Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-09-09

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

All times shown according to UTC.

Time Nick Message
00:04 sangy joined #git
00:06 aapaaal joined #git
00:06 MattMaker joined #git
00:07 romerocesar joined #git
00:07 ochorocho__ joined #git
00:08 aapaaal joined #git
00:09 johnmilton joined #git
00:19 kaldoran joined #git
00:20 shauna joined #git
00:22 SourTrout joined #git
00:25 rafalcpp joined #git
00:25 Geo joined #git
00:26 inflames joined #git
00:26 Geo If I create a branch and make/commit several changes to ten files, and now I realize I only want to keep changes to one file- is there an easy way to do that?
00:27 kaldoran_ joined #git
00:29 kadoban Geo: Are most of the commits separated by file that they changed, or is it all mixed up?
00:30 Geo mixed up
00:31 kadoban Geo: Either do 'git rebase -i' and manually edit each (use checkout to remove changes you don't want), or you could like ... make a copy of the branch, and then use filter-branch to edit each so they only change the one file (can write a simple script for that), and then rebase the commits you want onto the branch you want.
00:31 blaster joined #git
00:31 kadoban If there's less than a ton of commits, the first is likely saner.
00:32 kadoban If there's merges anywhere in there ... it gets more complicated.
00:32 Geo no merges
00:32 Geo but that first seems doable
00:32 Geo most of the 'other' changes are in a single commit
00:33 Geo thank you
00:33 gargola_ joined #git
00:33 kadoban 'welcome
00:33 kadoban rebase -i is cool as hell once you get used to it
00:34 mingrammer joined #git
00:35 tesuji_ joined #git
00:42 SourTrout joined #git
00:44 r3c0d3x joined #git
00:46 SourTrout joined #git
00:47 rominronin joined #git
00:47 ansimation joined #git
00:52 shmuli joined #git
00:52 CussBot joined #git
00:52 anonymuse joined #git
00:56 EvilPenguin joined #git
00:59 CussBot joined #git
00:59 mizu_no_oto joined #git
01:00 e14 joined #git
01:02 mizu_no_oto joined #git
01:02 arescorpio joined #git
01:06 relipse joined #git
01:06 shmuli joined #git
01:08 ochorocho__ joined #git
01:12 Dispatch joined #git
01:15 cqi joined #git
01:21 shmuli joined #git
01:21 rbern joined #git
01:23 anuxivm left #git
01:24 pks joined #git
01:25 moneylotion joined #git
01:28 mingrammer joined #git
01:29 mingrammer joined #git
01:31 dsdeiz joined #git
01:31 dsdeiz joined #git
01:32 madewokherd joined #git
01:33 nettoweb joined #git
01:34 _sjs joined #git
01:34 CussBot joined #git
01:36 blackwind_123 joined #git
01:37 duderonomy joined #git
01:37 citizen-stig joined #git
01:37 theskillwithin joined #git
01:39 evanwang joined #git
01:46 HoierM joined #git
01:46 d^sh joined #git
01:46 Vampire0_ joined #git
01:49 ilbot3 joined #git
01:49 Topic for #git is now Welcome to #git, the place for git help and NP-hard problems | Public logs at http://goo.gl/BuUi5o | Current stable version: 2.10.0 | First visit? Read: http://jk.gs/git | Getting "cannot send to channel"? /msg gitinfo .voice | Now with more indirect cycling giraffes!
01:52 charlesr joined #git
01:53 red82 joined #git
01:53 fakenerd joined #git
01:55 durham_ joined #git
01:55 SourTrout joined #git
01:57 ksefchik joined #git
01:58 ruby_ joined #git
02:00 israelzuniga joined #git
02:02 romerocesar joined #git
02:05 tax joined #git
02:05 mizu_no_oto joined #git
02:06 constant joined #git
02:09 hahuang61 joined #git
02:10 Dougie187 joined #git
02:11 ochorocho__ joined #git
02:11 krs93 joined #git
02:12 dmto joined #git
02:15 hexagoxel joined #git
02:15 mizu_no_oto joined #git
02:18 fstd_ joined #git
02:19 boombatower joined #git
02:19 ruby_ joined #git
02:19 Dispatch joined #git
02:20 binaryc joined #git
02:20 perlpilot joined #git
02:21 mycelium joined #git
02:21 fahadash joined #git
02:23 Catachan_ joined #git
02:23 learath2 joined #git
02:24 Goplat joined #git
02:24 DolpheenDream joined #git
02:25 mar77i joined #git
02:25 dreiss joined #git
02:26 A5101 joined #git
02:26 ojdo joined #git
02:27 cyphase joined #git
02:29 tesuji joined #git
02:30 david007 joined #git
02:32 ozymandias_ joined #git
02:32 Raed|Mobile joined #git
02:33 spike joined #git
02:34 mizu_no_oto joined #git
02:34 shmuli joined #git
02:34 HappyLoa- joined #git
02:34 lagothri1 joined #git
02:35 spike what's the right time for using a git submodule? I have a project I'm working on where I need to use static dependencies, which means whenever I want to use a python library off github I need to "download as zip" and include in the project. problem is that means any time the library is updated i need to go back to its repo. is that the kind of use case where setting it up as a submodule would be useful?
02:35 spike I could be entirely wrong too...
02:36 ash_workz joined #git
02:36 Dispatch joined #git
02:38 JamesBaxter joined #git
02:38 david007 joined #git
02:38 bket joined #git
02:38 chrisshattuck joined #git
02:40 fakenerd joined #git
02:42 cyphase joined #git
02:49 rominronin joined #git
02:49 dstarh joined #git
02:52 mycyber joined #git
02:52 watabou joined #git
02:53 mizu_no_oto joined #git
02:54 zmachine joined #git
02:54 cqi joined #git
02:54 diogenese joined #git
02:56 chachasmooth joined #git
02:56 diogenese joined #git
02:57 binaryc left #git
02:58 cyphase joined #git
02:59 SourTrout joined #git
03:04 mizu_no_oto joined #git
03:08 eizua joined #git
03:11 mizu_no_oto joined #git
03:12 texasmynsted joined #git
03:13 cyphase joined #git
03:13 ruby_ joined #git
03:15 Vortex34 joined #git
03:16 Dougie187 left #git
03:17 chrisshattuck joined #git
03:20 rchavik joined #git
03:20 t4nk249 joined #git
03:22 t4nk249 git trouble: i'm having a hard time committing to my own personal repo with my personal account on a computer that contains my work account
03:22 cdg joined #git
03:22 t4nk249 i've changed `git config user.name` and `git config user.email` to the personal account info
03:23 t4nk249 but any commit still states "Permission to personal/repo.git denied to work"
03:23 t4nk249 banging my head against the wall at this point, though that isn't helping as much as i would hope
03:23 kadoban t4nk249: On what command? Sounds like filesystem permission issues.
03:24 mingrammer joined #git
03:24 kadoban Oh you said commit ... yeah likely you don't have permission to write the right files.
03:24 t4nk249 `git push`
03:24 kadoban Oh. What is the exact message?
03:25 t4nk249 remote: Permission to PERSONALACCOUNT/REPO.git denied to WORKACCOUNT.
03:25 t4nk249 PERSONALACCOUNT = <my personal account>; WORKACCOUNT = <my work account>
03:26 kadoban t4nk249: Sounds like a hook ... really depends what they're doing.
03:26 t4nk249 `git config user.name` == PERSONALACCOUNT
03:26 chrisshattuck joined #git
03:26 citizen-stig joined #git
03:26 kadoban t4nk249: It would almost certainly not be depending on that. That just determines what name and email address gets put in commits you write.
03:26 kadoban That would be insecure as hell, nobody would be using it for permissions.
03:27 kadoban It doesn't change how push works either, you'd have to completely rewrite commits to change that. Almost certainly not what's required.
03:27 cyphase joined #git
03:27 govg joined #git
03:27 kadoban t4nk249: What git server is this? Is it a private one? One of the public ones?
03:28 t4nk249 public
03:28 kadoban Which?
03:28 kadoban How are you connecting, via SSH? Other?
03:28 mizu_no_oto joined #git
03:29 t4nk249 have tried both https & ssh to https://github.com/PERSONALACCOUNT/REPO.git and git@github.com:PERSONALACCOUNT/REPO.git
03:30 kadoban t4nk249: github determines who you are based on what SSH keypair you're using.
03:31 kadoban If you have two separate accounts, you should probably use two separate aliases for github (in ~/.ssh/config typically), each specifying a different keypair. Then use those aliases in git, whichever is appropriate.
03:32 pe3z joined #git
03:32 t4nk249 I have created a config file in ~/.ssh that includes "Host personal" and "Host work" with respective "IdentityFile"s
03:33 kadoban t4nk249: And are you using the "personal" and "work" names in git, instead of github.com?
03:33 kadoban t4nk249: What is 'git remote -v'?
03:33 mehola joined #git
03:34 pe3z left #git
03:34 kadoban t4nk249: Also check 'ssh work' and 'ssh personal'  it should spit back what user it thinks you are, make sure they match up correctly.
03:34 kadoban (it'll be an error message because it won't give you a tty, but it'll tell you who it thinks you are anyway)
03:35 t4nk249 oh no...`ssh work` and `ssh personal` now both say "Hi personal!"
03:35 aavrug joined #git
03:35 t4nk249 i've used this PC for work for a long time
03:35 t4nk249 it didn't know anything about "personal" until tonight
03:36 navidr joined #git
03:36 kadoban Are the aliases set up with separate IdentityFile? Are the actual files different?
03:37 kadoban You could check in github that it has the correct public key for both too I guess, but I don't think it'll actually let you give two different accounts the same public key.
03:37 t4nk249 yep; personal = "IdentityFile ~/.ssh/id_rsa" and work = "IdentityFile ~/.ssh/github_rsa"
03:38 kadoban And the files actually differ, you didn't overwrite them or something, right?
03:38 kadoban You can check 'ssh -v work' and see what's up with that, it'll tell you somewhere in there what auth it's actually doing
03:39 kadoban "Offering RSA public key" or something
03:39 mingrammer joined #git
03:39 vuoto joined #git
03:40 cyphase joined #git
03:40 t4nk249 they "look" different; personal file is ~twice as long as work
03:40 kadoban Hmm. Ya I guess check ssh -v? That sounds weird so far though.
03:42 t4nk249 `ssh -v work` tells me "debug1: Offering RSA public key: .\\id_rsa" as I would expect
03:43 t4nk249 oh, wait...that's not what i would expect...
03:43 kadoban t4nk249: Yeah, that's the opposite of above
03:44 kadoban Does it give any errors for your config file maybe? Maybe something is malformed and it's just ignoring it?
03:46 t4nk249 ok...i have fixed that...but my issue persists in trying to push to PERSONAL but it still sees me as WORK
03:46 t4nk249 (thanks for all of your help so far by the way)
03:47 t4nk249 how do i tell it that I am PERSONAL and not WORK?
