Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2017-02-16

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

All times shown according to UTC.

Time Nick Message
00:01 Rodya_ joined #git
00:03 alexandre9099 joined #git
00:04 Rodya_ joined #git
00:04 a_thakur joined #git
00:06 CountryNerd joined #git
00:06 BenLubar left #git
00:09 madewokherd joined #git
00:10 flashpoint9 joined #git
00:11 peterbecich joined #git
00:13 Rodya_ joined #git
00:13 DolpheenDream joined #git
00:13 askb joined #git
00:15 elijah joined #git
00:17 howitdo joined #git
00:20 DIDAVISION joined #git
00:20 BackEndCoder joined #git
00:20 nevodka joined #git
00:24 ecuanaso joined #git
00:24 peterbecich joined #git
00:26 thebope joined #git
00:27 SirOliver joined #git
00:27 orzo joined #git
00:29 thebope joined #git
00:29 VladGh joined #git
00:30 snowalpaca joined #git
00:30 snowalpa_ joined #git
00:31 digidog joined #git
00:32 ToBeCloud joined #git
00:35 jstimm joined #git
00:37 ckb joined #git
00:39 orzo joined #git
00:46 matoro joined #git
00:47 Peetz0r joined #git
00:48 nettoweb joined #git
00:50 Lyqyd joined #git
00:51 Rodya_ joined #git
00:52 venmx joined #git
00:53 manuelschneid3r joined #git
00:53 thebope_ joined #git
00:54 Peetz0r joined #git
00:57 crose joined #git
00:58 Guest26 joined #git
00:59 thebope joined #git
01:01 cqi joined #git
01:02 thebope_ joined #git
01:03 gugah joined #git
01:07 qpdb joined #git
01:09 overlord_tm joined #git
01:10 dsantiago joined #git
01:11 arescorpio joined #git
01:13 roygbiv joined #git
01:13 nickabbey joined #git
01:15 zacts joined #git
01:20 danimal2 joined #git
01:20 inflames joined #git
01:23 DolpheenDream joined #git
01:24 prg3 joined #git
01:24 Rodya_ joined #git
01:25 ckb joined #git
01:28 Vampire0_ joined #git
01:28 snowalpaca joined #git
01:29 anuxivm joined #git
01:29 zzz joined #git
01:29 thebope joined #git
01:31 dave0x6d joined #git
01:33 acetakwas joined #git
01:36 jbitdrop joined #git
01:37 thebope joined #git
01:40 thebope joined #git
01:43 thebope joined #git
01:45 Gunni joined #git
01:48 jameser joined #git
01:49 DolpheenDream joined #git
01:52 Kulrak joined #git
01:52 Kulrak joined #git
01:56 Gsham joined #git
01:57 JamesBaxter joined #git
01:58 matoro joined #git
01:59 livingstn joined #git
02:03 livingstn joined #git
02:03 winfli joined #git
02:06 raijin joined #git
02:09 sbulage joined #git
02:10 nowhere_man joined #git
02:11 a_thakur joined #git
02:14 livingstn joined #git
02:14 snowalpa_ joined #git
02:16 Gunni joined #git
02:16 r0gerth4t joined #git
02:17 snowalpaca joined #git
02:18 Vampire0_ joined #git
02:20 freddymungo joined #git
02:21 NwS joined #git
02:22 r0gerth4t any good resources for git 'best practices' or how it's frequently used in big group environments? ive only used git on small projects but just joined a bigger company. interested in how branches are frequently used for features, etc
02:23 jstimm joined #git
02:27 xall joined #git
02:29 thebope joined #git
02:29 ResidentBiscuit joined #git
02:31 Derperperd joined #git
02:31 cqi joined #git
02:32 ecuanaso joined #git
02:32 suttonwilliamd joined #git
02:32 moritz r0gerth4t: there's a certain trend away from feature branches, and towards trunk-based development
02:32 moritz r0gerth4t: mostly because it plays nicer with CI/CD
02:33 nivag r0gerth4t, You could do worse than have a look at how some big open source projects use git!  e.g. https://www.libreoffice.org/about-us/source-code
02:34 moritz if you do have to support older versions of some software, you can use something like "git flow"
02:34 moritz (though it's quite some overhead, IME)
02:34 nivag moritz,  what is 'CI/CD'
02:34 variable joined #git
02:35 moritz nivag: Continous Integration / Continuous Delivery
02:35 mizu_no_oto joined #git
02:35 trout is there a way to say "find all the branches that modify a particular file" ?
02:36 trout that isn't a really well defined questio
02:36 revoltingPeasant joined #git
02:36 trout since I need to define "changed against what", but some thoughts
02:36 trout * I want some thoughts
02:36 nivag moritz, thanks
02:37 cdg joined #git
02:38 shinnya joined #git
02:38 Puffball joined #git
02:39 moritz trout: maybe git log --all --decorate -- path/to/file can help?
02:40 trout moritz good point; I forgot about --all
02:40 brent__ joined #git
02:41 trout I suspect the answer will involve a script and for-each-ref :\
02:41 basiclaser joined #git
02:48 anuxivm left #git
02:51 nevodka joined #git
02:51 d^sh joined #git
02:53 xall joined #git
02:53 thebope_ joined #git
02:56 jimklo_ joined #git
02:57 rkazak_ joined #git
02:57 livingstn joined #git
02:59 thebope joined #git
02:59 ratdeptrai joined #git
03:02 livingstn joined #git
03:02 thebope joined #git
03:02 ojdo joined #git
03:03 wxg joined #git
03:04 tusharmalik joined #git
03:05 lektrik joined #git
03:06 Gloomy joined #git
03:06 cdg joined #git
03:09 Forbidd3n joined #git
03:09 bruce_lee joined #git
03:10 thebope joined #git
03:10 Forbidd3n Hey everyone. I am using bitbucket and need to use public/private key. I am on osx, but want to store the key in the project directory on my local computer instead of the standard ~/.ssh directory
03:10 Forbidd3n How would I tell git to look at a specific directory for validation?
03:14 s3td3m0n joined #git
03:14 wxg joined #git
03:14 moritz Forbidd3n: you can set GIT_SSH_COMMAND='ssh -i path/to/your/private/key'
03:14 Forbidd3n moritz: do I do this in the git directory?
03:14 moritz or use your ssh config to specify the key file
03:15 moritz Forbidd3n: it's an environment variable that git respects
03:15 Forbidd3n ok I will add it to the config
03:15 moritz so you'd have to export it in the shell where you use git
03:15 inflames joined #git
03:15 syg_ joined #git
03:16 sarbs joined #git
03:17 Vortex34 joined #git
03:19 wxg joined #git
03:20 fstd_ joined #git
03:20 Derperperd joined #git
03:22 hexagoxel joined #git
03:22 systemovich joined #git
03:22 Derperperd joined #git
03:22 dendazen joined #git
03:25 Vampire0 joined #git
03:28 brent__ joined #git
03:30 thiago joined #git
03:31 thebope joined #git
03:36 stevenhong joined #git
03:37 thebinary joined #git
03:38 thebope joined #git
03:39 durham joined #git
03:43 lexi933 joined #git
03:43 adymitruk hi lexi933
03:43 MarioBranco joined #git
03:44 lb1c joined #git
03:44 lexi933 hi
03:44 gitinfo lexi933: 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.
03:46 thebinary joined #git
03:49 rossome joined #git
03:54 peterbecich joined #git
03:56 brent__ joined #git
03:57 peterbec` joined #git
03:57 xall joined #git
03:57 ffilozov joined #git
03:59 sunri5e joined #git
04:00 a_thakur joined #git
04:01 peterbecich joined #git
04:01 pur3eval joined #git
04:02 danimal2 joined #git
04:03 chachasmooth joined #git
04:03 Forbidd3n hmm I keep getting this error - The following paths are ignored by one of your .gitignore files: and the file is not listed in the current dirctory .gitignore file
04:05 lexi933 left #git
04:05 puff Nobody's home on #github;  I did my initial push for setting up a github repo, but I want to undo that, clear it all out, and start from scratch, i.e. push a significantly different version of the files as the initial import for the repo?  The best I've found is this approach: http://stackoverflow.com/a/18116141/1813403
04:06 puff Whoops, actually this answer: http://stackoverflow.com/a/18113182/1813403
04:06 puff Ah, okay, I pasted the right URL to begin with :-)
04:07 puff No wait, actually that's a different url. Okay, the second one.
04:07 puff My head is spinning a bit.  Git is twisty, but git + eclipse is twistier.
04:08 Goplat joined #git
04:08 lexi933 joined #git
04:09 lexi933 hi
04:10 adymitruk hello lexi933
04:10 shanson joined #git
04:14 Wulf joined #git
04:15 blunaxela joined #git
04:16 aw1 joined #git
04:16 rwp puff, That answer should be fine. For all intents and purposes doing a force push over master will reset the repository to the new master.
04:17 hahuang61 joined #git
04:18 puff rwp: Thans.
04:18 puff Thanks, that is.
04:19 Klumben joined #git
04:19 chachasmooth joined #git
04:19 puff There should be a big button/link on the git homepage, to a page titled "THESE ARE THE THINGS YOU CANNOT DO IN GIT"
04:19 puff :-)
04:19 puff With links to "but you can fake it this way..."
04:19 thiago Forbidd3n: it's listed in another .gitignore
04:20 Forbidd3n thiago: I figured it out, thanks
04:20 Guest12196 joined #git
04:20 Guest17873 joined #git
04:20 Forbidd3n It was in my darn global ignore
04:20 miczac\away joined #git
04:23 thebope_ joined #git
04:24 govg joined #git
04:25 thebope joined #git
04:27 puff rwp: Okay, so next step is to push the new/good version of the initial import in, with all the files.  But previously there was nothing there.  Can I just locally git init, git add ., git commit -m "Initial import, and then: git remote add origin https://github.com/gitub/etc, and then git push -u origin master
04:27 puff ...or do I need to do git push origin --mirror  ?
04:28 rwp Yes to your first. I don't think you need the --mirror. That was an attempt by that author to clear everything. But I don't think it will have that desired effect.
04:28 puff Okay, thanks.
04:29 rwp Github allows you to force push non-fastforward commits on top of master. I think that is really all that you needed.
04:29 lhx puff: never tried that with github... but don't see why that won't work assuming you haven't protected any specific branches
04:29 rwp Technically files are not removed from the object store by that. But they won't be cloned out and they will be garbage collected later. So effectively they will evaporate.
04:29 lor913 joined #git
04:30 lhx yup... there's not a way to view github's reflog right?
04:30 rwp The question I don't know about are if there are any tags in the github repository. Not sure about those.
04:30 puff Yeah, they'll be left as... what's the opposite of a danglig pointer? An unreferenced object, which will hang around until gc'd by git gc.
04:30 Cabanoss- joined #git
04:31 puff Actually, since they're all stored by hash, most of these will be pointed to by the new push, which is mostly the same but missing a bunch of cruft.
04:31 rwp Then those won't matter. But the cruft you have removed from the pointer tree will then be orphaned and garbage collected later.
04:32 rwp You didn't say if you looked to see if you have any tags pushed to the remote though.
04:33 heap_ joined #git
04:35 d^sh joined #git
04:38 freddymungo_ joined #git
04:39 chachasmooth joined #git
04:40 nowhere_man joined #git
04:43 endiruna3 joined #git
04:43 watabou joined #git
04:45 sbulage joined #git
04:47 glspi joined #git
04:47 snowalpaca joined #git
04:48 snowalpaca joined #git
04:53 nevyn joined #git
04:54 patarr joined #git
04:55 dendazen joined #git
04:57 thebope joined #git
04:58 iamthad joined #git
05:03 nickabbey joined #git
05:04 a_thakur_ joined #git
05:08 drizbit joined #git
05:09 puff rwp: Hm, that failed.
05:09 puff rwp: There are no tags, all there was, was my initial commit.
05:09 murph joined #git
05:09 thebope joined #git
05:09 puff rwp: So that shouldn't have created any tags, right?
05:11 thebope joined #git
05:12 pxf_god joined #git
05:14 puff rwp: https://gist.github.com/stevenjowens/799258f5f03f0f9a223ecfb75b5229bd
05:14 pxf_god joined #git
05:16 jimi_ joined #git
05:17 xall joined #git
05:17 SwiftMatt joined #git
05:18 jimklo joined #git
05:22 durre joined #git
05:22 dave0x6d joined #git
05:23 ecuanaso joined #git
05:23 sbulage joined #git
05:24 MineCoins joined #git
05:24 shgysk8zer0 joined #git
05:25 aidalgol joined #git
05:28 thiago puff: you forgot the --force option
05:29 neeri joined #git
05:30 NimeshNeema joined #git
05:43 j_r joined #git
05:47 sbulage joined #git
05:47 a_thakur joined #git
05:47 justanotheruser joined #git
05:49 Nitesh joined #git
05:51 YuGiOhJCJ joined #git
05:53 puff` joined #git
05:53 sbulage joined #git
05:58 Vampire0_ joined #git
05:59 Cabanossi joined #git
06:02 chipotle joined #git
06:04 shanks_ joined #git
06:04 mehola joined #git
06:08 lordjancso joined #git
06:08 pxf_god joined #git
06:08 MineCoins joined #git
06:09 dcy joined #git
06:09 GT4066 joined #git
06:10 lor913 joined #git
06:10 swords joined #git
06:15 peterbecich joined #git
06:15 rscata joined #git
06:17 jozwior joined #git
06:18 yrseo joined #git
06:19 Dayne2 joined #git
06:22 sbulage joined #git
06:24 yrseo .........
