Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-07-29

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

All times shown according to UTC.

Time Nick Message
00:01 f3r70rr35f joined #git
00:01 johntramp ah i see, thanks
00:04 krstf joined #git
00:05 Literphor joined #git
00:11 anuxivm joined #git
00:15 leehambley joined #git
00:16 tesuji_ joined #git
00:16 neeca joined #git
00:17 mingrammer joined #git
00:20 boombatower joined #git
00:22 crayon joined #git
00:24 __Myst__ joined #git
00:24 __Myst__ Question
00:24 __Myst__ What's the difference between a rebase and a merge?
00:24 kadoban What history looks like afterwards, mostly.
00:25 __Myst__ Should I use a merge 90% of the time?
00:25 kadoban Depends on your workflow I guess.
00:25 lucasem rebase 95% of the time
00:25 EvilPenguin joined #git
00:25 kadoban I doubt that much rebase is generally a good idea.
00:25 lucasem what makes you think that?
00:26 lucasem merging is almost always a bad idea
00:26 __Myst__ lucasem: Why?
00:27 lucasem __Myst__:  if you granularize your work into commits, then you can play your commits on top of a different base.
00:27 kadoban __Myst__: People like very linear histories for some reason, even when that's not what happened ...
00:27 __Myst__ lucasem: Don't I lose the history of there being a branch there?
00:27 osse yes
00:28 __Myst__ Using github
00:28 __Myst__ rather
00:28 __Myst__ Using any git server
00:28 lucasem osse: no
00:28 __Myst__ Won't I have to force push after a rebase?
00:28 __Myst__ I'm not a Jedi
00:28 lucasem you don't lose history of your branch as long as you keep it around
00:28 lucasem you don't have to force push after a rebase, generally. Though it depends on context
00:28 osse sure if you keep the unrebased version around...
00:29 osse __Myst__: !merge_vs_rebase
00:29 gitinfo __Myst__: [!merge_or_rebase] For a detailed discussion of the dis-/advantages of merge and rebase, and when to best use which, see: https://coderwall.com/p/ny1hia -- and here's how Linus explains which one to use: http://www.mail-archive.com/dri-devel@lists.sourceforge.net/msg39091.html
00:30 rgrinberg joined #git
00:30 aidalgol joined #git
00:32 kadoban That stuff from linus looks nice and coherent.
00:32 red__ joined #git
00:33 lucasem merge makes sense for maintainers who deal with other peoples' code, yeah. If you aren't a maintainer and directly handling someone else's commits, you should really be rebasing.
00:34 edude03 joined #git
00:35 __Myst__ I have a 3 person project
00:35 __Myst__ Should we rebase or merge?
00:35 lucasem I work on large codebase with many contributors. I never merge.
00:35 lucasem I rebase my changes on top of master and push.
00:35 mellernoia joined #git
00:37 __Myst__ do you delete the branch afterwards?
00:37 kadoban lucasem: I disagree. I use merges even on solo projects. It's nice to keep track of sepearate feature work.
00:37 lucasem Then what are you doing in your commits? <1.0> <2.0> or something?
00:37 kadoban Huh?
00:38 lucasem Like, your commits should say what they do
00:38 lucasem you should still use branches, though.
00:38 kadoban They do, that doesn't mean there isn't a higher level of what the mid-level goal of a group of commits is.
00:38 lucasem (while you work on your features)
00:39 __Myst__ What about the issue of a feature branch getting lagged behind?
00:39 __Myst__ A sin
00:39 __Myst__ While one person works on their branch
00:39 __Myst__ Another merges their own branch into master
00:39 __Myst__ and the first person's branch doesn't have the new changes?
00:40 lucasem Then the first person can rebase and merge
00:40 __Myst__ lucasem: how?
00:40 __Myst__ `git checkout my_branch && git rebase master && git merge master`?
00:40 lucasem first two yes, not the last one.
00:41 __Myst__ doesn't git rebase master play my_branch on top of master?
00:41 lucasem yes
00:41 interne7y joined #git
00:42 lucasem no merge is necessary, but as kadoban said, some people use it to keep track of "mid-level goal[s]"
00:42 __Myst__ lucasem: Oh no
00:43 __Myst__ I want to play MASTER's new commits on my branch
00:44 lucasem __Myst__ on top, not on bottom? why?
00:44 rgrinberg joined #git
00:44 __Myst__ lucasem: Imgine this situation
00:44 __Myst__ There are two people working on a project, three branches
00:44 __Myst__ master, my_branch, their_brainch
00:44 jlebar What's the plumbing magic that I should use to translate a user-provided rev to a range?  That is, the function would turn "HEAD" into "HEAD^..HEAD" (or something equivalent), but would leave "foo..bar" unchanged.
00:44 __Myst__ their_branch gets merged into master
00:44 __Myst__ Now I want to replay the latest commits to master on my_branch
00:44 __Myst__ How can I do that?
00:45 ojacobson jlebar: most rev-list-accepting commands use `git rev-list` to parse specs
00:45 ojacobson __Myst__: `git merge`, most conservatively
00:45 lucasem __Myst__: you should rebase your branch to have the new changes on master.
00:45 __Myst__ `git checkout my_branch && git merge master` is what I do rn
00:45 ojacobson (`git fetch origin && git merge origin/master` for example)
00:45 kyan joined #git
00:46 ojacobson the resulting history will record all of the incremental integration decisions you made, which may be what you want
00:46 ojacobson and may not
00:46 __Myst__ ojacobson: sure, I can keep the commits
00:46 ojacobson (git is an authorial tool for imposing order on the primary sources that are editor sessions; it's not so much that there are wrong ways to use it as there are incomprehensible histories you can write with it)
00:46 __Myst__ I just want to get the latest update sinto m branch
00:46 ojacobson Do you want to record that you did that, or not?
00:46 ojacobson Is that fact of future interest, basically
00:47 __Myst__ ojacobson: I'm fine with either
00:47 lucasem (internet cut out, oops)
00:48 ojacobson If you believe the fact will be of future interest, or if reauthoring the history of your branch is a problem for other reasons, then `git merge` is probably the right authoring tool
00:48 lucasem Why do you want to replay their commits onto yours?
00:48 lucasem Why not merge after you complete?
00:48 ojacobson if you believe it will only be a distraction, and if reauthoring your branch is otherwise a non-problem, then `git rebase` is probably the right tool
00:48 jimi_ joined #git
00:48 ojacobson If your beliefs are at neither point, then you have more work to do :)
00:48 ojacobson git can author nearly any history you want, provided you can articulate what it should look like.
00:49 daveomcd joined #git
00:50 crayon joined #git
00:50 __Myst__ lucasem: Sometimes I want to use some new functions in my code
00:50 qpdb joined #git
00:50 __Myst__ To simplify my workflow
00:51 vassagus joined #git
00:51 jlebar ojacobson, It doesn't seem like git rev-list will do this thing?  I think I would like behavior like 'git show', where you can pass a single commit or a range.
00:51 ojacobson I believe git-show has its own logic. I strongly suspect it's "try rev-parse, and if that fails, try rev-list" or similar, but read the source :)
00:52 jlebar Yeah, I was looking, then I remembered why I'm a C++ programmer.  :)
00:52 * jlebar will look harder
00:54 lucasem __Myst__:  then you should really either (a) make a commit after your merge which rewrites your stuff to use the new function; (b) use those new functions when fixing the merge conflicts
00:55 anuxivm left #git
00:56 lucasem (I prefer (a) in that scenario. Nothing like a good 'ol refactor!)
00:56 __Myst__ joined #git
00:57 Gsham joined #git
00:57 jlebar Or I can just call "git show --quiet --pretty=%h USER_INPUT"
00:57 * jlebar likes that
01:02 hahuang61 joined #git
01:04 edude03 joined #git
01:06 afusion joined #git
01:07 PioneerAxon joined #git
01:08 zivester joined #git
01:08 PresidentBiscuit joined #git
01:08 EvilPenguin joined #git
01:11 Furai joined #git
01:12 mingrammer joined #git
01:12 crayon joined #git
01:12 EduardoMartins joined #git
01:13 cdg_ joined #git
01:14 durham_ joined #git
01:15 shgysk8zer0 joined #git
01:16 cqi joined #git
01:17 adino joined #git
01:17 oich joined #git
01:20 citizen-stig joined #git
01:21 takwas joined #git
01:21 Furai joined #git
01:22 Mannn joined #git
01:24 Furai joined #git
01:24 lindenle joined #git
01:27 PioneerAxon joined #git
01:28 thiago joined #git
01:30 Furai joined #git
01:32 crayon joined #git
01:35 Doginal joined #git
01:36 edude03 joined #git
01:36 jost_ joined #git
01:36 ekinmur joined #git
01:36 Derperperd joined #git
01:41 docnvk joined #git
01:46 adino joined #git
01:46 durham joined #git
01:46 snowkidind joined #git
01:49 pks joined #git
01:49 sleepee joined #git
01:50 eijk joined #git
01:50 harish_ joined #git
01:52 _sjs joined #git
01:52 pwnz0r joined #git
01:56 vassagus joined #git
01:57 doebi1 joined #git
01:58 irregular joined #git
01:58 irregular hi guys I'm running into this error fatal: unable to access 'https://user@github.com/myrepo.git/': Failed to connect to {IP address} port 8080: Connection refused
01:59 irregular when I run `git push`
01:59 irregular I've been pushing to this repo for weeks previously but am suddenly running into it now today
01:59 harish_ joined #git
02:01 d^sh joined #git
02:01 kadoban port 8080? That doesn't make sense.
02:01 kadoban irregular: What's 'git remote -v' ?
02:03 rewt !github
02:03 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)
02:06 edude03 joined #git
02:06 mingrammer joined #git
02:07 NeverDie_ joined #git
02:07 oich_ joined #git
02:09 devilboy joined #git
02:11 ahmed_elgabri joined #git
02:12 byteflame joined #git
02:13 Asiajey joined #git
02:15 hmmmm joined #git
02:15 hmmmm can anyone please help me??
02:15 lucasem !ask
02:15 gitinfo Yes, it's okay to ask questions here.... in fact, you just asked one! ;-) Pretty much any question is fine. We're not terribly picky, but we might be asleep. Please be patient and you should get an answer soon.
02:15 hmmmm well the last time i tried to ask everybody just ignored my question
02:16 hmmmm i am trying to restore a file from an earlier commit in history, by doing `git checkout HEAD~4 filepathhere`
02:16 hmmmm the expected result is that it restores the older version of that file to the staging area, under "changes to be committed"
02:16 hmmmm if i do this with 99% of the files, it works absolutely perfectly as expected
02:17 NeverDie joined #git
02:17 PioneerAxon_ joined #git
02:17 hmmmm however if i do this with one file in particular, it will both add the old revision of the file to the staging area like i want, but then immediately change the working area to the file at the current revision
02:17 hmmmm how can i stop it from doing this?  it's causing a major problem for me
02:18 pwnz0r joined #git
02:18 mehola joined #git
02:18 lucasem what do you mean by you second-to-last message? Could you elaborate?
02:19 kadoban hmmmm: Post a !transcript that shows it? Intersperse with many 'git status' calls and whatever shows the behavior you're saying.
02:19 gitinfo hmmmm: Please paste (using https://gist.github.com/ or https://vomitb.in/ ) a transcript ( https://gist.github.com/2415442 ) of your terminal session so we can see exactly what you see
02:19 rchavik joined #git
02:19 hmmmm i would love to but it's on another computer
02:19 lindenle joined #git
02:19 bluezone joined #git
02:19 hmmmm i'll try to re-create the message
02:19 kadoban Trying to diagnose a problem on a computer you can't even test things on is rather fruitless, yeah.
02:19 hmmmm by manually typing it
02:19 altendky joined #git
02:19 hmmmm i can test it, i just can't paste it
02:20 kadoban The other computer can't connect to the internet?
02:20 hmmmm correct
02:20 hmmmm it's basically air-gapped
02:21 EduardoMartins joined #git
02:21 jimi_ joined #git
02:21 hmmmm i can make up git messages that describe what's happening, but it can't be the exact thing
02:21 kadoban A manually typed transcript isn't terribly interesting honestly ... make one on the other computer and then copy it on a USB drive or something?
02:21 hmmmm that's especially not possible
02:21 fahadash joined #git
02:21 lucasem hmmmm: just write it out in english, but be very clear
02:22 kadoban If not ... I mean do what you can, but no promises I'll look that much at it.
02:22 hmmmm writing to a flash drive is strictly prohibited and is not allowed by software installed on the operating system
02:25 byteflame joined #git
02:25 sdaugherty joined #git
02:25 shaggycat_ joined #git
02:27 hmmmm okay I typed it up
02:27 hmmmm https://paste.fedoraproject.org/397163/46975924/
02:27 lucasem oh i see
02:28 lucasem hmmmm: did you already have something staged for that file?
02:28 hmmmm no
02:28 hmmmm absolutely nothing at all
02:28 hmmmm no modifications, at HEAD
02:28 bdonnahue joined #git
02:28 bdonnahue is there a way to change the commit mesage after the fact? typos are annoying
02:28 kadoban hmmmm: What is 'git diff -- src/foobar' ?
