Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-08-25

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

All times shown according to UTC.

Time Nick Message
00:00 justanotheruser joined #git
00:01 Spec-Chum joined #git
00:01 Rodya_ joined #git
00:02 red_ joined #git
00:06 fuchstronaut joined #git
00:06 sebboh joined #git
00:07 hwrdprkns joined #git
00:11 lvns joined #git
00:11 red_ joined #git
00:13 TyrellWellick000 joined #git
00:13 SpecChum joined #git
00:14 stamina joined #git
00:16 rubyonrailed joined #git
00:17 whomp joined #git
00:20 TyrellWellick000 left #git
00:20 duderonomy joined #git
00:27 MattMaker joined #git
00:28 astrozyk joined #git
00:29 whomp joined #git
00:29 jstein Hi, I want to use two different repositories for a project with the following structure: /prj/doc /prj/src  and more folders in repository A, which already exists and /prj/bigdata in repository B. This would be a good task for submodule, but after reading
00:30 Thorn__ joined #git
00:30 jstein https://chrisjean.com/git-submodules-adding-using-removing-and-updating/ I do not want to try this now, but later when I have some more time for testing.
00:30 rchavik joined #git
00:31 smithbone joined #git
00:31 jstein My plan was now to use a git in git and add some folders to the gitignore list and manually sync two git repositories. Is this possible, or will it fail at some point?
00:34 sbeller jstein: it sort of works, but you you cannot expect any syncing between the two
00:34 sbeller i.e. try bisecting a thing
00:34 sbeller in one and the other will stay static
00:35 vrand joined #git
00:36 jstein sbeller: if that is all, it is no problem. Thank you
00:40 LeBlaaanc joined #git
00:40 hwrdprkns joined #git
00:41 anonymuse joined #git
00:41 energizer left #git
00:42 MattMaker joined #git
00:44 vrand joined #git
00:44 tilerendering joined #git
00:44 tilerendering joined #git
00:47 Jon28 joined #git
00:48 anarchos2 joined #git
00:48 whomp joined #git
00:48 tcurdt joined #git
00:49 fakenerd joined #git
00:49 tesuji_ joined #git
00:56 vrand joined #git
00:57 MattMaker joined #git
00:59 arescorpio joined #git
00:59 tyoc213 joined #git
00:59 tyoc213 is git pull --rebase origin develop
00:59 tyoc213 correct?
01:00 milki thats a strange question. what part of the command are you checking?
01:02 dsantiago joined #git
01:02 Vampire0_ joined #git
01:02 tyoc213 all, Im making an "auto pull" .sh but I have always used front ends.... so :S
01:02 rubyonrailed joined #git
01:03 vrand joined #git
01:03 Mannequin joined #git
01:04 M-meznak joined #git
01:07 ojacobson "correct" is probably the wrong frame
01:07 vrand joined #git
01:07 ojacobson `git pull` is a frontend for either `git merge` or `git rebase` with `git fetch` run in front; which one is "correct" depends entirely on your intent and goals
01:07 sebboh joined #git
01:08 ojacobson I will say that "I know, I'll write a robot that does merges" is an idea that's fraught with peril
01:08 raijin joined #git
01:08 HoierM joined #git
01:08 ojacobson both merges and rebases can fail for non-automatable reasons, so you need to be prepared for that
01:08 atrigent joined #git
01:12 MattMaker joined #git
01:12 ochorocho__ joined #git
01:15 ytl joined #git
01:16 dsantiago joined #git
01:16 tyoc213 ojacobson, well... it should only have the develop branch up to date... no one will programm there
01:17 tyoc213 so it only need to update itself... maybe it doesnt need the rebase now that I think of it, but m so used to it on my worflow.... that I forgot I dont need it for a thing like this
01:17 ojacobson !deploy might be some interesting reading
01:17 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
01:18 ojacobson I also have some notes in http://grimoire.ca/git/stop-using-git-pull-to-deploy
01:19 daynaskully joined #git
01:21 tyoc213 ojacobson, thanks... now I see that I have missed a lot of points... xD wtf!
01:23 Torrone joined #git
01:23 zhiwliu joined #git
01:24 zhiwliu left #git
01:25 hahuang61 joined #git
01:27 Jon28 joined #git
01:28 evanwang joined #git
01:32 MattMaker joined #git
01:32 harish joined #git
01:34 mingrammer joined #git
01:34 _August_ joined #git
01:35 kurkale6ka joined #git
01:37 hwrdprkns joined #git
01:38 dreamyspell joined #git
01:41 hahuang61 joined #git
01:42 Gsham joined #git
01:45 xaviergmail joined #git
01:45 justanotheruser 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'?
01:49 ahr3n joined #git
01:49 rgrinberg joined #git
01:53 Gsham joined #git
01:54 thiago joined #git
01:54 UTAN_dev joined #git
01:55 pks joined #git
01:55 SwiftMatt joined #git
01:57 MattMaker joined #git
01:58 evanwang joined #git
01:58 nydel07 joined #git
02:00 Vampire0 joined #git
02:01 d^sh joined #git
02:01 thiago joined #git
02:01 JanC_ joined #git
02:05 mingrammer joined #git
02:05 cqi joined #git
02:06 smithbone joined #git
02:07 fuchstronaut joined #git
02:09 evanwang joined #git
02:12 MattMaker joined #git
02:13 anonymuse joined #git
02:13 motivic joined #git
02:15 atrigent_ joined #git
02:16 sebboh joined #git
02:17 Manuel15 joined #git
02:17 Manuel15 joined #git
02:20 hwrdprkns joined #git
02:20 evanwang joined #git
02:22 lightstalker joined #git
02:22 chid joined #git
02:22 Torrone joined #git
02:24 _August_ joined #git
02:25 dsantiago joined #git
02:25 moop joined #git
02:25 lagothri1 joined #git
02:25 thiago_ joined #git
02:25 happy-dude joined #git
02:26 anonymuse joined #git
02:27 huhlig joined #git
02:27 durham_ joined #git
02:27 NoOutlet joined #git
02:28 slayerjain joined #git
02:28 fcrick joined #git
02:31 hexagoxel joined #git
02:31 rkazak joined #git
02:32 MattMaker joined #git
02:33 atrigent_ joined #git
02:33 mrpackethead joined #git
02:34 chid joined #git
02:34 ggherdov joined #git
02:35 daishan joined #git
02:36 motivic joined #git
02:37 happy-dude joined #git
02:39 Goplat joined #git
02:40 thiago joined #git
02:45 zivester joined #git
02:45 moop joined #git
02:47 andschwa joined #git
02:48 kuldeep joined #git
02:48 fstd_ joined #git
02:49 sdothum joined #git
02:51 hahuang61 joined #git
02:52 MattMaker joined #git
02:52 atrigent_ joined #git
02:55 chachasmooth joined #git
02:58 mizu_no_oto joined #git
02:59 hwrdprkns joined #git
03:00 beyertech joined #git
03:02 MattMaker joined #git
03:02 dsantiago joined #git
03:03 aavrug joined #git
03:04 aavrug joined #git
03:09 lachm joined #git
03:09 KevinMGranger joined #git
03:11 hyperair joined #git
03:12 EduardoMartins joined #git
03:13 ochorocho__ joined #git
03:14 teadrop_ joined #git
03:15 a_thakur joined #git
03:19 mehola joined #git
03:19 zacts joined #git
03:20 sebboh joined #git
03:20 _August_ joined #git
03:21 a_thakur joined #git
03:22 MattMaker joined #git
03:24 theverbg_ joined #git
03:25 jumpman joined #git
03:29 Finnoloid joined #git
03:30 _August_ joined #git
03:31 jumpman joined #git
03:33 hobodave joined #git
03:34 ASOLAr joined #git
03:35 ericsupreme joined #git
03:35 rivarun joined #git
03:37 MattMaker joined #git
03:38 anarchos2 joined #git
03:40 arooni joined #git
03:43 ASOLAr joined #git
03:45 phanimahesh joined #git
03:46 happy-dude joined #git
03:52 MattMaker joined #git
03:55 tjone270 joined #git
03:55 Hink joined #git
03:56 new joined #git
03:57 new left #git
03:58 Rodya_ joined #git
03:59 hqkLChMUKc joined #git
04:00 thecomedian joined #git
04:02 Andrew_K joined #git
04:05 abyx joined #git
04:07 rgrinberg joined #git
04:08 rkazak joined #git
04:08 zacts joined #git
04:09 PioneerAxon joined #git
04:09 fuchstronaut joined #git
04:12 theverbg_ joined #git
04:12 MattMaker joined #git
04:12 Cabanoss- joined #git
04:16 ExoUNX_ joined #git
04:16 watabou joined #git
04:18 vassagus joined #git
04:19 sebboh joined #git
04:20 Hink_ joined #git
04:21 netj joined #git
04:23 govg joined #git
04:24 nights joined #git
04:27 MattMaker joined #git
04:29 mingrammer joined #git
04:30 govg joined #git
04:30 moop joined #git
04:30 jbu joined #git
04:31 hwrdprkns joined #git
04:31 vuoto joined #git
04:33 jbu all: I want to ignore a dir and its subdirs (recursively) so I created the .gitignore at the root dir of my project and entered "/bowser_components" but "git status" still shows the bowser components there.  Also tried "/bowser_components/**" as suggested online.  Also played with the leading slash. Using git 2.8.2.  Am I doing something wrong?
04:33 moneylotion joined #git
04:35 preaction jbu: are you sure it's not bower_components? also, if it's already in git, gitignore won't work. you have to remove it first
04:35 romerocesar joined #git
04:36 jbu preaction, typo into chat...Yes, I did use bower_components
04:36 jbu "/bower_components/**" and "/bower_components/" are what I used, and played with leading slash
04:37 MattMaker joined #git
04:37 preaction just /bower_components is what you should need. is it already in git?
04:38 jbu_ joined #git
04:38 l4v2 joined #git
04:38 romerocesar_ joined #git
04:38 theverbg_ joined #git
04:39 rubyonrailed joined #git
04:41 hahuang61 joined #git
04:49 navidr joined #git
04:53 moneylotion joined #git
04:54 phanimahesh joined #git
04:55 beyertech_ joined #git
04:55 jaafar joined #git
04:56 freimatz joined #git
04:57 d4rklit3 joined #git
04:59 subhojit777 joined #git
05:00 ASOLAr joined #git
05:00 renlo joined #git
05:01 rkazak joined #git
05:03 vassagus_ joined #git
05:04 syndikate_ joined #git
05:10 fuchstronaut joined #git
05:12 griffindy joined #git
05:14 ochorocho__ joined #git
05:14 leehambley joined #git
05:16 howdoi joined #git
05:16 cagmz joined #git
05:21 durham joined #git
05:23 mingrammer joined #git
05:24 braderhart joined #git
05:24 abyx joined #git
05:24 durham joined #git
05:27 johnmilton joined #git
05:27 abyx joined #git
05:29 zarel joined #git
05:30 alip_ left #git
05:31 bocaneri joined #git
05:32 MattMaker joined #git
05:32 sssilver joined #git
05:32 UserOO7 joined #git
05:33 sebboh joined #git
05:37 tvw joined #git
05:42 sssilver joined #git
05:44 zacts joined #git
05:46 qt-x joined #git
05:46 adac joined #git
05:47 MattMaker joined #git
05:47 ShekharReddy joined #git
05:50 phanimahesh joined #git
05:50 Muzer joined #git
05:52 gynter joined #git
05:52 Raging_Hog joined #git
05:52 apipkin__ joined #git
05:56 abyx joined #git
05:57 nickstenn joined #git
05:57 gynter Hello, when rebasing my-branch on top of master then master is ours/local and my-branch (working branch) is theirs/remote?
05:58 _ikke_ yes
05:58 gynter So when local has deleted file and remote has modified file then master has deleted file and working branch has modified file?
05:58 _ikke_ correct
05:58 phanimahesh joined #git
05:59 ome joined #git
05:59 gynter Ok, thanks!
