Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2017-02-23

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

All times shown according to UTC.

Time Nick Message
00:00 cqi joined #git
00:02 cqi joined #git
00:03 humboldt joined #git
00:03 peacememories joined #git
00:04 gugah joined #git
00:04 ash_workz joined #git
00:07 humboldt joined #git
00:07 elsevero is there a solution for this: https://gist.github.com/alexszilagy​i/9362ab6bcc54d536ec8afcd823353848 ?
00:07 whaley joined #git
00:08 gigq joined #git
00:09 elsevero StrongBit: you connect it via its IP (if its local) or hostname which you’ve choosed when you have installed it
00:10 elsevero StrongBit: on what Operating System have you’ve installed it ?
00:10 ESphynx left #git
00:11 GT4066 joined #git
00:11 m0viefreak joined #git
00:13 ShapeShifter499 joined #git
00:13 isbotnetcool joined #git
00:16 NwS joined #git
00:17 tang^_ joined #git
00:18 humboldt joined #git
00:18 Gsham joined #git
00:18 suy joined #git
00:19 KnightsOfNi joined #git
00:20 bvcosta joined #git
00:20 cjbrambo joined #git
00:21 Rodya_ joined #git
00:22 pur3eval joined #git
00:23 LiENUS left #git
00:24 matoro joined #git
00:26 shinnya joined #git
00:29 ResidentBiscuit joined #git
00:30 oaao joined #git
00:30 retroj joined #git
00:31 humboldt joined #git
00:32 lurkashf1ake joined #git
00:33 lurkashf1ake does having a git make sens for two person configuring and installing package in a server
00:34 rudi_s lurkashf1ake: I use git for /etc on my servers with etckeeper. Should work for two persons too ;-)
00:35 lurkashf1ake I am noob to git and realy want to start having different version and getting stuff portable and well commented between user in a readable workflow
00:35 rudi_s And a few other repos to document stuff when there are multiple admins so everybody knows how it works.
00:35 lurkashf1ake different version of files we edit
00:36 lurkashf1ake oh
00:36 rudi_s lurkashf1ake: The git used with etckeeper is only stored on a single host and normally not shared between multiple hosts.
00:37 lurkashf1ake so if I make a git of /etc/httpd, where will it save the other files. Do I choose?
00:37 rudi_s If you want to do that, I'd recommend using a configuration managment software like ansible, salt, etc. to track the config of the servers and deploy it. YOu can then store the config in git.
00:37 rudi_s lurkashf1ake: Git stores its data in .git in the root of the repository. You can synchronize that over multiple hosts by using remotes.
00:38 rudi_s !progit
00:38 gitinfo [!book] There are several good books available about git; 'Pro Git' is probably the best: http://git-scm.com/book but also look at !bottomup !cs !gcs !designers !gitt !vcbe and !parable
00:38 diego1 joined #git
00:39 ResidentBiscuit joined #git
00:39 xissburg joined #git
00:42 joeco joined #git
00:44 marenz joined #git
00:45 Levex joined #git
00:47 MattMaker joined #git
00:47 pks joined #git
00:47 Rodya_ joined #git
00:49 venmx joined #git
00:51 thebope joined #git
00:53 mischat joined #git
00:55 humboldt joined #git
00:57 nickabbey joined #git
00:58 rivarun joined #git
00:59 chipotle joined #git
01:01 rebrec joined #git
01:03 Levex joined #git
01:05 pur3eval joined #git
01:06 t-mart_ joined #git
01:06 nowhereman joined #git
01:08 xaviergmail_ joined #git
01:13 thebope joined #git
01:13 jon-mac joined #git
01:15 cjbrambo joined #git
01:27 Derperperd joined #git
01:30 Cthalupa joined #git
01:32 matoro joined #git
01:32 SwiftMatt joined #git
01:38 chardan joined #git
01:44 chardan joined #git
01:48 MattMaker joined #git
01:49 tjbp joined #git
01:53 moei joined #git
01:55 DARSCODE joined #git
01:56 chachasmooth joined #git
01:57 a_thakur joined #git
01:59 prsn left #git
01:59 _UniFreak joined #git
01:59 svm_invictvs joined #git
02:04 Sasazuka joined #git
02:04 zzz joined #git
02:04 raijin joined #git
02:06 a_thakur joined #git
02:09 Cthalupa joined #git
02:09 SwiftMatt joined #git
02:09 peacememories joined #git
02:14 MarioBranco joined #git
02:14 dreiss joined #git
02:15 Raed joined #git
02:15 SheldonCooper ok, I have a problem
02:16 glspi joined #git
02:16 SheldonCooper I have a small git submodule specifically for some test data used in unit testing
02:16 SheldonCooper running locally is not a problem
02:17 SheldonCooper but some custom install/check system we use copies everything to a /tmp directory
02:17 johnny56_ joined #git
02:17 SheldonCooper there, there is no git config --global user.name user.email
02:18 SheldonCooper is there a way to tell it to use a local git config in the submodule?
02:18 SheldonCooper which is only used for the test suite
02:19 SheldonCooper so that it doesn't matter if I run them from my development checkout or they are run from some other system/installer
02:20 Vampire0_ joined #git
02:23 d0nn1e joined #git
02:23 cdg joined #git
02:24 rrios joined #git
02:24 rrios joined #git
02:29 cagedwisdom joined #git
02:29 dviola joined #git
02:31 SheldonCooper hmm GIT_CONFIG environment variable may be a viable option
02:33 Anja joined #git
02:33 jstimm joined #git
02:33 a_thakur joined #git
02:34 gelei007 joined #git
02:35 madewokherd joined #git
02:38 matoro joined #git
02:39 gelei007 joined #git
02:40 Sasazuka joined #git
02:41 ResidentBiscuit joined #git
02:46 Wulf4 joined #git
02:46 d^sh joined #git
02:46 amdi_ joined #git
02:46 LiftLeft joined #git
02:48 MattMaker joined #git
02:49 xissburg joined #git
02:49 DarkPsydeLord joined #git
02:49 wcpan joined #git
02:52 clemf joined #git
02:53 Rodya_ joined #git
02:56 mischat joined #git
02:56 JanC_ joined #git
02:57 humboldt joined #git
02:58 mehola joined #git
02:58 raijin joined #git
02:58 pks joined #git
03:02 lanking joined #git
03:02 pur3eval joined #git
03:02 humboldt joined #git
03:03 jameser joined #git
03:03 ojdo joined #git
03:07 justan0theruser joined #git
03:08 humboldt joined #git
03:08 fstd_ joined #git
03:08 ffilozov joined #git
03:09 Rodya_ joined #git
03:10 rivarun joined #git
03:12 cdg joined #git
03:12 gopar joined #git
03:14 hexagoxel joined #git
03:16 t-mart joined #git
03:17 aavrug joined #git
03:17 Vortex34 joined #git
03:18 aavrug joined #git
03:19 humboldt joined #git
03:19 cdg joined #git
03:20 lowercaseman joined #git
03:24 msisay joined #git
03:24 microbuild3 joined #git
03:24 cqi joined #git
03:25 matoro joined #git
03:27 xall joined #git
03:29 duderonomy joined #git
03:29 jameser joined #git
03:30 cdg joined #git
03:41 MarioBranco joined #git
03:42 Ryanar joined #git
03:44 chachasmooth joined #git
03:44 lb1c joined #git
03:46 xall_ joined #git
03:47 Ryanar joined #git
03:47 humboldt joined #git
03:49 MattMaker joined #git
03:51 brent__ joined #git
03:51 sunri5e joined #git
03:52 Rodya_ joined #git
03:54 exitcode1 joined #git
03:57 xall joined #git
03:57 mischat joined #git
03:59 mehola joined #git
04:00 humboldt joined #git
04:00 SwiftMatt joined #git
04:02 pur3eval joined #git
04:02 roelmonnens joined #git
04:03 Goplat joined #git
04:04 clemf_ joined #git
04:06 hahuang65 joined #git
04:11 Anja joined #git
04:13 brent__ joined #git
04:15 thebope joined #git
04:17 dreiss joined #git
04:21 hexagoxel joined #git
04:21 Atm0spher1c joined #git
04:24 chipotle joined #git
04:24 a_thakur joined #git
04:33 rchavik joined #git
04:38 SwiftMatt joined #git
04:41 govg joined #git
04:43 ThomasTang joined #git
04:44 glspi left #git
04:45 rj1 joined #git
04:46 ResidentBiscuit joined #git
04:46 a_thakur joined #git
04:47 svm_invictvs joined #git
04:47 jstimm joined #git
04:50 MattMaker joined #git
04:55 Cabanoss- joined #git
04:55 acln joined #git
04:57 jeffreylevesque joined #git
04:58 ayogi joined #git
04:58 mischat joined #git
04:58 ozmage joined #git
05:00 menip joined #git
05:00 dermoth joined #git
05:03 pur3eval joined #git
05:05 ac joined #git
05:10 chipotle joined #git
05:10 thebope joined #git
05:14 raijin joined #git
05:15 ozmage joined #git
05:16 cdg_ joined #git
05:17 a_thakur joined #git
05:20 xissburg joined #git
05:27 Xenophon1 joined #git
05:27 durham joined #git
05:28 jimi_ joined #git
05:28 ozmage joined #git
05:29 safe joined #git
05:32 ash_workz joined #git
05:34 elementalest joined #git
05:35 hhee joined #git
05:36 Rodya_ joined #git
05:38 a_thakur joined #git
05:39 noteness joined #git
05:39 bocaneri joined #git
05:40 xaviergmail_ joined #git
05:40 elsevero joined #git
05:41 MarioBranco joined #git
05:45 ShapeShifter499 joined #git
05:47 kexmex joined #git
05:48 quite joined #git
05:48 quite joined #git
05:48 ckruczek joined #git
05:48 Tatou joined #git
05:49 Tatou Are there any issues with using the git-svn bridge?
05:50 ckruczek jop, svn
05:50 Tatou What.
05:51 MattMaker joined #git
05:53 pur3eval joined #git
05:53 _ikke_ Tatou: #git != #github
05:53 _ikke_ Or do you mean just using plain git-svn?
05:54 xall joined #git
05:55 netj joined #git
05:57 peterbecich joined #git
05:59 mischat joined #git
06:00 DaveTaboola joined #git
06:00 mehola joined #git
06:02 MarioBranco joined #git
06:02 gelei007 joined #git
06:04 thebope joined #git
06:04 navidr joined #git
06:05 sbulage joined #git
06:06 dminuoso joined #git
06:09 peterbecich joined #git
06:12 cyphase joined #git
06:13 rj1 joined #git
06:13 dec0n joined #git
06:14 Sample joined #git
06:14 bill99 joined #git
06:15 overlord_tm joined #git
06:16 xissburg joined #git
06:20 murph joined #git
06:21 venmx joined #git
06:32 xall joined #git
06:32 qt-x joined #git
06:37 MrC joined #git
06:37 expl015s_ joined #git
06:41 ahr3n joined #git
06:43 zeroed joined #git
06:47 ResidentBiscuit joined #git
06:47 circ-user-3LZMN joined #git
06:48 mehola joined #git
06:49 fletom joined #git
06:50 Raging_Hog joined #git
06:52 MattMaker joined #git
06:53 zefferno joined #git
06:53 a_thakur joined #git
06:56 a_thakur joined #git
06:57 zeroed joined #git
06:57 Darren_ joined #git
06:58 thebope joined #git
06:59 mischat joined #git
07:00 xissburg joined #git
07:03 Junior joined #git
07:07 a_thakur joined #git
07:07 rlb joined #git
07:08 pR0Ps joined #git
07:11 rlb Is there some recommended way to have all of my commit invocations verify "diff --check" by default before comitting (in addition to anything else a given repository might do in its own pre-commit hook)?
07:11 gelei007 joined #git
07:12 _ikke_ rlb: You'd add that to the pre-commit hook too
07:12 rlb Thanks, though I don't want to need to modify the pre-commit hook (if any) for every repo I work with.
07:13 rlb Worst case, I can just create my own commit subcommand, but I thought I'd see if there was something preferable.
07:13 rlb subcommand(wrapper)
07:14 _ikke_ rlb: You can have a pre-commit hook wrapper that calls multiple other checks
07:14 rlb You mean a "global" one that runs my check and then looks for the per-repo hook (if any) and runs that?
07:15 rlb I suppose as long as I can find the "current repo"'s hook dir from within my global hook, that might work.
07:16 MattMaker joined #git
07:17 scarabx joined #git
07:17 scarabx left #git
07:17 rlb Suppose I'll also have to see if hooksPath is all-or-nothing, or if it shadows.  Ideally, I don't want to have to keep a complete set (as the set evolves) of mostly pass-through hooks, just to wrap pre-commit.
07:19 rlb Hmm, the docs seem to imply all-or-nothing.