06:24 yrseo Is it possible to integrate GOGS with ECLIPSE?
06:25 Xe MAYBE
06:25 Xe it depends on what you mean by "integrate"
06:31 Xe in terms of cloning, pushing and commiting, yes it'd work perfectly
06:31 qt-x joined #git
06:31 Xe in terms of the issue management, i have no idea, there might be an eclipse plugin but there might not be
06:35 dreiss joined #git
06:36 yrseo thanks. I'll try a little more and ask you a question.
06:36 zeroed joined #git
06:36 Raging_Hog joined #git
06:38 sQVe joined #git
06:39 venmx joined #git
06:39 thebope_ joined #git
06:42 overlord_tm joined #git
06:43 thebope joined #git
06:43 ertesx joined #git
06:44 thebope joined #git
06:47 jozwior_ joined #git
06:48 thebope joined #git
06:49 jozwior joined #git
06:49 ThomasLocke joined #git
06:50 thebope joined #git
06:51 acetakwas joined #git
06:51 zulutango joined #git
06:54 patarr joined #git
06:57 ronskisika joined #git
06:58 SteffanW joined #git
07:05 pihpah joined #git
07:06 ThomasLocke joined #git
07:06 chachasmooth joined #git
07:07 jimklo joined #git
07:09 govg joined #git
07:12 variable joined #git
07:13 Learath2 joined #git
07:15 ecuanaso joined #git
07:16 dreiss joined #git
07:20 durham joined #git
07:21 olbat joined #git
07:22 thebope_ joined #git
07:25 durham joined #git
07:25 Atm0spher1c joined #git
07:26 chachasmooth joined #git
07:27 hexagoxel joined #git
07:28 ihradek joined #git
07:29 sbulage joined #git
07:30 nd joined #git
07:31 Vampire0 joined #git
07:31 thebope joined #git
07:31 wiesel joined #git
07:32 eroux joined #git
07:32 noteness joined #git
07:32 thebope joined #git
07:32 gelei007 joined #git
07:33 thebird joined #git
07:33 JeroenT joined #git
07:35 alexandre9099 joined #git
07:35 Atm0spher1c joined #git
07:36 freimatz joined #git
07:39 chele joined #git
07:40 dreiss joined #git
07:40 pur3eval joined #git
07:41 hahuang61 joined #git
07:42 oskarkv joined #git
07:42 JeroenT joined #git
07:43 pitastrudl joined #git
07:44 sarri joined #git
07:44 sarri joined #git
07:46 GodGinrai joined #git
07:47 jnavila joined #git
07:48 Qilibrun joined #git
07:49 sbulage joined #git
07:50 eroux joined #git
07:52 a_thakur joined #git
07:53 govg joined #git
07:54 Vampire0_ joined #git
07:55 beatdown joined #git
07:55 a_thakur joined #git
07:57 venmx joined #git
07:57 Darcidride joined #git
07:58 bomb-on joined #git
07:58 dc5ala joined #git
08:00 puff` thiago: Thanks.
08:01 Tobbi joined #git
08:04 puff I'm having trouble finding a URL that confirms this, "git checkout -d branchname" versus "git branch -d branchname", the first just deletes the label while the second deletes both the label and the commits?  Is that correct?
08:04 nickabbey joined #git
08:05 babilen joined #git
08:05 grawity no and no
08:05 osse puff: the second just deletes the label. the first doesn't exist
08:05 grawity first, there isn't *any* option in git that also deletes commits
08:05 grawity second, `git checkout -d` doesn't exist in the first place
08:05 thebope joined #git
08:06 JattMannu joined #git
08:07 JattMannu hi
08:07 gitinfo JattMannu: 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.
08:07 thebope joined #git
08:07 JattMannu i have a issue
08:07 JattMannu i have a file format that ends with .ini
08:07 JattMannu and git treats it as binary
08:07 JattMannu i want git to treat it as text file
08:07 grawity is it actually text, and if it's text, what encoding/charset does it use?
08:07 JattMannu what should i configure to get that working
08:09 JattMannu ucs-2 little endian
08:11 thebope joined #git
08:11 puff grawity: Ah, thank you.  Shoulda checked the man pages first.
08:12 moritz JattMannu: you can specify the encoding through an attribute. See man gitattributes
08:12 gitinfo JattMannu: the gitattributes manpage is available at http://jk.gs/gitattributes.html
08:12 puff osse: Thank you, too.
08:12 grawity hmm looks like git won't recognize that as text no matter what
08:13 grawity well, to the storage layer it's all the same
08:13 moritz and there's a "text" attribute too
08:13 grawity but if you want diffs you'll probably need to configure a textconv filter
08:14 grawity i.e. something like http://sprunge.us/gTZj
08:15 thebope_ joined #git
08:21 Alienpruts joined #git
08:22 TomyWork joined #git
08:23 g5pw joined #git
08:23 xall joined #git
08:24 ihradek joined #git
08:24 runlevl4 joined #git
08:25 a_thakur joined #git
08:26 theoceaniscool_ joined #git
08:28 jagob joined #git
08:29 dvdchr joined #git
08:31 JattMannu grawity place that line in the gitattributies location?
08:32 sbulage joined #git
08:32 wiesel joined #git
08:33 JattMannu wont it be difficult to update everone's gitconfig file?
08:34 moritz you can commit your .gitattributes file too
08:34 dcy joined #git
08:34 jast sure, configuring the filters everywhere will be a hassle
08:36 moritz dealing with non-UTF-8 files in general is a hassle, IMHO
08:36 aw1 joined #git
08:36 JattMannu legecy code. i cant tell the impact of the change
08:37 fkiejs joined #git
08:37 moritz JattMannu: been there, felt that. My sympathies :-)
08:38 ngm joined #git
08:39 blahdodo joined #git
08:39 Tobbi joined #git
08:42 dersand joined #git
08:42 Balliad joined #git
08:45 grawity tbh, git could at least recognize UTF-16 with BOM :|
08:45 grawity it's also unicode and it's rather common
08:46 JattMannu i am getting this "text-utf-16"] is not a valid attribute name: .gitattributes:34
08:46 JattMannu i have never done this before
08:46 jast what's in that line right now?
08:47 TomyWork joined #git
08:47 Whiskey git archive HEAD -- Start.tcl test.tcl | tar -x -C <-- is it possible to do same export buy give it commit id?
08:48 jast Whiskey: sure, HEAD is just one particular type of reference to a commit after all
08:48 grawity JattMannu: that's not supposed to go to gitattributes
08:48 jast well you're going to have to put *something* in .gitattributes
08:49 Whiskey git archive HEAD -- 18a82c759bc88dfa5f4a02e6d981685c6a056584 | tar -x -C /home/whiskey/www/git/source/ <- not working its latest commit id
08:49 jast yeah, that's because HEAD is the commit reference. if you don't want HEAD, *replace* it with a commit ID.
08:49 jast or a branch name or whatever
08:50 flaviodesousa joined #git
08:50 Vampire0 joined #git
08:50 grawity it's `git archive <commit> <files>`
08:51 Dumblez0r joined #git
08:52 a3Dman joined #git
08:54 thebope joined #git
08:55 Whiskey git archive <commit> <files> ?
08:55 patarr joined #git
08:56 Whiskey but i was hope i don´t need to give it files, and just id to get all files for that id exported
08:56 JattMannu thanks guy! help alot
08:58 grawity ok, `git archive <commit> [files]`
08:58 govg joined #git
08:58 diogenese joined #git
08:58 Whiskey i don´t falloe
08:58 b100s joined #git
08:58 Whiskey git archive -- 18a82c759bc88dfa5f4a02e6d981685c6a056584 | tar -x -C /home/whiskey/www/git/test/ export all files
08:59 Whiskey i just need the files from the commit that actully was modifed
08:59 jast that's not what archive is for, it gives you full snapshots
08:59 Whiskey is it any other tool to get them out
09:00 Sp4rKy_ joined #git
09:00 _ikke_ Whiskey: We already told you how to do it yesterday
09:00 jast in a scripting context you can automatically figure out the list of files
09:01 Whiskey _ikke_: that way don´t work as i espected i get problems whit it
09:01 _ikke_ Whiskey: What kind of problems
09:01 Whiskey Im not sure i if its TCL that don´t like the command
09:02 _ikke_ You can do it in two steps, or execute it in a shell
09:02 Whiskey _ikke_: if {[catch {exec git archive HEAD -- $git(diff) | tar -x -C $directories(source-commit)} error]} { <-
09:02 Whiskey _ikke_: get this errors ":2 fatal: pathspec 'Start.tcl test.tcl' did not match any files, tar: This does not look like a tar archive, tar: Exiting with failure status due to previous errors"
09:02 Whiskey i works fine in the terminal
09:02 leeN joined #git
09:03 Whiskey same command whit same vars i mean :)
09:03 bongjovi joined #git
09:03 mikecmpbll joined #git
09:04 jaith left #git
09:04 nowhere_man joined #git
09:04 jast you're going to have to figure out how to get Tcl to split the value in $git(diff) into multiple parameters for exec
09:04 Whiskey im not sure how you mean
09:05 jast right now exec is treating the content of $git(diff) as a single parameter, i.e. a single file called "Start.tcl test.tcl"
09:05 Whiskey git archive HEAD -- Start.tcl test.tcl | tar -x -C /home/whiskey/www/git/source/ <-- is the command it try to run
09:05 Whiskey aha
09:05 Whiskey :=)
09:05 p4trix joined #git
09:05 jast you need to somehow get it to turn into multiple parameters, but I don't know Tcl enough to help you
09:05 Whiskey its fine, thanks :)
09:06 nowhere_man joined #git
09:06 bomb-on joined #git
09:07 nowhereman joined #git
09:07 manuelschneid3r joined #git
09:07 Whiskey lets see if thats the problem manually
09:07 _ADN_ joined #git
09:08 nowhereman joined #git
09:09 nowhereman joined #git
09:09 Whiskey jast, it was the problem thanks :D
09:10 dimi1947 joined #git
09:11 nowhereman joined #git
09:12 Derperperd joined #git
09:12 zeroed joined #git
09:12 zeroed joined #git
09:13 SwiftMatt joined #git
09:14 rical left #git
09:15 clmsy joined #git
09:17 Masber_080 joined #git
09:18 xpk joined #git
09:18 bryngo joined #git
09:18 BackEndCoder joined #git
09:20 sfix joined #git
09:22 aw1 joined #git
09:22 clmsy joined #git
09:23 skunkz joined #git
09:24 Tobbi joined #git
09:25 bryngo_ joined #git
09:26 grayjoc joined #git
09:26 bryngo_ left #git
09:28 thebope joined #git
09:28 pur3eval joined #git
09:29 wachpwnski-mobi joined #git
09:29 wachpwnski-mobi Where are you supposed to run git gc? Is that something that you should run on your git server?
09:30 grawity most of the time, git runs it automatically when it notices the object store having grown too large
09:30 grawity you can run it manually whenever you want to re-compress stuff
09:31 grawity can be useful after large history rewrites, I guess?
09:31 rossome joined #git
09:31 thebope joined #git
09:32 wachpwnski-mobi So is it accurate to say that it just cleans up and dedupes object files on disk?
09:34 MattMaker joined #git
09:35 grawity the object files don't get duplicated, but yeah
09:35 grawity NAME
09:35 grawity git-gc - Cleanup unnecessary files and optimize the local repository
09:35 acetakwas joined #git
09:36 elect joined #git
09:37 govg joined #git
09:37 thebope joined #git
09:39 Nerbrun joined #git
09:39 Vampire0 wachpwnski-mobi, do you try to understand what `git gc`does or do you want to achieve something and are trying to find out whether `git gc` is the right command for that?
09:39 MrMojit0 joined #git
09:41 seni joined #git
09:42 drodger joined #git
09:42 nettoweb joined #git
09:43 chll_ joined #git
09:43 hahuang61 joined #git
09:44 seni1 joined #git
09:45 wachpwnski-mobi Vampire0: I ran across it. Was wondering if it's something I should be doing on a self managed git system.
09:45 aw1 joined #git
09:45 seni1 joined #git
09:47 afuentes joined #git
09:47 GodGinrai joined #git
09:47 garethdaine joined #git
09:48 thebope joined #git
09:50 aax joined #git
09:54 svs joined #git
09:57 hussam joined #git
09:58 Tobbi joined #git
09:59 SirOliver joined #git
10:01 sQVe joined #git
10:02 mischat joined #git
10:03 tabakhase git on windows - since about 3 weeks my pulls often "freeze" after "unpacking objects" - i need to ctrl+C then, git pull again and it actually plops the branches
10:03 tabakhase what is that, or rather, where to start digging?
10:04 aw1 joined #git
10:06 anioth left #git
10:06 nickabbey joined #git
10:08 venmx joined #git
10:10 bvtton joined #git
10:13 Sp4rKy joined #git
10:13 jfr joined #git
10:13 ych joined #git
10:15 thebope joined #git
10:15 afuentes joined #git
10:15 Vampire0 wachpwnski-mobi, ah, ok, then no, usually you don't need to do this manually
10:16 Vampire0 tabakhase, what is "to plop"?
