Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-08-18

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

All times shown according to UTC.

Time Nick Message
00:01 cagomez joined #git
00:04 Gitzilla joined #git
00:04 mingrammer joined #git
00:05 vassagus joined #git
00:06 yoan joined #git
00:07 cqi joined #git
00:09 beyertech joined #git
00:09 Gitzilla joined #git
00:10 beyertech joined #git
00:13 glebihan_ joined #git
00:16 MattMaker joined #git
00:17 Gr33nW0lf joined #git
00:17 geko joined #git
00:19 sagerdearia joined #git
00:19 linuxmint joined #git
00:20 Gr33nW0lf joined #git
00:21 Finnoloid joined #git
00:21 Rodya_ joined #git
00:23 d0nn1e joined #git
00:25 jeffreylevesque joined #git
00:26 mniip joined #git
00:27 mostlybadfly joined #git
00:28 HoierM joined #git
00:28 cqi joined #git
00:33 weox joined #git
00:34 vassagus joined #git
00:37 bambams Hmmmm, wouldn't `git fetch origin master' only store the result in FETCH_HEAD, not updating refs/remotes/origin/master? There's no <dst> part of the refspec.
00:38 bambams `git fetch origin +refs/heads/master:refs/heads/master && git diff HEAD origin/master' should work though. But you could also just compare FETCH_HEAD right after fetching a specific ref.
00:39 Gamecubic joined #git
00:41 constant joined #git
00:42 danimal2 joined #git
00:43 linuxmint joined #git
00:44 hahuang61 joined #git
00:51 Rodya_ joined #git
00:53 durham joined #git
00:53 CihanKaygusuz joined #git
00:53 skylite joined #git
00:54 tesuji joined #git
00:55 meanderer joined #git
00:59 steven_a_s joined #git
00:59 fahadash joined #git
00:59 Jon28 joined #git
01:01 thiago joined #git
01:04 rorro joined #git
01:05 fish3485 joined #git
01:05 b1tchcakes joined #git
01:05 malabaris joined #git
01:06 malabaris left #git
01:06 fish3485 joined #git
01:08 rkazak joined #git
01:08 pks joined #git
01:09 zumba_addict joined #git
01:10 geko joined #git
01:10 d^sh joined #git
01:12 urbanmonk joined #git
01:12 atomicb0mb joined #git
01:13 urbanmonk left #git
01:13 urbanmonk joined #git
01:14 cyphase joined #git
01:15 vassagus joined #git
01:17 urbanmonk joined #git
01:18 cloud_forecast joined #git
01:18 tanuki joined #git
01:18 zacts joined #git
01:19 stonerfish joined #git
01:20 Flattened joined #git
01:22 Flattened I fail to understand  something. git is supposed to be on a file level version control or is supposed to be on a content level?
01:23 RustyShackleford joined #git
01:23 Flattened sorry for my english. I am not a native speaker
01:24 Flattened I mean the difference between commits is on the number of files (components, headers, implementations, etc) or is it on the content of the files?
01:24 Flattened It seems like it is going to be messy if it is about the content of the files
01:24 constant joined #git
01:25 Flattened When you merge, you are supposed to merge files on different branches so that their content is also merges!!!? isn't this a bit out of control when dealing with many files?
01:25 Dougie187 joined #git
01:26 Vampire0 joined #git
01:29 rchavik joined #git
01:34 linuxmint joined #git
01:35 Gr33nW0lf joined #git
01:35 Rodya_ joined #git
01:37 urbanmonk joined #git
01:46 arescorpio joined #git
01:46 Genome36 joined #git
01:46 eka joined #git
01:47 cyan__ joined #git
01:47 kn-928 joined #git
01:48 ckubrak joined #git
01:48 wandererX joined #git
01:49 ckubrak Hi guys, if I have the commit ID (the short version), how can I see the difference between an old commit and the actual?
01:49 vassagus_ joined #git
01:50 wandererX when you merge between two branches, and when both branches have the same files a,b,c, and d. If the content of the files is different then a conflict occurs. Does that mean identical file names on both two branches should be the same? and that their content should also be identical?
01:50 Manuel15 joined #git
01:50 Manuel15 joined #git
01:50 beyertech joined #git
01:50 wandererX That changes actually occur on file level instead of file content leve?
01:50 Thorn joined #git
01:51 wandererX Meaning, when you add new features, you add them as new files?
01:51 CrazyBonz joined #git
01:52 PaulCapestany joined #git
01:52 Genome36 hello guys
01:52 mingrammer joined #git
01:53 Genome36 so i had used the command filter-branch to create a submodule from a folder
01:53 Genome36 it all works fine
01:53 Genome36 but now i need to create another submodule from FILES not folder
01:53 Genome36 http://stackoverflow.com/questions/1593620/how-to-make-part-of-an-existing-git-repository-a-submodule
01:54 AbstractArtist joined #git
01:54 Genome36 how can i do it ?
01:54 ResidentBiscuit joined #git
01:55 CihanKaygusuz left #git
01:55 alansaul joined #git
01:55 Sasazuka_ joined #git
01:55 Goplat joined #git
01:58 ahr3n joined #git
02:04 vaibhavsagar joined #git
02:06 checkItOut_ joined #git
02:07 crayon joined #git
02:07 Cihan left #git
02:09 darkbit joined #git
02:09 ResidentBiscuit joined #git
02:09 soLucien joined #git
02:10 soLucien hello guys ! as far as i can see , it is possible to use the git:// protocol on a local repository
02:10 soLucien so let me go one step forward from that
02:10 justanotheruser joined #git
02:11 soLucien is it possible to get the latest files (which are uncommitted) from it? or would this simply fetch head
02:11 chachasmooth joined #git
02:12 dreiss joined #git
02:12 dhollinger joined #git
02:13 yoh joined #git
02:14 snowkidind joined #git
02:18 yoh I experience "error: Unable to find ... under http://..."  I have tried everything (sure thing running update-server-info on server), e.g. removing .temp files under .git/objects on receiving side etc -- just doesn't work (FWIW: fresh clone works).  Any clues on what could be going wrong?
02:22 yoh oh hoh -- I thought I did try it before -- apparently if used more recent git (2.7.0) within sid env, instead of 2.1.4 from jessie, it worked out just fine!
02:22 fakenerd joined #git
02:23 * yoh often finds useful to complaint to get problems resolved automagically ;)
02:23 checkItOut joined #git
02:24 kn-928 joined #git
02:30 Gr33nW0lf joined #git
02:31 ASOLAr joined #git
02:32 rgrinberg joined #git
02:33 MattMaker joined #git
02:33 Gr33nW0lf joined #git
02:34 lagothri1 joined #git
02:35 hexagoxel joined #git
02:35 fakenerd joined #git
02:36 vaibhavsagar joined #git
02:38 thiago joined #git
02:47 mingrammer joined #git
02:54 tgunr joined #git
02:58 fozniku joined #git
03:00 krs93_ joined #git
03:01 lightstalker joined #git
03:02 aavrug joined #git
03:02 fstd_ joined #git
03:02 geko joined #git
03:10 ash_workz joined #git
03:12 cqi joined #git
03:13 i7c joined #git
03:18 watabou joined #git
03:22 filcab joined #git
03:24 nettoweb joined #git
03:25 ShekharReddy joined #git
03:26 l4v2 joined #git
03:31 Dougie187 left #git
03:32 hahuang61 joined #git
03:34 mehola joined #git
03:37 cqi joined #git
03:37 MattMaker joined #git
03:39 chandu20 joined #git
03:43 ahr3n joined #git
03:46 arescorpio joined #git
03:46 fahadash joined #git
04:00 thecomedian joined #git
04:04 nettoweb joined #git
04:06 Derperperd joined #git
04:07 lachm joined #git
04:12 HappyLoa- joined #git
04:13 mehola joined #git
04:14 constant joined #git
04:15 hjubal joined #git
04:15 constant joined #git
04:16 constant joined #git
04:16 constant joined #git
04:18 cyan__ joined #git
04:20 Cabanossi joined #git
04:23 CrazyBonz joined #git
04:31 mmlb joined #git
04:31 ytl joined #git
04:32 Dschingiz left #git
04:32 Spydar007 joined #git
04:32 Spydar007 joined #git
04:33 hahuang61 joined #git
04:35 mingrammer joined #git
04:36 khmarbaise joined #git
04:36 hahuang65 joined #git
04:36 Sasazuka joined #git
04:37 HappyLoa- joined #git
04:39 checkItOut joined #git
04:40 Eugene joined #git
04:40 fozniku joined #git
04:47 johnny56_ joined #git
04:51 dandaman joined #git
04:51 dandaman hey there
04:52 dandaman can anyone help me out with my .gitignore file? https://gist.github.com/dsauerbrun/de8d62be69fa452810dd9ce80130425f
04:52 dandaman for some reason the files in my vendor/gmapsjs and vendor/highcharts folders aren't showing up when i run a git status
04:52 dandaman that .gitignore file is sitting at the root of my git repo
04:52 stonerfish joined #git
04:52 dandaman https://github.com/dsauerbrun/climbcation-front there's the repo
04:55 i7c dandaman, try vendor/* and !vendor/gmapjs
04:56 dandaman i7c: no luck, however, when i do git status it shows vendor/ as a folder i need to add
04:56 i7c yep
04:56 dandaman but not vendor/gmapsjs/gmaps.js
04:56 i7c That’s normal.
04:56 i7c Just run git add vendor/gmapjs/gmaps.js anyway
04:57 dandaman ahh ok you're right
04:57 dandaman thanks!
04:57 johntramp joined #git
04:57 i7c :)
04:58 i7c git doesn’t really track directories so that hint means somehow that nothing under vendor/ is tracked yet.
04:59 cagedwisdom joined #git
04:59 i7c Not that it sounds intuitive. :D
05:00 sagerdearia joined #git
05:02 dhollinger joined #git
05:03 leeN joined #git
05:03 glebihan joined #git
05:05 _ikke_ It just summarizes everything as one dir
05:06 _ikke_ git status -uall
05:06 _ikke_ git status -u all
05:07 govg joined #git
05:08 i7c You decorrected, _ikke_.
05:09 i7c Imprworsified.
05:09 _ikke_ yeah, was doubting
05:09 _ikke_ Only parameter that break convention iirc
05:10 _ikke_ Though, in this case, -u would have worked (default to all)
05:10 dhollin3 joined #git
05:10 vassagus joined #git
05:10 i7c I don’t quite get why break convention. :)
05:10 i7c Was it an accident?
05:15 _ikke_ I don't know, possibly something historical
05:16 thiago joined #git
05:23 bocaneri joined #git
05:30 mingrammer joined #git
05:33 zumba_addict joined #git
05:39 MattMaker joined #git
05:40 mariuscc joined #git
05:43 jstimm joined #git
05:44 checkItOut joined #git
05:45 checkItOut joined #git
05:47 average joined #git
05:47 average is there any way to get a partial-history git clone ?
05:47 average http://stackoverflow.com/a/2587244/827519 ?
05:48 Raging_Hog joined #git
05:53 AbstractArtist joined #git
05:53 i7c yes, the --depth option is probably what you want.