07:20 tripton joined #git
07:21 roelmonnens joined #git
07:21 thiago joined #git
07:24 _UniFreak_ joined #git
07:24 JeroenT joined #git
07:26 JeroenT joined #git
07:27 a_thakur joined #git
07:27 Anja joined #git
07:28 redhedded1 joined #git
07:28 roelmonnens joined #git
07:28 Anja joined #git
07:29 Vampire0 rlb, you are aware that a repo you clone never had any hooks, until you add them manually, aren't you. hooks are not part of the clone process
07:30 pur3eval joined #git
07:35 oaao joined #git
07:38 fewspider joined #git
07:39 a_thakur joined #git
07:42 King_Hual joined #git
07:45 jnavila joined #git
07:45 sarri joined #git
07:45 sarri joined #git
07:48 Andrew_K joined #git
07:48 troulouliou_div2 joined #git
07:52 a_thakur joined #git
07:52 thebope joined #git
07:52 mrkake joined #git
07:52 bernardio joined #git
07:54 overlord_tm joined #git
07:54 bill99 joined #git
07:55 s1scha joined #git
07:56 gelei007 joined #git
07:56 miczac joined #git
07:56 ozmage joined #git
07:59 jagob joined #git
08:00 mischat joined #git
08:04 zeroed joined #git
08:07 nothingnew joined #git
08:07 peterbecich joined #git
08:08 roelmonn_ joined #git
08:09 miczac joined #git
08:10 mischat joined #git
08:12 a_thakur joined #git
08:13 nothingnew left #git
08:13 a_thakur joined #git
08:14 zulutango joined #git
08:16 TomyWork joined #git
08:16 Tobbi joined #git
08:18 Alienpruts joined #git
08:18 Anja joined #git
08:18 elect joined #git
08:19 zeroed joined #git
08:19 zeroed joined #git
08:20 User458764 joined #git
08:20 rlb joined #git
08:20 xaviergmail_ joined #git
08:22 venmx joined #git
08:22 miczac joined #git
08:22 vuoto joined #git
08:25 dedicated_ joined #git
08:26 theoceaniscool joined #git
08:30 donnib joined #git
08:30 elsevero joined #git
08:30 donnib do i understand correct that in a large repository not only will cloning be slow checkout will also since the pack files are large to unpack ?
08:34 a_thakur joined #git
08:36 dminuoso donnib: I have not experienced slow checkouts in large repositories before
08:38 chachasmooth joined #git
08:39 a_thakur joined #git
08:39 chachasmooth_ joined #git
08:41 a_thakur joined #git
08:41 dvaske joined #git
08:42 chachasm- joined #git
08:42 dvaske_ joined #git
08:43 Anja joined #git
08:44 sasamil joined #git
08:45 Balliad joined #git
08:45 kexmex joined #git
08:46 thebope joined #git
08:47 mikecmpbll joined #git
08:48 chachasmooth joined #git
08:48 donnib dminuoso: so it's only the Clone who is impacted by big repo ?
08:49 kurkale6ka joined #git
08:49 _ikke_ checkouts can take a long time too
08:49 _ikke_ but not necessarily due to the packfile format
08:49 Tobbi joined #git
08:50 _ikke_ Usually the slowness is in writing the individual files to disk
08:54 tvw joined #git
08:54 ozmage joined #git
08:56 a_thakur joined #git
08:57 MattMaker joined #git
08:57 chachasmooth joined #git
08:58 hhee joined #git
09:01 p4trix joined #git
09:02 bongjovi joined #git
09:03 dersand joined #git
09:04 ResidentBiscuit joined #git
09:04 fortin joined #git
09:05 govg joined #git
09:07 fortin hello folks. I'm trying to compile 2.11.1 statically linked using this parameters: CFLAGS="${CFLAGS} -static-libgcc" NO_OPENSSL=1 NO_CURL=1. Build succeeds but if I inspect the binary with "file" tool it says it's dynamically linked: "dynamically linked (uses shared libs)". Following suggestions from http://stackoverflow.com/questions/11570​188/how-to-build-git-with-static-linking
09:08 bernardio joined #git
09:10 dminuoso _ikke_: but usually a checkout would only be slow if there's huge differences right?
09:12 diogenese joined #git
09:14 brent__ joined #git
09:17 kexmex joined #git
09:17 gtristan joined #git
09:18 marcogmonteiro joined #git
09:18 gtristan How can I get the diff of a specific file between 2 revisions ?
09:18 gtristan git diff sha1...sha2 file.txt
09:18 gtristan seems not to be correct
09:18 gtristan nor is git diff file.txt sha1...sha2
09:19 grayjoc joined #git
09:19 _ikke_ gtristan: without the dots
09:19 _ikke_ git diff sha1 sha2 file
09:20 clmsy joined #git
09:20 afuentes joined #git
09:20 redeemed joined #git
09:21 a_thakur joined #git
09:21 gtristan _ikke_, Ah thanks !
09:22 gtristan Oh
09:22 gtristan fatal: ambiguous argument 'file.txt': unknown revision or path not in the working tree.
09:22 gtristan _ikke_, can I do it in a bare repo ?
09:22 MasterNayru joined #git
09:22 zeroed joined #git
09:22 _ikke_ should be possible
09:22 miczac joined #git
09:22 _ikke_ try adding --
09:22 gtristan ah
09:22 _ikke_ git diff sha1 sha2 -- file
09:22 gtristan Just did that :)
09:22 gtristan Indeed
09:23 gtristan that worked
09:23 _ikke_ alright
09:23 gtristan thanks again :)
09:24 snoo joined #git
09:24 Tobbi joined #git
09:24 snoo Hi, what is the config which determines that "git checkout foo" will automatically create a local branch "foo" tracking "origin/foo" if it doesn't already exist?
09:24 _ikke_ snoo: not sure if it's a config option
09:24 snoo Seems my Jenkins server isn't having this behaviour, and not sure how to enable it, since it's default in git on my desktop.
09:24 canton7 snoo, no config: it's a hard-coded bit of behaviour
09:25 _ikke_ snoo: If you have multiple remotes with the same branches, it won't work
09:25 snoo well on this one machine "git checkout master" gives an error " pathspec 'master' did not match any file(s) known to git." even though remotes/origin/master is fetched
09:25 snoo ah
09:25 _ikke_ snoo: i would not rely on this feature anyway in this case
09:26 _ikke_ snoo: it's better to be explicit
09:26 a_thakur joined #git
09:26 snoo I don't think it's my code, it's Jenkins behaviour
09:26 canton7 from the man page, "
09:26 canton7 If <branch> is not found but there does exist a tracking branch in exactly one remote (call it <remote>) with a matching name, treat as equivalent to $ git checkout -b <branch> --track <remote>/<branch>"
09:26 _ikke_ snoo: Did you change the default branch on the remote?
09:26 _ikke_ snoo: I think jenkins assumes master exists somehow
09:26 canton7 but yeah, the only reason that bit of behaviour exists is to be a bit nicer to new users, I think. 'git checkout -b foo origin/foo' is going to be better
09:26 snoo no, the issue is that jenkins is fetching multiple remotes which have "master"
09:27 snoo I will hack it to explicitly check out remote_name/branch_name
09:27 _ikke_ snoo: why doesn't master exist in the first place?
09:27 aard_ joined #git
09:28 snoo not sure. jenkins sets up a repo and then our job adds the remote
09:28 snoo our job doesn't do git init, so it must be baked into jenkins
09:29 a_thakur joined #git
09:32 _ikke_ snoo: I would expect jenkins to do a clone
09:32 _ikke_ which would also create the default branch (which by default is master)
09:32 snoo I'll look into it, I understand the behaviour now anyway so thanks.
09:33 MrWoohoo joined #git
09:34 raijin joined #git
09:37 xall_ joined #git
09:38 acidjnk22 joined #git
09:40 mischat joined #git
09:40 thebope joined #git
09:43 jimi_ joined #git
09:44 Tobbi joined #git
09:46 codebam how can I configure a webhook to auto update my local repo every time my friend pushes by adding a git hook on his repo?
09:47 seni1 joined #git
09:47 codebam ie he pushes, sends a put to me, code is updated, server is restarted
09:48 cqi joined #git
09:48 chll_ joined #git
09:49 Blkt joined #git
09:56 rafalcpp joined #git
09:58 Levex joined #git
09:58 MattMaker joined #git
10:02 irqq joined #git
10:02 Tobbi joined #git
10:03 babilen joined #git
10:05 khfeng joined #git
10:05 Levex joined #git
10:05 metalraiden34 joined #git
10:06 Dumblez0r joined #git
10:09 nothingnew joined #git
10:11 Guest55219 left #git
10:11 valkyr2e joined #git
10:12 Tobbi joined #git
10:13 acetakwas joined #git
10:16 mrtn joined #git
10:17 mrtn left #git
10:18 ferr1 joined #git
10:20 mrkake joined #git
10:20 joeco_ joined #git
10:22 venmx joined #git
10:24 grayjoc joined #git
10:25 oaao joined #git
10:25 chachasmooth joined #git
10:26 marcogmonteiro joined #git
10:27 StrongBit joined #git
10:28 nettoweb joined #git
10:29 a_thakur joined #git
10:30 miczac joined #git
10:31 RxMcDonald joined #git
10:31 intellix joined #git
10:31 Snugglebash joined #git
10:32 Snugglebash joined #git
10:32 Tobbi joined #git
10:32 User458764 joined #git
10:33 nd joined #git
10:34 thebope joined #git
10:38 pks joined #git
10:42 djb-irc joined #git
10:45 metachr0n joined #git
10:46 oskarkv joined #git
10:47 bvcosta joined #git
10:48 nettoweb1 joined #git
10:48 ahr3n joined #git
10:50 rgb-one joined #git
10:50 rgb-one Hey
10:52 r__rios joined #git
10:52 qt-x joined #git
10:54 r_rios joined #git
10:54 r_rios joined #git
10:54 davimore joined #git
10:55 gitinfo rgb-one: [!welcome] Welcome to #git, a place full of helpful gits. If you have a question,  feel free to just go ahead and ask—somebody should answer shortly.  For more info on this channel, see http://jk.gs/git/  Take backups (type !backup to learn how) before taking advice.
10:55 rgb-one Whenever I change a branch one specific file has a CRLF instead of LF and it shows up as modified
10:55 rgb-one why is this?
10:56 gtristan So I have an automated task which will be checking out repos, and one of the steps is to make a diff between two commits for a given file (as I was asking about above)
10:56 bernardio joined #git
10:56 gtristan My original thinking, was; I will do nothing if there is no diff output
10:56 gtristan But, now I want to parse the diff output because, it's going to be common that a one line change occurs which I want to ignore
10:57 gtristan + Version: abc
10:57 joeco__ joined #git
10:57 gtristan - Version: def
10:57 gtristan Like that
10:57 leeN joined #git
10:57 gtristan So I can get to parsing, but; is there a way I can tell `git diff` to give me more brief output ?
10:57 gtristan So that I only have the 2 lines, if possible
10:58 gtristan or at least not the `diff --git bla bla` line and not those ---a/path ---b/path lines
10:58 gtristan What would be the most practical git diff invocation for this ?
10:59 MattMaker joined #git
11:01 c0c0 joined #git
11:02 tcorneli joined #git
11:04 rnsanchez joined #git
11:05 StrongBit Hello. I am using GOGs. how can I backup my projects?
11:05 selckin good old games?
11:06 cdg joined #git
11:07 Gustavo6046 joined #git
11:07 gtristan fwiw, git diff --unified=0 gets me much closer to something useful :)
11:07 gtristan think that's good enough for me
11:10 theoceaniscool joined #git
11:11 kimorigadommoy joined #git
11:13 XenophonF joined #git
11:14 ThomasLocke joined #git
11:17 tristanp joined #git
11:18 marcogmonteiro joined #git
11:19 kimorigadommoy Hey guys! Me and my colleagues are having an argument about wether ‘cache’ in git is used as ‘cache’ (storing data, for easier access) or not. Can please someone give us a clean answer like: Yes or No! :) Thanks in advance!
11:20 osse cache? what cache?
11:20 selckin as in rm --cached ?
11:20 kimorigadommoy yes
11:20 ResidentBiscuit joined #git
11:20 osse the index used to be called the cache
11:21 selckin so not really a 'cache'
11:21 kimorigadommoy Thas is "No" right ?
11:21 selckin yes
11:22 flaviodesousa joined #git
11:22 rudi_s And now it's called staging area.
11:23 enckse joined #git
11:24 storrgie joined #git
11:28 thebope joined #git
11:28 marenz joined #git
11:29 a_thakur joined #git
11:30 okjdpi joined #git
11:30 _ikke_ Not really
11:30 _ikke_ It's still called the index
11:30 kimorigadommoy why index is used?
11:30 selckin because it was better then cache
11:31 kimorigadommoy what is index in this context ?
11:31 selckin the staging area
11:31 _ikke_ Index of file blobs
11:31 kimorigadommoy no, why is it used*
11:31 kimorigadommoy the question is... for easier access, or just to have an understanding of changes?
11:32 osse kimorigadommoy: man gitglossary
11:32 gitinfo kimorigadommoy: the gitglossary manpage is available at http://jk.gs/gitglossary.html
11:32 _ikke_ kimorigadommoy: git ls-files --debug gives you a little bit of insight of what it contains (but it's not complete)\
11:32 selckin you stage the changes you want to commit (which files, or which hunks from files), then you can review then and commit
11:32 _ikke_ But the index contains a complete snapshot
11:33 _ikke_ "A collection of files with stat information, whose contents are stored as objects. The index is a stored version of your working tree."
