Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-09-28

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

All times shown according to UTC.

Time Nick Message
00:01 JamesBaxter joined #git
00:02 McNozes joined #git
00:02 snowkidind joined #git
00:04 codfection joined #git
00:05 pothepanda joined #git
00:06 gravitron joined #git
00:08 ayjay_t joined #git
00:13 ahr3n joined #git
00:14 vassagus joined #git
00:15 shachaf joined #git
00:15 cbreak joined #git
00:16 shachaf Say I make a branch off master, and commit a few temporary commits onto it, and master is already updated in the meanwhile. I want to see the combined diff of all the commits I made. Is there an easy way to do that? I.e. to refer to the first commit which is an ancestor of the current one which can also be reached from another branch.
00:16 shachaf (Not necessarily a proper ancestor. So if the current commit is on another branch then I just want the current commit.)
00:16 chloe joined #git
00:18 qpdb joined #git
00:18 biotim left #git
00:19 romerocesar joined #git
00:19 whitby_ joined #git
00:19 mattcen joined #git
00:19 barajasfab joined #git
00:20 EisNerd joined #git
00:22 chloe joined #git
00:26 ojacobson shachaf: git diff master...your-branch
00:26 ojacobson (three dots)
00:26 ojacobson that will be the diff between, on the left, the last common ancestor of master and your-branch, and on the right, your-branch
00:27 shachaf Is there no way to do it without specifying master? It's good enough for me that it appears in any branch.
00:27 ojacobson you need to pick a reference point to identify what git should consider "the start" of your branch
00:28 ojacobson git has no inherent knowledge of that kind of relationship
00:28 shachaf Also I don't want to diff against my branch or even HEAD but the current state of the files.
00:28 ojacobson foo@{upstream}...foo is good if foo tracks some other branch
00:28 shachaf Well, I just want it to be reachable via any branch.
00:29 ojacobson No, for fairly good (if possibly frustrating) reasons
00:30 LBV_User joined #git
00:32 EvilPenguin joined #git
00:34 Milijus joined #git
00:35 kjsage[m] joined #git
00:35 ResidentBiscuit joined #git
00:38 jaguarmagenta joined #git
00:43 vulture joined #git
00:43 pizzaops joined #git
00:43 johtso joined #git
00:43 rxo joined #git
00:43 Gsham_ joined #git
00:43 moestevens joined #git
00:44 qws-user-1228 joined #git
00:44 hdon joined #git
00:44 msm joined #git
00:44 osxi joined #git
00:45 constantinexvi joined #git
00:46 chimeracoder joined #git
00:46 emlee joined #git
00:46 sangy joined #git
00:47 nedbat shachaf: can you say again what two things you want to diff? One is the current state of the files. What's the other thing you want to diff against?
00:47 unforgiven512 joined #git
00:47 ribasushi joined #git
00:48 leprechau joined #git
00:48 gnarf joined #git
00:48 boombatower joined #git
00:51 ojacobson I parsed it as "b, and the last common ancestor of both b and _any_ topologically-connected branch"
00:52 bjpenn joined #git
00:52 shachaf Hmm.
00:52 shachaf I was going to say: If HEAD is reachable via any branch except the current one, I want to diff with HEAD. Otherwise I want to do the same thing with HEAD's parent and so on.
00:53 shachaf But of course a branch can have multiple parents. Though that seems irrelevant to my use case.
00:53 rewt HEAD is the commit you're on, so it's always reachable
00:53 ojacobson Note the constraint
00:54 mischat joined #git
00:54 shachaf Maybe I should say what I want to accomplish a bit more.
00:54 nedbat shachaf: that would help
00:54 _Patrick_ joined #git
00:55 shachaf I often write a fair amount of code without committing it. I use `git diff` to see the current state of what I've written.
00:55 raijin joined #git
00:55 shachaf That's not great because the code isn't tracked by git. I'd rather git commit -a -m 'WIP' or something like that.
00:55 shachaf But if I do that more than once, I can't use `git diff` to see all the changes I've made in my branch anymore.
00:56 nedbat shachaf: the second time you do it, use:  git commit -a --amend
00:56 shachaf Yes, that's what I've been doing.
00:56 nedbat shachaf: or, just diff against where you diverged from master
00:56 shachaf But that isn't great either because I might checkpoint at conceptually distinct points and not want to lose that history.
00:56 wloncom joined #git
00:57 shachaf Yes. "where you diverged from master" is the thing I want to specify.
00:57 ojacobson If you tell git what branch you consider to be the upstream of your local branches, then you can use git diff b@{u}...b
00:57 ojacobson for branch b
00:57 shachaf Except it might not be master but some other branch.
00:57 ojacobson (or @{u}...HEAD for the current branch)
00:57 CtrlC I created a new repo on gitlab. I want to push the project I have on my local system which belongs to another repo, to this repo. How can I do it?
00:57 shachaf Yes, that's one approach. Maybe it's the right thing to do.
00:57 ojacobson git branch -u sets that information (`git branch -u master my-feature`), or you can have git branch set it automatically when creating branches
00:58 shachaf Rather than expect git to do a search from every root in the graph.
00:58 CtrlC I read some articles on stackoverflow but couldn't figure exactly what to do.
00:58 rewt you can use gitk to see your git history and easily diff between any 2 commits
00:58 ojacobson CtrlC: git push my-gitlab-remote some list of branches
00:58 ojacobson would be the easiest way
00:58 ojacobson (easiest = simplest to reason about)
00:58 shachaf It seems like it would be a nice feature nonetheless but I can see why it might not be implemented.
00:59 nedbat shachaf: another option: when you create your branch, create two branches. one will stay at the divergence point, the other you will commit to
00:59 CtrlC my-gitlab-remote is the url. right? ojacobson
00:59 ojacobson The URL, or a remote name created with `git remote add NAME URL`
00:59 shachaf nedbat: That seems like way too much overhead, and functionally the same as tracking a branch for what I'm doing.
00:59 ojacobson shachaf: well, it wouldn't be hard to script
01:00 ojacobson so if it's useful to you, throw it at the !list and see who else bites :)
01:00 gitinfo [!mailing_list] The mailing list can be reached via git@vger.kernel.org. You don't need to subscribe to the list, you will always be put in cc on reply. Read archives at http://j.mp/gitlist
01:00 afuentes joined #git
01:01 nedbat shachaf: so you have a solution?
01:01 lolmac joined #git
01:02 cqi joined #git
01:02 shachaf Not for my current situation.
01:02 CtrlC Thank you. ojacobson
01:02 ojacobson shachaf: starting from `git show-ref --contains COMMIT 'refs/heads/*'` and iterating COMMIT over `git rev-list` would be a starting point
01:03 nedbat shachaf: how often do you make one of these branches?
01:03 shachaf But maybe I'll try git checkout --track or something in the future.
01:03 ojacobson that'd find you branches that intersect your current branch, eventually
01:03 MissionCritical joined #git
01:03 shachaf Or I can write master...HEAD because the other branch is master right now.
01:03 shachaf Hmm, what I want isn't to diff with HEAD but with the actual state of the files.
01:04 Milijus joined #git
01:04 ojacobson `git diff $(git merge-base master HEAD)`
01:04 nedbat shachaf: you mentioned overhead: what kind of overhead?
01:04 shachaf The overhead of making a second branch.
01:04 ojacobson git diff of a single commit diffs the commit (left) and the work tree (right)
01:04 fhackdroid joined #git
01:04 nedbat shachaf: put it in a shell script or something
01:05 shachaf ojacobson: OK, so I can specify it as master... or something?
01:06 prawnzy joined #git
01:06 ojacobson unfortunately no, there's no dot notation for "work tree vs some merge base"
01:06 shachaf Ah.
01:07 shachaf Then something like "git diff $(git merge-base master HEAD)"
01:07 jnavila joined #git
01:14 durham joined #git
01:15 snthsnth joined #git
01:16 thierryp joined #git
01:17 ahr3n joined #git
01:19 vassagus joined #git
01:22 d^sh joined #git
01:23 bjpenn joined #git
01:23 clemf joined #git
01:26 LeBlaaanc joined #git
01:26 yaymukund joined #git
01:27 d^sh joined #git
01:29 ahr3n joined #git
01:30 Milijus joined #git
01:32 LeBlaaanc joined #git
01:35 yanome joined #git
01:41 LeBlaaanc joined #git
01:42 barajasfab joined #git
01:46 fstd_ joined #git
01:46 Lor_ joined #git
01:48 ilbot3 joined #git
01:48 Topic for #git is now Welcome to #git, the place for git help and electronically-induced comas | Please take the git survey: https://jk.gs/gsurv | Public logs at http://jk.gs/glog | First visit? Read: http://jk.gs/git | Current stable version: 2.10.0 | Getting "cannot send to channel"? /msg gitinfo .voice | git-daemon: the evil rises
01:53 hahuang62 joined #git
01:54 Dougie187 joined #git
01:54 mischat joined #git
01:55 therealGent joined #git
01:55 linuxmint joined #git
01:56 harwood joined #git
01:56 fuchstronaut joined #git
01:57 therealGent joined #git
01:58 harish joined #git
01:58 harwood joined #git
02:00 mizu_no_oto joined #git
02:00 JamesBax_ joined #git
02:01 LeBlaaanc joined #git
02:02 timvisher joined #git
02:03 star_prone joined #git
02:04 wxg joined #git
02:06 Milijus joined #git
02:09 ruby_ joined #git
02:09 jon-mac joined #git
02:10 HappyLoaf_ joined #git
02:11 ruby_ joined #git
02:12 timvisher joined #git
02:16 bjpenn joined #git
02:17 cdg joined #git
02:19 LeBlaaanc joined #git
02:21 cdg joined #git
02:21 Goplat joined #git
02:21 sanketdg joined #git
02:23 HalfEatenPie joined #git
02:23 vassagus joined #git
02:25 t0n1c joined #git
02:26 M-shine joined #git
02:27 yaymukund joined #git
02:29 evanwang joined #git
02:29 LeBlaaanc joined #git
02:29 ahr3n joined #git
02:30 thiago joined #git
02:31 dodobrain ok, my dev team has done some stupid things in the past. in a django requirements.txt file where each line is of the form: name==x.y.z  , they have randomly (through the years) sorted the file alphabetically in commits that have no real changes to this file
02:31 dodobrain is there a way i can track a particular line content through git history?
02:31 nilsi joined #git
02:32 HappyLoaf_ joined #git
02:33 fedepad joined #git
02:33 Milijus joined #git
02:35 dreiss joined #git
02:35 jab416171 joined #git
02:37 chachasmooth joined #git
02:38 lagothri1 joined #git
02:39 jaguarmagenta joined #git
02:43 timvisher joined #git
02:44 duderonomy joined #git
02:44 LeBlaaanc joined #git
02:47 Limes joined #git
02:48 danimal2 joined #git
02:51 JanC_ joined #git
02:51 hexagoxel joined #git
02:51 GT4066 joined #git
02:53 snthsnth joined #git
02:54 ahr3n joined #git
02:54 neilthereildeil joined #git
02:54 neilthereildeil how can i use difftool to compare 2 commits?
02:55 mischat joined #git
02:56 Tonkers joined #git
02:57 Tonkers Hello! I have a website that I want to turn into a git repository. How do I do this without breaking the live site?
02:57 Tonkers or like how do I do this at all haha
02:57 Tonkers the other thing is that it's on a VPS so I can't create a git user, but I have a user with SSH access to the server
02:58 fuchstronaut joined #git
02:58 brokensyntax Like, you want to back it up to a git, and do updates by commiting to git?
02:58 brokensyntax You should have the ability to make users on a VPS (unless it's not yours.)
03:01 Tonkers brokensyntax: yep, I want to update the site through commits instead of just overwriting files through FTP as it's handled right now. and the server is a shared host, I don't have root access. as far as I can tell, my account is not authorized to make new users, but I could poke around cPanel some more?
