Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-11-22

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

All times shown according to UTC.

Time Nick Message
00:00 ayjay_t joined #git
00:00 Darren_ joined #git
00:00 heftig joined #git
00:02 austin987 joined #git
00:02 peeps[lappy] i did a git clone, then tried git checkout branch_name (which exists in remote), and it says it created a new branch.  is it required to specify origin?  i don't remember having to do this before
00:03 m1dnight_ joined #git
00:03 mikecmpbll joined #git
00:04 venmx joined #git
00:05 clemf joined #git
00:06 m1dnight_ joined #git
00:06 clemf joined #git
00:06 peeps[lappy] nevermind, i guess i misunderstood the output messages
00:09 thiago joined #git
00:09 clemf joined #git
00:12 nfk joined #git
00:16 fmcgeough joined #git
00:17 oaaao joined #git
00:19 mischat joined #git
00:24 Doginal joined #git
00:27 courrier joined #git
00:27 snowkidind joined #git
00:32 sammyo joined #git
00:34 nanga joined #git
00:35 jimi_ joined #git
00:39 madewokherd joined #git
00:46 aidalgol joined #git
00:46 nicksloan joined #git
00:48 railssmith joined #git
00:49 cdg joined #git
00:52 Lyqyd joined #git
00:58 Learath2 joined #git
00:59 anonymuse joined #git
01:00 nicksloan joined #git
01:00 lolmac joined #git
01:01 acetakwas joined #git
01:05 mikecmpbll joined #git
01:10 Sasazuka joined #git
01:12 kyan_ joined #git
01:14 Nilesh_ joined #git
01:20 mischat joined #git
01:20 jason237 joined #git
01:22 qws-user-1228 joined #git
01:27 Olipro joined #git
01:29 t-mart_ joined #git
01:32 Peetz0r joined #git
01:32 Noldorin joined #git
01:33 anonymuse joined #git
01:33 Noldorin joined #git
01:34 pks joined #git
01:36 grift joined #git
01:37 Darren_ joined #git
01:38 manfredn joined #git
01:41 publio joined #git
01:42 vdv joined #git
01:42 cdg joined #git
01:43 Resident_ joined #git
01:45 PresidentBiscuit joined #git
01:48 Pyrogerg joined #git
01:49 t-mart joined #git
01:50 pwnz0r joined #git
01:55 jmpp_ joined #git
01:55 jmpp_ joined #git
01:59 Darren_ joined #git
01:59 lolmac joined #git
02:01 jpX joined #git
02:03 waterCreature joined #git
02:07 VampiricPadraig joined #git
02:08 athaller joined #git
02:08 Hobbyboy joined #git
02:09 d^sh joined #git
02:11 sgtbigman joined #git
02:13 finalbeta joined #git
02:14 athaller joined #git
02:14 VampiricPadraig joined #git
02:14 VampiricPadraig joined #git
02:15 heftig joined #git
02:15 barteks2x joined #git
02:15 Blkt joined #git
02:15 Olipro joined #git
02:20 VampiricPadraig joined #git
02:20 athaller joined #git
02:20 pflanze joined #git
02:21 leprechau joined #git
02:21 mischat joined #git
02:21 Hobbyboy joined #git
02:21 Blkt joined #git
02:21 heftig joined #git
02:22 trake joined #git
02:29 pwnz0r joined #git
02:30 VampiricPadraig joined #git
02:30 VampiricPadraig joined #git
02:30 athaller joined #git
02:33 anonymuse joined #git
02:33 nettoweb joined #git
02:33 Blkt joined #git
02:33 heftig joined #git
02:34 venmx joined #git
02:35 pks joined #git
02:38 cdg joined #git
02:39 Gsham joined #git
02:41 mycyber joined #git
02:48 ilbot3 joined #git
02:48 Topic for #git is now Welcome to #git, the place for git help and serious business | Public logs at http://jk.gs/glog | First visit? Read: http://jk.gs/git | Current stable version: 2.10.2 | Getting "cannot send to channel"? /msg gitinfo .voice | The git-jokes project has been suspended for lack of contribution. Help revive it! https://madeitwor.se/git-jokes
02:49 thiago joined #git
02:54 steelsw0rd joined #git
02:57 PresidentBiscuit joined #git
02:57 dodobrain joined #git
02:57 Vampire0_ joined #git
02:58 ixjlyons joined #git
03:03 aavrug joined #git
03:03 gmn8719 joined #git
03:08 qws-user-1228 joined #git
03:09 sunri5e joined #git
03:09 dreiss joined #git
03:11 Mobutils_ joined #git
03:13 ixjlyons joined #git
03:14 PresidentBiscuit joined #git
03:16 Goplat joined #git
03:16 JanC_ joined #git
03:19 thierryp joined #git
03:20 hexagoxel joined #git
03:21 mischat joined #git
03:23 chachasmooth joined #git
03:26 yrzr joined #git
03:29 byte512 joined #git
03:31 dreiss joined #git
03:33 pwnz0r joined #git
03:34 trake joined #git
03:34 rahtgaz joined #git
03:34 dpower joined #git
03:36 Anthaas joined #git
03:37 KnownSyntax joined #git
03:37 KnownSyntax joined #git
03:37 nebg joined #git
03:38 Bilge joined #git
03:38 midgetspy joined #git
03:38 tvl joined #git
03:38 ilmari joined #git
03:38 c0ded joined #git
03:39 niax joined #git
03:40 devster31 joined #git
03:40 tgy joined #git
03:41 bremner joined #git
03:41 bremner joined #git
03:44 vrishab joined #git
03:45 a_thakur joined #git
03:45 clemf joined #git
03:45 A5101 joined #git
03:46 jstimm joined #git
03:46 ilmari joined #git
03:47 justanotheruser joined #git
03:47 sangy joined #git
03:48 markelite joined #git
03:49 JSharpe joined #git
03:51 dsdeiz joined #git
03:51 dsdeiz joined #git
03:52 svm_invictvs joined #git
03:56 Shakeel_ joined #git
03:56 dsdeiz joined #git
04:05 mycyber joined #git
04:05 venmx joined #git
04:05 anonymuse joined #git
04:06 svara1 joined #git
04:06 Love4Boobies joined #git
04:07 Love4Boobies hi. can submodules be imported into the project's root directory? if yes, how are things like .gitignore treated?
04:09 cqi joined #git
04:10 sbulage joined #git
04:13 jaguarmagenta joined #git
04:14 Vortex35 joined #git
04:23 mischat joined #git
04:26 sdelic joined #git
04:28 t-mart joined #git
04:29 kadoban If I understand the question correctly, no.
04:29 dreiss joined #git
04:30 bernardio joined #git
04:40 robotroll joined #git
04:42 ayogi joined #git
04:42 shams joined #git
04:43 pks joined #git
04:45 Love4Boobies kadoban: that's unfortunate. i mean basically that i don't want to import submodule in proj/submodule. i want to import it in proj/, which is where .git is.
04:45 Love4Boobies it sounds silly, perhaps, but that's because the submodule is basically my build system so it makes sense to place it there
04:46 Love4Boobies can submodules at least be imported in the *parent* directory?
04:46 Love4Boobies rather than a subdirectory, i mean
04:46 Love4Boobies this i expect is possible
04:47 t-mart_ joined #git
04:47 kadoban Love4Boobies: Doubt that would make sense
04:47 kadoban Just do it in a subdirectory, why would it not make sense for your build system to live there?
04:48 Love4Boobies so people don't go searching for the makefile
04:48 Love4Boobies i guess i can write it in README.md or put a .sh file that executes make
04:49 Love4Boobies why wouldn't it make sense, anyway?
04:49 Love4Boobies or you mean from git's perspective?
04:49 Love4Boobies actually, you are right
04:49 Love4Boobies i don't want someone cloning the repo and then having the makefile written in *their* parent directory
04:50 Love4Boobies because they have the dir structure unprepared for this
04:50 Love4Boobies so yes, you are right
04:50 subhojit777 joined #git
04:50 bhuvanaurora joined #git
04:52 fstd_ joined #git
04:53 tmokros joined #git
05:00 pks joined #git
05:04 mycyber joined #git
05:06 Aces_Charles joined #git
05:09 ixjlyons joined #git
05:10 Peng_ joined #git
05:14 Cabanossi joined #git
05:16 akushner joined #git
05:18 dreiss joined #git
05:21 Andrew_K joined #git
05:22 netj joined #git
05:22 t-mart joined #git
05:23 sdelic joined #git
05:30 tjone270 joined #git
05:34 Raging_Hog joined #git
05:36 venmx joined #git
05:41 sbulage joined #git
05:42 lindenle joined #git
05:49 raspado joined #git
05:53 msonntag joined #git
05:56 Shakeel_ joined #git
05:56 pwnz0r joined #git
06:00 clemf joined #git
06:01 aavrug joined #git
06:02 sdelic joined #git
06:06 rscata joined #git
06:07 ToBeCloud joined #git
06:12 Qria joined #git
06:14 Peng_ joined #git
06:16 rominronin joined #git
06:17 bhuvanaurora joined #git
06:17 done joined #git
06:19 Vangelis joined #git
06:25 mischat joined #git
06:26 cqi joined #git
06:27 phanimahesh joined #git
06:28 pks joined #git
06:29 mycyber joined #git
06:32 meauses joined #git
06:32 clemf joined #git
06:33 dsantiago joined #git
06:34 Vangelis left #git
06:39 acetakwas joined #git
06:40 Rickmasta joined #git
06:43 Kulrak joined #git
06:43 Kulrak joined #git
06:43 Raging_Hog joined #git
06:45 LucaTM joined #git
06:47 kuldeep joined #git
06:47 bhuvanaurora joined #git
06:48 ThomasLocke joined #git
06:48 ThomasLocke joined #git
06:49 kuldeep is it a good practise to: my project require some (submodule) external library that need some modification. periodically i rebase the modification to upstream and force push to the submodule repo branch the new changes.
06:51 kuldeep others who checkout might have to force update their submodule. (i do understand that in this process, previous submodule commit of the branch are lost for forever)
06:52 PioneerAxon joined #git
06:52 kuldeep what would be the recommended method? manage my own branch, and merge instead of rebase from upstream.
06:53 dan2k3k4 joined #git
06:53 kuldeep "periodically" = 6month or so
06:56 aidalgol joined #git
06:57 dodobrain joined #git
06:58 kuldeep i do have the code public in case someone want to see the real project in effect.
