Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-08-29

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

All times shown according to UTC.

Time Nick Message
00:00 anttim joined #git
00:03 dreiss joined #git
00:03 constant joined #git
00:04 rahtgaz joined #git
00:06 Gsham joined #git
00:06 texasmynsted joined #git
00:07 inflames joined #git
00:07 MattMaker joined #git
00:08 jimi_ joined #git
00:11 jrahmy_ joined #git
00:14 romerocesar joined #git
00:17 cqi joined #git
00:17 Gitzilla joined #git
00:19 mingrammer joined #git
00:22 Literphor joined #git
00:22 MattMaker joined #git
00:23 kus joined #git
00:25 mikeh left #git
00:28 ToBeCloud joined #git
00:31 jimi_ joined #git
00:32 clemf joined #git
00:32 raijin joined #git
00:33 MattMaker joined #git
00:33 mizu_no_oto joined #git
00:34 nidr0x joined #git
00:36 nidr0x joined #git
00:37 afuentes joined #git
00:37 jimi_ joined #git
00:39 spudowiar joined #git
00:41 romerocesar joined #git
00:45 justanotheruser joined #git
00:46 dreiss joined #git
00:47 MattMaker joined #git
00:49 rubyonrailed joined #git
00:50 Marbug joined #git
00:53 MattMaker joined #git
00:58 MattMaker joined #git
01:00 darkbit joined #git
01:00 kaldoran joined #git
01:01 kaldoran joined #git
01:01 jaafar joined #git
01:04 Gsham joined #git
01:05 l4v2 joined #git
01:07 PioneerAxon joined #git
01:12 MattMaker joined #git
01:15 mizu_no_oto joined #git
01:15 jkt__ left #git
01:17 woodruffw joined #git
01:17 woodruffw joined #git
01:20 Guest81501 joined #git
01:22 kaldoran joined #git
01:22 martin_g joined #git
01:23 yoyoyo___ joined #git
01:23 Lewix joined #git
01:24 pfg joined #git
01:24 aendrew joined #git
01:24 paperziggurat joined #git
01:24 cloudcell_ joined #git
01:25 fahadash joined #git
01:25 cstrahan joined #git
01:25 unholycrab joined #git
01:26 ASOLAr joined #git
01:26 ulkesh joined #git
01:27 pizzaops joined #git
01:27 janx joined #git
01:27 bbhoss joined #git
01:27 cory_ joined #git
01:27 kaldoran joined #git
01:27 joseph_arnstein joined #git
01:27 rfv joined #git
01:27 huhlig joined #git
01:27 devhost joined #git
01:27 lebster joined #git
01:27 sebhoss joined #git
01:27 kwmiebach joined #git
01:27 MattMaker joined #git
01:28 barq joined #git
01:28 kaldoran joined #git
01:30 jimi_sanchez joined #git
01:33 MattMaker joined #git
01:34 arescorpio joined #git
01:36 A5101 joined #git
01:38 zacts joined #git
01:41 raijin joined #git
01:45 MattMaker joined #git
01:45 anuxivm left #git
01:45 raijin joined #git
01:45 cdown joined #git
01:46 mizu_no_oto joined #git
01:47 pks joined #git
01:48 PioneerAxon joined #git
01:52 raijin joined #git
01:52 monoprotic joined #git
01:52 monoprotic joined #git
01:53 MattMaker joined #git
01:53 danslo joined #git
01:53 raijin joined #git
01:55 clandestino joined #git
01:56 gfixler joined #git
01:56 Vampire0 joined #git
01:56 d^sh joined #git
01:57 chrisshattuck joined #git
01:59 Corosus joined #git
02:00 dreiss joined #git
02:00 sssilver joined #git
02:01 boombatower joined #git
02:02 rchavik joined #git
02:03 andlabs joined #git
02:03 MattMaker joined #git
02:08 mellernoia joined #git
02:08 mingrammer joined #git
02:11 mehola joined #git
02:13 MattMaker joined #git
02:18 drbean joined #git
02:19 harish joined #git
02:20 Gsham joined #git
02:22 hexagoxel joined #git
02:23 MattMaker joined #git
02:28 MattMaker joined #git
02:35 afuentes when i edit a windows file on linux, it saves the eol as 0a. but in git diff it shows a ^M at the end of the +lines
02:35 afuentes shouldnt it show it at the end of the -lines?
02:35 afuentes meaning it was on the old line but not in the new one? i checked with hd... no CR in there
02:38 chachasmooth joined #git
02:39 darkbit joined #git
02:40 fstd_ joined #git
02:40 rewt are you diffing old vs new, or new vs old?
02:41 afuentes git diff, where - us is old and + is new
02:41 javajoe joined #git
02:42 kaldoran joined #git
02:42 Dougie187 left #git
02:42 MattMaker joined #git
02:43 afuentes here, check it out yourself http://paste.debian.net/794326/
02:44 afuentes i added B1 to show what a regular line looks like... 0d 0a as you can see, windows file
02:44 afuentes i edited it in vim and saved the new line as 0a... okay, thats vims fault to not save on the same format for some reason
02:45 afuentes but the diff suggest i have a ^M on the changed line... :S
02:45 sssilver joined #git
02:46 kaldoran joined #git
02:47 jstimm joined #git
02:48 lagothri1 joined #git
02:48 jrahmy_ joined #git
02:49 kaldoran joined #git
02:50 cqi joined #git
02:51 drbean joined #git
02:51 kaldoran joined #git
02:52 MattMaker joined #git
02:53 PioneerAxon joined #git
02:53 rewt how about with -A1 or -C1 instead of -B1 to show the line after that one and possibly the line ending of the line you edited
02:53 inflames joined #git
02:55 rewt vim is usually good with the the line endings, but maybe it got messed up somehow
02:56 anonymuse joined #git
02:57 afuentes okay, its ack being an smart ass when you pipe stuff
02:58 afuentes if i hd the file directly it shows 0d 0a
02:58 afuentes but that still doesnt really explain why it shows ^M on the new line but not on the old one
02:58 afuentes :S
02:59 afuentes both have 0d 0a as far as i can tell
03:00 sandstrom joined #git
03:00 afuentes (ack is in perl, thank god i left perl not long ago... this do what you think i want, not what i asked you to is kinda terrible design ;s)
03:05 PioneerAxon joined #git
03:06 preaction ... so a questionable bug in a program means the language it was written in is "terrible"?
03:10 krs93 joined #git
03:14 rewt isn't that how that works?
03:14 rewt transitive properties and all
03:14 parallel21 joined #git
03:18 beyertech joined #git
03:19 beyertech joined #git
03:19 Puhe_ joined #git
03:21 polyzen joined #git
03:24 afuentes its not a bug... i bet if i report it i get some works as intended. Ive known enough perl developers to know is a state of mind... it goes beyond programs, it creeps on people too
03:25 afuentes from my pov, theres more than one way to do it is a terrible premise too
03:26 afuentes sorry for the rant ^^
03:26 a10c joined #git
03:28 a10c joined #git
03:28 daftbun joined #git
03:28 daftbun hi git.
03:28 gitinfo daftbun: 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.
03:29 daftbun if i do git merge [brancha] [branchb] it's not going to commit merge anything right?
03:29 _ikke_ daftbun: first, you probably never want to give git merge more then one branch
03:29 a10c joined #git
03:29 daftbun ok maybe i don't quite understand how to do it
03:29 _ikke_ (it  always merges in the current branch, given more then one branch it will do an octopus merge)
03:30 daftbun ah ok
03:30 daftbun so if i'm in branch A
03:30 daftbun i jut do git merge B
03:30 daftbun :)?
03:30 _ikke_ And yes, it will commit by default
03:30 _ikke_ or fast-forward if possible
03:30 daftbun but then i have to push it
03:30 daftbun so there's some safety
03:30 daftbun :D
03:31 rewt you can ask it not to commit
03:31 a10c joined #git
03:31 daftbun how do i do that
03:31 _ikke_ -n
03:31 _ikke_ (--no-commit)
03:32 _ikke_ oops, wrong one
03:32 _ikke_ ignore last line
03:32 daftbun ah cool
03:32 daftbun wait is it -n or --no-commit?
03:33 _ikke_ Both
03:33 daftbun ok let me try
03:33 _ikke_ one is the short form
03:33 a10c joined #git
03:34 rgrinberg joined #git
03:36 javajoe joined #git
03:37 PioneerAxon joined #git
03:38 aavrug joined #git
03:39 aavrug joined #git
03:41 a_thakur joined #git
03:41 SwiftMatt joined #git
03:42 beyertech joined #git
03:42 a_thakur joined #git
03:42 Darren_ joined #git
03:44 mmlb joined #git
03:44 zacts joined #git
03:44 Andrew_K joined #git
03:46 sagerdearia joined #git
03:51 clandestino joined #git
03:54 cqi joined #git
03:56 mingrammer joined #git
04:00 constant joined #git
04:01 constant joined #git
04:02 constant joined #git
04:02 vigilvindex joined #git
04:03 thecomedian joined #git
04:06 daey_ joined #git
04:08 jaguarmagenta joined #git
04:18 LUMIA930 joined #git
04:18 diogenese joined #git
04:19 constant joined #git
04:20 evanwang joined #git
04:22 phanimahesh joined #git
04:25 constant joined #git
04:26 constant joined #git
04:26 constant joined #git
04:26 azerus joined #git
04:27 PioneerAxon joined #git
04:28 javajoe_ joined #git
04:30 antiTORTURE joined #git
04:31 mingrammer joined #git
04:32 jrahmy_ joined #git
04:32 Anupkumar joined #git
04:33 azerus joined #git
04:34 antiTORTURE joined #git
04:35 azerus joined #git
04:36 antiTORTURE joined #git
04:36 azerus joined #git
04:38 antiTORTURE joined #git
04:38 moneylotion joined #git
04:42 antiTORTURE joined #git
04:45 beyertech joined #git
04:46 beyertech joined #git
04:47 beyertech joined #git
04:48 beyertech joined #git
04:48 adambrenecki joined #git
04:49 beyertech joined #git
04:50 freimatz joined #git
04:52 subhojit777 joined #git
04:52 PioneerAxon joined #git
04:53 azerus joined #git
04:54 netj joined #git
04:55 sssilver joined #git
04:56 anttim joined #git
04:59 azerus joined #git
05:01 daftbun hey
05:01 Andrew_K joined #git
05:02 daftbun i was just resolving some of the conflicts in my code
05:02 daftbun and now i'm trying to create a new branch with the changes
05:02 daftbun error: you need to resolve your current index first
05:02 daftbun do i need to flag that it has been resolved?
05:02 _ikke_ git status
05:02 daftbun then?
05:02 beyertech joined #git
05:03 _ikke_ What does it mention
05:03 constant joined #git
05:03 daftbun it says what has been modified deleted and renamed
05:03 daftbun i don't see any conflict?
05:03 daftbun oh
05:03 daftbun it says unmerged paths
05:04 _ikke_ ah
05:04 daftbun do i have to git add them?
05:04 _ikke_ yes
05:04 daftbun ok cool
05:04 _ikke_ and then continue the operation
05:04 _ikke_ Did you do a git merge?
05:05 daftbun i did and then it conflicted
05:05 daftbun so i resolved them manually
05:05 daftbun that's wherei was up to
05:05 daftbun before i git added
05:05 _ikke_ right
05:05 daftbun did i do something wrong?
05:05 _ikke_ then git commit to finish the merge
05:05 _ikke_ No
05:05 _ikke_ conflicts happen
05:05 daftbun thing is
05:05 daftbun i don't want to put it in this branch
05:05 daftbun i want to checkout a new branch
05:05 daftbun and keep this one as is
05:05 _ikke_ then git merge --abort
05:06 daftbun should i stash changes first?
05:06 govg joined #git
05:06 _ikke_ daftbun: All those changes came from the merge, right?
05:06 daftbun let me start from the beginning so u can understand in full
05:06 daftbun i have branch: upgrade which is where i'm on
05:06 daftbun and another branch: dev
05:06 daftbun i'm on upgrade
05:07 daftbun i then merged dev to upgrade
05:07 daftbun and it conflicted
05:07 daftbun and merged a bunch of other stuff
05:07 _ikke_ right
05:07 daftbun i then resolved upgrade's conflicts
05:07 daftbun and i want to branch off this
05:07 daftbun but i think this may be the wrong approach
05:07 l4v2 joined #git
05:07 daftbun should i have branched off dev
05:07 daftbun and then merged upgrade
05:07 daftbun then resolved?
