Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-08-17

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

All times shown according to UTC.

Time Nick Message
00:01 Neico joined #git
00:06 arescorpio joined #git
00:07 qpdb joined #git
00:11 rkazak joined #git
00:13 vassagus joined #git
00:15 shgysk8zer0 joined #git
00:21 eka joined #git
00:22 fstd_ joined #git
00:23 stux joined #git
00:26 cyphase joined #git
00:26 Lipp joined #git
00:27 JamesBaxter joined #git
00:33 nullp0inter joined #git
00:39 RustyShackleford So i'm working on my own feature branch. All developers are in one repository, no forks
00:39 RustyShackleford this is a rewrite, so lots of initial development. I would like to keep my branch up to date with master, daily perhaps
00:40 RustyShackleford I also back up my work remotely often since I work from another computer
00:40 durham joined #git
00:40 RustyShackleford is there a way to do this without needed to force push?
00:41 nedbat RustyShackleford: without a force push to where?
00:41 RustyShackleford my own feature branch
00:41 kadoban RustyShackleford: You can merge master into your branch occasionally.
00:41 nedbat RustyShackleford: you could merge master into your branch. why don't you want to force push?
00:41 RustyShackleford it makes coworkers uncomfortable
00:41 kadoban Without needing to force push, I think that's really the old way.
00:42 kadoban s/old/only/
00:42 JamesBaxter joined #git
00:42 RustyShackleford I don't think its risky. Nobody else is using my branch
00:42 RustyShackleford but I'm not sure how common the rebase and force push tactic is
00:43 kadoban Quite common
00:44 nedbat RustyShackleford: i do it all the time.
00:44 nedbat RustyShackleford: but you do want to be careful that you are the only one using the branch.
00:49 bobdobbs left #git
00:51 crayon joined #git
00:54 durham_ joined #git
00:54 tesuji_ joined #git
00:54 danimal2 joined #git
00:54 kulyzu joined #git
00:55 ytl joined #git
00:55 ShekharReddy joined #git
00:55 kulyzu joined #git
00:56 xall joined #git
00:57 b1tchcak_ joined #git
00:57 rewt RustyShackleford, why do you need to force push?
01:01 Anderson69s_Pi joined #git
01:05 eka joined #git
01:06 fractalizator_ joined #git
01:06 rossome joined #git
01:07 abyx joined #git
01:08 ahr3n joined #git
01:10 rj11 joined #git
01:10 pks joined #git
01:11 clemf joined #git
01:11 d^sh joined #git
01:11 nedbat rewt: if you rebase your branch, you need to force push
01:12 zivester joined #git
01:12 rewt could just merge instead
01:14 nullp0inter i have a master and a beta branch, i work out of my home dir, but i want to be able to look at both beta and master on the web...i have two dirs /var/www/html/webapp and /var/www/html/beta how do you set these up to pull from the repo? im going to explore automatic deployment soon, just want to get some previews going
01:14 nullp0inter do i clone in there?
01:15 vassagus joined #git
01:15 rewt make a script to pull one branch and copy the files over, then pull the other branch and copy the files to the other dir
01:16 rj1 joined #git
01:16 nullp0inter rewt, do i do a git init in those dirs?
01:17 rewt no git in those dirs
01:17 rewt copy from your git dir to those dirs with cp/scp
01:18 abyx joined #git
01:18 RustyShackleford joined #git
01:20 mingrammer joined #git
01:21 Orville joined #git
01:22 a3Dman joined #git
01:24 l4v2 joined #git
01:34 clarens0000 joined #git
01:35 theskillwithin joined #git
01:43 rchavik joined #git
01:48 ilbot3 joined #git
01:48 Topic for #git is now Welcome to #git, the place for git help and doorbells | Public logs at http://goo.gl/BuUi5o | Current stable version: 2.9.3 | First visit? Read: http://jk.gs/git | Getting "cannot send to channel"? /msg gitinfo .voice | Knock knock. Who's there? Git. Git-who? Sorry, 'who' is not a git command - did you mean 'show'?
01:50 hahuang61 joined #git
01:53 Vampire0 joined #git
01:54 RustyShackleford so you guys are not against frequent rebase-and-force-pushing?
01:54 RustyShackleford seems to be a controversial topic
01:55 xall joined #git
01:56 shaggycat joined #git
01:57 SporkWitch didn't we cover this rather thoroughly last night?
01:57 settermjd joined #git
01:57 RustyShackleford SporkWitch, sure. I got different answers when I asked again though
01:58 RustyShackleford not that I'm saying you're wrong
01:58 SporkWitch not a problem, was just wondering why i fell through a time warp
01:58 RustyShackleford IIRC you suggest working in my own fork and then rebasing as I please
01:58 SporkWitch when i time warp, i like there to be tim curry around
01:59 SporkWitch yup!
01:59 SporkWitch the history that matters is the history others could potentially see BEFORE the rebase
01:59 theskillwithin joined #git
02:00 SporkWitch if it's in a private repo, there's no harm done, the worst case is they see terrible commit comments and you potentially introduce some reversions.  It's when others see it before the rebase and again after that problems can crop up, because now we have conflicting history between repos
02:00 SporkWitch s/reversions/regressions/
02:01 RustyShackleford but if each developer works in his own branch, all of which branch from master
02:01 Shakeel__ joined #git
02:01 RustyShackleford rewriting history shouldn't affect anyone but me
02:02 SporkWitch RustyShackleford: fetch grabs 'em all
02:03 RustyShackleford okay good point. So they would potentially run into issues if they check out my branch?
02:03 SporkWitch RustyShackleford: more than that
02:04 SporkWitch RustyShackleford: check the manpage: man git fetch
02:04 gitinfo RustyShackleford: the git-fetch manpage is available at http://jk.gs/git-fetch.html
02:04 cyphase joined #git
02:04 * SporkWitch loves this bot
02:06 RustyShackleford nobodies complained about a broken git repo yet
02:06 mingrammer joined #git
02:06 RustyShackleford read the manpage. What am I supposed to conclude?
02:08 SporkWitch and now the reason you've asked three times is clear: you didn't LIKE the answer lol
02:08 SporkWitch if you're just looking for the answer you wanted from the start, go into your bathroom, you've got a mirror in there most likely.
02:09 SporkWitch the manpage is fairly self explanatory, and explains all of what it pulls.  When you rebase and force-push, you're changing that history.  Now the history in the repo you pushed to doesn't match the history in anyone else's.
02:09 Rodya_ joined #git
02:14 RustyShackleford SporkWitch, got me there :p
02:14 RustyShackleford doesn't the next git fetch fix their history issue?
02:14 mingrammer joined #git
02:15 Literphor joined #git
02:16 vassagus joined #git
02:16 SporkWitch RustyShackleford: potentially, or maybe they take something you did to some file because it helps them
02:16 SporkWitch RustyShackleford: end of the day, you're rewriting history in a shared repository, and that is BAD™
02:17 RustyShackleford I think I'll get this branch merged in and then start working in a fork just to be safe
02:17 SporkWitch that would typically be the best way.  storage is cheap.
02:19 shaggycat joined #git
02:19 RustyShackleford welp, gonna consult with myself in the bathroom mirror real quick
02:19 watabou joined #git
02:20 RustyShackleford managing a fork is a pain in the butt. Not that difficult, but an extra step
02:21 SporkWitch it really isn't lol
02:22 rwswitch joined #git
02:23 SporkWitch done right it's a grand total of ONE extra step: forking the repository in the first place.  After that, your workflow wouldn't actually change or add any steps, but you WOULD be removing the risk of buggering up the history on the shared repo
02:24 chachasmooth joined #git
02:25 RustyShackleford i've screwed up by confusing the two repos. Granted, it was my fault for having them both on my computer
02:25 RustyShackleford both open in a terminal window
02:26 nettoweb joined #git
02:26 SporkWitch well, that's a problem with the main repo; it shouldn't even be ALLOWING you to force-push to it, in a sane configuration
02:27 RustyShackleford nothing is sane where I work hah
02:27 RustyShackleford we're all pretty new to git, recent migration. I've used it before, but on my own projects not in a team
02:27 SporkWitch but that's about the only thing that would change: you push to origin (which should be your repo, not the one you forked FROM), and you submit pull requests to upstream (the one you forked from)
02:28 SporkWitch you don't push upstream, you ask upstream to pull from you; end result is the same, but the agency is different, and that difference is essential for maintaining integrity and preventing issues (like accidentally damaging the upstream because you pushed to the wrong place)
02:29 rwswitch joined #git
02:30 RustyShackleford do pretty much all git servers allow for pull requests and forks?
02:31 RustyShackleford github and bitbucket do
02:32 SporkWitch RustyShackleford: entirely dependent on the settings of the system
02:33 SporkWitch RustyShackleford: github private repos require auth to clone (which is all a fork is, initially; formal "forking" in things like github and bitbucket is merely a tracking tool to maintain the relations and allow for easy visibility and collaboration, within those that are authorized to view it)
02:34 nullp0inter joined #git
02:35 cdown joined #git
02:35 lagothrix joined #git
02:37 durham joined #git
02:39 rads joined #git
02:40 hexagoxel joined #git
02:41 tgunr joined #git
02:43 atomicb0mb joined #git
02:46 Goplat joined #git
02:46 darkbit joined #git
02:48 dodobrain joined #git
02:48 moei joined #git
02:49 snowkidind joined #git
02:51 skylite joined #git
02:52 crayon joined #git
02:59 Dougie187 left #git
03:00 rwswitch joined #git
03:00 aavrug joined #git
03:01 aavrug joined #git
03:02 dodobrain joined #git
03:02 aavrug joined #git
03:04 fstd_ joined #git
03:06 boombatower joined #git
03:07 dhollinger joined #git
03:07 nullp0inter left #git
03:09 nunchuck joined #git
03:12 cyphase joined #git
03:14 b1tchcak_ joined #git
03:14 clemf joined #git
03:16 vassagus joined #git
03:20 OrvilleWright joined #git
03:21 rossome joined #git
03:22 hahuang61 joined #git
03:22 annoymouse joined #git
03:26 mehola joined #git
03:26 Pinkamena_D joined #git
03:28 b1tchcakes joined #git
03:31 b1tchcak_ joined #git
03:36 Literphor joined #git
03:38 xall joined #git
03:40 texinwien_ joined #git
03:42 JanC_ joined #git
03:44 robogoat joined #git
03:45 rwswitch joined #git
03:46 mehola_ joined #git
03:47 The_Phoenix joined #git
03:48 The_Phoenix joined #git
03:48 Lipp joined #git
03:49 The_Phoenix joined #git
03:51 The_Phoenix joined #git
03:54 rwswitch joined #git
03:55 Derperperd joined #git
03:55 chandu20 joined #git
03:56 scott_sd joined #git
03:58 bayman joined #git
03:59 GT4066 joined #git
04:00 thecomedian joined #git
04:03 clemf joined #git
04:07 johnny56 joined #git
04:07 Andrew_K joined #git
04:11 netj joined #git
04:13 rkazak joined #git
04:14 dsantiago joined #git
04:17 vassagus joined #git
04:18 _sjs joined #git
04:21 Cabanossi joined #git
04:23 snowkidind joined #git
04:24 nettoweb joined #git
04:27 ljc joined #git
04:27 Pinkamena_D joined #git
04:30 realz joined #git
04:38 MrWoohoo joined #git
04:41 bayman left #git
04:47 ljc joined #git
04:54 dr_lepper joined #git
04:57 subhojit777 joined #git
04:58 xall joined #git
04:59 YourTheBombNow23 joined #git
05:00 cqi joined #git
05:04 eivarv joined #git
05:06 kutsuya joined #git
05:07 tjayanth joined #git
05:09 Lipp joined #git
05:20 vassagus joined #git
05:21 chandu20 joined #git
05:23 hahuang61 joined #git
05:25 nunchuck joined #git
05:26 leeN joined #git
05:28 jrobeson joined #git
05:28 Kaisyu joined #git
05:30 chachasmooth joined #git
05:33 OrvilleWright left #git
05:33 diogenese joined #git
05:34 Khisanth joined #git
05:35 jrobeson left #git
05:36 thiago joined #git
05:38 nidr0x joined #git
05:39 zeroed joined #git
05:40 circ-user-RNfpp joined #git
05:49 mariuscc joined #git
05:49 eivarv joined #git
05:52 Raging_Hog joined #git
05:54 nunchuck joined #git
05:56 _sjs joined #git
05:56 bocaneri joined #git
05:57 settermjd joined #git
05:58 Manuel15 joined #git
05:58 Manuel15 joined #git
05:59 rscata joined #git
06:00 kappy joined #git
06:03 LucaTM joined #git
06:03 satya4ever joined #git
06:05 dodobrain joined #git
06:07 TomyLobo joined #git
06:09 tgunr joined #git
06:10 jceb joined #git
06:12 tgunr joined #git
06:13 roelmonnens joined #git
06:16 mehola joined #git
06:18 tgunr joined #git
06:20 govg joined #git
06:21 vassagus joined #git
06:25 tgunr joined #git
06:26 alansaul joined #git
06:28 qt-x joined #git
06:34 Lipp joined #git
06:34 jknetl joined #git
06:37 th3Dotor joined #git
06:38 freimatz joined #git
06:38 Jibz joined #git
06:41 inflames joined #git
06:43 Jibz I'm on Windows, and I've converted a mercurial repo to git (hg-fast-export), and did a rebase to reword most of the commit messages. Now I wish to set the commit date back to the author date, but `git rebase --committer-date-is-author-date --root` appears to set the commit date to now. Anything I am doing wrong?