03:01 aavrug joined #git
03:02 brokensyntax Who's the hosting provider?
03:03 texasmynsted joined #git
03:03 Orion3k joined #git
03:03 notmike joined #git
03:04 Orion3k joined #git
03:05 notmike I have a real dumb question. I am using Git command line tools on a Mac. When I open up a terminal my prompt used to be ~. Now, my prompt is ~ git:(Master) while I am under my home directory.
03:06 notmike Well, that's not really accurate either. SOmetimes I see Git:(Master) and other times I don't. Why?
03:07 barajasfab joined #git
03:07 hahuang62 joined #git
03:08 phroa` you'll only see it when the directory you're in, or one of its parents, is a git repository. (meaning there's a .git directory in it.)  if you see that in ~, that means you ran 'git init' in your home folder, which is a reallybadthing™
03:08 notmike How do I undo that?
03:10 notmike Just rm the .git directory?
03:10 Tonkers brokensyntax: it's JaguarPC. I'm an independent developer brought into this project, so I'm not 100% clear on the details yet. So any ideas on how to migrate the site into a git repository?
03:10 neilthereildeil left #git
03:12 brokensyntax So you're on a service like this one? https://www.jaguarpc.com/vps-hosting-service/
03:12 Tonkers correct! brokensyntax
03:13 Tonkers at least that's what I've been told.
03:13 cagedwisdom joined #git
03:13 Tonkers Though now that you mention it, I'm in a jail shell. So that doesn't quite make sense for being on a VPS..?
03:13 moei joined #git
03:14 brokensyntax Looks like a standard VPS host... You should have full control of that box. Have you tried to "su" or "sudo" when shelled in?
03:14 Mobutils joined #git
03:14 * brokensyntax nods
03:14 Lyqyd joined #git
03:14 brokensyntax VPS, P = Private. Should be yours to do with.
03:14 cspack joined #git
03:15 Tonkers brokensyntax: yeah, huh. I guess i'm on a shared host then, odd. not what I was told
03:15 ahr3n joined #git
03:16 brokensyntax *nod*
03:16 brokensyntax I wipe mine and redeploy at will, just wish I had a little more flexibility on the distro.
03:17 brokensyntax I imagine still should be able to make things work, but I don't know git well enough.
03:17 arescorpio joined #git
03:18 thierryp joined #git
03:20 timvisher joined #git
03:22 jab416171 joined #git
03:27 yaymukund joined #git
03:28 Tonkers brokensyntax: so turns out my client just didn't give me root access welp. assuming that I can make the user etc, how do I get the files from public_html into a git repository?
03:28 Tonkers and then pull that locally
03:28 brokensyntax su or sudo work?
03:29 Tonkers yep!
03:29 brokensyntax Golden.
03:29 ahr3n joined #git
03:29 Tonkers in a couple minutes it will that is. just so I know what I'm doing when I get there though
03:29 vassagus joined #git
03:31 re1 joined #git
03:31 alex1a joined #git
03:32 HappyLoaf_ joined #git
03:33 kyan joined #git
03:33 cspack joined #git
03:33 jaguarmagenta joined #git
03:35 cdg joined #git
03:35 mangolisk joined #git
03:36 yaymukund joined #git
03:40 boombatower joined #git
03:40 Andrew_K joined #git
03:43 mizu_no_oto joined #git
03:47 ahr3n joined #git
03:47 Tonkers brokensyntax: any ideas?
03:49 romerocesar joined #git
03:51 prawnzy joined #git
03:51 yaymukund joined #git
03:52 titankiller joined #git
03:53 notmike joined #git
03:55 Remram[m] joined #git
03:55 Remram[m] joined #git
03:56 Limes__ joined #git
03:56 titankiller joined #git
03:56 mischat joined #git
03:56 HoloPed_ joined #git
03:58 timvisher joined #git
03:58 cdg_ joined #git
03:58 fuchstronaut joined #git
03:59 cspack joined #git
03:59 LeBlaaanc joined #git
04:00 circ-user-HSkrH joined #git
04:01 Atm0spher1c joined #git
04:02 EmeraldExplorer joined #git
04:05 a_thakur joined #git
04:06 a_thakur joined #git
04:07 ahr3n joined #git
04:08 Rodya_ joined #git
04:08 laumann joined #git
04:08 notmike I'm using git command line tools on OSX. Right now my .gitconfig file and a .git repository are both in ~ is that normal?
04:09 notmike well, I just deleted the repository. So, now it's really just the .gitconfig file
04:10 the_cybersphinx joined #git
04:10 HappyLoaf_ joined #git
04:10 snthsnth joined #git
04:12 Sucks joined #git
04:12 romerocesar joined #git
04:12 daey_ joined #git
04:13 thomasreggi joined #git
04:15 ahr3n joined #git
04:19 thierryp joined #git
04:24 HappyLoaf_ joined #git
04:26 nilg joined #git
04:28 titankiller joined #git
04:29 Mobutils joined #git
04:30 red82 joined #git
04:30 LeBlaaanc joined #git
04:34 vassagus joined #git
04:35 romerocesar joined #git
04:35 Sucks joined #git
04:35 timvisher joined #git
04:35 circ-user-RzWdn joined #git
04:35 circ-user-RzWdn left #git
04:40 eizua joined #git
04:40 DaleK5whr joined #git
04:40 preaction joined #git
04:41 DaleK5whr left #git
04:41 HappyLoaf_ joined #git
04:41 LeBlaaanc joined #git
04:45 titankiller joined #git
04:46 Vinnie_win joined #git
04:47 titankiller left #git
04:49 jaequery joined #git
04:50 HappyLoaf_ joined #git
04:52 DeltaHeavy joined #git
04:52 a_thakur_ joined #git
04:52 cdg joined #git
04:54 lolmac joined #git
04:57 mischat joined #git
04:57 freimatz joined #git
04:58 thierryp joined #git
04:58 prawnzy joined #git
05:00 dermoth joined #git
05:00 subhojit777 joined #git
05:02 HappyLoaf_ joined #git
05:02 PioneerAxon joined #git
05:03 Cabanossi joined #git
05:04 glebihan joined #git
05:10 bocaneri joined #git
05:11 a_thakur joined #git
05:12 hellohill joined #git
05:13 titankiller joined #git
05:14 romerocesar joined #git
05:18 star_prone joined #git
05:22 nydel07 joined #git
05:23 Khisanth joined #git
05:24 Takiito joined #git
05:24 Takiito joined #git
05:25 polyzen joined #git
05:25 dreiss joined #git
05:26 phaleth joined #git
05:27 elementalest joined #git
05:27 elementalest joined #git
05:29 bob14 joined #git
05:29 noyb joined #git
05:30 peaches joined #git
05:35 cdown joined #git
05:38 HappyLoaf_ joined #git
05:39 vassagus joined #git
05:39 fedepad joined #git
05:40 cdown_ joined #git
05:43 dsantiago joined #git
05:44 pijiu joined #git
05:47 rchavik joined #git
05:48 lolmac joined #git
05:48 stuh84 joined #git
05:49 mrgood joined #git
05:50 ThomasLocke joined #git
05:50 ThomasLocke joined #git
05:51 dsantiago joined #git
05:52 elementalest joined #git
05:52 elementalest joined #git
05:53 yaymukund joined #git
05:54 Raging_Hog joined #git
05:57 rscata joined #git
05:57 ShapeShifter499 joined #git
05:58 mischat joined #git
05:59 HappyLoaf_ joined #git
05:59 ShapeShifter499 joined #git
06:00 fuchstronaut joined #git
06:00 ShapeShifter499 joined #git
06:02 a_thakur_ joined #git
06:07 zeroed joined #git
06:07 zeroed joined #git
06:07 qt-x joined #git
06:09 diogenese joined #git
06:11 King_Hual joined #git
06:12 Macaveli joined #git
06:14 HappyLoaf_ joined #git
06:14 rendar joined #git
06:15 tax joined #git
06:16 LeBlaaanc joined #git
06:17 lb1c joined #git
06:17 cyan__ joined #git
06:18 roelmonnens joined #git
06:18 chele joined #git
06:24 sophiag joined #git
06:25 LeBlaaanc joined #git
06:27 star_prone joined #git
06:27 danielitit joined #git
06:31 JeroenT joined #git
06:32 sophiag hi. i'm pretty new to git and had to have github recover a branch from a repo i created on there with no local directory and then overwrote with a force push. now i'm wondering if i can merge them, but it says i can't submit a pull request because they have entirely different commit histories. is there a way around this?
06:32 PioneerAxon joined #git
06:33 perrier-jouet how can I add a repository in another repository ? I have my dotfiles in which I want to add 2bwm repository. when I tried and checked on github, the repo in another repos contains no files and is not even a link
06:34 fuzzmz perrier-jouet, look into git submodules and git subtrees
06:34 fuzzmz you'd basically add 2bwm as a submodule/subtree in your own repo
06:35 JeroenT_ joined #git
06:37 Sceorem joined #git
06:41 star_prone joined #git
06:42 Alienpruts joined #git
06:43 t0by joined #git
06:43 therealGent joined #git
06:43 vassagus joined #git
06:44 perrier-jouet subtree or submodule which one is right ?
06:44 AndreSomers joined #git
06:44 therealGent joined #git
06:45 fuzzmz I'm partial to subtree in most situations
06:45 fuzzmz easier to understand than submodules
06:46 mithenks joined #git
06:46 therealGent joined #git
06:46 fuzzmz check https://help.github.com/artic​les/about-git-subtree-merges/ out, it's a pretty good tutorial
06:46 fuzzmz guide
06:48 therealGent joined #git
06:49 thierryp joined #git
06:49 therealGent joined #git
06:50 perrier-jouet is it possible to add contents of a link ? or I add more working directories ?
06:51 perrier-jouet I have /home/dagnachew/.cfg as working directory
06:53 therealGent joined #git
06:54 fuzzmz by link do you mean a symlink? or?
06:54 perrier-jouet fuzzmz: yes
06:54 fuzzmz not sure tbh, haven't tried that usecase
06:55 fedepad joined #git
06:55 therealGent joined #git
06:56 JeroenT joined #git
06:56 bjpenn joined #git
06:57 fuzzmz from what I'm reading I'd be wary of using symlinks with git: http://stackoverflow.com/a/954575/1598323 and http://stackoverflow.com/a/15465497/1598323
06:59 marlinc joined #git
07:00 thiago joined #git
07:01 fuchstronaut joined #git
07:01 Mowgli joined #git
07:01 Macaveli joined #git
07:03 flobin joined #git
07:04 lexxis joined #git
07:05 rbr joined #git
07:05 sophiag joined #git
07:05 submitnine joined #git
07:05 thekrynn joined #git
07:05 neunon joined #git
07:06 albel727 joined #git
07:09 gshrikant joined #git
07:11 titankiller joined #git
07:12 danslo joined #git
07:13 star_prone joined #git
07:14 perrier-jouet why is that when I add a repos in another repo it does not appears on github ? it just a black folder
07:14 testild joined #git
07:14 encod3 joined #git
07:15 TomyWork joined #git
07:16 star_prone joined #git
07:16 Dougie187 joined #git
07:16 fuzzmz because they're basically references to other repositories
07:16 fuzzmz if you click on them it should take you to that other repo
07:19 Seveas (only if the repo is properly configured as a submodule, and the submodule also lives on github)
07:21 star_prone joined #git
07:21 Mobutils joined #git
07:26 Balliad joined #git
07:27 jaeckel joined #git
07:28 marlinc joined #git
07:29 dc5ala joined #git
07:30 timewalker joined #git
07:31 romerocesar joined #git
07:32 laumann joined #git
07:32 star_prone joined #git
07:35 plexigras joined #git
07:36 testild joined #git
07:36 star_prone joined #git
07:38 watabou joined #git
07:39 eduardas_m joined #git
07:40 plexigras how can i set the difftool for a user?
