IRCloggy #git 2012-06-28

Logs Search ←Prev date Next date→ Channels Documentation

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

2012-06-28

← blendedbychris left00:00
mns its accessible from the internet. so I work on ~/.emacs.d, make changes, do git commit, no need to do a push. I go to work, do a git pull and I should have my changes.00:01
→ simesy joined00:01
← simesy left00:01
← ben_alman left00:01
cmn yes, pull will download and merge the changes00:02
→ intransit joined00:02
cmn you won't be able to push to your ~/.emacs.d's active branch directly00:02
sitaram cjz: you didn't have to look in the log file. The push output should have shown you a nice "FATAL: git config .... not allowed"00:03
cjz sitaram: it did00:03
im just a dummy00:03
sitaram cjz: then why would you think your permissions and repos would work if you got a FATAL on the push?00:03
shouldnt you investigate *that* before going into chmod and all that?00:04
cjz well i thought it was doing everything just warning me about something00:04
i was going to look at it later..00:04
mns cmn: thanks for helping and clearing up things00:04
sitaram cjz: you have a different meaning of FATAL -- in all caps too00:04
sitaram notes this down for future support issues00:04
cjz yes thats probably true00:04
again, im a dummy00:04
i should have figured it out00:04
but cmn helped.. thank you again00:05
sitaram cjz: either that or English is not your first language. The latter is common enough00:05
cjz nah, i just assumed that FATAL was on something else00:05
sitaram does not actually *say* it but the former is even more so00:05
cjz like some post_hook…00:05
sitaram who cae *what* it was on. A FATAL is something tobe investigated00:06
gaaah!00:06
cjz ok ok00:06
sitaram gives up on humanity00:06
cjz anyway, have to head home00:06
← intransit left00:06
→ davidbanham joined00:06
Owner how do you remove curse words from prior commit messages00:06
cjz its ok, im not a systems engineer, so maybe you can give up on devs doing system engineering work00:06
nevyn Owner: rewrinting it is the only real option00:07
cjz anyhoo, see you tomorrow00:07
← cjz left00:07
← avpx left00:07
Owner nevyn~# ??00:07
→ enmand joined00:07
nevyn rewriting rather00:07
so you rewrite that commit.00:07
how far back is it?00:07
→ orzo joined00:08
Owner many are far back00:08
how would you automate it00:08
basicallyu do you guys know if egit ignores .gitattributes file?00:08
orzo so, trivial question, I have some changes i just made and realize now i want to save them in a new branch rather than the current one. What do i do?00:08
nevyn orzo: git branch -b newbranch00:09
oops00:09
← AaronMT left00:09
nevyn git checkout -b newbranch (make a new branch of what you've got now)00:09
→ ben_alman joined00:10
nevyn git checkout oldbranch (go back to where it shouldn't be but is)00:10
git reset --hard HEAD^ (reset back one commit)00:10
← twisted` left00:11
→ AaronMT joined00:11
← AaronMT left00:11
← oriba left00:11
← Myhro left00:12
→ AaronMT joined00:12
← AaronMT left00:12
orzo okay, well ive not commited anything anywhere, so i think all i want is that first command "git checkout -b newbranch" to start a new branch with altering the working tree and then i would commit -a. right?00:12
s/with/without/00:12
nevyn right00:12
→ AaronMT joined00:12
← AaronMT left00:12
nevyn I thought you'd committed.00:12
yes git checkout -b and commit00:13
→ twisted` joined00:13
→ AaronMT joined00:13
← AaronMT left00:13
← squentin left00:13
orzo thanks00:13
→ AaronMT joined00:14
← AaronMT left00:14
← w0m left00:14
← Mowsh left00:14
→ AaronMT joined00:14
← AaronMT left00:14
→ AaronMT joined00:15
← AaronMT left00:15
→ AaronMT joined00:16
← thornofmight_ left00:16
basicallyu I am badly failing to use "* text" inside .gitattributes in the root of the workdir and actually have the .gitattributes itself stored in the repo on commit with LF lineendings , they are not touched00:17
but I am using egit...00:17
→ Myhro joined00:18
← cakehero left00:19
→ squentin joined00:20
← squentin left00:20
→ squentin joined00:20
basicallyu now i notice that if i run git status it shows a lot of modified files which aren't seen by egit meh00:20
← cbreak left00:21
→ w0m joined00:21
← AaronMT left00:22
← Davey left00:23
bitkiller regarding my problem with pub key copying to the gitolite host, i found that copying the key using ssh-copy-id does work, but with scp doesn't00:23
basicallyu I'm taking it that means egit doesn't consider .gitattributes ... meaning there's no way i can ensure that devs will commit files with LF endings regardless00:23
bitkiller work -> ssh-keygen -lf validades the key00:24
sitaram you said it was already valid before00:24
→ oriba joined00:24
← Sebo left00:27
→ jcromartie joined00:28
← lightcap left00:28
← dreiss left00:31
→ dmiles_afk joined00:34
← kenperkins left00:34
bitkiller sitaram, yes...and i have just found the problem00:35
thanks to #openssh guys00:35
selinux disabled in the host, enabled in the VM00:35
← beatak left00:35
← adamm left00:35
→ cody_ joined00:35
← DrNick left00:36
→ DrNick joined00:37
→ macabre joined00:37
dmiles_afk i have this git repository I am trying to make using git2svn ... but I only want a sub-bit of the SVn repository00:37
should i make the entire reprository to GIT then.. somehow only use what i want?00:37
or is there a way of only snatchign the sub part of the SVN00:38
offby1 dmiles_afk: I don't know, but if you get stuck, google for "reposurgeon"00:38
it's advanced but might be useful anyway00:38
← harish left00:39
dmiles_afk looks good.. i cna probly make the whole thing into the GIT .. then use reprosurgious to trim it up00:41
i have thjis project on googlecode that has a submodule thqt deserves its very onw project on github00:42
deserves its very own project on github*00:42
← berserkr left00:42
dmiles_afk but the submodule/subproject is like 4% of the content and history00:43
bitkiller sitaram, i suppose i can place the repositories directory anywhere. am i correct?00:44
orzo i have a situation where there are two popular forks of the same software maintaining separate histories. Although they are both forks of the same base code and they both have git histories, they are two indepently created repositories (not cloned). There's been many changes and particularly name-changes for files and directories. Still, I'd like to merge changes from either into a single00:46
combined repository. Is that doable?00:46
perhaps having my local repository having two separate remote tracking branch anda personal branch that i can merge patches into?00:48
→ luxurymode joined00:48
← keltvek left00:49
→ beneggett joined00:50
→ carpediembaby joined00:50
carpediembaby hi. if i create a new branch on my local machine and commit and then push my changes, does it also create the same new branch on the remote server?00:51
cmn carpediembaby: see man git config push.default00:51
gitinfo carpediembaby: the git-config manpage is available at http://git-scm.com/docs/git-config.html00:51
← andref_dias left00:51
orzo that's not typically tehdefault i do not think00:51
cmn if you push to a new branch explicitly, it will create it00:51
→ Hemebond joined00:51
← oriba left00:52
cmn orzo: you can try to merge histories even if they're unrelated, but the chances of success aren't that great00:52
Hemebond Why would "git fetch upstream" not fetch the latest commits?00:52
→ keltvek joined00:53
cmn that depends completely on what upstream is, and what you mean by latest commits00:53
and whether you're using the term fetch correclty00:53
Hemebond I mean that exact command in quotes.00:53
carpediembaby cmn: what precisely do i look at in the config file? i just don't want to push to master. so i just say, git push my-branch?00:53
Hemebond And using "git log" to see what commits it contains.00:54
cmn carpediembaby: what I told you, push.default00:54
it explains what happens00:54
Hemebond: git log only? that won't be affected by a fetch unless you're in a bare repo00:54
→ AaronMT joined00:54
Hemebond I mean "git log upstream/master"00:55
cmn actually, forget about the bare repo, the settings would have to be changed for that to do anything sane00:55
Hemebond: and you think there's stuff missing why?00:55
← jcromartie left00:55
Hemebond Diff shows wrong changes.00:55
cmn does git ls-remote upstream refs/heads/master and git rev-parse refs/remotes/upstream/master return the same00:55
Hemebond: what's diff got to do with this?00:55
Hemebond Trying to compare local branch to the upstream branch00:56
cmn that is unrelated to fetch working00:56
Hemebond git diff master upstream/master00:56
Really?00:56
cmn does git ls-remote upstream refs/heads/master and git rev-parse refs/remotes/upstream/master return the same00:56
?00:56
Hemebond Sorry, trying to help someone else with this.00:57
It seems we had the fetch command wrong.00:57
We were doing "git fetch upstream master"00:58
cmn that won't update the remote-tracking branch00:58
Hemebond What does it do?00:58
cmn it fetches the master branch and doesn't store it anywhere other than FETCH_HEAD00:58
because you told it not to store it anywhere00:58
← w09x left00:58
Hemebond Oh, we did?00:58
cmn yes00:59
→ w09x joined00:59
Hemebond What do you add to specify where to store it?00:59
cmn a right-hand side to the refspec00:59
see your config file00:59
EugeneKay !fetchfour01:00
gitinfo Never use the four-word version of git-fetch or git-pull (e.g. git fetch remote refspec). It always ends in tears. Yes, if you understand the implications of FETCH_HEAD it can technically be done, but really it is easier to just fetch the whole remote (or perhaps edit the fetchspec if you never want other bits).01:00
EugeneKay I thought there was a writeup attached to that. Hrm.01:00
EugeneKay must need more booze01:00
Hemebond aha01:00
Okay.01:00
cmn man git fetch should say about refspecs01:00
gitinfo the git-fetch manpage is available at http://git-scm.com/docs/git-fetch.html01:00
← etcetera left01:00
EugeneKay The four-word version gives you FETCH_HEAD, which, if you look at the output, it states.01:01
Hemebond I see01:01
so01:01
← fuzzybunny69y left01:01
Hemebond +refs/heads/*:refs/remotes/upstream/*01:02
EugeneKay Correct01:02
Hemebond says to store the stuff from refs/heads into refs/remotes/upstream/01:02
EugeneKay Bravo for not using 'origin'01:02
← acl_ left01:02
← keltvek left01:03
Hemebond I don't really understand what our "master" argument actually told it to do.01:03
(if anything)01:03
← neurodrone left01:03
EugeneKay Grab the master branch from the remote named upstream and put it into FETCH_HEAD01:03
cmn not to store it anywhere01:03
Hemebond Oh01:03
cmn you never gave it the right-hand side01:03
Hemebond Could we have put a refspec after it?01:03
EugeneKay Not anywhere permanent, anyway01:03
cmn you did put a refspec01:03
EugeneKay Sure. master:bacon01:03
cmn that's why you overrode the config01:04
← duckxx left01:04
Hemebond Ah, I see.01:04
It was the left side of a refspec.01:05
Is it possible to see the currently stored... stuff?01:05
bitkiller sitaram, well, i've read the g2 migration doc and found the REPO_BASE change. i'll use symlink as suggested01:05
Hemebond Oh01:05
It's not permanent, so it's lost at the end of the fetch?01:05
cmn as fetch told you, it stored the ref in FETCH_HEAD01:07
→ zheng joined01:07
→ keltvek joined01:08
→ aantix_ joined01:08
Hemebond Right.01:08
Ah, neat.01:08
← aantix left01:09
← mtx_init left01:11
← xinming left01:11
← erkules left01:11
← aspotashev left01:11
→ erkules joined01:11
→ Vampire0 joined01:12
Hemebond Thank you, again, for the help :-)01:12
← wilmoore left01:12
→ xinming joined01:12
← Hemebond left01:13
→ mtx_init joined01:13
← aantix_ left01:13
← Vampire0_ left01:15
← kanch left01:15
→ kanch joined01:16
← ben_alman left01:17
→ iori joined01:17
→ ben_alman joined01:18
← smw left01:18
→ smw joined01:18
→ v0n joined01:19
carpediembaby so i get that i need to create a remote branch. now, i have one remote branch and another local branch which tracks the remote. how do i now push changes only to the new remote branch?01:19
← Lgb left01:20
← palm_vue left01:20
← palyboy left01:20
→ palm_vue joined01:21
→ Lgb joined01:21
← osmosis left01:21
frogonwheels carpediembaby: I just read back up, and I'm still not quite sure what you are asking by that.01:22
→ cooloney joined01:22
cmn tell push to only push that one branch01:22
git push <where> <what>01:22
← SkAzZ left01:22
→ kenperkins joined01:24
carpediembaby frogonwheels: i created a remote branch (git push origin origin:refs/heads/new_feature_name) and also a local branch which tracks the remote one. but i don't know how to push changes i make in the new local branch to the new remote branch01:24
frogonwheels carpediembaby: wow, you like doing it the hard way don't you !01:25
→ palyboy joined01:25
carpediembaby cmn: git push origin local_branch ?01:25
frogonwheels carpediembaby: ..and that syntax doesn't look right anyway.01:25
cmn if the name is the same, you can omit the right-hand side01:26
← hwrd|work left01:26
frogonwheels carpediembaby: and unless you have a very unusual case, you should keep it the same. things will run better.01:26
carpediembaby in my frogonwheels what's the easy way?!01:27
frogonwheels carpediembaby: git push -u origin local_branch (set upstream as well)01:27
carpediembaby: if you need a different name on the remote, then create the branch with the same name locally01:28
carpediembaby cmn, i'm not sure i understand. you mean if the name of the remote branch and the local branch is the same? i can just say "git push origin" and it will push from my current branch to the corresponding branch on remote?01:28
cmn no01:28
← xxthink left01:28
cmn for that, see man git config push.default01:28
gitinfo the git-config manpage is available at http://git-scm.com/docs/git-config.html01:28
frogonwheels carpediembaby: git push origin local_branch is what he. This is the equivalent of git push origin branch_name:branch_name .. or close to.01:28
← diamonds left01:29
frogonwheels he *meant01:29
err.. cmn meant.01:29
cmn yes, if you omit the right-hand side of the refspec, it will assume you mean the same name01:29
just as you can omit the refs/heads/ because you usually push branches01:30
→ harish joined01:30
carpediembaby cmn, frogonwheels : what i am confused about is the use of origin. i created a branch named 'guides'. and i can't use that as a remote. in my .git/config file, 'guides' is there with a remote = origin01:31
frogonwheels carpediembaby: you mean you can't use that as a remote branch name?01:31
cmn a remote isn't a branch01:31
frogonwheels carpediembaby: a remote is a reference to a repository01:32
carpediembaby frogonwheels: yes, from what i'm thinking, i should be saying git push guides01:32
← zheng left01:32
← Targen_ left01:32
frogonwheels carpediembaby: you have to specify where you push it to (which repository, or 'remote')01:32
carpediembaby: so git push origin guides (if you are using the default remote)01:32
→ centipedefarmer joined01:34
→ intripoon_ joined01:34
→ apok_ joined01:35
→ jhardinee joined01:35
← apok_ left01:35
← w09x left01:36
→ w09x joined01:36
← intripoon left01:37
→ etcetera joined01:37
← apok left01:38
← mocramis left01:38
→ mocramis joined01:39
→ Davey joined01:39
→ metcalfc joined01:40
→ jackh joined01:42
→ jackh_ joined01:44
infogulch → infogulch|afk01:44
→ kvda joined01:44
← alegen left01:44
← mcluskydodallas left01:45
← tonsofpcs left01:46
→ jpmelos joined01:47
← segher left01:50
← thearthur left01:50
→ mritz joined01:50
← jpmelos left01:50
→ jpmelos joined01:51
→ segher joined01:51
← carpediembaby left01:53
→ tonsofpcs joined01:53
← antgel left01:55
→ fuzzybunny69y joined01:57
← jpmelos left01:59
← basicallyu left01:59
→ macrover joined02:01
← cody_ left02:02
→ _d4rklit3 joined02:03
→ ferdna joined02:06
→ delight joined02:07
← metcalfc left02:07
