IRCloggy #git 2009-08-11

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.

2009-08-11

argnl → argonel00:00
← enherit left00:00
← offby1 left00:00
← bugfux left00:02
← Derander left00:02
→ porcnick_ joined00:02
→ Squee-D joined00:02
almostautomated what causes a test to skip vs pass/fail ?00:02
← qrush left00:03
← porcnick left00:03
→ qrush joined00:05
→ Derander joined00:05
→ agile joined00:05
← Aikawa left00:06
← dasy2k1 left00:09
← porcnick_ left00:09
→ ProLoser|Work joined00:09
→ JNR___ joined00:09
← crack_head left00:09
ProLoser|Work a file got wiped and my gitignores keep getting deleted, how do i revert to unwipe the file and recreate the git ignores? all other changes have been commited (revert those files to previous commit however)00:09
zenwryly If git just tracks content, I'm a little confused what git-mv does?00:10
jast it renames the file, removes the index entry for the old name, creates an index entry for the new name00:10
→ trapni^3 joined00:11
jast ProLoser|Work: you can use git checkout HEAD <file> to get the previous version of a file00:11
→ galderz joined00:12
jast and, well, I'm not the one deleting your .gitignore files. ;)00:12
ProLoser|Work ya i just locked the permissions, hopefully it stops00:12
← d0k left00:12
← mithro left00:13
zenwryly jast: ah, so it's all just cli sugar for the index?00:13
→ AkumaStreak joined00:13
jast pretty much.00:13
→ developernotes joined00:14
zenwryly jast: but status shows "renamed" as if it's some special state00:14
jast that's rename detection00:14
← RayMorgan_ left00:14
→ RayMorgan joined00:14
zenwryly jast: I guess what I'm asking is does a "mv" have any sort of designation in the objects once committed?00:14
jast try "renaming" the same file manually00:14
mv a b; git rm --cached a; git add b00:14
← Dashkal left00:14
← Derander left00:15
binjured|afk → binjured00:15
← galderz left00:15
zenwryly jast: ok, so no special representation in the commit or tree objects then?00:16
jast no00:16
← ProLoser|Work left00:16
zenwryly jast: so if I move *and* change a file, does that break rename detection? Does that affect delta compression at all?00:17
jast delta compression is usually not very much affected by renames00:17
rename detection is hit and miss if the file gets changed in the same commit00:17
← neoeinstein left00:18
zenwryly jast: where can I find rename detection documetation?00:18
→ mw_ joined00:18
jast source code :)00:18
← developernotes left00:18
← mw left00:19
→ developernotes joined00:19
← developernotes left00:20
→ neoeinstein joined00:20
zenwryly jast: I was afraid of that. :) Well thanks a lot for telling me by hand then, exactly the info I needed.00:21
← qhoxie left00:21
jast you're welcome. and now I need sleep.00:22
→ javatexan joined00:23
← javatexan left00:23
→ mithro joined00:26
← trapni left00:27
← meanburrito920_ left00:29
← JNR___ left00:32
→ JNR___ joined00:32
→ mw joined00:32
← smashwilson left00:33
← binjured left00:33
→ binjured joined00:33
← mw left00:33
← Chaser left00:35
← binjured left00:35
→ binjured joined00:35
← justin-george left00:36
← tedoc2000 left00:38
← nd33r left00:38
→ crack_head joined00:39
← mw_ left00:40
Squee-D The git plugin im using in hudson is checking out to a revision, eg 'git checkout -f <shar here>' - i'm not seing some of the files added in this commit (any in subdirs) but i am seing those in the root of the repository getting added. Whats wrong here?00:40
→ plediii joined00:41
→ mm_202 joined00:42
← cloudhead left00:45
← cemerick left00:46
← bsy__ left00:49
→ magcius_ joined00:50
← s0ber left00:50
← magcius left00:51
← JNR___ left00:51
→ JNR___ joined00:52
→ s0ber joined00:53
→ bugfux joined00:53
← Circuitsoft left00:54
→ sitaram joined00:55
→ leanucci joined00:56
← plediii__ left00:57
← Geralt left00:58
← RayMorgan left01:00
← kennethlove_ left01:00
← sgrimm left01:00
← spaceonline left01:02
→ [KirinoxzerO] joined01:03
← [KirinoxzerO] left01:04
← sunoano left01:04
→ sunoano joined01:06
→ danielcadete joined01:06
→ jinzhu joined01:08
→ Chaser joined01:10
← sitaram left01:11
← pigeon left01:12
← plediii left01:12
→ plediii joined01:14
→ azeroth-work joined01:14
→ nd33r joined01:15
← nd33r left01:15
→ plediii_ joined01:20
→ pigeon joined01:21
← Chaser left01:22
→ hydrogen joined01:22
→ spaceonline joined01:22
← namidark left01:22
← salty-horse left01:25
← bugfux left01:26
← dreiss left01:27
→ cemerick joined01:27
cpg → cpg|biab01:28
← psynaptic left01:29
← leanucci left01:30
cpg|biab → cpg01:32
← tongueroo left01:33
← fualo left01:35
← plediii left01:35
→ plediii joined01:36
→ porcnick joined01:37
→ FunkeeMonk joined01:38
→ marvil07 joined01:39
← owen1_ left01:43
→ MrMeek joined01:44
→ josesito joined01:44
josesito If I already started some work on the master branch (but haven't added any changes nor committed anything yet) how can I make all my changes be committed to a different branch?01:45
binjured → binjured|afk01:45
← hobodave left01:45
cpg → cpg|biab01:45
tomoj josesito: git checkout differentbranch01:47
← agile left01:47
← plediii_ left01:48
← MrMeek left01:48
→ MrMeek joined01:51
← bdiego left01:52
→ plediii_ joined01:53
← plediii left01:53
← scarabx left01:56
← hydrogen left01:58
→ bugfux joined01:58
← relix left01:59
← JNR___ left02:01
→ JNR___ joined02:05
→ hydrogen joined02:07
→ developernotes joined02:07
← agnel left02:08
developernotes how do you respond when someone tells you that git isn't 'safe' b/c it allows you to commit content from your index, when you might have changes that are in your working directory?02:12
← crack_head left02:12
→ ctusar joined02:12
developernotes to me it seems like they are confusing the notion that centralized repos. models often have (i.e., svn) and thinking they need to operate the same way. Thoughts?02:12
reaVer developernotes: sorry, what?02:13
it allows you to commit your index while your working directory has been changed?02:13
I don't think git allows that by default02:14
developernotes exactly, it does.02:14
I can edit a file, add it, then edit it again. When I issue a git commit, the content that was staged in the index is committed, not what hhas changed in the working directory.02:15
reaVer I don't remember it doing that02:16
usually it complains about an unsynchronized index and what not02:16
joshthecoder_afk → joshthecoder02:16
developernotes reaVer: My sentiment is that when you issue just a git commit without the -m, it displays both the staged files as well as the files that have been modified but not added to the index02:17
reaVer yes02:17
hmmm, I might be wrong about the commiting but anyways02:17
what's unsafe about it?02:18
developernotes But w/ the -m you are aware of how the command works, so therefore it doesn't need to share this information.02:18
reaVer usually I just use git commit -am "My message here"02:18
developernotes reaVer: I don't think it's unsafe, but my co-worker does. I think he thinks it's misleading b/c there are changes in the working directory.02:19
reaVer developernotes: then he should use git commit -a02:19
developernotes reaVer: Just wondering if anyone has had this discussion with anyone else before.02:19
reaVer developernotes: if you tell git to blow itself up it will do it02:19
→ murph joined02:19
reaVer it's not git's fault if that's actually been used02:19
developernotes reaVer: Sure.02:19
← cemerick left02:19
reaVer your routine of adding the file first was done to commit specific parts of changes02:20
so that stuff you didn't actually want to commit yet doesn't get commited02:20
makes it look a bit cleaner02:20
→ cytrinox_ joined02:21
reaVer what you're doing with adding and editing after adding basically comes down to that same thing02:21
→ jjuran joined02:21
reaVer generally you just get the howitzer out though and just use git commit -a02:21
← bugfux left02:22
← WALoeIII left02:23
→ Glutnix joined02:25
developernotes reaVer: I agree.02:25
← danielcadete left02:25
Glutnix is this the appropriate room if i have a git support question?02:25
← plediii_ left02:25
reaVer Glutnix: yes02:26
but I'm going to sleep02:26
Glutnix i've made a clone on gitorious of the master repository, and have cloned that to my local. how do i switch my local to the master's 0.8.x branch to work on a fix for current stable?02:26
(of the laconica repostiory)02:27
→ tfarina joined02:29
← aziz left02:29
developernotes Glutnix: Have you tried git checkout <branchname>?02:31
Glutnix error: pathspec '0.8.x' did not match any file(s) known to git.02:31
http://www.gitorious.org/~glutnix/laconica/glutnix has the branch listed though02:32
developernotes Glutnix: taking a look at it, just a sec.02:33
Glutnix thanks in advanice02:34
→ WALoeIII joined02:35
developernotes Glutnix: try:02:36
Glutnix: git checkout origin/0.8.x -b myBranchForFix02:36
Glutnix: that worked for me.02:36
← cytrinox left02:37
cytrinox_ → cytrinox02:37
Glutnix that worked, thanks02:37
→ jackylee joined02:38
developernotes Glutnix: no problem.02:38
→ scientes joined02:38
developernotes how do people respond when others say that git is not discoverable.02:44
→ fualo joined02:45
← fualo left02:48
→ offby1 joined02:48
→ leanucci joined02:48
← vuf left02:50
→ simNIX joined02:51
← jackylee left02:52
→ tjafk1 joined02:53
→ meanburrito920_ joined02:55
cpg|biab → cpg02:59
→ jjuran_ joined03:00
← hydrogen left03:00
cxreg developernotes: what does that even mean?03:00
http://git-scm.com <- look, I discovered it!03:01
DrNick "discoverable" probably involves a GUI with buttons and menus to explore03:01
god forbid we have to read a manual03:01
→ dreiss joined03:01
SamB DrNick: having easy to navigate --help might also work ...03:02
← jjuran left03:02
cxreg git --a-clue03:02
back to his first question, anyone who says "git isn't 'safe'" is speaking entirely out of their rear-end03:03
it's provably more safe than just about any other scm03:03
developernotes cxreg: It came from a developer I work with today.03:04
SamB cxreg: it doesn't corrupt data often, true03:04
developernotes DrNick: haha....funny. :-)03:04
← offby1 left03:04
cxreg forget corruption, it's *tamper proof*03:04
can you say the same for svn?03:04
← mfilipe left03:04
developernotes cxreg: For example, he said that he was looking to revert local changes, he was looking for a revert command and was surprised when the command was reset.03:05
SamB but it's easy to lose your data in the usual sense -- that is, lose track of where it *is*03:05
or just do crazy stuff to your working tree/index03:05
← jchris left03:05
cxreg he's surprised about what commands are named, and that makes it less safe?03:06
developernotes cxreg: To me, I think he is used to commands from other tools and is therefore always looking for a corresponding command within git.03:06
cxreg that's fine; most of the time there is just such a command03:06
it's just sometimes spelled differently03:07
→ mw joined03:07
developernotes cxreg: Agreed.03:07
→ mfilipe joined03:07
← timj left03:08
cxreg btw, this is not the first time this has come up in here03:08
cxreg greps logs03:08
← mfilipe left03:10
cxreg May 04 2009 is one03:11
→ mfilipe joined03:11
developernotes cxreg: I'm sure.03:11
cxreg http://colabti.org/irclogger/irclogger_log/git?date=2009-05-04#l260903:11
magcius_ → magcius03:12
→ scarabx joined03:13
← mfilipe left03:13
→ jackdempsey joined03:15
→ ryandotsmith joined03:16
→ justin-george joined03:16
→ offby1 joined03:18
← __iron left03:18
→ __iron joined03:19
← jackdempsey left03:21
developernotes cxreg: Thanks for the link and discussion.03:22
→ Lemao joined03:23
→ enherit joined03:23
← ryandotsmith left03:24
← developernotes left03:24
→ ryandotsmith joined03:24
→ plediii joined03:24
→ crack_head joined03:25
← spaceonline left03:25
zenwryly How can I disable the invocation of a pager (more)?03:27
→ mfilipe joined03:27
offby1 I assume there's an environment variable so you can do something like "GIT_PAGER=cat git status"03:27
JNR___ --no-pager03:27
offby1 that too03:27
← jjuran_ left03:28
→ moconnor joined03:28
zenwryly well there's core.pager but I'm not understanding how to just disable it all together03:28
I suppose I could just use cat as my pager :)03:29
→ t11m joined03:29
← mw left03:29
offby1 crude but effective03:29
← mfilipe left03:29
zenwryly offby1: what other channel do I know you from?03:30
→ mfilipe joined03:30
← t11m left03:30
→ fualo joined03:30
← Lemao left03:31
← Glutnix left03:32
← hyperair left03:33
→ aspotashev|eeepc joined03:34
offby1 points and intones: STALKER!!03:34
zenwryly hehe03:34
offby1 aah, you're not even on the list!03:34
zenwryly list?03:34
offby1 M-x erc-rank-stalkers RET :-)03:35
→ Marmouset joined03:35
offby1 http://gist.github.com/16560803:35
→ ashleyw joined03:35
zenwryly offby1: is that a real command or are you meesing with me, caue I don't have it :)03:35
← mm_202 left03:36
← ctusar left03:36
← leanucci left03:37
meanburrito920_ zenwryly: to turn of the pager, set core.pater to "", an empty string03:38
*core.parser03:38
urg03:38
zenwryly meanburrito920_: gotcha03:38
meanburrito920_ *core.pager03:38
zenwryly meanburrito920_: thanks!03:38
rock03:38
→ dduncan joined03:38
← dduncan left03:38
← markelikalderon left03:39
→ releod joined03:43
offby1 zenwryly: it's real in the sense that it exists; it's not real in the sense that it doesn't ship with emacs. Gimme five seconds03:44
← JNR___ left03:45
zenwryly offby1: found it I think03:45
erc-rank-stalkers03:45
http://paste.lisp.org/display/3573103:46
offby1 jinx :)03:46
bet that paste was from me03:46
zenwryly offby1: yours is longer03:46
err the gist one03:46
→ leanucci joined03:47
offby1 use the gist one03:47
it being newer03:47
zenwryly notes that gist seems like the best git based masturbation yet03:47
← moconnor left03:47
← Hideki left03:48
→ gasull joined03:50
meanburrito920_ agrees with zenwryly03:50
→ cedricv joined03:50
← fualo left03:53
← leanucci left03:53
← scarabx left03:54
← eletuchy left03:55
→ ryandots_ joined03:55
← mfilipe left03:56
← eMyller left03:57
← meanburrito920_ left03:57
← gasull left03:57
→ eMyller joined03:58
← tfarina left04:00
→ tedoc2000 joined04:01
← petercoulton left04:01
→ Derander joined04:01
→ markelikalderon joined04:06
alus hey. so I have some local changes04:10
uncommitted04:10
I realize it's getting out of hand, so I'd like to make a remote branch, commit and push this stuff there04:10
how do I do that?04:10
cxreg git branch new_branch; git add stuff; git commit; git push remote new_branch;04:12
offby1 or ...04:12
→ ryandot__ joined04:12
offby1 git stash; git checkout -b new_branch; git stash pop04:12
cxreg there's little point to that stash04:12
but sure, why not04:12
oh right, i suppose i forgot to checkout the new branch in my example04:13
offby1 now that you mention it04:13
alus I seem to recall something about making "git push" automatically push to the server, instead of having to specify it like that04:13
cxreg alus: it will magically push matching branches once it's on the remote04:13
but you'll need to name it the first time04:14
alus but why did you type "git push remote new_branch"04:14
cxreg because the remote doesn't know about it yet04:15
alus oh, maybe it's pulls04:15
it was extra args to git checkout, as I recall04:15
cxreg checkout -b or -t, probably04:16
alus why do I need "git branch" at all?04:16
← juan_arandaalvar left04:16
alus if git push makes the remote branch04:16
doesn't git checkout create the new local branch?04:16
cxreg checkout -b would, as given by offby1. yes.04:16
tmtowtdi04:17
→ JNR___ joined04:17
alus but offby1 didn't push to the remote server04:17
← ryandotsmith left04:17
offby1 laziness04:18
alus laziness as in not needed or as in not typed04:19
cxreg notices in the debian changelogs that gnu screen is now git hosted04:19
← mithro left04:21
→ mithro joined04:21
← porcnick left04:22
→ tweak joined04:22
tweak before I go and screw up this massive repo, just wondered if this was possible04:23
offby1 alus: as in not typed.04:23
← eMyller left04:24
tweak hard to explain tho :|04:24
cxreg tweak: ok, yes. possible.04:24
:)04:24
tweak assuming I were to check out a revision from somewhere in the past and create a new branch off that revision04:25
could I merge the 'newer' commits which occurred after the point I've branched into the new branch, and then revert the master branch back to the point where I branched?04:25
so04:25
→ girishr joined04:26
tweak basically moving a selection of commits into a new branch, but almost in a retroactive fashion04:26
cxreg sure, you don't even have to merge in order to do it04:27
tweak does that even make sense?04:27
lol04:27
cxreg the only real gotcha is that if someone else somewhere has a "master" with those commits on it, it could be tricky for you04:27
← toxx__ left04:27
tweak nah, my own private repo04:27
cxreg git branch new_branch master; git reset master <sha>; # done04:28
tweak ....surely it's not that easy04:28
cxreg er, wait, that reset is bogus syntax04:28
tweak nothing is ever that easy04:28
cxreg git update-ref master <sha>;04:28
but yes, it's that easy :)04:28
← ryandots_ left04:28
→ toxx__ joined04:28
cxreg no wait, still wrong. git update-ref refs/heads/master <sha>;04:28
tweak that sir, is 10 shades of awesome all mixed together04:29
cpg → cpg|biab04:29
cxreg (or you could just checkout master and git reset --hard <sha> and avoid the weird plumbing)04:29
tweak yeah that's what I was intending for the revert part04:29
cxreg but regardless of my inane bumbling, this is a normal thing for git04:29
tweak doing it this way without asking I would have expected git to have a spaz about how the new branch had been made off a commit that no longer existed04:30
cxreg of course the commit exists04:31
and if you create the new branch where master is, then they never even go unreferenced04:31
but if they did, they'll still stick around for a good while04:31
tweak see, why can't the rails channel be this helpful04:31
:D04:31
cxreg *cough*04:31
unlike some communities, this one is actually helpful ;)04:32
tweak amen!04:32
cxreg fwiw, dangling commits and other objects typically stick around for ~90 days or so. man git-reflog for more info on that.04:32
Gitbot fwiw: the git-reflog manpage can be found at http://git.or.cz/man/git-reflog04:32
→ sparc joined04:32
tweak awesome04:33
many thanks04:33
cxreg silly Gitbot04:33
np04:33
← Squee-D left04:33
← scientes left04:33
→ hyperair joined04:34
→ ajonat joined04:35
→ scientes joined04:38
→ porcnick joined04:39
← porcnick left04:39
← hyperair left04:42
almostautomated → almostautomated|04:42
→ pencilo joined04:42
→ fualo joined04:42
← neoeinstein left04:44
← doener left04:45
→ doener joined04:45
→ neoeinstein joined04:45
← markelikalderon left04:45
← kyxzme left04:46
→ hydrogen joined04:46
→ kyxzme joined04:47
→ daaku joined04:48
← AkumaStreak left04:48
→ fualo_ joined04:49
→ hyperair joined04:50
→ hobodave joined04:50
← kyxzme left04:50
→ tongueroo joined04:51
← JNR___ left04:54
← xelados left04:54
→ Guest62990 joined04:55
Guest62990 → jnr___05:00
cpg|biab → cpg05:02
binjured|afk → binjured05:02
← jonmasters left05:02
← hydrogen left05:03
→ Luca_Peck joined05:04
→ mw joined05:05
→ bugfux joined05:06
← tweak left05:06
← fualo left05:06
← mw left05:06
← Luca_Peck left05:10
→ seangrov` joined05:10
→ jonmasters joined05:11
← seangrove left05:12
→ ben_h joined05:12
← MetaCosm left05:16
→ ankit9 joined05:18
← ryandot__ left05:25
→ ryandotsmith joined05:25