07:40 j416 "for a user"?
07:40 Infamous joined #git
07:41 watabou joined #git
07:41 osse plexigras: set it in /home/user/.gitconfig
07:45 plexigras osse, just put "tool = meld" in there?
07:45 tvw joined #git
07:47 overlord_tm joined #git
07:48 laumann joined #git
07:48 plexigras dosent seam to work
07:49 ahmedelgabri joined #git
07:49 vassagus joined #git
07:51 plexigras did it $ git config --global diff.tool meld
07:51 bket joined #git
07:52 omx joined #git
07:53 fuzzmz don't do global, that adds it to all users
07:53 fuzzmz nvm, I'm dumb, ignore me
07:53 osse fuzzmz: no, that's --system
07:53 fuzzmz ayup
07:53 fuzzmz need to wake up before talking tech
07:53 HappyLoaf_ joined #git
07:55 osse fuzzmz: https://open.spotify.com/tr​ack/4hXqWYGyJYHrGawZOK8Rqq
07:56 fuzzmz osse, that sounds so nice
07:56 yaymukund joined #git
07:57 fedepad joined #git
07:58 osse you're welcome
08:00 JanC joined #git
08:01 fuchstronaut joined #git
08:03 _ADN_ joined #git
08:04 JanC joined #git
08:04 BloqueNegro2 joined #git
08:05 ocbtec joined #git
08:05 BloqueNegro2 Hi all together. I experience a strange git behaviour. I have three Gits: one @customer, one @cloud and one @bitbucket
08:06 BloqueNegro2 I have to clone the @customer git to our $site1 - which won't work. it will start, but then get slow and stop
08:06 BloqueNegro2 however, I can clone from @customer git to $site2
08:07 BloqueNegro2 and from there push it to @bitbucket or @cloud git - and from there i can now clone to $site1
08:08 grawity sounds like a general connection problem between '@customer' and '$site1'
08:08 grawity maybe MTU issues or something
08:08 BloqueNegro2 any idea what could cause this behaviour? only the combination of "@customer" and "$site1" won't work, and i'm running out of ideas why
08:08 BloqueNegro2 grawity: yeah, i'm looking for that kind of suggestions
08:09 brokensyntax joined #git
08:09 fuzzmz BloqueNegro2, if you ping @customer from $site1 what do you get?
08:10 BloqueNegro2 ping is disabled because paranoid admin
08:10 BloqueNegro2 but i ran tcpdump and see the packet flow stopping
08:10 BloqueNegro2 so i assume it's something related to the network and not to the git itself
08:10 grawity 'paranoid admin' is likely to be part of the issue, tbh
08:10 fuzzmz yeah
08:10 BloqueNegro2 :>
08:10 BloqueNegro2 i know
08:11 BloqueNegro2 but there is no way i could fix it
08:11 grawity when they blocked ping (ICMP Echo), they probably also whacked PMTUD (ICMP Too Big) and such
08:11 grawity that's my guess at least
08:11 fuzzmz BloqueNegro2, can you try doing a shallow clone? Something like git clone --depth=1 to get only the latest commit?
08:12 grawity anyway, this one might be easy to test; `ip link set eth0 mtu 1400` or such
08:12 fuzzmz then you can get more history using git fetch --depth=X and gradually increase X
08:12 BloqueNegro2 this results in the same problem
08:12 fuzzmz or if you have git 1.8 or newer use git fetch --unshallow to get all the remaining revisions
08:13 fuzzmz damn
08:13 BloqueNegro2 grawity: just checked, mtu on $site1 and $site2 is both 1500
08:14 grawity BloqueNegro2: checked how?
08:14 Snugglebash joined #git
08:14 BloqueNegro2 ip addr
08:14 grawity that's just the MTU for your immediate link
08:14 grawity and MTU problems generally happen when something *in between* has a lower one
08:14 HappyLoaf_ joined #git
08:15 BloqueNegro2 and when they block the "ICMP Too Big" they won't know about it
08:15 BloqueNegro2 this may be an approach
08:16 BloqueNegro2 (atleast that's what you said, right?)
08:16 grawity yes
08:16 grawity well, it's just one possible cause, but it's a common one.
08:16 BloqueNegro2 but why would it work towards $site2 then?
08:16 fuzzmz different route probably
08:16 aidalgol joined #git
08:16 grawity I assume both sites are at different locations (hence their name) and use different paths to '@customer'
08:17 fuzzmz that or maybe $site1 has a proxy or something in front of it which fucks everything up
08:17 grawity (is '@customer' an array or what)
08:17 chll_ joined #git
08:17 BloqueNegro2 no, that's just my shitty notation to show the difference between site and git :/ sorry for that ;)
08:18 grawity between site and git... server?
08:18 BloqueNegro2 there is an internet between with lots of devices
08:18 fuzzmz site is the webserver, git is the server holding the code for that website I think
08:18 BloqueNegro2 yeah basically
08:19 BloqueNegro2 the problem is, they store docker images (don't even ask...) in their git
08:19 noyb joined #git
08:19 fuzzmz hah, I take it git's trying to download hundreds+ megs doing the clone, eh?
08:19 BloqueNegro2 yes
08:19 clmsy joined #git
08:20 BloqueNegro2 which works fine unless @customer -> $site1
08:20 fuzzmz it might also come down to timeouts
08:21 fuzzmz if $site1 has a slower connection to @customer than $site2, then @custome might just drop the git connection because it's taking too long
08:21 BloqueNegro2 but not after a few seconds
08:21 sin3t joined #git
08:21 BloqueNegro2 the whole repo is < 200mb
08:21 fuzzmz oh, it's that fast?
08:21 BloqueNegro2 yes
08:22 fuzzmz what error does it throw?
08:22 BloqueNegro2 basically it loads about 20-30mb and then it will just hang til timeout
08:22 BloqueNegro2 and with tcpdump i see no more packet flow (and nload doesn't show traffic)
08:22 grawity can you tcpdump on the server?
08:22 BloqueNegro2 only on $site, not on @customer
08:23 vls-xy joined #git
08:24 EY joined #git
08:24 marlinc joined #git
08:25 titankiller joined #git
08:26 tlaxkit joined #git
08:26 BloqueNegro2 just reduced mtu to 1000, no success
08:27 BloqueNegro2 i'll be afk for 30min, but i'll be back after that :)
08:28 mikecmpbll joined #git
08:28 ronny hi
08:29 ronny is there a way to figure if a repo is shallow for a script/program?
08:30 wololokatamari[m joined #git
08:30 fuzzmz ronny, from what I remember if the repo is shallow then it has a shallow file in git-dir
08:31 fuzzmz you could do something like [ -f $(git rev-parse --git-dir)/shallow ] && echo true || echo false
08:33 mischat joined #git
08:34 aax joined #git
08:35 a_thakur joined #git
08:35 Kitty- joined #git
08:36 Mobutils joined #git
08:37 Serpent7776 joined #git
08:38 Darcidride joined #git
08:39 Mobutils joined #git
08:41 jokke joined #git
08:43 flobin joined #git
08:44 pijiu joined #git
08:46 jaeckel joined #git
08:47 felixjet i don't know if this is a good answer here but does anyone migrated from bitbucket to gitlab and can give some feedback? i'm using bitbucket because of private repos but no idea how it compares to gitlab
08:47 felixjet a good question*
08:48 fuzzmz I'm using GitLab. The main downside is that it can be pretty slow at times
08:48 fuzzmz otherwise I can't find any downsides to it, but I was coming from Github
08:49 fuzzmz GitLab is more open about development, and seems to get more features faster than bitbucket and github (though github started speeding up recently)
08:49 testild joined #git
08:49 felixjet you using gitlab self-hosted? or their service
08:50 PioneerAxon_ joined #git
08:50 fuzzmz their service
08:50 fuzzmz gitlab.com
08:50 testild joined #git
08:50 felixjet probably a self-hosted installation won't have the speed downside
08:50 fuzzmz yep
08:51 felixjet any reason why did you choose gitlab and not bitbucket?
08:51 fuzzmz we're using an atlassian stack at work, and stash/bitbucket server left me a pretty meh impression
08:52 fuzzmz you can't configure PR merging strategy without being an admin, less info displayed in the repository view
08:52 fuzzmz and my move was from github to gitlab self-hosted (bitbucket wasn't an option here) to gitlab.com
08:52 al-damiri joined #git
08:53 fuzzmz and once I was used to gitlab it made less sense to move to bitbucket
08:53 felixjet i see, thanks for sharing!
08:54 vassagus joined #git
08:55 spykins joined #git
08:57 martinsb joined #git
09:00 martinsb hi. I have branch that i merged into master and it broke everything. After that i did revert that made commit that removed all changes that previous commit did. I now know what were the problems with the branch and i want to put it at top of my master but when i try to merge it it says "Already up to date" since those commit id's are in parent chain. How do i force the merge?
09:01 spykins_ joined #git
09:01 tobiasvl martinsb: !faulty_merge
09:01 gitinfo martinsb: If you have published a merge commit that turned out to be faulty and you'd like to get rid of it, you're probably not going to have a good time. Here's the full story: https://raw.github.com/git/git/master/Docu​mentation/howto/revert-a-faulty-merge.txt
09:01 fuzzmz martinsb, first you would need to switch to that branch and make a new commit fixing those problems; then you'd rebase that branch on top of master, after that you'd merge it into master
09:02 JeroenT joined #git
09:02 spykins Hello everyone, I am reading Pro Git. I finished the chapter one today, I learnt about how to git config —global user.name spykins.. My question is, how do I change the global git config and folder git config
09:02 fuchstronaut joined #git
09:03 fuzzmz spykins, what do you mean by change the global git config and folder git config?
09:03 HappyLoaf_ joined #git
09:04 fuzzmz git config --global option value modifies the global gitconfig, git config option value modifies the repo-specific one
09:05 fuzzmz for example git config --global user.name spykins tells git to update the global gitconfig (the one in your users home) and set the user name to spykins
09:05 fuzzmz you can see the contents of the files by hand. the one in your repo is stored in the config file in the .git folder
09:05 fuzzmz the global one is stored in your users home in the .gitconfig file
09:06 fuzzmz just open them with a text editor
09:06 Ezzy joined #git
09:06 watersoul joined #git
09:07 Snugglebash joined #git
09:09 daynaskully joined #git
09:10 JeroenT_ joined #git
09:10 martinsb fuzzmz: tobiasvl: Reading tobiasvl provided link from what i understand i can do revert on my previous revert i did to remove the merge. That will result in latest code + code that was in merge. Then on top of it i fix the problem the merge had and i should be good to go
09:10 rominronin joined #git
09:12 tango_ joined #git
09:12 Masber_080 joined #git
09:16 achadwick joined #git
09:17 chachasmooth joined #git
09:19 fuchstronaut joined #git
09:21 HappyLoaf_ joined #git
09:23 Pieplay joined #git
09:24 idk_njr joined #git
09:29 idk_njr left #git
09:29 idk_njr joined #git
09:30 neunon joined #git
09:33 ahr3n joined #git
09:34 shinnya joined #git
09:35 kurkale6ka joined #git
09:36 Wind0r joined #git
09:37 a_thakur joined #git
09:39 rorro joined #git
09:40 Sceorem_ joined #git
09:41 DelphicOkami joined #git
09:42 Sceorem__ joined #git
09:42 wololokatamari[m joined #git
09:44 Mobutils joined #git
09:45 WayToDoor joined #git
09:47 weckl joined #git
09:47 weckl joined #git
09:48 weckl joined #git
09:49 weckl joined #git
09:50 weckl joined #git
09:50 stamina joined #git
09:50 weckl joined #git
09:53 titankiller joined #git
09:53 ninjacodr joined #git
09:56 heftig joined #git
09:58 ssitani joined #git
09:58 vassagus joined #git
09:58 yaymukund joined #git
09:59 tHEMtZ joined #git
10:01 a_thakur joined #git
10:01 shafun joined #git
10:05 phaleth joined #git
10:10 rafalcpp_ left #git
10:11 Cynede joined #git
10:12 Cynede hi
10:12 gitinfo Cynede: 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.