03:48 t4nk249 i thought `git config user.name 'PERSONAL'` did that but, as you mentioned, that would be incredibly insecure
03:48 jeffreylevesque joined #git
03:50 rominronin joined #git
03:51 kadoban t4nk249: It's entirely based on SSH keys, yeah. So does 'ssh work' and 'ssh personal' work correctly now?
03:52 kadoban If so, we should move on to 'git remote -v', likely it's not using the right alias
03:52 cyphase joined #git
03:54 t4nk249 i...i actually have no idea if `ssh work` and `ssh personal` are working properly
03:55 kadoban t4nk249: Hmm. It tells you what username it thinks you are for each, right? It should be near the start, like uhm "Hi whatever!"
03:55 t4nk249 `ssh -v work` tells me "debug1: Offering RSA public key: /c/Users/me/.ssh/github_rsa" which is correct
03:56 t4nk249 but `ssh work` says "Permission denied (publickey)."
03:56 diogenese joined #git
03:57 thiago joined #git
03:57 kadoban t4nk249: Ah, interesting. On your work github account, have you set up the SSH public key? It should be in the settings on the web interface.
03:57 t4nk249 `ssh -v personal` = "debug1: Offering RSA public key: /c/Users/me/.ssh/id_rsa"
03:57 kadoban You should try without the -v too, just to be sure, it's pretty verbose, you might be losing the message from the server in all the nonsense.
03:57 dmj joined #git
03:58 t4nk249 yes; work was working fine - has been for years with the SSH key
03:58 t4nk249 i think i should probably try to push to a repo that worked before i started mucking with this personal stuff
03:58 kadoban t4nk249: Oh, did you set the user in the .ssh/config for both? It should be using the 'git' user
03:58 t4nk249 literally "git"?
03:58 kadoban Permissing denied (publickey) I think means nothing at all will ever work on github.
03:59 kadoban Yep, literally git. Everyone logs into github with that username and then it figures out who you actually are by publickey
03:59 kadoban (via ssh anyway)
04:00 kadoban https://gist.github.com/kadoban/2fa6144d68fd17b32150278b83963da2 so like that's my github section of my ~/.ssh/config, if it helps.
04:00 dermoth joined #git
04:01 kadoban I don't think the last bit actually matters, just makes it not ask me for a password in case my public key fails or whatever, which shouldn't ever come up?
04:03 t4nk249 ok...i'm pretty sure i didn't do anything to mess with my "work" key, but just to eliminate that from the equation i generated a new ssh key for work
04:03 t4nk249 i added that to github web settings
04:03 t4nk249 also updated config
04:04 t4nk249 ssh work STILL says "Hi PERSONAL!"
04:04 kadoban And with -v is it offering the correct keypair?
04:05 t4nk249 no it isn't
04:06 ochorocho__ joined #git
04:06 kadoban Felt like we were getting somewhere, darn :-/ Anything suspicious in the output before that? Can you lpaste the part of your alias config maybe? You have saved the file right? I've forgotten to do that sometimes.
04:06 cyphase joined #git
04:07 t4nk249 https://gist.github.com/kadoban/2fa6144d68fd17b32150278b83963da2
04:07 kadoban Are you using an SSH agent? What happens if you "ssh-add whateverkey" first, does that fix it? Though I think it should ask you instead of just giving a random other key so that's kind of a longshot.
04:07 krillr joined #git
04:09 Cabanossi joined #git
04:09 star_prone joined #git
04:11 thiago I think ssh offers all keys that are in the agent, if there's no default one
04:11 t4nk249 omg
04:11 t4nk249 i think that fixed that part
04:12 kadoban thiago: An alias overrides that though, right? Or is that what you mean by "default one"?
04:12 t4nk249 `ssh work` = "Hi WORK!"
04:12 kadoban I mean IdentityFile overrides that.
04:12 t4nk249 `ssh personal` = "Hi PERSONAL!"
04:12 kadoban t4nk249: Yay!
04:12 thiago the alias sets the default
04:12 kadoban Okay, right, that matches my understanding.
04:12 t4nk249 "ssh-add whateverkey" got me by that one
04:13 star_prone joined #git
04:13 thiago I think the key needs to be loaded in the agent, though
04:13 kadoban It doesn't ask you if it's not?
04:13 thiago otherwise, it offers first the keys in the agent, then it tries to load the file
04:13 Geo left #git
04:13 kadoban Wow ... that sounds kinda broken. There's no way to get it to not do that?
04:16 t4nk249 not to be greedy, but back to the original issue
04:16 kadoban t4nk249: So yeah, if both of those are working, then see what 'git remote -v' says. It should be using your alias, not github.com
04:17 t4nk249 `git remote -v` is currently "https://PERSONAL@github.com/PERSONAL/REPO.git"
04:18 t4nk249 changed that to the SSH
04:18 t4nk249 and it pushed
04:18 t4nk249 alrighty then...
04:19 t4nk249 could have sworn i tried that before
04:20 cyphase joined #git
04:21 t2mkn joined #git
04:21 t4nk249 thanks so much for your patience and help!
04:22 kadoban Yay. 'welcome, glad it worked.
04:22 realz joined #git
04:22 kadoban The syntax I use, by the way is just uhm   myalias:kadoban/whatever
04:23 kadoban I find the actual ssh:// URL syntax kind of confusing for whatever reason
04:25 checkItOut_ joined #git
04:29 imacinto joined #git
04:33 hahuang61 joined #git
04:38 phanimahesh joined #git
04:43 cyphase joined #git
04:43 star_prone joined #git
04:44 subhojit777 joined #git
04:45 Andrew_K joined #git
04:49 romerocesar joined #git
04:51 rominronin joined #git
04:51 Saitaku joined #git
04:56 Saitaku left #git
04:56 greendo joined #git
04:56 daey_ joined #git
04:58 dec0n joined #git
05:01 Bl4ckpajamas joined #git
05:02 israelzuniga joined #git
05:03 cyphase joined #git
05:09 dreiss joined #git
05:10 superteece joined #git
05:10 nwkj86 joined #git
05:15 citizen-stig joined #git
05:20 cyphase joined #git
05:21 star_prone joined #git
05:22 justan0theruser joined #git
05:22 dreamyspell joined #git
05:28 jungsubk joined #git
05:34 bocaneri joined #git
05:36 MattMaker joined #git
05:37 hahuang61 joined #git
05:42 ThomasLocke joined #git
05:42 ThomasLocke joined #git
05:43 nidr0x joined #git
05:44 Marbug_ joined #git
05:44 krs93 joined #git
05:45 romerocesar joined #git
05:46 cyphase joined #git
05:49 MattMaker joined #git
05:50 fakenerd_ joined #git
05:51 jceb joined #git
05:52 rominronin joined #git
05:53 chele joined #git
05:53 qt-x joined #git
05:56 PCatinean joined #git
05:56 ShekharReddy joined #git
05:59 freimatz joined #git
06:00 rscata joined #git
06:03 cyphase joined #git
06:03 L-11 joined #git
06:04 pouet joined #git
06:04 romerocesar joined #git
06:07 ochorocho__ joined #git
06:13 JeroenT joined #git
06:14 romerocesar joined #git
06:17 roelmonnens joined #git
06:18 MattMaker joined #git
06:21 cyphase joined #git
06:23 mycelium joined #git
06:24 djb-irc joined #git
06:26 d0nn1e joined #git
06:30 jknetl joined #git
06:31 citizen-stig joined #git
06:34 saedelaere joined #git
06:35 yuhlw_ joined #git
06:35 cyan__ joined #git
06:36 Sceorem joined #git
06:36 Kurvivor joined #git
06:38 cyphase joined #git
06:40 imacinto joined #git
06:40 MattMaker joined #git
06:41 JeroenT joined #git
06:43 lordjancso joined #git
06:43 star_prone joined #git
06:44 jiahao joined #git
06:46 jiahao left #git
06:46 eduardas_m joined #git
06:46 Silenced joined #git
06:47 satya4ever joined #git
06:49 Silenced joined #git
06:50 aidalgol joined #git
06:53 imacinto joined #git
06:53 zeroed joined #git
06:53 rominronin joined #git
06:54 Brain joined #git
06:56 submitnine joined #git
06:56 MattMaker joined #git
06:56 apotry joined #git
06:56 Kurvivor hello! What should i do to make a new repository from existing one?
06:56 Kurvivor i want to base new one on old, but i do not need its history
06:58 cyphase joined #git
06:59 kadoban Kurvivor: Copy the directory, delete the .git and 'git init' I guess?
06:59 ruby_ joined #git
06:59 roelmonn_ joined #git
07:00 MattMaker joined #git
07:01 DieguezZ joined #git
07:01 Kurvivor kadoban: huh, that's a way to do that, yes
07:02 kadoban If you really want to do it within git, I guess you could clone and then 'git checkout --orphan whatever' and go from there, etc.
07:02 kadoban It'd require some fiddling with branch names most likely.
07:04 roxlu_ joined #git
07:04 watabou joined #git
07:05 imacinto joined #git
07:07 ochorocho__ joined #git
07:10 Darcidride_ joined #git
07:11 star_prone joined #git
07:14 snowkidind joined #git
07:18 elect joined #git
07:22 t0by joined #git
07:22 imacinto joined #git
07:23 sevenfourk joined #git
07:27 sarink_ joined #git
07:27 Balliad joined #git
07:31 fefa2k left #git
07:31 eduardas_m joined #git
07:35 clmsy joined #git
07:37 imacinto joined #git
07:37 roelmonnens joined #git
07:37 cyphase joined #git
07:39 hahuang61 joined #git
07:39 roelmonn_ joined #git
07:42 pouet joined #git
07:43 clime joined #git
07:43 GavinMagnus joined #git
07:46 hhfemh joined #git
07:47 danslo_ left #git
07:47 clime left #git
07:47 danslo joined #git
07:52 jceb joined #git
07:52 freimatz joined #git
07:53 encod3 joined #git
07:56 mikecmpbll joined #git
07:57 Oebele joined #git
07:57 cyphase joined #git
07:59 pouet joined #git
07:59 MattMaker joined #git
08:00 twizsted joined #git
08:00 sanjoyd left #git
08:01 aax joined #git
08:03 ahmed_elgabri joined #git
08:04 TomyWork joined #git
08:09 GavinMagnus left #git
08:09 lolmac joined #git
08:09 watabou joined #git
08:10 kurkale6ka joined #git
08:12 star_prone joined #git
08:13 chll_ joined #git
08:15 marcogmonteiro joined #git
08:19 roelmonnens joined #git
08:20 citizen-stig joined #git
08:20 Rodya_ joined #git
08:24 MadHatter42 joined #git
08:25 evie_hammond joined #git
08:27 imacinto joined #git
08:29 bongjovi joined #git
08:29 MattMaker joined #git
08:31 laumann joined #git
08:31 jceb joined #git
08:32 spinningarrow The other option would be to `git clone` the repo and then remove the `.git` folder inside the repo :P
08:33 fees joined #git
08:34 checkItOut joined #git
08:37 kulelu88 joined #git
08:38 groundnuty hey, I'm tryint to delete this commit form my repo without any success
08:38 groundnuty https://github.com/onedata/onedata-documentation/commit/7c57752b13d52acc604969c6350bb89caa92703d
08:39 groundnuty It was no parent and to my knowledge is not references ny anyting
08:39 groundnuty gc does'nt wante to remove it
08:39 groundnuty *it has no parent
08:39 osse groundnuty: gc is a local thing. you'll have to wait for github's own gc to kick in
08:40 groundnuty but I can't even do it localy
08:40 osse git gc --prune=all
08:41 groundnuty still there, I've spent few hourse on it :)
08:41 groundnuty tried this, rebasing...
