IRCloggy #git 2011-02-01

Logs Search ←Prev date Next date→ Channels Documentation

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

2011-02-01

Blazento so is there a command to say "scrap the mergetool and just push my changes"00:00
Ze_M SethRobertson: but i have another problem00:03
← chrischambers left00:03
SethRobertson Oh?00:03
← linusoleander left00:03
jdahm ok figured it out. If anyone cares I had to add an environment variable CVS_RSH=ssh00:04
← steveoh left00:04
← normanrichards left00:04
← eletuchy left00:06
← ph^ left00:07
→ b14ck joined00:09
b14ck Hi guys, quick questoin. I've got a git repo here. I'd like to 'view' all changes made to a single file since it was created. How can I do that?00:09
bremner git log file00:10
b14ck d'oh00:10
← bhosie left00:10
b14ck thanks00:10
→ linusoleander joined00:10
→ eean joined00:10
← radarek left00:11
→ nyceane joined00:13
← nyceane left00:13
→ nyceane joined00:13
← bitkiller left00:13
nyceane hey guys00:13
on github00:13
if the master has changed00:13
← b14ck left00:14
← zombor left00:14
nyceane how do i update it with branch?00:15
guys?00:16
← bigpeted left00:16
→ normanrichards joined00:18
→ mhall joined00:19
→ banisterfiend joined00:19
nyceane anybody here?00:19
mhall No00:19
:-P00:19
nyceane if forked branch has updated on github00:19
i mean if master have updated00:19
how do i update it on the branch?00:19
→ eletuchy joined00:21
← ISF left00:21
mhall I have got two people upstream of me from the same company. One of them maintains the current branch, and the stable branch. Another guy at the company has a branch forked from the current branch, which contains a feature I'd like to backport from current to stable so I can use it safely in production here. Given a clone of stable, and a clone of the current fork with the feature, I'm wondering what procedure I should follow to get the changes from the curren00:23
However I'm a bit hazy on (2) because I'm short on git experience.00:23
→ t3i joined00:23
← JohnFlux left00:24
← snearch left00:25
→ MikhX joined00:27
← priidu left00:27
← hobodave left00:28
← jrmuizel left00:28
← banisterfiend left00:33
← beatak left00:34
→ ISF joined00:35
Thell Anyone else get vexed with all the various project management apps and IDEs/editors temp files and various other cruft that we need to either add into .gitignore files or manually deal with?00:37
→ alex__c2022 joined00:37
Unknown[OFF] → Unknown[NF]00:38
→ project2501a joined00:39
Thell I'm pondering the idea of a patch to allow a #include token in .gitignore. Waste of time or possibly a good idea?00:39
← giallu left00:39
← malumalu left00:40
← Blazento left00:41
← bentob0x left00:42
← rayners left00:42
→ macrohauler joined00:43
macrohauler hi all00:43
macrohauler is in dire need of help00:43
← muep left00:43
ayust mhall: if the clone of the fork with the feature has a branch specifically for that feature, you can just rebase the branch on your stable branch00:43
otherwise, you'd have to either rebase a range of commits, or cherry-pick them00:43
→ banisterfiend joined00:43
← nadavoid left00:45
← Guest12638 left00:45
macrohauler anyone?00:46
Thell macrohauler: I think you will need to state the problem before anyone can offer assistance00:47
macrohauler ok sure00:47
im trying to revert to an old commit00:47
but git keeps giving me an error which i tried to follow00:47
← farktronix left00:47
macrohauler and it still gives me the same error00:47
← linusoleander left00:47
→ Neo joined00:48
macrohauler which is:00:48
error: could not revert 2daec18... Added method: num00:49
hint: after resolving the conflicts, mark the corrected paths00:49
hint: with 'git add <paths>' or 'git rm <paths>'00:49
so i tried adding the files, commiting then reverting00:49
still the same error00:49
so yeh00:51
→ amondracorp joined00:51
→ rgr joined00:52
← marcamilly left00:52
→ marcamilly joined00:52
← marcamilly left00:52
→ marcamilly joined00:52
← nyceane left00:53
→ farktronix joined00:54
← hyperair left00:55
← dreiss left00:55
→ squi joined00:55
squi hi00:55
how can i find out who edited a specific line via command line?00:56
Eridius git blame00:56
← urbanmonk left00:57
ayust git blame -L<start>,<end> /path/to/file00:58
if <start>==<end>, then you'll blame a single line00:58
← intripoon left01:01
→ wde joined01:02
← airborn left01:02
← jogla left01:02
squi thanks a lot01:03
→ intripoon joined01:03
← sako left01:03
squi when you said git blame, i just checked the help01:03
easy enough01:03
← spearce left01:05
→ boombatower joined01:06
← boombatower left01:06
→ boombatower joined01:06
→ nyceane joined01:09
← alvaro_o left01:10
→ zombor joined01:10
← banisterfiend left01:14
→ bhosie joined01:14
← tvw left01:15
Unknown[NF] → Unknown[OFF]01:15
← RaceCondition left01:15
← amerine left01:17
← gozoner left01:20
← shruggar_ left01:21
wde hi, i'm new to git coming from the svn world … originally i was told to do "git fetch <repo>; git checkout <branch>" … switching the to other branch was necessary because it contained some changes not yet in the central repo … so then i made some local modifications, and then ran "git diff" expecting to see some differences … however, it gives me nothing … any ideas?01:21
← ddkilzer left01:22
→ ahammond joined01:22
← eletuchy left01:22
Thell wde does git status give you anything useful?01:22
← bhosie left01:22
ahammond is there any way I can set change git log to default to doing --decorate?01:22
PerlJam wde: did you "git add" those items before you ran "git diff"?01:22
→ sdboyer_ joined01:22
← sdboyer_ left01:22
→ sdboyer_ joined01:22
PerlJam ahammond: setup an alias in your ~/.gitconfig01:23
← sdboyer left01:23
← MikhX left01:23
mhall ayust: as you posted, i was in another window experimenting01:23
PerlJam ahammond: or a shell alias01:23
wde Thell: git status does indeed list my modified files01:23
→ MikhX joined01:23
ahammond PerlJam: can I alias log as log --decorate? or will it recurse?01:23
wde PerlJam: most of my changes were in existing files01:23
mhall ayust: it seems that with the code i'm up against, some crap came along for the ride on the supposedly pristine branch for the feature01:23
Thell ahammond: check out making an alias for formatting the way you like01:23
mhall ayust: git merge had some relatively OK conflicts01:23
PerlJam wde: aye, but did you add them to the index prior to running "git diff"?01:23
mhall ayust: but i also picked up some added files that were causing build failures01:23
← macrohauler left01:23
mhall so what's the right way of taking files that come from a merge that you don't want, and kicking them out before committing to local HEAD? just git rm the bad guys?01:24
Thell wde: in the git status it will tell you if they are 'staged' or not, if they are (meaning you did 'git add') then git diff --cached01:24
ahammond nope... can't alias log to log --decorate. grr. I guess I'll go with l and see if I can learn to use it. :)01:24
wde PerlJam: i'm not sure what you mean, but probably no, as i didn't do anything else in git after the fetch and checkout01:24
ayust mhall: the branch might be pristine relative to the point in devel it was branched from, but that doesn't mean it's pristine relative to your stable branch01:24
sdboyer_ → sdboyer01:25
Thell ahammond: make something like git logd01:25
wde Thell: I see stuff like this:01:25
# On branch <foo>01:25
# Changed but not updated:01:25
ayust if the merge-base between feature and devel is some point after your stable head, it'd pull in stuff from devel that was added before the merge-base01:25
wde # ...01:25
# modified: MyFile.C01:25
ahammond Thell: yeah, it'd be really nice to have a logoptions variable I can set. Kinda like mergeoptions...01:25
wde # … etc...01:25
mhall ayust: indeed it's not ;)01:26
← bronson left01:26
mhall ayust: i'm alright with it, if i have an idea how to cherry pick. is it doc'ed somewhere?01:27
ayust well01:27
ahammond oh... I'm looking to set a global mergeoptions=--no-ff --no-commit... do I have to do it by branch?!01:27
ayust what version of git are you running? something fairly recent?01:27
if so it's pretty simple01:27
git cherry-pick A..B (where A is the merge-base between feature and devel, and B is feature)01:27
→ rlb joined01:28
wde Thell: git diff —cached shows nothing either01:28
mhall ayust: i have ubuntu 10.04 lts vm for this job, let me check version, should be fairly good01:28
Thell wde: yeah, if it has been changed but not updated, then it hasn't been staged, and git diff _should_ show the diff01:29
mhall ayust: git version 1.7.0.401:29
ayust i think 1.7.0 has support for cherry-picking ranges01:29
so should work01:29
Thell I don't know why it wouldn't wde, perhaps someone else does...01:29
→ bronson joined01:30
mhall ayust: yeah it's in there01:30
ayust: let me try this out01:30
← smolyn left01:31
← larsmoil left01:31
wde i guess part of my problem, is that the git terminology is throwing me off a bit … i found mercurial relatively easy to understand because it borrows a lot from svn … i don't quite understand what a "branch" means to git, nor do i get what it means to be "staged" or "unstaged"01:32
← aah left01:32
ayust wde: do you come from a computer science background at all?01:32
wde i look up the help for the git commands and i have no idea what "staged" means01:32
ayust: yes01:32
ayust http://eagain.net/articles/git-for-computer-scientists/01:32
that explains branches01:33
as Git sees them01:33
bremner except, that doesn't deal with staging does it?01:33
ayust getting to that bremner :)01:33
← kandjar left01:33
wde but i don't care about the internals01:33
→ muep joined01:33
ayust just skim it then01:33
the important part is01:33
wde i know what DAGs are, i use want to *use* it01:33
ayust right01:34
branches in Git are just leaves in the DAG01:34
when you create a new branch, you create another leaf that points at a node01:34
← jfkw left01:35
ayust as for what 'staging' means in Git01:36
http://www.gitready.com/beginner/2009/01/18/the-staging-area.html01:36
← matjas left01:36
wde alright, i'll do some reading and come back … thanks all01:36
mhall ayust: alright, getting closer. ran the cherry pick but i have conflicts. and it didn't print out what they were so not sure how to display them01:36
ayust mhall: if you're running into conflicts, it might be best to do a full rebase01:37
→ Tuller joined01:38
mhall ayust: is there a way to at least display them so i can see how bad they are01:38
ayust: or does it not create the merge conflict markers like a merge does01:38
→ sako joined01:38
ayust cherry-pick I don't think does01:38
mhall ayust: hmm it makes markers01:39
ayust: egrepped for them01:39
ayust try git status then?01:39
mhall sweet01:39
yeah that seems to work01:39
← leo2007 left01:39
← marcamilly left01:40
mhall what command will show me the diff for my changes to fix the conflicts so I can BS check01:41
NM, git diff worked01:41
hehe01:41
→ owen1 joined01:42
owen1 i try to deploy my project with capistrano and get: fatal: unable to create '.git/index.lock': File exists. i rm this file from the server and now i get: fatal: Could not reset index file to revision 'aef5cf446437682e64adf8dabed731203c1538a1'. any clues?01:42
→ eletuchy joined01:43
← SimonP86 left01:43
mhall oh sweet this is looking good now01:44
ayust: i think your advice saved my ass01:44
ayust cool01:45
→ fukuyamaggoooo72 joined01:45
mhall autogen.sh and configure both ran ok01:45
now time to try a compile01:45
← alberto56 left01:45
mhall if i want to send a patch to upstream to get it added to next rev of their stable branch (if they will permit it) what would i need to do then01:46
→ jrmuizel joined01:46
mhall that way i could try to avoid having to maintain a fork01:46
→ hobodave joined01:47
→ rayners joined01:48
← boombatower left01:48
wde ok, i've the git newbie links given to me and still don't understand why "git status" shows my modified files but "git diff" nor "git diff —cached" shows no changes01:48
er, i've *read the git newbie pages ...01:48
ayust can you pastebin the output of 'git status'?01:48
or simpler - what heading do your files show up under on git status01:48
→ montylounge joined01:49
wde % git status01:49
# On branch <foo>01:49
# Changed but not updated:01:49
# modified: MyFile.C01:49
ayust did you 'git add' the file you changed?01:49
→ banisterfiend joined01:49
SethRobertson Could the change be whitespace?01:50
What happens if you say `git diff HEAD`01:50
wde ayust: no … do i need to? it's not untracked … it sees several other files which are untracked (which is fine because i know i haven01:50
(i know haven't added them yet)01:50
ayust "git add" is what marks changes to be committed01:50
← bousquet left01:51
ayust unlike subversion, "commit" doesn't automatically commit every change that you made locally01:51
it only commits the ones you've specifically marked as wanting to commit01:51
wde <slaps face> really?01:51
ayust right.. that's why in the gitready article i linked you01:51
you notice the flow of working -> staging -> repository01:51
→ zenazn joined01:52
ayust 'git add' is what takes changes from working to staging01:52
then git commit takes whatever's in staging and commits it to the repository01:52
now, that said01:52
wde ayust: that link didn't say i needed to add before before i could diff01:52
ayust at least one of (git diff, git diff --cached) *should* have shown your chagnes01:52
changes*01:52
since 'git diff' shows changes from your working that aren't staged01:52
and git diff --cached shows changes that are staged01:53
wde git diff HEAD shows nothing either, btw01:53
ayust so I *am* kind of confused about why neither is apparently showing output01:53
wde does git care what my pwd is when i run git diff --cached?01:53
ayust no, as long as it's inside the repository01:54
→ mahata joined01:54
wde ok, i'm definitely in the repo … otherwise git status would show nothing as well01:54
ayust right01:54
← tty2 left01:54
ayust what happens if you type01:54
git diff MyFile.C01:54
wde nothing01:55
ayust ok01:55
← ahammond left01:55
ayust how about "git diff HEAD^ HEAD"01:55
wde ok, so let me try adding it … i can undo that easily right?01:55
ayust yes01:55
Unknown[OFF] → Unknown[NF]01:55
ayust the opposite of 'git add <path>' is 'git reset <path>'01:55
→ tty2 joined01:55
wde so i can't do "git add MyFile.C" ?01:56
rlb wde: and git status will remind you about reset.01:56
wde if i run "git branch", I get :01:56
master01:56
* bnranch01:56
where "* <branch>" is what i did "get checkout" with01:57
ayust what error do you get with 'git add MyFile.C'01:58
wde I did "git add MyFile.C; git diff" and nothing01:58
→ dfr|mac joined01:58
wde when i do git status, I now have "MyFile.C" listed under "Changes to be committed"01:58
ayust what kind of changes did you make to MyFile.C?01:59
← noahcampbell left01:59
wde lots of the changes, new functions, etc.01:59
ayust well, now that it's added01:59
it should be showing up under "git diff --cached", not "git diff"02:00
wde "git diff —cached" gives me nothing … also nothing for "git diff —cached MyFile.C"02:00
oh crap02:01
this is on solaris02:01
doesn't "git diff" just shell out to diff?02:01
hmm … no … diff <foo> <bar> seems to work … unless git is passing some options to solaris diff that it doesn't understand02:02
← shiba_yu36 left02:02
→ dl9pf joined02:03
wde when you do a "git diff", does it show in unified format? … my diff has no -u option02:04
← Evious left02:04
→ boombatower joined02:04
← dl9pf_ left02:04
← tjholowaychuk left02:05
→ fholgado joined02:06
← montylounge left02:07
→ gretch joined02:07
← pheaver left02:07
→ shabble joined02:08
← ldleworker left02:08
ayust it's -U###02:09
(case-sensitive)02:09
where ### is the number of lines of context to show02:09
but unified diff format is the default02:09
(with 3 lines of context)02:09
and no, 'git diff' doesn't invoke 'diff'02:10
it's a separate diff program implemented by git itself02:10
wde whoever installed git on this machine doesn't seem to have even installed the help files for it02:11