05:57 average i7c: that worked pretty well, yeah
05:57 xall joined #git
05:57 average i7c: i wonder if I can tell PHP's composer to use that
05:59 rscata joined #git
05:59 circ-user-RNfpp joined #git
05:59 LucaTM joined #git
06:02 Manuel15 joined #git
06:02 Manuel15 joined #git
06:03 zeroed joined #git
06:05 i7c I wouldn’t know anything about PHP, sorry.
06:05 rossome joined #git
06:05 average ok
06:07 fakenerd joined #git
06:08 hahuang65 joined #git
06:09 ASOLAr joined #git
06:11 eivarv joined #git
06:13 netj joined #git
06:14 ncd|leeN joined #git
06:24 stuh84 joined #git
06:25 roelmonnens joined #git
06:25 jagob joined #git
06:27 nilg joined #git
06:29 rkazak joined #git
06:29 jknetl joined #git
06:33 vassagus joined #git
06:33 mrkurtz joined #git
06:34 rkazak joined #git
06:42 qt-x joined #git
06:44 dsantiago joined #git
06:50 submitnine joined #git
06:55 MattMaker joined #git
06:56 MattMaker joined #git
06:56 MattMaker joined #git
06:57 Sceorem joined #git
06:59 ronskisika joined #git
07:01 eivarv joined #git
07:04 encod3 joined #git
07:09 Hounddog joined #git
07:10 vassagus_ joined #git
07:11 checkItOut joined #git
07:12 jceb joined #git
07:16 Derperperd joined #git
07:17 a_thakur joined #git
07:17 a_thakur joined #git
07:18 mingrammer joined #git
07:20 saedelaere joined #git
07:21 ki0 joined #git
07:21 MattMaker joined #git
07:22 ealvarez joined #git
07:24 hahuang65 joined #git
07:26 ASOLAr joined #git
07:27 Balliad joined #git
07:31 nilg` joined #git
07:31 zincrokx joined #git
07:32 zincrokx joined #git
07:33 zincrokx joined #git
07:33 chandu20 joined #git
07:35 rts-sander joined #git
07:36 rts-sander Hey I do: git checkout -p,  answer y on some hunks, but it doesn't change!
07:37 jagob joined #git
07:38 anttim joined #git
07:38 weox joined #git
07:38 t0by joined #git
07:40 dimi1947 joined #git
07:40 lb rts-sander !repro
07:40 gitinfo [!transcript] Please paste (using https://gist.github.com/ or https://vomitb.in/ ) a transcript ( https://gist.github.com/2415442 ) of your terminal session so we can see exactly what you see
07:42 yuhlw_ joined #git
07:43 vassagus joined #git
07:44 xaa joined #git
07:47 dreiss joined #git
07:49 jceb joined #git
07:51 chandu20 joined #git
07:51 ahmedelgabri joined #git
07:51 TomyWork joined #git
07:54 Raging_Hog joined #git
07:55 rads joined #git
07:55 Mikerhinos joined #git
07:58 ytl joined #git
08:00 aidalgol joined #git
08:03 ytl joined #git
08:05 bongjovi joined #git
08:08 ASOLAr joined #git
08:12 mingrammer joined #git
08:13 tomog999_ joined #git
08:13 eijk_ joined #git
08:14 kurkale6ka joined #git
08:16 anttim joined #git
08:18 leeN joined #git
08:19 inflames joined #git
08:22 mikecmpbll joined #git
08:24 cyan__ joined #git
08:24 GavinMagnus joined #git
08:24 chandu20 joined #git
08:29 afuentes joined #git
08:29 paws_ joined #git
08:31 lordjancso joined #git
08:32 GenZai joined #git
08:32 jceb joined #git
08:32 anttim joined #git
08:33 Andrew_K joined #git
08:34 mahendra_ joined #git
08:34 mrkat joined #git
08:35 jstein_ joined #git
08:37 Rish joined #git
08:39 Whitelye joined #git
08:44 bruce_lee joined #git
08:44 bruce_lee joined #git
08:44 eivarv joined #git
08:45 boltR joined #git
08:50 gsingh93 joined #git
08:51 venmx joined #git
08:52 netcarver joined #git
08:52 konrados joined #git
08:53 ikmaak joined #git
08:54 davisonio joined #git
08:57 venmx joined #git
09:00 ifra201 joined #git
09:01 a_thakur joined #git
09:07 timmu joined #git
09:10 garethdaine joined #git
09:11 robotroll joined #git
09:13 fuchstronaut joined #git
09:13 fuchstronaut joined #git
09:14 ytl joined #git
09:17 joko joined #git
09:19 vassagus_ joined #git
09:20 joko Hello, could anyone tell me how to sync the branches of two remote repositories?
09:21 Seveas joko: fetch from the one, push to the other
09:22 jceb joined #git
09:23 joko Seveas: there was a way to do it more automatically, like grab the heads and push them, sth like that
09:23 joko The first remote repository is the original one, the second one is a fork
09:24 checkItOut joined #git
09:24 joko Branches on the first one are deleted and created all the time, I just need to sync them
09:24 tobiasvl !sync ?
09:24 gitinfo [!pull_upstream] To incorporate changes from an "upstream" or forked-from repository: git remote add upstream $URL; git fetch upstream; git merge upstream/$BRANCH. https://help.github.com/articles/syncing-a-fork/
09:25 checkItOut joined #git
09:25 joko tobiasvl: I was under the impression there was a better way
09:25 subhojit777 joined #git
09:26 subhojit777 joined #git
09:26 solars joined #git
09:28 solars if I rebased a local branch which was already pushed (so there is 1 incoming change as history differs) is push --force enough to rewrite the remote history?
09:30 notebox joined #git
09:31 osse yes
09:31 solars osse, thanks
09:31 xall joined #git
09:31 osse Woah, that's imprssive. About three minutes after you asked I say yes without highlighting you and you see it that quickly :O
09:32 solars haha
09:33 solars I'm using a tiling window manager, the window I'm working on atm is right next to the irc window :)
09:33 Ezzy joined #git
09:34 pcre joined #git
09:36 choki joined #git
09:37 davisonio joined #git
09:38 cqi joined #git
09:39 checkItO_ joined #git
09:41 checkItOut joined #git
09:47 ikmaak joined #git
09:49 vassagus joined #git
09:51 ilmari aah, debian finally has git 2.9.3, so I can enable diff.compactionHeuristic
09:52 i7c solars, --force-with-lease is generally safer than --force.
09:52 i7c Just an arbitrary hint.
09:52 artemisart joined #git
09:53 howdoi joined #git
09:53 * ilmari likes that magit makes -f a shortcut for --force-with-lease instead of --force
09:55 FunkyAss joined #git
09:56 jceb joined #git
09:58 a3Dman joined #git
09:58 eivarv joined #git
09:59 interrobangd joined #git
10:01 mingrammer joined #git
10:03 ikmaak joined #git
10:07 jeffreylevesque joined #git
10:08 16WAA1XTP joined #git
10:10 fyrelyte joined #git
10:11 alansaul joined #git
10:12 solars thanks i7c
10:13 anuxivm joined #git
10:15 Manuel15 joined #git
10:15 Manuel15 joined #git
10:18 teemo joined #git
10:18 tlaxkit joined #git
10:18 teemo I get this problem today when pushing, never seen this before http://paste.rory.sh/2z1u
10:19 tobiasvl teemo: it's a hook on the remote
10:20 tobiasvl so the remote repo is set up to run some ruby script to verify the payload before accepting it, and there seems to be an error in that script
10:21 m0viefreak joined #git
10:21 teemo this is just github... and I don't think we set up anything like that. I wouldn't even know how.
10:21 teemo Do you know where to check?
10:22 teemo huh weird I just tried again and it worked
10:22 teemo github snafu I guess
10:22 Zarthus remote: ruby-jemalloc: symbol lookup error: /data/github/current/vendor/gems/2.1.7/ruby/2.1.0/gems/json-1.8.3/lib/json/ext/generator.so: undefined symbol: rb_data_typed_object_alloc
10:22 Zarthus right there :)
10:22 teemo I don't work for github
10:22 teemo anyway it went away
10:22 Zarthus i'd still notify support of it
10:22 Zarthus they might appreciate the bug report
10:23 tobiasvl !sgc
10:23 gitinfo For bug reports, feature requests, suggestions for improvement, or specific problems with an individual repo (or gist, etc.), please contact support via support@github.com or https://github.com/support - support tries to answer all questions within 24 hours, including weekends and holidays.
10:23 teemo OK just did, thanks.
10:30 stamina joined #git
10:31 Hink joined #git
10:33 bartosz_ joined #git
10:33 ikmaak joined #git
10:34 beyertech joined #git
10:36 ytl joined #git
10:40 Diabolik joined #git
10:42 zeroed joined #git
10:42 zeroed joined #git
10:42 artemisart_ joined #git
10:42 eivarv joined #git
10:44 nebg joined #git
10:45 nebg hello everyone... is there some way with git to manage tasks to be done ?
10:47 _ikke_ Not within git, no
10:47 _ikke_ That doesn't mean there isn't a way to build something around git
10:48 submitnine joined #git
10:48 ikmaak joined #git
10:50 filcab joined #git
10:51 noidea joined #git
10:51 noidea joined #git
10:52 dodobrain joined #git
10:55 weckl joined #git
10:55 mingrammer joined #git
10:57 ikmaak joined #git
11:05 jstimm_ joined #git
11:05 ShekharReddy joined #git
11:06 fuchstronaut joined #git
11:09 mehola joined #git
11:10 vassagus joined #git
11:11 jstimm joined #git
11:13 lightstalker joined #git
11:17 ichux joined #git
11:17 ichux_ joined #git
11:18 ichux joined #git
11:19 Impaloo joined #git
11:19 govg joined #git
11:20 ichux_ joined #git
11:20 thm joined #git
11:21 grift joined #git
11:21 thm how can I do a cherry-pick of multiple commits, but combine them into a single commit?
11:22 ilmari thm: use -n, then commit manually
11:22 sdothum joined #git
11:23 niax do single cherry picks, then rebase and squash?
11:23 ilmari that works too
11:24 thm thanks
11:24 ichux_ joined #git
11:29 kus joined #git
11:29 nitin joined #git
11:29 nikivi joined #git
11:29 HoierM joined #git
11:32 jceb joined #git
11:34 davisonio joined #git
11:35 vassagus joined #git
11:36 Akimb joined #git
11:36 nettoweb joined #git
11:40 ichux_ joined #git
11:43 WayToDoor joined #git
11:43 mohabaks joined #git
11:45 tchan joined #git
11:46 EisNerd hi
11:46 xcesariox joined #git
11:46 WayToDoor joined #git
11:47 jimi_ joined #git
11:47 EisNerd is it possible to have in the local network a read mirror for a remote git to avoid unecessary external traffic and speeding up clones etc. and allow them to work when external connection is down
11:47 rorro joined #git
11:48 EisNerd and is there some documentation howto implement this (remote git is also our administrative domain, so using hooks etc is possible)
11:48 bremner EisNerd: just use a local clone?
11:49 anttim joined #git
11:49 bremner or set up some kind of mirroring using e.g. gitolite.
11:49 bremner not really sure what you want...
