Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-06-20

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

All times shown according to UTC.

Time Nick Message
00:00 MrJones yea I'm not disputing nobody needs it, but all the tutorials covering it seem to have some strange priorities :)
00:00 MrJones gladly this place exists tho and you saved me :D
00:01 kadoban Hehe, glad to help.
00:06 ToxicFrog MrJones: based on my experience, people generally assume that either you have a priori knowledge that you want to throw away half of the merge (e.g. "this branch is obsolete but we need to mark it merged for workflow reasons")...
00:07 ToxicFrog ...or that you're using a mergetool, which itself gives you the option of either resolving conflicts individually or applying various strategies file-by-file
00:07 ToxicFrog Like, if you're using Meld, once you have conflicts you 'git mergetool' and then foreach conflict file you have an "accept all ours" and "accept all theirs" button
00:08 liquidjamm joined #git
00:09 Gitzilla joined #git
00:09 armyriad joined #git
00:10 uhmgswa^2 joined #git
00:15 karstensrage hi branches arent working for me and i thought of another idea and i would like thoughts on it
00:15 karstensrage the idea is to tag master with a "starting feature X"
00:16 karstensrage the log - tag date kind of indicates how long master has been good...
00:16 karstensrage and the tag lets you go back to the last known good state
00:20 aidalgol joined #git
00:22 tesuji_ joined #git
00:25 kadoban What's wrong with branches?
00:28 karstensrage they just dont work for me
00:28 karstensrage they sit around too long and get hard to merge
00:28 karstensrage and i forget what they are
00:29 karstensrage they make you think you can work on two features at the same time but in practice its way too messy if the features take a long time
00:31 sangy joined #git
00:32 texinwien joined #git
00:33 karstensrage i mean i think branches are fine
00:33 karstensrage they just arent working for me personally
00:34 Orbitrix joined #git
00:34 Orbitrix joined #git
00:37 raijin joined #git
00:38 Schwarzbaer__ joined #git
00:38 nettoweb joined #git
00:38 sdothum joined #git
00:39 Goplat joined #git
00:41 astrofog joined #git
00:46 hahuang62 joined #git
00:46 Umeaboy left #git
00:47 t2mkn joined #git
00:51 Rodya_ joined #git
00:54 jonath-n joined #git
00:55 nedbat karstensrage: perhaps you just need to rebase them frequently.
00:56 ochorocho__ joined #git
01:00 fstd_ joined #git
01:02 pks joined #git
01:05 ajf- joined #git
01:06 t2mkn1 joined #git
01:12 alex1a joined #git
01:12 solrize joined #git
01:14 solrize crap i typed "git branch master" instead of "git checkout master" then did the same thing for "remotes/origin/master"
01:14 GandalfTheWizard joined #git
01:14 solrize now i have two remotes/origin/master branches
01:14 solrize how do i get rid of the new one?
01:14 rewt git branch -d
01:14 solrize will that get rid of the right one?
01:15 BenderRodriguez joined #git
01:15 rewt yes; you can't delete remote branches with that
01:15 xall joined #git
01:15 solrize but there are two local branches with that name
01:15 rewt the original "remotes/origin/master" is a remote branch
01:15 rewt the one you made is local
01:16 solrize ok i got rid of it
01:16 rewt and you shouldn't really try to checkout a remote branch, as that will result in a detached head
01:16 solrize aha yes that's what i did and that's what happened
01:17 rewt your local master should be tracking remotes/origin/master
01:17 solrize but when i push from master it says there's no remote master branch even though there is one
01:17 rewt `git pull` to update it
01:17 cqi joined #git
01:18 rewt you probably made the local master branch without setting it to track the remote
01:18 solrize hmm
01:18 solrize yeah i have master and remotes/origin/master
01:18 rewt delete your local master, and checkout master
01:19 solrize git branch -d master?  then git checkout master?
01:19 rewt yes
01:19 solrize error: Cannot delete the branch 'master' which you are currently on.
01:19 rewt checkout to something else so you're not on it
01:19 solrize aha ok.  made a new master that's set up to track remote
01:20 solrize git push now says everything up to date, lemme check that on the remote
01:20 rewt don't forget to fetch before pushing
01:20 eka joined #git
01:21 tesuji joined #git
01:21 solrize oh i see what happened with the push, the remote side had that branch checked out
01:21 rewt your local branch may be up to date with the local copy of the remote master, but the remote may have advanced since the last time you fetched/cloned
01:21 solrize fetch to pull new changes?  hm ok
01:21 solrize fetch didn't seem to do anything
01:21 rewt fetch to update local copy of remote repo
01:22 solrize yeah i think the remote side hasn't changed
01:22 solrize the stuff i was doing was on local and i'm the only committer to this thing
01:23 Noldorin joined #git
01:24 codebam hey, so I'm trying to use mergetool to merge commits, when I use it with -t vimdiff its doing a 3 way merge, I'm not sure why though because all that I did was merge from upstream. So it should be a 2 way merge?
01:24 codebam Also, is <<<<<< start and >>>>>> end?
01:24 codebam or the other way around
01:24 solrize that stuff is where there was a merge conflict
01:25 xaviergmail joined #git
01:25 codebam well yes, thats why I am doing the mergetool. but why is it 3 way?
01:25 rewt 3 way merge between local (your branch), remote (other branch), and base (common ancestor)
01:26 solrize anyway thanks i think i can fix this now.  i'm getting tons of compile errors on the remote side but i think that's because it's an older compiler.  the file checksums match
01:26 rewt stuff between <<<<<< to ====== is in one of the branches, and stuff between ====== and >>>>>> is in the other branch
01:29 codebam okay so I'm seeing duplicates of lines and I dont know what to delete and what to keep. now I know I should be looking at the left top and right top. what is the difference between a <<<<<< and a >>>>>>
01:30 uhmgawa^3 joined #git
01:30 osse joined #git
01:30 codebam also what is BASE
01:31 rewt in that 4-pane layout, top left is your local branch before the merge, top right is the other branch before the merge, top middle is the common ancestor, and bottom is the working copy that you can save
01:32 codebam okay but what does common ancestor mean? what is common between the two files?
01:32 rewt <<<<<< and ====== and >>>>>> are just markers to show you where it found conflicts
01:32 rewt base = common ancestor = the file as it was before the 2 branches diverged
01:33 codebam so that would be like what they both have in common?
01:35 rewt no, that's how the file looked before changes were made to it in either branch
01:37 codebam oh so like essentially changes were made to my branch at the same time as changes were being made to the other branch. this is the commit where they were both the same?
01:37 durham joined #git
01:37 rewt yes
01:37 codebam oh cool okay that makes sense
01:37 rewt if there were no changes in one of the branches, it should be identical to that one ... but then there wouldn't be any conflicts either
01:38 codebam so for the most part I can just ignore that because I'm just bringing something up to date with upstream
01:38 durham joined #git
01:39 rewt yeah, it's set up to give you as much info as possible, even if the conflict is easy to resolve
01:39 rewt just edit the bottom pane, remove the conflict markers, and save
01:39 codebam okay cool, :)
01:40 karstensrage nedbat, well yeah there are things i could be doing but am not
01:40 codebam thanks rewt, I get this now I think
01:40 rewt np
01:40 karstensrage so im looking for something that might work better for me
01:41 safrabods joined #git
01:42 rewt karstensrage, some people create a branch when they start a new feature, and when they're done, they do a fast-forward merge, so the history stays linear with no evidence it was branched at all
01:46 Asiajey joined #git
01:46 nedbat karstensrage: have you tried rebasing?
01:46 T4nkAbb0t joined #git
01:46 mehola joined #git
01:49 ilbot3 joined #git
01:49 Topic for #git is now Welcome to #git, the place for git help and 20% time | Public logs at http://goo.gl/BuUi5o | Current stable version: 2.9.0 | First visit? Read: http://jk.gs/git | Getting "cannot send to channel"? /msg gitinfo .voice | git stash drop: Hey! That stuff's expensive.
01:51 karstensrage nedbat, the problem is i have branches and i have no idea what they are
01:51 nedbat karstensrage: how will tagging master help with that?
01:51 karstensrage so im going to delete them and hope it was nothing important
01:52 karstensrage nedbat, cause going forward i dont want to get into messes with multiple branches
01:52 nedbat karstensrage: so now you will hack on master, and lose track of what you were doing, and then what?
01:52 ajf- joined #git
01:52 karstensrage i never lose track of master
01:52 Wulf4 joined #git
01:52 nedbat karstensrage: why are branches hard to keep track of?
01:53 karstensrage master gets pushed to production its always whats running
01:53 rewt if you're not sure of what went into a branch, just diff it against master
01:53 liuzhen joined #git
01:57 nedbat karstensrage: how will you know when master is ok to push to production if you're doing feature development on it?
02:00 karstensrage Cause I'll wait until the feature is done
02:00 pks joined #git
02:00 zumba_addict joined #git
02:01 nedbat karstensrage: it sounds like you have a plan. good luck with it.  i don't think branches are your problem. :)
02:01 drbean_ joined #git
02:03 karstensrage Hmm you don't like the idea. Not sure what to do I've tried branches and it's not working for me
02:03 nedbat karstensrage: try the new plan, it may be good for you.
02:03 karstensrage Ok
02:04 mic4ael1 joined #git
02:04 nedbat karstensrage: at least you will learn more about your own development style
02:06 codebam okay so now that I've done all that, how do I merge while keeping ownership of all the files that weren't conflicts?
02:06 codebam because if I git add * and git commit -m "merged upstream" it wont keep ownership?
02:07 kadoban "ownership"? git doesn't have a concept of ownership for files. What are we talking about, the filesystem?
02:08 codebam like who wrote the original commit on those files
02:09 jonath-n joined #git
02:09 codebam if I just add * and commit it wont keep that history
02:10 rewt the merge commit will have your name on it, but the original commits will keep the original committers' names
02:10 kadoban codebam: git is pretty good at recording both the author and the commiter of changes, piece at a time. That's a lot of what it does. Unless you're doing something very strange, you don't even have to worry about it.
02:11 codebam oh okay, so I can just add and commit?
02:11 rewt yes
02:11 codebam oh cool thanks
02:11 Khisanth joined #git
02:15 chachasmooth joined #git
02:16 Brain joined #git
02:17 zturner_ joined #git
02:18 fracting joined #git
02:25 solrize left #git
02:27 durham joined #git
02:36 zumba_addict joined #git
02:39 livingstn joined #git
02:41 sauvin joined #git
02:41 fahadash joined #git
02:46 Goplat joined #git
02:46 mingrammer_ joined #git
02:47 ahr3n joined #git
02:47 hexagoxel joined #git
02:49 jceb joined #git
02:50 robot joined #git
02:54 arescorpio joined #git
02:54 eletuchy joined #git
02:55 watabou joined #git
02:58 ochorocho__ joined #git
02:59 mwhooker joined #git
03:01 Lyqyd joined #git
03:03 durham joined #git
03:03 phanimahesh joined #git
03:09 aavrug joined #git
03:10 backnforth joined #git
03:10 zturner_ joined #git
03:17 fscala joined #git
03:18 xall joined #git
03:19 ki0 joined #git
03:20 jacob_ joined #git
03:20 jacob_ left #git
03:20 phanimahesh joined #git
03:21 k3rn31 joined #git
03:24 eletuchy joined #git
03:25 backnforth joined #git
03:30 xelxebar joined #git
03:30 eletuchy joined #git
03:32 eletuchy_ joined #git
03:35 eletuch__ joined #git
03:38 mingrammer_ joined #git
03:38 dreiss joined #git
03:38 phanimah1sh joined #git
03:41 ki0 joined #git
03:48 boombatower joined #git
03:48 sagerdearia joined #git
03:50 robotroll joined #git
03:55 rgrinberg joined #git
03:55 blackwind_123 joined #git
03:55 Xz joined #git
03:56 t2mkn joined #git
03:56 Xz hi there, I have a git repo hosted outside of github. Now I want to fork that onto github - how do I do that?
