Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-12-03

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

All times shown according to UTC.

Time Nick Message
00:04 YuGiOhJCJ joined #git
00:05 chardan joined #git
00:08 tvw joined #git
00:09 YuGiOhJCJ joined #git
00:13 rwb joined #git
00:14 SourTrout joined #git
00:14 crayon joined #git
00:16 shashin joined #git
00:17 ResidentBiscuit joined #git
00:21 mischat joined #git
00:25 texinwien_ joined #git
00:29 dsantiago joined #git
00:29 dodobrain joined #git
00:31 aw1 joined #git
00:33 aw1 git diff --quiet && git diff --cached --quiet; echo $? should be sufficient to determine if my working directory is clean or not ... right? not bothered about untracked files
00:33 osse aw1: you can also do git diff --quiet HEAD
00:34 dendazen joined #git
00:34 osse aw1: this compares disk files against HEAD directly (ie. both unstaged and staged)
00:34 aw1 osse, great. thanks.
00:34 aw1 btw this is to be done inside of a script ...
00:34 osse aw1: you might want to source this:  https://github.com/git/git/blo​b/master/git-sh-setup.sh#L210
00:34 osse it's what the official git scripts use
00:35 astrofog joined #git
00:38 meauses joined #git
00:39 madewokherd joined #git
00:39 meth joined #git
00:40 scarabx joined #git
00:43 Orbitrix joined #git
00:46 steelsw0rd joined #git
00:46 angelica_ joined #git
00:46 Orbitrix joined #git
00:46 steelsw0rd joined #git
00:47 Orbitrix joined #git
00:47 alex1a joined #git
00:48 steelsw0rd joined #git
00:48 sargas joined #git
00:49 crayon joined #git
00:49 steelsw0rd joined #git
00:51 steelsw0rd joined #git
00:52 Naan joined #git
00:52 chachasmooth joined #git
00:55 ki0 joined #git
00:55 rorro joined #git
00:55 Learath2 joined #git
00:58 ramsub07 joined #git
00:58 ResidentBiscuit joined #git
00:59 Naan hi I'm new to git and just set up my first repo. I just pushed a commit to a branch on the command line but when I went on the webpage there was no pull request just an updated branch. So I had to create one before before merging it with master. Can someone explain where I went wrong?
01:02 Hello71 !pr
01:02 Hello71 !github
01:02 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)
01:03 Hello71 git does not recognize pull requests
01:03 Naan ok thank you!
01:03 joshszep joined #git
01:04 Naan left #git
01:04 joshszep joined #git
01:05 joshszep joined #git
01:05 joshszep joined #git
01:06 fling joined #git
01:06 joshszep joined #git
01:07 ramsub07 joined #git
01:07 joshszep joined #git
01:08 jon-mac joined #git
01:08 joshszep joined #git
01:09 joshszep joined #git
01:09 joshszep joined #git
01:10 fstd_ joined #git
01:10 BlueProtoman joined #git
01:11 anonymuse joined #git
01:11 Shentino joined #git
01:11 nidr0x joined #git
01:13 HoierM joined #git
01:21 hahuang65 joined #git
01:22 mischat joined #git
01:26 crayon joined #git
01:28 jaguarmagenta joined #git
01:32 nettoweb joined #git
01:33 arescorpio joined #git
01:38 mohan joined #git
01:40 hariel joined #git
01:44 durham_ joined #git
01:57 johnnyfive joined #git
02:00 anonymuse joined #git
02:01 Limes_ joined #git
02:02 ochorocho__ joined #git
02:06 dilfridge joined #git
02:06 Vampire0_ joined #git
02:06 steelsw0rd joined #git
02:08 svm_invictvs joined #git
02:08 durham joined #git
02:10 durham joined #git
02:14 pks joined #git
02:14 finalbeta joined #git
02:16 Puffball joined #git
02:17 dikiaap joined #git
02:21 meauses joined #git
02:23 mischat joined #git
02:24 thiago joined #git
02:28 d^sh joined #git
02:29 alex1a_ joined #git
02:31 steelsw0rd joined #git
02:33 dsantiago joined #git
02:34 sssilver joined #git
02:40 nettoweb joined #git
02:40 masuberu joined #git
02:46 Dougie187 joined #git
02:48 ilbot3 joined #git
02:48 Topic for #git is now Welcome to #git, the place for git help and serious business | Public logs at http://jk.gs/glog | First visit? Read: http://jk.gs/git | Current stable version: 2.11.0 | Getting "cannot send to channel"? /msg gitinfo .voice | The git-jokes project has been suspended for lack of contribution. Help revive it! https://madeitwor.se/git-jokes
02:49 nettoweb joined #git
02:52 railssmith joined #git
02:54 Vangelis joined #git
02:56 Samual joined #git
02:59 ResidentBiscuit joined #git
03:01 anonymuse joined #git
03:01 Rodya_ joined #git
03:02 sarbs joined #git
03:03 inflames joined #git
03:04 chachasmooth joined #git
03:05 d0nn1e joined #git
03:07 Goplat joined #git
03:10 sunri5e joined #git
03:11 hexagoxel joined #git
03:17 SourTrout joined #git
03:19 diogenese joined #git
03:23 Vangelis joined #git
03:24 mischat joined #git
03:25 govg joined #git
03:25 texasmynsted joined #git
03:29 jaguarmagenta joined #git
03:30 safe joined #git
03:31 holodoc joined #git
03:31 noecc left #git
03:34 Serus left #git
03:37 flaguy48 joined #git
03:38 sssilver joined #git
03:39 texinwien_ joined #git
03:40 rbern joined #git
03:43 kyan joined #git
03:44 chachasmooth joined #git
03:45 jgrasser joined #git
03:51 venmx joined #git
03:52 adymitruk joined #git
03:52 impermanence joined #git
03:52 impermanence Is anyone here that develops on a Mac?
03:52 impermanence and that uses eclipse, lol?
03:54 a_thakur joined #git
04:00 a_thaku__ joined #git
04:01 ResidentBiscuit joined #git
04:02 fahadash joined #git
04:03 ochorocho__ joined #git
04:03 impermanence Is anyone here that develops on a Mac?
04:03 impermanence and that uses eclipse, lol?
04:03 mizu_no_oto joined #git
04:04 nettoweb joined #git
04:14 Orbitrix joined #git
04:15 Orbitrix joined #git
04:15 Vortex34 joined #git
04:16 pks joined #git
04:17 adymitruk impermanence: why are you using eclipse?
04:21 impermanence adymitruk: it's a work thing.
04:21 impermanence I of course first went to #eclipse.
04:21 impermanence but no one is there.
04:22 impermanence so, then I hopped over to Java.  But it seems to be like one Windows guy that's a channel admin and then a bunch of students, so...wrong knowledge domain.
04:22 impermanence so then I thought...well...who will be up right now or in right now that actually knows about development and of course I immediately thought of #git.
04:23 Hello71 how is spamming here an improvement
04:23 Hello71 [04:22:42] -alis(alis@services.)- #eclipse                                           121 :Neon.1 and Eclipse 4.6.1 have been Released! Stay connected after you ask a question; different people check the channel log at different times.
