Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-06-24

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

All times shown according to UTC.

Time Nick Message
00:02 inflames joined #git
00:04 johnnyfive joined #git
00:08 dunpeal joined #git
00:12 githogori joined #git
00:13 githogori joined #git
00:13 cqi joined #git
00:18 axiom_1 joined #git
00:19 tesuji_ joined #git
00:21 romerocesar joined #git
00:21 mwhooker joined #git
00:22 fahadash joined #git
00:22 xaviergmail joined #git
00:26 Celelibi joined #git
00:29 rominronin joined #git
00:30 durham joined #git
00:30 durham joined #git
00:31 vacho joined #git
00:32 vacho gents. I made some file changes but I have not commited anything, how do I revert my changes? git reset --hard will do?
00:44 jhass vacho: git checkout .
00:44 dopesong joined #git
00:45 cdg joined #git
00:49 raijin joined #git
00:50 watabou joined #git
00:52 fstd_ joined #git
00:53 livingstn joined #git
00:59 codekyle joined #git
01:01 vacho jhass: thanks!
01:02 dopesong joined #git
01:03 dstarh joined #git
01:04 tgunr joined #git
01:06 hecliunyx joined #git
01:07 st0opkid joined #git
01:08 st0opkid joined #git
01:10 nick123 joined #git
01:11 GandalfTheWizard joined #git
01:12 kasper93_ joined #git
01:17 dmto joined #git
01:24 MrJones left #git
01:29 Literphor joined #git
01:30 rominronin joined #git
01:30 romerocesar joined #git
01:31 EvilPenguin joined #git
01:33 Dougie187 joined #git
01:34 nonconvergent joined #git
01:35 texinwien joined #git
01:41 sangy joined #git
01:46 dmto joined #git
01:48 ilbot3 joined #git
01:48 Topic for #git is now Welcome to #git, the place for git help and plumbing advice | Public logs at http://goo.gl/BuUi5o | Current stable version: 2.9.0 | First visit? Read: http://jk.gs/git | Getting "cannot send to channel"? /msg gitinfo .voice | Remember to keep your porcelain clean
01:52 pks joined #git
01:53 robogoat joined #git
01:53 z8z joined #git
01:53 eletuchy joined #git
01:54 xall joined #git
01:56 boombatower joined #git
01:58 boombatower joined #git
01:58 watabou joined #git
01:58 i7c joined #git
02:00 thiago joined #git
02:01 eletuchy_ joined #git
02:02 raijin joined #git
02:03 eletuch__ joined #git
02:04 drbean_ joined #git
02:05 rubyonrailed joined #git
02:10 pierre365 joined #git
02:11 sangy joined #git
02:13 rubyonra_ joined #git
02:14 rahtgaz joined #git
02:16 zivester joined #git
02:19 tgunr joined #git
02:21 romerocesar joined #git
02:22 hhee joined #git
02:22 ejb joined #git
02:22 nonconvergent joined #git
02:24 fahadash if my current working branch is A and I pull master from the remote; it pulls it and merges into A but never merges anything into master. I don't want that. I want it to pull and merge into local master and not touch my current working branch. How do I do that?
02:24 fahadash I don't want to have to checkout master and do the pull
02:24 eletuchy joined #git
02:24 rahtgaz joined #git
02:25 dstarh joined #git
02:27 dmto joined #git
02:29 pierre365 joined #git
02:29 docnvk joined #git
02:31 Sonicbit joined #git
02:31 rominronin joined #git
02:35 eletuchy joined #git
02:35 pierre365 joined #git
02:36 Goplat joined #git
02:36 rewt fahadash, a `git fetch` should get everything from the remote repo
02:36 diogenese joined #git
02:39 Ordentlig joined #git
02:40 Ordentlig Hi, I would like to back up all my git repos from server, what would be considered a good approach?
02:40 jchonig joined #git
02:40 Ordentlig I have about 60 repos or more on my private server.
02:41 rewt should be able to just back up the directories
02:42 hexagoxel joined #git
02:43 Ordentlig rewt, like if I have 3 harddrives, wouldn't it be better to `pull` all repos from server every night or something?
02:43 Ordentlig rewt, rather than copy --replacing, or creating multiple versioned copies.
02:44 rewt i thought you meant on tape or something
02:44 rewt but yeah, doing a fetch or clone are good
02:47 Ordentlig rewt, hmm, so I am using gitolite, so perhaps just cloning gitolite-admin and then keep pulling gitolite-admin for changes and whenever it changes I read the gitolite.conf and extract all repos and start cloning the new repos. Something like that, perhaps, hmm.
02:48 harish joined #git
02:48 Ordentlig "thinking intesifies"
02:48 rewt i'm not familiar with gitolite... does it have any buil-in backup functionality?
02:49 rewt built-in*
02:50 Ordentlig I doubt it, it is quite lightweight, can't seem to find anything about it in the docs.
02:51 tjone270 joined #git
02:51 Ordentlig you basically have an admin repo, with user keys and a configuration file for your repos looking something like this: https://i.imgur.com/gsHnoEK.png
02:52 Ordentlig so I could just, `grep repo ./gitolite.conf | sed 's/repo//` and then iterate those repos when cloning
02:53 tjone270 joined #git
02:54 b1tchcakes joined #git
02:55 acetakwas joined #git
02:55 Literphor joined #git
02:56 rewt yeah, that could work
02:57 acetakwas joined #git
02:57 jaguarmagenta joined #git
02:57 rewt may want to fetch instead of clone, so you don't lose anything in case something happens on the remote
02:59 b1tchcakes joined #git
03:01 dmto joined #git
03:02 chachasmooth joined #git
03:03 b1tchcak_ joined #git
03:04 aavrug joined #git
03:05 aavrug joined #git
03:07 b1tchcakes joined #git
03:11 eletuchy joined #git
03:11 nick123 joined #git
03:13 sitaram Ordentlig: gitolite doesn't have "backup" but it has a very powerful and flexible mirroring capability
03:13 Literphor joined #git
03:13 sitaram which, as the KDE guys found out once (!) is not the same as "backup" but if you squint, and you set reflogs on bare servers, then it's close enough
03:14 sitaram Ordentlig: getting the list of repos: "gitolite list-phy-repos" (because "wild repos")
03:14 Ordentlig sitaram, would that be applicably if I have the harddrives on the same machine?
03:14 Ordentlig applicable*
03:14 pierre366 joined #git
03:14 sitaram Ordentlig: if it's a different Unix userid, sure
03:15 sitaram Ordentlig: what version of gitolite are you using
03:15 Ordentlig the newest I guess
03:15 dstarh joined #git
03:16 sitaram look in contrib/triggers/file_mirror
03:16 sitaram instructions within the file
03:16 cdg joined #git
03:16 sitaram Ordentlig: don't forget a different userid on the same box is equivalent to "external"
03:19 pierre365 joined #git
03:20 Ordentlig sitaram, seems too complicated for me right now, I think I'll go with something like this for now, http://sprunge.us/QYWP
03:21 k3rn31 joined #git
03:22 Ordentlig sitaram, thanks for the tip though I'll try to learn about it later!
03:22 romerocesar joined #git
03:23 rubyonrailed joined #git
03:25 sitaram oh good lord is this a one-time thing?  You're only doing a clone
03:26 sitaram at least use "--mirror" on the clone man!  Otherwise the refs are all wonky and if you ever have to restore (this *is* a backup right?) you have to fix it up manually
03:26 sitaram and seriously, why are you parsing the gitolite-admin conf file; just use "gitolite list-phy-repos" it'll give you a nice list
03:27 Ordentlig sitaram, I am going to do a `pull` too, that's the `....`
03:27 SimonNa joined #git
03:28 Ordentlig sitaram, oh, thanks
03:29 Ordentlig what would the --mirror do? How would it affect me?
03:29 Ordentlig can't find --mirror in `man git`
03:29 gitinfo the git manpage is available at http://jk.gs/git.html
03:30 Ordentlig nvm found it
03:30 Ordentlig sitaram, but how can I use `gitolite list-phy-repos` if I want to backup on a non-local machine?
03:31 Ordentlig That won't work.
03:31 a_thakur joined #git
03:31 loui joined #git
03:32 rominronin joined #git
03:32 sitaram I thought you said "same machine" above
03:33 Ordentlig sitaram, yeah, but this makes it more portable, incase I ever need to back it up somewhere else.
03:33 sitaram anyway, a "same machine, same user" backup is best done like this (simplest code possible): http://paste.fedoraproject.org/383960/67391031
03:34 Ordentlig sitaram, Thank you so much!
03:36 sitaram Ordentlig: the problem with your way of pulling repo names from gitolite-admin is that you're not considering repo groups, and you're not considering wild repos.  If you're sure you don't use those features, that's fine, but I wouldn't call it "portable" (in a different sense) for those reasons
03:37 sitaram Ordentlig: please note that the script I gave you is untested
03:37 sitaram but you can adapt those two commands to your script; that's the important bit I guess
03:38 sitaram (i.e., clone --mirror, and fetch later)
03:43 muthu joined #git
03:44 Ordentlig sitaram, Oh, I see. I don't think I'll ever use repo groups, but duly noted for the occasion whenever I am about to use it, thanks for the heads up!
03:47 fscala joined #git
03:47 git2342332 joined #git
03:48 Literphor joined #git
03:48 qq joined #git
03:52 xaviergmail joined #git
03:52 boombatower joined #git
03:53 Ordentlig sitaram, I get this `fatal: The remote end hung up unexpectedly` from some repos when fetching, should I be concerned?
03:53 sitaram check if they're empty
03:53 sitaram if they're not, it needs investigating
03:55 ensyde__ joined #git
03:55 netj joined #git
03:55 Phylock joined #git
03:56 Ordentlig sitaram, they are not empty, I don't get any problems when I `git clone repo && cd repo && git pull` however.
03:58 dstarh joined #git
03:58 Ordentlig it is the `fetching` part that breaks on the bare repos
03:59 sitaram does "git ls-remote" work?
03:59 Ordentlig ye
03:59 sitaram if that works, and fetch doesn't, I'm lost... (and grasping at straws like "network problem")
04:00 sitaram because from gitolite's point of view both do essentially the same thing
04:01 sitaram in fact gitolite cannot distinguish between those two; any differences are deep inside git, at git's packet protocol level
04:01 taylorm joined #git
04:01 Ordentlig oh, nvm, it seems that the stderr doesn't flush at the appropriate time, giving misleading error messages from my script
04:02 sitaram also, are you still doing a clone without "--mirror"?  Other than the "future inconvenience" I mentioned, a plain clone creates a non-bare repo
04:03 Ordentlig sitaram, I followed your advise, with the exception to the `gitolite` approach to retrieving the repos
04:04 Ordentlig sitaram, http://sprunge.us/VGcc
04:04 GodGinrai joined #git
04:05 GodGinrai If I run `git remote update` under a different user than the one who made the local repo, (including different ssh creds) is there anything that could go wrong?
