Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-08-15

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

All times shown according to UTC.

Time Nick Message
00:00 chrunty joined #git
00:06 arooni smart man
00:09 Prisoner-2460_1_ joined #git
00:10 jimi_ joined #git
00:12 Raed|Mobile joined #git
00:14 jcadduono joined #git
00:14 jcadduono is it possible to drop all history before a specific commit while keeping the commit ids the same
00:14 p4trix joined #git
00:15 bremner no
00:15 jcadduono i thought about just git cloning to a specific depth, but then i'm not allowed to push a shallow clone to github
00:15 jcadduono reason being is this repo is like 10 GB and it only needs to be ~400 MB
00:16 cbreak jcadduono: rewrite history
00:18 soLucien joined #git
00:20 bremner of course that changes commit hashes
00:20 jcadduono how can i find the depth of a commit id (so i can git clone --depth it)
00:21 cd-rum_ joined #git
00:22 hahuang61 joined #git
00:22 cbreak jcadduono: probably something with git rev-list --count
00:23 cbreak jcadduono: probably something with git rev-list --count XXX..
00:26 The_Phoenix1 joined #git
00:28 ASOLAr joined #git
00:29 noteness joined #git
00:29 energizer joined #git
00:29 Prisoner-2460_1_ I jsut squashed two commits. now i’m gonna have to squash this newest one with that. poor planning i know. is this going to turn in to a mess
00:29 energizer How can i add a remote using an ssh alias that i have previously set up?
00:30 kadoban Prisoner-2460_1_: Doesn't sounds like that much of a mess.
00:30 kadoban sound*
00:30 SporkWitch all depends on how well you comment
00:31 Prisoner-2460_1_ your approval is all i needed
00:31 Prisoner-2460_1_ what worries me is that the commit message for the last one includes both messages, and now this one will have those two plus this one. is it a better idea to overwrite the message altogether?
00:32 energizer git remote add origin myalias:/srv/data/.git
00:32 mingrammer joined #git
00:32 monoprotic joined #git
00:32 energizer fatal: Not a git repository (or any parent up to mount point /mnt/TeraTwo)
00:32 energizer Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).
00:33 jcadduono left #git
00:33 kadoban energizer: Sounds pretty self-explanatory. You're not in a git repository.
00:34 energizer kadoban: oh i thought it was saying the target wasn't a git repo - indeed im not in one
00:35 catag87 joined #git
00:36 skered joined #git
00:37 SporkWitch lol
00:38 skered Does git keep any type of history of the last pull?  I want to know all the commits between the last pull command.  If there's nothing builtin I guess I could just save the latest commit and run git log afterwards.
00:39 catag87 joined #git
00:40 Gitzilla joined #git
00:40 SporkWitch skered: pull == fetch+merge.  To know about commits since the pull, do a fetch.
00:40 catag87 joined #git
00:42 skered I think I remember reading an article about that.  Don't use pull - use fetch and merge.
00:42 skered Thanks I'll look at that thanks
00:42 SporkWitch skered: there's nothing wrong with using pull that i've ever heard, it's just not what you want if you don't also want a merge.
00:43 mikecmpbll joined #git
00:44 nidr0x joined #git
00:44 catsup joined #git
00:44 a__ joined #git
00:45 rik316 joined #git
00:48 sim590 joined #git
00:49 garethdaine joined #git
00:49 mikecmpbll joined #git
00:50 monoprotic joined #git
00:52 monoprotic joined #git
00:56 tesuji_ joined #git
00:57 johnny56_ joined #git
00:59 Rodya_ joined #git
01:00 rkazak joined #git
01:00 Guest29537 joined #git
01:01 Anderson69s_Pi joined #git
01:02 johnny56 joined #git
01:02 safe joined #git
01:03 HoierM joined #git
01:04 Derperperd joined #git
01:12 monoprotic joined #git
01:12 fuchstronaut joined #git
01:13 d^sh joined #git
01:15 pks joined #git
01:15 SwiftMatt joined #git
01:21 dlaboratory joined #git
01:23 dsdeiz joined #git
01:26 skered Does git -C not work or do I just not know how to use it?  git -C ~/path/to/dir/ log   is running as if I'm running from cwd.  It seems like --git-dir is being set before -C is taken in to account.
01:26 mingrammer joined #git
01:27 rkazak joined #git
01:27 skered It seems to be acting as if I'm running in cwd because it's asking for my dotfiles' host ssh key phrase which is ~/.git/'s remote ssh://
01:31 fish3485 joined #git
01:33 Dougie187 joined #git
01:37 zacts joined #git
01:40 harish_ joined #git
01:41 Chunk2 joined #git
01:45 danimal2 joined #git
01:48 ilbot3 joined #git
01:48 Topic for #git is now Welcome to #git, the place for git help and doorbells | Public logs at http://goo.gl/BuUi5o | Current stable version: 2.9.3 | First visit? Read: http://jk.gs/git | Getting "cannot send to channel"? /msg gitinfo .voice | Knock knock. Who's there? Git. Git-who? Sorry, 'who' is not a git command - did you mean 'show'?
01:49 rkazak joined #git
01:49 tubaman joined #git
01:50 cyphase joined #git
01:51 tubaman joined #git
01:55 kyan joined #git
01:55 Vampire0 joined #git
01:57 codebam I'm doing a merge that works cleanly, but I don't want to merge everything. how do I revert one of the commits when I have files staged?
01:58 codebam never mind
02:01 livingstn joined #git
02:01 ericsupreme joined #git
02:02 spudowiar joined #git
02:11 tesuji joined #git
02:12 thiago joined #git
02:13 fuchstronaut joined #git
02:13 zivester joined #git
02:15 soLucien hi guys ! Is GitFlow a subject that i can ask about here ?
02:17 SporkWitch bit late now, isn't it?
02:18 soLucien i guess so , it's 4AM here
02:20 weox joined #git
02:21 mingrammer joined #git
02:23 diogenese joined #git
02:26 chachasmooth joined #git
02:30 theShe joined #git
02:31 l4v2 joined #git
02:32 energizer joined #git
02:32 catag87 joined #git
02:33 vassagus joined #git
02:34 catag87 joined #git
02:35 Rodya_ joined #git
02:37 Rodya_ joined #git
02:38 lagothrix joined #git
02:38 rossome joined #git
02:39 Rodya_ joined #git
02:40 Rodya_ joined #git
02:42 hexagoxel joined #git
02:42 Rodya_ joined #git
02:44 Rodya_ joined #git
02:46 Rodya_ joined #git
02:50 zumba_ad_ joined #git
02:50 skylite joined #git
02:51 kyan joined #git
02:52 johnny56 joined #git
02:54 justanotheruser joined #git
02:54 arooni hey folks;  i use git to track my personal journal.  is there any reason why i would *not* want to write a simple script and call it from cron that would basically run: git add -u; git commit -m "auto generated backup"; git push ?
02:55 arescorpio joined #git
02:56 SporkWitch arooni: git is not a backup tool, it is a version control system
02:56 SporkWitch arooni: if you want something for backups, look into rsync
02:58 SporkWitch arooni: (git produces a lot of additional data so that you can revert to any previous version, and this is non-desirable for backups, where you typically only one a couple versions back maintained).  What's more, automated commits would clutter your history, reducing the ease with which you could figure out where changes came from (commit comments should ALWAYS include meaningful information about
02:58 SporkWitch what changed; you shouldn't have to go digging through a diff to find out every single difference)
03:00 zifnab joined #git
03:02 arooni thanks for the advice
03:02 arooni this is just until i get my regular backup system working
03:03 arooni as i've already lost some data in this directory due to my own stupidity
03:03 SporkWitch arooni: i still wouldn't automate it the way you described.  Honestly, for your purposes, something like Google Drive (or if you prefer maintaining more control, owncloud) would be ideal
03:04 SporkWitch arooni: they have windows, mac, and linux clients that will automatically keep a folder on you computer sync'd with the remote storage
03:04 SporkWitch (also android clients; dunno about apple)
03:05 SporkWitch i believe you can even enable version control in both of those, if you wish
03:05 tristanp joined #git
03:05 SporkWitch (i haven't used the latest release of owncloud, so can't be sure; apparently they improved the collaborative editing features a LOT a month ago)
03:05 harish_ joined #git
03:08 fstd_ joined #git
03:09 arooni i just dont like the idea of uploading my personal journal to the cloud
03:09 arooni plus i like being able to edit in vim
03:09 arooni i suppose putting it on github or something is still the cloud lol
03:12 SporkWitch arooni: github free is definitely public lol
03:12 SporkWitch arooni: google drive uses google+ permissions, so it defaults to private unless you change the sharing settings on the item
03:12 SporkWitch arooni: owncloud depends on how you set it up (it's a self-hosted solution)
03:13 SporkWitch arooni: if you're really worried, though, you should look into pgp.  It's not nearly as intimidating as it seems at first glance, especially for your purposes, and you could use it to encrypt your journal
03:14 fuchstronaut joined #git
03:14 SporkWitch (there's even a really sweet vim plugin that can edit pgp-encrypted files without having to write a plaintext copy to disk)
03:15 dtrott joined #git
03:15 SporkWitch there's also keybase.io, which is a front-end of sorts to pgp, with a bunch of value-added features, including 8 or 10 GB of free remote storage, but it's all encrypted, and they don't have the keys
03:15 SporkWitch (if it's of interest let me know and i'll pm you an invite link, i've got like 20 of them)
03:16 marcm joined #git
03:17 Bizkit joined #git
03:17 phroa oh, it does more than exist as a place to send people your key ID now?
03:23 SporkWitch phroa: yup, they rolled out keybase filesystem recently
03:24 SporkWitch i haven't messed with it yet, but someone said it offers more than dropbox free accounts, and it's all encrypted (and can be used locally on your own machine, not just on the remote storage)
03:24 SporkWitch phroa: what little i read on it, it seems like an incredibly convenient (and SECURE!) filesharing tool
03:28 inflames joined #git
03:32 Andrew_K joined #git
03:33 Prisoner-2460_1_ joined #git
03:34 inflames joined #git
03:34 clarens0000 joined #git
03:34 Dougie187 joined #git
03:38 cagedwisdom joined #git
03:39 Prisoner-2460_1_ git rebase -i HEAD~2 gave me 20 commits. i just want to squash the latest two. do i just edit pick to squash on the second one and save?
03:40 ShekharReddy joined #git
03:42 SporkWitch i THINK so? the interactive rebase should ask you what you want to do with each commit as you're going along.  I'm still learning the correct answers to it to make it do what i want, though, heh
03:43 Prisoner-2460_1_ it didnt work
03:43 Prisoner-2460_1_ not even sure what happened. i got out of there and back to where i started
03:44 Prisoner-2460_1_ i was on a stale state of develop when i commited, had to pull and merge, and want to squash that into a single commit
03:46 SporkWitch that's a bit beyond my git-fu :(
03:46 SporkWitch when it comes to git, i'm still fairly basic
03:46 chandu20 joined #git
03:51 Prisoner-2460_1_ aren’t we all
03:57 Rodya_ joined #git
04:00 Manuel15 joined #git
04:00 Manuel15 joined #git
04:01 thecomedian joined #git
04:03 pwnz0r_ joined #git
04:09 afuentes joined #git
04:09 mingrammer joined #git
04:14 fuchstronaut joined #git
04:24 Cabanossi joined #git
04:24 ytl joined #git
04:33 Maple__ joined #git
04:36 frogonwheels joined #git
04:37 Andrew_K|2 joined #git
04:42 Limes_ joined #git
04:42 csd_ joined #git
04:44 sanketdg joined #git
04:44 GT4066 joined #git
04:45 khmarbaise joined #git
04:52 cdown_ joined #git
05:03 Andrew_K joined #git
05:03 mingrammer joined #git
05:04 nilg joined #git
05:08 jceb joined #git
05:10 hahuang61 joined #git
05:15 fuchstronaut joined #git
05:16 Raging_Hog joined #git
05:16 lamont joined #git
05:18 [cust] joined #git
05:18 watabou joined #git
05:18 crose joined #git
05:29 t-mart joined #git
05:34 annoymouse joined #git
05:35 evanvarvell joined #git
05:39 stuh84 joined #git
05:42 csd_ joined #git
05:46 noteness joined #git
05:46 sulky joined #git
05:48 subhojit777 joined #git
05:50 Guest29537 joined #git
05:50 hahuang61 joined #git
05:53 mohabaks joined #git
05:54 ASOLAr joined #git
05:55 JanC_ joined #git
05:57 catag87 joined #git
05:58 catag87 joined #git
06:00 LucaTM joined #git
06:02 klean joined #git
06:03 JustAnotherIdiot joined #git
06:03 JustAnotherIdiot how do I delete the last commit I made?
