Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-08-19

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

All times shown according to UTC.

Time Nick Message
00:00 wilhelmtell jnewt: what's the value of your TERM environment variable?
00:00 whomp_ joined #git
00:00 jnewt wilhelmtell, xterm
00:01 wilhelmtell try setting it to XTERM-256color and connecting again
00:02 dsantiag_ joined #git
00:02 inflames joined #git
00:03 wilhelmtell jnewt: do you see the escapes if you run
00:03 wilhelmtell git log --color=never
00:03 wilhelmtell ?
00:05 stickperson joined #git
00:09 jnewt wilhelmtell, it's not sticking.  i am reading up, i think i need to create a ~/.bashrc to get the environment variable to "stick" export TERM="XTERM-256colors" changes it, but not the behavior, an the disconnect / connect cycle sets it back to xterm
00:10 HoierM joined #git
00:11 wilhelmtell are you using putty? you can set it there.
00:11 dandaman joined #git
00:12 jnewt no, no putty, just the bash program that comes with the git installation on windows (git bash or git-bash, also shows MIINGW64 in the title bar)
00:15 vassagus joined #git
00:16 thiago the problem is the terminal application you're using
00:20 lebara joined #git
00:21 EvilPeng1 joined #git
00:22 Ozd joined #git
00:22 Ozd hi there
00:22 Ozd i had some conflicting files when updating a branch from master
00:23 Ozd i fixed them, but they are still being listed as conflicting when using that command "git diff --name-only --diff-filter=U"
00:23 Ozd therefore i cannot commit the merge
00:23 Ozd what should i do?
00:24 wilhelmtell The only escapes I can imagine you getting from Git are color escapes.
00:25 wilhelmtell jnewt:
00:25 wilhelmtell So if you want to fix this quickly, you run git config --global color.ui never
00:26 wilhelmtell and of course git config --edit, to verify the local config doesn't override that.
00:26 wilhelmtell But I think the problem is in the terminal setting.
00:27 wilhelmtell Although I also think Git should should see the issue and assume it's dealing with a dumb terminal if it doesn't recognize the terminal
00:27 wilhelmtell The next thing I'd try is setting TERM to xcolor-256color in your local machine as well as the remote.
00:28 wilhelmtell and then re-connecting.
00:28 wilhelmtell do that in .bashrc
00:28 wilhelmtell also make sure it gets picked up. sometimes windows places the files elsewhere
00:28 mrBen2k2k2k joined #git
00:29 wilhelmtell like, %HOME%/bashrc rather than .bashrc
00:29 wilhelmtell I don't know I don't have a Windows machine
00:29 wilhelmtell left #git
00:29 cdg joined #git
00:30 arescorpio joined #git
00:31 jstimm joined #git
00:33 Lyqyd joined #git
00:35 netj joined #git
00:36 roxtrongo joined #git
00:36 EarlBanana joined #git
00:39 eka joined #git
00:40 alansaul joined #git
00:43 vrishab joined #git
00:44 cjsarette joined #git
00:44 fuchstronaut joined #git
00:46 jefus joined #git
00:47 Rothschild_666 joined #git
00:47 Rothschild_666 hello git
00:47 gitinfo Rothschild_666: 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.
00:47 checkItOut_ joined #git
00:47 netj joined #git
00:47 Rothschild_666 when i am issuing a pull request, is it ok to change the short description i entered when initially comitting it to my forked repo?
00:48 jnewt wilhelmtell.  ok, i have color on the ssh session now.  i am still seeing several control characters.  <E2><80><94> for example
00:49 EvilPeng1 joined #git
00:50 lebara joined #git
00:50 SporkWitch Rothschild_666: do you mean changing what's put in the pull request's text? or what's in the repo?  You can do an amend commit to change the comment if you want; do so before submitting a pull request
00:50 SporkWitch Rothschild_666: shouldn't be any problem if you change the text of the pull request, though; it shows up as a separate comment in the history anyway, IIRC
00:52 SporkWitch Rothschild_666: so when the pull request gets merged in, the second to last commit will be the actual comment from your commit, and the most recent will be the merge comment, which will just include the description sentence
00:52 ryouma left #git
00:53 Rothschild_666 SporkWitch: no the text when you do git commit blah, that like short description it asks for
00:53 wgrant joined #git
00:53 Rothschild_666 i made mine a little too long so it carried over into the comment section, so i was wondering, can i shorten that text and just put the actual details in the comment?
00:54 SporkWitch Rothschild_666: you have two choices: you can do an amend commit and then force-push to your repo to change the comment (as long as no one's pulled from you, this should be fine)
00:55 rchavik joined #git
00:55 SporkWitch or as you said, you can just make the description in the pull request different; tools like github and bitbucket just auto-fill the fields from your last commit comment, there's nothing stopping you from completely rewriting the title and description of the pull request
00:56 tesuji_ joined #git
00:57 Rothschild_666 SporkWitch: ah ok
01:01 hahuang61 joined #git
01:03 button joined #git
01:06 inflames joined #git
01:06 pks joined #git
01:07 Bizkit joined #git
01:09 d^sh joined #git
01:09 mrBen2k2k2k joined #git
01:10 Literphor_ joined #git
01:10 Sonder joined #git
01:14 tesuji joined #git
01:17 kingarmadillo joined #git
01:17 fakenerd joined #git
01:18 mehola joined #git
01:18 jottr_ joined #git
01:19 lebara joined #git
01:22 vassagus joined #git
01:23 mingrammer joined #git
01:24 cqi joined #git
01:24 Vampire0 joined #git
01:25 user083 joined #git
01:26 dreiss joined #git
01:27 liuzhen joined #git
01:31 dandaman joined #git
01:33 optraz joined #git
01:34 ytl joined #git
01:36 chigang_ joined #git
01:38 dsantiago joined #git
01:38 rossome joined #git
01:38 stickperson left #git
01:40 JamesBaxter joined #git
01:42 mycyber joined #git
01:47 Gianormagantrous joined #git
01:47 Gianormagantrous joined #git
01:50 ilbot3 joined #git
01:50 Topic for #git is now Welcome to #git, the place for git help and doorbells | Public logs at http://goo.gl/BuUi5o | Current stable version: 2.9.3 | First visit? Read: http://jk.gs/git | Getting "cannot send to channel"? /msg gitinfo .voice | Knock knock. Who's there? Git. Git-who? Sorry, 'who' is not a git command - did you mean 'show'?
01:51 mrBen2k2k2k joined #git
01:51 lebara joined #git
01:52 dhollinger joined #git
01:53 ResidentBiscuit joined #git
01:53 Dougie187 joined #git
01:55 eizua joined #git
01:56 tjone270 joined #git
02:01 cjsarette joined #git
02:04 fakenerd joined #git
02:05 inflames joined #git
02:09 durham_ joined #git
02:10 anttim joined #git
02:10 chachasmooth joined #git
02:16 cqi joined #git
02:17 mingrammer joined #git
02:20 red__ joined #git
02:21 lebara joined #git
02:26 vassagus joined #git
02:33 lagothrix joined #git
02:37 mrBen2k2k2k joined #git
02:37 hexagoxel joined #git
02:44 Manuel15 joined #git
02:44 Manuel15 joined #git
02:46 fuchstronaut joined #git
02:47 xall joined #git
02:47 Dougie187 left #git
02:49 diogenese joined #git
02:50 Impaloo joined #git
02:52 lebara joined #git
02:54 Ozd_ joined #git
02:55 watabou joined #git
02:56 boombatower joined #git
03:00 fstd_ joined #git
03:02 aavrug joined #git
03:02 hahuang61 joined #git
03:03 aavrug joined #git
03:03 kn-928 joined #git
03:04 constant joined #git
03:06 kn-928 joined #git
03:06 JamesBaxter joined #git
03:06 Ozd_ left #git
03:06 durham joined #git
03:07 jefus joined #git
03:08 rgrinberg joined #git
03:09 mohabaks joined #git
03:10 blackwind_123 joined #git
03:11 red__ joined #git
03:12 mingrammer joined #git
03:12 dreiss joined #git
03:13 ELFrederich what all is available to me within a "git filter-branch --commit-filter"?  some examples I see use $GIT_AUTHOR_EMAIL and stuff.  But what else is available?  Is the message available?
03:14 lolmac joined #git
03:18 mehola joined #git
03:19 ahr3n joined #git
03:21 lebara joined #git
03:21 harleyyyyymayes joined #git
03:23 mrBen2k2k2k joined #git
03:27 TSS_ joined #git
03:27 VladGh joined #git
03:31 vassagus joined #git
03:32 raijin joined #git
03:34 mehola joined #git
03:35 jadew joined #git
03:36 darkbit joined #git
03:37 nydel07 joined #git
03:38 raijin joined #git
03:40 a_thakur joined #git
03:40 spudowiar ok, what's the best way to merge some commits locally but never push them to the server?
03:40 jnewt i'm trying to delete a folder from my repository (moved to a new repo.  i tried git filter-branch --tree-filter 'rm -rf <DIR_NAME>' -- --all and it's still there in git log.
03:41 a_thakur joined #git
03:41 jnewt spudowiar, create a new branch, do your merging and whatnot there, then don't push it
03:41 spudowiar :(
03:41 spudowiar not what I wanted to hear :)
03:41 justanotheruser joined #git
03:42 spudowiar I was hoping I could essentially create a branch that follows my current branch but excludes a merge commit or something
03:42 jnewt spudowiar, why not? or just do the merges, then don't push.  i don't know why you need this.
03:42 jnewt spudowiar, clone your repo, set up a tracking branch there to follow your branch, then don't commit the new repo.
03:43 jnewt you can do whatever you want and then not commit
03:43 jnewt or don't push. or whatever.  what is the problem
03:43 spudowiar jnewt: so, it's an embedded device I'm developing on and we have this hack to allow faster firmware flashing. I want to be able to commit and push, etc. but not push the commit containing the hack
03:46 crayon joined #git
03:46 jnewt spudowiar, can you use git add --interactive to exclude that hack as a hunk that you never commit?
03:47 spudowiar jnewt: well, it does get very annoying doing that :( if it were in its own file, I could use git-update-index to make it act as if it were not modified
03:47 fuchstronaut joined #git
03:47 spudowiar But I actively modify the file the hack lives in
03:47 jnewt spudowiar, that's fine, you can add all the other hunks, just not the hack
03:48 spudowiar jnewt: I know, but I work about that hunk so Git decides to treat it as one and I have to manually edit, etc.
03:48 spudowiar It's annoying
03:49 jnewt spudowiar, why can't the hack be committed?
03:49 spudowiar jnewt: because it's a hack! :)
03:49 spudowiar jnewt: it can be committed, just not pushed to the server
03:50 jnewt sounds like a feature to add faster operation.  i haven't heard anything about why it's detrimental to the project
03:50 Motoko joined #git
03:50 jefus joined #git
03:51 spudowiar jnewt: it compromises the security of the device
03:51 danimal2 joined #git
03:52 jnewt spudowiar, use a preprocessor option in your compiler to enable it when building locally, don't pass the option when doing a production build
03:53 raijin joined #git
03:53 lebara joined #git
03:54 spudowiar1 joined #git
03:54 spudowiar1 oops, I got disconnected :(
03:54 spudowiar1 jnewt: it compromises the security of the device
03:54 spudowiar1 oops, I got disconnected :(
03:54 spudowiar1 jnewt: it compromises the security of the device
03:55 spudowiar1 jnewt: when we push to the server, that code is considered "production" and can be signed. Code that can be signed shouldn't allow the sort of stuff this hack does
03:55 spudowiar1 jnewt: for example, flashing new unsigned firmware without wiping storage
03:58 miklcct joined #git
04:01 thecomedian joined #git
04:01 Andrew_K joined #git
04:02 jnewt compiler option
04:02 l4v2 joined #git
04:02 spudowiar jnewt: also, the code quality is terrible
04:03 spudowiar jnewt: because there's no reason to put the effort in to make it nicer
04:03 spudowiar but it's not the sort of thing you want committed
04:06 kingarmadillo joined #git
04:09 phanimahesh joined #git
04:10 rewt spudowiar, refactor that code out to another file that is included in the compiled source only if present, and don't commit it
04:11 thiago joined #git
04:12 mrBen2k2k2k joined #git
04:14 dandaman joined #git
04:16 rkazak joined #git
04:17 dhinojosa joined #git
04:18 jefus joined #git
04:19 Cabanossi joined #git
04:19 dhollinger joined #git
04:21 AbstractArtist joined #git
04:22 telephone joined #git
04:22 telephone joined #git
04:24 lebara joined #git
04:29 dhinojos1 joined #git
04:29 mohabaks joined #git
04:32 raijin joined #git
04:32 phanimahesh joined #git
04:35 chid joined #git
04:36 lalilulelo joined #git
04:36 jnewt i'm trying to delete a folder from my repository (moved to a new repo.  i tried git filter-branch --tree-filter 'rm -rf <DIR_NAME>' -- --all and it's still there in git log.