03:56 Xz I have that repo checked out
03:56 Xz in other words I need to push all the branches to my new github repo somehow
03:56 eletuchy joined #git
03:56 Xz and all tags, everything
03:57 rewt add github as a remote and push
03:58 Xz rewt: but push what exactly?
03:58 rewt everything
03:58 Xz push github *
03:58 Xz ?
03:59 rewt no, you don't push files
03:59 rewt you push the repo
03:59 Xz rewt: my remotes now are 'origin' and 'github'
03:59 Xz rewt: git push github origin?
03:59 Xz rewt: I have never pushed everything
03:59 Xz rewt: i know how to push single branch, or tag
03:59 rewt same as pushing anything
03:59 rewt !git push
04:00 rewt !man git push
04:00 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 !) ?
04:00 gitinfo the git-push manpage is available at http://jk.gs/git-push.html
04:00 rewt git push --all github
04:00 Xz rewt: --all, that's the trick, thanks
04:01 ki0 joined #git
04:02 Xz rewt: that didn't push all...
04:02 Xz rewt: there is plenty of remote branches it didn't even try to push to new remote
04:04 Xz rewt: i guess it only works on local stuff
04:05 Xz rewt: git push origin --mirror seems to do the job better
04:05 Xz rewt: git push github --mirror I meants
04:06 toogley joined #git
04:06 k3rn31 joined #git
04:07 Cabanossi joined #git
04:14 madewokherd joined #git
04:15 Ahad joined #git
04:16 rkazak joined #git
04:19 FuzzySockets joined #git
04:20 Ahad left #git
04:20 zturner_ joined #git
04:20 shgysk8zer0 joined #git
04:23 ki0 joined #git
04:27 _ng joined #git
04:27 eka joined #git
04:31 mingrammer_ joined #git
04:36 durham joined #git
04:36 hahuang62 joined #git
04:38 subhojit777 joined #git
04:42 WayToDoor joined #git
04:42 fracting joined #git
04:43 jungsubk joined #git
04:44 cqi_ joined #git
04:47 ajf- joined #git
04:48 fracting joined #git
04:48 cyan__ joined #git
04:49 Codfection joined #git
04:51 d0nn1e joined #git
04:54 backnforth joined #git
04:56 Schwarzbaer__ joined #git
04:58 liuzhen joined #git
04:59 ochorocho__ joined #git
05:05 cyphase joined #git
05:08 backnforth joined #git
05:10 blackwind_123 joined #git
05:10 howdoi joined #git
05:11 cqi joined #git
05:11 zumba_ad_ joined #git
05:13 mda1 joined #git
05:14 elementalest joined #git
05:16 stuh84 joined #git
05:24 emPi joined #git
05:25 sostenuto joined #git
05:27 catsup joined #git
05:27 _28_ria joined #git
05:29 nilg joined #git
05:30 fracting joined #git
05:32 Junior joined #git
05:33 zturner_ joined #git
05:34 mwhooker joined #git
05:38 hahuang62 joined #git
05:40 bgamari joined #git
05:40 mehola joined #git
05:41 cyphase joined #git
05:42 sword joined #git
05:44 ToBeCloud joined #git
05:48 xanadu_ joined #git
05:50 cyphase joined #git
05:50 ThomasLocke joined #git
05:51 JanC_ joined #git
05:54 albel727 joined #git
05:57 mehola joined #git
05:58 pwnz0r joined #git
05:58 Raging_Hog joined #git
05:59 cyphase joined #git
06:01 liuzhen joined #git
06:03 milki pushing remote branches is not a normal thing to do o.O
06:04 mariuscc joined #git
06:06 fracting joined #git
06:07 jason237 joined #git
06:09 ahmedelgabri joined #git
06:10 tango_ joined #git
06:11 cyphase joined #git
06:12 vuoto joined #git
06:15 D-Boy joined #git
06:16 checkItOut joined #git
06:17 nevodka joined #git
06:19 cbreak Xz: careful with --mirror
06:20 lindenle joined #git
06:21 xall joined #git
06:22 Limes joined #git
06:22 Oebele joined #git
06:23 absk007 joined #git
06:25 eijk joined #git
06:25 paul424 joined #git
06:28 zeroed joined #git
06:28 zeroed joined #git
06:29 ki0 joined #git
06:32 tonyj joined #git
06:33 JeroenT joined #git
06:37 Limes joined #git
06:38 semantipc joined #git
06:39 yuhlw joined #git
06:40 JeroenT joined #git
06:48 ewet joined #git
06:52 p1k hello
06:52 ki0 joined #git
06:52 p1k so normally I can see the differences of just the relevant files after a rebase with git diff <remote> (git diff HEAD~N HEAD --name-only)
06:53 p1k but if files were deleted locally during the rebase that exist in the remote git will complain about specifying a path which doesn't exist locally
06:53 p1k is there a way to ignore that ?
06:53 ki0 joined #git
06:54 zeemz joined #git
06:55 cyphase joined #git
06:57 barteks2x joined #git
06:58 LeMike joined #git
06:58 thiago p1k: define ignore
06:58 thiago p1k: do you want to tell git to match the file that was deleted?
06:58 lordjancso joined #git
06:58 thiago p1k: or do you want to ignore the files that were deleted?
07:00 ochorocho__ joined #git
07:00 k3rn31 joined #git
07:02 Fijit joined #git
07:03 Zarthus joined #git
07:03 encod3 joined #git
07:05 rchavik joined #git
07:05 sgfgdf joined #git
07:07 Sceorem joined #git
07:09 Hounddog joined #git
07:09 Sceorem_ joined #git
07:10 grift joined #git
07:11 Kicer86 joined #git
07:11 cdown joined #git
07:11 Sceorem__ joined #git
07:11 sgfgdf hello, guys! i have a bad branch (with more than one commit) merged in the develop branch, which is merged in master when deploying. what is the best way to remove the bad commits from develop?
07:11 grawity has it been merged into master yet?
07:11 moritz sgfgdf: revert them
07:12 p1k thiago: Yeah I'd like to match the files even though they don't have a local path
07:12 moritz (unless they introduce huge files that blow up the repo)
07:13 p1k otherwise I suppose something like ( (git diff HEAD~N HEAD --name-only)  | xargs ls -d 2>/dev/null) would work
07:19 liuzhen joined #git
07:19 Rish joined #git
07:20 neanderslob joined #git
07:20 neanderslob_ joined #git
07:20 xall joined #git
07:21 grift joined #git
07:21 neanderslob joined #git
07:22 zturner_ joined #git
07:25 Xethron joined #git
07:25 Sleepy63 joined #git
07:26 sgfgdf grawity: not yet.
07:26 grawity go nuts with `git rebase -i` then
07:26 sgfgdf moritz: how should i revert them?
07:26 sgfgdf grawity: but they are public in develop.
07:27 moritz sgfgdf: git revert <sha1>
07:27 sgfgdf grawity: isn't rebasic in public branches bad idea?
07:27 sgfgdf *rebasing
07:27 fracting joined #git
07:27 mouseover joined #git
07:28 kadoban sgfgdf: Usually, yes.
07:29 mouseover newcomer to git (long time mercurial user): I am using meld as my diff tool. Now, when I do `git difftool`, the modified files are shown individually one at a time. In mercurial, I could navigate and select the modified files. How can I do this with git?
07:29 Balliad joined #git
07:30 fscala joined #git
07:30 xall_ joined #git
07:33 JeroenT joined #git
07:35 grift joined #git
07:35 jaguarmagenta joined #git
07:38 cdown_ joined #git
07:38 leeN joined #git
07:39 nnyk_ joined #git
07:40 cyphase left #git
07:40 newlifer joined #git
07:41 grift joined #git
07:42 mehola joined #git
07:42 mehola joined #git
07:43 sgfgdf kadoban: so i cannot do rebase in develop and removemy bad commits if i don't want to have problems with commit history?
07:43 kadoban sgfgdf: If you're already shared/published develop, you shouldn't.
07:43 kadoban (usually)
07:43 kadoban Sometimes it's still the least bad of bad options.
07:44 frogonwheels joined #git
07:44 afuentes joined #git
07:46 kanzure joined #git
07:46 nevyn win 3
07:46 mehola joined #git
07:47 multi_io joined #git
07:49 absk007 joined #git
07:50 absk007 joined #git
07:50 grift joined #git
07:51 settermjd joined #git
07:53 a_thakur joined #git
07:59 jbitdrop joined #git
08:00 sctskw joined #git
08:03 tvw joined #git
08:03 kanzure joined #git
08:04 barteks2x joined #git
08:07 imjacobclark joined #git
08:07 MrMojito joined #git
08:07 rominronin joined #git
08:07 dsdeiz joined #git
08:08 eletuchy joined #git
08:09 jknetl joined #git
08:09 elementalest joined #git
08:11 kedare joined #git
08:11 a_thakur joined #git
08:11 _pix joined #git
08:13 _pix joined #git
08:13 _28_ria joined #git
08:15 mehola joined #git
08:16 _pix joined #git
08:17 eletuchy_ joined #git
08:17 _pix joined #git
08:18 marcogmonteiro joined #git
08:19 FrontPsych joined #git
08:19 LeMike joined #git
08:20 cyan__ joined #git
08:21 eletuchy joined #git
08:22 mithenks joined #git
08:23 eletuchy_ joined #git
08:24 sanketdg joined #git
08:25 keltvek joined #git
08:26 eletuchy joined #git
08:28 JeroenT joined #git
08:28 Orion3k joined #git
08:29 Xz left #git
08:31 hackal joined #git
08:31 hackal !deploy
08:31 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
08:33 eletuchy joined #git
08:36 feignix joined #git
08:36 weckl joined #git
08:37 eletuchy_ joined #git
08:39 mikecmpbll joined #git
08:40 eletuchy joined #git
08:40 Segfault_ joined #git
08:41 eletuch__ joined #git
08:43 Meteorhead joined #git
08:43 eletuchy_ joined #git
08:44 Chill_Surf joined #git
08:45 qwebirc860673 joined #git
08:45 ws2k3 joined #git
08:46 JeroenT joined #git
08:47 GavinMagnus joined #git
08:49 watabou joined #git
08:49 acetakwas joined #git
08:49 mooru joined #git
08:50 mischat joined #git
08:50 Keridos joined #git
08:51 BlackPanx joined #git
08:51 JeroenT joined #git
08:52 acetakwas joined #git
08:54 psprint_ joined #git
08:57 ocbtec joined #git
08:58 hhee joined #git
08:59 GavinMagnus left #git
09:00 ochorocho__ joined #git
09:02 shgysk8zer0 joined #git
09:02 eijk joined #git
09:04 mischat joined #git
09:04 encod3_ joined #git
09:06 JeroenT joined #git
09:08 settermjd joined #git
09:08 King_Hual joined #git
09:08 cdown joined #git
09:09 mariuscc2 joined #git
09:11 dcpc007 joined #git
09:11 guardianJ joined #git
09:12 vuoto joined #git
09:12 notebox joined #git
09:12 texinwien joined #git
09:17 circ-user-ZHfj9 joined #git
09:20 eletuchy joined #git
09:20 Kingsy joined #git
09:20 infra-red joined #git
09:20 khelkun joined #git
09:21 acetakwas What does this line mean: `git pull -s recursive -X theirs upstream master`
09:21 Kingsy can someone tell me why var/ is still appearing in my git status to be added with this .gitignore --> http://hastebin.com/fixulesaye.coffee
09:21 Kingsy its the stock gitignore from a framework..