05:08 dreiss joined #git
05:08 _ikke_ I can't tell which branch you should merge into what
05:08 daftbun dev has the latest changes
05:08 daftbun upgrade has custom modified changes
05:08 daftbun i need to have all of dev
05:08 daftbun and upgrade in a branch
05:08 _ikke_ But in a new branch?
05:08 daftbun yeah
05:08 _ikke_ ok
05:09 _ikke_ then first create a new branch (based on either dev, or upgrade, whatever makes sense to you)
05:09 _ikke_ then merge the other branch in
05:10 linuxmint joined #git
05:10 daftbun ok
05:10 daftbun so what i did
05:10 daftbun i checkout dev
05:10 daftbun i branched that to upgrade-2.0
05:10 daftbun i then merged upgrade
05:11 daftbun and got the conflicts
05:11 daftbun so
05:11 a_thakur joined #git
05:11 daftbun should i manually resolve them?
05:11 daftbun or should i do it another way
05:11 _ikke_ Yes, no matter what you do, the conflicts *will* happen
05:11 _ikke_ both branches have conflicting changes
05:11 daftbun yep
05:11 _ikke_ Resolve them, and finish the merge
05:11 daftbun so resolve manually
05:11 daftbun let me do the manual step now
05:12 _ikke_ I got to go now
05:12 daftbun waiti
05:12 daftbun after i manually resolve conflicts
05:12 daftbun do git add on the files?
05:12 daftbun right?
05:13 _ikke_ yes
05:13 _ikke_ then git commit
05:13 daftbun thanks
05:13 daftbun :)
05:14 parallel21 left #git
05:17 lvns joined #git
05:17 chrisshattuck joined #git
05:18 linuxmint joined #git
05:19 Rish joined #git
05:22 zacts joined #git
05:29 polyzen joined #git
05:29 mingrammer joined #git
05:30 nilg joined #git
05:31 chrisshattuck joined #git
05:32 linuxmint joined #git
05:35 gelei007 joined #git
05:36 chrisshattuck joined #git
05:41 ISmithers joined #git
05:41 ffabi joined #git
05:42 Ozd joined #git
05:43 a10c joined #git
05:44 Raging_Hog joined #git
05:46 dsantiago joined #git
05:47 PleaseReadThis joined #git
05:48 qt-x joined #git
05:49 Literphor joined #git
05:50 steven_a_s joined #git
05:50 ThomasLocke joined #git
05:53 steven_a_s joined #git
05:54 jknetl joined #git
05:55 steven_a_s joined #git
05:55 ISmithers I have a .gitignore that has a rule for: [Bb]in/ at the start of the file, and at the end I have an exclusion for: !Engonet.Web.Site.Azure/Engonet.Web.SiteContent/bin
05:55 ISmithers But the file is still excluded unless I comment out [Bb]in/ - any ideas how to properly override this?
05:58 vuoto joined #git
05:58 selckin if you remove the [bB] and and the '!' is it ignored? (ie check if there is a mistake/type in your long path)
05:59 mingrammer joined #git
06:00 mingrammer joined #git
06:00 beyertech joined #git
06:01 konrados joined #git
06:01 beyertech joined #git
06:02 Rish joined #git
06:02 armyriad joined #git
06:03 eijk joined #git
06:04 ISmithers Yes it still gets ignored if I do that selckin
06:05 selckin don't know then!
06:05 sssilver joined #git
06:05 selckin maybe another rule is matching it
06:05 rscata joined #git
06:06 robinsmidsrod joined #git
06:06 _ng joined #git
06:07 jceb joined #git
06:09 sword joined #git
06:11 ISmithers Is there a way to print the global ignore contents?
06:11 ISmithers Print/cat/output
06:11 justanotheruser joined #git
06:13 ISmithers OK it has committed now. I had to delete it and then re-add it. Really weird.
06:14 sandstrom joined #git
06:14 moei joined #git
06:14 selckin tracked files are never ignored
06:16 ISmithers Then it should have been there already, but it wasn't.
06:16 ISmithers So it wasn't tracked.
06:17 ISmithers Actually it looks like it _was_ tracked, but it was also ignored.
06:17 selckin impossible
06:17 ISmithers So I don't know. I had to delete it, commit that change, then add it back.
06:17 selckin impossible
06:17 ISmithers Well I can only tell you what I'm seeing. Git shows a delete.
06:17 selckin there are people here very day asking how to gitingore tracked files, it can't be done
06:18 ISmithers I was looking in the repo and it wasn't there. I know this as by myself and another developer pulled the project down and couldn't build due to the missing file.
06:18 ISmithers Apparently it can be done, through some black magic. I've tried the same thing and would normally agree with you, but this appears to have bugged out.
06:18 selckin well if you use black magic without knowing what you're doing you're on your own
06:19 selckin and should mention you tried black magic
06:19 King_Hual joined #git
06:19 jaguarma_ joined #git
06:19 ISmithers Well like most normal people, I don't actually believe in magic. I was using the term to suggest that something very strange happened, as Git was not exhibiting normal behaviour.
06:19 ISmithers It only started exhibiting normal behaviour again after I removed the file and re-added the file.
06:20 ISmithers I'm looking at the repo history now. I can see the file added in a commit, but repository pulled down at that commit SHA doesn't have the file. I can see the file removed in the next commit, then added in the subsequent one and on a re-pull the file is not present. *shrugs*
06:21 freimatz joined #git
06:27 Sceorem joined #git
06:28 fakenerd joined #git
06:29 jaguarmagenta joined #git
06:33 star_prone joined #git
06:34 blackwind_123 joined #git
06:34 jmd joined #git
06:36 star_prone joined #git
06:38 satya4ever joined #git
06:39 javajoe joined #git
06:40 yuhlw_ joined #git
06:41 fees joined #git
06:41 beyertech joined #git
06:42 jknetl joined #git
06:43 ASOLAr joined #git
06:46 d^sh joined #git
06:51 star_prone joined #git
06:51 d^sh_ joined #git
06:52 hellyeah joined #git
06:52 hellyeah quick question
06:53 hellyeah goes git have central repository approach or distributed repository approach?
06:53 selckin did you even try google
06:53 a10c hellyeah: really?
06:53 hellyeah git is dvcs
06:53 hellyeah :D
06:54 a10c https://git-scm.com/book/en/v2/Getting-Started-About-Version-Control
06:55 jceb joined #git
06:56 _ikke_ Note that there is a difference between the architecture and workflow
06:56 _ikke_ Many projects do use a single central repository to colaborate
06:57 _ikke_ But everyone still has a complete repository
06:57 hellyeah in distributed version control system like git right?
06:57 hellyeah everyone has complete repository
06:57 _ikke_ yes
06:58 d^sh joined #git
06:58 mariuscc joined #git
06:58 hellyeah i understand thanks
06:59 hellyeah i am watching a git vids from codeschool
06:59 danslo joined #git
07:00 hellyeah the vids meantion some command like git config --global user.name "..", git config --global user.email "..", git config --global color.ui true
07:00 hellyeah are you using those? Sorry for newbie question
07:01 vmiklos joined #git
07:01 submitnine joined #git
07:02 ASOLAr joined #git
07:04 GavinMagnus joined #git
07:04 planet_en joined #git
07:04 j416 joined #git
07:06 aidalgol joined #git
07:07 Darcidride__ joined #git
07:10 sandstrom joined #git
07:11 navidr joined #git
07:12 ASOLAr joined #git
07:13 armin hellyeah: most of us use those i'd say. why are you asking, afterall?
07:13 dreiss joined #git
07:15 t0by joined #git
07:16 ronskisika joined #git
07:16 nwkj86 joined #git
07:17 JeroenT joined #git
07:17 average joined #git
07:17 average if a Git object has size zero..
07:17 average does that represent a file that has not changed?
07:18 average and so, its object file (at that specific commit) is empty?
07:18 PioneerAxon_ joined #git
07:19 ASOLAr joined #git
07:19 grift joined #git
07:19 Hounddog joined #git
07:20 Raging_Hog joined #git
07:20 alansaul joined #git
07:22 _ikke_ average: no
07:22 encod3 joined #git
07:22 _ikke_ It means the file is empty
07:22 average i see
07:23 average _ikke_: that's a bit strange tho
07:23 _ikke_ git always stores the full contents of files (but only once)
07:23 star_prone joined #git
07:23 average because there's a lot of those
07:23 _ikke_ average: all with different hashes?
07:23 average and considering that Git does not let you add empty files.. i find it unusual
07:23 _ikke_ That would mean corrupt objects
07:23 _ikke_ average: git does allow empty files
07:23 average well i think they may have the same hash
07:23 average still, i remember that Git does not allow you to add an empty file by default
07:23 average maybe i'm wrong..
07:23 _ikke_ average: it does..
07:24 _ikke_ touch foo && git add foo
07:24 _ikke_ no problem
07:24 average i see
07:27 mingrammer joined #git
07:27 elect joined #git
07:28 star_prone joined #git
07:32 JeroenT joined #git
07:32 stuh84 joined #git
07:33 cyan__ joined #git
07:35 ahmedelgabri joined #git
07:37 Rish joined #git
07:37 dsantiago joined #git
07:37 rchavik joined #git
07:41 ferr joined #git
07:42 Gitzilla joined #git
07:44 Sceorem joined #git
07:45 hellyeah armin:  i am watching video it is meantioned there. So i am newbie thats why i asked
07:47 TomyWork joined #git
07:47 GavinMagnus joined #git
07:47 courrier joined #git
07:48 d1b joined #git
07:49 Rish_ joined #git
07:51 Balliad joined #git
07:52 netcarver joined #git
07:56 daumie joined #git
07:57 lvns joined #git
07:57 Fernando-Basso joined #git
07:59 t0by joined #git
08:00 danslo joined #git
08:02 harish joined #git
08:10 a_thakur joined #git
08:13 xkr47 joined #git
08:13 Oebele joined #git
08:14 beyertech joined #git
08:16 ome joined #git
08:18 govg joined #git
08:19 marcogmonteiro joined #git
08:20 duderonomy joined #git
08:21 Nitesh joined #git
08:21 mingrammer joined #git
08:23 moritz average: the thing that git doesn't track is empty directories
08:24 moritz (or directories in general; it just creates them on the fly when they are necessary for files)
08:24 average i see
08:26 jas4711 joined #git
08:29 stasmakarov joined #git
08:30 star_prone joined #git
08:32 JanC_ joined #git
08:33 MrWoohoo joined #git
08:33 carlo joined #git
08:37 Erjiin joined #git
08:41 chll_ joined #git
08:46 al-damiri joined #git
08:47 aidalgol joined #git
08:48 arup_r joined #git
08:48 arup_r left #git
08:50 Macaveli joined #git
08:50 govg joined #git
08:50 mago__ joined #git
08:52 mago__ joined #git
08:53 mago__ hey. in my linux tree, i have a branch where i put keep my own kernel work. Sometimes i want to backport stuff from newer kernels into my branch. What is the best way to manage this using git? should i cherry pick? if i do, what will happen later on when i merge that same branch into my branch as I upgrade to the kernel version from which i backported stuff?
08:55 tvw joined #git
08:57 raijin joined #git
08:57 _ikke_ mago__: is it usually just single features / patches you want to backport?
08:57 mago__ _ikke_: yeah, features.. but in some cases, there are 10+ commits belonging to that feature
08:57 ugjka_ joined #git
09:01 LeRieur joined #git
09:02 _ikke_ sure, but I mean, not just get everything from upstream
09:03 _ikke_ Because in that case, cherry-pick is the only option
09:03 _ikke_ When merging back, nothing should happen, except that you'll see a commit twice in the history (with different hashes)
09:03 Repox joined #git
09:05 mago__ _ikke_: and git is fine with that? a cherry picked change, and then i merge the same change with another hash. will it cause a conflict?
09:05 dsdeiz joined #git
09:06 _ikke_ No, because it will see the same content on both sides, which is not a conflict
09:08 evanwang joined #git
09:10 harish joined #git
09:13 armin mago__: in any case, make an own branch for it, try to do your work (read: "apply those 10 cherry-picks") there. if that works out for you, you're ready to merge/rebase. in any case, branch often if you do work like that.