06:06 eivarv joined #git
06:07 JustAnotherIdiot nvm
06:07 JustAnotherIdiot I figured it out
06:11 a_thakur joined #git
06:12 a_thakur_ joined #git
06:14 mariuscc joined #git
06:16 fuchstronaut joined #git
06:17 lb JustAnotherIdiot :)
06:17 lb JustAnotherIdiot best learning experience ever
06:18 lb just for the record: git reset HEAD^
06:18 JustAnotherIdiot I dunno git reset HEAD~ worked
06:19 lb JustAnotherIdiot same stuff
06:19 lb JustAnotherIdiot HEAD~ = HEAD~1 = the commit before HEAD
06:20 lb HEAD^ = the first parent of the current commit, which can differ if the current HEAD is a merge commit. otherwise if HEAD is not a merge commit then HEAD~ = HEAD~1 = HEAD^
06:21 hahuang61 joined #git
06:21 JustAnotherIdiot i dunno
06:21 JustAnotherIdiot don't really care
06:21 JustAnotherIdiot it worked
06:23 Rish joined #git
06:23 tristanp_ joined #git
06:24 tristanp joined #git
06:26 nnyk joined #git
06:27 lb that reminds me of https://xkcd.com/1597/
06:30 JustAnotherIdiot lb: lol
06:30 aubreypwd joined #git
06:32 vmiklos joined #git
06:35 rchavik joined #git
06:35 xaa joined #git
06:39 vassagus joined #git
06:39 EnergyCoffee joined #git
06:40 rkazak joined #git
06:41 EnergyCoffee joined #git
06:42 Rish joined #git
06:42 rj1 joined #git
06:42 EnergyCoffee joined #git
06:43 EnergyCoffee joined #git
06:45 csd_ joined #git
06:45 EnergyCoffee joined #git
06:45 lukebyrne joined #git
06:46 EnergyCoffee joined #git
06:48 EnergyCoffee joined #git
06:49 aidalgol joined #git
06:51 MrMojito joined #git
06:51 MrMojito joined #git
06:51 mingrammer joined #git
06:51 lukebyrne joined #git
06:52 cdesai joined #git
06:52 submitnine joined #git
06:53 djb-irc joined #git
06:55 mozzarella joined #git
07:00 vassagus_ joined #git
07:01 nnyk joined #git
07:03 LiohAu joined #git
07:03 djb-irc joined #git
07:06 nitin joined #git
07:07 armin joined #git
07:08 jknetl joined #git
07:09 mohabaks joined #git
07:10 dlaboratory joined #git
07:12 nanozer2 joined #git
07:14 rorro joined #git
07:14 alansaul joined #git
07:14 dtrott joined #git
07:17 fuchstronaut joined #git
07:18 dsdeiz joined #git
07:18 dsdeiz joined #git
07:20 t0by joined #git
07:24 rageshkrishna joined #git
07:27 grift joined #git
07:27 Balliad joined #git
07:34 Anticom joined #git
07:35 vassagus joined #git
07:37 jagob joined #git
07:38 lohfu joined #git
07:38 lohfu joined #git
07:43 iskander joined #git
07:43 a_thakur_ joined #git
07:46 mingrammer joined #git
07:46 csd_ joined #git
07:47 Ardethian\\work joined #git
07:49 thomas1 joined #git
07:50 hellyeah joined #git
07:50 hellyeah i commanded git init
07:51 hellyeah i got message empty git repo initialized
07:51 hellyeah but there is no .git folder
07:51 hellyeah something wrong?
07:51 hellyeah ah it there
07:51 hellyeah i am sorry
07:51 ahmedelgabri joined #git
07:56 dtrott joined #git
07:57 fished joined #git
07:58 lss8 joined #git
07:58 j416 hell yeah.
07:59 iskander joined #git
08:03 vassagus_ joined #git
08:04 quenode joined #git
08:05 UTAN_dev_ joined #git
08:13 tomog999 joined #git
08:17 fuchstronaut joined #git
08:18 wrouesnel joined #git
08:20 zeroed joined #git
08:27 jstein_ joined #git
08:28 dreiss joined #git
08:28 alansaul joined #git
08:28 ahmed_elgabri joined #git
08:31 liuzhen joined #git
08:31 Pursuit joined #git
08:32 nuc joined #git
08:33 PCatinean joined #git
08:34 Shentino joined #git
08:34 rj1 joined #git
08:34 GavinMagnus joined #git
08:35 filcab joined #git
08:35 nikivi joined #git
08:36 hahuang61 joined #git
08:40 mingrammer joined #git
08:44 mohabaks joined #git
08:48 MrTrick joined #git
08:50 MrTrick I've got files in my repo like
08:50 MrTrick And resized files like wp-content/uploads/(year)/(​month)/(name)-1024x768.jpg
08:50 MrTrick Is there a way to ignore any of the resized files? All different resolutions, but numeric.
08:51 osse MrTrick: are there unresized files whose names have dashes in them?
08:51 nuc joined #git
08:51 kurkale6ka joined #git
08:51 MrTrick osse: probably.
08:51 noodle joined #git
08:51 osse MrTrick: are there unresized files whose names have dashes in them and there is an x in the part after the dash?
08:52 SlizardWizard left #git
08:53 robotroll joined #git
08:54 cbreak MrTrick: just ignore the whole uploads directory
08:54 LiohAu joined #git
08:54 lordjancso joined #git
08:54 cbreak it's probably not part of the source code
08:54 dodobrain joined #git
08:55 OmNom joined #git
08:55 leeN joined #git
08:56 cyan__ joined #git
08:56 nnyk joined #git
08:56 TomyWork joined #git
08:58 MrTrick cbreak: I'm ignoring the source code and keeping the content.
08:58 cbreak that sounds... dumb
08:58 MrTrick cbreak: the source code is wordpress, there's no value in tracking it.
08:58 jast wouldn't a backup tool be a better solution, then?
08:59 cbreak MrTrick: if the source code isn't valuable, why bother with git in the first place?
08:59 MrTrick well, I want to be able to see changes, roll/branch/deploy etc.
08:59 jast isn't the majority of the metadata and content in a database when using wordpress?
08:59 jast databases are tricky to version
08:59 lb jast correct
09:00 jast rhetorical question, I used to run wordpress myself ;)
09:00 MrTrick jast: yes correct, was going to export/import
09:01 iskander joined #git
09:01 cbreak sounds like a job for... zfs snapshot!
09:01 lb sounds like a job for a xkcd
09:03 LeMike joined #git
09:04 cbreak https://xkcd.com/627/?
09:05 constant joined #git
09:06 qsx i need this in german for my mother
09:06 lb cbreak perfekt
09:06 lb s/perfekt/perfect/
09:08 nuc joined #git
09:14 EisNerd moin
09:15 EisNerd is there a function which would do the similar thing as git fomat-patch / git apply would do, to transport changes from one branch to another?
09:16 tobiasvl EisNerd: man git cherry-pick perhaps?
09:16 gitinfo EisNerd: the git-cherry-pick manpage is available at http://jk.gs/git-cherry-pick.html
09:16 lb yup, sounds like cherry-pick to me
09:17 EisNerd hm but cherry pick here behaves more like rebase
09:17 tobiasvl how so?
09:17 dreiss joined #git
09:18 fuchstronaut joined #git
09:18 EisNerd I'm currently bringing over the history to a recreated git repo, as the old one was messed up, by throughout the history imported large binary packs and multiple thousands of unneeded files
09:19 notebox joined #git
09:19 cdown joined #git
09:19 EisNerd therefore I decided to move some subtrees into submodules as they aren't intended to be changed throughout the project
09:19 EisNerd now rebase / cherry-pick always gives me conflicts on those
09:21 EisNerd futhermore rebase always forces me to go the way back to the old polluted state, resulting in several minutes of reviving waste on my disk, just to drop it right after
09:22 EisNerd a very unpleasant situation
09:22 cbreak cherry-pick does something like format-patch internally
09:22 cbreak EisNerd: why don't you use filter-branch?
09:22 EisNerd but here it also complains over the submodules
09:22 cbreak instead of wasting time with patching?
09:23 cbreak EisNerd: of course it does
09:23 cbreak it'd be quite dumb if it didn't complain about conflicts
09:23 EisNerd cbreak: but the commits cherry-picked have no chnages in thoose directories
09:24 cbreak EisNerd: but one of the bases has
09:24 texinwien joined #git
09:24 cbreak EisNerd: I recommend using filter-branch
09:24 cbreak don't try to do large scale history rewriting with rebase or cherry-pick
09:24 cbreak anything patch-based is too fragile, avoid :)
09:25 nuc joined #git
09:26 EisNerd hm ok
09:26 EisNerd but I rebuild the base from scratch carefully this time
09:27 EisNerd and we talk about roughly 80000 files no longer included
09:27 cbreak :(
09:27 EisNerd and even more when it comes to the submodules
09:27 cbreak that'll produce monster conflicts
09:27 EisNerd but I need to see potential changes in the submodules to update them accordingly
09:28 EisNerd those files where never chnaged
09:30 EisNerd so if you could tell me howto use filter branch to introduce submodules into the history, it would be great, but some rebase would still be needed as we cleaned out some initial source base missfits, the source base was created during some initial lectures on the stuff
09:30 lpapp joined #git
09:31 EisNerd by just dropping some working folder with the intended base under git control
09:31 lpapp hey, I got remote: warning: only found copies from modified paths due to too many files and remote: warning: you may want to set your diff.renamelimit variable to at least 2781 and retry the command when pushing, but I would not know why! This was my push output: https://paste.kde.org/pbxvscudu/fvhzdt/raw Why is this happening to me?
09:32 EisNerd later we noticed, the fact that the tree is lot large and even more complex in the state it was handed out to us, than we expected
09:32 EisNerd ending up with 3.5G bare repo
09:32 EisNerd and even more
09:32 Xizor joined #git
09:34 LucaTM joined #git
09:34 mingrammer joined #git
09:36 alansaul joined #git
09:36 lpapp I mean we have not done any massive renaming... is it possible to get examples what renames it detects, etc?
09:37 davisonio joined #git
09:39 roelmonnens joined #git
09:40 cbreak EisNerd: filter-branch: write a script to change every commit to how it should be
09:40 cbreak i.e. introduce a submodule
09:41 Anticom joined #git
09:42 nuc joined #git
09:42 roelmonnens joined #git
09:42 QwertyDragon joined #git
09:45 EisNerd would filter brnch stop/complain if there is an offending change, like a change in a subtree that is moved to a submodule?
09:46 roelmonnens joined #git
09:47 ljc joined #git
09:48 csd_ joined #git
09:49 roelmonnens joined #git
09:51 roelmonnens joined #git
09:51 harish_ joined #git
09:58 hjubal joined #git
09:59 zeroed joined #git
10:03 stamina joined #git
10:03 cbreak EisNerd: no
10:03 cbreak EisNerd: it stops when you want to. You write a script for it
10:04 inflames joined #git
10:05 nuc joined #git
10:07 jast lpapp: I believe this doesn't run on push except if a custom hook is used on the server. in that case the output is misleading because it's actually kind of directed at the server admins. :)
10:07 jast (your paste 404s, though)
10:07 Muzer joined #git
10:09 lpapp jast: aye, I put it up for half an hour only.
10:09 vassagus joined #git
10:10 elementalest joined #git
10:10 jast I probably don't need to look at it anyway :)
10:10 lpapp jast: https://paste.kde.org/pjbzlcedx/eongm7/raw
10:10 lpapp ah, ok.
10:10 jast what I guess is happening is that the folks who run the server are using one of the diff commands in their hook to do a few sanity checks, and they're not suppressing the warning
10:10 lpapp I am one of those folks, so let me have a brief look. ^_^
10:11 jast that sure makes things easier :)
10:13 lpapp which hook to check?
10:13 quenode_ joined #git
10:13 jast could be any of: pre-receive, update, post-receive, post-update
10:13 lpapp post-receive has diff in it
10:14 jast that's the one, then, probably
10:14 lpapp git diff-tree --stat --summary --find-copies-harder $oldrev..$newrev
10:14 lpapp could that cause any issues? ^^^
10:14 jast that one could cause that warning, yes
10:14 jast if the server has plenty of resources you could go: git -c diff.renamelimit=(any number here) diff-tree ...
10:14 lpapp jast: https://paste.kde.org/p5rttsejs/wabxmu/raw
10:15 artemisart joined #git
10:15 lpapp just to give some context
10:15 nnyk joined #git
10:15 jast hmm, that doesn't even seem to output
10:15 jast the "Summary of changes:" doesn't show in your paste
10:15 jast is the hook marked executable?
10:15 rahtgaz joined #git
10:16 lpapp it is just part of the hook script
10:16 lpapp but you are right.
10:16 lpapp I could not see that in the output
10:16 lpapp so maybe this is not the root cause of those warning
10:16 lpapp s
10:18 lpapp jast: yes, the script is executable
10:19 fuchstronaut joined #git
10:20 lpapp jast: seems we have a copy of this, more or less: http://www.scs.stanford.edu/his​tar/gitrepo/hooks/post-receive
10:21 lpapp and the other files in the hooks directory are suffixed as ".sample"
10:23 jast oh, duh, of course it wouldn't output to you, it's being included in the e-mail sent by the hook
10:24 inflames joined #git
10:24 lpapp yes, so what is outputting it to my console at push time?
10:24 jast the warning goes to STDERR and so isn't captured by the pipe that sends it as an e-mail
10:25 lpapp I see. Actually, I do not even know to which email it is sending these to.
10:25 lpapp I wonder how to figure that out.
10:25 The_Phoenix joined #git
10:26 jast look in 'config' in the repo
10:26 jast hooks.announcelist / hooks.mailinglist
10:27 The_Phoenix joined #git
10:28 lpapp ah, I see, thank you. Why does every commit need sending an email? This was "adjusted" by a colleague of mine who left.
10:28 artemisart1 joined #git
10:28 jast it's a matter of taste
10:28 lpapp and to basically a no longer existing company email address :-)
10:30 mohabaks joined #git
10:30 jast anyway, the two settings can be used to distinguish between normal commits (mailinglist) and tags (announcelist)
10:31 jast so, if you think one mail for each commit is a bit much, you can still use the same hook for getting mails only when a tag is created/updated
10:31 fuchstronaut joined #git
10:31 tvw joined #git
10:31 sdothum joined #git
10:32 lpapp sure, so the problem is that the post-receive script sends the emails from stdout, but not from stderr and so the stderr goes to the client?