06:58 thebope joined #git
07:00 jknetl joined #git
07:01 venmx joined #git
07:03 rendar joined #git
07:05 JeroenT joined #git
07:09 jfr joined #git
07:09 troulouliou_div2 joined #git
07:10 robattila256 joined #git
07:10 pks joined #git
07:13 xall joined #git
07:13 netj joined #git
07:15 thierryp joined #git
07:17 hhee joined #git
07:17 zeroed joined #git
07:18 mushan joined #git
07:18 qt-x joined #git
07:21 _ng joined #git
07:24 freimatz joined #git
07:24 chele joined #git
07:25 shmuli joined #git
07:25 hussam joined #git
07:26 mischat joined #git
07:27 pks joined #git
07:28 JeroenT joined #git
07:30 seishun joined #git
07:31 Silenced joined #git
07:31 Guest39285 joined #git
07:37 Alienpruts joined #git
07:37 leehambley joined #git
07:37 Anupkumar joined #git
07:37 rominronin joined #git
07:38 Macaveli joined #git
07:38 JeroenT joined #git
07:38 jnavila joined #git
07:39 lordjancso joined #git
07:41 Macaveli_ joined #git
07:42 Akimb joined #git
07:44 Silenced joined #git
07:45 Sceorem joined #git
07:46 GT4066 joined #git
07:53 fees joined #git
07:56 pks joined #git
07:58 pwnz0r joined #git
07:58 donmichelangelo joined #git
07:58 bhuvanaurora joined #git
07:59 JSharpe joined #git
08:00 Alienpruts joined #git
08:02 a_thakur joined #git
08:03 adac joined #git
08:04 danslo joined #git
08:06 masuberu joined #git
08:07 PtxDK_ joined #git
08:09 ahmedelgabri joined #git
08:11 xall joined #git
08:11 cagedwisdom joined #git
08:14 Mir4g3 joined #git
08:15 djb-irc joined #git
08:17 TomyWork joined #git
08:17 xall joined #git
08:19 JeroenT joined #git
08:20 _ng joined #git
08:20 manuels joined #git
08:22 overlord_tm joined #git
08:23 encod3 joined #git
08:24 CtrlC how can I pass the private key to git when trying to clone using ssh?
08:25 kadoban CtrlC: Use your ssh agent
08:25 danslo joined #git
08:25 CtrlC kadoban, the keys are already added if that's what you mean.
08:25 hhee joined #git
08:25 CtrlC or maybe you mean doing something localy?
08:27 kadoban CtrlC: Locally, ya. Usually you do 'ssh-add' to unlock your private key locally, and then the ssh agent will use it when needed. You also usually set up ssh aliases in ~/.ssh/config which configure which hosts get which key.
08:27 kadoban Anyway, git just uses whatever ssh stuff your os has, there's not really anything git-specific.
08:28 CtrlC I wanted something like "-i id-rsa"
08:29 kadoban Not a thing you should use, I don't think it exists anyway. SSH aliases and your ssh agent are where it's at.
08:30 yqt joined #git
08:30 rominronin joined #git
08:31 tvw joined #git
08:31 Repox joined #git
08:32 venmx joined #git
08:32 rchavik joined #git
08:35 ngperfex joined #git
08:35 derk0pf joined #git
08:36 Silenced joined #git
08:38 Olipro joined #git
08:41 wrouesnel1 joined #git
08:45 grawity GIT_SSH_COMMAND= could sort of be used for that
08:45 grawity but just set up .ssh/config
08:46 deepy Is there anything I can do to get git to be smarter when it comes to merges?
08:46 deepy Like with java I don't really care if two different import statements have appeared, if they're different then that is likely fine(tm)
08:47 PCatinean joined #git
08:49 adac joined #git
08:50 acetakwas joined #git
08:52 acetakwas joined #git
08:53 mda1 joined #git
08:54 denisMone joined #git
08:54 kedare joined #git
08:54 d0nn1e joined #git
08:55 brillenfux joined #git
08:55 Macaveli joined #git
08:55 bhuvanaurora joined #git
08:59 kurkale6ka joined #git
09:00 sbulage joined #git
09:01 brk_ joined #git
09:02 raijin joined #git
09:02 ToBeCloud joined #git
09:02 rominronin joined #git
09:03 Sasazuka joined #git
09:04 meauses_ joined #git
09:04 GavinMagnus joined #git
09:05 Macaveli joined #git
09:05 Guest39285 joined #git
09:06 Macaveli joined #git
09:07 Macaveli joined #git
09:08 cbreak deepy: you can make git dumber, but not smarter, without writing your own merge driver
09:09 halbomin joined #git
09:09 coasterz joined #git
09:09 cbreak deepy: see man gitattributes
09:09 gitinfo deepy: the gitattributes manpage is available at http://jk.gs/gitattributes.html
09:09 cbreak in the merge driver section
09:12 ngperfex joined #git
09:13 jfr_ joined #git
09:13 marcogmonteiro joined #git
09:13 Macaveli joined #git
09:14 acetakwas joined #git
09:16 Vinnie_win joined #git
09:16 apotry joined #git
09:17 jrm joined #git
09:18 Guest39285 joined #git
09:19 dreiss joined #git
09:19 JeroenT joined #git
09:20 zeroed joined #git
09:21 chll_ joined #git
09:24 raspado joined #git
09:24 ayjay_t joined #git
09:26 deepy I should probably write a merge driver
09:27 re1 joined #git
09:27 Charliechin joined #git
09:28 kadoban It's likely going to be fairly non-trivial and limited in usefulness, unless your workflow is kinda weird.
09:30 deepy Dunno, if I can just have it ignore some of the current silliness I can cut down on the effort
09:30 deepy (that is, add usefulness, not make it trivial to write)
09:31 KnownSyntax joined #git
09:31 KnownSyntax joined #git
09:32 venmx joined #git
09:33 Silenced joined #git
09:33 mkoske joined #git
09:33 nettoweb joined #git
09:35 mischat joined #git
09:36 Miika-- joined #git
09:36 Miika-- Hello, if I have repository with a lots of commits in it, how I could get all versions of one specific file? I probably need to list all commits where the file was modified and then read the file at that commit, right?
09:36 mkoske left #git
09:38 ProbabilityMoon joined #git
09:38 raspado joined #git
09:39 Akimbas joined #git
09:40 JeroenT_ joined #git
09:40 osse Miika--: yes
09:41 selckin git log file; git show sha -- file
09:42 osse Miika--: git rev-list -- file | while read hash; do git show $hash:file > /tmp/file_$hash; done
09:42 nd joined #git
09:42 CrazyBonz joined #git
09:44 subhojit777 joined #git
09:47 GavinMagnus joined #git
09:47 mablae joined #git
09:48 JeroenT joined #git
09:48 xall joined #git
09:50 venmx joined #git
09:52 Miika-- osse: thanks
09:52 Miika-- I'll try that
09:53 leftriver87 joined #git
09:54 leftriver87 May I ask for help?
09:55 Endarked joined #git
09:55 leftriver87 I have a commit in GitHub that I want to pull, but it is (no longer) part of any branch
09:56 leftriver87 I want to do something like 'git fetch --commit f9e6f969c0a82e75ab6facaf81b7bc6ce639a2ce'
09:56 Macaveli joined #git
09:56 rominronin joined #git
09:57 osse leftriver87: no can do
09:57 rominron_ joined #git
09:58 Macaveli joined #git
09:58 Charliechin Hi guys! I am a totally noob using git. I am doing a tutorial about it and still dont get the concept of push and pull request
09:58 tobiasvl leftriver87: that's not possible, but you can view it in GitHub's web UI at least: https://help.github.com/articles/commit-ex​ists-on-github-but-not-in-my-local-clone/
09:58 tobiasvl Charliechin: in context of what? GitHub?
09:58 phanimahesh joined #git
09:59 Charliechin tbiasvl yes
09:59 Charliechin let's say, I have downloaded a repostory, i made changes and i want to uploaded to the remote repository
09:59 leftriver87 What? It's not possible?
09:59 nowhere_man joined #git
09:59 Charliechin so, first I commit and aftwerwards pull request?
09:59 leftriver87 I thought, maybe I can somehow create a remote tag and then fetch?
09:59 leftriver87 Or create a remote branch?
10:00 tobiasvl Charliechin: this is something #github can probably help better with, but you can read their documentation: https://help.github.com/art​icles/about-pull-requests/
10:00 tobiasvl Charliechin: but first you commit, then you push to your fork, then you make a pull request
10:00 JeroenT joined #git
10:00 Charliechin Understood, I will have a good look at that link
10:02 tobiasvl leftriver87: you can't branch or make tags directly in the web UI of GitHub afaik… but I'm no GitHub expert, maybe #github can help. I see you can make a "release" from a "recent" commit SHA, whatever that entails, but I'm not sure if that also creates a tag
10:03 ronskisika joined #git
10:03 Macaveli joined #git
10:04 Silex^^ joined #git
10:04 tobiasvl hm, yes, a release does create a tag. not sure if you can make a release from an arbitrary SHA though.
10:04 NimeshNeema joined #git
10:05 Serpent7776 joined #git
10:07 djb-irc joined #git
10:09 rnsanchez joined #git
10:11 cdown joined #git
10:11 InfoTest joined #git
10:12 clauds joined #git
10:12 TheOrangeOne joined #git
10:13 leftriver87 tobiasv1: I tried to manually create a tag and then force push that, but it wouldn't let me
10:14 tobiasvl hmmm? push? I thought you didn't have the commit locally?
10:15 WoodLeg joined #git
10:18 Nilesh_ joined #git
10:21 acetakwas joined #git
10:23 zulutango joined #git
10:23 zacts joined #git
10:23 _ng joined #git
10:25 Phylock joined #git
10:25 bhuvanaurora joined #git
10:28 _ng joined #git
10:31 WayToDoor joined #git
10:32 xaa joined #git
10:35 blunaxela joined #git
10:35 clmsy joined #git
10:35 musicmatze Good morning everyone. If I want to create a custom tool for git that needs to create refs, where do I store these refs? Is it a good idea to put them in .git/refs/<toolname>/... ?
10:36 xocket joined #git
10:37 musicmatze or is that a bad idea? I see that git-annex creates .git/refs/heads/synced/master for example,... that's why I wonder whether I should store refs in one of the existing .git/refs/ sub folders or whether I'm allowed to create a new subfolder for my specific tool...
10:38 mycyber joined #git
10:38 xall joined #git
10:39 netj joined #git
10:39 cagedwisdom joined #git
10:43 grawity in the case of git-annex, synced/master more or less works as an actual branch that gets pushed to and merged from
10:43 venmx joined #git
10:43 grawity (just never directly checked out)
10:43 grawity therefore it's under refs/heads
10:43 leftriver87 left #git
10:44 grawity if you merely need refs that *aren't* real branches, put them wherever you like
10:44 grawity under one condition:
10:44 grawity do not assume that refs are actually stored in .git/refs/
10:44 grawity use git update-ref, git symbolic-ref, etc.
10:47 elect joined #git
10:49 _ng joined #git
10:52 subhojit777 joined #git
10:53 tvw joined #git
10:56 qws-user-1228 joined #git
10:58 mischat joined #git
10:58 acetakwas joined #git
11:04 King_Hual joined #git
11:07 raijin joined #git
11:09 jbitdrop joined #git
11:10 mikecmpbll joined #git
11:11 aax joined #git
11:11 sbulage left #git
11:11 raijin joined #git
11:12 ocbtec joined #git
11:14 LEE__ joined #git
11:14 Guest39285 joined #git
11:15 LEE__ left #git
11:16 ShaunGreiner[m] joined #git
11:18 acetakwas joined #git
11:19 xaa__ joined #git
11:19 musicmatze grawity: yep, thanks, that sounds good!
11:21 musicmatze grawity: just to be sure "put them wherever you like" means "wherever you like in .git/refs/", right?
11:21 rominronin joined #git
11:21 venmx joined #git
11:22 thebope joined #git
11:22 grawity wherever you like under the refs/* namespace
11:23 grawity ah, another thing
11:23 grawity only refs with 3+ components can be pushed/fetched
11:23 grawity e.g. git will refuse a push to refs/foo
11:23 grawity so you almost always need a subfolder
11:23 CrazyBonz joined #git
11:23 Miika-- osse: doesn't that rev-list command require hashes of starting commit and ending commits? Like git rev-list start..stop -- file
11:23 musicmatze grawity: perfect, so refs/<toolname>/<refname> is perfect
11:24 grawity sounds okay
11:24 DieguezZ joined #git
11:24 grawity Miika--: it accepts anything that `git log` accepts
11:25 Snugglebash joined #git
11:25 Miika-- grawity: guestion was that how do I get all the revisions of some file and osse gave me oneliner for that, but it didn't work directly...
