Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-12-09

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

All times shown according to UTC.

Time Nick Message
00:00 wizardedit joined #git
00:00 wizardedit joined #git
00:01 polyrob_ joined #git
00:05 Gsham joined #git
00:06 jstimm joined #git
00:06 railssmith joined #git
00:09 mischat joined #git
00:10 howitdo joined #git
00:12 nettoweb joined #git
00:12 macivy joined #git
00:13 fbrSohbet-278 joined #git
00:13 fbrSohbet-278 left #git
00:14 kfoonamalik joined #git
00:14 askb joined #git
00:14 mellernoia joined #git
00:14 LaydGirL joined #git
00:15 macivy I'm getting this branch missing error, never seen this before. Any clues? Here are the details: http://pastebin.com/8t4dQvtp
00:16 madewokherd joined #git
00:17 King_Hual joined #git
00:18 King_Hual joined #git
00:19 thiago macivy: what git command did you run and what was the exact error message?
00:19 Lipp joined #git
00:20 Shentino joined #git
00:21 venmx joined #git
00:24 oaaao joined #git
00:24 thiago macivy: still around?
00:28 King_Hual joined #git
00:29 Lipp joined #git
00:29 Derderderd joined #git
00:30 thiago assuming the problem has been solved
00:31 MrSamuel joined #git
00:33 MrSamuel I've got a repo which had a private key, and now I want to put it on github. I've rebased and removed the offending files, but when ?I run git grep I still see some commits with the private key. However, git branch --contains is empty
00:33 MrSamuel What else can I check?
00:33 Lipp joined #git
00:33 NeverDie joined #git
00:34 guest____ joined #git
00:36 cyan__ joined #git
00:36 thiago MrSamuel: git grep doesn't search commits
00:36 MrSamuel thiago, what does it search?
00:36 thiago the current commit only
00:36 thiago so you found files in the current commit
00:36 MrSamuel % git grep "PRIVATE KEY" $(git rev-list --all)
00:37 MrSamuel I'm running that
00:37 thiago ha, interesting
00:37 thiago I didn't know it could do that
00:37 kadoban MrSamuel: You should use bfs or filter-branch to remove the troublesome data from history
00:37 thiago anyway, find the oldest of those commits and rebase to its parent
00:37 thiago remove the file there
00:38 MrSamuel kadoban, it's removed, it's no longer in the master branch
00:38 MrSamuel but the data is still showing up in some kind of dangling commmit
00:38 MrSamuel I tried clearing the reflog too
00:38 MrSamuel then git fsck and git gc git rebase etc
00:39 guest____ joined #git
00:40 MrSamuel okay nailed it
00:40 MrSamuel .git/info/refs or something similar was still pointing at the old commit head
00:40 MrSamuel i think it's because i didn't update one of the remotes yet.. odd
00:40 thiago use git rev-list master
00:40 MrSamuel git rev-list master also would not show anything
00:40 thiago is master your only branch?
00:41 MrSamuel yes
00:41 MrSamuel but I need to be sure there are no private keys :D
00:41 MrSamuel ANYWHERE :)
00:41 thiago ok, so push to your remote to update it, then try your command again
00:41 Motoservo joined #git
00:41 MrSamuel Haha.. yeah, great idea.. push the private keys by accident :)
00:41 Naan joined #git
00:42 thiago push to that remote that currently already has the key
00:42 thiago you can't make it worse, only better
00:42 MrSamuel it's not that simple
00:42 MrSamuel I'm already deploying to a private server and I don't care what goes there
00:42 irqq joined #git
00:42 MrSamuel but I also had github as a remote and I think it was doing something weird
00:42 MrSamuel anyway, it's all sorted
00:42 thiago ok
00:47 Shentino joined #git
00:47 King_Hual joined #git
00:48 nettoweb joined #git
00:50 cqi joined #git
00:50 buffal0 joined #git
00:52 kfoonamalik joined #git
00:53 guest_____ joined #git
00:54 Motoservo joined #git
01:03 arescorpio joined #git
01:08 einnjo joined #git
01:10 mischat joined #git
01:19 shgysk8zer0 joined #git
01:20 tmsmith joined #git
01:28 italoacasas joined #git
01:33 zerozero joined #git
01:34 Nilesh_ joined #git
01:34 Lipp joined #git
01:34 King_Hual joined #git
01:40 wxg joined #git
01:41 qpdb joined #git
01:43 sdothum joined #git
01:47 tnecniv joined #git
01:48 baryon joined #git
01:55 peeps[lappy] hi, i have a machine running windows that I would like to use for editing source code, however the build process for this code is run in a Linux VM on the same machine.  I guess what I'd like to do is have a shared folder on the VM client and use a text editor directly on those files from the host machine
01:56 peeps[lappy] just wondering if this is likely to cause any problems with the git structure or something
01:56 jumpman joined #git
01:57 crose joined #git
01:57 peeps[lappy] or if anyone is doing something similar.  otherwise i would need to run my own git server from the host, that the VM client would update from, which seems complicated
01:57 Sleepy63 joined #git
01:59 Eugene peeps[lappy] - I found VMware's Shared Folders feature does not work well with git, but your mileage may vary.
01:59 peeps[lappy] :(
02:00 peeps[lappy] Eugene, in what way?
02:00 Eugene It corrupted my repo and ate my lunch
02:00 peeps[lappy] nooooo
02:01 peeps[lappy] what about a not VMware shared folder but editing over something like samba or NFS
02:01 Eugene I have separate clones of my repos on my desktop and my SSH host, depending on what I'm working on. It generally works fine to push/pull via an external git server
02:01 shinnya joined #git
02:01 Eugene git is decentralized, embrace it
02:02 peeps[lappy] i guess i don't like to push to origin until i've tested my build
02:02 Eugene That's what feature branches are for ;-)
02:03 Eugene I would keep the git repo entirely on the Linux VM, and use Windows tools to interact/edit files
02:03 Ardethian joined #git
02:03 Celelibi joined #git
02:03 peeps[lappy] um, how would you do that without a shared folder?
02:03 rewt you can push between the host/guest local repos without going through an external repo
02:03 Eugene A simple Samba share that's mounted on the Host should do it
02:04 peeps[lappy] i'm not sure i would put any faith in samba more than vmware shared folders
02:05 peeps[lappy] its always seemed incredibly buggy and insanely slow to me
02:05 thiago peeps[lappy]: ALWAYS use git on local disks only
02:05 thiago never shared via NFS, SMB or Virtual machines
02:05 thiago if you're using a virtual machine, run git only on the host, never the guest
02:06 peeps[lappy] why?
02:06 thiago because remote FS are known to corrupt Git databases
02:08 peeps[lappy] ok, but that last point at least doesn't make sense.  as if git can't be used in a vm at all?
02:08 peeps[lappy] seems extreme
02:08 Eugene Not inherently: It's the un-flushed writes, synchronization of key files, and (lack of) locking that gets you
02:08 King_Hual joined #git
02:09 Eugene git can be used in a VM just fine, its the non-local filesystem that doesn't behave like git assumes a filesystem ought to
02:09 Eugene If you write a file it should be written. That isn't true with (very many) shared filesystems
02:10 mischat joined #git
02:11 thiago peeps[lappy]: like Eugene said, you can use git in the guest. On a local disk, not a shared folder.
02:11 ResidentBiscuit joined #git
02:14 kbs left #git
02:21 fahadash joined #git
02:24 l_l joined #git
02:26 l_l Hey guys! How do I commit now with a date in the past?
02:27 LordRyan joined #git
02:27 phroa --date="date" : man git-commit
02:27 gitinfo the git-commit manpage is available at http://jk.gs/git-commit.html
02:27 LordRyan When trying to sign commits, I have to use `gpg --sign > /dev/null` to unlock the key before I'm able to sign the commit. I've tried the top answer http://superuser.com/questions/520980/how-to-force-gpg-to-use-console-mode-pinentry-to-prompt-for-passwords by gra\wity but it doesn't seem to be working.
02:28 Gsham joined #git
02:29 inflames joined #git
02:29 l_l phroa: Would I have to take care of any other dates like the file timestamp? This sets only the GIT_COMMITTER_DATE right?
02:29 phroa I don't think git records file c/a/mtime
02:30 phroa if it does, --date almost certainly wouldn't touch them.
02:30 menip joined #git
02:30 l_l I see. Alternatively if someone was to set the commit date using --date, is there any way to find the actual commit date?
02:30 jstimm joined #git
02:31 phroa well, you could look at server logs to see when they pushed it, but... no
02:31 Ryanar joined #git
02:32 l_l Pushing doesn't really tell me anything if a bunch of commits were pushed together, right? Although I think this suffices.
02:32 phroa it would be kinda useless to be able to change the commit date if there was a "real" commit date people could look at instead
02:35 finalbeta joined #git
02:36 l_l phroa: I get that though things aren't always so 'easy'. Something almost always comes up in sleuthing.
02:36 f4 joined #git
02:36 l_l phroa: Another question. Is it possible to commit to a date before your last commit date?
02:37 phroa there's tons of commits floating around github from the seventies and eighties.  sure!
02:38 l_l So how do conflicts get resolved?
02:38 phroa just based on file contents, keeping in mind any merge options/strategies you might have specified beforehand
02:38 l_l I could change files heavily resulting in conflicts and commit it to previous date than the last commit such that the latest commit doesn't make sense?
02:40 phroa you wouldn't get a conflict unless you're merging or rebasing; in your case, the "bad" commit wouldn't conflict unless you inserted it before the "good" commit in the git history, which would mean !rewrite ing history in general
02:40 gitinfo Rewriting public history is not recommended. Everyone who has pulled the old history will have to do work (and you'll have to tell them to), so it's infinitely better to just move on. If you must, you can use `git push --force-with-lease <remote> <branch>` to force (and the remote may reject that, anyway). See http://goo.gl/waqum
02:40 BlueProtoman joined #git
02:40 l_l So if A is the last commit and B is the commit(the actual latest commit) whose date I want to change such that history is now B->A, would this create a merge commit like B->A->C? or does A's commit get modified?
02:41 brent__ joined #git
02:41 BlueProtoman Is there a good way to set a file's access permissions on a file system that doesn't support Unix-style permissions (e.g. NTFS)?
02:47 l_l Ah. I see. So basically --date is only recommended if it to set a commit back in time after the last commit? Otherwise it's rewriting history?
