IRCloggy #git 2011-05-05

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.

2011-05-05

SethRobertson You may be thinking of a squash merge00:00
→ jmcantrell joined00:01
chrisf|work SethRobertson: i never really grokked why stgit makes things easier. care to explain?00:01
marchdown I'll go read up on squash-merge and cherry-picking from multiple sources. Thank you guys.00:01
← owst left00:02
← seivan left00:02
subpsyke SethRobertson: so, this would do it? git reset --hard $(git rev-list -1 $(git rev-parse --until=2011-04-29) HEAD)00:02
SethRobertson If you want to fixup patch/commit 2 of a 10 patch/commit series, you can "super-stash" commits 3-10, update 2, and the reapply 3-10. It is possible to do it with raw git, but the accounting gets annoying00:03
→ jawwad joined00:03
→ etcetera joined00:03
SethRobertson subpsyke: I recommend `git checkout @{2011-04-29 00:00:01}`00:04
chrisf|work SethRobertson: surely you'd just interactive rebase, mark the commit 2 as `edit`?00:04
subpsyke SethRobertson: don't the checkout logs get removed after a certain time?00:04
← f0i left00:04
SethRobertson You are requesting to check out the state of the repository at a certain time, not looking at the reflog.00:05
subpsyke ok00:05
← Evious left00:06
← lukeman left00:07
SethRobertson chrisf|work: Another example, you might want to insert a new commit between 2 and 3, which could be difficult (commiting at the end and then reordering might cause more conflicts). None of this is impossible with raw git, just more annoying.00:07
→ Evious joined00:07
→ alex__c2022 joined00:07
← jjuran left00:07
← jettero left00:07
← nick_h left00:07
chrisf|work SethRobertson: ok00:07
→ jettero joined00:07
← jettero left00:07
→ jettero joined00:07
← tgunr_ left00:07
→ jjuran joined00:08
← rhl left00:08
→ tgunr joined00:08
→ lukeman joined00:08
← jfkw left00:09
→ nick_h joined00:09
← nadirvar_ left00:09
→ jfkw joined00:09
← simosx left00:10
← chombee left00:11
Chib is there a special way you have to remove .gitignore files?00:12
→ PovAddict joined00:12
canton7 Chib, they're no different to any other file in that respect, no00:12
PovAddict can I rely on the format of git rebase -i input files? I want to do it automatedly, with 'sed' as $EDITOR00:12
Chib hmm, odd00:13
canton7 Chib, what's up?00:13
PovAddict EDITOR='sed -i "s/^pick \(75f5da2\|4660d50\).*//"' git rebase -i ...00:13
Chib it's still showing up on my github repo00:13
SethRobertson subpsyke: Looks like I was wrong. My reading of the man page suggested that it was not using reflog, but clearly it is when I actually tested it00:13
canton7 Chib, you've git rm'd it, committed, and pushed?00:14
← Corren left00:14
Chib yep00:14
← MikhX left00:14
Chib all looked successful00:14
canton7 Chib, different branches?00:14
Chib nah only have a master branch00:15
← beatak left00:15
→ alberto56 joined00:15
canton7 sha-1's of latest commits match?00:15
**sha1's00:15
→ Corren joined00:16
← adamholt left00:16
← jnylen left00:17
← rideh left00:17
→ adamholt joined00:17
Chib weird, i added a test file for the sake of it and commit/pushed it and now the .gitignore is gone too, oh well!00:17
← resixian left00:18
← grey_wolf left00:18
← kinlo left00:18
← clever_ left00:18
canton7 strange... check the diffs for those two commits00:18
← jjuran left00:18
→ grey_wolf joined00:18
→ jjuran joined00:19
→ clever joined00:19
Chib looks ok, i guess maybe github might've just been running a bit slow00:19
→ kinlo joined00:19
→ jnylen joined00:19
→ rideh joined00:19
→ resixian joined00:20
← MacGyverNL left00:20
Chib didn't actually check the github activity page before doing the test commit, probably just server slowness and me being a newb00:20
→ n8o-mba joined00:20
→ MacGyverNL joined00:20
canton7 fair enough. I guess that explains it00:21
← tatsuya left00:21
→ arunce_ joined00:21
← beyondcreed left00:21
← Chaser left00:22
← arunce left00:22
→ tgunr_ joined00:22
acidjazz what is going on here00:23
Your branch is behind 'origin/dev' by 7 commits, and can be fast-forwarded.00:23
wtf is this00:24
→ MikhX joined00:24
canton7 origin/dev has 7 more commits than your local branch00:24
but the history's nice and linear, so if you merge origin/dev into your local branch, it'll be able to do a fast-forward merge with no funky branchy stuff00:25
← tgunr left00:25
← larsl left00:25
← tty1 left00:25
→ Chaser joined00:26
← sraue left00:28
← aalex-home left00:30
PovAddict use merge --ff-only to be sure00:30
→ Textmode joined00:32
→ aalex-home_ joined00:33
← aalex-home_ left00:33
→ aalex joined00:34
→ savoxis joined00:35
→ AkumaStreak joined00:35
→ dyer_out joined00:35
→ d1zzy joined00:36
d1zzy is it expected that "git rev-list --no-walk" output may show the given commits in a different order than they show in "git rev-list" output? I was hoping that one can use "git rev-list --no-walk" to a set of order commits00:36
err to order a set of commits00:37
← alex__c2022 left00:41
patrikf d1zzy: try git rev-list | grep00:43
← PovAddict left00:44
← Chaser left00:46
d1zzy patrikf: that works but I wonder if it is expected that just adding --no-walk changes the order (I also noticed to depend on the order given on the command line, even more strange)00:46
patrikf apparently it is not "expected", or you had not asked00:47
→ sraue joined00:48
← hyperair left00:48
patrikf I don't think it's a bug, possibly poorly documented. But rev-list can't know the order without walking down the graph until it has found all commits you gave it00:48
d1zzy yeah, I was hoping it would walk the graph as much as it needs to determine the order but not display, but it's fine00:49
now, changing subject00:49
any idea why my freshly compiled Git seems to fail any test related to ISO-2022-JP (that is, it fails 4 tests in t3900, they all are involving that encoding, no other tests are involved with it)00:50
→ crash\ joined00:50
d1zzy I even tried the test commands manually, a commit created with that encoding (for the message), doesn't get converted as UTF-8 after switching to UTF-8 encoding (with git config i18n.commitencoding UTF-8)00:51
seems that works for other encodings, strange00:51
← banghouse left00:52
← rchavik left00:52
← Corren left00:52
← alberto56 left00:52
← tjholowaychuk left00:53
→ dyer_ joined00:54
← dyer_ left00:54
→ dyer_ joined00:54
← enherit left00:56
→ shevy joined00:57
← dyer_out left00:58
ChanServ set mode: +v00:58
savoxis → JesusLovesYou00:58
→ warlock_mza joined00:59
→ sunnyohno joined01:00
← aroop left01:01
→ cooloney joined01:04
← vanmik left01:04
→ Rotham joined01:06
→ tty1 joined01:08
→ CareBear\ joined01:09
→ DrDogwelder joined01:10
→ dyer_out joined01:10
→ jablko joined01:11
← n8o-mba left01:12
→ mosno joined01:12
← whyking left01:12
← jtaby left01:14
← arunce_ left01:14
← dyer_ left01:14
→ dreiss joined01:14
← zivester left01:15
← jablko left01:17
→ jablko joined01:17
← jfkw left01:17
→ CannedCorn joined01:18
← banisterfiend left01:21
← bitkiller left01:21
→ crakdmirror joined01:22
← adaptee left01:22
→ zivester joined01:23
→ banisterfiend joined01:23
← canton7 left01:26
← justin-george left01:27
← warlock_mza left01:27
← banisterfiend left01:27
→ n8o-mba joined01:29
← sraue left01:29
← n8o-mba left01:29
← dreiss left01:30
→ Adaptee joined01:31
→ dyer_ joined01:31
← dyer_ left01:31
→ dyer_ joined01:31
← abourg left01:32
← spearce left01:32
← deekayen left01:32
→ abourg joined01:34
← dyer_out left01:35
→ dyer_out joined01:36
→ rchavik joined01:37
← rchavik left01:37
→ rchavik joined01:37
← tchan left01:37
← mdel left01:38
→ tchan joined01:38
→ phaer joined01:39
← tystr left01:39
← dyer_ left01:40
← metcalfc left01:40
← MacGyverNL left01:41
← jophish__ left01:41
phaer Is it possible to rewrite the history to remove a certain String? I keep config files which included passwords in a repo and now i want to share my config but - if possible - without losing history.01:41
→ jophish joined01:42
← iAmerikan left01:42
→ MacGyverNL joined01:42
← MacGyverNL left01:42
→ MacGyverNL joined01:42
→ gretch joined01:43
← Neo left01:43
← flamingspinach left01:43
← kinesis left01:43
→ flamingspinach joined01:43
← sunnyohno left01:43
→ Neo joined01:44
→ iszak joined01:44
← Buglouse left01:44
→ angelsl joined01:45
← Anubisss left01:45
← Lgb left01:45
← jelmer left01:45
→ CrazyGir joined01:45
iszak So I'm trying to add a post-checkout, post-merge and post-receive hook but none of them are executing, there's no .sample prefix, they're simply http://pastie.org/1866465 and yet none of them work, I can't understand why.01:46
CrazyGir → Guest7469201:46
iszak s/prefix/suffix/01:46
← jophish left01:46
→ jophish joined01:46
→ thoth_mes joined01:47
→ hubertchang joined01:47
← Rotham left01:47
SethRobertson phaer: See git-filter-branch01:47
→ aidalgol joined01:48
→ Vampire0_ joined01:48
→ sraue joined01:48
SethRobertson iszak: Is it executable? What happens if you type the pathname to the hook by itself?01:49
← Vampire0 left01:49
iszak SethRobertson, yeah they're set to execute, and when I execute it, they do as they're required01:51
SethRobertson Then I suggest using `GIT_TRACE=1 git checkout master` and see what it says01:51
iszak in the hook?01:52
SethRobertson No, as the command to test the post-checkout hook01:52
iszak how would I execute that..01:52
just in terminal or what?01:52
SethRobertson Yes.01:53
← kermit left01:53
→ kermit joined01:53
← socketwiz left01:53
← jophish left01:54
iszak SethRobertson, I don't see any hooks executing01:54
← bartman` left01:55
SethRobertson Next try using `strace -o/tmp/tr -f -s 80 git checkout master` (hopefully you have strace)01:55
iszak btw I'm not trying to do a checkout01:55
SethRobertson Well, whatever.01:56
you will want to `grep hooks /tmp/tr`01:56
iszak no such directory01:57
→ bartman` joined01:57
SethRobertson It couldn't find the strace command?01:57
What OS?01:57
iszak not that01:57
→ Buglouse joined01:58
iszak just so we're on the same level this is what I'm doing, I've done git init, and now I'm pulling from another repo, I've added the hook post-merge to execute on git pull, all seems correct?01:58
→ jophish joined02:00
iszak the strace worked but it didn't output all that much02:01
← __sorin__ left02:01
← Ryback_ left02:01
← squiggy left02:02
→ xiangfu joined02:03
← rchavik left02:04
→ __sorin__ joined02:05
→ Lgb joined02:07
← chrisdev left02:08
→ Rotham joined02:08
← hynek left02:08
→ nb_ joined02:08
→ boombatower joined02:08
← mikepack left02:09
← wsuff left02:09
iszak SethRobertson, do I need to make the file like post-merge.sh or just post-merge ?02:09
→ astruasdh joined02:11
→ kpreid joined02:12
← nb left02:15
nb_ → nb02:15
→ yuriks_ joined02:15
← MikhX left02:16
→ hobodave joined02:17
← subsonic left02:18
← yuriks left02:18
→ bryanl joined02:20
SethRobertson post-merge02:20
→ tty2 joined02:20
← tty1 left02:20
SethRobertson iszak: I'd recommend doing something very simple. Have a post-checkout hook and fight getting that running so it is easier to test.02:21
If the strace worked, what did `grep hooks /tmp/tr` print?02:21
iszak SethRobertson, /tmp/tr doesn't exist02:21
oh wait now it does, maybe it was a typo - nothing in there02:22
SethRobertson Nothing at all or no hooks lines?02:22
iszak nope02:23
SethRobertson That was not a yes/no question.02:23
iszak there was nothing in that folder at all02:23
→ intripoon_ joined02:23
SethRobertson /tmp/tr should be a file, not a "folder"02:23
iszak oh right02:24
← jophish left02:24
iszak no hooks are executed in there.02:24
→ evanmeagher joined02:24
← astruasdh left02:25
SethRobertson Let me try again. Is /tmp/tr file non-empty? Did `grep hooks /tmp/tr` print anything?02:26
→ f0i joined02:26
iszak yes it's a file, no it's not empty, no the grep returned nothing.02:26
→ jophish joined02:26
→ majonto joined02:26
← mtx_init left02:26
← ki__ left02:27
→ mtx_init joined02:27
← intripoon left02:27
SethRobertson OK, then let us try a different hook and get that working first. Can you do the following: echo "#!/bin/sh" > .git/hooks/post-checkout; echo "date > /tmp/pc-ran" >> .git/hooks/post-checkout; chmod 755 .git/hooks/post-checkout; strace -o/tmp/tr -f -s80 git checkout HEAD^; cat /tmp/pc-ran; grep hooks /tmp/tr02:28
← jophish left02:28
→ dyer_ joined02:28
← dyer_ left02:28
→ dyer_ joined02:28
← evanmeagher left02:29
iszak SethRobertson, !/bin/sh": event not found02:29
SethRobertson Well, I suppose that is progress. It is trying to run the hook. I'm not entirely sure how the "#" disappeared in front of the !, but whatever02:30
Did the grep return several lines?02:30
iszak also can I replace git checkout with git clone?02:30
← Yuuhi left02:31
← mtx_init left02:31
← dyer_out left02:31
iszak SethRobertson, okay I changed checkout to clone and I've got you some output!02:32
→ evanmeagher joined02:33
→ amerine joined02:33
→ Escherial joined02:33
← thoth_mes left02:33
Escherial hey, is it alright to use --assume-unchanged to allow a file to remain in my repo, but not push changes to it when i push to a remote?02:33
→ jophish joined02:33
iszak SethRobertson, here you go http://pastie.org/private/e8ivallkwqgclmbehf5p7a02:34
Escherial i am actually changing the file (it's the classic example of a settings file that contains sensitive data that i don't want to appear on github, yet won't work at all if i omit it using .gitignore)02:34
→ rchavik joined02:35
Escherial i am aware that attempting to merge changes from a pull onto my changed (but ignored) settings file will fail, which i imagine is good?02:35
← jophish left02:35
SethRobertson iszak: You cannot have any hooks run on clone. When it is creating the .git directory there is no hooks present to run02:35
← ketas left02:36
SethRobertson It certainly seems as if the checkout version *did* run a hook. Did the grep really not return anything?02:36
→ n8o-mba joined02:36
SethRobertson Escherial: Common wisdom is assume-unchanged is evil to use unless the file is actually unchanged.02:37
Escherial: http://thomasrast.ch/git/local-config.html02:37
Escherial i've heard such common wisdom, which makes me uncomfortable :| i'm not sure what else to do, though02:37
ah, that link seems to help :) thanks02:37
SethRobertson Best practice is not to commit config files but rather configfile.sample and .gitignore configfile02:39
→ jophish joined02:39
← rchavik left02:39
Escherial that makes a *huge* amount of sense02:39
i was thinking that all of this mucking about for such a common thing was way too complicated for the naive user02:39
this'll also force people to actually read the config file, too :) or at least be aware of its existence02:40
iszak SethRobertson, but it says post-checkout "It is also run after git clone, unless the --no-checkout (-n) option is used. The first parameter given to the hook is the null-ref, the second the ref of the new HEAD and the flag is always 1."02:40
Escherial (since they'll have to rename it if they expect anything to work)02:40
anyway, i have to go running as a penance to the fitness gods x_x thank you for the advice02:40
→ konr joined02:41
→ jfkw joined02:41
→ hyperair joined02:41
SethRobertson iszak: It may be called after git-clone, I am just dubious that someone can get the necessary file in there. The checkout case is a good test case02:41
→ mando joined02:41
konr what's the right way to remotely pull a repo using ssh?02:42
SethRobertson pull or clone?02:42
← Escherial left02:42
← jophish left02:42
konr pull02:42
→ jophish joined02:43
konr my `ssh $host git --git-dir=konr.mobi/.git pull` somehow breaks the git repository02:43
it doesn't get updated but can no longer pull02:43
→ alester joined02:43
SethRobertson Wait, where are you trying to run the git command? Locally or on $host?02:44
konr I'm locally trying to pull in a remote host02:44
iszak SethRobertson, so I'm not sure how I'm meant to checkout.. I only know how to clone, I've got a remote repo how would I checkout that?02:44
SethRobertson That command was trying to run git on $host, not pull $host. Try this instead: git pull $host:knor.mobi02:45
konr: If you plan do that often, you should add that as a git-remote02:45
iszak: git clone, create the hook, then run `git checkout <something>`02:45
← dyer_ left02:45
→ dyer_out joined02:46
← DrDogwelder left02:46
→ mtx_init joined02:46
offby1 konr: did you figure it out?02:46
iszak SethRobertson, I try and do that with checkout and get.. error: pathspec 'HEAD' did not match any file(s) known to git.02:48
→ mdel joined02:48
konr SethRobertson: no, that's what I'm trying to do: the repo konr.mobi is on github, so the workflow, after updating it locally, push to github and then, on $host, pull it.02:48
offby1: not yet02:48
*is to push it to github02:49
← mdel left02:49
offby1 yeah. You're connecting to $host -- which is remote to you -- and then pulling from github to it.02:49
That sounds perfectly reasonably.02:49
reasonable.02:49
SethRobertson konr: Ah, that was not well understood before. ssh $host 'cd knor.mobi; git pull'02:49
iszak konr, I would echo it, get the output, run it directly, see if there is a syntax error.02:49
offby1 If I were you, I'd do "ssh $host" and get a shell, then try the pull from an interactive shell prompt. I can't explain why, but I think it'd be more illuminating.02:49
→ stliu joined02:50
SethRobertson iszak: What happens if you try to run the post-checkout hook directly?02:51
iszak SethRobertson, it executes, no errors02:51
pc-ran file is created02:51
← AkumaStreak left02:51
→ banisterfiend joined02:52
→ leo2007 joined02:52
SethRobertson Then when you say `git checkout HEAD` what happens?02:52
konr thanks for the suggestions, guys!02:53
SethRobertson What worked?02:53
→ LongBeach joined02:54
→ etcetera_ joined02:54
→ gnufied joined02:54
→ gnufied_ joined02:54
konr SethRobertson: I used your suggestion!02:54
iszak SethRobertson, we02:55
offby1 idly wonders why the original command didn't work02:55
iszak we've got progress, dumping new data02:55
← SerajewelKS left02:56
iszak SethRobertson, http://pastie.org/private/s5lfh7eiyuv9hwzdmldg02:56
SethRobertson And /tmp/pc-ran does not exist?02:57
iszak no it does now, nvm02:57
SethRobertson OK, so that is an example of a successful hook execution. strace being your friend.02:57
iszak yeah02:58
SethRobertson Now you can try post-merge that you have the nits worked out. Rename the post-checkout hook and try `strace -o/tmp/tr -f -s80 git pull`02:58
← etcetera left02:58
SethRobertson Of course, you will need to ensure that there is a change in the upstream remote to actually merge!02:58
← Rotham left02:59
→ shabble joined02:59
← aidalgol left03:00
shabble is there a way to tell which files will be committed (to go inside a pre-commit hook script)?03:00
SethRobertson shabble: Presumably you can look at git-status for the bits in the index03:01
Or git-diff the index vs HEAD or...03:02
← kukks left03:04
→ ketas joined03:04
shabble yeah, git status --porcelain is probably the way to go, once I figure out what the prefixes mean03:04
→ fivetwentysix joined03:04
→ cylence joined03:08
→ Corren joined03:08
← caseymcg left03:09
→ caseymcg joined03:10
→ fivetwentysix_ joined03:10
iszak SethRobertson, can't I have a hook execute on the initial cloning?03:12
I kind of need it.03:12
→ ajpiaNOU joined03:12
SethRobertson I do not believe so, because the git-clone is creating the .git directory and you will not be able to get the hook in place at the right moment03:12
iszak can't I git init; something; something else?03:13
← fivetwentysix left03:13
→ xerora joined03:14
← fivetwentysix_ left03:15
← xiangfu left03:15
← n8o-mba left03:15
→ xiangfu joined03:16
← rpg left03:18
← alester left03:19
← cschneid_ left03:20
← infid left03:20
iszak SethRobertson, will git init; git remote origin username@hostname:/path; git pull origin master; execute the post-checkout hook?03:20
→ rchavik joined03:21
→ infid joined03:22
← subpsyke left03:23
← xerora left03:23
→ hynek joined03:24
→ banghouse joined03:24
← JesusLovesYou left03:28
→ savoxis joined03:29
iszak SethRobertson++ nvm got it :) thanks for all your help!03:32
← aef left03:34
← blorbx left03:34
← boombatower left03:34
→ B2oba joined03:35
→ aef joined03:36
← mosno left03:39
← CannedCorn left03:39
← iszak left03:39
→ kinesis joined03:39
→ mosno joined03:39