06:03 dreiss joined #git
06:03 krs93_ joined #git
06:04 _ng joined #git
06:04 nilg joined #git
06:05 jim_wang joined #git
06:10 anttim joined #git
06:11 fuchstronaut joined #git
06:13 ams__ joined #git
06:15 ochorocho__ joined #git
06:15 romerocesar joined #git
06:16 Muzer joined #git
06:17 cqi joined #git
06:21 EisNerd joined #git
06:21 renlo joined #git
06:24 jknetl joined #git
06:24 roelmonnens joined #git
06:26 Aces_Charles joined #git
06:26 johnmilton joined #git
06:27 cdown joined #git
06:28 f-a joined #git
06:28 xaviergmail joined #git
06:28 d0nn1e joined #git
06:28 f-a hello: I forked a repo and made a merge request https://gitlab.com/fffaaa/mailman/commits/master
06:29 f-a after that, I fixed tests and documentation. Can I do a git rebase now or will it 'confuse' the developers on the main repo?
06:29 theskillwithin joined #git
06:30 Manuel15 joined #git
06:30 Manuel15 joined #git
06:30 f-a or alternatively: to 'collapse' all those commits, is it better to close the merge request, rebase, open a new one, or can I just do it without problems?
06:31 phanimahesh joined #git
06:32 Rish joined #git
06:32 cdown_ joined #git
06:32 govg joined #git
06:33 Electrom_ joined #git
06:33 ozzloy joined #git
06:33 ozzloy joined #git
06:35 VladGh joined #git
06:36 theverbg_ joined #git
06:37 paperziggurat joined #git
06:37 alcohol rebase, force push
06:37 satya4ever joined #git
06:37 alcohol since it is a fork, no issue
06:38 f-a thanks, alcohol
06:40 i7c Not many people can say that.
06:41 _ikke_ haha
06:41 t0by joined #git
06:41 konrados joined #git
06:42 konrados Morning :)
06:42 JeroenT joined #git
06:42 watersoul joined #git
06:42 i7c hey konrados
06:43 alcohol morning
06:43 hahuang61 joined #git
06:44 Pulp joined #git
06:44 sebboh joined #git
06:44 ASOLAr joined #git
06:45 konrados How about moving a directory, in a project founded on git? Can I simply move it by e.g. the "mv" command or should I do something git'ish? With svn I didn't care, now with git I'm not sure.
06:46 daishan joined #git
06:46 circ-user-cmOBz joined #git
06:47 mariuscc joined #git
06:47 watersoul joined #git
06:47 f-a left #git
06:48 i7c You can simply move it, which will lead to get telling you that the original directory got deleted and there’s a new untracked directory.
06:48 i7c Then you can stage these changes. Or you can use git mv to do it all at once.
06:49 i7c For the outcome it doesn’t really matter.
06:49 i7c A move is like a rename, git does not really care about those (does not track them specially).
06:50 konrados Thank you, i7c!
06:50 Sceorem joined #git
06:51 i7c :]
06:55 ArchNoob joined #git
06:55 jceb joined #git
06:57 fractalizator joined #git
06:58 blackwind_123 joined #git
06:59 rubyonrailed joined #git
06:59 diogenese joined #git
07:02 MattMaker joined #git
07:03 fees joined #git
07:03 mingrammer joined #git
07:04 JeroenT joined #git
07:05 Repox joined #git
07:06 GavinMagnus joined #git
07:07 xaa joined #git
07:08 ASOLAr joined #git
07:09 elect_ joined #git
07:09 GavinMagnus left #git
07:10 grift joined #git
07:12 fuchstronaut joined #git
07:14 stamina joined #git
07:16 ochorocho__ joined #git
07:17 Macaveli joined #git
07:18 ASOLAr joined #git
07:18 wxg joined #git
07:19 TomyWork joined #git
07:19 mangolisk joined #git
07:20 johnmilton joined #git
07:27 leeN joined #git
07:27 submitnine joined #git
07:27 MattMaker joined #git
07:29 watabou joined #git
07:30 davisonio joined #git
07:32 sagerdearia joined #git
07:35 jikz joined #git
07:37 ASOLAr joined #git
07:37 encod3 joined #git
07:42 ASOLAr joined #git
07:42 Dev0n joined #git
07:43 Torrone joined #git
07:44 adac joined #git
07:45 leehambley joined #git
07:48 danslo joined #git
07:48 danslo is there a "cleaner" way of doing something like: git diff 4c0860eab0341288d3fe9c3c1ed56fe4dbbd6c15^ 4c0860eab0341288d3fe9c3c1ed56fe4dbbd6c15
07:48 jceb joined #git
07:49 _ikke_ git show 4c0860eab0341288d3fe9c3c1ed56fe4dbbd6c15
07:49 danslo thank you
07:50 i7c _ikke_, same question for difftool?
07:50 GavinMagnus joined #git
07:51 i7c I usually do git difftool f0f0f0f0{~,}  for the lack of a better way, unless I’m mistaken and there is a better way.
07:51 i7c Then again, I don’t use difftool that often anymore.
07:52 Spec-Chum joined #git
07:53 ahmedelgabri joined #git
07:53 a_thakur joined #git
07:54 osse danslo, i7c:  4c0860e^!
07:57 sebboh joined #git
07:58 JyZyXEL joined #git
07:59 watabou joined #git
07:59 Balliad joined #git
08:02 i7c osse, ok true. Tho in zsh I have to type \^\! which is as horrible as {~,} :D
08:02 osse i7c: only the ^ :)
08:03 nights joined #git
08:03 i7c What.
08:03 bongjovi joined #git
08:03 alansaul joined #git
08:03 i7c Not sure what you’re saying.
08:04 ki0 joined #git
08:04 danslo i7c: I use (oh-my-)zsh and ^! seems to work fine
08:05 tilerendering collaboration workflow bug found. I was pulling from the repository into the repository of a colleague with my own user...then wondered why he couldnt pull
08:05 ki0 joined #git
08:05 i7c Mhm I just get "no matches found" without the escaping. ^^
08:05 dodobrain joined #git
08:05 roelmonnens joined #git
08:06 rominronin joined #git
08:08 MattMaker joined #git
08:08 p4trix joined #git
08:09 watabou joined #git
08:09 courrier joined #git
08:11 newcomer25 joined #git
08:11 newcomer25 The whole Law is fulfilled in one statement: ‘You’ll love your neighbour as much as yourself’ - Galatians 5:14
08:11 newcomer25 God bless you all and have fun chatting!
08:11 wizonesolutions joined #git
08:11 jaeckel joined #git
08:12 ASOLAr joined #git
08:12 i7c The only religion here is git.
08:12 fuchstronaut joined #git
08:13 aidalgol joined #git
08:15 osse i7c: I mean you can type  \^!  (no escaping of !)
08:15 i7c osse, you very wise person.
08:15 c0ded joined #git
08:16 JeroenT joined #git
08:16 marcogmonteiro joined #git
08:17 garethdaine joined #git
08:19 maxorator joined #git
08:20 tesuji joined #git
08:22 roelmonn_ joined #git
08:22 Torrone joined #git
08:23 zeroed joined #git
08:24 renlo joined #git
08:29 danslo joined #git
08:29 Darcidride joined #git
08:29 t0by joined #git
08:29 PioneerAxon joined #git
08:34 Ezzy joined #git
08:34 aca20031 joined #git
08:35 newcomer25 left #git
08:35 bremner joined #git
08:35 bremner joined #git
08:36 c0ded joined #git
08:38 DieguezZ joined #git
08:38 geko joined #git
08:39 robotroll joined #git
08:40 ertes-t6v joined #git
08:40 alansaul joined #git
08:43 MattMaker joined #git
08:44 hahuang61 joined #git
08:45 Torrone joined #git
08:50 jaeckel joined #git
08:52 jtprog_ joined #git
08:55 sebboh joined #git
08:57 axl_ joined #git
08:59 JeroenT joined #git
08:59 stamina joined #git
09:02 ahmed_elgabri joined #git
09:03 renlo joined #git
09:04 JeroenT joined #git
09:04 konrados Am I right that "checkout" is useful only if one uses branches? If we don't, then I can skip docs about it right now, and read it later, when I'll need it?
09:04 Dro joined #git
09:06 osse konrados: it is also useful for reverting changes to files
09:06 osse but git status tells you how to do that when you need it
09:07 axl_ joined #git
09:08 roelmonnens joined #git
09:09 davisonio joined #git
09:09 bernardio joined #git
09:11 konrados osse - thanks!
09:13 LucaTM joined #git
09:13 fuchstronaut joined #git
09:14 nikivi joined #git
09:15 chll_ joined #git
09:15 _nwkj86 joined #git
09:16 anarchos2 joined #git
09:16 JeroenT joined #git
09:17 ochorocho__ joined #git
09:21 cdown joined #git
09:23 lordjancso joined #git
09:23 fuchstronaut joined #git
09:24 jceb joined #git
09:26 abyx joined #git
09:26 mikecmpbll joined #git
09:27 aidalgol joined #git
09:27 MattMaker joined #git
09:28 davisonio joined #git
09:28 cdown_ joined #git
09:30 rubyonrailed joined #git
09:38 iskander joined #git
09:40 jeremM34 joined #git
09:40 ljc joined #git
09:41 cyan__ joined #git
09:42 MattMaker joined #git
09:43 JeroenT joined #git
09:46 nwkj86 joined #git
09:46 ljc joined #git
09:49 JeroenT joined #git
09:49 Rish_ joined #git
09:51 shymega joined #git
09:56 sebboh joined #git
09:58 romerocesar joined #git
10:02 JeroenT joined #git
10:02 MattMaker joined #git
10:03 beyertech joined #git
10:03 zeroed joined #git
10:04 Torrone joined #git
10:06 ljc joined #git
10:08 konrados I found an interesting SO answer, which opened my eyes a bit, although I don't understand everything. Here it is: http://stackoverflow.com/a/15733096/5733557 - for example: "In order to support this model git maintains a local repository with your code and also an additional local repository that mirrors the state of the remote repository. " does it do it constantly, without me doing anything?
10:09 grawity the "mirror" thing? it's updated whenever `git fetch` (and by extension `git pull`) is done.
10:10 grawity (and when the fetch says "master -> origin/master", it means "downloaded the 'master' branch to the local mirror 'origin/master'"
10:10 konrados grawity, but then how is this possible (the very next sentence): "y keeping a copy of the remote repository locally, git can figure out the changes needed even when the remote repository is not reachable." ?
10:10 konrados *by keeping...
10:10 a_thakur joined #git
10:11 osse it is possbile because git downloads everything
10:11 grawity when you ask it to show changes, it can compare against the local copy, without having to contact the server
10:13 xaviergmail joined #git
10:14 konrados OK, so... in a "normal" workflow, without any branches etc, I just "fetch" and "commit" and "push", right?
10:14 tobiasvl no branches? that doesn't sound like a normal workflow :)
10:14 _joes___ joined #git
10:14 tobiasvl konrados: you need "merge" as well (or "pull"), or perhaps just "rebase" if you have a very basic workflow
10:14 tobiasvl to integrate changes from the upstream
10:15 tobiasvl but well, depends what you do I guess
10:15 tobiasvl but you have branches even though you don't branch. master is a branch, so is origin/master
10:16 Torrone joined #git
10:16 ArchNoob joined #git
10:17 grawity konrados: yeah, more commonly it's fetch + merge (aka git pull)
10:17 MattMaker joined #git
10:18 konrados hmm... we implement things, test them on dev website, everyone reviews their code.... I meant we don't create "manually" those branches like "beta", "alpha" etc, tobiasvl. Thanks everyone, tobiasvl osse grawity, I'm trying to understand git's philosophy :)
10:18 grawity "beta" and "alpha" are kinda silly branch names anyway
10:18 grawity feature branches are a bit more useful
10:20 ShapeShifter499 joined #git
10:24 pampeho joined #git
10:26 _joes_ joined #git
10:27 beyertech joined #git
10:29 msm joined #git
10:29 ToxicFrog joined #git
10:31 konrados grawity - true, I just don't like branching... we once did it, the concept was not mine - we make branches based on developers, like "Konrad", "Someoneelse" and at the end of the day we merge it. That seemed crazy for me and ridiculous, becase I expected a lot of conflicts and I said I won't be mering anything, the guy how "invented" this agreed that he'll be doing that. Anyway, 10 minutes before the end of work, he took care about the conflicts and when
10:31 konrados he saw how many of them there is said f... He was sitting almost entire next day on this. We never came back to this idea :)
10:31 tobiasvl konrados: maybe read up on !flow
10:31 gitinfo konrados: [!gitflow] The description of the gitflow branch workflow model is at http://nvie.com/posts/a-successful-git-branching-model/ while a tool to help implement this workflow is at https://github.com/petervanderdoes/gitflow  See http://sethrobertson.github.com/GitBestPractices/#workflow for other workflow suggestions/references
10:31 tobiasvl but there are many workflows
10:32 tobiasvl making branches named after developers sounds like a bad workflow
10:32 tobiasvl merging them in at the end of the day regardless of their contents also sounds like a bad idea
10:32 beyertech_ joined #git
10:32 Torrone joined #git
10:33 konrados tobiasvl - thanks!