09:14 testild joined #git
09:14 armin mago__: also consider rebasing over merging for such things.
09:16 leeN joined #git
09:18 dimi1947 joined #git
09:23 elect joined #git
09:31 harish joined #git
09:35 bjoe2k4 joined #git
09:42 n-st joined #git
09:45 swift1 joined #git
09:45 brianrhude joined #git
09:47 brianrhude hello, how do you "git show" a list of revisions from a single branch? say you have this history A, B, C, D, E, F, can you "show" D, E and F in a single command? I cannot grasp how to specify ranges
09:47 _ikke_ brianrhude: git log
09:48 _ikke_ git show is for single things
09:48 _ikke_ if you add --no-walk to git log, then you can select multiple commits by specifying them explicitly
09:48 _ikke_ (otherwise it would show everything reachable from the commits)
09:49 brianrhude _ikke_: my need is actually to export a list of patches - show would output the diff - can you do it via log?
09:51 _ikke_ brianrhude: -p
09:51 _ikke_ you can also use git format-patch, which would result in mbox patch files
09:51 brianrhude _ikke_: super, thank you!
09:52 pijiu2 joined #git
09:53 rahtgaz joined #git
09:54 mingrammer joined #git
09:54 timothy joined #git
09:54 pijiu joined #git
10:02 stfn joined #git
10:03 brianrhude _ikke_: still didn't manage to specify the range, so I solved by looping: for i in {4..0}; do git show HEAD~$i >> /tmp/bundle.patch; done
10:03 OnkelTem joined #git
10:03 OnkelTem Hi all
10:04 OnkelTem I was trying to update commit messages of my last commits and used this instruction: https://help.github.com/articles/changing-a-commit-message/
10:04 OnkelTem but as it always happens with git - everything went wrong
10:04 OnkelTem http://apaste.info/N4r
10:05 jast !conflict
10:05 gitinfo [!eekaconflict] Merge conflicts are a natural part of collaboration. When facing one, *don't panic*. Read "How to resolve conflicts" in man git-merge and http://git-scm.com/book/ch3-2.html#Basic-Merge-Conflicts then carefully go through the conflicts. Picking one side verbatim is not always the right choice! A nice video explaining merge conflicts: https://www.youtube.com/watch?v=zz7NuSCH6II
10:05 OnkelTem So I've launched rebased, corrected commit messages in text editor and got this
10:05 ljc joined #git
10:05 achlys joined #git
10:05 govg joined #git
10:05 jast when 'git rebase -i' opened an editor for you to change the list of actions, did you remove or reorder any lines?
10:06 OnkelTem No! of course
10:06 OnkelTem I just used "r" (for reword) in few lines
10:06 jast okay, good :)
10:06 OnkelTem adding #task_id where needed (I'm always forgetting to do this)
10:06 jast I think I can tell what happened there
10:06 OnkelTem Oh thanks!
10:07 jast rebase and merge don't mix particularly well. one of the commits between HEAD~4 and HEAD was probably a merge commit
10:07 jimi_sanchez joined #git
10:07 jast normally, rebase "linearizes" merges, meaning it replaces them with a curiously ordered sequence of all the commits that got merged in in that place
10:08 eijk_ joined #git
10:08 jast it tends to be a good idea to avoid rebasing merges altogether, but for something as simple as rewording, it's probably fine here, assuming you haven't pushed the commits-to-be-altered yet (as usual)
10:08 OnkelTem Well, I see now a "both modified" file
10:08 OnkelTem in git status
10:09 jast 'git rebase --abort' and start over, this time include the -p flag in your rebase to make it *p*reserve merges
10:09 brianrhude left #git
10:09 jast yeah, as I said, linearizing merges does an interesting reordering on commits, so conflicts are almost guaranteed
10:09 Macaveli joined #git
10:09 jast with -p you can basically avoid that
10:10 karoal joined #git
10:10 OnkelTem I didn't know I could somehow "start over" :)
10:10 OnkelTem Trying
10:10 OnkelTem but before that I'd like ask: currently my working tree is not clean
10:10 jast it even told you about 'git rebase --abort' :}
10:11 OnkelTem what am I supposed to do with a new file in stage and one - both modifiled?
10:11 jast well, rebase refuses to start if you had any uncommitted changes in your tree, so any stuff smudging it up right now should either be untracked changes (irrelevant) or part of the rebase
10:11 OnkelTem so I have to resolve the merge conflict I got, right? A file which is "both modified"
10:12 CheckDavid joined #git
10:12 jast nah. the conflict is part of the rebase, so aborting the rebase will restore the old file(s), too
10:12 OnkelTem ok
10:12 OnkelTem jast: I love you!
10:12 OnkelTem seems aborted
10:12 Gonzo89 joined #git
10:12 jast that's nice... but I'm not single :P
10:13 OnkelTem jast: :)
10:13 jast in fact you can even undo a rebase if you notice it screwed things up and it's already finished
10:14 jast so, all in all, while rebases can do very strange things, they're not too dangerous if you take a moment to review the result (or notice things aren't going as planned while it's still in progress)
10:15 Aces_Charles joined #git
10:15 jast of course, rebase published commits at your own (and everyone else's) peril
10:15 OnkelTem Now when I started rebase with -p as you said - I see those merges, a lot of
10:17 jast side note, some people here might not appreciate the "things always go wrong with git" mentality, and be much less motivated to help you. I do understand this stuff can be frustrating, but on the other hand we can fix the vast majority of issues fairly quickly.
10:17 jast with -p you shouldn't be seeing that many entries in the rebase list
10:18 jast actually I might be wrong about that, I haven't actually ever used rebase with -p...
10:19 OnkelTem jast: I did it!
10:19 eijk_ joined #git
10:20 OnkelTem jast: thanks for the help
10:20 jast you're welcome
10:20 Furai joined #git
10:21 OnkelTem and sorry for the emotions :-/
10:22 kbs joined #git
10:24 jrnieder joined #git
10:25 Smirnov joined #git
10:25 Smirnov joined #git
10:30 Raging_Hog joined #git
10:31 anuxivm joined #git
10:31 star_prone joined #git
10:32 Kuukunen joined #git
10:35 l4v2 joined #git
10:39 phanimahesh joined #git
10:40 ljc joined #git
10:44 Puhe_ joined #git
10:45 stamina joined #git
10:46 Ozd joined #git
10:47 jmd joined #git
10:47 lagothrix joined #git
10:47 jmd How can I list all the changed files in a commit ?
10:47 osse jmd: git show --name-only {commit}
10:47 osse jmd: git diff --name-only commit~ commit
10:48 jmd osse: Thanks
10:48 star_prone joined #git
10:48 mingrammer joined #git
10:51 govg joined #git
10:52 sandstrom joined #git
10:52 freimatz joined #git
10:53 ash_workz joined #git
10:53 Raging_Hog joined #git
10:55 tilerendering joined #git
10:55 tilerendering joined #git
10:55 selckin joined #git
10:57 Andrew_K joined #git
10:57 harish joined #git
10:59 JeroenT joined #git
11:01 anttim joined #git
11:03 ronskisika joined #git
11:05 ash_workz joined #git
11:07 rafalcpp joined #git
11:09 planet_en joined #git
11:12 mizu_no_oto joined #git
11:14 harish joined #git
11:16 sdothum joined #git
11:18 HoierM joined #git
11:19 ojdo joined #git
11:19 tvw joined #git
11:23 tabrez joined #git
11:24 JeroenT joined #git
11:27 JeroenT_ joined #git
11:28 asd5a joined #git
11:28 vuoto joined #git
11:29 JeroenT_ joined #git
11:30 bongjovi joined #git
11:30 _joes___ joined #git
11:32 bauruine joined #git
11:36 vuoto joined #git
11:36 d0nn1e joined #git
11:36 harish joined #git
11:40 star_prone joined #git
11:41 achlys joined #git
11:42 javajoe joined #git
11:44 jceb joined #git
11:44 phanimahesh joined #git
11:48 DolphinDream joined #git
11:53 Gamecubic joined #git
11:53 rnsanchez joined #git
11:53 star_prone joined #git
11:54 rf314 joined #git
11:54 rf314 left #git
11:55 ljc joined #git
11:57 tomog999_ joined #git
12:01 Darcidride joined #git
12:01 danslo joined #git
12:06 iateadonut joined #git
12:06 anonymuse joined #git
12:06 ahmed_elgabri joined #git
12:07 holodoc joined #git
12:07 iateadonut i want to deploy to a staging site.  the staging site has checked out branch 'master'.  when i do: 'git push staging master' i have to login to staging and hard reset the latest commit.
12:08 iateadonut i wanted it to just checkout automatically.  is there a way to do that?
12:08 Vampire0 iateadonut, !deploy
12:08 gitinfo iateadonut: 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
12:08 jceb joined #git
12:09 moritz iateadonut: other ways to build a deployment system: https://leanpub.com/deploy
12:10 Sceorem joined #git
12:11 cyan__ joined #git
12:12 a_thakur joined #git
12:13 rgrinberg joined #git
12:13 tlaxkit joined #git
12:15 rgrinberg joined #git
12:15 javajoe joined #git
12:15 evanwang_ joined #git
12:17 str joined #git
12:19 _ikke_ joined #git
12:25 sanketdg joined #git
12:28 Prisoner-2460_1 joined #git
12:28 constant joined #git
12:30 andlabs joined #git
12:31 Darren_ joined #git
12:32 jeffreylevesque joined #git
12:32 elect joined #git
12:35 ahmed_elgabri joined #git
12:36 mingrammer joined #git
12:38 bjoe2k4 joined #git
12:40 jimi_ joined #git
12:40 BrianBlaze420 joined #git
12:41 anuxivm joined #git
12:43 jimi_ joined #git
12:43 harish joined #git
12:43 JeroenT joined #git
12:43 JeroenT joined #git
12:46 anuxivm joined #git
12:52 _ikke_ joined #git
12:53 LUMIA930 joined #git
12:55 beyertech_ joined #git
12:55 kadoban joined #git
12:56 digidog joined #git
12:58 emptynine joined #git
12:58 anonymuse joined #git
13:00 ronskisika joined #git
13:03 Darcidride_ joined #git
13:04 ExoUNX joined #git
13:04 mithenks joined #git
13:05 orfeo left #git
13:09 boushi joined #git
13:09 Gsham joined #git
13:13 JeroenT joined #git
13:14 cdg joined #git
13:17 howdoi joined #git
13:19 mkopriva joined #git
13:20 dimi1947 joined #git
13:24 Gianormagantrous joined #git
13:25 Drzacek joined #git
13:25 Drzacek Hi
13:25 i7c ohai
13:26 lucz joined #git
13:26 Drzacek If I'm in my git project dir, and I create new branch using "git branch newbranchname", and then I checkout to that new branch, what files do I have there?
13:26 achlys joined #git
13:26 fmcgeough joined #git
13:26 i7c The working dir won’t change at all, I think.
13:26 rpd Probably whatever files existed in the git project dir
13:26 Drzacek Is it copy of current branch?
13:27 rpd I wouldn't call it a copy
13:27 i7c Tracked files are the same because you just created the new branch, and untracked files will remain untracked files.
13:27 tobiasvl Drzacek: !float
13:27 gitinfo Drzacek: If you have made a change in your working directory and have NOT YET COMMITTED, you may "float" that change over to another (`git checkout oldbranch`) or new (`git checkout -b newbranch`) branch and commit it there.  If the files you changed differ between branches, the checkout will fail.  In that case, `git stash` then checkout, and `git stash apply` and go through normal conflict resolution.
13:27 rpd but something like that
13:27 tobiasvl that what you're wondering about?
13:27 Drzacek No, I had no uncommited changes
13:27 tobiasvl OK, then what's your question?
13:28 tobiasvl remember that a branch is just a pointer at a commit. so if you create a new branch, that will point at the commit you branched from
13:28 Drzacek I checkouted to new branch, and git log shows me all commits I made
13:28 Drzacek I thought it would be clean new branch
13:28 rpd why?
13:28 i7c You thought wrong. Usually that’s not what you want.