06:46 Sceorem joined #git
06:49 Jibz hmm looks like --root implies -i (for no apparent reason), so --committer-date-is-author-date silently does nothing .. sigh
06:50 tgunr joined #git
06:56 freddymungo joined #git
06:57 jstein_ joined #git
06:57 submitnine joined #git
06:58 rkazak joined #git
06:59 jeffreylevesque_ joined #git
07:04 niranjan joined #git
07:09 MainB joined #git
07:13 jguillen joined #git
07:14 alansaul joined #git
07:14 chigang_ joined #git
07:16 freddymungo joined #git
07:17 t0by joined #git
07:18 styler2go joined #git
07:21 ome joined #git
07:22 vassagus joined #git
07:22 xaa joined #git
07:23 hahuang61 joined #git
07:24 saedelaere joined #git
07:25 Freeze16 left #git
07:26 Balliad joined #git
07:29 vuoto joined #git
07:30 Lomex joined #git
07:30 sddwdw joined #git
07:31 sddwdw How do I remove the latest commit from master?
07:31 mingrammer joined #git
07:32 selckin rewrite history so it never happened remove, or revert by commiting the inverse?
07:33 Impaloo joined #git
07:33 mingrammer joined #git
07:36 16WAA1H2P joined #git
07:36 Guest93616 joined #git
07:37 ki0 joined #git
07:39 spudowiar joined #git
07:40 Xin joined #git
07:41 Xin hey all im just getting used to git command line lol
07:41 Xin we want to track everyones stuff so we can see who is doing what using git remotes
07:41 Xin so we want to do something like a dev comes in to program for the day
07:41 ahmedelgabri joined #git
07:42 Xin they clone/fork the current dev branch from a public project
07:42 ljc joined #git
07:42 Xin now I have say 10 users with clones/forks...
07:42 Xin how do I get all their individual changes into the public dev branch
07:42 th3Dotor joined #git
07:43 Xin like what do they do at the end of their sessions?
07:43 alansaul joined #git
07:44 selckin they go home
07:44 Xin lol but how do they merge it in?
07:44 Xin and how do I allow them access to do so?
07:44 selckin they work on features/changes and create logical commits for those, and when they are ready to get usually get pushed/merged into the central repo
07:45 Xin so how do they push that to a public git though without say github
07:46 selckin by typing git push
07:46 Xin or can I allow them to push to github somehow?
07:46 selckin !github
07:46 gitinfo Note that git != github. Feel free to ask us about Github-specific features (Forks, Pull Requests, Wikis, etc), but there are no guarantees. There is a #github channel, which might help too (again, no guarantees)
07:46 Xin right im asking how to do it without github
07:46 Xin -_-
07:46 selckin standard feature
07:46 selckin notting special
07:46 Impaloo !man git-push
07:46 gitinfo The git man pages are available online at http://jk.gs/git.html. Or were you looking for the "man git-foo" syntax(without the !) ?
07:46 gitinfo the git-push manpage is available at http://jk.gs/git-push.html
07:47 Xin they do something like git push git:git@originhost
07:47 Xin right?
07:47 selckin !man git-remote
07:47 gitinfo The git man pages are available online at http://jk.gs/git.html. Or were you looking for the "man git-foo" syntax(without the !) ?
07:47 gitinfo the git-remote manpage is available at http://jk.gs/git-remote.html
07:51 Lipp joined #git
07:53 Xin yep so I add the correct remote origin and branch details
07:53 Xin and if they push over ssh then they just need a user account on the server eh?
07:54 selckin !gitolite
07:54 gitinfo 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/
07:54 selckin but yes
07:55 Xin yeah im not sure I need anything like that, this is great
07:55 Xin thanks for your advice
07:57 nikivi joined #git
07:57 kurkale6ka joined #git
07:57 rchavik_ joined #git
07:57 armin i'm just writing a script to create/delete/list repositories on a remote server accessible via ssh because i got too annyoed by gitolite actually.
07:57 armin shoot me.
07:57 selckin gitolite is awesome, you're the problem
07:58 armin selckin: i've actually used it for a long time.
07:58 mehola_ joined #git
07:58 selckin maybe it doesn't match your usecase, but its still awesome
07:59 KeelOfSteel joined #git
07:59 sin3t joined #git
07:59 Xin left #git
07:59 EduardoMartins joined #git
08:00 armin selckin: it surely is awesome. that doesn't mean i could implement something like that on my own, though, so let's try. i'm not even saying i won't fail at trying to do so. :)
08:01 armin selckin: i used it. for a long time. probably what annoyed me most was debian's default, but i don't have other distributions or operating systems to compare with.
08:01 mohabaks joined #git
08:04 dreiss joined #git
08:07 TomyWork joined #git
08:08 Oebele joined #git
08:11 ki0_ joined #git
08:11 ki0_ joined #git
08:12 bongjovi joined #git
08:13 tomog999 joined #git
08:14 sitaram armin: https://paste.fedoraproject.org/409478/14714216 is some ~/.gitconfig aliases I use when even I feel gitolite is too heavy :-)
08:15 sitaram but honestly, gitolite with a simple wild card rule would be the same end-result, but more flexible if things change later
08:16 slugmandrew joined #git
08:17 nikivi joined #git
08:17 Phlogistique left #git
08:17 sitaram armin: what was the "debian default" issue?  I'm interested in details (even if I don't use debian)
08:17 WayToDoor joined #git
08:18 hjubal joined #git
08:21 jceb joined #git
08:22 robotroll joined #git
08:22 infra-red joined #git
08:22 vassagus joined #git
08:23 lynkd joined #git
08:25 Guest93616 joined #git
08:29 Virox joined #git
08:32 chll_ joined #git
08:34 fakenerd joined #git
08:34 ahmed_elgabri joined #git
08:34 maaarghk joined #git
08:36 maaarghk Hi, quick question - I want to do a sparse checkout of a single folder. Is it possible to effectively chroot the working copy to that single folder?
08:36 venmx joined #git
08:37 ams__ joined #git
08:37 maaarghk It would be as if the following were possible; `git checkout https://github.com/git/git/Documentation`
08:37 slixpk joined #git
08:37 selckin no
08:37 lb maaarghk sparse checkout? you could play with die env variable GIT_DIR and then do the normal git checkout -- /path/to/folder
08:38 selckin checkout or clone?
08:38 alansaul joined #git
08:38 lb ahh right, selckin spotted the key question. checkout works only with local clones...
08:39 maaarghk sorry
08:39 maaarghk my brain is back to front
08:39 maaarghk i meant as if `git clone https...` were possible
08:40 selckin not possible
08:40 maaarghk alright
08:40 maaarghk guess ill use symlinks as a workaround
08:40 lb maaarghk no that's not possible. you can only clone with the --depth parameter used, to limit the cloned amount of data
08:41 lb man git clone
08:41 gitinfo the git-clone manpage is available at http://jk.gs/git-clone.html
08:41 maaarghk yeah that's as in depth of history right?
08:41 fate_ joined #git
08:41 lb yes
08:41 lb i don't recommend that though ^^
08:41 maaarghk yeah it's not what i'm after anyway
08:41 lb but that's more my guts feeling. i don't have a technical reason for it
08:42 lb great \o/
08:42 lb !botsnack
08:42 gitinfo Om nom nom
08:44 maaarghk i think i was probably looking for a --relative paramete ron git read-tree
08:45 armin sitaram: hi. i found "gitolite3" to be not the user name i would want on a git server as i like simple user- and host-names, so i ended up changing a lot of those defaults for the user-name to be "git". also what's actually happening is just a wrapper that is being called from authorized_keys as command= which reads $SSH_ORIGINAL_COMMAND so i considered i could roll my own.
08:48 lb armin there is #gitolite for gitolite specific questions
08:49 armin lb: that's quite awesome. i don't have one, though.
08:49 cyan__ joined #git
08:51 roelmonnens joined #git
08:53 Guest93616 joined #git
08:54 davisonio joined #git
08:55 _sjs joined #git
08:55 Shentino joined #git
08:57 sitaram armin: yeah that maybe a debian default.  But I suspect, after you install it, you can simply ignore "gitolite3" and start *using* gitolite on some other userid (most people prefer "git").  Gitolite is designed to allow any Unix user to be used as a "hosting user"
08:57 sitaram (and "designed to allow" sounds grand but it's actually ssh that's doing it so, no credit to me!)
08:57 WayToDoor joined #git
09:05 notebox joined #git
09:05 chll_ joined #git
09:06 eijk_ joined #git
09:06 eijk joined #git
09:09 Anticom joined #git
09:11 mahendra_ joined #git
09:12 ki0 joined #git
09:12 ki0 joined #git
09:13 ki0 joined #git
09:14 fakenerd joined #git
09:15 fakenerd joined #git
09:16 Lipp joined #git
09:19 Silenced joined #git
09:19 darkbit joined #git
09:21 Silenced joined #git
09:21 Thorn joined #git
09:23 Guest93616 joined #git
09:23 vassagus joined #git
09:24 hahuang61 joined #git
09:25 GenZai joined #git
09:25 _W_ joined #git
09:26 sylario joined #git
09:28 bouncy joined #git
09:29 dmpeters63 joined #git
09:31 fakenerd joined #git
09:39 eijk_ joined #git
09:39 mSSM joined #git
09:41 mSSM I have a branch “bar” off of “master” with commits a, b, c; is it possible to move a, b to a new branch “foo”, and have bar now branch off of foo?
09:41 mSSM Or is that too insane?
09:44 mSSM (reason: I want to keep c in bar, because I want to keep a PR on github associated with it...)
09:45 Rish joined #git
09:55 fuchstronaut joined #git
09:55 Guest93616 joined #git
09:56 _ikke_ mSSM: rebasing can help with that
09:56 _ikke_ !fixup
09:56 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!
09:56 jast mSSM: you don't actually need to do any moving, because 'branching off' isn't recorded in git
09:56 jast if you create 'foo' and have it point at b, the overall outcome is the same as though you'd created 'bar' based on 'foo' at b
09:57 jast so basically all you need to do is this: git branch foo b
09:58 jast all that assumes that you committed in this chronological order: a first, then b, then c
09:59 mSSM jast: Ah, that makes sense.
10:02 rezrov joined #git
10:02 djb-irc joined #git
10:03 phanimahesh joined #git
10:05 Silenced_v2 joined #git
10:05 Silenced_v2 joined #git
10:06 ix007 joined #git
10:06 _Patrick_ joined #git
10:08 fozniku joined #git
10:12 stamina joined #git
10:13 inflames joined #git
10:13 gechr joined #git
10:15 Manuel15 joined #git
10:15 Manuel15 joined #git
10:16 mingrammer joined #git
10:17 ASOLAr joined #git
10:18 aguimaraes1986 joined #git
10:21 jceb joined #git
10:23 sitaram armin: sorry was called away... to continue... "just a wrapper [that] reads $SSH_ORIGINAL_COMMAND" is kinda missing the point.  *Any* access control system can be (modulo how the OS/app/whatever actually wants to see it) thought of as a wrapper that just reads the **request** being made (which is what the SSH_ORIGINAL_COMMAND has)
10:23 weckl joined #git
10:24 vassagus joined #git
10:27 Silenced joined #git
10:27 Silenced joined #git
10:28 ifra201 joined #git
10:32 dmpeters63 joined #git
10:37 ToBeCloud joined #git
10:41 chid joined #git
10:47 a3Dman joined #git
10:50 lizzin joined #git
10:50 eka joined #git
10:50 lizzin ive installed git-lfs and i've tracked the large file, but i still get the large file detected error when i try to push. what might i be missing here?
10:52 armin sitaram: and the point is?
10:52 armin sitaram: don't get me wrong, i love gitolite. i just consider it way too complex for most my use cases already.
10:53 armin sitaram: also i love writing shellscripts.
10:54 a3Dman joined #git
10:55 Thorn__ joined #git
10:55 xall joined #git
10:56 Hounddog joined #git
10:57 Hounddog Good Morning, i need to ignore all files with in a directory but need to have certain directories added... how would i accomplish this? I could show what i require
10:58 Lipp joined #git
10:58 ronny joined #git
10:58 ronny hi
10:58 ronny i have a branch i want ro resuffle with interactive merge without moving it to the head its relative to at the same time
10:58 ronny how would i do that
10:58 selckin Hounddog: you can 'include' with '!some/path'
10:59 mostlybadfly joined #git
11:00 armin sitaram: actually i'm hacking on this since 2 hours now and got a fully working way to create/delete/list repositories from remote already, 100% customizable because i know 100% of the code that i wrote. i think that's fairly great and hope you won't consider this offending or as criticism against gitolite.
