Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2017-02-01

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

All times shown according to UTC.

Time Nick Message
00:00 AlexMax B is identical to A except it has additional "protocol breaking" fixes
00:00 AlexMax Somebody accidentally committed and pushed something to A that breaks protocol
00:00 AlexMax What is the proper way of "transplanting" those two changesets from one branch to the other?
00:01 f0rks joined #git
00:01 kadoban cherry-pick probably, and then either revert it on the branch it doesn't belong, or edit history if you're okay with annoying the shit out of your peers
00:01 AlexMax I regularly merge from A to B, but NEVER the other way around
00:02 AlexMax kadoban, what I'm trying to avoid is reverting it on branch A, but then that "revert" occuring on branch B when I merge A into B
00:02 rwp Has the breaking commits been published yet? Or is this completely internal?
00:02 AlexMax it's been published
00:02 AlexMax no putting the genie back in the bottle
00:02 rwp Bummer. :-(
00:03 solenodic joined #git
00:03 AlexMax I guess what I need to do is revert on A FIRST
00:03 AlexMax then cherry pick the commits from A to B
00:03 AlexMax that way, I don't get "Oh, you revertsed those on A, let me revert those on B too"
00:03 kadoban AlexMax: revert it, merge A into B, and then revert the revert on B
00:03 rwp If you have already been merging from A to B then wouldn't reverting it on A then follow through with the next merge of A into B?
00:03 Nysar joined #git
00:04 Bizkit joined #git
00:04 AlexMax kadoban, If this breakage took place over two non-contiguous separate commits, is there a good "order" to revert them in?
00:04 AlexMax do I revert the older one first, or the newer
00:05 kadoban I'm not sure actually
00:05 valar joined #git
00:05 AlexMax kadoban, I think maybe cherry picking might make more sense in the context of history than reverting the revert
00:06 nowhereman joined #git
00:06 lindenle joined #git
00:07 sammyo joined #git
00:09 DolphinDream joined #git
00:09 red82 joined #git
00:10 endiruna joined #git
00:11 _xor joined #git
00:12 nostrora joined #git
00:12 Darren_ joined #git
00:12 spacelord_ joined #git
00:14 govg joined #git
00:15 theskillwithin joined #git
00:17 dendazen joined #git
00:20 invisbl_ joined #git
00:23 Cavallari joined #git
00:28 mehola joined #git
00:33 barteks2x joined #git
00:33 spacelord_ joined #git
00:34 shgysk8zer0_ joined #git
00:40 Gsham joined #git
00:41 nckpz joined #git
00:45 blackwind_123 joined #git
00:46 Ender13 joined #git
00:48 matoro joined #git
00:50 Hobbyboy joined #git
00:51 a_thakur joined #git
00:51 Lyqyd joined #git
01:00 jccn joined #git
01:01 solenodic joined #git
01:04 Atm0spher1c joined #git
01:06 solrize joined #git
01:08 fuzzybear3965 Gitlab.com is down so I decided to zip up a local copy of a repository and file transfer it over RDP (SSH isn't an option).
01:08 fuzzybear3965 So, now I have some/stale/repo and some/fresh/repo and I want to fetch the changes from some/fresh/repo into some/stale/repo (all locally).
01:08 fuzzybear3965 How would I accomplish this?
01:08 menip joined #git
01:09 Atm0spher1c joined #git
01:09 tang^ you can add a remote with a file path
01:09 arescorpio joined #git
01:13 shgysk8zer0_ joined #git
01:15 fuzzybear3965 Okay, tang^, that's on option.
01:15 fuzzybear3965 The problem I'm having right now is being able to specify a branch of that file:/// repository in my fetch request.
01:15 fuzzybear3965 Adding a remote would fix this problem, I think.
01:16 SwiftMatt joined #git
01:16 fuzzybear3965 But, if I wanted to avoid adding a remote (because I'm only going to use this once), do you know how I might specify a branch without a remote being added?
01:17 Eugene fuzzybear3965 - `git fetch /path/to/repo <branch>` will give you a FETCH_HEAD. Merge that.
01:17 fuzzybear3965 Eugene, I want to rebase that....
01:17 fuzzybear3965 I don't know how to use the FETCH
01:17 fuzzybear3965 _HEAD that I have :(.
01:18 fuzzybear3965 Ope, I think I have it!
01:18 fuzzybear3965 One sec.
01:18 Eugene Use it just like a remote/foo branch name
01:18 fuzzybear3965 Yuhhs.
01:18 fuzzybear3965 Yeah, it worked.. Thanks!
01:19 fuzzybear3965 For the future versions of me: `git fetch path/to/latest/copy master` followed by `git rebase FETCH_HEAD master` worked perfectly!
01:19 fuzzybear3965 (when the working directory is path/to/stale/copy)
01:20 fuzzybear3965 Eugene, tang^, gitlab.com is still down so I'm happy that you all helped me.
01:21 fuzzybear3965 I still would have been stuck :).
01:23 Guma joined #git
01:24 Eugene !beer
01:24 gitinfo Beer! It's what's for dinner!
01:24 CheckDavid joined #git
01:24 ^7heo Well, something tells me you're not exactly having a six pack
01:24 infernix joined #git
01:28 b1tchcakes joined #git
01:30 GodGinrai joined #git
01:33 z3t0 joined #git
01:33 fuzzybear3965 joined #git
01:35 jccn joined #git
01:39 peeps[lappy] joined #git
01:41 streetwitch joined #git
01:41 Eugene Oh jeeze, I wasn't aware of the serious of "Gitlab is down" https://docs.google.com/document/d/1GCK53Y​DcBWQveod9kfzW-VCxIABGiryG7_z_6jHdVik/pub
01:44 rivarun joined #git
01:47 rchavik joined #git
01:50 rivarun joined #git
02:01 ploop joined #git
02:02 mark4_ joined #git
02:03 dendazen joined #git
02:03 Nilesh_ joined #git
02:03 mark4_ left #git
02:03 mark4_ joined #git
02:04 Guest55 joined #git
02:04 SwiftMatt joined #git
02:04 mark4_ left #git
02:05 I440r joined #git
02:05 I440r so im trying to locate an error that ws introduced into the linux kernel somewhere between versions 4.5.7 and 4.6.0. i know 4.6.0 is bad and 4.5.7 is good. i told git bisect this and... its having me build 4.5.0
02:05 I440r ive never done a bisect operation before
02:06 I440r 4.5.0 is not between 4.5.7(good) and 4.6.0(bad)
02:08 adino joined #git
02:09 I440r can someone tell me why git bisect has selected 4.5.0 for testing when i told it 4.5.7 was good and 4.6.0 was bad?
02:13 bernardio joined #git
02:13 pks joined #git
02:13 I440r can someone tell me why git bisect has selected 4.5.0 for testing when i told it 4.5.7 was good and 4.6.0 was bad?
02:13 I440r ugh sorry
02:13 pur3eval joined #git
02:13 I440r just as i hit enter i hit cursor up. thats not what i typed :/
02:14 I440r but if theres anyone here who can help?
02:14 I440r i dont understand this bissect operation at all
02:14 I440r its like its having me test something outside the good/bad window
02:14 rmbeer joined #git
02:14 rmbeer hello
02:15 ecuanaso joined #git
02:15 finalbeta joined #git
02:16 ljc joined #git
02:16 rmbeer i view in the help of 'git log', but can't found the parameter for search by user name or mail
02:17 I440r is there anyone here who can help?
02:17 I440r all assleep :(
02:19 HedgeMage rmbeer: git log isn't that smart.  You can grep it or you can push the fast-import stream through reposurgeon if you are doing something extremely complex.
02:19 Darren_ joined #git
02:19 I440r why did get check out kernel version 4.5.0 when i told it 4.5.7 was good and 4.6.0 was bad?
02:20 rgrinberg joined #git
02:20 rkazak_ joined #git
02:20 jeffreylevesque_ joined #git
02:20 I440r or are all the bisected versions going to be 4.5.0?
02:20 ShapeShifter499 joined #git
02:21 mizu_no_oto joined #git
02:21 SpeakerToMeat left #git
02:23 HedgeMage I440r: If you are kernel hacking, you're at the level where you can read http://catb.org/~esr/faqs/smart-questions.html and then try again.  You've asked nearly simultaneously in multiple channels, which is rude, and not given enough information to let us approach your question usefully (such as the EXACT commands you are using).
02:23 HedgeMage I440r: Please don't expect the kid gloves we sometimes give the newbies.
02:24 I440r im not a kernel developer. this is absolutely the first time ive ever done a git bisect
02:24 ResidentBiscuit joined #git
02:24 I440r git checkout v4.6
02:24 I440r git bisect start
02:24 I440r git bisect bad
02:24 I440r git bisect good 4.5.7
02:25 I440r now im building 4.5.0 which is what it selected from between good and bad
02:25 I440r err no. thats outside the good/bad window. so... naturally im confused
02:25 I440r good v4.5.7 i mean
02:25 bernardio joined #git
02:25 HedgeMage I440r: Okay, in that case please still read the article -- as it outlines how to make your support requests easy to answer, albeit a bit sharply -- and I'll be gentler :)
02:26 HedgeMage And yes, I know you're not on the core team or I'd recognize you, but lots of people do kernel hacking...driver maintainers, infosec folks, etc.
02:26 HedgeMage :D
02:27 HedgeMage I440r: git bisect can sometimes get confused by cherry picks or other twists in the history, especially in something as complex as the kernel.
02:27 I440r i avoid kernel level development like the plague. i consider the kernel source tree to be one of the WORST source trees ive ever seen in my entire life.  almost totally devoid of comments. the fact that it works so damned well notwithstanding :)
02:27 I440r hegge, so git IS confused and im actually building something outside the good/bad window?
02:27 dreiss joined #git
02:27 I440r souds like a very serious bug in git bisect
02:27 I440r making it a useless tool
02:28 HedgeMage I've seen MUCH worse...but I'm in infosec, so a large portion of my work is "it's all hopeless, let's call Susan!!!"
02:28 I440r and "it works sometimes" is not a saving grace
02:28 * HedgeMage is jaded, cynical, and biased.
02:28 I440r no idea what infosec is or who susan is
02:28 HedgeMage I440r: bisect is reliable, once you know how it treats all the edge cases...the problem is that there are nearly a dozen edge cases.
02:29 I440r im just a lowbie asm coder who write forth compiler for linux in assembler
02:29 HedgeMage Susan is me, and "infosec" is short for "information security" aka computer security
02:29 I440r aha
02:29 I440r ooooh! lol
02:29 I440r your "tech support" like it or not :P
02:29 I440r got it lol
02:29 HedgeMage Meh, I try not to have to do user support.
02:30 HedgeMage More like "the internet is falling down and no one remembers how to read this code" support.
02:30 I440r been there, done that. i worked for the company that did the access control and alarm monitoring for the whitehouse, smthsonian, every single us marine base in the country.... TOTALLY unmaintainable
02:30 HedgeMage Mostly older C dialects, the occasional Ada / Fortran / asm stuff
02:30 msonntag joined #git
02:30 * HedgeMage nods.