11:33 _ikke_ Note that .git/index does not contain the contents itself, just a reference to the blobs
11:33 acetakwas joined #git
11:34 User458764 joined #git
11:34 sunri5e joined #git
11:35 kimorigadommoy if git was a daemon :D it would save those reference in memory? :D
11:35 dcpc007 joined #git
11:36 _ikke_ kimorigadommoy: Some things, but it would still need to store things on disk
11:36 _ikke_ You don't want your staged changes to be gone when you restart the daemon
11:36 DBag joined #git
11:37 DBag left #git
11:37 kimorigadommoy makes sense, but the main question still exists, is it for speed purposes, or it is a logic thing
11:37 _ikke_ combination
11:37 _ikke_ the stat info is for speed
11:38 _ikke_ The reference snapshot of the working tree is for logical reasons (though, it's also optimized to easily create tree objects from it)
11:38 dgonzo joined #git
11:40 _ikke_ If you delete the index, you lose state (the staged changes)
11:40 _ikke_ without the index, certain operations would take much longer (git status for example)
11:41 Torrone joined #git
11:41 bvcosta joined #git
11:43 thadtheman joined #git
11:43 kimorigadommoy _ikke_  thank you for your response!
11:43 kimorigadommoy it was what needed
11:43 thadtheman Can you clone multiple repos in one command? git clone repo1 repo2 repo3 ?
11:44 _ikke_ no
11:44 whaley joined #git
11:44 thadtheman Thank you.
11:44 bremner !myrepos
11:45 dedicated joined #git
11:45 nettoweb joined #git
11:46 Anja joined #git
11:48 Torrone Hello, this is git-related, but not strictly a git question: I work in a company, I am one of the employees who work from home, I'm on the main project and I'm practically doing it all by myself, I can handle doing everything but sometimes I really would like to have someone else working with me, at least to discuss decisions. There's this guy who works at the office, he's nice and he's supposed to be good at javascript stuff, also sup
11:48 Torrone posed to work on the project at some point. When I told him he just needed to git clone he said he couldn't use git, that's fine to me but it makes me doubt his competence, and makes me worry about having to do everything on my own without having anyone else to know the project ( this one involves payments, it's pretty important..) how should I deal with this?
11:48 cdg joined #git
11:49 donnib Torrone: why can he not use Git ?
11:49 nettoweb1 joined #git
11:50 donnib Torrone: and what do you need him for ? Review ? Send the code thru other channels if he is not confident to use git
11:50 Torrone I think he's not accustomed to work at the same standards I'm used to
11:51 donnib Torrone: unless he starts to code there might be no reason to push him to use your tool just because you need his help
11:51 Torrone since I've been hired I've struggled to have competent confrontations
11:51 donnib Torrone: well you are asking for help so he must be skilled in some way or ?
11:51 Torrone some guys I know for sure that are incompetent, I'm not sure about this one
11:52 Torrone I'm worried to do all the work by myself, I am up to higher standards but it's the first time I have to deal with sensitive stuff like payments
11:53 donnib Torrone: take it up with the manager or the team and agree on a process of work like integrating Git and integrating review as part of the process to ensure that others also knows about others code
11:54 Torrone donnib, I just introduced the joel test t my manager
11:54 donnib Torrone: now, i don't know the country you are in and what culture you are up agains but in a west european culture i would take this issue (formulated most likely different) with the team like guys i am new and i works some things which i am not comfortable with, how can i involve you into this
11:56 Torrone I'm from western europe
11:56 donnib Torrone: it's really is a bad practice in a team to sit with something on your own without sharing it with other especially if it's areas that are new to you,
11:57 roelmonnens joined #git
11:57 donnib Torrone: good, talk to the team. you might have a meeting (retrospective) where you can discuss freely and tell the other your opinion in a constructive way without pointing fingers
11:58 elsevero_ joined #git
11:58 Tobbi joined #git
11:58 osse FWIW Git is one of those tools it makes sense to use on a team level. Either everyone is in on it, or you use something else.
11:58 donnib Torrone: and if you have a competent manager this will be de dealt with, you can suggest to the team to go to use Source Control (i pressume not all do) and choose Git
11:59 donnib Torrone: then help team adapt Pull Request aka Reviews, let the team know you want this for the best interest for the company/product and help them
11:59 UniFreak joined #git
11:59 MattMaker joined #git
12:00 donnib Torrone: i meet so many teams not knowing they could do better both process wise but also tools/technology because they don't know about them
12:00 selckin sounds like its a small company tho
12:00 donnib selckin: yeah i have the same feeling
12:00 Torrone donnib, yeah, I'm already doing this with the client's IT guy, but I can't  always involve him for obvious reasons
12:00 selckin those things don't really work in small companies
12:00 Levex joined #git
12:00 donnib selckin: what doesn't work in samll companies ?
12:01 bremner having managers?
12:01 osse private armies
12:01 bremner outsource!
12:01 selckin teams and talking to team lead, and setting up reviews
12:02 metalraiden34 joined #git
12:02 osse The nice thing about private armies is that they can't hang you for treason if you no longer wish to work for them.
12:02 donnib selckin: i see no reason that it wouldn't work
12:02 selckin they don't exist
12:02 donnib selckin: what doesn't exist ? be precise
12:03 dviola joined #git
12:03 oaao joined #git
12:03 roelmonnens joined #git
12:04 Torrone donnib, the process is fairly organized, I already use source control, pull requests ( since I'm working on other people's code ) and I think I have a pretty decent workflow, my problem is I would sometimes need another competent pair of eyes and I'm not sure I can find them in my company, because I've seen other works from my colleagues and they damn sure are not up to the standards and second because I'm the only one in the company w
12:04 Torrone ho knows the server side language of this particular project
12:07 selckin having expertise in a differend area as you does not make them incompetent
12:07 selckin be careful before everyone hates you
12:07 oaao joined #git
12:08 joeco_ joined #git
12:08 donnib I agree with selckin
12:09 Torrone selckin, I'm not saying they're incompetent ( well, I've been handed code from one guy and I can tell that he's incompetent at writing css, but appartently it's the only thing he does). I barely know them and it's difficult for me to estimate their competence
12:11 a_thakur joined #git
12:12 YuGiOhJCJ joined #git
12:12 Torrone all I know is that they're nice people and they're all seniors
12:12 xissburg joined #git
12:12 Torrone anyway I'll try and talk to my manager, thanks for the advice
12:13 Torrone donnib,
12:13 donnib Torrone: sure no problem if not revert to my first comment, use them where they have knowledge experience and share your code/whatever you want input on with them in the way THEY feel comfortable
12:14 marcogmonteiro joined #git
12:15 cyberz joined #git
12:16 ismithers joined #git
12:19 a_thakur joined #git
12:22 psyb0t joined #git
12:22 thebope joined #git
12:25 explody joined #git
12:25 whaley joined #git
12:25 mischat_ joined #git
12:27 al-damiri joined #git
12:30 donnib osse: i was discussing with someone else earlier on perfomance cost of big repositories and checkout operations, do you know for sure that there is no gain in performance on a checkout operation on a big vs small repo ?
12:32 gtristan joined #git
12:37 seni1 joined #git
12:37 nettoweb joined #git
12:38 mahakal joined #git
12:38 Masber joined #git
12:42 levex_ joined #git
12:42 zeroed joined #git
12:44 osse donnib: what do you mean by performance? MB/s of checked out data during the operation?
12:48 marcogmonteiro joined #git
12:48 jimi_ joined #git
12:49 gugah joined #git
12:49 ayogi joined #git
12:51 MrMojit0 joined #git
12:54 xaviergmail_ joined #git
12:55 Ardethian joined #git
12:56 DolphinDream joined #git
12:57 StrongBit joined #git
12:57 xissburg joined #git
13:00 Tobbi joined #git
13:00 MattMaker joined #git
13:00 donnib osse: i am talking about time to checkout
13:01 donnib osse: i mean does it take more time to checkout in a big repo vs small repo
13:01 osse i don't see how it couldn't
13:01 Seveas repo size matters a bit, but worktree size matters more
13:01 r_rios joined #git
13:01 osse ohhh
13:01 donnib osse: my theory without knowing the details in git is that the pack file is bigger so file operations are slower
13:01 osse "big repo" = more history?
13:01 donnib osse: not necessary but big files
13:02 Seveas pack files are indexed
13:02 selckin can test this quite easy with like the kernel git repo ? time a bunch of checkouts, then blow away .git, make a single commit, and time again?
13:02 Seveas so while a big packfile is slower to read, it's not much slower.
13:02 osse I don't know
13:03 Seveas "amount of files that need to be checked out" (and their size) will be the biggest cotributing factor
13:03 donnib so i am thinking that a repo that has let's say 1.5 gb and i clean it (remove some of the big files that was inside git which devs have removed but are stuck) should give me both a clone performance but also checkout but i am not sure
13:03 Seveas opening, writing and closing 100 files is a lot slower than reading a packfile index.
13:05 synthroid joined #git
13:06 lss8 joined #git
13:06 lss8 joined #git
13:06 rgb-one joined #git
13:11 peacememories joined #git
13:12 cyberz joined #git
13:13 gp5st left #git
13:13 Sample joined #git
13:13 jeffreylevesque joined #git
13:15 jameser joined #git
13:16 PrashantJ joined #git
13:16 thebope joined #git
13:17 PrashantJ left #git
13:18 Timvde I'm suddenly totally confused about rebase :( If I have a merge conflict during rebase, do I have to commit manually after fixing, or does git rebase --continue do this for me?
13:19 selckin believe continue does it, have to stage it tho
13:19 Timvde I guess I could set up a test repo and try, though... But it annoys me that I lack the intuition.
13:19 selckin you can also commit if you want
13:20 jophish_ left #git
13:20 Timvde https://stackoverflow.com/questions/6457044/forgot​-git-rebase-continue-and-did-git-commit-how-to-fix
13:20 Timvde Okay, looks like I need to just do a rebase --continue
13:21 Torrone joined #git
13:21 selckin hmm i was sure that was valid to the docs
13:21 Tobbi joined #git
13:22 Timvde selckin: 2011 question, maybe the answers are outdated
13:22 venmx joined #git
13:23 star_prone joined #git
13:24 joeco__ joined #git
13:25 selckin differend behavior in rebase and rebase -i, that what trips me on it
13:28 acetakwas joined #git
13:29 Nilesh_ joined #git
13:29 kritzikratzi joined #git
13:30 explody joined #git
13:31 Levex joined #git
13:33 rvgate joined #git
13:33 JeroenT_ joined #git
13:34 kexmex joined #git
13:34 reynierpm joined #git
13:34 Dumblez0r joined #git
13:35 MattMaker joined #git
13:35 guardian hello, when using git rev-list --before=<date> is it the author date or the commiter date?
13:35 reynierpm I have a few branches locally with some work but I don't want to push them to the origin because it's an uncompleted work and shouldn't be there, now I am moving from one laptop to a brand new and I need to move also such work, what's the best way to achieve this without push all those branches to the origin?
13:35 jameser_ joined #git
13:36 rgb-one joined #git
13:36 hussam joined #git
13:36 ResidentBiscuit joined #git
13:38 jameser joined #git
13:38 pk64 joined #git
13:39 Ryanar joined #git
13:40 pk64 How do I rebase my patch branch? When I go to rebase, I see commits from what seems to be the master branch, where my patch branch lists my 4 commits, the master has 2 from me + the upsream branch
13:41 jameser_ joined #git
13:41 marcan joined #git
13:41 xissburg joined #git
13:43 GodGinrai joined #git
13:43 mindfart joined #git
13:44 thethorongil joined #git
13:45 oaao joined #git
13:46 marcogmonteiro joined #git
13:46 PrashantJ joined #git
13:47 bvcosta joined #git
13:47 moritz pk64: generally by checkiing out the branch and doing a "git rebase master"
13:48 PrashantJ I recently made a PR on github and ran into merge conflicts. This is my very first pool request and I am newbie. Please suggest ways to resolve. Some reference link is  appreciated.
13:48 grayjoc joined #git
13:49 PrashantJ For your kind reference : https://github.com/rtnpro/opencabs/pull/8
13:49 pk64 moritz: wouldn't that affect master? I just want to squash the commits I see in the patch-4 branch
13:50 rwb joined #git
13:50 moritz pk64: no, it affects the branch you're on.
13:50 pk64 ah ok, thank you. i'll review the rebase doc again as well
13:53 jameser joined #git
13:54 dedicated_ joined #git
13:54 scoobertron joined #git
13:55 rafalcpp sha1 is broken in practice now.   https://shattered.it/
13:55 rafalcpp when will git move to sha256 or sha512 checksums?
13:56 _ikke_ rafalcpp: It's not that easy
13:57 _ikke_ rafalcpp: Work is underway to not rely on a fix hash lenght, but it's still a lot of work to actually change the hash algorithm
13:57 corentin joined #git
13:57 corentin hello
13:57 gitinfo corentin: 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.
13:58 corentin what about this? http://shattered.io/
13:58 _ikke_ corentin: you're #2
13:58 _ikke_ rafalcpp: corentin Best place is the !mailing_list
13:58 gitinfo rafalcpp: The mailing list can be reached via git@vger.kernel.org. You don't need to subscribe to the list, you will always be put in cc on reply. Read archives at http://j.mp/gitlist
13:58 corentin will git use a stronger hash algorithm?