→ xiangfu joined02:08
← davidbanham left02:09
← kukks left02:09
→ jpmelos joined02:09
→ tmcmahon joined02:10
← jeyanth left02:10
→ davidbanham joined02:10
→ nadavoid joined02:10
← zanneth left02:11
← mtytel left02:14
→ weiyang joined02:16
→ Anxi80_ joined02:16
→ dhruvasagar joined02:16
← btanaka left02:17
← Anxi80_ left02:19
→ Ecesis joined02:21
← lolfrenz left02:22
← Ecesis left02:22
← davidbanham left02:22
← frogonwheels left02:23
→ davidban_ joined02:23
← JEEB left02:24
→ JEEB joined02:25
← Bombe left02:25
→ Targen joined02:26
→ jnewt joined02:27
← Fandekasp left02:27
← v0n left02:27
→ Fandekasp joined02:27
← acrocity left02:28
jnewt what is the correct way to reset a folder to a previous commit? I have some support headers / source files that i want to "rollback" to a previous state (I alsow want to remove created and add deleted since a specific commit) I don't want to lose modifications to other folders02:29
← weiyang left02:29
← mdhopkins left02:30
← nadirvardar left02:31
cmn checkout in file mode02:31
checkout commit -- dir02:31
→ leeight_ joined02:34
← SirFunk left02:35
← fuzzybunny69y left02:36
→ Parth_ joined02:36
← xinming left02:36
jnewt i take it that can't be done via git gui, or there's no support here or anywhere else for non-command line git02:37
cmn git-gui can probably do it somehow02:37
→ SirFunk joined02:38
→ edude03 joined02:38
cmn but I never use it; and it's not like checkout is compliated02:38
→ wilmoore joined02:38
← TheDudeKnows left02:38
jnewt i never could figure out how to use git from windows and the command line. works fine when i'm using my linux computer, but most of my work is on windows based software02:39
← MattRichardson left02:39
← delight left02:39
cmn you either tell the installer to put it in your usual path; or you click the so-called "git bash"02:39
→ v0n joined02:40
→ delight joined02:41
→ weiyang joined02:42
← jhford left02:43
→ neurodrone joined02:44
→ blendedbychris joined02:45
← blendedbychris left02:45
→ blendedbychris joined02:45
→ mvrilo joined02:46
→ thursday joined02:47
→ Bombe joined02:48
→ MattRichardson joined02:48
← mvrilo left02:48
→ jhford joined02:49
← macabre left02:50
← redii left02:50
← AaronMT left02:50
jnewt i can't believe you can't copy that commit code from git bash and have to type the whole thing02:51
← bgerber left02:51
← Parth_ left02:52
← edude03 left02:53
← simesy_work left02:53
→ macabre joined02:54
→ xinming joined02:55
→ simesy_work joined02:55
cmn of course you can02:55
cmd is just stupid and you have to activate edit mode02:55
→ bgerber joined02:55
← dhruvasagar left02:56
← jackh_ left02:56
← ehsan left02:56
← williamcotton left02:57
jnewt i can't even type in this thing, it's aweful. i start to type at $, and my typing jumps half way up the scrollback. what a pos02:58
← Anxi80 left02:58
cmn what?02:58
jnewt this git-bash is the most unfriendly thing i've ever used.02:58
cmn it's cmd02:59
← iori left02:59
cmn it's just what git gives you02:59
milki try linux. i hear its great.02:59
cmn what Windows gives you02:59
jnewt i use linux milki, it is great, but it doesn't run 90% of the software I need02:59
← relurk left03:00
milki virtualization!03:00
→ iori joined03:00
jnewt milki: ever tried to virtualize solidworks. good luck03:00
cmn use powershell if you like it better03:00
mns you're using it for CAD type work jnewt ?03:00
milki doesnt have hardware that can handle virtualization :P03:01
jnewt mns: i'm using git for embedded source code, but yes, i do hardware design as well03:01
→ nareshgbhat joined03:02
mns jnewt: I was presuming that when you said you couldn't run 90% of the applications you need and then you mentioned solidworks ..03:02
→ jeyanth joined03:03
jnewt i just know solidworks doesn't render right when virtulized. lots of people have tried it, as have I, and it is buggy, slow, crashes, etc. it just doesn't work right. I also know that mplab (my ide), doesn't run right, and although there is a fix in the works, it doesn't either. my pcb sofware doesn't either. or my accounting. i use excel alot, and the open office version whatever it's called is not as good. so, i'm bas03:06
ically stuck on windows other than my servers (freebsd), which runs great, but severly limited in the scope of software i need03:06
→ Yuuhi` joined03:07
→ relurk joined03:07
jnewt and i see an os as just a place to use software, nothing more than a container, so regardless of how well a specific os completes performance or stability tests, if it can't run software i need, it's worthless to me.03:08
← thursday left03:08
← Yuuhi left03:08
→ cakehero joined03:08
mns to each their own, but I understand what you're saying03:09
← sharenothing left03:10
→ mvrilo joined03:11
→ andref_dias joined03:11
mns for accounting have you looked at GnuCash or jGnash ?03:11
← tyatpi_ left03:11
← mvrilo left03:11
← Aleric left03:12
→ tyatpi_ joined03:12
← macrover left03:12
→ ipoval joined03:13
← marhaban left03:13
← xiangfu left03:14
← tyatpi left03:15
→ tyatpi joined03:16
← aruiz left03:16
→ techhelp2 joined03:16
→ xiangfu joined03:16
jnewt mns: installed it, ran it, poked around the code to see if i could modify it. it's amateur hour.03:17
mns which one GnuCash ?03:18
or jGnash ?03:18
jnewt oh, my bad, never heard of jGnash03:18
← techhelp left03:18
mns try jGnash ... its Java based .. you know, that whole write once run anywhere thingamajig ...03:19
jnewt mns: i'm aware java. At least from an android development perspective03:20
→ chintanparikh joined03:20
→ thursday joined03:20
chintanparikh Hey people. I have a branch, lets call it 'a_branch' that's based off develop03:21
I then checkout develop and checkout -b a new branch 'another_branch'03:21
jnewt mns: jGnash is really elementary too.03:21
chintanparikh Now for some reason, some changes from a_branch have stuck themselves in another_branch03:21
Is there anyway to remove all those changes from another_branch?03:21
I imagine I could somehow do it via rebase?03:21
But I'm not entirely sure how03:21
I've also committed another_branch a few times already03:22
infogulch|afk → infogulch03:22
mns I dont use either anyway.03:22
jnewt could have re-written all the headers in the time it's taking to get them back03:22
← leeight_ left03:23
jnewt every time i type gi, then press t, it says "No next tag (press RETURN)"03:24
← delight left03:24
→ ISF joined03:26
→ chaowang joined03:26
→ techhelp joined03:27
← macabre left03:27
→ duckxx joined03:28
→ aantix joined03:28
← techhelp2 left03:29
← aantix left03:29
← bitkiller left03:30
← duckxx left03:30
→ duckxx joined03:31
← bwidmer left03:34
← chrisgeorge left03:36
→ the_cybersphinx joined03:36
→ Eridius joined03:37
← v0n left03:37
← duckxx left03:38
← cybersphinx left03:39
← bgerber left03:40
→ EricKamsky joined03:40
← centipedefarmer left03:40
→ bgerber joined03:42
→ centipedefarmer joined03:44
→ mikepack joined03:46
→ ehsan joined03:48
← centipedefarmer left03:49
← thursday left03:50
← Davey left03:51
← Eridius left03:52
→ redii joined03:52
→ duckxx joined03:53
→ subhojit777 joined03:54
→ aantix joined03:54
← redii left03:55
→ hwrd|work joined03:56
→ macabre joined03:57
← lacrymology left03:57
← tyatpi_ left03:58
← luxurymode left03:59
→ gavin_huang joined04:00
→ DofLeFou joined04:01
→ tyatpi_ joined04:01
← cakehero left04:02
→ adamm joined04:03
→ thursday joined04:03
→ Zenith77|2 joined04:03
→ d4rklit3 joined04:04
← xiangfu left04:04
← MattRichardson left04:04
← _d4rklit3 left04:04
← erkules left04:04
← Zenith77 left04:04
← cagedwisdom left04:04
← teotwaki left04:04
← roland left04:04
← brentratliff left04:04
← Werring left04:04
← Axsuul left04:04
← RichardBronosky left04:04
← surfdue left04:04
← trout left04:04
→ roland joined04:04
← tmcmahon left04:04
← AlbireoX`Laptop left04:06
→ alexukf joined04:06
← chintanparikh left04:06
← nadavoid left04:07
← davidban_ left04:08
→ Joseph_ joined04:08
← Kuukunen left04:09
← duckxx left04:09
→ xiangfu joined04:09
→ MattRichardson joined04:09
→ erkules joined04:09
→ cagedwisdom joined04:09
→ teotwaki joined04:09
→ brentratliff joined04:09
→ Werring joined04:09
→ Axsuul joined04:09
→ RichardBronosky joined04:09
→ surfdue joined04:09
→ trout joined04:09
← jhardinee left04:09
→ zdk joined04:10
← cilly left04:12
→ BiggFREE joined04:12
← andref_dias left04:13
→ cilly joined04:14
← tyatpi_ left04:16
← tyatpi left04:16
→ sangi joined04:18
→ tmcmahon joined04:18
← macabre left04:18
← ISF left04:18
→ rchavik joined04:19
→ rtjure joined04:20
← xiangfu left04:22
← MattRichardson left04:22
← erkules left04:22
← cagedwisdom left04:22
← teotwaki left04:22
← brentratliff left04:22
← Werring left04:22
← Axsuul left04:22
← RichardBronosky left04:22
← surfdue left04:22
← trout left04:22
→ abstrusenick joined04:22
← mikepack left04:23
← tmcmahon left04:23
→ duckxx joined04:24
→ Axsuul joined04:25
← d4rklit3 left04:25
← alexukf left04:26
→ davidbanham joined04:26
→ tyatpi joined04:26
→ dfr|mac joined04:26
→ teotwaki joined04:27
→ xiangfu joined04:28
→ MattRichardson joined04:28
→ cagedwisdom joined04:28
→ brentratliff joined04:28
→ Werring joined04:28
→ RichardBronosky joined04:28
→ surfdue joined04:28
→ trout joined04:28
→ frogonwheels joined04:28
← rtjure left04:28
← jeyanth left04:29
← bgerber left04:30
← duckxx left04:30
→ cbreak joined04:31
→ igotnolegs joined04:31
→ bgerber joined04:32
→ erkules joined04:38
→ grandy joined04:38
→ tyatpi_ joined04:40
→ PRETTY_FUNCTION joined04:41
→ grandy` joined04:42
← grandy left04:42
→ davidban_ joined04:42
← davidbanham left04:43
← nareshgbhat left04:44
← grandy` left04:45
→ grandy` joined04:46
← Soliton left04:46
→ TheDudeKnows joined04:48
→ Soliton joined04:48
← grandy` left04:49
← davidban_ left04:49
→ grandy` joined04:49
gitinfo set mode: +v04:49
→ duckxx joined04:49
grandy` hello!04:50
gitinfo grandy`: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.04:50
grandy` hello, trying to use git diff to see the changes between two spots in the tree... as in git diff hash..HEAD .... but it's showing a lot of stuff that did not actually change between those two points... confused.04:50
→ davidbanham joined04:50
← davidbanham left04:51
← frogonwheels left04:51
→ davidbanham joined04:51
grandy` when I use gitx to view the commits in HEAD it shows what i'd expect all the way back to the hash in question... oddly the diff shows other files that were not modified by any of the commits04:51
i'm obviously doing something wrong but not sure what... any ideas?04:52
→ hexaphixylix joined04:53
← bgerber left04:55
← toothr left04:55
→ bgerber joined04:56
→ toothr joined04:57
→ airborn joined04:57
→ arvind_khadri joined04:57
← dmiles_afk left05:00
→ Zenmaker__ joined05:00
→ dmiles_afk joined05:01
← madewokherd left05:02
← infogulch left05:03
→ infogulch joined05:03
← duckxx left05:04
← Zenmaker_ left05:04
→ williamcotton joined05:05
← kpreid left05:06
infogulch → infogulch|afk05:11
→ spaceonline__ joined05:12
grandy` hmm, looks like maybe it's b/c i cherry picked one commit... not sure05:13
← spaceonline left05:14
→ JeremyWei joined05:16
→ Eridius joined05:16
← andrewhl left05:16
→ acl_ joined05:17
← BiggFREE left05:17
← dfr|mac left05:18
gitinfo set mode: +v05:18
← hexaphixylix left05:19
→ marciodias__ joined05:21
← spaceonline__ left05:21
← acl_ left05:21
← delinquentme left05:22
→ acl_ joined05:23
→ toabctl joined05:25
→ maleknet joined05:26
→ ThomasLocke joined05:27
→ dpino joined05:29
← aantix left05:30
→ bluesnow joined05:33
← relurk left05:33
→ jaredoconnor joined05:34
jaredoconnor is there any way to remove an intermediate commit and merge from our source as if they never happened?05:35
bluesnow Hi, I'm wondering what Git considers a conflict when it merges two branches05:35
jaredoconnor i dont care whether they stay in the log, i just want the code to reflect that they were never comitted05:35
it seems that if we just revert those two commits, we get a lot of problems due to one of them being a merge05:35
sorry let me put that a better way...05:35
→ harshpb joined05:36
← bluesnow left05:36
→ relurk joined05:36
jaredoconnor there was a commit and a merge made a while ago05:37
we want to revert all those changes as if they never occurred, but its fine if they stay in the git log/history/etc05:37
← jwmann left05:38
→ jwmann joined05:38
jaredoconnor if we jsut revert both those commits, a lot of stuff gets changed/removed that we dont want to05:38
← jwmann left05:38
← bgerber left05:38
jaredoconnor any suggestions?05:38
→ bluesnow joined05:39
← adamm left05:42
← xtagon left05:43
← bluesnow left05:44
→ lightcap joined05:45
→ pantsman joined05:45
← pantsman left05:45
→ pantsman joined05:45
nevyn so if a revert doesn't work (which it won't on a merge that old..) your only real option is rebase I think05:46
md_5|away → md_505:46
← beneggett left05:47
← airborn left05:47
← fisted left05:48
← lightcap left05:49
→ fisted joined05:50
→ dreiss joined05:50
← grandy` left05:54
→ AlbireoX`Laptop joined05:55
jaredoconnor thanks nevyn we've sorted it out05:58
← toabctl left06:01
← oogatta left06:02
→ oogatta joined06:03
← ehsan left06:03
← blorbx left06:03
← ferdna left06:05
→ nareshgbhat joined06:06
← intripoon_ left06:08
→ intripoon joined06:09
→ dzonder joined06:09
→ dvide joined06:10
← damnick left06:11
→ dc5ala joined06:11
← Targen left06:12
→ damnick joined06:12
→ marthinal joined06:15
→ dhruvasagar joined06:15
→ cbreak-work joined06:17
→ ilogger2 joined06:18
→ soee joined06:20
→ tsdh joined06:22
→ theCole joined06:25
→ codezombie joined06:25
→ jceb_ joined06:25
→ bgerber joined06:29
→ jh3_ joined06:31
→ toabctl joined06:32
→ twem2 joined06:33
ChanServ set mode: +v06:36
→ txdv joined06:36
→ afuentes joined06:37
→ dhruvasagar joined06:38
← toabctl left06:39
→ cesc joined06:42
→ shruggar joined06:44
→ acrocity joined06:44
→ Dave^| joined06:44
→ milli joined06:44
→ sangxiaolong joined06:46
← jceb_ left06:47
→ dpino joined06:50
→ ashphy joined06:51
→ Strogg joined06:54
← ashphy left06:55
→ saju_m joined07:00
→ pom_ joined07:02
→ hoschi joined07:05
→ jdpo joined07:05
→ palam joined07:05
→ stodan joined07:06
→ yizumi joined07:06
← afuentes left07:06
→ doug_carmichael joined07:06
→ Swathe joined07:06
palam is there a way to force conflict resolution by asking git to choose HEAD over the other changes?07:07
that's basically what i'm about to do, but it'll take forever manually...07:07
→ afuentes joined07:07
→ luke joined07:07
→ nareshgbhat joined07:08
→ jceb_ joined07:10
wereHamster checkout --ours ...07:11
→ nadirvardar joined07:12
palam wereHamster: thanks. does that have to be done for each file?07:14
→ spaceonline joined07:15
→ Kuukunen joined07:16
wereHamster no07:16
→ Littlex joined07:17
palam ahh, 'git checkout --ours .' thanks wereHamster07:17
→ mishok13 joined07:19
← cesc left07:19
→ na3r joined07:19
→ chrisf|work joined07:20
→ ph^ joined07:21
→ palm_vue joined07:21
→ frogonwheels joined07:21
→ Lgb joined07:21
→ bwidmer joined07:21
→ PRETTY_F_ joined07:21
→ jonnu joined07:23
→ psoo joined07:24
← Littlex left07:25
→ thiago joined07:30
← yizumi left07:31
← nadirvardar left07:31
→ teknopaul joined07:31
→ iwoj joined07:32
→ QaDeS joined07:34
→ shredding joined07:34
← palam left07:35
→ und3f joined07:35
→ spaceonline__ joined07:37
← spaceonline left07:39
→ polymar joined07:41
→ Sebo joined07:42
→ whitman joined07:43
← theCole left07:44
→ fsvend joined07:45
→ flip214 joined07:47
flip214 Are feature requests ok here, or should I send an email to [email@hidden.address]07:48
→ caseymcg joined07:48
→ rtuin joined07:48
cmn that stuff belongs in the mailing list; it's the only way to know you're reaching developers07:48
flip214 ok, thanks07:49
← flip214 left07:49
→ yizumi joined07:49
← jdpo left07:51
EugeneKay We're mostly just gits.07:51
→ berserkr joined07:53
→ jkremser joined07:54
→ ipalaus joined07:55
→ rurufufuss joined07:56
→ agile joined07:56
← mishok13 left07:57
← ipalaus left07:57
→ ipalaus joined07:59
→ simesy joined07:59
→ ToBeFree joined08:00