02:31 I440r wait! you code asm? you heretic!
02:31 HedgeMage I440r: Not on purpose...only when shit's falling apart.
02:31 HedgeMage I'm rather bad at it, actually.
02:32 HedgeMage I'm a baby of the era of high-level languages like C and friends.
02:32 I440r if i were dictator of the universe i would require all software engineers to serve a 20 year apprenticeship during which time they would only be allowed to code in pur assembler
02:32 HedgeMage But, my life happens in whatever this week's emergency happens to be in. :)
02:32 I440r any deviation to be punishable by death
02:32 I440r no appeal
02:32 HedgeMage lol
02:33 HedgeMage It didn't take me 20 years, but I had an impeccable apprenticeship.
02:33 I440r i learned 6502 in 2 weeks. then bought my first compiler. and i mean the MACHINE code not just the assembler
02:33 druonysus_ joined #git
02:33 HedgeMage I've been personally yelled at (well, more like condescended to) by Brian Kernighan.  I deserved it.  ESR is a close friend and mentor.  Worked with Dave Reed and many of the other early internet guys.
02:35 I440r C is quirky, flawed and an enormous success
02:35 I440r was that him or richie that said that?
02:35 HedgeMage I'm pretty sure that was Ritchie.
02:35 I440r i consider C to be one of the single worst programming languages ever devised
02:35 I440r not because its bad
02:36 I440r but because it has convinced every moron out there and his autistic brother that 24 hours of book lernin == expert coder
02:36 HedgeMage Meh, TBH, I'm informally taught and I'm WAY more skilled than most of the CS PhDs I work with.
02:36 HedgeMage Academia jumped the shark long ago.
02:36 I440r i hate c with a passion which is why i dont do it unless im being paid and im for sure NOT the best c coder you ever saw
02:36 I440r i have no degree
02:37 I440r but im the author of what i believe is one of the fastest compilers of any non trivail programming language
02:37 I440r 4 megs of sourc per second
02:37 I440r best i can fudgestimate
02:37 * HedgeMage nods
02:37 I440r could be as much as 11 or 12 per second
02:38 I440r on my todo list is a random do nothing source code generator that can generate 20 or 30 megabytes of syntatically correct but dont try run it code
02:38 I440r split up into lots of modules of course
02:38 I440r the i can do time ./build-this
02:39 d^sh joined #git
02:39 om_henners joined #git
02:39 I440r what kind of security flaws do you handle?
02:39 I440r in code?
02:39 I440r we all know buffer overflows are bad mojo but... how do you determine that something is a flaw?
02:39 I440r i mean not every flaw is that obvious
02:41 HedgeMage Mostly code, but a bit of everything.
02:41 HedgeMage I'm multi-specialty, and tend to by my center's answer to "we don't know what this is yet"
02:41 I440r you poke around inside other random peoples code and say "THIS is bad!"
02:42 I440r or you observe someone doing an exploit and track down the exploitable feature ?
02:46 HedgeMage So, part of my time is as information security officer for Open Science Grid (mediates access and accounting for most of the supercomputing and high-throughput clusters in north America), and part is available to the National Science Foundation for projects and facilities with security problems they don't know how to solve.  Telescopes with control systems they can't secure, malware hits the south pole station,
02:46 HedgeMage whatever...
02:46 I440r oh like washington who had to pay randsome to get their security cameras back recently
02:46 HedgeMage Part of my time goes to weird stuff that lands on our doorstep (defense stuff, medical devices, whatever).  Part of it goes to DHS or DOE or NSF projects that need me to straighten out their software engineering (usually bad, and the resulting code is usually worse)
02:46 Vampire0 joined #git
02:46 HedgeMage The fun part is saving the internet, though.
02:46 I440r sounds like an interesting job
02:46 HedgeMage When NTP went crisis mode in 2015, I got to do a rescue of the code base and now we have NTPSec.
02:49 ilbot3 joined #git
02:49 Topic for #git is now Welcome to #git, the place for git help and alternative commits | Public logs at http://jk.gs/glog | First visit? Read: http://jk.gs/git | Current stable version: 2.11.0 | Getting "cannot send to channel"? /msg gitinfo .voice | git-slushie: submit your best punch line to https://madeitwor.se/git-jokes
02:49 I440r comments are not for you, they are for the poor shmuck that needs to understnd your scratchings later
02:49 Limes joined #git
02:49 I440r kernel devs are of the opinion that if you need comments you should get another job
02:50 I440r anyway time to reboot and test this bisect
02:50 I440r ty for the help
02:50 HedgeMage Meh, not entirely.  Just keep in mind that the kernel code base is actually of a size that downloading it is prohibitive in some places.
02:50 HedgeMage ttyl, and you are welcome.
02:54 dreiss joined #git
02:55 duderonomy joined #git
02:55 govg joined #git
02:55 aidalgol joined #git
03:00 Anja joined #git
03:02 Nysar joined #git
03:05 rmbeer left #git
03:12 fstd_ joined #git
03:13 madduck joined #git
03:13 madduck joined #git
03:17 Vortex34 joined #git
03:18 pur3eval joined #git
03:20 serialoverflow joined #git
03:27 zacts joined #git
03:30 madduck joined #git
03:30 madduck joined #git
03:30 GodGinrai joined #git
03:31 chipotle joined #git
03:31 Ryanar joined #git
03:32 ShapeShifter499 joined #git
03:34 a_thakur joined #git
03:35 sagerdearia joined #git
03:35 mostlybadfly joined #git
03:41 rivarun joined #git
03:45 a_thakur joined #git
03:45 madduck joined #git
03:46 a_thakur joined #git
03:47 jj96 joined #git
03:47 jj96 How to make a branch even with the master branch in same repo?
03:52 chachasmooth joined #git
03:55 perlpilot jj96: git branch even master  # now they're even  :-)
03:56 perlpilot jj96: If I can guess what you mean, it sounds like you want to merge master into your branch?
03:57 jj96 perlpilot, yes
03:57 jost__ joined #git
03:57 Goplat joined #git
03:58 sunri5e joined #git
03:58 thiago joined #git
03:58 hexagoxel joined #git
03:58 perlpilot jj96: git checkout branchname; git merge master  # unless I'm crazy
03:58 Ryanar joined #git
03:59 kyan Hi. How can I revert changes, but without getting rid of any commits?
03:59 rivarun joined #git
03:59 kyan (I just want to replace the current working tree with a previous state of the working tree, and commit that.)
04:00 kyan Googling just turns up loads of ways to get rid of my commits, which totally defeats the reason I'm using git in the first place
04:00 kyan Thanks!
04:00 shgysk8zer0_ joined #git
04:03 phroa kyan: man git revert - generates a new commit that's basically an inverse of a selected range of commits
04:03 gitinfo kyan: the git-revert manpage is available at http://jk.gs/git-revert.html
04:03 kyan Sweeet, thanks! :D
04:04 skered joined #git
04:05 madduck joined #git
04:06 kyan Oh, wow, I can pull out a single commit and reverse it without reverting the intermediate changes, too. That's pretty slick
04:10 boombatower joined #git
04:14 skered Is this possible via somet builtin or an extension?  with 'git commit -vp' I can interactivly go thru changes and choose what to stage.  However, would it be possible to queue multiple stagings in one pass?  ex: There are 5 changes to question, 'git commit -vp...', "Stack this hunk" yes to stageA, ... yes to stageB, ..., yes to stageA, ..., finished git will open $EDITOR for stageA, open $EDITOR for stageB.
04:16 variable joined #git
04:17 svm_invictvs joined #git
04:18 Atemu joined #git
04:18 pur3eval joined #git
04:23 om_henners joined #git
04:24 red82 joined #git
04:24 sbulage joined #git
04:24 aw1 joined #git
04:25 bannakaffalatta joined #git
04:28 fragMental skered: Since there is only one index/staging area, I don't think that is possible.
04:29 mehola joined #git
04:31 Darren_ joined #git
04:32 rivarun joined #git
04:34 skered Maybe but it seems possible git could at least prep all the commands to perform each commit with all the proper hunks.
04:36 tang^ joined #git
04:37 mehola joined #git
04:39 madduck joined #git
04:39 Atm0sphe1 joined #git
04:44 ayogi joined #git
04:47 dreiss joined #git
04:49 Atm0sphe1 joined #git
04:50 Anja joined #git
04:51 armyriad joined #git
04:51 mahendratech joined #git
04:52 wilornel joined #git
04:53 wilornel Hey #git! I had just done `git apply -3 /tmp/some.diff` and it worked as expected. All conflicts remained in the files with the expected <<< === >>> sections!
04:53 wilornel However, I am trying it now again and the command outputs some "error: patch failed:... Failling back to three-way merge..."
04:54 wilornel And then `git status` does not show that any of the files got modified
04:54 wilornel Also, the files I am modifying are not being ignored by .gitignore
04:56 Emperor_Earth joined #git
04:56 robotroll joined #git
04:59 Cabanoss- joined #git
05:01 Atm0sphe1 joined #git
05:02 Darren_ joined #git
05:08 pR0Ps joined #git
05:19 lucasem joined #git
05:19 acetakwas joined #git
05:20 f0rks joined #git
05:20 Andrew_K joined #git
05:26 sbulage joined #git
05:26 ckruczek joined #git
05:26 chipotle joined #git
05:31 GodGinrai joined #git
05:32 BlackMaria joined #git
05:37 f0rks joined #git
05:44 rivarun joined #git
05:45 _ikke_ wilornel: git apply --reject
05:45 lucasem joined #git
05:46 a_thakur joined #git
05:47 gigq joined #git
05:50 frobnic joined #git
05:51 fragMental joined #git
05:53 solenodic joined #git
05:54 msonntag joined #git
05:54 wilornel thank you _ikke_
06:00 freimatz joined #git
06:01 Charliechin joined #git
06:05 dec0n joined #git
06:08 gigq joined #git
06:09 jaafar joined #git
06:11 rscata joined #git
06:12 orzo what's the easiest way to tell git ot merge a branch as a single new non-merge commit?