08:41 jast go through the 'checklist for shrinking a repository' in man git filter-branch
08:41 gitinfo the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html
08:41 groundnuty partly because I want to remove ti, partely as an excercise
08:41 osse gitinfo: is it referenced by the reflog maybe?
08:42 jast some of the points may not be relevant to your situation
08:42 groundnuty osse: any nice command that would show mw that?
08:42 BrianBlaze420 joined #git
08:42 groundnuty jast: I will read it, thx
08:43 vano joined #git
08:43 jast be aware that expiring the reflogs means you can't use them to restore anything else from the past, either :}
08:43 osse groundnuty: is that commit listed by 'git fsck' as unreachable?
08:44 jast you'll want to use git fsck --unreachable, I think
08:44 MattMaker joined #git
08:44 groundnuty osse: hmm.. no!
08:44 groundnuty that's a clue
08:45 vano How does one go about this? I have a config file which is in git and that I want it present so when it is cloned, the default values are used. However in a specific clone I want to change this config file, but I don’t want it version controlled in that clone, because I don’t want to override the default values in the git repository. What is a good solution for this case?
08:45 jast it's very likely a reflog is still referencing your commit
08:45 jast check 'git log -g' and 'git log -g yourbranch' for mentions of the commit ID
08:45 hahuang61 joined #git
08:45 osse groundnuty: then it is reachable from a reflog. try again with git fsck --no-reflogs. is it listed now?
08:45 vano Forgot to mention that I also version control my .gitignore, so changing that doesn’t seem ideal either.
08:46 jast vano: you can use git's sparse checkout/skip-worktree feature. see the section about sparse checkout in man git-read-tree
08:46 gitinfo vano: the git-read-tree manpage is available at http://jk.gs/git-read-tree.html
08:46 Ipe_ joined #git
08:47 Macaveli joined #git
08:48 bauruine joined #git
08:48 vano thank will look into it
08:49 jast .trigger_edit sparse_checkout Sparse checkout can be used to restrict which files/folders are updated with all checkout and status operations. It cannot change the tree the checkout is rooted from. One common use case is locally keeping Git from updating some of the files tracked in your repository. See the "Sparse Checkout" section in man git-read-tree (http://jk.gs/git-read-tree.html).
08:49 gitinfo jast: Okay.
08:51 citizen-stig joined #git
08:52 SwiftMatt joined #git
08:52 knittl is there a reverse of filter-branch --subdirectory-filter? I want to add a dir-prefix to all files. is there an easy way?
08:52 knittl hi jast :]
08:53 imacinto joined #git
08:53 imacinto joined #git
08:53 groundnuty osse: fsck --no-reflogs does not list it, not git reflog
08:54 StatutoryApe My build scripts use a tool called GitVersion to determine the SemVer-compliant version for a commit based on things like the number of commits since the last merge to master.  Upon successful build, my scripts then push a tag to the repo with that version.
08:54 groundnuty jast: I can see it in git --no-pager log  --all  --format=raw
08:54 MattMaker joined #git
08:54 groundnuty and --graph option shows it's not connected to anything
08:55 Oebele joined #git
08:55 StatutoryApe However, GitVersion will not generate a "release" version from the master build by itself, so I have to tag a merge or commit in master with the appropriate version before the regular build script kicks off.
08:55 jast groundnuty: that's different, the -g option I mentioned views reflogs instead of history. according to your commands it shows up even in your history, i.e. it's still referenced by some branch or tag...
08:55 jast try git branch --contains <commit> and git tag --contains <commit>
08:56 StatutoryApe On Tuesday I worked out a powershell script to determine the correct version for a merge based on the tags on the commits that have been merged.
08:56 Oebele joined #git
08:57 StatutoryApe My remaining questions now are: a) How do I determine whether a commit in master is a merge or not? b) If it is not a merge, how do I get a chronological list of commit hashes in master since the last merge to master?
08:57 groundnuty jast: so, git log -g does not list it
08:58 groundnuty jast: git branch --contains return nothing
08:58 jast hi knittl :) no there isn't. you could do it using --commit-filter but it requires some knowledge of stuff like 'git write-tree'
08:58 knittl ah. found it in the filter-branch manpage
08:58 knittl »To move the whole tree into a subdirectory, or remove it from there: […]«
08:58 groundnuty jast:  git tag --contains also nothing
08:58 knittl uses git-ls-files
08:59 jast yeah, that'll do it, too, but it will probably be slightly slower
08:59 jast on the plus side you can just copy & paste :)
08:59 knittl index-filter should be fast enough (and it's only 100 commits)
08:59 osse groundnuty: references by the stash maybe
08:59 jast right
09:00 jast stash is a reflog, I wouldn't expect it to be included in  --all
09:00 star_prone joined #git
09:00 jast I suppose it has a ref, too...
09:00 _ikke_ There is one stash ref
09:00 osse refs/stash is included
09:00 osse as _ikke_ learned the other day \o/
09:00 jast nasty :)
09:00 _ikke_ osse: :D
09:00 groundnuty osse: no stash, I use this thing often :)
09:01 irqq joined #git
09:01 jast hmm, if you don't have any stashes, the stash ref doesn't exist
09:01 groundnuty osse: also not in git reflog --all
09:01 _ikke_ jast: makes sense
09:02 _ikke_ jast: There is nothing to point to
09:02 groundnuty stash empty
09:02 groundnuty I do try to keep it clean
09:02 jast well my thinking was it might not clean up properly ;) but it does
09:02 knittl \o/ worked
09:03 ferr joined #git
09:03 _ikke_ jast: so little expectations
09:03 jast well if it shows up in 'git log --all', you can always use 'git log --graph --decorate' and follow the line up until the next ref label
09:04 jast _ikke_: I spend most of my job time fixing bugs :)
09:04 _ikke_ heh
09:05 imacinto joined #git
09:05 groundnuty jast: its not in  git --no-pager log --graph --decorate, but its in  git --no-pager log --graph --decorate --all
09:05 knittl rebase -i got so much faster on windows, thanks to the sequencer rewrite <3
09:06 groundnuty it shows as a unconnected node in a graph
09:06 berndl joined #git
09:06 _ikke_ knittl: Send an e-mail to git@vger.kernel.org, the maintainer would appreciate it
09:07 _ikke_ (the git for windows maintainer, to be clear)
09:08 _ikke_ https://public-inbox.org/git/alpine.DEB.2.20.1609082018410.129229@virtualbox/T/#t
09:08 ochorocho__ joined #git
09:08 MattMaker joined #git
09:09 GavinMagnus joined #git
09:09 jceb joined #git
09:09 GavinMagnus left #git
09:09 mooru joined #git
09:10 jast groundnuty: if it's unconnected it should have some kind of ref label (e.g. branchname or ref/foo/bar), otherwise there's really no way for it to show up at all
09:12 groundnuty jast: even in git --no-pager log --graph --decorate  --all ?
09:14 groundnuty jast: what are refs/replace/ ?
09:14 mooru_ joined #git
09:14 MattMaker joined #git
09:15 jast groundnuty: a special feature you can use to do open heart surgery on your git history
09:15 jast myself I've never really seen a need to use them
09:16 jast man git-replace
09:16 gitinfo the git-replace manpage is available at http://jk.gs/git-replace.html
09:16 groundnuty jast: this commit is referenced by such refs/replace/
09:16 groundnuty so we got it!
09:16 groundnuty we got it Sherlock! :)
09:17 jast wonder how that ended up there
09:17 GavinMagnus joined #git
09:17 groundnuty jast: other users...
09:17 groundnuty life
09:17 jast AFAIK replace refs aren't normally covered by push/fetch, so normally you should never have these if you didn't create them yourself
09:18 _ikke_ correct
09:19 groundnuty jast: so any tips how to remoce them from reflog?
09:19 jast they're not actually in the reflog (remember, that's with -g and you said they didn't show up there) :)
09:19 ShekharReddy joined #git
09:20 jast and that's a good question, actually. it would really help to know where these replace objects are coming from, who created them, and why
09:20 groundnuty jast: yes yes I hold on to that answer :)
09:20 jast because if they're not actually needed you could just delete them, but who knows
09:21 groundnuty jast: there was som subtree thing going on in that repo, and somone tired to get rid of the subtree
09:21 groundnuty it me by a product of those activities
09:22 jast if it's no longer at all relevant to your rewritten history, you can probably delete the replace ref (git replace -d)
09:22 jast and since it's not referenced anywhere else, that's _probably_ the case... (keep in mind, though, that I don't use replace objects and so am not familiar with all of the subtleties)
09:24 sanketdg joined #git
09:26 ahmed_elgabri joined #git
09:26 [Brain] joined #git
09:27 afuentes joined #git
09:27 al-damiri joined #git
09:28 jceb joined #git
09:29 GavinMagnus left #git
09:30 Fuco joined #git
09:31 Fuco how can I do 'git diff hash^1..hash' in some less annoying way? That is, give me difference between hash and its parent
09:31 MadHatter42 joined #git
09:31 moritz Fuco: git show hash
09:32 moritz it also gives you the commit message (and meta data), but that ususally doesn't annoy me :-)
09:32 Fuco moritz: amazing, thank you :)
09:32 nivag joined #git
09:32 nopevms joined #git
09:32 chid joined #git
09:33 groundnuty jast: ech, git replace -d says there is no such reflog :/
09:33 _ikke_ what did you pass/
09:33 _ikke_ ?
09:33 _ikke_ You have to specify the object that is being replaced
09:33 groundnuty fried firs and second 7c57752b13d52acc604969c6350bb89caa92703d refs/replace/83f737d8db345e1d506e471201fd16a49855182e
09:34 groundnuty a left side of ref and right
09:34 groundnuty its a output ofgit show-ref
09:34 _ikke_ I would expect git replace -d 83f737d8db345e1d506e471201fd16a49855182e to work
09:35 thm Fuco: git diff hash^! is also possible, shorter albeit still annoying to type
09:35 King_Hual joined #git
09:35 King_Hual joined #git
09:35 star_prone joined #git
09:35 groundnuty _ikke_: ok so thaer was a third option :)
09:36 Kurvivor in org-mode, how does one make links to directories?