11:26 Miika-- so I just did git log for that file and took first and last sha and used those as start and end points
11:27 ahmed_elgabri joined #git
11:28 chachasmooth joined #git
11:28 ruby_on_tails joined #git
11:28 jeffreylevesque joined #git
11:28 xall i forked a repo, pushed changes, amended locally, and want to push again, but now the branches are diverged so i have a rebase conflict. how can i easily resolve using my local changes?
11:30 ruby_on_tails joined #git
11:32 boz_v1 joined #git
11:33 JeroenT joined #git
11:34 xall i guess git push --force works
11:34 ^7heo use the force, git.
11:35 tobiasvl xall: yep, that works, but be aware of !rewrite
11:35 gitinfo xall: 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
11:35 ferr joined #git
11:35 leeN joined #git
11:37 xall tobiasvl: right. how does it work with pull request revisions? if you have several revisions as part of pr feedback, does the maintainer usually squash them to cleanup at the end?
11:38 denisMone joined #git
11:38 tobiasvl xall: the pull request will be changed to reflect your force push
11:38 xall i didn't actually make the pr yet
11:39 tobiasvl a pull request is just one branch against another branch, if the branch is changed, the PR will reflect that
11:39 tobiasvl OK
11:39 xall i'm just curious in general
11:40 nowhere_man joined #git
11:40 MaBunny joined #git
11:41 MaBunny guys is there a git command tht adds an issue to my remote repo in github?
11:41 tobiasvl MaBunny: check out !hub
11:41 gitinfo MaBunny: Hub is a wrapper around git which provides specific github features to the git cli. See https://github.com/github/hub for details.
11:41 MaBunny i want to do it from the terminal in future so tht ...
11:41 MaBunny ok
11:42 MaBunny is it available for debian?
11:42 rominronin joined #git
11:42 tobiasvl not sure it can actually open an issue though…
11:42 MaBunny hmm
11:43 tobiasvl it can browse issues, open PRs, do lots of stuff. but unsure about opening issues. check it out
11:43 tobiasvl yes you can build it yourself or download a binary
11:43 MaBunny ok cool
11:43 MaBunny thnx for the tip
11:44 zarubin joined #git
11:45 Chaser joined #git
11:45 Love4Boobies joined #git
11:46 lindenle joined #git
11:47 Love4Boobies hi. where does git track submodule-related information? it can't just be .gitmodules because that file doesn't say which revision of the submodule is being used.
11:47 Love4Boobies oh, nvm
11:48 Love4Boobies i was sure i looked inside .git and didn't find it but i looked again now and it seems i just missed it
11:48 _ikke_ In tree objects for the record as a gitlink
11:49 MaBunny left #git
11:49 lss8 joined #git
11:49 lss8 joined #git
11:50 _ng joined #git
11:51 mischat joined #git
11:56 mycyber joined #git
11:56 mischat joined #git
11:57 paul424 joined #git
11:58 mischat_ joined #git
11:58 mikecmpbll joined #git
11:59 mischat_ joined #git
12:06 j`ey joined #git
12:07 osse Miika--: yes, my bad. You can give simply HEAD, or --all
12:07 cdown joined #git
12:07 Khisanth joined #git
12:07 j`ey if I do git log, and see a commit X, and then I scroll down, and see commit Y
12:07 j`ey if I checkout X, should that always contain Y?
12:07 grawity not necessarily
12:08 grawity `git log` by default shows flattened history, and a branch without Y might get listed before a branch which introduced Y
12:08 j`ey bah
12:08 grawity try `git log --oneline --graph`
12:11 Snugglebash joined #git
12:12 Peng_ joined #git
12:12 j`ey so even commit date isn't important?
12:12 j`ey cos Y (a merge) has a commit date earlier than X
12:13 Guest39285 joined #git
12:13 qws-user-1228 joined #git
12:13 raspado joined #git
12:13 grawity by default git tools show the 'author' date, which might be different from the actual commit date
12:13 j`ey that's why i said commit date :)
12:13 grawity e.g. a patch written half a year ago but applied yesterday
12:14 ronskisika joined #git
12:14 Love4Boobies What's 'author' date?
12:14 Love4Boobies Sorry, new to git.
12:14 manuels joined #git
12:14 xock_stream joined #git
12:14 fbwnd joined #git
12:15 fbwnnd joined #git
12:15 j`ey grawity: in the graph it looks like X should contain Y, but git merge-base --is-ancestor X Y and Y X shows nothing
12:15 grawity it doesn't show anything, it uses exit status ($?)
12:16 j`ey oh lol
12:16 zeroed joined #git
12:16 Snugglebash joined #git
12:18 Miika-- osse: thanks, it worked also by manually looking up first and last sha in the git log history
12:18 j`ey this is confusing. it shows that Y is an ancestor of X.. but some of the commits merged in by Y are not an ancestor of X?
12:19 raspado joined #git
12:20 Peng_ joined #git
12:22 osse Miika--: sure. The argument is
12:22 f3r70rr35f joined #git
12:22 xo_odin joined #git
12:23 j`ey does --is-ancestor work for commits that are brought in via a merge?
12:23 Puffball joined #git
12:23 Akimb joined #git
12:23 osse Miika--: Sure. Those arguments work exactly the same for log and rev-list. The only difference is that log assumes "HEAD" if you don't provide any, and rev-list doesn't for some reason
12:24 cek joined #git
12:24 cek guys, I have an unconnected branch that I want to "glue" to some branch. How do I do that?
12:24 cek (never seen anything like that, was achieved with git fast-import)
12:25 wgrant joined #git
12:25 ^7heo #define 'unconnected branch'
12:25 ^7heo #define glue
12:25 ^7heo #define that
12:25 moritz you generally glue togther branches with 'git merge'
12:25 ^7heo ah that kind of glue
12:25 ^7heo rebase works too, otherwise.
12:26 ^7heo depending on your workflow.
12:26 cek it seems fast-export just detached it from base branch
12:26 zeroed joined #git
12:26 cek effectively, branch1..branch2.
12:27 osse cek: what do you mean by glue? Instinctively I'm thinking rebase --onto
12:28 darkbit joined #git
12:28 cek i'm not sure myself.
12:30 osse cek: by unconnected, to do you mean the the first commit of your branch is an initial commit ?
12:30 noecc joined #git
12:30 cek yes!
12:31 osse cek: man git-replace. See the --graft option
12:31 gitinfo cek: the git-replace manpage is available at http://jk.gs/git-replace.html
12:31 osse cek: IOW: pretend the initial commit has a parent, and the parent should be a commit on the branch you want to "glue" it to.
12:32 courrier joined #git
12:32 cek I also didn't know fast-export would change commit hash
12:32 j`ey oh nice, someone committed something that said 'merge commit blah blah', but it wasnt actually a merge commit
12:33 mischat joined #git
12:35 TangoTango joined #git
12:37 _ng joined #git
12:37 mischat joined #git
12:38 Guest39285 joined #git
12:38 catphish_ joined #git
12:39 catphish_ how does git store trees on disk? they appear to be binary, or compressed
12:39 a_thakur_ joined #git
12:40 cek now the problem is that initial commit contains hell lot of diff ( essentially branch1), but I need only diff against branhc1' tip
12:40 osse catphish_: they are compressed with zlib
12:40 catphish_ osse: thanks
12:41 selckin https://git-scm.com/book/en/​v2/Git-Internals-Git-Objects
12:42 JeroenT joined #git
12:43 freimatz joined #git
12:43 catphish_ thank you
12:44 j`ey left #git
12:44 dsdeiz joined #git
12:44 dsdeiz joined #git
12:44 m1m3-50 joined #git
12:46 venmx joined #git
12:46 JeroenT_ joined #git
12:51 xall joined #git
12:55 johto joined #git
12:55 str joined #git
12:56 johto left #git
12:56 johnmilton joined #git
12:56 Guest39285 left #git
12:56 TangoTango left #git
12:57 kulelu88 joined #git
12:59 GavinMagnus joined #git
13:02 g_zen joined #git
13:03 isysd joined #git
13:03 thebope joined #git
13:04 synthroid joined #git
13:06 ngperfex joined #git
13:06 sdothum joined #git
13:08 WayToDoor joined #git
13:08 paladinn joined #git
13:14 nikivi joined #git
13:15 djb-irc joined #git
13:15 xocket joined #git
13:16 afuentes joined #git
13:16 GavinMagnus joined #git
13:17 griffindy joined #git
13:23 cek How do i edit a commit?
13:23 cek git replace --edit can't edit a replace reference
13:24 tobiasvl cek: is it the most recent commit?
13:24 catphish_ left #git
13:24 m1m3-50 joined #git
13:24 cek no, that's first commit
13:24 tobiasvl OK, then !interactive_rebase is your friend
13:24 gitinfo Interactive rebase sounds similar to rebase but has completely different abilities. It can do this to commits: change the order, squash some of them together, remove some, add random existing commits from other branches, split them, and more... it's very powerful. Documentation is in the section "Interactive Rebase" in 'man git-rebase'.
13:24 cek I can't rebase to commit^
13:24 cek parent doesn't exist.
13:25 cek http://stackoverflow.com/a/1186549 won't work.
13:25 Starky joined #git
13:25 tobiasvl oh, yeah, quite. the very first commit
13:25 tobiasvl cek: https://en.wikipedia.org/wiki/Chess_boxing
13:25 Starky joined #git
13:26 qqx cek: git rebase -i --root
13:27 tobiasvl oooh nice.
13:27 cek git replace --edit firstcmt  wont' show me the diff
13:27 danslo_ joined #git
13:27 cek it shows only metadata
13:27 osse cek: if you graft like i said then the diff will probably be more correct
13:27 cek graft didn't work.
13:27 cek my first commit is wrong
13:27 osse because the diff will no be against the parent instead of nothing
13:28 cek osse, are you sure?
13:28 osse yes
13:28 johnmilton joined #git
13:28 cek because git show firstcommit shows large diff
13:28 cek it doesn't show parent..firstcommit
13:28 osse is the graft in place?
13:29 cek oh wait, hmmm
13:29 osse a new commit is created. if you do git show {original sha} then you get the original diff
13:29 jeffreylevesque joined #git
13:30 JeroenT joined #git
13:30 cek well, something is still wrong
13:31 cek it shows less files affected, butstill, > 1 file that it supposed to contain
13:32 djb-irc joined #git
13:33 anonymuse joined #git
13:33 cek git rebase -i --root dropped me into empty working tree
13:33 acetakwas joined #git
13:33 cek while I see for sure this first commit contained hell lot of stuff
13:33 cdown joined #git
13:34 alhariel joined #git
13:35 acetakwas joined #git
13:36 jaguarmagenta joined #git
13:39 cek funny thing is that on git commit while rebasing, it shows a lot of "new file: " , but git status doesn't show them!
13:41 delboy1978uk joined #git
13:41 osse do you know that the initial commit of the separate branch is based on a particular commit in the main branch?
13:41 delboy1978uk how to reset just one file?