04:23 Hello71 there are 121 users in #eclipse
04:24 impermanence It isn't.  I figured maybe there was actually no one in here that was active.
04:25 mischat joined #git
04:25 impermanence My issue is: no matter what key I specify eclipse to use, it always grabs the wrong ssh key when attempting to connect to github.  It is...elusive.
04:25 impermanence I've tried many things and they've all failed.
04:26 blackwind_123 joined #git
04:30 bn0t joined #git
04:30 fstd joined #git
04:32 Rodya_ joined #git
04:36 robotroll joined #git
04:37 impermanence left #git
04:43 diogenese joined #git
04:45 Poppa joined #git
04:48 diogenese joined #git
04:49 anonymuse joined #git
04:51 venmx joined #git
04:51 Motoko joined #git
04:54 chachasmooth joined #git
05:02 ResidentBiscuit joined #git
05:02 aidalgol joined #git
05:03 brent__ joined #git
05:05 dodobrain joined #git
05:05 dodobrain joined #git
05:07 justanotheruser joined #git
05:08 netj joined #git
05:10 chachasmooth joined #git
05:10 Rodya_ joined #git
05:15 SwiftMatt joined #git
05:19 sssilver joined #git
05:23 d0nn1e joined #git
05:26 mischat joined #git
05:29 ki0 joined #git
05:30 jaguarmagenta joined #git
05:31 hahuang65 joined #git
05:37 InfoTest joined #git
05:38 RustyShackleford suppose you have a commit that changes several files
05:38 RustyShackleford is there a way to pick out the changes from one file?
05:38 Nilesh_ joined #git
05:39 RustyShackleford so, remove those changes from the original commit, and create a new commit that has only the changes from that one file
05:40 ShapeShifter499 joined #git
05:50 bn0t_ joined #git
05:55 chachasmooth joined #git
05:59 Cabanossi joined #git
06:01 xberg joined #git
06:02 chachasmooth joined #git
06:03 ResidentBiscuit joined #git
06:03 ochorocho__ joined #git
06:04 aw1 joined #git
06:05 kesslerfrost joined #git
06:06 jleclanche As it turns out, git gets super slow when you rewrite a root commit extremely often :)
06:06 jleclanche yay for gc
06:08 TomyWork joined #git
06:11 chachasmooth joined #git
06:16 CEnnis91 joined #git
06:16 cyphase joined #git
06:16 romerocesar joined #git
06:21 bocaneri joined #git
06:21 b1tchcakes joined #git
06:24 SwiftMatt joined #git
06:24 cdown joined #git
06:24 jaguarmagenta joined #git
06:25 netj joined #git
06:26 kesslerfrost joined #git
06:26 mischat joined #git
06:27 chachasmooth joined #git
06:29 GT4066 joined #git
06:32 anonymuse joined #git
06:34 inflames joined #git
06:34 RustyShackleford suppose you have a commit that changes several files
06:35 RustyShackleford I want remove the changes for one file from the original commit, and create a new commit that has only the changes from that one file
06:35 RustyShackleford or maybe I want to seperate that commit so that each file's changes are in a different commit
06:37 dan2k3k4 joined #git
06:41 kesslerfrost joined #git
06:42 aw1 joined #git
06:44 jaguarmagenta joined #git
06:49 dan2k3k4 joined #git
06:51 venmx joined #git
06:52 bn0t__ joined #git
06:55 jmpp joined #git
06:55 jmpp joined #git
06:56 thiago RustyShackleford: is it the last commit you've made?
07:00 RustyShackleford thiago: sure thats how I was picking it
07:00 neunon joined #git
07:00 dermoth joined #git
07:01 RustyShackleford if theres a way to do that, you can always interactive rebase and reduce to a known solution
07:02 jstimm joined #git
07:03 TomyLobo joined #git
07:07 LucaTM joined #git
07:07 chachasmooth joined #git
07:08 sla3k joined #git
07:10 sla3k Hi people, I have a situation here, I added my key to ssh-agent and upon doing 'ssh -T git@github.com' I get the welcome message as mentioned in the helping article at github, but whenever I try to push my code to github (git push), it ask for my username and password.
07:10 sla3k What I might be doing wrong?
07:12 jaguarmagenta joined #git
07:15 kadoban sla3k: You are probably using HTTPS urls to access github via push? SSH can't work over those.
07:15 thiago RustyShackleford: git reset HEAD~ filename; git commit --amend; then create your new commit
07:15 thiago sla3k: what URL did you use when you pushed?
07:16 meauses joined #git
07:18 lindenle joined #git
07:18 qws-user-1229 joined #git
07:22 netj joined #git
07:24 a_thakur joined #git
07:26 ygeLN joined #git
07:27 mischat joined #git
07:30 mkilivan joined #git
07:30 jknetl joined #git
07:30 ki0 joined #git
07:33 Shentino joined #git
07:33 iliv joined #git
07:35 sea-gull joined #git
07:39 Oebele joined #git
07:41 sla3k kadoban: thiago: I just use 'git push' but inside the 'config' file, I see 'https' being used.
07:42 kadoban sla3k: Ya, don't do that then. Use the SSH style URLs, like  git@github.com:user/whatever
07:42 thiago sla3k: SSH keys only apply to SSH
07:43 mseyne joined #git
07:43 sla3k Got it, Thanks it is working now...
07:44 vuoto joined #git
07:45 breezy_ joined #git
07:48 Rodya_ joined #git
07:51 bmahe joined #git
07:54 flavius joined #git
07:55 tissuesamples2 left #git
07:55 Vangelis joined #git
07:56 jmpp joined #git
07:56 jmpp joined #git
07:57 Dragonbird joined #git
07:59 Mobutils joined #git
08:01 Caelum joined #git
08:04 ResidentBiscuit joined #git
08:06 ochorocho__ joined #git
08:09 qws-user-1228 joined #git
08:09 Caelum joined #git
08:10 dreiss joined #git
08:11 sweatsuit joined #git
08:13 neunon joined #git
08:13 AndrewAlexMac joined #git
08:14 www-bukolay-net joined #git
08:16 govg joined #git
08:21 bn0t_ joined #git
08:23 AndrewAlexMac joined #git
08:28 mischat joined #git
08:35 dikiaap joined #git
08:42 a_thakur joined #git
08:43 honggaruy joined #git
08:46 neunon joined #git
08:46 vrlx joined #git
08:48 ShekharReddy joined #git
08:48 neunon joined #git
08:49 Endarked joined #git
08:52 venmx joined #git
08:52 sea-gull joined #git
08:53 Oatmeal joined #git
08:53 xall_ joined #git
08:56 ResidentBiscuit joined #git
08:59 robotroll joined #git
09:00 jstimm joined #git
09:01 ki0 joined #git
09:01 neunon joined #git
09:03 Eryn_1983_FL joined #git
09:04 Spec-Chum joined #git
09:10 hahuang61 joined #git
09:13 spacelord_ joined #git
09:15 ochorocho__ joined #git
09:17 ki0 joined #git
09:21 _Soto joined #git
09:22 _Soto Hey everyone. I recently needed to rename two files and I used git mv to do it. The commit diff still shows the whole files being added anew and the old ones being deleted. Is that what's expected?