13:58 _ikke_ corentin: Probably in the future, but it's not something you can switch over night
14:01 Raging_Hog joined #git
14:02 mischat joined #git
14:03 JeroenT joined #git
14:04 pks joined #git
14:04 aspiers joined #git
14:04 star_prone joined #git
14:07 metalraiden34 joined #git
14:09 jimi_ joined #git
14:09 jameser joined #git
14:09 kinlo joined #git
14:10 thebope joined #git
14:11 Rodya_ joined #git
14:12 jeffreylevesque joined #git
14:15 peacememories joined #git
14:16 bgerber joined #git
14:17 DarkPsydeLord joined #git
14:17 Learath2 _ikke_: seems the archive link is broken
14:19 jameser joined #git
14:19 peacememories joined #git
14:19 nioncode joined #git
14:20 _ikke_ Hmm, right
14:20 r_rios joined #git
14:20 _ikke_ http://public-inbox.org/git
14:20 noctux joined #git
14:21 aielima joined #git
14:22 star_prone joined #git
14:23 Vampire0 !list
14:23 gitinfo [!mailing_list] The mailing list can be reached via git@vger.kernel.org. You don't need to subscribe to the list, you will always be put in cc on reply. Read archives at http://public-inbox.org/git
14:23 mahakal joined #git
14:23 _ikke_ Vampire0: thanks
14:26 Darcidride joined #git
14:26 alexandre9099 joined #git
14:26 aielima joined #git
14:27 Murii_ joined #git
14:27 jsho joined #git
14:27 finalbeta joined #git
14:29 Vampire0 _ikke_, yw
14:29 Eryn_1983_FL joined #git
14:32 King_Hual joined #git
14:33 texinwien_ joined #git
14:34 cqi joined #git
14:35 Stummi joined #git
14:36 Stummi hey there, just wondering are there any plans for git to rely on other hash algorithms than SHA-1 for commit hashes? (Or maybe is it already possible at all?)
14:37 r_rios joined #git
14:37 r_rios joined #git
14:37 osse Stummi: work has been done. not sure how complete it is
14:37 osse e.g. stop hardcoding 20 everywhere
14:37 grawity see the mailing list for the heaps of object_id patches
14:39 Ryanar joined #git
14:39 _ikke_ Stummi: You're #3 :P
14:40 ShekharReddy joined #git
14:41 oaao joined #git
14:41 xissburg_ joined #git
14:44 publio joined #git
14:44 dragoonis joined #git
14:45 Es0teric joined #git
14:45 dragoonis Hey! I'm building a CI script and would like some advice .. I want to clone down the specific target branch specified by the CI job.. I can clone down the first time, but on the second run I don't want to re-clone the entire repository every time, is there a quicker way? This is my command right now
14:45 dragoonis git clone -b ${FURY_BRANCH} --depth=1 --single-branch git@repo.url/fury.git
14:46 dragoonis Let's say FURY_BRANCH changes from 'master' to 'develop'
14:46 dragoonis I need it to quickly trash all [potential] dirty changes and switch to the new branch and pull down its code
14:46 dragoonis right now I'm doing "rm -rf fury && git clone -b ..."
14:46 dragoonis but it's slow ..
14:47 dragoonis this repo's .git folder is 900MB
14:47 ResidentBiscuit joined #git
14:48 BadHorsie joined #git
14:49 PaulCapestany joined #git
14:50 linuxmodder joined #git
14:50 khfeng joined #git
14:50 theoceaniscool joined #git
14:51 yanome oh boy, https://security.googleblog.com/2017/0​2/announcing-first-sha1-collision.html
14:51 crose joined #git
14:52 mahakal joined #git
14:52 xissburg joined #git
14:52 peacememories joined #git
14:53 Vampire0 yanome, you're #4
14:53 nickabbey joined #git
14:53 bremner clearly we need a bot command
14:54 moritz !collision
14:54 Vampire0 bremner, the bot triggers cannot increase the count on each invocation *g*
14:54 yanome wait, really?
14:54 yanome sorry
14:54 Vampire0 no problem
14:54 yanome what?
14:55 yanome nobody in my scrollback has posted that link
14:55 github051 joined #git
14:55 yanome ah, but i wasn't the first to mention it
14:55 osse dragoonis: fetch
14:55 Vampire0 yanome, http://shattered.io/
14:55 moritz dragoonis: git fetch && git checkout -f origin/FURY_BRANCH
14:55 moritz dragoonis: as a good first step
14:55 osse dragoonis: the --depth thing will bite you in the end. do a regular clone first and fetch thereafter
14:55 yanome Vampire0: wneat
14:55 BadHorsie Is there a way to have git fetch use the same flag as clone --reference ? Trying to use Jenkins but it uses git fetch instead of clone so not sure how to use a local cache (mirror --clone)
14:56 dragoonis osse, what will happen, the --depth thing saves me SOOO much clone time
14:56 imack joined #git
14:56 osse dragoonis: yes, but only the once
14:56 rgb-one joined #git
14:56 osse subsequent fetches are fast
14:56 osse and depth make them slower (AFAIK)
14:57 kbs joined #git
14:57 tristanp joined #git
14:59 dragoonis osse, moritz for my initial clone can I still keep --single-branch, yea?
14:59 osse yes
14:59 dragoonis cool deal
15:00 AaronMT joined #git
15:02 texinwien_ joined #git
15:02 sbulage joined #git
15:04 r_rios joined #git
15:04 basy joined #git
15:05 thebope joined #git
15:05 ash_workz joined #git
15:05 dragoonis osse, moritz this is what i've whipped up - what do you think?
15:05 dragoonis https://gist.github.com/dragoonis​/cd1d45f21121a9f70cbe2f9273689092
15:06 _ikke_ yanome: Nice btw that you link to the blog post instead of the marketing page
15:06 mahakal joined #git
15:06 jameser joined #git
15:06 BlueMatt joined #git
15:07 osse dragoonis: you should checkout origin/blah like moritz said
15:07 dragoonis cool
15:08 shinnya joined #git
15:09 BlueMatt heh, re: collision: is there some way to coerce git into giving you a tree hash re-calculated with something other than sha1 (even if super slow)?
15:09 Gsham joined #git
15:09 kbs wondering if that would help (even if done)
15:10 jj- joined #git
15:10 lowercas_ joined #git
15:10 BlueMatt kba: not for many things, but eg signed tags you could just sign that tree hash and publish that elsewhere
15:10 ResidentBiscuit joined #git
15:11 BlueMatt would be much nicer than going and sha256ing every file in your repo
15:11 ravi__ joined #git
15:12 blackwind_123 joined #git
15:12 bannakaffalatta joined #git
15:12 kbs ah, I see
15:12 ravi_ joined #git
15:12 gugah joined #git
15:13 msonntag joined #git
15:14 Sample joined #git
15:14 Hozy joined #git
15:14 vimal2012 joined #git
15:15 gugah_ joined #git
15:15 MacGyver joined #git
15:17 Torrone joined #git
15:17 jeffreylevesque_ joined #git
15:17 LeBlaaanc joined #git
15:17 malt3 joined #git
15:18 shgysk8zer0_ joined #git
15:19 varun joined #git
15:19 aw1 joined #git
15:20 dsdeiz joined #git
15:20 dsdeiz joined #git
15:20 texinwien_ joined #git
15:21 whaley joined #git
15:21 kbs Seems reasonable :-) and thinking (naively) correct to assume something like this would need to reformat + then recalculate every object in the full history?
15:22 denisMone joined #git
15:22 kbs (i.e. the embedded 20 byte hashes in each object reference would need to be redone all the way to the starting point)
15:23 marcan I imagine forward-migration should be possible
15:23 Hozy left #git
15:23 jameser joined #git
15:24 e14 joined #git
15:24 marcan you'd have to tag and allow both kinds of hashes, then at some point force a cutoff
15:24 marcan you could do that by naming every leaf SHA1 in a SHA-256 object and forbidding any child commits not using SHA-256 after that
15:24 MattMaker joined #git
15:24 Darren_ joined #git
15:25 marcan then, as long as there are no preimage attacks in the future (unlikely), or secret chosen-prefix collision attacks in the present (probably unlikely), you're pretty much safe
15:26 marcan I guess git already implements the parent/child commit mechanics to allow this, even if the signed trees are dummies. you'd just have a special commit that has as parents all leaf SHA-1 commits and basically "caps" the SHA1 history
15:26 marcan so the main thing would be adding support for more than one hash algorithm and somehow tagging that in the various binary formats
15:27 varundath joined #git
15:27 gtristan joined #git
15:28 kbs is a "leaf" sha-1 commit the same as a blob object?
15:28 _ikke_ No
15:28 marcan I mean a commit object
15:28 marcan dealing with trees and blobs is easy
15:29 marcan you just enforce that sha-256 commits must reference sha-256 trees and blobs
15:29 _ikke_ marcan: there has been a discussion about this on the mailing list
15:29 marcan _ikke_: got a link?
15:30 Dumblez0r joined #git
15:30 _ikke_ marcan: Not sure if this contains what I mean, but this is at least a discussion about it: http://public-inbox.org/git/CAPp-Vrb_n6z39RLHZ4A​eUaBFiJfL3_xX8Utfq7+bTgzZrza58Q@mail.gmail.com/
15:30 berndj what practical steps can high-stakes projects take? identify versions by an n-tuple of the most recent commits, n > 1? (would that even help?)
15:31 marcan keep infra secure and sign release tarballs is the obvious way
15:31 marcan and don't allow force pushing
15:31 _ikke_ Note that git doesn't overwrite existing objects with the same hash
15:32 marcan right, this is mostly about fresh clones
15:32 marcan also keep in mind that with the *current* attack you're mostly safe if you use git for *code*
15:32 _ikke_ right, it's not a pre-image attack
15:32 marcan no, it's also not a chosen-prefix attack
15:32 _ikke_ so it's still hard to find a collision for an existing object
15:32 marcan it's a fixed-prefix attack
15:32 marcan so you'd have to have *malicious code* in there from the beginning
15:33 marcan if you're using git for binaries then you're at risk unless you're carefully vetting those binaries
15:33 marcan if you use git for code presumably you actually read said code (if you don't then you're already screwed) so you should be fine
15:33 marcan for *now*
15:33 LeBlaaanc joined #git
15:33 basy joined #git
15:33 marcan using the current attack with code would require malicious code to already be present in the repo, triggered conditionally on the collision block
15:34 marcan that's easy with binaries nobody goes around decompiling. it's easy with PDFs. it's rather hard with source code.
15:34 berndj marcan, how does signing help? is the signature over something other than the commit id?
15:34 marcan you'd have to sign the whole tree, not the commit object
15:34 varundath What is the difference between these two git fetch commands.?
15:34 varundath Command 1: `git fetch origin refs/pull/ID/merge:BRANCHNAME`
15:34 varundath Command 2: `git fetch origin pull/ID/head:BRANCHNAME`
15:34 marcan I mean traditional gpg signatures on a tarball
15:34 Droolio joined #git
15:35 berndj ok, but making sure that gpg doesn't just use sha1 under the hood
15:35 marcan sure
15:35 berndj but yes, perspective: it's a collision attack, not (yet) a preimage attack
15:35 moritz I think you can configure git to use a message digest algorithm of your choice
15:35 marcan again, it's not an arbitrary collision attack
15:35 marcan it's a fixed prefix collision attack
15:35 marcan the difference is *very* important
15:35 berndj although that isn't a guarantee of safety
15:35 marcan a preimage attack on SHA-1 will likely never happen
15:36 marcan we don't even have collision attacks on MD2
15:36 marcan but a *fixed prefix* collision attack on MD5 you can compute on a smartphone
15:36 Dougie187 joined #git
15:36 marcan a *chosen prefix* collision attack on MD5 is what broke x.509 certs and what the Flame malware used, and you still need a cluster to compute that
15:36 berndj an attacker could be (or could compromise) a developer with push access, then generate two colliding trees, one benevolent, one malevolent
15:36 marcan this is a *fixed prefix* attack on SHA-1
15:36 marcan berndj: no
15:37 marcan an attacker could generate two colliding trees, *both malevolent*, except one might not actually do anything evil
15:37 * moritz wouldn't sign "a preimage attack on SHA-1 will likely never happen"
15:37 berndj well both malevolent wouldn't be useful, as it'd stand out like a sore thumb
15:37 basiclaser joined #git
15:37 osse I blame capitalism
15:37 berndj unless you mean malevolent at the philosophical level
15:37 Eryn_1983_FL joined #git
15:38 marcan moritz: okay, in my lifetime, at least not until brute force compute power advances to the point where it's nearly meaningless
15:38 moritz cryptoanalysis *has* made some impressive breakthroughs before
15:38 roelmonnens joined #git
15:38 marcan berndj: I mean both malevolent
15:38 marcan today's attack lets you create two files which *only* differ in one tiny section
15:38 marcan you can *make* the rest of the file check that section and behave differently. that's how you get the "evil binary, good binary" or "evil pdf, good pdf" case - buth both are evil, one just doesn't behave in an evil way
15:39 marcan but you can't do that with code because the if(collision == 1) evil(); else not_evil(); would stick out like a sore thumb
15:39 marcan you *cannot* collide good(); //collision1 with evil(); //collision2 with today's attack
15:39 marcan (you *can* with the MD5 collision attack that Flame used)
15:39 marcan *that* is the important difference
15:40 berndj printf("Example nonce: 458cbd7fe\n"); /* Show user allowed values. */
15:40 berndj printf("Example: yes\n"); system("rm -rf /"); /* 7n29c2==28c7 */
15:40 marcan moritz: keep in mind that we don't have a preimage attack on MD2 yet
15:40 berndj ^ like that, marcan
15:40 marcan berndj: you don't get to pick the contents of the section that changes
15:41 berndj marcan, that's why it's disguised as a harmless random number
15:41 marcan where does the "system("rm -rf /");" come from?
15:41 berndj ok, i'll come up with a better example
15:42 ispn joined #git
15:42 kbs Just to offer a gentle diversion from the "is the sky really falling" discussion :-) _ikke_ or others in the know - likely to see something like the mailing list proposal emerge in the next versions or two of git?