13:41 osse if not, then the diff will be just an approximation
13:41 tobiasvl delboy1978uk: !revert
13:41 gitinfo delboy1978uk: That's a rather ambiguous question... options: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard uncommitted changes in the working tree [git reset --hard]; c) undo committing [git reset --soft HEAD^]; d) restore staged versions of files [git checkout -p]; e) move the current branch to a different point(possibly losing commits)[git reset --hard $COMMIT]?
13:42 cek osse, there's no parent branch
13:42 delboy1978uk checkout then i guess
13:42 osse cek: i didn't say parent branch
13:42 cek well why the heck "git show" shows me something then?
13:42 cek so the commit contains the diff
13:43 cek I want to edit the damn diff
13:43 osse it shows a diff against nothing.
13:43 cek You are currently editing a commit while rebasing branch 'local' on 'd06af8b'.
13:43 osse as if there was a commit with nothing in it
13:43 cek $ git diff
13:43 cek empty
13:43 JeroenT joined #git
13:43 cek why nothing, there are a lot of files in there
13:44 osse 'git diff' diffs the working directory against the current commit
13:44 osse that totally different
13:44 cek how do I edit d06af8b diff?
13:44 osse sounds like you have already run git rebase -i
13:45 cek yes, --root, i'm on first step
13:45 osse so make changes to the actual files, then commit --amend; then git rebase --continue
13:45 osse like git rebase told you
13:45 cek i don't see what files are modded
13:45 osse currently there are no modded files. you're supposed to make the modifications
13:45 nowhere_man joined #git
13:45 cek oh, it reset to pre-commit state
13:46 cek that's good.
13:46 nfk joined #git
13:46 osse not pre-commit. commit
13:46 osse the files on disk are like they are id d06af8b
13:46 Learath2 joined #git
13:50 cek they are not.
13:50 cek they are pre-d06af8b
13:53 fmcgeough joined #git
13:53 elect joined #git
13:54 szgyg joined #git
13:55 thethorongil joined #git
13:56 delboy1978uk joined #git
13:56 delboy1978uk left #git
13:57 thebope joined #git
13:57 rnsanchez_ joined #git
13:57 safernandez joined #git
13:57 mischat joined #git
13:59 cek well, after git reset parent, then making proper mods, it finally worked.
13:59 cek although, I had to correct some post commits
14:00 cek I didn't know moving a branch from one repo to another same repo is such a headache
14:00 cek apparently, people use only pull/push but don't move stuff manually (on usb for ex.)
14:01 nikivi joined #git
14:01 GavinMagnus joined #git
14:02 meauses joined #git
14:05 anonymuse joined #git
14:05 LionsMane joined #git
14:06 codeiscode joined #git
14:06 bremner git bundle lets you do both
14:07 lindenle joined #git
14:07 qqx Don't necessarily even need to use git bundle. You could also have a full repo (either bare or with worktree) on a USB drive.
14:07 lindenle joined #git
14:08 Gsham joined #git
14:08 cek yeah, git bundle sounds good but binary
14:09 elsevero joined #git
14:09 al-damiri joined #git
14:10 a_thakur joined #git
14:10 ixjlyons joined #git
14:10 bremner qqx: yeah, but then you have to deal with crap filesystems, typically
14:11 ybden joined #git
14:11 Darcidride joined #git
14:11 denisMone joined #git
14:11 qqx True, the filesystem on USB drives is generally pretty bad. But I wouldn't expect that to be much of an issue for a bare repo.
14:11 qqx Haven't actually tried that though.
14:12 cdg joined #git
14:13 ExoUNX joined #git
14:15 xock_stream joined #git
14:15 cdg joined #git
14:19 mischat joined #git
14:21 BrianBlaze420 joined #git
14:31 todd_dsm joined #git
14:31 ShekharReddy joined #git
14:31 nettoweb joined #git
14:34 nettoweb joined #git
14:37 Es0teric joined #git
14:38 anuxivm joined #git
14:39 Peng_ joined #git
14:40 inflames joined #git
14:42 Snugglebash joined #git
14:46 Gsham joined #git
14:50 AaronMT joined #git
14:51 thebope joined #git
14:53 crose joined #git
14:55 masuberu joined #git
14:56 elfknowski joined #git
14:56 nettoweb joined #git
14:57 Gsham joined #git
14:57 bannakaffalatta joined #git
14:58 jstein_ joined #git
14:58 a_thakur joined #git
14:59 bjoe2k4 joined #git
15:00 boombatower joined #git
15:00 mischat_ joined #git
15:00 sangy joined #git
15:01 xall joined #git
15:03 thebope joined #git
15:03 Dougie187 joined #git
15:07 EricPoe joined #git
15:07 e14 joined #git
15:11 kotk joined #git
15:13 shinnya joined #git
15:13 Drzacek joined #git
15:13 Drzacek Hi
15:14 jimi_ joined #git
15:15 tchan joined #git
15:15 tchan joined #git
15:15 Drzacek I have a merge commit that went wrong (pulled changes from remote that didn't mixed well with what I have locally). I want now to revert to last commit and pull changes from remote again and resolve conflicts - how do I do that?
15:15 xocket joined #git
15:16 selckin !reflog
15:16 gitinfo The git reflog (`git log -g`) records the SHAs of your HEADs for 2+ weeks. `git checkout -b myrestore OLDSHA` and `git reset --hard OLDSHA` will relink to that state via a new and current branch respectively, see http://sethrobertson.github.com/GitFixUm/ for full details.  WARNING: reset --hard will trash any uncommitted changes!  Visualize with: gitk --all --date-order `git log -g --pretty=%H`
15:16 selckin !undo
15:16 gitinfo [!fixup] So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!
15:18 Snugglebash joined #git
15:20 davesidious joined #git
15:20 mischat joined #git
15:20 davesidious Hey folks!  What's the most sane way to try to programatically determine the branch structure of a repository (which branches exist, from which branches are they branched, etc.)?
15:21 selckin gitk --all
15:21 selckin oh programaticlly, nevermind
15:21 davesidious :)
15:23 Rodya_ joined #git
15:23 davesidious If I could figure out how to list which branches are direct children of a given branch, that would be all I'd need.
15:23 davesidious my git-fu is weak.
15:25 eivarv joined #git
15:26 d10n-work joined #git
15:27 Drzacek okay, I think I did, now how do I solve conflicts with git?
15:28 osse davesidious: for branch in lots of branches; do git merge-base --is-ancestor $given_branch $branch && echo "omg $branch is direct child"; done
15:28 tang^ joined #git
15:28 davesidious osse: thanks very much!  I'll have a look.
15:29 re1 joined #git
15:30 IrishPrime joined #git
15:31 atomicb0mb joined #git
15:31 davesidious osse: that doesn't seem to be working for me, but thanks.
15:32 mischat joined #git
15:33 osse davesidious: What goes wrong?
15:34 davesidious It doesn't seem to work - child branches are not indicated as such.
15:35 davesidious It works with master and develop, but for any children of develop checked against develop it doesn't indicate they are children.
15:35 osse Maybe they aren't
15:35 nicksloan joined #git
15:36 davesidious I'm looking at a visualisation of them, and they are :)
15:36 osse davesidious: Does develop have commits that the children don't?
15:36 davesidious yes
15:37 osse Then develop is not an ancestor anymore. Some older commit on develop is, but not the tip.
15:37 davesidious I'm not worried about the tip - I just want to know from which branch the branch in question was branched from
15:37 osse That information is not stored
15:37 osse You'll have to do heuristics
15:39 chigang joined #git
15:40 davesidious The visualisation is managing to do it rather well, so it must be possible somehow.  That's what I'm here to find out :)
15:40 davesidious But thanks for the information :)
15:42 osse As in develop is shown as a straight line and the other branch are on the side?
15:42 osse Probably based on dates etc. It's not reliable
15:42 davesidious But it knows where things branched from.  It's never been wrong yet, for me at least.
15:43 jknetl joined #git
15:43 osse Been wrong for me several times.
15:44 thebope_ joined #git
15:44 davesidious git log --graph seems to be able to manage it, too...
15:44 davesidious and gitk
15:44 davesidious either they're all lying (and have yet to be caught), or it's possible :)
15:44 osse It might be correct. It might be correct 95% of the time. But that information isn't directly available, so it's based ob heuristics one way or another
15:46 osse Doesn't develop have the most recent commit?
15:46 davesidious It has the latest commit.
15:47 mischat_ joined #git
15:48 leeN joined #git
15:48 jbeeze joined #git
15:49 a_thakur joined #git
15:50 osse try making a commit on one of the other branches and see what happens. log --graph will display that in a straight line here, and master off to the side
15:51 davesidious Yup - that's what happens.
15:51 davesidious (still showing the structure perfectly :))
15:52 BrianBlaze420 joined #git
15:52 osse but what is now the parent branch and what are the children?
15:52 davesidious The branch is still the child of the branch it was before.
15:52 osse why?
15:52 davesidious Nothing has changed except the graph has been drawn slightly differently.  The relationships are still precisely the same as before, except with one commit.
15:53 jaguarmagenta joined #git
15:53 davesidious one commit *more*, I should have said.
15:53 osse yes, one branch pops out to the side and one is straight.
15:53 osse but which is the parent and which is the child ?
15:53 davesidious The existing feature branch from develop is still the child of develop.
15:53 osse why?
15:54 lss8 why does everyone append ".git" to a github clone url? e.g. https://github.com/torvalds/linux.git instead of https://github.com/torvalds/linux
15:54 lss8 I always clone via the non-.git url, works fine
15:54 davesidious why what?
15:54 osse davesidious: why do you think that the feature branch is a child of develop?
15:54 Anja joined #git
15:54 osse what in the visual output makes you say that
15:55 davesidious because it has a line from "develop", from their last shared common commit, and that line contains the commits from that feature branch.
15:55 robotroll joined #git
15:55 osse wouldn't that still be true if develop was a child of the feature branch ?
15:55 madewokherd joined #git
15:56 e lss8: because it's tradition? also i don't know if the latter has always worked
15:56 osse there is a line from "feature" from their last shared common commit, and that line contains the commits from that develop branch
15:57 osse lss8: bare repos usually have a directory name foo.git, since there is no hidden .git inside it. that's just convention.  My guess is github treats the two URLs the same.
15:57 davesidious True.  I see what you're getting at, and even if I have to use dates to figure this out,  it is still exactly what I'm looking to do.
15:57 WayToDoor joined #git
15:57 osse davesidious: but you can't trust dates, as we've discovered
15:57 davesidious But you can.
15:58 osse how?
15:58 davesidious At least I can, in my situation.
15:58 CrazyBonz joined #git
15:58 Raging_Hog joined #git
15:58 gr1sha joined #git
15:58 osse In that case you alreadu have your answer: develop is the main branch, and all the others are children of it.
15:59 davesidious That's a partial solution at best :)
15:59 clemf joined #git
15:59 gr1sha is my assumption that git protocol is weak for huge (70GB) repos cloning is somewhat right?
15:59 osse davesidious: there will never be a full one
15:59 davesidious I'm not saying there is.
15:59 osse then we are in agreement
15:59 gr1sha I have a mirror of AOSP repo on my LAN and I still get network errors when I try to access it by git protocol
15:59 pikajude_ joined #git
15:59 davesidious But there is most likely something between "naaah can't be done mate sorry" and "oh look Jesus just gave this perfect solution to me" :)
16:00 venmx joined #git
16:00 davesidious seeing as I can see many examples in various different libraries, and they are all in accordance when viewing the same repository
16:01 osse Usually when you start a feature branch, you start at what is the tip of develop at that point right. So if you compare the dates of the *first* unique commit in two branches, then it's likely that the branch containing the oldest commit is a child of the other.
