Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-07-27

| Channels | #git index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 Lomex joined #git
00:01 romerocesar joined #git
00:02 neanderslob joined #git
00:06 zacts joined #git
00:08 MarconM joined #git
00:12 fahadash joined #git
00:15 Yarne joined #git
00:16 l4v2 joined #git
00:17 autoboxer joined #git
00:17 tesuji_ joined #git
00:18 SegFaultAX joined #git
00:19 neeca joined #git
00:21 elfuego joined #git
00:22 LeBlaaanc joined #git
00:23 meLon joined #git
00:25 mwhooker joined #git
00:26 cyphase joined #git
00:26 PioneerAxon joined #git
00:26 elfuego joined #git
00:29 mischat joined #git
00:33 IWishIKnew joined #git
00:38 hahuang61 joined #git
00:39 moei joined #git
00:40 f3r70rr35f joined #git
00:42 NeXTSUN joined #git
00:42 Literphor joined #git
00:44 Electrometro joined #git
00:50 dopesong joined #git
00:53 jaguarmagenta joined #git
00:54 lucasem joined #git
00:55 chigang_ joined #git
00:56 Rodya_ joined #git
00:59 Rodya_ joined #git
01:00 kn-928 joined #git
01:01 thiago joined #git
01:05 leehambley joined #git
01:10 mwhooker joined #git
01:11 dtrott joined #git
01:12 afusion joined #git
01:15 dmto joined #git
01:16 adino joined #git
01:17 afusion_ joined #git
01:18 mingrammer joined #git
01:20 kn-928 joined #git
01:21 mingrammer joined #git
01:21 safe joined #git
01:22 Lyqyd joined #git
01:22 LeBlaaanc joined #git
01:30 fuchstronaut joined #git
01:33 Sucks joined #git
01:34 ochorocho__ joined #git
01:36 sangy joined #git
01:38 Sucks joined #git
01:39 d^sh joined #git
01:41 durham_ joined #git
01:46 fairuz joined #git
01:46 ome joined #git
01:46 cdg joined #git
01:46 tabrez joined #git
01:46 Rodya_ joined #git
01:48 elfuego joined #git
01:52 fairuz joined #git
01:52 drbean_ joined #git
01:53 pks joined #git
01:53 nilg joined #git
01:54 Dougie187 joined #git
01:55 jeffreylevesque joined #git
01:55 Limes_ joined #git
01:55 Literphor joined #git
01:57 jaguarma_ joined #git
01:59 bunniez left #git
02:00 Rodya_ joined #git
02:01 romerocesar joined #git
02:02 lagothri1 joined #git
02:03 nettoweb joined #git
02:04 lindenle joined #git
02:05 justanotheruser joined #git
02:06 inflames joined #git
02:07 Mannn joined #git
02:08 lucasem is there no issue-tracking system for git? (just the email list?)
02:08 drbean joined #git
02:13 LeBlaaanc joined #git
02:14 Rodya_ joined #git
02:14 afusion joined #git
02:14 titanium17 joined #git
02:15 thiago lucasem: there are several
02:15 thiago in fact, most issue tracking systems work with Git these days
02:15 lucasem thiago: I mean for the git project itself, though
02:17 zrts alright so here is the situation: i've some modifications that I left on my work root but now i forgot to push them and @ home i dont have access to it. question: is it possible to make some sort of global project mapping ? or should I simply move the whole project to a removable media?
02:18 mingrammer joined #git
02:18 zrts % git diff > foo.patch
02:19 zrts cool
02:20 rubie joined #git
02:21 xameada joined #git
02:22 stantonk joined #git
02:23 cqi joined #git
02:24 ajf- joined #git
02:26 dundee joined #git
02:27 LeBlaaanc joined #git
02:28 BenderRodriguez joined #git
02:28 MattMaker joined #git
02:28 chachasmooth joined #git
02:31 skylite joined #git
02:33 shgysk8zer0 joined #git
02:35 ochorocho__ joined #git
02:35 Derperperd joined #git
02:39 MattMaker joined #git
02:40 Goplat joined #git
02:41 Mannn git help does not list a "sync" command, so in GUI git clients, what exactly is the sync doing?
02:42 Literphor joined #git
02:43 rinwa_ joined #git
02:45 Mannn I think its just git pull
02:45 Derperperd joined #git
02:47 crazyhorse18 joined #git
02:49 sagerdearia joined #git
02:53 red_ joined #git
02:55 fairuz joined #git
02:59 cagmz joined #git
02:59 hexagoxel joined #git
03:00 rinwa_ joined #git
03:01 b_rare joined #git
03:09 phroa Mannn: probably followed by a git push
03:11 Dougie187 left #git
03:11 texinwien_ joined #git
03:11 leehambley joined #git
03:12 codybear joined #git
03:13 mingrammer joined #git
03:17 mingrammer joined #git
03:19 dtrott left #git
03:20 johnmilton joined #git
03:20 dreiss joined #git
03:20 red__ joined #git
03:20 k3rn31 joined #git
03:21 dstarh joined #git
03:22 RedneckDonut joined #git
03:23 steverandy joined #git
03:26 steverandy left #git
03:30 fuchstronaut joined #git
03:30 danimal2 joined #git
03:33 boombatower joined #git
03:33 blackwind_123 joined #git
03:38 BSAlb joined #git
03:42 Twirl joined #git
03:44 jstimm joined #git
03:46 fstd_ joined #git
03:48 Darren_ joined #git
03:49 mehola joined #git
03:55 danslo joined #git
03:58 elementalest joined #git
03:59 siaw23 joined #git
04:00 siaw23 hello. how do i get rid of this https://gist.github.com/siaw23/ad5bd3316​f4703d8b5e8f4792e0b9db9#file-git-txt-L2 i started getting it after changing head to an earlier commit
04:01 snappy joined #git
04:02 romerocesar joined #git
04:03 lucasem siaw23: that's expected. What are you trying to do/what state are you trying to be in?
04:03 jaguarmagenta joined #git
04:03 siaw23 fixed. had to checkout
04:03 siaw23 lucasem: thanks :)
04:05 snappy is it possible to set a specific remote ref to a certain commit _without updating the remote_?
04:06 snappy basically i want to undo a "git fetch" and then apply it again to test a few scenarios
04:07 rubie joined #git
04:09 drbean_ joined #git
04:10 danimal2 joined #git
04:10 danimal2 joined #git
04:11 danimal2 joined #git
04:12 phanimahesh joined #git
04:12 denimsoft joined #git
04:15 k-man joined #git
04:17 k-man joined #git
04:17 case39 joined #git
04:18 phanimahesh joined #git
04:18 k-man joined #git
04:22 stantonk joined #git
04:24 kadoban snappy: For testing you can just edit the file in .git/refs/remotes/ I guess. Make a backup first just in case.
04:25 kadoban There's probably some plumbing command to do it, but I can't remember which.
04:25 barajasfab joined #git
04:29 kadoban snappy: 'git update-ref' might be the plumbing command, looks right at least. It's still not friendly though, so still make a backup.
04:36 ochorocho__ joined #git
04:36 tgunr joined #git
04:38 anu0 joined #git
04:39 Faylite joined #git
04:39 dreiss joined #git
04:40 stantonk joined #git
04:41 nnyk joined #git
04:44 duderonomy joined #git
04:45 Lvjasha joined #git
04:46 jeffreylevesque_ joined #git
04:46 danimal2 joined #git
04:48 Cabanossi joined #git
04:51 jaguarma_ joined #git
04:52 snappy kadoban: cheers
04:53 freimatz joined #git
04:58 jaafar joined #git
04:59 fahadash joined #git
04:59 Literphor joined #git
05:04 subhojit777 joined #git
05:05 rubie joined #git
05:07 TomyLobo joined #git
05:12 liuzhen joined #git
05:12 f3r70rr35f joined #git
05:14 BSab joined #git
05:16 A5101 joined #git
05:16 romerocesar joined #git
05:19 tango_ joined #git
05:21 adino joined #git
05:21 A5101_ joined #git
05:22 jaguarmagenta joined #git
05:26 wxg joined #git
05:30 _xor joined #git
05:31 fuchstronaut joined #git
05:31 clemf joined #git
05:32 narendraj9 joined #git
05:34 jost joined #git
05:35 danimal2 joined #git
05:35 cyan__ joined #git
05:35 red__ joined #git
05:36 danimal2 joined #git
05:36 dreamseeker joined #git
05:38 d^sh_ joined #git
05:40 emPi joined #git
05:40 ThomasLocke joined #git
05:40 ThomasLocke joined #git
05:41 MrWoohoo joined #git
05:41 clemf joined #git
05:42 zeroed joined #git
05:42 NeverDie joined #git
05:45 satya4ever joined #git
05:45 fairuz joined #git
05:47 Alienpruts joined #git
05:48 albel727 joined #git
05:50 qt-x joined #git
05:51 armyriad joined #git
05:52 danimal2 joined #git
05:52 drbean joined #git
05:52 Macaveli joined #git
05:53 Captain_Ramius joined #git
05:53 moneylotion left #git
05:54 Captain_Ramius OK, easy question (for those of you who already know git): how do i revert only the last commit? (note that the "git log" is blasting my window and I can't get the commit id)
05:54 grawity revert as in "undo via yet another commit", or revert as in "get rid of, completely"
05:54 grawity also what's wrong with `git log`?
05:54 Captain_Ramius get rid of: I added my vendor folder and then realized I didn't want to
05:54 doebi joined #git
05:55 mariuscc joined #git
05:55 Captain_Ramius git log just flooded (figuratively) my cmd window and I can't scroll back far enough to read the commit id ("hash"?)
05:55 grawity a) reset to HEAD^, i.e. one commit before HEAD
05:55 danslo joined #git
05:56 grawity b) rm -rf --cached vendor/ && commit --amend, to re-do the same commit but without the vendor folder
05:56 grawity you were asking for a), I guess
05:56 grawity but b) is probably the end result you were looking for
05:57 Captain_Ramius I'm not sure. I'm a little fuzzy this late and not very skilled with git when I'm not fuzzy.
05:57 adino joined #git
05:57 Captain_Ramius Disclaimer: I'm not quite sure what the HEAD is actually.
05:57 Captain_Ramius *Disclosure
05:58 grawity in short, an alias for the commit you're on
05:58 ochorocho__ joined #git
05:58 Captain_Ramius that makes sense, thanks!
05:58 Captain_Ramius I'm trying to learn everything at once it seems.
05:59 Captain_Ramius When I did the commit --amend, it worked great but why does it insist on showing me ALL the code in the commit?
05:59 Captain_Ramius The "flood" I mentioned earlier was what-looked-like every bit of code from the vendor/ folder!
05:59 grawity I dunno, I can't exactly see what you're seeing
06:00 Captain_Ramius *Correction: when I tried "git log"
06:00 grawity by 'code' do you mean a list of filenames, or do you literally mean the contents of each and every file?
06:00 Captain_Ramius oh, definitely contents.
06:00 Esya You can use "-n" with git log to specify the number of commits you want to see
06:00 grawity well then, again, I can't see what you're seeing on your screen, and this is not what Git normally does
06:00 Esya Also git log has options to see or not see the files impacted, and even the content
06:00 grawity so I'm going to answer with "it doesn't."
06:01 grawity unless you're actually typing `git log -p` and not just `git log`
06:01 Esya Captain_Ramius: A plain git log output should look like this : http://2.bp.blogspot.com/--Oj7Ocw5dd​A/UQavWR8_1oI/AAAAAAAAI-M/g1hbsrRRbJ​4/s1600/git-log-default-output.png
06:01 Repox joined #git
06:01 Captain_Ramius Oh, because a search result said to use "git log -p -2" to show 2 results.
06:02 * grawity facepalms.
06:02 Captain_Ramius I didn't realize the -p wasn't part of that.
06:02 Captain_Ramius My teachers always did that a lot too. You're not alone.
06:03 Captain_Ramius Whoa, time to be done. Enough damage for one night.
06:03 armyriad joined #git
06:03 Captain_Ramius Thanks for helping me roll that back.