06:13 orzo meaning, i had an experiemental branch and did a lot of nonsense commits that i don't want to remember now that i got it figured what i want
06:13 _ikke_ git merge -n <branch> && git commit
06:13 pR0Ps joined #git
06:14 _ikke_ sorry
06:14 _ikke_ git merge --squash
06:14 _ikke_ not -n
06:15 orzo thanks
06:16 zacts joined #git
06:17 lindenle joined #git
06:18 Atm0sphe1 joined #git
06:25 dreiss joined #git
06:27 jagob joined #git
06:30 pR0Ps joined #git
06:30 sbulage joined #git
06:38 overlord_tm joined #git
06:38 ThomasLocke joined #git
06:39 mda1_ joined #git
06:40 Atm0sphe1 joined #git
06:44 MineCoins joined #git
06:45 dersand joined #git
06:49 ianmethyst joined #git
06:51 Anja joined #git
06:54 zeroed joined #git
06:57 NimeshNeema joined #git
06:58 brent__ joined #git
06:59 sQVe joined #git
07:00 ertes joined #git
07:00 bocaneri joined #git
07:00 Raging_Hog joined #git
07:03 SwiftMatt joined #git
07:03 howdoi joined #git
07:04 spacelord_ joined #git
07:06 sQVe_ joined #git
07:07 seni joined #git
07:08 JeroenT joined #git
07:11 algun joined #git
07:11 variable joined #git
07:14 qt-x joined #git
07:17 zeroed joined #git
07:17 zeroed joined #git
07:19 sQVe_ joined #git
07:21 sQVe_ joined #git
07:22 solenodic joined #git
07:25 seni joined #git
07:26 pur3eval joined #git
07:26 ecuanaso joined #git
07:26 chele joined #git
07:26 nukeu666 joined #git
07:27 nukeu666 I've been working on a branch by mistake the last few commits, do I merge or to master or rebase it?
07:27 lucasem joined #git
07:30 PaulePanter joined #git
07:30 dreiss joined #git
07:31 freimatz joined #git
07:32 GodGinrai joined #git
07:34 chipotle joined #git
07:34 jnavila joined #git
07:34 theoceaniscool joined #git
07:36 eroux joined #git
07:43 vj4 joined #git
07:45 synthmeat joined #git
07:46 zefferno joined #git
07:47 ihradek joined #git
07:47 tobiasvl nukeu666: what's the situation exactly? you want to move some commits to a different branch?
07:48 nukeu666 yes
07:48 kteja joined #git
07:48 Alienpruts joined #git
07:48 GreenJello joined #git
07:48 GreenJello So I always do git push; git push --tags; is there a better way?
07:48 nukeu666 tobiasvl: want to move last 5 commits to master
07:49 kteja how to add multiple remotes??
07:50 tobiasvl nukeu666: the easiest is probably man git cherry-pick
07:50 gitinfo nukeu666: the git-cherry-pick manpage is available at http://jk.gs/git-cherry-pick.html
07:50 GreenJello I'm seeing some half solutions
07:50 tobiasvl kteja: just add multiple remotes. they need to have different names
07:50 sarri joined #git
07:50 sarri joined #git
07:51 Singmyr joined #git
07:51 GreenJello eh I guess I'll just make an alias for it in my shell
07:53 azerus joined #git
07:54 tobiasvl GreenJello: yeah, probablt best. you can set the push.followTags config to automatically push annotated tags on the branches you push, but that's all I think
07:54 ecuanaso joined #git
07:55 GreenJello I'm not sure if that's better or not
07:55 tobiasvl me neither
07:55 madduck joined #git
07:56 tobiasvl GreenJello: you can always set the "push" setting for the remote to push "refs/tags/*"
07:56 GreenJello I have >100 repos :/
07:57 _ikke_ GreenJello: git push --follow-tags
07:57 tobiasvl GreenJello: git push origin 'refs/tags/*' 'refs/heads/*' ?
07:57 GreenJello yeah, maybe follow is better; I don't know why I'd want to push tags from other branches
07:57 _ikke_ It's mainly about tags not reachable by the refs you are pushing
07:58 madduck joined #git
07:58 GreenJello tobiasvl, doesn't that push all the things, like commits on other branches?
07:58 tobiasvl GreenJello: maybe that works then. I assume you don't want --mirror
07:58 tobiasvl GreenJello: yeah it does, sorry. I somehow thought you wanted "git push --all; git push --tags"
07:58 GreenJello I'll go with -follow-tags, thanks all!
07:59 _ikke_ kteja: git remote add <name> <url>; repeat for each remote
08:00 kteja _ikke_ thank u
08:01 a_thakur joined #git
08:01 Atm0spher1c joined #git
08:02 Tobbi joined #git
08:05 zeroed joined #git
08:05 zeroed joined #git
08:07 asdgadsg joined #git
08:08 ecuanaso joined #git
08:09 asdgadsg left #git
08:11 p4trix joined #git
08:11 Murii joined #git
08:15 TomyWork joined #git
08:16 jsolano joined #git
08:17 Random832 joined #git
08:17 Macaveli joined #git
08:20 lordjancso joined #git
08:21 Atm0spher1c joined #git
08:25 elect joined #git
08:26 Wind0r joined #git
08:26 elect_ joined #git
08:26 iamtakingiteasy joined #git
08:26 pur3eval joined #git
08:27 lb1c joined #git
08:28 elect__ joined #git
08:28 seni joined #git
08:28 mikecmpbll joined #git
08:28 ozmage joined #git
08:29 gareppa joined #git
08:31 thierryp joined #git
08:31 thierryp joined #git
08:33 Dumblez0r joined #git
08:34 ozmage joined #git
08:35 jknetl joined #git
08:36 Anuj__ joined #git
08:36 gigq joined #git
08:37 tvw joined #git
08:37 DieguezZ joined #git
08:37 alcohol joined #git
08:38 seni joined #git
08:39 _xor joined #git
08:39 PHPanos joined #git
08:41 djb-irc joined #git
08:42 nowhere_man joined #git
08:43 Balliad joined #git
08:44 shgysk8zer0_ joined #git
08:49 netj joined #git
08:51 nostrora joined #git
08:51 a_thakur joined #git
08:51 Charliechin joined #git
08:52 amrits joined #git
08:52 kurkale6ka joined #git
08:53 ozmage joined #git
08:55 seni joined #git
08:56 ShalokShalom joined #git
08:57 apotry joined #git
08:58 mikecmpbll joined #git
09:00 Darcidride joined #git
09:05 nowhere_man joined #git
09:05 seni joined #git
09:08 causasui joined #git
09:09 marcogmonteiro joined #git
09:11 seni joined #git
09:11 WayToDoor joined #git
09:13 tvw joined #git
09:16 clmsy joined #git
09:17 robattila256 joined #git
09:20 seni joined #git
09:20 nowhere_man joined #git
09:21 lindenle joined #git
09:23 ozmage joined #git
09:25 cdown joined #git
09:28 seni joined #git
09:29 Tobbi joined #git
09:32 adamru joined #git
09:33 zeroed joined #git
09:33 GodGinrai joined #git
09:34 Impaloo joined #git
09:36 durham joined #git
09:37 finalbeta joined #git
09:40 govg joined #git
09:41 Snugglebash joined #git
09:45 solenodic joined #git
09:45 clmsy joined #git
09:46 Snugglebash joined #git
09:46 Snugglebash joined #git
09:47 elect joined #git
09:48 manuelschneid3r joined #git
09:49 Vgr_ joined #git
09:49 Vgr_ joined #git
09:52 grift joined #git
09:57 nowhere_man joined #git
09:58 theoceaniscool joined #git
09:59 Tobbi joined #git
09:59 jmibanez joined #git
10:04 reactormonk How do I grep for commit messages everywhere in my local repo, not just the active history?
10:05 tobiasvl reactormonk: git log --all --grep='foo'
10:06 reactormonk cool, thanks
10:06 RxMcDonald joined #git
10:07 RxMcDonald how do you add changes to commit after removingn the files from .gitignore?
10:07 djb-irc joined #git
10:08 mehola joined #git
10:08 RxMcDonald git diff shows nothing and I created a file and changed the .gitignore, why>
10:08 a3pq51 joined #git
10:08 tobiasvl RxMcDonald: git add file
10:08 RxMcDonald nvm, was looking at the fork
10:08 tobiasvl ah ok
10:09 marcogmonteiro joined #git
10:10 Serpent7776 joined #git
10:11 seni joined #git
10:15 pur3eval joined #git
10:15 aleb left #git
10:16 magey joined #git
10:17 solenodic joined #git
10:20 Hudu joined #git
10:20 magey what is the proper way to set up a scenario where my team maintains our own branch of a particual kernel version (lets say linux-4.9.y from linux-stable), which is periodically rebased with the upstream kernel.org tree so we can keep our local changes on top the upstream history
10:21 achadwick joined #git
10:21 tobiasvl magey: rebasing is going to get tiresome quickly because the team's local branches will get out of date all the time. why not merge?
10:22 magey currently i have to fetch the upstream branch (kernel.org linux-4.9.y), rebase our own 4.9.y with it, then push my local 4.9.y to origin (our central git repo)
10:22 a_thakur joined #git
10:22 magey but when you try to do that git rightfully complains that you're rewriting public history
10:22 tobiasvl yep, that's why I'd merge instead
10:23 magey reason i wanted to keep it rebased is so that we can easily see what changes we've made
10:24 JeroenT joined #git
10:25 nettoweb joined #git
10:25 mischat joined #git
10:26 djb-irc joined #git
10:28 mischat_ joined #git
10:29 Junior joined #git
10:31 vapula joined #git
10:33 bruce_lee joined #git
10:33 bruce_lee joined #git
10:34 ronskisika joined #git
10:34 nostrora joined #git
10:35 cyan__ joined #git
10:35 JeroenT joined #git
10:37 cyan__ joined #git
10:37 mischat joined #git
10:39 cyan__ joined #git
10:39 eduardas_m joined #git
10:41 mischat_ joined #git
10:44 mahendratech joined #git
10:55 wiesel joined #git
10:55 Darcidride joined #git
10:57 LBV_User joined #git
10:58 arnoldoree joined #git
10:58 pbrewczynski joined #git
11:00 acetakwas joined #git
11:01 parasite_ joined #git
11:04 JeroenT joined #git
11:04 arnoldoree joined #git
11:04 aspiers joined #git
11:07 Darcidride joined #git
11:07 acetakwas joined #git
11:10 Garen_ joined #git
11:11 m4sk1n joined #git
11:13 ferr1 joined #git
11:14 ignatiz- joined #git
11:14 mda1 joined #git
11:15 ignatiz- hello, I want to list all branches / tags for a remote repo (ssh://git.company.com) but it requires credentials, is this possible?
11:15 osse by supplying valid credentials, yes
11:16 pur3eval joined #git
11:20 dsdeiz joined #git
11:20 ignatiz- osse: this will be ran in an automated build, can I add credentials with ls-remote?
11:21 pity joined #git
11:21 osse ignatiz-: man gitcredential
11:22 osse ignatiz-: man gitcredentials
11:22 gitinfo ignatiz-: the gitcredentials manpage is available at http://jk.gs/gitcredentials.html
11:22 osse I am not 100% sure, but I assume so
11:23 MineCoins joined #git
11:23 ignatiz- osse: thanks :)
11:25 sword` joined #git
11:26 sawk joined #git
11:27 Tobbi joined #git
11:29 mischat joined #git
11:33 johnnyfive joined #git
11:33 GodGinrai joined #git
11:33 wagle joined #git
11:37 jfr joined #git
11:42 mischat_ joined #git
11:43 JeroenT joined #git
11:45 MineCoins joined #git
11:46 causasui joined #git
11:47 diogenese joined #git
11:48 dendazen joined #git
11:50 stamina joined #git
11:50 a_thakur joined #git
11:51 seni joined #git
11:52 dumacdev joined #git
11:52 ertesx joined #git
11:59 p4trix joined #git
12:00 arnoldoree joined #git
12:01 atcagn joined #git
12:01 Dev0n hey, would you only --squash into the master branch or should you be squashing your commits all the way from your feature branch to dev, staging then master as well?