11:00 shams joined #git
11:00 diogenese joined #git
11:01 Hounddog selckin: ahh that might be something  http://pastebin.com/F5WWn8t0 this was what i need but without any files in it
11:01 cory_ joined #git
11:01 felowe joined #git
11:01 joseph_arnstein joined #git
11:02 selckin Hounddog: git doesn't track directories
11:03 Hounddog selckin: i know but i can add a gitignore in there
11:04 selckin i have no idea what you mean anymore
11:04 Hounddog inside the directory add .gitignore but ignore everything else
11:04 Lipp joined #git
11:04 Hounddog that way i can track the directory
11:04 Lipp joined #git
11:05 selckin sure
11:05 selckin or you can fix your shit the create the directories it needs on startup
11:05 selckin you can force add files that are ignored, so can have any random file in there
11:06 Hounddog Well to answer in your rude language, That shit is already setup and i am fixing this shit so people who have not  a real clue of git can also use this as easily as possible
11:06 j416 there there, calm down children
11:06 _sjs joined #git
11:07 ljc joined #git
11:08 j416 ronny: I don't quite understand what you mean
11:08 sitaram armin: "and the point is"?  Reading the *request* is not even half the deal
11:08 j416 ronny: you want to change the ordering of commits, yes?
11:10 sitaram armin: are you doing this for a single user?  It sounds like it if you've got it done with a few shell scripts in 2 hours.  (Or I bow my head in reverence at your superior hacking skills)
11:10 peib joined #git
11:11 ronny j416: i havea branch i cant rebase to master without conflict that i need to reorder first
11:11 sitaram armin: hmm wait I just noticed "way too complex for my use cases"; never mind
11:11 j416 ronny: alright.
11:11 ronny also - is there a easy way to split the commits in one branch into 2 distinct branches?
11:11 j416 ronny: one question at a time, please
11:11 ronny ok, lets go for reorder first
11:11 j416 ronny: I have not even understood your first question yet.
11:12 armin sitaram: actually, yes. for a single user. exactly. my use case is: i have ssh access to some host. i want to host git repositories on that host easily. i'm eventually not root there. just store my damn git repos.
11:12 boombatower joined #git
11:13 sitaram armin: for that, I use 3 "~/.gitconfig" aliases.  I pasted a link to a pastebin sometime ago.  I'm pretty sure they do most of what you want :)
11:13 armin sitaram: would you mind grabbing me a link to it?
11:13 sitaram armin: https://paste.fedoraproject.org/409512/47143243
11:14 sitaram (just fpaste-d again; not the same link as before but same content)
11:14 armin sitaram: almost what i came up with...
11:14 sitaram I kinda guessed :)
11:16 ronny j416: i have a branch, that has not been rebased to master, and i want to reorder its commits without rebasing over to master
11:17 HoierM joined #git
11:18 armin sitaram: http://core.m2m.pm/gitctl.txt (beware, work in progress, considered harmful and buggy until proven otherwise)
11:19 skylite joined #git
11:19 armin sitaram: basically i was just playing around with that but will probably use that on an every day basis soon probably.
11:20 selckin set -e
11:20 j416 ronny: just interactively rebase the branch onto a commit earlier in the branch.
11:20 mingrammer joined #git
11:20 ronny j416: whats the canonical way to rebase a branch onto its common ancestor with master?
11:22 j416 ronny: you can say: git rebase $(git merge-base master HEAD)
11:22 j416 (add -i for interactive)
11:22 ronny thanks :)
11:22 omegak joined #git
11:22 j416 (otherwise this is a no-op)
11:22 j416 might be able to do it with dot syntax, I always forget so I use the above
11:23 d0nn1e joined #git
11:23 zeroed joined #git
11:23 zeroed joined #git
11:24 afuentes joined #git
11:25 vassagus joined #git
11:25 cyan__ joined #git
11:27 sitaram armin: looks ok on first glance but if it's really meant for yourself it's too verbose.  Plus you really need to check/enforce fundamental things like ".git" at the end.  And speaking for myself, I never want to delete remotely; it has to be "log on and do it manually"
11:27 armin sitaram: good points. thank you!
11:27 lizzin left #git
11:29 armin sitaram: it's still perfectly valid to do "git init --bare foobar" (without the .git at the end), though. does the ".git" suffix mean something more than "this is a directory that will likely have a .git folder inside"?
11:31 jast armin: it means the opposite, actually. convention says that foo.git contains what would normally be inside the '.git' subdir in a normal (non-bare) git repo
11:32 sitaram armin: ".git" at the end for bare repos is only a convention but a good one to follow; no hard and fast rules enforced by git itself.  Sorry if I made it sound like that
11:32 sitaram ".git folder inside" means it is NOT a bare repo, then it (again, convention) should NOT have a directory name ending in ".git"
11:33 sitaram or... what jast said (I should have read that before I posted!)
11:33 * sitaram blames irssi for highlighting messages containing "sitaram" or "gitolite" in red and making me react to them first, in some Pavlovian manner
11:33 armin sitaram: you didn't - i'm just trying to understand things. all good. :)
11:34 mycyber joined #git
11:34 ronny hmm
11:34 ronny anyone aware of a tool that would take a style fixup commit and then integrated it to n previous commits that did the actual changes so that hte result is a style fixed history?
11:34 selckin just curious what do you do that has you creating git repos this often?
11:36 selckin ronny: you can use rebase -i to merge fixup commits into older commits or i don't understand what you mean
11:40 ronny selckin: i have one fixup commit and i want to split it over to 5 older commits since each of the older commits touches different bits that the fixup commit fixes up
11:41 selckin you wanna create 5 fixup commits then, (git add -p ) is big help
11:41 selckin or manual edit
11:46 armin selckin: did that question relate to me?
11:46 selckin yeah
11:47 ronny selckin: the commit is already done now i want to automatically split it
11:47 selckin ronny: pick edit, unstage, and use git add -p to stage the hunks how you want to split them and commit each part
11:48 armin selckin: since you consider the problem to be myself, my motivation to answer this in an honest way is quite low, but i'll try: i like to enhance my knowledge and am bored at work and script things just for fun.
11:48 G1eb joined #git
11:49 selckin armin: i don't consider anything a problem, just can't think of a reason where this would be needed, and you don't need it so i wasn't that wrong, but learning is always a good excuse
11:49 G1eb heya, I've accidentally removed a file, how do I put it back with git?
11:49 G1eb I've tried git reset HEAD <file> but that does not do anything
11:49 selckin G1eb: git checkout <sha> -- file
11:50 G1eb ah, yes thats it
11:50 G1eb thanks! :)
11:50 G1eb left #git
11:51 selckin G1eb: the reset will unstage it, and he left already
11:51 jast ronny: I don't think you can do this automatically :(
11:51 ronny sometimes i wish i could just drag&drop comits/hunks in a ui
11:51 selckin ronny: git gui
11:52 ronny selckin: can that drag&drop patch hunks/commits?
11:52 jast not exactly, from what I recall
11:52 selckin you can stage hunks in the gui then commit
11:52 selckin then rebase and merge
11:53 jast actually in this specific case a magic special rebase with -Xours (or something?) might work
11:53 armin selckin: 09:58:03 selckin gitolite is awesome, you're the problem
11:53 jast but that would be a rather special way of doing it :)
11:53 armin well...
11:53 selckin armin: don't take everything so seriously, you bashed it for a bullshit reason so i countered with that
11:54 sdothum joined #git
11:55 jast maybe I missed something, but the annoyance seems to be with debian's defaults for gitolite and not gitolite itself?
11:56 selckin at taht point he just said he got "too annoyed" by it
11:56 jast if so, I can empathize with that. I've often been dissatisfied with how git-related packages are handled in debian
11:56 subhojit777 Is there any way to bypass resolve conflict using mergetool for specific files. My case - I usually get merge conflicts for compiled + minified CSS files, without resoolving the conflict I can easily autogenerate the CSS by  running compass.
11:56 Seveas subhojit777: don't commit the generated files.
11:57 selckin subhojit777: in general you don't commit those files
11:57 synthroid joined #git
11:57 jast subhojit777: there are attributes you can set on files to customize how merging is handled, see man gitattributes
11:57 gitinfo subhojit777: the gitattributes manpage is available at http://jk.gs/gitattributes.html
11:58 jast also, what the other people said :)
11:58 subhojit777 Thanks, I will try to suggest  this to my superior. But, is there any other way?  Like I  will manually mark those files as resolved.
11:58 bartosz_ joined #git
11:59 selckin what jast said
12:00 subhojit777 jast, "other people"? sorry didn't get you? do you mean Google for the problem?
12:00 selckin line above that
12:01 jast no, I meant the responsess from selckin and Seveas :)
12:01 subhojit777 ohkk..  Thanks :)
12:02 styler2go Hello. I have a problem with git. My git project is on a samba drive and i laways have file permission differences and thus all files are marked as changed. config.fileMode false did not help. any idea?
12:03 selckin samba confnig issue
12:05 Sceorem joined #git
12:05 jast it's core.fileMode, actually :)
12:05 saedelaere joined #git
12:06 Seveas styler2go: don't pubt repos on samba drives.
12:06 selckin you can configure samba not todo that
12:06 styler2go hmm.. couldn't find any option in samba
12:06 styler2go jast: yes, tried that
12:06 selckin enable unix extensions or disable something or the other, been a few years
12:07 selckin i'm sure google knows
12:08 ASOLAr joined #git
12:09 jast of course UNIX extensions don't help if the share is accessed by windows clients
12:09 selckin can do it with that too
12:10 selckin *without
12:10 jast you can tell your client to disable +x flags for all files, then you'll only run into trouble for files that are *supposed* to be executable :)
12:11 styler2go hmm
12:11 Maple__ joined #git
12:11 styler2go it always says old 0755 and new 0644
12:12 YourTheBombNow23 joined #git
12:12 ronny hmm
12:12 ronny got cola seems halfway there
12:13 styler2go tried a lot of things now... nothign worked sadly
12:14 snowkidind joined #git
12:14 nettoweb joined #git
12:16 rwb joined #git
12:16 garethdaine joined #git
12:19 Diabolik joined #git
12:19 Lipp joined #git
12:19 Diabolik joined #git
12:20 Diabolik joined #git
12:20 Diabolik joined #git
12:21 hahuang61 joined #git
12:22 johnmilton joined #git
12:22 dsdeiz joined #git
12:26 fees joined #git
12:26 vassagus joined #git
12:27 tvw joined #git
12:28 jimi_ joined #git
12:29 The_Phoenix joined #git
12:29 Pinkamena_D joined #git
12:31 kim_bruning joined #git
12:32 shams joined #git
12:33 TooLmaN joined #git
12:33 johnny56_ joined #git
12:33 kim_bruning Hello, so once upon a time I was clever, and decided to do git init in some of my /etc subdirs (like /etc/asterisk and /etc/samba).   Today, I am cleverer, and realize I really would like to just have all of /etc in git ;-) Other than deleting my old git history, what's a smart way to git init in /etc, and include subdirs properly?
12:34 tobiasvl that sounds like a strange idea
12:34 selckin !etc
12:34 gitinfo [!fetch] When you work with remote repositories, Git stores copies of the remote's branches in !tracking_branches (basically mirrors). You can use 'git fetch' to update those. If you want to actually apply changes from the remote to your checked out branch, a second step is needed, usually 'git merge' or 'git rebase'. There's also 'git pull' which combines both steps.
12:34 selckin wrong one
12:35 selckin !etckeeper
12:35 gitinfo etckeeper is a collection of tools to let /etc be stored in a git, mercurial, darcs, or bzr repository. It hooks into various package managers. http://etckeeper.branchable.com
12:35 tobiasvl well there you go
12:35 tobiasvl you learn something every day
12:35 zivester joined #git
12:36 selckin !subtree
12:36 gitinfo The subtree merge method is great for incorporating a subsidiary git repo into your current one with "unified" history. Read http://git-scm.com/book/en/v2/Git-To​ols-Advanced-Merging#_subtree_merge for more info, see also !git-subtree and !git-stitch-repo.
12:36 kim_bruning tobiasvl: keeping /etc in git is sheer genius I tell you.
12:36 tobiasvl apparently!
12:37 kim_bruning selckin : is git going to be confuzzeled because the ostensible subtree is a subdir of the new repo?
12:38 kim_bruning also worried about git add *
12:38 * kim_bruning scratches head
12:38 tvw joined #git
12:38 selckin not sure what you mean, can't do much wrong tho
12:39 kim_bruning well, okay then
12:39 Hko joined #git
12:39 kim_bruning I should probably practice on something none-etc on none-production first :-P
12:39 ShekharReddy joined #git
12:40 kim_bruning Oh! And I should really have the main repo for etc on my backup box (so I can git push and have a sort of backup already)
12:40 johnmilton joined #git
12:41 vaibhavsagar joined #git
12:41 selckin need a backup for the rest of the system anyway that includes /etc, the git is just nice for history
12:42 vaibhavsagar hi, can anyone explain how to convert an integer into the offset encoded representation?
12:42 bremner uhm. Connection with git?
12:42 kim_bruning selckin: ofcourse. this way just means more granular backups for that particular dir
12:42 vaibhavsagar I can parse it just fine, but I don't understand how to serialize it
12:42 vaibhavsagar bremner: https://github.com/git/git/blob/master/Docum​entation/technical/pack-format.txt#L119-L124
12:43 kim_bruning selckin: *in addition to* the regular ;-)
12:43 mycyber joined #git
12:43 bremner vaibhavsagar: oh, ok.
12:43 bremner vaibhavsagar: you might need to try the list for internals questions
12:43 f0rblaze joined #git
12:43 vaibhavsagar sure, thanks bremner
12:45 JamesBaxter joined #git
12:45 vaibhavsagar is that just the git mailing list?