06:03 jknetl joined #git
06:03 grawity so yes, -2 (shortcut for -n 2) means two commits max, and -p means "show the commit diffs, in addition to regular log"
06:03 grawity `git log --stat` might be a better middle ground
06:04 afusion joined #git
06:06 Captain_Ramius Good deal.
06:07 rubie joined #git
06:07 byteflame joined #git
06:08 ThomasLocke joined #git
06:08 ThomasLocke joined #git
06:08 mischat joined #git
06:09 byteflame joined #git
06:10 zeroed joined #git
06:10 JeroenT joined #git
06:10 Captain_Ramius To remove that last commit completely, do I reset hard or soft?
06:11 grawity doesn't matter
06:11 byteflame joined #git
06:11 grawity well, --hard would throw away the vendor/ directory in reality, so maybe stick with the default
06:12 Captain_Ramius Oh man. I just went through all the trouble of GETTING the vendor/ directory.
06:12 vuoto joined #git
06:13 Captain_Ramius When your volunteer PHP tutors require you to use composer before they will answer your original question.
06:13 dopesong joined #git
06:13 Captain_Ramius Uhhhh...I typed "git reset HEAD^" and it says "More?"
06:13 yitz YHeh
06:14 Captain_Ramius It's a prompt evidently
06:14 grawity ah, cmd.exe
06:14 grawity well, git reset HEAD~1 then
06:14 grawity same thing but avoids the ^
06:14 Captain_Ramius some syntax issue in cmd?
06:16 byteflame joined #git
06:16 Captain_Ramius Geez, what a mess. Thanks again.
06:17 armyriad joined #git
06:17 Hi-Angel joined #git
06:18 Ocnod joined #git
06:18 eijk joined #git
06:18 [cust] joined #git
06:18 ThomasLocke joined #git
06:20 Literphor joined #git
06:25 Xack left #git
06:28 WayToDoor joined #git
06:32 realz_ joined #git
06:34 star_prone joined #git
06:35 djb-irc joined #git
06:35 zeroed joined #git
06:35 jceb joined #git
06:41 subhojit777 joined #git
06:41 armyriad joined #git
06:41 stantonk joined #git
06:42 Alienpruts joined #git
06:45 Sceorem joined #git
06:48 metze joined #git
06:49 cyphase joined #git
06:49 i7c One should probably mention that only for the case of 1 ~1 and ^1 are the same thing.
06:50 tvw joined #git
06:50 danslo joined #git
06:50 Faylite_ joined #git
06:51 dminca joined #git
06:51 etcinit joined #git
06:51 beckyconning_ joined #git
06:51 GameGamer43 joined #git
06:52 pfg joined #git
06:52 lucasem joined #git
06:52 cspotcode joined #git
06:52 janx joined #git
06:52 barq joined #git
06:52 wizonesolutions joined #git
06:53 iota_pi joined #git
06:53 SwiftMatt joined #git
06:53 lmatteis joined #git
06:54 cloudcell__ joined #git
06:56 jstein_ joined #git
06:56 dpower joined #git
06:57 Literphor joined #git
07:00 dimi1947 joined #git
07:00 LiohAu joined #git
07:00 jaguarmagenta joined #git
07:01 MattMaker joined #git
07:08 nilg joined #git
07:08 lb i just love xkcd http://m.xkcd.com/1654/
07:11 Codfection joined #git
07:14 MattMaker joined #git
07:16 p4trix joined #git
07:16 JeroenT joined #git
07:17 Lomex joined #git
07:20 Silenced_v2 joined #git
07:23 star_prone joined #git
07:23 elijah joined #git
07:24 ffledgling joined #git
07:24 thiago lb: it's scary
07:25 lb thiago but effective ^^
07:25 ffledgling Hi, I'm trying to transplant a subdirectory from one repository to another while trying to maintain all history. What's the best way to do this?
07:26 ffledgling I've done this in the past with simpler cases where I created a list of patches from history and applied them in the other repo, but this doesn't work as well on files that have been moved around.
07:27 MrMojito joined #git
07:27 MrMojito joined #git
07:27 ffledgling I'm looking at subtree merges, but they seem to move the entire project/repo to the new repo as a subdirectory instead of just a particular subdirectory/folder... anyone have any suggestions on how to do this right?
07:27 Balliad joined #git
07:27 canton7 ffledgling, you can extract the subfolder with filter-branch --subdirectory-filter. To integrate it into the new repo, I'd probably go with a subtree merge (https://github.com/git/git/blob/mast​er/contrib/subtree/git-subtree.txt)
07:27 moriko joined #git
07:28 lb canton7 ahh thanks i just typed a lot dirty hack text, but forgot that subtree merge exists
07:28 canton7 might be worth posting anyway, for interest's sake
07:28 Softeisbieger joined #git
07:29 lindenle joined #git
07:29 lb i've done it once, i created a new repo with filter-branch that only containted the history of this subdir, then added that as a remote to the "second" repo and merged it in. resulted in a history that had 2 root commits but worked... not the best option i believe
07:30 ffledgling canton7: so something like this -- http://stackoverflow.com/q/1365541/1220089 ?
07:31 Rish joined #git
07:31 canton7 he's pulling out a subdirectory and moving it into the root of another repo, I think?
07:31 canton7 whereas you're pulling out a subdirectory and transplanting it into a subdirectory of another repo?
07:32 fuchstronaut joined #git
07:32 WayToDoor joined #git
07:32 reactormonk canton7, looking for git subtree?
07:32 ffledgling canton7: correct, but the pulling out the subdirectory part is the same, and then I can replace it with a subtree merge possibly?
07:32 ffledgling I need to check how much history the filter-branch will save
07:33 canton7 reactormonk, read the backlog
07:33 canton7 ffledgling, yeah, sounds about right
07:34 RustyShackleford joined #git
07:34 ffledgling canton7: problem is, if the files in that directory have been moved around before... /that/ history might now show up (git log /path/to/subdir, for example doesn't show it)
07:34 _xor joined #git
07:34 canton7 ffledgling, (man git filter-branch has examples too)
07:34 gitinfo ffledgling: the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html
07:35 ffledgling I was reading side-by-side :)
07:35 jeffreylevesque joined #git
07:35 canton7 no, it won't show up
07:35 danslo joined #git
07:35 duderonomy joined #git
07:35 ffledgling Yeah, then it doesn't have any real advantage over the patch list approach I did earlier... any idea how I get around that?
07:35 canton7 and how could that work with your scenario? you presumably only want to add commits to the new repo which concern a particular subdirectory - you don't want commits which edit files in other directories, potentially overwriting pre-eisting files
07:35 monoprotic joined #git
07:36 jaafar joined #git
07:36 Captain_Ramius joined #git
07:36 ffledgling That's not technically what I want to do. I want to move all the files currently in a particular subdirectory into a new repository, while maintaining the ability to git-blame correctly
07:37 canton7 the only thing I can think of would be to edit the history of the old repo, before doing a --subdirectory-filter, to make sure that all edits to the files you're interested in happen in the subdirectory you're going to extract
07:37 canton7 ffledgling, yes, but in order to blame correctly you need the history, and the history is a series of commits, and commits change files
07:37 lmatteis joined #git
07:37 ffledgling right
07:37 canton7 anyway sorry, got to commute. good luck!
07:38 ffledgling thanks
07:38 tgunr joined #git
07:38 kanzure joined #git
07:39 leprechau joined #git
07:39 icedp joined #git
07:39 romerocesar joined #git
07:39 j416 joined #git
07:39 tmz joined #git
07:39 Cyclohexane joined #git
07:39 Silenced_v2 joined #git
07:40 DolphinDream joined #git
07:40 h12o joined #git
07:40 tobiasBora joined #git
07:40 spriz joined #git
07:40 larrybolt joined #git
07:41 ahmedelgabri joined #git
07:41 __main__ joined #git
07:42 Pawky joined #git
07:42 M-meznak joined #git
07:42 JeroenT joined #git
07:43 rbr joined #git
07:44 stantonk joined #git
07:45 Pawky how can I fetch files from another branch into a current where, in the other branch some files in current are deleted (such as symlinks)?
07:46 selckin git checkout branch -- file
07:46 Pawky but I want all files :-)
07:46 ninsei joined #git
07:46 selckin git checkout branch -- .
07:46 selckin (maybe)
07:46 Pawky testing...
07:46 fairuz joined #git
07:48 netj joined #git
07:49 Pawky selckin: nope.. that will remove all files in current...
07:49 i7c What?
07:50 i7c It shouldn’t remove any files.
07:50 aidalgol joined #git
07:50 cyphase joined #git
07:51 Pawky like  i have 10 files in master, i branch  to brtest. Then i switch back to master, deletes 5 files.. then adds another 3, now i want br  test to get these files.
07:51 Pawky if I do as above, brtest will end up only having 5 files....
07:51 selckin that sounds completely differend to your original explaintation or i'm too sleepy
07:51 Pawky thats probably true..
07:51 qpdb joined #git
07:52 i7c You switch back to master, delete 5 files and commit the deletion? Or not?
07:52 texinwien_ joined #git
07:52 Pawky my idea is as:... having my files in master, then having several branches with  .ini files  depending on environment.
07:53 Pawky thus master will have like 5 files... and the other ones a bunch more.. then I maybe add stuff in master and want it in the other branches as well..
07:53 Pawky weird ha?...
07:53 i7c Sounds like you want to maintain a patch set for your configurations AKA rebase the environment specific branches onto new commits on master.
07:54 selckin !xy
07:54 gitinfo Woah, slow down for a bit. Are you sure that you need to jump through that particular hoop to achieve your goal? We suspect you don't, so why don't you back up a bit and tell us about the overall objective...
07:54 selckin wish that actually linked to xy
07:55 kedare joined #git
07:55 Pawky ok... my objective is... having  all code in master, then have branches for  like, prod, test, staging having unique .ini files the code is dependent upon.
07:56 selckin like i7c said, you want to keep rebasing those branches on master
07:56 mikecmpbll joined #git
07:56 selckin or just have a seperate repo with config
07:56 selckin !deploy
07:56 Brain joined #git
07:56 gitinfo Git is not a deployment tool, but you can build one around it (in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/deploy.html
07:56 clmsy joined #git
07:57 Pawky well... funny as it is... what I am deploing into the different environments are rex deployment scripts... :-D
07:57 daynaskully joined #git
07:57 cbreak Pawky: don't call it master
07:57 ffledgling Why not just have prod.ini, test.ini and staging.ini in the same branch but as different files?
07:58 Pawky because the the environments should be clones...
07:58 kadoban Or ... not have any of those tracked, since it sounds like config to me.
07:58 Pawky so, test.ini must be the same in all environmnets, but with different data..
07:58 ffledgling kadoban: why would you not want your configs version controlled?
07:58 Pawky but I want to track config
07:59 dopesong joined #git
08:00 kadoban ffledgling: Because most meanings for config work *terribly* with how git works. !config
08:00 gitinfo ffledgling: [!configfiles] It is recommended to store local configuration data in a file which is not tracked by git, but certain deployment scenarios(such as Heroku) may require otherwise. See https://gist.github.com/1423106 for some ideas
08:00 Pawky however weird my ideas sound, i am actually trying to clean things
08:00 grohne left #git
08:00 Pawky gitinfo: and this is so in my case...
08:01 Pawky but I am a newbie at git itself, thus my whole idea might be bad.. though i thought it might be a good one :-)
08:01 ffledgling kadoban: I don't quite get why configs work poorly with git
08:02 ffledgling Unless you put deployment secrets/passwords in them ofcourse, but that's a different case
08:02 Pawky scripts in master, config in different branches... then fetch correct branch for each environment,.
08:02 Pawky in my case the ini files describe the local environments.. like.. web server is abc, etc..
08:03 kadoban ffledgling: Because as soon as two different people want to have their own separate version, that doesn't work at all. You either have a file sitting around you can't ever commit without screwing over the other person, but can't be ignored because it's tracked, or you end up with goofy hacks that don't work.
