Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2017-05-15

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

All times shown according to UTC.

Time Nick Message
00:00 transhuman git log --diff-filter=D --summary | grep "delete mode 100" | cut -c 21- | sort > deleted.txt
00:03 cydrobolt joined #git
00:03 cydrobolt joined #git
00:04 CDanU I don't get that also. Better try writing it again. I looks different from what you wanted initially.
00:04 CDanU * verbally writing it out
00:04 nachoman joined #git
00:05 miha_S7_ joined #git
00:07 SkyPatrol_ joined #git
00:08 msonntag joined #git
00:10 transhuman CDanU, supposedly that gets a list of all the files deleted in a branch, I want to get a list of the files deleted and when they were deleted in all of the history. I assume maybe if I had a hash of the files that are deleted I could search for that hash in subsequent commits to find when they were added perhaps?
00:11 transhuman errr. re-added if they were infact re-added somewhere else
00:11 TheSimonator joined #git
00:12 durham joined #git
00:13 cam27 joined #git
00:14 Stumbler joined #git
00:16 CDanU Ok I actually tested your command and id does seem to show all deleted files. That hash only shows you at which point in time the files where deleted. You can't use that one to find them again.
00:21 maroloccio joined #git
00:21 NeXTSUN joined #git
00:22 MattMaker joined #git
00:22 CDanU I doubt that git has something that keeps track of when a deleted file is re added. Btw. do you need to be sure that is has the same/similar content when it is re added or only if the same file name is re added ?
00:22 anuxivm left #git
00:22 Stumbler_ joined #git
00:23 somiaj joined #git
00:23 kwic joined #git
00:24 cdown joined #git
00:25 bitmod why is it than when i clone a repo from a local file, into two seperate dirs and give them each different .gitconfigs, they end up syncing
00:25 qndel joined #git
00:25 bitmod *from  a local bare repo
00:25 qndel hey, any1 here?
00:25 gitinfo qndel: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.
00:26 qndel Is it possible to push code to some git repo directly from a website ?
00:27 Gsham joined #git
00:27 transhuman CDanU need to see if it has the same content. I am assuming if a file is re-added that in most cases its a move from one directory to another so that finding the exact same content is the only way if its changed /has been updated before the re-add its going to be impossible to find
00:29 bitmod why is it than when i clone a bare repo (from a local directory)into two seperate dirs and give them each different .gitconfigs, they end up syncing?
00:30 kadoban bitmod: What?
00:31 CDanU transhuman: well not impossible but time consuming with flaky results. You could make a feature vector of some sorts from t:delete-1 and than check that with every re added file feature fector after t:delete.
00:31 bitmod kadoban: i created a --bare repo, then "git clone"d it twice, with each clone having its own .gitconfig. For some reason these .gitconfig files keep syncing
00:31 bitmod (i.e. i edit one, the other changes so they match)
00:31 qndel can somebody help me with my question?
00:31 justanotheruser joined #git
00:33 kadoban bitmod: What's a .gitconfig? Are you talking about the config file in .git/ in the repo ?
00:34 CDanU qndel: well the website has to support such a feature, on github for example you can make small changes to a file and than make a commit from that
00:34 bitmod kadoban: crap, you're right. thanks
00:36 CDanU qndel: Or do you actually want do build such website? That is also possible, your web server would execute the git commands that it gets via the website.
00:36 qndel Well I think it will be nice if I add some more info
00:36 gunnaro joined #git
00:36 Cabanossi joined #git
00:37 qndel I have a python script and I've found some free host where I can run it, it has only 7 days trial and after that I have to reupload my script to a new git repo, I do it from a virtual machine but would like to move the whole process to some website so it wouldn't need a virtual machine
00:37 wxg joined #git
00:39 CDanU qndel: why do you need the vm in the first place or a website.
00:39 qndel to push my code from there ?
00:40 qndel I'm running ubuntu and using the git there
00:40 CDanU you can push your code from your pc into a remote git repo
00:41 qndel yeah but that'd require me to install git anyway
00:41 qndel I want to be able to do it anywhere, not only from my PC
00:42 transhuman CDanU, git rev-parse $REVISION:$PATH  <--- this gets the hash but what is meant by $REVISION?
00:43 CDanU qndel: As far as I know every linux box has ssh, ssh into a box that has git and push from there. Installing git will be by far the easiest solution.
00:43 boombatower joined #git
00:43 durham joined #git
00:44 transhuman skip that CDanU I found better doc on it
00:44 qndel yeah but using a website would be easier ;)
00:45 ISmithers joined #git
00:45 peepsalot joined #git
00:47 TheSimonator joined #git
00:47 watabou joined #git
00:48 CDanU I know that there was some project on git that let you run a small linux in your web browser completely locally. I am not sure if that on had a network connection though. Search for that.
00:48 CDanU *github
00:48 cagedwisdom joined #git
00:49 jameser joined #git
00:49 transhuman qndel can you explain what you want to do in  a different way...? I might be able to help
00:50 qndel yeah, sorry for that, might be because I don't really know what I want haha
00:50 qndel I create an app on my hosting website, they give me a git repo
00:50 qndel like this Repository git@git.evennode.com:71c6197316b56120e47bbdc36a1a.git, I push my code there and it works
00:51 qndel I'd like to create a website where I just put this thing into a text field and it pushes my code to that repo
00:52 transhuman put what thing in a text file?
00:53 transhuman so your saying you have a page where you want to enter the code ...then click a button and have it update the git location?
00:53 qndel yeah
00:53 qndel no I don't have it, I'd like to have it
00:53 qndel :D
00:54 dsdeiz joined #git
00:56 transhuman qndel https://github.com/kevclark/cgit
00:57 qndel yeah but where am I going to install it?
00:58 fstd_ joined #git
00:58 transhuman install it in a VM on the cloud azure or something get a small instance for 5 bucks a month
00:59 transhuman install ubuntu
01:00 transhuman (ubuntu 16.04 server)
01:00 transhuman or a windows instance if its a windows only piece of software
01:01 qndel 5 bucks a month :D
01:01 qndel if I wanted to pay 5 bucks a month, I'd have bought the python host and I wouldn't have to redeploy every week :P
01:01 transhuman maybe 10
01:01 transhuman ok well cant help you unless you want to self host from your pc
01:01 transhuman how much bandwidth do you have at home
01:02 dodobrain joined #git
01:02 qndel the whole point is I dn't want to self host from my PC
01:03 qndel I have 64 mb/s
01:03 transhuman do it from a rasberri pi
01:04 qndel it would make everything I've done so far pointless
01:05 transhuman why? sorry you probably said but I didn't understand
01:05 SkyPatrol joined #git
01:06 qndel my goal is to get things running remotely
01:06 qndel without needing to turn on any stuff at home
01:06 miha_S7_ joined #git
01:06 qndel so far I've achieved what I wanted but I just have to turn on a VM once a week to redeploy the code
01:07 CDanU here it is: https://bellard.org/riscvemu/js/ not internet unfortunately
01:07 qndel yep already found it :P
01:08 Gsham joined #git
01:08 transhuman whats wrong with leaving a raspberry pi or raspberry pi zero running? You could even put it on a timer to start
01:09 qndel well ... let's begin with the fact that I don't have a raspberry pi ;)
01:09 transhuman so 5 bucks is too much money or 10 bucks to get wireless?
01:09 CDanU well at some point you have to throw around some coins
01:09 qndel wireless?
01:10 Husky_Apocalypse joined #git
01:10 transhuman there is a raspberry pi that has wireless for 10 bucks plus 10 bucks for a good size usb flash drive
01:10 qndel could you give me a link to that 10 bucks one?
01:10 qndel the cheapest ones I see in my country would be around 50$
01:11 Bock joined #git
01:11 CDanU raspberry pi zero w
01:11 livingstn joined #git
01:11 CDanU i guess they should be sold out at this point
01:11 transhuman spend 75 bucks you can use it for much more odroid but here it is
01:11 transhuman http://www.popularmechanics.com/technology/a25433/raspberry-pi-zero-w-10-dollars/
01:11 rchavik joined #git
01:12 transhuman http://www.hardkernel.com/main/products/prdt_info.php
01:12 qndel well it's nope anyway
01:12 qndel my script is a discord chat bot
01:12 qndel one day I might stop caring about it
01:13 Sasazuka joined #git
01:13 qndel but I'd like to be able to just give some website to a guy and say "Hey, if you want to keep it running, rehost once a week"
01:13 Bock joined #git
01:13 transhuman why is it a no?
01:13 nachoman joined #git
01:13 qndel so a 10$ raspberry pi has bluetooth
01:14 qndel and my 30$ tablet doesn't :D
01:14 cads joined #git
01:14 transhuman you must either be real cheap or even poorer than me ...how about the odroid?
01:15 cebula joined #git
01:15 qndel I have a pc with gtx 1080 xD
01:17 svm_invictvs joined #git
01:21 finalbeta joined #git
01:21 transhuman hey for 30 bucks I have an Alix computer I can sell you with an 800 mhz geode cpu?
01:22 transhuman https://pcengines.ch/alix2d3.htm
01:22 transhuman 3 network ports keep it running all the time
01:23 qndel I have a 30$ tablet ;)
01:23 qndel and no :P
01:23 qndel I have a pc/laptop/2 tablets/4 phones
01:23 qndel I don't need more stuff
01:23 Kuukunen joined #git
01:23 qndel seems I can't push the code without a VM then
01:24 cqi joined #git
01:24 om_henners joined #git
01:25 git868 joined #git
01:28 beerdrop joined #git
01:29 Stumbler joined #git
01:29 sunri5e joined #git
01:30 * bremner wonders wth he wandered into
01:30 raijin joined #git
01:32 PEMapModder joined #git
01:34 Nilesh_ joined #git
01:37 lagothri1 joined #git
01:43 svm_invictvs joined #git
01:47 lagothrix joined #git
01:47 svm_invictvs joined #git
01:47 mizu_no_oto joined #git
01:49 ilbot3 joined #git
01:49 Topic for #git is now Welcome to #git, the place for git help. Just git help, apparently. | Public logs at https://gitirc.eu/log | First visit? Read: https://gitirc.eu | Current stable version: 2.13.0 | Getting "cannot send to channel"? /msg gitinfo .voice | This is serious business. https://madeitwor.se/git-jokes
01:50 TheSimonator joined #git
01:54 transhuman qndel what kind of phones do you have?
01:54 qndel XD
01:54 qndel it's irrelevant lol
01:54 transhuman got a link to it
01:54 transhuman XD what
01:55 transhuman its not irrelevant... you can run linux on a rooted phone
01:55 transhuman if you can run linux you can run LAMP if you can run LAMP you can run the software I found
01:57 d^sh joined #git
01:58 transhuman http://danielpetty.com/android-lamp-linux/
01:58 transhuman Root your system. Install the following packages: Superuser, Android Terminal Emulator, and Linux Installe follow this guide
01:59 dreiss joined #git
01:59 danimal2 joined #git
02:00 transhuman or without rooting https://play.google.com/store/apps/details?id=com.alfanla.android.pws&amp;hl=en
02:00 bitmod joined #git
02:00 dermoth joined #git
02:04 Stumbler joined #git
02:05 Cabanossi joined #git
02:06 svm_invictvs joined #git
02:08 qndel nah
02:08 qndel you are still missing the point
02:08 lagothri1 joined #git
02:11 transhuman ok well I think you arn't explaining it very well, sorry I did my best
02:11 netj joined #git
02:19 safe joined #git
02:20 Vortex34 joined #git
02:20 maroloccio joined #git
02:21 qndel I know, thanks for your help :)
02:24 ojdo joined #git
02:26 transhuman Hi! find -name .git -prune -o -exec git log --pretty=tformat:'' --numstat --follow --find-copies-harder --reverse {} ';' | cut -f3- | grep '{.* => .*}' <== does this screw up the tree or does it as I think it does, just find files that have been renamed/moved?
02:27 transhuman (at the current commit level?)
02:28 mizu_no_oto joined #git
02:29 Gsham joined #git
02:31 hexagoxel joined #git
02:32 ShapeShifter499 joined #git
02:35 leeN joined #git
02:35 adac joined #git
02:42 dsdeiz_ joined #git
02:44 lb1c joined #git
02:46 aavrug joined #git
02:48 aavrug joined #git
02:56 chachasmooth joined #git
02:56 aw1 joined #git
02:57 AlexKvazos joined #git
02:59 leeN joined #git
03:01 AlexKvazos left #git
03:05 jon-mac joined #git
03:15 lightstalker joined #git
03:17 dsdeiz joined #git
03:18 bitmod what does git diff do?
