Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2017-05-08

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

All times shown according to UTC.

Time Nick Message
00:04 ravndal joined #git
00:04 Atomic_qTAx7 joined #git
00:06 grazfather joined #git
00:07 grazfather hey guys, I am trying to change the layout of my github page. I have forked the layout i want, and i could just add my assets and _posts, but i want the post history. how do i merge the two unrelated histories? --allow-unrelated-histories doesn't work
00:07 Atomic_qTAx7 joined #git
00:08 durham joined #git
00:09 northfurr joined #git
00:12 tristanp joined #git
00:16 dzho !github
00:16 gitinfo Note that git != github. Feel free to ask us about Github-specific features (Forks, Pull Requests, Wikis, etc), but there are no guarantees. There is a #github channel, which might help too (again, no guarantees)
00:16 Cabanossi joined #git
00:17 pavlix :)
00:20 fstd_ joined #git
00:22 YuGiOhJCJ joined #git
00:24 Sadeukko joined #git
00:27 dave0x6d joined #git
00:35 mentazoom joined #git
00:37 sturner joined #git
00:37 telephone joined #git
00:37 telephone joined #git
00:39 Guest18947 joined #git
00:39 chipotle joined #git
00:41 grazfather it's a merging problem, not a github problem
00:41 grazfather dzho:
00:41 sturner joined #git
00:41 justan0theruser joined #git
00:43 LordRyan i don't suppose there's a way to make git work with symlinks in a way that allows two files in the same repository linked to each other, is there?
00:43 LordRyan without having absolute paths
00:43 cads joined #git
00:43 bremner "two files in the same repository linked to each other" sounds easy with relative paths
00:44 bremner ln -s foo ../bar/foo
00:44 bremner or so.
00:44 LordRyan if i understand symlinking relative paths correctly, won't that error if i'm not in the same directory that i was in when making the link?
00:44 bremner I don't understand the question.
00:44 bremner symlinks exist at some place in the file system.
00:45 LordRyan one second, i might be doing the symlinks wrong
00:45 LordRyan brb
00:46 duderonomy joined #git
00:46 Cabanossi joined #git
00:47 LordRyan right, so i totally don't understand relative symlinking correctly. my bad :P
00:48 bremner roughly speaking, symlinks are files with absolute or relative paths in them
00:48 fission6 joined #git
00:49 LordRyan i get how it works now; the way i was doing it was from a misunderstanding of symlinks that somehow worked.
00:49 LordRyan the path has to be relative to the symlink, not relative to the current directory
00:50 mel00010 joined #git
00:52 tnecniv joined #git
00:58 moei joined #git
01:03 miha_S7_ joined #git
01:04 jameser joined #git
01:06 d^sh joined #git
01:12 Fairy joined #git
01:13 dsdeiz joined #git
01:13 osse joined #git
01:16 rkazak joined #git
01:18 MattMaker joined #git
01:18 livingstn joined #git
01:19 Wulf4 joined #git
01:38 sunri5e joined #git
01:47 nevodka joined #git
01:47 h12o joined #git
01:47 duderonomy If my next feature has a dependency on feature-branch-1, is it common to rebase feature-branch-2 on top of feature-branch1; then later when PR for feature-branch-1 is approved and merged to master, then rebase feature-branch-2 on master (after fetch and pull on master)? Thanks very much for taking the time to read.
01:47 Cabanossi joined #git
01:49 ilbot3 joined #git
01:49 Topic for #git is now Welcome to #git, the place for git help. Just git help, apparently. | Public logs at https://gitirc.eu/log | First visit? Read: https://gitirc.eu | Current stable version: 2.12.2 | Getting "cannot send to channel"? /msg gitinfo .voice | This is serious business. https://madeitwor.se/git-jokes
01:50 durham joined #git
01:53 dsdeiz joined #git
01:53 dsdeiz joined #git
02:05 prg3 joined #git
02:14 lagothri1 joined #git
02:14 Gurkenglas joined #git
02:14 johnnyfive joined #git
02:15 Hello71 read
02:16 Hello71 did not understand
02:16 Cabanossi joined #git
02:17 fish3485 joined #git
02:20 rchavik joined #git
02:22 mel00010 joined #git
02:26 northfurr joined #git
02:30 fish3485 joined #git
02:34 miha_S7_ joined #git
02:39 hexagoxel joined #git
02:41 thallada joined #git
02:44 fhackdroid joined #git
02:44 lb1c joined #git
02:45 nd joined #git
02:46 Cabanossi joined #git
02:49 dumnut joined #git
02:52 blunaxela joined #git
02:54 Vortex35 joined #git
02:54 rkazak joined #git
03:04 kexmex joined #git
03:08 onehrxn joined #git
03:08 chachasmooth joined #git
03:09 davimore joined #git
03:14 fahadash joined #git
03:15 Cabanossi joined #git
03:17 aavrug joined #git
03:17 dermoth joined #git
03:19 aavrug joined #git
03:19 iamasamosa joined #git
03:24 northfurr joined #git
03:26 durham_ joined #git
03:31 Emperor_Earth joined #git
03:36 ShapeShifter499 joined #git
03:37 robotroll joined #git
03:37 pijiu joined #git
03:38 Anja joined #git
03:41 pijiu2 joined #git
03:43 marianina8 joined #git
03:45 Cabanossi joined #git
03:45 durham joined #git
03:47 Starcraftmazter joined #git
03:47 Starcraftmazter hi
03:47 Starcraftmazter how can i remove any untracked files but not ignored files?
03:47 venmx joined #git
03:49 kadoban Some invocation of 'git clean' should be able to. Read carefully before running.
03:49 kadoban Starcraftmazter: ^
03:50 durham_ joined #git
03:56 pifon joined #git
04:00 Starcraftmazter ah yeh thanks kadoban
04:00 jaziz joined #git
04:00 watabou joined #git
04:03 miha_S7_ joined #git
04:09 xall joined #git
04:15 Xenophon1 joined #git
04:18 xall joined #git
04:21 dumnut joined #git
04:22 prg3 joined #git
04:24 zefferno joined #git
04:28 Doginal joined #git
04:30 blackwind_123 joined #git
04:31 Doginal joined #git
04:33 marianina8 joined #git
04:37 h12o joined #git
04:38 ski7777 joined #git
04:40 Raging_Hog joined #git
04:41 davidomanfredo joined #git
04:41 xall_ joined #git
04:41 redoverture joined #git
04:41 nickel_dime_look joined #git
04:42 darkbit joined #git
04:44 xall joined #git
04:45 Cabanossi joined #git
04:54 FuzzyWuzzyPanda joined #git
04:57 MattMaker joined #git
04:57 treaki_ joined #git
05:01 prg3 joined #git
05:01 Silenced joined #git
05:02 xall joined #git
05:07 D630 joined #git
05:12 pks_ joined #git
05:13 dec0n joined #git
05:13 shgysk8zer0 joined #git
05:16 lindenle joined #git
05:16 a_thakur joined #git
05:16 tourdownunder joined #git
05:16 miha_S7_ joined #git
05:16 m4sk1n joined #git
05:20 xall joined #git
05:23 dumnut joined #git
05:25 xall_ joined #git
05:25 spm_draget joined #git
05:25 overlord_tm joined #git
05:26 spm_draget If I use bfg/filter-branch, by HEAD will be rewritter for sure, or? So people working on the same repository *MUST* re-clone it to keep working on it, even if they would not touch the directory that was removed, or?
05:26 spm_draget Or - as long as they have no changes in this directory - can they simply do an 'git pull' and be good again working on their same repo?
05:27 davidomanfredo joined #git
05:28 kadoban It's not as simple as them pulling. Recloning would be the easy way, there's also ways to recover from a peer's history rewrite using 'git reset --hard', etc.
05:32 Brski joined #git
05:33 pijiu2 joined #git
05:33 HurricaneHarry joined #git
05:33 sz0 joined #git
05:35 lordjancso joined #git
05:35 venmx joined #git
05:36 dersand joined #git
05:37 xall joined #git
05:37 tristanp joined #git
05:39 Gurkenglas joined #git
05:40 ayogi joined #git
05:44 todd_dsm joined #git
05:45 Cabanossi joined #git
05:48 davidomanfredo joined #git
05:49 xall joined #git
05:50 pijiu3 joined #git
05:51 spm_draget kadoban: So I should tell them to re-clone I guess. The thing is… the repo is HUGE. Re-cloning takes a while and is not that easy to organize with 5+ external people also working on it x.x
05:51 spm_draget I need a different strategy =/
05:52 Junior joined #git
05:52 pijiu3 joined #git
05:53 pijiu3 joined #git
05:55 a_thakur joined #git
05:58 MattMaker joined #git
05:59 grock joined #git
06:03 ThomasLocke joined #git
06:05 visual hey /gits/
06:06 visual all i can find regarding diffing remote & local, are files
06:06 visual git diff origin/master -- /file_name/
06:06 visual but what abotu all changes?
06:06 tarkus joined #git
06:07 regedit joined #git
06:08 rscata joined #git
06:08 Nilesh__ joined #git
06:09 osse visual: just remove "-- ..."
06:09 lipsumar joined #git
06:09 visual :O!
06:09 visual thanks osse, worked like a charm
06:11 a_thakur_ joined #git
06:11 Gurkenglas joined #git
06:12 xall_ joined #git
06:14 JeroenT joined #git
06:15 Cabanossi joined #git
06:15 zeroed joined #git
06:19 grock_ joined #git
06:19 XenophonF joined #git
06:19 adymitru1 joined #git
06:19 Anja_ joined #git
06:19 Starcraftmazter joined #git
06:19 mmlb_ joined #git
06:19 pks joined #git
06:19 batrick_ joined #git
06:19 lucido joined #git
06:19 watabou joined #git
06:19 miha_S7_ joined #git
06:19 a_thakur joined #git
06:19 ThomasLocke joined #git
06:19 ThomasLocke joined #git
06:19 malide joined #git
06:19 onehrxn joined #git
06:20 sz0 joined #git
06:20 timwis joined #git
06:20 d^sh joined #git
06:20 thallada joined #git
06:20 pifon joined #git
06:21 dmj` joined #git
06:21 yena joined #git
06:21 justan0theruser joined #git
06:22 User458764 joined #git
06:22 jay-m joined #git
06:23 tnecniv joined #git
06:23 Lowell joined #git
06:24 roelmonnens joined #git
06:24 roelmonnens joined #git
06:26 h12o joined #git
06:28 freimatz joined #git
06:30 qt-x joined #git
06:31 roelmonn_ joined #git
06:31 hussam joined #git
06:33 frojnd joined #git
06:33 frojnd joined #git
06:33 frojnd Hi there.
06:33 chele joined #git
06:35 frojnd I had a local branch that I set origin for, let's call it "supervisors" I was working in it and then I decided to merge it to develop. So I did: git co develop. git pull (inside develop) and finally git merge supervisors. The problem is I see many commits from supervisors branch not just "merge" commit unfortunatelly I already pushed it. Did I make a big mistake?
06:35 visual no
06:35 visual by default it shows all the commits from the source branch
06:36 telephone joined #git
06:36 telephone joined #git
06:36 frojnd visual: aha
06:36 visual if you want the target branch to only show 1 commit, look up "commit squash"
06:36 frojnd ah ok, I think I heard about squash
06:37 visual np, gitlab has this functionality as a checkbox thing
06:37 iamashishkebab2 joined #git
06:37 visual http://prntscr.com/f5d6fi
06:37 jozwior joined #git
06:39 frojnd visual: that's weird,.. I don't see my changes in one file that I was working on my local branch (also set origin)
06:39 frojnd visual: ah forget it
06:39 frojnd visual: looking at wrong file :P
06:41 visual :D happens
06:42 fluffystub joined #git
06:42 jay-m joined #git
06:43 infernix joined #git
06:45 Cabanossi joined #git
06:46 imack joined #git
06:48 _ng joined #git
06:49 ertes-w joined #git
06:50 QwertyDragon joined #git
06:51 Doginal joined #git
06:53 schleppel joined #git
06:53 xaa__ joined #git
06:54 rvgate joined #git
06:54 dvaske joined #git
06:57 JeroenT joined #git
06:58 murii_ joined #git
07:00 albel727 joined #git
07:01 TomyWork joined #git
07:04 elect joined #git
07:08 xall_ joined #git
07:09 a_thakur joined #git
07:09 dvaske_ joined #git
07:15 Cabanossi joined #git
07:16 leeN joined #git
07:19 pbandark joined #git
07:20 arbu joined #git
07:20 ertes joined #git
07:23 User458764 joined #git
07:24 Levex joined #git
07:26 Puffball joined #git
07:27 jackrandom joined #git
07:31 sevensidedmarble joined #git
07:33 lipsumar joined #git
07:34 prg3 joined #git
07:35 kelvah joined #git
07:35 davidomanfredo_ joined #git
07:36 JeroenT joined #git
07:40 djb-irc joined #git
07:45 chardan joined #git
07:45 mSSM joined #git
07:45 zulutango joined #git
07:45 MattMaker joined #git
07:45 Levex joined #git
07:47 Introoter joined #git
07:48 apotry joined #git
07:50 miha_S7_ joined #git
07:52 sea-gull joined #git
07:53 wandering_vagran joined #git
07:54 sbasso joined #git
07:54 acetakwas joined #git
07:55 davidomanfredo joined #git
07:55 miczac joined #git
07:56 lipsumar joined #git
07:56 blahdodo joined #git
07:57 JeroenT joined #git
07:57 prg3 joined #git
07:59 User458764 joined #git
07:59 JeroenT_ joined #git
08:01 marcogmonteiro joined #git
08:03 roelmonnens joined #git
08:04 cdown joined #git
08:05 MattE joined #git
08:06 manuelschneid3r joined #git
08:07 linuxmodder- joined #git
08:10 navidr joined #git
08:10 YuGiOhJCJ joined #git
08:12 dec0n joined #git
08:13 vamiry joined #git
08:14 jay-m joined #git
08:18 permalac Morning everyone, any suggestion on a GUI for my local repos ? (using ubuntu by now)
08:23 g5pw permalac: gitkraken is pretty :)
08:24 otisZart joined #git
08:24 permalac g5pw: can I use it with my own gitlab?
08:25 g5pw permalac: it doesn't have any assumptions on upstream, so yes :)
08:25 permalac nice, thanks. :)
08:28 _ADN_ joined #git
08:29 cads joined #git
08:31 dfadsf joined #git
08:31 Levex joined #git
08:31 MrWoohoo joined #git
08:32 sbasso joined #git
08:35 JeroenT joined #git
08:36 JeroenT joined #git
08:36 flaviodesousa joined #git
08:45 jay-m joined #git
08:47 User458764 joined #git
08:48 permalac g5pw: I like this gitkraken very much. thank you. :)
08:48 mischat joined #git
08:48 mentazoom joined #git
08:49 Vampire0 permalac, !guis
08:49 gitinfo permalac: Popular GUIs for Git are listed on the website: http://git-scm.com/downloads/guis
08:49 Vampire0 permalac, but remember, if you seek help here, you should use the commandline client. !gui
08:49 gitinfo permalac: Graphical user interfaces are not supported here. If you want to get support, it needs to be through the git CLI. Reasons: 1) Because very few people here use the graphical interface. 2) Because giving instructions for GUI's is difficult. 3) The command line gives you a history of what commands you have executed.