13:28 rpd A branch is a branch off of where you came from
13:29 NeXTSUN joined #git
13:29 rpd like think of a tree branch, everything before the split will still exist on every branch you could create
13:29 rpd the trunk doesn't go away because of a branch, so to speak
13:29 Drzacek so if my HEAD was on Branch1 when I made new branch, then the new branch is based on Branch1 last commit?
13:29 stuh84 joined #git
13:30 tobiasvl yes
13:30 tobiasvl what else would you expect? an orphan branch?
13:30 tobiasvl !orphan
13:30 gitinfo To create an orphaned branch(unrelated to existing history): `git checkout --orphan newbranchname`. This will leave your index/worktree as-is(use `rm .git/index; git clean -dfx` to delete EVERYTHING). If you get 'error: unknown option `orphan`' see !orphan_old. For an empty/null commit see !orphan_null.
13:30 tobiasvl I don't understand the scenario you thought would happen
13:30 tobiasvl what's a "clean new branch"?
13:31 Drzacek this is exactly what I expected, just wanted to confirm
13:31 tobiasvl ah ok.
13:31 mingrammer joined #git
13:31 BSaboia joined #git
13:32 lss8 joined #git
13:33 axisys joined #git
13:33 fk left #git
13:33 mozzarel1 joined #git
13:35 kpease joined #git
13:35 fk joined #git
13:35 Rish joined #git
13:36 star_prone joined #git
13:36 WayToDoor joined #git
13:36 WayToDoor joined #git
13:38 rnsanchez_ joined #git
13:40 hellyeah joined #git
13:40 hellyeah sup
13:41 jaguarmagenta joined #git
13:42 GodGinrai joined #git
13:42 osse whaddu pbruh
13:44 crelix joined #git
13:44 DieguezZ joined #git
13:45 paws_ joined #git
13:46 bremner notmuch
13:46 RichiH bremner: fuenf mark in die schlechte wortspielkasse
13:46 bremner yay!
13:49 anonymuse joined #git
13:50 cjbrambo joined #git
13:52 hackel joined #git
13:54 jaafar joined #git
13:56 AaronMT joined #git
13:59 phaleth joined #git
14:03 azwieg104 joined #git
14:04 danslo joined #git
14:04 sarbs joined #git
14:07 robotroll joined #git
14:07 JeroenT joined #git
14:08 M-shine joined #git
14:08 v60r joined #git
14:09 Raging_Hog joined #git
14:09 griffindy joined #git
14:11 fusionx86 joined #git
14:12 Prisoner-2460_1 joined #git
14:12 grift joined #git
14:13 tvw joined #git
14:13 terminal_echo joined #git
14:15 subhojit777 joined #git
14:15 hashpuppy joined #git
14:18 rahtgaz joined #git
14:18 Raging_Hog joined #git
14:20 terminal_echo joined #git
14:21 ShekharReddy joined #git
14:24 prasanth joined #git
14:25 the193rd left #git
14:25 mingrammer joined #git
14:25 tabrez joined #git
14:26 PioneerAxon joined #git
14:27 grift joined #git
14:28 Faylite joined #git
14:28 crose joined #git
14:29 v60r PSA:  https://youtu.be/E9swS1Vl6Ok
14:30 Gsham joined #git
14:30 mangolisk joined #git
14:30 PioneerAxon joined #git
14:30 Gurkenglas joined #git
14:30 sssilver joined #git
14:31 NightStrike joined #git
14:32 cjbrambo joined #git
14:33 eijk joined #git
14:36 kanyeezy joined #git
14:36 kanyeezy joined #git
14:37 morenoh149 joined #git
14:37 achlys_ joined #git
14:38 Raging_Hog joined #git
14:38 sargas joined #git
14:39 t0by joined #git
14:40 nidr0x joined #git
14:41 johnmilton joined #git
14:42 fakenerd joined #git
14:43 nidr0x joined #git
14:44 eftov joined #git
14:45 Gianormagantrous joined #git
14:45 sanketdg joined #git
14:46 sanketdg joined #git
14:46 fahadash joined #git
14:48 UTAN_dev joined #git
14:48 elect joined #git
14:49 harish joined #git
14:53 daynaskully joined #git
14:54 tom[] joined #git
14:55 shinnya joined #git
14:55 Raging_Hog joined #git
14:57 clemf joined #git
14:58 dmto joined #git
15:00 dsdeiz joined #git
15:00 dsdeiz joined #git
15:01 jknetl joined #git
15:01 choki joined #git
15:03 watabou joined #git
15:05 bsanford joined #git
15:07 LUMIA930` joined #git
15:08 KrasPvP joined #git
15:09 vassagus joined #git
15:10 zivester joined #git
15:11 harish joined #git
15:12 rubie joined #git
15:14 Gitzilla joined #git
15:14 cyan__ joined #git
15:15 thiago joined #git
15:15 miklcct joined #git
15:17 _ikke_ dsfsdfsdfsd
15:17 i7c That’s what I thought.
15:19 Darren_ joined #git
15:20 Remram[m] joined #git
15:20 Ineentho[m] joined #git
15:20 M-mistake joined #git
15:20 M-meznak joined #git
15:20 f3r70rr35f joined #git
15:21 marcogmonteiro joined #git
15:22 BSAlb joined #git
15:22 glebihan joined #git
15:22 BSAlb joined #git
15:22 BSAlb joined #git
15:23 kini joined #git
15:23 BSAlb joined #git
15:25 GavinMagnus left #git
15:25 thiago joined #git
15:26 snowkidind joined #git
15:26 ahmedelgabri joined #git
15:27 exarch joined #git
15:27 courrier I've modified a file for test purposes and then pulled, creating a conflict, following git howtos I've checkout this_file --theirs but it stills wants me to commit, I don't want to commit a confict solving for a silly test, what's the cleanup way to recover from this?
15:28 ochorocho__ joined #git
15:28 _ikke_ git merge --abort
15:28 j416 courrier: you want to throw that change away, I suppose?
15:29 NeXTSUN joined #git
15:29 SwiftMatt joined #git
15:30 editshare-ericf joined #git
15:30 joki joined #git
15:30 jceb joined #git
15:30 Gurkenglas joined #git
15:31 vassagus joined #git
15:31 courrier j416: the silly change on that file, yes I want to throw it away, but there are other changes I'd like to keep (they don't conflict though), I would only restore that file to their version and pretend no conflict happened with it
15:31 settermjd joined #git
15:31 courrier that was not a pull, but a pull and a stash pop
15:31 courrier the stash pop created the conflict
15:32 courrier only on one file
15:32 Macaveli joined #git
15:32 courrier I want to keep other modifications of the stash though
15:32 j416 courrier: start with !backup
15:32 gitinfo courrier: Worried about your data while trying stuff out in your repo? The repository in its entirety lives inside the .git directory in the root of your work tree so to backup everything `cp -a path/to/workdir path/to/backup` or equivalent will suffice as long as the repo is not modified during backup. See also http://sethrobertson.github.com/GitBestPractices/#backups
15:32 j416 courrier: then do what _ikke_ says
15:33 j416 courrier: then commit your things before you try to merge/rebase
15:33 badloop i'm guessing i just having found the correct way to do this, but i'm trying to automate the syncing of code from my repository to different tiers of my environment.
15:34 badloop example:  I put a CI script in my personal repo, and it executes a script that runs a "git pull" on the repository that sits on my DEV server
15:34 j416 courrier: (avoiding committing the change to that file, by either selectively adding it, or by replacing it with an unchanged file first (checkout with the file pathe can do this))
15:35 badloop however every time i make a change and try to issue a git pull, git acts as if my DEV server tried to make changes to the file, since i had to manually set the script file to executable on the server
15:35 j416 badloop: you _could_ simply fetch from the CI regularly, to avoid unnecessary contact between dev and CI
15:35 j416 badloop: also, use fetch + reset instead of pull if on an unattended server, to avoid potential conflicts
15:36 badloop j416: was hoping to preserve the change times of individual files... :-(
15:36 stamina joined #git
15:36 badloop though i could rely on git for that i suppose...
15:36 jrahmy_ joined #git
15:36 badloop but its far less elegant
15:36 j416 badloop: they'll be as preserved as if you would have done a pull.
15:36 courrier j416 git merge --abort fails because There is no merge to abort (MERGE_HEAD missing)
15:36 badloop j416: but not with a reset? correct?
15:36 Dougie187 joined #git
15:36 j416 badloop: not correct
15:36 courrier j416 git checkout config/speech_mapping.json fails because path 'web_asker_api/config/speech_mapping.json' is unmerged
15:37 badloop j416: hmm... let me run a few tests... one sec
15:37 j416 courrier: `git status` should give you a hint on how to get out of the situation
15:37 j416 courrier: (were you rebasing?)
15:37 badloop j416: well ok... it is correct, in my case, but there's a valid reason....
15:37 j416 courrier: (applying stash?)
15:38 badloop part of my scripting alters the files to be executable, since in this case i am versioning shell scripts
15:38 dreiss joined #git
15:38 j416 badloop: git tracks the executable bit
15:38 badloop and git doesn't maintain file permissions
15:38 badloop j416: uh... ok maybe i read wrong
15:38 j416 badloop: executable bit is the only permission git tracks.
15:38 badloop interesting...
15:38 badloop i wonder where its getting lost then....
15:38 badloop hmmmmm
15:39 courrier j416: no rebasing, just pulled and then applyied stash, the conflcit happened there because that silly change in my stash on a fil that had been updated by the pull
15:39 courrier it advises git reset or git add
15:39 courrier so maybe reset
15:39 j416 courrier: yep
15:40 Noldorin joined #git
15:40 vassagus joined #git
15:41 courrier j416: but reset restored my silly test, so maybe now I need a checkout on that file?
15:41 j416 courrier: nod
15:41 j416 courrier: or, simply "git add" the changes you want to keep, commit those, then reset --hard to remove anything else
15:42 j416 courrier: ("git diff" to check what would be removed by a reset)
15:42 badloop j416: ahhhh! i think i found it
15:42 j416 badloop: good for you
15:44 courrier j416: ok thanks that's fixed
15:44 courrier but why are git conflicts so obscur to solve compared to svn...
15:44 j416 courrier: good rule of thumb: always commit changes you care about; it's hard or impossible to recover changes that have never been committed, easier to recover thcanges that were once committed, and easy as pie to remove commits that you don't need
15:44 j416 courrier: obscure?
15:47 thiago courrier: conflicts are easy to solve .They have the same format as on svn.
15:47 j416 courrier: another rule of thumb: avoid trying to merge stuff that has not been committed, it'll only cause headache. commit first, then merge.
15:49 Erati joined #git
15:51 tyreld joined #git
15:53 courrier In my souvenir, when a svn conflict happens, you can then choose to take version "ours", "theirs", or manually fix the conflict, only the last option modifying the files to add these <<<<<< theirs and so on
15:53 SourTrout joined #git
15:53 courrier Most of the time when I meet a conflict this is for a silly reason so I would be happy in forcing their version
15:53 thiago I don't remember svn offering any help in solving conflicts when you svn up'ed with uncommitted changes
15:54 j416 courrier: git has that, or if you prefer you can use a visual merge tool. But editing the file directly is often the most straightforward, in my experience.
15:54 thiago courrier: use a tool like kdiff3 to solve the conflict then
15:54 thiago courrier: run: git mergetool
15:55 zacts joined #git
15:55 hobodave joined #git
15:56 silverfall joined #git
15:57 vassagus joined #git
15:57 courrier thiago: I already tried mergetool, but you have to fix and then commit, I don't want a useless commit for a silly test, an option "just ignore local changes of this file and pretend I didn't change it" would suit perfectly
15:58 sssilver_ joined #git
15:58 thiago courrier: if you want to ignore local changes, you need to drop the commits you made
15:59 Dougie187 joined #git
15:59 j416 thiago: (not sure he made commits, did he, hm)
15:59 thiago mergetool doesn't operate on uncommitted changes
15:59 ASOLAr joined #git
15:59 thiago merge doesn't operate on uncommitted changes
15:59 j416 ah, they're just floated unless conflict
15:59 j416 right
15:59 madewokherd joined #git
15:59 thiago rebase refuses to run if there are uncommited changes
15:59 j416 thanks for reminding me
16:00 thiago it might be a case of stash pop or checkout -m, though
16:00 maxorator joined #git
16:00 cdg joined #git
16:01 courrier thiago: this is precisely a stash pop that created my conflict, one part of the stash was safe, the rest was silly and I wanted to ignore it
16:02 thiago git checkout HEAD -- thatfile
16:02 hobodave_ joined #git
16:02 thiago I think --ours also works instead of HEAD, but I never use that
16:03 badloop j416: ah that got it.  i guess those scripts must not have had +x when i originally committed them.