10:18 tabakhase plop as in "Updating ...." - where it pops out branches and such
10:18 tabakhase no idea how to call that ;D
10:19 Sp4rKy_ joined #git
10:20 tabakhase http://pastebin.com/kBBEfjYr  - maybe that explains it best =)
10:20 rnsanchez joined #git
10:21 stevenhong joined #git
10:22 Sp4rKy_ joined #git
10:23 BooK I'm using sha1 in an application to quickly decide if files are different (so not for security), and I'm wondering if sha1(size(data),data) (which is was git does) is more likely to protect against collisions than sha1(data), given that I have a lot of files to consider
10:24 BooK also, if there's some literature about that. my first internet searched have pointed to interesting but unrelated discussions :-)
10:24 BooK searches
10:24 canton7 BooK, I don't think it's going to make an ounce of difference. sha1 is so incredibly unlikely to collide by accident
10:25 freddymungo_ joined #git
10:25 bernardio joined #git
10:25 borkr joined #git
10:26 BooK canton7: yeah, I was leaning towards that too. even for a million file or so
10:26 oskarkv_ joined #git
10:26 BooK I wonder why git takes the extra steps, though, using sha1(type, size, data)
10:26 wachpwnski-mobi Vampire0: https://paste.ofcode.org/qp63iKFAuXppsdwaWKiaKi <- I ran it and got this. To me it looks like it cleaned up something.
10:26 osse BooK: It's stored therefore for efficiency, so that Git doesn't need to figure it out for every object
10:27 osse s/fore//
10:27 osse Not really related to the usage of sha1 as such
10:28 canton7 BooK, yeah, you'd need fifty thousand billion billion files for a 0.1% probability of a collision
10:28 BooK osse: I don't understand what you mean
10:28 Vampire0 tabakhase, maybe try running git through strace to see what exactly it does during the waiting period
10:28 nettoweb joined #git
10:29 canton7 BooK, and I think, actually, adding the length onto the beginning does nothing to help avoid a collision anyway, mathematically
10:29 BooK osse: or rather, what I understand (you can get the size back from the sha) is impossible, so I must understand wrong :-)
10:29 ChampagneHouse joined #git
10:29 pur3eval joined #git
10:29 Hetman joined #git
10:30 Hetman Hello any way to play situation like that: I've got zip file downloaded from gitlab (files in my repo, no meatada) , I want unzip it on remote side and re-create as repo ... but dont want to download whole repo again as this is reason why i've created zip (poor internet) Any tips how to do that ?
10:31 selckin explain "unzip on remote side"
10:31 moritz Hetman: you can do that, but you won't be able to collaborate with the original repo if you do that
10:31 grawity better clone with --depth=1
10:31 moritz git init; git add .; git commit -m 'Initial import'
10:32 moritz right, what grawity said
10:32 Hetman selckin: got zip with all files in repo but no metadata, unzip this on remote host and I want re-create repo from it . So basically I want to avoid git clone
10:32 grawity that way you save on transferred data *and* still keep the metadata
10:32 BooK reading the README in the original commit in git to see if there are some explanations there :-)
10:32 grawity but recreating a repo from .zip is
10:32 aw1 joined #git
10:32 Vampire0 wachpwnski-mobi, of course it does something if you call it. But git gc is called automatically every so often, so it is not really necessary to be called manually. If you want more often or more thorough cleanup, you can call it manually like you did or with --aggressive to even do  more but need longer
10:32 grawity well, it's *not impossible*, but takes a bit of hacking if you want to preserve commit-IDs and everything
10:32 svs left #git
10:33 selckin the remote also has bad internet?
10:33 grawity hrmm
10:33 sQVe joined #git
10:33 tabakhase Vampire0 no such thing on mysgit/winbash :D
10:33 * grawity experiments
10:33 Hetman ok so unzip ,git init, git add . , git commit , git push . Not have other choice as customer got shit internet link . Remote had good internet
10:33 grawity oh you definitely have other choices
10:33 moritz no git push
10:33 Vampire0 wachpwnski-mobi, if this is about disk size saving, you can also first expire the reflogs and after an aggressive gc repack the whole repository with a wide windows
10:33 _ikke_ Hetman: note that when you git init and commit in a new repo, you cannot just push to another existing repostiory
10:34 courrier joined #git
10:34 Hetman ok this is not perfect
10:34 Vampire0 tabakhase, well, maybe you should not use msysgit. msysgit is just the development environment for Git for Windows which is what end-users should use
10:34 grawity Hetman: how large is the .zip?
10:34 Vampire0 tabakhase, I don't know if it is related to your problem though
10:34 Hetman grawity: 2GB
10:35 grawity Hetman: and how large would the whole repo be?
10:35 Hetman 2.5GB
10:35 grawity hmm
10:35 Hetman its not big but they on some shit adsl in timezone +11h vs remote server :D
10:35 grawity well, really should have downloaded a clone when you had good internet, tbh
10:35 grawity but...
10:35 aw1 joined #git
10:35 acetakwas joined #git
10:36 selckin wonder if he could use like sshfs or if git likes to read all blobs anyway
10:36 tabakhase Vampire0 thats a first time hearing... i remember a time where that was the suggested method o.O -- and i need my msys for a few other things too.. so win-cmd is not rly an option
10:36 BooK canton7, osse: https://github.com/git/git/blob/e83c5163316f89bfbde7d9ab23ca2e25604af290/README#L35-L40 this seems to be the reason behind it
10:37 Vampire0 tabakhase, Git for Windows has a Git Cmd *and* a Git Bash and I would never use the former
10:37 Vampire0 Besides that I use Cygwin anyway
10:38 tabakhase hm maybe that is waht im talking about... *scrubs head*
10:38 joeco joined #git
10:38 aw1 joined #git
10:39 aw1 joined #git
10:39 joeco Hi. What am I thinking wrong about using "git rm --cached -r..."? I want all the subdirs with files of a certain type to be untracked.
10:40 aw1 joined #git
10:40 moritz git rm doesn't know about about file types
10:40 joeco weird. And thanks, moritz
10:40 joeco but it caught a couple within the first nested dir o.O
10:40 joeco just not any beneath that.
10:40 grawity if you meant the filename extensions, then yes
10:41 grawity but remember that wildcards are expanded *by the shell*
10:41 grawity and the shell doesn't know anything about the -r option when it does that
10:41 joeco the actual command I used: "git rm --cached -f -r client/*.js"
10:41 grawity it just expands the command to "git rm --cached -f -r client/foo.js client/bar.js"
10:41 bremner you can use find(1) for this kind of thing
10:41 grawity with git specifically, you can quote the wildcards and let git itself expand them (since it can do that)
10:42 joeco right.. so I need a wildcard for "add directories below too"..?
10:42 moritz if you use zsh, you can say client/**.js
10:42 thebird joined #git
10:42 joeco yes bremner I'm thinking I might end up going that route.. zsh not on this machine.
10:42 Vampire0 joeco, how about `find -type f -name '*.js' -exec git rm --cached {} +`?
10:42 moritz otherwise find client/ -name '*.js' | xargs git rm --cached
10:42 grawity in bash, that works after a `shopt -s globstar`
10:42 moritz (assumes well-behaved file names)
10:42 grawity otherwise quote it and let git expand it: `git rm ... 'client/**.js'`
10:43 bremner no real reason to use xargs with find, other than find has terrible syntax.
10:43 aw1 joined #git
10:43 moritz grawity: a, nice to know
10:43 Vampire0 joeco, or maybe even `git rm --cached -f -r 'client/*.js'` might work
10:43 thebope_ joined #git
10:43 joeco trying '../**.js'
10:43 moritz bremner: or if performance is an issue, and you don't want to invoke your command for each file separately
10:44 bremner I guess. but then you should at least use -print0
10:45 thebope joined #git
10:46 joeco 'git rm --cached -f -r 'client/**.js' (and also git rm --cached -f -r 'client/**.map') did the trick. Thx all
10:47 Fissure joined #git
10:47 raijin joined #git
10:49 netj joined #git
10:49 thebope_ joined #git
10:49 acetakwas joined #git
10:51 yrseo joined #git
10:54 govg joined #git
10:55 bvtton joined #git
10:56 Tobbi joined #git
10:56 patarr joined #git
10:58 navidr joined #git
10:58 dsdeiz joined #git
10:58 muthu joined #git
11:04 mehola joined #git
11:05 mehola joined #git
11:05 mehola joined #git
11:06 mehola joined #git
11:07 mehola joined #git
11:08 mehola joined #git
11:09 mehola joined #git
11:09 mehola joined #git
11:10 mehola joined #git
11:12 Derperperd joined #git
11:14 sbulage joined #git
11:15 Celelibi I was previously tracking only two branches from a remote.
11:15 Celelibi Now I want to delete one. How do I do?
11:15 pxf_god joined #git
11:18 rchavik joined #git
11:18 theoceaniscool_ joined #git
11:19 moritz what do you want to delete? the local branch? the remote branch? the tracking?
11:19 Celelibi The tracking and the local refs to the remote branch.
11:20 moritz is the branch gone on the remote server?
11:20 Celelibi Well, to remove the tracking I could just edit .git/config
11:20 Celelibi Yes, the branch is gone.
11:20 moritz git fetch --prune
11:21 aw1 joined #git
11:21 Celelibi "fatal: Couldn't find remote ref refs/heads/modifs"
11:22 PioneerAxon joined #git
11:22 nettoweb joined #git
11:23 SirOliver joined #git
11:23 tvw joined #git
11:24 dviola joined #git
11:25 thebope joined #git
11:27 thebope__ joined #git
11:28 dimi1979 joined #git
11:29 xall_ joined #git
11:29 a3Dman joined #git
11:31 dimi1979 joined #git
11:31 lowercaseman joined #git
11:33 grayjoc joined #git
11:34 ni291187 joined #git
11:35 corrado joined #git
11:35 MattMaker joined #git
11:36 thebope joined #git
11:36 dispix joined #git
11:37 SirOliver joined #git
11:38 ggaeit joined #git
11:38 ni291187 left #git
11:40 SwiftMatt joined #git
11:48 GodGinrai joined #git
11:50 jimi_ joined #git
11:50 itsgokul joined #git
11:51 inflames joined #git
11:51 syg_ joined #git
11:52 Darcidride joined #git
11:52 DBag joined #git
11:53 albel727 joined #git
11:54 noecc joined #git
11:54 patarr joined #git
11:55 lss8 joined #git
11:55 sQVe joined #git
11:59 sQVe joined #git
12:02 SirOlive_ joined #git
12:06 PtxDK joined #git
12:08 dimi1979 I have made a change to a file but I do not want to create a separate commit for it...instead I want to link it to a past commit..what must I do?
12:10 venmx joined #git
12:11 nedbat dimi1979: do you want to add it to the latest commit?
12:11 mehola joined #git
12:12 dimi1979 nedbat:not to the lasted commit but to a commit 2-3 positions back
12:12 temhaa joined #git
12:14 canton7 dimi1979, have you pushed any commits since the commit you want to link it to?
12:14 canton7 or done any merges since the commit you want to link it to?
12:15 thebope joined #git
12:17 mischat_ joined #git
12:18 dimi1979 canton7:Yes...I do have push some commits since then
12:18 dimi1979 that makes it harder
12:18 canton7 dimi1979, yeah - does anyone else have access to the repository you pushed to?
12:18 pur3eval joined #git
12:18 jimi_ joined #git
12:19 askb joined #git
12:19 dimi1979 canton7:no...it is a local repository and I am the only one using it...
12:20 canton7 dimi1979, ok, so you pushed to another local repository?
12:20 sQVe joined #git
12:20 dimi1979 canton77:there is only one local repo
12:20 KnightsOfNi Hey, all of a sudden I get this message: fatal: Not a git repository (or any parent up to mount point /home/vagrant/Projects)
12:20 KnightsOfNi Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set). - while the .git directory is still there...
12:20 nettoweb joined #git
12:20 canton7 dimi1979, you said you pushed some commits since the commit you want to edit. now you're saying that there is only one repo, and you haven't pushed anywhere? Which one of those is true?
12:21 moritz KnightsOfNi: where exactly is the .git directory?
12:21 KnightsOfNi in the projects/myproject/ directory where it has always been
12:21 moritz but that's not where git is looking
12:21 xall_ joined #git
12:22 moritz so it seems you've invoked git in /home/vagrant/Projects
12:22 dimi1979 canton7:I said I pushed some comits to one repo and that repo is a local one...what is the problem with that statement?
12:22 moritz but when it worked, it was in /home/vagrant/Projects/myproject
12:22 KnightsOfNi no I invoke it in /myproject/
12:22 canton7 dimi1979, you said there's only one local repo. There must be at least two in your description - the repo you're currently working inside, and the repo you pushed to
12:22 KnightsOfNi Like I've been doing for years
12:23 KnightsOfNi the error message just says:  fatal: Not a git repository (or any parent up to mount point /home/vagrant/Projects)
12:23 moritz then maybe some path mapping in vagrant has changed
12:23 dimi1979 canton7:there is only one repo...I want to be clear about that...
12:23 canton7 dimi1979, anyway, commit the change to the file, and pass '--fixup=hash_of_the_commit_to_add_the_file_to' to the commit command
12:23 KnightsOfNi moritz, where can I check on that?