← girishr left05:26
← spearce left05:26
→ Tuomas joined05:28
→ sitaram joined05:31
→ sprasanna joined05:32
→ plediii_ joined05:32
← plediii left05:33
→ harinath joined05:34
→ Luca_Peck joined05:38
→ ngng joined05:38
← scientes left05:38
← orafu left05:39
→ orafu joined05:40
→ mital joined05:41
→ MIR100 joined05:41
mital hi... I am facing an issue related to proxy... can I set proxy authentication05:42
→ knittl joined05:42
← hyperair left05:45
← tedoc2000 left05:45
← ngng left05:46
sitaram what is the issue? and what are you trying to do? clone or pull over HTTP? over git:// or ssh?05:48
some details would help05:48
mital I am doing "git clone url"05:48
mital is on windows05:49
← releod left05:49
mital git clone git://github.com/appcelerator/kroll.git05:49
mital is behind proxy... I am not sure weather its a proxy issue or not05:49
sitaram it will be a proxy issue05:50
mital do we have any configuration settings for proxy ??05:50
oh cool... can I set proxy authentications using git global config ?05:50
sitaram does your proxy server allow proxying arbitrary ports? (other than 443 and 80)?05:50
→ roop joined05:50
← tongueroo left05:51
mital is not sure...05:51
sitaram are you allowed to install arbitrary programs on your machine?05:51
→ digitalhobbit joined05:52
→ JasonWoof joined05:52
mital I am administrator in my comp.05:52
btw git uses which port ?05:52
sitaram what is comp?05:53
company? computer?05:53
git uses 9418 I think05:53
mital computer05:53
sitaram ok05:53
→ jchris joined05:53
sitaram then install something like corkscrew or something; there's a github help page about that if you search05:53
do you have a github account?05:53
mital yeah...05:54
← ryandotsmith left05:55
mital sitaram: I found core.gitproxy parameter in config05:55
← qrush left05:56
← wolog left05:57
sitaram if you have a github account you can (1) use the web interface to fork the kroll project (2) use your ssh credentials to grab it. This would require you to use a local proxy (I use "socat", though "corkscrew" is more popular) to proxy connections to port 443 of ssh.github.com. Any proxy server will allow port 44305:58
this is all the help I can give just now; I just tried it by the way, and got the kroll repo down to my machine behind a proxy05:58
good luck05:58
→ drizzd joined05:59
→ eMyller joined06:00
→ thiago_home joined06:02
← eMyller left06:02
→ saurabhverma joined06:02
← dreiss left06:04
→ Beket joined06:06
Beket Hi people. Is it possible for git to remember set uid bit ?06:06
remember=track06:06
→ Modius joined06:07
Nafai I think the only file permissions it tracks is executable06:07
→ sde_antrix1 joined06:08
sitaram specifically, is it executable to the effective UID of the person adding the file, I think06:08
Beket nice, thanks people06:10
→ tongueroo joined06:11
← tongueroo left06:11
Beket I will add a chmod in the Makefile target which generated the binary06:11
← fualo_ left06:12
→ fualo joined06:12
→ dreiss joined06:13
← aspotashev|eeepc left06:15
← josesito left06:23
← Luca_Peck left06:23
← MrMeek left06:23
→ gitte joined06:24
→ Luca_Peck joined06:25
← benn_d left06:26
→ hyperair joined06:27
← Beket left06:28
← bugfux left06:28
→ agile joined06:29
→ bugfux joined06:29
← digitalhobbit left06:30
← ankit9 left06:31
→ Aikawa joined06:32
→ mae joined06:33
← __iron left06:34
← crack_head left06:34
→ __iron joined06:35
→ galderz joined06:37
← thiago_home left06:37
→ fivetwentysix joined06:38
fivetwentysix how do i go revert to the previous commit?06:38
sitaram the word is reset06:39
fivetwentysix reset and?06:39
sitaram revert means a new commit that reverts the actions of the older one06:39
fivetwentysix: I'll back up a little.. what did you mean by "go revert to the previous commit"06:40
fivetwentysix well i made a mistake with my newest commit06:40
now the information is live on the production server06:40
so i need to rollback one commit06:40
git can do this right?06:40
sitaram git reset --hard HEAD^06:40
this will rollback one commit, as well as throw away all that stuff. If you want to preserve it, make a temp branch out of it before the reset06:41
git branch savethis; git reset --hard HEAD^06:41
→ wolog joined06:42
fivetwentysix ty sitaram06:42
sitaram at some later point, when you're ready to pick up that erroneous piece of work and fix it up, "git checkout savethis", work on it, fix it, "git commit --amend", and then merge it into master06:42
fivetwentysix so if i want to go back two commits its HEAD^^?06:42
sitaram yw fivetwentysix06:42
→ JNRo__ joined06:43
sitaram fivetwentysix: yes; and man git-rev-parse gives you other ways to name ancestor commits, especially when merges are involved06:43
Gitbot fivetwentysix: the git-rev-parse manpage can be found at http://git.or.cz/man/git-rev-parse06:43
← jnr___ left06:43
sitaram has a nice ASCII art picture too06:43
→ girishr joined06:44
← JNRo__ left06:46
← Bass10 left06:46
→ JNR___ joined06:46
← almostautomated| left06:49
→ ankit9 joined06:52
← JNR___ left06:52
→ JNR___ joined06:52
→ bentob0x joined06:53
← knittl left06:55
→ bdiego joined06:55
→ eletuchy joined06:55
→ [Po]lentino joined07:00
→ Yewbacca joined07:00
← JasonWoof left07:03
← ashleyw left07:05
→ ashleyw joined07:08
mital sitaram: hey it got solved I asked system engineers to open port 9418... thanks for your help :)07:08
← ashleyw left07:08
sitaram nice; you're lucky -- my guys need permission from the president of India or something to do that :)07:09
sitaram exaggerate a leeetle bit for effect :)07:09
[Po]lentino → [Po]lentino|away07:10
← __iron left07:12
→ __iron joined07:13
← JNR___ left07:14
→ JNR___ joined07:15
joshthecoder → joshthecoder_afk07:16
→ ph^ joined07:17
← bugfux left07:18
← jegade left07:18
← multani left07:18
← enherit left07:21
→ priidu joined07:22
hyperair how does one terminate a git fetch operation and then resume it later?07:24
→ shingo joined07:25
shingo → Guest3646207:25
→ unknown` joined07:26
unknown` → ryoma07:27
ciaranm hyperair: unless i've missed that being implemented recently, one doesn't07:27
ryoma hi, "up-to-date" is a modifier, not a noun. has this misspelling been corrected?07:27
hyperair it's not supported eh..?07:27
ciaranm last time i looked fetch just downloaded everything as one single big blob, and started over if that blob failed07:28
ryoma fwiw, i am using 1.6.1, and it incorrectly says up-to-date instead of up to date07:29
→ vals_ joined07:29
← fivetwentysix left07:29
← JNR___ left07:29
→ JNR___ joined07:30
→ ianmcorvidae|alt joined07:33
← ianmcorvidae left07:34
→ harinath1 joined07:36
← harinath left07:38
harinath1 → harinath07:38
← dreiss left07:38
charon ryoma: where? if it's in the docs or a porcelain command, have you considered writing a patch? if it's plumbing, odds are it's too late to change it07:40
← Aikawa left07:40
→ dpino joined07:41
→ evaryont joined07:42
evaryont is there a git channel for interesting/cool git stuff? or am I already in it? :)07:43
→ lohrun joined07:43
← tango_ left07:45
→ ntoll joined07:45
→ khmarbaise joined07:46
← mital left07:50
→ multani joined07:52
charon evaryont: this is it. you arrived at a time of day where the channel is usually fairly quiet, though07:52
→ psoo joined07:54
← ryoma left07:54
→ dwave joined07:54
← khmarbaise left07:58
← hyperair left07:59
← altrux left07:59
→ shruggar joined08:01
← Derander left08:01
← lu_zero_ left08:02
→ psankar joined08:04
charon doener: remember the huge blame-stats command? git ls-files | while read f; do git blame -M -- "$f"; done | perl -pe 's/^\^?[a-f0-9]* (?:[^(]* )?\(([^2]*?) *20.*/$1/' | sort | uniq -c | sort -n | tee stats; perl -ne '$s += $1 if /(\d+)/; END{print "$s\n"}' < stats08:06
→ NET||abuse joined08:07
charon any idea why it blames more lines to me than there are +lines in the diffstat between when i first started committing on the project and now?08:07
i can't see it :(08:07
→ koolhead1 joined08:08
koolhead1 hi all08:08
← chad left08:08
charon in fact, there are _still_ fewer lines if i count the number of + lines in 'git log -p' for the relevant lines, which makes no sense to me. i'd have expected quite the opposite (meaning that i patched the same stuff over and over)08:09
← WALoeIII left08:09
charon s/relevant lines/relevant history/08:09
→ Teo` joined08:10
← agenteo left08:16
→ Tobin joined08:18
→ mattions joined08:18
← Marmouset left08:23
→ halfline_ joined08:24
← Adlai left08:25
→ gonzojive joined08:27
← sitaram left08:28
gonzojive I have a git repository that I have cloned to two different directories and made separate modifications to each. I have committed the changes to both cloned repositories, but when I try to git-pull one into the other, I get "Already up-to-date."08:29
← ianmcorvidae|alt left08:32
→ ianmcorvidae joined08:32
← JNR___ left08:32
Ilari gonzojive: What branch are you trying to pull? Into what branch? Getting those wrong is the most common cause of "mysterious" already up-to-date messages.08:32
shruggar gonzojive, did you push the changes?08:33
gonzojive shruggar: I am pulling directly from the modified repository, not the original one.08:33
shruggar ah08:34
then yeah, what Ilari said08:34
gonzojive Ilari: how can I determine what branch I am working with?08:34
Ilari gonzojive: 'git branch' output. Or if you have it set up, its in shell prompt when you are in repo directory.08:34
gonzojive Ilari: they are both "master"08:35
Is it possible for the git repository to be "up to date" but the files are not identical to those in the repository from which I pulled?08:36
← halfline left08:37
→ jaeckel joined08:37
→ Octalot joined08:37
← bobesponja left08:39
Ilari gonzojive: "up to date" means current branch is superset of the branch to fetch + merge.08:40
voker57__ → Voker5708:40
→ prusnak joined08:41
→ HG` joined08:41
dwave → asksol08:42
→ khmarbaise joined08:45
→ hyperair joined08:47
← azeroth-work left08:49
→ kyxzme joined08:49
← Strogg left08:50
← marvil07 left08:50
→ Strogg joined08:50
← Luca_Peck left08:51
← nano- left08:52
← prusnak left08:52
→ tono joined08:53
gonzojive Ilari: ah, so my pulling repo has already merged whatever was in the other repo, and I have overwritten the changes?08:54
i must not properly understand. I just made a trivial change and committed it, and now I am still getting "up to date" when I pull08:56
shruggar "overwritten" is more destructive of a word than I would use08:56
can you pastebin a log of what you are doing?08:57
← eldragon left08:58
→ obiwan_chinobi joined08:58
→ eldragon joined08:59
→ prusnak joined09:00
→ jceb81 joined09:01
← ajonat left09:02
→ offby1` joined09:02
← fualo left09:03
← prusnak left09:05
← sunoano left09:05
Ilari gonzojive: Pulling from the another repo, right?09:06
← cpg left09:06
gonzojive Ilari: yes09:06
→ sunoano joined09:07
Ilari gonzojive: Get 'git rev-parse master' from both repos, then do 'git rev-list HEAD | grep <rev-parse-from-other-repo>' (assuming you are doing 'git pull <other-repo-path> master'.09:07
gonzojive shruggar: I have two repositories: Pulling Repo and Other Repo. I have changes in Other Repo that I want to merge into Pulling Repo. I just made a change (added a few lines to a file and committed) in Other Repo. When I pull Other Repo into Pulling Repo, it says "already up-to-date"09:08
shruggar: is there a more technical log that I can generate from git that would be helpful?09:09
Ilari gonzojive: What's the exact pull command you are using?09:09
gonzojive Ilari: red@condi:/git/webfunk$ git pull ~/projects/webfunk/ master09:10
Ilari gonzojive: Ok 'GIT_DIR=~/projects/webfunk/.git git rev-parse master' (hope I got that syntax right).09:11
gonzojive: Then in git/webfunk, 'git rev-list HEAD | grep <id-from-rev-parse>'.09:13
← krawek left09:14
→ Raydiation joined09:15
Raydiation isnt git coded in C? why does it pull in 2 perl libs then?09:15
stdorb it's a mixed bag09:16
Ilari Raydiation: Most parts are in C, some in shell and few in perl.09:16
← koolhead1 left09:19
gonzojive Ilari: git rev-parse master in ~/projects/webfunk gives only one 40-char object id: a21c730993ba8d166c8486cd946c3ad4b6fe7261. This appears in the middle of the git rev-parse in /git/webfunk09:19
← offby1 left09:20
← psankar left09:21
Ilari gonzojive: Rev-list in ~/git/webfunk you mean?09:22
← hyperair left09:22
→ hyperair joined09:22
Ilari gonzojive: 'cat ~/projects/webfunk/.git/HEAD'09:23
gonzojive Ilari: My two repositories are in the directories /git/webfunk and ~/projects/webfunk09:23
Ilari: ref: refs/heads/master is the result of that09:24
Raydiation is the git version in lenny "young" enough?09:24
Ilari Raydiation: Bit oldish.09:24
gonzojive Ilari: thanks for the help with this, by the way09:24
Raydiation are there any major changes in recent versions?09:25
Ilari gonzojive: If the ID from rev-parse appears in rev-list of another repo, it means the commits are already in there.09:25
Raydiation: Well combine 5 changelog summaries (and none of them is exactly short).09:26
→ psankar joined09:26
gonzojive Ilari: I see. thanks09:27
can anyone recommend a tool similar to Araxis Merge (gui for merging files) for linux?09:28
broonie meld09:29
→ GarethAdams joined09:29
→ Fullmoon joined09:32
dwmw2_gone → dwmw209:32
← mattions left09:34
← funkycat90210__ left09:35
← Theravadan left09:35
→ parasti joined09:35
doener charon: there are some cases where blame fails to follow some lines further back in the history, seems to happen when a) they have moved in the file and/or b) the lines around that line have been changed09:38
charon: if you recall Dan McGee's question about blaming deleted/renamed files, that was hit by such a case too09:40
charon: http://article.gmane.org/gmane.comp.version-control.git/124922/match=blame+deleted+renamed+file09:40
hm, that's better: http://article.gmane.org/gmane.comp.version-control.git/12492209:40
→ Chaser joined09:40
jkp question: i built and installed git using 'make all; make install' ut it didnt install the server components09:42
what command do i need to do this?09:42
wereHamster jkp: it should have. Does ~/bin/git-daemon exist?09:42
→ calebr joined09:42
doener git-daemon should be in libexec/git-core/09:43
jkp doener: ahhhh09:43
doener jkp: you're supposed to run it as "git daemon"09:43
jkp: the "git" wrapper knows where to look09:43
jkp doener: hmmm09:43
wereHamster oh, I thought it's one of those commands that remain in PATH09:43
jkp im getting this message in my system log when i try to clone09:44
doener jkp: bin/ should just have git, git-shell, git-receive-pack, git-upload-archive, git-upload-pack, git, gitk, git-cvsserver09:44
→ d0k joined09:44
jkp http://pastie.textmate.org/57975009:44
doener: it coule be because the daemon is running as root though09:45
maybe i should set it to run as the git user and ensure /usr/local is in the path09:45
← mcdebugger left09:45
calebr Hi, I have a commit which was pushed on remote servers long time ago. Maintainer asks me to split it into several small commits, but I red rebase is not recommended when commit was shared. how can I split this commit smoothly?09:46
doener jkp: IIRC git should add the right thing to PATH when trying to spawn a subprocess09:47
← ankit9 left09:47
_rane calebr: it's indeed not "recommended" but if no one tracks your repo it's not a big deal09:47
calebr _rane they are tracking09:47
→ halfline2 joined09:48
Chaser calebr, undo that commit using git-revert, and then push your changes by splitting etc ...09:48
that way pulls will be clean by other people and what you did is visible too ...09:49
doener calebr: more important question: has anyone that commit in his branches, i.e. did they merge your stuff, or build their own stuff upon it?09:49
← jceb81 left09:49
calebr doener: yes. maitainer just want to have small commits09:49
→ Sordran joined09:50
→ DavidKlein joined09:50
→ prusnak joined09:51
[Po]lentino|away → [Po]lentino09:52
jkp doener: indeed it does: thanks for the headsup09:53
← cedricv left09:54
jkp i went wrong because i saw that git-daemon had not been installed in bin and copied it there09:54
using it the way you suggested works great09:54
doener jkp: ah, so running git-daemon directly failed, but "git daemon" works, right?09:54
→ hal_line2 joined09:54
jkp y09:55
doener ok09:55
gonzojive broonie: thanks, meld is splendid09:56
→ lu_zero_ joined09:56
charon doener: ah, thanks. i tried blaming with -M5 -C5 -C5 -C5 as you suggested in that mail, and now the numbers are more reasonable. weird that blame -M would miss what diff (without any -M/-C) found...09:57
→ quoin joined09:58
doener charon: well, 5 as the limit might be a bit too low, finding idiotic "copies" like "int i;"09:58
charon: but the -C -C -C might make a difference, as it looks at "everything" to find copies.09:59
→ mcdebugger joined10:00
charon doener: OTOH diff shouldn't have found these copies either, and used a "creation" type of diff10:00
→ quoin- joined10:02
← halfline_ left10:02
doener charon: yep, the plain -M mode should (according to my totally unknowing assessment) always blame the same or less lines to an author than what "git log -p --author=foo | grep -c ^+" reports10:02
charon well, point is, it blames far more.10:02
(on the order of a factor of two as many)10:02
doener charon: yeah, that was meant as an agreement, saying "Your results are weird"10:03
charon ok :)10:03
doener unfortunately, I'm totally clueless about the blaming code10:04
→ chris2 joined10:05
← Q| left10:06
stdorb b0e450df (Canada 1999-06-30 1) everything10:07
← halfline2 left10:07
→ cemerick joined10:09
galderz → galderz_brb10:10
→ hal_lin_2 joined10:13
→ Q| joined10:13
← hal_line2 left10:13
← Fullmoon left10:15
← justin-george left10:16
→ joeytwiddle joined10:17
→ Fullmoon joined10:17
[Po]lentino → [Po]lentino|away10:19
→ icwiener joined10:20
← cemerick left10:22
doener charon: that git:// url from your mail gets me a "Connection refused"10:27
charon hmph10:28
← quoin left10:29
→ naeu joined10:29
charon doener: thanks for pointing that out. seems there's some iptables issue, but i'll have to defer until after lunchtime10:30
→ aliceinwire joined10:31
→ sitaram joined10:32
→ Adlai joined10:32
→ mutable joined10:33
← _Vi left10:35
→ psynaptic joined10:35
→ _phil2974 joined10:37
→ relix joined10:37
→ WinterWeaver joined10:38
→ klbate joined10:39
WinterWeaver hi all... a colleague and I are working on a project. I've been committing and pushing changes to our public repo all weekend, and he's dont the same. Yet everytime we pull, we are not getting each other's changes. Why is this?10:40
← eletuchy left10:40
David-T WinterWeaver: no idea, but i suggest you look at what's happening in your "public repo" for clues...10:43
syd_ and check your remotes..10:43
WinterWeaver unfortunately the guy that manages the public repo is not availablt at this time :(10:43
syd_: what exactly do you mean by "check your remotes" ?10:44
David-T well, i guess check that you're not pushing to different branches10:44
broonie Or different repositories.10:44
sitaram WinterWeaver: compare hashes from "git branch -v" and "git ls-remote origin" -- may help10:44
WinterWeaver thanks10:44
← khmarbaise left10:47
← ntoll left10:50
← aliceinwire left10:51
← calebr left10:51
← jinzhu left10:51
← hyperair left10:58
WinterWeaver ok, he has 2 remotes, I have only one10:58
→ paulp joined10:59
mutable Hello. Is it possible to `export' subtree to some location? I've already asked here similar question but I forgot to save log :(.10:59