04:37 vassagus joined #git
04:41 tanuki joined #git
04:42 jadew joined #git
04:48 fuchstronaut joined #git
04:53 dreiss joined #git
04:54 bttf joined #git
04:55 mrpackethead joined #git
04:55 lebara joined #git
04:56 bttf I did a `git pull origin pr/:id/head` to checkout a PR locally. I can't figure out how to go back to my version of master. my remote hasn't changed, and when I do a regular pull, I get 'up to date', even though the PR's commits are still part of my tree
04:56 kyan joined #git
04:56 _ikke_ bttf: That would have merged the PR into the current branch..
04:57 rkazak joined #git
04:57 bttf ooooh
04:58 cyan__ joined #git
04:58 _ikke_ git pull is not meant for changing branches or something like that
04:58 bttf right right
04:58 _ikke_ Is the pull the last thing you did?
04:59 subhojit777 joined #git
04:59 bttf i should have fetched presumably
05:00 bttf it is, yes. i just did a reset --hard
05:00 mingrammer joined #git
05:01 anttim joined #git
05:02 hahuang61 joined #git
05:04 mrBen2k2k2k joined #git
05:06 _ikke_ good
05:06 _ikke_ and yes, fetch just gets you history, letting you check it it
05:06 _ikke_ it out*
05:07 rj1 joined #git
05:08 SteffanW joined #git
05:08 Andrew_K|2 joined #git
05:09 jefus joined #git
05:10 rj1 joined #git
05:12 diogenese joined #git
05:13 dandaman joined #git
05:13 phanimahesh joined #git
05:14 bttf danke _ikke_
05:18 a_thakur joined #git
05:21 cqi joined #git
05:22 a_thakur_ joined #git
05:22 mrpackethead joined #git
05:26 nidr0x joined #git
05:26 lebara joined #git
05:29 zeroed joined #git
05:32 jnewt when was git subtree added.  i have it on my windows 2.9.3 version, but not my freebsd 2.9.2 install
05:32 ShekharReddy joined #git
05:32 kadoban jnewt: It's in the contrib stuff in the git.git repo. It probably depends on the packager if they include it or not.
05:32 kadoban It's far older than 2.9.2 of course.
05:32 alansaul joined #git
05:34 githogori joined #git
05:38 mariuscc joined #git
05:39 dreamyspell joined #git
05:40 jnewt kadoban, does it matter where HEAD or which commit / branch you have checked out when running subtree?
05:40 dandaman joined #git
05:40 vassagus joined #git
05:40 Rish joined #git
05:41 jnewt i want to extract the whole tree of one folder (all branches) into a different repo
05:43 mrpackethead joined #git
05:45 SimpleName joined #git
05:45 SimpleName how to show commit date by a commit sha1
05:45 harleyyyyymayes joined #git
05:46 mgedmin git log --pretty=%ci $sha1
05:46 mgedmin oh and tack -1 at the end
05:48 fuchstronaut joined #git
05:49 crayon joined #git
05:52 anttim joined #git
05:52 stuh84 joined #git
05:53 ome joined #git
05:54 mingrammer joined #git
05:55 Raging_Hog joined #git
05:55 LucaTM joined #git
05:56 tvw joined #git
06:02 circ-user-FePRT joined #git
06:02 mrBen2k2k2k joined #git
06:03 jceb joined #git
06:10 jefus joined #git
06:10 rscata joined #git
06:13 Darren_ joined #git
06:20 satya4ever joined #git
06:21 roelmonnens joined #git
06:22 timvishe` joined #git
06:22 qballer joined #git
06:23 crayon joined #git
06:23 Macaveli joined #git
06:27 cbreak jnewt: use git filter-branch
06:27 cbreak jnewt: git subtree is not a git command
06:27 nilg joined #git
06:29 Raging_Hog joined #git
06:34 crayon joined #git
06:34 jefus joined #git
06:36 xaa joined #git
06:36 kingarmadillo joined #git
06:39 lebara joined #git
06:47 bocaneri joined #git
06:48 portwolf left #git
06:48 t0by joined #git
06:48 AbstractArtist joined #git
06:48 Sceorem joined #git
06:48 ferr joined #git
06:48 elect joined #git
06:49 vassagus joined #git
06:49 fuchstronaut joined #git
06:51 dhollinger joined #git
06:53 jefus joined #git
06:54 dreiss joined #git
06:55 leeN joined #git
06:58 Manuel15 joined #git
06:58 Manuel15 joined #git
06:59 FunkyAss joined #git
07:02 freimatz joined #git
07:03 elect we have this situation
07:03 elect http://i.imgur.com/XOzpxwv.png
07:03 hahuang61 joined #git
07:03 elect the selected commit "baseVertex bug" is the last good commit without the bug
07:04 elect I just checked it out
07:04 noobish-noob joined #git
07:04 elect now I'd like to merge with all the previous commits from my collegues, commit by commit to see where is the bugging commit
07:04 elect how can I do that?
07:04 _ikke_ elect: look at man git bisectr
07:05 gitinfo elect: the git manpage is available at http://jk.gs/git.html
07:05 _ikke_ elect: look at man git bisect
07:05 gitinfo elect: the git-bisect manpage is available at http://jk.gs/git-bisect.html
07:05 noobish-noob NIGGERS SPICS AND JEWS - THEY ALL TEAMED UP AND THEY ARE TRYING TO HACK MY
07:05 noobish-noob I see twenty tcp tries
07:05 noobish-noob every second
07:05 noobish-noob lol
07:05 noobish-noob i am voiced
07:05 noobish-noob anyhow, they are going to compromise us all
07:06 noobish-noob was kicked by _ikke_: noobish-noob
07:06 klan-noob joined #git
07:06 klan-noob hey
07:06 gitinfo klan-noob: 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.
07:06 klan-noob they are going to kill the klan
07:06 klan-noob if they hack the main databasr
07:07 klan-noob they see all the members' addresses
07:07 klan-noob the Ku Klux Klan needs your help!
07:07 klan-noob was kicked by _ikke_: klan-noob
07:07 mrBen2k2k2k joined #git
07:07 nchambers _ikke_, I would try /mode #git +b *!*@*139.64.40.188*
07:07 nchambers if they come back
07:07 _ikke_ right
07:08 _ikke_ But I think he joined before the generic +b
07:08 nchambers oh I missed the generic one
07:08 rossome joined #git
07:09 rkazak joined #git
07:10 TomyWork joined #git
07:10 elect _ikke_, I'd like to merge previous commits from my collegues into mine one after each other
07:11 elect and see where is the problem
07:11 elect because it's not really a bug, but a performance regression
07:11 submitnine joined #git
07:12 encod3 joined #git
07:13 _ikke_ Do you understand the idea behind bisect?
07:13 elect kind of
07:13 elect but I am not really confident
07:13 elect never did it before
07:14 elect I am afraid to mess things up
07:14 osse elect: bisect tells you to saw a commit is "good" or "bad". what the requirements are is up to you. doesn't have to be a bug in the true sense
07:14 osse *to say a commit
07:15 elect may I do things manually?
07:15 osse what things?
07:15 elect that is checking out commit by commit manually?
07:15 rads joined #git
07:16 osse sure. in that case bisect is not for you
07:16 elect I tried to check out another commit that is not mine
07:17 elect (everything by netbeans gui)
07:17 elect If I don't checkout as new branch it says it will create a detached head
07:17 elect shall I create the new branch, right?
07:17 osse yes, that's expected, and harmless
07:18 osse no need if you don't intend to make changes you want to keep
07:18 _ikke_ detached HEAD just to view commits is not a problem
07:18 elect ok
07:18 elect then I'll with it
07:18 AL13N_work left #git
07:18 elect *go
07:18 _ikke_ Bisect is basically a binary search. It tried to reduce the amount of commmits you need to test
07:18 _ikke_ tries
07:20 elect ok
07:21 Curly060 joined #git
07:22 KnightsOfNi joined #git
07:26 DevAntoine joined #git
07:26 Balliad joined #git
07:27 styler2go Hello. Can someone help me? I have git repos on a network drive and it always tells me that all files changed from old mode 100755 to new mode 100644, what can i do to fix that?
07:27 styler2go "git config --global core.filemode false" did not help
07:27 lebara joined #git
07:30 rads_ joined #git
07:30 Curly060 Hi, lets say my commit history is A B C and in B the files b1 and b2 were modified. How can I undo the changes introduced by B but *only on file b1* and not b2? I know git revert, but that reverts all files of a given commit.
07:31 rads_ joined #git
07:32 mgedmin you can git checkout B^ -- b1 to get the contents of file b1 as it was before B
07:32 mgedmin then you can git commit -am "Revert changes to b1 made by B because <reason>"
07:32 howdoi joined #git
07:35 vuoto joined #git
07:35 vuoto joined #git
07:37 Curly060 mgedmin: but what if file b1 was also modified in C? would those change not be lost in that case?
07:37 mgedmin you're right
07:37 ytl joined #git
07:37 Seveas Curly060: they would be. You can also try git revert -n B to revert without committing, then git add -p to make sure only changes to b1 are added, commit that, and reset --hard the rest
07:37 mgedmin in that case try git revert --no-commit B && git checkout -- b2
07:38 jknetl joined #git
07:39 dreiss joined #git
07:40 osse another alternative: git checkout -p B^ -- b1
07:40 osse another alternative: git diff diff B B^ -- b1 | git apply
07:42 Seveas Tim Toady in the house :)
07:43 ASOLAr joined #git
07:44 viki joined #git
07:45 Curly060 osse: that is exactly what I have been looking for! your solutions actually only modify the file in question, which is exactly what I need!
07:47 SteffanW joined #git
07:50 fuchstronaut joined #git
07:51 vassagus joined #git
07:51 jceb joined #git
07:52 ahmedelgabri joined #git
07:54 osse I win again!
07:56 Andrew_K joined #git
07:58 Seveas Have some internet points 💯
07:59 osse I'd rather you donate them to charity in my name
08:00 Seveas I'll send it to the home for orphaned commits
08:03 AbstractArtist joined #git
08:03 osse the Fletcher Memorial Home for incurable tyrants and kings
08:05 mingrammer joined #git
08:05 mohabaks joined #git
08:05 osse styler2go: you're adjusting ~/.gitconfig, but what does .git/config say?
08:06 dreiss joined #git
08:06 styler2go how can i check that osse
08:07 osse styler2go: cat .git/config
08:07 osse styler2go: git config core.filemode
08:07 osse git config --list --show-origin
08:08 chll_ joined #git
08:09 styler2go osse, git config core.filemode says true.. although i want it to be false
08:10 osse that's what i suspected
08:10 osse fix it by removing --global
08:10 styler2go so why does this happen?
08:10 osse styler2go: because .git/config takes precedence over ~/.gitconfig
08:10 styler2go ah
08:11 vuoto joined #git
08:12 rossome joined #git
08:12 styler2go that one worked
08:12 styler2go thank you
08:13 mrBen2k2k2k joined #git
08:13 osse styler2go: you might want to set the global one back to true
08:15 styler2go yep, already set it back. wow and i was so mad that nothing worked :D
08:15 bongjovi joined #git
08:16 _ADN_ joined #git
08:16 vuoto joined #git
08:17 wrouesnel1 joined #git
08:20 lebara joined #git
08:24 maxorator joined #git
08:24 qballer joined #git
08:24 submitnine joined #git
08:25 jnewt is there any way to convert a merge --no-ff into a ff merge after the fact (ie there's been commits on top of the merge commit.