09:21 digidog joined #git
09:22 khelkun joined #git
09:22 acetakwas Kingsy::  `var/` != '/var/'
09:22 Kingsy hmm actually
09:22 Kingsy perhaps I need to ask the framework
09:22 JeroenT joined #git
09:23 acetakwas Kingsy::  By the way, if I simply need to ignore a folder in the root directory of my repo, I just type the name like so: [name]
09:24 acetakwas Without the square brackets of course
09:24 acetakwas Notice there's no trailing forward slash
09:24 fstd joined #git
09:25 MrMojito joined #git
09:27 ferr1 joined #git
09:29 a_thakur joined #git
09:29 mehola joined #git
09:32 zturner_ joined #git
09:33 MrWoohoo joined #git
09:34 infinite joined #git
09:35 _enigma_ joined #git
09:38 gajus I don't understand the purpose of the 'git rebase --skip'. I am rebasing my branch on to develop and I get about ~20 conflicts (out of 50) tat require me to simply `git rebase --skip`
09:39 selckin i believe skip drops/removes that commit
09:39 selckin so probably not what you want
09:39 JeroenT joined #git
09:40 acetakwas joined #git
09:40 gajus well, what options do I have?
09:40 selckin fix the conflicts
09:40 selckin !mergetool
09:40 gajus http://paste.linux.chat/view/6acce877
09:40 jast sometimes if the same change(s) have been applied on the base branch, --skip is the right thing to do
09:40 gajus there are no conflicts
09:41 jast if, after fixing conflicts and using --continue, rebase tells you you probably need --skip, it tends to be right about that
09:41 gajus jast is there a way to automate this?
09:42 Timvde joined #git
09:43 ahmedelgabri joined #git
09:43 jast git already automatically skips commits if it can detect that they can be skipped
09:43 jast unfortunately it's not always possible
09:43 jast FWIW this only happens a lot if you copy commits around a lot, e.g. using cherry-pick
09:44 gajus that is the case here
09:44 jast or if you use squash merges, the bane of merging and rebasing
09:44 jast right, in that case you probably will have to live with this
09:44 gajus but, when I run into this message, it is not like I have any other option than to `git rebase --skip` or `git rebase --abort`, then why not default to skip?
09:44 selckin because it doesn't like to lose data
09:44 jast well maybe you made a mistake fixing the conflict?
09:45 jast in which case it's a good thing git doesn't just go ahead and throw everything away
09:45 jast this is not *supposed* to be a regular occurrence, so git errs on the side of being careful
09:45 gajus are we really talking about this message? http://paste.linux.chat/view/6acce877
09:46 jaguarmagenta joined #git
09:46 gajus note that `git add -A; git rebase --continue` is for a previous conflict, not the current
09:47 dsdeiz joined #git
09:47 dsdeiz joined #git
09:54 Sylvain31 joined #git
09:56 Sylvain31 hi, how to manage two issue working on two feature in two branch? I would like to have a branch with both, do I need to have a third branch with merged branches?
09:56 MrMojito_Busy joined #git
09:57 liuzhen joined #git
09:57 feignix joined #git
09:58 jhass yes
09:59 gajus whats the simplest way to list all unresolved conflicts?
09:59 grawity git status
09:59 watabou joined #git
09:59 gajus yeah, that shows all modified files, etc.
09:59 gajus I need basically, `grep -H -r "<<<<<<< HEAD"`
10:00 grawity scripting or human consuption?
10:00 gajus human consumption
10:01 Sylvain31 jhass: and if I have more feature and more branch, I can maintain a merged branch with everything merged from each?
10:02 jhass sure
10:03 MrMojito joined #git
10:03 eletuchy joined #git
10:05 a3Dman joined #git
10:07 Sylvain31 how to see from  where branch did I forked a branch?
10:07 grawity git merge-base --fork-point ...
10:08 jaguarmagenta joined #git
10:11 DomeMaster joined #git
10:11 Timvde joined #git
10:12 leandrosansilva joined #git
10:16 leandrosansilva Hello guys, I have some code on a branch B which was forked from a branch A. Developement on both branches continued, but on branch B the commits were squashed into a single one, including the ones also used on A. Now I want to rebase B on A, but git obviously complain and is unable to merge the conflicts. Is there a clean way of solving this issue?
10:17 leandrosansilva and rebasing B on A normally?
10:17 leandrosansilva Oh, I am on git 2.7.4
10:17 ahmedelgabri joined #git
10:18 gajus I am utterly confused. I have a branch 'A1' and 'A2'. 'A2' branched from 'A1'. I had to make some changes to 'A1'. I have rebased 'A1' onto 'develop'. I have merged 'A1' into 'develop'.
10:18 cdown joined #git
10:18 gajus Now I have rebased 'A2' on 'develop'
10:18 gajus but I do not see commits that have been applied to 'A1'?
10:18 gajus I don't even see them in the 'develop' branch
10:18 Macaveli joined #git
10:19 gajus oh
10:19 gajus All changes have been squashed into a single commit when merging.
10:20 mdw joined #git
10:20 fstd joined #git
10:20 jast leandrosansilva: no, you'll have to brave the conflicts
10:21 jast i.e. a happy day of manual resolution
10:21 KeelOfSteel joined #git
10:22 rgrinberg joined #git
10:22 leandrosansilva jast: thx, it'll be a long day :-)
10:22 texinwien joined #git
10:24 rgrinberg joined #git
10:27 fees joined #git
10:28 anth0ny joined #git
10:30 encod3 joined #git
10:30 diegoaguilar joined #git
10:30 ams__ joined #git
10:33 cdown joined #git
10:33 jaguarmagenta joined #git
10:33 sdothum joined #git
10:35 Repox joined #git
10:36 wodim joined #git
10:37 wodim hello, I have two commits with other commits in between, I'd like to merge them into one only commit. it should be possible (the commits in between haven't touched that file). what's the name of this procedure, and can somebody link me to a good tutorial about this matter?
10:37 emPi joined #git
10:38 _ikke_ wodim: interactive rebase + squashing
10:40 wodim ah, sounds like what I need, thank you.
10:42 wodim oh, it failed.
10:44 wodim I'm picking the oldest commit, and then squashing the more recent ones
10:45 jaguarmagenta joined #git
10:45 ahmedelgabri joined #git
10:46 j416 wodim: is there still a question?
10:49 Keridos joined #git
10:49 cdown joined #git
10:49 JeroenT joined #git
10:50 wodim yes.
10:50 wodim https://paste.pound-python.o​rg/raw/kMYplAwP1zHzapmnS1Iu/
10:50 wodim I'd like to make it so those three commits are a single one, but I don't know how git rebase works, I guess?
10:51 _ikke_ What failed?
10:51 cdown joined #git
10:51 _ikke_ Ah, merge conflict
10:52 wodim those commits in between do not even touch the file
10:53 jast wodim: did you remove any other lines in the editor? because anything you remove will actually be removed and then later commits might not apply cleanly...
10:53 wodim the editor, by default, comes with a "noop"
10:54 jast how did you run rebase --interactive? any args?
10:54 wodim no
10:55 jast okay, in that case it should list all commits "newer than" wherever the upstream branch is at
10:55 jast if that does not include the commits you want to squash, you may need an argument, e.g. the most recent commit that is an ancestor of the commits you want to squash
10:56 jast in your case that seems to be 'git rebase -i 6bb3da6' or simply 'git rebase -i e673f29^' which comes out to the same thing
10:56 johnmilton joined #git
10:57 ahmedelgabri joined #git
10:58 benycze joined #git
10:59 wodim what you mean is that the entire git log (at least until that commit) should have appeared in the editor?
10:59 Firzen joined #git
11:00 wodim is this what I want, then? https://paste.pound-python.o​rg/raw/kwQjD5DIOZYSShammpfW/
11:01 ochorocho__ joined #git
11:01 diegoaguilar joined #git
11:01 zarel joined #git
11:02 Firzen I did a git checkout and it aborted with a fatal error, because it couldn't create a directory. It now says I'm still in the branch I wanted to switch from, but it created some files. Are these the files it created for switching to the other branch until it hit the error? Why didn't it clean up those after the error? And how would I get rid of them now to retry to checkout?
11:02 watabou joined #git
11:06 fracting joined #git
11:06 Panacea joined #git
11:06 Firzen Okay, it appears checkout -f did the trick.
11:07 tobiasvl perhaps you overwrote some local changes now
11:08 Firzen Nah, I had a clean working dir before.
11:09 Firzen But I got a bit confused. I thought git would clean up half finished business in case of error.
11:10 airmac joined #git
11:10 tobiasvl git doesn't like throwing away stuff for you
11:10 chrfle joined #git
11:11 chrfle Is it possible to make 'git fetch' always fetch all remotes so I can omit the --all every time I fetch?
11:12 airmac joined #git
11:13 airmac joined #git
11:15 wodim jast: ah, I could make it work, I was supposed to reorder the commits. Thank you
11:15 _pix joined #git
11:16 airmac joined #git
11:17 airmac left #git
11:17 cdown joined #git
11:18 benycze left #git
11:18 xnrand joined #git
11:18 mkjackson joined #git
11:18 JeroenT joined #git
11:18 Firzen chrfle: I don't know, but maybe an alias would help you as well. Like git config alias.fall 'fetch --all' to have git fall fetch all. add --global to add it to your user config (or --system for system config).
11:23 chrfle Firzen: true.. I'll do that, thanks
11:24 sol56 joined #git
11:25 Xethron joined #git
11:26 ahmedelgabri joined #git
11:26 mkjackson joined #git
11:27 imjacobclark joined #git
11:27 MrMojito_Busy joined #git
11:27 diegoaguilar joined #git
11:28 imjacobclark joined #git
11:30 sol56 joined #git
11:32 sol56 joined #git
11:33 diegoaguilar joined #git
11:33 psprint I'm writing a software that will handle subset of allowed branch names. Currently [a-zA-Z0-9_-]. Would you recommend any more character, is there additional popular one?
11:33 encod3 joined #git
11:35 settermjd joined #git
11:35 Seveas psprint: / is very popular
11:35 ojdo psprint: this answer mentions the '/' as a separation character http://stackoverflow.com/q/273695/2375855
11:36 grawity psprint: huh? the branch names are *already* rather limited, why bother subsetting it further?
11:37 ojdo psprint: it seems that git-check-ref-format can be use to check whether a ref (branch name) is acceptable for git
11:38 psprint Seveas: thanks
11:39 Seveas in the main repo at work, all those characters are used plus the .