← hobodave left03:39
→ justin-george joined03:39
← caseymcg left03:41
← mosno left03:41
→ mosno joined03:42
→ vsayer joined03:44
← ajpiaNOU left03:46
→ blorbx joined03:46
← pyroscope left03:47
→ pyroscope_ joined03:47
→ aixl joined03:48
← gnufied_ left03:50
← gnufied left03:50
→ eletuchy joined03:51
→ joshteam joined03:51
← p3rror left03:52
→ CannedCorn joined03:53
← pyroscope_ left03:56
← paeva left03:57
→ pyroscope joined03:57
← savoxis left03:57
→ flazz joined03:57
flazz i have a submodule that says the difference is the same sha but with -dirty appened to it, what could be wrong?03:58
→ Error404NotFound joined03:59
→ gasull joined04:00
← jophish left04:02
gasull Is there a way I can delete everything, including history, in a remote repo? Thanks.04:02
mmattice flazz: cd into it, check git status04:02
← hagabaka left04:03
flazz mmattice: adding ignore = dirty to .gitmodule fixed it04:03
btw what does -dirty mean at the end of a commit-id?04:03
mmattice for some value of "fixed"04:03
flazz: cd into it, check git status04:03
← aixl left04:03
→ boombatower joined04:03
← boombatower left04:03
→ boombatower joined04:03
flazz untracked files exist04:04
generated files04:04
← ISF left04:04
← OOPMan left04:04
flazz will top-level .gitignore apply to submodules?04:04
mmattice shouldn't.04:04
→ seivan joined04:05
→ hagabaka joined04:05
← hagabaka left04:05
→ hagabaka joined04:05
← bryanl left04:06
flazz right04:07
← kenneth_reitz left04:07
← theoretical left04:10
→ envi_ joined04:11
→ EricInBNE joined04:11
→ ISF joined04:12
→ savoxis joined04:13
← envi left04:14
← [intra]lanman left04:15
← leo2007 left04:18
← joshteam left04:18
← smolyn left04:20
← phantomcircuit left04:21
← seivan left04:22
→ OOPMan joined04:23
→ GothAlice joined04:24
ivey → ivey_off04:24
← ivey_off left04:25
→ joshteam joined04:25
→ aminpy joined04:26
← joshteam left04:27
GothAlice It's getting late here and I'm starting to forget simple things. ^_^ How do you merge changes from a branch (say, master) into the 'current' branch (say, fuji) without removing the first (updated) branch?04:27
← mosno left04:28
→ phantomcircuit joined04:29
← savoxis left04:29
→ erichammond joined04:31
→ slackorama joined04:32
→ zhengyang joined04:34
→ ivey joined04:36
← gebbione left04:36
milki merging doesnt remove branches04:38
whats the problem? o.O04:38
→ glcrazy joined04:39
← amerine left04:40
→ drev1_ joined04:40
← drev1 left04:41
drev1_ → drev104:41
← jawwad left04:41
→ savoxis joined04:42
← phantomcircuit left04:42
→ gebi joined04:42
← crakdmirror left04:42
→ Tobbe joined04:43
← Error404NotFound left04:43
← gasull left04:43
→ joshteam joined04:43
→ mdel joined04:44
← glcrazy left04:44
Tobbe Hi. I'm trying to do "git diff stash@{0}^1 stash@{0} -- <filename>" but I get "fatal: bad revision 'stash@0^1'". What am I missing?04:44
→ jawwad joined04:44
← quiliro left04:45
← joshteam left04:45
← wonginator1221 left04:46
milki you can diff stashes? i didnt know that04:47
i though states were diffs themselves04:47
stashes*04:47
GothAlice AFIK stashes are branches.04:47
(Or at least implemented in a very branch-like way.)04:48
milki hm that makes sense from what i see in qgit04:48
Tobbe: try using the reference directly?04:49
resf/stash04:49
→ dannas_ joined04:49
Tobbe this is where I got the command from, you should read it, it's interesting :) http://stackoverflow.com/questions/1105253/how-would-i-extract-a-single-file-or-changes-to-a-file-from-a-git-stash04:49
← CannedCorn left04:50
← AtomicSpark left04:52
→ gnufied_ joined04:53
→ gnufied joined04:53
→ roop joined04:55
← evanmeagher left04:55
Tobbe I just tried "git show stash@{0}:<full filename>" but got "fatal: ambiguous argument 'stash@0:<filename>: unknown revision or path not in the working tree. Use '--' to separate paths from revisions"04:56
→ freimatz joined04:56
→ Error404NotFound joined04:57
→ steph021 joined04:58
← steph021 left04:58
→ steph021 joined04:58
Bombe Tobbe, have you tried using “--” to seperate paths from revisions?04:58
Tobbe :p04:59
yes :) Then I get "fatal: bad revision 'stash@0'"04:59
wereHamster Tobbe: on windows?04:59
Tobbe (I'm not liking how it strips the { } from everything I type)04:59
→ phantomcircuit joined04:59
Tobbe wereHamster: Solaris04:59
wereHamster you have a weird shell05:00
quote the argument05:00
← gigamonkey left05:00
→ gigamonkey joined05:00
→ dr0id joined05:00
Tobbe that worked (at least for "git show")05:00
I'll try the stash diff too05:01
← eletuchy left05:01
→ caseymcg joined05:02
Tobbe didn't work. "fatal: ambiguous argument 'stash@{0}^1 stash@{0} -- <filename>: unknown revision or path not in the working tree. Use '--' to separate paths from revisions"05:02
wereHamster please paste the exact command05:03
dr0id → depressed05:05
← caseymcg left05:06
Tobbe (not allowed to paste the exact path, made some changes to it) git diff "stash@{0}^1 stash@{0} -- dev-network/src/com/company/product/tree/conditions/IpAddress.java"05:06
→ caseymcg joined05:06
← towski left05:06
Tobbe (replaced the actual company named with "company" and the product name with "product")05:07
← phantomcircuit left05:08
← caseymcg left05:08
→ caseymcg joined05:08
→ ThomasLocke joined05:10
← gigamonkey left05:10
wereHamster you need to quote each argument separately, not everything together05:10
← shevy left05:11
Tobbe "--" "<filename>" or "-- <filename"?05:11
→ eletuchy joined05:11
wereHamster -- and <filename> are different arguments05:12
(otherwise there woudn't be space between them)05:12
Tobbe yeah, got it05:12
what a pita...05:12
wereHamster solaris. Tell me bout it..05:13
Tobbe heh05:13
→ maleknet joined05:13
→ canuck3141 joined05:13
← slackorama left05:15
← caseymcg left05:16
→ caseymcg joined05:16
Tobbe So this is my workflow up until now: git checkout master; git pull; git checkout -b bug_id master; "writing testcase"; "fixing bug"; git rebase master; "cannot rebase: you have unstaged changes"; "what to do?"; "lets stash everything"; git rebase master;05:20
is stashing the best thing to do at that point so that I can rebase?05:20
→ fr0stbyte joined05:20
→ TrineLocke joined05:20
Tobbe I have changes in files that I'm not going to commit, so I didn't want to stage them05:20
fr0stbyte i've added my ignore directories to /.git/info/exclude, and git status shoes no untracked files or changes. good. but when i go to checkout a different branch: The following untracked working tree files would be overwritten by checkout05:21
Tobbe But I also don't want to lose my changes because I need them to be able to run the code05:21
← Pupeno left05:21
→ _Qman joined05:23
wereHamster Tobbe: ask yourself: why do you have uncommitted changes?05:23
fr0stbyte: the files are tracked in one branch and untracked in the other branch. Either track or not track them, but don't mix it05:23
fr0stbyte wereHamster: genius. i forgot i accidentally committed them to the other branch. accident.05:24
wereHamster: thanks. appreciated05:24
Tobbe wereHamster: because I had to work around a bug in an old version of a third party library that I'm not in control over. And the people in charge of the code uses a newer version of said library that doesn't have the bug05:24
← fr0stbyte left05:25
wereHamster Tobbe: so what do you want to do with those change?05:25
→ fr0stbyte joined05:25
Tobbe I just want to keep them local to myself05:25
should I check them in, and just not push them?05:25
→ shevy joined05:26
Tobbe (how do I select what commits to push?)05:26
wereHamster can't you updatet he library?05:26
← canuck3141 left05:27
← etcetera_ left05:28
Tobbe No, I don't have write permission to the path where it's stored05:28
← phaer left05:28
→ phaer joined05:28
→ bindaas joined05:28
wereHamster so, you're using some kind of hook which restricts the paths to which you can commit?05:29
Tobbe google tells me I can't commit my changes and then later not push it, as any newer commit would depend on the earlier commit05:29
wereHamster: No. I have a library in my classpath that's in a directory I'm not allowed to touch05:30
wereHamster is that library part of your git repo?05:30
Tobbe no05:30
wereHamster so where do these uncommitted changes come from?05:30
banisterfiend wereHamster: can you give me an intuitive explanation of rebase --onto05:31
wereHamster Tobbe: "cannot rebase: you have unstaged changes"05:31
where are these changes?05:31
banisterfiend: I believe man git-rebase explains it pretty well05:31
jast banisterfiend: [automatic message] the 'git-rebase' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-rebase.html05:31
wereHamster banisterfiend: just read the description. carefully05:31
→ ExtraSpice joined05:31
banisterfiend wereHamster: i prefer an explanation from someone with a good heart, and you look like a man with a good heart05:32
wereHamster no time05:32
← blorbx left05:32
Tobbe wereHamster: what do you mean "where"? they are in the file that uses the library with a bug05:32
→ blorbx joined05:32
wereHamster Tobbe: so the library code *is* tracked in git05:32
Tobbe no, the library is just a .jar05:32
the code that *uses* the library is tracked in git05:33
→ evanmeagher joined05:33
Tobbe and it's the code that *uses* the library I did my work-around in05:33
← cooloney left05:33
ivey → ivey_off05:33
Tobbe and I don't use the same version of the lib/.jar as the team responsible for the upstream repo05:33
(because I can't change the version of the lib I use)05:34
→ joshteam joined05:34
GothAlice Is 'rebase' the correct operation here? https://gist.github.com/956586#L61 These per-server branches are linear (no merging or additional branching; all immediate descendants of master)05:34
← joshteam left05:34
GothAlice I chose rebase since it appears to be reversible on failure, whereas merge leaves potentially invalid files with markers in them.05:35
← giallu left05:35
← blorbx left05:36
milki i dont see why you need to merge back to master if each server has their own branch already05:36
do the branches share files?05:37
GothAlice They do.05:37
← evanmeagher left05:37
GothAlice And "master" is treated as files common to every server, vs the files modified within the branch which are server-specific.05:37
→ brienna joined05:38
milki oo ok, so you are updating the branch with new master stuff05:38
ya, rebasing makes sense then05:38
GothAlice \o/ Thank you. :)05:39
wereHamster Tobbe: looks like you need to change your setup rather than the code05:39
← brodo left05:39
milki do you need git branch though? the branch should already exist in origin05:39
git checkout might auto create a tracking branch05:40
GothAlice milki: This script is executed on each boot for both existing servers and new. The branch statement there ensures the branch exists before continuing.05:40
milki or git branch should be modified to do tracking05:40
oo05:40
GothAlice Yay for virtual machine images, eh? ;)05:40
milki i still think there needs to be some if somewhere if the branch already exists05:41
Tobbe wereHamster: Yeah, I agree. But I can't. I can ask management to do it, but it will take weeks, and they want this bug fixed asap...05:41
milki probably dont need git push --all since you know exectly what branch you modified05:41
← f0i left05:42
GothAlice milki: the test / ewarn line seems sufficient enough, and simpler than grep magic to parse the output of 'git branch'.05:42
milki git branch will succeed whether or not the remote branch exists05:42
→ f0i joined05:42
milki git push will fail if it exists in remote but your branch isnt tracking i think05:43
← brienna left05:43
milki o, so its ok for the script to fail?05:43
← mdel left05:43
→ brienna joined05:44
→ rndbit joined05:44
← rndbit left05:44
GothAlice milki: git branch <name> returns a non-zero status code if the branch exists. That's what I'm trapping.05:45
← brienna left05:45
← xiangfu left05:45
← jfkw left05:45
→ rndbit joined05:45
GothAlice milki: For the most part I have lots of exit-early returns from the script to handle errors (unexpected non-zero return codes).05:45
milki if it exists locally yes05:45
but not for remote05:45
GothAlice Define "it".05:46
The branch?05:46
→ harshu_ joined05:46
milki ya05:46
if its on origin/, git branch doesnt care05:46
GothAlice Hmm. 99% of the time either the remote branch doesn't exist (new server), or does exist and is also present locally (existing long-running server). Only a new server with an old name would encounter that.05:47
→ cooloney joined05:47
milki oo05:48
that works then05:48
GothAlice Wait; am I missing something between fetch and switching branches? Do I need to fast-forward master first?05:49
milki o ya, that would be useful05:50
or maybe you can rebase off of origin/master?05:50
GothAlice Hmm; that would be handy.05:50
← Corren left05:50
← maleknet left05:50
GothAlice Quick Q, too: how do I delete a branch both locally and up-stream?05:51
← f0i left05:51
milki uh05:51
two commands05:51
branch -d/-D05:51
push origin :branchname05:51
GothAlice Thanks. ^_^05:52
← Adys left05:54
→ airborn joined05:55
GothAlice I wonder if there's anyone else who publicly uses Git in this way (for server configuration management in a cluster)?05:55
→ Adys joined05:56
← Error404NotFound left05:58
→ jutaro joined05:58
← envi_ left05:58
→ envi_ joined05:59
← justin-george left06:00
← LongBeach left06:00
← Octalot left06:01
← rndbit left06:01
→ namemon joined06:01
← savoxis left06:02
→ rndbit joined06:02
← ISF left06:03
→ savoxis joined06:04
→ glcrazy joined06:04
← tty2 left06:04
→ thiago_berlin joined06:05
→ msponge joined06:06
→ n1ck3 joined06:08
← thiago_berlin left06:09
→ thiago_berlin joined06:09
→ cbreak_work joined06:09
→ ISF joined06:10
→ Swaagie joined06:10
→ n1ck31 joined06:11
← n1ck3 left06:11
→ f0i joined06:11
→ Error404NotFound joined06:12
← rchavik left06:12
← hagabaka left06:12
← bronson left06:13
← mando left06:15
→ hagabaka joined06:16
→ opalka joined06:16
← intripoon_ left06:17
← vdv left06:17
→ intripoon joined06:18
← gretch left06:18
→ tvw joined06:18
→ urkud joined06:19
← __sorin__ left06:21
→ kylemcgill joined06:22
→ Titosemi joined06:23
← ISF left06:24
← rndbit left06:24
milki well06:25
you can checkout out confman06:25
but it uses svn as a backend06:25
→ aalex-home_ joined06:25
milki http://www.timesinks.net/projects/confman06:25
it might do well with a git backend06:25
but its not supported as of yet06:25
← jettero left06:27
← henrikbjorn left06:27
→ jettero joined06:28
← jettero left06:28
→ jettero joined06:28
← sbell left06:29
← konr left06:30
milki GothAlice: ^06:30
← envi_ left06:31
→ rndbit joined06:31
← rndbit left06:31
→ sbell joined06:31
→ MikhX joined06:32
→ ISF joined06:32
← hyperboreean left06:32
→ rndbit joined06:32
← Chib left06:32
→ ph^ joined06:33
← sbell left06:33
→ __sorin__ joined06:33
→ psoo joined06:34
← tvw left06:35
→ dim_ joined06:36
← kokx left06:37
← urkud left06:37
→ kokx| joined06:37
→ envi joined06:37
← babilen left06:37
← harv left06:39
← jjuran left06:39
→ rosco_y joined06:39
→ Spockz joined06:39
→ workmad3 joined06:40
→ strk joined06:41
strk git svn clone http://dolphin-emu.googlecode.com/svn/trunk/ -r 750006:41
^^^ quietly seems to do nothing, ideas ?06:41
← Spockz left06:42
← guardian left06:42
← GothAlice left06:43
← envi left06:44
→ envi joined06:44
rosco_y This is the output of my "git status" command. How do I need to do to commit my "additions?" : http://paste2.org/p/139998506:44
This is the output of my "git status" command. How do I commit my "additions?" : http://paste2.org/p/1399985, (fixed grammo)06:45
I do recall running "git reset "HEAD", which might be the problem....06:46
strk git commit06:46
oh, you're not on a branch06:47
where do you want to commit those changes ?06:47
→ Badiss joined06:47
strk git checkout -b newthings; git commit -m "my new things"06:47
rosco_y my .git directory is in the root of my working directory --is that bad?06:47
strk no, it's expected to be like that06:47
rosco_y first I need to check them out?06:48
→ dpino joined06:48
→ kenneth_reitz joined06:48
rosco_y this seems a bit more complicated than it should be....I modifed the files, and they should be checked in.06:48
strk you did something else06:49
or you wouldn't be in "(no branch)"06:49
check 'git branch'06:49
→ giallu joined06:49
← giallu left06:49
→ giallu joined06:49
strk tells you which local branches you have06:49
rosco_y strk: yes, I'll bet it's because I ran "git reset" the other day. Things haven't been the same since06:49
strk what argument did you pass to "git reset" ?06:49
rosco_y HEAD, I believe06:50
it was the other day, and I'm just learning git06:50
strk uhm, shouldn't be in no-branch if you used HEAD06:50
milki history!06:50
rosco_y :) I'll take a look06:50
strk anyway, since you're learning... see what 'git branch' tells you06:50
add a -v for more06:50
and also, I'd recommend checking out 'git show-ref'06:50
rosco_y git06:51
ross@linux-aoty:~/Projects/Thumprint> git branch -v06:51
* (no branch) efc4ec2 initial commit06:51
master 57610cb deleted: www/www/Views/Tree/Index_private_error.aspx06:51
→ pdelgallego joined06:51
← amirite left06:51
strk rosco_y: so it tells you have two local branches06:52
one is unnamed "(no branch)" and is the current one "*"06:52
the other is named "master"06:52
rosco_y should I just delete my .git directory, and re-run "git init"06:52
strk no06:52
rosco_y how can I move back to the master?06:52
→ radarek joined06:52
strk you can switch between a branch and another using 'git checkout <branchname>"06:53
← airborn left06:53
rosco_y so, "git checkout "master" ?06:53
strk but note that unnamed branches would be lost06:53
yes06:53
→ amirite joined06:53
rosco_y cool, that's what I was looking for06:53
thanks strk!06:53
strk are you sure you don't need those changes saved ?06:53
let me know how it proceeds, as there's always somethign to learn for me :)06:54
rosco_y strk: the changes are trivial at this point--I'm just trying to learn how to use git before continuing in my project06:54
← Tobbe left06:54
strk so go one step further and see how to "backup" your changes (if any)06:54
rosco_y I ran the checkout, now git status tells me I'm not on a branch06:54
strk maybe the checkout failed06:55
no error/warning message ?06:55
something like "can't checkout for this and that"06:55
rosco_y yup, error: your local changes to the following files wouel be overwritten by the checkout06:55
strk eh :)06:55
so it's telling you that you have some files changed06:55
so that checking out master would override those changes06:56
rosco_y can I force the checkout?06:56
strk good06:56
→ Chib joined06:56
strk probably06:56
rosco_y :)06:56
strk those changes are in the so-called "index"06:56
rosco_y right, -f06:56
strk (changes to be committed)06:56
let me know if -f did it06:57
rosco_y so to keep my current changes, I should probably delete my ".git" dir, and rerun git init?06:57
strk no06:57
never wipe the .git dir :)06:57
→ D5 joined06:57
strk it's full of helpful things06:57
← kenneth_reitz left06:57
strk you could give your branch a name06:58
git branch aname # will create a branch from your current branch06:58
rosco_y well, I doubt mine has too much helpful stuff, so far it's just been an exercise in frustration06:58
→ d0k joined06:58
rosco_y ah, let me try that06:58
strk but hold on a sec06:58
thing is your branch "(no branch)" isn't really interesting06:59
as those changes are not in it yet06:59
→ rchavik joined06:59
← rchavik left06:59
→ rchavik joined06:59
strk it'd take a "git commit" to really register those changes into a branch06:59
← OOPMan left06:59
strk or "git reset" to take them off the "index"06:59
→ RevoOf1 joined06:59
← RevoOf left06:59
← ISF left07:00
strk try "git reset HEAD www/www/www.csproj"07:00
rosco_y I'm trying to read ProGit online07:00
strk doesn't know if that will work07:00
← __sorin__ left07:00
rosco_y doesn't know if reading ProGrit will work either :P07:01
← thiago_berlin left07:01
rosco_y do you use a gui?07:01
← vsayer left07:02
→ ISF joined07:03
→ OOPMan joined07:03
→ sebbi joined07:03
rosco_y Thanks strk, I think I'll just have to read the manual....:(07:04
← pdelgallego left07:06
strk I try to avoid the gui07:07
sorry, got a system crash07:08
did you try the reset command ?07:08
← lresende left07:09