16:03 a_thakur joined #git
16:03 DANtheBEASTman is it possible to have git `cd` into a directory from `git clone` ?
16:04 j416 badloop: probably
16:04 crayon joined #git
16:04 j416 DANtheBEASTman: you can clone into the current directory, if that is what you mean
16:05 j416 DANtheBEASTman: git clone some/remote .
16:05 danslo joined #git
16:05 thiago DANtheBEASTman: you mean after it's done?
16:06 DANtheBEASTman j416: that wouldn't be ideal. thiago: yeah
16:06 phanimahesh joined #git
16:06 thiago DANtheBEASTman: it's impossible. A child process cannot affect the parent process.
16:06 j416 DANtheBEASTman: git clone some/remote && cd blabla
16:07 DANtheBEASTman j416: well of course. i'm trying to save myself from having to `&& cd repo` all the time
16:07 thiago if you want that, write a shell function (not a script!) that clones and cds
16:07 vassagus joined #git
16:08 j416 DANtheBEASTman: another alternative is to figure out why you clone so often that it becomes a problem to type cd repo
16:08 DANtheBEASTman i currently have `gcl () { if [[ -n "$1" ]]; then case "$1" in; (http* | https* | git* | ssh*) repo="$1"  ;; (*) repo="https://github.com/$1"  ;; esac; shift; git clone "$repo" "$@"; fi; }`
16:08 DANtheBEASTman j416: i try out a lot of different things.
16:09 j416 DANtheBEASTman: with a lot of different repos?
16:09 DANtheBEASTman yes. so many fun things on github
16:09 j416 interesting
16:09 DANtheBEASTman i can't think of a reliable way to get the directory cloned into... i'll see what #bash says
16:10 thiago you can get that github behaviour with: git config --global url.https://github.com/.insteadOf gh:
16:10 thiago then you can do: git clone gh:foo/bar
16:10 settermjd joined #git
16:10 j416 thiago: will that set the remote to gh:foo/bar as well or will it resolve it and put the real url in the config?
16:11 j416 (yes I could try but perhaps you know off top of head)
16:11 thiago j416: I don't remember
16:11 j416 alright
16:11 thiago j416: I don't use this. I used to, some 8 years ago.
16:11 thiago I think it doesn't rewrite in the config.
16:11 j416 thiago: any reason why you stopped?
16:12 thiago yep
16:12 thiago stopped contributing to the project that used it
16:13 j416 checked now; it actually uses the insteadof-url in the config, i.e. gh:foo/bar instead of https://github.com/foo/bar
16:13 mingrammer joined #git
16:13 j416 thiago: I see
16:13 thiago how come no one has registered the user foo with repository bar yet?
16:14 j416 user foo exists at least
16:14 j416 indeed,irresponsible of him to not have a bar repo
16:14 LeRieur joined #git
16:14 thiago should be a copy of GNU Hello
16:14 j416 poor guy who nicks foo
16:15 j416 must get a ton of highlights, all the time
16:15 ash_workz I am faced with sort of a difficult situation
16:15 ash_workz I should make a diagram, I think
16:15 IrcChlame550 joined #git
16:16 IrcChlame550 left #git
16:16 InfoTest joined #git
16:23 sshine joined #git
16:23 sshine I'm using git-svn and the SVN repo changed server. it seems that if I just update the '[svn-remote "svn"]' url in .git/config, then 'git svn rebase' will just hang. not sure if it's an authentication problem. how do I find out?
16:26 mastro joined #git
16:27 SourTrou_ joined #git
16:29 rishi joined #git
16:30 dreiss joined #git
16:30 tristanp joined #git
16:30 ToBeCloud joined #git
16:30 vbgunz joined #git
16:30 l4v2 joined #git
16:31 Alenah joined #git
16:31 vbgunz hello fellas, I have a question. When git is actually ignoring files, these files are no longer affected by git reset --hard? in other words, when I git reset --hard I expected these files to disappear but they don't. It's my .gitignore keeping them around?
16:33 rpd Those files are no longer tracked by git
16:33 rpd So git reset doesn't consider them as much as any other git command doesn't consider them
16:34 rpd If you want to delete a file, you can just delete it. It won't remove them for you because you told it not to.
16:34 vbgunz reset is not acting like I thought it did. I thought a git reset --hard would bring the entire working directory to the exact state it was in. this isn't true.
16:34 kadoban vbgunz: .gitignore has no effect at all on 'git reset', AFAIK. Perhaps you're looking for 'git clean'?
16:34 vbgunz well, I just added a regular file that is not ignored and did another git reset --hard. that file is still there
16:35 vbgunz kadoban: thanks, I think I am
16:35 rpd git reset doesn't do anything to untracked files, though?
16:35 tlaxkit joined #git
16:35 vbgunz I was just sort of expecting for new files to vanish. that's not happening, I'll look into clean
16:35 vbgunz thanks fellas!
16:37 vbgunz clean -f -x is what I think I'm looking for
16:37 vbgunz thanks, that was it!
16:37 Dro joined #git
16:37 NeXTSUN joined #git
16:39 cagmz joined #git
16:39 emPi joined #git
16:39 rgrinberg joined #git
16:40 diogenese joined #git
16:41 Remram[m] joined #git
16:41 Remram[m] joined #git
16:41 nitric joined #git
16:41 durham joined #git
16:42 Rish joined #git
16:44 SourTrout joined #git
16:44 chrisshattuck joined #git
16:44 EvilPenguin joined #git
16:44 LionsMane joined #git
16:44 Kicer86 joined #git
16:45 Raging_Hog joined #git
16:46 choki joined #git
16:46 tabrez joined #git
16:49 choki joined #git
16:51 Dougie187 joined #git
16:51 Prisoner-2460_1 joined #git
16:53 kline joined #git
16:57 ahammond joined #git
16:59 freimatz joined #git
17:01 ahammond I don't know exactly where to ask this question. I'm looking for a best practices document around repository granularity. This seems painfully obvious to me: a single repository should contain only the code directly relevant to the artifact it generates when built, and that artifact should be composed of only a very few RPMs which are extremely deeply coupled. Related artifacts (for example libraries upon which the other build depends) should be in
17:01 ahammond their own repos with their own builds.
17:01 ahammond However, apparently I need to explain this to a team and having something well worded with pretty graphics might help.
17:01 HoierM joined #git
17:03 peepsalot joined #git
17:03 chrisshattuck joined #git
17:06 Groscheri joined #git
17:07 madduck_1 left #git
17:07 madduck joined #git
17:08 mingrammer joined #git
17:08 Alenah joined #git
17:10 anuxivm joined #git
17:11 mkoskar joined #git
17:11 _ikke_ ahammond: I don't know any articles describing it
17:12 lvns joined #git
17:13 thiago___ joined #git
17:14 skylite joined #git
17:14 SourTrou_ joined #git
17:15 SteffanW joined #git
17:15 lvns_ joined #git
17:15 TSS_ joined #git
17:17 SourTrout joined #git
17:19 tabrez joined #git
17:21 achlys joined #git
17:21 dviola joined #git
17:22 achlys joined #git
17:23 AnarchyAo joined #git
17:24 Spec-Chum joined #git
17:25 achlys_ joined #git
17:26 Sasazuka joined #git
17:28 shurnormal What do people use? git? gitk? gitg? gitv? tig=
17:29 bremner shurnormal: use for what?
17:29 deepy I suspect I use at least half of those
17:29 shurnormal for log inspection
17:29 chrisshattuck joined #git
17:30 shurnormal pre merge reviews, stuff like that
17:30 bremner git log, gitk, magit, tig
17:31 ochorocho__ joined #git
17:31 kline joined #git
17:31 thiago I usually review on websites
17:32 thiago most of the good review tools have web tools that you can use to submit comments, most often on a per-line basis.
17:32 thiago joined #git
17:33 thiago the one exception I have is patches by email. So I just review in my mail client.
17:33 srcerer joined #git
17:33 shurnormal On which project do you receive patches by email?
17:33 _ikke_ git.git
17:33 _ikke_ linux
17:33 thiago Subsurface
17:34 thiago Wayland, but I don't usually read the emails
17:34 thiago only when the subject catches my attention
17:34 ArchNoob joined #git
17:35 shurnormal I just developed a hack for gitk and sent an email infoing the author, if he doesn't reply in some days I'll remail but pasting the patch.
17:36 rubyonrailed joined #git
17:36 m0viefreak joined #git
17:36 shurnormal I wanted to do this small survey in order to find out if most you kept using gitk.
17:37 planet_en joined #git
17:37 _ikke_ I mostly use tig
17:38 thiago I haven't used gitk in at least 8 years
17:38 shaggycat joined #git
17:39 shurnormal If I tell `tig`, it doesn't show the stash, if I tell `tig --all` it shows the most recent stash, but what about older ones?
17:39 thiago they will not be searched
17:41 _ikke_ stashes are not reachable through refs
17:41 planet_en left #git
17:41 achlys joined #git
17:41 osse one of them is
17:41 sangy joined #git
17:41 osse tee hee
17:41 sangy Hello, is there a way to compute the merge of two blob objects?
17:42 thiago sangy: yes. Git merge does it.
17:42 thiago that's also the only way to do it: via git merge
17:42 anonymuse joined #git
17:43 Antares joined #git
17:43 sangy thiago: cool, thanks!
17:43 Antares https://vk.com/rexant
17:43 shurnormal Correct! And those tools tend to parse `show-ref` when requested --all, but you can merge that list with `git stash list --pretty=format:"%H %gD"` and then you have all the stashes in the tree! ^^
17:43 _ikke_ osse: what ref?
17:43 sangy thiago: I got this though :[ error: ab436b3d7bfbf09f25286a3ae43e45bee4db52ff: expected commit type, but the object dereferences to blob type
17:44 thiago sangy: do as it says: use a commit name
17:44 thiago sangy: git merge merges histories
17:44 osse _ikke_: refs/stash
17:44 shurnormal er, graph!
17:44 sangy thiago: but I just want to compute the resulting blob, using the git's diff algorithm
17:44 _ikke_ osse: TIL
17:44 osse :O
17:44 thiago sangy: the diff algorithm has nothing to do with this
17:45 chrisshattuck joined #git
17:45 thiago sangy: like I said, the only way to reach the merging algorithm is to merge two branches.
17:45 thiago sangy: this sounds like !xy
17:45 gitinfo sangy: Woah, slow down for a bit. Are you sure that you need to jump through that particular hoop to achieve your goal? We suspect you don't, so why don't you back up a bit and tell us about the overall objective...
17:45 sangy thiago: oh, ok. That's wha I was worried about
17:45 sangy thiago: it's really not an xy, I do want to compute the blob. It's for a research project
17:45 AnarchyAo joined #git
17:45 thiago sangy: why do you want to compute the blob?
17:46 sangy thiago: We're trying to see if there could be ways to optimize merge of highly related histories
17:47 thiago sangy: then you don't want what you asked. You want to merge histories.
17:47 sangy e.g., 1M files are common but 1 differs. We want to compute the merge tree over the network. Ideally, we would only need the two differing blob objects to get the resulting tree
17:47 thiago sangy: that's what git merge does
17:47 rgrinberg joined #git
17:47 thiago merging is a local operation
17:47 thiago you already have all the objects.
17:47 sangy thiago: in the current usecase, yes
17:48 kfunk joined #git
17:48 sangy thiago: this is a research question.
17:48 thiago sangy: I don't see how it would change
17:48 thiago sangy: if the remote does the merge for you, then it has computed the whole history
17:48 thiago sangy: the 3+directory depth objects are all that would need to be sent when you fetched again
17:49 dreiss joined #git
17:49 sangy thiago: I agree.
17:50 sangy thiago: and that's the deal, could I have a trusted party to compute the merge with the minimal set of objects?