12:23 moritz KnightsOfNi: if your Vagrantfile is under version control, check its history
12:24 moritz check your installer's log if there were any vagrant updates
12:24 canton7 dimi1979, then 'git rebase -i hash_of_the_commit_to_add_the_file_to^'. You should see your latest commit added with the 'fixup' line. Save and exit the editor, and let rebase do its thing
12:24 moritz etc.
12:24 dimi1979 canton7:give me a minute or so cause I have never run that command again...
12:24 canton7 dimi1979, I'm still confused. You're currently inside a repository. That's one. You pushed to another local repository. That's two. How is there only one repo?
12:24 KnightsOfNi It's not under version control
12:25 dimi1979 canton7:I never said I pushed to another repo...I cannot understand how you reach this conclusion
12:25 canton7 <dimi1979> canton7:Yes...I do have push some commits since then
12:25 canton7 ^ right there
12:26 dimi1979 canton7:and how this statement leads to the conclusion that there are more than one repo
12:26 moritz KnightsOfNi: then compare with your last backups
12:26 canton7 dimi1979, when you push, you have to push from one repository to another repository. For you to be able to push, there has to be at least two repositories
12:26 SirOliver joined #git
12:27 mizu_no_oto joined #git
12:27 dimi1979 canton7:I really do not understand...anyway I want to focus now the command
12:27 Murr joined #git
12:27 canton7 dimi1979, before you run the command, wait
12:27 dimi1979 ok
12:28 canton7 dimi1979, I need to make sure that noone else has access to the commits you pushed
12:28 canton7 dimi1979, if someone else has fetched the commits you pushed, and then you edit them to add this file in, that's going to cause all sorts of trouble
12:28 KnightsOfNi Going to try to upgrade vagrant to the latest version first
12:29 moritz right, change more variables, that'll make it easier to track down /s
12:29 canton7 dimi1979, let me try and reword the question using different words then... You pushed your latest commits. Could anyone else have fetched them from wherever you pushed them to?
12:30 dimi1979 canton7:NO!!!
12:30 canton7 ok, great
12:30 canton7 please don't shout :)
12:30 dimi1979 canton7:can we focus on the command now please?
12:30 GustavoCarvalho joined #git
12:31 canton7 sure, go for it
12:31 dimi1979 canton7:give me a second
12:34 Whiskey can git tell me the root dir of repositorie?
12:34 osse git rev-parse --show-toplevel
12:35 SirOliver joined #git
12:35 Whiskey it don´t output anything
12:35 canton7 maybe you're not in a repository?
12:36 Whiskey i am
12:36 Learath2 joined #git
12:36 canton7 !repro
12:36 gitinfo [!transcript] Please paste (using https://gist.github.com/ or similar) a transcript ( https://git.io/viMGr ) of your terminal session so we can see exactly what you see
12:36 dimi1979 canton7: is this command correct from a syntax point of view? commit --fixup=Frontend/profile_regular.php     Frontend/profile_regular.php is the file to be commited
12:36 osse no
12:36 canton7 dimi1979, no
12:36 Eryn_1983_FL joined #git
12:36 canton7 dimi1979, I said '--fixup=hash_of_the_commit_to_add_the_file_to'
12:36 Whiskey well i don´t see anything, i do the command and then it come a new line that tell me next command to do
12:36 Whiskey nothing to telll
12:36 Whiskey no error nothing
12:37 dimi1979 canton7:oh I forgot that
12:37 Masber joined #git
12:37 osse Whiskey: what does git rev-parse HEAD say
12:37 Whiskey it give me the id of last commit i think
12:38 osse what's your git version
12:38 sQVe joined #git
12:39 Whiskey hmm
12:39 Whiskey 1.91
12:39 Whiskey 1.9.1
12:40 osse good enough
12:40 osse try git rev-parse --git-dir
12:40 Vampire0 Whiskey, and !repro
12:40 gitinfo Whiskey: [!transcript] Please paste (using https://gist.github.com/ or similar) a transcript ( https://git.io/viMGr ) of your terminal session so we can see exactly what you see
12:40 helmet33 joined #git
12:40 Whiskey that work
12:40 dimi1979 canton7:I have one question...is it natural to assume that git must know which file I want to add to the commit?this command does not tell git that
12:40 osse now type --show-toplevel again
12:40 DolpheenDream joined #git
12:41 Whiskey nope nothing
12:41 canton7 dimi1979, I was assuming that you'd either use 'git add' as appropriate, or add '-a' to the commit command
12:41 Whiskey strange that it not put any error or something
12:42 dimi1979 canton7:do not assume anything...just tell me what to after I made this clarification
12:42 Whiskey Vampire0: whats your problem its nothing to show
12:42 osse GIT_TRACE=1 git rev-parse --show-toplevel
12:42 canton7 dimi1979, I have to assume you have a level of basic knowledge. I assume you know how to use a computer, that you know how to use a console, that you can type commands and use your editor. Why not assume that you've read the first chapter of a git book, too?
12:42 Whiskey osse: https://gist.github.com/anonymous/0502ec894441c071ee88442a29b54015
12:43 Whiskey as i told, nothing
12:43 osse looks like a bare repository
12:43 Whiskey bare?
12:43 osse in which case there is no top level directory, because there are no checked out files
12:44 canton7 !bare
12:44 gitinfo A bare repository is used to push/fetch (useful for running a git server), and contains only the contents of .git/ from a "normal" repo. Read more: http://bare-vs-nonbare.gitrecipes.de/
12:44 canton7 (this, by the way, is why repros are useful - we can spot things that you either miss or don't think are important enough to mention)
12:45 Vampire0 Whiskey, be polite if people try to help. The transcript was essential to right away see that it is a bare repository which is why it doesn't output anyhting just like osse said
12:45 osse run ''git config core.bare''
12:45 zzz joined #git
12:45 Vampire0 Whiskey, being unpolite is not good if you are hunting for help from volunteers
12:46 Whiskey Vampire0: its just make me angry as i did not actully have anything to show
12:46 osse you did, you just didn't know it
12:46 canton7 Whiskey, it *did* though - it told us that you're in a bare repo
12:46 Whiskey No output means no output and it was exactly what it is
12:46 canton7 Whiskey, no! the line "/home/whiskey/b/vcs/gitea/repositories/whiskey/hidden.git" told us that
12:46 Vampire0 Whiskey, a transcript not only contains the result, but also the command
12:46 Whiskey How the hell can it show that?
12:46 Whiskey How would the none bare looks like?
12:46 Vampire0 And as you are in a Gitea repository that is bare, there is no top level to show
12:47 canton7 the folder name ends in .git, which is a common pattern for bare repos
12:47 Whiskey Vampire0: in that case im sorry for that
12:47 osse a non-bare repo looks like a bunch of files and directories you can modify
12:47 osse with a .git directory next to them
12:47 dimi1979 canton7:Unfortunately I have not read the first chapter of git book...
12:47 Vampire0 Whiskey, as canton7 said, the .git is common for bare, additionally you have gitea in your path which is a git hosting software that should always have the repos bare, otherwise you should not push into it
12:47 canton7 dimi1979, can I recommend you do that now, before trying to do more advanced stuff like editing history? !book is good
12:47 gitinfo dimi1979: There are several good books available about git; 'Pro Git' is probably the best: http://git-scm.com/book but also look at !bottomup !cs !gcs !designers !gitt !vcbe and !parable
12:48 jameser joined #git
12:48 Whiskey :=)
12:48 Vampire0 Whiskey, typical case of !just ;-)
12:48 gitinfo Whiskey: [!just_ask] You can just ask your question. If anybody knows the answer, they will answer soon (most of the time)
12:48 Vampire0 ah, wrong trigger
12:48 texinwien joined #git
12:49 Vampire0 Whiskey, I meant typical case of !justdo ;-)
12:49 gitinfo Whiskey: [!justdoit] If you go to the trouble of asking questions, please go to the trouble of taking the advice offered, Take backups (type "!backup" as a reply to learn more) if necessary -- or in fact, it's a good idea anyway. You can almost always censor log messages if you have to.  We are literally helpless unless you tell us what we need or do what we ask.  Please, Just Do It™.
12:49 canton7 Whiskey, so, in future, you're only here asking questions because you can't figure out the answer. That means you don't know what information is important. When we ask for something, there's a reason for it - we have a *lot* of experience debugging git issues
12:49 marcogmonteiro joined #git
12:51 nettoweb joined #git
12:51 dimi1979 canton7:what segment of git book will help me do what I want?
12:51 canton7 dimi1979, some later chapter, if at all. That's not the point. The point is that you should have a basic working knowledge of git before trying to do things like edit old commits
12:51 canton7 that's why I suggested that you read up on the basic stuff furst
12:51 canton7 *first
12:52 ShapeShifter499 joined #git
12:52 dimi1979 canton7:of course I have some basic knowledge of git...
12:52 Gustavo6046 joined #git
12:52 osse dimi1979: you have managed to create some commits, apparantly. git commit --fixup is no different. you need to use git add, or use commit -a or something to indicate what you want to record as changed
12:52 canton7 you wanted explicit instruction to 'git add' your file when committing it. That to me implies the opposite
12:53 NullableTruth joined #git
12:53 dimi1979 canton7:the file is already added to the index...if that is what you mean
12:54 canton7 great, yes, why do you want an explicit instruction to 'git add' it then?
12:54 dimi1979 it is being tracked
12:54 dispix joined #git
12:54 jameser joined #git
12:55 dimi1979 canton7:you said that...not me...I just asked how am I going to link the file to be commited to an older commit
12:56 canton7 <canton7> dimi1979, I was assuming that you'd either use 'git add' as appropriate <dimi1979> canton7:do not assume anything...just tell me what to after I made this clarification
12:56 osse dimi1979: git rebase takes care of that, after the new commit has been created
12:57 dimi1979 osse:so you are saying to commit the file and then using rebase to merge the 2 commits?
12:57 osse yes
12:57 osse that is what i am saying
12:57 canton7 and using commit --fixup will tell rebase which two commits to merge
12:57 osse ... if autosquash is set, but that's different matter
12:58 canton7 (it is by default)
12:58 osse oh?
12:58 dimi1979 I will proceed with osse's recommendation and see what happens
12:58 osse well that's something that can be removed from my .gitconfig then
12:58 dispix Hello everyone, I am looking for some insights around rebasing multiple branches that stem from one another after un update on the "master" branch ?
12:58 dimi1979 give me a second
12:58 canton7 hmm, I might be misremembering
12:59 canton7 dimi1979, we suggested the same thing :)
12:59 canton7 dimi1979, you should pass --autosquash to the rebase -i command, as well
12:59 osse dispix: git rebase master firstbranch; git rebase --onto firstbranch firstbranch@{1} secondbranch; git rebase --onto secondbranch secondbranch@{1} thirdbranch
13:00 Whiskey Hmm i have a problem
13:00 osse dispix: https://github.com/Osse/git-stuff/blob/master/rebase-onto-2.txt
13:00 MattMaker joined #git
13:00 Whiskey I have commit-msg/post-commit that i made, it runs fine in my own repo when i commit, but can´t it run from the remote repo?
13:01 osse hooks are a per repo configuration and cannot be pushed
13:01 dimi1979 osse:remind me please...DO i have to stash changes before using rebase?
13:01 dispix thanks osse :)
13:01 canton7 Whiskey, no, the post-commit hook is only run in the repo it's configured in, and only when the user types 'git commit'
13:01 osse dimi1979: can't remember
13:01 SwiftMatt joined #git
13:01 osse dispix: actually, my link doesn't show the --onto thing. it might in fact not be needed
13:05 dimi1979 osse:I am into rebase -i mode..what to go from there?
13:05 osse dimi1979: what do you see?
13:06 synthroid joined #git
13:06 Whiskey Is there any hook the can be run from the remote?
13:06 mischat joined #git
13:07 osse several
13:07 osse post-receive for example
13:07 Whiskey yeh, but thats after or before commit is accepted?
13:07 masuberu joined #git
13:07 helmet33 joined #git
13:08 dimi1979 osse:just some commented instructions about the available commands and at the top something saying:rebase xxxx...xxxx onto xxxx where xxxx's are hashes
13:08 dimi1979 osse:the goal is to merge two comits
13:09 qqx Whiskey: man githooks
13:09 gitinfo Whiskey: the githooks manpage is available at http://jk.gs/githooks.html
13:09 MineCoins joined #git
13:10 jameser joined #git
13:10 Tobbi joined #git
13:10 Whiskey yeh i just don´t find what i search, i guss its same as commit-msg/post-commit but for remote side
13:11 venmx joined #git
13:11 osse Whiskey: what do you want it to do?
13:12 osse dimi1979: jsut that? nothing else?
13:12 hamza joined #git
13:12 dimi1979 osse:yes...do you want me to prepare a snapshot?
13:12 osse yes
13:12 qqx Whiskey: No commits are made on a remote. So there something like a commit-msg or post-commit ho9ok makes no sense.
13:13 osse or tell me, what exactly did you run when you ran rebase?
13:13 mehola joined #git
13:13 re1 joined #git
13:13 Derperperd joined #git
13:13 Whiskey Reject commit massage if its not in any "way" that scripted
13:13 hamza left #git
13:13 tobiasvl post-receive then
13:14 tobiasvl or pre-receive
13:14 tobiasvl probably the latter to reject it
13:14 bschindler joined #git
13:14 osse definitely the latter
13:14 bschindler Hi - I realized that I'm having a mess within my repository with line endings. Some files are committed as crlf, some are lf
13:15 bschindler As we have the autocrlf option, is there a way to find out for a given file, whether it is stored as crlf or lf in the repository?