← mihar left07:09
→ lresende joined07:10
← lresende left07:10
→ lresende joined07:10
→ mihar joined07:11
→ dreiss joined07:12
→ mastroDani joined07:12
← depressed left07:12
→ vsayer joined07:12
← lresende left07:12
→ akheron joined07:13
→ amerine joined07:15
→ saschagehlich joined07:16
← oc80z left07:17
← alip left07:17
→ oc80z joined07:17
← savoxis left07:17
← mihar left07:18
← opalka left07:18
→ henrikbjorn joined07:18
← f0i left07:18
← henrikbjorn left07:19
→ henrikbjorn joined07:19
← blast_hardcheese left07:19
← tokkee left07:20
→ mihar joined07:20
→ jugg joined07:20
→ opalka joined07:21
→ tokkee joined07:21
→ alip joined07:21
→ tvw joined07:22
→ f0i joined07:23
→ blast_hardcheese joined07:25
← caseymcg left07:26
jugg pushing a new branch to a remote is failing ( https://gist.github.com/956665 ) this used to work: $ git push origin HEAD:newbranch07:26
strk what is it supposed to do ?07:26
jugg I've recently upgraded my workstation to 1.7.5, but I don't believe the server version (1.7.2.3) has changed recently.07:26
strk knows $ git push origin newbranch07:27
jugg takes the ref in HEAD and creates 'newbranch' ref on origin.07:27
← Edulix_ left07:28
→ caseymcg joined07:28
→ Rotham joined07:29
strk didn't know you could do that w/out checking out HEAD07:29
→ mikl joined07:29
← envi left07:29
← mikl left07:29
→ mikl joined07:29
→ envi joined07:29
→ savoxis joined07:30
→ avinashhm joined07:32
rob_w_away → rob_w07:33
→ jds joined07:33
→ legumbre_ joined07:33
→ dr_ick joined07:34
← legumbre left07:35
← dreiss left07:40
← MikhX left07:40
← workmad3 left07:43
← RevoOf1 left07:43
← txxt left07:43
← rosco_y left07:45
→ harv joined07:46
← amerine left07:47
← gebi left07:48
← saschagehlich left07:49
→ nevyn^ joined07:49
← ph^ left07:50
→ ph^ joined07:50
← nevyn^ left07:51
→ nevyn^ joined07:51
→ brodo joined07:52
→ ch077179 joined07:54
→ ph^_ joined07:54
← ph^_ left07:54
→ guardian joined07:54
← ph^ left07:54
← ch077179 left07:54
→ ph^ joined07:54
→ shiba_yu36 joined07:55
→ Luca_Peck joined07:55
→ ch077179 joined07:55
→ dharrigan joined07:56
← dharrigan left07:56
→ dharrigan joined07:56
→ dr0id joined07:58
dr0id → depressed07:58
← mastroDani left07:58
← f0i left07:59
→ eijk joined07:59
← Tuomas left08:00
← fs-sylr left08:00
→ ph^_ joined08:00
→ fs-sylr joined08:00
→ troessner joined08:00
← cthuluh left08:01
→ nevyn^_ joined08:02
→ noonoooo joined08:02
← ph^ left08:03
← cylence left08:03
→ mastroDani joined08:03
→ davzie joined08:03
→ Henrik_ joined08:04
→ sechrist joined08:04
← dharrigan left08:05
← henrikbjorn left08:05
← troessner left08:05
→ dharrigan joined08:05
← nevyn^ left08:06
nevyn^_ → nevyn^08:06
← D5 left08:07
← steph021 left08:07
→ D5 joined08:08
→ crashanddie joined08:08
→ mklappstuhl joined08:08
→ Ryth joined08:09
← shiba_yu36 left08:10
← jawwad left08:11
→ genbattle joined08:11
← gnufied left08:12
← gnufied_ left08:12
→ LuckyNeo joined08:12
← sigh left08:12
→ jawwad joined08:14
jayne can I rebase -i through merges, or does that still not work?08:15
→ xiangfu joined08:15
empity jayne: yes you can08:17
→ alus joined08:17
alus I accidentally ran 'git stash apply' when I didn't mean to08:17
how do I undo that?08:18
it caused merge conflicts in a few files, but I can't "git checkout foo" because "error: path 'foo' is unmerged"08:18
→ konr joined08:18
jayne did you have any unstaged/uncommited changes before the apply?08:18
strk git reset HEAD ?08:18
alus jayne: no08:19
jayne if you're sure, you can "git reset --hard HEAD" to wipe out everything back to what you had at HEAD08:19
→ avdg1 joined08:19
alus alright08:19
seems to have worked08:19
→ shiba_yu36 joined08:19
alus thanks!08:20
← rndbit left08:21
← alus left08:21
← dr_win left08:21
→ dr_win joined08:22
← avinashhm left08:22
← avdg left08:23
→ rndbit joined08:24
→ vanmik joined08:25
← shiba_yu36 left08:25
→ jatt joined08:25
→ willvarfar joined08:26
← jawwad left08:27
→ RaceCondition joined08:28
willvarfar I am working in a branch; "git show-ref -s HEAD" gives the master head hash however; how do I get the hash of the current branch (which may or may not be master)08:29
?08:29
← jugg left08:29
→ avinashhm joined08:30
→ jawwad joined08:31
→ dpy joined08:32
selckin you know both hashes can be the same?08:32
willvarfar yeap08:33
I want the hash that I'd get as a comment in git archive08:33
jast HEAD = whatever is currently checked out08:33
selckin show-ref --head HEAD, works for me08:33
→ d3vic3 joined08:33
jast it has nothing to do with master unless master is checked out08:33
strk it's broken here08:33
← roop left08:33
strk -s, --hash[=<n>] only show SHA1 hash using <n> digits08:33
→ snearch joined08:33
strk --hash=2 works; -s 2 doesn't08:33
jast well, yes08:34
willvarfar if you are on a branch, git show-ref HEAD seems to not show the current branch's hash for me08:34
jast hence the "-s, --hash[=<n>]", rather than "-s[ <n>], --hash[=<n>]"08:34
willvarfar git show-ref HEAD08:34
5ad7c90b19af703584ec6f55f6de1710b3d92525 refs/remotes/origin/HEAD08:34
← MissionCritical left08:35
willvarfar git show-ref08:35
5ad7c90b19af703584ec6f55f6de1710b3d92525 refs/heads/devel08:35
a0426e7a03ecd6a8f0d91e758f9c65ff6a625e36 refs/heads/will_108:35
I'm on will_108:35
strk jast: I don't get it... so what's -s supposed to do ?08:35
← avinashhm left08:35
jast strk: show only the SHA1 ID, don't include the ref name08:35
as the manpage explains08:35
quoting: "Only show the SHA1 hash, not the reference name."08:35
strk the help screen shouldn't put it on the same line as "using <n> digits"08:36
but thanks :)08:36
willvarfar my show-ref lists all my branches etc as expected; if I ask it for HEAD, however, it's not returning just the head of the current branch08:36
jast it's the very first sentence in the description of the option, though08:36
← dr_ick left08:36
willvarfar ok coming at this another way, is there a way to get the name of my current branch?08:37
jast that's because show-ref accepts a pattern and displays everything that matches it08:37
strk the very first sentence is fine08:37
willvarfar (am writing a bash script)08:37
jast git symbolic-ref HEAD08:37
strk willvarfar: git branch | grep ^* ?08:37
| something else to strip the *08:37
jast or just what I said :)08:38
willvarfar git show-ref -s `git symbolic-ref HEAD` works08:38
now to get double `ticks into my bash script08:38
jast uhh08:38
that's a bit verbose08:39
strk git show-ref --head -s HEAD08:39
jast you probably want: git rev-parse HEAD08:39
→ Spockz joined08:39
strk git show-ref git-svn # fine08:40
git rev-parse git-svn # warning: refname 'git-svn' is ambiguous.08:40
why ?08:40
jast I've already said it08:40
show-ref shows all matches, rev-parse expects something that matches only one object08:41
strk but I don't find other matches for git-svn08:41
ie: git show-ref | grep svn # single match08:41
→ achadwick joined08:41
jast then presumably there is something else that's called "git-svn"08:42
→ mastroWork joined08:42
strk how can I find out what ?08:43
→ shiba_yu36 joined08:44
strk seems to be a commit, which is not any ancestor of my HEAD08:44
→ curtana joined08:45
strk uhm, it is08:45
← Rotham left08:45
strk Apr 13, 74 commits ago08:45
← mastroDani left08:46
← boombatower left08:47
→ Rotham joined08:47
strk maybe .git/refs/remotes/git-svn08:47
nope, that's the one I know about08:47
→ avinashhm joined08:48
← Luca_Peck left08:48
willvarfar my verbose thing works08:50
git show-ref --head -s HEAD will list several because there are heads on all my branches08:50
it happens that the first one is the current branch, but I don't want to get the results and try and get just the first line for my script08:51
→ _Ryth joined08:51
willvarfar git rev-parse HEAD looks right, thanks jast08:52
→ _pingu joined08:52
← willvarfar left08:52
← Ryth left08:53
← Zephyrus left08:53
→ sigh joined08:53
→ alextheger joined08:54
alextheger Hello there08:54
→ MissionCritical joined08:54
← Rotham left08:54
← _pingu left08:54
alextheger how can I use "git format-patch" to get patches for the hole repository. I'm especially interested in the first commit. git format-patch 00000 didn't work.08:55
FauxFaux You have a repository for tracking holes?08:55
jast alextheger: you'll need the --root option08:55
← msponge left08:56
jast e.g. git format-patch --root HEAD to get patches for everything on the current branch08:56
or use the ID of the first commit to only get that08:57
alextheger jast: great. "git format-patch --root" works wonderful08:57
← Henrik_ left08:58
albel727 why patches? why not git-bundle?08:58
← durre left08:59
→ larsl joined08:59
← lanthan left08:59
← lahwran left09:01
← kenyon left09:02
→ g0bl1n joined09:02
← ruskie left09:04
→ hyperboreean joined09:05
→ lahwran joined09:06
→ roop joined09:06
← DrNick left09:07
→ kenyon joined09:07
→ Behi joined09:07
← B2oba left09:08
← jds left09:12
← snearch left09:13
tsp How can I tag the current revision that I'm on, for example just before a git pull, so I can use it in operations? e.g. tag, git pull, git log tag..master, delete tag09:18
← alextheger left09:19
wereHamster tsp: git pull && git log @{1}..09:19
← vanmik left09:19
wereHamster tsp: or: git remote update && git log ..@{u}09:19
→ DrNick joined09:19
tsp Where can I find documentation on these shortcuts?09:20
wereHamster tsp: man git-rev-parse09:20
jast tsp: [automatic message] the 'git-rev-parse' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-rev-parse.html09:20
tsp thanks09:20
← nevyn^ left09:20
wereHamster tsp: @{<number>} refers to the reflog (man git-reflog), @{u} is short for @{upstream}09:21
jast tsp: [automatic message] the 'git-reflog' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-reflog.html09:21
→ ruskie joined09:21
← namemon left09:21
→ namemon joined09:22
← resmo left09:22
→ resmo joined09:23
ludde how do I make a new local branch that automatically tracks the server so 'git push' works?09:24
← jawwad left09:25
→ nevyn^ joined09:26
wereHamster ludde: man git-push -> --set-upstream09:26
jast ludde: [automatic message] the 'git-push' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-push.html09:26
→ canton7 joined09:26
strk ludde: git branch localname remote/remotename09:26
ludde wereHamster: ok thanks09:27
→ jjido joined09:27
wereHamster oh, a new branch. checkout --track09:27
→ kalusn joined09:27
ludde wereHamster: the branch doesn't exist on the server yet, so 'git branch -t -b new_branchname' started tracking the repo itself.09:28
and then weird things happened when i pushed09:28
← namemon left09:28
wereHamster weird as in .. ?09:29
→ xro joined09:30
jjido I am new to git. how do I tell it to use the latest revision of a file when rebase stops on a 3-way merge?09:30
wereHamster there is no 'latest'. There is only yours and theirs09:30
which one do you want?09:30
xro Hi, is there a way to do a "git pull" but only for ONE folder? "select wich files to update"09:30
wereHamster xro: noonoooo09:30
xro: sorry. Just no.09:31
jjido wereHamster: I don't want mine09:31
xro wereHamster, ok, so i have to update all my local repository? i cannot update just a part of it... boring09:31
← canton7 left09:32
cmn xro: you can update part of it, but you can't do a pull, because that's a merge09:33
remember, pull is fetch+merge09:33
xro cmn, so i could only fetch a folder... git fetch folder name?09:34
cmn no, you have to fetch the whole thing09:34
← mklappstuhl left09:34
wereHamster jjido: checkout --ours gives you the file which is in the repo, checkout --theirs gives you the file from the patch that rebase is trying to apply09:34
xro cmn, ok.... i'll do it09:35
wereHamster xro: if you ever feel the need to clone only part of the repository, you should split the repository up into two or more repos09:35
xro wereHamster, yes... it could be a solution...09:35
← rndbit left09:35
← zhengyang left09:37
xro thanks for your help... have a nice day09:38
← xro left09:38
← d0k left09:41
ivan how do I use filter-branch to remove everything except for two files?09:42
← ebi left09:42
ivan can I use zsh's extendedglob?09:43
cmn c09:43
sorry09:43
← caseymcg left09:44
→ ebi joined09:44
wereHamster ivan: I believe filter-branch uses `sh`. so if you've set up your computer so that sh is zsh, then it should work09:47
jjido after checkout --theirs, git tells me to do an add. I found I get an add/add conflict later, should I solve it the same way? (doing rebase)09:47
ivan wereHamster: thanks09:47
wereHamster jjido: yes09:48
jjido thanks09:48
→ workmad3 joined09:49
→ caseymcg joined09:54
← dpino left09:56
← jjido left09:58
← clever left10:00
→ fs-absynthe joined10:01
← strk left10:03
← dwmw2__ left10:04
→ eregon_ joined10:05
← fahadsadah left10:05
← lamont left10:05
← up_the_irons left10:05
→ dwmw2__ joined10:05
→ whyking joined10:05
← Neo left10:05
← ptl left10:05
← dsop left10:06
← milli left10:06
← madduck left10:06
→ up_the_irons joined10:06
→ dsop joined10:06
← silfreed_ left10:06
→ fahadsadah joined10:06
← darkskiez left10:06
← eregon left10:06
→ darkskiez joined10:06
iaj To exclude the folder yikes/ with all of its contents from the index... dont I just have to put yikes/ into .gitignore?10:06
→ silfreed joined10:07
→ madduck joined10:07
→ clever joined10:07
← fs-absynthe left10:07
wereHamster you do10:07
→ ptl joined10:07
→ milli joined10:08
→ jmkeyes joined10:08
→ lamont joined10:08
jmkeyes Hey, I've looked everywhere and can't figure this out: How can I recreate a tree object that was lost in a corrupted repository?10:08
wereHamster you restore it from backup10:09
→ Neo joined10:09
→ saschagehlich joined10:09
ndim Where the backup might be a remote clone.10:09
jmkeyes wereHamster: http://dpaste.com/539008/10:09
iaj wereHamster; oh.. it was still in the index I had to git rm --cached firs.t10:10
→ tydeas joined10:10
tydeas I have `git add a_file` I have not commited yet and want to get this file out of the commit. How can i achieve this?10:10
jmkeyes That is what I get when I attempt to clone the repository. Can you tell me how to restore it from a backup?10:10
← kalusn left10:11
← Textmode left10:12
tydeas is the git rm --cached a_file correct?10:12
← caseymcg left10:13
← TrahDivad left10:14
→ boombatower joined10:15
→ TrahDivad joined10:15
← larsl left10:16
iaj tydeas; i think you need git reset HEAD a_file10:16
tydeas iaj: resolved thanks10:16
iaj it should say so when you do `git status` as well10:17
→ engla joined10:17
← sechrist left10:20
→ mastro joined10:20
→ mikelangelo joined10:22
← mastroWork left10:24
→ ReshadN joined10:25
→ sunnyohno joined10:26
← avinashhm left10:26
← ReshadN left10:26
jmkeyes wereHamster: How can I restore a tree from a backup? I've done 'git ls-tree -z $treeID | git mktree -z' which emitted the same tree ID again when it finished. The problem still exists.10:27
← xiangfu left10:29
← RaceCondition left10:31
→ arunce joined10:34
→ seivan joined10:35
→ rndbit joined10:35
← rndbit left10:35
← saschagehlich left10:35
→ rndbit joined10:36
→ saschagehlich joined10:36
← genbattle left10:37
→ marcob joined10:37
→ alextheger joined10:38
→ NoNoNo joined10:40
ChanServ set mode: +v10:42
NoNoNo Hi! is it possible in "git diff" to highlight inserted text in a changed line? (some kind of "sub line diff")10:42
FauxFaux NoNoNo: man git diff --word-diff10:43
jast NoNoNo: [automatic message] the 'git-diff' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-diff.html10:43
← seivan left10:45
→ seivan joined10:45
← JEEB left10:48
← ISF left10:50
→ avinashhm joined10:50
NoNoNo when I run 'git rebase -i --autosquash HEAD~6' with 'git version 1.7.5.70.g0b9dee',10:54
how can there be a line 'pick 697b7a6 fixup!' instead of 'fixup 697b7a6 fixup! '?10:54
← dyer_out left10:55
← seivan left10:55
FauxFaux NoNoNo: I haevn't used fixup and the manpage isn't clear; does it only work if the commit message is the same, like squash!?10:56
→ dyer_out joined10:56
→ NotTooDumb3 joined10:56
NoNoNo I read about it on <http://technosorcery.net/2010/02/07/fun_with_the_upcoming_1.7_release_of_git_rebase_--interactive_--autosquash.html>10:57
→ gonemad3 joined10:58
FauxFaux Thanks for linking me to a non-canonical piece of documentation for something either of us could have tried in 30 seconds!10:58
gonemad3 → workmad3_10:58
workmad3_ → workmad3__10:58
← workmad3 left10:58
workmad3__ → workmad310:58
NoNoNo http://www.kernel.org/pub/software/scm/git/docs/git-rebase.html10:59
FauxFaux Nice.05/11:56:16 < FauxFaux> NoNoNo: I haevn't used fixup and the manpage isn't clear; does it only work if the commit message is the same, like squash!?10:59
-Nice.10:59
NoNoNo ??10:59
→ Fuu joined10:59
→ Kulrak joined11:00
← Kulrak left11:00
→ Kulrak joined11:00
→ ixti joined11:02
NoNoNo Sarcasm over IRC doesn't work well11:03
← jmkeyes left11:03
wereHamster no?11:04
← Behi left11:05
→ JEEB joined11:05
→ paeva joined11:05
→ canton7 joined11:06
→ brian_g joined11:08
cbreak_work NoNoNo: it totally does...11:09
→ _johnny joined11:09
← envi left11:11
→ envi joined11:11
→ Dave^| joined11:11
← madsy left11:13
← alextheger left11:14
→ dyer_ joined11:16
← dyer_ left11:16
→ dyer_ joined11:16
← dyer_out left11:20
← cooloney left11:20
→ oriba joined11:25
→ urkud joined11:26
← oriba left11:26
← mikelangelo left11:27
← NoNoNo left11:28
→ mikelangelo joined11:28
→ NoNoNo joined11:28
← ph^_ left11:29
← dim_ left11:30
→ lanthan joined11:31
← arunce left11:32
→ RaceCondition joined11:33
← jmcantrell left11:34
depressed → dr0id11:34
→ lmcdowell joined11:35
→ rendar joined11:35
→ soc42 joined11:35
← RaceCondition left11:36
← amirite left11:38
→ chombee joined11:41
→ schweizer joined11:41
schweizer hi folks11:41
→ xshyamx joined11:41
→ jitesh_shetty joined11:41
schweizer how can i set my working dir back to a certain revision?11:41
→ namemon joined11:41
mjt checkout?11:42
← sebbi left11:43
→ sebbi joined11:44
schweizer mjt: git checkout HASH does not work11:44
cbreak_work it does11:44
but it will detach your head11:45
so don't commit anything in that state11:45
← sebbi left11:45
wereHamster ach die schweizer...11:45
cbreak_work wir sind ganz liebe11:45
engla it will detach your head! pretty dangerous stuff11:46
wereHamster I know. I have to deal with you every day :)11:46
→ falican joined11:46
wereHamster (you = die schweizer)11:46
engla especially on case-insensitive filesystems11:46
→ bracki joined11:47
cbreak_work I don't think the filesystem makes things worse11:47
in this context11:47
← banisterfiend left11:47
→ Titosemi1 joined11:48
engla git only detaches head on caseinsensitive filesystems11:48
thm woot?11:48
← hyperair left11:48
cbreak_work there's no reason to yell here11:48
engla there's a big difference between losing your HEAD and losing your head, to me11:49
← Titosemi left11:49
thm ah that was intended to be funny.11:49
→ banisterfiend joined11:49
← NotTooDumb3 left11:50
wereHamster which brings us aback to what nonono said a while back: Sarcasm over IRC doesn't work well :)11:50
→ gonemad3 joined11:50
engla I guess it doesn't11:51
Titosemi1 → Titosemi11:51
engla imagine git newbs reacting to "but it will detach your head"11:51
thm and that is better,m if you SHOUT it?11:51
albel727 lol11:52
← workmad3 left11:52
engla it's not shouting when that's the proper spelling11:52
like USA11:52
thm ah right, much shouting there, too.11:52
← urkud left11:52
albel727 lol11:52
engla exactly what I was saying11:52
cbreak_work I don't think the usa cares if they are yelled or spoken.11:53
← hvgotcodes left11:53
→ tatsuya joined11:53
engla maybe a newcomer will be even more scared to have the detached HEAD thing shouted at him/her11:53
cbreak_work that's a good point.11:54
let's scare people away from doing suicidal stuff, until we run out of work :)11:54
thm the git boot camp.11:54
← canton7 left11:55
→ sebbi joined11:55
albel727 yeah, because we all know, that git should be extinct and superceded by other better vcs-s, like mercurial. at least mercurial doesn't make people shout "DETACHED HEAD" at strangers. </troll>11:56
← mikelangelo left11:56
thm or monotone, which git stole its core ideas from anyway.11:57
engla Hg is poisonous11:57
cbreak_work and compared to git, it's not solid at all11:58
albel727 it's better to die rather than to shout about various detached organs for the rest of your life, don't you think?11:58