10:12 Cynede how can I show not added files?
10:13 nedbat Cynede: doesn't git status do that for you?
10:14 Mobutils joined #git
10:14 danslo joined #git
10:16 Cynede nedbat yes ) thanks >__M
10:23 HappyLoaf_ joined #git
10:24 tvd joined #git
10:24 cyan__ joined #git
10:25 thebope joined #git
10:31 dodobrain joined #git
10:32 dsdeiz joined #git
10:32 dsdeiz joined #git
10:34 PCatinean joined #git
10:35 compeman joined #git
10:36 yaymukund joined #git
10:36 tlaxkit joined #git
10:40 titankiller joined #git
10:41 bruce_lee joined #git
10:42 tvd left #git
10:42 HappyLoaf_ joined #git
10:44 maret joined #git
10:44 andy___ joined #git
10:44 jaeckel joined #git
10:44 andy1254 joined #git
10:45 andy1254 left #git
10:49 shock- joined #git
10:52 _ng joined #git
10:53 eshlox joined #git
10:54 JeroenT_ joined #git
10:54 winterchip joined #git
10:54 MrWoohoo joined #git
10:54 JeroenT joined #git
10:55 basy_ joined #git
10:57 qt-x joined #git
10:57 flobin joined #git
10:57 mischat joined #git
11:00 prawnzy joined #git
11:01 JeroenT_ joined #git
11:01 cloudbud joined #git
11:03 ahmed_elgabri joined #git
11:03 cloudbud how can we check on which bran ch the head is pointing
11:04 vassagus joined #git
11:05 _ikke_ git symbolic-ref HEAD
11:05 timvisher joined #git
11:06 JeroenT joined #git
11:07 jaguarmagenta joined #git
11:08 acetakwas joined #git
11:08 Cynede left #git
11:10 dunaeth joined #git
11:11 JeroenT joined #git
11:12 midnightmagic joined #git
11:13 rnsanchez joined #git
11:13 dunaeth Hi, let's say I made a stash keeping all files but one in index with -k, that for some reason I worked on the indexed files and committed them. Is git show stash@{n} the best way to retrieve the real changes (git stash show seems to show me all the changed at the stash time, even if I already have them in the current index) ?
11:14 watabou joined #git
11:15 _ikke_ dunaeth: git stash stores the commit and index it was based off, so it uses those to diff against (so comitting changes has no effect on the output)
11:17 dunaeth _ikke_ , yup so git stash show will show me all the changes stashed between its ref commit and the stash itself, but if I assume the files I wanted to keep in index where somehow discarded
11:17 moritz you can also do a  git diff stash@{2} HEAD if you want to diff against the current state of the repo
11:17 dunaeth could git show stash@{n} returns me only the file I first wanted to stash ?
11:23 HappyLoaf_ joined #git
11:23 heroux joined #git
11:24 mischat joined #git
11:25 Mobutils joined #git
11:26 dunaeth Maybe my question could be summarized as "Is it a [possible / a good idea / stupid] to use git show stash -p stash@{n} | git apply to patch from an old stash ?"
11:26 mischat_ joined #git
11:26 ^7heo dunaeth: why not git stash pop stash@{n}
11:26 ^7heo ?
11:27 sdothum joined #git
11:28 dunaeth cause the ref context has changed (I had several modifications and did not want to commit them all at once, so I (mis)used git stash -k to keep modification I wanted to commit, commit'em, fix some bugs and the pop the other modifications)
11:28 ojdo joined #git
11:28 lexxis joined #git
11:28 theskillwithin joined #git
11:29 dunaeth I misused git stash -k since I stated that it would not stash the files I intended to keep, where it actually stashed them, but kept'em in index
11:29 ^7heo dunaeth: the way I pop the stash when the context has changed is the following:
11:29 ^7heo I apply the patch restoring the state it was in
11:29 ^7heo I add that to the index
11:30 ^7heo I pop the stash
11:30 ^7heo I reset
11:30 rominronin joined #git
11:30 ^7heo can be summarized as: git checkout $SHA -- $FILE; git add .; git stash pop stash@{$STASH}; git reset
11:31 ^7heo and then you can start resolving the delta.
11:31 moritz but why?
11:31 ^7heo moritz: why what?
11:31 moritz ^7heo: why do you check out files from other versions just to create a stash?
11:32 psprint joined #git
11:32 ^7heo moritz: where did I create a stash?
11:32 moritz ^7heo: sorry, I was confused
11:32 ^7heo np
11:32 psprint Hello. git -C . rev-parse --show-toplevel will show path that is a resolved simlink. I.e., my PWD=/path/with/simlink. Can it not resolve simlinks?
11:32 ^7heo moritz: this is to pop a stash while the history has diverged from the state the stash was in.
11:32 mizu_no_oto joined #git
11:33 ^7heo moritz: that way, you have a `git diff` showing all the changes since then, but as deletions
11:33 ^7heo moritz: and your stash as addition.
11:33 ^7heo moritz: so it's pretty easy to see what is what.
11:33 ^7heo moritz: and rather trivial to manage.
11:34 dunaeth So the git show stash@{n} that seemed promising (actually it seemed to give me an expected result) is not an option
11:34 dunaeth I supposed it could be some shortcut to what you are doing
11:35 sin3t joined #git
11:35 dunaeth treating the stash as a raw patch, not looking at its context
11:37 HappyLoaf_ joined #git
11:37 nuck joined #git
11:38 rorro joined #git
11:38 ^7heo dunaeth: the problem with the git stasg show -p | git apply is that it doesn't differ from git stash pop
11:38 ^7heo s/sg/sh/
11:38 romerocesar joined #git
11:38 dunaeth please note it's not git stash show but git show stash@{#n}
11:38 ^7heo same.
11:39 ^7heo I have dully noted that.
11:39 nettoweb joined #git
11:39 ^7heo Duly
11:39 bsanford joined #git
11:39 dunaeth thus, the result of the 2 commands is really different
11:39 ^7heo no
11:40 ^7heo git stash show -p stash@{$STASH} | git apply == git stash pop stash@{$STASH} afaik.
11:40 ^7heo gosh, you make me write a lot.
11:42 dunaeth :) git stash show -p stash@{n} gimme the whole differences between HEAD at stash time and the stash, whereas git show stash@{n} only returns me the differences between my current index and the stash, without the modification I made to the old HEAD, that's why I find it weird
11:43 qt-x joined #git
11:44 ^7heo dunaeth: oh, I expected you to have made a typo and never mean to type `git show stash@...`
11:44 ^7heo dunaeth: sorry, it's not often that on IRC, people type what is correct.
11:44 dunaeth :D
11:45 ^7heo I actually have never bothered checking git show stash@...
11:45 ^7heo but I would expect it to be the bare patch
11:45 sunri5e joined #git
11:45 ^7heo which in any case will probably not apply to your current HEAD.
11:49 dunaeth I suppose that the risk is if the patch target has been modified
11:50 dunaeth I could give it a try in my use case though
11:51 flobin joined #git
11:51 dunaeth ok I now know what u meants saying it would not apply to the current HEAD
11:53 cloudbud git log --oneline --decorate command is used for ?
11:53 grawity for getting a one-line, decorated log
11:53 grawity of course
11:55 nedbat cloudbud: that's one of those questions that makes me think I don't understand the question.
11:55 cloudbud grawity : But i was reading that it is used to find the where the branch pointers are pointing.
11:55 cloudbud but i cant understand the output
11:56 nedbat cloudbud: can you pastebin the output?
11:56 grawity cloudbud: in an indirect way, I guess
11:57 mischat joined #git
11:57 cloudbud http://pastebin.com/ecvB0j6G
11:57 cloudbud nedbat grawity : output pasted
11:58 fuzzmz cloudbud, for example be7bf10 (tag: 2.8.26, origin/master, origin/HEAD) Merged change f82ed89a-d0c5-4a
11:58 nedbat cloudbud: add --graph to the command, and it might be clearer
11:58 fuzzmz this lets you know that for that origin/master and a tag point to that commit
11:58 CheckDavid joined #git
11:59 fuzzmz I usually make an alias to git log --oneline --all --graph --decorate
12:00 ahmed_elgabri joined #git
12:00 rominronin joined #git
12:00 dan2k3k4 joined #git
12:02 johnmilton joined #git
12:03 raijin joined #git
12:03 cloudbud I cant understand the output
12:03 lamont joined #git
12:03 johnmilton joined #git
12:04 fuzzmz it basically tells you what commits follow which branch
12:04 cloudbud but doc says it tells which branch head is pointing to
12:04 nedbat cloudbud: if you have new output, put it in a pastebin, and tell us which line in particular you want to know about.
12:05 fuzzmz cloudbud, if you drop the --decorate you won't have branch info on commits
12:06 fuzzmz basically --decorate adds to the appropriate commits information such as tags and branch info (if it's the latest commit on a particular branch)
12:06 fuzzmz the things inside ( ) in your pastebin
12:06 HoloPed_ joined #git
12:06 aavrug joined #git
12:07 fuzzmz be7bf10 (tag: 2.8.26, origin/master, origin/HEAD) Merged change f82ed89a-d0c5-4a
12:07 fuzzmz 09cca47 (origin/berks_bump) Bump berkshelf gem version
12:07 fuzzmz for example --decorate tells you that be7bf10 is the latest commit on the origin/master branch (the master on the server)
12:07 fuzzmz where as 09cca47 is the latest commit on the berks_bump branch on the server
12:07 aavrug joined #git
12:08 mischat_ joined #git
12:09 vassagus joined #git
12:09 zivester joined #git
12:10 nedbat cloudbud: are these explanations helping?
12:10 johnmilton joined #git
12:13 dsdeiz_ joined #git
12:14 Surfer2010 joined #git
12:16 HappyLoaf_ joined #git
12:17 bruce_lee joined #git
12:17 bruce_lee joined #git
12:17 LBV_User joined #git
12:18 watabou joined #git
12:18 Dougie187 left #git
12:21 d10n-work joined #git
12:21 JeroenT joined #git
12:22 JeroenT joined #git
12:23 Pyrogerg joined #git
12:24 WayToDoor joined #git
12:24 PCatinean joined #git
12:26 flaviodesousa joined #git
12:28 dunaeth Is there any way to know why applying a patch fails with a fatal : unrecognized input ?
12:29 HappyLoaf_ joined #git
12:29 King_Hual joined #git
12:29 theskillwithin joined #git
12:32 mohabaks joined #git
12:33 DeltaHeavy joined #git
12:33 King_Hual joined #git
12:33 King_Hual joined #git
12:34 mischat joined #git
12:35 serialoverflow joined #git
12:38 Impaloo joined #git
12:40 UTAN_dev joined #git
12:41 elementalest joined #git
12:41 elementalest joined #git
12:42 cdrf joined #git
12:42 averell joined #git
12:42 bauruine joined #git
12:42 isysd_ joined #git
12:42 testild joined #git
12:42 alcohol how could i get a list of all moved files in a commit? git show sha --stat almost gets me there, but it doesnt detect all renames
12:43 kugel joined #git
12:44 bauruine joined #git
12:44 matthiaskrgr git diff -M100% ?
12:45 matthiaskrgr -M[<n>], --find-renames[=<n>]
12:45 matthiaskrgr Detect renames. If n is specified, it is a threshold on the similarity index (i.e. amount of
12:45 matthiaskrgr addition/deletions compared to the file’s size). For example, -M90% means Git should consider a
12:45 matthiaskrgr delete/add pair to be a rename if more than 90% of the file hasn’t changed.
