Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-11-28

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

All times shown according to UTC.

Time Nick Message
00:01 Derperperd joined #git
00:04 anonymuse joined #git
00:04 pbrewczynski joined #git
00:04 dsantiago joined #git
00:05 pigeon joined #git
00:05 tvw joined #git
00:08 cool_boy joined #git
00:08 nidr0x joined #git
00:08 buffal0 joined #git
00:08 hfp left #git
00:09 rebrec joined #git
00:12 jmpp joined #git
00:18 rafaeldelucena joined #git
00:18 PioneerAxon joined #git
00:19 Sleepy63 joined #git
00:20 cyan__ joined #git
00:21 Derperperd joined #git
00:22 Derperperd joined #git
00:22 anonymuse joined #git
00:24 harrymoreno joined #git
00:24 leb joined #git
00:24 jmpp joined #git
00:25 rebrec joined #git
00:25 jnewt_ joined #git
00:27 jaguarmagenta joined #git
00:29 jmpp_ joined #git
00:29 jmpp_ joined #git
00:33 a3Dman joined #git
00:35 nettoweb joined #git
00:36 CrashOverride joined #git
00:37 qws-user-1228 joined #git
00:40 CussBot joined #git
00:40 jnewt__ joined #git
00:40 rebrec joined #git
00:41 Mobutils joined #git
00:41 dodobrain joined #git
00:46 ISmithers joined #git
00:49 mischat joined #git
00:53 arescorpio joined #git
00:54 a3Dman joined #git
00:55 Sonderblade joined #git
00:57 jimi_ joined #git
00:58 buffal0 joined #git
00:59 TheMoonMaster left #git
01:00 ResidentBiscuit joined #git
01:00 PresidentBiscuit joined #git
01:10 GavinMagnus joined #git
01:13 fling joined #git
01:14 jaguarmagenta joined #git
01:17 derk0pf joined #git
01:18 cqi joined #git
01:19 SpeakerToMeat Question, is there any way to find change sets that change a specific line number in a file?
01:19 t-mart joined #git
01:20 tnecniv joined #git
01:21 a3Dman joined #git
01:24 CEnnis91 joined #git
01:27 tnecniv joined #git
01:34 courrier_ joined #git
01:36 jeffreylevesque joined #git
01:38 t-mart joined #git
01:38 z8z joined #git
01:38 paramourne joined #git
01:40 jnewt joined #git
01:41 paramourne joined #git
01:43 hahuang65 joined #git
01:44 a3Dman joined #git
01:46 jaguarmagenta joined #git
01:47 paramourne joined #git
01:50 mischat joined #git
01:50 sagerdearia joined #git
01:55 jnewt joined #git
01:56 jnewt joined #git
01:56 cyan__ joined #git
01:56 Nilesh_ joined #git
01:57 cdg joined #git
02:06 jaguarmagenta joined #git
02:07 jaguarmagenta joined #git
02:09 jmpp joined #git
02:10 Rodya_ joined #git
02:14 buffal0 joined #git
02:15 d^sh joined #git
02:17 ryouma joined #git
02:18 ryouma i keep a set of changes.  when i pull, they automatically rebase.  this sometimes causes conflicts.  what i want to see when i resolve these conflicts are my changes to the old version, so that i can use those as a guide to edit the conflict so that i am making a similar change to the version i just pulled from upstream.  is this possible?
02:21 finalbeta joined #git
02:23 LiENUS joined #git
02:23 pks joined #git
02:27 malber joined #git
02:30 malber_ joined #git
02:30 acetakwas joined #git
02:33 Vampire0 joined #git
02:35 snowkidind joined #git
02:37 Hello71 ryouma: set merge.conflictStyle=diff3
02:38 mizu_no_oto joined #git
02:39 Mixologic_ joined #git
02:40 Rodya_ joined #git
02:42 howdoi joined #git
02:46 Rodya__ joined #git
02:48 ilbot3 joined #git
02:48 Topic for #git is now Welcome to #git, the place for git help and serious business | Public logs at http://jk.gs/glog | First visit? Read: http://jk.gs/git | Current stable version: 2.10.2 | Getting "cannot send to channel"? /msg gitinfo .voice | The git-jokes project has been suspended for lack of contribution. Help revive it! https://madeitwor.se/git-jokes
02:50 Mixologic_ joined #git
02:51 mischat joined #git
02:52 aavrug joined #git
02:53 aavrug joined #git
02:54 svm_invictvs joined #git
02:55 Jameser joined #git
02:55 Mixologic_ joined #git
02:56 ljc joined #git
02:56 Rodya_ joined #git
02:56 hahuang65 joined #git
02:57 dreiss joined #git
02:57 moei joined #git
02:58 raspado joined #git
03:00 GavinMagnus left #git
03:00 tnecniv joined #git
03:03 Mixologic joined #git
03:08 hahuang65 joined #git
03:08 ResidentBiscuit joined #git
03:10 chachasmooth joined #git
03:10 krillr joined #git
03:13 hexagoxel joined #git
03:15 sunri5e joined #git
03:19 Rodya_ joined #git
03:19 nezZario I really hate asking this, ... Does anyone know of a decent 3rd party/"cloud" solution that supports pre-receive hooks (and rejecting commits) ?