17:50 sangy thiago: to merge this case, the minimal set of objects is only those two blobs
17:51 thiago 3 + directory depth count
17:51 thiago it's never less than 3 objects
17:51 thiago well, it could be 2, but only if the merge result is identical to an existing tree state
17:51 thiago but no, then it would be 1
17:52 thiago I don't think 2 is a possible case
17:52 thiago anyway, I don't see what you want that isn't already there
17:52 thiago the perty that performs the merge computation has access to all objects and the entire history
17:53 thiago the result is a new commit, with a new tree and the resulting, merged blob objects
17:53 thiago git fetch performs the graph slicing and sends you only the new objects
17:55 multi_io joined #git
17:55 sangy thiago: wouldn't you just need blob and blob' (if it matches an existing tree state)?
17:55 thiago no
17:55 thiago you need the commit object, the tree object(s)
17:55 thiago the minimum transfer is 1 object and that's a commit object
17:56 thiago or a tag object, if you tagged an existing commit. But never is it anything below commit.
17:56 _ikke_ But that would be an 'empty' commit, right?
17:56 sangy thiago: you're assuming the default pack protocol. I'm telling you it's a somewhat theoretical/research question. I wanted
17:56 thiago no
17:56 sangy _ikke_: it could be a commit that matches a tree that you already ahve
17:56 thiago a merge commit can result in an existing tree. Example: git merge -s ours.
17:57 badloop hmmm...
17:57 inflames joined #git
17:57 badloop git index perms aren't getting updated as my filesystem perms change...
17:58 badloop even though i have the config var set
17:58 chrisshattuck joined #git
17:58 _ikke_ badloop: what perms?
17:58 _ikke_ git only tracks +x
17:58 badloop _ikke_: +x
17:58 thiago sangy: well, the transfer protocol is secondary to your question
17:58 badloop i'm wondering if eclipse has anything to do with it
17:58 thiago sangy: the point is that, in order to perform the merge in the first place, one must have access to all the history and all objects
17:58 thiago sangy: so that's already implemented in Git
17:58 morenoh149 joined #git
17:59 thiago sangy: now, if you want to improve the transfer protocol, sure. But I don't understand what's wrong with the current one.
17:59 badloop it creates the file originally as 644, but then when i save it, it changes the mode to 755
17:59 badloop but that change isn't getting communicated to git
17:59 curvv joined #git
17:59 thiago sangy: if you merged two trees with 1M files and only 1 file got changed in that (let's say it's in the root dir), then the transfer would be 3 objects: commit, root tree, blob object
18:00 sangy thiago: i could merge without creating a commit (yet). And I want to compute this new blob to create the new tree
18:00 thiago sangy: that
18:00 thiago sangy: that's the same as committing
18:00 thiago I don't see the difference
18:01 sangy thiago: ideally, no difference :). I just don't want to have the whole history for both sides. Are you telling me that I need all unrelated blobs to compute the new blob?
18:01 ahammond left #git
18:01 thiago sangy: I'm saying you need to split your research
18:01 badloop ugh... it looks like its a windows issue.  *sigh*
18:01 thiago sangy: one side is merging. That requires the entire history.
18:02 _ikke_ sangy: you need 3 blobs per file
18:02 badloop (windows eclipse that is)
18:02 thiago sangy: the other is the transfer. Currently, Git transfers entire histories. It sounds like you don't want to get the history of the other side of the merge.
18:02 thiago sangy: Git doesn't support that because it would break git log
18:02 sangy thiago: ok, If I want to compute only the novel blob from two blobs, is there a plumbing command for it?
18:03 thiago sangy: no, because merging requires at least three.
18:03 sangy thiago: I understand the normal case, and it's a way more robust way of doing things. I agree with that
18:03 sssilver joined #git
18:03 thiago sangy: in fact, it requires the full history
18:03 _ikke_ sangy: what do you want to achieve with your alternate scheme?
18:03 LionsMane joined #git
18:04 thiago sangy: a diff of two files doesn't contain enough information to determine which of the two sides to choose
18:04 _ikke_ You need the history to find the blobs
18:04 thiago you need at least 3 to do a 3-way merge
18:04 thiago it's better with history, though
18:04 _ikke_ technically, one or more of the blobs can be missing
18:04 _ikke_ (newly added file, removed file)
18:04 thiago that's the same as the empty blob
18:05 _ikke_ yeah
18:05 sangy thiago: ok so the history is only required for priorities? What happens if I do --allow-unrelated histories. It will prioritize the base of the merge right?
18:06 thiago then everything that isn't equal is a conflict
18:06 thiago and needs to be solved manually
18:08 sangy thiago: I see. I guess I'll have to go read some more on git's internals
18:08 sangy thiago: thanks!
18:09 zorph2 joined #git
18:10 romankl joined #git
18:10 hoe` joined #git
18:10 Corosus joined #git
18:11 zorph2 how would I duplicate develop - on master..without a conflict merge...?
18:11 hoe` I'm using "Ubuntu on Windows" and when I access a git directory it runs git-lfs for ages. When it completes, if I access it using git on windows, the same thing happens. I'm guessing there's a file that has a line ending difference that matters for LFS. Any guesses?
18:12 romankl joined #git
18:12 nettoweb joined #git
18:12 _ikke_ zorph2: If there are merge conflicts, there are merge conflicts, nothing you can do about it
18:12 _ikke_ except solving them
18:13 kyan joined #git
18:13 cyphase joined #git
18:14 zorph2 okee
18:19 Balliad joined #git
18:19 settermjd joined #git
18:20 bittin joined #git
18:20 bluezone joined #git
18:20 mozzarella joined #git
18:21 shinnya joined #git
18:21 sevenfourk joined #git
18:21 t4nkd joined #git
18:21 vrishab joined #git
18:22 vrishab couple of queries.
18:23 osse couple of answers.
18:23 vrishab how to know which url a repo is cloned from ?
18:23 vrishab :)
18:23 osse vrishab: git remote -v
18:23 osse it's what it says next to origin
18:24 vrishab works!
18:24 SourTrou_ joined #git
18:24 vrishab q2. I always want to check what is the difference between origin and my local clone ( svn stat -u )
18:24 vrishab before pulling in changes.
18:25 ToxicFrog Do you want a diff or a log?
18:25 vrishab 1. list of files changed. 2. diff
18:26 ToxicFrog Files changed: git diff --stat origin
18:26 ToxicFrog Diff: leave off the --stat
18:26 ToxicFrog Log: git log origin..
18:26 chris2 and fetch first ;)
18:26 ToxicFrog (which will show all commits in the current branch but not in the corresponding branch in origin)
18:26 ToxicFrog Note that you need to "git fetch" before any of these so it has an up-to-date view of the state of origin; unlike svn it won't automatically fetch that info from the server when you want a dif.
18:27 ToxicFrog There are also graphical tools you can use to examine history, like gitk, which can be useful for more complicated tasks.
18:29 vrishab checking them. just a sec.
18:30 Darren_ joined #git
18:30 ekinmur joined #git
18:32 bittin joined #git
18:33 tvw joined #git
18:33 vrishab ToxicFrog, diff / File changed work, log doesn't.
18:33 Jellyg00se joined #git
18:33 vrishab git log origin.. gives empty output
18:36 ahmed_elgabri joined #git
18:37 govg joined #git
18:38 ToxicFrog vrishab: huh. works here. git log origin/$BRANCHNAME..$BRANCHNAME ?
18:39 aard_ joined #git
18:42 vrishab ToxicFrog, http://paste.debian.net/794875/
18:42 tvw joined #git
18:42 Ph_ joined #git
18:42 Spec-Chum joined #git
18:44 romerocesar joined #git
18:44 emPi joined #git
18:45 drawkula joined #git
18:46 cbreak joined #git
18:46 Corosus joined #git
18:47 boshhead joined #git
18:47 ToxicFrog vrishab: what's the output of 'git branch -a'?
18:48 Dougie187 joined #git
18:50 piling joined #git
18:50 AaronMT joined #git
18:51 teroshan joined #git
18:51 AaronMT joined #git
18:51 vrishab ToxicFrog, http://paste.debian.net/794877/
18:52 boshhead joined #git
18:53 SwiftMatt joined #git
18:54 teroshan joined #git
18:56 mingrammer joined #git
18:56 AaronMT joined #git
18:56 IanV0rn2341 joined #git
18:57 vassagus joined #git
18:57 ToxicFrog vrishab: that's why diffing origin/gnome-maps against gnome-maps didn't work; there's no such branch.
18:57 ToxicFrog You wanted "git diff origin/master..master"
18:57 ToxicFrog gnome-maps appears to be the name of the repo, not the name of any branch within it.
18:58 ToxicFrog Er
18:58 ToxicFrog git log origin/master..master, rather
19:00 vrishab ToxicFrog, git log master..origin/master, rather
19:00 sentriz joined #git
19:01 vrishab that worked. http://paste.debian.net/794881/
19:01 ToxicFrog vrishab: oh, right, I got the ordering you wanted backwards. Sorry.
19:02 ToxicFrog I'm so used to asking "what have I done that isn't on origin" I reflexively answered that.
19:02 ToxicFrog "git log ..origin" will probably also work.
19:02 vrishab great. thanks much !
19:04 tlaxkit joined #git
19:04 bsanford_ joined #git
19:07 star_prone joined #git
19:07 kpease joined #git
19:07 krs93 joined #git
19:08 settermjd joined #git
19:09 danslo_ joined #git
19:09 MattMaker joined #git
19:09 Macaveli joined #git
19:09 SourTrout joined #git
19:10 afuentes joined #git
19:12 Pulp joined #git
19:13 Jon28 joined #git
19:14 Macaveli joined #git
19:15 sythe joined #git
19:17 sandstrom joined #git
19:21 tomasm- joined #git
19:23 tomasm- hey, so is it possible  when multiple  people are working on the  same repository that one could accidentally erase all history of some changes in a file? even from the past? I'm looking at some code where I know I did changes and i just cant find anything in the history showing what I did. oddly, I'm looking at code i committed a month ago and what should have been there is wrong/incomplete.
19:23 davis joined #git
19:23 davis hello
19:24 _ikke_ tomasm-: It's possible for someone to push an alternate history, yes
19:25 tomasm- so if  I didn't notice it and all the  code was pushed to all the locations, am i pretty much screwed?
19:25 davis i have checkedout a specific revision and it says detached. I would like to start a new branch. can I now do git branch foo, git checkout foo, and now all my work will be with a new branch?
19:26 chabil joined #git
19:26 tomasm- the odd thing is thing is it wasnt noticed until just now, so they had code for several weeks that had the  changes i made.
19:28 vassagus joined #git
19:28 _ikke_ tomasm-: if it's less then 90 days ago, you should still have it locally
19:30 SourTrou_ joined #git
19:30 tomasm- _ikke_, why 90 days? how  do I find it? I'm using gitk to visualize the branch/history and its not there. is there a command line version to browse all changes touching a file?
19:30 rgrinberg joined #git
19:31 _ikke_ tomasm-: your local reflog keeps entries for 90 days
19:31 _ikke_ man git reflog
19:31 gitinfo the git-reflog manpage is available at http://jk.gs/git-reflog.html
19:31 shurnormal tomasm-: Chances are some members of the team anticipated you could screw up and didn't merge all changes fetched from you without review? Duplicity backups or AWS snapshots of the repositories..?
19:32 al-damiri joined #git
19:32 King_Hual joined #git
19:32 ToBeCloud joined #git
19:32 tomasm- shurnormal, well the problem of having some backup somewhere is that unless I'm reviewing the others' changes (which i stupidly didnt do), the backup could be affected as well
19:32 shurnormal gitk --all --date-order FILE but I guess it wont suit files not existing in the rev checked out.
19:33 stamina joined #git
19:33 tomasm- the file is there, the lines modified within it not
19:34 shurnormal If the file is there and there is any ref including the changes, then it will hint. If not, I think in Stack Overflow you can find a command returning all revs changing the file.
19:34 shurnormal Then you can pass the revs list to a git log or something
19:35 tomasm- gitk doesnt show those specific changes. the specific commit is there, but the lines changed are wrong.
19:36 shurnormal The commit is not wrong, the hashing function protects it.
19:36 shurnormal You committed wrongly, or you are searching for a commit orphan from any ref.