12:45 matthiaskrgr [..]
12:45 sanketdg joined #git
12:45 Remram[m] joined #git
12:45 kikijiki[m] joined #git
12:45 Gsham joined #git
12:45 kjsage[m] joined #git
12:45 hallogallo[m] joined #git
12:45 M-shine joined #git
12:45 M-meznak joined #git
12:45 M-mistake joined #git
12:45 aviraldg joined #git
12:45 Ineentho[m] joined #git
12:46 matthiaskrgr git show should accept -M also
12:47 jeffreylevesque joined #git
12:48 hahuang61 joined #git
12:49 Bowel_ left #git
12:51 WayToDoor joined #git
12:51 alcohol so, git show 5a3f4b7ca01a650df241a0246ab8e9032175ba01 --stat -M25% kind of gives me what i want
12:51 alcohol but when i redirect it to a tmp file
12:51 alcohol it truncates the output
12:51 heftig truncates?
12:51 alcohol as in, all the paths that are too long start with ...
12:52 cdown joined #git
12:52 alcohol every line in the file is max 62 characters long
12:52 heftig git show --stat=200
12:52 heftig or whatever
12:52 alcohol lol ok, that works. didnt know about that
12:57 f3r70rr35f joined #git
12:58 jguillen joined #git
12:58 cdrf joined #git
13:01 brokensyntax joined #git
13:01 prawnzy joined #git
13:01 jnavila joined #git
13:01 rchavik_ joined #git
13:02 cdrf joined #git
13:03 PioneerAxon joined #git
13:03 Shentino joined #git
13:03 Cyclohexane joined #git
13:03 dunaeth Hi, is there a way to remove commit informations in git show, only keeping diff ?
13:04 osse dunaeth: use git diff
13:04 grawity use --format=
13:04 ^7heo grawity: format what?
13:05 grawity ^7heo: format the commit information, of course
13:05 grawity you don't want any – specify an empty format string
13:05 HappyLoaf_ joined #git
13:05 dunaeth format with empty value is the trick, thanks
13:05 Dazzyp joined #git
13:06 grawity though, it *is* generally identical to diffing against commit^
13:06 grawity except for merges, in which case *shrug*
13:06 osse it's interesting that --format= works but --format=x generates an error
13:06 ^7heo grawity: I didn't know that format could take a string.
13:06 osse in any case i think usnig git diff makes more sense if it's a diff you want
13:06 ^7heo grawity: I didn't know that format could take a string. Thanks for pointing it out.
13:07 grawity osse: it generally expects either a predefined name, or an arbitrary format prefixed by "format:"
13:07 filcab joined #git
13:07 grawity though the latter part seems to be magic when a % is present
13:07 jimi_ joined #git
13:07 osse ah
13:07 osse --format=format:x works
13:07 ^7heo osse: --format= also works
13:07 jaguarmagenta joined #git
13:08 ^7heo osse: oh you meant to add an arbitrary string.
13:08 osse yes I said so :p
13:08 grawity --pretty[=<format>], --format=<format>
13:08 grawity Pretty-print the contents of the commit logs in a given format, where <format> can be one of oneline, short, medium, full, fuller, email, raw, format:<string> and
13:08 grawity tformat:<string>. When <format> is none of the above, and has %placeholder in it, it acts as if --pretty=tformat:<format> were given.
13:08 ^7heo osse: yeah nevermind, brainfart.
13:08 ryez joined #git
13:09 osse Farts of the brain
13:09 yaymukund joined #git
13:10 sangy joined #git
13:10 JamesBaxter joined #git
13:10 cdg joined #git
13:12 rgrinberg joined #git
13:12 staafl joined #git
13:14 yaymukund joined #git
13:14 vassagus joined #git
13:15 harish joined #git
13:16 nilg joined #git
13:18 red82 joined #git
13:20 HappyLoaf_ joined #git
13:21 watabou joined #git
13:21 mischat joined #git
13:23 somaReverse joined #git
13:23 ExoUNX joined #git
13:24 LionsMane joined #git
13:26 dothebart joined #git
13:27 serialoverflow joined #git
13:30 ash_workz joined #git
13:30 fmcgeough joined #git
13:31 anonymuse joined #git
13:31 dothebart Hi, I'm using git for windows, and since the last days it will give me: Error in GnuTLS initialization: Failed to acquire random data.\n fatal: curl_global_init failed\n error: Could not fetch origin
13:32 dothebart I've tried the usual install/uninstall/reboot cycles... tried looking at procmon, but didn't find anything suspective
13:35 jost joined #git
13:37 GodGinrai joined #git
13:40 HappyLoaf_ joined #git
13:41 spykins joined #git
13:41 zivester joined #git
13:43 acetakwas joined #git
13:44 hexagoxel joined #git
13:44 ruby_ joined #git
13:45 rnsanchez_ joined #git
13:46 keltvek joined #git
13:46 JeroenT_ joined #git
13:46 oedgok joined #git
13:47 JeroenT joined #git
13:47 t4nkd joined #git
13:48 raijin joined #git
13:48 hexagoxel joined #git
13:49 PCatinean joined #git
13:51 fedepad joined #git
13:53 hexagoxel joined #git
13:55 patarr joined #git
13:56 sin3t joined #git
13:57 hexagoxel joined #git
14:01 Gsham joined #git
14:01 AaronMT joined #git
14:02 edwlarkey joined #git
14:03 codfection joined #git
14:04 lolmac joined #git
14:08 ResidentBiscuit joined #git
14:09 edwlarkey joined #git
14:10 jaafar joined #git
14:10 DeltaHeavy joined #git
14:10 e14 joined #git
14:11 nettoweb joined #git
14:11 hanthings joined #git
14:15 Mobutils joined #git
14:15 ShekharReddy joined #git
14:16 Rodya_ joined #git
14:16 charlesr joined #git
14:16 watabou joined #git
14:19 vassagus joined #git
14:19 Snugglebash joined #git
14:19 cyan__ joined #git
14:21 TomyWork can git config items be passed via environment variables?
14:21 fedepad joined #git
14:21 dothebart left #git
14:21 TomyWork "git -c rebase.autoStash=true svn rebase" doesn't seem to do anything
14:22 madewokherd joined #git
14:22 mischat joined #git
14:22 TomyWork my guess is the "git rebase" command is invoked by "git svn rebase" without passing on config overrides
14:23 TomyWork so maybe there's an env var i can use?
14:23 elementalest joined #git
14:23 elementalest joined #git
14:25 subhojit777 joined #git
14:26 bluepixel joined #git
14:26 fees joined #git
14:27 shafun joined #git
14:28 dendazen joined #git
14:30 Zamyatin joined #git
14:30 flobin joined #git
14:30 wrouesnel1 joined #git
14:31 noyb joined #git
14:32 synthroid joined #git
14:32 crose joined #git
14:32 matsaman joined #git
14:34 mischat_ joined #git
14:34 mohabaks joined #git
14:34 vassagus joined #git
14:34 tlaxkit joined #git
14:35 jaeckel joined #git
14:35 emg joined #git
14:38 Dong__ joined #git
14:39 Dong__ Is there any tool you can use to show two commit tree overlapping each other? I can use gitk to show one.
14:39 canton7 what do you mean, "two commit tree overlapping each other"?
14:39 canton7 you're aware that gitk can show multiple branches at the same time, if that's what you're asking?
14:40 sunri5e joined #git
14:40 Dong__ like say a clone I made yesterday compare with a clone I made today.
14:40 Remram[m] joined #git
14:40 Remram[m] joined #git
14:40 canton7 ah, two completely different repositories
14:40 canton7 you can fetch one into the other, I guess
14:41 Dong__ but then how to tell which part is new from gitk?
14:41 acetakwas joined #git
14:42 hobodave joined #git
14:43 ochorocho__ joined #git
14:44 skyres joined #git
14:45 brokensyntax joined #git
14:46 ech0s7 joined #git
14:47 ech0s7 there is any chance to keep local changes if i change current branch ?
14:47 Pyrogerg joined #git
14:47 HedgeMage ech0s7: sure, use git stash
14:47 HedgeMage :)
14:47 HedgeMage it's for stashing stuff while you change your working tree around
14:48 HedgeMage 'man git-stash' for instructions
14:48 gitinfo the git-stash manpage is available at http://jk.gs/git-stash.html
14:48 canton7 Dong__, one will appear as remote-tracking branches
14:48 ech0s7 should be simple
14:48 ech0s7 git stash save;git checkout branch;git stash pop
14:48 ech0s7 should works ?
14:48 sunri5e joined #git
14:49 HedgeMage yep
14:50 inflames joined #git
14:51 ewet joined #git
14:51 PioneerAxon joined #git
14:51 Cyclohexane joined #git
14:52 michaelsharpe joined #git
14:53 logikos joined #git
14:54 logikos when i do a git push how do i make it not ask me for username and password.  I created a rsa key and added it to my account on github
14:54 ilmari logikos: make sure you're pushing over ssh, not https
14:54 synruby joined #git
14:54 synruby I git it now.
14:54 logikos ilmari how do i make sure of that...
14:55 ilmari logikos: to see what it currently is: git remote -v
14:55 ilmari logikos: to set it: git remote set-url origin git@github.com:username/repo.git
14:55 mohabaks joined #git
14:55 ilmari replace origin/username/repo with the appropriate values
14:55 synruby I pulled my sh!t together pushed myself harder and now I finally git it.
14:56 Hobbyboy joined #git
14:57 ahmedelgabri joined #git
14:57 kpease joined #git
14:58 logikos ilmari: I updated .git/config and changed https: and then when i do git push it says: fatal: remote error: You can't push to git://github.com/... Use https://github.com/...
14:58 thebope joined #git
14:59 ilmari logikos: git@github.com:, not git://github.com/
14:59 ilmari user@host: is ssh
14:59 canton7 git:// is an authenticated protocol, and is typically only used for anonymous fetching
14:59 ilmari *un*authenticated
14:59 star_prone joined #git
14:59 canton7 bleh, yep, I meant to type that
15:00 Celmor joined #git
15:01 kpease joined #git
15:02 prawnzy joined #git
15:03 Darren_ joined #git
15:03 roelmonn_ joined #git
15:04 roelmonn_ joined #git
15:05 InfoTest joined #git
15:05 spykins_ joined #git
15:05 Celmor can someone suggest a deploy mechanism for getting the source fil, destination, permissions and applying corresponding operation? my best method currently is using such script https://pastebin.com/pv5PuS0t
15:06 grawity um, what do you need it for
15:06 Celmor deploying config files to another location
15:07 `slikts joined #git
15:08 `slikts I have a bare repo ~/bla.git on my server; how do I clone it over ssh?
15:08 grawity `git clone [user@]host:path`
15:08 jaguarmagenta joined #git
15:08 grawity just like git@github.com:user/repo
15:08 `slikts if I do git clone ssh://server:~/bla.git, I get "/bla.git does not appear to be a git repository"
15:09 grawity don't mix URL style addresses and rcp style addresses
15:09 le_melomane joined #git
15:09 `slikts same difference
15:09 grawity by which you mean
15:09 `slikts same error
15:09 grawity same error with what
15:09 `slikts "/bla.git does not appear to be a git repository"
15:10 grawity either server:bla.git or ssh://server/~/bla.git should work
15:10 `slikts ah, nvm
15:10 `slikts I had the wrong path now; removing ssh:// worked
15:10 `slikts thanks
15:10 grawity which is why I said "don't mix URL style and rcp style addresses"
15:10 grawity whatever you were using was half one, half other
15:11 clemf joined #git
15:11 `slikts where can I read about rcp addresses? googling didn't help
15:11 clemf joined #git
15:12 magicbit joined #git
15:12 grawity man git fetch
15:12 gitinfo the git-fetch manpage is available at http://jk.gs/git-fetch.html
15:12 grawity or man git clone, has all the same stuff
15:12 gitinfo the git-clone manpage is available at http://jk.gs/git-clone.html
15:12 `slikts that doesn't mention rcp
15:13 notebox joined #git
15:13 spykins_ joined #git
15:13 grawity it mentions scp
15:13 grawity same difference
15:13 `slikts so I'm guessing rcp is scp(1)
15:13 Dong__ the git bot is a bit interesting, would alphaGo do a better job I wonder
15:13 lbv-user joined #git
15:14 tyreld_ joined #git
15:14 rominronin joined #git
15:14 spykins fuzzmz: Thank you boss.