08:50 ToeSnacks Is there a recommended way to prevent git from saving credentials locally on windows?
08:50 permalac thanks, I should have googled. I did not want to start a discusion, is Monday morning and I am not thinking straight yet. :)
08:50 Vampire0 ToeSnacks, use SSH keys for authentication?
08:51 ToeSnacks I would like to use https if possible
08:51 Vampire0 ToeSnacks, besides that, Git for Windows should have the credential.help manager pre-set and thus store the credentials safely in the Windows authentication store
08:51 ToeSnacks Can it be set to never remember?
08:52 g5pw permalac: you're very welcome :)
08:52 xall_ joined #git
08:52 Vampire0 ToeSnacks, if you configure to not use any credential helper, then you are asked every time and nothing is stored locally
08:52 ToeSnacks I want the user to have to log in each time they perform a git action
08:52 Vampire0 ToeSnacks, man git config
08:52 gitinfo ToeSnacks: the git-config manpage is available at https://gitirc.eu/git-config.html
08:52 lipsumar joined #git
08:52 Vampire0 ToeSnacks, man git credentialhelper
08:52 gitinfo ToeSnacks: the git manpage is available at https://gitirc.eu/git.html
08:52 ToeSnacks Thank you I will look into that
08:52 Vampire0 hm, no, the last one was fwrong
08:52 Vampire0 ToeSnacks, man git credentia
08:52 gitinfo ToeSnacks: the git manpage is available at https://gitirc.eu/git.html
08:52 Vampire0 ToeSnacks, man git credential
08:52 gitinfo ToeSnacks: the git-credential manpage is available at https://gitirc.eu/git-credential.html
08:53 rchavik joined #git
08:53 Vampire0 ah, that one was it I think
08:53 ToeSnacks And as you just stated guis are not supported here right?
08:53 Vampire0 ToeSnacks, you can ask questions, but you will most likely not get much helpful comments
08:54 Tobbi joined #git
08:54 ToeSnacks Not sure if this is pushing that rule but do you know if that setting is adhered to by any of the guis?
08:54 mSSM joined #git
08:54 Vampire0 ToeSnacks, man git credentials as it actually
08:54 gitinfo ToeSnacks: the git manpage is available at https://gitirc.eu/git.html
08:54 Vampire0 ToeSnacks, man gitcredentials as it actually
08:54 gitinfo ToeSnacks: the gitcredentials manpage is available at https://gitirc.eu/gitcredentials.html
08:54 Vampire0 Now
08:55 ToeSnacks I have artists at work and they currently use SVN but I am trying to migrate to git with lfs
08:55 ToeSnacks Artists and the command line usually lead to pain
08:55 Vampire0 ToeSnacks, you can set the credential.helper to cache, so that the credentials are at least stored in cache for a short period, or unset it to require the credentials on each remote operation
08:56 ToeSnacks Cool I will read the links you gave
08:56 ToeSnacks Thank you
08:56 ToeSnacks Is git lfs supported at all here?
08:56 Vampire0 ToeSnacks, it could of course be, that a GUI does its own credentials storing beside the credential helper, I don't know
08:57 Vampire0 ToeSnacks, you're welcome
08:57 gitinfo ToeSnacks: This channel tracks karma based on who has gotten lots of thanks for being helpful. If you want to help someone reach karmic nirvana, please mention their name when thanking them with "thank you", "thankyou", "thanks", "thx", "ty" or "cheers". Try ".karma <nick>" or ".topkarma" to show karma status of a person but don't expect immediate increase. Ten thanks make up one karma point.
08:57 Vampire0 ToeSnacks, you can ask, if someone can help he will, I will not
08:57 ToeSnacks Thank you Vampire0
08:58 irqq joined #git
08:59 ToeSnacks I was mainly curious if anyone knew a command line script to return all binary blob file extensions in a directory to ease lfs tracking
08:59 cdown joined #git
09:00 marcogmonteiro joined #git
09:00 Cabanossi joined #git
09:02 cads joined #git
09:02 Starcraftmazter left #git
09:03 fatalhalt joined #git
09:04 romank joined #git
09:05 ams__ joined #git
09:05 D4R5C0D3 joined #git
09:06 glebihan_ joined #git
09:06 glebihan joined #git
09:07 Snugglebash joined #git
09:07 seni joined #git
09:07 Acerific joined #git
09:07 Acerific joined #git
09:08 tourdownunder joined #git
09:09 Gloomy joined #git
09:11 misosoup joined #git
09:14 marcogmonteiro joined #git
09:14 PasiFist joined #git
09:14 madduck joined #git
09:15 Brando753-o_O_o joined #git
09:16 ephimetheus joined #git
09:18 bronson joined #git
09:20 miha_S7_ joined #git
09:23 arussel joined #git
09:26 Snugglebash joined #git
09:26 JeroenT joined #git
09:31 D4R5C0D3 joined #git
09:32 Snugglebash joined #git
09:33 tvw joined #git
09:35 Stace joined #git
09:37 Stace Hi. If I'm on a detatched head and want to create a commit that I can cherry-pick from elsewhere, do I need to create a branch on that detatched head first?
09:37 Stace I'm probably going to make a branch in any case but was curious
09:38 sarri joined #git
09:38 sarri joined #git
09:38 maarhart joined #git
09:38 moritz you don't need to
09:38 skelterjohn joined #git
09:38 moritz you can just create a commit, and cherry-pick it by sha1
09:39 Stace moritz: ok. does the detatched head stay the same (on that previous commit) or move to the new one?
09:40 Stace I know the branch means I'm always on the head commit, but with the detatched head there is no branch
09:40 moritz Stace: move to the new commit
09:40 Stace ok
09:40 tobiasvl Stace: you're always on the head commit no matter what, by definition. a detached head is also a head
09:41 Stace so then why is creating commits on a detatched head a bad idea?
09:41 D4R5C0D3 joined #git
09:42 moritz because it's very easy to lose track of
09:42 tobiasvl they're not pointed to by anything, so they can easily become lost
09:42 tobiasvl !dangling
09:42 gitinfo Dangling objects represent things added to git which are no longer needed based on the git commands you typed in. This can be normal workflow (rebase, reset, add, etc) or errors you made. Typing "!dangling_commit" "!dangling_blob" and "!dangling_tree" into a query to me will get you more information about each type.
09:42 Stace oh, I got it
09:43 Stace so if i *do* decide to checkout master, the only place that the sha1 is going to be is in the log
09:43 Stace so working in the detatched head like this is just a very temporary place to work
09:43 Stace ... or just make a branch
09:44 tobiasvl yeah
09:44 tobiasvl just making a branch is the easiest. branches are cheap
09:44 Stace I'm using branches as a stach mechanism too now
09:44 tobiasvl yeah why not
09:44 Stace stash
09:44 Stace more control and stashing is messy in any case
09:45 kelvah joined #git
09:46 MattMaker joined #git
09:48 Xenophon1 joined #git
09:50 D4R5C0D3 joined #git
09:51 Vampire0 Stace, in what regard is stashing messy?
09:52 Stace Vampire0: sometimes I want to un-stash just one file
09:53 Vampire0 Stace, and this is easier with branches than with stashes because?
09:53 Stace It's probably just because I am more familiar with working with branches/making commits than stashing
09:54 Vampire0 Stace, a stash is basically a commit (actually two to differentiate changes in the worktree and on the index) where a ref is pointing to
09:54 Vampire0 Stace, so exactly the same as a branch
09:54 tvw joined #git
09:54 Stace I usually just make a branch, commit everything I would stash, switch back to master
09:55 Stace how do you stash everything but not one file?
09:56 Stace brb testing my erc config
09:57 Stace joined #git
09:59 Vampire0 Stace, I never needed this, but I'd say either `git add file && git stash save --keep-index` or `git stash save -p` should work
10:00 olso111 joined #git
10:00 Stace Vampire0: I've used keep-index before, when I apply the stash it conflicts with what's in my working tree. Don't know what -p does, perhaps I should take a look
10:00 Stace it's possible stashing is more convenient, perhaps I just don't know the commands well enough
10:01 Vampire0 Stace, -p is --patch, the same as for add or reset. You can choose for each hunk or file whether it should be stashed or not
10:01 theoceaniscool joined #git
10:01 Vampire0 Stace, dunno why you get conflicts just because of keep-index, shouldn't be I think
10:01 Stace oh, that's nice. I didn't know about that
10:02 Vampire0 yes, it is very helpful, I use it constantly for add and sometimes for reset
10:02 Stace I should give -p a try
10:03 Kaisyu joined #git
10:03 rudrab joined #git
10:03 olso111 Is there a git config that allows push to remote but only for concrete branch? "pushurl" works, but it targets whole repo. I'd like to only enable push to remote for one branch
10:03 Vampire0 Stace, one thing you should keep in mind though, stashes are for short-time temporary use
10:03 Vampire0 Stace, the latest stash is a ref, previous stashes are just the reflog of the stash ref
10:04 Vampire0 Stace, so they will be expired and garbage collected like any other reflog entry (after 90 days by default)
10:04 Stace Vampire0: yeah, I usually keep my stash empty. it messes up my git history otherwise
10:05 Stace oh, I didn't know they would be cleaned up with reflog. I usually clean the stash myself after using it
10:05 Stace otherwise I can't remember what is in there and if it's important
10:05 Timvde Is there any way to alias git push -f to use --force-with-lease by default?
10:05 Vampire0 olso111, if a remote is defined, you can push any ref to it. If you have the permission to push to the repository. If you need any restrictions you should implement them on the remote repo where push is done into
10:06 olso111 Vampire0, ok, so ill just protect the branch on server
10:06 Vampire0 Stace, sure, I usually use stash pop, that automatically throws it from the stash if there were no conflicts when applying
10:07 Vampire0 Timvde, no, you would need to invent a new name like `git pushf` which then expands to `git push --force-with-lease`
10:08 aielima joined #git
10:08 kelvah joined #git
10:08 D4R5C0D3 joined #git
10:09 ploop joined #git
10:10 jay-m joined #git
10:10 borkr joined #git
10:10 Stace I want to create a branch that's the same as my master except it doesn't have 2 commits in the history. What I've done is I've branched before those commits and then cherry-picked after them. This means that every time I do a master commit I must cherry-pick the commit to my "duplicate" branch. is there a nicer way of doing this?
10:11 King_Hual joined #git
10:12 Silenced joined #git
10:13 Vampire0 Stace, what is the purpose of this branch. Do you just want to not have the changes in there to test something or is it an alternative to master that will replace it, ...?
10:13 Stace Vampire0: the two commits are a tool upgrade. I want a branch without the tool upgrade. It won't replace it, it's just sometimes I need to to a build for the old tools with the same source
10:13 D4R5C0D3 joined #git
10:14 Snugglebash joined #git
10:14 Stace by "tool upgrade" I mean the software I'm using got upgraded and the project scripts for that software changed
10:15 Vampire0 Stace, how about simply doing a `git revert` for those two commits on the separate branch instead of eradicating them from the history
10:15 Vampire0 Stace, then you can merge or rebase as usual from master
10:16 arussel I've got a dev branch with A1, A2, A3, A4, A5, A6. I had to get few of it into master, so I've cherry picked some of: master: A1, A2, A5. We've got a policy that say that all master should be in dev, checking with 'git log dev..master'. But because I've cherry picked the commit, git  show A1, A2 and A5. Is there anyway to fix that ?
10:16 Stace Vampire0: that's a good idea
10:16 D4R5C0D3 joined #git
10:17 Stace I don't care if the history has a revert in it
10:17 Vampire0 arussel, merge or rebase from master to dev? as everything else is there already you will not get additional changes in and those commits will not change anything as they introduce the same changes
10:18 xall_ joined #git
10:18 arussel will give it a shot
10:18 kiltzman joined #git
10:19 elmcrest joined #git
10:19 bronson joined #git
10:19 Vampire0 arussel, you cannot ruin much as it is just local and easily reversible. :-) just !tryit
10:19 gitinfo arussel: [!tias] Try it and see™. You learn much more by experimentation than by asking without having even tried. If in doubt, make backups before you experiment (see !backup). http://gitolite.com/tias.html may help with git-specific TIAS.
10:20 Timvde Vampire0: I'm aware of that option, I was hoping to not having to break my muscle memory :)
10:20 Timvde But thanks for confirming
10:22 Vampire0 Timvde, well, you can maybe do it outside git by providing a git-push script that checks for -f, adds --force-with-lease and otherwise forwards to the real command
10:22 User458764 joined #git
10:22 gitinfo Timvde: This channel tracks karma based on who has gotten lots of thanks for being helpful. If you want to help someone reach karmic nirvana, please mention their name when thanking them with "thank you", "thankyou", "thanks", "thx", "ty" or "cheers". Try ".karma <nick>" or ".topkarma" to show karma status of a person but don't expect immediate increase. Ten thanks make up one karma point.
10:25 Timvde Eh, sure :P
10:25 Timvde ty Vampire0
10:25 miczac joined #git
10:28 iamashishkebab2 joined #git
10:30 ayogi joined #git
10:31 visual :!U
10:32 kelvah joined #git
10:33 varo joined #git
10:33 varo hello, here's a dumb question
10:34 varo is there an easy way to cherrypick a range of commits, but skip those that are already present in the branch?
10:34 varo i.e they were already cherrypicked into the branch in the past
10:34 Vampire0 varo, yes, just do it. What is already present will be skipped automatically
10:35 miczac joined #git
10:35 varo Oh, thanks Vampire0
10:35 Vampire0 varo, or if I remember wrongly it will at least say that the commit will be empty and then you can skip it
10:38 seni joined #git
10:38 silverdust joined #git
10:38 a_thakur joined #git
10:39 xsddz joined #git
10:39 lipsumar joined #git
10:39 silverdust I made a PR it got merged and I was told to make a minor fix. Is there a way to send the fix to the merged PR or I have to create a new PR for this?
10:39 _ikke_ If it's already merged, you need to make a new PR
10:40 xsddz left #git
10:40 Vampire0 silverdust, but you can do it from the same branch
10:40 silverdust Thank you _ikke_ . Was quite confusing. If it wasn't merged and I pushed to my forked origin it would automatically appear in the PR right?
10:40 _ikke_ yes
10:40 silverdust Okay great! Thanks y'all
10:41 earnestly joined #git
10:44 silverdust Since I'm here, I might just ask this already. I've never had to properly use rebase. This looks like a great explanation https://wincent.com/wiki/Git_rebase_explained except that could be done with git pull
10:45 dsdeiz joined #git
10:45 dsdeiz joined #git
10:45 _ikke_ silverdust: git pull = git fetch + (merge | rebase)
10:45 _chrispop_ joined #git
10:45 Vampire0 silverdust, git pull == git fetch + git merge or if specified by config or parameter git fetch + git rebase
10:45 silverdust So when do I really have to use rebase. If you may, please I'd like a scenario when I'm working alone and one wehn I'm working with a team
10:46 Vampire0 silverdust, the difference between merge and rebase is that the history looks like a ladder or linear
10:46 _ikke_ silverdust: It's mostly a matter of style
10:46 ColdKeyboard[a] joined #git
10:46 _ikke_ you have _have_ to use rebase
10:47 Snugglebash joined #git
10:47 _ikke_ but espescially when making pull requests, rebasing your branch on the latest changes from develop/master is nice because it lets you fix the conflicts instead of the one doing the merge
10:48 visual so wait, if saveOrFail() and save() just return false when they fail, how do you know the error?