02:47 phroa BlueProtoman: git doesn't actually store permissions at all except for 644 (normal file) and 755 (executable or directory).  you could use a post-checkout hook to set permissions after you check something out, but that would mean trusting the hook to do its job and not get tampered with
02:47 umpc joined #git
02:47 phroa l_l: "history" and "git history" aren't really the same thing; git history is the sequence of commits.  the particular date each commit was made doesn't really matter.
02:47 l_l So if I was to make the commit with the previous date would it still register as A->B? Then what's the point of setting the date since one can easily find out that B occured after A and not when it's specified?
02:47 BlueProtoman phroa: That's fine, I need to store the fact that a file is executable
02:47 thiago l_l: why do you set a name? why do you write a message?
02:47 thiago l_l: same answer as the date
02:49 ilbot3 joined #git
02:49 Topic for #git is now Welcome to #git, the place for git help and serious business | Public logs at http://jk.gs/glog | First visit? Read: http://jk.gs/git | Current stable version: 2.11.0 | Getting "cannot send to channel"? /msg gitinfo .voice | The git-jokes project has been suspended for lack of contribution. Help revive it! https://madeitwor.se/git-jokes
02:49 phroa if you want, I guess
02:50 l_l phroa: Something like this git commit --amend --date="Fri Nov 6 20:00:00 2015 -0600" followed by git commit --date="Sat Nov 7 20:00:00 2015 -0600"?
02:52 phroa ok
02:56 l_l phroa: Okay thanks.
02:58 SwiftMatt joined #git
02:58 JamesBaxter joined #git
02:59 realz joined #git
03:00 ace_33 joined #git
03:00 Siegfried joined #git
03:00 franks-reich joined #git
03:01 pks joined #git
03:03 aavrug joined #git
03:07 robics joined #git
03:08 meauses joined #git
03:08 dodobrain joined #git
03:11 mischat joined #git
03:13 einnjo joined #git
03:13 _Patrick_ joined #git
03:13 jaguarmagenta joined #git
03:14 albel727 joined #git
03:15 breezy_ joined #git
03:15 LordRyan I hate to post again, but:
03:15 LordRyan When trying to sign commits, I have to use `gpg --sign > /dev/null` to unlock the key before I'm able to sign the commit. I've tried the top answer http://superuser.com/questions/520980/how-to-force-gpg-to-use-console-mode-pinentry-to-prompt-for-passwords by gra\wity but it doesn't seem to be working.
03:18 dreiss joined #git
03:22 kfoonamalik joined #git
03:25 meauses joined #git
03:26 chachasmooth joined #git
03:26 Motoservo joined #git
03:36 sdothum joined #git
03:36 thiago LordRyan: works here
03:36 thiago but I had to select a key
03:36 LordRyan Right, that's normal. Mine's failing though and I'm not sure how to debug why.
03:36 LordRyan I'm going to switch over to GNOME and see if I can reproduce the behavior.
03:38 LordRyan https://i.imgur.com/PMcowKK.png nope, still fails.
03:38 LordRyan so it's not i3 causing it.
03:38 Rodya_ joined #git
03:38 Motoservo joined #git
03:39 sunri5e joined #git
03:39 LordRyan I actually don't see a 'gpg-agent.conf' under https://www.gnupg.org/documentation/manuals/gnupg/GPG-Configuration.html
03:41 venmx joined #git
03:42 ergzay joined #git
03:42 ergzay Hi I'm trying to install git from source and it keeps failing to make the docs
03:42 ergzay i hit xmlto and it fails
03:42 LordRyan so, https://i.imgur.com/oqtW38t.png is what happens when I try `gpg --sign > /dev/null`
03:42 ergzay It tries to fetch some file "docbook.xsl" from the internet and then fails to access it
03:43 LordRyan thiago: ^^
03:44 blunaxela i have a directory in my repo that i'd like to make a copy of, but i'd like to copy/keep the history for those new files. What would be the most sane way to do that? I'm thinking git-subtree-split to get the directory history and then replay those commits on the new directory.
03:44 BUSY joined #git
03:46 Orion3k joined #git
03:47 hendry i accidentally committed a bunch of files into my home git directory sometime ago. Do I just rebase and omit the commit to remove them ?
03:48 phroa or, !bfg
03:48 gitinfo A tool designed to remove large files, or passwords from history: https://rtyley.github.io/bfg-repo-cleaner/ (!rewrite applies)
03:48 ergzay hendry: http://stackoverflow.com/questions/9529078/how-do-i-use-git-reset-hard-head-to-revert-to-a-previous-commit
03:49 ergzay hendry: oh it wasn't most recent commit? then yeah i'd just rebase, as long as you're the only one working on that git repository
03:49 venmx joined #git
03:49 thiago LordRyan: strace -f the commit command. Does it even connect to the agent socket?
03:49 ergzay interactive rebase and delete the commit
03:50 ergzay and then force push to make sure the remote is up to date (warning this will mess with pulls from all repos)
03:51 LordRyan thiago: it works but it doesn't prompt me for my password
03:51 LordRyan so i would assume so
03:51 Goplat joined #git
03:52 tnecniv joined #git
03:54 hexagoxel_ joined #git
03:58 JamesBaxter joined #git
03:59 ploop joined #git
03:59 davimore__ joined #git
04:02 moei joined #git
04:03 thiago LordRyan: it connects to the agent? reading what the conversation with the agent is, you'll have to determine why it refused to sign
04:03 blackwind_123 joined #git
04:04 l_l left #git
04:12 mischat joined #git
04:14 ploop joined #git
04:15 Vortex34 joined #git
04:16 Andrew_K joined #git
04:21 railssmith joined #git
04:22 BUSY joined #git
04:24 fstd_ joined #git
04:24 novas0x2a|laptop joined #git
04:24 aw1 joined #git
04:28 robotroll joined #git
04:28 raspado joined #git
04:30 thesha joined #git
04:33 red82 joined #git
04:33 cdg joined #git
04:34 diogenese joined #git
04:34 buffal0 joined #git
04:38 texinwien_ joined #git
04:39 buffffal00 joined #git
04:39 cdg joined #git
04:39 subhojit777 joined #git
04:40 ayogi joined #git
04:40 r4nd0m0ne joined #git
04:41 sbulage joined #git
04:41 g5pw joined #git
04:43 Andrew_K joined #git
04:49 thesha joined #git
04:49 svm_invictvs joined #git
04:52 realz joined #git
04:52 mar77i joined #git
04:59 fahadash joined #git
04:59 BasdeKort[m] joined #git
04:59 imgglasses joined #git
04:59 imgglasses I didn't know of this channel. I'm tracking with git a folder named `sunflower` and inside the folder there's `locale` and `girasol`. Ok. When I push this to github obviously `locale` and `girasol` are inside `sunflower`. How can I move them up when pushing to github?
04:59 arnsa joined #git
05:00 diogenese joined #git
05:03 GT4066 joined #git
05:06 diogenese joined #git
05:08 blunaxela imgglasses: Are you saying that you want those two files/directories to be moved up on the github repo and remain in the same location locally?
05:09 imgglasses blunaxela: yes, that's what I wanted. But I deleted my old .git and started over the way I wanted it. Thanks anyway.
05:09 imgglasses blunaxela: out of curiosity, is there a way to do that?
05:10 meauses joined #git
05:11 askb joined #git
05:12 donught joined #git
05:14 BasdeKort[m] left #git
05:15 ahmedelgabri joined #git
05:15 Dragonbird joined #git
05:17 Sample joined #git
05:20 railssmith joined #git
05:21 acetakwas joined #git
05:24 Motoservo_ joined #git
05:29 irqq joined #git
05:29 menip joined #git
05:30 SwiftMatt joined #git
05:30 thesha joined #git
05:31 bmahe joined #git
05:31 lindenle joined #git
05:31 blunaxela imgglasses: well, git is a verstion control system, so what you're describing wouldn't make much sense to me, unless you move them up, commit them, and then move them back
05:32 pks joined #git
05:36 howitdo joined #git
05:36 dsantiago joined #git
05:40 navidr joined #git
05:46 menip joined #git
05:49 howitdo joined #git
05:51 svm_invictvs joined #git
05:53 Cabanossi joined #git
05:54 cdg_ joined #git
05:56 _ikke_ imgglasses is correct, there is no way to do this in transit
05:56 brillenfux joined #git
05:56 imgglasses blunaxela, _ikke_: thanks a lot! I'm a real beginner with git
05:56 mablae_ joined #git
05:56 pks joined #git
05:57 Limes_ joined #git
05:58 ChiLL_SurF joined #git
05:59 mentazoom joined #git
06:02 breezy_ joined #git
06:02 xberg joined #git
06:05 venmx joined #git
06:07 blunaxela imgglasses: you're welcome, and have fun with it!
06:14 mischat joined #git
06:14 rscata joined #git
06:17 OnBrokenWings joined #git
06:18 mar77i_ joined #git
06:20 ogny joined #git
06:23 OnBrokenWings joined #git
06:24 jrnieder joined #git
06:25 overlord_tm joined #git
06:27 sbulage joined #git
06:28 subhojit777 joined #git
06:28 durham joined #git
06:29 Furai joined #git
06:29 _leb joined #git
06:38 debeaure joined #git
06:40 davimore_ joined #git
06:45 m___ joined #git
06:46 thierryp joined #git
06:50 shinnya joined #git
06:51 jrnieder joined #git
06:51 ThomasLocke joined #git
06:51 bocaneri joined #git
06:52 zerorax joined #git
06:53 zerorax hi, i looked in the man page but i don't see this
06:54 zerorax is there a way to supply password and username in the git push command? I want to make a script that pushes my project
06:54 phroa use an ssh key
06:54 zerorax ok
06:56 Dragonbird joined #git
06:58 howdoi joined #git
06:59 Raging_Hog joined #git
07:00 Puffball joined #git
07:00 dermoth joined #git
07:02 qt-x joined #git
07:04 cyan__ joined #git
07:08 kfoonamalik joined #git
07:13 Lope joined #git
07:14 Lope sometimes I'm busy with editing a bunch of files as part of a big commit. then I do some little thing that may involve some of those files. and would like to do the small change as it's own commit.
07:14 Lope How can I add only certain changed lines to my new commit?