10:39 zeroed joined #git
10:40 jstein_ joined #git
10:41 ferr joined #git
10:44 ToxicFrog joined #git
10:45 hahuang61 joined #git
10:46 PioneerAxon_ joined #git
10:50 KnightsOfNi joined #git
10:52 cjbrambo joined #git
10:53 forrestv joined #git
10:53 rafalcpp before merging in someone/branch, how to see the diff between current state of code, and the code if we would do the merge?
10:54 tobiasvl rafalcpp: man git diff
10:54 gitinfo rafalcpp: the git-diff manpage is available at http://jk.gs/git-diff.html
10:54 jast that's actually tricky because merges can require conflict resolution which needs manual input
10:54 grawity HEAD..someone/branch – use that with git diff, log, log -p
10:54 jast you *can* easily compare the two versions, though
10:54 rafalcpp diff HEAD..someone/branch seems to be not showing that (it shows in diff also - that seem to delete some newest changes,  I guess things done after someone/branch)
10:54 ToxicFrog joined #git
10:54 grawity or with tig
10:54 grawity ah, hmm
10:54 grawity was it with three dots
10:54 rafalcpp grawity: tobiasvl yeah of course, but it doesn't seem to show that
10:54 grawity the merge-base thing
10:54 osse rafalcpp: unless it's a fast forward the thing you ask for isn't possible. do the merge, then diff, then undo the merge if you wish
10:54 jast that's three dots, yeah
10:55 osse ... or maybe I misunderstand
10:55 rafalcpp osse: well that would work. No other way really? for one I must merge first, then review, then merge again but -S this time
10:56 grawity merge --no-commit, review, commit -S
10:56 rafalcpp indeed
10:56 grawity though I usually start with reviewing individual commits in `tig ..foo/branch`
10:57 rafalcpp grawity: merge more often :P
10:57 grawity why
10:58 davisonio joined #git
10:58 ToxicFrog` joined #git
10:59 jeremyabel joined #git
10:59 sdothum joined #git
10:59 Endarked joined #git
10:59 rafalcpp grawity: oh just saying, some people prefer to review summary of PR changes anyway
11:01 osse rafalcpp: maybe the triple dot diff is all you need
11:01 osse it's what git will attempt to merge
11:02 sebboh joined #git
11:04 mingrammer joined #git
11:04 mizu_no_oto joined #git
11:04 HoierM joined #git
11:04 Torrone joined #git
11:05 rubyonrailed joined #git
11:05 danslo joined #git
11:05 jeffreylevesque joined #git
11:06 watabou joined #git
11:07 JeroenT joined #git
11:07 Sapphomo joined #git
11:13 encod3 joined #git
11:18 ochorocho__ joined #git
11:18 ojdo joined #git
11:20 beyertech joined #git
11:21 vuoto joined #git
11:22 beyertech joined #git
11:27 Torrone joined #git
11:27 ljc joined #git
11:28 bakabakabaka left #git
11:35 antranigv joined #git
11:35 iron_houzi joined #git
11:36 Akimb joined #git
11:36 iron_houzi If I rewrite history to change my email, will my peers have to wipe their repos and reclone, or can they do something like a `git pull -f` ?
11:37 osse neither
11:37 encod3 joined #git
11:37 osse they will have to be careful and rebase their local branches, but there's no need to wipe.
11:37 osse pull -f doesn't exist
11:37 tobiasvl well yeah it does but it just runs fetch -f when fetching
11:37 osse or at least doesn't do what the question implies
11:37 iron_houzi it does in my help file ..
11:38 tobiasvl there's no merge -f
11:38 iron_houzi I see
11:38 osse if they have no local work worth keeping then wiping is maybe easier
11:38 osse write a .mailmap file and forget about it
11:38 iron_houzi Why will they need to rebase their local branches? Is it because the parent commits will have changed?
11:38 osse yes
11:39 deadnull joined #git
11:39 watabou joined #git
11:39 iron_houzi I see .. Never heard of `.mailmap`, I'll see if that can solve my problem.
11:40 osse man git-shortlog
11:40 gitinfo the git-shortlog manpage is available at http://jk.gs/git-shortlog.html
11:40 osse it will solve the problem of git log etc. showing the wrong email for past commits
11:41 osse it will NOT solve the problem of hiding your old email as if it ever existed
11:42 Sceorem joined #git
11:42 iron_houzi I see .. we're only two people using the repo and we're pretty early at this point, so perhaps it makes sense to just rewrite history..
11:42 str joined #git
11:42 jast probably... get it over with
11:44 rnsanchez joined #git
11:46 iron_houzi Thanks for the great help!
11:55 ArchNoob joined #git
11:57 tomog999_ joined #git
11:57 harish joined #git
11:58 learath2 joined #git
11:59 learath2 Is there a command I can use to get which commit im detached from in a clean way
11:59 learath2 ?
11:59 romerocesar joined #git
12:00 bremner I don't think so. You have to pay attention when it happens and a message is printed on the console.
12:01 bremner I might be missing some clever trick with e.g. the reflog
12:01 learath2 i kinda need it in a script currently im parsing the output of git branch but thats just meh
12:01 _ikke_ git status in newer versions of git should show it
12:01 bremner save before you switch sounds like the most robust
12:02 _ikke_ it uses git reflog to find out
12:03 roelmonn_ joined #git
12:04 p4trix joined #git
12:04 ljc joined #git
12:04 learath2 nothing seems to expose wt_status_get_detached_from to the outside except git status or git branch so i guess ill keep paring eithers output
12:05 anonymuse joined #git
12:05 ome joined #git
12:07 fuchstronaut joined #git
12:07 sebboh joined #git
12:07 bremner or run "git symbolic-ref HEAD" before you detach.
12:09 raijin joined #git
12:09 learath2 its a script that gets the state of a repository thus its not an option
12:11 sanketdg joined #git
12:11 bremner I don't see why not, but OK.
12:12 choki joined #git
12:12 rahtgaz joined #git
12:12 ArchNoob joined #git
12:12 bremner you mean the script doesn't do the detaching, but tries to pick up the pieces?
12:13 stamina joined #git
12:13 alansaul joined #git
12:13 drbean joined #git
12:14 learath2 the script just observes it doesnt do anything
12:14 jeffreylevesque joined #git
12:14 bremner ok
12:15 mohabaks joined #git
12:16 wrouesnel1 joined #git
12:16 Gamecubic joined #git
12:16 anonymuse joined #git
12:16 phanimahesh joined #git
12:16 Torrone joined #git
12:17 fuchstronaut joined #git
12:18 iamdevnul joined #git
12:19 badloop_ joined #git
12:19 musl_ joined #git
12:19 grawity heh, sha-1 list discussion: "We are practically limited to 256-bit hashes because anything longer will wrap on an 80-column terminal when in hex form."
12:19 boltR_ joined #git
12:20 rahtgaz and we can't have that
12:20 mrkurtz joined #git
12:20 daniel_rose joined #git
12:20 Lewix joined #git
12:21 daniel_rose joined #git
12:21 _ikke_ learath2: Was thinking about @{-1}, but no way apparently to get rev-parse to resolve that
12:21 unreal joined #git
12:21 thomas1 joined #git
12:21 pikajude_ joined #git
12:22 zero7 joined #git
12:22 joseph_arnstein joined #git
12:22 learath2 i do git status 2> /dev/null | head -n 1 | awk 'END {print $NF}' for now
12:22 hahuang61 joined #git
12:22 grawity well, the output of @{-1} is a branch name and the output of rev-parse is an object ID
12:22 ealvarez__ joined #git
12:22 worstadmin joined #git
12:22 sarbs joined #git
12:22 discopatrick joined #git
12:22 Smirnov joined #git
12:22 Smirnov joined #git
12:22 rgrinberg joined #git
12:23 koro_ joined #git
12:23 bedouin joined #git
12:23 jgornick joined #git
12:23 sfan786_ joined #git
12:23 _ikke_ grawity: things like --symbolic can make it output refs
12:24 Torrone joined #git
12:24 _ikke_ and abbrev-ref
12:24 twisted` joined #git
12:24 Tee_Pee joined #git
12:24 slax0r_ joined #git
12:24 Zimmedon joined #git
12:24 flaguy48 joined #git
12:24 Hello71 joined #git
12:25 cyphase joined #git
12:25 PalTale joined #git
12:25 dpower joined #git
12:25 iota_pi joined #git
12:25 GodEater joined #git
12:25 GodEater joined #git
12:25 sweatsuit joined #git
12:25 bsanford joined #git
12:25 slax0r_ joined #git
12:26 Awesomecase joined #git
12:26 NeverDie_ joined #git
12:26 Anupkumar joined #git
12:26 ploop joined #git
12:26 SporkWitch joined #git
12:26 troyt joined #git
12:26 slax0r_ left #git
12:27 faheem joined #git
12:27 oxsyn joined #git
12:27 cory_ joined #git
12:27 rgrinberg joined #git
12:27 VampiricPadraig joined #git
12:27 VampiricPadraig joined #git
12:27 dsdeiz joined #git
12:28 jhass joined #git
12:28 ghostlight joined #git
12:28 Akimb joined #git
12:29 ghostlight joined #git
12:32 pro_metedor joined #git
12:32 mizu_no_oto joined #git
12:33 fuchstronaut joined #git
12:36 pro_metedor can I squash commits when branch is already sent to remote ?
12:36 osse yes, but you have to push --force
12:36 _ikke_ You can, but with the necessary caveats
12:36 osse depending on circumstances it might be no problem or it might be a shitshow
12:38 MattMaker joined #git
12:38 tobiasvl pro_metedor: !rewrite
12:38 gitinfo pro_metedor: 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
12:39 andlabs joined #git
12:40 watabou joined #git
12:41 pro_metedor osse: I know, it is my ‚personal’ branch I can only commit to, so I want to clean up commit history before pull request
12:42 osse pro_metedor: in that case it's fine.
12:43 fuchstronaut joined #git
12:43 pro_metedor I use remote only to share code between my desktop and laptop. No one else is allowed to contribute there.
12:44 pro_metedor and after Im done - pull request goes on
12:47 MattMaker joined #git
12:48 mjuszczak joined #git
12:49 mjuszczak joined #git
12:50 synthroid joined #git
12:50 koro_ left #git
12:53 rubyonrailed joined #git
12:53 zivester joined #git
12:54 iTeck joined #git
12:55 ams__ joined #git
12:56 fmcgeough joined #git
12:57 sebboh joined #git
13:02 synthroid joined #git
13:03 specing joined #git
13:03 altendky joined #git
13:03 Curly060 joined #git
13:04 specing Hi -- Is there a way to make "git add -e" make diff windows of max 15 lines or so?
13:04 Torrone joined #git
13:04 specing "diff window" the thing between two @@
13:04 badloop_ left #git
13:04 badloop joined #git
13:04 badloop test
13:04 cdg joined #git
13:04 badloop hmm... weird
13:05 _ikke_ specing: what if the diff exceeds 15 lines?
13:05 nettoweb joined #git
13:06 TooLmaN joined #git
13:06 xaviergmail joined #git
13:07 raijin joined #git
13:08 rominronin joined #git
13:08 ExoUNX joined #git
13:08 ExoUNX joined #git
13:09 AaronMT joined #git
13:10 Curly060 joined #git
13:11 ash_workz joined #git
13:11 bernardio joined #git
13:11 mda1_ joined #git
13:13 kpease joined #git
13:13 mingrammer joined #git
13:13 andlabs joined #git
13:14 ArchNoob joined #git
13:15 johnmilton joined #git
13:17 ebarry hey, I'm not entirely sure how to tell git to ignore newlines for certain type of files. I added .gitattributes and all, but I'm not sure if I did it right. I have some files which have the wrong line endings locally, and I can't push .gitattributes to check that I did it right. how should I proceed?