10:48 _ikke_ visual: what?
10:48 visual like for instance you try to save a dublicate value for a column set as unique
10:48 visual it'll fail
10:48 silverdust So I could `git rebase upstream/master` instead of `git pull upstream master`
10:48 _ikke_ visual: #git <--
10:48 visual oh shit
10:48 visual :D
10:48 visual sorry
10:48 _ikke_ np
10:49 dvaske joined #git
10:49 silverdust Thanks again
10:50 miha_S7_ joined #git
10:52 masuberu joined #git
10:55 User458764 joined #git
10:57 iamasamosa joined #git
10:58 ferr joined #git
10:59 Vampire0 silverdust, if you fetch first, yes
10:59 Vampire0 silverdust, or you use --rebase parameter to the pull command
11:00 Cabanossi joined #git
11:02 sbasso joined #git
11:05 iamashishkebab2 joined #git
11:07 Sceorem joined #git
11:08 miczac joined #git
11:11 ayogi joined #git
11:11 chipotle joined #git
11:12 ayogi joined #git
11:13 Bookwormser joined #git
11:13 fahadash joined #git
11:15 digidog joined #git
11:17 dsdeiz joined #git
11:17 dsdeiz joined #git
11:17 _rht joined #git
11:18 marianina8 joined #git
11:20 bronson joined #git
11:22 jackrandom joined #git
11:22 d10n-work joined #git
11:24 cebor joined #git
11:26 Levex joined #git
11:28 HoierM joined #git
11:36 ToBeCloud joined #git
11:38 Stace oh, cool, so I made a detatched commit and then switched to master and it warned me " you are leaving 1 commit behind..."
11:38 Stace with the has
11:38 Stace hash
11:39 pitastrudl joined #git
11:39 Vampire0 Stace, sure, Git is pretty pedantic on warning about possible losses :-)
11:40 Vampire0 Stace, the same if you try to !float changes that would conflict
11:40 gitinfo Stace: If you have made a change in your working directory and have NOT YET COMMITTED, you may "float" that change over to another (`git checkout oldbranch`) or new (`git checkout -b newbranch`) branch and commit it there.  If the files you changed differ between branches, the checkout will fail.  In that case, `git stash` then checkout, and `git stash apply` and go through normal conflict resolution.
11:40 Snugglebash joined #git
11:40 Stace yeah I've been through that before
11:41 Stace I'm pretty happy on the git command-line. the only thing I'm bad at is conflict resolution. I usually use meld for that
11:42 Stace I haven't found it worth getting good at cli conflict resolution yet
11:42 Stace ... until I'm on a pc with no meld, that is
11:43 roelmonn_ joined #git
11:44 Vampire0 Stace, for conflict resolution I'd heavily recommend using P4Merge. It is the best free tool available that I know for diffing and merging. :-)
11:44 Vampire0 Stace, in my opinion of course
11:45 _pix joined #git
11:46 Stace Vampire0: hever heard of P4Merge before, I'll take a look at that
11:46 Seveas I just use vim :)
11:47 MattMaker joined #git
11:48 Stace oh, it's not in the debian repo. I see there is a linux release though
11:48 Vampire0 Stace, it is a client tool for Perforce. The client tools are free, only Perforce server users need to be paid for. So P4Merge can freely be used for any diffing / merging. The only major missing thing is diffing directories. For that I use kdiff3.
11:49 Gloomy joined #git
11:49 Vampire0 Stace, no, not in the repo. It is not free as in FOSS, but free as in no costs involved
11:49 Stace Seveas: I find myself using emacs and vim for different uses (neither for conflict resolution though) and the number of times I've typed :wq in vim... I really should pick one
11:49 Vampire0 Stace,  but there are installers for all major platforms on perforce.com
11:49 Stace I mean in emacs
11:49 Seveas just ditch emacs :)
11:50 Seveas emacs is for terrorists
11:50 Vampire0 Stace, in vim you can also type "ZZ" instead of ":wq", same effect, two keystrokes less :-)
11:50 Stace Seveas: typical vim user... I use it for the VHDL/Verilog mode tools which are *marvellous*
11:50 Stace oh, cool, I didn't know that
11:51 Stace my most recent find is Ctl-A
11:52 miha_S7_ joined #git
11:52 sbasso joined #git
11:53 wandering_vagran joined #git
11:53 sturner joined #git
11:53 Impaloo89 joined #git
11:54 Vampire0 Stace, what does it do?
11:54 apax joined #git
11:57 chrisM_1 joined #git
11:57 Chrism13 joined #git
11:57 Stace Vampire0: it increments the number on the line.
11:58 Stace I use it to increment my version number
11:58 chrisM_1 joined #git
11:58 Stace I believe Ctl-z decrements
11:58 Vampire0 In vim?
11:58 Stace oh, no, silly me
11:59 Stace Ctl-z kills the process
11:59 JanC joined #git
11:59 Vampire0 ah, interesting
11:59 Stace it's Ctl-a and Ctl-x increment and dec
11:59 Stace in vim
12:01 Vampire0 amazing. Not too useful for me I think, but interesting. And in edit mode it inserts something, not sure what though
12:01 Stace I usually do my version increments on cmd line so I use it daily
12:01 Vampire0 It inserts what you inserted before in the same edit session or something like that
12:02 Stace oh, cool
12:02 Stace I tried "ZZ" but it doesn't save and prompts for recovery when opened again
12:02 Stace I don't know if I'm doing it wrong
12:02 Vampire0 Stace, not Ctrl+Z
12:02 Stace I tried Shift-Z
12:02 Vampire0 Stace, just Shift+z+z
12:03 Vampire0 Stace, sounds strange
12:04 dglambert joined #git
12:04 dglambert yo
12:04 Vampire0 oy
12:04 Stace Vampire0: oh I got it working
12:05 Stace there was a swap file there from before that I missed
12:05 Stace deleted that and it worked. so it wasn't the ZZ after all
12:05 Vampire0 ok
12:05 Stace that's nice, thanks. I'll use it instead
12:05 synthroid joined #git
12:06 Vampire0 Stace, you're welcome
12:06 gitinfo Stace: This channel tracks karma based on who has gotten lots of thanks for being helpful. If you want to help someone reach karmic nirvana, please mention their name when thanking them with "thank you", "thankyou", "thanks", "thx", "ty" or "cheers". Try ".karma <nick>" or ".topkarma" to show karma status of a person but don't expect immediate increase. Ten thanks make up one karma point.
12:07 rivarun joined #git
12:07 Tobbi joined #git
12:08 romank joined #git
12:09 dvaske joined #git
12:09 Stace thanks, Vampire0 :p
12:11 lipsumar joined #git
12:12 skelterjohn joined #git
12:13 armyriad joined #git
12:13 re1 joined #git
12:19 ephimetheus joined #git
12:23 kelvah joined #git
12:24 kexmex joined #git
12:25 D630 joined #git
12:25 kristhian joined #git
12:26 kristhian somebody here familiar with hospitalrun.io?
12:26 Cavallari joined #git
12:27 Daviey joined #git
12:30 Vampire0 kristhian, what should that be?
12:30 circ-user-zNiD2 joined #git
12:31 mentazoom joined #git
12:32 sbasso_ joined #git
12:32 jason237 joined #git
12:33 bdesemb joined #git
12:34 Stumbler joined #git
12:34 robertparkerx joined #git
12:37 bdesemb Hi everyone. Is it possible to handle a subfolder which is a git repo like a normal subfolder and not like a submodule?
12:38 TomyWork joined #git
12:40 diogenese joined #git
12:40 maroloccio joined #git
12:43 northfurr joined #git
12:45 Vampire0 bdesemb, !subr
12:45 gitinfo bdesemb: [!subprojects] So, you want to add git repositories inside of other git repositories? Well, you have four main options. First is to just do it, add the repo to the outer project's .gitignore, and treat them entirely separately. Best if they are entirely separate. Otherwise your best options are "!submodule", "!gitslave", and "!subtree". Try those commands in this channel, or in a PM to avoid flooding.
12:45 Starky joined #git
12:45 Cabanossi joined #git
12:46 seni joined #git
12:48 Snugglebash joined #git
12:49 macivy joined #git
12:51 Guest18947 joined #git
12:53 h12o joined #git
12:53 jackrandom joined #git
12:55 banzajus joined #git
12:58 macrover joined #git
13:02 livingstn joined #git
13:03 ExoUNX joined #git
13:05 banzajus_ joined #git
13:06 nowhereman joined #git
13:07 notebox joined #git
13:07 jameser joined #git
13:10 Gurkenglas joined #git
13:10 macrover joined #git
13:10 karols146 joined #git
13:12 kurkale6ka joined #git
13:13 Snugglebash joined #git
13:15 Cabanossi joined #git
13:15 arbu left #git
13:18 _ng joined #git
13:20 madduck joined #git
13:21 bronson joined #git
13:21 fmcgeough joined #git
13:21 onehrxn_ joined #git
13:21 otisZart joined #git
13:22 dstw joined #git
13:24 onehrxn__ joined #git
13:24 finalbeta joined #git
13:26 Oatmeal joined #git
13:26 Silenced joined #git
13:27 TheSimonator joined #git
13:27 roelmonnens joined #git
13:27 onehrxn joined #git
13:27 theoceaniscool joined #git
13:30 sunri5e joined #git
13:31 bernd27 joined #git
13:32 Snugglebash joined #git
13:32 jimmyrcom_ joined #git
13:32 Murii joined #git
13:32 _ADN_ joined #git
13:33 dstw left #git
13:33 redoverture joined #git
13:35 BSAlb joined #git
13:35 kexmex joined #git
13:36 rts-sander joined #git
13:36 otiose joined #git
13:37 Dougie187 joined #git
13:37 raynold joined #git
13:39 hashpuppy joined #git
13:39 Silenced joined #git
13:39 onehrxn_ joined #git
13:41 styx_ joined #git
13:43 jon-mac joined #git
13:43 cdown joined #git
13:44 bsanford joined #git
13:45 regedit joined #git
13:46 sbasso joined #git
13:47 shinnya joined #git
13:47 guampa joined #git
13:48 MattMaker joined #git
13:49 leeN joined #git
13:50 jameser joined #git
13:52 sunri5e joined #git
13:52 tmsmith joined #git
13:54 Es0teric joined #git
13:54 sturner_ joined #git
13:55 tmsmith_ joined #git
13:56 mizu_no_oto joined #git
13:58 aard joined #git
13:58 davidomanfredo joined #git
13:58 Qilibrun joined #git
13:58 otiose joined #git
13:58 dglambert joined #git
13:59 romank joined #git
14:00 govg joined #git
14:02 AJ_Michels joined #git
14:03 sbasso_ joined #git
14:04 redoverture joined #git
14:04 prg3 joined #git
14:04 mda1 joined #git
14:08 MDA2 joined #git
14:08 sunri5e joined #git
14:10 h12o joined #git
14:10 HardlySeen joined #git
14:12 redoverture joined #git
14:13 weylin joined #git
14:13 leeN joined #git
14:13 pks joined #git
14:14 pks joined #git
14:15 cebula joined #git
14:17 davidfetter_ge joined #git
14:18 IRQ` joined #git
14:19 m4sk1n joined #git
14:21 rasto joined #git
14:25 visual !subtree
14:25 gitinfo The subtree merge method is great for incorporating a subsidiary git repo into your current one with "unified" history. Read http://git-scm.com/book/en/v2/Git-Tools-Advanced-Merging#_subtree_merge for more info, see also !git-subtree and !git-stitch-repo.
14:25 visual what was the one with multiple folders, same git repo
14:25 visual tree something
14:25 visual multiple folders belonging to multiple branches, but the same /repo/
14:25 davidomanfredo joined #git
14:26 _ikke_ visual: man git work-tree
14:26 gitinfo visual: the git manpage is available at https://gitirc.eu/git.html
14:27 _ikke_ visual: man git worktree
14:27 gitinfo visual: the git-worktree manpage is available at https://gitirc.eu/git-worktree.html
14:27 gtvz joined #git
14:27 User458764 joined #git
14:28 dansan yeah, I couldn't make it without git worktree anymore
14:28 kelvah joined #git
14:28 _ikke_ never used it
14:28 jameser joined #git
14:29 Ploppz joined #git
14:29 dansan Well, especially when you're at the testing phase of a project, where you want to create branches and check them out, but you don't want to have to rebuild your main worktree after all of those headers have their date changed from switching branches
14:29 misosoup joined #git
14:29 Ploppz when using submodules - each commit has information about which commit should be checked out in each submodule right?
14:29 _ikke_ Ploppz: correct
14:30 * dansan googling git submodules :)
14:30 dansan oh yeah!
14:30 _ng joined #git
14:30 dansan yeesh, I remember doing that in svn.....
14:30 dansan Does anybody know why gcc hasn't switched to svn yet? :)
14:30 dansan I haven't dared to actually ask in #gcc :)
14:31 dansan err, I mean FROM svn (to git)
14:31 visual dansan: how do you handle multiple branches = 1 server?
14:31 Vampire0 dansan, maybe because they done *want* to? :-D
14:31 visual multiple working directories -> sftp to 1 server?
14:31 cdg joined #git
14:31 dansan visual: Well you only havwe the worktrees on your client
14:31 prg3 joined #git
14:31 visual yse
14:31 visual yes
14:32 visual but to test them out, you have to upload each new change
14:32 dansan visual: you first clone.  Then from within that tree, you say "git worktree add <path> [branch]"
14:32 dansan You cannot have the same branch checked out in multiple worktrees
14:33 visual i understand, my question is how do you not make a mess of things on the other end?
14:33 visual other end = remote server (considering you do the testing on remote)
14:33 miczac joined #git
14:33 visual or local server, doesnt matter, it still points to 1 folder
14:33 Gloomy joined #git
14:34 visual but in order to see the /changes/, you need to upload the changed files from your current working branch to the server, so my question is
14:34 visual how do you not make a mess of things?
14:35 JeroenT_ joined #git
14:35 fluffystub joined #git
14:35 visual like you change and upload 2 files from branch 1 (B1) / Working Tree 1 (WT1), then switch to B2/WT2, change some other file, upload it to the server
14:35 dansan visual: Well you would have to cherry pick it, rebase your main tree or something
14:35 visual in the end, the server will end up with a mess of a merge between the branches
14:36 visual cherry pick what?