02:29 hmmmm kadoban:  the entire file is deleted and then readded
02:29 hmmmm in other words, some kind of change in the line endings
02:29 NeverDie_ joined #git
02:29 lucasem hmmmm: what's in 'git diff --staged'?
02:30 hmmmm all the lines in the file being deleted, and then all the lines in the file being readded with "^M" appended to them all
02:30 kadoban hmmmm: Is the file open in a text editor maybe?
02:30 MZAWeb_ joined #git
02:30 hmmmm it is not
02:31 lucasem are you using windows?
02:31 hmmmm yes.
02:31 Sucks joined #git
02:31 hmmmm i did open it in visual studio to look at it
02:31 * lucasem sighs at carriage returns
02:31 chachasmooth joined #git
02:31 hmmmm and then all of a sudden it's marked as modified in git
02:31 hmmmm i didn't touch it at all!
02:31 hmmmm and i closed it out of visual studio
02:31 lucasem hmmmm: do you have cygwin?
02:31 hmmmm tried to revert it
02:31 hmmmm not installed, no
02:31 kadoban hmmmm: Is it marked as modified before you open it in anything?
02:32 hmmmm yes
02:32 hmmmm okay so
02:32 hmmmm about that
02:32 Goplat joined #git
02:32 hmmmm i needed to make a commit really quick and so i stopped git add "individual path here" for all the files i want to commit and leaving out that file that got changed somehow
02:32 hmmmm so i just did git commit -a --amend
02:32 jpX joined #git
02:33 hmmmm but now that i have more time, i need to go back and fix this
02:33 hmmmm so the idea was to bring this file back to a commit before HEAD, and amend HEAD
02:34 hmmmm now it doesn't show up as modified not staged for commit anymore
02:34 hmmmm but now it's like it's impossible to change this file
02:34 kadoban I'm totally lost what you mean ...
02:34 lucasem hmmmm: is anything about that file staged?
02:34 lucasem if so, git reset it
02:35 lucasem are there any changes in your working dir? If so, git checkout.
02:35 hmmmm okay
02:35 leehambley joined #git
02:35 hmmmm there's a backstory to all this
02:35 hmmmm i was explaining how it got into this state before
02:35 Sucks joined #git
02:35 hmmmm but as of right now i am at HEAD on a clean branch
02:35 hmmmm no modifications whatsoever
02:35 hmmmm and i try to revert that file to an earlier revision and it will not let me
02:36 lucasem cool. So 'git checkout <old-commit> <my-file>' , then what?
02:36 kyan joined #git
02:36 hmmmm what i pasted in the pastebin link...
02:36 lucasem even without opening the file or anything?
02:36 hmmmm correct
02:36 lucasem are you sure/can you try again?
02:36 hmmmm i have tried this a zillion times
02:36 NeverDie joined #git
02:37 hmmmm and i am about to chuck it at a wall if it were physically possible to
02:37 kadoban Will not let you? I don't understand. Is this a new thing, or the same thing you were asking about before?
02:37 hmmmm this is the same thing i'm asking about before
02:37 lucasem lol then just try 'git add <my-file>'
02:37 hmmmm lucasem:  ???
02:37 lucasem then 'git diff --staged HEAD <old-commit> <my-file>'
02:37 lucasem should be nothing in the diff
02:38 hmmmm can I stop for a second and explain what I think is going on?
02:38 kadoban lucasem: What's in 'git config --list', I guess? Maybe some weird line endings settings.
02:38 hmmmm it's like all these checkout, reset, etc. operations succeed as expected
02:38 hmmmm but then the working set is instantly overwritten
02:39 hmmmm i had core.autocrlf = false to begin with when this all started
02:39 hmmmm and then i was trying to fix it on my own so i changed autocrlf to input
02:39 hmmmm and i core.safecrlf was unset, but i set that to false
02:40 kadoban hmmmm: Is there anything in .gitattributes anywhere, or in .git/info/attributes ?
02:40 kadoban I'd have to look up what the crlf settings all do, they're pretty specific and annoying ...
02:40 hmmmm yes
02:40 sostenuto joined #git
02:40 hmmmm in .gitattributes there's * text=auito and *.bat text eol=crlf
02:40 hmmmm * text=auto rather
02:41 hmmmm there is no attributes file in .git/info however
02:41 lucasem so you are intentionally using carriage returns, yes?
02:41 hmmmm uhhh hrmm
02:42 hmmmm i think these files were originally CRLF
02:42 hmmmm i don't know what they got modified to though
02:42 kadoban I think on windows you usually have core.autocrlf=true, so that part is a bit suspicious
02:42 cqi joined #git
02:42 kadoban What *should* be the endings of this file, the one you want commited?
02:42 kadoban And what should it be in the working directory, if you care?
02:43 sangy joined #git
02:43 hmmmm i don't know
02:43 hmmmm is there a way i can tell?
02:43 hmmmm i want this to be just back to normal
02:44 hmmmm is there some kind of option for git diff to show line endings?
02:44 lucasem you could write the diff to a file, and inspect with vim
02:44 kadoban Well, you could see what's in history before this I guess. I dunno the windows commands to do that though. Usually you'd just view the file in an editor configurable enough to be able to show you I guess, or a hex editor or something in the worst case.
02:45 kadoban But I don't really understand what's up with the weird checkout behavior. Kind of confusing.
02:45 hmmmm i've seen 2 stackoverflow posts about what seems to be the same situation
02:45 hmmmm but literally nothing works
02:46 kadoban hmmmm: This problem came up before you started messing with line ending settings, right?
02:46 hmmmm how do i write the diff to a file?  just redirect stdout?
02:47 hmmmm kadoban, yes
02:47 hmmmm the whole problem started when i tried to look at the file in visual studio
02:48 kadoban hmmmm: Yeah, just direct it to a file with >   You can also try getting the contents of the file directly from history with 'git show thecommit:thefile > somewhere'
02:48 hmmmm okay
02:48 hmmmm that's probably better
02:49 hmmmm if the file has tabs in it, will that ruin the tabs?
02:49 jpX__ joined #git
02:49 hmmmm or should that not matter for what we're doing?
02:50 edude03 joined #git
02:50 kadoban Shouldn't matter.
02:50 jstimm joined #git
02:50 ekinmur joined #git
02:51 hmmmm if git diff fileA fileB returns immediately with no output that means they're identical, correct?
02:51 kadoban Correct
02:51 hmmmm .... what on earth
02:51 hmmmm well they are the same then
02:51 PioneerAxon joined #git
02:52 hmmmm and they do have the same line endings, both CRLF
02:52 optraz left #git
02:52 hmmmm weird, i knew it was windows line endings originally but i didn't want to answer that because it'd be contradictory
02:52 hmmmm now i'm sure it's contradictory
02:52 optraz joined #git
02:53 optraz how do i get the name of a last committer in the repository?
02:53 optraz is it possible?
02:53 kadoban optraz: Define "last commiter"?
02:53 kadoban hmmmm: I'm not really sure what's going on there offhand, I'd have to read a decent amount about line endings.
02:54 optraz kadoban: i commit file to the repository , i want my username, nwo dev commit again to the git, i want to show his username
02:54 jaguarmagenta joined #git
02:54 hmmmm so let me get this straight... git show HEAD~3:src/foobar.c > oldfoobar.c && git diff oldfoobar.c src/foobar.c shows no difference
02:55 hmmmm there are NO changes staged for src/foobar.c, it is unmodified at HEAD
02:55 hmmmm yet git diff --name-only HEAD~1 clearly shows it's been modified
02:55 edude03 joined #git
02:57 hexagoxel joined #git
02:57 MZAWeb_ joined #git
02:57 sagerdearia joined #git
02:58 lagothri1 joined #git
03:00 jpX joined #git
03:00 mingrammer joined #git
03:01 rgrinberg joined #git
03:01 vassagus joined #git
03:01 ee joined #git
03:02 edude03 joined #git
03:02 hahuang61 joined #git
03:05 netj joined #git
03:06 jaguarmagenta joined #git
03:07 red_ joined #git
03:11 Douhan joined #git
03:12 Literphor joined #git
03:12 GT4066 joined #git
03:13 mingrammer joined #git
03:13 notadrop joined #git
03:14 notadrop Hi. When I clone repo foobar, do I run "clone" from inside ~/git/ or ~/git/foobar/ ?
03:14 notadrop i.e. do I have to create the directory foobar manually first, and then run "clone" from inside of it?
03:17 hmmmm git clone ssh://git.foobar.com/foobar foobar is going to create a directory called foobar in your CWD
03:17 notadrop thanks, hmmmm
03:17 notadrop left #git
03:18 rewt you can also add a target directory on the end of that to specify which dir it should be cloned to.  if you want it cloned to the current dir (you're inside the one you want the root to be), you can add . on the end
03:18 hmmmm kadoban, lucasem, i didn't figure out what happened but i sorta fixed it by doing git reset HEAD~1, adding only the files i want to be included in the commit, committing it, then git rm --cached -r . && git reset --hard
03:19 edude03 joined #git
03:20 __Myst__ joined #git
03:20 k3rn31 joined #git
03:21 hmmmm only problem is now i somehow have a bunch of unstaged changes in files from a certain directory
03:21 hmmmm oh god no, they all have the same problem
03:21 hmmmm this is a curse
03:22 MrWoohoo joined #git
03:27 doebi1 joined #git
03:28 danslo joined #git
03:33 A5101 joined #git
03:35 irregular joined #git
03:37 edude03 joined #git
03:39 fairuz joined #git
03:40 mehola joined #git
03:40 MattMaker joined #git
03:42 MattMaker joined #git
03:42 fstd_ joined #git
03:42 Duder9999 joined #git
03:43 ee joined #git
03:44 A5101_ joined #git
03:47 shaggycat_ joined #git
03:48 fission6 joined #git
03:48 __Myst__ joined #git
03:49 a_thakur joined #git
03:49 fairuz joined #git
03:50 Limes_ joined #git
03:51 zacts joined #git
03:51 jpX joined #git
03:56 sagerdearia joined #git
03:57 moei joined #git
04:01 leehambley joined #git
04:06 ericsupreme joined #git
04:09 __Myst__ joined #git
04:09 jpX joined #git
04:11 ahmed_elgabri joined #git
04:14 MacWinner joined #git
04:24 anu0 joined #git
04:25 mingrammer joined #git
04:26 safe joined #git
04:30 k3rn31 joined #git
04:33 dsdeiz joined #git
04:36 d4rklit3 joined #git
04:42 justanotheruser joined #git
04:42 mingrammer joined #git
04:43 mycyber joined #git
04:43 interne7y joined #git
04:45 Cabanossi joined #git
04:47 subhojit777 joined #git
04:51 leehambley joined #git
04:55 danimal2 joined #git
04:57 freimatz joined #git
04:57 red_ joined #git
05:00 drbean joined #git
05:00 sostenuto joined #git
05:01 cdown joined #git
05:02 cdown_ joined #git
05:03 hahuang61 joined #git
05:03 Andrew_K joined #git
05:04 rscata joined #git
05:08 moneylotion joined #git
05:11 mingrammer joined #git
05:11 netj joined #git
05:15 drbean_ joined #git
05:18 LeMike joined #git
05:19 Derperperd joined #git
05:22 hahuang61 joined #git
05:22 emPi joined #git
05:25 TomyLobo joined #git
05:26 Zamerick joined #git
05:29 danslo joined #git
05:30 phanimahesh joined #git
05:33 ahmed_elgabri joined #git
05:35 xall joined #git
05:37 Ph_ joined #git
05:39 Fijit joined #git
05:43 Finnoloid joined #git
05:44 ThomasLocke joined #git
05:45 titanium17 joined #git
05:45 Alienpruts joined #git
05:46 jaguarma_ joined #git
05:48 jceb joined #git
05:49 durham joined #git
05:49 Macaveli joined #git
05:52 Khisanth joined #git
05:52 rtyu1120 joined #git
05:54 GavinMagnus joined #git
05:56 qt-x joined #git
05:57 jaguarmagenta joined #git
05:59 rominronin joined #git
05:59 zeroed joined #git
05:59 zeroed joined #git
06:00 MattMaker joined #git
06:02 nilg joined #git
06:02 cdown_ joined #git
06:03 circ-user-hroqE joined #git
06:04 a_thakur joined #git
06:05 howdoi joined #git
06:06 freimatz joined #git
06:08 jstimm joined #git
06:10 cdown_ joined #git
06:11 vassagus joined #git
06:14 moneylotion left #git
06:14 rtyu1120 left #git
06:21 a_thakur joined #git
06:27 takwas joined #git
06:32 Softeisbieger joined #git
06:35 cqi joined #git
06:36 King_Hual joined #git
06:36 crayon joined #git
06:38 nilg` joined #git
06:38 GavinMagnus left #git
06:39 lucas_ai__ joined #git
06:41 lucas_ai__ left #git
06:41 lucas_ai joined #git
06:41 lucas_ai Hello. When I do: "git diff", instead of seeing a few line changes, it looks like my whole text has been removed and re-written. Why is this?