09:23 Peng try git diff -M
09:26 msonntag joined #git
09:27 _Soto Peng: Is it supposed to explicitly show the renames? It doesn't.
09:28 hahuang61 joined #git
09:29 mischat joined #git
09:29 _Soto Peng: http://pastebin.com/4AFdBWqF
09:30 Mobutils_ joined #git
09:33 chachasmooth_ joined #git
09:35 cdown_ joined #git
09:38 hahuang61 joined #git
09:39 _Soto It shows 'new file mode' for the newname file and 'deleted file mode' for the oldname file
09:41 ygeLN joined #git
09:42 Peng You should Google it, but Git doesn't really do renames. It's tracked as a delete and an add, and renames are statistically inferred after the fact, with commands like 'git diff -M'.
09:42 venmx joined #git
09:43 Peng If the file was also significantly modified, pass a lower percentage. E.g. git diff -M60
09:43 Peng wait, 50 is the default
09:43 Peng -M30 or something
09:48 dsdeiz joined #git
09:49 sea-gull joined #git
09:50 King_Hual joined #git
09:52 LucaTM joined #git
09:56 mischat joined #git
09:57 Nikitaw99 joined #git
09:59 function joined #git
10:00 function if I'm in middle of a rebase how can I programmatically tell if the current patch is empty ?
10:00 jost joined #git
10:01 cheater joined #git
10:01 eivarv joined #git
10:03 function git diff-index --quiet HEAD --
10:03 chachasmooth joined #git
10:04 osse function: depends... how did you get there? did you mark a commit 'edit' ?
10:04 maret joined #git
10:05 _Soto Peng: Thing is I significantly modified the file after renaming it, so something like git diff -M2% HEAD~1 HEAD works and shows me the rename because the similarity is very less. The problem occurs when I push to Github and it shows the changes as file deletion and addition.
10:05 cdown joined #git
10:06 diogenese joined #git
10:07 _Soto Peng: Thanks!
10:07 mischat joined #git
10:08 function osse I marked the commit keep; a previous edit make the new patch empty
10:09 function what I'm actually trying to do is script through parts of rebase: "attempt to skip empty patches, and abort if you find any conflicts"
10:09 osse "keep" isn't an option :O
10:09 YuGiOhJCJ joined #git
10:09 osse but yes, that git diff should do the trick
10:09 function sorry, I meant pick
10:09 function not keep
10:10 function :-)
10:10 function osse thanks!
10:10 function now to figure out how to write the loop that will abort,or continue until the end
10:10 function :)
10:11 spacelord_ joined #git
10:12 giraffe joined #git
10:13 seishun joined #git
10:15 bn0t__ joined #git
10:15 spacelord_ joined #git
10:15 ochorocho__ joined #git
10:17 Limes_ joined #git
10:18 afuentes joined #git
10:20 byte512 joined #git
10:21 Nikitaw99 joined #git
10:26 ChiLL_SurF joined #git
10:26 Chill-Surf joined #git
10:27 ki0 joined #git
10:28 ahmedelgabri joined #git
10:36 spacelord_ joined #git
10:36 satifant joined #git
10:39 cheater joined #git
10:45 spacelord_ joined #git
10:47 spacelord_ joined #git
10:50 CheckDavid joined #git
10:54 spacelord_ joined #git
10:56 ki0 joined #git
10:56 Jackneill joined #git
10:57 ResidentBiscuit joined #git
10:57 _Soto joined #git
11:01 bn0t_ joined #git
11:03 mischat joined #git
11:04 irqq_ joined #git
11:04 matumbo joined #git
11:09 [0xAA] joined #git
11:09 maxb joined #git
11:13 kier joined #git
11:18 saml joined #git
11:18 ochorocho__ joined #git
11:19 ResidentBiscuit joined #git
11:23 phanimah1sh joined #git
11:23 YuGiOhJCJ joined #git
11:23 ki0 joined #git
11:24 Darcidride joined #git
11:25 robotroll joined #git
11:26 FederationOfNULL joined #git
11:29 xberg joined #git
11:32 Shentino2 joined #git
11:33 Duckily joined #git
11:33 yohnnyjoe joined #git
11:34 Duckily Hey there, I output my $PATH on osx and see /usr/local/git/bin added but it’s not in .bash_profile or .bashrc
11:34 Duckily Where would git add itself
11:35 xberg_ joined #git
11:38 _ikke_ Duckily: How do you mean?
11:38 _ikke_ Ah, now I understand
11:39 Duckily _ikke_: I’m not sure where it’s coming from
11:39 Duckily not in /etc/paths either
11:39 _ikke_ Duckily: This is not something git does itself. I think brew/macports is doing this
11:42 Puffball joined #git
11:42 m0viefreak joined #git
11:44 vuoto joined #git
11:46 finalbeta joined #git
11:48 shinnya joined #git
11:49 ispn joined #git
11:49 osse Duckily: it's a mouthful, but try the following:
11:49 osse PS4='+$BASH_SOURCE:$LINENO:$FUNCNAME: ' bash -xlic : 2>&1 | grep PATH
11:50 Duckily im scared
11:50 Duckily osse: what does it do?
11:50 xberg joined #git
11:51 osse Duckily: prints everything bash does when it starts up
11:51 osse with file and line number info
11:51 osse and then just quits
11:52 CEnnis91 joined #git
11:52 Duckily thats nifty
11:54 osse Duckily: x: print what it's doing. l: pretend it's a login shell.  i: pretend it's interactive. c: run the following command.  the colon is the dummy command. it could be 'true' or 'exit'
11:54 cyphase joined #git
11:54 osse Duckily: PS4 is like PS1 (a prompt), but is used for what the -x prints
11:54 xberg_ joined #git
11:57 Duckily osse: the only files it lists are /etc/profile and .bashrc
11:57 Duckily but neither of those contain anything about git
11:58 osse Duckily: what does  echo $0  say ?
11:58 osse and what did it print about PATH
11:58 mindriot101 joined #git
11:58 Duckily -bash
11:58 rorro joined #git
11:59 Duckily this: /usr/bin:/bin:/usr/sbin:/sbin:/u​sr/local/bin:/usr/local/git/bin
11:59 osse I mean what did my command print
11:59 dsfreak93 joined #git
12:01 xberg joined #git
12:02 Muzer joined #git
12:02 ramsub07 joined #git
12:02 Duckily osse: http://pastie.org/10975526
12:03 osse /etc/profile then
12:03 osse that's what sets it.
12:06 jknetl joined #git
12:06 jerv joined #git
12:07 jerv How can I use git rev-list to count the total amount of commits up to a hash? By default, git rev-list --count <rev> only counts revisions that are reachable from rev, but I want to count all revisions. There exist --all, but it doesn't work relative to a revision.