→ mishok13 joined08:00
← simesy left08:00
→ segher joined08:01
→ LimeBlast joined08:02
→ shirish joined08:05
→ _Qman joined08:05
← shirish left08:05
→ OOPMan joined08:06
→ erasmospunk joined08:06
→ BarryCrayon joined08:06
→ SkAzZ joined08:09
→ marcogmonteiro joined08:09
← na3r left08:10
← BarryCrayon left08:11
→ angelsl joined08:11
→ eno joined08:11
← eno left08:11
→ eno joined08:11
→ BarryCrayon joined08:13
→ timotei joined08:14
timotei hello. I've git reset --hard on an older commit, before implementing some breaking functionality. Made a tag, and now I want to get back to the latest head (last commit before reset)08:14
→ puneetgoyal joined08:14
timotei Is that possible without specifying that commit's sha?08:14
← nareshgbhat left08:15
thiago timotei: yes08:15
do you have a branch or another ref pointing to the commit you want to get to?08:15
→ blacktulip joined08:16
timotei A... nope08:16
thiago git reflog then08:16
find it there08:16
timotei thiago, thanks. It worked!08:19
I love git. I fell safe whatever it would happen :D08:19
cbreak-work timotei: it was a stupid thing to do08:20
you should have done git checkout -b tmpname commit instead08:20
timotei cbreak-work, stupid things what?08:20
s/things/thing/08:20
cbreak-work to use reset --hard without the intention to want to kill your current branch's history08:20
thiago I agree08:21
reset --hard has the intention of killing work08:21
or of permanently moving the branch, at least08:21
nevyn yep08:21
timotei Hmm... well, I knew that my commits won't be lost.08:22
I just wanted to tag that previous state.08:22
What would have been a better way of doing it?08:22
thiago checkout08:22
→ tcurdt joined08:23
← tcurdt left08:23
thiago yes, your work wasn't lost, but you're not using the tool the way it was meant to08:23
for example, you've reset again back to where you were, right?08:23
timotei Yes08:23
And tagged that too08:23
thiago the work you've done, where is it now?08:23
timotei Ok, I have a branch: master (with unstable changes) and safe (with stable ones)08:23
2 branches*08:23
and 2 tags pointing to the same08:23
thiago why do you need those tags?08:24
timotei So yes, I've used reset to get back to the commit previous to unstable work, tag/branch it, and then back to normal08:24
Well... to be able to get back quick to the safe changes\08:24
cbreak-work timotei: you can just git tag name commithash08:24
thiago note that a tag is not supposed to be moved08:24
cbreak-work no need to switch anything if you want to tag only08:24
thiago once created, it should stay pointing to the same commit forever08:25
timotei cbreak-work, I see08:25
thiago so you need to create a LOT of tags08:25
instead, you should use something that is like a tag but that can easily be moved: a branch08:25
timotei thiago, Yeah. Well, only the safe tag would be ... important08:25
I see08:25
thiago if you wanted to create a branch out of the work you're about to do, you should create the branch upon moving08:25
e.g., git checkout -b new_work <the commit where it applies to>08:25
once you're done, git checkout master08:26
timotei Ah. Without the need to reset all that stuff08:26
Right?08:26
thiago right08:26
timotei Good to know xD08:26
thiago the way you've done it accomplished the same goal, but it's not how git was designed08:26
also, reset --hard will destroy uncommitted changes08:26
timotei Yes. I know that08:26
thiago that checkout will complain if it can't move08:27
it will not destroy uncommitted changes08:27
→ robotmay joined08:27
→ madsy joined08:28
→ AlbireoX`Laptop joined08:29
→ BiggFREE joined08:29
timotei Is it possible to make the safe branch a remote one?08:30
So I can push it?08:30
thiago to make a branch a remote one, you need to push it08:30
so your first question doesn't make sense with the second08:31
cmn pushing is what makes a branch appear on a remote08:31
and git can't do anything with remote branches other than fetch them, so they're pretty useless08:31
thiago except for the uses they were meant to have, of course :-)08:32
timotei Well, currently the master branch is shared with the others08:32
I created a new branch which I want to make visible to the others too08:32
thiago push it08:32
← OOPMan left08:32
timotei Ok.08:32
→ Nirvanko joined08:33
Nirvanko Is it a good practice to store binary files in the Git repository?08:33
→ Zespre joined08:34
thiago Nirvanko: no08:34
Git was invented to track source code08:34
→ marc_kdab_ joined08:35
frogonwheels Nirvanko: What type of files were you thinking?08:35
thiago source code often has some binary files (like images and other artwork), so storing them *works*08:35
but it's not meant to store huge files that change often08:35
frogonwheels Nirvanko: git-annex provides a solution by not storing the files themselves,but softlinks that it has some funky way of having alternate repositories of the large files08:37
which means it doesn't work in windoze08:37
thiago softlink != symlink08:38
frogonwheels well yeah, they are symlinks08:38
timotei Symlinks work in Windows actually :P08:39
cmn under lab conditions, sure08:39
timotei No. NTFS supports symlinks actually08:40
frogonwheels timotei: but not in msys or msysgit (yet - I've got some work on it i need to get back to) and the implementation in cygwin of symlinks doesn't involve windows ntfs sy mlinks08:40
cmn yes, but you need to be an admin to create them08:40
← AlbireoX`Laptop left08:40
cmn which is why you can't use them on the field08:40
frogonwheels timotei: one of the big stumbling blocks is that you need certain permissions which only admins have by default..08:40
cmn: I believe you can give other users permissions to create them... but I've not really explored that properly08:41
cmn my Windows guy just mentions having to be admin08:41
but even if you can, you can't rely on it08:42
frogonwheels cmn: I think it's more complex than that - like I think there's some special permission you need .. which admin users have.08:42
Nirvanko frogonwheels, images08:42
cmn sure, just like in Linux you need some capabilities, but generally only root has them08:43
frogonwheels Nirvanko: are they generated? big? change often?08:43
→ raatiniemi joined08:43
cmn (not for symlinks, but it's related)08:43
← iwoj left08:43
→ kpl joined08:43
frogonwheels cmn: http://superuser.com/questions/104845/permission-to-make-symbolic-links-in-windows-7 .. buried deep!08:43
Nirvanko Not often08:43
frogonwheels Nirvanko: are they generated?08:43
Nirvanko: Sure, you can store them. Merging is obviously not handled.08:45
→ nareshgbhat joined08:45
Nirvanko Very well then08:45
thx08:45
cmn frogonwheels: as with most paths on Windows, once you know which one it is, every path element makes sense, but you'd never think of looking there08:45
frogonwheels cmn: lol. soooo true08:45
→ spyro joined08:47
← spyro left08:47
→ spyro joined08:47
← bgerber left08:49
kpl If by mistake a number of files where committed at some point, and thereafter some new commits have been done, I assume that the files added by mistake are only present in the object created by the commit where those files were added. Right? Do the objects created by later commits depend on the object with all the unwanted files, so that it is not possible to remove the object with the...08:50
...unwanted files form the repo?08:50
→ jatt joined08:51
kpl Is it possible to remove objects from the repo at all?08:51
← BiggFREE left08:52
kpl I understand that an object cannot be changed, because it is uniquely identified by the SHA1 of its contents. But can it be removed?08:52
DrNick not without producing a corrupted repository08:53
but this problem is one of the reasons git filter-branch exists08:53
→ bgerber joined08:55
→ andref_dias joined08:57
→ uraimo joined08:59
← hoschi left09:02
← jkremser left09:03
→ stkrzysiak joined09:03
→ leeight_ joined09:03
← teknopaul left09:04
→ nax joined09:07
→ maknz joined09:09
← maknz left09:09
→ maknz joined09:09
← nax left09:09
→ marhaban joined09:10
← thiago left09:13
← _Qman left09:13
← spaceonline__ left09:14
→ hoschi joined09:14
→ variousnefarious joined09:20
→ cortana2 joined09:20
→ jatt` joined09:21
← bgerber left09:24
→ Vile joined09:24
← Vile left09:24
← jatt left09:27
← jatt` left09:27
→ jatt joined09:27
→ rtjure joined09:29
→ Lord_DeathMatch joined09:30
→ aspotashev joined09:34
→ eijk joined09:38
→ _Qman joined09:38
← caseymcg left09:43
→ Cinecity joined09:45
← Cinecity left09:45
← shredding left09:47
← sangxiaolong left09:48
← madsy left09:49
→ mklappstuhl joined09:49
← aspotashev left09:50
→ newbie999 joined09:51
→ jn_ joined09:51
→ danslo_ joined09:54
← luke left09:55
← marhaban left09:57
← newbie999 left09:57
← danslo_ left09:58
→ danslo joined09:58
→ naquad joined09:59
danslo In an interactive rebase, a commit was changed - however, those changes were unwanted. I can edit the commit again during a rebase, but how do I undo those changes relative to some remote? I want to do something like; git reset --hard origin/development COMMIT_SHA ; how can I achieve something like this?09:59
(does that make any sense?)09:59
preferably without the reflog10:00
← puneetgoyal left10:00
danslo ooohh10:01
git reset --hard ORIG_HEAD10:01
frogonwheels danslo: or use git checkout branchname -- filename10:01
← raatiniemi left10:01
danslo frogonwheels: the thing is I don't know specific filenames10:02
← naquad left10:02
danslo and ORIG_HEAD won't cut it either actually, due to HEAD being actual HEAD and not the 'virtual' head you get during a rebase10:02
→ tomislater joined10:03
→ sebrock joined10:03
→ harish joined10:05
danslo frogonwheels: never mind me… i can just 'git fetch ssh://some_repo refs/changes/85/17085/2 && git checkout FETCH_HEAD' thanks to gerrit being so awesome :)10:05
→ puneetgoyal joined10:06
→ rbarraud joined10:06
→ naquad joined10:07
← jn_ left10:07
← dhruvasagar left10:08
← stkrzysiak left10:08
→ dhruvasagar joined10:08
→ webben joined10:11
→ heroux joined10:11
→ cedeon joined10:14
← cedeon left10:15
→ arietis joined10:16
→ cedeon joined10:16
← rbarraud left10:19
→ rbarraud joined10:19
→ jkremser joined10:21
→ arti_t joined10:22
→ PapaSierra joined10:27
→ Codif joined10:28
← Codif left10:28
→ noisebleed joined10:29
→ wolog joined10:29
PapaSierra morning, i've added "system" as a submodule. i don't want to tamper with its contents. from my main repo i must have done something wrong because i get the following: http://www.hastebin.com/fetewowura.mel what can i do to "reset" this change or whatever's gone wrong?10:30
→ dkannan joined10:30
← cedeon left10:30
← arti_t left10:30
→ arti_t joined10:31
cbreak-work PapaSierra: cd into it and diff again10:32
dkannan hi. i have deleted a remoted branch and on local as well. can i remover it ?10:32
eh. can i recover it10:33
i tried fsck and reflog. but it does not tell me the SHA for the branch10:34
→ newbie999 joined10:34
→ alansaul joined10:34
→ SkAzZ1 joined10:35
← afuentes left10:35
← SkAzZ left10:36
→ nareshbhat joined10:36
← erasmospunk left10:37
gitinfo set mode: +v10:37
→ rcu joined10:38
← puneetgoyal left10:38
→ argoneus joined10:41
← rbarraud left10:46
→ gusnan joined10:48
→ afuentes joined10:48
→ puneetgoyal joined10:50
→ Vile joined10:50
← kpl left10:50
→ skylamer` joined10:50
→ tvw joined10:51
→ RaceCondition joined10:55
← tomislater left10:55
→ glennpratt joined11:00
← glennpratt left11:01
→ dcarper joined11:01
→ glennpratt joined11:01
← Sebo left11:04
→ jcromartie joined11:04
← dcarper left11:05
→ enmand joined11:06
→ tomislater joined11:08
→ Tinti_ joined11:09
← skylamer` left11:14
→ raatiniemi joined11:16
→ Nibato joined11:16
→ thiago joined11:18
→ avinashhm joined11:19
→ wbednarski joined11:19
→ Eridius joined11:21
→ Sebo joined11:22
charon dkannan: if you delete a branch, the reflog is deleted along with it. so you can either dig up the tip commit from the reflog of HEAD (or another branch), or you can look in the unreachable commits. perhaps !fixit has instructions for that11:23
gitinfo dkannan: [!fixup] So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!11:23
→ tjadc joined11:24
→ shredding joined11:25
← avinashhm left11:25
← rcu left11:26
← nareshgbhat left11:26
← Eridius left11:26
→ avinashhm joined11:27
← jcromartie left11:29
dkannan thanks @charon. i do not see my branch name in "git reflog --all'11:29
is that because as u said, deleting the local branch deleted the reflog as well11:30
how do i, "dig up the tip commit from the reflog of HEAD (or another branch"11:30
← tsdh left11:30
→ luke joined11:33
charon dkannan: HEAD's reflog will only contain this commit if you had the branch checked out recently. in this case, you would stare at 'git reflog' until you find the commit11:34
dkannan: you can also search for the branch name in 'git reflog', git-checkout records it11:34
→ Alir3z4 joined11:35
Alir3z4 is git-subtree merged into main git or not?11:35
cmn Alir3z4: it's in contrib/11:36
← afuentes left11:36
Alir3z4 Because when i run "$ git subtree" i get => "git: 'subtree' is not a git command. See 'git --help'"11:36
cmn then you don't have it11:37
your distribution might not have activated it11:37
if you built it yourself, you need to install it explicitly11:37
Alir3z4 i'm on archlinux11:37
→ cehteh joined11:37
cmn then it looks like the maintainer didn't decide to install it11:38
Alir3z4 cmn: I have to check, maybe it's in testing repo11:39
danslo I installed git just yesterday, 1.7.11, that command is not available for me either11:40
downloaded from git-scm11:40
cmn installed how? from where?11:40
danslo from git-scm, by executing installer11:40
cmn then whoever makes the package for your OS didn't install it either11:41
danslo I suppose. I don't need it, thought it would be relevant ^11:41
Alir3z4 cmn: That maintainer have to be effected by zombies :|11:42
cmn :S11:42
→ JeremyWei joined11:43
← JeremyWei left11:43
cmn you should go complain to him then11:44
← mklappstuhl left11:44
→ tatsuya_o joined11:45
← shredding left11:46
Alir3z4 cmn: I'm complaining right now on #archlinux, which i know it's not take me anywhere!11:46
cmn I don't see why it would11:47
← RaceCondition left11:47
→ tyatpi joined11:48
Alir3z4 cmn: Is subtree module in stable ?11:48
Oh yes11:49
cmn meaning what?11:49
→ levi501d joined11:49
← shruggar left11:49
tabakhase hi folks11:50
levi501d is it possible to fetch a pull request into my forked repository ( a pull request from the originating repository)?11:50
tabakhase shouldnt 1.7.10 be able to clone into a nonempty folder? =(11:50
cmn tabakhase: no version lets you do that11:51
tabakhase cmn hmpf? i thing that where one of these things that didint worked one one but on another machine...11:51
levi501d tabakhase: http://stackoverflow.com/questions/2411031/git-how-do-i-clone-into-a-non-empty-directory might help you out11:51
→ toothr joined11:51
tabakhase so the deal is still clone it bare, make it unbare, copy the .git dir into another one and checkout? sic...11:51
cmn what are you actually trying to achieve?11:52
→ afuentes joined11:53
tabakhase cmn ma repo is ma homedir11:55
where "my" is "the company product project thing on our servers" ;-)11:55
levi501d nvm figured my question out11:55
← levi501d left11:55
→ rgsteele joined11:55
cmn that doesn't seem like it's anything close to a good idea; but what it's definitely not is a goal11:56
← Alir3z4 left11:56
→ cedeon joined11:56
tabakhase cmn not a god idea io totaly agree, but a project running sice 2008 with 60k files and about 2.8 million lines of code is not that easy to reorganise (specialy without a vcs) ;-)11:57
cmn that is still not a goal11:58
and if you're not using version control, what is the problem with git?11:58
← Lord_DeathMatch left11:59
tabakhase so when a new webnode popsUp puppet creates a user and then should be able to clone the repo "in him" (ignoring server individual files like ~/.ssh and so on whats just fine)11:59
→ Lord_DeathMatch joined11:59