03:19 nezZario BitBucket nor AWS CodeCommit support rejecting commits despite supporting webhooks and AWS supports even sending it straight to Lambda functions .. but just doesn't give you the simple ability to say "no"
03:20 acetakwas joined #git
03:24 andlabs joined #git
03:28 ugjka joined #git
03:32 sdothum joined #git
03:32 Andrew_K|2 joined #git
03:32 ryouma Hello71: hmm, good idea.  so showing the ancestor would allow me to diff.  but is there a trick to showing the actual diffs?  as opposed to you copy pasting then diffing?
03:33 Hello71 wat
03:34 ResidentBiscuit joined #git
03:40 A5101 joined #git
03:41 ryouma Hello71: i find fixing merge conflicts a little confusing and stressful.  it would be a better experience if i not only saw the theirs mine ancestor, but also can show the diff between ancestor and mine.  at merge time.
03:41 ryouma then i will know what i changed, so i can make similar changes to theirs.
03:42 ryouma i can hack all this in emacs if necessary i suppose, but i'm not aware that git or magit do this
03:42 Hello71 probably your commits should be small enough that you can do that mentally
03:42 ryouma (smerge does it, but maybe not exactly what i want)
03:43 ryouma Hello71: they are quite small, but sometimes big changes occur on upstream
03:43 boombatower joined #git
03:43 mizu_no_oto joined #git
03:44 tnecniv joined #git
03:44 atesc joined #git
03:45 justan0theruser joined #git
03:46 Gsham joined #git
03:47 kesslerfrost joined #git
03:48 svm_invictvs joined #git
03:49 Rodya_ joined #git
03:51 mischat joined #git
03:53 Helithumper joined #git
03:53 Rodya_ joined #git
03:56 cdown_ joined #git
03:57 a_thakur joined #git
03:57 a_thakur joined #git
04:00 hhee joined #git
04:00 cqi joined #git
04:02 cqi joined #git
04:04 jost__ joined #git
04:07 jnewt_ joined #git
04:09 davimore_ joined #git
04:10 jason237 joined #git
04:11 Gsham joined #git
04:12 donmichelangelo joined #git
04:14 Vortex35 joined #git
04:18 gmn8719 joined #git
04:20 jason237 joined #git
04:22 venmx joined #git
04:26 Rodya_ joined #git
04:28 hahuang65 joined #git
04:29 govg joined #git
04:31 svm_invictvs joined #git
04:31 vonnieda joined #git
04:32 vonnieda Hi folks, I have a merge conflict that is confusing me. If anyone has a moment, could you help me understand why develop does not merge cleanly onto master in https://github.com/openpnp/openpnp.git?
04:32 vonnieda The last commit to master is a merge from develop
04:33 vonnieda And there are odd conflicts in the merge, like completely new files.
04:33 Rodya_ joined #git
04:34 jason237 joined #git
04:35 ResidentBiscuit joined #git
04:37 Helithumper joined #git
04:39 robotroll joined #git
04:40 fstd joined #git
04:46 ayogi joined #git
04:47 subhojit777 joined #git
04:48 Derperperd joined #git
04:50 dsdeiz joined #git
04:51 cdg joined #git
04:52 mischat joined #git
04:55 dreiss joined #git
04:58 Rodya_ joined #git
04:58 OnBrokenWings joined #git
04:59 navidr joined #git
05:00 hahuang65 joined #git
05:07 Cabanossi joined #git
05:07 SwiftMatt joined #git
05:17 kerrick joined #git
05:17 Akardi joined #git
05:29 cheater joined #git
05:31 Sonderblade joined #git
05:34 SwiftMatt joined #git
05:43 abak joined #git
05:46 vonnieda left #git
05:48 jfr joined #git
05:48 Gsham joined #git
05:51 harrymoreno joined #git
05:52 Gsham joined #git
05:53 mischat joined #git
05:55 inflames joined #git
05:57 Rodya_ joined #git
06:00 Anja_ joined #git
06:00 SwiftMatt joined #git
06:05 GT4066 joined #git
06:07 ok91 joined #git
06:08 rwb joined #git
06:13 overlord_tm joined #git
06:17 ToBeCloud joined #git
06:17 Rodya_ joined #git
06:19 inflames joined #git
06:22 sssilver joined #git
06:23 jozwior joined #git
06:24 b_rare joined #git
06:27 cdown joined #git
06:27 svm_invictvs joined #git
06:30 _28_ria joined #git
06:31 shgysk8zer0 joined #git
06:31 _UniFreak joined #git
06:33 UniFreak I use composer installed a package, but it won't be added to commit
06:33 UniFreak cause some thing about sub-module
06:34 UniFreak what's that? how can I remove it from sub-module?