15:43 Levex joined #git
15:43 texinwien_ joined #git
15:43 _ikke_ kbs: No, this is a long term thing
15:44 mischat_ joined #git
15:46 mischat_ joined #git
15:47 mischat__ joined #git
15:48 Kumool joined #git
15:48 kpease joined #git
15:48 shgysk8zer0_ joined #git
15:49 PCatinean joined #git
15:49 PCatinean Hello everyone
15:50 PCatinean I got myself into a very complex scenario (not sure how I landed here)
15:50 Wind0r joined #git
15:50 PCatinean I have a branch that's 20 commits behind the main one, and 48 ahead, after rebase it's 24 commits after, how can thi sbe?
15:51 Wind0r joined #git
15:51 berndj marcan, preferred pastebin? pastebin.com ok?
15:52 marcan whatever works
15:52 miczac joined #git
15:52 marcan I run my own private one :P
15:52 moritz PCatinean: does "after" mean "ahead" or "behind"?
15:53 PCatinean ahead moritz sorry
15:53 berndj marcan, here, my better attempt: http://pastebin.com/UGZMJMZn
15:53 PCatinean just trying to be sure I don't lose any work
15:53 moritz PCatinean: sounds imposible
15:53 jstimm joined #git
15:53 hezekiah joined #git
15:53 wizeman joined #git
15:53 moritz PCatinean: did you git fetch or push in the mean time?
15:54 marcan berndj: I'd argue that ought to raise eyebrows (also the junk would probably have to be binary)
15:54 moritz s/impossible/implausible/
15:54 berndj granted, this pseudo-example uses C, and not all 8-bit values are valid even inside a C comment, and the collision-grinding would need to be repeated until it happens to generate the right prefix to those ifdef / comment bits
15:54 moritz PCatinean: or did git rebase ask about empty commits several times during the rebase?
15:54 berndj and it's just a 5-minute proof of concept from a relative n00b, not a serious assault
15:55 marcan berndj: this is the kind of thing the underhanded C code contest is for
15:55 PCatinean moritz, tehere were 2 conflicts
15:56 PCatinean I solved them and then pushed to a separate branch instead of pushing with force on the original one
15:57 Salz joined #git
15:58 PCatinean I wonder what's going on
15:58 rgrinberg joined #git
15:58 PCatinean moritz, maybe it omits the merges?
15:59 thebope joined #git
15:59 robotroll joined #git
15:59 overlord_tm joined #git
15:59 moritz the rebase might have flattened out some merges
15:59 moritz and thus lost some merge commits
15:59 moritz which might account for a lower total numer of commits afterwards
16:00 PCatinean hmm
16:00 PCatinean The thing is I would want to later merge the changes from dev to master
16:00 PCatinean But right now I need some commits from master on dev
16:00 PCatinean And I think rebase is the best way to go?
16:01 moritz it's an option
16:01 moritz "best" really depends on many factors
16:01 jost__ joined #git
16:02 rgb-one joined #git
16:02 tlaxkit joined #git
16:02 NeXTSUN joined #git
16:02 DreadKnight joined #git
16:02 roelmonn_ joined #git
16:03 jagob joined #git
16:03 DreadKnight hey, when cloning a repo in a folder I want, how to stop git from cloning the whole thing into a subfolder?
16:04 moritz DreadKnight: git clone <url> destination/
16:04 moritz DreadKnight: and if into the current directory, git clone <url> .
16:05 moritz (. = current directory)
16:05 DreadKnight moritz, did just that and it made a nodejs folder with the whole thing, sigh; will try again
16:05 metalraiden34 joined #git
16:08 Atemu joined #git
16:09 clemf joined #git
16:09 kexmex joined #git
16:10 guest3456 joined #git
16:10 guest3456 is what they say about Git on this page true?
16:10 guest3456 https://shattered.it/
16:12 moritz guest3456: it's been discussed here before; please read the logs starting from https://irclog.perlgeek.de​/git/2017-02-23#i_14150968
16:12 guest3456 will do
16:13 moritz (note that there were at least two blocks of discussion about this)
16:13 GodGinrai joined #git
16:14 chardan joined #git
16:15 nickabbey joined #git
16:16 whaley_ joined #git
16:19 ayogi joined #git
16:19 hasc joined #git
16:22 basy joined #git
16:22 l4v2 joined #git
16:23 Emperor_Earth joined #git
16:23 qqx joined #git
16:23 dragoonis moritz, osse hey
16:23 dragoonis git fetch && git checkout -b origin/dockerization
16:23 dragoonis fatal: A branch named 'origin/dockerization' already exists.
16:24 dragoonis If I'm already on the target branch, it isn't pulling in latest changes for the target branch
16:24 GodGinrai joined #git
16:25 GodGinrai dragoonis: -b creates a branch.  You should not be using that flag
16:26 pie__ joined #git
16:26 CheckDavid joined #git
16:26 texinwien_ joined #git
16:27 de-vri-es joined #git
16:27 xaviergmail_ joined #git
16:27 pie__ so, given googles practical collision attack, can we get a push for deprecating sha1 in git now :I
16:27 pie__ or if you prefer,  "practical"
16:27 lordjancso joined #git
16:28 GodGinrai pie__: sha1's use in git never had anything to do with security, did it?  I thought it was just a reliable way to create unique hashes
16:28 GodGinrai for git's purpose
16:28 nowhereman joined #git
16:29 Topic for #git is now Welcome to #git, the place for git help. Just git help, apparently. | We are aware of the SHA-1 collision |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 | This is serious business. https://madeitwor.se/git-jokes
16:29 Ryanar joined #git
16:29 dragoonis GodGinrai, i have a CI script, i'm avoiding deleting and re-cloning the repo every time .. so was looking to have the system just quickly check out the target branch
16:29 Es0teric joined #git
16:29 Topic for #git is now Welcome to #git, the place for git help. Just git help, apparently. | We are aware of the SHA-1 collision | 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 | This is serious business. https://madeitwor.se/git-jokes
16:29 perlpilot heh
16:29 dragoonis GodGinrai, so I ended up with this: cd $DIR/fury && git fetch && git checkout -b origin/${FURY_BRANCH}
16:30 BSAlb joined #git
16:30 GodGinrai dragoonis: you should never use the -b flag with any remote branch name as an argument
16:30 GodGinrai dragoonis: remote branches are created by git.  You are not supposed to create them yourself
16:30 dragoonis I need something to cover this : a) you're already on the target branch but you want to pull latest changes. b) you're on a branch and want to switch to a new one and pull its latest changes
16:31 _ikke_ GodGinrai: Note that this would create a local branch called origin/..
16:31 perlpilot dragoonis: you could keep the git clone someplace separate and use rsync to make the branch CI uses match what's in the repo
16:31 GodGinrai dragoonis: like I said, get rid of the -b.  It has nothing to do with what you want to do
16:31 _ikke_ refs/heads/origin/<name> vs refs/remotes/origin/name
16:31 GodGinrai _ikke_: I know it will.  That's a problem
16:31 _ikke_ Just to be sure
16:32 dragoonis GodGinrai, so "git fetch && git checkout origin/{$TARGET_BRANCH}" will suffice my requirements that I listed ?
16:32 pie__ GodGinrai, well, yeah, i guess. but why not actually make it secure? :/
16:32 _ikke_ dragoonis: as long as you are not planning on comitting, then that works
16:32 Gsham joined #git
16:32 ayogi joined #git
16:32 mischat joined #git
16:32 GodGinrai pie__: Would you feel the need to secure a stack of free newspapers?
16:32 dragoonis _ikke_, no committing here.. pulling down the code .. running some tests .. then passing/failing the jenkins build based on the test suite result
16:33 pie__ GodGinrai, secure them from meddling sure
16:33 pie__ i.n.t.e.g.r.i.t.y.
16:33 dedicated joined #git
16:33 _ikke_ Right
16:33 GodGinrai pie__: you are telling me this attack could allow someone to ruin the integrity of the repo?  How?
16:34 _ikke_ pie__: Note that this specific attack is not very usable in git
16:34 pie__ GodGinrai, i havent done my homework, but even if this concrete attack isnt of interest....
16:34 _ikke_ pie__: It's not a pre-image attack, which lets you find a collision for existing data
16:34 pie__ though i suppose i did imply that it has immediate raifications, so i apologize
16:35 pie__ *ramifications
16:35 ToBeCloud joined #git
16:35 dragoonis GodGinrai, _ikke_ i removed the -b and I get this now: "warning: refname 'origin/dockerization' is ambiguous."
16:35 _ikke_ dragoonis: because you already created a branch called origin/dockerization
16:35 _ikke_ dragoonis: git update-ref -d refs/heads/origin/dockerization
16:36 dragoonis Is it safe to ignore this ? it will be repeatedly checking out origin/dockerization on every built
16:36 GodGinrai dragoonis: see the command _ikke_ just gave you
16:36 dragoonis yeah - msg delays
16:36 dragoonis thanks i will do this before running the checkout
16:36 _ikke_ dragoonis: You only need to do this once
16:36 dragoonis fetch && update-ref && checkout
16:36 dragoonis _ikke_, not on every CI build?
16:37 _ikke_ dragoonis: no
16:37 _ikke_ dragoonis: it's because you used -b before
16:37 dragoonis _ikke_, alright
16:37 InfoTest joined #git
16:38 dragoonis How's this? https://gist.github.com/dragoonis​/ee18195bcca15c2687457fcfb05812e9
16:38 MattMaker joined #git
16:38 GodGinrai dragoonis: why are you using a pre-existing repo in your CI instead of cloning?  Normally, CI is supposed to run stateless
16:38 pie__ _ikke_, its not chosen prefix but its getting there i think?
16:38 pie__ *preimage
16:39 dragoonis GodGinrai, coz if you clone it, it takes like 10 minutes bcoz the repo is 1GB in size :P
16:39 _ikke_ pie__: Eventually it will, the question is how long it would take
16:39 dragoonis so it's a one-time clone then a speedy "switch to this branch" thereafter
16:39 chachasmooth joined #git
16:39 GodGinrai dragoonis: but if you are just getting the code to test, you could do a shallow clone
16:39 Fallen0223 joined #git
16:39 _ikke_ GodGinrai: What is wrong with this approach?
16:40 brent__ joined #git
16:40 pie__ i mean its not like it matters how i feel, but i feel git should have the built in capability to upgrade its algorithm if necessary
16:40 pie__ and that it should do so before something happens not after...
16:40 GodGinrai _ikke_: if something happens to the repo outside of CI, then the CI can fail for reasons that have nothing to do with anything that actually happened in the code
16:40 _ikke_ pie__: They are working on it, it just takes time
16:40 vimal2012 left #git
16:40 pie__ oh they are?
16:40 _ikke_ GodGinrai: same with shallow clones..
16:40 pie__ last time i looked into it i got the feeling they didnt want to
16:41 GodGinrai _ikke_: that's only if something happens to the remote repo.  I'm talking about the repo local to the CI
16:41 _ikke_ pie__: Someone has been working over the last months / years removing all the hardcoded references to sha1 (length)
16:41 GodGinrai _ikke_: the one he is fetching into
16:41 pie__ ok,thats coforting :P
16:41 _ikke_ GodGinrai: pie__ The main challenge is backwards compattibility, which they are not sure about yet how to handle)
16:41 pie__ comforting
16:42 osse cloning takes a long time. fetching takes not as long.
16:42 dragoonis the .git folder on this repo is 950MB
16:42 _ikke_ I would use git archive then instead of a shallow clone, if you do not require a repo at all..
16:42 dragoonis mental
16:42 _ikke_ unless you use github, but then you can get an archive from the webinterface
16:42 pie__ _ikke_, i presume its more nuanced but id assume you should just upgrade old repos
16:42 _ikke_ pie__: Like you said, it's more nuanced
16:43 yushyin joined #git
16:43 GodGinrai osse: shallow clone wouldn't take as long.  Sure, it might still be longer than fetch.  But is it worth it to jeopardize the integrity of the tests for faster speeds?
16:43 _ikke_ pie__: It's not that the developers are clueless :)
16:43 osse what integrity?
16:43 dragoonis I know for a fact that the client won't delete their repository.
16:43 pie__ _ikke_, ;)
16:43 dragoonis (if that makes a diff)
16:44 pie__ well, thanks i guess
16:44 pie__ do you happen to have any mailinglist archive links by any chance?
16:44 pwndave joined #git
16:44 _ikke_ pie__: https://public-inbox.org/git/CAPp-Vrb_n6z39RLHZ4​AeUaBFiJfL3_xX8Utfq7+bTgzZrza58Q@mail.gmail.com/
16:44 GodGinrai _ikke_: Continuous integration is supposed to be automatic.  Where would you get the archive?  Set up a hook that uploads an archive whenever the repo changes?
16:44 _ikke_ GodGinrai: git archive
16:45 nettoweb joined #git
16:45 Vampire0_ joined #git
16:45 matsaman joined #git
16:45 GodGinrai _ikke_: git archive only works from within a git repo.  You can't use it to pull the code from a remote repo. (unless there is something I'm missing)
16:45 Ranhir joined #git
16:45 Gsham joined #git
16:46 _ikke_ GodGinrai: It does work with remote repos
16:46 oaao joined #git
16:46 _ikke_ Only some services disabled it
16:46 _ikke_ like github
16:46 GodGinrai _ikke_: wow.  I was not aware of this.
16:46 GodGinrai then yes
16:46 _ikke_ see --remote= option of git archive
16:46 GodGinrai that would obviously be preferable to a shallow clone
16:47 GodGinrai assuming that the remote server has this functionality enabled
16:47 osse I am confused. Our jenkins at work does git fetch && git checkout origin/master. What's bad about that?
