Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-11-30

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

All times shown according to UTC.

Time Nick Message
00:04 enleeten joined #git
00:06 dirtyroshi joined #git
00:08 Rodya_ joined #git
00:08 azerus joined #git
00:09 anuxivm joined #git
00:16 Rodya_ joined #git
00:20 Charliechin joined #git
00:21 SwiftMatt joined #git
00:22 moestevens joined #git
00:24 al-damiri joined #git
00:25 arand joined #git
00:26 anuxivm left #git
00:27 howdoi joined #git
00:28 hahuang65 joined #git
00:29 madewokherd joined #git
00:29 SwiftMatt joined #git
00:30 chachasmooth joined #git
00:35 Intee joined #git
00:38 ejb joined #git
00:38 arescorpio joined #git
00:41 dodobrain joined #git
00:43 drbean joined #git
00:44 sargas joined #git
00:45 rominronin joined #git
00:46 quackgyver joined #git
00:47 Darren_ joined #git
00:52 azerus joined #git
00:53 hahuang65 joined #git
00:53 NeverDie joined #git
00:54 mischat joined #git
00:55 sangy joined #git
00:55 keviv joined #git
00:56 revoltingPeasant joined #git
00:57 HoierM joined #git
00:58 inflames joined #git
00:59 Lyqyd joined #git
00:59 venmx joined #git
01:03 rwb joined #git
01:03 Sasazuka joined #git
01:03 xwid joined #git
01:04 nickabbey joined #git
01:04 PresidentBiscuit joined #git
01:07 chigang joined #git
01:08 bamb left #git
01:12 tterrag joined #git
01:13 theskillwithin joined #git
01:13 SwiftMatt joined #git
01:14 CubiK joined #git
01:15 keviv I want to "hide" hide a bunch of files in a different branch (they're not needed in the project right now but might be a few months down the road). The only way I can think of is to delete the files in the index, commit to master, branch off that commit and restore the files. Then master doesn't contain them, but I can merge the second branch down the road. Is there a better way?
01:16 przemoc joined #git
01:17 kadoban keviv: Seems like it'd work fine. You can use 'git revert' for restoring the files, but otherwise there's not much there to improve that I see
01:18 einnjo joined #git
01:18 keviv kadoban: Ah that would make it simpler
01:18 keviv Thanks
01:24 axl_ joined #git
01:25 buffal0 joined #git
01:28 jacob123 joined #git
01:29 eb0t joined #git
01:31 CubiK joined #git
01:35 floatingman joined #git
01:39 Nilesh_ joined #git
01:39 weaksauce joined #git
01:39 Rodya_ joined #git
01:39 hariel joined #git
01:41 malber joined #git
01:43 jaguarmagenta joined #git
01:47 weaksauce is there a way to tell what branch a branch was branched from originally?
01:48 weaksauce i have a branch named default and i don't know if it was branched off of central/default or inbound/default
01:48 Oatmeal joined #git
01:52 Random832 weaksauce, not really as such, but if you do "git log --decorate --oneline --graph default central/default inbound/default" you might figure out what you need to know
01:52 Random832 or if what you really want to know is what remote it is tracking, git branch -vv
01:56 _Soto joined #git
01:56 cqi joined #git
01:56 cqi joined #git
01:58 justanotheruser joined #git
01:59 Gsham joined #git
02:01 weaksauce second one was perfect Random832
02:02 UniFreak joined #git
02:02 _Soto Hey I was adding collaborators to my repository when I realised I probably don't want to give them push access. Is there a way to change permission levels for collaborator and set up a review system?
02:04 durham_ joined #git
02:04 HedgeMage _Soto: it depends a lot on how you are providing access to your git repos.  I use gitolite, which has the most granular access control of the options I'm familiar with.  Gitlab has less granularity, but does allow you to force review, etc.
02:04 HedgeMage they're all different
02:06 _Soto Okay. Doesn't Github inherently provide support for permission levels?
02:09 bremner probably. but you didn't ask in #github ;)
02:09 HedgeMage They do, but not with nearly the granularity I'm used to on gitolite.  Every time I set up a repo on there I struggle with it. :(
02:09 * HedgeMage == spoiled.
02:10 Vampire0 joined #git
02:11 _Soto bremner: Right you are. My mistake. HedgeMage: Gitolite sounds interesting! :) I'll give it a go. Although my team probably can work with the one on Github. Thanks!
02:12 bremner hrm.
02:12 Derperperd joined #git
02:13 d^sh joined #git
02:15 finalbeta joined #git
02:16 _genuser_ joined #git
02:17 _genuser_ hey folks. is there any way to search thru a repo for a given string literal in all the versions in the repo?
02:18 Qria joined #git
02:18 rominronin joined #git
02:18 _genuser_ so if someone includes some information on a commit, then deletes it on a subsequent commit, can one look for that and see if it's present an in which commit?
02:19 HedgeMage _genuser_: You'll want to use reposurgeon to operate on the git repo as a stream. :)
02:20 anonymuse joined #git
02:20 pks joined #git
02:20 HedgeMage _genuser_: http://www.catb.org/~esr/reposurgeon/
02:20 _genuser_ HedgeMage: thanks. and if it's on github, does that change the equation? (does GH have any tools, Or should I just work on my local copy and forget GH tools)?
02:21 HedgeMage I do not know if GH has tools for that.  I don't use github that often, and the only time I've done what you describe, it's on stuff I can't have on github (cleaning secret or classified stuff from a repo that some idiot accidentally committed then stitching together a clean repo history)
02:22 _genuser_ HedgeMage: heh, same thing on this one. I think I may have put my bintray api key in there at one point and the GPG private key password in there without ignoring that settings files.
02:23 _genuser_ now, I'm wondering if I should just nuke the entire repo, git init again and upload a fresh copy to GH.
02:23 _genuser_ alternative I thought of was to just delete my api key in bintray (re-issue another one) and nuke my gpg pub/priv key set.
02:23 HedgeMage _genuser_: well, if it's been in github, then revoke and renew the api key to be safe, someone may already have it.
02:24 _genuser_ ok, good then, I have revoked and re-issue the api key.
02:24 sargas joined #git
02:24 HedgeMage Also, if you committed your GPG privkey *passphrase* but not the privkey itself, you can change the private key's passphrase without revoking the key
02:24 _genuser_ and I'll throw out that pub/priv key combo also now that the upload on bintray/jcenter is done.
02:24 HedgeMage (assuming you have good key management and know where every damn copy of the privkey is)
02:24 * HedgeMage nods
02:24 _genuser_ hmm, what you mentioned above requires me to go and look at the gpg options again and learn the entire pub/priv mindset again.
02:25 _genuser_ to only do someone with it once and then not have to do it for another 3years. lol. and then I've forgotten the whole thing.
02:25 _genuser_ re: key mgmt - I just saved a copy of each set in a folder named after the use of it. other than that, I don't use them much.
02:27 Orbitrix joined #git
02:27 prg3 joined #git
02:27 Orbitrix joined #git
02:28 Orbitrix joined #git
02:28 _genuser_ guess the most anybody could do with those is change my lib on jcenter and it needs an upload.
02:28 HedgeMage Don't feel bad.  gpg is ill maintained and overused...it's a usability mess.  I do infosec full time, and 1/3 of the people I work with struggle with it.
02:29 HedgeMage I wonder if my "securing the programmer" article from linuxjournal ever got released publicly...I lose track
02:29 HedgeMage I should look when I get less busy
02:30 DolpheenDream joined #git
02:30 derk0pf joined #git
02:31 _genuser_ HedgeMage: I understand the concept that there's a public key and a private key. and using your pub key stuff can be encrypted and only you can decrypt it with your priv key (because of the pub/priv keys' mathematical relationship to each other).
02:32 _genuser_ BUT, the cause for stumbling becomes protecting the private key itself with a password, and which password. And what's being protected by the password and what's being protected by the key. And what the software using the key is doing.
02:32 _genuser_ heh, that's where I always lose track of which software is doing what with it.
02:34 tnecniv joined #git
02:34 a3Dman joined #git
02:37 nicksloan joined #git
02:37 nezZario Hey guys, ... I hate to ask .. But does anyone know of a hosted solution that allows for real pre-receive hooks (even if they are webhooks or whatever)?
02:38 nezZario Quite disappointed I've went through all the trouble of migrating to BitBucket since we also use JIRA, .. seemed like a logical choice, but can't reject pushes with their webhook system.  Can't with AWS CodeCommit either
02:39 nezZario I won't mind hosting it myself (and then mirroring) but it's just managing per-repo+per-branch permissions and keeping up with SSH keys that I have a problem with
02:41 tnecniv joined #git
02:41 afuentes joined #git
02:42 HedgeMage _genuser_: I'm really really slammed right now but if you care to ping me thursday or later I'd have time to explain, or hop on oftc and go to #grokmenot and the infosec geeks I've been teaching can probably give you a decent rundown :)
02:46 _Soto_ joined #git
02:46 Celelibi Can't I make git stash pop merge the modifications?
02:46 bremner _genuser_: there is also #gnupg
02:46 _Soto_ Regarding gitolite, is there a way for me to use it on top of regular github? Not enterprise.
02:46 bremner I'd guess not
02:46 _Soto_ bremner: Was that for me?
02:46 bremner yes
02:49 ilbot3 joined #git
02:49 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: 2.11.0 | 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
02:51 AciD` hey, having a classic setup where I compile my js library in src/ and output the result in the dist/ directory, I want to allow the github users (for instance) to download the dist/*.js files directly without having to setup the environment to compile those files from src/. No, this is annoying from a dev point of view since every single merge is poluted by conflits on those ever-changing files.
02:51 AciD` is there a way to 'ignore' the dist/ files conflits during a merge?
02:52 AciD` or a better repository organisation?
02:52 tnecniv joined #git
02:55 nezZario yes .. ignore them with gitignore and force-add them on releases
02:56 nezZario Still sounds like you're doing something wrong, however ... I would branch from master as like release/v2.0.5, build your library, force-add dist/, commit, and switch back to master (or develop, it should be really)
02:56 pypi_ joined #git
02:56 treehug88 joined #git
02:56 eb0t joined #git
02:58 mischat joined #git
02:58 aavrug joined #git
03:00 venmx joined #git
03:00 yohnnyjoe joined #git
03:00 aavrug joined #git
03:01 chachasmooth joined #git
03:02 mizu_no_oto joined #git
03:07 Goplat joined #git
03:09 railssmith joined #git
03:11 hexagoxel joined #git
03:11 _genuser_ HedgeMage: check, will remember to ping later. :)
03:12 _genuser_ bremner: thanks, I'll pop by.
03:13 sunri5e joined #git
03:14 winny joined #git
03:14 winny is it possible using git-pull to determine if the command fetched (& merged) andy changes from a remote?
03:17 navidr joined #git
03:21 chachasmooth joined #git
03:23 CEnnis91 joined #git
03:23 drbean_ joined #git
03:23 parrotriver winny, git pull && git log ORIG_HEAD..
03:25 Andrew_K joined #git
03:27 winny can I use that in a shell script?
03:29 bmahe joined #git
03:30 Andrew_K joined #git
03:30 alyptik joined #git
03:37 gchristensen joined #git
03:43 athaller_ joined #git
03:44 bjoe2k4 joined #git
03:44 jaguarmagenta joined #git
03:45 chris2 joined #git
03:46 bittyx joined #git
03:46 quizzi joined #git
03:46 serialoverflow joined #git
03:47 Shentino joined #git
03:48 ski7777 joined #git
03:49 rominronin joined #git
03:54 blackwind_123 joined #git
03:56 a_thakur joined #git
03:58 mischat joined #git
03:59 dsantiago joined #git
04:01 jaguarmagenta joined #git
04:06 Rodya_ joined #git
04:14 Vortex35 joined #git
04:16 johnnyfive joined #git
04:17 xall joined #git
04:21 tnecniv joined #git
04:22 phanimahesh joined #git
04:23 jaguarmagenta joined #git
04:24 cqi joined #git
04:24 User458764 joined #git
04:26 jaguarmagenta joined #git
04:29 bmahe joined #git
04:35 dreiss joined #git
04:35 jstimm joined #git
04:35 kulelu88 joined #git
04:36 d^sh joined #git
04:36 fstd joined #git
04:37 durham joined #git
04:41 subhojit777 joined #git
04:43 Anja joined #git
04:45 pypi__ joined #git
04:47 jaguarmagenta joined #git
04:49 xall joined #git
04:50 azerus joined #git
04:50 svm_invictvs joined #git
04:54 zapb_ joined #git
04:54 ayogi joined #git
04:59 mischat joined #git
05:00 venmx joined #git
05:04 strugee joined #git
05:07 User458764 joined #git
05:07 Rodya_ joined #git
05:08 arnsa joined #git
05:09 willlondon joined #git
05:13 lindenle joined #git
05:19 peepsalot joined #git
05:21 rominronin joined #git
05:22 jfr_ joined #git
05:22 watabou joined #git
05:35 InfoTest joined #git
05:38 d0nn1e joined #git
05:41 cheater joined #git
05:42 xberg joined #git
05:46 sbulage joined #git
05:46 QwertyDragon joined #git
05:48 holodoc joined #git
05:49 pks joined #git
05:51 sargas joined #git
05:52 cbegin joined #git
05:54 tnecniv joined #git
05:55 Gallardo joined #git
05:55 fugiman joined #git
06:00 xberg joined #git
06:00 mischat joined #git
06:01 meloc joined #git
06:01 meloc joined #git
06:02 acetakwas joined #git
06:02 Cabanossi joined #git
06:04 SwiftMatt joined #git
06:06 overlord_tm joined #git
06:06 dreiss joined #git
06:08 Rodya_ joined #git
06:08 cheater joined #git
06:08 waterCreature joined #git
06:09 waterCreature hi, i dont want to make the same mistake as git reset --hard.
