-end. You have been provided with a copy of our database, scrubbed of customer data. Use that."10:03
← aristidesfl left10:03
FauxFaux I really like !parable, but it's obviously far too long for anyone of that type to read.10:03
gitinfo 'The git parable' provides some good reasoning behind git. http://tom.preston-werner.com/2009/05/19/the-git-parable.html10:03
adhawkins → adhawkins-test10:04
→ maleknet joined10:04
→ aristidesfl joined10:04
→ Yuuhi joined10:05
← shruggar left10:05
← adhawkins_ left10:07
← aristidesfl left10:07
→ adhawkins joined10:07
← raek left10:09
→ aristidesfl joined10:11
← RaceCondition left10:11
← ashphy left10:14
→ Zespre joined10:14
→ robotmay joined10:14
← apok left10:14
← aristidesfl left10:16
← maleknet left10:17
← d3vic3 left10:17
→ aristidesfl joined10:18
→ d3vic3 joined10:19
→ maleknet joined10:19
→ amitprakash joined10:23
← ipalaus left10:23
amitprakash Hi... I made a couple of changes to a repo, and commited them.. is it possible to revert back from the commit to the staged state?10:24
← herdingcat left10:24
FauxFaux amitprakash: !reset Read the whole thing before running anything or you WILL LOSE ALL YOUR FILES.10:25
gitinfo amitprakash: A good resource explaining git-reset is http://progit.org/2011/07/11/reset.html10:25
andres amitprakash: not sure if its what you want but git reset --soft might be what youre looking for10:25
FauxFaux (And then we'll just point and laugh.)10:25
cortana2 i think --soft is the right one10:25
you will feel like indiana jones when picking out the cup of christ10:25
heytrav, the git-reset man page has been rewritten since i last read it. it's a lot better now10:26
amitprakash andres, well I already did a --soft HEAD^ .. but I am unable to checkout or git-rm the files I have removed using rm command instead10:27
jast note that you can just amend your commit, too, if all you want to do is fix it10:27
amitprakash FauxFaux, also, so long as reflog is there.. I think I should be able to get back to a commit :)10:28
← adityamenon90 left10:28
FauxFaux You keep believing that. I'll prepare to laugh. :)10:28
amitprakash :(10:28
jast 'unable to'? !work10:29
gitinfo [!doesntwork] Sorry to hear it doesn't work. What happened? Did it fall asleep on your couch, with your sister? Eat your homework? What did you want it to do? What happened instead? Please be specific!10:29
→ noisebleed joined10:32
← noisebleed left10:32
→ noisebleed joined10:32
→ roychri joined10:32
← adhawkins-test left10:32
→ babilen joined10:32
← jacobrask left10:33
← chaowang left10:33
→ plastics joined10:33
← babilen left10:34
← nasso left10:35
→ yatiohi joined10:36
→ adityamenon joined10:37
linduxed can i write comments in the .gitmodules file?10:37
and add some empty lines to make stuff more categorized?10:37
jast sure. prefix comments with #10:37
← plastics left10:37
linduxed ok, good thx10:37
jast same as in .git/config10:37
← tqmz left10:38
→ tqmz joined10:38
→ nasso joined10:39
← flijten left10:39
→ jacobrask joined10:40
→ ak5 joined10:40
← mklappstuhl left10:40
→ federic joined10:40
ak5 hi if I want to delete a branch on the origin, how can I do this? I deleted it locally already10:40
→ crazyrohila joined10:41
← crazyrohila left10:41
← flaguy48 left10:42
adhawkins ak5: git push --delete origin branch10:42
(I think)10:42
→ shwaiil joined10:42
shwaiil hi10:42
amitprakash jast, sorry.. by unable to I meant that on checkout of files deleted via rm -fR such as dojo library, git checkout path/ git rm path leads to 'fatal: pathspec $path did not match any files'10:42
→ RaceCondition joined10:42
shwaiil How to pull from repository and forget about anything done locally ? (suposely i'm ahead 10 commits, but this is old and not needed anymore)10:43
jast amitprakash: since you already did the soft reset, try: git checkout HEAD path/10:43
shwaiil: git fetch; git reset --hard origin/master (note that this will destroy uncommitted stuff, too)10:43
← Guest33096 left10:43
→ adhawkins2 joined10:43
ak5 adhawkins: that works, as does git push origin :branch10:43
amitprakash jast, that works.. thanks10:43
→ totte joined10:43
adhawkins ak5: Yes, the second syntax you quote is the original, hard to remember one.10:43
:)10:43
← adhawkins2 left10:44
ak5 thanks!10:44
← maleknet left10:44
amitprakash jast, could you please also provide an explanation as to why the latter works but the former doesn't?10:44
jast amitprakash: git reset --soft updates the ref but not the index. git checkout without commit arg tries to get the files from the index, where they were evidently deleted earlier.10:44
shwaiil jast: thanks a lot! I was trying to do 'git reset --hard HEAD', I was on master branch, what's the diference ? your command worked, maybe I were doing something wrong, not sure.10:45
jast shwaiil: !HEAD10:45
gitinfo shwaiil: HEAD is a 'pointer' in the repository that refers to whatever is currently checked out (or the default branch in bare repositories). It's not specific to any branch.10:45
amitprakash jast, ah, that makes sense10:45
thanks10:45
jast shwaiil: so, resetting to HEAD doesn't change your branch, it only throws away uncommitted changes10:45
← leeight_ left10:45
shwaiil I see! Well, when I do 'git push origin OTHER_BRANCH_NOT_MASTER' I get error: pathspec 'development' did not match any file(s) known to git.10:46
jast and you do have that 'development' branch locally, right? (I'm guessing that's what you tried to push)10:46
→ ISF joined10:47
→ naught101 joined10:48
← OOPMan left10:48
shwaiil I need to create it first and then pull ? ehh 'git checkout -b NEW_BRANCH_NAME origin/NEW_BRANCH_NAME something like this ?10:48
reald world 'git checkout -b development origin/development10:48
:T10:48
→ cheetsz joined10:49
← chrisward left10:51
→ chrisward joined10:51
← enmand left10:54
ziro` given a valid reference, how can i obtain the commit that this reference points to?>10:54
wereHamster ziro`: git rev-parse <ref>10:55
← chrisward left10:55
wereHamster ziro`: git rev-parse <ref>^{commit}10:55
ziro` does that return all the revs?10:55
wereHamster no, just one. a ref can point only to one commit10:55
← madsy left10:55
→ chrisward joined10:55
ziro` awesome ta :)10:55
wereHamster like a pointer. You can't have it point to multiple things. How would that work?10:55
ziro` alright, i just wondered if rev-parse returned all the commit id's10:56
that are children of the reference10:56
etc10:56
wereHamster that's what git rev-list is for, or git log10:56
ziro` ok, well I wasn't aware, but thank you :)10:56
→ puneet joined10:57
← sejo left10:57
→ cedr joined10:58
→ flijten joined10:58
← puneetgoyal left10:58
← aristidesfl left10:59
← ziro` left11:00
→ maleknet joined11:01
→ ipalaus joined11:02
→ aristidesfl joined11:02
→ spaceonline joined11:03
→ babilen joined11:04
Marcux83|away → Marcux8311:05
← chrisward left11:06
← d3vic3 left11:06
→ shruggar joined11:06
→ d3vic3 joined11:08
← maleknet left11:08
→ Sil4nc4 joined11:10
← flijten left11:11
→ netzmonster joined11:11
→ angelsl joined11:12
← cedr left11:12
→ chrisward joined11:12
→ cedr joined11:12
→ flyz joined11:13
→ RaceCondishun joined11:13
← RaceCondishun left11:13
→ RaceCondishun joined11:14
← roychri left11:14
→ RaceCondishun_ joined11:14
← ottoshmidt left11:16
← RaceCondition left11:16
← RaceCondishun left11:18
← shruggar left11:19
→ rtuin joined11:20
← marthinal left11:22
shwaiil Q: Is there a way to RM all deletes files on working directory ? I usually do it one by one11:24
← daniel-s left11:25
thiago git rm *11:25
FauxFaux shwaiil: man git add, flags like -u and -A.11:25
gitinfo shwaiil: the git-add manpage is available at http://git-scm.com/docs/git-add.html11:25
shwaiil thiago: won't that rm the other files ?11:25
FauxFaux: I'll take a read, tks11:25
_ikke_ shwaiil: yes11:25
→ daniel-s joined11:25
← Fandekasp left11:25
shwaiil _ikke_: yes what ? =)11:25
thiago shwaiil: you asked how to delete all files11:26
shwaiil: there are no "other files" besides "all files"11:26
→ Fandekasp joined11:26
→ roychri joined11:26
_ikke_ thiago: RM all deleted files11:26
→ maleknet joined11:26
shwaiil thiago: yeah, what I really ment is, I deleted some files in the working directy, I got a list of "deleted files" and I need to "GIT RM" all on that ls.11:26
→ marthinal joined11:27
shwaiil not sure if made more sense :T11:27
I'm taking a read onf FauxFaux -u and -A tip11:27
thiago shwaiil: you can just commit11:27
← jargon- left11:27
→ programer1234 joined11:27
FauxFaux thiago: Only with -a, which people here genrally complain about.11:28
_ikke_ shwail: From the man: git ls-files -z | xargs -0 rm -f11:28
← programer1234 left11:28
_ikke_ Oh wait11:28
→ n123 joined11:28
_ikke_ shwail: From the man: git ls-files -d -z | xargs -0 rm -f11:28
FauxFaux She almost certainly wants "git add -u". :p11:28
_ikke_ shwail: From the man: git ls-files -d -z | xargs -0 git rm11:28
got the wrong one11:28
→ mklappstuhl joined11:29
_ikke_ Hmm, here is a better version: git diff --name-only --diff-filter=D -z | xargs -0 git rm --cached11:29
→ enmand joined11:30
shwaiil FauxFaux: -u works fine11:30
EmilyS heya enmand11:31
long time :)11:31
enmand Hey EmilyS11:31
gitinfo enmand: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.11:31
enmand Do I know you EmilyS?11:31
EmilyS <-- Noia11:31
enmand Ahh11:31
md_5 → md_5|away11:31
enmand Well, that makes more sense :p11:31
EmilyS :D11:31
enmand :D11:31
How're things?11:32
EmilyS it probably raises more questions than answers but still :P11:32
Good! career is taking a turn for the awesome11:32
enmand Nice! What are you working on these days?11:32
→ sejo joined11:33
← roychri left11:34
→ fsvend joined11:36
→ flijten joined11:37
← ak5 left11:37
← Milossh left11:37
→ simesy joined11:38
← maleknet left11:38
← spaceonline left11:39
→ maleknet joined11:39
← simesy left11:39
→ beautifulmind joined11:39
→ OOPMan joined11:40
→ JeremyWei joined11:41
← maleknet left11:41
→ spaceonline joined11:42
→ raatiniemi joined11:43
← zdk left11:45
← bubbles10101 left11:46
Marcux83 → Marcux83|away11:46
cedr → zombiecedr11:49
→ jjbohn joined11:50
← toothrot left11:51
→ AdrienBrault joined11:51
→ toothr joined11:53
← harshpb left11:54
← tango_ left11:54
→ harshpb joined11:55
← ketas left11:55
→ tango_ joined11:56
← puneet left11:57
→ puneetgoyal joined11:57
→ maleknet joined11:58
→ ketas joined11:59
← AdrienBrault left11:59
← quazimodo left12:00
→ harv_ joined12:01
← harv left12:01
harv_ → harv12:01
← toothr left12:01
← marthinal left12:02
← n123 left12:02
→ simesy joined12:03
→ Tinti joined12:03
← maleknet left12:04
→ mntzn_ joined12:04
→ toothr joined12:04
→ axt_work_ joined12:04
→ maleknet joined12:05
→ ryez joined12:06
← mntzn_ left12:06
← flyz left12:06
← philips left12:06
→ mntzn joined12:06
mntzn Hi, If I've just checked out an empty repository(git branch gives no output), how do I then push to a specific repository that should not be named master?12:06
wereHamster mntzn: what do you want to push if there is nothign?12:07
you can not push nothing. how would that work.. ?12:07
→ intransit joined12:07
mntzn touch qwe; git add . ; git commit -m 'whatever'; git push12:08
wereHamster also, master is the name of a branch, not a repository12:08
and.. that doesn't work?12:08
mntzn it goes pushesh to a master branch12:08
→ jargon- joined12:08
wereHamster and how would you like to name your local branch?12:08
mntzn can I make it so that the firsh push will be to my named branch other than master?12:09
wereHamster sure you can.12:09
mntzn great, how?12:09
→ philips joined12:09
→ ziro` joined12:09
← crippa left12:09
wereHamster git push origin HEAD:whateverzzzzz12:09
jast you could just rename your local branch12:09
← mklappstuhl left12:11
mntzn and that will make that branch a default one to be pulled from ?12:11
← ehsan left12:11
mntzn or clones12:11
cloned*12:11
wereHamster non12:12
_ikke_ mntzn: No, you would have to use git symbolic-ref on the remote repository12:12
→ shruggar joined12:12
→ SkAzZ joined12:12
mntzn ok, thanks12:13
← zombiecedr left12:15
← alanhollis left12:16
→ flyz joined12:17