13:15 svs joined #git
13:16 osse bschindler: git show HEAD:file | file -
13:16 dimi1979 osse:https://1drv.ms/i/s!AjCBrCpLQye-g09YLGCme8amSwQV
13:16 osse it will say "with CRLF line terminators" if that's the case
13:16 svs hi all. I have a remote project which only contains a readme.md file. I did a pull, then added some other folders with my data. then I committed and did a push. my folder gets created remotely, but no other subdirectories or files are uploaded. any hints, please?
13:17 bschindler osse, I suppose I can replace HEAD with a commit hash right?
13:17 osse bschindler: yes
13:17 Whiskey are hooks shared. if do git pull? like post-commit if its in the remote repo?
13:17 bschindler osse, thanks a lot. That will save me... :o
13:17 osse dimi1979: it also said noop. please don't leave things out.
13:17 osse dimi1979: you've invoked rebase wrongly
13:17 CheckDavid joined #git
13:17 osse Whiskey: they are not shared
13:18 Whiskey So every memeber i add, need to manually add it
13:18 dimi1979 osse:I knew that I must have said that also..it puzzled me
13:18 tobiasvl svs: did you remember to add the files inside the subdirectories?
13:18 wiesel joined #git
13:19 svs tobiasvl: yes. I created a folder, then used git init, add,then copied my entire fulders/files substructure to the local repo, and then committed and pushed. I can confirm my files are where they should belong... locally. but not on remote
13:20 thebird joined #git
13:20 sonOfRa svs: are you sure you committed anything?
13:20 sonOfRa Did you stage your changes via git add?
13:20 tobiasvl svs: you did "git add" and THEN copied the folders and files to the local repo and THEN committed? you did not do "git add" after copying in the folders?
13:21 tobiasvl svs: instead of looking at what files and folders you see visually locally, just run "git status" and see what git thinks of those files
13:21 svs # On branch master \ nothing to commit, working directory clean
13:21 b3kN joined #git
13:21 b3kN joined #git
13:21 svs the order was: init, pull, added my files, then add command, then commit and lastly push
13:22 svs I see that a subfolder has been added remotely, but nothing else
13:22 tobiasvl ok, that's a different order
13:22 revoltingPeasant joined #git
13:22 tobiasvl svs: !repro
13:22 gitinfo svs: [!transcript] Please paste (using https://gist.github.com/ or similar) a transcript ( https://git.io/viMGr ) of your terminal session so we can see exactly what you see
13:23 svs sure. give me a minute
13:25 a_thakur joined #git
13:25 qqx svs: Are you expecting push to actually create files in the remote working copy? It doesn't, at least by default.
13:25 govg joined #git
13:26 nedbat i made a tag that I don't need any more. It seems impossible to get rid of it, because it's been pulled into other working trees, and any time one of them pushes tags, they will push it back to the shared repo.  Is there a way to truly get rid of an unwanted tag?
13:26 svs qqx: yes, I'm expecting it to do so
13:26 qqx It can be made to do so by setting `receive.denyCurrentBranch` to `updateInstead`.
13:26 DaveTaboola joined #git
13:28 svs thank you
13:28 tobiasvl svs: is the remote a bare repo? github/gitlab/bitbucket?
13:28 tobiasvl ah there you go
13:28 re1_ joined #git
13:28 osse Whiskey: yes, that's why it's a good idea to have verification stuff on the server
13:29 seni1 joined #git
13:29 Gsham joined #git
13:30 absinthe joined #git
13:31 svs tobiasvl: it's repository inside my current intranet
13:33 Darcidride joined #git
13:34 ws2k3 what does this exacly do?         denyCurrentBranch = ignore
13:36 GodGinrai joined #git
13:37 qqx ws2k3: I'm assuming you're referring to receive.denyCurrentBranch there. That allows pushing to the branch that is currently checked out in a non-bare repository.
13:38 qqx It's generally a bad idea, since the working tree and index will be out of sync.
13:39 qqx So it will appear that the new commits which were pushed have been reverted (but not committed).
13:41 wiesel joined #git
13:41 dendazen joined #git
13:42 nowhere_man joined #git
13:45 hahuang61 joined #git
13:45 canton7 ws2k3, all of the config options are in man git config
13:45 gitinfo ws2k3: the git-config manpage is available at http://jk.gs/git-config.html
13:46 Derperperd joined #git
13:46 treia joined #git
13:46 ephimetheus joined #git
13:46 svs tobiasvl: https://gist.github.com/snevs/90b052021a076a26dd93ef2a2d6523a8
13:46 svs I have deleted everything and reproduced it
13:46 marcogmonteiro joined #git
13:47 svs he only thing I can see remotely is the password-change folder and nothing else
13:48 tobiasvl svs: you can notice that locally too: 1 file changed, 1 insertion(+) and create mode 160000 password-change
13:48 thebope joined #git
13:48 svs is there something preventing sub-directories from being added to the repo?
13:49 DaveTaboola joined #git
13:49 svs what's the difference between 'add .' and 'add *' ?
13:50 qqx Normally `add *` wouldn't pick up files in the current directory which start with `.`. Otherwise should do the same.
13:50 svs thanks
13:50 qqx Although in completely different manners.
13:50 svs is there a solution for me to add an entire tree to the repo?
13:51 Gsham joined #git
13:51 merlac joined #git
13:51 qqx `add .` says to add the current directory. With `add *` the * gets expanded by the shell to all (non-dot) content of the current directory.
13:52 qqx Either of those should add the entire tree, other than ignored files.
13:53 Learath2 joined #git
13:53 suttonwilliamd_ joined #git
13:54 NeXTSUN joined #git
13:55 bschindler|work joined #git
13:56 qqx svs: In the transcript from your link is password-change a submodule?
13:57 svs it's a sub-directory
13:57 jeffreylevesque joined #git
13:58 jamick joined #git
13:58 adymitruk joined #git
14:00 thebope joined #git
14:00 qqx A submodule would also be a sub directory.
14:00 Learath2 joined #git
14:00 qqx But a directory by itself would not be listed as being created in the output of `git commit`.
14:01 bvtton joined #git
14:03 tinanoo joined #git
14:03 temhaa left #git
14:04 tobiasvl to add on that: git does not track empty directories, so it's not possible to add just a direct to git, and none of its contents
14:04 tobiasvl so password-change is not a directory
14:04 tobiasvl is it a symbolic link? a submodule?
14:05 Tobbi joined #git
14:05 tobiasvl mode 160000 is apparently a submodule, so it's a submodule
14:05 tobiasvl to be precise it's a special file that points to a commit object, so I guess it could be something else, but my bet is a submodule
14:05 tobiasvl svs: ^
14:06 thebird joined #git
14:07 pur3eval joined #git
14:07 King_Hual joined #git
14:08 Learath2 joined #git
14:08 nickabbey joined #git
14:08 dispix joined #git
14:08 kfunk joined #git
14:12 venmx joined #git
14:12 svs I see... so it converts my directory
14:13 svs to a submodule to my main project
14:13 theoceaniscool joined #git
14:14 jameser_ joined #git
14:14 tobiasvl well, presumably the directory is a git repo
14:15 Derperperd joined #git
14:15 tobiasvl it's kind of the other way around: it's NOT converting your git repo to a flat subdirectory :)
14:16 throstur joined #git
14:16 throstur I wanted to merge one commit from another branch into master, but I ended up accidentally pushing the whole branch, and now on github when offered to do a pull request it wants to merge the stuff I don't want merged too. What can I do?
14:17 Darcidride joined #git
14:17 Learath2 joined #git
14:18 tobiasvl throstur: well, what did you want to do? "merge one commit" does not make sense in git
14:18 tobiasvl throstur: presumably you wanted to cherry-pick one commit from another branch or something instead? man git cherry-pick
14:18 gitinfo throstur: the git-cherry-pick manpage is available at http://jk.gs/git-cherry-pick.html
14:18 synthroid joined #git
14:18 DDBag joined #git
14:19 ash_workz joined #git
14:19 throstur tobiasvl: I guess I could have said "merge one file"
14:19 osse even worse!
14:19 Hetman left #git
14:19 tobiasvl even that does not make sense. merge in git is a merging of two git HISTORIES
14:20 tobiasvl throstur: this is just a fork on github that nobody else dabbles with? if so, it can be fixed quite easily
14:20 throstur well then, I want to rip a file from another branch and replace master's file with that one
14:20 madhur joined #git
14:20 d-fence joined #git
14:20 throstur it's my private repo for stuff
14:20 chipotle joined #git
14:21 kanzure joined #git
14:21 tobiasvl throstur: so what command did you run? "git merge something" followed by "git push"?
14:21 marcogmonteiro joined #git
14:22 mischat joined #git
14:22 throstur hmm... I'm not quite sure anymore, history shows various things, some of which failed and I cannot recall which
14:22 throstur however, it seems git checkout other_branch filename did what I wanted
14:22 xmj joined #git
14:22 xmj moin
14:22 jimi_ joined #git
14:22 xmj can anyone tell me how i can see a list of branches available on remote?
14:23 xmj i'm currently trying to grab code from a gerrit review that i've pushed, that's been added to, and that i'd like to CRF - but, somehow, i don't get which branch to use for pulling
14:23 JamesBaxter joined #git
14:23 jameser joined #git
14:23 tobiasvl throstur: yep, that'll do it. how you undo what you did by mistake earlier kind of depends on what you did, but you can probably just !rebase_interactive away the wrong history regardless
14:23 tobiasvl ugh. !interactive_rebase
14:23 gitinfo Interactive rebase sounds similar to rebase but has completely different abilities. It can do this to commits: change the order, squash some of them together, remove some, add random existing commits from other branches, split them, and more... it's very powerful. Documentation is in the section "Interactive Rebase" in 'man git-rebase'.
14:23 qqx xmj: `git ls-remote <remote-name-or-url>`
14:24 Learath2 joined #git
14:28 fyrril joined #git
14:28 Tobbi joined #git
14:28 throstur thanks
14:29 hobodave joined #git
14:29 xmj zing thanks qqx
14:29 xmj left #git
14:30 Derperperd joined #git
14:31 ShekharReddy joined #git
14:31 Cabanossi joined #git
14:31 jameser joined #git
14:32 troulouliou_div2 joined #git
14:32 suttonwilliamd__ joined #git
14:33 thebope joined #git
14:33 freimatz joined #git
14:34 nowhere_man joined #git
14:35 raijin joined #git
14:36 shanson joined #git
14:37 fyrril left #git
14:38 NeXTSUN joined #git
14:39 mustmodify__ joined #git
14:40 mustmodify__ can I get a log of commits for a specific line of a file? kind of like a git log / git blame thing?
14:41 karthik joined #git
14:41 eroux joined #git
14:41 _ikke_ mustmodify__: git log accepts -L
14:41 mustmodify__ Ohhhh
14:41 Learath2 joined #git
14:41 mustmodify__ hunh. ok thanks.
14:41 * moritz tends to use tig blame
14:42 moritz oh, -L seems to be new-ish
14:42 GodGinrai joined #git
14:42 moritz 1.7 didn't have it, at least :-)
14:42 texinwien_ joined #git
14:43 * moritz uses 1.7 on some old systems and up to 2.7.4 on newer ones
14:44 eroussel joined #git
14:44 griffindy joined #git
14:44 nettoweb joined #git
14:46 Tobbi joined #git
14:46 snowalpaca joined #git
14:48 mustmodify joined #git
14:48 DBag joined #git
14:48 ecuanaso joined #git
14:50 troulouliou_divn joined #git
14:51 Anticom joined #git
14:51 le_melomane joined #git
14:51 koneko joined #git
14:55 _Marek_ joined #git
14:56 Learath2 joined #git
14:56 Dougie187 joined #git
14:56 phy1729 I have a repo that I converted from svn and then ran the bfg to move some globs to lfs. I've run  git reflog expire --expire=now --all && git gc --prune=now --aggressive  but there are still some large blobs. The weird thing about them is that the object id doesn't show up in  git rev-list --objects --all  and  git fsck --unreachable  lists nothing (as expected after the expire and prune). Any idea wtf is
14:56 phy1729 keeping this blob around?
14:57 mustmodify left #git
14:59 svs left #git
14:59 nedbat i made a tag that I don't need any more. It seems impossible to get rid of it, because it's been pulled into other working trees, and any time one of them pushes tags, they will push it back to the shared repo.  Is there a way to truly get rid of an unwanted tag?
14:59 suttonwilliamd_ joined #git
15:01 canuk joined #git
15:01 wiesel joined #git
15:01 MattMaker joined #git
15:02 AaronMT joined #git
15:03 Es0teric joined #git
15:04 bvtton joined #git
15:05 dcpc007 joined #git
15:05 bernardio joined #git
15:06 Es0teric joined #git
15:06 JamEngulfer joined #git
15:07 nicks_85 joined #git
15:07 Peuc joined #git
15:08 acetakwas joined #git
15:08 nicks_85 Hi , I have two repositories one is under /puppet/git/puppet.git and another is in /puppet-2016.4/git/puppet.git I would like to compare the production branch changes and copy it to text file , How do I do it ?