12:04 chll_ joined #git
12:04 kurkale6ka joined #git
12:09 Vampire0 Dev0n, I personally would never use a squash merge. I would either do a real merge, or a fast-forward merge, depending on whether I want linear history or want to see the parallel development
12:10 Dev0n oh, but you end up taking a lot of "noise" from feature branches into master though?
12:10 Dev0n Vampire0
12:11 CheckDavid joined #git
12:14 netj joined #git
12:15 Vampire0 Dev0n, depends on that you consider noise
12:17 Dev0n Well, say a feature ends up having over 100 commits, of which only a few commits without actually made an impact in the feature, without squashing, you are sending all those commits up to master?
12:17 Dev0n few commits which*
12:20 PHPanos joined #git
12:21 Tobbi joined #git
12:22 Vampire0 Dev0n, if there are only 5 commits which made an impact in the feature, what do the other 95 commits do?
12:23 Dev0n was a bad example, maybe it was a junior developer that was going back and forth between changes, the point is that should those micro changes ever hit a master branch?
12:23 lindenle joined #git
12:25 chipotle joined #git
12:25 aw1 joined #git
12:27 aard_ joined #git
12:30 rorro joined #git
12:33 Dumblez0r joined #git
12:40 pl joined #git
12:40 Bombe Dev0n, there’s also a middle ground: rework those 100 commits into 5 nice-looking commits (with an interactive rebase), then use a real merge for those.
12:40 Bombe That’
12:40 Bombe That is what I usually do.
12:40 Dev0n yea true, so I get the impression that squashes are not in favour?
12:43 mischat joined #git
12:43 Bombe Dev0n, no… most of what you do is a logical progression of changes; why would you want to throw away that information?
12:43 cyphase_eviltwin joined #git
12:43 Dev0n valid point, thanks for clearing that up Bombe  and Vampire0
12:46 Darren_ joined #git
12:47 Vinnie_win_v joined #git
12:47 DolphinDream joined #git
12:50 marcogmonteiro joined #git
12:51 red82 joined #git
12:52 Dumblez0r joined #git
12:52 Gsham joined #git
12:53 _Cyclone_ joined #git
12:53 cdown joined #git
12:55 Dumblez0r joined #git
12:56 kettlewell joined #git
12:57 regedit joined #git
12:59 ojdo joined #git
12:59 eroux joined #git
12:59 synthroid joined #git
13:00 JanC_ joined #git
13:01 JeroenT joined #git
13:02 JeroenT_ joined #git
13:02 philipsd6 joined #git
13:04 Bombe If you do 100 commits but the real change is only a single logical unit that is also very small it would be okay to squash them all down to a single commit.
13:04 eroux joined #git
13:05 pur3eval joined #git
13:05 MattMaker joined #git
13:06 axisys joined #git
13:06 GodGinrai joined #git
13:07 boombatower joined #git
13:07 matoro joined #git
13:07 JeroenT joined #git
13:07 HD|Laptop joined #git
13:08 nedbat joined #git
13:08 nedbat joined #git
13:08 HD|Laptop joined #git
13:08 saintaquinas[m] joined #git
13:08 HD|Laptop Hi
13:08 dsop joined #git
13:08 Learath2 hi
13:08 altendky joined #git
13:08 HD|Laptop I have two machines, one running git 2.1.4 and one 2.11.0
13:08 HD|Laptop git clone [git@git.company:16122]:user/repo.git
13:09 troulouliou_div2 joined #git
13:09 HD|Laptop this command works on 2.11, but on 2.1.4 it fails with Could not resolve hostname
13:09 Learath2 uhm you sure it can resolve the hostname?
13:10 sborza joined #git
13:10 NvpkD1y7Ez joined #git
13:10 nellicus_ joined #git
13:10 PCatinean_ joined #git
13:10 selckin or doesn't support [] syntax yet
13:11 Learath2 or that but wanted to make sure he can actually resolve it before i go checking :P
13:11 selckin configure the port in  ~/.ssh/config
13:11 HD|Laptop Learath2: ping does work
13:11 selckin 3 years old
13:11 Tsutsukakushi joined #git
13:11 Learath2 yeah just specify the port in config
13:11 HD|Laptop selckin: no way, the 2.1.4 git is on a jenkins docker image...
13:11 HD|Laptop can't customize the port
13:12 selckin dec 17 2014
13:12 HD|Laptop okay I could by persisting /root/.ssh, but I really dont want to go down that route
13:12 selckin try if it works with like GIT_SSH='ssh -p 16..'
13:13 selckin then you're sure of the reason and can find optimal solution
13:13 Learath2 ssh://git@git.company:16122/user/repo.git maybe
13:13 nukeu666 joined #git
13:14 DolphinDream joined #git
13:15 patrickr[m] joined #git
13:15 sQVe joined #git
13:15 GodGinra1 joined #git
13:15 Sheogorath[m] joined #git
13:16 JeroenT joined #git
13:18 M-meznak joined #git
13:19 HD|Laptop Hmm the syntax without the bracket seems to halfway work, the problem is that somehow jenkins doesn't supply the ssh with the correct credentials
13:19 srkrishna[m] joined #git
13:20 musicmatze[m] joined #git
13:20 ecuanaso joined #git
13:21 Learath2 HD|Laptop: did you try ssh:// ?
13:21 HardlySeen joined #git
13:21 Raed joined #git
13:21 mischat joined #git
13:23 Tobbi joined #git
13:24 HD|Laptop Learath2: the problem isn't in git any more (I hope)
13:24 Learath2 well hf working with the cancer that is jenkins
13:26 aw1 joined #git
13:28 mSSM joined #git
13:28 serialoverflow joined #git
13:29 nellicus_ joined #git
13:29 geoid_ joined #git
13:29 HD|Laptop Hmm
13:29 HD|Laptop Was GIT_SSH available in 2.1.4?
13:33 sunri5e joined #git
13:33 PCatinean_ joined #git
13:35 liberpater joined #git
13:35 MattMaker joined #git
13:36 perlpilot joined #git
13:38 djb-irc joined #git
13:38 Impaloo joined #git
13:39 rush joined #git
13:39 Oatmeal joined #git
13:40 jeffreylevesque joined #git
13:41 e14 joined #git
13:42 serialoverflow joined #git
13:44 _ikke_ HD|Laptop: I believe so
13:45 Darcidride joined #git
13:45 _ikke_ Hmm
13:45 osse GIT_SSH_COMMAND is the recent one
13:45 dendazen joined #git
13:46 solenodic joined #git
13:46 _ikke_ right
13:46 _ikke_ I have references going back to 1.8.1 for GIT_SSH
13:47 mischat joined #git
13:51 Gsham joined #git
13:52 Snugglebash joined #git
13:52 Vaelatern joined #git
13:52 Tobbi joined #git
13:54 MineCoins joined #git
13:56 stamina joined #git
13:57 djb-irc joined #git
13:57 PHPanos joined #git
13:58 brent__ joined #git
13:59 nowhere_man joined #git
14:00 ThomasLocke_ joined #git
14:01 Dumblez0r joined #git
14:02 jimi_ joined #git
14:03 cdg joined #git
14:03 Deb_ joined #git
14:04 robogoat joined #git
14:04 HD|Laptop Argh. Seems like I found the problem. Specifying git@git.company:16122:user/repo.git makes Jenkins invoke "ssh -i /tmp/ssh1741958454019629805key -l git -o StrictHostKeyChecking=no git@git.company.de git-upload-pack '16122:user/repo.git'"
14:05 thebishop joined #git
14:05 dermoth_ joined #git
14:05 tvw joined #git
14:07 TheMontyChrist joined #git
14:08 TheMontyChrist joined #git
14:08 TheMontyChrist joined #git
14:08 TheMontyChrist my repository was cloned from my thumb drive.  There's another machine with the same repo that's a little more up to date how do I pull from it?
14:08 TheMontyChrist I gues generally, how do I pull from an arbitrary repo?
14:08 _ikke_ add it as a remote
14:08 osse TheMontyChrist: git pull <URl>
14:08 TheMontyChrist \o/
14:08 TheMontyChrist thanks
14:09 elect joined #git
14:10 mehola joined #git
14:10 grawity see also, `man git remote`
14:10 gitinfo the git-remote manpage is available at http://jk.gs/git-remote.html
14:11 bannakaffalatta joined #git
14:15 salamanderrake joined #git
14:15 Tobbi joined #git
14:15 theoceaniscool joined #git
14:16 sts_ joined #git
14:16 sts_ left #git
14:16 sts_ joined #git
14:16 sts_ left #git
14:17 Hudu joined #git
14:17 sts joined #git
14:17 sts Hello folks. I'd like to squash all commits made to a branch before I create a pull request.
14:18 sts The problem is, that during development I sometimes already merged commits from master in my feature branch.
14:18 sts My squashed commit should only contain all commits which happened in the feature branch.
14:19 sts I read I should rebase my feature branch to the newest commit of master.
14:19 sts But then, how do I squash all of them?
14:20 Shashikant86 joined #git
14:20 Ryanar joined #git
14:21 perlpilot rebase -i and reorder the commits so that all of non-merge commits are next to one another, then squash them all
14:22 jhasse_ left #git
14:24 sts perlpilot: rebase -i gives me a 'noop' and doesn't list any commits as usual
14:25 tinanoo joined #git
14:27 JeroenT joined #git
14:29 theoceaniscool joined #git
14:29 ShekharReddy joined #git
14:33 thiago joined #git
14:36 OMSQ joined #git
14:37 e14 joined #git
14:37 tinanoo joined #git
14:38 YuGiOhJCJ joined #git
14:38 raijin joined #git
14:38 hashpuppy joined #git
14:39 dviola joined #git
14:39 ash_workz joined #git
14:42 ikelso joined #git
14:42 Snugglebash joined #git
14:43 Dougie187 joined #git
14:44 ShalokShalom joined #git
14:44 red82 joined #git
14:45 Takumo joined #git
14:45 Takumo joined #git
14:46 JeroenT joined #git
14:47 aw1 joined #git
14:47 griffindy joined #git
14:48 neunon joined #git
14:48 nickabbey joined #git
14:50 dunpeal joined #git
14:51 jknetl1 joined #git
14:52 dunpeal Hi. I have to update the author (both name and email) for an entire repo. I.E. any commit authored by "John Smith <john.smith@yahoo.com>" should be attributed to "Jane Doe <jane.doe@gmail.com>". What's the best way?