06:36 dsdeiz joined #git
06:36 dsdeiz joined #git
06:36 ResidentBiscuit joined #git
06:38 lindenle joined #git
06:39 UniFreak I removed the .git folder under the package, now git status show there is no new change
06:39 dan2k3k4 joined #git
06:40 jozwior joined #git
06:41 UniFreak so I moved it back, now git status will hang up
06:42 mischat joined #git
06:42 svm_invictvs joined #git
06:43 Abbadonnet joined #git
06:44 hhee guys, which GUI programs available for linux?
06:44 Jameser joined #git
06:44 WayToDoor joined #git
06:44 Siamaster joined #git
06:45 Siamaster Any idea how I can solve this error when pushing to Gerrit?
06:45 Siamaster Pushing merges in commit chains with 'all not in target' is not allowed,
06:45 Siamaster to override please set the base manually
06:45 Abbadonnet left #git
06:46 ThomasLocke joined #git
06:46 ThomasLocke joined #git
06:50 LucaTM joined #git
06:51 Raging_Hog joined #git
06:51 qt-x joined #git
06:53 ResidentBiscuit joined #git
06:55 valkyr2e joined #git
06:59 vdv joined #git
07:00 alcohol Siamaster: seems like you merged something manually but the gerrit controlled repo does not allow that?
07:00 Siamaster I merged master on my topic branch
07:00 rscata joined #git
07:00 Siamaster and then tried to continue push to that branch
07:00 Siamaster then I got this error
07:00 alcohol should probably have rebased instead
07:00 Siamaster I see
07:00 Siamaster thanks
07:01 alcohol merge commits on topic branches make for a messy history once you merge back to master (or so most gerrit fans would say i reckon)
07:02 inflames joined #git
07:02 17SAANZQO joined #git
07:02 syg_ joined #git
07:02 Hobbyboy joined #git
07:04 Macaveli joined #git
07:04 byte512 joined #git
07:04 chele joined #git
07:05 _ng joined #git
07:05 govg joined #git
07:06 jnewt joined #git
07:06 elect joined #git
07:07 barteks2x joined #git
07:08 ResidentBiscuit joined #git
07:11 lb1c joined #git
07:12 nowhereman joined #git
07:17 freimatz joined #git
07:17 troyt joined #git
07:18 jnavila joined #git
07:20 seishun joined #git
07:22 spacelord_ joined #git
07:23 roelmonnens joined #git
07:24 roelmonnens joined #git
07:24 ResidentBiscuit joined #git
07:27 cheater joined #git
07:28 UniFreak joined #git
07:29 UniFreak after the git status hang up
07:29 UniFreak I checked the task monitor, git at least have three hundred of thread
07:29 UniFreak WTF
07:29 venmx joined #git
07:29 dreiss joined #git
07:32 PCatinean joined #git
07:35 jnewt_ joined #git
07:35 dsdeiz joined #git
07:35 hahuang65 joined #git
07:36 g_zen joined #git
07:38 lordjancso joined #git
07:39 Pseudocrat joined #git
07:39 Pseudocrat joined #git
07:39 Sceorem joined #git
07:40 ResidentBiscuit joined #git
07:43 mischat joined #git
07:43 Alienpruts joined #git
07:43 ferr1 joined #git
07:48 moei left #git
07:49 paradox_ joined #git
07:49 blahdodo joined #git
07:50 grayjoc joined #git
07:52 paradox_ left #git
07:53 JeroenT joined #git
07:53 paradox111_111 joined #git
07:55 paradox111_111 left #git
07:55 fees joined #git
07:56 ResidentBiscuit joined #git
07:56 brillenfux joined #git
07:59 Lotlot joined #git
08:01 _28_ria joined #git
08:02 vmiklos joined #git
08:03 thierryp joined #git
08:03 cheater joined #git
08:04 troyt joined #git
08:05 nilg joined #git
08:05 srigi joined #git
08:07 lb1c parallel ftw
08:08 paradox111_111 joined #git
08:09 Darcidride joined #git
08:10 paradox111_111 Hi guys, I have a question. If I want to use a github project in my own project, how should I do that?