06:42 tvw joined #git
06:44 thiago lucas_ai: windows?
06:44 interne7y joined #git
06:45 subhojit777 joined #git
06:45 star_pro_ joined #git
06:47 lucas_ai thiago, ubuntu
06:47 Alienpruts joined #git
06:48 SpeakerToMeat joined #git
06:50 jaguarmagenta joined #git
06:50 jknetl joined #git
06:50 thiago lucas_ai: the only thing I can think of that shows everything removed and everything added again is change to the line ending
06:50 thiago lucas_ai: is it possible your text editor changed the line ending in that file?
06:51 hikanndefff joined #git
06:53 hikanndefff left #git
06:53 Sceorem joined #git
06:54 brandonw joined #git
06:54 lucas_ai thiago, oh, I fixed it. Yes, that was it
06:54 star_pro_ joined #git
06:54 moriko joined #git
06:55 jknetl joined #git
06:56 gfixler1 joined #git
06:57 xall joined #git
06:57 jknetl joined #git
06:58 chachasmooth joined #git
06:58 gfixler joined #git
07:00 jknetl joined #git
07:01 gfixler2 joined #git
07:05 Corosus joined #git
07:08 red_ joined #git
07:09 subhojit777 joined #git
07:10 Alienpruts joined #git
07:11 Repox joined #git
07:12 vassagus joined #git
07:13 p4trix joined #git
07:15 abhinav joined #git
07:15 abhinav hi
07:15 abhinav i commited last few changes to wrong branch. how to bring them to correct branch?
07:18 k3rn31 joined #git
07:18 Grievre cherry-pick them to the correct branch and then revert them
07:19 Grievre Did you push already or did you only commit locally?
07:19 abhinav locally. not pushed yet.
07:19 LeMike joined #git
07:21 Grievre you can switch to the correct branch, cherry-pick them, then switch back and rewind the wrong branch to before the commits
07:21 Grievre at least I think that works
07:21 abhinav ok
07:23 ahmedelgabri joined #git
07:26 Balliad joined #git
07:30 danslo joined #git
07:31 cqi joined #git
07:32 danslo joined #git
07:32 leeN joined #git
07:34 kurkale6ka joined #git
07:41 occup4nt joined #git
07:42 pikajude_ joined #git
07:43 likbardio joined #git
07:43 mra90 joined #git
07:45 crayon joined #git
07:46 xaa joined #git
07:46 MatthewAllan93 joined #git
07:47 GreatPotato joined #git
07:47 GreatPotato Can I do “git apply update.diff” on something that isnt a git repo, even if i have git installed?
07:47 GreatPotato Im trying it but doesnt seem to work
07:47 t0by joined #git
07:48 jceb joined #git
07:48 sans_s3r1f joined #git
07:49 _ikke_ lkdjfgl
07:50 _ikke_ It does not require a git repository
07:51 GreatPotato So any idea why “git apply update.diff” just returns nothing, and the changes dont seem to apply?
07:52 cbreak GreatPotato: tried --verbose --apply?
07:52 cbreak GreatPotato: I've tried to use git apply recently, it didn't work reliably :(
07:53 cbreak patch worked though
07:53 GreatPotato I tried “git apply -v security-update.diff” which again returns sod all :/
07:53 GreatPotato I’ll try patch, ive heard its more reliable
07:54 pikajude_ left #git
07:54 RedShift joined #git
07:54 maroloccio joined #git
07:56 hmmmm yay
07:56 hmmmm i totally fixed my problem
07:57 hmmmm all i needed to do was repeat the last two steps again
07:58 djb-irc joined #git
07:58 PCatinean joined #git
07:58 _ikke_ GreatPotato: I've never had problems with git apply
08:00 neeca joined #git
08:00 Darcidride joined #git
08:01 lucas_ai_ joined #git
08:03 cYmen how do I conveniently browse through files and see who changed every file last and in which commit?
08:04 ki0 joined #git
08:05 chll_ joined #git
08:06 GreatPotato _ikke_: wish i knew what you were doing, because this is doing absolutely nothing
08:06 t0by joined #git
08:07 _ikke_ Would you be able to share the diff?
08:07 marcogmonteiro joined #git
08:07 Macaveli joined #git
08:12 abhinav while cherry picking comments from another branch to this branc, it conflicts at one commit and ask me to resolve conflict. how to resolve conflict? do i need to manually edit some file or git has some command?
08:13 Literphor joined #git
08:13 tomog999_ joined #git
08:13 zacts joined #git
08:15 texinwien_ joined #git
08:15 ycon_ joined #git
08:15 dakotahawkins joined #git
08:16 ycon_ Hi all, if I have this in my .gitignore "*__pycache__" will that ignore any folder called __pychche__?
08:16 GreatPotato _ikke_: http://pgrm.co/18151V1Q3h0J
08:16 GreatPotato Althought i think its vagrant causing problems - seems like its trying to apply it to the vagrant repo possibly...
08:17 canton7 ycon_, that will ignore any file or folder *ending* with __pycache__
08:17 interne7y joined #git
08:17 ycon_ canton7: so I should remove the *?
08:17 GavinMagnus joined #git
08:17 canton7 ycon_, if you want it to match "__pycache__" but not "foo__pycache__", yes
08:17 citizen-stig joined #git
08:18 ycon_ ok cool
08:18 vassagus joined #git
08:21 lucas_ai_ joined #git
08:21 eijk joined #git
08:22 abhinav i deleted last 9 commits in my local branch. how to sync remote branch with this hcange?
08:22 clmsy joined #git
08:24 lucas_ai__ joined #git
08:25 jceb joined #git
08:27 dakotahawkins I have a question which may (I'm skeptical) be a bug, but here goes:
08:27 dakotahawkins In my global .gitconfig, I have "user.useConfigOnly = true" and user.email isn't set there (I set it on a per-repo basis, as I use different e-mails for work and personal repos). I ALSO have "pull.rebase = preserve" set.
08:27 dakotahawkins An example of the problem I have is with tools like golang (I filed an issue there: https://github.com/golang/go/issues/16516#issuecomment-235800085) that use git to pull in package repos without any real user interaction.
08:28 dakotahawkins When something like that runs a git pull for me (to update a package repo) my global config makes it try to rebase, which fails because git doesn't know who I am (no e-mail set globally, remember?)
08:28 dakotahawkins I'd like to know if anybody can tell me why a "git pull --rebase" needs to get my author info from my config, when presumably it could just grab that info from the pre-rebase commits.
08:29 canton7 .man_update
08:29 gitinfo canton7: Manpage index updating. Please allow a few seconds before using again.
08:32 canton7 dakotahawkins, each commit has not one, but two people associated with it. One is the author: that's the person who wrote the patch originally. The other is the committer: that's the person who committed the patch (git was designed around sending patches by email). When you rewrite a commit (rebase, --amend, cherry-pick, what have you) you then become the committer, but the author is preserved
08:33 dakotahawkins That's kind of what I thought, but I tried digging through the code to see what specifically was triggering it. It was hard to follow without a lot of experience, and looked a little like it wasn't able to find the author, but I may have just followed the trace to the wrong location.
08:34 flavius joined #git
08:35 dakotahawkins In these cases, nothing generally gets re-written (since they're package repos not really designed for working on), but I guess "pull --rebase" doesn't detect it can fast-forward instead and do that. Do you happen to know why that is?
08:35 Oebele joined #git
08:35 canton7 'fraid I'm not familiar with that bit of code
08:36 canton7 the mailing list might be a better place to ask a question which pertains to git's internals
08:36 dakotahawkins OK, I didn't want to go there first :)
08:36 dakotahawkins Thanks for the help!
08:38 _ikke_ canton7: Ah, didn't know that command
08:38 _ikke_ .version
08:38 Topic for #git is now Welcome to #git, the place for git help and nursery rhymes | Public logs at http://goo.gl/BuUi5o | Current stable version: 2.9.2 | First visit? Read: http://jk.gs/git | Getting "cannot send to channel"? /msg gitinfo .voice | Ring around the repo, pockets full of plumbing, objects, objects, we all commit.
08:39 canton7 _ikke_, it's not documented: I have to go digging through the source each time (https://github.com/jast/gitinfo/blob/master/plugins/man.pm#L10)
08:39 lucas_ai joined #git
08:41 dopesong joined #git
08:42 lucas_ai_ joined #git
08:45 _ikke_
08:46 yuhlw_ joined #git
08:47 rolha joined #git
08:48 crayon joined #git
08:48 mra90 hmm I have noticed that rebasing performs hard merge, I mean just don't notify about conflict in files and apply changes from rebase branch anyway
08:48 garethdaine joined #git
08:49 _ikke_ That's not default behaviour
08:49 _ikke_ rebase will drop to console and tell you it encountered a conflict
08:49 canton7 no: that's not the correct behaviour
08:49 sostenuto joined #git
08:49 _ikke_ Unless you specified an option or config that overrides this
08:53 mra90 hmm it did it without any notification
08:54 TomyWork joined #git
08:55 yuhlw_ joined #git
08:57 _ikke_ Are you sure it was a conflict then?
08:57 chid joined #git
08:58 chid_ joined #git
09:00 nnyk joined #git
09:00 mra90 _ikke_: sure, no doubt baout it
09:01 wrouesnel1 joined #git
09:01 mra90 after rebase it just apllied rebasing branch commit
09:02 mra90 btw, how to remove all commit beside last two ?
09:02 mra90 commits*
09:05 mehola joined #git
09:06 _ikke_ From the entire repository?
09:07 mra90 _ikke_: yes or deletae are commit before certain one
09:07 mra90 they are simply no longer needed
09:07 _ikke_ combination of git replace + git filter-branch
09:10 mra90 _ikke_: no deletions ?
09:10 sostenuto joined #git
09:12 a_thakur joined #git
09:13 fairuz joined #git
09:14 _ikke_ git replace --graft <last_commit>
09:14 _ikke_ This wil temporarily (for your local repository) set the parent to nothing (making it a root commit)
09:18 delboy1978uk joined #git
09:19 dsdeiz joined #git
09:19 mahendrak joined #git
09:20 delboy1978uk hi guys, someone has rejected a ticket i worked on because they said there isn't a branch found! yet i can check it out from my working folder! when i git push it says everything up to date, but my colleague did a git fetch and git branch, and sure enough my branch isnt listed! what has happened and how can i sort it?
09:21 mra90 fatal: replace ref 'refs/replace/ed6f926b7340efe85e6c202c44a0c34158ec0163'
09:21 mra90 _ikke_: ^
09:21 canton7 delboy1978uk, if you've got an older git version, 'git push' will (by default) push all branches for which a remote branch already exists, but won't push any "new" branches
09:22 canton7 (this behaviour was changed a while back, but maybe you've got an old git version or non-default config?)
09:22 canton7 'git push remotename branchname' anyway, or 'git push remotename HEAD' as a shortcut if you've got branchname checked out
09:22 _ikke_ mra90: what command did you execute?
09:22 canton7 add '-u' in there to set up tracking at the same tmie
09:22 david007 joined #git
09:22 vassagus joined #git
09:22 david002 joined #git
09:22 mra90 git replace --graft ed6f926b7340efe85e6c202c44a0c34158ec0
09:23 delboy1978uk canton7: i did  git push origin S5353, but it just says everything up to date
09:23 canton7 delboy1978uk,  is the branch called 'S5353'?
09:23 delboy1978uk yes
09:23 canton7 delboy1978uk, what does 'git branch -a' show?
09:23 mra90 I thought removind series of commit would be easy
09:24 _ikke_ mra90: No, because everything is tied together
09:24 canton7 delboy1978uk, do 'git ls-remote origin' while you're at it
09:24 mra90 _ikke_: so how to do this ?
09:24 delboy1978uk canton7: it shows a big list, some green, some red. remotes/origin/S5353 is one of the ones in red
09:25 jknetl joined #git
09:25 mra90 I think it's simple - just remove whole repo and do git init
09:26 canton7 delboy1978uk, ok, so that branch exists on the remote repo
09:26 xall joined #git
09:26 delboy1978uk canton7: ls-remote gives this 5383b2016f0b807f512a0bca88d09e5f647a7584        refs/heads/S5353
09:26 canton7 it definitely definitely exists in the remote repo then
09:26 delboy1978uk why were some of the branches listed red?
09:28 _ikke_ delboy1978uk: remote tracking branches
09:28 phanimahesh joined #git
09:29 mra90 _ikke_: solved by what I said
09:29 mra90 just rebase to some commit and remove repo
09:29 mra90 then git init
09:29 mra90 problem solved ;)
09:35 delboy1978uk canton7: thanks for your help. I updated my ticket slapping the guy into place :P
09:35 canton7 delboy1978uk, I was going to suggest, maybe you're looking at a different repository to the one you're pushing to?
09:36 delboy1978uk hmm you think?
09:36 delboy1978uk that would be in .git/config, right?
09:37 canton7 git remote -v
09:37 jceb joined #git
09:37 canton7 (or yes, look in the config)
09:37 geoid joined #git
09:38 gtristan joined #git
09:38 gtristan Hi
09:39 gtristan What is the command again to set the remote url for origin ?