14:36 tazle_ joined #git
14:36 dansan A working tree is just a way to have multiple (local) branches checked out on the same machine w/o cloning (and wasting extra space) and then having to push your changes back up to somehwere in order to access them from your other worktrees
14:36 redoverture joined #git
14:36 dansan visual: No, you can merge locally
14:36 sbasso joined #git
14:36 visual i hear you, so i guess it's not as much of a working tree question, but more of a branching question in general
14:37 dansan Oh yes
14:37 telephone joined #git
14:37 telephone joined #git
14:37 dansan worktrees do not alter how that works at all
14:37 visual assuming local machine has the repository, with branches and stuff, and the server is remote (or local for that matter)
14:37 visual when you switch branches, the server doesnt switch any branch
14:37 dansan correct!
14:38 dansan Well, you don't really ever switch branches on the server
14:38 thiago joined #git
14:38 visual in the end, while your local files and repo might be sorted and shit, the server isnt, so you end up with a mish-mash of changes uploaded to the server
14:38 visual exacltly
14:38 borkr joined #git
14:38 dansan You simply check out a branch *from* the server, but the server doesn't care what branch you're on, it's never "on" a branch
14:38 varo vishal: the remote also has branches, it doesn't mix them together when you push
14:38 visual what do you mean you check out a branch from the server?
14:38 dansan --- well I need to clarify!! I'm assuming that you're server is "bare"
14:38 varo though I'm 99% sure you're trolling
14:39 visual usually the server (prod), doesnt even have a git repo on it
14:39 dansan visual: git checkout <remote>/<branch>
14:39 visual varo: im not talking about remote _repo_, but remote _server_
14:39 visual apache
14:39 varo oh lol
14:39 visual sorry if i left that ambiguous ))
14:40 dansan visual: I'm only talking about git.  If the server doesn't have a git repo then it can't be a git server
14:40 varo why would you upload more than 1 branch to a prod server?
14:40 visual you dont
14:40 visual idk what you call that server, just.. a server that you work on
14:40 dansan varo: there are a lot of good reasons for that
14:40 onehrxn joined #git
14:40 visual not prod, not dev, not stage, but all 3 kind of
14:40 dansan varo: because you have more than one person working on a logical branch that you aren't ready to merge
14:41 varo dansan: he's not talking about git servers
14:41 njbair joined #git
14:41 visual what im asking is, how do you keep track of the changes uploaded to the server?
14:41 dansan varo: lol
14:41 visual not git servers, apache
14:41 visual or nginx, the remote metal machines ))
14:41 dansan visual: That's a WHOLE different topic of production change control
14:41 visual i feel you
14:41 overlord_tm joined #git
14:42 visual soooooo
14:42 dansan I spent many years at AT&T and GTE (now verizon) where they had fairly good change control proceedures.  That's not a topic for #git really
14:42 visual dansan: which channel would be more appropriate for it then?
14:42 dansan unless I misunderstand (I've only had half a cup of coffee) :)
14:43 visual well it's half related to git
14:43 dansan I dunno really, I can try to help point you to some good resources.
14:43 visual because im asking how you keep track of the git branches/changes that get uploaded to server :I
14:43 dansan Anybody know of some good books on the topic of change control
14:43 dansan ?
14:43 visual change control eh
14:43 visual never heard of it tbh
14:43 mischat_ joined #git
14:43 visual oh, like jenkins
14:44 dansan visual: Well, you know the general theory of versionining and software development cycle and releases I presume?
14:44 visual yeah
14:44 hsiktas joined #git
14:44 dansan Usually, you start with a branch and developers work on things and submit patches and/or pull requests and you add commits to that branch, test, etc.
14:44 mischat_ joined #git
14:44 visual vaguely, altho i've been doing it for many years
14:45 dansan Then you declare at some point that you are only comitting bug fixes to that branch
14:45 dansan You continue to fix bugs and eventually decide to put out a release candidate "RC"
14:45 Cabanossi joined #git
14:45 UTAN_dev joined #git
14:45 visual dansan: yeah, or you can use jenkins, write unit tests, and push to prod only after those branches pass the tests
14:45 dansan You have your guys test that like hell, maybe fix a few bugs and put out an RC1
14:45 dansan haven't heard of jenkins
14:45 visual :O!
14:45 mischat_ joined #git
14:45 dansan :)
14:46 visual google it, it's some magical thingy
14:46 dansan Anyway, you generally use "git tag" to declare a release candidate
14:46 visual ok pls continue
14:46 dansan lol!
14:46 dansan finally, you decide that it's great and you tag it "release-1.0"
14:47 Vampire0 Jenkins is not a magical thing, just a quite popular build / CI server like there are many others out there. I like much better TeamCity
14:47 dansan From there, you probably will create a maintence branch "something-1.x"
14:47 davidfetter_ge left #git
14:47 dansan So you start a new development branch for "something-2.x" (dependent upon your versioning scheme).  New features go there, bug fixes to go "something-1.x"
14:48 dansan You fix a lot of bugs after customers scream, cry, etc and release something-1.1, etc...
14:48 dansan I think that's the general theory
14:49 dansan You'll usually end up backporting bugfixes from your something-2.x branch into something-1.x
14:49 dansan Oh, that "something-1.1" is a TAG, not a branch
14:49 dansan I said "release-1.0" earlier, inconsistient, sorry
14:49 visual oh man now im lost
14:49 visual ))
14:50 dansan tags aren't suposed to move.  A branch changes it's reference every time you make a commit
14:50 visual when you release it, and start the something-2.x branch
14:50 dansan yeah
14:50 visual and at the same time you start a tag, not a branch, of 'maintainence-1.x'
14:50 visual right?
14:50 dansan Your "something-2.x" branch is for development of your next version.  That way you have new feature development (which can easily break it) and bugfixes
14:51 dansan Yes, I just changed my naming, sorry
14:51 visual that is so cool
14:51 dansan oh wait
14:51 dansan no
14:51 visual >inb4 another name change
14:51 varo what a clusterfuck
14:51 dansan lol!!
14:51 dansan your "maintainence-1.x" is a branch that changes
14:51 Starky joined #git
14:51 davidomanfredo joined #git
14:52 dansan but if you're project is called "jack", it would be something like "jack-1.x", or "jack-1"
14:52 Meow-J joined #git
14:52 dansan or just "branch-1.x"
14:52 dansan as long as you have sorted out what your naming shceme will be and are consistient (unlike I was :) it doesn't matter what you call it
14:53 visual got it
14:53 visual )
14:53 visual or
14:53 visual or
14:53 visual if there's 1 person working per project
14:53 visual is there any reason at all to use branching?
14:54 dansan The same can apply for good organization
14:54 hobodave joined #git
14:54 visual yes but if you use branching, you need to separate the dev server from the prod server, right?
14:54 dansan And if you're working on new features, but then you're boss comes back and says "omfg, fix this bug"
14:54 TbobbyZ joined #git
14:54 jeffreylevesque joined #git
14:54 dansan no
14:54 dansan git server?
14:54 varo left #git
14:54 visual no, http server
14:54 dansan oh, yeah!
14:54 onehrxn joined #git
14:55 TbobbyZ joined #git
14:55 dansan you don't HAVE to seperate them... but VERY good idea to do so!
14:55 visual why?
14:55 visual in layman terms))
14:55 dansan I suppose it depends upon how important your production http server is :)
14:55 dansan Because in development, nobody can hear you scream.  If you crash your production server, the whole world hears it.
14:56 visual i hear you
14:56 visual but in this scenario, a server you show your clients the changes
14:56 visual what is it?
14:56 visual not the one that hosts the final product, or even the running product, just one to show your clients changes
14:56 dansan "staging"
14:57 dansan You generally have "development", "staging" and "production"
14:57 visual staging eh
14:57 dansan yup
14:57 plos joined #git
14:57 D4R5C0D3 joined #git
14:57 plos joined #git
14:57 dansan Maybe read git pro
14:58 visual how..bad of an idea is it to have the dev & staging on the same machine?
14:58 visual git pro, eh
14:58 dansan err, pro git: https://git-scm.com/book/en/v2
14:58 visual oh, i read it
14:58 dansan oh! there's a second edition!
14:58 onehrxn_ joined #git
14:58 dansan I thought it was in the topic, but I don't see it anymore
14:58 dansan oooh! git jokes
14:59 UTAN_dev joined #git
14:59 dansan Well thanks visual, you helped me repay some of my karma for all of the awesome people that have helped me out here :)
14:59 onehrxn__ joined #git
14:59 mischat joined #git
14:59 visual haha, my pleasure man, always fun to meet somebody with the patience to explain stuff the way you did
14:59 d10n-work joined #git
14:59 visual alright, it's 6 pm and im out of here, thanks again and have a good one
15:00 dansan see ya!
15:00 visual bye!
15:01 roelmonn_ joined #git
15:03 misosoup sigh.. I put ".vs" in my .gitignore but .vs/slnx.sqlite is still listed in git status. What I do wrong
15:03 redoverture joined #git
15:04 Vampire0 misosoup, is this file already tracked?
15:04 mizu_no_oto joined #git
15:04 redoverture_ joined #git
15:04 misosoup oh, maybe. Yeah it is
15:04 Vampire0 misosoup, !ignore_tracked
15:04 gitinfo misosoup: Git only applies ignore patterns to untracked files. You can't use ignore patterns to ignore changes to files that are already tracked by git. To remove files only from git, but keeping them on disk, use git rm --cached <file>. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.
15:05 al-damiri joined #git
15:05 psprint What do you guys think about this? https://github.com/zdharma/giturl
15:05 theoceaniscool joined #git
15:06 misosoup vamiry, thanks
15:06 misosoup * Vampire0
15:07 redoverture_ joined #git
15:08 Vampire0 misosoup, you're welcome
15:08 gitinfo misosoup: This channel tracks karma based on who has gotten lots of thanks for being helpful. If you want to help someone reach karmic nirvana, please mention their name when thanking them with "thank you", "thankyou", "thanks", "thx", "ty" or "cheers". Try ".karma <nick>" or ".topkarma" to show karma status of a person but don't expect immediate increase. Ten thanks make up one karma point.
15:08 Vampire0 psprint, I don't understand what use this should have
15:08 Vampire0 psprint, it is an unreadable version of some data to spare some characters, but whatfor?
15:09 Vampire0 psprint, to confuse people?
15:09 northfurr joined #git
15:10 psprint Vampire0: ok, you might be right. I would account a thing – you have consistency in single place, normally you need to read sentences and verify if everything is in check, here you just grab single string, and know it doesn't lack or confuse something
15:11 redoverture joined #git
15:12 Vampire0 psprint, but you cannot really determine whether it references what you are told it references except by using an additional command. I don't like using URLs with encrypted information I don't understand
15:12 Vampire0 psprint, besides that using that many characters in the encoding could lead to errors in copy / paste due to encoding stuff
15:13 psprint that's true, when storing such url a description is also required, but it can be free language not very carefully written
15:13 psprint Vampire0: utf copy pasting is quite rock-solid today, I have choosen unicode *letters* so that double clicking selects whole thing, doesn't stop on a symbol
15:14 marianina8 joined #git
15:15 Cabanossi joined #git
15:15 marianina8 joined #git
15:15 Wulf joined #git
15:16 Es0teric joined #git
15:16 Vampire0 psprint, ah, so you're the author, I see. :-) Sorry then if I offensed :-D
15:16 cyphase joined #git
15:16 psprint no problem, I was looking for true opinions
15:17 ableto joined #git
15:17 Vampire0 psprint, utf copy pasting might be rock-solid, but the webpage using that link has to reliably specify the page encoding and the browser has to reliably use it for displaying, ... I'm just not too confident it will work seamlessly, but I might be wrong
15:17 redoverture joined #git
15:18 Vampire0 psprint, regarding the description, yes, you need a description besides the URL, but what I meant is, that I do not trust a textual description for an encrypted URL. Before I can use it without doubt I first have to decode it and verify the information in it visually.
15:18 Vampire0 psprint, but that might just be my paranoia :-D
15:19 guardianJ joined #git
15:19 reactormonk[m] joined #git
15:20 TikityTik joined #git
15:20 psprint yes that are problems I'm not sure how would be using this with good support, like web browser extensions, windows app, *nix gui app not only terminal, and suppose even github support, an extreme scenario, and still unsure if it would be nice to grab repo data this way
15:20 Retropikzel[m] joined #git
15:21 elwan7[m] joined #git
15:21 monomon[m] joined #git
15:21 dsdeiz joined #git
15:21 dsdeiz joined #git
15:21 kirb joined #git
15:21 kirb joined #git
15:22 bronson joined #git
15:22 Vampire0 psprint, btw. what is "Revision"? I only know that term in SVN context, not in Git context
15:23 psprint I thought about such thing: have refs/patches/patch[0-9]+, encode such revision in the gcode, send patches this way, not via inlining / attaching
15:23 psprint also, I can encode 1st, 2nd, ..., commit in the URL, I mean: encode revision REV, then select 1st, 3rd commit starting from that revision, and send to someone – "here are the commits in question", but that's uncommon
15:23 psprint Vampire0: branch, tag, SHA
15:23 tarkus joined #git
15:23 jzeus joined #git
15:24 psprint maybe a "ref" is more git-word
15:24 Vampire0 psprint, more a commit-ish, so maybe "commit". "ref" would not include using explicit SHAs
15:25 raspado joined #git
15:28 imanc joined #git
15:28 marianina8 joined #git
15:29 marianina8 joined #git
15:29 Emperor_Earth joined #git
15:29 Kronuz joined #git
15:29 bdesemb joined #git
15:29 kyan joined #git
15:32 s1scha joined #git
15:33 thecomedian joined #git
15:34 Hello71 a SHA is a ref
15:34 Hello71 sort of.
15:34 lindenle joined #git
15:34 Hello71 you could argue that a SHA is a ref, and "master" is the name of a ref
15:34 _ikke_ a SHA is not a ref
15:35 _ikke_ a ref is a pointer to a commit (hash)
15:35 Gustavo6046 joined #git
15:35 Vampire0 Hello71, a sha is a sha. A ref has a name and points to something
15:35 Hello71 but a SHA is also a pointer to a commit (object)
15:36 _ikke_ It's more the identity of a commit, not a pointer
15:36 osse no, it's the name of a commit object!
15:36 osse let's fight!
15:36 Vampire0 Hello71, https://git-scm.com/book/en/v2/Git-Internals-Git-References
15:36 Vampire0 Hello71, first two paragraphs
15:36 _ikke_ A pointer can be made to point something else, a hash is tied to that specific object
15:36 Rxi joined #git
15:36 osse not a const std::unique_ptr
15:37 Random832 git documentation never uses "ref" to refer to command line arguments where a hash is permitted - most often it says <commit> or <tree-ish>
15:37 Vampire0 osse, ah, just make a stack overflow and change it with some embedded code :-D
15:38 Vampire0 It's as simple as this: Everthing that can potentially be in .git/refs is a ref, hence its name (can potentially be in it, because it could be stored in a pack file)
15:39 svm_invictvs joined #git
15:39 gaucheph joined #git
15:40 _ikke_ Vampire0: refs can even live outside of refs/
15:40 _ikke_ (git update-ref test <hash> creates a ref)
15:41 Vampire0 _ikke_, ah, right, HEAD, FETCH_HEAD and alike are also outside refs/
15:41 _ikke_ right
15:41 brent__ joined #git
15:41 _ikke_ though FETCH_HEAD is not really a ref
15:42 brent__ joined #git
15:43 Vampire0 _ikke_, why not?