11:50 EisNerd we work on a large project, with an externally provided git lab, which should be obviously the authorative one
11:50 Blackb|rd joined #git
11:50 Sadeukko joined #git
11:51 EisNerd but I think it might be desirable to have a local pull url as our internal net is much faster
11:51 EisNerd so push to gitlab and pull local from http
11:52 Blackb|rd Hey, I have a quick question about git-attributes. I am doing analysis of a pretty large tree (both by file/dir count and commits, it's the Linux kernel tree) and it does *not* use git-attributes at all. I am doing a lot of "git show --short-stats --pretty=... <commit id>" calls
11:52 dsdeiz joined #git
11:52 nilg joined #git
11:52 Diabolik joined #git
11:52 Blackb|rd Unfortunately, that lstat()s every single dir in the tree for .git-attributes., making it very slow.
11:53 Blackb|rd Is there a way to tell git that I only want the builtin diffing stuff and don't care if .git-a defines something special?
11:54 EisNerd btw I came across a performance pitfall on windows, line ending normalisation, default all text files will be converted to crlf and back when git starts working with them. This gets a hell when you have 60000 and more text files in your project
11:55 j416 silly default
11:55 Blackb|rd The commands I'm running take ~10s on average, which seems quite slow, given this is not exactly a slow machine (I am working on spinning rust, though)
11:56 cyan__ joined #git
11:56 tomog999 joined #git
11:57 ljc joined #git
11:58 Sceorem joined #git
11:58 khmarbaise joined #git
11:59 robotroll joined #git
12:01 AbstractArtist joined #git
12:02 celdia joined #git
12:02 skalpin what's the proper way to get "some" changes from another branch. I don't want to merge because if I undo the changes I don't want, the next time the owner of the branch merges with develop it will undo his changes that I didn't want.
12:04 rnsanchez joined #git
12:04 eVolutive joined #git
12:04 * selckin brain explodes trying to parse that
12:05 skalpin selckin: sorry :(
12:05 elimode joined #git
12:05 i7c skalpin, you wanna have "some changes" in the form of code changes in your working directory or entire commits?
12:06 skalpin in the form of code changes in my working directory, I don't need the entire commits
12:06 Softeisbieger joined #git
12:06 ytl joined #git
12:06 i7c You can have entire files from a different commit: git checkout <commit> -- path/to/file
12:07 i7c Or you can use git checkout -p which lets you choose changes hunk by hunk.
12:07 Rodya_ joined #git
12:07 skalpin i7c: interesting, give me a minute to try checking out a file from a different commit
12:08 i7c Take your time.
12:08 synthroid joined #git
12:11 boyen joined #git
12:11 synthroid joined #git
12:13 flaguy48 left #git
12:15 fuchstronaut joined #git
12:16 govg joined #git
12:16 TooLmaN joined #git
12:17 rwb joined #git
12:18 skalpin i7c: that is awesome, thank you. I also tried out checkout -p and it looks like it get changes from the whole branch and puts it in my working directory. Is that correct?
12:18 drbean joined #git
12:18 davisonio joined #git
12:18 i7c uhm it’s supposed to be interactive.
12:19 i7c git checkout -p <branch>
12:19 skalpin it was interactive sorry
12:19 i7c ah
12:19 i7c you can limit it to files as well
12:19 i7c git checkout -p <branch> -- path/to/file/or/dir
12:19 i7c That way it only asks you about files in that path
12:20 skalpin and it's all the changes from the branch since we diverged which is good, is there a way I can auto accept all the changes?
12:20 stamina joined #git
12:21 i7c Well if you want aaaaall changes you can do git checkout <branch> -- .  while in the root of your repository. (note the dot)
12:21 i7c Also note, that it won’t delete any files even if they are missing in the other branch.
12:21 huptoka joined #git
12:22 skalpin i7c: perfect thank you. I need to read the manual for checkout. It seems more powerful than I though
12:22 skalpin thought*
12:22 huptoka how do you pull and merge? Tried git pull origin branch and now it says that my branch is diverged and I have  unmerged paths
12:23 i7c skalpin, sometimes it seems it does everything but checking out revisions! ;)
12:23 skalpin i7c: ha
12:24 i7c huptoka, that’s probably because you have merge conflicts. You should configure a merge tool of your liking (git config --global merge.tool <tool>) and fire it up (git mergetool)
12:25 huptoka i7c: no I just want to pull and forget my merge
12:25 huptoka I made my own branch, merged it with the one from origin then pulled, that's why it's complaining
12:25 huptoka then I will re-merge it with my branch
12:27 celdia joined #git
12:28 huptoka Anyone knows how to pull and forget my merge?
12:28 huptoka tried git pull "It's not possible because you have unmerged files"
12:28 i7c You cannot forget the merge, unless you mean you want to forget your changes completely.
12:29 huptoka I'm not forgetting any changes I said I will merge again
12:29 huptoka I have my changes in another branch all I have to do is to merge it
12:29 Softeisbieger joined #git
12:29 Ineentho[m] joined #git
12:30 portwolf joined #git
12:30 portwolf hello
12:30 gitinfo portwolf: 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.
12:31 mooncup hello
12:31 mooncup :(
12:31 mar77i joined #git
12:32 portwolf you look sad m8
12:32 i7c huptoka, you can undo a merge by resetting your branch to the commit before the merge.
12:33 i7c You best use !lol to inspect the branches, identify the right commit and do git reset --hard <commit>
12:33 gitinfo A nifty view of branches, tags, and other refs: git log --oneline --graph --decorate --all
12:35 FunkyAss joined #git
12:35 huptoka now it won't let me checkout to my modified branch, "you need to resolve your current index first" great
12:36 mooncup portwolf: the bot didn't say hello to me :(
12:36 huptoka1 joined #git
12:36 portwolf thats because you are not new here then :)
12:37 davisonio joined #git
12:37 i7c huptoka1, git merge --abort
12:38 i7c How rude.
12:38 huptoka joined #git
12:38 huptoka error: 'merge' is not possible because you have unmerged files.
12:38 Blackb|rd huptoka: "Hit alt-f4 to solce your git problems"
12:38 Blackb|rd solve*
12:38 i7c huptoka, how old is your git?
12:39 huptoka Blackb|rd: nice typo there pro person
12:39 huptoka ignored btw
12:39 Blackb|rd huptoka: I was not poking fun at you
12:39 huptoka i7c: what?
12:39 i7c merge --abort should work since 1.6 or so I think
12:39 i7c at least the interwebz say so
12:39 i7c too lazy to check :)
12:40 celdia joined #git
12:43 johnny56 joined #git
12:44 mingrammer joined #git
12:45 Impaloo joined #git
12:45 synchronous joined #git
12:46 vassagus_ joined #git
12:49 spiritz joined #git
12:50 flaguy48 joined #git
12:52 armin joined #git
12:54 Gamecubic joined #git
12:54 dimi1947 joined #git
12:55 KeelOfSteel joined #git
12:56 eivarv joined #git
12:56 jceb joined #git
12:57 rominronin joined #git
12:58 kadoban joined #git
13:01 Lucretia joined #git
13:01 f0rblaze joined #git
13:01 rgrinberg joined #git
13:03 anuxivm left #git
13:04 grayjoc joined #git
13:09 cdown joined #git
13:11 Lucretia hi, i'm looking for advice on managing a linux repo. I'm wanting to maintain, temporarily, a repo which contains certain patchsets and fixes. I have a fork of linus' tree from/on github and added another repo as a remote and linus' as upstream. the remote, called alex, after the name of the person, has a number of branches in which there are various versions of the kernel. I'd like to take the patchsets from those and rebase them onto my
13:11 Lucretia local branches for 4.6.6, 4.7.x, 4.8.x-rcy , etc. am I ok to do this? I keep seeing mentions of don't rebase other's public trees and don't back-merge, etc.
13:11 airdisa joined #git
13:12 zivester joined #git
13:12 buffal0 joined #git
13:13 CEnnis91 joined #git
13:14 GenZai joined #git
13:17 veilig2000 joined #git
13:17 fuchstronaut joined #git
13:19 buffal0 joined #git
13:19 JamesBaxter joined #git
13:19 cdg joined #git
13:19 DevAntoine joined #git
13:19 DevAntoine hi
13:20 sangy joined #git
13:20 ELFrederich joined #git
13:20 DevAntoine I've done multiple "work in progress" commits on a working branch. Now I want to merge some of the files on this branch to develop. Is it doable? If so, how?
13:20 ELFrederich are there any scripts out there which re-write history as if a particular directory never existed?
13:21 compeman joined #git
13:21 buffal0 joined #git
13:22 Diabolik left #git
13:22 bremner git filter-branch can do that.
13:23 Blackb|rd ELFrederich: git filter-branch --tree-filter 'rm filename' HEAD would make it appear that "filename" never existed
13:23 Blackb|rd ELFrederich: this can be adjusted for whole directory hierarchies, I'd guess
13:23 ELFrederich Blackb|rd: and it would keep the committer info, correct dates and times, etc?
13:24 Blackb|rd ELFrederich: only if the commit being filtered still has actual changes
13:25 bremner ELFrederich: what it won't do is preserve commit hashes.
13:25 Blackb|rd ELFrederich: if it was only editing files under the deleted dir, the commit would be dropped entirely
13:25 ELFrederich Blackb|rd: okay... so I'd run that command on each branch that contains history with that directory
13:25 Blackb|rd And yes, as bremner mentioned, all of your refs are going to be different
13:25 zeroed joined #git
13:25 zeroed joined #git
13:25 ELFrederich bremner: yeah, I understand
13:26 sonOfRa joined #git
13:27 ljc joined #git
13:28 roelmonnens joined #git
13:30 huptoka left #git
13:30 jaafar joined #git
13:31 happy-dude joined #git
13:31 JamesBaxter joined #git
13:31 ExoUNX joined #git
13:31 ExoUNX joined #git
13:32 eivarv joined #git
13:34 ki0_ joined #git
13:34 s0m30n3ud0ntkn0w joined #git
13:34 SupaHam joined #git
13:35 rnsanchez_ joined #git
13:35 mgedmin joined #git
13:36 mgedmin I'm having trouble coming up with a git alias to show me the diff of all changes made on a topic branch (that's already been merged into master)
13:36 ki0 joined #git
13:36 mgedmin I can assume a non-ff merge
13:36 mgedmin and I can assume the branch ref still points to the last commit made on that branch prior to the merge
13:38 osse mgedmin: git diff master...branch
13:38 osse triple dot = profit
13:38 buffal0 joined #git
13:38 mgedmin tripple dot produces an empty diff (probably because the branch's already been merged?)
13:39 AaronMT joined #git
13:39 osse oh, right
13:39 osse you'd rather not depend on the merge commit etc.?
13:39 BigWillie joined #git
13:39 mgedmin I want 'git show-changes-on <branch-name>'
13:39 osse if you know the merge commit then it's  git diff foo^1...foo^2
13:39 mgedmin without having to look up the merge commit hash manually
13:40 ichux_ joined #git
13:40 osse Kiff, we have a conundrum!
13:40 alansaul joined #git
13:41 AaronMT joined #git
13:41 vuoto joined #git
13:41 mgedmin hmm, one stackoverflow answer involves $(git log --grep="^Merge branch '$1'" master)
13:41 mgedmin which is ICK but maybe it could work
13:42 mgedmin ah the joys of working with gitolite with no comfortable web UI
13:42 eVolutive joined #git
13:42 cdown_ joined #git
13:43 jast comfortable web UI? I have yet to see any such thing ;)
13:43 sitaram me too!
13:43 chid joined #git
13:43 notebox joined #git
13:43 sitaram mgedmin: there actually *is* a native GUI for gitolite...