07:14 thiago git add -p
07:14 Lope Is there an easy way to do it? like a GUI?
07:15 Lope or some interactive CLI interface at least.
07:15 mischat joined #git
07:15 Lope thiago: thanks, will look
07:15 durham joined #git
07:15 Lope Oh, I see it's interactive.
07:16 Lope Ok I'll give it a try :)
07:17 wrouesnel1 joined #git
07:17 realtime joined #git
07:17 JeroenT joined #git
07:17 diogenese joined #git
07:18 Lope Great, now how can I see the changes that I've staged?
07:19 freimatz joined #git
07:20 SwiftMatt joined #git
07:20 thiago git diff --cached
07:20 elect_ joined #git
07:20 Lipp joined #git
07:20 thief joined #git
07:24 Lope thanks will try it.
07:25 Lope with `git add -p foo` the one hunk it's offering me has some lines that I want, and some lines that I don't want in ONE hunk. How can I get the part I want?
07:25 thiago split it
07:25 Lope it's offering me Stage this hunk [y,n,q,a,d,/,K,g,s,e,?]
07:25 jrnieder joined #git
07:25 Lope how can I split it?
07:25 thiago press ? and read
07:25 Lope thanks :)
07:26 GodGinrai joined #git
07:27 Lope awesome :) thanks for the help my friend!
07:27 thiago you're welcome!
07:30 iamtakingiteasy joined #git
07:30 wrouesnel1 joined #git
07:32 muelli joined #git
07:32 quetz joined #git
07:35 lordjancso joined #git
07:35 jknetl joined #git
07:37 Sceorem joined #git
07:40 arnsa joined #git
07:43 jnavila joined #git
07:44 JeroenT joined #git
07:45 rominronin joined #git
07:49 venmx joined #git
07:51 Lope I tried to edit a hunk manually. I cut out the + and - lines I don't want, then pressed Ctrl X. Then it said: Error: file.txt patch does not apply
07:51 Lope What does that mean?
07:52 jrnieder joined #git
07:54 JeroenT joined #git
07:56 mmattice joined #git
07:56 Darcidride joined #git
07:58 davimore__ joined #git
07:59 Balliad joined #git
07:59 brk_ joined #git
08:01 mischat joined #git
08:02 m___ joined #git
08:05 khfeng joined #git
08:05 johnny56 joined #git
08:06 Sceorem joined #git
08:07 Sample joined #git
08:09 thierryp joined #git
08:09 eduardas_m joined #git
08:10 derk0pf joined #git
08:12 tarkus joined #git
08:13 tarkus On "merge -s subtree" step in this tutorial I always get "refusing to merge unrelated histories". Is something wrong in the tutorial? https://git-scm.com/book/en/v1/Git-Tools-Subtree-Merging
08:17 howitdo joined #git
08:18 blunaxela joined #git
08:20 Lipp joined #git
08:20 pdgwien joined #git
08:20 zeroed joined #git
08:20 Alienpruts joined #git
08:23 usm4n joined #git
08:26 derk0pf joined #git
08:28 Limes_ joined #git
08:30 chele joined #git
08:30 raspado joined #git
08:31 ahmedelgabri joined #git
08:32 Balliad joined #git
08:37 jnewt joined #git
08:38 OOPMan joined #git
08:40 jaguarmagenta joined #git
08:41 acetakwas joined #git
08:43 m___ joined #git
08:46 dan2k3k4 joined #git
08:47 t0by joined #git
08:51 Atm0spher1c joined #git
08:52 Sonderblade joined #git
08:54 flaviodesousa joined #git
08:57 SpeakerToMeat joined #git
08:58 SaberArthur joined #git
08:59 thierryp joined #git
09:00 re1 joined #git
09:01 ahmed_elgabri joined #git
09:03 Learath2 joined #git
09:04 kurkale6ka joined #git
09:05 SwiftMatt joined #git
09:06 willlondon joined #git
09:06 alcohol joined #git
09:09 ahmedelgabri joined #git
09:10 JuliusN joined #git
09:11 apeiros joined #git
09:12 apeiros hi. I can do `git --since='3 days ago' -- some/path` to see all commits which changed files in a given path. but is there a way to see the diffs too? right now I manually `git show SHA -- some/path`
09:14 mindfart joined #git
09:17 Sasazuka joined #git
09:17 davimore_ joined #git
09:19 mikecmpbll joined #git
09:20 encod3 joined #git
09:20 overlord_tm joined #git
09:21 rominronin joined #git
09:21 O47m341 joined #git
09:23 ElAngelo joined #git
09:23 systemovich joined #git
09:23 ElAngelo hi, trying something with git submodules
09:23 ElAngelo i have a 'master' repo
09:23 ElAngelo and 3 submodules
09:24 ElAngelo in there
09:24 ElAngelo i want the master - develop branch to track the develop branches of the submodules
09:24 ElAngelo and when i release version 1.0 i want that master release to track the version 1.0 of all of those submodules
09:25 ElAngelo i'm not entirely sure if that is a scenario that is supported by git?
09:25 marcogmonteiro joined #git
09:25 Lipp joined #git
09:27 tvw joined #git
09:27 UncleCJ[m] left #git
09:27 thierryp joined #git
09:28 Sample joined #git
09:28 grantwu left #git
09:29 SaberArthur Hey, can I install more linux tools on linux bash?
09:30 rom1504 yes
09:30 SaberArthur how? there is no apt-get or anything
09:30 SaberArthur I mean git bash sorry
09:30 SaberArthur I install it on windwos
09:30 rom1504 export PATH=/some/dir/where/you/put/binaries:$PATH
09:30 ElAngelo it's just cygwin
09:30 rom1504 ah
09:30 mentazoom joined #git
09:30 SaberArthur Yeah, cygwin.. So I can't?
09:31 ElAngelo mingw i mean
09:31 ElAngelo sure you can... but chances are you completely break it
09:31 JeroenT_ joined #git
09:31 SaberArthur Well. that's not good
09:31 ElAngelo if you want more tools, just install mingw seperatly
09:31 SaberArthur I thought mingw has special repos or something
09:31 osse SaberArthur: you can't. git for windows is a bundle of tools and that's it.  if you get MSYS2 (which it's based on)  you can do whatever you want
09:32 SaberArthur They have vim and everything tho :/
09:32 osse and you can install git for windows inside there
09:32 JeroenT joined #git
09:32 SaberArthur okay, thanks
09:33 lindenle joined #git
09:33 ElAngelo msys2.github.io
09:33 ElAngelo or cygwin
09:33 chll_ joined #git
09:33 ElAngelo or mingw
09:33 _ADN_ joined #git
09:33 ElAngelo or .... LSW
09:33 ElAngelo which would be my preferred way
09:33 regedit joined #git
09:33 SaberArthur anyone use it?
09:34 SaberArthur or do you get enough tools just from git windows installer?
09:34 * ElAngelo is using it
09:34 SaberArthur msys2?
09:34 ElAngelo no LSW
09:34 ElAngelo don't see why you would bother with all the recompiled stuff if you can just run lsw
09:35 osse I use this: https://msys2.github.io/ - works just fine
09:35 SaberArthur thanks
09:36 JeroenT_ joined #git
09:36 osse how come I get no relevant results when I google lsw
09:36 ElAngelo linux subsystem for windows
09:36 ElAngelo wasn't that the official name?
09:37 ElAngelo oh shit
09:37 ElAngelo it's WSL
09:37 ElAngelo windows subsystem for linux
09:37 SaberArthur haha, I don't get it at first either
09:37 ElAngelo my bad, i always reverse them letters
09:37 osse are you french by any chance? :p
09:37 ElAngelo not at all
09:38 SaberArthur wsl is the bash on win10?
09:38 ElAngelo yes
09:38 mazert joined #git
09:39 jeffreylevesque joined #git
09:39 ElAngelo just realized i assumed you were running windows10
09:39 SaberArthur I still use win 7
09:39 SaberArthur I guess it is msys2 for me, then
09:40 kedare joined #git
09:43 jmpp_ joined #git
09:43 jaguarmagenta joined #git
09:45 osse ElAngelo: does wsl have a compiler ?
09:45 ElAngelo it's just ubuntu
09:45 ElAngelo apt install gcc
09:45 ElAngelo apt install whatever you like
09:45 ElAngelo some stuff doesn't work
09:46 ElAngelo but most of it is just working
09:46 xaa joined #git
09:46 ElAngelo obviously nothing that requires hardware access... like controlling wifi or something
09:46 cyberz_ joined #git
09:46 mahendratech joined #git
09:47 nettoweb joined #git
09:48 rorro joined #git
09:51 aleray joined #git
09:51 aleray hi, is there a way to get the hash of an git notes object?
09:51 zeroed joined #git
09:52 rnsanchez joined #git
09:53 baryon joined #git
09:57 satya4ever joined #git
09:58 ams__ joined #git
10:03 MrMojito joined #git
10:08 manuel joined #git
10:08 bmahe joined #git
10:08 vaesper joined #git
10:11 WakiMiko_ joined #git
10:14 muelli joined #git
10:15 sbulage joined #git
10:15 qt-x joined #git
10:19 irqq joined #git
10:20 arnsa joined #git
10:22 SwiftMatt joined #git
10:22 mablae_ joined #git
10:23 RxMcDonald joined #git
10:24 RxMcDonald joined #git
10:24 RxMcDonald Hello, anyone knows if it's possible to track a file in a different folder?
10:24 RxMcDonald i.e. a folder outside the one containing the .git file
10:24 _ikke_ git does not track files outside of the repo root di
10:24 _ikke_ dir
10:24 moritz nope
10:24 moritz but you can have a file inside it, and from the outside set a symlink to it
10:24 systemovich joined #git
10:25 nd joined #git
10:25 moritz it would be a huge security risk if cloning a repo could write a file outside of the repo
10:25 RxMcDonald what do you mean?
10:25 moritz for example your ~/.ssh/authorized_keys
10:25 _ikke_ move file to repo, then symlink the file from the original location to the repo
10:25 RxMcDonald yea but I don't really care I just want to track my emacs config and it's in my come dir and I don't want to make it a git repo
10:26 RxMcDonald oh yea that sounds good
10:26 RxMcDonald my come dir lmao, home*
10:28 JeroenT joined #git
10:29 JeroenT joined #git
10:30 netj joined #git
10:30 moritz there are also project that specialize in tracking dotfiles in git repos
10:31 afuentes joined #git
10:33 xall joined #git
10:33 RxMcDonald well looks like it didn't work?