09:36 _ikke_ groundnuty: It expected an object, not a ref
09:36 _joes_ joined #git
09:37 Fuco thm: ah, that's good to know.  I'm slowly getting lost in all these special shortcuts
09:37 _ikke_ man gitrevisions
09:37 gitinfo the gitrevisions manpage is available at http://jk.gs/gitrevisions.html
09:37 pouet joined #git
09:37 _ikke_ it lists them all
09:41 imacinto joined #git
09:42 groundnuty _ikke_, osse, jast: it worked! thank your for your help and your time!
09:42 groundnuty I learned a lot
09:44 MattMaker joined #git
09:45 phanimahesh joined #git
09:45 Anticom joined #git
09:46 hahuang61 joined #git
09:47 lb1c joined #git
09:48 netj joined #git
09:49 jast Kurvivor: that sounds like a question you might get more answers to in an emacs-related channel...
09:50 Kurvivor jast: it seems i have just wrote to wrong channel
09:50 Kurvivor extermely sorry
09:50 jast no worries :)
09:50 jast questions unrelated to git are not forbidden here, anywya
09:52 SteffanW joined #git
09:52 cyphase joined #git
09:53 GavinMagnus joined #git
09:54 GavinMagnus left #git
09:54 MattMaker joined #git
09:54 nivag joined #git
09:57 star_prone joined #git
09:58 cdown_ joined #git
09:58 Arguggi joined #git
10:00 SwiftMatt joined #git
10:02 Arguggi joined #git
10:04 MattMaker joined #git
10:04 a3pq51 joined #git
10:05 krzysiekj joined #git
10:10 nivag joined #git
10:10 red82 joined #git
10:12 leeN joined #git
10:13 krzysiekj Hello. A lot of sites reference kerneltrap.org when citing Git mailing list, but this site is down. Any replacements?
10:13 freimatz joined #git
10:14 grawity krzysiekj: marc.info, https://public-inbox.org/git/, or Gmane (back up again)
10:14 MattMaker joined #git
10:14 krzysiekj Thanks. :)
10:14 grawity krzysiekj: and of course http://web.archive.org/http://kerneltrap.org/whatever
10:18 stamina joined #git
10:18 sin3t joined #git
10:21 garyserj grawity: holy mackeral, the genius himself
10:21 garyserj i think i've seen some of your contributions on a stackexchange site
10:23 jast *mackerel ;)
10:23 garyserj sounds even better!
10:23 Rodya_ joined #git
10:23 Gurkenglas joined #git
10:23 jast ... and now I'm hungry
10:23 garyserj How do you checkout a file giving it a different name? So copying a file from a commit into the working directory, but giving the copy of the file a new filename?
10:23 jast garyserj: trickery: git show commit:path/to/file >fancyname
10:24 garyserj thanks
10:26 knittl I cannot apply a svn patch :( seems to be EOL related
10:27 knittl my working dir files end with crlf, and the patch file also contains crlf
10:27 jast what are you using to apply the patch?
10:27 knittl stil, git aply cannot apply the patch "patch failed. does not apply"
10:27 knittl git apply
10:27 jast have you tried with --ignore-whitespace?
10:27 knittl patch seems to work, but removes all crlf
10:27 knittl yes
10:28 ahmed_elgabri joined #git
10:28 knittl git apply -p0 --whitespace=warn < generator.patch; is what I use right now. I tried with --ignore-whitespace --ignore-space-change before
10:28 mehola joined #git
10:28 jast --recount might help, too, depending on what exactly makes it not apply
10:29 knittl hm. looks like the svn patch format is not liked
10:29 MattMaker joined #git
10:29 knittl unexpected line: Index: ...
10:29 knittl will remove those
10:29 Fuco This is not exactly git related, but does someone know some neat tooling for "making releases"... that is, replacing keywords with version IDs/tag in the sources/configuration and so on.
10:30 jast Fuco: you might want to look at how git itself does it
10:30 jast there's a script called GIT-VERSION-GEN or something like that, used in the Makefile
10:30 Fuco ouk, will check that you
10:31 knittl svn diff format is just messed up. I thought they were also using unified diff
10:34 imacinto joined #git
10:34 imacinto joined #git
10:36 sevenfourk joined #git
10:39 MattMaker joined #git
10:39 poweruser joined #git
10:41 groundnuty jast: I'm trying to push -f this changed I made to the remote
10:41 groundnuty but every time I clone from it, this just as it was before:/
10:44 MattMaker joined #git
10:49 j416 groundnuty: well what does push -f output?
10:49 sdothum joined #git
10:53 imacinto joined #git
10:54 MattMaker joined #git
10:55 DieguezZ joined #git
10:56 ruby_ joined #git
10:58 thierryp joined #git
10:59 MattMaker joined #git
11:00 rnsanchez joined #git
11:01 garyserj where is a definition of uncommitted change?
11:01 j416 garyserj: what do you mean?
11:02 j416 garyserj: an uncommitted change is a change that has not been committed.
11:02 garyserj well how would one know if uncommitted change includes untracked files or not?
11:02 JeroenT joined #git
11:02 garyserj is there a definition that says so?
11:02 _ikke_ I file is either tracked or untracked
11:03 garyserj I file?
11:03 _ikke_ s/I/A/
11:03 garyserj I know a file is either tracked or untracked
11:03 _ikke_ Uncomitted changes can only happen to tracked files
11:04 garyserj is that written in any official source?
11:04 i7c So can comitted changes. :)
11:04 _ikke_ garyserj: I don't think there are any formal definitions defined
11:04 _ikke_ but man gitglossary has some definitions
11:04 gitinfo the gitglossary manpage is available at http://jk.gs/gitglossary.html
11:05 garyserj alright so what from there, told you that uncommitted changes can only happen to tracked files?
11:05 greendo joined #git
11:06 kristofferR joined #git
11:06 _ikke_ That's not based on that resource, more from my understanding of git
11:06 _ikke_ garyserj: What's the reason you're interested in this? :)
11:06 garyserj it makes git easier to understand if things are properly defined
11:07 _ikke_ https://jk.gs/gitglossary.html#def_clean
11:07 _ikke_ https://jk.gs/gitglossary.html#def_workingtree
11:08 _ikke_ man git add
11:08 gitinfo the git-add manpage is available at http://jk.gs/git-add.html
11:08 _ikke_ also contains info
11:08 aviraldg joined #git
11:09 nwkj86 joined #git
11:09 aviraldg joined #git
11:09 aviraldg joined #git
11:09 ochorocho__ joined #git
11:09 Megaf joined #git
11:11 garyserj https://jk.gs/gitglossary.html#def_clean  "A working tree is clean, if it corresponds to the revision referenced by the current head. Also see "dirty"."  <---    When it says "corresponds" that seems a bit loose.. Does it mean completely equal? so, if an untracked file is there, then it's not clean
11:12 circ-user-F6zEO joined #git
11:13 roelmonnens joined #git
11:13 MattMaker joined #git
11:16 _ikke_ You could say that. But what if you have ignored untracked files? git add for example would report things to be clean, while the working tree is not strictly equal to the HEAD tree
11:16 hexagoxel_ joined #git
11:16 netj_ joined #git
11:17 leeN joined #git
11:17 pouet joined #git
11:17 chll_ joined #git
11:17 sweatsuit joined #git
11:17 nanga joined #git
11:18 _ikke_ git cares only about tracked files (but does show untracked files for your convenience)
11:18 jast for the purpose of determining cleanness/dirtiness, untracked files are typically not considered
11:18 aviraldg I remember there was a way, while merging to interactively choose per chunk whether to keep the local version, or remote, etc. Is there a similar way to quickly "select" some of the currently staged changes to stash or move them to a separate branch.
11:18 _ikke_ Right, though some git prompts also indicate untracked, not ignored files
11:19 GavinMagnus joined #git
11:19 MattMaker joined #git
11:19 _ikke_ aviraldg: git add -p the changes for the current branch, commit, and checkout another branch, repeat
11:19 garyserj whenever i've done git add, it has never outputted anything.. how do you get it to report whether the working directory is clean or not?
11:19 _ikke_ garyserj: it's not git adds task to do that
11:20 jast 'git add' doesn't care about that
11:20 _ikke_ garyserj: git status does
11:20 garyserj ok
11:20 flopsec joined #git
11:20 jast aviraldg: git stash -p
11:20 garyserj so what did you mean when you said "git add for example would report things to be clean"?
11:20 garyserj did you mean git status?
11:20 _ikke_ garyserj: Yes, I meant git status
11:20 aviraldg ah, thanks jast and ikke
11:21 rahtgaz joined #git
11:23 abra0 joined #git
11:24 Rodya_ joined #git
11:24 King_Hual joined #git
11:24 King_Hual joined #git
11:29 MattMaker joined #git
11:29 mehola joined #git
11:30 ToBeFree joined #git
11:31 JeroenT joined #git
11:33 stuh84 joined #git
11:33 stonercoder joined #git
11:34 cyphase joined #git
11:37 clemf joined #git
11:39 mingrammer joined #git
11:39 MattMaker joined #git
11:40 deepy Can I somehow list all files in the repository that are ascii and not UTF-8?
11:40 _ikke_ all ascii is utf8
11:40 anonymuse joined #git
11:41 deepy to rephrease, can I somehow list all files that aren't UTF-8?
11:41 _ikke_ Not through git alone
11:42 _ikke_ the file tool can make some destinction
11:42 groundnuty j416: --mirror -f solved it
11:43 groundnuty probably nit the best t do it
11:44 tilerendering joined #git
11:45 ash_workz joined #git
11:45 imacinto joined #git
11:45 tilerendering hey
11:45 SteffanW joined #git
11:45 imacinto joined #git
11:45 antranigv guys, need an advice :) what I do is make a new branch for each "feature", but after I merge it with master/dev, I delete the branch. should I keep em? :))
11:45 _ikke_ Up to you
11:45 Jellyg00se joined #git
11:46 antranigv say I want to know what differences I've done for a feature, will a deleted branch help with it? or I should keep it for that?
11:46 tilerendering a colleague has made about 100 changes (saved) to a file he didnt commit. the file is not new in the repo though. anyhow: I guess it is impossible for him to revert that file to a state like… 50 changes ago, right ?
11:46 antranigv tilerendering: rebase to a commit?
11:46 _ikke_ git does not track uncomitted changes
11:46 antranigv ah diden't commit
11:47 tilerendering yep
11:47 hahuang61 joined #git
11:47 _ikke_ rebase to a commit does not make sense btw
11:47 tilerendering so I m right, it wont be possible
11:47 _ikke_ tilerendering: unless your editor undo buffer still contains it, no
11:47 antranigv tilerendering: if he's luckey, say uses vim, he can use vim's history :3
11:47 _ikke_ intellij also keeps history
11:47 tilerendering yeah I hope aptana studio which he uses has sthg like local history tracking....