cmn do something sane like rsync or git-archive11:59
tabakhase then im gone stay on the old cp way12:01
cmn git isn't for deployment12:03
and if copying the files is enough, using git is most likely a huge overkill12:03
← Sebo left12:04
danslo to be fair, rsync isn't really for deployment either (though Im not sure you implied such a thing)12:04
cmn rsync is a hundred times better than git for that12:04
and for simple things, it's good enough12:05
→ tyatpi_ joined12:05
cmn as long as you choose the flags carefully12:05
tabakhase cmn reprase webnode with devnodes12:06
webnodes are synced from the stage system12:06
cmn neither mean anything without context12:06
danslo cmn: I know, we actually use rsync with jenkins, but there is the odd chance of race conditions with db migrations in our case … want to swap to something symlink based (capistrano) asap12:06
you said simple, I know12:07
cmn if you need to do migrations, you're in a completely different game12:07
← dpino left12:08
→ Fullmoon joined12:08
→ yaymukund joined12:09
← arti_t left12:10
→ AeroNotix joined12:11
→ tqmz joined12:12
→ TheJH joined12:13
→ jcromartie joined12:14
← alansaul left12:14
→ Sebo joined12:16
← toothr left12:17
→ toothr joined12:17
toothr → toothrot12:17
→ mabrand joined12:18
→ tkboy joined12:19
→ d0k joined12:19
← jatt left12:20
← afuentes left12:20
→ afuentes joined12:21
← leeight_ left12:24
← afuentes left12:24
← PRETTY_F_ left12:24
→ project2501a joined12:25
→ afuentes joined12:25
← danslo left12:25
→ Lulzon joined12:28
← AeroNotix left12:29
→ bitkiller joined12:30
← tkboy left12:30
→ bgerber joined12:31
→ AeroNotix joined12:32
← AeroNotix left12:34
→ leeight_ joined12:37
← yaymukund left12:38
→ danslo joined12:40
→ yaymukund joined12:40
project2501a good morning!12:40
is there a way i can take a pre-installed OS on a host computer and cp it, some way, into a virtual image?12:41
→ ISF joined12:41
→ jargon- joined12:42
charon project2501a: huh, how is that a git question?12:43
(i would also imagine that if the OS is linux, simply making a disk image should work to some extent)12:44
← afuentes left12:44
tabakhase can i revert a checkout? as in "transform to bare"? git-ls | xargs rm is "..."12:44
wereHamster tabakhase: that doesn't make a repo bare12:44
cmn checkout doesn't transform anything to non-bare, so undoing it wouldn't make it bare12:44
wereHamster tabakhase: why don't you git clone --bare ?12:44
tabakhase not a real bare, thats why its in quotes ;-)12:46
wereHamster what do you mean by bare then?12:47
what are you actually trying to do.. ?12:47
tabakhase "no working tree" or better "all files that are 'in git' removed from the working tree12:47
for getting 'files that are in there but not in git' or so...12:47
→ Maccer joined12:48
wereHamster eh?12:48
tabakhase wereHamster a git clone --no-checkout is kind of what im looking for12:48
but "fixing a forgotten --no-checkout" is what im looking for12:49
wereHamster tabakhase: why not make a bare clone?12:49
what problem are you trying to solve?12:49
MacGyver wereHamster: I think what he means is he did a checkout into a non-empty folder and now wants to isolate the files that are not in git.12:49
Or something.12:49
(Not even sure that doing that directly is possible)12:50
tabakhase MacGyver quite yes - the "direct possible" is my question in here ;-)12:50
MacGyver No, I mean, if it's directly possible to checkout into a non-empty folder.12:51
tabakhase atm im ls -la | xargs deletewhennotingit.sh $0 where deletzewhennotingit checks for "non empty output" of a git log for the particular file12:51
works, but its a shame12:51
charon tabakhase: git ls-files -z | xargs -0 rm12:51
← fsvend left12:52
tabakhase charon hm that seems great12:53
→ shruggar joined12:53
tabakhase and ls-files -o gives the counterprobe, wonderfull12:54
expect heas eating directorys on the default settings12:58
→ mxweas_ joined12:58
→ Targen joined12:58
→ stuntmachine joined12:58
→ bartek joined12:59
→ dcarper joined13:01
→ jackh joined13:02
→ Vert joined13:03
← dcarper left13:05
← mxweas_ left13:05
Rhonda huhm, why did my fingers let me mistype "git diff" as "big diff".13:05
… especially when the diff is empty currently. :)13:05
→ ThePremier joined13:05
tango_ lol13:06
← Targen left13:07
→ mxweas_ joined13:07
→ benjam_es joined13:08
← mxweas_ left13:12
→ TeckniX joined13:12
→ macabre joined13:12
→ centipedefarmer joined13:13
→ mxweas_ joined13:13
→ AaronMT joined13:14
→ mklappstuhl joined13:14
→ mns joined13:16
→ juanfer joined13:18
→ dv310p3r joined13:18
→ dcarper joined13:19
← palm_vue left13:20
← Lgb left13:20
→ Ecesis joined13:21
→ Lgb joined13:21
→ palm_vue joined13:21
← mns left13:22
→ bambams joined13:22
→ neurodrone joined13:22
← ph^ left13:23
→ rendar joined13:23
bambams I rebased a (private) upstream and am trying to figure out how to rebase a local branch on it in one command. There are 9 commits different in each branch.13:23
I tried: git rebase --onto master~9 my-remote/master~9 master13:23
→ Zenmaker_ joined13:24
→ ph^ joined13:24
bambams That's not working. :\ I also tried with my-remote/master, which does something, but still leaves 2 and 9 commits different, respectively. :\13:24
← jcromartie left13:25
→ hellopat joined13:25
→ ph^_ joined13:25
← ph^_ left13:26
→ superdmp joined13:26
→ erichynds joined13:27
→ ph^_ joined13:27
← ph^ left13:28
bambams Hmmm, maybe I just have to do it in two steps. First rebase master to remove those 9 commits, and then rebase or fast-forward merge the remote..13:29
superdmp i use https://github.com/divio/django-cms/, but I also need to use my custom version of it (which implements a couple of new features). What I need is a convenient way to keep up to date with their develop branch, and merge my own feature code into each time their is updated.13:29
← mxweas_ left13:29
← Ecesis left13:29
→ axl_ joined13:29
nevyn rebase? or a munge filter13:29
superdmp And later, I might like to make my features a commit for their master version13:29
so what would be a good strategy for that13:30
bambams git rebase --onto master~9 master && git merge my-remote/master # This was a triumph...13:30
← TheJH left13:31
charon bambams: starting on which branch?13:31
← ph^_ left13:31
charon bambams: i'd have to look, but if you start on master, that would seem to *discard* 9 commits from master13:32
→ mns joined13:32
→ jcromartie joined13:33
← rgsteele left13:33
→ eueu joined13:34
→ kaputtmacher joined13:34
kaputtmacher hi all13:34
gitinfo kaputtmacher: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.13:34
kaputtmacher i have some strange message since i tried the command "git add .gitattributes"13:35
now i'm getting this one each time:13:35
behaviour, is not a valid attribute name: .gitattributes:113:35
bambams charon: Yes, that is basically what I wanted to do. Throw away those 9 commits in master and then add those rebased commits back from the rebased remote branch.13:35
→ thornofmight joined13:35
kaputtmacher the 1st line of my gitattributes is a comment and looks like this: # Set default behaviour, in case users don't have core.autocrlf set.13:35
← benjam_es left13:35
kaputtmacher any hints how to get rid of that message every time i make some commits13:36
??13:36
bambams I'm not sure, but perhaps git-pull --force would do that too? I'm trying to learn to rebase though (I rebase interactively all the time, but I want to do it non-interactively when possible).13:36
→ consti joined13:37
→ mxweas_ joined13:37
gitinfo set mode: +v13:38
consti Hi! We have a very old and pretty large repository and we would like to reduce its size by removing everything in the history (including objects) before a certain commit13:38
→ gour joined13:38
thiago consti: my suggestion is to choose "a certain commit" to be "the latest right now"13:38
→ chrisgeorge joined13:38
thiago consti: rm -rf .git; git add .; git commit -m "New history starts"13:38
consti: push to a different repository and people will clone13:39
consti would we still be able to cherry pick things from the old repository?13:39
← toothrot left13:39
← raatiniemi left13:39
bambams kaputtmacher: I'm not familiar with .gitattributes, but the docs don't say anything about comments in that file (that I can see). Are you sure that file supports comments?13:40
→ raatiniemi joined13:40
→ intransit joined13:40
consti thiago: I tried something like this http://eggandham.com/2010/08/clear-old-git-history/ - but that didn't help13:40
→ beniamin joined13:40
kaputtmacher bambams: i'm sure it supports comments, because of all the other gitattribute files out there being commented somehow13:40
i just think it was some kind of a typo of mine13:41
→ guampa joined13:41
kaputtmacher what does <filename>:1 stand for?13:41
does it mean line 1 on file <filename> ?13:41
bambams I assume that means line 1.13:41
kaputtmacher well okay13:41
consti our ./.git/objects/pack is 233M, but the current repository is only around 45 MB13:41
kaputtmacher i will remove the comment and read it again after a commit :)13:41
→ toothr joined13:41
bambams kaputtmacher: Perhaps file encoding. :\ (guessing...)13:41
kaputtmacher hmm13:42
→ thornofmight_ joined13:42
← thornofmight left13:42
kaputtmacher i had a lot trouble with crlf stuff… therefore i use a gitattributes now13:42
consti I would have guessed that there is some straight forward way to remove everything in history before a certain commit. I don't mind checking out a new repository for that -13:43
kaputtmacher since i added it to the repo via git add .gitattributes, the message appears..13:43
← mxweas_ left13:44
jast kaputtmacher: my guess is you have a comma somewhere in .gitattributes13:44
kaputtmacher i do13:44
jast yeah. remove it. ;)13:44
kaputtmacher # Set default behaviour, in case users don't have core.autocrlf set.13:45
→ Pupeno_W joined13:45
kaputtmacher lol, that's all?!13:45
bambams o_O13:45
kaputtmacher <- keeps in mind never to use commas again… :-P13:45
jast .gitattributes doesn't support comments, apparently13:45
so that line is treated as an attribute assignment13:45
← jackh left13:45
kaputtmacher aah.. thx wise man :)13:45
bambams Pffft, grammar. Who needs it? errrr i mean grammar who needs it13:45
→ mxweas_ joined13:45
consti :)13:46
→ arvind_khadri joined13:46
→ kpreid joined13:46
→ heptadecagram joined13:47
→ quazimodo joined13:47
← wbednarski left13:47
consti anyone else who has an idea how I could remove everything (including objects) before a certain commit?13:48
jast sure. use a graft, then filter-branch.13:48
finally, clone the result to get rid of the unreferenced objects easily13:48
consti thanks! that sounds doable :) (reading up on graft) - any 'tutorial' you can point me to?13:49
→ dpino joined13:49
→ lledet joined13:50
jast it's pretty easy, really. the 'grafts' file is documented very briefly in man gitrepository-layout13:50
gitinfo the gitrepository-layout manpage is available at http://git-scm.com/docs/gitrepository-layout.html13:50
← puneetgoyal left13:51
jast it just means you put a line in there with nothing but the ID of the commit you want to change13:51
since there's nothing else in that line, git will take that to mean you want that commit to pretend to have no parent commit(s)13:51
then you just need to run git filter-branch -- --all and that should be it13:51
bambams recommends a tarball backup first. xD13:52
← mxweas_ left13:52
→ shirish joined13:52
consti wow13:52
jast filter-branch backs up all the refs itself, too... but the tarball is quite a bit easier :)13:52
→ puneetgoyal joined13:54
→ revagomes joined13:54
consti jast: will try all of this as soon as I cloned the repository again (since I destroyed my local copy with filter-branch.. ;)) - thanks!13:54
← puneetgoyal left13:55
→ asedeno_work joined13:58
→ ixti joined13:58
← timotei left13:58
→ Davey joined13:59
rendar is the git code released under gpl? why there is a LGPL file in the root dir? :)13:59
← Zespre left13:59
← eueu left14:00
→ Zespre joined14:00
← tyatpi left14:01
→ john2496 joined14:01
← soee left14:01
cmn probably because that file is under the LGPL14:03
wereHamster rendar: see for yourself. Find the commit which added the file, it's described why.14:03
jast indeed so14:04
git itself is GPL'd14:04
→ ehsan joined14:05
→ shawngo joined14:05
wereHamster except for the parts thart are LGPL :)14:06
project2501a charon: mis-messge, sorry14:09
← Fullmoon left14:09
rendar wereHamster: i'm searching, btw i need to know if the sha1.c is under LGPL or GPL, do you know that?14:12
→ madsy joined14:12
← madsy left14:12
→ madsy joined14:12
→ Fandekasp joined14:12
→ archdori joined14:12
wereHamster rendar: gpl, I would assume14:13
rendar: LGPL is only for those tihngs that are described in the commit message of the commit which added the LGPL file14:13
→ TheHaven joined14:13
rendar wereHamster: i see14:13
→ fidothe joined14:14
→ dfr|mac joined14:14
← shirish left14:14
→ alester joined14:15
→ Joseph__ joined14:15
→ yatiohi joined14:16
→ kerrick joined14:17
→ beatak joined14:18
← tjadc left14:18
→ jjbohn joined14:19
← danslo left14:19
→ andrewhl joined14:20
← Fandekasp left14:23
← archdori left14:23
→ giallu joined14:24
← giallu left14:24
→ giallu joined14:24
jast wereHamster: the commit message says "and so on" ;)14:24
→ plastics joined14:25
→ GuestQ283503 joined14:25
← Sebo left14:27
→ Pupeno_W_ joined14:28
→ fly9 joined14:28
← Pupeno_W_ left14:28
← Pupeno_W left14:29
→ joshteam joined14:29
→ dpino_ joined14:29
→ danslo joined14:30
→ shirish joined14:30
→ Synthead joined14:30
← jjbohn left14:32
← dpino left14:32
← shirish left14:33
← ISF left14:34
→ arturaz joined14:35
← thiago left14:36
→ Destos joined14:36
arturaz Hello. Is there a way to only clone one revision from remote repo?14:36
→ senny joined14:36
→ kukks joined14:37
kpreid git clone --depth 1, but note that such a clone is mostly useless -- see man git-clone14:37
gitinfo the git-clone manpage is available at http://git-scm.com/docs/git-clone.html14:37
wereHamster arturaz: it's easier if you download a tarball14:37
→ alansaul joined14:40
arturaz wereHamster, tarball from where?14:40
← psoo left14:40
→ devn joined14:41
→ irqq joined14:41
wereHamster how am I supposed to know what repo you're tryign to clone14:41
← andrewhl left14:41
→ sylr joined14:41
← luke left14:43
← senny left14:44
→ Sebo joined14:47
→ kermit joined14:47
← Lord_DeathMatch left14:48
→ macrover joined14:50
← dfr|mac left14:51
→ Silex joined14:51
consti jast: I added the SHA1 to grafts and ran the filter-branch command14:52
jast: but the objects/pack is still very large14:52
← ThePremier left14:53
consti ..and gitk still shows history way before my SHA1 -14:54
so I'm not sure if that helped :(14:54
superdmp can I use git (or github) to show me a diff between two branches? I don't care about all the intervening commits, it's only the end result of changed files that I am interested in14:54
PerlJam superdmp: yes14:55
Silex give Silex !tias (ignore this it's for me)14:55
gitinfo Try it and see™. You learn much more by experimentation than by asking without having even tried. If in doubt, make backups before you experiment (see !backup). http://sitaramc.github.com/1-basic-usage/tias.html may help with git-specific TIAS.14:55
→ mxweas_ joined14:56
→ styx_ joined14:56
superdmp PerlJam: can you elucidate? the best I have got so far is GitHub's compare view: https://github.com/evildmp/django-cms/compare/84f1212d44a147f8bb6f4c985b05a6aa5e30c63a...pageflags for example, but it's giving me a lot more than I need or want14:56
→ Laura__ joined14:58
cbreak-work superdmp: man git-diff14:58
gitinfo superdmp: the git-diff manpage is available at http://git-scm.com/docs/git-diff.html14:58
cbreak-work just git diff a b and you're done14:58
if you want it to look better, use --patience and maybe --color-words14:59
superdmp thanks cbreak-work14:59
← jcromartie left15:00
← yizumi left15:00