15:43 telephone joined #git
15:43 telephone joined #git
15:43 _ikke_ check its contents
15:44 hobodave joined #git
15:45 theresajayne joined #git
15:45 ivo_ joined #git
15:45 Wulf joined #git
15:46 boombatower joined #git
15:47 PHPanos joined #git
15:48 davidomanfredo joined #git
15:48 Vampire0 _ikke_, it has additional information like branch and URL, but  it still is a symbolic name that points to a SHA, isn't it?
15:49 Vampire0 _ikke_, so I'd say it is a special ref with more info, but still a ref
15:49 reactormonk[m] joined #git
15:49 MattMaker joined #git
15:49 _ikke_ yeah, you can treat it like one
15:51 hobodave_ joined #git
15:51 osse FETCH_HEAD might have several
15:52 TomyLobo joined #git
15:53 Vampire0 osse, several what? referencees?
15:53 osse lines
15:53 osse not sure what you'd call them :p
15:53 Vampire0 osse, ok, didn't know that. Which one is used if you use FETCH_HEAD in place of a commit-ish?
15:53 osse no idea
15:53 livingstn joined #git
15:54 osse Vampire0: the only line that does NOT say "not-for-merge" is my guess
15:54 mikecmpbll joined #git
15:54 osse I moved the line down and now rev-parse says something else.
15:55 bronson joined #git
15:55 mloy joined #git
15:56 mel00010 joined #git
15:56 Hogofwar joined #git
15:57 synthroid joined #git
15:58 Vampire0 hm
15:58 macrover joined #git
15:58 Introoter joined #git
15:59 kegster joined #git
15:59 osse huh, if I read this code right what it does is read the first 40 bytes of the file
16:00 osse "the file" = .git/argument  after all the other checks fail
16:01 Vampire0 so, I'd still say it *is* a ref with additional information :-D
16:01 Vorap joined #git
16:01 durham joined #git
16:02 osse at this point, what isn't?
16:02 leeN joined #git
16:02 Vampire0 huh?
16:02 durham_ joined #git
16:02 livingstn joined #git
16:02 osse It's like the opposite of "no true scotsman"
16:03 thiago joined #git
16:03 Vampire0 I also do not know that phrase, sry. I'm not a native speaker
16:04 zmachine joined #git
16:04 mizu_no_oto joined #git
16:04 osse Vampire0: i mean if you count this as a ref and that as a ref, what *don't* you count as a ref? :p
16:04 YuGiOhJCJ joined #git
16:04 madewokherd joined #git
16:05 marianina8 joined #git
16:05 livingst_ joined #git
16:06 rogl joined #git
16:06 TbobbyZ_ joined #git
16:07 davidomanfredo joined #git
16:09 Gurkenglas joined #git
16:10 dreiss joined #git
16:10 l2y joined #git
16:11 jozwior joined #git
16:12 finalbeta joined #git
16:14 Archrover joined #git
16:14 publio joined #git
16:15 Cabanossi joined #git
16:16 heftig joined #git
16:16 ajeremias joined #git
16:16 saintaquinas[m] joined #git
16:16 am2on joined #git
16:16 deepakbn[m] joined #git
16:16 flo[m] joined #git
16:16 KevinMGranger joined #git
16:16 dyce[m] joined #git
16:16 M-meznak joined #git
16:16 unclechu joined #git
16:16 Elliott[m]1 joined #git
16:16 jascot[m] joined #git
16:16 timlyo[m] joined #git
16:16 swalladge[m] joined #git
16:16 aviraldg joined #git
16:16 SpEcHiDe joined #git
16:16 Remramm joined #git
16:16 musicmatze[m] joined #git
16:16 dikiaap joined #git
16:16 monomon[m] joined #git
16:16 Vampire0 osse, I don't count as ref what not fits the definition of a ref. A symbolic name that refers to a commit-ish or tree-ish
16:16 dbn[m] joined #git
16:16 UncleCJ joined #git
16:16 logos[m] joined #git
16:16 patrickr[m] joined #git
16:16 Retropikzel[m] joined #git
16:16 elwan7[m] joined #git
16:16 waltervargas[m] joined #git
16:16 hiq[m] joined #git
16:16 spinningarrow[m] joined #git
16:16 Satchmo1[m] joined #git
16:17 Atm0spher1c joined #git
16:18 circ-user-zNiD2_ joined #git
16:18 otiose joined #git
16:19 lipsumar joined #git
16:19 jason237 joined #git
16:21 xall_ joined #git
16:21 nickenchuggets joined #git
16:21 LionsMane joined #git
16:21 halftroll joined #git
16:22 halftroll Hello, I tried to change branch and users permissions of the filesystem were wrong.. now things went a little crazy :\
16:22 ojdo joined #git
16:22 NeverDie joined #git
16:23 mischat joined #git
16:23 TbobbyZ joined #git
16:23 kadoban halftroll: If you fix your permissions, there's usually not anything more you have to do except repeat what didn't work.
16:23 FuzzyWuzzyPanda joined #git
16:23 d10n-work joined #git
16:23 nitric joined #git
16:24 halftroll kadoban: there's a lot of files removed.. so should I just change branch and then change back ?
16:24 macrover left #git
16:25 kadoban Sounds like a plan. Or you can just checkout the files, I assume you had no uncommited changes when you were trying to switch branches?
16:26 jzeus joined #git
16:26 bronson joined #git
16:27 marianina8 joined #git
16:27 halftroll kadoban: that's right, thanks a lot !
16:27 Bookwormser joined #git
16:28 kadoban 'welcome. If changing branches back and forth doesn't work, it should just be  'git chekcout -- ." (note the .), which is safe if you have no uncommited changes.
16:28 rkazak joined #git
16:28 ams__ joined #git
16:30 otiose joined #git
16:32 SteffanW joined #git
16:32 Sound joined #git
16:33 bongjovi joined #git
16:34 zarubin joined #git
16:35 tobie joined #git
16:35 PowaBanga joined #git
16:36 PowaBanga hi all
16:36 gitinfo PowaBanga: 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.
16:36 StuartMI joined #git
16:37 rpi joined #git
16:37 Doginal joined #git
16:39 Sound_ joined #git
16:39 cseder joined #git
16:40 halftroll kadoban: changing back and forth did work!, thanks again :)
16:41 PowaBanga I have a problem,  with git, so it sayd that it was work that i don't have in local, so i did "git pull origin master", but at the end, it sayd "Fatal: refusal to merge unrelated histories"
16:41 PowaBanga and I can't use push -u origin master, i have the same error
16:41 PowaBanga how to reslve that ?
16:41 thiago PowaBanga: the two repositories don't have common history
16:41 thiago is this a new repository you've just created?
16:42 endiruna joined #git
16:42 marianina8 joined #git
16:42 PowaBanga no i just did a other push there is 2 minutes
16:42 hobodave joined #git
16:42 PowaBanga without problem
16:42 kyan_ joined #git
16:44 mischat_ joined #git
16:46 Es0teric how do i go back one commit on local? is it git reset HEAD~ ?
16:46 Gustavo6046_ joined #git
16:46 thiago PowaBanga: then confirm you've got the right URL for the right project
16:47 thiago Es0teric: define "go back": permanently, or just to see what it was?
16:47 otiose joined #git
16:47 PowaBanga yeah thank you, I just modifer the url in config file, it was wrong
16:47 PowaBanga now it work, thank you
16:47 Es0teric thiago perm
16:47 Es0teric this same commit is on origin so i want to go back on origin as well
16:47 thiago Es0teric: yeah, reset. You can use --hard to make git check out the commit after it's done resetting.
16:47 Es0teric so git reset --hard HEAD~ ?
16:48 lafleurdubien joined #git
16:48 Es0teric thiago
16:49 peepsalot joined #git
16:52 thiago Es0teric: yes
16:52 Es0teric thiago ok i have a problem though... for some reason there is a folder that registers as a git submodule
16:52 Es0teric how do i remove it?
16:53 thiago the folder or the fact that it's a submodule?
16:53 Tobbi joined #git
16:54 circ-user-zNiD2 joined #git
16:54 mda1 joined #git
16:55 otiose joined #git
16:58 sea-gull joined #git
16:59 kexmex joined #git
16:59 cjohnson If you cloned a git repo inside another repo it'll detect it as you trying to add a submodule
17:00 cjohnson if you're just wanting to add the other repo's files to the parent repo you can delete .git dir in the subdirectory
17:00 grawity or do `git add inside/dir/`
17:00 grawity the trailing slash is interpreted as "add contents"
17:00 cjohnson I don't think leaving behind a .git dir is a good idea though
17:01 grawity depends on whether you'll want to update it in the future
17:01 cjohnson files belonging to two repos seems wrong
17:01 grawity I do that with some vim plugins under ~/dotfiles/
17:01 cjohnson you heathen
17:02 dbb joined #git
17:03 dbb hi - my project lead has insisted we use Github..  any current "cheat sheet"  for useful git <-> Github.com tips?
17:03 Xe don't overthink your use of github
17:04 jaziz joined #git
17:04 cjohnson dbb: just use github as your origin copy of your git repo
17:04 cjohnson use the github features if/when you feel like it
17:04 thiago dbb: git is git anywhere
17:04 cjohnson the only thing i would really brush up on is how to use pull requests
17:04 dbb thiago: that is completely not the case
17:04 thiago dbb: any github.com cheat sheets would be related to how you use the web interface
17:04 cjohnson nah it is. github is just hosting a standard git repo
17:05 cjohnson it just provides some pretty UI around some of it
17:05 dbb there is no CLI access
17:05 thiago dbb: if your project uses pull requests, if you use their task lists, etc.
17:05 cjohnson if you've used git you know how to use github
17:05 cjohnson dbb: yes there is, git push
17:05 thiago dbb: correct. GitHub is a website. You use your web browser for that.
17:05 dbb oh that is not true!!!
17:05 cjohnson pushes to github
17:05 thiago dbb: but Git is git.
17:05 cjohnson no need to ever open the github website
17:05 cjohnson other than to initiate
17:05 gopar joined #git
17:05 thiago or to use their task list, close pull requests, etc. That is, stuff that is *not* Git.
17:06 thiago dbb: maybe if you tell us what your concerns are, we can better help you
17:06 cjohnson even PRs are implemented as some sort of branch now, though Im' not sure if you can totally bypass the web interface yet
17:06 grawity PRs have always been branches
17:06 MattMaker joined #git
17:06 thiago right, but you can't comment on them from git
17:07 thiago you can't close them from git aside from merging the commit that it created for you
17:07 cjohnson You can merge with a commit message and I think it leaves that message on the PR and closes it
17:07 cjohnson for simple "looks good!" it works
17:07 jagob joined #git
17:07 GT4066 joined #git
17:08 thiago but you can't close with rebases or cherry-picks
17:08 dbb I think the current situation on our giant project is.. You clone the Master Master into your own Github acount clone, then clone THAT into something you control.. make your changes in a branch you create, push that branch back to the Acct Clone, and use Github.com to make a PR
17:08 thiago dbb: right. So that's GitHub.com, not Git.
17:08 ij left #git
17:08 thiago like I said, it depends on your project
17:08 dbb I asked for git to Github tips.. please recall
17:08 ReenignE dbb: if you want to use the github specific features through the cli you can use https://hub.github.com
17:08 thiago the way the workflow works is that you have to have a reopsitory on GitHub so that you can create a pull request from
17:09 thiago The Git parts are unchanged
17:09 dbb ReenignE: the first useful feedback I have heard so far
17:09 thiago you clone a repository, you push branches
17:09 thiago if you don't have push access to the master repository, you push elsewhere and you ask someone else to review it and pull it
17:09 ReenignE dbb: that said, Github does host normal git repos, they just provide some web UI around certain Github specific features
17:09 thiago that's all
17:10 DuckHuntDog joined #git
17:10 saml joined #git
17:10 WillMoogle joined #git
17:11 DuckHuntDog Hey. I've got 2 staged files which I wanna stash and 10 unstaged files which I wanna keep
17:11 StuartMI joined #git
17:11 sgen joined #git
17:12 DuckHuntDog I just did `git stash --keep-index` followed by `git stash save`
17:12 DuckHuntDog this did stash those staged files, but my unstaged files are now gone
17:12 DuckHuntDog How do I recover the unstaged files?
17:12 saml I have a github pull request (multiple commits). I can get a patch out of it. how do I revert it?
17:12 mlearner joined #git
17:12 thiago saml: revert as in git revert (create reverse patches)?
17:12 thiago saml: or in some other meaning?
17:13 grawity saml: are you trying to do this in the PR itself, or just your own repo?
17:13 saml git revert each commit?   I do not care of actual mechanics. i want to undo things that's done by those commits
17:13 DuckHuntDog saml: just as in "i want to get the unstaged files back :O" not as in the command name
17:14 saml i want to create a separate PR that undos previous PR (that is already merged)
17:14 grawity the question wasn't about mechanics, but about the result you want
17:14 kfoonamalik joined #git
17:14 DuckHuntDog saml: oh sorry for the confusion. you were replying to someone else..
17:14 DuckHuntDog never mind me
17:15 northfurr joined #git
17:15 Starky joined #git
17:16 mlearner left #git
17:18 manuelschneid3r joined #git
17:19 physicalist joined #git
17:20 User458764 joined #git
17:21 iamashishkebab2 joined #git
17:21 tyreld joined #git
17:21 TheSimonator joined #git
17:22 Vampire0 DuckHuntDog, you first stashed the non-indexed files, then the rest, so I'd say you just pop stash@{1} from the stash
17:24 orbyt_ joined #git
17:26 satifant joined #git
17:27 cdown joined #git
17:27 jwest joined #git
17:31 Starky joined #git
17:31 imack joined #git
17:34 finalbeta joined #git
17:34 denisMone joined #git
17:34 mizu_no_oto joined #git
17:34 a_thakur joined #git
17:36 gugah joined #git
17:37 davidomanfredo joined #git
17:38 krainboltgreene joined #git
17:39 hussam joined #git
17:39 dbb "you need to track to upstream repository (as origin) for the active development, and your fork on GitHub for the write permissions so your commits can be turned into Pull Requests, your local repository is where you do the actual work, and has two remotes to the respective repositories"
17:40 ayogi joined #git
17:40 phinxy joined #git
17:40 dbb .. says the DD on our project channel
17:41 phinxy Is it possible to compile a difference on a cloned checkout directory and then import it on another machine with the same checkout?
17:41 dbb If I run my own http endpoint, I could issue a PR to that endpoint also, I believe
17:42 marianina8 joined #git
17:42 iamasamosa joined #git
17:42 Vampire0 dbb, what is a DD and what do you mean with "my own http endpoint"?
17:42 gopar_ joined #git
17:43 a_thakur_ joined #git
17:43 dbb DD => Debian Developer
17:43 dbb .. which is a formal status within the Debian project
17:43 Vampire0 phinxy, sure, man git format-patch and man git am or man git diff and man git apply
17:43 gitinfo phinxy: the git-format-patch manpage is available at https://gitirc.eu/git-format-patch.html
17:43 gitinfo phinxy: the git-am manpage is available at https://gitirc.eu/git-am.html
17:43 gitinfo phinxy: the git-diff manpage is available at https://gitirc.eu/git-diff.html
17:43 gitinfo phinxy: the git-apply manpage is available at https://gitirc.eu/git-apply.html
17:44 Vampire0 dbb, and what do you mean with "my own http endpoint"?