12:47 fmcgeough joined #git
12:47 notebox joined #git
12:50 davisonio joined #git
12:50 sanketdg joined #git
12:51 m0viefreak joined #git
12:52 CuriousMind joined #git
12:52 kim_bruning ok so to simulate my fun situation I made a dir test   and a dir   test/foo.d    , I did git init in test/foo.d   , then git init in test ... so far so bad ( I now have 2 different logs ;-) )
12:52 rkazak joined #git
12:52 kim_bruning let's see if I can do the merge
12:53 CheckDavid joined #git
12:53 CuriousMind My question is, if I have more than one remote, will git or whatever push my work to both repos? I'm trying to deploy my app on heroku http://xomf.com/cmzlx
12:53 CuriousMind That's a screenie of the remotes I have
12:54 kim_bruning $ git read-tree --prefix=foo.d/ -u foo fatal: Not a valid object name foo
12:54 kim_bruning ok, clearly I actually need to grok what I'm doing ... magical invocations won't work without that ^^;;
12:54 SporkWitch CuriousMind: you can script it, but no, normally you push to one at a time
12:55 selckin CuriousMind: if ust do 'git push' it will push to the remote tracking branch (not sure if you can have multiple), but you can always push explicitly to both git push origin; git push heroku;
12:56 CuriousMind ohh that's how I would do that
12:56 CuriousMind Oh right you push to one at a time, that makes sense
12:57 CuriousMind Ok so I could still push to heroku, I would have to do I don't know "git push heroku master"?
12:57 selckin sure
12:58 netj joined #git
12:58 MadHatter42 joined #git
12:59 Pinkamena_D joined #git
13:00 kim_bruning ok, you know what... I don't think I have too many commits in the /etc subdirs... I'm just gonna cp -a out those subdirs just in case, and then apply etckeeper
13:00 SporkWitch selckin: you can't (re: tracking multiple).  Though some tools will cheat and do it anyway, to show your relative position to your listed remotes
13:00 chid_ joined #git
13:00 kim_bruning selckin: neat that etckeeper is in the debian repo! :-)
13:01 m1dnight_ joined #git
13:02 CuriousMind Thanks SporkWitch selckin
13:03 CuriousMind I'm trying to deploy my app to heroku
13:03 armin kim_bruning: best thing is that it actually works out of the box, with apt-integration.
13:03 MadHatter42 joined #git
13:04 SporkWitch etckeeper?
13:04 selckin yeah commits after upgrades etc
13:04 pcre_ joined #git
13:04 alansaul joined #git
13:07 talin hello. i need to compare two files on different branches and in different dirs
13:07 talin i tried git diff branch1:file branch2:file, but it claims that it can't find the first file
13:07 polyzen talin, try omitting branch1:
13:09 garethdaine joined #git
13:09 cdg joined #git
13:09 polyzen by the lack of a response, I can only assume Git has crashed your system
13:10 cdg joined #git
13:10 ASOLAr joined #git
13:10 talin no such path in the working tree
13:12 talin if i add -- between them, i get some output
13:12 talin not sure if it's correct
13:13 cyan__ joined #git
13:13 rahtgaz joined #git
13:13 Peuc joined #git
13:15 RustyShackleford joined #git
13:18 bsanford joined #git
13:18 SporkWitch selckin: ok, that tool sounds pretty sweet lol
13:18 _sjs joined #git
13:18 al-damiri joined #git
13:20 jstein Hi, I read a discussion about version tagging recently, but do not understand what changed in git in the past. https://github.com/smxi/inxi/issues/92 Do you know, what was bad about tags in the past and changed?
13:20 fahadash joined #git
13:21 rgrinberg joined #git
13:22 ws2k3 joined #git
13:24 kim_bruning is etckeeper init supposed to be slow?
13:24 stux|away joined #git
13:25 kim_bruning and git status?
13:25 kim_bruning oh boy
13:26 zeroed joined #git
13:26 zeroed joined #git
13:27 vassagus joined #git
13:27 jast kim_bruning: with huge directories it can be, though if you have enough RAM the filesystem cache should be able to take the edge off it after the first time
13:27 netj joined #git
13:27 jstein kim_bruning: no, it should not be slow. How long did it take?
13:27 kim_bruning 45 megs
13:27 kim_bruning I stopped it after a minute
13:27 jast you have to keep in mind that 'git status' has to look at the metadata of every single file in the directory
13:27 kim_bruning suspect a symlink loop or something
13:28 kim_bruning ok, there be a lot of files in that dir ;-)
13:28 jast the size of the files shouldn't matter as much. if anything, the number of files does.
13:28 kim_bruning 7798
13:28 ExoUNX joined #git
13:28 kim_bruning can't be that bad
13:28 jast still shouldn't take a minute unless you've got a tape recorder as your storage system
13:28 jstein kim_bruning: that is a lot, but no problem for git
13:28 jast or I/O saturation from unrelated processes
13:29 jast (especially if it's not an SSD)
13:29 kim_bruning hum, running etckeeper init twice is probably the wrong move :-P
13:29 kim_bruning got an ssd and 8 gigs of ram. Not fastest, but definitely not slowest :-P
13:30 kim_bruning anyway, this time it was near-instant
13:30 selckin time for a backup, your ssd is about to die
13:30 kim_bruning apparantly apt-get had run etckeeper init for me already
13:30 kim_bruning Ok, let's set up for a remote push
13:32 ash_workz joined #git
13:35 kim_bruning ok, let's push this thing
13:36 sangy joined #git
13:36 kaldoran joined #git
13:39 romerocesar joined #git
13:39 kadoban joined #git
13:41 Shentino joined #git
13:42 kim_bruning hum, for some reason git doesn't accept a remote like ssh://rbackup@pi:myserver-backup
13:42 kim_bruning it has to be   ssh://rbackup@pi/home/rbackup/myserver-backup
13:43 kim_bruning am i Doing It Wrong(tm) ?
13:43 ASOLAr joined #git
13:44 kim_bruning also, now I'm doing git push backup .... but wasn't there a way to make a default for git push
13:44 kadoban kim_bruning: I always specify SSH URLs as just   me@whatever:/some/path   usually leaving off the me@ part as well, since I set that stuff in the alias config in SSH
13:45 jeffreylevesque joined #git
13:45 mycyber joined #git
13:45 cqi joined #git
13:45 fairuz joined #git
13:47 shams joined #git
13:49 AaronMT joined #git
13:51 AaronMT joined #git
13:51 jimi_ joined #git
13:51 macivy joined #git
13:53 kim_bruning ohhhhh, etckeeper immediately runs after I do an apt run
13:53 kim_bruning <3<3<3
13:53 kim_bruning selckin: that was a really neat tip. Why doesn't everyone know about this?
13:54 selckin because nutella is to breakfast food
13:54 kim_bruning it just doesn't automatically do a push... though that might actually be a Good Thing (tm)
13:54 jast kim_bruning: you may be getting the two different SSH URI schemes mixed up. one is user@host:relative/path. the other is ssh://user@host/absolute/path.
13:55 kim_bruning jast: indeed!
13:55 kim_bruning both are valid for git?
13:55 jast yes
13:56 kim_bruning learn something new every day
13:56 jast in fact possibly git even allows ssh://foo@bar/~/relative/to/homedirt
13:56 kim_bruning this is why you gotta *do* stuff, RTFM might be necessary, but not sufficient
13:56 roflincopter joined #git
13:57 kim_bruning so now that I've thoroughly shaved this yak, I'm off to shave the next one ;-)
13:58 kim_bruning geeze it's 4 pm already
13:59 Rodya_ joined #git
13:59 nikivi joined #git
14:00 jaafar joined #git
14:02 roflin joined #git
14:04 fairuz2 left #git
14:06 mingrammer joined #git
14:06 GodGinrai joined #git
14:06 ncd|leeN joined #git
14:07 phaleth joined #git
14:07 kim_bruning laters o/
14:08 Finnoloid joined #git
14:10 delboy1978uk joined #git
14:10 MainB_ joined #git
14:10 delboy1978uk whatever_file: needs merge
14:10 delboy1978uk then i nano it, search for <<<, nothing
14:10 delboy1978uk no conflicts
14:10 delboy1978uk so what has broken?
14:11 ResidentBiscuit joined #git
14:12 delboy1978uk its cool, think ive fixed it
14:13 pioiytr joined #git
14:13 jast there are several possibilities, e.g. file mode conflict, delete/modify conflict
14:14 jast tools like 'git merge' and 'git rebase' tell you, as part of their output, what kind of conflict they detected. unfortunately you can't see that information later on, so it helps to pay attention to the output
14:14 delboy1978uk jast: interesting, thanks, i'll investigate a little more
14:14 garethdaine joined #git
14:16 sanketdg joined #git
14:17 sanketdg joined #git
14:20 hahuang65 joined #git
14:21 Dougie187 joined #git
14:21 rgrinberg joined #git
14:21 l4v2 joined #git
14:22 rgrinberg joined #git
14:22 notfrosty joined #git
14:22 Darren_ joined #git
14:23 notfrosty How do I set master up to date from branch without actually changing branch ?
14:25 GodGinrai so you want to merge your branch into master
14:25 GodGinrai but not change branch?
14:25 GodGinrai to reword: but not switch branches?
14:26 notfrosty No, I wanna see what changes were done in master without switching to it. (git diff --name-status master..branch)
14:27 selckin think you might mean like, git fetch; git diff master..origin/master, else i'm lost
14:27 delboy1978uk left #git
14:27 GodGinrai ^ this
14:27 zivester joined #git
14:27 vassagus joined #git
14:28 madewokherd joined #git
14:28 notfrosty git fetch doesn't update master (git log master doesn't show latest commits)
14:29 notfrosty I wanna do git branch master; git pull origin master; git branch branch; Is it possible without actually changing branch ?
14:29 osse notfrosty: fetch updates origin/master
14:30 Manuel15 joined #git
14:30 Manuel15 joined #git
14:30 przemoc joined #git
14:30 przemoc joined #git
14:30 selckin not in a sane way and only if its a fast forward
14:30 osse notfrosty: sort of. you can "push" to the current repo. but it's a bit awkward
14:30 osse use origin/master in place of master. it probably gives you want to want
14:31 stamina joined #git
14:32 GodGinrai notfrosty: it updates origin/master
14:32 GodGinrai notfrosty: it would be better to stash, checkout master, pull, checkout branch, and then stash pop
14:33 notfrosty ok I got it, thx guys
14:33 GodGinrai if you want to update master and be back on your branch
14:33 ash_workz joined #git
14:33 osse git push . origin/master:master
14:33 osse aww yiss
14:34 lmat When I clone a remote
14:34 notfrosty left #git
14:34 lmat I clone a remote (git clone git4 git4clone;)  and the new repo's branches don't match the old! :o
14:34 eric97477 joined #git
14:35 lmat The old repo had a branch pointing to a particular merge point, but in the new one, that branch is pointing somewhere else!!
14:35 lmat This seems very bad.
14:35 lmat What's the difference between a green ref and a red ref in git log --decorate?
14:35 bremner also somewhat unlikely,
14:36 MainB_ left #git
14:36 osse lmat: usually local vs remote branches
14:36 kaldoran_ joined #git
14:36 _sjs joined #git
14:36 lmat osse: okay
14:38 lmat Maybe the problem is how I Get those branches in the original repo.
14:38 lmat I use     for remote in $(git branch -r); do git branch --track "$remote"; done;
14:38 osse then you clone the original branches become remote branches in the new clone
14:38 lmat Does this make all local tracking branches track at HEAD?
14:38 clemf joined #git
14:39 osse lmat: i suggest you only create the local branches you need. otherwise it just becomes a mess
14:39 lmat osse: No-can-do. I'm converting a repo and I need to get all the branches.
14:39 vassagus joined #git
14:39 osse maybe you want clone --mirror then
14:39 lmat osse: This will be a "central" repo. Of course, my local
14:39 thiago_ joined #git
14:39 lmat HAH!
14:40 * lmat kisses osse
14:40 lmat I'm pretty sure that's the one. Thanks!
14:41 lmat osse: Tell me about no-hardlinks. What happens whet I don't pass this?
14:41 osse git hardlinks
14:42 lmat osse: so du . -sh;
14:42 lmat What is the point of passing no-hardlinks?
14:43 lmat *with* hardlinks, du . -sh; shows the total repo size.
14:44 osse "This may be desirable if you are trying to make a back-up of your repository."
14:44 lmat osse: hmm, okay. Thanks!
14:45 osse i don't know more than that
14:45 rgrinberg joined #git
14:45 osse and I would guess du counts the size of hard linked files
14:45 osse du can't tell which is the "original" and which is the "linked" files, since that distinction is not present with hardlinking
14:46 osse (i think)
14:46 lmat My thoughts exactly.
14:46 Tobarja joined #git
14:49 eric97477 I am not sure how to solve this problem. I need to transition two git repos ( X & Y ) from server A to B. I did a git clone —mirror from A and a git push —mirror to B. This all seemed to work correctly and without issue. However, repo X has many branches and repo Y is a submodule on every branch. As I understand it, the .gitmodules file for X is tracked per branch. So, what is the best way to update the location of the submodule Y across all branches in X?
14:49 eric97477 Hopefully this makes sense. I am hoping that I do not need to individually checkout each branch, update the location of the submodule, and then commit/push .gitmodules.
14:50 Lipp joined #git
14:50 astrofog joined #git
14:51 TomyLobo joined #git
14:52 dmto joined #git
14:55 ToxicFrog eric97477: git filter-branch?