devn Any suggestions for options other than submodules? I have a repo where devs switch back and forth quite a bit between the master and foo branches. In master there are a few subdirectories like "bin" and "lib". In the foo branch bin and lib have been `git rm -rf bin`ed, and added to .gitignore. I then cloned bin and lib into the foo branch. Everything looked fine (probably see where this is headed) until I switched back to master, and then back to the foo ...15:00
... branch again. Oops. If I descend into bin or lib, they show that all the files have been deleted due to switching branches. So, this is obviously a less than happy solution. Submodules seem like a poor fit particularly given the constant branch switching by devs. What about git subtree? Does that get me closer to heaven?15:00
← giallu left15:00
← mxweas_ left15:00
← jonnu left15:01
← dhruvasagar left15:02
→ blendedbychris joined15:02
← blendedbychris left15:02
→ blendedbychris joined15:02
→ mxweas_ joined15:03
← Maccer left15:03
← centipedefarmer left15:03
→ wilmoore joined15:03
→ stkrzysiak joined15:04
→ kevwil joined15:06
← mxweas_ left15:07
→ marhaban joined15:07
← jceb_ left15:07
→ mxweas_ joined15:08
→ _iron joined15:09
→ FabTG_ joined15:10
← raatiniemi left15:11
jast consti: there will still be the backup refs created by filter-branch that contain the old history. the only easy way to get rid of everything is to clone the repository and look at the clone.15:12
→ cakehero joined15:12
jast consti: if, however, you're in the correct branch and the old history is still shown there, that means something didn't work15:12
devn: if you follow the hints in !subpro, you get brief explanations of the various options15:13
gitinfo devn: [!subprojects] So, you want to add git repositories inside of other git repositories? Well, you have four main options. First is to just do it, add the repo to the outer project's .gitignore, and treat them entirely separately. Best if they are entirely seperate. Otherwise your best options are "!submodule" "!gitslave" and "!subtree" Try typing those commands into this IRC channel.15:13
→ dermoth|work joined15:13
← Laura__ left15:13
→ i42n joined15:13
→ diwic joined15:14
→ ipoval joined15:14
← mxweas_ left15:14
← guampa left15:14
consti thanks jast! - I have to run now, but I think it didn't work (I still see the old history) - will look into it tomorrow15:15
← Zenmaker_ left15:15
← Davey left15:15
→ ph^ joined15:15
→ Zenmaker_ joined15:15
→ etcetera joined15:15
gitinfo set mode: +v15:15
→ Pupeno_W joined15:15
diwic Hi! Is it possible to remove all merge commits with e g "git filter-branch"?15:15
← danslo left15:15
→ mxweas_ joined15:16
jast sure... but note that that won't remove the *effects* of the merge15:16
→ davidg joined15:17
diwic jast, exactly, I want to remove the merge commit but not the effects of it15:17
→ danslo joined15:17
→ bluj joined15:17
jast diwic: then filter-branch is your friend15:18
bluj i have a local commit with 2 files in it... and i would like to remove 1 of those 2 files from the commit (but still keep the changes... e.g. just keep it in the staging area again), how do i do this?15:18
→ kenperkins joined15:18
→ Gizmokid2005 joined15:19
← Synthead left15:19
→ jcromartie joined15:19
diwic jast, so that if I have commits A1 and A2 in branch A, B1 and B2 in branch B, then branch A has merged branch B in a merge commit. After the filter-branch, it would look like B1 and B2 were directly committed to branch A.15:19
wereHamster bluj: completely remove the file frmo the next commit or just unstage the changes?15:20
← ipoval left15:20
→ alvaro_o joined15:20
→ Zenmaker__ joined15:20
bluj wereHamster: i want to eventually still have 1 local commit, but it to only contain changes to 1 file instead of two. however, that second file i would like to retain the changes (e.g. just kept it staged, i'll add it to a separate commit later)15:20
diwic jast, any hint about how to go about doing such a filter?15:21
→ Quanteek joined15:21
← Quanteek left15:21
wereHamster bluj: are you realizing that what is staged goes into the next commit? You can't keep it staged and not have it go into the next commit15:21
jast diwic: okay, for that you actually want to rebase the branch. this sounds tricky to do with filter-branch.15:21
bluj wereHamster: sorry not staged then, untracked i guess15:21
→ egarcia joined15:22
→ tr-808_ joined15:22
→ tr-808 joined15:22
jast diwic: try git rebase with --root. note that rebase incantations can be tricky to figure out, so do look at the examples in the manpage.15:22
→ Synthead joined15:22
diwic jast, but there are thousands of these merges, won't I then have to make thousands of rebases?15:22
wereHamster bluj: untracked? git rm --cached <file>15:22
bluj thanks15:22
jast diwic: it should be possible to have rebase rewrite the complete branch at once15:22
but actually I think it squashes the merges15:22
that's not what you want, is it?15:22
the problem if you mangle the merged commits into a linear history is that the result will look rather nonsensical15:23
← Zenmaker_ left15:23
jast I mean, suppose B1 and B2 were made after A1 and A2 were made. then you have the merge, and you insert B1 and B2 in that place instead15:24
→ thiago joined15:24
← arvind_khadri left15:24
jast now B1 will look as if it magically reverted the effects of A1 and A2, and you'll have to add an extra commit after B2 that contains the actual effects of the merge (including fixed conflicts etc.)15:24
← alvaro_o left15:24
diwic jast, hmm, yeah that sounds a little strange I guess15:25
→ tjadc joined15:25
jast the only sensible way to linearize merges is to turn them into regular commits15:25
assuming that you think linearizing merges can be sensible at all ;)15:25
← mxweas_ left15:26
superdmp when I push: git push <remote> <some new branch> that should create the new branch on the remote, is that correct? if so, why: error: src refspec new-pageflags does not match any ?15:26
milki superdmp: does it exist as a local branch?15:27
→ Yuuhi joined15:27
→ sternocera joined15:27
superdmp milki: yes, though it has a differnet name locally15:27
← Zenmaker__ left15:27
→ mxweas_ joined15:27
milki thats your problem15:28
← tr-808_ left15:28
← tr-808 left15:28
→ Zenmaker__ joined15:28
→ tr-808 joined15:28
milki superdmp: read the definition for refspec in man git-push15:28
superdmp thanks milki15:28
gitinfo superdmp: the git-push manpage is available at http://git-scm.com/docs/git-push.html15:28
wereHamster superdmp: so, please explain how you expect git to figure out which branch you want to push? I mean.. it can't read your mind..15:28
→ tr-808_ joined15:28
superdmp wereHamster: it does sometimes15:29
← blendedbychris left15:29
wereHamster superdmp: from your command, 'git push origin new-pageflags', how should git know that you want to push your local branch frotz?15:29
→ plastics- joined15:29
→ Zenmaker_ joined15:30
← wilmoore left15:31
→ tonyj joined15:31
diwic jast, yeah I guess A1 and A2 could be conflicting with the changes from B1 and B2...15:31
← plastics left15:31
← mxweas_ left15:32
superdmp wereHamster: actually I assumed it would push the current branch to origin new-pageflags15:32
← tr-808 left15:33
→ mxweas_ joined15:33
← Zenmaker__ left15:33
← shawngo left15:34
← tr-808_ left15:34
→ flijten joined15:34
wereHamster superdmp: a refspec 'foo' is the same as 'foo:foo'15:34
→ dfr|mac joined15:35
→ tr-808 joined15:35
← flijten left15:36
→ tr-808_ joined15:36
← doug_carmichael left15:36
← Sebo left15:36
→ doug_carmichael joined15:37
← mxweas_ left15:38
→ mxweas_ joined15:39
← tr-808 left15:40
← polymar left15:40
→ tr-808 joined15:40
← mishok13 left15:41
→ RESNO joined15:41
RESNO there's a way to use git without ssh-agent, passing the private key directly to it15:41
← doug_carmichael left15:42
← danslo left15:42
RESNO is there way to use git without ssh-agent passing theprivate key toit?15:42
RandalSchwartz sure15:42
← tr-808_ left15:42
milki if thats a question, the answer is yes via .ssh/config15:42
← AaronMT left15:43
→ tr-808_ joined15:43
→ danslo joined15:43
← rtuin left15:43
← dj_tjerk left15:43
← Lulzon left15:43
RESNO RandalSchwartz: milki cool, im helping with a project piggybacking on git :)15:43
← tyatpi_ left15:44
RESNO milki: how would you set it?15:44
→ eijk_ joined15:44
← eijk left15:45
milki RESNO: man ssh_config15:45
i dont use it myself15:45
RESNO ah ok15:45
→ dj_tjerk joined15:45
→ Lulzon joined15:45
← tr-808 left15:46
← rtjure left15:46
→ AeroNotix joined15:46
→ tr-808 joined15:47
→ cambrian joined15:48
→ AaronMT joined15:48
← AaronMT left15:48
← tr-808_ left15:48
→ AaronMT joined15:49
← AaronMT left15:49
← saju_m left15:49
→ ISF joined15:49
→ AaronMT joined15:49
← AaronMT left15:49
← tvw left15:49
→ tr-808_ joined15:49
→ cryptocompress joined15:50
→ AaronMT joined15:50
← AaronMT left15:50
← cryptocompress left15:50
→ AaronMT joined15:50
← Lulzon left15:51
← tr-808 left15:52
→ tr-808 joined15:53
→ robbyoconnor joined15:53
← nareshbhat left15:54
→ techhelp joined15:54
→ xtagon joined15:54
→ maletor joined15:54
→ Maccer joined15:54
← LimeBlast left15:55
← tr-808_ left15:55
→ nareshbhat joined15:55
← lledet left15:55
→ tr-808_ joined15:56
→ Cerise joined15:56
→ tyatpi_ joined15:56
← mxweas_ left15:56
→ ferdna joined15:56
← fly9 left15:56
→ mxweas_ joined15:58
← robotmay left15:59
→ aruiz joined16:00
← kerrick left16:00
devn !gitslave16:01
gitinfo gitslave (http://gitslave.sf.net) is useful to add subsidiary git repositories to a git superproject when you control and develop on the subprojects at more or less the same time as the superproject, and furthermore when you typically want to tag, branch, push, pull, etc. all repositories at the same time.16:01
devn !subtree16:01
gitinfo The git subtree merge method is ideal to incorporate a subsidiary git repositories directly in to single git repository with "unified" git history, where you only need to pull changes in from external sources not contribute your own changes back (which if technically possible is at least difficult). See http://progit.org/book/ch6-7.html Type "!subtree_alt" for more options16:01
← leeight_ left16:01
← whitman left16:02
← Nirvanko left16:03
← mxweas_ left16:05
→ barneybook joined16:05
← barneybook left16:05
→ mxweas_ joined16:06
→ williamcotton joined16:07
← dj_tjerk left16:07
→ Eridius joined16:07
→ andrewhl joined16:08
← maletor left16:08
→ TheJH joined16:08
→ barneybook joined16:08
← mklappstuhl left16:09
→ hwrd|work joined16:09
→ fly9 joined16:09
→ apok joined16:09
→ dj_tjerk joined16:10
→ mmc1 joined16:10
→ Targen joined16:11
← macrover left16:12
← barneybook left16:12
← mxweas_ left16:13
→ aantix joined16:13
← uraimo left16:13
→ mastro joined16:13
← aantix left16:14
→ aantix joined16:14
→ mxweas_ joined16:15
→ devth joined16:15
→ wilmoore joined16:16
elik I'm reading http://dotfiles.github.com and am left wondering why it's better to put your dotfiles in a subdir rather than just git init and git add the home directory directly16:16
← Pupeno_W left16:16
← kaputtmacher left16:17
→ maletor joined16:18
sente what scripts/hooks do people use such that the shell's prompt always shows the current branch and/or other git related info?16:19
RandalSchwartz depends on the shell16:19
sente bash16:19
within (usually) a virtualenv16:19
← tjadc left16:20
RICHIH sente: if you are serious about this, try zsh with vcs_info (built-in)16:20
sente RICHIH: I'm serious about learning git16:21
RICHIH is not aware of anything that works with bash and is as good16:21
sente I prefer one thing at a time16:21
also I use a lot of systems which don't have zsh installed I believe16:21
RICHIH sente: understandable16:21
but then, i can't help you, sorry16:21
PerlJam just uses git-completion.bash16:21
sente thanks for the feedback anyhow16:21
→ dvide_ joined16:21
RICHIH sente: no worries16:22
← axl_ left16:22
RICHIH PerlJam: that does not give you any info though, oes it?16:22
→ axl__ joined16:23
m1sc sente: check the comment: https://github.com/gitster/git/blob/master/contrib/completion/git-completion.bash16:23
PerlJam RICHIH: what sort of "info" are we talking about?16:23
← ixti left16:24
RICHIH PerlJam: 18:19:00 < sente> what scripts/hooks do people use such that the shell's prompt always shows the current branch and/or other git related info?16:24
sente m1sc: thanks16:24
PerlJam RICHIH: git-completion.bash shows the current branch and other info.16:24
devn I don't have a git subtree command. Is that part of an extra package or something?16:24
RICHIH at least in my case: branch, uncommited changes, ongoing merges, tag/name/commit id etc16:24
PerlJam: then its name is misleading :p16:25
PerlJam well, it also allows for completion of branch/tag names and git subcommands16:26
:)16:26
m1sc RICHIH: true, will be a separate script soon16:26
← etcetera left16:26
RICHIH wonders why that's not integrated into bash proper, but oh well16:26
m1sc devn: I think you have to install it separately: https://github.com/gitster/git/tree/master/contrib/subtree16:26
→ ajpiano joined16:27
← tyatpi_ left16:28
→ KBme joined16:28
→ Kicer86 joined16:29
← mxweas_ left16:30
← stkrzysiak left16:31
← angelsl left16:31
← i42n left16:31
→ mxweas_ joined16:33
→ mikepack joined16:33
→ mwrosen joined16:33
← robbyoconnor left16:34
→ robbyoconnor joined16:34
← _Qman left16:36
← mxweas_ left16:37
→ Chel joined16:38
→ ixti joined16:38
← yaymukund left16:38
Chel question: if i made git checkout rev , how can i undo this action to HEAD before this command ?16:38
→ mxweas_ joined16:38
→ sporous joined16:39
→ boombatower joined16:39
→ yaymukund joined16:39
→ nadavoid joined16:40
← danslo left16:41
→ Shusshu joined16:42
m1sc Chel: you can allways check the reflog for HEAD to see what happened before16:42
*always16:42
→ mklappstuhl joined16:43
→ JEEB joined16:43
← mklappstuhl left16:43
→ Goplat joined16:45
→ wubino joined16:47
→ QaDeS_ joined16:48
← BarryCrayon left16:48
← QaDeS_ left16:48
← robbyoconnor left16:49
← ixti left16:49
→ robbyoconnor joined16:49
← Targen left16:50
← yaymukund left16:50
← nadavoid left16:52
← mikepack left16:52
← Vile left16:52
→ nadavoid joined16:53
← nadavoid left16:53
→ AlbireoX`Laptop joined16:53
→ tewecske joined16:54
← fly9 left16:55
← plastics- left16:55
→ txomon joined16:56
txomon hi, I have a branch that comes from the master's head, and I would like to have *some* lines merged, not all of them. Is there anyway git doesn't use a merge strategy and represents all differences as conflicts?16:57
← AlbireoX`Laptop left16:57
← kevwil left16:57
shruggar you shouldn't use git's merge for partial merges16:59
txomon but I want to have that branch integrated in master16:59
merged*16:59
← andref_dias left17:01
→ kevwil joined17:01
txomon what would you suggest shruggar ?17:01
shruggar if I was doing a fake merge, I'd probably make the tree look however I wanted, then record the merge manually using git hash-object, I'm lame like that :/17:02
← sylr left17:02
shruggar not the most user-friendly solution17:03
→ AlbireoX`Laptop joined17:03
→ delight joined17:04
← alansaul left17:04
txomon definetly that is not the way I want to do it ;)17:04
found this17:05
http://www.gelato.unsw.edu.au/archives/git/0701/37964.html17:05
← neurodrone left17:05
→ etcetera joined17:07
← delight left17:10
→ mishok13 joined17:12
← stuntmachine left17:13
→ danslo joined17:13
← Shusshu left17:15
← robbyoconnor left17:15
→ martinjlowm joined17:15
→ raatiniemi joined17:15
→ JStoker joined17:15
→ Macca- joined17:16
← project2501a left17:17
← mxweas_ left17:17
Macca- if someone makes changes on a branch (branch X) then merges it back into branch Y.. is it possible to reverse all changes that were merged in, if development continued on branch Y after the merge, without losing those subsequent changes?17:17
→ btree joined17:18
→ mxweas_ joined17:18
→ rking joined17:18
txomon Macca-: see man git revert17:19
gitinfo Macca-: the git-revert manpage is available at http://git-scm.com/docs/git-revert.html17:19