11:47 tilerendering yeah intellij netbeans eclipse do
11:48 antranigv I never knew about aptana thingie
11:49 j416 groundnuty: no idea what you are doing
11:49 j416 groundnuty: can't help without info, sorry
11:49 MattMaker joined #git
11:50 m0viefreak joined #git
11:51 mcalavera joined #git
11:51 LUMIA930 joined #git
11:52 gniourf joined #git
11:52 mcalavera left #git
11:53 achlys joined #git
11:53 imacinto joined #git
11:53 imacinto joined #git
11:54 Curly060 joined #git
12:01 ksefchik joined #git
12:02 tlaxkit joined #git
12:03 imacinto joined #git
12:03 romerocesar joined #git
12:05 eftov joined #git
12:07 synthroid joined #git
12:09 garyserj would it be correct to say that a tracked file, is a file that is not in ignore, and is not in the current commit?
12:09 garyserj sorry..
12:09 garyserj would it be correct to say that an untracked file, is a file that is not in ignore, and is not in the current commit?
12:10 ochorocho__ joined #git
12:10 tilerendering left #git
12:11 osse garyserj: not in the current commit is the main part.
12:11 cyphase joined #git
12:11 osse ignored files are technically untracked too. they just don't show up in git status etc.
12:13 garyserj well, is there any technical definition that technically says that ignored files are technically considered untracked?
12:13 osse not that I know of
12:13 osse I'd simply say a file is untracked if it's not tracked.
12:13 berndl joined #git
12:14 garyserj well, git definitions don't always go like that though.. e.g. uncommitted change, doesn't just mean any change that is not committed
12:14 osse it doesn't?
12:14 garyserj 'cos uncommitted change, is a shorthand for "changes to be committed", by which it means files in staging, and it doesn't include untracked files.
12:14 bruce_lee joined #git
12:14 garyserj even though untracked files are changes to the working directory
12:14 garyserj and are changes that have not been committed
12:15 garyserj so in that kind of technical english sense, untracked files are uncommitted changes.
12:15 osse since I'm in a picky mood: "change" implies there's a previous version. you cannot "change" an untracked file because there's nothing to compare it to
12:16 garyserj working directory compared to last commit, and the file is existent in the working directory, so it's an additional file
12:16 garyserj an A rather than an M
12:16 GodGinrai joined #git
12:16 garyserj added files still count when comparing
12:16 garyserj s/last commit/current commit/
12:16 osse not according to git diff
12:17 NimeshNeema joined #git
12:17 garyserj i will have another look at git diff.. but git status does compare
12:17 garyserj and considers additional files as something that comes out of its comparison
12:17 osse I mean git diff won't show untracked files as additions
12:18 ayjay joined #git
12:18 GodGinrai sure, just like git status
12:19 MattMaker joined #git
12:19 romerocesar joined #git
12:20 nd joined #git
12:22 star_prone joined #git
12:22 garyserj well.. i can't argue my point re git diff 'cos i'm not familiar enough with it...
12:23 ahmed_elgabri joined #git
12:23 garyserj GodGinrai: so what would you say is the definition of untracked file.  A)A file that is in the working directory and not in staging or the current commit, and not ignored. or B)A file that is in the working directory, and not in staging or the current commit, and including ignored files?
12:24 garyserj i'm taking my definition from git status, and saying 'A'.
12:24 garyserj definition from the output of git status, which doesn't list ignored files under untracked or anywhere.
12:24 osse i take my definition from git status --ignored :P
12:24 osse untracked = not tracked. I don't see how a different definition would be useful
12:25 tabrez joined #git
12:25 GodGinrai garyserj: well, technically B) iswhat untracked files, and A) is just the untracked files git tells you about, because it believes you are unaware of them
12:25 imacinto joined #git
12:25 imacinto joined #git
12:25 yoh joined #git
12:25 GodGinrai * is what
12:25 Rodya_ joined #git
12:25 j416 asterisk
12:25 GodGinrai yes
12:25 j416 !next
12:25 gitinfo Another satisfied customer. NEXT!
12:26 anonymuse joined #git
12:26 garyserj thanks
12:28 osse garyserj: it doesn't mention tracked/untracked specifically, but you might be interested in man gitglossary
12:28 gitinfo garyserj: the gitglossary manpage is available at http://jk.gs/gitglossary.html
12:28 osse maybe track deservers an entry
12:29 garyserj does uncommitted changes exclude untracked files?
12:29 osse where? when?
12:29 garyserj good question.. i'm not sure, I hear the term sometimes. i guess maybe it depends where and when
12:30 GodGinrai garyserj: uncommitted changes is a vague term.  They could mean "staged, but not committed", or "the changes I made and haven't done anything with"
12:30 anonymuse joined #git
12:30 GodGinrai of which only the second has the possibility of including untracked files
12:30 cyphase joined #git
12:31 garyserj what about 'changes to be committed' also vague?
12:31 GodGinrai "changes to be committed" sounds like staged changes to me
12:31 GodGinrai because if you ran commit, those are the changes that would be committed
12:33 yoh is there a way (without --batch, which is slowish with big number of files) to 'cat-file' all files (while knowing filenames) in a tree?    so smth in effect of   git ls-tree -r --name-only git-annex | sed -e "s/^/git-annex:/g" | git --git-dir=.git cat-file --buffer --batch
12:33 roelmonnens joined #git
12:33 GodGinrai !xy
12:33 gitinfo Woah, slow down for a bit. Are you sure that you need to jump through that particular hoop to achieve your goal? We suspect you don't, so why don't you back up a bit and tell us about the overall objective...
12:34 jeffreylevesque joined #git
12:34 roelmonnens joined #git
12:35 zivester joined #git
12:36 GodGinrai yoh: btw, I believe you're looking for the --buffer flag
12:36 GodGinrai yoh: assuming you just want the output
12:36 yoh I am already using --buffer flag
12:36 GodGinrai oh whoops
12:36 GodGinrai didn't see that in the one-liner
12:37 yoh there is --batch-all-objects but that one dumps all the objects, not only blobs ... and I think there is no way to identify clearly the boundary between items
12:37 yoh GodGinrai: no problem -- it was long indeed ;)
12:37 MadHatter42 joined #git
12:37 GodGinrai honestly, I don't see how you could make it faster.  I mean, --batch is made for a group of objects, and calling it w/o batch for each file would probably incur more overhead
12:38 GodGinrai yoh: unless you loop through and background all your processes
12:39 GodGinrai that might speed it up
12:39 GodGinrai maybe
12:39 yoh ha -- aactually!   git archive -- produce a tar and process it!
12:39 yoh works much faster! ;)
12:40 rbr joined #git
12:40 GodGinrai yea, this is why I said xy. <.<
12:41 aniceguy joined #git
12:42 fmcgeough joined #git
12:43 aniceguy Hello
12:44 aniceguy i have merge commit A    from a dev branch to master trunk. But i realize if i amend commit A to update a file, the merge between commit A to master trunk is broken?
12:44 eroussel joined #git
12:45 aniceguy I try internet solution, but to no avail. :(
12:45 aniceguy Not sure if there is no workaround it? That is even if i amend the commit A, the merge should be maintain
12:45 skylite joined #git
12:46 GodGinrai aniceguy: you aren't supposed to amend after a merge.  Just create a new commit
12:49 cyan__ joined #git
12:49 skylite_ joined #git
12:49 MattMaker joined #git
12:50 jnewt joined #git
12:50 aniceguy <GodGinrai>:Thanks for the reminder. But for learning purpose, i think its good to know how to achieve it.
12:51 cyphase joined #git
12:52 GodGinrai aniceguy: it's not achievable.  Merge commits are special.
12:52 GodGinrai They represent two branches becoming one.  Not "changes" being made to a particular branch
12:52 GodGinrai (two or more, actually)
12:56 mizu_no_oto joined #git
12:56 aniceguy cos i saw on internet, people talking about git rebase --preserve-merges. but i try it, not work.
12:57 aniceguy if merge commits cannot be altered, then wonder what is git rebase --preserve-merges
12:57 aniceguy maybe i misinterpret this command....:(
12:57 GodGinrai aniceguy: I'm fairly certain that was only for changing the message.  And you'll also notice that in one of those SO questions, it was suggested to recreate the merge.
12:58 stonercoder joined #git
12:59 red82 joined #git
12:59 aniceguy <GodGinrai> oh ok.
12:59 MattMaker joined #git
12:59 Oebele joined #git
13:01 watabou joined #git
13:03 cdg joined #git
13:03 cdg joined #git
13:04 rgrinberg joined #git
13:07 romerocesar joined #git
13:08 ahmed_elgabri joined #git
13:08 JamesBaxter joined #git
13:08 ExoUNX joined #git
13:09 johnmilton joined #git
13:09 MattMaker joined #git
13:09 andlabs joined #git
13:11 star_prone joined #git
13:12 roelmonnens joined #git
13:13 cdown_ joined #git
13:14 Bob8989|2 joined #git
13:14 stonercoder joined #git
13:17 shinnya joined #git
13:18 SourTrout joined #git
13:19 MattMaker joined #git
13:20 AaronMT joined #git
13:20 ahmed_elgabri joined #git
13:21 SourTrou_ joined #git
13:24 MattMaker joined #git
13:25 sharkman joined #git
13:26 Rodya_ joined #git
13:28 lordjancso joined #git
13:30 cyphase joined #git
13:31 ahmedelgabri joined #git
13:31 nickthetait joined #git
13:31 mehola joined #git
13:33 eroussel joined #git
13:33 ahmed_elgabri joined #git
13:34 MattMaker joined #git
13:34 PaulePanter joined #git
13:36 jaafar joined #git
13:37 VagabondAzulien joined #git
13:38 Derega joined #git
13:38 romerocesar joined #git
13:38 bluezone joined #git
13:39 PrashantJ joined #git
13:40 crelix joined #git
13:40 rnsanchez_ joined #git
13:42 mizu_no_oto joined #git
13:44 sevenfourk joined #git
13:44 MattMaker joined #git
13:44 Dougie187 joined #git
13:47 hahuang61 joined #git
13:47 overlord_tm joined #git
13:47 cyphase joined #git
13:48 fahadash joined #git
13:48 harish joined #git
13:48 sangy joined #git
13:49 overlord_tm Hey, can someone tell me when is better to use "--merge" flag with "git rebase" and when is better not to use it?
13:50 ruby_ joined #git
13:53 anonymuse joined #git
13:53 Rodya_ joined #git
13:54 MattMaker joined #git
13:55 Vict0r joined #git
13:55 heftig joined #git
13:57 mablae joined #git
13:57 jast overlord_tm: from what I know, the main benefit is rename detection. I can't think of any disadvantages right now, but the code paths are quite different so if there's no need for you to use --merge it might be better to, well, walk the path most traveled
14:00 geoid joined #git
14:01 gusnan joined #git
14:01 overlord_tm jast, I started researching available flags because last week I had many stupid conflicts during rebase. I call it stupid conflict when git reports conflict like both added, but files are actually the same. So i want to avoid this situations :D
14:02 tabrez joined #git
14:03 griffindy joined #git
14:03 watabou joined #git
14:04 MattMaker joined #git
14:04 cyphase joined #git
14:05 texasmynsted joined #git
14:05 drizztbsd joined #git
14:09 DomeMasterYolo left #git
14:09 Ryanar joined #git
14:09 deadprez joined #git
14:10 ochorocho__ joined #git
14:11 laumann joined #git
14:12 ResidentBiscuit joined #git
14:12 star_prone joined #git
14:12 ResidentBiscuit joined #git
14:13 dtrott joined #git
14:14 texasmynsted joined #git
14:14 qws-user-1228 joined #git
14:14 MattMaker joined #git
14:15 BSaboia joined #git
14:16 ronny joined #git
14:16 ronny hi
14:16 ronny im wondering, will there ever be support for fully operational shallow clones ?