06:09 waterCreature i made a few commits that i could not push just now.
06:09 waterCreature how do i push just the last commit to the branch?
06:10 waterCreature can i just do git push origin branchname?
06:10 waterCreature oh, btw, i just delete the .git folder
06:10 waterCreature and did git init, added the remote
06:11 kadoban waterCreature: What do you want the branch to look like on the remote? Do you want to nuke all of the old stuff?
06:11 waterCreature yes
06:12 waterCreature i want to nuke everything left on that branch
06:12 _ikke_ git push --force <remote> <branch>
06:13 _genuser_ left #git
06:13 waterCreature i am scared that it would interfere with other branches
06:13 _ikke_ If you just specify a single branch, only that branch is updated
06:14 sbeller joined #git
06:14 bongo joined #git
06:17 waterCreature I see <<<< HEAD stuff, in the git repo
06:17 waterCreature is that normal?
06:17 _ikke_ No
06:17 _ikke_ It means someone ignored conflict and comitted it
06:18 cheater joined #git
06:18 waterCreature i was the only one pushing to that branch
06:18 kadoban Then that someone was almost certainly you
06:19 waterCreature how do I fix that?
06:19 waterCreature even if I did not push it, just committing will create those HEADs?
06:20 waterCreature is it because I did not do git pull?
06:20 waterCreature after redoing the git init
06:21 kadoban waterCreature: They are conflict markers. They happen when you try to merge and git can't figure out how different things should be merged. It creates those markers around the places you're supposed to fix. You're supposed to fix the indicated spots, remove the markers, and stage the changes.
06:21 kadoban waterCreature: Likely you didn't do that, you just staged the files and commited.
06:21 waterCreature is there a gui tool or something to do that?
06:21 waterCreature :D sorry for being so noob
06:21 kadoban There are merge tools you can use
06:22 kadoban meld is one, there's a few
06:22 ToBeCloud joined #git
06:22 kadoban You can choose which one you like and have git run it for you with 'git mergetool', you'd have to look up how to do the config for that though, I never remember.
06:25 waterCreature kadoban, so i am supposed to run it like git mergetool meld?
06:25 waterCreature oh
06:25 waterCreature git meld?
06:26 tnecniv joined #git
06:26 kadoban waterCreature: Look up man git mergetool
06:26 gitinfo waterCreature: the git-mergetool manpage is available at http://jk.gs/git-mergetool.html
06:29 waterCreature i added using git config --global merge.tool meld
06:29 waterCreature then i run git mergetool
06:29 waterCreature and it say no files need merging.
06:29 meauses joined #git
06:29 kadoban That'd be because no files need merging.
06:29 waterCreature but on the repo, those conflict markers are there?
06:30 kadoban waterCreature: The time to do mergetool was when git required your assistance to fix the conflicts. You already fixed the conflicts as far as git is concerned, you told it that the fix was just to commit files including the conflict markers.
06:30 kadoban waterCreature: You probably should just manually edit the files and do whatever to them to make them sane.
06:31 kadoban You might be able to manually run meld on the file(s), I don't know what options are needed or not though.
06:32 venmx joined #git
06:32 waterCreature ahh, Thanks I did not realize it affected my local files as well.
06:33 mseyne joined #git
06:34 strugee joined #git
06:37 derk0pf joined #git
06:38 dreamseeker_ joined #git
06:38 davidosomething joined #git
06:39 _28_ria joined #git
06:40 davidosomething joined #git
06:40 govg joined #git
06:40 oskarkv_ joined #git
06:41 msonntag joined #git
06:41 rath joined #git
06:41 rath left #git
06:42 tnecniv joined #git
06:43 ThomasLocke joined #git
06:44 rath joined #git
06:45 rath left #git
06:46 PtxDK_ joined #git
06:47 qws-user-1228 joined #git
06:47 overlord_tm joined #git
06:48 OOPMan joined #git
06:52 overlordtm joined #git
06:52 zeroed joined #git
06:53 rominronin joined #git
06:55 qhp_ joined #git
06:55 SwiftMatt joined #git
06:56 Raging_Hog joined #git
06:58 LucaTM joined #git
06:58 TomyLobo joined #git
07:00 dermoth joined #git
07:01 mischat joined #git
07:03 mkilivan joined #git
07:06 venmx joined #git
07:06 Vinnie_win_s joined #git
07:11 _cyril_ joined #git
07:11 zeroed joined #git
07:11 zeroed joined #git
07:12 bocaneri joined #git
07:12 sbeller joined #git
07:14 cheater joined #git
07:14 derk0pf joined #git
07:14 lindenle joined #git
07:14 eb0t joined #git
07:15 chele joined #git
07:15 mkilivan joined #git
07:21 mindriot101 joined #git
07:22 PCatinean joined #git
07:22 overlord_tm joined #git
07:23 nidr0x joined #git
07:24 cdown joined #git
07:25 mindriot101 joined #git
07:25 zeroed joined #git
07:25 zeroed joined #git
07:26 freimatz joined #git
07:27 seishun joined #git
07:27 Vinnie_win joined #git
07:29 overlord_tm joined #git
07:29 xall joined #git
07:31 nezZario How do branch permission plug ins work?
07:31 nezZario Or, - - is it possible to add branch level permissions with hooks alone?
07:31 kadoban nezZario: Not exactly sure what you mean in particular, but sounds like they'd be implemented with hooks.
07:31 kadoban nezZario: That's what gitolite does for instance
07:31 thiago plugins?
07:31 thiago Git doesn't have plugins
07:32 nezZario Referring to gitolite , specifically
07:32 kadoban Then yeah, SSH magic and hooks I believe.
07:33 maret joined #git
07:33 jnavila joined #git
07:34 Raging_Hog joined #git
07:35 lb Oo read the docs if you care? http://gitolite.com/gitolite/concepts.html
07:37 Alienpruts joined #git
07:39 _28_ria joined #git
07:40 t-mart joined #git
07:40 hhee joined #git
07:44 enleeten joined #git
07:44 hfp_work joined #git
07:47 PaulCapestany joined #git
07:49 elect joined #git
07:50 spacelord_ joined #git
07:50 Kartagis left #git
07:51 overlord_tm joined #git
07:52 nezZario So I still don't understand one simple thing..  Gitolite is a wrapper around the git-xxxx-pack commands that are ran by git over ssh during push/pull/etc?
07:53 JeroenT joined #git
07:53 ojacobson Yep, plus some access control.
07:53 nezZario It's that right?  I'm just trying to figure out how gitolite digs it's claws in and actually puts itself in front of git
07:53 ojacobson And a set of git hooks.
07:53 nezZario Ah OK
07:53 ojacobson man githooks
07:53 gitinfo the githooks manpage is available at http://jk.gs/githooks.html
07:54 Raging_Hog joined #git
07:54 ojacobson Most of what it does is hooks; the access control is (by default) done using command="" pinning in ssh
07:54 kadoban nezZario: Using the command= stuff in the authorized_keys file, which SSH lets you do, and hooks.
07:55 iliv joined #git
07:55 _28_ria joined #git
07:56 Sceorem joined #git
07:57 TomyWork joined #git
07:57 Balliad joined #git
07:59 overlord_tm joined #git
07:59 mindriot101 joined #git
07:59 Balliad joined #git
08:02 jozwior joined #git
08:02 cbreak ssh does authentication and gitolite enforcement
08:02 cbreak gitolite's script does authorization
08:03 cbreak and the hooks do part of the write authentication afaik
08:03 cbreak authorization I mean :)
08:04 mkilivan joined #git
08:04 roelmonnens joined #git
08:04 nezZario Lol...  Wait...
08:05 nezZario Does it assign a magic env variable for each key (using command=) and then check that in the hook?
08:06 nezZario I was literally just looking at how to accomplish what gitolite does in a simple manner and that was what I came up with
08:06 cdown_ joined #git
08:06 ojacobson magic command line arg, but, basically.
08:06 nezZario I'll dig around and find out myself...  I do appreciate the help
08:06 ojacobson command="path/to/gitolite-shell username" plus some options
08:07 ojacobson sitaram's here (and in #gitolite)
08:08 ojacobson https://github.com/sitaramc/gitolite/blob/master/src/triggers/post-compile/ssh-authkeys#L127-L141 is the relevant bit that generates authorized_keys entries
08:09 nezZario I see,..  No that's brilliantly simple
08:10 qhp joined #git
08:10 ojacobson gitolite is, bar none, my favourite bit of perl software - and I use apt
08:10 qws-user-1228 joined #git
08:11 nezZario Long story short I am just frustrated with third party hosted repos and thinking of writing up some home-cooked solutions at this point
08:12 ojacobson gitolite's great if what you need is _git_
08:12 ojacobson p much The Right Thing
08:12 ojacobson if you want workflow tooling, bitbucket server (formerly "Stash") and gitlab are both available in self-hosted form
08:12 _UniFreak joined #git
08:12 ojacobson for less than the infinite number of dollars gihub enterprise costs :)
08:12 nezZario Between three very large hosted solutions none of them offer the simplest of simple ability to have custom pre receive hooks
08:12 dminuoso ojacobson: stash is ridiculously expensive for medium/large sized teams though.
08:13 ojacobson True. Atlassian knows where to apply the squeeze.
08:13 alcohol stash is dead.
08:13 ojacobson I mean, Github have them beat by miles, but
08:13 osse nezZario: gitlab does
08:13 alcohol when it comes to interface, github > gitlab > bitbucket > vsts (this is like looking at a childs drawing)
08:14 selckin gitlab is impressively easy to install too for such a complex beast
08:14 dminuoso The one thing great about Atlassian is confluence. Period.
08:14 dminuoso But now I feel like it's not worth introducing Jira/Bitbucket for it.
08:14 ojacobson the one great thing is understanding your users and picking the tool that meets their needs
08:14 ojacobson that could well be gitolite :)
08:14 belak dminuoso: I may be biased, but once Jira is set up, it's pretty amazing
08:14 sitaram 13:42:02  <       ojacobson> gitolite's great if what you need is _git_
08:15 ojacobson mornin'
08:15 dminuoso belak: Jira is not bad don't get me wrong, but it all feels really heavy and cumbersome to set up.
08:15 sitaram ojacobson: or, as I like to call it: Social Coding.  Without the 'Social' :)
08:15 sitaram ojacobson: morning :)
08:15 ojacobson my pull request is an email, and so can you
08:15 belak dminuoso: Setting up most Atlassian products is a huge pain... but once they're set up, they're pretty nice
08:15 a_thakur joined #git
08:15 belak That being said, I work there, so I'm a bit biased
08:15 ojacobson (I'm looking a lot harder at divesting from US services in the last, uh, two weeks. For some reason.)
08:15 alcohol the cloud solutions they offer are pretty nice
08:16 dminuoso belak: We're considering migrating to gitlab currently from our Jira/Stash setup - but Confluence is probably going to stay.
08:16 alcohol ojacobson: well, dont move to the UK :p
08:16 dminuoso Nothing beats Confluence. :)
08:16 jnavila joined #git
08:16 nezZario Well initially I had just a plain instance hosting plain git...  Then I looked into git lab, didn't care for it...  Then landed on BitBucket hosted and currently using that
08:16 alcohol ojacobson: also i believe atlassian is aussie based?
08:16 ojacobson Atlassian's main office is in SF these days
08:16 alcohol or is that jetbrains?
08:16 ojacobson I don't mind buying US products, but I want the data to live here.
08:17 nezZario Right.. And we use Confluence and JIRA extensively so we're already paying Atlassian quite a bit
08:17 belak Atlassian's main engineering office is still Sydney. There are two SF offices, one engineering and one marketing...
08:17 belak And there's on in Austin as well
08:17 sbeller joined #git
08:17 belak There are a few other small offices scattered around as well
08:17 nezZario We have a 7 man dev team and a 5 man marketing by  team
08:17 dminuoso belak: Honestly we've had too many issues setting up Bamboo - and when you end with this huge hetereogenous stack of Jira, Stash, Confluence and Bamboo and a gazillion of settings and plugins to make them all work together - it's too much administrational overhead.
08:17 belak As far as I know, all our servers are in the US, but I can't guarantee that
08:17 belak All the BB servers are currently in the US
08:18 belak dminuoso: that's fair. Bamboo is pretty painful.
08:18 nezZario So GitLab would be like 300 bucks a year and have to host it ourselves
08:18 selckin gitlab is free if you host yourself
08:18 ojacobson nezZario: you can host it yourself for free, no? It's largely open-source.
08:19 belak Gitlab has been slow and cumbersome for me in the past
08:19 dminuoso belak: But I have to hand it to you, the user interface is still great. :)
08:19 belak I usually end up using gogs for personal stuff I want to self-host
08:19 Darcidride joined #git
08:19 selckin atlassian guy saying something is slow, lol
08:19 nezZario Huh?  Maybe I'm Missing something...  I thought GitLab was $$ for more than X users
08:19 dminuoso belak: And did I say that Confluence was your best product?