→ Linc joined17:19
← Linc left17:20
Macca- ta17:20
→ Linc joined17:20
txomon ta?17:20
Macca- thanks17:21
txomon ah ok17:21
← rking left17:22
→ rking joined17:22
← davidg left17:23
→ dgasson joined17:23
← mishok13 left17:24
→ tinti joined17:24
← mxweas_ left17:24
← jkremser left17:25
→ kerrick joined17:25
← dpino_ left17:27
→ mxweas_ joined17:27
→ whitman joined17:28
→ Xizor joined17:28
→ tyatpi joined17:30
← heroux left17:30
strobert loves git-svn17:30
→ mtytel joined17:31
← tomislater left17:31
← mxweas_ left17:32
← marcogmonteiro left17:32
→ heroux joined17:32
→ Vile joined17:33
← jargon- left17:33
← dgasson left17:33
→ dgasson joined17:34
→ mxweas_ joined17:34
EugeneKay BURN THE HERETIC17:35
← Macca- left17:36
→ linguini joined17:36
← mxweas_ left17:38
← heroux left17:39
← Joseph__ left17:39
→ heroux joined17:39
← ehsan left17:39
→ ehsan joined17:40
← ipalaus left17:40
← Tinti_ left17:40
→ rgsteele joined17:40
← tinti left17:41
→ tinti joined17:41
→ cjz joined17:42
cjz what is handling the git+ssh type url ?17:42
→ spaceonline joined17:42
cjz is that gitolite or just plain old git handling that17:42
→ OOPMan joined17:43
← heroux left17:44
→ mxweas_ joined17:44
← tinti left17:45
→ tinti joined17:46
← tinti left17:46
→ tinti joined17:47
← OOPMan left17:47
← ferdna left17:48
PerlJam cjz: git17:48
cjz ah17:49
← mxweas_ left17:49
→ dreiss joined17:49
→ mxweas_ joined17:50
→ stuntmachine joined17:50
← marhaban left17:51
→ jargon- joined17:51
→ EvanDotP1o joined17:52
→ heroux joined17:52
← EvanDotP1o left17:53
→ ffwacom joined17:55
← mxweas_ left17:55
ffwacom git17:55
→ Sigma joined17:55
→ mxweas_ joined17:55
→ rryoumaa joined17:56
gitinfo set mode: +v17:56
← danslo left17:57
→ toabctl joined17:57
→ danslo joined17:58
→ neurodrone joined17:58
← neurodrone left17:58
→ neurodrone joined17:58
← toabctl left17:58
← mxweas_ left18:00
← heroux left18:01
→ ciao joined18:01
→ mxweas joined18:02
→ OOPMan joined18:03
← kevwil left18:04
← danslo left18:05
← ciao left18:06
← cortana2 left18:06
→ puneetgoyal joined18:06
← mxweas left18:06
→ mxweas_ joined18:08
→ imachuchu joined18:08
→ heroux joined18:08
→ tsuMi joined18:08
← mns left18:09
← tatsuya_o left18:10
← intransit left18:10
→ intransit joined18:11
→ miceiken joined18:11
miceiken is it any way to force git to remove files matching .gitignore after they've been pushed?18:11
← ffwacom left18:12
← mxweas_ left18:12
→ kevwil joined18:12
→ yaymukund joined18:12
→ fisted joined18:12
← Sigma left18:12
PerlJam miceiken: what do you mean?18:13
miceiken I already commited all the files in my git folder, then just now I made a .gitignore file. However all the files I don't want in the remote repo are already pushed18:14
so is there any way to remove them from the repo but still keep them locally, without having to do that manually?18:14
→ mxweas_ joined18:14
PerlJam miceiken: git rm --cached18:15
miceiken what more?18:15
→ Sigma joined18:16
→ andref_dias joined18:16
miceiken ah thanks PerlJam18:18
← mxweas_ left18:19
← yaymukund left18:20
→ mxweas_ joined18:20
← andrewhl left18:21
→ danslo joined18:22
→ mns joined18:22
← noisebleed left18:23
strobert EugeneKay: I love pure git better, but better git-svn than having to use svn directly :)18:23
EugeneKay raises pitchfork18:23
← mxweas_ left18:24
→ Ste-Jay joined18:25
← SkAzZ1 left18:25
gitinfo set mode: +v18:26
→ rtjure joined18:26
← miceiken left18:26
Ste-Jay Hi all - I'm working on a backup system using rsync, and I'm considering using GIT to track changes between snapshots (thus allowing restoration of old versions of the files or files which have been deleted)...18:27
I read somewhere that GIT shouldn't be used for projects larger than 2GB - My backup currently stands at 81GB - this will be reduced somewhat as I am pruning our files down and removing a lot of unneeded data - none the less I expect to have several GB of data which would need processing18:28
Are there any special considerations using GIT on such a large directory?18:28
← gusnan left18:29
← tsuMi left18:29
← ph^ left18:30
EugeneKay !GIT18:30
gitinfo Git is not an acronym. Writing it in all caps is heresy and will be punished by Subversion or SCCS for more grievous offences. It's "git" or "Git". Thank you for your help in making the world a better place.18:30
EugeneKay !big18:30
gitinfo [!large_files] Git isn't yet great at large files(larger than RAM). Look at !annex for solutions. You can find them (after gc) with: git verify-pack -v .git/objects/pack/pack-*.idx | grep blob | sort -k3nr | head | while read s x b x; do git rev-list --all --objects | grep $s | awk '{print "'"$b"'",$0;}'; done18:30
→ mxweas_ joined18:30
EugeneKay !big_repos18:30
Oh, wtf is it.18:30
!large_repos18:30
gitinfo Git can be slow in the face of large repositories. There are git-config options which can help. pack.threads=1; pack.deltaCacheSize=1; pack.windowMemory=512m; core.packedGitWindowSize=16m; core.packedGitLimit=128m. Other likely ones exist.18:30
→ danslo_ joined18:30
EugeneKay !annex18:31
gitinfo Storing lots of binary files in git causes repo balloon. git-annex and git-media are two solutions to work around this by keeping the blobs out of the repo. http://git-annex.branchable.com/ and https://github.com/schacon/git-media18:31
→ gusnan joined18:31
Ste-Jay awesome - many thanks18:31
EugeneKay Ste-Jay - git can be used as a file store for a backup system, but rsync and hardlinking works better for it, IMO.18:31
→ andrewhl joined18:32
Ste-Jay *nods* i'm using rsync for the actual synchronisation - I thought git might have been a more elegant solution than hardlinking though, especially in the face of the Windows servers I'm forced to tolerate18:32
EugeneKay It will balloon up quickly, and that causes issues.18:33
Especially if you're talk about binaries18:33
← danslo left18:34
danslo_ → danslo18:34
← mxweas_ left18:35
→ gitinner joined18:35
Ste-Jay *nods* i suspect about 90% of our data won't need to be tracked for changes, so it's still possible that we could use it - mostly I'm interested in tracking our staff's files (excel, word, etc) and some dbf databases, so it is mostly binary info18:36
→ mxweas_ joined18:36
EugeneKay As silly as it sounds, svn will probably do a better job of that.18:36
Ste-Jay yeah, i did think of svn first but like most people I much prefer git - sometimes it's hard choosing the right tool for the job rather than the one that fits best in your hand :)18:37
anyway, you've been very helpful and therefore are a star. Thanks very much :D18:37
← gitinner left18:38
→ gitinner joined18:39
→ cbreak joined18:39
EugeneKay Sure.18:40
→ kyphi joined18:40
← mxweas_ left18:40
EugeneKay git-annex MIGHT work, but for windows homedirs.... well.... no. :-p18:40
cjz lord gitlab is difficult to setup18:40
jast lemme guess... rails-based?18:41
← stodan left18:41
cjz yeah18:41
← kyphi left18:42
→ mxweas_ joined18:43
→ fly9 joined18:44
← Chel left18:44
→ ph^ joined18:45
← gitinner left18:45
→ thoraxe joined18:46
EugeneKay jast - there was some complaint about using the git-scm man page reference. Apparently they stripped the command directory from 'man git'18:47
gitinfo jast: the git manpage is available at http://git-scm.com/docs/git.html18:47
← mxweas_ left18:47
→ Axsuul joined18:47
→ mxweas_ joined18:49
→ Macca- joined18:49
→ BarryCrayon joined18:49
Macca- is there any way to query a git repository for a list of commits without cloning it?18:49
→ j416 joined18:49
EugeneKay No, not really.18:49
← toothr left18:49
→ Chillance joined18:50
→ itsrachelfish joined18:51
itsrachelfish Hi #git!18:51
gitinfo itsrachelfish: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.18:51
itsrachelfish Oh thanks mister git robot18:51
→ toothr joined18:51
← JStoker left18:52
itsrachelfish Is there any option for git diff to only display additions?18:52
← mxweas_ left18:53
→ JStoker joined18:53
bremner itsrachelfish: you can probably use filterdiff for that18:53
→ stickyboot joined18:53
← BarryCrayon left18:53
bremner or grep18:53
→ \toothrot joined18:53
j416 itsrachelfish: git diff --diff-filter=A18:53
for added files18:54
added lines? not sure if A applies18:54
itsrachelfish Yeah added lines18:54
I suppose *could* just regex ^+18:54
j416 git diff | egrep '^\+'18:54
or something18:54
→ mxweas_ joined18:54
j416 heh18:55
Ste-Jay ok guys and gals - thanks for all the help, see you soon :D18:55
← Ste-Jay left18:55
→ gdoteof joined18:55
→ Targen joined18:56
← toothr left18:56
→ yaymukund joined18:56
→ Gitzilla joined18:56
← txomon left18:57
← tr-808_ left18:57
← tr-808 left18:57
→ tr-808 joined18:58
← mxweas_ left18:59
→ drizzd_ joined18:59
→ thearthur joined19:00
← kevwil left19:00
→ mxweas_ joined19:00
← kerrick left19:01
→ kerrick joined19:02
← diwic left19:02
→ v0n joined19:02
← \toothrot left19:03
→ toothr joined19:04
← mxweas_ left19:05
→ avpx joined19:05
Olipro if I do a "git clone --mirror" of a git-svn repo, how do I update the mirror after an svn rebase19:06
→ flijten joined19:06
← flijten left19:06
← thearthur left19:06
→ jeyanth joined19:06
→ Joseph__ joined19:07
← fisted left19:08
← rgsteele left19:08
→ codeM0nK3Y joined19:09
→ scanf joined19:10
→ mxweas_ joined19:10
→ gjeoffroy joined19:11
← toothr left19:11
→ spaceonline__ joined19:11
← Macca- left19:11
← spaceonline left19:12
← gjeoffroy left19:12
← dfr|mac left19:13
→ toothr joined19:13
→ teknopaul joined19:14
→ dfr|mac joined19:14
← mxweas_ left19:14
heoa How to use Milestones in Ticgit?19:15
← Sigma left19:15
← AeroNotix left19:16
→ mxweas_ joined19:16
→ SmogMonkey joined19:16
avpx Olipro: Couldn't you just a) mirror again, or b) fetch the changes from your repo with git fetch?19:16
→ mdhopkins joined19:18
avpx Olipro: The latter option is a bit harder because you will have some sequence on the mirror like:19:19
A -- B -- C19:19
And on your working repo like:19:19
A' -- B' -- C' -- D19:19
Olipro git push --mirror worked19:19
← Lgb left19:20
← palm_vue left19:20
avpx Oh, good.19:20
← mxweas_ left19:20
→ kevwil joined19:21
→ palm_vue joined19:21
→ mxweas_ joined19:21
→ Lgb joined19:21
← SmogMonkey left19:22
← spaceonline__ left19:23
← mxweas_ left19:25
→ rtjure_ joined19:27
→ spaceonline joined19:27
→ mxweas_ joined19:28
← danslo left19:28
← toothr left19:29
→ danslo joined19:29
← rtjure left19:29
← mxweas_ left19:32
→ ivan12345 joined19:32
→ ivan12346 joined19:33
← jeyanth left19:33
→ mxweas_ joined19:34
→ bronson joined19:34
→ toothr joined19:34
← boombatower left19:35
→ Garrick joined19:35
→ davidjdagino joined19:36
Garrick how should I approach a situation where I want to create a new git repo from an existing repo's branch19:37
→ dorkmafia joined19:37
← andrewhl left19:37
→ AeroNotix joined19:37
dorkmafia how come araxis isn't loading a three way merge by default anymore?19:37
→ ivan12347 joined19:38
Owner Garrick~# uhhh couldnt it be as simple as changing to the branch, copying all files to a new directory, changing ot that directory, and deleting .git/, then git init ?19:40
← mxweas_ left19:40
→ Jay_Levitt joined19:40
← jcromartie left19:41
canton7 Garrick, duplicate (clone) the repo, and delete the other branches19:41
Owner Garrick~# that would be a new git repo...if you want some kind of linkage, i dont know the best practicies19:41
Garrick Owner: but how would you retain the history?19:41
→ jcromartie joined19:41
→ Yaodin_ joined19:41
Garrick canton7: that could do it.19:42
→ andrewhl joined19:42
→ mxweas_ joined19:42
→ Nikoli_ joined19:42
→ booginga joined19:43
→ harshpb joined19:43
→ matthavener joined19:44
Olipro if I'm using git-svn, can I safely take two branches say "branchA" and "branchB" and simply merge B into A and then take that and rebase against master?19:44
matthavener is it possible to list the files in a remote (http://) repo without cloning it?19:44
cehteh no, but you could do a shallow clone19:45
matthavener cehteh: ah, ok, i think that'll do. thanks :)19:46
Owner what is a shallow clone19:46
matthavener Owner: git clone --depth 119:47
← kerrick left19:47
matthavener basically you get the repo, but you don't get any history, just the last commit (that's my limited understanding anyways)19:47
→ apok_ joined19:47
Owner hmm and what is --bare19:48
← apok left19:48
apok_ → apok19:48
Olipro Owner: it builds the git repo but doesn't actually dump out the files themselves19:49
← shruggar left19:49
Owner ah, so the opposite19:49
→ teweWORK joined19:49
Olipro no19:49
you could have a shallow, bare repo19:49
Owner also, how is this useful http://git-scm.com/search/results?search=--bare19:50
Olipro the point of a bare repo is when you want a location for everyone to be able to push and pull to/from19:50
← andref_dias left19:50
→ kerrick joined19:50
← jh3_ left19:50
Olipro for a common commit point, you're unlikely to want the files in the working tree to be written to the FS19:50
← kerrick left19:51
heoa How can I see Github -style punchcards and timeline of git -repo?19:51
→ kerrick joined19:51
→ jeyanth joined19:51
Owner Olipro~# interesting......19:52
← rryoumaa left19:52
→ pantsman joined19:52
← pantsman left19:52
→ pantsman joined19:52
→ bartman`` joined19:53
→ kaen joined19:53
→ comps joined19:54
← ajpiano left19:54
→ ajpiano joined19:55
→ aidan joined19:56
aidan I tried to do a git cherry-pick and it conflicted, how do I make it like I never typed that command?19:56
→ u37 joined19:56
aidan I tried to revert it but it won't, because it's conflicted and I have local changes19:56
finch git reset --hard HEAD19:57
comps cherry-pick --abort19:57
or git rebase --abort or whatever it uses as "backend"19:58
← matthavener left19:58
← kaen left19:58
EugeneKay !bare19:58
gitinfo an explanation of bare and non-bare repositories (and why pushing to a non-bare one causes problems) can be found here: http://bare-vs-nonbare.gitrecipes.de/19:58
aidan comps: no --abort on cherry-pick and "no rebase in progress" on git rebase19:59
→ Renari joined19:59
← bgerber left19:59
→ thearthur joined19:59
comps then just hard reset19:59
and possibly clean -fdx20:00
aidan reset --hard HEAD^ seemed to work20:00
lucky I backed up my local changes20:00
otherwise tears.20:00
← Targen left20:01
gitinfo set mode: +v20:01
Owner git...as if programming wasn't already hard enough20:01
u37 probably not a git question. anyway, i have 2 repo(s) that use the same config file, where do i save it? Do people create another repo for storing config files?20:01
Olipro so, anyone have comment on whether it's safe to merge two branches together and then rebase against master for the purposes of git-svn20:01
→ danslo_ joined20:02
comps Olipro: -p, --preserve-merges20:02
for git-rebase20:02
imachuchu Owner: what? Git is fairly simple once you realize what's all going on.20:02
Olipro comps: and that's desirable when subsequently doing the "git svn dcommit" ?20:03
Owner imachuchu~# funny...its easier to pickup a new language than realize what's all going on20:03
comps Olipro: well, I always dcommited only linear history, so no idea on that20:04
Olipro right20:04
comps perhaps dcommit can make svn branch merges as well20:04
Olipro right, I want to do the same for SVN20:04
so, I'm planning to be using master as the branch to make svn dcommits from20:05
← cedeon left20:05
Olipro so say I've got BranchA20:05
I do some stuff20:05
→ plastics joined20:05
imachuchu Owner: !simple20:05
gitinfo Owner: At it's heart, git is made up of many concepts that are individually simple. Getting the whole picture right is often tricky, and it's usually about breaking up the complex concept into its simple, individual parts and grokking those.20:05
Olipro then branch to BranchB from BranchA20:05
do more stuff20:05
can I then merge BranchB into BranchA20:05
then, rebase BranchA against master20:05
← mxweas_ left20:05
Olipro then svn dcommit master20:05
← danslo left20:06
→ danslo__ joined20:06
→ wabash joined20:07