14:53 OMGOMG look into filter-branch
14:53 y_sekino joined #git
14:53 canton7 there's an example in man git filter-branch to do just that
14:53 gitinfo the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html
14:53 pur3eval joined #git
14:54 ThomasLocke_ joined #git
14:54 ThomasLocke_ joined #git
14:54 jknetl1 joined #git
14:54 Es0teric joined #git
14:54 rivarun joined #git
14:55 rewt dunpeal, if you do that, you'll be altering history, which means everyone will have to reset all their local branches to match the new ones
14:55 rewt in their own local repos
14:56 amagawdd joined #git
14:56 texasmynsted if I git push origin --delete <branch_name> that only deletes the remote branch and not the local, right?
14:56 rewt texasmynsted, that sounds about right
14:57 rewt you could make another local branch at the same point with a different name and try it out, and see if the local branch gets deleted too
14:57 rewt if it does, just recreate it
14:57 dunpeal rewt: yeah I know.
14:57 dunpeal canton7, OMGOMG: thanks.
14:58 kpease_ joined #git
14:59 A124 joined #git
14:59 A124 Hey. How does one get up to date with remote repo without merging anything?
15:00 perlpilot A124: git fetch
15:01 texasmynsted k
15:01 _ikke_ git pull --rebase
15:02 dec0n joined #git
15:03 A124 Thanks.
15:03 dave0x6d joined #git
15:04 AaronMT joined #git
15:06 seni joined #git
15:07 Darcidride joined #git
15:08 Tobbi joined #git
15:09 pity joined #git
15:09 Gsham joined #git
15:11 cyberik joined #git
15:11 Darcidride joined #git
15:14 LeBlaaanc joined #git
15:16 stoopkid joined #git
15:16 nostrora joined #git
15:18 workisfun joined #git
15:18 dunpeal Why does Git care so much that authors have emails?
15:18 workisfun hi guys, what is the problem with no newline at the end of a file. (github diffs show this as a red stop sign at the bottom of the file's diff
15:19 canton7 dunpeal, so you can contact them and tell them their code's rubbish? :P
15:19 selckin because of its origins from lkml
15:19 canton7 workisfun, no problem
15:19 canton7 workisfun, it's just a way of indicating that there's no trailing newline in a clear way
15:20 selckin workisfun: trailing newline is nice, becuase if you add a new line it doesn't have to change the last line and mess up diffs/merges
15:20 synthroid joined #git
15:20 workisfun selckin:right, ok so it messes up diffs/merges
15:21 Darcidride joined #git
15:25 Tobbi joined #git
15:25 noecc joined #git
15:26 rorro joined #git
15:26 Hedgework dunpeal: it's a UUID, basically
15:27 overlord_tm joined #git
15:28 _ikke_ github and the like can use it to match the author to an account
15:29 kbaegis joined #git
15:30 kbaegis Hi all. What's the best way to grab month-by-month stats for a project on github?
15:31 Jackneill_ joined #git
15:31 Snugglebash joined #git
15:32 publio joined #git
15:32 jedahan joined #git
15:33 moritz kbaegis: do you want stats from github, or stats from a git repo you cloned from github?
15:34 kbaegis moritz: I actually want a csv output for the openstack projects in aggregate, month by month
15:34 kbaegis total number of commits
15:34 razi joined #git
15:35 rwb joined #git
15:35 kbaegis Github produces graphs, but I have no clean way of retrieving that data as far as I can tell.
15:36 moritz kbaegis: try asking in #github if there's an API for that
15:36 sbulage joined #git
15:36 MattMaker joined #git
15:37 kbaegis thanks
15:38 Tobbi joined #git
15:38 PHPanos joined #git
15:41 solenodic joined #git
15:41 ocbtec joined #git
15:41 jbitdrop joined #git
15:44 Gsham joined #git
15:46 nckpz joined #git
15:46 jstimm joined #git
15:49 Es0teric joined #git
15:49 Atemu joined #git
15:51 seni joined #git
15:52 ResidentBiscuit joined #git
15:53 mda1 joined #git
15:54 pur3eval joined #git
15:55 WayToDoor joined #git
15:55 AlexMax left #git
15:56 TankBo joined #git
15:57 Tobbi joined #git
15:58 workisfun left #git
16:06 jstimm joined #git
16:07 Shashikant86 joined #git
16:07 sborza joined #git
16:08 tvw joined #git
16:11 mehola joined #git
16:12 Tobbi joined #git
16:13 Ryanar joined #git
16:14 jknetl joined #git
16:14 jknetl1 joined #git
16:15 Wind0r joined #git
16:16 MeltedLux joined #git
16:16 tcorneli joined #git
16:17 mischat joined #git
16:18 RxMcDonald left #git
16:19 shinnya joined #git
16:20 jstimm joined #git
16:21 netj joined #git
16:22 Lunatrius joined #git
16:24 seni joined #git
16:24 a_thakur joined #git
16:25 Endarked joined #git
16:26 str joined #git
16:27 overlord_tm joined #git
16:28 rafalcpp joined #git
16:28 tvw joined #git
16:29 re1 joined #git
16:29 peeps[lappy] joined #git
16:29 Snugglebash joined #git
16:31 dreiss joined #git
16:35 DolphinDream joined #git
16:36 pur3eval joined #git
16:37 matsaman joined #git
16:37 Murii joined #git
16:40 texasmynsted I see 'git branch —edit-decription'  I used this, but now how may I see the description for a branch?
16:40 vbsales joined #git
16:42 mischat joined #git
16:42 bongjovi joined #git
16:42 texasmynsted hmm maybe the only way is to do this for a particular branch by 'git branch —edit-description' again
16:47 Aleric utils>git commit -a - m 'Use CW_NON_THREADED/CW_THREADED conditionals.'
16:47 Aleric fatal: Paths with -a does not make sense.
16:47 jstimm joined #git
16:47 thiago joined #git
16:47 tlaxkit joined #git
16:47 Aleric What am I missing here?
16:47 Aleric Oh, I see it...
16:47 Aleric nm
16:54 lss8 joined #git
16:55 eroux joined #git
17:00 WayToDoor joined #git
17:00 nowhere_man joined #git
17:01 ocbtec joined #git
17:01 mohabaks joined #git
17:02 theoceaniscool joined #git
17:03 Cabanossi joined #git
17:04 relipse joined #git
17:04 relipse i made one good commit to a branch and about 15 debug commits, how can I go back to the one good commit
17:05 relipse and blow away all the debug commits
17:06 thierryp joined #git
17:06 tcorneli joined #git
17:07 madewokherd joined #git
17:08 TomyLobo joined #git
17:08 tang^ joined #git
17:12 osse git reset --hard HEAD~15
17:12 relipse too dangerous
17:12 relipse i used 15 as ane xample there could be more or less
17:13 osse so did I
17:13 ThiefMaster git rebase -i HEAD~n
17:13 ThiefMaster with n being the number of commits you want to go back. then delete whatever you don't need anymore in the editor
17:13 relipse yeah thats a good idea
17:13 relipse thank you both
17:13 dumacdev joined #git
17:14 Aleric I was going to say that, then Thief stole my idea.
17:14 xall joined #git
17:14 osse or git log, find sha of good commit, git reset --hard commit
17:14 osse if you do git reset -i HEAD~n then make sure to pick n large enough; if you want to remove all commit sin the editor nothing will happen
17:15 e14 joined #git
17:15 jagob joined #git
17:15 Puffball joined #git
17:17 nilg joined #git
17:18 rafalcpp joined #git
17:19 Puffball joined #git
17:19 Es0teric joined #git
17:20 mischat_ joined #git
17:20 chardan joined #git
17:21 nickabbey joined #git
17:21 mischat__ joined #git
17:21 Gsham joined #git
17:21 mischat__ joined #git
17:23 cdown_ joined #git
17:23 jccn joined #git
17:24 svm_invictvs joined #git
17:25 ianmethyst joined #git
17:26 a_thakur joined #git
17:27 a_thakur joined #git
17:27 chardan joined #git
17:28 mischat joined #git
17:29 mischat_ joined #git
17:31 tyreld joined #git
17:32 jstimm joined #git
17:34 nowhere_man joined #git
17:35 nowhere_man joined #git
17:36 rivarun joined #git
17:37 MattMaker joined #git
17:39 g105b joined #git
17:40 harry1 joined #git
17:41 Guest85 joined #git
17:43 peepsalot joined #git
17:43 cdg joined #git
17:44 dreiss joined #git
17:45 stux|RC-only joined #git
17:45 a_thakur joined #git
17:46 shgysk8zer0_ joined #git
17:51 MineCoins joined #git
17:51 israelzuniga joined #git
17:51 lindenle joined #git
17:53 malt3 joined #git
17:54 brent__ joined #git
17:55 PettanShoutaKun joined #git
17:55 PettanShoutaKun I'm on my local branch
17:55 PettanShoutaKun I have a bunch of changed files.... but I realized a few of them were wrong and rechanged them
17:56 PettanShoutaKun I want to sort of complete undo my last commit and then repush all my changes as a new single commit. How would I do this?
17:57 PettanShoutaKun I'd like to not have to reset --hard because then I lose my changes
17:57 geoid_ joined #git
17:58 PettanShoutaKun I really want to like take all my files as is and push them to HEAD~ ? I think
17:59 nickabbey joined #git
18:01 vuoto joined #git
18:02 texasmynsted So for branch descriptions, the only way to see them is to edit them or have them automatically show when merging or something?
18:04 jaafar joined #git
18:05 xall joined #git
18:06 Es0teric joined #git
18:06 jstimm joined #git
18:07 nedbat texasmynsted: i hadn't heard of them before: the help files seem to indicate that they are used for the email workflow mostly.
18:08 texasmynsted hm
18:08 geoid_ joined #git
18:09 kbaegis joined #git
18:09 * texasmynsted shrug
18:09 texasmynsted well I will use it and see how it goes
18:09 Tobbi joined #git
18:09 duderonomy joined #git
18:10 pur3eval joined #git
18:10 nowhere_man joined #git
18:10 a_thakur joined #git
18:13 rorro joined #git
18:14 mehola joined #git
18:14 overyander joined #git
18:15 nowhere_man joined #git
18:16 thierryp joined #git
18:16 govg joined #git
18:17 fuchstronaut I have messed up big time and now I need to take added files from a commit ontop of another branch and commit them again, cleanly, on my current (new) branch
18:18 chipotle joined #git
18:18 mikecmpbll joined #git
18:18 solenodic joined #git
18:19 jknetl joined #git
18:20 rwp fuchstronaut, Have those bad commits been published? Or are they just in your own private local repository?
18:20 fuchstronaut they are just locally
18:20 rwp Whew! Then you are in luck and all is easily accomplished.
18:20 fuchstronaut good to hear
18:21 hahuang61 joined #git
18:21 rwp From what you said so far it actually seems very straight forward. Just pull those commits from the old branch to the new branch. What are the details making that difficult?