15:08 Learath2 joined #git
15:09 mischat joined #git
15:11 thebope joined #git
15:12 venmx joined #git
15:13 Iacobus__ left #git
15:13 Peetz0r joined #git
15:15 nedbat nicks_85: if those are really both the same repo, but at different commits, it might be easier to just diff the commits in one repo
15:16 jwest joined #git
15:16 Iacobus joined #git
15:17 chra94 joined #git
15:18 SirOliver joined #git
15:18 le_melomane joined #git
15:22 RobertS joined #git
15:23 morenoh149 joined #git
15:23 morenoh149 how do I get tab completion for branch names
15:25 Tobbi joined #git
15:26 ckb joined #git
15:27 shgysk8zer0 joined #git
15:27 shgysk8zer0_ joined #git
15:28 adymitruk joined #git
15:28 lexi933 joined #git
15:28 nettoweb joined #git
15:30 nettoweb1 joined #git
15:34 Darren_ joined #git
15:36 netj joined #git
15:38 JamesBaxter joined #git
15:41 Learath2 joined #git
15:43 crose joined #git
15:43 ResidentBiscuit joined #git
15:43 zoharsacks[m] joined #git
15:44 LeBlaaanc joined #git
15:45 Derperperd joined #git
15:45 troulouliou_div2 joined #git
15:46 hahuang61 joined #git
15:46 jsho joined #git
15:46 atomi joined #git
15:46 mischat_ joined #git
15:49 jsho joined #git
15:49 bongjovi joined #git
15:52 chardan joined #git
15:53 chardan joined #git
15:53 Ryanar joined #git
15:56 pur3eval joined #git
15:57 hobodave joined #git
15:58 j7k6 joined #git
15:58 Hink joined #git
15:59 chardan joined #git
15:59 Learath2 joined #git
16:00 eroux joined #git
16:01 boombatower joined #git
16:02 mischat joined #git
16:02 cdg joined #git
16:03 Hink joined #git
16:03 Es0teric joined #git
16:07 cjohnson nedbat: you can't stop people from pushing new things
16:07 nedbat cjohnson: yeah, that's what I was afraid of.
16:07 cjohnson nedbat: I would probably suggest you instruct people not to push all branches
16:07 cjohnson only push what they intend to push
16:07 nedbat cjohnson: these are tags
16:07 patarr joined #git
16:07 cjohnson tags, branches, same diff
16:09 irqq joined #git
16:10 xissburg joined #git
16:12 bannakaffalatta joined #git
16:14 mehola joined #git
16:16 zeroed joined #git
16:16 zeroed joined #git
16:16 nedbat cjohnson: is there anyway to create the tag in the shared repo so that people trying to push the tag wouldn't be able to change it?
16:16 GT4066 joined #git
16:16 armin anyone of you a clue what i'm overlooking here? https://paste.fedoraproject.org/paste/Efhr-V7cZfMraaU~MpzDgV5M1UNdIGYhyRLivL9gydE=
16:18 dminuoso armin, did you by accident do some "sudo git" before? Check permsissions on these files/directories
16:18 salamanderrake jmesmon: sorry for the long wait, https://gist.github.com/salamanderrake/a8cac5ab742e60572a3173c64f705a39
16:18 salamanderrake jmesmon: that was git status
16:18 Glooomy joined #git
16:19 armin dminuoso: looks good to me, permissions seem to be fine.
16:19 dminuoso armin, and ownership?
16:19 dminuoso :-)
16:19 armin dminuoso: also fine, i related to user/group as well when you said permissions.
16:19 dminuoso armin, right. and whoami ?
16:20 armin dminuoso: armin
16:20 dminuoso You didn't happen to fiddle around with user ids right?
16:20 Tobbi joined #git
16:20 armin no.
16:20 dminuoso armin, do you have atlassian software on your server?
16:21 armin dminuoso: there is no server. it's just a git init i did on a directory.
16:21 dminuoso This is quite strange.
16:21 solarfly joined #git
16:22 user121212 joined #git
16:23 ankit01ojha joined #git
16:23 jimklo joined #git
16:23 user121212 Hi, I'm going through a fairly big git diff. Is there anyway speedup the scroll down key in less?
16:24 dave0x6d joined #git
16:24 dminuoso user121212, if you have something specific to search for, you can use / to search for the string. scrolling is quite slow naturally.
16:24 dminuoso armin, did you check the entire directory hierarchy for these path specs?
16:25 cjohnson nedbat: seems pointless and extra confusing. but yeah sure if you were to push a tag whose history is incompatible with other people's tags, it wouldn't let them push until they resolved
16:25 osse user121212: press 'h' to see available movements
16:25 cjohnson nedbat: but then, they can always just pull and rebase and fix whatever intentional conflicts you introduced
16:25 bttf joined #git
16:25 nedbat cjohnson: i've resigned myself to having these tags lying around, but i'm interested to know more about what esoteric things are possible.
16:26 Darren_ joined #git
16:26 cjohnson nedbat: policy is the best approach imo
16:26 nedbat cjohnson: i didn't realize that pushing a tag would require compatibility with a previous version of the tag.
16:26 cjohnson like I said, treat a tag like a branch
16:26 cjohnson imagine if you tried pushing a diverged branch
16:26 cjohnson it would fail
16:26 nedbat cjohnson: how do I enforce treating a tag like a branch?
16:28 cjohnson No I'm saying
16:28 cjohnson think of it like a branch
16:28 cjohnson tags are a lot like branches
16:29 shinnya joined #git
16:29 nedbat cjohnson: but if i push a tag, and it's referring to something completely unrelated to the existing tag, git won't complain.
16:29 armin i guess what cjohnson is trying to say is: a branch is just a pointer to the sha. so is a tag.
16:29 qqx nedbat: You could install an `update` hook on the repo which would reject the unwanted tag.
16:29 xall_ joined #git
16:30 nedbat armin: yes, but a crucial difference is that branches are meant to move, and there are checks for their proper movement, which raises errors if the branch is moved wrong.  Not so with tags, unless I'm missing something.
16:30 nedbat qqx: the repo is github :(
16:30 armin nedbat: you can as well have a tag that moves. you should not do that, of course...
16:31 armin nedbat: you should also not reference any state to by a branch name when communicating with other humans.
16:32 armin nedbat: also beware, you could have a branch and a tag with the same name. if you want to avoid major confusion, don't do that.
16:32 nedbat armin: yes, thanks.  i was exploring ways to get rid of obsolete tags.
16:32 armin nedbat: what's an "obsolete tag"?
16:32 nedbat but it seems like there is not a good way, so I will let them be.
16:32 theoceaniscool joined #git
16:33 armin nedbat: depends on your view of "obsolete", i guess. ;)
16:33 nedbat armin: a tag i needed for just an hour or so.  It has to do with testing releases, where a tag name and a branch name are not interchangable due to ansible bugs.
16:33 raspado joined #git
16:33 armin nedbat: ah ok, well, i would just not get rid of tags, ever, once you published them to a remote.
16:33 hahuang61 joined #git
16:34 nedbat armin: i think that is what i am left with, yes.
16:34 SilverBe joined #git
16:34 KnightsOfNi My .git directory is over 700MB, why is it so big and how do I reduce this?
16:34 armin KnightsOfNi: stop commiting your movies collection to it?
16:34 armin ;)
16:35 armin KnightsOfNi: no, really, try to find out what you committed what is so huge, and if you didn't publish that already revert the commits.
16:35 KnightsOfNi There isn't much content in the repo itself anymore, it's really the .git directory only
16:35 qqx Git can compress an entire movie collection to 700MB? Guess I need to move mine into git. :)
16:35 armin KnightsOfNi: that's because git keeps a history of what you had committed there, of course, so you have the possibility to go back.
16:36 armin qqx: :D
16:36 KnightsOfNi the project itself is only 150MB in total
16:36 svm_invictvs joined #git
16:36 armin KnightsOfNi: you could check with "ncdu" what's so huge there.
16:37 brent__ joined #git
16:37 KnightsOfNi .git/objects/
16:37 KnightsOfNi is where all the size is
16:37 Emperor_Earth joined #git
16:37 armin KnightsOfNi: and what actual object in there is so huge?
16:37 KnightsOfNi .git/objects/pack/
16:37 armin that's a directory.
16:38 KnightsOfNi pack-261ee....pack
16:38 KnightsOfNi 1 file 702MB
16:38 Pcatinean_ joined #git
16:38 armin then there's a commit object with a commit id that starts with 261ee. check what it is with git log and see what the commit included as a change.
16:39 qqx armin: No, that's a pack file which contains a lot of objects.
16:39 armin ah.
16:39 armin thanks for clarifying.
16:40 PCatinean joined #git
16:40 PCatinean hey guys
16:40 PCatinean how do I merge 10-12 commits from one branch to another?
16:40 PCatinean a certain range I meant
16:40 PCatinean Instead of cherry-pinking each of them?
16:42 nicks_85 nedbat its the same repo with different commits , How do I diff the commits ?
16:42 Murii joined #git
16:42 PCatinean wait, you're a python wizard and also help on #git? cheers nedbat :))
16:43 nedbat PCatinean: well, i was here to get help, but... :)
16:43 PCatinean hahaha :)))
16:44 nedbat nicks_85: http://stackoverflow.com/a/29374476/14343
16:45 safinaskar joined #git
16:45 livingstn joined #git
16:45 safinaskar left #git
16:46 safinaskar joined #git
16:46 safinaskar i am sitting on local branch "dev". how to push from local branch "dev" to remote branch "dev" on github repo called "origin"?
16:46 YuGiOhJCJ joined #git
16:46 _ikke_ safinaskar: git push origin dev
16:47 Random832 probably just git push will work
16:47 _ikke_ depends on the push.default setting
16:47 mischat joined #git
16:47 safinaskar _ikke_: thanks
16:47 safinaskar left #git
16:48 Random832 I've gotten in the habit of using git push {remote} HEAD:refs/heads/{branch} because the shorter forms had some kind of problem in gerrit
16:50 brntspawn joined #git
16:51 synthroid joined #git
16:52 Gsham joined #git
16:53 elijah_ joined #git
16:53 tlaxkit joined #git
16:54 KnightsOfNi can I delete that .pack file?
16:55 nutron joined #git
16:55 nutron joined #git
16:55 jeremyhall joined #git
16:55 Exagone313 joined #git
16:56 kiltzman joined #git
16:56 zoharsacks[m] left #git
16:56 pur3eval joined #git
16:58 _ikke_ I would not do that
16:58 _ikke_ It contains most of your history probably
16:58 KnightsOfNi damn
16:58 KnightsOfNi Guess I need to start a new repo then
16:59 _ikke_ KnightsOfNi: Did you commit large objects in the past?
16:59 KnightsOfNi yes _ikke_
16:59 KnightsOfNi but I removed them later, they were images
16:59 KnightsOfNi git rm -r -f /images/
16:59 KnightsOfNi I don't need them in the history anymore either
16:59 skered joined #git
17:00 nutron joined #git
17:00 _ikke_ Right, but just removing them doesn't remove them from history
17:00 bernardio joined #git
17:00 thiago joined #git
17:00 _ikke_ !bfg can remove them (but that is rewriting history)
17:00 gitinfo A tool designed to remove large files, or passwords from history: https://rtyley.github.io/bfg-repo-cleaner/ (!rewrite applies)
17:00 KnightsOfNi apparently
17:01 madewokherd joined #git
17:01 KnightsOfNi I'll look into that, thanks
17:02 MattMaker joined #git
17:03 nutron joined #git
17:05 jccn joined #git
17:06 matoro joined #git
17:07 leeN joined #git
17:07 jccn joined #git
17:10 Ryanar joined #git
17:10 jccn joined #git
17:12 atomi joined #git
17:13 jagob joined #git
17:13 tyreld joined #git
17:13 venmx joined #git
17:14 tusharmalik joined #git
17:15 Vampire0_ joined #git
17:16 kpease joined #git
17:16 gingitsune joined #git
17:18 nickabbey joined #git
17:23 bket joined #git
17:24 Vampire0 joined #git
17:26 Ozymandy joined #git
17:27 svm_invictvs joined #git
17:29 glspi joined #git
17:29 xall_ joined #git
17:30 tang^ joined #git
17:33 Learath2 joined #git
17:36 Derperperd joined #git
17:37 diogenese joined #git
17:37 jost_ joined #git
17:38 ToBeCloud joined #git
17:39 MattMaker joined #git
17:40 metachr0n joined #git
17:40 joki joined #git
17:40 govg joined #git
17:41 durham joined #git
17:42 robotroll joined #git
17:43 jeremyhall joined #git
17:45 denisMone joined #git
17:46 mikecmpbll joined #git
17:46 raynold joined #git
17:46 patarr joined #git
17:46 BackEndCoder joined #git
17:46 ChiLL_SurF joined #git
17:49 thebird joined #git
17:50 crayon joined #git
17:55 g105b joined #git
17:56 nowhere_man joined #git
17:57 pur3eval joined #git
17:59 Synthead joined #git
18:02 ckb joined #git
18:02 patarr joined #git
18:03 _ADN_ joined #git
18:03 robattila256 joined #git
18:04 Darcidride joined #git
18:07 thiago joined #git
18:08 bernardio joined #git
18:10 overlord_tm joined #git
18:11 physikoi joined #git
18:12 physikoi Hello #git! Quick question, does the Travis service allow for adding production files to the repo automatically so that production files need not be tracked?