08:03 Pawky but when playing around with git, i have noticed.. that if files are deleted in.... master (its what its called), after branching,  whatever type of sync I do will delete the files in branch
08:04 ffledgling kadoban: ah, I see what you mean. Those weren't the kind of configs I was thinking of, more of configs that control deployments on a server, which will not/should not be independent for different people
08:04 Pawky (in my case, i am the only one using git for this, as well as manipulating the scripts)
08:05 rinwa_ joined #git
08:06 Pawky so... my ideas are all stupid?
08:06 jaeckel joined #git
08:07 ffledgling Pawky: you probably just want to create a fresh repo to store your configs or store them in some kind of object store and then pull them in via your scripts from there
08:07 tmz joined #git
08:08 jaafar joined #git
08:08 grawity hm, if you remove a file from master, then separately add it to a branch
08:08 Pawky ffledgling: maybe...  but is there a way to fetch all current files from one branch into another?
08:08 grawity further merges shouldn't remove it again
08:09 drbean_ joined #git
08:09 Pawky grawity: its this behaviour i havent really figured out yet.
08:09 Slush- joined #git
08:10 Pawky so master <- add 10 files; branch to brtest; delete 5 in master; add another 3; switch to brtest, now trying to sync master stuff into brtest results in brtest loosing the 5  deleted files
08:11 docnvk joined #git
08:11 kanzure joined #git
08:11 chll_ joined #git
08:11 marcogmonteiro joined #git
08:12 crayon joined #git
08:12 PCatinean joined #git
08:12 spriz joined #git
08:12 nellicus_ joined #git
08:12 Pawky i want whatever changes and new files from mater -> brtest, but not sync the removal of the 5 files.
08:12 Kobaz joined #git
08:12 Pawky (master)
08:12 dcpc007 joined #git
08:13 tomog999_ joined #git
08:13 steveccc joined #git
08:14 JeroenT joined #git
08:15 steveccc are there any web developers (just html, css, javascript, jquery, php) who wouldnt mind chatting with me for a short while about your workflow - i would like some help with how i can split up work in a team of three so we arent tripping over each other
08:22 vassagus joined #git
08:25 t0by joined #git
08:26 p4trick joined #git
08:26 fuchstronaut joined #git
08:28 MrMojito joined #git
08:28 BryzNSTY joined #git
08:29 Slush- joined #git
08:29 bongjovi joined #git
08:35 ffledgling What's the expected behaviour when using git log with both --follow and --reverse?
08:35 kurkale6ka joined #git
08:38 _ikke_ Good question
08:38 _ikke_ Not sure if the follow semantics would be reversed too
08:39 ffledgling It seems like they are
08:39 ffledgling I actually don't know what it's doing, it shows me all commits without --follow and just one (empty? or rename) commit with --reverse
08:40 _ikke_ The man page implies that it would only affect the order the commits are output
08:40 ffledgling I thought so too, but that doesn't seem to be the case
08:40 drbean joined #git
08:41 sagerdearia joined #git
08:42 ramsub07 joined #git
08:44 fairuz joined #git
08:48 subhojit777 joined #git
08:57 ahmedelgabri joined #git
08:59 skylite joined #git
08:59 mischat joined #git
09:00 romerocesar joined #git
09:02 MattMaker joined #git
09:07 drbean_ joined #git
09:08 nnyk joined #git
09:14 ahmedelgabri joined #git
09:17 lindenle joined #git
09:17 jaguarmagenta joined #git
09:19 ryez joined #git
09:19 Derperperd joined #git
09:22 boombatower joined #git
09:25 Derperperd joined #git
09:25 cdown joined #git
09:26 star_prone joined #git
09:27 fuchstronaut joined #git
09:28 lohfu joined #git
09:28 neeca joined #git
09:31 mahendra_ joined #git
09:34 dsantiago joined #git
09:35 dopesong joined #git
09:37 JJjack joined #git
09:37 JJjack hi
09:37 JJjack I am following these intsructions:
09:38 JJjack http://www.bartread.com/2014/02/12/gi​t-basics-how-to-merge-changes-from-a-​different-fork-into-your-own-branch/
09:38 JJjack I got stuck at the git merge <remote_name>/<remote_branch_name>
09:38 JJjack i get error: The following untracked working tree files would be overwritten by merge:
09:38 JJjack readme.md
09:38 JJjack Please move or remove them before you can merge.
09:38 _ikke_ JJjack: what does git status say?
09:39 JJjack how do I access the fetched repo?
09:39 ryez JJjack: your workspace isn't clean, stash or commit first
09:39 JJjack ok
09:39 dopesong joined #git
09:39 danslo joined #git
09:41 JJjack nothing to commit, working directory clean
09:41 JJjack still same errror
09:41 MattMaker joined #git
09:41 _ikke_ JJjack: is readme.md ignored (in a .gitignore file)?
09:42 Angs joined #git
09:42 JJjack no
09:46 Derperperd joined #git
09:47 Derperperd joined #git
09:47 dsdeiz joined #git
09:47 Synthead joined #git
09:48 fuchstronaut joined #git
09:48 JJjack git rm readme.md
09:48 JJjack fatal: pathspec 'readme.md' did not match any files
09:48 JJjack ok stuck
09:51 osse JJjack: it's untracked. use regular rm
09:52 Pursuit joined #git
09:52 JJjack I don't even see this file
09:52 daumie joined #git
09:53 daumie joined #git
09:54 osse maybe you already deleted it :p
09:54 JJjack so why isn't it merging?
09:54 MrWoohoo joined #git
09:55 heftig JJjack: rm readme.md
09:57 JJjack heftig thanks that did it
09:57 JJjack now I have conflicts
09:59 afiskon joined #git
09:59 cdown_ joined #git
10:00 RedShift joined #git
10:02 daumie I fucked up guys... I 'git rm'  and 'rm -r' a folder then pushed changes...this was due to a duplicate folder which i had earlier 'rm -r'
10:02 daumie Can i recover the folder?
10:04 heftig daumie: look through git reflog
10:06 dopesong joined #git
10:06 sdothum joined #git
10:07 CussBot joined #git
10:08 drbean joined #git
10:09 weckl joined #git
10:09 MattMaker joined #git
10:10 King_Hual joined #git
10:10 kashyap joined #git
10:10 mingrammer joined #git
10:10 reactormonk left #git
10:10 pluszak joined #git
10:11 pluszak I've moved some code into a different function in the same file. Is there a switch for git diff that would suggest that happened? I mean, I want to know what I did other than moving the code
10:12 pluszak And currently there are too many lines to see if I didn't mess something up by accident
10:14 kashyap How to list all commits for a particular Git tag?  Something like `git log --oneline --tag=1.4.0`  My look-up-foo seems not sufficient
10:16 nikitasius joined #git
10:16 piling joined #git
10:17 kashyap Oh, this is it: $ `git log --simplify-by-decoration 1.4.0 --oneline`
10:18 MattMaker joined #git
10:18 kashyap Hmm, not quite...The above shows the commits at which various tags were created
10:20 f5 joined #git
10:20 elfuego joined #git
10:20 yoan_ joined #git
10:22 snappy left #git
10:23 vassagus joined #git
10:25 fish1 joined #git
10:26 fish1 joined #git
10:27 fish8 joined #git
10:29 fish86546 joined #git
10:29 daumie joined #git
10:29 fish86546 joined #git
10:30 dopesong joined #git
10:30 kashyap Okay, it is this (diff the commits by comparing with the previous tag) - `git log --oneline --no-merges 1.3.0..1.4.0`
10:31 fish86546 joined #git
10:38 spudowiar joined #git
10:41 ShapeShifter499 joined #git
10:42 harish_ joined #git
10:46 stantonk joined #git
10:47 JeroenT joined #git
10:48 fuchstronaut joined #git
10:48 netcarver joined #git
10:50 rahtgaz joined #git
10:52 rahtgaz hello everyone. git log --graph is showing me my mainline branch "master" to the right of the development branch (http://dpaste.com/2FB8GP7). How can I set it so that master is shown on the left?
10:52 rudi joined #git
10:54 MattMaker joined #git
10:55 dopesong joined #git
10:56 reactormonk joined #git
10:56 reactormonk How can I git grep through all the diffs in the current repo?
10:56 reactormonk Looking for a specific commit
10:56 _ikke_ git log -S
10:56 _ikke_ That looks for commit that either added or removed a specific text
10:57 daumie joined #git
10:58 reactormonk nice, thanks
10:58 texinwien_ joined #git
11:05 mingrammer joined #git
11:05 lindenle joined #git
11:05 Siilwyn joined #git
11:06 danslo joined #git
11:07 Siilwyn Hi people, I have quite an unusual task, well for me atleast. I'm moving my old repositories to a new account, I'd like to replace certain words in all of the history, like it was always written that way while keeping the original commit history. Is that possible?
11:08 _ikke_ look into man git filter-branch
11:08 gitinfo the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html
11:08 _ikke_ That can do it
11:08 Siilwyn looks very powerful
11:08 pothepanda joined #git
11:09 Siilwyn thank you _ikke_ going to try it out!
11:09 _ikke_ YW
11:13 WayToDoor joined #git
11:13 fuchstronaut joined #git
11:14 Lomex joined #git
11:16 dopesong joined #git
11:17 ojdo joined #git
11:18 reactormonk Can I git stash files that are not checked in?
11:19 dminca joined #git
11:19 selckin -u
11:19 reactormonk Thanks
11:20 MrMojito1 joined #git
11:20 dminca joined #git
11:21 reactormonk How can I set a branch to a specific commit?
11:21 reactormonk Via checkout or reset?
11:21 cyan__ joined #git
11:21 HoierM joined #git
11:22 _ikke_ reactormonk: either git checkout -B <name> <commit>, and/or if the branch is already checked out: git reset --hard <commit>
11:22 selckin reset --hard (will delte uncommited changes)
11:23 rnsanchez joined #git
11:24 pothepanda joined #git
11:26 yoan_ joined #git
11:30 kashyap left #git
11:35 fahadash joined #git
11:36 ikmaak joined #git
11:37 Grievar joined #git
11:38 hamid_ joined #git
11:40 fuchstronaut joined #git
11:40 Guest78273 hi there
11:40 gitinfo Guest78273: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.
11:40 rahtgaz hello everyone. git log --graph is showing me my mainline branch "master" to the right of the development branch (http://dpaste.com/2FB8GP7). How can I set it so that master is shown on the left?
11:43 ki0 joined #git
11:44 rahtgaz is this just by design? I notice most git gui applications, make a distinction between branches, and master is usually placed on the left with branches branching out to the right. gitg (git gnome) for instance, even allows me to change the mainline branch and alter the view
11:45 spudowiar1 joined #git
11:45 _ikke_ rahtgaz: The branch currently checked out it shown on the left
11:45 cdown joined #git
11:46 rahtgaz _ikke_: I thought of that before. but when I checkout master, I get the same view
11:47 rahtgaz This is what I get after checking out master http://dpaste.com/25Q47YJ
11:51 ajf- joined #git
11:52 _ikke_ I don't think git makes any promises how that graph is rendered
11:53 Sceorem joined #git
11:53 canton7 yeah, git doesn't see any branch as being more special than any other in that graph
11:54 rahtgaz Yeah, thanks for the confirmation you both. I do notice that sometimes it renders it differently. So it's just by design.
11:59 cYmen is it possible to have both ids and relative time stamps shown in the stash list?
12:02 fuchstronaut joined #git
12:02 jimi_ joined #git
12:05 stoopkid joined #git
12:05 kedare_ joined #git
12:06 ffledgling How do I get to output a patch that shows renames/moves?
12:06 ffledgling *get git to
12:06 moritz uhm, does the standard patch format even have a notation for renames?