12:08 osse jerv: try git rev-list --count --all --not hash
12:08 dan2k3k4 joined #git
12:09 HoierM joined #git
12:10 Balliad joined #git
12:10 jerv hmm doesn't appear to be correct, will look into --not
12:11 Duckily osse: interesting it looks like its being added by the osx path_helper application
12:12 Duckily osse: really interesting, it concocts a path based on what’s in  /etc/paths.d
12:13 Duckily I vaguely remember setting the path somewhere in another file outside of bash but I can’t seem to find it
12:13 osse jerv: hmm, maybe you have to explicity list all the stuff instead of using --all
12:13 osse jerv:  git rev-list --count foo bar baz --not quux
12:14 osse which afaik is equivalent to:  git rev-list --coint quux..foo quux..bar quux..baz
12:15 jerv osse: given that I want to start with the first commit, is there a symbolic name for this one?
12:15 xberg joined #git
12:15 jerv osse: I'm basically looking for a local revision number for each commit.
12:17 manuels joined #git
12:19 eb0t joined #git
12:21 osse jerv: that's just  git rev-list --count commit
12:21 jstimm joined #git
12:21 osse there is no name for it. there might be more than one
12:22 jerv osse: the problem with --count commit is that it will only count commits reachable from commit. So it will not count branched commits.
12:22 osse Duckily: grep git /etc/paths.d/*
12:22 HandheldPenguin joined #git
12:23 osse jerv: it will count the branches that were merged in before the given commit
12:23 Duckily osse: yeah git is in there, but i swear there’s another file that has the other strings somewhere
12:23 Duckily export $PATH=“/bin:/usr/sbin:/sbin”
12:23 osse no $ in front
12:24 Duckily something like that, i’ve seen it
12:24 Duckily and i can’t for the life of me remember where, and the engineer in me is saying it’s impossible given the script you showed me
12:24 osse unless it's already set in the environment at boot (before even the first shell starts) it has to be that one
12:24 ChiLL_SurF joined #git
12:24 osse do you have a /etc/environment ?
12:25 jerv osse: the repo has branches that are not merged into the main line. So commit numbers will not be unique.
12:25 osse jerv: that's true
12:25 osse that's why git doesn't use them
12:25 Duckily nope
12:25 Duckily this is driving me insane
12:25 jerv osse: I guess I could git rev-list --all and then do a hash look-up.
12:26 xberg joined #git
12:26 osse jerv: a local revision number for each commit *is* git rev-list --count commit. Unless you mean something else
12:26 osse that commit has count-1 commits before it.
12:27 jerv osse: I'm looking for a revision number that has the property that it is (locally) unique and monotonically increasing
12:27 jerv osse: relative to _all_ other commits in the repo, not just commits on the same branch
12:27 linduxed joined #git
12:27 jerv osse: e.g. mercurial has hg log -T {rev}
12:28 Duckily osse: is there a /bash_profile outside of ~
12:29 nettoweb joined #git
12:29 osse jerv: AFAIK for the purposes of this discussion mercurial works the same as git: commits have parent(s), it's a directed acyclic graph etc. etc.  so I suggest you investigate how hg calculates that number
12:30 osse Duckily: not one that's read by your shell at least. otherwise it would have appeared in the output
12:30 Cavallari joined #git
12:31 Duckily osse: http://pastie.org/10975548
12:31 Duckily here’s my /etc/profile so there’s a good chance the applciation could be loading something from somewhere else
12:31 nowhereman joined #git
12:32 osse I have no idea what path_helper does. BUt it seemingly prints a line of the form "PATH=..." which is then evaluated in the current shell
12:32 osse from the output we see that line has /usr/local/git/bin in it
12:32 osse not sure what else you're looking for
12:32 Duckily its an application osx uses to construct the path
12:33 Duckily im 80% sure that i’ve edited a file that contains a single export line with various /usr/bin’s coloned together
12:33 Duckily and that i edited it again recently and can’t remember for the life of me where that was
12:34 Duckily but im becoming less and less sure
12:34 jerv osse: I guess my problem is that it's efficient for git to run down the commit graph child -> parent, but not the other way around, which is what I need.
12:34 Duckily seems like the path is being created completely there
12:34 osse jerv: the other way around does not exist in git. afaik it doesn't in hg either, but maybe that's more efficient somehow
12:35 osse Duckily: maybe you edited some file that has no effect :p
12:35 osse Duckily: was it for a different shell maybe? ksh or zsh ?
12:35 jerv osse: I guess what I could do is `get rev-list --all`, which prints all commits in topological order. Then I basically have a map hash => number.
12:35 Duckily osse: im beginning to think you’re right
12:36 cheater joined #git
12:36 Duckily any magic grep line for searching for the string in a file without killing my computer, maybe only hidden files
12:37 Duckily Something like egrep -ir --include=".*" --exclude="Applications" "/usr/sbin" .
12:37 jerv osse: the difference with hg is that it literally associates the number at commit time, which makes it efficient. It's stored in the index.
12:38 xall_ joined #git
12:39 Duckily zsh has a path but nothing excisiting there
12:40 lindenle joined #git
12:42 ramsub08 joined #git
12:43 mindriot101 joined #git
12:44 Duckily osse: alright im going to give up for now, thanks for all the help!
12:44 overlord_tm joined #git
12:44 FederationOfNULL joined #git
12:46 FederationOfNULL joined #git
12:47 jerv osse: fwiw, this is my "solution" for now: `git rev-list --all | grep -n <HASH>`
12:49 spacelord_ joined #git
12:55 Starky joined #git
13:01 dan2k3k4 joined #git
13:09 jstein_ joined #git
13:09 qws-user-1228 joined #git
13:10 mephistorium joined #git
13:11 dan2k3k4 joined #git
13:17 courrier joined #git
13:17 steelsw0rd joined #git
13:20 ResidentBiscuit joined #git
13:22 dan2k3k4 joined #git
13:23 sylario joined #git
13:24 sdothum joined #git
13:25 scarabx joined #git
13:28 mseyne joined #git
13:30 ace_33 joined #git
13:30 govg joined #git
13:30 dendazen joined #git
13:32 inflames joined #git
13:32 rwb1 joined #git
13:32 cdown joined #git
13:34 ResidentBiscuit joined #git
13:35 dan2k3k4 joined #git
13:36 jeffreylevesque joined #git
13:36 mda joined #git
13:37 jaguarmagenta joined #git
13:44 mkoskar joined #git
13:45 isysd joined #git
13:45 dan2k3k4 joined #git
13:47 CrazyBonz_ joined #git
13:48 spacelord_ joined #git
13:50 seishun joined #git
13:52 kesslerfrost joined #git
13:58 p4trix joined #git
13:58 daishan joined #git
14:00 Nilesh_ joined #git
14:01 dan2k3k4 joined #git
14:03 spacelord_ joined #git
14:04 snowkidind joined #git
14:05 User458764 joined #git
14:05 _Soto joined #git
14:08 DrRobotic joined #git
14:11 manuels joined #git
14:13 mephistorium joined #git
14:15 YuGiOhJCJ joined #git
14:16 dan2k3k4 joined #git
14:17 Nikitaw99 is there an irc bot that posts in a channel when there are new commits in a specific repository?
14:18 Dragonbird joined #git
14:19 Zarthus the github hook and notifico are popular, but github hooks are about to be deprecated and I don't think anyone has contributed a new integration solution yet
14:20 Zarthus if you're looking for something local; that really depends on your configuration.