13:17 MattMaker joined #git
13:18 UrsoBranco joined #git
13:19 pro_metedor joined #git
13:19 ochorocho__ joined #git
13:19 anarchos2 joined #git
13:19 [cust] joined #git
13:21 _ikke_ gitattributes and endline normalizations works when adding things from the worktree to the index and visa-versa
13:21 _ikke_ so you can check even before comitting what the result is
13:21 ebarry well, the issue is that I didn't add the files, they're already present upstream
13:22 GodGinrai joined #git
13:22 MattMaker joined #git
13:22 ebarry they have the wrong line ending on my machine but are fine upstream
13:23 ebarry so basically I want git to go "oh I have new attributes what about upstream", but I can't make it do that...
13:23 ebarry is that any way to alter files already in the tree?
13:24 ebarry rather, the local ones
13:26 specing _ikke_: split it up like git stash --patch offers
13:26 AaronMT joined #git
13:26 sangy joined #git
13:27 l4v2 joined #git
13:27 umbSublime joined #git
13:32 ams__ Does git clone --reference save disk space? The docs suggest it does, but I'm confused - I cloned my repo to /eg/a and then cloned it again to /eg/b referencing /eg/a and the disk usage is the same
13:32 osse ams__: how do you measure it?
13:33 _ikke_ specing: I don't thing it wil split up larger continuous hunks
13:34 fuchstronaut joined #git
13:34 [cust] joined #git
13:35 jaafar joined #git
13:36 JeroenT joined #git
13:36 ams__ osse: du
13:37 osse ams__: try du -hsc /eg/a /eg/b
13:37 osse ie. give both at the same time
13:37 Dro joined #git
13:37 ams__ osse: Same results
13:38 Akimb joined #git
13:38 osse ams__: meaning that it shows double the size of each one?
13:38 rnsanchez joined #git
13:39 ams__ osse: It shows they are both the same size
13:39 osse what about the total
13:39 ams__ It's the sum of the 2 sizes
13:40 osse hmm
13:40 osse maybe you need --reference /eg/a/.git
13:41 JeroenT joined #git
13:43 ams__ So, you're sure that --reference should do what I think it does?
13:43 osse no
13:43 osse but I think it would do what you think it does
13:43 zeroed joined #git
13:43 ams__ ok fair enough
13:44 Eryn_1983_FL joined #git
13:44 osse the reason for the du thing was to see if it it was fooled by hardlinks etc.
13:45 ebarry anyone knows how to tell git to apply the newline preferences I gave it in my yet-to-be-pushed .gitattributes to files already in upstream?
13:46 zincrokx joined #git
13:46 canton7 ebarry, so you'll have to fix your local files, commit, then push: that's how you get stuff in upstream
13:46 canton7 read https://help.github.com/articles/dealing-with-line-endings/, in particular "Refreshing a repository after changing line endings"
13:47 ebarry thanks for the link!
13:47 ebarry canton7: oh, I'm afraid this is backwards from what I want though :(
13:47 specing _ikke_: yeah, it will not
13:48 specing _ikke_: that is why I'm asking how to make it so
13:48 canton7 "backwards"? please explain
13:48 MattMaker joined #git
13:48 ebarry the files upstream are fine, it's my local ones which are not. I created .gitattributes locally, and want the files upstream to be re-cloned locally so that the line endings are fine. I can't push anything upstream
13:49 kadoban joined #git
13:49 canton7 ebarry, wait, so you fetched commits from upstream, something happened to change the endings, you changed .gitattributes in an attempt to fix it, but your local files are still out of whack?
13:50 shatnerz joined #git
13:51 foodtooth joined #git
13:51 ebarry more or less; basically the files upstream are and always have been fine, but git converts newlines from lf to crlf locally, and in the case of some (binary) files this screws stuff up. I'm adding .gitattributes (it wasn't present before) so that my local newlines no longer get screwed up, but I want to make sure the newlines get properly converted before I submit the diff
13:51 shatnerz Not entirely sure if this is relevant here, but the company I work at is using SCM manager. Is it possible to give another user read or write permission through the git cli?
13:51 canton7 ebarry, follow steps 1-3 of my link then
13:52 canton7 after those, there shouldn't be anything to commit
13:52 ebarry thanks, will give it a shot!
13:53 sebboh joined #git
13:54 foodtooth left #git
13:54 tabrez joined #git
13:55 alexsystemf__ joined #git
13:55 jh3 joined #git
13:55 ams__ osse: It seems to me that it uses the --reference instead of asking the remote git server.
13:55 ams__ oh wait
13:55 * osse waits
13:56 EisNerd_ joined #git
13:56 neurodrone_ joined #git
13:56 exarch joined #git
13:56 ams__ aha! it does work! (with and without .git fyi)
13:56 ams__ It's not working when I try and use submodules
13:57 Torrone joined #git
13:58 canton7 submodules have an additional level of complexity, where the contents of their .git folder are actually stored in the .git folder of their parent repository
13:58 canton7 so I don't know whether that's interfering
13:58 _ikke_ specing: There is no way afaik, except for a custom merge driver
13:58 a3Dman joined #git
13:58 ams__ canton7: Yeah, but I'm running "git submodule update --reference <path to super project>
13:58 ams__ Which I thought would do the trick
14:01 ebarry canton7: thanks! that seems to have worked :)
14:01 canton7 cool :)
14:02 ebarry now I have a file while got all its line endings changed for seemingly no reason, but I'll deal with that :P
14:02 MattMaker joined #git
14:03 phaleth joined #git
14:04 vuoto joined #git
14:06 Torrone_1 joined #git
14:06 a_thakur joined #git
14:07 krillr joined #git
14:07 MattMaker joined #git
14:07 Lynxium joined #git
14:10 daniel_rose joined #git
14:10 alansaul joined #git
14:11 anarchos2 joined #git
14:12 mjuszczak joined #git
14:12 sssilver joined #git
14:13 MattMaker joined #git
14:13 vuoto joined #git
14:13 holdsworth joined #git
14:14 andrewSC joined #git
14:14 andrewSC hi all
14:14 Torrone_1 joined #git
14:14 jceb joined #git
14:14 andrewSC I've setup signing my commits with my GPG key and I was reading briefly online that if my teammembers don't also adopt gpg commit signing then their workflows become much more difficult?
14:15 andrewSC is there any truth to that?
14:15 moneylotion joined #git
14:15 _ikke_ Do you knw where you've read that/
14:15 ResidentBiscuit joined #git
14:15 andrewSC I don't see why if I sign my commits to prove they're coming from me, why anyone else would have trouble adding additional commits to the branch?
14:16 andrewSC _ikke_: https://git-scm.com/book/en/v2/Git-Tools-Signing-Your-Work#Everyone-Must-Sign
14:16 grawity they wouldn't
14:16 andrewSC grawity: hi
14:16 grawity unless the signatures rule becomes enforced
14:17 andrewSC so basically that paragraph is just poorly worded
14:17 andrewSC gotcha
14:17 grawity e.g. via pre-merge hooks or server-side
14:17 _ikke_ It's mainly about rewriting history
14:17 andrewSC gotcha gotcha
14:17 grawity _ikke_: as in?
14:17 _ikke_ "If you don’t, you’ll end up spending a lot of time helping people figure out how to rewrite their commits with signed versions. "
14:17 andrewSC yeah I could see people getting hassled if the rules became enforced but if I'm just signing my commits, I would think I'll be fine
14:18 _ikke_ andrewSC: they only get trouble when they would rewrite your commits
14:18 grawity ah
14:18 grawity well, rewriting somebody else's commits already sounds like trouble
14:18 andrewSC like if they tried to squash and they included my commit in the squash?
14:18 andrewSC grawity: right
14:19 Torrone joined #git
14:19 grawity hmm not sure how signatures would make that a problem
14:19 andrewSC or if they tried to do a reset soft to my commit or something?
14:19 Mo0O hi there
14:19 grawity that one is not even history-rewriting
14:20 grawity well, not to that extent
14:20 Gailby joined #git
14:20 andrewSC yeah
14:20 _ikke_ soft resetting is history rewriting
14:20 _ikke_ But only moving refs
14:20 grawity well partially
14:20 andrewSC either way if you're doing that you're already f-ing up hard
14:20 neunon joined #git
14:20 Mo0O I'm looking for a solution to ignore a given folder on every branch except one, do you know how to do that?
14:20 grawity it throws away some history, but doesn't "recreate" any
14:21 mjuszczak joined #git
14:21 UncleCJ joined #git
14:21 andrewSC so tl;dr I'll be fine, just follow best practices
14:21 andrewSC etc
14:21 _ikke_ Mo0O: Not in a sensible way
14:22 Gailby left #git
14:22 stamina joined #git
14:22 Mo0O that's what I though too
14:22 MattMaker joined #git
14:22 hahuang61 joined #git
14:23 chachasmooth joined #git
14:23 mjuszczak joined #git
14:23 bakibour joined #git
14:24 bakibour left #git
14:24 bourbaki joined #git
14:24 bourbaki Hey
14:24 bourbaki What can i do to speed up a git repository?
14:24 _ikke_ speed up what exactly?
14:24 bourbaki A git clone takes about 1h our our repository these days.
14:25 Mo0O --depth
14:25 joegreen88 joined #git
14:25 tango_ also pack server side
14:25 bourbaki There are no submodules
14:25 Mo0O git clone --depth 10 ...
14:25 bourbaki or what is the depth for?
14:25 Mo0O depth is not about submodules
14:25 bourbaki Ok
14:25 _ikke_ How many commits to clone
14:25 Mo0O --depth only the last 10 commits
14:25 _ikke_ Which as limits the usefulness
14:25 bourbaki It just clones a part of the history?
14:25 Mo0O --depth 10
14:25 Mo0O yes
14:25 bourbaki Ok
14:25 anticom joined #git
14:26 bourbaki So for a build server --depth 1 would be totally ok?
14:26 Mo0O yeah
14:26 _ikke_ Mo0O: Note that for github adding --dept=1 caused a lot of performance issues for them (although, there were other factors contributing to that too)
14:26 _ikke_ bourbaki: For a build server, using git archive would perhaps even be better
14:26 bourbaki performance issues as in good or bad?
14:26 bourbaki What does that do exactly?
14:27 Mo0O thanks for the advice _ikke_
14:27 bourbaki Why are the repos get so slow?
14:27 _ikke_ bourbaki: performance issues are bad
14:27 bourbaki Would it be wise to split it up more and abandon a repo once every x years?
14:29 _ikke_ bourbaki: I'm not that well version in those details, but on the !mailing_list they could probably give you some helpfull advise speeding things up if you give enough information about the repo
14:29 gitinfo bourbaki: The mailing list can be reached via git@vger.kernel.org. You don't need to subscribe to the list, you will always be put in cc on reply. Read archives at http://j.mp/gitlist
14:29 fakenerd joined #git
14:30 bourbaki Is this not more or less a general problem that i can read about somewhere?
14:31 _ikke_ bourbaki: there probably is, but I don't know from the top of my head
14:31 yaun joined #git
14:31 Jellyg00se joined #git
14:32 rubyonrailed joined #git
14:33 TPwFEgmQkn joined #git
14:33 watabou joined #git
14:35 fuchstronaut joined #git
14:36 encod3 joined #git
14:36 sebboh joined #git
14:37 ArchNoob joined #git
14:38 chachasmooth joined #git
14:38 subhojit777 joined #git
14:43 al-damiri joined #git
14:44 QwertyDragon joined #git
14:44 kaldoran joined #git
14:44 pkiller joined #git
14:44 Tsutsukakushi joined #git
14:45 tomboy65 joined #git
14:45 Torrone joined #git
14:47 mikecmpbll joined #git
14:48 bjoe2k4 joined #git
14:48 LionsMane joined #git
14:48 crayon joined #git
14:49 thiago joined #git
14:50 shinnya joined #git
14:51 stomplee joined #git
14:52 raijin joined #git
14:53 crayon joined #git
14:57 UTAN_dev joined #git
14:57 jceb joined #git
14:58 JamesBaxter joined #git
14:58 _Curly060_ joined #git
14:58 _Curly060_ jow: you are the man! thanks so much. it works again!