15:14 spykins I am sorry for the late reply fuzzmz , I had to rush out in the morning
15:15 HappyLoaf_ joined #git
15:15 lbv-user left #git
15:16 Mobutils joined #git
15:17 raijin joined #git
15:20 NeverDie joined #git
15:21 spykins joined #git
15:22 davix joined #git
15:22 King_Hual joined #git
15:25 boxrick1 Howdy, I am trying to get "sharedrepository = 2" to appear in my config when I do a git init, how would I go about that?
15:25 boxrick1 --shared=group  doesnt seem to do it
15:25 p4trix joined #git
15:26 AndreSomers joined #git
15:26 rominronin joined #git
15:26 HappyLoaf_ joined #git
15:27 rnsanchez_ joined #git
15:28 johnmilton joined #git
15:28 boxrick1 Seems it needs to be --shared=all
15:28 raijin joined #git
15:31 einnjo joined #git
15:31 thiago joined #git
15:31 thiago_ joined #git
15:31 kulelu88 joined #git
15:32 AndreSomers joined #git
15:32 BloqueNegro2 left #git
15:33 e14 joined #git
15:33 le_melomane joined #git
15:34 eduardas_m joined #git
15:38 Pyrogerg joined #git
15:39 mangolisk joined #git
15:40 romerocesar joined #git
15:40 durham joined #git
15:42 annoymouse joined #git
15:42 durham joined #git
15:44 ash_workz how do I view the changes for files not yet added?
15:44 t0by joined #git
15:44 ash_workz actually, `git status` just shows a directory (modified content)
15:44 ash_workz so the files should already be there
15:45 Snugglebash joined #git
15:45 ash_workz I mean, that's weird to me
15:46 dreiss joined #git
15:46 ash_workz it should say the actual file, shouldn't it?
15:46 canton7 when you add the file, it'll be shown individually iirc
15:47 davix joined #git
15:49 ash_workz canton7: but it's not new
15:49 ash_workz canton7: it's just modified
15:49 ash_workz cant I view the changes somehow with HEAD?
15:50 canton7 'git diff' will show the changes
15:50 ash_workz git diff HEAD <dirname> just shows that there are changes... not what they are
15:50 _sjs joined #git
15:50 canton7 !repro
15:50 gitinfo [!transcript] Please paste (using https://gist.github.com/ or similar) a transcript ( https://git.io/viMGr ) of your terminal session so we can see exactly what you see
15:51 ash_workz it says
15:51 ash_workz -Subproject commit 3c...
15:51 canton7 oh, it's a submodulie
15:51 canton7 it's not a normal directory
15:51 ash_workz +Subproject commit 3c...-dirty
15:51 canton7 it's an entirely separate git repository, tracked as a submodule
15:51 ash_workz oh
15:51 ash_workz I get it now
15:51 canton7 you'll have to cd into it to see what's changed
15:51 ash_workz no wonder nothing was making any sense
15:51 canton7 remember that submodules have different sets of commits, and different histories, to their parent, etc
15:51 synthroid joined #git
15:52 ash_workz I know nothing of submodules
15:52 qws-user-1229 joined #git
15:52 ash_workz but I will keep that in minde
15:52 ash_workz -e
15:52 canton7 you probably don't want to be using them, then :P
15:52 canton7 if you have a git repository, you can't normally have any .git files in subfolders of it
15:52 ash_workz yeah, I didn't realize I was 1 directory too high
15:52 madewokherd joined #git
15:53 ash_workz my project is confusingly nested
15:53 canton7 funtimes
15:53 g105b joined #git
15:55 dan2k3k4 joined #git
15:55 Pyrogerg joined #git
15:56 LeBlaaanc joined #git
15:58 boxrick1 Can a .git folder reference something which is not in the same location
15:59 boxrick1 Lets assume /opt/.git for /opt/fish/bob
15:59 boxrick1 Or would it always be /opt/fish/bob/.git
16:00 raijin joined #git
16:01 intellix joined #git
16:03 ash_workz canton7: thanks, I was freakin' out because I needed to push this thing that should've been done months ago and I was like, "aaaaaaah"
16:03 canton7 no worries
16:04 troulouliou_div2 joined #git
16:06 ahmedelgabri joined #git
16:08 boxrick1 If I have a .git folder within than a secondary folder, does that mean it references a different file location?
16:09 jost_ joined #git
16:10 Rodya_ joined #git
16:12 preyalone joined #git
16:12 Nd-47-M joined #git
16:12 crayon joined #git
16:12 nettoweb joined #git
16:12 crayon joined #git
16:13 actinide joined #git
16:14 preyalone left #git
16:14 eihabarabia joined #git
16:14 thiago boxrick1: /opt/.git applies to everything in /opt
16:15 thiago boxrick1: if you want to store the .git dir elsewhere, you can. But you have to set the GIT_DIR environment variable. And possibly the GIT_WORKTREE one too, unless you set the tree in the config file.
16:16 boxrick1 I dont specifically want to store it elsewhere its just that in this case I am looking at someone elses setup which seems a bit none standard
16:16 boxrick1 And it is set up in the way I just explained which is odd
16:16 boxrick1 Yet to clone it I need to do git clone /opt/fish/bob
16:17 boxrick1 when the .git dir lives in /opt
16:17 a_thakur joined #git
16:17 * brokensyntax wants to 'mv bob bulb'
16:18 * GodGinrai windows why brokensyntax doesn't just do so
16:18 GodGinrai *wonders
16:18 GodGinrai no idea how I made that typo
16:18 brokensyntax Because it's boxrick1's box.
16:18 brokensyntax LOL, are you on a Windows machine right now?
16:19 GodGinrai lol
16:20 ruby_ joined #git
16:20 boxrick1 Seems its just random .git litter left around
16:20 boxrick1 None of it is actually bob/fish I was just using that as a silly example
16:22 shgysk8zer0 joined #git
16:22 kyan joined #git
16:22 brokensyntax see... fish/bulb just seems like a Simpsons reference to me ;)
16:23 * GodGinrai subscribes to the practice of foo/bar for example writing
16:23 mischat joined #git
16:23 brokensyntax But yeah... cloning is wroking on subdir instead of rootdir of the .git path...
16:24 phaleth joined #git
16:26 VagabondAzulien joined #git
16:26 vls-xy left #git
16:27 EvilPenguin joined #git
16:28 NeverDie joined #git
16:30 gadonj18 joined #git
16:30 dinoocch joined #git
16:30 cyan__ joined #git
16:31 Vaelatern joined #git
16:31 gadonj18 is this the right place to ask for help about what specific strategy would work for my current predicament? I am trying to upgrade an old web application to the newest version but the old version has heavy customizations that I'd like to carry forward if possible
16:32 a_thakur_ joined #git
16:33 rominronin joined #git
16:34 gadonj18 anyone think they could help me?
16:34 Dong__ anything to do with git?
16:34 gadonj18 yeah
16:34 Dong__ in which way?
16:35 Salander27 joined #git
16:35 EricPoe joined #git
16:36 mikecmpbll joined #git
16:36 gadonj18 if I install an off the shelf PHP application, track it in git, and make a bunch of customizations to it, when version 2 comes out it will presumably look like all of my custom code has been removed when git tracks it, how do I maintain those customizations moving forward with new versions of the application
16:37 ruby_ joined #git
16:37 gadonj18 is there a specific workflow or technique that I can google for that kind of situation? I'm pretty much at a loss even trying to articulate it to people
16:38 freimatz joined #git
16:38 ojacobson Start a new branch at the commit matching the version you're upgrading from.
16:38 ojacobson Empty the work tree and install the new version, clean. Commit that on the new branch.
16:38 ojacobson Return to the branch with your customizations.
16:38 ojacobson git-merge the branch with the upgrade.
16:38 ojacobson Resolve conflicts, enjoy.
16:39 ojacobson (Tagging the commits matching upstream clean versions may be helpful in the future)
16:39 Zamyatin joined #git
16:39 gadonj18 is it really that easy? lol
16:39 ojacobson Depends on how many customizations you have, but yes, that's the basics
16:39 gadonj18 someone else was suggesting patches and whatnot and I was super concerned it was going to be more overwhelming
16:39 ojacobson "Resolve conflicts" is a step with a lot of hidden complexity
16:40 ojacobson This approach lets Git compute the appropriate patches
16:41 gadonj18 would the "resolve conflicts" step mark all of my custom code as deleted that I'd then have to manually indicate that I want to keep it?
16:41 actinide_ joined #git
16:41 ojacobson No, it'll do the same 3-way merge that any other merge would do
16:41 Boobuigi joined #git
16:41 gadonj18 alright, I'll create a new test repo and try out that technique and see exactly how things get handled
16:41 ojacobson it'll attempt to keep your code, where there's some automated way to do so, and present merge conflicts where it's not automatable. (You should review the full result, not just the conflicts, because what's automated and what's intended don't always line up.)
16:42 gadonj18 thanks a bunch ojacobson
16:44 f3r70rr35f joined #git
16:44 jorj joined #git
16:44 Sucks joined #git
16:44 rominronin joined #git
16:45 naughtypeaches joined #git
16:45 naughtypeaches joined #git
16:48 rominron_ joined #git
16:49 HappyLoaf_ joined #git
16:49 shafun joined #git
16:51 eihabarabia joined #git
16:52 nitric joined #git
16:52 dreiss joined #git
16:54 fuchstronaut joined #git
16:54 Pyrogerg_ joined #git
16:57 leeN joined #git
16:58 subhojit777 joined #git
16:58 sangy joined #git
16:59 f3r70rr35f joined #git
16:59 thebope_ joined #git
17:00 LionsMane joined #git
17:01 p4trix joined #git
17:01 EvilPeng1 joined #git
17:02 mischat joined #git
17:03 prawnzy joined #git
17:04 SteffanW joined #git
17:04 thiago_ joined #git
17:04 csd_ joined #git
17:05 jaequery joined #git
17:05 csd_ What allows git's tests to call something like `git clone myhost:src ssh-clone` without an error? E.g. https://github.com/git/git/blo​b/master/t/t5601-clone.sh#L350 . I'm guessing some env flag is set but anyone know specifically what?
17:06 _sjs joined #git
17:07 thiago_ csd_: maybe there's a test harness that sets up the environment
17:07 thiago_ it may be setting ~/.ssh/config too
17:07 chachasmooth joined #git
17:07 csd_ thiago_: hard to find it when you dont know exactly what to look for :-/
17:08 thiago_ search for "myhost"
17:08 csd_ doesnt appear outside the that file
17:08 romerocesar joined #git
17:08 mischat_ joined #git
17:09 Pyrogerg joined #git
17:09 DeltaHea_ joined #git
17:09 jaguarmagenta joined #git
17:10 dreiss joined #git
17:10 mingrammer joined #git
17:12 thiago anywhere in the repo?
17:12 thiago not in sources? Not in documentation?
17:13 duderonomy joined #git
17:13 csd_ thiago: nowhere other than that particular bash test file
17:14 DeltaHea_ Would having hundreds of local branch references really slow down git bash on Windows?
17:14 VagabondAzulien joined #git
17:14 mikecmpbll joined #git
17:14 barteks2x joined #git
17:14 DeltaHea_ I'm using the git-for-windows SDK.