10:34 RxMcDonald added the symlink in my mac but it doesn't get read by the program
10:34 moritz can 'cat' or 'head' read it?
10:35 RxMcDonald cat says too many symlinks
10:35 RxMcDonald head as well
10:35 jaguarmagenta joined #git
10:36 RxMcDonald it's a mac tho
10:37 systemovich joined #git
10:37 venmx joined #git
10:38 RxMcDonald nvm had the wrong link
10:39 msonntag joined #git
10:41 processkiller joined #git
10:42 _JoySpring_ joined #git
10:43 thierryp joined #git
10:43 O47m341 joined #git
10:44 Oatmeal joined #git
10:49 JeroenT joined #git
10:49 ismithers joined #git
10:51 mischat joined #git
10:51 Ediz joined #git
10:51 Lipp joined #git
10:52 cyphase_eviltwin joined #git
10:53 Learath2 joined #git
10:56 meLon joined #git
10:57 rominronin joined #git
10:58 rominron_ joined #git
11:02 rorro joined #git
11:02 rchavik joined #git
11:04 GavinMagnus joined #git
11:09 jaguarmagenta joined #git
11:10 King_Hual joined #git
11:13 GavinMagnus left #git
11:17 thierryp joined #git
11:18 jeffreylevesque joined #git
11:18 brillenfux_ joined #git
11:19 bejeb joined #git
11:19 brillenfux__ joined #git
11:22 star_prone joined #git
11:23 brillenfux_ joined #git
11:28 achadwick joined #git
11:32 Snugglebash joined #git
11:32 re1_ joined #git
11:34 aleray How do I find the hash reference of a `git notes` note object?
11:36 rudi_s aleray: git notes list $commit
11:36 dan2k3k4 joined #git
11:37 JeroenT joined #git
11:38 [0xAA] joined #git
11:40 buffal0 joined #git
11:41 cqi joined #git
11:42 dsdeiz joined #git
11:42 dsdeiz joined #git
11:42 buffffal00 joined #git
11:44 zerozero joined #git
11:45 deadnull joined #git
11:45 AndrewAlexMac joined #git
11:45 ferr1 joined #git
11:47 deadnull joined #git
11:48 [0xAA] joined #git
11:48 rads joined #git
11:51 courrier joined #git
11:52 anth0ny_ joined #git
11:52 aleray rudi_s, thanks!
11:58 buffal0 joined #git
11:59 m___ joined #git
12:02 svm_invictvs joined #git
12:06 JeroenT joined #git
12:09 howitdo joined #git
12:09 howitdo joined #git
12:14 YuGiOhJCJ joined #git
12:15 Snugglebash joined #git
12:16 fahadash joined #git
12:17 leeN joined #git
12:18 mazert joined #git
12:19 Jackneill joined #git
12:19 Darcidride joined #git
12:21 holodoc joined #git
12:22 dan2k3k4 joined #git
12:22 sbeller joined #git
12:24 JuliusN if i want to gitignore all files in directory '.idea/', is the correct .gitignore line '.idea/*'?
12:24 TikityTik joined #git
12:25 TikityTik Lets say I'm working on a file away from git working directory because virtual machine, then do a git pull where the file I was working on changed.
12:25 TikityTik How do I merge the files properly?
12:27 moritz TikityTik: have you committed the local changes?
12:27 Lope joined #git
12:27 TikityTik moritz: not yet, so how do i do this merge
12:27 moritz TikityTik: you can first commit, then pull; or git stash, pull, git stash apply; git stash drop
12:28 TikityTik moritz: You don't understand. My git working directory is in a VM. My work is being copied from the host machine
12:29 TikityTik how can I merge changes from a pasted file?
12:29 moritz TikityTik: are you looking for git format-patch and git am?
12:29 moritz or git diff and git apply?
12:30 TikityTik moritz: I already did a git pull, I want to paste my changes
12:30 moritz saying that something is in a VM provides zero information; a VM is just a regular OS instance
12:30 moritz TikityTik: paste where?
12:31 moritz also https://git-scm.com/book/en/v2/Git-Basics-Working-with-Remotes
12:31 TikityTik it's in /mnt/file.tex from a shared VM folder, and my git is in ~/the-git
12:31 iliv joined #git
12:31 TikityTik I want to bring what I have from /mnt/file.tex <-(my own work off an older commit) and bring my changes to ~/the-git <-(just pulled)
12:32 zeroed joined #git
12:32 TikityTik ~/the-git is ahead by 1 commit compared to my /mnt/file.tex
12:32 raspado joined #git
12:32 TikityTik do you know what I mean?
12:32 moritz TikityTik: and is /mnt/ a git repo?
12:32 TikityTik moritz: No, it's not connected with git.
12:32 moritz TikityTik: use cp
12:33 TikityTik moritz: But if I paste my changes over a newest commit, doesn't that remove the changes that were just pushed?
12:33 howdoi joined #git
12:33 moritz TikityTik: pushed where? why "paste"?
12:33 moritz TikityTik: somehow everything you say seems to be very vague
12:34 moritz TikityTik: if you want help, please try some more precise terminology
12:34 moritz so, you have a git repo, in ~/the-git
12:34 moritz that one has a new commit
12:34 moritz and then you have /mnt/file.tex
12:34 TikityTik moritz: Yes. And it's 1 commit ahead
12:34 moritz TikityTik: what is 1 commit ahead of what?
12:34 TikityTik moritz: ~/the-git is 1 commit ahead compared to /mnt/file.tex
12:35 moritz TikityTik: but you said /mnt/file.tex isn't part of a git repo
12:35 TikityTik moritz: ~/the-git just pulled and is now 1 commit ahead.
12:35 moritz TikityTik: how can it be "1 commit ahead"?
12:35 TikityTik sorry
12:35 moritz or behind, in that case
12:35 TikityTik renaming ~/the-git to ~/git
12:35 TikityTik moritz: ~/git/file.tex just got pulled and is 1 commit ahead of /mnt/file.tex
12:36 gemie joined #git
12:36 moritz TikityTik: so, cp ~/git/file.tex /mnt/file.tex
12:36 gemie hi, I just learned git add -p, but I am missing something. I want to "discard" a chunk, meaning delete this. Is it possible?
12:36 TikityTik moritz: /mnt/file.tex is 1 commit behind, but that was where I was working.
12:36 TikityTik moritz: I will lose my work
12:36 moritz TikityTik: ah, so it has local modification too
12:37 TikityTik moritz: Yes
12:37 moritz gemie: git checkout -p discards hunks from the working copy
12:37 nettoweb joined #git
12:37 moritz TikityTik: so, make /mnt a proper git repo. Commit your files in there; pull from ~/git/
12:37 moritz TikityTik: or in ~/git/ make a branch with the changes from /mnt; commit, merge, then copy the result back
12:38 TikityTik moritz: it's too late :(
12:38 TikityTik for the 1st method
12:38 moritz why?
12:38 TikityTik moritz: the 2nd method is preferable, thank you
12:38 moritz cp -r ~/git/.git /mnt/
12:39 TikityTik moritz: Isn't that the same as pasting over my changes?
12:40 TikityTik I mean pasting over the newest commit's changes
12:40 moritz no
12:41 moritz because copying .git doesn't change file.tex
12:41 sdothum joined #git
12:42 Outsider__ joined #git
12:42 iliv joined #git
12:44 tobie joined #git
12:46 jnewt joined #git
12:49 tarkus joined #git
12:50 tarkus joined #git
12:51 evanc joined #git
12:52 SaberArthur joined #git
12:53 zero7 joined #git
12:54 Limona joined #git
12:55 Limona hello
12:55 jaguarmagenta joined #git
12:55 Limona got a question
12:55 Gsham joined #git
12:55 aruns joined #git
12:55 Limona how does a typical workflow look like
12:56 Limona who do i have to commit before resolving merges?
12:57 _ikke_ Limona: What do you mean with your last question?
12:57 Limona i got a file
12:57 trig-ger joined #git
12:57 Limona i mean i pulled from master
12:57 Limona and now i want to resolve conflicts
12:57 Limona but i have to merge? first?
12:57 _ikke_ Limona: pull = fetch + merge
12:57 Limona ahh
12:57 _ikke_ It's the merge part where the conflict happened
12:58 Limona why didn't it want to resolve then
12:58 johnmilton joined #git
12:58 _ikke_ what do you mean, didn't want to resolve?
12:58 Limona why did i have to commit
12:58 Limona before resolving
12:58 Limona the merge
12:58 _ikke_ You have to commit *after* you resolved the conflict
12:58 Limona that's the confusing part
12:58 Limona it didn't let me do that
12:58 Limona untill i've commited changes
12:59 Limona :/
12:59 Limona they were staged, just not commited
12:59 synthroid joined #git
12:59 _ikke_ a merge normally creates a commit
12:59 _ikke_ when a conflict happens, git does not create the commit yet
12:59 Limona i had to commit before making a merge?
13:00 _ikke_ Limona: The commit finishes the merge
13:00 Limona for some reason
13:00 yena joined #git
13:00 Limona no, i've fetched, and then it told me to commit before merging
13:00 _ikke_ right, because you were still in a conflit
13:00 Limona i have a weird "branchy" 2 commit ook now
13:00 Limona right?
13:00 Limona why was i in conflict?
13:01 osse Limona: because both branches modified the same piece of code
13:02 Limona i c
13:02 nikivi joined #git
13:04 Learath2 joined #git
13:04 hanetzer joined #git
13:06 hanetzer so, I think I'm in a bit of a pickle. I'm working on adding a new board to coreboot, and in doing so I need to add a new superio chip to it. right now I have two branches, one for the mainboard and one for the chip (plus master); problem is, I need the chip branch to build the board branch.
13:06 g_zen joined #git
13:07 g_zen joined #git
13:07 hanetzer reason i have them both on separate branches is to aid with gerrits workflow
13:08 hanetzer so, what does one suggest I do to handle such a situation?
13:11 eb0t joined #git
13:11 zeroed joined #git
13:14 fmcgeough joined #git
13:14 ploop joined #git
13:15 mkoskar joined #git
13:16 tarkus Q: If I squash a subtree into parent "git merge -s subtree --squash --prefix=foo/" the following commits merges would result in no common histories error?