08:26 a_thakur joined #git
08:29 cdown joined #git
08:31 vuoto joined #git
08:34 renlo joined #git
08:34 alansaul joined #git
08:34 ToBeCloud joined #git
08:34 MattMaker joined #git
08:35 GavinMagnus joined #git
08:36 mohabaks joined #git
08:36 osse jnewt: g rebase merge~1  i think should do the trick
08:37 osse oops
08:37 osse git rebase merge^2 I guess
08:39 iahbxx joined #git
08:40 mohabaks joined #git
08:42 Kulrak joined #git
08:42 Kulrak joined #git
08:42 GenZai joined #git
08:42 Gonzo89 joined #git
08:44 ahmedelgabri joined #git
08:47 MattMaker joined #git
08:48 cdown joined #git
08:50 lebara joined #git
08:51 fuchstronaut joined #git
08:53 tgunr joined #git
08:53 mahendra_ joined #git
08:54 iamtakingiteasy joined #git
08:54 ichux_ joined #git
08:55 kurkale6ka joined #git
08:56 nilg joined #git
08:56 caswitch joined #git
08:56 caswitch left #git
08:57 weckl joined #git
08:57 quite joined #git
08:58 infra-red joined #git
08:58 quite I'm standing somewhere in a repo and want to list all files under source control, with paths. (something like `find /reporoot`. How?
08:58 vassagus joined #git
08:59 AbstractArtist joined #git
09:01 leeN joined #git
09:01 osse quite: git ls-files
09:02 quite osse: would not list all files unless I'm actually standing the "repo-root". I guess I'll have to script cd:ing to there and back..
09:03 osse quite: git ls-files --full-name :/
09:03 osse that last thing is not a smiley
09:03 osse in general you can use git -C foo to avoid having to cd
09:03 quite osse: ah, thanks! missed the need for colon there
09:04 denimsoft joined #git
09:04 hahuang61 joined #git
09:05 MattMaker joined #git
09:05 antipsychiatry- joined #git
09:05 antipsychiatry- left #git
09:06 lordjancso joined #git
09:09 mohabaks joined #git
09:09 osse quite: i added the --full-name because I assume that's what you want. but maybe since you're inside a subdir you want relative paths
09:10 anttim joined #git
09:12 davisonio joined #git
09:12 a_thakur joined #git
09:13 submitnine joined #git
09:14 ORiON-- joined #git
09:15 cqi joined #git
09:17 MattMaker joined #git
09:17 quite osse: yes, in the end omitting --full-name worked best
09:17 grayjoc joined #git
09:20 netj joined #git
09:22 LeMike joined #git
09:22 robotroll joined #git
09:22 lebara joined #git
09:23 fuchstronaut joined #git
09:23 stasmakarov joined #git
09:25 mrBen2k2k2k joined #git
09:26 a_thakur joined #git
09:26 Darcidride joined #git
09:27 WayToDoor joined #git
09:27 vrishab joined #git
09:29 vuoto_ joined #git
09:30 vuoto joined #git
09:30 vuoto joined #git
09:30 rbern joined #git
09:32 linuxmint joined #git
09:37 mingrammer joined #git
09:37 jceb joined #git
09:38 delboy1978uk joined #git
09:38 meZee joined #git
09:40 prestorium joined #git
09:41 jefus joined #git
09:41 MattMaker joined #git
09:42 meZee joined #git
09:42 delboy1978uk i startted a new branch to merge my branch with a colleagues. however i didnt realise his bvranch has code that's for a different release, and i've pushed my branch to another which shouldnt contain his code! how can i undo all my commits for this ticket?
09:44 spudowiar joined #git
09:44 nikivi joined #git
09:45 delboy1978uk if it helps, all my commit messages start with the same text, the release version and ticket number
09:45 SteffanW joined #git
09:45 meZee joined #git
09:46 madknight joined #git
09:46 matthiaskrgr joined #git
09:46 meZee joined #git
09:46 smile999 joined #git
09:47 meZee joined #git
09:47 j416 delboy1978uk: rebase and/or reset
09:50 Guest61233 joined #git
09:50 Guest61233 joined #git
09:50 lebara joined #git
09:54 eycsigfy joined #git
09:54 freimatz joined #git
09:55 freednars_ab joined #git
09:58 gechr joined #git
10:00 m0viefreak joined #git
10:01 elect I'd like to revert on one commit, but I get an error because of a single file, what should I do?
10:01 vassagus joined #git
10:02 delboy1978uk elect: what's the error?
10:03 elect I am using the gui from netbeans
10:03 elect it just say "because of: file.java"
10:03 tobiasvl !gui
10:03 gitinfo Graphical user interfaces are not supported here. If you want to get support, it needs to be through the git CLI. Reasons: 1) Because very few people here use the graphical interface. 2) Because giving instructions for GUI's is difficult. 3) The command line gives you a history of what commands you have executed.
10:03 elect ok i try by cli
10:05 elect git revert -m 1?
10:05 MattMaker joined #git
10:05 elect is the commit before
10:06 ahmedelgabri joined #git
10:06 delboy1978uk git reset --hard HEAD^ goes back one commit
10:06 MattMaker joined #git
10:07 charlespl joined #git
10:07 elect it worked
10:07 elect thanks
10:08 tobiasvl that's not the same as a revert though
10:08 tobiasvl !revert
10:08 gitinfo That's a rather ambiguous question... options: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard uncommitted changes in the working tree [git reset --hard]; c) undo committing [git reset --soft HEAD^]; d) restore staged versions of files [git checkout -p]; e) move the current branch to a different point(possibly losing commits)[git reset --hard $COMMIT]?
10:08 charlespl Hello. I got question. What if I add to staged area some files and then I will checkout to another branch. If I come back to the branch from which I checkout I will receive all changes I have staged?
10:08 charlespl or it will give me files from recent HEAD?
10:09 tobiasvl charlespl: the changes will !float over to the other branch when you check it out
10:09 gitinfo charlespl: If you have made a change in your working directory and have NOT YET COMMITTED, you may "float" that change over to another (`git checkout oldbranch`) or new (`git checkout -b newbranch`) branch and commit it there.  If the files you changed differ between branches, the checkout will fail.  In that case, `git stash` then checkout, and `git stash apply` and go through normal conflict resolution.
10:09 elect so after the reset, if I try to push it wants to pull first
10:10 elect I would like to avoid that
10:10 elect I'd like to just overwrite it
10:10 tobiasvl elect: you did a !rewrite of history so now you will need to !force_push to overwrite the remote history
10:10 gitinfo elect: Rewriting public history is not recommended. Everyone who has pulled the old history will have to do work (and you'll have to tell them to), so it's infinitely better to just move on. If you must, you can use `git push --force-with-lease <remote> <branch>` to force (and the remote may reject that, anyway). See http://goo.gl/waqum
10:10 gitinfo elect: If you need to overwrite the history of a remote git repository (very bad idea, see !rewrite), you can do so with `git push --force-with-lease`. Note the remote server may reject this. See man git-config and search for receive.denyNonFastForwards. Best practice is for upstream servers to denyNonFastForwards.
10:11 tobiasvl delboy1978uk: I would probably not advise people to reset --hard without also explaining how it differs from a revert
10:11 netj joined #git
10:11 elect I dont care about rewriting history
10:11 tobiasvl all right, then git push --force-with-lease
10:12 elect I would just like to push without pull and pollute my current status with the last wrong commit
10:12 elect ok
10:12 infra-red joined #git
10:13 vuoto joined #git
10:14 AbstractArtist joined #git
10:15 Kulrak joined #git
10:15 Kulrak joined #git
10:16 roelmonnens joined #git
10:17 circ-user-FePRT joined #git
10:17 stamina joined #git
10:19 delboy1978uk tobiasvl: ok, thanks and noted.
10:20 vuoto_ joined #git
10:21 vuoto__ joined #git
10:23 vuoto joined #git
10:24 alansaul joined #git
10:24 thekrynn_ joined #git
10:25 elect ok, history seems rewritten
10:25 elect and the last commit is correct
10:25 ahmed_elgabri joined #git
10:25 elect we did all of this from mine collegue pc
10:25 bruce_lee joined #git
10:25 dsdeiz_ joined #git
10:25 lebara joined #git
10:26 elect how can I now checkout that commit without detaching HEAD?
10:26 elect I just want to get that copy as it is
10:27 elect I am pulling
10:27 elect it has for a rebase/merge
10:27 elect shall I rebase, right?
10:27 elect I don't want to keep any of my changes
10:28 Mikerhinos joined #git
10:28 tubaman joined #git
10:29 tobiasvl elect: on the branch, git fetch; git reset --hard origin/<branch>
10:29 mingrammer joined #git
10:29 nikivi joined #git
10:31 tvw joined #git
10:31 tgunr joined #git
10:31 TinkerTyper joined #git
10:32 nikivi joined #git
10:33 Zoxc joined #git
10:34 nikivi joined #git
10:35 mahendra_ joined #git
10:36 delboy1978uk hi guys, trying to do a revert, it's asking for the parent commit for -m, but it wants a number, not 8ec2d4b, any help?
10:36 chandu20 joined #git
10:36 delboy1978uk or am i typing it wrong?
10:37 osse delboy1978uk: It's the parent number. It means you are reverting a commit
10:37 _ikke_ delboy1978uk: 1
10:37 delboy1978uk git revert 9af0a37 -m 8ec2d4b
10:37 osse In most cases you want 1
10:37 _ikke_ delboy1978uk: mainstream is always 1
10:37 delboy1978uk error: switch `m' expects a numerical value
10:37 _ikke_ osse: Do you know a case where this is not the case?
10:38 delboy1978uk i have two parents
10:38 _ikke_ Me too :P
10:38 delboy1978uk dont we all? :-P
10:39 _ikke_ delboy1978uk: A merge commit has two parents, you need to specify the number of the parent
10:39 osse _ikke_: The only case I can think of is when someone pulls and pushes, so their work becomes the first parent
10:39 mrBen2k2k2k joined #git
10:39 elect tobiasvl, it worked, thanks!
10:40 _ikke_ osse: But would that affect the parameter you specify?
10:40 zacts joined #git
10:40 M-shine joined #git
10:40 ahmed_elgabri joined #git
10:40 fuchstronaut joined #git
10:40 osse _ikke_: Yes, if you want to revert the work that guy did.
10:41 _ikke_ osse: iirc, if you would specify -m2, you would then revert your own changes
10:41 osse Because that guy's "feature branch" would be the first parent,
10:41 _ikke_ -m is mainline, the version you want to keep basically if I understand correctly what it does
10:41 osse Yes
10:43 osse _ikke_: What I mean is: let's say I don't merge my feature into master, but merge master into feature. Then I fast forward master and push. Later on if we wanted to revert my feature we would need -m2 because the feature branch is the first parent
10:44 osse This is weird, but the same thing can apply if you merge origin/master into master
10:45 osse And that often happens when newbies work and then pull & push
10:45 _ikke_ yeah
10:45 _ikke_ I see what you mean
10:45 delboy1978uk ok so revert -m. This option specifies the parent number (starting from 1) of the mainline and allows revert to reverse the change relative to the specified parent.
10:45 delboy1978uk so how do i know which is parent 1, which is parent 2?
10:46 _ikke_ delboy1978uk: The branch being merged into is always 1
10:47 delboy1978uk ok so in this case i need 2
10:48 _ikke_ delboy1978uk: Out of curiousity, why do you think that's the case?
10:50 sdothum joined #git
10:50 osse delboy1978uk: Do you want to revert what was merged in? Use -m1. Do you want to revert what was on the branch initially and only keep what was merged in? Use -m2
10:50 vuoto joined #git
10:50 _ikke_ right, that's a good description
10:51 rnsanchez joined #git
10:51 osse _ikke_: In that Scott Chacon talk I like to link to he suggests this workflow: When working on a feature you regularly merge from master to keep up to date. Once the feature is done you merge one final time, then merge feature into master. Consequently master is only ever fast-forwarded.  But I've never heard of anyone else using this workflow.
10:52 _ikke_ Right
10:52 _ikke_ osse: it makes sense if you are cooperating with meany people in a feature branch
10:52 _ikke_ making rebasing hard
10:52 rahtgaz joined #git
10:53 osse Imho the DAG (at least when visually presented) because less descriptive. It just sounds weird to me. But merging from master regularly itself makes senes
10:53 osse s/because/becomes/
10:54 cyan__ joined #git
10:55 geoid joined #git
10:55 lebara joined #git
10:57 cbreak osse: sounds very dumb.