14:17 cyphase joined #git
14:18 makinen left #git
14:18 phaleth joined #git
14:23 ResidentBiscuit joined #git
14:23 heftig joined #git
14:24 VagabondAzulien joined #git
14:24 MattMaker joined #git
14:24 star_prone joined #git
14:25 romerocesar joined #git
14:26 wrouesnel joined #git
14:26 ResidentBiscuit joined #git
14:27 qsx there is.
14:30 dmto joined #git
14:32 ochorocho__ joined #git
14:32 mehola joined #git
14:32 GavinMagnus left #git
14:34 MattMaker joined #git
14:34 Gamecubic joined #git
14:34 cyphase joined #git
14:34 Darcidride__ joined #git
14:35 deadprez joined #git
14:36 deadprez joined #git
14:36 Darcidride joined #git
14:39 Darcidride_ joined #git
14:39 ycyclist joined #git
14:40 charlesr joined #git
14:41 rwb joined #git
14:43 Rodya_ joined #git
14:44 strk joined #git
14:44 strk what's a POST to git-upload-pack via http ?
14:44 strk is it necessarely a push ?
14:45 strk as I'm seeing "frequent" such POSTs with very high transfer size
14:47 morenoh149 joined #git
14:47 cyphase joined #git
14:49 MattMaker joined #git
14:49 strk by "high" I mean 8-9 MB
14:49 r4z joined #git
14:50 le_melomane joined #git
14:51 dunpeal joined #git
14:51 snowkidind joined #git
14:53 dunpeal How do I check if a branch foo was merged into branch bar?
14:54 thm left #git
14:55 thiago_ joined #git
14:55 star_prone joined #git
14:57 sharkman joined #git
14:59 MattMaker joined #git
14:59 ResidentBiscuit `git branch --merged` will show all branched merged into your checked out branch
14:59 ResidentBiscuit There's probably a way to check for a specific branch
15:00 star_prone joined #git
15:01 dunpeal ResidentBiscuit: yup, found it here: http://stackoverflow.com/questions/18345157/how-can-i-tell-if-one-commit-is-an-ancestor-of-another-commit-or-vice-versa
15:01 dunpeal TLDR `git merge-base --is-ancestor <commit1> <commit2>` is the most proper way.
15:01 ResidentBiscuit That seems like overkill when `git branch --merged` works
15:02 cagmz joined #git
15:02 poweruser joined #git
15:05 watabou joined #git
15:05 sysanthrope joined #git
15:05 strk uhm, I just tried a "git clone" and did see a POST resulting from it
15:06 strk with size 29805958 (~30MB) -- is that normal ?
15:06 strk I mean, to POST 30MB on "git clone" ?
15:08 grift joined #git
15:08 sanketdg joined #git
15:09 sanketdg joined #git
15:10 zivester joined #git
15:11 notebox joined #git
15:11 Brain joined #git
15:11 sanketdg joined #git
15:12 romerocesar joined #git
15:13 charlesr joined #git
15:13 cyphase joined #git
15:15 richardsmd joined #git
15:15 israelzuniga joined #git
15:16 eivarv joined #git
15:16 richardsmd is there a way to get `git ls-files` to show all files (not just files at/below current dir)? I have a repo where the worktree is a level above the main repo dir
15:16 clemf joined #git
15:16 dunpeal ResidentBiscuit: for specific branches, I really liked the inverse: git branch --contains some_branch
15:20 nwkj86 joined #git
15:20 Gitzilla_ joined #git
15:20 anonymuse joined #git
15:23 MaBunny joined #git
15:23 MaBunny hello
15:24 OtakuSeenpai can anyone help me with this error msg: unable to store sha1 filename
15:24 MattMaker joined #git
15:24 OtakuSeenpai fatal: cannot store packfile
15:24 OtakuSeenpai pack file*
15:25 OtakuSeenpai im in a console environmeent in my debian system
15:27 mikecmpbll joined #git
15:28 satbyy joined #git
15:28 satbyy joined #git
15:28 Noldorin joined #git
15:30 perlpilot joined #git
15:33 JeroenT joined #git
15:34 actinide joined #git
15:38 actinide joined #git
15:38 cyphase joined #git
15:38 thiago OtakuSenpai: check permissions in .git
15:38 thiago make sure that everything is owned by your user
15:39 Sna4x8 joined #git
15:40 synthroid joined #git
15:40 synthroid joined #git
15:42 raspado joined #git
15:44 specing joined #git
15:45 specing I am still dissapointed in git add -e  diff view
15:45 specing the sections are just too large
15:45 AaronMT joined #git
15:46 peepsalot joined #git
15:47 thiago specing: if the distance between two modifications is 3 lines or less, it's shown as one chunk
15:47 thiago just like git diff
15:48 hahuang61 joined #git
15:49 ahmed_elgabri joined #git
15:49 strk left #git
15:49 romerocesar joined #git
15:51 specing thiago: would it be possible to limit each diff chunk to a maximum of 20 lines?
15:51 thiago no
15:51 specing i.e. so that the distance between two @@ is at most 20 lines
15:51 chimay joined #git
15:51 thiago if you added 21 lines, then 21 lines need to be shown
15:52 d5sx43 joined #git
15:52 skarn_ joined #git
15:53 specing thiago: it could split the chunk into two
15:54 specing Note that git stash save --patch has this option to further split it up
15:54 MattMaker joined #git
15:54 diogenese joined #git
15:54 cyphase joined #git
15:56 deadprez joined #git
15:56 spudowiar joined #git
15:57 watabou joined #git
15:57 richardsmd left #git
16:00 nidr0x joined #git
16:01 deadprez joined #git
16:02 madewokherd joined #git
16:02 nikivi joined #git
16:04 overlord_tm joined #git
16:04 Rodya_ joined #git
16:05 mikecmpb_ joined #git
16:06 theskillwithin joined #git
16:06 tailriver joined #git
16:07 mycelium joined #git
16:08 cyphase joined #git
16:10 flavius joined #git
16:11 kpease joined #git
16:11 chrisshattuck joined #git
16:11 nitric joined #git
16:12 stonercoder joined #git
16:12 rnsanchez_ joined #git
16:14 Khisanth joined #git
16:19 hweaving joined #git
16:19 hweaving Is there any documentation of the behavior of "git format-patch" with no arguments?
16:20 hweaving It returns silently, and all man pages I've found have no information on this.  I saw a patch that attempted to change the behavior, but I'm not sure it was accepted.
16:20 imacinto joined #git
16:20 imacinto joined #git
16:20 Electrometro joined #git
16:21 thiago hweaving: yes, there's documentation
16:21 hweaving thiago:  Which section then -- I am blind :(
16:22 irax joined #git
16:22 thiago the "DESCRIPTION"
16:22 notebox joined #git
16:23 thiago it describes the two ways of specifying the range of commits you want
16:23 thiago you didn't specify any, so there are no commits to be shown
16:24 thiago it's the same revision range as that listed by git rev-parse with no arguments: empty range
16:24 MattMaker joined #git
16:25 hweaving thiago: So effectively it's described by omission then -- I didn't notice null revision ranges were legal in man gitrevisions
16:25 gitinfo thiago: the gitrevisions manpage is available at http://jk.gs/gitrevisions.html
16:26 hweaving I would expect git format-patch to print an error due to no revisions, but I think I understand now, thank you
16:27 InfoTest joined #git
16:27 CheckDavid joined #git
16:28 cyphase joined #git
16:28 watabou joined #git
16:29 crayon joined #git
16:29 Zamerick joined #git
16:30 le_melomane joined #git
16:31 hweaving left #git
16:31 daivyk joined #git
16:31 freimatz joined #git
16:31 rgrinberg joined #git
16:33 mehola joined #git
16:33 crayon joined #git
16:33 fakenerd joined #git
16:34 greendo joined #git
16:35 clemf joined #git
16:36 fxrs joined #git
16:36 rgrinberg joined #git
16:37 ahmed_elgabri joined #git
16:38 MattMaker joined #git
16:40 dviola joined #git
16:40 DTekk joined #git
16:42 durham joined #git
16:42 cyphase joined #git
16:42 shgysk8zer0 joined #git
16:44 MattMaker joined #git
16:44 chrisshattuck joined #git
16:48 Groscheri joined #git
16:49 tabrez joined #git
16:50 mingrammer joined #git
16:51 Impaloo joined #git
16:51 stonercoder joined #git
16:52 durham joined #git
16:54 PrashantJ joined #git
16:54 ruby_ joined #git
16:54 MattMaker joined #git
16:55 [Brain] joined #git
16:56 EvilDMP joined #git
16:56 edaught joined #git
16:56 EvilDMP I have a project with a large .git directory, probably because at some point I added lots of image files by mistake
16:57 jungsubk joined #git
16:57 lacrymology joined #git
16:57 d0nn1e joined #git
16:58 lacrymology does someone have a 'clean-merge' command that deletes _BASE_, _BACKUP_, _LOCAL_, _REMOTE_ files?
16:58 EvilDMP er, how does one undo a git clean -f
16:58 nullie File recovery software
16:58 EvilDMP nullie: thought so
16:58 cyphase joined #git
16:58 EvilDMP ho hum
16:59 lacrymology EvilDMP: XD
16:59 lacrymology thanks
17:00 mikecmpbll joined #git
17:01 dreiss joined #git
17:02 mycelium joined #git
17:04 MattMaker joined #git
17:05 EvilDMP Well, that could have been worse
17:05 EvilDMP This time I will ask #git, not Stack Overflow: my .git directory is very large, probably because at some point I added lots of image files by mistake
17:06 EvilDMP Is there some way of removing files from it that don't need to be tracked in git?
17:07 Rodya_ joined #git
17:07 d0nn1e joined #git
17:09 kevwil joined #git
17:09 MacGyver Well you can rewrite history, carve out the offending files, and then repack.
17:09 darkbit joined #git
17:09 MattMaker joined #git
17:10 kevwil does anyone know why "git clean -fdx" would return a 143 error code?
17:11 osse kevwil: that might happen if it's killed by signal 15
17:11 osse (SIGTERM)
17:11 MacGyver EvilDMP: What you basically want to do is the same procedure as outlined at https://help.github.com/articles/remove-sensitive-data/
17:11 osse kevwil: shells usually report the exit code as 128+n when the process is killed by signal n
17:11 MacGyver EvilDMP: Buyer beware, though, it's easy to mess up and lose stuff you don't want to lose.
17:12 kevwil osse: ah, that makes sense.