ziro` http://a4.sphotos.ak.fbcdn.net/hphotos-ak-ash4/311589_467733126588009_148568698_n.jpg12:18
woopps12:18
sorry!12:18
was a little bit hasty with that one :P12:18
← ziro` left12:18
→ marthinal joined12:20
→ RaceCondition joined12:20
← soee left12:20
→ cedr joined12:20
← _br_ left12:21
→ _br_ joined12:21
→ n123 joined12:22
← _br_ left12:22
← toothr left12:23
← RaceCondishun_ left12:23
← jjbohn left12:23
→ soee joined12:23
← Torsten85 left12:23
→ cedeon joined12:23
→ _br_ joined12:24
→ toothr joined12:24
→ jackh joined12:25
→ Torsten85 joined12:25
← jackh left12:25
→ jackh joined12:27
→ jn_ joined12:28
→ Sebo joined12:29
→ canton7 joined12:29
← arti_t left12:30
Sebo Hi, How can I swich the commits which the branch points to?12:30
→ adhawkins2 joined12:30
FauxFaux git branch -f branchname newcommitname12:30
← nisseni left12:31
← adhawkins2 left12:32
Sebo FauxFaux: thats it :)12:33
→ jcromartie joined12:33
← ph^ left12:37
← jn_ left12:37
→ ph^ joined12:38
→ fatpowder joined12:38
gitinfo set mode: +v12:38
fatpowder hi. in a git repository, whenever someone commits a specific config file, how can i reject that?12:39
cmn you can use a pre-commit hook, but you can't enforce any checks at commit time12:39
the user can either remove the hook or use --no-verify12:40
you can refuse an update in a central repo under your control12:40
fatpowder cmn in pre-retrieve hook?12:41
cbreak-work what's that?12:41
fatpowder: man githooks12:41
gitinfo fatpowder: the githooks manpage is available at http://git-scm.com/docs/githooks.html12:41
cmn that doesn't exist12:41
→ pyhead joined12:41
cmn and how would something pre-retreival help with committing or updating a repo?12:42
→ eurek joined12:42
← Tinti left12:42
← ph^ left12:42
cbreak-work just reject pushes that touch the config file12:42
(don't see the point in that though, honestly...)12:43
_ikke_ You shouldn't track config files anyway12:43
!config12:43
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.12:43
← _Vi left12:43
→ buggaboo joined12:43
→ Ryback_ joined12:45
← Ryback_ left12:45
→ Ryback_ joined12:45
→ imexil joined12:45
legumbre_ → legumbre12:46
← imexil left12:46
→ d0k joined12:48
→ cakehero joined12:49
→ bombetta98888 joined12:49
← maleknet left12:49
← bombetta98888 left12:50
→ ehsan joined12:50
→ TrahDivad joined12:50
← yacc left12:50
← ipalaus left12:50
← TrahDivad left12:51
→ fsvend_ joined12:51
← d0k left12:51
→ tommyvyo_ joined12:51
→ ToBeFree joined12:51
→ d0k joined12:51
→ jaseemabid joined12:52
→ kanzie joined12:53
kanzie how can I export all commit messages into a change log?12:54
jast redirect output of git log or something12:54
it has many options to control the format of the output, notably --pretty=...12:54
→ madsy joined12:54
FauxFaux !lg !lol12:54
gitinfo git config --global alias.lg "log --graph --pretty=format:'%Cred%h%Creset -%C(yellow)%d%Creset %s %Cgreen(%cr) %C(bold blue)<%an>%Creset' --abbrev-commit --date=relative"12:54
git config --global alias.lol "log --oneline --graph --decorate"12:54
← anatolyspektor left12:54
_ikke_ FauxFaux: That's probably not what kanzie wants12:55
← fsvend left12:55
jast graphs aren't that useful for changelogs, now are they :)12:55
FauxFaux I'd rather have a graph than a linear changelog. :)12:55
_ikke_ lol12:55
← JeremyWei left12:55
kanzie yeah, I kind of need one liners12:55
jaseemabid Can I check for whitespace errors in an existing commit? I applied a patch with some w/s issues. Now I want to fix it up12:55
← juvenal left12:55
cmn kanzie: check out shortlog12:56
BinGOs https://metacpan.org/source/BINGOS/CPANPLUS-YACSmoke-0.82/tools/git-log.pl12:56
→ keller joined12:56
→ ij joined12:56
jast --pretty=oneline12:56
→ glennpratt joined12:56
← dr_diamond left12:56
jaseemabid jast git log --oneline12:56
→ juanfer joined12:57
ij When I cherry-pick, the picked commit in dest branch is a different, new one. How do I make them the same one later? Merging later just makes two commits appear in the source branch.12:59
cmn that's what a cherry-pick is12:59
ij So you mean I can't union them later as one commit?13:00
→ hellopat joined13:00
← hellopat left13:00
cmn no, because cherry-pick necessarily creates a new commit13:00
→ oriba joined13:00
→ cesc joined13:00
ij Hmm, okay.13:00
cmn if you want the same commits, you mere13:00
merge13:00
→ hellopat joined13:01
→ lnostdal_ joined13:01
← harshpb left13:02
← cilly left13:02
← rvsjoen left13:03
→ rvsjoen joined13:03
← afuentes left13:03
← Targen left13:03
→ plastics joined13:03
← jhardinee left13:04
→ bitkiller joined13:04
← lnostdal__ left13:05
→ maleknet joined13:06
→ kerrick joined13:07
→ macabre joined13:08
→ TeckniX joined13:08
← eurek left13:09
→ ipalaus joined13:10
→ Jon30 joined13:11
← d3vic3 left13:11
← enmand left13:11
→ d3vic3 joined13:12
← fsvend_ left13:13
← dhruvasagar left13:13
→ enmand joined13:13
← maleknet left13:14
→ dhruvasagar joined13:14
→ quazimodo joined13:15
← kermit left13:15
→ AaronMT joined13:15
← AaronMT left13:15
→ phantomtiger joined13:15
→ AaronMT joined13:16
← AaronMT left13:16
→ kermit joined13:16
← PRETTY_FUNCTION left13:17
→ AaronMT joined13:17
← AaronMT left13:17
→ AaronMT joined13:17
← iori left13:18
→ iori joined13:18
→ ziro` joined13:18
← iori left13:18
ziro` is there something like $Format:%d$ which will only give me tags, not branch refs etc ?13:19
→ iori joined13:19
← cooloney left13:19
cmn for what command? what kind of syntax is that?13:19
→ eurek joined13:19
→ nisseni joined13:19
← palm_vue left13:20
← Lgb left13:20
ziro` that's a substitution string for archive, but git log --format="%d"13:20
→ rgsteele joined13:20
→ palm_vue joined13:21
→ Lgb joined13:21
kerrick Git GUI is missing the right-side panes (the diff and the commit/push controls). I can't figure out how to restore them. This is on OS X, if it matters.13:22
→ ph^ joined13:22
← iori left13:22
← ij left13:22
selckin tried deleting the config file?13:23
← d3vic3 left13:23
← Torsten85 left13:24
→ aruiz joined13:24
← toothr left13:27
← intransit left13:28
← nisseni left13:28
→ dv310p3r joined13:28
← durre left13:28
→ intransit joined13:28
← MiJyn left13:29
→ benjo2 joined13:30
→ d3vic3 joined13:30
→ MiJyn joined13:30
→ toothr joined13:30
← aristidesfl left13:31
← p3rror left13:32
← aruiz left13:32
→ abledea_ joined13:33
← abledea_ left13:33
→ loin joined13:34
→ puneet joined13:34
loin can i have the files changed in lists only display the file names13:34
and not the full abstract path13:34
← OOPMan left13:35
← puneetgoyal left13:35
← ehsan left13:37
→ dcarper joined13:37
→ jn_ joined13:38
← yongleflibbit left13:39
→ Torsten85 joined13:39
→ tudorstudio joined13:39
selckin there is --short, else you'l have to use --porcelain and script it13:40
→ Syzygy_ joined13:40
← jackh left13:40
→ jackh joined13:41
← flyz left13:41
← dhruvasagar left13:42
→ milky joined13:42
→ bartek_ joined13:43
← tubbo left13:43
← ThomasLocke left13:43
→ neurodrone joined13:44
→ ehsan joined13:45
jaseemabid Patch application failed. error: patch failed: <file>:1 . Is there anyway I can somehow apply the patch. Its a bit big change to apply manually13:45
→ aruiz joined13:45
jaseemabid And what does that <file>:1 indicate ?13:45
← raatiniemi left13:47
→ ManDay joined13:47
ManDay How can I force overwrite the remote after I did a git commit --amend AFTER I pushed before?13:47
← cagedwisdom left13:47
← plastics left13:47
canton7 ManDay, !rewriting_public_history13:48
gitinfo ManDay: 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)13:48
→ v0n joined13:48
cbreak-work jaseemabid: what does git have to do with patches?13:48
loin selckin, --short doesn't work, i get usage 'myworkingdirectory' [{blame|browser|citool}]13:48
cbreak-work ManDay: force push...13:48
jaseemabid cbreak-work, git am13:48
← linduxed left13:49
cbreak-work note that this won't kill remote history that was already fetched to somewhere else13:49
ManDay cbreak-work: thanks, --force worked13:49
canton7: yes, I know, it's bad practice13:49
→ converge joined13:49
ManDay it's my tiny personal repo13:49
← Sebo left13:49
canton7 ManDay, the trigger also tells you how to force-push ')13:49
cbreak-work jaseemabid: so, and are you on the commit that the patch was formated to?13:49
→ pandeiro joined13:49
cbreak-work jaseemabid: and was the patch made with format-patch?13:49
→ mklappstuhl joined13:49
converge someone knows a free private git repository ?13:50
← loin left13:50
canton7 converge, bitbucket13:50
jaseemabid cbreak-work, It was a patch generated long time back. Its an important change and stayed as a txt file there. Was trying to apply it. But the problem is that the repo changed a *bit*13:50
→ leeight_ joined13:50
converge canton7: thanks13:50
→ cilly joined13:50
→ linduxed joined13:51
jaseemabid cbreak-work, Yes, made with fomat-patch, It was supposed to be applied on <commit:XXX> but there is no XXX anymore. I have a *very similar* YYY. I was trying to apply it ignoring the minor diff b/w XXX and YYY13:51
← jaseemabid left13:52
→ jaseemabid joined13:52
cbreak-work jaseemabid: then make a branch from the point when the patch was made on13:52
← und3f left13:52
← yizumi_ left13:52
cmn jaseemabid: have you tried telling it to ignore whitespace and/or to use the three-way merge?13:52
cbreak-work commits in git don't vanish.13:52
just git checkout -b parentofyourpatchcommit13:52
then apply it there13:52
jaseemabid cmn, I tried both. Not working. I am applying it manually after playing with it for a while.13:53
cbreak-work, History rewrite, BIG TIME13:53
cmn is the patch really in mbox format?13:53
it was complaining about the first line, wasn't it?13:53
jaseemabid cmn, Yes. Any other kickass patching tools?13:53
→ bgerber joined13:54
cbreak-work jaseemabid: ...13:54
→ harshpb joined13:54
cmn for mbox, that's am; for normal patches, apply13:54
jaseemabid What i am really looking for is something like meld. I use it for merge conflict fixing. I need an interactive patcher13:54
cbreak-work why didn't you commit it if it was so important?13:54
that'd make everything easier13:54
maybe you can find the pre-rewrite history in the reflog13:54
← ManDay left13:55
jaseemabid cbreak-work, ... means hopeless? lol. This is a patch i *can* manually apply. Not that big. I was trying to learn it this time.13:55
Though, manually applying a patch it a pita13:55
cbreak-work not really hopeless13:55
but not a git job13:55
→ revagomes joined13:55
cbreak-work you'll probably have to resort to external tools like text editors13:56
jaseemabid cbreak-work, Know any good patch tools then?13:56
cbreak-work I don't use patches.13:56
← philips left13:57
jaseemabid cbreak-work, This is just another ugly instance. Nevermind13:57
← amitprakash left13:58
_ikke_ jaseemabid: Doesn't get apply work for applying patches?13:58
→ davidjdagino joined13:58
→ yangchang joined13:59
_ikke_ git*13:59
← DofLeFou left13:59
← yangchang left13:59
jaseemabid _ikke_, Conflicts conflicts13:59
← soee left13:59
→ ij joined13:59
_ikke_ Well, doesn't any patch tool have conflicts?13:59
← BarryCrayon left13:59
ij Why's git stash list -p showing different results? Because I've added code since I've saved them?14:00
→ yangchang joined14:00
_ikke_ ij: different from what?14:00
→ philips joined14:00
ij Different from what git show stash shows for each stash commit.14:00
← cakehero left14:00
→ _md joined14:00
jaseemabid _ikke_, error: patch failed: <file>:1 is all I know. Its more a patch bug than a git issue.14:01
← oo_miguel left14:01
→ oo_miguel joined14:01
← puneet left14:01
ziro` if a git pull results in conflict, does it return an error code?14:02
error status14:02
→ booginga joined14:02
bremner yes. But you probably don't want to script pull. At least fetch and merge seperately.14:03
← gour left14:03
← babilen left14:03
→ Zenmaker_ joined14:04
gdoteof I made two commits into master that should have been on a new branch14:05
_ikke_ gdoteof: !fixup14:06
gitinfo gdoteof: 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, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!14:06