10:32 jast yeah
10:33 jast you could *probably* just add a '2>&1' to that line
10:33 jast but then of course all of the recipients get to see that warning instead :)
10:35 vassagus joined #git
10:36 lpapp yes, but what generates that warning in the first place?
10:36 lpapp that goes into the email or to the client?
10:36 lpapp can we avoid that warning by doing things right?
10:36 lpapp I mean I do not get this output all the time
10:37 lpapp I have had almost 1000 commits and this is the first time I see it and I wonder why. I have not done any massive renaming as far as I can tell, so I would like to understand why this is happening to me.
10:37 LeMike joined #git
10:38 hahuang61 joined #git
10:38 fuchstronaut joined #git
10:39 bongjovi joined #git
10:41 jast well, --find-copies-harder makes git go the extra mile and search a lot of objects, potentially
10:42 filcab joined #git
10:42 jast normal -C (copy detection) only considers files as the source that have been modified in the same commit. with --find-copies-harder it can go back in history and look there, too
10:43 jast so, potentially it can look at every revision of every file :}
10:43 jast if you don't care about copy detection that much, you can replace --find-copies-harder with -C
10:44 artemisart1 joined #git
10:45 lpapp ok, thank you.
10:51 palus joined #git
10:55 yoan joined #git
10:57 boombatower joined #git
10:58 Ezzy joined #git
11:01 ruskie joined #git
11:01 Oebele joined #git
11:01 Rodya_ joined #git
11:07 wrouesnel1 joined #git
11:07 fmcgeough joined #git
11:07 a_thakur joined #git
11:07 Rodya_ joined #git
11:11 a_thakur joined #git
11:13 a_thakur joined #git
11:16 ahmed_elgabri joined #git
11:18 rorro joined #git
11:20 HoierM joined #git
11:23 jimi_ joined #git
11:23 mingrammer joined #git
11:23 a_thakur joined #git
11:31 chandu20 joined #git
11:34 Gianormagantrous joined #git
11:34 ash_workz joined #git
11:37 a_thakur joined #git
11:38 a3Dman joined #git
11:45 chandu20 joined #git
11:47 jeffreylevesque joined #git
11:49 csd_ joined #git
11:50 johnmilton joined #git
11:51 bartosz_ joined #git
11:54 fmcgeough joined #git
11:54 rageshkrishna joined #git
11:55 linuxmint joined #git
11:56 spriz joined #git
11:56 Munt joined #git
11:57 wrouesnel1 joined #git
11:57 fahadash joined #git
12:03 zeroed joined #git
12:03 GavinMagnus joined #git
12:04 mohabaks joined #git
12:04 Kobaz sooo
12:04 i7c Yes?
12:04 Kobaz yeap
12:04 Kobaz i have an issue with a local repo
12:04 Kobaz git diff origin/2.0..2.0
12:05 Kobaz i'm showing differences.... i git pull and i'm up to date
12:05 Kobaz so what i tried is git diff origin/2.0..2.0 > patch
12:05 fuchstronaut joined #git
12:05 Kobaz and then apply the path, but patch is saying it's already applied
12:05 Kobaz my current branch is 2.0
12:06 Kobaz i want the changes applied to my repo that are showing in git diff origin/2.0..2.0
12:06 Kobaz er
12:06 Kobaz whoops
12:06 Kobaz that's the wrong way around
12:06 Kobaz haha
12:07 kyan joined #git
12:07 Kobaz should be good, okay thanks
12:08 i7c Why would you write the diff into a patch and then apply it instead of just running git merge?
12:08 Kobaz trying to just fix it for now
12:08 Kobaz get it to "just work" (tm)
12:08 Kobaz but that does sound like a good way
12:09 Kobaz git merge origin/2.0 Already up-to-date.
12:10 LiohAu joined #git
12:12 UrsoBranco joined #git
12:12 hjubal joined #git
12:13 Kobaz that's why
12:16 InfoTest joined #git
12:20 Maple__ joined #git
12:21 dlaboratory joined #git
12:21 ahmed_elgabri joined #git
12:22 hahuang61 joined #git
12:22 Manuel15 joined #git
12:22 Manuel15 joined #git
12:25 dsdeiz joined #git
12:27 vassagus joined #git
12:29 TooLmaN joined #git
12:31 star_prone joined #git
12:35 artemisart joined #git
12:38 zivester joined #git
12:39 cdown joined #git
12:39 stasmakarov joined #git
12:40 GenZai joined #git
12:40 chandu20 joined #git
12:41 FuzzySockets joined #git
12:44 Gamecubic joined #git
12:46 ASOLAr joined #git
12:47 sentriz joined #git
12:49 spriz joined #git
12:50 csd_ joined #git
12:52 algae joined #git
12:53 jimi_ joined #git
12:55 cdg joined #git
12:56 mohabaks joined #git
12:56 vassagus joined #git
12:58 nnyk joined #git
12:58 ShekharReddy joined #git
12:59 flavius joined #git
13:00 cd-rum joined #git
13:00 kadoban joined #git
13:00 Faylite joined #git
13:01 Thorn__ joined #git
13:01 daniel_rose joined #git
13:03 cd-rum_ joined #git
13:04 perlpilot joined #git
13:04 tvw joined #git
13:04 Thorn joined #git
13:05 veilig2000 joined #git
13:08 bsanford joined #git
13:11 mingrammer joined #git
13:11 chid joined #git
13:12 Guest5 joined #git
13:12 zeroed joined #git
13:15 robotroll joined #git
13:16 izabera left #git
13:17 ExoUNX joined #git
13:17 byteflame joined #git
13:20 zeroed joined #git
13:21 harish_ joined #git
13:21 fuchstronaut joined #git
13:23 cdg joined #git
13:24 rgrinberg joined #git
13:25 rgrinberg joined #git
13:28 cdg joined #git
13:30 cyphase joined #git
13:30 codeman joined #git
13:31 nullp0inter joined #git
13:32 nullp0inter i want to setup a staging repo for testing in between development and production, is this the best way to achieve this? https://www.digitalocean.com/commun​ity/tutorials/how-to-set-up-automat​ic-deployment-with-git-with-a-vps
13:33 callumacrae joined #git
13:35 kbs joined #git
13:35 morenoh149 joined #git
13:38 ams__ joined #git
13:39 Sonderblade joined #git
13:39 al-damiri joined #git
13:39 rossome joined #git
13:40 lagothrix joined #git
13:41 ahmed_elgabri joined #git
13:44 TSS_ joined #git
13:44 CheckDavid joined #git
13:46 slayerjain joined #git
13:47 YourTheBombNow23 joined #git
13:48 lb that's one way to achieve it. "best" is a rather ambiguous term
13:48 lpapp joined #git
13:48 snowkidind joined #git
13:49 lpapp hi, I apparently checked in some library (build artefacts) into my repository as per https://paste.kde.org/pyyrdifnn/4cb0ji/raw. When I try to do git rm libfoo.so.dev or even git checkout libfoo.so.dev, it says it cannot find it. It is a timestamped file in its name and there have been rebuilds locally since the commit, so new timestamped lib files were generated, but still, I would like to remove that
13:49 lpapp checked-in file from the repository. How can I do that?
13:49 snowkidind joined #git
13:49 mikeym joined #git
13:50 flobin joined #git
13:50 joki joined #git
13:50 chid joined #git
13:51 snowkidind joined #git
13:52 flobin hi everyone, I just undid a commit with git reset --soft a0cf530
13:52 flobin then deleted a file that was supposed to not be there
13:52 flobin then git add .
13:52 flobin git commit -m "some message"
13:53 flobin and then I tried to git push, but I got this
13:53 flobin error: failed to push some refs to 'https://github.com/Togethere-com/togethere.git'
13:53 flobin hint: Updates were rejected because the tip of your current branch is behind
13:53 _ikke_ That's because you are rewritingh history
13:53 flobin I understand why this happens (well, I think I do)
13:53 Dougie187 joined #git
13:53 snowkidind joined #git
13:54 _ikke_ flobin: You can use git push --force-with-lease <remote> <branch> to force it if you think it's safe enough to do
13:54 flobin my question is now: how do I push my commit?
13:54 HanClinto joined #git
13:54 AaronMT joined #git
13:55 flobin _ikke_ ah, I think it's safe, but what if it isn't?
13:55 AaronMT joined #git
13:55 lpapp flobin: the people you are working with may need to do a fresh clone.
13:55 _ikke_ A fresh clone is not necessary
13:55 flobin so far it's just me, just a personal project
13:55 _ikke_ But they need to be careful when updating from that branch
13:56 lpapp flobin: I would not worry too much then
13:56 g105b joined #git
13:56 flobin probably just git push --force-with-lease then?
13:56 vassagus joined #git
13:56 lpapp it is like marriage or being single... when you are single, you have freedom :D
13:56 veilig2000 joined #git
13:57 flobin haha, right
13:57 flobin thanks _ikke_ and lpapp
14:00 lpapp I am still interested in an answer to my question, by the way :-)
14:00 lpapp it kinda puzzles me.
14:00 flobin what is your question?
14:00 sebras joined #git
14:01 star_pro_ joined #git
14:01 lpapp ah, it is not a question to your question :)
14:01 lpapp but this was it before you joined: https://paste.kde.org/pp5fuztr9/tgxd23/raw
14:01 notebox joined #git
14:01 _ikke_ lpapp: git ls-files | grep libfoo.so.dev
14:03 lpapp _ikke_: empty output with exit code 0
14:03 _ikke_ So according to git it's not tracked
14:03 NightStrike I'm going to say this wrong, but bear with me.. I have a local checkout of "master" to which I added a commit, and I want to push it to a new branch on github so that I can do a pull request.  I tried "git push origin/newbranchname", but that didn't work.  How do I do this?
14:03 phaleth joined #git
14:03 lpapp _ikke_: :/ gitweb shows me that it is.
14:03 _ikke_ lpapp: different branch?
14:04 _ikke_ NightStrike: git push origin newbranchname
14:04 lpapp both master
14:04 diego2 joined #git
14:04 NightStrike _ikke_: that didn't work, either
14:04 _ikke_ NightStrike: git push origin master:newbranchname
14:04 NightStrike I got "error: src refspec does not match any."
14:04 NightStrike colon?  ok
14:05 NightStrike with the colon, I got an ssh error
14:05 NightStrike "could not resolve hostname origin"
14:05 mingrammer joined #git
14:06 osse NightStrike: what does 'git remote -v' say ?
14:06 nullp0inter lb, by best i mean that a. it works, and b. isnt a maintenance nightmare
14:06 lpapp _ikke_: did a branch new checkout and git ls-files lib/libfoo.so.dev does return it
14:07 NightStrike origin githubURL (fetch), followed by the same line but with (push)
14:07 nullp0inter lb, from what you said a looks to be a thumbs up
14:07 aboccia joined #git
14:07 lpapp I wonder why it is not returning in my old checkout with further development on top of that .... ?
14:07 jaafar joined #git
14:08 lpapp _ikke_: perhaps because that is the file with two symlinks?
14:08 crelix joined #git
14:08 lpapp and I git rm'd the symlinks first?
14:08 lpapp yay for in-source builds \o/
14:09 lpapp unfortunately I checked in the library with its generated symlinks using git add on Friday, so just trying to clean up..
14:09 sebras git add -p works well for tracked files for me, but what if I want to do similar things for _untracked_ files? i.e. only stage a subset of the lines in an untracked file (but leave the unstaged lines in the working directory to be commited later).
14:09 Vinnie_win joined #git
14:09 acetakwas joined #git
14:10 lpapp hmm, odd, if I remove the lib, the symlinks are removed automatically by git as it seems?
14:10 lpapp and vice versa
14:10 lpapp interesting stuff
14:10 blackwind_123 joined #git
14:10 lpapp and strangely enough, git reset --hard HEAD does not revert such a git rm on libs with symlinks
14:10 ahmedelgabri joined #git
14:11 lpapp oh, it does, I am silly, never mind, but still, it seems that git cleans up automatically
14:11 lpapp if you remove a checked in library, it will remove its symlinks to it as well at git rm time
14:12 osse NightStrike: and the url is valid?
14:12 NightStrike yeah
14:12 osse retry with GIT_TRACE=1 git push ...
14:13 Peuc joined #git
14:13 NightStrike with a space, colon, or slash between origin and the branch name?
14:13 lpapp left #git
14:13 NightStrike osse: ^^
14:14 osse NightStrike: space
14:14 NightStrike I should note that the branch I'm trying to push to doesn't exist yet on github
14:14 NightStrike I was trying to push to push to a new branch there
14:14 osse the problem now is that git doesn't find origin for some reason
14:15 SporkWitch NightMonkey: push -u, IIRC
14:15 Spacew00t joined #git
14:15 vassagus joined #git
14:16 catag87 joined #git
14:16 NightStrike SporkWitch: was that meant for me?
14:17 SporkWitch probably
14:17 SporkWitch come up with a more unique name :P
14:17 anuxivm joined #git
14:18 NightStrike hah
14:18 NightStrike git push -u has same result
14:20 nilg joined #git
14:21 GodGinrai joined #git
14:22 hahuang61 joined #git
14:23 fuchstronaut joined #git
14:23 davisonio joined #git
14:24 csd_ joined #git
14:24 catag87 joined #git
14:25 robinsmidsrod joined #git
14:25 airdisa joined #git
14:26 zacts joined #git
14:27 fahadash joined #git
14:28 Peuc joined #git
14:28 nettoweb joined #git
14:29 shaggycat joined #git
14:29 catag87 joined #git
14:32 BSaboia joined #git
14:32 kutsuya joined #git
14:32 cdown joined #git
14:33 aard_ joined #git
14:36 Manuel15 joined #git
14:36 Manuel15 joined #git
14:36 skapunker joined #git
14:36 vassagus_ joined #git
14:36 zeroed joined #git
14:37 catag87 joined #git
14:38 hashpuppy joined #git
14:38 b1tchcakes joined #git
14:40 gfixler1 joined #git
14:40 Sonderblade joined #git
14:40 whallz joined #git
14:41 whallz hello, how can i list only deleted files of git status's output?