11:41 rnsanchez joined #git
11:42 xnrand joined #git
11:43 fscala joined #git
11:43 same_ol_geez joined #git
11:44 Atlantic777 joined #git
11:44 Atlantic777 joined #git
11:45 tesuji_ joined #git
11:45 choki joined #git
11:52 eletuchy joined #git
11:53 reactormonk left #git
11:53 aard joined #git
11:53 espindola joined #git
11:54 Anticom joined #git
11:54 SirOliver joined #git
11:55 gelei007 joined #git
11:55 encod3 joined #git
11:56 nnyk_ joined #git
11:57 gelei007 hi, guys,i have a question to ask,hope someone help me ,what does git rebase do, my english is not very good,i cant understand what the office file explanation,hope someone help me
11:58 Sceorem__ joined #git
11:59 tobiasvl gelei007: did you read man git rebase
11:59 gitinfo gelei007: the git-rebase manpage is available at http://jk.gs/git-rebase.html
11:59 tobiasvl gelei007: in simple terms, a rebase takes a commit and changes which commit it is based on. hence "re-base". it can be very powerful, see also !interactive_rebase
11:59 gitinfo gelei007: Interactive rebase sounds similar to rebase but has completely different abilities. It can do this to commits: change the order, squash some of them together, remove some, add random existing commits from other branches, split them, and more... it's very powerful. Documentation is in the section "Interactive Rebase" in 'man git-rebase'.
11:59 ahmedelgabri joined #git
12:00 feignix joined #git
12:01 d0nkeyBOB joined #git
12:03 guardianJ joined #git
12:04 synthroid joined #git
12:07 Sylvain31 how to import all change to a new branch like for a merge, but witout merging, so I can discard change I don't want before commit?
12:07 tobiasvl Sylvain31: git merge --no-commit ?
12:09 Sylvain31 tobiasvl: that was that, thanks. :)
12:11 newlifer_ joined #git
12:12 bluezone joined #git
12:12 ahmedelgabri joined #git
12:13 belak joined #git
12:13 JeroenT_ joined #git
12:14 boudicca joined #git
12:14 boshhead joined #git
12:14 nowhere_man joined #git
12:14 t-mart_ joined #git
12:15 Balliad joined #git
12:15 marko___ joined #git
12:16 sweatsuit joined #git
12:16 fmcgeough joined #git
12:16 johnmilton joined #git
12:16 ndim joined #git
12:16 blackwind_123 joined #git
12:17 bluesilver joined #git
12:17 joki joined #git
12:18 Sylvain31 tobiasvl: is it possible not to add file too?
12:18 alip joined #git
12:18 MikeChernev joined #git
12:18 tobiasvl Sylvain31: !unstage ?
12:18 gitinfo Sylvain31: To unstage a file from the index(in other words, to undo an accidental `git add foo/bar.txt`), use: git reset -- foo/bar.txt
12:19 tmz joined #git
12:19 rawtaz joined #git
12:19 christophler joined #git
12:21 rmrrn joined #git
12:21 rmrrn joined #git
12:21 SebastianFlyte joined #git
12:21 Karunamon joined #git
12:22 e-dard_afk joined #git
12:23 hfp joined #git
12:23 Ram-Z_ joined #git
12:23 topi` joined #git
12:23 Sylvain31 of course, thanks, I mean not to add any file after merging changes (--no-stage like…)
12:24 albel727 joined #git
12:24 nilg` joined #git
12:25 DomeMaster joined #git
12:26 tobiasvl I dunno, why do you want that? git marks the merge conflicts as resolved by staging
12:26 larsrustand joined #git
12:27 Hello71 joined #git
12:27 Sylvain31 to recreate a clean branch from part spread into some other branch, I just want the final code, but may be some folder, not the full branch
12:28 rwb joined #git
12:28 SirOliver joined #git
12:29 DoniRS joined #git
12:29 aguimaraes1986 joined #git
12:30 toogley joined #git
12:31 imjacobclark joined #git
12:32 feignix joined #git
12:32 livingstn joined #git
12:32 mda1 joined #git
12:32 sharpy joined #git
12:33 Firzen Sylvain31: Maybe with a clean working dir, do the merge with --no-commit, then git reset HEAD~1, which should reset the index to its state before the merge, then add the merged files that are still in your working dir as you desire.
12:33 wviana joined #git
12:35 _W_ joined #git
12:35 Firzen Or move all changed (/merged) files to a directory, if you want that, commit that directory and then checkout HEAD to get the old unmerged files back at where they were.
12:35 Firzen (after doing the above)
12:37 toogley joined #git
12:37 Wulf joined #git
12:37 rahtgaz joined #git
12:37 infra-red joined #git
12:39 Firzen Sylvain31: git reset HEAD after the merge, I'm sorry.
12:41 armyriad joined #git
12:41 psprint How to list branches of a remote repo without cloning it?
12:41 subhojit777 joined #git
12:41 grawity git ls-remote
12:42 danchr joined #git
12:42 danchr joined #git
12:43 shinnya joined #git
12:43 mikeBr joined #git
12:45 mattp_ joined #git
12:46 eletuchy joined #git
12:46 goTAN joined #git
12:47 jason237 joined #git
12:48 xcesariox joined #git
12:50 mariuscc joined #git
12:53 cdown joined #git
12:56 italoacasas joined #git
12:56 whaley joined #git
12:58 shaggycat joined #git
13:02 Khisanth joined #git
13:02 ochorocho__ joined #git
13:02 CEnnis91 joined #git
13:03 Silenced joined #git
13:04 al-damiri joined #git
13:04 docnvk joined #git
13:05 ahmedelgabri joined #git
13:07 ShekharReddy joined #git
13:08 d0nkeyBOB joined #git
13:11 johnny56 joined #git
13:13 imjacobclark joined #git
13:13 jimi_ joined #git
13:13 tinanoo joined #git
13:14 Marbug joined #git
13:15 Godfather_ joined #git
13:15 Godfather_ hi
13:16 Godfather_ if i have a conflict in file1 after a merge, how can i use a merge union only in this file ?
13:17 Godfather_ its a large file, and i want to use both parts
13:17 Marbug Hi I have some git-repo in a server, and I configured rights so people can push and pull, but can't write something specific in the .git dir. although, when you push we get "fatal: Unable to create '.git/packed-refs.lock': Permission denied" can that be a problem? and is there a way to solve it besides giving write rights on the root dir? We have that repo for 2 years now but someone is asking to make it
13:17 Marbug writable
13:19 tabrez joined #git
13:19 ash_workz joined #git
13:20 Oebele joined #git
13:20 tobiasvl you want git to be able to write to .git
13:21 Dougie187 left #git
13:21 Marbug but is it something you want? if you just want people to push and pull?
13:22 imjacobclark joined #git
13:22 Silenced joined #git
13:22 toogley joined #git
13:23 M-sbts joined #git
13:23 rnsanchez_ joined #git
13:24 tobiasvl if you want them to pull, then you need to be able to write to .git
13:24 tobiasvl obviously
13:24 cqi joined #git
13:24 CheckDavid joined #git
13:25 troulouliou_div2 joined #git
13:25 Hounddog joined #git
13:27 semantipc joined #git
13:27 Godfather_ After a merge i have a file with conflicts. I want to resolve all of the using union (use both sides). How can i do this automatically by console?
13:28 tobiasvl union?
13:28 tobiasvl is that a git merge strategy?
13:28 mhinz joined #git
13:28 CJKinni joined #git
13:28 Codfection joined #git
13:29 Pioupiou joined #git
13:29 watabou joined #git
13:30 BrianBlaze420 joined #git
13:30 CJKinni joined #git
13:30 King_Hual joined #git
13:30 Arsonide joined #git
13:30 jamiejackson joined #git
13:31 Marbug tobiasvl, you mean if you want them to push :). But I have set group permissions on multiple files, just not the .git dir so nothing can be changed there. Though when you push it wants to make the packed-ref.lock file, and that isn't possible because the .git isn't writable (other files and dirs have sometime write permissions when needed)
13:31 Noldorin joined #git
13:31 basiclaser joined #git
13:32 tobiasvl Marbug: oh, you mean push _to_ the repo
13:32 tobiasvl is it a bare repo?
13:32 Marbug yes indeed tobiasvl :)
13:33 Marbug how you mean a bare repo?
13:33 frem_ joined #git
13:33 tobiasvl "bare
13:33 tobiasvl !bare
13:33 gitinfo A bare repository is used to push/fetch(useful for running a git server), and contains only the contents of .git/ from a "normal" repo. Read more: http://bare-vs-nonbare.gitrecipes.de/
13:33 Sc0rpius was it setup with git init --bare ?
13:33 tobiasvl since you say it has a .git repo I assume it's not bare
13:33 Marbug indeed tobiasvl :)
13:33 Marbug no no
13:33 tobiasvl it probably should be though
13:33 Marbug it's a bare repo
13:33 Sc0rpius In my mind, a bare repo is a repo, and a normal repo is a working directory.  I think Linus really chose very wrong the names.
13:34 tobiasvl Sc0rpius: uh, isn't that exactly what a bare repo and normal (non-bare) is?
13:34 Marbug so it's a central repo where people can push their changes and pull branches from
13:34 tobiasvl Marbug: oh, ok. but you have .git directory?
13:34 Sc0rpius but calling a normal repo a repo is wrong
13:34 selckin because you're thinking in svn terms
13:34 Marbug yes I have
13:34 Sc0rpius it should be "working directory" not "repo"
13:35 Sc0rpius that's why I said Linus chose badly
13:35 Marbug the names are always confusing to me :)
13:35 selckin in the git flow linus uses there is no repo then
13:35 Sc0rpius they are the worst names ever.
13:35 tobiasvl Sc0rpius: sure. so what is colloquially called a non-bare repo is a working directory + a repo. a bare repo is a repo without a working directory. seems ok by me.
13:36 Marbug so that is sorted out, it's a bare repo ;-)
13:36 Sc0rpius ok :)
13:36 newlifer joined #git
13:36 tobiasvl all right. so which directory do you not have group permissions on?
13:37 dsdeiz joined #git
13:37 circ-user-eHrBN joined #git
13:38 Marbug tobiasvl, maybe I'll pastebin the ls and the commads I used to set the permissions?
13:38 Marbug it'll be easier to clarify :)
13:38 M-shine joined #git
13:38 M-mistake joined #git
13:38 M-meznak joined #git
13:39 nettoweb joined #git
13:39 fracting joined #git
13:39 tobiasvl sure
13:40 Marbug tobiasvl, http://apaste.info/kco
13:40 eletuchy joined #git
13:41 Marbug so tobiasvl as you can see, .git has no group writes, so packed-refs.lock can't be written to the root dis as there are no permissions to write on that dir
13:42 mahendra_ joined #git
13:42 ysyk joined #git
13:42 moritz !gitolite
13:42 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/
13:42 testild joined #git
13:43 moritz it is much easier to set up reliably than group permissions
13:43 selckin indeed recommand gitolite aswel
13:43 lb +1 gitolite
13:43 j416 gitolite++
13:44 feignix_ joined #git
13:44 Marbug okido will look into it later on, seems good :) as I want to integrate gitlab too so maybe both things can be done at the same time
13:45 moritz gitlab has permissions management similar to gitolite
13:45 moritz so if you use gitlab for the repos, no need for gitolite
13:45 Marbug but what you someone suggest for the packed-refs.lock atm? I assume let it be as it isn't doing any harm?
13:45 roundabout22 joined #git
13:46 Marbug oh nice, so gitlab has it's own git server ?
13:46 j416 gitlab does not have nearly as fine-grained control as gitolite, but it may be enough.
13:46 j416 we moved from gitolite to gitlab to get something to do code reviews in, it works ok.