14:20 Zarthus you could write something really simple and install a post-commit hook, I suppose.
14:21 Dragonbird joined #git
14:23 ahmedelgabri joined #git
14:26 shashin joined #git
14:27 dan2k3k4 joined #git
14:28 cheater joined #git
14:34 xall_ joined #git
14:34 mkoskar joined #git
14:35 navidr joined #git
14:49 hfp_work joined #git
14:51 aw1 joined #git
14:54 spacelord_ joined #git
14:54 musicmatze Hi #git! I have a question: Why does git-log not work with both --graph and --reverse - are there technical reasons for this or is it just not implemented?
14:55 GavinMagnus joined #git
14:55 grawity I'm guessing nobody just bothered to implement the drawing of an upside-down graph
14:55 grawity technically it doesn't make much of a difference
14:56 BlessJah except for it being harder to reverse-walk history graph
14:56 grawity well, I imagine it'd work the same way as regular --reverse
14:57 musicmatze BlessJah: you wouldn't even do this, would you? You would just print the graph in reverse order, so the commits are already fetched, just the printing mechanism would be another...
14:57 grawity which already cannot "reverse-walk" the history
14:57 grawity so it has to obtain the entire graph upfront
14:57 Limes_ joined #git
14:57 grawity and more or less just renders stuff upside-down
14:58 BlessJah you could always use tac
14:59 musicmatze BlessJah: yes, but then you can only have one-line commits printed. Not the whole commit message, or even some nice two/three/n-line output per commit
15:00 musicmatze you know... sometimes it is a bit hard to read a history bottom-to-top if you want to read commit messages :-)
15:02 breezy_ joined #git
15:02 hfp_work joined #git
15:02 grawity tbh, I usually just run `tig` and use its split view
15:03 grawity https://i.imgur.com/6vSYGjU.png might well be the best git tool ever
15:05 ResidentBiscuit joined #git
15:06 musicmatze grawity: yes, I use that too ... it is really awesome!
15:06 musicmatze but there are times when I want to use plain git for doing these things
15:06 seishun joined #git
15:06 * canton7 just fires up gitk
15:07 musicmatze naah, I don't want to click around! :-)
15:08 ResidentBiscuit joined #git
15:10 Nikitaw99 yeah, we want to type around instead ;))
15:13 bn0t__ joined #git
15:13 jeffreylevesque joined #git
15:14 bn0t joined #git
15:21 pdgwien joined #git
15:21 DolphinDream joined #git
15:21 iliv joined #git
15:22 k6ekb joined #git
15:24 vrlx joined #git
15:26 xall_ joined #git
15:27 dsfreak93 joined #git
15:28 robattila256 joined #git
15:29 function joined #git
15:32 Cathyhav joined #git
15:32 Cathyhav left #git
15:33 Ijcor joined #git
15:33 anonymuse joined #git
15:44 lvns joined #git
15:44 _Vi joined #git
15:45 xcesariox joined #git
15:49 seishun joined #git
15:51 jstein_ joined #git
15:52 ramsub07 joined #git
15:52 sea-gull joined #git
15:57 Spec-Chum joined #git
15:58 Derperperd joined #git
16:00 jstimm joined #git
16:03 vuoto joined #git
16:03 beny joined #git
16:05 lvns joined #git
16:05 LucaTM joined #git
16:07 masuberu joined #git
16:10 a3Dman joined #git
16:11 mizu_no_oto joined #git
16:12 jaguarmagenta joined #git
16:12 ojdo joined #git
16:14 jnavila joined #git
16:16 sq-one joined #git
16:17 lvns joined #git
16:20 texinwien_ joined #git
16:20 aollier joined #git
16:22 jnavila aollier: c'est toi ?
16:27 aollier je suis là
16:27 Nikitaw99 joined #git
16:31 _ikke_ Que?
16:32 a_thakur joined #git
16:34 cdg joined #git
16:34 ResidentBiscuit joined #git
16:35 franl joined #git
16:36 jnavila _ikke_: no problem, I was waiting for a newcommer
16:40 bremner _ikke_: manuel from barcelona died
16:42 robotroll joined #git
16:42 scarabx joined #git
16:43 YuGiOhJCJ joined #git
16:47 paws_ joined #git
16:49 lvns joined #git
16:50 kaldoran joined #git
16:52 nikivi joined #git
16:52 Seveas and Basil is very sad.
16:54 nowhereman joined #git
16:55 gchristensen left #git
16:56 jstimm joined #git
16:57 xwid joined #git
16:58 Naan joined #git
17:01 Literphor joined #git
17:01 madewokherd joined #git
17:01 Nikitaw99 joined #git
17:02 lvns joined #git
17:03 Learath2 joined #git
17:04 romerocesar joined #git
17:05 CEnnis91 joined #git
17:06 HandheldPenguin` joined #git
17:11 Nikitaw99 joined #git
17:11 dviola joined #git
17:13 ahmedelgabri joined #git
17:15 Nikitaw99 joined #git
17:20 cdg_ joined #git
17:20 dreaddymck joined #git
17:21 jknetl joined #git
17:21 chachasmooth joined #git
17:24 jason237 joined #git
17:25 xcesariox joined #git
17:25 chachasmooth_ joined #git
17:28 diogenese joined #git
17:30 sargas joined #git
17:31 ResidentBiscuit joined #git
17:35 kaldoran joined #git
17:35 dikiaap joined #git
17:36 User458764 joined #git
17:37 ShekharReddy joined #git
17:37 Emperor_Earth joined #git
17:37 chachasmooth joined #git
17:38 ybden joined #git
17:41 Derperperd joined #git
17:44 nikivi joined #git
17:49 romerocesar joined #git
17:52 seishun joined #git
17:53 wgrant joined #git
17:54 mda1 joined #git
17:56 jaguarmagenta joined #git
17:58 fstd joined #git
18:00 cdg joined #git
18:04 bodhi joined #git
18:08 jaguarmagenta joined #git
18:10 Dazzyp joined #git
18:14 NimeshNeema joined #git
18:17 anonymuse joined #git
18:18 muelli joined #git
18:18 texinwien_ joined #git
18:20 steelsw0rd joined #git
18:21 cyberz joined #git
18:22 kbs joined #git
18:23 Ryanar joined #git
18:24 Vangelis_ joined #git
18:25 darkcar joined #git
18:26 jaafar joined #git
18:29 adymitruk joined #git
18:29 Literphor joined #git
18:29 jaguarmagenta joined #git
18:30 nikivi joined #git
18:33 amdi_ joined #git
18:35 ryouma joined #git
18:38 chachasmooth joined #git
18:38 ryouma i have git set up so that when i pull, it rebases my local patches on top of upstream.  this works.  but i want to get the sha for the head of vanilla upstream.  i.e. just before my patches go on top.  head and orig_head do not point there.  what does?
18:39 Sna4x8 joined #git
18:40 chachasmooth_ joined #git
18:41 jstimm joined #git
18:42 Sna4x8 I have a repository and corresponding node module that has some DB functionality.  The repository/module is currently tied to MySQL; in the new version I need to support other databases.  In the new version, I want to make a module for each database I support (my-module-mysql, my-module-mssql, my-module-postgres, etc.).