14:42 whallz sorry, dumb question, i piped to grep :)
14:43 bremner try "git diff --diff-filter=D"
14:44 daniel_rose joined #git
14:45 csd_ joined #git
14:47 jason237 joined #git
14:47 charlesr joined #git
14:48 osse --name-only
14:49 sangy joined #git
14:50 shinnya joined #git
14:51 cyan__ joined #git
14:51 daniel_rose joined #git
14:54 Darren_ joined #git
14:54 zeroed joined #git
14:54 zivester joined #git
14:56 jceb joined #git
14:57 diego2 joined #git
14:57 diego2 joined #git
14:58 dtrott joined #git
15:00 mingrammer joined #git
15:03 Gitzilla joined #git
15:03 clemf joined #git
15:04 eivarv joined #git
15:04 Eryn_1983_FL 6/join #zfs
15:06 Sonderblade joined #git
15:06 slugmandrew joined #git
15:09 johnmilton joined #git
15:12 roxtrongo joined #git
15:14 bittin joined #git
15:14 dmto joined #git
15:14 johnmilton joined #git
15:14 nettoweb joined #git
15:16 irco joined #git
15:22 abyx joined #git
15:22 tyreld joined #git
15:23 fuchstronaut joined #git
15:27 foist joined #git
15:27 rubyonrailed joined #git
15:28 Electrometro joined #git
15:28 nullp0inter i followed this guide https://www.digitalocean.com/commun​ity/tutorials/how-to-set-up-automat​ic-deployment-with-git-with-a-vps but i realize now i want to also use bitbucket for the production code. i want to be able to go from beta to bitbucket, and have my production repo pull from bitbucket...how is that configured
15:28 Derperperd joined #git
15:29 Guest5_ joined #git
15:33 irco joined #git
15:37 alansaul joined #git
15:39 dtrott joined #git
15:39 rorro joined #git
15:41 clemf joined #git
15:41 Faylite_ joined #git
15:43 eivarv joined #git
15:44 ResidentBiscuit joined #git
15:44 LeBlaaanc joined #git
15:45 ahmedelgabri joined #git
15:45 macivy joined #git
15:45 nikivi joined #git
15:46 kpease joined #git
15:46 a_thakur joined #git
15:47 YourTheBombNow23 joined #git
15:48 mkoskar joined #git
15:50 unbalancedparen joined #git
15:50 rubyonrailed joined #git
15:51 acetakwas joined #git
15:52 reactormonk Can I create a merge commit where I ignore commits on one side?
15:52 Junior joined #git
15:52 kutsuya joined #git
15:53 _ikke_ That's not a merge commit :P
15:53 _ikke_ Do you want to ignore the changes from one side?
15:53 Rodya_ joined #git
15:53 axl_ joined #git
15:54 unbalancedparen joined #git
15:54 reactormonk One specific commit, yes.
15:55 _ikke_ When git merges, it does not look at all the commits, it just looks at the end result
15:56 reactormonk Any way to cherry-pick a commit and create a revert at the same time?
15:57 LionsMane joined #git
15:57 reactormonk Got a commit on another branch that I want to ignore, but not force-push to kill it.
15:57 AaronMT joined #git
15:59 Derperperd joined #git
16:00 p4trix joined #git
16:00 yates joined #git
16:01 yates is there a way to get a list of only modified files in a "git status" command?
16:01 afuentes joined #git
16:01 yates i.e, exclude untracked files
16:03 osse reactormonk: do a merge, do the revert, squash the revert into the merge
16:03 osse git merge foo; git revert --no-commit abc1234; git commit --amend
16:04 a_thakur_ joined #git
16:04 thiago_ joined #git
16:06 livingstn joined #git
16:08 rubyonrailed joined #git
16:08 snits joined #git
16:09 constant joined #git
16:10 l4v2 joined #git
16:10 LeMike joined #git
16:17 shiriru joined #git
16:19 shiriru joined #git
16:21 vassagus joined #git
16:22 dhollinger joined #git
16:23 TomyLobo joined #git
16:23 hahuang61 joined #git
16:24 daniel_rose joined #git
16:24 fuchstronaut joined #git
16:28 venmx joined #git
16:29 dk0r joined #git
16:29 Derperperd joined #git
16:30 dk0r Can't seem to ignore specific dir's in my ~/.gitignore
16:30 dk0r Can't I just add `node_modules/*` to my .gitignore to ignore the contents of the node_modules directory?
16:31 ReenignE osse: better to separate those commands with && instead of ; in case one fails
16:31 ReenignE dk0r: the entire node modules directory is often ignored 'node_modules'
16:32 ReenignE dk0r: if it's already being tracked you'll have to remove it 'git rm --cached node_modules'
16:32 zacts joined #git
16:32 dk0r ReenignE, I'm running Github Desktop 3.2 on Windows and the node_modules dir is always included
16:32 ReenignE & let others know when they do a pull the delete will be merged
16:33 dk0r ReenignE, is there any reason node_modules dir would be tracked?
16:33 gechr joined #git
16:33 ReenignE it's generally bad practice
16:33 ReenignE npm install generates it
16:33 ReenignE a reason it would be tracked is due to it not being in the gitignore
16:34 dk0r I'll try starting again by deleting the .git dir, .gitattribs and .gitignore
16:34 gechr joined #git
16:34 ReenignE why?
16:35 ReenignE Simple matter to include 'node_modules' in the repo level .gitignore
16:36 ReenignE If it still informs you of changes, then it's likely already tracked. Which can be resolved by staging the deletion of the directory.
16:36 anuxivm joined #git
16:36 ReenignE After that has been done, changes to it will be ignored.
16:36 Guest5 joined #git
16:36 dk0r did as you suggested ReenignE --thanks
16:36 subhojit777 joined #git
16:38 robotroll joined #git
16:38 ReenignE No problem, might want to let others know they may have to 'npm install' after they pull down the commit if it was removed.
16:38 dk0r ReenignE, of course. Thanks man!
16:40 dk0r ReenignE, we usually indicate this by having an empty node_modules dir in the git repo.
16:40 dk0r however.. when I add `node_modules/*` to my .gitignore, it is also ignoring the directory, not just the files
16:41 dk0r How would I go about including an empty node_modules directory via .gitignore?
16:41 ReenignE place a .gitignore inside the node_modules directory
16:41 ReenignE and commit it
16:41 ReenignE have it ignore everything in that directory except itself
16:41 SteffanW joined #git
16:42 ReenignE That's if you actually want to keep the node_modules directory tracked at least.
16:42 dk0r oh, I see. Seems I was erroneously trying to do it from the repo level. How do you exclude itself? !.gitignore node_modules
16:42 a_thakur joined #git
16:43 daniel_rose joined #git
16:43 ReenignE The .gitignore inside the directory would be two lines
16:43 ReenignE *
16:43 ReenignE !.gitignore
16:44 morenoh149 joined #git
16:44 ReenignE git will not track empty directories, tracking the .gitignore inside that directory is about as close as you can get.
16:46 nitric joined #git
16:47 daniel_rose joined #git
16:48 mingrammer joined #git
16:49 gmh joined #git
16:49 dk0r ReenignE, ty. I tried adding npm-debug.log* to my repo level .gitignore and then doing `git rm --cached npm-debug.log.1562832053` but it's still being included in my commits. Any ideas?
16:50 abyx joined #git
16:50 vassagus joined #git
16:50 ReenignE dk0r: did you commit the deletion and update to the gitignore?
16:50 a_thakur joined #git
16:51 ReenignE 'git rm --cached' stages the deletion of a file for commit without removing it from your working files
16:51 ReenignE needs to be committed from there
16:51 dk0r Yes. The repo lvl .gitignore shown in github.com includes my npm-debug.log* line
16:52 dk0r http://i.imgur.com/H2zz3oV.png
16:52 ReenignE Was the deletion also committed?
16:52 rwb joined #git
16:52 daynaskully joined #git
16:52 davisonio joined #git
16:53 HoierM joined #git
16:53 dk0r I don't see anything in my commits about a deletion --so, no?
16:53 ReenignE That would be why.
16:54 ReenignE .gitignore is not considered for files that are already tracked. They need to be 'deleted' and that deletion committed first. Using the --cached option for 'git rm' stages a deletion without actually removing the file.
16:54 a_thakur_ joined #git
16:54 ReenignE After running git rm, you should notice you have a deletion staged.
16:54 ReenignE That will need to be committed.
16:54 dk0r Yeah, I notice it and I get this: http://i.imgur.com/Apyibgz.png
16:55 ReenignE Do you have the git shell available?
16:55 dk0r yep
16:56 ReenignE What's the output from 'git status -sb'
16:56 dk0r ## master...origin/master
16:56 dk0r D  npm-debug.log.1562832053
16:56 ReenignE git commit -m 'Remove npm-debug.log'
16:57 dk0r error: pathspec 'npm-debug.log'' did not match any file(s) known to git.
16:58 ReenignE Oh sorry you said Windows right?
16:58 ReenignE git commit -m "Remove npm-debug.log"
16:58 ReenignE Think double quotes are necessary on Windows.
16:59 dk0r Actually, I was in my own shell, sorry about that.
16:59 dk0r Opened gitshell specifically now
17:00 artemisart joined #git
17:00 dk0r Here's what I get from the remove using ' (not ")
17:00 dk0r 1 file changed, 0 insertions(+), 0 deletions(-)
17:00 dk0r delete mode 100644 npm-debug.log.1562832053
17:00 dreiss joined #git
17:00 eivarv joined #git
17:00 ReenignE That should be fine.
17:00 dk0r file is still present on github.com
17:00 paul424 joined #git
17:01 ReenignE No special characters so they weren't necessary, just good practice to always quote.
17:01 ReenignE You'll need to push the commit now.
17:01 madewokherd joined #git
17:01 dk0r can I do that on the windows gui client, or..?
17:01 ReenignE Your local master is tracking origin/master
17:01 ReenignE git push
17:02 dk0r ah, okay great --ty
17:02 dk0r You work here or?
17:02 ahmed_elgabri joined #git
17:02 ReenignE I just idle in this channel and help out on every now and then =)
17:02 SporkWitch dk0r: if you're on windows or mac, you might also look into sourcetree; lot of value-added features (GUI git interface), and you can use it with your own repos, github, and of course it integrates very well with bitbucket (since it's the same company)
17:03 SporkWitch (it's so nice even us linux CLI geeks have been begging them for years to port it to linux)
17:03 ReenignE SporkWitch: ever use tig?
17:03 SporkWitch ReenignE: i've not; comparable?
17:03 ReenignE https://github.com/jonas/tig
17:04 dk0r thanks again ReenignE. Much appreciated
17:04 SporkWitch so, not comparable :P lol
17:04 ReenignE SporkWitch: It's a cli curses based interface heh
17:04 skered left #git
17:05 SporkWitch ReenignE: what does it actually bring to the table? i honestly don't have a problem using the CLI, i just like the added features sourcetree brings to the table lol
17:05 ReenignE Easily extensible, haven't found myself wanting a gui for anything except image diffs.
17:05 thiago_ joined #git
17:05 ReenignE The way I see it, SourceTree misses a lot of a features that cli git has.
17:05 ReenignE Like bisect (afair)
17:05 SporkWitch i've not used bisect, so i wouldn't have noticed.
17:05 jimi_ joined #git
17:05 t0by joined #git
17:06 ReenignE A cli interface also works on remote headless servers which I've managed repos on from time to time
17:06 SporkWitch and yeah, you're almost always going to sacrifice something with a GUI, it's more the convenience and the excellent visualization tools for keeping track of what's going on
17:06 ReenignE Yeah, guis are convenient for finding options though usually lack full support
17:06 SporkWitch true, i've just ended up using normal commands in those situation
17:06 SporkWitch (re: headless)
17:07 ReenignE I've seen SourceTree and Tower(Mac) crash or take a very long time to execute a command
17:07 ReenignE Have the devs here open a terminal and use the git commands without any pain
17:07 ReenignE So there's an extra layer of potential issues that come with them as well
17:07 SporkWitch i've not had such issues, but i've also not spent a ton of time with it, since i'm usually not on windows.  When I have been, it's been excellent.
17:07 SporkWitch well that's true of anything you layer on something else
17:07 ReenignE For the most part they work well enough.
17:09 MrWoohoo joined #git
17:09 nullp0inter im trying to push to my beta repo, and am getting this Bad owner or permissions on /home/dave/.ssh/config
17:09 nullp0inter its on localhost, just not sure why its doing that
17:09 abyx joined #git
17:10 _ikke_ ls -l /home/dave/.ssh/config
17:10 _ikke_ whoami
17:11 _ikke_ ls -ld /home/dave/.ssh/config
17:11 nullp0inter _ikke_, the output looks right i think it might be the way i setup the remote
17:12 unbalancedparen joined #git
17:12 nullp0inter beta    ssh://dave@xx.xx.xx/var/beta/beta.git (fetch)
17:12 nullp0inter beta    ssh://dave@xx.xx.xx/var/beta/beta.git (push)
17:12 _ikke_ This is an ssh problem, not a git problem
17:12 nullp0inter ok
17:13 _ikke_ https://serverfault.com/questions/253313/ssh-hostn​ame-returns-bad-owner-or-permissions-on-ssh-config
17:14 satifant joined #git
17:14 eivarv joined #git
17:15 livingstn joined #git
17:15 aguimaraes1986 joined #git
17:18 Spec-Chum joined #git
17:20 tgunr joined #git
17:21 snowkidind joined #git
17:22 sagerdearia joined #git
17:23 LeBlaaanc joined #git
17:23 nikivi joined #git
17:24 Literphor joined #git
17:25 fuchstronaut joined #git
17:25 jimi_ if i want to perform some action on the files that I am adding to commit, like lint check them, etc... that would be a hook right? a pre-commit hook?