12:07 _ikke_ I don't think so
12:07 ffledgling No, it show file1 > /dev/null and then dev/null > file2
12:07 ffledgling *shows
12:07 griffindy joined #git
12:07 ffledgling Unfortunately I want it to output a particular subset of moves/renames and output them
12:09 sangy joined #git
12:09 Lomex joined #git
12:10 st0opkid joined #git
12:10 eijk joined #git
12:10 synthroid joined #git
12:11 fuchstronaut joined #git
12:11 sto0pkid joined #git
12:13 neeca joined #git
12:13 jknetl joined #git
12:14 stoopkid joined #git
12:15 jost joined #git
12:15 kyan joined #git
12:16 jknetl joined #git
12:16 CheckDavid joined #git
12:19 bongjovi joined #git
12:20 Gamecubic joined #git
12:20 Ph_ joined #git
12:21 dimi1947 joined #git
12:21 rgrinberg joined #git
12:23 vassagus joined #git
12:24 anu1 joined #git
12:25 Harzilein joined #git
12:25 Harzilein hi
12:29 Harzilein i have the following scenario and would like to discuss extensions/porcelain to make this easy to implement: i have two upstreams, where one (something/master) sometimes gets rebased onto/merged onto origin/master. now occasionally i want the newest stuff from something/master and would like to merge it. when origin gets around to merge it as well, i'd like to get rid of my personally created bits of merge history
12:29 rwb joined #git
12:29 cbreak joined #git
12:30 eijk_ joined #git
12:30 Harzilein anything that can enhance my eexperience googling it for myself is also appreciated
12:32 crazyhorse18 joined #git
12:34 star_prone joined #git
12:34 psprint joined #git
12:35 psprint Hello. If I fetch from a repo, can I remove the downloaded objects if I will not decide to merge?
12:35 st0opkid joined #git
12:35 rolha joined #git
12:35 selckin probably have to remove the remote
12:36 moritz and then the garbage collector will take care of them
12:38 Pawky is there a way to checkout files from another branch only if they do not exist in current?
12:38 selckin what you need is a build
12:39 stantonk joined #git
12:39 Pawky selckin: a build?
12:40 moritz [ -e file ] || git checkout otherbranch file
12:40 Pawky hmm.. :-)
12:40 Harzilein hmm
12:40 Harzilein is my question unclear?
12:40 nilg joined #git
12:40 zeroed joined #git
12:40 zeroed joined #git
12:41 moritz Harzilein: not really, I just don't have a good idea
12:41 selckin Harzilein: i think you just want to rebase onto the other master, but can't parse it
12:41 mingrammer joined #git
12:42 moritz if there are siginficant differences between the merges, that'll become painful very quickly
12:42 eijk joined #git
12:43 osse Harzilein: i think you want (at least in the simple cases):  git rebase --onto=origin/master something/master
12:43 Harzilein okay, let's be concrete: pcbsd's idea of tracking their ports tree is using their github fork of the github mirror of the freebsd svn ports tree
12:43 Twirl joined #git
12:43 Harzilein sometimes they lag behind wrt recent ports
12:43 osse Harzilein: IOW: take what you have on your local master that is not on something/master and "transplant it" into origin/master
12:43 Harzilein for now i just copied them to my home directory
12:44 Harzilein now i want to make my own local branch that tracks the pcbsd branch and occasionally update single ports from their freebsd commits
12:44 Harzilein s/their freebsd/freebsd's commits/
12:45 Harzilein but when pcbsd gets around to updating those new ports i want to eliminate the history that diverges from pcbsd
12:45 freimatz joined #git
12:46 moritz Harzilein: and that's all in one big git repo?
12:46 Harzilein of course that could just be resetting my local branch to be like pcbsd's, but i might just like to add my own stuff. it's okay to then have merges, but i'd like to have the merges from freebsd resemble what is done at pcbsd's master
12:46 Harzilein yeeess... and its's slooooow :D
12:47 Harzilein ~150k files
12:47 moritz well, I guess the "one big repo" thing is the main problem
12:47 Harzilein and i only have a consumer harddrive
12:47 moritz sounds painful
12:47 moritz and yes, then you need to rebase on pcbsd's master
12:48 Harzilein meh, so local changes would be too complicated?
12:49 marble_visions joined #git
12:49 jeffreylevesque joined #git
12:49 anuxivm joined #git
12:49 moritz you mean uncommitted local changes?
12:49 marble_visions hi all, is there a tool that can alert you if a remote is trying to rewrite history?
12:50 marble_visions (by you, for example, issuing a pull/fetch)
12:50 Harzilein moritz: no, something like local ports, i keep them in my home directory as well atm, which is a bit of a nuisance when slave ports have relative directories
12:51 moritz Harzilein: no idea
12:51 _ikke_ marble_visions: You can change the default refspec to not do force update by default
12:51 vaibhavmule joined #git
12:51 citizen-stig joined #git
12:52 shinnya joined #git
12:52 cbreak joined #git
12:52 marble_visions _ikke_: i suppose this is documented here, right?: https://git-scm.com/book/en/​v2/Git-Internals-The-Refspec
12:52 marble_visions thanks for the suggestion as well!
12:52 Harzilein harzi@toybox:/usr/ports $ git reset --soft toybox
12:52 Harzilein harzi@toybox:/usr/ports $ git read-tree toybox
12:53 cyan__ joined #git
12:53 _ikke_ marble_visions: yes, remote the + and it will complain when something is not a fast-forward
12:53 _ikke_ s/remote/remove
12:53 lindenle joined #git
12:54 Harzilein harzi@toybox:/usr/ports $ git checkout-index -a -f
12:55 Harzilein now to cherry pick the port update. let's take git as an example (2.9.0 in pcbsd, 2.9.2 in freebsd)
12:55 marble_visions _ikke_: i need to read more about the refspec, but you imply that all history rewrites are non-fast-forwards. can I have non-fast-forwards that are not changes to the history?
12:56 stoopkid joined #git
12:56 _ikke_ marble_visions: No, you can't
12:56 _ikke_ If people just push (without --force), it will always be a fast-forward
12:58 segment-fault joined #git
13:01 ffledgling canton7: So I figured out how to do what I wanted re: transplanting a subdirectory that's been moved before
13:01 canton7 ok, cool. What did you decide on?
13:01 citizen-_ joined #git
13:02 ffledgling The problem seemed to be that git log <path/to/dir> was able to do a --follow because that path didn't exist before the commit that did the git mv
13:02 st0opkid joined #git
13:02 ffledgling So I checked out to {sha of git mv commit}^1
13:03 SourTrout joined #git
13:04 TooLmaN joined #git
13:04 ffledgling Created patches of all the history of the prior path. Created a patch for the move commit. checked out to HEAD, created patches from from the git mv commit till the HEAD. Modified the patch that did the git mv manually, trimmed the patch list for the aftermove patchset. Then applied them in the new repo in the following order: history.beforemove.patches, gitmv.modified.patch, history.aftermove.modified.patches
13:04 ffledgling If that makes sense :s
13:05 canton7 haha, nice
13:05 canton7 pretty manual, but should work weel
13:05 canton7 *well
13:05 afusion joined #git
13:06 ffledgling I feel like I could put it in a loop, since the modifications to all patches are sed commands I ran to modify paths, remove diffs of files that don't need to be moved, but yeah it's kind of manual
13:06 ffledgling Okay for a one-off I guess, but I still haven't found a better way
13:07 canton7 I reckon a filter-branch --index-filter before the -subdirectory-filter might have worked
13:07 canton7 so mv the file from where it was into the subdir before filtering the subdir
13:07 marble_visions _ikke_: thank you!
13:07 marble_visions left #git
13:08 SourTrout joined #git
13:08 nnyk joined #git
13:08 nnyk joined #git
13:09 ffledgling canton7: don't quite grok how --index-filter/--tree-filter work, might try that to simply to see what works better
13:09 reynhout left #git
13:09 Ph_ joined #git
13:10 Harzilein harzi@toybox:/usr/ports $ git cherry-pick a475a7cfa84a1de8fbd65965041f160449a47a97
13:10 vmiklos joined #git
13:10 fuchstronaut joined #git
13:11 jast ffledgling: basically tree-filter checks out all the files in a temporary dir and lets you do whatever on them, then collects it back together into a git structure, while index-filter just lets you use any git commands that affect the index (e.g. git rm --cached). the latter is much faster but some things are way easier with tree-filter.
13:12 jason237 joined #git
13:12 ffledgling jast: when you say collects them back into a git structure, what does that mean? Does it create a new commit with the changed files?
13:12 cYmen I have somehow gotten my head detached. :D
13:12 cYmen great things to say when working with git ^^
13:12 mikecmpb_ joined #git
13:12 cYmen can somebody tell me how to forward master with my commits?
13:13 ffledgling assuming your current branch is forked from master and they haven't diverged; git checkout master; git rebase <mybranch>
13:14 ffledgling Or are you trying to do something else?
13:14 cYmen I don't know
13:14 jason237 joined #git
13:14 cYmen my master is 4 commits behind head
13:14 cYmen I don't know where I screwed up to leave it behind but I want it back. :p
13:14 Didac joined #git
13:15 jast ffledgling: yes. index-filter does, too... tree-filter just has the extra work of writing the files into the index, after that the two both make a commit out of the index
13:15 moriko joined #git
13:15 lohfu joined #git
13:16 d1b joined #git
13:16 lohfu joined #git
13:16 Harzilein okay, this worked out okay, i now have:
13:16 ffledgling jast: hm, not quite sure how that applies to what I'm trying to do, but I might know more after I play around withit
13:17 Harzilein harzi@toybox:/usr/ports $ git log
13:17 Harzilein commit 4aca7dc04d59eb28af79a10d41a368167e49a827
13:17 Harzilein ...
13:18 Harzilein now when the change hits my upstream as well, will something detect that my local change is now empty?
13:18 Harzilein (on rebase)
13:19 jaguarmagenta joined #git
13:19 rubie joined #git
13:19 bsanford joined #git
13:20 MZAWeb joined #git
13:21 star_prone joined #git
13:21 sto0pkid joined #git
13:22 elfuego joined #git
13:22 ash_workz joined #git
13:23 jimi_ joined #git
13:23 jimi_ joined #git
13:24 crose joined #git
13:24 lohfu joined #git
13:24 stoopkid joined #git
13:25 gechr joined #git
13:26 byteflame joined #git
13:27 ExoUNX joined #git
13:28 MattMaker joined #git
13:29 st0opkid joined #git
13:30 rolha joined #git
13:30 ramsub07 joined #git
13:30 AaronMT joined #git
13:31 pluszak left #git
13:31 narendraj9 joined #git
13:32 Gsham joined #git
13:33 l4v2 joined #git
13:35 _ikke_ Harzilein: What do you mean, your local change being empty?
13:35 JeroenT joined #git
13:35 ffledgling TIL git cross-checks it's patches by comparing the +++ b/file/path line against the diff -- b/file/path line
13:35 mikecmpbll joined #git
13:35 johnmilton joined #git
13:35 mingrammer joined #git
13:36 Harzilein _ikke_: i'll try to make a minimal example, but i need more caffeine :D
13:36 Harzilein _ikke_: perhaps watching it makes me understand it
13:36 rnsanchez joined #git
13:37 fuchstronaut joined #git
13:38 settermjd joined #git
13:38 fmcgeough joined #git
13:40 nilg` joined #git
13:41 theelous3 joined #git
13:42 MattMaker joined #git
13:43 danslo joined #git
13:44 Harzilein _ikke_: hmm... it looks like it worked
13:44 nickthetait joined #git
13:45 ki0 joined #git
13:46 _ikke_ Harzilein: ok
13:47 dimi1947 joined #git
13:49 Mixologic_ joined #git
13:49 Harzilein _ikke_: dunno how it worked though
13:49 _ikke_ Can you explain? I haven't been following
13:51 bsanford joined #git
13:51 Harzilein _ikke_: i made a minimal example: http://paste.debian.net/785592/
13:52 MattMaker joined #git
13:53 Harzilein _ikke_: toybox cherry-picked from master, while pcbsd eventually merged the change from master
13:53 Harzilein _ikke_: when toybox rebases from pcbsd, the cherry-pick is gone
13:53 leeN joined #git
13:54 pR0Ps joined #git
13:55 _ikke_ I find that example hard to follow
13:56 _ikke_ but rebase recognizes duplicate patches, and drops the duplicated ones (cherry-picked commits)
13:56 Pulp joined #git
13:57 elfuego joined #git
13:57 danslo joined #git
13:58 Harzilein _ikke_: is there a way i can see this in action? adding -v was not very explicit
13:59 LionsMane joined #git
13:59 _ikke_ I'm not sure, but I though it would mentioning something about skipping a commit
14:00 johnmilton joined #git
14:01 JeroenT joined #git
14:04 GodGinrai joined #git
14:05 kedare joined #git
14:08 ki0 joined #git
14:09 stoopkid joined #git
14:09 Codfection joined #git
14:10 kaldoran joined #git
14:11 MattMaker joined #git
14:14 neovesta joined #git
14:14 barajasfab joined #git
14:15 ^7heo joined #git
14:15 ^7heo hey
14:15 ^7heo I'm trying to init some submodules, but without a .git folder
14:15 ^7heo is that even possible?