03:21 maroloccio joined #git
03:23 BorkStick joined #git
03:23 msonntag joined #git
03:23 titankiller joined #git
03:25 xyphoid_ joined #git
03:26 miczac joined #git
03:29 meLon joined #git
03:34 Cabanossi joined #git
03:34 star_prone joined #git
03:35 Random832 joined #git
03:35 robotroll joined #git
03:36 Andrew_K joined #git
03:36 TikityTik joined #git
03:37 Garen joined #git
03:38 Garen_ joined #git
03:41 cqi joined #git
03:42 theiostream joined #git
03:42 renkon left #git
03:46 hhee joined #git
03:48 salamanderrake joined #git
03:49 xyphoid_ so we've got a master branch and a CI process that does build stuff. for (annoying) reasons we want to set up a branch that contains the built version of master - eg a commit to master will trigger a commit to this one automatically, but with production-ready files etc. is there a standard name for that kind of thing?
03:50 xyphoid_ thinking about master-built or master-compiled or something, but if there's a standard name for that kind of thing i'd rather use that
03:51 theiostream joined #git
03:51 netj joined #git
03:53 ShapeShifter499 joined #git
03:55 robattila256 joined #git
04:03 theiostream joined #git
04:04 Anja_ joined #git
04:06 jimmyrcom joined #git
04:13 Stumbler joined #git
04:14 theiostream joined #git
04:15 overlord_tm joined #git
04:17 miha_S7_ joined #git
04:18 clickjack joined #git
04:24 thiago xyphoid_: "master-built" sounds like you've checked in generated/compiled files
04:24 thiago which you shouldn't do
04:26 xyphoid_ that's where the 'annoying reasons' comes in (an external deployment system that has input of 'a git tag' and can't do the compilation part for you)
04:26 kadoban I don't think there's a standard name for that type of thing.
04:27 thiago there's clearly no standard name for a practice that is frowned upon
04:28 kadoban Just having some "here be dragons" crappy deploy branch isn't really that bad. The bad parts are when you're generated stuff is intermixed with content in real branches.
04:28 kadoban Though it's not that great either, but w/e, you can't avoid it anymore.
04:31 jidar xyphoid_: why not just have artifacts that exist in a repo (not git, say yum/rpm/tar.gz) outside of git
04:32 jidar ¯\_(ツ)_/¯
04:32 kadoban Because some tools make it easier to use a git repo somewhere
04:32 Cabanossi joined #git
04:33 valhala_22 joined #git
04:33 thiago well, UGFIINI (Use of Git For What It Is Not Intended)
04:33 jidar I hear what you're saying, I just think it's bad design
04:34 kadoban It is. But the user rarely has much say in that.
04:38 nutron|w joined #git
04:39 om-henners joined #git
04:43 aw1 joined #git
04:44 ayogi joined #git
04:45 ych joined #git
05:01 fggrf joined #git
05:01 lordjancso joined #git
05:02 freimatz joined #git
05:04 tymczenko joined #git
05:07 dec0n joined #git
05:07 dcpc007 left #git
05:14 tmz joined #git
05:24 dmc joined #git
05:26 tymczenko joined #git
05:26 dmc joined #git
05:33 robattila256 joined #git
05:34 navidr joined #git
05:35 overlord_tm joined #git
05:37 dersand joined #git
05:43 chandu20 joined #git
05:47 nd joined #git
05:47 qt-x joined #git
05:48 manuelschneid3r joined #git
05:53 rscata joined #git
05:55 jaziz joined #git
05:55 jozwior joined #git
05:55 User458764 joined #git
06:01 ertes-w joined #git
06:02 Raging_Hog joined #git
06:02 matoro joined #git
06:03 ThomasLocke joined #git
06:03 ThomasLocke joined #git
06:03 valhala_22 joined #git
06:04 lowercaseman joined #git
06:06 salamanderrake joined #git
06:08 ExpiredDev joined #git
06:09 jameser_ joined #git
06:10 JeroenT joined #git
06:10 User458764 joined #git
06:14 gry joined #git
06:15 theoceaniscool joined #git
06:15 gry hi, i've added a remote and it's a few commits ahead; how do i view each of them and approve or reject? is there a standard comfortable process for this? thanks in advance.
06:18 zeroed joined #git
06:18 freimatz joined #git
06:18 dsantiago joined #git
06:19 roelmonnens joined #git
06:19 Cabanossi joined #git
06:19 MineCoins joined #git
06:20 zeroed joined #git
06:21 oskarkv joined #git
06:21 dreiss joined #git
06:22 zeroed joined #git
06:24 sea-gull joined #git
06:27 zefferno joined #git
06:30 cqi joined #git
06:30 JanC_ joined #git
06:31 aw1 joined #git
06:33 aw1 joined #git
06:35 slidercrank joined #git
06:35 finalbeta joined #git
06:37 ogny joined #git
06:39 finalbeta joined #git
06:39 JC_Yang joined #git
06:39 star_prone joined #git
06:39 King_Hual joined #git
06:39 King_Hual joined #git
06:40 JC_Yang left #git
06:40 JC_Yang joined #git
06:43 JC_Yang how to properly restore a git repository if I tar and transfer it? after I unpack the tarball, no matter how I do, git reset --hard, git always complain files are modified and not stage for commit
06:44 AndreasLutro check file permissions
06:45 selckin maybe fs that forces +x/-x or maybe some crlf madness
06:48 JC_Yang the file permission of the sources? hmm... maybe that's when I tar, I specified -p...
06:48 rosek joined #git
06:50 selckin can you pastebin a transcript
06:50 aw1 joined #git
06:50 selckin i usually just tar up .git/ and then git checkout -- . after untarring it to get my workdir back
06:50 selckin but reset should also work
06:52 JC_Yang but all files seems to have rwx permissions for all users, it that problematic?   I use "tar pcJf mytar.tar.xz src_dir/"
06:54 sostenuto joined #git
06:54 ferr joined #git
06:54 selckin git doesn't care about any permission except +x
06:54 selckin what does git diff show
06:54 xall joined #git
06:56 sz0 joined #git
06:58 chele joined #git
07:01 JC_Yang thank u for your suggestion, selckin. git config core.filemode false solve it, now the working dir is clean, but is it a good solution?
07:01 Murii joined #git
07:01 selckin if one of them is like a vfat or ntfs volume mounted from linux it'll force +x on every file
07:02 selckin so you need to fix that in the mount config
07:03 mkalfon_ joined #git
07:06 lindenle joined #git
07:07 Milos joined #git
07:08 elect joined #git
07:08 TomyWork joined #git
07:09 YuGiOhJCJ joined #git
07:11 jackrandom joined #git
07:14 schleppel joined #git
07:15 BorkStick joined #git
07:15 Vampire0 joined #git
07:18 j08nY joined #git
07:19 adac joined #git
07:21 chandu20 joined #git
07:22 armyriad joined #git
07:22 freddy_ joined #git
07:22 Introoter joined #git
07:23 dec0n joined #git
07:27 furrykef joined #git
07:27 JeroenT joined #git
07:28 JeroenT joined #git
07:29 User458764 joined #git
07:30 sea-gull joined #git
07:33 kelvah joined #git
07:36 _ng joined #git
07:37 Levex joined #git
07:43 Junior joined #git
07:43 rcn joined #git
07:45 chrisM_1 joined #git
07:45 djb-irc joined #git
07:47 kurkale62 joined #git
07:48 mkalfon_ joined #git
07:48 MrMojit0 joined #git
07:49 acetakwas joined #git
07:51 chardan joined #git
07:51 dodobrain joined #git
07:51 freddy__ joined #git
07:53 jameser joined #git
07:53 kurkale62 joined #git
07:55 clickjack joined #git
07:55 ShapeShifter499 joined #git
07:56 mikecmpbll joined #git
07:56 User458764 joined #git
08:00 JeroenT joined #git
08:03 stepho joined #git
08:05 aielima joined #git
08:06 zefferno joined #git
08:07 _ikke_ and git also probes the fs and set some default config
08:08 _ikke_ (late response)
08:11 pbandark joined #git
08:11 Spherical joined #git
08:12 freddy_ joined #git
08:13 skunkz joined #git
08:13 freddy_ joined #git
08:14 nutron|wirk joined #git
08:15 fatalhalt joined #git
08:16 wrouesnel joined #git
08:17 chardan joined #git
08:18 rominronin joined #git
08:19 freddy_ joined #git
08:19 RxMcDonald joined #git
08:20 _ADN_ joined #git
08:20 freddy__ joined #git
08:21 RxMcDonald how do you merge a commit from the past into head? I created a new branch at said commit, then try git merge branch and it says up to date
08:21 cbreak RxMcDonald: it is already merged
08:21 cbreak RxMcDonald: why do you want to merge it again?
08:23 RxMcDonald cbreak: you think this is not normal?
08:23 cbreak RxMcDonald: what?
08:23 _ikke_ RxMcDonald: Merging something that has been already merged is kind of pointless
08:23 cbreak RxMcDonald: not being able to merge something that is already part of history is normal
08:23 _ikke_ RxMcDonald: But I assume you want to achieve something different
08:24 cbreak RxMcDonald: because it is obviously already merged
08:24 _ikke_ RxMcDonald: So can you please describe what you want to achieve?
08:24 JeroenT joined #git
08:24 Jan-_ left #git
08:25 ams__ joined #git
08:26 KekSi joined #git
08:27 feng joined #git
08:27 aidalgol joined #git
08:27 Levex joined #git
08:30 JC_Yang left #git
08:30 timewalker joined #git
08:30 freddy_ joined #git
08:33 fletom joined #git
08:33 jrahmy_ joined #git
08:36 Mr_Muh joined #git
08:37 KekSi joined #git
08:38 borkr joined #git
08:39 chandu20 joined #git
08:39 _ADN_ joined #git
08:39 nocbreak joined #git
08:39 Mr_Muh git 2.13 crashes on solaris (core dump) with a simple git status. I switched back to the old version. Tried GIT_TRACE but it gives no more clue... any suggestions?
08:41 avar Mr_Muh: Did you just build 2.13 yourself?
08:41 Mr_Muh avar: yes
08:42 Mr_Muh on a new git repository it does not crash
08:42 avar I'm one of the git devs and have done solaris fixes in the past, but I lost access to my solaris build boxes long ago, I'd be interested to help you with this if you're interested.
08:42 avar i.e. bisect git, see what the bug is..
08:43 tvw joined #git
08:44 Mr_Muh avar: okay, I have to set it up first then... hopefully I can make automate the bisect
08:45 avar Are you using gcc or suncc, also is this sparc or x86?
08:45 Mr_Muh avar: are there any other things I can try first (enable more debugging)?
08:45 avar You likely won't get much info out of trace, but if you  can run it with gdb or a similar debugger that would be very useful
08:45 Mr_Muh avar: sparc gcc
08:45 avar I can walk you through that if you don't know how..
08:46 Mr_Muh avar: first I need to get a full git git repository to that machine, so I can bisect
08:46 JeroenT joined #git
08:46 tvw joined #git
08:47 cstratak joined #git
08:47 avar right, but just a simple backtrace might make it easy to find the patch right away
08:47 avar Anyway, if you want to bisect first, go ahead :)
08:49 cjb73 joined #git
08:52 strk left #git
08:52 yaewa joined #git
08:54 tcorneli joined #git
08:56 cagedwisdom joined #git
08:57 moei joined #git
08:59 ThiefMaster is there any nice way of applying the current (unstaged) changes on another worktree? i can do `git diff > /tmp/whatever.diff && cd /.../myworktree && patch -p1 < /tmp/whatever.diff` but i'm hoping for a nicer way of doing that
09:01 mentazoom joined #git
09:03 sunri5e joined #git
09:06 theoceaniscool joined #git
09:07 JeroenT joined #git
09:07 slidercrank joined #git
09:07 beerdrop joined #git
09:08 star_prone joined #git
09:08 AndreasLutro maybe git stash, switch branch, git stash pop
09:09 re1 joined #git
09:09 tobiasvl what's "another worktree" though?