16:47 _ikke_ github allows you also to download archives through the webinterface, but that's more limited
16:47 r_rios joined #git
16:49 zeroed joined #git
16:49 zeroed joined #git
16:50 _ikke_ http://github.com/org/proj​ect/archive/branch.tar.gz
16:50 _ikke_ Get's you a tarball of the last commit on the branch
16:51 nothingnew joined #git
16:52 e14 joined #git
16:52 moritz but you want meta data like commit hash, last commit etc. for CI purposes, so IMHO that's a poor choice
16:52 _ikke_ pie__: Note that not everyone has the possibility to upgrade, so it cannot be a hard cut-off point
16:53 thebope joined #git
16:53 moritz I also expect a CI system to keep a reference cache of a repo around so that it avoids a full clone each time.
16:53 pie__ _ikke_, what would stop them? though i do like the sound of this zsolt fellows idea but i see in the followup emails that thats probably not whats going to happen
16:53 jstimm joined #git
16:54 Hozy joined #git
16:54 synthroid joined #git
16:54 _ikke_ pie__: compattibility
16:55 _ikke_ pie__: centos 7 still ships git 1.8.x
16:55 _ikke_ Going to take a while before all git installations are up-to-date
16:55 le_melomane joined #git
16:56 pie__ i guess you cant force people to update but....
16:56 pie__ give a deprecation deadline
16:56 matoro joined #git
16:57 d0nn1e joined #git
16:57 redhedded1 joined #git
16:58 Cyp___ joined #git
17:03 kbs joined #git
17:03 Es0teric joined #git
17:03 texinwien_ joined #git
17:05 Gsham joined #git
17:05 User458764 joined #git
17:06 _ikke_ pie__: https://public-inbox.org/git/alpine.DE​B.2.20.1607180905320.28832@virtualbox/
17:06 _xor joined #git
17:07 thiago joined #git
17:10 clemf joined #git
17:10 yehowyada joined #git
17:11 madewokherd joined #git
17:11 multi_io joined #git
17:15 Sample joined #git
17:15 _ikke_ pie__: https://public-inbox.org/git/Pine.LNX.4.​58.0504291221250.18901@ppc970.osdl.org/
17:15 imack joined #git
17:15 _ikke_ (This is what Junio, the git maintainer, replied on the mailing list)
17:17 NeXTSUN joined #git
17:17 orbyt_ joined #git
17:17 osse do you mean linus?
17:18 bvcosta joined #git
17:18 _ikke_ No, I mean junio replied with a link to that post
17:19 nickabbey joined #git
17:19 osse it's amazing that i didn't see you posted *two* links
17:19 Topic for #git is now Welcome to #git, the place for git help. Just git help, apparently. | We are aware of the SHA-1 collision. See: http://jk.gs/t/a3 | 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 | This is serious business. https://madeitwor.se/git-jokes
17:19 le_melomane joined #git
17:20 pie__ osse, spacing \o/
17:21 _ikke_ "In other words, the security isn't in the hash. The hash is an added level
17:21 _ikke_ to make it much harder to fool, but it's not "the security". "
17:21 durham joined #git
17:23 le_melomane joined #git
17:23 _ikke_ Junio: "IOW, we want to continue the work to switch from SHA-1, but today's announcement does not fundamentally change anything and we do not panic. "
17:23 chachasmooth joined #git
17:24 nostrora joined #git
17:24 nostrora Git have to change SHA-1 to another hash function
17:24 nostrora http://shattered.io/
17:24 nostrora Git use SHA-1 right ?
17:24 _ikke_ nostrora: See topic
17:24 _ikke_ nostrora: https://public-inbox.org/git/CA+55aFxJGDpJXqpco​Pnwvzcn_fB-zaggj=w7P2At-TOt4buOqw@mail.gmail.co​m/T/#m57b6730ba5afee7100870396504edb8c909d5aad
17:25 tyreld joined #git
17:26 _ikke_ nostrora: Junio: "IOW, we want to continue the work to switch from SHA-1, but today's announcement does not fundamentally change anything and we do not panic. "
17:26 MattMaker joined #git
17:27 markmarkmark joined #git
17:27 texinwien_ joined #git
17:28 seemikehack joined #git
17:28 rj1 joined #git
17:29 aard_ joined #git
17:30 texinwien_ joined #git
17:32 debian joined #git
17:32 seemikehack Hey all. We ran across a strange scenario that I was tasked with tracking down, and even though I now know what's happening, I still don't know why. There is a change to a single line in a file that did not get merged from Branch A to Branch B at some point, and no subsequent merges into Branch B from any branch that has the correct line have brought over the change. The line was not modified on Branch B at the time of the merge, so seemingly no
17:32 seemikehack conflict would have been generated that could have been improperly handled. Why does the change refuse to propagate to this branch? What could have happened during the initial merge to prevent this? What can be done now to fix it?
17:35 nickabbey joined #git
17:35 svm_invictvs joined #git
17:38 texinwien_ joined #git
17:38 mgoodwin left #git
17:39 marcogmonteiro joined #git
17:39 rlb Vampire0_: I wasn't (either way) - thanks, but thinking about it; certainly makes sense.  Haven't messed with hooks much.
17:39 Eugene seemikehack - if the line was deleted as part of conflict-resolution for the merge
17:39 Eugene Its "merged", so re-merging to bring it in doesn't make sense. Add the line manually and move on with life.
17:41 seemikehack Eugene, so basically, there had to have been a merge conflict that was resolved incorrectly that I just don't know about.
17:41 Eugene seemikehack - or whoever did the merge deleted it. ;-)
17:41 seemikehack Eugene, OK, thanks for the tip. There really doesn't seem to be any other way it could have happened.
17:42 r00twrh joined #git
17:42 boombatower joined #git
17:43 thiago joined #git
17:44 i7c joined #git
17:47 thebope joined #git
17:51 Ryanar joined #git
17:51 keep_ joined #git
17:52 coreyfarrell joined #git
17:53 ertesx joined #git
17:53 alexandre9099 joined #git
17:55 gtrmtx joined #git
17:55 Rodya_ joined #git
17:55 dustinm joined #git
17:55 chachasmooth joined #git
17:56 gtrmtx how would one go about configuring git to automatically push any time a new commit happens?
17:56 coreyfarrell left #git
17:56 texinwien_ joined #git
17:56 seemikehack gtrmtx, I've seen clients have an option for that, but otherwise you could probably set up a script to do it for you.
17:57 gtrmtx a script?
17:57 seemikehack gtrmtx, a bash script, or an alias, or something.
17:57 thiago joined #git
17:58 gtrmtx ok
17:58 gtrmtx ill have to do some googling
17:58 seemikehack gtrmtx, see here: http://stackoverflow.com/a/23328996/5269153
17:58 gtrmtx what about clients? i was unaware that they existed other than gogs...anytime i run git i just do it via ssh
17:59 SwiftMatt joined #git
17:59 seemikehack gtrmtx, SourceTree has an option to always commit and push, egit has two buttons for "Commit" and "Commit and Push"
17:59 chardan joined #git
17:59 gtrmtx which one do you like more?
18:01 seemikehack gtrmtx, I prefer SourceTree, but it's only available on Windows and Mac, so I just use egit in Eclipse. There are some standalone clients that wrap egit. If you're on Linux and doing open source work, or are willing to pay, there's also SmartGit, which I *really* like.
18:02 seemikehack gtrmtx, almost forgot, I ran across GitKraken the other day, but haven't investigated it much. I know they have free and paid tiers.
18:02 kritzikratzi hi! is there a good trick to prevent me from checking in by adding a comment in a source file? (something like /*git-blocki regularly tweak things, and then for
18:02 cagomez joined #git
18:02 kritzikratzi ooops, hit enter by accident…. um… yea, something that prevents me from committing, as long as a specific text appears in source
18:02 kadoban kritzikratzi: You could have a pre-commit hook that checks for banned text of whatever format you like
18:02 kritzikratzi kadoban: that sounds like what i want. are you using something like this? i don’t want to reinvent the wheel
18:03 seemikehack gtrmtx, also, surprised I haven't been flamed for this already, but there's also ye olde faithefule git gui and gitk. The other solutions are just more self-contained and easier to get started with.
18:03 kadoban https://www.emptypath.com/git_pre_commit_testing is what I use (warning: of my own invention). I'm sure there's more well known tools for the job.
18:04 gtrmtx seemikehack, running everything in linux so sourcetree wont work
18:04 kadoban It's essentially functionality that's available just out of the box in git, but it doesn't work too well unless you add some stuff on top of it. You don't have to read that whole page, you can skip to the end to see the final result.
18:04 jalr[m] left #git
18:04 gtrmtx seemikehack, so i would install smartgit on server, and access it via web interface? or what
18:04 shgysk8zer0_ joined #git
18:04 seemikehack gtrmtx, GitEye is the standalone app that wraps egit. It's just an Eclipse RCP app that exposes egit, AFAIK.
18:05 seemikehack gtrmtx, all of these options I mentioned are local clients. Are you looking to administrate your own git repository on a server you control?
18:05 gtrmtx yes
18:05 gtrmtx i control server
18:06 gtrmtx all i want to do is automatically push to github anytime i do a git commit on the server
18:06 kadoban kritzikratzi: Let me know if you need assistance beyond what's provided there. If you're familiar with the 'grep' command that'll be enough to do the test you have to set up.
18:07 finalbeta joined #git
18:07 marcogmonteiro joined #git
18:07 seemikehack gtrmtx, Hmm, I think this may be an X/Y problem. Why don't you just set up GitHub as your remote?
18:10 hahuang65 joined #git
18:10 ych joined #git
18:10 grayjoc joined #git
18:11 ResidentBiscuit joined #git
18:11 Torrone joined #git
18:12 dsantiago joined #git
18:15 User458764 joined #git
18:16 tomaw joined #git
18:17 MattMaker joined #git
18:18 seemikehack gtrmtx, I hate to do this to you, but I've gotta go pick the kiddo up from school and head to lunch. I'll be back in about an hour.
18:20 Tobbi joined #git
18:20 orbyt_ joined #git
18:21 buckowski joined #git
18:21 DaveTaboola joined #git
18:22 le_melomane joined #git
18:22 dreiss joined #git
18:23 hahuang65 joined #git
18:23 jnavila joined #git
18:23 crayon joined #git
18:25 jimi_ joined #git
18:27 chachasmooth joined #git
18:29 kpease joined #git
18:30 theoceaniscool joined #git
18:30 gtrmtx seemikehack, youre good...when you get back, i think maybe thats how i did it?
18:31 HardlySe1n joined #git
18:31 durham_ joined #git
18:32 durham_ joined #git
18:32 gtrmtx my first command was git remote add origin git@github.com:username/repository.git
18:33 jedahan joined #git
18:36 j7k6 joined #git
18:36 oleksiyp joined #git
18:38 d0nn1e joined #git
18:40 Gsham joined #git
18:40 metalraiden34 joined #git
18:41 thebope joined #git
18:41 matoro joined #git
18:43 Spydar007 joined #git
18:43 Spydar007 joined #git
18:44 texinwien_ joined #git
18:45 MattMaker joined #git
18:45 kpease joined #git
18:46 robotroll joined #git
18:47 fatalhalt joined #git
18:47 prg3 joined #git
18:47 nidr0x joined #git
18:50 CAPRA-MAU joined #git
18:50 le_melomane joined #git
18:51 CAPRA-MAU left #git
18:51 texinwien_ joined #git
18:52 aspiers joined #git
18:52 mikecmpbll joined #git
18:54 PrashantJ joined #git
18:55 iliekcomputers joined #git
18:55 mablae joined #git
18:56 Es0teric joined #git
18:56 duderonomy joined #git
18:56 timetrex joined #git
18:58 TomyLobo joined #git
18:58 iliekcomputers left #git
18:58 NullableTruth joined #git
19:00 joki joined #git
19:01 alexandre9099 joined #git
19:03 kritzikratzi kadoban: thx, that’s a great start. already playing with it.
19:05 OerHeks joined #git
19:05 durham joined #git
19:05 jwest joined #git
19:05 texinwien_ joined #git
19:06 kadoban kritzikratzi: Cool, glad to help. Feel free to let me know if you run into any trouble with it.
19:07 kritzikratzi a few things are different as i’m on osx, but i’m getting there :)
19:07 matoro joined #git
19:07 cagomez joined #git
19:07 eroux joined #git
19:08 j7k6_ joined #git
19:08 gopar joined #git
19:08 jedahan joined #git
19:09 kadoban Ahh osx, interesting. *crosses fingers* I definitely never tested there, though as I recall I made some effort to make it vaguely platform-neutral.
19:10 kritzikratzi simple things are different, like ln doesn’t support „dash-dash“  arguments
19:10 m0viefreak joined #git
19:10 kadoban Ooo, hmm.