09:39 delboy1978uk canton7: i just looked on the GitLab site too, it is most certainly there
09:39 canton7 man git remote set-url
09:39 gitinfo the git-remote manpage is available at http://jk.gs/git-remote.html
09:39 canton7 delboy1978uk, ok, cool
09:39 gtristan canton7, thanks !
09:39 delboy1978uk canton7: so i'm cool and the other guy isn't.
09:39 delboy1978uk canton7: Thanks for all your help!
09:41 canton7 no worries
09:44 mingrammer joined #git
09:44 case39 joined #git
09:46 PCatinean joined #git
09:47 sostenuto joined #git
09:49 jubalh joined #git
09:49 jubalh i am merging a branch into master via git merge origin/develop (being on my local master) it seems to force me to write a commit stating "merge branch X" is that necessary?
09:49 grawity it's recommended, but can be turned off
09:50 delboy1978uk keep it that way
09:50 delboy1978uk left #git
09:50 jubalh what benefit will it bring me?
09:50 kutsuya joined #git
09:52 grawity a high-level description of what feature got merged (if it's spread out across a dozen of commits)
09:52 grawity generally kinda depends on how the project is structured
09:53 grawity it's more useful for the "feature branches" style than "merge from dev whenever it's green" style
09:54 jubalh i see
09:55 duderonomy joined #git
09:55 jknetl joined #git
09:58 Klumben joined #git
09:58 duderonomy joined #git
10:01 mra90 does anybody here use smartgit ?
10:01 Lomex joined #git
10:03 mflow joined #git
10:04 murph joined #git
10:05 chid joined #git
10:05 chid_ joined #git
10:08 cdown_ joined #git
10:09 SwiftMatt joined #git
10:10 duderonomy joined #git
10:16 quackgyver joined #git
10:19 acetakwas joined #git
10:21 chid joined #git
10:21 chid_ joined #git
10:24 rom1504 doing stats on git client usage in #git ?
10:25 harish_ joined #git
10:26 jimi_ joined #git
10:27 lucas_ai__ joined #git
10:28 vassagus joined #git
10:29 playfullyExist joined #git
10:30 mra90 should I always do pull from server before commit there ?
10:30 anu0 joined #git
10:34 neeca joined #git
10:35 SwiftMatt joined #git
10:35 garethdaine joined #git
10:36 rom1504 it's probably better yes
10:37 cdown_ joined #git
10:37 rom1504 it's not required though, you can always merge/rebase afterwards
10:39 achadwick joined #git
10:43 lucas_ai joined #git
10:46 mra90 rom1504: well rebase on remote server isn't good idea
10:46 rom1504 yes, that's unrelated to your question
10:47 MZAWeb_ joined #git
10:48 nettoweb joined #git
10:48 david007 joined #git
10:48 sostenuto joined #git
10:49 rnsanchez joined #git
10:51 Derperperd joined #git
10:55 tvw joined #git
10:57 mra90 when we delete branch all its commits are also removed ?
10:57 texinwien_ joined #git
10:58 _ikke_ Not immediately
10:59 mra90 no? What do you mena ?
10:59 scoopr joined #git
11:00 mra90 mean*
11:00 _ikke_ The reflog keeps a reference for 90 days. After that, the gc will remove the commits
11:01 mra90 but it will happen automaticly right ?
11:01 scoopr Hi! I'm doing a clone of a repo of mine.. at the final step the clone errors out with: fatal: did not receive expected object 1889a...
11:01 scoopr cat-file -t shows its a commit, branch --contains doesn't give any branches
11:01 shinnya joined #git
11:01 scoopr but pruning and gc:ing doesn't get rid of it
11:01 _ikke_ mra90: yes, not something you should have to worry about
11:02 mra90 ok
11:03 paws_ joined #git
11:04 Guest36751 joined #git
11:04 chrisml joined #git
11:05 Lomex joined #git
11:05 Alienpruts joined #git
11:05 Guest36751 left #git
11:05 sdothum joined #git
11:06 ffs joined #git
11:06 ffs how do i tell git to screw all other commits and just jump ahead to what's currently in my directory?
11:06 danslo joined #git
11:07 _ikke_ What do you mean, jump ahead to what's currently in your directory/
11:07 ffs i'm not wasting my time merging what my coworkers wrote. just replace everything they have with what's in my directory and make that the latest commit.
11:08 bluezone joined #git
11:08 _ikke_ Is what you have in your directory already comitted?
11:08 chachasmooth joined #git
11:08 iota_pi joined #git
11:09 ffs there are some merge conflicts git wants me to take care of
11:09 ffs i want to say "screw merging, just use what i already have"
11:10 interne7y joined #git
11:10 _ikke_ So you're currently in a merge conflict?
11:10 ffs yeah
11:10 mra90 hmm what's the difference between merge and patch ?
11:10 rnsanchez joined #git
11:10 fuchstronaut joined #git
11:10 fuchstronaut joined #git
11:11 _ikke_ Merge merges two divergent branches together
11:11 _ikke_ patch is a diff that you can apply to files
11:12 freimatz joined #git
11:12 ffs i'm about to just reclone the repo and then copy my files over theirs
11:14 _ikke_ no need to
11:14 _ikke_ git checkout --ours . should help
11:14 _ikke_ (note the dot)
11:15 cbreak joined #git
11:16 GavinMagnus joined #git
11:17 ojdo joined #git
11:18 ffs is there a good irc channel for venting
11:18 HoierM joined #git
11:19 _ikke_ I don't think so
11:20 HoierM joined #git
11:20 mra90 so, to create patch file for i.e last two commit I will have to do git diff commit_id_1 commit_id_2 -p ?
11:21 _ikke_ -p is not necessary for diff
11:21 _ikke_ git diff HEAD~2 >name_of_patch.patch
11:22 HoierM joined #git
11:22 _ikke_ What is your intention for this patch?
11:23 _ikke_ (there is also git format-patch, if it's mean to be applied in some other repository through mailing-list for example)
11:23 Pawky odd behaviour, having a branch test, checkout -b test-ini, delete all but ini files, checkout test, all but ini files are gone, checkout master, all files there, checkout test, all files there, checkout test-ini only ini files... duh?
11:24 Pawky how come when moving from one branch with lesser files, to the older with more files I do not see them, unless moving to master, and then back?
11:25 Pawky Oh dear... never mind....   test != testing... :-D
11:25 HoierM joined #git
11:25 _ikke_ alright
11:25 * Pawky secretly sneaking away
11:27 HoierM joined #git
11:28 doebi joined #git
11:28 tomreyn joined #git
11:28 skunkz joined #git
11:29 zeroed joined #git
11:30 phanimahesh joined #git
11:30 HoierM joined #git
11:30 nilg joined #git
11:33 vassagus joined #git
11:35 cdown_ joined #git
11:38 Macaveli joined #git
11:39 mra90 _ikke_: I do this just for learning purposes
11:39 rolcam joined #git
11:39 jimi_ joined #git
11:39 mra90 and I think now I know enough to start working with git
11:40 GreatPotato left #git
11:40 _ikke_ alright
11:41 bolovanos joined #git
11:42 mra90 _ikke_: and the purpose of patch instead of merge is to give somebody else ability to apply changes without tinkering into server repository right ?
11:43 ki0 joined #git
11:43 phanimahesh joined #git
11:43 _ikke_ mra90: patch and merge two completely different things with different purposes
11:44 mra90 _ikke_: patch to allow somebody apply specyfic changes withut pull fro mmaster server ?
11:45 dminca joined #git
11:45 drbean joined #git
11:45 dminca joined #git
11:46 _ikke_ mra90: the linux project and git itself use a patch based workflow, where patches are sent to a mailing-list for review
11:47 mra90 _ikke_: I guess they also use commits :P but was my description of patching correct ?
11:47 BryzNSTY joined #git
11:47 rafalcpp joined #git
11:47 _ikke_ mra90: To be clear, you have two types of patches
11:48 _ikke_ just simple diffs, which only contains file changes, which are applied to your working dir
11:48 _ikke_ (eg, the output of git diff)
11:49 _ikke_ typical use would be source codde patches before compiling
11:49 freddieptf joined #git
11:49 mra90 ok
11:50 _ikke_ The second type is what git format-path would generate
11:50 _ikke_ which are commits serialized to a mail format
11:50 _ikke_ git format-patch&
11:51 _ikke_ These contain also the commit message and author information, which, when applied, would directly result into commits
11:52 Mir4g3 joined #git
11:52 __Myst__ joined #git
11:54 _ikke_ So those get sent to a mailing list, where they can be reviewed, and applied
11:55 fuchstronaut joined #git
11:55 Sceorem joined #git
11:55 mra90 hmm my description was correct
11:55 Macaveli joined #git
11:55 _ikke_ Just wanted to make sure you understood the difference between two different type of patches
11:55 mra90 instead of merge changes with remote server you send patch file so the developer can apply changes on his own
11:56 mra90 this would be also the way to coperate without centralized server, right ?
11:56 BlaXpirit !orphan
11:56 gitinfo To create an orphaned branch(unrelated to existing history): `git checkout --orphan newbranchname`. This will leave your index/worktree as-is(use `rm .git/index; git clean -dfx` to delete EVERYTHING). If you get 'error: unknown option `orphan`' see !orphan_old. For an empty/null commit see !orphan_null.
11:56 _ikke_ mra90: The commits that are sent as patches will still be merged
11:56 _ikke_ so it's not an or / or
11:56 _ikke_ mra90: but yes, it does not require a central server
11:57 mra90 I wonder where this solution - patching isn't error prone
11:58 liuzhen joined #git
12:00 RustyShackleford joined #git
12:01 _ikke_ solution for hwat?
12:01 _ikke_ what?
12:04 griffindy joined #git
12:06 mra90 _ikke_: maybe solution wasn't good word there I was rather wondering wheter patching is always safe ;)
12:07 _ikke_ safe in what sense?
12:07 alt513 joined #git
12:08 eijk joined #git
12:09 circ-user-hroqE joined #git
12:10 mra90 file corruption things like thata
12:10 mra90 that
12:11 mra90 now I have a problem at some point I did patch file then few commits and now I try to restore by using this patch file with git apply patchfile.patch
12:11 mra90 and I am getting error |....patch does not apply"
12:12 _ikke_ That's the patch version of a merge conflict
12:13 tazle_ joined #git
12:13 Mikaela joined #git
12:14 courrier joined #git
12:14 MZAWeb_ joined #git
12:15 ShekharReddy joined #git
12:17 mra90 _ikke_: ee could you please say something more ?
12:18 _ikke_ A patch requires the files to be in the state that the patch was created
12:18 _ikke_ basically a two-way merge
12:19 ki0 joined #git
12:21 mra90 so without branches
12:22 rgrinberg joined #git
12:22 mingrammer joined #git
12:23 _ikke_ each patch-series would be its own branch
12:23 _ikke_ applied to the original commit they ware made of
12:25 rgrinberg joined #git
12:25 david007 joined #git
12:25 mra90 so I cannot apply patch from one commit to another ?
12:26 _ikke_ You can, as long as they don't conflict
12:26 mra90 hmm I cannot grasp where do you see this conflict
12:27 mra90 I am testing this feature on just on file
12:27 mra90 I did patch several commits ago and now I try to apply it
12:27 mra90 I don't see any potential conflict
12:27 jimi_ joined #git
12:27 mra90 the error message is also very laconic
12:28 _ikke_ It requires the context (the lines around the diffs) to be the same
12:30 fooobarrr joined #git
12:32 Mikaela joined #git
12:32 dopesong joined #git
12:33 mra90 _ikke_: lines around diff?
12:33 _ikke_ yes
12:33 _ikke_ the lines around the + and - lines
12:34 sanketdg joined #git
12:36 sdaugherty right, diff provides those as "context" so that patch can verify it's changing the same, or at least, very close to the same file.
12:37 sdaugherty they also serve as readability aids for someone just looking at a diff and trying to figure out what it's doing
12:37 vassagus joined #git
12:38 mra90 now I did patch from two last commits haven't commited/changed anything and patching failed also same error as before
12:39 sdaugherty try to get a new diff if you can, or manually apply the page
12:39 sdaugherty patch
12:40 phanimahesh How do I go to a commit approximately a month ago? I remember using a flag to git checkout or similar, can't recollect.
12:40 phanimahesh In a single command, I don't want to check the log, get the id and use it.
12:41 mra90 sdaugherty: manualy, how ?
12:41 sdaugherty mra90, by reading the diff - diffs are designed to be human readable.
12:42 jeffreylevesque joined #git
12:42 mra90 so you want me to manualy copy paste content ?
12:42 sdaugherty as a last resort when a diff doesn't apply, you can do that.
12:42 sdaugherty if you don't have any other way
12:43 phanimahesh nvm, found it. git checkout branch@{time} does it
12:43 sdaugherty It's basically copying and pasting, but you actually want to look at what you are doing as you do it, obviously,
12:44 rbern joined #git
12:44 __Myst__ left #git
12:44 sdaugherty A diff is simply a text representation of one or more changes (replacement of one text with another)
12:44 jstein_ joined #git
12:45 tazle_ joined #git
12:45 Mikaela joined #git
12:45 sdaugherty They can be interpreted by a person almost as easily as by a machine.