08:19 dminuoso :-)
08:19 ogny joined #git
08:19 nezZario Regardless it's almost embarrassing at this point ...
08:20 shashin joined #git
08:20 belak selckin: yeahhhh
08:21 belak I work on BB Cloud... quite a bit of work goes into trying to keep it fast... I can't really speak for other products though
08:21 nezZario Didn't know but BitBucket was sending emails everytime I changed repo permissions so I spammed everyone's inbox with BitBucket emails trying to get their silly permissions to setup only to figure out their web hooks have no ability to  reject a push
08:21 jost joined #git
08:21 irqq joined #git
08:22 nezZario Oh you work for Atlassian?   I'm sorry :)
08:22 dminuoso Don't be. They make Confluence.
08:22 * dminuoso is a fan
08:23 dminuoso If only they had confluence shirts and coffee mugs
08:23 belak I like it. There's quire a few cool back-end challenges here.
08:24 nezZario Well regardless..  Point being after all of this,  looks like I'm going to have to tell my team to update all of their repo remotes again
08:24 cqi joined #git
08:24 ojacobson dminuoso: at one point in the very dim and distant past, a buddy and I got Bamboo shirts by conclusively demonstrating a fairly major bug in Bamboo
08:24 ojacobson (we're talking around ten years ago no)
08:25 belak I haven't seen any new confluence shirts
08:25 belak But that doesn't mean they don't exist
08:25 ojacobson secret unix protip: do not rewrite files in place, write a new file and rename() it over the original. You'll save your users a bunch of hassle if, for example, the disk turns out to be full and you can't actually complete the write.
08:25 rominronin joined #git
08:25 ojacobson (The bug was that, when the disk was full, changing a build configuration would instead delete it)
08:25 cqi joined #git
08:25 SwiftMatt joined #git
08:26 selckin general good practice to end up with corrupted files
08:26 selckin *not
08:26 belak Rewriting files in place is just asking for race conditions too
08:26 alcohol but.. but... sed -i is so hawt
08:26 dminuoso ojacobson: Or use flock if your system supports it.
08:26 ojacobson Yeah. To their credit, the bamboo team at the time figured it out instantly and fixed it :)
08:26 selckin overwriting in place is just the same
08:26 ojacobson dminuoso: not sufficient
08:26 nezZario ojacobson: honest question... If you know - - is there a honest reason why bb's web hooks don't have a simple way to return like a 401 instead of 2xx to reject a push?
08:26 ojacobson flock stops someone else from doing a conflicting write to the same file, but doesn't guarantee that your write will succeed
08:27 ojacobson nezZario: I don't _know_, but I can speculate - they're probably asynchronous, to avoid slowing down pushes when someone's hook is slow or down
08:27 cqi joined #git
08:27 dminuoso ojacobson: That would depend on the lock I guess.
08:27 belak Hooks are definitely handled by our background task system
08:27 ojacobson there is no lock that will prevent someone from filling or disconnecting the disk.
08:27 ojacobson No amount of software-side prep makes in-place writes safe.
08:27 nezZario Honestly that was my own speculation
08:28 dminuoso ojacobson: Fair enough.
08:28 selckin you can even get kill -9d during the write, renaming always better
08:28 ojacobson oomkill!
08:28 ojacobson It's fine for trivial stuff, but it's good practice to use rename for atomic replacement instead: the kernel promises that rename calls are uninterruptible even if the machine powers off in the middle
08:29 ojacobson (you'll either get the original file back or the new file, barring unrecoverable disk corruption)
08:29 belak We have a few types of hooks... there are pre-hooks (which do double checks on repo permissions, etc) and then post-hooks (most of which just fire off background tasks... here we update pull requests and send webhooks, among other things)
08:29 nezZario But id be happy for a +500ms push if I could stop these lazy devs from writing stupid commit  messages, or junk/passwords into the repositories,  and so on
08:29 ojacobson You can't fix all of that with hooks, sadly :)
08:30 belak nezZario: it has been something we've considered... but for a cloud service, it's next to impossible to scale that properly
08:30 ojacobson Mechanical message style enforcement will lead the least cooperative to figuring out the shortest acceptable message and using i
08:30 snufft joined #git
08:30 belak And it essentially hands off performance of our own app to external services
08:30 ojacobson (At one place they instated "javadoc comments must contain at least one sentence of description." There were, when I left, a number of methods whose docs were /** . */.
08:30 adac joined #git
08:30 nezZario Well we so distribute client side hooks too
08:30 belak Which, if how people abuse our service is any indication, would be an awful idea
08:32 belak As an example, we have a huge number of people who clone repos at the top of the hour, half hour, etc... our biggest hour is midnight UTC... presumably from people who just set cron jobs to run at midnight UTC.
08:32 belak Which is a super lazy time to run things...
08:33 xberg joined #git
08:33 belak Anyway, that's the big complication with writing a cloud VCS service at the moment... you have to build for the thundering herd.
08:33 nezZario In all reality I don't really even "badly want" pre receive hooks..  Just the fact that no service offers them as a hosted solution
08:33 mindriot101 joined #git
08:33 ojacobson Yeah, it's tricky, because a lot of hooks are equivalent to "let me run code on your server"
08:33 nezZario It bugs me badly
08:33 belak Yeah, it's not all that realistic for a hosted service to add something like that
08:33 ojacobson even if you isolate it as far as you can, it can still run slowly and consume memory
08:33 belak Yep
08:34 ojacobson You can build a business around that (my own employer _has_) but a code-hosting service is not that business
08:34 nezZario Well,... I gotta disagree...
08:34 belak nezZario: oh?
08:36 nezZario I don't mean a code hosting solution...  (although Amazon pulled it off, it's pretty terrible at best, speaking of lambda)...  But just simply calling a remote http endpoint before writing things..
08:37 nezZario If times out (default accept | default reject)
08:37 selckin feels like something that should be done in front of it
08:37 ojacobson Eh. You could, in principle, allow pushes to hang pending an outgoing http request. You run into problems with timeouts and with various HTTP libs' assumptions about threading, though.
08:37 ojacobson If BB Cloud is java, then they're presumably stuck with the Servlet API's one-request-per-thread model, or its broke-ass resumption model, for example
08:37 belak You would pretty much have to default to "default reject" if there are hooks installed.
08:37 nowhere_man joined #git
08:37 belak ojacobson: it's python
08:37 ojacobson and ssh is nearly impossible to make asynchronous without rewriting whole swaths of it
08:37 ojacobson oh, neato
08:38 Vangelis joined #git
08:38 ojacobson openssh, at least, uses one unix process per client, which gets expensive if those processes can hang around for a long time and you have a lot of them. That's a money problem, not an impossible problem, but money problems matter
08:39 belak multiple processes actually simplifies quite a few things
08:39 SwiftMatt joined #git
08:39 ojacobson (I'm not sure I'd _trust_ an sshd that hosted multiple users in the same unix process, but that's irrelevant for something like bitbucket, where the only ssh user is `git` most of the time)
08:39 belak We use a custom ssh server right now
08:39 * ojacobson nods.
08:40 belak That was rewritten over the past 2 years and just released that earlier this year
08:40 ojacobson I dunno what Github Cloud uses, but Enterprise uses openssh :)
08:40 belak github cloud uses libssh
08:41 rominron_ joined #git
08:41 xberg_ joined #git
08:42 brillenfux joined #git
08:43 belak Or they did when I last checke
08:43 belak *checked
08:43 ojacobson Neato
08:43 ojacobson You're a wealth of info, thanks for being you :)
08:44 belak If you run with ssh -vvv, it'll show all protocol level requests, and reporting the version of the server is one of them
08:44 belak Actually, wait, it's a part of the handshake
08:44 ojacobson Anyways, to nezZario: my day job _runs_ a service that will stall git pushes based on http calls to third-party sites, and let me tell you, I dread my weeks on call
08:44 belak Ah, any time. :)
08:44 mischat joined #git
08:44 moei joined #git
08:45 OOPMan joined #git
08:45 ojacobson (it's the heroku git server, which, among other things, runs your entire build in pre-receive, and does git submodule setup before that)
08:45 mischat joined #git
08:45 ojacobson (and, for the record, we use https://godoc.org/golang.org/x/crypto/ssh)
08:45 belak Hey, so do we!
08:45 belak I recently got a fix in there for dropbear. :)
08:45 nezZario Interesting...  Heroku git server?
08:46 ojacobson Did you get a bunch of nasty tickets when the last DH vuln dropped asking for support for EC keys?
08:46 ojacobson ECC*
08:46 mar77i could someone assist me in fixing my github repository? warning: remote HEAD refers to nonexistent ref, unable to checkout.
08:46 ojacobson nezZario: 'git push heroku master' is one of the ways to deploy an app to heroku
08:46 belak We've had requests for ecdsa and ed25519 for a while
08:46 mar77i I tried googling and didn't come to an understanding what exactly might be wrong with my repo
08:46 ojacobson we build your code (run packaging steps, etc) and then slap it out to containers to serve web requests from
08:46 belak It made me happy when support was added
08:46 nezZario I just want something that a year from now I don't need to tell my team to update remotes...... Again
08:47 ojacobson Not useful to you, just an example of "yes, you could do that, but it's a serious problem" :)
08:48 ojacobson belak: what I think is hilarious, to bring this full circle, is that Heroku's git server is the _exact_ same architecture as gitolite
08:48 iliv joined #git
08:48 nezZario Which is why I am considering just self hosting....  Heck I could do anything needed, get cheaper builds, and probably just mirror it to a SaaS for all the pretty interfaces
08:48 ojacobson a front-end wrapper doing access control, which invokes real git(1), and then some git hooks to handle git lifecycle stuff
08:48 syednoorali joined #git
08:49 belak Yeah...
08:49 belak We were using a patched version of openssh, along with a replacement shell and an authorized keys script
08:49 belak But, as it turns out, that has performance issues
08:49 ojacobson patched to accelerate authorized_keys lookup with millions of keys, I would guess?
08:49 belak yeah
08:49 ojacobson because last time I looked, the authorized_keys script interface was missing some key features
08:49 ojacobson like "give me the key matching this fingerprint"
08:49 belak haha
08:50 Andrew_K joined #git
08:50 belak It's improved a ton
08:50 ojacobson yay
08:50 belak Like, it would work just fine now
08:50 belak You can specify arguments to the script now
08:50 belak Our patch just passed the key via stdin
08:52 encod3 joined #git
08:52 oskarkv joined #git
08:53 hhee joined #git
08:55 solf1re joined #git
08:56 buzzLightBeer joined #git
08:58 Charliechin joined #git
08:59 ejnahc joined #git
09:01 willlondon joined #git
09:06 kristofferR joined #git
09:07 mikecmpbll joined #git
09:08 xberg joined #git
09:08 netj joined #git
09:11 Rodya_ joined #git
09:11 xberg joined #git
09:11 chachasmooth joined #git
09:11 filcab joined #git
09:12 arnsa joined #git
09:14 shashin joined #git
09:14 kurkale6ka joined #git
09:16 Marbug_ left #git
09:20 kanzure joined #git
09:21 User458764 joined #git
09:23 vrlx joined #git
09:23 oskarkv joined #git
09:25 garethdaine joined #git
09:26 zeroed joined #git
09:26 zeroed joined #git
09:28 chachasmooth joined #git
09:30 mischat joined #git
09:31 as_g5pw joined #git
09:31 Darcidride joined #git
09:32 venmx joined #git
09:32 dviola joined #git
09:36 tnecniv joined #git
09:36 User458764 joined #git
09:37 Sonderblade joined #git
09:37 sitaram ojacobson: any URLs? (I could search but if you have any handy...)
09:37 User458764 joined #git
09:37 sitaram (about Heroku's git server architecture)
09:37 Darcidride_ joined #git
09:37 clmsy joined #git
09:38 sitaram ojacobson: although, if you mean the authentication side, I copied it from gitosis, which copied it from some ssh-based SVN access control tool... there's really no *other* way to do it if you're using ssh
09:38 a_thakur joined #git
09:38 muelli joined #git
09:39 hahuang65 joined #git
09:39 muelli can I configure git s.t. format-patch will be run with a certain To address? Would smth like git config format-patch.to foo@bar  work?
09:41 bongjovi joined #git
09:42 chachasmooth joined #git
09:43 Darcidride joined #git
09:44 Serpent7776 joined #git
09:45 Ferdroid joined #git
09:46 Ariadeno joined #git
09:46 OOPManZA joined #git
09:46 Peng_ joined #git
09:47 ahmedelgabri joined #git
09:47 roelmonn_ joined #git
09:48 inflames joined #git
09:49 _ikke_ muelli: git config format.from 'e-mail'
09:49 _ikke_ see man git config
09:49 gitinfo the git-config manpage is available at http://jk.gs/git-config.html
09:49 texinwien joined #git
09:51 chll_ joined #git
09:51 rnsanchez joined #git
09:54 navidr joined #git
09:54 OOPMan joined #git
09:54 muelli _ikke_: cool. thanks.