17:44 phinxy thanks man
17:44 dbb "http endpoint" means that the http protocol can be used to make git commands.. if I understand, git alone does not have this
17:45 dbb so, on my own machine, I have a setup called Kalithea
17:45 phinxy diff and apply sounds better in my ears and makes more sense than format-patch and am
17:45 dbb Kalithea has an http visible endpoint, and can be the target or source of PRs
17:45 dbb I do not believe I can do that without Kallithea
17:46 Levex joined #git
17:46 dbb however, it means I have four copies of the main project, instead of three, to think about .. so I have not been using it
17:47 shiv joined #git
17:47 dbb I would strongly prefer not to do FOSS work on Githuib.com, but that does not seem to be possible now
17:47 Vampire0 phinxy, depends on what you want. With format-patch / am you transfer commits including author, authordate, commit message, ..., with diff / apply you only transfer changes
17:48 phinxy Before trying to apply a patch should the host clone be updated to the same version as the receiving clone?
17:49 phinxy either way an update might be in place, either now for the host or later for the receiver
17:49 Vampire0 dbb, well, if your repository is available to the one who should pull your changes and it is accessible for the project to do pull requests like that, you are fine with providing your own HTTP endpoint. If the project demands that PRs are handled by the GitHub website or GitLab website or Bitbucket website or whatever, then you need your fork on that page and push to it to being able to create a PR
17:49 dbb yes 'demand' is right !
17:50 Vampire0 phinxy, if you apply the patch to the same version you diffed against it will apply cleanly. If not, the receiver might need to resolve any conflicts.
17:50 Vampire0 dbb, this is all about project workflow definition
17:51 cbreak phinxy: diff isn't meant to create applyable diffs
17:51 Vampire0 dbb, you can also handle PRs by using git format-patch -> email -> git am
17:51 cbreak it's meant to create readable ones
17:51 roelmonnens joined #git
17:51 cbreak you might have some data loss ifyou rely on it
17:51 dbb so your long reply does not include "of course you can make a PR with command line git"
17:51 Vampire0 cbreak, sure diff is for applyable diffs
17:51 cbreak Vampire0: you can coax it to create them
17:52 cbreak but by default it will not
17:52 dbb git format-patch might be useful yes
17:52 cbreak much safer to just do it properly and use format-patch or git bundle
17:52 dbb what is git bundle?
17:52 Vampire0 dbb, well, a PR is just some workflow definition. If you tell someone hey, at that URL is my repo, pull from it, it is a PR.
17:53 dbb "URL"  -> http endpoint
17:53 Vampire0 dbb, if you click the create PR button on GitHub it is also a PR
17:53 cbreak dbb: ssh url, git url, ...
17:53 Vampire0 dbb, no, url is url, doesn't have to be http, can be ssh, git, ftp, whatever
17:53 dbb I have never seen anyone every use those other protocols, even once
17:53 Vampire0 dbb, while ftp is deprecated and shouldn't be used
17:53 dbb s/ever/severy/
17:54 dbb never never not once
17:54 cbreak you've not seen anyone use ssh?
17:54 cbreak sheltered life you've been living
17:54 dbb to make a PR
17:54 dbb I run secure servers
17:54 cbreak they aren't used to make pull requests
17:54 cbreak they are used to pull
17:54 cbreak pull requests can be made by phone, signal, email, what ever
17:54 dbb I was talking about PRs no?
17:54 Vampire0 dbb, sure, the git protocol is quite common for read-only access and the ssh protocol is the best when also wanting to push, because you can identify with SSH keys instead of username / password
17:54 cbreak I usually yell across the table to make pull requests
17:54 Vampire0 dbb, this has nothing to do with PRs or not
17:55 Vampire0 <dbb> what is git bundle? -> man git bundle
17:55 gitinfo the git-bundle manpage is available at https://gitirc.eu/git-bundle.html
17:55 dbb okok- less noise, more light from me.. so if I want to push with ssh, where do I see an example of that?
17:55 AndrewGazelka joined #git
17:55 dbb you mean that ssh thing where you pass a single command inline ?
17:56 cbreak git remote add origin ssh://server/path.git
17:56 cbreak git push -fu origin master
17:56 phinxy its a difference to master git repo id like to apply on another machine
17:56 Vampire0 dbb, what do you mean? Whether you push via ssh, http, or whatever is just a matter of URL you use
17:56 mischat joined #git
17:56 dbb git push -fu !!
17:56 dbb LOL
17:57 dbb I will try that !
17:57 Vampire0 dbb, -fu are common options, not specific to SSH
17:57 Vampire0 cbreak, what do you need to coax to make diff applyable o_O
17:58 dbb git remote add <remote_name> ssh://server/path.git  # yes?
17:58 cbreak Vampire0: enable binary diffs
17:58 TonCherAmi joined #git
17:58 cbreak as a minimum
17:59 cbreak otherwise you'll lose non-text changes
18:00 finalbeta joined #git
18:00 njbair joined #git
18:00 circ-user-zNiD2 joined #git
18:00 Cabanossi joined #git
18:02 vacho joined #git
18:02 Vampire0 cbreak, yeah, granted if there are binary changes
18:02 irqq joined #git
18:02 Vampire0 dbb, probably yes, depends on the question
18:03 yena joined #git
18:03 vacho hello all, I am new to git and I might have done something wrong where I lost my files. I ran "git checkout -b sprint-11-b" and then "git add ." and then "git commit -m 'msg'", and not suddendly that branch is not existing
18:04 _ikke_ If you just did that, then there is nothing lost
18:04 _ikke_ vacho: !transcript
18:04 gitinfo vacho: 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
18:05 Sasazuka joined #git
18:05 dbb honestly I work on a copy of files outside of source control quite often, to avoid the constant needs of the git "OS"
18:06 cbreak sounds dumb.
18:06 cbreak like driving without seatbelt
18:06 dbb so vacho -- personally, I would delete the whole directory, get the files are you did.. copy them and do work.. unless the idea is to test your git skills
18:06 dbb it is exactlylike driving without a seatbelt, which I do when I feel like it
18:07 cbreak as I said, dumb :P
18:07 dbb I have writeen several large software systems from scratch
18:07 cbreak git won't lose changes you tell it to keep until you tell it to lose them.
18:07 tarkus joined #git
18:08 vacho hope this is better - https://gist.github.com/wisercapital/d45a5ad944b9ed80112959d3136c6a3c
18:09 _ikke_ vacho: what does git branch return?
18:09 vacho _ikke_: master and sprint-11 (older branch I am not using)
18:10 notebox joined #git
18:10 vacho I want to find my files committed to "sprint-11-b"
18:10 vacho I made the mistake of checking out a new branch with the same name "sprint-11-b", hoping files would come back, but no
18:11 _ikke_ vacho: Those commands you typed in do not make a branch disappear
18:11 mkoskar joined #git
18:11 _ikke_ But it helps to give an actual terminal transcript, including the output of each command
18:11 vacho _ikke_: I agree, it shouldn't..but it did
18:11 vacho _ikke_: I don't have that anymore :(
18:12 _ikke_ git log --all
18:12 vacho _ikke_: but I didn't get any error outputs
18:12 AndreasLutro check git reflog to see if your commit is there somewhere. I would guess that one of your commands failed and you didn't notice
18:12 Es0teric joined #git
18:12 phinxy If a repo on github.com is cloned will a git commit push the updates to master. obviously not, but why?
18:13 _ikke_ git commit pushes nothing
18:13 dbb git commit - from where ?  to where?
18:13 vacho _ikke_: https://gist.github.com/wisercapital/ba47f766c7b6d982d3b1b1ba9c72cbbb
18:15 vacho git commit locally, then I did git push that will push it to github.com
18:15 vacho sorry, I am pretty new to git
18:15 MLM__ joined #git
18:15 vacho I hope I am answering your questions
18:15 _W_ joined #git
18:19 phinxy git apply --stat didnt show any new files added, only changes.
18:20 vacho phinxy: so there might be a risk I lost all my files?
18:20 vacho or changes I mean
18:20 cbreak vacho: not if you only did that
18:20 cbreak vacho: type git reflog and paste the output
18:21 dbb cbreak:  losing all your changes sounds "dumb"
18:21 cbreak dbb: that's why people should use git
18:21 cbreak using git avoids such problems
18:21 dbb or not
18:21 vacho cbreak: https://pastebin.com/iYDdJHsJ
18:22 tmsmith joined #git
18:22 cbreak vacho: those are the changes you made to HEAD
18:22 dbb every single undergraduate has lost a nights worth of work in some simple error.. I suggest that using a COPY of the files in a "dumb" way, prevents that, while other skills build..
18:22 vacho cbreak: ok?
18:23 dbb git is great .. git is hard.. no surprises
18:23 shgysk8zer0 joined #git
18:23 vacho I am just hoping I can recover my files and changes, it's about two weeks of work
18:23 cbreak vacho: doesn't look like you actually did commit anything
18:23 phinxy ugh theres a ton of useless change because master decided to indent code a bit to the right. any way to ignore these?
18:23 dbb TWO WEEKS!
18:23 cbreak vacho: haha, you didn't commit for two weeks?
18:24 mizu_no_oto joined #git
18:24 vacho cbreak: I was working on a single ticket, wanted to commit when it was done..I know that's a bad practice
18:24 dbb hey dont laugh - that guy/gal has a serious problem
18:24 Murii joined #git
18:24 cbreak vacho: anyway, it seems the most recent commit is eb3db09
18:24 govg joined #git
18:25 vacho dbb: it's okay, it's not a big deal..I can rewrite it in 2 days probably
18:25 vacho but I rather recover my files if possible
18:26 cbreak vacho: do you think there should be a newer commit than that?
18:26 h12o joined #git
18:26 vacho cbreak: cool if I write you in private?
18:26 cbreak this channel is fine
18:26 vacho ok
18:26 cbreak I'm not the only one with git knowledge :)
18:27 Es0teric joined #git
18:27 cbreak if you type git log -g, it will list all reflog commits with commit message
18:27 vacho I did a git stash also at some point about 4 days ago, does that get stored?
18:27 cbreak if you don't see your commit there, then you didn't make it in the last 30 days or so
18:27 cbreak stashes get stored in their own reflog
18:27 cbreak git stash list will show them
18:28 cbreak but git stash pop or git stash drop will remove them
18:28 TheSimonator joined #git
18:28 vacho git stash list has very old stashes, not the recent one
18:29 vacho so I guess it got dropped
18:29 dbb so - Move SLowly !   git stash list and stop
18:29 orbyt_ joined #git
18:29 vacho cbreak: https://pastebin.com/y3UR8XJt
18:29 dbb ugh - too much at stake for a single commandline to delete it
18:30 cbreak I doubt anything got deleted
18:30 cbreak at the moment it looks as if it was never committed
18:30 vacho cbreak: glad to hear
18:30 cbreak because if it were committed, it would be in the reflog
18:30 cbreak and would easily be recoverable
18:30 evanc joined #git
18:30 dbb hey vacho  - did you change branches at all ?
18:31 cbreak dbb: yes he did
18:31 dbb that rewrites the whole tree
18:31 vacho cbreak: something happened, cause the brandh was not even in the list when I did "git status"
18:31 cbreak dbb: all the checkout in his reflog show that he did
18:31 vacho dbb: yes, I was changing between sprint-11-b and master
18:31 dbb clue
18:31 cbreak vacho: git status only shows working dir state and some meta data
18:31 cbreak if you want to see branches, type git branch -a
18:32 dbb ok I am in the presence of deeper knowledge.. I *think* .. that changing branches, writes the files that have been changed back to the last common branch, not the whole tree.. yes?
18:32 dbb common commit ID, that is
18:33 cbreak dbb: neither
18:33 cbreak dbb: in this case, changing branches doesn't touch the filesystem at all
18:33 cbreak since the branches he switched between pointed at the same commit
18:33 dbb why
18:33 cbreak as you can see from the commit hash
18:33 dbb right - the same common commit node
18:34 dbb which is 'current'  in this case perhaps
18:34 vacho when I do "git add .", does that state get saved somehow?
18:34 tamatar joined #git
18:34 cbreak if there were differences in his files, git would refuse to change branches
18:34 cbreak vacho: yes
18:34 raspado joined #git
18:34 cbreak vacho: it gets recorded in the staging area
18:34 arussel left #git
18:34 vacho cbreak: any suggestions on what I can try next?
18:34 phinxy format-patch seems to make a nice patch with changes. Although inside the file there is no mention of some changes. Im looking in the .patch with vim
18:35 cbreak vacho: why do you think your changes are gone?
18:35 dbb vacho: I use  git branch -vva
18:35 dbb .. to see branch info, because the debian people have super complex things in there
18:38 cbreak vacho: did you look at the files in the filesystem? the output of git status?
18:39 vacho cbreak: please explain your last question again?
18:40 vacho cbreak: the new files I added to the file-system are gone
18:40 cbreak :/
18:41 cbreak so they are not in the filesystem?
18:41 cbreak did you run something like git clean?
18:41 vacho cbreak: No
18:41 cbreak or git reset?
18:41 karstensrage joined #git
18:41 cbreak the purpose of both commands is to delete things
18:41 karstensrage can you git subtree split into a bundle?
18:42 D630 joined #git
18:42 vacho cbreak: no, I did following: https://pastebin.com/f62xE62v
18:42 bronson joined #git
18:43 cbreak karstensrage: you can push which ever history you want into a bundle as far as I know
18:43 Vgr joined #git
18:43 monokrome joined #git
18:43 fahadash joined #git
18:43 cbreak vacho: there is no record of that git commit
18:43 cbreak vacho: are you sure you really made it and it didn't give an error?
18:43 Acerific joined #git
18:43 Acerific joined #git
18:43 cbreak because if git would know that you made it, which is to be expected if it happened, then it would be in the reflog
18:45 phinxy applying .patch with am did not show a resolve tool when error
18:45 phinxy it just tells "fix it"
18:45 KG4PEQ____ joined #git
18:45 vacho cbreak: I am pretty sure I did it..even if it's not in the reflog, I should see that branch at the very least, that branch was not there either
18:45 cbreak vacho: maybe you did it in a different repository?
18:47 zeroed joined #git
18:47 zeroed joined #git
18:47 cbreak vacho: your branch switching was recorded
18:47 kyan joined #git
18:47 cbreak you switched to sprint-11-b twice
18:48 vacho cbreak: ok, will that overwrite it? I am sure it will just warn saying it already exists?
18:48 cbreak vacho: ... just to be sure
18:48 hobodave joined #git
18:48 cbreak your git repository is on a sane filesystem?
18:48 cjohnson vacho: none of those commands would delete/reset working copy changes
18:48 cbreak not something shitty like smb?