17:26 _ikke_ right
17:26 _ikke_ that runs right before the commit is made
17:27 jimi_ ok perfect, i found one in .git/hooks/pre-commit.sample
17:27 jimi_ Going to work on making one to run the php files through a linter for syntax, and against a ruleset for style etc
17:29 SporkWitch jimi_: don't forget to kill all those evil tabs people like to use instead of spaces
17:29 jimi_ SporkWitch, good idea :)
17:29 SporkWitch and UNIX line endings
17:29 SporkWitch ftw
17:30 abyx joined #git
17:30 joki joined #git
17:30 garethdaine joined #git
17:31 Darren_ joined #git
17:32 kaldoran joined #git
17:32 vuoto joined #git
17:37 GodGinrai "evil tabs"
17:37 energizer joined #git
17:38 diegoviola joined #git
17:38 _ikke_ expand command fwt
17:38 _ikke_ ftw
17:40 GodGinrai why even expand?  Tabs allow better depth control in your editor of choice
17:40 FuzzySockets joined #git
17:41 Dougie187 GodGinrai: because otherwise people mix tabs and spaces, and the formatting of your code is wrong,
17:41 SporkWitch ^
17:41 GodGinrai Dougie187: so why not just make rules preventing them from mixing tabs and spaces?
17:41 GodGinrai No need to outlaw tabs for that reason.
17:41 Dougie187 Because no one follows those.
17:41 SporkWitch different editors (and settings) will display tabs differently, resulting in inconsistent code layouts (ESPECIALLY with mixed tabs and spaces).  Until truly elastic tabs are invented (there have been attempts over the years; no one supports them), spaces are preferable
17:41 vjacob joined #git
17:42 crayon joined #git
17:42 SporkWitch GodGinrai: because what is readable with the tab settings YOU wrote it in, is NOT readable with the tab settings in MY editor
17:42 GodGinrai Dougie187: that's why hooks are great.  They let you enforce rules.
17:42 GodGinrai SporkWitch: it is if I am not mixing tabs and spaces.
17:42 mingrammer joined #git
17:42 SporkWitch GodGinrai: so sure, you prevented mixed tabs and spaces, but because you're allowing tabs you've now caused the code to display differently to each person
17:42 nedbat GodGinrai: seems like people suggest that tabs let you use personal preference for width, but i'm not sure i've ever seen it done.
17:43 nedbat GodGinrai: and how do you set a rule about line length when people can change the tab width? Or don't you?
17:43 Dougie187 I like tabs because you get to chose the format of them, but even to make things look pretty with tabs you have to mix with spaces too
17:43 nedbat GodGinrai: and all the tools outside your control are going to display the tabs as 8 spaces anyway.
17:43 GodGinrai SporkWitch: Instead of forcing it to display a way that is harder to read for everyone whether they like it or not.  Got it.
17:44 Dougie187 How is it harder to read?
17:44 SporkWitch GodGinrai: no, instead we're enforcing a standard convention and consistent layout.  4 spaces is always 4 spaces.
17:44 GodGinrai nedbat: The whole reason to use tabs is so that people can use the tab width they want.
17:44 nedbat GodGinrai: until they look at a diff on github, and then they get 8 spaces, tough luck.
17:45 nedbat GodGinrai: do you have a rule about line-length limits?
17:45 GodGinrai nedbat: github supports modelines
17:45 SporkWitch all you're really doing with tabs is forcing me to use the tab width YOU want, if i want the code to appear in a sane layout
17:45 SporkWitch much easier to just enforce it with spaces
17:45 Dougie187 And I don't understand how spaces are harder to read.
17:46 GodGinrai Dougie187: When you are in a group that uses 2 spaces for indention, it gets really hard to read.
17:46 Dougie187 But even if you want to use tabs, you almost always need to mix in spaces.
17:46 Dougie187 GodGinrai: So, agree on using 4
17:46 LeBlaaanc joined #git
17:46 GodGinrai Dougie187: Some people prefer 2 so that they can see deeper.
17:46 GodGinrai That's my point
17:46 GodGinrai with tabs
17:46 GodGinrai those who want 2 can have 2
17:46 Dougie187 Either way, you need to have a format.
17:46 GodGinrai and those who want 4 can have 4
17:47 GodGinrai either way, it takes up 1 tab
17:47 Dougie187 With tabs, you'll use spaces too, so you'll have mixed.
17:47 GodGinrai Dougie187: you won't use spaces for indention
17:47 Dougie187 You will though
17:47 Jon28 joined #git
17:47 GodGinrai No
17:47 GodGinrai you won't
17:47 SporkWitch GodGinrai: so appeal for 4.  2 would bug me too, as 2 spaces (# ) is where my comments are aligned.  8 wastes too much space (stupid linux kernel...).  4 is a lovely compromise, and is pretty much standard outside linux kernel dev
17:47 Dougie187 You need to use spaces to line things up nicely when they are smaller than tabs
17:47 SporkWitch ^
17:47 GodGinrai Example?
17:48 Dougie187 For example, you'll use 4 spaces for a tab, but eventually you'll have a variable name that won't line up cleanly the way you want, so you'll have to use spaces.
17:48 Dougie187 Let me write something, one sec.
17:49 Dougie187 my_var_1 = other_var_1 + other_var_2
17:49 Dougie187 * other_var_3;
17:49 ojacobson the only tab policy that both works and allows tabs is "tabs for indentation, spaces for alignment"
17:49 nedbat GodGinrai: gist.github.com doesnt seem to support modelines: https://gist.github.com/nedbat/3​b34653aaf90d5571ddd2684d5af0b68  can you show me where you see it?
17:49 ojacobson the second line of that expression has the same indentation as the first (thus the same # of tabs)
17:49 ojacobson the remaining whitespace before the * would be spaces
17:49 SporkWitch def meowLearnToSpace():
17:49 SporkWitch lookAtMyList = [meow
17:49 SporkWitch meow #tab only
17:49 SporkWitch meow #space+tab
17:49 SporkWitch meow #tab only
17:49 Dougie187 But, if you change the width of a tab, the format is off.
17:50 GodGinrai Dougie187: I guess that's where you and I would code differently.  I don't focus on lining up the operators with the first line.  Only for the subsequent (tabbed-in) lines
17:50 Dougie187 My expression is assuming tabs are 4 spaces, btw.
17:50 SporkWitch as an aside, i've just uncovered an issue with kvirc, and must now investigate why alignment in the text entry area differs from the chat frame
17:50 Dougie187 :P
17:50 Dougie187 GodGinrai: There are projects in existence that require you to line up lines like that.
17:50 Dougie187 And generally, that is the preferred way to align operators.
17:51 SporkWitch (as i was entering that stuff, the first tab only is slightl left of the bracket, the space+tab is aligned, and the second tab only is a mile to the right of the other meows)
17:51 nedbat GodGinrai: a github repo also doesn't seem to support them: https://github.com/nedbat/nedb​atcom/blob/master/modeline.py
17:52 nedbat GodGinrai: maybe it needs to be a modeline in a different format?
17:52 ForeignBiscuit joined #git
17:53 jstein_ joined #git
17:55 GodGinrai Dougie187: Yes.  There are also projects that require me to use spaces instead of tabs.  All of those come down to the rules defined by whoever created the source. (Or whoever is currently maintaining it)
17:55 Dougie187 GodGinrai: Yes, and a very good point from this conversation is that everyone has different preferences. So the standards should reflect that and also take into account (1) what's maintainable, (2) what's easy.
17:56 nwkj86 joined #git
17:56 Dougie187 It's going to be hard to say "Everyone, code like me." So, tabs are going to be difficult.
17:56 Dougie187 (using pure tabs)
17:56 GodGinrai nedbat: looks like linguist doesn't support spacing settings for modelines yet.  You could open an issue against it. https://github.com/github/linguist
17:57 nedbat GodGinrai: what is linguist?
17:57 nedbat GodGinrai: oops, just read it, sorry.  doesn't sound like it's responsible for displaying code, maybe?
17:57 concerti joined #git
17:58 GodGinrai Well, I know that's where they determine how to syntax highlight files based on modelines: https://github.com/github/linguist/blob/​master/lib/linguist/strategy/modeline.rb
17:58 nedbat or maybe it does? it says it will read a modeline to decide how to syntax-color.
17:58 nedbat yes.
17:58 dreiss joined #git
17:58 nedbat yikes, look at that regex!
17:59 GodGinrai lol yea
17:59 GodGinrai Does ruby use Thompson's nfa or pcre?
18:00 nullp0inter i think im going to start this all over.. what is the flow for developer(local repo) -> beta(local repo) -> live(bitbucket) with a live local repo for the stable web app - i.e. which ones are bare?
18:00 _ikke_ You don't need multiple repositories
18:01 _ikke_ Ideal git would not involved in deployment at all
18:01 GodGinrai nullp0inter: the "flow" really depends on what you find best.
18:01 ki0 joined #git
18:01 Guest5 joined #git
18:01 nullp0inter _ikke_, how woudl it work in your pov
18:02 GodGinrai git (developer) <--> bare git repo (server)
18:03 GodGinrai \___ deploy tools
18:03 GodGinrai that's how I see it working
18:03 nullp0inter how do i fit in bitbucket into that
18:03 GodGinrai bitbucket is your server for storing your source code, isn't it?
18:04 nullp0inter yeah
18:04 GodGinrai then that is the "bare git repo (server)"
18:04 nullp0inter how is it bare if it will contain the files
18:05 GodGinrai Sure, bitbucket might have the files checked out
18:05 GodGinrai but it still is taking the same place the "bare" repo would
18:05 davisonio joined #git
18:06 nullp0inter GodGinrai, i was trying to setup a beta part so we can have a beta version of the web app to test changes
18:06 nullp0inter instead of going live right away
18:06 artemisart1 joined #git
18:06 rpd nullp0inter: why not just have two branches: say "beta" and "live"
18:06 rpd beta.whatever has the "beta" branch checked out and the live.whatever has the "live" branch checked out
18:07 GodGinrai ^ this
18:07 LeMike joined #git
18:07 GodGinrai although there's no need to have the branches "checked out"
18:07 alansaul joined #git
18:07 GodGinrai since the deploy tools should take the correct files and put them on the beta server for testing
18:07 GodGinrai as _ikke_ said, don't use git for deployment
18:07 GodGinrai just for storing the code
18:08 nullp0inter GodGinrai, what deploy tool would work well for this kind of setup?
18:08 GodGinrai nullp0inter: I don't know the specifics of your setup, but it could be something as simple as scp
18:08 GodGinrai and some scripts
18:09 nullp0inter ok cool ill have to look into this some more, ty
18:09 nullp0inter ty rpd too
18:12 star_prone joined #git
18:16 SporkWitch nullp0inter: were you able to fix your push issue?
18:16 SporkWitch nullp0inter: also, you should have said you were trying to set up website deployment, i could have told you what to do, i've done exactly what GodGinrai told you before lol
18:17 nullp0inter SporkWitch, :)
18:17 mooru joined #git
18:18 nullp0inter so when a dev commits new changes you would do something like git push bitbucket beta
18:18 nullp0inter and that would put it into the beta branch?
18:18 nullp0inter or would you default it to beta for all devs
18:20 SporkWitch nullp0inter: what you'd normally do is a relase-based branching model on the repo.  You'd have repo/releases/stable repo/releases/testing repo/releases/beta (each is a named branch)
18:21 nullp0inter whats the difference between testing and beta
18:22 GodGinrai SporkWitch: namespacing your branches w/ "releases/" or am I misreading that?
18:22 SporkWitch GodGinrai: yup
18:22 SporkWitch GodGinrai: repo/releases/release-branch, repo/hotfixes/datetimestamp is the naming convention i've found works well for me
18:23 GodGinrai I see
18:23 SporkWitch nullp0inter: preference.  I like having a "development" branch that isn't guaranteed to be in a working state
18:23 SporkWitch nullp0inter: testing would be when you THINK it's ready, and need to test it
18:23 nullp0inter SporkWitch, got ya
18:23 SporkWitch nullp0inter: and then production, for once you're confident it's ready for public use and consumption
18:23 nullp0inter SporkWitch, I just made a beta branch off the master, is that right
18:23 SporkWitch that's fine, it's all a matter of preference and workflow
18:24 kpease joined #git
18:24 nullp0inter im pretty new to branches
18:24 StuartMI joined #git
18:24 hahuang61 joined #git
18:24 GodGinrai nullp0inter: they're awesome
18:24 LeBlaaanc joined #git
18:24 nullp0inter ive just done local git using a bare repo, first time im using a hosted git too
18:25 SporkWitch so in your model you just have your testing and production branches.  That makes it a little easier.  So you'd just set a post-commit hook on the repo that looks to see if someone pushed a commit to the beta branch.  When they do, you can have it automatically check out that commit into a working directory of your choice (presumably the testing portion of your webserver setup)
18:25 mutex joined #git
18:25 nullp0inter so i made this beta branch off the master, if i check out the beta branch, do i do that where my local repo is?