19:37 _ikke_ tomasm-: you need the reflog
19:37 shurnormal The first has solutions, and the second too :)
19:37 LionsMane joined #git
19:38 _ikke_ check the reflog if you find your commits
19:38 _ikke_ use either git reflog, or git log -g
19:38 tomasm- hmmm, ok well now that i think about this i think it was  a wild goose chase
19:38 tomasm- i just woke up from a nap =(
19:39 tomasm- crisis averted
19:40 tomasm- QA team said "I noticed page X stopped working over the weekend".... and im thinking  it never worked and is an entirely different issue and they never tested it in the  first place
19:40 leeN joined #git
19:40 nonconvergent joined #git
19:41 nonconvergent Trying to diagnose a problem that happened after I used git-flow-avh to merge a hotfix.
19:41 nonconvergent I think it merged master to develop. How would I verify that?
19:43 cagedwisdom joined #git
19:43 settermjd joined #git
19:43 nonconvergent Also, it shows the hotfix branch as deleted on origin when I browse it on our github, and when I try to checkout origin/hotfix/myHotfix it says the pathspec doesn't match anything, but my devops guy says he can pull it from a fresh clone.
19:44 tomasm- thanks guys, i appreciate it
19:44 nonconvergent What fresh hell have I created?
19:44 _ikke_ nonconvergent: merging master in develop is not bad
19:47 shurnormal I wish you actually called it myHotfix. Check it with gitk, gitg, sourcetree...
19:48 shurnormal tig --all, git log --all --graph...
19:49 rgrinberg joined #git
19:49 eijk joined #git
19:50 newbiee joined #git
19:50 King_Hual joined #git
19:50 mingrammer joined #git
19:51 shurnormal git log --all --graph --remotes --decorate=full maybe
19:51 newbiee we have a server with cloned git. Now we are getting one more version of the uploaded code from one more server and want to upload this receievd code. How can we do git rebase and upload again?
19:52 Wysi joined #git
19:53 nettoweb_ joined #git
19:54 shurnormal You can check out master from the patched outdated, push it as outdated-master, remove master, then pull master from upstream, push to server one as master, checkout both master and outdated-master, and merge as like?
19:54 eijk_ joined #git
19:55 _ng joined #git
19:56 emPi joined #git
19:57 newbiee shurnormal: can i do this? I download both the codes and do some kind of rebase?
19:57 dave0x6d joined #git
19:57 newbiee basically doing rebase on a remote master
19:58 shurnormal If you can discard server one contents you can maybe do the rename of serverone/master to serverone/old-master and force push of servertwo/master to serverone/master?
19:58 hahuang65 joined #git
19:58 iota_pi joined #git
19:59 cdg joined #git
19:59 shurnormal I understand rebase and rebase interactive but don't know what you mean by rebase on remote.
19:59 alansaul joined #git
19:59 shurnormal Rebase is local as is merge.
20:00 rpd I think they mean rebasing a local branch off of a remote master
20:01 Jellyg00se joined #git
20:02 rpd anyways you can probably do something like git pull --rebase <remote> master and push that to your server
20:02 raijin joined #git
20:03 dsantiago joined #git
20:03 newbiee shurnormal: so what i mean is this. I have a server_1 where we have pushed the code which we have received from one or our vendors. We do our own development and merge with the received code and push it to server_1. This goes on. After some time our vendor release new code and we basically need to push it again to our server_1. However we want to keep our history of both the old + new code.
20:04 shurnormal I think You can do the first tip I told.
20:04 alansaul joined #git
20:04 nikivi joined #git
20:05 shurnormal Copying master from server 1 to itself as old-master, then you can move new master to server 1 while keeping the old master as a branch.
20:05 nonconvergent shurnormal: is there a way to list files updated by a commit in that log graph?
20:06 nonconvergent without doing it manually
20:06 shurnormal for sure
20:06 shurnormal I learned it today. Like...
20:07 newbiee shurnormal: so basically you are saying is this: change the master branch to the latest code from vendor and keep the old master as a branch. Am I right in my understanding?
20:07 jorj joined #git
20:07 rpd nonconvergent: --name-only will show the files affected by a commit; not sure how that combines with --graph though
20:08 nonconvergent rpd: works
20:08 shurnormal nonconvergent: git log --all --graph --remotes --decorate=full --stat
20:08 nonconvergent shurnormal: what's remotes do?
20:09 shurnormal newbiee: yes, but maybe you want better to hold master and push new code as version2 or something for not breaking master.
20:09 mkoskar joined #git
20:10 aard_ joined #git
20:10 cagedwisdom left #git
20:10 shurnormal nonconvergent: I'm not sure, but you got a lot of information! You won't miss your rogue commits!
20:11 nonconvergent Does ref/heads/master mean it was applied to master? "commit a78785ed8d7c423275fdfadbb3ed69330900bef3 (tag: refs/tags/v3.0.1, refs/heads/master)"
20:11 drizztbsd joined #git
20:11 shurnormal ref/heads/master IINM is your master
20:12 ResidentBiscuit joined #git
20:12 shurnormal As refs/heads/origin/master is what you know of origin's master from the last time you fetched from it.
20:12 zumba_addict joined #git
20:13 zumba_addict hey folks, I accidentally moved a directory within an ide and now git status is showing it. I'm not sure where it's from. How can I undo? The editor cannot undo it
20:13 Sample joined #git
20:13 shurnormal You can git checkout -- FILES.
20:13 Sample joined #git
20:14 shurnormal And/or git reset HEAD FILES.
20:14 zumba_addict ah i missed --
20:14 nettoweb joined #git
20:14 zumba_addict i only typed git checkout directory
20:15 zumba_addict i have other changes that i don't want to undo
20:16 Macaveli joined #git
20:17 zumba_addict looks like it doesn't work if i supply directory
20:17 cjohnson works fine here
20:17 cjohnson git checkout -- .  will recursively restore all files in .
20:17 cjohnson are you supplying the right path?
20:18 zumba_addict i figured out where it came from
20:18 zumba_addict i can just rm -rf it since it's in our ,gitignore
20:18 zumba_addict yes, I supplied the right path
20:18 zumba_addict let me try it again
20:19 chabil joined #git
20:19 cdg joined #git
20:19 shurnormal If you moved a tracked into some path that is ignored, I guess it appears in git status as "removed", then git checkout -- DIR should do.
20:19 zumba_addict so I guess, i am missing the something in my path, i only typed git checkout -- dirname maybe it should be git checkout -- dirname/* ?
20:20 zumba_addict it came from a directory that is in our .gitignore then I accidentally moved it to the root which is tracked
20:20 shurnormal Idk I never needed any star in git checkout.
20:20 garethdaine joined #git
20:20 stamina joined #git
20:21 shurnormal Ah then in status it shows in the added region, so git reset HEAD DIR should do.
20:21 alansaul joined #git
20:21 johnmilton joined #git
20:21 zumba_addict i tried that too and it failed
20:21 zumba_addict let me grab the error
20:21 King_Hual joined #git
20:21 King_Hual joined #git
20:22 zumba_addict no error but looks like git won't move it to where it was previously. It's in untracked now when I run git status
20:23 shurnormal OK! Now you move it back with mv or the IDE and you are done.
20:23 zumba_addict maybe because it's not being tracked and it's in .gitignore?
20:23 zumba_addict got i
20:23 zumba_addict cool
20:23 zumba_addict so looks like it's only applicable to files that are being tracked
20:24 hahuang65 joined #git
20:24 cjohnson correct, git can't do things to files it doesn't know about
20:24 kaldoran joined #git
20:24 satifant joined #git
20:25 MrHall joined #git
20:25 johnmilton joined #git
20:26 courrier joined #git
20:27 eijk_ joined #git
20:28 sangy joined #git
20:30 fission6 joined #git
20:32 MattMaker joined #git
20:34 MrHall Hello
20:34 gitinfo MrHall: 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.
20:35 MrHall i need help to connect eclipse with gitlab
20:37 mkoskar joined #git
20:40 omegak joined #git
20:41 hoe` The issue I came here to talk about was when git on Ubuntu on Windows accesses a repo it rebuilds the index, and then when git on Windows accesses the directory it needs to also do this.
20:41 hoe` It becomes an issue because of LFS not properly using cache.
20:42 hoe` After parsing out the structs in the index file, I noticed windows is setting the file modes to 0's, and Ubuntu on Windows is giving them real values, and that the file times in the structs were different.
20:42 hoe` "git config --global core.filemode false" did not fix my issue,so I'm guessing it's the file times?
20:42 kadoban hoe`: You should probably have separate repos depeding on what OS you're on, I would assume. They can be linked by using a "central" bare repo you push to and pull from or whatever.
20:43 kadoban Not sure that's correct, but it's what I'd do. I suspect you'll lose your mind trying to make it work in both without weird things happening.
20:43 hoe` The workflow I was hoping for was CLI access in Ubuntu for Windows, and "git gui" access from normal Windows (where I like to review my commits)
20:44 hoe` There's a lot of workarounds but none of them are as convenient as just using git.exe for everything.
20:44 hoe` was hoping someone else here had done my homework for me though :)
20:45 cjohnson It would have to be a huge amount of homework to make the two systems play with the same set of meta data
20:46 Iw1ll1nj3ct joined #git
20:46 hoe` I'm guessing the time stamps not being identical is still an issue?
20:47 crayon joined #git
20:47 hoe` ohh, I missed core.trustctime
20:49 hoe` nope :)
20:52 jeffreylevesque joined #git
20:53 Iw1ll1nj3ct thanks cjohnson
20:54 zumba_addict thanks cjohnson
20:54 zacts joined #git
20:54 kulelu88 joined #git
20:56 romerocesar joined #git
21:01 Emperor_Earth joined #git
21:01 zumba_addict joined #git
21:01 aidalgol joined #git
21:04 newbiee joined #git
21:05 AnarchyAo joined #git
21:06 nurey joined #git
21:06 t4knd joined #git
21:07 hoe` is there a way to enable git verbose logging that explains why it's updating the index?
21:07 eijk joined #git
21:08 newbiee_ joined #git
21:08 DieguezZ joined #git
21:09 kadoban GIT_TRACE=1 git whatever     possibly. I have no idea how verbose that gets though, I've rarely used it.
21:09 _ikke_ It shows all processes git executes
21:10 kadoban Oh is that all it does? Nevermind then probably xD
21:10 hoe` that does still help me though.
21:11 hoe` GIT_TRACE=1 git -C /mnt/c/ml2 update-index --refresh --verbose
21:11 hoe` that --verbose does not help me, but the git_trace is showing that it's the git-lfs clean calls that are a bit of an issue.
21:11 hoe` guess I'll wait for this https://github.com/github/git-lfs/issues/1329
21:15 achlys joined #git
21:16 clemf joined #git
21:19 zumba_addict joined #git
21:20 _ikke_ Yeah, I think the patch is cooking currently
21:21 zacts_pi joined #git
21:21 HoierM joined #git
21:23 nopevms joined #git
21:24 furrykef joined #git
21:27 aartist joined #git
21:33 zacts joined #git
21:35 andrewray joined #git
21:35 moumitude joined #git
21:35 andrewray I want to stay on the commit i'm on, and not modify the history, but update all the files to a specific commit
21:36 andrewray I thought reset --soft would do it but that appears to take history with it
21:36 _ikke_ "git checkout <commit> ."
21:37 cdg joined #git
21:38 vassagus joined #git
21:38 rvaldes_ joined #git
21:39 mingrammer joined #git
21:39 eijk_ joined #git
21:43 i7c Wouldn’t bet my life on it, but I think reset always changes HEAD (unless you reset to HEAD).
21:43 i7c At least that’s how I tried to remember it.
21:45 andrewray this is what I was looking for, I was just on the wrong branch when I did it http://stackoverflow.com/a/17354830/743464
21:46 Dougie187 left #git
21:48 i7c Uhm yeah that’s keeping the working dir/index while moving to another commit, not staying on one commit and changing the working dir/index.
21:49 andrewray yeah convoluted question was my fault
21:49 i7c Or do you mean the second answer?
21:49 andrewray my mental model was in the wrong place
21:49 i7c Yeah well, that happens. :)
21:50 lvns joined #git
21:52 lvns_ joined #git
21:52 javajoe joined #git
21:53 justanotheruser joined #git
21:54 ochorocho__ joined #git
21:54 lvns joined #git
21:54 justanotheruser joined #git
21:55 lvns_ joined #git
21:55 ahr3n joined #git
21:57 vrishab joined #git
21:57 crayon joined #git
21:58 radsy joined #git
21:59 nullp0inter joined #git
21:59 cjohnson _ikke_: can you explain why it's  git checkout -- path  ?