18:42 Sna4x8 Should I fork my repository to a new one, and rename that forked version?
18:44 Sna4x8 I need to keep the current repository and module (my-module) intact so I don't break my users' code bases.
18:44 YuGiOhJCJ joined #git
18:46 krillr joined #git
18:47 thunderbirdtr joined #git
18:47 ryouma hmm, can i get the sha for master even if i am not on master?
18:47 osse ryouma: @{u} does
18:48 ryouma osse: what's u?
18:48 osse you can get the sha of anything at any time
18:48 ryouma osse: what command does this?
18:48 osse git rev-parse master.  git rev-parse master@{u}.  git rev-parse master@{upstream}
18:49 hahuang61 joined #git
18:49 BlessJah why not origin/master (i.e. name remote explicitely)
18:50 osse why not? why? :p
18:50 mda joined #git
18:50 osse what if it's origin/maestro
18:50 osse also you can do HEAD@{u}  so it works for any branch
18:51 kaldoran joined #git
18:51 ryouma osse: works.  thank you.
18:55 qws-user-1229 joined #git
18:56 snowkidind joined #git
18:58 Keiz joined #git
18:58 snowkidind joined #git
18:59 function joined #git
18:59 cagedwisdom joined #git
19:00 Keiz Hey! I want to use amend to add a trivial change to my previous commit. This commit has already been pushed to a feature branch which no one has branched off and neither has it been been merged to the master branch. I'll probably have to force push as well. Is it completely okay in this scenario to force push?
19:00 osse Keiz: you will definitely have to force push. it's ok if you're sure
19:01 Keiz osse: Sure of? Nobody branching off? What are the common pitfalls in force pushing?
19:01 osse Keiz: yes, that.
19:01 chachasmooth joined #git
19:02 osse Keiz: the common pitfall is that the history can become a godawful mess
19:02 osse you merge an "old" branch with the "new" (rebased version) branch leading to what seems to be duplicate commits in the log
19:02 osse other people who have base their work also risk this, unless they're handy with rebase --onto
19:03 osse !rewrite
19:03 gitinfo Rewriting public history is not recommended. Everyone who has pulled the old history will have to do work (and you'll have to tell them to), so it's infinitely better to just move on. If you must, you can use `git push --force-with-lease <remote> <branch>` to force (and the remote may reject that, anyway). See http://goo.gl/waqum
19:04 Keiz Ah I see. May reject?
19:06 kadoban joined #git
19:07 osse reject what?
19:07 BSAlb joined #git
19:07 Keiz 'and the remote may reject that, anyway'
19:08 Zarthus it'll be easier for everyone involved if you don't amend & force push if this is a popular project and branch, unless the commit is very recent
19:09 Keiz Zarthus: Thanks I get it now.
19:10 osse Keiz: it just means that the server may reject the push even if you --force
19:10 jason237 joined #git
19:10 BSab joined #git
19:11 osse as in, the push won't complete.
19:11 btree joined #git
19:11 Keiz osse: I understand the language but I meant in what circumstances doesn't the push work?
19:12 osse Keiz: That's a broad question. The server may implement a hook to inspect what's being pushed. Based on that it may deny your push.
19:12 osse The reason might be anything. The most common one is probably to deny force-pushes
19:13 Keiz osse: Okay I'll read up on those then when I can. Thanks :)
19:14 mkoskar joined #git
19:14 osse Another one might be to deny pushes that modify specific files, or deny pushes whose commit messages don't meet a particular standard
19:15 rorro joined #git
19:19 anonymuse joined #git
19:19 msonntag joined #git
19:20 pbrewczynski joined #git
19:23 romerocesar joined #git
19:25 isysd joined #git
19:26 fiddlerwoaroof left #git
19:29 steelsw0rd joined #git
19:31 Rodya_ joined #git
19:33 msonntag joined #git
19:33 ResidentBiscuit joined #git
19:34 mda1 joined #git
19:35 ResidentBiscuit joined #git
19:38 sea-gull joined #git
19:40 seishun joined #git
19:46 g105b_ joined #git
19:48 mseyne joined #git
19:50 kaldoran joined #git
19:51 Saturn812 joined #git
19:53 dendazen joined #git
19:55 JanC_ joined #git
19:55 GavinMagnus joined #git
19:56 lindenle joined #git
19:59 tnecniv joined #git
19:59 trout how can I run something like
19:59 trout git log -1 --pretty=format:"%<(15)%C(green)%cr%C(white) | %C(cyan)%h%C(white) | %<(12)%C(yellow)%an%C(white) | %s" -- peval
19:59 trout on multiple files
19:59 trout I if I add multiple files at the end nothing shows up
20:00 trout and if I remove the -1, then I get the complete history
20:00 trout I want the "latest commit for each of the files I list"
20:00 thiago probably because there's no commit that modified all of those files at the same time
20:00 trout right now I run the above in a loop
20:01 trout I'm looking to see if I could run it as a single command
20:01 thiago hmm.. no, that's not how git log works
20:01 trout I'm happy to use something other than git log
20:01 thiago listing multiple files there will show the latest commit that touches any of those files
20:01 trout I just want "for each file, the last sha that modified it"
20:01 trout does something like that exist ^ ?
20:01 m0viefreak joined #git
20:01 thiago no, you have to write your own loop
20:01 osse either one git log per file name, or one complete file name and keep track of the files
20:01 _Patrick_ joined #git
20:01 trout okay - that's what I have, thanks
20:02 osse ehh, one complete log
20:02 trout any suggestions to make this better/faster would be great: https://github.com/unixdotfiles/b​in/blob/master/git-last-modified
20:04 prg3 joined #git
20:04 osse use git log --name-only so it lists all modified files. for each new file found note the current log. keep going until all wanted files have been seen, then stop
20:04 osse should only require one git log run.
20:04 thiago use git whatchanged instead of git log, so it lists the file names too
20:05 btree left #git
20:05 thiago then use one single run of it, and get the last time the file was modified
20:05 trout I think that would be actually slower for a Very Large Log
20:05 trout where its possible that a file last changed ~ 15 years ago
20:05 thiago and drop those colours
20:05 Starky joined #git
20:05 hahuang61 joined #git
20:06 thiago it would be faster than your code
20:06 osse trout: how so? the git log -- oldfile command would have to go back so far anyway
20:06 osse hmm, but it doesn't have to deal with other trees I guess
20:06 osse not sure if there's an optimization there
20:06 trout yeah, exactly
20:07 trout its not a real problem: script takes about 30 seconds
20:07 trout on the Very Large Repo I'm talking about
20:07 trout was hoping to make it faster, but 30 seconds is within the realm of usable
20:07 trout thanks all for the help
20:07 osse "yeah, exactly" what? i'm arguing against you :p
20:08 trout [12:06:47]  <osse>not sure if there's an optimization there
20:08 trout that?
20:08 trout I think my mental map of how "git log" works when given a single file is not correct
20:08 thiago if you have two files that were modified 15 years ago, your code will scan back in history twice
20:08 thiago with our suggestion, only once
20:08 trout I'll try it with the single "git log --name-only" format
20:08 osse I think 'git log -- file' is implemented by 'git log -p' then just hiding commits that don't touch the file.