16:02 vir_ joined #git
16:02 osse best I can come up with on the spot
16:02 vuoto joined #git
16:02 inflames joined #git
16:03 vuoto joined #git
16:04 subhojit777 joined #git
16:05 zeroed joined #git
16:05 zeroed joined #git
16:05 sawk do you think the world really needs an open source user interface for git? I mean something well done with all the features, but not as ugly as gitk :-)
16:06 selckin many exist
16:06 qqx joined #git
16:06 bremner gitk is not really a complete user interface
16:06 sawk I seen many of them, but none is working well on desktop
16:07 sawk At least not in Linux
16:07 UTAN_dev joined #git
16:07 sawk https://www.gitkraken.com/
16:07 ayogi joined #git
16:07 raspado joined #git
16:07 ToxicFrog I've been pretty happy with gitk for history exploration and gitcola for making commits
16:08 sawk This gitkraken works good and it's fully functional, but it looks like they are going to make it non-free pretty soon
16:08 sawk it already needs the login to the website. That's crazy to me
16:09 selckin so buy it or use something else
16:09 sawk :-)
16:09 xall joined #git
16:09 sawk i'm asking because that kraken program looks what i would like to find, but open source
16:10 osse there's also git gui, and git cola
16:10 selckin sourcetree
16:10 sawk sourcetree is not for linux ;-)
16:10 morenoh149 joined #git
16:10 oskarkv joined #git
16:10 selckin linux people are too smart for guis
16:10 morenoh149 is there a way to tell git-diff to ignore changes in the working tree?
16:11 ojdo joined #git
16:11 selckin morenoh149: --cached
16:12 davesidious Thanks for your help, osse - I'll look into this :)
16:12 e morenoh149: that question almost sounds as if you think a plain `git diff` includes staged changes
16:12 e so er, be aware that it doesn't, in that case
16:12 sawk <selckin> linux people are too smart for guis
16:13 sawk that's true
16:13 sawk but still, Linux is growing day by day, and there are people who needs something easy to learn for mass production
16:13 e i think a more sane CLI would be a lot beter for git users than any GUI
16:14 sawk e: what do you mean?
16:14 selckin notting will help people that just come from svn and refuse to learn any of the concepts
16:14 e understanding git--like you get from learning the CLI--is good
16:14 morenoh149 e: I was thinking `git diff head~1 --cached`
16:14 morenoh149 I wanted to review the last commit but ignore my local changes
16:14 e `git show`
16:15 selckin !xy
16:15 gitinfo Woah, slow down for a bit. Are you sure that you need to jump through that particular hoop to achieve your goal? We suspect you don't, so why don't you back up a bit and tell us about the overall objective...
16:15 selckin bah
16:15 sawk selckin: sure, but we know how it works in big companies. Time = training = Money
16:15 selckin sawk: they still fuck everything up even with a gui client that someone with a clue has to go fix
16:15 e so yeah. ...and i don't think it is the problem with the CLI, the problem is that the commands have a fairly random relationship with functionality
16:15 morenoh149 e: cool!
16:17 alexmajor joined #git
16:17 thebope joined #git
16:17 xock_stream joined #git
16:17 sawk I'm working with people who are not able to use a VS in command line, just because they always used svn, perforce (fuck...I know..), and other VS which already come with a UI
16:17 e can they really not or are they just scared to
16:18 sawk They fuck up everything, that's true, but it's also true that the world is going in a really specific way: git. And Linux is growing up..it's a matter of time and people will use Linux for mass production
16:18 aax joined #git
16:19 mischat__ joined #git
16:19 morenoh149 information theoretic pov
16:19 xaa joined #git
16:19 sawk And companies which develop on Linux will switch to git, at certain point. And they need something easy to use, because (unfortunately) a lots of engineers are hard to teach :-)
16:20 e don't they kind of suck at engineering, then?
16:20 thiago it's not just Linux
16:20 BrianBlaze420 joined #git
16:20 moritz that's why many vendors develop (or have developed) git integrations for their IDEs
16:20 sawk e: yes!
16:20 thiago it's all of Open Source switching to Git
16:20 moritz I don't quite see the problem
16:20 e learning is the main skill involved
16:20 thiago and even Microsoft
16:20 surfnturf_ joined #git
16:20 e well, the problem is that all of them are terrible
16:21 e they try to be simple, but they don't know how to simplify the complicated cases git has, so they just break stuff or refuse to work
16:22 surfnturf_ i merged a branch with a lot of commits onto my main branch. But now i want to go back to before the merge. However, I want to keep a separate branch of the new edits for later use
16:22 sawk exactly, but how a big company, composed by 1000+ employeers, can fire someone just because doesn't want to learn git in command line? ;-)
16:22 bremner sawk: well, such a company can pay
16:22 sawk e: a lot of engineers are conservative
16:22 surfnturf_ should i rebase or merge?
16:22 e surfnturf_: no
16:23 mischat joined #git
16:23 moritz sawk: they can absolutetly fire somdebody unwilling to learn what they need to know for their job
16:23 bremner sawk: either pay for non-free software, or pay for training, or pay to have free software written
16:23 sawk bremner: yes, and they because they are good on what they do, but they suck at git
16:23 e surfnturf_: why do you want to go back before the merge?
16:23 sawk they do*
16:23 moritz sawk: have you pushed the merge already?
16:23 surfnturf_ we merged a big set of changes onto our dev server. but i realize now it will take a lot of edits before its ready
16:24 surfnturf_ in the meantime i can't make small fixes to the codebase
16:24 surfnturf_ unless i edit them on staging
16:24 surfnturf_ and then push to production
16:24 moritz sorry, meant surfnturf_
16:24 moritz but it seems like "yes"
16:24 mischat_ joined #git
16:24 sawk moritz: how can you fire someone is really really good at (for example) electronic eng, but he sucks at versioning?
16:24 surfnturf_ i have pushed the merge to dev
16:25 sammyo joined #git
16:25 sawk moritz: the difficulty is to teach to good employers something in _that_ way, so they can become better engineers
16:25 moritz sawk: it's not about being good at versioning, it's about whether they are willing and able to learn
16:25 e first of all, the branch for keeping this in is easy, you can just use git branch to make another branch pointing there
16:25 moritz sawk: if somebody was a good electrical engineer 30 years ago, and they refuse to learn new tools, they'll likely be obsolete by now
16:26 surfnturf_ e: yes I did that
16:26 e okay, so i guess all that remains is to establish whether more people will want to kill you if you forcepush your dev branch or make a huge mess of the history
16:27 sawk someone with 30y old experience on embedded systems can't be fired just like that. Maybe he's using windows 98 with oldish systems, but still they can amazing on what they do and they are perfectly matching into this world
16:27 surfnturf_ well we're a 2 person operation
16:27 surfnturf_ so not really
16:27 e oh, then i'd forcepush every time, i guess
16:27 surfnturf_ forcepush != rebase?
16:27 e Not Exactly
16:27 nicksloan joined #git
16:27 tyreld joined #git
16:27 e rebasing published things will usually mean forcepushes
16:27 e but in this case no
16:28 sawk so, back to question: is there a multi-platform ui for git, which is user-friendly enough and not as ugly as git gui/gitk that can be used for mass-production and teach people using git in a graphical way? :-)
16:28 e i take it you have made a branch of the point you wanted to remember, and committed or stashed any other changes?
16:28 nettoweb joined #git
16:28 surfnturf_ yes
16:28 surfnturf_ then i checked out the staging branch
16:28 surfnturf_ so i could push that onto dev
16:28 surfnturf_ to go backwards
16:28 jstimm joined #git
16:29 e oh, there was nothing else between dev and staging but that merge?
16:29 surfnturf_ yes but that merge had like 20 commits
16:30 inflames joined #git
16:30 jstein_ joined #git
16:31 a_thakur_ joined #git
16:31 e then if there is definitely nothing of value in dev, you could probably do something like: git branch -f dev staging ; git push -f origin dev
16:32 mischat joined #git
16:32 e for correct values of origin and staging
16:32 brent__ joined #git
16:32 crayon joined #git
16:34 dreiss joined #git
16:35 xall is there a 'go to' post explaining why git is so great? i don't need convinced but i'm a young blood and never experienced other scms
16:35 selckin linus his original video did it for me
16:35 xall the google talk?
16:35 grawity it isn't particularly great
16:35 e i don't think git is so great, it's a tool, it has advantages and disadvantages
16:35 grawity it's just that everything else sucks
16:35 grawity except hg
16:35 e hg is pretty sucky too
16:36 e just in wildly different ways
16:36 grawity hg is probably easier to learn
16:36 xall so its like life?
16:36 e yes
16:36 grawity git is more scriptable/hackable
16:36 nikivi joined #git
16:36 grawity (maybe not quite in the same way as hg's plugins, but)
16:36 grawity also, git rebase
16:37 xall i see
16:37 e anyway, i guess the point is that i think you should learn several things
16:37 xall so linus + the fact that it took off
16:37 pikajude joined #git
16:37 e it was pretty early on the dvcs scene too
16:37 sawk git is a stable system with good advantages. And it's popular, so it's easier to find documentation on it
16:38 sawk and trust me, it's much better that centralized versioning systems ;)
16:38 Gobbert joined #git
16:38 xall is there anything particular about how it works vs hg or something?
16:38 e most of the things i dislike about git are peripheral UI things
16:39 _ikke_ xall: Mostly the very transparent object model
16:39 grawity xall: probably the way people hacked a VCS around the object-addressed storage system >_>
16:39 grawity er, on top of
16:39 _ikke_ xall: And how branches are implemented (hg later copied that)
16:40 xall right. i think i've read that git is sort of unique in that branching wasn't just a pointer in other vcs
16:40 cjbrambo joined #git
16:40 grawity that's part of it, yeah
16:40 joki- joined #git
16:40 grawity compared to other vcs'es, branching and tagging is really cheap
16:41 grawity hg branches are permanent and kinda scary
16:41 grawity (so now it has 'bookmarks' which are like git branches)
16:41 grawity svn branches are...
16:42 SwiftMatt joined #git
16:42 sawk and git scales really well on huge projects
16:43 grawity which is why facebook uses hg
16:43 NeverDie joined #git
16:44 sawk what git misses, in my opinion, is to include a native system for accesses administration (something like gitolite plugin)
16:44 _ikke_ The main reason was they found hg better adjustable (They have their own code around hg to make it scale for them)
16:44 sawk so code review can be done easier
16:44 e14 joined #git
16:45 synthroid joined #git
16:45 _ikke_ How would you do that without a central server?
16:46 moritz with a blockchain-like structure? :-)
16:46 grawity oh, like Fossil does/
16:46 _ikke_ How does fossil do it?
16:46 grawity hell if I know, other than their repos include bugtracker and wiki functionality (synced the same way commits are)
16:46 NeverDie joined #git
16:47 xall how are binary assets handled relative to git? is there some other software to handle it?
16:48 sawk i din't know fossil, it looks interesting
16:48 grawity xall: people use stuff like git-lfs or git-annex
16:48 moritz xall: you can either commit them, which blows up the size of the repo, or use external storage with integration into git
16:48 moritz what grawity said
16:48 grawity git-lfs is more for binary assets in regular repos, git-annex is more for keeping track of ~/Movies/ Git-like
16:49 Snugglebash joined #git
16:49 sawk btw, code review can be done easily by giving access to one specific branch to the code reviewer (the master for example). And the others to the developers (develop branch for example)
16:49 sawk code reviewer has gets notified that someone merged modifications in the develop branch, check the code and merge it into master
16:50 sawk oh nevermind...