04:06 eletuchy joined #git
04:10 dreiss joined #git
04:11 unforgiven512 joined #git
04:11 eletuchy joined #git
04:11 lindenle joined #git
04:12 unforgiven512 joined #git
04:12 unforgiven512 joined #git
04:12 sctskw joined #git
04:13 rscata joined #git
04:14 romerocesar joined #git
04:15 eletuchy_ joined #git
04:18 aidalgol joined #git
04:19 xorox90 joined #git
04:20 eletuchy joined #git
04:20 aidalgol joined #git
04:20 xorox90 joined #git
04:21 eletuch__ joined #git
04:23 xorox90 joined #git
04:24 eletuchy joined #git
04:28 dunpeal joined #git
04:29 eletuchy_ joined #git
04:29 a_thakur joined #git
04:30 b1tchcakes joined #git
04:31 eletuch__ joined #git
04:32 nonconvergent joined #git
04:33 rominronin joined #git
04:35 eletuchy joined #git
04:35 xall joined #git
04:36 dreiss joined #git
04:36 eletuchy_ joined #git
04:38 eletuch__ joined #git
04:40 aidalgol joined #git
04:41 phanimahesh joined #git
04:42 eletuchy joined #git
04:44 WakiMiko joined #git
04:44 zumba_ad_ joined #git
04:46 eletuchy_ joined #git
04:48 blackwind_123 joined #git
04:49 eletuchy joined #git
04:50 settermjd joined #git
04:51 eletuchy_ joined #git
04:52 settermjd joined #git
04:53 cyphase joined #git
04:53 iveqy joined #git
04:54 pierre365 joined #git
04:57 eletuchy joined #git
04:59 eletuch__ joined #git
04:59 Cabanossi joined #git
05:00 Myrl-saki joined #git
05:03 eletuchy joined #git
05:05 hhee joined #git
05:06 eletuchy_ joined #git
05:07 rchavik joined #git
05:07 eletuch__ joined #git
05:08 mangolisk joined #git
05:09 cyphase joined #git
05:09 eletuchy joined #git
05:10 romerocesar joined #git
05:11 subhojit777 joined #git
05:12 nick123 joined #git
05:12 eletuchy_ joined #git
05:14 eletuch__ joined #git
05:15 Andrew_K joined #git
05:16 eletuchy joined #git
05:18 maestrojed joined #git
05:19 taylorm joined #git
05:19 lordjancso joined #git
05:20 eletuchy_ joined #git
05:20 feignix joined #git
05:22 eletuch__ joined #git
05:25 scoobertron joined #git
05:27 eletuchy joined #git
05:27 nidr0x joined #git
05:28 dimi1947 joined #git
05:29 stuh84 joined #git
05:29 a_thakur joined #git
05:29 eletuchy_ joined #git
05:33 a_thakur_ joined #git
05:34 rominronin joined #git
05:34 StuartMI joined #git
05:34 tinajohnson___ joined #git
05:36 nidr0x joined #git
05:43 dreiss joined #git
05:46 ThomasLocke_ joined #git
05:46 a_thakur joined #git
05:49 emPi joined #git
05:49 TomyLobo joined #git
05:50 githogori joined #git
05:54 aliasd joined #git
05:54 zumba_ad_ joined #git
05:54 jason237 joined #git
05:56 _ikke_ qsx: Northern part, only from a distance
05:57 sauvin joined #git
05:58 jceb joined #git
06:00 noyb joined #git
06:03 Repox joined #git
06:04 skirk joined #git
06:04 Oebele joined #git
06:08 Ordentlig joined #git
06:10 freimatz joined #git
06:11 romerocesar joined #git
06:13 zumba_addict joined #git
06:15 moei joined #git
06:15 cdown joined #git
06:15 ShekharReddy joined #git
06:22 raindev joined #git
06:22 mithenks joined #git
06:23 MrMojito2 joined #git
06:24 dopesong joined #git
06:24 dminca joined #git
06:25 dminca joined #git
06:25 rominronin joined #git
06:25 dminca joined #git
06:28 zeroed joined #git
06:28 zeroed joined #git
06:28 eijk joined #git
06:29 h0bbit joined #git
06:32 cyan__ joined #git
06:33 Alienpruts joined #git
06:34 lindenle joined #git
06:36 Andrew_K|2 joined #git
06:37 Sceorem joined #git
06:38 qt-x joined #git
06:40 dopesong joined #git
06:40 nnyk_ joined #git
06:40 jaguarmagenta joined #git
06:41 dopesong_ joined #git
06:42 FroMaster joined #git
06:43 DieguezZ joined #git
06:45 eijk joined #git
06:46 eletuchy joined #git
06:47 douglascorrea joined #git
06:48 Andrew_K joined #git
06:49 dopesong joined #git
06:50 nonconvergent joined #git
06:50 ki0 joined #git
06:51 liuzhen joined #git
06:54 FroMaster joined #git
06:55 jstein joined #git
06:56 xall joined #git
06:56 noyb joined #git
06:58 vuoto joined #git
06:58 zamba joined #git
07:04 Oebele joined #git
07:04 Rynomster joined #git
07:04 xaviergmail joined #git
07:06 encod3 joined #git
07:07 nezZario Can anyone let me know if my plan sounds rational?  I have a small team I'm trying to organize.  We have maybe 10 repositories.  I have 2 developers that are under me.  I am the lead dev.  Aside from us three, there is a designer and the owner with little technical knowledge.  They know how to use HTML but no server side langauges, and certainly not
07:07 nezZario linux/cli
07:07 dopesong joined #git
07:08 dan2k3k4 joined #git
07:09 nezZario (^sorry for the wrap..) .. So what I'm wanting to do is setup all of our repositories in a staging area and just use git hooks and branches.  People would commit to staging, which would cause the staging area to pull -- and same for master/production.  The issue I have is that our code doesn't run well without a lot of "fluff" (redis servers, database, etc)
07:09 nezZario .. so what I'm thinking is in addition to the staging area that I'd just host the repos on the same server and give the 2 non-techies a 5  minute web interface to commit, push, pull, etc ...
07:10 cyphase joined #git
07:13 nick123 joined #git
07:14 dopesong joined #git
07:14 ahmedelgabri joined #git
07:15 romerocesar joined #git
07:16 Hounddog joined #git
07:16 baraba joined #git
07:17 elastix joined #git
07:17 newlifer joined #git
07:21 King_Hual joined #git
07:21 King_Hual joined #git
07:22 jungsubk joined #git
07:25 King_Hual joined #git
07:25 King_Hual joined #git
07:25 TomyWork joined #git
07:25 Balliad joined #git
07:28 HD|Laptop joined #git
07:31 vuoto joined #git
07:32 cyphase joined #git
07:33 HD|Laptop Hi all
07:33 HD|Laptop So, I have cloned a repo using git svn clone -s
07:33 iveqy joined #git
07:34 texinwien joined #git
07:34 HD|Laptop its two branches (besides trunk) and one tag show only up with git branch -a / git show-ref
07:34 HD|Laptop naming e.g. refs/remotes/origin/Layout_Design
07:34 HD|Laptop Now, how do I make these branches (and the tag) show up properly?
07:35 gigq joined #git
07:37 jceb joined #git
07:38 binaryplease1 joined #git
07:40 Literphor joined #git
07:44 quakephil joined #git
07:47 multi_io joined #git
07:47 FrontPsych joined #git
07:48 moritz HD|Laptop: well, it's a remote branch; if you want to develop on it, you start a local branch based on the remote branch
07:48 moritz so git branch --set-upstream Layout_Design origin/Layout_Design
07:49 moritz or shorter: git checkout Layout_Design # also switches to the branch
07:51 raindev joined #git
07:52 clmsy joined #git
07:52 Anticom joined #git
07:52 ToBeCloud joined #git
07:53 dmto joined #git
07:53 davisonio joined #git
07:53 fees joined #git
07:56 HD|Laptop moritz, well the point is that the repository is entirely local and the "remote" is an old svn server which I want to burn in a bonfire
07:57 vuoto joined #git
07:57 moritz HD|Laptop: anyway, same commands to do
07:57 HD|Laptop thanks :)
07:58 rkazak joined #git
08:00 rasto joined #git
08:00 Balliad joined #git
08:00 _noblegas joined #git
08:01 stonerfish joined #git
08:02 eletuchy_ joined #git
08:02 PCatinean joined #git
08:02 HD|Laptop moritz, how do I convert the tag that's currently a branch "origin/tags/connection"?
08:03 DieguezZ joined #git
08:04 eletuch__ joined #git
08:04 Sanle joined #git
08:05 mdw joined #git
08:07 eletuchy joined #git
08:08 gigq joined #git
08:09 fracting joined #git
08:09 HD|Laptop moritz, hmm git checkout branchname worked for one branch, but another branch "frontend_templates" gives error: pathspec 'frontend_templates' did not match any file(s) known to git.
08:09 DistantStar joined #git
08:09 kurkale6ka joined #git
08:09 lss8 joined #git
08:10 eletuchy_ joined #git
08:11 iLembus joined #git
08:11 Softeisbieger joined #git
08:13 eletuchy joined #git
08:14 romerocesar joined #git
08:15 eletuch__ joined #git
08:18 tristanp_ joined #git
08:18 eletuchy joined #git
08:19 circ-user-Oetft joined #git
08:20 eletuchy_ joined #git
08:22 jceb joined #git
08:23 eletuchy joined #git
08:24 jason237 joined #git
08:26 marcogmonteiro joined #git
08:27 jceb joined #git
08:29 eletuchy_ joined #git
08:31 eletuch__ joined #git
08:34 dunpeal joined #git
08:34 acetakwas joined #git
08:34 eletuchy joined #git
08:37 eletuchy_ joined #git
08:37 a_thakur joined #git
08:38 DARSCODE joined #git
08:39 HD|Laptop joined #git
08:39 HD|Laptop joined #git
08:40 cdown joined #git
08:40 eletuchy joined #git
08:41 dsdeiz joined #git
08:42 texinwien joined #git
08:42 eletuchy_ joined #git
08:44 eletuch__ joined #git
08:47 qswz joined #git
08:47 vuoto joined #git
08:47 qswz is it possible to git diff 2 branches but only showing intersecting files?
08:47 eletuchy joined #git
08:47 harish joined #git
08:49 eletuchy_ joined #git
08:49 Balliad joined #git
08:51 eletuch__ joined #git
08:51 wrouesnel joined #git
08:53 notebox joined #git
08:53 Repox joined #git
08:54 courrier joined #git
08:54 djb-irc joined #git
08:54 eletuchy joined #git
08:56 eletuchy_ joined #git
08:58 eletuch__ joined #git
09:00 eletuchy joined #git
09:00 carell joined #git
09:02 voiter2 joined #git
09:02 dsantiago joined #git
09:04 hk238 joined #git
09:05 ahmedelgabri joined #git
09:06 HD|Laptop joined #git
09:07 voiter2 how can i isolate two or more commits that are not consecutive?
09:07 voiter2 reset soft, stash, repeat?
09:07 eletuchy joined #git
09:09 eletuchy_ joined #git
09:10 qswz left #git
09:10 amitprakash joined #git
09:10 amitprakash Hi, like git clone --recursive, is there a git archive --recursive?
09:10 zeroed joined #git
09:13 eijk joined #git
09:14 nick123 joined #git
09:15 romerocesar joined #git
09:15 eletuchy joined #git
09:18 mischat joined #git
09:18 eletuchy_ joined #git
09:23 eletuchy joined #git
09:25 eletuch__ joined #git
09:28 eletuchy joined #git
09:30 notarobot joined #git
09:32 eletuchy_ joined #git
09:33 eletuch__ joined #git
09:34 dunpeal joined #git
09:34 FunkyAss joined #git
09:36 eletuchy joined #git
09:37 barteks2x joined #git
09:39 ahmedelgabri joined #git
09:39 qt-x joined #git
09:40 eletuchy_ joined #git
09:40 fornax joined #git
09:41 eletuch__ joined #git
09:43 hk238 joined #git
09:45 feignix joined #git
09:45 eletuchy joined #git
09:45 Sanle left #git
09:46 Tryop88 joined #git
09:46 Sanle joined #git
09:46 gsingh joined #git
09:46 lss8 joined #git
09:47 eletuchy_ joined #git
09:50 jceb joined #git
09:50 iLembus joined #git
09:50 eletuchy joined #git
09:50 nonconvergent joined #git
09:52 Guest8082 joined #git
09:52 iveqy joined #git
09:53 eletuchy_ joined #git
09:54 digidog joined #git
09:55 pierre365 joined #git
09:56 ImJune joined #git
09:59 Fijit joined #git
09:59 dsdeiz joined #git
09:59 dsdeiz joined #git
10:01 keltvek joined #git
10:04 phanimahesh voiter2: Define isolate.
10:04 leehambley joined #git
10:05 gechr joined #git
10:05 dunpeal joined #git
10:09 howdoi joined #git
10:10 djb-irc joined #git
10:11 choki joined #git
10:12 ahmedelgabri joined #git
10:14 eletuchy joined #git
10:14 afuentes joined #git
10:16 sdothum joined #git
10:17 phanimahesh amitprakash: You have to use a helper script, http://stackoverflow.com/questions/​14783127/git-archive-export-with-su​bmodules-git-archive-all-recursive looks good
10:17 stratos joined #git
10:18 nnyk_ joined #git
10:18 jeffreylevesque joined #git
10:18 Ordentlig left #git
10:20 settermjd joined #git
10:24 romerocesar joined #git
10:24 ntonjeta joined #git
10:27 mdw joined #git
10:28 fornax joined #git
10:29 subhojit777 joined #git
10:32 fuchstronaut joined #git
10:36 lindenle joined #git
10:37 raijin joined #git
10:37 chachasmooth joined #git
10:39 Segfault_ joined #git
10:40 jaguarmagenta joined #git
10:43 amitprakash phanimahesh, yeah, did that :)
10:44 _ikke_ git has nothing built-in
10:45 jason237 joined #git
10:46 vuoto joined #git
10:48 ElChicoNube joined #git
10:50 diegoaguilar joined #git
10:54 dsantiago joined #git
10:55 johnmilton joined #git
10:56 cbreak voiter2: cherry-pick them somewhere
10:58 a_thakur joined #git
10:59 nivag joined #git
10:59 Pulp joined #git
11:01 texinwien joined #git
11:03 jimi_ joined #git
11:06 dunpeal joined #git
11:07 playfullyExist joined #git
11:07 meicode joined #git
11:09 psprint When I'm switched to branch X, will git push origin push to that branch?
11:12 barteks2x joined #git
11:12 cbreak psprint: depends
11:12 cbreak psprint: look at man git-config, search for push.defaults
11:12 gitinfo psprint: the git-config manpage is available at http://jk.gs/git-config.html
11:12 isxek joined #git
11:13 ojdo joined #git
11:15 nick123 joined #git
11:16 psprint thanks
11:17 djb-irc joined #git
11:23 ImJune joined #git
11:24 voiter2 cbreak, thanks
11:25 dopesong joined #git
11:26 allcentury joined #git
11:26 dopesong_ joined #git
11:27 raindev joined #git
11:28 subhojit777 joined #git
11:29 rnsanchez joined #git
11:31 romerocesar joined #git
11:33 zivester joined #git
11:35 paul424 joined #git
11:36 ahmedelgabri joined #git
11:36 acetakwas joined #git
11:39 acetakwas joined #git
11:43 raininja joined #git
11:45 dsdeiz_ joined #git
11:47 b1tchcakes joined #git
11:47 rbern joined #git
11:47 toogley joined #git
11:49 muthu joined #git
11:49 cyphase joined #git
11:50 notebox joined #git
11:51 romerocesar_ joined #git
11:52 dopesong joined #git
11:52 davisonio joined #git
11:52 rominronin joined #git
11:53 sobersabre joined #git
11:54 sobersabre hi, someting stupid.
11:55 sobersabre I'm trying to push using a permitted key to github.com, but the error says "ERROR: Permission to user1/repo1.git denied to user2."
11:55 sobersabre how do I override this "user" ?
11:55 sobersabre Is it some ssh connecting host signature I need to delete?
11:57 dopesong joined #git
11:58 rgrinberg joined #git
11:59 _ikke_ git remote -v show
11:59 _ikke_ does it mention user2 there?