17:14 thiago csd_: then it may be a case of "you'll figure it out when you try to run this test"
17:16 Darcidride joined #git
17:16 shgysk8zer0 joined #git
17:17 ansimation joined #git
17:18 refried_ joined #git
17:18 ansimation joined #git
17:19 flobin joined #git
17:20 nikivi joined #git
17:21 VagabondAzulien joined #git
17:22 mingrammer joined #git
17:23 dsantiago joined #git
17:24 kpease joined #git
17:29 timvisher joined #git
17:30 cdrf joined #git
17:30 cdrf left #git
17:31 laumann joined #git
17:32 Wulf joined #git
17:32 Wulf Hello
17:33 Sasazuka joined #git
17:34 Limes__ joined #git
17:34 Wulf I cloned a repo with --bare.
17:35 ojacobson 👏
17:35 Wulf Now when I do fetch --all, it's not actually fetched from the remote. Why not?
17:35 Wulf prints a strange message " * branch            HEAD       -> FETCH_HEAD"
17:36 Hobbyboy joined #git
17:37 NeXTSUN joined #git
17:39 miklcct joined #git
17:39 Rodya_ joined #git
17:39 Pyrogerg joined #git
17:40 tang^ joined #git
17:44 kpease joined #git
17:44 nidr0x joined #git
17:44 sunri5e joined #git
17:46 a_thakur joined #git
17:46 nug700 joined #git
17:47 danslo_ joined #git
17:48 nug700 so I pushed the .gitignore file with the rest of the project in the initial commit, which I though would already follow the ignore rules to what's being pushed. It didn't and pushed everything to the repo.  how do I remove those files from the repo?
17:48 kpease joined #git
17:50 nug700 without removing them from my local project.
17:51 WayToDoor joined #git
17:52 miklcct joined #git
17:53 MadHatter42 joined #git
17:55 m0viefreak joined #git
17:55 fuchstronaut joined #git
17:55 nidr0x joined #git
17:56 star_prone joined #git
17:57 jnavila joined #git
17:58 jberkus joined #git
17:58 shock- joined #git
17:58 jberkus I'm having a strange thing in one repo, where if I change a file and save the changes, I need to "git add" on the file again in order for it to be included in the next commit
17:59 jberkus it seems like I'm in some kind of "don't stage changes mode" by accident; where should I look?
18:01 lolmac joined #git
18:02 multi_io_ joined #git
18:03 oncall-pokemon joined #git
18:04 perlpilot jberkus: I don't understand.  You do have to add changes to the index each time you want them committed.
18:04 timvisher joined #git
18:05 jberkus so, there's a file, index.html, which is already under git
18:05 jberkus I edit the file and save
18:05 jberkus normally, if I do "git commit -a" those changes get picked up in that commit, no?
18:06 mingrammer joined #git
18:06 dviola joined #git
18:06 perlpilot yes, if you use the -a option and git is already tracking the file.
18:06 jberkus right.  so for this one repo, that's not happening
18:06 Hobbyboy joined #git
18:06 jberkus every time I save a change, I have to do "git add" on the file again for it to be picked up
18:07 spacelord joined #git
18:08 perlpilot jberkus: can you produce a transcript of this behavior?
18:09 jberkus perlpilot: ok, so this isn't some kind of mode I set by accident?
18:09 perlpilot Not that I'm aware of.  It sounds really strange.
18:09 jberkus I've certainly never encountered it before.  but it's very consistent
18:09 jberkus and just for this one repo
18:10 jberkus hence thinking I accidentally set some special flag
18:10 jberkus I'll bundle up a transcript and post it to gist
18:10 jberkus later today, thanks
18:10 nitric jberkus: maybe something went sideways in the .git directory, maybe try a fresh clone...?
18:11 ShekharReddy joined #git
18:12 VagabondAzulien joined #git
18:13 perlpilot or, perhaps there's an alias for "git commit" that makes it forget about the -a option just for this repo?
18:17 jimi_ joined #git
18:17 clemf joined #git
18:17 sunri5e joined #git
18:18 HappyLoaf_ joined #git
18:22 MrWoohoo joined #git
18:22 kulelu88 joined #git
18:24 thebope joined #git
18:24 winterchip joined #git
18:24 NeXTSUN joined #git
18:24 gadonj18 how do you normally represent a patch in those git branch diagrams?
18:25 aard_ joined #git
18:26 thebope joined #git
18:26 jaequery joined #git
18:27 nikivi joined #git
18:28 shachaf left #git
18:29 Cata|Work joined #git
18:29 Cata|Work I need help figuring out how to fix a pack file
18:29 Cata|Work I think the issue is a corrupted commit that I tried to remove by doing a reset
18:30 ProbabilityMoon joined #git
18:30 Cata|Work but it appears to still be present as a dangling commit
18:31 nug700 creating git repo rolled back my code somehow.
18:31 nug700 just creating one for it.
18:31 govg joined #git
18:31 michaelsharpe joined #git
18:32 Cata|Work I've run 'git fsck --full' which returned the output in the following pastebin: http://pastebin.com/ZEKEY1WD
18:32 _ikke_ Cata|Work: A dangling commit is usually harmless
18:32 Cata|Work I know
18:33 Cata|Work I have plenty of dangling commits in other repos
18:33 _ikke_ Hmm,  I see
18:33 _ikke_ somehow git cannot unpack it (what it usually does with dangling objects)
18:33 Cata|Work right
18:33 jstimm joined #git
18:34 kpease joined #git
18:34 _ikke_ This needs some more hardcore git-foo, which you can best get through the mailing list
18:34 PCatinean joined #git
18:35 mischat joined #git
18:35 Cata|Work oh boy
18:35 Cata|Work do you have a link to the mailing list?
18:36 nug700 how does creating a git repository for a projet role back the code to a previous version?
18:36 _ikke_ !mailing_list
18:36 gitinfo The mailing list can be reached via git@vger.kernel.org. You don't need to subscribe to the list, you will always be put in cc on reply. Read archives at http://j.mp/gitlist
18:36 crawler joined #git
18:36 _ikke_ nug700: It doesn't
18:36 joelsmith joined #git
18:36 nug700 yes it did
18:36 _ikke_ git init only creates a .git directory with contents
18:36 Cata|Work thanks
18:36 _ikke_ nug700: !transaction
18:36 _ikke_ !transcript
18:36 gitinfo Please paste (using https://gist.github.com/ or similar) a transcript ( https://git.io/viMGr ) of your terminal session so we can see exactly what you see
18:36 EvilPeng1 joined #git
18:37 nug700 I created the repo last night. this morning I'm looking at my code, months old.
18:37 nug700 I'm having to go through my decompiled .Net app to update it
18:38 _ikke_ Why do you think creating the repo did it (and not something else)?
18:38 nug700 because the .Net app was built yestarday, before creating the repo
18:39 _ikke_ How can git even roll back code if it doesn't have any history.. ?
18:39 nug700 unless the windows update did something
18:39 _ikke_ Why not something in visual studio?
18:39 NeverDie joined #git
18:39 _ikke_ It's just that I cannot imagine anything in git causing this
18:41 nug700 so strange
18:42 sssilver joined #git
18:42 timvisher joined #git
18:43 nug700 think it was just one script, thankfully
18:44 crawler joined #git
18:44 EvilPenguin joined #git
18:47 UrsoBranco joined #git
18:48 Salander27 joined #git
18:50 p4trix joined #git
18:51 anti-terrorism joined #git
18:52 h1d joined #git
18:52 satifant joined #git
18:53 brokensyntax joined #git
18:54 WayToDoor joined #git
18:54 kpease joined #git
18:56 rsukla joined #git
18:56 synthroid joined #git
18:56 fuchstronaut joined #git
18:59 Jupelius joined #git
19:00 mingrammer joined #git
19:01 aviraldg joined #git
19:01 aviraldg joined #git
19:01 nettoweb joined #git
19:03 derpsauce__ joined #git
19:04 prawnzy joined #git
19:05 romerocesar joined #git
19:06 mycelium joined #git
19:07 WaReZ joined #git
19:09 aidalgol joined #git
19:10 sangy joined #git
19:10 jaguarmagenta joined #git
19:10 blackwind_123 joined #git
19:10 dreiss joined #git
19:11 cagedwisdom joined #git
19:12 johnmilton joined #git
19:17 QwertyDragon joined #git
19:18 Murii joined #git
19:19 jberkus joined #git
19:19 buzzLightBeer joined #git
19:20 Esya joined #git
19:20 SourTrout joined #git
19:21 stamina joined #git
19:22 sythe joined #git
19:25 snthsnth joined #git
19:26 glebihan joined #git
19:27 yaymukund joined #git
19:32 red82 joined #git
19:33 HappyLoaf_ joined #git
19:35 dsdeiz joined #git
19:35 mablae joined #git
19:37 JamesBaxter joined #git
19:43 Groscheri joined #git
19:43 theonefoo joined #git
19:44 leeN joined #git
19:45 cdown joined #git
19:45 thierryp joined #git
19:47 thebope joined #git
19:47 olalonde joined #git
19:49 Atm0spher1c joined #git
19:49 _ADN_ joined #git
19:49 aax joined #git
19:49 danslo_ joined #git
19:50 ech0s7 joined #git
19:50 gamer joined #git
19:50 jaequery joined #git
19:52 Zamyatin joined #git
19:54 mingrammer joined #git
19:56 hashpuppy joined #git
19:57 hashpuppy i'm looking to get the current tag name or revision if there is no tag name.  i was doing this:  git describe $(git rev-parse HEAD), but someone on stackoverflow said to avoid "porcelain" commands.  is there really anything wrong with that?
19:57 ochorocho__ joined #git
19:58 HappyLoaf_ joined #git
19:59 _ikke_ why add git rev-parse HEAD?
19:59 hahuang62 joined #git
19:59 HoloPed_ joined #git
20:00 hashpuppy just pasting what i saw.  i just assumed the author meant that it doesn't always operate on HEAD
20:00 _ikke_ git name-rev --tags --name-only --always HEAD
20:01 _ikke_ That is the plumbing command that should give it
20:01 hashpuppy i should pass to describe?
20:01 _ikke_ No, just this command
20:01 _ikke_ man git name-rev
20:01 gitinfo the git-name-rev manpage is available at http://jk.gs/git-name-rev.html
20:02 hashpuppy _ikke_: that shows "undefined" when the HEAD has not tag
20:02 prawnzy joined #git
20:02 _ikke_ The idea is that the output for porcelain command can and will change
20:02 hashpuppy ohh
20:02 hashpuppy i see
20:02 hashpuppy weird that always didn't work
20:03 SourTrou_ joined #git
20:03 _ikke_ Hmm, odd, according to the man-page, --always should make it fall back to an object hash
20:04 _ikke_ add --no-undefined
20:04 hashpuppy only seemed to work when i passed in --no-undefined
20:04 m0viefreak --tags forces *only* tags
20:04 _ikke_ m0viefreak: that's the idea
20:04 _ikke_ but --always should make it fall back to commit revs
20:04 m0viefreak oh, i see, sorry
20:05 _ikke_ but apparently only does so with --no-undefined
20:05 cjohnson Is there a way, native to git or otherwise, to do a git log a..b using commit log messages instead of ids?
20:05 _ikke_ cjohnson: man gitrevisions names a method
20:05 gitinfo cjohnson: the gitrevisions manpage is available at http://jk.gs/gitrevisions.html
20:06 wrapash joined #git
20:07 _ikke_ :/commit message
20:07 cjohnson I'm not looking to find one commit, but rather see a list of commits which don't show up in a yet that are in b
20:07 cjohnson I've been rebasing b so the commit IDs won't match up
20:07 cjohnson so I cherry pick the changes into a but somewhere along the way I forgot one
20:07 _ikke_ You can combine these two things
20:08 wrapash Hello there! I want to manage my simple website with git, but really don't kown much about git apart from git clone. I don't seem to have enough google-fu to be able to find a good beginners' guide. Can you give me some pointers?