09:10 tobiasvl I'm thinking ThiefMaster doesn't mean "another branch"
09:10 ThiefMaster man git-worktree ;)
09:10 gitinfo the git-worktree manpage is available at https://gitirc.eu/git-worktree.html
09:10 ronskisika joined #git
09:11 AndreasLutro ah my bad
09:11 ThiefMaster but using a stash sounds like a good idea.. maybe with a script like this to avoid actually resetting the changes: http://stackoverflow.com/questions/6315459/git-command-to-save-a-stash-without-modifying-working-tree
09:12 dasidiom joined #git
09:12 YuGiOhJCJ joined #git
09:12 visual hello /gits/
09:13 visual how would i go about tracking _only_ specific folders?
09:13 visual like add * gitignore to gitignore _except_ a few folders?
09:13 tobiasvl visual: ignore everything (*), then unignore specific folders (!foo\n!bar)
09:13 gitinfo visual: bar
09:13 visual googling suggests ignoring *, and unignoring the ones
09:13 tobiasvl gitinfo: please
09:13 tobiasvl visual: yes
09:14 visual tobiasvl: thanks, that's exactly what i needed
09:14 ZyX-I joined #git
09:14 visual !folder1 {enter} !folder2 ?
09:14 visual or * !folder1 {enter} !folder 2? :I
09:14 visual not rly a .gitignore gur
09:14 visual y
09:15 visual s/y/u
09:15 tobiasvl * {enter} !folder1 {enter} !folder2
09:15 stfn joined #git
09:16 visual right, thanks!
09:19 tobiasvl visual: you need to unignore some more stuff if you have subdirectories under folder1 and folder2 though
09:19 tobiasvl see for example http://stackoverflow.com/questions/5533050/gitignore-exclude-folder-but-include-specific-subfolder/
09:19 visual i was thinking !folder1/*
09:19 visual oh that's the right way :D, thanks tobiasvl
09:20 metalraiden34 joined #git
09:22 tobiasvl yeah
09:28 kelvah joined #git
09:33 DaGoaty joined #git
09:34 cebula joined #git
09:34 sbasso joined #git
09:37 sbasso_ joined #git
09:37 PioneerAxon joined #git
09:38 ki0 joined #git
09:38 SkyPatrol_ joined #git
09:39 mindfart joined #git
09:44 bakibour joined #git
09:44 bakibour left #git
09:45 bourbaki joined #git
09:45 bourbaki Hi i try to run "git clean -dxf -e "somedirectory/"" works fine. Now i want to exclude *.someext from the clean also how do i do that?
09:47 Cabanossi joined #git
09:48 njbair joined #git
09:50 kurkale6ka joined #git
09:55 tymczenko joined #git
09:55 bitmod joined #git
09:55 ertes-w joined #git
09:58 dsdeiz joined #git
09:58 ertesx-w joined #git
10:00 kurkale6ka joined #git
10:01 Schrostfutz joined #git
10:02 jnavila joined #git
10:03 Random832 joined #git
10:06 barteks2x I have a small problem: I'm bisecting a repository that at one point had a submodule, but doesn't have it anymore, and there was also a time before that submodule. I'm bisecting from the beginning to the end. And the submodule doesn't seem to update it's contents when I run git bisect next. Any way to make it update and appear/disappear as necessary?
10:06 dsdeiz joined #git
10:07 Kaisyu joined #git
10:07 star_prone joined #git
10:09 Nilesh_ joined #git
10:13 alcohol barteks2x: just run git submodule update, ignoring any errors from this command?
10:13 barteks2x it just doesn't seem to do anything
10:14 barteks2x it returns instantly with no meassage and contents of that submodule are unchanged
10:15 tlaxkit joined #git
10:15 barteks2x it seems like the submodule was no longer there after git bisect next so submodule update didn't remove it
10:15 barteks2x any way to remove it as needed?
10:17 barteks2x what git... warning: unable to rmdir fml: Directory not empty
10:17 barteks2x that's what happened on git bisect next when the submodule got removed
10:18 chandu20 joined #git
10:24 robattila256 joined #git
10:24 xanadu joined #git
10:25 romank joined #git
10:25 AfroThundr joined #git
10:26 AfroThundr joined #git
10:27 star_prone joined #git
10:28 dsdeiz joined #git
10:28 dsdeiz joined #git
10:29 dendazen joined #git
10:34 bourbaki Ok so 'git clean -dxf -e somedir/ -e *.xxx' works it will not delete anything from somedir and no *.xxx files in the root. But i can not get it to work with subdirectories for excluding .xxx files.
10:34 tourdownunder joined #git
10:34 cdown joined #git
10:34 bourbaki Is there any special format or somesuch? Is it because the folder the .xxx files are in is in the .gitignore?
10:35 _ikke_ if they are in .gitignore, it will not delete them by default
10:35 bourbaki That is what the x is for right?
10:36 _ikke_ Oh, right, you are correct
10:36 bourbaki I really do not get the problem yet :(
10:37 bourbaki **/*.xxx does not work either not even when i explicitly specify the file
10:38 IRQ` joined #git
10:38 freimatz joined #git
10:39 _ikke_ git ls-files -o '**/*.xxx', does that return the file?
10:39 cdown_ joined #git
10:41 gunnaro joined #git
10:41 cdown__ joined #git
10:41 bourbaki sec
10:42 tvw joined #git
10:42 dwts joined #git
10:42 bourbaki It lists this file.
10:42 _ikke_ but with -e '*.xxx', it would still delete that file?
10:42 star_prone joined #git
10:42 bourbaki Yes
10:43 bourbaki Well it says it deletes the directory
10:43 bourbaki With the file in it, the file is not listed explicity when i dryrun
10:43 bourbaki -e **/*.xxx does not work either.
10:43 _ikke_ right
10:43 Snugglebash joined #git
10:44 _ikke_ I think that's a bug that's going to be fixed
10:44 bourbaki _ikke_: can you try if this behaviour is the same for you?
10:44 bourbaki I need some workaround thoug.
10:45 bourbaki +h
10:46 _ikke_ https://public-inbox.org/git/CABfTTAchc61aB02sCD=Oa9gRMGr94h7mC53B9q6Qy2k2hDqzAQ@mail.gmail.com/
10:46 _ikke_ Can you confirm this is the issue?
10:47 bourbaki Sec
10:47 jast did you quote the pattern in your commands ('*.xxx' instead of *.xxx) to avoid the shell messing it up?
10:48 bourbaki I thought i did but will try that also.
10:48 bourbaki My issue is a bit different i use x and the files are in the .gitignore.
10:49 jast you didn't use quotes in your examples, that's why I asked :)
10:49 bourbaki But i tried it without the folders or files in the .gitignore with the same outcome.
10:49 _ikke_ yeah, with -x, .gitignore should not be considered at all
10:49 jast with -x .gitignore and friends are ignored, so that much shouldn't make a difference
10:50 bourbaki neither ' nor " make a difference
10:50 _ikke_ I think it's the same underlying issue
10:50 Gustavo6046 joined #git
10:50 jast let me try and reproduce
10:50 Vampire0 jast, I also tried it. clean just does not descend into folders it will delete for checking the -e rules
10:51 jast ohh, I see
10:51 jast that's... unfortunate :)
10:51 bourbaki Ok i just setup a new repo and the problem is the same.
10:51 jast you could work around that with the usual cascade of ignore and unignore patterns, I guess, but that gets unwieldy fast
10:52 Vampire0 jast, `mkdir a && touch a/b && git clean -fd -e b -e '**/b'` will still delete the whole a folder
10:52 Anja_ joined #git
10:52 jast actually I guess that cascade wouldn't help much, either
10:53 _ikke_ jast: I think so neither
10:53 bourbaki Vampire0: so you got the same issue on *nix?
10:53 _ikke_ adding a tracked dummy file in the directory might help?
10:54 jast that's nasty :) but might help
10:54 bourbaki Maybe the problem is that it is in the bin folder and the file does not have the same name all the time.
10:54 Vampire0 bourbaki, actually this was in Cygwin with a Git for Windows called, but I don't think it would be any different on *nix
10:54 bourbaki How long does it usually take for bugs like this to be fixed?
10:56 _ikke_ depends a bit in how involved the fix is going to be
10:56 _ikke_ might be in the next minor release, but it can also take a while before it actually makes a release
10:57 jast I'm looking at the code right now
10:58 bourbaki jast: Cool
10:58 sbasso joined #git
10:59 jast okay, looks like our guesses were right
10:59 jast basically it looks at the index entries and applies the ignore rules to them
11:00 bourbaki So it is a bug? Or is there a way to make this work?
11:00 Vampire0 bourbaki, `git add -f '**/*.xxx' && git clean -dxf -e somedir/ -e *.xxx && git reset '**/*.xxx'` should work as a workaround
11:00 bourbaki Vampire0: Sec
11:01 jast I'll keep you updated as I read the code and try to understand it
11:01 Vampire0 bourbaki, sorry, almost: `git add -f '**/*.xxx' && git clean -dxf -e somedir/ -e '**/*.xxx' && git reset '**/*.xxx'` I meant
11:02 jast are you sure the standard glob parser supports '**'? as I recall only ignore patterns do
11:03 Vampire0 jast, at least it worked with the a/b example I wrote above
11:03 star_prone joined #git
11:03 Vampire0 jast, there I used '**/b'
11:03 Vampire0 at all three places
11:03 jast did you check it works in more nested dirs, too? :)
11:03 Vampire0 But might have been a coincidence
11:04 Vampire0 nope
11:04 jast I'm not completely sure either way
11:04 jast just something to consider
11:04 bourbaki Vampire0: The workaround works, thanks.
11:04 bluepixel joined #git
11:05 finalbeta joined #git
11:06 Vampire0 jast, yes, it works
11:07 Vampire0 jast, I just tried it with a/b/f and a/b/c/d.xxx and with the exact command I gave bourbaki last. Works fine
11:07 Vampire0 jast, It just deleted a/b/f
11:08 bourbaki Thanks guys and sorry for the hassle :)
11:08 jast oh okay, good to know
11:09 star_prone joined #git
11:09 Vampire0 bourbaki, you're welcome
11:09 gitinfo bourbaki: This channel tracks karma based on who has gotten lots of thanks for being helpful. If you want to help someone reach karmic nirvana, please mention their name when thanking them with "thank you", "thankyou", "thanks", "thx", "ty" or "cheers". Try ".karma <nick>" or ".topkarma" to show karma status of a person but don't expect immediate increase. Ten thanks make up one karma point.
11:09 zerow joined #git
11:10 bourbaki thank you Vampire0  jast  _ikke_ :)
11:11 sarri joined #git
11:11 sarri joined #git
11:11 pbandark1 joined #git
11:11 bremner what's the most pleasant git2mercurial tool these days? (accessing mercurial repos via git)
11:13 * RichiH hands bremner the eyefork
11:13 Vampire0 RichiH, is that a tool, or do you want bremner to stick the fork into his eye?
11:14 RichiH neither; it's the more permanent cousin of eyebleach and bremner might need it
11:14 bremner only for messages that you send me, mainly
11:14 jast I believe it's a proven method to distract from the pain caused by... other things
11:14 m0nk3h joined #git
11:14 Tobbi joined #git
11:15 Vampire0 jast, ah, ok, like the foot-hammer then :-D
11:16 bourbaki Or the Acme(tm) anvil.
11:16 jast I think the Acme(tm) anvil doubles as a feedback device
11:17 dasidiom joined #git
11:18 Vampire0 bremner, at least the "official" way is https://github.com/git/git/tree/master/contrib/remote-helpers, but I have no idea how good it works :-)
11:19 m0nk3h If I have a git repo on a server and a local git repo that started at the same place as the server but was 'git init-ed' rather than pulled/cloned. Is there any way to 'append' my local changes to those on the server?
11:19 xanadu joined #git
11:19 nineg joined #git
11:20 Vampire0 m0nk3h, add the server with man git remote and then use rebase or merge to combine those histories. How exactly heavily depends on the exact situation
11:20 gitinfo m0nk3h: the git-remote manpage is available at https://gitirc.eu/git-remote.html
11:20 jast m0nk3h: so the server already has commits but they're different from what you have locally?
11:20 Stace joined #git
11:21 m0nk3h server has commits 1, 2, 3, ... n. client started with the same code as commit n and went from there
11:21 m0nk3h no further changes in the server
11:21 m0nk3h so I could just clone, copy over the new code and push
11:21 m0nk3h but then I'd lose the log
11:21 jast okay. do you have a commit with the same contents as 'n' in your client, or does your first commit already include changes?