11:59 carell joined #git
11:59 jhass sobersabre: why do you have more than one user anyway..... well, add something like Host ghsuser1 Hostname github.com IdentityFile usera Host ghuser2 Hostname github.com IdentityFile userb to your ~/.ssh/config
12:00 jhass and use git@ghuser1 or git@ghuser2 respectively in your remotes
12:00 Andrew_K joined #git
12:02 eletuchy joined #git
12:02 synthroid joined #git
12:03 ensyde__ joined #git
12:04 dan2k3k4 joined #git
12:04 ahmedelgabri joined #git
12:05 Tryop88 joined #git
12:06 dunpeal joined #git
12:07 quakephil joined #git
12:08 phanimahesh joined #git
12:08 dan2k3k4 joined #git
12:08 Sceorem joined #git
12:10 d0nkeyBOB joined #git
12:11 ramsub07 joined #git
12:11 douglascorrea joined #git
12:12 fscala joined #git
12:13 armyriad joined #git
12:13 davisonio joined #git
12:13 d0nn1e joined #git
12:17 mattcen joined #git
12:17 quakephil joined #git
12:17 piattino joined #git
12:17 dopesong joined #git
12:19 raindev joined #git
12:21 armyriad joined #git
12:24 Gamecubic joined #git
12:28 gechr joined #git
12:30 rwb joined #git
12:31 jceb joined #git
12:32 mdw joined #git
12:34 sobersabre jhass: I got it, thanks!
12:34 sobersabre I not only have more than one user, but also more than one server :-] life is tough :)))
12:35 CEnnis91 joined #git
12:36 a_thakur_ joined #git
12:37 dunpeal joined #git
12:38 jh3 joined #git
12:38 jeffreylevesque joined #git
12:39 gechr joined #git
12:39 ensyde__ joined #git
12:40 ahmedelgabri joined #git
12:41 ramsub08 joined #git
12:41 jaguarmagenta joined #git
12:41 cbreak sobersabre: use ssh-add, and only add the key you want to use
12:42 phanimahesh Also just add a `User git` in there.
12:42 phanimahesh I map github to host gh, bitbucket to host bb and so on, so clones are usually git clone gh:user/repo
12:43 phanimahesh In your case, you need multiple host entries, one with each user's IdentityFile and appropriate Hostname, User
12:43 playfullyExist joined #git
12:44 BaconWings joined #git
12:44 ramsub07 joined #git
12:45 nonconvergent joined #git
12:45 cbreak phanimahesh: why bother with separate keys?
12:46 nettoweb joined #git
12:46 LeMike joined #git
12:46 Macaveli joined #git
12:46 phanimahesh cbreak: because usually the ssh user is `git` on most hosting servers, common to all user accounts. Identification is done via ssh keyprovided
12:46 cbreak phanimahesh: so?
12:47 phanimahesh So to auth as different users, they need different keys
12:47 cbreak you also have separate hosts
12:47 liuzhen joined #git
12:47 phanimahesh Well, I was mentioning the general case, considering the possibility of multiple users on same host
12:47 ahmedelgabri joined #git
12:47 docnvk joined #git
12:47 fracting joined #git
12:48 romerocesar joined #git
12:49 freimatz joined #git
12:50 fracting joined #git
12:51 heptadecagram joined #git
12:54 xaviergmail joined #git
12:54 davisonio joined #git
12:55 iLembus_ joined #git
12:57 BaconWings left #git
13:04 infra-red joined #git
13:04 heybaro joined #git
13:05 carell joined #git
13:05 darkfrog joined #git
13:05 italoacasas joined #git
13:05 ToBeCloud joined #git
13:06 darkfrog I want to create a 2.0 branch of my project, but I want to start the code over from scratch.  Is the only way to do that to branch from 1.0 and delete everything?
13:06 heybaro itself*
13:06 heybaro Hey there; If I create a symlink to something and then push it, what will be pushed, the soft link itself, or the linked file?
13:06 darkfrog heybaro: soft link
13:06 infra-red joined #git
13:06 eijk joined #git
13:07 harish joined #git
13:07 italoacasas joined #git
13:07 cbreak darkfrog: you can git checkout --orphan
13:07 heybaro Ah, woops, so what;s the solution! I want update my font folder on my computer as an upstream for my git repos
13:08 cbreak that'll create a new branch without history
13:08 heybaro what's*
13:08 cbreak you can make a new initial commit
13:08 darkfrog cbreak: awesome...thanks. :)
13:08 allcentury joined #git
13:09 cbreak darkfrog: your current state will be staged, but you can remove that of course
13:09 d0nkeyBOB joined #git
13:09 playfullyExist left #git
13:11 subhojit777 joined #git
13:11 gfixler joined #git
13:12 heybaro left #git
13:14 Junior joined #git
13:15 acetakwas joined #git
13:16 TooLmaN joined #git
13:16 nick123 joined #git
13:20 phanimahesh heybaro: you want to push a copy instead of a symlink? Use a hardlink/reflink and git add everything in there
13:21 happy-dude joined #git
13:21 nonconvergent joined #git
13:21 feignix joined #git
13:22 fmcgeough joined #git
13:22 bongjovi joined #git
13:22 jimi_ joined #git
13:23 notebox joined #git
13:23 rnsanchez_ joined #git
13:24 douglascorrea joined #git
13:25 cdg joined #git
13:26 diegoviola joined #git
13:29 livingstn joined #git
13:30 linx joined #git
13:30 ash_workz joined #git
13:33 leehambley joined #git
13:35 jstimm joined #git
13:35 a_thakur joined #git
13:36 a_thaku__ joined #git
13:36 Khisanth joined #git
13:38 linx joined #git
13:38 andlabs joined #git
13:38 dunpeal joined #git
13:38 andlabs Hi. Is there a way I can exclude a file in the repo from git commit -a?
13:38 andlabs In this case it's a configuration file I don't want to push my local changes
13:38 andlabs accidentally
13:38 Eryn_1983_FL joined #git
13:39 _ikke_ Don't use commit -a then
13:39 docnvk joined #git
13:40 _ikke_ git -a means commit every change to a tracked file
13:40 _ikke_ (note that we usually recommend not to track these kind of config files, just because of these reasons)
13:40 gechr_ joined #git
13:42 BlaXpirit _ikke_, the alternative would be  tracking a  config.example  file?
13:42 andlabs not my repo, alas
13:42 _ikke_ BlaXpirit: correct
13:43 phanimahesh andlabs: just git add whatever you want to track.
13:43 _ikke_ andlabs: right, that's why I also recommend people not to use commit -a, because it makes people not looking at what they commit
13:43 _ngz_ngzz joined #git
13:43 _ikke_ (just like add -A, add ., etc)
13:44 _ikke_ It's a good habbit to inspect what you are about to commit
13:44 shinnya joined #git
13:44 andlabs of course, I always do that anyway
13:44 a_thakur joined #git
13:44 phanimahesh andlabs: or git add . to add everything and then unstage the config with git reset HEAD -- configfile
13:45 rubyonrailed joined #git
13:46 andlabs also I wonder why the log file is in the repo
13:46 Dougie187 joined #git
13:46 andlabs but that's probably saying too much in this point
13:46 cdown joined #git
13:47 djb-irc joined #git
13:47 zivester joined #git
13:47 phanimahesh Logfiles in the repo? -_-
13:47 GodGinrai joined #git
13:48 a_thakur_ joined #git
13:48 romerocesar joined #git
13:49 GodGinrai Would running `git remote update` to update from upstream using a different ssh key (and different user) than the one who cloned the repo cause any problems in a git repo?
13:50 f3r70rr35f joined #git
13:50 linx joined #git
13:50 minjaeKim joined #git
13:53 JeremyM_ joined #git
13:55 a_thakur joined #git
13:55 robotroll joined #git
13:55 f3r70rr35f joined #git
13:56 phaleth joined #git
13:58 weox joined #git
13:58 D-Boy joined #git
13:59 ejb joined #git
13:59 raindev joined #git
14:01 vuoto joined #git
14:02 a_thakur_ joined #git
14:03 leeN joined #git
14:09 gechr_ joined #git
14:14 Tim-SEM joined #git
14:15 CussBot joined #git
14:15 UTAN_dev joined #git
14:16 Lunatrius joined #git
14:16 tgunr joined #git
14:16 davisonio joined #git
14:18 Rhonda joined #git
14:19 tgunr joined #git
14:20 mtcrutch joined #git
14:20 sgrover joined #git
14:20 ochorocho__ joined #git
14:20 cevrial joined #git
14:21 nonconvergent joined #git
14:22 tvw joined #git
14:22 Tim-SEM joined #git
14:22 Tim-SEM left #git
14:23 caudill joined #git
14:23 docnvk joined #git
14:25 feignix joined #git
14:25 settermjd joined #git
14:26 ShekharReddy joined #git
14:28 cevrial joined #git
14:30 nonconvergent joined #git
14:31 mmmveggies joined #git
14:31 livingstn joined #git
14:32 cevrial joined #git
14:33 aarobc joined #git
14:33 MrMojito2 joined #git
14:34 ekinmur_ joined #git
14:36 romerocesar joined #git
14:37 nonconvergent joined #git
14:37 lindenle joined #git
14:39 dunpeal joined #git
14:41 TomyLobo joined #git
14:41 perlpilot joined #git
14:42 dopesong joined #git
14:42 jaguarmagenta joined #git
14:42 Polo_ joined #git
14:43 dopesong joined #git
14:43 refried_ joined #git
14:43 eukn joined #git
14:43 dopesong_ joined #git
14:43 eijk_ joined #git
14:44 NeverDie joined #git
14:45 jimi_sanchez joined #git
14:45 a_thakur joined #git
14:46 Ajac joined #git
14:47 romerocesar joined #git
14:48 raindev joined #git
14:49 Ajac I'm having an issue, we're just using the git kraken push and pull commands, when we make changes and we seem to be losing files we want to keep sometimes without it even being indicated in the shown in the commit history that the files were removed
14:51 raijin joined #git
14:52 clemf joined #git
14:53 GodGinrai what is git kraken?
14:53 sangy joined #git
14:54 Ajac it's a git gui
14:54 GodGinrai ah
14:55 swalladge joined #git
14:55 nonconvergent joined #git
14:56 swalladge hi, i'm wanting to backup some git repos - what's the advantages/disadvantages of 'git clone <repo>' vs 'git clone --bare <repo>' ?
14:56 infra-red joined #git
14:56 fracting joined #git
14:56 mingrammer_ joined #git
14:57 swalladge (or something else altogether?) - basically i want to backup some repos in such a way that i can cd into one of them and start work straight away
14:57 Dougie187 `--bare` makes the repo not have an actual checkout
14:57 Dougie187 it's more useful for a server than something you want to develop in
14:57 Dougie187 `git clone <repo>` has a checkout.
14:57 Sonicbit joined #git
14:58 BlaXpirit bare is just the  .git  folder, right?
14:58 swalladge Dougie187: ah ok. so either way will result in all branches, tags, refs, etc?
14:59 Dougie187 swalladge: yes.
14:59 Dougie187 BlaXpirit: Yeah, but it's the .git folder where you would expect a checkout instead.
15:00 swalladge cool, thanks :)
15:00 nnyk_ joined #git
15:01 andlabs left #git
15:01 ResidentBiscuit joined #git
15:02 crose joined #git
15:05 diegoaguilar joined #git
15:05 rominronin joined #git
15:06 sgfgdf left #git
15:07 Spacew00t joined #git
15:09 diegoaguilar joined #git
15:10 irco joined #git
15:10 gechr_ joined #git
15:10 diegoaguilar joined #git
15:11 ahmedelgabri joined #git
15:13 feignix joined #git
15:13 cevrial joined #git
15:14 fracting joined #git
15:15 feignix joined #git
15:15 livingstn joined #git
15:17 nick123 joined #git
15:19 cdown_ joined #git
15:20 mdw joined #git
15:20 JeremyM_ joined #git
15:22 vuoto joined #git
15:23 dmto joined #git
15:23 choki joined #git
15:23 madewokherd joined #git
15:23 taylorm joined #git
15:23 aard_ joined #git
15:25 editshare-ericf joined #git
15:26 _h264 left #git
15:26 h264 joined #git
15:28 jetpack_joe joined #git
15:28 thiago joined #git
15:29 raijin joined #git
15:32 heptadecagram joined #git
15:34 dopesong joined #git
15:35 dopesong_ joined #git
15:38 playfullyExist joined #git
15:38 PCatinean joined #git
15:38 Lvjasha joined #git
15:38 phanimahesh joined #git
15:39 aep-shoutlet joined #git
15:39 Meteorhead joined #git
15:39 dunpeal joined #git
15:40 aep-shoutlet it is possible to git pull an arbitrary remote branch into an arbitrary active local branch? i realize the more typical usage is fetching/pulling then merging, but i'm curious if git pull could also effectively do this in a pinch
15:41 Eugene aep-shoutlet - sure; !fetch4 will do that
15:41 gitinfo aep-shoutlet: [!fetchfour] [pre 1.8.4 only] We recommend against using 'git fetch/pull <remote> <refspec>' (i.e. with branch argument), because it doesn't update the <remote>/<branch> ref. The easy way to fetch things properly is to get everything: 'git fetch' or 'git pull' are sufficient if you have one remote; otherwise we recommend 'git fetch <remote>' (plus 'git merge <remote>/<branch>' if you wanted to pull/merge).
15:41 Phylock joined #git
15:41 Eugene (though it isn't a normal way to do things)
15:42 Eugene pull will take the arguments just the same as fetch
15:42 durham joined #git
15:42 aep-shoutlet what does it mean that git-pull would no update the remote branch ref, compared to git-fetch, git-merge?