13:43 al-damiri joined #git
13:43 ELFrederich what is this .git/refs/original/refs/
13:44 ELFrederich I've never seen that before.  it's not a tag or a remote.  How do I delete it?  It was created after doing a "git filter-branch"
13:44 sitaram mgedmin: actually more than one; the one I use is called "gvim" :)
13:44 jast ELFrederich: git filter-branch backups all altered refs (branches/tags/...) in refs/original in case you want to roll back. if you did a filter-branch in the past and are satisfied with its results, you can remove it.
13:44 jast if you want the quick and dirty way: rm -rf .git/refs/original/
13:44 satya4ever joined #git
13:44 ELFrederich jast: okay... not through a Git command, just delete the refs/original directory
13:44 jast you can do it through git commands but it's more work :)
13:45 joko left #git
13:45 efco joined #git
13:46 osse mgedmin: it's easy enough to find the merge commit itself from the log
13:46 osse mgedmin: give me a minute or two
13:47 osse have an appropriate container of your favorite beverage in the mean time
13:47 Dougie187 joined #git
13:47 Noldorin joined #git
13:47 ELFrederich jast: just out of curiosity, what is that called in Git termanology?  It's not a remote or a tag.
13:47 cdown_ joined #git
13:48 mgedmin ok I've a solution of sorts
13:48 mgedmin branch-changes = !f() { b=origin/$1; m=$(git log --pretty=%H -1 --first-parent --merges --grep="^Merge branch '$1'" master); git diff $m^...$m^2; }; f
13:48 mgedmin I'd much rather avoid using git log --grep
13:50 buffal0 joined #git
13:50 Softeisbieger joined #git
13:51 wrouesnel1 joined #git
13:53 buffal0 joined #git
13:54 osse mgedmin: https://gist.github.com/Osse/f104ea3b8bb015b72b24aca1db6928fa
13:54 Rodya_ joined #git
13:55 osse This should me more reliable since it uses the sha1s and the DAG itself instead of parsing messages
13:55 mgedmin osse: neat!  now to convert that into a oneliner
13:55 ddavidd joined #git
13:56 eivarv joined #git
13:56 osse mgedmin: unless you're into self-harm I suggest you name this script git-something and put it in $PATH
13:56 osse git can execute it for you
13:56 mgedmin good plan
13:56 ddavidd So I did a rebase locally on commits that are upstream. I'm pretty much the only one using this repo right now, can I just push it up or do I have to do something different?
13:57 buffal0 joined #git
13:57 _Patrick_ joined #git
13:57 osse mgedmin: one more file to carry around of course, but easier to edit and extend instead of those very large !f() aliases.
13:58 buffal0 joined #git
13:58 mgedmin I already have a few ~/bin/git-* scripts in my scripts git repo
13:58 osse then this should fit right in
13:59 RustyShackleford joined #git
13:59 osse Also feel free to remove those "ez". I put them there mostly to amuse myself since I am thirteen years old mentally
13:59 sfeinste joined #git
14:00 sfeinste joined #git
14:01 osse oh, and it is probably a good idea to do  git log --blah "$1"..master  so that git log doesn't run forever in case there's no match or there's a bug
14:03 mgedmin thanks a lot, osse!
14:04 rgrinberg joined #git
14:08 cdown_ joined #git
14:09 brandonw joined #git
14:10 cdown_ joined #git
14:10 osse mgedmin: de nada. now I just wonder if it actually works
14:11 GodGinrai joined #git
14:12 cdown__ joined #git
14:12 buffal0 joined #git
14:13 Blackb|rd left #git
14:14 dhollinger joined #git
14:15 grayjoc joined #git
14:15 cdown joined #git
14:17 alansaul Hey guys, how do I list commit date of all the files in a directory?
14:18 buffal0 joined #git
14:18 fuchstronaut joined #git
14:18 sangy joined #git
14:19 ddavidd alansaul: git help log
14:20 rahtgaz joined #git
14:20 alansaul Sweet thanks
14:20 alansaul ;)
14:20 cdown_ joined #git
14:21 ddavidd If I rebase locally, is there a way I can force that upstream? I realize the ramifications, but I'm the only one using the upstream.
14:22 ResidentBiscuit joined #git
14:22 ash_workz joined #git
14:23 The_Phoenix joined #git
14:24 BlaXpirit ddavidd, well you just force it...  push -f
14:24 cqi joined #git
14:24 BlaXpirit whenever you use  push -f , a good idea to also specify  remote and branch  so there are no surprises
14:25 ddavidd BlaXpirit: I was just wondering that. Okay.
14:25 ddavidd BlaXpirit: What do you mean by remote and branch?
14:25 BlaXpirit like   push -f origin master
14:25 flaviodesousa joined #git
14:26 cagomez joined #git
14:26 ddavidd Oh, gotcha.
14:26 cdown_ joined #git
14:28 johnny56_ joined #git
14:29 cdown joined #git
14:30 datasoop joined #git
14:30 phaleth joined #git
14:31 buffal0 joined #git
14:31 datasoop hi, i have ~2GB of files from a github repo, i have it in a local folder in my HDD without any .git; how i "sync" files from github?
14:32 mingrammer joined #git
14:32 ELFrederich to re-write history as if a particular commit never happened.  How can I do this?  Doesn't "git rebase -i" also turn your history into a linear set of commits?  It doesn't preserve merges and such right?
14:32 kaldoran joined #git
14:35 _ADN_ joined #git
14:36 cdown_ joined #git
14:37 morenoh149 joined #git
14:38 gewoonm joined #git
14:39 thiago_ joined #git
14:39 mycyber joined #git
14:39 veilig2000 joined #git
14:40 crelix joined #git
14:42 buffal0 joined #git
14:42 Spec-Chum joined #git
14:43 cdown joined #git
14:43 clemf joined #git
14:44 buffal0 joined #git
14:44 axisys joined #git
14:46 freimatz joined #git
14:47 YourTheBombNow23 joined #git
14:49 thiago joined #git
14:50 BSaboia joined #git
14:51 fahadash joined #git
14:51 AL13N_work joined #git
14:52 flavi0 joined #git
14:52 dmto joined #git
14:53 flavi0 i just commited something while not being on a branch and switched to another branch afterwards. how can i list the id of the non-reachable commit?
14:53 cjohnson flavi0: well first thing I'd do is scroll up in your terminal
14:53 AL13N_work i have a weird thing with my --bare git repos; i'm doing a fetch --all regularly, but it's broken now, i get: "Unable to find <hash> under <url> .cannot obtain needed tree <hash> while processing commit <hash>."
14:53 caffeinatedcode joined #git
14:53 cjohnson when you run commit it prints the commit hash
14:53 kadoban flavi0: git reflog
14:53 flavi0 cjohnson: using magit
14:53 kadoban Or cjohnson's suggestion is easier, heh.
14:53 cjohnson ah
14:53 caffeinatedcode joined #git
14:54 flavi0 kadoban: thanks!
14:54 cjohnson I do that a lot :P
14:54 AL13N_work does this mean my git repos is corrupt? or anyone can help me trying to fix this?
14:54 cjohnson especially in submodules
14:54 kyoayala joined #git
14:55 eivarv joined #git
14:56 dandaman joined #git
14:59 rgrinberg joined #git
14:59 GenZai joined #git
14:59 roelmonnens joined #git
15:01 buffal0 joined #git
15:02 buffal0 joined #git
15:03 zivester joined #git
15:04 rgrinberg joined #git
15:05 BSAlb joined #git
15:07 varnie joined #git
15:07 varnie hello.
15:07 nebg hello everyone what do you use to keep track of what to do ? s a simple TXT file with a todo list ?
15:07 OnkelTem joined #git
15:07 nebg or is there within git which allows me to simplify the task of what to do
15:07 annoymouse_ joined #git
15:07 OnkelTem Hi all
15:07 osse nebg: that's a human affair that git does not concern itself with
15:08 cjohnson nebg: you use git to track history of changes
15:08 OnkelTem How to remove "deleted" file from the stage? I tried 'reset' but it didn't help
15:08 cjohnson if you care about tracking that history on your todo.txt you could use git
15:08 osse OnkelTem: rm
15:08 varnie I had a branch which I realized that it is redundant and then I successfully deleted it. now when I run: "git branch --no-merged release" it shows up that branch among the others. how do I completely discard that branch from all git's "internals"?
15:08 osse doesn't git status tell you this?
15:08 cjohnson nebg: personally if it's a short-list item it goes to a text file, long-list items I create a ticket for
15:08 kadoban nebg: http://todotxt.com/ is one idea
15:08 i7c "remove" means you want the file back?
15:08 i7c Please clearify. :)
15:08 varnie btw, "git branch -r" doesn't show it.
15:09 osse varnie: then you didn't actually delete it. git branch -D release
15:09 nebg cjohnson, what's a ticket for  ?
15:09 OnkelTem osse: it says "fatal: pathspec '<path>' did not match any files
15:09 cjohnson nebg: for something that I need to do. I create a ticket in our ticket tracking system at work (redmine)
15:09 varnie osse, when I try that command from my local computer, it gives me: "error: branch 'feedback' not found."
15:09 osse OnkelTem: then either add or rm --cached
15:09 cjohnson assign it to myself and mark priority. then when I've got nothing else going on I check the ticket list and work on it
15:09 varnie because I deleted it already.
15:10 osse varnie: what *exactly* does git branch --no-merged say ?
15:10 varnie osse, on my local machine or a remote one?
15:10 varnie does it matter btw?
15:10 osse varnie: it matters greatly. branches are local only.  master and origin/master are different things
15:11 OnkelTem osse: hm, rm --cached says the same thing :/
15:11 osse OnkelTem: !repro
15:11 gitinfo OnkelTem: [!transcript] Please paste (using https://gist.github.com/ or https://vomitb.in/ ) a transcript ( https://gist.github.com/2415442 ) of your terminal session so we can see exactly what you see
15:11 varnie osse, hm...on my local machine your commands gives: "release".
15:11 varnie ha, the same goes for the remote one.
15:11 OnkelTem https://vomitb.in/ -- LOL
15:11 osse varnie: i thought that "release" was the branch name, but you said "feedback" so I'm not sure
15:12 varnie sorry, let me clarify that what I want.
15:12 grayjoc joined #git
15:12 varnie I want to check out that my deleted branch named "feedback" was completely removed from everywhere in git.
15:13 ELFrederich hmm... running git filter-branch and noticed the new history it created diverged earlier than it should have.  I see that it modified a commit that used to have "encoding US-ASCII
15:13 varnie yes, it  was not merged into release branch before I deleted it. (at least tried to)
15:13 mehola joined #git
15:13 OnkelTem osse: https://vomitb.in/rtS5KIgnEy
15:13 osse varnie: Say there are two PCs with a repo each, with a branch feedback each, and a central repo that both refer to as origin. If on one PC you do 'git branch -d feedback' (or -D) *and* you do git push --delete origin feedback, then the second PC will *still* have a branch named feedback that needs to be removed manually
15:13 varnie but if it is physically deleted, why does 'git branch --no-merged' shows it?