12:47 mra90 sdaugherty: the problme is machine fails to apply those changes
12:47 ahmedelgabri joined #git
12:47 mra90 I mean git apply fails
12:48 rbern joined #git
12:48 fmcgeough joined #git
12:49 sdaugherty mra90, right, but there's nothing stopping you from applying them by hand if get apply fails, and the exercise of doing so will likely reveal to you WHY it's failing.
12:51 cdown_ joined #git
12:51 rik316 joined #git
12:52 dopesong joined #git
12:52 mmlb joined #git
12:53 dopesong_ joined #git
12:54 dopesong_ joined #git
12:56 CheckDavid joined #git
12:56 Furai Hey, is there anything in git that could limit the cloning/fetching speed? For some reason my co-workers get limited to around 100Kb/s where for me it's going with full speed.
12:56 Furai I'm suspecting some routing issue.
12:57 Furai Something along the way.
12:57 Furai I'm just living around 200km north from my colleagues and everything works fine for me.
12:59 _ikke_ iirc nothing in git
13:01 crayon joined #git
13:01 afusion joined #git
13:02 sdaugherty Furai, I'd look at network throughput and latency outside of git
13:02 fahadash joined #git
13:02 sdaugherty at least, before you look for issues in git itself - those are much less likely
13:02 Darren_ joined #git
13:02 SourTrout joined #git
13:03 kyan joined #git
13:03 sdaugherty Furai, also remember git uses very efficient delta compression for push/pull/fetch, but initial clones don't get much benefit
13:04 mmlb joined #git
13:09 SirCmpwn joined #git
13:10 tonyyarusso joined #git
13:11 l4v2 joined #git
13:11 attente joined #git
13:11 Furai Yeah, I'm aware of that. And I've been looking at our server settings and network but there's literally nothing.
13:11 i7c So
13:11 i7c wann mach ich heute Feierabend?
13:12 i7c sorry wrong channel :D
13:12 Furai It has to be fault of their ISPs and some link being slow.
13:12 SourTrout joined #git
13:12 rwb joined #git
13:14 mingrammer joined #git
13:15 garethdaine joined #git
13:15 mmlb joined #git
13:15 dopesong joined #git
13:19 ColdKeyboard joined #git
13:20 duarteper joined #git
13:20 ColdKeyboard joined #git
13:24 neca joined #git
13:24 Rodya_ joined #git
13:26 abhinav i want to change branch without comming changes, but i need those changes so that i can add those to changed branch
13:26 abhinav commiting*
13:27 abhinav how can i change?
13:27 tango_ abhinav: git checkout otherbranch
13:27 abhinav tango_: it says there are changes not commited.
13:28 xiaokang joined #git
13:28 tango_ abhinav: that happens when the uncommitted changes conflict with the state of the tree in the other branch
13:28 tango_ abhinav: a solution for this is to use git stash
13:28 Stummi abhinav, stash is your friend
13:28 abhinav tango_: ok
13:28 Anticom joined #git
13:29 Anticom Hi all. Is there a way to do "git subtree pull" in a dirty wc?
13:30 Anticom i made some changes in my repo which require some new files in that subtree repo and i want to have those changes in a single commit
13:30 edude03 joined #git
13:30 jaafar joined #git
13:33 deadnull joined #git
13:33 nydel07 joined #git
13:33 AaronMT joined #git
13:36 grawity stash, pull, unstash, commit --amend
13:43 sfeinste joined #git
13:43 vassagus joined #git
13:44 edude03 joined #git
13:44 ExoUNX joined #git
13:44 byteflame joined #git
13:45 kedare joined #git
13:46 ramsub07 joined #git
13:48 edude03 joined #git
13:48 phanimahesh joined #git
13:49 leehambley joined #git
13:52 phaleth joined #git
13:53 _ikke_ Anticom: funny, wc means toilet in dutch :P
13:53 Anticom _ikke_: in german too. Afaik even in eng it's for water closet
13:54 _ikke_ yeah
13:54 Anticom but as a dev i'm lazy anyway... so i don't care :D
13:54 _ikke_ But it's hardly used in english
13:54 edude03 joined #git
13:54 Anticom hm strange... although there's an update in that subtree repo `git subtree pull...` says "Already up-to-date."
13:54 Anticom o.O
13:55 osse But what's wc here anyway? Working... Catalog?
13:55 _ikke_ working copy
13:55 osse Oooh
13:55 Anticom lol
13:55 Anticom thought it was obvious
13:56 osse It is in hindsight as always.
13:56 _ikke_ working tree is more common to use in git
13:56 Gsham joined #git
13:56 Anticom ah okay
13:56 Anticom so i'll use wt next time
13:56 osse I know the term working copy, just never seen it abbreviated.
13:57 osse But I got it from context due to "dirty" and "pull" :P
13:57 str joined #git
13:57 _ikke_ wt is also going to confuse people
13:57 Anticom However.. still having an issue with 'git subtree pull'
13:57 _ikke_ Anticom: I have not used subtrees yet, so no clue
13:58 Anticom is there any meta-data stored in .git/ ?
13:58 Anticom oh wait, maybe i'm just stupid
13:59 edude03 joined #git
14:00 Anticom yep, i am
14:02 _ikke_ Almost all meta data is stored in .git
14:03 fstd joined #git
14:03 ramsub07 joined #git
14:07 Anticom _ikke_: Well my question was more like whether git subtree stores some meta-data *at all* ;)
14:08 _ikke_ Oh, no clue
14:08 plasticle joined #git
14:08 _sjs joined #git
14:10 edude03 joined #git
14:10 Sceorem joined #git
14:11 Rodya_ joined #git
14:11 interne7y joined #git
14:13 psprint Can it be said that after git fetch, FETCH_HEAD will contain only one entry marked as for merge (i.e. not with "not-for-merge") ?
14:13 Chandak01 joined #git
14:14 kaldoran joined #git
14:14 fstd joined #git
14:16 _ikke_ It would make sense, but I can't say whether it's guaranteed
14:16 danslo_ joined #git
14:16 ColdKeyboard joined #git
14:17 ash_workz joined #git
14:17 osse man git-fetch says that all fetched refs are written to FETCH_HEAD
14:17 gitinfo the git-fetch manpage is available at http://jk.gs/git-fetch.html
14:18 axl_ joined #git
14:18 _ikke_ but what format?
14:19 Zoxc joined #git
14:19 _ikke_ ie, will only one ref not have not-for-merge mentioned?
14:20 edude03 joined #git
14:20 livingstn joined #git
14:21 danslo joined #git
14:23 Rodya_ joined #git
14:24 _ikke_ apparently that's fairly new
14:24 _ikke_ 2.5.0
14:26 rafalcpp joined #git
14:26 edude03 joined #git
14:27 kutsuya joined #git
14:29 edude03 joined #git
14:32 kpease joined #git
14:32 pcre joined #git
14:33 ffs joined #git
14:33 theskillwithin joined #git
14:34 al-damiri joined #git
14:36 edude03 joined #git
14:36 kpease joined #git
14:36 psprint I did the following: had second branch "Branch2", backuped whole repository, added commit to Branch2 and to master, pushed the two commits, removed local repo, restored it from backup
14:37 tazle_ joined #git
14:37 psprint then issued git fetch. And the result is: single "not-for-merge" entry for Branch2, single `merge` entry for master
14:37 Mikaela joined #git
14:37 psprint this means: there will be no fast forward or other merge for Branch2, which seems overall pointless, I wanted to update two branches
14:37 _ikke_ "git merge FETCH_HEAD" learned that the previous "git fetch" could be to create an Octopus merge, i.e. recording multiple branches that are not marked as "not-for-merge"; this allows us to lose an old style invocation "git merge <msg> HEAD $commits..." in the implementation of "git pull" script; the old style syntax can now be deprecated (but not removed yet).
14:39 sfvkro joined #git
14:40 ikmaak joined #git
14:42 edude03 joined #git
14:42 Mikaela joined #git
14:43 ResidentBiscuit joined #git
14:43 psprint git pull does the same, fetches and merges current checkouted branch, only fetches other branches, however updating tip of remotes/origin/Branch2, so at least the fetched objects for Branch2 aren't in-moment-dangling (after another git fetch with e.g. singular refspec)
14:44 _ikke_ psprint: yes, pull will only merge one branch at the most
14:44 _ikke_ because it can only merge the currently checked out branch
14:45 zacts joined #git
14:45 catag87 joined #git
14:46 catag87 joined #git
14:46 edude03 joined #git
14:46 psprint just checked what happens if I order pull to get not checked branch – it will update origin/Branch2, also do Branch2 -> FETCH_HEAD, and then try to merge FETCH_HEAD with master, which is pretty superficial logic, resulting in conflicts in my case
14:48 vassagus joined #git
14:48 vaibhavsagar joined #git
14:48 p4trix joined #git
14:50 dopesong joined #git
14:51 psprint good that in my libgit2 software I fetch only single branch, will be able to manage and e.g. inform user that no merge will be done because the branch fetched isn't checked out, and then present git log <branch>..FETCH_HEAD nicely explaining what's new and ready to merge after required checkout
14:52 iLembus joined #git
14:52 itissid joined #git
14:53 edude03 joined #git
14:53 altendky joined #git
14:54 fuchstronaut joined #git
14:55 ekinmur joined #git
14:56 nydel07 joined #git
14:57 jimi_ joined #git
14:57 m0nkmaster joined #git
14:59 fstd joined #git
15:00 psprint weird that even git pull origin Branch2:Branch2 produces merge conflicts the same as when merging Branch2 with master, even though there is no "Branch2 -> FETCH_HEAD" in pull's output, however there is the same "master" oriented FETCH_HEAD `for-merge` entry, so git pull's output lies, there is Branch2 -> FETCH_HEAD operation
15:01 vaibhavsagar Why does a git packfile use 3 different variable-length integer encodings?
15:01 darkbit joined #git
15:03 InfoTest joined #git
15:04 SwiftMatt joined #git
15:04 diegoviola joined #git
15:04 ahmedelgabri joined #git
15:05 harish_ joined #git
15:05 _ikke_ psprint: why is that strange?
15:05 shinnya joined #git
15:05 _ikke_ THe same branch gets merged
15:05 _ikke_ vaibhavsagar: no clue
15:05 _ikke_ vaibhavsagar: the !mailing_list is a better place to ask
15:05 gitinfo vaibhavsagar: The mailing list can be reached via git@vger.kernel.org. You don't need to subscribe to the list, you will always be put in cc on reply. Read archives at http://j.mp/gitlist
15:07 anth0ny joined #git
15:09 ki0_ joined #git
15:10 madewokherd joined #git
15:10 mingrammer joined #git
15:11 edude03 joined #git
15:11 vaibhavsagar _ikke_: thanks :)
15:11 tazle_ joined #git
15:12 Mikaela joined #git
15:13 kbs vaibhavsagar: not super familiar with the packfile format but curious - in what contexts are you seeing them? I assumed some of the are simply an artifact of wanting to reduce the packfile sie
15:14 unbalancedparen joined #git
15:15 vaibhavsagar I've been teaching myself haskell by implementing a parser for the packfile format
15:15 vaibhavsagar there are some good blog posts/articles out there
15:15 cyan__ joined #git
15:16 nickthetait joined #git
15:16 vaibhavsagar http://stefan.saasen.me/articles/git-clone-in-haskell-from-the-bottom-up/#pack_file_objects
15:16 vaibhavsagar https://codewords.recurse.com/issues/three/unpacking-git-packfiles
15:16 vaibhavsagar https://github.com/vincenthz/hs-git/blob/44cee79bb701bd71deb40d39774c117c025e8829/Data/Git/Delta.hs
15:17 BlessJah joined #git
15:17 vaibhavsagar and of course https://github.com/git/git/blob/master/Documentation/technical/pack-format.txt
15:17 _ikke_ vaibhavsagar: nice
15:17 vaibhavsagar the size of a packed object (after inflation) is stored as a little-endian var-length int
15:18 vaibhavsagar as per https://github.com/git/git/blob/master/Documentation/technical/pack-format.txt#L98
15:19 vaibhavsagar the offset of an OFS_DELTA is stored as a big-endian var-length int with some extra stuff as per https://github.com/git/git/blob/master/Documentation/technical/pack-format.txt#L113
15:19 itissid joined #git
15:20 thiago joined #git
15:21 vaibhavsagar and delta copy instructions are stored in a compressed format with the bits of one byte 0..3 determining how many bytes to read to construct an offset and the bits 4..7 determining how many bytes to read to construct a size
15:21 sanketdg joined #git
15:21 itissid Testing ..