15:01 dreiss joined #git
15:02 nwkj86 joined #git
15:04 zivester joined #git
15:04 vassagus joined #git
15:04 griffindy joined #git
15:05 invisbl joined #git
15:07 notebox joined #git
15:07 yaun joined #git
15:09 Torrone joined #git
15:10 rominronin joined #git
15:12 WJMUqqJZRD joined #git
15:13 InfoTest joined #git
15:14 heroux_ joined #git
15:15 yaun left #git
15:16 Darren_ joined #git
15:16 cyjobes joined #git
15:17 danslo joined #git
15:17 MattMaker joined #git
15:17 cyjobes Greetings
15:20 ochorocho__ joined #git
15:20 cyjobes I am new ti GIT, as will be obvious by my question - I had a branch with uncommitted changes and new files. I create a branch from the one I was working on. Played around, then decided to go back to all original branch. Sourcetree asked if I wanted to save local changes and I said no. Now all the uncommitted files in the first branch are gone. any way to get them back?
15:21 _ikke_ cyjobes: did you git add / stage the files?
15:21 cyjobes No
15:21 dmto joined #git
15:21 _ikke_ Undo buffers in your editor?
15:22 _ikke_ or history
15:22 MattMaker joined #git
15:23 cyjobes _ikke_: Using PHPStorm. It also is connected to GIT. No undo available fromwhat Ican see.
15:23 _ikke_ phpstorm keeps a history of files
15:23 _ikke_ right click on the files, and look for history
15:24 chimay joined #git
15:24 sebboh joined #git
15:26 snowkidind joined #git
15:26 realz joined #git
15:27 ertes-t6v left #git
15:27 ertes joined #git
15:27 ertes why was i banned?  did my connection suck?
15:27 hobodave joined #git
15:28 MattMaker joined #git
15:28 Gitzilla joined #git
15:29 daniel_rose joined #git
15:29 osse ertes: I can't find any trace of a ban
15:30 _sjs joined #git
15:31 unbalanced joined #git
15:31 ertes weird
15:32 ertes *** morgan.freenode.net 435 ertes-t6v ertes #git Cannot change nickname while banned on channel
15:32 ertes couldn't find a matching +b or +q
15:32 osse Could it come from not being authed with nickserv?
15:35 Juzzika joined #git
15:35 ochorocho__ joined #git
15:35 _ikke_ ertes: This is because we require registered users
15:35 _ikke_ which prevents you from changing names while in this channel (and others)
15:36 fuchstronaut joined #git
15:37 MattMaker joined #git
15:37 ertes ah, that makes sense
15:39 ahmed_elgabri joined #git
15:39 mingrammer joined #git
15:40 chimay joined #git
15:42 LebedevRI joined #git
15:43 jordila1 joined #git
15:43 jordila1 left #git
15:43 LebedevRI hi. in git 1:2.9.3-1, was there any changes in how git-gui handles gpg-signing of commits? i think right after update to this version, it no longer does sign them.
15:43 sandstrom joined #git
15:43 jordila1 joined #git
15:43 jordila1 left #git
15:44 moneylotion joined #git
15:44 grawity that was reported in the mailing list, I think
15:45 nettoweb joined #git
15:46 LionsMane joined #git
15:47 diogenese joined #git
15:47 morenoh149 joined #git
15:48 dreiss joined #git
15:48 anonymuse joined #git
15:48 jangeador joined #git
15:49 Lachezar joined #git
15:50 Lachezar Hey all. I have some repositories that have been created with 'clone --bare'. Is it possible to make them the same as if I had 'clone --mirror' instead?
15:50 _ikke_ Lachezar: It's a matter of changing the refspec
15:50 Lachezar _ikke_: How?
15:50 cyjobes _ikke_: Great news. Thanks to PHPStorm's history of changes, I was able to restore all files that were reverted or deleted to the older versions. LOVE PHPStorm!
15:51 _ikke_ cyjobes: :-)
15:51 elect joined #git
15:52 cyjobes _ikke_: Lesson learned - COMMIT changes every day or hour from now on.
15:52 _ikke_ Lachezar: git config remote.origin.fetch "+refs/*:refs/*"
15:52 Lachezar _ikke_: Thank you.
15:53 Lachezar _ikke_: And then do a 'fetch' I presume?
15:53 _ikke_ yes
15:53 harish joined #git
15:54 _ikke_ and optionally remove all the refs/remotes/* refs from the initial clone
15:54 Zimmedon left #git
15:54 jangeador joined #git
15:54 ertesx joined #git
15:54 _ikke_ Might need to do that first
15:55 Darcidride joined #git
15:55 GavinMagnus left #git
15:56 chrisshattuck joined #git
15:56 Lachezar _ikke_: That sounds like I need to 'git config --unset remote.origin.fetch' and then 'git config remote.origin.fetch "+refs/*:refs/*"'
15:57 Nugget joined #git
15:57 le_melomane joined #git
15:58 fusionx86 joined #git
15:58 EvilPeng1 joined #git
15:58 vuoto joined #git
15:58 mjuszczak joined #git
15:59 madewokherd joined #git
15:59 _ikke_ No, it will override it
15:59 _ikke_ But you need to remove the refs self
15:59 durham joined #git
15:59 Nugget joined #git
16:00 preyalone joined #git
16:00 _ikke_ git for-each-ref "refs/remotes" --format="delete %(refname)" | xargs git update-ref --stdin
16:00 Lachezar _ikke_: Hm… I'm lazy: is removing …/.git/refs sufficient?
16:00 OnBrokenWings joined #git
16:01 romerocesar joined #git
16:01 _ikke_ ^^
16:01 _ikke_ I provided a command
16:01 zacts joined #git
16:01 mjuszczak joined #git
16:02 preyalone Which git config settings should I use for a repository that doesn't have a preference for LF or CRLF; it just passes through files as-is? E.g., in a multi-platform project
16:02 _ikke_ echo "* -text" >>.gitattributes
16:04 Fissure joined #git
16:04 LebedevRI grawity: :( i was hoping it is a local configuration issue. reported to my distro https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=835428
16:04 LebedevRI thanks
16:05 vassagus_ joined #git
16:05 Phlogistique joined #git
16:05 Phlogistique Hi
16:06 Phlogistique I'm pretty sure my Git send-email settings for GMail used to work
16:06 Phlogistique Now they do not work anymore. Instead, git send-email waits for a long time and then prints:
16:06 tshirtman left #git
16:06 invisbl joined #git
16:07 morenoh149 how can I use git-diff to compare two files, same commit+branch ?
16:07 Phlogistique Unable to initialize SMTP properly. Check config and use --smtp-debug. VALUES: server=smtp.gmail.com [...]
16:07 Phlogistique adding --smtp-debug does not change anything
16:08 Phlogistique morenoh149: git diff foo:bar.txt foo/baz.txt
16:08 _ikke_ Phlogistique: Not sure it helps, but try GIT_TRACE=1 git send-email
16:08 rominronin joined #git
16:09 Phlogistique morenoh149: typo; I meant: git diff foo:bar.txt foo:qux.txt
16:09 Phlogistique _ikke_: trying that, thanks
16:10 lungaro joined #git
16:11 UTAN_dev_ joined #git
16:11 lungaro Hello. I am writing a script to do a automated git pull. I want to hide the output, but still see just the files that have changed. Is there a way to do this? I can't seem to get "git pull --dry-run --stat" to function
16:12 morenoh149 Phlogistique: what's foo in that case?
16:12 sebboh What do you call all the terminal points on the DAG?  Like, HEAD (normally), or other commits that don't have children yet?
16:12 lungaro sebboh, guessing.. but leafs?
16:12 Phlogistique morenoh149: the commit/branch you want to take the files from
16:12 sebboh ok.  I want to search all leafs in a repo for a word.  What do I do?
16:12 Phlogistique _ikke_: that does not print anything interesting
16:13 morenoh149 Phlogistique: can it infer from the current environment?
16:13 Phlogistique morenoh149: what do you mean? you want to compare files in HEAD?
16:13 morenoh149 in working tree
16:14 kadoban morenoh149: Why don't you just use diff itself?
16:14 morenoh149 kadoban: okay
16:14 harish joined #git
16:16 ReenignE lungaro: separate commands
16:16 ReenignE Assuming a branch of develop: git fetch && git diff --stat develop..origin/develop && git pull &> /dev/null
16:16 lungaro fetch -q first
16:16 nitric joined #git
16:16 lungaro I still see a few lines I dont want still but it should be ok
16:16 ReenignE Which?
16:16 ReenignE The summary at the end?
16:17 lungaro Updating 43c9c87..421111c
16:17 lungaro Fast-forward and the summary
16:17 lungaro all I really want is a list of what changed, like git status -s
16:17 raijin joined #git
16:18 UTAN_dev joined #git
16:18 lungaro its ok, i am just being too pedantic on this scripts output
16:18 ReenignE Silencing the fetch and pull should prevent the fast forward and updating messages from showing.
16:18 ReenignE The git diff --stat shows the files affected with a one line summary at the end.
16:19 lungaro oh I see. Let me try doing what you showed. I was silencing just the fetch
16:19 vassagus joined #git
16:19 ReenignE I didn't include -q in the fetch part of my command.
16:19 vuoto joined #git
16:20 ReenignE Also not certain if it prevents all output, &> /dev/null should.
16:20 lungaro fantastic
16:21 vuoto joined #git
16:22 lungaro thanks, this is nice. I can live w/ the summary
16:22 sdothum joined #git
16:22 ReenignE No problem glad it helped.
16:22 watabou joined #git
16:22 synthroid joined #git
16:23 hahuang61 joined #git
16:23 PHPanos joined #git
16:25 cschneid joined #git
16:26 raijin joined #git
16:27 a3Dman joined #git
16:28 ertes how can unreachable commits survive `git gc --prune=all --aggressive`?  they still show up in my reflog, and i can still `checkout` them
16:28 danslo joined #git
16:28 ertes i also tried `git prune`, but it didn't help either
16:29 Seveas ertes: they survive because they're in your reflog
16:29 phanimahesh joined #git
16:29 Seveas in the reflog == reachable enough to not be gc'ed
16:29 canton7 search for "If you really don’t want to clone it" in man git filter-branch
16:29 gitinfo the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html
16:30 ertes ah
16:30 joki joined #git
16:31 glebihan joined #git
16:32 Seveas ReenignE: try 'git diff --stat develop..origin/develop' is a bit brittle. Try git diff @{upstream}, works on other branches too
16:33 UTAN_dev_ joined #git
16:33 MattMaker joined #git
16:33 thiago joined #git
16:34 Nugget joined #git
16:34 iron_houzi left #git
16:34 mingrammer joined #git
16:35 ertes thanks
16:35 ertes that makes sense
16:38 dreiss joined #git
16:38 anarchos2 joined #git
16:38 xaviergmail joined #git
16:39 zincrokx joined #git
16:39 beyertech joined #git
16:40 vuoto joined #git
16:40 vuoto joined #git
16:44 rafalcpp osse: yeah that   git diff HEAD...devel/branch is showing nicelly what will change after merge.  Is there any situation in which it would not show accuretly silently?  Other then merge conflicts, but they will be noticed ofc
16:44 rafalcpp *when it would silently show other diff and missinform the reviewer
16:46 ReenignE Seveas: @{upstream} is more flexible yeah, curious if it works when a local branch is not tracking a remote branch though which I've seen often.
16:47 dmto joined #git
16:47 ReenignE Co-workers using guis have no upstream set for their branches, git branch -vv only shows the local ref
16:47 MattMaker joined #git
16:48 jeremyhall joined #git
16:49 aard_ joined #git
16:49 HoierM joined #git
16:50 romerocesar joined #git
16:52 synthroi_ joined #git
16:52 MattMaker joined #git
16:52 daynaskully joined #git
16:52 vuoto joined #git
16:53 ttvsesh joined #git
16:53 vuoto joined #git
16:54 vuoto joined #git
16:54 boombatower joined #git
16:54 vuoto joined #git
16:54 ReenignE Yeah unsurprisingly @{upstream} requires that an upstream be set for the current local branch which isn't always the case.