15:13 ELFrederich when it ran, it removed that "encoding US-ASCII" from the commit object.  I saw it using a "git cat-file -p" of the first point of divergence
15:13 varnie s/shows/show/
15:13 aard_ joined #git
15:14 osse OnkelTem: woops. I misunderstood everything. you want to unstage right? do what git status says
15:14 osse OnkelTem: git reset HEAD modules/custom/im_import/config/install/migrate_plus.migration.im_file_imageset.yml.txt
15:14 OnkelTem Ah, I see. Hm.. But what 'git reset' worked for other files? W/o HEAD I mean
15:14 varnie osse, the central repo still thinks that branch is alive and not merged. the local PC DOES NOT.
15:14 OnkelTem why*
15:15 osse varnie: local and remote branches are separate things
15:15 varnie is it possible to remove that branch from the central repo as well?
15:15 osse varnie: push --delete
15:15 osse like in my example above
15:15 davisonio joined #git
15:15 i7c The commits will still float around tho. :)
15:16 dreiss joined #git
15:16 varnie osse, pardon, I don't get it.
15:17 varnie "error: unable to delete 'feedback': remote ref does not exist"
15:18 osse varnie: then it has already been deleted from the remote. This might not be reflected in the local repo (yet). Run git fetch --prune and see if it disappears
15:19 buffal0 joined #git
15:19 fuchstronaut joined #git
15:20 varnie okay, but "git branch --no-merged" still shows "feedback" on the central repo
15:20 osse varnie: central, as in you're logged in on the server and run commands?
15:20 varnie yes, exactly.
15:20 choki joined #git
15:20 InfoTest joined #git
15:21 osse varnie: and this is the same place as the URL shown in
15:21 osse shown in ''git remote -v'' refers to on the local machine?
15:21 varnie hmm, no.
15:22 khmarbaise joined #git
15:22 varnie you mean, it is not  the central repo, correct?
15:22 varnie it is the test server! I am sorry.
15:22 LionsMane joined #git
15:23 sanketdg joined #git
15:23 rnsanchez joined #git
15:24 sanketdg joined #git
15:24 eivarv joined #git
15:24 varnie osse, okay, I figured it out. I deleted that branch from the test server as well. now "git branch --no-merged" shows an empty output.
15:24 varnie thank you for your time, osse.
15:26 varnie how do people usually go about dealing with this situation then: " then the second PC will *still* have a branch named feedback that needs to be removed manually"?
15:27 mingrammer joined #git
15:27 varnie I mean if John deletes non-needed branch, it remains on the computers of Jane and Peter.
15:27 buffal0 joined #git
15:27 varnie should he tell them to perform delete manually or what?
15:27 slayerjain joined #git
15:28 varnie maybe there's some way to perform delete from the central repo, and then Jane and Peter will have no branch after some "fetch" or whatever?
15:29 b1tchcakes joined #git
15:29 Finnoloid joined #git
15:29 ELFrederich where do I get the command git_commit_non_empty_tree from?
15:30 mozzarella joined #git
15:30 davisonio joined #git
15:30 Rodya_ joined #git
15:31 thiago ELFrederich: why do you think it exists? What are you trying to do?
15:31 ELFrederich thiago: this mentions it https://git-scm.com/docs/git-filter-branch
15:32 thiago ELFrederich: it's a shell function provided by git-filter-branch itself
15:32 ELFrederich thiago: I'd like to have my --commit-filter as a file that I can edit in an editor rather than inlining it in a bash command
15:32 mohabaks joined #git
15:33 OnkelTem osse: thanks!
15:33 synthroid joined #git
15:33 osse OnkelTem: not sure what your question about git reset for other files mean
15:33 ahmedelgabri joined #git
15:33 OnkelTem osse: I meant that when a file is just modified, I do: git reset to remove it from the stage
15:34 zmachine joined #git
15:35 govg joined #git
15:35 kbs joined #git
15:36 buffal0 joined #git
15:37 osse OnkelTem: seems to work the same for deleted files
15:37 buffal0 joined #git
15:38 OnkelTem osse: not im my specific case though. Probably that's because of pre-history:
15:38 OnkelTem I had a file, then I renamed it
15:38 OnkelTem So first I saw somthign like:
15:39 OnkelTem old_name > new_name  (in the stage)
15:39 OnkelTem then I did 'git reset new_name', and then only the old_name left in the stage
15:39 buffal0 joined #git
15:39 OnkelTem deleted: old_name
15:40 OnkelTem and that old_name I couldn't reset w/o HEAD word
15:40 EvilPenguin joined #git
15:41 cortexman joined #git
15:41 DeJoker joined #git
15:41 stamina joined #git
15:42 cortexman has anyone used android's git-repo tool?
15:42 cortexman i have some questions about the manifest, and general usage
15:43 DeJoker left #git
15:43 kadoban cortexman: You should probably just ask your followups. Non-standard tools don't get amazing support here though. Try in whatever the android channel is if you don't get answers, maybe.
15:44 cortexman in general i am just not finding this file to be entirely self-documenting: https://android.googlesource.com/platform/manifest/+/master/default.xml
15:44 Gsham joined #git
15:45 cortexman as far as i can tell, this must be the place to specify the main repo url: <remote name="aosp" fetch=".." review="https://android-review.googlesource.com/" />
15:46 cortexman but i don't understand why the attr is called review
15:46 dimi1947 joined #git
15:47 varnie msg gitinfo .voice
15:47 varnie oops
15:47 WayToDoor joined #git
15:48 ahmed_elgabri joined #git
15:49 oich joined #git
15:49 durham joined #git
15:50 byteflame joined #git
15:50 darkbit joined #git
15:51 htafdresgi joined #git
15:51 htafdresgi how do I rollback to the previous commit?
15:51 nitric joined #git
15:51 htafdresgi I did some stuff and now it's broken and I don't know why.
15:51 durham joined #git
15:51 shinnya joined #git
15:51 BSAlb joined #git
15:52 BSAlb joined #git
15:52 nchambers joined #git
15:53 f0rblaze joined #git
15:53 cjohnson you can git reset to rewind history without affecting the working tree
15:54 fish3485 joined #git
15:54 htafdresgi cjohnson: okay I'll try that
15:55 cjohnson you can use git revert $COMMIT to commit a new change which is a negative of the commit you want to rewind
15:55 cjohnson depends if you've pushed or not
15:57 htafdresgi cjohnson: I haven't pushed, and I haven't commited (the new changes).
15:57 htafdresgi just want to get back to how things were at the previous commit around noon yesterday
15:58 htafdresgi I tried git reset but program still doesn't work.. maybe it's some system libraries I installed since then or something.
15:58 kyan joined #git
15:59 varnie git clean -df; git checkout -- . possibly
15:59 osse OnkelTem: hmm, yes I see what you mean. it works without HEAD if the file is still on disk (like rm --cached), but not when it's actually not there (like rm or if you rename)
16:00 OnkelTem osse: ah, I see!
16:00 clemf_ joined #git
16:00 dsantiago joined #git
16:00 nidr0x joined #git
16:00 osse OnkelTem: but if it's any comfort this was news to me too, and i find it a bit weird
16:01 sangy joined #git
16:01 htafdresgi hmm, nope didn't do it, I guess I'll have to try it some other way
16:01 cyan__ joined #git
16:01 varnie htafdresgi, what does 'git status' show?
16:01 davisonio joined #git
16:01 varnie it should give you a hint what to do, anyway
16:01 htafdresgi varnie: on branch master, with a bunch of uncommited changes in other parts of the project
16:02 leeN joined #git
16:02 brandonw joined #git
16:02 varnie hmm, git checkout didn't clear them?
16:03 nidr0x joined #git
16:03 SpecChum joined #git
16:03 l4v2 joined #git
16:04 johnny56_ joined #git
16:04 grimel joined #git
16:05 grimel Hi!. Is it OK that after I did git-branch (removed some sensitive data) - "git log --all" shows every commit twice? Without --all everything is OK
16:05 htafdresgi varnie: well I'm in a subdirectory of the pertinent parts of the project, it's probably cause I was trying to compile pocl yesterday and apt-getted a bunch of packages in that effort
16:05 roxtrongo joined #git
16:06 KevinMGranger Then you still have (a) branch(es) pointing to whatever you filtered / rebased
16:06 KevinMGranger just as a guess
16:06 zacts joined #git
16:07 tyreld joined #git
16:07 grimel KevinMGranger - there is only one branch - "master" and if I do "git log" - everything is OK - one commit - one entry, but with prefix --all every commit is shown two times
16:07 grimel different references, though
16:07 grimel exactly the same time but different references
16:08 cjohnson htafdresgi: if you're in a subdirectory, replace .  with the path to the root of the repo
16:08 cjohnson htafdresgi: git checkout -- $PATH
16:08 ORiON-- joined #git
16:08 grimel maybe that shouldn't bother me at all...
16:08 cjohnson htafdresgi: if you're in root, . is the top of repo, but if you aren't, you'll need to specify top of repo
16:10 buffal0 joined #git
16:12 grimel I got it for a bit - filter-branch saves references to old commits. Here is a question - how do I delete them?
16:12 nebg guys i have my local repo of a sftware i wrote
16:12 nebg i want to upload it to github... and i'm finding it difficult...
16:12 htafdresgi cjohnson: it's okay I figured it out, there were some packages I isntalled yesterday that I removed, now it works fine :-D
16:12 htafdresgi i'll add them back one by one to find out which one,
16:13 nebg is there a guide steèp by step which explains how to do it...without removing all my project and without having synchronization problems
16:13 nebg ?
16:13 nebg should i do first a fetch and then a push ?
16:13 nebg or what '
16:13 _ikke_ !book is agoodo one
16:13 gitinfo There are several good books available about git; 'Pro Git' is probably the best: http://git-scm.com/book but also look at !bottomup !cs !gcs !designers !gitt !vcbe and !parable
16:13 nebg ?
16:13 nebg _ikke_, don't have time to read an entire book
16:13 _ikke_ https://git-scm.com/book/en/v2/Git-Basics-Working-with-Remotes
16:13 xall joined #git
16:13 _ikke_ More specific
16:13 grimel nebg: when you create a repo on github you have instruction
16:14 ahmed_elgabri joined #git
16:14 htafdresgi it was ocl-icd-opencl-dev
16:14 HappyLoa- joined #git
16:14 rossome joined #git
16:15 kbs nebg: does the standard https://help.github.com/articles/adding-an-existing-project-to-github-using-the-command-line/ not help?
16:15 nidr0x joined #git
16:15 AbstractArtist joined #git
16:16 grimel In case if somebody is interested how to remove old history after filter-branch - http://stackoverflow.com/a/5985276/3434076 helped me
16:16 nebg kbs, sorry this https://help.github.com/articles/adding-an-existing-project-to-github-using-the-command-line/ does not contain anything...
16:16 nebg i don't find the instructions
16:17 kbs nebg: curious, hm - oh well.