gonemad3 → workmad311:58
→ gnufied joined11:58
→ gnufied_ joined11:58
engla cbreak_work: I heard it's pretty mercurial in its operation11:58
thm can sb help me with cgit here?11:59
cbreak_work is that some kind of git fork?11:59
← ormaaj left11:59
thm no, a web frontend11:59
→ sproaty joined12:00
sproaty how can I request push access12:00
wereHamster sproaty: to where?12:01
→ ph^ joined12:01
cbreak_work ask the person that controls the repository12:01
wereHamster sproaty: usually you ask the repository owner or admin.12:01
sproaty fair enough12:01
thought there'd be some request button on the site12:01
→ urkud joined12:02
cbreak_work site?12:02
wereHamster there is none on my site...12:02
→ RaceCondition joined12:03
→ ningia joined12:03
sproaty on here sorry https://github.com/happygiraffe/jslint4java12:03
my bad, I'm being awfully vague :p12:04
→ ormaaj joined12:04
wereHamster sproaty: make a fork and then send back a pull request. You don't need access to his repo12:04
engla in fact, you're not likely to ever get it (or need it)12:04
sproaty good thinking12:05
I've yet to play around with git much so I'm a little lost, thanks12:05
← ph^ left12:05
← stliu left12:06
→ mikelangelo joined12:07
wereHamster sproaty: read a tutorial.12:07
→ bsod1 joined12:07
← dyer_ left12:08
← bsod1 left12:13
← drubies left12:14
← RaceCondition left12:14
← TrahDivad left12:15
→ vdv joined12:16
→ Juo joined12:17
→ Wizek joined12:20
← hubertchang left12:24
→ bartek joined12:24
bartek Hi. What is the purpose of tagging outside of adding relevant snippets to points in your commit log? Can I create a tag and then check out that tag specifically?12:25
cbreak_work yes12:26
or you can use it for diffing12:26
git describe can use tags for version name generation12:26
you can use it as base for rebasing12:26
← _Ryth left12:27
bartek Great, good to know. thanks!12:27
← urkud left12:28
Rhonda Well, you can check out any commit nevertheless.12:29
See a tag as a symbolic name for a commitish, just like a domain name points to an IP address. ;)12:29
bartek Right, an easy way to identify a point in history12:29
→ jsick3 joined12:29
→ shiba_yu_ joined12:30
Rhonda But you can also leave a message with the tag.12:30
Rhonda . o O ( leave a message after the tag. *tag* )12:30
← shiba_yu36 left12:32
→ mklappstuhl joined12:32
→ shiba_yu36 joined12:33
← shiba_yu_ left12:33
→ gusnan joined12:34
ivey_off → ivey12:35
→ jrabbit joined12:35
→ urkud joined12:36
← mklappstuhl left12:38
→ canton7 joined12:39
→ TeckniX joined12:41
→ mklappstuhl joined12:41
→ fs-absynthe joined12:41
← frerich left12:42
→ frerich joined12:42
← frerich left12:42
→ frerich joined12:42
cmyers BAM!12:43
java -jar jgit-cli.jar describe => v0.11.3-241-gd65553e12:43
and it only takes 50x longer to run! =)12:43
← canton7 left12:44
→ envi_ joined12:44
→ whyking_ joined12:45
← whyking left12:45
← bazz left12:45
← urkud left12:45
→ [intra]lanman joined12:46
→ drubies joined12:46
← NetHawk left12:46
→ NetHawk joined12:47
yaotti_aw → yaotti12:47
cbreak_work cmyers: normal git ftw12:47
← mastro left12:48
← envi left12:48
→ Sembiance joined12:48
→ dyer_out joined12:49
Sembiance Hello :) So I branched from master to a new feature branch. I've been working on it. Meanwhile someone else just merged a bunch of changes into master. I have more work to do on my branch, but I'd like to merge master into my branch so I can do the cleanupwork now before doing more work, and then later after I'm done merge my branch back into master.12:49
← Niedar left12:49
Sembiance Is there any problem with this approach?12:49
→ JP-Dery joined12:49
Sembiance It sounds sane to me :)12:49
← silfreed left12:50
→ bushwakko joined12:50
← _Qman left12:50
← Fuu left12:50
→ silfreed joined12:50
cmn Sembiance: that's pretty much the way it's meant to be done12:51
→ Guedes joined12:51
→ Fuu joined12:51
Sembiance cmn: awesome! thanks :)12:51
→ Niedar joined12:51
→ mastro joined12:52
← erichammond left12:52
→ erichammond joined12:52
← workmad3 left12:53
← shiba_yu36 left12:53
← chombee left12:53
← mklappstuhl left12:53
Guedes i've cherry-picked some specifics commits from branchA to branchB, branchC , branchD and branchE randomly. now i want to knows on which branch some commit is... since SHA1 changes i can't (?) use git branch --contains SHA1 ... there is other way?12:53
→ zhengyang joined12:54
→ Vadtec joined12:54
→ CannedCorn joined12:56
→ workmad3 joined12:56
zomg Argh :(12:56
→ cschneid joined12:56
zomg Battling with this damn index pack failed error again12:56
Vadtec I know this is frowned upon, but I'm the only developer using this particular repo, so...25 commits ago, I inadvertantly added a HUGE directory to my repo that isnt necessary. I need to remove it from the repo history, because its making the repo size much to big, and it exists in another repo. Im not familiar with git-rebase, can anyone offer some help?12:56
→ shiba_yu36 joined12:56
zomg How is it possible that fetching fails but cloning works just fine..12:56
Rhonda Sembiance: Some prefer rebase to merge, but that only works out if your feature branch is local only12:56
→ Dr4g joined12:57
Rhonda Vadtec: I think that's more a job for git filter-branch instead12:57
← Dr4g left12:57
→ Dr4g joined12:57
Vadtec Rhonda: thanks, ill look there12:57
Dr4g I got some files that have changes but i want to disgard these changes. - How can i achieve this ?a12:57
I want to switch branch and want a clean slate, i don't want to take the changes with me.12:58
Rhonda Dr4g: See the hint in "git status".12:58
[exa] Vadtec: you need to rewrite history. You can either a] reset branches so that they don't reference the commit and GC the repository afterwards, or b] use git-filter-branch to remove any additions to that directory12:58
I'd go with b].12:58
Dr4g Rhonda: i'm looking at it, what are you hinting12:58
Vadtec [exa]: git-filter-branch sounds like what i need12:58
Rhonda Dr4g: The git checkout HEAD -- path thing.12:59
→ snearch joined12:59
← erichammond left12:59
← jitesh_shetty left12:59
→ erichammond joined12:59
Dr4g excellent Rhonda thanks12:59
Rhonda Hmm, wait, that's for unstaging changes, not throwing them away.13:00
But git checkout file should do the trick nevertheless.13:00
← orafu left13:00
Vadtec makes an entire copy of the repo for backup just incase he royally SNAFUs13:03
→ Vile joined13:03
← shiba_yu36 left13:03
→ bsod1 joined13:03
← erichammond left13:04
bsod1 I added some files with git -A accidentally, now how can I remove them?13:04
→ erichammond joined13:04
→ mklappstuhl joined13:05
← crashanddie left13:05
→ bkozal joined13:06
→ xiangfu joined13:06
FauxFaux bsod1: git status tells you how.13:07
bsod1 FauxFaux, it says "nothing to commit"13:08
← albel727 left13:08
← zhengyang left13:08
FauxFaux Well, you haven't added any files with add -A, then, congrats.13:08
→ zhengyang joined13:08
→ orafu joined13:09
→ Prism joined13:09
← paeva left13:09
→ simplechat joined13:10
← simplechat left13:10
→ simplechat joined13:10
→ albel727 joined13:11
← mklappstuhl left13:11
→ v0n joined13:11
yaotti → yaotti_aw13:12
← xiangfu left13:12
→ jonatasnona joined13:13
→ petrux joined13:13
→ hvgotcodes joined13:13
→ shiba_yu36 joined13:13
→ lolfrenz joined13:14
← shevy left13:14
← dwmw2__ left13:14
→ keystr0k joined13:14
← CannedCorn left13:15
Vadtec Rhonda: hrm...i did git-filter-branch, but I still see the directory listed in the branch history13:15
← sunnyohno left13:17
→ crashanddie joined13:17
← crashanddie left13:17
→ crashanddie joined13:17
→ dv310p3r joined13:17
Rhonda Maybe you fumbled with the syntax?13:18
→ sunnyohno joined13:18
→ bibinou joined13:18
← TLMonitor left13:19
Vadtec Rhonda: git filter-branch --index-filter 'git rm -r --cached --ignore-unmatch bin-img' bad-img13:19
got the syntax from the man page13:19
hmmm....13:21
→ leitao joined13:21
← jonatasnona left13:21
Vadtec think i know what it was13:21
→ emf^ joined13:21
← glcrazy left13:22
→ Ryback_ joined13:22
← Ryback_ left13:22
→ Ryback_ joined13:22
Vadtec instead of giving the tag i created for the bad commit, i needed to give HEAD for the starting point13:23
now its removed13:23
:D13:23
→ xiangfu joined13:23
Rhonda :)13:23
You need to git gc it to make it really disappear from your objects.13:23
← mastro left13:23
Vadtec yeah13:23
just read that13:23
Rhonda Potential with --prune=now or something like that.13:24
Vadtec yeah, man pages ftw :D13:24
rob_w → rob_w_away13:24
← erichammond left13:25
→ erichammond joined13:25
→ shevy joined13:26
← sigh left13:26
→ seivan joined13:27
→ AkumaStreak joined13:27
← workmad3 left13:27
← harshu_ left13:27
← bkozal left13:30
→ __sorin__ joined13:30
→ lb joined13:30
← LuckyNeo left13:31
Vadtec hmmm, now i just have to figure out how to git rid of this one tag that was branched off during this filter-branch and ill be good to go13:31
→ jonpacker joined13:32
← madewokherd left13:33
→ deekayen joined13:34
← deekayen left13:34
→ deekayen joined13:34
← seivan left13:34
→ bryanl joined13:34
Vadtec or...better idea...restore the backup, and do things one at a time13:34
← tydeas left13:34
→ jfkw joined13:35
← albel727 left13:36
→ jelmer joined13:36
← jelmer left13:36
→ jelmer joined13:36
← erichammond left13:37
→ erichammond joined13:37
← konr left13:38
→ jonatasnona joined13:38
← sonnym left13:39
→ albel727 joined13:39
← frerich left13:39
→ ymasory_ joined13:40
← ymasory left13:41
→ m_3 joined13:41
ymasory_ → ymasory13:41
← shiba_yu36 left13:42
→ dajhorn joined13:42
→ shiba_yu36 joined13:42
→ Freakshow joined13:42
→ TheWarden joined13:42
← sunnyohno left13:43
TheWarden Hi, I'm a newbie to using Git. I installed gitcore and gitweb packages for debian. Now I'm confused how to create a repository that can be accessible remotely in NetBeans v7. In NetBeans I have Git plugin installed.13:44
Can anyone direct me?13:44
→ glcrazy joined13:44
nevyn TheWarden: have you used any DVCS before?13:46
→ frerich joined13:47
← frerich left13:47
→ frerich joined13:47
← giallu left13:47
← albel727 left13:47
SethRobertson TheWarden: The GitPro book is very good !book13:48
jast TheWarden: [automatic message] 'Pro Git' is a complete book about git that is available online for free: http://progit.org/book/13:48
← ixti left13:48
→ tstclair joined13:48
SethRobertson TheWarden: You need to pick a server to store the git location. It can be on an ssh accessible system, or perhaps on GitHub or another git hosting provider13:49
s/git location/git repository/13:49
← jonpacker left13:49
→ seivan joined13:49
→ mrball joined13:50
→ jonpacker joined13:50
TheWarden SethRobertson: oh cool, I'll check out that book. Thanks. I have Git installed already along with gitweb (web interface just for those non-git users to browse).13:50
← dr0id left13:51
← soc42 left13:51
→ soc42 joined13:51
← Niedar left13:51
← jonpacker left13:51
→ jonpacker joined13:52
→ albel727 joined13:52
← ExtraSpice left13:52
← jonpacker left13:52
← mikl left13:54
← EricInBNE left13:54
→ madsy joined13:54
← madsy left13:54
→ madsy joined13:54
← v0n left13:54
→ TomyLobo joined13:55
TomyLobo hi13:55
FauxFaux I've seen you in enough channels enough times that, by now, you must know how rude that is, TomyLobo.13:56
TomyLobo what, saying hi?13:56
← petrux left13:56
yaotti_aw → yaotti13:56
← erichammond left13:57
TomyLobo i only say that when i join a channel and want to ask a question right as the first thing :)13:57
anyway. it's kind of a silly question, with no real implications or anything:13:58
→ sheikra joined13:58
TomyLobo what happens if there is ever a revision of something with the hash 000000...? would that confuse only the users or also git itself?13:58
← noonoooo left13:58
→ canton7 joined13:58
sheikra hello. Is there a way to push an existing git repository to another server?13:58
TomyLobo since i see git using 000000 to denote a non-existant file13:58
sheikra sure, just add the other server as a remote, then push it there13:59
git remote --help && git push --help13:59
→ jjido joined13:59
sheikra TomyLobo: It says "No refs in common and none specified; doing nothing."13:59
TomyLobo sheikra what says that?13:59
sheikra TomyLobo: git push ...14:00
TomyLobo git push what?14:00
just git push?14:00
sheikra TomyLobo: oh ,,, got it14:00
TomyLobo: I didn't put "master" in the end14:00
TomyLobo i think it's git push local_branch remote_repository_name:remote_branch14:01
cmn no14:01
lb TomyLobo: just the other way round14:01
→ tearsong joined14:01
cmn gt push repository local-branch:remote-branch14:01
lb git push remote_repo local_branch:remote_branch14:01
← savoxis left14:01
lb damn cmn :)14:02
sheikra got it .. I didn't put the branch name14:02
jjido I am running git 1.7.5 on Mac. I get Authorization failed when I run git svn dcommit. the error is at line 4459, which is a call to $opts->{ra}->get_commit_editor. Any suggestion?14:02
cmn lb: I made a typo, so I didn't really win :)14:02
lb cmn: strike!14:02
TomyLobo yeah i didnt do it in a while14:02
← soc42 left14:02
cmn sheikra: if you don't specify any branches, git will push any common ones14:02
TomyLobo that's why i pointed sheikra to the manual :)14:03
nevyn TheWarden:14:03
TheWarden nevyn: yes14:03
→ giallu joined14:03
Vadtec now for the fun part...git push --force so I can update my remote storage repo, weeee!14:04
nevyn so netbeans has made a git repo and converted it's own project directory into a git repo as I understand that page.14:04
→ noonoooo joined14:04
nevyn TheWarden: what platform are you running netbeans on?14:04
Vadtec makes one more backup....just in case14:04
TheWarden nevyn: mmm I think I found all the options now. I right clicked on my project and clicked connect then I got a whole bunch of icons beside my files/directories... mm14:04
nevyn: Windows 7 for now but I do have NetBeans on Debian too.14:05
→ n8o-mba joined14:05
g0bl1n I've done "git add ." but there are files I don't want to commit. How can I do this ?14:06
← joel_falcou left14:06
→ dwmw2_gone_ joined14:06
selckin read git status output14:06
g0bl1n selckin, yes, they are there14:07
selckin no read it.14:07
← dajhorn left14:07
g0bl1n git clean ?14:07
fairuz git checkout14:08
→ md-2 joined14:08
fairuz git checkout <file>14:08
g0bl1n or git -rm ?14:08
→ mastroDani joined14:08
sheikra TomyLobo: that problem solved, but another one prevented me from push14:08
By default, updating the current branch in a non-bare repository14:08
FauxFaux g0bl1n: git status tells you.14:08
sheikra is denied, because it will make the index and work tree inconsist ...14:08
→ jitesh_shetty joined14:08
FauxFaux g0bl1n: git status tells you.14:08
PerlJam g0bl1n: "unstage" means "don't commit"14:09
→ dajhorn joined14:09
g0bl1n ok, thanks14:09
← jfkw left14:10
g0bl1n git reset HEAD <file> thanks14:10
→ jophish joined14:10
← lectrick left14:12
← Error404NotFound left14:12
← jophish___ left14:12
← aangel left14:13
← roop left14:14
FauxFaux TomyLobo: I didn't see anyone answer your question; if a commit has a hash of 40 0s then it'd probably confuse git, but the chances are about 1-in-2^128, which is pretty low. ;) If you mean 6 or 8 0s, try it: http://blog.prelode.com/2010/12/git-set-commit-id/14:14
q0rban|afk → q0rban14:14
← sproaty left14:15
Vadtec just an FYI to all: if you do a git filter-branch and have tags that are going to get filtered, you will end up with a branch containing those tags, you might want to remove the tags, do your filter-branch, then create the tags again14:15
→ tanuki_ joined14:15
→ babilen joined14:16
TomyLobo FauxFaux i think that only creates alias ids14:16
engla FauxFaux: :-)14:17
TomyLobo: nope14:17
TomyLobo how does it work? extra files?14:17
sheikra Hmm... why can't I see the files after a successful push?14:17
engla TomyLobo: it's "rerolling" the commit until the desired hash prefix is found. How, I don't know14:18
FauxFaux It appends whitespace to the commit message.14:18
← n8o-mba left14:18
→ joshteam joined14:18
→ Dynetrekk joined14:18
TomyLobo lol14:18
→ mdel joined14:19
← bushwakko left14:19
← whyking_ left14:19
Dynetrekk hi, I'm relatively new at git and I would like to try github. However, I can't find the 'create repository' thingy. where is it? (I just have a "free" account)14:20
engla Dynetrekk: "New Repository" on the right-hand side of github.com when logged in?14:21
fairuz Dynetrekk: https://github.com/repositories/new14:21
← avinashhm left14:21
sheikra Any ideas whether I can delete my repositories from github?14:21
→ tty1 joined14:22
Dynetrekk heh, funny, I've been looking and I can't find it...14:22
TomyLobo FauxFaux oh, you made that :D14:22
fairuz sheikra: of course you can14:22
Dynetrekk ah, I had a lot of "you are new here" junk hiding it :P14:22
sheikra fairuz: I tried to find it, but ..14:23
engla sheikra: 'Admin' button on your repo's home view14:23
← albel727 left14:23
fairuz sheikra: admin, then choose Delete Repo14:23
sheikra: Dynetrekk: http://help.github.com/ make this your friend =)14:24
→ Error404NotFound joined14:24
← Vadtec left14:25
← jjido left14:25
→ blorbx joined14:26
sheikra I found it . thanks14:26
→ jjido joined14:26
→ tydeas joined14:26
tydeas how can i add a dir to the repository but gitignore it's content ?14:27
→ ShooterMG joined14:27
avar add a .gitignore with * in it to the dir14:27
→ kukks joined14:28
avar but why the hell would you want to do that?14:28
fairuz :D14:28
→ etcetera joined14:28
→ PabloM joined14:28
yaotti → yaotti_aw14:28
canton7 tydeas, you can't. git tracks files, not folders. if you're not tracking any files in the folder, git won't track the folder14:28
yaotti_aw → yaotti14:29
tydeas canton7: so i can not have an empty folder ?14:29
canton7 tydeas, can't track an empty folder, no14:29
tydeas it's a spool folder14:29
see14:29
cmn tydeas: you can if you /really/ want to, put a .gitignore file in it and track it14:29
canton7 cmn, point. you see a lot of .empty files in otherwise-empty folders, too14:29
tydeas cmn: what will this gitignore content be?14:29
cmn it doesn't matter14:30
canton7 tydeas, can leave it empty if you want14:30
cmn or put *14:30
→ sonnym joined14:30
_johnny hi, not *really* a git question, but i hope one of you might be able to help me. i've edited motd for my (ssh) box, and it looks like it's suppose to when i ssh into it, however, when i use my git user (which i've set to git-shell - hence the relevance for this channel, heh) i get the motd *plus* "The programs included with ... NO WARRANTY" rubbish also, and then finally the only message i'm interested in "what do you think i am? a shell". anyone know14:30
→ Corren joined14:30
← opalka left14:31
_johnny as in, do i need to mod the git-shell binary?14:31
← lb left14:32
cmn _johnny: git-shell isn't meant to be used as an interactive shell, I think14:32
what do you want to accomplish?14:32
_johnny oh, right. i appologize. i know. i just connect through ssh to it to make sure the git user can't do anything14:33
what i want is for it just to reply with what do you think i am bla bla, and then quit14:33
← kukks left14:34
_johnny not tell the git user about some no warrranty OS stuff14:34
if i ssh to a user which has bash it's not displayed, so i'm just wondering if git-shell is the one retreiving this info14:34
→ rayners joined14:35
→ mastroWork joined14:36
cmn the no warranty thing may be in /etc/issue.net14:36
but I think either ssh or login print it out14:36
tydeas cmn: .empty means something to git or it's just a self-explained name ?14:36
→ mikl joined14:36
← mikl left14:37
→ mikl joined14:37
cmn tydeas: it's just some name14:37
← psoo left14:37
cmn git doesn't care about filenames much14:37
← Corren left14:37
← angelsl left14:37
kylehayes ugh, got over an argument about git guis with a coworker and why its more efficient to learn the command line14:37
_johnny cmn: ok, thanks :)14:38
kylehayes He said we won't be able to get it adopted here if we don't provide a decent gui14:38
rndbit → zz_rndbit14:39
→ Polysics joined14:39
Polysics hello14:39
sorry if this is probably a stupid question14:39
fr0sty-away "decent" == ???14:39
← mastroDani left14:39
Polysics but how do i clone a specific branch off github, please?14:39
fr0sty-away Polysics: you clone repositories, not branches.14:39
zz_rndbit → rndbit14:39