gdoteof i created a new branch where it "should" have been, but i am not sure how to move the commits there14:06
!fixup_hints14:06
gitinfo Hints for fixing commits are: (1) NOT PUSHED/PUBLISHED: `git rebase -i $COMMIT^` or perhaps `git commit --amend` (or `git reset HEAD^`). (2) OTHERWISE, `git revert $COMMIT` to make a reverse commit. (3) If you have pushed and MUST remove it, use rebase or filter-branch and type !rewriting_public_history in IRC.14:06
_ikke_ gdoteof: How did you create that branch?14:06
cbreak-work you can't move commits.14:06
best you can do is cherry-pick or merge them14:06
_ikke_ or rebase them14:06
gdoteof _ikke_: git branch my_branch 18734cc14:06
where 18734cc is the first commit that should have been on the new branch14:07
_ikke_ combination of cherry-pick and git reset14:07
← marthinal left14:07
← benjo2 left14:07
← kkszysiu_work left14:07
→ Sebo joined14:08
← Zenmaker__ left14:08
_ikke_ Something like git checkout my_branch; git cherry-pick master..master~2; git checkout master; git reset --keep HEAD~2;14:08
→ Destos joined14:08
← pandeiro left14:09
← oriba left14:09
→ plastics joined14:09
← ill1cit left14:10
← harshpb left14:10
→ ill1cit joined14:11
ziro` bremner: I'm trying to put together a command to update my local master, update the changelog, commit, tag then package for release14:11
→ harshpb joined14:11
→ cakehero joined14:11
← Sebo left14:12
→ beatak joined14:12
→ Sebo joined14:12
→ chrisgeorge joined14:13
← jettero left14:14
← mntzn left14:15
← kcj left14:15
→ randomubuntuguy joined14:16
Olipro does a "git svn rebase" get performed against the currently checked out git branch or just against master14:16
gdoteof _ikke_: thanks a lot that was almost it exactly14:16
← lnostdal_ left14:16
_ikke_ gdoteof: What did you do different?14:17
← ehsan left14:17
blameless how can I see which is the most recently created of 2 particlar branches14:18
→ ferdna joined14:18
jast Olipro: against the first SVN branch found in a 'git-svn-id' line in commits of the currently checked out git branch14:18
(don't ask!)14:19
→ jjbohn joined14:19
Olipro so basically, if I check out an SVN repo and that branch is master14:19
if I then create branch "mybranch" and make a few commits14:20
→ Davey joined14:20
Olipro doing a "git svn rebase" will presumably rebase against the creation point of my branch14:20
→ rurufufuss joined14:20
→ pandeiro joined14:21
← leilaM left14:21
jast yes14:22
← RaceCondition left14:22
→ fly9 joined14:22
jast you should base your new branch on the SVN branch of your choice14:22
and you can use 'git svn branch' to create a new one (iirc)14:22
← jatt left14:23
← gavin_huang left14:23
← tasslehoff left14:23
← axt_work_ left14:24
→ gavin_huang joined14:24
→ hwrd|work joined14:25
→ codeM0nK3Y joined14:25
→ Thanatermesis joined14:26
← yangchang left14:27
→ bilel joined14:27
converge I had add a wrong file to the repository, is there some way to delete it to never be listed in the history ?14:27
← _br_ left14:27
converge *there is a password in the file14:27
gitinfo set mode: +v14:27
converge already added it to gitignore14:28
RandalSchwartz converge, !fixup14:28
gitinfo converge: 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, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!14:28
RandalSchwartz and !rewrite14:28
gitinfo [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)14:28
RandalSchwartz so hopefully, you haven't pushed it anywhere, right?14:28
Thanatermesis im going to buy a wordpress plugin so i would like to made changes on it under git, one branch for the original code and another for the modifications (so that i can switch to the original sources to import updates from the original sources), from there is not a problem so i have already do that, but my question is that i want to have this as a part of an already-existing git repository (with other plugins), how i can should do that ?14:28
FauxFaux converge: 1) Change the password. 2) Get on with your life.14:28
RandalSchwartz waits for converge14:29
← kenperkins left14:29
→ ManDay joined14:29
ManDay How can I undo the lats commit I already pushed to remote?14:29
← harshpb left14:30
converge FauxFaux, RandalSchwartz, thanks14:30
RandalSchwartz ManDay, !fixup and !rewrite14:30
gitinfo ManDay: 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, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!14:30
ManDay: [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)14:30
← randomubuntuguy left14:30
RandalSchwartz I should just have that on a key. :)14:30
bilel hi, I made some test changes on my project, it's like a sandbox, so I've put them in a stash, currently it's stashed, but in the meanwhile I've made a lot of changes (new untracked files and changes), not staged for a commit, I don't want to commit this changes yet, but I'd like to go to my sandbox stash, witouth mixing this changes to the changes of the sandbox stash. how should I do? thx14:30
RandalSchwartz bilel - you must be a guy. afraid of comitting14:31
→ harshpb joined14:31
RandalSchwartz I make throw-away commits all the time14:31
easy to undo later14:31
git commit -a -m 'WIP' and go somewhere else14:31
ManDay thanks RandalSchwartz14:31
RandalSchwartz just. don't. push. them. :)14:31
ManDay Not an issue here14:32
I don't have a big audience14:32
I rather prefer a clean (false ;)) history14:32
RandalSchwartz that's for bilel14:32
ManDay come again14:32
RandalSchwartz I was talking to him not you14:32
→ opdynamiks joined14:32
→ ipoval joined14:32
bilel RandalSchwartz: yes because usually I commit only when the feature is ready, but you're right I could do some wip commits and undo, thx ;)14:32
RandalSchwartz I make commits whenever I've made forward progress14:33
"feature ready" is usually about 18 steps away :)14:33
if I need to squash those before push, easy enough14:33
_ikke_ bilel: git stash has a branch command, in order to create a branch for the stash, based on the commit where the stash was made on14:34
← Lord_DeathMatch left14:34
→ guampa joined14:36
→ mikepack joined14:37
→ randomubuntuguy joined14:37
← mocramis left14:37
→ kkiero joined14:38
← ManDay left14:38
→ airborn joined14:39
← harshpb left14:40
← tqmz left14:40
RandalSchwartz I was playing with git long before git-stash was added14:40
so I still think in terms of WIP commits for temp parking places.14:40
→ harshpb joined14:41
← ISF left14:41
FauxFaux doesn't really like the stash workflow, either, if the data has to exist for more than a few minutes.14:41
jast same here14:42
→ raatiniemi joined14:42
→ ottoshmidt joined14:42
← naught101 left14:42
ij What's the stash workflow?14:42
What do you mean by that?14:43
bilel When I do git stash list, I can see my stash, it's written : "stash@{0}: On master: form form sandbox" , when I created it, the name was "form" and the comment "form sandbox", it looks ok, but when I do git stash pop form it says the argument is ambigous, do you know why?14:43
ij Are there any more stashes?14:43
← Axsuul left14:43
bilel no14:43
→ hoschi joined14:44
← simesy left14:45
_ikke_ bilel: " When no <stash> is given, stash@{0} is assumed, otherwise <stash> must be a reference of the form stash@{<revision>}"14:45
bilel: You can't reference a stash by it's name14:45
→ sdeboey joined14:45
→ lotus joined14:46
→ duckxx joined14:47
→ ehsan joined14:47
← bgerber left14:47
← tvw left14:48
→ tvw joined14:49
lotus hey, I have an interesting situation. I have a remove devsite and a local devsite. I have a "master" and a "dev" branch. My dev branch is for my local devsite and has a .htaccess RedirectBase of /local/path and a database.php file of user='localDbUser' and pass='localDbPass'. My master branch has a .htaccess file with RedirectBase /remote/path and a database.php with user='remoteDbUser' and pass='remoteDbPass'. What I need is to be able to merge the dev bran14:49
← ipoval left14:49
Thanatermesis im going to buy a wordpress plugin so i would like to made changes on it under git, one branch for the original code and another for the modifications (so that i can switch to the original sources to import updates from the original sources), from there is not a problem so i have already do that, but my question is that i want to have this as a part of an already-existing git repository (with other plugins), how i can should do that ?14:49
← randomubuntuguy left14:49
→ ThomasLocke joined14:50
→ randomubuntuguy joined14:50
→ TheJH joined14:50
bilel _ikke_ : ok so I should reference it by stash@{form}, it doesn't work too, but I've made git stash pop and it worked, it looks like when I save the stash I can't give it a name, that was the issue, I had done git stash save form "form sandbox", it is wrong?14:51
lotus Thanatermesis: I'm thinking you should pull your two branches into your existing repository perhaps14:51
milki Thanatermesis: its common to use !sub14:51
gitinfo Thanatermesis: [!subtree] 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 options14:51
jast that's not a name for the stash, it's more of a description14:51
← eurek left14:52
← beautifulmind left14:52
→ banghouse2 joined14:52
jast there is no 'name' argument for 'git stash save'14:52
→ Talisman joined14:53
_ikke_ bilel: revision is a number14:53
← Talisman left14:54
bilel jast: oh yes you're right, thanks14:54
← mklappstuhl left14:54
bilel so we only identify stashes by descriptions, then we have to check the number doing git stash list14:54
jast yeah... you really don't want to have a ton of stashes. it gets chaotic.14:55
← lotus left14:55
→ lotus_ joined14:55
→ brentdanley joined14:55
← lotus_ left14:55
infogulch|afk → infogulch14:55
← pandeiro left14:56
→ lotus joined14:56
← glennpratt left14:56
→ glennpratt joined14:56
bilel yes indeed, I'll only have one, to put all the dirty tests and play with the project, don't see other case :)14:56
← harshpb left14:57
→ harshpb joined14:57
← Cam left14:57
← lotus left14:57
→ lotus joined14:58
→ yacc joined14:58
→ Talisman joined14:58
→ xckpd7 joined14:58
→ giallu_ joined14:58
→ raek joined14:59
xckpd7 question.... let's say I have a master branch... I introduce massive changes... let's say that pretty much any branch I would work on afterwards would conflict with. Let's say I would tell someone to merge the major changes first before anything else14:59
what would be the correct way to circumvent that?14:59
have all the subsequent branches (before that branch is merged into master) be rebased (or merged in) with the risky branch?15:00
→ leilaM joined15:01
milki i dont understand the last hypothetical statement with merging the major changes first15:01
o.O15:01
← maloik left15:02
milki are you trying to circumvent the conflicts?15:02
variable → Guest800015:02
→ mocramis joined15:02
← constant left15:02
→ constant joined15:02
constant → trout15:02
milki even if you rebase, you will still have the conflicts15:02
← harshpb left15:02
milki you should be rebasining before merging anyways15:03
→ harshpb joined15:03
← Sebo left15:04
← jonnu left15:04
xckpd7 milki: yes but put it like this... you have a branch with 1000 lines of pretty drastic code based off of master... and then you have like 2 lines of code based off of master... I want the branch with the 1000 lines merged first, because even if I have conflicts, I want the branches with the least changes to conflict15:04
← Talisman left15:04
→ Milossh joined15:04
xckpd7 I don't really have the ability to give an order to the person that I give pull requests to15:05
← harshpb left15:06
xckpd7 I figure for all the small commits (I don't know git as well as you guys so bear with me), I want to operate as if that one 1000 line commit was already merged in with them, except again I don't want to have to manage the conflict of 1000 lines so I want the bigger thing to appear earlier in history15:06
← bilel left15:06
xckpd7 as if I created that small branch directly off of the 1000 line branch15:06
→ harshpb joined15:06
← fatpowder left15:07
xckpd7 so no matter what order the guy pulls the branches in, it will merge cleanly. that's my goal15:07
→ YesJustWolf joined15:07