13:46 moritz Marbug: yes, though it reuses the ssh daemon, afaict
13:46 moritz (which is a good thing :-)
13:47 Marbug ait :)
13:48 Marbug and migration from just a "bare" git repo to gitolite is easy I assume? or do you need to start over with the git ?
13:48 testild joined #git
13:49 moritz migration is easy
13:51 reactormonk joined #git
13:51 reactormonk How do I list all remote branches?
13:51 Marbug and for now, do you think I might set group write permissions on the .git dit itself? or leave it like it is? (I just want to know what you guys would do :))
13:51 phanimahesh joined #git
13:51 Marbug git branch -a reactormonk
13:52 osse git branch -r to be more specific
13:53 DARSCODE joined #git
13:53 T4nkAbb0t joined #git
13:56 mmlb joined #git
13:56 mmmveggies joined #git
13:57 charlesr joined #git
13:57 EduardoMartins joined #git
13:58 sangy joined #git
14:00 kaldoran joined #git
14:01 T4nk5bb0t joined #git
14:02 e3c joined #git
14:03 phaleth joined #git
14:05 watabou joined #git
14:06 acetakwas joined #git
14:06 JeroenT joined #git
14:07 d10n-work joined #git
14:07 WayToDoor joined #git
14:07 Juzzika joined #git
14:07 WayToDoor joined #git
14:07 sword` joined #git
14:08 settermjd joined #git
14:08 zivester joined #git
14:09 sword`` joined #git
14:10 toogley joined #git
14:12 axl_ joined #git
14:13 boudicca joined #git
14:14 ToBeCloud joined #git
14:15 MrMojito_Busy joined #git
14:17 dangerou_ joined #git
14:19 ResidentBiscuit joined #git
14:20 Wind0r joined #git
14:20 settermjd joined #git
14:23 jaguarmagenta joined #git
14:23 ash_workz hey, does anyone know of http://learngitbranching.js.org/ and if so, do you know how to get to just the sandbox where you can mess around?
14:23 ash_workz oh
14:23 ash_workz jeez
14:24 ash_workz read much? >ash_workz
14:24 ronskisika joined #git
14:24 ash_workz it's right on the homepage
14:24 diegoviola joined #git
14:25 todd_dsm_work joined #git
14:26 todd_dsm_wk joined #git
14:26 hlysig joined #git
14:26 todd_dsm_wk joined #git
14:26 Silenced joined #git
14:27 lb ;)
14:27 lb your private sandbox -> cd /tmp/ && git init sandbox && cd sandbox
14:30 netj joined #git
14:31 kaldoran joined #git
14:32 mdw joined #git
14:32 nnyk_ joined #git
14:33 mdw joined #git
14:35 eletuchy joined #git
14:36 marcogmonteiro joined #git
14:37 Wind0r joined #git
14:37 mikecmpb_ joined #git
14:41 eijk joined #git
14:45 dougquaid joined #git
14:45 fuchstronaut joined #git
14:46 dougquaid I'd like to do an automatic git pull every couple of days using cron, but the repo requires my username and password. I can manually type them in when I run "git pull" myself, but can I automate the input of the username and password?
14:46 semantipc joined #git
14:47 mooru joined #git
14:47 BlaXpirit dougquaid, can't you use a ssh key?
14:47 lindenle joined #git
14:48 dougquaid BlaXpirit: Well, it's github, so I assumed I couldn't
14:48 BlaXpirit dougquaid, you sure can
14:50 Firzen I've a baseless commit (remote branch I added and didn't rebase) that just gets merged, but I've already rewritten the history so that the merge commit is empty (the changes of said baseless commit aren't needed anymore). How would I get rid of the baseless commit? (afterwards I could remove the empty merge commit, which would have only one parent then, with --prune-empty)
14:53 watabou joined #git
14:53 jason237 joined #git
14:55 clemf joined #git
14:56 lindenle joined #git
14:57 LionsMane joined #git
14:59 vuoto joined #git
15:00 Ahmed90 joined #git
15:01 CJKinni` joined #git
15:02 semantipc joined #git
15:02 ochorocho__ joined #git
15:02 Raven667 joined #git
15:03 Firzen Then again, I have a commit with no parents and children. --prune-empty did not remove it, although it's empty. Any ideas how to clean that up?
15:03 shruggar joined #git
15:04 tesuji_ joined #git
15:05 shruggar what's the best way to tell git diff to ignore certain types of changes? Specifically, I want to ignore any changes that match a particular regex, then view the --name-status of what remains
15:05 Illusioneer joined #git
15:06 Illusioneer left #git
15:06 Raven667 joined #git
15:06 selckin impossible
15:07 kpease joined #git
15:09 grawity could be hacked together with a clean/smudge filter, maybe
15:10 Milkweed joined #git
15:10 grawity or a diff something-or-other
15:10 Topic for #git is now Welcome to #git, the place for git help and plumbing advice | Public logs at http://goo.gl/BuUi5o | Current stable version: 2.9.0 | First visit? Read: http://jk.gs/git | Getting "cannot send to channel"? /msg gitinfo .voice | Remember to keep your porcelain clean
15:11 eijk joined #git
15:12 mangolisk joined #git
15:12 madewokherd joined #git
15:12 bango2 joined #git
15:12 lucido-cwl joined #git
15:12 Wind0r86 joined #git
15:13 Raven667 joined #git
15:13 watabou joined #git
15:13 uhmgswa^2 joined #git
15:14 tyreld joined #git
15:14 bango2 joined #git
15:14 axiom_1 joined #git
15:14 kpease joined #git
15:15 nd joined #git
15:16 mikecmpbll joined #git
15:16 sol56 joined #git
15:17 skapunker joined #git
15:19 Firzen Just found out the commit is actually connected, but my terminal displayed the backslash of the --graph output in a color that completely vanished from my perspective...
15:19 grawity use tig
15:21 semantipc joined #git
15:22 dmto joined #git
15:22 subhojit777 joined #git
15:23 nitin joined #git
15:25 zumba_ad_ joined #git
15:25 encod3 joined #git
15:27 diogenese joined #git
15:27 rnsanchez joined #git
15:27 ki0_ joined #git
15:27 tesuji_ left #git
15:28 diegoviola how do I create a patch of the commit in HEAD
15:28 tesuji joined #git
15:28 diegoviola with git-format-patch
15:29 diegoviola I think it was git format-patch -1 or something... I forgot
15:29 Spacew00t joined #git
15:29 eijk_ joined #git
15:29 thiago joined #git
15:30 fuchstronaut joined #git
15:31 psprint From github, I can clone repositories regardless of letters case in user and repository name. Does this mean Git is case insensitive?
15:31 artemisart joined #git
15:32 ilmari psprint: no, it mean's github's git server is case insensitive
15:32 ilmari *means
15:32 k3rn31 joined #git
15:33 tabrez joined #git
15:33 eijk joined #git
15:33 Gitzilla joined #git
15:35 psprint from the alternatives to github, which ones are case sensitive, does someone know?
15:35 rom1504 try it
15:36 aarobc joined #git
15:36 psprint yeah just registered at bitbucket
15:37 Eugene psprint - git uses the case sensitive functions; various filesystems(or a wrapper layer, like github) can choose whether to do case-matching or not
15:37 zumba_ad_ joined #git
15:38 Firzen Say I have three commits A > B > C in a row. I change C's parent to A now. What could go wrong?
15:39 Eugene Firzen - That can count as a !rewrite, but that's about it.
15:39 gitinfo Firzen: Rewriting public history is not recommended. Everyone who has pulled the old history will have to do work (and you'll have to tell them to), so it's infinitely better to just move on. If you must, you can use `git push --force-with-lease <remote> <branch>` to force (and the remote may reject that, anyway). See http://goo.gl/waqum
15:39 garethdaine joined #git
15:39 uhmgawa^3 joined #git
15:39 AaronMT joined #git
15:40 AaronMT joined #git
15:41 semantipc joined #git
15:41 efdetonator joined #git
15:41 qsx i just lost the game
15:45 osse diegoviola: have faith in yourself. try your own guess.
15:45 hlysig joined #git
15:45 diegoviola -1 worked
15:45 diegoviola thanks
15:46 diegoviola I just wonder what -1 means in this context
15:46 synthroid joined #git
15:48 Firzen Eugene: Do you happen to know what happens to B in that case?
15:48 Firzen Will it be gc'd?
15:48 ItSANg___ joined #git
15:49 Eugene If unreferenced by any other branches, it will eventually be GC'ed into the ether, yes.
15:49 astrofog joined #git
15:50 Sylvain31 how can I import the diff of a branch I already partially merged in a branch? I've done a merge --no-commit, removed extra file, now, I'm creating an new branch with every feature merged.
15:50 joaumg What is the best way to cast a long double to an unsigned long long ?
15:51 Eugene joaumg - wrong channel / language ;-)
15:52 Sylvain31 can I make a path with git diff other_branch and reapply?
15:52 imjacobclark joined #git
15:52 joaumg Eugene: oopsy !
15:52 Eugene We don't mind much
15:55 Firzen So, in general, why wouldn't you just replace the parent of a commit with its grandparent to delete a commit? I found a page that tells me to use rebase for that.
15:55 rnsanchez joined #git
15:55 Firzen (http://sethrobertson.github.io/​GitFixUm/fixup.html#remove_deep)
15:55 lindenle joined #git
15:56 Firzen Only for local repositories, no remotes.
15:56 fscala joined #git
15:57 BombStrike joined #git
15:57 Sylvain31 Firzen: for who?
15:58 Firzen Sylvain31: I don't understand that question in this context.
15:58 tijko joined #git
15:59 Sylvain31 Firzen: about your answer, was it for my question?
16:00 Firzen Sylvain31: Yeah, a while ago I answered to a question from your. So your nick is there on purpose, not by accident. ;)
16:00 bhjcps joined #git
16:01 rominronin joined #git
16:01 rkazak joined #git
16:01 Firzen And that last thing from me was a question of my own.
16:01 mooru_ joined #git
16:01 Sylvain31 Firzen: I'm lost. Can I apply a selected patch with the link you posted?
16:03 ShapeShifter499 joined #git
16:03 watabou joined #git
16:04 maucar joined #git
16:04 lindenle joined #git
16:04 Firzen Sylvain31: No, my link is when you want to delete a commit.
16:05 Sylvain31 Firzen: OK. So it wasn't for me. So my question, "for who?" ;)
16:06 dhollinger joined #git
16:06 corebits joined #git
16:09 nabax joined #git
16:10 eletuchy joined #git
16:11 nabax hi! Novice question: I've just committed my changes, and before pushing to origin I've found out I made some mistakes. I'd like to cancel the commit without losing any changes in the working dir. Is git reset what I need or will it revert my changes?
16:12 SimonK joined #git
16:13 ki0 joined #git
16:13 audy left #git
16:14 imachuchu joined #git
16:14 ash_workz joined #git
16:15 Fingel joined #git
16:15 cdg joined #git
16:15 Firzen nabax: Just do NOT use the --hard flag and it shouldn't change your working dir.
16:15 Fingel hey all, I'm trying to pipe my git diffs through diff-highlight. Without piping my diffs have color, with diff-highlight it removes color. Anyone know why this might be?
16:16 ojacobson git-diff detects whether stdout is a tty or not
16:16 nabax Firzen, thanks! That's what I suspected! :)
16:16 ojacobson if it is, you get color (unless you configure git not to). if it's not, you don't (unless you force it).