14:56 bsanford joined #git
14:56 eric97477 I honestly have no idea ToxicFrog. I have to assume someone has faced this issue before and I am hoping someone has an easy solution.
14:57 artemisart joined #git
14:57 darkbit joined #git
14:57 ToxicFrog eric97477: so, the two approaches that immediately come to mind for me are to use filter-branch (if you want to rewrite the historical contents of .gitmodules) or just a simple for loop over the output of 'git branch' (if you just want to make new commits updating it without changing the historical values)
14:58 peepsalot joined #git
14:58 suchness joined #git
14:58 eric97477 ok. so, essentially, I need to write a little script that will do the checkout for each branch, modify .gitmodules, and then push the results? I think I like that option better then rewriting history.
14:59 suchness Can someone tell me what this means: warning: symbolic ref is dangling: refs/remotes/origin/HEAD
14:59 suchness It's been reporting that when i switch branches
14:59 selckin !dangl
14:59 gitinfo [!dangling] Dangling objects represent things added to git which are no longer needed based on the git commands you typed in. This can be normal workflow (rebase, reset, add, etc) or errors you made. Typing "!dangling_commit" "!dangling_blob" and "!dangling_tree" into a query to me will get you more information about each type.
14:59 lmat suchness: I guess that means HEAD doesn't point to a
14:59 lmat nevermind
14:59 LionsMane joined #git
14:59 selckin you most likely made a commit on a detached head (maybe)
14:59 osse eric97477: hmm, maybe I misunderstand but can't you just copy them over?
14:59 lmat suchness: You need to create a *local* branch to track that remote.
15:00 osse eric97477: oh.. urls in .gitmodules and that I suppose?
15:00 eric97477 osse…yes, the url in the .gitmodules need to change for each branch.
15:01 InfoTest joined #git
15:02 eric97477 osse…as ToxicFrog mentioned, I could write a script to do this, but if there is an easier way, without rewriting history (unless that truly is the recommended method in this case), I would be interested to learn what that is.
15:02 cagomez joined #git
15:02 eric97477 If I didn’t have 30 some branches to update, this wouldn’t be much of an issue.
15:03 osse 30 branches! :O
15:03 eric97477 do not ask.
15:03 roelmonn_ joined #git
15:03 osse move repo X first. wait for the whole thing to blow over. forget about moving Y. profit.
15:04 eric97477 osse….I like your plan. Sadly, it is not one I can follow. So, a script it shall be.
15:04 Guest5 joined #git
15:04 ToxicFrog eric97477: the script seems pretty easy -- for branch in $(git branch); do git checkout $branch; sed -Ei '...' .gitmodules; git commit -am "update repo Y URL"; done
15:04 ResidentBiscuit I've never heard of a good story coming out of submodule usage
15:04 a_thakur joined #git
15:05 osse eric97477: git for-each-ref --format=%(refname:short) 'refs/heads' | while read branch; do git checkout $branch && sed -i .gitmodules blabla && git commit -am 'Update URL for module'; done
15:05 osse eric97477: e<
15:06 osse $(git branch) :O  for i in * master feature; do... Woops!
15:06 osse you rekt yourself
15:06 eric97477 thanks. I will go that route and get myself a script to do the work for me.
15:06 ASOLAr joined #git
15:07 algae joined #git
15:07 roelmonnens joined #git
15:08 JanC_ joined #git
15:08 ToxicFrog osse: if for some reason you have a branch named *, it won't be expanded in the output of $(git branch)
15:08 chrisgeorge joined #git
15:08 osse ToxicFrog: git branch prints a '*'. and the shell will expand it
15:09 osse at least bash/dash/sh will. zsh won't
15:11 ToxicFrog osse: aha. That explains it. I tested on zsh.
15:11 boyen joined #git
15:11 ASOLAr joined #git
15:11 ToxicFrog And totally forgot about the *
15:13 ech0s7 joined #git
15:13 ech0s7 hi!
15:13 GodGinrai good morning
15:14 ech0s7 i would create a different patch file for each commit into my repo, one by one in separate file ordered by time (from oldest to newest)
15:14 ech0s7 how can i do ?
15:14 cfxd joined #git
15:15 rossome joined #git
15:16 Guest93616 joined #git
15:17 victorbjelkholm_ left #git
15:17 GodGinrai osse: wouldn't for-each-ref be applicable here? ^
15:18 jguillen joined #git
15:18 dhollinger joined #git
15:19 sanketdg joined #git
15:20 tyreld joined #git
15:20 Darren_ joined #git
15:22 liamonade joined #git
15:22 Guest5 joined #git
15:22 Seveas ech0s7: man git format-patch
15:22 gitinfo ech0s7: the git-format-patch manpage is available at http://jk.gs/git-format-patch.html
15:24 Seveas ech0s7: though do be aware that merge commits are not representable as patches, so if you have merges, this cannot be done.
15:24 osse GodGinrai: see my suggestion :p
15:25 osse ToxicFrog: btw, git forbids ref names with a * in it. but I think the classic shell fork bomb is allowed
15:25 ech0s7 Seveas: i haven't merges
15:25 GodGinrai osse: I didn't see a suggestion in response to ech0s7
15:25 osse GodGinrai: oh, i misunderstood
15:25 kaldoran joined #git
15:25 sebboh joined #git
15:26 osse that just sounds like git format-patch --all or somesuch
15:26 EduardoMartins joined #git
15:28 ech0s7 thanks osse
15:28 ech0s7 just looking a way to exclude the first commit :)
15:28 fakenerd joined #git
15:28 khmarbaise joined #git
15:30 osse git format-patch firstcommit..HEAD
15:30 synthroid joined #git
15:31 synthroid joined #git
15:32 ahmed_elgabri joined #git
15:32 atomicb0mb joined #git
15:33 rwswitch joined #git
15:34 wilhelmtell joined #git
15:35 LionsMane joined #git
15:36 sebboh Hi!  I'm trying to build some open source project, but that's not what my question is about.  My question is, what knowledge am I missing about git which is preventing me from understanding what is happening?  I'll describe what is happening...  I've cloned a project called uzbl from github.  I'm using git 2.8.1.  Like some other projects, there's a master branch and a 'next' branch.  So I tried `git checkout origin/next`.  It says I am
15:36 sebboh in a detached HEAD state.  I don't recall seeing that message when I did the same thing on other projects.  But I think too that I only did `git checkout next`, when switching to the 'beta' on other projects.  Why did this one specify 'origin/next' in the install docs?  Why am I in a detached head state?  Does not each branch have its own HEAD?
15:36 eric97477 joined #git
15:37 Gitzilla joined #git
15:37 Seveas sebboh: 'origin/next' is not a branch, so if you check it out, your HEAD gets detached
15:38 wilhelmtell joined #git
15:38 wilhelmtell left #git
15:38 Seveas 'git checkout next' is a bit magical, it actually translates to 'git checkout -b next origin/next', so 'create branch next, tracking origin/next. And then check it out'
15:39 sebboh I didn't download a list of branches when I cloned the repo?
15:39 davisonio joined #git
15:40 Seveas you did, but those are stored as refs in refs/remotes, local branches don't get created automatically
15:40 shams joined #git
15:40 sebboh Does a simple git clone <Address> download everything that could possibly be downloaded?  Is the result suitable for use as a historical archive of the repo at the point in time the clone was made?
15:40 Seveas see https://git.seveas.net/recover​ing-from-a-detached-head.html and https://git.seveas.net/the-mea​ning-of-refs-and-refspecs.html for background info
15:41 sebboh ok, thanks
15:41 Seveas sebboh: it downloads everything downloadable, but that misses things like repo config and hooks, so it's not a complete backup
15:42 sebboh ok, understood.
15:42 shinnya joined #git
15:42 romerocesar joined #git
15:43 liamonade joined #git
15:43 leeN joined #git
15:43 rossome joined #git
15:43 byteflame joined #git
15:46 mrkat joined #git
15:49 abyx joined #git
15:49 kbs oh, that looks a great site Seveas - nice; and one more handy link for me :)
15:50 ShekharReddy joined #git
15:51 Muzer joined #git
15:51 Dougie187 Seveas: Doesn't it also miss unreachable commits?
15:52 hashpuppy joined #git
15:52 ksolv joined #git
15:53 Dougie187 Maybe those are considered not downloadable, but I think that distinction missing.
15:54 Electrometro joined #git
15:55 Derperperd joined #git
15:55 Lomex joined #git
15:58 madewokherd joined #git
15:58 bambams I was hoping I could use GIT_DIFF_OPTS to get -w --ignore-blank-lines into git commit -v. It doesn't appear to work.
15:59 bambams ack'ing the source it looks like it only looks for --unified and -u specifically...?
15:59 morenoh149 joined #git
15:59 Seveas Dougie187: it does. And the reflogs
16:00 Seveas https://git.seveas.net/how-to-​back-up-a-git-repository.html :)
16:01 cagomez joined #git
16:01 Guest5 joined #git
16:01 mingrammer joined #git
16:01 irco joined #git
16:02 Guest5 joined #git
16:04 dreiss joined #git
16:04 ikmaak joined #git
16:05 ahmed_elgabri joined #git
16:07 eivarv joined #git
16:08 oich joined #git
16:10 ksolv left #git
16:10 preyalone joined #git
16:16 subhojit777 joined #git
16:17 artemisart joined #git
16:17 johntramp joined #git
16:17 Guest5 joined #git
16:18 spudowiar joined #git
16:19 mingrammer joined #git
16:20 m0viefreak joined #git
16:21 xall joined #git
16:22 blackwind_123 joined #git
16:24 SteffanW2 joined #git
16:24 HappyLoa- joined #git
16:25 mSSM joined #git
16:26 dsdeiz_ joined #git
16:28 Timvde joined #git
16:31 Noldorin joined #git
16:32 nitric joined #git
16:34 bambams http://explosm.net/comics/4388/
16:35 kpease joined #git
16:35 Vampire0_ joined #git
16:36 dreiss joined #git
16:37 nettoweb joined #git
16:38 HappyLoa- joined #git
16:41 rossome joined #git
16:42 albel727 joined #git
16:43 jstein_ joined #git
16:48 rubyonrailed joined #git
16:49 jheckman joined #git
16:49 ahmed_elgabri joined #git
16:51 xall joined #git
16:52 daynaskully joined #git
16:53 HoierM joined #git
16:53 durham joined #git
16:53 jheckman I am running into a weird problem that I can't figure out. Earlier I cloned an internal repo, specifying the directory it should clone into. It cloned just fine but for some reason when I would fetch and reset nothing would actually change (but the log was correct). I deleted the folder and tried to clone again without specifying the folder and instead of cloning like normal, it is just showing the content of .git in the root of the che
16:54 jheckman If I try to checkout in a different place (like my home directory) it complains about the checkout in a different directory. This is on debian. I can't reproduce with the same repo on my mac
16:56 _ikke_ jheckman: A transcript of what you are doing would help a lot
16:57 davisonio joined #git
16:57 grawity hmm you aren't cloning with --bare or --mirror, are you?
16:57 Impaloo joined #git
16:57 HappyLoa- joined #git
16:58 jheckman not doing bare or mirror, one sec on transcript
16:58 lss8 joined #git
16:59 jost_ joined #git
17:00 jheckman jheckman@vm101:/var/git$ git clone git@internal_repo:ops/docker_config Cloning into 'docker_config'... remote: Counting objects: 41, done. remote: Compressing objects: 100% (32/32), done. remote: Total 41 (delta 9), reused 0 (delta 0) Receiving objects: 100% (41/41), 10.68 KiB | 0 bytes/s, done. Resolving deltas: 100% (9/9), done. Checking connectivity... done.
17:00 jheckman that formatting is less than ideal though
17:01 grawity hmm
17:01 jheckman jheckman@vm101:/var/git/docker_config$ ls branches  description  hooks  info            logs     packed-refs config    HEAD         index  jenkins_master  objects  refs
17:01 UTAN_dev joined #git
17:01 jheckman jenkins_master is one of the folders within the repo, that for some reason that is where the code is
17:01 grawity check that 'config' file, does it say "[core] bare = true"?
17:02 jheckman jheckman@vm101:~$ pwd /home/jheckman jheckman@vm101:~$ git clone git@internal_repo fatal: working tree '/var/git/docker_config/jenkins_master' already exists.
17:02 grawity hmm
17:02 _ikke_ type git
17:03 jheckman bare is false in the config
17:03 grawity env | grep GIT
17:04 jheckman GIT_WORK_TREE=/var/git/doc​ker_config/jenkins_master -- this was set for another script
17:04 grawity kindly please unset it
17:05 hdon joined #git
17:05 _ikke_ grawity++
17:05 jheckman reset my session, That fixed my problem. Thank you, should probably use a better variable name
17:06 eivarv joined #git
17:06 _ikke_ GIT_WORK_TREE has a meaning to git
17:06 phanimahesh joined #git
17:06 Jonuz joined #git
17:08 skylite joined #git
17:08 jheckman good to know. I did not write the other script so maybe there was a reason it used that instead of something unique
17:08 roelmonnens joined #git
17:10 Dougie187 Seveas: Oh sorry, I meant `git clone`. Not the proper backup method.