wabash Is there a way to merge through mergetool when there are no conflicts?20:07
← Renari left20:07
Owner imachuchu~# so how do i remove all cusswords from all previous commit messages ?20:08
→ mxweas_ joined20:08
← danslo_ left20:08
← j416 left20:08
→ nelk joined20:08
comps Olipro: "The merge/dcommit problem is that whenever you 'git svn dcommit' a branch, the merge history of that branch is 'flattened': git forgets about all merge operations that went into this branch: Just the file contents is preserved, but the fact that this content (partially) came from a specific other branch is lost."20:09
← ISF left20:09
imachuchu Owner: that's a very complex question, what is a cussword?20:09
Owner imachuchu~# like vulgar words20:09
comps Olipro: http://stackoverflow.com/questions/190431/is-git-svn-dcommit-after-merging-in-git-dangerous , quite long answer, 4th post20:09
← cjz left20:10
→ Beryllium joined20:11
→ delinquentme joined20:11
Olipro comps: right, thus why I want to rebase branchA onto master and dcommit whilst on master20:11
delinquentme forked a repo ... repo was updated ... want to get the most recent version of whats running on the repo .. on my fork20:11
→ Renari joined20:11
delinquentme how do efficient?20:11
I could delete my fork and refork .. but im pretty sure theres a better way.20:11
← erichynds left20:11
Beryllium hey, um. I just did a "git add" to a local repository and it added some .svn folders that were hidden in subfolders. Is there any way I can remove them from my git add (I haven't committed yet) without removing them from disk?20:11
bremner Owner: the short answer is man git-filter-branch20:12
gitinfo Owner: the git-filter-branch manpage is available at http://git-scm.com/docs/git-filter-branch.html20:12
comps Olipro: ? .. I thought your question was purely theoretical, since one would probably rebase branchA on top of master and branchB on top of the result, and then use dcommit anyway20:12
u37 well looks like subtree merges is one way20:12
Olipro no, not theoretical20:12
Beryllium nevermind, I fixed it :) Purged the empty git repo folder and made a .gitignore and redid my stuff. Nice.20:12
Olipro say I make a bunch of structural changes on branch B that will require conflict resolution and merging20:13
imachuchu Owner: ahh, well git doesn't know that the word "poop" is acceptable while other derivatives of it are not. Do you have a list of all words you think are cusswords?20:13
Olipro *and would benefit from merging20:13
Owner imachuchu~# how can we teach git that poop is acceptable? can you make a patch?20:13
comps dcommit probably commits the merge resolution in the merge commit20:13
Owner bremner~# interesting20:13
Olipro if I merge A and B, I would hope I can safely rebase that onto master20:14
← mxweas_ left20:14
imachuchu Owner: I'll get right on that, a bigger problem is the word "Hell". Is it a cussword?20:14
comps Olipro: you can, with -p20:14
← heptadecagram left20:14
comps but you can't push that nonlinear history to svn20:14
Olipro but20:15
when I rebase it onto master20:15
imachuchu Owner: if you know all words you want to replace in all history the command filter-branch can be used to remove things in the past. All to often though it's moot, just change it today and let the past be the past20:15
comps right, conflicts, hmm20:15
Olipro master should still be linera20:15
*linear20:15
all the crap I've done on BranchA and BranchB will just be a single commit on the master branch20:15
courtesy of the rebase20:15
→ edude03 joined20:16
EugeneKay !enter20:16
gitinfo The enter key is not a punctuation mark. Please use fewer lines, with complete sentences on them.20:16
Owner imachuchu~# i removed all history to solve it20:17
→ mxweas_ joined20:17
Owner is each commit treated as a branch or something?20:17
because i wanted to rewrite commits not branches20:18
← nelk left20:18
← enmand left20:18
imachuchu Owner: a solution to be sure20:18
→ Eridius_ joined20:19
← Beryllium left20:19
EugeneKay Which commit are you wanting to fiddle? The last one?20:19
Owner imachuchu~# yes there was lots of wailing and gnashing of teeth before version 1.020:19
PerlJam Owner: git filter-branch is perhaps a little misnamed, but it does what you want20:19
← dvide_ left20:19
→ fisted joined20:19
Owner awesome....good to know there is an easy way20:20
← dfr|mac left20:21
→ nelk joined20:21
→ codewrangler joined20:21
← Renari left20:21
→ Renari joined20:21
← mxweas_ left20:21
Garrick thx for the help, guys.20:21
→ boombatower joined20:22
← Garrick left20:22
→ mxweas_ joined20:23
← dorkmafia left20:24
→ mcluskydodallas joined20:24
→ dorkmafia joined20:24
← harshpb left20:25
→ Koganei joined20:25
avpx PerlJam: It seems like it should be map-branch20:25
jast but it can remove commits, too20:26
or add new ones20:26
both 'filter' and 'map' are similarly inaccurate :)20:27
← u37 left20:27
← mxweas_ left20:27
→ Lubinski joined20:28
→ Kogsworth joined20:28
← nelk left20:28
→ starsinmypockets joined20:29
starsinmypockets Is it possible to do a `git pull remote master` while maintaining other local development branches?20:30
→ shruggar joined20:30
cmn other branches won't be affected by the merge20:30
← Koganei left20:30
cmn also, !fetchfour20:30
gitinfo Never use the four-word version of git-fetch or git-pull (e.g. git fetch remote refspec). It always ends in tears. Yes, if you understand the implications of FETCH_HEAD it can technically be done, but really it is easier to just fetch the whole remote (or perhaps edit the fetchspec if you never want other bits).20:30
jast personally I think it's okay with pull, if not ideal20:30
→ mxweas_ joined20:31
jast anyway, merges only affect the branch checked out at the time20:31
starsinmypockets Thanks, so if I go fetch the remote HEAD and merge it with the master branch, my dev branches remain unchanged?20:32
cmn it's OK if you know what it does20:32
but most people expect the remote-tracking branch to be updated, and then come asking why 'git pull origin master && git diff origin/master' is confusing them20:32
→ aspotashev joined20:33
cmn starsinmypockets: other branches are not affected by a merge you do in a branch20:33
← Renari left20:33
→ ew0 joined20:34
← mabrand left20:34
← chrisf|work left20:34
← mxweas_ left20:35
→ Renari joined20:36
→ mxweas_ joined20:36
→ nelk joined20:36
Hunner Hmm. Why does git-svn not create a remote name for the svn remote branches?20:36
Super confusing when I have local branches named the same as the remote branches20:37
jast it's git-svn20:37
'nuff said20:37
cmn they're not remote branches, they're refs with odd names20:37
Hunner :P20:37
cmn it's a bit of tape that's holding way too many moving parts together20:37
Hunner Oh! You can add a remote name with --prefix when doing the clone20:38
Stackoverflow is invaluable sometimes :)20:38
← stuntmachine left20:38
cmn then man git svn should also tell you20:39
gitinfo the git-svn manpage is available at http://git-scm.com/docs/git-svn.html20:39
→ theCole joined20:39
cmn and you don't have to scour a possibly unreliable source20:39
Hunner cmn: Yeah, but it's usually a case of not knowing exactly what I'm looking for20:39
Human search engine for the win :)20:39
cmn you can still, read, can't you?20:39
jast to be fair, reading git manpages from top to bottom is usually a pretty time-intensive task :)20:40
Hunner cmn: I'm not a support leach. No need to get testy :/20:40
cmn huh?20:40
Hunner Never mind... thanks for the help :)20:40
← mxweas_ left20:40
← gour left20:41
→ gour1 joined20:41
← Destos left20:41
→ mxweas_ joined20:42
← ajpiano left20:42
Hunner is this close -| |- to actually getting them add 'fully convert from svn to git' to the roadmap20:42
→ harshpb joined20:42
jast my workplace switched over to git shortly before I joined20:42
now I just have to get them to have a separate repository for each component20:42
→ dfr|mac joined20:43
Hunner is a consultant, so it changes from week to week :)20:43
jast in fact I think I've already convinced them... just need to make it happen when I find some time20:43
Hunner jast: Hah :)20:43
wabash So, say I want to merge something, and it ends up being a fast forward merge. No conflicts. Now, is there a way to review and alter it, like using mergetool, before the FF is performed?20:44
jast wabash: no, but you can undo it if you don't like it (just don't push it yet)20:45
wabash jast: Yes, that will work.20:45
← segher left20:45
→ lightcap joined20:45
← ph^ left20:45
← lightcap left20:45
→ ajpiano joined20:45
→ UukGoblin joined20:46
jast unless you change something else between the merge and the undo, 'git reset --hard HEAD@{1}' should do the trick (destroys uncomitted changes etc.)20:46
wabash jast: But I'm looking for something like either the mergetool, or else a way to do something similar to add --patch, where I can review and patch line by line, or chunk by chunk.20:46
jast: Yes, I thought about that too.20:46
→ ph^ joined20:46
← cakehero left20:46
→ segher joined20:46
wabash jast: Just that, diff is sometimes hard to review for goodness, especially if there are so many lines changed!20:46
← mxweas_ left20:46
jast well, I suppose you could force a non-ff merge and use -n20:46
wabash jast: what's -n? For non-op?20:47
jast: how do I force a non-ff merge?20:47
jast "don't actually commit the merge yet"20:47
--non-ff20:47
err, --no-ff20:47
→ bgerber joined20:47
jast as mentioned in the manpage :)20:47
→ stkrzysiak joined20:47
wabash jast: Sure.20:48
jast: Ok, thank you.20:48
← segher left20:48
jast you're welcome20:49
→ SmogMonkey joined20:49
→ segher joined20:49
← segher left20:50
← ToBeFree left20:50
← ph^ left20:50
→ segher joined20:50
→ jh3 joined20:50
→ mxweas_ joined20:52
← consti left20:52
comps Olipro: http://pastebin.com/Ps36sGN2 .. the hashes in both examples are different because of the rebase and because of the fact that I had to commit --allow-empty each commit, since I made empty commits for this example and rebase doesn't pass that flag by default20:54
← segher left20:54
comps however you're right that it can't be produced using simple rebase command with basic syntax, which really surprised me20:54
→ segher joined20:55
← segher left20:56
comps it could be probably simplified using various revspecs20:56
← etcetera left20:56
← mxweas_ left20:56
← Xizor left20:56
← und3f left20:57
→ mxweas_ joined20:57
→ Koganei joined20:57
→ jjbohn joined20:58
→ und3f joined20:58
← Koganei left20:58
→ segher joined20:58
← TheHaven left20:59
← Kogsworth left21:00
→ opdynamiks joined21:00
← gour1 left21:00
← nelk left21:01
← mxweas_ left21:02
→ marhaban joined21:02
← Renari left21:02
→ hayward joined21:02
→ yakovv joined21:02
→ mxweas_ joined21:03
→ nelk joined21:03
→ Renari joined21:03
Olipro how would I go about copying/cloning a git-svn repo from a remote server such that everything is identical save for the URI to the SVN store21:03
would it be sufficient to just tarball up the whole project folder, extract at my end and then manually alter the URI to the SVN repo?21:04
← _iron left21:05
→ jexmex joined21:05
← hayward left21:07
jexmex I do some modifications inside the submodule directory, but for some reason my branch was not checked out (it was before, I know submodules are headless by default). Anyways I now need to get my changes over to the branch, what can I do?21:07
← mxweas_ left21:07
comps Olipro: I don't understand the question, if you want to transfer "commits", you can use git-bundle21:08
→ mxweas_ joined21:08
→ airborn joined21:08
comps Olipro: if you git-svn clone the svn repo to another location, commit hashes will remain the same, feature of git-svn21:08
Olipro cloning an svn repo into git is slow - I want to clone from svn on the local machine and then transfer to the remote21:09
comps then just push git-svn ref21:09
(or pull from it, that is)21:09
← intransit left21:09
Olipro ok, it's a little more complex; I have git branches on this repo already that I've been working on21:09
I don't want the git repo on my local machine to treat those git branches as remote21:10
← heroux left21:10
Olipro I want it to *only* interact with the SVN repo21:10
comps you can probably use git-svn in combination with bare repos, ie. "git-svn fetch" ("git-svn rebase" is basically "git-svn fetch ; git-rebase git-svn"21:10
→ intransit joined21:10
comps sort of a cron-automated svn->git bot21:11
← bgerber left21:11
← hoschi left21:11
→ goshawk joined21:11
← SmogMonkey left21:11
→ nicc joined21:11
Olipro I think I'll just tarball the dir21:12
comps git for-each-ref should list that git-svn ref, git branch -a should do so as well21:12
Olipro transfer to myself21:12
then modify the URI from file:// to ssh21:12
→ kogent joined21:12
→ heroux joined21:12
haarg modify the svn uri?21:13
comps I probably didn't understand the question correctly21:13
devn I have a sha and I want to get its symbolic name. It looks like `git name-rev 123fbbcdefg` will work, but is there anything that gives back ONLY the symbolic name and not the SHA along with it?21:13
← mxweas_ left21:13
Olipro haarg: right21:13
← cambrian left21:13
haarg did you use rewrite-root?21:13
because otherwise that sounds problematic21:13
Olipro comps: do a git-svn clone on a remote server, put the repo on my local machine in such a way that it was as if I ran the clone from my local machine in the first place21:13
haarg the URI is embedded in all of the commit messages21:14
cmn devn: pipe it to cut and select the first field21:14
comps Olipro: then you're right, removing all the files (except .git), doing git-gc and using tar is probably the easiest way21:14
Olipro [svn-remote "svn"]21:15
url = file:///svn/myrepo21:15
is it not sufficient to just change that?21:15
← avinashhm left21:15
→ etcetera joined21:15
→ mxweas_ joined21:15
← heroux left21:16
→ plasticle joined21:16
haarg the URL becomes part of the commit messages, which impacts the hashes of every commit. i'm not sure what happens if you just change it, but i suspect bad things.21:16
→ ph^ joined21:17
haarg there is the --rewrite-root option to clone and related config setting to use a different URL than where you are actually fetching from in the commit messages, and after the initial clone you can remove the rewrite-root config option and change the url.21:18
← yakovv left21:18
→ HyperNerdV2 joined21:18
← argoneus left21:18
cmn devn: actually, see man git name-rev, it's right there21:18
gitinfo devn: the git-name-rev manpage is available at http://git-scm.com/docs/git-name-rev.html21:18
← Jay_Levitt left21:19
← mxweas_ left21:20
→ bgerber joined21:20
← harshpb left21:21
→ mxweas_ joined21:23
← bartek left21:23
← hellopat left21:23
← sebrock left21:24
Olipro haarg: the crux of my issue is that I need to clone in such a way that the branches on the remote repo are considered local in the clone, rather than remote21:24
→ ISF joined21:25
→ Destos joined21:25
← nelk left21:25
→ nelk joined21:26
← nicc left21:26
haarg just create a local branch for each remote branch after fetching?21:26
← plasticle left21:26
← martinjlowm left21:26
→ Tor_ joined21:26
Olipro ok, and what about also pulling it across so that it's svn savvy21:26
wereHamster Olipro: you mean clone --mirror ?21:27
Olipro no, because that will result in a bare repo21:27
← mxweas_ left21:27
Olipro although...21:27
what if I do a --mirror and change the bare flag to 0 afterwards?21:27
wereHamster then create a non-bare repo with a fetch refspec of refs/*:refs/*21:27
→ nicc joined21:28
→ mxweas_ joined21:29
→ avinashhm joined21:29
ndim I am trying some operations with grafts or git-replace - and in both cases, "gitk --all" shows the intended new inheritance information, but a "git clone ssh://..../" of the modified repo does not contain the graft/replace information any more.21:29
Is it intentional that "git clone" does not copy that information?21:30
← mcluskydodallas left21:30
← nicc left21:30
→ eletuchy joined21:31
← Destos left21:32