16:17 Fingel so piping it through diff-highlight removes the tty? I wasn't aware thats how unix pipes worked
16:17 mwhooker joined #git
16:17 dougquaid left #git
16:17 xall_ joined #git
16:18 tomog999_ joined #git
16:18 Sylvain31 is there shortcut to name the current branch in git diff? git diff <current> other_branch
16:18 ojacobson a pipe is not a tty, yes
16:18 ojacobson diff-highlight's output may be going to a tty but git's isn't
16:18 Fingel is there a config option to force color all the time?
16:18 ojacobson you can force git-diff to colorize if you want, yes
16:19 ojacobson you don't want it always (diffs with terminal control sequences aren't patch(1) or git-apply(1) friendly)
16:19 ojacobson git diff --color will force it
16:19 holodoc joined #git
16:19 ojacobson You can set color.ui or color.diff to change the default
16:19 Fingel that's annoying to have to type all the time
16:20 Fingel oh I see, the [color] section
16:20 Fingel or I can just make an alias
16:20 Fingel thanks!
16:20 Sylvain31 To myself: "git diff HEAD other_branch" do the trick ;)
16:22 lohfu joined #git
16:22 netj joined #git
16:23 flavius joined #git
16:23 TomyLobo joined #git
16:23 jaguarmagenta joined #git
16:24 durham joined #git
16:24 LeBlaaanc joined #git
16:24 diegoaguilar joined #git
16:25 mingrammer_ joined #git
16:26 jstimm_ joined #git
16:27 samson joined #git
16:27 Fingel left #git
16:28 thebloggu joined #git
16:29 mariuscc joined #git
16:29 mwhooker joined #git
16:30 robotroll joined #git
16:31 algun joined #git
16:31 amsharma joined #git
16:31 amsharma are emails a necessary part of a commit's author?
16:32 amsharma left #git
16:32 amsharma joined #git
16:32 amsharma so, can't "justsomestring" be used as an author string?
16:32 BlaXpirit amsharma, probably not
16:32 ojacobson you can use any string there, but git's UI layer will fight you on it.
16:32 ojacobson Nearly any string, anyways
16:32 ki0 joined #git
16:33 amsharma i'm actually converting a repo from bzr to git
16:33 amsharma and bzr has weird committer strings
16:33 circ-user-eHrBN_ joined #git
16:33 BlaXpirit ojacobson, you can use  '<>'  as the email, that has worked out OK for me
16:33 nidr0x joined #git
16:33 amsharma BlaXpirit: oh, let me try that out
16:33 amsharma btw, what bzr to git scripts are available in the wild?
16:34 Gsham joined #git
16:37 fuchstronaut joined #git
16:38 eijk joined #git
16:39 d0val71 joined #git
16:40 amsharma left #git
16:42 BrianBlaze420 joined #git
16:43 LeBlaaanc joined #git
16:44 OliverJAsh joined #git
16:44 pgdagenais joined #git
16:44 imjacobclark joined #git
16:47 clemf_ joined #git
16:48 mooru joined #git
16:49 dougquaid joined #git
16:50 dougquaid I've cloned a repo and I do an automatic pull of that repo every few days. I'd like to run a script when my "git pull" command pulls down new commits. Can I accomplish this with a hook?
16:50 d0val71 joined #git
16:51 BlaXpirit dougquaid, i'd be interested in it.  otherwise  grep up to date  :|
16:52 texinwien joined #git
16:53 d0val71 joined #git
16:53 samson joined #git
16:56 samson joined #git
16:58 am55 joined #git
16:59 ojacobson scripting merges is hazardous at the best of times
17:02 pgdagenais left #git
17:02 BlaXpirit it's just fast-forward merges.  could even do reset --hard, no prob
17:02 rominronin joined #git
17:02 ojacobson break it up into steps
17:03 ochorocho__ joined #git
17:03 ojacobson fetch, then check if HEAD and @{upstream} are the same commit, then take action as appropriate
17:04 musl joined #git
17:04 mwhooker joined #git
17:04 joshc_ I've wanted the same functionality to implement "post-receive"-like behaviors for an active upstream project who's repo I don't own.  I ended up setting up '.' as a remote (with some special namespace mappings), and doing 'git remote update && git push .' in my cronjob
17:04 eletuchy joined #git
17:07 shinnya joined #git
17:13 Oebele joined #git
17:14 kadoban joined #git
17:14 johnnyfive joined #git
17:15 mikecmpbll joined #git
17:15 encod3 joined #git
17:16 zeroSteiner left #git
17:19 srcerer joined #git
17:20 watabou joined #git
17:20 Macaveli joined #git
17:20 EvilPenguin joined #git
17:20 semantip_ joined #git
17:20 Firzen http://firzen.neon-gaming.de/shar​e/mintty_2016-06-20_19-20-01.png :(
17:21 fracting joined #git
17:23 jrnieder joined #git
17:23 Gsham joined #git
17:23 sorenb_ joined #git
17:24 srcerer joined #git
17:25 elastix joined #git
17:25 rgrinberg joined #git
17:25 ochorocho__ joined #git
17:25 oblivioncr joined #git
17:26 uok joined #git
17:27 watabou joined #git
17:33 nitin_ joined #git
17:35 Gsham joined #git
17:35 thiago_ joined #git
17:36 Lynxium joined #git
17:36 aard joined #git
17:38 Impaloo64 joined #git
17:38 notebox joined #git
17:38 Impaloo joined #git
17:39 _28_ria joined #git
17:39 dreiss joined #git
17:41 Sasazuka joined #git
17:45 Impaloo joined #git
17:47 shaggycat joined #git
17:47 fuchstronaut joined #git
17:48 sanketdg joined #git
17:51 m0viefreak joined #git
17:52 xall joined #git
17:53 skylite joined #git
17:53 MutantMahesh joined #git
17:53 mischat_ joined #git
17:55 lcarlson joined #git
17:55 lcarlson any idea why I get a subcommand command missing on Fedora
17:55 lcarlson ?
17:56 lcarlson woops, submodule, not subcommand
17:57 rkj_ left #git
17:59 eletuchy joined #git
17:59 AaronMT_ joined #git
18:00 taylorm joined #git
18:01 rominronin joined #git
18:02 LeBlaaanc joined #git
18:03 setuid joined #git
18:03 setuid joined #git
18:04 jimi_ joined #git
18:05 ki0 joined #git
18:06 digidog joined #git
18:06 jbitdrop joined #git
18:07 boudicca joined #git
18:09 fracting joined #git
18:12 sbeller lcarlson: what did you do?
18:13 Balliad joined #git
18:15 cdown_ joined #git
18:15 ocbtec joined #git
18:16 akushner joined #git
18:17 zumba_addict joined #git
18:17 zumba_addict gheez, need help. I was told to undo the git push changes I made in package.json in master. How do I do that?
18:18 barteks2x joined #git
18:18 Seveas zumba_addict: revert the commit that made those changes and push the revert
18:19 zumba_addict !git revert
18:19 zumba_addict what is our bot?
18:19 Seveas don't be lazy, just read the manpage if you don't know how to use it.
18:19 jost_ joined #git
18:19 zumba_addict i'm checking it now
18:20 BlaXpirit !revert
18:20 gitinfo That's a rather ambiguous question... options: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard uncommitted changes in the working tree [git reset --hard]; c) undo committing [git reset --soft HEAD^]; d) restore staged versions of files [git checkout -p]; e) move the current branch to a different point(possibly losing commits)[git reset --hard $COMMIT]?
18:20 BlaXpirit ah perfect. zumba_addict^
18:20 zumba_addict am I right that i'm going to get the last commit sha which is mine? git revert shanumber?
18:20 zumba_addict cool
18:21 zumba_addict wow, that's complex
18:21 BlaXpirit zumba_addict, the popular suggestion is to not delete any commits and just add a commit that does the opposite
18:21 zumba_addict so i'll edit package.json again and repush?
18:21 BlaXpirit zumba_addict, the other option is  git reset --hard HEAD~4    -  that is to revert 4 last commits    && git push -f
18:21 zumba_addict got it
18:21 BlaXpirit reset --hard IS VERY DANGEROUS
18:21 BlaXpirit and force push, especially to master is very frowned upon
18:22 zumba_addict what about git revert to a specific commit?
18:23 BlaXpirit zumba_addict, git revert automatically makes a commit that has the opposite changes,  just read the manual
18:23 Ankhers joined #git
18:23 Robdgreat joined #git
18:23 BlaXpirit zumba_addict, man git-revert EXAMPLES 2nd one
18:23 gitinfo zumba_addict: the git-revert manpage is available at http://jk.gs/git-revert.html
18:24 zumba_addict ok
18:24 qswz joined #git
18:24 eight58 joined #git
18:24 strayArch joined #git
18:24 BlaXpirit probably drop the -n
18:24 jaguarmagenta joined #git
18:24 zumba_addict thanks
18:25 qswz I'd like to have a new branch (ok for that) and on them I'd like to squash the 89first commits
18:25 qswz with got rebase -i it would be long
18:25 qswz git*
18:25 BlaXpirit qswz, meh, find&replace in the editor :p
18:25 wviana joined #git
18:25 zumba_addict when I ran git revert 5df06d454e2, it opened vi editor and I saw my commit message
18:26 qswz how do you acess that rebase file?
18:26 zumba_addict looks like I referenced to the write file
18:26 qswz I'm always scared when rebasing
18:26 BlaXpirit and rightfully so
18:27 zumba_addict i ran git status and it's just showing i'm head by 1 commit
18:27 zumba_addict Your branch is ahead of 'origin/master' by 1 commit.
18:27 BlaXpirit zumba_addict, sounds about right. you made a commit that does the opposite of commit 5df06d454e2
18:27 zumba_addict got it
18:27 qswz git*
18:27 qswz :)
18:27 zumba_addict i'll git push -u origin master now
18:28 zumba_addict that's correct right?
18:28 zumba_addict do I need -f
18:28 BlaXpirit zumba_addict, no, you just added onto the history without editing existing commits, dont need f
18:28 zumba_addict cool
18:28 zumba_addict i'm verifying gitlab to make sure nobody made changes to it
18:28 BlaXpirit -u, --set-upstream   is also something you just use the very first time
18:29 zumba_addict verified, i'm the last to commit
18:29 BlaXpirit zumba_addict, if you weren't it would tell you
18:29 BlaXpirit (unless, of course, you're force pushing, in which case do the check like that)
18:29 zumba_addict it worked!
18:29 zumba_addict woohoo
18:30 watabou joined #git
18:31 zumba_addict sigh of relief
18:31 zumba_addict thanks a lot :)
18:32 netcarver joined #git
18:34 zumba_addict am I right that if we git revert backtoprevious5commits, the commits later than that commit will also be deleted?
18:34 BlaXpirit zumba_addict, no
18:34 zumba_addict oh
18:35 zumba_addict so only that commit we specified will be undo'ed?
18:35 zumba_addict undone
18:35 nidr0x joined #git
18:35 qswz What does squash do actually, concatenate commit messages?
18:35 BlaXpirit zumba_addict, yes, that's why i linked you to a specific example
18:35 qswz f, fixup = like "squash", but discard this commit's log message
18:35 qswz it's confusing to me
18:36 zumba_addict got it, thanks!
18:36 qswz what is I just remove the first 89 lines?