ayust "git help diff" doesn't work?02:11
← gretch left02:11
ayust if so, perhaps maybe the git installation is messed up?02:11
wde that's right … i get "No manual entry for git-diff"02:11
→ pheaver joined02:12
wde if i got to /usr/local/bin (where git shows up in my PATH), I see: git, git-cvsserver, git-receive-pack, git-shell, git-upload-archive, git-upload-pack, gitk02:12
← pheaver left02:12
→ pheaver joined02:13
wde is there some other place i should be looking to see if git-diff is supposed to work?02:13
(or installed)02:13
← justin-george left02:14
ayust try /usr/lib/git/git-core02:14
and manpages typically in /usr/share/man/man102:15
wde neither /usr/lib/git nor /usr/local/lib/git exists02:15
ayust (note - these may vary for solaris?)02:15
i haven't worked with solaris a ton, so I don't know if those are the expected paths for Git there02:15
well, if you have 'locate' available, try 'locate git-diff'02:16
← Orbixx left02:16
owen1 i use git 1.7.1 on ubuntu and try to deploy to debian with git 1.5.6.5. is it fine? i get funny error: fatal: Out of memory, malloc failed02:16
→ Orbixx joined02:16
owen1 (debian lenny)02:16
→ justin-george joined02:17
ayust owen1: pushing with git invokes processes that package up your objects, these can require at least a moderate amount of memory02:17
what kind of ulimits are you working with owen1 ?02:17
→ jfelchner joined02:18
← silanus_ left02:18
owen1 ayust: what is ulimits?02:19
jfelchner Anyone know what I need to do to configure git to diff WITHIN a line instead of entire lines?02:19
ayust http://ss64.com/bash/ulimit.html02:19
→ zenazn_ joined02:19
ayust jfelchner: --word-diff02:19
owen1 ayust: unlimited02:19
ayust owen1: and on the other end?02:19
owen1 ayust: that one was on the server02:20
ayust what about locally02:20
owen1 ayust: the same on the client, my desktop02:20
ayust odd, then... pretty sure those git versions should be compatible02:21
how much memory does the server have available?02:21
jfelchner ayust: thank you much :)02:21
← vitor-br left02:21
→ Heimidal joined02:22
→ smolyn joined02:22
owen1 ayust: Mem: 512 464 4702:22
← zenazn left02:22
Unknown[NF] → Unknown[OFF]02:23
← mithun left02:23
owen1 ayust: i have 7 thin servers, each one about 150MB.. WTF02:23
→ crowbar` joined02:24
← verto left02:25
← MetaCosm left02:26
→ silanus joined02:27
→ John___ joined02:27
→ Davey_ joined02:28
← Davey_ left02:28
→ Davey_ joined02:28
zenazn_ → zenazn02:29
← justin-george left02:30
← pheaver left02:30
← crowbar` left02:31
→ Jamie joined02:32
Jamie → Guest3105302:32
← Guest31053 left02:32
→ tekmosis joined02:33
→ hyperair joined02:33
tekmosis does git have a version of vendor tags that cvs has?02:33
mhall the jk.gs domain in the channel topic seems to be down02:33
→ Sidnicious joined02:34
Sidnicious Greetz. So, I accidentally ^C'd git in the middle of checking out a different branch. Now, `git checkout` prints "error: Unable to open .git/HEAD.lock for writing". How do I fix this situation?02:35
→ lanwarder44 joined02:35
→ livingstn joined02:36
← lanwarder44 left02:37
livingstn Hey everyone. Is there a command to overwrite any files that I have locally if there's a newer repo version?02:37
→ lanwarder44 joined02:38
wde thx ayust … gnight all02:38
← wde left02:38
← tekmosis left02:38
← lanwarder44 left02:38
← bmahe left02:40
← John___ left02:41
→ gretch joined02:41
ayust Sidnicious: sudo rm .git/HEAD.lock ?02:42
Sidnicious ayust: I did that a minute ago :) still concerned that my repo is ballsed in some subtle way, though.02:43
how likely is that?02:43
→ lanwarder44 joined02:43
Sidnicious (it appears to have worked)02:43
← dfr|mac left02:43
ayust can always 'git fsck' if you desire02:45
← eletuchy left02:45
→ drl joined02:45
← welly left02:45
← mhall left02:45
← Swimming_bird left02:46
← lanwarder44 left02:47
← Vampire0 left02:47
← rgr left02:48
→ jfkw joined02:48
→ zenazn_ joined02:49
← zenazn left02:51
← peshalto left02:51
← fukuyamaggoooo72 left02:51
→ aah joined02:54
zenazn_ → zenazn02:54
→ tizzo joined02:54
← Tuller left02:57
← Rupan left02:59
← smolyn left03:00
→ Rupan joined03:01
← zombor left03:01
→ Vampire0 joined03:05
→ pgib joined03:07
→ mithun joined03:07
← spaceonline left03:07
← utu left03:08
← mithun left03:08
← tizzo left03:08
← Orbixx left03:09
→ mithun joined03:10
→ Orbixx joined03:11
← Orbixx left03:11
→ Orbixx joined03:11
→ heeton joined03:14
← chrisdarroch_ left03:14
→ chrisdarroch joined03:15
← drl left03:15
← heeton left03:15
→ heeton joined03:15
→ SirFunk_ joined03:15
← heeton left03:15
ben_alman is there an wasy way to see if a specified path has any modifications since the last commit?03:15
→ xanderbeedle joined03:16
← Elfe left03:16
xanderbeedle I'm trying to add 'time since last commit' to my bash prompt.03:16
Is there a command to get that time since last commit?03:16
→ intripoon_ joined03:17
SethRobertson ben_alman: git diff <path>03:18
→ Chillance joined03:19
ben_alman ok03:19
SethRobertson xanderbeedle: You can parse git show or git log to pull out the timestamp. git log would be easier to get only the value you wanted03:19
(seconds since the epoch)03:19
→ zenazn_ joined03:19
dwmw2_gone → dwmw2_SYD03:20
← intripoon left03:20
← engla left03:21
← Orbixx left03:21
← zenazn left03:21
← alnewkirk left03:21
← zenazn_ left03:23
→ Orbixx joined03:24
← Orbixx left03:24
→ Orbixx joined03:24
← bronson left03:24
← alex__c2022 left03:26
← nyceane left03:26
→ friskd joined03:26
← ldh left03:28
← Orbixx left03:28
← Error404NotFound left03:29
→ drl joined03:29
→ Elfe joined03:29
→ alex__c2022 joined03:29
→ Orbixx joined03:31
← Orbixx left03:31
→ Orbixx joined03:31
← ROBOTARMY left03:33
← Amorphous left03:33
ben_alman if i do this: git --git-dir=/ath/to/repo/.git status --porcelain subdir-name03:34
it works when i'm in the root of the repo03:34
but not when i'm outside03:34
even though i'm specifying --git-dir03:34
→ justin-george joined03:34
← alex__c2022 left03:34
ben_alman is there a way i can have it runnable from otuside the git repo?03:34
← flaguy48 left03:34
Sidnicious ayust: forgot to says thanks :). I completely forgot about git-fsck03:35
good call03:36
← Sidnicious left03:36
← livingstn left03:36
← Orbixx left03:37
ben_alman i'm trying to see if there aren't just modified files but also new, untracked files03:37
like git status --porcelain03:37
but it doesn't give me info i need when run from outside the git repo03:37
← drl left03:38
← fholgado left03:38
← Schmallon left03:39
ben_alman any ideas?03:39
Chillax → Chillaway03:42
← m_3 left03:42
→ Orbixx joined03:43
← dermoth_ left03:44
→ dermoth_ joined03:44
→ Amorphous joined03:47
→ gozoner joined03:48
Chillaway → Chillax03:49
→ letomuaddib joined03:49
← JohnFord left03:50
→ drl joined03:51
← gozoner left03:52
→ gozoner joined03:52
→ timj__ joined03:52
← letomuaddib left03:53
← timj_ left03:56
← killerchicken_ left03:56
← SirFunk_ left03:56
dwmw2_SYD → dwmw2_gone03:59
→ killerchicken_ joined04:01
← drl left04:01
xanderbeedle SethRobertson: do you know if it's possible to format the date that git log displays?04:01
or any other way to get the raw timestamp (seconds since epoch or something) of the commit04:02
offby1 xanderbeedle: there are a bunch of options04:02
oh sure04:02
--pretty=%at04:02
man git-log04:02
xanderbeedle I've found the date options, awesome04:03
→ RLN joined04:04
→ Schmallon joined04:04
→ roop joined04:05
→ ROBOTARMY joined04:06
→ pheaver joined04:08
→ davetoo joined04:10
← friskd left04:10
← SethRobertson left04:12
→ friskd_ joined04:13
→ timholum1 joined04:13
timholum1 hello, I am having trouble's using push and pull, I cloned my git repo edited a file, and commited it, now I am trying to merge that back into the main repo,04:14
→ drl joined04:15
offby1 ... and ... ?04:15
timholum1 I have tryed from the master repo git pull, and git pull /path/to/other master. and I have tryed from the other one git push and git push /path/to/other master04:15
← friskd_ left04:16
timholum1 the file I edited does not change in the primary repo04:16
but it tells me everything up-to-date04:16
offby1 timholum1: can you paste a transcript?04:16
I suspect you're just misunderstanding something04:16
timholum1 http://pastebin.com/z886JRD404:19
offby1 hole oj04:20
on04:20
timholum1 ?04:20
offby1 timholum1: oh, but that doesn't show me the various pushes that you did.04:20
I don't suppose you could paste everything from when you edited the file, including all of the relevant pushes and pulls?04:21
timholum1 how do I get that info?04:21
offby1 if you happened to have the consoles open in which you issued those commands ...04:21
← xanderbeedle left04:21
offby1 timholum1: anyway, try running "gitk --all"; it may well explain everything.04:21
perhaps you pushed into one branch, but are looking for your changes in another.04:21
← Gryllida left04:22
→ MattDiPasquale joined04:23
→ adiabatic joined04:24
timholum1 oh, gitk is a gui, the server does not run x at all04:24
offby1 if you do "ssh -X server gitk --all", it'll display on your local machine.04:24
anyway, "git log --graph --pretty=oneline --abbrev-commit --decorate" is a decent text-mode substitute.04:25
oops, make sure you add "--all" to that already-complex command line.04:25
← Shentino left04:27
timholum1 http://pastebin.com/qYHcwcuq this is git log --all --decorate04:27
→ urbanmonk joined04:27
← Davey_ left04:28
timholum1 http://pastebin.com/Jhdu0P3s this is the same command on the cloned repo04:29
offby1 those aren't helpful04:29
plus I don't know what the commit you "lost" looks like04:30
← drl left04:31
← sivy left04:31
← rettub left04:32
← roop left04:32
timholum1 do you know is that the way I am suppose to push and pull from the repo's?04:33
that is my main question04:34
→ SethRobertson joined04:34
→ smolyn joined04:34
timholum1 I can wipe my demo repo out, that is not a problem I am more conserned about how to correctly do it for the future04:34
← muep left04:35
→ rettub joined04:35
← gebi left04:36
timholum1 here is every git command I ever did http://pastebin.com/G31emcH3 ( simply did a history|grep git )04:36
→ kenichi joined04:37
timholum1 the first 30 lines are not about this at all thought04:38
RLN would this error: "error: git was compiled without libcurl support." be my host's error? or potentially mine? my host has git 1.6.4.4 prepackaged and i have curl enabled. i am assuming based on the error that they compiled git without libcurl support?04:39
← MattDiPasquale left04:40
RLN i'm using: git clone https://github.com/fabriceb/sfFacebookConnectPlugin.git plugins/sfFacebookConnectPlugin04:41
offby1 timholum1: sorry, that's too hard to digest04:42
RLN: well, I'm sure github has a properly working version of git.04:42
RLN: so the blame is surely on the box on which you ran the "clone" command.04:43
RLN so basically i'm screwed huh?04:43
timholum1 I did a git checkout master -f in the original ( the one i was trying to push to ) and it came throught,04:43
offby1 RLN: not at all. Get a git that has curl support.04:44
RLN: or use the ssh flavor of the URL; github always supplies both.04:44
→ drl joined04:44
offby1 e.g. try git://github.com/fabriceb/sfFacebookConnectPlugin.git04:44
(OK, granted, that's not SSH; but it should work as long as you don't need to push)04:45
and if you have write access, then you can indeed get an SSH URL.04:45
→ harshu_ joined04:45
RLN i've tried git://github.com/fabriceb/sfFacebookConnectPlugin.git04:46
still same thing04:46
and its shared hosting so i don't exactly have control over the version of git04:46
offby1 RLN: seriously? you typed "git clone git://..." and you saw "error: git was compiled without libcurl support"?04:47
that's baffling04:47
← timholum1 left04:48
RLN yea, i'll double-check real quick though04:48
← Pacman_Home left04:48
← tty2 left04:49
← kukks left04:49
→ SirFunk_ joined04:51
← kenichi left04:51
← DrNick left04:51
→ Pacman_Home joined04:52
→ harshu___ joined04:53
→ DrNick joined04:54
← drl left04:54
← davetoo left04:54
← mithun left04:57
→ mithun joined04:57
RLN offby1, i went to go try and somehow everything was there already :|04:58
that's even more baffling04:58
offby1 laughs cruelly04:59
→ pratz joined04:59
offby1 we git ninjas break into your hard drive and add missing commits.04:59
RLN hahaha04:59
i think the tech guy i was talking to did it for me after i disconnected04:59
→ muep joined05:00
← hagabaka left05:01
offby1 hm05:01
→ hagebake joined05:01
RLN but05:02
i tried it just to be sure05:02
github.com[0: 207.97.227.239]: errno=Connection refused05:03
fatal: unable to connect a socket (Connection refused)05:03
that's with git://05:03
← j0su3 left05:04
→ j0su3 joined05:04
← smolyn left05:04
→ drl joined05:07
wereHamster rns firewall?05:08
RLN: ^^05:09
← mahata left05:09
RLN wereHamster, idk i'm just gonna be ignorant here and just be happy its working for now.. i've tried it on a different server and worked fine so i'm guessing tech did something magical05:13
→ dreiss joined05:15
RLN thx for help offby105:16
← RLN left05:16
← dermoth_ left05:19
← adamm left05:19
← Tommy[D] left05:19
→ dermoth_ joined05:19
→ dfr|mac joined05:19
→ bhosie joined05:19
→ Tommy[D] joined05:20
← Eridius left05:24
← mutex left05:25
→ jorn joined05:25
← drl left05:26
jorn hey, short question: I'm currently in the end phase of my of a project, it's getting hectic, committing a lot...05:27
now what i want to ask is: is it possible to make git be a bit more verbose on commit?05:28
like after a commit this line:05:28
17 files changed, 5943 insertions(+), 72 deletions(-)05:28
← ki__ left05:28
→ envi joined05:28
jorn can i somehow get it to instead print a git diff --stat HEAD^1 ?05:28
wereHamster make an alias05:28
or run the diff manually05:29
jorn thought maybe there's a user wide hook, which configures what git commit prints by default05:29
→ TanEk joined05:30
wereHamster jorn: man githooks05:30
→ thermal joined05:30
← ISF left05:32
jorn yupp, knew that already, that's why i mentioned "user wide hook"05:32
so for all repositories05:32
wereHamster there are no user wide hooks05:32
make an alias05:32
jorn thanks05:34
→ amerine joined05:36
← MacGyverNL left05:36
→ nothingHappens_ joined05:37
← ceej left05:38
→ marry joined05:38
→ drl joined05:40
← jorn left05:40
← banisterfiend left05:41
← rayners left05:41
→ psankar joined05:43
← psankar left05:43
→ psankar joined05:43
← TanEk left05:43
← berto- left05:43
← juanjoc left05:46
← Heimidal left05:48
← marry left05:52
← jrmuizel left05:53
→ jrmuizel_ joined05:54
Unknown[OFF] → Unknown[NF]05:54
→ devanjan joined05:56
← sako left05:57
← thermal left05:57
← t3i left05:57
→ maleknet joined05:58
→ MacGyverNL joined05:58
← SirFunk_ left05:58
← SirFunk left05:58
→ tty1 joined05:58
→ freimatz joined05:58
→ sako joined05:58
← zwuwmad left05:59
→ zwuwmad joined05:59
→ SirFunk_ joined05:59
→ SirFunk joined05:59
← drl left06:00
→ Bruce_Wayne joined06:05
← nothingHappens_ left06:05
← dreiss left06:05
→ ThomasLocke joined06:06
→ nataraj joined06:07
← Ze_M left06:08
← pgib left06:10
→ LongBeach joined06:11
→ DrNick2 joined06:12
→ j416 joined06:12
← Mage left06:13
← DrNick left06:13
← DrNick2 left06:14
→ banisterfiend joined06:14
→ drl joined06:14
← jfelchner left06:16
yaotti_aw → yaotti06:17
← noyb left06:18
→ berto- joined06:21
← berto- left06:22
→ berto- joined06:22
Chillax → Chillaway06:23
← ThomasLocke left06:24
← berto- left06:24
→ marcamilly joined06:25
→ eletuchy joined06:25
→ berto- joined06:26
← mcbonz left06:27
→ eletuchy_ joined06:27
→ DrNick joined06:27
→ ThomasLocke joined06:28
→ bronson joined06:28
← gozoner left06:28
→ gozoner joined06:29
→ mahata joined06:29
← eletuchy left06:30
→ roop joined06:30
→ jfelchner joined06:32
→ ki__ joined06:32
→ bousquet joined06:32
← bronson left06:33
→ jony joined06:34
← ymasory left06:34
→ bronson joined06:34
← bronson left06:38
→ bronson joined06:39
→ Sgeo joined06:39
← agile left06:40
← tro left06:40
← Adys left06:41
← drl left06:41
→ Adys joined06:41
→ tro joined06:43
→ opalka joined06:44
← adiabatic left06:46
→ SIGBUS joined06:46
← jdav_gone left06:47
→ shiba_yu36 joined06:48
← j416 left06:50