16:56 chachasmooth joined #git
16:58 LUMIA930 joined #git
16:58 osse rafalcpp: conflicts git is able to solve by itself
16:58 osse rafalcpp: e.g. the hunk is applied 10 lines higher or lower. or some of the context might have changed
17:02 MacWinner joined #git
17:02 theverbg_ joined #git
17:02 invisbl joined #git
17:02 MattMaker joined #git
17:03 absk007 joined #git
17:03 chrisshattuck joined #git
17:04 Sasazuka joined #git
17:05 ttvsesh i'm trying to figure out exactly what happened with some mysteriously disappearing commits
17:06 ttvsesh i wrote a branch, created a pull request on github enterprise, merged it
17:06 LUMIA930 Hi
17:06 gitinfo LUMIA930: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.
17:07 ttvsesh person B somehow created a branch, created a pull request and got some funny messages, their branch somehow now included my changes
17:07 dreiss joined #git
17:07 dviola joined #git
17:07 ttvsesh so they deleted my changes as they didn't recognize them, and merged the pull request. now my code is out of master and my commits don't exist
17:07 invisbl joined #git
17:07 ttvsesh i'm trying to figure out exactly what went wrong
17:08 MattMaker joined #git
17:08 thiago joined #git
17:08 ttvsesh like maybe they rebased their branch to an old local master?
17:08 invisbl joined #git
17:09 ttvsesh it seems like what they did should not be possible without a forced push
17:09 rominronin joined #git
17:09 Electrometro joined #git
17:12 diphtherial left #git
17:12 MacWinner joined #git
17:13 rominronin joined #git
17:13 sagerdearia joined #git
17:14 freimatz joined #git
17:14 fakenerd joined #git
17:15 Lachezar _ikke_: Didn't quite work: either the 'xargs' or the --stdin is superfluous :(
17:16 beyertech joined #git
17:16 rahtgaz joined #git
17:16 jtprog_ joined #git
17:17 beyertech joined #git
17:17 [cust] joined #git
17:18 MattMaker joined #git
17:19 newbiee joined #git
17:20 newbiee is it possible to apply same patch file multiple times? I am getting this weird problem where my patch is getting applied multiple times without failing the second time.
17:21 nullie I think this can work if you have repeating lines
17:22 newbiee nullie: didn’t get you.
17:23 shaggycat joined #git
17:23 MattMaker joined #git
17:23 nullie If you patch deletes foo and file has several foo lines
17:23 nullie With same context
17:25 AnarchyAo joined #git
17:27 LUMIA930 git checkout <LATEST cartodb-postgresql tag>
17:27 LUMIA930 bash: syntax error near unexpected token `newline'
17:27 LUMIA930 What am getting this error
17:29 grift joined #git
17:29 osse LUMIA930: what is the exact command?
17:30 Rodya_ joined #git
17:30 SteffanW joined #git
17:31 Mobutils joined #git
17:31 LUMIA930 osse am trying to check out the CartoDB package
17:31 LUMIA930 so thats check will drops the latest files right ?
17:32 MattMaker joined #git
17:35 LUMIA930 osse: git checkout <LATEST cartodb-postgresql tag>
17:35 LUMIA930 Thats the command i have used to get it
17:35 LUMIA930 but that gave me error as newline
17:36 ojacobson with unquoted < and > ?
17:36 ojacobson git-tag doesn't read stdin, and redirecting stdout requires a destination like a filename
17:36 bourbaki joined #git
17:37 fuchstronaut joined #git
17:37 ojacobson I'm almost certain the docs you're reading expect you to replace the whole string "<LATEST cartodb-postgresql tag>" with a tag name
17:38 bourbaki earlier i was told that i can use clone in conjunction with --depth to speed up cloning a repo. I read a bit about realated issues though.
17:38 bourbaki When i pull after lets say i cloned with a depth of 100 commits will that work?
17:39 bourbaki I would like to be able to clone shallowly for a build server but also would like to be able to pull stuff in so i can keep the workspace and build incrementally
17:39 rpd bourbaki: shouldn't be a problem, you truncated the past history, not the future
17:40 jnewt joined #git
17:40 bourbaki I read stuff, that as i have to admit can not verify, someone cloned with a depth of 1 and pulled changes in which git was not able to merge then
17:41 jnewt i want to clone the subdirectory of a project, but that subdirectory's name was changed somewhere in history.  how do i get the clone to include all of the relevant history, but only clone the subdirectory as the root directory?
17:41 phroa joined #git
17:41 rpd bourbaki: in that situation, that's way too shallow. To merge, git needs to be able to find a common ancestor.
17:41 rpd bourbaki: in that person's history book, the common ancestor never exists so there's no relation
17:41 sdothum joined #git
17:42 LUMIA930 ojacobson so i need to find the file name in git web page ?
17:42 bourbaki i see
17:42 rpd 100 commits is prolly okay, but you could also easily run into problem with merging things that branches before those 100 commmits existed
17:42 bourbaki well at work we have a shit ton of branches
17:42 bourbaki maybe that might introduce some problems also
17:43 bourbaki It is just that cloning our repo takes 1h
17:43 bourbaki is that normal
17:43 bourbaki ?
17:43 bourbaki for a 5 year old project
17:44 Noldorin joined #git
17:44 jnewt if it takes that long here, we almost always have someone that committed some binary or images or something that they shouldn't have and it's just huge.
17:44 rpd Maybe, really depends on what's stored in the repo. A 5-year old project with only a README and no commits will take no time. A 2 day old project with 100k commits over 1k branches will take somewhere closer to an hour
17:45 bourbaki at some time we had a lot of binaries in there
17:45 fakenerd joined #git
17:45 bourbaki and still have some test data in there
17:45 bourbaki we had 3rd party libs and stuff in there
17:45 bourbaki now we have them in nuget, like 11gb which takes another hour to fetch ...
17:46 jnewt i'd clean it out, but your policy may prevent it
17:46 rpd Sometimes it is unavoidable. I clone and pull gcc constantly and it similarly takes forever.
17:46 bourbaki no idea if we even have a policy :)
17:46 rpd Instead of shallow cloning history, do you need every single branch?
17:46 bourbaki Personally i would like to not build any branches on the build server
17:46 rpd For example, I found it was more useful to clone --single-branch than to limit by --depth
17:47 rpd in terms of clone size and time taken to clone
17:47 bourbaki ppl can do whatever they want i branches but if i only had to build like develop, release and whatever branches it would be al ot easier to build incrementally
17:47 rpd but gcc has like 50 branches with a ton of history
17:47 bourbaki i read that also
17:47 mariuscc joined #git
17:47 bourbaki i think im going to test that out tomorrow
17:47 rpd bourbaki: sounds like you should only pull those branches and see if that helps your clone speed
17:48 jnewt bourbaki, i'd explain the problem, then explain that you'd like to clean them out (or have someone else do it).  i've had situations where a repo would go from 1hr to 2min just by cleaning out a build binary that was getting regenerated every commit and therefore taking 90% of the space
17:48 bourbaki we only have two guys at work who are somewhat familiar with git
17:48 bourbaki what would cleaning the repo involve?
17:49 bourbaki i really would like to get smaller modules of the project anyway and put every module in its own repo
17:49 jnewt well, i'd test locally first (git filter-branch) and see what the result was.
17:49 kadoban joined #git
17:49 bourbaki ok
17:49 bourbaki and then build packages for those, no subrepos
17:50 jnewt then if the results were good, i'd have the other couple of guys that use it clean up their local repos, commit, push and delete their repo's
17:50 bourbaki we use git like svn anyway
17:50 jnewt then i'd backup the server bare repo, run the filter-branch there on those files, and have everyone else clone back their local work
17:50 jnewt bourbaki, we use it centralized too.  I think it's more common that way.
17:50 bourbaki so setting up a new repo is not needed?
17:51 bourbaki yeah well i think it is ok, but this way you loose a lot of the good stuff
17:51 jnewt not for filter-branch --treefilter and something like 'rm bigassfile.binary'
17:51 bourbaki and dont get a lot of the good stuff centralized repos have
17:51 ki0 joined #git
17:51 bourbaki for example i am not able to always tell in what branch i am
17:52 realz joined #git
17:52 bourbaki which would be nice in terms of testing, because i could use the branch name to configure tests
17:53 bourbaki do you guys at work test a lot of branches?
17:53 vassagus_ joined #git
17:53 cschneid joined #git
17:54 jnewt why would you not know what branch you are on?
17:54 multi_io joined #git
17:54 xender joined #git
17:54 bourbaki I had the problem that when i pulled my own repo with unpushed comits in it to test some stuff i get HEAD instead of a branch name.
17:55 bourbaki And i was told by our git dude that since commits can also be in several branches i can not always get a useful answer.
17:55 kadoban joined #git
17:56 xender Hi. In a repository in two remotes, let's say 'upstream' and 'origin', is there a way to automatically push all branches from upstream to origin?
17:56 xender My use case is updating a forked repository
17:57 rpd xender: branch-name is optional, right? Couldn't you just do git push upstream or git push origin?
17:57 xender Bonus points for it being compatible with or equivalent to --force-with-lease, i.e. force updating everything on origin to be a mirror of upstream, but only if a branch on origin has not diverged from remote-tracking branch
17:57 rpd Assuming you have all of those branches on your local workspace
17:57 BlaXpirit rpd, without branch name  current branch is pushed
17:57 jnewt borbaki: don't clone your own repo to do tests, just do git checkout -b mybranch_test, do your stuff, then do git checkout mybranch and git branch -D mybranch_test
17:57 xender Well, depends on the git config, but I don't want to mess with the actual config
17:58 xender rpd, BlaXpirit: ^
17:58 BlaXpirit xender, do you actually care about updating all branches on origin, or is it OK to only update everything locally?
17:58 antipsychiatry joined #git
17:59 rpd uh anyways i think the command you're looking for is git push upstream --all or git push upstream --mirror, etc
17:59 bigcat joined #git
17:59 theverbg_ joined #git
17:59 xender BlaXpirit: I want to make origin a mirror to upstream, but not the branches where remote origin diverges from remote-tracking branches of origin in my local repo
17:59 mikecmpbll joined #git
18:00 BlaXpirit i think the most important first step to maintain forks is to make upstream the primary remote
18:00 romankl joined #git
18:00 xender For updating remote-tracking branches in local repo I just use git remote -v update
18:00 BlaXpirit "origin" becomes "fork", "upstream" becomes "origin",   then you can  'git pull'  to update everything anytime
18:00 xender I don't use git pull anyway
18:00 BlaXpirit and maybe then  pushing everything as rpd suggests will work
18:01 msonntag joined #git
18:01 antipsychiatry joined #git
18:02 [cust] joined #git
18:02 xender Hm, even if I set default push mode to 'matching', it will push matching local branches to there remote-tracking ones, not remote_a/branches to remote_b/branches
18:02 pikajude_ left #git
18:02 synthroid joined #git
18:02 GenZai joined #git
18:03 MattMaker joined #git
18:03 xender And I used names remote_a and remote_b because it doesn't really matter which one is named 'origin' AFAIK, because only local branches can be set to push to a remote-tracking branch by default, right?
18:04 linuxmodder joined #git
18:04 jnewt so git-filter-branch --subdirectory-filter for a subdirectory which had it's name changed somewhere in history.  possible?
18:04 pikajude joined #git
18:05 sparr joined #git
18:05 bittin joined #git
18:05 jnewt does this remap to ancestor thing work for my situation?
18:06 vassagus joined #git
18:06 sdothum joined #git
18:08 jnewt yeah, doesn't work
18:09 mingrammer joined #git
18:09 [cust] joined #git
18:10 voidfire joined #git
18:11 Gsham joined #git
18:12 kadoban joined #git
18:14 adac joined #git
18:14 rominronin joined #git
18:14 daniel_rose joined #git
18:18 MattMaker joined #git
18:18 NomNKDJIvx joined #git
18:20 LeBlaaanc joined #git
18:20 mahesh joined #git
18:20 alexsystemf__ joined #git
18:21 falkenb0t joined #git
18:22 vassagus joined #git
18:22 falkenb0t left #git
18:22 zeroed joined #git
18:22 zeroed joined #git
18:24 hahuang61 joined #git
18:25 greymalkin joined #git
18:26 jnewt let's say i do the subdirectory-filter in two different clones (one for each folder name).  could i then combine the two repos to have full history again?
18:29 h264 joined #git
18:29 jokester joined #git
18:30 alansaul joined #git
18:31 _alex90982_ joined #git
18:32 clemf joined #git
18:33 fusionx86 joined #git
18:34 _alex123123_ joined #git
18:34 kpease joined #git
18:34 fahadash joined #git
18:35 _alex123123_ Hello, I have a question about git-diff and the man pages didn't realy help.