engla Polysics: you can use git clone -b branchname URL14:39
Polysics or how do i swtich t oa branch after cloning, that is14:40
engla but yeah, you still get the whole repository14:40
Polysics: use -b or -n options to clone. See git help clone for more! :-)14:40
fr0sty-away you can 'git fetch URL branchname' though...14:40
← tydeas left14:40
← gusnan left14:41
← g0bl1n left14:41
kylehayes fr0sty-away: I couldn't get out of him what decent meant. I asked why a GUI would be better and he said that's just the way we think....14:41
← mikl left14:41
← radarek left14:41
→ brienna joined14:42
→ ceej joined14:42
← brienna left14:43
fr0sty-away well, then any gui is 'decent'...14:43
→ tystr joined14:44
kylehayes I see the advantage in diffing tools, but I don't see it in normal day to day operations of add, committing and branching14:44
→ excess^ joined14:45
→ minijupe joined14:45
rane_ I have one directory in the root of a repo and want to make that directory root of the repo, how do I move it?14:45
git mv foo/* . doesn't seem to work14:45
minijupe how do I get a diff between my current working dir and the last commit?14:46
yaotti → yaotti_aw14:46
teuf minijupe: git diff HEAD14:46
minijupe teuf: tnx14:46
engla kylehayes: I think he is right. problem is that git's creators also created git after how they think, and that's different14:46
→ mando joined14:47
← Polysics left14:47
← dannas_ left14:47
← minijupe left14:48
kylehayes engla: I don't see how a gui makes it more efficient. I feel if one learns the basic commands, they will be able to use them a lot quicker than running the mouse over a interface14:48
→ Textmode joined14:49
kylehayes then again, I can touch type, he can't so perhaps he doesn't like typing so slow14:49
→ aangel joined14:49
engla well if he said "that's just how I think" then it wasn't a question of what's possibly most efficient14:49
← MacGyverNL left14:50
rane_ fatal: source directory is empty, source=devhub/log, destination=log14:51
;[14:51
→ MacGyverNL joined14:51
thm when I rebase a branch, how can I also on-the-fly rebase refs (tags or branches) that point to commits on the old branch?14:51
→ jfkw joined14:51
sitaram kylehayes: a gui makes it more efficient for someone who looks at git as a chore and not central to his work. He feels no need to "learn" it and just wants to get the job done. Ergonomics of mouse versus keyboard don't (usually) enter into it14:51
→ airborn joined14:52
sitaram kylehayes: I know/have heard of/deal with several people like that at work14:52
→ Vadtec joined14:52
Vadtec i must be dense...how do I do a clone of a tag?14:52
← jonatasnona left14:53
sitaram you can *fetch* a tag14:53
→ deuterium joined14:53
Vadtec so i have to clone the entire repo first? i cant just clone everything up to the tag point?14:54
← deuterium left14:54
Vadtec or can i use the tag as my REV?14:54
engla Vadtec: it's normally not worth it to bother about the difference. clone, then check out the revision you are interested in14:54
→ deuterium joined14:54
Vadtec ok, so how do i use the tag to specify the rev im after?14:55
Vadtec reads some docs14:56
→ bkosborne joined14:56
→ alberto56 joined14:56
engla Vadtec: just use the name of the tag14:56
→ mklappstuhl joined14:56
engla Vadtec: git checkout v2.2 #for example14:56
← gnufied_ left14:56
← gnufied left14:56
→ albel727 joined14:57
Vadtec engla: thanks, i just found that in the docs *facepalm* i over looked it the first time14:57
shamess I try to push to my origin, but it says "rejected, non-fast forward". To fix that I usually git-fetch then check it out and then pull and then I can push. But the diff for the origin/master looks like it will get rid of all my changes on my current repo... Almost like the origin thinks it's ahead of my repo.14:57
Could anyone explain what's going on?14:57
engla Vadtec: you will probably want to create a branch there to have something to work on14:57
Vadtec engla: not so much, this is just a tag for a release point14:58
gonna use it to package up the source files into a tarball14:58
canton7 thm, no direct way to do it. see http://blog.nilsjonsson.com/post/4421450571/rebasing-tags-in-git-repositories for a solution14:58
engla Vadtec: you can possibly use git archive <revision> without even needing to check it out at all (or have a worktree)14:59
→ workmad3 joined14:59
engla Vadtec: except of course if you need more than just the repository files14:59
← ajt left14:59
Vadtec engla: in this case, it would be not adding files im after, there are some files that are only used by myself that dont need to be packages14:59
packaged*14:59
q0rban → q0rban|afk14:59
→ ajt joined15:00
→ ISF joined15:00
← lolfrenz left15:00
← KKA left15:00
→ kukks joined15:00
Vadtec i was thinking there was a way to clone the tag, but if i want to do that, i might as well create a branch and then merge it back into master15:00
→ KKA joined15:01
← kpreid left15:02
← lahwran left15:02
← sheikra left15:02
← engla left15:03
→ beatak joined15:03
→ lahwran joined15:04
→ kold joined15:04
→ montylounge joined15:04
kold Hi all, is there a way to have git-diff ignore certain file types? I have minified javascript files that I'd like to ignore. I can't seem to determine if there are any options I can pass to git-diff to make this happen15:05
← Vadtec left15:05
← eijk left15:06
→ joel_falcou joined15:06
thm canton7: if one notes the old hashes, it would be easy or rather mechanical task to do that, not complicated, but error-prone15:07
→ g0bl1n joined15:07
← deuterium left15:08
← airborn left15:08
← Spockz left15:08
→ deuterium joined15:09
→ hyperair joined15:09
_johnny cmn: decided to do a little hex editor magic to take care of it. so i also changed the reference of "who do you think i am" to a region of free space and typed in something more "appropriate": "fatal: Hi! You've successfully authenticated, but you're denied shell access." :)15:09
→ hubertchang joined15:10
→ savoxis joined15:10
→ eijk joined15:11
← xshyamx left15:11
_johnny it may have taken me a couple of hours, but setting up a git "server" without git server daemons like gitorious etc. *is* possible :P15:11
etcetera what's the best way to fetch and merge from a remote repo?15:11
cmn _johnny: you can do that in the source, which should be much safer15:11
_johnny true, but with my chrooted setup, i'd prefer not to compile it (again) hehe15:11
← vmiklos left15:12
← nevyn left15:12
cmn and at any rate, you can use gitolite if you want users not to be able to do anything else on the server, they don't even need a login, I think, just the gitolite user15:12
← jjido left15:13
_johnny cmn: i'll try that :)15:13
← kold left15:13
_johnny i have too much time at hand.. lol15:13
→ vmiklos joined15:13
→ allaire joined15:13
← alberto56 left15:14
→ mm_202 joined15:14
← namemon left15:15
→ mikl joined15:15
→ RaceCondition joined15:15
→ smolyn joined15:16
← __sorin__ left15:17
allaire Hey guys, I'm do you know if svn merge files line-by-line like git does? I'm doing an oral presentation about why using git15:18
→ tjholowaychuk joined15:18
allaire I remember I had much more trouble doing merge with svn back in the days15:18
→ alex__c2022 joined15:19
→ alberto56 joined15:19
← xiangfu left15:19
→ nevyn joined15:19
→ kpreid joined15:19
→ Niedar joined15:19
← cbreak_work left15:20
← werdan7 left15:20
→ roop_ joined15:21
→ BiggFREE joined15:21
→ sunnyohno joined15:22
← leitao left15:23
savoxis → SavoxUnity15:23
← jophish left15:24
← metze left15:25
← mklappstuhl left15:25
← jutaro left15:25
← ShooterMG left15:26
← rchavik left15:26
→ petrux joined15:26
← Error404NotFound left15:27
← sunnyohno left15:27
→ n8o-mba joined15:28
→ flaguy48 joined15:28
← shiba_yu36 left15:30
→ sou_ joined15:30
← envi_ left15:30
→ envi_ joined15:30
→ masolino joined15:30
← sou_ left15:31
← giallu left15:31
→ metze joined15:32
← aangel left15:32
← kermit left15:33
← orafu left15:34
← freimatz left15:35
→ sechrist joined15:35
→ leitao joined15:35
← hubertchang left15:35
→ kipras_ joined15:35
→ hubertchang joined15:36
→ SimonP86 joined15:37
→ jutaro joined15:37
→ Error404NotFound joined15:40
→ mklappstuhl joined15:40
← Sembiance left15:40
← dpy left15:40
← sebbi left15:41
→ shiba_yu36 joined15:41
→ booginga joined15:42
→ beiter joined15:42
booginga hey all is qucik question is there something like git hub that allows you to visualize and trak changes in a self hosted repo that is open source and can be deployed on a private server?15:43
← deekayen left15:43
teuf gitorious ? or maybe cgit/gitweb15:43
← nigelb left15:43
→ devth joined15:44
→ thiago_home joined15:44
→ deekayen joined15:44
cmn if you want something social-like in the github style, gitorious15:45
→ crakdmirror joined15:45
→ deviantintegral joined15:46
deviantintegral what would be the best command to get a parseable version of "git branch --merged"?15:46
← infid left15:47
← BiggFREE left15:47
← phaer left15:48
→ BiggFREE joined15:48
→ alvaro_o joined15:48
booginga thanks im going to look at them now15:48
→ infid joined15:49
← smolyn left15:49
← envi_ left15:49
→ envi_ joined15:49
← D5 left15:50
← alvaro_o left15:50
→ D5 joined15:50
← pizza_ left15:50
→ conan joined15:50
← mklappstuhl left15:51
← crakdmirror left15:51
→ cal__ joined15:52
← ddilinger left15:52
cmn deviantintegral: what do you mean with 'parseable'? parseable with what?15:52
→ mklappstuhl joined15:53
deviantintegral cmn: I want a script that will delete all branches that have been merged into the current branch15:53
→ pizza_ joined15:53
← albel727 left15:54
← D5 left15:54
cmn the output from git branch --merged should be quite usable for that15:54
just separate on \n15:54
← kipras_ left15:54
→ kipras joined15:55
→ ddilinger joined15:55
← n1ck31 left15:55
deviantintegral cmn: yeah, right now I have http://pastebin.com/w70nDNiM, but I don't like having to grep out the current branch15:56
→ radarek joined15:56
ChanServ set mode: +v15:56
← ISF left15:56
→ albel727 joined15:56
→ quiliro joined15:57
→ ISF joined15:57
← Badiss left15:57
← mastroWork left15:57
cal__ if i have a massive repo, say "project", that contains "project/lib" and now want to make three repos (project, project2, lib) and use submodules, is it possible to preserve the history for "lib" when I make a new repo out of it?15:58
← dajhorn left15:58
← roop_ left15:58
→ mastroDani joined15:58
← mastroDani left15:58
fr0sty-away deviantintegral: git symbolic-ref HEAD15:58
→ sunnyohno joined15:59
cmn cal__: man git-filter-branch15:59
jast cal__: [automatic message] the 'git-filter-branch' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-filter-branch.html15:59
deviantintegral thanks fr0sty-away I'll look into that15:59
booginga teuf: thanks c git is close to what im looking for15:59
← mklappstuhl left16:00
booginga sorry cgit16:00
← Titosemi left16:00
cal__ cmn: jast: thanks will investigate16:00
cmn np16:00
→ mklappstuhl joined16:00
→ israel_dahan joined16:01
→ CannedCorn joined16:01
→ psoo joined16:01
← davzie left16:01
← sunnyohno left16:01
← lahwran left16:01
→ bkosborne_ joined16:02
→ dajhorn joined16:02
← rndbit left16:03
→ LiamH joined16:04
← q0rban|afk left16:04
→ zz_rndbit joined16:04
← zz_rndbit left16:04
← BiggFREE left16:04
← bkosborne left16:04
bkosborne_ → bkosborne16:04
← hubertchang left16:04
→ lahwran joined16:05
→ necromancer joined16:05
→ amerine joined16:05
← mklappstuhl left16:06
→ rndbit joined16:06
← Pilif12p left16:07
fr0sty-away !jast16:08
jast [automatic message] jast runs a script that provides all kinds of information, but rumour has it that he's also human. Sometimes.16:08
← rndbit left16:08
→ envi__ joined16:09
→ rndbit joined16:09
← sechrist left16:10
→ HG` joined16:10
← envi_ left16:12
→ t0rc joined16:12
necromancer om16:13
omg16:13
git-svn is a godsend16:13
however, has anyone here had trouble with nested repos?16:13
cmyers heh, it's something alright =)16:13
→ mikepack joined16:14
cmyers necromancer: git-svn is easily confused because SVN doesn't have real branches or structure or *anything*16:14
→ grahamc joined16:14
necromancer cmyers: yeah but the way you're supposed to use svn is do work->make sure the build functions->commit16:14
cmyers if you tell me more about what your repo layout is, I may be able to suggest a fix.16:14
← conan left16:15
necromancer well right now i clone some repos into a dir called lib/16:15
i don't REALLY need the code for the app to be functional, but i do need it for the tests to run since that's where i keep my bdd framework16:15
when git attempts to commit to the svn repo, it runs into an error. i can't remember what the error was and i don't want to fuck up the repo right now so i can't exactly make it fail16:15
the only way i could get around it was recreating the whole thing16:16
← seivan left16:16
necromancer because even though lib/ was destroyed, git kept trying to commit it to svn16:16
→ spearce joined16:16
cmyers necromancer: yeah, git doesn't track dirs and svn does16:16
so that creats some parity problems sometmies also16:16
→ mklappstuhl joined16:16
cmyers the main thing to understand about git-svn16:16
→ engla joined16:16
cmyers is that it uses the metdata in the comments to figure out where to commit to16:17
← abourg left16:17
cmyers so, if you have trunk and stable16:17
← envi__ left16:17
→ v0n joined16:17
cmyers and you cherry-pick from trunk to stable, and don't change the commit message, then dcommit16:17
→ envi__ joined16:17
necromancer oh yeah we're not that complicated here16:17
cmyers it will try to commit it to trunk - because the commit message has "git-svn://..." pointing to trunk16:17
necromancer ah i see16:17
cmyers are you doing any branches, or just flat checkouts of a single line of history?16:17
← t0rc left16:18
→ chombee joined16:18
necromancer not really. the only time we really work in branches is if one of us needs to work on a project and doesn't want to mess with the actual repo since it's in production use. or if one of us is adding a new feature16:18
it's only 3 people committing to the repo so we don't get that complicated in branching and shit16:18
otherwise i'm pretty sure we'd be using git by now ;)16:18
cmyers yeah, git-svn is very good at just checking out trunk or whatever16:19
necromancer because my boss would be so pissed off at svn's inability to handle branching16:19
cmyers when you have lots of branches flying, that is when git-svn tends to get confused16:19
→ enherit joined16:19
necromancer well one of the cool things is it *does* allow you to branch, locally.16:19
just like git does16:19
and i use local branches that i never commit to remote all the time...for experimenting with new techniques/features/etc.16:19
that way if i fuck up i can just delete the branch and start fresh with a working build16:19
cmyers yup. the other thing that can confused git-svn is if you have an svn-copy from one path to another that doesn't make sense to it.16:19
if it sees a copy it tries to follow it backwards in history16:19
necromancer a copy?16:20
→ konr joined16:20
cmyers so if you ever "reorganized the repo" or anything, it gets confused16:20
necromancer oh shit yeah16:20
cmyers in svn, "cp" == "new branch" remember16:20
necromancer well, svn gets confused too16:20
cmyers it sure does16:20
necromancer you really shouldn't EVER use shell commands to manipulate shit in svn. because if you make a mistake and svn can't commit, you're fucked16:20
gotta start over16:20
ever try to rm -rf a file without telling svn? it doesn't like that very much :)16:21
cmyers < necromancer> you really shouldn't EVER use svn. because if you make a mistake and svn can't commit, you're fucked16:21
fixed16:21
necromancer haha16:21
→ d0k joined16:21
necromancer nah it's not THAT bad. i don't really think it's that different from git, conceptually. but git works differently behind the scenes, much like a filesystem (how surprising, coming from mister linux!), so it handles non-standard situations with ease16:22
so that's why i like it16:22
cmyers the more you understand git, the more you will understand just how different they are16:22
and how *great* that difference is.16:22
reliability not withstanding16:22
← thiago_home left16:23
← petrux left16:24
→ panike joined16:25
→ mysphyt joined16:26
→ subsonic joined16:26
fr0sty-away necromancer: they _are_ different conceptually. distributed vs. centralized being the main difference..16:26
necromancer i guess that's true16:26
← fs-absynthe left16:27
necromancer i meant at their most basic level, however, they're still both scms and they follow the same paradigms: commit changes, pull/update from the server, etc.16:28
→ Plumber joined16:28
necromancer except git is way better since it's not centralized :)16:28
← achadwick left16:28
necromancer so i can do whatever i want to the repo over here and it won't affect anyone else in the office16:28
→ Titosemi joined16:28
ChanServ set mode: +v16:28
→ gigamonkey joined16:29
Plumber Hi!16:29
→ Corren joined16:29
mysphyt Heyo. I'm having an odd trouble with a pull. If I set the remote up via ssh, everything's fine. If I set the remote up via FTP (better for me for network reasons), I get a 550 error trying to get repo/info/refs . But repo/info/refs doesn't exist. repo/refs does. Any guesses as to what's going on?16:29
Plumber Why are my untracked files copies when I checkout a new branch ?16:29
copied*16:29
cmyers I would agrue their paradigms are drastically different. Monotonic increasing repo-wide revs versus an acyclic directed graph of objects which contain metadata. Also, svn properties versus the git metadata which is cryptographically consistent16:29
Plumber anybody can help me ?16:30
cmyers Plumber: they aren't copied, they are "still there"16:31
git never touches untracked files unless you run a `git clean` with certain arguments16:31
Plumber Ok right16:31
I will try this :D16:31
→ dreiss joined16:31
← vsayer left16:31
Plumber Although, they 're untracked because of my gitignore so they will be there a day or another16:32
→ justin-george joined16:32
→ lossytossymaster joined16:32
← PigeonFriend left16:32
→ Anubisss joined16:32
← dharrigan left16:33
→ smolyn joined16:33
→ urkud joined16:33
cmyers Plumber: if you want to remove *everything*, including untracked files *and* ignored files, `git clean -d -f -x`, -f means "yes, really delete stuff", -d means "recurse", and -x means "also delete ignored files"16:33
← jutaro left16:33
mysphyt Anybody? What files should a pull be looking at?16:33
Plumber Thanks cmyers :D16:33
See you16:34
cmyers mysphyt: can you paste your [remote "foo"] section into a pastebin for both remotes?16:34
← md-2 left16:34
→ theoretical joined16:35
→ dAnjou joined16:35
→ trobrock joined16:36
Guedes Hi all... i've cherry-picked some specifics commits from branchA to branchB, branchC , branchD and branchE randomly. now i want to knows on which branch some commits are... since SHA1 changes i can't (?) use git branch --contains SHA1 ... there is other way?16:36
Plumber cmyers : Ok, it works, but if I add the same untracked file (because of a compilation for instance ), and checkout again, they reapear16:36
not add but recreate **16:36
← lossytossymaster left16:36
→ phantomcircuit joined16:37
→ envi_ joined16:37
← dajhorn left16:37
← NoNoNo left16:38
mysphyt cmyers: Yeah, two shakes.16:38
→ g105b joined16:38
cmyers Guedes: my understanding is not really - I would write a script to compare commit messages and try to heuristically figure out which commits line up16:38
→ phishy joined16:38
g105b hi gits! I've got a large filestructure and want to add all files to my repo (I think that is `git add *`) but I want to ignore certain files (only a few). How can I do this?16:39
mysphyt cmyers: https://gist.github.com/957386 . Somewhat vague, I recognize, but I hope you don't mind my obscuring my credentials.16:40
← envi__ left16:40
devth i'm going to answer "acyclic directed graph" every time a nontechnical person asks me a question16:40
mysphyt cmyers: (It's a personal project, and so not a public repo.)16:40
Guedes cmyers, i'm doing something like this, in time the filename is unique so i can "infer" from filename via git log and rev-list.. but i want to be sure about the commit by itself...16:40
→ nadirvardar joined16:40
← mklappstuhl left16:40