15:42 aep-shoutlet *not
15:43 axiom_1 joined #git
15:43 jetpack_joe joined #git
15:44 raininja joined #git
15:44 ToxicFrog aep-shoutlet: if you have a remote tracking branch "origin/master" and a local branch "master", both of which are (say) 4 commits behind the state on origin itself:
15:44 ToxicFrog - "git fetch" will upgrade origin/master and leave master alone;
15:45 ToxicFrog - "git pull" or "git pull origin" will update origin/master and fast forward master to match;
15:45 ToxicFrog - "git pull origin master" will fast forward master to match the state on origin without touching origin/master
15:45 ams__ joined #git
15:46 ceze joined #git
15:48 lindenle joined #git
15:49 KnightsOfNi joined #git
15:49 troulouliou_div2 joined #git
15:50 bango2 joined #git
15:50 dopesong joined #git
15:52 cevrial joined #git
15:52 dopesong_ joined #git
15:52 perlpilot joined #git
15:52 jjasinski joined #git
15:52 MrMojito joined #git
15:53 AaronMT joined #git
15:54 AaronMT joined #git
15:54 kpease joined #git
15:55 sgen joined #git
15:57 dingus joined #git
15:58 concerti joined #git
15:59 LionsMane joined #git
16:00 dsdeiz joined #git
16:00 TomyLobo has anyone here had any success in making git-lfs not redownload everything with "git worktree add" or anything in that direction?
16:00 TomyLobo basically what i need is a plain copy of one particular commit in one particular directory
16:01 TomyLobo be it an alternate checkout, a worktree, an export, whatever
16:01 refried_ joined #git
16:02 fornax joined #git
16:02 SegunOyebode joined #git
16:03 rkazak joined #git
16:05 CheckDavid joined #git
16:05 cydrobolt joined #git
16:05 cydrobolt joined #git
16:06 TomyLobo oooooh just had an idea and it seems to be working
16:06 rominronin joined #git
16:06 TomyLobo git archive | tar x
16:07 skalpin If I accidentally committed into the master branch instead of a topic branch. Can I create the topic branch, and then rebase the first commit onto the topic branch?
16:07 SwiftMatt joined #git
16:07 TomyLobo skalpin, did you push yet?
16:07 skalpin TomyLobo: no
16:07 TomyLobo good
16:07 _ikke_ No rebase needed even
16:08 TomyLobo no assumptions please
16:08 TomyLobo did the topic branch equal master before that commit?
16:08 skalpin I haven't made a topic branch yet
16:08 TomyLobo ah, even better
16:08 TomyLobo git checkout -b yourtopicbranch
16:08 TomyLobo git branch -f master HEAD~1
16:09 skalpin that will only go back 1 commit correct? so if I need to go back 3 I would do git branch -f master HEAD~3?
16:09 EduardoMartins joined #git
16:09 romerocesar joined #git
16:09 TomyLobo basically
16:09 TomyLobo but at that point, use gitk
16:09 TomyLobo gitk --all&
16:10 TomyLobo then right-click the commit you want master to be at, click "create new branch" and enter "master"
16:10 TomyLobo click yes when it asks you whether to overwrite
16:11 skalpin ah, perfect! thanks!
16:11 TomyLobo if gitk tells you to set DISPLAY, direct your hatemail at #cygwin
16:11 skalpin hahaha
16:11 skalpin no, gitk works fine
16:12 courrier_ joined #git
16:14 rath joined #git
16:15 durham joined #git
16:16 User458764 joined #git
16:17 tomog999 joined #git
16:21 lindenle joined #git
16:21 Tryop88 joined #git
16:21 NaN joined #git
16:21 phanimahesh joined #git
16:22 flaguy48 left #git
16:22 MistahKurtz joined #git
16:23 romerocesar_ joined #git
16:23 mda1_ joined #git
16:24 mingrammer_ joined #git
16:24 Tryop88 joined #git
16:26 fscala joined #git
16:28 samson joined #git
16:29 doritos joined #git
16:29 Tryop88 joined #git
16:32 zeroed joined #git
16:32 zeroed joined #git
16:33 netcarver joined #git
16:33 romerocesar joined #git
16:35 lindenle joined #git
16:37 ToBeCloud joined #git
16:37 NeverDie joined #git
16:39 jast joined #git
16:39 snips joined #git
16:40 dunpeal joined #git
16:40 eletuchy joined #git
16:41 dopesong joined #git
16:41 sanketdg joined #git
16:41 TomyLobo git-lfs is NOT cygwin-compatible -_-
16:42 TomyLobo that piece of crap wrote into C:\home\$USER instead of /home/$USER
16:42 steven- joined #git
16:42 jaguarmagenta joined #git
16:43 TomyLobo let's see if there's git-lfs in the cygwin repos
16:44 TomyLobo negative
16:45 feignix joined #git
16:46 feignix joined #git
16:46 skalpin git-lfs is in chocolatey
16:46 subhojit777 joined #git
16:47 Repox joined #git
16:49 ochorocho__1 joined #git
16:50 LeBlaaanc joined #git
16:53 TomyLobo probably not a cygwin version
16:53 attente joined #git
16:53 TomyLobo https://github.com/github/git-lfs/issues/1326
16:54 TomyLobo i reported the bug
16:55 samson joined #git
16:56 EvilPenguin joined #git
16:57 ochorocho__ joined #git
16:58 sctskw joined #git
17:03 NeverDie joined #git
17:04 rahtgaz joined #git
17:04 Lvjasha joined #git
17:05 dunpeal joined #git
17:05 m0viefreak joined #git
17:05 nnyk joined #git
17:05 rominronin joined #git
17:07 stuh84 joined #git
17:07 romerocesar joined #git
17:07 durham_ joined #git
17:08 ochorocho__ joined #git
17:08 a_thakur joined #git
17:09 pwnz0r joined #git
17:09 skalpin maybe the new windows 10 bash can do it
17:10 jeffreylevesque_ joined #git
17:10 mase-tech joined #git
17:11 CJKinni joined #git
17:11 inflames joined #git
17:11 SwiftMatt joined #git
17:12 xall joined #git
17:13 feignix joined #git
17:14 ramsub07 joined #git
17:14 User458764 joined #git
17:14 ochorocho__ joined #git
17:15 dopesong_ joined #git
17:15 nonconvergent joined #git
17:17 fracting joined #git
17:19 nick123 joined #git
17:19 thiago joined #git
17:20 fornax joined #git
17:22 xaviergmail joined #git
17:23 CJKinni` joined #git
17:25 ToBeFree joined #git
17:25 geekbri joined #git
17:26 romerocesar joined #git
17:28 MacWinner joined #git
17:28 tyreld joined #git
17:29 GodGinrai Would running `git remote update` to update from upstream using a different ssh key (and different user) than the one who cloned the repo cause any problems in a git repo?
17:30 chilversc joined #git
17:30 kasper93_ joined #git
17:31 stonerfish joined #git
17:31 skylite joined #git
17:32 grassass joined #git
17:33 sbeller_ joined #git
17:33 PaulCapestany joined #git
17:33 mns I have my master and a build-ci branch both of which have a .travis.yml file.  I dont ever want to merge master/.travis.yml to build-ci/.travis.yml.  I want git to ignore those changes.  Is there a way to do that ?
17:33 ochorocho__ joined #git
17:35 GodGinrai mns: this sounds relevant: https://medium.com/@porteneuve/how-to-make-git-pr​eserve-specific-files-while-merging-18c92343826b#.dkh6c8vit
17:35 eletuchy joined #git
17:35 mns GodGinrai: thanks I'll look at that.
17:35 GodGinrai also these: https://git-scm.com/book/en/v​2/Git-Tools-Advanced-Merging https://git-scm.com/book/en/v2/​Customizing-Git-Git-Attributes
17:36 SmashingX joined #git
17:36 WizJin joined #git
17:38 diegoaguilar joined #git
17:38 grindhold joined #git
17:39 ochorocho__ joined #git
17:39 liuzhen joined #git
17:41 gabi_ joined #git
17:41 dreiss joined #git
17:42 SmashingX joined #git
17:44 mehola joined #git
17:45 elastix joined #git
17:46 a_thakur joined #git
17:48 Literphor joined #git
17:49 cdown joined #git
17:50 kpease joined #git
17:50 ochorocho__ joined #git
17:52 jh3 joined #git
17:52 d10n-work joined #git
17:56 SwiftMatt joined #git
17:56 astrofog joined #git
17:57 j416 GodGinrai: ssh key does not matter. User only matters when it comes to permissions on the files in the repo.
17:57 jstein___ joined #git
17:57 raijin joined #git
17:57 vacho joined #git
17:57 vacho in github, is the deployment key the public key?
17:57 romerocesar joined #git
17:58 thiago what's a deployment key?
17:58 thiago Git is not a deployment tool
17:58 thiago you may want to check #github
17:58 j416 vacho: always send your public key, never share your private key.
17:58 GodGinrai j416: so if I am only doing a remote update, then it should be fine?
17:58 vacho j416: got it..thanks.
17:58 j416 GodGinrai: if you user has the correct permissions, yes
17:59 vacho I am trying to pull from my git repo from amazon ec2, not sure how I would get my private key onto my amazon ec2 server
17:59 j416 vacho: use ssh agent forwarding
17:59 acetakwas joined #git
17:59 j416 vacho: -A to ssh.
17:59 SmashingX_ joined #git
17:59 vacho j416: please tell me more
17:59 j416 vacho: man ssh
17:59 GodGinrai j416: has the correct permissions meaning?  Do you mean if the user has read access to the repo?
17:59 j416 GodGinrai: read and write
18:00 GodGinrai j416: for the local files?
18:00 nd joined #git
18:00 j416 GodGinrai: yes
18:00 blackwind_123 joined #git
18:00 tom___ joined #git
18:00 Sasazuka joined #git
18:01 GodGinrai j416: This is because it would need to update files in the .git folder?
18:01 j416 GodGinrai: yes
18:02 MickyMicks joined #git
18:02 tom___ has anyone recovered cleared stashes? I've tried git fsck and then using git show to check the latest unreachable commits, but it wasn't there.  It may be because the stash I cleared had no committed work
18:02 narendraj9 joined #git
18:03 j416 tom___: a stash is a commit behind the covers
18:03 thiago tom___: did you fsck find unreferenced commit objects?
18:03 tom___ yep, i found the unreferenced/unreachable commits
18:03 GodGinrai j416: thanks
18:04 j416 welcome
18:04 dmto joined #git
18:04 raininja joined #git
18:04 j416 tom___: look for "dangling commit "
18:04 ochorocho__ joined #git
18:04 tom___ the ones listed at the top should be the most recent, correct?
18:04 thiago tom___: inspect each and everyone of them. One of them might be your cleared stash.
18:04 ybden joined #git
18:04 thiago tom___: note that stash commits are often merges, so you may not see the change immediately on that commit
18:05 j416 thiago: not always?
18:05 tom___ i've inspected the top few with git show, but after the first few, i know i'm going way too far back
18:05 thiago j416: I think always, but I don't know for sure
18:05 j416 ok :)
18:05 j416 ty
18:05 tom___ ok i'll give it another go
18:06 thiago tom___: your changes are usually in the second parent of that merge commit
18:06 rominronin joined #git
18:06 EvilPenguin joined #git
18:06 j416 tom___: commit message of the commit will be "WIP on <branchname>"
18:06 j416 tom___: (with some following text)
18:06 timthowtdi_ left #git
18:06 tom___ alright, thanks for the info
18:06 timthowtdi joined #git
18:07 GodGinrai j416: I'm just going to float this by you... Is there some way to tell if a branch is up-to-date with the remote branch without having to update the repo?  The fact that I don't know how to do this is why I am looking at using `git remote update`
18:07 j416 GodGinrai: git status
18:07 j416 GodGinrai: after git fetch, yes
18:07 j416 GodGinrai: so, not reallty
18:07 j416 really*
18:07 maestrojed joined #git
18:07 j416 GodGinrai: what is the issue with running fetch?
18:07 tom___ once i do git fsk, is git show the correct command to look at the commit?
18:07 j416 GodGinrai: (remove update is fetch)
18:08 j416 tom___: git show will work.
18:08 thiago GodGinrai: let's try this
18:08 thiago GodGinrai: if I tell you my branch is at b31cdb3582b0a32c2e09d151b4370f27990efdb4, can you tell me whether it's up-to-date or not?
18:09 thiago I always remember advice from a teacher in elementary school, "when comparing, you always compare at least two things"
18:09 GodGinrai thiago: I knew that, I more meant a way to check in-memory
18:10 thiago I can check whether it's up to date relative to the last time I got the commit from upstream
18:10 GodGinrai thiago: like, pull the data of the remote in memory, and compare
18:10 thiago that doesn't mean that upstream hasn't changed
18:10 thiago is that sufficient for you?
18:10 j416 GodGinrai: why can you not run fetch?
18:10 thiago oh, you do want to contact the remote, you just don't want the local refs to be updated because of that
18:10 thiago is that it?
18:11 j416 thiago: seems so
18:11 GodGinrai j416: Basically, I am trying to prevent the user from performing certain actions in a docker container (that mounts their git repo) if their master is not up to date.
18:11 GodGinrai j416: I *can* run the fetch, but I am just paranoid and want to take the path of least resistance (or least chance of breaking things)
18:12 j416 GodGinrai: and if their connection is down, you will prevent them from working?
18:12 thiago why would fetch break things?
18:12 thiago what are you afraid of?
18:12 thiago what you're asking for is possible, just not easy
18:12 GodGinrai thiago: the user in the docker is not the same user as the host system, I don't want to cause problems like the "whoops I saved a file as root" problem
18:12 j416 there we go, a reason
18:12 fornax joined #git
18:13 thiago GodGinrai: a) why would you run anything as root; b) isn't the point of using a container so that the host system can't be screwed up?