14:15 ^7heo or do I have to get a git repo in order to have submodules in it?
14:16 _ikke_ ^7heo: Not sure I understand you, but submodules to require a git repository, yes
14:17 ^7heo _ikke_: I downlodaded https://github.com/concourse/c​oncourse/archive/v1.6.0.tar.gz
14:17 ^7heo (yes, downlodaded, I like doing dad stuff)
14:17 ^7heo and then I tried to run git submodule update --init in it.
14:18 ^7heo (after extraction)
14:18 _ikke_ That's not going to work
14:18 ^7heo why do they even use submodules for distribution then? =/
14:18 GodGinrai ^7heo: There's no way to know about the submodules if you don't have the .git folder
14:19 ^7heo GodGinrai: aren't they in .gitmodules?
14:19 ^7heo GodGinrai: because the are in the .gitmodules file of that archive...
14:19 GodGinrai ^7heo: If you are going to download it like a dad, then you need to go full on dad and download each required piece.
14:20 ^7heo GodGinrai: haha. Yeah. But I could also git init in the folder of the extracted archive
14:20 _ikke_ The linked commit is stored in the repository itself (which is why you typically use submodules, to have a strong binding to specific commits)
14:20 ^7heo GodGinrai: and then use git submodule.
14:20 TheLawyer joined #git
14:20 ResidentBiscuit joined #git
14:21 psprint Where on the remote will git push if the refspec is "refs/heads/master:" ?
14:21 ResidentBiscuit joined #git
14:21 JanC_ joined #git
14:21 mikecmpb_ joined #git
14:21 TheLawyer I created .pub key and saved it into github account ... now I don't know how to add that into git config I suppose and tutorial from github doesn't show what to do next!
14:22 _ikke_ psprint: Not sure if that's a valid push refspec
14:22 _ikke_ TheLawyer: git does not need to know about it, just ssh
14:22 fahadash joined #git
14:22 psprint _ikke_: 190  cl_git_pass(git_push_add_refspec(push, "refs/heads/master:"));     here: https://git.kaarsemaker.net/libgit2/b​lob/fd536d29c127648abb2ce5f6f619135ce​69b9800/tests/network/remote/local.c
14:22 axl_ joined #git
14:23 jost joined #git
14:23 starkhalo joined #git
14:23 GodGinrai TheLawyer: the key is for ssh, not git.
14:23 Vinnie_win joined #git
14:24 elfuego joined #git
14:24 psprint _ikke_: I could do fetch refspec as "master:master" ? They do it in the code
14:24 _ikke_ psprint: I think the default is refs/heads then
14:24 vassagus joined #git
14:24 ChristopheT joined #git
14:25 felixjet joined #git
14:25 d0nkeyBOB joined #git
14:25 ChristopheT joined #git
14:25 ChristopheT left #git
14:26 harish_ joined #git
14:26 TheLawyer ok, so now I've created a new local testing repo .. how can I push it remotely directly? "without having to use github GUI"
14:26 sanketdg joined #git
14:26 tabrez joined #git
14:27 _ikke_ you need to add a remoete
14:28 phaleth joined #git
14:28 axl_ joined #git
14:28 elfuego joined #git
14:29 madewokherd joined #git
14:29 SourTrout joined #git
14:29 ^7heo disclaimer: it's the first time I use submodules
14:29 ^7heo but
14:29 ^7heo $ git submodule status src/github.com/BurntSushi/migration
14:29 ^7heo error: pathspec 'src/github.com/BurntSushi/migration' did not match any file(s) known to git.
14:29 ^7heo if I put a file in there
14:30 ^7heo and then add that path in git's index, commit
14:30 kbs ^7heo: just curious (as your approach seems more tricky) why you did you choose to download a tarball and try to splice a .git store, rather just git clone the repo in the first place?
14:30 theskillwithin joined #git
14:30 ^7heo kbs: because I need a release.
14:30 kbs ^7heo: you can checkout to the same tag
14:30 _ikke_ You can use --depth=1
14:30 TheLawyer shit I forgot all what I've learned couple of weeks ago, I must go back to my git notes
14:31 ^7heo kbs: the only alternative would be for me to clone the tag of the release; but our build system isn't designed to do that as easily as downloading tgz files
14:31 kbs ^7heo: I see, okay
14:31 dopesong joined #git
14:31 ^7heo kbs: yeah, using tgz files would reduce the amount of code to audit, maintain, the amount of possible bugs, and reuse more code.
14:32 ^7heo kbs: so it'd be much better if I could do it.
14:32 st0opkid joined #git
14:32 _ikke_ ^7heo: How do you know what commits of the submodules were comitted in the parent repo?
14:32 ^7heo however, I'm not sure about how to correctly fetch submodules in that case.
14:32 psprint _ikke_: that (you need to add a remote) seems to also apply to me:  if I fetch from URL, there will be no refs/remotes/NAME/BRANCH, i.e. no remote tracking branch, and I would pull the changes where... to refs/heads/BRANCH ?
14:32 ^7heo _ikke_: good question?
14:32 kbs ^7heo: hm, I don't know all your constraints - it does strike me though, that some part of the system will still need to understand git (in order to fetch the submodules) so it can't all work with tg files
14:32 ^7heo _ikke_: is that information supposed to be stored in the .git folder?
14:32 _ikke_ ^7heo: It is
14:33 ^7heo damn
14:33 ^7heo looks like I can't "spoof" the git folder.
14:33 _ikke_ in the tree to be precise
14:33 ^7heo meh, ok.
14:33 ^7heo that explains why gh knows what rev the submodules are at
14:33 ^7heo and not me =/
14:33 ^7heo ok, thanks _ikke_ :)
14:33 sto0pkid joined #git
14:33 ^7heo thanks also kbs for your suggestion that using git is anyway better.
14:34 ^7heo it is right in the end.
14:34 fairuz joined #git
14:34 vassagus joined #git
14:34 ^7heo now it's a build system problem; I'll be back should I hit another roadblock.
14:35 _ikke_ psprint: I think just to FETCH_HEAD
14:36 MattMaker joined #git
14:36 psprint _ikke_: and normally, if I would fetch via symbolic name (like "origin"), to FETCH_HEAD and ... ?
14:36 SourTrout joined #git
14:37 _ikke_ if you fetch from a named remote, both FETCH_HEAD and the remote/remotes/<remote>/<branch> refs are updated
14:37 courrier joined #git
14:38 psprint _ikke_: you mean "refs/remotes/..." or indeed "remote/remotes/..." ?
14:38 LionsMane joined #git
14:38 stoopkid joined #git
14:39 _ikke_ psprint: refs/remotes
14:39 psprint thanks
14:40 pool joined #git
14:40 mikecmpbll joined #git
14:41 g5pw joined #git
14:41 solf1re joined #git
14:42 sylvain joined #git
14:42 lindenle joined #git
14:42 sangy joined #git
14:42 moriko joined #git
14:48 The_Phoenix joined #git
14:49 leehambley joined #git
14:50 sangy joined #git
14:50 MattMaker joined #git
14:50 cYmen joined #git
14:51 psprint how to display what unmerged objects are in FETCH_HEAD?
14:52 elfuego joined #git
14:53 jast git log ..FETCH_HEAD
14:53 Macaveli joined #git
14:53 jast in case you're talking about commits you haven't merged yet
14:54 jim__ joined #git
14:54 psprint yes
14:56 psprint I see that doing  `git fetch http://... somebranch:` indeed downloads objects and set tip pointer FETCH_HEAD, and doing that again for `... otherbranch:` just overwrites FETCH_HEAD, meaning that the previous objects will be garbage collected?
14:56 noidea joined #git
14:56 _ikke_ psprint: Eventually, yes
14:57 _ikke_ As long as they're not somehow referenced
14:57 ribasushi joined #git
14:58 Derperperd joined #git
14:58 mcqr joined #git
15:00 eijk joined #git
15:00 thiago joined #git
15:01 psprint well I ran gc and it wasn't very talkative, had 70 objects downloaded with fetch somebranch:, no common commits, overwritten FETCH_HEAD by fetching other commit, so the 70 should be stray, but git gc showed just "Counting objects: 30, done.", and stats on compression, so no number "70" there, weird
15:01 PCatinean joined #git
15:02 psprint s/other commit/other branch/
15:02 danslo why worry about it?
15:02 Eryn_1983_FL joined #git
15:04 sbeller_ joined #git
15:04 leeN joined #git
15:04 stoopkid joined #git
15:04 RksJ joined #git
15:05 cyphase joined #git
15:05 RksJ joined #git
15:05 cyan__ joined #git
15:05 dopesong_ joined #git
15:06 romerocesar joined #git
15:07 clemf joined #git
15:07 dopeson__ joined #git
15:08 cYmen joined #git
15:09 Darren_ joined #git
15:09 burtons joined #git
15:10 MattMaker joined #git
15:12 Gsham joined #git
15:12 tyreld joined #git
15:13 dimi1947 joined #git
15:15 jast psprint: in any case, 'git gc' doesn't *immediately* garbage collect. objects are kept around for a certain number of days (configurable), and also for as long as they are listed in a reflog
15:15 wrouesnel joined #git
15:18 MrMojito joined #git
15:18 weckl_ joined #git
15:19 MattMaker joined #git
15:20 eijk_ joined #git
15:20 dmto joined #git
15:21 bsanford_ joined #git
15:21 subhojit777 joined #git
15:22 _ikke_ defaults (object age: 14 days, reflog expiry: 90 days)
15:22 _ikke_ but the reflog does not matter in this case
15:23 mingrammer joined #git
15:24 IWishIKnew joined #git
15:25 netcarver joined #git
15:26 Manuel15 joined #git
15:26 Manuel15 joined #git
15:26 gargola joined #git
15:27 SourTrout joined #git
15:27 clemf joined #git
15:27 cyan__ joined #git
15:28 weckl__ joined #git
15:29 MattMaker joined #git
15:32 citizen-stig joined #git
15:33 skroon joined #git
15:33 skroon is it valid to have this in .gitattributes:      * text eol=lf       ?
15:34 neovesta left #git
15:34 Gitzilla joined #git
15:35 neovesta joined #git
15:35 clemf joined #git
15:37 SourTrout joined #git
15:38 xaa joined #git
15:38 The_Phoenix joined #git
15:39 netcarver joined #git
15:40 neovesta left #git
15:40 oich joined #git
15:40 neovesta joined #git
15:41 Lomex joined #git
15:41 thiago joined #git
15:41 thiago joined #git
15:42 _ikke_ Sure, but you have to watch out for binary files (like images)
15:42 shinnya joined #git
15:42 ghoti joined #git
15:43 PCatinean joined #git
15:43 dreiss joined #git
15:44 MattMaker joined #git
15:47 lindenle joined #git
15:50 InfoTest joined #git
15:50 synthroid joined #git
15:51 meLon joined #git
15:51 cdg joined #git
15:52 Lomex joined #git
15:52 rinwa joined #git
15:52 madewokherd joined #git
15:52 FuzzySockets joined #git
15:54 MattMaker joined #git
15:54 TheLawyer < back > I got the remote add thing ... now I have a tottally different question .. what if I don't have any browser to create a repo on github ... I want to create it locally then push it directly to github so that git is the resposible for creating it.