19:10 kadoban That's unfortunate. I should figure those out, probably not too too hard to be more general
19:11 kritzikratzi no, but a pain… and no big deal for me to change
19:12 intellix joined #git
19:13 kadoban I bet that means it's borked on BSDs as well
19:13 irqq joined #git
19:14 Sasazuka joined #git
19:14 pjdowson joined #git
19:14 overlord_tm joined #git
19:14 peterbec` joined #git
19:14 jimi_ joined #git
19:14 pjdowson left #git
19:15 dedicated_ joined #git
19:16 Sample joined #git
19:16 nickabbey joined #git
19:17 oleksiyp joined #git
19:18 Mof joined #git
19:19 jnavila joined #git
19:19 Gsham joined #git
19:20 jdl joined #git
19:20 oleksiyp joined #git
19:20 metalraiden34 joined #git
19:20 nickabbey joined #git
19:21 j7k6 joined #git
19:22 thiago joined #git
19:23 hahuang65 joined #git
19:23 synthroid joined #git
19:23 m4sk1n joined #git
19:24 texinwien_ joined #git
19:26 joeco joined #git
19:26 orbyt_ joined #git
19:27 kritzikratzi kadoban: i think it does something funny to empty directories
19:27 kritzikratzi (like…delete them)
19:28 thiago left #git
19:28 thiago joined #git
19:29 ojacobson Git does not have a representation of an empty directory (or of directories at all, really). Common convention is to put a .gitkeep or empty .gitignore file in any directory that needs to exist on checkout but has no content
19:29 kadoban kritzikratzi: Hmm, interesting. I wonder if I've just never run into that or if the behavior is different.
19:29 cdg joined #git
19:29 ojacobson (The database representation can store empty directories, but the staging area can't.)
19:29 mischat joined #git
19:30 Atm0spher1c joined #git
19:30 b1tchcakes joined #git
19:32 grawity joined #git
19:32 MattMaker joined #git
19:33 dvaske_ joined #git
19:34 kadoban kritzikratzi: I can't tell if I'd expect that behavior or not, or if so if there's a way I can work around that. I'll take a peek tonight though.
19:34 cworo joined #git
19:35 Murii_ joined #git
19:35 kritzikratzi kadoban: i don’t think it matters at all! was just a random funky case i came across
19:35 thebope joined #git
19:35 kritzikratzi here’s what i got so far: https://asdfg.me/up/git-hooks.zip
19:35 oleksiyp joined #git
19:35 kritzikratzi seems to do what i want, maybe i’ll put it on github soon if it turns out stable and fast enough
19:36 kritzikratzi now i installed it globally, which i find more convenient
19:36 kritzikratzi kk, gotta run… dinner time,
19:36 kritzikratzi kadoban: thx again for your help!
19:36 kadoban Ya, if it's a case I can handle better I'd like to though. Cool beans, have a good dinner. Ya, anytime.
19:36 sharpdream joined #git
19:37 e14 joined #git
19:37 Murii_ joined #git
19:37 elsevero joined #git
19:38 Murii_ joined #git
19:39 Sample joined #git
19:39 Murii_ joined #git
19:40 Murii_ joined #git
19:40 roelmonnens joined #git
19:41 kexmex joined #git
19:42 sieve joined #git
19:42 Murii_ joined #git
19:42 durre joined #git
19:43 lamont` joined #git
19:43 Murii_ joined #git
19:44 sieve We currently have some microservices (5 or so) which we are currently deploying with docker and puppet to AWS EC2 instances. I'm wondering if it might be socially acceptable to just deploy the code with git and then have a cron job that pulls master every x mins.
19:44 durham joined #git
19:44 hahuang65 joined #git
19:44 sieve Throw away all the docker and puppet stuff.
19:44 sieve Which seems to mostly cause annoyance and complications
19:45 lamont` joined #git
19:46 texinwien_ joined #git
19:47 aw1 joined #git
19:47 sonOfRa Does signing a git commit actually sign the entire object, or just the hash of the commit?
19:47 grawity it signs the object minus the signature header
19:47 dedicated joined #git
19:48 bremner sieve: well, you can have a look at !deploy
19:48 gitinfo sieve: Git is not a deployment tool, but you can build one around it (in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/deploy.html
19:48 Virox joined #git
19:48 Sasazuka joined #git
19:49 thiago joined #git
19:49 grawity not that it matters much, since the commit references the *data* by tree hash
19:49 sonOfRa grawity: it matters if you insert an evil colliding SHA1 :)
19:50 sonOfRa The question is whether a signature for a good commit is also valid for an evilly forged commit with the same SHA1, but actually different code changes
19:50 grawity sonOfRa: it doesn't, because you can insert colliding sha1's at *any* place
19:50 grawity if you sign the commit, you can collide the tree hash
19:50 grawity or a subtree hash
19:50 grawity or a blob hash
19:51 synthroid joined #git
19:51 dedicated_ joined #git
19:53 OerHeks left #git
19:56 whaley joined #git
19:56 durham joined #git
19:56 dsantiag_ joined #git
19:57 texinwien_ joined #git
19:58 FilipNortic joined #git
19:58 Anja_ joined #git
20:02 kexmex joined #git
20:03 le_melomane joined #git
20:03 rumble joined #git
20:04 sharp_dream joined #git
20:04 livingstn joined #git
20:08 CrypticGator joined #git
20:08 bket joined #git
20:09 khfeng joined #git
20:09 Flaghacker joined #git
20:09 meLon joined #git
20:10 CussBot joined #git
20:10 Fallen0223 joined #git
20:12 le_melomane joined #git
20:12 jedahan joined #git
20:14 CussBot_ joined #git
20:16 Flaghacker Hey guys! I have a sequence of commit, A -> B -> C, and I would like to remove B. I already tried "git rebase -p --onto B^ B" but that didn't change anything. How can I do this?
20:16 Decorater joined #git
20:16 Decorater uh
20:16 Eugene Flaghacker - !fixup
20:16 gitinfo Flaghacker: So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!
20:16 Decorater Where can I find the git bugtracker
20:16 Flaghacker Eugene, Yea that was the site that told me to execute that command, but it didn't work.
20:17 moritz !bug
20:17 gitinfo [!bugtracker] the developers of git don't use a bug tracker. If you want to report a bug, send an e-mail to the mailing list at git@vger.kernel.org (no subscription required; everyone uses "reply to all" when responding); you can review past discussions at http://public-inbox.org/git
20:17 Eugene !repro
20:17 gitinfo [!transcript] Please paste (using https://gist.github.com/ or similar) a transcript ( https://git.io/viMGr ) of your terminal session so we can see exactly what you see
20:17 cagomez joined #git
20:17 Decorater I would like to report 2 bugs that I experianced in Windows before on cygwin, cygwin64,mingw, and the mingw64 bit versions
20:17 Decorater ah
20:19 rgb-one joined #git
20:21 ahr3n joined #git
20:23 texinwien_ joined #git
20:23 peterbec` joined #git
20:26 Flaghacker Eugene, I'm sorry that took so long, rebase is very slow for some reason: https://gist.github.com/flaghacker​/546bbc4d415c1cbaeb3b7c78d10be3c2
20:26 thebope joined #git
20:27 Rodya_ joined #git
20:27 Decorater there is 2 bugs I encountered with rebase that I am emailing now.
20:27 Decorater that is on Windows
20:27 Decorater ok emailed
20:28 Decorater Also anyone know if I can somehow use MSVC to compile git to be standalone so users do not have to have cygwin, cygwin64, mingw, or even mingw64?
20:28 aidalgol joined #git
20:29 Decorater Also my git version
20:29 Decorater $ git --version git version 2.11.0.windows.1
20:29 snowalpaca joined #git
20:31 Wza joined #git
20:32 Flaghacker Eugene, Any idea what I'm doing wrong?
20:32 texinwien_ joined #git
20:32 Es0teric joined #git
20:33 perlpilot Flaghacker: you could try an interactive rebase  (I don't see that you're doing anything wrong)
20:33 perlpilot Flaghacker: what version of git are you using?
20:33 LiftLeft joined #git
20:33 hasc joined #git
20:33 Flaghacker perlpilot, git version 2.5.1.windows.1
20:34 jedahan joined #git
20:34 Flaghacker I'll an interactive rebase (once I figured out whta that id)
20:34 hasc joined #git
20:34 perlpilot Flaghacker: git rebase -i d70f2ca  # then just remove the commit that you don't want anymore
20:35 Sample joined #git
20:35 perlpilot Flaghacker: It will put you in your editor to edit the commits.  Dunno what that's likely to be for you.
20:35 gugah joined #git
20:36 perlpilot Flaghacker: random thought ... is your shell eating the ^ somehow?
20:36 grawity joined #git
20:37 Flaghacker perlpilot, The one that has ":wq" for save-and-exit. I thought it was Linux thing...
20:37 mloy vi/m
20:38 perlpilot according to my 15 seconds of googling, ^ is special to cmd.exe in some way
20:39 texinwien_ joined #git
20:40 UTAN_dev joined #git
20:40 perlpilot Flaghacker: so ... you could try git rebase --onto "769c251^" 769c251     (or whatever the appropriate quoting is)
20:40 Eugene Yup, windows ^ would be it
20:40 Flaghacker perlpilot, That would expalian a lot of stuff... I've had problems in the past but I always ended up using a workaround instead.
20:40 Eugene : and { are also special
20:41 perlpilot Eugene: what does it do with ^ ?  I haven't used Windows in forever.  : and { make sense, but I don't have a clue about ^
20:41 jimsio joined #git
20:41 griffindy joined #git
20:42 Eugene http://www.robvanderwoude.com/escapechars.php
20:42 Flaghacker Yes it's rebasing, taking a loooong time yet again.
20:42 kritzikratzi kadoban: i’ve just put in on github under the same license now: https://github.com/kritzikratzi/git-forbid
20:42 perlpilot ah, they use ^ instead of \ because of the path thing.  Makes some sense
20:42 Eugene Unless you're using a literal \\!
20:43 Eugene Or worse, a literal "\\\!"
20:43 Eugene (Which is \!)
20:43 Eugene (Maybe?)
20:43 Eugene Lesson: don't trust cmd
20:43 perlpilot I thought the lesson was to use Linux?  ;)
20:44 Eugene Gods no, linux is way worse about insane made-up standards
20:44 Eugene If you want a well-documented understandable system, find an old copy of System V Unix and type it in yourself
20:44 Eugene Bonus points if it was photocopied
20:45 perlpilot no love for NetBSD or FreeBSD?
20:45 Decorater or even OpenBSD^
20:45 robattila256 joined #git
20:45 bvcosta joined #git
20:45 Eugene Their source doesnt' fit in a 3-ring binder
20:45 Decorater also anyone got an idea with these issues on git rebase (2.11.0.windows.1) http://public-inbox.org/git/CAHikyLoK-h4tY_rxGik​aSSv6AmcrBAXiayDFTtLa44A9qMZDqA@mail.gmail.com/
20:47 Flaghacker perlpilot, Interactive rebase worked, thanks!
20:47 texinwien_ joined #git
20:48 Flaghacker Does anyone happen to know why git rebase would be very slow? It doesn't use that much CPU, memory or disk, it just takes a couple of minutes to fininsh.
20:48 Decorater https://github.com/AraHaan/TinyURL an repo where the rebase bugs on as well
20:48 _ikke_ Flaghacker: What OS?
20:49 Flaghacker _ikke_, Windows 10
20:49 perlpilot Flaghacker: You aren't using git over an SMB share are you?  (or CIFS, if that makes more sense)
20:49 askb joined #git
20:49 Decorater I am using Windows 7 Ultimate SP1 Build 7601
20:49 Decorater x64^
20:50 Flaghacker perlpilot, The whole project (including .git etc) is 1.6MB if that's what you're asking.
20:50 _ikke_ Flaghacker: What version of git are you using?
20:50 Flaghacker oops I misread SMB as 5MB
20:50 Flaghacker _ikke_, git version 2.5.1.windows.1
20:50 perlpilot Flaghacker: get a better font :)
20:50 Flaghacker And no, perlpilot, it's not on a share.
20:51 _ikke_ Flaghacker: You might want to use a newer version. THe maintainer of git for windows ported some part of git rebase to native c, increasing it's performance a lot
20:51 peterbec` joined #git
20:51 Mindi joined #git
20:51 Decorater on Windows try to use git v 2.11.0.windows.1
20:51 MicronXD joined #git
20:51 _ikke_ Flaghacker: Rebase is/was written in bash, causing a lot of forks. In linux, this is cheap, in windows, not so
20:51 Decorater and then you will notice some bugs I got with rebase
20:51 Decorater it is not slow for me, it just has bugs
20:51 synthroid joined #git
20:51 _ikke_ Decorater: What kind of bugs?
20:52 Flaghacker Is there an easy way to update git?
20:52 moritz depends on your OS and/or package manager
20:52 Decorater 1 that if I do git rebase -i -root and on all commits change pick to edit and append ``-S`` to every one that all of the commits get merged
20:52 Flaghacker moritz, W10 so no package manager :(
20:52 ryant joined #git
20:52 _ikke_ Flaghacker: Just download the latest installer
20:53 _ikke_ should be pretty easy
20:53 Decorater and then there is another one where if I clone a repo on Windows and not on github desktop and that I placed commits to the repo on github web and then when I rebase to squash the commits to 1 commit (some repos are doing it as a requirement for 1 commit PR's) that all of my commits on the remote (fork in this case) that is linked to an open pull request are discarded and then the pull request is somehow and oddly closed. It is super
20:53 Dougie187 left #git
20:53 perlpilot Flaghacker: I found https://git-scm.com/download/win from google  :)
20:53 MicronXD I'm afraid I'm totally screwed. Is there anyone who's managed to unscrew an accidental `stash pop` during a rebase?
20:53 moritz Flaghacker: good news for you: https://chocolatey.org/
20:53 peterbec` joined #git
20:53 _ikke_ MicronXD: Can you post a !transcript?