18:13 j416 GodGinrai: seems it is not the container's responsibility to do this at all.
18:13 j416 GodGinrai: then again I don't know the issue you are solving.
18:14 tom___ @j416 thanks for the help, I found the commit, luckily it was about 1/4 the  way down a list of about 100... Also found out that the most recent is not listed at the top
18:14 GodGinrai j416: I'm using docker to contain our build system so that there aren't any weird "wfm" situations.
18:14 GodGinrai among devs
18:14 j416 GodGinrai: and what if you need to build an older version?
18:14 j416 tom___: cool
18:14 GodGinrai j416: checkout the older version in git, and then spin up the container
18:15 rbr joined #git
18:15 j416 GodGinrai: but that would compare it to the remote, and it would fail because it's not up-to-date with remote master.
18:15 j416 GodGinrai: I don't see the point in checking with the remote.
18:15 j416 GodGinrai: fetch the latest things, spin up, build.
18:16 j416 GodGinrai: no check needed
18:16 thiago and I don't see how checking the remote will help with preventing stuff as root or doing other stupid things
18:16 GodGinrai j416: no, because the build is not what I'm disallowing.  We have a task that pushes the built code to a test server.  Since we only have one test server, we don't want people pushing old code to it.
18:16 j416 GodGinrai: you never want to test an old version? ever?
18:16 thiago there's a race condition in there
18:16 GodGinrai j416: If I wanted to, I could easily make a flag that allows for that use case.
18:16 thiago what if you start the build process, then the remote updates?
18:17 thiago and if you always want to build the latest, why is there even an option to tell the test system to build something else?
18:17 j416 GodGinrai: it seems your problem is better solved by actually talking to your team
18:17 thiago make the test system check out the latest
18:18 GodGinrai thiago: that's a good point, but master only updates when we accept PRs, so devs should be able to time that correctly.
18:18 GodGinrai thiago: I'm all for making our CI do that.  But it doesn't currently, and I'm not on the team that does that.
18:18 j416 GodGinrai: have the test server fetch continuously then?
18:18 thiago so you're fixing the horn because the horn doesn't work
18:18 thiago er, fixing the horn because the brakes don't work
18:18 thiago fix the brakes
18:19 jstimm joined #git
18:19 j416 GodGinrai: sorry for not helping out much in your workaround, but it seems like a waste of time on your part
18:19 GodGinrai thiago: I'm fixing what I can :P
18:19 j416 GodGinrai: why not do it right instead?
18:19 thiago GodGinrai: fix the right problem
18:20 j416 GodGinrai: you're clearly doing CI stuff, and as you said, you're not on that team :)
18:20 thiago GodGinrai: you don't have a problem with "uptodateness"
18:20 j416 GodGinrai: so either do it properly anyway, or don't :P
18:20 lindenle joined #git
18:20 thiago anyway, if you REALLY want to do what you're asking, do 'git fetch <url-of-the-remote> <upstream-branch>"
18:20 thiago you cannot use the remote name. You have to use the URL.
18:21 j416 thiago: that won't write objects? o_O
18:21 thiago that will download the latest upstream to FETCH_HEAD. Then compare FETCH_HEAD to HEAD: if they are different, it's not up-to-date.
18:21 thiago of course it will
18:21 ShekharReddy joined #git
18:21 j416 I thought that was his whole problem
18:21 thiago ok, then git ls-remote remotename branchname
18:21 j416 he doesn't want to write things to disk
18:21 GodGinrai thiago: Yea, the problem is devs not paying attention to what they are doing.  But I figure making the process tell them "don't do that! BAD!" would help
18:21 thiago there are some options to print only the SHA-1
18:21 dranve joined #git
18:22 thiago I thought you may want to actually download so because not-up-to-date could mean behind or after. Eventually you'll want to know that.
18:22 j416 GodGinrai: if your devs cannot understand what you just said here, it seems you have bigger problems :)
18:22 thiago anyway, fix the CI itself
18:22 BSaboia joined #git
18:23 GodGinrai j416: we do. <.<
18:23 thiago if you want to allow testing only the latest, then don't allow selecting anything but the latest for testing
18:23 Literphor joined #git
18:23 thiago solve the right problem, at the right time, in the right place (with the correct solution)
18:23 j416 GodGinrai: if you only test when merging a PR, why don't you have one person dedicated to merging and testing?
18:23 thiago knowingly doing anything other than that is just patchwork and will come to bite you in the rear later
18:24 j416 GodGinrai: when you have a proper CI, perhaps you can allow more people in.
18:24 GodGinrai j416: Well, testing goes on concurrently, which is why people pushing up old masters causes problems
18:25 j416 concurrently?
18:25 j416 I thought you had only one test server?
18:25 GodGinrai yes
18:25 GodGinrai all QA tests on that one server
18:25 elastix1 joined #git
18:25 j416 I see, so you run multiple tests in parallell on limited hardware
18:25 GodGinrai yup
18:25 j416 I see
18:25 mkoskar joined #git
18:25 rsukla joined #git
18:26 thiago well, that's what containers are for
18:26 thiago I remember when we did that on bare hardware at Trolltech
18:26 thiago you couldn't touch the machine while it was running the tests. Not even moving the mouse or pressing the space bar.
18:27 j416 haha I remember that
18:27 GodGinrai thiago: We use a proprietary CMS/javaserver solution that we get from a vendor.  We've been bothering them for docker containers, but they've basically just ignored us.
18:27 j416 we had the same thing
18:27 thiago GodGinrai: choose a better solution
18:27 ahmedelgabri joined #git
18:27 thiago would that be Pulse, btw?
18:27 j416 GodGinrai: until better solution arrives, just don't run the tests all the time on ti?
18:27 j416 it*
18:27 GodGinrai thiago: That's a decision I am not in a position to make
18:28 Orbitrix joined #git
18:28 eijk_ joined #git
18:28 GodGinrai j416: we can't.  Office politics
18:28 GodGinrai and no, not pulse
18:28 j416 GodGinrai: then escalate
18:28 j416 GodGinrai: buy another server, etc.
18:29 thiago you're really in the horn-instead-of-brakes case
18:29 FuzzySockets joined #git
18:29 thiago I'm sorry for you
18:29 thiago good luck, though
18:29 GodGinrai thiago: thanks
18:30 cdg joined #git
18:30 bamb joined #git
18:31 bamb left #git
18:31 neilthereildeil joined #git
18:31 eletuchy joined #git
18:31 neilthereildeil hi i am trying to clone a certain commit of my git tree
18:31 ash_workz joined #git
18:31 neilthereildeil i gopied .git to a new dir, and ran git log
18:31 j416 neilthereildeil: you clone a repo, not a commit
18:32 neilthereildeil then i did "git checkout <OLDEST_HASH>"
18:32 neilthereildeil what will this do for me?
18:32 ynanm joined #git
18:32 j416 neilthereildeil: it'll update your work tree to reflect the state at that hash
18:32 thiago neilthereildeil: sorry, didn't understand. Let's go one step back.
18:32 j416 neilthereildeil: what do you want to do?
18:32 thiago neilthereildeil: what are you trying to accomplish?
18:32 neilthereildeil i want to clone my tree at a certain point in time
18:32 j416 step back one more
18:32 thiago neilthereildeil: don't say clone
18:33 linx joined #git
18:33 neilthereildeil oops i mean copy
18:33 GodGinrai j416: thiago: a much simpler solution for me might be to just check if the last pull was recent enough.  Is there a way to check the last time the remote was updated?
18:33 thiago neilthereildeil: what are you trying to do with that tree at that point in time?
18:33 raijin joined #git
18:33 anddam GodGinrai: I'm not upstream, I'm downstream
18:33 neilthereildeil move it to another machine to build it
18:33 anddam j416: I see, thanks
18:33 j416 GodGinrai: not that is sane, I think.
18:33 neilthereildeil i dont want any of the other changes available there
18:34 thiago GodGinrai: if your repository has logallrefupdates turned on, then you can check the reflog of that upstream commit.
18:34 GodGinrai anddam: I'm not talking about your problem
18:34 thiago neilthereildeil: ok, you want git export
18:34 thiago neilthereildeil: sorry, git archive
18:34 neilthereildeil thiago: ok. copying .git and running "git checkout OLDEST_HASH" wont do it?
18:34 thiago neilthereildeil: no moving of .git, direct stdout stream of the files you want
18:34 rahtgaz joined #git
18:35 thiago neilthereildeil: it will, but it will also create problems by updating the index. Git archive is better.
18:35 GodGinrai j416: I had one dev asking me to force a pull before this code was pushed to the test servers, so checking that they have at least checked recently doesn't seem that far out there.
18:35 neilthereildeil when i do checkout, it doesnt give me the whle tree
18:35 neilthereildeil is it just populating files from that commit?
18:35 thiago neilthereildeil: because the index isn't matching
18:36 thiago neilthereildeil: try my solution
18:36 neilthereildeil ahh ok
18:36 j416 GodGinrai: horn and brakes
18:36 j416 GodGinrai: I wouldn't try to solve it using git at all
18:36 neilthereildeil what does it mean that index doesnt match?
18:36 thiago neilthereildeil: the "index" is a local cache of what has been checked out and what will be in the next commit
18:37 thiago neilthereildeil: by moving the .git dir elsewhere, the index and the worktree won't match
18:37 thiago neilthereildeil: git checkout will refuse to work because it thinks it's overwriting your changed files.
18:37 thiago neilthereildeil: they *are* changed as far as git is concerned: you deleted them all.
18:37 neilthereildeil ohh ok
18:37 neilthereildeil yea git status shows EVERYTHING deleted
18:37 thiago yeah
18:38 thiago now, please apply the Engineer's Rule: the right tool for the right job
18:38 thiago you want git archive, not git checkout
18:38 neilthereildeil thiago: ok, what should i do to git archive that one snapshot without changing the source tree at all?
18:38 xall joined #git
18:38 thiago neilthereildeil: git archive never changes the source tree, so that is safe
18:38 thiago neilthereildeil: for the rest, man git-archive
18:38 gitinfo neilthereildeil: the git-archive manpage is available at http://jk.gs/git-archive.html
18:39 thiago neilthereildeil: you said you want to send it to another machine. How are you going to transport the tree there?
18:39 j416 neilthereildeil: I'd go one step further back, it sounds like an xy problem
18:39 neilthereildeil ill use sftp
18:39 thiago neilthereildeil: ok, with git archive you can just use ssh and pipe through, telling tar to extract on the other side
18:40 neilthereildeil no, i need to test it locally first
18:40 neilthereildeil like if there are any buiild issues etc
18:40 thiago neilthereildeil: but can't the other machine simply do a git fetch && git checkout of the commit you wanted checked out?
18:40 j416 neilthereildeil: also, !deploy
18:40 gitinfo neilthereildeil: 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
18:40 thiago git archive <commit> | tar -xC /other/dir will do it for you
18:41 thiago but I would recommend simply having another clone at the place you want to test
18:41 Literphor joined #git
18:41 thiago you can also use git new-workdir and keep it always in a detached HEAD commit.
18:41 thiago but no, actually archive is a better idae, for two reasons:
18:42 thiago 1) it ensures a *completely* clean tree, even with no .git dir
18:42 arooni joined #git
18:42 thiago 2) you'll use git archive to make your releases, so that's what your users will be seeing. That's what you should be testing.
18:42 cdown joined #git
18:42 nedbat neilthereildeil: if you don't mind me asking, why don't you want the other changes on the other machine?
18:43 neilthereildeil theres other unrelated work that went on in the newer commits
18:43 dopesong joined #git
18:43 dmto joined #git
18:43 jaguarmagenta joined #git
18:43 nedbat neilthereildeil: sure, you want the working tree to reflect the desired commit.  But you're ok with the repo being there with history?
18:43 linx joined #git
18:44 neilthereildeil no, that is unrelated
18:44 thiago I'd say there's nothing wrong with those commits being there, but there's also nothing right with that.
18:45 thiago the test machine doesn't need them
18:45 nedbat neilthereildeil: i'm just wondering what's wrong with having a clone there, and updating to the commit you want?
18:45 thiago I still recommend using the git archive route when testing.
18:45 thiago actually, more than that: use the whole release process to test
18:46 thiago if you don't test what your users will use, then you may never catch issues that they have and you don't.
18:46 neilthereildeil thiago: and by release process you mean git archive?
18:46 thiago I do recommend using git archive as your release process
18:46 thiago but you don't have to. Nor does it mean that it's the only step.
18:47 thiago if possible, you should make your release process simply tagging the release and running git archive with the proper parameters.
18:47 nedbat neilthereildeil: i think what thiago is getting at is this:  if you already have a way that you ship software when it's ready, then use that same process to get it on the test machine.
18:48 ejb joined #git
18:48 thiago yup
18:48 thiago but consider using git archive
18:49 thiago if you have no process established yet, use git archive
18:49 EvilPeng1 joined #git
18:49 diogenese joined #git
18:50 neilthereildeil why is git archive better than git clone?
18:51 nedbat neilthereildeil: can you give us some more background? What kind of software is this, and how does it get to the real users when it is ready?
18:52 fus joined #git
18:52 likewhoa joined #git
18:52 neilthereildeil we deploy it internally, so we build and install bins
18:53 jyjon joined #git
18:54 lindenle joined #git
18:54 neilthereildeil thiago: why is git archive better than git clone?
18:55 fus joined #git
18:56 TomyLobo it isnt, they serve different purposes. git clone gives you a local git repository you can commit to and push from, whereas git archive gives you a .tar
18:56 nedbat neilthereildeil: why wouldn't you install bins on the test machine?