paulp would anyone happen to know where github stores the info that a repo was forked at creation (thus the "fork of foo/bar" message on its main page) and whether this can be altered/deleted without recreating the repo?10:59
mutable I need to update test installation but I don't want to export whole tree11:00
→ ntoll joined11:02
wereHamster mutable: it seems as you should be able to git archive any tree, so find out which tree corresponds with your subdirectory and export that11:02
→ khmarbaise joined11:03
mutable wereHamster: yeah, that's it, thank you11:03
wereHamster or even easier: git archive HEAD:<subdirectory>11:03
← khmarbaise left11:07
→ froschi joined11:07
← _phil2974 left11:08
→ naith joined11:10
→ klbate_ joined11:11
→ cemerick joined11:15
→ gele joined11:22
→ _patton joined11:23
→ sea-gull joined11:23
← klbate left11:25
← gele left11:28
charon doener: turns out there's some switch-level filtering in place that i'll have to wait for tech support to change. you can use git://thomasrast.ch/domjudge.git in the meantime11:29
→ killerchicken_ joined11:31
killerchicken_ hey, how can I selectively add stuff to be staged for a new file? I don't want to add the entire file in the first commit.11:32
→ UUncia joined11:32
killerchicken_ My usual method of doing this (git add -p) fails me here11:32
doener killerchicken_: add -p also has (e)dit mode, where you can edit the patch in your editor11:33
killerchicken_: and in "git gui", you can add single lines to the index11:33
charon in other news, git@vger and the @ethz.ch main mailserver aren't getting along with each other any more after said tech support also saw fit to install some new version of MS exchange...11:33
→ hyperair joined11:33
[Po]lentino|away → [Po]lentino11:33
killerchicken_ doener: well, I did this:11:33
← klbate_ left11:34
killerchicken_ git add -p lib/thandy/bt_compat.py11:34
No changes.11:34
was all I got11:34
I would love to use the edit mode11:34
→ gele joined11:34
charon killerchicken_: first 'git add -N $file'11:34
killerchicken_ that's what my plan was, actually.11:34
gele → klbate11:35
charon so that it looks like an addition of all the lines in the file, instead of a creation of a file.11:35
killerchicken_ heh.11:35
WinterWeaver my colleague and I both work on the same remote repo. I have the url starting as, ssh:// he has, git+shh:// will this affect he's permissions etc?11:35
killerchicken_ nice11:35
hrm. /me forgot how to remove stuff from the staging area. Does anyone have a hint? *hides in shame*11:36
wereHamster killerchicken_: git status11:36
killerchicken_ was it rm --cached?11:36
hah11:36
thanks Weiss11:36
wereHamster*11:36
→ gwoo joined11:36
← WildPikachu left11:36
← roop left11:37
killerchicken_ awesome. Big thanks, charon and wereHamster.11:38
galderz_brb → galderz11:38
→ roop joined11:38
→ msafi joined11:40
→ ab3 joined11:41
msafi Is git an overkill for a single developer, is there a simpler versioning system?11:41
drizzd I use git solo all the time. I wouldn't trade it for any other system.11:41
phako I'd say choose whatever fits you best11:42
Chaser msafi, does nt matter if its overkill or not, no other vcs provides reliability of your repo data as git afaik11:42
zomg msafi: I'd say git is much simpler than some alternatives11:43
for one you don't need a server to use git11:43
Raydiation hi i got a folder on a webserver which id like to manage with git11:44
can i declare that folder with all existing data as git repo?11:44
msafi Anyone use git with Zend Studio I can PM?11:45
cehteh Raydiation: yes, git init it and commit all stuff11:45
Raydiation cehteh: ty11:46
cehteh Raydiation: but in case you want to push updates there, read the documentation about pushing to non bare repos .. that needs some git-fu11:46
zomg msafi: not sure if Zend Studio has anything to do with Git =) Last time I checked they didn't even provide Git integration11:46
← dodo_the_last left11:46
← ab3 left11:46
→ Fullmoon_ joined11:46
← Fullmoon left11:47
→ ab3 joined11:47
doener charon: git log -p doc/logos/DOMjudgelogo.pdf11:47
charon: vs. git blame doc/logos/DOMjudgelogo.pdf11:47
cehteh !non-bare11:47
← Fullmoon_ left11:47
doener faq non-bare it is11:47
Gitbot doener: Pushing to non-bare repositories is discouraged. See http://git.or.cz/gitwiki/GitFaq#non-bare11:47
cehteh .. i never understand how this bots work :)11:47
but nice to have doener-as-bot-proxy :)11:47
→ dodo_the_last joined11:48
msafi zomg, but there got to be a way to integrate it because Zend Studio is basically an Eclipse IDE11:49
→ knittl joined11:49
zomg msafi: you can probably just get an Eclipse Git plugin11:49
if you use the Eclipse version of ZS11:50
charon doener: and same for other binaries like bin/sh-static, apparently. *sigh311:50
Raydiation do i have to do that on my local pc or on the server: git config --global user.name "Your Name Comes Here"11:52
charon doener: and of course the upstream has the same binaries under other names, so with enough -C it finds that instead11:53
charon bangs his head against the table11:53
→ tiny joined11:54
doener at least we got it sorted out11:54
Raydiation this tutorial is kinda hard to understand if youre drunk11:54
stdorb then, don't be drunk11:55
← naeu left11:57
← hyperair left11:57
→ Hendu joined11:57
→ klbate_ joined11:57
→ wu-lee joined11:58
→ Jinzhu joined11:59
→ hyperair joined11:59
← HG` left12:00
msafi Since Zend Studio comes with built-in support for SVN, I'll just get started with that. Thanks for the advice anyway, bye!12:01
← msafi left12:01
Guest36462 → Aikawa12:01
→ HG` joined12:01
← paulp left12:03
→ fpcfan joined12:03
tiny Any estimates on how long will "fsck --strict --full" take on a current linux kernel tree?12:06
zomg Depends on how fast your PC is ;>12:06
(no idea)12:06
(I don't even know what that does)12:06
→ iron joined12:07
tiny Well I have a broken repo so I thought I ran fsck on it first12:07
→ qrush joined12:08
shruggar what are the symptoms of brokenness?12:08
tiny and it's doing it for about 10 minutes taking all of CPU time12:08
← ben_h left12:09
→ qrush__ joined12:09
tiny shruggar: git gc: error: Could not read 3c2e4b82ffa4827e65c4e7d5a2e6d2bc299c4257, fatal: bad tree object 3c2e4b82ffa4827e65c4e7d5a2e6d2bc299c4257, error: failed to run repack12:09
→ ben_h joined12:09
doener tiny: which git version?12:09
tiny 1.6.2.412:10
shruggar have you checked the specific object?12:10
(ie: it could be just that one, though fsck is probably a good idea anyway)12:10
tiny shruggar: not yet. Let me describe what I wan't to achieve.12:10
_patton If I have a repository which pulls say the dev branch from another bare repository, will the branch name stay as dev once I've pulled it or will it automatically be renamed master? For me it seems that it is being renamed master...12:11
→ markelikalderon joined12:11
tiny I have a local repo from which now broken repo was cloned. I lost data on my repo and backup was old.12:11
New data is in the broken repo from which I can't "push" anymore.12:11
shruggar _patton, what did you do to set up the remote tracking branch, and what are you doing to pull?12:11
tiny I would like to keep commits12:12
but first I need to fix it if possible12:12
shruggar what commands are working? can you dump a backup of the commits you want to keep with format-patch?12:13
→ gwoo_ joined12:13
drizzd _patton: if you do something like "git checkout -b master origin/dev", then "git checkout master; git pull" will pull from origin/dev. That does not mean that dev was renamed. It just means that master is tracking origin/dev.12:14
← klbate left12:14
doener tiny: real 6m3.986s12:14
tiny: that's how long it took here12:14
(the fsck)12:14
tiny shruggar: how would I do that? I work on a separate branch. I just pulled a current linux kernel. But I'm afraid merge will fail.12:15
_patton shruggar: I have set up a bare repository to act as a hub, then there is a post update hook which changes directory to the tracking repository and does a git pull hub dev12:15
tiny doener: I just aborted fsck.12:15
→ eletuchy joined12:15
tiny doener: with --full and --strict options?12:15
doener yep12:15
_patton drizzd: can I do a git pull on the tracking repository and pull in say all 3 branches and have them stay intact on the tracking repository. I tried doing this and the files from all three branches seemed to merge and create conflicts...12:16
← Sordran left12:16
← pencilo left12:16
→ pencilo joined12:17
drizzd _patton: you're doing it wrong. git pull is used to merge stuff and can only be used interactively.12:17
if you simply want to propagate changes (without merging), use "git remote update"12:18
← gwoo left12:18
← froschi left12:18
← mithro left12:18
_patton drizzd: really, ok. I was using git pull in the post update hook. basically I want to use a hub with 3 branches for dev/ staging/ master/ and then pull these branches down into the separate folders for each version of the website12:19
tiny drizzd: yes but then: "git diff mybranch..master" reports all kind of stuff.12:19
← [Po]lentino left12:21
drizzd _patton: ok. You still cannot use git pull for that. If you really need a "hub", set up a mirror repository (see the --mirror options to clone or remote) and use git reset --hard to update your branches (note that will remove local changes, even committed ones).12:21
tiny: are you talking to me?12:21
← qrush left12:22
→ qrush joined12:22
← __iron left12:22
tiny drizzd: sorry I thought you were referring to my problem with that propagation changes12:22
drizzd tiny: nope, I didn't pay attention to you, sorry :-)12:23
_patton drizzd: cheers, I'll give that a shot12:23
tiny :)12:23
drizzd _patton: there should be example hooks for things like that out there12:23
← Yewbacca left12:24
drizzd _patton: http://git.or.cz/gitwiki/GitFaq#Whywon.27tIseechangesintheremoterepoafter.22gitpush.22.3F12:24
vals_ → tango_12:25
→ cloudhead joined12:25
← neoeinstein left12:25
_patton drizzd: thanks, the examples I found used a bare "hub" repo, then the files on the web server had a git repository which were updated with a post-update hook on the bare repo which did a pull from the bare repo to update the files on the web server.12:25
drizzd: there was also a post-commit hook on the files on the server so that when changes were made directly on the server they were pushed to the bare hub repository12:26
→ spaceonline joined12:27
→ neoeinstein joined12:27
drizzd _patton: it works if the upstream branches never change history, and if downstream branches do not change at all.12:29
(because then git pull always results in a fast-forward merge)12:29
dsop Pieter: will there be a signed off option in gitx soon?12:30
← qrush left12:30
_patton drizzd: ok, cheers. I'm going to give it one last try, then I might just try using github and manually pulling down branches from the repo stored on github to each folder or something12:33
Pieter dsop: yes, it's already in the nightly12:33
← qrush__ left12:34
← TDJACR left12:37
→ kumbayo joined12:38
→ mithro joined12:40
gwoo_ → gwoo12:41
→ TDJACR joined12:41
→ mw joined12:43
← wu-lee left12:43
← xenoterracide left12:44
→ xenoterracide joined12:45
← xenoterracide left12:45
→ xenoterracide joined12:46
← Tobin left12:47
→ Tobin joined12:48
← Tobin left12:48
→ cmarcelo joined12:48
← girishr left12:49
→ wu-lee joined12:50
→ lorandi joined12:51
← mcdebugger left12:51
← ia left12:51
→ aziz joined12:53
→ jrmuizel joined12:54
← mhoefli left12:54
→ piero joined12:55
piero hi everyone! I need to embed the git tag inside my C code. How can I do that?12:56
charon piero: man git-describe; and perhaps look at <HEAD:GIT-VERSION-GEN>12:56
piero e.g. If my tag is VERSION_1_0_0, there should be a string "VERSION_1_0_0" created at build-time12:56
Gitbot [git GIT-VERSION-GEN]: http://tinyurl.com/lowbne [blob]12:56
piero: the git-describe manpage can be found at http://git.or.cz/man/git-describe12:56
→ Niamor joined12:57
spb the standard approach here is to have your C code include a header named something like version.h, which is generated by a shell script at build time12:57
piero thank you :-) I'll have a look!12:57
→ cilly joined12:59
→ ctusar joined13:02
→ boto joined13:02
→ MrRio joined13:04
Ilari And then maybe have a file that is used as default version if describe fails.13:04
piero I looked at the git-describe documentation. I couldn't find any way to grab the tag name directly from C code13:05
Ilari piero: Have makefile do it.13:06
← knittl left13:06
piero Ilari: thank you. I was investigating this solution, but I wanted to be sure that there's nothing similar to what happens e.g. in SVN, where you can get the version directly from your code13:07
but the makefile solution is ok, thank you again13:07
→ ia joined13:08
← agib left13:09
Ilari piero: There are keywords, but the only that's not export-time-expanded is $id$, which is _BLOB_ ID. And there are many export-time-expanded ones (commit IDs, timestamps, names, etc) but sadly (unless this has recently changed), no keyword for describe output.13:10
piero: Basically, any checkout-expanded keyword that depends on anything else than the checked out file itself would be disasterous to performance.13:12
MrRio hey, i'm using svn at the moment for web development, i'm switching to git, i've imported my svn repo into a git repo on my staging server, and my production server is a working copy checkout of my svn repo, can i remove the .svn files and make it into a live git repo without deleting everything and starting again?13:12
piero Ilari: yes I understand the problem13:12
Ilari MrRio: Why didn't you use git svn for the import?13:12
MrRio Ilari: I did, I'm just wondering how I deploy this to my production machine13:13
Ilari MrRio: There should not be .svn directories then...13:13
MrRio: Any branches? Tags? Non-standard layout? Merges?13:14
← sprasanna left13:14
MrRio Ilari: The import is on my staging machine13:14
parasti your svn checkout is pretty much useless, you need to clone the converted repo13:14
→ sprasanna joined13:15
MrRio parasti: Can I clone the converted repo over the top of the live website?13:15
parasti: trying to avoid downtime, i'll experiment with it in another directory before I do it to the live website of course13:15
Ilari MrRio: Hopefully you didn't do anything really stupid in import like not specifying layout when one should have...13:16
MrRio Ilari, standard layout, trunk, tags, branches, told the import about all that13:16
Ilari MrRio: Any merges?13:16
MrRio Ilari, yes, but pre svn 1.4, so not tracked merges as far as i understand13:17
← piero left13:18
Ilari MrRio: One could make them tracked at import time. And if the previous merge from A to B was unmarked merge and you try to merge further development of A to B, you are SOL.13:18
→ mm_202 joined13:19
Ilari MrRio: And oh, the absolutely worst thing in imports: Layout changes...13:20
MrRio Ilari: yeah, I imagine that's quite painful13:20
→ benn_d joined13:20
MrRio Ilari: I'm half tempted to leave the history for the project where it is, and make a fresh git repo from the live site13:21
→ hydrogen joined13:21
Ilari MrRio: The basic idea of marking merges is: Find merge commits, find the merged tips and mark merged tips as parents of merge commit using grafts. Then use filter-branch to make it permanent.13:22
→ mw_ joined13:22
← quoin- left13:22
← mw left13:23
Ilari MrRio: And the nastiest part is: Finding tip of what was merged in.13:24
MrRio Ilari: Hmm, yeah13:24
Ilari: I may just start afresh in git13:24
fpcfan Ilari: what is the advantage of finding the merges?13:25
Ilari Gives proper history "as it happened" and merging further development works right.13:26
MrRio: If you start afresh by reimporting you can only bring one branch (if you bring more, try to merge them back and you are SOL)13:26
MrRio: Actually, only one version of one branch.13:27
MrRio Ilari: Yeah that's fine, this project is currently at a maintenance level of development, with some major changes coming in the next few months13:27
→ fargiola` joined13:27
Ilari MrRio: Maybe archive the git-svn import in case it proves useful.13:28
MrRio Ilari: good plan13:28
→ Bass10 joined13:28
→ Ryback_ joined13:28
← wolog left13:30
→ wolog_ joined13:30
Ilari Then there is SVN braindamage like "partial branches" that git svn might not be able to handle properly at all.13:31
← danl left13:31
← hyperair left13:32
← wolog_ left13:32
→ wolog__ joined13:32
← fargiolas left13:32
fargiola` → fargiolas13:32
→ hyperair joined13:33
← mw_ left13:33
hal_lin_2 → halfline13:35
→ khmarbaise joined13:35
→ mnick joined13:35
← agile left13:36
Raydiation can i extend the information of registered users?13:36
add a jabber adress for instance13:36
plus can i share this information across applications?13:37
→ bremner_ joined13:38
← NET||abuse left13:39
Ilari Raydiation: Err? What that has to do with git?13:39
← obiwan_chinobi left13:39
Raydiation Ilari: ou wrong channel :)13:39
tiny Hello ... having problems with git repo while optimizing. Can't push either. http://pastebin.com/d49f7f43113:39
← Raydiation left13:40
← bremner_ left13:40
→ rb____ joined13:41
Ilari tiny: Does 'git rev-list --all --objects' complain?13:41
tiny Ilari: yes13:41
Ilari: error: Could not read 3c2e4b82ffa4827e65c4e7d5a2e6d2bc299c425713:42
fatal: bad tree object 3c2e4b82ffa4827e65c4e7d5a2e6d2bc299c425713:42
→ AkumaStreak joined13:42
← shd_ left13:42
← khmarbaise left13:42
← hydrogen left13:43
tiny If there's a way to fix this I'd like to do it. If not I would like to start with a fresh pull and apply a patch of my branch. If so ... is it possible to keep commit history?13:43
Ilari tiny: Does 'git rev-list --all --objects | grep "3c2e4b82ffa4827e65c4e7d5a2e6d2bc299c4257"' yield anything?13:43
tiny Ilari: yes the above two lines ...13:44
← psynaptic left13:45
→ shd joined13:45
Ilari tiny: Well, if that object is part of upstream history, its possible to fix the repo. If not, the local changesets might not be recoverable...13:45
tiny: Maybe run 'git fsck --full' to see if there's more than that one thing wrong.13:46
→ Adlai` joined13:46
← Adlai left13:47
tiny Ilari: ok ... this will take a while13:47
shruggar tiny: as I said, use git format-patch to make a patch which keeps commit history13:47
tiny shruggar: thank you13:47
Ilari tiny: The ability to preserve commit history would imply ability to fix the repo.13:47
→ froschi joined13:47
shruggar depends on how much is broken13:47
→ MisterN joined13:47
→ ryandotsmith joined13:47
tiny exactly13:47
→ [Po]lentino joined13:50
← dodo_the_last left13:52
→ hugoxrosa joined13:52
→ cannonball joined13:52
Adlai` → Adlai13:53
→ zmbmartin joined13:55
→ AkumaStreak_ joined13:55
offby1` → offby113:55
← klbate_ left13:56
← psankar left13:56
→ drizzd_ joined13:56
zmbmartin If using git for websites that I am working on is it better to ignore image files?13:56
← cilly left13:57
killerchicken_ zmbmartin: imho, nope. It doesn't really make sense not to version them13:57
→ broheem joined13:57
hobo many sites keep media on a separate server, so they'd be under their own repository13:57
→ bcardarella joined13:58
dirker hi all, how can I get a list of commits no branch or tag is pointing to currently13:58
commits lost in space so to say :-)13:58
zmbmartin ok I have been so far but the sites I have under git right now are fairly small. thanks killerchicken_13:58
killerchicken_ zmbmartin: I'd create a .gitattributes entry13:59