18:35 _alex123123_ To be precise, I do not understand what git diff prints as outpout when 3 commits are given as arguments
18:35 jordan__ joined #git
18:37 _alex123123_ Is there anyone who can help?
18:37 kpease_ joined #git
18:37 WhiskerBiscuit joined #git
18:38 fuchstronaut joined #git
18:39 Darcidride joined #git
18:40 shinnya joined #git
18:40 LeBlaaanc joined #git
18:41 ash_workz joined #git
18:41 Torrone joined #git
18:41 ahmed_elgabri joined #git
18:43 jstein_ joined #git
18:46 fakenerd joined #git
18:46 YqwHkAMaYX joined #git
18:46 sssilver joined #git
18:47 miklcct joined #git
18:47 mikecmpbll joined #git
18:47 daniel_rose joined #git
18:47 vF3hNGxc47h8 joined #git
18:47 GavinMagnus joined #git
18:48 MattMaker joined #git
18:49 kadoban joined #git
18:51 WhiskerBiscuit joined #git
18:51 Gsham joined #git
18:54 sandstro_ joined #git
18:54 Torrone joined #git
18:55 kfunk joined #git
18:56 zeroed joined #git
18:57 nullie can I somehow commit tree state without modifying index?
18:57 nullie like maybe backup index and then restore it
18:58 nettoweb joined #git
18:58 roelmonnens joined #git
18:59 AaronMT joined #git
19:00 ojacobson nullie: in principle, yes
19:00 ojacobson in practice, stash, or back up the index (it's just a file) and restore it
19:00 nullie okay
19:01 GavinMagnus left #git
19:01 zmachine joined #git
19:02 MattMaker joined #git
19:02 AaronMT joined #git
19:03 mingrammer joined #git
19:04 newbiee_ joined #git
19:05 daniel-wtd joined #git
19:06 netcarver joined #git
19:06 _alex123123_ I have two branches a and b, and a commit used as merge base between those two
19:07 rgrinberg joined #git
19:07 _alex123123_ I'm trying to find which files appear in "git status" if while on branch a i do "git merge --no-commit b"
19:08 brb joined #git
19:08 _alex123123_ I thought that those are the files modified from merge-base to b
19:08 _alex123123_ However this does not seem to be the case.
19:09 _alex123123_ If branch a and branch b both have modified file foo, but the end result of file foo is the same in both branches, foo will not appear in "git status"
19:09 matumbo I'm having a problem with "git config commit.gpgsign true" having no effect. I have specified key with "git config --global user.signingKey xxxxxxx". When commiting with: "git commit -m "very good comment" . " the commit is not signed
19:09 matumbo Does anyone have an idea of what I'm missing?
19:10 rgrinberg joined #git
19:11 ProbabilityMoon joined #git
19:11 crelix joined #git
19:11 ArchNoob joined #git
19:12 kadoban joined #git
19:12 nwkj86 joined #git
19:13 LebedevRI left #git
19:13 crelix joined #git
19:15 rominronin joined #git
19:16 matumbo When using "git commit -S -m "very good comment" . " I'm prompted for my key and the signing operation works as intended, but my goal is to sign every commit for a specific repo without the "-S" option
19:16 docnvk joined #git
19:19 JamesBaxter joined #git
19:20 docnvk joined #git
19:21 sandstrom joined #git
19:22 OMGilluno joined #git
19:22 anonymuse joined #git
19:22 OMGilluno left #git
19:23 Torrone joined #git
19:23 Rodya_ joined #git
19:23 t0by joined #git
19:25 antipsychiatry joined #git
19:26 zeroed joined #git
19:27 roelmonnens joined #git
19:28 Darcidride Hi, im trying to use the command line gcp (graphical copy file, for progress bar, with percentage, etc...) and it seems my git packages overrite this command, because "gcp" is an alias of git cherry-pick... so, how can i say to my gcp command line, it's the gcp tool, not the git tool i want to call ? thank you in advance for your help
19:29 bremner Darcidride: that's not a built in alias
19:29 bremner so, change your bashrc / aliases?
19:30 ProbabilityMoon Darcidride use fully qualified path for gcp command, like: /usr/bin/gcp
19:30 terminal_echo joined #git
19:31 Darcidride bremner, Well, i don't create any aliases in my bashrc/zshrc file
19:31 Darcidride ProbabilityMoon, good idea, i didn't have it yet, it should work, but it's a workaround, a definitive solution doesn't exist ?
19:32 ProbabilityMoon Darcidride either that or change your alias like bremner said
19:34 ProbabilityMoon see unalias command
19:36 nettoweb joined #git
19:36 pasturesofplenty left #git
19:38 MattMaker joined #git
19:38 shymega joined #git
19:39 rgrinberg joined #git
19:41 newbiee joined #git
19:42 jnewt i am trying to split a commit in my history.  i do git rebase -i <SHA TO SPLIT>^, select edit on the one i want to commit, git reset HEAD^, then add and commit until my directory is clean.  all good but..
19:42 daniel_rose joined #git
19:43 jnewt if i diff the HEAD with the <SHA TO SPLIT> it isn't the same, (shows modified files) and i can't continue the rebase without conflicts.
19:44 rgrinberg joined #git
19:44 Torrone joined #git
19:47 clemf_ joined #git
19:47 fakenerd joined #git
19:47 crayon joined #git
19:48 johnmilton joined #git
19:51 ki0 joined #git
19:56 netj joined #git
19:56 fuchstronaut joined #git
19:58 crayon joined #git
19:58 pothepanda joined #git
19:59 davisonio joined #git
19:59 adac joined #git
20:00 anonymuse joined #git
20:01 hobodave_ joined #git
20:02 m0viefreak joined #git
20:06 anonymuse joined #git
20:06 xaviergmail joined #git
20:07 fuchstronaut joined #git
20:07 stamina joined #git
20:08 MattMaker joined #git
20:08 p4trix joined #git
20:08 nwkj86 joined #git
20:09 johnmilton joined #git
20:11 Munt joined #git
20:14 p4trick joined #git
20:14 theverbg_ joined #git
20:14 foist joined #git
20:14 dsdeiz joined #git
20:14 cdg joined #git
20:15 skalpin jnewt: you can't continue the rebase without conflicts? meaning you cant git rebase --continue?
20:16 jnewt skalpin, correct
20:16 tvw joined #git
20:16 rominronin joined #git
20:16 skalpin is it telling you to rebase --skip instead?
20:16 jnewt it's the line endings.  i can't seem to find a setting that allows me to work without headaches on both win and linux.
20:17 jnewt i have to turn it off in this case to get it to continue, turn it on other times to avoid having diffs and committing changes just because of a CR
20:17 MattMaker joined #git
20:17 GodGinrai jnewt: what kind of dev do you do on Windows?
20:17 cpg joined #git
20:18 cpg hello, i tap on this most friendly channel once again. I keep on needing this and I’m not sure how to do it:
20:19 cpg i want a diff of the branch i have just checked out against the “root” of that branch,
20:19 madewokherd my solution has always been commit/checkout as-is, and use a text editor that writes unix line endings; but that depends on the tools you're using and who else uses the repository
20:19 jnewt GodGinrai: all kinds.  most of my git stuff is for microcontroller firmware & pcb hardware.  i have some software that won't work on my linux machine, and some stuff that we only have on linux.  it's a real pita
20:20 _ikke_ cpg: You need to know the branch this branch was made on
20:20 LionsMane joined #git
20:20 _ikke_ cpg: then you can do: git diff <parent-branch>...
20:20 cpg the root is typically master and has moved on from that root point
20:20 _ikke_ note the three dots
20:20 _ikke_ so git diff master...HEAD
20:20 jnewt madewokherd, the problem is some of these win tools automatically put a CR in there.  i wish i could just flush out all the CR's automagically, as they all will read a LF only, they just like to put the CR in there.
20:21 _ikke_ (HEAD is optional)
20:21 cpg _ikke_: niiice. worked well
20:21 GodGinrai jnewt: I would just commit as-is, checkout as-is.  Anything made by those windows-specific apps don't need to be viewed in linux, and anything made in linux can just be viewed in a proper text editor on windows
20:21 cpg thanks much _ikke_
20:21 _ikke_ cpg: no problem
20:22 cpg this channel is remarkably friendly, consistently. no pompous neckbeards telling people to RTFM. maybe it’s linus’ character being channeled
20:22 madewokherd somehow I thought there was a setting for converting the line-endings when you commit only
20:22 cpg (pun was not intended in “channelled”, but welcome!)
20:23 madewokherd though you potentially have to convert the whole repo if files are already committed with CR's
20:23 astrozyk joined #git
20:23 _ikke_ cpg: Yes, we try to be friendly here
20:24 GodGinrai jnewt: wait, if they all will read lf-only, there's plenty of ways to fix this
20:24 cjohnson Is there a command to tell me which branches are prunable? Or is this a GH feature?
20:25 hahuang61 joined #git
20:25 cjohnson I remember seeing a list of branches which were all already merged into master with no additional commits
20:25 raijin joined #git
20:25 _ikke_ cjohnson: git branch --merged
20:25 _ikke_ shows branches merged into the current branch
20:25 _ikke_ (use git branch -r merged to show remote tracking branches)
20:26 cjohnson ok so invert that list, run from master
20:26 GodGinrai jnewt: jnewt set core.autocrlf to input to convert to lf on commit, but checkout as-is
20:26 cjohnson er nevermind no inversion
20:26 _ikke_ cjohnson: --no-merged also exists, but yes, no inversion needed
20:27 GodGinrai whoops, double highlight
20:27 MattMaker joined #git
20:27 cjohnson oh sexy
20:28 dodobrain joined #git
20:28 vuoto joined #git
20:28 johnmilton joined #git
20:29 cjohnson ok one more question for now. When I type git push origin :some-b<tab> instead of completing to some-branch, it completes to :origin/some-branch
20:29 cjohnson is that the same thing?
20:29 cjohnson (and could I just type git push :some-<tab> -> git push :origin/some-branch  for the same result?)
20:30 _ikke_ no, that's something different
20:30 cjohnson to which
20:31 gusnan joined #git
20:33 davisonio joined #git
20:33 _ikke_ pushing to :some-branch is different from pushing to :origin/some-branch
20:34 nwkj86 joined #git
20:34 clemf joined #git
20:36 cjohnson got it, wonder why it completes like that
20:37 osse hmm, the sensible thing to do would be to complete your remote branches (e.g. origin/some-branch) but strip off the origin/
20:41 _nwkj86 joined #git
20:42 raijin joined #git
20:43 anonymuse joined #git
20:43 jstvz joined #git
20:44 napping joined #git
20:44 EY joined #git
20:44 napping I want to do git filter-branch --subdirectory-filter ... in a repository I just made with clone -n. Is there any way to make it run despite unstaged/uncommited changes?
20:45 _ikke_ napping: run git stash first?
20:46 napping the problem is it has tons of files. I'd like to filter without unpacking them all into a working directory
20:46 Torrone joined #git
20:48 fakenerd joined #git
20:49 EY Hi. Is there anyway to make `git rebase -i ` faster?
20:50 _ikke_ EY: Get faster storagE?
20:50 EY _ikke_, faster storage on client or server?
20:50 _ikke_ rebase only does things on client
20:50 EY hmm, client runs on SSD
20:50 osse tmpfs !
20:51 EY _ikke_, having said that, I do `git rebase -i origin/blah` usually. Could it be the remote access?
20:51 _ikke_ That does no remote access
20:51 mingrammer joined #git
20:52 dmto joined #git
20:52 jeffreylevesque joined #git
20:52 EY _ikke_, ok. apart from faster disk?
20:52 MattMaker joined #git
20:53 _nwkj86 joined #git
20:53 napping setting a sparse checkout to avoid avoid too many files worked out okay
20:53 _ikke_ EY: what part is slow?
20:53 EY Between each rebase command.
20:56 Rodya_ joined #git
20:56 EY Between each command of the rebase, that is.
20:57 _nwkj86 joined #git
20:58 Celelibi joined #git
20:59 timewalker joined #git
20:59 Jellyg00se joined #git
21:00 LeBlaaanc joined #git
21:00 Manuel15 joined #git
21:00 Manuel15 joined #git
21:02 _ikke_ large repo?