← jrmuizel_ left06:51
→ jdav_gone joined06:52
← Bruce_Wayne left06:52
← jfkw left06:55
→ drl joined06:55
→ ben225 joined06:57
← bronson left07:00
→ doener joined07:00
→ guardian joined07:00
→ bronson joined07:00
→ odol joined07:01
→ iband joined07:03
← nataraj left07:03
→ OOPMan joined07:03
← drl left07:03
← dfr|mac left07:05
→ noyb joined07:05
← ROBOTARMY left07:06
→ dreiss joined07:06
→ ludde joined07:06
← gozoner left07:06
→ ROBOTARMY joined07:09
→ matjas joined07:10
← ben225 left07:10
← jhoepken left07:11
← dreiss left07:12
→ thiago_home joined07:13
← ROBOTARMY left07:13
← emrebey left07:13
→ Bruce_Wayne joined07:13
← rettub left07:13
→ mastro joined07:13
→ Error404NotFound joined07:15
← Error404NotFound left07:15
→ Error404NotFound joined07:15
← banisterfiend left07:16
← bronson left07:16
→ drl joined07:16
→ bronson joined07:16
→ banisterfiend joined07:16
→ rettub joined07:17
→ jhoepken joined07:17
← guardian left07:18
← Sonder left07:18
→ joshsmith joined07:19
← sako left07:19
← LouisJB left07:21
→ keyvan joined07:24
← ki__ left07:24
→ dvaske joined07:26
→ shentino joined07:26
← shentino left07:26
→ shentino joined07:26
→ joel_falcou joined07:27
← madewokherd left07:29
→ ben225 joined07:30
← AlexP left07:31
← drl left07:32
→ ntoll joined07:32
→ ki__ joined07:32
→ oriba joined07:33
→ xiong joined07:35
→ babbage joined07:35
← bronson left07:37
← pheaver left07:37
→ bronson joined07:37
→ kreemer joined07:37
← roop left07:38
← amondracorp left07:39
→ Sgeo_ joined07:40
← shiba_yu36 left07:40
→ Titosemi joined07:40
← Sgeo left07:43
← joshsmith left07:43
→ drl joined07:44
→ bronson_ joined07:44
← bronson left07:44
bronson_ → bronson07:44
eighty4 I've changed the permission of a folder in my repo to 775 but whenever I pull from a remote repo the permission is reset to 755. Is there a way to avoid this?07:44
→ giallu joined07:45
→ resmo joined07:47
→ shiba_yu36 joined07:48
← ben225 left07:50
← sassyn left07:50
→ ph^ joined07:51
Rhonda eighty4: git does only store minimum permission flags, like execute or not. The rest is guided by your umask.07:53
→ engla joined07:53
← engla left07:53
→ engla joined07:53
← mahata left07:54
→ Badiss joined07:54
→ wtfness joined07:55
← drl left07:56
← SIGBUS left07:56
→ TrineLocke joined07:58
→ cjs joined07:58
cjs What's the simplest syntax to cat a specific revision of a specific file?07:58
← Bruce_Wayne left07:58
→ Deesl joined07:59
← Deesl left07:59
→ Deesl joined07:59
cjs (I.e., send it to stdout.)07:59
mastro cjs, it's not a "revision" anyway: git show <commit>:path/to/file > /tmp/myFile.recovered07:59
cjs Sorry, does "commit" have a different meaning from "revision"?07:59
mastro cjs, example: git show tag-1.2:build.xml > build.xml.old08:00
cjs Thanks for the syntax, by the way; that's exactly what I needed.08:01
→ saschpe joined08:01
mastro cjs, revision is something related to subversion :) a revision in subversion is an incremental number identify the "commit" in an unique way, in git we talk about commits and hashes because there is no such thing as a linear, incremental number in a distributed system08:01
cjs, a commit is both a snapshot of the working directory and a "patch" on the previous commit08:02
→ nixness joined08:02
mastro cjs, what I said is not really true, it's "a patch" only if there if there is a single parent commit08:03
→ teweWork joined08:03
cjs mastro: Allow me to disagree in a minor way about the definition of "revision," having used SCCS, RCS and CVS for many years before Subversion even existed. :-)08:03
→ chrischambers joined08:04
← wtfness left08:04
mastro cjs, you are probably right :) I associate the term "revision" with subversion, I'm probably the one in error here :P08:04
cjs But "both a commit and a patch," well, I'm going to say "no," even though I'm not a git expert. A commit is a particular snapshot, or vertex in the graph, a patch is an edge in the graph.08:05
mastro: Well, it depends on whose revision you're using. :-)08:05
Mostly, I just want to make sure it's not unfair to call a commit a revision, if we can first agree that I have a reasonable basic understanding of git.,08:05
Actually, that patch vs. snapshot thing is a bit tricky. This again comes back to git's user interface; it's actually a rather simple and elegant system that hides under a somewhat complex UI.08:06
mastro cjs, actually a commit is a snapshot an nothing else... patches are computed by git automatically by comparing two commits, but since you have commands like git cherry-pick I tend to think of a commit as a "patch" somewhat...08:08
→ drl joined08:09
mastro cjs, of course that apply only when you don't have a merge commit :08:09
dean[w] What does "unmerged paths" represent in 'git status'08:09
→ thermal joined08:11
thiago_home dean[w]: conflicts08:11
← bronson left08:11
thiago_home dean[w]: it means the merge didn't merge everything (unmerged)08:11
dean[w]: you need to fix the conflicts08:11
→ bronson joined08:11
→ radarek joined08:12
dean[w] Okay that's what I expected. So if I were to manually resolve a conflict in vim, the file remains in "unmerged paths"08:12
Is it in "staging". Do I have to git add?08:12
thiago_home dean[w]: yes, git add08:12
dean[w] I'm just trying to fully understand what's going on rather than guess ;-)08:12
okay, perfect08:12
→ mcbonz joined08:12
→ Jippi_mac joined08:12
→ roop joined08:14
← bronson left08:14
eighty4 Rhonda: k, so any idea why it would change the permissions?08:14
cjs mastro: Exactly my point: patches are the paths between nodes in the graph of snapshots. This applies even to merge commits.08:14
→ darkredandyellow joined08:14
→ bronson joined08:14
← darkredandyellow left08:14
→ cbreak_work joined08:15
← ki__ left08:15
Rhonda eighty4: Because your umask is 022 and not 002 like you want it?08:15
← engla left08:15
→ borwar joined08:16
← borwar left08:16
→ borwar joined08:16
eighty4 Rhonda: let me rephrase that :) I need to read up on umask and how git uses it.08:17
→ Xaphiosis joined08:17
eighty4 Rhonda: thanks for the help08:17
Xaphiosis when I set g+rws on a dir and init a --shared=group repo in there, why are some of the directories not accessible by the group?08:17
← keyvan left08:18
→ cirwin joined08:18
Xaphiosis specifically, the permissions are correct, but the group is my own username rather than the group of the directory the repo is in08:18
← drl left08:22
← Schmallon left08:24
mastro cjs, yes but merge commits have 2 or more commits involved, so if you want to see them as "patch" you have to decide if you want to see the patch from parent A or parent B08:25
→ ningia joined08:25
→ selud joined08:25
← selud left08:25
← roop left08:26
← amerine left08:26
cjs Right. And those are two separate arcs in the graph.08:26
← xiong left08:27
→ simoncpu joined08:27
simoncpu hi08:27
what's a good git web ui that i can install on my server?08:27
cjs This is why Git's porcelain is confusing: you do something like look at a particular commit, and it doesn't show you that commit, it shows you the arc between that commit and a previous commit, which is not actually something that exists in the database (except implicitly) at all.08:28
→ guardian joined08:28
cirwin agreed, it's a horribly leaky abstraction08:28
cbreak_work you can look at the commit if you want08:29
you just have to ask nicely :)08:29
cirwin but that isn't a very useful thing to do :p08:29
cjs In fact, to say that "merge commits have 2 or more commits involved" is also confusing. A merge commit is just a single commit, no different from any other snapshot when viewed as a snapshot of work. It's just that there's an automated way to say, "give me a diff between that and snapshot A or snapshot B...."08:30
cbreak_work: Yes, I know. cirwin: I dunno, I've checked out previous commits from time to time. :-)08:30
cirwin that looks at the tree, not the commit :p08:31
mastro simoncpu, have a look here: https://git.wiki.kernel.org/index.php/Interfaces,_frontends,_and_tools08:32
simoncpu mastro: okidoki, thanks08:32
cjs cirwin: Am I wrong, or is a checkout of the tree not pretty much the commit (with the exception of excluding the commit message and similar metadata?)08:33
thiago_home the checkout of a commit is the checkout of the tree associated with the commit08:34
mastro simoncpu, I'd personally go with http://www.redmine.org/ either if it is not git specific nor in that list08:34
cirwin well, you could have the same tree with multiple commits (if you tried reallly hard) at which point a checkout would show you no information08:34
(or multiple commits with the same tree to put it forwards)08:34
mastro simoncpu, http://www.redmine.org/projects/redmine/wiki/Features08:34
cirwin i tend to think of commits as just convenient labels08:34
all the useful stuff they point to08:34
← Thumper_ left08:35
→ Thumper_ joined08:35
mastro cjs, the checkout of a commit is actually just pointing the HEAD to that commit and changing your working directory to resemble that snapshot08:35
→ drl joined08:36
simoncpu whoa, google/stackoverflow has answered my questions08:36
http://stackoverflow.com/questions/1130466/trac-vs-redmine-vs-jira-vs-fogbugz-for-one-man-shop08:36
=)08:36
simoncpu <3 gooooogllllle08:36
thiago_home one-man quartet?08:37
simoncpu yeah08:38
← tyteen4a03 left08:38
→ goshawk joined08:38
mastro cjs, since you HEAD point to that commit you are on THAT commit, you can have multiple commits with the same working directory / same snapshot08:38
← babbage left08:39
→ talespin_kit joined08:39
← marc_kdab left08:39
mastro simoncpu, actually I think the best software around is github, too bad you can't have the source code to install it locally08:39
→ tyteen4a03 joined08:39
→ ppawel joined08:39
talespin_kit how to remove the file only from the index. i tried git rm --cached filename , but it did not work.08:39
← devanjan left08:39
mastro in my opinion this is something really missing with git08:40
→ bradleyayers joined08:40
← tyteen4a03 left08:40
→ bigpeted joined08:40
talespin_kit ie. opposite to git add08:40
thiago_home talespin_kit: that's how you do it08:40
→ tyteen4a03 joined08:40
mastro talespin_kit, git status tell you08:40
→ rgr joined08:41
→ tyteen4a03_ joined08:41
simoncpu mastro: yepp, github is cool08:41
the paid hosting is too high though08:42
← bigpeted left08:42
→ bigpeted joined08:42
← bronson left08:45
→ bronson joined08:45
← tyteen4a03 left08:45
← rgr left08:47
→ nevyn^ joined08:48
→ cenuij joined08:48
← cenuij left08:48
→ cenuij joined08:48
← matjas left08:49
← Xaphiosis left08:49
→ devanjan joined08:50
← theoros left08:51
← krosenvold_ left08:51
→ psoo joined08:51
egrouse just out of interest, how can i reset or revert a remote repo? earlier on i did a reset and attempted to push back to it and got an error - is the some specific method?08:51
← chrisdarroch left08:52
thiago_home what do you mean reset or revert?08:52
what do you want to do? explain that, not what your solution is08:53
→ nicoulaj joined08:53
egrouse rollback to an earlier commit08:53
← drl left08:53
← ruied left08:53
thiago_home do you mean erase the new commits completely, from history, forever?08:54
egrouse yes, that would be ideal08:54
total obliteration08:54
thiago_home do other people have those commits already?08:54
egrouse no08:55
thiago_home then reset --hard to the commit you want, then push -f08:55
egrouse push -f is force i guess?08:55
tango_ yes08:55
egrouse cool.08:55
thanks08:55
→ snearch joined08:56
← mikepack_ left08:57
silex → Silex08:57
→ t3i joined08:58
→ gebi joined09:00
← gebi left09:00
→ gebi joined09:00
← sagsousuke left09:00
→ matjas joined09:00
→ Anubisss joined09:01
← oriba left09:03
→ torvalamo joined09:04
← tyteen4a03_ left09:04
→ mziulu joined09:04
→ drl joined09:05
← bauruine left09:07
talespin_kit mastro: thanks, its git reset.09:07
→ ruied joined09:07
← Anti-X left09:08
talespin_kit mastro: why did the git-rm man page misleads. i see these lines under --cached option, --cached09:08
Use this option to unstage and remove paths only from the index.09:08
→ priidu joined09:09
talespin_kit Working tree files, whether modified or not, will be left alone.09:09
→ davidc_ joined09:09
→ tyteen4a03 joined09:10
→ LouisJB joined09:10
→ metoikos joined09:10
→ krfs joined09:10
← justin-george left09:11
← krfs left09:12
← ruied left09:12
→ sgillet joined09:13
← drl left09:14
→ serard joined09:14
← tyteen4a03 left09:14
→ The_Ball joined09:15
The_Ball What is the easiest way to split up a previous commit into two commits?09:15
selckin get rebase -i, edit, make 2 commits, git rebase --continue09:16
→ marc_kdab joined09:17
← thiago_home left09:18
→ maattd joined09:18
→ tyteen4a03 joined09:19
→ darkredandyellow joined09:19
→ fr0sty|mobile joined09:19
→ ayanes joined09:20
← ayanes left09:20
→ ayanes joined09:20
→ SIGBUS joined09:20
← The_Ball left09:20
← berto- left09:22
→ mahata joined09:23
→ ruied joined09:24
← cirwin left09:25
→ vdv joined09:25
mastro talespin_kit, you should understand what "cache" or "index" means in a git repository: it is the *ready_to_commit* state09:25
→ pfrenssen joined09:25
yaotti → yaotti_aw09:26
← tyteen4a03 left09:26
mastro talespin_kit, git rm --cached is used to remove a file from the index but not from the working directory: the file is removed in the next commit but kept in the filesystem as untracked file09:27
→ drl joined09:27
→ The_Ball joined09:27
mastro talespin_kit, when you git add or git rm you update the index state, you can see it as a "stage" for the next commit09:27
→ dpino joined09:28
← bradleyayers left09:28
mastro talespin_kit, if you change your mind and want to do the reverse of a git add or a git rm you just have to reset the index to the current commit with git reset HEAD <file>09:28
→ sgronblo joined09:29
sgronblo Why does git say that trunk is ambiguous when doing checkout trunk on an svn repository that i cloned with standard layout?09:30
Is it because of the remotes/trunk branch?09:30
selckin you never wanna checkout trunk anyway09:30
→ PigeonFriend joined09:30
selckin unless you renamed master to trunk?09:30
doener sgronblo: yes. refs/remotes/trunk and refs/heads/trunk both have the same valid shortname "trunk"09:31
sgronblo: that's why you should always use something like --prefix=svn/ with git-svn09:32
← drl left09:32
talespin_kit mastro: thanks again, very clear description.09:33
mastro selckin, if you _really_ want to checkout a remote branch (which is not useful anyway) you can do it with git checkout remotes/trunk09:33
→ tyteen4a03 joined09:33
→ zoran119 joined09:34
zoran119 is it possible to 'patch ignore' a file09:34
mastro selckin, sorry, it was mean for Sgeo_09:34
sgronblo,09:34
sorry Sgeo_ :)09:35
zoran119 i edit a config file (say database password) and don't want it to come up in 'git status' and 'git diff', but if i edit someting else i do want it to be reported in 'git status' and 'git diff'09:35
is this possible?09:35
→ kunthar joined09:37
FauxFaux Something else in the same file? No.09:38
zoran119: man git-update-index --assume-unchanged for the whole file.09:38
kunthar hi, i have a question regarding mirror supported fork of gitosis: https://github.com/dinoboff/gitosis.git09:38
anyone used mirror option?09:39
sgronblo Why would you not want to check out trunk?09:39
← SIGBUS left09:39
→ tatsuya joined09:39
sgronblo I think I might have renamed master to trunk09:39
mastro sgronblo, what do you want to do after checking out trunk? what are you trying to achieve?09:39
sgronblo It was a while ago. Time for some merge hell.09:40
→ simplechat joined09:40
← GaVrA left09:41
← kunthar left09:41
→ tvw joined09:42
← tyteen4a03 left09:42
← Neo left09:43
→ Neo joined09:43
→ HG` joined09:44
→ arnis joined09:44
→ drl joined09:46
arnis hi09:47
→ tyteen4a03 joined09:47
← thermal left09:47
← tyteen4a03 left09:47
→ tyteen4a03 joined09:48
← maattd left09:48
← tyteen4a03 left09:48
→ tyteen4a03 joined09:49
← gretch left09:49
→ RaceCondition joined09:49
← RaceCondition left09:51
→ SIGBUS joined09:51
← tyteen4a03 left09:54
← drl left09:54
→ anjen joined09:55
← SIGBUS left09:56
← priidu left09:56