09:55 marcogmonteiro joined #git
09:55 donught joined #git
09:59 iliv joined #git
10:06 mablae joined #git
10:07 nanga joined #git
10:07 King_Hual joined #git
10:08 User458764 joined #git
10:09 Qria joined #git
10:09 OOPManZA joined #git
10:10 rooftopjoe joined #git
10:10 Qria left #git
10:11 rooftopjoe what happens if i "git add foo", modify foo and then "git commit" without adding foo again? will foo not be committed at all or will the initial change be committed?
10:11 selckin 'git diff --cached' will be commited
10:12 Rodya_ joined #git
10:12 rooftopjoe is the cache deleted on commit?
10:12 Timvde Which is the initial change that you have added
10:12 Snugglebash joined #git
10:12 Timvde Nothing in your working directory will be deleted
10:12 Timvde If you do a git status before committing, you will see that you have both staged and unstaged changes to foo
10:12 rooftopjoe no, the cache (which i expect is in .git). is that deleted once i commit?
10:12 osse rooftopjoe: there is no cache as such. git add does all the heavy lifting, commit just creates the commit object
10:13 rooftopjoe so is there a way to see all the added versions i did before a commit?
10:13 Timvde rooftopjoe: yes, "git diff --cached", as has been said
10:14 Timvde Or "git diff --staged", which is a synonym, and might match your expectations more closely
10:14 paul424 joined #git
10:14 rooftopjoe and are those kept in .git forever?
10:14 rooftopjoe even after the commit?
10:15 Timvde What is "those"?
10:15 rooftopjoe all the versions i added in between commits
10:15 rooftopjoe you said they are all recorded
10:15 Timvde Everything that has ever been committed is in .git, everything that has not been committed is not in .git
10:15 rooftopjoe of course
10:15 rooftopjoe but you said git caches everything i add even if i don't commit it
10:16 rooftopjoe "git add foo", edit foo, "git add foo", edit foo, "git add foo", edit foo
10:16 overlord_tm joined #git
10:16 rooftopjoe [12:13] <rooftopjoe> so is there a way to see all the added versions i did before a commit?
10:16 rooftopjoe [12:13] <Timvde> rooftopjoe: yes, "git diff --cached", as has been said
10:16 Ferdroid joined #git
10:16 Timvde Ah, I misinterpreted that
10:16 rooftopjoe oh, i understand what you meant then
10:16 smithbone_ joined #git
10:16 Timvde As far as I know, no, you cannot see older added versions
10:17 hahuang61 joined #git
10:17 milki_ joined #git
10:17 jmpp_ joined #git
10:17 jmpp_ joined #git
10:17 Timvde You should commit if you want to store the state of a file
10:17 unbalanced joined #git
10:17 mrkake joined #git
10:17 rooftopjoe i know. but i wanted to see if "git add" could work as a way to trick git into thinking in terms of files
10:17 hanthings_ joined #git
10:18 Timvde I am not sure what you mean by that
10:18 rooftopjoe well, a commit is more like a changeset or something like that
10:18 fahadash_ joined #git
10:18 iliv joined #git
10:18 chrivers joined #git
10:18 Timvde Not really, a commit is a set of blobs which can be used to restore the state of your working directory
10:19 tm512` joined #git
10:19 rooftopjoe yes, i'm talking about the structure of those blobs
10:19 kadoban_ joined #git
10:19 tobiasvl you can think of a commit as a kind of changeset if you want, but what does it mean to trick git into thinking in terms of files?
10:19 chrivers hey everybody. I'm working on cleaning up the history of a multi-project repository, in preparation for splitting it with --subdirectory-filter for filter-branch.
10:19 k_sze[work] joined #git
10:19 snowinferno joined #git
10:19 chrivers How can I use filter-branch to undo the effect of a "git rm"?
10:19 Remramm joined #git
10:19 PityDaFool joined #git
10:20 ndim joined #git
10:20 chrivers since I'm sanitizing unclean history, I would really like to make it appear as though the git rm never happened, so it is easier to reason about the reamining history
10:20 rooftopjoe argh, so annoying that the freenode web client can't be used to filter out all this netsplit joins/parts
10:20 PerlJam joined #git
10:20 davidstrauss joined #git
10:20 rooftopjoe i will return in a few minutes when the noise is over
10:20 CussBot joined #git
10:20 dilfridge joined #git
10:20 gigq joined #git
10:20 atomi_ joined #git
10:20 jlewis joined #git
10:20 msm joined #git
10:20 marcaddeo joined #git
10:20 Esya joined #git
10:21 jbondhus joined #git
10:21 kozie joined #git
10:21 __name__ joined #git
10:21 Forkk joined #git
10:21 * tobiasvl is blissfully unaware of such noise
10:21 enckse joined #git
10:21 acln_ joined #git
10:21 tobiasvl do we ban people with unstable connections temporarily here?
10:21 Fiouz joined #git
10:21 devhost_ joined #git
10:21 Lowell joined #git
10:21 holgersson joined #git
10:21 daishan joined #git
10:21 systeem joined #git
10:21 nanga joined #git
10:21 navidr joined #git
10:21 Lunatrius joined #git
10:21 FabTG joined #git
10:22 CheckDavid joined #git
10:22 Peng I haven't been banned yet
10:22 nixdork joined #git
10:22 llamapixel joined #git
10:22 tobiasvl that can be arranged
10:22 acln_ joined #git
10:22 cjohnson_ joined #git
10:22 mroth joined #git
10:22 phz__ joined #git
10:22 paperziggurat joined #git
10:23 chrivers As far as I understand it, filter-branch with --tree-filter cannot undo removes
10:23 Peng ...only one of my connections is unreliable :<
10:23 boshhead_ joined #git
10:23 chrivers but could index-filter be used for that?
10:23 thecomedian joined #git
10:23 spr0tsuki joined #git
10:23 twisted` joined #git
10:23 nanga joined #git
10:23 mehdi___ joined #git
10:23 miklcct joined #git
10:24 rideh joined #git
10:24 d^sh joined #git
10:24 d0nn1e joined #git
10:24 kutenai joined #git
10:24 sim642 joined #git
10:24 hide joined #git
10:24 TooLmaN joined #git
10:24 jaguarmagenta joined #git
10:24 n-st joined #git
10:24 angular_mike_ joined #git
10:24 timothy joined #git
10:24 vrishab joined #git
10:24 BrAsS_mOnKeY joined #git
10:24 synx joined #git
10:24 mukmuk joined #git
10:24 devster31 joined #git
10:24 cheater joined #git
10:24 boxrick1 joined #git
10:24 phroa joined #git
10:24 nedbat joined #git
10:24 hexagoxel joined #git
10:24 APK joined #git
10:24 PioneerAxon joined #git
10:25 valkyr2e joined #git
10:25 zifnab06 joined #git
10:25 musca joined #git
10:25 chrisml joined #git
10:25 HalfEatenPie joined #git
10:25 bremner` joined #git
10:25 tabakhase__ joined #git
10:25 codebam joined #git
10:25 tobiasvl chrivers: yes, don't see why not
10:25 LeMike joined #git
10:26 maqr joined #git
10:26 bremner` joined #git
10:26 SirCmpwn joined #git
10:26 Korvin joined #git
10:26 Korvin joined #git
10:26 idr0p joined #git
10:26 uidzer0 joined #git
10:26 chrivers tobiasvl: I'm still learning about filter-branch. What would be the right command to run in index-filter? git reset?
10:26 andschwa joined #git
10:27 chrivers on the right path, I mean
10:27 ensign joined #git
10:27 Anja joined #git
10:27 Khisanth joined #git
10:27 hussam joined #git
10:28 tobiasvl chrivers: what do you want to do? can't you just use !interactive_rebase?
10:28 gitinfo chrivers: 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'.
10:28 User458764 joined #git
10:28 tuor joined #git
10:28 chrivers tobiasvl: That would be my go-to tool, but in this case I'm cleaning up history from colleagues who are, shall we say, new to git
10:28 gmh joined #git
10:28 chrivers so one commit changes 17 files, removes 5, and adds 4
10:28 User458764 joined #git
10:29 chrivers I need to undo the 5 deletes, but keep the rest of the commits
10:29 fiete joined #git
10:29 chrivers s/commits/changes/
10:29 lungaro joined #git
10:29 Guest96457 joined #git
10:29 coasterz joined #git
10:29 User458764 joined #git
10:29 storrgie joined #git
10:30 chrivers I'm trying to give everything cleaned up, so we can all have a fresh start, with a repository per project, instead of a big tangles mess
10:30 Ferdroid joined #git
10:30 Ferdroid joined #git
10:30 waterCreature joined #git
10:30 chrivers and the history is too important to give up on
10:30 de-vri-es joined #git
10:30 unclechu joined #git
10:30 ExeciN joined #git
10:30 Eryn_1983_FL joined #git
10:31 tribly joined #git
10:31 mischat joined #git
10:31 iliv joined #git
10:31 jason237 joined #git
10:31 errr joined #git
10:31 cagedwisdom joined #git
10:31 jesk joined #git
10:31 richardlitt joined #git
10:31 ispn joined #git
10:31 davix[matrix] joined #git
10:31 oskarkv joined #git
10:32 moop joined #git
10:32 ribasushi joined #git
10:32 asd5a joined #git
10:32 perrier-jouet joined #git
10:32 Stravy joined #git
10:32 doppo joined #git
10:32 markelite joined #git
10:32 FilipNortic joined #git
10:33 chrivers tobiasvl: does it sound like a hopeless project? :)
10:33 Vampire0 joined #git
10:35 InfoTest joined #git
10:36 ShoaibJahejo joined #git
10:36 fhackdroid joined #git
10:36 Eryn_1983_FL joined #git
10:36 ExeciN joined #git
10:36 fhackdroid joined #git
10:36 ShoaibJahejo joined #git
10:36 ShoaibJahejo joined #git
10:37 Raging_Hog joined #git
10:38 kedare joined #git
10:39 [0xAA] joined #git
10:39 alcohol joined #git
10:39 venmx joined #git
10:40 zacts joined #git
10:40 InfoTest joined #git
10:41 ghostlight joined #git
10:42 mischat joined #git
10:43 irqq joined #git
10:44 ExeciN joined #git
10:44 hhee joined #git
10:45 achadwick joined #git
10:46 hhee guys, tried git clone to web server directory, but got fatal: working tree '/var/www/html/example.org/html' already exists.
10:47 hhee if i wrote git status
10:47 hhee in webserver dir /var/www/html/example.org/html - fatal: Not a git repository (or any of the parent directories): .git
10:47 acln joined #git
10:48 _ikke_ 2 problems: 1. git does not clone into a directory that already contains files, 2. You should not use git to manage a live website
10:48 cqi joined #git
10:49 paramourne joined #git
10:49 hhee _ikke_, clean dir, without any files
10:49 hhee _ikke_, it just test project. what can i use for this?
10:49 [0xAA] joined #git
10:50 cqi joined #git
10:50 Lope joined #git
10:50 hhee _ikke_, which approach do i need to?
10:50 courrier joined #git
10:51 Lope how can I get an output like `git log -p filename` but only lines that contained 'foo' before or after the edit?
10:51 _ikke_ My approach of choice is git-archive
10:51 _ikke_ hhee: ^
10:51 Raging_Hog joined #git
10:51 derk0pf joined #git
10:51 hhee _ikke_, got it, thx
10:52 Lope hhee: your nick is genius.
10:52 Lope I thought ikke was laughing.
10:53 hhee Lope :)
10:54 xall joined #git
10:54 chrivers Can I undo a "git rm" in history, with filter-branch? If so, any pointers? :)
10:56 NimeshNeema joined #git
10:58 donught joined #git
11:00 hhee got tar: html: Cannot utime: Operation not permitted; tar: html: Cannot change mode to rwxrwxr-x: Operation not permitted; tar: Exiting with failure status due to previous errors;
11:00 hhee work as root, using sudo
11:00 sbulage joined #git
11:00 maret joined #git
11:01 Virox joined #git
11:01 chrivers hhee: "cannot utime" seems to suggest a problem with file date/times. is this on some non-standard filesystem?
11:01 chrivers maybe vfat, fuse-ftp, sshfs or somesuch?
11:01 dminuoso hhee: You can use  --no-overwrite-dir
11:02 mikecmpbll joined #git
11:02 dminuoso hhee: It usually happens when you try to unpack into a directory hierarchy into an existing hierarchy to which you have write permissions but are not an owner.