13:16 andi_ joined #git
13:16 mischat_ joined #git
13:17 ToBeCloud joined #git
13:19 SaberArthur joined #git
13:19 cyberz_ left #git
13:19 cyberz joined #git
13:21 lxnch joined #git
13:21 noahsussman joined #git
13:25 chedda joined #git
13:26 parasite_ joined #git
13:27 sbeller_ joined #git
13:30 iliv joined #git
13:33 murgi joined #git
13:34 prosody joined #git
13:35 mladenp joined #git
13:35 mladenp hi guys. Is it enough to change repo url just in .git/config file?
13:35 murgi joined #git
13:35 mladenp i want to replace repo to do some testing
13:36 irqq joined #git
13:36 _ikke_ sure, but recommended is to use the git config tool for that
13:36 _ikke_ or even git remote set-url <remote> <url>
13:37 b3kN joined #git
13:37 tnecniv joined #git
13:38 mladenp ok so change url to bitbucket instead of github and if i push it will push to new url? nothing else needs to be changed?
13:38 canton7 yeah, it'll work (assuming that the remote repo exists, that the remote side doesn't reject the push because it's non-fast-forward, etc)
13:39 Silenced joined #git
13:39 mladenp ok, thanks will try now
13:39 theresajayne joined #git
13:40 JamesBaxter joined #git
13:41 WillMoogle joined #git
13:41 imanc joined #git
13:42 MLM_ joined #git
13:46 Averell- joined #git
13:48 Peetz0r joined #git
13:49 anonymuse joined #git
13:49 ams__ joined #git
13:50 da5id joined #git
13:51 Gsham joined #git
13:51 neuro_sys joined #git
13:52 neuro_sys Hmm, isn't git `fetch origin master && git reset --hard origin/master` supposed to make the local same as remote? But even after doing that, if I do a `pull origin master` I receive new commits.
13:54 Limona left #git
13:58 bremner neuro_sys: fetch first
13:59 neuro_sys `git fetch origin master && git reset --hard origin/master` sometimes results in `git pull origin master` merging more commits.
13:59 neuro_sys as though the former doesn't fully reset the local to remote
14:01 rnsanchez_ joined #git
14:01 kadoban neuro_sys: Typically one doesn't specify a refspec with 'git fetch', 'git fetch' is as much as is necessary most of the time. I'm not 100% sure what the behavior actually is when you specify a branch like that.
14:02 kadoban neuro_sys: Also, depending on how long you're taking before doing the pull, that could have an effect of course, remotes don't stay the same forever.
14:02 neuro_sys I observed if I don't specify master, it simply fetches all the remote branches, which I usually don't need.
14:02 neuro_sys kadoban: that's right, but I do it immediately. However it might still be a confusion on my part, I just wanted to make sure if I'm missing something very obvious.
14:03 kadoban Unless it's a very busy remote, does it matter that it fetches all branches?
14:03 qqx neuro_sys: When you do `git fetch origin master` it only updates the FETCH_HEAD ref, not the origin/master remote-tracking branch.
14:03 neuro_sys kadoban: it doesn't, so I'll do `git fetch origin` as of now.
14:03 neuro_sys qqx: ah!
14:03 kadoban If you really want to just fetch some branch, you'd really want to read up on the refspec part of fetch's manpage, which is ... involved.
14:04 kadoban git fetch origin +refs/heads/master:refs/remotes/origin/master      should work, but ... no idea which parts can be shorter
14:05 yottabyte joined #git
14:06 yottabyte what is the right approach for merging master into a feature branch? git merge master, or git rebase master? do they both update history in git log? e.g. I make a commit on 1/1 and 1/3 in my feature branch, but in master someone merged something on 1/2. When I do a merge or rebase of master, will my log then show commits on 1/1, 1/2, and 1/3?
14:06 Bengi joined #git
14:06 cdg joined #git
14:10 neuro_sys You would want to rebase master, I'd say, because a merge commit in your feature branch is undesired. After you rebase, your feature branch commit history will have everything up to master, and your feature commits applied one by one. So it will be 1/2, 1/1, 1/3 I think.
14:10 yottabyte why is a merge commit undesired?
14:11 netj joined #git
14:12 hanetzer yottabyte: because it would be a more or less empty commit with no real info polluting the history?
14:12 kadoban Because some people define it to be undesired for some reason. Which you pick really depends on workflow. If you're sharing the feature branch with others, you usually want to merge. If you're not, you may prefer to rebase. Rebase changes history, which is usually not friendly with shared code.
14:12 hanetzer also depends on the project; some have very particular rules
14:12 lucianangelus joined #git
14:13 _ikke_ random merge commits are undesired (ie, caused by pulling because you wanted to update your branchh)
14:15 JackMc joined #git
14:15 neuro_sys It all depends on the workflow, but that particular merge commit caused by a feature branch that was left behind by the master during developmente is not really relevant or eligible to be included in the master history IMO.
14:15 neuro_sys If you want your master history to reflect all the blood and sweat, and complications during the development, then go for it.
14:16 jimi_ joined #git
14:16 EmeraldExplorer joined #git
14:16 _ikke_ By merging in the feature branch, you clearly separate what commits were part of the feature
14:16 _ikke_ the master branch would only consist of the merge commmit
14:17 mischat joined #git
14:18 neuro_sys Anyways, I need to read up on refspec to understand the git fetch issue, so far I've only lived with the concepts of branches and remotes.
14:19 kadoban neuro_sys: It's not something people mess with much really
14:19 neuro_sys "When you do `git fetch origin master` it only updates the FETCH_HEAD ref, not the origin/master remote-tracking branch." This is still a black box to me.
14:19 _ikke_ This is not true anymore
14:19 _ikke_ It got changes somewhere
14:19 _ikke_ changed*
14:21 neuro_sys I always expected `git fetch origin && git reset --hard origin/master` to make the local same as remote, so apparently refspec is involved, and is nearly part of common use.
14:21 neuro_sys I realized git pull isn't always what I want (because it merges, and my history somehow might have got changed).
14:23 qqx neuro_sys: `git fetch origin` is not the same as `git fetch origin master`. The former has always updated remote-tracking branches.
14:24 qqx And as _ikke_ said, the former has been updated to do that as well. The system that I used to double check is still using a pretty old version of git.
14:24 neuro_sys oh, that is the culprit then... in my case.
14:24 neuro_sys I have 2.10.2
14:25 _ikke_ Yeah, that should definitely also update the local tracking branches
14:25 al-damiri joined #git
14:26 qqx The fetch operation should show what it's updating. With a recent version (such as 2.10.2) even if you specify a branch it should show that it's updating both the remote-tracking branch and FETCH_HEAD.
14:28 neuro_sys Right now, when I do `git fetch origin`, it outputs nothing. If I do `git fetch origin master`, it outputs: "* branch    master   -> FETCH_HEAD".
14:29 neuro_sys Passing -v makes it a bit more understandable.
14:33 boombatower joined #git
14:33 SaberArthur_ joined #git
14:34 Snugglebash joined #git
14:34 ShekharReddy joined #git
14:35 Dougie187 joined #git
14:35 italoacasas joined #git
14:36 Milos joined #git
14:37 saml joined #git
14:37 tarkus joined #git
14:38 llamapixel joined #git
14:40 danemacmillan joined #git
14:40 Ryanar joined #git
14:41 _Vi joined #git
14:41 altendky joined #git
14:43 overlord_tm joined #git
14:43 tnecniv joined #git
14:46 refried_ joined #git
14:46 Guest46178 joined #git
14:47 tnecniv joined #git
14:47 jottr joined #git
14:48 yottabyte hanetzer: a merge from master into a feature branch doesn't include history?
14:48 mahendratech joined #git
14:48 parliament joined #git
14:49 hanetzer yottabyte: I mean specifically the 'merge branch blah' commit. sometimes you just have a branch for tinkering purposes and don't actually mean for its history to end up on master
14:49 nickabbey joined #git
14:49 italoacasas joined #git
14:50 Siegfried joined #git
14:52 Es0teric joined #git
14:52 garethdaine joined #git
14:53 HoierM joined #git
14:54 KevinDF joined #git
14:55 jleclanche joined #git
14:55 SteveCharleston joined #git
14:57 noecc joined #git
14:57 Gsham joined #git
14:58 tarkus Would you recommend using subtree or submodule for subprojects in the main repo (e.g. /app - web app frontend, /server - data api server).. most likely subpeojcts will be developed in isilation with just a couple of minor tweaks inside the parent repo
14:59 Lyqyd joined #git
15:00 jaafar joined #git
15:00 cbreak tarkus: are they closely related? if so, then don't split them
15:00 cbreak if not, then it's ok to split them
15:00 PtxDK joined #git
15:01 tarkus cbreak, they are related but not closely (can be developed separately - front-end part of a web app + graphql data API part)
15:01 cagedwisdom joined #git
15:02 eduardas_m joined #git
15:02 tarkus cbreak, if I split them, how bring them back together in a single parent repo via submodules or subrees?
15:02 synthroi_ joined #git
15:05 Lipp joined #git
15:06 m___ joined #git
15:08 shinnya joined #git
15:09 ExoUNX joined #git
15:10 jozwior joined #git
15:10 cbreak I'd use submodules
15:12 snowkidind joined #git
15:12 kadoban How do you do "minor tweaks" of submodules in a super-repo? Isn't that kind of outside of the submodules model?
15:13 jottr joined #git
15:14 cbreak kadoban: not at all, indeed outside of spec
15:14 cbreak the only control a super repository has about its submodules is which hash it gets to check out
15:14 kadoban Right, so that doesn't necessarily sound like it'd match that well.
15:16 cbreak the point of a parent repo is not to make tweaks
15:16 cbreak it's to have the submodules
15:16 cbreak if you want to change X, then you'll have to do that in the repository that contains X
15:17 cdg_ joined #git
15:17 mar77i_ joined #git
15:17 makerops joined #git
15:17 jstein_ joined #git
15:18 Gsham joined #git
15:19 tarkus cbreak good point.. i was leaning towards subtree initially.. but it will pollute the revision history in a super-repo
15:19 chadhinton left #git
15:19 makerops is there a way to query a remote for a particular commit?