and then not worry about them13:59
→ cilly joined13:59
→ RaceCondition joined13:59
parasti dirker: man git-fsck13:59
Gitbot dirker: the git-fsck manpage can be found at http://git.or.cz/man/git-fsck13:59
dirker parasti: thanks14:00
RaceCondition I'm trying to decide whether I should set pull --rebase as the default setting for all developers... might it cause issues that won't occur with regular pull?14:00
i.e. something that other devs might have to deal with that they wouldn't with pull --no-rebase?14:00
hobo 10:00 < dirker> parasti: thanks14:00
10:00 < RaceCondition> I'm trying to decide whether I should set pull --rebase as the default setting for all developers... might it cause issues that won't occur with regular pull?14:00
10:00 < RaceCondition> i.e. something that other devs might have to deal with that they wouldn't with pull --no-rebase?14:01
→ coachz joined14:01
hobo Crap sorry about that, I hate putty >_>14:01
and windows for that matter14:01
coachz putty rocks :-)14:01
→ intripoon joined14:02
Grum how would you even set --rebase as the default setting?14:02
ugh google power -- git config branch.autosetuprebase always14:02
zmbmartin killerchicken_: I will have to read up on .gitattributes, as I am not familiar with it yet. Thanks14:02
→ name joined14:02
→ spearce joined14:03
sitaram man git-config and look for branch.autosetuprebase14:03
Gitbot sitaram: the git-config manpage can be found at http://git.or.cz/man/git-config14:03
→ sprasanna_ joined14:03
← ctusar left14:04
→ duncanmv joined14:05
← drizzd left14:06
duncanmv guys, I have repoA.git/ with some files and history. I have repoB.git with some files. Is it possible to take all repoA history and put it into repoB/subdir directory?14:06
wereHamster duncanmv: google 'git subtree merge'14:06
→ khmarbaise joined14:06
← roop left14:06
duncanmv wereHamster: thanks, I will look :-)14:07
→ offby1` joined14:07
wereHamster duncanmv: the other possibility is to use git submodules14:07
← Octalot left14:08
sitaram duncanmv: is this onetime, or do you expect to keep making changes to repoA both in the original place *and* in the repoB/subdir14:09
if its onetime, dont bother with submodules14:09
duncanmv sitaram: no, we want to make a one time import and delete repoA14:09
← stuffcor1se left14:09
sitaram duncanmv: yeah then submodules dont buy you anything; you need subtree merge14:09
duncanmv I would never bother with submodules, they are a broken feature IMO14:09
← cilly left14:09
→ stuffcorpse joined14:10
→ cilly joined14:10
sitaram duncanmv: http://progit.org/book/ch6-7.html is nice14:10
sitaram agrees about submodules -- could never understand them anyway :)14:10
← AkumaStreak left14:10
sitaram duncanmv: you'll like the first line of that page :)14:10
RaceCondition can I tell pull --rebase to fall back to --no-rebase/merge if any problems would occur that wouldn't occur with merge?14:11
or is this the default behavior even?14:11
→ loincloth joined14:12
← sprasanna left14:12
← ph^ left14:13
→ crack_head joined14:14
→ ctusar_ joined14:15
offby1` no idea14:15
→ bugfux joined14:16
zmbmartin I am trying to learn git but am having a problem14:16
I git rm a file but I only wanted it ignored14:17
I added it to the ignore list but now it wont show in the branch that I want it to14:17
offby1` what do you mean by "won't show"?14:18
← offby1 left14:18
offby1` → offby114:18
zmbmartin I have a settings file that changes depending on the comp so I did want it tracked14:18
problem is I made a branch and removed it from the branch so it shows in the master but not in the branch14:19
→ doM^ joined14:19
doM^ hi14:19
→ almostautomated joined14:19
offby1 zmbmartin: so add it back to the branch.14:19
zmbmartin I did add it back but when I switch branches it goes away again when I go back to that branch14:19
doM^ I was searching a way to only view the file under the section 'Changed but not updated' in git status14:20
zmbmartin I want the file in there but I want it ignored14:20
wereHamster zmbmartin: if a file is in one branch but not the other, git will delete it when switching from the first branch to the second14:21
coachz does .gitignore effect all branches or just the current one ?14:21
→ Octalot joined14:21
wereHamster coachz: just the current one, .git/info/excludes is global for the whole repository14:21
coachz nice14:21
→ agile joined14:22
← tiny left14:22
zmbmartin wereHamster: that is the problem I am having. so if I have it on the gitignore it will still track the file just not the changes?14:22
trapni^3 → trapni14:22
wereHamster zmbmartin: no, gitignore simply causes git to ignore the files (not show in git status, ignore in git add -a ietc)14:23
→ petercoulton joined14:23
wereHamster also, gitignore has no effect on already tracked files14:23
→ tedoc2000 joined14:24
wereHamster maybe you want git update-index --assume-unchanged?14:24
→ agib joined14:25
neoeinstein If it's already in the index, it will still track it, even if it's in .gitignore.14:26
You can use a fun hack via git update-index --assume-unchanged to ignore local modifications of files in your index.14:26
But it's a brittle hack at times.14:26
→ Turias joined14:27
← krisfremen left14:28
zmbmartin ok I think I am starting to screw things up I just got this --> ! [rejected] master -> master (non-fast forward)14:28
wereHamster zmbmartin: faq non-ff14:28
Gitbot zmbmartin: Your push would lose changes on the remote. See http://git.or.cz/gitwiki/GitFaq#non-ff14:28
wereHamster zmbmartin: did you use git rebase or git commit --amend?14:28
← cilly left14:29
→ glauber_ joined14:29
→ cilly joined14:32
→ ericindc joined14:32
→ mw44118 joined14:33
← ryandotsmith left14:34
mw44118 How do I see what bar.py looks like in branch foo?14:34
Adlai → Adlai-AWAY14:34
zmbmartin so if I was tracking the settings.py file but no longer want that tracked how do I had git honor when I add it to gitignore but not remove it from the directory14:34
wereHamster mw44118: git show foo:bar.py14:34
PerlJam git show foo:bar.py14:34
← aatanasov left14:34
← hyperair left14:35
zmbmartin I added the settings.py to the ignore file but still when I make changes to it git status is showing it as modified14:35
wereHamster zmbmartin: you have to tell git to stop tracking it, otherwise gitignore has no effect: git rm --cached settings.py; git commit;14:35
→ hyperair joined14:35
zmbmartin so that is what I did wrong before, the --cached flag.14:36
wereHamster git rm --cached removes the file from the index, but leaves it in the working tree14:37
mw44118 thanks!14:38
← broheem left14:39
zmbmartin Thanks very much wereHamster14:39
wwalker is there a way to tell git to always ignore local changes to a file? such as config/db_data.conf . I want to keep that file in the git master, but never want to commit, let alone push, changes to it.14:40
stepnem Is it possible to know if there are any remote changes to pull and act upon it? I'm currently doing 'git pull && git whatchanged -p', but I'd like to abort the command if there are no new commits sice the last pull14:41
wwalker: if I understand you correctly, you don't in fact want to version control the file? Then just use the ignore facilities14:43
wwalker stepnem: no, I want to keep that file in the git master14:44
stepnem wwalker: what do you mean?14:44
wwalker but most devs' local changes should never be committed.14:44
stepnem if you're never going to commit it, what's the point of tracking it?14:44
wwalker stepnem: the file has 30 config values in it. 4 of them are changed to make a dev's local system tests work.14:45
Grum so you mean: config/db_data.conf.sample and everyone having their own version?14:45
← khmarbaise left14:45
→ tiny joined14:45
Grum and echo "config/db_data.conf" >> .gitignore14:45
wwalker Grum: that is one way to handle it and works for some cases. There are also times when one changes a local file for expediency (like to skip 4300 slow tests in the test suite) but don't ever want those local changes commit'd / push'd14:46
Grum then dont commit them? :D14:47
← crack_head left14:47
wwalker Grum: I use that a lot (default behavior in a rails app).14:47
← prusnak left14:47
wwalker It's a royal pain when there are 300 files changed to not be able to say commit -a14:47
tiny Hi again. Is there a way to tell git to accept patches that will not get upstream without conflicts? I'm working on a separate branch and there are conflicts if I edit files that are part of official repo.14:47
Grum blame the framework then for changing 300 files per tiny commit :(14:48
wwalker because of 1 or 2 files with local, non committable changes.14:48
Grum: back to my question. can it be done? I'm slowly working to change the source base, but can't do that immediately (it's currently svn with 80 very active externals...<shiver>)14:49
Grum you want git add . to skip files that are tracked?14:50
and it has to guess which ones by itself?14:50
← tedoc2000 left14:50
wwalker Grum: I'm asking if I can tell it which ones to skip, without removing them from the master on my next push. .gitignore only controls new files14:51
← hobodave left14:51
Grum .gitignore works on untracked files14:51
and well, you can just 'unadd them' :)14:51
Ilari tiny: Of course, it's good idea to take a backup of corrupt repo before trying to do more than just take a peek at what's corrupted.14:51
wwalker Grum: if I "unadd" them will that affect the master in any way on my next push?14:51
Grum you can just manually unstage the items so they do not get recorded in the commit you are making14:52
wwalker is "unadd" == "git rm --cached config/with_local_only_changes"14:53
Grum so yes, if they are not in the commit they will not be pushed14:53
nonono14:53
unadd would be git reset HEAD <file>14:53
tiny Ilari: did backup, I have the history of changes (patches), started fresh with git clone, applying patches now ... but I hate the conflicts.14:53
wwalker Grum: ok,...reads man page14:53
Grum git rm --cached will remove the file from git but leave it on disk14:54
→ tfarina joined14:54
Grum ea it alters the staging area not to 'track' the file anymore14:54
→ crack_head joined14:54
Grum git reset HEAD <file> just tells the staging area that for the next commit you should take the file as it is in HEAD14:55
wwalker Grum: so I would run "git reset --mixed config/with_local_only_changes" before every commit?14:55
Grum no14:55
wwalker Grum: so I would run "git reset --mixed HEAD config/with_local_only_changes" before every commit?14:55
Grum you run get reset HEAD <file>14:55
*git -- you dont have to use --mixed14:55
wwalker man page sounds like if I run "git reset HEAD" it will change the working tree file.14:56
→ naeu joined14:56
Grum no, git checkout HEAD <file> does that14:56
from git status:14:57
# (use "git reset HEAD <file>..." to unstage)14:57
# (use "git checkout -- <file>..." to discard changes in working directory)14:57
← sde_antrix1 left14:58
← cilly left14:59
→ usomaki joined14:59
stepnem any idea on my question above? it would be nice if git-pull returned different exit status depending on whether it actually pulled something or not; or is there some other way short of parsing the text output?14:59
wwalker Grum: thanks. my git status doesn't output anything like that. what options did you use?14:59
← usomaki left14:59
← kpreid_ left14:59
Grum it does O.o15:00
→ kpreid_ joined15:00
→ tedoc2000 joined15:00
Grum mkdir test; cd test; git init; echo a > x; git add .; git commit -m"first"; echo aa >> x; git status; git add .; git status15:00
→ cilly joined15:01
← AkumaStreak_ left15:02
wwalker Grum: I was in a simple local working tree without any modified files (the tree in question is on the desktop at work) I see it now.15:02
← DrNick left15:03
→ RobertMeta joined15:04
← tim_jenssen left15:04
Grum i have to say, if 300 files change in a single commit something is odd15:04
duncanmv sitaram: uhm, works nice in a subdir, however I get errors on subtree merge in a prefix like subdir1/subdir215:04
mm_202 is there a way to copy a file from a remote branch, but without checking it out first?15:05
Grum mm_202: try git archive perhaps?15:05
wwalker Grum: the code base is 1.5GB I suspect it can be shrunk by 3X. That refactoring will often require changes across multiple files. 300, probably very seldom. More than I want to type on the command line, very often.15:06
mm_202 Grum: okay, I'll check it out. Thanks!15:06
Grum wwalker: what on earth is 1.5gb worth of code? O.o15:06
you making google orso?15:07
← cilly left15:07
wwalker but it sounds like I can add a pre-commit and post-commit hook scripts to git reset the few files that should always be ignored.15:07
Grum or just unadd them manually?15:07
or fix your setup by making the config files actually not tracked ?15:08
or by providing ways of pointing at different configs conveniently?15:08
wwalker think like an enterprisy wireshark. it's not really like that, but similar. includes everything from kernel drivers to java report graphers.15:08
like I said Grum, LOTS of refactoring on the horizon....15:08
duncanmv anyone knows how to do subtree merge with a two level prefix?15:09
wwalker Grum: config files was a bad example. lots of these will be local modifications to test suites.15:09
Grum wwalker: what kind of project is it?15:09
wwalker one with an NDA :(15:10
Grum 1.5gb worth of code? O.o15:10
→ cilly joined15:10
← cilly left15:11
wwalker Grum: makes me dizzy too.15:11
sitaram duncanmv: sorry just came back "to life"; what sort of errors (I admit I haven't actually tried it or used it on a real case; just hapened to read that chapter today)15:11
Grum wwalker: its bigger than the biggest thing i've seen sofar :D15:12
duncanmv sitaram: solved, I found out that prefix needs to be empty ;-)15:12
sitaram aah!15:12
wwalker Grum: biggest project source tree I've worked on too. and I've been in the game for 20+15:13
← jhelwig left15:13
Grum wwalker: and its all RoR ? O.o15:13
→ Sho_ joined15:14
→ DrNick joined15:14
wwalker no. python, C, C++, Java, ruby, and RoR15:14
→ cilly joined15:14
sitaram didnt know RoR has been around long enough to collect COBOL-size software suites :)15:14
wwalker that I know of so far.15:14
sitaram aah15:14
Grum ehhe15:14
and its all under a NDA?15:14
meaning its not even connected to the internet? or 'yet to be displayed' ? :D15:15
wwalker RoR is slowly replacing the existing UI. Yup.15:15
→ ryandotsmith joined15:15
→ coldwaterlover joined15:15
→ malumalu joined15:16
Grum wwalker: without going in specifics, is it public somewhere? or supposed to become public?15:16
wwalker It's out there but internals of a company are NDA'd even though the product is available. I try to never attach one of my client's names to the problem I'm solving for them15:16
→ altrux joined15:17
mm_202 → mm_work15:18
Grum ah i guess that is nice :D15:18
← name left15:19
wwalker requirement for staying in business long term. even clients that "fire" me (as a consultant, officially the consulting simply stops) know that their internal dirty laundry won't be told by me.15:20
well, not associated with them.15:20
jast I think that's the only decent way to handle things.15:21
Grum indeed15:21
sitaram its the only legal way too, often :)15:21
but it helps if you have a repotation for it, without requiring legalities!15:21
reputation*15:21
→ WebDragon joined15:22
Grum so next time start with the: 'zomg i have 1.5gb sourcecode -- companyX product Y' and then just reconnect with another nick and appear like nothing happened!15:22
wwalker Grum: :) I've found that our community is too small to hide for long.15:23
Grum :p15:23
we'll play dumb as well!15:23
wwalker sitaram: yes, I quite often start on a project just by agreeing to sign a reasonable NDA later and pre-date the effective date.15:24
← malumalu left15:24
sitaram wwalker: that means you already have a reputation for doing the right thing, because for that (short) duration they're exposed... good for you!15:24
wwalker Grum: thanks for the help15:24
→ rudi_s joined15:25
Grum yw, good luck :D15:26
→ jhelwig joined15:26
← wwalker left15:27
← Adlai-AWAY left15:28
→ Adlai-AWAY joined15:28
→ Lemao joined15:29
← coldwaterlover left15:29
→ stringo0 joined15:30
→ Adlai-AW` joined15:31
← HG` left15:31
← Akufen left15:32
← Adlai-AWAY left15:32
← doM^ left15:33
→ hobodave joined15:37
← Adlai-AW` left15:41
→ Adlai-AW` joined15:41
→ summatusmentis joined15:43
→ eMyller joined15:43
summatusmentis I screwed up, I was trying to update a dev branch, and ended up commiting changes that had already been made to the master repo. In trying to fix this, I deleted my branch I'd been doing dev work on, is there a way to get this back?15:44
← UUncia left15:45
← Lemao left15:45
broonie summatusmentis: git reflog may be of assistance15:45
sitaram summatusmentis: you deleted a branch? well the branch reflog gets deleted along with the branch :( but the HEAD reflog may still show you what you want... try "git reflog show HEAD -10"15:47
(the -10 is to limit to the top 10 entried; adjust to taste)15:47
→ fernandor joined15:47
joshthecoder_afk → joshthecoder15:48
shruggar sitaram, isn't there a way to list unreachable commits?15:48
sitaram shruggar: sure but there would be far too many. "git reflog --date=relative HEAD" is far better -- the relative timestamps (like "20 minutes ago") and the text accompanying the reflog give you a solid clue what is where15:49
← summatusmentis left15:49
→ summatusmentis joined15:49
summatusmentis I screwed up, I was trying to update a dev branch, and ended up commiting changes that had already been made to the master repo. In trying to fix this, I deleted my branch I'd been doing dev work on, is there a way to get this back?15:49
sitaram shruggar: in contrast, nothing distinguishes an unreachable commit made 2 months ago from one made 2 seconds ago15:49
← loincloth left15:49
← _patton left15:50
shruggar I suspect that a union of "all unreachable commits" and "the last howevermany reflog entries" would be the right answer...15:50
union? intersection!15:50
→ selckin joined15:50
sitaram shruggar: actually I take back what I said. Unless the reflog expired, you wont even see it in unreachable. Intersection is null :)15:50
shruggar dies15:51
sitaram shruggar: unless you specified one or more heads in the "git fsck --unreachable" command, then it finds stuff unreachable from there -- in which case of course even a named branch can show up as unreachable15:53
← ab3 left15:54
shruggar git fsck --unreachable $(git show-ref -s) ?15:54
sitaram or you say "git fsck --no-reflogs" (facepalm -- how did I forget that one!)15:54
shruggar ah, there you go then15:55
← priidu left15:55
shruggar so much dangling, so much I'm afraid to touch15:55
sitaram shruggar: but really, all this is moot in his scenario. You know the branch name, except you deleted it, so the best way is the reflog. Unless you deleted it more than 90 days ago, or you explicitly expired reflogs15:56
try it... delete a branch and then try "git reflog show --date=relative HEAD | grep your_branch_name"15:56
← zmbmartin left15:57
→ coldwaterlover joined15:57
← ryandotsmith left15:57
shruggar I suppose I should be more trusting of the reflog's storing of branch names.. I never am because it does so in plain english15:58
→ klbate joined15:58
sitaram :)15:58
dont worry15:58
← coldwaterlover left15:58