11:02 davimore joined #git
11:03 hhee chrivers, i'll check it. thx. dminuoso, thx a lot. dminuoso, i'll check it too. thx a guys. you are awesome :)
11:04 rossome joined #git
11:04 chrivers hhee: in this context, an actual permission problem is probably more likely. you should check out the other suggestions first :)
11:05 hhee chrivers, got it :)
11:05 rooftopjoe left #git
11:05 hhee you are guys - the best teachers :)
11:07 willlondon joined #git
11:07 tvw joined #git
11:07 daishan joined #git
11:08 waterCreature left #git
11:11 phz__ left #git
11:11 romerocesar joined #git
11:13 Rodya_ joined #git
11:14 ToBeCloud joined #git
11:14 Charliechin joined #git
11:15 arnsa joined #git
11:15 mischat joined #git
11:15 Darcidride joined #git
11:17 kenansulayman joined #git
11:17 epist3me joined #git
11:17 gniourf joined #git
11:19 xberg joined #git
11:21 mischat_ joined #git
11:30 dimi1947 joined #git
11:31 paladinn joined #git
11:31 paladinn joined #git
11:31 daishan joined #git
11:32 cyan__ joined #git
11:33 mischat joined #git
11:35 stamina joined #git
11:37 xberg_ joined #git
11:38 revoltingPeasant joined #git
11:41 Caelum joined #git
11:41 midnightmagic joined #git
11:41 robotroll joined #git
11:42 dsdeiz joined #git
11:42 jordila joined #git
11:43 barteks2x is there a limit on how far git reflog can go? (I don't currently need it for anything, just want to know, it doesn't show some of older parts)
11:43 tobiasvl barteks2x: yes. the default is that it's pruned after 90 days
11:44 barteks2x what git does it? I don't remember doing any command that does it
11:44 barteks2x *when
11:44 tobiasvl man git gc
11:44 gitinfo the git-gc manpage is available at http://jk.gs/git-gc.html
11:44 tobiasvl man git reflog
11:44 gitinfo the git-reflog manpage is available at http://jk.gs/git-reflog.html
11:45 * jordila finds some git commands syntax difficult to keep in mind ... lol
11:45 _ikke_ jordila: which ones?
11:45 jordila When 'git reset --hard HEAD~1 ' in mind ... ¿is there any other syntactically simpler command leading to the same result ?
11:45 barteks2x I don't remember using git reflog expire of git gc
11:45 barteks2x *of->or
11:46 tobiasvl barteks2x: no, but like I said and like the man pages say, pruning automatically happens after 90 days (by default)
11:46 barteks2x but when exactl? When I run some random command?
11:46 _ikke_ jordila: if the HEAD~1 is the hard part, then there is no easy way except finding the commit hash manually
11:46 barteks2x it's not like git is running in background so it can do that
11:47 User458764 joined #git
11:51 Guest57513 joined #git
11:52 stamina joined #git
11:54 _ikke_ barteks2x: git runs it automatically during some command calls
11:56 _ikke_ "When there are approximately more than this many loose objects in the repository, git gc --auto will pack them. Some Porcelain commands use this command to perform a light-weight garbage collection from time to time. The default value is 6700. Setting this to 0 disables it."
11:56 a_thakur joined #git
11:58 leehambley joined #git
11:59 Raging_Hog joined #git
12:00 JeroenT joined #git
12:00 pypi__ joined #git
12:01 nowhere_man joined #git
12:01 Snugglebash joined #git
12:02 rominronin joined #git
12:03 DolpheenDream joined #git
12:03 InfoTest joined #git
12:05 cheater joined #git
12:05 justanotheruser joined #git
12:05 paladinn joined #git
12:06 [0xAA] joined #git
12:06 qws-user-1229 joined #git
12:06 zero85 hey guys, i pushed a crap commit to github, how can i remove the commit and the changes in it?
12:07 xberg joined #git
12:07 justanotheruser joined #git
12:08 Peng Is it just slightly embarrassing, or a lot embarrassing, or does it contain your credit card number or something?
12:09 overlord_tm joined #git
12:10 mischat joined #git
12:10 zero85 its just that i wanted to pull from upstream and that didnt work but i dont noticed, than i pushed my additions to my fork and now i have everything as i wanted but there is a commit that should fix it but it didnt, when i want to pull request there is this crappy commit i dont want to pull request ^^
12:10 Masber joined #git
12:11 zero85 as i can see, its the last commit
12:11 Rodya_ joined #git
12:11 justanotheruser joined #git
12:15 noodle joined #git
12:18 shinanyenzo joined #git
12:19 xall joined #git
12:23 mischat joined #git
12:23 willlondon joined #git
12:24 jaguarmagenta joined #git
12:26 al-damiri joined #git
12:26 utrack joined #git
12:30 _ikke_ zero85: basically force pushing the previous commit
12:31 _ikke_ that moves the branch back
12:31 wrouesnel joined #git
12:31 zero85 ok
12:34 zeroed joined #git
12:34 zeroed joined #git
12:37 Myris joined #git
12:37 WayToDoor joined #git
12:37 Jackneill joined #git
12:37 ace_33 joined #git
12:38 imprfcto joined #git
12:38 biberu joined #git
12:38 imprfcto left #git
12:39 re1 joined #git
12:40 venmx joined #git
12:41 arnsa joined #git
12:42 willlondon joined #git
12:43 tvw joined #git
12:44 www-bukolay-net joined #git
12:48 gniourf joined #git
12:49 dreiss joined #git
12:49 noecc joined #git
12:49 cdown joined #git
12:49 rominronin joined #git
12:50 Rodya_ joined #git
12:51 ncthom91 joined #git
12:52 Darcidride joined #git
12:55 zeroed joined #git
12:55 zeroed joined #git
12:55 maret joined #git
12:57 mindriot_ joined #git
12:58 a_thakur joined #git
13:03 cdown joined #git
13:03 Flowinho joined #git
13:04 boombatower joined #git
13:04 chris2 joined #git
13:05 cdown_ joined #git
13:05 DolpheenDream joined #git
13:06 mischat joined #git
13:07 johnmilton joined #git
13:08 sdothum joined #git
13:08 nilg joined #git
13:08 venmx joined #git
13:10 anonymuse joined #git
13:10 mischat_ joined #git
13:12 FederationOfNULL joined #git
13:13 Peng_ joined #git
13:13 mischat joined #git
13:15 roelmonnens joined #git
13:16 deadnull joined #git
13:17 venmx joined #git
13:18 nowhere_man joined #git
13:19 venmx joined #git
13:21 biberu left #git
13:21 hjbq joined #git
13:24 sbulage joined #git
13:24 mindriot101 joined #git
13:27 iliv joined #git
13:27 synthroid joined #git
13:27 mikecmpb_ joined #git
13:28 User458764 joined #git
13:28 systemovich joined #git
13:32 meauses joined #git
13:33 jordila joined #git
13:33 willlondon joined #git
13:35 venmx joined #git
13:36 mikecmpbll joined #git
13:36 ace_33 joined #git
13:39 mischat joined #git
13:40 dunpeal_ joined #git
13:41 rbern joined #git
13:41 byte512 joined #git
13:41 User458764 joined #git
13:42 Karazhan joined #git
13:43 Karazhan joined #git
13:45 dunpeal joined #git
13:46 mkilivan joined #git
13:46 jeffreylevesque joined #git
13:47 leeN joined #git
13:49 xall joined #git
13:49 mindriot101 joined #git
13:50 mkilivan joined #git
13:51 Gloomy joined #git
13:51 Siegfried joined #git
13:51 WizJin joined #git
13:53 ShekharReddy joined #git
13:53 arand__ joined #git
13:53 mischat joined #git
13:56 User458764 joined #git
13:56 benzoo joined #git
13:57 cdown joined #git
13:58 mischat_ joined #git
14:01 xberg_ joined #git
14:01 encod3 joined #git
14:02 johnny56 joined #git
14:03 rorro joined #git
14:03 fuzzmz joined #git
14:03 jimi_ joined #git
14:04 encod3 joined #git
14:04 archaic joined #git
14:04 Nilesh_ joined #git
14:05 daynaskully joined #git
14:05 shinnya joined #git
14:05 encod3 joined #git
14:05 fahadash_ joined #git
14:06 cdg joined #git
14:06 fahadash joined #git
14:07 cdg joined #git
14:08 encod3_ joined #git
14:09 oxsyn joined #git
14:10 mischat joined #git
14:10 willlondon joined #git
14:11 BrianBlaze420 joined #git
14:11 benzoo left #git
14:12 soc42 joined #git
14:12 encod3 joined #git
14:13 Gsham joined #git
14:14 ExoUNX joined #git
14:15 xberg joined #git
14:15 bsanford joined #git
14:15 encod3_ joined #git
14:16 mischat joined #git
14:17 JeroenT joined #git
14:19 encod3 joined #git
14:20 solf1re joined #git
14:21 encod3_ joined #git
14:25 Urbany joined #git
14:25 jaguarmagenta joined #git
14:26 nickabbey joined #git
14:27 jimi_sanchez joined #git
14:27 Enekk joined #git
14:27 mischat joined #git
14:28 User458764 joined #git
14:31 blackwind_123 joined #git
14:33 jaguarmagenta joined #git
14:34 todd_dsm joined #git
14:35 daynaskully joined #git
14:36 p4trix joined #git
14:36 ace_33 joined #git
14:36 cdg joined #git
14:37 jimi_sanchez joined #git
14:40 swift1 joined #git
14:41 xall i find that certain config/package stuff pollutes what i really want to look at in git log diffs. i was thinking a good solution might be to only show .js files in a git log diff. is there a way to do this? (i don't want gitigore)
14:42 selckin git log -p -- '*.js'
14:43 cdown joined #git
14:45 xall selckin: thanks. i think i'll make a shell function to easily pass args
14:46 GodGinrai joined #git
14:46 Dougie187 joined #git
14:47 xall does double dash (--) usually mean, we're passing params rather than options now?
14:47 quackgyver joined #git
14:47 vktec Yes, usually
14:47 selckin files instead of sha/ref
14:48 cdg joined #git
14:51 porsche944 joined #git
14:52 jaguarmagenta joined #git
14:54 AaronMT joined #git
14:55 porsche944 joined #git
14:56 encod3 joined #git
14:56 porsche944 so cherry-picking creates a new commit on the other branch?
14:57 Snugglebash joined #git
14:57 porsche944 I thought the whole point was to apply that commit to the other branch
14:57 tobiasvl cherry-pick copies a commit and applies it to another branch, yes
14:57 porsche944 by creating a new commit
14:57 porsche944 or did I screw something up
14:57 tobiasvl sure. what do you mean?
14:57 selckin yes
14:58 porsche944 I just thought the commit wwould have the same hash
14:58 tobiasvl oh. that would be impossible
14:58 porsche944 oh
14:58 tobiasvl the hash is derived from what parents the commit has
14:58 porsche944 I've seen commmits say contained in branch x, y
14:58 Flowinho joined #git
14:58 porsche944 oh right
14:58 porsche944 its the checksum
14:58 tobiasvl yes
14:58 jost joined #git
14:59 sbulage joined #git
14:59 anonymuse joined #git
15:00 Gloomy joined #git
15:00 Noldorin joined #git
15:00 e14 joined #git
15:00 arooni joined #git
15:03 tvw joined #git
15:05 crose joined #git
15:05 dunpeal left #git
15:06 axl_ joined #git
15:06 snowkidind joined #git
15:07 jimi_sanchez joined #git
15:07 govg joined #git
15:09 GavinMagnus joined #git
15:09 nicksloan joined #git
15:10 jimi joined #git
15:10 mischat joined #git
15:10 jost joined #git
15:12 alhariel joined #git
15:12 xall joined #git
15:13 thebope joined #git
15:14 IrishPrime joined #git
15:14 Blkt joined #git
15:15 hhee joined #git
15:15 ash_workz joined #git
15:17 ski7777 joined #git
15:17 lindenle joined #git
15:18 albel727 joined #git
15:19 computer2000 joined #git
15:21 computer2000 Hi, I have a completely new codebase locally, different than the one on my repo - is there an option to force-overwrite-push my new stuff onto the repo, also deleting orphaned files on the repo?
15:21 bremner` why not just make a new repo?
15:21 computer2000 If I had access to the repo as an admin (like on github) I'd delete the repo and make a new one but it's hosted somewhere else and I can't delete / make new one
15:22 computer2000 bremner`: ^
15:22 bremner` "deleting orphaned files" sounds like some kind of web deployment scenario
15:23 computer2000 anyone?
15:24 bremner` well, of course you can force push, but the other part of your question is harder to understand.
15:25 bremner` the "normal" case is pushing to a bare repository.
15:25 jost joined #git
15:25 grapheti_ joined #git
15:25 dfaught joined #git
15:26 jimi_sanchez joined #git
15:26 FederationOfNULL joined #git
15:27 nikivi joined #git
15:27 jbeeze joined #git
15:28 sangy joined #git
15:28 sssilver joined #git
15:29 jimi joined #git
15:29 Random832 or he didn't know how to delete the files in the commit
15:30 bremner` perhaps. In any case, more explanation is needed.
15:31 jimi joined #git
15:33 Gsham joined #git
15:33 jimi joined #git
15:33 adac joined #git
15:36 alhariel joined #git
15:36 jost joined #git
15:37 waterCreature joined #git
15:37 lss8_ joined #git
15:37 waterCreature hi, just to be sure, if i do "git add profiles/."
15:37 waterCreature it will add all the files and folders inside the profiles folder, right?
15:37 FilipNortic_ joined #git
15:38 lss8__ joined #git
15:38 selckin mkdir test/; touch test/1; touch test/2; git add test/.; git status
15:38 inflames joined #git
15:38 Guest96457 left #git
15:38 NeverDie_ joined #git
15:39 karstensrage joined #git
15:42 waterCreature selckin, I see this in the terminal http://i.imgur.com/u4TLRzq.png
15:42 waterCreature the green color means that these files will be sent to the repo, if I do commit and push?