16:17 kadoban nebg: Click on your OS at the top
16:17 nebg ok
16:18 nebg ok thanks
16:19 spudowiar joined #git
16:19 GenZai joined #git
16:20 fuchstronaut joined #git
16:22 jnewt anyone know how to get tab completion for git bash on windows?  or if it should already be in the install, why it may not be working
16:23 buffal0 joined #git
16:24 Gsham joined #git
16:30 Iw1ll1nj3ct joined #git
16:30 dreiss joined #git
16:31 Goplat joined #git
16:31 zumba_addict joined #git
16:34 thiago joined #git
16:34 diegoviola joined #git
16:34 buffal0 joined #git
16:35 rath joined #git
16:36 cnagy joined #git
16:36 GenZai joined #git
16:38 buffal0 joined #git
16:40 leeN joined #git
16:40 Electrometro joined #git
16:41 ytl joined #git
16:44 nebg jnewt,
16:44 nebg no idea on windows
16:47 TomyLobo joined #git
16:49 buffal0 joined #git
16:49 Zarthus cygwin already has tabcompletion afaik
16:49 Zarthus that's something I would definitely notice missing in the time i've used it
16:49 nebg Zarthus, i think he is using the git bash
16:49 nebg on windows
16:49 SteffanW joined #git
16:49 Zarthus yeah, but git bash is cygwin/mingw
16:49 nebg and not cygwin
16:50 nebg ahh ok
16:50 nebg didn't know
16:50 jaafar joined #git
16:50 z8z joined #git
16:51 MeAndCesarLeo joined #git
16:52 Balliad joined #git
16:52 buffal0 joined #git
16:52 daynaskully joined #git
16:52 madewokherd joined #git
16:55 dreiss joined #git
16:56 nikivi joined #git
16:56 Sasazuka joined #git
16:57 hahuang65 joined #git
16:57 GavinMagnus left #git
16:59 algae joined #git
16:59 jaafar joined #git
17:06 fakenerd joined #git
17:06 bongjovi joined #git
17:06 jimi_ joined #git
17:07 inflames joined #git
17:07 Manuel15 joined #git
17:07 Manuel15 joined #git
17:07 ResidentBiscuit joined #git
17:08 inflames joined #git
17:09 skylite joined #git
17:10 naw joined #git
17:12 alansaul joined #git
17:13 moei joined #git
17:14 clemf joined #git
17:14 mehola joined #git
17:14 l4v2 joined #git
17:14 Fissure joined #git
17:14 clemf joined #git
17:14 vuoto joined #git
17:14 mingrammer joined #git
17:15 vuoto joined #git
17:16 vuoto joined #git
17:16 Dyonisos joined #git
17:17 vuoto_ joined #git
17:17 Dyonisos joined #git
17:17 InfoTest1 joined #git
17:18 JanC_ joined #git
17:18 tgunr joined #git
17:18 ddavidd left #git
17:19 vuoto joined #git
17:19 ahmedelgabri joined #git
17:20 whomp joined #git
17:21 vuoto joined #git
17:21 whomp how can i add just one line from a set of lines that i changed? i tried "s" in git add -p but it won't let me split it up further
17:22 osse whomp: 'e'
17:22 beyertech joined #git
17:22 osse then wrap your head around how to do it (help at the bottom)
17:24 dsantiago joined #git
17:24 cagomez joined #git
17:25 whomp osse, great, thanks :)
17:27 jaafar joined #git
17:29 buffal0 joined #git
17:29 diegoviola joined #git
17:30 ahmed_elgabri joined #git
17:30 sanketdg joined #git
17:31 buffal0 joined #git
17:32 nilg joined #git
17:32 pandeiro joined #git
17:32 johnmilton joined #git
17:33 hahuang65 joined #git
17:36 ryouma joined #git
17:38 kpease joined #git
17:39 cdg joined #git
17:39 ryouma is .merge_file_4IJdJn something that git creates?  it shows up in git status for no obvious reason.
17:39 johnmilton joined #git
17:40 cdg joined #git
17:41 blackwind_123 joined #git
17:42 grawity it is
17:42 Balliad joined #git
17:42 ryouma grawity: what i did was stage hunks and delete them.  it remains.
17:42 jstein__ joined #git
17:43 Balliad joined #git
17:44 buffal0 joined #git
17:47 BrianJ joined #git
17:47 GodGinrai joined #git
17:48 celdia joined #git
17:49 BrianBlaze420 joined #git
17:49 AbstractArtist joined #git
17:50 joki joined #git
17:51 TalkToMe joined #git
17:52 Literphor joined #git
17:54 subhojit777 joined #git
17:55 subhojit777 joined #git
17:56 morenoh149 joined #git
17:56 buffal0 joined #git
17:57 MrMojito joined #git
17:57 paul424 joined #git
17:59 cschneid_ joined #git
18:00 EricPoe joined #git
18:02 dandaman joined #git
18:02 xall joined #git
18:05 buffal0 joined #git
18:09 mingrammer joined #git
18:09 buffal0 joined #git
18:12 msonntag joined #git
18:13 dodobrain joined #git
18:14 kaldoran joined #git
18:14 labels joined #git
18:16 buffal0 joined #git
18:16 AbstractArtist joined #git
18:17 labels left #git
18:17 buffal0 joined #git
18:18 labels joined #git
18:19 zacts joined #git
18:19 labels ...
18:19 labels I would like place an attribute on a commit, showing the CI test results. Is there a way to do this?
18:19 bremner git-notes might work
18:20 Darren_ joined #git
18:20 stuh84 joined #git
18:21 fuchstronaut joined #git
18:23 rossome joined #git
18:23 energizer joined #git
18:23 labels @bremner yes, it looks like git-notes may be what I'm looking for...
18:23 labels although I was hoping for something that looks more like labels on issues on github
18:24 energizer I'm working with a few other guys on a website, which has a testing server. We have everything in git, but I don't understand how we can test out different branches and see the result
18:24 netcarver joined #git
18:24 jaafar joined #git
18:25 energizer Do we have to just merge everything into master in order to see the results? If so, seems like branching doesn't get us much.
18:25 energizer I don't understand how the workflow is supposed to be
18:26 GodGinrai energizer: when you change branches, it changes the files that are in your working directory.
18:27 ToxicFrog energizer: this is a how-is-your-test-server-set-up question, I think, not a git question.
18:27 GodGinrai If you deploy those files to your test server (or if you are lazy, are just changing the branches on the test server) then you will see the changes
18:27 whomp_ joined #git
18:27 ToxicFrog I mean, at my work, the answer is "git isn't available on the test server anyways; you 'git checkout' the branch you want to test and then push the configs to the test server using a separate tool"
18:27 ToxicFrog (or just bring up a local instance and connect to localhost)
18:27 GodGinrai local instances are the best
18:28 GodGinrai for testing
18:28 energizer so each person would just have their own copy of the site that they'd be running on their pc?
18:28 ToxicFrog GodGinrai: they are, but I work on some things that require more resources to start up than my workstation has, so sometimes you have to test remotely
18:29 GodGinrai ToxicFrog: or sometimes you have to test remotely because of other reasons... (like a vendor-run server who won't give you a local instance to use... -_-)
18:29 ToxicFrog Yeah :(
18:30 lebara joined #git
18:30 energizer so in those circumstances, there's only one instance, and only master is ever checked out on it, and everyone just pushes to it?
18:30 energizer to test out new setuff
18:30 energizer stuff
18:30 dreiss joined #git
18:31 GodGinrai energizer: no.  We never once said anything about pushing to master
18:32 energizer how do you test out new stuff on a remote server
18:32 GodGinrai That depends on your setup
18:32 mrBen2k2k2k joined #git
18:32 GodGinrai Normally you'd just deploy the files to the remote server
18:32 neanderslob joined #git
18:33 energizer no git on the remote at all
18:33 ToxicFrog Exactly. Like I said.
18:33 GodGinrai it's not necessary on the remote
18:33 GodGinrai unless you are lazy and trying to use git as a deploy tool (which it isn't)
18:34 ToxicFrog Here we have a separate tool for "package all your local configs/code, upload it to a remote server, and run it"; you use git locally, check out the branch you want to test, and then run that tool and it does the thing. The remote server doesn't care whether the code is coming from a git checkout or a release archive or whatever.
18:34 Vampire0_ joined #git
18:35 jimi_ joined #git
18:35 prestorium joined #git
18:35 GodGinrai for us, we have a gulp task that builds our code and pushes it to the server
18:36 davisonio joined #git
18:37 armin so we have a company that thinks that feature toggles are the answer to merge conflicts. they use subversion. i would love to move them to git but they see no benefit and consider feature branches in favor of feature toggles contradictionary to continuous delivery. what should i do?
18:37 guardian hello, is there a way to use diff-highlight for git add -p ?
18:38 GodGinrai armin: start looking for more progressive companies
18:38 armin i already told them that copy-on-write is invevitable and that using branches for different stages is a huge benefit.
18:38 buffal0 joined #git
18:39 armin GodGinrai: even MORE progressive? i already consider the way they work aggressive.
18:39 dsantiago joined #git
18:39 armin GodGinrai: can you elaborate?
18:39 mrBen2k2k2k joined #git
18:39 armin mrBen2k2k2k: oh chatting as a logstash user. did you hack your logging server?
18:40 GodGinrai armin: progressive means "willing to adopt new ideas".  If they are eshewing git in favor of feature toggling, then I don't see them as very keen on new things
18:40 fuchstronaut joined #git
18:40 daniel_rose joined #git
18:40 armin GodGinrai: do i have a chance to convince them? what would you think?
18:40 stickperson joined #git
18:41 GodGinrai armin: you only have a chance to convince them if you have a lot of say.  If you don't, then they'll just keep ignoring you.
18:41 stickperson can i reuse a tag?
18:44 buffal0 joined #git
18:44 lb stickperson define reuse.
18:45 stickperson lb: tag a commit with tag “a” and push. make a change, commit, tag that commit with “a” and push
18:45 stickperson i could tag it “b”, sure, but just curious
18:46 lb nope
18:46 lb but delete a tag an recreate a new tag with same name, yes
18:47 lb (but you would have to force push)
18:47 lb not recommended
18:47 armin GodGinrai: so we would have to have a group of people trying to convince them?
18:48 armin stickperson: be careful, you can use the same name for both a tag and a branch.
18:48 mrBen2k2k2k joined #git
18:48 armin stickperson: you should, by convention, avoid this.
18:48 stickperson armin: yea, i keep branch and tag names separately. i’ll just create different tags instead of deleting/force pushing
18:49 armin force pushing is evil.
18:49 armin only do it if you know nobody other already obtained a copy.
18:51 Cassiopaya joined #git
18:51 UTAN_dev_ joined #git
18:52 MrMojito joined #git
18:53 IlNominePatre joined #git
18:54 1JTAA105Z joined #git
18:54 ExoUNX joined #git
18:55 buffal0 joined #git
18:55 jimi_ joined #git
18:55 GodGinrai armin: you could try
18:56 Macaveli joined #git
18:57 f0rblaze joined #git
18:58 bmispelon joined #git
18:58 AbstractArtist joined #git
18:59 hahuang61 joined #git
18:59 mrBen2k2k2k joined #git
19:00 EricPoe joined #git
19:01 annoymouse joined #git
19:01 kingarmadillo joined #git
19:02 kingarmadillo is there a way to see the additions/subtractions since the branch was made?
19:03 astrozyk joined #git
19:03 maxb joined #git
19:04 mariuscc joined #git
19:04 ToxicFrog kingarmadillo: git diff is your all purpose "see changes between two commits" command; git diff starting-point..branch-name
19:06 AbstractArtist- joined #git
19:06 jnewt hwo do i ignore all of the subdirectories in a top level folder, but include all the files in that folder (in .gitignore file)
19:06 jnewt ie ./dir/dir should be ignored, but ./dir/file should be included
19:07 kadoban /thedir/*/   ?
19:07 jimi_ joined #git
19:08 jnewt kadoban, well, now I feel dumb.  thanks.