18:15 wilbert joined #git
18:15 mehola joined #git
18:20 Eugene That sounds like a Travis question, not git
18:20 Eugene Unless you're really asking about !config_files ?
18:20 gitinfo [!configfiles] It is recommended to store local configuration data in a file which is not tracked by git, but certain deployment scenarios(such as Heroku) may require otherwise. See https://gist.github.com/1423106 for some ideas
18:20 pmatulis joined #git
18:20 le_melomane joined #git
18:22 Tobbi joined #git
18:22 pmatulis say i've branched, done some work, and committed the changes. later, maybe much later, i go back to this branch and would like to remind myself of the changes i made. what command would i use?
18:22 e01 joined #git
18:22 e01 joined #git
18:23 kadoban pmatulis: 'git log'
18:23 pmatulis oh ok :)
18:23 pmatulis but that won't give the actual changes, just the commit messages?
18:24 pmatulis i'd like to see the actual changes
18:24 kadoban gitk --all or !lol can be helpful too.  Well, git log takes all kinds of options. If you want to see actual changes, --stat or --patch should help.
18:24 gitinfo A nifty view of branches, tags, and other refs: git log --oneline --graph --decorate --all
18:24 [Brain] joined #git
18:25 sorressean left #git
18:26 qqx pmatulis: `git log -p`
18:27 pmatulis qqx, thanks, looks promising!
18:28 thebird joined #git
18:29 cr34ton joined #git
18:31 nowhereman joined #git
18:33 nowhere_man joined #git
18:35 eroussel joined #git
18:38 ecuanaso joined #git
18:39 Es0teric joined #git
18:40 matoro joined #git
18:40 dimi1979 joined #git
18:40 morenoh149 joined #git
18:42 kevr joined #git
18:43 manuelschneid3r joined #git
18:45 finalbeta joined #git
18:47 thebird joined #git
18:51 le_melomane joined #git
18:52 CussBot joined #git
18:52 sathed joined #git
18:53 ekinmur joined #git
18:55 jimi_ joined #git
18:55 jimi_ joined #git
18:56 SwiftMatt joined #git
18:57 aswen joined #git
18:57 physikoi hello Eugene , thanks for the input. much appreciated
18:57 EricPoe joined #git
18:57 livingstn joined #git
18:57 physikoi left #git
18:58 jsho joined #git
18:59 khmarbaise joined #git
19:00 jmcantrell_ joined #git
19:03 b100s joined #git
19:04 j7k6 joined #git
19:06 treehug88 joined #git
19:08 leeN joined #git
19:09 TomyLobo joined #git
19:10 miczac\away joined #git
19:10 morenoh149 joined #git
19:11 ecuanaso joined #git
19:11 Sasazuka joined #git
19:12 pmatulis left #git
19:12 Sasazuka joined #git
19:14 venmx joined #git
19:14 moL-Ou joined #git
19:16 Whiskey If i always make a backup of my .git on the server, it always possible to restore it if something go wrong
19:16 Vampire0_ joined #git
19:17 cdg joined #git
19:18 saml_ joined #git
19:18 yottabyte joined #git
19:19 saml_ there was a commit that introduced a file  foo.txt.   but due to history rewrite, the file's gone. is there a way to search for the file in orphan commits or objects?
19:19 osse saml_: yes, but not directly
19:19 bernardio joined #git
19:20 yottabyte every time I try get fetch I get some unable to update local ref, there a lot of branches being created, I don't know if this has any bearing. I'm on windows 8.1. I have to do git gc --prune=now like every time I want to do it, then the git fetch succeeds
19:20 osse find all non-referenced objects, of those find the trees, view those to see which contains file.txt and its sha1
19:20 saml_ it's a github project (private repo). can there be commits in github?
19:20 osse what do you mean
19:20 saml_ ah i see
19:20 grawity much better chance of finding them on your local repo, tbh
19:21 snowalpaca joined #git
19:21 jnavila joined #git
19:21 yottabyte I do this because rather than checking out the branch locally and then git pull for something I want to merge with mine (say A → B), in B I just do "git fetch && git pull origin A" — I don't need a local checkout of A, I can just fetch the latest remote changes or however the right way to phrase it is (please tell me how to say it)
19:22 systemovich joined #git
19:22 BlackFox left #git
19:22 saml_ git rev-list --all --objects  this is all objects?
19:22 aw1 joined #git
19:22 saml_ i mean, un-referenced trees
19:23 grawity no
19:23 grawity as the manpage says, --all lists *from all refs*
19:23 nowhereman joined #git
19:23 grawity you need `git fsck` (e.g. with --lost-found, or just grep the default output) to find the unreferenced ones
19:23 saml_ git fsck --unreachable
19:23 ecuanaso joined #git
19:24 grawity nah, just `git fsck | grep "dangling commit"`
19:24 joshszep joined #git
19:24 grawity and maybe go through `git reflog` first
19:25 DBag joined #git
19:26 BlackFox joined #git
19:26 BlackFox I have a kind of a stupid question
19:26 BlackFox Why can't git root folder be a symlink on NTFS/Windows?
19:26 BlackFox neither symlink nor junction
19:27 moL-Ou joined #git
19:28 Vampire0 joined #git
19:28 DDBag joined #git
19:29 treia joined #git
19:29 Eugene BlackFox - No good reason in particular; probably a bug and Patches Welcome. NTFS symlinks/junctions are not POSIX-style that `git` expects, so things are "funky"
19:30 m0viefreak joined #git
19:31 grawity NTFS symlinks were deliberately added to be a POSIX-style alternative to junctions
19:31 grawity but I guess msys doesn't quite understand them
19:31 grawity anyway, you can use a gitfile instead
19:31 grawity i.e. make .git a regular file with contents "gitdir: /path/to/whatever"
19:32 BlackFox The thing is, for Windows apps normally these paths are transparent
19:32 grawity git isn't a full Windows app
19:32 BlackFox but alas
19:32 BlackFox "fatal: Invalid symlink 'E:/Repo': Function not implemented"
19:32 grawity since it uses the msys shim
19:33 saml_ it's not there
19:33 BlackFox I'm not sure why it has to check if root folder is a symlink
19:33 BlackFox Why not just naively do its thing
19:35 jccn joined #git
19:36 joshszep joined #git
19:37 Vampire0_ joined #git
19:43 pur3eval joined #git
19:43 robotroll joined #git
19:43 Gsham joined #git
19:44 todd_dsm joined #git
19:46 jsho joined #git
19:48 nettoweb joined #git
19:48 rewt`` joined #git
19:52 Derperperd joined #git
19:53 cd-rum joined #git
19:59 gopar joined #git
20:01 aidalgol joined #git
20:03 trewe joined #git
20:05 fatalhalt joined #git
20:06 SirOliver joined #git
20:08 irqq joined #git
20:09 knuth0 joined #git
20:11 satifant joined #git
20:11 iamrohit7 joined #git
20:11 iamrohit7 how do i undo "git checkout .."?
20:12 iamrohit7 i accidentally deleted all the working changes
20:12 iamrohit7 that i have been working on
20:12 iamrohit7 since a week
20:12 cjohnson step 1 invent a time machine
20:12 cjohnson step 2 go back and start staging and committing more often
20:12 ojacobson iamrohit7: do you have backups? git doesn't retain work tree contents when overwriting them, and generally assumes you meant what you asked for
20:12 cjohnson I usually go to my editor and hit save in all the files that I can when I do that
20:12 cjohnson and then make a note to stop making so many changes without commiting
20:13 iamrohit7 cjohnson: so, all that is gone?
20:13 ojacobson I have a very complicated protip
20:13 ojacobson whose punch line is "install a pre-commit hook that runs your tests, and get in the habit of using `git commit` to run them"
20:13 cjohnson iamrohit7: yep sorry
20:13 cjohnson ojacobson: good idea actually
20:13 ojacobson I have a stack of aliases, including `git wip` which snapshots the work tree verbatim, involved in this
20:14 ojacobson so I actually use `git wip` to run tests ;)
20:14 cjohnson git can't track changes git deosn't know about
20:14 cjohnson and it doesn't know until you git add
20:15 venmx joined #git
20:15 cjohnson iamrohit7: really though, you don't have an open editor you can save your files in?
20:15 iamrohit7 cjohnson: i did a git diff just now and the content is still in the terminal. i am so damn lucky
20:15 cjohnson HA nice
20:15 cjohnson very good
20:15 Vampire0 joined #git
20:16 GodGinrai joined #git
20:16 cjohnson so yeah, stage/commit more often, and if you don't have one already, switch to a filesystem that supports snapshots
20:16 iamrohit7 cjohnson: i use atom and it didn't do anything like that
20:16 cjohnson and enable hourly snapshots
20:16 * osse sometimes rely on that intentionally
20:16 cjohnson oh you mean, if you change a file, atom reloads instantly?
20:16 ojacobson atom tends to reload on focus, yeah
20:16 ojacobson most OS X editors do unless they have unsaved changes
20:17 cjohnson ouchie
20:17 cjohnson vim doesn't :D
20:17 cjohnson lol
20:17 ojacobson textmate, helpfully, lets you cmd-z the reload to get the pre-reload state back
20:17 cjohnson that's good
20:17 ojacobson yeah, it's a nice compromise, I wish it were better-documented
20:17 ojacobson I found out because I tried it once on a lark
20:18 iamrohit7 usually, i get by because i use vim.
20:18 iamrohit7 but i was editing js files, so someone suggested me to use atom.
20:19 cjohnson vim does JS
20:20 grayjoc joined #git
20:20 iamrohit7 cjohnson: i can't get away without suggestions and vim doesn't do it well.
20:20 iamrohit7 at least, for me
20:22 johnny56 joined #git
20:22 cjohnson yeah there are various plugins to do it but I have admittedly not spent too much time on them. ##javascript has some good advice though
20:23 ResidentBiscuit joined #git
20:24 hasc joined #git
20:25 iamrohit7 ojacobson: cjohnson: thanks
20:25 DDDBag joined #git
20:26 bernardio joined #git
20:26 iamrohit7 left #git
20:28 guampa joined #git
20:30 buckowski joined #git
20:31 SirOliver joined #git
20:31 miczac\away joined #git
20:36 bluezone joined #git
20:37 thebird joined #git
20:37 lexi933 joined #git
20:38 nettoweb1 joined #git
20:39 m4sk1n joined #git
20:44 Vampire0 joined #git
20:44 tlaxkit joined #git
20:47 winfli joined #git
20:52 cagomez joined #git
20:52 dreiss joined #git
20:55 BlackFox ojacobson, "install a pre-commit hook that runs your tests" <-- what should happen if these tests fail?
20:55 BlackFox Commit should be refused?
20:57 cjohnson hell yeah
20:57 Sleepy63 joined #git
20:57 cjohnson 23 hour lockdown
20:59 brent__ joined #git
21:00 dminuoso BlackFox, here we use webhooks to run tests after pushes.
21:00 diogenese joined #git
21:00 dminuoso Meaning you can push all you want, but we dont accept a pull request unless all tests have passed. :-)
21:00 dminuoso or merge request rather.
21:01 BlackFox Sounds good
21:01 snowalpaca joined #git
21:01 ojacobson BlackFox: yep
21:01 snowalpaca joined #git
21:01 Es0teric joined #git
21:01 ojacobson you can always suppress the hook when you need to with `--no-verify`
21:03 BlackFox sounds kinda mean
21:03 BlackFox I don't neccessarily keep integrity between commits, only integrity after final commit
21:05 j7k6 joined #git
21:05 ojacobson Sure. That's why I said it was a complicated protip: it's actually hiding a lot of conceptual detail and complexity under a simple-sounding bit of advice
21:05 dsdeiz joined #git
21:05 dsdeiz joined #git
21:07 osse BlackFox: also --amend/--fixup exists :P
21:07 cdg joined #git
21:07 ojacobson Yeah. I should rewrite my `wip` alias to generate commit serieses that are fixup-friendly
21:08 ojacobson (rn it gives each commit a unique message, whereas it should probably give the first wip commit a unique message and all subsequent commits that are continuous with it `fixup:` messages)
21:09 qqx cjohnson: vim can be configured to reload files when they change on disk (but haven't been changed in vim) with 'set autoread'.
21:10 cjohnson yeah I know but by default it does the (imo) better approach of not doing that
21:10 qqx IMO the autoread behavior is better.
21:12 qqx Much nicer than if I make a bunch of changes to a file, then when I try to save it tells me that it's changed on disk leaving me wondering if I should force the save or not.
21:13 Guest85 joined #git
21:14 pur3eval joined #git
21:15 cdg_ joined #git
21:16 tvw joined #git
21:16 GodGinrai joined #git
21:16 nettoweb joined #git
21:20 nickabbey joined #git
21:21 zoli joined #git
21:21 Rodya_ joined #git
21:22 howitdo joined #git
21:22 howitdo joined #git
21:29 le_melomane joined #git
21:35 Murii joined #git
21:35 netj joined #git
21:36 chipotle joined #git
21:36 JeremyD joined #git
21:38 zoli left #git
21:39 boritek joined #git
21:39 boritek hello, how can I version control /etc/hosts under a different project folder like /opt/git/project/?
21:40 jnavila joined #git
21:41 perlpilot boritek: with symlinks?