sitaram shruggar: besides, when fsck finds you 20 candidate commits (and God help you if you did a large rebase somewhere!) how would you determine which is the correct one? Just pretend reflog is giving you a head start on that exercise :)15:59
shruggar I think I just want the before and after ref names to be part of the "real" log instead of just the message (they aren't at all for commits and rebases, for example, just skimming right now)15:59
→ coldwaterlover joined15:59
sitaram shruggar: what's the "real" log?15:59
the commit log?15:59
shruggar I don't know a good name for it. The non-english part15:59
→ jackdempsey joined16:00
doener shruggar: rebases happen on a detached HEAD, only when it finishes is the branch head updated. And IIRC that one says which ref was affected16:00
← [Po]lentino left16:00
shruggar the part that looks like <sha> <sha> <whodunnit> <when> <and-I-don't-mean-the-description>16:00
← Adlai-AW` left16:00
→ Adlai-AW` joined16:00
sitaram doener: you mean a complex rebase will not flood reflog? cool!16:01
doener sitaram: HEAD's will be flooded, but not the ref's16:01
sitaram shruggar: the description is in English but not human generated, so you can trust it16:01
shruggar doener, doesn't seem to, unless it's the wonderful "updating HEAD" message that seems to be repeated 47 times here16:01
sitaram doener: in this case we're looking at someone who deleted a branch accidentally and is trying to recover it. shruggar was considering "git fsck --unreachable" and I said that will give you lots of junk16:02
sitaram loves reflog...16:02
shruggar What I tend to want from the reflog is: 1) where was I, 2) where am I, 3) what did I do to get there. And while all that information /tends/ to be there, none of it seems to /always/ be there16:03
doener shruggar: hm, indeed, there's no "checkout" message at the end (which I thought there was)16:03
summatusmentis is there a way to recover a deleted branch?16:04
← psoo left16:04
shruggar that's what we been talk'n 'bout he-ah!16:04
sitaram summatusmentis: I already told you: git reflog --relative-date HEAD | less and look for stuff16:04
summatusmentis: paste if needed16:04
summatusmentis: you asked the question twice, identically, didn't listen to either answer, and then asked again :)16:05
← elliottcable left16:05
→ enherit joined16:06
summatusmentis sitaram: I missed the context, and thought you were talking to someone else, I'm sorry16:06
← enherit left16:06
→ enherit joined16:06
→ jjuran joined16:06
shruggar git co branchname $( git reflog | grep branchname | head -1 | awk '{print $1}' ) ? :D16:06
sitaram 21:17:08 <sitaram > summatusmentis: you deleted a branch? well the branch reflog gets deleted along with the branch :( but the HEAD reflog may still show you what you want... try "git reflog show HEAD -10"16:06
shruggar git co -b branchname $( git reflog | grep branchname | head -1 | awk '{print $1}' ) ? :D16:06
sitaram 21:17:08 <sitaram > summatusmentis: you deleted a branch? well the branch reflog gets deleted along with the branch :( but the HEAD reflog may still show you what you want... try "git reflog show HEAD -10"16:06
← ntoll left16:06
summatusmentis sitaram: then my client is being crappy, and I didn't get either of those16:07
sitaram even though I addressed you by name? you need a better IRC client16:07
← jackdempsey left16:07
sitaram my sentiments exactly; what are you using?16:07
summatusmentis neither of them showed up, I must have a laggy connection16:07
irssi16:07
sitaram oh thats what I use too; lag... then -- my apologies for being a little, er, sharp :)16:07
→ kjg joined16:08
summatusmentis no worries, my fault16:08
joshthecoder → Kroll16:08
→ tim joined16:08
summatusmentis if I've found what I'm looking for, and it's a recursive merge, can I cherry pick that?16:08
Kroll → joshthecoder16:09
← glauber_ left16:09
sitaram summatusmentis: why? you said you deleted a branch? Give that SHA the same name, and you're in business: git branch my_long_lost_branch SHA16:09
kjg hey, is it a known problem to get the git push warning message when pushing into a bare repo with git 1.6.3.3 or 1.6.3.1?16:10
summatusmentis oh, hrm, let me see if I can cherry-pick this commit from the log, and see if that does it16:10
sitaram kjg: a "warning" is never a "problem" :)16:10
kjg haha i suppose other than I don't like looking at it warning me about something that I shouldn't have to worry about16:10
→ cpg joined16:11
kjg does it warn on all pushes and not just pushes to non-bare repos?16:11
sitaram summatusmentis: don't do that... if you deleted a branch, undeleted it first, then do whatever you want. cherry-pick etc are not really a good idea most of the time, except as one-offs16:11
kjg: any warning in git can be squelched by some appropriate config variable16:12
kjg: the way you tell git "yeah I know, dont tell me OK?" is to set it :)16:12
→ dcurtis joined16:13
← froschi left16:13
→ jackdempsey joined16:13
dcurtis Is there a way to copy a file from one barnch to another easily with git?16:13
→ fualo joined16:14
kjg yeah, the warning message is very verbose. it says how squelch the warning, but unless i'm reading it wrong, it seems to indicate that if i squelch the warning that it will accept pushes to non-bare repos.16:14
summatusmentis sitaram: that doesn't seem to be working16:14
sitaram dcurtis: git checkout b1; git checkout b2 -- path/to/file16:15
kjg: pastebin please?16:15
summatusmentis: pastebin please?16:15
kjg: sorry that was not for you; you can just say which config variable it is talking about...16:15
kjg sitaram: receive.denyCurrentBranch16:16
summatusmentis sitaram: of which?16:16
→ Luca_Peck joined16:16
summatusmentis sitaram: I've gotta run, will you be around way later?16:16
sitaram kjg: pushes to non-bare repos are always OK; as long as it is to some other (non-checked-out) branch16:16
summatusmentis: maybe, how long? Its almost 10pm here and I may hit the hay soon but there'll be someone around dont worry16:17
kjg right, so I would think I shouldn't get a warning about pushing to a non-bare repo if it doesn't apply to me, it thought maybe it was a known issue that the warning got displayed all the tiem.16:17
summatusmentis sitaram: oh, I was thinking in 10 hours or so :)16:17
sitaram summatusmentis: make it 12 and you're on :)16:18
kjg I don't want to squelch it on the off change that I do push into a non-bare repo. The warning would be useful in that case16:18
→ bobesponja joined16:18
sitaram 1116:18
kjg s/change/chance16:18
→ kazzbott joined16:18
kazzbott good morning!16:18
← bobesponja left16:18
→ prusnak joined16:19
→ mw joined16:19
→ Geralt joined16:19
sitaram kjg: I kind ofkjgset it to refuse, not warn, and you'll be fine for all cases16:19
oops16:19
kjg: set it to refuse, not warn, and you'll be fine for all cases16:20
← naith left16:20
sitaram kjg: and refuse will become the default at a later stage, so you're future proof :)16:20
kazzbott Would someone mind checking out a pastie? http://pastie.org/58015816:20
← Jinzhu left16:20
kazzbott I forked a submodule and for some reason I can't deploy it, I get ERROR: Permission to tones95/active_merchant denied to tones95/hnd.16:20
kjg sitaram: ok cool. i'll do that. Do you disagree that it is misleading to show the warning for bare repos though?16:21
kazzbott but it works fine locally16:21
kjg s/misleading/confusing16:21
→ bobesponja joined16:21
sitaram if the remote was bare you should not see the warning. Did you see it on a bare repo? Thats a bug16:21
kjg: ^16:21
kjg yes! haha thats what i've been trying to say.16:22
→ ryandotsmith joined16:22
sitaram kjg: what version of git on either side?16:22
kjg local: 1.6.3.3 remote: 1.6.3.116:22
← FunkeeMonk left16:22
sitaram kjg: and does the warning start with "warning: updating the current branch"?16:22
kjg yes16:23
sitaram kjg: let me test... give me a minute16:23
← dcurtis left16:23
← Luca_Peck left16:24
kazzbott Anyone else having issues doing a pull from this public repository? git clone [email@hidden.address]16:24
← jackdempsey left16:25
→ Luca_Peck joined16:25
sitaram kjg: I dont see it here, but my versions are a bit newer... are you positive your remote is bare?16:25
wereHamster kazzbott: permission denied16:25
kjg $ git status16:26
fatal: This operation must be run in a work tree16:26
wereHamster kazzbott: is that a private repository?16:26
kazzbott hmmm16:26
it's public16:26
→ zalun joined16:26
kazzbott Privacy16:26
This repo is viewable by everyone.16:26
wereHamster kazzbott: the public fetch url is git://github.com/tones95/active_merchant.git, not [email@hidden.address]16:27
kazzbott ohhh16:27
hmm16:27
← zalun left16:28
sitaram kjg: ok that settles it I guess :) maybe you should send a mail to the list16:28
kjg well if it only happens in an old version, then i'm happy to upgrade. I just didn't see the fix in the change log16:29
→ loincloth joined16:29
sitaram kjg: but I cant reproduce it with 1.6.4 on both sides (and I dont have anything else anywhere) so not sure what the deal is; maybe it was fixed recently16:29
kjg: I didn't either, offhand16:29
→ juanjoc joined16:29
→ lhz joined16:30
← tiny left16:31
→ EruditeHermit joined16:31
→ kennethlove joined16:32
EruditeHermit how does one set the merge conflicts tool?16:32
→ Theravadan joined16:33
← ben_h left16:33
sitaram git config merge.tool meld16:33
→ funkycat90210__ joined16:33
sitaram (example)16:33
← ia left16:34
→ gaffo joined16:36
→ MrGando joined16:37
MrGando Hello :) , I need some help with git, how can I see which files did I commited on an specific revision ?16:37
→ mutable_ joined16:38
coachz isn't that in the manual ?16:38
← shruggar left16:38
→ ia joined16:38
→ Bira joined16:38
MrGando coachz: maybe, but I can't find it , could you point me on what to search for ?16:38
sitaram MrGando: git log --stat is a good start16:38
← naeu left16:39
MrGando sitaram: thanks16:39
→ justin-george joined16:39
← Bira left16:40
→ jerbear joined16:42
tokkee oO16:42
ddollar gaffo: your irc client going nuts?16:43
sitaram is everyone else seeing a mini-flood16:43
spb yes16:43
PerlJam aye16:43
← Chaser left16:43
← gaffo left16:43
tokkee Yep.16:43
EruditeHermit sitaram: thanks =)16:43
tokkee root@…16:43
tokkee shakes his head …16:43
sitaram EruditeHermit: you may want to add --global to that; typically that's a global setting, not per repo16:44
EruditeHermit does it take any text editor?16:44
sitaram tokkee: :) I see what you mean now. Full hostname too...16:44
EruditeHermit: there's a defined list of them; man git-mergetool16:45
Gitbot EruditeHermit: the git-mergetool manpage can be found at http://git.or.cz/man/git-mergetool16:45
→ Bira joined16:45
sitaram EruditeHermit: you can make it work for any by setting other config variables or writing wrapper scripts etc; it's all in man git-config16:45
→ mmw joined16:45
Gitbot EruditeHermit: the git-config manpage can be found at http://git.or.cz/man/git-config16:45
gbacon ?ty fromListWith (+) . map (flip (,) 1)16:46
← cilly left16:46
→ kraymer joined16:46
→ krawek joined16:46
← mutable left16:48
← moh left16:48
→ AkumaStreak joined16:50
→ smotchkkiss joined16:53
← gitte left16:53
← Bira left16:53
← MisterN left16:53
→ bsy joined16:55
→ brain0 joined16:56
→ MisterN joined16:56
brain0 hello. I have a git repository which I cloned from one location and want to push it to another. I cloned the "master" branch. Now I want to push the "master" and the "for-upstream" branch to the other location. so, I want it to pull master from the "upstream" remote when running "git pull" and when I run "git push --repo=publish" I want it to push the for-upstream AND the master branch to "publish"16:59
is this possible?16:59
→ moh joined16:59
→ ramlev joined17:00
sitaram brain0: first, you can't clone a branch; you clone the whole repo, or you fetch a branch into your repo17:00
→ meanburrito920_ joined17:00
brain0 I cloned the whole repository17:01
then I created a new branch17:01
wereHamster brain0: you can fetch and push from/to different urls since git-1.6.417:01
sitaram brain0: git push defaults to "matching" (or you can make it do that explicitly by setting some config). So -- one time -- create matching branches on the other side, and sure git push will work.17:01
brain0 I have 1.6.417:01
→ knittl joined17:01
brain0 hmm, I set it to "tracking"17:01
wereHamster set remote.<name>.pushurl to the url you want to push to17:01
sitaram brain0: or do it explicitly (my preferred method these days)17:02
duncanmv thanks for all the help guys, you rock17:02
← duncanmv left17:02
brain0 I am confused17:03
let me make this more explicit17:03
→ Sigma joined17:03
← sunoano left17:04
brain0 this is what I have right now: http://pastebin.com/d2198c97517:04
now I run git push --repo=publish17:04
it will push the for-upstream branch, but obviously not the master branch17:04
← daaku left17:04
→ eMyller|busy joined17:04
rylar_off → rylar17:05
→ murphl joined17:05
→ corni joined17:05
brain0 what I thought would be great is if I run git push --repo=publish, then both master and for-upstream will be pushed to publish17:05
→ sunoano joined17:06
sitaram git config remote.publish.push "master for-upstream" # may work; try it17:06
sitaram is not sure how a refspec consisting of two different branches is written; space separated?17:06
→ adeodatus joined17:06
← Octalot left17:07
→ ph^ joined17:07
brain0 sitaram: hmm, nope:17:08
sitaram brain0: there has to be a way to do this; give me a few minutes17:09
brain0 hmm, even if I say remote = publish on branch "master" it doesn't do anything17:10
← Hendu left17:10
brain0 I could try wereHamster's suggestion but it seems unclean to me17:10
← MrGando left17:11
← knittl left17:11
← murph left17:12
→ knittl joined17:13
coachz To run git on my xp box i see "Git-1.6.4-preview20090730.exe". It says featured Beta. Is that what I should be running or is there a non-Beta ?17:13
sitaram they're all called beta17:14
coachz so that's the right one ?17:14
← DrNick left17:15
← Luca_Peck left17:16
brain0 sitaram: for now, running "git push publish master" has done what I wanted, it's not nicely configured though17:18
sitaram brain0: well I always do explicit pushes anyway nowadays -- so that's not too bad actually17:18
brain0 I love to be able to "git push" and it will push everything where I want it to be17:19
that only works right if the repository I fetch from is the same one I push to17:20
← galderz left17:20
← kyxzme left17:20
meanburrito920_ so I know that if I want to get the return status of a command, i can use $? but what variable contains the error message itself?17:20
← justin-george left17:20
→ ctusar joined17:21
sitaram meanburrito920_: there isn't one. Each program has its own convention for what the codes mean and you have to read that program's manual17:21
meanburrito920_ sitaram, no, i mean whatever it prints to stdout17:22
or stderr17:22
rather17:22
← lu_zero_ left17:23
← kumbayo left17:23
→ kumbayo_ joined17:23
← corni left17:23
sitaram brain0: got it...17:24
brain0: git config remote.b.push b1:b1; git config --add remote.b.push b2:b217:24
brain0: then "git push b" :)17:24
← eMyller left17:25
← mutable_ left17:26
← fualo left17:27
coachz on windows install should i pick Use OpenSSH or Use Tortoise Plink and why please ?17:29
← sprasanna_ left17:29
sitaram coachz: msysgit I assume? if so, pick openssh. I've heard of various problems with the other ones. But I'm not an expert17:30
coachz correct17:30
yeah i don't know what plink even is17:30
brain0 sitaram: aaah, it seems to support several push = directives in one section17:30
sitaram: thanks (although I can't test it anymore, as there is nothing to push)17:31
sitaram coachz: there are often reports of incompatibilities between the various types of ssh too so if you've got more than one kind installed they sort of fight :(17:31
brain0: yep17:31
→ ppawel joined17:31
coachz i don't know if i have another, i have apache installed, mysql, php stuff17:31
→ jed_ joined17:32
← ramlev left17:32
← ctusar_ left17:33
stepnem :'( I can't believe there is no simple scriptable way to know if a local repository is up-to-date or not -- do I really have to parse the text output of 'git-pull' for that?17:33
← Theravadan left17:33
→ fualo joined17:33
← funkycat90210__ left17:33
jed_ what is the best way to deal with config files you don't want to watch but were added to the repository?17:33
wshimmy stepnem: git diff origin/master?17:33
→ candybar joined17:34
→ tongueroo joined17:35
stepnem wshimmy: ah... that automatically fetches from the remote and diffs it?17:36
brain0 you have to fetch yourself17:37
wshimmy stepnem: no, you have to fetch it first17:37
brain0 the check the return value of git diff origin/master master17:37
wshimmy git doesn't hit the network without explicit orders to, in most (all?) cases17:37
stepnem hm... so there's no way to just *check* for changes to fetch/pull?17:37
before fetching them?17:38
wshimmy no, but if you just fetch them it doesn't affect your local branches at all17:38
stepnem yes, I know17:38
sitaram stepnem: git ls-remote may help; at least if the hashes are not equal you know there's some difference, whether push or pull17:39
→ legendulo joined17:39
sitaram hits the network of course but very lightly17:39
← sparc left17:40
← jed_ left17:41
coachz are there any advantages to me picking tortioseplink during install? i already have tortoise svn installed so plink is there.17:41
← jaeckel left17:42
brain0 sitaram: thanks for your help again (this is one of the few channels on IRC that is actually helpful)17:43
and have a nice day17:43
coachz it's a great channel !17:43
← zenwryly left17:43
brain0 bye17:43
← brain0 left17:43
← wolog__ left17:43
← tono left17:45
→ MrMeek joined17:45
← WinterWeaver left17:46
stepnem sitaram: thank you!17:46
sitaram stepnem: did it work?17:47
stepnem sitaram: I'm only trying to figure it out, my understanding of git is still limited :|17:49
← ryandotsmith left17:49
sitaram ok :)17:49
troyt Quick question on fast-import: The 'merge' keyword - is it more to be used along the lines of merging one branch to another, or what? (I've got a SCM i'm importing from where I'll get what amounts to several cherry-picks in one 'commit', and wanted to get it to map into git as best as possible)17:50
→ khmarbaise joined17:52
→ dreiss joined17:53
← kraymer left17:54
← kumbayo_ left17:54
← legendulo left17:54
← markelikalderon left17:55
← adeodatus left17:55
→ owen1_ joined17:55
→ yann joined17:57
→ pacific-blue joined17:58
Ilari troyt: Presumably 'merge' means to do a merge in sense git understands merges.17:59
→ patmaddox joined17:59
← rb____ left17:59
pacific-blue When specifying the "git remote add origin ssh://user@server..." I must specify the full path to my user account. It seems to me it should default to the users home directory. Is there a way to specify it so it defaults to the home directory just like an rsync does?18:01
Ilari troyt: I.e. all commits on that branch that aren't on this branch become part of this branch.18:01
doener pacific-blue: user@server:path/relative/to/home18:01
→ markelikalderon joined18:01
→ qhoxie joined18:01
pacific-blue doener: so the key here is to specify a ':' after the user?18:02
sitaram pacific-blue: man git-clone and look for ssh: to see all possible syntaxes18:02
Gitbot pacific-blue: the git-clone manpage can be found at http://git.or.cz/man/git-clone18:02