08:11 paradox111_111 Should I download the repo using my IDE and then use it.
08:11 tobiasvl paradox111_111: it's not clear what your question is
08:11 ResidentBiscuit joined #git
08:12 paradox111_111 There is a project on github with a GNU GPL v3 license which I want to use in my own project. So, how can I do that?
08:12 tobiasvl use it how?
08:13 jnewt joined #git
08:13 tobiasvl do you want to make it a submodule of your project?
08:13 paradox111_111 yes
08:13 tobiasvl !submodule
08:13 gitinfo git-submodule is ideal to add subsidiary git repositories to a git superproject when you do not control the subprojects or more specifically wish to fix the subproject at a specific revision even as the subproject changes upstream. See http://www.git-scm.com/book/en/Git-Tools-Submodules
08:14 JeroenT_ joined #git
08:14 thierryp joined #git
08:14 axisys joined #git
08:15 paradox111_111 Thanks a bunch
08:16 srigi joined #git
08:16 anddam joined #git
08:17 paradox111_111 Let's say I am not using git, then how can I proceed?
08:17 tobiasvl uh
08:17 anddam if I stash on master, and then pull from origin are the stashes supposed to be automatically reapplied?
08:17 tobiasvl proceed with what?
08:17 tobiasvl paradox111_111: what's your question if it's not git related?
08:17 tobiasvl anddam: no
08:17 JeroenT_ joined #git
08:17 anddam tobiasvl: then I must have double-screwed something
08:18 JeroenT joined #git
08:18 LukasT joined #git
08:18 paradox111_111 so here's the thing, can I use git to download the online repo and then use it.
08:19 cerkauskas joined #git
08:19 srigi left #git
08:19 tobiasvl paradox111_111: yes. it was on github, you said? there should be a green button on the repo page saying "Clone or download"
08:19 paradox111_111 That's it thank you very much.
08:19 tobiasvl paradox111_111: there you can either clone the repo with git, or download a zip file
08:20 paradox111_111 got it
08:20 paradox111_111 good night
08:20 r00t^2 joined #git
08:20 cerkauskas hello. I have this case: before me, developer would deploy using ftp. I recognize that git pull is superior way compared to ftp. So, is there a way to init repository and overwrite all changes with pull, no matter what is on the server (no git repository initialized) ?
08:21 seishun joined #git
08:21 mar77i_ joined #git
08:22 Masber joined #git
08:23 r00t^2 ...wow, that nick scroll. question: is it possible to only have submodules tied to a *local* branch? e.g.:
08:23 UniFreak joined #git
08:23 r00t^2 i have two branches in a local repo: 'master' and 'dev'. i want to use submodule 'sub' with URI git://git.domain.tld/sub, but only have it "active", if you will, on banch 'master'. is there a "right" way of going about this or should i just never init it on branch 'dev'?
08:23 djb-irc joined #git
08:27 derk0pf joined #git
08:27 ResidentBiscuit joined #git
08:27 seishun joined #git
08:28 sostenuto joined #git
08:30 nowhereman joined #git
08:32 djb-irc joined #git
08:36 djb-irc joined #git
08:37 barduck joined #git
08:38 Anja_ joined #git
08:38 encod3 joined #git
08:40 bernardio joined #git
08:43 Gleb` joined #git
08:43 Gleb` hiya
08:43 ResidentBiscuit joined #git
08:43 Gleb` I want to download the .git folder of a repository without cloning it, is that possible?
08:44 tobiasvl Gleb`: not really, unless you have direct access to the server. but why do you want that?
08:44 overlord_tm joined #git
08:45 venmx joined #git
08:46 finalbeta joined #git
08:46 LukasT Hi, how do I do git line normalization?
08:47 LukasT I checked what is the status of the files with:
08:47 LukasT git ls-files --eol
08:47 Snugglebash joined #git
08:47 LukasT So I followed this tutorial https://help.github.com/articles/dealing-with-line-endings/
08:47 LukasT but after git reset --hard
08:47 LukasT I do not see any file changes
08:47 anddam I'm on master on a project clned from github, I added a remote with a fork's url in order to check a change
08:48 anddam I then merged remotes/other/master and now I have a commit
08:48 anddam can I delete this commit?