19:09 kadoban 'welcome
19:10 mrBen2k2k2k joined #git
19:11 astrozyk joined #git
19:12 astrozyk joined #git
19:12 buffal0 joined #git
19:13 astrozyk joined #git
19:13 astrozyk joined #git
19:14 sgen joined #git
19:14 cyan__ joined #git
19:15 fahadash joined #git
19:15 arand_ joined #git
19:16 buffal0 joined #git
19:16 neanderslob joined #git
19:18 rossome joined #git
19:18 DolphinDream joined #git
19:18 zacts joined #git
19:18 daniel_rose joined #git
19:19 LionsMane joined #git
19:19 Balliad joined #git
19:20 jimi_sanchez joined #git
19:21 tgunr joined #git
19:21 daniel_r_ joined #git
19:23 buffal0 joined #git
19:23 Akimb joined #git
19:23 mrBen2k2k2k joined #git
19:24 crayon joined #git
19:25 meth joined #git
19:26 mahendra_ joined #git
19:27 buffal0 joined #git
19:28 Rodya_ joined #git
19:32 thiago joined #git
19:32 YourTheBombNow23 joined #git
19:35 thm left #git
19:35 buffal0 joined #git
19:35 nwkj86 joined #git
19:37 jimi_ joined #git
19:37 buffal0 joined #git
19:38 clemf_ joined #git
19:40 bart_ joined #git
19:40 circ-user-FePRT joined #git
19:40 mrBen2k2k2k joined #git
19:40 bart_ hi
19:41 gitinfo bart_: 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.
19:41 bart_ Quick question for you guys.
19:42 bart_ I have NFS home accounts with Git repos in them.
19:42 bart_ Doing a "git gc" does not work, it's exepecting a share in "/" not in a "/home/johndoe/dev/git..."
19:43 bart_ Has anyone been able to work around that ?
19:43 jnewt i split out a folder from my repository into it's own, and now want to remove it from the original   stackoverflow had an answer that was git filter-branch --prune-empty --tree-filter 'rm -rf <name-of-folder>' HEAD which failed:
19:44 ToxicFrog bart_: git doesn't care where the repo is; how are you invoking it?
19:44 jnewt a couple of these:  0 [main] sh 3356 fork: child -1 - forked process 4520 died unexpectedly, retry 0, exit code 0xC0000005, errno 11
19:45 NeverDie joined #git
19:45 bart_ @ToxicFrog That's the thing... it seems the command "git gc" is messed up
19:45 jnewt then a Segmentation Fault
19:45 bart_ I get this "Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set)"
19:46 jnewt C:\Program Files\Git\mingw64/libexec/git-core\git-filter-branch: line 412:  4900 Segmentation fault
19:46 bart_ the command is run, it actually deletes almost every file in the current git repo, it's a mess....
19:46 BadApe_ joined #git
19:46 jnewt git version 2.9.3.windows.1
19:46 morenoh149 joined #git
19:47 bart_ If I move the repo to the local drive it works... It's an NFS issue
19:47 ToxicFrog bart_: that means it's walking up the directory tree to try to find the .git directory and hitting a mount boundary before it finds .git
19:47 bart_ I don't get why the 'git gc' command treats it differently
19:47 ToxicFrog Perhaps try setting GIT_DISCOVERY_ACROSS_FILESYSTEM as it suggests?
19:47 ToxicFrog Differently how? Do other git commands work as expected?
19:47 yoan joined #git
19:48 bart_ yes all other git commands work, except branching seems broken as well
19:48 sdothum joined #git
19:49 bart_ regarding "git gc" I understand that it's moving up the tree and hits the top. But I am in the proper folder, so weird
19:49 ToxicFrog Weird. I've got nothing. Maybe someone who actually uses NFS would have more insights.
19:50 bart_ Also I'm afraid that If I set the value of "GIT_DISCOVERY_ACROSS_FILESYSTEM" it will find other repos
19:50 bart_ not the proper one
19:50 Macaveli joined #git
19:50 buffal0 joined #git
19:51 paul424 joined #git
19:51 bart_ TOxicFrog: Thank you for your time thought :-)
19:52 DRiOscar1 joined #git
19:53 buffal0 joined #git
19:53 Electrom_ joined #git
19:54 buffal0 joined #git
19:55 Macaveli joined #git
19:56 _ADN_ joined #git
19:57 mingrammer joined #git
19:58 djb-irc joined #git
19:59 EvilPeng1 joined #git
19:59 davisonio joined #git
20:00 buffal0 joined #git
20:00 mrBen2k2k2k joined #git
20:01 buffal0 joined #git
20:05 jefus joined #git
20:05 macivy joined #git
20:06 ExoUNX joined #git
20:07 buffal0 joined #git
20:09 JacobsLadd3r joined #git
20:11 Spec-Chum joined #git
20:11 CheckDavid joined #git
20:14 dodobrain joined #git
20:15 jeremyhall joined #git
20:15 dodobrain joined #git
20:15 neanderslob joined #git
20:15 neanderslob_ joined #git
20:15 johnmilton joined #git
20:16 cdg joined #git
20:16 neanderslob joined #git
20:19 EvilPeng1 joined #git
20:19 EricPoe joined #git
20:19 daniel_rose joined #git
20:20 buffal0 joined #git
20:20 johnmilton joined #git
20:21 g5pw joined #git
20:22 mrBen2k2k2k joined #git
20:22 arand_ joined #git
20:26 tgunr joined #git
20:27 buffal0 joined #git
20:28 blr joined #git
20:29 buffal0 joined #git
20:30 morenoh149 joined #git
20:31 venmx joined #git
20:32 buffal0 joined #git
20:32 badloop joined #git
20:33 badloop If I am on an owner of a remote git repository, is it possible to restrict push access to only certain users on the remote repo?
20:33 badloop or would there need to be a service on top of git for that type of authentication control
20:34 j416 badloop: ssh and unix users
20:34 j416 badloop: but, your life would be easier if you used e.g. gitolite
20:34 j416 badloop: or, go with some fullblown web ui thing
20:36 zacts joined #git
20:36 badloop j416: perfect.  thanks!
20:36 jwest joined #git
20:37 buffal0 joined #git
20:38 venmx joined #git
20:38 jefus joined #git
20:40 buffal0 joined #git
20:41 fuchstronaut joined #git
20:41 buffal0 joined #git
20:45 clemf joined #git
20:45 buffal0 joined #git
20:46 GavinMagnus joined #git
20:46 jagob joined #git
20:46 mrBen2k2k2k joined #git
20:47 jaafar joined #git
20:47 jefus joined #git
20:50 kingarmadillo joined #git
20:51 buffal0 joined #git
20:51 GavinMagnus left #git
20:52 mingrammer joined #git
20:52 aidalgol joined #git
20:53 t0by joined #git
20:53 renlo joined #git
20:54 buffal0 joined #git
20:55 renlo is there a way to source control a file without 'diffing it' / doing merges, etc? We have a compilation process where we output some JavaScript and CSS files. These are minimized and everything is on one line. I would like to source control these files so that when someone pulls or clones the repository, these files are there, BUT I dont want to migrate these files; I just want a snapshot of whatever the file was for that commit. Is there a way to do this?
20:55 anttim joined #git
20:55 osse renlo: git does snapshots
20:55 osse renlo: all diffs are produced on the fly
20:56 _ikke_ "I don't want to migrate these files" What do you mean with this?
20:56 renlo is there a way to tell git that I dont want to merge that file? that I want it to grab one or the other?
20:57 _ikke_ yes
20:57 _ikke_ add <filename> -diff to a .gitattributes file
20:57 renlo that will remove the diffing if you do `git diff ...` ?
20:57 _ikke_ This will cause git to treat the file as binary
20:57 renlo perfect!
20:58 _ikke_ renlo: It will just say they're different
20:59 _ikke_ renlo: there is also a binary attributes, which does a bit more
20:59 _ikke_ so <filename> binary
20:59 renlo that might be exactly what I want
20:59 dsfreak93 joined #git
20:59 hahuang61 joined #git
21:00 faheem joined #git
21:00 _ikke_ binary -diff -merge -text
21:00 vassagus joined #git
21:01 vassagus joined #git
21:01 davisonio joined #git
21:01 a3Dman joined #git
21:02 vassagus joined #git
21:02 renlo that works perfectly, thank you _ikke_
21:02 _ikke_ yw
21:03 atomicb0mb joined #git
21:05 filcab joined #git
21:05 neanderslob joined #git
21:05 roubix joined #git
21:07 buffal0 joined #git
21:08 robotroll joined #git
21:08 filcab joined #git
21:08 mr_quincy joined #git
21:08 mr_quincy left #git
21:09 djb-irc joined #git
21:09 buffal0 joined #git
21:10 buffal0 joined #git
21:11 crose joined #git
21:12 mrBen2k2k2k joined #git
21:13 Goplat joined #git
21:15 buffal0 joined #git
21:15 mehola joined #git
21:17 davisonio joined #git
21:19 buffal0 joined #git
21:19 xall joined #git
21:25 djb-irc joined #git
21:27 Jon28_ joined #git
21:27 xall joined #git
21:28 grazfather joined #git
21:28 grazfather Any idea why this would fail? `git reset --hard`, but i still have a local change
21:28 grazfather same with git checkout -- .
21:28 grazfather i just cannot get rid of this change
21:28 osse what is the change?
21:29 grazfather it's a binary
21:29 grazfather probably something made from a madefile. a codebase i am unfamiliar with
21:29 grazfather i just want to trash it
21:29 osse the change or the file?
21:29 grazfather the change is to a binary file so i can't diff
21:29 grazfather and i don't care i just want to trash it
21:31 neanderslob joined #git
21:32 Linkandzelda joined #git
21:32 Linkandzelda joined #git
21:36 rwb joined #git
21:36 thiago joined #git
21:38 Dyonisos joined #git
21:39 ahr3n joined #git
21:40 mrBen2k2k2k joined #git
21:43 weox joined #git
21:45 Dougie187 left #git
21:46 aard_ joined #git
21:46 rbern joined #git
21:48 _ikke_ grazfather: I suspect eol normalization
21:49 whomp joined #git
21:51 mostlybadfly joined #git
21:57 Dong__ joined #git
21:58 cagomez joined #git
21:58 Dong__ quick question, what are the two network speed means in git clone?
21:59 cd-rum joined #git
22:00 atomicb0mb joined #git
22:02 realz joined #git
22:04 thiago Dong__: what two speeds?
22:04 jlebar joined #git
22:05 byteflame joined #git
22:06 vassagus joined #git
22:09 mrBen2k2k2k joined #git
22:10 eivarv joined #git
22:11 Dong__ thiago: Receiving objects: 100% (6887/6887), 6.11 MiB | 3.37 MiB/s, done.
22:11 kadoban MiB isn't a speed. MiB/s is
22:11 thiago technically, MiB/s isn't a speed, it's a transfer rate
22:12 LeBlaaanc joined #git
22:12 thiago Dong__: but like kadoban said, one of them is a volume, not a rate
22:14 Dong__ Thanks guys, know I see it, the first is the size of the pack file, the second is the size divide by time spend receiving it.
22:14 mooru joined #git
22:14 grayjoc joined #git
22:16 jaafar joined #git
22:17 BSaboia joined #git
22:23 bronson joined #git
22:24 t0by joined #git
22:24 polyzen joined #git
22:27 jefus joined #git
22:31 Manuel15 joined #git
22:31 Manuel15 joined #git
22:32 shinnya joined #git
22:34 sangy joined #git
22:36 alcaveman joined #git
22:36 alcaveman can i add empty directory?