18:36 zeemz joined #git
18:36 qswz it will be broken I guess since git is difference-based
18:36 BlaXpirit qswz, squash makes a commit that does the changes of multiple commits.  in rebase -i  it lets you edit the commit message, but initially in the editor you get all the commit messages of the referenced commits
18:37 qswz ok
18:37 BlaXpirit the difference with fixup probably is that you don't get this in the editor, just the message of first commit
18:37 BlaXpirit qswz, don't remove lines in git rebase -i,  that's just removing commits
18:37 qswz ok
18:38 qswz but I'd like to merge of those 89 first commits to be with the date of the last one
18:38 BlaXpirit well thats not so easy - dont know about the date part
18:38 qswz hmm
18:39 qswz I think I could just git co at the right point
18:39 fuchstronaut joined #git
18:39 hahuang62 joined #git
18:39 qswz then git co -b --orphan
18:39 aob joined #git
18:40 qswz well.. then how could I merge only further commit from the other branch
18:44 qswz just to make sure, if I do a git rebase on a branch, it won't affect other branches history?
18:46 ynanm joined #git
18:49 EvilPenguin joined #git
18:52 elementalest joined #git
18:52 elementalest joined #git
18:52 fuchstronaut joined #git
18:52 feignix joined #git
18:53 ahmedelgabri joined #git
18:53 LeBlaaanc joined #git
18:55 xall_ joined #git
18:55 nwkj86 joined #git
18:56 miroshi joined #git
18:56 mooru_ joined #git
18:58 mikecmpbll joined #git
18:59 slayerjain joined #git
19:02 rominronin joined #git
19:02 vuoto joined #git
19:02 ochorocho__ joined #git
19:02 jstein joined #git
19:02 zeroed joined #git
19:04 tgunr joined #git
19:04 MeXTuX joined #git
19:05 editshare-ericf joined #git
19:06 jstein Hi, I want to undo the not yet comitted renames: https://bpaste.net/show/2650fbebd454 line 10..19. I tried
19:06 jstein git checkout -- blob/article/Baker1992a-Stacking_power_MOSFE​Ts_for_use_in_high_speed_instrumentation.pdf
19:06 jstein but git status did not change. What is the correct way?
19:06 zarel joined #git
19:06 qswz left #git
19:07 dsdeiz joined #git
19:07 dsdeiz joined #git
19:08 Seveas jstein: you'll need to git mv them back
19:12 lindenle joined #git
19:13 jstein is there really nothing more sophisticated? Can I commit all but the renames and undo all renames then?
19:13 Seveas sure, but that's not what you asked :)
19:13 jstein you are right ;-)
19:13 Gianormagantrous joined #git
19:13 Gianormagantrous joined #git
19:13 Seveas git add all the files you do want to commit, git commit, git reset --hard to wipe the other changes
19:16 Raging_Hog joined #git
19:18 jstein yes! that is what I was looking for. Thank you.
19:18 kappy joined #git
19:18 * jstein adds this line to his zim
19:19 BlessJah git reset -- <file> ; git checkout -- <file>, much better than git reset --hard
19:20 apotry joined #git
19:20 Impaloo joined #git
19:21 SwiftMatt joined #git
19:21 kfunk joined #git
19:23 SimonNa joined #git
19:26 robotroll joined #git
19:28 lcarlson joined #git
19:31 holodoc joined #git
19:32 Balliad joined #git
19:32 jstein BlessJah: so my command would have worked, if I had put git reset -- <file> && in front... I will try that next time. Thank you too.
19:36 adgtl_ joined #git
19:37 mooru joined #git
19:37 clemf joined #git
19:37 watabou joined #git
19:38 dmto joined #git
19:39 BlessJah may get confusing, that reset
19:40 BlessJah reset --soft/--hard/--mixed/-keep modifies HEAD pointer, reset -- <file> modifies staging area (removes files from staging, including renames/removals)
19:40 cdg joined #git
19:41 mikecmpbll joined #git
19:43 LeBlaaanc joined #git
19:44 Wulf joined #git
19:45 shaggycat joined #git
19:48 fuchstronaut joined #git
19:48 LeBlaaanc joined #git
19:48 Levitator joined #git
19:49 akiva joined #git
19:50 sapiens_vector joined #git
19:54 lordPoseidon joined #git
19:55 ResidentBiscuit joined #git
19:57 moei joined #git
19:58 hahuang62 joined #git
20:02 cdown_ joined #git
20:02 rkazak joined #git
20:03 mingrammer_ joined #git
20:03 rominronin joined #git
20:03 lasko joined #git
20:05 ResidentBiscuit joined #git
20:06 Rynomster joined #git
20:07 Repox joined #git
20:11 troulouliou_dev joined #git
20:11 clemf joined #git
20:11 sol56 joined #git
20:12 nairwolf joined #git
20:13 nairwolf hi, in the release note of git 2.9, it is said that we can use "--no-expand-tabs" to disable the indentation in git log, but the option doesn't seem to be present. Is it an option of git log ?
20:13 spudowiar joined #git
20:13 nairwolf It's in the third paragraph : https://raw.githubusercontent.com/git/git​/master/Documentation/RelNotes/2.9.0.txt
20:14 nnyk_ joined #git
20:14 johnmilton joined #git
20:15 Samy__ joined #git
20:15 Samy__ hi , Can I get previous commits after doing git -am ?
20:15 backnforth joined #git
20:16 BlaXpirit Samy__, no such thing as "git -am"
20:16 BlaXpirit use  git log,  whatever
20:16 Samy__ git -a
20:16 Samy__ git commit -a
20:16 BlaXpirit Samy__, so you made a commit, what does that change?
20:17 Samy__ it puts a single commit for all my commits
20:17 Samy__ so say I had 17 commits now I have just one and if I want to go to the 11th commit is there a way?
20:17 tobiasvl what do you mean you just have one Samy__
20:18 tobiasvl did you squash?
20:18 _ikke_ nairwolf: seems to work for me
20:19 Samy__ I did a git rebase --soft actually then I did git commit -am
20:19 cdg joined #git
20:19 tobiasvl rebase --soft ?
20:19 BlaXpirit Samy__, maybe  reset --soft
20:19 BlaXpirit sounds likely then
20:20 Samy__ so basically it just reverted all my commits :( yes sorry :D
20:20 BlaXpirit your commits are lost but they still may be in reflog
20:20 Samy__ So is there a way i can see a diff/
20:20 Samy__ ?
20:21 jonath-n joined #git
20:22 jonath-n joined #git
20:23 Macaveli joined #git
20:25 jaguarmagenta joined #git
20:25 nairwolf _ikke_: yes, it works indeed.
20:25 nairwolf in fact, it's the autocompletion which doesn't work
20:25 nairwolf is it something expected ?
20:25 BlaXpirit Samy__, what diff?  chances are, the commit you made incorporates all of the commits you had squashed into one
20:26 nabblet joined #git
20:27 nairwolf also, I'm reading the line about git-p4 and git-p4 doesn't seem to be available in my package. Is it because it's not available by default ? I'm using git from Ubuntu package (with the 2.9 version)
20:27 nabblet hi, what's the best way to deal with file paths that lead to files not in the repository? (E.g. location of binary data files)
20:27 nabblet I don't want to sync that data though. It's probably more a languaage specific yestion.
20:27 nabblet question
20:28 jyjon joined #git
20:30 aidalgol joined #git
20:30 Samy__ BlaXpirit: yes that is what happened now basically I just want to see my previous commits
20:32 BlaXpirit Samy__, well they are lost, unless you had them on another server or another branch
20:32 BlaXpirit !reflog
20:32 gitinfo The git reflog (`git log -g`) records the SHAs of your HEADs for 2+ weeks. `git checkout -b myrestore OLDSHA` and `git reset --hard OLDSHA` will relink to that state via a new and current branch respectively, see http://sethrobertson.github.com/GitFixUm/ for full details.  WARNING: reset --hard will trash any uncommitted changes!  Visualize with: gitk --all --date-order `git log -g --pretty=%H`
20:33 Samy__ BlaXpirit: I didnt have them anywhere else sadly so are they lost?
20:33 tobiasvl Samy__: they are probably in the reflog like BlaXpirit alluded to
20:33 artemisart joined #git
20:33 BlaXpirit Samy__, you can try recovering old commits, I have never done that so I can't help
20:34 tobiasvl just run "git reflog"
20:36 Goplat joined #git
20:36 Samy__ tobiasvl: It showed me the commits but how to access them?
20:37 tobiasvl like gitinfo said above, just checkout a branch at the commit
20:37 tobiasvl and then you can do whatever with it
20:37 tobiasvl http://sethrobertson.github.com/GitFixUm/
20:38 efdetonator joined #git
20:38 watabou joined #git
20:40 MeXTuX Is it possible to search in code      git log -S "window.alert"      but only for files under the specific directory "js"???
20:41 ojdo joined #git
20:41 BlaXpirit MeXTuX, tried appending    `-- directory`  to your command?
20:41 BlaXpirit just a guess, otherwise i don't know
20:42 tobiasvl yes that should work
20:42 MeXTuX mmm no. I didn't. I didn't know about "--directory" switch. I'm gonna try it Tnx :D
20:43 Phylock joined #git
20:43 diegoaguilar joined #git
20:43 tobiasvl not --directory, it's not a switch
20:44 tobiasvl at the end, put -- to separate switches/options/arguments from the path
20:44 tobiasvl so "git log -S "window.alert" -- js/
20:44 BlaXpirit tobiasvl,  git log --help  and see the very top
20:44 tobiasvl I assume you meant MeXTuX
20:44 BlaXpirit oops
20:45 zifnab question: how many lines does git look at to figure out a merge conflict?
20:45 zifnab is it just before/after 1, or something like that?
20:46 ojacobson the modified lines.
20:46 ojacobson If two chunks modify the same lines in different ways, then they conflict.
20:46 ojacobson The context may be used to deal with chunks that have moved within the file, but it's not used for deciding what conflicts. (A sufficiently modified context will stop one or the other chunk from applying at all, though.)
20:47 radsy joined #git
20:56 inflames joined #git
20:56 johnmilton joined #git
21:00 uhmgswa^2 joined #git
21:03 d0nkeyBOB joined #git
21:03 zarel joined #git
21:04 leeN joined #git
21:04 sonOfRa joined #git
21:04 jooni joined #git
21:04 Asiajey joined #git
21:04 Asiajey joined #git
21:05 ahmedelgabri joined #git
21:05 echapman_ joined #git
21:05 durham joined #git
21:05 rominronin joined #git
21:06 echapman joined #git
21:08 echapman left #git
21:10 foodSurprise joined #git
21:15 mwhooker joined #git
21:18 foodSurprise hey
21:18 foodSurprise let's say I do a pull and I'm like 3 commits behind. How can I find out exactly what commit ID I was on before the pull so I can do a diff?
21:18 foodSurprise So I can compare the HEAD to the place I was before I did the pull. (I didn't do the diff prior to pull)
21:19 BlaXpirit foodSurprise, do you still have console output from the pull?
21:19 foodSurprise yes
21:19 BlaXpirit it usually prints out things like  a161fae..b1df863   on the last line, doesn't it?
21:19 foodSurprise yes I see it
21:19 i7c If not, you could have a look at reflog.
21:19 Rodya_ joined #git
21:19 perlpilot_tpc foodSurprise: reflog
21:19 BlaXpirit foodSurprise, so copy and paste that.  like   git diff a161fae..b1df863
21:19 * i7c so fast. /me very satisfied.