15:54 TheLawyer is that possible?
15:55 _ikke_ look at !hub
15:55 gitinfo Hub is a wrapper around git which provides specific github features to the git cli. See https://github.com/github/hub for details.
15:55 _ikke_ Note sure if it allows you to create projects though
15:56 Nethereal joined #git
15:56 EvilPenguin joined #git
15:56 ericho joined #git
15:58 GodGinrai TheLawyer: github has an API you can use: https://developer.github.com/v3/repos/#create
15:59 GodGinrai you would use that API to create your repo, then use git to push your local repo to it once it is created
15:59 ericho joined #git
16:00 ^7heo thanks for the answers GodGinrai and _ikke_
16:00 ^7heo left #git
16:00 stantonk joined #git
16:01 meLon joined #git
16:07 danimal2 joined #git
16:08 TomyWork joined #git
16:09 SteffanW2 joined #git
16:10 stantonk joined #git
16:10 Electrometro joined #git
16:10 subhojit777 joined #git
16:10 moritz TheLawyer: there are also command line clients available for the github API
16:10 moritz TheLawyer: like https://github.com/ingydotnet/git-hub
16:12 kpease joined #git
16:14 Mannn joined #git
16:14 griffindy joined #git
16:17 st0opkid joined #git
16:17 skroon joined #git
16:18 digidog joined #git
16:18 mingrammer joined #git
16:18 NeverDie joined #git
16:18 luxflow joined #git
16:19 MattMaker joined #git
16:19 ansimation joined #git
16:19 romerocesar joined #git
16:20 vuoto joined #git
16:20 segment-fault joined #git
16:20 LeMike joined #git
16:22 luxflow left #git
16:23 aep-shoutlet joined #git
16:25 elfuego joined #git
16:25 durham joined #git
16:26 ash_workz joined #git
16:27 elfuego joined #git
16:27 star_prone joined #git
16:28 burtons joined #git
16:29 LiohAu joined #git
16:30 stantonk joined #git
16:31 danslo joined #git
16:31 stantonk joined #git
16:31 phanimahesh joined #git
16:32 vuoto joined #git
16:32 MattMaker joined #git
16:32 dopesong joined #git
16:32 cdg joined #git
16:34 vuoto joined #git
16:34 HoierM joined #git
16:36 leehambley joined #git
16:37 mablae joined #git
16:43 MattMaker joined #git
16:43 hexa- joined #git
16:44 hexa- hi, is there a git hook that triggers on fetch & reset --hard origin/master?
16:44 IWishIKnew joined #git
16:45 cagedwisdom joined #git
16:46 sto0pkid joined #git
16:47 tvw joined #git
16:47 weckl_ joined #git
16:47 GodGinrai hexa-: nope
16:47 GodGinrai I'm surprised there isn't one for the reset
16:47 GodGinrai but it makes sense for the fetch
16:48 GodGinrai since fetch doesn't actually *change* your code
16:48 clemf joined #git
16:51 vuoto joined #git
16:51 vuoto joined #git
16:52 vuoto joined #git
16:53 durham_ joined #git
16:53 ResidentBiscuit Can't you also just do create a new remote pointing to something on GitHub that doesn't exist, and then push to that remote?
16:54 chrispg21 joined #git
16:57 Twirl left #git
16:57 _ikke_ no
16:58 GodGinrai ResidentBiscuit: if you could do that, then Github's instructions for "uploading a new project from the CLI" wouldn't start with "create a new repository on the github website"
16:58 _ikke_ Any typo would create a new reoositry
16:59 GodGinrai ^ that too
16:59 ResidentBiscuit Maybe I'm crazy. Time to try!
16:59 _ikke_ Lots of confused people that don't their repos updated
16:59 _ikke_ +see
17:00 stantonk joined #git
17:00 vavovi joined #git
17:00 GodGinrai and lots of filespace used up with copy typo repos
17:01 jim__ I had a project that i decided needed a total rewrite. are there any risks to grafting the old repo on to the first commit with info/grafts ?
17:02 nd joined #git
17:02 durham joined #git
17:02 ahmedelgabri joined #git
17:05 irco joined #git
17:05 Lomex joined #git
17:06 Guest55 joined #git
17:06 MattMaker joined #git
17:08 texinwien_ joined #git
17:08 peepsalot what's the simplest way to just get the md5 of the head of current branch?
17:08 peepsalot er sha1 or whatever it is
17:08 sanketdg joined #git
17:08 ShekharReddy joined #git
17:10 Literphor joined #git
17:11 dzho git log?
17:11 ilmari git rev-parse HEAD
17:12 dzho nice
17:12 emPi joined #git
17:12 mingrammer joined #git
17:14 dopesong joined #git
17:15 peepsalot thanks ilmari
17:19 subhojit_ joined #git
17:20 jaguarmagenta joined #git
17:20 LeBlaaanc joined #git
17:21 MattMaker joined #git
17:21 vuoto joined #git
17:22 mattcen joined #git
17:22 LeBlaaanc joined #git
17:24 vuoto joined #git
17:24 Ryanar joined #git
17:25 Ryanar hey guys, I forked a repository, pushed a commit accidentally to the master branch, and now everytime I fetch upstream or make branches off my forked master branch it includes this extra unwanted commit, how can I remove this one commit from the history?
17:25 jason237 joined #git
17:26 SwiftMatt joined #git
17:26 Ryanar I tried `git rebase --onto 08519ea8e5de22f764d04664995eec99dca03f33^ 08519ea8e5de22f764d04664995eec99dca03f33 HEAD` and it left me in a state of `HEAD detached from refs/heads/develop`
17:29 PioneerAxon joined #git
17:29 osse Ryanar: would another way to phrase it be: "make your master identical to the upstream one" ?
17:29 Ryanar yeah
17:30 skalpin Ryanar: git reset <id of commit>
17:30 LeMike joined #git
17:30 osse Ryanar: git reset --hard upstream/master
17:30 osse or origin/master or whathaveyou
17:30 Ryanar if
17:30 shinnya joined #git
17:30 Ryanar I have a branch that has work in it
17:30 Ryanar will I lose it
17:30 _sjs joined #git
17:30 skalpin make sure you have master checked out
17:30 osse a different branch? no. the current one? yes
17:31 Ryanar so now if I have a separate branch that has work on it that has this extra commit in it that I dont want
17:32 Ryanar what would be the best way to get rid of it in the branch without making a new branch from the properly reset master and redoing it?
17:32 osse then you have to do some trickery along the lines of what you tried
17:32 skalpin if it's just one commit, you could try reverting it
17:32 osse git rebase -i HEAD~6 (or some suitably large number), remove the line corresponding to the unwanted commit. save and close
17:33 GodGinrai Ryanar: `git fetch origin; git reset --hard origin/master` on your master branch
17:33 GodGinrai where "origin" is your upstream remote
17:33 GodGinrai (so if origin isn't your upstream remote, then just replace origin with the correct remote)
17:34 quetz joined #git
17:34 elementalest joined #git
17:34 elementalest joined #git
17:34 Ryanar osse: since I already pushed the commit, should I make a new branch off of it so its only in local before pushing
17:34 osse Ryanar: branches are always local
17:35 lindenle joined #git
17:35 Ryanar yeah I think that worked, thanks!
17:36 MattMaker joined #git
17:36 adino joined #git
17:36 romerocesar joined #git
17:37 emb joined #git
17:37 emb left #git
17:38 emb2 joined #git
17:38 sangy joined #git
17:41 adino joined #git
17:43 mostlybadfly joined #git
17:44 ktosiek joined #git
17:44 nitric joined #git
17:47 star_prone joined #git
17:48 NeXTSUN joined #git
17:49 Limes_ joined #git
17:51 emb2 If I want to make a branch's head look like another branch's head, but still keep all the history it currently has, what is the best way to do that?
17:51 emb2 Or is that even possible
17:52 preaction sounds like you want to merge
17:52 emb2 But merging won't do anything because the branch already has the changes
17:52 osse emb2: git checkout otherbranch -- . && git commit -m 'Change branch to be equal to otherbranch'
17:52 emb2 Basically, I have the master branch and a staging branch
17:52 preaction then what's the problem?
17:52 emb2 and I staging to look like master
17:52 emb2 As it has gotten messy
17:53 osse does it have history worth keeping that master doesn't have?
17:53 _ikke_ What has become messy?
17:53 preaction why has it gotten messy?
17:53 daivyk joined #git
17:53 emb2 It is causing merge conflicts that don't make sense (that don't occur on master)
17:54 kadoban joined #git
17:54 andschwa joined #git
17:54 GodGinrai osse: doesn't he also have to use -a?
17:54 osse GodGinrai: checkout also stages
17:54 GodGinrai oh, does it?
17:54 GodGinrai I wasn't aware
17:55 emb2 Thank you osse :) I will consider this but I assume the history will also change?
17:55 osse emb2: just the one new commit
17:55 GodGinrai emb2: no.
17:55 GodGinrai emb2: you will get a new commit
17:55 emb2 Oh!
17:55 MattMaker joined #git
17:55 emb2 That is perfect then
17:55 GodGinrai emb2: but you are not checking out a different branch, just "files" from another branch
17:55 GodGinrai emb2: so the rest of history will remain the same
17:55 fahadash joined #git
17:55 emb2 So do I do that while on master?
17:55 emb2 Or on staging
17:56 GodGinrai emb2: on the one you want to change
17:56 GodGinrai and "otherbranch" is the branch you want to pull the files from
17:56 Pulp joined #git
17:56 ska joined #git
17:56 GodGinrai osse: correct me if I'm wrong, but this won't remove any files that don't exist in otherbranch, right?
17:56 ska How do i show the log of branch feature/xyz when I'm in develop branch?
17:57 rgrinberg joined #git
17:57 _ikke_ git log develop..feature/xyz
17:57 _ikke_ (assuming the feature branch is based off develop)
17:57 osse GodGinrai: correct
17:57 ska ok..
17:58 SourTrou_ joined #git
17:58 emb2 Is there any way to remove the files that don't exist in the other branch?
17:59 Sasazuka__ joined #git
17:59 osse emb2: git rm -r --cached . && git checkout otherbranch -- .
17:59 osse run from top level
18:00 emb2 Thanks :)
18:01 vassagus joined #git
18:01 osse emb2: are you sure you don't want to just nuke staging and make it identical to masteR?
18:01 MattMaker joined #git
18:01 GodGinrai osse: that would make too much sense :P
18:02 krstf joined #git
18:03 riceandbeans joined #git
18:04 andschwa joined #git
18:04 jstimm joined #git
18:04 vassagus_ joined #git
18:04 riceandbeans what causes a git merge to require a commit msg vs not?
18:04 riceandbeans I thought branch merging always required a commit message
18:04 riceandbeans I just merged a branch into my master branch, the other branch had only one commit, and it basically just did a cherry-pick
18:05 TomyLobo joined #git
18:06 TheLawyer joined #git
18:06 _ikke_ riceandbeans: it probably did a fast-forward merge
18:06 zacts joined #git
18:06 MZAWeb joined #git
18:06 _ikke_ (it tells you when it does)
18:06 _ikke_ (it's not actually a merge
18:06 _ikke_ )
18:07 mikecmpbll joined #git
18:07 dopesong joined #git
18:08 riceandbeans yes, it was a fast forward merge
18:08 riceandbeans it felt like a rebase+cherry-pick
18:08 vassagus joined #git
18:09 LiohAu joined #git
18:09 _ikke_ when history is not diverging, git by default will just fast-forward the branch instead of actually merging
18:10 WayToDoor joined #git
18:10 Gsham joined #git
18:12 riceandbeans is that a new thing?
18:13 riceandbeans and am I off in my description that it was basically a rebase+cherry-pick?
18:13 dopesong joined #git
18:13 mutex joined #git
18:13 _ikke_ riceandbeans: no, that's not a new thing
18:14 _ikke_ riceandbeans: rebase and cherry-pick are almost the same
18:15 osse riceandbeans: http://sprunge.us/ZIPj
18:16 stantonk joined #git
18:16 _ikke_ osse: Do you have something to draw those diagrams, or is that all by hand?