17:12 EvilDMP MacGyver: you're talking to the man who just ran git clean -f "because it seemed like a nice command"
17:12 kevwil it was on a jenkins server, the command hit the 10 minute timeout, so jenkins probably killed the command
17:13 kevwil now I wonder why "git clean" would take > 10 minutes.
17:14 GodGinrai EvilDMP: it sounds so... clean :P
17:14 kevwil osse: thanks
17:15 falcom joined #git
17:16 falcom do anyone know a way way to git log which will display a few commits after a particular <COMMIT_HASH> ?
17:16 shgysk8zer0 joined #git
17:16 cyphase joined #git
17:16 osse falcom: after as in newer? that's not easy because of how git stores the history
17:17 osse falcom: git log --reverse <COMMIT_HASH>.. maybe
17:17 osse kevwil: ten minutes sounds like a long time
17:18 kevwil osse: yes, I'm quite sure that slave has issues, I just don't know what they are yet. Timing out on clean is not normal.
17:18 Mikerhinos joined #git
17:19 vuoto joined #git
17:20 duderonomy joined #git
17:24 MattMaker joined #git
17:28 spudowiar joined #git
17:30 Brain joined #git
17:31 VagabondAzulien joined #git
17:34 MattMaker joined #git
17:35 israelzuniga joined #git
17:35 AustinMatherne joined #git
17:38 mizu_no_oto joined #git
17:39 Sasazuka joined #git
17:41 glebihan joined #git
17:44 MattMaker joined #git
17:44 deadprez joined #git
17:45 Marbug joined #git
17:45 mingrammer joined #git
17:46 satifant joined #git
17:49 hahuang61 joined #git
17:50 ToBeCloud joined #git
17:50 tabrez joined #git
17:50 cyphase joined #git
17:51 tabrez joined #git
17:51 deadprez joined #git
17:53 mehola joined #git
17:54 irqq joined #git
17:54 MattMaker joined #git
17:55 mingramm_ joined #git
17:56 multi_io joined #git
17:56 guampa joined #git
17:58 mikecmpbll joined #git
17:58 sin3t joined #git
17:59 Sasazuka__ joined #git
18:02 PrashantJ left #git
18:02 liuzhen_ joined #git
18:03 SteffanW joined #git
18:03 deadprez joined #git
18:03 charlesr joined #git
18:04 lukky513 joined #git
18:04 MattMaker joined #git
18:07 Rodya_ joined #git
18:08 saml how can I check if there has been commits to a file since I branched off from master?
18:08 star_prone joined #git
18:08 saml or how can I find which commit on master branch i branched off from?
18:09 dendazen joined #git
18:09 cyphase joined #git
18:14 MattMaker joined #git
18:17 emg joined #git
18:17 Balliad joined #git
18:19 zivester joined #git
18:20 parasite_ joined #git
18:20 satifant saml, man git-merge-base
18:20 gitinfo saml: the git-merge-base manpage is available at http://jk.gs/git-merge-base.html
18:21 saml yup thanks
18:21 drbean joined #git
18:22 Xizor joined #git
18:22 phoenixz joined #git
18:24 yoh is it a known issue somehow that git (2.9.3) clone fails from dummy http://   some times but not always? (some race condition)  it is on a local net, so connection is quite good
18:24 _ng joined #git
18:24 MattMaker joined #git
18:24 dmto joined #git
18:25 yoh see e.g.   http://pastebin.com/hEndRFbq
18:30 EvilPenguin joined #git
18:31 yoh and GIT_TRACE_PACKET=true  doesn't anyhow trigger tracing of packets for me :-/
18:33 erguven joined #git
18:34 MattMaker joined #git
18:35 chrisshattuck joined #git
18:37 ahmed_elgabri joined #git
18:38 fahadash joined #git
18:39 crayon joined #git
18:39 dsantiago joined #git
18:40 LUMIA930 joined #git
18:45 ekinmur joined #git
18:45 DTekk joined #git
18:48 tyreld joined #git
18:52 cyphase joined #git
18:54 crayon joined #git
18:58 chrisshattuck joined #git
18:58 nidr0x joined #git
18:58 fxrs joined #git
19:02 chrissha_ joined #git
19:02 spriz joined #git
19:04 MattMaker joined #git
19:04 chrisshattuck joined #git
19:07 Kicer86 joined #git
19:08 Rodya_ joined #git
19:08 dsdeiz joined #git
19:09 jeremyabel left #git
19:09 BlaXpirit when doing git rebase autosquash, is there a way to automatically pick the rebase base to be the closest commit possible to still get all the fixups?
19:10 stamina joined #git
19:11 fractalizator joined #git
19:11 BlaXpirit hm, seems like just `rebase -i` does it
19:13 cyphase joined #git
19:14 BlaXpirit but it might have picked up one extra commit
19:15 BlaXpirit it must have rebased on origin/master
19:15 osse BlaXpirit: with no other argument rebase will default to the upstream
19:15 BlaXpirit yay i guessed it :|
19:16 osse sometimes for rewriting stuff I like to not actually rebase (ie. "rebase" on HEAD~763 or whatever)
19:16 BlaXpirit I don't understand
19:17 osse BlaXpirit: git config alias.rewrite '!git rebase -i $(git merge-base @ @{u}) && :'
19:18 markmarkmark joined #git
19:18 BlaXpirit osse, I don't see how this differs from just  `git rebase -i`
19:18 MattMaker joined #git
19:19 osse BlaXpirit: it doesn't if the upstream branch doesn't have any unique commits
19:19 osse (ie. your branch is strictly ahead)
19:20 sargas joined #git
19:20 osse but if it does then a regular rebase will move all your commits on top
19:20 mingrammer joined #git
19:24 MattMaker joined #git
19:26 osse BlaXpirit: http://sprunge.us/dXMM
19:27 chrissha_ joined #git
19:27 BlaXpirit ok thanks
19:28 rossome joined #git
19:29 osse BlaXpirit: I was merely saying that sometimes the lower thing is useful. I think what you asked about is one of those cases.
19:30 osse so the trick is to find out what to do instead of HEAD~4. turns out $(git merge-base @ @{u}) is usually good
19:31 osse so I'd advice you to do that in your situation
19:32 MrWoohoo joined #git
19:34 shinnya joined #git
19:34 deadprez joined #git
19:34 MattMaker joined #git
19:34 cyphase joined #git
19:35 hahuang61 joined #git
19:35 satbyy joined #git
19:35 satbyy joined #git
19:35 Gfurst joined #git
19:35 BlaXpirit so then what's the point of  !     && :
19:35 BlaXpirit as I understand, it's to suppress the return status,  but why?
19:35 Gfurst hey guys, so how would one go about cleaning project history?
19:36 osse BlaXpirit: no, it's to consume extra arguments
19:36 osse BlaXpirit: so if you run 'git write blah blah blah
19:36 GT4066 joined #git
19:36 osse then what ends up being executed is:  git rebase -i $(git merge-base @ @{u}) && : blah blah blah
19:37 BlaXpirit ok thanks
19:37 osse BlaXpirit: if you wanted to suppress exit status you'd do  || :  :
19:37 Gfurst I was thinking about completely starting up from scratch on a project repository, the project is relatively new but already has a lot of garbage collected, including a bunch of heavy binary files
19:37 Gfurst does it make sense to start a fresh from that?
19:38 inflames joined #git
19:38 osse Gfurst: pretend the files are sensitive and follow this guide: !sensitive
19:38 gitinfo Gfurst: [!filter_sensitive] You can use filter-branch to remove sensitive data from a repository's history. https://help.github.com/articles/remove-sensitive-data/
19:39 osse starting over is definitely easier, though. depends on how much you value the history so far. (and other people have to reclone etc)
19:41 jenzza joined #git
19:41 cdown joined #git
19:42 deadprez joined #git
19:42 kyan joined #git
19:43 actinide joined #git
19:43 Gfurst osse: thanks, that bfg seems like a useful tool, but really I don't care much for history, I actually dislike the history, specially cause I already redid a bunch of stuff back and forth it was a mess
19:43 Gfurst osse: thus wanting to actually clean it up
19:43 osse Gfurst: in that case... https://www.youtube.com/watch?v=aCbfMkh940Q
19:43 Gfurst I wonder, can I just de-init the repo on a working tree
19:44 osse Gfurst: yes:  rm -r .git
19:44 MattMaker joined #git
19:44 GodGinrai lol
19:45 Gfurst osse: nice reference
19:46 Gfurst but doesn't git have a de-init command to make it cleaner? I'm also worried with sub repos and etc
19:46 sharkman joined #git
19:46 osse no
19:47 cdown_ joined #git
19:47 Gfurst in which case I think it would be best, to keep the legacy repository backed up, and start anew with another one, including the github, and point people there instead, (I also want to change the name)
19:47 imacinto joined #git
19:48 cagedwisdom joined #git
19:48 clr Gfurst: git branch new-start $(echo 'initial commit' | git commit-tree HEAD^{tree})
19:49 cyphase joined #git
19:49 lb1a joined #git
19:49 clr that will start a new branch in your repo with a single commit reflecting the current state (i.e., no history)
19:50 osse clr: i think reflecting the current state in not wanted since it contains a bunch of crap
19:50 Gfurst but then wouldn't git garbage collect all of the history stuff?
19:51 clr osse: that's just one more commit on the old branch.  thought that might be a little easier than rm -rf .git
19:51 osse clr: that' true. good point
19:51 clr easier to see what you're removing before cutting over
19:51 osse git checkout --orphan new-start is also an alternative
19:52 TheVoid joined #git
19:52 laumann joined #git
19:53 GenZai joined #git
19:54 chrisshattuck joined #git
19:54 MattMaker joined #git
19:54 Gfurst clr: can you elaborate that? I didn't quite understand
19:55 osse Gfurst: instead of nuking everything right now, rather make a new commit with stuff removed, then afterwards make that the new initial commit
19:56 osse makes it easier to see that you've removed the correct things and such
19:56 cdg joined #git
19:57 red82 joined #git
19:57 ivo_ joined #git
19:58 chrisshattuck joined #git
19:59 Gfurst osse: so in that case it starts off with the initial commit of the main history? and still you have the new-start branch apart from the main?
19:59 Gfurst In my case I don't really matter about previous things, eg the working tree would be in the state I want it in
20:00 ivo_ joined #git
20:00 deadprez joined #git
20:00 ivo_ joined #git
20:00 Synthead joined #git
20:00 osse Gfurst: no after the dust has settled the history will start off with the commit you just created
20:01 chrisshattuck joined #git
20:01 osse Gfurst: basically: clean it up as if you didn't plan to nuke it all to hell. Once everything is clean, you nuke it all to hell.
20:04 ivo_ joined #git
20:04 raijin joined #git
20:04 MattMaker joined #git
20:06 dilfridge joined #git
20:06 jason237 joined #git
20:07 Gfurst osse: but I don't get it, the first half of the command is creating a new branch right?
20:08 osse yes
20:08 Gfurst where did you get $initial commit from?