18:22 fuchstronaut I have a couple of branches where I really messed up, merged things, rebased things and cherry picked. It is a huge mess
18:22 fuchstronaut so now I made a new branch on top of master, commit a couple of things
18:22 fuchstronaut created a new branch on top of that
18:22 fuchstronaut cherry picked two commits
18:23 fuchstronaut and now I want changes that are stored in a commit ontop of one of my messy branches
18:23 fuchstronaut so I can commit them step by step
18:23 fuchstronaut its a commit where I added 4 files
18:23 fuchstronaut I want to add them file by file
18:23 jnavila joined #git
18:23 fuchstronaut after that just delete the messy branches and push the clean version
18:23 fuchstronaut so everyone thinks I know what I am doing
18:24 yottabyte joined #git
18:24 rwp That is actually one of the best features of git that *I* like best. The ability to reshape my commits before publish so that I look better than I am. :-)
18:24 yottabyte when I create a branch, it's off whatever branch I'm currently in and matches that with any local uncommitted/unpushed changes I have, right?
18:24 Noldorin joined #git
18:25 rwp However I think this is just something you need to work through. I am not sure how anyone other than you can do the work or help out.
18:25 rwp Are you familiar with rebase -i which I think would be useful to you here? !interactive_rebase
18:25 gitinfo Interactive rebase sounds similar to rebase but has completely different abilities. It can do this to commits: change the order, squash some of them together, remove some, add random existing commits from other branches, split them, and more... it's very powerful. Documentation is in the section "Interactive Rebase" in 'man git-rebase'.
18:25 fuchstronaut rwp: my strategy would be: * checkout messy branch, reset last commit, checkout clean branch, commit stuff
18:26 rwp yottabyte, yes
18:26 fuchstronaut Or: cherry pick messy commit, reset, commit
18:26 rwp yottabyte, Unless you specify a specific ref to base the new branch from explicitly.
18:27 rwp fuchstronaut, I think my best advice would be to read about interactive rebase from man git-rebase and then play a little. Very useful for doing the type of cleanup you are needing.
18:27 gitinfo fuchstronaut: the git-rebase manpage is available at http://jk.gs/git-rebase.html
18:27 nowhereman joined #git
18:27 rwp Then come back and ask further specific questions when you run into them.
18:28 adino joined #git
18:28 fuchstronaut rwp: I fiddled with rebase a little, though I don't know how to do thinks like pull in commits from other branches and undo them
18:28 yottabyte rwp: how do I explicitly base the new branch?
18:28 cagmz joined #git
18:28 fuchstronaut but right now I just can't read up
18:28 fuchstronaut I am working for too long and I fear I will screw up worse
18:28 fuchstronaut Would my strategies work?
18:28 matoro joined #git
18:28 Es0teric joined #git
18:29 rwp The typical action would be to check out the branch you want to modify. Then rebase it against another. Example "git checkout mytopic;git rebase master"
18:30 Es0teric_ joined #git
18:30 leftriver59 joined #git
18:30 rwp And interactive rebasing allows you to "squash" commits that you choose, or edit commits, or skip commits, or many other very useful things.
18:30 synthroid joined #git
18:30 fuchstronaut I get that, rebase is the right way to do this thing, but right now I just need that ugly quick fix
18:31 fuchstronaut cherry pick and reset wont do any harm, right?
18:31 rwp You only modify the branch you are on.  Not other branches.  So your changes on other branches are always safe.
18:31 rwp fuchstronaut, To be completely safe you should always make a !backup
18:31 gitinfo fuchstronaut: Worried about your data while trying stuff out in your repo? The repository in its entirety lives inside the .git directory in the root of your work tree so to backup everything `cp -a path/to/workdir path/to/backup` or equivalent will suffice as long as the repo is not modified during backup. See also http://sethrobertson.github.c​om/GitBestPractices/#backups
18:32 leftriver59 hi, can anyone help with setting git using symlinks on windows for dir-diff? i have enabled everything but it still does not seem to be working
18:32 rwp fuchstronaut, Sorry but you have a task in front of you that is tedious to accomplish. But none of us can see your work. Therefore we can't help you beyond cheering you on at this point.
18:33 rwp fuchstronaut, What you are asking is definitely doable. You just need to do it.
18:33 fuchstronaut rwp: I know, thank you for your help
18:34 freimatz joined #git
18:34 rwp I know it seems like you have no time to learn the entire flow right at this moment. We are often in the time squash. But sometimes one just needs to take the time to sharpen the axe. It pays off in the long run.
18:35 rwp And the only other possibility is to find a buddy that you trust with more knowledge than you who will mentor you locally. Someone who can look at your screen and see what you see. And can help you along.
18:36 rwp leftriver59, Not a clue about windows. Sorry. Good luck!
18:36 fuchstronaut rwp: you are right, but sometimes there really IS NO time
18:36 grayjoc joined #git
18:36 fuchstronaut so i will take the wooden hammer and just cherry pick and git reset my way through it
18:37 fuchstronaut and rename branches after that
18:37 fuchstronaut :p
18:37 mark4 joined #git
18:37 peepsalot joined #git
18:37 barteks2x joined #git
18:37 rwp Good luck! If you have specific questions please ask them.
18:37 brent__ joined #git
18:38 mark4 if i know an exact commit ID how do i determine the author of that commit?
18:38 rwp git log ??
18:39 rwp git log --pretty=full ??
18:39 fuchstronaut git log -n1 <sha1>
18:39 mark4 aha ty :)
18:40 _ikke_ git show --format="%an" <commit>
18:40 chardan joined #git
18:42 mark4 left #git
18:44 Hobbyboy joined #git
18:44 dreiss joined #git
18:45 daelras joined #git
18:45 bannakaffalatta joined #git
18:46 pity joined #git
18:47 synthmeat joined #git
18:47 BackEndCoder joined #git
18:48 milki joined #git
18:49 satifant joined #git
18:50 redhedded1 joined #git
18:53 Puffball joined #git
18:54 jedahan joined #git
18:54 emg joined #git
18:57 Mikerhinos joined #git
18:57 nowhereman joined #git
18:57 matoro joined #git
18:58 jedahan joined #git
18:59 jknetl joined #git
18:59 manuelschneid3r joined #git
18:59 MonsterKiller Hey, wondered if someone could help. I'm stuck in a place where I cant push local changes because theres a pull need doing but I cant pull because a local file would be overwritten, however I dont care if it overwrites that local file
18:59 manuelschneid3r joined #git
19:00 Balliad joined #git
19:01 Vampire0 MonsterKiller, that it overwrites the local file is because you have uncommitted changes. Revert those changes by `git checkout -- file` and then do the pull again
19:01 spacelord_ joined #git
19:01 Atm0spher1c joined #git
19:02 xall_ joined #git
19:02 e14 joined #git
19:04 cdown joined #git
19:06 nowhereman joined #git
19:08 MonsterKiller Vampire0, thanks. I've never used that command before but its coming back with "error: pathspec '<path_to_file>' did not match any file(s) known to git."
19:08 manuelschneid3r joined #git
19:08 manuelschneid3r joined #git
19:09 byte512 joined #git
19:09 Gsham joined #git
19:09 Sasazuka joined #git
19:11 brent__ joined #git
19:12 fission6 joined #git
19:12 solenodic joined #git
19:13 elesmod joined #git
19:13 nidr0x joined #git
19:13 cdg joined #git
19:14 xissburg joined #git
19:15 madduck joined #git
19:15 elesmod hello, my "git log" doesn't work; issuing that command doesn't produce the list of commits, but a message "Value is required after -x (--tabs)", I couldn't find solution to this problem on the internet
19:16 _ikke_ What is the complete command you are executing?
19:16 bket joined #git
19:17 elesmod just git log
19:17 madduck joined #git
19:17 _ikke_ did you alias git log?
19:17 MrMojito joined #git
19:17 _ikke_ what does "type git" return?
19:17 PettanShouta-Kun joined #git
19:18 _ikke_ http://octave.1599824.n4.nabble.com/octave​-told-quot-Value-is-required-after-x-quot-​when-I-typed-some-commands-td3942868.html
19:18 _ikke_ Do you use octave?
19:18 elesmod i've thought about it too, but no; i've even checked on another user which doesn't have any aliases set
19:18 jimi_ joined #git
19:19 elesmod type git: [after translation to english]: path to git is remembered (/usr/bit/git)
19:19 elesmod checking the link, don't know what octave is yet
19:19 cdg_ joined #git
19:19 _ikke_ me neither
19:19 Sasazuka joined #git
19:19 ResidentBiscuit joined #git
19:19 _ikke_ but google returns results for octave for that error message
19:20 _ikke_ https://www.gnu.org/software/octave/
19:20 elesmod doesn't seem like octave is installed on my system
19:21 _ikke_ elesmod: export | grep LESS
19:21 Es0teric_ joined #git
19:21 inflames joined #git
19:21 elesmod other git commands work; i've just initialized with "git init"; i've added a file "git add file", i've commited the change with "git commit", only git log doesn't work
19:22 _ikke_ did you run that command?
19:22 elesmod export | grep LESS: declare -x LESS="-erx"; declare -x LESSOPEN="|lesspipe.sh %s"
19:23 xissburg joined #git
19:23 _ikke_ That x in there requires a parameter\
19:23 cdg joined #git
19:23 _ikke_ git is spawning a pager, and that pager is returning that error
19:25 elesmod oh! i've recently added "export LESS="-erx" so that less would work with colors instead of showing "ESC[
19:25 _ikke_ only -r is needed for that
19:26 djb-irc joined #git
19:27 cdg__ joined #git
19:27 elesmod i'm amazed how quickly you've found a solution that wasn't really related to git; thank you very much _ikke_
19:28 sathed joined #git
19:28 Sasazuka joined #git
19:28 pwnz0r joined #git
19:31 _ikke_ elesmod: googling around let me on the right track
19:31 _ikke_ https://lists.gnu.org/archive/html​/help-octave/2011-10/msg00571.html
19:31 _ikke_ This message in partical
19:31 _ikke_ particular
19:31 str joined #git
19:32 wilbert joined #git
19:34 githubmgw joined #git
19:36 Macaveli joined #git
19:38 MattMaker joined #git
19:39 Puffball joined #git
19:40 Ryanar joined #git
19:41 adino joined #git
19:42 jnavila joined #git
19:43 mSSM joined #git
19:47 Macaveli joined #git
19:50 Vampire0 MonsterKiller, well, then it is not a file with local changes, but an untracked file just as Git told you. Simply delete it and go on
19:52 hahuang65 joined #git
19:55 Puffball joined #git
19:55 Ryanar joined #git
19:55 aidalgol joined #git
19:56 MineCoins joined #git
19:57 stux|RC-only joined #git
20:00 jsolano joined #git
20:00 fuzzybear3965 This happened today at github.com:jekyl/jekyll: https://github.com/jekyll/jekyll/issues/5850 .
20:00 brent__ joined #git
20:00 fuzzybear3965 How do you guys make sure that your issue/PR isn't duplicated?
20:00 thiago we don't
20:00 fuzzybear3965 I don't like wasting people's (even my) time.
20:01 thiago users can'report the same thing
20:01 fuzzybear3965 thiago, so these things happen?