doener pacific-blue: and not to use ssh://, the user@host:syntax comes from scp18:02
→ _Vi joined18:02
→ transplant joined18:03
transplant hi. anyone knows of any repo that's in http and not packed? I want to do some testing18:03
pacific-blue doener: thanks that worked perfectly. :)18:04
stepnem sitaram: umh, I'm confused: both the output of 'git ls-remote -h origin master' and the contents of .git/refs/heads/master don't change (and are different), even though I did a 'git pull'; I must be missing something :|18:04
→ JackWinter3 joined18:05
doener stepnem: "git pull" or "git pull <remote> <refspec>"?18:05
→ kumbayo_ joined18:05
stepnem just git pull18:06
ah, sorry, I'm not on the master branch18:06
*blush*18:06
doener well, just the remote configured for the branch matters18:07
← kazzbott left18:07
→ pdobrogost joined18:07
pdobrogost Hi all!18:07
stepnem doener: indeed, it's the same for the branch I'm on18:08
[so I'm still confused, or even more]18:08
doener stepnem: and it's set to "origin"?18:08
← EruditeHermit left18:08
doener stepnem: git config --get branch.<name>.remote18:08
→ juan_arandaalvar joined18:08
→ sshc joined18:09
pdobrogost Need some help with git. Have local repository with remote-tracking branches. How do I now which of these remote-tracking branches my local branch synchronizes with?18:09
stepnem doener: yes, origin18:09
doener stepnem: and: git config --get remote.origin.fetch18:09
sshc how do I rename a branch without creating a new one and removing the old one?18:09
doener sshc: git branch -m18:09
← sitaram left18:10
doener pdobrogost: "git config --get branch.<name>.remote" and "git config --get branch.<name>.merge" show the relevant "git pull" defaults18:10
pdobrogost: if you know which remote you care about, you can also do "git remote show <remote>"18:11
→ shingo joined18:11
shingo → Guest8913618:11
doener pdobrogost: which shows, besides other information, which branches have pull defaults, and what they are18:11
← RaceCondition left18:11
stepnem doener: +refs/heads/*:refs/remotes/origin/*18:11
doener stepnem: oh wait... you looked at refs/heads/master, not refs/remotes/origin/master18:12
stepnem: you most likely just have local commits that don't exist in the remote repo, so obviously master references a different commit than origin/master18:12
doener gotta go18:13
stepnem doener: yes, but I thought that after I pull from the origin, it should be updated to the same hash18:13
← Aikawa left18:13
stepnem doener: and no, I don't have any local commits18:13
→ panpainter joined18:14
← ^Einstein left18:14
coachz on windows, i don't see a .gitconfig file in my C:\Documents and Settings\$USER directory18:16
→ Lemao joined18:16
← GarethAdams left18:17
← patmaddox left18:17
→ patmaddox joined18:17
coachz but i do see it here: C:\Program Files\Git\etc18:17
← dpino left18:17
coachz just gitconfig not .gitconfig18:17
→ marvil07 joined18:18
stepnem doener: oh, sorry, I was just confused -- now it shows the same hash as expected18:19
pdobrogost doener: after git branch -r I have origin/4.5 origin/HEAD -> origin/master origin/master origin/master-stable18:20
← JackWinter2 left18:20
stepnem coachz: presumably you haven't made any personal config changes yet?18:21
pdobrogost doener: I want to follow master-stable. what should I do?18:21
stepnem but I don't use windows, so not sure18:21
coachz stepnem, i just made some form dos in c:\ and it wrote the config to c:\18:22
hmm18:22
if i write to my home dir, will git find it ?18:22
even from home it wrote the new file to c:\18:23
so now i have .gitconfig in c:\18:24
→ ryandotsmith joined18:24
stepnem coachz: there are repo-specific `config' files in each .git directory, and then there's a global .gitconfig, in the user's home directory on Unix18:24
coachz doesn't seem like the right place for it since the book says it goes in home file:///C:/My%20Web%20Sites/Pro%20Git%20Book/progit.org/book/ch1-5.html18:24
← ppawel left18:24
coachz i'm on windows xp though18:24
i have just done a new install of windows git and have no repos yet18:25
← jjuran left18:25
→ SolarWar joined18:26
coachz when i run this from C:\Documents and Settings\$USER it writes to c:\.gitconfig git config --global user.name "John Doe"18:26
18:26
pdobrogost doener: ping18:26
→ dodo_the_last joined18:29
→ gma joined18:29
gma how easy is it to undelete a branch?18:30
→ thiago_home joined18:30
← Evanlec left18:30
← dreiss left18:30
coachz read this page git checkout -b new_branch_name HEAD@{1} http://andrzejonsoftware.blogspot.com/2008/06/git-working-with-branches.html18:31
coachz total noob fwiw18:31
gma coachz: cheers. think I may just have found the relevant sha with the reflog18:31
phew18:31
coachz yay18:32
gma yep, it's all good. that'll teach me to tidy up after myself...18:32
coachz one thing i read about git is that it basically always adds so you can't really delete and can always go back18:32
gma coachz: yeah. the reflog is a nice piece of work.18:32
→ Suhail joined18:32
coachz what's a reflog18:32
gma coachz: try "git reflog"18:33
Suhail In git how do I get rid of 2 commits to a branch, if I haven't pushed them yet18:33
coachz i guess i'll need a repo first :-)18:33
gma coachz: it'll show you the sha every time you switch branches. I went back far enough to find the last time I switched onto that branch, then did "giti checkout -b mybranch abcd1234"18:33
stdorb coachz: like a log of where you did your time travel to.18:33
← Guest89136 left18:33
coachz nice18:33
gma Suhail: git reset --hard HEAD^^18:33
Suhail: so long as you don't really want 'em18:34
stdorb and then they are still available through the reflog ;)18:34
coachz this is the quote in the Pro Git book: When you do actions in Git, nearly all of them only add data to the Git database. It is very difficult to get the system to do anything that is not undoable or to make it erase data in any way.18:34
gma coachz: yeah, that's true18:35
coachz smart18:35
← ryandotsmith left18:35
gma coachz: if you do "git reset --hard HEAD^" you throw a commit away18:35
← marvil07 left18:35
thiago_home not really18:35
the commit is still there18:35
← transplant left18:35
gma coachz: then if you type "git reflog" you'll see the sha of your old HEAD, and can then type "git reset --hard abcd1234" to go back to where you were18:35
→ jigfox joined18:36
→ patmaddo_ joined18:36
coachz fun stuff18:36
gma coachz: eventually though it'll get garbage collected, so it won't stay there forever18:36
coachz k18:36
→ Evanlec joined18:36
coachz i'm just doing a new windows install so i have to make my ssh key now18:36
gma though you may have to run git-gc to do the collecting, I'm not totally clear on that18:36
→ dschulz joined18:37
← jigfox left18:38
← Lemao left18:39
→ Dashkal joined18:39
← lohrun left18:41
coachz hmm, ssh-keygen put the .ssh dir in my c:\ root18:41
→ Akufen joined18:41
← gma left18:41
coachz and my .gitconfig is there too18:41
SRabbelie1 → SRabbelier18:41
coachz for some reason it's not putting them in my home dir18:41
thiago_home probably $HOME is empty18:41
coachz ahh18:42
env shows HOME=/c whatever that means HOMEDRIVE=C: HOMEPATH = \18:42
seems odd for xp18:42
← _Vi left18:43
coachz and HOME system var is set to C:\18:43
so i'll change that to be right18:43
and start over18:43
→ jigfox_ joined18:43
→ Bira joined18:43
Ilari coachz: Just copy the files to proper place after changing the variables?18:44
coachz ok18:44
→ ryandotsmith joined18:44
→ dreiss joined18:48
→ transplant joined18:48
← jigfox_ left18:50
rylar → rylar_off18:50
→ eMyller joined18:50
→ jigfox joined18:50
→ Guest89136 joined18:51
← owen1_ left18:51
→ DasEi joined18:51
transplant any news on allowing branches to point to tag objects?18:51
→ g3d joined18:53
→ dandean joined18:53
→ lohrun joined18:53
patmaddo_ → patmaddox_18:54
← patmaddox left18:54
→ montylounge joined18:54
patmaddox_ → patmaddox18:54
→ justin-george joined18:55
← Adlai-AW` left18:56
← WebDragon left18:56
→ Adlai joined18:56
dandean dumb GitHub-related question: when editing a file with their web-editor, git pull doesn't pull change down to local files. Is git pull the wrong command?18:57
→ WebDragon joined18:59
← pacific-blue left19:03
cpg → cpg|biab19:03
← khmarbaise left19:04
→ wolog joined19:04
meanburrito920_ dandean, i would suggest asking on #github19:05
does anyone know how to limit who is allowed to push to a repo (for a bare repo on a server that many users have access to)19:06
dandean I'm doing that as well.19:06
meanburrito920_ dandean, ah :)19:06
git pull seems like it would be correct. does github have an updated sha for the edit you made? did it register as a commit?19:07
Nafai meanburrito920_: There is an example here under "Enforcing a user-based ACL system" http://progit.org/book/ch7-4.html19:07
If I'm understanding correctly what you are asking19:07
→ galderz joined19:08
← montylounge left19:08
← bentob0x left19:09
dandean meanburrito920_: I'll look into that. thanks!19:10
← markelikalderon left19:10
← eMyller|busy left19:10
← jrmuizel left19:13
← DasEi left19:14
← saurabhverma left19:14
→ tiny joined19:15
→ Gof joined19:16
Gof there is the wonderfull command 'git grep' , is there a command like 'git sed' ?19:16
→ DasEi joined19:16
→ lu_zero_ joined19:16
stdorb in either case you would want perl instead of sed if you're going to do regexp substitutions19:16
→ Derander joined19:17
meanburrito920_ Gof, that sounds dangerous :) imagine what kind of screwups could happen...19:17
Gof meanburrito920_: refactoring, you know19:18
meanburrito920_ true enough. but why would you bother with previous revisions? why not just use sed as normal?19:18
Gof ok, i'm doing find -name \*.cpp -or -name \*.h | xargs sed -i s/foo/bar/g that will work :-)19:18
meanburrito920_ i think you need the s//g19:18
statement to be in quotes19:18
also, the -i needs an argument for what the backup file extension will be19:19
Gof meanburrito920_: it is just to have access to all source files. I always use git grep on main because it is faster19:19
HEAD i mean19:19
meanburrito920_ Gof, right, but grep and sed will work on all the source files in your working copy, which is all you should really be messing with19:20
→ jrmuizel joined19:21
Gof meanburrito920_: agreeed19:21
→ porcnick joined19:23
coachz if i'm developing on windows and running the app on linux do i need "git config core.autocrlf false"19:23
transplant error: untracked working tree file 'xxx' would be overwritten by merge.19:23
?19:23
→ mcdebugger joined19:24
meanburrito920_ coachz, turn it on for the windows box, and off for the linux box19:24
coachz thanks19:24
→ mjf joined19:24
← mjf left19:24
→ camwest joined19:25
→ spaes joined19:25
→ sgrimm joined19:26
→ mjf joined19:26
← Adlai left19:27
→ Adlai joined19:27
→ Octalot joined19:29
coachz on my new windows repo, i did git init, then git add ., then just to try to unstage them i did git reset and i got fatal: Failed to resolve 'HEAD' as a valid ref.19:29
thiago_home unstaqge what?19:30
→ kyxzme joined19:30
coachz ahh19:30
add didnt stage anything :-)19:30
doh19:30
← g3d left19:30
coachz buries his head in the pro git book in shame and continues reading19:30
meanburrito920_ coachz, actually, the reason why was because HEAD didn't exist, because it is created on the first commit19:31
→ malumalu joined19:31
coachz cool, i just did git commit -a -m "first commit" and then did git reset and it worked :-)19:32
← Adlai left19:32
→ Adlai joined19:32
thiago_home of course it did :-)19:32
coachz so with commit, did i just stage them or are they committed locally ?19:32
thiago_home when you use Git properly, it works19:32
meanburrito920_ :D19:32
coachz, stage what?19:33
thiago_home coachz: they're committed locally19:33
coachz well, i just created my local repo on my xp box19:33
git init19:33
then added the files git add .19:33
so that added them to the index (staged) ?19:33
meanburrito920_ then if you commit the files are commited, not staged19:33
coachz makes sense19:33
meanburrito920_ :)19:34
coachz but to git reset, you must have done one commit19:34
meanburrito920_ staging a file is git add /file/19:34
right19:34
coachz so you can't unstage until you do one commit19:34
meanburrito920_ becuase it has to have something to reset to, right?19:34
:)19:34
coachz yes19:34
:-)19:34
i like it !19:34
← transplant left19:35
→ Sordran joined19:35
← asksol left19:36
coachz has 3 million dollar houses being auctioned off across the street in an hour. oh the joy of foreclosures, i bet they go for 500k19:38
← Adlai left19:38
→ Adlai joined19:38
coachz not mine19:39
meanburrito920_ haha19:39
→ markelikalderon joined19:39
thiago_home checks his pockets for 500k in loose change19:39
nick_h i've read the manpage for git-diff, and the "specifying revisions" part of git-rev-parse. however, i can't seem to get git-diff to do what i want19:39
coachz one house has no reserve19:39
meanburrito920_ nick_h, which is?19:39
coachz, where do you live?19:40
coachz charleston, sc19:40
meanburrito920_ theres million dollar homes in south carolina?19:40
nick_h if i have commits 422bc927d and a05c9bfe, how do i display the differences between them? i thought this would achieve that: git diff 422bc927d a05c9bfe19:40
coachz tons of em19:40
those are the average houses19:40
thiago_home nick_h: that's how you do it19:40
meanburrito920_ nick_h, put two periods in between them19:40
coachz starter house is 250k, basic nice house is 450k19:40
meanburrito920_ git diff 422bc927d..a05c9bfe19:41
coachz, yeah, i live in ca and the house prices here are going crazy19:41
← markelikalderon left19:42
coachz yeah ca is like wild19:42
nick_h meanburrito920_, thiago_home: ack, brain fart. i was in the wrong directory =/ thanks, guys!19:43
coachz rookie mistake .....minus 5 points :-)19:44
← Adlai left19:45
→ Adlai joined19:45
meanburrito920_ coachz, minus 5 points from you :)19:45
← dschulz left19:46
coachz i'm at -10000000019:46
→ hobodave_ joined19:46
← ryandotsmith left19:46
meanburrito920_ coachz, no, not that bad19:46
more like -999999919:46
coachz ha19:46
meanburrito920_ jk19:46
:P19:46
← galderz left19:47
← dandean left19:47
← Suhail left19:47
coachz i'm teaching an 86 year old man to do web design and he forgets everything i teach him everyday so we just start over everyday, talk about fun19:47
he's bored19:47
meanburrito920_ apparently.19:47
← tiny left19:49
→ markelikalderon joined19:49
→ alezandro joined19:51
alezandro hi all, i accidentally pushed to a server and created a new branch, how can I delete it? tried "git branch -D origin/my_bad_branch" but no luck19:53
← Sordran left19:53
→ ppawel joined19:53
alezandro "git branch -a" shows it as remotes/origin/my_bad_branch19:53
coachz git branch -D branchname is for unmerged branches right ?19:54
→ ntoll joined19:55
RobertMeta → MetaCosm19:55
← hobodave left19:55
hobodave_ → hobodae19:55
→ dwave joined19:55
hobodae → hobodave19:55
Bira alezandro: git push origin :my_bad_branch19:55
alezandro Bira: worked perfectly, thanks19:56
would've never figured that out19:56
coachz so how does push delete ?19:57
Bira "git help push" will show you a few of the less intuitive options that the command supports.19:57
I think the idea here is that you're pushing "nothing" on top of a remote branch.19:57
coachz ahh19:57
→ Suhail joined19:59
← cloudhead left19:59
← mnick left19:59
→ ryandotsmith joined20:00
→ psynaptic joined20:02
→ Theravadan joined20:03
→ funkycat90210_ joined20:03
→ __iron joined20:05
coachz what do i put in .gitignore to ignore all _notes directories but not files20:05
→ dzwell joined20:06
← ilteris_ left20:06
davegb3 coachz: i think it's just _notes/20:06
coachz thanks davegb320:06
dzwell could anyone remind me of the proper place to put comments when submitting a patch? (not the commit message, but other comments related to the e-mail)20:07
is it between the diff stat and the "---"?20:07
← Bira left20:13
charon dzwell: yes20:13
actually the diffstat has nothing to do with it, it's simply that everything between '---' and the first line starting 'diff ...' is ignored20:14
(so the diffstat is a special case of such comments)20:14
← coldwaterlover left20:16
→ saurabhverma joined20:16
← Geralt left20:17
→ Geralt joined20:17
coachz git diff tells me the differences between files that are staged and unstaged yes ?20:19
meanburrito920_ n20:20
no20:20
coachz hrm20:20
meanburrito920_ it tells you the difference between HEAD and the working copy20:20
← simNIX left20:20
meanburrito920_ although that may be what you meant20:20
← iron left20:20
charon no20:20
thiago_home git diff, without any arguments, gives you the difference between staged and the working copy20:20
the staged files may be different from HEAD20:20
charon ... bah, i'm too slow at typing. thiago wins20:20
dzwell charon: thanks!20:21
meanburrito920_ really? so staged is not diffed?20:21
→ ilteris joined20:21
meanburrito920_ wow, i never knew that20:21
always more to learn, i guess :)20:21
coachz so what is working copy? is that staged, unstaged ?20:21
coachz ducks20:21
meanburrito920_ coachz, the working copy is the physical files on the file system20:22
coachz ok20:22
that's clear20:22
thiago_home there's the checkout (the working copy), the staging area (the index) and the commits20:22
→ g0blin joined20:22
coachz the Pro Git books says git diff tells you What have you changed but not yet staged? And what have you staged that you are about to commit?20:22
meanburrito920_ thiago_home, but git diff includes HEAD for the files that are not staged, correct?20:23
coachz kinda confusing to a noob20:23
thiago_home files that aren't staged aren't tracked20:23
if they are tracked, they are staged20:23
coachz git add filename makes then tracked and stage if i'm right20:23
cpg|biab → cpg20:24
thiago_home yes20:24
← kjg left20:24
→ coldwaterlover joined20:25
meanburrito920_ ah, so i was getting hung up on semantics20:25
→ porcnick_ joined20:25
coachz there's about 15 cars next door for the auction so i'm gonna check it out bbl guys and thanks for being patient with noobs :-) ......still reading :-)20:26
meanburrito920_ thiago_home, so any tracked files are considered staged?20:26
→ koltroll joined20:27
meanburrito920_ so what do you call it when you add a file to be committed?20:27
_rane staging?20:27
thiago_home staging20:28
meanburrito920_ but you said that all tracked files are already staged20:28
thiago_home going to be committed = staged20:28
yes20:28
meanburrito920_ that makes no sense20:28
thiago_home sure it does :-)20:28
meanburrito920_ what if i have a file that is tracked by the repo, but I'm not going to commit it for that commit?20:29
charon meanburrito920_: you're thinking in *changes*. we're thinking in *files*. the stage (= index = cache) is the set of files/file contents that will be in the next commit20:29
meanburrito920_ it shows up in git status as being tracked20:29
charon meanburrito920_: so even if you do not stage the *changes*, you still have the (old) file contents staged, so that they end up in the next commit "unchanged"20:29
(note that git only talks about "changes" after the fact, i.e., it generates them from the tree snapshots, not vice versa)20:29
thiago_home does it make sense now?20:30
meanburrito920_ charon, ooh. ok, that makes sense.20:30
→ DrNick joined20:30
→ neynenmo joined20:30
meanburrito920_ i didnt realize that changed files had their last unchanged version staged20:31