→ maattd joined09:57
→ tyteen4a03 joined09:57
← maattd left09:57
→ maattd joined09:58
← MikhX left09:59
← maattd left10:00
→ maattd joined10:00
→ henrikbjorn joined10:00
→ peshalto joined10:00
← peshalto left10:01
← tyteen4a03 left10:01
→ bluenovember joined10:02
← bronson left10:03
→ bauruine joined10:04
→ bronson joined10:04
← shiba_yu36 left10:06
→ tyteen4a03 joined10:06
→ RaceCondition joined10:07
← bronson left10:07
→ curtana joined10:07
→ bronson joined10:07
← ruied left10:08
→ drl joined10:09
← t3i left10:10
← bousquet left10:12
→ aeiou joined10:12
→ shiba_yu36 joined10:12
aeiou i have a directory of files, (not under git) i also have a directory inside this directory that _is_ a git repo ( http://pastebin.com/vLFQzre5 )10:12
i now want to add the non-git-dir to git, and add the directory-under-git to be a submodule10:13
← funfun left10:13
→ funfun joined10:13
aeiou if i do git init inside non-git-dir what do i do to add teh directory-under-git as a submodule?10:13
malesca aeiou: Probably easiest to just move the inner directory somewhere else. Like, move it to /inner_dir, then do as you would with any other submodule10:15
IIRC something like: git submodule add /inner_dir where/it/should/go/inner_dir10:16
aeiou yeah that is a pain though, because the inner_dir has lots of untracked files10:16
i.e cache files, uploads etc10:16
thus i was wondering if it is possible to add a subdirectory manually, using the repo that is already there?10:16
malesca well, if it has gitignores and you add it as a submodule, I don't think that should be a problem10:16
aeiou i guess i could do: mv inner inner_bak; git submodule add x inner; rm -rf inner; mv inner_bak inner10:17
?10:17
if i knew anything about submodule internals i would guess this is possible without this hack10:17
← dpino left10:18
yaotti_aw → yaotti10:18
→ robzor joined10:18
malesca aeiou: What is x? Some remote repo?10:18
aeiou yeah10:18
[email@hidden.address]10:19
(ssh)10:19
malesca Oh, now I get you. So you want the cache files etc to be around afterwards. Yeah, what you said seems like it should work. I don't know submodules intimately. I'm sure someone will suggest something cooler if you stick around ;)10:19
← kpreid left10:21
→ chrischambers_ joined10:21
aeiou malesca, ah yeah cool :)10:21
i think it will work, just feels hacky :010:22
← chrischambers left10:22
chrischambers_ → chrischambers10:22
→ drizzd_ joined10:22
→ ruied joined10:24
← drizzd left10:26
→ j416 joined10:26
→ felixge joined10:26
→ airborn joined10:26
felixge Question: Sometimes I see a hunk I don't want to keep when using "git add -p". Is there any way to not only ignore the hunk, but rather remove it from my file completely?10:27
j416 o/ if I create a branch with one initial commit, then in a copy of that branch I change something and commit one; how come when I merge back this copy into the first branch, there are suddenly two identical "Initial commit"s ?10:28
← ruied left10:29
j416 felixge: I think not, since add is only about adding things to the index, not modifying your working copy. If you commit however, you can use git checkout or git reset to remove other changes.10:29
felixge j416: well, the problem is that while I want to "reset --hard" on most changes not staged after the commit, sometimes I want to keep a few hunks :)10:31
← gebi left10:31
← Neo left10:31
felixge it always felt like telling git add -p to "nuke" the hunk in question would feel most natural to my workflow10:31
but I guess I'm out of luck10:31
:)10:31
j416 felixge: well, not adding it means it's not going into your commit10:32
Weust`afk → Weust`10:32
j416 when you're done running git add you can run git diff to see what you didn't add10:32
felixge j416: Yeah, that's what I have been doing so far. And git reset/checkout works great if I want to nuke all remaining hunks in the tree / a single file. But if I want to nuke a particular hunk in a file without commiting the others, I have to use my text editor :(10:33
← devanjan left10:33
felixge I guess if I could tell 'git stash' to ignore my index, life would be good10:34
then I could stash the bad / good stuff away10:34
→ Neo joined10:34
← hobodave left10:35
j416 felixge: git stash save --keep-index10:35
felixge I guess I'm in luck10:35
yeah10:35
just saw that10:35
:)10:35
j416: thanks for the help!10:35
j416 you're welcome. :)10:35
felixge j416: that will help my flow :)10:35
<3 gi10:35
* git10:35
j416 felixge: so optimally you would like a git reset -p option?10:36
felixge j416: yes10:36
→ MikhX joined10:36
felixge j416: but I just saw there is git stash -p10:36
j416 well there is one. :)10:36
(I just found out myself)10:37
felixge ohhhh10:37
that's even much more perfect10:37
doesn't get my stash dirty10:37
:)10:37
j416 so you should be able to add everything, then reset just that hunk10:37
I guess.10:37
felixge hmm10:37
j416 thanks, I learnt something new.10:37
felixge nah doesn't work that way10:37
git reset -p --hard does not work10:38
fatal: --patch is incompatible with --{hard,mixed,soft}10:38
← mahata left10:38
felixge but git stash -p will work just fine10:38
j416 not --hard10:38
felixge well, I want hard10:38
I want to modify my files10:38
j416 what do you need that for10:38
felixge not the index10:38
Scenario: I have 1 file with 2 changes10:38
and I want to remove one change from the file10:39
j416 git isn't a text editor.. :P10:39
felixge while keeping the other10:39
j416: no, but it applies and reverts patches10:39
j416 which apply to entire projects :)10:39
← hgb left10:39
← roger_padactor left10:39
← opalka left10:40
felixge so you're telling me I should not embrace my awesome discovery of git stash ?10:40
)10:40
:)10:40
j416 I see your point, but having git editing workdir files like that sounds rather dangerous10:40
felixge git stash -p10:40
that is10:40
→ kpreid joined10:40
j416 do whatever works best for you. :)10:40
felixge well in my scenario I usually just want to nuke debugging statements while keeping one or two temporary changes for testing10:41
← matjas left10:41
felixge so even if I loose all dirty files, no biggie10:41
j416 felixge: you could define your own command that does git stash -p then immediately drops that stash10:41
?10:41
felixge j416: sure, probably won't even need it10:41
I mean I don't care about my stash being overwritten10:41
it has a history10:41
and my stashes are shortlived anyway10:42
j416 felixge: why not have a debug branch for such things that you rebase on top of your dev branch?10:42
might be overkill10:42
felixge j416: yeah, it'd be more manual work in my case10:42
j416 ok.10:42
→ opalka joined10:42
felixge j416: the real problem is that the project I work on keeps the config under version control without having a non-versioned config that can be merged over the main one10:43
← hyperair left10:43
← boombatower left10:43
felixge ;)10:43
j416 hm ok10:43
anyway10:43
time for lunch10:44
o10:44
o/10:44
← j416 left10:44
→ verdon joined10:45
→ gfto joined10:45
→ tcurdt joined10:47
← maattd left10:47
→ matjas_ joined10:50
→ mabrand joined10:50
← DaveIngram left10:50
→ _pingu joined10:52
← Thell left10:53
→ Titosemi1 joined10:53
→ devanjan joined10:53
→ avinashhm joined10:53
← bran left10:54
← Titosemi left10:54
→ petrux joined10:54
→ trasz joined10:54
trasz git-p4 submit10:55
fatal: Not a valid object name HEAD~18510:55
Command failed: git cat-file commit HEAD~18510:55
what might be wrong?10:55
→ krfs joined10:55
→ k04n joined10:58
→ fr0sty-away_ joined10:58
k04n quick question...i just initialized a new repo on server (git init), then on my local machine cloned it via ssh. then, i added all local files and did git push origin master10:58
← fr0sty-away left10:58
← drl left10:59
k04n everything was uploaded, and on ther server it even looks like files are in .git... git log shows my commit10:59
(forgot to mention i committed between add and push origin master)10:59
doener faq non-bare10:59
→ gebi joined10:59
k04n however, on the server, how do i get the files out of the .git directory?10:59
doener hm, ok, we're apparently out of faq bots...10:59
k04n do i have to clone it somewhere else or is there a way to make the server copy function similarly to local10:59
→ kkszysiu joined11:00
kkszysiu Hello11:00
How to merge just some revisions from one branch to another?11:01
fr0sty|mobile trasz: there is no commit 185 generations before HEAD.11:01
but i don't know git-p4 at all11:01
doener k04n: https://git.wiki.kernel.org/index.php/GitFaq#non-bare11:01
malesca kkszysiu: git cherry-pick11:01
→ jast joined11:02
k04n doener: awesome. got it :) thanks.11:02
← Jippi_mac left11:04
trasz fr0sty|mobile: could you explain the 'there is no commit'?11:05
← anjen left11:05
← ketas left11:09
→ krosenvold_ joined11:10
← devanjan left11:10
→ ketas joined11:10
→ tizzo joined11:11
→ drl joined11:12
fr0sty|mobile your history is not that deep11:12
← VladDrac left11:12
← psankar left11:13
← k04n left11:13
→ VladDrac joined11:13
→ sonnym joined11:13
fr0sty|mobile why that is or whether that is a problem, I can't say.11:14
→ ruied joined11:14
→ Jippi_mac joined11:14
← bronson left11:14
→ bronson joined11:15
← _pingu left11:16
← talespin_kit left11:16
fr0sty|mobile trasz: there seem to be some good questions about this issue on stackoverflow.com11:18
that might be of some use.11:18
→ _Vi joined11:19
← ruied left11:19
← drl left11:19
→ hobodave joined11:21
arnis can someone tell me in few words what is branch and whats the point of it?11:22
→ priidu joined11:23
→ hyperair joined11:24
→ Ryback_ joined11:25
→ leo2007 joined11:25
bremner arnis: man gittutorial11:25
jast arnis: the 'gittutorial' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/gittutorial.html [automatic message]11:25
← radarek left11:25
squi i think there is a spam bot here11:25
08:56:15][ <marcamilly> hey, how'd you get voice just now ?11:25
08:56:19][ <marcamilly> did you use the link ?11:25
→ bradleyayers joined11:26
← bronson left11:26
→ bronson joined11:27
← snearch left11:28
← LouisJB left11:29
→ drl joined11:31
← hyperair left11:34
← felixge left11:34
→ Puck` joined11:34
Puck` hi everyone11:34
could anyone recommend an irc bot to get updates to a channel?11:34
selckin cia11:35
→ koltroll joined11:35
koltroll mac.se11:35
bremner or kgb, if you want to host it yourself11:36
Puck` thank you (:11:36
koltroll bah. that wasn't intended to get printed out.11:36
← kreemer left11:37
← Neo left11:39
→ Neo joined11:40
← drl left11:42
← bradleyayers left11:42
← Thumper_ left11:45
← bronson left11:45
→ bronson joined11:45
← Puck` left11:47
→ newbiz joined11:47
← picasso left11:51
→ zorzar joined11:51
→ alin-gym joined11:52
← zorzar_ left11:54
→ drl joined11:54
→ marcob joined11:54
→ JohnFord joined11:55
→ Thumper_ joined11:58
→ JoeCoolDesk joined11:59
← JoeCoolDesk left11:59
→ xanderbeedle joined11:59
← maleknet left12:00
← banisterfiend left12:00
→ banisterfiend joined12:00
→ maleknet joined12:01
← bronson left12:03
→ bitkiller joined12:03
→ bronson joined12:03
← sudonotplease left12:03
← drl left12:03
→ sudonotplease joined12:04
→ stamina joined12:05
→ spaceonline joined12:05
→ bradleyayers joined12:06
← bradleyayers left12:06
← ch077179 left12:06
→ hyperair joined12:07
→ lucs joined12:07
← bronson left12:07
→ bronson joined12:08
→ orafu joined12:09
→ tty2 joined12:10
→ bentob0x joined12:12
→ vitor-br joined12:12
← priidu left12:13
← sudonotplease left12:13
← tty1 left12:13
→ sudonotplease joined12:13
← mishok13 left12:14
← rns left12:15
→ priidu joined12:15
→ ExtraSpice joined12:15
→ mishok13 joined12:16
← bauruine left12:16
← SirFunk_ left12:17
→ drl joined12:17
← tizzo left12:18
← davidc_ left12:20
→ davidc_ joined12:20
← tyteen4a03 left12:21
← jony left12:21
← banisterfiend left12:22
→ banisterfiend joined12:22
→ telemachus joined12:22
← drl left12:22
→ ch077179 joined12:23
→ tyteen4a03 joined12:27
← mikemeyer left12:27
→ Yuuhi joined12:30
← EricInBNE left12:30
→ JohnFlux joined12:30
→ ruied joined12:31
← kpreid left12:31
JohnFlux git commit --interactive .12:31
fatal: Paths with --interactive does not make sense.12:31
It makes perfect sense !12:31
why can't I specify some paths with interactive?12:31
← tyteen4a03 left12:33
→ rns joined12:34
→ ramusara joined12:34
→ dpino joined12:37
→ felixge joined12:38
← felixge left12:38
→ felixge joined12:38
← JohnFord left12:38
→ MattDiPasquale joined12:38
→ kpreid joined12:38
→ tyteen4a03 joined12:39
→ picasso joined12:40
matjas_ → matjas12:40
→ magnet joined12:42
→ arialdo joined12:42
arialdo When I have a branch and a tag with the same name (ambiguousname), I can checkout the tag with git checkout refs/tags/ambiguousname. Is it the right way? Why does git checkout refs/heads/ambiguousname take me to a detached state?12:44
kpreid detached = not on a branch12:44
arialdo kpreid: yes. That's my doubt. My question is: how can I absolutely refer to a branch and checkout it, using something like I use with tags, in git checkout refs/tags/ambiguousname?12:45
← magnet left12:46
arialdo kpreid: I thought that git checkout refs/heads/somebranch was equivalent to git checkout somebranch. But the first takes to detached, the second one to the right branch. I'm a bit confused12:47
kpreid hm, I see12:47
→ radarek joined12:48
kpreid Well, the manual contradicts the behavior you're seeing; it claims that you get the branch if there is a branch, else detached.12:48
(but that may be bad writing...)12:48
arialdo kpreid: I know that git checkout ambiguousname takes me to the branch, and not to the tag with the same name, but it prints out a warning. Then, I suppose there's a clearer way to impose the checkout of a branch.12:48
→ lundtor joined12:49
arialdo kpreid: see what I mean? https://gist.github.com/80580712:51
← shentino left12:51
arialdo kpreid: git checkout ambiguous prints a warning, but checkout a branch12:52
← _Vi left12:52
arialdo kpreid: git checkout refs/heads/ambiguous takes me to a detached state12:52
→ kukks joined12:52
← maleknet left12:52
arialdo kpreid: may be I'm missing something... I'm not that expert12:52
→ ded joined12:54
arialdo kpreid: git checkout master^0 also gets you a detached HEAD12:54
→ pmabie joined12:54
← lucs left12:54
fr0sty|mobile arialdo: git --version ?12:55
→ lucs joined12:55
← tyteen4a03 left12:55
ded I want to do something like 'git archive HEAD:dir/0002* > snapshot.tar' i.e. restrict the files going into the archive according to a pattern/glob. How?12:55
→ mahata joined12:56
bremner ded: man git-archive12:57
jast ded: the 'git-archive' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-archive.html [automatic message]12:57
→ roger_padactor joined12:57
→ wsgndh joined12:57
← krfs left12:58
bremner ded: is particular, check the synopsis12:58
s/is/in/12:58
ded bremner, jast: what makes you think I would ask here without looking at the manpage first?12:58
bremner because the answer is on line 4 of the man page?12:58
jast ded: what makes you think that the "[automatic message]" at the end of what I said indicates that I was speaking to you myself?12:59
arialdo fr0sty|mobile: 1.7.112:59
jast anyway. basically, git archive HEAD:dir '0002*'12:59
bremner err, no :12:59
ded jast, thankyou12:59
jast (does it support globbing there?)12:59
→ d0m joined13:00
→ tyteen4a03 joined13:00
bremner oh, I think just git archive HEAD dir/002* ?13:00
arialdo fr0sty|mobile: Do you have an idea which can help me?13:00
jast depends on what you want the base path of the archive to be, I suppose13:00
arialdo fr0sty|mobile: In the meanwhile, I'm reading http://sitaramc.github.com/concepts/detached-head.html13:00
← verdon left13:01
jast I would assume (haven't tested it) that HEAD:dir 0002* creates a tarball that contains 0002*/ directly, whereas HEAD dir/0002* gives you entries like dir/0002*/13:01
xanderbeedle Is there a bash method that returns true / false depending on whether the current dir is a git folder or not?13:03
fr0sty|mobile arialdo: i think you are stuck with the 'ambiguous' warning if you want non-detached state.13:03
ded jast, bremner: I'm in a bare repo in case that's relevant13:03