08:49 Bombe Generally, yes and no. It pretty much depends on what you want to achieve.
08:49 fmcwlt joined #git
08:49 Gleb` tobiasvl, i want to do some text mining on all commit messages
08:50 Gleb` cloning all repositories when all i need is the history in master branch seems overkill
08:50 SwiftMatt joined #git
08:50 acetakwas joined #git
08:51 kerrick joined #git
08:54 Sasazuka joined #git
08:54 dimi1947 joined #git
08:55 djb-irc joined #git
08:57 ngperfex joined #git
08:57 tobiasvl Gleb`: git clone --single-branch
08:58 mikecmpbll joined #git
08:58 upjyij joined #git
08:58 anddam Bombe: was that for me?
08:58 tobiasvl anddam: you can delete it, it just depends on what you want to achieve, as Bombe said. see !revert
08:58 gitinfo anddam: That's a rather ambiguous question... options: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard uncommitted changes in the working tree [git reset --hard]; c) undo committing [git reset --soft HEAD^]; d) restore staged versions of files [git checkout -p]; e) move the current branch to a different point(possibly losing commits)[git reset --hard $COMMIT]?
08:59 ResidentBiscuit joined #git
08:59 anddam tobiasvl: I have push right on origin but I think this commit doesn't need to go there, yet, so I want to obliterate it
09:00 anddam I merged but I actually just wanted to check that remotes/other/master difference on top of origin/master to see how it worked
09:00 venmx joined #git
09:00 anddam I think I should have rebased
09:00 anddam I'm still quite confused by git commands specific meaning
09:01 r00t^2 left #git
09:02 jab416171 joined #git
09:03 tobiasvl anddam: sounds like you can just git reset --hard to the commit before the merge
09:05 anddam tobiasvl: do I need to esplicitly meantion the commit's hash?
09:06 anddam git help will tell
09:06 anddam yep
09:06 anddam thanks
09:07 tobiasvl you can just say HEAD^ if the merge commit is the most recent one
09:07 anddam it was, I went with has
09:07 anddam hash
09:07 Macaveli joined #git
09:08 ayogi joined #git
09:10 marcogmonteiro joined #git
09:12 Siegfried joined #git
09:12 lss8 joined #git
09:13 JeroenT joined #git
09:14 sunri5e joined #git
09:15 ResidentBiscuit joined #git
09:18 sssilver joined #git
09:19 kurkale6ka joined #git
09:19 ahmedelgabri joined #git
09:23 derk0pf joined #git
09:24 sssilver joined #git
09:27 Balliad joined #git
09:27 d0nn1e joined #git
09:29 garethdaine joined #git
09:30 dsantiago joined #git
09:31 ResidentBiscuit joined #git
09:31 giraffe joined #git
09:33 mindriot101 joined #git
09:36 gareppa joined #git
09:38 mischat joined #git
09:39 chimay joined #git
09:41 bongjovi joined #git
09:42 Balliad joined #git
09:42 roelmonn_ joined #git
09:44 clmsy joined #git
09:45 krillr joined #git
09:46 ResidentBiscuit joined #git
09:46 derk0pf joined #git
09:47 Serpent7776 joined #git
09:53 davimore__ joined #git
09:55 MrMojito joined #git
09:55 MrMojito joined #git
09:56 rominronin joined #git
09:57 systemovich joined #git
09:59 JeroenT joined #git
10:02 Kodax joined #git
10:02 xaa__ joined #git
10:02 tvw joined #git
10:02 ResidentBiscuit joined #git
10:04 GT4066 joined #git
10:06 InfoTest1 joined #git
10:06 chll_ joined #git
10:08 xall joined #git
10:08 mablae joined #git
10:09 noidea joined #git
10:10 [0xAA] joined #git
10:10 [0xAA] left #git
10:10 [0xAA] joined #git
10:10 [0xAA] how to push to multiple remotes in a single command ?
10:10 jaguarmagenta joined #git
10:11 mischat joined #git
10:11 Bombe [0xAA], Git does not offer that.