17:10 Dougie187 Maybe I was plugging in to the wrong part of a conversation though. :P
17:11 tgunr joined #git
17:12 roelmonnens joined #git
17:13 _ikke_ jheckman: It's usually used in combination with git checkout on an existing repository to put the files in some directory
17:14 jimi_ joined #git
17:16 bambams (Sorry for the explosm link.. Wrong channel)
17:16 eciohc joined #git
17:16 weedtm joined #git
17:17 InfoTest1 joined #git
17:18 filcab joined #git
17:19 dan2k3k4 joined #git
17:20 TooLmaN joined #git
17:21 abyx_ joined #git
17:22 cyphase joined #git
17:23 abyx_ left #git
17:27 sanketdg joined #git
17:30 notebox joined #git
17:31 mostlybadfly joined #git
17:32 fahadash joined #git
17:34 roelmonnens joined #git
17:34 siaw23 joined #git
17:35 siaw23 in my git log i have only 2 messages. the very first commit messages is wrong. i want to fix it. i tried ‘git rebase -i HEAD~2’ but it won’t show the 2 commit messages so i can pick which one i want to edit. it shows the last commit which is not what i want
17:35 tgunr left #git
17:35 siaw23 any idea how to fix old commit messages?
17:36 grayjoc joined #git
17:36 Ahmed90 joined #git
17:36 aard_ joined #git
17:36 siaw23 better if i can use the hash code to edit a particular commit message
17:36 GodGinrai wouldn't HEAD~2 mean to go to a commit that doesn't exist?
17:36 GodGinrai considering you only have 2 commits
17:37 bremner siaw23: pass the commit _before_ the one you want to edit to rebase -i
17:37 bremner oh. root commit. Yeah, there's a way, but I forget
17:37 GodGinrai siaw23: http://stackoverflow.com/questions/​2119480/edit-the-root-commit-in-git
17:37 siaw23 bremner: ok what if i have about 90 commits and i want to edit the 5th commit message?
17:38 siaw23 how would i do it?
17:38 YourTheBombNow23 joined #git
17:38 bremner then pass the commit before (so the 4th hash)
17:38 romankl joined #git
17:39 siaw23 so it would look like
17:39 siaw23 git commit hast_of_4th_commit —amend?
17:39 _ikke_ siaw23: git rebase --root
17:39 bremner no, git rebase -i hash_of_4th
17:40 siaw23 bremner: ah ok let me try
17:42 siaw23 _ikke_: git rebase —root does nothing for me
17:42 airdisa joined #git
17:42 siaw23 i want to edit the root commit message
17:42 infra-red joined #git
17:43 dreiss joined #git
17:44 _ikke_ siaw23: What do you mean, does nothing?
17:44 annoymouse_ joined #git
17:44 Spacew00t joined #git
17:44 _ikke_ AH, I thought it implied -i
17:44 _ikke_ git rebase -i --root
17:47 LarArT joined #git
17:50 iota_pi joined #git
17:52 mingrammer joined #git
17:53 blackwell joined #git
17:53 todd_dsm_wk joined #git
17:54 ASOLAr joined #git
17:55 Dyonisos joined #git
17:55 Guest5 joined #git
17:57 siaw23 joined #git
17:58 kinetik_ joined #git
17:58 BrianBlaze420 joined #git
17:58 siaw23 _ikke_: works now thanks :)
17:59 nettoweb joined #git
18:00 mSSM left #git
18:00 govg joined #git
18:00 iota_pi joined #git
18:01 dpower joined #git
18:01 moop joined #git
18:01 MrMojito joined #git
18:01 _ikke_ yw
18:02 kinetik_ Hello guys. I am interesting on understand the inner workings of git revert. Can we say that git revert works the same way as the svn reverse merge ?
18:02 fakenullie it just applies commit in reverse
18:03 _ikke_ Take a patch, change all + to - and - to +, apply
18:03 kadoban joined #git
18:03 kinetik_ I see.
18:04 _ikke_ (where patch is the diff of the commit to revert
18:05 BSaboia joined #git
18:06 b1tchcakes joined #git
18:06 subhojit777 joined #git
18:08 infra-red joined #git
18:08 kinetik_ Let's say we have 3 sequential commits, and all change the first character of the file. Why reverting the second commit makes git throw a conflict cause it can't decide between the first and the third change? Why git doesn't assume that the last commit should be preferred when managing the conflict?
18:08 fakenullie revert doesn't know anything about other commits
18:10 _ikke_ kinetik_: Why is the last commit prefered?
18:11 StuartMI joined #git
18:11 _ikke_ kinetik_: When you revert a commit, you want the changes of that commit gone
18:11 kadoban kinetik_: That specific revert seems like it would make no sense at all. Even thinking about it myself, I'd have no idea what you want the result to be. How could git know?
18:12 roelmonnens joined #git
18:14 ahmed_elgabri joined #git
18:14 nanga joined #git
18:14 eivarv joined #git
18:15 kinetik_ _ikke_: because the last commit is closer to the right code, since it was the last change
18:15 _ikke_ Not necessarily
18:15 hseg joined #git
18:16 kinetik_ what I really want to do (regardless of which command), it's to take a out a specific commit from the repository. In a real-case scenario I want to take out a specific feature from the repository by removing the commit, as the feature doesn't belong to the code anymore avoiding manual intervention
18:17 kinetik_ I thought revert was designed for that
18:17 fakenullie git is stupid content tracker
18:17 durham_ joined #git
18:18 davisonio joined #git
18:18 durham_ joined #git
18:18 _ikke_ Sure, but does not mean that reverting a commit would not commit with the current history
18:18 fakenullie not feature management tool
18:18 btree_ joined #git
18:19 _ikke_ (last commit == conflict)
18:19 kadoban kinetik_: You can certainly do that, with revert or otherwise. git doesn't know what your code means though, you have to resolve the conflicts that arise.
18:19 AaronMT_ joined #git
18:19 vaibhavsagar joined #git
18:21 Didac joined #git
18:22 kbs I also wonder if you mean revert in the sense of not having it in the history at all, as I wonder if git-revert actually does that?
18:22 _ikke_ kbs: No
18:23 _ikke_ kbs: That would be rewriting history
18:23 kbs right
18:23 _ikke_ kbs: git is basically creating a new 'undo' commit
18:23 kbs but reading log above, I sorta wonder if that's the actual intent, and was curious
18:23 kbs because git-revert is easier to understand (in the not-actually-removing-history sense) as being a merge
18:26 EvilPeng1 joined #git
18:26 kinetik_ I appreciate that, but in my perspective if you take a commit in the middle of a pool of commits and conflicts arise shouldn't the commit after the left out one be the deciding base-source to manage any conflicts?
18:27 kbs I think _ikke_ would know best, might strategy=recursive -Xours roughly approximate a merge that wanted to prefer HEAD on conflicts?
18:27 * kbs stays away from debates about about what should and should not be default behavior :)
18:28 yan left #git
18:28 _ikke_ kinetik_: git plays safe and lets you decide
18:29 kadoban kinetik_: Probably not. It requires human intervention to figure it out.
18:29 roelmonnens joined #git
18:29 SonikBoom joined #git
18:30 kinetik_ Ok. Thank you for your quick answers fakenullie, _ikke_ and kadoban
18:31 veilig2000 joined #git
18:31 kinetik_ I guess it makes also sense because every commit is based on the previous one, so git plays safe not assuming more than it should
18:32 fakenullie you can try branch from that commit, revert it in that branch and merge that branch
18:32 fakenullie but there will still be conflict, because commits in both branches changed that line
18:34 _ikke_ in any case, git revert does not look at the history at all
18:34 cagomez joined #git
18:35 hseg Hi. I'm trying to get a feel for my options w.r.t. subproject management. I have a couple of libraries that are used in multiple projects. How would I manage them?
18:35 _ikke_ git can be summarized as: git diff <commit> <commit>~1 | git apply && git commit
18:36 _ikke_ git revert*
18:36 hseg It seems like I could use submodules, but I've heard a lot of bad stuff about it.
18:36 fakenullie I've used git submodules successfully
18:37 fakenullie it's good when each project can break from library upgrade
18:37 kbs ah, that's makes it very clear (re: git revert) thanks
18:37 mohabaks joined #git
18:37 Maple__ joined #git
18:37 hseg fakenullie: What do you mean by "break from library upgrade"?
18:37 fakenullie hseg: when library interface is unstable
18:38 fakenullie and needs to change
18:38 fakenullie or you just don't want to break other projects by changes in library
18:38 fakenullie we had dormant projects, so that worked good
18:38 hseg Ah, OK.
18:38 hseg also, is it correct to summarize safe submodule usage as: "clients of a library only change the SHA they point to, all changes to the library are made in the library repo itself"?
18:39 fakenullie it's only possible usage
18:39 artemisart joined #git
18:39 _ikke_ yes, that's correct
18:39 hseg OK.
18:39 _ikke_ Although nothing prevents you from comitting changes in the submodule repository
18:39 _ikke_ It's your responsibility to share those changes
18:40 romankl joined #git
18:40 fakenullie you can refer them by write-only urls in project
18:40 fakenullie read-only
18:40 dsantiago joined #git
18:40 fakenullie that's how we did it, git:// url on github
18:41 hseg How does a git:// github URL make it read-only?
18:41 _ikke_ It doesn't
18:41 hseg Then what did you mean by that remark?
18:42 ASOLAr joined #git
18:42 fakenullie github does not allow write via git://
18:42 fakenullie or you can specify read-only user in url
18:42 hseg Really?
18:42 bremner almost nobody allows write via git://
18:42 hseg Oh, for organizations I suppose.
18:42 _ikke_ You can still commit in the repository
18:42 hseg _ikke_: If you commit changes to the submodule, how do you keep them from being lost?
18:42 _ikke_ and if you commit that commit in the parent repo, no one is able to find hte commit
18:43 fakenullie hseg: you push them
18:43 Manuel15 joined #git
18:43 Manuel15 joined #git
18:43 grayjoc joined #git
18:44 hseg OK. Then what's this about commits being lost?
18:44 fakenullie I'm not sure
18:44 _ikke_ You need to make sure they're pushed
18:45 hseg Otherwise, what happens? Suppose I committed a local patch to a library I use, then pull from upstream?
18:45 diegoviola joined #git
18:45 sonOfRa joined #git
18:45 fakenullie the change will be merged or rebased into upstream
18:45 _ikke_ It gets merged
18:46 fakenullie maybe "lost" change are then someone forgets to commit and push submodule changes
18:46 _ikke_ yes
18:46 hseg assuming I'm tracking a remote branch. So submodules are no different from regular repos?
18:46 _ikke_ or commit them, commit them in the parent repo, but never push
18:46 fakenullie yeah
18:46 _ikke_ Then other ones try to check out the submodule but cannot find the recorded commit
18:47 davisonio joined #git
18:47 hseg OK, so don't update the library version used in the project without pushing the library commits first.
18:48 _ikke_ correct
18:49 hseg And seeing as submodules are considered binary blobs w.r.t. git, resolve merge conflicts in the submodules before fetching from upstream in the project.
18:49 hseg Doesn't seem like they deserve the bad press they seem to be getting.
18:49 hseg (he said, unaware of the hours of sadness ahead of him)
18:50 ash_workz joined #git
18:50 fakenullie well, unpushed submodules happen quite often
18:51 hseg Oh? Why?
18:51 satifant joined #git
18:51 fakenullie people forget
18:51 axelitus joined #git
18:51 fakenullie iirc you have to go into submodule and push explicitly
18:52 hseg Eh. OK.
18:52 fakenullie so if someone does that on goes on vacation it can be tough
18:53 hseg Right. As usual, make sure to leave history in consistent state before getting hit by a bus.
18:53 _ikke_ It's better to not directly work in the submodule dir
18:53 hseg OK, will avoid doing that.
18:53 * _ikke_ thinks read-only submodules should be a thing
18:54 veilig2000 joined #git
18:55 hseg Yeah... If you actually need to patch the submodule, wouldn't it be saner to clone the submodule repo as a repo in its own right, develop on it, and use the clone as a submodule?
18:55 fakenullie it's more convenient to make changes together
18:56 _ikke_ yes, but also more error prone
18:56 hseg Why?
18:56 _ikke_ For the reasons we just mentioned
18:56 zacts joined #git
18:57 hseg I develop libfoo-v1-internal, then set project bar to use that version of libfoo instead of the upstream version.
18:57 NightStrike I just did a "git checkout -b XX".... how do I rename that to YY ?
18:58 hseg Keeping track of work within libfoo works just as usual, and the only thing to keep in mind is to only change library version used once the corresponding commit is pushed
18:58 hseg NightStrike: git branch -m XX YY
18:59 FuzzySockets joined #git
18:59 NightStrike hseg: thanks
18:59 davisonio joined #git
18:59 nettoweb joined #git
18:59 Maple__ left #git
18:59 jnewt joined #git
19:00 MrMojito joined #git
19:00 MrMojito joined #git
19:00 Maple__ joined #git
19:00 jnewt i'd like to set the author & commit date of one commit to the same as another (i split a commit buried in history, but didn't preserve the acutal dates of the file changes (which i would like to))
19:01 jnewt i have the sha of the old commit and the sha of the new commits.