11:21 Stace ...an accurate representation of my first 6 months of git: https://cdn-images-1.medium.com/max/1400/1*kTvU2RZdpTaR2JMUADkVNw.png
11:22 m0nk3h the first commit is "initial commit" so adding stuff mostly. probably included a slight change...so n+1 I guess
11:22 Vampire0 m0nk3h, then the same as I said, but use man git replace then to replace the start commit from the client with your server n commit
11:22 gitinfo m0nk3h: the git-replace manpage is available at https://gitirc.eu/git-replace.html
11:22 dege joined #git
11:22 m0nk3h oops, OK, missed your reply sorry
11:23 Vampire0 m0nk3h, ah, so not identical, then us a graft to set the parent
11:23 * m0nk3h googles graft ^_^
11:23 jast the first two steps are the same in any case: git remote add origin (URL); git fetch origin
11:23 jast this will give you a local mirror of the server stuff to work from
11:23 jast let me explain a little from the technical side
11:24 adac joined #git
11:25 jast one of the tools for history mutilation that we have is rebase. it takes commits as patches and re-applies them to a different base commit. this might seem like a perfect solution to your situation, except that your first commit is basically a patch that "adds" everything. since the server already has everything, too, there will be a conflict
11:25 MineCoins joined #git
11:25 jast and a huge one at that. it's not that big a deal because we can just override the conflict with what's in your first commit and then continue, and rebase will happily chug along for the rest of your commits
11:25 Vampire0 jast isn't a graft the best way if one repository does 1..n and the other n+1..HEAD?
11:26 jast grafts are a more straightforward approach to do this but there are more small hassles (you have to mess around more with commit IDs and the 'filter-branch' tool gets involved which means there's a cleanup phase at the end, too)
11:27 Vampire0 ok
11:27 m0nk3h thank you chaps
11:28 jast it's a matter of taste to some extent. both will do the job, I wouldn't mind using either, but I think rebase might be slightly friendlier
11:28 m0nk3h got to switch back to work network to try this out -_-
11:28 xanadu joined #git
11:29 leeN joined #git
11:29 jast Stace: less complaining, more Git godhood :P
11:30 lss8 joined #git
11:31 cstratak joined #git
11:31 Stace jast: I'm *finally* getting to the point where I know enough that I'm not *terrified* of randomly losing work. I've also found that git usually warns you when you do something potentially dangerous.
11:32 Stace perhaps terrified is a bit strong. more like weary
11:32 jast well that's certainly a point of less stress at least :)
11:32 Cabanossi joined #git
11:33 Stace git reset -xfd sets the same alarm bells off that rm -rf * does
11:33 jast okay, but those flags are from 'git clean' :)
11:33 Stace oh, git clean. lol
11:33 Stace see, one never reaches god-hood
11:34 jast never doubt the God-making skills of this channel
11:34 Stace oh yeah, the people around here have been really helpful
11:35 nutron|wirk joined #git
11:36 Stace once you get there, you start to just version control ALL the things. I have my linux user config and bashrc under version control now. although that may just be me being odd
11:36 jast (and of trying things out on dummy repos)
11:36 jast there are plenty of people who keep their dotfiles in Git repos
11:36 jast personally I've so far resisted that siren's call, seems like too much of a mess to me
11:37 alcohol why? it is quite easy
11:38 Stace I started because I started customising my linux user significantly (custom systemd startup scripts, vpn settings, bash scripts for all kinds of things) and knew that I'd only want to have to do all of that once
11:38 jast I'm definitely not disputing that
11:38 Stace for a vanilla user setup it's  not really necessary
11:38 robattila256 joined #git
11:39 jast but I'm a fan of the 'one thing per repository' philosophy, because history becomes much more useful that way... and putting a bunch of unrelated dotfiles in a repo is the opposite of that
11:39 jast so I started writing a thing to simplify managing dotfiles in several repos (or using other sources) but it's slow going because there's not that much of an actual itch :)
11:40 Stace in my head that one thing is *config*. it's mostly all config, sh, tcl, and perl
11:40 Stace so scripting
11:41 jast yeah, but then you have a mix of "change editor config" commits and "change display manager config" commits etc.
11:41 Stace one does question whether it's a good idea to put a private ssh key in the repo. I haven't done that
11:41 jast one use case I have in mind is having slightly different environments on different machines
11:41 dglambert joined #git
11:41 dglambert yo
11:41 jast with one big dotfiles repo that's complicated
11:42 jast well, if you like sharing your credentials with everyone everywhere, sure, why not upload your private key to github :}
11:42 Stace jast: well, obviously not a public repo
11:43 jast still, the idea of private keys is to not share them with helpful service providers...
11:43 Stace I'm just pointinh out a limitation ...
11:43 Stace one can't just blindly add dotfiles to the repo
11:44 dsdeiz joined #git
11:44 dsdeiz joined #git
11:44 jast true. blindly adding is rarely the best way to go, in fact
11:44 Stace what's wrong with separate commits for different things? Even my work repo has those ("module x: bug x fixed"
11:45 Levex joined #git
11:45 sbasso_ joined #git
11:45 Stace ) not that my naming convention is the best
11:45 jast there are plenty of public repos with config files that contain API keys and passwords, great examples of that...
11:45 Stace oh, yeah, that's not good
11:45 jast there's nothing wrong with having multiple modules of the same thing in the same repo, *iff* you never need them separately
11:46 jast for me, my work env and my private env at home have many similarities and I keep large parts of them in sync, but there are some important differences, too
11:46 jast if I just stuff everything in a big bag, there's no sane way for me to account for that
11:46 Stace oh, yeah, I have scripts to keep my repos in sync too. I have a private one which is mine and then I have repos for my clients
11:47 jast in short, the big bag hurts reusability
11:47 dasidiom joined #git
11:47 Stace and then I also have a work one which is for my personal IP
11:47 Stace it can be quuite a deep rabbit hole to get lost down
11:51 cagedwisdom joined #git
11:56 NeverDie joined #git
11:56 rchavik joined #git
11:56 TheCycoONE joined #git
12:00 PHPanos joined #git
12:00 jameser joined #git
12:00 JeroenT_ joined #git
12:01 kurkale61 joined #git
12:02 valhala_22 joined #git
12:02 jozwior joined #git
12:03 zulutango joined #git
12:07 Snugglebash joined #git
12:09 _ng joined #git
12:11 jackrandom joined #git
12:12 Murr joined #git
12:12 courrier joined #git
12:14 xanadu joined #git
12:14 sbasso joined #git
12:14 dfas joined #git
12:15 blackwind_123 joined #git
12:17 kelvah joined #git
12:17 jameser joined #git
12:17 Cabanossi joined #git
12:18 cstratak joined #git
12:19 stepho joined #git
12:22 jameser joined #git
12:22 kurkale61 joined #git
12:23 theoceaniscool joined #git
12:25 theiostream joined #git
12:27 jameser joined #git
12:27 cdown joined #git
12:28 chandu20 joined #git
12:28 cstratak joined #git
12:29 plos joined #git
12:30 m4sk1n joined #git
12:32 dendazen joined #git
12:34 PaulCapestany joined #git
12:35 star_prone joined #git
12:36 shinnya joined #git
12:37 Snugglebash joined #git
12:37 leeN joined #git
12:38 fission6 joined #git
12:39 kurkale61 joined #git
12:41 BSaboia joined #git
12:41 Snugglebash joined #git
12:42 Altiare joined #git
12:43 theiostream joined #git
12:47 tsdev joined #git
12:47 Masber joined #git
12:49 kexmex joined #git
12:53 noidea joined #git
12:56 jnavila joined #git
12:56 adac joined #git
12:57 theiostr_ joined #git
13:00 ExoUNX joined #git
13:04 gugah joined #git
13:04 theiostream joined #git
13:05 Marquisk2 joined #git
13:07 Nilesh_ joined #git
13:08 Stumbler joined #git
13:10 rasto joined #git
13:11 leeN joined #git
13:12 fission6 joined #git
13:12 SkyPatrol joined #git
13:13 clickjack joined #git
13:15 om_henners joined #git
13:16 sircortly joined #git
13:16 Acerific joined #git
13:16 Acerific joined #git
13:17 star_prone joined #git
13:17 yehowyada joined #git
13:18 sz0 joined #git
13:19 cdg joined #git
13:20 Bhootrk_ joined #git
13:26 IanLiu left #git
13:27 h12o joined #git
13:28 TheSimonator joined #git
13:30 flaviodesousa joined #git
13:30 patarr joined #git
13:32 Cabanossi joined #git
13:34 Bigcheese joined #git
13:34 ResidentBiscuit joined #git
13:35 fragMental joined #git
13:38 _ng joined #git
13:38 shiriru joined #git
13:38 nowhere_man joined #git
13:43 Es0teric joined #git
13:44 lindenle joined #git
13:48 jikz joined #git
13:49 theoceaniscool joined #git
13:51 ilithium joined #git
13:51 Gurkenglas joined #git
13:52 al-damiri joined #git
13:53 flash_ joined #git
13:53 ilithium Hey all. I have a repository with two branches: master and other. I'd like to swap these around so that master becomes 'archive' and 'other' becomes master. Can anyone advise on the correct way to do this?
13:54 tobiasvl ilithium: you can rename branches with "git branch -m master archive", but note that when you rename "other" to "master" you might not be able to push that without forcing, if it's not a fast forward merge
13:55 flash_ Hi guys. I'm automating a build process which as input takes either a commit hash or branch name. Assuming the repo can be in any (old) state, I first do a git fetch, but then I need a way to checkout the specified commit or branch, with the newest code guaranteed. How would I do that?
13:56 flash_ Any local changes may and must be overwritten and the repo cleaned. I basically want a snapshot of the code at the point specified.
13:56 Stace joined #git
13:56 jast the procedure is a little different depending on whether you want to checkout a commit or a branch
13:56 IanLiu joined #git
13:57 jast actually I guess it doesn't matter that much. this is for a clone that isn't used for working/developing in?
13:57 nuc joined #git
13:57 tymczenko joined #git
13:57 tobiasvl well, if local changes must be thrown away, "git reset --hard" will do the trick for both commits and branches
13:57 catern hey #git
13:57 catern I want to get a list of the my most recent commits on this branch, so I do git log -n10 --author=me
13:57 catern --author=catern
13:58 catern but most of the commits on this branch are merge commits, with meaningless commit messages
13:58 thebope joined #git
13:58 catern how can I see the real commits instead?
13:58 jast flash_: 1) git fetch 2) git checkout -f yourthingy^0. obviously you don't want to do this normally, only for very destructive purposes. :)
13:58 tobiasvl catern: --no-merges
13:58 tobiasvl jast: why not reset --hard?
13:59 flash_ jast, what does the ^0 do?
13:59 jast because reset --hard could accidentally update a local branch pointer
13:59 tobiasvl that's true
13:59 jast it makes sure HEAD is always detached. I guess actually you don't need the ^0 here
13:59 jikz flash_, to fetch the specific version of the file of that specific commit,
13:59 jast it's an extra precaution to avoid any branch from being modified though it probably doesn't really make a difference if that does happen
14:00 flash_ tobiasvl, the branch may already have been checked out previously with newer changes in origin.
14:00 jikz git show [commitid]:[COMMITFILE]
14:00 jikz this should print the version of the commitfile for the commitid
14:00 jast note that if you want to checkout a branch, you'll want the remote branch (e.g. origin/master), not the local (e.g. master)
14:00 jast oh, and I guess since that implies detaching HEAD the ^0 is completely superfluous :)
14:01 catern tobiasvl: but that just hides the merges, I don't think it shows the actual commits merged in
14:01 jast if they are "real" merges, the actual commits are shown
14:01 IanLiu joined #git
14:01 flash_ Is there a command or series of commands that would work for both a commit and branch?
14:01 jast if someone made squash merges or something nasty like that, Git can't find the corresponding individual commits
14:02 tobiasvl catern: not following – do you want to list commits that you merged in, but didn't author yourself, but not list the merge commits?
14:02 jast flash_: the one I outlined, if you make sure that you use the remote-tracking branch, not the local branch
14:02 jast git fetch; git checkout -f <commit/remote-tracking branch>
14:03 flash_ I would still need to distinguish between commit and branch and if possible I want to avoid this. git checkout -f origin/branch works, but git checkout -f origin/commit throws an error.