by default20:31
thiago_home yes20:31
charon that last thing about changes also goes for "unstaged/uncommitted changes" btw. by definition, unstaged changes are the diff between worktree and index (git diff) and the staged-but-uncommitted changes are the diff between index and HEAD (git diff --cached)20:32
thiago_home when you check out a commit, the contents of that commit are staged20:32
if "git status" shows no changes, that the commit matches the staging are (the index), which matches the working copy20:32
← neynenmo left20:32
thiago_home if it shows a file changed but not staged, it means the commit matches the staging area, but the working copy differs20:33
if it shows the file changed and staged, then the staging area and the working copy match, but the commit differs20:33
it's also possible to have all 3 differ. It happens if you use git add -p, for example.20:33
← ericvw left20:33
← malumalu left20:34
jaalto I use external too git-buildpackage (Debian), which magically imports source archive to branches 'upstream' and 'pristine-tar' and then merges the chnages to 'master'. Well, I want to go back to state *before* all that took place. I know the previous 'master' commit it where I want to go, but I'd like to reset all other branches as well to the state of that (can I use DATE somehow)? What's that correct approach?20:34
davegb3 thiago_home, charon: just want to say that explanation was brilliant... that really helps me understand the staging area - i thought only new changed files got copied in. makes perfect sense now.20:34
thiago_home davegb3: it is confusing for new users20:35
jaalto commit it =< commit id20:35
← porcnick_ left20:35
→ porcnick_ joined20:36
jaalto I think I just need to 'for BRANCH in list-of-branches; do git reset --hard HEAD~1 ; done', but what do you think?20:37
cxreg you would need to check the branches out first20:38
← mmw left20:38
cxreg (which then changes the meaning of HEAD)20:38
jaalto cxreg: Yes, naturally, but the general concepts? I can't use some git reflog trickery?20:38
cxreg you could git update-ref refs/heads/whatever directly20:39
← mm_work left20:39
→ saurabhverma1 joined20:39
cxreg that's a plumbing command20:39
→ Chaser_ joined20:39
← mw left20:40
jaalto cxreg: still per branch basis?20:40
cxreg a branch is a ref, so yes20:40
branch "pristine-tar" is refs/heads/pristine-tar20:40
try "git for-each-ref" to see what i mean20:40
← porcnick left20:40
jaalto cxreg: No chance to say "reset everything to the stae of <date>"?20:41
cxreg everything?20:41
no\20:41
wereHamster jaalto: if you know that @{1} of each branch points to the commit just before the last update, then you can use for-each-ref/update-ref to loop through all branches and update them20:41
jaalto cxreg: Eveything, yes. As of before the "magick import"20:41
← plediii_ left20:41
← jtaby left20:41
wereHamster jaalto: you could also use @{yesterday} or another date instead of @{1}20:42
jaalto wereHamster: ah, that was it. What's the format for day, hour, min?20:42
→ mw joined20:42
← Suhail left20:43
← vmiklos left20:43
→ EruditeHermit joined20:43
→ ilteris_ joined20:44
wereHamster jaalto: man git-rev-parse -> specifying revisions20:44
Gitbot jaalto: the git-rev-parse manpage can be found at http://git.or.cz/man/git-rev-parse20:44
wereHamster fourth bullet point20:44
→ mmw joined20:45
jaalto checking wereHamster's manual20:45
→ Suhail joined20:45
jaalto wereHamster: chnage spaces to dots on shell cmd line? as in: {1979-02-26.18:30:00} ?20:46
charon or just quote them from the shell20:47
wereHamster that, or quote the whole command20:47
charon reset to HEAD@{teatime}20:47
jaalto wereHamster: Ok, thanks. Let's see.20:47
← icwiener left20:48
← Gof left20:50
← bdiego left20:51
← cemerick left20:52
← __iron left20:54
→ Thiago__ joined20:54
jaalto wereHamster: Works fine, thanks again.20:54
EruditeHermit what is the recommended mergetool?20:54
→ __iron joined20:54
jaalto cxreg: Do you have experience with pristone-tar?20:55
tine20:55
← saurabhverma left20:57
→ _Vi joined20:57
← Ryback_ left20:57
→ usomaki joined20:58
usomaki hi20:58
→ vmiklos joined20:59
← jerbear left21:01
cxreg jaalto: no21:02
← ilteris left21:02
← dzwell left21:02
cxreg i didn't even know there was a git-buildpackage21:03
→ schmrz1 joined21:04
schmrz1 Hello Everyone. I have created a bare repo in /var/www/, pushed to it from a local repo and now when I try to clone from the bare repo I get "warning: remote HEAD refers to nonexistent ref, unable to checkout."21:04
schmrz1 → schmrz21:05
schmrz I would appreciate any help. Thanks21:05
← lohrun left21:05
wereHamster schmrz: can you post the output of git ls-remote <url> to a pastebin?21:07
schmrz From which repo?21:07
The local one, or the bare one?21:07
cxreg on the local one, do "git ls-remote <remotename>"21:07
wereHamster you don't have to be in a repo, but <url> should point to the bare repo in /var/www21:08
cpg → cpg|brb21:08
→ quoin- joined21:08
schmrz will do21:09
← tfarina left21:09
← Chaser_ left21:10
schmrz 4bf3c197ffeab68b3df0f532be0f1575aa9abab9 HEAD21:10
4bf3c197ffeab68b3df0f532be0f1575aa9abab9 refs/heads/master21:10
wereHamster, cxreg, this is it.21:10
→ porcnick joined21:11
Ilari schmrz: git --version ?21:11
wereHamster schmrz: when you try to clone the repository, which url do you use? http://, ssh://, git:// or from the local filesystem?21:12
schmrz 1.6.0.421:12
wereHamster: http21:12
wereHamster schmrz: you probably need to run git update-server-info in the bare repository21:13
schmrz will do that now.21:13
← hpa left21:13
wereHamster (and enable the post-receive hook to do it automatically after each push)21:13
← porcnick left21:13
← porcnick_ left21:13
→ porcnick joined21:14
← DasEi left21:14
cxreg or simply use git-daemon?21:15
why does anyone even bother with http21:15
schmrz For some reason my friend can foward port 9418, but he can forward port 80...21:15
Don't ask me...21:15
← koltroll left21:15
cxreg heh21:16
Adlai schmrz: which one of those should be a "can't"?21:16
schmrz can't for 9418 :)21:16
Adlai does git:// always run on 9418?21:16
← seangrov` left21:16
schmrz Adlai: no but he needs to have port 80 available for http21:17
→ an0maly joined21:17
← bugfux left21:17
wereHamster it's the default port, but you can change that with --port=X21:17
→ bugfux joined21:18
← g0blin left21:19
← loincloth left21:21
→ RaceCondition joined21:24
← schmrz left21:24
→ WebDragon|away joined21:24
← RaceCondition left21:25
← robinr left21:26
← WebDragon left21:26
WebDragon|away → WebDragon21:26
← rudi_s left21:31
→ Tobin joined21:31
← EruditeHermit left21:32
← kumbayo_ left21:33
→ RaceCondition joined21:34
smotchkkiss is it difficult to setup your own git server?21:35
← Geralt left21:35
← Sigma left21:36
→ [1]intripoon joined21:36
→ Chaser_ joined21:37
← quoin- left21:37
Ilari smotchkkiss: Depends mainly on how co-operative SSHD is and also how substle mistakes one makes...21:38
→ jceb81 joined21:38
jayne is there a way to show differences (git diff) that don't match one or a few regular expressions?21:39
← camwest left21:39
← saurabhverma1 left21:39
jayne in this case, a function was renamed, and it's used all over the bloody place, so I want to see differences that _aren't_ the rename21:40
Ilari smotchkkiss: There are instructions for setting it up. Some mistakes can be maddeningly difficult to track down.21:40
hobodave is it me or does github have daily outages?21:40
smotchkkiss Ilari: thank you21:40
i will look into it i guess21:40
Ilari smotchkkiss: And SSHD being so paranoid doesn't help anything debugging those...21:41
→ DrNick2 joined21:42
← DrNick left21:43
DrNick2 → DrNick21:43
← spaceonline left21:45
→ elliottcable joined21:46
→ coldwate1lover joined21:46
← coldwaterlover left21:47
RaceCondition why does git remote show remotename hang when remotename is down?21:47
← pdobrogost left21:47
→ ^Einstein joined21:48
← ntoll left21:49
cpg|brb → cpg21:50
Ilari RaceCondition: IIRC, add -n option so it doesn't try to contact it?21:53
← Tobin left21:53
RaceCondition Ilari: OK, but why does it by default?21:53
← intripoon left21:53
WebDragon → WebDragon|away21:54
← spaes left21:54
meanburrito920_ does anyone know why a pre-recieve hook would not be called?21:54
← ppawel left21:55
← lhz left21:55
← _Vi left21:56
← fernandor left21:57
→ fernandor joined22:01
→ plediii joined22:01
← ryandotsmith left22:03
← tim left22:03
meanburrito920_ I have a pre-receive hook in my bare repo. I push to my bare repo and the pre-recieve hook is never called. is there a config setting I need to enable this hook?22:03
Ilari meanburrito920_: Not enabled? Pushing via some dumb protocol (like http://)?22:03
meanburrito920_ I also noticed that this hook does not appear under the hook samples22:03
→ tim_jenssen joined22:03
meanburrito920_ Ilari, no, on the local file system22:03
Ilari meanburrito920_: What 'ls -l' says about that hook?22:03
cxreg meanburrito920_: it must be +x ;)22:03
that's the "setting"22:04
meanburrito920_ that would be it :)22:04
thanks22:04
← nitrate left22:04
→ lil-murph joined22:04
→ nitrate joined22:05
← lorandi left22:05
→ RayMorgan joined22:05
← [1]intripoon left22:05
→ khmarbaise joined22:06
→ ryandotsmith joined22:06
← AkumaStreak left22:07
coachz $ git commit22:07
error: There was a problem with the editor 'pspad'.22:07
Please supply the message using either -m or -F option.22:07
smotchkkiss coachz: i used to use pspad22:07
coachz i set an EDITOR env var to my path to pspad.exe22:07
in xp22:07
smotchkkiss yea22:07
coachz env22:07
→ robinr joined22:07
coachz EDITOR=C:\Program Files\PSPad editor\PSPad.exe22:07
→ AkumaStreak joined22:08
Ilari smotchkkiss: And as hint about setting git server. Use some Unix if at all possible. But if one has to set it up on windows, install cygwin sshd and cygwin git.22:08
→ ajonat joined22:08
smotchkkiss Ilari: it'd be on CentOS22:08
coachz so how come it can't find my editor ?22:09
→ jjuran joined22:09
smotchkkiss coachz: do you need quotes around that path?22:09
← bcardarella left22:09
coachz none of the others have them22:09
smotchkkiss coachz: ok i'm not sure then22:09
← AkumaStreak left22:10
coachz i have core.editor=pspad22:10
should that be pspad.exe ?22:10
← khmarbaise left22:10
meanburrito920_ coachz, shouldnt that be pspad.exe22:10
haha22:10
too slow22:10
coachz i try it22:10
← ctusar left22:11
→ khmarbaise joined22:11
← Evanlec left22:11
← bugfux left22:12
← candybar left22:12
← Thiago__ left22:13
coachz nope, didn't work22:13
core.editor=pspad.exe22:13
in global .gitconfig22:13
← murphl left22:14
Ilari coachz: Maybe try 'pspad.exe' from the prompt? Does that work?22:14
coachz not because it's not in the path, just in the env var22:14
maybe i need a command line option in gitconfig22:14
RayMorgan Hello, is there anyway to create a branch that is always kept up to date with another branch? (other than just symlinking .git/refs/heads/foo -> .git/refs/heads/)22:14
← knittl left22:15
RayMorgan .git/refs/heads/bar**22:15
→ psytek joined22:15
← hugoxrosa left22:15
← thiago_home left22:15
davegb3 coachz: i did this in windows for notepad++ as the editor, and i couldn't get any combination of quotes to work - in the end i used a batch file to launch the absolute path, and put the batch file in a dir with no spaces (like c:\windows)22:15
funkycat90210_ when I do a git diff, not all modified files show up. Why is that?22:16
coachz this didn't work either EDITOR=C:\Program Files\PSPad editor\PSPad.exe %122:16
do i need %1 ?22:17
Ilari coachz: If whatever you put into that option doesn't work on command line, it's probably not going to work as editor setting.22:17
coachz well an env var is different than a path22:17
→ Evanlec joined22:17
Ilari coachz: Leave out the %122:17
coachz the instructions for git didn't say to add it to the path22:17
i tried before without %122:17
i'm just grasping22:17
Ilari coachz: Does 'C:\Program Files\PSPad editor\PSPad.exe' from command line work?22:18
coachz yeah22:18
i'll add to path now and try22:18
← smotchkkiss left22:18
fpcfan → fpcfan-away22:19
← fernandor left22:19
coachz adding to path didn't help either22:20
→ mw_ joined22:20
coachz i'm supposed to be lost asking rebase questions, not setting up a stupid editor22:21
← john39091 left22:22
→ john3909 joined22:23
Ilari coachz: Maybe the space in path causes trouble?22:23
coachz let me try with notepad first22:24
→ jnareb joined22:24
← MrMeek left22:24
→ cloudhead joined22:26
coachz that works22:26
%SystemRoot%\system32\notepad.exe22:26
just that in the EDITOR env22:26
→ Hideki joined22:26
jnareb Grrr... I hate not-blogs (looks like a blog, but doesn't have comments) like http://blog.zacharyvoase.com/post/147813314/why-mercurial-git-or-how-i-learned-to-stop-censoring22:26
coachz and core.editor notepad without the exe22:26
→ JEEBcz joined22:27
← usomaki left22:27
jnareb Ever heard about operating system written in higher-level, dynamic language?22:27
← mmw left22:28
coachz so i put pspad in c:\pspad\pspad.exe and it runs fine from there, change the env var for EDITOR to that and core.editor to pspad and it doesn't run22:29
$ git commit22:29
error: There was a problem with the editor 'pspad'.22:29
Please supply the message using either -m or -F option.22:29
SRabbelier jnareb: ping22:29
Ilari coachz: If you have core.editor set to 'pspad', does 'pspad' (as such) from command line work?22:30
cxreg coachz: path issue?22:30
← ph^ left22:30
coachz yeah, looks like git is to blame and i need to run it from a batch file as mentioned earlier http://stackoverflow.com/questions/10564/how-can-i-set-up-an-editor-to-work-with-git-on-windows22:30
SRabbelier jnareb: that blogpost is.... wrong, to say the least :P22:30
Ilari Cpp???22:31
jnareb SRabbelier: yep. And no comments means no way to (try) to correct them (like fogetting to exclude contrib/ from software metric, at least)22:31
cxreg SRabbelier: the "5 features" post is wrong too22:31
SRabbelier jnareb: write him an email :P22:31
jnareb: and/or poke on irc22:32
Ilari Wonder what files it detects as CPP and Python. AFAIK, there's not supposed to be any.22:32
cxreg i submitted a correction to that one weeks ago, and was completely ignored22:32
jnareb cxreg: well, perhaps with exception of hg-serve, and "git init directory" (which is in master, he, he , he)22:32
cxreg no doubt the author has no interest in a fair comparison22:32
pardon me, your bias is showing :p22:32
← khmarbaise left22:32
jnareb cxreg: alleged-blog? supposedly has comments, but can't post them?22:32
cxreg they're moderated22:33
jnareb Ilari: contrib/ is not excluded22:33
SRabbelier jnareb: hg-serve doesn't seem very useful to me tbh, look at the usage stats of git-instaweb, why would a git-instaserve be used more :P22:33
jnareb cxreg: moderated is not censored...22:33
SRabbelier: hg-serve == git-instaweb + git-daemon; I think the instadaemon part might be more important22:34
→ naeu joined22:34
jnareb SRabbelier: and usage stats for instaweb aren't that bad, I think22:34
Ilari In version I have, I can find three contrib python scripts. But no idea what that CPP was for?22:34
SRabbelier jnareb: didn't the survey say that like, most people have never used it?22:35
→ ben_h joined22:35
cxreg oh, heh. it looks like the author removed the part i objected to, with no indication that a correction was made22:35
← mw left22:35
SRabbelier omg22:35
the five features22:35
Sensible defaults part22:35
the part about incremental commits22:35
that's just so retarded22:35
just because _you_ don't understand something (because you haven't bothered to try and learn how to use it), doesn't mean it's useless22:36
→ _Vi joined22:36
wereHamster at least he admitted that it's not a fair comparison22:36
jnareb Ilari: CPP (C Pre-Processor ?) is probably misdetection22:37
SRabbelier he did22:37
→ JEEB__ joined22:37
SRabbelier oh, btw, did the meaning of 'HEAD~' change since 1.5.4 and 1.6.4?22:37
it seems that in 1.6.4 'HEAD~' and 'HEAD~1' are the same22:37
but that in 1.5.4 'HEAD~' and 'HEAD' are identical22:38
wereHamster probably a bug in 1.54, because HEAD~ is supposed to mean HEAD~122:38
← kpreid_ left22:38
jnareb SRabbelier: in 1.6.3.3 is the same22:38
← coldwate1lover left22:38
→ coldwaterlover joined22:38
jnareb SRabbelier: I mean HEAD~ == HEAD~122:38
SRabbelier jnareb: right22:38
annoying bug then22:38
SRabbelier mutters and wishes distro's would update to 1.6.x already22:39
← Chaser_ left22:39
Q| SRabbelier: Jaunty has 1.6.0.422:39
jnareb "Quick Local References" i.e. 'hg checkout 18' for '# 18:a432bc' in Mercurial... I don't think that it is very useful at least in succesfull project, where it would be rather 'hg checkout 6280'... or was it 6820 :-)22:41
→ Geralt joined22:41
coachz how obvious ! git config --global core.editor "'C:\Program Files\PSPad editor\PSPad.exe' -m"22:41
that works22:41
SRabbelier Q|: sweet, good thing I'm on Hardy then :(22:41
coachz no batch file needed22:41
SRabbelier jnareb: so true ^_^22:42
← Geralt left22:42
SRabbelier jnareb: sides, it creates ambiguity22:42
jnareb: does 6280 mean r6280 or is it an abbreviation22:42
→ MrUnleaded joined22:42
Q| SRabbelier: :D Maybe you could use a backport? I don't think Jaunty is really worth the update, except for it's fast boot up, maybe.22:42
jnareb SRabbelier: well, author admits that 'These numbers don't travel across clones'22:43
coachz does that help anyone else use their fav editor ? git config --global core.editor "'C:\Program Files\PSPad editor\PSPad.exe' -m"22:43
← JEEB left22:43
Q| SRabbelier: But on the other hand there's simply TO much drawbacks, like that pulseaudio thing. Gnah22:43
SRabbelier Q|: nah, I compiled my own22:43
Q|: it's just annoying that I have to22:43
Q| SRabbelier: ^^22:43
SRabbelier: That's the price I guess ;)22:43
jnareb errr... you means is 6280 shortened shortened SHA-1 or number?22:43
SRabbelier jnareb: yes22:43
jnareb I guess that there is some order of precedence, like for git: branches, tags, and tags named deadbeef ;-)22:44
← jrmuizel left22:44
SRabbelier jnareb: hehhe, that's just asking for trouble ^_^22:44
← mjf left22:44
← naeu left22:45
jnareb but there at least you know that you are shooting in the direction of foot with such names, and how to resolve ambiguity...22:45
← cbreak left22:45
→ gerberb joined22:45
jnareb refs/heads/deadbeef22:46
stdorb s/.*/bbq/22:46
MrUnleaded if one were to start making something similar to git-svn to import from one SCM to git....are there any existing tools to help with that?22:46
SRabbelier jnareb: how about we add a `git push --delete somebranch` ?22:47
jnareb MrUnleaded: git-fast-import22:47
SRabbelier jnareb: as sugar for 'git push :somebranch' ?22:47
MrUnleaded: but wait, there's more22:47
MrUnleaded: what SCM are you speaking of, there are many such efforts already?22:47
jnareb SRabbelier: I think that was proposed already, and Junio or Linus are for that. Now to find suck^W volunteer who submits patch ;-)22:47
SRabbelier MrUnleaded: p4, hg, csv to name a few22:47
jnareb: sure, I'll do it22:48
jnareb SRabbelier: _csv_ ?22:48
SRabbelier jnareb: I'll resend my 'fast-import options' patch in about a week too22:48
jnareb ;-)))))22:48
SRabbelier: I rather liked this patch22:48
jaalto Can I undo a "git stash apply". I mistakenly ran in wrong branch?22:48
SRabbelier jnareb: I'm quite sure there's git<->csv bridges :P22:48
jnareb is working on AJAX-y interactive blame22:49