20:08 bremner !deploy
20:08 gitinfo Git is not a deployment tool, but you can build one around it (in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/deploy.html
20:08 cjohnson _ikke_: you mean in a git command? or write a shell script which loops over each commit and tehn checks git log a to see if that commit exists?
20:08 _ikke_ cjohnson: git log :/"one commit"..:/"Other commit" works for me
20:09 cjohnson there are about 50 commits
20:09 cjohnson sounds like I would have to script
20:09 _ikke_ cjohnson: man git cherry
20:09 gitinfo cjohnson: the git-cherry manpage is available at http://jk.gs/git-cherry.html
20:09 _ikke_ might help
20:10 Wulf joined #git
20:10 utrack joined #git
20:10 cjohnson hrm
20:10 cjohnson interesting
20:10 crawler joined #git
20:10 leeN joined #git
20:10 _ikke_ but it relies on the patch id staying the same
20:11 _ikke_ although, for a missing commit this wouldn't matter
20:11 Faylite joined #git
20:11 cjohnson ah then yeah that won't work because I'm rebasing, no ids remain the same
20:11 _ikke_ no, not commit id
20:11 _ikke_ patch id
20:11 cjohnson oh
20:12 _ikke_ a way to get a stable has of only the patch, the changes it introduces
20:12 _ikke_ so even if the parent and comitter informatiom changes (and so the commit hash), the patch-id would stay the same
20:12 cjohnson nice!
20:12 cjohnson that helped
20:13 _ikke_ cool
20:13 ahmedelgabri joined #git
20:13 cjohnson git cherry a b  listed a whole bunch of - and 2 +
20:14 cjohnson the one I was trying to cherry pick which conflicted, and the one a little higher up which caused it to conflict
20:14 cjohnson applied higher one and now it comes over clean
20:14 _ikke_ yay
20:18 cjohnson so is a patch id determined by basically just hashing the diff?
20:18 cjohnson so if two entirely different commits were to happen to have an identical diff they would share the same patch id
20:19 m0viefreak if they have the same diff -- they *are* the same patch :P
20:19 _ikke_ right
20:19 _ikke_ It does some normalization, but basically, that's it
20:20 cjohnson gotchya
20:20 cjohnson that's cool
20:22 wrapash Thanks bremner
20:25 ochorocho__ joined #git
20:25 Atm0spher1c joined #git
20:26 Balliad joined #git
20:29 HappyLoaf_ joined #git
20:30 dsdeiz joined #git
20:30 dsdeiz joined #git
20:30 danimal2 joined #git
20:33 dsantiago joined #git
20:35 Wulf joined #git
20:36 e14 joined #git
20:37 Dougie187 joined #git
20:39 hahuang62 joined #git
20:40 alex1a joined #git
20:40 star_prone joined #git
20:41 lnbb joined #git
20:43 lmat I would like to rebase my commits, but the history contains merges, too.
20:44 cdg joined #git
20:48 cdg joined #git
20:49 jnavila joined #git
20:49 ochorocho__ joined #git
20:50 dpdv joined #git
20:53 sandroqz joined #git
20:54 MacWinner joined #git
20:54 lnbb joined #git
20:54 shgysk8zer0 joined #git
20:54 acetakwas joined #git
20:55 sandroqz joined #git
20:57 ochorocho__ joined #git
20:58 _ikke_ That's not recommended, but possible with --preserve-merges
20:59 fuchstronaut joined #git
20:59 ahr3n joined #git
20:59 lmat _ikke_: Thanks!
20:59 lmat How can I tell what commits will be rebased when i run    git rebase otherbranch;
21:01 nds_ joined #git
21:02 _ikke_ lmat: git log --oneline otherbranch...
21:02 _ikke_ (three dots)
21:03 lmat _ikke_: Ahh I was thinking somethingl like   git refspec HEAD..otherbranch
21:04 whitby joined #git
21:04 refried_ joined #git
21:06 h1d joined #git
21:07 nd joined #git
21:07 romerocesar_ joined #git
21:07 jberkus joined #git
21:08 nilg joined #git
21:10 []\[] joined #git
21:11 jaguarmagenta joined #git
21:11 snthsnth joined #git
21:11 bernardio joined #git
21:11 elastix joined #git
21:13 Atm0spher1c joined #git
21:13 michaelsharpe joined #git
21:15 anonymuse joined #git
21:16 crawler joined #git
21:17 romerocesar joined #git
21:17 Blkt joined #git
21:19 mingrammer joined #git
21:20 Darren_ joined #git
21:20 demize joined #git
21:20 mattcen joined #git
21:20 demize Hmm, interesting bug, doing a git command with sudo and then accidentally running one without is giving me a "fatal: BUG: get_tempfile_fd() called for inactive object"
21:21 demize Oh, no, it's just doing git-pull that causes it.
21:21 demize (Without sudo)
21:21 gambler joined #git
21:23 mischat joined #git
21:23 hexagoxel joined #git
21:24 DeltaHeavy left #git
21:24 [ADN] joined #git
21:26 lamont joined #git
21:26 RikkiMongoose joined #git
21:28 hexagoxel joined #git
21:28 yaymukund joined #git
21:28 theonefoo joined #git
21:28 e14 joined #git
21:29 treehug88 joined #git
21:30 hexagoxel joined #git
21:32 Zamyatin joined #git
21:32 dsdeiz joined #git
21:32 dsdeiz joined #git
21:33 e14 joined #git
21:34 UTAN_dev_ joined #git
21:36 d5sx43 joined #git
21:37 jimi_ joined #git
21:37 jimi_ joined #git
21:38 e14 joined #git
21:39 danslo joined #git
21:39 romerocesar joined #git
21:40 HappyLoaf_ joined #git
21:40 rnsanchez_ joined #git
21:41 romerocesar_ joined #git
21:41 a_ joined #git
21:41 ochorocho__ joined #git
21:42 cdown joined #git
21:44 a3Dman joined #git
21:45 UTAN_dev joined #git
21:47 Dougie187 left #git
21:47 aidalgol joined #git
21:54 jeffreylevesque joined #git
21:56 dodobrain joined #git
21:59 theonefoo left #git
22:00 ochorocho__ joined #git
22:02 snowkidind joined #git
22:03 HappyLoaf_ joined #git
22:06 dendazen joined #git
22:08 mustmodify___ joined #git
22:08 mustmodify___ left #git
22:09 mustmodify___ joined #git
22:10 aarobc joined #git
22:12 mischat joined #git
22:13 mustmodify Let's say I'm at revision A. I pull down revision B, then try to add 'merge' revision C, and some other stuff D, and then I'm like, "This is bogus." And I want to create a commit that just takes us back to A, overwriting B because it's stupid. Is there an easy way to do that other than reverting the three commits? it doesn't want to let me revert a merge.
22:13 mischat joined #git
22:13 mischat joined #git
22:14 mingrammer joined #git
22:14 michaelsharpe joined #git
22:14 mischat joined #git
22:15 prawnzy joined #git
22:17 chachasmooth joined #git
22:19 thebope joined #git
22:22 crawler joined #git
22:23 Atm0spher1c joined #git
22:23 mycelium joined #git
22:24 rbern joined #git
22:26 naughtypeaches joined #git
22:28 Wulf joined #git
22:29 Milijus joined #git
22:30 madewokherd joined #git
22:30 mrpackethead joined #git
22:31 xaa__ joined #git
22:34 jberkus joined #git
22:34 Guest66710 left #git
22:35 cyphase joined #git
22:36 cyphase joined #git
22:37 Sasazuka__ joined #git
22:38 hexagoxel mustmodify: sounds like you might want some 'reset' variant. not sure what "pull down" means and what exactly you want to revert or keep though.
22:38 gambler joined #git
22:39 timewalker joined #git
22:39 sangy joined #git
22:40 danslo joined #git
22:40 anuxivm joined #git
22:40 cyphase joined #git
22:40 dilijev_ joined #git
22:40 cyphase joined #git
22:41 dilijev_ I just upgraded to Git 2.10 and read the changelog -- was excited to see: "log --graph --format=" learned that "%>|(N)" specifies the width relative to the terminal's left edge. ... Negative N ... means the column limit is relative to the right border.
22:41 dilijev_ I tried adding this to my format string and haven't had much success
22:41 HappyLoaf_ joined #git
22:42 Zenrium left #git
22:42 StormLrd joined #git
22:43 raijin joined #git
22:44 mustmodify left #git
22:45 thebope joined #git
22:48 harish joined #git
22:48 dilijev_ Oh I see, it looks like on windows it's just assuming the right margin of the console is 80 columns
22:49 dilijev_ is there any way to make this be detected dynamically?
22:50 dsdeiz_ joined #git
22:51 kyan joined #git
22:51 kyan joined #git
22:53 Derperperd joined #git
22:53 kisonecat joined #git
22:56 Milijus joined #git
22:56 cdown_ joined #git
22:57 csd__ joined #git
22:57 cspack joined #git
22:58 ResidentBiscuit joined #git
23:00 fuchstronaut joined #git
23:01 thiago_ joined #git
23:01 romerocesar joined #git
23:03 aax joined #git
23:03 crayon joined #git
23:07 flobin joined #git
23:08 anuxivm left #git
23:09 laumann joined #git
23:11 Eugene dilijev_ - git passes it off to $PAGER, which you can customize to anything you like
23:11 Eugene I suspect that less.exe isn't smart enough to detect a modern(non-fixed-width) terminal on Windows, because that's such a new invention
23:11 dilijev_ I assume $PAGER is less.exe in the Git tree on Windows
23:11 jaguarmagenta joined #git
23:11 dilijev_ yeah
23:12 HappyLoaf_ joined #git
23:12 Eugene Use a smarter pager or submit a bug/patch to msysgit and/or msys
23:12 dilijev_ any suggestions for smarter pagers?
23:13 Eugene Nope, less has always worked good for me
23:13 movetodevnull joined #git
23:15 michaelsharpe joined #git
23:15 mischat joined #git
23:18 _rnsanchez_ joined #git
23:20 movetodevnull joined #git
23:20 thebope joined #git
23:21 dodobrain joined #git
23:22 laumann joined #git
23:22 mablae_ joined #git
23:23 rpg joined #git
23:24 rpg Is there a git workflow for "I'd like a local branch that has multiple topic branches merged into it?" For keeping a locally configured version with multiple enhancements that are under review?
23:24 rpg Pointer would be fine -- this smells like a FAQ, but my google fu wasn't strong enough to find a solution.
23:26 sdothum joined #git
23:26 howitdo joined #git
23:29 yaymukund joined #git
23:29 CrazyBonz joined #git
23:31 cyphase joined #git
23:31 fedepad joined #git
23:32 azerus joined #git
23:32 t0by joined #git
23:32 buzzLightBeer joined #git
23:34 sdothum- joined #git
23:34 Synthead joined #git
23:35 HappyLoaf_ joined #git
23:35 bronson joined #git
23:36 _ADN_ joined #git
23:38 crayon joined #git
23:38 crayon joined #git
23:41 calloc joined #git
23:41 danslo joined #git
23:42 NeXTSUN joined #git
23:43 sin_ joined #git
23:43 einnjo joined #git
23:44 calloc "git fetch origin; git reset --hard origin/master" does not seem to leave me with a clean index
23:45 calloc How would I revert my master to be exactly like the remote master
23:45 calloc ?
23:47 rewt man git clean
23:47 gitinfo the git-clean manpage is available at http://jk.gs/git-clean.html
23:47 calloc I ran "git clean -xdf" as well.
23:53 cyphase joined #git
23:54 mizu_no_oto joined #git
23:55 calloc I pasted the output here: "https://ptpb.pw/4KKi"
23:56 star_prone joined #git
23:57 egorsmkv joined #git
23:59 rewt`` joined #git
23:59 canuk joined #git

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