14:03 IanLiu left #git
14:03 IanLiu joined #git
14:03 jast yeah, but there isn't really any way around that... because the local branch doesn't match the remote branch if there have been updates remotely
14:04 Vampire0 flash_, if the "repo cleaned" also means untracked and ignored files should be removed, you might also need to do `git clean -fdx` after the checkout
14:04 hollusion joined #git
14:04 hollusion i am having a strange problem with my .gitignore file
14:04 flash_ Vampire0, gotcha.
14:04 hollusion line 20: config.inc.php line 21: config.inc.php*
14:05 hollusion it is ignoring config.inc.php_live
14:05 hollusion but not config.inc.php
14:05 Vampire0 flash_, is it essential that you can give "master" instead of "origin/master"?
14:05 hollusion any ideas?
14:05 catern tobiasvl: I guess there's some weird thing with how we are using git at my company but when I pass --no-merges it doesn't show some commits I know I did
14:05 jast hollusion: maybe you 'git add'ed config.inc.php earlier? tracked files bypass ignore patterns
14:05 catern regular commits that is
14:05 flash_ What is the best way to determine if a string is a commit or branch?
14:05 hollusion jast, thank you very much
14:05 hollusion how do i untrack the file without deleting it from the main repo?
14:05 jast you can't
14:06 flash_ Vampire0, if a branch is specified, origin/branch is assumed.
14:06 jast untracking is the same as deleting
14:06 hollusion okay  i got you
14:06 tobiasvl hollusion: !ignore_tracked
14:06 gitinfo hollusion: Git only applies ignore patterns to untracked files. You can't use ignore patterns to ignore changes to files that are already tracked by git. To remove files only from git, but keeping them on disk, use git rm --cached <file>. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.
14:06 Vampire0 hollusion, !ignore_untracked
14:06 hollusion thank you again
14:06 Vampire0 ah, right, wrong trigger :-D
14:06 jast flash_: you could just check that origin/<thingy> is valid (git rev-parse --verify -q origin/<thingy> and look at the exit code)
14:07 nuc left #git
14:07 hollusion would it be a good idea to rename the tracked file to rename_me_to_config.inc.php?
14:07 jast hollusion: that's a common way people do this kind of thing, yeah
14:07 jast !configfile
14:07 gitinfo [!configfiles] It is recommended to store local configuration data in a file which is not tracked by git, but certain deployment scenarios(such as Heroku) may require otherwise. See https://gist.github.com/1423106 for some ideas
14:07 Vampire0 flash_, you could as jast said, verify whether origin/thingy exists and then do either or the other, or you could just try with origin/thingy and if it fails just thingy
14:08 fk joined #git
14:08 mizu_no_oto joined #git
14:08 flash_ That seems reasonable.
14:08 jast but in the latter case suppressing the error output is tricky
14:08 Vampire0 jast, is it? `2>/dev/null`?
14:09 jast that suppresses success output, too :)
14:09 TbobbyZ joined #git
14:09 jast depends on what output you want, I guess
14:09 flash_ Good points, both ways. Thanks for the advice, I'll get it going from here.
14:09 Vampire0 jast, success output is written to stderr? o_O
14:09 jast stderr is for all status output
14:09 linuxmodder- joined #git
14:09 jast stdout is for data
14:10 Vampire0 jast, ah, ok. So yes, depends on what you expect to see or not. :-)
14:10 xanadu joined #git
14:10 jast I had a lengthy discussion about this a while ago in this channel, in fact, with someone being adamant that only errors belong on stderr... but none of the common tools do it that way :)
14:10 ChiLL_SurF joined #git
14:10 jast and it would be stupid for anything that outputs data you want to pipe somewhere else
14:11 Vampire0 jast, that's true, but what actual data does checkout output beside the status messages?
14:11 jast this isn't really relevant to checkout, it's a more general thing
14:12 chigang joined #git
14:12 j08nY joined #git
14:12 Vampire0 jast, well, I am used to and would find it reasonable that a command (subcommand) that outputs actual data does status or progress data to stderr to not pollute stdout. But a command (subcommand) that does not output anything else should output status to stdout I'd say. But that is philosophical I guess. :-D
14:13 furrykef really the problem is there are only two standard output streams, I think.
14:13 mage__ left #git
14:13 furrykef Sometimes it'd be handy to have a third
14:13 jast many programs have options to control what output is generated in the first place
14:13 jast and more output streams exist, it's called "the program opens a file" ;)
14:14 jameser joined #git
14:14 _ikke_ furrykef: if you look at fd's, there are many more you can open :P
14:14 jast or a pipe or whatever
14:14 AaronMT joined #git
14:16 Murii joined #git
14:17 robotroll joined #git
14:17 clickjack joined #git
14:18 Hello71 Or pass in more FDs.
14:19 iffraff joined #git
14:19 cdown joined #git
14:21 jameser joined #git
14:22 mizu_no_oto joined #git
14:22 stepho_ joined #git
14:22 xall joined #git
14:24 fission6 joined #git
14:24 _ng joined #git
14:25 Altiare joined #git
14:26 raijin joined #git
14:26 HardlySeen joined #git
14:27 SkyPatrol_ joined #git
14:27 JeroenT joined #git
14:29 jameser joined #git
14:30 docmur joined #git
14:30 docmur if I use git stash list, i can see 6 stashes in the list.  How do I clear the stash or can I clear the stash?
14:30 PityDaFool joined #git
14:31 kelvah joined #git
14:31 Chill-Surf joined #git
14:32 s17 git stash something
14:32 s17 man git-stash
14:32 gitinfo the git-stash manpage is available at https://gitirc.eu/git-stash.html
14:33 docmur Perfect :) it was git stash clear!
14:33 docmur thx
14:34 JeroenT joined #git
14:34 tabakhase__ joined #git
14:34 star_prone joined #git
14:35 LeBlaaanc joined #git
14:35 JeroenT_ joined #git
14:35 tabakhase__ joined #git
14:36 jwest joined #git
14:37 RxMcDonald left #git
14:37 ski7777 joined #git
14:39 Hunter1 joined #git
14:39 TheSimonator joined #git
14:40 psprint Hello. I've just did: git remote add bb https://...@bitbucket.org/... After `git push -u bb master`, I've got 'Branch master set up to track remote branch master from bb.' Why did it associate "master" with not-origin, i.e. not-main site, "bb"?
14:40 zerow joined #git
14:40 jameser joined #git
14:40 jast psprint: because you asked git to. that's precisely what -u does.
14:41 jast if you use -u git assumes you know what you're doing :)
14:41 psprint I thought it's "don't worry no common references"
14:42 demize -u is short for --set-upstream
14:42 psprint jast: how to associate master back with origin?
14:42 jast git branch -u origin/master
14:42 psprint demize: hyh
14:42 jast assuming you're on master
14:42 Snugglebash joined #git
14:42 Emperor_Earth joined #git
14:43 psprint ok, all works, thanks
14:44 jt joined #git
14:44 jast so, for future reference, don't use -u with push if you don't want to set up new tracking info :)
14:45 psprint jast: but it is needed for pushing to empty repository I think.. When you create some repo on github, bitbucket, etc.
14:45 jast nah
14:45 psprint I like notabug.org because it has "Mirror" repository type, and just copies things from github
14:45 transhuman Hi! find -name .git -prune -o -exec git log --pretty=tformat:'' --numstat --follow --find-copies-harder --reverse {} ';' | cut -f3- | grep '{.* => .*}' <== does this screw up the tree or does it as I think it does, just find files that have been renamed/moved? At the current commit level. Where I found this command they were talking about it screwing things up...just wanted to be sure...thanks
14:47 robotroll joined #git
14:47 groverdill joined #git
14:47 jast that looks pretty complicated
14:47 aspiers_ joined #git
14:47 MarioBranco joined #git
14:49 transhuman it produces output like this http://paste.debian.net/932541/
14:49 rivarun joined #git
14:49 ChiLL_SurF joined #git
14:50 Hunter1 joined #git
14:50 jast in any case I don't see how it could mess anything up, other than potentially giving you a lot more output that you need :)
14:50 dave0x6d joined #git
14:50 MarioBranco joined #git
14:51 transhuman I take it most of the changes were not just simple renames but copies or splits?
14:51 relipse joined #git
14:51 King_Hual joined #git
14:51 King_Hual joined #git
14:51 notebox joined #git
14:52 stana joined #git
14:52 relipse I made a commit on develop branch but I do not want it anymore how do I go back to origin/develop instead I didn't push
14:52 MarioBranco joined #git
14:52 jast relipse: assuming you've got develop checked out, git reset --hard origin/develop will erase all traces, including any uncommitted changes
14:54 Hunter1 joined #git
14:54 PHPanos joined #git
14:54 transhuman what my intention is is to make a history/patch of all the changes to files throughout the tree and throughout all the commits/versions/tags/etc (tracing these changes) its for data to train a neural net. I have other commands for doing this but I was worried about renames and moves preventing a good and accurate flow
14:55 karols146 joined #git
14:55 relipse I tried that and now it says HEAD is now at 7e210a4 merge conflict
14:55 relipse Does that mean my changes got pushed?
14:55 jeffreylevesque joined #git
14:56 furrykef I just did "git remote add -f old git@github.com:SNES-SDK/llvm-65816.git" in anticipation of merging a subdirectory from it into a different repository. However, it seems to be taking a long time. It's probably been five minutes now. Is that normal?
14:56 rmesta1 joined #git
14:57 jast furrykef: the 'git remote add', or a subsequent 'git fetch'?
14:57 ChiLL_SurF joined #git
14:57 furrykef Well, it says:
14:58 furrykef Updating old
14:58 furrykef Warning: Permanently added 'github.com,192.30.253.112' (RSA) to the list of known hosts.
14:58 furrykef That's where it's stuck.
14:58 dodobrain joined #git
14:58 furrykef "updating" means it's fetching, right?
14:58 jast oh, I didn't know -f implies fetching :)
14:58 furrykef yeah
14:58 jast my guess is github was stuck for a moment, I'd cancel and do a separate fetch to try again
14:59 jast normally you do get progress updates during fetching, so the lack of those likely means it never got that far
14:59 star_prone joined #git
15:00 manuelschneid3r joined #git
15:01 hherman1 joined #git
15:01 furrykef OK. I did ctrl+c and "git fetch old", and got the same message about adding the host to the list of known hosts. That seems weird.
15:01 hherman1 Hey guys! I have a question
15:01 furrykef It still doesn't seem to be doing anything
15:01 furrykef hherman1: just ask your question
15:01 miczac joined #git
15:01 irqq joined #git
15:02 zivester joined #git
15:02 TheSimonator joined #git
15:02 hherman1 Suppose I have the directory structure /A/a/src and /B/b/src and In my upstream I want to have the structure App/a/src and App/b/src ... so I want to get around the A vs. B issue without consolidating the folders. Is this possible? Thanks
15:03 aboccia joined #git
15:03 roelmonn_ joined #git
15:03 jast furrykef: yeah, that *is* weird. you could try a 'ssh git@github.com' and see if properly terminates with a friendly message from github...
15:03 MarioBranco joined #git
15:04 Snugglebash joined #git
15:04 furrykef Yeah, that works. But it still gives me the "Warning: Permanently added..." message, even if I do it twice in a row.
15:04 MarioBranco joined #git
15:05 Masber joined #git
15:05 jast maybe your machine is possessed?
15:05 kevwil joined #git
15:05 furrykef I also got a "PTY allocation request failed on channel 0" before the friendly message, both times.
15:06 jast that's probably normal
15:06 Vampire0 furrykef, with the same IP or with different IPs?
15:07 furrykef same
15:07 grawity don't mind the pty error; it's something that 'batch' connections don't use at all
15:07 grawity so they've disabled it on the server entirely
15:07 jast from my end, there are only two IP addresses associated with the hostname
15:08 Vampire0 furrykef, do you maybe miss write permission on the file it mentions
15:08 jast it doesn't mention a file, does it
15:08 furrykef The ssh weirdness probably has to do with me using GitHub Desktop on Windows
15:09 hobodave joined #git
15:09 ilithium left #git
15:09 Vampire0 jast, hm, I may remember wrongly, I thougt it mentions the known_hosts file it tried to write to
15:09 furrykef nope
15:10 jast oh, yeah, indirectly
15:10 Vampire0 furrykef, nope what? does not mention the file or you don't have write permission?