18:56 GodGinrai neilthereildeil: git archive gives you a file tht can easily be passed around
18:56 nedbat neilthereildeil: git archive is just a snapshot of the current revision with no history or connection to git. it's a tarball of "now"
18:56 nedbat (or some other point in time)
18:56 neilthereildeil ohh
18:57 thiago neilthereildeil: I think I explained
18:59 neilthereildeil also it wouldnt be |tar -xC, would it?
18:59 neilthereildeil -x is extract
18:59 neilthereildeil i am writing your guys' advice in notes
18:59 neilthereildeil i think its just tar -C, right?
19:00 dopesong joined #git
19:02 linx joined #git
19:02 ki0_ joined #git
19:02 mdw joined #git
19:03 neilthereildeil that didnt create a tar file for me
19:03 neilthereildeil i just ran -xC
19:03 neilthereildeil it copied the files out
19:06 jason237 joined #git
19:06 eijk_ joined #git
19:07 SmashingX joined #git
19:07 rominronin joined #git
19:08 SmashingX joined #git
19:08 paul424 joined #git
19:08 SmashingX left #git
19:12 linx joined #git
19:12 SimonNa joined #git
19:13 mehola joined #git
19:17 durham joined #git
19:19 hashpuppy joined #git
19:19 mehola joined #git
19:19 carell joined #git
19:19 kpease joined #git
19:19 nick123 joined #git
19:20 dreiss joined #git
19:20 jwynn6 joined #git
19:20 SwiftMatt joined #git
19:26 weylin joined #git
19:27 raijin joined #git
19:29 texasmynsted I may have asked this before, but is there a way to pull all the commit messages into a merge commit when a merge commit is required rather than a ff ?
19:30 haris3301 joined #git
19:31 glauxosdever joined #git
19:31 glauxosdever Hi, I'm wondering how can I reliably remove objects that are not in current master, that is that are not in either HEAD and its parents.
19:32 glauxosdever Because I did a lot of forcable pushes, so the old objects are still there.
19:33 skylite joined #git
19:35 TomyLobo neilthereildeil, x is extract, C is change directory. C takes an additional parameter
19:35 ResidentBiscuit glauxosdever: Garbage collector will get them in due time
19:35 ResidentBiscuit You can also force the gc with some command... I can look it up
19:35 kugel joined #git
19:36 TomyLobo git gc --prune=now
19:36 ResidentBiscuit There we go
19:36 glauxosdever Yes, I looked it up, but not sure how to do it reliably.
19:36 ResidentBiscuit I've never had to use it
19:36 TomyLobo and --reflog=expire i think
19:36 glauxosdever Is it indeed that?
19:36 TomyLobo that'll get rid of your reflog too
19:36 TomyLobo which you usually dont want
19:36 ResidentBiscuit reflog is very helpful
19:36 TomyLobo which brings us back to ResidentBiscuit's original suggestion: wait
19:36 ResidentBiscuit Unless you have some weird situation, I'd just leave the garbage there
19:37 ResidentBiscuit It'll go away after awhile (30 days by default?)
19:37 fahadash joined #git
19:40 glauxosdever Ok, I ran this. Now, where are the HEAD and its parent objects?
19:41 glauxosdever (I wanted to keep those as they were.)
19:41 jwynn6 anyone ever seen a security audit of git anywhere?
19:41 dopesong joined #git
19:41 xorox90 joined #git
19:45 kugel joined #git
19:45 FuzzySockets joined #git
19:45 texasmynsted If I have a merge the require a commit, I can grab the commit message with -e, but I want to pull in the commit messages from the feature branch I am pulling in.  Is there no way to do that?
19:45 barteks2x joined #git
19:49 justinmrkva joined #git
19:51 fracting joined #git
19:53 TomyLobo jwynn6, audit of what specifically?
19:55 nnyk joined #git
19:56 jwynn6 TomyLobo, I'm looking to prove to my boss that there are no inherent security risks in installing GIT and using it. He wanted to see any code audits or security reviews if they exist.  My google-fu has failed me thus far.
19:56 jyjon joined #git
19:56 glauxosdever Ok, seems I'll better off rewriting it by hand (along with the times, so hashes don't change).
19:57 TomyLobo glauxosdever look at "gitk --all". everything you see there is referenced
19:57 glauxosdever Since I want to keep 5objects, but get of 80.
19:57 TomyLobo everything else, possibly minus a few fringe cases will be garbage-collected once the reflog expires
19:57 glauxosdever s/get/get rid/
19:58 TomyLobo jwynn6, it's probably better than that unpatched SVN server your company likely has :)
19:58 TomyLobo glauxosdever, why do you care?
19:59 TomyLobo and if you really care, just re-clone it
19:59 glauxosdever Because these commits don't exist in current master.
19:59 glauxosdever TomyLobo: These objects are cloned too.
19:59 TomyLobo so what?
19:59 TomyLobo so what if they dont exist in current master
19:59 TomyLobo they will be gone in a few weeks
20:00 jwynn6 TomyLobo, I wish we had even that. How about the unmonitored smb share?  does it trump that :)
20:00 TomyLobo why do you want to get rid of them now, potentially getting rid of a useful recovery tool in the process?
20:00 CussBot joined #git
20:00 TomyLobo jwynn6, ok here's the pitch:
20:00 glauxosdever I'll backup it manually.
20:01 hhee joined #git
20:01 TomyLobo jwynn6, Git employs cryptographical algorithms to ensure integrity of its repository structures. it is at least extremely hard to even change a commit message without someone noticing
20:01 TomyLobo if not impossible
20:02 TomyLobo he's concerned about security risks? that smb share is a business risk!
20:02 feignix joined #git
20:04 lindenle joined #git
20:05 TomyLobo but if he insists on using it, in a pinch, you can put your master repo onto it
20:05 TomyLobo git init --bare and go
20:05 jwynn6 yeah its making my eyes cross. ok thats a good bit to add to my pitch.   any good arguments to put out against hg that a drinking buddy showed him? you know, other than git is the way the industry has decided to go?
20:06 jknetl joined #git
20:06 TomyLobo hg wasn't designed by the creator of the linux kernel
20:06 TomyLobo hg lies about having a meaningful numbering scheme
20:06 TomyLobo no one uses hg
20:07 AndroidLoverInSF joined #git
20:07 jwynn6 yeah that bit about nobody using it is my primary. its virtually unsupported next to git.
20:07 kbs joined #git
20:07 primosdace joined #git
20:07 TomyLobo there is no mercurialhub, no mercuriallab, no gohgs
20:07 Arsonide joined #git
20:08 TomyLobo btw, for sanity and PR reasons, dont recommend cygwin git
20:08 primosdace hello guys may someone with a little bit exprience with github send me priv msg for help
20:08 rominronin joined #git
20:09 TomyLobo primosdace, try the whole of #github :)
20:09 jwynn6 i try to stay away from cygwin, cool tool but not for the faint of heart lol
20:09 primosdace ok thanks
20:09 abc123 joined #git
20:09 TomyLobo jwynn6, well i've noticed two major problems with cygwin git so far
20:09 primosdace left #git
20:09 TomyLobo 1. the gui tools wont work without a display server
20:09 TomyLobo 2. git-lfs stores stuff in the wrong directories
20:10 TomyLobo if you want to use git-lfs
20:10 jwynn6 no 2. seems like a kinda big oversight
20:10 TomyLobo it still works, just the object cache is in the wrong directory
20:10 TomyLobo it breaks down as soon as you use "git worktree" or anything like that
20:12 kbs Sorry for a several year-old question - could a kindly soul point me to a summary of why opinions changed from http://git.661346.n2.nabble.com/GPG-s​igning-for-git-commit-td2582986.html to "git commit -S" ?
20:12 SwiftMatt joined #git
20:14 jwynn6 TomyLobo, thanks!
20:14 b1tchcakes joined #git
20:15 TomyLobo jwynn6, gpg-signing is not what i meant by cryptographical algorithms, btw
20:15 TomyLobo that's on top of it
20:16 justinmrkva Regarding svn2git: I'm trying to split a large SVN repo with multiple projects. Converted Git repos contain just the contents of the SVN subdirectory specified (as I'd expect), but in the history, I see a handful of old commits from other unrelated projects from other directories. Is it common for this to happen? Is there any hope of fixing it reasonably quickly, or doth madness lie in wait for those foolish enough to try?
20:16 a_thakur joined #git
20:17 rbr joined #git
20:17 jwynn6 TomyLobo, i figured. I'm assuming you meant the way the node hashing is done.
20:17 vacho joined #git
20:17 TomyLobo yep
20:18 cyan__ joined #git
20:18 TomyLobo s/node/object/
20:18 vacho I am new to git and I am hoping someone can take a couple of minutes of their time and guide me.
20:18 TomyLobo justinmrkva, did you git svn clone the subdirectories or did you git svn clone everything and then move?
20:19 perlpilot vacho: if you can articulate a cohesive question, we will endeavor to answer it
20:19 vacho I forked a github project and did some work to it. I want to be able to fetch the updates made to the original repo if I need. This has been working, but I just realized my fork is public, and I cannot turn it into private.
20:19 TomyLobo !book
20:19 gitinfo 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
20:19 eletuchy joined #git
20:19 zumba_ad_ joined #git
20:19 GodGinrai vacho: sounds more like a github problem than a git problem
20:19 GodGinrai vacho: was the upstream private?
20:19 vishal vacho: yeah, !github
20:19 gitinfo vacho: Note that git != github. Feel free to ask us about Github-specific features (Forks, Pull Requests, Wikis, etc), but there are no guarantees. There is a #github channel, which might help too (again, no guarantees)
20:19 vacho GodGinrai: yes you are right
20:19 TomyLobo vacho, on github, anything private involves money
20:20 vacho TomyLobo: I don't mind spending. I just to be able to update my version with the changes done to original
20:20 perlpilot vacho: you can do private repos on bitbucket
20:20 vacho want*
20:20 GodGinrai vacho: if you pay for one of their paid options, you should be able to convert your public repos into private repos
20:21 vacho GodGinrai: I tried, it says you cannot turn a fork into private
20:21 GodGinrai vacho: so it is a fork of a public repo?
20:21 vacho GodGinrai: "Public forks can’t be made private. Please duplicate the repository."
20:21 vacho GodGinrai: so I went ahead and duplicated it. But now it's not based of the original..so I won't be getting the updates.
20:22 vacho GodGinrai: I hope I am making sense.
20:23 TomyLobo vacho, it is kinda bad style to benefit from a public project and not contribute back, anyway
20:23 TomyLobo or at least leave it open for others to pick up
20:23 vacho TomyLobo: well, in this case, I am using a Laravel Starter Kit to built my own private app. I can't share my private app with the world. :)
20:24 vacho TomyLobo: the point of this repo is for people to fork and built their own apps with.
20:24 GodGinrai vacho: sure, but why is your private app sitting in the starter kit's code base.
20:24 GodGinrai * ?
20:24 TomyLobo vacho, they're GPLv3, btw
20:24 TomyLobo https://github.com/sroutier/laravel-5.1-ent​erprise-starter-kit/blob/master/LICENSE.md
20:24 justinmrkva TomyLobo I cloned subdirectories using `svn2git https://domain/repo/projectFoo`. In the resulting Git repo, I found a bunch of commits from repo/projectBar. The thing is, directories in the SVN repo have moved around a few times, but projectFoo has never been in a directory with the same name as projectBar.
20:24 Arsonide joined #git
20:25 pwnz0r joined #git
20:25 TomyLobo justinmrkva, sorry, i've never used svn2git... i usually use git-svn
20:26 justinmrkva TomyLobo I think svn2git uses git-svn under the hood... I'll give git-svn a try directly and see if it's different.
20:26 mehola joined #git
20:26 codeman joined #git
20:26 GodGinrai vacho: TomyLobo's right.  You'd be violating gpl if you keep the code private
20:26 TomyLobo justinmrkva, are these commits empty, btw?
20:26 vacho GodGinrai: ok, maybe I got it wrong then. I forked it so I could fetch any bug fixes updated to the main repo
20:26 TomyLobo GodGinrai, only if you release binaries
20:26 nedbat vacho: which repo did you fork?
20:27 vacho https://github.com/sroutier/lar​avel-5.1-enterprise-starter-kit
20:27 TomyLobo GodGinrai, if you never release anything, you dont need to release the source with that nothing :)
20:27 GodGinrai TomyLobo: Oh, I guess if it is an in-house app, then this wouldn't violate gpl
20:27 TomyLobo my non-lawyer opinion
20:27 frem_ joined #git
20:27 mischat joined #git
20:27 justinmrkva Ah, good old GPL discussions...
20:28 TomyLobo including the "i aint a lawyer" prephase :D
20:28 GodGinrai TomyLobo: IANAL? :P
20:28 BlaXpirit good that you got this into the right direction. GodGinrai, TomyLobo, you're correct
20:28 nedbat vacho: who will be using your application?
20:28 TomyLobo GodGinrai, where's the N?
20:29 TomyLobo oh, am not
20:29 vacho nedbat: me and a few other internal people
20:29 TomyLobo i wouldnt phrase it that way
20:29 GodGinrai TomyLobo: that's the common acronym for it online
20:29 vacho " combining a set of features that can kick start any Web application for the Internet or on an Intranet"
20:29 liuzhen joined #git
20:29 vacho that's the whole point of this project, I am sure people can use it
20:29 vacho without sharing their entire app with the world.
20:29 nedbat whatever sroutier intended with that license, it's silly to use GPL on a web app anyway.