10:57 gechr_ joined #git
10:58 cbreak merges of branches into which master was merged should not be fast-forward merged into master
10:58 osse Yeah that's what I'm thinking too
10:58 cbreak this would break features like first-parent, and create a very undesirable topology
10:58 cbreak I'd make sure to always specify --no-ff when merging such branches
10:59 mooru joined #git
10:59 vuoto joined #git
11:00 Munt joined #git
11:01 kingarmadillo joined #git
11:01 vuoto joined #git
11:02 vuoto joined #git
11:04 ShadeS joined #git
11:05 hahuang61 joined #git
11:05 vassagus joined #git
11:06 roelmonnens joined #git
11:08 jil joined #git
11:09 Dazzyo joined #git
11:10 lss8 joined #git
11:10 lss8 joined #git
11:11 malesca joined #git
11:11 Manuel15 joined #git
11:11 Manuel15 joined #git
11:11 tanuki joined #git
11:11 constantinexvi joined #git
11:12 HoierM joined #git
11:13 Perseus0 joined #git
11:13 submitnine joined #git
11:13 Remram[m] joined #git
11:14 Ineentho[m] joined #git
11:14 M-mistake joined #git
11:14 M-meznak joined #git
11:14 jeffreylevesque joined #git
11:14 lebara joined #git
11:14 nikivi joined #git
11:14 Perseus0 trying to push a file that had its size changed but it still acknowledges the previous file size, what to do ?
11:15 phanimahesh joined #git
11:16 osse Perseus0: what is "it" that acknowledges, and what does it matter?
11:16 Perseus0 matters because im pushing to github and it's over 100mb and doesn't let me
11:17 Perseus0 it's a .flv file
11:17 cbreak Perseus0: git doesn't care about file sizes
11:17 cbreak it just tracks them
11:17 cbreak (and it breaks if things are too big :)
11:17 Perseus0 github not git
11:17 bremner !github
11:17 gitinfo Note that git != github. Feel free to ask us about Github-specific features (Forks, Pull Requests, Wikis, etc), but there are no guarantees. There is a #github channel, which might help too (again, no guarantees)
11:17 osse Perseus0: you need to rewrite it so the file was never that big in the first place. changing it in a new commit won't help
11:18 Perseus0 i need to maintain its name tho
11:18 cbreak names don't matter for this
11:18 cbreak you can reuse it as much as you want
11:18 ojdo joined #git
11:18 Perseus0 so how do I remove the trace that it was x amount of size?
11:18 cbreak rebase -i for example
11:19 cbreak go to the commit where you added it
11:19 cbreak and remove it
11:19 cbreak (or replace it with the smaller version)
11:19 LeMike joined #git
11:19 chandu20 joined #git
11:19 cbreak (rebase -i, chose edit, then commit --amend, rebase --continue, ..., the usual)
11:19 Perseus0 the problem is that it's not there, github didn't allow me to push it because of the size
11:20 Perseus0 it's locally
11:20 davisonio joined #git
11:21 submitnine joined #git
11:21 cbreak Perseus0: doesn'tmatter
11:22 _tgunr joined #git
11:22 cbreak Perseus0: all of this is local op
11:22 Perseus0 ok let me give that a go ty
11:23 jil joined #git
11:24 ghoti_ joined #git
11:24 jil hello
11:24 delboy1978uk left #git
11:25 Junior joined #git
11:25 jil I've been doing a few tutorial on git and using it to clone stuff but It's still not clear how I can use it for developpement.
11:26 jil I want to developpement localy a wordpress site and commit changes to an online dev.site.  how do I do that ?
11:26 cbreak jil: make changes, stage changes for commit (with git add or similar), then commit.
11:26 lebara joined #git
11:26 cbreak all of this is local
11:26 cbreak use git push to push changes to somewhere else
11:26 tazle_ joined #git
11:27 tanuki joined #git
11:28 zeroed joined #git
11:28 zeroed joined #git
11:28 jil hum..
11:29 jil when I push the changes are they automaticaly taken in account ?
11:31 ljc joined #git
11:33 fahadash joined #git
11:33 ewet joined #git
11:34 cbreak jil: what?
11:34 cbreak changes are taken into account as soon as you add them, and they become history when you commit
11:34 cbreak pushing only transfers them into an other repository
11:34 Darren_ joined #git
11:34 cbreak usually on an other computer, i.e. a server
11:35 Akimb joined #git
11:37 mohabaks joined #git
11:38 jil cbreak: I probably do  not visualise correctly how git works.  On the local side I think I got it.
11:39 jil but on the distant site, it's not clear.  when I push my commit distantly files are updated on the distant site right ?
11:39 bremner jil: see !deploy
11:39 gitinfo jil: 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
11:39 nikivi joined #git
11:39 jil ok bremner I guess that's where my confusion was.
11:40 veilig2000 joined #git
11:40 jil thank you both
11:40 a_thakur_ joined #git
11:41 fuchstronaut joined #git
11:42 Mowgli joined #git
11:42 indistylo joined #git
11:44 chandu20 joined #git
11:44 cbreak jil: usually, no user visible files are touched
11:44 veilig2000 joined #git
11:45 Macaveli joined #git
11:48 anticom_ joined #git
11:48 beyertech joined #git
11:52 anticom_ left #git
11:53 Sceorem joined #git
11:53 anticom joined #git
11:55 mrBen2k2k2k joined #git
11:56 tomog999_ joined #git
11:58 jimi_ joined #git
12:04 huzhengchuan joined #git
12:04 synthroid joined #git
12:04 davisonio joined #git
12:07 denimsoft joined #git
12:07 jceb joined #git
12:08 Guest0 joined #git
12:09 checkItOut_ joined #git
12:11 vassagus joined #git
12:12 jeffreylevesque joined #git
12:14 rgrinberg joined #git
12:15 rgrinberg joined #git
12:16 zacsek joined #git
12:16 beyertech joined #git
12:17 buffal0 joined #git
12:17 Guest0 left #git
12:20 tvw joined #git
12:21 mohabaks joined #git
12:21 Gamecubic joined #git
12:23 buffal0 joined #git
12:23 rwb joined #git
12:24 diogenese joined #git
12:25 buffal0 joined #git
12:26 mingrammer joined #git
12:26 Impaloo joined #git
12:27 submitnine joined #git
12:27 ljc joined #git
12:29 tsia joined #git
12:29 buffal0 joined #git
12:29 rcc joined #git
12:29 rcc joined #git
12:30 buffal0 joined #git
12:31 _ADN_ joined #git
12:32 darkbit joined #git
12:34 ivan_on_trac joined #git
12:36 buffal0 joined #git
12:37 denimsoft joined #git
12:37 buffal0 joined #git
12:39 buffal0 joined #git
12:40 daniel_rose joined #git
12:40 buffal0 joined #git
12:41 phanimahesh joined #git
12:41 FunkyAss joined #git
12:41 nikio_ joined #git
12:42 nikio_ so i made a new github account and im trying to push a repo but it tells me permission denied to my old username
12:42 nikio_ i forgot how to fix this
12:42 nikio_ any tips?
12:42 airdisa joined #git
12:42 fuchstronaut joined #git
12:42 osse nikio_: #github. but that said, did you configure ssh keys on the new account?
12:43 nikio_ nope
12:43 osse oh, you use http? maybe wipe the credentialcache thingy
12:45 nikio_ yeh not sure what it was
12:49 jeffreylevesque joined #git
12:50 Akimbas joined #git
12:51 robotroll joined #git
12:51 raijin joined #git
12:52 daynaskully joined #git
12:53 ahmed_elgabri joined #git
12:55 Mowgli joined #git
12:55 WayToDoor joined #git
12:55 denimsoft joined #git
12:56 lebara joined #git
12:57 JamesBaxter joined #git
13:03 pcre joined #git
13:04 RustyShackleford joined #git
13:05 denimsoft joined #git
13:05 hahuang61 joined #git
13:05 f0rblaze joined #git
13:06 ExoUNX joined #git
13:06 pcre joined #git
13:07 cdg joined #git
13:10 submitnine joined #git
13:11 KnightsOfNi joined #git
13:11 nilg joined #git
13:11 cdown joined #git
13:12 cjbrambo joined #git
13:12 atomicb0mb joined #git
13:12 shaggycat joined #git
13:12 TooLmaN joined #git
13:13 satya4ever joined #git
13:14 sin3t joined #git
13:16 vassagus joined #git
13:19 veilig2000 joined #git
13:19 vuoto joined #git
13:20 gfixler joined #git
13:21 mingrammer joined #git
13:24 al-damiri joined #git
13:25 djb-irc joined #git
13:25 ahmed_elgabri joined #git
13:26 GodGinrai joined #git
13:27 circ-user-FePRT joined #git
13:28 mrBen2k2k2k joined #git
13:30 nikivi joined #git
13:30 zacts joined #git
13:32 ramsub07 joined #git
13:32 ahmed_elgabri joined #git
13:32 nikivi joined #git
13:34 lmat I just installed git for windows (linked to from git-scm.com/download/win, pointing to git-for-windows.github.io) for work. What is this?
13:34 lmat It put its own bash in...where is the package manager so I can install more stuff? Is this based on cygwin?
13:35 rnsanchez_ joined #git
13:36 ToxicFrog lmat: I'm pretty sure it's based on either gnuwin32 or mingw. There is no package manager. If you want a package manager you should be using cygwin; g4w only comes with what it needs to make git work.
13:36 lmat ToxicFrog: It has bash...
13:36 lmat or something that calls itself bash ^_^
13:36 ToxicFrog Yes, I know.
13:37 ToxicFrog Bash is not a package manager and does not require one.
13:37 phanimahesh joined #git
13:37 lmat ToxicFrog: It has vim.
13:37 ToxicFrog Looking at the g4w page, it's based on mingw.
13:37 lmat ToxicFrog: How do I install more packages?
13:38 bremner asked and answered?
13:38 ToxicFrog Does mingw have a package manager? It didn't last time I looked, but that was a while ago. If so, g4w might come with it, and it might work.
13:38 lmat ToxicFrog: I see. okay, thanks!
13:38 ToxicFrog If g4w doesn't come with it, you might be able to download it from the mingw website and add it to your g4w install.
13:38 ToBeFree joined #git
13:39 lmat ToxicFrog: aye.  What a miserable web we weave trying to keep everyone using Microsoft Windows...
13:39 ToxicFrog My approach has always just been to use cygwin, because it comes with git *and* with a bunch of other stuff not available in g4w, like an X server.
13:39 lmat ToxicFrog: right. I'm looking for tmux ^_^
13:40 ToBeCloud joined #git
13:40 lmat ToxicFrog: Oh, cygwin will come with the real git, right? Not some Microsoft Windows wrapper/port/emulation whatever?
13:40 ToxicFrog I can't even interpret that question.
13:41 ToxicFrog What would make it "not the real git"?
13:41 lmat ToxicFrog: :-) Not important
13:41 bremner Unreal Git Tournament
13:41 ash_workz joined #git
13:41 ToxicFrog That said, before you leap into cygwin, double check that it has tmux! I don't actually know if it does.
13:42 lmat ToxicFrog: Git (like, the real git: https://github.com/git/git) isn't compilable on Microsoft Windows, so some magic souce has to be added. But with Cygwin, this magic sauce isn't necessary?
13:42 Rish_ joined #git
13:42 osse M-m-m-m-merge commit!
13:42 osse Godlike!
13:42 EricPoe joined #git
13:42 ORiON-- joined #git
13:43 fuchstronaut joined #git
13:43 ToxicFrog lmat: so, first of all, if it's format and interface compatible with "real git", I don't see why it should matter
13:43 ToxicFrog But as it happens, yes, it is "real git"
13:43 lmat ToxicFrog: I see.
13:43 ToxicFrog Both g4w and cygwin take the approach of installing enough stuff that git has an environment to run in that "looks like" linux
13:44 ToxicFrog (and based on some extremely perfunctory googling, there are tmux packages for both mingw and cygwin)
13:45 lmat ToxicFrog: excellent.
13:45 Akimb joined #git
13:45 mooru joined #git
13:46 datasoop joined #git
13:46 datasoop left #git
13:46 diogenese joined #git
13:47 Dougie187 joined #git
13:47 FilipNortic_ joined #git
13:47 fahadash joined #git
13:48 staafl joined #git
13:50 lmat ToxicFrog: Sorry to bother you again; do you have a link for tmux/mingw? :-)
13:50 SporkWitch lmat: https://google.com/
13:50 johnmilton joined #git
13:51 lmat SporkWitch: I tried, I can only find tmux for cygwin :-( (and lots of references to alternate Microsoft Windows terminals). Do you know where to find it?