Guedes cmyers, thank you... i'm seeing that i'm not so wrong at moment... i'll figure other better ways to do that.16:41
cmyers Guedes: alternatively, you can change your process for cherry-picking so that they say "cherry-picked from original SHA XXXXXX..."16:41
then your script can look for that.16:41
← tvw left16:41
cmyers mysphyt: wow, those look pretty similar16:42
I've never used FTP before but I am suprised that isn't working16:42
perhaps you have found a bug =/16:42
→ aroop joined16:42
cal__ → calimocho16:43
mysphyt cmyers: Right? Everything is the same, the ftp account has permission for the file, and using a regular FTP client I can fetch the files fine. It just seems to be looking in the wrong place.16:43
fr0sty-away Guedes: look at man git-cherry...16:43
jast Guedes: [automatic message] the 'git-cherry' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-cherry.html16:43
Guedes cmyers, hum... can i append a message to the commit cherry-picked?16:44
→ rbriggsatuiowa joined16:44
fr0sty-away mysphyt: according to man git-clone ftp:// urls don't accept user/pass16:44
jast mysphyt: [automatic message] the 'git-clone' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-clone.html16:44
rbriggsatuiowa is there a short command to rebase a branch since it started diverging?16:44
fr0sty-away maybe mess with your .netrc...16:44
cmyers Guedes: you can manupulate the commit message however you like16:45
see `git help cherry-pick`16:45
mysphyt fr0sty-away: AH! That'll do it. Well, that's craptacular.16:45
dAnjou hey, i didn't get what git graphs are showing. is it explained somewhere?16:45
fr0sty-away rbriggsatuiowa: where do you want the commits to end up? based on what commit?16:45
dAnjou *don't16:45
fr0sty-away "graphs"?16:45
← trobrock left16:46
cmyers he probably means gitk or decorated git log16:46
→ Zephyrus joined16:46
rbriggsatuiowa I branched off of master to topic - made a bunch of changes and want to alter the history since I branched from master before I merge topic back to master16:46
mysphyt fr0sty-away: Any way around that, that you know of?16:46
Guedes cmyers, i'm seeing it... thanks.16:46
→ trobrock joined16:46
cmyers dAnjou: maybe http://marklodato.github.com/visual-git-guide/index-en.html ?16:46
adamholt → adamholt_away16:46
fr0sty-away rbriggsatuiowa: git checkout topic, git rebase master will place all of your topic commits on top of master16:46
adamholt_away → adamholt16:46
fr0sty-away besides, .netrc? no...16:47
cmyers mysphyt: what are your requirements? you could have a pw in .git/config but not in .netrc?16:47
→ cylence joined16:48
→ towski joined16:48
mysphyt cmyers, fr0sty-away: I'm not familiar with .netrc--it sounds like that's the obvious solution I'm missing out on.16:48
Let me do some googling.16:48
thm or man netrc16:48
dAnjou cmyers: no, i don't think this helps16:48
cmyers mysphyt: yeah, that should explain it. if you still have trouble let us know.16:48
fr0sty-away just 'man ftp' and look at the "THE .netrc FILE" section...16:48
mysphyt I'll do that. Thanks a lot.16:49
cmyers dAnjou: can you ask a specific question? I'm not sure what you need help understanding.16:49
→ conan joined16:49
→ icwiener joined16:49
← SavoxUnity left16:49
← g105b left16:49
→ savoxis joined16:49
rbriggsatuiowa fr0sty-away: that was exactly what I wanted - thank16:51
← masolino left16:51
← bkosborne left16:51
dAnjou cmyers: i'll make a screenshot of my graph and you will understand my confusion (maybe because i did some stupid things)16:51
cmyers dAnjou: cool16:52
← rbriggsatuiowa left16:53
← vdv left16:53
← envi_ left16:53
Plumber Do you know if I can trigger a git clean -D -f -X when I checkout ?16:53
→ envi_ joined16:53
→ joe-mac joined16:53
booginga y16:53
mysphyt cmyers, fr0sty-away: Same problem. .netrc configured correctly, ftp server connects me to the correct user account, git pull still looks for repo/info/refs instead of repo/refs .16:53
booginga sorry to many windows16:53
joe-mac what is the syntax to git diff a remote branch via url if it isn't set up as a git remote?16:54
← envi_ left16:54
joe-mac tells me ambiguous argument when I do "git diff URL master"16:54
mysphyt I get the same message: "error: RETR response: 550 while accessing ftp://server/path/to/repo/info/refs"16:54
→ slonopotamus joined16:54
→ cakehero joined16:55
→ phaer joined16:56
← phaer left16:56
fr0sty-away mysphyt: http://stackoverflow.com/questions/2930371/set-a-ftp-repository-with-git16:56
→ phaer joined16:56
← joe-mac left16:56
fr0sty-away update-server-info is needed, it seems...16:56
dAnjou cmyers: don't laugh at me :P http://img6.imagebanana.com/img/pv2vr1u6/screenshot_007.png16:57
fr0sty-away if joe-mac was still here, I'd tell him he needs to fetch then diff...16:57
mysphyt fr0sty-away: Aha! Thanks heaps. Solved the problem.16:57
← zhengyang left16:57
mysphyt cmyers, fr0sty-away: Thanks a bunch. Now I can actually do work, instead of doing work in preparation to do work. :P16:57
(Except now I have to go to an unrelated and probably irrelevant meeting. Huzzah.)16:58
Thanks again, folks. I'm sure I'll be back with another dumb question later. :)16:58
← mysphyt left16:58
devth dAnjou: that graph looks nicer than my gitx. what app is that?17:01
← Plumber left17:01
dAnjou devth: http://digilander.libero.it/mcostalba/17:01
→ ph^ joined17:02
dAnjou devth: so you can read those graphs?17:02
devth dAnjou: i can make sense of them, yes17:03
fr0sty-away dAnjou: yes, what's the issue...17:03
?17:03
← lahwran left17:03
fr0sty-away looks like you are doing a bunch of random merging which is not generally a good idea.17:03
← aroop left17:03
dAnjou fr0sty-away: exactly17:03
fr0sty-away actually, I'll bet you are blindly using "git pull"17:03
← d3vic3 left17:03
fr0sty-away but I may be wrong...17:04
← phaer left17:04
dAnjou fr0sty-away: no you don't17:04
i did stupid stupid things17:04
my actual problem is, how to revert some pulls and commits17:05
fr0sty-away probably not too bad,17:05
→ lahwran joined17:05
dAnjou but i can't figure out which ones17:05
fr0sty-away what do you mean by 'revert'? "create a commit which ondoes another" or "remove a commit from history"17:05
← phantomcircuit left17:05
→ phantomcircuit joined17:05
fr0sty-away is the final-result of all this work broken?17:05
or are you just trying to clean up after yourself?17:06
← radarek left17:06
→ vanmik joined17:06
→ cornfeed joined17:06
dAnjou i want to clean up17:07
fr0sty-away why?17:07
← subsonic left17:07
dAnjou because i did some commits on a branch without pulling before17:07
fr0sty-away why does that matter?17:07
dAnjou and some commits on some other branches that use the work of the pulled branch17:08
cmyers dAnjou: ok, looking17:08
← Zephyrus left17:08
cmyers believe it or not, that looks pretty normal to me.17:08
dAnjou it's really messed up17:08
fr0sty-away in what way? is the code in the branches broken?17:09
were the merges done incorrectly?17:09
→ alextheger joined17:09
cmyers you have a few topic branches, and they are diverged from the remote branches. I see merges where you convert in various places too.17:09
fr0sty-away what problem are you trying to solve, other than the graph is not 'pretty'?17:09
← israel_dahan left17:09
dAnjou no, but they are not up to date, we did some changes that .. well actually broke thing i did later17:09
→ d0k_ joined17:09
→ sunnyohno joined17:09
cmyers dAnjou: which branch is not up to date? now that we are looking at the picture, say exactly what you need to do17:10
→ bran joined17:10
fr0sty-away bows out...17:10
← d0k left17:10
d0k_ → d0k17:10
dAnjou ok, i think i should become clear in my mind first17:10
← sunnyohno left17:12
→ Zephyrus joined17:13
→ sylr joined17:13
← KKA left17:13
dAnjou maybe it would help if you can tell me what the edges of the graph mean17:14
→ tydeas joined17:14
tydeas I have this `strange` ( for me atleast caz there must be an explanation ) workflow http://pastebin.com/GTS9K6pa17:15
← beiter left17:15
→ aroop joined17:15
tydeas The summary is that I add a .gitignore file to assets folder ( line 31 of pastebin ) but this does not appear in the git status.17:15
Is this because file test I added is an empty one ?17:16
cmyers dAnjou: git is fundementally different from SVN/CVS/perforce17:16
each commit is an object which contains a pointer to a tree of files17:16
and a pointer to the version before it17:17
→ KKA joined17:17
cmyers so the "entire history" is represented as an acyclic directed graph - which is what you are looking at17:17
each dot is a revision and each edge is a pointer from one revision to a parent revision17:17
← ph^ left17:17
cmyers when you merge, you get a commit with more than one parent17:17
← conan left17:17
← excess^ left17:18
cmyers does that clarify things a bit?17:18
alextheger I have a rebasing problem. I diskribed it on http://pastebin.com/DRfQ5VqF . Any idea how I can rebase that correctly?17:18
→ ph^ joined17:18
→ beyondcreed joined17:19
dAnjou cmyers: can i see which commits belong to which branches in the graph?17:19
← mdel left17:19
engla alextheger: rebase help lists --preserve-merges that you can try17:20
→ MikhX joined17:21
→ israel_dahan joined17:21
alextheger engla: thx. I'll try that. Sounds promising.17:21
← empity left17:21
cmyers dAnjou: you can find out which tags contain a given commit using a flag to `git describe` you can see if a particular commit has another commit in its history by running `git log <FOO> | grep <BAR>`17:22
← ningia left17:22
→ giallu joined17:22
→ radarek joined17:22
→ squiggy joined17:22
← jatt left17:23
→ thiago_home joined17:23
→ oriba joined17:23
← oriba left17:23
→ excess^ joined17:24
→ kenneth_reitz joined17:24
← alextheger left17:24
→ metcalfc joined17:24
→ trueg joined17:24
trueg hi17:24
→ WebDragon joined17:25
← drev1 left17:25
← allaire left17:25
dAnjou cmyers: i think i just found the needle in the haystack17:25
→ drev1 joined17:25
trueg does anyone know how I could insert the last modification date into a file the way one could do in svn?17:25
← excess^ left17:25
trueg maybe a commit hook would even be sufficient?17:25
dAnjou cmyers: thanks for your explanations17:25
→ excess^ joined17:26
Guedes cmyers, for my problem i found an option for cherry-pick "-x When recording the commit, append a line that says "(cherry picked from commit …)" to the original commit message in order to indicate which commit this change was cherry-picked from." is that what i need!17:26
cmyers no prob =)17:26
WebDragon I rebased a branch that had merely been checked out at a particular state with no additional commits on it, to the tip of develop, but this seems to have wiped out the 'state' things were in for the one file in question that was why I'd created that branch in the first place. now I would like to know how to 'undo' that rebase, or if there's any way I can dig out that bit of copy from git's insides, so that I can re-apply them to this17:27
temporary branch and make an actual commit so as to prevent this problem in the future17:27
→ Quentarez joined17:27
cmyers sweet17:27
← excess^ left17:27
→ excess^ joined17:27
ChanServ set mode: +v17:27
← cakehero left17:28
WebDragon http://fpaste.org/O3Ot/17:28
→ Plouj joined17:28
Plouj hi17:28
WebDragon which is strange as Products/Radial_Flow_Settler.php had different copy on it than the one in the main develop tree17:28
kylehayes sitaram: thank you for your sentiments re: gui17:28
I guess there are some I just can't change17:29
← rien left17:29
→ sechrist joined17:29
Plouj I made the mistake of doing git co -b origin/release_2.6. How do I now get rid of the local origin/release_2.6 branch? Just rm refs/heads/origin/release_2.6 ?17:29
← radarek left17:29
Quentarez Greetings! I'm fairly new to git (came from SVN, git is so much better) and I've just discovered .gitignore files. I've updated my .gitignore with the recommendations on github/gitignore, but I already have tracked files in there which would have been ignored.17:29
Is there a way for me to "apply" my new gitignore to my repository and remove all the files that should not be tracked?17:29
← trueg left17:29
→ __sorin__ joined17:30
cmyers Plouj: `git branch -d origin/release_2.6`17:30
Plouj cmyers: that says it can't find it17:30
cmyers hmm, maybe try refs/origin/release_2.617:30
erm, or refs/heads/...17:30
Plouj cmyers: I mean it says "is not fully merged"17:30
cmyers oooh17:30
right17:31
-D then17:31
capital D17:31
Plouj I'm kinda scared that it says it's not fully merged17:31
cmyers before you delete it17:31
make another branch pointing to it17:31
Plouj git log refs/heads/origin/release_2.6..refs/remotes/origin/release_2.6 is empty17:31
dAnjou ok, now i got a concrete question: how can i delete 2 of my commits without deleting things in my working copy?17:32
cmyers maybe it is just confused. look at gitk or equivalent to make sure you aren't going to lose anything when it goes away17:32
then delete with prejudice =)17:32
Plouj k17:32
cmyers dAnjou: which commits?17:32
→ avinashhm joined17:32
← ph^ left17:32
→ avinash_hm joined17:32
→ nielx joined17:32
← ew0 left17:32
fr0sty-away tydeas: .gitignore doesn't work on tracked files.17:32
← mikelangelo left17:32
dAnjou cmyers: in my picture the latest on the black edge17:33
← boutcher left17:33
dAnjou the 2 latest17:33
← tatsuya left17:33
cmyers so yuo want to delete "it's firstName, not surName" and "Publisher has been added"? ?17:33
tydeas fr0sty-away: i know that17:33
WebDragon can anyone assist with my question?17:33
tydeas the files under the assets was untracked17:33
WebDragon: ask17:33
WebDragon repeats his question17:34
→ boutcher joined17:34
WebDragon http://fpaste.org/O3Ot/17:34
I rebased a branch that had merely been checked out at a particular state with no additional commits on it, to the tip of develop, but this seems to have wiped out the 'state' things were in for the one file in question that was why I'd created that branch in the first place. now I would like to know how to 'undo' that rebase, or if there's any way I can dig out that bit of copy from git's insides, so that I can re-apply them to this17:34
temporary branch and make an actual commit so as to prevent this problem in the future17:34
← __sorin__ left17:34
dAnjou cmyers: no, the "firstName" one and the "nullable, not null" one17:34
WebDragon the results in the fpaste are strange as Products/Radial_Flow_Settler.php had different copy on it than the one in the main develop tree17:34
cmyers dAnjou: oh, I see why I missed it, it was highlighted17:35
ok17:35
so what is your goal - you want to remove these commits from the grails-refactoring branch?17:35
← slonopotamus left17:35
fr0sty-away tydeas: 'git add assets' tracked it. That or I misunderstand your question...17:36
dAnjou cmyers: actually i want to make the changes i did in these two commits after the latest commit of my collegue17:36
← nevyn^ left17:36
dAnjou because they collide17:36
← shevy left17:37
cmyers there are three "primatives" to edit history. `git reset --hard` says "make this ref point to this commit", `git rebase -i` lets you interactively edit, reorder, or remove commits from the middle of a history, and `git filter-branch` or `git filter-tree` lets you change all commit objects in a branch by a given rule17:37
which commit conflicts with them?17:37
→ radarek joined17:37
fr0sty-away WebDragon: you can check that commit out from the reflog: 'git checkout -b <newbranch> HEAD@{<somenumber>}'17:38
cmyers the one pointed to by hte upload branch?17:38
fr0sty-away get <somenumber> from 'git reflog'17:38
dAnjou cmyers: you see after the "Publisher" commit there are 3 commits of my collegue that diverged17:38
← bsod1 left17:38
tydeas fr0sty-away: ok17:39
fr0sty-away: thanks17:39
→ ew0 joined17:39
cmyers dAnjou: ahah! at last the actual goal is revealed17:39
so you want to merge those three commits into the grails-refactoring branch, but your last two commits conflict with them - is that right?17:40
→ cybernd joined17:40
← Quentarez left17:40
→ jtza8 joined17:40
dAnjou cmyers: that's right .. i think i finally figured out what my problem is17:41
jtza8 If I moved files manually, is there a way to let git know, so that the changes can be tracked, rather than me adding a new file?17:41
fr0sty-away jtza8: git rm <oldfile>; git add <newfile>17:41
cmyers are you sure you can't merge them by fixing the conflict? if you `git checkout grails-refactoring` then `git merge origin/grails-refactoring` you should get a merge conflict and an opportunity to fix it17:42
fr0sty-away git does no explicit rename tracking17:42
but it does do on-the-fly rename detection afterward.17:42
savoxis → SavoxUnity17:42
jtza8 thanks fr0sty-away, so there isn't much to worry about, I mean, most of the text in the file is still the same..17:42
fr0sty-away so "git mv file newname" is essentially "git rm file; git add newname"17:43
jtza8 Once again, git rocks! :)17:43
WebDragon fr0sty-away: lovely. perfect. JUST what I needed. (makes mental notes)17:43
dAnjou cmyers: i'm pretty sure i can resolve conflicts .. i didn't make severe changes in my commits17:43
cmyers dAnjou: are you thinking it would be easier to fix it in the other direction - if you applied your changes, one by on, on top of his?17:43
tydeas fr0sty-away: yes... now i am somehow screwed... I have used git rm --cached to rm the assets folder17:44
now i need them back17:44
I `rm ./assets` i create them touch files17:44
but git status can not find it.17:44
WebDragon fr0sty-away: is there a way for me to have 'a branch checked out, as a stopping point but with no additional commits', rebase onto the main develop tree but prefer the content state at that time ?17:44
fr0sty-away WebDragon: I have no idea what this means: "as a stopping point but with no additional commits"17:45
or "the content stat at that time"17:45
dAnjou cmyers: that would be easier17:45
cmyers dAnjou: k...git is super sweet, so that is eas17:46
easy*17:46
← ch077179 left17:46
fr0sty-away tydeas: if .gitignore is set to ignore them, git status will not show them.17:46
cmyers here is what you do - first, create a local branch pointing to your collegue's changes - `git checkout -b fix origin/grails-refactoring`17:46
dAnjou cmyers: and it's not much to do. it would be ok to completely remove my commits and changes17:46
← SavoxUnity left17:46
tydeas fr0sty-away: I have removed the folder as well as the data inside of it.17:46
fr0sty-away you need to 'git add -f <files>' in that case, though I am completely unsure of what you are trying to accomplish...17:46
→ savoxis joined17:46
cmyers now, you are on a branch with those changes. You can ask git to "cherry-pick" each one of your changes, one by one, resolving conflicts as you go17:47
→ Saur joined17:47
← savoxis left17:47
WebDragon fr0sty-away: 'or the contents at that time' it was supposed to read17:47
← gigamonkey left17:47
tydeas fr0sty-away: I am trying to accomplish having assets folder as empty. Just to have it in structure.17:47
→ savoxis joined17:47
cmyers `git cherry-pick <firstName commit>` then `git cherry-pick grails-refactoring`17:47
after each cherrypick if there are conflicts you will have to fix them.17:47
dAnjou i'll try17:47
← dyer_out left17:47
→ dim_ joined17:47
fr0sty-away tydeas: git cannot track empty directories17:48
tydeas fr0sty-away: I know17:48
→ SavoxUnity joined17:48
tydeas fr0sty-away: trying to work around this17:48
by having a gitignore in it.17:48
cmyers when you are done, the fix branch will point to a commit that has both your changes and your college's and it will be a fast-forward push to send it to the repo when you are ready.17:48
fr0sty-away touch assets/.gitignore; git add assets/.gitignore;17:48
← savoxis left17:48
← SavoxUnity left17:48
dAnjou cmyers: i get "error: you need to resolve your current index first17:48
fr0sty-away problem?17:48
WebDragon fr0sty-away: say at some point along develop-tree's line I stop, checkout a branch that may lead to being merged in at some point but the verbiage hasn't been approved yet (awaiting proofing). then continue on develop-tree's branch deleting the copy that hasn't yet been approved and making other changes. holding that branch aside as a 'state' that may be merged into develop in the future17:48