21:59 cjohnson why -- ?
21:59 cjohnson -- = HEAD ?
21:59 nullp0inter i am using a composer dependancy that pulls from a remote git project..when i try to add the cir manually to my repo i get "fatal: Pathspec 'src/vendor/spacenate/surveygizmo-api-php/composer.json' is in submodule 'src/vendor/spacenate/surveygizmo-api-php'" what does that mean?
22:02 eijk joined #git
22:03 eijk_ joined #git
22:04 blackwind_123 joined #git
22:05 beyertech joined #git
22:06 newbiee joined #git
22:06 nullp0inter and whats the best way to keep this dir as is and track it in my main repo
22:08 hoe` left #git
22:12 osse !submodule > nullp0inter
22:12 gitinfo git-submodule is ideal to add subsidiary git repositories to a git superproject when you do not control the subprojects or more specifically wish to fix the subproject at a specific revision even as the subproject changes upstream. See http://www.git-scm.com/book/en/Git-Tools-Submodules
22:14 nullp0inter osse, its a dependency managed by composer so trying to figure out what to do there
22:14 ToxicFrog cjohnson: the actual format is "git checkout [options] [commit] [-- paths]; the -- is there to differentiate paths from commits.
22:15 ToxicFrog So that it can tell the difference between "switch to the branch named master" and "check out, from HEAD, the file named master"
22:16 furrykef left #git
22:17 osse nullp0inter: i've no idea. just trying to explain what a submodile is
22:17 osse maybe that's the best way
22:17 cyphase joined #git
22:17 nullp0inter ok ty
22:19 andlabs joined #git
22:19 romerocesar joined #git
22:20 osse nullp0inter: if it's in a submodule you either have to deal with that, or nuke the .git dir within
22:21 nettoweb joined #git
22:25 nullp0inter osse, when i did a mv on the .git dir, everything disappeared
22:25 HoierM joined #git
22:25 nullp0inter in the source dir
22:25 vassagus joined #git
22:26 nwmcsween joined #git
22:27 zartu joined #git
22:28 nwmcsween has anyone made a script that rewrites history based on a $cur_file $new_location pair? e.g script dir/file.c dir2/renamed.c dir/another.c dir3/renamed2.c
22:29 rwb joined #git
22:29 eijk_ joined #git
22:29 nwmcsween or easier is there someway to get GIT_INDEX_FILE via a cmd?
22:29 osse nwmcsween: I mean src/vendor/spacenate/surveygizmo-api-php/.git
22:29 osse ehh, that was for nullp0inter
22:29 wagle joined #git
22:31 eijk_ joined #git
22:32 refried_ joined #git
22:32 boombatower joined #git
22:33 mingrammer joined #git
22:33 eijk_ joined #git
22:34 cjohnson ToxicFrog: oh! that makes more sense. I've never passed multiple paths
22:34 cjohnson just called checkouta bunch of times or found a common ancesotr
22:35 tjbp joined #git
22:35 cjohnson ToxicFrog: so why does "git checkout <commit> ."  work
22:36 MatthewAllan93 joined #git
22:36 cjohnson Oh I guess git checkout master somefile  isn't ambiguous
22:36 cjohnson weird API choice
22:36 spudowiar joined #git
22:38 thiago joined #git
22:41 eijk joined #git
22:41 tjone270 joined #git
22:41 eijk_ joined #git
22:43 eijk_ joined #git
22:44 Raed|Mobile joined #git
22:45 star_prone joined #git
22:47 MatthewAllan93 joined #git
22:47 preyalone joined #git
22:49 tjbp joined #git
22:53 inflames joined #git
22:56 jimi_ joined #git
22:58 MatthewAllan93 joined #git
22:58 tjbp joined #git
22:58 NaN joined #git
22:59 NaN is it possible to do a "cherry pick" from a stash?
23:00 durham_ joined #git
23:00 GyrosGeier yes
23:01 NaN great, I will RTFM then :D
23:01 GyrosGeier but it is annoying to do so
23:01 NaN pfff
23:02 GyrosGeier I just took a few minutes to look up how
23:02 NaN So what do you recommend?
23:02 GyrosGeier check out some state that the stash applies to
23:02 GyrosGeier apply the stash
23:02 GyrosGeier create a proper commit
23:02 GyrosGeier go back to the original branch
23:02 GyrosGeier cherry-pick the commit
23:02 NaN seems a lot of work
23:03 NaN any interactive stash apply?
23:03 durham joined #git
23:03 GyrosGeier gitg has something like that
23:04 GyrosGeier stashes should normally not live that long
23:04 GyrosGeier these are just for small interruptions
23:04 GyrosGeier anything that takes longer should live in a separate branch
23:04 Gsham joined #git
23:04 NaN hehe you're right, but my interruptioin was 2 weeks ago, so now I only need some of those lines not all the stash
23:05 tjbp joined #git
23:05 NaN technically a stash is a separate branch :B
23:05 NaN but I get the point
23:06 ShekharReddy joined #git
23:06 hfp joined #git
23:07 GyrosGeier gitg shows you which stashes exist
23:09 jstein_ joined #git
23:09 MatthewAllan93 joined #git
23:09 jnewt joined #git
23:10 eijk joined #git
23:10 jnewt how do you rebase a whole tree instead of just a branch? (ie i want to just specify the "oldbase" and "newbase", and move every commit and branch pointer and tag.
23:11 holgersson joined #git
23:12 sgen joined #git
23:13 vassagus joined #git
23:14 tgunr joined #git
23:16 nurey joined #git
23:16 GyrosGeier jnewt,
23:16 GyrosGeier jnewt, "for i in refs/branches/*; do ..."
23:18 jnewt GyrosGeier: not every branch in the whole repo.  just the ones with some commit in their history
23:20 jstein__ joined #git
23:21 rkazak joined #git
23:21 bernardio joined #git
23:22 MaddHatter joined #git
23:24 diphtherial joined #git
23:24 MaddHatter Is there some way to tell checkout-index "ignore errors. These are not the errors you're looking for. *jedi hand wave*"? It's correctly failing when it can't create a directory, but I want it to just ignore the uncreated directory and keep going with the rest of the checkout. Instead it gives fatal error and dies.
23:24 diphtherial left #git
23:25 evanwang joined #git
23:27 GyrosGeier jnewt, for i in refs/branches/*; do if git rev-list $i | grep -q ^abcdef12345; then ...
23:27 DieguezZ joined #git
23:27 GyrosGeier you can just list the branches with
23:28 GyrosGeier for i in refs/branches/*; do if git rev-list $i | grep -q ^abcdef12345; then echo $i; fi; done
23:29 parallel21 joined #git
23:29 tyreld joined #git
23:29 aidalgol joined #git
23:29 GyrosGeier erm
23:29 GyrosGeier you need to be in .git
23:30 GyrosGeier and use refs/heads, not refs/branches
23:30 rgrinberg joined #git
23:30 parallel21 I just removed a bunch of large files from my git repo and reduced the repo from 500mb down to 90mb. I've pushed the changes up to my remote git, but when I clone from the remote, I'm not seeing a reduced file size
23:30 GyrosGeier these files are still in the history
23:31 parallel21 Looks like, but not locally
23:31 GyrosGeier you need to rewrite history if you want them gone permanently
23:31 Raed|Laptop joined #git
23:31 dave0x6d joined #git
23:31 parallel21 git commit hashes to tags updated to new ones
23:31 GyrosGeier also, all branches?
23:32 parallel21 And hashes show changes have carried in the remote and when I clone back to my local. Just the newly created clone is still just as fat as it was
23:32 GyrosGeier and even if you caught everything, the reflog will still keep them around for two weeks
23:32 GyrosGeier ah
23:32 parallel21 :(
23:32 GyrosGeier new clone should not have them
23:33 GyrosGeier because the reflog is local
23:33 parallel21 new clone should not have updated hashes?
23:33 GyrosGeier no, it shouldn't have reflog entries
23:33 GyrosGeier so the objects should be really gone
23:33 parallel21 I'm lost at why they're not getting purge on the remote repo
23:34 GyrosGeier did you clone through git protocol (ssh or not), or via http?
23:34 GyrosGeier http uses prebuilt packfiles
23:34 parallel21 git://
23:34 GyrosGeier hmmk
23:34 GyrosGeier the remote repo will keep them for another two week
23:34 GyrosGeier s
23:34 GyrosGeier but should not hand them out anymore
23:35 rbern joined #git
23:35 tgunr joined #git
23:36 durham joined #git
23:36 parallel21 What causes the remote repo to hold onto files for two weeks?
23:37 parallel21 Is that git-gc --auto defaults?
23:37 GyrosGeier that is the reflog
23:37 GyrosGeier the reflog keeps a log of where every branch pointed to, in time
23:37 durham joined #git
23:37 shruggar joined #git
23:38 GyrosGeier this still counts as a reference, so gc will not delete the objects yet
23:38 GyrosGeier in case you decide you want them back
23:38 GyrosGeier after two weeks, the reflog cycles, and the ref is deleted
23:38 GyrosGeier git gc --aggressive leaves you with a single pack, and loose objects for everything that is unreferenced
23:39 GyrosGeier if the large files end up in the single object files, then everything is fine
23:40 GyrosGeier if they end up in the packs, there are still references left
23:41 parallel21 but why would the local copy not have them after `git filter` , reflog, prune etc. But the remote still hangs onto them?
23:41 Atrumx joined #git
23:42 GyrosGeier you updated the remote branch
23:43 GyrosGeier which changes what is referenced from the current branch tips
23:43 parallel21 I think I see... I didn't checkout remote branches locally when doing the git-filter?
23:43 parallel21 might that be it?
23:43 GyrosGeier no
23:43 raijin joined #git
23:43 parallel21 hrmm...
23:43 tax joined #git
23:43 nurey joined #git
23:43 GyrosGeier as long as you did overwrite all the remote branches with filtered ones, the remote has forgotten about the old state
23:44 GyrosGeier except for the reflog
23:44 GyrosGeier which will keep a record that you pushed new versions for all the branches, and what they were before
23:44 GyrosGeier but the reflog forgets after two weeks
23:44 GyrosGeier so when you run gc in two weeks, the objects should go away
23:45 parallel21 You mean, when the remote runs gc?
23:45 GyrosGeier yes
23:45 parallel21 and no way to trigger manually?
23:45 GyrosGeier and I think before that, the gc should already isolate the objects from the others
23:45 parallel21 I've read some places like bitbucket, requiring a ticket to be put in
23:45 GyrosGeier sure, you can clean the reflog
23:45 GyrosGeier but that is generally a last-resort thing
23:46 madewokherd joined #git
23:46 GyrosGeier because this is what protects you from stupidity
23:46 parallel21 thanks for helping build a better understanding for the problems
23:46 belak Yeah, normally a gc on bitbucket's end will happen automatically, but when you go over the limit, I think we have to run it manually.
23:46 cqi joined #git
23:46 GyrosGeier when you find out that your filter command was broken, and you changed all the line endings in the process, the reflog gives you a chance to go back
23:47 parallel21 belak, you mean like size limit? This repo was sitting at close to 600mb at it's largest
23:47 GyrosGeier or someone accidentally pushed to master
23:47 parallel21 heh
23:47 GyrosGeier that happened to me in a customer project once
23:47 GyrosGeier the reflog saved the day
23:47 belak parallel21: yeah, this is also bitbucket specific...
23:47 parallel21 I see, so though i curse reflog today
23:47 belak The size limit is 2G as well
23:47 parallel21 That's what I thought... some fool put a bunch of mp4
23:48 d4rklit3 joined #git
23:48 parallel21 and then readded changes... and cloning updates has been a nightmare
23:49 GyrosGeier there should be an option "reject pushes that are descended from x"
23:49 GyrosGeier so after you rewrite history, people merging the old history back get rejected
23:50 parallel21 +1 for that feature
23:50 parallel21 git bad-blood
23:50 skylite joined #git
23:51 spudowiar1 joined #git
23:53 rbern joined #git
23:54 a10c joined #git
23:54 pijiu2 joined #git
23:56 ochorocho__ joined #git

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