15:42 navidr joined #git
15:43 selckin yes, staging area
15:43 raspado joined #git
15:43 selckin it also has the header "changes to be comiitted"
15:44 a_thakur joined #git
15:46 AndroUser2 joined #git
15:47 elsevero joined #git
15:47 sweatsuit joined #git
15:48 JanC_ joined #git
15:48 aarobc joined #git
15:50 hashpuppy joined #git
15:50 User458764 joined #git
15:51 Lauven joined #git
15:51 e14 joined #git
15:51 _Cyclone_ joined #git
15:52 Enekk joined #git
15:52 Enekk joined #git
15:52 Gsham joined #git
15:53 FederationOfNULL joined #git
15:53 hashpuppy i have a coworker with a fork of project1.  we develop in the develop branch.  he will merge upstream/develop into his develop branch and then add a new feature in his develop branch and submit a pull request.  sometimes the pull requests will have "merge branch 'develop' of company:project1 into develop".  and we will see commits already merged in his pull
15:53 hashpuppy request, making it hard to figure out what's been changed.  what should i tell him he's doing wrong?  would asking him to always create a new branch for each pull request help solve this issue?
15:54 _ikke_ yes, he should not merge anything into develop
15:54 GodGinrai joined #git
15:54 _ikke_ or commit anything in there
15:54 Snugglebash joined #git
15:55 tnecniv joined #git
15:55 hashpuppy i'm confused about the first statement.  should he always be branching off of upstream/develop?  and never keep his own 'develop' up to date w/ upstream/develop?
15:56 _ikke_ hashpuppy: He can keep his local develop up-to-date, but never by merging
15:56 NeverDie joined #git
15:56 hashpuppy what's the alternative?
15:56 hashpuppy rebase?
15:56 _ikke_ If he's not comitting on develop, there should not be anything to merge in develop anyway
15:56 hashpuppy true
15:56 a_thakur joined #git
15:57 hashpuppy i always do git checkout develop && git fetch upstream && git merge upstream/develop
15:57 _ikke_ So he can just fast-forward (ff-only) merge his develop branch to upstream/develop
15:57 hashpuppy i've probably been doing this wrong myself
15:57 hashpuppy git merge --ff upstream/develop?
15:57 _ikke_ yes
15:57 LionsMane joined #git
15:57 hashpuppy ok.  thanks
15:57 GodGinrai hashpuppy: yea, if his develop is supposed to mirror yours, then he should never be developing in it
15:57 _ikke_ (fast-forward is the default, but with --ff-only, it will abort if it cannot fast-forward)
15:58 hashpuppy thanks
15:58 GodGinrai if he sets upstream/develop as the upstream of his develop branch, he should just be able to pull to update it
15:59 hashpuppy yeah, i''m wondering if that's recommended
15:59 GodGinrai hashpuppy: ^
15:59 hashpuppy probably easier
15:59 subhojit777 joined #git
16:00 troulouliou_div2 joined #git
16:01 _ikke_ hashpuppy: He can set git config branch.develop.mergeOptions '--ff-only'
16:01 _ikke_ That way he can never do a non-ff merge in that branch
16:01 hashpuppy i wonder if that's better than just having develop be upstream/develop?
16:01 PaulCapestany joined #git
16:01 tvw joined #git
16:01 hashpuppy using upstream/develop seems a little odd, but does work
16:02 mischat joined #git
16:02 PCatinean joined #git
16:02 jnavila joined #git
16:03 JanC joined #git
16:03 jimi_ joined #git
16:04 _ikke_ Sure, there is nothing wrong with just using upstream/develop
16:04 _ikke_ as a base for other branches
16:04 roelmonn_ joined #git
16:05 Guest99883 left #git
16:06 jost joined #git
16:09 hashpuppy if develop already points to refs/heads/develop, how can i have it point to upstream/develop instead?  git branch -u develop upstream/develop?
16:11 hashpuppy remotes/upstream/develop
16:11 hashpuppy hmm... maybe not.  lol
16:11 ojdo joined #git
16:13 hashpuppy nm
16:15 thiago joined #git
16:15 aarobc joined #git
16:16 meauses joined #git
16:18 mischat joined #git
16:19 chardan joined #git
16:23 crose joined #git
16:26 venmx joined #git
16:27 emg joined #git
16:29 ejb joined #git
16:29 meauses joined #git
16:30 eivarv joined #git
16:30 mischat joined #git
16:31 jccn joined #git
16:31 tang^ joined #git
16:32 derk0pf joined #git
16:35 iliv joined #git
16:37 tyreld joined #git
16:37 shgysk8zer0 joined #git
16:38 osse hashpuppy: what are you trying to do?
16:38 osse refs/heads/X is the full name of a local branch commonly referred to as X.
16:38 Akimb joined #git
16:39 svm_invictvs joined #git
16:39 osse oh i am late to the party. excusez-moi
16:40 lss8 joined #git
16:40 FederationOfNULL joined #git
16:40 lss8 joined #git
16:44 modin joined #git
16:48 lss8_ joined #git
16:48 xall joined #git
16:48 anonymuse joined #git
16:51 ams__ joined #git
16:53 jaguarmagenta joined #git
16:53 mischat joined #git
16:54 GavinMagnus left #git
16:55 charlesr joined #git
16:56 ejb joined #git
16:57 venmx joined #git
16:58 User458764 joined #git
16:58 ochorocho__ joined #git
17:00 hanthings joined #git
17:00 Virox urghhh I checked out an individual file before committing changes and it's not showing in git frelog
17:00 Virox *reflog
17:00 Virox Im guessing there's no way of getting that back?
17:01 GodGinrai Virox: so you had changes in a file.  Then checked out that file.  And now you want the changes back?
17:01 Virox Yeah
17:01 GodGinrai You don't happen to still have your text editor open, do you?
17:01 Virox I do... I tried hitting undo/redo to try and get something backl
17:02 GodGinrai I know with vim that you can undo the load of the checked out file to get back your changes
17:02 GodGinrai I don't know about other text editors
17:02 GodGinrai other than that, you are probably up a creek without a paddle
17:02 Virox I'm using xcode. It was only like 4 lines so I can redo but I was curious for future reference if git covered this kind of stupidity
17:03 GodGinrai I don't believe it does
17:04 Virox Thanks anyway :)
17:04 _ikke_ Virox: have you git added the file before?
17:04 _ikke_ those changes I mean
17:04 Virox Nah
17:04 _ikke_ Ok, then git does not know about them
17:05 seishun joined #git
17:06 mischat joined #git
17:06 nfk joined #git
17:09 SilenceDogood joined #git
17:09 SilenceDogood Hey there folks
17:09 gitinfo SilenceDogood: 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.
17:10 SilenceDogood I am having problems trying to go back to a previous commit
17:10 modin Any good suggestion on how to visualise a git workflow? Something like http://learngitbranching.js.org/?NODEMO or https://onlywei.github.io/explain-git-with-d3/#zen but I like to be able to save/share/collaborate and easier edit my commands than to start over
17:10 GodGinrai SilenceDogood: what's the problem?
17:10 SilenceDogood I said "git checkout [commit number]"
17:10 SilenceDogood then deleted master branch and recreated master branch
17:11 SilenceDogood that should, I thought, point master to where head currently is
17:11 SilenceDogood But when I look at my git log, I still see all the commits I want to get rid of
17:11 SilenceDogood btw this is on a project using github
17:12 GodGinrai SilenceDogood: use `git-reset` instead of `git-checkout`
17:12 SilenceDogood ok
17:12 SilenceDogood so its just "git reset [commit number]"?
17:12 GodGinrai yea
17:12 SilenceDogood k thanks i'll try that
17:12 GodGinrai that will reset HEAD to that commit number
17:13 SilenceDogood that was it. thanks!
17:13 crose joined #git
17:13 jccn joined #git
17:14 madewokherd joined #git
17:15 GodGinrai np
17:15 SilenceDogood That did not work
17:15 GodGinrai hrm?
17:16 ok91 joined #git
17:16 Puffball_ joined #git
17:17 dpower joined #git
17:17 SilenceDogood saying git log still shows the commits i want to get rid of
17:18 GodGinrai hrm
17:18 durham joined #git
17:18 xberg joined #git
17:18 GodGinrai can you throw up this process in a gist so I can see what you are doing?  Show the log before and after the reset
17:18 rnsanchez joined #git
17:19 SilenceDogood oh wait, no i've got it
17:19 bluepixel joined #git
17:19 OnBrokenWings joined #git
17:22 GodGinrai aha :)
17:22 Snugglebash joined #git
17:22 Jackneilll joined #git
17:22 SilenceDogood Don't know why I was seeing that because I tried the process again and it worked this time. I have the changes I'm discarding safely stored in an alternate branch while my master is back in the past before I made those changes ready to work :)
17:22 SilenceDogood had to do a git push --force
17:23 GodGinrai yup, I've done that before
17:23 GodGinrai "oh crap, I committed to master!" `git branch newbranch; git reset HEAD~1`
17:24 Random832 you just reset the new branch GodGinrai
17:24 GodGinrai Random832: nope.  git-branch doesn't checkout the new branch
17:24 Random832 oh oops
17:25 venmx joined #git
17:26 SilenceDogood GodGinrai well actually this was bigger than that.  but thanks
17:26 GodGinrai np
17:26 robotroll joined #git
17:27 GodGinrai SilenceDogood: just make sure that whoever is pulling from the same repo is aware that their history is changing
17:27 GodGinrai SilenceDogood: otherwise, you will have a mess on your hands
17:28 SilenceDogood only me
17:28 flavius joined #git
17:28 GodGinrai perfect
17:28 sangy joined #git
17:28 ehynds joined #git
17:39 milki_ left #git
17:39 milki_ joined #git
17:39 milki_ left #git
17:39 milki joined #git
17:42 _Cyclone_ joined #git
17:42 nidr0x joined #git
17:45 Gsham joined #git
17:45 Ryanar joined #git
17:46 lindenle joined #git
17:46 e14 joined #git
17:46 watabou joined #git
17:49 Derperperd joined #git
17:52 SilenceDogood left #git
17:53 b_rare joined #git
17:54 nikivi joined #git
17:55 Jackneillll joined #git
17:57 Akimb joined #git
17:58 tnecniv joined #git
17:58 derk0pf joined #git
17:59 tabakhase__ left #git
18:00 nickabbey joined #git
18:00 tabakhase joined #git
18:01 afuentes joined #git
18:02 aarobc joined #git
18:03 oskarkv joined #git
18:04 cqi joined #git
18:04 grapheti__ joined #git
18:05 muelli joined #git
18:06 cagmz joined #git
18:06 Gloomy joined #git
18:06 HM joined #git
18:06 Spec-Chum joined #git
18:09 ohcibi joined #git
18:15 marcogmonteiro joined #git
18:16 mikecmpbll joined #git
18:17 garethdaine joined #git
18:18 mindriot101 joined #git
18:21 spacelord_ joined #git
18:21 emg joined #git
18:21 shout-user27 joined #git
18:22 derk0pf joined #git
18:22 shout-user27 hi folks - so I'm trying to truly understand git rebase... I've got pretty developed feature branch for a proof-of-concept, and I want to rebase on our latest code
18:22 Akimbas joined #git
18:23 MajorPayne joined #git
18:23 grapheti_ joined #git
18:23 cjohnson git rebase master  will find the point where your branch diverged from master (assuming your branch is off of master), rewind all your branch's commits, apply all the latest from master, and replay your commits one by one
18:23 _Cyclone_ joined #git
18:23 cjohnson bringing your branch 100% up to date with master. note that this will change all of your commits so they will get a new hash and you will have to force push this branch
18:23 cjohnson if anybody else is working on it that will cause them problems when they pull this branch
18:24 shout-user27 yeah... that's interesting - because when I rebase on the branch we call "qa" (basically master - latest) which this was created from, I get merge conflicts. and I need to resolve these 20 times
18:24 shout-user27 it includes a lot of files that I don't think should have conflicts
18:24 shout-user27 I'm the only one using this feature branch
18:25 leonarth joined #git
18:25 cjohnson right, so the way I work around that problem is: never merge qa to master, and always branch features from master
18:25 cjohnson when I'm testing, I merge feature to qa
18:25 cjohnson when it's ready, merge it to master directly, dont' just merge qa to master
18:25 shout-user27 so I followed a bit of advice at http://stackoverflow.com/a/7241759/4200039 and tried onto
18:25 shout-user27 --onto
18:26 wget joined #git
18:26 shout-user27 and if I'm on my feature branch and do git rebase qa (the original branch)
18:26 shout-user27 all my work is destroyed
18:26 shout-user27 but if I create a new branch off of qa and then rebase onto my feature branch, all my work is at the top
18:26 shout-user27 and no merge conflicts
18:26 [0xAA] joined #git
18:26 shout-user27 that seems opposite what I'd expect
18:26 cjohnson Yeah that works too in some cases but I still suggest not branching off qa
18:27 Vorisi joined #git
18:27 cjohnson because you are including other people's testing changes in your branch
18:27 cjohnson which may not be ready for production
18:27 shout-user27 cjohnson: I think our naming might be different than yours
18:27 shout-user27 we're basically using gitflow
18:27 cjohnson well presumably whatever qa is is for testing yes?