20:53 gitinfo MicronXD: 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
20:54 Flaghacker moritz, That's look very interesting, I've been wishing for this a long time. I'll check it out, thanks.
20:54 fatalhalt_ joined #git
20:54 Tobbi joined #git
20:54 perlpilot MicronXD: *Usually* git saves you from screwing yourself.  Have hope!
20:54 moritz Flaghacker: full disclosure, I haven't used it myself, but have come across it in several podcasts and/or blog posts
20:54 Decorater yeah unless you need to edit every commit from root to gpg sign them to append -S then it acts up and merges all commits you have to 1
20:54 moritz (not a windows user)
20:55 Decorater (where you lose all the history)
20:55 perlpilot Decorater: even then you didn't *lose* history, just made it slightly harder to get to.
20:55 peterbec` joined #git
20:56 MicronXD perlpilot: is there a way to restore my stash and abort my rebase?
20:56 Decorater yeah, at least I did not do git pull --force it before git log to show the commits and signitures
20:56 Decorater so then all I had to do was reclone
20:57 perlpilot MicronXD: do what _ikke_/gitinfo said.
20:57 peterbec` joined #git
20:57 perlpilot MicronXD: it's harder to tell without a transcript.
20:58 MicronXD perlpilot: _ikke_: https://gist.github.com/MicronXD/​db55d48f3ad6ac81eaa3bedaa91d860e
20:58 perlpilot MicronXD: btw, I've found "git stash pop" annoying, so I tend to "git stash apply" and "git stash drop" separately.
20:58 perlpilot MicronXD: you might want to start doing the same :)
20:59 _ikke_ MicronXD: I don't think the stash got applied
20:59 diogenese joined #git
20:59 _ikke_ MicronXD: You can use git stash list to verify
20:59 xissburg joined #git
21:00 Decorater also is there a way I can drop an extra merge commit that is shown on github but not in git?
21:00 _ikke_ I think you are just in a regular rebase conflict
21:00 Decorater even if I open it in rebase
21:00 perohig joined #git
21:00 _ikke_ Decorater: rebase by default never shows merges
21:00 Decorater hmm could you somehow force rebase to show them though?
21:01 snowalpaca joined #git
21:01 _ikke_ --preserve-merges
21:01 _ikke_ But if you want to get rid of a merge, then you obviously don't want that
21:01 Decorater Well lets say the merge commit does nothing
21:02 Decorater as a few commits on github has the data it has and well when I got the changes from github it wanted to do an merge commit instead
21:02 Decorater so essentially that merge commit should be deleteable
21:02 MicronXD _ikke_: I updated this gist. https://gist.github.com/MicronXD/​db55d48f3ad6ac81eaa3bedaa91d860e I see a message reading "stash@{0}: WIP on Settings-Flux-Additions: 4bce9bd Setting-Flux" followed by a list of messages that I presume were generated by the GitHub desktop app
21:03 perohig left #git
21:03 peterbec` joined #git
21:04 LiftLeft joined #git
21:04 _ikke_ MicronXD: You can run git rebase --abort to get back to the state before the rebase
21:04 jeffreylevesque joined #git
21:04 MicronXD I tried `git stash show -p`, but it's only showing 3 modifications (there were a probably 10-15)
21:04 CussBot joined #git
21:04 sieve joined #git
21:04 MicronXD _ikke_: will that restore the stash though?
21:05 Decorater like the top commit on this screenshot contains all the data that I pushed before it and those commits are also in the screenshot. (The top one should be deleted) http://puu.sh/ugO93/6a195377df.png
21:06 _ikke_ MicronXD: git stash pop doesn't drop in case of conflicts
21:06 livingstn joined #git
21:06 _ikke_ MicronXD: And in this case, it didn't even apply I believe
21:07 MicronXD _ikke_: oh thank god
21:08 texinwien_ joined #git
21:08 Es0teric joined #git
21:08 _ikke_ Decorater: Did you just sent something to the mailing list?
21:08 Decorater yeah
21:09 _ikke_ Note that git for windows has it's own bug tracker
21:09 Uesg joined #git
21:10 rt joined #git
21:10 Budyn joined #git
21:10 cdesai joined #git
21:10 _ikke_ https://github.com/git-for-windows/git/issues
21:10 _ikke_ Decorater: Or do you believe this is not just related to git for windows?
21:10 Decorater I think it is related to git in general
21:11 finalbeta joined #git
21:12 exitcode1 joined #git
21:12 Flaghacker I updates to 2.11 and rebase still is very slow. It'll probably be something with my pc I gueess.
21:12 _ikke_ alright
21:13 _ikke_ Decorater: It would have been better if you included what version of git you are using
21:13 Decorater 2.11.0
21:14 _ikke_ Right, but the people reading the mailing list don't read this :)
21:14 xaviergmail_ joined #git
21:15 mwfires joined #git
21:15 cagomez joined #git
21:16 MicronXD joined #git
21:16 cdg joined #git
21:16 _ikke_ Decorater: So you do git rebase --root -i, and change everything to edit, and then save+exit, and then everything is squashed?
21:17 Decorater yes
21:17 Decorater I do edit to add gpg to the commits
21:17 _ikke_ Seems to work for me
21:17 mwfires joined #git
21:18 Decorater hmm
21:18 matoro joined #git
21:18 _ikke_ I had a repo with 2 commits
21:18 _ikke_ did git rebase --root -i
21:18 _ikke_ changed both to edit
21:18 _ikke_ and it allowed me to edit both commits
21:18 _ikke_ and the resulting branch still had 2 commits
21:18 _ikke_ (I'm on linux btw)
21:19 Budyn joined #git
21:20 [Brain] joined #git
21:20 Budyn joined #git
21:21 ronny joined #git
21:23 Decorater hmm
21:23 Decorater I see
21:24 Decorater can you reporduce the clone part too where you placed commits using git web and then when you rebase those that are for an PR that it discards those commits magically?
21:26 rmb joined #git
21:27 texinwien_ joined #git
21:28 cgdub joined #git
21:29 ponyofdeath joined #git
21:29 ponyofdeath hi, trying to figure out how I can change where my cloned master is pointing to? it seems to be pointing to a branch's ref and i want to change that to track what the remote origin master is.
21:32 _ikke_ ponyofdeath: How did you determine where it's pointing to?
21:34 ponyofdeath did a git ref-list
21:34 ponyofdeath or just cat .git/config
21:35 _ikke_ Just trying to determine what you mean by pointing to
21:35 ponyofdeath it seems that when it was cloned it's master was set up to be a branch
21:35 _ikke_ master is a branch
21:35 le_melomane joined #git
21:35 _ikke_ Just a label pointing to a commit
21:35 Phylock joined #git
21:36 _ikke_ ponyofdeath: what does git branch -vv show?
21:36 ponyofdeath yes but not it a bit confusing for people wokring that commiting to master on that branch is actually not master on the remote
21:36 ponyofdeath * master d8ea8f0 [origin/feat/saltmole: ahead 1]
21:36 _ikke_ right
21:36 ponyofdeath and now trying to merge master back into that branch
21:37 _ikke_ git branch master --set-upstream-to origin/master
21:37 ponyofdeath i dont know how to do :)
21:37 _ikke_ ^^\
21:38 ponyofdeath _ikke_: thanks!
21:39 hagridaaron joined #git
21:40 texinwien_ joined #git
21:40 v0z_ joined #git
21:41 chardan joined #git
21:43 Gsham joined #git
21:46 Electrometro joined #git
21:49 texinwien_ joined #git
21:50 sieve joined #git
21:50 livingstn joined #git
21:52 AciD- joined #git
21:52 tg joined #git
21:56 texinwien_ joined #git
21:57 cagomez joined #git
21:58 jedahan joined #git
21:59 sieve joined #git
22:02 chipotle joined #git
22:05 scoobertron joined #git
22:06 kevwil joined #git
22:07 texinwien_ joined #git
22:07 Shakeel_ joined #git
22:08 shgysk8zer0_ joined #git
22:10 sieve joined #git
22:12 Masber joined #git
22:12 elsevero joined #git
22:14 not-an-aardvark joined #git
22:18 GodGinrai joined #git
22:19 ISmithers joined #git
22:21 ych joined #git
22:23 Sasazuka joined #git
22:23 texinwien_ joined #git
22:23 rgb-one left #git
22:23 dave0x6d joined #git
22:25 oleksiyp joined #git
22:26 DARSCODE joined #git
22:33 v0z_ left #git
22:33 tang^ joined #git
22:34 Sasazuka joined #git
22:35 Abbott left #git
22:37 gugah joined #git
22:37 joeco joined #git
22:39 d^sh joined #git
22:39 dan2k3k4 joined #git
22:42 bcc joined #git
22:42 bcc joined #git
22:44 jdl_ joined #git
22:44 exitcode1 joined #git
22:44 cbreak so, when's git switching to sha2 or sha3?
22:45 anuxivm joined #git
22:45 joshszep joined #git
22:45 cbreak or are annotated signed tags left out to rot?
22:45 Eugene !list
22:45 gitinfo [!mailing_list] The mailing list can be reached via git@vger.kernel.org. You don't need to subscribe to the list, you will always be put in cc on reply. Read archives at http://public-inbox.org/git
22:45 Eugene That is a great question for the mailign list ;-)
22:45 rwb joined #git
22:46 jdl_ I'm glad I'm not on the malign list!
22:46 pfrench joined #git
22:46 cbreak well, I don't really use annotated signed tags
22:46 durre joined #git
22:46 jdl_ There is still some work to do before a conversion to SHA256 can be done.
22:46 cbreak jdl_: sounds like the git devs are slow :(
22:46 jdl_ It is in the works, but not fully baked yet.
22:47 cbreak sha1 has been suspect for over 5 years
22:47 jdl_ I'm not telling them that.
22:47 Sasazuka joined #git
22:47 jdl_ Or, rather, if they are too slow, they could use your help?
22:47 cbreak I doubt that
22:48 cbreak I'm a C++ dev, I don't do C :)
22:48 jdl_ Version, um, 5 of the patch series was posted a day or two ago.
22:48 * jdl_ nods
22:48 aidalgol joined #git
22:48 cbreak but I'd expect the change to not be that big
22:48 cbreak you'd only have to add a new hashing algorithm, change the repo format on disk and in transit
22:48 cbreak and maybe add transition code
22:48 jdl_ The notion of OID == SHA1 is being fully abstracted out.
22:49 cbreak the bigger problem would be that there'd be incompatible repositories
22:49 jdl_ You would likely have to commit to a conversion or so.  Not sure what the expected path there is yet.
22:49 cbreak maybe they make object IDs versioned
22:50 cbreak git describe already uses a g prefix for git hashes
22:50 cbreak maybe it'll be h for sha2 or 3 :)
22:50 jdl_ Dunno.
22:50 nettoweb joined #git
22:50 Sasazuka__ joined #git
22:51 cbreak well, what ever, good to hear they're not deciding to give up on protecting repository integrity
22:51 jdl_ They are not.
22:53 livingstn joined #git
22:55 pfrench joined #git
22:55 SwiftMatt joined #git
22:56 finalbeta joined #git
23:00 orbyt_ joined #git
23:00 livingstn joined #git
23:01 dragoonis I'm looking to forcibly check out a target branch, but i'm being told I can't due to dirty working tree changes .. I'd like to ignore these and just blow away any dirty changes and take whatever is in the branch
23:01 dragoonis see my command here: https://gist.github.com/dragoonis​/6c331414a26e984faa0ec6537728917f
23:01 fmeerkoetter joined #git
23:02 afuentes joined #git
23:02 dragoonis will "git checkout -f " suffice?
23:03 venmx joined #git
23:10 tg joined #git
23:10 nettoweb joined #git
23:16 kevwil_ joined #git
23:17 Narwhaal joined #git
23:18 bannakaffalatta joined #git
23:18 ToBeCloud joined #git
23:19 gugah joined #git
23:20 cagomez joined #git
23:22 tertu joined #git
23:22 metalraiden34 joined #git
23:23 tertu quick question: i'm having some trouble seeing remote branches on my local copy of a repo that's on github
23:24 cagomez joined #git
23:25 qqx tertu: How are you trying to see them? `git branch`? If so use `git branch -r`.
23:26 scoobertron joined #git
23:27 tertu i still only see master when i do that
23:27 xissburg_ joined #git
23:28 Krenair joined #git
23:29 qqx What does `git config --get-all remote.origin.fetch` show?
23:32 dgonzo joined #git
23:33 tertu +refs/heads/master:refs/remotes/origin/master
23:36 sentriz joined #git
23:36 Gustavo6046 joined #git
23:38 qqx Looks like you used --single-branch when cloning. In which case it's doing what you told it to do.
23:38 thiago joined #git
23:38 qqx You should be able to use `git remote set-branches '*'` to reconfigure it to get everything. Then `git fetch` to actually do that.
23:39 tertu got it
23:41 dmj` joined #git
23:41 tekniq joined #git
23:43 menip joined #git
23:43 tertu thanks!
23:48 cdg joined #git
23:48 Kaisyu joined #git
23:48 Munt joined #git
23:50 Sasazuka joined #git
23:51 cwre joined #git
23:54 cwre joined #git
23:55 Munt Hey folks I've made a big mess of my online git repo.  I have a backup of the local git repo before I tried to do a lot of merging and inappropriate commands.  Is there a way to just replace the online repo with my local backup ?   Would it be reasonable to delete the online repo and recreate it with the backed up local copy ?  Maybe there's a better way
23:55 chipotle joined #git
23:57 explody joined #git
23:58 madewokherd joined #git

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