15:10 Vampire0 furrykef, it should be ~/.ssh/known_hosts I think
15:11 furrykef It does not mention the known_hosts
15:11 mizu_no_oto joined #git
15:11 jagob joined #git
15:11 jnewt joined #git
15:11 Vampire0 Well search for the file and see if you can edit it. As I said, should be ~/.ssh/known_hostss
15:11 Vampire0 -s
15:11 furrykef I'm on Windows. I don't think that directory structure is relevant here
15:12 furrykef ah
15:12 furrykef actually I can do cd ~
15:12 Vampire0 furrykef, sure, it is your user folder :-)
15:12 Vampire0 furrykef, but where the known hosts are saved depends on the SSH you use
15:12 furrykef well I don't know how this thing is set up. It uses PowerShell, which I'm not familiar with
15:13 cdown joined #git
15:13 kevwil joined #git
15:13 jast doesn't matter
15:13 Vampire0 furrykef, the "Git Cmd" or what?
15:13 Vampire0 furrykef, I'd recomment using "Git Bash" instead, but shouldn't matter much
15:13 jast if ~/.ssh/known_hosts exists, chances are that's what's used by your client
15:13 furrykef OK, there is a .ssh with a github_rsa and and a github_rsa.pub, but no known_hosts
15:13 furrykef Should I maybe touch it?
15:14 Vampire0 furrykef, do you have write permission in this folder?
15:14 jast well, the bad news is that none of us here know what magic things the Github client does :)
15:14 Vampire0 furrykef, so **can** you actually touch it?
15:14 jast so feel free to experiment
15:14 Vampire0 furrykef, yeah, like !gui :-)
15:14 gitinfo furrykef: Graphical user interfaces are not supported here. If you want to get support, it needs to be through the git CLI. Reasons: 1) Because very few people here use the graphical interface. 2) Because giving instructions for GUI's is difficult. 3) The command line gives you a history of what commands you have executed.
15:15 furrykef It let me touch it and the file is there, but nothing is added to it when I ssh
15:15 Vampire0 jast, but if he uses `ssh` like you suggested to test and it gives that message, this has nothign to do with the GitHub client I'd say
15:15 wrouesnel joined #git
15:15 jast well who knows how they patched the SSH client they ship with GitHub Desktop
15:16 jast and AFAICS there is no source code for their GitHub Desktop package
15:16 furrykef yeah, it's a fair point.
15:16 jast which is an interesting implied statement from a company that professes to be all for open source, but eh whatever
15:16 Vampire0 furrykef, maybe use the Sysinternal Process Monitor (not Process Explorer) to watch the ssh executable trying to open files and registry entries to see where it tries to write to
15:16 furrykef jast: indeed
15:17 Vampire0 furrykef, maybe that helps to find the file it cannot write to
15:17 jast assuming it does even try to write
15:17 Vampire0 yes, that assumed
15:18 MineCoins joined #git
15:19 Gurkenglas joined #git
15:19 Vampire0 furrykef, ah, you can even make it easier
15:19 furrykef hm?
15:20 Vampire0 furrykef, add -vvv to the ssh command. It will output on the debug3 level the path to the file it tries to read or write for known hosts
15:20 Vampire0 furrykef, Process Monitor will show the filesystem accesses though, -vvv is just the debug output of the SSH client, so it might not do what it tells you to do. But maybe you find the file it tries to read / write there
15:21 roelmonnens joined #git
15:21 h12o joined #git
15:21 orbyt_ joined #git
15:22 furrykef debug3: hostkeys_foreach: reading file "/dev/null"
15:22 furrykef debug3: hostkeys_foreach: reading file "/dev/null"
15:22 furrykef Warning: Permanently added 'github.com,192.30.253.113' (RSA) to the list of known hosts.
15:22 furrykef If that's the "file" it's trying to write to, then it's no wonder it won't remember :P
15:23 furrykef There's also: debug3: hostkeys_foreach: reading file "/dev/null"
15:23 jast yeah, that would explain it
15:23 jast but we still don't know why it freezes
15:23 furrykef indeed...
15:23 jast I'm going to be lazy and just blame your weird client and leave it at that ;)
15:24 furrykef I'm thinking of just deleting my local repo and checking it out again. GitHub Desktop has been having lots of crashes with it anyway
15:24 rominronin joined #git
15:24 furrykef and I have no idea if there's something wrong with my repo or if the client is just being a douche
15:24 furrykef (I've fscked it but nothing came up)
15:24 jast if fsck runs through in less time than your fetch hangs, I don't think your repo is to blame
15:25 furrykef well it does take a little while. The repo is LLVM.
15:27 xall joined #git
15:29 Endarked joined #git
15:29 ravndal joined #git
15:29 mmlb joined #git
15:29 thiago joined #git
15:30 furrykef Still less time than that command hangs, though.
15:31 furrykef I don't imagine there's a similar way to debug a git fetch command?
15:32 grawity GIT_TRACE=1
15:32 tlaxkit joined #git
15:32 SteffanW joined #git
15:32 Vampire0 grawity, with github desktop?
15:33 Levex joined #git
15:33 furrykef I wonder, what if I maybe use https instead of the git protocol for the remote. Let's try that.
15:33 Vampire0 furrykef, you didn't use the git protocol, but the ssh protocol, that is a huge difference ;-)
15:34 Vampire0 furrykef, if you use the https protocol it might work, but then you have to identify with username / password, not with ssh keys anymore
15:34 furrykef oops.
15:34 furrykef That's OK, this is just a one-time thing.
15:34 valhala_22 joined #git
15:35 BSAlb joined #git
15:35 hherman1 For those of you familiar with Go (and the GOROOT), how would you structure a repo containing a client and a server (where the server is Go)?
15:35 monte joined #git
15:36 Vampire0 hherman1, is that related to Git? *wondering*
15:37 Seveas hherman1: to start with, I'd stick them in different repos
15:38 hherman1 I would do that too, but the projects are so small that it seems not worth it for them to each be on their own
15:38 dreiss joined #git
15:38 furrykef Using https instead of ssh seems to have made no difference. There's no visible progress being made.
15:39 Vampire0 hherman1, I'd say if they are highly coupled and always released together, it is fine to do them in one repo. If they are independent and talking with each other via a stable interface, releasing the two separately, then it is probably better to separate them out in different repos if not doing it anyway to logically separate them better
15:39 stana joined #git
15:40 ertes joined #git
15:40 Husky_Apocalypse joined #git
15:40 furrykef I take that back. It's doing something now.
15:40 scarabx joined #git
15:41 duderonomy joined #git
15:44 mattcen joined #git
15:48 iffraff joined #git
15:49 scoobertron joined #git
15:49 Introoter joined #git
15:50 cam27 joined #git
15:50 madewokherd joined #git
15:52 furrykef OK, that worked.
15:52 valhala_22 joined #git
15:56 valhala_22 joined #git
15:58 somiaj left #git
16:01 kegster joined #git
16:02 Es0teric joined #git
16:03 tyreld joined #git
16:05 tristanp joined #git
16:05 rebrec joined #git
16:05 cdg joined #git
16:06 ravndal joined #git
16:07 Husky_Apocalypse joined #git
16:09 mel00010 joined #git
16:09 soc42 joined #git
16:09 brk_ joined #git
16:10 durham joined #git
16:10 kfoonamalik joined #git
16:10 trobotham joined #git
16:10 deepakbndeepu joined #git
16:11 jackrandom joined #git
16:11 Jitta joined #git
16:11 brent__ joined #git
16:11 BSab joined #git
16:12 MuffinMedic joined #git
16:12 clickjack joined #git
16:12 svm_invictvs joined #git
16:12 rory joined #git
16:13 theoceaniscool joined #git
16:13 jpX joined #git
16:14 relipse joined #git
16:14 Framedragger joined #git
16:14 cads joined #git
16:16 cstratak joined #git
16:17 ki0 joined #git
16:17 Cabanossi joined #git
16:17 phinxy joined #git
16:18 ssarah joined #git
16:21 crose joined #git
16:24 TheSimonator joined #git
16:24 rominronin joined #git
16:25 raspado joined #git
16:25 miha_S7_ joined #git
16:25 zumba_addict joined #git
16:25 stepho joined #git
16:25 Husky_Apocalypse joined #git
16:25 zumba_addict hi. Anyone here uses gitlab
16:26 TomyLobo joined #git
16:27 karols146 joined #git
16:27 Seveas zumba_addict: probably some people do. It even has its own #gitlab channel :)
16:27 zumba_addict oh ok, I wasn't aware. Thanks! :)
16:27 Framedragger left #git
16:28 nitric joined #git
16:29 overlord_tm joined #git
16:29 TomyLobo2 joined #git
16:30 webdev007 joined #git
16:31 dasidiom joined #git
16:32 orbyt_ joined #git
16:34 BSaboia joined #git
16:34 bdbaddog joined #git
16:35 bdbaddog Greetigns.  I'm doign some work on a branch, have 3 local files changed. I'd like to create a new branch and pull just one of those 3 locally changed files to that branch and check them in, without losing the other 2 changed files. How best to do that?
16:36 valhala_22 joined #git
16:36 groverdill joined #git
16:36 Seveas bdbaddog: did you commit any of the changes to those 3 files? And do you want to have the changes to other 2 files committed on the current branch?
16:37 bdbaddog haven't committed yet.
16:37 bdbaddog yes commit on current branch, but not ready yet for that. they need soem more work.
16:37 bdbaddog hmm. could commit one of the 2.
16:38 rivarun joined #git
16:38 Seveas bdbaddog: git checkout -b name-of-new-branch; git add file3; git commit; git checkout name-of-current-branch
16:39 bdbaddog k. thanks! that's what I was thinking, but didn't want to pull the trigger and then have to unravel it all.
16:41 freimatz3 joined #git
16:41 brent__ joined #git
16:41 boombatower joined #git
16:42 ogny joined #git
16:42 GT4066 joined #git
16:44 valhala_22 joined #git
16:44 wizardedit joined #git
16:45 snits joined #git
16:48 brent__ joined #git
16:48 eehackr joined #git
16:50 brent__ joined #git
16:51 cam27 joined #git
16:52 Stace` joined #git
16:52 meLon joined #git
16:52 rkazak joined #git
16:56 freimatz3 joined #git
16:56 strk joined #git
16:56 strk somehow my git repo was corrupted by an external event (I think): git branch -D master_2 # error: Couldn't look up commit object for 'refs/heads/master_2'
16:56 strk this exists: ./.git/refs/heads/master_2
16:56 strk but indeed it contains an hash which is not known by git
16:57 xall joined #git
16:57 furrykef Have you tried "git fsck"?
16:57 strk trying it now
16:58 strk it reports many danglign commits and warns about that missing ref
16:59 Masber joined #git
16:59 strk but still leaves me with the inability to delete the branch
16:59 strk git branch -D master_2 # error: Couldn't look up commit object for 'refs/heads/master_2'
17:00 sea-gull joined #git
17:00 furrykef mm, unfortunately that's beyond my area of expertise
17:00 strk I'm rm -rf .git/refs/<stuff>
17:01 Seveas strk: plumbing time! git update-ref -d refs/heads/master_2
17:01 strk hopefully my way wasn't too destructive
17:01 Seveas might need to also remove something in .git/logs
17:02 Seveas in general, manually poking at files inside .git is not recommended
17:02 strk I did .git/logs too
17:02 netj joined #git
17:02 peterbecich joined #git
17:02 Seveas if git fsck and git for-each-ref are happy, you've not messed up too much :)
17:05 Adylas Hello ! With 2.13.0. I'm trying my HTTP negociate authentication with git. It works fine when I do clone but when I do submodule update, it fail. Seems that libcurl is not used the same way when doing a submodule update than a clone operation. Anyone has an idea ? See debug : https://www.diffchecker.com/P02uQH1t
17:05 rebrec joined #git
17:07 Vampire0 Adylas, maybe run with GIT_TRACE=1 to see the difference?
17:08 furrykef I think I'm good for now. Thanks for your help, guys
17:08 furrykef left #git
17:10 strk Seveas: nope, still not happy, full of dangling commit reports
17:11 njbair joined #git
17:11 robattila256 joined #git
17:11 gnixev joined #git
17:11 Adylas Vampire0, Hello ! Yeah, I did without much success. However, that debug does not display the call made to cURL. Both call to "git-remote-https" are very similar.