15:20 KevinDF joined #git
15:20 ash_workz joined #git
15:20 cdown joined #git
15:21 bumpy joined #git
15:22 garethdaine joined #git
15:22 tang^ joined #git
15:24 Learath2 joined #git
15:25 noecc joined #git
15:26 Seveas makerops: depends on your definition of query. All you can get from a remote is information about which refs point where (and of course you can fetch all those...)
15:28 menip joined #git
15:28 vrlx joined #git
15:29 makerops I have several (40 or 50) distributed git repos that live on management sysytems, and servers that use the management systems as remotes
15:29 rwb joined #git
15:29 makerops I sync to the management servers via a cental repo
15:29 makerops what Im trying to do, is before a change is made on the "end server" is query the upstream mgmt server to make sure the commit has been synced out
15:31 nettoweb joined #git
15:31 absinthe joined #git
15:31 jaguarmagenta joined #git
15:34 blackwind_123 joined #git
15:34 bouncy joined #git
15:35 ResidentBiscuit joined #git
15:35 absinthe Is there a way to log git commits of a submodule when being in the parent directory?
15:35 ResidentBiscuit joined #git
15:37 MarkFletcher joined #git
15:38 mostlybadfly joined #git
15:38 crose joined #git
15:39 qqx absinthe: You can use `git -C path/to/submodule` log
15:39 eduardas_m joined #git
15:39 LionsMane joined #git
15:39 lindenle joined #git
15:39 qqx There I meant to have the closing quote at the end.
15:40 subhojit777 joined #git
15:40 C00D3R joined #git
15:40 bdeferme joined #git
15:41 absinthe qqx ok, i'll give it a try. thanks
15:43 Silenced joined #git
15:45 _Patrick_ joined #git
15:46 Gsham joined #git
15:46 cdown_ joined #git
15:46 shgysk8zer0 joined #git
15:47 rorro joined #git
15:47 bouncy joined #git
15:48 noidea joined #git
15:48 Es0teric joined #git
15:51 unbalanced joined #git
15:51 unbalanced joined #git
15:52 rchavik joined #git
15:53 breezy_ joined #git
15:54 moerkfaen joined #git
15:55 Gsham joined #git
15:57 dikiaap joined #git
15:58 tarkus joined #git
15:58 unbalanced joined #git
15:59 xlegoman joined #git
16:00 nicksloan joined #git
16:00 GodGinrai joined #git
16:05 eduardas_m joined #git
16:05 gareppa joined #git
16:05 muelli joined #git
16:05 kumarsuraj joined #git
16:06 star_prone joined #git
16:07 C00D3R joined #git
16:08 C00D3R left #git
16:08 Zamerick joined #git
16:13 regedit joined #git
16:15 eduardas_ joined #git
16:16 Shentino joined #git
16:18 ejb joined #git
16:18 Es0teric joined #git
16:18 m___ joined #git
16:19 italoacasas joined #git
16:20 aarobc joined #git
16:21 Lipp joined #git
16:21 bsanford joined #git
16:22 Gsham joined #git
16:22 mar77i joined #git
16:23 mischat joined #git
16:24 mischat joined #git
16:25 mischat joined #git
16:26 mischat joined #git
16:27 tvw joined #git
16:27 mar77i_ joined #git
16:30 synthroid joined #git
16:31 synthroi_ joined #git
16:31 Shentino joined #git
16:33 Shentino joined #git
16:35 publio joined #git
16:37 cjohnson I accidentally just did this: git checkout commithash && git cherry-pick master
16:37 Lipp joined #git
16:37 inovas joined #git
16:37 cjohnson it seems to have created a new commit, am I right in thinking I can just git checkout master and that other detached commit is lost?
16:37 kadoban cjohnson: Ya
16:37 navidr joined #git
16:38 cjohnson I meant to git checkout master && git cherry-pick that-hash
16:38 cjohnson cool
16:39 red82 joined #git
16:40 synthroid joined #git
16:41 derk0pf joined #git
16:42 Lipp_ joined #git
16:43 Es0teric joined #git
16:46 elvay joined #git
16:47 elvay left #git
16:48 Es0teric joined #git
16:48 _Vi joined #git
16:50 dreiss joined #git
16:50 johnny56 joined #git
16:52 mischat joined #git
16:53 str joined #git
16:53 mischat_ joined #git
16:53 einnjo joined #git
16:53 Lipp joined #git
16:53 raspado joined #git
16:54 zeroed joined #git
16:54 zeroed joined #git
16:55 ikonia win 8
16:55 ikonia oops, sorry
16:56 mischat_ joined #git
16:57 madewokherd joined #git
17:00 hashpuppy joined #git
17:01 finalbeta joined #git
17:02 noecc left #git
17:02 logikos joined #git
17:02 bruce_lee joined #git
17:03 nicksloan joined #git
17:04 garethdaine joined #git
17:04 floatingman joined #git
17:04 jccn joined #git
17:04 _Cyclone_ joined #git
17:05 zeroed joined #git
17:09 howitdo joined #git
17:09 howitdo joined #git
17:11 svm_invictvs joined #git
17:12 Lipp joined #git
17:12 cyan__ joined #git
17:12 robotroll joined #git
17:13 m___1 joined #git
17:13 Rodya_ joined #git
17:16 howitdo joined #git
17:17 czart joined #git
17:21 Lipp joined #git
17:21 PioneerAxon_ joined #git
17:21 msonntag joined #git
17:21 czart Hi, how to enable displaying the actual author identity of every commit he made on HitGub webpage? Since currently my commits are marked with the github account name (github login) only.
17:22 durham joined #git
17:22 grawity that's what github always does
17:23 czart grawity: ok, so I assume this can not be changed...
17:24 novas0x2a|laptop joined #git
17:24 nd_ joined #git
17:24 xlegoman joined #git
17:24 pks joined #git
17:25 czart grawity: Consider scenario: there are several commits coming from different authors and only single github is used to push those commits. How to distinguish (show on GitHub) which author is associated with respective commits?
17:25 czart Or, as you said - this is not possible?
17:26 czart I meant: single github account
17:27 chardan joined #git
17:27 BlessJah czart: github recognises user based on email saved in author and commiter fields
17:28 Lipp joined #git
17:28 algun joined #git
17:29 ploop joined #git
17:29 BlessJah when it does not know the author - I would guess it will simply use author field instead of linking to some user
17:30 italoacasas joined #git
17:31 czart BlessJah: thanks, I will seek to set it up.
17:32 BlessJah czart: an example of commit with author not known to github: https://github.com/git/git/commit/c1e0dc59bddce765761a6f863c66ee0cd4b2ca09
17:34 Motoservo joined #git
17:34 unbalanced joined #git
17:35 elsevero joined #git
17:37 grawity czart: what BlessJah said – it's taken from the commits' own metadata; it has absolutely nothing to do with who pushed the commits
17:37 lindenle joined #git
17:38 hanetzer czart: could institute a signed-off by policy similar to the linux kernel or coreboot
17:38 czart BlessJah: Also, I commit with using PyCharm, so maybe the IDE overrides some setting, do you know?
17:39 Gsham joined #git
17:39 aruns joined #git
17:39 czart grawity: yes, you can see who is the author of commit by invoking: git show in command line.
17:40 treehug88 joined #git
17:40 gareppa joined #git
17:40 czart hanetzer: I do not know those policies.
17:40 BlessJah czart: go to project directory and type git log, you will see author/mail as seen by github
17:40 czart BlessJah: OK.
17:40 BlessJah git show is also good command, but it will show only last commit
17:43 jstein_ joined #git
17:43 mda1 joined #git
17:44 shgysk8zer0 joined #git
17:44 byte512 joined #git
17:44 sea-gull_ joined #git
17:45 refried_ joined #git
17:48 peeps[lappy] is it bad to rebase a branch thats already been pushed?
17:48 guest_____ joined #git
17:48 resmo joined #git
17:48 resmo hi
17:49 acetakwas joined #git
17:49 resmo is there a way to append a msg to the current commit msg non-interactively?
17:50 corn13read joined #git
17:51 corn13read what's the best way to check if a branch is merge-able but not do the merge? I want a proper test for my bash script
17:51 cjohnson peeps[lappy]: has anybody else used it?
17:51 cjohnson or is it deployed anywhere that isn't disposable
17:51 peeps[lappy] cjohnson, no.  i guess i just have to do a push --force?
17:51 vdv joined #git
17:51 xaa joined #git
17:51 cjohnson yep exactly
17:51 Lipp joined #git
17:51 cjohnson I do that regularly to refresh feature branches from master
17:51 cjohnson while I'm working on them
17:52 cjohnson then once you merge make sure to stop rebasing
17:52 cjohnson (make a hotfix branch instead)
17:52 peeps[lappy] ok, yeah was just confused for a sec because it said my branch was behind remote, but that's expected?
17:53 cjohnson yep, remember when you rebase you blow away all your local commits and then recreate with new ones
17:53 cjohnson so when it compares to remote, technically you are behind
17:53 peeps[lappy] ok, thanks that makes sense
17:57 cuku joined #git
17:57 BlessJah peeps[lappy], cjohnson: never push --force
17:58 BlessJah !force
17:58 gitinfo [!force_push] If you need to overwrite the history of a remote git repository (very bad idea, see !rewrite), you can do so with `git push --force-with-lease`. Note the remote server may reject this. See man git-config and search for receive.denyNonFastForwards. Best practice is for upstream servers to denyNonFastForwards.
17:58 czart BlessJah: yeah, it's show the proper author now. It is as you said: 'github recognises user based on email', i.e. user.email local setting can not be associated with any account at GitHub, otherwise GitHub uses the github account associated with the e-mail.
17:59 BlessJah czart: depending on what you want to achieve, you may want to create separate accounts for developers
17:59 czart BlessJah: yes, thats another option, thanks :)
17:59 fahadash joined #git
18:01 BlessJah corn13read: I believe there is better option, but you can go to detached head state, perform merge, abort if it failed, and get back to branch
18:03 mikecmpbll joined #git
18:05 enleeten joined #git
18:06 Es0teric joined #git
18:06 tang^ joined #git
18:07 cjohnson BlessJah: er, force pushing is the only way to rewrite history of a branch that's already been pushed
18:07 cjohnson of course you can force push, you just need to know the context of hte branch you are force-pushing
18:07 cjohnson as I had explained...