13:52 SporkWitch lmat: what distro are you on? lol
13:52 lmat SporkWitch: Microsoft Windows :-(
13:52 GodGinrai SporkWitch: "for cygwin".  Obviously windows
13:52 johnmilton joined #git
13:52 ToxicFrog lmat: I googled for "tmux mingw package" and ended up at https://github.com/Alexpux/MSYS2-packages/ which contains a tmux package. I assume this means that msys has something package-manager-esque you can use to install it.
13:53 GodGinrai lmat: If you can find it for cygwin, then isn't that good?  You should be using cygwin anyways
13:53 ToxicFrog I can't be more specific than that because I haven't used mingw/msys in at least five years, and last time I did, it didn't have a package manager at all
13:53 ToxicFrog So, consult your mingw/msys documentation.
13:53 lmat GodGinrai: I'm using git-for-windows, so I don't think I'm using cygwin.
13:53 osse those packages are for MSYS2
13:53 lmat ToxicFrog: no sweat. Thanks for the link!
13:53 osse which use a port of pacman for package management
13:53 SporkWitch GodGinrai: i misunderstood what he meant by it, i saw "i can only find tmux for windows" with the emphasis on "for windows" implying he needed it NOT on windows lol
13:53 ToxicFrog osse: is that different from the MSYS that git4win is based on?
13:53 osse it's nice but I prefer Git for Windows because it's much faster
13:54 lmat SporkWitch: ah; I got tmux elsewhere no problem: pacman -Syu tmux; done. ^_^
13:54 ToxicFrog Clearly the solution here is just to give up entirely on developing on windows~
13:54 SporkWitch lmat: hence my reaction lol
13:54 osse ToxicFrog: the difference is using msys.dll versus using mingw directly. I don't know much about it, but mingw is much faster, but sacrificing some unixy stuff that msys ports
13:54 SporkWitch ToxicFrog: it's the sane decision most people make ;)
13:54 lmat ToxicFrog: That's been my stance for some time now!
13:54 osse lmat: https://msys2.github.io/
13:54 SporkWitch though windows DOES have sourcetree
13:54 ToxicFrog osse: I thought msys was just a collection of utilities on top of mingw?
13:54 lmat osse: Ah, so that msys.dll is some emulation layer?
13:55 kaldoran joined #git
13:55 pcre_ joined #git
13:55 osse ToxicFrog: msys is. msys2 is more like, but not the same as, cygwin
13:55 ToxicFrog osse: like, my information is way out of date, but last time I used it mingw32 was a collection of *nix utilities built to run natively on windows, and msys was a "development environment" (gcc, make, bash, a terminal emulator) built using mingw
13:55 ToxicFrog aah
13:55 SporkWitch ToxicFrog: that seems so frankensteinian...
13:56 lmat osse: Thanks; very helpful!
13:56 osse I am confused by this as well
13:56 lmat So git-for-windows is mingw, and not compatible with msys2, right?
13:56 ToxicFrog I note with some irritation that "msys2.github.io" does not at any point explain what msys2 is or why I would it over cygwin or plain msys >.<
13:56 osse but with msys2 you get stuff like tmux and zsh which msys won't offer. but it's a bit slow, and git in particular is muc hslower
13:56 sangy joined #git
13:56 AaronMT joined #git
13:57 lmat osse: Thanks.
13:57 osse ToxicFrog: that's just the home page of the installer. https://sourceforge.net/projects/msys2/
13:57 ToxicFrog SporkWitch: it makes sense for what it was built for. mingw32 was "I'm stuck on windows but still want grep". msys, the Minimal SYStem, was "I'm stuck on windows but still want to build programs with make and gcc"
13:57 fakenullie Git status was annoyingly slow on Windows
13:57 SporkWitch ToxicFrog: icic
13:57 ToxicFrog Unlike Cygwin it doesn't want to be a complete "linuxalike" environment, but be just complete enough to use Linux development tools to build windows native programs
13:58 SporkWitch ToxicFrog: and i'm guessing it dates back to when VMs weren't cheap and practical
13:58 SporkWitch oh! gotcha
13:58 phanimahesh joined #git
13:58 ToxicFrog (and that emphasis on "windows native" means there's no msys1 equivalent to cygwin.dll; stuff you build with it links directly against MSVCRT or equivalent with no translation layer)
13:58 SporkWitch icic
13:59 ToxicFrog The mingw32 project also produced the linux-host, windows-target gcc cross compiler, if you wanted to do all your development on linux but still get a windows .exe or .dll at the end of it.
13:59 lucasem joined #git
14:00 osse I want to build git on mingw myself, so it works better with msys2. right now if I avoid installing git in msys2 and adjust paths and whatnot to find Git for Windows I sometimes get weird DLL errors when git wants to launch a pager for example
14:00 phaleth joined #git
14:01 ToxicFrog (you could also build windows native programs with cygwin, but by default everything you built would depend on cygwin.dll to provide windows implementations of things like fork(). You can turn that off -- -mno-cygwin, IIRC -- but if you want windows binaries without extraneous dependencies it was often easier to just use the mingw/msys compiler.)
14:01 jefus joined #git
14:04 Impaloo joined #git
14:06 crose joined #git
14:07 veilig2000 joined #git
14:08 hiroki I'm setting up unattended GIT clone/fetch from a BASH script (no need for jenkins or similar for this purpose). To make sure there are no issues with connectivity, authentication, etc. I've created a 'test-connectivity' repository and perform a 'git ls-remote --exit-code -h' to make sure all is well before I actually continue with any clone/fetch for the configured real repositories. I'm wondering if there are
14:08 hiroki more efficient ways of doing this, anyone has any tips or similar setups they like to share ?
14:09 bremner why is this better than just cloning the repo you want, and seeing if it fails?
14:10 l4v2 joined #git
14:10 robotroll joined #git
14:12 hiroki Because at one point the script deletes the repository from disk (and execute a repository specific terminate script that deletes the changes it has made, before actually deleting the repository from disk) if it decides it does not exist anymore on the remote because it got removed intentionally. I'd want to make sure it will detect the intentional removal of the remote repository without a failed network
14:12 hiroki connection or failed credentials to get in the way of that automated decision
14:12 hiroki In short: I have my reasons.
14:16 chandu20 joined #git
14:16 ResidentBiscuit joined #git
14:19 askhader left #git
14:21 vassagus joined #git
14:21 roelmonnens joined #git
14:24 Darren_ joined #git
14:25 shinnya joined #git
14:26 denimsoft joined #git
14:27 SporkWitch hiroki: you're talking about the repo as a whole, or a specific branch? if branches, man git prune
14:27 gitinfo hiroki: the git-prune manpage is available at http://jk.gs/git-prune.html
14:27 * SporkWitch loves this bot
14:29 SporkWitch hiroki: if you mean for a whole repo, that seems like a really bad idea; you can't prove a negative, so the only *safe* way to do this would be to place a file on the remote saying "hey there, this repo was deleted"
14:29 SporkWitch (or some similar kind of POSITIVE result we can check for).  This would of course result in lots of cruft, over time.
14:30 nehaljwani joined #git
14:30 roelmonn_ joined #git
14:30 nehaljwani joined #git
14:30 nehaljwani Hi! Is there a mechanism in git to show a file in a remote repository? Something on the lines of ls-remote?
14:31 dcpc007 left #git
14:31 enleeten joined #git
14:32 SporkWitch nehaljwani: if you do a fetch, you have a local copy you can check out, diff against, etc.
14:32 SporkWitch nehaljwani: not sure about listing files, though
14:32 SporkWitch (i assume it must be possible, since zsh offers autocompletion for files in a branch)
14:32 hiroki SporkWitch: that's a good point indeed, i'm tryign to incude a positive by verifying accessibility is OK, assuming that a removed repo is indeed enough to base it's decision on, which in this case is OK as i'm now and ever will be the only one with GIT access
14:32 hiroki induce*
14:33 davisonio joined #git
14:34 SporkWitch hiroki: the problem is that git doesn't distinguish between non-git and non-existent directories, AFAIK
14:34 CheckDavid joined #git
14:35 eivarv joined #git
14:36 daniel_rose joined #git
14:36 hiroki SporkWitch: What do you mean exactly with a non-git directory ? It's for a whole repo BTW. AFAIK a git can report the fact a remote repository has not been found ?
14:37 fuchstronaut joined #git
14:37 SporkWitch hiroki: a repo IS just a directory.  All it is is a directory with some extra files in a particular, git-defined format, that tells git about previous versions, commit comments, etc.
14:37 SporkWitch hiroki: it detects a git dir by whether it can find /path/you/give/.git
14:38 clemf joined #git
14:38 SporkWitch hiroki: if that .git can't be found, for any reason, it gets back a non-exist error AFAIK
14:38 SporkWitch hiroki: AFAIK, it can't distinguish between foo/.git not existing, and foo/ not existing
14:39 hiroki SporkWitch: I see what you mean. That should be fine as i'm meaning to check for foo, not foo/.git or foo/<somethingelse>
14:39 SporkWitch hiroki: further, it would not be able to distinguish between it not existing, and ../../foo getting -x on it so you can't traverse; the files are still there, you just can't see them or get there anymore
14:40 prestorium joined #git
14:40 hiroki SporkWitch: I see your point, I need to further re-think this
14:40 daniel_rose joined #git
14:40 timvishe` left #git
14:41 SporkWitch hiroki: sorry to burst bubbles, i just see a LOT of potential problems here, since you're taking positive action based on negative feedback
14:41 timvisher joined #git
14:41 hiroki SporkWitch: That's why i'm here, to get a second opinion from more experienced people ;-)
14:42 SporkWitch hiroki: honestly, i'm still a git noob, when it comes to much more than basic usage and theory lol (though I have copied and modified some hooks lol).  What I AM experienced with is breaking things, and thinking about how to :) lol
14:42 hiroki SporkWitch: Always good to meet a fellow-break-lover
14:43 stuh84 joined #git
14:43 ahmed_elgabri joined #git
14:43 SporkWitch hiroki: especially the beats ;)
14:43 hiroki lol
14:43 * SporkWitch wonders if that "breakbeat" joke is rimshot-worthy... lol
14:44 SporkWitch hiroki: out of curiosity, what's the real end goal?
14:44 SporkWitch hiroki: to nuke your locals when the remotes are gone?
14:46 paws_ joined #git
14:47 hiroki SporkWitch: I wrote a GIT fetch/reset/deploy wrapper containing a list of secondary GIT repos to fetch/reset/deploy. It traverses that list in the wrapper script and for loops over it to do perform some checks and based on the outcome of it fetch/reset/deploy a repo or terminate/delete_repo_dir/remove_repo_crontab. That way I can control deployment or termination using only GIT commits / GIT repo_removals on
14:47 hiroki pfSense nodes that don't have access to Jenkins setups. And with deployment I mean running a deployment script in the repo applying changes, adn with termination I mean running a termination script in the repo reverting those changes.
14:48 kuldeep joined #git
14:49 tsia joined #git
14:49 SporkWitch interesting... i feels like there's probably a better way to do this
14:49 kuldeep how can i make `git diff` only show diff of non-delete files only?
14:49 SporkWitch *it feels
14:49 grawity kuldeep: --irreversible-delete
14:49 SporkWitch hiroki: have you looked into things like puppet / ansible / spacewalk / chef ?
14:49 grawity kuldeep: aka -D
14:50 hiroki If there is, i'm surely interested in that
14:50 SporkWitch hiroki: it sounds like those might be better suited for your remote configuration needs, particularly with multiple systems
14:50 YourTheBombNow23 joined #git
14:50 kuldeep thanks grawity
14:50 hiroki I did, and I ruled them out as i'd want the maintenance of the pfSense nodes to be as independent and autonomous as possible. Therefore no dependencies on cfg. mgmt. tools, jenkins, etc. -> KISS
14:51 SporkWitch hiroki: ansible in particular is pretty awesome, since the remotes don't need anything more than ssh and an account with the necessary permissions.  ansible runs EXCLUSIVELY on the "control" machine, and uses ssh (preferably with key-based auth) to send remote commands and files
14:51 hiroki The less I have to change and incur in actually having done by scripts, the less can go wrong, the better
14:51 mrBen2k2k2k joined #git
14:51 SporkWitch that really sounds like ansible is the way to go, lol
14:52 hiroki SporkWitch: In that scenario the Ansible controller is a SPOF and from a security perspective very bad
14:52 SporkWitch you'd have one machine that installs ansible, you don't need to install anything on the remotes ll
14:52 SporkWitch hiroki: no more so than with the git implementation; anyone with that SSH key would be able to achieve the same thing
14:52 hiroki SporkWitch: additionally, I tend to never push, but only fetch, based on the logic that a push is never guaranteed to do what you intend, but a fetch will at least go through because it's initiated from a working machine
14:52 ice9 joined #git
14:53 hiroki If a machine is down or inaccessible, a push will be wasted resources
14:53 morenoh149 joined #git
14:53 ice9 is it possible to pass dir path to git to perform an action, or i must cd first into that dir?