18:16 durham joined #git
18:16 nwkj86 joined #git
18:16 p4trix joined #git
18:17 osse _ikke_: by hand
18:17 _ikke_ ok
18:17 osse _ikke_: and a lot of copy paste
18:17 osse i have a whole dir of them
18:18 eivarv joined #git
18:18 _ikke_ I notice
18:19 durham_ joined #git
18:19 GodGinrai osse: impressive
18:20 osse and many more than in that git-stuff repo
18:20 osse i've lost control :/
18:21 rgrinberg joined #git
18:22 nilg joined #git
18:22 durham joined #git
18:22 altendky joined #git
18:24 riceandbeans so basically if you do --no-ff
18:26 MZAWeb_ joined #git
18:27 altendky i just did a merge and it had no conflicts, but it also doesn't work.  is there a simple way to launch a three way diff as would have happened if there had been a conflict?
18:28 astrofog joined #git
18:28 GodGinrai altendky: why do you need a 3-way diff when you could just diff the 2 non-conflicting branches?
18:30 MattMaker joined #git
18:30 altendky GodGinrai: perhaps i don't but i thought the visualization of the diff between the successful merge and it's two parents would be helpful for me to understand the correct merge that wouldn't be broken.
18:31 GodGinrai I see
18:32 altendky GodGinrai: i think i can resolve this without, but it seemed like it would be achievable even without a conflict
18:34 danslo_ joined #git
18:37 Codfection joined #git
18:39 romerocesar joined #git
18:40 editshare-ericf joined #git
18:43 SourTrout joined #git
18:43 airdisa joined #git
18:43 elsevero joined #git
18:44 Derperperd joined #git
18:46 SourTrou_ joined #git
18:48 Dro joined #git
18:49 riceandbeans left #git
18:50 splud joined #git
18:51 cyan__ joined #git
18:52 garethdaine joined #git
18:53 Cryp71c joined #git
18:53 mwhooker joined #git
18:53 splud I’m trying to sort out how to set up a workflow where master is managed by an integrator, but other branches can be commited to by other developers.  Ideally, devs can manage their contributions on the dev branch(es), and the integrator can approve those into master (where the CI system picks them up from).  Any good docs on how to achieve this?
18:53 splud Emailing patch files around isn’t a convenient workflow...
18:54 MattMaker joined #git
18:54 m0viefreak joined #git
18:56 mikecmpbll joined #git
18:56 kadoban splud: I'd likely just set up gitolite on a server somewhere, and each user could have their own namespace that they're allowed to create branches in. Then you can just do pull requests, either my email or whatever. You can give per-branch permissions for reading/writing branches, so it should work nicely.
18:56 kadoban There's fancier systems all set up I'm sure, I just don't know them.
18:57 GodGinrai splud: you could have a pre-receive hook that aborts all pushes it receives to master that aren't from the correct user.
18:58 GodGinrai or calls some bigger script that selectively aborts or allows based on branch chosen and user pushing
18:58 kadoban If you're going to manually do a pre-receive thing, gitolite would likely just be better, it at least abstracts a lot of that fun for you.
18:58 nitric and depending on how clever/malicious your developers are, they can easily spoof their author past a hook, couldn't they?
18:59 GodGinrai I mean, honestly, if you want master to be protected, then everyone should have RO access to the repo, and RW access to forks
18:59 neeca joined #git
18:59 GodGinrai then you just set up the permissions for the forks with standard unix permissions per user
18:59 nitric splud: another alternative to gitolite (besides github obviously) is gitlab
18:59 usm4n joined #git
19:00 GodGinrai nitric: *user*, not author
19:00 mingrammer joined #git
19:02 emb2 joined #git
19:02 nitric oh, the actual ssh user is in the environment, good point.
19:02 star_prone joined #git
19:02 GodGinrai yup
19:05 SourTrout joined #git
19:06 splud this is locally administered git, not hosted by service.
19:06 garethdaine joined #git
19:06 splud I’ve been hedging towards the forking workflow, was just thinking it’d be nice if it was one repo, not multiple.
19:07 kadoban None of the suggested things have been hosted services.
19:09 nitric splud: gitlab's self-hosted setup (apt-get install gitlab-ce) is fantastic if gitolite is too lite
19:09 MattMaker joined #git
19:09 robotroll joined #git
19:10 GodGinrai splud: there is nothing nice about one repo for multiple people.  <.<
19:10 usm4n joined #git
19:11 kadoban There can be, if it's set up right.
19:11 kadoban I've use a single central gitolite setup for a fairly large team and it worked great, you just have to keep people's playgrounds separate.
19:13 mikecmpbll joined #git
19:13 GodGinrai We've got over 40 already merged branches just sitting on remote because my teammates won't delete them after a merge.  And we're a small team
19:13 MattMaker joined #git
19:14 sdaugherty The most common setup I see is one "canonical" repo, and everyone having a local "working" repo  cloned from it.
19:14 sdaugherty That forces them to deal with merges locally, therefore preventing them from mucking up things for everyone else.
19:15 kadoban GodGinrai: As long as they're namespaced away somewhere, that doesn't seem to matter. I don't care what branches joe has in the joe/whatever namespace on the server. If I set up my fetch specs correctly, I don't even see them (unless joe specifically tells me to look at one, because it's ready to merge or whatever)
19:16 GodGinrai kadoban: we name them after our ticket numbers, so they aren't namespaced
19:16 ahmedelgabri joined #git
19:17 SourTrout joined #git
19:17 kadoban Well if it's something like tickets, it should be easy to tell when they can go away I guess. Just giving someone the job of doing that occasionally doesn't seem too bad, or automating it if you really want.
19:17 jimi_ joined #git
19:17 LeBlaaanc joined #git
19:18 sdaugherty everyone working out of their own clone of the repo and push/pull actually provides a great deal of security
19:18 GodGinrai or my team could just click the "delete branch" button on github that shows up right after they click the "merge this PR" button like they've been told to
19:18 afiskon joined #git
19:18 FuzzySockets joined #git
19:19 GodGinrai the point is that people can't be trusted to follow the rules, and forcing them to muck up their own forks instead of the main repo at least keeps them from mucking it up for everyone else
19:19 MattMake_ joined #git
19:20 sdaugherty if I have  a cloned repo that I work from, and someone tries to rewrite history in  the "main" repo, then I'll immediately know because push//pulll will fail.
19:20 sdaugherty That's a  very useful security property
19:20 GodGinrai sdaugherty: if you have an integration manager, they can't even rewrite history unless they *are* the integration manager
19:22 dsdeiz joined #git
19:22 dsdeiz joined #git
19:23 unbalancedparen joined #git
19:27 d0nkeyBOB joined #git
19:28 LeMike joined #git
19:28 fairuz joined #git
19:28 nitric joined #git
19:30 star_prone joined #git
19:31 adino joined #git
19:32 nitric joined #git
19:32 dsdeiz_ joined #git
19:33 SourTrout joined #git
19:35 danslo_ joined #git
19:35 ash_workz joined #git
19:35 Derperperd joined #git
19:35 iota_pi joined #git
19:37 MattMaker joined #git
19:37 elastix joined #git
19:40 usm4n joined #git
19:40 ki0 joined #git
19:42 romerocesar joined #git
19:42 ejb joined #git
19:43 leehambley joined #git
19:45 cyan__ joined #git
19:46 vuoto joined #git
19:47 aguimaraes1986 joined #git
19:48 lindenle joined #git
19:49 GavinMagnus joined #git
19:50 eijk_ joined #git
19:51 IWishIKnew joined #git
19:54 sagerdearia joined #git
19:55 paul424 joined #git
19:55 mingrammer joined #git
19:57 rgrinberg joined #git
20:02 mwhooker joined #git
20:03 SourTrout joined #git
20:03 splud another 500MB per developer for this one repo.  Ick.
20:03 Faylite_ joined #git
20:03 star_prone joined #git
20:05 cdown joined #git
20:07 _ikke_ splud: have you looked at !gitolite
20:07 gitinfo splud: Gitolite is a tool to host git repos on a server. It features fine-grained access control, custom hooks, and can be installed without root. Download: https://github.com/sitaramc/gitolite Docs: http://gitolite.com/gitolite/
20:10 aboseley joined #git
20:11 abs25 joined #git
20:11 abs25 can I have git alias for: git pull origin <current_branch>
20:11 ojacobson you almost certainly don't want that, exactly
20:11 ojacobson "git pull" without arguments is very likely what you want.
20:12 ojacobson git pull merges (rebases, with options) the upstream of the current branch
20:12 _ikke_ If you setup remote tracking information
20:12 eivarv joined #git
20:12 ojacobson "git pull origin <current branch name>" will merge the branch with the same name as your local branch, as fetched from origin, regardless of whether that makes any sense
20:12 Mannn joined #git
20:14 MattMake_ joined #git
20:14 cdown_ joined #git
20:15 a3Dman joined #git
20:15 Mannn joined #git
20:17 Mannn joined #git
20:18 nwkj86 joined #git
20:18 zacts joined #git
20:18 abs25 ojacobson: and what does git push origin <branch name >do?
20:18 abs25 because when I do just git push and if there is no remote branch, it will give that message
20:18 ojacobson copies <branch name> from your local repo to origin
20:18 abs25 and block my script from working
20:18 abs25 ojacobson: so how would I do git pushO alias for git push origin <current_branch>
20:18 ojacobson (or fails if <branch name> in origin is not a strict ancestor of <branch name> locally
20:19 ojacobson git push origin HEAD
20:19 abs25 that would do the same?
20:19 Mannn joined #git
20:20 lucasem joined #git
20:23 Arban joined #git
20:24 anu0 joined #git
20:25 nitric joined #git
20:28 BSaboia joined #git
20:31 dopesong joined #git
20:35 cdown_ joined #git
20:35 danslo_ joined #git
20:38 davisonio joined #git
20:40 usm4n joined #git
20:40 PioneerAxon joined #git
20:42 Kicer86 joined #git
20:43 MattMaker joined #git
20:44 romerocesar joined #git
20:45 neanderslob joined #git
20:46 HoierM joined #git
20:47 neanderslob_ joined #git
20:49 neanderslob joined #git
20:49 Repox joined #git
20:49 sargas joined #git
20:50 mingrammer joined #git
20:52 SourTrout joined #git
20:53 ramsub07 joined #git
20:54 boombatower joined #git
20:55 aidalgol joined #git
20:55 dtrott joined #git
20:57 PioneerAxon joined #git
20:57 garethdaine joined #git
20:58 mwhooker joined #git
20:58 kaldoran joined #git
21:01 TheLawyer joined #git
21:02 LionsMane joined #git
21:02 MattMaker joined #git
21:04 gargola joined #git
21:04 Finnoloid joined #git
21:05 Literphor joined #git
21:08 LeBlaaanc joined #git
21:09 SourTrout joined #git
21:13 neanderslob joined #git
21:13 neanderslob_ joined #git
21:13 MattMaker joined #git
21:14 AndreasLutro left #git
21:15 neanderslob joined #git
21:16 st0opkid joined #git
21:18 tvw joined #git
21:19 stoopkid joined #git
21:19 leehambley joined #git
21:20 Duder9999 joined #git
21:20 yoan_ joined #git
21:21 MiniMax joined #git
21:21 jeffreylevesque joined #git
21:21 fairuz joined #git
21:22 jaguarmagenta joined #git
21:23 eijk_ joined #git
21:27 King_Hual joined #git
21:28 Dougie187 joined #git
21:30 zacts_raincity joined #git
21:31 vuoto_ joined #git
21:34 Mannn joined #git
21:37 WaReZ joined #git
21:37 catag87 joined #git
21:39 SwiftMatt joined #git
21:39 moriko joined #git
21:40 FuzzySockets joined #git
21:43 sunny256 joined #git
21:44 mingrammer joined #git
21:44 fairuz joined #git
21:44 MattMaker joined #git
21:45 ramsub07 hello, how do i see the changes before doing "git add" ?