18:08 BlessJah cjohnson: no, --force flag is evil regardless of context and intention, always use --force-with-lease
18:10 Shentino joined #git
18:11 mentazoom joined #git
18:11 jnavila joined #git
18:12 * cjohnson shrugs
18:12 cjohnson I disagree
18:12 venmx joined #git
18:13 d0nn1e joined #git
18:13 cjohnson I advocate informed decision making
18:13 BlessJah sure, what is then valid use case that is not covered by with-lease version?
18:13 nicksloan joined #git
18:14 * BlessJah is genuinely interested if this flag has any use
18:14 Spec-Chum joined #git
18:15 cjohnson If I know for sure that I have the authoritative version of this branch and want to push it
18:15 cjohnson especially in cases where I am the only developer working on it
18:15 nandi_ge___ joined #git
18:15 BlessJah covered by --force-with-lease
18:15 cjohnson If I have screwed something up then my approach is generally to clean things up locally, get it to a known correct state, and push that
18:16 cjohnson er, no it's not
18:16 cjohnson but I can see this is a waste of time. I don't even need to be able to justify it to make my point about advocating informed decision making
18:16 cjohnson simply knowing about it is enough
18:16 BlessJah cjohnson: lease prevents you from making mistake, this is not waste of time
18:17 cjohnson The dogmatic discussion is what I'm referring to as a waste of time
18:18 BlessJah I would be fine with current --force being longer (e.g. --do-careless-force) and current lease beahaviour being invoked by --force
18:19 grawity the old behaviour could be under --darth-vader
18:19 cjohnson lol
18:19 cjohnson I'd be ok with that
18:20 BlessJah yep, something hard to find in manual, unless you exactly now what you are looking for
18:21 mikecmpbll joined #git
18:21 nickabbey joined #git
18:21 BlessJah cjohnson: while thinking about it, I finally found reason for --darth-vader to stay: someone pushed gigabytes of binary data you want to get rid of, without downloading first
18:22 dmc joined #git
18:22 mmlb joined #git
18:23 moerkfaen JAH
18:23 moerkfaen rastafariiiiiiii
18:23 brent__ joined #git
18:24 tyreld joined #git
18:24 nickabbey joined #git
18:24 mizu_no_oto joined #git
18:25 cjohnson lol
18:25 durham joined #git
18:26 Spec-Chum joined #git
18:28 durham_ joined #git
18:30 nicksloan joined #git
18:30 dreiss joined #git
18:31 robattila256 joined #git
18:32 Motoservo left #git
18:32 e14 joined #git
18:34 tmsmith joined #git
18:34 nanga joined #git
18:34 joshszep joined #git
18:41 Domhack joined #git
18:42 muelli joined #git
18:43 mizu_no_oto joined #git
18:43 SwiftMatt joined #git
18:45 romerocesar joined #git
18:47 dermoth joined #git
18:47 shgysk8zer0 joined #git
18:47 davimore joined #git
18:48 italoacasas joined #git
18:49 Linux-Ninja01 joined #git
18:51 geoid joined #git
18:55 ajf- joined #git
18:56 Domhack_ joined #git
18:56 railssmith joined #git
19:02 finalbeta joined #git
19:02 hanthings_ joined #git
19:02 wikkie joined #git
19:04 ahmedelgabri joined #git
19:07 str joined #git
19:09 Sasazuka joined #git
19:10 cagmz joined #git
19:13 corn13read joined #git
19:13 meLon joined #git
19:15 dsantiago joined #git
19:17 star_prone joined #git
19:24 Lipp joined #git
19:24 [Saint] joined #git
19:26 [Saint] Potentially obscure question I can't/haven't yet found an answer to:
19:26 [Saint] "Can I set the 'config --global' location?"
19:26 [Saint] And if I can:
19:26 [Saint] "How?"
19:26 [Saint] (operating systems are debian-esque)
19:26 vrlx joined #git
19:26 thiago yes, you can choose between two options:
19:26 thiago ~/.gitconfig and ~/.config/git/config
19:27 [Saint] Hmmmm...ok, the latter will do in a pinch for what I'm aiming for.
19:27 [Saint] What governs the behavior selection?
19:27 [Saint] errr, selection behavior.
19:28 phroa they're read in this order https://jk.gs/git-config.html#FILES
19:28 phroa meaning ~/.gitconfig overrides ~/.config...
19:29 [Saint] OK. Right. So if the former doesn't exist and the latter does, 'config --global foo.bar baz' automagically picks the latter up and won't create the former?
19:29 tzuu joined #git
19:29 tang^ joined #git
19:30 thiago [Saint]: the presence or absence of the files
19:30 [Saint] also - thanks.
19:30 [Saint] right.
19:31 thiago if none exists, using git-config will write to ~/.gitconfig
19:31 thiago otherwise, it writes to the one that exists
19:31 thiago if both exists, it's ~/.gitconfig
19:32 romerocesar joined #git
19:32 [Saint] just trying to tidy up/compartmentalize dotfiles, this is helpful, thanks.
19:32 arand joined #git
19:33 einnjo joined #git
19:33 Jackneill_ joined #git
19:33 weam joined #git
19:33 thiago yeah, I've decided to only use ~/.config/git/config too
19:34 [Saint] the ~/.config/git/ option is actually really handy. Now I essentially just have .bashrc.d, .config, and .profile.d instead of ten bazillion independent dotfiles.
19:36 irco joined #git
19:38 [Saint] Thanks again thiago and phroa, appreciated.
19:38 [Saint] Turns out I'm an idiot and it was /right/ there in front of me, right at the top of the gitconfig git-scm doc page. My apologies.
19:39 synthroi_ joined #git
19:40 spacelord_ joined #git
19:41 italoacasas joined #git
19:41 watabou joined #git
19:41 aruns joined #git
19:44 irqq joined #git
19:44 przemoc joined #git
19:46 neminis joined #git
19:47 aarobc joined #git
19:52 jost_ joined #git
19:53 acetakwas joined #git
19:55 PioneerAxon joined #git
19:56 knod joined #git
19:57 HoierM joined #git
19:57 mischat joined #git
19:57 WaReZ joined #git
19:57 knod I have a file that I can't seem to stage for commit. The git diff shows the commit ID removed and the commit ID + '-dirty' added. What should I do?
19:57 durham joined #git
20:03 jccn joined #git
20:07 rominronin joined #git
20:07 durham joined #git
20:08 nickabbey joined #git
20:08 cjohnson sounds like you're talking about a submodule not a file?
20:09 cjohnson in which case that shows up when you modify a file within a submodule and the working tree in that submodule is dirty
20:09 knod cjohnson: Hmm, I did copy a git clone in there
20:09 cjohnson well yeah that is a problem...
20:10 knod Ok, how do I remove git from there?
20:10 Linux-Ninja01 joined #git
20:10 knod Eh, maybe I'll just get rid of the old folder, npm install the module from scratch, and re-edit to the changes that I want.
20:11 knod That would probably be simpler
20:11 knod It's at least something I can do without any further research.
20:11 red82 joined #git
20:11 flavius joined #git
20:12 fatalhalt joined #git
20:16 dsdeiz joined #git
20:19 rorro joined #git
20:20 acetakwas joined #git
20:20 [Saint] knod: My understanding is that the 'right' answer is:
20:20 [Saint] 'git $command --ignore-submodules=dirty'
20:21 m0viefreak joined #git
20:21 prg318 If I have a repo with a submodule that I would like to be a part of the main repo - is there a way to convert the submodule to the main repo without losing history?
20:21 knod [Saint]: Thanks for the info
20:21 knod It would be nice to just push my changes to that repo too as I edit…
20:21 [Saint] knod: well, that's the 'ignore it' option.
20:22 griffindy joined #git
20:22 [Saint] There's a 'fix it' option too.
20:22 [Saint] Can't quite remember the syntax though.
20:22 red82 joined #git
20:22 [Saint] Something along the lines of:
20:22 [Saint] 'git submodule foreach --recursive git add .'
20:23 [Saint] that'll walk your submodules recursively and add the 'dirt'.
20:23 Naan joined #git
20:23 [Saint] I /may/ be screwing the syntax up there, but it's surely enough breadcrumbs to follow.
20:23 knod Hmm. I couldn't just go to that folder dir in terminal and git add . in that folder?
20:23 [Saint] Yes.
20:24 knod Yeah, definitely plenty of breadcrumbs. I appreciate that
20:24 Efffe joined #git
20:24 [Saint] I was assuming multiple dirty modules, sorry. Didn't realize it was just one errant complainer.
20:24 cjohnson this isnt a submodule
20:24 cjohnson they just blondly moved one git repo in another
20:24 hwoarang joined #git
20:24 [Saint] Oooooooh.
20:25 [Saint] I see.
20:25 cjohnson if you want to maintain the nested dir as its own repo knod then submodules are the answer
20:25 jimi_ joined #git
20:26 [Saint] git turtles all the way down.
20:27 knod lol
20:27 knod ok, thanks cjohnson
20:28 Gsham joined #git
20:30 absinthe joined #git
20:32 Sample joined #git
20:33 Sample joined #git
20:33 Ardethian joined #git
20:33 rmbeer joined #git
20:34 rmbeer left #git
20:36 aw1 joined #git
20:36 Lxuta joined #git
20:38 mischat joined #git
20:40 anonymuse joined #git
20:41 Lipp joined #git
20:42 nanga joined #git
20:47 refried_ joined #git
20:50 carpii joined #git
20:50 vcarl joined #git
20:51 carpii can a git alias accept params to be embedded in the substituted command?   Basically im trying to turn 'git stashdiff 3' into 'git stash show -p stash@{3}'
20:53 cvitullo joined #git
20:55 git517 joined #git
20:56 elsevero joined #git
20:56 Es0teric joined #git
20:57 dreiss joined #git
20:57 vcarl joined #git
20:57 dsdeiz joined #git
20:59 tarkus joined #git
21:00 geoid joined #git
21:00 texinwien_ joined #git
21:02 vcarl is there a good way to sync up 2 branches that have had commits cherrypicked instead of merged/rebased for years?
21:02 vcarl one is the "live" branch, the other is the "development" branch, i'm trying to get us on a sane branching strategy and bringing history together would make that a lot easier
21:05 kadoban vcarl: I think you can probably just merge one into the other. Though ... not sure, make a backup first in case that goes sideways.