14:53 hiroki SporkWitch: Sure, but at least the damage will be limited to that specific node, instead of gaining access to a whole infrastructure
14:53 osse ice9: git -C dir status
14:54 ice9 thanks osse
14:54 hiroki SporkWitch: I use CFengine BTW for hosts, puppet is horrible considering it's memory footprint, and I dislike Ansible's PUSH philosophy.
14:54 badloop joined #git
14:54 badloop joined #git
14:54 osse ice9: there's also --git-dir and --work-tree (and GIT_DIR and GIT_WORK_TREE env vars9
14:54 SporkWitch hiroki: i would debate that, but i do understand your concerns.  I'm just trying to come up with something simpler, as your solution feels unnecessarily complex.
14:55 Impaloo joined #git
14:56 SporkWitch hiroki: no matter what your solution is going to have a single point of failure, unless you're seriously proposing that you go to different machines to update different machines lol; presumably you have all of these machines listed as remotes, and even if you're using different keys for each, you've still got all the keys there.  That's what i mean when i say it's not really any different from
14:56 SporkWitch ansible.
14:56 hiroki SporkWitch: appreciate the efforts. Based on your input I think i'll adapt the wrapper to not decide to remove a repository local GIT clone based on the remote existence, but instead on the listed array of GIT repo's that's already defined in the wrapper itself coupled with a local directory existence check to decide the discrepancies
14:56 Impaloo joined #git
14:57 SporkWitch that sounds like a sane workaround; i'd also have it log the issue somewhere, so you can check up on it and decide if it just couldn't connect, or if it was really gone (maybe comment out the line for that machine in the script, instead of deleting it?)
14:57 SporkWitch (and i'm sorry, but it really does sound like you're reinventing ansible lol)
14:58 badloop left #git
14:58 mingrammer joined #git
14:59 mohabaks joined #git
15:00 Juventus joined #git
15:00 hiroki SporkWitch: Logging including locking etc, is and should always be included
15:01 fozniku joined #git
15:01 hiroki SporkWitch: You can check the contents of the script later and decide for yourself the resemblence with Ansible ;-)
15:01 Juventus left #git
15:01 The_Phoenix joined #git
15:01 SporkWitch hiroki: fair enough :)
15:02 dmto joined #git
15:02 SporkWitch hiroki: i'm just going by how it sounds and the concerns you've raised lol
15:02 SporkWitch and yes, logging is your friend
15:02 The_Phoenix joined #git
15:02 hiroki SporkWitch: That's why seeing things from a different perspective sometimes helps
15:02 prestorium joined #git
15:02 hiroki Like seeing someone with long blond hair from the front instead of the back, to understand it's actually a dude and not a woman
15:02 hiroki blonde*
15:03 Akimb joined #git
15:03 ramsub07 joined #git
15:04 SporkWitch hiroki: it's funny you mention long hair lol https://keybase.pub/sporkwitch/me.jpg
15:04 The_Phoenix joined #git
15:04 SporkWitch hiroki: it's about halfway to my ass now, lol
15:04 hiroki SporkWitch: At least it's not blonde :p
15:05 SporkWitch hiroki: and i have been catcalled from behind; was hilarious.  Carload of guys "hey baby, wanna go for a ride?" so i turned around with a full beard and in the gruffest voice i could muster, "how much?" lol
15:05 The_Phoenix joined #git
15:05 SporkWitch hiroki: they peeled out, blew through the light, and got pulled over by the cop that was around the corner lol; was priceless
15:05 hiroki See, always helps to look at things from a different perspective
15:06 The_Phoenix joined #git
15:06 encod3 joined #git
15:06 hahuang61 joined #git
15:07 SporkWitch hiroki: what if it's a line segment? :P
15:07 hiroki SporkWitch: You read it in Arabic
15:07 SporkWitch #pedantry
15:07 SporkWitch hehe
15:08 The_Phoenix joined #git
15:08 skapunker joined #git
15:09 The_Phoenix joined #git
15:12 le_melomane joined #git
15:12 ice9 left #git
15:14 zacts joined #git
15:15 ahmedelgabri joined #git
15:16 submitnine joined #git
15:16 bernardio joined #git
15:16 fuchstronaut joined #git
15:18 raijin joined #git
15:18 subhojit777 joined #git
15:18 rossome joined #git
15:19 subhojit777 joined #git
15:19 InfoTest joined #git
15:19 jstein_ joined #git
15:19 thiago joined #git
15:20 subhojit777 joined #git
15:20 dreiss joined #git
15:20 daniel_rose joined #git
15:21 subhojit777 joined #git
15:21 sanketdg joined #git
15:22 dpower joined #git
15:22 subhojit777 joined #git
15:22 Balliad joined #git
15:22 cyan__ joined #git
15:23 subhojit777 joined #git
15:24 subhojit777 joined #git
15:26 vassagus joined #git
15:28 kadoban joined #git
15:29 BSaboia joined #git
15:31 kpease joined #git
15:31 YourTheBombNow23 joined #git
15:31 keyan_ joined #git
15:31 vvaris joined #git
15:32 darkpassenger joined #git
15:33 spudowiar joined #git
15:33 darkpassenger I am on a branch with multiple commits (including some merges) and i'd like to keep all the change but do only one clean commit
15:33 darkpassenger how can I do this without resetting the merge
15:34 t0by joined #git
15:34 netcarver joined #git
15:34 darkpassenger i tried git reset --soft HEAD~x_num_of_commits
15:34 darkpassenger but it undoes the merge ?
15:34 Rish joined #git
15:34 vvaris joined #git
15:34 darkpassenger and then resolving the conflicts, recommit and force push on my remote branch
15:34 darkpassenger that works, but its not fun to do at all
15:35 madewokherd joined #git
15:36 GavinMagnus joined #git
15:36 GavinMagnus left #git
15:36 YourTheBombNow23 joined #git
15:37 GavinMagnus joined #git
15:37 daniel_rose joined #git
15:39 charlespl joined #git
15:39 kpease joined #git
15:39 eivarv joined #git
15:40 t0by joined #git
15:41 jfelchner joined #git
15:52 durham joined #git
15:55 Finnoloid joined #git
15:56 glebihan joined #git
15:58 rkazak joined #git
15:59 vrishab joined #git
16:00 sanketdg how do I get the commits of the last 3 months?
16:01 osse sanketdg: git log --since='3 months ago'
16:03 sanketdg osse: ah thanks!
16:03 dendazen joined #git
16:03 xlegoman joined #git
16:04 _ikke_ date -d @$(git rev-parse --since='3 months ago' | cut -d= -f2)
16:04 _ikke_ To see where it get's parse to
16:05 daniel_rose joined #git
16:05 osse smart
16:05 osse TIL
16:06 _ikke_ I just found out
16:06 _ikke_ Was wondering if there was a way to do it
16:06 _ikke_ And figured if there was such a thing, it would be in rev-parse
16:06 kpease joined #git
16:07 _ikke_ It does require a git repo though
16:07 osse at first i though this would find the oldest commit itself, not the interpretation of the argument
16:07 osse but it's dank
16:08 osse testgit() ( cd "$(mktemp -d)" && git init &>/dev/null && git "$@"; }
16:08 osse shit
16:08 osse testgit() ( cd "$(mktemp -d)" && git init &>/dev/null && git "$@"; )
16:08 mrBen2k2k2k joined #git
16:08 crayon joined #git
16:09 dendazen guys i have a directory 'AB' in branch A with files and subdirectories, git log -- directory 'AB' shows me all the commits done to that directory. Now what would be the best way to merge all those commits to another branch 'AC'?
16:09 bremner you can cherry-pick individual commits, but really only merge whole refs (e.g. branches)
16:11 dendazen oh so what would be the best way.
16:11 osse what do you want to accomplish, exactly?
16:11 dendazen i do not want to mess up branch 'AC' as it's prod branch
16:12 astrofog joined #git
16:12 dendazen i want that directory 'AB' to be present in branch 'AC'
16:12 dendazen but not introduce all the other stuff in branch A
16:13 dendazen to branch 'AC'
16:13 osse dendazen: git checkout AC; git checkout A -- AB; git commit -m 'Add dir AB from branch A'
16:13 bronson joined #git
16:13 GodGinrai yea, that's probably simpler than trying to merge the separate commits in
16:13 durham joined #git
16:14 vrishab joined #git
16:14 renlo joined #git
16:14 dendazen hmm.
16:16 ramsub07 joined #git
16:18 fuchstronaut joined #git
16:18 LeBlaaanc joined #git
16:19 ash_workz joined #git
16:19 rgrinberg joined #git
16:20 rbr joined #git
16:20 nitric joined #git
16:20 tyreld joined #git
16:20 ytl joined #git
16:22 Mowgli joined #git
16:22 jil left #git
16:23 durham joined #git
16:24 Gsham joined #git
16:27 cjohnson So I've got a branch and we have been working on our git workflow, and the commits are rather wonky from branch to branch. When I try to rebase master from my branch it's a huge drawn out process and after about 20 manual conflict resolutions I aborted. Instead I just want to take a diff of my branch vs master, which accurately shows my changes, and commit that as a new commit in a new branch off master, so when I go to merge I don't bring all
16:27 cjohnson How can I create that patch?
16:27 cjohnson would it just be git diff master..branch   ?
16:28 cjohnson and git apply thatfile.patch
16:28 osse cjohnson: if you apply that then master will be *identical* with branch
16:29 ramsub07 joined #git
16:29 cjohnson er well, git checkout master && git checkout newbranch && git apply olddiff.patch
16:29 cjohnson now newbranch contains my changes without any tainted history
16:30 cjohnson git checkout -b newbranch
16:30 osse ok, in that case newbranch is identical to branch
16:30 cjohnson perfect
16:31 vassagus joined #git
16:32 eivarv joined #git
16:32 gargola joined #git
16:36 fullstack joined #git
16:39 daniel_rose joined #git
16:40 Electrometro joined #git
16:40 davisonio joined #git
16:44 Serus joined #git
16:44 ASOLAr joined #git
16:44 Serus how do I force crlf line endings on .appveyor.yml?
16:45 Serus google has failed me
16:45 Electrometro joined #git
16:46 choki joined #git
16:47 ojacobson Serus: man gitattributs
16:47 ojacobson Serus: man gitattributes
16:47 gitinfo Serus: the gitattributes manpage is available at http://jk.gs/gitattributes.html
16:47 ojacobson sorry, keyboard issues
16:48 GavinMagnus left #git
16:48 Serus oh, I found my problem
16:48 Serus using * text=auto interferes with it
16:49 Serus how do I make it ignore specific files? order doesn't seem to matter
16:51 EvilPeng1 joined #git
16:53 fate_ joined #git
16:53 Serus oh, I found it
16:53 Serus that works
16:53 thiago joined #git
16:53 thiago joined #git
16:54 Serus well, I'm out again, my poor pi is about to die if I stay in this channel
16:54 Serus left #git
16:55 HoierM joined #git
16:55 DRiOscar joined #git
16:55 hahuang61 joined #git
16:56 Sasazuka joined #git
16:56 ASOLAr joined #git
16:58 avar Why when I do "echo a >file; commit; echo b >file; commit; git show" would git show "Binary files..differ"
16:58 skapunker joined #git
16:58 avar Probably some configuration at a distance, but I can't find what it is
16:58 osse avar: .gitattributes maybe
16:59 avar Ah, there, I found it, thanks
16:59 osse git check-attr -a -- file
16:59 spudowiar How can I apply a commit that isn't pushed to the server?
17:00 spudowiar e.g. work on a branch, committing as normal etc. but preventing certain commits from being pushed to the server
17:00 avar osse: great, thanks!
17:00 kadoban spudowiar: Don't push them?