21:42 perlpilot boritek: surely etckeeper or ... (that's its name, see https://github.com/ingydotnet/...) would have some instruction on how to do this.
21:42 DDBag joined #git
21:42 Atm0spher1c joined #git
21:42 dminuoso perlpilot, that won't do it as git does not follow them.
21:42 dminuoso perlpilot, etckeeper works by directly versioning /etc
21:42 dminuoso boritek, you either use hardlinks or mount bind
21:43 dminuoso or well, you have a symlink at /etc/hosts if you can live with that.
21:43 dminuoso boritek, generally it sounds like a bad idea though.
21:44 mikecmpbll joined #git
21:45 pks joined #git
21:46 boritek dminuoso: why is it a bad idea?
21:48 Gsham joined #git
21:48 dviola joined #git
21:48 perlpilot boritek: you could rsync your git working dir to /etc  assuming you can keep a working dir like you want /etc to be.
21:49 Tobbi joined #git
21:50 miczac joined #git
21:54 dminuoso boritek, can you explain your intentions?
21:56 cdg joined #git
21:56 ecuanaso joined #git
21:56 Somelauw joined #git
21:57 Somelauw I have stashed something, but don't remember what. I tried git stash shows, but all it tells me is 25 insertions, one deletion. No way to see the stash without applying it?
21:57 _ikke_ git show stash -p
21:57 _ikke_ git stash show -p
21:59 le_melomane joined #git
21:59 Somelauw _ikke_: thanks, code turned out to be a literal copy of code i already have
22:00 Somelauw so i'll drop it
22:00 brent__ joined #git
22:00 metachr0n joined #git
22:01 venmx joined #git
22:02 Es0teric joined #git
22:02 hanshenrik__ joined #git
22:05 UTAN_dev joined #git
22:05 durham_ joined #git
22:09 blackwind_123 joined #git
22:10 fckfreenode joined #git
22:10 BackEndCoder joined #git
22:11 boritek dminuoso: i want to monitor the hosts file on host, and automatically push it, if it has changed, so that i can have it, if i pull the project
22:12 freddymungo_ joined #git
22:12 fckfreenode what's the best resource for learning git?
22:12 kadoban fckfreenode: This place. You should probably prime your knowledge first with one of the available books though, maybe pro git, which is available free on the website.
22:13 _ikke_ fckfreenode: !book
22:13 gitinfo fckfreenode: There are several good books available about git; 'Pro Git' is probably the best: http://git-scm.com/book but also look at !bottomup !cs !gcs !designers !gitt !vcbe and !parable
22:13 jeffreylevesque joined #git
22:13 fckfreenode awesome, thanks
22:14 * RichiH is somewhat amused by the helpless rage of fckfreenode evident in their nick
22:15 _ikke_ didnt even notice that :P
22:15 dminuoso boritek, yeah hardlink it into your git repository.
22:15 gopar joined #git
22:15 dminuoso should be fine in your caser.
22:16 brent__ joined #git
22:16 boritek dminuoso: how is it possible if I only have a bare repo?
22:17 mehola joined #git
22:17 glspi joined #git
22:18 boritek I guess it is only possible with a working repo
22:18 dminuoso boritek, do you need this to be completely automated?
22:19 meZee joined #git
22:19 dminuoso boritek, if so look into either etckeeper and/or inotify
22:19 DDDBag joined #git
22:21 Somelauw If someone clones your repo and puts your repo in a subfolder, what is the right way to piggy-back a commit from that?
22:21 Rodya_ joined #git
22:22 kadoban Somelauw: If it's only once in a while, maybe ... clone theirs, use filter-branch to have it match your structure, and cherry-pick the commit from it?
22:22 kadoban If it's something you're going to want to do repeatedly ... maybe look into git subtree (contrib command), though I'm not 100% sure how well that'd work without thinking about it more.
22:23 DBag joined #git
22:23 Somelauw kadoban: is it okay to copy the changes and give credit in source code saying taken from ...?
22:24 AustinMatherne joined #git
22:24 _ikke_ depends on the license
22:24 kadoban Somelauw: That's rather beyond a git question, but it ... ya that ^
22:24 _ikke_ and on who owns the copyright
22:24 kadoban If there's no license, typically no, unless you can ask for one.
22:24 Somelauw it's GPL3
22:25 kadoban Then probably yes, as long as you follow GPL3 license terms, which are ... lengthy, though not all that strict really. You'd want to read up on what that means though.
22:26 snowalpaca joined #git
22:29 durham joined #git
22:29 pur3eval joined #git
22:36 peterbecich joined #git
22:41 roxlu joined #git
22:41 Noldorin joined #git
22:43 cdg_ joined #git
22:43 roxlu hi! I just did something like `git submodule add giturl` and it cloned the repository, though it failed to add it as a submodule because of my .gitignore file. I have updated my .gitignore but I'm wondering if I can add the cloned repository as a submodule now?
22:43 morenoh149 joined #git
22:45 roxlu oh I can just execute the same command ... nvm
22:48 jccn joined #git
22:48 snowalpaca joined #git
22:49 absinthe joined #git
22:53 cdown joined #git
22:56 _ADN_ joined #git
22:58 askb joined #git
23:01 fmeerkoetter joined #git
23:02 howlinbash joined #git
23:04 fckfreenode do you guys use git for every single piece of software you write? It's something i discoved relatively late, and it's never really been a habit for me (also due to the fact that i'm still only beginner level). Is git something i should be using whenever i code? I feel like it's a good habit to get into
23:04 preaction yes
23:04 Thorn joined #git
23:04 ToxicFrog fckfreenode: pretty much, yes
23:05 Insentience joined #git
23:05 Insentience Hello
23:05 om_henners joined #git
23:05 ToxicFrog Not everything gets its own *repo* -- A lot of stuff ends up in ~/bin/ or ~/devel/misc/ -- but it all gets versioned.
23:05 ToxicFrog (unless it can't be)
23:05 j7k6 joined #git
23:05 Insentience How am I able to set a remote branch to track another remote branch
23:06 fckfreenode ok, thanks guy
23:08 dminuoso fckfreenode, yes.
23:08 cdg joined #git
23:08 dminuoso fckfreenode, git adds value with the first commit.
23:08 dminuoso and more value with every subsequent commit.
23:08 howlinbash hello #git i'm trying to use a --bare repo to track my dotfiles like this: http://bit.ly/2gu7yvA However i'd like to create 2 repos from it on gitlab so one can be private (with all tracked dotfiles) and one could be public (with the few i wish to share). my thought was to have a master branch (with less files; for public) and a private branch with all files. each branch would be attached to a different remote repo.
23:08 howlinbash so master goes to my-public-dotfiles and private goes to my-private-dotfiles. Is this possible or have i got it all wrong? thanks
23:09 preaction howlinbash: i would advise against that and instead have two completely-separate repos. you could, if you wish, have the private one _include_ the public one via submodules or subtrees
23:09 nedbat howlinbash: just curious: why are some of them secret?
23:10 Emperor_Earth joined #git
23:11 howlinbash nedbat, good question. i was thinking of uploading my public ssh keys to the private repo cause i would like to access them from computers i work on. i was being cautious
23:11 ToBeCloud joined #git
23:11 le_melomane joined #git
23:11 Dougie187 left #git
23:12 [ADN] joined #git
23:12 nedbat howlinbash: hmm, i see
23:12 nedbat howlinbash: public keys are public. it's the private keys you have to worry about.
23:12 howlinbash preaction, that makes me think. i have dot aliased to git specifically for this repo. if i understand what you're saying i should do the same again so have dot aliased for my public files and dots for my private or something like that... ?
23:12 preaction and putting your private keys on any other machine at all is _BAD_
23:13 preaction i, for example, have hooks in my dot files so that i can override things on a per-machine basis (so my work machine has work things on it)
23:14 howlinbash sure, i have no intention of moving or sharing my private keys anywhere, i guess i would like to include ALL (except stuff like private keys) my dotfiles in private repo so i can track and move my system and share just the few i want to in an unrelated repo
23:15 howlinbash preaction, i don't really know what hook are
23:15 preaction for example, my .zshrc has if [[ -f ~/.zshrc.after ]]; then source ~/.zshrc.after; fi
23:16 svm_invictvs joined #git
23:17 howlinbash preaction, does that mean "if .zshrc exists, load it"?
23:17 cdown joined #git
23:17 preaction if .zshrc.after exists, yes
23:17 preaction so i can do machine-specific overrides that don't need to exist on any other machine
23:17 preaction (also, because they're private)
23:18 inovas joined #git
23:19 howlinbash something your saying is not making sense to me. sorry. i see that snippet around. but i don't understand what your explaining to me as to how that works and why its useful. sorry if i'm just being a bit thick
23:19 Takumo joined #git
23:19 Takumo joined #git
23:21 nickabbey joined #git
23:21 howlinbash and i guess, because i see this snippet around, i'm very curious as to what is does. it seems there is a cool trick i don't know about!
23:22 b1tchcakes joined #git
23:22 preaction it allows me to have private things without making them public
23:23 howlinbash fckfreenode, by the way git's amazing. a smart fella taught me to code with it. make a small change work then 'commit' that change with a description explaining what you fixed. and then keep coding and commiting one baby step at a time. very cool tool
23:24 fckfreenode howlinbash: yeah i can definitely see the potential, and i guess on some basci level i understand this, but it hasn't "sunk in" enough yet to make me use it (although i'm trying to change that now)
23:24 Darren_ joined #git
23:24 howlinbash like in a sort of modular way. so you could split your .zshrc into 2 differnt files. 1 you share the other never leaves your pc...
23:24 howlinbash ?
23:25 preaction howlinbash: yes, exactly
23:25 howlinbash preaction, gotchya! thanks for taking the time!
23:25 fckfreenode i guess on a logical level i know it makes sense, but there's a certain amount of uncomfortability that comes with integrating git into my everyday programming toolset
23:25 preaction howlinbash: https://github.com/preaction/dot-files/blob/master/zshrc#L120-L123
23:25 preaction most of the things in here are encased in if statements to make them optional to the situation
23:26 howlinbash fckfreenode, read the first 3 chapters of this book and you will be FLYING! https://git-scm.com/book/en/v2/
23:27 fckfreenode howlinbash: yep, i'll start reading through pro git tomorrow
23:29 howlinbash preaction, it's a handy method. i'll have a play around with it another time. meanwhile, do you know if what i first proposed is possible/advisable. dividing one --bare repo across 2 remote repos on gitlab via 2 different branches. specifically without compromising the private data of the private branch in the public repo
23:30 preaction i don't know anything about gitlab. it sounds doable, but it also sounds terrible, which is why i suggested what i suggested
23:30 freddymungo_ joined #git
23:31 howlinbash preaction, gitlab is just github, no different. by terrible i know you secretly mean brilliant right :-D
23:32 preaction i know what gitlab is, but since i've never used it, i don't know anything else about it. i also don't really care how it works. git was totally designed to have local repos be an amalgam of different remote repos, it's just going to happen that you push the wrong branch to the wrong place. also, you have to somehow sync up the "public" stuff on the "public" branch, which will suck and you'll stop doing it
23:32 howlinbash fckfreenode, as i read those chapters i made a note of the commands i really liked (created aliases - about 30) and i keep that file open while i code for reference. saves time and brain power :-D
23:33 leah2 joined #git
23:34 Spuri0us joined #git
23:34 blokeybloke joined #git
23:34 cdown joined #git
23:35 johnny56 joined #git
23:35 SwiftMatt joined #git
23:36 howlinbash the public stuff doesn't have to be synced so frequently... but in theory i like the idea is cause they will share the commits. i dont have to commit the same file twice. i was hoping that if i tie master with public-remote and private with private-remote then pushing from either branch would automatically only ever send it to the correct/ already connected remote. if not i could just create an alias. but perhaps git
23:36 howlinbash wont work like this
23:37 Kaisyu joined #git
23:37 howlinbash if i what i want is not possible, then its not possible
23:37 howlinbash i'll just keep hacking about, see where it gets me
23:38 preaction they cannot, indeed, share the same commits, at all
23:38 cagomez joined #git
23:38 preaction because as soon as you add a non-public thing, the commits from there on out change (commits refer to a "parent")
23:39 afuentes joined #git
23:40 Spuri0us joined #git
23:40 NeXTSUN joined #git
23:40 dendazen joined #git
23:40 howlinbash aaahh you're right
23:41 howlinbash aw heck
23:42 LeBlaaanc joined #git
23:42 howlinbash left #git
23:43 helmet33 joined #git
23:43 DolpheenDream joined #git
23:46 mischat joined #git
23:49 sofly joined #git
23:50 anuxivm joined #git
23:51 SleePy I'm trying to capture the output of a "git verify-commit" for use in a script (php script actually using shell_exec).  It seems that I can capture the raw output, but the [GNUPG:] lines are not captured.  Why is this?  I even specified --raw and -v
23:55 kutenai How is git branch -m (or -M) different from -f? i.e. "git branch -f test_branch HEAD" moves test_branch to head. What does 'move' d?
23:57 LeBlaaanc joined #git
23:58 Derperperd joined #git
23:58 revoltingPeasant I just did a `git -A; git commit` when I really shouldn't have. I've written a lot of code since my last commit and now I've added a lot of junk files along with it. Is there any way to revover from this blunder without trawling through everything manually or rewriting everything?
23:59 hagridaaron joined #git

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