→ fernandor joined22:49
→ AkumaStreak joined22:49
MrUnleaded SRabbelier: well actually i have been having troubles using a TFS server the TFS has an SVN interface but it doesnt play nicely with git svn so I wanted to see if maybe i could find a way to either import via TFS or maybe analyze whats going on with git-svn/svn communication22:49
SRabbelier jaalto: if you don't have any important changes "git reset --hard" should work just fine22:49
davegb3 jaalto: or just re-stash it? "git stash"22:49
jnareb jaalto: or stash again ;-)22:49
jaalto SRabbelier: Ok, but doesn't "stash apply" move stash pointer backward. I mean, can I still re-apply in another branch?22:50
SRabbelier jnareb: that would create two stash entries22:50
jaalto: no, you're confused with 'git stash pop'22:50
← fernandor left22:50
SRabbelier MrUnleaded: mhhh, I dunno about TFS, I think there's a MVSS one maybe22:50
MrUnleaded: but TFS I don't recall22:51
→ djgera joined22:51
jnareb MrUnleaded: I don't think there is TFS importer for git. Perforce and ClearCase, yes. See also http://git.or.cz/gitwiki/InterfacesFrontendsAndTools for (incomplete) list of various importers.22:51
jaalto SRabbelier: I see now that "stash list" has it still. Good.22:51
← ryandotsmith left22:51
MrUnleaded ok thanks22:51
→ cbreak joined22:52
SRabbelier MrUnleaded: btw, writing an exporter is extremely easy :D22:52
MrUnleaded: you can get something basic working in about a weekend22:52
MrUnleaded: so if it's important to you, give it a shot22:52
MrUnleaded: see http://www.kernel.org/pub/software/scm/git/docs/git-fast-import.html for the format22:52
← JEEBcz left22:53
MrUnleaded SRabbelier: ok thanks22:53
jnareb SRabbelier, MrUnleaded: the problem might be with fast _extracting_ data from TFS. Is there any language binding / library for it?22:53
JEEB__ → JEEB22:53
stdorb TFS = ?22:54
MrUnleaded microsofts attempt at a SCM22:55
SRabbelier jnareb: true enough22:56
jnareb stdorb: TFS == MS Team Foundation Server. Something MS actually use themselves, as opposed to the mockery that is MS Visual SourceSafe...22:56
← ben_h left22:56
→ ben_h joined22:57
← juanjoc left22:58
→ drizzd joined22:58
coachz VSS is a toy22:58
Nafai VSS is awful22:59
← bgerber left22:59
cxreg jnareb: i'd rather use csv as version control than cvs22:59
stdorb haha23:00
coachz but it corrupts at no extra charge23:00
→ kpreid_ joined23:00
stdorb cxreg: that could almost come from Linus itself :p23:00
← cannonball left23:00
Fissure i realized a while ago that i feel about itunes the same way linus feels about cvs23:01
SRabbelier Fissure: how's that?23:01
coachz like you're being screwed23:02
Fissure well, not *quite*, but i feel like lots of player authors want to be just like itunes and wind up cloning all the bad things about it along with the good23:02
coachz i dragged a video iso file onto vlc player today and it played it, blew me away23:02
even the menus worked with the mouse23:02
Fissure like, why the hell is album art downloads put as a higher priority than actually playing music properly instead of introducing random skips?23:03
coachz haha23:03
you mean like why does windows 7 start skipping like crazy when you plug in an external sata drive during playback? hahahaha23:03
they never learn23:04
Fissure i'm talking about linux players, actually23:04
→ fernandor joined23:04
coachz but the just don't prioritize the playback of media properly23:04
Fissure try listening to dark side of the moon; the half-second breaks between tracks somewhat kills the effect23:04
coachz no kidding23:04
davegb3 → davegb3|away23:04
coachz but it makes rap sound normal23:04
davegb3|away → davegb3|away|23:04
Fissure eh?23:04
coachz instead of "yo baby" you get "yo yo yo yo yo baby"23:05
and it sounds the same23:05
← fernandor left23:05
← parasti left23:06
jnareb http://xkcd.com/411/ (well, it is about techno, but same difference)23:07
← boto left23:08
Fissure ha, wasn't talking about that kind of skipping23:08
→ montylounge joined23:08
→ bcardarella joined23:08
Fissure can't remember actually having that happen on linux23:08
← drizzd_ left23:08
→ gma joined23:09
← wshimmy left23:10
← summatusmentis left23:10
jnareb Let's somebody remind me why we allow Perl but forbid Python (in git core, not in contrib)?23:11
→ summatusmentis joined23:11
stdorb python seems slower :p23:12
SRabbelier jnareb: cos... perl is old... and... linus likes old?23:12
SRabbelier has no clue23:12
← agile left23:13
→ Rolenun joined23:13
SRabbelier jnareb: oh, and because Dscho doesn't like it :P23:13
← montylounge left23:13
coachz my pspad editor is working for commit messages....yay23:13
jnareb well, that might have been different in 2005, when Git started: portability and having Python installed23:13
stdorb in 2005? My system did not require a python interpreter (distro config scripts perl or something)23:13
← jigfox left23:13
jnareb and because Git is not mercurial, eh... :-PPPPP23:14
stdorb merc has strange terminology.23:14
SRabbelier jnareb: maybe because we want to not have too many dependencies?23:14
stdorb CVS has files, dirs, commits, tags, SVN has them too, Git has blobs, trees, commits, tags and Hg has , uhm remind me, manifests and something23:15
→ wshimmy joined23:15
jnareb what's one more dependency ;-) besides, we can^Wcould get rid of Perl ;-PPPPPP23:16
SRabbelier jnareb: yesplease.jpg23:16
stdorb make it in C23:16
jnareb stdorb: Mercurial has manifests (like trees, but not hirearchical) and changerevs (like commits). It doesn't have tag objects (and lightwieght tags have very... hmmm... 'interesting' implementation)23:17
SRabbelier git-add--interactive.perl git-difftool.perl git-relink.perl git-send-email.perl git-svn.perl23:17
jnareb: wha'ts relink even do?23:17
jnareb SRabbelier: and gitweb.perl23:18
← summatusmentis left23:18
SRabbelier jnareb: you don't seriously want to rewrite that though?23:18
→ summatusmentis joined23:18
stdorb jnareb: that's what I meant. It's too different from Linus's thinking (which, eventually, led to Git)23:19
→ jrmuizel joined23:19
jnareb SRabbelier: relink looks like relic from the times of unpacked repositories (and "git clone --local" hardlinking)23:19
coachz one day i would like to know enough to give a talk on git and i would teach commits by having a guy come in with a flash camera and take snapshots and yell commit each time and then leave the room....hehe23:19
jnareb SRabbelier: instead of redoing "git clone --local" you can git-relink... (or use alternates)23:19
SRabbelier jnareb: ok, so no-one ever uses that :P23:20
jnareb: that leaves 'add -i', which I <3 :P23:20
so want to keep that :P23:20
jnareb SRabbelier: let me check...23:20
← Turias left23:20
stdorb Git now has so much porcelain that the cupboard on the wall runs risk to crash down.23:20
SRabbelier and send-email could be easily rewritten in python too23:20
→ Tapnzout joined23:20
SRabbelier stdorb: true :)23:20
→ d0k_ joined23:21
→ agenteo joined23:22
← stringo0 left23:22
← Tapout left23:22
jnareb SRabbelier: 5 often, 11 sometimes, 70 rarely / 2406 in this year survey: http://www.survs.com/shareResults?survey=2PIMZGU0&rndm=678J66QRA223:22
→ danggun joined23:22
→ Tapout joined23:23
jnareb SRabbelier: I guess that 'add -i' would get rewritten in C as soon as API stabilizes and features freeze23:23
SRabbelier jnareb: which one?23:23
← summatusmentis left23:23
→ montylounge joined23:23
stdorb how long is the survey still open?23:23
danggun Hey everyone. I accidentally deleted a directory from master while also making changes in master. What's the best way to handle this? Should I commit the changes and then cherrypick (?) from the previous commit?23:23
jnareb errr... that was about git-relink (question 17 = 16 cont)23:23
SRabbelier danggun: 'git checkout path/to/deleted/directory'?23:24
err23:24
danggun: 'git checkout -- path/to/deleted/directory'?23:24
(you need the --)23:24
jnareb stdorb: survey is open till 15 September (1 month and a bit)23:24
← RaceCondition left23:24
SRabbelier jnareb: oh, the relink one, right23:24
← jjuran left23:24
stdorb jnareb: question 2 should not have been a freeform text23:24
← MisterN left23:24
Hideki is the survey result considered to be "publicized" info, or just internal to IRC?23:24
danggun Sweet! But I'm curious. Now all the directories in that tree have a 0 value instead of 4096. What does that mean?23:25
SRabbelier Hideki: there's no such thing as "internal to IRC"23:25
danggun I do have all my fiels and they have data but the directories say 0.23:25
files*23:25
jnareb hmmm... looks like git-relink is no less rare than git-cvsserver and git-imap-send...23:26
Hideki SRabbelier: I know what you mean, what I meant is if someone will be mad if I write some remarks about it on my blog, etc.23:26
← jrmuizel left23:26
← prusnak left23:26
jnareb Hideki: you mean http://www.survs.com/shareResults?survey=2PIMZGU0&rndm=678J66QRA2 URL? this is ongoind on-the-fly analysis of current results23:26
SRabbelier Hideki: no, we won't :)23:26
almostautomated → almostautomated|23:26
SRabbelier Hideki: as jnareb said though, survey is still open23:26
coachz i created a repo, added all the files and committed them. at this point all my files are considered tracked right ? and I will have no files staged right ?23:27
← Tapnzout left23:27
Hideki ok, I will hold off until survey's over :-)23:27
→ qrush joined23:27
jnareb Hideki: but please if you provide some analysis of survey, or partial survey results, post on git mailing list an announcement at least (git@vger.kernel.org, doesn't require to be subscribed), and add link to http://git.or.cz/gitwiki/GitSurvey2009 wiki page, please23:28
← d0k left23:28
Hideki I'm mainly advocating git to Japanese speaking demography and I'm always looking for some materials :-)23:28
jnareb: will do :-)23:28
jnareb Hideki: even if it is not in English23:28
stdorb We will assimilate and translate your review. Resistance is futile... :p23:29
← mithro left23:29
jnareb Hideki: Junio C Hamano, git maintainer, IIRC is publishing some Git book in Japaneese (c.f. http://gitster.livejournal.com)23:29
→ summatusmentis joined23:29
→ plediii_ joined23:30
jnareb Hideki: after survey ends I would publish CSV and Excel (XLS) export of data, similar to http://git.or.cz/gitwiki/GitSurvey200823:30
jaalto I'd like to ignore STDOUT, but send STDERR to a filter: "command > /dev/null | post-process-stderr" What do I need?23:31
Hideki jnareb: yeah, I think there were fairly big coverage of git in the magazine, unfortunately, I'm catering to Japanese demography from the US, so I don't have a copy :-(23:31
→ stianhj joined23:32
← klbate left23:32
almostautomated| Hi all! Could someone help me troubleshoot a patch issue I'm having with alpine? The patch works fine if I save locally and apply it, if I import and then directly export it, yet if I send it to myself it wont apply. :/23:32
http://git.pastebin.com/d3fef76d23:32
The highlighted line needs @@ added to the front because of pastebin's substitution of @@ for a highlighted line23:32
ohhh, that is the version that worked locally... I'll grab the sent to self23:33
Hideki jnareb: interesting thing I found about the survey, the answering survey itself was quite informative; for trying out commands I've never used before :-)23:33
jnareb jaalto: nothing, it should work (perhaps '> /dev/null' --> '1> /dev/null')23:34
→ plediii__ joined23:34
← summatusmentis left23:35
Ilari almostautomated|: Diff the pre- and post-send versions of patch and see what's the difference is?23:35
jnareb suddenly remembers that he should take the survey he created (with help of git users)...23:35
→ summatusmentis joined23:35
SRabbelier jnareb: lol, you didn't answer it yet?23:35
coachz i created a repo, added all the files and committed them. at this point all my files are considered tracked right ? and I will have no files staged right ?23:35
← MrUnleaded left23:36
jnareb SRabbelier: nope, I forgot23:36
well, I try to prepare for free-form essay questions, but...23:36
coachz: after "git commit" or "git commit -a23:37
coachz: you should not have any staged changes, yes23:37
coachz just git commit cuz i already added all the fiels23:37
files23:37
so how do i now remove a file from being tracked without deleting it from my working dir23:38
← crack_head left23:38
jnareb coachz: "git rm --cached <file>" I think is what you want (and then commit removal, of course)23:38
→ plediii___ joined23:39
coachz doesn't that remove it from staging23:39
but i dont have anything staged23:39
jnareb coachz: this removes it from index (stages removal)23:40
coachz but if i don't have anything in my index how can it be removed23:40
almostautomated| Ilari: it is showing differences... when I pipe the diff to a file and then cat -E it I'm seeing a '$' marker on each line marked as a difference...23:40
jnareb 'nothing staged' deoepnds on the definition of staged, if it is for example staged == version in index / staging area is != version in working directory23:41
coachz i guess i just don't get it then23:41
jnareb coachz: after commit you have in index what you have in commit == what you have tracked in working directory23:42
coachz i created a repo, added all the files and committed them. at this point all my files are considered tracked right ? and I will have no files staged right ?23:42
jnareb coachz: after commit working area == index == HEAD23:42
coachz after commit, doesn't the index / staged area become empty ?23:42
ahhh23:42
→ plediii____ joined23:42
jnareb coachz: no, it becomes same as commit (HEAD)23:42
→ azeroth-work joined23:42
jnareb staged file <=> file with version in index different23:43
coachz so to remove it from tracking i would remove it from staged as you mentioned cool23:43
→ summatus1entis joined23:44
← summatusmentis left23:44
coachz thanks for helping a noob out jnareb !23:44
← Tuomas left23:45
jnareb coachz: you are welcome23:45
← plediii left23:46
Hideki are there any hooks executed for when executing 'git status"23:47
← juan_arandaalvar left23:47
Fissure nope23:47
considering it's a non-mutating operation, i don't know why you'd want one23:48
Hideki Fissure: ok, well, I'm just trying to ponder some cross-platform filename conversion, because, Linux uses UTF-8 for Japanese file names, and Windows uses cp920 or something (basically Shift-JIS, I believe)23:48
and git stores these index in byte-stream23:49
← plediii_ left23:49
← DavidKlein left23:49
← agenteo left23:49
jnareb Hideki: i18n.commitEncoding ?23:50
Hideki: file namse are stored as byte stream23:50
Hideki: with no encoding information23:50
Hideki jnareb: yeah, so I thought it had to be converted at front-end, but i18n.commitEncoding does that?23:51
← Guest89136 left23:51
Hideki Windows filename <--> Shift-JIS to/from UTF <--> git <--> repository <--> git <--> UTF to/from Shift-JIS <--> Linux filename23:51
← coldwaterlover left23:52
← drizzd left23:52
almostautomated| Ilari: duh, I gues cat -E _would_ show the '$' symbols. Here's the pastebin diff of the local vs sent to self:: http://git.pastebin.com/pastebin.php?diff=d4812978923:52
jnareb Hideki: errr... no, i18n.commitEncoding is about commit encoding. Git treats file names (and file contents) as bag of bytes, and doesn't store encoding information anywhere. Which means that you are f*** if you work on filesystems with different filename encoding (and non-ASCII filenames).23:52
stdorb it would be much easier if you just used UTF on Linux23:52
kibombo if i want to start cloning and committing from a git repo, how do i do that?23:53
from and to23:53
stdorb kibombo: you just clone23:53
kibombo git clone <git-url>23:53
then what?23:53
← hyperair left23:53
← jceb81 left23:53
kibombo git init?23:53
stdorb then you start vi :p23:53
kibombo or git push origin master23:53
Hideki stdorb: well, it's more of it'd be much easier if Windows uses UTF :-) I am in fact using UTF on Linux23:54
wereHamster kibombo: it would probably be best if you first read one of the tutorials on http://git-scm.com/23:54
jnareb Hideki: there are some beginnings of code to deal with filesystem encoding issues, IIRC currently only for case-preserving case-insensitive filesystems, but planned also for MacOS X nonstandard NDF decomposition of Unicode23:54
stdorb Hideki: last time I checked, Windows was using UTF-16 for its filesystem stuffage on my side23:54
← plediii__ left23:54
coachz git clone <git-url> brings the repo down so you don't need to do git init right guys ?23:54
kibombo well, after moving a file into the dir, i know you do git add <file> ; git status; git commit -a -m "comment"; then what?... git push origin master?23:54
stdorb coachz: git-clone = git-init + git remote add + git fetch + git checkout ...23:54
→ smotchkkiss joined23:55
Hideki stdorb: internally, yes, but they can't lose it on the front-end because of compatibility issue, I think23:55
kibombo wereHamster: i've read some tutorials on git already23:55
stdorb kibombo: The manual page, quick!23:55
coachz stdorb, nice23:55
stdorb Hideki: frontend?23:55
jnareb Hideki: and there was idea (but not accepted or no implementation) of pre-add hook23:55
Hideki stdorb: cp932, is the actual encoding they use...23:55
kibombo my problem is..23:55
Fissure kibombo: that should work, though you should just be able to 'git push'23:55
kibombo fatal: protocol error: expected sha/ref, got23:55
→ crack_head joined23:55
wereHamster kibombo: what command did you run?23:55
kibombo git push23:55
Fissure kibombo: what kind of url did you clone?23:55
kibombo git push origin master does the same23:55
← ryanakca left23:56
Hideki stdorb: cp for code page, I think23:56
stdorb well yeah but codepage becomes pretty irrelevant if you use CJK in filenames23:56
kibombo git://gitorious.org/project/sub.git23:56
← summatus1entis left23:56
→ summatusmentis joined23:56
kibombo Fissure ^23:56
Fissure kibombo: you can't push to git://23:56
wereHamster kibombo: you can't push through git://, you have to use ssh://23:56
stdorb and since I can store Unicode chars in Windows filenames, there's no codepage crap I am bound to. Sort of23:56
kibombo so..?23:56
Fissure well, you *can* but only if it's enabled and people would be crazy to do that on a publicly-accessible server23:56
Hideki jnareb: I see, yeah it seems to be quite weird one, as it won't be problem if you are pushing/pulling across same OSes23:57
kibombo do i have to rename remote or something?23:57
wereHamster kibombo: if you got to the project homepage on gitorious, there's the 'Push URL'23:57
Fissure kibombo: you can just edit the url in .git/config23:57
Hideki stdorb: looking at how exactly it handles, empirically this has been a problem23:57
stdorb I may need to inspect how cygwin deals with Unicode filenames tho23:57
→ ProperNoun joined23:57
wereHamster kibombo: in your local repository, run: git config remote.origin.url <the-push-url-from-gitorious>23:57
kibombo the push url given is git://gitorious.org/project/sub.git23:58
Fissure that's... not right23:58
← plediii___ left23:58
← bcardarella left23:58
kibombo i meant... [email@hidden.address]23:58
there23:58
jnareb Hideki: or if all filesystems use UTF-8 (in sane way)23:59
← gma left23:59
wereHamster kibombo: see, totally different URLs ;)23:59
jnareb Hideki: or you don't use non-portable filenames23:59
kibombo there it goes :)23:59

Logs Search ←Prev date Next date→ Channels Documentation