19:02 _ikke_ jnewt: look into git commit -c / -C
19:02 siaw23 i did a rebase and i have my profile pic appearing on commits fromt he point i did the rebase. how do i revert this? https://gyazo.com/6294feabb16524e043b36d3a16fc31cd
19:02 dan2k3k4 joined #git
19:02 _ikke_ (you do have to restore the message though)
19:02 xdinomons joined #git
19:03 _ikke_ siaw23: This is probably because comitted date is different from author date
19:04 NeonSynth joined #git
19:04 siaw23 but this has affected my app and tests are failing now
19:04 siaw23 but there’s no code change :(
19:04 _ikke_ Why?
19:04 jnewt _ikke_, is there a way to get the commit date of a commit by sha?
19:04 _ikke_ Why should your app fail on this?
19:04 siaw23 i don’t know why the test are failing after. since there’s not code change
19:04 _ikke_ jnewt: look at man git show --format
19:04 gitinfo jnewt: the git-show manpage is available at http://jk.gs/git-show.html
19:05 l4v2 joined #git
19:05 siaw23 i just edited an older commit using rebase and reword
19:05 xdinomons how do i make a bare repository available to anyone like a webserver
19:06 _ikke_ xdinomons: dumb http is easiest, but requires you to execute git update-server-information regularly
19:06 _ikke_ siaw23: Why would that break your app?
19:07 siaw23 _ikke_: i don’t know why it’s doing that honestly
19:07 _ikke_ siaw23: find out why the tests are failiing
19:07 _ikke_ failing
19:07 MrMojito joined #git
19:07 monk12 joined #git
19:08 eivarv joined #git
19:09 xdinomons ohh just read an article that ssh is best
19:09 monk12 Hey all, Im using git with Eclipse. I was curious... i forked a project in BitBucket (so its my personal playground for a project)... in Eclipse i had the original origin Remote. but also just added a new Remote for my personal fork... how does Eclipse know which Remote to send it to by default (pretend its the same branch in both remotes)?
19:09 _ikke_ xdinomons: yeah, imo it's the best
19:10 siaw23 _ikke_: is there a way to revert the rebase i made?
19:10 _ikke_ monk12: git has so called tracking information
19:10 monk12 how does it know which Remote to send to currently. Im wondering how I can find that out or change it via GUI/Eclipse and command line git.
19:10 siaw23 i honestly don’t know why my tests are failing
19:10 xdinomons left #git
19:11 monk12 so if the checked out branch in Eclipse was originally associated with the origin remote... how can i change to new remote be4 my Push?
19:11 _ikke_ siaw23: A beter thing would be to diff the changes from before and after
19:11 _ikke_ monk12: I don't know how to do it in eclipse
19:11 siaw23 _ikke_: i thought of that but i’m not sure what to diff against what
19:11 _ikke_ siaw23: look at man git reflog
19:11 gitinfo siaw23: the git-reflog manpage is available at http://jk.gs/git-reflog.html
19:12 _ikke_ This keeps track of HEAD and your branches so that you can retrieve back commits
19:12 monk12 _ikke_: its might be Right CLick->Properties in the configuration... u can change... _ikke_: how about on the command line? how can you check and change a remote for a branch.
19:12 monk12 just curious how to do it the proper way on command line
19:13 _ikke_ monk12: easiest is git push -u <remote> <branch>
19:13 _ikke_ the -u will make git update the tracking info
19:14 monk12 _ikke_: cool... i think i also got a way to report it tooo..
19:14 monk12 _ikke_:  git branch -vv    (two v's)
19:14 monk12 seems to work
19:15 _ikke_ monk12: correct
19:15 merpnderp joined #git
19:15 monk12 sweet... thanks _ikke_ ... this really helps
19:15 _ikke_ git status will also mention it for the current branch
19:15 _ikke_ git checkout
19:16 mingrammer joined #git
19:16 merpnderp We have a react project in git. But we don't want to regularly check in our builds. But we do want to deploy those builds with git. Are there sub projects that we'd only commit and push to a sub repo for that build folder?
19:18 GodGinrai merpnderp: git is not a deploy tool
19:18 GodGinrai merpnderp: you shouldn't use it as one
19:19 siaw23 _ikke_: i think the head somehow moved to another point
19:19 Pinkamena_D joined #git
19:20 filcab joined #git
19:20 mr_quincy joined #git
19:21 jeffmjack joined #git
19:24 merpnderp GodGinrai: git is actually a great deploy tool if you want to use it as one.
19:24 merpnderp Lots of people happily use it that way day in and day out.
19:24 merpnderp Github et al have special git deployment keys just for using a special key for deployments.
19:25 palmira67 joined #git
19:25 GodGinrai merpnderp: A lot of people use excel to keep track of their work, but JIRA et al do a much better job
19:26 GodGinrai merpnderp: github deploy keys are for deployment tools to use to clone the repo *during* the deploy process.
19:26 GodGinrai git itself isn't the deployment tool
19:27 Derperperd joined #git
19:27 mr_quincy I inherited a "non-bare" git repo, and I'm getting a strange result from running "git status" from the master branch. It says "On branch master" and on the next line says "Your branch is ahead of origin/master by 7 commits". This should not be. The master branch should be identical to the origin/master branch.
19:28 davisonio joined #git
19:28 tobiasvl mr_quincy: well apparently it's not, but your origin/master might not be up to date with origin's master
19:28 tobiasvl do a git fetch
19:28 GodGinrai mr_quincy: They might not be identical in commits
19:28 merpnderp GodGinrai: so the deployment tool does git clone or git pull. Like my fingers will do on the CLI.
19:28 GodGinrai what tobiasvl said, too
19:28 siaw23 _ikke_: ok the problem is this “Your branch is ahead of 'origin/rails4_clone' by 125 commits.
19:29 siaw23 (use "git push" to publish your local commits)
19:29 siaw23 nothing to commit, working directory clean”
19:29 siaw23 _ikke_: how can i fix that please. i don’t want to risk to get things worse
19:29 siaw23 is it because i used rebase?
19:29 GodGinrai merpnderp: the deployment tool will clone so that it can get the source for building/deploying
19:29 siaw23 because i don’t see how my branch can be ahead 125 commits
19:30 siaw23 that’s why my tests are failing :(
19:30 GodGinrai merpnderp: if you don't intend to "control" your source, there is no reason to be using git
19:30 paul424 joined #git
19:30 merpnderp GodGinrai: What do you mean? I want a submodule for our build folder. We run a build that we decide is ready for deployment. Check it in to the submodule's master. And then on prod we can do a git pull.
19:30 merpnderp What's wrong with that?
19:31 GodGinrai siaw23: did you try the fetch command like tobiasvl suggested?
19:31 tobiasvl that was for mr_quincy
19:31 tobiasvl but applies for siaw23 too probably
19:31 siaw23 GodGinrai: i didn’t see his “fetch” suggesting
19:32 LionsMane joined #git
19:32 siaw23 git fetch is not the same as git pull?
19:32 GodGinrai !deploy
19:32 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
19:33 arand joined #git
19:33 GodGinrai siaw23: `git pull` === `git fetch; git merge origin/master` essentially
19:33 GodGinrai when run from master
19:35 mniip joined #git
19:35 _Cyclone_ joined #git
19:35 kpease joined #git
19:35 fractalizator joined #git
19:36 siaw23 how do i set the head to the current commit?
19:37 siaw23 i think that’s what’s bugging me
19:37 Bizkit joined #git
19:37 GodGinrai siaw23: HEAD is where the current commit is
19:37 GodGinrai that's what HEAD means
19:38 siaw23 i thought you could set it to which commit you would like
19:38 dodobrain joined #git
19:38 GodGinrai siaw23: when you do, that changes which commit is the "current commit" for a branch
19:38 siaw23 yes
19:39 siaw23 my tests is failing i think for this reasong “Your branch is ahead of 'origin/rails4_clone' by 125 commits.”
19:39 siaw23 but i don’t understand why my local repo can be AHEAD
19:39 siaw23 and i want it to be the same as the remote repo
19:39 fakenullie you pulled commits from other branch
19:40 yitz ls
19:40 fakenullie . ..
19:41 yitz If someone checked a massive binary into a repo and I did a 'git rm', how do I clear it out of the .git to shrink my repo size?
19:41 fakenullie you can't
19:41 fakenullie it's forever there unless you change the history
19:41 perlpilot yitz: rewrite the commits such that it's not there anymore
19:41 yitz >.>
19:41 khmarbaise joined #git
19:43 yitz How do I do that?
19:43 shaggycat joined #git
19:44 fakenullie https://help.github.com/arti​cles/remove-sensitive-data/
19:46 l4v2 joined #git
19:47 yitz If I got to a detached HEAD, how do I get back to the ... origin HEAD?
19:47 yitz Oh. origin/HEAD :)
19:47 arooni i added a file to not be ignored via a .gitignore file; yet when it try to add the file it tells me its still part of the git ignore file.. why?
19:47 kapu joined #git
19:48 arooni do i need to commit the changes to the ignore file firsT?
19:48 fakenullie arooni: something is wrong, you don't need to commit it
19:48 yoan joined #git
19:48 arooni why does git hate me
19:49 circ-user-RNfpp joined #git
19:49 fakenullie I guess you made a mistake
19:49 arooni in my ignore file i have * ;  then  later i have !bin/pbcopy-remote  .. located in ~/bin/pbcopy-remote
19:49 gitinfo [!binary] Storing binary files in git causes repo balloon, because they do not compress/diff well. In other words, each time you change a file the repo will grow by the size of the file. See !annex for some solutions
19:50 paul424 joined #git
19:51 mr_quincy tobiasvl: Thanks for the tip. I want my master to become origin/master. Ordinarily I would git add, git commit, and git push origin master. When I try to push however, I get an error message related to non-bare repositories. I don't want to create an inconsistent work tree, but I do want to push my local master branch to the origin/master branch.
19:51 Sasazuka__ joined #git
19:51 fakenullie mr_quincy: your remote is non-bare repository
19:53 zeroed joined #git
19:53 zeroed joined #git
19:53 mr_quincy fakenullie: that seems consistent with the messages I'm seeing. Why would anyone use a non-bare repo? I haven't seen one before I inherited this.
19:54 fakenullie mr_quincy: did you clone from non-bare repository?
19:54 tobiasvl mr_quincy: perhaps the remote has only been pulled from, not pushed to in the past?
19:54 Sasazuka__ joined #git
19:55 mr_quincy fakenullie: unfortunately I'm not certain how this clone was initially created. It does seem likely that no one has ever pushed to this. It was kind of an abandoned undergraduate project
19:55 tobiasvl do you need to push to it?
19:56 tobiasvl do you have access to the remote somehow?
19:56 yitz ...how do I attach my HEAD? I did a checkout origin/HEAD
19:56 mr_quincy Ideally, I'd like to push the changes that haven't been committed yet, just so I have a complete history. Once that's done I'll wipe this .git out and create a bare repo for my future developments.
19:57 mr_quincy I do have access to the remote
19:57 mr_quincy admin rights as well
19:57 mr_quincy on the fs where the remote is located
19:57 grawity yitz: checkout an actual branch
19:59 fakenullie mr_quincy: you can clone it into bare repo and then push to it
19:59 fakenullie or you can push to new .git
19:59 yitz grawity: Thanks
20:02 cdg joined #git
20:02 nettoweb joined #git
20:03 thiago joined #git
20:05 foist joined #git
20:05 mr_quincy fakenullie: Please feel free to tell me if I'm doing something stupid, but it seems cleaner to push these changes to the non-bare origin as a kind of "final backup" before I create a bare repo. This repo contains a subset of the source I wish to control moving forward.
20:06 fakenullie mr_quincy: create new bare repo and push to it, without changing old repo
20:06 t-mart joined #git
20:07 Guest5 joined #git
20:08 xall joined #git
20:08 hjubal joined #git
20:09 vaibhavsagar joined #git
20:09 JacobsLadd3r joined #git
20:11 mr_quincy fakenullie: thanks, I'll do that. What's the use case for a non-bare repo?
20:11 zumba_addict joined #git
20:11 le_melomane joined #git
20:11 fakenullie mr_quincy: work with it?
20:11 zumba_addict folks, when my ofcmate was doing git revert on  a sha 3 commits ago, it's failing
20:11 zumba_addict i only told to do a git log and find the sha
20:12 fakenullie failing?
20:12 zumba_addict then also told him to do a git revert the_sha_key_here
20:12 zumba_addict yeah
20:12 zumba_addict it's failing
20:12 rossome joined #git
20:12 fakenullie what error
20:13 zumba_addict one sec, let me ask him to forward the error to me
20:13 zumba_addict he's sending it
20:13 grayjoc joined #git
20:13 Narwhaal joined #git
20:16 nwkj86 joined #git
20:18 javier_ joined #git
20:18 Sasazuka joined #git
20:19 zumba_addict gheez, i LOL hard, he git revert someone's code
20:19 zumba_addict can we undo the revert?
20:20 fakenullie you can revert a revert
20:20 zumba_addict so the sha of the last revert right?
20:20 GodGinrai and revert a revert of a revert
20:20 fakenullie yes
20:20 zumba_addict hahaha :D
20:20 zumba_addict he is friggin scared man! LOL
20:20 zumba_addict and teammates where wondering why i was laughing so hard
20:20 zumba_addict brb
20:21 merpnderp Thanks GodGinrai that was good stuff.
20:21 mr_quincy fakenullie: I misspoke, sorry. Reading a bit to ask a better question. Is there a use case for establishing non-bare remote origin? I hope I asked that correctly...