→ m_3 joined13:04
← sudonotplease left13:04
fr0sty|mobile xanserbeedle: git rev-parse might help you.13:04
jast xanderbeedle: git rev-parse --is-inside-*13:04
arialdo fr0sty|mobile: Ok. It's ok for me. I just wanted to know if there's a better way. If it's the right way, I'll adopt it. Thank you!13:04
→ d0k joined13:04
xanderbeedle jast: that returns a fatal error when not inside git folder13:05
jast: nevermind, i'll catch the error or something instead13:05
→ sudonotplease joined13:05
jast xanderbeedle: 2>/dev/null :)13:05
← nixness left13:06
xanderbeedle jast: no idea how to use dev null (bash newbie)13:06
I'll work it out, thanks!13:07
ben_alman how can i get a diff of the current folder state, including un-added files, against the latest commit?13:07
jast xanderbeedle: just append that to the command to suppress the error output13:07
cbreak_work ben_alman: git diff does not care about untracked files13:08
ded bremner: 'git archive HEAD dir/002*': I'm in a bare repo so that glob matches nothing.13:08
ben_alman cbreak_work: can it?13:08
cbreak_work no13:08
jast ded: the idea is that git matches the glob rather than the shell13:08
cbreak_work you can do git diff HEAD .13:08
→ vaishali joined13:08
jast ded: if you shell complains, quote the argument13:08
cbreak_work but that will compare only tracked files13:08
jast ded: if it still doesn't work, git probably doesn't interpret globs in that command13:08
cbreak_work if you want untracked, just add them :)13:08
git archive wants a treeish13:09
not a shell glob13:09
← pmabie left13:09
cbreak_work hmm... but if the shell evaluates those, it should work13:10
← ruied left13:10
ded cbreak_work, right but I'm in a bare repo so no checked out dirs13:10
jast cbreak_work: it wants a treeish PLUS paths13:10
(paths are optional)13:10
cbreak_work indeed13:10
jast the idea was that maybe the paths use globs13:10
xanderbeedle jast: but i want a response if the command succeeds13:11
jast I haven't checked13:11
xanderbeedle only trying to avoid the error13:11
← arialdo left13:11
← trasz left13:11
jast xanderbeedle: that 2>/dev/null only suppresses the error message, not the actual output on success13:11
xanderbeedle oh, i see13:11
jast look up stdout and stderr to see why13:11
I think even Wikipedia explains them :)13:11
→ lanthan_afh joined13:13
ded jast, cbreak_work: I'm still stuck: it seems (a) we don't know how to create a treeish pointing at a set of files matching a glob, and (b) I can't use a globbed path because I'm in a bare repo. Any advance on that?13:13
→ sivy joined13:14
ded bremner: where on line 4?13:14
← hyperair left13:15
cbreak_work ded: a treeish is a tree13:15
no glob support13:15
you can just mention all the paths you want in13:15
like, manual shell glob evaluation13:15
→ LouisJB joined13:16
→ agile joined13:16
ded cbreak_work: right, thanks. I'll try that approach.13:17
→ _pingu joined13:18
_pingu How can I push a new branch to a bare repo?13:18
tried git bush origin test, but got: error: failed to push some refs to ...13:19
s/bush/push13:19
→ jaql joined13:20
← normanrichards left13:20
← miloshadzic left13:20
→ blahdeblah joined13:20
→ Mystique joined13:20
→ nixness joined13:20
→ envi_ joined13:21
← envi left13:21
jast _pingu: try git push origin test:test13:21
Mystique hey all, is there a totaly newbie guide to git somewhere? I have something that is now in git format that I am trying to update; and I'm more than a little lost..13:21
jast Mystique: man gittutorial @13:22
Mystique: the 'gittutorial' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/gittutorial.html [automatic message]13:22
assuming you haven't found that yet :)13:22
it covers several different workflows13:22
Mystique I have not, thank you..13:22
jast look at "using git for collaboration"13:22
Mystique so, trying to figure out what I did wrong..13:23
blahdeblah Hi. I'm trying to create a local branch which tracks a remote branch. When i run 'git checkout --track origin/next', i get 'fatal: git checkout: updating paths is incompatible with switching branches.' What's up? Does my local repo not know about all the branches on origin?13:23
← tyteen4a03 left13:24
Mystique I'm running 5.1.1 ; did git status saw 5.1.2 and 5.213:24
blahdeblah Mystique: I've been using http://progit.org/book/ - seems pretty good13:24
← squi left13:24
Mystique did git checkout 5.1.2 and tried to do a git pull13:24
← nixness left13:24
Mystique and git update..13:24
neither of which did what I thought it would do..13:25
→ jfkw joined13:25
← MattDiPasquale left13:25
RandomSort A workflow/git question, I'm trying to start using git on the CMS my work is maintaining/developing13:25
There are a central solution that gets customized and developed, I'm interested in begin able to develop custom modules and merging them back in the master branch as well as using different templates for different solutions13:26
_pingu jast: the feedback looks fine, but the I cannot see the new branch after a new git clone.13:26
← lucs left13:26
← koo5 left13:26
→ dc5ala joined13:26
RandomSort that is have some files that will be merged back in the master branch while others are preserved from the master branch, is this possible?13:27
jast _pingu: git clone creates only one local branch, but it still clones everything hidden in the background; you can see all cloned remote branches using: git branch -r13:27
RandomSort: merges are all-or-nothing. if, in your side branch, you only change the files that will be merged back eventually, that will work. if you change other files, too, a simple merge won't cut it anymore.13:28
→ tyteen4a03 joined13:28
jast Mystique: you probably checked out a tag there. tags are not meant to be worked on directly; consider working off of a branch instead. use 'git branch -r' to see a list of existing branches, then tell me which git version you're using so I can give you the simplest command to create a new branch off of one of those branches. :)13:29
RandomSort jast would it be possible to have a folder ignored?13:30
then I could just have a folder for the "site specific" files13:30
_pingu jast: ok, thanks. Is there an option in git clone, that I can use git branch to show all branches like ever?13:30
jast RandomSort: no. generally, site-specific files should not be in the repository in the first place.13:30
_pingu: you mean so that git clone automatically creates local branches from all remote ones?13:30
_pingu jast: yes13:31
→ Mannequin joined13:31
RandomSort okay, thanks jast, I'm back to the drawing board then :)13:31
← blahdeblah left13:31
jast _pingu: I'm afraid not... but you could use this command to create all of them at the same time after the clone: git push . refs/remotes/origin/*:refs/heads/*13:31
it's a bit of an ugly hack, though :)13:31
RandomSort: however, you could simply never commit changes in that directory. requires a bit of discipline, though.13:32
_pingu jast: it's ok, I have to learn... . thanks13:32
← sudonotplease left13:32
RandomSort jast, both discipline and refraining from using commit -a which I'm not confident will be a success :P13:32
← tyteen4a03 left13:32
→ tizzo joined13:33
Mystique jast: http://slexy.org/view/s2TIaALLF813:33
jast RandomSort: you could also create a hook that automatically strips out changes in particular files from all new commits as they are made13:33
Mystique: ah. what version of git are you using?13:33
RandomSort jast, that seems interesting, got a link where I can enknowledge me on that topic?13:33
Mystique 1.7.2.2 (freebsd ports from whenever ago..)13:33
jast Mystique: ah, right. seems like everything should have been okay then. so, what did you expect git pull to do?13:34
Mystique well I was on 5.1.1 (which I got using git clone path/to/file.git) and then I was trying to 'update my branch to 5.1.2' and then pull the local branch up to the 5.1.213:35
← Pewpewarrows left13:35
jast RandomSort: basically, create an executable shell script in .git/hooks/pre-commit that executes something like "git reset -- <list of paths>" to automatically unstage all changes in those paths prior to committing. the hooks are documented at http://git.or.cz/man/githooks13:35
→ ldh joined13:36
jast Mystique: git checkout switched your working tree to a different branch, so "git pull" didn't really do anything because that branch was completely up-to-date :)13:36
RandomSort thanks very much jast13:36
Mystique hmm, ok :/ so how do accomplish what I was trying to do? how do I update my current branch then..13:36
_pingu git fetch origin; git status says: Your branch is behind 'origin/hello' by 1 commit, and can be fast-forwarded. how do I resolve that?13:37
→ tyteen4a03 joined13:37
jast Mystique: so, you were on 5.1.1, right? did you make and commit any changes?13:37
_pingu: git pull13:37
Mystique jast: it's readonly to me13:37
jast _pingu: or git merge origin/branchname, since you've already fetched13:37
Mystique: uh, what? your local copy is read-only?13:37
Mystique my remote copy, I'm not allowed to push into only allowed to pull from..13:38
_pingu ok13:38
jast Mystique: yes, but you can commit locally, can you not?13:38
Mystique I can commit locally yes,13:38
jast which is what I was asking about13:38
Mystique but I'm not doing that..13:38
jast so have you made any changes to 5.1.1 yet?13:38
Mystique I'm a 'user' of this code branch13:38
← metoikos left13:38
← cjs left13:39
→ shentino joined13:40
← shentino left13:40
→ shentino joined13:40
← d0m left13:40
→ rp2 joined13:41
← tyteen4a03 left13:41
← _pingu left13:42
← petrux left13:42
→ nothingHappens_ joined13:42
→ angelsl joined13:43
rp2 git newbie question: I have made a local git repository clone, now how do I extract the code base with a particular tag?13:44
avar git checkout that-tag13:44
→ koo5 joined13:44
rp2 where will git put it?13:45
hmm13:45
→ Dave^| joined13:45
avar everywhere13:45
→ tyteen4a03 joined13:45
fr0sty in your working copy.13:45
← ben_alman left13:45
rp2 but the working copy already has the latest codebase13:45
i don't want to replace that13:46
← tyteen4a03 left13:46
→ tyteen4a03_ joined13:46
fr0sty what _do_ you want to happen, then?13:46
← bronson left13:47
← Mystique left13:47
→ bronson joined13:47
← spaceonline left13:48
→ sudonotplease joined13:48
rp2 I don't know, because I'm still learning how git works13:48
→ normanrichards joined13:48
fr0sty rp2: if you checkout the tag, you can checkout the 'latest' again13:49
you can switch between whatever versions you want.13:49
← Textmode left13:49
rp2 i guess i need to learn the ifference between clone and checkout13:49
checkout appears to be some sort of switch, it updates the files to a particular branch or tag13:50
← tyteen4a03_ left13:51
rp2 hm13:51
fr0sty you only clone once. That gets you every commit, tag and branch into your local repository.13:51
rp2 i guess i can just cp -r the clone i made and then run git checkout on the copy13:51
fr0sty that works.13:52
rp2 that way i have one copy with the latest codebase and one with the codebase with the given tag13:52
i wanted to avoid pulling everything over the network twice13:52
→ tizbac joined13:53
rp2 ok thanks i'll try that13:53
← tvw left13:54
← harshu_ left13:54
← harshu___ left13:54
→ pmabie joined13:54
pmabie Hello to all13:55
→ rseifert joined13:55
→ tyteen4a03 joined13:57
← tizbac left13:57
pmabie Here's what I am try to do , I have a git repo on a mounted partition , I am user clone to copy it to my home . Know where I make a check to the clone I can't push it back to the master13:58
trying13:58
thanks anyway13:59
← pmabie left13:59
→ coderdad joined14:00
→ chittoor joined14:00
→ tizbac joined14:01
← chittoor left14:01
→ petrux joined14:01
→ ben_alman joined14:02
← tyteen4a03 left14:02
← OOPMan left14:02
← avinashhm left14:04
→ t2gs joined14:04
← Yuuhi left14:04
← eean left14:05
t2gs how can I go back 10 commits on my master branch to check if my code was working and then come back to where I currently am?14:06
→ tyteen4a03 joined14:06
← tyteen4a03 left14:07
→ tyteen4a03 joined14:07
selckin git checkout <sha1>; make; git checkout master14:07
← tyteen4a03 left14:07
→ tyteen4a03 joined14:08
→ JohnFord joined14:08
← henrikbjorn left14:08
← teweWork left14:09
t2gs selckin so it's "like" checking out to a branch?14:09
→ jrmuizel joined14:10
← sudonotplease left14:11
→ sudonotplease joined14:11
← bronson left14:12
→ Pewpewarrows joined14:12
→ bronson joined14:12
selckin will put you on a detached head, so don't commit there etc14:12
you can also create a real branch from the point if you want14:13
→ ruied joined14:13
t2gs selckin so I can checkout -b back in time and be a happy child?14:14
← kpreid left14:14
selckin sure14:14
t2gs another thing, I'm using gitk but it just sucks on UI, how do I get the sha1? :P14:14
selckin git log?14:14
t2gs thanks14:15
→ montylounge joined14:15
selckin sha1 is in the middle somewhere in gitk14:15
opalka → opalka_away14:15
dwmw2_gone → dwmw2_HKG14:15
→ rsewill joined14:15
t2gs selckin cool, log worked, thanks a lot14:16
→ ilogger2 joined14:27
→ _ollie joined14:28
→ MrSamuel_ joined14:31
MrSamuel_ I have a sub-directory of files, some have been changed, some deleted, I want to reset to head14:31
→ DiGitalX joined14:31
MrSamuel_ only this sub directory14:31
how do I do it?14:31
→ charlietanksley joined14:32
t2gs you mean undo latest commit?14:32
MrSamuel_ no14:32
revert the chages14:32
Pieter git checkout -- path/to/dir14:32
t2gs or forget that mess you just did?14:32
MrSamuel_ revert the changes14:32
well actually it isn't a mess14:32
Pieter or, git checkout HEAD -- path/to/dir14:32
t2gs MrSamuel_ I was ABOUT to ask that same question o_o14:32
MrSamuel_ the library I updated was broken14:32
→ avinashhm joined14:32
t2gs Pieter what do you mean pathtodir?14:33
Pieter t2gs: the name of the subdir you want to revert14:33
t2gs Pieter say I wanna revert it all14:33
→ ded``` joined14:33
t2gs git add -A; git checkout HEAD14:33
Pieter git reset --hard14:33
t2gs ?14:33
oh I see14:34
no commits changed by that, right?14:34
Pieter nope14:34
just your working copy14:34
→ fenrig joined14:37
→ JohnFord joined14:38
→ tyteen4a03 joined14:38
fenrig If I want to set up a GIT server what guide and program do I best use (gitosis and gitolite don't seem to create a existing repo for some reason)14:38
existing => new (sorry)14:39
→ bronson joined14:39
→ carli joined14:39
spb if you are sensible, you use one of those two14:39
fenrig Then I need a good guide cause I've got a lot of troubles setting it up :( (btw is it better to set up git using ssh or with http?)14:41
← tyteen4a03 left14:42
Pieter both are fine nowadays, but http might be a bit more work14:42
if you want user accounts, you'll also need ssl etc14:42
fenrig Pieter: is there a good howto for setting git up using http and ssl?14:43
→ bronson_ joined14:43
Pieter no14:43
← bronson left14:43
bronson_ → bronson14:43
→ tyteen4a03 joined14:43
→ sudonotplease joined14:44
Pieter well, you can take a look at http://anders.com/cms/384/git/https/RPC.failed.result.3514:44
← carli left14:44
← tyteen4a03 left14:44
→ tyteen4a03 joined14:44
Pieter and also, man git-http-backend14:44
jast the 'git-http-backend' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-http-backend.html [automatic message]14:44
← ded``` left14:45
→ hyperair joined14:45
→ berto- joined14:47
→ panike joined14:47
→ drl joined14:48
→ t0rc joined14:48
← t0rc left14:48
→ t0rc joined14:48
← tyteen4a03 left14:48
→ dhruvasagar joined14:51
→ masonkramer joined14:52
→ tyteen4a03 joined14:52
→ mahata joined14:53
→ rudi_s joined14:54
→ spaceonline joined14:54
← drl left14:56
← tyteen4a03 left14:57
→ dpino joined14:57
→ d0m joined15:00
← d0m left15:00
→ tyteen4a03 joined15:01
→ RJD22 joined15:01
→ skoop joined15:02
RJD22 does someone know if there is a way to automate the login procedure for a git-smart-http request in bash?15:02
→ Swimming_bird joined15:05
← tyteen4a03 left15:06
← masonkramer left15:07
→ bartek joined15:07
→ PerlJam joined15:07
→ drl joined15:09