20:09 trout ah, I thought it was more optimized than that
20:09 osse for it to be awesome you'd have to run it in parallel, though. so you can kill it asap
20:09 trout lol true
20:09 osse it may be. I don't know
20:09 trout I'll read the code
20:09 thiago of course, that was implied
20:09 thiago open a pipe, read as it progresses, close it when you're done
20:10 trout thiago ofc
20:10 thiago git log will receive SIGPIPE and exit
20:10 trout I'll do that next
20:10 * trout goes to finish unrelated stuff
20:11 osse very simple test on my VM:  git log -- git.c takes 0.37s and git log --name-only 2.25s
20:12 thiago for all files?
20:12 osse so if I do git log -- some_file.c more than 5 times the --name-only approach is faster
20:12 osse thiago: yes
20:12 rorro joined #git
20:12 thiago can't confirm
20:13 thiago git log -- filename > /dev/null => 0.40s
20:13 thiago git log --name-only -- filename > /dev/null => 0.40s
20:13 InfoTest joined #git
20:13 blunaxela joined #git
20:13 osse in that case the answer to your question was "no" :p  I didn't use -- filename when I tried --name-only
20:14 thiago well, yeah, then you listed all files
20:14 thiago the point is that if you want to know the last time that N files changed, you don't need to list all other files in the repo
20:14 cdown_ joined #git
20:14 thiago you restrict it to the files whose information you want
20:14 osse sure
20:15 osse if you don't want all files you'd do  git log --name-only -- list of files here
20:15 Silenced joined #git
20:15 Goplat joined #git
20:16 re1 joined #git
20:16 romerocesar joined #git
20:16 cheater left #git
20:17 osse Or maybe not?!  Here 'git log --name-only -- *.c' is much slower than 'git log --name-only'
20:18 thiago hmm
20:18 thiago sorry, still can't confirm
20:18 thiago 'git log --name-only -- some files here' is faster than both:
20:19 thiago a) 'git log --name-only'
20:19 thiago b) the sum of 'git log --name-only -- each file'
20:19 thiago hmm, wait, I forgot -1 for the latter
20:19 thiago the -1 improves, but not sufficiently
20:19 osse how many files? turns out I tried with 175
20:20 thiago 14
20:20 thiago let me try with more
20:20 rena_ joined #git
20:20 Noldorin joined #git
20:20 osse yes, 10 here is faster than without a file list
20:20 osse but a lot of files is slower
20:21 osse FYI I used some files from the root dir of git.git.  Maybe subdirs and stuff has a big effect?
20:21 re1_ joined #git
20:22 mindriot101 joined #git
20:23 thiago I'm running on a subdir with files on sub-subdirs
20:23 thiago 1034 files, still going
20:23 ramsub07 joined #git
20:23 thiago ok, it ran for 75 seconds
20:23 thiago 75.94 s, running again
20:24 rena_ joined #git
20:24 rena_ Hi there, is there a way to add a file to my repo without having git counting its lines? It's a pretty big file and was wondering if there's another way maybe?
20:24 osse what do you mean? when does this counting happen?
20:25 thiago rena_: Git doesn't count the lines
20:25 thiago therefore, the strict answer to your question is "simply run git add"
20:26 thiago new run was 74.59 s
20:27 thiago git log --name-only -- the same files here: 19.5 seconds, two runs
20:28 thiago osse: ^^
20:28 rena_ oops sorry. I meant the number of lines shown on github...
20:28 thiago rena_: sorry, we can't control GitHub
20:28 thiago rena_: do you have any problems with git?
20:29 thiago does it happen during a push? Or in a browser?
20:30 rena_ browser
20:30 thiago then please ask #github
20:30 rena_ okay;(
20:31 rena_ ;) *
20:31 rena_ left #git
20:32 d0nn1e joined #git
20:32 oskarkv joined #git
20:34 WayToDoor joined #git
20:39 osse thiago: hmm, now i'm not sure which numbers you're comparing.  19.5 secs for which same files? the 1034 ?
20:42 King_Hual joined #git
20:43 Discovery joined #git
20:44 tristanp joined #git
20:45 jstimm joined #git
20:46 IlNominePatre joined #git
20:47 SteffanW joined #git
20:49 systemovich joined #git
20:49 msonntag joined #git
20:49 mindriot101 joined #git
20:50 kaldoran joined #git
20:52 irqq joined #git
20:59 astrofog joined #git
21:00 ahmed_elgabri joined #git
21:01 Nikitaw99 joined #git
21:05 systemovich joined #git
21:07 anonymuse joined #git
21:07 Derperperd joined #git
21:08 aidalgol joined #git
21:10 cdown joined #git
21:15 NeverTired joined #git
21:16 khfeng joined #git
21:17 jnavila joined #git
21:18 AfroThundr joined #git
21:19 tmsmith joined #git
21:20 GavinMag1us joined #git
21:22 tmsmith joined #git
21:23 seishun joined #git
21:23 Limes_ joined #git
21:24 Literphor joined #git
21:25 meth joined #git
21:26 meth_ joined #git
21:27 tmsmith joined #git
21:28 tmsmith joined #git
21:29 tmsmith joined #git
21:30 GavinMagnus joined #git
21:30 jaguarmagenta joined #git
21:30 TikityTik joined #git
21:31 GavinMagnus joined #git
21:31 TikityTik Is it possible to avoid merges after I commit and want to push?
21:31 TikityTik Meaning I can have the merge commit combine with the commit I made?
21:32 TikityTik Do I have to git pull before I do git commit?
21:33 ToBeCloud joined #git
21:33 tobiasvl TikityTik: !rebase
21:33 gitinfo TikityTik: 'git rebase' takes away your local commits, updates your branch with new stuff from <upstream> (argument), then re-applies your local commits on top. This makes it look like your commits were created "after" the new stuff, and it can look cleaner than doing a !merge. Beware of !rewriting_public_history, though. Not to be confused with !interactive_rebase.
21:34 tmcmahon joined #git
21:34 TikityTik Is the only way through rebasing?
21:34 jstimm joined #git
21:34 TikityTik Or is it because I should always git pull before doing git commits?
21:34 thiago if you want to avoid merges, don't merge
21:34 TikityTik Is that the only way*
21:35 thiago what is the problem with avoiding merges?
21:35 Xe TikityTik: git pull --rebase will 99% of the time do what you need in that case
21:35 TikityTik thiago: I just don't want the commit history cluttered with merge commits, isn't it messy?
21:35 Rodya_ joined #git
21:35 thiago TikityTik: it usually is
21:35 thiago TikityTik: so avoid them
21:35 tobiasvl TikityTik: rebase is the way, yes. otherwise you can always pull before committing, to be sure, but that gets busy fast (in git you usually commit often)
21:35 tmsmith joined #git
21:35 TikityTik Should I not have merges in my commit history? Or is it a normal thing to have commit history cluttered with merges?