21:45 _ikke_ git diff
21:46 Gsham joined #git
21:46 neanderslob joined #git
21:46 romerocesar joined #git
21:46 neanderslob_ joined #git
21:47 jimi_ joined #git
21:47 SourTrou_ joined #git
21:48 SourTrou_ joined #git
21:51 yitz I got a git repo I want to "move" into a new repo. How do I go about doing that?
21:51 adino joined #git
21:51 mikecmpbll joined #git
21:51 BSAlb joined #git
21:51 _ikke_ What do you want the end result to be?
21:51 yitz As much state/content/info from repo A showing up in the new repo B
21:52 _ikke_ Does the new repository already have commits?
21:52 SourTrout joined #git
21:52 neanderslob joined #git
21:52 adino joined #git
21:52 yitz No. The new repo doesn't exist yet
21:53 _ikke_ Why do you need to have a new repository while still having the old history?
21:53 yitz The old repo is in a personal space in gitlab and I want to move it into a team owned repo
21:53 _ikke_ No need to create a new repository then
21:53 _ikke_ Just add a remote and push it
21:54 SourTrout joined #git
21:54 aidalgol New git alias: git nuke => git reset --hard && git clean -fdxq
21:54 _ikke_ You can even transfer it in gitlab
21:54 yitz Ooh. Transfer sounds nice :D
21:54 yitz I'll ping the owner and see if they can just do that
21:54 _ikke_ If you own the current repository, you can do it yourself
21:54 yitz With a new remote, I'd need to make the new repo and do an init?
21:54 _ikke_ No
21:55 _ikke_ just git remote add <name> <url>
21:55 _ikke_ in the existing repo
21:55 yitz And that magically creates the repo when I push it?
21:55 _ikke_ no, that reposistory needs to exist
21:56 _ikke_ no, that reposistory needs to exist
21:56 _ikke_ ugh
21:56 _ikke_ So transfering the project is the easiest way
21:56 yitz Yeah. Pinging the person that made the original repo about a transfer
21:57 eka joined #git
21:57 _ikke_ ok
21:57 dopesong joined #git
21:57 byteflame joined #git
21:58 clemf_ joined #git
21:58 yitz If transfering is possible, that's definitely the cleanest
21:58 yitz Thanks
21:59 spudowiar joined #git
22:00 kbs _ikke_: curious - what's your own opinion on self-hosted code-review systems on top of git for small teams? [context: I've mostly know/use gerrit, like the code review model, but not the UI or the CI "integration". Poking at gitlab, and you mentioned gitolite earlier - haven't looked at it yet.]
22:01 _ikke_ kbs: gitolite is only repository hosting, no web UI
22:02 cdown_ joined #git
22:02 kbs ah, okay I see.
22:02 dopesong joined #git
22:02 _ikke_ If you want a web UI, gitlab is worth looking into
22:03 _ikke_ (though I have no experience setting it up myself)
22:05 eka joined #git
22:05 Darren_ joined #git
22:06 kbs Cool, thanks.
22:08 mwhooker joined #git
22:10 SourTrou_ joined #git
22:11 drbean joined #git
22:12 SourTrou_ joined #git
22:12 neanderslob joined #git
22:12 fahadash joined #git
22:15 ResidentBiscuit joined #git
22:15 joki joined #git
22:19 xaa joined #git
22:20 MattMaker joined #git
22:21 Cryp71c joined #git
22:22 harish_ joined #git
22:23 fairuz joined #git
22:24 f3r70rr35f joined #git
22:26 yitz If I got a directory with a bunch of files in it and I want to move that all into a git repo, what's the correct/clean way to do that? No conflict between the files in the git and git history vs the directory
22:26 Spec-Chum joined #git
22:27 _ikke_ Just move it in there?
22:27 SourTrout joined #git
22:28 yitz Well, yeah. I was hoping I could ... check out the repo over/into the existing tree
22:28 Spec-Chum Hiya, complete noob to git.  I've cloned a full repo but have now decided to fork it instead.  How do I point my already downloaded repo to the fork instead?  Seems a waste to download it again, it's fully up to date
22:28 mikecmpbll joined #git
22:28 _ikke_ Spec-Chum: git remote set-url origin <new_url>
22:28 Spec-Chum will git clone myforkedrepo on the same dir just work?  I've not tried in case I bork something
22:29 _ikke_ No
22:29 Spec-Chum _ikke_, oh, easy as that?
22:29 _ikke_ yes
22:29 nincha joined #git
22:29 Spec-Chum incidentally, what would happened if I'd have done 'git clone myforkedrepo' on the same directory?
22:30 TimDix joined #git
22:30 Spec-Chum obviously the files would already be 100% there and the same
22:30 n-st joined #git
22:30 _ikke_ I think it would create a subdirectory in there with a new clone
22:31 PabloR joined #git
22:31 Spec-Chum ah, a clone in a clone, that won't end well
22:31 moriko joined #git
22:32 amdi_ joined #git
22:32 nincha left #git
22:32 TimDix Hello, I'm having a problem installing a project with npm, and it's failing on a git line... I'm not sure why, but I'm not getting anything when using https://github.com, but I do get it with git://, I'm trying to figure out how to make the https:// version work... suggestions? Here's me running it: https://gist.github.com/TimDix/1​78e9cc4fc46614131f9b3532ecd4b4b
22:32 Spec-Chum thanks _ikke_ :)
22:33 yitz So, could I check a repo out over an existing directory tree?
22:34 holodoc joined #git
22:34 mostlybadfly joined #git
22:35 ShekharReddy joined #git
22:36 danslo joined #git
22:36 red__ joined #git
22:37 kbs yitz: I think what _ikke_ suggested is to move/copy your directory over to a checked out rep, and git add/commit the newly copied directory
22:39 neanderslob joined #git
22:39 safe joined #git
22:41 nitric TimDix: hmm, i get the same output for both git:// and https:// with git 2.5.4, are you behind a proxy that's doing weird things with https traffic...?
22:42 TimDix Not that I know of, is there a way I can confirm?
22:42 dopesong joined #git
22:44 neanderslob joined #git
22:44 Timvde I'm doing a clone via https. What could be wrong if it hangs at "Cloning into 'my-repo'..."?
22:44 Timvde (-v doesn't give any more output)
22:44 nitric TimDix: stick `GIT_CURL_VERBOSE=1 GIT_TRACE=1 ` in front of the command and see if you're getting a clean http response
22:45 Timvde nitric: stuck at "* Trying 104.192.143.2..."
22:46 Timvde (That's Bitbucket)
22:46 Timvde Oh, maybe firewall...
22:46 TimDix nitric: https://gist.github.com/TimDix/1​78e9cc4fc46614131f9b3532ecd4b4b - I ran the command and got this response, see anything that could indicate the issue?
22:47 PabloR git checkout <file> vs git checkout -- <file>. Is the difference between them that the 1st one modifies the working-tree’s file to match the repo’s version of that file and the 2nd one just discards all the changes in the working-tree instead?
22:47 kugel_ joined #git
22:47 kugel_ joined #git
22:47 dopesong joined #git
22:48 nitric TimDix: can you re-format that gist? i don't see the http request to `> GET /ember-cli/ember-cli-test-loader.gi​t/info/refs?service=git-upload-pack HTTP/1.1` that i would expect. what version of git are you using?
22:48 Literphor joined #git
22:49 romerocesar joined #git
22:49 nnyk joined #git
22:50 TimDix nitric: Updated for better formatting.
22:50 TimDix https://gist.github.com/TimDix/1​78e9cc4fc46614131f9b3532ecd4b4b
22:50 kyan joined #git
22:51 TimDix $ git --version => git version 1.8.3.4
22:52 texinwien_ joined #git
22:53 cuechan joined #git
22:53 packetswitcher joined #git
22:54 cyan__ joined #git
22:55 TimDix Ok, weird... Why does git --version tell me 1.8.3.4, but the trace response says "User-Agent: git/1.7.1"
22:55 nitric TimDix: the only difference between your debug output and mine is the "* Expire cleared" bit at the very end, and i can't find a good explanation for that message
22:56 nitric oh good catch, that is odd
22:56 nitric github does claim 1.7.10 is the minimum stable git: https://help.github.com/art​icles/https-cloning-errors/
22:57 TimDix I removed and reinstalled git on the server that was being problematic, and now it's working.
22:57 TimDix So... somehow bad install? I've no explanation for why there's a discrepency in the versions.
22:57 TimDix But now both places agree with 1.
22:57 TimDix But now both places agree with 1.7.1
22:58 TimDix I wonder if I somehow got two versions of git installed on here.
22:58 MattMaker joined #git
22:59 nitric very possible, my cron environment is always grabbing my osx git at /usr/bin which is 2.5.4, while my shell git is Homebrew's 2.9.2, it does occasionally make for confusion
23:00 Gsham joined #git
23:00 mwhooker joined #git
23:02 APK joined #git
23:02 jimi_ joined #git
23:03 clemf joined #git
23:04 TimDix nitric: Well, that got the rest of the package to install properly, thank you very much for your help.
23:05 nitric you're welcome!
23:06 LeBlaaanc joined #git
23:10 romerocesar joined #git
23:10 Electrometro joined #git
23:10 Nethereal joined #git
23:12 Gsham joined #git
23:12 PabloR Anybody? git checkout <file> vs git checkout -- <file>. Is the difference between them that the 1st one modifies the working-tree’s file to match the repo’s version of that file and the 2nd one just discards all the changes in the working-tree instead?
23:13 bamb joined #git
23:13 bamb left #git
23:14 neilthereildeil joined #git
23:14 neilthereildeil hi
23:14 kadoban PabloR: The first in ambiguous. Is 'git checkout something' changing branches, or checking a file out from the index?
23:14 kadoban s/in ambiguous/is ambiguous/
23:14 neilthereildeil i want to create a branch at a few commits ago, but i also want to take a select few commits that are more recent than the branch point with me. how can i do that?
23:15 kadoban If there is no branch/tag/ref named 'something' then 'git checkout something' and 'git checkout -- something' are identical in behavior.
23:15 Gitzilla joined #git
23:17 AKPWD joined #git
23:17 PabloR kadoban I have no branches yet other than the master. I committed only once for now. (learning from a book). So after committing I made changes to the file and want to know how the 2 commands would differ
23:19 kadoban I think I just answered that, no?
23:19 LeBlaaanc joined #git
23:20 eehacker joined #git
23:20 PabloR kadoban Ah, just saw that. For what I read it seems that both modify the file, but one to match the already committed one and the 2nd one just undo changes. So that means it's the same in this case, I guess?
23:21 nitric neilthereildeil: git checkout -b your-new-branch old-commit && git cherry-pick some-commit && git cherry-pick another-commit
23:21 ffabi joined #git
23:21 neilthereildeil nitric: thanks. how is this different than git rebase?
23:21 kadoban PabloR: Reread what I said.
23:21 SwiftMatt joined #git
23:22 PabloR ok. got it. Thanks
23:22 nitric neilthereildeil: the end result would be the same, so if you can rebase, that's probably easier
23:22 neilthereildeil but rebase would not allow me to select individual commits to merge, right?
23:24 nitric well, a rebase would involve you picking which commits to keep, rather than picking which commits to merge, but the end result is the same
23:24 mablae_ joined #git
23:24 PresidentBiscuit joined #git
23:24 madewokherd joined #git
23:27 unbalancedparen joined #git
23:30 LeBlaaanc joined #git
23:30 romerocesar joined #git
23:32 PioneerAxon_ joined #git
23:32 mingrammer joined #git
23:32 MattMaker joined #git
23:38 LeBlaaanc joined #git
23:38 t-mart joined #git
23:40 MattMaker joined #git
23:41 kyan joined #git
23:41 zacts joined #git
23:47 LeBlaaanc joined #git
23:49 zzz_ joined #git
23:51 romerocesar joined #git
23:57 mwd joined #git

| Channels | #git index | Today | | Search | Google Search | Plain-Text | summary