10:11 [0xAA] then a script will do that for me
10:14 [0xAA] joined #git
10:14 finalbeta joined #git
10:14 King_Hual joined #git
10:15 nowhere_man joined #git
10:15 adac joined #git
10:15 Puffball joined #git
10:18 a3Dman joined #git
10:18 ResidentBiscuit joined #git
10:19 brillenfux joined #git
10:20 [0xAA] joined #git
10:20 cheater left #git
10:21 rnsanchez joined #git
10:22 achadwick joined #git
10:22 shashin joined #git
10:23 mischat joined #git
10:27 [0xAA] joined #git
10:30 dsdeiz joined #git
10:30 dsdeiz joined #git
10:34 ResidentBiscuit joined #git
10:36 moritz uhm, you can do it
10:36 moritz you can add a remote with several push URLs
10:40 [0xAA] joined #git
10:44 manuel joined #git
10:44 kier joined #git
10:47 cebor_ joined #git
10:50 ResidentBiscuit joined #git
10:50 [0xAA] joined #git
10:53 Balliad joined #git
10:54 [0xAA] joined #git
10:55 hahuang65 joined #git
11:00 Puffball joined #git
11:00 [0xAA] joined #git
11:03 aax joined #git
11:06 ResidentBiscuit joined #git
11:08 leeN joined #git
11:09 dodobrain joined #git
11:10 roelmonnens joined #git
11:12 nanga joined #git
11:12 nanga joined #git
11:15 buffal0 joined #git
11:16 donught joined #git
11:17 ResidentBiscuit joined #git
11:17 buffffal00 joined #git
11:23 www-bukolay-net joined #git
11:24 freimatz joined #git
11:25 hanthings_ joined #git
11:25 TomyLobo joined #git
11:26 venmx joined #git
11:27 Mir4g3 joined #git
11:28 alhariel joined #git
11:29 nathdwek joined #git
11:29 ManiacTwister joined #git
11:29 nathdwek Hi, how can I delete remote empty directories with git svn?
11:30 nathdwek I ran into this issue: https://spin.atomicobject.com/2014/08/17/git-svn-empty-directories/ , but I only realized that after the empty directories were left on the remote
11:30 mindriot101 joined #git
11:30 nathdwek So now I don't have those in my local repo, but there are empty dirs left on the remote, and I don't know how to cleanly remove them
11:31 Macaveli joined #git
11:31 cbreak nathdwek: prune
11:32 cbreak hmm... no, that's something else
11:32 cbreak --rmdir does it I think
11:32 nathdwek ...
11:32 cbreak there are details on that flag and the config in man git-svn
11:32 gitinfo the git-svn manpage is available at http://jk.gs/git-svn.html
11:32 Puffball joined #git
11:33 nathdwek That's not exactly what I'm asking, cbreak
11:33 nowhere_man joined #git
11:33 nathdwek Yeah, if I were aware of it, it would have helped
11:34 nathdwek but now I did dcommit without --rmdir
11:34 buffal0 joined #git
11:34 nathdwek and the next commits, even with --rmdir don't seem to remove those leftovers empty dirs
11:34 Guest21509 joined #git
11:34 nathdwek So I'm wondering if there is a way to force the removal of those, without special rights
11:35 cbreak you can't just make a new commit that removes them?
11:35 davimore_ joined #git
11:36 Balliad joined #git
11:36 nathdwek That's what I was hoping
11:36 nathdwek but doesn't seem like it
11:37 nathdwek I'm doing a fresh clone to see if the empty dirs will be cloned again, maybe I could delete them if they do
11:37 nathdwek they are not :'(
11:38 Puffball joined #git
11:38 JeroenT joined #git
11:41 osse Git is fairly oblivious to empty directories. Try making a commit in svn instead
11:42 paladinn joined #git
11:44 CheckDavid joined #git
11:44 JeroenT joined #git
11:46 re1 joined #git
11:48 shinnya joined #git
11:49 Puffball joined #git
11:50 shashin joined #git
11:50 hariel joined #git
11:54 pbrewczynski joined #git
11:59 dimi1947 joined #git
12:04 dimi1947 I have installed npm  and its folder is installed in project dir...as a result all the files contained in it are tracked and are considered by git as unstaged changes..how I can make git stop tracking these files? I think git rm cached should be used here but we are talking about  many files...not just one
12:05 osse rm -r --cached
12:05 osse are they listed as modified files or new files?
12:08 dimi1947 osse:how am I going to understand that they are listed as new ones...