→ mklappstuhl joined15:08
selckin doesn't mather15:08
you'll get same amount of conflicts no mather the order15:08
xckpd7 selckin: but there is no way for me to manage the conflicts by just merging in the 'problem' branch into the smaller commits?15:09
→ surendra joined15:09
← harshpb left15:09
← cbreak-work left15:09
→ harshpb joined15:10
selckin maybe i don't understand what you're describing, but you can rebase all the other branches on top of that one instead of master sure15:10
milki you will always get the conflicts where there are conflicts15:11
you can change your branch tips to those with least conflicts if you want15:11
then it will be easier to merge15:11
→ bloony- joined15:11
xckpd7 milki: ?15:11
milki but then you will eventually need to rebase the commits that will conflict15:11
← ben_alman left15:12
xckpd7 selckin: ok so for all the smaller branches, just rebase off the problematic branch (which sorta merges) and I should be good15:12
milki if you can identify the commits that wont result in conflicts, reorder those commits so they are on a single branch and rebase that branch15:12
basically, move the troublesome commits to a later point so you merge them last15:13
← pyhead left15:13
← bloony- left15:13
→ andrewhl joined15:13
→ bloony- joined15:13
← john_anderson left15:14
← harshpb left15:14
xckpd7 milki: the troublesome commits are in that one branch basically. by rebasing, am I not doing what you just described? Thanks for the help by the way15:14
gitinfo set mode: +v15:14
gitinfo set mode: +v15:14
→ harshpb joined15:14
→ sebrock joined15:14
→ john_anderson joined15:14
← john_anderson left15:15
surendra hi, i have a branch called release and I want my master branch to complete sync to it and undo any extra commits in master branch. how do i do that?15:15
← duckxx left15:15
→ heptadecagram joined15:16
milki if you rebase all your branches on that single branch, you are effectively dealing with the conflicts now instead of later15:16
← psoo left15:16
xckpd7 milki: yes thanks15:16
milki surendra: you could do a branch --force15:16
→ branduren joined15:17
← plastics left15:17
← MiJyn left15:17
→ MiJyn joined15:17
← shreya left15:17
Thanatermesis so correct me if im wrong with that: with subtree-merging im able to "git pull" and automatically update the sources from the extra sources into my actual git, right ? then if i push this complete-git it includes the subtree so that others can get it too ?15:18
RandalSchwartz you might want to look at git-subtree for that15:18
→ ben_alman joined15:18
RandalSchwartz we're using that here in production (nearly)15:18
← v0n left15:19
Thanatermesis RandalSchwartz, this ? http://git-scm.com/book/en/Git-Tools-Subtree-Merging15:19
→ puneetgoyal joined15:19
← hwrd|work left15:20
RandalSchwartz no15:20
git-subtree - it's in contrib15:20
possibly in mainstream soon15:20
→ duckxx joined15:20
RandalSchwartz git-subtree is what subprojects should have been. :)15:21
← giallu_ left15:21
← flijten left15:22
→ AeroNotix joined15:22
← maletor left15:23
Mkaysi_ → Mkaysi15:23
← kidsalsa1 left15:24
→ kenperkins joined15:25
Thanatermesis now im more confused :/15:25
because i dont know what to do lol15:25
← n123 left15:25
← banghouse2 left15:26
→ marthinal joined15:26
← raatiniemi left15:27
surendra milki: thanks15:28
← surendra left15:28
→ erichynds joined15:28
← zackiv31 left15:28
→ juvenal joined15:28
milki np15:28
→ kpl joined15:31
← rtuin left15:31
← converge left15:31
← mikepack left15:32
← lamont left15:34
← neurodrone left15:34
→ lamont joined15:35
→ hwrd|work joined15:35
← txdv left15:36
→ txdv joined15:36
thorbjornDX how do subtrees differ from submodules?15:37
→ bubbles10101 joined15:37
← harshpb left15:37
→ psoo joined15:38
→ harshpb joined15:38
jast !submod15:38
gitinfo [!submodule] 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.html15:38
→ phishy joined15:38
jast !subtree15:38
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 options15:38
← phishy left15:38
→ notmypudding joined15:39
→ brentratliff joined15:39
← matt_taws left15:39
→ matt_taws joined15:40
thorbjornDX jast: thanks. So do I have full checkout/branch control over a subtree? (and will it all show up in the superproject's history?)15:40
← adityamenon left15:41
jast well. with subtree, you basically merge updates from the upstream/original subproject into your main branch.15:41
→ thearthur joined15:41
thorbjornDX and the subtree will exist as it's own directory?15:42
jast so the history will be preserved, but it's basically fused together with your main project's history (i.e. you can't stay in master and switch the subtree to an different branch)15:42
that's the idea, yeah15:42
thorbjornDX jast: Okay, so the files are more closely tied w/ my actual repository (superproject)15:42
jast yeah15:42
← dwmw2_gone left15:42
→ mdhopkins joined15:42
thorbjornDX jast: what about pushing changes back to the subtree's host? (is that a no-no?)15:43
jast you'd have to do that separately, i.e. in a separate clone of the original subproject repository15:43
thorbjornDX jast: okay, good to know15:43
jast there's one other approach to subprojects you might be interested in: !gitslave15:43
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.15:43
← d3vic3 left15:43
→ kevwil joined15:44
→ mikepack joined15:44
→ d3vic3 joined15:46
→ dwmw2_gone joined15:46
→ edude03 joined15:46
← barmaglot left15:47
→ kukks joined15:47
← kukks left15:47
→ kukks joined15:47
→ n123 joined15:47
← arvind_khadri left15:48
← kerrick left15:49
← posciak left15:49
← psoo left15:49
→ avpx joined15:49
← cakehero left15:49
← senny left15:49
← sonig left15:50
→ kerrick joined15:50
← dwmw2_gone left15:50
→ xtagon joined15:51
→ mostafaxx joined15:51
→ Goplat joined15:51
→ markgifford joined15:53
← jackh left15:53
→ cviebrock joined15:54
→ bgerber joined15:54
cviebrock any idea why "git push" would try to go to a remote I have set up, instead of my origin?15:55
← kerrick left15:55
canton7 because the current branch has tracking info set up?15:55
selckin because the remote is the tracking branch15:55
avpx Ding ding ding ding!15:55
→ simesy joined15:55
cviebrock ok … not sure how that happened, but can i switch it back? (sorry … bit of a git noob)15:56
← tatsuya_o left15:56
canton7 git branch -vv to see what tracking is set up15:57
push -u and branch --set-upstream to change it15:57
s/and/or15:57
cviebrock branch -vv gives: * master 34f839f [laravel/develop: ahead 4079] remove Fontdeck15:57
laravel is a remote, obv.15:57
← bartek_ left15:58
→ Syzygy__ joined15:58
canton7 which means that the local master is set up to track the remote branch laravel/develop15:58
→ bartek joined15:58
cviebrock ok … and it should be the local master tracks remote origin/master15:59
so, git branch --set-upstream origin/master ?15:59
canton7 git push -u origin master (if you want to push at the same time)15:59
→ fsvend joined15:59
canton7 the branch --set-upstream syntax is more unintuitive, and I can never remember it (but it's not that)15:59
'git branch --set-upsteam origin/master master' maybe? might be the other way around16:00
cviebrock i should never be pushing to laravel/develop … only pulling and merging16:00
lemme RTFM :)16:00
← simesy left16:00
canton7 man git-branch16:00
gitinfo the git-branch manpage is available at http://git-scm.com/docs/git-branch.html16:00
← Syzygy_ left16:01
→ codewrangler joined16:01
avpx cviebrock: Not trying to be a PITA here, but you mean "fetching and merging."16:02
cviebrock: "Pulling and merging" is redundant.16:02
← Per4u3e left16:02
cviebrock yes, of course … too quick typing :)16:02
canton7 cviebrock, oops it's the other way around: git branch --set-upstream master origin/master16:02
→ neurodrone joined16:02
← robotmay left16:02
cviebrock before i do that … http://pastie.org/415512716:03
→ Targen joined16:03
cviebrock that seems correct already, no?16:03
canton7 looks sensible16:04
← thiago left16:04
cviebrock i haven't set the upstream yet though16:04
to me, it looks like it's already correct :/16:04
canton7 that pastie didn't show tracking info16:04
→ dwmw2_gone joined16:04
→ an0maly joined16:05
adhawkins cviebrock: git branch -vv will show which branches are tracking16:05
← whitman left16:05
← mostafaxx left16:05
→ _iron joined16:05
cviebrock just the first one16:05
← kanzie left16:06
canton7 cviebrock, do you know what a) remote-tracking branches and b) trackinfo is/are ?16:06
cviebrock reload http://pastie.org/415512716:06
well, kinda16:06
← bgerber left16:07
← stodan left16:07
cviebrock i had this working fine until yesterday. I have my repo, and I added laravel as a remote (its a MVC framework) and would occasionally merge in changes from it16:07
canton7 remote-tracking branches (e.g. 'origin/master') are local records of the states of remote branches. tracking info is a bit of config which says what branch is merged in when you run 'git pull', which remote to push to, and some other stuff16:07
← rath left16:07
cviebrock but, by default, all my pushes and stuff would go to my repo (without me needing to specify "origin/master" or anything)16:07
canton7 there's a bit of config in your .git/config which says that the local 'master' is tracking the remote branch 'laravel/develop'16:08
→ jhardinee joined16:08
avpx When it comes down to it, I prefer to get fetch, inspect the changes from origin, and then git merge origin/master when I'm ready16:08
canton7 therefore, when you type 'git push' with the 'master' branch checked out, git tries to push to the remote 'laravel', because that's what the tracking info is telling it to do16:08
cviebrock [branch "master"]16:09
remote = laravel16:09
merge = refs/heads/develop16:09
canton7 that's the bit16:09
cviebrock (sorry for the paste flood)16:09
avpx 3 lines is hardly flood :p16:09
cviebrock :)16:09
← jn_ left16:09
canton7 now, you said you wanted your master branch to track origin/master, not laravel/develop16:09
so either change that config manually, or use branch --set-upstream to do it for you16:09
cviebrock so … I guess i could either remove those 3 lines, or "change remote = origin"16:09
→ redii joined16:09
→ adityamenon joined16:10
← leilaM left16:10
avpx I'd recommend using the right tool for the job, in this case branch --set-upstream16:10
canton7 you could remove those 3 lines. if you changed the remote, you'd also want to change the 'merge' line to 'refs/heads/master'16:10
bear in mind that if you remove those 3 lines, a 'git pull' (without any other arguments) will fail, as git won't know what to merge in16:10
cviebrock got ti16:11
it16:11
→ aantix joined16:11
cviebrock ok, so it now reads: [branch "master"]16:11
remote = origin16:11
merge = refs/heads/master16:12
canton7 looks right16:12
← erichynds left16:12
cviebrock thank you!16:12
avpx cviebrock: Solid16:12
cviebrock not sure how that got swapped16:12
i blame the GUI git tools :)16:12
→ alester joined16:12
avpx Bah!16:12
jast good heuristic16:12
canton7 if you did something like 'git checkout -b master origin/develop', or something else which created a local branch from a remote one, tracking is normally set up automatically16:12
cviebrock … which i never use :)16:12
anyway … thanks a lot everyone!16:13
avpx cviebrock: Enjoy16:14
Stay safe out there, etc16:14
← d3vic3 left16:14
cviebrock hehe16:14
← cviebrock left16:14
← oo_miguel left16:15
adhawkins → adhawkins-away16:15
adhawkins-away → adhawkins16:15
→ d3vic3 joined16:15
→ imrobert joined16:16
← Vile left16:16
← mishok13 left16:16
← thearthur left16:17
← fsvend left16:18
adhawkins → adhawkins-away16:18
← dwmw2_gone left16:18
← jargon- left16:18
→ Vile joined16:18
← LimeBlast left16:19
adhawkins-away → adhawkins16:19
→ _huoxito joined16:19
→ AlbireoX`Laptop joined16:20
← PigeonFriend left16:20
adhawkins → adhawkins-away16:20
← rtjure left16:20
← leeight_ left16:20
← harshpb left16:21
← rchavik left16:21
→ bgerber joined16:21
← ill1cit left16:21
→ ill1cit joined16:21
→ harshpb joined16:22
→ shreya joined16:22
→ dwmw2_gone joined16:22
→ john_anderson joined16:22
← totte left16:22