→ Mowsh joined21:32
← Tor_ left21:32
← mxweas_ left21:33
← juanfer left21:34
← Renari left21:34
← plastics left21:34
→ mxweas_ joined21:34
← nelk left21:34
→ tatsuya_o joined21:34
→ Destos joined21:35
→ psoo joined21:35
→ zero1337 joined21:35
→ nicc joined21:36
→ enmand joined21:37
→ heroux joined21:38
← Lubinski left21:38
← mxweas_ left21:39
→ mxweas_ joined21:40
← booginga left21:42
← mxweas_ left21:44
← rtjure_ left21:45
→ mxweas_ joined21:45
← jjbohn left21:49
← mxweas_ left21:50
→ s0ber_ joined21:50
→ mxweas_ joined21:51
→ DanielConvissor joined21:51
← neurodrone left21:51
→ avandenhoven joined21:52
s0ber_ → s0ber21:52
← egarcia left21:52
→ nelk joined21:52
→ _W_ joined21:52
← imachuchu left21:53
DanielConvissor hi folks. just published some handy hooks and shell scripts for deploying websites and synchronizing databases via git and ssh. perhaps some folks here will find it useful: https://github.com/convissor/git_push_deployer21:53
← rendar left21:54
→ JoshuaA joined21:55
← mxweas_ left21:55
← AeroNotix left21:55
→ egarcia joined21:55
→ mxweas_ joined21:56
→ Renari joined21:57
← TheJH left21:57
→ ben_alman joined21:59
← JoshuaA left22:00
← mxweas_ left22:01
← superdmp left22:01
← beniamin left22:01
← ew0 left22:02
avandenhoven DanielConvissor: Looks interesting. I may have to borrow large chunks of it for my own deploy scripts.22:02
→ mxweas_ joined22:02
→ superdmp joined22:02
← superdmp left22:02
→ superdmp joined22:02
→ ew0 joined22:03
→ madewokherd joined22:03
← blacktulip left22:04
DanielConvissor avandenhoven: go for it. fork away!22:05
→ axl_ joined22:05
← pantsman left22:05
DanielConvissor avandenhoven: if you have any feedback, or pull requests, that'd be great.22:06
avandenhoven DanielConvissor: it seems that this expects you to have the repo cloned on the remote22:06
← andrewhl left22:06
← mxweas_ left22:06
DanielConvissor avandenhoven: yeah, need to start on the remote so the hooks are in place.22:06
arooni-mobile why does git insist on using nano; i want to use vim in commit messages22:07
avandenhoven DanielConvissor: I mean you can't use a bare repo22:07
← ph^ left22:07
→ mcluskydodallas joined22:08
→ mxweas_ joined22:08
avandenhoven DanielConvissor: I doubt anything I do will be generally useful. I'm trying to take a _rational_ git repo and ftp changes to an irrational server.22:08
→ ph^ joined22:08
← axl__ left22:08
← axl_ left22:08
haarg arooni-mobile: set GIT_EDITOR, VISUAL, EDITOR, or the core.editor config option22:08
arooni-mobile thx22:08
→ Fandekasp joined22:08
→ archdori joined22:08
→ axl_ joined22:09
← dv310p3r left22:09
→ shruggar1 joined22:10
← shruggar left22:10
→ chrisf|work joined22:11
→ axl__ joined22:11
DanielConvissor avandenhoven: need a checked out repo to reflect the files out to.22:11
← whitman left22:11
→ Anxi80 joined22:12
← mxweas_ left22:12
← ph^ left22:12
→ mxweas_ joined22:13
← tqmz left22:13
→ MrMagic joined22:13
avandenhoven DanielConvissor: yeah. I suppose. I have 2 branches to do this for so that means I have the full repo twice at all times (and I only need the files when I do a build since I have to FTP files over) but this arrangement might not be a bad idea.22:13
← mns left22:14
← axl_ left22:14
axl__ → axl_22:14
← airborn left22:14
← wilmoore left22:16
← john2496 left22:16
← macabre left22:16
→ ph^ joined22:17
← mxweas_ left22:18
→ stkrzysiak_ joined22:18
DanielConvissor the pushes go into the index for all branches / refs pushed. the checked out ref/branch is the one that gets the reflection out to the checked out file system. that's done mainly via "git reset --hard"22:19
avandenhoven: ^22:19
→ mxweas_ joined22:19
← stkrzysiak left22:21
← ph^ left22:21
← stkrzysiak_ left22:23
← Linc left22:23
← mxweas_ left22:24
← superdmp left22:24
← Anxi80 left22:24
← Eridius_ left22:25
→ mxweas_ joined22:25
→ mns joined22:26
← kogent left22:27
→ rizzuh joined22:27
rizzuh Created a private key, added public to github, added to pageant, I get "Permission denied (publickey)." when trying to clone. Why? SSH works for Codebase and Bitbucket O_o22:28
RandalSchwartz did you test it by ssh'ing in yet?22:29
→ Anxi80 joined22:29
← revagomes left22:29
RandalSchwartz and this is #git, not #github :)22:29
rizzuh ssh'ing in to here?22:29
RandalSchwartz, I know xD Trying on both channels.22:29
tomek ssh -T [email@hidden.address]22:29
RandalSchwartz step 5 of https://help.github.com/articles/generating-ssh-keys22:29
you apparently aren't following the instructions. :(22:29
← mxweas_ left22:30
rizzuh I don't have openssh though. I'll try through putty.22:30
frogonwheels rizzuh: or plink22:30
RandalSchwartz . o O (putty? ugh)22:30
tomek to use putty22:30
you need to convert your key to putty format22:30
rizzuh I do have it in that format.22:30
frogonwheels rizzuh: it's ok. I use putty for git. Mostly so it's easier for my team to manage :|22:31
→ mxweas_ joined22:31
rizzuh Yep, works to ssh directly.22:31
RandalSchwartz but... windows. :)22:31
frogonwheels RandalSchwartz: yeah. i hear ya22:31
rizzuh Can't get to clone a repo though. I know SSH is set up properly, I use it all the time with Mercurial.22:31
RandalSchwartz, no, don't go there.22:31
frogonwheels rizzuh: so what happens whenyou use plink ?22:32
← nicc left22:33
rizzuh frogonwheels, same as through putty gui, "Hi cmircea! You've successfully authenticated, but GitHub does not provide shell access."22:33
← starsinmypockets left22:33
rizzuh frogonwheels, so it accepts my key it seems.22:33
frogonwheels rizzuh: well that's a start.22:33
← TeckniX left22:33
frogonwheels rizzuh: are you using ssh to access your own repo?22:33
rizzuh yeah22:34
frogonwheels have you got git setup properly to use putty?22:34
← kerrick left22:34
frogonwheels msysgit presumably?22:34
rizzuh frogonwheels, on mysys installed I left the default, to use plink/pageat.22:34
← egarcia left22:34
rizzuh Yup22:34
← Dave^| left22:34
DanielConvissor rizzuh: what's your remote url look like? here's one i use for pushing to git via ssh: [email@hidden.address]22:35
→ Eridius_ joined22:35
frogonwheels rizzuh: you actually saw that option?22:35
rizzuh [email@hidden.address]22:35
that's mine22:35
frogonwheels, yeah.22:35
← mxweas_ left22:35
frogonwheels rizzuh: echo IT_SSH% (or ${GIT_SSH} if you are in bash )22:36
that was meant to be % GIT_SSH % without the gaps22:36
→ mxweas_ joined22:36
← d0k left22:37
rizzuh frogonwheels, doesn't seems to exist, well cmd says %GIT_SSH% on the output22:37
→ neurodrone joined22:37
→ kerrick joined22:37
frogonwheels rizzuh: ok. then it's setup in the bash file22:37
DanielConvissor rizzuh: tomek may be on to something. putty and openssh use different key formats. fix it and submit it to github again.22:37
frogonwheels .bat file22:37
rizzuh according to bash, nothing.22:37
frogonwheels DanielConvissor: nah.. plink is validating22:37
rizzuh DanielConvissor, I gave github the ssh2 format, don't worry.22:38
frogonwheels rizzuh: have a look in the git.bat file then22:38
DanielConvissor col22:38
rizzuh frogonwheels, where would I find it?22:38
← davidjdagino left22:38
frogonwheels in your path.. um22:38
DanielConvissor gotta go. good luck.22:38
→ philips joined22:38
← DanielConvissor left22:38
tomek thats why i use cygwin & openssh + git from cygwin22:38
← dorkmafia left22:38
frogonwheels rizzuh: C:\Program Files\Git\cmd\git.cmd22:38
rizzuh alright, what about it?22:39
← kevwil left22:39
← fly9 left22:39
→ cagedwisdom joined22:39
frogonwheels rizzuh: check to see if it sets GIT_SSH22:39
← Anxi80 left22:39
rizzuh no it doesn't22:39
→ Anxi80 joined22:40
frogonwheels rizzuh: hmm.. you know, why don't you just try setting it...22:40
rizzuh To... uh... what?22:40
← gusnan left22:40
← teweWORK left22:40
frogonwheels rizzuh: set GIT_SSH=c:\path_to\plink.exe22:40
← mxweas_ left22:40
TML something like 'C:\Program Files\TortoiseSVN\bin\TortoisePlink.exe'22:41
← marhaban left22:42
→ mxweas_ joined22:42
TML (That just happened to be the only copy of Plink on my only Windows VM - your path may vary)22:42
→ revagomes joined22:42
rizzuh frogonwheels, thanks, set an env variable with the path to plink, works now!22:43
frogonwheels rizzuh: :)22:43
TML: I thin kthat might be the only good thing about tortoise - a plink that presents a gui 'do you want to accept this key' dialog.22:44
← revagomes left22:44
← kerrick left22:45
← mxweas_ left22:46
← thearthur left22:46
→ ipalaus joined22:47
← ben_alman left22:48
← boombatower left22:49
← Destos left22:49
→ boombatower joined22:50
→ mxweas_ joined22:51
frogonwheels rizzuh: by any chance did you not restart your cmd shell since installing msysgit?22:52
rizzuh frogonwheels, I never use cmd.22:52
frogonwheels rizzuh: ok. then bash ?22:52
rizzuh frogonwheels, no, I use PowerShell. I did restart it.22:52
→ nicc joined22:53
→ thearthur joined22:53
frogonwheels rizzuh: was GIT_SSH set in your env variables?22:53
rizzuh no, I set it manually22:53
frogonwheels rizzuh: I mean the system/user ones22:53
rizzuh no22:53
frogonwheels hm.. that's bad. oh well, fixed now.22:54
← boombatower left22:54
← theCole left22:54
← dfr|mac left22:55
← mxweas_ left22:56
← Fandekasp left22:56
← archdori left22:57
→ ben_alman joined22:57
← raatiniemi left22:57
→ mxweas_ joined22:58
→ Destos joined22:58
← Anxi80 left22:58
→ Anxi80 joined22:59
avandenhoven Does git "know" when a file is binary or text? and if so, how can I find out what Git knows?22:59
← teknopaul left23:00
→ dfr|mac joined23:00
← eijk_ left23:02
← tewecske left23:02
← mxweas_ left23:02
frogonwheels avandenhoven: yeah, git knows.. um23:03
← delinquentme left23:05
→ mxweas_ joined23:05
← toothr left23:05
avandenhoven frogonwheels: Ok, I'm building a post-recieve script to do a deploy over FTP to a crappy server I don't have ssh access to. I know how to get the files that were changed, i just need to know whether a file that I'm putting on the see is text or binary.23:05
frogonwheels avandenhoven: I'm no sure exactly how to find out sorry um.23:06
→ kerrick joined23:07
→ osmosis joined23:07
← shruggar1 left23:08
→ zackiv31 joined23:08
← zero1337 left23:08
thiago avandenhoven: upload everything as binary23:08
avandenhoven thiago: that works?23:08
thiago sure23:08
→ toothrot joined23:08
avandenhoven facepalm - Thanks23:09
← arietis left23:09
→ _huoxito joined23:09
← HyperNerdV2 left23:10
← dfr|mac left23:11
→ dfr|mac joined23:12
← dfr|mac left23:12
cbreak avandenhoven: text is just a special case23:14
(or in other words: text is binary)23:15
← intransit left23:15
→ SkAzZ joined23:16
← danslo__ left23:16
→ danslo joined23:16
→ stkrzysiak joined23:17
← stkrzysiak left23:17
← ajpiano left23:19
← mxweas_ left23:20
→ marhaban joined23:20
→ ajpiano joined23:20
← thiago left23:21
→ mxweas_ joined23:22
← tatsuya_o left23:24
← thornofmight_ left23:25
← marhaban left23:25
→ cjz joined23:25
cjz hmmm, my .gitolite.rc definitely has UMASK = 0007,23:25
but gitlab seems to not see that23:25
← neurodrone left23:26
← toothrot left23:27
cjz gitolite doesnt run per se, there is nothing to restart23:27
right?23:27
← etcetera left23:29
→ toothr joined23:29
← Zenmaker_ left23:29
→ schnatta joined23:29
← techhelp left23:30
→ Zenmaker_ joined23:30
→ \toothrot joined23:31
→ beneggett joined23:33
→ etcetera joined23:33
arosen if you change some files and do git add file ; How do you see what you have changed before you do a git commit?23:33
← toothr left23:34
milki diff --cached23:34
arosen: ^23:35
arosen thanks23:35
← rizzuh left23:36
→ imachuchu joined23:37
← MrMagic left23:40
→ MrMagic joined23:41
→ jjbohn joined23:41
→ tazle_ joined23:41
→ bijan0770 joined23:41
→ bfig joined23:42
bfig hello. i was having conflicts and i wanted to redownload a whole folder on my repo. how can i download a specific folder with all it's contents?23:42
(since the new pull only brought a fraction of the files i wanted to get)23:42
← codewrangler left23:43
← cjz left23:44
→ PatF joined23:45
← kerrick left23:46
imachuchu bfig: "a whole folder", has the whole folder been checked into the repo (i.e. in the past were all of the files in the folder added then commited to some commit)?23:46
bfig imachuchu, yes23:46
← mxweas_ left23:46
bfig i have this: repo/ -> .git,other,folders23:47
i erased the contents of "folders"23:47
→ mxweas_ joined23:47
bfig i pulled and it only brought those who were new, but didn't redownload the old ones23:47
i have other changes in the "other" folder, so i don't want to reset --hard (that's as far as my git arsenal goes)23:48
→ erichynds joined23:48
← Destos left23:48
PatF → Destos23:48
→ centipedefarmer joined23:48
← bijan0770 left23:49
imachuchu bfig: "git checkout HEAD folders", though you may want to create a copy of other elsewhere just in case23:50
← opdynamiks left23:50
bfig imachuchu, excellent! thanks23:50
imachuchu bfig: you actually don't need to download anything (git already knows about it), what you actually need to do is tell git to ignore the changes you've made (the deletion of folders)23:51
bfig ohh ok23:51
← enmand left23:52
← ipalaus left23:52
→ callen joined23:53
callen so, I had a branch "Experimental"23:53
← psoo left23:53
callen it diverged from master23:53
I want to reset it to equal the current HEAD of master without losing the "Experiment" history23:53
→ subrosa joined23:53
callen the git-reset appeared to work, but when I try to push it, I'm told I need to merge23:53
if I pull, the reset gets undone.23:54
ideas?23:54
→ bolowni joined23:54
← dcarper left23:54
subrosa Anyone know how to run a webhook from one level under public folder?23:54
imachuchu callen: "I want to reset it to equal the current HEAD of master", what? Are you asking to basically add a new commit on experimental with the current state of master?23:54
callen imachuchu: basically, yes.23:55
imachuchu: am I using the wrong tool or idiom to accomplish that?23:55
I know I don't want to rebase.23:55
imachuchu callen: no, it's just not something I've heard usually done. Quite often a new branch will be created off of master, or the old branch could be renamed/tagged instead, but there is nothing wrong with what you are trying to do (just a bit odd)23:57
callen imachuchu: there are good reasons for it.23:57
imachuchu callen: as for how, I'm not sure. give me a sec23:57
callen: what reasons?23:57
← dgasson left23:57
callen experimental is a canonical branch that is tracked by a server, but we want to ditch the experiment we tried recently without losing the full history.23:57
I thought it would be a matter of git checkout experimental; git reset --hard origin/master23:57
SethRobertson !fixup23:57
gitinfo So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!23:57
callen which seems to work at first, but then I can't push it without it forcing me to pull and merge which ditches the reset.23:58
SethRobertson You want to delete the branch? git branch -D experimental23:58
callen no!23:58
imachuchu SethRobertson: actually not useful in this situation.23:58
callen we want to keep the full history.23:58
just change it to equalling the current HEAD of master.23:58
← ISF left23:58
kevlarman callen: git revert then23:58
callen: if it's a simple list of commits23:59
callen er, no.23:59
no.23:59
we don't want to revert the commits23:59
we want to keep the history23:59
SethRobertson A better approach would be a new commit from master to experimental making experimental a directory. The best approach would be to use a new branch23:59
kevlarman callen: revert keeps the history23:59
imachuchu callen: first off branches can be thought of as throw away. If you have something you want to keep in archive why not tag it? Second, while not a git command, a checkout of master, copy that to somewhere else, checkout experimental, copy back, git add -u will do it (but messy)23:59
kevlarman callen: git revert generates inverse commits of whatever you tell it to23:59
callen I don't want that.23:59

Logs Search ←Prev date Next date→ Channels Documentation