dAnjou cmyers: for the checkout17:48
→ SavoxUnity joined17:48
cmyers dAnjou: `git status` - do you have a half-done merge you tried before or something?17:49
dAnjou cmyers: maybe i should undo the pull first17:49
WebDragon but since it was the active content in the file at the time the branch was checked out, there were no additional commits on the branch17:49
cmyers yes - if you have NO changes in your working tree you need to save, then do `git reset --hard`17:49
if you need to save them, then `git stash`17:49
→ kipras_ joined17:49
WebDragon i.e. I got told to remove the as-yet-unapproved content, take the page live on the main site, and keep that extra stuff live in devel.* site so the client can proof-and-approve it17:49
this was the simplest way I could think of to preserve the content. I hadn't expected the rebase to eat it.17:49
primarily, I presume, because there were no additional commits on that branch17:50
just the 'state' at which it was checked out, and which had some additional copy that was later removed in the main develop tree17:50
fr0sty-away WebDragon: still not following you, sorry. can you fake up some ascii art, or similar?17:50
GodEater_ → GodEater17:51
← curtana left17:51
→ p3rror joined17:52
→ ki__ joined17:52
WebDragon easy, main develop has four paragraphs of description. we find out client's approved two of them but not all four and wants to go live with the page NOW without having approved the other two paragraphs. I checkout the current state (containing all four) as a separate branch. then switch back to develop, remove the two paragraphs that were as yet unapproved. push the develop tree live on the website, and go on hacking, while that page17:52
awaits approval17:52
← nielx left17:52
WebDragon easier this time?17:52
→ nielx joined17:53
← kipras left17:54
fr0sty-away I think so, then what happened?17:54
← preflex left17:54
→ tty2 joined17:54
WebDragon I had a series of changes on develop, pushed that live to the devel.* and live websites respectively, then boss said, don't forget to keep that extra copy -- it's still awaiting approval17:55
so I rebased that branch to the top of develop so I could push develop + branch and just have that extra two paragraphs in that one file -- only it ate the changes and it was as if they hadn't existed (when I went to check out the branch and push it to the devel server -- those two p's were missing)17:56
fr0sty-away extra copy? it's just a pointer to a commit and that commit is in the ancestry of the current "develop" branch.17:56
WebDragon yeah there was no actual commit on the branch17:56
fr0sty-away rebasing that branch on top of a descendent of itself just moves the pointer.17:56
WebDragon so I saw17:56
← tty1 left17:56
fr0sty-away "no divergent" commits you mean...17:56
WebDragon NOW I have an actual commit on that branch that contains the additional copy17:56
fr0sty-away besides, commits are not 'on a branch' anyways...17:56
← sraue left17:57
WebDragon so it shouldn't be a problem again17:57
→ spetrea joined17:57
fr0sty-away not sure what you mean, did you re-add the two paragraphs?17:57
spetrea any way to git-add nonrecursive ?17:57
-n ?17:57
fr0sty-away then make a commit on top of master?17:57
→ sraue joined17:57
WebDragon I was just curious whether when rebasing, I could have preferred 'theirs' to 'ours' when diffing17:57
fr0sty-away err, develop17:57
ki__ Your branch is behind 'origin/master' by 3 commits, and can be fast-forwarded.17:57
fr0sty-away WebDragon: there was no diffing, it just moved the pointer17:57
ki__ How do I do fast forward?17:57
fr0sty-away ki__: git merge origin/master17:57
thiago_home git pull will be enough17:57
WebDragon no I stayed on the extraneous branch, and by checking the diff with the reflog info I got, obtained the extra two paragraphs and made a commit on that branch containing them17:58
which was now at the tip of develop17:58
→ rosco_y joined17:58
ki__ error: Your local changes to 'bars.php' would be overwritten by merge. Aborting.17:58
Please, commit your changes or stash them before you can merge.17:58
I don't need to keep local changes though I don't think, but if i did, would I have to commit local changes before I can merge with origin/master?17:58
→ amirite joined17:58
fr0sty-away so: commit---commit---develop(with two paragraphs)---extraneous(with four)17:59
right?17:59
thiago_home if origin/master has changes to the same file, yes17:59
WebDragon fr0sty-away: yeah17:59
fr0sty-away ki__: or stash them (like the message says)17:59
ki__ and if I don't need them, is there a faster way to revert them than git checkout -- onefile.file at.file a.file time.file ?17:59
WebDragon now I can rebase that (if it becomes necessary) without losing that data17:59
fr0sty-away or merge, but yes...18:00
WebDragon it's just a silly temporary branch that needs to be preserved for two paragraphs of data that need to stay in devel.* site until approved/denied18:00
yea, or merge18:00
fr0sty-away you could also have tagged the commit which removed the paragraphs and 'git revert <tag>' later to reinstate them.18:00
or whatever...18:00
WebDragon hm never thought of that18:01
good to know/be reminded of18:01
thanks fr0sty-away :)18:01
or fr0sty-not-so-away18:01
:D18:01
cmyers for i in `complicated list of files`; do git checkout <sha1> -- $i; done18:01
=D18:01
where the backticks could even just be `cat filelist.txt`18:02
→ slonopotamus joined18:02
thiago_home git checkout <sha1> -- `complicated list of files`18:02
cmyers and you could > them into the list in advance18:02
← v0n left18:02
cmyers oh, yes, or that.18:02
→ caseymcg joined18:02
→ erichammond joined18:02
WebDragon yes! another 'useless use of something like cat' has been resolved18:02
cmyers could also be a `find . -type f -name '*.xml'` or something like that18:03
→ Yuuhi joined18:03
cmyers not useless if you want to do more than one command, like `checkout` followed by `add` or something18:03
WebDragon backticks ftw18:03
← Saur left18:03
→ q0rban joined18:03
WebDragon agreed the for i in ..; do ; done is useful18:04
← Neo left18:04
← phantomcircuit left18:04
WebDragon I typically use it when using imagemagick to create thumbs of snaps for the image galleries or whatever, and have to have a consistent naming convention when done18:05
→ phantomcircuit joined18:05
→ Neo joined18:06
ki__ fr0sty-away: hey i stashed but it didn't really move the files i had changed locally. How can I get rid of these files locally so i can merge?18:06
← caseymcg left18:06
← guardian left18:06
← erichammond left18:06
thiago_home ki__: untracked files?18:07
→ pantsman joined18:07
← pantsman left18:07
→ pantsman joined18:07
ki__ thiago_home: no, Changed but not updated:18:07
thiago_home ki__: stash them or commit them18:07
spetrea any way to git-add nonrecursive ?18:07
ki__ it says: use "git checkout -- <file>..." to discard changes in working directory18:07
but that isn't working, like it normally does!18:08
and git stash says it stashed, however, i still have those files listed as "Changed but not updated"18:08
i think something is wrong ... ?18:08
thiago_home is the same file listed in the other section of git status?18:08
← psoo left18:11
← Adaptee left18:11
← thiago_home left18:11
adamholt → adamholt_away18:11
adamholt_away → adamholt18:11
← ddilinger left18:11
← bryanl left18:12
← MikhX left18:12
→ jutaro joined18:14
← avinashhm left18:15
← avinash_hm left18:15
→ bonaventura joined18:18
→ mklappstuhl joined18:18
← tjholowaychuk left18:19
→ codeshah joined18:20
→ murtada joined18:21
→ bitkiller joined18:21
→ tjholowaychuk joined18:21
fr0sty-away suspects line endings...18:22
← mikl left18:23
→ jmcantrell joined18:23
aroop → aroop|away18:24
aroop|away → aroop18:24
← codeshah left18:25
← Corren left18:25
← bonaventura left18:27
etcetera is .gitignore supposed to be in the root directory?18:27
← snearch left18:27
nielx say I have a remote git repo, and a local clone (which might not be a direct clone). My clone differs, and I want to find out which parent revs are in common with the remote repo. Could I find out?18:27
dr_lepper etcetera, it can be in any directory18:27
etcetera, it provides rules for the directory, where it is, and below18:28
etcetera ok.18:28
what does this do: 'git remote add origin'18:30
?18:30
I am trying to push a locally init'd repo to github.18:30
I created the repo using the github ui already though.18:30
I'm guessing I don't need that as I've already completed that step.18:30
?18:30
or does it just alias 'origin' to the remote repo?18:30
cbreak git remote add adds a new remote18:31
dr_lepper etcetera, you'd want to read man git-remote18:31
jast etcetera: [automatic message] the 'git-remote' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-remote.html18:31
dr_lepper etcetera, basically, it creates an entry in your repository's remote list18:32
← jdav_gone left18:32
etcetera dr_lepper: so I don't need it, I can just issue the push to the repo?18:32
dr_lepper it's like an entry in an addressbook18:32
→ Corren joined18:32
etcetera right, 'alias' ;-)18:32
dr_lepper well, technically you can push to the address itself18:32
etcetera sure, but why do they call it "origin"18:33
dr_lepper it's just a convention18:33
etcetera sounds very...stargate.18:33
by which I mean, fanatical and useless ;-)18:33
→ Chillance joined18:33
dr_lepper when you use "git clone" it creates an entry named "origin" in the cloned repository's remote list18:33
etcetera ah.18:34
dr_lepper of course you can name the remote any way you want18:34
→ drizzd joined18:34
etcetera makes sense. thanks.18:34
← lanthan left18:34
→ Pilif12p joined18:34
→ dannas joined18:35
→ mdel joined18:35
← Titosemi left18:36
→ vdv joined18:36
etcetera darn, I hate this pipe so much!18:36
the download / upload rate keeps on diminishing till it's basically nothing.18:37
← lsolesen left18:37
→ gebi joined18:38
etcetera does github offer shorturls for repo names?18:39
← scrooloose left18:39
→ scrooloose joined18:41
dr_lepper etcetera, what do you mean?18:41
etcetera dr_lepper: https://github.com/SomeUserName/SomeRepositoryName -> git.ub/as948jX18:42
→ HG`` joined18:42
← Textmode left18:42
← Vile left18:42
etcetera or...for a given gist.18:42
dr_lepper ask at #github18:43
rndbit → zz_rndbit18:43
dAnjou cmyers: i messed things up more, so i decided to go the hard way and discard all my local changes (that's not really tragic) and clone a clean working copy. but i want to give you a huge thanks for your time .. also i learned a bit about git today :)18:44
← HG` left18:44
cmyers dAnjou: I'm glad it was helpful.18:45
albel727 cloning anew is not the hard way. fixing it all is harder. )18:45
cmyers albel727: that depends.18:45
albel727 well, fixing it all is actually easier, if you know what you're doing. but not in that case, obviously. =)18:46
cmyers fetch + reset is exactly equivaelnt to cloning unless the server F'd something up, but you still have commits around to pull diffs from instead of rewriting changes from scratch18:46
→ ExtraSpice joined18:47
albel727 fetch + reset = cloning? I wouldn't call it that. fetch is usually only a small incremental update, while clone is a one-shot get-it-all heavy operation.18:47
hyperair albel727: git init; git remote add origin git://blah; git fetch origin; git reset --hard origin/master18:48
cmyers what hyperair said.18:48
← necromancer left18:49
← ivey left18:49
dr_lepper hyperair, also --set-origin18:49
albel727 I'm well aware of it, thanks. =) but again, fetching big repo != further incremental fetches, as anyone with a slow connection can tell )18:50
dr_lepper s/origin/upstream/18:50
hyperair dr_lepper: which git command is that?18:50
ah18:50
→ Captain`Ramen joined18:50
→ Textmode joined18:50
← Captain`Ramen left18:50
→ captainramen joined18:51
← nielx left18:51
← AkumaStreak left18:51
→ ShooterMG joined18:53
→ intripoon_ joined18:53
→ thews_ joined18:54
→ fakewaffle joined18:55
← icwiener left18:56
← thews left18:56
thews_ → thews18:56
← intripoon left18:56
← thm left18:56
→ caseymcg joined18:59
→ LongBeach joined18:59
→ nenjordi joined19:00
→ nielx joined19:01
→ ymasory_ joined19:01
→ tzervo joined19:02
→ cheezus joined19:02
→ pyrony joined19:02
cheezus hey guys, I've created a branch by doing git checkout -b mybranch . Now, how do I make it so that when I do "git push", it pushes my new branch as well?19:02
dAnjou cheezus: it does so automatically .. afair19:03
SavoxUnity → SavoxGlutton19:03
cheezus dAnjou: doesn't seem to… When I say "git push" it says everything up to date19:04
→ sterlo joined19:04
dAnjou did you commit things?19:04
→ berserkr joined19:04
cheezus dAnjou: I did19:04
→ preflex joined19:05
← caseymcg left19:05
cmyers cheezus: if you are ever confused, give git push more info. `git push <remote> <localbranch>:<remotebranch>`19:05
→ necromancer joined19:05
cmyers then you know exactly what you are telling it and see if it still says that19:05
cheezus cmyers: hrm, good point, let me try that19:05
→ caseymcg joined19:05
cmyers the data in .git/config tells git what to push when you do `git push`, but I never use that, it is too confusing.19:05
← noonoooo left19:05
cmyers besides, good tools don't read your mind, you tell them exactly what you want.19:06
→ erichammond joined19:06
nenjordi hi all, I'm doing a "git rebase -i 12345^" and now I realized the log is too big to go line by line today. Is there a way to store my changes and continue later or when I leave the editor it will be too late?19:06
← tzervo left19:07
cmyers nenjordi: you can type literally anything into that19:07
it's a domain language19:07
cheezus cmyers: when I do it that way, it says: " * [new branch] …" as I expect19:07
now how do I get it to do this automagically?19:07
cmyers so just save the output, then later, do another rebase and load the file into your editor19:07
dAnjou cheezus: maybe you have to do it the first time manually19:07
cmyers cheezus: see `git branch -t` and `git pull -u` I think, among others.19:07
← digitteknohippie left19:07
cmyers there is more than one way to set up "tracking branches"19:07
dAnjou i remember doing a git push and it pushed all branches that had something to push19:08
nenjordi cmyers: thanks a lot19:08
→ alester joined19:09
cheezus so I did git push <host> <local>:<remote> and then git branch —set-upstream remotes/<branch>19:09
that looks stupid19:09
legumbre_ → legumbre19:09
canton7 cheezus, git push -u is handy19:10
nenjordi cmyers: now I get commit messages in editor, is that normal?19:10
btw: it is my first rebase19:10
:D19:10
cmyers nenjordi: yes, the first editor is teh "script" it is going to run19:10
cheezus canton7: ah, interesting19:10
cmyers if you did any edits or squashes, it will pop open an editor for each one so you can edit the msg or squash the two messages together19:11
"-i" means "interactive"19:11
canton7 the bit of "modern" git magic i particularly like19:11
← phantomcircuit left19:11
cheezus so would this be ok: git branch -b whatever master && git push -u ?19:12
nenjordi so, it will show me the commits for all the ones I modified right?19:12
cheezus that question is for canton719:12
nenjordi the ones I didn't modify will be ignored?19:12
canton7 cheezus, assuming that the branch doesn't yet exist in <remote>, you have to tell gut what branch to push where (remember you might have two remotes, and git doesn't want to assume which you want it to push to), so it's more like git push -u <remote> <new_branch>19:13
cheezus, there's no option -b for git branch though. you're getting confused with checkout :)19:14
→ digitteknohippie joined19:14
canton7 → canton7_away19:14
cheezus canton7_away: of course19:14
canton7_away → canton719:15
← berserkr left19:16
→ Spockz joined19:16
← keystr0k left19:17
→ owst joined19:17
← Error404NotFound left19:19
← workmad3 left19:19
→ psoo joined19:19
→ berserkr joined19:21
← Swaagie left19:21
nenjordi cmyers: should I have changed the file name?19:21
cmyers: :s19:21
← aminpy left19:22
→ simosx joined19:22
→ phantomcircuit joined19:22
← dreiss left19:23
nenjordi have to go, thanks anyway :D19:23
← nenjordi left19:23
→ chrisdev joined19:25
← montylounge left19:26
← psoo left19:26
aroop → aroop|away19:26
→ jds joined19:26
→ tyrone_slothrop joined19:28
← mando left19:28
→ jdav_gone joined19:30
→ Eaven joined19:30
→ Error404NotFound joined19:32
→ Ziaix joined19:33
← Ziaix left19:34
← ki__ left19:34
→ legumbre_ joined19:35
← jds left19:36
tyrone_slothrop my git repo seems to have gotten into quite the strange state. a git diff-tree on two commits shows a file as deleted but there's no commits on the file when I do a git log with the same commits so I'm curious why git deletes the file when I merge.19:37
← legumbre left19:38
tyrone_slothrop 'git diff-tree --stat --summary A..B -- file.txt' shows it as deleted yet 'git log A...B -- file.txt' shows nothing.19:38
→ hobodave joined19:39
→ res0nat0r joined19:40
→ xerora joined19:41
← murtada left19:42
→ MikhX joined19:42
→ tech13 joined19:45
tech13 Is there a way to define the order for 'git remote update' ? Have a couple of remotes, that tend to have many of the same things, but some have notably better network connections then others.19:46
← res0nat0r left19:47
patrikf tech13: try setting remotes.default (you'll have to update it manually though for new remotes)19:47
← Vampire0_ left19:49
serverhorror hmm when I create a tag with git and do a git svn dcommit afterwards there's no magic that will somehow create a tag in svn, correct?19:49
→ dreiss joined19:50
wereHamster serverhorror: corect19:50
← alester left19:51
d1zzy if you want magic you should buy some iStuff19:51
← beatak left19:52
drizzd and if you tag previous to dcommit, the tagged commit does not generally exist in svn19:52
← OOPMan left19:52
→ beatak joined19:52
← ymasory_ left19:53
tech13 patrikf, looks like that would do it. Don't see much about what format should be in that option though. Happen to have a link that explains it, or gives a sample value?19:53
← jutaro left19:53
→ tty3 joined19:54
← owst left19:54
patrikf tech13: probably either space-separated or giving the same config key twice (or using git config --add)19:54
→ swtaarrs joined19:55
← MikhX left19:56
← tty2 left19:57
← rosco_y left19:57
← drubies left19:58
→ airborn joined19:59
← RaceCondition left20:00
→ alester joined20:00
→ RaceCondition joined20:00
← RaceCondition left20:00
→ RaceCondition joined20:01
← RaceCondition left20:01
→ dl9pf_ joined20:01
← dl9pf left20:01
← chrisf left20:02
chrisf|work → chrisf20:02
→ chrisf_ joined20:02
← alester left20:02
→ RaceCondition joined20:03
→ tyfighter joined20:03
→ Vampire0_ joined20:03
→ alester joined20:04
Vampire0_ → Vampire020:04
→ guns joined20:06
→ OOPMan joined20:08
→ MikhX joined20:08
← nielx left20:10
→ dnorman joined20:10
← dnorman left20:11
← deekayen left20:11
→ flavius joined20:11
SavoxGlutton → JesusLovesYou20:12
→ deekayen joined20:12
flavius Hi. How can I check out a remote 'abc' into a new branch?20:12
← [intra]lanman left20:14
← cornfeed left20:14
→ [intra]lanman joined20:14
← guns left20:15
canton7 flavius, check out a remote? do you mean a branch from a remote?20:15
← ping-pong left20:17
← hvgotcodes left20:17
← LongBeach left20:17
flavius canton7: I've followed the instructions http://bradlyfeeley.com/2008/09/03/update-a-github-fork-from-the-original-repo/ however i would like that remote to be "attached" to a new branch20:17
and not merge directly from the remote20:18
canton7 ah, so you mean the *master* branch on the remote. a remote is an entire git respository, with potentially multiple branches20:18
flavius yes20:18
→ Vile joined20:19
→ mando joined20:19
← jrabbit left20:19
canton7 just git checkout -b local_branch_name remote_name/master will do it20:19
that won't set up any tracking20:19
so a git pull won't merge anything in from that remote20:19
← Juo left20:20
← mklappstuhl left20:22
flavius thanks it worked20:23
and if i want to see the diff between that branch and the currently active master branch?20:23
← swtaarrs left20:23
canton7 addendum - if you do want tracking, use -t instead of -b (-t implied -b)20:23
**implies20:23
flavius what does tracking mean?20:23