bartek Hi. This might be going too far but can I use git cherry-pick in such a way that I pick out a single commit, but additionally pick out select files from that commit? Or .. perhaps split that commit on the branch into two commits (possible?) and then cherry pick the right one15:09
→ berserker joined15:09
→ tyteen4a03 joined15:09
→ tvw joined15:10
→ xiong joined15:11
→ pdkl joined15:11
→ gozoner joined15:11
ChanServ set mode: +v15:12
_rane cherry pick it without committing15:12
you'll have the diff then, and you can do whatever you please with it15:12
i bet it's git cherry-pick -n15:12
or --no-commit15:12
cbreak_work or... both!15:12
Pieter or cherry-pick it, then ammend the commit to whatever you want it to be15:12
← avinashhm left15:13
patrikf cbreak_work: actually, either ;-)15:13
← _ollie left15:15
→ jrmuizel joined15:15
← fenrig left15:17
← sudonotplease left15:17
← berserker left15:17
→ sudonotplease joined15:18
→ Helios joined15:19
← pdkl left15:20
→ bauruine joined15:21
← dpino left15:22
→ priidu joined15:23
← gozoner left15:23
→ gozoner joined15:23
→ Dave^| joined15:25
← drl left15:25
→ juangiordana joined15:27
→ MattDiPasquale joined15:27
→ jankoprowski joined15:27
→ Error404NotFound joined15:29
→ nona joined15:31
→ avinashhm joined15:31
← jrmuizel left15:34
→ normanrichards joined15:35
← tyteen4a03 left15:35
→ stringoO joined15:36
→ timholum1 joined15:37
→ dfr|mac joined15:37
→ E-shark joined15:37
→ drl joined15:38
→ alvaro_o joined15:38
→ tyteen4a03 joined15:39
→ slowneutron joined15:39
→ jxpx777 joined15:41
← MrSamuel_ left15:42
→ Sgeo joined15:45
rmk Is there a way to pull one file and all its revision history from one repo to another?15:46
Assuming the file does not exist in the destination repo15:46
→ fowlduck joined15:47
← sudonotplease left15:47
→ bhosie joined15:47
→ alberto56 joined15:50
← stringoO left15:50
→ sebersole joined15:50
← sebersole left15:50
charon rmk: yes and no. you cannot just merge it in the git-pull sense. you can however use filter-branch to create a reduced history that only cares about this file, and then merge that history.15:51
→ beatak joined15:51
→ sebersole joined15:51
rmk I see -- ok, thanks15:51
charon rmk: the fastest way to do the filter-branch is to use an index filter like: git ls-files -s filename | GIT_INDEX_FILE=tmpindex git update-index --stdin --add && mv -f tmpindex $GIT_INDEX_FILE15:53
or some such15:53
(i.e., i haven't tested it)15:53
→ ceej joined15:55
→ pgib joined15:56
→ jnoos4 joined15:57
→ rcu joined15:58
→ d0m joined16:00
→ kpreid joined16:00
→ Tac joined16:00
← avinashhm left16:01
← slowneutron left16:02
→ ivan-kanis joined16:02
ivan-kanis say i have a local branch foo that tracks a remote a remote branch foo16:03
when i do a 'git pull origin foo' it doesn't do what I expect16:03
→ trobrock joined16:03
→ sonnym joined16:03
ivan-kanis git status tells me I am ahead of origin/foo by one commit16:04
→ alex__c2022 joined16:05
PerlJam ivan-kanis: you've got a commit in your local repo that you haven't pushed upstream yet.16:05
→ dcamp joined16:05
→ ROBOTARMY joined16:05
→ Linc joined16:05
→ flazz joined16:07
→ iAsterisk joined16:08
→ alester joined16:08
→ bkosborne joined16:09
ivan-kanis i haven't made any commmit, i am trying to synchronise with the server16:09
→ cvk_cvk joined16:09
→ jfkw joined16:09
→ mikesusz joined16:10
← trobrock left16:10
← normanrichards left16:11
ChanServ set mode: +v16:11
← drl left16:12
← dfr|mac left16:13
→ SirFunk_ joined16:18
ChanServ set mode: +v16:18
→ jrmuizel joined16:19
→ normanrichards joined16:22
← jnoos4 left16:22
→ setmeaway1 joined16:23
← setmeaway1 left16:23
→ setmeaway joined16:23
→ tjholowaychuk joined16:23
→ }dec3pti0n{ joined16:23
ivan-kanis git pull outputs : * branch foo -> FETCH_HEAD16:24
}dec3pti0n{ hi I got a small svv repo with some tags and I would like to copy that to github (including the log history) I've been reading about git-svn and git2svn , which should be better ?16:25
→ resmo joined16:25
→ drl joined16:25
→ hagebake joined16:26
← cvk_cvk left16:26
→ sagsousuke joined16:26
← rcu left16:27
← bauruine left16:28
→ teweWork joined16:29
← }dec3pti0n{ left16:29
RJD22 → RJD22|away16:30
→ dr_lepper joined16:32
→ rcu joined16:33
ivan-kanis i have narrowed down the problem to a branch that's never been pushed16:33
← bkosborne left16:34
ivan-kanis for some reason the sha1 of remotes/origin/foo is not updated16:34
→ bentob0x joined16:34
→ jfelchner joined16:35
← panike left16:35
→ ISF joined16:35
→ hagabaka joined16:35
← hagabaka left16:35
→ hagabaka joined16:35
← hagebake left16:36
→ Guest78064 joined16:37
→ tototot joined16:38
→ bcardarella joined16:39
ivan-kanis it easy to reproduce16:39
create repo1 add a branch, commit in the branch16:40
do a clone of repo1 to repo216:40
in clone1 check in a file and commit16:40
← bcardarella left16:40
← Guest78064 left16:41
ivan-kanis now go to clone2, do a "git pull origin branch" then git status16:41
→ RobertBerger joined16:42
ivan-kanis it show one commit ahead, git branch -rv shows that origin/foo is one commit too early16:42
← gozoner left16:42
ivan-kanis doing a git push magically fixes things... Have I found a bug?16:42
→ gozoner joined16:43
→ hagebake joined16:44
← hagebake left16:44
→ hagebake joined16:44
← sebersole left16:44
← hagabaka left16:44
→ ruied joined16:45
← drl left16:46
→ hobodave joined16:46
→ NickyP joined16:46
fr0sty ivan-kanis: no. 'git pull' doesn't update refs/remotes/origin/<branch>16:47
→ banisterfiend joined16:47
→ dr_win joined16:47
→ necromancer joined16:48
mikesusz does anyone pre-process files before commit? i'm working on a way to run compass compile on my sass dir before commit, but running then re-adding the files in git-precommit doesn't seem to work16:48
i end up pushing the (version of the file before the pre-commit script ran)16:49
fr0sty doing 'git fetch origin', 'git remote update' or similar would have fixed it too.16:49
→ justin-george joined16:49
fr0sty mikesusz: re-adding?16:49
→ dreiss joined16:49
← MattDiPasquale left16:49
mikesusz fr0sty - i thought since i was changing the contents of the file in pre-commit i had to git add <file> again (in that script) bad assumption?16:50
← ROBOTARMY left16:50
→ Thell joined16:51
→ Schmallon joined16:51
→ sako joined16:51
→ sjohnson joined16:52
→ oriba joined16:52
fr0sty no, you would have to add the change.16:52
sjohnson Q: is there a way to find out the absolute base path of which repo you are in?16:52
fr0sty are you sure the git add is being called and succeeding?16:52
mikesusz fr0sty - this is what i'm trying to do https://gist.github.com/80612916:52
fr0sty sjohnson: man git-rev-parse and look at --show-top-level16:52
jast sjohnson: the 'git-rev-parse' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-rev-parse.html [automatic message]16:52
sjohnson thanks16:52
→ MetaCosm joined16:52
fr0sty mikesusz: exec never returns to the shell.16:53
don't do that.16:53
mikesusz ah!16:53
fr0sty drop the 'exec ' and just start that line with 'compass...'16:53
→ nadavoid joined16:54
mikesusz k, trying it out now. i may not need the wait then; i thought maybe the git add was firing out of sequence before the compass command returned16:54
sjohnson sweet16:54
← sjohnson left16:54
→ MattDiPasquale joined16:55
mikesusz fr0sty - most excellent. :) thank you!16:56
wasn't sure if i was failing at shell scripting, git hooks, compass, or all of the above.16:56
→ tgunr joined16:58
→ tgunr_ joined16:58
→ MacGyverNL_ joined16:58
← MacGyverNL_ left16:58
→ drl joined16:59
← kpreid left16:59
← ruied left16:59
← jankoprowski left17:00
← SirFunk_ left17:00
← jfelchner left17:01
← teweWork left17:02
← tgunr left17:02
← DiGitalX left17:03
→ stringoO joined17:03
→ noahcampbell joined17:03
← dreiss left17:04
→ Andy19782 joined17:05
← Andy19782 left17:05
→ guardian joined17:06
← bronson left17:06
→ bronson joined17:06
→ amerine joined17:07
→ kipras_ joined17:07
← stringoO left17:08
→ mrconnerton joined17:09
← mrconnerton left17:09
→ excess^ joined17:09
← kipras_ left17:09
→ mrconnerton joined17:09
→ kipras joined17:10
mrconnerton crap17:11
hi :-)17:11
I'm having and issue with a git repo that has gotten fairly large17:11
← drl left17:11
mrconnerton when I run git gc I'm getting: Counting objects: 25368, done.17:11
fatal: Out of memory, malloc failed9)17:11
error: failed to run repack17:11
← bronson left17:11
→ radarek joined17:12
→ err_ok joined17:12
← mrconnerton left17:13
→ mrconnerton joined17:13
← radarek left17:13
mrconnerton hello17:13
I'm not sure if I actually connected. People around in here?17:13
fr0sty patience...17:14
← MattDiPasquale left17:14
mrconnerton fr0sty, sorry my app rebooted just as I asked the question, I wasn't sure if it went though17:14
fr0sty also, check the header: Users:72617:14
→ OOPMan joined17:14
mrconnerton fr0sty ya thanks.17:14
→ pheaver joined17:15
← pheaver left17:15
→ DiGitalX joined17:15
→ pheaver joined17:16
fr0sty mrconnerton: how much memory do you have?17:16
how big is the repo's .git dir?17:17
mrconnerton 2gig on the dev server17:17
fr0sty what are you storing in it?17:17
mrconnerton .git - 2.6G17:18
fr0sty go on...17:18
← Linc left17:18
mrconnerton Fr0sty en entire drupal site. unfortunately the form devs put a bunch of files and sql dumps in there that didn't need to be17:19
← OOPMan left17:19
mrconnerton I have since added those directories to the .gitignore17:19
→ boombatower joined17:19
→ Cyd joined17:21
← rcu left17:21
mrconnerton however git gc, pull, etc still fail because of memory (and what I suspect is all that crap in the history)17:21
fr0sty I see.17:21
→ kenichi joined17:21
fr0sty btw, adding them to .gitignore doesn't stop the from being tracked or remove them from the history.17:22
mrconnerton fr0sty thats what I assumed, at least the history part17:22
→ rcu joined17:23
← necromancer left17:23
→ necromancer joined17:23
→ drl joined17:24
→ courtain joined17:24
← courtain left17:25
← justin-george left17:25
fr0sty mrconnerton: man git-filter-branch and get that crap out of the history. that will help.17:25
jast mrconnerton: the 'git-filter-branch' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-filter-branch.html [automatic message]17:25
← sagsousuke left17:26
→ chrischambers_ joined17:26
fr0sty in the meantime, setting pack.windowMemory and pack.packSizeLimit (man git-config for details) might help your other problems.17:26
jast the 'git-config' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-config.html [automatic message]17:26
chrischambers_ → chrischambers17:27
mrconnerton fr0sty got it. I will look into both of those17:28
← pgib left17:29
→ koo6 joined17:29
→ maattd joined17:29
→ MattDiPasquale joined17:29
→ tstclair joined17:30
← drl left17:30
← rcu left17:30
→ corni joined17:30
→ rcu joined17:31
→ numeromancer joined17:32
← Schmallon left17:32
→ pielgrzym joined17:32
pielgrzym could someone explain me main difference between remote tracking branch and remote branch checked out locally?17:33
mrconnerton fr0sty pack.windowMemory documentation is vague. What should I be doing here? increasing the value or lowering it. If I'm running out of memory I suspect i wan to increase it?17:33
→ GaVrA joined17:33
→ Linc joined17:33
→ Bruce_Wayne joined17:34
→ ChrisBuchholz joined17:36
← err_ok left17:36
→ spearce joined17:36
→ OOPMan joined17:37
→ cirwin joined17:37
ChrisBuchholz Hello. I am looking for a git book that covers all the basics as well as more advanced topics. I have used git for a year or so, but I dont feel that i "know it" - i just do what i know work. Which book are recommended?17:37
→ steph021_mac joined17:38
spearce ProGit (progit.org) is pretty good17:38
ChrisBuchholz spearce: i had that on my wishlist already - guess ill go with that one then17:41
→ ChUrP joined17:42
→ panike joined17:43
→ drl joined17:43
→ bauruine joined17:43
fr0sty mrconnerton: I think lower is better.17:43
rp2 nice name17:44
→ keystr0k joined17:44
→ gnat42 joined17:45
mrconnerton fr0sty last question, what is an ideal pack size when your .git is > 2gb?17:45
← dcamp left17:45
gnat42 suppose I have a project in git, and I'm refactoring a file into two files in two different directories17:45
is there a way to keep the 'history' link between them?17:46
should I even care?17:46
Nugget git tracks information, not files.17:46
gnat42 is that a svn move type thing that doesn't really apply?17:46
Nugget: ok - good to know, I don't personally care about tracking the 'history' but just wanted to make sure I wasn't doing it 'wrong'17:46
Nugget you're doing it right. :)17:47
gnat42 perfect - thanks17:47
jast git never even stores information about renames. it's detected dynamically as needed.17:47
→ mrenouf|work joined17:48
fr0sty mrconnerton: beats me. 100M ?17:48
mrconnerton fr0sty sounds good to me17:48
→ VladDrac joined17:49
→ pedrobelo joined17:49
fr0sty gnat42: the only thing that fouls it up is deleting a file, making a bunch of commits and then re-adding the file under a different name.17:49
gnat42 fr0sty thanks17:50
fr0sty as long as content is just being shuffled around git can follow it.17:50
← ChrisBuchholz left17:52
→ iband joined17:52
→ Titosemi joined17:53
← berto- left17:54
→ starcoder joined17:55
← rcu left17:58
→ rcu joined17:58
→ aah joined17:58
← drl left17:59
← keystr0k left18:01
→ ROBOTARMY joined18:02
Titosemi i've a remote repository which i want to change the path. would it just work in my local repo with all the branches if i run a git remote set-url? does git remote set-url also work with local paths?18:03
← ChUrP left18:03
← rcu left18:03
→ rcu joined18:04
→ torbenh3 joined18:05
← cirwin left18:06
→ smolyn joined18:07
← ivan-kanis left18:07
← steph021_mac left18:08
← rcu left18:09
→ steph021_mac joined18:09
← steph021_mac left18:09
→ steph021_mac joined18:09
→ rcu joined18:09
Titosemi ok, i already got it!18:11
→ drl joined18:11
→ LiamH joined18:13
← steph021_mac left18:14
→ CAP6478 joined18:14
→ psoo joined18:15
→ ruied joined18:15
← CAP6478 left18:15
→ CAP6478 joined18:16
← CAP6478 left18:16
→ ELREY joined18:16
ELREY anyone knows why it may happen that files modified by teammates show up in my working copy and even makes ME commit them instead of the actual file author?18:16
makes no sense really18:17
← rcu left18:17
→ rcu joined18:17
→ Yuuhi joined18:17
← d0m left18:18
→ marc_kdab joined18:18
→ Schmallon joined18:20
bremner merges?18:20
→ coderdad joined18:20
← rcu left18:20
← ruied left18:21
→ rcu joined18:21
ELREY hows that18:23
→ gusnan joined18:23
→ ShooterMG joined18:25
← marc_kdab left18:25
yrlnry Suppose I've made a whole bunch of changes to my working tree, and my regression tests all pass. I would like to split up my changes into several commits, to make them easier to understand.18:26
But I don't want to make any commits that break the regression tests.18:26
What's a good workflow for arranging that?18:26
SethRobertson ELREY, are they editing in your working copy? If not, you are getting some kind of merge conflicts or the like18:26
yrlnry: git add -p18:27
fr0sty yrlnry: git add -p; git stash; test; git stash pop; repeat18:27
yrlnry Oh, git stash! Duh.18:27
Thanks, that was the part I was missing.18:27
fr0sty s/test/test && commit || retry/18:27
→ berto- joined18:27
yrlnry And if the tests fail I can pop the stash and add some more stuff to the index.18:27
→ AlexP joined18:28
yrlnry No, wait, I'm missing something.18:28
Won't the git-stash command stash everything, including the changes I added to the index?18:28
→ steph021_mac joined18:28
yrlnry Oh, I need a git-commit in between the add and the stash.18:28
Then if I don't like it I can commit --amend.18:29
fr0sty yrlnry: or git stash --keep-index18:29
agile I've the weirdest thing going on in one of my repos, where several commands (rebasing, applying stashes etc) tell me my tree is dirty but I can't tell why.. git status shows it to be clean, even tried deleting the branch and checking it back out from origin and still insists the tree is dirty.. any ideas what might cause that?18:30
→ ChUrP joined18:30
← rcu left18:30
← steph021_mac left18:30