20:30 GodGinrai vacho: of course they can use it.  Just not for non-open-source projects that have releases
20:30 TomyLobo GodGinrai, and totally misses its purpose as a disclaimer if the other person doesnt know what it even means
20:30 nedbat vacho: you're copying a template. I wouldn't try to keep up as it changes anyway: you'll be making your own changes which will be very difficult to merge automatically.
20:31 justinmrkva TomyLobo No, the commits aren't empty. The file tree at the time of the commits matches the file tree for projectBar. They're all stuck at the beginning of the commit history. So anyone looking at the repo over time will see some random commits from projectBar and suddenly a massive commit (the first commit of projectFoo) where the whole structure changes to projectFoo. After that it appears projectBar never intrudes on the history again.
20:31 vacho nedbat: so the best way is to manually go through the commits and make updates to my app?
20:31 justinmrkva I'll try a few more things and see if git-svn does the same thing.
20:32 nedbat vacho: if you even need the changes
20:32 Repox joined #git
20:32 TomyLobo justinmrkva, so restrict the revision range
20:32 TomyLobo git svn clone can do that
20:32 GodGinrai TomyLobo: I'd argue most GPL-licensed projects don't accurately describe what you can or can't do outside of the LICENSE file
20:32 nedbat vacho: i might be wrong, I don't know laravel, so the template may not work as I'm expecting
20:32 vacho nedbat: ok got it. I don't think I will have any merge conflicts, because he is providing a base, I have not made edits to his base.
20:32 justinmrkva TomyLobo The thing is, projectFoo and projectBar were developed in parallel for a time, so a single range does't cut it.
20:32 mkoskar joined #git
20:32 TomyLobo justinmrkva, in the same directory?
20:33 justinmrkva TomyLobo No, different directories
20:33 TomyLobo so give it a try
20:33 TomyLobo search "man git-svn" for --revision
20:33 gitinfo the git-svn manpage is available at http://jk.gs/git-svn.html
20:37 ahmedelgabri joined #git
20:37 AndroidLoverInSF joined #git
20:38 romerocesar joined #git
20:39 LeBlaaanc joined #git
20:39 justinmrkva TomyLobo Doesn't work, for example, projectFoo would have revisions 40, 41, 43, 45, and 48, and projectBar would have 42, 44, 46, and 47. You can't define a range that excludes all of projectBar while keeping all of projectFoo.
20:40 justinmrkva TomyLobo Thanks for the help, but at this point I think I just need to try a few more things to figure out what exactly is causing it to happen in the first place.
20:41 LeBlaaanc joined #git
20:43 settermjd joined #git
20:44 TomyLobo justinmrkva, try it!
20:44 jaguarmagenta joined #git
20:44 TomyLobo you identified that first commit and the directory your project is in
20:44 TomyLobo limit your clone to the directory and start it at the first commit
20:45 dopesong_ joined #git
20:45 TomyLobo the directory limit will serve to weed out everything after you moved stuff around on svn
20:45 TomyLobo the revision limit will serve to weed out everything before you moved stuff around on svn
20:46 diogenese joined #git
20:49 romerocesar joined #git
20:51 netj joined #git
20:51 fracting joined #git
20:52 Guest87932 joined #git
20:54 akushner joined #git
20:55 fracting joined #git
20:57 glauxosdever TomyLobo: Seems when cloning only reachable objects get fetched?
20:58 glauxosdever Since the manually created git repo is the same to the cloned one?
20:58 glauxosdever (Yes, I checked with diff.)
21:01 CheckDavid joined #git
21:02 l0rdPE joined #git
21:02 dopesong joined #git
21:03 xaviergmail joined #git
21:09 Impaloo57 joined #git
21:09 Impaloo57 joined #git
21:09 rominronin joined #git
21:10 TomyLobo glauxosdever, stop caring or tell us why you care
21:10 glauxosdever I care because I don't want to have old objects lying there.
21:11 allcentury joined #git
21:11 TomyLobo that's stupid
21:11 Impaloo57 joined #git
21:11 TomyLobo the garbage collection mechanism is there to clean those up
21:11 glauxosdever Having them locally isn't really a problem. Having them on remote is.
21:11 pur3eval joined #git
21:11 King_Hual joined #git
21:11 TomyLobo the remote also has git gc
21:11 nedbat glauxosdever: it sounds like it's working as you want it to?
21:12 dunpeal joined #git
21:12 glauxosdever I'll see if this is the case for gitlab too.
21:12 glauxosdever (I tested my own site first.)
21:12 TomyLobo and no, having orphanned objects remotely is not a problem
21:12 TomyLobo sometimes it can even save you
21:12 Impaloo57 joined #git
21:13 TomyLobo and if you have disk space constraints, buy a new disk
21:13 TomyLobo they're cheap these days
21:13 eletuchy joined #git
21:13 ShekharReddy is there a way so that i don have to type the username everytime i do a push
21:13 ResidentBiscuit Cant you adjust how often gc runs?
21:13 nedbat TomyLobo: doesn't it sound like it's already working the way glauxosdever hopes it will?
21:14 pur3eval hello.  i am running the command "git cherry-pick -n mybranch".  it works successfully; however, i get a merge conflict message and then i have to use a merge tool.  when i stash the same changes and apply the stash, i get a merge conflict message but it automatically merges for me.  how can i get the git cherry-pick to auto merge the way that git stash apply does it?
21:14 TomyLobo nedbat, i think he wants automatic git gc after every command
21:14 ResidentBiscuit You can create a hook to run git-gc everytime a commit is pushed? Those orphaned objects can be incredibly helpful and shouldn't be deleted prematurely, though
21:14 TomyLobo which is a waste of cpu cycles and his time
21:14 jetpack_joe joined #git
21:14 ResidentBiscuit That's similar to just insta-deleting all your logs
21:14 TomyLobo and since he's bothering us with it, also our time
21:14 ResidentBiscuit Because they take space
21:14 ShekharReddy is there a way so that i don have to type the username everytime i do a push
21:15 nedbat TomyLobo: you're talking about it more than glauxosdever is :)
21:15 ShekharReddy any config kindaa
21:15 TomyLobo maybe
21:15 TomyLobo some people just dont understand garbage collection
21:15 _ikke_ ResidentBiscuit: Do they take enough space for it to be really a problem?
21:15 TomyLobo and why doing it right now is not the best idea
21:15 pur3eval ShekharReddy: http://stackoverflow.com/questions/114034​07/git-asks-for-username-everytime-i-push
21:15 glauxosdever TomyLobo: I'm a low level programmer.
21:16 TomyLobo i figured as much
21:16 ToBeCloud joined #git
21:16 _ikke_ ResidentBiscuit: Even github does not gc immediately
21:16 nedbat TomyLobo: you're being really harsh.
21:16 TomyLobo probably C programmer who despises java because of the GC
21:16 glauxosdever Nope. I'm an operating systems developer.
21:16 glauxosdever That's why.
21:16 pwnz0r joined #git
21:16 glauxosdever I'm not an expert with git internals.
21:16 ResidentBiscuit _ikke_: No, they are worth the space the take (which isn't much)
21:17 efdetonator joined #git
21:17 TomyLobo glauxosdever, do you want to become an expert with git internals?
21:17 glauxosdever Maybe later. I have scheduler now.
21:17 TomyLobo then stop caring about the gc
21:17 ResidentBiscuit ^
21:17 _ikke_ ResidentBiscuit: Note that objects only become unreachable after something like a forced push
21:17 TomyLobo and for later:
21:17 TomyLobo !book
21:17 gitinfo 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
21:18 ResidentBiscuit _ikke_: That's exactly what we're talking about, yes
21:18 mehola joined #git
21:18 _ikke_ ResidentBiscuit: So doing gc after every push only wastes resources
21:18 axiom_1 joined #git
21:18 TomyLobo _ikke_, what about incremental "git add"s
21:18 _ikke_ TomyLobo: He's talking about the remote
21:19 ResidentBiscuit _ikke_: That wasn't a real suggestion. My next comment was "That's like insta-deleting all logs because they take space"
21:19 TomyLobo yeah, for the remote it's probably true
21:20 _ikke_ ResidentBiscuit: Sorry, got confused about who was asking the question
21:20 ResidentBiscuit Oh yeah, wasn't me
21:21 nick123 joined #git
21:21 dsdeiz joined #git
21:21 glauxosdever Why I care: Others will see the unreachable objects and will not seem clean to them.
21:22 _ikke_ They won't
21:22 glauxosdever git clone repo
21:22 _ikke_ when cloning, only reachable objects will be sent
21:22 glauxosdever Turned out I forgot to diff it recursively.
21:23 glauxosdever _ikke_: Are you sure about this?
21:23 ShekharReddy pur3eval:  thanks it worked than i expected
21:23 _ikke_ glauxosdever: yes
21:23 glauxosdever If so, then it's indeed not a problem.
21:23 _ikke_ git walks down all the refs to find out objects that are reachable and will create a pack-file for only those objects
21:24 _ikke_ That's the whole counting objects phase you see when cloning
21:24 glauxosdever This is how these pack files are explained.
21:24 glauxosdever (I thought they were unreachable objects, for some reason.)
21:24 boombatower joined #git
21:25 nedbat glauxosdever: people will only "see" the files in the working tree anyway.
21:25 dunpeal joined #git
21:26 glauxosdever nedbat: Yes. What if someone tries to see the previous commits, anyway?
21:26 Gast268__ joined #git
21:26 glauxosdever Only the reachable ones exist after cloning then?
21:26 glauxosdever Good to know.
21:27 _ikke_ yes
21:27 _ikke_ And after cloning, they will be all in a single packfile
21:28 glauxosdever All of the reachable ones, right?
21:28 glauxosdever It's good then.
21:29 _ikke_ yes
21:29 romerocesar joined #git
21:30 settermjd joined #git
21:31 ResidentBiscuit The average user is not going to see any orphaned objects ever. If they do, they know what they're doing
21:31 t0by joined #git
21:34 orbitphreak joined #git
21:34 mdw joined #git
21:35 zarel joined #git
21:37 glauxosdever Goodnight!
21:38 EvilPeng1 joined #git
21:38 ahmedelgabri joined #git
21:39 eletuchy_ joined #git
21:40 SwiftMatt joined #git
21:42 eletuchy joined #git
21:45 eletuchy_ joined #git
21:46 fracting joined #git
21:46 kerrick joined #git
21:46 ajf- joined #git
21:47 eletuch__ joined #git
21:48 kerrick When I rebase and get a merge conflict "CONFLICT (modify/delete): server/request_handler.cpp deleted in 3ef936217fdd62ed2a0cd67c8d8bd0e8adcb1b5c and modified in <message>. Version <message> of server/request_handler.cpp left in tree", is there any way to get the SHA of the commit referenced by <message>?
21:48 kerrick I can log the branch I'm rebasing, search for the message, and copy the SHA
21:48 pwnz0r joined #git
21:48 kerrick but it seems like it should be easier
21:48 HerrK joined #git
21:50 Dougie187 left #git
21:51 eletuchy joined #git
21:53 eletuchy_ joined #git
21:54 neilthereildeil left #git
21:55 jimi_ joined #git
21:56 eletuchy joined #git
21:57 joki joined #git
21:58 eletuch__ joined #git
22:00 durham joined #git
22:02 eletuchy joined #git
22:02 diegoaguilar joined #git
22:04 eletuchy_ joined #git
22:04 ekinmur joined #git
22:05 clemf_ joined #git
22:06 mdw joined #git
22:07 datasoop joined #git
22:08 dingus joined #git
22:08 datasoop i can sign a previous commits with my gpg key?
22:08 eletuchy joined #git
22:09 mehola joined #git
22:10 eletuchy_ joined #git
22:10 rominronin joined #git
22:11 rasto joined #git
22:11 j416 datasoop: only if you rewrite history
22:12 italoacasas joined #git
22:12 j416 datasoop: the sign is part of the commit, and you cannot change a commit, only replace it (and subsequently everything that comes after it)
22:12 j416 datasoop: more info https://git-scm.com/book/en/v​2/Git-Tools-Signing-Your-Work
22:12 docnvk joined #git
22:13 italoacasas joined #git
22:13 datasoop i don't wanna touch things, they are perfect, thanks j416
22:13 j416 datasoop: in that case, don't sign
22:13 bluepixel joined #git
22:13 fracting joined #git
22:14 datasoop j416: i sign, but only in my next actions.
22:14 j416 datasoop: in my opinion, signing commits is overkill.
22:14 j416 datasoop: signing tags is good enough
22:15 j416 datasoop: but ymmv.
22:15 eletuchy joined #git
22:16 fracting joined #git
22:16 cdown_ joined #git
22:17 MattMaker joined #git
22:18 eletuch__ joined #git
22:18 datasoop if i make tag in my local repo, i can make it look like a new release in github.com and be able to edit the text and add files?
22:19 j416 I don't know about github. But you tag things that don't change, that is the point.
22:20 j416 If you need them to change, you need to re-create the tag, and by that you are changing history
22:20 romerocesar joined #git
22:21 j416 datasoop: you tag a commit. A commit records the state of your repository at a given time.
22:21 lownin joined #git
22:22 eletuchy joined #git
22:22 CameronShorter joined #git
22:23 j416 datasoop: it seems github gives the option to add data that is associcated with a tag.
22:23 liuzhen joined #git
22:23 mehola joined #git
22:23 j416 datasoop: I don't know the details about how that works. #github or the github help probably knows better.
22:24 datasoop ok
22:25 benwbooth joined #git
22:25 eijk_ joined #git
22:26 kbs j416: just my current curiosity - what might be a reason to sign individual commit, rather than just stick with tags?