17:11 valhala__ joined #git
17:13 duderonomy joined #git
17:15 thiago joined #git
17:16 nowhere_man joined #git
17:19 sathed joined #git
17:19 chardan joined #git
17:20 adac joined #git
17:21 Vampire0 strk, if those are the same as before, it is expected. dangling commits means you have commits that are not referenced by any ref anymore and will potentially be deleted on next garbage collection.
17:22 nixjdm joined #git
17:23 Vampire0 strk, if you deleted some non-merge branches or tags or did some history rewriting e. g. with interactive rebase or filter-branch, or cleared the reflogs, it is expected that you find quite some dangling commits
17:23 h12o joined #git
17:24 Vampire0 Adylas, if you want to know the exact call to curl, you could maybe add a /usr/local/bin/curl script that logs the parameters to some file and then calls the real curl.
17:25 thebope joined #git
17:26 lafleurdubien joined #git
17:26 relipse joined #git
17:28 om_henners joined #git
17:29 jnavila joined #git
17:29 Seveas strk: dangling commits are fine. any commit --amend creates one for instance
17:29 bdbaddog left #git
17:29 gopar joined #git
17:29 Puffball joined #git
17:32 Adylas Vampire0, Humm, interesting approach. However I guess that git is not calling the binary itself but use the lib and call it directly via code .. no ?
17:32 fission6 joined #git
17:32 flavius joined #git
17:32 karols146 joined #git
17:33 _ikke_ Adylas: RIght, git uses libcurl
17:33 jaziz joined #git
17:33 grawity GIT_CURL_VERBOSE=1
17:34 orbyt_ So i have some changes stored in git stash which I forgot to apply when I switched back to my working branch. I've since made a few changes and only just now realized I forgot to apply the stash. Can I apply it without losing my current changes?>
17:36 rory left #git
17:37 jnavila joined #git
17:37 irqq joined #git
17:38 Adylas grawity, Yeah, used that in my diff : https://www.diffchecker.com/P02uQH1t
17:39 miczac joined #git
17:40 cahoots joined #git
17:40 acidvegas joined #git
17:41 cahoots hi, i noticed that git has a noticeable delay when i do a git add --verbose or git commit --verbose before it starts printing anything out. is there some sort of git startup sequence that i can speed up, e.g. it running things from a file like bash does with .bashrc?
17:42 rebrec joined #git
17:44 orbyt_ To answer my own question, apparently using git stash apply on a dirty working directory is just fine. See: https://git-scm.com/book/en/v1/Git-Tools-Stashing
17:44 kexmex joined #git
17:44 thiago cahoots: what OS?
17:45 Atm0spher1c joined #git
17:47 cwre joined #git
17:47 tonyj joined #git
17:47 Puffball joined #git
17:47 Cabanossi joined #git
17:47 strk Seveas: git gc should drop those dangling, right ?
17:48 MineCoins joined #git
17:48 Levex joined #git
17:48 Jackneill joined #git
17:49 _ikke_ strk: Only if they are older then two weeks
17:49 Seveas strk: not necesseraily
17:49 cahoots thiago, os x
17:49 Seveas they should not appear in any reflog (or be reachable by something in the reflog), and not be too new (default is 2 weeks, as _ikke_ says)
17:49 thiago cahoots: you can try to dtruss th execution to see what it is doing in that time
17:50 Seveas cahoots: git add --verbose does a git diff. This can take a while in large worktrees.
17:50 TheSimonator joined #git
17:50 eloquenza joined #git
17:51 eehackr joined #git
17:52 ok91 joined #git
17:55 gnixev joined #git
17:55 orbyt_ joined #git
17:56 rominronin joined #git
17:56 xanadu joined #git
17:57 stepho joined #git
17:58 TheSimonator joined #git
17:58 nidr0x joined #git
18:01 romank joined #git
18:02 hobodave joined #git
18:02 Sasazuka joined #git
18:03 hahuang65 joined #git
18:04 livingstn joined #git
18:05 dreiss joined #git
18:09 maarhart joined #git
18:11 tvw joined #git
18:15 crose joined #git
18:16 mar77i joined #git
18:17 diogenese joined #git
18:18 eloquenza joined #git
18:20 _ikke_ rip gitinfo
18:21 xall joined #git
18:22 eloquenza joined #git
18:22 misosoup joined #git
18:24 raffs joined #git
18:24 joshszep_ joined #git
18:24 Vampire0 _ikke_, I guess gitinfo runs on the same host jast is connecting from. He also reconnected 2 minutes ago
18:25 _ikke_ right, makes sense
18:26 joshszep joined #git
18:26 Dougie187 joined #git
18:26 AaronMT joined #git
18:27 SkyPatrol joined #git
18:29 notebox joined #git
18:30 ahrs joined #git
18:31 misosoup joined #git
18:31 NeXTSUN joined #git
18:33 TbobbyZ_ joined #git
18:33 nixjdm joined #git
18:34 Es0teric joined #git
18:34 ravndal joined #git
18:34 jast joined #git
18:35 Seveas fortunately we're all smart and can do without bot :)
18:35 shiriru joined #git
18:35 Vampire0 !no
18:36 _ikke_ what, I don't know anything
18:36 Pulp joined #git
18:36 Seveas !git
18:37 Gsham joined #git
18:37 gnixev joined #git
18:38 hseg joined #git
18:39 hseg Hi. git submodule status still prefixes - to the SHAs even after running git submodule init and git submodule sync
18:39 hseg Moreover, these two commands seem to not have gotten git to check out the relevant SHA
18:40 cahoots_ joined #git
18:41 Es0teric ok so i ran this command -- `git reset HEAD~` to go back a commit i had locally
18:41 Es0teric i did not push to origin
18:41 lindenle_ joined #git
18:41 _ikke_ right
18:41 hseg Forgot git submodule update
18:41 Es0teric so how do i go back up to the commit?
18:42 Es0teric because now my history is screwed up
18:42 _ikke_ git reset HEAD@{1}
18:42 tymczenko joined #git
18:42 _ikke_ (this uses the reflog to go back to the previous commit HEAD pointed to
18:42 _ikke_ )
18:42 pks_ joined #git
18:42 Es0teric shit
18:42 Es0teric _ikke_ now its worse =_=
18:43 jackrandom joined #git
18:43 marlinc joined #git
18:44 _ikke_ use git reflog to find the commit that you initially had checked out
18:44 Es0teric _ikke_ the problem is that i commited something
18:44 Es0teric before i did a pull
18:44 Es0teric so there was a conflict
18:45 Es0teric _ikke_ this is what my git status looks like now https://puu.sh/vQXWZ/18060d6deb.png
18:45 _ikke_ Es0teric: 'You are currently rebasing..'
18:45 Es0teric how do i get rid of that
18:45 _ikke_ git rebase --abort
18:46 Es0teric ok
18:47 Es0teric did that
18:47 Es0teric _ikke_ https://puu.sh/vQYaf/6fab257331.png
18:48 _ikke_ Ok, what do you want to achieve now?
18:48 davimore joined #git
18:49 gopar joined #git
18:49 sea-gull joined #git
18:50 Es0teric _ikke_ they diverged
18:50 Es0teric i want to fix that
18:50 Seveas merge or rebase
18:51 fees joined #git
18:52 madewokherd joined #git
18:55 h12o joined #git
18:57 Es0teric _ikke_, Seveas nvm my boss is fixing it
18:57 Es0teric thanks though
19:02 jast joined #git
19:03 MattMaker joined #git
19:03 gitinfo joined #git
19:07 jaziz joined #git
19:07 acidvegas joined #git
19:08 Snugglebash joined #git
19:09 gopar joined #git
19:11 ash_workz joined #git
19:11 crose joined #git
19:11 orbyt_ joined #git
19:12 LionsMane joined #git
19:12 ash_workz can I pull to overwrite (discard) changes?
19:13 j08nY joined #git
19:13 _ikke_ ash_workz: That's not what pull does
19:13 ash_workz yeah, I know... but thought maybe... :\
19:14 rxc joined #git
19:14 rxc joined #git
19:14 apotry joined #git
19:14 _ikke_ ash_workz: what do you want to achieve?
19:14 osse ash_workz: sounds like you want  git fetch; git reset --hard @{u};  but that's sort of destructive
19:14 LionsMane joined #git
19:15 hobodave_ joined #git
19:15 maxime_ joined #git
19:17 yqt joined #git
19:17 jstein_ joined #git
19:22 ash_workz osse: that's okay; I did something stupid on the server and I just want to push changes through my local machine and then pull them on the server
19:22 ash_workz stupid as in just modified code directly to see if it works
19:23 ash_workz I forgot what I was doing just now
19:23 Flelmo joined #git
19:23 gesh joined #git
19:23 ash_workz too much ☕
19:26 Sound joined #git
19:26 Sound joined #git
19:27 rominronin joined #git
19:27 gopar joined #git
19:28 Flelmo left #git
19:29 gesh left #git
19:31 kristofferR joined #git
19:32 kisonecat joined #git
19:32 jstein__ joined #git
19:32 Cabanossi joined #git
19:33 bitmod joined #git
19:33 tymczenko joined #git
19:34 nixjdm joined #git
19:34 leeN joined #git
19:35 gopar joined #git
19:36 Acerific joined #git
19:36 TbobbyZ joined #git
19:39 Phylock joined #git
19:41 gopar joined #git
19:41 TheSin joined #git
19:42 TheSin hey guys I'm trying to move a bitbucket repo to a local repo, I've got what I ened to do that, but how cause I do an authors file to link my local users to a commit like you can do when you migrate svn to git? or is that even possible?
19:43 netj joined #git
19:43 grawity probably not in the way that you're thinking
19:44 TheSin kk, I gues sit really shoudln't matter in the end just thought I'd ask before I import so that I don't mess it up ;)
19:44 grawity in git, commit authors aren't named after system accounts in the first place
19:44 ixian joined #git
19:44 grawity the github/bitbucket web UIs map the author emails to github accounts using their own user database, but it's a pretty UI-specific thing to do
19:45 BSAlb joined #git
19:45 grawity as far as git cares, the author is "Fred Foobar <fred@example.com>"
19:45 PHPanos joined #git
19:45 grawity now, it is possible to
19:45 TheSin right I'm I'm doing bitbucket to gitlab which is kinda the same but the gui version uses OAuth to do it and I don't want to set that up just for 4 imports ;)
19:45 grawity 1) rewrite old commits and replace garbage authors (like "jim@localhost")
19:46 meLon joined #git
19:46 grawity or 2) add a .mailmap file to help UIs associate those with more up-to-date info
19:46 Eugene TheSin - git repo to git repo? no rewriting or special stuff is needed.
19:46 grawity #2 is generally recommended unless you're already rewriting the repo
19:46 TheSin K thanks guys, I just wont' worry about it then
19:47 Eugene In each repo you want to migrate, do: `git remote add gitlab <gitlab URL>:myrepo/foo.git; git push --mirror gitlab`
19:47 Gurkenglas joined #git
19:47 Eugene This will push everything i your current repo to that new one
19:47 strk is there any way to reconstruct git-svn commit mappings from metadata ?
19:47 Eugene strk - Sure, polish up on your sed skillz ;-)
19:48 strk really ? no git-svn command for that ?
19:48 rorro joined #git
19:49 Yondering joined #git
19:49 Eugene I don't know what metadata you mean. Generally creating an authors.txt is not that hard
19:49 nocbreak joined #git
19:49 grawity probably the svn url that git-svn adds to every commit message
19:49 TheSin okay so one more question if i already have a clone, is there a way to convert it to --mirror or do I need to reclone?