20:01 thiago yes
20:01 fuzzybear3965 This is a "s*** happens" moment?
20:01 thiago bugzilla usually does a search as you type the summary and shows similar issues
20:01 thiago then users might see "oh, it's reported"
20:04 dumacdev joined #git
20:04 durham joined #git
20:05 cdg_ joined #git
20:05 sweatsuit joined #git
20:07 githubmgw1 joined #git
20:07 mischat joined #git
20:07 annoymouse joined #git
20:08 kyan joined #git
20:09 porsche944 joined #git
20:10 kbaegis joined #git
20:10 codemann joined #git
20:12 HandheldPenguin joined #git
20:12 Macaveli_ joined #git
20:13 jstein_ joined #git
20:13 joshszep joined #git
20:13 DrSlony Hey, I need to "rename" our branch 'master' to 'gtk2'. This is a public repo. All the devs are aware of the upcoming change. What is the best way to go about it? There is (1) this http://stackoverflow.com/a/3790682 and (2) this http://www.benjaminlhaas.com/blog/loc​ally-and-remotely-renaming-branch-git
20:13 harry1 joined #git
20:15 fission6 joined #git
20:15 druonysus_ joined #git
20:16 durham joined #git
20:16 zapb_ joined #git
20:17 YuGiOhJCJ joined #git
20:17 Gsham joined #git
20:18 GodGinrai joined #git
20:18 Puffball joined #git
20:18 jsolano joined #git
20:19 thiago DrSlony: will there be a master branch after this?
20:19 solenodic joined #git
20:19 Ryanar joined #git
20:20 kadoban https://www.theregister.co.uk​/2017/02/01/gitlab_data_loss/  :-/
20:20 DrSlony thiago depends on the risks. I could go without a branch called "master". Would things break if someone then made a branch called "master"?
20:20 MattMaker joined #git
20:21 tinanoo1 joined #git
20:21 _ADN_ joined #git
20:22 durham joined #git
20:22 Es0teric_ joined #git
20:22 DrSlony thiago there are several open pull requests for merging things into master
20:23 J32RY joined #git
20:24 jnavila joined #git
20:24 matoro joined #git
20:26 Puffball joined #git
20:26 XVar joined #git
20:26 durham joined #git
20:26 Singmyr joined #git
20:27 J32RY joined #git
20:27 damccull joined #git
20:27 thierryp joined #git
20:27 thierryp joined #git
20:28 thiago DrSlony: my recomendation: don't delete it
20:28 thiago just move on with your new development (which I guess is a Gtk3 port)
20:28 thiago so: git branch gtk2; git push -u origin gtk2
20:28 developernotes joined #git
20:28 DrSlony thiago I'm thining that too, I'm too nervous deleting it and hoping github figures stuff out and the PRs dont die
20:29 plushy joined #git
20:29 thiago so don't delete it
20:29 thiago do as I suggested
20:31 invisbl joined #git
20:31 durham joined #git
20:31 invisbl joined #git
20:32 Cassiopaya joined #git
20:32 DrSlony thiago do I understand correctly that after doing what you wrote, gtk2 would share the same latest commit as master?
20:32 _ADN_ joined #git
20:34 plushy Okay, it's time for stupid question: How do I manage multiple users working in the same directory on a server?
20:34 plushy 1) The changes need to be in the same directory on the staging server while the users are working on them before being commited
20:34 plushy 2) Every user should have a seperate index of changes they're about to commit
20:34 plushy Due to technical issues I can't set up multiple staging environments right now nor can I split the repo into 3k+ repositories for each directory
20:34 dunpeal joined #git
20:34 MattMaker joined #git
20:35 bannakaffalatta joined #git
20:35 m0viefreak joined #git
20:35 rwp DrSlony, After a 'git branch gtk2' from master then gtk2 is identical to master. So yes to your question.
20:35 dunpeal Hi!  In a script, I want to check if a branch "foo" exists in a repo, and if so - get its hash. What's the best way?
20:35 adino joined #git
20:35 wilbert joined #git
20:35 developernotes joined #git
20:36 rwp plushy, Sorry. Can't have separate indexes per user. Indexes are a part of the current working copy.
20:36 durham joined #git
20:36 rwp plushy, And what you are asking for is not a target design. You are dancing close to !deploy and should look at that reference.
20:36 gitinfo plushy: Git is not a deployment tool, but you can build one around it (in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/deploy.html
20:37 XVar joined #git
20:37 rwp dunpeal, ?? 'git checkout foo' ??
20:38 plushy Yes, I do know it's not a deploy tool. But it's been in use like one for 8 years now and I've been aware of the problem for 3 months
20:38 yqt joined #git
20:38 dunpeal rwp: that also performs an action.
20:38 dunpeal Which I don't want to perform.
20:38 Puffball joined #git
20:38 rwp dunpeal, git rev-list foo ??
20:39 chardan joined #git
20:39 rwp dunpeal, "git rev-list -n1 foo" seems like what you want.
20:39 overyander joined #git
20:39 dunpeal rwp: or git rev-parse... thanks.
20:40 Gloomy joined #git
20:40 icey joined #git
20:40 icey joined #git
20:40 codemann- joined #git
20:40 plushy Well, in fact the deploy is done using a script fired from gitlab on commit, but I still have a problem with multiple people in the same working dir
20:44 DrSlony rwp so then, in the future, after the PRs are merged into gtk2, master could be deleted and everything would be safe and sound in gtk2, right? not talking about PRs.
20:45 rwp Yes. Because operations on different branches do not affect other branches. Operations taken on gtk2 do not affect master. And the reverse.
20:46 DrSlony ive been using git for a year now but im asking anyway because i dont want to screw this up by some assumption.
20:46 rwp But I don't like the idea of deleting master. Nor renaming it, which drags other things around. I would simply branch as thiago suggested and keep moving there.
20:46 durham joined #git
20:46 DrSlony yes, i like that the most too
20:47 skered left #git
20:47 tvw joined #git
20:47 DrSlony though, considering we must likely wont commit anything on master/gtk2 anymore, we might as well just leave it as 'master' and forget about it.
20:48 DrSlony its just that having it called 'gtk2' would be clear to users who checkout the repo, but then if i create gtk2 and dont delete master, we're back to square one with regard to clarity
20:48 plushy Okay, I'll just go with spliting it by first letter of dirs and praying this minimizes the trouble
20:48 elesmod joined #git
20:50 durham joined #git
20:52 Balliad joined #git
20:55 Naan joined #git
20:56 pity joined #git
20:56 matoro joined #git
20:59 MattMaker joined #git
20:59 solenodic joined #git
21:03 bgerber joined #git
21:04 icey joined #git
21:04 icey joined #git
21:05 a0x10c joined #git
21:06 dunpeal damn Xubuntu is snappy after OS X.
21:06 DrSlony thiago why do you recommend not deleting the original branch?
21:08 brent__ joined #git
21:08 yonatankoren joined #git
21:09 simon816 joined #git
21:09 askb joined #git
21:13 dsdeiz joined #git
21:13 dsdeiz joined #git
21:13 bannakaffalatta joined #git
21:18 Tobbi joined #git
21:20 a_thakur joined #git
21:22 NightMonkey joined #git
21:23 causasui joined #git
21:27 Es0teric_ joined #git
21:27 Ryanar joined #git
21:28 nickabbey joined #git
21:32 DrSlony thiago rwp what problems could I expect if I branch off under a new name and delete the original branch?
21:32 DrSlony please help, I'm having a tough time making a decision
21:32 rwp DrSlony, Why do you want to delete the master branch?
21:33 BlackMaria_netsp joined #git
21:34 anuxivm joined #git
21:34 rwp DrSlony, For that matter why do you want to abandon master and use gtk2?  Let's get back to the root cause of the trouble...
21:34 om_henners joined #git
21:35 rwp Sure master is just a name. But by convention everyone expects there to be a master branch. Why go against convention here?
21:35 balu_23 joined #git
21:35 ikelso joined #git
21:36 unbalanced joined #git
21:36 DrSlony rwp because what happens in master in the future will be very different to what happens in it now
21:36 DrSlony currently master is a development branch using gtk2
21:37 * rwp never knows how long to wait for other people to type. :-)
21:37 DrSlony in the future there will be a branch into which we merge stable release branches and only tag in it, we could call that branch "master" or "builds" or whatever
21:38 rwp Then I think what you want to do is to branch the current master over to gtk2.  Then reset master back to whatever earlier fork point you want to pick. Then move forward from there.
21:40 solenodic joined #git
21:41 Aleric DrSlony: The right thing to do is to branch off a branch named gtk2 and then only commit stuff to that to fix bugs for the gtk2 release, while starting work on master to make a gtk3 branch.
21:41 acetakwas joined #git
21:42 rwp DrSlony, I like Aleric's suggestion quite a bit too. Simpler.
21:43 Aleric to make it*
21:43 DrSlony There already is a branch called gtk3. That will become the main development branch, so it would be nice to rename that one to "development"
21:43 yottabyte left #git
21:43 Aleric You could merge that into master no?
21:44 govg joined #git
21:44 SwiftMatt joined #git
21:44 * rwp thinks there is a lot of reorganization just for naming that isn't doing anything actually useful.
21:44 Guma joined #git
21:45 Electrometro joined #git
21:45 Aleric The usual way projects work with branches is to have 'stable' (old) branches for long time support of old code (security fixes etc), have 'experimental' heavily broken branches that a few people are working on (ie, adding new features) and the master branch where all work comes together is the main synchronization branch for all developers.
21:46 Aleric As all your developers now will start to work on gtk3, they should use master to synchronize their work, and the experimental gtk3 branch now should be merged into master (after forking off a gtk2 "stabel" branch)
21:47 mehola joined #git
21:47 matsaman joined #git
21:47 Aleric Some projects, most notably those who insist on a linear history and have everyone constantly rebase everything before it is added to the HEAD of master have problems here: they can't merge an experimental branch into master. That's why you should use that method with rebasing everything :p
21:48 DrSlony Aleric we have a model, we just need to figure out whether and how to adapt branch names so they're more intuitive.
21:48 Aleric But if you use git the way it was intended, then the gtk3 devs already merged master into gtk3 often, and merging it back into master will be easy.
21:48 DrSlony I merge master into gtk3 regularly
21:49 Aleric ok, so now do it the other way around and announce that people should continue to work with master instead of gtk3.. in fact, you CAN delete gtk3 after merging it into master
21:49 OverCoder joined #git
21:49 OverCoder Hi
21:49 OverCoder I did git reset HEAD^
21:50 OverCoder Now how do I push to remote server or whatever so the commit is also gone there?
21:50 SunOS joined #git
21:51 Aleric OverCoder: you can't really remove the commit from a remote server (without hacking; I think), but you can remove any branch pointers pointing to it, so nobody will ever find it back.
21:51 BlackMaria joined #git
21:51 Aleric If you reset your branch one commit and push -f, (note the -f) then it will be like that...