16:50 SwiftMatt joined #git
16:50 WayToDoor joined #git
16:51 sawk i figured out that's divagation
16:52 publio joined #git
16:53 brillenfux joined #git
16:55 xall is there anything fancy about how git diff works? names are based on hash so it knows when there are changes. otherwise, it compares the blobs. is that right?
16:55 Es0teric joined #git
16:55 _ikke_ yess, git diff is mainly the xdiff algorithm
16:56 bsanford joined #git
16:56 svm_invictvs joined #git
16:56 _ikke_ Though they're currently experimenting with heuristics to make diffs more natural
16:57 Snugglebash joined #git
16:57 xall i saw https://github.com/so-fancy/diff-so-fancy recently. i like it
16:58 anuxivm joined #git
16:58 _ikke_ Right, the normal git diff output is machine readable, this is mainly human readable
16:58 grawity that just colourizes the pregenerated diff output
16:59 xall yeah
17:00 surfnturf_ e: thx for the help. it worked
17:04 tcorneli joined #git
17:04 GavinMagnus left #git
17:04 VagabondAzulien joined #git
17:05 govg joined #git
17:08 e14 joined #git
17:08 ochorocho__ joined #git
17:11 cjbrambo joined #git
17:12 seishun joined #git
17:14 spykins joined #git
17:15 spykins Can anyone help? I am trying to clone a private repository and I get remote: Repository not found.
17:15 spykins fatal: repository
17:16 _ikke_ spykins: Check the url you are using
17:17 moritz spykins: if you're using ssh keys, make sure your ssh-agent has only ssh keys that the remote associates with your account
17:18 spykins I am using https
17:18 moritz for example gitlab will always accept the first SSH key that it is offered, and then checks if this ssh key has access to the repo
17:18 xocket joined #git
17:18 spykins I have never used ssh before
17:18 moritz time to get started! :-)
17:18 spykins This is on github
17:18 moritz double-check the URL and credentials
17:19 spykins I have done this before but I have forgotten
17:19 spykins I need to add my username somewhere
17:20 nicksloan joined #git
17:20 cjbrambo joined #git
17:21 cdown joined #git
17:22 SegFaultAX joined #git
17:26 NeverDie joined #git
17:26 eltopo[m] joined #git
17:27 spykins Please help.. Say for instance I have this https://github.com/CDRussell/Mockito.git and it’s a private repository and my user name is spykins, please how do I get it into this format it@github.com:username/repo.git ?
17:27 kriskropd joined #git
17:29 astrofog joined #git
17:30 railssmi_ joined #git
17:30 SegFaultAX spykins: git@github.com:spykins/Mockito.git
17:30 SegFaultAX spykins: Not really sure what you're asking for beyond that. This is assuming you already have a fork on Github.
17:31 Rodya_ joined #git
17:31 spykins No, I don’t have a fork
17:32 spykins I just want to clone a private repository SegFaultAX
17:32 SegFaultAX spykins: You can't clone a repo you don't have access to.
17:32 spykins I have been added to the repo
17:33 Dougie187 spykins: If you have access to it though, you do `git clone git@github.com:CDRussell/Mockito`
17:33 _leb joined #git
17:33 SegFaultAX spykins: If you have access to the repo, then Github will generate the clone URIs for you
17:34 spykins I have the clone url Dougie187
17:34 SegFaultAX spykins: Then what's the problem?
17:34 spykins git clone theUrl
17:34 spykins It says fatal error
17:34 akushner joined #git
17:34 Dougie187 What is the fatal error?
17:34 thebope joined #git
17:35 SegFaultAX spykins: What happens when you do `ssh git@github.com`?
17:35 spykins emote: Repository not found.
17:35 spykins fatal: repository
17:35 sea-gull_ joined #git
17:35 CheckDavid joined #git
17:35 Dougie187 What command are you executing?
17:35 porsche944 joined #git
17:35 spykins git clone thePrivateRepoHttpUrl
17:35 Dougie187 ...
17:36 Dougie187 I hope you didn't execute that command, because that wouldn't make sense.
17:36 InfoTest joined #git
17:36 SegFaultAX spykins: Try the ssh command I posted. You should get a response like this: "Hi SegFaultAX! You've successfully authenticated, but GitHub does not provide shell access."
17:36 Dougie187 So, if that's not the command you executed, what command did you execute?
17:36 SegFaultAX It's an easy way to test that Github knows who you are.
17:37 spykins SegFaultAX: I get Permission denied (publickey).
17:37 acbg joined #git
17:37 Dougie187 spykins: Then your public key isn't setup correctly.
17:37 SegFaultAX spykins: You haven't correctly setup your ssh keypair with github.
17:37 Dougie187 https://help.github.com/arti​cles/generating-an-ssh-key/
17:37 spykins Thanks Dougie187
17:38 spykins I am going through the link now
17:38 a1l3erto joined #git
17:39 SegFaultAX spykins: If you're on Windows or macOS, you could use one of the dedicated Github clients. It will simplify this stuff for you.
17:39 ayjay_t joined #git
17:39 SegFaultAX https://desktop.github.com/
17:39 spykins I am on Mac
17:39 SegFaultAX You might prefer that to the command line interface.
17:39 spykins SegFaultAX: funny enough, I have cloned private repository before on my mac
17:44 Dougie187 spykins: If you're using the `git@github.com` syntax, you need private keys. But if you use the `https://` syntax, then you'll authenticate with a username and password.
17:45 Dougie187 However, I personally find the username/password more annoying, since you have to enter it every time you want to fetch / push
17:45 spykins Dougie187: I want to use https://
17:45 bremner I agree, although there are credential helpers to make https less painful
17:45 emg joined #git
17:45 spykins git clone https://username:password@github.com/*****/******.git
17:46 spykins Can I do this?
17:46 Dougie187 I'm not sure you can specify your username / password like that.
17:46 Dougie187 But even if you could, it's not a great idea, because it would be in plain text in your remote name.
17:46 Dougie187 I think you would do `git clone https://username@github.com/*****/*****.git` and then it would propmt you for your password.
17:47 m0viefreak joined #git
17:48 Dougie187 Github will give you the clone URI though.
17:49 skift joined #git
17:50 spykins Thank you Dougie187
17:50 spykins Exactly what I needed to do
17:50 Dougie187 np
17:50 spykins I have cloned the repo successful
17:50 encod3 joined #git
17:50 surfnturf_ left #git
17:50 spykins You are a life saver
17:52 neca joined #git
17:52 nettoweb joined #git
17:53 nidr0x joined #git
17:53 mischat joined #git
17:53 jaguarmagenta joined #git
17:54 manuels joined #git
17:56 panzon joined #git
17:56 railssmith joined #git
17:57 panzon Hi, Do you know if in git is enabled the https authentication method?
17:57 panzon I always have used ssh
17:58 panzon and now in  my company that has many strict rules, maybe will become a problem the usage of ssh
17:59 Dougie187 panzon: It's more a question of if the server has it setup.
17:59 Dougie187 At least afaik
18:01 thebope_ joined #git
18:01 grawity surely the company won't forbid SSH if it's for accessing work repos
18:03 masuberu joined #git
18:03 jeffreylevesque joined #git
18:04 ugjka joined #git
18:05 akushner joined #git
18:05 nickthetait joined #git
18:06 kpease joined #git
18:10 lamont joined #git
18:10 lamont joined #git
18:11 bjoe2k4 joined #git
18:11 vuoto joined #git
18:12 cyan__ joined #git
18:12 kba I have a private repository that I want to push incremental changes to, so I've added that using `git remote add pdev https://github.com/...`, then done `git fetch`
18:12 eivarv joined #git
18:12 kba I now created a commit, but when I try to push it with `git push pdev`, I get "Everything up-to-date", but it's clearly not
18:13 JanC_ joined #git
18:13 kba `git log` shows 2 commits I don't see on github
18:13 trelane joined #git
18:13 trelane joined #git
18:13 kba the first being a merge, the other being my actual commit
18:13 Macaveli joined #git
18:14 kba What's going on? How do I push my commit?
18:16 spacelord joined #git
18:17 VagabondAzulien joined #git
18:20 barduck joined #git
18:21 encod3 joined #git
18:22 mindriot101 joined #git
18:23 Cavallari joined #git
18:23 panzon joined #git
18:25 akkad joined #git
18:26 sdothum joined #git
18:27 vuoto joined #git
18:32 ugjka joined #git
18:32 vuoto joined #git
18:33 dviola joined #git
18:34 Es0teric joined #git
18:35 zeroed joined #git
18:37 lmat I nearly always work in a detached head...is that normal?
18:37 pikajude what are you, a wasp?
18:38 dreiss joined #git
18:38 matsaman joined #git
18:38 Anupkumar joined #git
18:39 ResidentBiscuit You should really see a doctor asap
18:39 kadoban lmat: Not at all
18:39 thiago_ joined #git
18:40 Dougie187 lmat: In some repos I like to work in a detached head a lot.
18:40 Dougie187 It depends on what I'm doing a lot.
18:40 abyx joined #git
18:41 Dougie187 Reviewing PRs is generally easier (imo) in a detatched head, as is performing merges.
18:41 Dougie187 but, personal preference.
18:41 ugjka joined #git
18:41 Sonderblade joined #git
18:42 sssilver joined #git
18:44 lmat This repo is shared with several others, and I generally have several somewhat short-lived feature branches.
18:45 WoodLeg joined #git
18:45 lmat The major factor is probably that the upstream is updated fairly frequently. If I had a real branch for each feature, I would have several branches that all need to be rebased onto the new tip.
18:45 lmat It's much easier to simply   git checkout upstream/devel; and work from there.
18:45 lmat It *is* annoying however, that you can't create a remote branch from a detatched head.
18:46 lmat you can    git push origin somelocal:someremote;  and this can create a new remote. Also, you can  git push origin HEAD:someremote;  when 'someremote' already exists.
18:46 lmat You can't    git push HEAD:someremote; when someremote doesn't exist :-(
18:46 lmat git gets confused or something.
18:46 Faylite joined #git
18:46 a1l3erto joined #git
18:47 lmat If I could do that, I don't see why I would ever create a local branch ^_^
18:47 pwnz0r joined #git
18:47 cagedwisdom joined #git
18:49 aard_ joined #git
18:50 VagabondAzulien joined #git
18:50 steelsw0rd joined #git
18:52 SegFaultAX lmat: That's a terrible workflow.
18:52 zeroed joined #git
18:52 mindriot101 joined #git
18:53 SegFaultAX lmat: Also, you can definitely `git push <remote *or* repo> refspec` if you want.
18:53 PCatinean joined #git
18:54 SegFaultAX Eg, where remote "foo" is git@github.com:segfaultax/foo, I could `git push foo my_local_branch:a_remote_branch`
18:54 railssmith joined #git
18:54 ahmedelgabri joined #git
18:54 dsdeiz joined #git
18:56 lmat SegFaultAX: Right, that's what I said.
18:56 lmat SegFaultAX: And you can    git push foo HEAD:remotebranch; *unless* remotebranch doesn't exist yet :-(
18:57 SegFaultAX lmat: You can do that whether or not the branch exists remotely.
18:57 kriskropd joined #git
18:57 SegFaultAX You may have to force it, or use + in the refspec to ignore non-fast forward state.