20:22 GodGinrai merpnderp: np
20:22 btree_ left #git
20:22 fakenullie mr_quincy: well, he may worked on it, and then he cloned it in a new place
20:22 sbeex joined #git
20:22 fakenullie mr_quincy: he did not plan to push
20:22 sbeex good evening
20:22 fakenullie clone just kept source repository as origin
20:23 fakenullie mr_quincy: or he pushed to non-checked out branches
20:23 SirCmpwn joined #git
20:24 ajf- joined #git
20:26 mablae joined #git
20:27 netcarver joined #git
20:27 zumba_addict fakenullie: finally received the error message - error: Commit ef058f25291fecf952abf027d3f797ec4 is a merge but no -m option was given.
20:27 mr_quincy got it. Thanks fakenullie, you've been very helpful.
20:27 _ikke_ zumba_addict: says it all
20:27 zumba_addict then next line fakenullie fatal: revert failed
20:27 zumba_addict what do we have to do?
20:27 _ikke_ zumba_addict: Do you actually want to revert a merge?
20:28 zumba_addict he just wants to revert all the things he updated
20:28 jimi_ joined #git
20:28 cyan__ joined #git
20:28 zumba_addict git log shows it
20:28 _ikke_ git revert would not remove it from git log
20:29 zumba_addict yup, I'm aware
20:29 fakenullie git reverting a merge is not trivial
20:29 zumba_addict he just want to remove the errors he introduced
20:29 Sasazuka joined #git
20:29 zumba_addict so he wants to git revert the sha that git log is showing but we're getting that error above
20:29 artemisart joined #git
20:29 fakenullie you have to specify which branch you want to get rid of
20:29 zumba_addict we are in master
20:30 zumba_addict so if it's a merge, we have to specify that branch that we merged into master?
20:30 fakenullie fakenullie: look at a commit, see which branch is master
20:30 fakenullie which parent of it is master
20:31 zumba_addict which command would show it?
20:31 zumba_addict will git log show it?
20:31 fakenullie yes, search for -m in git revert
20:31 fakenullie man git revert
20:31 gitinfo the git-revert manpage is available at http://jk.gs/git-revert.html
20:31 zumba_addict ok
20:31 zumba_addict -m parent-number so parent-number is the sha?
20:31 fakenullie I can't remember which command shows parents
20:31 _ikke_ zumba_addict: parent number is always 1
20:32 zumba_addict so -m 1 ?
20:32 _ikke_ yes
20:32 zumba_addict so git revert -m 1 theshaofthecommit
20:32 _ikke_ It tells git revert which side of the revert to 'keep'
20:34 jadew joined #git
20:35 jadew hey, is it still not possible to add an empty folder to the index?
20:35 milki jadew: it will never be allowed
20:35 jadew milki, why never?
20:35 milki folders arent tracked by git at all
20:35 jadew well, yeah, but why would it never be allowed
20:35 milki jadew: this is a very low-level git specification
20:36 jadew it seems like a big oversight
20:36 jadew I see, so no plans to fix it
20:36 milki jadew: a question we would ask is why do you need a folder?
20:36 NeverDie joined #git
20:36 ResidentBiscuit It's nothing to "fix"
20:36 ResidentBiscuit It's not a bug
20:36 jadew milki, because an empty folder or folder structure for that matter can be part of a project
20:36 jadew it's not always the contents of the files that matter
20:36 milki jadew: we believe its not the responsbility of git to track non-code things like folders
20:36 Pinkamena_D joined #git
20:37 milki jadew: why does the project require empty folders?
20:37 tjbp joined #git
20:37 jadew milki, for structure
20:37 milki what does that mean?
20:37 jadew it means that some stuff may go in there in the future and there's a place for that
20:38 jadew it may also be as part of a template
20:38 perlpilot jadew: a common mechanism is to put a .gitignore file in the directory and add that to the repo
20:38 jadew perlpilot, yeah, that doesn't seem like a decent solution
20:38 perlpilot jadew: another is to have your deployment mechanism build the needed directory structure
20:38 milki jadew: this sounds like a workflow responsibility rather than a code repository responsibility
20:39 jadew milki, the code may rely on those folders being there
20:39 crayon joined #git
20:39 jadew anyway, I see your point
20:39 milki your application should be able to handle the creation of folders during runtime or configuration
20:39 crayon joined #git
20:39 milki jadew: and its best not to run your applications straight out of the repo
20:40 Ineentho[m] joined #git
20:40 zumba_addict sorry was in his cube
20:40 jnewt just put a blank .gitignore in the folder
20:40 zumba_addict so I saw the branch that was merged into master
20:40 jadew jnewt, yeah, silly solution
20:40 jadew how is that part of the code?
20:41 jadew the point is that various projects need this sort of functionality and forcing the user to do shitty stuff like putting .gitignoore files or dummy files in there doesn't seem to make sense
20:41 zumba_addict ah _ikke_ since it's a merge say branch_A into master with a sha of abc123, the command will be git revert -m 1 abc123?
20:41 jnewt it's not, but it will solve the problem of not having folders you create stick around.
20:42 jadew jnewt, I know, I've been using dummy files in the past, right now I was just countering the point that "it shouldn't do it"
20:43 perlpilot jadew: It is a little weird that git doesn't treat directories as files (just like unix does).  But, it is what it is.
20:43 ResidentBiscuit git tracks files, not directories. If that doesn't suit your needs, don't use git
20:43 jadew ResidentBiscuit, how's that an argument?
20:43 dmto joined #git
20:43 ResidentBiscuit It's not an argument, it's a fact of life. If a tool doesn't do what you need, don't use the tool
20:43 jadew perlpilot, I agree
20:43 zumba_addict i found this in stackoverflow,i it
20:44 jnewt are we arguing about something?
20:44 * jnewt is late to the party
20:44 zumba_addict it's git log output, Am i right that first value is parent? Merge: 8989ee0 7c6b236
20:44 morenoh149 how do I list the current changes when committing?
20:44 zumba_addict i mean 8989ee0 is the parent 1 which is master?
20:44 jnewt morenoh149, git status
20:44 morenoh149 I think it's a thing added by zsh and I miss it
20:44 kbs jadew: basically, you're right that git doesn't support empty folders; and for fairly basic choices made in its design. It's reasonable to want it to do it if it can, and the main reason it doesn't exist is fundamentally a symptom, not an explicit goal per-se.
20:45 morenoh149 jnewt: I used to be able to do `git commit` and I'll be prompted with vim and a diff below my commit message
20:45 jnewt git diff .
20:45 morenoh149 this way I can write the commit message while reviewing the changes to be commited in a single vim window
20:45 jadew kbs, makes sense, thanks
20:45 jadew I thought it's a design decision
20:45 zumba_addict he couldn't wait, he edited the file again, lol
20:46 lamont joined #git
20:46 thiago joined #git
20:46 morenoh149 jadew: git commit -v
20:46 morenoh149 https://gist.github.com/bkono/6358506
20:47 morenoh149 now I need colors 8)
20:47 CheckDavid joined #git
20:50 cagomez joined #git
20:51 morenoh149 dash vee son
20:53 morenoh149 ah that's a vim setting actually
20:54 nettoweb joined #git
20:59 EvilPenguin joined #git
21:01 thibaultcha joined #git
21:05 mingrammer joined #git
21:08 Cihan joined #git
21:11 SporkWitch joined #git
21:12 adron joined #git
21:13 kyan joined #git
21:14 paws_ joined #git
21:14 cyan__ joined #git
21:14 glebihan_ joined #git
21:15 osse pro tip, if you ever forget -v you can do :$r!git diff --staged
21:15 mingrammer joined #git
21:15 osse nnoremap SHIT :$r!git diff --staged<CR>``
21:17 osse hmm
21:17 osse might be easier to :q and git commit -v
21:17 dodobrain joined #git
21:20 ericP joined #git
21:20 ericP left #git
21:20 mda1 joined #git
21:28 mattcen joined #git
21:31 Literphor joined #git
21:33 aard_ joined #git
21:33 blackwind_123 joined #git
21:33 cd-rum joined #git
21:36 veilig2000 joined #git
21:37 cd-rum_ joined #git
21:37 sagerdearia joined #git
21:37 Manuel15 joined #git
21:37 Manuel15 joined #git
21:38 mycyber joined #git
21:39 texinwien_ joined #git
21:39 mellernoia joined #git
21:41 veilig2000 joined #git
21:44 cyphase joined #git
21:46 robotroll joined #git
21:51 Dougie187 left #git
21:52 unbalanced joined #git
22:01 veilig2000 joined #git
22:02 cortexman joined #git
22:03 cortexman is there a way to have a meta-repository that has the behavior when you "git clone" it clones all the repositories that it points to, and that when you "git pull" it updates to the latest version of all the repositories that it points to?
22:03 DeJoker joined #git
22:06 jhass you might want the repo tool from the android project https://code.google.com/p/git-repo/
22:08 cortexman it wasn't clear to me how general that tool is
22:10 osse Poor man's alternative: use  git clone --recursive, and put 'git submodule update' in .git/hooks/post-merge and .git/hooks/post-checkout
22:10 cortexman osse, those hooks go in all submodules?
22:11 osse no
22:11 osse they go in the outer repo
22:11 spudowiar joined #git
22:11 adron joined #git
22:11 cortexman osse, how will i then update?
22:12 cortexman git pull?
22:12 osse yes
22:12 cortexman nice
22:12 DeJoker left #git
22:12 mycyber joined #git
22:13 StuartMI joined #git
22:13 osse but this requires some manual steps to set up. maybe the android repo tool is better in this regard
22:13 cortexman if i could set up the meta repository such that users just type repo init, etc, that would be ideal
22:14 vassagus_ joined #git
22:14 osse that's not happening with vanilla git
22:14 adron joined #git
22:14 osse but you could put a script in the outer repo that will set things up after they clone it
22:16 cagomez joined #git
22:17 Ravana joined #git
22:20 Ineentho[m] joined #git
22:21 jaafar joined #git
22:23 perlpilot joined #git
22:23 cdg_ joined #git
22:26 ahr3n joined #git
22:26 cd-rum joined #git
22:26 Shakeel_ joined #git
22:29 nettoweb joined #git
22:29 boombatower joined #git
22:29 thiago_ joined #git
22:34 madewokherd joined #git
22:34 artemisart1 joined #git
22:37 cjohnson joined #git
22:37 cjohnson When I make a new branch and git push origin newbranch
22:37 cjohnson I always have to come back and --set-upstream-to
22:37 cjohnson How can I tell git to be smarter about that?
22:38 osse cjohnson: use -u the first time you push
22:38 mingrammer joined #git
22:38 osse -u = --set-upstream
22:38 cjohnson sweet. is there also a shorthand to say "use the same branch name on remote"?
22:38 osse yes. you're already using it
22:39 osse git push origin newbranch is the same as git push origin newbranch:newbranch
22:39 osse for something even shorter and/or general I think git push origin HEAD will work
22:39 cjohnson Oh I guess I mean, even shorterhand to say "use the current branch"
22:39 cjohnson gotchya
22:39 cd-rum joined #git
22:39 osse git config alias.topkek 'push -u origin HEAD'
22:40 osse then git topkek
22:40 osse \o/
22:40 red_ joined #git
22:42 b1tchcakes joined #git
22:42 cd-rum_ joined #git
22:42 Spec-Chum joined #git
22:43 NightStrike I have a commit that has changes from 2 files... it's supposed to be 2 commits one for each file...  how do I split that?
22:44 mingrammer joined #git
22:44 garethdaine joined #git
22:45 l4v2 joined #git
22:46 nedbat NightStrike: is it the latest commit on your current branch?
22:47 NightStrike yeah
22:47 osse git reset HEAD~; git add file; git commit; git add file2; git commit
22:48 NightStrike osse: thanks!
22:53 le_melomane joined #git
22:54 Orion3k joined #git
22:54 rkazak joined #git
22:56 Derperperd joined #git
22:57 adron joined #git
22:58 crayon joined #git
22:58 thibaultcha left #git
23:01 crayon joined #git
23:01 Rodya_ joined #git
23:04 crayon joined #git
23:09 crayon joined #git
23:09 rwb joined #git
23:13 red_ joined #git
23:13 liamonade joined #git
23:13 yitz After a `git fetch origin master` how do I see the diff between local HEAD and origin/master/HEAD ?
23:13 zivester joined #git
23:14 osse git diff HEAD origin/master
23:15 osse origin/master/HEAD isn't a thing
23:16 yitz That command is showing no output >.>
23:17 osse Then there are no differences
23:17 ljc joined #git
23:17 yitz But `git fetch origin master` showed: remote: Total 31 (delta 20), reused 16 (delta 9)
23:17 osse It probably fetched more than just master
23:18 timewalker joined #git
23:20 dodobrain joined #git
23:21 kn-928 joined #git
23:29 dsdeiz joined #git
23:29 dsdeiz joined #git
23:30 alansaul joined #git
23:31 alansaul joined #git
23:35 JanC_ joined #git
23:38 vassagus joined #git
23:39 mablae_ joined #git
23:39 Derperperd joined #git
23:40 nilg joined #git
23:42 _sjs joined #git
23:42 AbstractArtist joined #git
23:44 afuentes joined #git
23:46 ahr3n joined #git
23:47 cdg joined #git
23:52 RustyShackleford joined #git
23:53 kulyzu joined #git
23:53 a3Dman joined #git
23:53 kulyzu joined #git
23:54 Spec-Chum joined #git

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