21:35 tobiasvl TikityTik: in your case, just "git pull --rebase" now and it will fetch and then rebase for you
21:36 thiago it's a choice
21:36 thiago you can develop entirely without merges, or with a lot of them
21:36 tobiasvl TikityTik: up to you, usually there are some merges in the history, but merges like one you face now can be avoided with rebase like I said
21:36 TikityTik well as a rookie git user, what should I do?
21:36 thiago your choice
21:36 Xe i choose to not give any attention to merges
21:36 TikityTik I don't know if it's important to have those merge commits
21:36 Xe i have bigger problems, like actually writing software
21:36 bn0t_ joined #git
21:36 thiago most people fall in between "a few are ok" and "a merge per feature contribution, bug fix, etc. is fine"
21:37 GavinMagnus left #git
21:37 TikityTik what if I already have git pulled?
21:37 thiago then it's already merged
21:37 TikityTik and then I would need to do git rebase?
21:37 thiago you'd have to first remove the merge commit
21:37 thiago which you can't if you've pushed to other people
21:38 TikityTik well if it's not pushed
21:38 thiago so you can remove that merge commit
21:38 thiago git reset --hard <the commit before the merge>
21:38 thiago remember: this will throw away any uncommitted changes too
21:38 TikityTik would that be HEAD~1?
21:39 TikityTik why not git rebase if it's not pushed?
21:39 thiago yes, provided the merge commit is HEAD
21:39 thiago because you have to remove the merge first
21:39 TikityTik thiago: would it be HEAD if i just git pull and merged?
21:40 thiago yes
21:40 TikityTik thanks
21:41 kaldoran joined #git
21:41 TikityTik typically have too many merges from editing the README through github
21:41 TikityTik like it's TODO list
21:41 TikityTik its*
21:41 shashin joined #git
21:41 thiago stop editing via github
21:42 TikityTik that is true, but I like being able to see the preview.
21:42 osse TikityTik: if it's important to have on record that different lines of development happened in parallel then merging is useful. With rebasing it just looks like things happened in sequence when in truth they didn't. As a newbie it's of course hard to tell whether you want merges or not.  I recommend doing a bit of both. Find out what you like
21:44 vuoto joined #git
21:47 HoierM joined #git
21:48 Literphor left #git
21:51 bn0t__ joined #git
21:51 jstimm joined #git
21:57 diogenese joined #git
22:00 neunon joined #git
22:00 tnecniv joined #git
22:03 dikiaap joined #git
22:03 cagedwisdom joined #git
22:06 Raed|Mobile joined #git
22:07 anonymuse joined #git
22:11 brent__ joined #git
22:13 aollier left #git
22:14 ski7777 joined #git
22:15 yohnnyjoe joined #git
22:19 tristanp joined #git
22:19 rwb joined #git
22:21 kaldoran joined #git
22:26 BPL joined #git
22:26 ski7777 joined #git
22:36 masak joined #git
22:37 masak remind me: how do I dry-run a merge? or rather, how do I ask whether two branches would merge cleanly without conflict?
22:37 diogenese joined #git
22:37 masak I leafed through `man git-merge`, but nothing stood out as helpful
22:37 gitinfo the git-merge manpage is available at http://jk.gs/git-merge.html
22:38 anonymuse joined #git
22:38 osse masak: try git merge-tree
22:39 mizu_no_oto joined #git
22:39 masak ah, makes sense
22:39 masak so I'd have to locate the base-tree myself, yes?
22:39 * masak ponders how much that'd end up re-inventing the 'recursive merge' wheel
22:40 osse masak: git merge-baase will tell you the base
22:41 masak ah, and `git merge-base` gives me... right
22:41 masak I can work with this :)
22:41 BPL hi, any recommended git command to use it as a package version?
22:41 osse BPL: git describe
22:41 BPL osse: I was trying that before asking... how can i use it effectively? (not using tags right now)
22:42 osse BPL: in that case, git describe --always
22:42 Rodya_ joined #git
22:42 BPL osse: ah, that's good one, git describe --all was giving head/master ;)
22:42 BPL osse: thx
22:43 thiago masak: the only way to be sure if the merge would have conflicts is to try it.
22:43 mindriot101 joined #git
22:43 thiago unless you can show it's a fast-forward. Those never fail.
22:47 medeirosthiiago joined #git
22:49 pothepanda joined #git
22:50 troulouliou_dev joined #git
22:51 medeirosthiiago joined #git
22:52 aw1 joined #git
22:54 nicksloan joined #git
22:54 buffal0 joined #git
22:55 masak thiago: you're saying merges fundamentally cannot be dry-run?
22:55 masak thiago: in the Github PR interface, there's somehow information about whether the merge would result in conflicts or not
22:56 ReptilianBrain joined #git
22:56 masak your statement is either wrong, or way above my head
22:57 masak what I'm most interested in is asking the question without disturbing the index or the working copy. git-merge-tree seems to do that.
22:58 CrazyBonz_ left #git
22:58 bn0t_ joined #git
22:58 ygeLN joined #git
22:59 bmahe joined #git
23:00 venmx joined #git
23:00 realz joined #git
23:00 ReptilianBrain joined #git
23:01 fmeerkoetter joined #git
23:02 Gsham joined #git
23:03 thinkpad joined #git
23:04 mizu_no_oto joined #git
23:05 svm_invictvs joined #git
23:07 chachasmooth joined #git
23:08 svm_invictvs joined #git
23:09 vuoto joined #git
23:16 cbegin joined #git
23:17 buffal0 joined #git
23:19 medeirosthiiago joined #git
23:20 thiago masak: note that the man page for git-merge-tree says it's not complete
23:20 thiago masak: I am right: the only command that tells whether merge will be successful is a successful merge
23:21 thiago masak: Github doesn't use git. Even if it did, how it could simply try the merge in another clone/working dir.
23:21 ResidentBiscuit joined #git
23:21 thiago masak: I recommend you try that: create a new checkout elsewhere (git new-workdir is useful for that) and then try to merge
23:23 venmx joined #git
23:23 pothepanda joined #git
23:25 Learath2 joined #git
23:25 mda joined #git
23:26 masak thiago: I hear what you're saying. not able to grok it all yet -- specifically, where git-merge-tree falls short and whether it'd be good enough for my purposes.
23:26 masak seems git-new-workdir has been superseded by git-worktree as of Git 2.5
23:27 masak (I am, however, running 1.9.1 for no good reason whatsoever)
23:27 thiago read the second paragraph of the man page
23:27 thiago and note how it takes only two arguments. Git merge can take more than one commit.
23:27 _Kyto joined #git
23:29 thiago I mean, it takes three, but it only merges two branches.
23:29 thiago git merge also has the -s and -X arguments, which git merge-tree doesn't.
23:30 jaguarmagenta joined #git
23:35 masak am I right in assuming that git-merge-tree is closer to the 'resolve' strategy than to 'recursive'?
23:36 kaldoran joined #git
23:38 moei joined #git
23:38 nickabbey joined #git
23:44 inflames joined #git
23:44 mizu_no_oto joined #git
23:47 chachasmooth joined #git
23:47 sssilver joined #git
23:50 re1_ joined #git
23:53 lindenle joined #git
23:54 Gabemo joined #git
23:59 pothepanda joined #git
23:59 mindriot101 joined #git

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