21:51 * OverCoder tries
21:52 DrSlony Aleric no, master is gtk2 code which will be killed off, gtk3 is gtk3 code. Master is merged into gtk3, but they are still very different in the ui code. But that's besides the point.
21:52 DrSlony My current question is what do I risk by branching off from foo into bar and deleting foo
21:52 OverCoder that worked, <333
21:52 davis joined #git
21:53 SunOS left #git
21:54 davis i have two branches i'm working on. i have currently on branch A. can I add items to branch B without doing a checkout of branch B?
21:54 Aleric DrSlony: very little, a branch is just a tag. If something went wrong you could always recover as long as you know the hash that the branch was pointing too. Although, I'm not sure about githubs PR's... I deleted a branch once that I asked a PR for, I think I had to redo the pull request.
21:56 Aleric davis: I don't think so... but you can switch to branch B, commit, push and go back to A without losing anything if you give the right commands.
21:56 davis Aleric: thanks
21:56 perlpilot davis: why do you not want to checkout B?
21:56 davis i merged some things by hand already
21:57 matoro joined #git
21:57 PHPanos joined #git
21:57 Aleric untracked files wouldn't be touched by checking out B... changed file would just stay changed, stuff in the index can end up reset though (so you have to add them to the index again). Maybe 'git stash' can help some how...
21:57 omarramo joined #git
21:58 Aleric I think a normal 'git stash' does reset the index, quite annoying...
21:58 davis it says your local changes will be overwriten by the checkout
21:59 Aleric The safest bet is to just do a commit on A, then checkout B and then go back to A and recover whatever you want to recover :p
21:59 omarramo hey, I had some files in my repo when I created a .gitignore and applied it, which obviously deleted a bunch of stuff. I did that because the repo size was too big. but now the package file is still too big because it still contains the files that were unwanted. is there some command to shrink it?
21:59 davis if I git stash, I can then recover the stash right?
21:59 serialoverflow joined #git
21:59 Aleric Provided you unstash it when on A again
22:00 mniip joined #git
22:00 Aleric if you try that on B you'll get collisions
22:00 Aleric or might get
22:00 omarramo for reference: my pack file 58 mb, the whole repo is 58.5 mb
22:01 Aleric A+untracked+changed+indexed --> git stash --> A+untracked.  git co B --> B+untracked  ..do stuff.. --> git co A --> A+untracked --> git stash pop --> A+untracked+changes, where the changes are the old changes plus what used to be in the index.
22:02 Aleric Actually, I'm not 100% sure :/
22:02 Es0teric_ joined #git
22:02 Aleric but if your 'git diff --cached' is empty (nothing in the index) then git stash is going to work
22:03 Aleric Maybe stash only puts things NOT in the index in the stash?
22:04 gajus joined #git
22:05 fuzzybear3965 thiago, I guess when you say "bugzilla" you mean "GitHub's issue database", in this context?
22:06 chardan joined #git
22:06 Es0teric_ joined #git
22:06 _28_ria joined #git
22:07 griffindy joined #git
22:08 ikelso joined #git
22:11 rwb joined #git
22:12 chardan joined #git
22:12 bannakaffalatta joined #git
22:13 redhedded1 joined #git
22:14 dendazen joined #git
22:16 thiago fuzzybear3965: no, I mean bugzilla
22:16 disi joined #git
22:16 davis that worked. many thanks.
22:16 davis left #git
22:19 nickabbey joined #git
22:19 miczac joined #git
22:20 fuzzybear3965 thiago, how would I have used bugzilla (where is it hosted for GitHub) to avoid this problem of duplicating issue/PRs?
22:20 jedahan joined #git
22:22 DrSlony The goal is to rename "master" to "gtk2" and to rename "gtk3" to "dev". Does this look correct? http://paste2.org/np7A3t7m
22:22 thiago you would use bugzilla outside of GitHub
22:22 fuzzybear3965 Oh. Do you do this? Is this why you suggested it?
22:23 brokensyntax joined #git
22:23 jeffreylevesque joined #git
22:25 malide Bugzilla is for reporting issues on Mozilla projects/products. https://bugzilla.mozilla.org/. It's not affiliated with GH at all.
22:26 fuzzybear3965 malide, I know this.
22:26 thiago Bugzilla is a free and open source tool that you can install and run. It is not exclusive for Mozilla.
22:26 fuzzybear3965 But, thiago is suggesting to use that to avoid duplicating GitHub issues.
22:26 fuzzybear3965 That's the context.
22:26 fuzzybear3965 So, that's what's confusing me.
22:27 malide thiago: fair point
22:27 thiago I am saying that there are better tools than GitHub's issue tracker feature
22:27 fuzzybear3965 Oh, okay.
22:27 fuzzybear3965 I see.
22:29 XVar joined #git
22:33 govg joined #git
22:36 JanC_ joined #git
22:36 kent\n is there a way to tell git to emit times in TZ=UTC without having to set it via ENV? Particularly because I want to have an alias that always uses UTC instead of localtime
22:36 Gsham joined #git
22:37 kent\n "ENV is the only way" seems so wrong
22:37 theonefoo joined #git
22:38 theonefoo how can I preserve the git repository when using `git archive HEAD --format=zip > archive.zip` ?
22:38 matsaman kent\n: what's the problem with setting it & unsetting it?
22:38 kulelu88 joined #git
22:39 kent\n matsaman: because I can't do that in an alias
22:39 kent\n its like there are so many options to the date options, but none of them that I actually need.
22:40 thiago kent\n: define "emit"
22:40 thiago kent\n: where are you seeing the times?
22:40 kent\n git log
22:40 thiago why do you care?
22:40 thiago are you parsing it?
22:40 kent\n no.
22:41 thiago just a person reading it?
22:41 kent\n yes.
22:41 thiago set TZ environment variable and pass --date=local
22:41 pl left #git
22:41 kent\n I'm aware of this.
22:42 matsaman kent\n: oh a git alias
22:42 matsaman TBH I've always viewed git aliases as a crutch for people who don't have bash or similar
22:42 thiago why can't you do it in an alias?
22:42 thiago you can set environment variables in them
22:42 matsaman haven't found a reason to use a git alias over a bash alias/function, yet
22:42 thiago $ git config --get alias.aa
22:42 thiago !GIT_EDITOR=/bin/true git rebase -i --autosquash --autostash $(git merge-base HEAD HEAD@{upstream})
22:42 kent\n my bash profile is too bloated as it is :/
22:43 theonefoo left #git
22:43 thiago git alias and a script in $PATH are identical
22:43 thiago shell alias or function aren't handled by the git binary
22:43 kent\n I guess I can do that.
22:44 kent\n but like, I try to avoid "!" unless its strictly necessary.
22:44 thiago aye
22:44 thiago setting environment variables or running subcommands require it
22:44 chipotle joined #git
22:44 acetakwas joined #git
22:46 miczac joined #git
22:46 Vampire0 matsaman, e. g. shell completion does work fine with git aliases
22:47 kent\n matsaman: most the reason for me is a "namespacing" issue, I don't need a shell alias visible everywhere, only to git
22:48 nowhereman joined #git
22:48 kent\n and I regularly have fun problems where I enter a shell, try to run a command and discover its not there, because something stupid happened and its not in my env :/
22:48 aw1 joined #git
22:48 mehola joined #git
22:52 relipse how do I completely undo a git merge and get out of the conflict resolution thing
22:52 preaction git merge --abort
22:53 Puffball joined #git
22:55 seni joined #git
22:55 stux|RC-only joined #git
22:55 oxiih joined #git
22:56 Naan joined #git
22:56 dansan joined #git
22:57 matsaman kent\n: well, you can use a function wrapper, but whomever said git will take an env var anyway so =)
22:57 kent\n format placeholder syntax is confusing :/
22:57 kent\n like, how do I render %b indented 4 spaces like git log does by default?
22:58 dansan Hello! Is there a nifty git way to see the differences in two different versions of a patch set that are on my_patchset-v1 and my_patchset-v2, where -v2 has been rebased and I don't want to see the upstream changes?
22:58 metalraiden34 joined #git
22:58 dansan hehe, I suppose I could try to rebase one or the other...
23:00 overlord_tm joined #git
23:00 kent\n ah. %w ftw I guess
23:01 fmeerkoetter joined #git
23:05 seni joined #git
23:08 Dougie187 left #git
23:08 ShalokShalom_ joined #git
23:12 Vampire0 kent\n, just add 4 spaces?
23:12 Walex joined #git
23:12 Vampire0 git log -1 --pretty=format:'    %b'
23:12 kent\n Vampire0: that will indent the first line only....
23:13 kent\n %w(0,4,4)
23:13 kent\n the documentation sucks here.
23:14 Vampire0 Well, [<width>[,<indent1>[,<indent2>]]]
23:14 Vampire0 Linewrap the output by wrapping each line at width. The first line of each entry is indented by indent1 spaces, and the second and subsequent lines are indented by indent2 spaces.  width,
23:14 Vampire0 indent1, and indent2 default to 76, 6 and 9 respectively.
23:14 mquin joined #git
23:14 Vampire0 Is not that hard to understand, is it?
23:15 kent\n it is if don't even know "%w" is what you want, because its documentation in `git help log` doesn't explain what it is beyond "wrapping" and "read git short-log"
23:15 kent\n "%w([<w>[,<i1>[,<i2>]]]): switch line wrapping, like the -w option of git-shortlog(1)." # verbatim
23:16 kent\n looks unrelated to indentation at first glance.
23:16 Vampire0 kent\n, yeah, maybe. But I guess it is 1 of 1000 that ever uses this, if not even fewer :-D
23:16 kent\n story of my life
23:17 kent\n I'm always doing things nobody thought to do for some reason
23:17 kent\n makes for great fun in discovering fractal breakage that everyone conveniently stepped over
23:18 kent\n X thing nobody uses itself uses Y thing nobody uses  and uses Z thing that nobody uses, and all are broken \o/
23:19 ShalokShalom joined #git
23:19 Darren_ joined #git
23:20 a_thakur joined #git
23:20 WaReZ joined #git
23:20 matoro joined #git
23:21 DieguezZ joined #git
23:22 _raynold_ joined #git
23:24 dsdeiz joined #git
23:24 dsdeiz joined #git
23:24 nd joined #git
23:26 invisbl joined #git
23:27 invisbl joined #git
23:27 __builtin left #git
23:28 adino joined #git
23:32 Kaisyu joined #git
23:33 om_henners joined #git
23:33 serialoverflow joined #git
23:37 matoro joined #git
23:38 om_henners joined #git
23:39 pioiytr_ joined #git
23:39 pioiytr joined #git
23:39 rivarun joined #git
23:40 DolpheenDream joined #git
23:42 Es0teric joined #git
23:52 dumacdev joined #git
23:54 shgysk8zer0_ joined #git
23:54 justanotheruser joined #git
23:55 Peng_ joined #git
23:56 p4trix joined #git
23:57 seni joined #git
23:58 Peng joined #git
23:58 dirtyroshi joined #git
23:59 jccn joined #git
23:59 Tobbi joined #git

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