20:08 clr Gfurst: what osse said.  doesn't matter much, though, either approach will get the job done
20:08 ivo_ joined #git
20:09 osse Gfurst: there's no $initial anywhere
20:09 Rodya_ joined #git
20:09 osse ok, i see the code doesn't match my explanation exactly.
20:09 ivo_ joined #git
20:10 Gfurst ahh right I didn't see the ending )
20:10 osse this creates a new initial commit which you then clean up.
20:10 ivo_ joined #git
20:11 Darcidride joined #git
20:11 ivo_ joined #git
20:12 chrisshattuck joined #git
20:12 Gfurst ahh, so is HEAD^{tree} the beginning of history?
20:12 osse no
20:12 osse it refers to the tree of the current commit directly
20:12 osse tree = "directory" basically.
20:12 cyphase joined #git
20:13 osse !tree
20:13 gitinfo [!subtree] The subtree merge method is great for incorporating a subsidiary git repo into your current one with "unified" history. Read http://git-scm.com/book/en/v2/Git-Tools-Advanced-Merging#_subtree_merge for more info, see also !git-subtree and !git-stitch-repo.
20:13 osse bleh, disregard that.
20:14 mehola joined #git
20:14 Gfurst he
20:14 MattMaker joined #git
20:15 Gfurst right so you create a new commit base on current working tree, then creates branch new-start from that
20:15 Gfurst right so far?
20:15 osse yup
20:15 Gfurst in that case doesn't the commit have a parent?
20:16 osse nope
20:16 clr Gfurst: ... as long as working tree agrees with the last commit
20:16 afe joined #git
20:16 Gfurst ah ok, so then you would still have your other branches along with the new-start one wouldn't it?
20:17 clr HEAD^{tree} is something immutable in the repo
20:17 osse Gfurst: yes
20:17 osse you can delete them after
20:17 Gfurst ahh right, delete the others rename master and so on....
20:18 osse Gfurst: but we don't know how big and complex this project is... maybe you'd be already done by now if you did rm -r .git when I told you to :p
20:18 Gfurst thus why you said I could check?
20:18 osse this is the super-duper safe alternative
20:19 Gfurst well, anyway I just wanted to understand the command for my own
20:19 osse no problem
20:19 osse btw. it's the use of commit-tree instead of commit that makes it a parent-less commit
20:19 Gfurst I think the safest alternative would be to start off from empty directory
20:20 mizu_no_oto joined #git
20:20 naw joined #git
20:21 Gfurst osse: I didn't get that understanding from the man page :p
20:21 Gfurst so I guess you do have to manually specify parents on commit-tree
20:21 osse yes
20:21 clr Gfurst: there are no -p <parents> specified, thus zero parents
20:21 Gfurst ahh right
20:22 Sadeukko joined #git
20:22 Gfurst actually thinking about it removing .git wouldn't be such a good idea since I could lose some meta data that may be important
20:22 osse more human alternative, imho.:  git checkout --orphan new-start; git commit -m 'initial commit'
20:22 HoierM joined #git
20:23 Gfurst creating the new-start branch is good idea indeed, the I could push to a different remote and start from there
20:23 Gfurst that one too should have the same result
20:23 Gfurst hehe cheers
20:23 clr Gfurst: to be totally clear, I should have said HEAD^{tree} *refers* to something immutable in the repo, whereas the working tree (the files you edit) are obviously mutable.
20:24 MattMaker joined #git
20:24 Gfurst clr: so what does it refers actually? the tree state from HEAD isn't?
20:25 clr Gfurst: right, a tree is the current state and a commit points to one tree
20:26 Gfurst alright thanks
20:26 Gfurst always learning something new in git
20:27 e14_ joined #git
20:28 clemf joined #git
20:29 vbgunz joined #git
20:30 Gfurst bw, quick question, if I'm starting from a fresh commit like, would it already know of submodules?
20:30 liquidjamm joined #git
20:30 osse yes
20:31 skylite_ joined #git
20:31 vhost- joined #git
20:32 saml i did   git rebase origin/master;   but some files in origin/master isn't on my HEAD
20:32 saml is there a way to find what files i missed during conflict resolution  and add them to my branch?
20:33 stamina joined #git
20:34 liquidjamm joined #git
20:34 MattMaker joined #git
20:34 toogley joined #git
20:37 Hink joined #git
20:39 hahuang61 joined #git
20:41 madebymarkca joined #git
20:41 osse saml: git diff --diff-filter=D origin/master HEAD
20:41 osse that's all files in origin/master not in HEAD
20:41 boombatower joined #git
20:42 elementalest joined #git
20:42 elementalest joined #git
20:42 cyphase joined #git
20:44 dendazen joined #git
20:44 MattMaker joined #git
20:45 jlebar joined #git
20:46 gfixler joined #git
20:55 greendo joined #git
20:56 cdown joined #git
20:57 Gianormagantrous joined #git
20:57 Gianormagantrous joined #git
20:57 Sasazuka joined #git
20:59 MattMaker joined #git
20:59 cyphase joined #git
21:00 [Brain] joined #git
21:00 red82 joined #git
21:01 dmto joined #git
21:02 noctux joined #git
21:03 spudowiar joined #git
21:06 madebymarkca joined #git
21:07 vbgunz anybody know of any cool tricks to making a commit and getting a version number to work with your project? I'm sure there's dozens of solutions for this, any one recommended over the other?
21:08 strk joined #git
21:08 vbgunz what I mean is, rather than write my version number into my project, getting git to manage it but making it available for use within the project itself outside of git
21:08 strk (how) is it possible to do something similar to "git clone --reference $path $remote" on fetch ?
21:08 vbgunz I'm thinking a hook but I'm curious if there's perhaps a recommended method to this
21:08 strk like: git init; git add remote origin $remote; git <what_here?> fetch origin
21:09 MattMaker joined #git
21:10 Salander27 joined #git
21:12 EvilPenguin joined #git
21:15 e14 joined #git
21:15 mehola joined #git
21:17 whitby joined #git
21:17 relipse joined #git
21:19 MattMaker joined #git
21:19 Duder9999 joined #git
21:22 guampa joined #git
21:24 cyphase joined #git
21:25 strk found: found: echo "xxx/.git/objects" >  .git/objects/info/alternates
21:25 strk right before the "git fetch"
21:25 strk is there a cleaner way to do the alternates fillup ?
21:26 strk GIT_ALTERNATE_OBJECT_DIRECTORIES, sounds like
21:29 MattMaker joined #git
21:31 MattMaker joined #git
21:33 neanderslob joined #git
21:34 clemf_ joined #git
21:36 clemf__ joined #git
21:39 clemf joined #git
21:40 fractalizator left #git
21:42 madewokherd joined #git
21:42 cyphase joined #git
21:46 RickLee joined #git
21:51 imacinto joined #git
21:51 Blkt joined #git
21:52 leehambley joined #git
21:53 Blkt joined #git
21:54 Blkt joined #git
21:56 drizztbsd joined #git
21:56 andlabs joined #git
21:56 deadprez joined #git
22:01 cyphase joined #git
22:04 imacinto joined #git
22:04 d0nn1e joined #git
22:04 nwkj86 joined #git
22:11 imacinto joined #git
22:11 shgysk8zer0 joined #git
22:16 justanotheruser joined #git
22:16 cyphase joined #git
22:17 aidalgol joined #git
22:18 Fijit joined #git
22:18 King_Hual joined #git
22:18 King_Hual joined #git
22:19 dendazen joined #git
22:19 MrWoohoo joined #git
22:19 clemf joined #git
22:21 e14 joined #git
22:26 King_Hual joined #git
22:26 King_Hual joined #git
22:28 red82 joined #git
22:29 BlaXpirit i made some commits, how can I push to a remote excluding a few last commits? (don't want to push yet)
22:31 zumba_addict joined #git
22:32 relipse joined #git
22:32 zumba_addict good afternoon folks. I remember I was able to do a git diff do different versions of remote master using ^^^^HEAD or HEAD^^^, i couldn't recall how it's done. I also remember I was able to specify a filename too
22:35 NightMonkey joined #git
22:35 zumba_addict found something, git diff HEAD^^ HEAD main.c   Am I right that HEAD^^ is remote 2 previous commits?
22:36 diogenese joined #git
22:37 clr BlaXpirit: git push <remote> HEAD~<N>:<branch>, where <remote> is probably "origin", <N> is the number of commits you don't yet want to push, and <branch> is the remote branch, probably the same name as your local branch (e.g., "master")
22:37 BlaXpirit clr, that is very clear, thank you
22:38 clr zumba_addict: HEAD^^ is two commits back on your local branch
22:39 zivester joined #git
22:40 clr zumba_addict: git diff origin/master master would diff the remote "master" branch to the local "master" branch
22:40 cyphase joined #git
22:41 boombatower joined #git
22:43 e14 joined #git
22:44 e14 joined #git
22:44 Peuc joined #git
22:47 ochorocho__ joined #git
22:47 strk left #git
22:48 phoenixz joined #git
22:51 clr zumba_addict: if your current branch has an upstream branch set, this should generally work: git diff HEAD@{upstream} HEAD
22:52 clr zumba_addict: so I think HEAD@{upstream} is what you're looking for, not HEAD^^.  but I believe the former is somewhat of a recent addition to Git.
22:57 DieguezZ joined #git
22:59 citizen-stig joined #git
23:00 Duder9999 joined #git
23:04 Vampire0 clr, I think HEAD@{upstream} will not work as HEAD is not a branch name, but  you can make it even shorter:         git diff @{u} @                @{u} is short for@{upstream} is short for <current branch>@{upstream} or <current branch>@{u}          @ is short for HEAD
23:06 Duder9999 I need some help
23:09 cdg joined #git
23:09 clr Vampire0: nice.  i only saw that syntax recently, and wondered if HEAD would work if currently on a branch (i.e., not detached.)  seemed to do the job, but clearly the syntax you suggest is superior.  thanks.
23:16 cdown_ joined #git
23:17 clemf joined #git
23:17 mehola joined #git
23:18 Vampire0 ok, wasn't sure about HEAD@{u}, good to know
23:19 mablae_ joined #git
23:23 ffabi joined #git
23:24 deadprez joined #git
23:27 ASOLAr joined #git
23:28 mingrammer joined #git
23:29 vbgunz_ joined #git
23:34 star_prone joined #git
23:34 disi joined #git
23:43 EvilPenguin joined #git
23:44 jwarren_ joined #git
23:46 deadprez joined #git
23:47 jwarren_ Hey folks, I'm trying to get a diff --name-only between my branch and master, but I only want additions my branch makes if that makes sense. Is there a way to get that output?
23:48 jwarren_ A regular diff --name-only works if your branch is even with master, but if master is ahead that doesn't work.
23:48 ochorocho__ joined #git
23:49 fmcgeough joined #git
23:51 cyphase joined #git
23:52 thiago joined #git
23:52 MrWoohoo joined #git
23:55 stonercoder joined #git
23:56 JanC_ joined #git
23:57 mizu_no_oto joined #git
23:58 osse jwarren_: --diff-filter
23:58 osse or maybe i don't understand the question

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