yrlnry Hey, the git-stash manual has a section that describes my exact situation.18:30
→ asmega joined18:30
yrlnry "Testing partial commits",18:30
Thanks again.18:30
→ steph021_mac joined18:31
← steph021_mac left18:31
→ steph021_mac joined18:31
← drl left18:31
→ loincloth joined18:31
→ phaedrix joined18:32
← panike left18:32
→ rcu joined18:33
→ justin-george joined18:33
← phaedrix left18:33
← AlexP left18:33
← guardian left18:33
→ phaedrix joined18:33
→ AlexP joined18:34
→ _psychic_ joined18:36
→ koltroll joined18:36
→ tizbac joined18:37
→ LouisJB joined18:38
← jfkw left18:39
→ jfkw joined18:39
→ marc_kdab joined18:40
→ theoros joined18:40
← rcu left18:42
→ drl joined18:44
→ Bass10 joined18:44
→ TheFox21 joined18:44
→ Unknown[NF] joined18:45
← Bass10 left18:45
← AlexP left18:45
→ Bass10 joined18:46
← ChUrP left18:46
← JohnFord left18:46
← resmo left18:47
→ bronson joined18:48
→ TheMoonMaster joined18:48
← MattDiPasquale left18:49
→ siprbaum joined18:50
← drl left18:51
→ lovy joined18:51
← lovy left18:52
← DiGitalX left18:54
→ archis joined18:54
← archis left18:54
→ archis joined18:54
Cyd → NotCyd18:55
→ larsmoil joined18:55
NotCyd → Cyd18:55
→ chops13 joined18:55
← archis left18:55
← tvw left18:57
→ rcu joined18:58
→ emptyflask joined18:59
→ dkasak joined19:00
→ SirFunk_ joined19:00
→ cirwin joined19:00
→ Evious joined19:00
← ISF left19:01
→ bkosborne joined19:02
→ heptadecagram joined19:03
← justin-george left19:03
→ drl joined19:04
→ justin-george joined19:04
→ joel_falcou joined19:05
heptadecagram I have a very unusual use case, and could use some help.19:06
I am using git parallel to a CVS repository that is large enough that tailor, cvsps, etc. all die if I try to convert them. So, I have the dirty/pristine model:19:07
The pristine WD is where CVS and git interact, and the dirty WD is where I do any development work.19:07
The CVS tree has just had two maintenance branches. I want to find a way for git to treat them as branches off of the trunk; I currently have them as separate checkouts (pristine WDs), that my dirty WD references as remotes.19:09
But because they have no common history, git gets confused.19:09
→ MattDiPasquale joined19:09
→ elyos joined19:09
heptadecagram How do I make my git-history of the CVS branches treat those CVS branches as branches off the same trunk, rather than completely distinct repositories?19:10
→ johnw joined19:10
→ montylounge joined19:10
→ Zephyrus joined19:12
→ JohnFord joined19:14
← iband left19:15
← Bruce_Wayne left19:18
→ stephenplatz joined19:18
← MetaCosm left19:20
→ bmahe joined19:20
→ SimonP86 joined19:20
← marc_kdab left19:22
← flazz left19:23
→ khmarbaise joined19:23
← Evious left19:24
→ guardian joined19:24
← rcu left19:24
→ Evious joined19:25
← bkosborne left19:26
→ AlexP joined19:26
← AlexP left19:26
→ sergio_ joined19:26
← priidu left19:27
→ engla joined19:27
← engla left19:27
→ engla joined19:27
→ j416 joined19:29
→ priidu joined19:29
→ deek0146 joined19:29
← LouisJB left19:30
fr0sty https://git.wiki.kernel.org/index.php/GraftPoint19:30
→ djn joined19:30
← E-shark left19:31
→ rcu joined19:31
→ cbreak_work joined19:31
← rcu left19:32
→ rcu joined19:32
← oriba left19:32
SethRobertson Isn't git-replace the recommended replacement for grafts?19:33
→ trobrock joined19:34
ChanServ set mode: +v19:35
→ radarek joined19:35
ArseneRe1 Is there a way to export a repository uncompressed? git archive's formats are only tar and zip.19:36
→ iAsteris_ joined19:36
ayust git archive | tar -x ?19:36
that's the typical way19:36
→ dcamp joined19:36
→ mitchellhislop joined19:36
← rcu left19:37
→ hunzinker joined19:37
ayust tar doesn't really do compression (hence why people gzip tar files)19:37
→ rcu joined19:37
ArseneRe1 ayust: Oh that's right.19:37
← koltroll left19:37
ArseneRe1 I had forgotten. Thanks!19:37
← dkasak left19:37
→ ruied joined19:38
→ dkasak joined19:38
← dkasak left19:38
← hunzinker left19:38
← iAsterisk left19:38
← djn left19:39
← bronson left19:39
j416 ArseneRe1: you are aware that git archive exports files, and not the repository, right? :)19:39
→ sftp joined19:39
ArseneRe1 j416: Yeah, sorry the misspeak. I just want the files.19:39
j416 just checking.19:40
← drl left19:40
→ bronson joined19:41
→ _Vi joined19:42
← steph021_mac left19:42
← gnat42 left19:42
← SirFunk_ left19:43
→ Bruce_Wayne joined19:43
← sergio_ left19:44
← gusnan left19:45
→ mikepack joined19:45
← radarek left19:46
heptadecagram Ah, fr0sty, I didn't realize you were addressing my question. Thanks!19:47
← _Vi left19:47
← rcu left19:49
→ rcu joined19:49
→ khmarbaise_ joined19:49
← khmarbaise left19:49
khmarbaise_ → khmarbaise19:49
→ marcob joined19:49
→ mutex joined19:49
→ AlexP joined19:49
← AlexP left19:49
← mrconnerton left19:50
→ dreiss joined19:50
→ drl joined19:53
→ AlexP joined19:54
→ necromancer` joined19:55
← necromancer left19:55
→ orafu joined19:56
← corni left19:58
← rcu left19:58
← deek0146 left19:58
→ rcu joined19:58
← coderdad left20:00
← ruied left20:02
→ bigpeted joined20:03
← orafu left20:03
← JohnFord left20:07
→ _ollie joined20:07
← berto- left20:07
← drl left20:07
→ brodo joined20:08
TrentonDAdams so, I was wondering. To get around the issue with git merging, and svn, couldn't git just do merges as "new" commits? i.e. for every branch you want to merge, it just goes in as a new commit, rather than it being tracked in SVN?20:08
For example, if you're merging to branchX into branchY, change the commit message to branchY.20:09
for the svn tracking purposes.20:09
→ rns joined20:09
→ coderdad joined20:09
← _ollie left20:09
→ huehnts joined20:10
← bartek left20:11
PerlJam TrentonDAdams: you're talking about "git merge --rebase" essentially.20:11
→ ki__ joined20:11
PerlJam TrentonDAdams: and always rebasing will help svn, yes. But it's not always the best way to work in git.20:12
← bigpeted left20:12
→ bartek joined20:12
← rcu left20:13
← Zephyrus left20:14
→ Buglouse joined20:15
← heptadecagram left20:15
necromancer` → necromancer20:16
→ nadim_ joined20:17
→ LouisJB joined20:17
→ _ollie joined20:17
→ rcu joined20:18
← charlietanksley left20:18
← MattDiPasquale left20:18
→ Zephyrus joined20:18
→ ISF joined20:20
→ drl joined20:20
← Bruce_Wayne left20:21
← rcu left20:21
→ speakman joined20:21
← speakman left20:21
← stephenplatz left20:23
→ MetaCosm joined20:24
← TheFox21 left20:24
→ joshsmith joined20:25
→ rcu joined20:26
jefferai hello, I know that you can find the name of the alias for insteadOf by running e.g. git config url.git://git.example.com/.insteadOf20:26
is there a way to do the opposite -- given the alias, find the URL?20:26
← alberto56 left20:26
← iAsteris_ left20:26
RaceCondition I have two main branches, master and devel, and then several local branches. how can I see which local branches are based on the master and which on the devel branch?20:27
→ giallu joined20:27
→ RobertMeta joined20:27
← RobertMeta left20:27
→ RobertMeta joined20:27
← rcu left20:28
→ iAsterisk joined20:28
→ welly joined20:29
→ koltroll joined20:29
← marcob left20:30
→ thiago_home joined20:30
← torbenh3 left20:30
btipling does git status hash all the files to check to see if there are changes or just check last modified time?20:31
→ brosner joined20:31
→ alberto56 joined20:31
→ rcu joined20:32
_rane and compare it to what?20:32
wouldn't it then show that there are changes if I simply save a file after committing20:33
→ saysjonathan joined20:33
→ marc_kdab joined20:33
→ Heimidal joined20:34
→ tvw joined20:34
← _ollie left20:34
ayust btipling: Git always deals with file contents20:35
btipling ah20:35
it's so fast20:35
fr0sty RaceCondition: 'git log <branch>..devel' will be empty if it is based on devel20:36
btipling how does it hash so many files so quickly20:36
_rane: it could be combination: last modified time + hash20:36
so it doesn't hash files that obviously haven't changed20:37
← joshsmith left20:37
→ _ollie joined20:37
btipling I dunno, just amazed at how good and fast it is20:37
FauxFaux btipling: Yes, it uses mtimes.20:37
fr0sty btipling: how big is your repository?20:37
btipling huge20:37
gigabytes20:37
← drl left20:38
← hobodave left20:38
btipling I think20:38
FauxFaux du -sh .git/objects/20:39
← _ollie left20:39
btipling oh much smaller than I thought20:39
109M20:39
← rcu left20:40
fr0sty btipling: sha hashing is fast by design.20:40
FauxFaux I do like the way the reference implementation has provably identical but significant algorithm (i.e. not implementation) ineffiencies.20:40
← koltroll left20:41
→ malinens joined20:41
→ rcu joined20:42
RJD22|away → RJD2220:42
→ koltroll joined20:42
→ Jiyn joined20:42
ChanServ set mode: +v20:43
Jiyn hi20:43
we could talk about gitosis-init here ?20:44
bremner you could, but we'll suggest you switch to gitolite.20:46
someone might answer your question, but the switch to gitolite message is mandatory.20:46
← rcu left20:47
→ rcu joined20:47
Jiyn ok, I will try gitolite to see if I have the same problem20:48
→ t3i joined20:48
← rcu left20:49
← tizbac left20:49
← marc_kdab left20:49
← malinens left20:50
→ drl joined20:50
← priidu left20:51
→ rcu joined20:51
← rcu left20:52
← nona left20:52
→ PKKid-Home joined20:53
PKKid-Home Hey guys..20:53
I acidentlly the whole thing..20:54
thiago_home PKKid-Home: you forgot the verb20:54
RaceCondition fr0sty: oh, good point.. thanks20:54
PKKid-Home Is there a way to revert bad push?20:54
there were 40 commits I pushed that I need to unpush.20:54
← koltroll left20:55
RaceCondition fr0sty: although it'd still be cool to get a list of all branches and what they are based on with one command20:55
PKKid-Home thiago_home, Its an internet meme. ;)20:55
← spaceonline left20:55
thiago_home PKKid-Home: "I accidentally the whole thing" <--- missing a verb, makes no sense20:56
fr0sty RaceCondition: 'based on' is not really a concept in git.20:56
thiago_home PKKid-Home: anyway, if no one has got those commits yet, simply git push -f the commit you want the remote to have20:56
→ bkosborne joined20:56
→ bradleyayers joined20:56
PKKid-Home thiago_home, don't I need to specify the branch as well?20:57
thiago_home PKKid-Home: yes20:57
PKKid-Home: git push remotename sha1:branchname20:57
← Heimidal left20:58
→ cjs joined20:58
→ spaceonline joined20:58
← joel_falcou left20:59
→ dsdjf joined21:00
→ Heimidal joined21:00
PKKid-Home thiago_home, thanks, I might have it.. pulling a new copy of it now to double check.21:01
I basically ran this: "git push -f origin 1bffa7840b61:master"21:01
→ EricInBNE joined21:03
← jrmuizel left21:04
← coderdad left21:04
→ PABRR3M joined21:07
→ k04n joined21:07
→ lightcap joined21:07
← drl left21:08
→ MattDiPasquale joined21:09
→ marc_kdab joined21:09
→ darkredandyellow joined21:09
← mahata left21:10
→ joshsmith joined21:10
→ Eridius joined21:10
→ novitololo joined21:11
→ coderdad joined21:12
← marc_kdab left21:13
→ marc_kdab joined21:14
→ henr_k joined21:14
→ suspectx joined21:14
→ pantsman joined21:16
← pantsman left21:16
→ pantsman joined21:16
→ bgerber joined21:17
← psoo left21:17
→ cesario joined21:19
← coderdad left21:19
→ csgeek joined21:19
→ ppawel joined21:21
→ drl joined21:21
← Zephyrus left21:21
→ tizbac joined21:22
← j416 left21:22
→ MikhX joined21:22
→ wtfness joined21:23
← sonnym left21:23
← bronson left21:23
→ berto- joined21:23
← suspectx left21:23
→ awilkins joined21:24
→ Agos joined21:25
← k04n left21:25
→ benco joined21:27
→ jesmith joined21:27
← alester left21:27
→ bronson joined21:27
← alberto56 left21:28
← bkosborne left21:28
→ chessguy joined21:29
→ bkosborne joined21:29
→ angelsl joined21:29
→ tewecske joined21:31
→ smolyn_ joined21:33
← smolyn left21:33
← benco left21:33
← Jiyn left21:33
→ nurey joined21:34
← smolyn_ left21:35
→ smolyn joined21:35
→ dfr|mac joined21:36
← drl left21:36
→ eirikb joined21:38
eirikb If I want my submodules to always be up-to-date with the dependant project, should I use something else? Or is there a trick? I don't even know how to update it (git checkout master;git pull?)21:38
← bhosie left21:38
→ bhosie joined21:38
← bkosborne left21:39
← marc_kdab left21:39
← Tac left21:39
→ Tac joined21:40
rudi_s eirikb: AFAIK subdmodules aren't really suited for this. If I need this I just use a normal git clone and update them both with a little script.21:40
PKKid-Home thiago_home, I never said thank you before. Thanks.21:41
eirikb rudi_s: I had that before, using the Makefile from jQuery21:41
→ patmaddox joined21:41
fr0sty eirikb: ask SethRobertson about git-slave21:41
PKKid-Home actually maybe I did, but i forgot. ;-P21:41
→ bkosborne joined21:41
eirikb SethRobertson: Can I msg you about git-slave? :)21:42
→ Sonderblade joined21:42
← Sonderblade left21:42
→ jsemar joined21:42
→ bhosie_ joined21:42
jsemar is there a way to re merge in or checkout a non-commit tree hash21:42
→ Sonderblade joined21:43
eirikb So using a Makefile is maybe even better than submodules? I changed to submodule because I felt it was more proper (built into git, linking in github etc)21:43
← necromancer left21:43
rudi_s SethRobertson, eirikb: I'm curious too, what is git-slave?21:44
→ moqq joined21:44
← bhosie left21:44
bhosie_ → bhosie21:44
moqq i have a .gitmodules file, `git --version` => "git version 1.7.0.4", but when i do `git submodule update --init`, `git submodule status`, or really `git submodule ANYTHING` it does nothing. no output, no error, just nothing21:45
and teh same repo works fine on other machines21:45
anyone have any suggestions?21:46
SethRobertson gitslave? I'm glad you asked21:46
fr0sty eirikb: if you care about dependencies between repositories you should track the relationship formally.21:46
SethRobertson rudi_s: eirikb: gitslave (http://gitslave.sf.net) is useful when you control and develop on the subprojects at more of 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.21:46
→ miloshadzic joined21:47
→ bhosie_ joined21:47
eirikb SethRobertson: Sounds like something I want21:47
SethRobertson Unlike git-submodules, there is only a strict relationship between the repositories at tags (meaning if you do not tag, you cannot trivially figure out what commit a submodule was at when a specific commit was made somewhere else)21:47
← skoop left21:47
rudi_s SethRobertson: Looks interesting, thanks.21:48
TrentonDAdams PerlJam: really? git rebase will prevent the commit messages from having a commit message referencing another branch?21:48
SethRobertson However, this gives you the ability to easily pull, checkout, commit, etc over all repos at the same time, unlike subtree or submodules where you have to do a little dance21:48
→ snearch joined21:48
← bhosie left21:49
bhosie_ → bhosie21:49
SethRobertson It is a simplier paradigm when you are doing forward activities. It is more complex when you are doing forensics due to the lack of linkage.21:49
→ drl joined21:50
SethRobertson Or when you don't control the other repos or are not otherwise willing you make your own branches/tags/etc on them (e.g. local master of a remote project)21:50
eirikb SethRobertson: Will this be added to git native?21:50
SethRobertson Never21:51
Buglouse So i would like to receive information (logs, size) without git clone, what are methods?21:51
SethRobertson Some of the concepts (like relative URLs, which were added in 1.7.4) might go in, but not the overall system. Perhaps in contrib at most.21:51
Buglouse maybe just grab the '.git'.21:52
SethRobertson Buglouse: ssh to the remote server and run the commands there? gitweb? Really you need a clone. Grabbing a .git is cloning21:52
← EricInBNE left21:52
← guardian left21:53
→ albel727_2 joined21:54
eirikb SethRobertson: Can this co-exist with submodules?21:54
Against the same dependencies21:55
Just if I hook into a system (like a PaaS) where I can't control third party applications this might be a problem (as well as hackish Makefiles, but they might actually randomly work)21:55
SethRobertson eirikb: It is possible in theory. However, you would have remember to only run gits commands when you have last done a gits checkout, and only run git submodule commands when you have last run git submodule update21:56