17:00 osse spudowiar: the only way is to rebase it so it's always the newest commit
17:00 benbergman joined #git
17:00 spudowiar osse: hmm, that's actually a brilliant idea! :)
17:01 cbreak joined #git
17:01 kingarmadillo joined #git
17:02 spudowiar osse: I could just commit each time then swap the two commits
17:02 spudowiar So the extra commits are always on top!
17:02 Noldorin joined #git
17:02 osse spudowiar: https://github.com/Osse/git-​scripts/blob/master/git-move
17:03 spudowiar thanks! :)
17:03 spudowiar I've got this horrible hack for faster (but insecure) firmware flashing
17:03 osse I sometimes use this. I'm on my branch "myown", make a commit, do git move master and it will apply the newest commit on master then rebase
17:03 spudowiar Awesome for development but firmware built with that CANNOT be signed EVER
17:03 osse or rebase whatever you give it.
17:04 osse sounds awesome
17:05 Darcidride joined #git
17:08 tristanp joined #git
17:09 skylite joined #git
17:10 renlo joined #git
17:10 inflames joined #git
17:10 ahmedelgabri joined #git
17:14 neersighted joined #git
17:16 Achylles joined #git
17:17 emPi joined #git
17:17 khmarbaise joined #git
17:18 Groscheri joined #git
17:18 InfoTest1 joined #git
17:18 fuchstronaut joined #git
17:19 renlo joined #git
17:21 Darcidride joined #git
17:23 Groscheri joined #git
17:23 Groscheri joined #git
17:25 neersighted joined #git
17:28 eivarv joined #git
17:29 LionsMane joined #git
17:29 bttf i'm having trouble figuring out how to fetch this PR submitted by another person
17:29 bttf i've tried `git fetch origin pr/6/master` to no avail, PR: https://github.com/bttf/em​ber-scroll-operator/pull/6
17:31 EvilPeng1 joined #git
17:32 kadoban bttf: Maybe something like ... git fetch origin refs/pull/6/head:refs/remotes/origin/pr/6  ? semi-guessing. I always add the fetch-spec to my config when I care about PRs, here's that if you want it: https://gist.github.com/piscisaureus/3342247
17:34 m0viefreak joined #git
17:34 _ADN_ joined #git
17:35 GodEater` joined #git
17:35 NeonSynth joined #git
17:36 vassagus joined #git
17:37 fate_ joined #git
17:38 diegoviola joined #git
17:38 YourTheBombNow23 joined #git
17:39 grayjoc joined #git
17:39 bttf kadoban much obliged
17:40 m0viefreak joined #git
17:41 kingarmadillo joined #git
17:41 fate_ joined #git
17:42 t0by joined #git
17:45 rtyler left #git
17:45 daniel_rose joined #git
17:46 morenoh149 joined #git
17:47 jstein__ joined #git
17:49 grayjoc joined #git
17:51 nioncode_ joined #git
17:51 fiete_ joined #git
17:51 githubmgw joined #git
17:52 githubmgw left #git
17:52 dreiss joined #git
17:53 githubmgw joined #git
17:53 githubmgw left #git
17:55 i7c joined #git
17:57 eivarv joined #git
18:01 atomicb0mb joined #git
18:02 chris349 joined #git
18:02 chris349 I did a git pull which caused a file to be deleted, how can I undo that?
18:03 osse chris349: resurrect the file or undo the whole pull ?
18:03 i7c chris349, you wanna undo the pull or just get that single file back?
18:03 i7c osse heh
18:04 Drugo joined #git
18:04 ahmedelgabri joined #git
18:05 chris349 I just want the file back. It was accidently commited and them removed from the repo (config file)
18:05 Drugo I see there are some projects to use git with shared hostings. Do you know which is the best one?
18:06 Manuel15 joined #git
18:06 Manuel15 joined #git
18:06 vishal Drugo: likely gitolite
18:06 i7c chris349, git checkout <commit-where-it-still-existed> -- path/to/file
18:06 osse chris349: git checkout X -- file, where X is a commit the file was in
18:06 i7c damn
18:06 i7c osse, I figured I wait some time and see if you go ahead.
18:07 i7c Then I thought you went for a dump so I'll help him.
18:07 i7c Doublewrong!
18:07 osse it was only a mysqldump
18:07 i7c Only? Those take longer often.
18:08 i7c osse, so I'll go for a walk now, support hard!
18:08 chris349 But if I do that will it complain next time I do a git pull? Or since it knows I did the checkout git pull will delete it again?
18:08 Drugo vishal, i'll take a look, thanks
18:08 osse chris349: git will not delete it again
18:08 osse unless someone else deletes it again again
18:10 dreiss joined #git
18:13 renlo left #git
18:14 lalilulelo joined #git
18:15 kingarmadillo joined #git
18:17 benbergman joined #git
18:18 ntnn joined #git
18:19 denimsoft joined #git
18:21 vassagus joined #git
18:21 Somelauw joined #git
18:23 Xizor joined #git
18:23 eivarv joined #git
18:25 l4v2 joined #git
18:27 AndrewAlexMac joined #git
18:28 hahuang61 joined #git
18:31 atomicb0mb joined #git
18:31 paws_ joined #git
18:32 HanClinto joined #git
18:35 rangergord joined #git
18:37 alansaul joined #git
18:39 mooru joined #git
18:39 rangergord Why wouldn't pull/fetch also run "git remote update origin --prune"? What's the justification for keeping out of date info about the remote?
18:40 osse rangergord: just safety I guess
18:41 osse you jsut expect something to be there, it disappears behind your back, you go "ooops"
18:41 osse But if you liked you should've put a branch on it
18:41 osse also, you can 'git config --global fetch.prune' to do this by default
18:41 rangergord aah, thanks. Updating my config.
18:42 HanClinto joined #git
18:43 rbr joined #git
18:45 rangergord I wish git had 2 APIs, simple mode for 99.9% of people, and full mode for Linus and friends
18:45 rangergord I wonder how many millions of hours of productivity have been lost to this API
18:46 flaguy48 left #git
18:47 t0by joined #git
18:47 grayjoc joined #git
18:49 KevinMGranger It does. Plumbing and porcelian.
18:49 KevinMGranger If you want to call that "full mode"
18:50 joki joined #git
18:50 rangergord the porcelain has a bad API too
18:51 rangergord Like if you're rebasing, during a conflict your options are --abort and --continue. If you're merging, during a conflict your options are --abort and...git add + git commit.
18:51 osse hmm, they could add merge --finish
18:51 rangergord The API is not designed with the user in mind, it's designed with the programmer (of git) in mind
18:51 mathu joined #git
18:52 rangergord there is a huge productivity cost. That XKCD is popular for a reason, it's 100% true
18:52 osse in git's defense this is just the sort of thing git status actually tells you about
18:52 j416 rangergord: there have been discussions about redesigning the user interface
18:53 leeN joined #git
18:54 bsanford joined #git
18:54 nopea joined #git
18:54 warthog9 joined #git
18:55 j416 rangergord: if I remember correctly, the conclusions have been along the lines of "yes, we might be able to improve it slightly, but keeping the flexibility it would still probably end up being on about the current level"
18:55 rangergord j416: well, they should consider 2 sets of APIs as I said. Don't try to make a one-size-fits-all API
18:55 nopea when I do a 'git status' it shows me which files I have deleted and modfied since the last commit.  However I do not see the directories or files I created.  How do I know for sure those will be created when I commit?
18:56 osse nopea: they should be listed as "untracked", unless they are ignored.
18:56 j416 rangergord: do you have a good suggestion of such an API?
18:56 stasmakarov joined #git
18:56 nopea osse: I will check again but I did not seem them at all
18:57 osse nopea: you can also configure git status to not care about untracked stuff.
18:57 osse make sure that's disabled
18:57 nopea osse: doh - yes I see them now... must have missed it before.
18:57 osse \o/
18:57 nopea osse: if I do 'git add .' then they should be added to the tracked files no?
18:57 osse yes
18:58 davisonio joined #git
18:58 nopea osse: perfect thanks.  And do you know if 'git rm filename' removes the file from the tracking?
18:58 osse nopea: it does, but also from disk.
18:58 osse nopea: use --cached to avoid physically removing it
18:59 nopea osse: cheers.  I have already deleted the file from the disk, so I need to remove the tracking
18:59 osse ah, then you can even use "add" I think
18:59 vuoto joined #git
18:59 osse as i "add this change"
18:59 osse *as in
18:59 mathu facebook put out a mercurial extension that i find myself wanting in git. `hg amend` will create a new amended commit (leaving the original and its descendents alone) and `hg amend --rebase` will do the same but rebase all the descendents onto the new commit. i know this changes the hashes of all the commits involved -- would this totally screw up a github PR if i `amend --rebase`d one of the commits and
18:59 nopea perfect thanks
18:59 mathu pushed?
19:00 rbr joined #git
19:00 rangergord j416: I don't have a full API planned in my head, no, but I definitely think "this would have been better if" often when I'm using git. For example, "git remote sync" instead of git pull + git remote update.
19:00 dan2k3k4 joined #git
19:00 osse mathu: the PR would simply reflect the new contents of the branch
19:00 osse mathu: it's actually pretty common to force-push to make a PR prettier
19:01 osse mathu: a PR is just "merge this branch into that", so there's a level of indirection, sort of. if "this branch" is updated (in any way) then the PR is too
19:02 mathu okay, cool. i didn't know if github was tracking commits by hash or anything; if it's just overlaying the source branch on the dest branch, all is well
19:02 johnmilton joined #git
19:02 osse nopea: if you forget, git status can remind you
19:03 rangergord j416: I don't think wanting to create a replacement API will work. All the people that went through the effort of learning git won't want to change to something that only helps other people. This includes all git devs. It needs a separate new API that's simpler.
19:03 rangergord *that only helps other people, and is less powerful
19:04 efco joined #git
19:04 osse rangergord: https://github.com/felipec/git#git-fc
19:06 vassagus joined #git
19:06 rangergord osse: maybe something like that. But a 3rd party option doesn't fix the issues, those things remain obscure forever. Git has to support it built-in for it to get mindshare.
19:06 eivarv joined #git
19:07 mathu another thing i miss from mercurial: `hg shelve [list of files]` where `hg shelve` is analogous to `git stash` :P
19:07 osse git config alias.shelve stash
19:08 osse hmm, except for the list of files I think
19:08 mathu osse: stash doesn't take filename arguments
19:08 mathu yeah
19:08 mathu can do by hunk but that is cumbersome :P
19:09 osse this sounds like yet another one of those things where I think "how hard could it be?", start hacking, get frustrated after a few hours and stop
19:09 rgrinberg joined #git
19:10 mathu haha yeah, i considered trying but decided i didn't want to find out if filename is accessible within whatever the code uses for hunks
19:11 daynaskully joined #git
19:14 j416 rangergord: there have been attempts at it.
19:14 j416 rangergord: not very successful attempts, I think
19:14 j416 rangergord: https://people.gnome.org/~newren/eg/ for instance
19:15 latenite joined #git
19:15 j416 I think it only makes it more complex
19:15 tgunr joined #git
19:15 j416 just adds another layer
19:15 javajoe joined #git
19:15 latenite Hi folks, I got this error (totaly out of the blue) : git add -A .
19:15 latenite error: bad index file sha1 signature
19:15 latenite fatal: index file corrupt
19:16 The_Phoenix joined #git
19:16 latenite What did I do wrong and how can I fix this?
19:17 The_Phoenix joined #git
19:18 The_Phoenix joined #git
19:19 ResidentBiscuit You might need to reset the index
19:19 ResidentBiscuit Reset the changes out of the index, kill the index file
19:20 fuchstronaut joined #git
19:20 The_Phoenix joined #git
19:22 The_Phoenix joined #git
19:22 buffal0 joined #git
19:23 The_Phoenix joined #git
19:25 The_Phoenix joined #git
19:25 vrishab joined #git
19:29 g3ntleman joined #git
19:29 g3ntleman I got 2 branches one is 1 year old, one is up to date since 1 minute ago, do I merge old one with new one, or new one with old one?
19:29 mingrammer joined #git
19:32 ojacobson you do whichever one implements your intentions
19:32 ojacobson what do you intend to do?
19:32 ojacobson (natural language description, not git description)
19:33 g3ntleman ojacobson: I intend to make that old branch up to date with new branch
19:34 g3ntleman old wiht new merge gives me 4k changed files, new with old gives me 14k
19:39 grayjoc joined #git
19:40 stamina joined #git
19:42 neca joined #git
19:43 leolrrj joined #git
19:44 leolrrj hey guys, I commit my work in a detached head... and then after that I ran git branch master... now I obviously cannot see my changes, but I also cannot find my detached head
19:44 leolrrj is it possible to recover it?