12:09 nd_ joined #git
12:09 dimi1947 osse:I do not see any clue related to that
12:10 osse dimi1947: in git status
12:10 nathdwek osse: Should've thought about that earlier, thanks! it worked, even though the repo history is a bit cluttered now. I hope it won't mess up later merges on the svn front
12:11 alcohol dimi1947: also add the npm dir to your local or global .gitignore
12:11 dimi1947 osse:I did use git status and as I did not find a clue that are considered new...so I am asking again how am I going to understand this
12:11 jaguarmagenta joined #git
12:11 dimi1947 alcohol:got that
12:12 osse dimi1947: git status spits it in your face. next to every file it either says "new file" or "modified"
12:12 dimi1947 osse:give me a second
12:12 alcohol osse: just pastebin the output for us
12:14 dimi1947 alcohol:I am preparing to do just that....
12:15 Flotho66 joined #git
12:15 Flotho66 left #git
12:15 Darcidride joined #git
12:18 dimi1947 https://1drv.ms/i/s!AjCBrCpLQye-g0qFZhLFS9cx2NIa as you will see it does not mention if it is new or modified
12:18 ResidentBiscuit joined #git
12:18 dsdeiz joined #git
12:18 osse you said they were tracked.
12:18 osse they aren't tracked
12:20 dimi1947 osse:in git gui they are mentioned as unstaged changes
12:20 osse that is strange
12:20 alcohol dimi1947: and what does "unstaged" mean to you?
12:20 osse can you show a screenshot of that as well
12:23 Mir4g3 joined #git
12:23 dimi1947 osse:https://1drv.ms/i/s!AjCBrCpLQye-g0uRBDstOSJndUsl  unstaged can be both tracked and untracked files
12:24 dimi1947 osse:and ok these files are untracked...where does that lead us
12:24 osse I think git gui is misleading
12:24 ohpu joined #git
12:25 osse regular git status shows unstaged and untracked separately
12:25 osse it leads us to you simply having to add these directories to .gitignore and you're done
12:26 dimi1947 osse:I was about to say just that...
12:26 dreiss joined #git
12:27 dimi1947 osse:so can the gitignore file accept entries for folders?
12:28 osse yes
12:28 Starky joined #git
12:29 dimi1947 osse:yes that did it...one last question,not that gitignore has been modified I suppose I should commit it...
12:29 ams__ joined #git
12:29 JeroenT_ joined #git
12:30 JeroenT__ joined #git
12:30 dimi1947 I have forgotten that last detail...
12:34 venmx joined #git
12:36 cool_boy joined #git
12:39 fmeerkoetter joined #git
12:39 Awoi joined #git
12:39 sea-gull hi all! It looks like there's no way to reset timestamp when using git commit with `-C/-c` options.
12:39 sea-gull is there other way of reseting git commit date?
12:40 noecc joined #git
12:41 kanyeezy joined #git
12:41 osse --date
12:41 ilmari or the GIT_AUTHOR_DATE and GIT_COMMITTER_DATE environment variables
12:42 sea-gull I was looking for setting it to `now`, but that also works. Thank you all!
12:44 osse --date="$(date)" probably does the trick
12:44 JeroenT joined #git
12:44 tobiasvl setting it to now can probably be done with --amend too
12:46 JeroenT joined #git
12:49 qqx joined #git
12:50 frostythesnowman joined #git
12:54 LukasT joined #git
12:55 robattila256 joined #git
12:56 Puffball joined #git
12:56 acuo joined #git
12:58 johnmilton joined #git
13:01 ahmed_elgabri joined #git
13:01 Gianormagantrous joined #git
13:03 aleb joined #git
13:05 ahmedelgabri joined #git
13:05 brillenfux joined #git
13:05 Puffball joined #git
13:05 GavinMagnus joined #git
13:06 xall joined #git
13:08 Nilesh_ joined #git
13:08 aleb The way we organize branches in our project is to prefix them with the Phabricator task name, for example "T7579-opengl-effects". The problem is that when I use tab completion "git checkout T<tab>" I see both my own and the others's branches which start with "T".  How can I configure it to show only the branches from the local repository, if any, and only
13:08 aleb when none are available to show branches from remotes?
13:10 aleb (Using .. oh my zsh)
13:11 isysd joined #git
13:19 ResidentBiscuit joined #git
13:20 Udozl joined #git
13:20 osse aleb: type git checkout then press Ctrl-X h
13:20 osse what do you see
13:20 aleb https://www.irccloud.com/pastebin/UALPM5fb/
13:21 osse ugh
13:21 aleb (What's this Ctrl-X?)
13:22 osse shows some inner workings of the completion system
13:22 aleb What is it that you don't like there?