19:49 eehackr joined #git
19:49 Es0teric joined #git
19:49 grawity TheSin: something roughly like https://github.com/grawity/bin/blob/master/git-mkbare
19:50 grawity frankly, just reclone, it's more reliable
19:50 roelmonnens joined #git
19:50 MarioBranco joined #git
19:50 TheSin kk thanks again
19:50 strk Eugene: the problem is I changed authors.txt over time
19:50 Eugene That looks smarter than what I was going to suggest, heh. I would also just make a new `git clone --bare <bitbucket URL>`
19:50 strk so now a single authors.txt cannot recreate the git history
19:50 Eugene strk - sounds like a problem with no solution. Sorry, but thats a crazy edge case
19:51 strk the metadata I mean is the  one that tells `git svn fetch` what's left to fetch
19:51 strk as it currently restarts from r1
19:51 Eugene You should treat authors.txt entries as immutable after addition to avoid this
19:51 strk didn't really sound crazy to me, but I dunno what kind of metadata does git-svn really need
19:52 aard_ joined #git
19:52 strk or, to be more clear, I dont' see why those `    git-svn-id: http://svn.osgeo.org/postgis/trunk@15390 b70326c6-7e19-0410-871a-916f4a2858ee` lines are added in git history if they have no purpose
19:53 Eugene That is informational. authors.txt is the metadata
19:54 MarioBranco joined #git
19:54 Eugene It is used for rev_map stuff, but if your authors.txt has changed then by definition so has your metadata
19:54 Eugene Soooooo
19:54 hahuang65 joined #git
19:55 jnavila joined #git
19:56 blunaxela joined #git
19:56 Dougie187 joined #git
19:57 cbreak joined #git
19:57 strk Soo, my only chance is finding all the changes of authors file, then write an authors-script that based on revision decides which version to use
19:57 strk does the authors-script receive information about the SVN revision the change comes from ?
19:57 strk or can it access it anyhow ?
19:58 h12o joined #git
19:58 strk (crazier solution)
19:58 strk --authors-prog=<filename>, that is
20:01 mar77i_ joined #git
20:02 gopar joined #git
20:04 finalbeta joined #git
20:04 TheSimonator joined #git
20:05 strk doesn't look authors-prog receives anything many interesting things
20:05 strk just a name
20:07 gopar joined #git
20:08 TheSin left #git
20:08 exotime joined #git
20:08 m0viefreak joined #git
20:13 tango_ joined #git
20:13 gopar joined #git
20:14 pbandark joined #git
20:16 strk I seem to be on the right track, just need to tweak off-by-one errors
20:17 cdg joined #git
20:18 jrahmy_ joined #git
20:20 tymczenko joined #git
20:25 arecaceae joined #git
20:25 damian_ joined #git
20:26 ResidentBiscuit joined #git
20:27 strk I cannot pass --reference to git-fetch, nor git-init, any way to add a local reference after init ?
20:28 AfroThundr joined #git
20:28 gopar joined #git
20:29 AfroThundr joined #git
20:30 strk what's a good --pretty=format:XXXX to retrive svn revision ?
20:30 strk to replace git show ${hash} | grep git-svn-id | cut -d@ -f2 | awk '{print $1}'`
20:30 joshszep joined #git
20:31 publio joined #git
20:31 joeco joined #git
20:32 Darcidride joined #git
20:33 nixjdm joined #git
20:34 raijin joined #git
20:34 meLon joined #git
20:35 jast strk: unfortunately there's no format token for it, because the svn ID is part of the commit message, not a separate field in the commit object
20:36 joshszep joined #git
20:36 strk I could at least reduce the input passed to grep by only retriving the commit log
20:37 strk how to extract just that ?
20:37 jast %b
20:37 jast the full list of placeholders is in man git-rev-list
20:37 gitinfo the git-rev-list manpage is available at https://gitirc.eu/git-rev-list.html
20:37 jast section "PRETTY FORMATS"
20:39 gopar joined #git
20:40 brent__ joined #git
20:40 strk git log -1 --pretty=format:%b 7
20:40 strk seems to work, with git-svn-id being also last line (which could be another thing to consider)
20:40 strk git log -1 --pretty=format:%b 7ca7d1b | tail -1
20:42 armyriad joined #git
20:42 jast since it's always the last line and you're looking at one commit only, how about this:     git cat-file commit 7ca7d1b | tail -1
20:42 _ADN_ joined #git
20:42 jast less overhead and all that
20:43 strk interesting
20:43 leeN joined #git
20:44 gopar joined #git
20:45 acidvegas joined #git
20:45 jast it's strange that git-svn doesn't have a simple subcommand to return the svn rev for a commit...
20:47 strk I really hope my trick to track authors.git changes works, but I'm afraid it won't be so easy
20:47 strk that file changed ~9 times
20:48 strk git log --reverse --pretty=format:%h svn-trunk authors.git  # 8 times, exactly
20:48 strk but I'm not sure that my mirroring machinery always tracked it as a symlink to the checkout rather than as a copy
20:48 strk for sure the initial one had to be a copy
20:48 strk but after that, it was probably at one point turned into a symlink
20:49 strk and initially the mirror used to use "git svn rebase" which would have updated the authors.git as changed in the repo
20:49 strk but later changed to a "git svn fetch" which would not change authors.git :(
20:49 Altiare joined #git
20:49 strk so this is likely going to be a manual trial & error to find the changes in authors.git (reflected in changes in git history hashes)
20:49 strk and then somehow find the correct version of authors.git to use for the specific range of revisions
20:50 strk and somehow encode all this craziness into a script :'(
20:50 cbreak joined #git
20:53 D630 joined #git
20:53 askb joined #git
20:54 raijin joined #git
20:55 dave0x6d joined #git
20:58 stepho joined #git
20:58 eehackr joined #git
20:58 rominronin joined #git
20:59 rebrec joined #git
21:00 durham_ joined #git
21:02 oleksiyp joined #git
21:02 orbyt_ joined #git
21:02 strk bah, big mess, not even history is the same betwene the mirror and svn/trunk remote, it's not just authors.txt that's being a mess
21:04 devvie joined #git
21:04 rebrec joined #git
21:05 ertes joined #git
21:06 sea-gull joined #git
21:06 mrcrjs joined #git
21:10 epsi10n joined #git
21:12 dsdeiz joined #git
21:12 dsdeiz joined #git
21:14 grumble joined #git
21:14 anuxivm joined #git
21:16 gopar joined #git
21:17 jaziz joined #git
21:17 gopar joined #git
21:18 awc737 joined #git
21:19 dtrott joined #git
21:21 gopar joined #git
21:22 JanC_ joined #git
21:23 raijin joined #git
21:25 TbobbyZ_ joined #git
21:25 hrehf joined #git
21:25 durham joined #git
21:26 orbyt_ joined #git
21:27 durham_ joined #git
21:27 AfroThundr joined #git
21:27 mrcrjs joined #git
21:27 rebrec joined #git
21:28 overlord_tm joined #git
21:28 diogenese joined #git
21:29 stepho joined #git
21:30 AfroThundr joined #git
21:32 AfroThundr joined #git
21:33 nixjdm joined #git
21:36 gopar joined #git
21:41 raspado joined #git
21:44 gopar joined #git
21:48 Cabanossi joined #git
21:48 gopar joined #git
21:50 gopar joined #git
21:52 aidalgol joined #git
21:52 P4nther joined #git
21:55 mattcen joined #git
21:55 irco joined #git
22:01 dreiss joined #git
22:01 Dougie187 left #git
22:02 Derperperd joined #git
22:02 gopar joined #git
22:03 msfh joined #git
22:05 blackwind_123 joined #git
22:06 YuGiOhJCJ joined #git
22:07 joshszep joined #git
22:08 rkazak joined #git
22:08 joshszep joined #git
22:08 webdev007 joined #git
22:09 joshszep joined #git
22:09 Sna4x8 joined #git
22:10 joshszep joined #git
22:11 joshszep joined #git
22:11 joshszep joined #git
22:12 Mitt joined #git
22:12 joshszep joined #git
22:17 oskarkv joined #git
22:17 Yondering joined #git
22:19 cagedwisdom joined #git
22:20 caper joined #git
22:21 dreiss joined #git
22:22 masuberu joined #git
22:24 Masber_080 joined #git
22:26 ToBeCloud joined #git
22:26 theiostream joined #git
22:26 TheFuzzball joined #git
22:27 kurdzman joined #git
22:27 Masber joined #git
22:28 tubaman joined #git
22:29 lxnr joined #git
22:29 lagothrix joined #git
22:30 rominronin joined #git
22:30 cstratak joined #git
22:32 orbyt_ joined #git
22:33 dsdeiz joined #git
22:33 dsdeiz joined #git
22:33 tubaman joined #git
22:34 tubaman joined #git
22:35 tubaman joined #git
22:35 chrysa joined #git
22:38 Aoaz joined #git
22:38 lxnr left #git
22:40 jstimm joined #git
22:43 Hunter1 joined #git
22:46 theoceaniscool joined #git
22:47 duderono_ joined #git
22:48 nidr0x joined #git
22:48 theiostream joined #git
22:48 mentazoom joined #git
22:49 NeXTSUN joined #git
22:50 Noldorin joined #git
22:50 mgoodwin joined #git
22:51 mgoodwin Hi, can someone explain to me why a simple `git status` causes a checkout to re-checkout all files on a custom work-tree? Example: https://hastebin.com/raw/tolemenama
22:51 Levex joined #git
22:53 TheFuzzball joined #git
22:56 s17 because you did not specify the work-tree
22:57 mgoodwin Yes that makes sense, so it's recording the work-tree location
22:58 mgoodwin I'm asking from the frame of mind of a deploy script that is somewhat outside of my control (gitlab-ci) and therefore it's doing something like this when I checkout -f and causes it to always fully checkout
22:58 Eugene Elon is lobbing another one at the stars in 20 minutes https://youtu.be/ynMYE64IEKs
22:58 s17 man page says you can set GIT_WORK_TREE environment variable
22:59 mgoodwin That's one thought, set that in the environment and let the CD handle it
22:59 mgoodwin but I'm also very curious what git status is "resetting"
22:59 mgoodwin or is it any command without a specified --work-tree?
23:00 s17 that's what I'm thinking
23:00 mgoodwin I guess I can diff the git dir to find out exactly
23:00 s17 I was just telling you my thoughts.  I'm not sure if I am correct about it
23:00 joshszep joined #git
23:01 gry hi, i've added a remote and it's a few commits ahead; how do i view each of them and approve or reject? is there a standard comfortable process for this? thanks in advance.
23:01 mgoodwin Well if the directories differ before and after, then you must be :)
23:02 gunnaro joined #git
23:02 mikecmpbll joined #git
23:02 tourdownunder joined #git
23:03 Cabanossi joined #git
23:03 mgoodwin .git/index differs after status
23:03 kfoonamalik joined #git
23:04 h12o joined #git
23:05 oskarkv joined #git
23:07 dsdeiz joined #git
23:07 dsdeiz joined #git
23:07 orbyt_ joined #git
23:08 Eugene gry - `git fetch` will grab new changes from a remote. Various commands can be used to examine these, most interesting are probably `git log`, `git show`, `git diff`, and `git checkout`
23:11 safe joined #git
23:12 madewokherd joined #git
23:13 mgoodwin So now that the abovementioned behavior is a little more wonky than I would have liked - the underlying question is raised which is - is checkout with a worktree the most effective way to synchronize a webroot or should i be using rsync?
23:14 thiago joined #git
23:15 diogenese joined #git
23:15 netj joined #git
23:16 Hello71 joined #git
23:18 TheFuzzball joined #git
23:19 DaGoaty joined #git
23:21 Hello71 !deploy
23:21 gitinfo Git is not a deployment tool, but you can build one around it (in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/deploy.html
23:21 cagedwisdom joined #git
23:25 Eugene I like rsync
23:26 xall joined #git
23:26 Eugene `git archive` is the most reliable way to object a copy of the worktree; merging/replacing your webapp with this is an exercise left up to the reader
23:26 DaGoaty joined #git
23:26 hashtagstrashtag joined #git
23:27 Hunter1 joined #git
23:29 mgoodwin The drawback with archive is that it doesn't handle deletions
23:29 mgoodwin ... to my knowledge
23:29 mgoodwin I like that checkout only operates on files git knows about
23:30 Eugene Yup. You would need to handle that logic yourself.
23:30 mizu_no_oto joined #git
23:33 wjwu joined #git
23:35 shinnya joined #git
23:37 stepho joined #git
23:39 CussBot joined #git
23:40 Hello71 joined #git
23:41 mostlybadfly joined #git
23:42 Hello71 joined #git
23:42 plos joined #git
23:43 Gsham joined #git
23:46 justan0theruser joined #git
23:46 mel00010 joined #git
23:49 Stumbler joined #git
23:52 prg3 joined #git
23:52 dsdeiz joined #git
23:52 dsdeiz joined #git
23:54 Hunter1 joined #git
23:56 raijin joined #git
23:56 eehackr joined #git
23:58 nitric joined #git

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