15:21 vaibhavsagar I'm not explaining it very well, I figured it out with help from http://stefan.saasen.me/articles/git-clone-in-haskell-from-the-bottom-up/#delta_encoding
15:22 vaibhavsagar so that's 3 different variable-length integer encodings
15:22 kbs cool, thanks for pointers :-) I was assuming that the latter two oddball representations are probably because the values are shorter to represent in those contexts
15:22 vaibhavsagar the last is specific to deltas so it makes sense that it's different from the others
15:23 vaibhavsagar but I fail to see the point of both the big-endian and the little-endian encodings :)
15:23 vaibhavsagar imho only the little-endian encoding should have been used
15:23 itissid Ok so my question is that if you merge from branch a to b and immediately merged it back to a. Can there ever be a merge conflict in the latter operation. Assume nothing changed after the first merge..
15:23 vaibhavsagar thanks for listening #git :)
15:24 kbs vaibhavsagar: interesting :) be curious about what you find out
15:24 Rodya_ joined #git
15:24 vaibhavsagar gmane is currently down but I will report back for sure
15:24 IWishIKnew joined #git
15:24 _ikke_ itissid: the latter merge should be a fast-forward
15:24 burtons joined #git
15:24 _ikke_ itissid: That can never result into conflicts
15:30 Exia7Blade joined #git
15:31 crayon joined #git
15:31 JanC joined #git
15:31 EduardoMartins joined #git
15:31 JJjack joined #git
15:33 eivarv joined #git
15:33 danslo joined #git
15:34 JanC joined #git
15:35 crayon joined #git
15:35 doebi joined #git
15:36 mohabaks joined #git
15:39 xaa joined #git
15:40 sangy joined #git
15:41 mingrammer joined #git
15:42 rgrinberg joined #git
15:45 cdg joined #git
15:48 crayon joined #git
15:49 dreiss joined #git
15:51 daivyk joined #git
15:52 nettoweb joined #git
15:52 vassagus joined #git
15:53 xaa joined #git
15:54 crayon joined #git
15:55 burtons joined #git
15:57 GavinMagnus joined #git
15:59 GavinMagnus joined #git
16:01 fstd joined #git
16:04 GavinMagnus left #git
16:05 hahuang65 joined #git
16:08 harish_ joined #git
16:10 phanimahesh joined #git
16:11 usm4n joined #git
16:12 Rodya_ joined #git
16:12 lindenle joined #git
16:13 Mikaela joined #git
16:13 eivarv joined #git
16:14 freimatz joined #git
16:14 subhojit777 joined #git
16:14 davisonio joined #git
16:16 burtons joined #git
16:18 BSaboia joined #git
16:22 LionsMane joined #git
16:23 sondr3 joined #git
16:25 ki0 joined #git
16:25 fuchstronaut joined #git
16:25 adino joined #git
16:25 bolovanos joined #git
16:27 anuxivm joined #git
16:29 ffs joined #git
16:30 nnyk joined #git
16:30 adino joined #git
16:34 burtons joined #git
16:34 mohabaks joined #git
16:34 Limes_ joined #git
16:35 IWishIKnew joined #git
16:35 leehambley joined #git
16:35 pcre joined #git
16:36 Zamerick joined #git
16:37 jumpman joined #git
16:41 le_melomane joined #git
16:42 WayToDoor joined #git
16:44 Pursuit joined #git
16:45 WayToDoor joined #git
16:45 doebi joined #git
16:45 burtons joined #git
16:45 tcurdt left #git
16:49 durham joined #git
16:49 semantipc joined #git
16:50 takwas joined #git
16:50 EvilPenguin joined #git
16:51 daynaskully joined #git
16:52 hahuang65 joined #git
16:54 sfeinste joined #git
16:55 jost joined #git
16:56 fuchstronaut joined #git
16:57 orbitphreak joined #git
16:58 vassagus joined #git
16:59 Fissure joined #git
17:01 liuzhen joined #git
17:01 sondr3_ joined #git
17:01 crayon joined #git
17:01 crayon joined #git
17:02 LeBlaaanc joined #git
17:06 crayon joined #git
17:07 TomyLobo joined #git
17:07 thiago joined #git
17:08 shgysk8zer0 joined #git
17:08 Sasazuka joined #git
17:09 tanuki_ joined #git
17:10 Sasazuka joined #git
17:12 Harwood joined #git
17:13 edude03 joined #git
17:13 fuchstronaut joined #git
17:16 LeBlaaanc joined #git
17:16 IWishIKnew joined #git
17:16 davis joined #git
17:16 davis hello
17:17 Dyonisos joined #git
17:18 emPi joined #git
17:19 _ikke_ hi
17:19 Dyonisos joined #git
17:22 LeBlaaanc joined #git
17:22 narendraj9 joined #git
17:23 burtons joined #git
17:24 ajf- joined #git
17:25 lindenle joined #git
17:27 rath joined #git
17:28 fuchstronaut joined #git
17:29 danslo_ joined #git
17:31 d5sx43 joined #git
17:31 LeBlaaanc joined #git
17:31 bdonnahue left #git
17:33 byteflame joined #git
17:33 disi hey
17:33 gitinfo disi: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.
17:35 edude03 joined #git
17:35 pwnz0r joined #git
17:36 mingrammer joined #git
17:36 byteflame joined #git
17:36 d5sx43 joined #git
17:39 dreiss joined #git
17:39 LeBlaaanc joined #git
17:40 mohabaks joined #git
17:40 eijk joined #git
17:41 blackwind_123 joined #git
17:41 jwmann joined #git
17:43 aep-shoutlet joined #git
17:44 ShekharReddy joined #git
17:45 Zamerick joined #git
17:45 Noldorin joined #git
17:46 Harwood_ joined #git
17:47 sagerdearia joined #git
17:47 byteflame joined #git
17:47 adino joined #git
17:48 ffs joined #git
17:49 davis so i am confused, ...
17:49 davis i did $ git branch try1
17:49 davis did a bunch of adds and a commit and then a push
17:50 edude03 joined #git
17:50 Zamerick_ joined #git
17:50 davis now when I reclone the repot, get checkout try1 is not there and my changes appear to be in master branch
17:50 davis did i need to do $push try1 master?
17:51 _ikke_ git branch only creates a branch, but your current branch remains the same
17:51 Literphor joined #git
17:51 _ikke_ You have to checkout the branch to create commits on that branch
17:51 davis ahh
17:52 _ikke_ git checkout -b <branch> is a shortcut for that
17:52 davis so $ git branch try1; git checkout try1; ...
17:52 davis many thanks
17:52 blackwind_123 joined #git
17:52 _ikke_ yes
17:53 eivarv joined #git
17:53 _ikke_ or git checkout -b try1
17:53 LeBlaaanc joined #git
17:53 edude03 joined #git
17:56 pootsmk joined #git
17:57 emPi joined #git
17:57 fuchstronaut joined #git
17:57 davis if i do git checkout -b newbranchname, after i do my adds and commit, do I do $git push or $git push newbranchname head?
17:57 edude03 joined #git
17:58 _ikke_ git push <remote> <branchname>
17:58 Electrometro joined #git
17:59 catag87 joined #git
18:02 pikajude joined #git
18:02 vassagus joined #git
18:03 Derperperd joined #git
18:03 anuxivm joined #git
18:04 LeBlaaanc joined #git
18:04 Spec-Chum joined #git
18:05 foist joined #git
18:06 croberts left #git
18:08 vuoto joined #git
18:09 xall joined #git
18:11 anuxivm joined #git
18:11 Gsham joined #git
18:11 livingstn joined #git
18:12 ansimation joined #git
18:12 ansimation joined #git
18:13 interne7y joined #git
18:13 byteflame joined #git
18:14 edude03 joined #git
18:15 cwailes joined #git
18:16 LeBlaaanc joined #git
18:18 anuxivm joined #git
18:19 BSaboia joined #git
18:20 davisonio joined #git
18:20 vuoto joined #git
18:21 LeBlaaanc joined #git
18:21 Codfection joined #git
18:22 nilg joined #git
18:23 davisonio joined #git
18:23 anuxivm joined #git
18:23 cwailes I would like to create a copy of a subset of the branches of a repository and then remove all references to the old repo from the new one.  My first though on how to do this was to clone the repo and use set-url to change the origin location (to a new, empty github repo), but this leaves me with a bunch of references to remotes/origin/reponames that I don't want.  What is the best way to do this?
18:24 kfunk joined #git
18:24 grawity clone, set-url, then fetch --prune
18:25 takwas joined #git
18:25 grawity ah, don't forget to actually create local branches for the stuff you want
18:25 grawity before using --prune
18:25 cwailes grawity: Thanks.  I'll give that a try.
18:25 byteflame joined #git
18:26 grawity it's always possible to use `git update-ref -d` to delete stuff, too
18:26 horusdev joined #git
18:26 cwailes Will that clean up any backing metadata?  I'm guessing yes, but I wanted to double check.
18:27 cwailes And before I do the `git fetch --prune` do I need to push my local copies of the branches I care about to the new origin?
18:27 _ikke_ Not necessarily
18:28 anu1 joined #git
18:28 cdown_ joined #git
18:29 LeBlaaanc joined #git
18:29 grawity there's not much "backing metadata", really
18:30 grawity just the refs and reflogs
18:30 grawity and yes, --prune will delete everything that's not on the new remote
18:30 mingrammer joined #git
18:30 grawity it *won't* touch actual local branches, though
18:30 _ikke_ only remote tracking branches
18:30 grawity just the remote/origin/ ones
18:30 _ikke_ so that doesn't matter for the new remote
18:32 yoan___ joined #git
18:34 adino joined #git
18:34 cwailes Thanks everyone.  That appears to have done it.
18:35 byteflame joined #git
18:35 davisonio joined #git
18:36 byteflame joined #git
18:37 anuxivm joined #git
18:37 editshare-ericf joined #git
18:39 oich_ I have rebased twice where there is only the addition of a file in the parent branch that is new. The previous rebase worked. Now, it fails including warnings about trailing whitespace when applying the target branch changes and apparently rebases correctly if I pass -Xignore-space-at-eol. Do you know how there can be conflict about whitespace at the end of a line when it is essentially...
18:39 byteflame joined #git
18:39 oich_ ...replaying commits that were previously committed?
18:40 oich joined #git
18:40 davisonio joined #git
18:40 oich left #git
18:40 oich joined #git
18:41 anuxivm joined #git
18:41 oich I have rebased twice where there is only the addition of a file in the parent branch that is new. The previous rebase worked. Now, it fails including warnings about trailing whitespace when applying the target branch changes and apparently rebases correctly if I pass -Xignore-space-at-eol. Do you know how there can be conflict about whitespace at the end of a line when it is essentially...
18:41 oich ...replaying commits that were previously committed?
18:41 eijk joined #git
18:43 LeBlaaanc joined #git
18:44 sfeinste joined #git
18:45 edude03 joined #git
18:46 ResidentBiscuit Do you need to enable rerere?
18:46 Derperperd joined #git
18:47 BSAlb joined #git
18:48 anuxivm joined #git
18:48 b1tchcakes joined #git
18:48 mohabaks joined #git
18:49 nilg joined #git
18:49 _ikke_ ResidentBiscuit: yes
18:49 anuxivm joined #git
18:50 ResidentBiscuit Why isn't rerere enabled by default?
18:53 anuxivm joined #git
18:54 LeBlaaanc joined #git
18:58 fuchstronaut joined #git
18:59 eijk joined #git
19:00 MZAWeb_ joined #git
19:01 BSaboia joined #git
19:02 davisonio joined #git
19:03 LeBlaaanc joined #git
19:07 hashpuppy joined #git
19:07 eijk joined #git
19:08 vassagus joined #git
19:11 Orion3k joined #git
19:13 maucar joined #git
19:13 LeBlaaanc joined #git
19:13 catag87 joined #git
19:15 edude03 joined #git
19:18 burtons joined #git
19:20 cwailes I'm now getting a warning about a broken ref `refs/remotes/origin/HEAD`.  What's the best way to fix this?
19:20 jimi_ joined #git
19:21 horusdev joined #git
19:21 livingstn joined #git
19:22 _ikke_ git update-ref -d refs/remotes/origin/HEAD
19:22 _ikke_ (assuming origin was the old remote)
19:23 cwailes I used `git remote set-url origin new-url` to point it at a new, empty, repository.
19:26 LeBlaaanc joined #git
19:26 horusdev joined #git
19:27 davisonio joined #git
19:29 Zamerick joined #git
19:29 cwailes _ikke_: I ran that command but I'm still getting the error.  If I do `git show-ref --head` I don't see a refs/remotes/origin/HEAD, only refs/remotes/origin/master.
19:30 shgysk8zer0 joined #git
19:32 danslo_ joined #git
19:34 LeBlaaanc joined #git
19:35 burtons joined #git
19:35 Zamerick_ joined #git
19:36 kyan joined #git
19:37 MZAWeb joined #git
19:38 Zamerick__ joined #git
19:38 tvw joined #git
19:39 elementalest joined #git
19:39 elementalest joined #git
19:40 Zamerick joined #git
19:42 LeBlaaanc joined #git
19:42 vuoto joined #git
19:44 rwb joined #git
19:48 mingrammer joined #git
19:48 ZedCee joined #git
19:49 Kicer86 joined #git
19:50 LeBlaaanc joined #git
19:50 acetakwas joined #git
19:52 Electrometro joined #git
19:52 grill__ joined #git
19:53 ZedCee left #git
19:53 grill__ question: let's someone without merge access makes a pull request. how can I get those changes?