18:58 SegFaultAX lmat: Anyway, I'm not really sure what you're complaining about. But rest assured that working off of a detached head is a terrible idea.
18:58 Sasazuka joined #git
18:58 lmat SegFaultAX: Thanks for that explanation.
18:59 SegFaultAX lmat: Are you using the reflog to find floating commits when you switch to something else?
18:59 mustmodify joined #git
19:00 svara1 joined #git
19:00 mustmodify I just did `git reset HEAD~1` unintentionally. I pushed that commit to my origin. How do I fix it without messing up everything?
19:00 akushner joined #git
19:00 mustmodify Just delete the changed files and git pull origin master?
19:00 _ikke_ Why do I need to sign into gitkraken in order to use it :-/
19:01 mustmodify so they can sell you stuff. :(
19:01 svara1 joined #git
19:01 ToxicFrog mustmodify: git reflog to get the commit ID of the commit you just threw away, git reset to recover it, git push?
19:01 lmat SegFaultAX: I don't leave commits floating. Either I push or put down a branch.
19:02 boombatower joined #git
19:02 SegFaultAX lmat: That's the crux of why it's a bad idea: it's a really brittle way to work. Either you're going to create the branch anyway, or you're going to hope that no other changes happen upstream such that your local state remains fast-forwardable.
19:03 johnmilton joined #git
19:03 lmat SegFaultAX: I do   git rebase upstream/devel; a lot, and git push. I don't remember creating a local branch so that I can switch to something else in a very long time :-)
19:04 lmat SegFaultAX: creating the branch, then maintaining it when upstream moves seems like more headache than it's worth; but I see your point.
19:04 SegFaultAX mustmodify: Did you force push?
19:04 kadoban It sounds like you go through quite a bit of extra work to use remote branches as local branches. Which I guess if you're careful and your workflow isn't very complicated works.
19:05 SegFaultAX lmat: I usually do --no-ff when merging into master anyway because I *want* the history of feature branches being merged to master.
19:05 SegFaultAX lmat: So even if I rebase and the history is fast forwardable, I'll create a merge commit anyway.
19:06 SegFaultAX But anyway, if you have the luxury of a) only working on exactly 1 thing at a time and b) only doing small enough amounts of work that you can afford to constantly rebase with minimal conflicts, I guess that system could be workable, although still highly ill-advised.
19:09 thebope joined #git
19:10 f3r70rr35f joined #git
19:10 spacelord joined #git
19:11 flavius joined #git
19:13 mustmodify SegFaultAX: NO! I didn't force push.
19:13 mustmodify I removed the committed files and pulled.
19:14 mustmodify It all worked out.
19:14 mustmodify left #git
19:14 jstimm joined #git
19:15 Gsham joined #git
19:15 matsaman joined #git
19:15 vktec joined #git
19:16 vktec How should I go about resetting a single file to clean state?
19:16 _ikke_ git checkout <file>
19:16 [twisti] joined #git
19:17 Faylite joined #git
19:17 vktec Thanks _ikke_
19:17 akushner joined #git
19:18 [twisti] is there a way to set up kind of a proxy repo ? say i have online repo A, and part of my build process is something like 'git clone A; mvn package', but A is very big, so i dont want to redownload it every build. but on the other hand i want to run builds in fresh, empty dirs. so i was thinking, i could set up a local repo B, that clones A, and then clone B. but then when A gets a commit, 'clone
19:18 [twisti] B' wont get it until someone pulls A to B
19:18 [twisti] and assume we cant use hooks with A
19:18 mindriot101 joined #git
19:18 [twisti] is there a way to say "clone B (after pulling all upstreams)" ?
19:18 _ikke_ You can make a --mirror clone
19:19 _W_ joined #git
19:20 dsantiago joined #git
19:20 cdg_ joined #git
19:21 xwid joined #git
19:21 matsaman joined #git
19:22 [twisti] ill read up on that, thanks
19:24 endiruna joined #git
19:25 barteks2x I'm in the middle of git add -Ap and I selected yes instead of no for one hunk, any way to go back?
19:27 e14 joined #git
19:27 barteks2x I don't want to start from the beginning
19:28 _ikke_ I think there is, hold on
19:28 barteks2x actually, nevermind
19:28 barteks2x I selected correctly
19:28 vktec left #git
19:28 dsantiag_ joined #git
19:29 barteks2x But good to know there is no way
19:30 barteks2x description for option "K" says "leave this hunk undecided, see previous hunk", but when I try to use it,it says "No previous hunk"
19:30 _ikke_ I think there is a file boundary
19:30 _ikke_ (someone reported it on the mailing list the other day)
19:30 _ikke_ So if the previous hunk is in a different file, you cannot go back
19:31 barteks2x is it going to be changed?
19:32 barteks2x I'm currently dealing with small amount of changes in many different files
19:32 rolha joined #git
19:33 JamesBaxter joined #git
19:33 _ikke_ barteks2x: I cannot find the thread anymore
19:33 _ikke_ barteks2x: But you can use git reset -p to unstage a hunk again
19:33 _ikke_ not ideal, but at least something
19:33 cdown joined #git
19:34 barteks2x but that means I either have to stop what I'm doing now, or write down the file I got wrong
19:34 _ikke_ correct, like I said, not ideal
19:35 barteks2x if I press ctrl+c, will it save the current progrss and next git add -Ap start where I left off, or will it start from the beginning?
19:35 gahbu_ joined #git
19:36 _ikke_ barteks2x: You can press q
19:36 gmn8719 joined #git
19:36 _ikke_ But, it will skip the already staged changes, but everything not staged it will show again
19:36 matsaman joined #git
19:37 barteks2x right, didn't see that option. ctrl+c is just the universal think that very frequently gets me out of where I'm now
19:37 _ikke_ correct
19:38 SwiftMatt joined #git
19:39 cdown joined #git
19:39 e14_ joined #git
19:40 clemf joined #git
19:43 WoodLeg joined #git
19:44 xwid joined #git
19:45 brandonw joined #git
19:46 nikivi joined #git
19:47 Faylite left #git
19:47 dreiss joined #git
19:52 elsevero joined #git
19:52 cdown joined #git
19:53 adac joined #git
19:54 jaguarmagenta joined #git
19:56 satifant joined #git
19:56 xall joined #git
19:56 finalbeta joined #git
19:57 Rodya_ joined #git
19:58 qws-user-1228 joined #git
19:59 qws-user-1228 joined #git
20:00 clemf joined #git
20:01 xwid joined #git
20:05 vdv joined #git
20:05 mindriot101 joined #git
20:09 NeverTired joined #git
20:10 t4nkd joined #git
20:11 Jackneill hey
20:11 Jackneill http://codepad.org/wa5g4jSN i have this error. i added previuously a submodule but forked it and want to add my own path. so i did a git rm subm-dir; and git submodule add myurl.git but i get this message
20:12 Jackneill if i use the --force option how can i check that the remote is correct?
20:12 dminuoso joined #git
20:12 Jackneill nvm
20:13 dminuoso Hi. Question came up whether you could commit without moving HEAD to the new commit in one command.
20:13 xocket joined #git
20:14 finalbeta joined #git
20:14 _ikke_ Nope
20:14 _ikke_ !xy
20:14 gitinfo Woah, slow down for a bit. Are you sure that you need to jump through that particular hoop to achieve your goal? We suspect you don't, so why don't you back up a bit and tell us about the overall objective...
20:15 kadoban You can as long as that command includes &&, assuming your shell is bash-like.
20:15 _ikke_ I don't consider that one command :P
20:15 _ikke_ Just two commands chained
20:15 griffindy joined #git
20:16 kadoban :)
20:16 Guest55 joined #git
20:19 thomasreggi joined #git
20:19 finalbeta joined #git
20:21 grayjoc joined #git
20:23 vuoto joined #git
20:23 aidalgol joined #git
20:25 cdg joined #git
20:28 Charliechin joined #git
20:28 qws-user-1228 joined #git
20:28 Jackneill i do an add_subdirectory(dir) from an included() CMakeLists.txt. i got an error that dir does not exist however it clearly does. what are paths relative to? also i guess then its good practice to always do ${CMAKE_CURRENT_LIST_DIR}/..?
20:29 crose joined #git
20:29 noecc left #git
20:29 thebope joined #git
20:29 thiago Jackneill: wrong channel?
20:30 Jackneill oh, sorry
20:30 averell joined #git
20:30 Jackneill #git is right under #cmake :)
20:30 thiago n/p
20:31 vuoto joined #git
20:31 aidalgol joined #git
20:33 davidfetter_ge joined #git
20:33 davidfetter_ge hi
20:33 arand joined #git
20:33 _ikke_ hello
20:33 davidfetter_ge git is guessing wrong about a rename. how do i tell it "change renamed foo -> bar to renamed foo -> baz"
20:33 davidfetter_ge ?
20:34 _ikke_ Git does not record that
20:34 davidfetter_ge oh
20:34 davidfetter_ge so it just looks like a delete + create in the history?
20:34 thiago yes
20:34 thiago and then compares content to guess
20:35 davidfetter_ge it's the content comparison it got wrong. any way to hint at that?
20:35 thiago no
20:35 qqx Git doesn't really store changes at all. Each commit is effectively a separate snapshot.
20:35 davidfetter_ge ah
20:35 davidfetter_ge then i guess that's for the commit message
20:36 anuxivm joined #git
20:38 clemf joined #git
20:39 zulutango joined #git
20:43 jstimm joined #git
20:44 thomasreggi joined #git
20:45 heroux joined #git
20:51 emg joined #git
20:51 Stadtpirat joined #git
20:52 dreiss joined #git
20:55 viny joined #git
20:55 alex1a joined #git
20:55 boombatower joined #git
20:55 converge joined #git
20:56 viny if amended a commit after pushing it to a remote repo, so I can no longer push to it because i'm stupid
20:56 viny can I diff origin/master to my current master to re-commit what I amended?
20:57 anonymuse joined #git
20:57 ^7heo viny: what I'd do is: reset from before the last commit
20:57 ^7heo (not reset hard, but reset)
20:58 ^7heo git stash
20:58 ^7heo wait no git add
20:58 ^7heo and pull
20:58 ^7heo then reset again.
20:58 mischat joined #git
20:58 mda1 joined #git
20:59 xall joined #git
21:00 dsdeiz joined #git
21:01 dminuoso _ikke_: Don't worry - not an xy problem. It was just a question that resulted from a humerous discussion and I just wondered if it was possible - purely hypothetically.
21:02 Learath2 joined #git
21:04 nettoweb joined #git
21:05 kelevra joined #git
21:05 thebope_ joined #git
21:07 Atm0spher1c joined #git
21:08 kelevra left #git
21:08 vuoto joined #git
21:11 boombatower joined #git
21:12 xock_stream joined #git
21:15 cdg joined #git
21:17 cdg_ joined #git
21:19 grayjoc joined #git
21:23 nfk joined #git
21:23 vdv joined #git
21:24 sea-gull_ joined #git
21:25 blackwind_123 joined #git
21:32 manuels joined #git
21:33 Es0teric joined #git
21:37 Darcidride joined #git
21:37 seishun joined #git
21:38 Darcidride joined #git
21:39 nowhere_man joined #git
21:40 venmx joined #git
21:43 mniip_ joined #git
21:44 Cavallari joined #git
21:45 BenderRodriguez joined #git
21:46 Orphis joined #git
21:46 ugjka joined #git
21:48 grayjoc joined #git
21:49 courrier joined #git
21:51 pothepanda joined #git
21:52 sea-gull_ joined #git
21:53 mindriot101 joined #git
21:53 ryouma joined #git
21:53 ryouma under what circumstances would a ~/.gitignore be ignored in a ~/whatever/.git?