→ totte joined16:22
← totte left16:23
← eicto left16:23
→ totte joined16:23
← mabrand left16:24
→ the_cybersphinx joined16:25
← theCole left16:25
← jkremser left16:25
→ diegoviola joined16:27
← cybersphinx left16:27
→ Mowsh joined16:28
→ maletor joined16:29
→ etcetera joined16:30
← harshpb left16:30
→ theCole joined16:31
← HyperNerdV2 left16:31
← mklappstuhl left16:31
← cesc left16:31
→ fsvend joined16:31
← francisfish left16:32
→ semigroup joined16:32
← _md left16:33
→ spyro joined16:34
→ fuzzybunny69y joined16:35
← YesJustWolf left16:36
← juvenal left16:37
→ j264 joined16:37
→ john2496 joined16:37
→ raykyri joined16:38
→ mcluskydodallas joined16:38
← j264 left16:38
→ kerrick joined16:39
→ apok joined16:39
← revagomes left16:40
← _W_ left16:40
→ juvenal joined16:40
← angelsl left16:40
→ jargon- joined16:42
→ Araxia joined16:42
← ziro` left16:42
→ superdmp joined16:43
← jceb left16:44
superdmp I have done git remote add <name> <address of remote repo>; now I want to merge a branch from <name> into my current working directory16:44
I have obviously missed a step somewhere though16:44
I though git fetch would pull the references from <name>16:45
canton7 git fetch <remote>; git merge <remote>/<branch>16:45
EugeneKay !repro16:45
gitinfo Please paste (using https://gist.github.com/ or similar) a transcript (https://gist.github.com/2415442) of your terminal session, or at least explain exactly what you did that led up to the problem. This will help immensely with troubleshooting.16:45
FauxFaux superdmp: "git remote update" is the friendly version of fetch that fetches everything.16:45
← chrisf|work left16:45
superdmp ace, thank you all16:45
← achadwick left16:45
→ chrisf|work joined16:47
superdmp so, which was Torvalds's greater gift to the world, Linux or Git?16:47
→ SuperMiguel joined16:47
→ HyperNerdV2 joined16:48
← _Qman left16:48
← bgerber left16:48
→ alvaro_o joined16:48
avpx I think just gifting us with His Divine Presence is enough16:49
→ bgerber joined16:49
→ Sebo joined16:49
← HyperNerdV2 left16:50
← dvide left16:50
← ipalaus left16:50
EugeneKay I was going to make a dirty joke, but i can't find any recent photographs of his daughters.16:50
← alvaro_o left16:50
← Goplat left16:50
avpx That's reasonably creepy.16:50
→ mishok13 joined16:51
→ tyatpi joined16:52
→ plastics joined16:52
IcePic EugeneKay: browser refuses to show such content? (oh snap!)16:53
EugeneKay Nah, laziness. I gave up after the first page of google results of biographies of Linus proper.16:54
← k-man left16:55
shruggar is there a good way, given an object hash, to get a list of commits which introduced that object? (I also have the file's path, but I specifically want to know when this file became this object, if ever)16:56
→ Sky[x] joined16:57
avpx shruggar: I wrote a script that does something similar16:57
Marcux83|away → Marcux8316:57
avpx shruggar: The basic idea is to get a list of commits using rev-list, and then use ls-tree -r to find all blobs contained therein.16:58
← mishok13 left16:59
← kevwil left16:59
← gusnan left16:59
→ williamcotton joined17:00
← john_anderson left17:00
→ leilaM joined17:00
→ john_anderson joined17:00
← tvw left17:00
→ Tinti joined17:01
← mikaelj left17:02
→ kevwil joined17:02
← kevwil left17:02
← project2501a left17:03
→ himanshu_m786 joined17:03
← semigroup left17:03
← himanshu_m786 left17:03
→ brentdanley_ joined17:04
← brentdanley_ left17:04
→ mishok13 joined17:04
← ottoshmidt left17:05
→ jajang joined17:05
← brentdanley left17:05
← cedeon left17:05
→ phishy joined17:05
← buggaboo left17:06
← john_anderson left17:07
→ avandenhoven joined17:07
← matt_taws left17:07
→ john_anderson joined17:08
← fisted left17:08
→ lindenle_ joined17:09
→ eijk joined17:09
→ mwrosen joined17:11
→ _md joined17:12
← williamcotton left17:13
← jajang left17:14
← Xunie left17:15
→ fisted joined17:15
→ HyperNerdV2 joined17:17
← superdmp left17:18
← marhaban left17:18
← Cerise left17:18
← SuperMiguel left17:18
← brentratliff left17:19
→ mikaelj joined17:20
← apok left17:20
→ apok joined17:20
← shreya left17:21
→ Cerise joined17:21
→ R4p70r joined17:23
← marthinal left17:24
→ noscript joined17:24
→ Xizor joined17:24
→ superdmp joined17:25
← kpl left17:25
→ maxagaz joined17:25
maxagaz hi17:25
gitinfo maxagaz: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.17:25
maxagaz how to list the differents version of one file in a branch ?17:26
and how to compare them ?17:26
← mcluskydodallas left17:27
→ whitman joined17:27
→ cakehero joined17:28
→ mcluskydodallas joined17:28
← superdmp left17:30
→ thearthur joined17:30
← macUzer left17:31
← marcogmonteiro left17:31
← notmypudding left17:32
← bgerber left17:32
→ martinjlowm joined17:33
← Silex left17:33
← theCole left17:34
← Thanatermesis left17:34
→ bgerber joined17:35
← Sebo left17:35
avpx maxagaz: Try git rev-list --objects <branch> | grep <path>17:35
maxagaz: The git diff tool is useful for comparison17:36
← rgsteele left17:36
→ osmosis joined17:36
← n123 left17:36
← Targen left17:37
bremner isn't git log -- path easier?17:37
← cortana2 left17:37
← mikaelj left17:37
→ dfr|mac joined17:37
avpx bremner: Yeah, or I guess you could do that.17:38
:p17:38
maxagaz: Ignore me, listen to bremner. All hail bremner17:38
← solrize left17:39
→ z0a joined17:40
← kerrick left17:40
← mikepack left17:42
→ notmypudding joined17:44
→ aselder joined17:44
→ _W_ joined17:44
→ pdtpatrick joined17:45
→ tewecske joined17:45
← tewecske left17:45
← marc_kdab_ left17:46
→ marc_kdab_ joined17:46
→ YesJustWolf joined17:46
→ mikepack joined17:47
→ rendar joined17:47
→ Targen joined17:47
→ toabctl joined17:48
→ giallu_ joined17:49
← mikepack left17:51
maxagaz thanks17:52
← kenperkins left17:53
→ jajang joined17:53
→ abhishek__ joined17:54
→ Chillance joined17:55
→ dsiu joined17:55
→ mikaelj joined17:55
→ n123 joined17:56
→ ziro` joined17:56
← pdtpatrick left17:58
→ lacrymology joined17:58
→ kerrick joined17:58
→ pdtpatrick joined17:58
→ marthinal joined17:59
→ jstout24 joined17:59
← Davey left17:59
← bubbles10101 left18:00
← abhishek__ left18:00
→ mritz joined18:01
← dfr|mac left18:01
← kerrick left18:02
← Vile left18:02
← mritz left18:02
→ dfr|mac joined18:03
→ kerrick joined18:04
← dc5ala left18:04
→ j416 joined18:04
← rurufufuss left18:05
← Owner left18:06
→ kenperkins joined18:06
→ Owner joined18:06
← Sky[x] left18:10
← giallu_ left18:10
← randomubuntuguy left18:13
← juvenal left18:14
→ cesc joined18:15
→ linguini joined18:15
→ sq-one joined18:16
linguini I'm constantly committing to the wrong branch, and then realizing it a second after the commit. What is the easiest way to fix a commit on the wrong branch? git rebase?18:16
→ Escherial joined18:18
linguini Or git cherry-pick?18:18
Escherial hey, i did a pull which resulted in my branch being conflicted...is there an easy way to just take the remote's version over my own rather than manually merging each file?18:18
→ juvenal joined18:19
jast linguini: cherry-pick IMO18:19
EugeneKay Escherial - man git-merge, see "merge strategies". You probably want recursive, with the "theirs" option18:19
gitinfo Escherial: the git-merge manpage is available at http://git-scm.com/docs/git-merge.html18:19
Escherial i know you can specify theirs as a strategy to merge, but i don't know what branch i'm merging into my main one (er, for reference, i'm on local branch master, attempting to merge origin/master)18:19
linguini jast: Thanks.18:19
→ superdmp joined18:19
→ dreiss_ joined18:19
Escherial i'm kind of lost on the relationship between remote branches and local ones...merge won't let me merge anything when my branch is in a conflicted state, it seems18:19
RandalSchwartz am I missing anything in this list for git code review: crucible, gerrit, phabricator, code collaborator, barkeep, github enterprise, gitlab ?18:20
FauxFaux mailman.18:20
← fly9 left18:20
FauxFaux Trusting developers to ask for code review.18:20
jast pair programming18:20
RandalSchwartz other than mail-based solutions. :)18:20
→ nwp_ joined18:20
RandalSchwartz developers *will* ask if it's easy18:20
jast innovative real world-based solutions18:20
RandalSchwartz like crucible or github18:21
← Wooga left18:21
adhawkins-away → adhawkins18:21
← ben_alman left18:21
FauxFaux Faxing patches to each other.18:21
avpx IP over avian carrier18:21
EugeneKay Escherial - !fixup will talk you through getting back to a sane state18:21
gitinfo Escherial: 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, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!18:21
RandalSchwartz print the patch on paper. take PDF of it on a wooden table. email PDF.18:21
FauxFaux I'm sorry, I only had a black and white printer.18:22
EugeneKay Escherial - a remote branch is a local pointer to the state of the branch on the remote repo as it sits, nothing more.18:22
Escherial - the point of this is that you can create a merge commit with the remote's current state as a parent, so you can push cleanly(without a -f)18:22
wereHamster !deploy18:23
gitinfo Git is not a deployment tool, but you can build one around it for simple environments. http://sitaramc.github.com/the-list-and-irc/deploy.html18:23
→ eletuchy joined18:23
RandalSchwartz !review18:23
gitinfo [!codereview] Gerrit is a code review tool for git. See http://code.google.com/p/gerrit/ You may want to enforce certain workflow to prevent people bypassing review. gitolite (and a branching workflow) or a integrator repository (distributed workflow) would be good for this.18:23
RandalSchwartz hmm. knew that already18:23
← superdmp left18:24
← hyperair left18:25
→ fly9 joined18:25
← dfr|mac left18:26
→ ben_alman joined18:28
→ mikepack joined18:30
→ mmc2 joined18:32
← ajpiano left18:34
← deebo left18:34
← d0k left18:35
→ ajpiano joined18:37
← mikepack left18:37
← nwp_ left18:38
cmn there isn't really much more for code review18:38
← Element9 left18:38
→ d0k joined18:39
cmn you can use github/gitorious/gitlab with pull requests for review, but you can't really enforce that much18:39
← akiress left18:40
← AeroNotix left18:41
→ mrbtie joined18:41
→ williamcotton joined18:42
RandalSchwartz yeah - I have those... but that's not the workflow that works here.18:45
→ revagomes joined18:45
→ semigroup joined18:45
← _iron left18:46
→ jceb joined18:46
← mmc2 left18:46
→ pandeiro joined18:47
← fly9 left18:47
→ fly9 joined18:47
→ Silex joined18:47
← mrbtie left18:48
→ pantsman joined18:48
→ _iron joined18:48
→ marcogmonteiro joined18:50
→ flyz_ joined18:52
← phishy left18:53
← shwaiil left18:53
→ Sky[x] joined18:54
← jjbohn left18:55
→ Vile joined18:55
← lacrymology left18:56
← _md left18:57
→ drizzd_ joined18:58
→ lacrymology joined18:58
→ RaceCondition joined19:00
← edude03 left19:00
→ eicto joined19:00
← drizzd left19:01
→ mmc2 joined19:01
← pandeiro left19:03
→ harshpb joined19:03
EugeneKay You can, in combination with an "out of band" review system, use gitolite to do review19:04
← booginga left19:04
← ph^ left19:05
EugeneKay Have people push to a user- or feature- specific branch(say, the same ID# as from the review system), and then somebody with the correct permissions can pull, merge, and push it back into your main branch19:05
→ venom00ut joined19:06
← notmypudding left19:07
venom00ut hi, I'm trying to diff a gzip compressed XML file (xournal's .xoj)19:07
I tought I could use an external diff script which decompresses it19:08
→ xref_ joined19:08
venom00ut but I'd also like to thread the file as text file, so I can do merge easily19:08
any smart way to do this?19:08
→ mikepack joined19:08
canton7 venom00ut there's a lot of cool stuff in man gitattributes to do with custom diff and merge drivers19:09
gitinfo the gitattributes manpage is available at http://git-scm.com/docs/gitattributes.html19:09
← BlindOracle left19:09
← mcluskydodallas left19:10
← xref left19:10
xref_ → xref19:10
← mikepack left19:10
venom00ut canton7, do you think that using an uncompressing diff script and a "uncompress, apply patch and recompress" merge script would work?19:11
← YesJustWolf left19:11