19:53 acetakwas joined #git
19:54 ZedCee joined #git
19:54 grawity you tell git to pull the specified branch from the specified repo
19:55 grawity is this a Github pull request, or another website pull request, or the email kind?
19:55 grill__ github
19:55 grill__ specifically, this guy
19:55 grill__ https://github.com/AHAAAAAAA/PokemonGo-Map/pull/2610/commits/f47ce459dba04d3a3890297fcdf138fe05f2d332
19:55 Electrom_ joined #git
19:55 yqt joined #git
19:56 Rodya_ joined #git
19:56 zumba_addict joined #git
19:56 grawity fetch refs/pull/2610/head from AHAAAAAAA's repo, or the 'adapt-radius-to-api-change' branch from EarthlingKira's fork
19:57 zumba_addict hey folks, i git stash pop and got conflicts. Now I want to undo since it's a newly created local branch
19:57 zumba_addict i'll just put my code manually
19:57 zumba_addict how do I undo it?
19:57 zumba_addict git reset --HARD right
19:57 _ikke_ did you have any changes unrelated to the stash?
19:58 _ikke_ if not, then yes, reset --hard will fix it
19:58 zumba_addict no
19:58 zumba_addict cool
19:58 Eugene Use apply+drop instead of pop
19:58 _ikke_ Doesn't make a difference in this case
19:58 _ikke_ when conflicts happen, it's the same as apply
19:58 Eugene It saves you if you forget that you pop'ed
19:58 zumba_addict i can try again
19:59 ZedCee joined #git
19:59 Eugene (or at least makes it less of a PITA to find the stash again)
19:59 zumba_addict it's weird that it's making a conflict but the changes I made is on a clean area
19:59 _ikke_ it provides the hash of the stash
19:59 zumba_addict what does that happen?
19:59 Eugene Probably because you're not applying the stash to the same thing that it was created from
20:00 zumba_addict all I do now is open the file and paste my new code which doesn't really make sense
20:00 shinnya joined #git
20:00 zumba_addict yeah
20:00 zumba_addict that is true Eugene
20:00 zumba_addict it's from a different branch
20:00 zumba_addict now i see why
20:00 dopesong joined #git
20:01 edude03 joined #git
20:01 zumba_addict is it still possible to stash though?
20:02 grill__ grawity cool thanks
20:02 Eugene Sure, you'll get a new stash
20:02 zumba_addict just want to learn new parameters
20:02 zumba_addict i mean, is it still possible to stash pop the previously one i made on a new branch
20:02 zumba_addict into the new branch
20:03 Eugene If it's still lasted in `git stash list`, then yes, you can Apply or Pop it
20:03 zumba_addict ah
20:03 _ikke_ if a conflict happens, it won't drop the stash
20:03 zumba_addict i tried pop earlier but had conflict
20:03 zumba_addict it still listed
20:03 zumba_addict that's why i ran git reset --hard
20:03 _ikke_ ^^
20:03 Eugene Then you're in luck
20:04 _ikke_ Otherwise, it outputs the hash
20:04 zumba_addict so try apply?
20:04 zumba_addict k
20:04 _ikke_ than it's just matter of git stash create <hash>
20:04 _ikke_ (if it had been dropped)
20:04 zumba_addict got it
20:04 _ikke_ s/than/then/
20:06 ZedCee joined #git
20:08 ZedCee left #git
20:08 ahmedelgabri joined #git
20:11 edude03 joined #git
20:13 vassagus joined #git
20:17 m0viefreak joined #git
20:21 LeBlaaanc joined #git
20:21 livingstn joined #git
20:21 mingrammer joined #git
20:22 elastix joined #git
20:23 t0by joined #git
20:26 Rodya_ joined #git
20:26 byteflame joined #git
20:26 lindenle joined #git
20:28 byteflame joined #git
20:28 TTachyon joined #git
20:28 Gsham joined #git
20:28 LeBlaaanc joined #git
20:28 TTachyon .
20:29 _ikke_ ..
20:29 Zarthus ...
20:29 Electrometro joined #git
20:30 d4rklit3 joined #git
20:31 ZedCee_ joined #git
20:31 PabloR joined #git
20:32 PabloR http://gitimmersion.com/lab_16.html is a method to revert changes committed to a repo. It uses ‘git revert HEAD’ (or hash) to create a new commit that removes these changes & ‘git reset —-hard <hash>’ to reset the branch to the desired commit. But it also requires tagging and later untagging the unwanted commits to allow them to be garbage collected. Why is the tag/untag part needed? Wouldn’t the unwanted commits have been
20:32 PabloR garbage collected if they were not tagged at all anyway?
20:33 _ikke_ yes
20:33 _ikke_ where does it mention that?
20:33 fahadash joined #git
20:33 PabloR _ikke_ is that for me?
20:33 PabloR This goes also with http://gitimmersion.com/lab_17.html
20:34 _ikke_ PabloR: yes
20:34 ZedCee_ I am receiving 'noop' in my git-rebase-todo after a 'git rebase mybranch -i HEAD~3'.  I've tried different
20:34 ZedCee_ shells/editors/terminals to no aavail.  Anyone might be able to help on that?  Following the direction of our lead
20:34 ZedCee_ dev, he is baffled too.
20:35 PabloR _ikke_ Why would the tags be needed at all?
20:35 jimi_ joined #git
20:35 _ikke_ PabloR: they wouldn't
20:35 PabloR _ikke_ That's what I thought. Not sure why they do it in the explanation
20:35 PabloR Seems useless
20:35 LeBlaaanc joined #git
20:36 _ikke_ But where do they explain that? I don't see it
20:36 PabloR Let me paste, just a sec
20:37 ShekharReddy joined #git
20:38 PabloR IN the point #5 Nothing is Ever Lost
20:38 PabloR But what happened to the bad commits? It turns out that the commits are still in the repository. In fact, we can still reference them.
20:38 grawity that is very temporary, however
20:38 PabloR then, "Commits that are unreferenced remain in the repository until the system runs the garbage collection software."
20:39 PabloR My understanding is that they untagged them to let them be garbage collected
20:39 grawity hmm, that article
20:40 grawity it could be that they just use a tag to avoid having to remember a commit ID
20:40 PabloR Unless I misunderstood. But why would the tag them anyway?
20:40 grawity or to have it show up in the commit log
20:40 PabloR ok, as long as it's not needed, I get it
20:41 PabloR Now, the other question is what’s the difference with ‘git commit --amend’
20:41 _ikke_ between what?
20:41 cwailes left #git
20:42 PabloR The first method to revert changes and the ‘git commit --amend’ method
20:42 anuxivm joined #git
20:42 PabloR Are they equivalent in results?
20:42 LeBlaaanc joined #git
20:42 grawity --amend doesn't revert changes
20:42 _ikke_ --amend is history rewriting
20:42 grawity it recreates the last commit
20:43 _ikke_ just like reset --hard
20:43 davisonio joined #git
20:43 _ikke_ git revert is not rewriting history
20:43 pwnz0r joined #git
20:43 PabloR ok, but you get to the same result, correct? Either you amend or you revert and then commit again
20:43 grawity hmm, if one wanted to amend 'manually', reset --soft might be more useful
20:43 grawity by 'revert' do you mean `git revert` or `git reset`?
20:44 PabloR git revert + git reset + git commit (the desired commit)
20:44 ribasushi joined #git
20:44 PabloR Like the 1st method + git commit
20:44 _ikke_ PabloR: git revert is its own thing
20:44 _ikke_ no reset necessary
20:45 grawity reset+commit? yeah, you could --amend that way
20:45 grawity but `git revert` is something else entirely
20:45 PabloR Ah, ok, I was just following th e method described in the website
20:46 nidr0x joined #git
20:46 _ikke_ Those are two seperate methods
20:46 _ikke_ not a single one combined
20:46 PabloR I understand. The website uses both methods separately to get to the desired result
20:47 grawity I have to say, though, that website is *confusing*
20:47 PabloR Ok, I was using it to learn how to actually use git in practical examples
20:48 nnyk joined #git
20:48 ZedCee_ Anyone know anything about "noop" showing up in git-rebase-todo after 'git rebase my branch -i HEAD~3'
20:48 PabloR I'll go back to reading a bit more. Thanks guys
20:48 JeroenT joined #git
20:49 vassagus joined #git
20:49 LeBlaaanc joined #git
20:49 livingstn joined #git
20:49 grawity those examples kinda lack context though
20:49 dhollinger joined #git
20:50 _ikke_ ZedCee_: try git rebase -i "HEAD~3"
20:50 PabloR <grawity> reset+commit? yeah, you could --amend that way.  Are these then equivalent?
20:50 grawity no
20:51 JeroenT joined #git
20:51 byteflame joined #git
20:51 ZedCee_ _ikke_: I'll take a look at that
20:52 ZedCee_ _ikke_: Still noop
20:52 ajf- joined #git
20:52 _ikke_ and git log HEAD~3.. ?
20:52 Cassiopaya joined #git
20:52 byteflam1 joined #git
20:53 adino joined #git
20:53 dasaj joined #git
20:54 sfeinste joined #git
20:55 IlNominePatre joined #git
20:55 editshare-ericf joined #git
20:55 ZedCee_ _ikke_: You lead me on to something, You were right after all
20:56 _ikke_ ZedCee_: Please explain
20:56 PabloR grawity With reset I reset the branch to the desired commit. If I then commit what I need. What would be different than just doing git --ammend?
20:57 LeBlaaanc joined #git
20:57 sfeinste joined #git
20:57 ZedCee_ _ikke_: User error.  Wrong syntax
20:58 _ikke_ Ok
20:58 _ikke_ What did you use?
20:58 sagerdearia joined #git
20:59 PabloR Wouldn't the amend be similar to the new commit after the reset?
21:00 ZedCee_ Exactly what you provided the second time.  Seems you don't specify as much as I was. Branch der
21:00 _ikke_ ah ok
21:00 _ikke_ syntax is git rebase [-i] <base> [checked_out_branch]
21:02 anuxivm joined #git
21:02 ZedCee_ Good to know.  You can look at it two ways, too much information, or not enough
21:02 fuchstronaut joined #git
21:02 ZedCee_ I feel like less of a dolt though now that you provided that
21:02 davisoni_ joined #git
21:03 atal421 joined #git
21:03 emPi joined #git
21:04 davisoni_ joined #git
21:04 burtons joined #git
21:05 davisoni_ joined #git
21:06 ZedCee_ _ikke_: Thanks again
21:06 ZedCee_ left #git
21:06 LeBlaaanc joined #git
21:09 atal421 left #git
21:10 ahmedelgabri joined #git
21:11 byteflame joined #git
21:11 horusdev joined #git
21:15 LeBlaaanc joined #git
21:15 millerti joined #git
21:16 Gsham joined #git
21:16 disi joined #git
21:17 sfeinste joined #git
21:17 elastix joined #git
21:19 sagerdearia joined #git
21:20 LeBlaaanc joined #git
21:21 horusdev joined #git
21:21 sfeinste joined #git
21:23 LeBlaaanc joined #git
21:23 le_melomane joined #git
21:26 phanimahesh joined #git
21:26 shum joined #git
21:30 nettoweb joined #git
21:30 elastix joined #git
21:33 danslo_ joined #git
21:33 ajf- joined #git
21:36 horusdev joined #git
21:36 Arban joined #git
21:37 elastix joined #git
21:37 boombatower joined #git
21:37 jlebar I'm trying to fetch git notes from https://github.com/llvm-project/llvm-project, but it seems to hang.  I ran "git fetch origin refs/notes/commits:refs/notes/commits" under strace, and it seems to make progress for a while, then it enters into a loop where it polls()'s an fd for a long time, making no progress.  Then there's a bit of progress, then another long wait poll()'ing...  Does github throttle pulling of notes or something?
21:39 LeBlaaanc joined #git
21:40 rgrinberg joined #git
21:40 mellernoia joined #git
21:42 NeXTSUN joined #git
21:44 elastix joined #git
21:44 Rodya_ joined #git
21:45 inflames joined #git
21:46 davisonio joined #git
21:46 LeBlaaanc joined #git
21:46 jimi_ joined #git
21:46 rik316 joined #git
21:47 edude03 joined #git
21:47 sythe joined #git
21:48 johnmilton joined #git
21:48 burtons joined #git
21:52 LeBlaaanc joined #git
21:52 _ikke_ jlebar: downloading with 3MB/s for me
21:52 _ikke_ about 30 seconds
21:52 Rodya_ joined #git
21:52 bremner that's a lot of notes.
21:52 _ikke_ yeah
21:52 _ikke_ 86MB
21:52 jlebar One per commit, I believe.
21:53 jlebar Giving an svn revision for every commit.
21:53 _ikke_ yeah
21:53 jlebar How are you measuring 3mb/s?
21:53 _ikke_ Notes: git-svn-rev: 277214
21:53 _ikke_ git output
21:53 jlebar hm, my git says nothing.
21:53 jlebar And I'm on 2.9.2.
21:54 _ikke_ remote: Counting objects: 785544, done.