18:48 vacho I can try restarting the server
18:48 prosody joined #git
18:48 vacho I am not using samba, it's an EC2 environment
18:48 vacho going to try a reboot
18:49 h12o joined #git
18:49 Bengi joined #git
18:49 cbreak vacho: "fatal: A branch named 'x' already exists."
18:49 cbreak that's what it says when checkout -b collides
18:52 vacho files still missing after reboot, seems like I have exhausted all alternatives?
18:52 vacho my files are lost
18:53 cbreak it doesn't seem like git recorded a commit
18:53 cbreak if your git add succeeded, git fsck --lost-found might get them
18:53 cbreak but ... that's more a last resort
18:53 cbreak it's also annoying to use
18:54 vacho ok, I see a list of dangling blobs and commits, what do I do next?
18:54 osse look at them
18:54 vacho how do I look at them?
18:54 cbreak they should be in lost-found inside .git, somewhere
18:54 cbreak with hash-names
18:55 vacho ok
18:55 osse .git/lost-found/
18:56 auv joined #git
18:56 vacho just found a few files there, nothing related to myw ork
18:56 vacho my work*
18:57 osse did you check all the files in .git/lost-found/other ?
18:57 vacho osse: yes I did
18:57 osse then you've run out of luck
18:57 vacho thanks everyone for helping, I guess my work got lost. I will make sure I commit daily from now on
18:58 bpna joined #git
18:59 phinxy first patch went fine but second said sha1 info lacking or useless
18:59 hoek joined #git
18:59 finalbeta joined #git
19:01 TonCherAmi Hi, I'm working on a PR and I need to remove an unneeded commit and make some changes to another one (which is newer than the unneeded one). Would the following steps be an okay way to handle this? https://ptpb.pw/ATWC
19:01 avar I'm looking for volunteers to print out git's source code, eat it, so we can certify it safe to eat (https://public-inbox.org/git/504208edac024b2dba0c1743dc86683e@AMUSATLWP20005.alereinc.com/) >:)
19:03 phinxy is it possible to remove all commits without deleting changes and make them all over
19:04 phinxy .. in one big commit so the patch perhaps works better
19:04 da5id joined #git
19:04 avar git reset <commit to start at>; see "man git-reset"
19:04 gitinfo the git-reset manpage is available at https://gitirc.eu/git-reset.html
19:04 avar i.e. the not---hard options
19:04 MACscr OK, so I have multiple servers that i work on and then i push the changes to a single branch. I do though have a case where I have changes already pushed to the master branch from server A on file example.php, but server B has changes to example.php as well. How do I merge what i have local with what i have in master?
19:04 MACscr for that one file
19:05 SlashLife joined #git
19:05 dviola joined #git
19:05 phinxy the manuals are pretty hard to understand
19:06 phinxy there are so many options for each commands!
19:06 Starky joined #git
19:07 phinxy Will a git patch contain new files?
19:07 karstensrage it keeps telling me no HEAD
19:08 karstensrage how can i bundle from a branch to "master"
19:09 karstensrage like i do git subtree split --prefix=foo -b foo
19:09 h12o joined #git
19:09 karstensrage so now i have a foo branch split from the main git repo
19:10 karstensrage now i want to bundle that up as if its the only thing
19:10 cfoch-always joined #git
19:10 cfoch-always is there a way to know if git is in detached state?
19:10 iamashishkebab2 joined #git
19:13 joshszep joined #git
19:13 n-st joined #git
19:17 avar "git status"
19:17 miha_S7_ joined #git
19:19 mspiggy joined #git
19:20 phinxy patch failed: patch does not apply. what now?
19:20 h12o joined #git
19:20 apax joined #git
19:21 dsdeiz joined #git
19:21 kiltzman joined #git
19:22 cjohnson well, ideally, get a patch that applies
19:22 cjohnson where did your patch come from
19:22 phinxy --reject --whitespace=fix seems to do more
19:22 finalbeta joined #git
19:22 phinxy only thing thats wrong is that the patch comes from a slightly older branch than master and there are whitespace
19:23 phinxy slightly older clone
19:23 cjohnson if that's not enough, yu can use the patch tool to work on it in smaller bits
19:23 cjohnson https://www.drupal.org/node/1129120
19:23 svm_invictvs joined #git
19:23 yqt joined #git
19:23 kiltzman joined #git
19:23 zeroed joined #git
19:23 zeroed joined #git
19:24 apax joined #git
19:24 phinxy a lot of rejects whatever that means. at quick glance it looks like my modifications patched over fine
19:25 phinxy will try compile now..
19:26 ploop joined #git
19:26 aielima joined #git
19:26 phinxy theres a '
19:26 phinxy <<<<< updated upstream' inside a .c file
19:27 miha_S7_ joined #git
19:29 durham joined #git
19:30 phinxy removed it and it compiles :)
19:30 apax joined #git
19:32 m0viefreak joined #git
19:34 mischat joined #git
19:35 Starky joined #git
19:35 mischat joined #git
19:35 Goplat joined #git
19:36 XenophonF joined #git
19:37 cagomez joined #git
19:38 northfurr joined #git
19:38 dbb ok back to the "change branch" question.. when git changes branches, it rewrites.. how many files?
19:38 mmattice dbb: the ones it needs to
19:39 dbb mmattice: political background?
19:39 dbb sounds like what a politician would say !
19:39 Bombe dbb, project manager background?
19:40 mmattice I'll just quote the song that just started playing... "It's a trap!"
19:40 Bombe Because that’s exactly what a project manager would sa<y.
19:40 h12o joined #git
19:40 dbb so there is a chain of hashed nodes..  which forks at some commit node.. and that common node is decided on the change branch command
19:40 bremner I'd say the number of files changed is undefined behaviour. But it usually works reasonable quickly.
19:41 dbb the branch has to find out which common node is there, and replays all files on all nodes past the common node
19:41 dbb I would guess
19:41 dbb I dont know how dirs are handled
19:42 bronson joined #git
19:43 sea-gull joined #git
19:44 Bombe I don’t think common commits are involved here.
19:44 Bombe You can have completely unrelated commit chains in one repository and switch between those without hassle.
19:44 Bombe That wouldn’t be possible if a common commit was required.
19:44 Pseudocrat joined #git
19:45 Cabanossi joined #git
19:46 dbb ok - but that implies a lot of files being written on branch switch
19:46 Bombe No, it doesn’t.
19:46 dbb so - there can be files in one branch that do not appear at all in another.. sure
19:46 bremner dbb: have you actually tried it?
19:47 dbb tried what -- the code that computes the needed files to write on branch switch  ?  no I havent
19:47 Bombe Every commit has a tree associated with it, and only the difference in the trees are re-created in the filesystem.
19:47 Bombe If there are common files in both trees they won’t be touched.
19:47 Bombe Because why?
19:47 bremner tried switching branches with your codebase/whatever and seen if it's fast enough
19:48 dbb "only the differences in the tree are re-created" .. thats the same thing I am saying
19:48 dbb but with details
19:48 Bombe There’s no details involved here.
19:48 dbb I am asking, dont get me wrong.. I definotely dont know lots about git
19:48 dbb heh
19:48 Bombe If files differ, of course they will be rewritten.
19:48 dbb you have got to be joking
19:48 Bombe If they don’t differ, they won’t be touched.
19:49 Bombe That’s all there is to it.
19:49 dbb how are the files to be written determined?
19:49 Bombe By comparing the trees associated with the commits.
19:49 dbb thats what I was describing
19:49 dbb the tree is composed of nodes, which are hash signed commits, no?
19:50 qqx dbb: No, that isn't what you were describing. The comparison has nothing to do with commits.
19:50 Vortex34 joined #git
19:50 qqx s/commits/history/
19:50 cagomez joined #git
19:50 Bombe dbb, no, the tree is a tree and has nothing to do with commits.
19:50 dbb when the branch is switched from one to another, WHICH files are touched?  crawl back the tree to find a common ancestor
19:51 grawity it doesn't have anything to do with common ancestors, really
19:51 qqx dbb: There is no crawling to find a common ancestor. Both the original and new commits point at the root tree to be used.
19:51 grawity diff the two commits directly and you get the result
19:51 Furai joined #git
19:52 borkr joined #git
19:52 dbb qqx: commits?  I am asking about branches
19:52 dbb when switching two branches
19:52 dbb what files are touched..
19:52 qqx You can do largely the same thing with "diff | patch"
19:52 Bombe A branch is a pointer to a commit.
19:52 grawity when switching two branches, you *are* switching two commits
19:52 grawity there is no difference as far as the checkout goes
19:52 qqx dbb: Branches are just a pointer to a commit.
19:52 Bombe Your mental model of Git is completely screwed.
19:53 grawity in both cases you're telling Git to go from snapshot A to snapshot B
19:53 dbb well thats why I am here asking quesitons.. but I am not convinced I have said anything actually wrong yet
19:53 grawity so it touches those files which are different between those two snapsots
19:53 dbb if you say that branches are a pointer to a commit thats partly true
19:53 orbyt_ joined #git
19:54 grawity I am not convinced I have said anything not entirely true, either
19:54 Bombe I am not convinced I have not not said anything not entirely not true, or not.
19:54 dbb I believe that!  but "snapshot" .. that sounds like a screwed mental model to me... there is no "snapshot" .. there is a tree of hashed commit nodes
19:54 durham joined #git
19:55 dbb with a pointer to the latest one
19:55 grawity each of which is a full snapshot of the entire worktree.
19:55 Bombe “Snapshot” is a very accurate term here.
19:55 dbb all the replayed nodes, there current state, are kept as a seperate thing from the nodes ?
19:55 Bombe “Replayed nodes?”
19:55 dbb I dont believe there is a snapshot, only nodes
19:55 grawity those nodes *are* snapshots
19:55 dbb yes, the nodes in the tree of commits
19:56 grawity you're probably thinking that commits describe changes?
19:56 Bombe dbb, there is nothing replayed about the commits.
19:56 dbb commits are a signed set, making a node I think
19:56 Bombe dbb, each commit contains a reference to a complete snapshot of the directory.
19:56 grawity nope, you're making shit up
19:56 Bombe A commit is _not_ a diff.
19:56 dbb hm
19:57 _ikke_ a commit *is* a snapshot
19:57 grawity (I have both `git cat-file` and https://git-scm.com/book/en/v2/Git-Internals-Git-Objects as support)
19:57 _ikke_ it contains the complete working tree
19:57 dbb that would make git repos pretty large, no?
19:57 _ikke_ there is deduplication
19:57 dbb the whole set of all files is in every branch, really?
19:58 Bombe No, why would it?
19:58 Bombe Gosh.
19:58 dbb each commit node?
19:58 grawity most of that set is identical across commits
19:58 _ikke_ git uses a content addressable system
19:58 Bombe Are you trying to be difficult? :)
19:58 grawity so the same file is stored just once, even if you have 100 commits with it
19:58 dbb _ikke_:  <--  that!
19:58 dbb see !!
19:58 dbb you guys are quick to call BS
19:58 dbb I am not wrong..
19:58 _ikke_ dbb: conceptually, the commit contains a complete snapshot
19:58 grawity neither are we
19:58 MineCoins joined #git
19:58 dbb but I dont know all of it of course
19:59 dbb I didnt say anyone was wrong
19:59 _ikke_ so saying a commit is a snapshot is correct
19:59 dbb ok  what _ikke_  is saying makes sense to me
19:59 heroux joined #git
19:59 telephone joined #git
19:59 telephone joined #git
19:59 _ikke_ You only need the commit and the tree + blobs
20:00 MarkusDBX joined #git
20:00 dbb ok - so back to the original quesiton then.. when switching branches, which files are rewriteen
20:00 h12o joined #git
20:00 Bombe Hahaha!
20:01 Bombe The ones that change between the two commits.
20:01 grawity back to the same answer – the ones which are different between those two branches
20:01 thebope joined #git
20:01 _ikke_ dbb: git calculates which files differ
20:01 Bombe That was the correct answer half an hour ago and it still isl
20:01 dbb how are they determined to be different
20:01 grawity by comparison?
20:01 MarkusDBX I'm teaching a fellow dev not too familiar with the cli git, is there any good tools for mac, that will make merging and commiting easier. Or is any tool that tries to abstract git into a gui, useless and a waste of time.
20:01 grawity or in practice, by their hash
20:01 _ikke_ dbb: diff against the checked out commit
20:01 bpna joined #git
20:01 thiago joined #git
20:01 grawity (as the snapshots record hashes of every file)
20:01 dbb yes .. diff has to have the code
20:01 _ikke_ ?
20:02 grawity not even a diff, just binary comparison
20:02 _ikke_ You can diff tree objects
20:02 Bombe git-checkout doesn’t compary any file contents, though, does it?
20:02 _ikke_ Doesn't need to
20:02 macrover joined #git
20:03 Bombe I know. :)
20:03 dbb "The ones that change between the two commits."  says nothing..
20:03 Bombe It says exactly what happens.
20:03 _ikke_ dbb: list all files in commit1, list all files in commit2
20:03 grawity what sort of explanation are you expecting though
20:03 _ikke_ find the differences
20:04 dbb thx _ikke_ .. I will do some reading and come back another day
20:04 masuberu joined #git
20:04 TbobbyZ_ joined #git
20:04 dbb git cat-file is new to me, I will look at that, and perhaps also  https://git-scm.com/book/en/v2/Git-Internals-Git-Objects
20:05 dbb left #git
20:05 grawity good, I nearly ran out of popcorn
20:05 _ikke_ haha
20:05 Bombe Yeah. Also, bedtime. Good night!
20:05 grawity (kidding... I absolutely hate popcorn)
20:05 _ikke_ https://imgur.com/gallery/0hQyd5L
20:09 macrover joined #git
20:11 cdg joined #git
20:12 hobodave joined #git
20:15 zeroed joined #git
20:15 zeroed joined #git
20:16 redoverture joined #git
20:18 misosoup joined #git
20:19 finalbeta joined #git
20:19 Es0teric joined #git
20:21 h12o joined #git
20:21 bariscant joined #git
20:23 mischat joined #git
20:23 livingstn joined #git
20:24 leeN joined #git
20:26 northfurr joined #git
20:28 synthroid joined #git
20:29 cagomez joined #git
20:32 lipsumar joined #git
20:37 jackrandom joined #git
20:39 heroux joined #git
20:41 h12o joined #git
20:44 mizu_no_oto joined #git
20:45 Cabanossi joined #git
20:45 CheckDavid joined #git
20:46 apax joined #git
20:51 cdown joined #git
20:53 TbobbyZ joined #git
20:53 apax joined #git
20:58 diogenese joined #git
20:59 madprops is git diff a  purely local operation?
21:00 apax joined #git
21:00 moritz madprops: yes
21:01 madprops hmm weird
21:01 madprops sometimes it lags quite a bit
21:01 moritz madprops: basically only fetch, pull, push, ls-remote talk with a remote
21:01 _ikke_ remote with some options also
21:01 moritz slow file system, maybe?