21:54 ryouma is ~/.gitignore only ever looked at when there is no local one?
21:54 ryouma or even /home/whatever/whatever/.git
21:54 Eugene if core.excludesFile is specified in the repo
21:55 Eugene That's unusual(and silly). Is this an !ignore_tracked situation?
21:55 gitinfo Git only applies ignore patterns to untracked files. You can't use ignore patterns to ignore changes to files that are already tracked by git. To remove files only from git, but keeping them on disk, use git rm --cached <file>. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.
21:55 jaguarmagenta joined #git
21:56 nwkj86 joined #git
21:56 Phylock joined #git
21:58 ToBeCloud joined #git
21:59 cdown joined #git
22:00 Darcidride joined #git
22:01 nicksloan joined #git
22:02 mindriot101 joined #git
22:05 adac joined #git
22:06 mindriot101 joined #git
22:07 panzon Do you know if the  "Smart HTTP" git support includes the usage of HTTPS?
22:08 CrazyBonz joined #git
22:08 Peng_ joined #git
22:09 macivy joined #git
22:10 tvw joined #git
22:11 dreiss joined #git
22:12 qqx panzon: Yes, you can use HTTPS for either smart HTTP or the older HTTP transport.
22:12 kadoban It's orthoganol. HTTP/HTTPS is a separate question from smart-HTTP/dumb-HTTP
22:12 xocket joined #git
22:13 macivy left #git
22:14 panzon in my company we usually authenticate through ssh, but right now they gonna forbid the ssh connection... So we need to use HTTPS authentication, and enable our server to support https
22:15 panzon do you have any link,  that explain how to figure out the HTTPS in a git server?
22:15 Rodya_ joined #git
22:15 kadoban So your company is arbitrarily ruining your already existing setup? Complain and tell them to not.
22:15 panzon the already setup, was not installed by me
22:16 panzon and the company just says that no body have to use ssh to connect to the outside (where our git server lives)
22:16 Peng_ joined #git
22:17 j416 panzon: that's just nuts :/
22:17 kadoban panzon: That's pretty dipshitted.
22:17 j416 panzon: I'd just put things locally for a while and not push
22:17 panzon I know!
22:17 * j416 shrug
22:18 geoid joined #git
22:18 j416 panzon: maybe have them realise it'll cost a day for you to set things up, and an hours worth of disruption and setup for each other dev to fix each repo (easy, but not everyone knows or remembers how to set proper url, fix auth etc)
22:21 panzon I know man, but it is a  big company where I have no voice in these kind of choices. Actually the guy who configured git said that https is not supported by git!
22:21 vuoto joined #git
22:21 Intee joined #git
22:21 j416 panzon: I'd buy a mobile router and use that instead
22:21 j416 panzon: or use my phone to connect
22:22 j416 they should be able to easily allow ssh to github but not elsewhere, for example
22:22 Darren_ joined #git
22:23 j416 there's always a way to opt out.. ;)
22:23 panzon the problem is that some CI scripts download some source from the repository and then connect to some servers that we have access just when we are connected on the internal network
22:23 kadoban Ya ... don't do that, that's probably grounds for getting fired.
22:24 Dougie187 left #git
22:24 p4trix joined #git
22:24 creppe joined #git
22:24 Dougie187 joined #git
22:25 j416 kadoban: I wouldn't want to work for a company that would rather fire a dev than set up ssh firewall rules properly :)
22:25 j416 but each to their own ;)
22:25 tcurdt joined #git
22:26 _Patrick_ joined #git
22:28 Snugglebash joined #git
22:28 kadoban j416: Well, better to just quit than get fired for intentionally going around IT rules really. That'd be ... not good.
22:28 kadoban They tend to take it badly when you don't follow their bullshit rules.
22:29 mizu_no_oto joined #git
22:33 netj joined #git
22:35 Dougie187 left #git
22:35 crazyphil joined #git
22:36 Dougie187 joined #git
22:38 crazyphil ok - trying to use git cli in the middle of a jenkins job to push some info to a repo - I'm trying to push to a git@dns.name.com type reference that has a special account that does not use a key - how/where do I give my git commands these credentials as my internet searches have only added confusion
22:39 crazyphil (repo is hosted on a gitlab enterprise platform)
22:39 nfk joined #git
22:41 Salander27 joined #git
22:41 mozzarella guys
22:41 mozzarella hello
22:42 tcurdt joined #git
22:45 stomplee joined #git
22:45 ochorocho__ joined #git
22:46 peeps[lappy] if you git clone from a local path, its only going to get data from latest commit, right?  no stash info, working dir changes, etc?
22:47 pwnz0r joined #git
22:47 anonymuse joined #git
22:48 dminuoso peeps[lappy]: well stash is just refs/stash which can be pushed at least
22:48 railssmith joined #git
22:49 anonymuse joined #git
22:51 anonymus_ joined #git
22:51 Dougie187 left #git
22:51 Dougie187 joined #git
22:52 creppe joined #git
22:54 cdown joined #git
22:56 LaptopBrain joined #git
22:56 peeps[lappy] and the new cloned repo will point to the same branch as whatever the local repo it was cloned from right?
22:56 j416 kadoban: I suppose.
22:56 j416 you've a point.
22:56 j416 :)
22:57 peeps[lappy] i'm trying to write a little script to help me run builds based on what i've been working on
22:57 peeps[lappy] just want to make sure i understand the behavior correctly
22:57 peeps[lappy] that i'm doing what i intended to do
22:58 mozzarella on github, if I want to work on some project and submit my changes, should I clone the repo right away or should I fork it first and clone my fork instead?
22:59 peeps[lappy] mozzarella, i would fork first, then you don't need push permissions.  also there is a #github
23:00 Gobbert joined #git
23:00 thomasreggi joined #git
23:00 mozzarella how can I then pull the changes from the original repo?
23:00 davidfetter_ge i was trying to get gitinfo to tell me about how to edit a commit, but i'm not getting the syntax right
23:01 Aboba joined #git
23:01 sangy joined #git
23:03 cdg joined #git
23:03 Dougie187 left #git
23:04 vuoto joined #git
23:04 Gobbert joined #git
23:04 WaReZ joined #git
23:07 cebor joined #git
23:07 rooftopjoe joined #git
23:08 rooftopjoe does anyone know whether downloading the master branch archive from github includes the source in the submodules? (i'm trying to find this out before i commit the relevant documentation and submodule).
23:09 rooftopjoe if anyone uses a submodule with their github project, link me to it so i can check :D
23:09 johnny56 joined #git
23:09 mozzarella honestly I don't know what kind of workflow should be used with github
23:09 morenoh149 joined #git
23:10 tcurdt joined #git
23:10 mozzarella you fork a repo, make some changes, commit and push them to your fork, then make a pull request on the original repo, is that it?
23:11 creppe joined #git
23:11 sbeller rooftopjoe: git archive doesn't include submodules
23:11 rooftopjoe thank you
23:12 ech0s7 joined #git
23:12 rooftopjoe mozzarella: sure, why not? nothing special about github.
23:13 mindriot101 joined #git
23:13 mozzarella I've only used git "locally"
23:13 ech0s7 i have this log: A - B - C - D - E - F - D(head), and i would remove the commit B and E, how can i do with rebase ?
23:13 vuoto joined #git
23:13 xock_stream joined #git
23:13 mozzarella rooftopjoe: are there other ways of accomplishing the same task?
23:13 jeffreylevesque joined #git
23:15 rooftopjoe mozzarella: well, you could just not use github and host your repo elsewhere, perhaps on your own machine.
23:15 Rodya_ joined #git
23:16 sbeller ech0s7: use --interactive rebase
23:17 rooftopjoe mozzarella: but ultimately, if people collaborate, there are central servers (where their stuff ought to eventually merge), annd local repos.
23:17 dodobrain joined #git
23:18 BrianBlaze420 joined #git
23:18 rooftopjoe if you want one commit to go from one repo to another, you need to pull.
23:18 rooftopjoe that's the gist of it
23:19 mozzarella but, do most contributors also create forks and pull requests, or do they directly push to the original repo?
23:20 rooftopjoe mozzarella: depends on whether they are associated with the project or not. typically, maintainers push but don't just allow anyone else to. if a volunteer writes a patch
23:20 rooftopjoe he does a pull request and a maintainer reviews and maybe accepts it
23:21 dminuoso mozzarella: It depends on the project structure in the end. Some projects require all code to be reviewed even from their own committers.
23:21 rooftopjoe pull requests are basically "can you please push this?"
23:21 dminuoso rooftopjoe: More like "can you please pull this"
23:21 dminuoso It's called "pull request" for a reason.
23:22 mozzarella alright, I think I get the gist of it
23:22 rooftopjoe i was trying to explain it in terms of pushing. the thing that makes it pulling is that it's from someone else's repo.
23:22 rooftopjoe but yeah.
23:23 rmbeer joined #git
23:23 rmbeer hello
23:23 rmbeer the git conserve the all dir and subdir that have in the list of .gitignore?
23:24 CrazyBonz joined #git
23:26 Spec-Chum joined #git
23:26 Xarodon joined #git
23:26 mozzarella should I use ssh or https?
23:26 Seveas mozzarella: yes.
23:26 tcurdt joined #git
23:26 ech0s7 sbeller: ok, so should be like this ? git rebase -i E ?
23:27 Seveas rmbeer: that question is a bit hard to understand. Could you maybe rephrase?
23:27 Seveas ech0s7: git rebase -i A, then change 'pick' to 'drop' for B and E (or remove the B and E lines entirely)
23:27 sbeller ech0s7: rather git rebase -i A, (then an editor pops up that shows A..E, which you remove the lines B and E from)
23:28 amaged__ joined #git
23:29 dsdeiz joined #git
23:29 dsdeiz joined #git
23:29 sharon_so joined #git
23:32 Gsham joined #git
23:33 ech0s7 thanks sbeller and Seveas
23:33 ech0s7 !
23:36 Lynxium joined #git
23:37 tcurdt joined #git
23:38 Vinnie_win_b joined #git
23:38 rmbeer Seveas, i have in the .gitignore a dir **/Datos , where i put in several directories of Datos/ for any datas. But something of this directory not exist in the other branchs, if i change with checkout, this directory is removed or is conserved? i have something like ./a/b/Datos/ where the dir+source is created like ./a/main.cpp, but ./a/ not exist in other branchs
23:38 Marbug_ joined #git
23:39 rafiki joined #git
23:40 SpeakerToMeat hello
23:40 cdown joined #git
23:40 SpeakerToMeat Is there any way to update the data of a remote origin without pulling in changes?
23:41 Derega_ joined #git
23:43 anonymuse joined #git
23:44 monokrome joined #git
23:46 rmbeer ok, i try copy all repo, test with this, and view that all directory with datas is conserved... thank.
23:47 rmbeer left #git
23:49 d0nn1e joined #git
23:51 cebor joined #git
23:52 davidfetter_ge how do i edit a patch that's already committed?
23:52 mozzarella Seveas: which one?
23:52 mozzarella can I push to my fork using https?
23:53 mozzarella or is that only possible with ssh?
23:55 raijin joined #git
23:56 Vangelis joined #git
23:56 jaguarmagenta joined #git
23:58 cagedwisdom joined #git
23:58 creppe joined #git

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