21:20 foodSurprise Thanks. How do I use the diftool?>
21:21 chris2 joined #git
21:22 foodSurprise is there a way to view the diff of only files that end in .cs
21:24 mjal joined #git
21:24 nwkj86 joined #git
21:26 Tryop88 joined #git
21:27 semantipc joined #git
21:30 mellernoia joined #git
21:31 axiom_1 joined #git
21:31 garethdaine left #git
21:32 ramsub07 joined #git
21:32 ramsub07 hello, how do i remove a commit when i have the commit hash ?
21:34 Wulf ramsub07: git revert .......
21:35 ramsub07 then do i need to commit ?
21:35 Wulf ramsub07: this automatically creates a new commit
21:35 ramsub07 so if i push it'd be alright ?
21:35 mikecmpbll joined #git
21:36 Wulf ramsub07: I guess so.  Why do you think you need to do this anyway?
21:36 ramsub07 i don't want a commit that i have pushed anymore
21:36 ramsub07 i don't want the changes in the commit*
21:36 BlaXpirit ramsub07, good, git revert is better than rewriting history and literally removing the commit
21:36 Wulf ramsub07: because they're wrong or because you don't want others to ever know it happened?
21:37 kutsuya joined #git
21:39 clemf joined #git
21:39 durham joined #git
21:40 watabou joined #git
21:40 uhmgawa joined #git
21:42 d0nkeyBOB joined #git
21:44 d0nn1e joined #git
21:46 Rodya_ joined #git
21:48 Gamecubic joined #git
21:50 mda1 joined #git
21:52 joki joined #git
21:53 orbitphreak joined #git
21:53 jaime_ joined #git
21:53 mda1_ joined #git
21:54 jaime_ left #git
21:56 h4k1m joined #git
21:56 h4k1m Hi everyone
21:56 h4k1m How can I find which version is master on Github?
21:57 T4nk5bb0t joined #git
21:57 nidr0x joined #git
22:01 BlaXpirit please clarify
22:02 h4k1m BlaXpirit: ok let me take numpy package on github as an exmaple: https://github.com/numpy/numpy
22:03 h4k1m there's a branch master on github website
22:03 h4k1m when I download the zip file, this is the version downloaded
22:04 h4k1m is there a way to tell which version is the master branch?
22:04 BlaXpirit h4k1m, if you download master, it's not really any version, it is in between versions
22:04 tobiasvl well, it might be a version
22:04 tobiasvl if you mean tag
22:04 tobiasvl a tag can point to master
22:05 tobiasvl that's incidental though. you might be downloading the repo in the same state as a version.
22:05 kadoban h4k1m: You should usually use https://github.com/numpy/numpy/releases instead
22:05 kadoban And if you want something stable, avoid "rc" and beta and alpha, though I don't think numpy uses those designations except the first one.
22:05 h4k1m kadoban: yes but the rc version isn't stable, is it?
22:05 h4k1m kadoban: exactly
22:06 blackwind_123 joined #git
22:06 ajf- joined #git
22:06 h4k1m kadoban: thank you
22:06 kadoban Sure
22:06 rominronin joined #git
22:07 h4k1m BlaXpirit: tobiasvl: in this particular case, does master mean a branch?
22:08 BlaXpirit h4k1m, sure... master is a branch
22:08 kadoban master is a branch, yes. It's often the default branch. What it actually means in terms of development workflow or stability is up to the project though.
22:09 kadoban It'd be a bad idea to assume it's what you're "supposed" to use.
22:09 foodSurprise joined #git
22:09 h4k1m I see, thanks again guys
22:10 cdown joined #git
22:10 kadoban h4k1m: By the way, usually there's some other way to get this stuff, like 'pip' for python. I'm not sure it really makes sense to get numpy from github, unless you're going to be modifying numpy itself.
22:11 h4k1m kadoban: no actually I was struggling with a python package (this one https://github.com/perrygeo/simanneal)
22:12 h4k1m when I installed this package using `pip`, the package installed is not in its last version
22:12 h4k1m I mean its developer made few modifications after the last release
22:12 Codfection joined #git
22:12 madewokherd joined #git
22:13 fscala joined #git
22:17 CJKinni` joined #git
22:17 ollien joined #git
22:18 ollien I need to implement a git filter to my repo. Silly me has been indenting python with tabs and I want to fix that. I understand i can edit /etc/gitconfig to use a filter, but how can I make a filter repository wide?
22:18 ollien i.e. anyone who clones it will have the filter as well
22:19 ollien or should I just rewrite history once and call it a day
22:19 Eugene ollien - tl;dr you can't: that would be an arbitrary-code-execution vulnerability to anybody who does `git clone`
22:20 ollien Eugene So is rewriting history my best option?
22:20 Eugene I don't understand the problem. !xy
22:20 gitinfo Woah, slow down for a bit. Are you sure that you need to jump through that particular hoop to achieve your goal? We suspect you don't, so why don't you back up a bit and tell us about the overall objective...
22:20 Eugene "should I use a filter or rewrite history" doesn't make sense ;-)
22:21 ollien haha, let me explain myself. So I was indenting python with tabs when I should have been using spaces. Using some research, I found I could use a filter to do that, but that requires it be on each machine. Would the better option be to rewrite history using a filter and let that be the end of it.
22:21 Eugene You don't need to do either of those things
22:21 ollien Oh really?
22:21 Eugene Just make the spaces/tabs how you want them and make a commit of it
22:22 ollien How would you suggest doing it?
22:22 ahmedelgabri joined #git
22:22 ollien Right, but it's not just one file.
22:22 ollien it's a multitude of py files
22:22 Eugene So do it a bunch? Or use sed or whatever?
22:22 ollien Shit I forgot about sed...
22:22 Eugene I don't see how "I need to rewrite all of history" comes into play
22:23 ollien Eugene You can use a filter to rewrite tabs into spaces in history
22:23 Eugene To what end? That's history
22:23 ollien Also, the question then becomes, how do I merge this change into my branches?
22:23 ollien I don't wanna be going through all my branches to do this
22:23 Eugene How many branches do you have that it's a concern?
22:24 ollien 82
22:24 Eugene Again, xy ;-)
22:24 ollien aha
22:24 ollien Are you saying I should be committing these changers into each branch?
22:24 ollien or is there a better way to do this across branches
22:24 Eugene I don't understand what you would need 82 branches for
22:25 ollien Eugene They're just branches in the repo. Some are stale, but quite a few aren't
22:25 Eugene It's a very simple `git merge` to bring the changes across all the branches, but you'd end up needing to lint+fix each one to catch all the missed parts
22:25 Eugene Most branching models don't end in "and now you have 82 branches"
22:25 ollien which is why I was thinking to use a filter that swaps tabs/spaces
22:25 Eugene !gitflow
22:25 gitinfo The description of the gitflow branch workflow model is at http://nvie.com/posts/a-succ​essful-git-branching-model/ while a tool to help implement this workflow is at https://github.com/petervanderdoes/gitflow  See http://sethrobertson.github.c​om/GitBestPractices/#workflow for other workflow suggestions/references
22:25 jaguarmagenta joined #git
22:26 ollien Eugene
22:26 ollien oops
22:26 Eugene See how everything diverges and then re-converges eventually?
22:26 ollien It does converge eventually, but that's not the problem. If one branch has tabs, and that code is merged, the file will be mixed tabs/spaces
22:26 ollien which doesn't run in python
22:27 Eugene So you'll get a merge-conflict. Re-run your sed command and move on with life.
22:27 ollien It won't be merge conflicted if it's a new line
22:28 Eugene Rewriting all history everywhere would "fix" the tabs problem, but you'll get innumerable other problems with changing all history
22:28 Eugene git can't fix bad practices, only you can do that.
22:28 ollien Running sed until the problem is fixed doesn't seem like the greatest option, either
22:28 ollien I know that. I'm hoping that I'll be able to correct it across branches
22:29 Eugene If you really want, man git for-each-ref. Chain in sed
22:29 gitinfo the git-for-each-ref manpage is available at http://jk.gs/git-for-each-ref.html
22:29 osse git merge --no-commit; fix files; git add -u; git commit
22:30 ollien osse that doesn't fix across branches
22:31 osse it'll fix it when each branch is merged
22:31 foodSurprise joined #git
22:31 osse seems like the sanest thing to do
22:31 ollien osse ah
22:31 ollien that is, I suppose, a decent enough solution
22:32 ollien not one I would love, but it's better than using sed on each file
22:32 ollien Eugene When you say chain in sed, what do you mean?
22:32 TheVoid joined #git
22:32 osse if you're gonna do this, might as well use a proper python formatter
22:33 ollien very true
22:33 ollien aight
22:33 ollien thanks
22:33 osse some sort of sed s/// thing is destined to screw up some edge case :p
22:33 osse also sed -i breaks links
22:34 Eugene `git for-each-ref refs/heads/ | cut -d ' ' -f 3 | (xargs git checkout; sed -i s/\t/    /; git add .; git commit -m 'Tabfix')`
22:34 Eugene Or so
22:35 Eugene It's ugly and I doubt that'll work as it sits
22:35 Eugene But that's bashisms
22:35 osse ain't no bashisms there
22:36 Eugene It's generically a shellism, and bash is the One False Shell
22:36 foodSurprise joined #git
22:37 osse so say we all
22:40 elastix joined #git
22:41 watabou joined #git
22:45 mischat joined #git
22:49 McNozes joined #git
22:49 Tryop88 joined #git
22:51 SwiftMatt joined #git
22:51 ash_workz joined #git
22:53 foodSurprise joined #git
22:54 johnnyfive joined #git
23:00 cdown joined #git
23:00 zumba_ad_ joined #git
23:01 ahr3n joined #git
23:02 sangy joined #git
23:06 Goplat joined #git
23:07 ResidentBiscuit joined #git
23:07 a3Dman joined #git
23:07 mikecmpbll joined #git
23:07 rominronin joined #git
23:10 Rodya_ joined #git
23:14 jimi_ joined #git
23:18 harish joined #git
23:18 diegoaguilar joined #git
23:18 oblivioncr joined #git
23:19 WaReZ joined #git
23:20 larsss joined #git
23:22 cyphase joined #git
23:24 toblogan joined #git
23:27 BSaboia joined #git
23:27 toblogan joined #git
23:27 cyphase joined #git
23:28 toblogan joined #git
23:28 toblogan joined #git
23:29 toblogan joined #git
23:30 pipegeek joined #git
23:31 maestrojed joined #git
23:34 digidog joined #git
23:34 maestrojed I use my difftool all the time (BBedit). It just stopped working. "git difftool" just spills a diff in the CLI. It no longer launches BBedit and It doesn't give an errors either. I was using git difftool successfully for hours in this same session so I am confused. I have restarted iTerm and BBedit. Any advice on how to debug this? https://gist.github.com/maestrojed​/bc2602e15f076a287a0c28616abe8b46
23:34 maestrojed I confirmed the path defined there too.
23:36 toblogan joined #git
23:36 pipegeek does anyone have any familiarity with the tool Rultor?
23:40 peepsalot joined #git
23:41 watabou joined #git
23:45 ajf- joined #git
23:46 jstimm_ joined #git
23:46 nanga joined #git
23:46 nanga joined #git
23:46 fahadash joined #git
23:47 rkazak joined #git
23:49 naitsirhc joined #git
23:52 diogenese joined #git
23:56 eletuchy joined #git

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