18:26 nullp0inter or do i need a new folder
18:26 SporkWitch the TRICKY bit is how to get the live/production one (in this case master) to auto-deploy; when you merge a pull request in bitbucket, it kind of "cheats" some stuff, and doesn't trigger git hooks normally.  You'd need to use a deployment solution, such as bamboo (which integrates nicely, like all the other atlassian products)
18:26 fuchstronaut joined #git
18:26 SporkWitch nullp0inter: you would clone the repository, and check out that branch locally
18:26 nullp0inter so one repo is for beta and one for the master?
18:26 SporkWitch do your work, commit locally, and then push to the same branch on the repo
18:26 nullp0inter master being the production
18:26 SporkWitch no, single repository for all of it
18:27 SporkWitch you're using branches, not respositories, to manage the differences.
18:27 nullp0inter so right now im in my home dir from which i pushed index.php to test. that worked to get to master. then i made a beta branch. now would i stay in the same directory and checkout the beta branch?
18:28 SporkWitch repositories are for projects (or on a large enough scale, logical subdivisions OF that project).  Branches are for releases (alpha, beta, golden-master), hotfixes, feature additions, etc. that apply to THAT project
18:28 GodGinrai SporkWitch: bitbucket breaks git-hooks? :(
18:29 SporkWitch GodGinrai: yes and no.  It doesn't provide any convenient interface to them (though there are free plugins you can use, i've had mixed results).  You can manually navigate to the actual repo directories and set up hooks, but then you run into what I mentioned about how bitbucket handles merging pull requests.
18:29 SporkWitch GodGinrai: it doesn't actually trigger the post-commit hook because of the "magic" it's doing behind the scenes
18:29 Guest5 joined #git
18:29 GodGinrai ugh
18:30 SporkWitch these are the "official" hooks supported in it: https://confluence.atlassian.com/bitbuckets​erver/using-repository-hooks-776639836.html
18:31 SporkWitch but like i said, there are plugins to offer different and more extensive functionality.  Realistically, if you're using bitbucket at scale, you've probably got a continuous deployment solution, such as bamboo, in place
18:31 SporkWitch (or if not, you'll probably want to) as it makes it easier to manage such things, and integrates nicely
18:31 mar77i_ joined #git
18:31 kpease joined #git
18:31 SporkWitch don't let the $10 small team license price fool you, we're talking about enterprise software designed for large scales
18:32 SporkWitch honestly, in spite of the small little issues like that, i love their products, i just wish they weren't java; i can't afford a supercomputer to host the stuff on
18:32 GodGinrai lol
18:33 nullp0inter SporkWitch, so Im in my webapp local dir which is a clone of my bitbucket repo...I ran git fetch && git checkout beta and it says "Branch beta set up to track remote branch beta from origin.
18:33 nullp0inter Switched to a new branch 'beta'"
18:33 nullp0inter what now?
18:33 GodGinrai SporkWitch: they do that so you pay them money to host it on their supercomputers
18:33 SporkWitch GodGinrai: actually, their hosted options for small team are free
18:33 SporkWitch GodGinrai: so that logic falls apart
18:34 GodGinrai SporkWitch: including bamboo?
18:35 SporkWitch GodGinrai: it just comes down to the fact that each of hte products is its own self-contained world in its own java bucket.  That's why the resource reqs are so high: java eats memory like crazy to begin with, and it ALWAYS takes that memory, even when it's not being used.  Since each product hosts its own webserver, it works out to around 2GB per product, which adds up quickly (and annoyingly).
18:35 SporkWitch i'm not sure about bamboo, as i hadn't played around with it
18:35 SporkWitch i don't have the money for apartment rent right now, let alone a server with decent RAM, so i'm not currently running them; i've worked with bitbucket and jira in the past, though, and absolutely love them
18:36 GodGinrai Everything I'm seeing indicates money involved
18:36 * SporkWitch still needs to find someplace to sleep tonight...
18:36 GodGinrai unless you were *just* talking about bitbucket
18:36 kpease joined #git
18:36 SporkWitch GodGinrai: was talking about bitbucket in particular; hipchat is also free for small team
18:37 SporkWitch (hipchat annoys me, it's the one product they don't offer as a purchase if you host yourself; it's actually MORE than going through them for some asinine reason, as there's a monthly sub for the lowest self-hosted hipchat package, but the lowest is free if they host it)
18:37 GodGinrai looks like JIRA is $10 for small teams
18:38 SporkWitch GodGinrai: not sure about jira.  I mostly care about the self-hosted prices, because i like and prefer running things myself
18:38 SporkWitch GodGinrai: looks right
18:40 SporkWitch dunno, my "stuff to buy that isn't life-essential" list right now is pretty much: yubikey neo, the one-time fees for the atlassian stuff, and then subscription on a few droplets (or maybe linodes; they seem to offer more RAM for the same price, but a much worse interface, and far less granularity or value-added features) to host the stuff lol
18:42 kpease joined #git
18:42 l4v2 joined #git
18:42 SporkWitch (and i wish i could cut the yubikey out, but my feitian card doesn't support openpgp; if they offered that yubikey in a credit-card formfactor, and standard contact + NFC on it, i'd be in heaven; prefer cards and readers to a USB plug)
18:42 Gitzilla joined #git
18:42 davisonio joined #git
18:43 SporkWitch (never mind that the price is insane; 50 bucks for that thing? my smartcard was 10 euro! and does nearly everything i want the yubikey for! lol)
18:44 cbreak USB Stick + 1Password + GPG Key + SSH Key + ... :D
18:46 SporkWitch cbreak: i don't need (or really want) the storage, PGP+SSH is covered in openpgp support, so is auth to any number of other things.  OTP could be neat, but i've got authenticator on my phone for that.
18:47 kpease joined #git
18:47 SporkWitch cbreak: feature-wise, the only thing i'm missing with this feitian card is openpgp, and that's what i find myself using more these days.  when i was in the military, the pkcs11 card made sense, since everything played nice with both my military CAC and my personal card.
18:48 SporkWitch cbreak: the biggest thing that bugs me though is just the formfactor; i REALLY don't like it as a keyfobe.  I much prefer a card in my wallet, and a reader on the machine (and i've got a little fold-out reader that's half the size of a 5-stick pack of gum)
18:48 vuoto joined #git
18:49 SporkWitch (the reader requirement also provides an extra barrier, though only a minor one; if i lose it or the card is stolen, they still need to find a reader for it; a usb device they just plug in.  it's more delays that i can use to notice the issue and revoke credentials)
18:49 LiohAu joined #git
18:50 vuoto_ joined #git
18:51 cr3 joined #git
18:51 nettoweb joined #git
18:51 zacts joined #git
18:52 cr3 hi folks, after I run `git rev-list <commit>`, is `git log -n1 --pretty=format:%H` the right command to run to get the <commit> string back?
18:54 daniel_rose joined #git
18:54 BigCalhoun joined #git
18:54 jussx joined #git
18:55 Orion3k joined #git
18:55 BigCalhoun Hi, I’m new to Git and building from source and have a basic question
18:55 SporkWitch BigCalhoun: don't ask to ask
18:55 mooru joined #git
18:56 BigCalhoun When changes are made to a project, are they automatically synced to your machine or do you have to resync/refresh/etc?
18:56 _ikke_ You have to explicitly commit, push, fetch, merge, etc
18:56 _ikke_ git is *not* a file sync program
18:57 flobin joined #git
18:57 BigCalhoun OK, makes sense. Thank you.
18:57 SporkWitch _ikke_: though it can be used (very inefficiently) as one :P lol
18:58 SporkWitch (sync'ing config files between personal machines is a not-uncommon use)
18:58 nkhoit joined #git
18:58 nullp0inter i did a git fetch && git checkout beta but i dont get what happens now?
18:58 _ikke_ Yes, but there are better alternatives
18:58 nullp0inter beta is my beta branch
18:58 rj1 joined #git
18:59 SporkWitch _ikke_: yup
19:00 crayon joined #git
19:00 nullp0inter i made a change but git push beta tells me about matching vs simple
19:02 nullp0inter ahh git push origin beta worked
19:03 _ikke_ nullp0inter: You need to set push.default config
19:04 khmarbaise joined #git
19:09 hg joined #git
19:10 nullp0inter what are the cons to using push.default matching...code you dont want to merge gets merged?
19:10 _ikke_ It's more about what gets pushed to the remote
19:10 _ikke_ (pushing never merges)
19:10 GodGinrai nullp0inter: technically, if your beta branch is tracking the server's beta branch, then you don't need args
19:11 nullp0inter GodGinrai, im a little slow on the terminology, what do you mean by tracking
19:12 GodGinrai 14:33 < nullp0inter> SporkWitch, so Im in my webapp local dir which is a clone of my bitbucket repo...I ran git fetch && git checkout beta and it says "Branch beta set up to track remote branch beta from origin.
19:12 _ikke_ GodGinrai: That entirely depends on the push.default setting
19:12 GodGinrai _ikke_: yes, but by default, doesn't push.default push to the tracking branch?
19:13 GodGinrai ugh so much same word usage
19:13 Literphor joined #git
19:13 eijk joined #git
19:13 SporkWitch GodGinrai: what's up?
19:13 _ikke_ GodGinrai: Only if the name is the same
19:14 GodGinrai SporkWitch: you got highlighted in a quote I pasted
19:14 SporkWitch Ah
19:14 GodGinrai _ikke_: which it is
19:14 GodGinrai see the pasted quote above
19:15 eijk joined #git
19:15 govg joined #git
19:17 hg__ joined #git
19:17 ForeignBiscuit joined #git
19:18 hg__ left #git
19:19 cyan__ joined #git
19:20 johnmilton joined #git
19:21 alansaul joined #git
19:21 clemf joined #git
19:24 nullp0inter thanks for all the help so far everyone you guys are awesome
19:26 mablae_ joined #git
19:26 hahuang65 joined #git
19:26 fuchstronaut joined #git
19:27 hg left #git
19:27 hg___ joined #git
19:30 Guest93616 joined #git
19:30 _ikke_ GodGinrai: Just trying to make it clear what the effect of the setting is
19:30 mingrammer joined #git
19:32 GodGinrai I see
19:33 _ikke_ Most settings look alike with small differences
19:35 LiohAu joined #git
19:36 Sample joined #git
19:36 Sample joined #git
19:37 KevinMGranger joined #git
19:41 dmpeters63 joined #git
19:43 energizer joined #git
19:43 mooru joined #git
19:46 Rodya_ joined #git
19:47 m0viefreak joined #git
19:51 BrianBlaze420 joined #git
19:51 paul424 joined #git
19:51 clemf joined #git
19:52 vassagus joined #git
19:55 ShapeShifter499 joined #git
19:55 grayjoc joined #git
19:57 dodobrain joined #git
19:57 MikeChernev joined #git
19:58 veilig2000 joined #git
19:59 LionsMane joined #git
20:00 vuoto joined #git
20:01 cdg joined #git
20:02 LeMike joined #git
20:03 superchar joined #git
20:10 stamina joined #git
20:12 vuoto joined #git
20:13 grayjoc joined #git
20:14 bongjovi joined #git
20:14 ahmed_elgabri joined #git
20:15 a_ joined #git
20:18 BSAlb joined #git
20:18 HappyLoa- joined #git
20:19 mooru joined #git
20:22 aidalgol joined #git
20:22 Darcidride joined #git
20:22 Neo joined #git
20:25 hahuang61 joined #git
20:25 mingrammer joined #git
20:25 DJ_Spies joined #git
20:27 fuchstronaut joined #git
20:28 mariuscc joined #git
20:30 FuzzySockets joined #git
20:35 clemf joined #git
20:37 Samual joined #git
20:38 Derperperd joined #git
20:39 BadApe joined #git
20:39 cyan__ joined #git
20:40 BadApe hello, i was wondering if there is a good git-lfs packaged server i could easily deploy
20:40 nnyk joined #git
20:42 * SporkWitch is a bitbucket fanboy
20:42 SporkWitch BadApe: ^
20:42 docnvk joined #git
20:43 BadApe i was going to use bitbucket but my game assets are huge and require more space than bitbucket supports for a sensible price
20:43 SporkWitch it's basically self-contained, so initial installation and setup is fairly trivial.  A little extra work if you want to configure SSL (and you should), particularly with your own keys (which, again, you should; letsencrypt provides free certs signed by a CA that's preloaded in most browsers)
20:44 SporkWitch ah, no LFS no the $10 one-time? didn't notice that :(
20:44 ericP left #git
20:44 SporkWitch (and the price is perfectly sensible for what the product is, just not for smaller teams, heh)
20:44 rbern joined #git
20:46 UTAN_dev joined #git
20:46 BadApe it is the amount of space,
20:46 BadApe i think i would be paying $50 a month
20:46 vuoto joined #git
20:46 codeman joined #git
20:47 SporkWitch ummmm, it lists LFS support on bitbucket server
20:47 SporkWitch oh!
20:47 SporkWitch i assumed you were hosting it yourself
20:47 BadApe but you get 1gb
20:47 BadApe oh i don't have hardware to support it
20:47 SporkWitch small teams (10 people or less) is $10/yr (and you only have to keep paying if you want to keep receiving updates)
20:47 veilig2000 joined #git
20:48 BadApe yes but you need some power to push stash, maybe bitbucket is better now
20:48 SporkWitch the 2GB of RAM is the killer, yeah
20:48 SporkWitch it's because it's java based
20:48 BadApe well for a virtual machine it will be
20:49 SporkWitch i'm honestly not sure; i've never bothered looking into hosted solutions, as i prefer running things myself :(
20:49 nitric BadApe: use Gitlab, it's better than Bitbucket, especially for just self-hosting something
20:50 alansaul joined #git
20:50 feignix joined #git
20:50 SporkWitch well running it locally on a VM should be cake; if your machine has the power to be dealing with large graphic assets, it's probably got the spare RAM to boot up the VM when it's needed.