canton7 venom00ut, that was my thought -- I've not tried anything similar, though, so I've got no idea how much success you'll have with it19:11
→ btree joined19:12
venom00ut I'll give a try19:12
← mrtazz left19:13
canton7 also maybe smudge/clean filters? I guess the encryption stuff in the man page is not *too* dissimilar to what you're doing19:13
→ kevwil joined19:13
← netzmonster left19:13
→ superdmp joined19:14
venom00ut canton7, but a filter works upon checkout/checkin, and I'd like to have them in compressed form on the disk19:15
→ mikepack joined19:16
← Escherial left19:16
canton7 venom00ut, fair play19:17
→ hyperair joined19:17
canton7 there's also the option of having them in compressed form when checked out, but check in the uncompressed versions? That would be a bit kinder on git's delta-compression algorithm too I guess19:17
but again, might not be what you're after. I'm just throwing ideas around19:17
← shruggar left19:18
← superdmp left19:18
← mikepack left19:18
← harshpb left19:18
← federic left19:19
→ mikepack joined19:19
→ zackiv31 joined19:19
→ dr_diamond joined19:19
← dr_diamond left19:19
→ dr_diamond joined19:19
venom00ut canton7, on the tree they should always be compressed, otherwise Xournal won't be able to open them19:19
← Lgb left19:20
← palm_vue left19:20
← tommyvyo_ left19:20
← mikepack left19:20
canton7 I was assuming xjournal would be reading them from the working copy, rather than the tree19:20
but hey -- I do think the diff/merge drivers would be a better choice, if they'll work19:20
venom00ut sorry, I meant the working copy19:20
I'm not an expert with the git terminology19:21
canton7 so I suggested having them compressed in the working copy but uncompressed in the tree. Not that I massively like the idea...19:21
→ Lgb joined19:21
→ palm_vue joined19:21
→ tommyvyo_ joined19:22
← cesc left19:22
← jason237 left19:22
→ cesc joined19:22
← Jackneill left19:22
← dr_lepper left19:23
→ Superlokkus joined19:23
→ marhaban joined19:23
← gavin_huang left19:23
← marhaban left19:24
→ marhaban joined19:24
→ mikepack joined19:25
← erasmospunk left19:26
← mikepack left19:27
← venom00ut left19:27
← jhardinee left19:28
→ imachuchu joined19:28
← bartman`` left19:29
→ mikepack joined19:30
← mikepack left19:30
→ rtjure joined19:31
→ codeM0nK_ joined19:31
→ ivan12348 joined19:31
→ ivan12349 joined19:32
← ivan12346 left19:32
← codeM0nK3Y left19:33
codeM0nK_ → codeM0nK3Y19:33
← ivan12345 left19:33
← pantsman left19:34
← ivan12347 left19:34
→ bartman`` joined19:35
← diegoviola left19:36
→ bartek_ joined19:36
← bartek left19:36
→ ivan12345 joined19:37
← opdynamiks left19:37
← Yuuhi left19:38
→ ColKurtz joined19:38
→ nielsbusch joined19:38
fly9 anyone use gitlab with ldap? need to turn off account creation + notification on a new ldap login19:39
← cyphase left19:40
→ cyphase joined19:41
→ btanaka joined19:43
→ afuentes joined19:43
← enmand left19:44
← marthinal left19:44
← bgerber left19:44
→ mikepack joined19:45
→ mrbtie joined19:47
→ Yuuhi joined19:51
→ mcluskydodallas joined19:52
→ nadirvardar joined19:52
→ i42n joined19:53
← mikepack left19:54
← toabctl left19:55
← hyperair left19:55
→ bgerber joined19:55
← Natureshadow left19:56
→ mikepack joined19:57
→ Natureshadow joined19:57
← blameless left19:57
← plastics left19:59
→ blameless joined19:59
→ Wooga joined20:01
→ hyperair joined20:01
→ mikew3 joined20:02
gitinfo set mode: +v20:03
← berserkr left20:03
→ cedeon joined20:04
← semigroup left20:05
→ Axsuul joined20:05
← Natureshadow left20:06
→ semigroup joined20:07
→ Natureshadow joined20:07
→ mxweas_ joined20:08
→ yoklov joined20:08
→ offby1 joined20:08
← offby1 left20:08
→ offby1 joined20:08
← Natureshadow left20:09
← mikew3 left20:09
← hyperair left20:09
← _huoxito left20:11
→ marthinal joined20:12
→ Natureshadow joined20:13
← cedeon left20:13
← bgerber left20:15
→ dr_lepper joined20:16
→ hyperair joined20:17
→ diegoviola joined20:17
→ pantsman joined20:18
← pantsman left20:18
→ pantsman joined20:18
← hyperair left20:18
→ shruggar joined20:18
← JPT left20:19
→ bgerber joined20:20
← heptadecagram left20:20
adhawkins → adhawkins-away20:20
← fsvend left20:22
→ JPT joined20:22
← Ryback_ left20:23
← jcromartie left20:24
→ hyperair joined20:24
→ jcromartie joined20:24
→ fsvend joined20:25
← mxweas_ left20:27
← Mowsh left20:28
← nielsbusch left20:29
→ randomubuntuguy joined20:29
← AaronMT left20:29
← randomubuntuguy left20:29
→ randomubuntuguy joined20:30
← ToBeFree left20:32
→ benjo2 joined20:32
→ heptadecagram joined20:33
→ tigrang joined20:33
← heptadecagram left20:33
→ nutron joined20:34
← nadirvardar left20:35
← duckxx left20:36
→ duckxx joined20:36
→ nadirvardar joined20:37
← jceb left20:38
← shruggar left20:39
→ Shusshu joined20:40
← Shusshu left20:41
← jaseemabid left20:41
← jajang left20:41
→ flazz joined20:42
← semigroup left20:43
flazz (on a topic branch T) git diff master, is there a way to export this to a visual diff tool to let me accept or reject differences from master? kinda like cherry picking but visually?20:44
← mishok13 left20:45
→ mishok13 joined20:45
canton7 flazz, checkout -p will probably do it20:45
← Natureshadow left20:47
← fsvend left20:48
→ Natureshadow joined20:48
→ minism joined20:48
← minism left20:48
→ delight joined20:48
→ bauruine joined20:49
delight i was wondering is there a difference between "git rm --cached FILENAME" and "git reset HEAD FILENAME" ?20:49
← Xizor left20:49
← martinjlowm left20:50
← Natureshadow left20:50
→ Natureshadow joined20:51
← d0k left20:51
cmn they're only equivalent if the file hasn't been committed yet20:51
→ hellopat_ joined20:51
← delight left20:51
← mikepack left20:51
→ snap_ joined20:51
→ delight joined20:51
← xref left20:52
← mishok13 left20:52
cmn or more exactly, if the file didn't exist in the last commit of the current branch20:52
→ xref joined20:52
snap_ hey, is there any way I can have 2 directories with identical repositories but different working branches; so I am able ot work on 2 branches at the same time and commit to one and see the changes on both without having to sync back and forth?20:53
← andref_dias left20:53
snap_ or some altogether easier way to do that?20:53
rendar what are those files in the "t" directory in the git's source tree ?20:53
← dsiu left20:53
← tyll left20:54
→ oops joined20:54
delight snap_: just clone your repo locally .. done20:54
oops What is the wildcard character to select .foo.swp, .bar.swp, .swp but not .barswp?20:54
← hellopat_ left20:54
cmn rendar: tests20:54
rendar ok20:54
delight cmn: what exactly is the difference if the file pre-existed in history ?20:54
← hellopat left20:54
→ mikepack joined20:55
snap_ delight: wouldnt i still have to pull from one to the other after a commit and from upstream on both?20:55
→ d0k joined20:55
← noisebleed left20:56
→ mishok13 joined20:57
cmn delight: reset will set the file's entry in the index to be that in HEAD20:57
which what happens in either case, but if the file isn't in HEAD, it gets deleted20:57
see man git reset for details on what reset is20:57
gitinfo the git-reset manpage is available at http://git-scm.com/docs/git-reset.html20:57
← kevwil left20:58
canton7 oops, *.swp maybe?20:59
→ raatiniemi joined21:00
oops canton7, that doesn't select any of them.21:00
canton7 snap_, what exactly are you after? even if there was some way to magically sync two repos, they've got different branches checked out -- committing in one wouuldn't directly affect the other anyway21:00
← mikepack left21:00
delight cmn: I'll have to double check that … but last time i used "git rm --cached FILE" I think it took it of only from the index … and the file in index was new … well I'll better double check that before averring something21:00
oops canton7, because the filenames start with a dot. there is nothing before the dot.21:01
cmn that's what rm --cached does21:01
→ psoo joined21:01
canton7 oops, well it's working fine for me...21:02
hang on, I'll do a pastie21:02
delight cmn: but reset HEAD does the same … all right I'll stop nagging :-D … I'll read the manual some more … thnx so far21:02
oops canton7, ok21:02
→ axl_ joined21:02
cmn delight: it will sometimes do it; I've already described the situation where it behaves that way21:02
→ Davey joined21:02
→ nielsbusch joined21:02
→ kevwil joined21:03
canton7 oops, http://pastie.org/415653021:03
← sq-one left21:04
→ aspotashev joined21:05
→ eletuchy_ joined21:06
→ mklappstuhl joined21:06
oops canton7, yes, made it work. I was confused between the shell's wildcard expansion and .gitignore's. They don't seem to match.21:06
canton7, are you able to select the same files using ls *.swp or echo *.swp?21:07
how can I know the rules for wildcard expansion of git?21:07
→ Drakken joined21:07
← cesc left21:07
← mishok13 left21:07
canton7 oops, no, they differ. I believe .gitignore uses fnmatch syntax21:07
EugeneKay man gitignore ;-)21:08
gitinfo the gitignore manpage is available at http://git-scm.com/docs/gitignore.html21:08
canton7 (presumably without FNM_PERIOD set)21:08
Drakken I'm having trouble pushing some commits onto a new github repo.21:09
← chrisf|work left21:09
Drakken If anyone here is on #github, I posted it there.21:09
EugeneKay !repro !work21:09
gitinfo Please paste (using https://gist.github.com/ or similar) a transcript (https://gist.github.com/2415442) of your terminal session, or at least explain exactly what you did that led up to the problem. This will help immensely with troubleshooting.21:09
[!doesntwork] Sorry to hear it doesn't work. What happened? Did it fall asleep on your couch, with your sister? Eat your homework? What did you want it to do? What happened instead? Please be specific!21:09
→ chrisf|work joined21:10
← eletuchy left21:10
Drakken EugeneKay is that for me?21:10
← noscript left21:10
oops ok so it first tries it as shell glob. If nothing found, then it tries fnmatch()21:10
Drakken output is here: http://pastebin.com/fiNW3gRH21:10
EugeneKay I make the assumption that people are smart enough to know when I'm speaking to them based upon the context. If this assumption is wrong for you, please let me know and I will alter my habits for you ;-)21:11
You appear to be using HTTPS:// to speak to Github. Please use !ssh instead.21:11
gitinfo Please use SSH to talk to remote repos. http:// and git:// are completely unsecured. You can try to configure "smart" https:// access, but it is usually much easier to just stick with ssh://21:11
EugeneKay Yes, github says to use HTTPS. They're wrong.21:11
cmn for them using username/password has less support burden than ssh21:12
EugeneKay The underlying problem is that your system's CA store(the one libcurl uses) does not contain the CA certificate that github.com is signed by. This is because SSL PKI is crappy.21:12
So, use SSH and enjoy. Github even has a page on how to generate a key, which you should google up and read.21:13
→ thiago joined21:13
← lotus left21:14
← linduxed left21:15
← psoo left21:16
← eletuchy_ left21:17
→ eletuchy joined21:17
← Riviera left21:17
← Owner left21:18
← marthinal left21:18
→ Owner joined21:18
← lacrymology left21:19
← i42n left21:19
← j416 left21:19
→ shruggar joined21:21
← Drakken left21:22
→ AeroNotix joined21:23
delight cmn: alright I understood the difference … thnx again … basically git rm --cached prepares a "delete-commit" for a file that was committed earlier21:24
← benjo2 left21:25
snap_ canton7: i just want to be able to have 2 arbitrary branches checkedout from the same repo and be able to commit to either without having to sync between them21:25
← mmc2 left21:27
← delight left21:27
← hakunin left21:27
← hoschi left21:27
→ hakunin joined21:28
→ linduxed joined21:28
→ kahmalo joined21:29
← mwrosen left21:29
→ gusnan joined21:29
→ hakunin_ joined21:30
gitinfo set mode: +v21:30
kahmalo what's a handy way to get git diff against HEAD but list files in random order, rather than alphabetical?21:30
← bartek_ left21:30
← hakunin left21:30
← Dave^|| left21:31
EugeneKay man sort; see the -R flag.21:32
A better question is *why* you need this. !xy ?21:33