22:27 CameronShorter left #git
22:27 j416 kbs: to prove that the commit has not been altered
22:28 j416 kbs: e.g. someone committed in your name
22:28 eijk_ joined #git
22:31 mischat joined #git
22:32 kbs the kernel approach to this is for the committer to tag-sign the head of their pull request is that right? This one I guess proves something slightly different, interesting...
22:32 j416 I don't know how linux does it.
22:32 kbs https://www.kernel.org/pub/software/scm/git/do​cs/howto/using-signed-tag-in-pull-request.html
22:33 j416 kbs: are you saying this is how the linux project does it?
22:34 kbs j416: right - or at least, that's how I read it :-)
22:34 j416 kbs: I read nothing about this in the article
22:35 fornax joined #git
22:36 kbs no? Hm, maybe I'm missing something... doesn't that article give a mechanism that essentially the contributor tag-signing the head of their feature branch?
22:37 kbs https://www.kernel.org/doc/Do​cumentation/SubmittingPatches also refers to this, if I'm reading this correctly. search for "signed tag"
22:37 j416 kbs: it seems so, yes
22:38 j416 kbs: maybe I am misunderstanding something; what is your question?
22:39 rahtgaz joined #git
22:39 docnvk joined #git
22:39 lebster i branched off a develop branch and say i named it xupdates, that is already ahead of a release branch. i rebased and then merged xupdates into develop. Now i need those same changes in release branch but do not want the other changes in develop branch
22:39 kbs j416: mostly, trying to understand whether the tags approach is sufficient in practice, to establish who commit'ed something.
22:39 lebster do i checkout xupdates and then do rebase release
22:39 lebster then merge into release
22:40 j416 kbs: in many cases, signing at all can be overkill
22:41 lebster right now when i just merge xupdates into release, it seems to have other files that i did not touch
22:41 kbs j416: *nod* agreed. Just trying to see if there's a non-zero use-case for signing commits, that's mostly out of academic interest :-)
22:42 kbs er, meaning I just have an academic curiousity if there's a non-zero use-case for signing commits, that isn't reasonably provided by signed tags - and thinking about your example around authenticating commits.
22:43 j416 kbs: the difference between an unsigned and a signed tag, is that in the former case you cannot be sure that no one with access to say the github servers re-tagged something to plant something malicious. If the tag is signed, you cannot re-create the tag with different content -- the signature would change.
22:43 fracting joined #git
22:44 kbs j416: yes - signing per-se makes sense :-) was just thinking whether signing tags was sufficient, instead of signing commits...
22:45 jaguarmagenta joined #git
22:45 j416 kbs: if you sign a tag, or a commit for that matter, you are effectively signing all history up until and including that point.
22:45 mehola joined #git
22:45 madewokherd joined #git
22:45 kbs j416: makes sense. Maybe a better context - was reading this long tirade back in 2009 http://git.661346.n2.nabble.com/GPG-s​igning-for-git-commit-td2582986.html about why signing commits was a bad idea, given tags could be signed
22:46 kbs but now I notice git does have the ability to sign commits, hence my curiosity around what use-cases have given birth to this feature
22:46 kerrick joined #git
22:47 texasmynsted I can use .git/info/exclude to exclude a directory from git without updating the .gitignore.   Is there a way to do this for one remote but not another, i.e. rather than excluding it manually from every push?
22:47 j416 kbs: in most cases, there is no need to be paranoid about someone changing the contents of your commit
22:51 j416 kbs: I have not seem this thread. Nice post by Torvalds.
22:52 j416 kbs: I understand what you are wondering now; I don't know the answer. But I am interested to find out.
22:52 texasmynsted There is no way to do what I am asking right?
22:52 clemf joined #git
22:53 kbs j416: indeed, me too :-)
22:53 j416 texasmynsted: that's not how it works, no. You push commit history, not files.
22:53 j416 texasmynsted: a file is either in a commit or not. Cannot eat cake and have it.
22:54 texasmynsted I think the work-around I found was to exclude from git, then go into the directory make it a git repo and push it to a different remote.
22:54 j416 texasmynsted: your workflow seems flawed.
22:54 texasmynsted indeed
22:55 mmlb joined #git
22:57 jstimm joined #git
22:58 texasmynsted snap.  I added it to exclude but it is still seen
22:59 allcentury joined #git
22:59 j416 texasmynsted: you need to remove it as well if it is tracked (committed)
22:59 texasmynsted it is not
23:00 j416 texasmynsted: define 'still seen'
23:00 stonerfish joined #git
23:00 texasmynsted git status — local
23:01 texasmynsted says it is on the branch and has nothing to commit
23:01 j416 texasmynsted: so the file is ignored
23:02 texasmynsted hmm.
23:03 texasmynsted why when I cd into that directory and do a git status does it give me status for the parent git repo
23:03 texasmynsted it should not know it is in a git repo right?
23:03 j416 texasmynsted: 'the parent git repo'?
23:03 j416 texasmynsted: you have a git repo inside another one?
23:03 texasmynsted say the repo is named foo
23:03 texasmynsted I created a directory called local
23:03 j416 texasmynsted: is this a submodule?
23:03 texasmynsted I added local/ to the exclude file
23:04 texasmynsted I do not want a submodule.  I want the foo repo to know nothing about local
23:04 texasmynsted I will make local its own repo
23:04 j416 I don't see why you want to place a repo inside a repo
23:04 rnsanchez_ joined #git
23:05 texasmynsted crazy?
23:05 docnvk joined #git
23:05 mmlb joined #git
23:05 j416 texasmynsted: what are you trying to accomplish?
23:05 texasmynsted ok when I did git init it seems to be working how I want
23:06 texasmynsted I do not want the foo repo to know anything about the local directory.
23:06 texasmynsted I will use the for local things like notes etc while I am working in foo.
23:06 j416 you're not answering the question
23:06 texasmynsted I can push it to its own remote if I want
23:07 texasmynsted The owner of the foo repo does not want or need my local notes etc.
23:07 j416 texasmynsted: so why put them in foo to begin with?
23:07 NeverDie joined #git
23:07 texasmynsted because I can find them there when I work in foo
23:08 j416 texasmynsted: put them in a separate branch, then
23:08 texasmynsted I am open to better ideas
23:08 j416 texasmynsted: a nested repo is not the right way.
23:08 ahmedelgabri joined #git
23:08 texasmynsted ok.
23:09 texasmynsted Lets explore your way.  Lets say I am in foo, and branch off of develop branch into a bug/123 branch
23:09 eletuchy joined #git
23:10 texasmynsted I could keep all my notes for bug/123 there.  But then when I go to merge my changes back into develop, I will need to cherry pick to keep my notes out, rather than being able to ff merge.
23:10 NeverDie joined #git
23:10 Tim-SEM joined #git
23:11 texasmynsted or perhaps there is a better way.  What do you recommend?
23:11 j416 texasmynsted: do you need to keep your notes after you merge?
23:11 texasmynsted yes
23:11 j416 for what purpose?
23:11 texasmynsted For next time I encounter that part of the code.
23:12 Tim-SEM left #git
23:12 j416 then document properly instead
23:12 rominronin joined #git
23:12 texasmynsted Notes would include insert scripts to push the data store into the state needed to reproduce the bug, etc.
23:12 j416 document your change in the commit message, document the code itself in the code when needed
23:12 texasmynsted Ha
23:12 j416 texasmynsted: replace your script to reproduce the bug with a test case.
23:12 texasmynsted Yes, but what about the insert scripts, and other artifacts?
23:13 texasmynsted I like that.
23:13 texasmynsted I do not think I can make that work for this bug however
23:13 eletuchy_ joined #git
23:13 fracting joined #git
23:14 texasmynsted Though that is a far better way of handling the problem
23:14 j416 texasmynsted: if the information is worth saving, make it part of history somehow. If not, throw it away.
23:15 j416 texasmynsted: I don't see the point in trying to connect info to some part of the code that only you will access
23:17 eletuchy joined #git
23:17 j416 kbs: this seems to be the commit that makes it possible to gpg-sign commits. Does not seem to reveal the reason why it was added. https://github.com/git/git/commit/ba3c69a9ee
23:17 mehola joined #git
23:17 texasmynsted Hmm.  Maybe the right thing to do in this case is to add a file and or directory to part of the test path, with the insert scripts needed to reproduce the problem.
23:18 j416 texasmynsted: figure out a way to test it
23:18 j416 texasmynsted: if the script is small, I would consider in-lining it in the commit message
23:18 j416 texasmynsted: (if this indeed cannot be put as a test case)
23:18 texasmynsted It is not small
23:19 eletuch__ joined #git
23:19 texasmynsted There might be a way to test it as a unit test, that I have simply not identified yet.
23:19 rbr joined #git
23:19 j416 texasmynsted: that would be the right way, discuss with your peers?
23:20 texasmynsted I think I will add the files to some place in the test path and move it to a unit test if possible.
23:20 j416 texasmynsted: in either case, it seems this is not a git question anymore :)
23:20 artemisart joined #git
23:20 kbs j416: haha - you're way ahead of me, thanks for the pointer to the commit - interesting...
23:20 j416 texasmynsted: I hope you figure out a nice way to keep it.
23:21 eletuchy_ joined #git
23:21 texasmynsted Heh thanks.  I like your ideas.  I will try to solve this without this local nonsense
23:22 nick123 joined #git
23:23 LeBlaaanc joined #git
23:23 Literphor joined #git
23:23 j416 kbs: release notes for the release that added the feature don't reveal much detail either http://git.kernel.org/cgit/git/git.git/​plain/Documentation/RelNotes/1.7.9.txt
23:24 j416 texasmynsted: good. :)
23:24 eletuchy joined #git
23:25 cdown__ joined #git
23:25 bluezone joined #git
23:27 eletuchy_ joined #git
23:28 kbs j416: curiosier and curiosier :)
23:29 eletuchy joined #git
23:30 docnvk joined #git
23:31 timwis joined #git
23:32 j416 kbs: the ability to sign tags has been around since the early days in 2005.
23:32 timwis Hey folks, I've made a pull request and several months later they just asked me to "rebase on top of master" since it's been updated and they'd like to merge it. I only know how to merge the latest master. Is rebasing different/better? How should I do that? git fetch and then just git rebase?
23:32 eletuchy_ joined #git
23:33 j416 kbs: https://github.com/git/git/commit/d727782eaf
23:34 kbs j416: early days indeed. In the context of Linus insights on what he wants out of signatures, also makes perfect sense
23:34 j416 timwis: 1) yes/sometimes; 2) git checkout master && git pull && git checkout yourbranch && git rebase master  # (there is a shorter way, this seems pedagogical) 3) see (2)
23:34 j416 kbs: nod
23:35 timwis thx!
23:35 j416 timwis: (shorter way assuming remote is named origin: git checkout yourbranch && git fetch && git rebase origin/master)
23:35 timwis yeah i was thinking the latter. upstream in this case
23:35 j416 timwis: then you will need to force-push (git push --force)
23:35 timwis but thx
23:36 xaviergmail joined #git
23:36 eletuchy joined #git
23:37 timwis do i have to force push if i use the first method j416?
23:37 j416 timwis: short answer to why rebase is better than merge in your case, is that you end up with a less cluttered history.
23:37 j416 timwis: define "first method"
23:37 timwis yeah i think that makes sense. it's as if i wrote the changes as of today, rather than 6 months ago, right?
23:37 timwis git checkout master && git pull && git checkout mybranch && git rebase master
23:37 timwis i'm getting "Updates were rejected because the tip of your current branch is behind" when i push
23:38 j416 timwis: the author timestamp will be unaltered, but the commits will be as if you based them on the current master, yes.
23:38 j416 timwis: did you read what I wrote?
23:38 j416 timwis: scroll up 3 minutes or so
23:39 timwis "then you will need to force-push" ? Yes - I was asking whether that applied to the first method you mentioned as well. Reading it, it sounded like it only applied to the 2nd method.
23:39 j416 it does, yes.
23:39 clemf joined #git
23:39 timwis okay, thanks
23:39 Lyqyd joined #git
23:39 fracting joined #git
23:39 j416 apologies for being unclear
23:39 timwis no worries, and thanks for explaining the benefits. i understand it now
23:39 eletuchy_ joined #git
23:39 j416 the only difference between the two is that the former will give you an updated branch "master" and the second won't.
23:40 Timvde timwis: You have to force push if the last remote commit isn't an ancestor of your current branch head
23:40 timwis yeah, i think i get that
23:41 j416 (upstream/master and master are two different branches, upstream/master is update when you do 'git fetch' and is read-only to you, master is updated when you do 'git pull' and have it checked out, or by other means, but never by 'git fetch')
23:41 j416 s/is update/is updated/
23:41 timwis interesting, it didn't update my existing pull request on github. could that be because of the --force?
23:41 eletuch__ joined #git
23:41 j416 timwis: probably because github is slow to refresh, check again in a minute or so
23:41 fracting joined #git
23:42 timwis oh interesting, that's always been pretty fast for me. i'll give it a wait though
23:42 j416 timwis: then again I have only used github pull requests a handful of times, so I'm not very familiar with them
23:43 j416 kbs: if there is information, I think it's in the mailing list
23:43 zumba_addict joined #git
23:43 eletuchy joined #git
23:43 cdown joined #git
23:44 kbs j416: *nod* I'll poke around there around that time period later, but should you find something interesting, would love to know what happened! And, thanks for the various pointers along the way :-)
23:44 fracting joined #git
23:45 eletuchy_ joined #git
23:46 ffabi joined #git
23:50 eletuchy joined #git
23:51 mkelly joined #git
23:51 romerocesar joined #git
23:51 ash_workz joined #git
23:54 xlegoman joined #git
23:56 datasoop left #git
23:59 ffabi joined #git
23:59 barteks2x joined #git
23:59 rsukla joined #git

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