20:51 BadApe nitric, anything a little more light weight, i had a small celeron with 4gb that crashed a lot with my repo
20:52 BadApe i wonder if anyone has used gitblit
20:52 UTAN_dev_ joined #git
20:53 vassagus joined #git
20:56 MacWinner joined #git
20:57 MacWinner do you use git stash when you need to switch to another branch, but you don't want to commit your updates to your current branch?
20:57 osse MacWinner: sure
20:57 MacWinner like i'm part way through implementing something, but i need to make a quick fix on master, is the standard process to stash the updates, switch to master, fix.. switch back and pop the stash?
20:57 osse yup
20:57 osse the prototypical use case
20:57 ForeignBiscuit joined #git
20:57 MacWinner are stashes stored in a remote location or only locally?
20:58 osse local
20:58 MacWinner is it possible to have it remote too?
20:58 SporkWitch no, that's what branches are for
20:58 zacts joined #git
20:58 rpd You could just make a real commit on some other branch for that
20:58 SporkWitch stash is just that: "stashing" changes while you do something else and still get back to it
20:58 MacWinner got it, thanks!
21:02 flopsec joined #git
21:03 rubyonrailed joined #git
21:04 UTAN_dev joined #git
21:07 Samual joined #git
21:07 Samual joined #git
21:08 FroMaster joined #git
21:08 artemisart joined #git
21:09 dtrott joined #git
21:10 cd-rum joined #git
21:12 cd-rum_ joined #git
21:14 dodobrain joined #git
21:19 crayon joined #git
21:19 EvilPeng1 joined #git
21:22 ehva joined #git
21:24 Samual joined #git
21:24 constant joined #git
21:26 jimi_ joined #git
21:27 nullp0inter if i have a branch coming out of the master, and there's a change in master, how can i update my branch with those changes?
21:27 robotroll joined #git
21:27 GodGinrai nullp0inter: you merge master into your branch
21:28 nullp0inter that was easy ty
21:28 nullp0inter i saw something online about a rebase
21:28 nullp0inter but that didnt seem right
21:28 osse nullp0inter: rebase is a different way of achieving the same result
21:28 GodGinrai rebasing is another way of doing it
21:28 GodGinrai I prefer merge
21:28 fakenullie not quite the same
21:28 osse nullp0inter: makes it so that it appears as if you redid your work on top of the updated branch
21:29 osse people fight about what's best all the time
21:29 Finnoloid joined #git
21:29 osse apathy always wins
21:29 nullp0inter so if i merge, i retain any of my changes in the branch which havent been merged with master? and rebase removes my changes?
21:29 fakenullie if you branch is public you should only do merges, because rebase will break things for other people
21:29 fakenullie nullp0inter: no, neither remove your changes
21:29 GodGinrai nullp0inter: no, rebase replays your changes with the new master as the new base
21:30 SporkWitch nullp0inter: it mostly comes down to history.  a rebase would allow you to make it appear in the history as if the branch forks from master at the new commit, rather than where it ACTUALLY forked.  In a large project, with lots of contributers, this can be desirable to keep the history on master clean and readable.  In a small or personal project, this is less important, and a merge is a LOT
21:30 GodGinrai nullp0inter: merge just makes a new commit that is the merger of the 2 branches
21:30 SporkWitch easier (and less dangerous and destructive) than a rebase
21:30 nullp0inter thanks GodGinrai that makes sense
21:30 GodGinrai np
21:30 nullp0inter ill stick with merges for now
21:30 SporkWitch nullp0inter: http://justinhileman.info/arti​cle/git-pretty/git-pretty.png
21:31 nullp0inter at quick glance, cant wait
21:31 osse nullp0inter: http://sprunge.us/OIdf
21:31 GodGinrai SporkWitch: lol @ the "do you hate them?"
21:32 SporkWitch GodGinrai: "enough to form a lynch mob?"
21:32 SporkWitch GodGinrai: those two lines are why i love that image so much.  And it's even better because it's actually a REALLY good flowchart
21:32 nitric osse: shouldn't that 3rd graph be labeled "MERGE: " rather than "REBASE: "....?
21:32 SporkWitch nullp0inter: it's a bit of a joke, but the advice in it is actually really good
21:32 nullp0inter SporkWitch, 10/10 would click on one of your links again
21:32 osse yes
21:32 nwkj86 joined #git
21:32 Dougie187 left #git
21:33 osse http://sprunge.us/WdAg
21:36 rorro joined #git
21:36 nnyk1 joined #git
21:37 nnyk1 joined #git
21:37 zacts joined #git
21:38 morenoh149 joined #git
21:39 nnyk1 joined #git
21:39 weckl joined #git
21:40 nnyk1 left #git
21:41 flobin joined #git
21:42 ki0 joined #git
21:42 xaa joined #git
21:45 superchar joined #git
21:47 mellernoia joined #git
21:48 cjohnson joined #git
21:48 abyx joined #git
21:49 cjohnson If I have an integration branch "qa" and a production branch "master", if I branch from qa, can I safely say that as long as I never directly merge qa -> master, I won't pull any unrelated stuff up to master?
21:49 cjohnson so feature-foo and feature-bar are branched from qa, when I test I merge into qa, test, and when I'm happy, I merge feature-foo into master
21:49 Guest93616 joined #git
21:49 cjohnson that for sure won't pull any of feature-bar into master right? Even if feature-bar has been merged into qa?
21:50 fakenullie if you don't merge qa into feature-foo
21:50 fakenullie or feature-bar into feature-foo
21:50 cjohnson ok so to do that I need to never ever merge qa into anything
21:50 cjohnson and yeah I wouldn't merge features into other features
21:51 SporkWitch in your model, yes, qa should always be a mirror of master, except for the feature branch currently being tested.
21:51 fakenullie well, you branch feature-foo from qa after you merged feature-bar into qa, you'll get feature-bar in feature-foo
21:51 SporkWitch at least that's how it sounds.  It feels like a bit of an odd model
21:52 a_ joined #git
21:52 cjohnson SporkWitch: what's odd? what would you recommend instead? I've got a 2 person team
21:52 cjohnson trying not to add too much complexity
21:53 fakenullie usually you don't branch feature of qa branch
21:53 fakenullie features go into master
21:54 vassagus joined #git
21:54 SporkWitch cjohnson: could just be a matter of what i'm used.  I've normally gone with the production/release/development model.  You have the "working copy" on development, with no guarantee that it works.  This is the main codebase being worked on for the next release.  All feature branches branch from here.  When we think it's ready, we merge that into testing.  Testing SHOULD be thought good, that's why
21:54 cjohnson That's what I was thinking, new guy says he does it off qa though and it's worked well for him. we're still playing with the format
21:54 rkazak joined #git
21:54 SporkWitch it's there to test: to see if it's ready for release.
21:54 SporkWitch if it has small issues, you can branch a hotfix and merge that into it.  You might end up merging testing back into dev in some circumstances.
21:55 SporkWitch once it's stable and good, merge testing into stable
21:55 SporkWitch that's how i'm used to doing things, anyway.  The closer you get to the release branch, the cleaner the history ends up, because it's hierarchical, with fewer individual changes at each step up the ladder
21:56 cfxd joined #git
21:57 cjohnson @Spork: I thought that at first but the problem is if something gets into testing that is abandoned, or say it's just not ready to go out yet, then it will go out when you merge
21:57 cjohnson that's why I was just using an "integration branch", test on that, and then for each feature that passes, merge that feature right into master
21:58 SporkWitch cjohnson: prepending characters to names will typically break highlights; i also don't have highlights on shortenings of my name.
21:58 jason237 joined #git
21:58 cjohnson I figured you were watching the window :P
21:59 ehva joined #git
21:59 SporkWitch cjohnson: you misunderstand, that kind of stuff would go on in development.  If something gets abandoned, you remove / revert it.  By the time it gets to the testing branch, it should be the "hopefully finished" product, you're just testsing it before merging it into the stable branch
21:59 GodGinrai SporkWitch: just tell him IRC isn't twitter :P
22:00 SporkWitch GodGinrai: it's a standard enough convention it warrants explanation; the twitter crack i reserve for
22:00 SporkWitch jerks, that
22:00 SporkWitch think that
22:00 SporkWitch newlines are
22:00 SporkWitch punctuation
22:00 GodGinrai lol
22:00 GodGinrai good
22:00 GodGinrai to
22:00 cjohnson my client highlights just fine with @ , has nothing to do with twitter
22:00 GodGinrai know
22:00 cjohnson I've used that on IRC for a decade
22:01 GodGinrai cjohnson: And I'm sure people have told you to stop referring to them with @ for a decade
22:01 cjohnson never once
22:01 cjohnson not a single time until now
22:01 cjohnson didn't realize it was such a trigger :P
22:01 GodGinrai lol
22:01 osse gaah why do people care so much about stuff!!
22:02 cjohnson anyway
22:02 SporkWitch cjohnson: it's just not consistent between clients, is all, that's why i say; back to your question, you can think of it like the early access week before launch; yeah, you might still have a hotfix or two, but it's basically there, there aren't going to be any significant changes (and if there are, we're talking about real HUGE blockers), it's just for
22:02 cjohnson my approach is like it is becasue I don't want to necessarily have to keep track of, and revert, every feature that doesn't make it up
22:02 cjohnson sometimes stakeholders just sit on it for 3 months
22:03 SporkWitch for seeing if it IS ready, possibly with a wider base to increase feedback.  The hope is that what you merge into testing won't need further modification, just a merge right into stable.
22:03 GodGinrai osse: well in my case, I hate it when I miss messages because people can't just use God<tab>
22:03 SporkWitch ^
22:03 osse get a better client
22:03 cjohnson ^
22:03 cjohnson hahahaha
22:03 SporkWitch that's the reason i actually take the time to explain the thing about @'s
22:04 GodGinrai that's not a client problem
22:04 cjohnson all that time wasted when you could have just fixed your client
22:04 GodGinrai the convention on IRC is <nick>, or <nick>:
22:04 cjohnson or @<nick>
22:04 GodGinrai no
22:04 SporkWitch osse: it's a trivial thing, if i wanted those things to highlight me i could make them.  it's not always desirable.  tabcompletion is present even in some WEB clients,there's no need for increasing the chances of false positives
22:04 GodGinrai that's forums and twitter
22:04 osse the only sensible thing to highlight is your nick with word boundaries at both ends
22:04 cjohnson well I've been doing it for a decade without issue, that's convention enough for me
22:04 GodGinrai osse: yes, and @ is not a word boundary
22:04 GodGinrai so that wouldn't highlight it
22:04 osse sure it is
22:05 cjohnson Anyway I think I got some good tips on my git questions despite triggering some anger about people's bad clients
22:05 osse a word is [_a-zA-Z0-9]+
22:05 cjohnson hahahaha ;P
22:05 cjohnson love you guys
22:05 SporkWitch osse: you clearly don't vim
22:05 SporkWitch osse: it may be a WORD, but it's not a word
22:05 osse hah
22:05 osse SporkWitch: ok, let me rephrase: \<nickhere\>
22:06 blackwind_123 joined #git
22:06 SporkWitch ok, i feel a little bad now, because apparently my vim config actually does consider @ to be a word boundary lol
22:07 SporkWitch osse: honestly, i think that's part of the reason that leading characters are not a thing on irc, since leading symbols are used to denote modes on a channel
22:07 osse anyway, it's only because I am a super-human selfless person who sees the entire human condition at once. war, starvation, and @ on irc
22:07 osse one of those is not like the others
22:07 cjohnson the war one
22:07 SporkWitch osse: my name isn't +sporkwitch in #freenode, it's still just sporkwitch
22:07 osse cjohnson: correct
22:09 osse SporkWitch: I'm not fond of @nick either. nick: feels more natural even from a purely lingistic point of view
22:09 GodGinrai @ is a hat.  Not a word boundary
22:09 osse but come on. who really gives a shit?
22:09 dmpeters63 joined #git
22:09 GodGinrai :P
22:10 SporkWitch GodGinrai: exactly
22:10 SporkWitch osse: again, i explained it politely, because it's NOT a big deal, the worst that'll happen is someone won't get a response.
22:10 osse I didn't say @ was a word boundary
22:10 osse oh, yes I did
22:10 SporkWitch osse: it's the use of newlines as punctuation that IS a problem lol
22:10 osse What I meant was that it's not a word character
22:10 osse SporkWitch: now that's a much bigger one
22:11 osse like
22:11 SporkWitch GodGinrai: my vim config considers it a word boundary, i had to concede that point because i made a stink about WORD vs word lol
22:11 osse i mean,
22:11 SporkWitch DIE
22:11 cjohnson I
22:11 cjohnson do
22:11 cjohnson agree
22:11 cjohnson with
22:11 cjohnson that
22:11 * cjohnson awaits kickban
22:11 GodGinrai SporkWitch: now you can tell him IRC is not twitter ;)
22:12 SporkWitch goddamn twits
22:12 SporkWitch (hey, if it's called twitter, the users should be called twits!)
22:12 GodGinrai of course
22:12 Narwhaal twitterrers *
22:13 cjohnson was kicked by bremner: sorry no ban today
22:13 GodGinrai Narwhaal: if you're really going to try to correct him.  at least make an amusing correction, like "twittors"
22:13 mingrammer joined #git
22:13 SporkWitch lol
22:13 osse sorry to burst your bubbles, but twit and twitter don't share etymololology
22:14 Narwhaal GodGinrai: it's what they be referred to in old api docs
22:14 SporkWitch osse: that is sooooo not the point lol
22:14 osse i know
22:14 jstein_ joined #git
22:14 osse you dislike twitter, I suppose
22:14 SporkWitch i do now, but that's recent; prior it was more indifference.