21:01 madprops it's not even a huge project
21:01 h12o joined #git
21:01 madprops i think it might be windows 10 or maybe cmder
21:01 durham_ joined #git
21:02 moritz right, 'remote update' is basically 'fetch'
21:03 Vampire0 moritz, `remote --set-head`
21:03 _ikke_ Vampire0: local operation
21:04 Vampire0 _ikke_, --set-head is not a local operation. It asks the remote for the current value of HEAD and sets refs/remotes/HEAD to it
21:04 _ikke_ Oh, right
21:04 Vampire0 s!refs/remotes/HEAD!refs/remotes/<remote>/HEAD!
21:04 moritz my point is that msot high-level commands except fetch/pull/push are local
21:05 Vampire0 yay, the one time I was right against _ikke_ :-D
21:05 apax joined #git
21:05 _ikke_ :D
21:05 _ikke_ Vampire0: Some people think it would set the remote HEAD
21:05 _ikke_ Vampire0: Only with -a|--auto though
21:06 Vampire0 yes, of course :-)
21:07 Vampire0 to set the remote HEAD I guess you should push foo:refs/HEAD ?
21:07 _ikke_ nope
21:07 apax joined #git
21:07 _ikke_ you cannot
21:07 Vampire0 ah, only on the remote itself then, or via some web interface
21:07 _ikke_ correct
21:08 dsdeiz joined #git
21:08 Vampire0 why is there no way to set the remote HEAD?
21:08 _ikke_ Vampire0: No one bothered yet to add it\
21:08 marianina8 joined #git
21:10 kyan joined #git
21:13 Vampire0 ah, ok. and pushing to refs/HEAD does not work because? What is allowed to be pushed to?
21:13 duderonomy joined #git
21:13 Vampire0 Only refs/heads and refs/tags?
21:13 lipsumar joined #git
21:13 miczac joined #git
21:17 cbreak Vampire0: HEAD is a symref
21:17 cbreak it cannot be pushed to
21:18 herr_barium joined #git
21:18 _ikke_ Vampire0: anything starting with refs I believe
21:18 _ikke_ refs/
21:18 marianina8 joined #git
21:20 MattMaker joined #git
21:21 aidalgol joined #git
21:22 h12o joined #git
21:23 Tobbi joined #git
21:25 mischat joined #git
21:27 nobo728x joined #git
21:28 dsantiag_ joined #git
21:29 northfurr joined #git
21:29 faisal joined #git
21:30 durham joined #git
21:30 marianina8 joined #git
21:30 brent__ joined #git
21:30 shinnya joined #git
21:31 regedit joined #git
21:31 bpna joined #git
21:33 Starky joined #git
21:35 Vampire0 ah, right, HEAD is outside refs and it is a sym ref, so pushing does not make much sense
21:36 Vampire0 thx cbreak _ikke_
21:37 Levex joined #git
21:38 transhuman joined #git
21:39 transhuman hi, just checking. I am cloning the entire linux torvals tree with git clone .... .git am I correct in saying this gets the entire git repository or do I need a different command?
21:41 ResidentBiscuit joined #git
21:41 thiago it gets everything
21:41 ResidentBiscuit joined #git
21:41 ResidentBiscuit joined #git
21:42 tourdownunder joined #git
21:42 h12o joined #git
21:42 cfoch-always when git is in detached state, is it possible to know what branch was being checked out before being detached?
21:43 Vampire0 cfoch-always, do you want to return to it or just look it up?
21:43 TheSimonator joined #git
21:43 transhuman thanks thiago
21:43 bronson joined #git
21:44 cfoch-always Vampire0: return it I think
21:44 Vampire0 cfoch-always, `git checkout -`
21:44 Vampire0 - always goes to the last checked out branch you can also switch back and forth between two branches with it
21:44 cfoch-always Vampire0: I missunderstood
21:45 cfoch-always I don't want to checkout
21:45 cfoch-always I just want to know the name of the branch before the status is detached
21:45 Cabanossi joined #git
21:45 Tobbi joined #git
21:45 cfoch-always Vampire0: ^
21:46 _ikke_ cfoch-always: git rev-parse --abbrev-ref HEAD@{-1}
21:46 thiago why?
21:47 marianina8 joined #git
21:47 marianina8 joined #git
21:48 northfurr joined #git
21:49 borge joined #git
21:50 mattcen joined #git
21:52 h12o joined #git
21:52 chipotle joined #git
21:53 askb joined #git
21:54 aielima joined #git
21:55 Dougie187 left #git
21:56 SCHAPiE joined #git
21:56 cfoch-always _ikke_: nope, that doesn't work
21:56 Vampire0 cfoch-always, not? it should, what does it give you?
21:57 sim642 joined #git
21:57 Vampire0 cfoch-always, you should also be able to leave out the HEAD
21:58 Vampire0 cfoch-always, ah, acutally you *have* to leave out HEAD. It is only @{-1}
21:58 tmsmith joined #git
21:58 cfoch-always awesome
21:58 Vampire0 _ikke_, probably confused it because in most cases @ is just short for HEAD
21:59 cfoch-always what is "git rev-parse --abbrev-ref @{-1}
21:59 cfoch-always sorry
21:59 Vampire0 cfoch-always, man git revisions
21:59 gitinfo cfoch-always: the git manpage is available at https://gitirc.eu/git.html
21:59 cfoch-always @{-1}
21:59 Vampire0 hm, the bot misses a page
21:59 _ikke_ yeah, @{-1} is not working either, hmm
21:59 _ikke_ man gitrevisions
21:59 gitinfo the gitrevisions manpage is available at https://gitirc.eu/gitrevisions.html
21:59 Vampire0 it is _ikke_
21:59 Vampire0 ah, ok, so cfoch-always man gitrevisions
21:59 gitinfo the gitrevisions manpage is available at https://gitirc.eu/gitrevisions.html
22:00 cfoch-always what does it mean?
22:00 cfoch-always @{-1}
22:00 Vampire0 cfoch-always, there you find @{-1} is short for "the last branch checked out before the current one"
22:00 cfoch-always nice!
22:00 tourdownunder joined #git
22:00 Vampire0 _ikke_, @{-1} works fine here
22:00 cagomez joined #git
22:00 cagomez joined #git
22:01 cfoch-always yes, it works
22:01 cfoch-always I didn't say it doesn't
22:01 Vampire0 cfoch-always, _ikke_ said it does not work
22:02 _ikke_ Vampire0: I head to checkout another commit for it to work
22:03 h12o joined #git
22:03 _ikke_ git rev-parse --abbrev-ref  @{-1}
22:03 _ikke_ This works, yeah
22:03 _ikke_ returns nothing if the previous thing you checked out was a commit instead of a branch
22:05 cagomez joined #git
22:06 regedit i'm adding a new origin; how do i setup "foo" branch of my first origin to track "master" branch of this new origin?
22:07 regedit or am i doing terrible, terrible things
22:07 regedit and should give up and go drive for Uber
22:08 _ikke_ regedit: s/origin/remote
22:08 regedit right sorry; i'm adding a new remote
22:09 _ikke_ regedit: And you want your local branch foo track remote branch master for this new origin?
22:09 regedit how do i setup "foo" branch of my 'origin' remote to track "master" branch of this new remote?
22:09 regedit yes, if that makes any sense, because the same local "foo" branch is already tracking remotes/origin/foo
22:10 _ikke_ regedit: remote tracking branches do not track other remote tracking branches
22:10 _ikke_ git branch --set-upstream-to remote_name/master foo
22:10 _ikke_ that would set local branch foo to track remote_name/master
22:11 regedit so i can't have "foo" simultaneously track origin/foo AND new-origin/master ?
22:12 _ikke_ No, a branch can track only one other branch
22:12 regedit i see
22:12 _ikke_ regedit: what do you want to achieve with it?
22:13 ResidentBiscuit joined #git
22:13 _ikke_ it tells git for example what remote to push to when you do just git push (depending a bit on your push.default setting)
22:13 _ikke_ That would not work when it can track multiple branches
22:14 regedit we have a branch "foo" which we now want to deploy on its own site (i.e. separate domain & remote git repo)
22:14 regedit we still want to keep foo's roots in the original repo, but we want to be able to push & deploy changes to a separate server
22:15 orbyt_ joined #git
22:15 _ikke_ tracking is not required for that
22:16 _ikke_ though, I would not recommend using git to deploy, but that's another story
22:16 regedit right i guess we're kinda using git as a deploy tool here.... :(
22:16 howitdo joined #git
22:16 howitdo joined #git
22:17 regedit despite how doing-it-wrong that is, any chance we can still do our usual "git push" command to deploy this foo branch to a new site?
22:17 regedit or is it just "git push new-origin foo"?
22:18 _ikke_ correct
22:18 _ikke_ tracking is not required to push
22:19 otiose left #git
22:19 gaucheph joined #git
22:22 jeffreylevesque joined #git
22:22 orbyt_ joined #git
22:23 stealthii joined #git
22:24 steeze joined #git
22:25 steeze im seeing a recommendation to keep hooks at root of the project and symlink them to the actual ./git/pre-commit, why is this?
22:25 regedit _ikke_ ok thanks ????
22:25 Vampire0 _ikke_, interestingly wihtout --abbrev-ref it outputs the sha
22:25 h12o joined #git
22:27 jaziz joined #git
22:29 mischat joined #git
22:32 cagomez joined #git
22:33 Vampire0 _ikke_, cfoch-always,  well I guess to get this reliably working one would have to code a loop that loops through @{-1}, @{-2}, ... using rev-parse and checking the result against show-ref --heads to find the last branch that was checked out. Or git reflog and then reading the entries
22:35 devhost joined #git
22:35 BSaboia joined #git
22:35 huhlig joined #git
22:35 northfurr joined #git
22:36 Acerific joined #git
22:36 Acerific joined #git
22:37 hussam joined #git
22:38 kexmex joined #git
22:42 mikecmpbll joined #git
22:45 orbyt_ joined #git
22:46 ulkesh joined #git
22:47 MrJoshua joined #git
22:47 Atm0sphe1 joined #git
22:47 pl joined #git
22:49 NeXTSUN joined #git
22:49 cqi joined #git
22:50 roar joined #git
22:50 masuberu joined #git
22:51 zturner joined #git
22:54 cyphase joined #git
22:56 gugah joined #git
22:56 Starky joined #git
22:57 NeverDie joined #git
22:59 dendazen joined #git
23:00 tourdownunder joined #git
23:00 xpika joined #git
23:02 druonysus joined #git
23:02 PettanShoutaKun joined #git
23:03 telephone joined #git
23:03 telephone joined #git
23:03 brent__ joined #git
23:04 jasom joined #git
23:07 cads joined #git
23:07 jasom is there a command to delete stale tmp_pack files, or should I just manually delete them?  I had a repack fail due to disk quotas and it left a bunch of those around.
23:07 xpika left #git
23:08 livingstn joined #git
23:09 YuGiOhJCJ joined #git
23:12 netj joined #git
23:13 Whiskey joined #git
23:14 ToBeCloud joined #git
23:15 sebhoss joined #git
23:17 Emperor_Earth_ joined #git
23:17 ertes joined #git
23:21 earnestly joined #git
23:21 mal_ joined #git
23:25 drupol joined #git
23:29 mischat joined #git
23:29 Emperor_Earth__ joined #git
23:30 Cabanossi joined #git
23:30 mda1 joined #git
23:30 malmalmal joined #git
23:31 Emperor_Earth_ joined #git
23:33 malmalmal joined #git
23:34 Emperor_Earth joined #git
23:35 northfurr joined #git
23:35 miha_S7_ joined #git
23:35 malmalmal joined #git
23:36 malmalmal Hello, I am starting with git, and have some doubts. In example, I inited a git repo on a remote server, as I understand it, if I 'git clone' that server, on a local machine, it will not only clone master branch but also any other existing branches is this correct ?
23:37 malmalmal I mean, it will clone, locally, all remote branches including master, on remote repo
23:39 Vampire0 malmalmal, yes, if you don't take special measures (parameters to clone) you will get a full-fledged repository with all the history that is available in the remote for all branches
23:40 malmalmal Thanks Vampire0
23:40 malmalmal Another question if possible. Once I do that, git pull will update and merge all remote branches including master, into my local branches including master, is this correct ?
23:41 Vampire0 malmalmal, no
23:41 malmalmal Oh
23:41 Vampire0 malmalmal, pull will fetch the upstream branch of the current branch and merge it into your local branch or if given via config or parameters rebase instead of merge
23:43 malmalmal Vampire0: which means that if checked out to master branch, locally, it will fetch and merge remote master branch into my local master branch ?
23:43 Vampire0 malmalmal, to update all remote tracking branches, use `git fetch --all` or `git remote update`
23:43 Vampire0 malmalmal, yes, if the remote master branch is the upstream branch of your local master branch
23:44 Vampire0 malmalmal, if you track another remote branch with your master branch, then that one is fetched and integrated
23:44 malmalmal Thanks Vampire0 , I think I am confused about the meaning of "upsteam", sorry if that is a RTFM
23:44 Vampire0 malmalmal, if you don't track any remote branch the command will complain about that
23:44 bronson joined #git
23:44 Vampire0 malmalmal, the upstream branch of a local branch is the remote branch you are tracking
23:45 Archrover joined #git
23:45 Vampire0 malmalmal, it is often the case that your master branch tracks the remote master branch, but it is not a must
23:45 malmalmal Ok Vampire0 , I assumed that local master would by default track remote master, and each local branch, track each remote branches
23:45 malmalmal Oh I see
23:45 Vampire0 malmalmal, if I work on a pull request I have e. g. a local branch fix-issue-5 that tracks upstream/master and pushes to origin/master (my fork)
23:46 Vampire0 malmalmal, by default after the clone there is only one local branch
23:46 Vampire0 malmalmal, either the one you told the clone command or otherwise the default branch of the repo you clone
23:46 malmalmal Ok, thank you Vampire0
23:47 Vampire0 malmalmal, this **can** be the master branch and most often will be, but doesn't have to
23:47 Vampire0 malmalmal, for a repo following git-flow principles the default branch should actually be set to the develop branch
23:47 malmalmal hmm, I see what you mean
23:48 malmalmal This is, I created a repo on a remote server. Now I should create a dev branch. Then I should clone that dev branch as my local master branch is this correct ?
23:49 malmalmal I am going to check git-flow as you suggest
23:50 oskarkv joined #git
23:50 druonysus joined #git
23:52 Vampire0 malmalmal, that was just an example, not a recommendation
23:52 malmalmal Vampire0: oh, ok
23:53 Vampire0 malmalmal, if you are interested in !workflows, follow the link and read. But for the start I'd start using basic Git usage first
23:53 gitinfo malmalmal: [!workflow] Finding the right workflow for you is critical for the success of any SCM project.  Git is very flexible with respect to workflow.  See http://sethrobertson.github.com/GitBestPractices/#workflow for a list of references about choosing branching and distributed workflows.
23:53 gyre007 joined #git
23:53 malmalmal Vampire0: Ok, thank you.
23:53 Vampire0 yw
23:53 malmalmal So, Vampire0 : As I understand 'git fetch -all' is what will fetch and merge all remote branches into their local branches
23:54 northfurr joined #git
23:55 malmalmal while 'git pull' will only fetch and merge the remote branch that is associated to the local branch I am actually checked out.
23:55 malmalmal Is this correct ?
23:56 cagomez joined #git

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