21:06 knod left #git
21:07 vcarl uh wow
21:07 vcarl it worked, not so much as a merge conflict
21:07 garethdaine joined #git
21:07 Linux-Ninja01_ joined #git
21:07 vcarl that's... surprising, i expected a lot of sideways-ness
21:07 star_prone joined #git
21:09 marlinc joined #git
21:09 dhollinger joined #git
21:11 howitdo joined #git
21:13 shgysk8zer0 joined #git
21:15 Gsham joined #git
21:16 jccn joined #git
21:17 manuels joined #git
21:27 chardan joined #git
21:27 durham joined #git
21:31 nickabbey joined #git
21:33 adamm joined #git
21:34 synthroid joined #git
21:34 durham joined #git
21:34 nickabbey joined #git
21:35 kyan joined #git
21:35 kayzon joined #git
21:35 manuels joined #git
21:36 Derderderd joined #git
21:37 f4 joined #git
21:38 dreiss joined #git
21:38 WebDragon joined #git
21:38 WebDragon hey folks. need some help with a corrupted git repo https://paste.fedoraproject.org/502716/14813194/
21:38 WebDragon been looking at https://git.wiki.kernel.org/index.php/GitFaq#How_to_fix_a_broken_repository.3F but the methodology there doesn't seem to outline my exact situation
21:39 acetakwas joined #git
21:39 selckin restore backup, clone again from source?
21:39 WebDragon this is the only copy of this repo (no remotes)
21:40 WebDragon otherwise I'd be all over that
21:40 WebDragon there's no tree object to inspect, in the git fsck output, so I'm sort of stuck
21:41 overlord_tm joined #git
21:41 WebDragon been puttering around stackexchange looking for similar situations but haven't yet found an 'exact' match close enough to my situation to proceed
21:43 WebDragon and of course it's almost 5pm here on the east coast
21:43 ojacobson WebDragon: gauging from the messages, aa5f0a6d3c26b23107f1afb4c3611510cb73d4b3 should be a commit object
21:44 ojacobson if you know the author, author date, committer, committer date, message, and tree, you can reconstruct the object
21:44 WebDragon it's all me
21:44 WebDragon the object file itself is emptuy
21:44 WebDragon empty*
21:45 WebDragon if I try one of the suggestions in the git wiki, I get  git commit -m "Fixing Git Repo"
21:45 WebDragon fatalhalt: could not parse HEAD
21:45 tnecniv joined #git
21:45 ojacobson yep. You're well into needing to use lower-level tools like git commit-tree
21:45 WebDragon (this after moving the empty/corrupt object out of the tree)
21:45 ojacobson do you know what tree object aa5f0a6d3c26b23107f1afb4c3611510cb73d4b3 should have been using?
21:46 ojacobson (odds of you reproducing that exact object: slim)
21:46 ojacobson (bordering on none)
21:46 WebDragon off the top of my head no. I'd just made one single commit to the tree, today (haven't worked on this project in ages) and can easily reproduce the commit from the working files
21:46 WebDragon and then the laptop crashed and when I went to make the next commit after rebooting, boom
21:47 WebDragon no ongoing committing or otherwise working with the repo was going on at the time -- I was just trying to view the website in a local browser
21:47 WebDragon browser crashed, and had to reboot
21:47 WebDragon locked up the whole box, annoyingly. (thanks Chrome)
21:48 yottabyte left #git
21:48 WebDragon if I could somehow restore to the previous commit, there's not much to recover, work-wise
21:48 WebDragon but I'm not 100% certain what that object is supposed to be referring to
21:48 ojacobson the previous commit presumably exists
21:50 WebDragon I'm still a bit shaky on the low-level git stuff as I've hardly ever had to use them
21:51 przemoc joined #git
21:53 manuels joined #git
21:54 WebDragon so the real question is, what do I try next?
21:54 manuels joined #git
21:56 mizu_no_oto joined #git
22:02 inflames joined #git
22:02 corn13read joined #git
22:06 m___ joined #git
22:09 zerozero joined #git
22:09 italoacasas joined #git
22:10 rwb joined #git
22:11 altendky joined #git
22:12 Derderderd joined #git
22:15 Lipp joined #git
22:15 unbalanced joined #git
22:18 ahmedelgabri joined #git
22:19 aard_ joined #git
22:19 _ikke_ use git reflog to find the previous commit
22:19 _ikke_ and then force the branch to that commit and check if that works
22:19 tnecniv joined #git
22:19 _ikke_ WebDragon: ^
22:20 zacts joined #git
22:20 WebDragon _ikke_: yeah but that gives me the same problem. git reflog
22:20 WebDragon fatalhalt: bad object HEAD
22:20 WebDragon cuz it reads HEAD
22:20 WebDragon ;)
22:20 fatalhalt oh noes
22:20 WebDragon bahahhahahaha @ fatalhalt
22:21 Es0teric joined #git
22:21 sssilver joined #git
22:22 _ikke_ WebDragon: Just look directly at .git/refs/HEAD then
22:23 _ikke_ It's human readable
22:23 aard__ joined #git
22:27 mischat joined #git
22:27 Lipp joined #git
22:28 tmsmith joined #git
22:30 geoid joined #git
22:31 chadhinton joined #git
22:31 WebDragon _ikke_: https://paste.fedoraproject.org/502736/14813226/raw/  so this is what I get from tail .git/refs/HEAD (edited for email substitution)
22:31 jottr joined #git
22:31 heftig joined #git
22:31 WebDragon current branch is 'develop'
22:32 joshszep joined #git
22:33 joshszep joined #git
22:36 rbern joined #git
22:36 WebDragon _ikke_: so what do I do with that info?
22:39 netj joined #git
22:41 hahuang62 joined #git
22:41 watabou joined #git
22:42 AfroThundr joined #git
22:44 _ikke_ WebDragon: So this is your last commit right? "fix for missing HTTP_HOST server variable spitting in error_log"
22:45 WebDragon there was one commit past that, to the best of my knowledge, that I had made today, prior to the laptop crashing
22:45 jottr_ joined #git
22:45 Linux-Ninja01_ joined #git
22:45 hahuang65 joined #git
22:45 WebDragon but it's easily recreated from the file contents as it was all edits to one file, plus the new pdfs they link to
22:46 WebDragon so if I can back up to that last commit and restore the repo to that state, I'm good
22:46 _ikke_ Ok
22:46 troyt joined #git
22:47 kfoonamalik joined #git
22:47 _ikke_ try git update-ref -m "Fix develop" refs/heads/develop 8dfd0c2e2a89a206df1c50d2d43e2e905147ee56
22:47 NightMonkey joined #git
22:47 _ikke_ This will change develop to point to the last commit you see in the reflog
22:47 netj joined #git
22:48 WebDragon and then what, git status, git fsck, what
22:48 _ikke_ Yeah, check if everything is alright now
22:49 _ikke_ git status may give confusing status (because you reset the branch, but not the files
22:49 _ikke_ like you did not make the commit you did today
22:50 WebDragon yeah git status is back to where it should be, exactly as you say -- as if I had not made that last commit
22:51 WebDragon so should I now run a git fsck --full ?
22:51 __main__ joined #git
22:51 WebDragon (and probably a git gc)
22:51 _ikke_ Yeah, might be a good idea
22:51 _ikke_ I suspect only the last commit got butched though
22:51 WebDragon fsck still reports dangling blob
22:52 _ikke_ dangling blob is innocent
22:52 WebDragon http://paste.fedoraproject.org/502740/13239171
22:52 _ikke_ At least, most of the time
22:52 _ikke_ whenever you git add a file, but never commit
22:52 _ikke_ then you get a dangling object for example
22:52 WebDragon http://www.funnysigns.net/files/bill-stickers-is-innocent-400x274.jpg heh made me think of this
22:53 WebDragon so you think git gc should clean that up, then?
22:53 _ikke_ hehe
22:54 dzho joined #git
22:55 steve joined #git
22:56 Salander27 joined #git
22:56 novas0x2a|laptop joined #git
22:56 WebDragon yeah git gc did clean up the danglers
22:56 WebDragon whew!
22:56 WebDragon thanks ever so much, _ikke_
22:57 jimi_ joined #git
22:57 * WebDragon slides a mug of _ikke_'s favorite beverage down the bar ~~[_]3
22:57 elsevero joined #git
22:57 dreiss joined #git
22:58 manuels joined #git
22:59 hagridaaron joined #git
23:00 miczac joined #git
23:01 fmeerkoetter joined #git
23:01 WebDragon time to hit the showers and head to the bar
23:03 miczac joined #git
23:04 str joined #git
23:04 JanC_ joined #git
23:04 dzho joined #git
23:05 Limes_ joined #git
23:06 corn13read joined #git
23:07 Shentino joined #git
23:11 kyan joined #git
23:12 Anja joined #git
23:12 ahmedelgabri joined #git
23:12 Derderderd joined #git
23:13 zerozero joined #git
23:14 jottr joined #git
23:15 elsevero joined #git
23:16 astrofog joined #git
23:18 ahmed_elgabri joined #git
23:18 Ediz joined #git
23:19 dearpanda joined #git
23:20 m0viefreak joined #git
23:21 durham joined #git
23:21 jmpp joined #git
23:21 jmpp joined #git
23:23 venmx joined #git
23:27 NeonSynth joined #git
23:27 mischat joined #git
23:28 Sasazuka joined #git
23:28 durham joined #git
23:29 Zamerick joined #git
23:31 geoid joined #git
23:31 whitby joined #git
23:32 buffal0 joined #git
23:35 alex1a_ joined #git
23:35 ski7777 joined #git
23:35 vrlx joined #git
23:37 przemoc joined #git
23:37 joshszep joined #git
23:38 xlegoman joined #git
23:38 baobao joined #git
23:39 Mamqa joined #git
23:39 faux___ joined #git
23:40 Mamqa Hello!
23:40 gitinfo Mamqa: 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.
23:40 chadhinton joined #git
23:41 Tsutsukakushi joined #git
23:43 geoid joined #git
23:43 chadhinton joined #git
23:44 cdg joined #git
23:45 fatalhalt joined #git
23:49 Shentino joined #git
23:52 pitastrudl joined #git
23:52 nikivi joined #git
23:52 cj_ joined #git
23:57 YuGiOhJCJ joined #git
23:59 ToBeCloud joined #git

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