22:15 osse You have to stop caring about pointless things.
22:16 GodGinrai osse: I'm glad you believe twitter is pointless as well :)
22:16 GodGinrai :P
22:17 SporkWitch GodGinrai: it's not like i sit awake at night lamenting it, i just don't use it, won't interact with it, and encourage others to do the same if it comes up (I don't bring it up)
22:18 Guest93616 joined #git
22:18 cjohnson joined #git
22:18 * cjohnson shakes fist
22:19 dreiss joined #git
22:19 osse A lot of things are pointless. Like computer games. But as someone said (who said John Lennon said it, which I don't think is true): Time you enjoy wasting is not wasted time.
22:20 osse At least that's what I use as an excuse
22:20 SporkWitch osse: how dare you! *strokes the screen where his Idris-M from star citizen is listed*
22:20 SporkWitch internet spaceships are srz bznz
22:21 jkt joined #git
22:21 osse is that for people who think eve online is too mainstream?
22:21 GodGinrai lol
22:21 GodGinrai wouldn't eve be less mainstream?
22:21 ToBeCloud joined #git
22:22 osse i don't know
22:22 GodGinrai it may have more users, but EVE's play is not focused to as broad of a userbase as Star Citizen's, IMO
22:22 catag87 joined #git
22:22 osse i just went off the fact that I had heard of eve online before, but not star citizen
22:22 osse a buddy made me buy elite dangerous. have i been coaxed into a snafu?
22:22 bremner cjohnson: heh, no hard feelings, I hope.
22:25 SporkWitch osse: if you're legitimately asking, Star Citizen is a little more mainstream since it's more action focused, but it's overdue and very ambitious (and requires a monster to run it).  I mentioned it because some of the stuff you could pledge was very pricey (my Idris-M was US$1k)
22:25 hahuang61 joined #git
22:25 SporkWitch osse: you have if you paid more than $30; it's boring, the ships all handle poorly, it's repetitive.  You'll play for 20 or 30 hour stretches over a couple weeks, then give it up for months, rinse and repeat.
22:26 SporkWitch osse: honestly, it's only a little more than a graphical update to the old 80's game it's based on.
22:26 osse Now I'm very confused, because wikipedia says that game is "upcoming"
22:29 fuchstronaut joined #git
22:30 SporkWitch really we should continue this topic elsewhere lol
22:30 Vampire0 joined #git
22:30 Darcidride joined #git
22:30 SporkWitch bremner: i had to PM him to let him know you didn't ban him, he apparently thought you did lol
22:31 EvilPenguin joined #git
22:32 rkazak joined #git
22:32 cjohnson no no hard feelings
22:32 cjohnson I just don't have auto-join enabled
22:32 osse "no no. hard feeling." ?
22:32 * cjohnson shakes fist harder
22:33 pasadinhas joined #git
22:33 SporkWitch lol
22:34 fahadash joined #git
22:37 gajus I want to validate commit messages of all the commits contained in a pull-request (GitHub). Correct me if I am wrong, but to do that, I would need to validate HEAD~{number of commits since branched out}, right?
22:38 kyan joined #git
22:38 SporkWitch gajus: define validate; do you mean pgp-signed commits?
22:38 gajus No, validate git messages format.
22:38 SporkWitch gajus: if so, you can basically go by the signature on the commit the pull request is for; effectively, that person is signing off on everything that merge would do.
22:39 SporkWitch ah, that's beyond me; text parsing is a PITA, and i've not delved into that on git :(
22:39 SporkWitch your math seems correct, though
22:39 SporkWitch either that or -[commits since branch - 1]
22:40 gajus but how would I actually get "number of commits since branched out"?
22:41 SporkWitch i would assume there's a value for where the branch diverged that can be polled, probably just a matter of figuring out what that is
22:41 nitric gajus: "git log master...feature" will list the commits you want, without worrying about how many there are
22:41 SporkWitch derp, duh
22:42 osse nitric: that with list commits on master not in feature too
22:42 osse git log master..feature is probably what's wanted
22:43 rpd You want to validate the message format? What are you going to do if it's wrong? Sounds like really want everyone to run a precommit script that doesn't allow them submit bad messages in the first place
22:44 SporkWitch ^
22:44 gajus rpd There is already pre-commit hook
22:44 gajus This is only done to ensure that you are not bypassing it.
22:45 osse server-side enforcement of da law = best enforcementof da law
22:46 nitric osse: oops, you're right, that extra dot slipped in there
22:46 osse that's what she said
22:46 SporkWitch *Stallone voice* I... Am the LAW!
22:47 sangy joined #git
22:49 gajus bash question: why this doesn't work? `git log "master..$(git branch)"`
22:49 gajus ah..
22:49 gajus git branch doesn't return what I thought it does
22:49 SporkWitch Correct
22:50 Guest93616 joined #git
22:50 SporkWitch You need a refspec iirc
22:50 rpd the arguments for a <commit-ish> range should be a SHA or a refspec
22:50 osse HEAD will do
22:50 gajus `git symbolic-ref --short HEAD`
22:50 osse but in a hook you already get the info you need on stdin, or maybe as args
22:50 rik316 joined #git
22:51 rpd gajus: anyways, try https://git-scm.com/book/en/v2/Custom​izing-Git-An-Example-Git-Enforced-Pol​icy#_enforcing_commit_message_format
22:54 SporkWitch If you Google how to do an auto deploy hook, it'll give you what you want to pull the branch name in one of the first results (some great guides and it's helpful for building other types of hooks
22:54 gajus Thanks rpd
22:54 nickthetait joined #git
23:00 arescorpio joined #git
23:01 star_prone joined #git
23:04 gajus Does this look right?
23:04 gajus for sha in $(git --no-pager rev-list master..$(git symbolic-ref --short HEAD)); do git cat-file commit $sha | sed '1,/^$/d' | conventional-changelog-lint; done
23:05 fairuz joined #git
23:05 rpd seems fine
23:06 gajus Right. Though, how do I make this command exit with status code 1 if either of the commit validation pipelines exit with status code 1?
23:06 catag87 joined #git
23:07 ASOLAr joined #git
23:08 mingrammer joined #git
23:08 gajus a question more suited for #bash, though
23:09 SporkWitch Gajus figure out the return statement in bash and wrap the command in it (a subprocess or whatever it's called in bash, might make sense and make this easier)
23:09 tcurdt let's say I want to add a file to a git repo ... but instead of committing it to a local repo and pushing the change to the remote I want to do this just through the git http protocol
23:09 tcurdt before I dive into git-receive-pack - should that be doable?
23:10 SporkWitch I'm not sure that's desirable...
23:11 SporkWitch Part of the strength of git is the distribution of repositories. Also, I'm not sure what error checking there would be if it was damaged in transit (hashing when you push would ensure any damage is immediately visible)
23:11 HoierM joined #git
23:12 tcurdt It's a very special use case - that's for sure.
23:12 MatthewAllan93 joined #git
23:15 rpd tcurdt: you do git commands over HTTP https://git-scm.com/book/en/v2/Git-on-the​-Server-The-Protocols#The-HTTP-Protocols
23:15 tcurdt rpd that's the idea
23:16 tcurdt but I was hoping to get some insights of what's happening under the covers
23:16 energizer joined #git
23:17 csd_ joined #git
23:17 cjohnson @SporkWitch
23:17 gajus I am confused, what is `def check_message_format` in Bash?
23:18 gajus My Google-foo tells me it is not a function. I cannot find docs for what it is.
23:18 belak Ok, what protocol does git:// imply? Is it simply dumb http?
23:18 rpd tcurdt: uh, what do you mean what's happening over the covers? You mean, what's different from using the ssh protocol?
23:19 rpd belak: git:// using implies dumb http
23:19 rpd usually*
23:19 rpd usually git:// implies read-only which implies dumb http I should say
23:19 gajus oh...
23:19 gajus that example was in Ruby, lol
23:20 Guest93616 joined #git
23:20 belak rpd: thanks
23:20 zivester joined #git
23:21 t-mart joined #git
23:21 ffabi joined #git
23:22 tcurdt rpd I am wondering if there are any specs for the output of git-send-pack
23:23 bignose joined #git
23:23 SporkWitch ah, now i remember why i never tried gitlab: ruby
23:23 bignose is there a solid convention for naming remotes, when using a “that's the central repo where changes come from, this is my personal fork where I push changes to” workflow?
23:24 tcurdt ha! found this https://github.com/git/git/blob/master/D​ocumentation/technical/pack-protocol.txt
23:24 bignose i.e. which should be ‘origin’, if any? what names should each remote be?
23:24 Nugget bignose: I prefer origin for my fork and "upstream" for the master repo
23:24 Nugget but I don't think there's any real consensus
23:24 SporkWitch bignose: origin is where you cloned from by default, and i would generally say it should stay that way (typically this will be your personal fork at the remote)
23:25 Nugget the github docs do it that way
23:25 SporkWitch bignose: upstream is a common one for the repo that you forked from
23:25 SporkWitch whoo, i'm doing it how git says to without having read them! lol
23:25 a3Dman joined #git
23:25 bignose SporkWitch: that isn't what I normally have, though. I typically clone from the central one, then only later decide I want a personal fork, so push to that later.
23:25 rpd bignose: for example, I maintain a gcc fork. So origin is my fork, <username>/gcc and upstream is gcc/gcc.
23:26 SporkWitch bignose: it's trivial to change remotes.  if you do that you can just do git remote set-url [remote name] [new url]
23:26 bignose as I thought, there is disagreement. ‘origin’ == central, is incompatible with ‘origin’ == personal fork.
23:26 bignose SporkWitch: I'm asking how best to wrk with the tools, without needing to change remote names all the time.
23:27 rpd the tools don't care at all though
23:27 SporkWitch ^
23:27 Nugget the git docs themselves shy away from this subject in any detail in a nod to the sense that a "central" repo is an invalid notion to begin with.
23:27 SporkWitch the naming convention should be what makes the most sense at the time; if it no longer makes sense, change it
23:27 vassagus joined #git
23:27 SporkWitch this is especially trivial with naming remotes, since it's a local alias anyway
23:27 ffabi joined #git
23:28 Nugget any names you choose are local to your repo so they can be whatever makes you happy
23:28 Nugget name them after star trek characters
23:28 SporkWitch or the machine they're stored on!
23:28 rpd If you're talking about how to help a team discuss these concepts, just pick something. Origin as the fork and Upstream as the original repository are fine things to pick.
23:28 _28_ria joined #git
23:28 SporkWitch (this is one i use a lot for my config repos)
23:28 ASOLAr joined #git
23:29 niranjan joined #git
23:30 fuchstronaut joined #git
23:32 niranjan Hi I am getting error error: pack-objects died of signal 7 when cloning the repository, but push/pull works fine. On Ubuntu 16.04, git 2.9.2
23:33 nullp0inter joined #git
23:33 Rodya_ joined #git
23:33 nettoweb joined #git
23:33 Dschingiz joined #git
23:33 rpd niranjan: what's signal 7, SIGILL??
23:33 al-damiri joined #git
23:34 RedneckDonut joined #git
23:34 tcurdt hm ... I am seeing a "003f945d499872f5de5bcf2ddf71263a29e352f8770e refs/heads/master" in the response to git-upload-pack - but it doesn't match any commits I have
23:34 niranjan @rpd - I have no idea, trying find it out
23:35 tcurdt any idea what else that could be?
23:35 tcurdt the spec says "The returned response is a pkt-line stream describing each ref and
23:35 tcurdt its current value"
23:35 niranjan @rpd - It seems to be sigbus
23:37 Literphor joined #git
23:38 EvilPeng1 joined #git
23:38 tcurdt ah ... git ls-remote
23:40 dsdeiz joined #git
23:40 dsdeiz joined #git
23:42 mablae__ joined #git
23:42 paws_ joined #git
23:44 sr___ joined #git
23:44 rpd niranjan: uh, I feel like git-pack-objects can be very memory intensive. Not sure what you'd do about that though
23:45 niranjan @rpd, thanks, I can clone chrome and webkit repository fine, but this one is local development repository. I don't think it's that big.
23:45 niranjan Something to do with my repository - that's what I am thinking.
23:45 ASOLAr joined #git
23:47 ASOLAr_ joined #git
23:48 thiago joined #git
23:49 nullp0inter is it normal practice to install something like bootstrap within your repo...of do people prefer to install dependencies separate from the application
23:49 nullp0inter or do*
23:49 SporkWitch nullp0inter: isn't "installing" bootstrap just a matter of putting the files there?
23:50 Guest93616 joined #git
23:50 nullp0inter SporkWitch, yep, but there are ways to do it like composer and such, just not sure if its normal to include all that
23:50 nullp0inter usually i do
23:50 nullp0inter but wondering if theres a better alternative
23:50 SporkWitch nullp0inter: AFAIK, the normal practice is to make the docroot of that site the base of the repo (that is, the root of your repo contains the same files as the docroot of the site)
23:51 _sjs joined #git
23:51 SporkWitch nullp0inter: i'm not much of a web guy though, i avoid it like the plague.  My thinking would be that I should be able to completely restore the site with a simple "checkout -f workdir=''"
23:51 RedneckDonut joined #git
23:52 Thorn__ joined #git
23:52 nullp0inter makes sense ty
23:52 SporkWitch that's just my perspective, though; i'm honestly the last person this particular question should be addressed to.  I'm just saying how i, personally would do it; it could be totally daft
23:53 nullp0inter haha
23:53 SporkWitch g2g though
23:53 nullp0inter take care
23:53 vassagus joined #git
23:55 anuxivm joined #git
23:58 stamina joined #git
23:59 jeffreylevesque joined #git

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