19:45 Seveas git fsck will find it
19:45 Seveas as will git reflog
19:46 leolrrj Seveas: ok I have my commit here on reflog: 6d884af --- how can I recover it at merge it to my branch?
19:46 leolrrj recover it and merge *
19:46 leolrrj Seveas: thank you buddy for the answer  :)
19:47 kbs joined #git
19:50 vrishab joined #git
19:50 leolrrj Seveas: so, it's just git checkout -b recovery-branch 6d884af --- right?
19:50 dsdeiz joined #git
19:51 buffal0 joined #git
19:52 Seveas that would work
19:52 Seveas but if it's just one commit, I'd simply cherry-pick it
19:53 leolrrj Seveas: can i run this: git merge 6d884af from my master branch?
19:56 underpantz joined #git
20:01 paul424 joined #git
20:04 nikivi joined #git
20:04 nopea can I do 'git rm .' to remove all deleted files from being tracked?
20:04 nopea or would that remove all files - I just want to remove the ones reported as 'deleted' from git status
20:04 osse nopea: git add -u
20:05 nopea osse: cheers - that will remove just the deleted files?
20:05 nopea from being tracked
20:05 osse no it will stage all changes, but not add untracked files
20:08 bartosz_ joined #git
20:10 cdg joined #git
20:11 nopea osse: thanks - I think that worked.  Have to remember those commands ;-)
20:14 buffal0 joined #git
20:17 vassagus joined #git
20:18 mrBen2k2k2k joined #git
20:20 javajoe joined #git
20:21 fuchstronaut joined #git
20:21 ahmedelgabri joined #git
20:22 EvilPeng1 joined #git
20:23 th3fallen joined #git
20:24 th3fallen Hey guys is there a way to remove all local branches that have been merged using githubs sqash merge method by chance?
20:24 th3fallen i used to use git fresh but that doesnt work anymore with that :(
20:25 atomicb0mb joined #git
20:25 prestorium joined #git
20:27 ojacobson you can identify the branches that _do not_ appear in `git cherry`, I suspect
20:27 _ikke_ I don't think that works with squash merge
20:27 _ikke_ It relies on patch id, which is lost with squashing
20:28 osse how about attempting a merge and then detecting if it did anything?
20:28 _ikke_ diffing with 3 dots is better
20:28 _ikke_ See if it introduces new changes
20:28 _ikke_ although, that can also fail
20:28 th3fallen i think i found a way
20:28 Guest55 joined #git
20:29 th3fallen by removing all branches tracking remotes that are no longer there
20:29 th3fallen git fetch --prune && git branch -r | awk '{print $1}' | egrep -v -f /dev/fd/0 <(git branch -vv | grep origin) | awk '{print $1}' | xargs git branch -D
20:32 kingarmadillo joined #git
20:36 th3fallen now to try to figure out how the hell to shove that into a git alias
20:37 _ikke_ th3fallen: when you start with a !, it's executed by a shell
20:37 th3fallen _ikke_ i have it in !sh -c 'code'
20:37 th3fallen but i get awk: syntax error at source line 1
20:37 th3fallen context is
20:37 th3fallen >>>  <<<
20:37 th3fallen awk: illegal statement at source line 1
20:37 th3fallen missing }
20:37 th3fallen whoops sorry for the spam
20:37 osse no need for the sh -c
20:38 osse only adds more quoting issues
20:38 osse besides <() is a bash feature
20:38 th3fallen so just !git fetch --prune && git branch -r | awk '{print $1}' | egrep -v -f /dev/fd/0 <(git branch -vv | grep origin) | awk '{print $1}' | xargs git branch -d
20:38 th3fallen will d it?
20:38 osse put it in a file named git-foo, put #!/bin/bash at the top and drop it in $PATH. Much saner
20:39 th3fallen putting a bat in front of it just echos it back out _ikke_
20:39 th3fallen am too anal for that osse must waste my time to make it an alias
20:39 th3fallen :(
20:41 th3fallen guess i could do git branch -vv | grep ': gone]' | awk '{print $1}' | xargs git branch -d as well
20:41 osse at least combine grep | awk
20:41 osse git branch -vv | awk '/: gone]/{print $1}'
20:42 th3fallen thanks osse
20:43 th3fallen git alias's has issues with awk though so now to figure out how to drop that from the command
20:43 cyphase joined #git
20:44 rossome joined #git
20:45 oncall-pokemon joined #git
20:46 osse th3fallen: nah
20:47 osse th3fallen:  git config --global alias.rekt '!git branch -vv | awk '\''/: gone]/{print $1}'\'' && :'
20:47 osse let git take care of the quoting in .gitconfig
20:47 osse git config alias.rekt '!git branch -vv | awk '\''/: gone]/{print $1}'\'' | xargs git branch -d && :'
20:48 th3fallen last one wont add
20:48 th3fallen first did though
20:48 osse what do you mean
20:48 th3fallen running the second command wont add the alias to my git config
20:48 osse i forgot --global
20:48 gargola joined #git
20:49 th3fallen oh son of
20:49 th3fallen that worked
20:49 th3fallen thanks osse
20:50 osse a nicer way to do with would be git for-each-ref, but I can't get the "gone" or equivalent information from it
20:50 falkenb0t joined #git
20:50 th3fallen also has to be a upper case D since git wont know if the branch is deleted
20:50 th3fallen but yeah it works
20:50 th3fallen thanks man
20:52 l4v2 joined #git
20:52 vassagus joined #git
20:54 preyalone joined #git
20:55 diegoaguilar joined #git
20:55 SwiftMatt joined #git
20:56 diegoaguilar Hello, is there any git built in way to grep a diff
20:56 circ-user-FePRT joined #git
20:56 _ikke_ git log -S
20:56 _ikke_ (pickaxe in the logs
20:56 _ikke_ man git log
20:56 gitinfo the git-log manpage is available at http://jk.gs/git-log.html
20:57 _ikke_ That would search for commits where either the string was added or removed
21:00 cyphase joined #git
21:00 EvilPeng1 joined #git
21:01 clemf joined #git
21:01 th3fallen left #git
21:02 davisonio joined #git
21:05 rahtgaz joined #git
21:10 crayon joined #git
21:10 cyan__ joined #git
21:16 RustyShackleford joined #git
21:17 offlim joined #git
21:17 mingrammer joined #git
21:18 ichux_ joined #git
21:23 ahmedelgabri joined #git
21:25 gruetzkopf joined #git
21:26 prestorium joined #git
21:26 veilig2000 joined #git
21:28 tgunr joined #git
21:28 javajoe joined #git
21:29 fk joined #git
21:29 Guest48460 left #git
21:33 ramsub07 joined #git
21:35 crayon joined #git
21:40 _ADN_ joined #git
21:42 disi i changed a file then committed (A). then i renamed that file and committed (B). now i want to revert A, but i am getting a conflict. how do i handle this?
21:43 boombatower joined #git
21:43 ToBeCloud joined #git
21:46 cd-rum joined #git
21:47 kaldoran joined #git
21:48 Dougie187 left #git
21:48 awt joined #git
21:49 NeverDie joined #git
21:49 disi this seems too hacky: http://stackoverflow.com/a/11894144/822302
21:49 vishal disi: in 'B', was the only thing you did the rename?
21:50 vishal if the hacky thing works for you, it doesn't sounds too bad
21:51 fakenullie revert B, revert A, revert revert A
21:51 Finnoloid joined #git
21:51 vishal I was going to suggest reset --hard to before B, and then revert/edit etc A
21:51 disi vishal: B has a few renames
21:51 fakenullie I mean revert revert B
21:51 fakenullie revert B, revert A, revert revert B
21:51 neanderslob joined #git
21:51 vassagus joined #git
21:51 fakenullie then squash
21:51 vishal ^ that works too
21:53 disi hmm... ok so im not too far off in the weeds. i thought of those, but i cant help but think git has some more automatic, history-preserving way
21:54 Sasazuka joined #git
21:54 clemf joined #git
21:55 Goplat joined #git
21:56 rossome joined #git
21:56 dan2k3k4 joined #git
21:56 anuxivm joined #git
21:57 cyphase joined #git
21:58 _ikke_ Most commands have follow-renames options
21:59 cd-rum_ joined #git
21:59 osse I just sent in a patch to add the "gone" logic to for-each-ref
21:59 fuchstronaut joined #git
21:59 osse so that that guy earlier doesn't need to awk and grep and whatever else the output of branch -vv
21:59 JamesBaxter joined #git
22:00 osse I am 85% sure that makes me an awesome person
22:04 andersoldahl joined #git
22:06 disi _ikke_: hmmm i dont see anything about that in the man page... was it added recently?
22:08 dendazen joined #git
22:08 gruetzkopf joined #git
22:08 ahmedelgabri joined #git
22:09 UTAN_dev joined #git
22:11 inflames joined #git
22:12 mingrammer joined #git
22:12 Achylles joined #git
22:12 rbern joined #git
22:15 neanderslob joined #git
22:15 neanderslob_ joined #git
22:17 M-ou-se joined #git
22:17 neanderslob joined #git
22:18 mycyber joined #git
22:21 sangy joined #git
22:23 vassagus joined #git
22:24 jstimm joined #git
22:28 Serus joined #git
22:28 Serus I'm back again
22:29 mrpackethead joined #git
22:29 gruetzko- joined #git
22:29 Serus I have some sensitive info in a few commits on the reflog, how can I remove them properly?
22:30 Serus Properly meaning the reflog stays usable so you can still restore/read stuff from it.
22:31 madewokherd joined #git
22:32 Serus Since the commits are not visible on history, would the command be git reflog delete --rewrite SHA-hash-here?
22:40 jnewt_ joined #git
22:43 veilig2000 joined #git
22:46 diegoaguilar joined #git
22:47 fk joined #git
22:51 NeonSynth_ joined #git
22:53 chid joined #git
22:55 vassagus joined #git
22:57 EvilPenguin joined #git
22:58 cdown joined #git
22:59 chid joined #git
23:00 cdown_ joined #git
23:02 alansaul joined #git
23:03 osse Serus: FYI the reflog is local to the repo only
23:03 ljc joined #git
23:03 Serus you mean to my locally checked out repo?
23:04 osse yes
23:04 Sasazuka__ joined #git
23:04 osse the reflog is not pushed. when other people clone they won't see your reflog. etc etc
23:06 Serus alright
23:06 cdown joined #git
23:06 osse the only way for other people to see your sensitive info is if you pushed before you rewrote AND they fetched before you force-pushed over it AND they would have to know what to look for in the first place
23:06 mingrammer joined #git
23:06 osse that said, you can nuke your whole reflog if you want to. i'm not sure how to remove individual entry from it
23:07 Serus ah, yes, there's only one entry in the log
23:08 Serus so what I would need to do is simply kill my fork after this
23:08 Xizor joined #git
23:08 awt joined #git
23:08 tgunr joined #git
23:09 ramsub08 joined #git
23:10 gruetzkopf joined #git
23:13 Derperperd joined #git
23:13 mrBen2k2k2k joined #git
23:14 cdown_ joined #git
23:15 fuchstronaut joined #git
23:15 _ADN_ joined #git
23:16 imanc joined #git
23:17 fk joined #git
23:18 aidalgol joined #git
23:18 airdisa joined #git
23:19 javajoe joined #git
23:19 zero7 joined #git
23:19 bedouin joined #git
23:19 javajoe left #git
23:19 da5id joined #git
23:20 falkenb0t joined #git
23:25 Darren_ joined #git
23:27 ffabi joined #git
23:27 jimi_ joined #git
23:28 hahuang61 joined #git
23:28 tgunr joined #git
23:29 The_Phoenix joined #git
23:30 A_F_K left #git
23:31 spudowiar joined #git
23:37 mablae_ joined #git
23:37 hexafraction joined #git
23:45 alansaul joined #git
23:45 user083 joined #git
23:46 EvilPeng1 joined #git
23:49 nanga joined #git
23:52 madewokherd joined #git
23:55 cdown joined #git
23:55 spudowiar joined #git
23:56 Remram[m] joined #git
23:56 Remram[m] joined #git
23:59 latenite joined #git
23:59 Derperperd joined #git

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