Well, as I said, you can make a local submaster for the third party. Let the remotes/thirdparty/branchnames do what they want and merge the right remote branch in your local branch (e.g. master, 2.0-release, etc)21:57
← MikhX left21:57
eirikb Hmm yeah, many balls in play?21:57
← TheMoonMaster left21:59
novitololo When using git diff: what @@ -1,7 +1,7 @@ means?22:00
← Heimidal left22:01
Nugget novitololo: http://stackoverflow.com/questions/987372/patch-format22:01
novitololo appreciate.22:02
← drl left22:02
← chops13 left22:03
← albel727_2 left22:03
→ acid joined22:04
← cesario left22:04
→ ice799 joined22:04
→ linusoleander joined22:05
ChanServ set mode: +v22:05
← linusoleander left22:06
→ linusoleander joined22:06
← pantsman left22:06
← lightcap left22:07
← RobertBerger left22:08
→ _Vi joined22:09
→ asdfasdf_ joined22:09
eirikb SethRobertson: So how would you use gitslave against github, where you don't want others to depend on third party applications for git? I could for instance use gitslave personally and hook up to the dependent projects outside the actual github repo22:09
→ PABRSEI joined22:10
→ codespectator joined22:10
← elyos left22:11
← asdfasdf_ left22:11
← ISF left22:11
← mikepack left22:12
← darkredandyellow left22:12
→ drl joined22:15
→ nanotube joined22:15
SethRobertson eirikb: Well, without actually knowing much about how github works and whether magic happens when you fork, I see two options. First is your own project is simple enough you only want to use the master branch, and the third party repos are stable enough that you can likewise only use the master branch. You would then not really tag or create branches (at least not with gitslave).22:16
→ agateau joined22:16
← asmega left22:16
eirikb SethRobertson: I'm the only guy commiting code, and they are small simple projects. master branch is fine22:17
SethRobertson eirikb: The other option is to create your own fork/copy of the third party repos on github with your own branch names so that your master branch might not correspond to the thirdparty master (it might be thirdparty/2.5.1_release_branch or whatever)22:17
→ tro joined22:17
eirikb SethRobertson: I haven't even released a 1.0 yet22:17
← agateau left22:18
SethRobertson eirikb: Then download gitslave http://gitslave.sf.net and do the prepare and attach commands and you will magically be all set up22:18
→ agateau joined22:18
→ Ze_M joined22:18
eirikb SethRobertson: And other people that might want to fork my code will not have dependency of gitslave?22:18
SethRobertson eirikb: No, you are not forced to use gitslave. You can use normal git commands as normal. gitslave is a simple foreach loop around all the repos in a project, with output summarization. Nothing tricky is going on22:19
eirikb Interesting, I have to look into it. Thanks22:21
→ workmad3 joined22:22
→ Textmode joined22:22
← tro left22:22
← tstclair left22:23
→ cking_nj joined22:23
← khmarbaise left22:24
→ nothingHappens_ joined22:24
→ tro joined22:24
→ MikhX joined22:25
← agateau left22:25
→ bigpeted joined22:25
← bigpeted left22:25
→ bigpeted_ joined22:25
→ agateau joined22:26
→ Illusioneer joined22:26
Illusioneer hey, has anybody had a problem where suddenly when you try to push gitorious says Connection Refused when it was working just fine before?22:28
← eirikb left22:28
→ airborn joined22:28
→ Giles joined22:29
→ necromancer joined22:29
← LiamH left22:29
← bartek left22:29
Giles why is git show --oneline <commitish> showing me one line followed by some diff output?22:29
SethRobertson It is stupid that way22:30
→ marc_kdab joined22:30
← bkosborne left22:30
SethRobertson git log is more accomidating22:30
Giles SethRobertson: ?22:30
SethRobertson !22:30
Giles so I want git log -1 --oneline <commitish>22:31
SethRobertson Something like that22:31
Giles gar22:31
← excess^ left22:32
← MikhX left22:32
charon Giles, SethRobertson: git-show takes -s to suppress the diff22:33
Giles charon: ta22:33
← drl left22:33
→ zoran119 joined22:33
charon git show is more or less git log -p --no-walk --cc22:34
→ kpreid joined22:34
charon at least with commit args22:34
← Titosemi left22:35
← alex__c2022 left22:35
Giles right22:36
← bronson left22:36
→ sonnym joined22:36
→ jdowdle joined22:38
→ jaql joined22:38
ChanServ set mode: +v22:38
← Illusioneer left22:39
← novitololo left22:39
→ hobodave joined22:40
→ SethRobertson joined22:41
← Sonderblade left22:41
→ goshawk joined22:41
← nurey left22:41
← jdowdle left22:42
← agateau left22:44
← montylounge left22:44
→ alberto56 joined22:44
← MetaCosm left22:44
← trobrock left22:45
→ agateau joined22:45
→ eletuchy joined22:46
← moqq left22:47
→ drl joined22:47
← hyperair left22:48
→ MetaCosm joined22:48
← MetaCosm left22:48
→ MetaCosm joined22:48
piti Hi there. I'd like to put git rev id into the code that is under (mainly to help debugging when user comes with a bug). Is there a best way to do ?22:49
ayust smudge/clean22:49
→ _ollie joined22:50
SethRobertson I do it during make install22:50
ayust having it done by make can also work22:50
piti I thought about using a hook (post-commit), but not sare how to do it, or fetching from git during the configure.in22:50
what do you mean by smudge/clean ?22:51
ayust http://www.kernel.org/pub/software/scm/git/docs/gitattributes.html search for 'smudge'22:51
← ki__ left22:52
→ ki__ joined22:53
piti hum, seems great (let me dig that a while :) )22:53
← fowlduck left22:53
→ priidu joined22:53
← drl left22:54
→ willmarshall joined22:54
→ lresende joined22:54
willmarshall How would I pull a list of branches from a remote git server?22:54
→ silence joined22:54
RobertLaptop git fetch remoteserver or git pull remoteserver22:54
silence → Guest510722:54
← Guest5107 left22:55
ki__ Hello. I have an id_rsa.pub and an id_dsa.pub key ... does that mean I can use EITHER key or is one of them configured? Where can I check which one is configured to be used?22:55
→ k04n joined22:55
← lresende left22:55
← jaql left22:56
→ MikhX joined22:57
→ elyos joined22:57
→ fr0sty-away joined22:58
← bigpeted_ left22:58
piti btw, I have another question: we use hooks (code formatting), but it seems to be local for each repository. is there a way to «attatch» and distribute those hooks, like just with a git clone repo ?22:58
cirwin piti: not really, it's pretty annoying — you could put them in the repo along with a script that symlinks the .git/hooks directory22:58
there's probably a tool that automates that process, but don't know what it is22:59
← willmarshall left22:59
piti ok, good to now anyway22:59
← tgunr_ left22:59
→ bauruine_ joined23:00
← bauruine left23:01
→ tgunr joined23:01
→ eletuchy_ joined23:01
← Linc left23:02
← tizbac left23:03
→ sjohnson joined23:03
← necromancer left23:03
sjohnson Q: i'm trying to run a git-whatchanged by variable dirnames. is it possible to do it without chdir? i'm finding --work-tree not working for what i thought was the solution23:04
← eletuchy left23:04
→ zombor joined23:06
→ bronson joined23:06
Giles hmm, I broke my server, when I try to push to it its asking for a password, despite having the ssh keys setup, what could have gone wrong?23:07
→ drl joined23:07
← snearch left23:07
johnw Giles: what OS?23:07
→ ISF joined23:07
cirwin Giles: what are the permissions on files in your .ssh directory?23:07
Giles pushing from Linux ubuntu, to Linux Lenny23:08
johnw this is more of an SSH question, but it could be DNS, permissions, PAM, SSH settings, user not having a home directory or shell, etc.23:08
cirwin (and are you sure it's asking for a server password, not a key password)23:08
Giles cirwin: yeah, I don't have a password on this key23:09
← Dave^| left23:09
← smolyn left23:09
cirwin ok, just verify that the files in .ssh are 600 or 40023:09
Giles johnw: well I recently added an rDNS record.23:09
cirwin any more permissive and ssh refuses to use them23:09
johnw also: The server cannot perform reverse DNS lookups on the inbound IP address23:09
→ smolyn joined23:09
johnw make sure .ssh itself is 70023:09
Giles .ssh is u+r and g+r23:10
johnw .ssh has to be u+rx23:10
→ makubi joined23:10
Giles johnw: hmm23:10
johnw: sorry yeah authorised keys is u+r and g+r23:11
.ssh is u+rwx and g+rwx23:11
oh no its g+rx23:11
Arrowmaster normally ~/.ssh should be 70023:12
Giles Arrowmaster: whats 700 I am not so good with bitmasks23:12
Arrowmaster u+rwx,g-rwx,o-rwx23:13
← drl left23:13
→ fowlduck joined23:13
Giles Arrowmaster: well that doesn't help, but thanks for the security tip.23:14
piti on the webpage we give to me a while ago, I found: «Ident - When the attribute ident is set for a path, git replaces $Id$ in the blob object with $Id:, followed by the 40-character hexadecimal blob object name, followed by a dollar sign $ upon checkout. Any byte sequence that begins with $Id: and ends with $ in the worktree file is replaced with $Id$ upon check-in»23:14
I guess that's I'm looking at, is it ?23:14
→ kulp joined23:15
johnw Giles: what I do is use sshd -d -d -p XXX on the server, and ssh -v -v -v -p XXX on the client23:16
→ archis joined23:16
← kipras left23:16
johnw because there can be *many* reasons why SSH might fail23:16
also, SSH has three different password prompts, so it's good not to assume which one you're seeing23:16
Arrowmaster piti: thats a per file hash, it only changes when the file content does, it has nothing to do with the commit hash23:16
johnw (there's publickey pass, password auth, and keyboard-interactive)23:16
← zombor left23:16
→ madewokherd joined23:16
→ loa joined23:16
← archis left23:17
piti ok, thank you avoiding me going into dark and dangerous ways23:17
→ ls_n joined23:19
→ Pewpewarrows joined23:19
→ ptl joined23:20
← johnw left23:20
MarcWeber How to list all files which are present at a given revision? something like git ls-files?23:20
→ necromancer joined23:20
← MattDiPasquale left23:20
cirwin ls-files --with <commit>23:21
MarcWeber cirwin: --with-tree ?23:22
cirwin --with works too I think, but yes23:22
MarcWeber cirwin: I searched the help file for commit and hash but didn't find it.23:22
yeah. it works.23:23
cirwin a commit is a type of tree-ish23:24
ls_n I'm getting a stream of errors from a bunch of operations all of a sudden. Starting with error: inflate: data stream error (incorrect header check)23:24
sound familiar?23:24
← spaceonline left23:24
ls_n https://gist.github.com/b7a9b58f4dcee16331c523:25
→ drl joined23:26
← chessguy left23:26
← bradleyayers left23:26
→ bgy joined23:27
bgy hi23:27
→ radarek joined23:27
← airborn left23:27
← _ollie left23:27
→ airborn joined23:28
fr0sty ls_n: repo certainly looks corrupt. what does git fsck --full tell you?23:29
→ shyam_iyer joined23:29
← goshawk left23:30
← Ze_M left23:31
ls_n fr0sty: running now. spitting out the same errors. Previous runs of git fsck (not --full) did the same.23:31
→ infogulch joined23:31
→ eean joined23:32
infogulch how do i download updates from a remote source into a cloned directory?23:32
MarcWeber infogulch: git fetch origin23:33
This will update the remote references show by git branch -r23:33
infogulch aah ok thanks :)23:33
ls_n fr0sty: https://gist.github.com/6bb608f2ab795c03080823:33
← jesmith left23:33
infogulch um is it normal for it to have no output?23:33
ls_n so... nuke the site from orbit and reclone?23:33
MarcWeber If you want to update your local references (branches) you have to use git pull, git rebase or git merge. The first runs one of the others automatically reading some info from ~/.git/config23:33
→ shiba_yu36 joined23:34
MarcWeber infogulch: try git checkout -tb new origin/remote-branch-name and watch .git/config. This way you learn fastest how to configure git pull. the -t means "--track" and sets up the relation between local branch and remote "source".23:35
if you add rebase = true rebasing will be used instead of merging when feeding remote changes into your local branch23:35
Giles Authentication refused: bad ownership or modes for directory /home/git23:36
← shyam_iyer left23:36
Sgeo Is there a point to deleting branches? Does it just modify history, or does it do something else?23:36
MarcWeber Sgeo: See git history as graph. braches are "pointers to commits" only.23:36
deleting branches means you remove the pointers - and git gc will drop the revisions later.23:37
← ISF left23:37
MarcWeber If there is a point in doing so depends on your workflows.23:37
Sgeo Besides deleting branches, and rebase, does git handle branches better than mercurial?23:37
MarcWeber I love this about git - because you can also drop branches easily keeping the repo clean.23:37
→ DAA3FGO joined23:37
MarcWeber Sgeo: I found that the remotes/* location is the most useful feature git has which mercurial does not have.23:37
mercurial has many plugins which to some degree mimic git behaviour.23:38
Sgeo I vaguely heard that somewhere, but don't particularly like history modification. I've played with Git via Git Extensions, haven't touched Mercurial23:38
MarcWeber However in mercurial you can't delete remote branches easily. You have to use the strip command.23:38
→ BiggFREE joined23:38
← starcoder left23:38
fr0sty ls_n: if you have nothing of value there, go ahead.23:38
MarcWeber Sgeo: It depends. In theory you should never get rid of data - but on the other hand human resources are limited. There is no point in keeping old bad code IMHO. But this heavily depends on the project23:39
ls_n Looks like a clean git status...don't think there's anything in branches. bleh.23:39
fr0sty: thx23:39
MarcWeber Eg for sharing quick and dirty ideas I can push, talk and tidy up later easily.23:39
← drl left23:39
→ mt3 joined23:40
← bhosie left23:40
Sgeo I'm sort of planning on using either Git or Mercurial for some single-person (me) projects. So far, I like what I've seen of Git with Git Extensions.23:40
→ starcoder joined23:40
Sgeo I've used SVN for a project with one other person before23:41
MarcWeber Sgeo: There are some minor differences. Eg mercurial supports cherry-picking writing down the patchs origin branch. But in practise git never got into my way.23:41
← ls_n left23:41
Sgeo MarcWeber, um.. not too familiar with cherry picking, please explain?23:42
MarcWeber spb: cherry-pick means you take a patch from one branch and apply it to another.23:42
← dreiss left23:42
MarcWeber in git you have to add manually "was taken from brach foo" or the like. Sometimes it may be helpful - however you can script these missing pieces easily.23:43
→ ISF joined23:43
← iAsterisk left23:43
→ MikhX_ joined23:44
Sgeo Ah23:44
By "add manually", you mean in the commit message, or the history graph?23:44
MarcWeber Sgeo: when cherry-picking you can tell git htat you want to edit the message.23:45
→ orafu joined23:45
← thiago_home left23:46
MarcWeber You can also edit the message later using git amend23:46
← awilkins left23:46
← MikhX left23:47
MikhX_ → MikhX23:47
Sgeo My question is, when you said Mercurial does that automatically.. does it put it in the commit message automatically, or is it visible by looking at history?23:47
→ iAsterisk joined23:47
→ keyvan joined23:48
MarcWeber Sgeo: http://mercurial.selenic.com/wiki/TransplantExtension I haven't used it yet.23:48
← dcamp left23:48
← Giles left23:49
MarcWeber Sgeo: in git its important to know that a commit does not store to which branch it belongs. THis is stored in the reference name only.23:49
However if you really care you could add a file storing this info. git is stupid by design - and I personally like this design choice.23:50
Don't worry to much. I'm pretty sure that you can convert from git to mercurial and back easily.23:50
← Bass10 left23:50
Sgeo At this point, I've played with Git. Don't really see much point in bothering with Mercurial23:51
Although I wonder if Git Extensions is warping my understanding23:51
→ shyam_iyer joined23:52
← shyam_iyer left23:52
→ silence joined23:52
→ shyam_iyer joined23:52
← silence left23:52
← radarek left23:53
→ drl joined23:54
← mt3 left23:56
← t3i left23:56
→ davidc_ joined23:57
→ SirFunk_ joined23:58
← sako left23:59

Logs Search ←Prev date Next date→ Channels Documentation