21:04 osse mkdir rebasedir; sudo sudo mount -t tmpfs -o size=100M tmpfs rebasedir; export GIT_DIR=$PWD/.git GIT_WORK_TREE=rebasedir; git checkout -f; git rebase -i; unset GIT_DIR GIT_WORK_TREE; git checkout -f;
21:04 alfism joined #git
21:05 osse Unless it's writing stuff in .git which is the slow part
21:06 osse napping: I suppose it would be possible to only use --index-filter but it's complicated
21:08 EY _ikke_, not even that large.
21:08 daniel-wtd joined #git
21:08 ki0 joined #git
21:08 _ikke_ How slow?
21:08 EY _ikke_, 1489 objects, 2637 kilobytes (if this measure is relevant)
21:09 napping wow, it took a long time to do a git gc that threw away almost all objects
21:09 crayon joined #git
21:10 EY I'd say up to 10 seconds from typing the command at the prompt to the last step for three commits.
21:10 osse EY: Windows?
21:10 EY osse, yes. so your tmpfs suggestion cannot be used.
21:10 osse I too notice that git rebase is slooow on Windows
21:11 dsdeiz joined #git
21:11 dsdeiz joined #git
21:11 osse I think it's mostly caused by git rebase being a shell script
21:11 osse Even in a virtual machine on the same computer you'd notice it's MUCH faster.
21:11 dsantiago joined #git
21:12 EY osse, do you think running it off cygwin would bring any improvement?
21:12 osse Afaik work is being done in preparation of a rewrite in C
21:12 sssilver_ joined #git
21:12 osse EY: no. slower if anything. That's the problem. Git for Windows relies on some sort of emulation of fork()+exec()
21:13 EY ah!
21:13 osse I notice that Git installed in MSYS2 is significantly slower than Git for Windows which is built with Mingw.
21:13 EY osse, are there alternatives? something that uses other git libraries? or do they all ultimately call the native C functions?
21:14 z3t0 joined #git
21:14 _ikke_ EY: alternatives to what?
21:15 osse As far as I understand it (Correct me if I'm wrong anyone). You can more or less compile code meant for Unix on windows if you link with msys2.dll or cygwin.dll. Or you can make changes to avoid that linkage so you only need to compile with a mingw toolchain, which will give a faster binary but is more hassle to get done.
21:15 z3t0 hi all, is there some place where i can start writing a commit message even before adding any files to git or trying to make a commit, that way i can keep track of what i am doing by writing it down and then just commit at the end without having to think back to what it was that i actually did
21:15 BlaXpirit osse, about right
21:15 BlaXpirit not sure about msys though
21:16 d0nn1e joined #git
21:16 osse z3t0: sure, you can copy stuff from a temp file when you're writing the final commit message
21:16 osse z3t0: but it's easier to use commit --amend
21:16 z3t0 osse: oh good idea, il just use commit --amend
21:17 Learath2 care about accidentally pushing your half commit tho
21:17 cjohnson z3t0: what I do is rebase before pushing. I'll make several small commits and then go back when I'm ready to push and squash any commits that don't really need to be on their own
21:17 rominronin joined #git
21:18 cjohnson git rebase -i
21:18 z3t0 cjohnson: ok i see, althogh i think it would be more efficient to just create a single commit in the first place
21:18 osse EY: the C stuff is fast enough. The fork-heavy shell script is the main problem
21:18 cjohnson yeah I just prefer to have as much history as possible when something goes wrong
21:18 cjohnson multiple commits makes that nice
21:18 cjohnson then I rebase so I don't pllute commit log
21:20 EY osse, ah, i see.
21:21 z3t0 cjohnson: makese sense
21:21 osse I'm not sure if the shell that comes with Git for Windows relies on msys2.dll or not
21:21 str joined #git
21:22 Learath2 osse, it does
21:22 osse Ah
21:22 mxmasster joined #git
21:22 mxmasster hello
21:22 osse EY: in that case rewrite the bash source ! Easy!
21:22 z3t0 mxmasster: hey
21:23 MattMaker joined #git
21:23 Learath2 yeah just 633 lines in git-rebase.sh
21:23 mxmasster rando question, in CVS there were keywords that could be included in a file that CVS would expand (log, author, date, etc...)
21:23 mxmasster is there anything like that with GIT?
21:23 osse Learath2: I was thinking of modifying bash itself :P
21:23 EY osse, rewrite it in powershell? :)
21:24 osse EY: if you're up for it!
21:24 osse I'd surely welcome it
21:24 EY :)
21:24 Learath2 mxmasster, think gitbook had a section about that
21:25 Learath2 yep chapter 8.2
21:25 mxmasster thanks!
21:27 sweatsuit joined #git
21:28 durham_ joined #git
21:29 moei joined #git
21:29 languitar joined #git
21:30 nwkj86 joined #git
21:30 durham__ joined #git
21:32 MattMaker joined #git
21:33 HoierM joined #git
21:33 danslo_ joined #git
21:37 jnewt GodGinrai, do i need to modify eol when setting core.autocrlf to input?
21:38 MattMaker joined #git
21:44 sdothum joined #git
21:46 l4v2 joined #git
21:46 specing git must hate me
21:46 mingrammer joined #git
21:46 specing I create a stash and immidiately pop it
21:46 specing error: Your local changes to the following files would be overwritten by merge:
21:46 specing >_>
21:46 rgrinberg joined #git
21:47 disi joined #git
21:47 ahr3n joined #git
21:48 MattMaker joined #git
21:48 fakenerd joined #git
21:52 mostlybadfly joined #git
21:54 dreiss joined #git
21:54 jason237 joined #git
21:56 raijin joined #git
21:57 geoid joined #git
21:57 terminal_echo joined #git
21:58 xaviergmail joined #git
21:58 MattMaker joined #git
22:00 monoprotic joined #git
22:00 monoprotic joined #git
22:01 aidalgol joined #git
22:01 andlabs joined #git
22:02 simulacr joined #git
22:03 simulacr is there a simple way to unpack pack files?
22:04 osse simulacr: no, but there is a complicated way. cd /tmp/unpack; git init; git unpack-objects < path/to/packfile.pack
22:04 red_ joined #git
22:10 hahuang61 joined #git
22:12 nettoweb joined #git
22:12 alansaul joined #git
22:14 geoid_ joined #git
22:16 simulacr seems something doing wrong, thought plain src in pack files, what's the right way to get plain src from copy of master which has pack files instead src?
22:19 rominronin joined #git
22:19 hweaving joined #git
22:20 hweaving Is there any way for a git repo to track checkouts, for the purpose of using a "central" repo?  As in, "Let's make sure no one has outstanding commits they need to do."
22:20 osse hweaving: no
22:21 osse at least not with vanilla git. maybe some hosting solutions implement this
22:21 madewokherd joined #git
22:22 hweaving osse:  Thanks, I didn't think so but wanted to check.
22:22 rkazak joined #git
22:23 osse hweaving: also it would be easily fooled:  git clone https://git.company.com/mainrepo.git; rm -rf mainrepo
22:23 osse now the central repo will be confused forever
22:23 rubyonrailed joined #git
22:24 HoierM joined #git
22:24 jason237 joined #git
22:24 hweaving osse: Fair enough :)
22:24 durham joined #git
22:25 pikajude joined #git
22:25 pikajude joined #git
22:26 hweaving left #git
22:30 MattMaker joined #git
22:32 daniel-wtd joined #git
22:33 sangy joined #git
22:34 spudowiar joined #git
22:34 danslo joined #git
22:37 terminal_echo joined #git
22:38 osse omg
22:38 osse was that the real Hugo Weaving ?
22:40 Bradipo joined #git
22:40 Bradipo Hello... I need to revert a merge, but that needs a -m.  How do I determine what the valid numbers are?
22:40 motivic joined #git
22:40 Bradipo I thought perhaps -m would take a revision hash, but apparently not.
22:41 osse Bradipo: it's the parent number. 1 2 3
22:41 Bradipo But how do I see the parents?
22:41 Bradipo I'm looking at gitk and it doesn't show any parent numbers as far as I can tell.
22:41 Bradipo It shows a parent commit hash.
22:41 osse Bradipo: they are literally one, two, three etc.
22:42 byteflame joined #git
22:42 Bradipo Right, but which is which?  In a merge, there are more than 1 parent, right?
22:42 osse yes
22:43 osse if you are on master and do 'git merge bugfix', then after merge the first parent is old master and the second parent is the bugfix branch
22:43 Bradipo Oh, sorry, this is an auto-merge...
22:43 Bradipo Not the result of git merge.
22:43 osse ie. the "mainline" is the first parent, the branch that was *merged in* is  the second
22:43 Bradipo It's the result of git pull.
22:43 osse same thing
22:43 Bradipo Ok.
22:43 osse whatever you pulled in is the second parent
22:43 Bradipo I see.
22:44 Bradipo So my repository will always be the first parent?
22:44 Bradipo When I pull.
22:44 osse yes
22:44 osse unless it's a fast forward and there is no merge at all, of course
22:44 Bradipo So 1 is my parent line, 2 is the merged in line?
22:44 clemf joined #git
22:44 osse if X is the merge commit:  X^1 is the first parent; X^2 is the second parent
22:44 Bradipo Don't think it was a fast forward, I see: Merge branch 'master' of ...
22:45 osse git show X^1 etc.
22:45 Bradipo Right, I've used X^n before.
22:45 osse n = parent number
22:45 Bradipo Thank you, very helpful
22:45 Bradipo In fact, I've always wondered how to do ``git show'' but never been able to find it.
22:46 osse (in X~n, n is the ancestor number. X~n is X's first parent's first parent's first parent's...)
22:46 pijiu joined #git
22:46 Bradipo Right, thank you.  Let me see if I can revert this now...
22:47 osse reverting a pull sounds a bit like unnecessary noise. why not reset?
22:47 osse or is it an old pull ?
22:48 Bradipo It's a fairly new pull.
22:48 Bradipo To be honest, I'm not really sure why I would prefer revert over reset.
22:48 Bradipo revert seems more VCSish.
22:48 Bradipo So 1 is my parent, 2 is the merged in parent?
22:48 osse reset = rewriting history. so if this pull has been pushed and whatnot it gets messy
22:49 Bradipo Oh, well I haven't pushed anything, but they did.
22:49 fakenerd joined #git
22:50 Bradipo Hmm, maybe I don't want revert.
22:50 pijiu joined #git
22:52 pijiu joined #git
22:54 pijiu joined #git
22:55 pijiu joined #git
22:57 Mixologic joined #git
23:00 osse Bradipo: also, pull = fetch + merge
23:01 Bradipo Yeah, hence the auto-merge.
23:02 Bradipo But thanks, I didn't know about fetch.
23:02 Bradipo By the way, revert seems to have worked (at least from my perspective).
23:02 osse but that's my point. it's not an "auto-merge". it's a merge
23:02 Bradipo Oh, I see what you mean.
23:03 Bradipo I call it auto-merge because I did not type ``git merge''
23:03 Bradipo It's part of git's strategy to avoid conflicts I imagine.
23:05 dsantiago joined #git
23:09 DrAwesome joined #git
23:10 dreiss joined #git
23:11 hazelux joined #git
23:13 monoprotic joined #git
23:14 clemf joined #git
23:19 Gsham joined #git
23:19 rominronin joined #git
23:22 Guest9 joined #git
23:29 spudowiar joined #git
23:31 cagedwisdom joined #git
23:32 cagedwisdom left #git
23:34 mingrammer joined #git
23:34 inflames joined #git
23:36 beyertech joined #git
23:36 linuxmint joined #git
23:37 dsdeiz joined #git
23:37 dsdeiz joined #git
23:40 hahuang61 joined #git
23:41 mar77i_ joined #git
23:46 dreiss joined #git
23:46 skylite_ joined #git
23:48 Mixologic joined #git
23:49 pikajude joined #git
23:49 pikajude joined #git
23:50 fakenerd joined #git
23:50 Finnoloid joined #git
23:53 WhiskerBiscuit joined #git
23:53 raijin joined #git
23:55 energizer joined #git
23:55 energizer I'm trying to clone a repo from a server, and it keeps hanging.
23:57 energizer at this point every timeReceiving objects:  21% (230/1089), 10.81 MiB | 2.61 MiB/s
23:58 energizer But I can successfully SSH to the server
23:59 str joined #git
23:59 stuh84 joined #git

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