gitinfo [!doinitrong] It sounds like you're approaching this problem in the wrong manner. Let's step back for a minute - What are you actually trying to achieve? Why are you trying to do it this way?21:33
← alester left21:33
kahmalo the scenario is that I run sed on a large number of files and need to verify that the result is somewhat sane; random sampling should be more reliable than looking at the first files alphabetically.21:33
as there is correlation between file names and contents21:33
EugeneKay Interesting problem. Take the output of git-status(in porcelain format), sort -R that, then git-diff the individual files, using !dashes21:34
gitinfo Use a double-dash(--) to separate refs from paths, especially when dealing with ambiguous file names. Ex: git checkout master -- origin (check out the file "origin" from branch "master")21:34
← marc_kdab_ left21:35
kahmalo anyway, I'm at 51% of the whole diff, so I guess I'll just read to the end now21:36
← sh0ne left21:36
← Anxi80 left21:37
→ basicallyu joined21:37
→ semigroup joined21:37
→ Drakken joined21:37
← andrewhl left21:37
kahmalo EugeneKay: what do you mean with porcelain format? I don't see such an option in the git-status or git-commit man pages.21:38
actually I'd have thought that git-ls-files was the appropriate program21:39
EugeneKay man git-status; ctrl-f for "porcelain"21:39
gitinfo the git-status manpage is available at http://git-scm.com/docs/git-status.html21:39
← aspotashev left21:39
→ aspotashev joined21:39
EugeneKay I presume you're diffing the index against HEAD; -status gives you a list of files staged relative to HEAD21:39
gitinfo set mode: +v21:40
basicallyu hi, how would I go about doing this: first time clone project it gives me the .classpath file, then any modifications on it will not be seen as unstaged and will never be commited ? but when I really want to commit it then i should can do it21:40
→ giallu joined21:40
← giallu left21:40
→ giallu joined21:40
← twisted` left21:40
kahmalo hm, I seem to have stale man pages in my home directory.21:41
jrdnn basicallyu: git add .classpath git commit21:41
→ Anxi80 joined21:41
basicallyu thing is, if it's already in index then .gitignore will have no effect when .classpath file is detected as modified21:41
→ twisted` joined21:41
basicallyu jrdnn: would be cool if when i try to commit next it wouldn't detect .classpath as modified at all, but still .classpath should exist in the repo21:41
jrdnn basicallyu: git rm --cached to remove from index but leave i t in working die21:42
basicallyu: If you want to leave it in index and working dir and just not update index to equal working dir, just don't' git add that file before you commit and don't use git commit -a21:42
basicallyu yeah that is what i want, but how can I ever have it in the repository if i never add it?21:43
← EricKamsky left21:43
← dzonder left21:43
basicallyu i still need the rm --cached you said above right?21:43
i'm looking into that :)21:43
jrdnn basicallyu: If it's indexed and you want to unindex it and you don't want it in the repository21:43
basicallyu it's indexed but i do want it in the repository21:44
and i want it on local too, but i just don't want the modifs I do in it to get detected when I do git commit (actually via Eclipse ie. egit)21:44
← kaen left21:44
jrdnn basicallyu: Just don't use git add on it then and don't use git commit -a. Then you can have the one in the repository unmodified, and the one in your working directory.21:45
basicallyu: Or do git branch -b LocalFile git add theFile git commit21:45
basicallyu and if I modify any other files but I want them commited21:45
→ _md joined21:45
basicallyu but can it maybe not detect at all that .classpath was modified ?21:45
jrdnn basicallyu: If you make a local branch for it and put all the changes there and don't push from that branch you can version control that file without affecting the other repository21:45
← etcetera left21:46
→ psoo joined21:46
basicallyu hmm interesting,21:46
jrdnn basicallyu: git checkout -b MyClasspath git add .classpath git commit -m "Local classpath" git checkout originalBranch21:46
← whitman left21:46
← maxagaz left21:46
→ acl_ joined21:46
basicallyu ok that i understand, but trying to figure out what happens when im on original branch and i modify the .classpath21:47
or would the original branch never actually have the file?21:47
← Drakken left21:47
← kevwil left21:48
← axl_ left21:48
→ lledet joined21:48
→ benjo2 joined21:48
basicallyu Here's what I actually want to do: first time user clones repo, it has a .classpath in the repo; but then if he modifies it, it shouldn't be detected by git commit -a ie. it should be ignored so that to prevent users from accidentally commiting a .classpath of their own21:48
jrdnn basicallyu: After you add it to MyClasspath, git checkout originalbranch git rm .myclasspath git commit -m "Don't track .myclasspath in this branch" echo '.myclasspath' >> .gitignore21:49
basicallyu and also when they do try to update, if they already have the .classpath file, then it should auto be ignored21:49
ok that would work but it'd mean that i'd have to pull both branches to get the .classhpath in my work dir?21:50
→ etcetera joined21:50
jrdnn basicallyu: Both branches yes since it wouldn't be tracked in the main one that people push21:50
← macabre left21:51
canton7 snap_, you can do that no problem. if you're committing to them independently, you don't need them to be synchronised all the time, no?21:51
→ dfr|mac joined21:51
basicallyu if the file exists only in the myclasspath branch would it show locally if i'm on the original branch ?21:51
true that canton721:51
trying to figure out if I'd still have the file on work dir while i'm on original branch21:52
but tipically this is a great idea jrdnn21:52
ty*21:52
→ superdmp joined21:52
jrdnn basicallyu: It wouldn't. Sorry. I know of no way to have a file in a branch and you get the file when you pull that branch and you modify that file in that branch and not have git commit -a add it. So I guess just only manually git add what you want to change21:53
← marcogmonteiro left21:53
→ madewokherd joined21:54
← sebrock left21:55
basicallyu yeah :/ unfortunatelly when modifying .classpath locally and say a few other files, it can just be that someone will forget to deselet .classpath from commiting (imagine in a gui, egit) but also this deselect would have to be done on every commit21:55
RandalSchwartz any experience with indefero here?21:55
basicallyu jrdnn thanks a lot for your help and time, I really do appreciate it21:55
jrdnn basicallyu: You're welcome. Sorry for no complete solution21:55
← nipnop left21:55
→ grantbdev joined21:55
basicallyu it's all good21:55
→ youngin joined21:56
← blacktulip left21:56
basicallyu oh snap, i think i just googled it :O and there's this solution:21:56
git update-index --assume-unchanged <file>21:56
← youngin left21:56
basicallyu If you wanna start tracking changes again run the following command: git update-index --no-assume-unchanged <file>21:57
→ mikkelb joined21:57
→ kaen joined21:57
basicallyu do you think this would work? sounds ok21:57
mikkelb Hey21:57
gitinfo mikkelb: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.21:57
mikkelb How do I view the date of a repo's last commit?21:57
sorry its probably a stupid question21:57
← TheHaven left21:58
← rtjure left21:58
← davidjdagino left21:58
canton7 mikkelb, git log --all21:58
(is one way)21:58
basicallyu aweome canton7 thx21:58
← chrisgeorge left21:59
← juanfer left22:00
→ chrisgeorge joined22:00
→ cagedwisdom joined22:00
Marcux83 → Marcux83|away22:01
→ Pupeno_W_ joined22:02
← Pupeno_W left22:02
Pupeno_W_ → Pupeno_W22:02
← Mkaysi left22:02
basicallyu jrdnn: that is pretty epic with assume unchanged, it works, now i only need a way to make that a default such that when cloning the repo, it would act as if the command git update-index --assume-unchanged .classpath was already executed; any ideas if this can be done?22:02
jrdnn basicallyu: I think that's what you're looking for22:02
mikkelb canton7: any way to get just the date?22:03
canton7: preferably YYYY-MM-DD22:03
→ masterkaen joined22:03
basicallyu yes but I'd need a way to automatically have that set for the .classpath when the user clones the repo, so that he doesn't have to execute that command (and 3 more for more files) right after cloning22:04
← kermit left22:04
← Zenmaker_ left22:04
jrdnn basicallyu: If you set assume-unchanged and push it, whenever somebody pulls it, it should be assume-unchanged already22:04
← kpreid left22:04
→ Zenmaker_ joined22:04
canton7 mikkelb, git log --date=short --format="%cd" -122:04
→ Mkaysi_ joined22:05
mikkelb canton7: sweeeeeeeeeet22:05
thanks man :D22:05
← imachuchu left22:05
→ kermit joined22:05
→ mikepack joined22:05
canton7 mikkelb, no problem. man git-log next time ;)22:05
gitinfo mikkelb: the git-log manpage is available at http://git-scm.com/docs/git-log.html22:05
mikkelb :D22:05
Hunner Why is there no man page for 'git show'?22:05
canton7 man git-show22:05
gitinfo the git-show manpage is available at http://git-scm.com/docs/git-show.html22:05
Hunner (or maybe my system is broken)22:05
← afuentes left22:06
→ Sebo joined22:06
Hunner Yep, my system is broken :P22:06
← twisted` left22:06
canton7 mikkelb, stick --all in there if you want the last commit across all branches22:06
→ ziggles joined22:06
basicallyu jrdnn: wow that would be epic then, i'll test22:07
→ neakretsam joined22:07
← markgifford left22:07
← mikepack left22:07
→ kcj joined22:07
← kaen left22:07
basicallyu jrdnn: did you mean commit after i set assume unchanged? and then push ? thing is, if i set that i cannot commit it(?)22:08
mikkelb canton7: i definitely want that yeah22:08
canton7: real quick though, is there a way to apply this same command but to the second last commit?22:09
RandalSchwartz anyone have any experience here with http://crew-cr.org/ ?22:09
basicallyu jrdnn: oh i see, If you git add the file directly, it will be added to the index.22:09
← nielsbusch left22:10
← shruggar left22:10
jrdnn basicallyu: If you git add the file it works22:10
basicallyu: Sorry didn't see last message22:10
basicallyu jrdnn: apparently not from withing egit, maybe works on cmd line, testing22:10
← masterkaen left22:10
← Destos left22:11
→ enmand joined22:11
← mattp_ left22:11
→ mattp_ joined22:11
← tubbo`cloud left22:11
← mitsuhiko left22:11
← MetaCosm left22:11
← claudiu__ left22:11
← desaila left22:11
← jmcantrell left22:11
→ Zenmaker__ joined22:11
→ MetaCosm joined22:11
jrdnn basicallyu: git add the file, update-index --assume-unchanged the file, git commit, and git push should work, don't know about GUI though22:11
→ mitsuhiko_ joined22:12
← neakretsam left22:12
← semigroup left22:12
→ berserkr joined22:12
basicallyu jrdnn: i see what you mean but git add there expects the file wasn't previously already added, which is my case22:12
← dv310p3r left22:12
→ sh0ne joined22:12
canton7 mikkelb, not directly for all branches, no. you can do it for the current branch by sticking 'HEAD^' on the end, or --all and -2, and just pick the second line for all branches22:13
→ jmcantrell joined22:13
→ mikepack joined22:13
mikkelb canton7: cool yeah thats a good idea22:13
→ tubbo`cloud joined22:13
jrdnn basicallyu: git add --force file22:14
← beatak left22:14
→ AaronMT joined22:15
← AaronMT left22:15
← Zenmaker_ left22:15
← dfr|mac left22:15
← neurodrone left22:15
→ AaronMT joined22:15
← AaronMT left22:15
→ shruggar joined22:15
→ AaronMT joined22:16
← AaronMT left22:16
← arietis left22:16
→ AaronMT joined22:17
← AaronMT left22:17
→ apok_ joined22:17
ij Give me something about dealing with push -f on the other end.22:17
→ AaronMT joined22:17
← AaronMT left22:17
mikkelb sorry last question y'all!22:17
← shruggar left22:18
→ claudiu__ joined22:18
→ AaronMT joined22:18
← AaronMT left22:18
← juvenal left22:18
→ desaila joined22:18
→ bryanmv`away joined22:18
basicallyu jrdnn: awesome, seems that's all i need; just have to test now with clone ;)22:18