18:27 shout-user27 we ship a weekly release
18:27 shout-user27 yep, qa is for testing - but all our latest is in qa
18:27 cjohnson if I'm working on branch foo and you're on branch bar and I merge my changes to qa for testing and you branch bar from qa
18:27 cjohnson now your branch includes my work
18:28 cjohnson which may not be ready to ship at the same time
18:28 cjohnson and also makes the commit history harder to track
18:28 shout-user27 yep, but all that work will be ready by the end of the week
18:28 shout-user27 we only merge things when we know what we want to release
18:28 cjohnson Sure but that's why you're getting mucked up commit history
18:28 cjohnson :)
18:28 cjohnson your branch includes my commits and vice versa, and when you merge, they conflict
18:28 shout-user27 no, I'm just trying to sync my old proof-of-concept to the latest code
18:29 aarobc joined #git
18:29 cjohnson Your branch has never ever merged to qa?
18:29 cjohnson and nobody else has either?
18:29 shout-user27 no, definitely not
18:29 shout-user27 it's a proof-of-concept
18:29 cjohnson maybe they're doing it then. they merge foo to qa, then they want to work on foo some more, so they rebase foo
18:29 shout-user27 we have feature branches which we sync with the latest code periodically ... it's not an uncommon pattern
18:29 cjohnson now foo contains a different set of commits than those merged to qa
18:30 cjohnson basically you cannot rebase after you merge, or not cleanly anyway
18:30 cjohnson once those commits have escaped that branch you have to stop rebasing that branch
18:30 shout-user27 none of my commits have ever "escaped" my branch
18:30 cjohnson Right, I'm saying maybe the other users are doing this
18:30 cjohnson which is why you see their conflicts
18:31 shout-user27 I'm not following... I was able to get all my commits into the codebase with zero conflicts by following a certain --onto approach
18:31 shout-user27 a git merge shows conflicts in 5 files, which is plausible
18:31 shout-user27 a git rebase shows conflicts in thousands of files, because we made a whitespace change
18:32 cjohnson I'm not sure, I'd have to know more about the history
18:32 shout-user27 (or at least probably)
18:32 cjohnson these are just the things that I have learned will definitely cause conflicts when merging and rebasing features
18:32 shout-user27 yeah, this is a proprietary codebase, I should come up with a small demonstration example
18:32 cjohnson I virtually never have conflicts, even for giant whitespace changes
18:32 invisbl_ joined #git
18:32 shout-user27 we almost never have conflicts either
18:32 cjohnson and that's because I: branch from master, merge feature to qa to test, then merge feature to master once it's ready
18:32 cjohnson rebase feature from master regularly
18:33 cjohnson but stop rebasing once it's merged to master
18:33 shout-user27 that was our prior approach, caused constant merge conflicts
18:33 GodGinrai shout-user27: --onto is used to exclude commits from an unrelated branch
18:33 cjohnson it shouldn't if you do it the way I described exactly
18:33 cjohnson it also results in PRs not being polluted with changes from other branches
18:34 GodGinrai shout-user27: for example, if you had branch "poc" based off of "qa" which was based off of "master" and you wanted to rebase onto master, you would have to use onto to exclude the commits in "qa" that weren't in "master"
18:34 ochorocho__ joined #git
18:34 BSaboia joined #git
18:35 dm03 joined #git
18:35 treehug88 joined #git
18:36 ejb joined #git
18:36 shout-user27 cjohnson: the trouble with that approach is that qa and master get out of sync
18:37 cjohnson we never merge qa to master so it's fine
18:37 shout-user27 cjohnson: so I merged this commit into qa, someone else is working in a similar area, branches from master, and also merges a commit into qa... boom, merge conflict
18:37 cjohnson about every few weeks I start looking at qa to see if it's got a ton of branches that aren't merged to master
18:37 cjohnson if it's crazy out of sync, I blow away qa and make it fresh from master. but that rarely happens
18:37 cjohnson we don't see too many abandoned features
18:37 shout-user27 whereas if everyone is working off the same base (qa), a merge conflict is impossible
18:38 shout-user27 how many people do you have working on the repo?
18:38 cjohnson 5, with pretty much constnat activity
18:38 cjohnson the only time there are merge conflicts are when they are legit
18:38 cjohnson like two features conflicted
18:38 shout-user27 but what you describe is definitely an ingredient for merge conflicts - you see what I mean about two people merging separate commits into qa, right?
18:38 cjohnson we never get any non-legit conflicts using this approach
18:39 zacts joined #git
18:39 shout-user27 are you familiar with gitflow? https://www.atlassian.com/git/tutorials/comparing-workflows/
18:39 ochorocho__ joined #git
18:40 cjohnson yes, not intimately but we read through it when deciding on our approach
18:40 cjohnson it wasn't working out perfectly as is
18:40 cjohnson I think we kept getting ugly PRs full of unrelated changes
18:40 shout-user27 what I'm calling qa is what gitflow calls develop https://www.atlassian.com/git/tutorials/comparing-workflows/gitflow-workflow
18:40 cjohnson whether that's git's fault or github's
18:40 shout-user27 well, how we prevent getting other people's commits is release planning
18:41 shout-user27 and feature branches
18:41 nikivi joined #git
18:41 shout-user27 the only stuff that is merged into develop is that which is planned for the next deploy
18:41 shout-user27 we also squash everything
18:41 cjohnson yeah but then you still can't do a PR for each branch right?
18:41 cjohnson you have to PR the whole release
18:41 cjohnson that's what we prefer, so devs can code review each other's features
18:42 User458764 joined #git
18:42 shout-user27 we hold pull requests out until they are ready to release but they're all feature branches, usually with only a commit or two
18:42 shout-user27 when we ship a deploy, we start merging these feature branches, which are generally code-reviewed
18:43 cjohnson got it. we prefer a PR to happen as soon as the branch is ready. then when we ship it only requires a small once-over
18:43 shout-user27 there's a a chance that a merge conflict might arise during that... we then have the person rebase onto develop, and it gets code reviewed again
18:43 cjohnson otherwise every dev has to be on deck for every release
18:43 cjohnson to justify their features
18:43 shout-user27 we make pull requests whenever the branch is ready, but we don't merge until we get the thumbs up that it will be in the release
18:44 [0xAA] joined #git
18:44 cjohnson and the PR is strictly limited to only that branch's commits? When you go to the "files changed" tab you never see anybody else's changes from any other branch at all?
18:44 shout-user27 so generally rebase isn't a big deal when you have a commit or two for something small
18:44 cjohnson We weren't getting that behavior with git flow
18:44 shout-user27 we never see anyone else's commits
18:44 shout-user27 but you have to do release planning
18:44 shout-user27 you're only going to get other people's commits if they've merged
18:45 shout-user27 but generally all the merging happens the day after our release
18:45 shout-user27 everything has been code-reviewed prior to that
18:45 cjohnson got it
18:45 shout-user27 note that we also use a lot of feature flags
18:45 cagedwisdom joined #git
18:45 shout-user27 we find those very helpful for preparing big features, it lets our product managers and customers test code a lot
18:45 shout-user27 we can also deploy our feature branches into isolated instances with a slack message
18:46 Exagone313 joined #git
18:47 shout-user27 we just did a big release so we destroyed most of our feature flags now tho
18:47 jimi_ joined #git
18:47 shout-user27 which is nice - too many feature flags is a nightmare, have to make sure to go back and clean them out
18:47 cjohnson what is a feature flag?
18:47 rena_ joined #git
18:48 shout-user27 http://stackoverflow.com/questions/7707383/what-is-a-feature-flag
18:48 shout-user27 or also https://en.wikipedia.org/wiki/Feature_toggle
18:48 cjohnson Oh I see
18:48 cjohnson thought it was a git concept
18:48 e14 joined #git
18:49 shout-user27 it's related in that there's different approaches to developing long-running features... feature flags is an alternative to long-lived feature branches
18:49 cjohnson I think we move a lot faster than you do, sounds like
18:50 Mobutils joined #git
18:51 sangy joined #git
18:52 shout-user27 probably - we're past the move fast and break things stage
18:52 tnecniv joined #git
18:52 shout-user27 but feature flags are an agile technique designed around moving fast
18:53 dreiss joined #git
18:53 nickabbey joined #git
18:53 cjohnson pft we never break things! ;P
18:54 jaguarmagenta joined #git
18:54 shout-user27 how do you handle long-running features? something that's in development for several months?
18:54 [0xAA] joined #git
18:54 anonymuse joined #git
18:55 derk0pf joined #git
18:56 cjohnson Nothing is ever so big, we break it down into smaller pieces in the project stage
18:56 cjohnson longest feature we had over the last 2 years was 4 weeks or so
18:56 cjohnson and it fit into the above workflow just fine
18:56 Sasazuka joined #git
18:56 TomyLobo joined #git
18:57 cjohnson 3 weeks heavy development so I just kept rebasing master, and then in the last week I merged it to qa
18:57 cjohnson 2 days before the end I rebased that branch once more from master, blew away and recreated qa so as not to get merge conflicts, merged to qa, had a thorough final run-through testing
18:59 MrWoohoo joined #git
19:04 sangy joined #git
19:04 ochorocho__ joined #git
19:04 Salander27 joined #git
19:05 talkJS joined #git
19:05 Snugglebash joined #git
19:06 talkJS I have two branches.  develop and a feature branch.  I need to merge my feature branch into develop.   I want to be able to unmerge the merge without causing other people issues.  How can I do this?
19:07 _ikke_ Why do you need to unmerge?
19:07 Gsham joined #git
19:07 talkJS because i need to test the branch, and possibly remove it if it's not approved
19:07 talkJS i want to be able to remove it, easily
19:07 vktec talkJS: Then test in your branch, not master
19:07 talkJS it's develop
19:07 vktec talkJS: Then test in your branch, not develop
19:08 talkJS I want to merge to develop, as that is where the hooks are to auto-deploy on our test environment
19:08 vktec Then it's badly set up.
19:08 vktec Either run the tests locally or get someone to fix the test environment
19:08 cjohnson talkJS: branch develop to develop-your-test
19:08 cjohnson merge your feature to that one
19:08 cjohnson deploy it somewhere
19:08 cjohnson as a hack
19:09 hhee joined #git
19:09 talkJS there's no way to undo a merge without causing issues?
19:09 vktec cjohnson: That won't solve the problem, afaict
19:09 sea-gull_ joined #git
19:09 cjohnson well it would let you test without mucking up develop's history
19:09 vktec talkJS: There probably is, but it's a stupid idea. Don't do it
19:09 vktec Fix the underlying issue
19:09 cjohnson agreed
19:09 talkJS okay
19:09 talkJS thx fellas
19:09 Sonderblade joined #git
19:09 vktec cjohnson: 19:08 <talkJS> I want to merge to develop, as that is where the hooks are to auto-deploy on our test environment
19:10 vktec It has to be called 'develop'
19:10 cjohnson got it
19:12 ochorocho__ joined #git
19:13 qws-user-1228 joined #git
19:13 chardan joined #git
19:13 kerrick joined #git
19:14 bket joined #git
19:15 enckse joined #git
19:15 mozzarella joined #git
19:16 GT4066 joined #git
19:19 finalbeta joined #git
19:20 specious joined #git
19:20 ochorocho__ joined #git
19:22 Puffball joined #git
19:23 meauses joined #git
19:24 aard_ joined #git
19:25 mindriot101 joined #git
19:26 afuentes joined #git
19:26 Jellyg00se joined #git
19:27 Jackneill joined #git
19:28 anonymuse joined #git
19:28 cdown joined #git
19:29 mdw joined #git
19:29 ok92 joined #git
19:29 gmn8719 joined #git
19:29 azerus joined #git
19:30 cdown_ joined #git
19:32 ochorocho__ joined #git
19:34 e14 joined #git
19:34 nowhere_man joined #git
19:35 satifant joined #git
19:37 cheater joined #git
19:38 Sasazuka_ joined #git
19:38 azerus joined #git
19:38 ejb joined #git
19:39 pandeiro joined #git
19:40 synthroid joined #git
19:40 jimi_ joined #git
19:47 relipse joined #git
19:49 azerus joined #git
19:50 relipse how do I check out a tag. I tried git fetch , nothing updated, and then I tried git checkout mytag and ti says error: pathspec 'mytag' ...
19:51 multi_io joined #git
19:52 _ikke_ relipse: does git tag return it?
19:54 urda joined #git
19:54 ochorocho__ joined #git
19:55 urda So I'm git 2.11.0, and I'm getting a "warning: empty strings as pathspecs will be made invalid in upcoming releases. please use . instead if you meant to match all paths" when running "git add -p ." ... which is confusing because I did use the "." as requested. I do NOT get this error when just running "git add -p"
19:55 mischat joined #git
19:56 irqq joined #git
19:56 _ikke_ urda: iirc, that's an older message..
19:56 urda Just started seeing it today _ikke_
19:56 cyan__ joined #git
19:56 urda when up'ing to 2.11
19:56 emg joined #git
19:57 neredsenvy joined #git
19:57 neredsenvy Anyone experienced a problem where git clone -b .. method is incredibly horribly slow but all other operations work fast ?
19:58 Sasazuka joined #git
19:58 rebrec joined #git
19:58 _ikke_ urda: https://github.com/gitster/git/commit/d426430e6ec2a05bf0a4ee88c319dd6072908504
19:58 Gsham joined #git
19:59 _Cyclone_ joined #git
19:59 [0xAA] joined #git
19:59 _ikke_ urda: right, is new in 2.11
19:59 urda interesting, even though I DID have the "." in my command I still got the error
19:59 urda git add -p "."