22:36 osse no
22:36 alcaveman why ?
22:36 osse Not sure
22:36 osse !empty
22:36 gitinfo [!directory] Git does not track directories, only files in directories.  Create a nonce file as a placeholder (eg .gitignore) to create the directory or create the directory as part of a post-checkout hook
22:37 roxtrongo joined #git
22:37 alcaveman wat
22:38 spudowiar joined #git
22:38 thiago alcaveman: exactly what it says
22:38 thiago alcaveman: git only tracks files, not directories
22:39 thiago in fact, it only tracks content, so some file attributes aren't kept
22:39 alcaveman make it work on directories too please
22:39 thiago no
22:39 osse the willpower is not there
22:39 thiago haven't needed it for 11 years, won't happen now
22:39 Gsham joined #git
22:40 osse barring any human objections etc, wouldn't it be a relatively small patch to make it handled an empty tree?
22:40 thiago possibly
22:40 mingrammer joined #git
22:40 mrBen2k2k2k joined #git
22:41 alcaveman it is useful to also track directories
22:42 thiago 4b825dc642cb6eb9a060e54bf8d69288fbee4904
22:42 alcaveman kwhat's taht?
22:42 thiago the SHA-1 of an empty tree
22:43 fuchstronaut joined #git
22:43 alcaveman sha512 please
22:43 alcaveman btw if a collision happens, does git report it?
22:43 thiago no
22:43 osse funny how git cat-file -t will say it's a tree at all since no git repo can possibly contain it :P
22:44 alcaveman linus torvalds doesn't believe in birthday attack?
22:44 osse hmm, there is some code do report detections, isn't there?
22:44 mikecmpbll joined #git
22:44 osse report collisions i mean
22:44 alcaveman have anyone spotted a collision yet?
22:44 thiago it's probably a by-product of your committing and git add'ing
22:44 alcaveman git probably doesn't scale well for long-term use cause collisions
22:44 thiago it exists in your repo, but not as part of the DAG
22:45 thiago alcaveman: no collision has ever been reported
22:45 thiago alcaveman: if you think a birthday attack is relevant, please calculate how many objects are necessary in a repo for the chance to climb over 1%
22:45 alcaveman any reason why collision detection isn't built in? just like how hash/dictionaries have it?
22:45 nwkj86 joined #git
22:45 thiago because it *isn't* a hashing table
22:46 alcaveman same concept
22:46 thiago it hashes the content and, if it detects that an object by that name exists, it assumes it's the same
22:46 thiago no, it isn't
22:46 alcaveman git uses these hashes to have a key-value structure
22:46 thiago a hashing table uses a shorthand that is easy to calculate in order to speed up searching
22:46 thiago unless it's a perfect hash, it knows collisions could happen
22:47 alcaveman these are technical differences in some details relating performance/efficiency. but fundamental concept is that git is using such hashes to implement some kind of key-value store
22:47 thiago anyway, please calculate the number of objects required for a collision chance of 1%
22:47 alcaveman okay hold on lemme calculate
22:47 osse https://github.com/git/git/blob/master/builtin/index-pack.c#L731-L757
22:47 alcaveman 160 bits right?
22:47 thiago alcaveman: yes
22:47 grazfather _ikke_: But why would `reset --hard` fail? I ended up manually rm-ing the file and check `co --`
22:48 rpd grazfather: did `reset --hard` actually fail or did it not do what you expected?
22:48 rpd reset --hard can only make modification to files it knows about afaik
22:48 rpd if that binary was never tracked, it wouldn't remove it
22:48 alcaveman thiago: between 2.6 × 10^18  and  4.8 × 10^37
22:49 alcaveman there must be that many commits to get 1%
22:49 thiago alcaveman: that many objects
22:49 alcaveman right
22:49 alcaveman but 1% is really huge. it's really bad luck to get 1% chance of having a silent collision that creates a blackhole
22:50 thiago alcaveman: so assuming that they're all 40-byte long files, we need a storage of 4*10^38 bytes
22:50 _nwkj86 joined #git
22:50 stephanie left #git
22:50 alcaveman 3.63797880709171e+26 tera bytes
22:51 osse git will have switched to sha-something-else by then
22:51 thiago if a 1 TB SSD (10^12 bytes) is 120 g, then this would weigh 4.8*10^25 kg
22:51 thiago that's 8 times more than the Earth
22:52 alcaveman osse: wasteful. just have an extra file per object that is named (say): 4b825dc642cb6eb9a060e54bf8d69288fbee4904_objectname -- every time you want to retrieve object 4b825dc642cb6eb9a060e54bf8d69288fbee4904, you also read 4b825dc642cb6eb9a060e54bf8d69288fbee4904_objectname to see if names match
22:53 madewokherd joined #git
22:54 thiago what are "names" for you?
22:54 thiago for Git, the SHA-1 *is* the name
22:54 alcaveman heh..
22:55 thiago don't say file name
22:55 alcaveman i am sure git knows file names
22:55 osse the file names are stored in the trees
22:55 thiago a file can be renamed. Its content remains.
22:55 thiago so it can't be the file name
22:55 alcaveman oh so git names contents..
22:56 alcaveman when i rename a file, will get copy its data again, and delete the old data?
22:56 thiago no
22:56 alcaveman just change the file name that is associated to the data hash?
22:56 thiago if the file's contents remained the same, the tree will point to the same blob object.
22:56 thiago yes
22:56 alcaveman interesting. smart
22:57 thiago try it: git ls-tree HEAD:
22:59 alcaveman i see permission, blob/tree, hash, filename
22:59 thiago that's what a tree object contains
23:00 thiago renaming a file without changing the file simply renames the entry in the tree object
23:00 hahuang61 joined #git
23:00 Groscheri joined #git
23:01 alcaveman i have an empty directory and its hash is 06a1843a854be35a1981dbd56dafd4cc22599896
23:01 alcaveman hmm sorry not empty
23:01 alcaveman has a .gitignore in it
23:03 Dong__ thiago: you are very patient. :)
23:03 * thiago is bored waiting for his flight
23:05 Dong__ i see.
23:06 Dong__ thiago: if you are so bored, any chance you can explain a bit of the diff strategy currently used in git?
23:07 Thorn__ joined #git
23:07 annoymouse joined #git
23:08 jottr joined #git
23:11 cbreak google histogram / patience diff
23:12 vassagus joined #git
23:13 mrBen2k2k2k joined #git
23:14 red__ joined #git
23:14 HappyLoa- joined #git
23:16 Dong__ cbreak: yes, I know that there are 3 diff used in git, but except looking at the source code, I failed to find any in depth documentation regarding which is used when.
23:17 mehola joined #git
23:19 algae joined #git
23:21 Atrumx joined #git
23:21 Groscheri joined #git
23:24 jefus joined #git
23:25 paws_ joined #git
23:25 thiago Dong__: the diff algorithms are pretty standard
23:26 thiago Dong__: I ve never read how they work, I'm sure that there are other resources online
23:26 jefus_ joined #git
23:26 thiago Dong__: man git-diff will tell you which one is the default
23:26 gitinfo Dong__: the git-diff manpage is available at http://jk.gs/git-diff.html
23:26 hiroki When performing an initial git clone, there is no '.git/refs/remotes/origin/master', however, once I commit and push an update (on a different machine), and then perform a fetch/pull, '.git/refs/remotes/origin/master' appears. I'm not sure I understand this behaviour, however, i'm now in search of a way to have the file appear during an initial clone, any tips ?
23:26 thiago hint, it's called "default"
23:26 thiago a.k.a. "myers"
23:27 dsdeiz joined #git
23:27 dsdeiz joined #git
23:27 Dong__ thiago: that doesn't explain what do they really mean the current diff is based on the standard but modified a bit, no where to find what is modified.
23:29 bremner hiroki: you shouldn't really rely on file structure. refs can be packed
23:30 hiroki bremner: i'm trying to do an initial clone, and then subsequently afterwards compare hashes between refs/remotes/origin/master and refs/heads/master to decide if there was a change, and therefore execute a deployment script. Any other methods to achieve a similar compare a different way ?
23:31 bremner hiroki: use "git rev-parse refs/remotes/origin/master"
23:33 sangy joined #git
23:34 hiroki bremner: tested and confirmed. Did not know rev-parse yet, thanks a million !
23:34 Thorn joined #git
23:34 Dong__ hiroki: that is what Jenkins use as well.
23:35 mingrammer joined #git
23:35 hiroki Dong__: good to know ! Jenkins is on the todo
23:36 Dong__ hiroki: if you look at Jenkins job log, you can see how it detect change.
23:36 hiroki Dong__: will definitely check
23:37 wilhelmtell joined #git
23:37 meanderer joined #git
23:38 mablae__ joined #git
23:39 LeBlaaanc joined #git
23:39 Hanumaan joined #git
23:39 Hanumaan left #git
23:40 jnewt when i run git log over ssh, the log output has a bunch of extra characters like ESC and [m  .  is this an issue with git settings,shell,ssh or what?
23:41 jnewt the things like ESC have inverted colors (ie my terminal background is black with white letters, but ESC is a white block with black letters in it
23:41 Dong__ jnewt: looks like you are using Windows?
23:41 jnewt Dong__, windows client, but i am running the commands on a freebsd server (problem doesn't occur on my local installation of git)
23:42 Dong__ those chars are control char which your windows client doesn't understand properly
23:42 cd-rum_ joined #git
23:42 lebara joined #git
23:42 cjsarette joined #git
23:43 Dong__ it's in the old days when you only got a terminal, to move cursor back and forth etc.
23:43 bremner if you're using ssh, it still is the old days
23:43 Dong__ :)
23:44 Dong__ I think ssh is pretty new in fact
23:44 ryouma jnewt: ansi control sequences
23:44 fuchstronaut joined #git
23:44 jnewt i don't seem to have an issue with standard commands, I only noticed it on git (but i don't use a lot of stuff over ssh).  how do i fix it?
23:44 ryouma jnewt: does git log use color?
23:44 bremner it can
23:44 ryouma turn off color
23:45 Dong__ or use a better client
23:45 ryouma (or figure out ansi sequences)
23:45 Dong__ I guess putty should understand that.
23:45 jnewt i don't have putty, i just have the windows git bash.
23:46 qballer joined #git
23:47 jnewt commented the [color] section in ~/.gitconfig.  problem remains
23:48 Dong__ oh just for fun, if you can get windows 10 anniversary update, you can use the Bash on Windows directly.
23:48 jnewt Dong__, I can't, i have software that runs on win7 that i can't afford to update.
23:48 Electrometro joined #git
23:49 jnewt i'd like to just fix what i have.  i disabled the colors.   what else do i need to do?
23:50 jnewt or at least i commented the color and color "status sections in ~/.gitconfig and verified the repository doesn't have a ./git/.gitconfig file.   whether this is enough to "disable colors" is unkonwn
23:51 mrBen2k2k2k joined #git
23:51 stamina joined #git
23:51 dandaman joined #git
23:53 rkazak joined #git
23:54 Dong__ jnewt: got to go sleep, hope someone knows how to solve this.
23:54 kn-928 joined #git
23:58 wilhelmtell joined #git

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