mikkelb is there a way to, before running a "git remote update", to check if there actually are any remote commits to update with?22:18
jrdnn crosses fingers22:18
→ AaronMT joined22:19
← AaronMT left22:19
canton7 mikkelb, you *could* use ls-remote. why not just fetch though? :)22:19
← Sebo left22:19
← flyz_ left22:19
→ AaronMT joined22:19
← _md left22:20
mikkelb canton7: hold up ill show u, its for this script to pull all repos in this huge collection of repos22:20
← apok left22:20
apok_ → apok22:20
basicallyu jrdnn: ok wait apparently lol I am unable to git add --force file if it's assume unchanged, or I am missing something22:21
canton7 doesn't like the words "pull" and "script" in the same sentence22:21
← superdmp left22:21
← TheJH left22:21
← tudorstudio left22:22
← rendar left22:22
← meLon left22:22
← john2496 left22:22
← revagomes left22:23
← n123 left22:24
→ ashphy joined22:24
← hyperair left22:24
← airborn left22:25
mikkelb hehe22:25
→ bubbles10101 joined22:25
← giallu left22:29
← AeroNotix left22:30
← mikkelb left22:30
→ hyperair joined22:31
← cakehero left22:32
→ dabd joined22:34
← QaDeS left22:34
→ twisted` joined22:34
← sh0ne left22:34
← Jon30 left22:35
→ theCole joined22:35
→ yaymukund joined22:35
← intransit left22:35
yaymukund can I fast forward a branch without checking it out?22:35
← twisted` left22:36
EugeneKay Sure. Overwrite the branch pointer with the ID of the revision you want it to be at.22:36
→ flaguy48 joined22:36
EugeneKay This is probably not the best of ideas, but hey....22:36
canton7 (git branch -f branch origin/branch for example. it won't ensure ff-only merges though -- you can tell it to move the branch head wherever)22:37
cmn if it's really a fast-forward, it shouldn't cause problems22:37
yaymukund haha, fair enough22:37
EugeneKay If youre just concerned about losing your current branch you can make a stash then checkout & merge as you like, switching back and popping when finished.22:37
→ twisted` joined22:37
canton7 and remember that if e.g. master isn't up to date, you can do a lot of work from origin/master instead, without having to go near master22:38
← twisted` left22:38
EugeneKay To do the overwrite, something like: echo $(git rev-parse origin/otherbranchname) > .git/refs/heads/otherbranchname22:38
Actually, you can drop the () and the echo and the > altogether..... I think I need a beer.22:39
→ twisted` joined22:39
← toothr left22:40
pragma- but don't drop the $? wat?22:40
→ toothrot joined22:40
EugeneKay $()*22:40
← d0k left22:40
EugeneKay My four key is sticky22:40
← yacc left22:40
pragma- I WONDER WHY THAT IS22:41
cmn erm... git branch -f otherbranchname origin/otherbranchname is much much easier to remember and type22:41
← grantbdev left22:41
EugeneKay Yeah, but it doesn't sound as scary.22:41
cmn ah, gotcha22:41
← linduxed left22:41
EugeneKay "You CAN do this, but it's a plumbing command, and could cause !weasels..... or you could just do it the right way."22:42
gitinfo The consequences of this proposal are not well-defined. A band of furious weasels may infest your undergarments, or it might work just fine. You should !backup then !tryit and let us know what happens.22:42
← AaronMT left22:42
→ n123 joined22:43
← eijk left22:43
→ wolfric joined22:44
← puneetgoyal left22:44
← mrbtie left22:44
→ yacc joined22:45
basicallyu jrdnn: wow get this, to actually force the assume unchanged file to be seen as changed i just execute git update-index --really-refresh However, this removes the assume unchanged flag apparently :/22:45
← thomas_b_ left22:45
milki interesting side effect22:45
jrdnn basicallyu: =(22:46
EugeneKay Which, the weasels?22:46
← pantsman left22:47
→ Lulzon joined22:49
← bubbles10101 left22:50
frogonwheels EugeneKay: you shouldn't update refs directly like that. you should use git update-ref instead. git update-ref refs/heads/otherbranchname origin/otherbranchname22:50
EugeneKay: but of course not on the your checked out branch. Who was it that you were helping?22:50
EugeneKay points to yaymukund, mentions the weasels again22:51
yaymukund I just stashed22:51
EugeneKay Good. ;-)22:51
yaymukund my git fu is not strong enough, and I know better :)22:51
← basicallyu left22:51
→ linduxed joined22:52
yaymukund but thank you :)22:52
EugeneKay I didn't know about update-ref, but I doubt fiddling with un-checked-out branches like that is a good idea, anyway. Best to use -merge so you KNOW it's a ff22:52
frogonwheels EugeneKay: another cool thing about using update-ref is that it actually updates reflog as well :) I often will give a -m "some comment" arg to update-ref whcih goes in the reflog comment22:52
EugeneKay Nifty.22:52
→ sirdancealot joined22:53
← yacc left22:53
EugeneKay If I'm gonna screw up .git/ I like to do it really, really good.22:53
frogonwheels EugeneKay: Meh, it's safe enough if you know it's a fast-forward. I pretty much have gitk open all the time anyway22:53
EugeneKay Nothing pisses it off quite like overwriting files blindly22:53
frogonwheels EugeneKay: the other thing about update-ref, is that it works in packed refs as well22:53
EugeneKay nods22:53
← aspotashev left22:54
← TeckniX left22:54
frogonwheels EugeneKay: (ie those that have been moved into .git/info/refs file22:54
EugeneKay: oh, worth mentioning, update-ref doesn't accept shortcuts in the first param. so you must have refs/heads/blah or refs/tags/whatever :|22:55
cmn those aren't packed refs22:55
.git/info/refs is for the dumb http protocol22:55
and update-ref really shouldn't change it22:56
← Superlokkus left22:56
frogonwheels cmn: woops. Well just as well I don't rely on that information and use update-ref to mess with them :)22:56
← Jiyuu|Srv left22:57
→ Lord_DeathMatch joined22:58
frogonwheels cmn: but thankyou, (actually one of the things I like about an active and friendly channel like #git is that you usually have oversite to keep you from misleading people )22:58
→ yacc joined22:58
← fly9 left22:59
← chrisgeorge left22:59
→ fly9 joined23:00
← john_anderson left23:01
← diegoviola left23:01
→ Jiyuu|Srv joined23:02
← RaceCondition left23:02
→ RaceCondition joined23:03
← yacc left23:03
← n123 left23:03
← randomubuntuguy left23:04
→ edude03 joined23:04
← eletuchy left23:04
← sdeboey left23:04
← thiago left23:05
→ simesy joined23:06
← berserkr left23:07
→ mikkelb joined23:10
mikkelb canton7: soon got my script ready for you23:11
real quick though - its been a while since i dabbled with this script23:11
← Syzygy__ left23:11
mikkelb (basically i have a huge collection of various github repos i store for fun and learning)23:12
any idea why i'd put git reset --hard @{u} after git remote update?23:12
tommyvyo_ → tommyvyo23:13
frogonwheels mikkelb: that's a forced fast-forwarding of the branch. Dangerous ina script i'd a thought.23:13
mikkelb canton7: heres what im working on though: https://gist.github.com/299733623:13
frogonwheels mikkelb: but fine enough if you're not doing any changes23:14
mikkelb frogonwheels: i remember having some issues with repos that wouldnt git remote pull even though i didnt touch them, but that would git remote pull after running git reset --hard @{u}. either way thought i thought a command like that would come *before* the git remote pull..23:14
yup, no changes, i keep my own repos elsewhere so.. this collection is all other people's stuff..23:15
frogonwheels mikkelb: well that's a remote update, not a pull.23:15
mikkelb sorry remote update23:15
(typo)23:15
frogonwheels mikkelb: !pull23:15
gitinfo mikkelb: 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/23:15
frogonwheels mikkelb: a remote-update is more a fetch i believe23:16
← simesy left23:16
mikkelb frogonwheels: for a script like https://gist.github.com/2997336 would pull be more suited?23:17
frogonwheels: so like git remote update doesn't actually update the repo, just stores the updates elsewhere is that it?23:19
← avpx left23:19
← Targen left23:19
mikkelb gonna read that doc tomorrow morning, am super tired right now23:19
← apok left23:21
→ apok joined23:21
frogonwheels mikkelb: it updates refs/remotes/remotename/branchname23:21
mikkelb: ie the remote tracking branch23:22
mikkelb frogonwheels: sorry im not familiar with this terminology23:22
← theCole left23:22
mikkelb frogonwheels: will my local clones (i.e. their files and folders) be updatedwith a git remote update?23:23
→ n123 joined23:23
frogonwheels mikkelb: ok - do a google tomorrow when you're awake ;)23:23
mikkelb: no. no files will be update.23:23
d23:23
mikkelb hehe will do23:23
what? no shit?23:23
→ basicallyu joined23:23
frogonwheels mikkelb: it updates the .git repo with the new stuff23:23
mikkelb oh i see i see23:23
← joshteam left23:24
frogonwheels the reset --hard @{u} is what makes the checked out branch match the remote branch23:24
the remote tracking branch23:24
mikkelb oh like that23:24
← teknopaul left23:24
← gusnan left23:24
frogonwheels here @{u} stands for the remote tracking branch of the currently checked out branch23:24
mikkelb now i understand23:24
gitinfo set mode: +v23:24
basicallyu hey guys, back, my internet died, meanwhile i checked and noticed that if I checkout a specific commit and then switch back to master i lose the assume-unchanged flag23:24
← z0a left23:24
frogonwheels basicallyu: I'm not quite shure what you are doing, but have you considdered using sparse checkout?23:25
basicallyu no idea what that is, but what I'm trying to do is23:25
cmn using assume-unchanged means that you can forget anything behaving sanely23:25
frogonwheels basicallyu: you can use sparse-checkout to restrict certain files from being noticed by git23:25
basicallyu have a default .classpath file from the repo on clone, and then totally avoid commit -a for ever staging it again unless somehow forcing it23:26
→ w09x joined23:26
← dr_lepper left23:26
basicallyu frogonwheels: ok can you man the page for me? :)23:26
frogonwheels basicallyu: !sparse ?23:26
basicallyu man sparsecheckout23:26
frogonwheels b*23:26
basicallyu !sparse23:26
frogonwheels basicallyu: nah. wait a sec.23:26
← etcetera left23:26
basicallyu :)23:26
frogonwheels basicallyu: man git-read-tree23:27
gitinfo basicallyu: the git-read-tree manpage is available at http://git-scm.com/docs/git-read-tree.html23:27
→ gavin_huang joined23:27
basicallyu cool23:27
→ etcetera joined23:28
← dr_diamond left23:28
→ Syzygy__ joined23:28
→ angelsl joined23:28
← dcarper left23:29
frogonwheels basicallyu: !sparse23:30
gitinfo basicallyu: [!sparse_checkout] Sparse checkout can be used to restrict which files/folders are updated with all checkout and status operations. It cannot change the tree the checkout is rooted from. See man git-read-tree23:30
← psoo left23:30
basicallyu that means i could exclude certain files from git commit -a ?23:31
i see that by having a list I could put in it !.classpath23:31
in that file $GIT_DIR/info/sparse-checkout23:32
← Lord_DeathMatch left23:33
← thornofmight left23:37
frogonwheels basicallyu: yep. don't forget you need to enable sparse-checkout..23:38
core.sparseCheckout i23:38
→ erasmospunk joined23:39
→ w09x_ joined23:40
← w09x left23:40
→ Zenith77 joined23:41
← ketas left23:41
← basicallyu left23:42
← kerrick left23:42
← kahmalo left23:42
← argoneus left23:43
← etcetera left23:45
→ vitor-br joined23:46
← lledet left23:47
→ basicallyu joined23:48
← erasmospunk left23:48
→ dachew joined23:48
← eedeep left23:48
← HyperNerdV2 left23:49
← dachew left23:49
gitinfo set mode: +v23:49
basicallyu my inet failed again :) is it possible to have a file on the repo which is untracked but would exist in workdir after you clone the repo ?23:49
canton7 basicallyu, not really... have you read !config_files23:49
gitinfo basicallyu: [!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.23:49
→ dachew joined23:51
→ etcetera joined23:52
basicallyu thanks canton7 that is quite relevant to what i need23:54
→ Sebo joined23:56
mikkelb So hey, anybody know how I'd pull this off?23:57
# Make sure repo needs updating23:57
#23:57
if git ls-remote ???; then23:57
→ eletuchy joined23:57
← dabd left23:58

Logs Search ←Prev date Next date→ Channels Documentation