19:59 urda warning: empty strings as pathspecs will be made invalid in upcoming releases. please use . instead if you meant to match all paths
19:59 urda also fails
20:00 _ikke_ urda: Right. Note that this is just a warning for now
20:00 _ikke_ But let me see if I can get it also
20:00 d0nn1e joined #git
20:00 dmc joined #git
20:01 urda thanks _ikke_ , it just seems to me since I'm already using "." that I shouldn't be seeing the warning, but I could be misunderstanding it :)
20:01 _ikke_ urda: From what I read, you are correct
20:02 Phylock joined #git
20:02 urda D:
20:03 pypi__ joined #git
20:03 cheater joined #git
20:03 davidkrauser joined #git
20:04 sssilver joined #git
20:05 LionsMane joined #git
20:07 _ikke_ urda: Yup, happening to me as well
20:07 relipse _ikke_:  yes git tag returns the tag
20:07 urda neat, I guess I need to hit the mailing list
20:08 seishun joined #git
20:08 azerus joined #git
20:10 Flowinho joined #git
20:11 dermoth joined #git
20:14 pandeiro joined #git
20:14 _ikke_ urda: Note that it's only when using -p
20:14 urda interesting
20:14 _ikke_ not when using plain git add .
20:15 david__ joined #git
20:16 noecc left #git
20:16 ok91 joined #git
20:17 Orion3k joined #git
20:18 axl__ joined #git
20:18 Charliechin joined #git
20:19 Dougie187 left #git
20:19 aidalgol joined #git
20:19 nwkj86 joined #git
20:19 MTecknology left #git
20:20 jccn joined #git
20:21 Dougie187 joined #git
20:21 Charliechin joined #git
20:23 Siegfried joined #git
20:24 cqi joined #git
20:24 Siegfried joined #git
20:26 enleeten joined #git
20:27 Siegfried joined #git
20:27 dsantiago joined #git
20:31 johnmilton joined #git
20:31 nikivi joined #git
20:31 invisbl joined #git
20:32 invisbl joined #git
20:36 finalbeta joined #git
20:36 Dougie187 joined #git
20:41 rolha joined #git
20:41 stamina joined #git
20:42 Dougie187 left #git
20:43 Dougie187 joined #git
20:44 govg joined #git
20:44 finalbeta joined #git
20:44 refried_ joined #git
20:46 anonymuse joined #git
20:47 [0xAA] joined #git
20:48 sangy joined #git
20:48 Remramm joined #git
20:48 Remramm joined #git
20:50 finalbeta joined #git
20:50 Dougie187 joined #git
20:52 sdlnv joined #git
20:53 sdlnv joined #git
20:54 jaguarmagenta joined #git
20:54 finalbeta joined #git
20:55 e14 joined #git
20:56 nickabbey joined #git
21:03 azerus joined #git
21:03 anonymus_ joined #git
21:03 Rodya_ joined #git
21:04 azerus joined #git
21:04 cdg joined #git
21:05 Masber joined #git
21:05 dsdeiz joined #git
21:05 dsdeiz joined #git
21:05 irqq_ joined #git
21:06 qws-user-1229 joined #git
21:09 fmcgeough joined #git
21:09 vuoto joined #git
21:12 Gloomy joined #git
21:12 Puffball_ joined #git
21:12 rena_ joined #git
21:13 Sasazuka joined #git
21:15 dan2k3k4 joined #git
21:16 m0viefreak joined #git
21:18 ejb joined #git
21:18 ahmedelgabri joined #git
21:19 tnecniv joined #git
21:19 hussam joined #git
21:20 joki- joined #git
21:20 e14 joined #git
21:21 railssmith joined #git
21:27 anonymuse joined #git
21:28 roygbiv joined #git
21:28 ghostlight joined #git
21:30 WebDawg joined #git
21:30 jason237 joined #git
21:31 muelli joined #git
21:32 chachasmooth joined #git
21:32 rolha joined #git
21:35 paul424 joined #git
21:38 Synthead joined #git
21:38 patarr_ joined #git
21:39 urda joined #git
21:39 eycsigfy joined #git
21:40 bronson joined #git
21:42 netj joined #git
21:43 crose joined #git
21:44 qhp joined #git
21:45 rena_ joined #git
21:48 WaReZ joined #git
21:49 Gsham joined #git
21:51 svm_invictvs joined #git
21:53 irqq joined #git
21:53 chachasmooth joined #git
21:54 irqq__ joined #git
21:58 irqq_ joined #git
21:59 d10n-work joined #git
21:59 eycsigfy joined #git
22:01 bernardio joined #git
22:01 irqq joined #git
22:01 CJLymar joined #git
22:02 whitby joined #git
22:03 tvw joined #git
22:05 irqq__ joined #git
22:05 rolha joined #git
22:06 irqq_ joined #git
22:06 stomplee joined #git
22:07 govg joined #git
22:07 irqq joined #git
22:08 mindriot101 joined #git
22:08 joshszep joined #git
22:09 elsevero joined #git
22:11 paul424 joined #git
22:11 Peetz0r joined #git
22:12 irqq_ joined #git
22:12 [0xAA] joined #git
22:14 davix[matrix] joined #git
22:15 eycsigfy joined #git
22:15 enleeten joined #git
22:16 chachasmooth joined #git
22:17 irqq joined #git
22:18 rena_ joined #git
22:18 jeffreylevesque joined #git
22:19 irqq___ joined #git
22:19 dreiss joined #git
22:20 rolha joined #git
22:20 irqq_ joined #git
22:22 [0xAA] joined #git
22:22 Limes_ joined #git
22:23 irqq____ joined #git
22:24 mindriot101 joined #git
22:26 M-nated joined #git
22:26 WillianPaiva[m] joined #git
22:26 kineticrootkit[m joined #git
22:26 Tiaan[m] joined #git
22:26 unclechu joined #git
22:26 Remramm joined #git
22:26 patrickr[m] joined #git
22:26 saintaquinas[m] joined #git
22:26 mtso[m] joined #git
22:26 hallogallo[m] joined #git
22:26 kikijiki[m] joined #git
22:26 malo[m] joined #git
22:26 samwise[m] joined #git
22:26 M-shine joined #git
22:26 M-meznak joined #git
22:26 M-mistake joined #git
22:26 Dhanya[m] joined #git
22:26 monomon[m] joined #git
22:26 am2on joined #git
22:26 mei[m] joined #git
22:26 Ineentho[m] joined #git
22:26 warrantyvoid joined #git
22:26 musicmatze[m] joined #git
22:26 aviraldg joined #git
22:26 srkrishna[m] joined #git
22:26 alaeri[m] joined #git
22:26 Carlos[m] joined #git
22:26 timlyo[m] joined #git
22:26 kjsage[m] joined #git
22:26 zdexin[m] joined #git
22:26 M64137[m] joined #git
22:26 eltopo[m] joined #git
22:28 bluepixel joined #git
22:30 Karazhan joined #git
22:31 chachasmooth joined #git
22:34 ghostlight joined #git
22:35 fracting joined #git
22:37 mischat joined #git
22:37 mischat joined #git
22:37 eycsigfy joined #git
22:38 mischat joined #git
22:39 mischat joined #git
22:40 meauses joined #git
22:40 mischat joined #git
22:41 mischat joined #git
22:41 [0xAA] joined #git
22:41 mischat joined #git
22:42 mischat joined #git
22:42 specious joined #git
22:43 emg joined #git
22:43 alex1a joined #git
22:43 mischat joined #git
22:43 shashin joined #git
22:44 mischat joined #git
22:44 d5sx43 joined #git
22:44 Paul__ joined #git
22:46 t-mart_ joined #git
22:47 Paul__ left #git
22:48 PaulM888 joined #git
22:48 chachasmooth joined #git
22:49 _Cyclone_ joined #git
22:49 GodGinrai joined #git
22:49 mizu_no_oto joined #git
22:50 enckse joined #git
22:55 enckse joined #git
22:55 jaguarmagenta joined #git
22:56 Vinnie_win_k joined #git
22:56 Charliechin joined #git
22:57 ochorocho__ joined #git
22:58 FLHerne joined #git
22:59 storrgie joined #git
23:01 fmeerkoetter joined #git
23:01 FLHerne What would be a better way to express "the currently staged (or committed) state"?
23:02 Charliechin joined #git
23:02 FLHerne We'd like to improve the string "Show changes between the staging area (if not clean or last commit on current branch, otherwise) and the last commit on the selected branch.", because it's unreadable
23:02 FLHerne (#kdevelop)
23:02 FLHerne My current best idea is "Show differences between the selected branch and the current staged (or committed) state", but that's only slightly better
23:03 ojacobson FLHerne: "staged changes" are diffs between the staging area and HEAD, "unstaged changes" are diffs between the work tree and the staging area, and … I dunno what to call the third leg of that triangle
23:04 troulouliou_div2 joined #git
23:04 ojacobson I'm unclear on the utility of a diff between "some random state-dependent thing on the current branch, and the tip of another branch" tho :)
23:05 ojacobson I tend to encourage people to think about the two problems separately (what haven't I commited yet, and how does my committed state differ from some reference point)
23:05 Vinnie_win joined #git
23:06 FLHerne I didn't invent the feature...
23:06 mizu_no_oto joined #git
23:07 Ryanar joined #git
23:07 FLHerne It shows the difference between some random branch and the latest not-unstaged state. I don't know why :P
23:07 davidfetter_ge joined #git
23:07 davidfetter_ge hi
23:08 davidfetter_ge is there any way to specify the ordering of changes in a commit?
23:08 neredsenvy joined #git
23:09 kyan joined #git
23:09 milki davidfetter_ge: no
23:09 neredsenvy Anyone else experienced the issue where git clone -b ... is horribly slow but all other operations are fast ? I have a 100Mbps connection on a decent PC : / fresh ubuntu 16.04 install, google public dns, ip6 disabled : /
23:10 davidfetter_ge milki, k. i won't waste time looking for it, then :)
23:10 anonymuse joined #git
23:10 kyan Hi! How can I get my submodules configured so that when the main repository is cloned, the submodules can be cded to and had git pull run on them?
23:10 FLHerne ojacobson: What I need to express is "the current HEAD, plus any staged changes"
23:10 milki davidfetter_ge: you can just split your changes into seperate commits
23:10 FLHerne Actually, that in itself ain't so bad
23:11 milki kyan: clone has a init submodules option i think
23:11 jstimm joined #git
23:11 milki kyan: [--recursive | --recurse-submodules]
23:11 davidfetter_ge milki, yeah, i know how to do that. it's just that the changes only make sense together in this case
23:11 milki kyan: that looks relevant
23:11 kyan milki: Yes, but when git pull is run within a submodule directory it doesn't work
23:11 kyan (gives a detached HEAD error)
23:11 milki davidfetter_ge: you can still organize your changes into a branch and merge the branch
23:12 kyan (after the main repository is already cloned)
23:12 milki kyan: you dont want to run git pull, but rather some git submodule command
23:12 milki kyan: maybe git submodule update or somethign
23:12 davidfetter_ge milki, yeah, i suppose. i have a topic branch, as usual. i guess i could make this into a flock of commits, even though it really only makes sense as a single one
23:12 kyan Mm, I'm not in charge of that unfortunately, it's an app that I'm trying to get my repository to work with, and it always runs git pull
23:12 davidfetter_ge as in separating them would break the build
23:13 nunchuck joined #git
23:13 kyan I guess my question is something like, how to re-attach the submodule head?
23:13 ojacobson FLHerne: that is the staging area, in total :)
23:13 ojacobson the staging area is a snapshot of the complete project, just like the commit that will be built from it is
23:14 ojacobson davidfetter_ge: ^ is probably relevant to you, too -- there are no "changes" in a commit, a commit is a snapshot of your project.
23:14 milki kyan: !reattach is still valid
23:14 gitinfo kyan: Letters refer to !detached. (a) and (b): 'git checkout branchname' to continue working on another branch, or 'git checkout -b branchname' to start a new one here; (c) git am --continue; (d) git rebase --continue
23:14 ojacobson Diffs are synthesized as needed, not stored and not elemental in any way
23:14 kyan Cool, thanks!
23:17 davidfetter_ge ojacobson, yeah, it would just be nice to be able to reorder changes for review without splitting them. #firstworldproblems
23:18 ojacobson For sure
23:18 Dougie187 left #git
23:22 buffal0 joined #git
23:22 Limes_ joined #git
23:23 chachasmooth joined #git
23:26 pbrewczynski joined #git
23:26 FLHerne davidfetter_ge: You could reorder the hunks in an exported patch, with your favourite text editor
23:27 venmx joined #git
23:27 davidfetter_ge FLHerne, good point
23:28 Derperperd joined #git
23:31 dsdeiz joined #git
23:31 azerus joined #git
23:32 Didac joined #git
23:38 Ciblia joined #git
23:38 User458764 joined #git
23:38 davidkrauser joined #git
23:39 thebope joined #git
23:39 ShekharReddy joined #git
23:39 ejb joined #git
23:39 davidkrauser left #git
23:43 Kaisyu joined #git
23:45 mischat joined #git
23:45 inflames joined #git
23:48 netj joined #git
23:49 Gsham joined #git
23:52 shgysk8zer0 joined #git
23:53 leftriver26 joined #git
23:55 HoierM joined #git
23:56 cagedwisdom joined #git
23:58 ochorocho__ joined #git

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