canton7 git diff remote_name/master master20:24
→ nielx joined20:24
legumbre_ → legumbre20:24
← Fuu left20:24
canton7 in this context, it means that git knows that you want a "git pull" to pull from that remote when that branch is checked out20:25
→ ping-pong joined20:26
→ Linkadmin joined20:26
← dannas left20:26
→ zivester_ joined20:26
Linkadmin when you do a git checkout -b myBranch, does it create the branch based on HEAD, or does it include your uncommitted changes?20:26
canton7 keeps everything uncommitted20:27
(and untracked)20:27
flavius when i use -t it says "fatal: Missing branch name; try -b"20:27
→ gebbione joined20:27
canton7 git checkout -t local_name remote/master ?20:28
flavius yes20:28
→ rcu joined20:30
canton7 my bad. git checkout -t origin/branch will create a new local-tracking branch called "branch"20:30
git checkout -t -b local origin/master will create a new local-tracking branch called "local"20:30
← d0k left20:30
← tanuki_ left20:30
flavius ook and how to create a new remote-tracking branch?20:31
→ ph^ joined20:31
← hobodave left20:31
canton7 ack, i need more sleep. substitude "local-tracking" with "remote-tracking" in my last message :$20:31
flavius i know how it is i am tired too ...20:31
→ tjh joined20:31
flavius :)20:32
canton7 goes off to make a cuppa20:32
flavius thanks bud good luck20:32
canton7 np. also, if you want to set up tracking on a branch which isn't currently tracking anything, git branch --set-upstream and git push -u (either will do) are your friends20:32
→ v4n joined20:33
aroop|away → aroop20:34
→ lanthan joined20:34
← q0rban left20:34
← ymasory left20:35
← tjholowaychuk left20:35
→ workmad3 joined20:35
← g0bl1n left20:36
→ mikl joined20:36
← mikl left20:37
→ runa joined20:37
runa heyas. I have a relatively big repo and a branch checkout takes ~1minute. I've tried git-gc --aggressive with no luck. any hints?20:37
[exa] how big is the actual repository?20:39
btw. isn't your filesystem somehow slow?20:39
→ WL6784 joined20:39
→ owst joined20:40
runa [exa]: 20k files, using ext4 on a 5400rpm disk20:40
→ qdw joined20:41
→ fbass joined20:42
fbass git is cool20:42
→ jophish joined20:42
← simplechat left20:42
← owst left20:43
→ mikl joined20:44
JesusLovesYou → JesusHatesYou20:45
JesusHatesYou → Xavierbrightness20:46
← tty3 left20:47
→ tty1 joined20:48
→ ki__ joined20:49
← glcrazy left20:49
aroop → aroop|away20:49
← nielx left20:49
aroop|away → aroop20:50
→ glcrazy joined20:50
→ guardian joined20:50
→ nielx joined20:51
← jitesh_shetty left20:51
[exa] runa: that might be a FS performance problem. Probably try it in tmpfs, that is usuable for debugging speed issues. Btw I assume that CPU usage is low during checkout, is that right?20:51
← bindaas left20:52
← squiggy left20:53
← v4n left20:53
→ simula joined20:54
← tech13 left20:55
← jsick3 left20:56
→ dnorman joined20:56
← zivester_ left20:56
runa [exa]: yep. I have the output of 'time git co stable' if you want to see it20:58
← leitao left20:58
← jnylen left20:59
← deekayen left20:59
→ mysphyt joined20:59
→ jnylen joined21:00
← tgunr_ left21:00
← boombatower left21:01
→ q0rban joined21:02
← dnorman left21:02
← resixian left21:03
→ deekayen joined21:03
← dv310p3r left21:03
← radarek left21:05
→ thiago_home joined21:06
→ v4n joined21:06
← flavius left21:07
← mysphyt left21:10
→ squiggy joined21:10
← bitkiller left21:11
→ Fuu joined21:11
→ gigamonkey joined21:11
← gigamonkey left21:12
← panike left21:12
← glcrazy left21:13
← booginga left21:13
→ radarek joined21:13
→ morganick joined21:14
← quiliro left21:15
← Freakshow left21:15
→ Freakshow joined21:15
pyrony → scalabot_21:16
scalabot_ → pyrony21:16
← dj_goku left21:16
→ dj_goku_ joined21:16
← dj_goku_ left21:16
→ dj_goku_ joined21:16
← mdel left21:17
← fbass left21:17
→ brodo_ joined21:18
← deuterium left21:18
→ malumalu joined21:19
→ fbass joined21:19
← brodo left21:20
brodo_ → brodo21:20
→ cthuluh joined21:21
← chombee left21:21
→ Zenopus_ joined21:21
← Zenopus left21:22
Zenopus_ → Zenopus21:22
→ robotarmy joined21:25
← TeckniX left21:26
→ ddilinger joined21:26
→ bkosborne joined21:26
← nielx left21:26
← brodo left21:26
→ bryanl joined21:27
→ EricInBNE joined21:27
Xavierbrightness → FckOTAs21:29
← tystr left21:30
→ orafu joined21:30
← morganick left21:30
← calimocho left21:30
← PabloM left21:31
→ tystr joined21:31
→ guns joined21:31
← alex__c2022 left21:31
→ shapr joined21:32
shapr Hi, is there a Linux Standard Base -style convention for filesystem location of shared git repos? Maybe /usr/src/ or /opt or something?21:32
← bartek left21:32
← erichammond left21:33
thiago_home shapr: probably /srv21:33
← Dave^| left21:34
← tystr left21:34
shapr thiago_home: Thanks. Any idea if there's any mailing list traffic / wiki page /etc on this subject?21:34
← Guedes left21:34
shapr thiago_home: Or more succintly, why /srv instead of some other location?21:34
thiago_home well, what does the LSB say?21:34
→ tystr joined21:34
← fbass left21:35
← runa left21:35
thiago_home http://www.pathname.com/fhs/pub/fhs-2.3.html#SRVDATAFORSERVICESPROVIDEDBYSYSTEM21:35
← metcalfc left21:35
→ boombatower joined21:35
shapr thiago_home: point, that explicitly mentions cvs21:36
→ mklappstuhl joined21:36
→ bulavard joined21:39
bulavard .21:40
→ fbass joined21:40
← bkosborne left21:40
← eijk left21:41
wereHamster /var/git21:41
→ ymasory joined21:41
← bulavard left21:41
← cheezus left21:41
→ jaimef joined21:41
→ paeva joined21:42
← sterlo left21:42
→ CollinPDX joined21:42
← justin-george left21:44
← boutcher left21:45
← ISF left21:45
→ justin-george joined21:46
shapr wereHamster: The rationale given in the FHS implies /srv is the right place rather than /var21:46
← radarek left21:46
shapr thiago_home: Thanks for the help!21:47
← shapr left21:47
→ mgorbach228 joined21:47
← mgorbach228 left21:47
→ Fuu` joined21:48
→ mgorbach228_ joined21:49
→ n1ck3 joined21:49
ChanServ set mode: +v21:49
mgorbach228_ Anyone to help with a git repository problem?21:50
I tried to push to a local branch, and it won't succeed.21:50
error: Ref refs/heads/shared-outlineView-work is at <hash> but expected 0000000000000000000000000000000000000000 remote: error: failed to lock refs/heads/shared-outlineView-work To . ! [remote rejected] destiny-outlineView-work -> shared-outlineView-work (failed to lock)21:50
← Fuu left21:50
thiago_home how did you try to push21:51
?21:51
← engla left21:51
mgorbach228_ git push21:52
I have it set up to track, and push to the tracking branch.21:52
thiago_home what is the branch it's tracking?21:52
itself?21:52
← rayners left21:53
← tyfighter left21:53
← ExtraSpice left21:53
→ erichammond joined21:53
← mklappstuhl left21:53
mgorbach228_ destiny-outlineView-work is tracking shared-outlneView-work21:54
I was pushing from destiny-outlineView-work to shared-outlineView-work21:54
thiago_home try again21:54
push to . works just fine21:54
thiago_home uses that to update a branch he hasn't used in a while just before checking it out21:54
FckOTAs → SeriouslyBRB21:55
← guardian left21:55
mgorbach228_ I think it's a permissions isssue.21:55
sudo works fine on the push21:55
← SimonP86 left21:55
mgorbach228_ but without sudo the push fails21:55
← blorbx left21:56
← Spockz left21:56
mgorbach228_ I think I might have messed up the permissions for the directories inside .git.21:57
SeriouslyBRB → Savoxis21:58
← dj_goku_ left21:59
← alester left22:00
← alip left22:01
→ alip joined22:01
→ whyking joined22:01
← MikhX left22:01
← deekayen left22:02
← TrineLocke left22:02
→ MikhX joined22:02
→ TLMonitor joined22:03
→ guardian joined22:04
← trobrock left22:06
→ MikhX_ joined22:06
← MikhX_ left22:07
→ lolfrenz joined22:07
← erichammond left22:07
→ erichammond joined22:07
← ThomasLocke left22:08
← MikhX left22:09
← excess^ left22:10
← airborn left22:10
← Dynetrekk left22:11
← Tommy[D] left22:11
simula \clear22:12
FauxFaux |clear22:12
← guardian left22:13
→ engla joined22:13
← HG`` left22:14
→ sigh joined22:15
← joel_falcou left22:16
→ MikhX joined22:17
← rendar left22:19
→ metcalfc joined22:21
→ Tommy[D] joined22:21
→ varioust joined22:22
← sylr left22:22
← WebDragon left22:24
→ joel_falcou joined22:24
← israel_dahan left22:25
← boombatower left22:25
→ madewokherd joined22:26
← etcetera left22:27
← WL6784 left22:28
← joel_falcou left22:28
← kipras_ left22:28
→ Chibby_ joined22:29
→ paeva_ joined22:29
→ angelsl joined22:30
→ mklappstuhl joined22:30
← Deck` left22:32
ichilton Can different branches in a repo have a submodule pointing to different commits/branches?22:32
the submodule chapter in Pro Git implies not...22:33
← tstclair left22:33
← n1ck3 left22:33
← paeva left22:33
← ebi left22:34
← saschagehlich left22:34
→ saschagehlich joined22:35
patrikf I've never used submodules, but from the technical side nothing speaks against it22:36
← workmad3 left22:36
← Error404NotFound left22:36
patrikf if you have a repo at hand, try it with a fresh clone?22:36
→ trobrock joined22:36
← ph^ left22:40
← v4n left22:42
← sigh left22:43
← trobrock left22:44
→ ebi joined22:44
→ Error404NotFound joined22:46
← necromancer left22:47
→ lightcap joined22:47
ichilton patrikf: yeah, I might do some tests - thanks22:49
banghouse → walkinDAWG22:50
← sonnym left22:50
→ rpg joined22:50
→ hackeron joined22:50
← grahamc left22:50
← falican left22:52
hackeron hey, I have a question about updating submodules. My submodule is configured to be url = git://git.videolan.org/x264.git -- I do git submodule init - then I do git submodule update - this clones x264 into dist/x264 -- when I do git submodule update again after the x264 repo is updated, it does nothing :/ -- how do I get it to pull the x264 changes?22:52
← tjh left22:53
→ trobrock joined22:55
harleypig When doing 'git status -s' I have two files listed as ' T path/to/file' I can't find the meaning for the T. What does this mean?22:57
canton7 hackeron, the parent repo holds the exact commit that it wants the submodule to be at22:57
so cd into the submodule, git pull22:57
then cd back to the root of your parent project, and git status22:57
→ drubies joined22:57
hackeron canton7: when I do that, it says: "You are not currently on a branch, so I cannot use any...."22:57
canton7 that should tell you that there are new commits in that submodule (or something, can't remember the exact message)22:57
then add and commit22:58
hackeron add what?22:58
← tearsong left22:58
canton7 hackeron, point. git checkout master && git pull mebbe?22:59
git add path/to/submodule22:59
hackeron canton7: I did cd dist/x264 && git checkout master && git pull -- that said already up to date :/23:00
← mrball left23:00
← drizzd left23:00
hackeron canton7: any ideas?23:02
canton7 are you sure that the remote has updated since you did git submodule update for the first time?23:02
← Corren left23:02
← rcu left23:04
hackeron canton7: actually, wait, maybe it hasn't - tried to git pull a module to which I have control - after adding a new file - that worked23:04
← TheWarden left23:04
→ tty2 joined23:04
canton7 hackeron, sweet. we're back on predictable territory :)23:04
hackeron canton7: so I have to manually cd into each submodule, and do git pull origin?23:04
canton7 unless there's a shortcut / flag I haven't come across yet23:04
albel727 git submodule foreach?23:05
canton7 as I say, the parent repo holds the exact commit that a submodule is at (try a git diff path/sub/submodule after updating the submodule to see what I mean)23:05
← simosx left23:06
hackeron hmm, also my submodule has only one branch, master -- the others, I don't know. How do I just get it to pull whatever branch it did in the first place?23:06
← engla left23:06
thiago_home submodules don't track branches23:06
they keep an exact commit23:06
hackeron so how do I update? -- git submodule update --init works somehow -- how do I tell git - do that again?23:06
← tty1 left23:07
canton7 albel727 has a point: http://stackoverflow.com/questions/1030169/git-easy-way-pull-latest-of-all-submodules has a discussion23:07
hackeron well, when I try that, it says: You are not currently on a branch, so I cannot use any23:07
'branch.<branchname>.merge' in your configuration file.23:08
ah, so I guess I need to do: git submodule foreach git pull origin master23:08
→ joel_falcou joined23:08
← gebbione left23:08
← alberto56 left23:09
← pantsman left23:09
← Amorphous left23:10
← jtza8 left23:10
→ yogurt_truck joined23:11
harleypig T is typechange ... still can't find docs on this ... does this mean the filetype has changed? Or something else?23:11
thiago_home where?23:12
albel727 suspects that means a file got executable/symlink bit set/cleared.23:12
harleypig thiago_home: on the assumption you're asking me, git status -s is giving me ' T' on two files.23:13
← n8o-mba left23:13
ki__ git reset --hard only resets files on my working directory, right?23:13
yogurt_truck is it possible to do a git push using a specific config file with different user.name and user.email than the ones in the global configuration?23:13
→ jaws joined23:13
albel727 of course. both answers.23:13
cbreak ki__: it also changes the current branch23:14
thiago_home I don't know, sorry. I have never used status -s.23:14
cbreak yogurt_truck: git push has nothing to do with user.name and user.email23:14
yogurt_truck: so you can change it to what ever you want, it's ignored anyway23:14
harleypig thiago_home: :] no problem ... I ran git status and the same two files were labeled as typechange23:14
albel727 needs some sleep. he completely misread both questions.23:15
thiago_home harleypig: did it say that exactly?23:15
← Plouj left23:15
ki__ cbreak: the current branch locally, right? I want to get my working folder back to what the remote/origin/master is23:15
yogurt_truck cbreak: oh i see. ok thx23:15
→ Deck` joined23:15
harleypig thiago_home: yeah ' T path/to/file'23:15
→ tty3 joined23:15
cbreak ki__: yes, but it is history rewriting23:16
ki__ meh thats not what i want then23:16
cbreak if you reset it to origin/master, that should not conflict though, since you only kill local history23:16
harleypig found the answer in help for git diff --diff-filter: "have their type (i.e. regular file, symlink, submodule, ...) changed (T)"23:16
ki__ git checkout origin/master ?23:16
cbreak no23:16
unless you want to visit the past23:17
→ alberto56 joined23:17
cbreak this will detach your HEAD, so that you are not on a branch23:17
→ etcetera joined23:17
← metcalfc left23:17
cbreak this is not a state where you should commit things, so git checkout a branch before you continue working23:17
← tty2 left23:18
→ metcalfc joined23:19
← guns left23:20
→ sunnyohno joined23:20
ki__ cbreak how can i reset to origin/master?23:20
sorry the reset concept isn't completely clear to me23:20
git reset origin/master23:20
← erichammond left23:20
thiago_home harleypig: how about without -s23:20
ki__ git reset [sha-of-origin/master] ?23:21
cbreak git reset --hard origin/master... simple23:21
← saschagehlich left23:21
→ jitesh_shetty joined23:21
← RaceCondition left23:22
harleypig thiago_home: without -s it was 'typechange: path/to/file'23:23
thiago_home I don't know what that is23:23
harleypig I found the answer to what it meant in the git-diff page under diff-filter section23:23
thiago_home if it's not a mode change, maybe you replaced a file with a submodule (or the other way around)23:24
→ sonnym joined23:24
→ tty1 joined23:24
harleypig no submodules ... probably changed a hardlink to an actual file23:24
← warthog9 left23:24
← lolfrenz left23:25
cbreak git does not track hard links23:25
← ericholscher left23:26
harleypig cbreak: right23:26
oh, good point23:26
← Savoxis left23:26
← jShaf left23:26
harleypig well, then a symlink <-> regular file23:26
one way or the other.23:26
→ jShaf joined23:26
→ Amorphous joined23:27
← crashanddie left23:27
← tty3 left23:27
→ cakehero joined23:28
→ ajay joined23:28
← jaws left23:28
→ lolfrenz joined23:28
→ whyking_ joined23:28
ki__ cbreak: it says I am now current with origin/master but has a list of files that are Changed but not updated. These are the files I want to remove the concepts23:29
*remove the contents23:29
← berserkr left23:29
ki__ git reset didn't do that. neither did git stash23:29
cbreak just uncommitted changes?23:29
git reset --hard HEAD23:29
and they'll be gone23:29
git stash is not for removing changes23:29
it's for storing changes23:29
← whyking left23:29
← aroop left23:29
→ ericholscher joined23:30
ki__ ah i see23:30
cbreak: that didn't remove the local changes either.23:31
cbreak it does.23:31
it removes all uncommitted changes23:31
ki__ the only way that is working is doing this file by file: git checkout -- file1 file223:31
cbreak no.23:31
ki__ i did a git reset --hard HEAD, then git status|more and it has the same list of Changed but not updated: files23:32
← Chillance left23:33
cbreak impossible23:33
albel727 have never seen anything like it.23:33
FauxFaux crlf failure.23:33
albel727 oh23:33
cbreak well, I guess if you have windows it is possible23:33
albel727 have seen something like it.23:33
FauxFaux Due to autocrlf being the worst thing ever.23:33
cbreak but on any non-shitty operating system...23:33
FauxFaux laughs at cbreak.23:34
FauxFaux cbreak: Name two Windows applications that can't cope with unix line endings.23:34
cbreak notepad, paint23:34
why?23:35
albel727 lol23:35
FauxFaux Paint is fine!23:35
→ xxneolithicxx joined23:35
ki__ i have autocrlf on, is that my problem?23:35
FauxFaux autocrlf being on is always a problem.23:36
← fbass left23:36
← LiamH left23:36
→ fbass joined23:38
← fbass left23:38
← yogurt_truck left23:38
← TomyLobo left23:39
→ jutaro joined23:41
albel727 I've have always thought, that it's committing windows line endings on linux system, that causes the problem. imho, git diff-index could have just ignored line endings discrepancies, when core.autocrlf is true.23:42
→ berserkr joined23:42
FauxFaux s/line endings/whitespace/ ;)23:43
→ conan joined23:43
← juangiordana left23:44
→ BiggFREE joined23:44
cbreak -w23:44
albel727 sorry, I'm unable to get anything reasonable when I replace "line endings" with "whitespace" in my sentences. for instance, what is a "windows whitespace"? =) can you elaborate?23:45
FauxFaux glares.23:45
→ linusoleander joined23:45
← metcalfc left23:45
FauxFaux Basically, I'd set diff-index to default to -w, so whitespace changes wouldn't show up in status etc.23:45
albel727 ah, that explains it. indeed.23:46
→ juangiordana joined23:46
cbreak but they have to show up23:46
FauxFaux Why?23:46
cbreak they are after all changes that need to be staged23:46
FauxFaux But I don't care about them at all.23:46
BiggFREE Sorry :(23:46
cbreak then why did you do them?23:46
why did you fix indentation?23:47
FauxFaux To increase the ease of making real changes.23:47
← mgorbach228_ left23:47
cbreak it is a real change23:47
if you would not commit it, the result would look horrible23:47
FauxFaux Obviously it is, but it doesn't mean I care about it.23:47
→ warthog9 joined23:47
← conan left23:48
→ rchavik joined23:51
→ mdel joined23:53
albel727 (...python code would certainly not just look horrible...)23:53
← BiggFREE left23:53
FauxFaux More good reasons to mock python programmers. ;)23:54
→ paeva__ joined23:54
albel727 I'd rather keep all whitespaces, except end-of-lines, and don't mock anyone. )23:56
ChanServ set mode: +v23:56
Vile That reminded me - is there a way to force line endings to particular type for cloned repository?23:57
→ paeva joined23:57
← paeva_ left23:57
Vile i.e. I have bare "central" repository, from which developers clone their ones, I need to make sure that their clones contain LF line endings23:57
→ metcalfc joined23:57
albel727 unfortunately, line endings on clones depend only on their local configuration. central repo cannot affect that. you may try core.eol on clones, if you want.23:58
Vile core.eol and core.safecrlf settings are not transferred to clone, obviously23:58
← berserkr left23:58
albel727 there are no settings at all, that get transferred, obviously.23:59
so I'm not sure, that I understand your question.23:59
d1zzy you culd use post-receiv-pack hooks on the server to validate the source files against your rules23:59

Logs Search ←Prev date Next date→ Channels Documentation