21:54 _ikke_ remote: Compressing objects: 100% (68/68), done.
21:54 _ikke_ remote: Total 785544 (delta 16), reused 0 (delta 0), pack-reused 785459
21:54 _ikke_ Receiving objects: 100% (785544/785544), 86.43 MiB | 3.60 MiB/s, done.
21:54 _ikke_ Resolving deltas: 100% (350927/350927), done.
21:54 _ikke_ From https://github.com/llvm-project/llvm-project
21:54 _ikke_ * [new ref]         refs/notes/commits -> refs/notes/commits
21:54 jlebar Well that's what I expected to see.  So...something is busted with my git.  Not unbelievable, I built it from source recently.
21:55 Rodya_ joined #git
21:55 sondr3 joined #git
21:57 spudowiar joined #git
21:57 Rodya_ joined #git
22:00 mellernoia-work joined #git
22:01 Rodya_ joined #git
22:01 aidalgol joined #git
22:01 LeBlaaanc joined #git
22:03 mangolisk joined #git
22:03 selckin hangs for like ~10 second for me then starts the output
22:04 selckin maybe your pc is even slower then mine :)
22:04 jlebar I very seriously doubt that.  :)
22:05 jlebar With -v it's spinning on git-upload-pack.  I see some old threads from git 2.5 with people having this problem on Windows, but the solution of increasing the POST buffer doesn't help.  How strange...
22:05 rbern joined #git
22:06 jlebar Aha.
22:06 ekinmur joined #git
22:06 jlebar I had a bunch of refs in my repo that weren't upstream.
22:06 jlebar Like, half a million or so.
22:06 jlebar Get rid of those and it works fine.
22:09 mingrammer joined #git
22:09 selckin thats a lot of working
22:11 LeBlaaanc joined #git
22:13 Rodya_ joined #git
22:13 SwiftMatt joined #git
22:14 flaguy48 left #git
22:14 jrnieder joined #git
22:15 interne7y joined #git
22:15 Rodya_ joined #git
22:19 joki joined #git
22:19 LeBlaaanc joined #git
22:21 Cyclohexane joined #git
22:23 lindenle joined #git
22:24 paul424 joined #git
22:27 madewokherd joined #git
22:27 davisonio joined #git
22:28 flaguy48 joined #git
22:29 avar TIL "git commit --date=tea" is a thing
22:29 avar tea time = 17:00:00
22:29 FilipNortic joined #git
22:30 edude03 joined #git
22:30 fuchstronaut joined #git
22:30 Zarthus avar: :D
22:30 _ikke_ lol
22:31 edude03 joined #git
22:32 adino joined #git
22:32 SwiftMatt joined #git
22:32 Zarthus If i'm reading the code right it doesn't support hours eleven and twelve either
22:32 Zarthus this might acxtually be a bug
22:32 Zarthus https://github.com/git/git/blob/5c589a73de4394ad125a4effac227b3aec856fa1/date.c#L989-L990
22:32 Zarthus https://github.com/git/git/blob/5c589a73de4394ad125a4effac227b3aec856fa1/date.c#L942
22:32 Sasazuka http://imgur.com/a/FGFO8
22:33 LeBlaaanc joined #git
22:33 _ikke_ Zarthus: You are right
22:33 _ikke_ It doesn't support them
22:33 Zarthus Sad face.
22:34 avar Also as far as I can tell this is documented nowhere
22:34 Zarthus avar: most easter eggs aren't
22:35 _ikke_ I think it's an existing time spec
22:35 avar well e.g. --date=now is quite useful
22:35 _ikke_ I do recall reading about it somewhere
22:35 avar "tea" is definitely not in some timespec
22:35 avar git-commit just documents that --date accepts various ISO/RFC dates, not these aliases
22:35 occup4nt left #git
22:36 Zarthus I think --date=now is usual behaviour  `date --date` as well, perhaps they just did it to mimic that
22:36 occupant joined #git
22:36 _ikke_ you can do something like "1 month ago noon"
22:36 arooni anyone use vcsh here with git?  i've got a new repo set up correclty; but whenever i do; vcsh weechat; it takes me to ~ and not ~/.weechat ;  everything else is fine.  how to fix?
22:37 _ikke_ ah approxidate
22:37 lucas_ai_ joined #git
22:39 Twirl joined #git
22:39 _ikke_ teatime was added by Linus himself
22:39 Twirl Hello, If i commit, then make changes but want to revert them without committing, can i just switch branches back and forth?
22:40 github423 joined #git
22:40 _ikke_ No, switching branches does not affect uncomitted changes
22:40 fractalized joined #git
22:40 _ikke_ those might even prevent switching branches
22:40 Twirl so how do i do it
22:40 _ikke_ Twirl: You can use git stash for example to temporarily stash the changes away
22:40 Twirl no i just want to rever
22:40 Twirl revert
22:41 Twirl to the state of the project at the last commit
22:41 _ikke_ git reset --hard HEAD # cannot be undone
22:42 Twirl alright
22:43 Twirl left #git
22:44 fuchstronaut joined #git
22:45 LeBlaaanc joined #git
22:46 snowkidind joined #git
22:47 PresidentBiscuit joined #git
22:47 qsx !next
22:47 gitinfo Another satisfied customer. NEXT!
22:48 snowkidind joined #git
22:48 Derperperd joined #git
22:49 rolha joined #git
22:50 vassagus joined #git
22:53 LeBlaaanc joined #git
22:57 fractalized left #git
22:58 bolovanos joined #git
22:58 fairuz joined #git
22:58 fractalized joined #git
22:58 jetpack_joe joined #git
23:00 fractalized left #git
23:00 fractalized joined #git
23:01 LeBlaaanc joined #git
23:03 bibble joined #git
23:03 mingrammer joined #git
23:04 LeBlaaanc joined #git
23:04 bibble .seen sitaram
23:05 bibble !seen sitaram
23:06 qsx i doubt gitinfo has such a feature.
23:06 bibble Ah ;_;
23:07 Zarthus they're online right now, though
23:07 Zarthus the idletime suggests 94 hours.
23:07 LeBlaaanc joined #git
23:07 bibble sitaram: Currently, only root can "ssh git@host create user/repo-name". What needs to be changed ?
23:08 sr105 joined #git
23:09 nettoweb joined #git
23:10 bibble The gitolite-admin conf: http://termbin.com/df3e
23:10 bibble Thanks Zarthus :)
23:11 sr105 how do I debug a failed ff-only merge? I'm getting "Not possible to fast-forward". The graph is a straight line. http://pastebin.ca/3668652 I'm going from HEAD (at the bottom) to the top-most commit.
23:11 Zarthus i couldn't help you with your issue, but you owe it to me to fix the mixture of tabs and spaces  bibble!
23:12 spudowiar lol
23:12 spudowiar you owe it to us all
23:12 rgrinberg joined #git
23:13 bibble @all: Will do, lol. Done it in vim. Need to set config so vim doesn't do that.
23:14 is_null joined #git
23:14 sr105 I've had trouble merging on this machine before. If I push the branches, I can merge on another machine without issue.
23:14 oich When rebasing "other-branch" after a file was added to "master", the merge of "other-branch" has conflicts with warnings about trailing whitespace. I can get around the problem by passing -Xignore-space-at-eol . Do you know how there can be conflict about whitespace at the end of a line when it is essentially replaying commits that were previously committed?
23:14 is_null joined #git
23:14 is_null hi all, is there a tool to mirror a git repo ? i mean, able to do complete and proper push/pull
23:15 qsx is_null: you mean, like, man git-clone?
23:15 gitinfo is_null: the git-clone manpage is available at http://jk.gs/git-clone.html
23:16 bibble is_null , gitolite has a mirror option in the config...
23:16 oich is_null --mirror in particular
23:16 Rodya_ joined #git
23:17 sr105 is_null : if I remember correctly, git doesn't do mirrored clones. But you can use --mirror and then with commands or by hand editing the .git/config, make the mirror into a regular working git repo.
23:18 ajf- joined #git
23:18 Derperperd left #git
23:19 NeXTSUN_ joined #git
23:19 Derperperd joined #git
23:20 circ-user-1aOox joined #git
23:20 NeXTSUN_ joined #git
23:20 is_null nice, thanks all !! i was investigating gitolite's capabilities, seems like the right path
23:21 is_null qsx: git-clone: Clone a repository into a new directory, we're far from "complete and proper push/pull" with just that ;)
23:21 LeBlaaanc joined #git
23:21 mablae_ joined #git
23:22 is_null http://gitolite.com/gitolite/mirroring.html perfect ^^
23:23 sr105 I ran into the mirroring problem a lot at first because I came from mercurial where a local clone is a full mirror.
23:24 sr105 I couldn't understand why git didn't "work"
23:24 Darren_ joined #git
23:24 qsx is_null: no.
23:24 qsx is_null: what do you think what a clone is?
23:24 lindenle joined #git
23:25 qsx well.  a “normal” clone is not a full clone, it has a limited !refspec and so on, but a clone --mirror is about everything you can get through the git protocol
23:25 gitinfo [!refspecs] Refspecs are used by fetch/push to *spec*ify which *ref*s to transmit where. They have the form "source:destination". They can be prefixed with a "+" to force the update, possibly displacing existing history. More info: http://jk.gs/git-fetch.html or http://i.qkme.me/3tke7r.jpg
23:26 is_null qsx: and then for updates you can use git fetch
23:26 qsx is_null: yes.
23:27 is_null i've already got a dirty script doing all that, i'm just investigating existing solutions before continuing to reinvent this ^^
23:27 qsx sr105: well, mercurial works a bit differently, it does not have git’s concept of refs
23:28 qsx sr105: a git clone is a clone of mostly everything as well, in contrast to subversion, where you just have one revision in your working directory
23:28 qsx but hg and git come pretty close.
23:28 osse is_null: how about a good old-fashioned cp -r or rsync
23:30 is_null never tried to rsync over on github, even if that's possible i'm not sure that's what should be done xD
23:30 is_null btw, is there a tool to help iterating over the different values a YAML key has had in a repo ?
23:30 osse not that's not possible
23:30 is_null or is that too specific ?
23:30 osse *no
23:31 LeBlaaanc joined #git
23:31 osse and you can't copy .git/config either that way
23:31 EvilPeng1 joined #git
23:31 osse but if the source is on github then clone --mirror is what you want
23:31 is_null the source depends on the user
23:32 is_null known users would push on the server and unkown users would make pull requests
23:32 texinwien_ joined #git
23:34 danslo_ joined #git
23:34 PioneerAxon joined #git
23:34 neca joined #git
23:34 nettoweb joined #git
23:35 sr105 qsx: the biggest difference was that with hg, you get all of the remotes and their refs when you clone.
23:35 sr105 just a different mindset
23:37 ffabi joined #git
23:37 sr105 anyone know how to get git to tell why my ff-only merge is aborting?
23:38 osse sr105: is there no output at all?
23:38 sr105 just "fatal: Not possible to fast-forward, aborting."
23:39 osse that's all that needs to  be said
23:39 dhollinger joined #git
23:39 sr105 and if you look at my paste it's a straight line
23:39 kfpratt joined #git
23:39 sr105 how can it not ff a straight line?
23:39 osse which paste?
23:39 sr105 http://pastebin.ca/3668652
23:39 sr105 it had already scrolled by
23:39 thiago don't argue with the evidence. If it's telling you it's not a straight-line, it isn't.
23:40 thiago if you think it is a straight line, maybe you're wrong or you know something Git doesn't (missing commits to be fetched)
23:40 osse line 21 looks weird. why is the hash more indented than the rest? bad copy-paste?
23:41 osse sr105: run git log 7c25be2..HEAD
23:43 sr105 If I run that, I get every commit back to the root of the repo. If I flip the order, I get 3 commits.
23:43 rubie joined #git
23:44 sr105 the indentation was because of the graph showing a merge I think
23:44 osse sr105: then it's not a straight line
23:44 osse if it were, the output would have been empty
23:45 fuchstronaut joined #git
23:45 osse sounds like you're trying to merge branches that have no common ancestor
23:46 osse so you have straight lines for sure. but you have two of them
23:46 Electrometro joined #git
23:46 sr105 maybe the repo is corrupted. I pushed the branches and they ff merged on another machine with no issue
23:46 LeBlaaanc joined #git
23:46 sr105 but I did get an error because I had a stash
23:47 sr105 I just wish it would give a reason.
23:48 sr105 It's acting weird. If I run that git log HEAD..branch command, I get 3 commits. If I add --graph and --all I get 10 more commits before HEAD
23:48 Mannn joined #git
23:49 sr105 I'll just blow away this local repo.
23:49 sr105 oh.... it's a shallow clone. Could that be the problem?
23:50 qsx shallow clones bring interesting problems with them, so yes, it could
23:52 LeBlaaanc joined #git
23:54 ffabi joined #git
23:54 fahadash joined #git
23:54 sr105 It's a repo in a giant android build tree. I make them 1 commit shallow to speed up builds. Thanks.
23:56 vassagus joined #git
23:58 mingrammer joined #git
23:59 EvilPeng1 joined #git
23:59 LeBlaaanc joined #git

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