13:23 osse that it looks very different from mine
13:24 osse I get a lot of info about specific completion functions that I think you can turn on and off with zstyle
13:25 acetakwas joined #git
13:25 osse I don't have __git_zsh_main at all
13:26 osse I'm starting to think it's omz spreading its vile diseases into the completion. in any case, ask #zsh
13:26 neeak joined #git
13:26 aleb hm, let me try with no .zshrc
13:28 osse you can start a clean zsh with 'zsh -f'. then inside it run 'autoload compinit && compinit'
13:30 osse this is what i get: http://sprunge.us/SATQ
13:30 shashin joined #git
13:30 aleb https://www.irccloud.com/pastebin/UoCxTzFe/
13:31 osse I think the clue is to remove the 'heads-remote' from the completion. but i don't know how :/
13:32 stoned is gitlab down for me or you too?
13:32 stoned 500
13:32 stoned Whoops, something went wrong on our end.
13:32 Akimb joined #git
13:32 venmx joined #git
13:33 Hfsen joined #git
13:34 WhereIsMySpoon joined #git
13:34 jeffreylevesque joined #git
13:36 kadoban joined #git
13:37 GraNNy- joined #git
13:38 finalbeta joined #git
13:40 WhereIsMySpoon So, I've done a Really Stupid Thing. I had a branch, that had 2 or 3 local commits in, nothing pushed to remote, was a purely local branch. I then rebased (which actually looking at the git extensions log, was a git reset --mixed HEAD command) branch onto a later commit, which left me with a large number of changed files from the new base, and also
13:40 WhereIsMySpoon all of the files I'd changed in this local branch. I decided I actually didnt want to rebase in the end, and reset ALL of the staged files. This has meant I am now left with nothing, all of the files I'd changed/created and committed earlier are gone.
13:40 WhereIsMySpoon My question is - is there a way to get those previous commits of my local branch back?
13:41 shashin joined #git
13:41 WhereIsMySpoon So that I can re-run them
13:42 eb0t joined #git
13:42 osse !reflog
13:42 gitinfo The git reflog (`git log -g`) records the SHAs of your HEADs for 2+ weeks. `git checkout -b myrestore OLDSHA` and `git reset --hard OLDSHA` will relink to that state via a new and current branch respectively, see http://sethrobertson.github.com/GitFixUm/ for full details.  WARNING: reset --hard will trash any uncommitted changes!  Visualize with: gitk --all --date-order `git log -g --pretty=%H`
13:43 osse WhereIsMySpoon: look at 'git reflog yourbranch'. Identify the commit it was on before the Really Stypid Thing happened. git reset --hard thatthing. For example, git reset --hard yourbranch@{1}
13:43 WhereIsMySpoon right, let me take a look
13:43 WhereIsMySpoon if git bash worked...that would be great...
13:43 Mir4g3 joined #git
13:43 osse WhereIsMySpoon: look for a line right after 'rebase finished'
13:43 WhereIsMySpoon i dont think it was -actually- a rebase
13:44 WhereIsMySpoon git extensions calls it "rebase"
13:44 WhereIsMySpoon but it looks like it was a git reset --mixed HEAD
13:44 WhereIsMySpoon oh shit
13:44 osse then git extensions can go to hell
13:44 WhereIsMySpoon thats because i did a reset >_>
13:44 osse in any case, do the reflog thing. find the correct thing. reset to that thing.
13:44 WhereIsMySpoon i pressed the wrong fucking button
13:44 rorro joined #git
13:46 ShekharReddy joined #git
13:46 xaa__ joined #git
13:46 Mir4g3 joined #git
13:47 borekb joined #git
13:48 borekb hi, I'm looking for confirmation that the behavior of `* text=auto eol=lf` changed in recent versions (certainly between 2.7 and 2.10).
13:49 borekb it used to run normalization on binary files (e.g., PNG) while in 2.10, it seems the way a human being would expect. has there been any change in this?
13:52 Snugglebash joined #git
13:52 snowkidind joined #git
13:52 kadoban borekb: Doesn't text=auto mean "use whatever heuristics you want?" Meaning that the behavior could change with the phase of the moon or anything anyway?
13:53 jimi_ joined #git
13:53 aax joined #git
13:54 jnhy joined #git
13:54 osse are you suggestion git decides randomly how to determine whether a file is binary? o_O
13:55 al-damiri joined #git
13:55 Raed joined #git
13:55 kadoban I'm suggesting that its behavior could change with very little provocation at all, so it's one of the things that I wouldn't really consider stable behavior, practically.
13:56 selckin i'd try grepping the release notes, other then that would need someone who just happens to know
13:57 osse the heuristic for determining binary files has been "is there a null byte in the first 8 kB?" as far as I can remember
13:57 WhereIsMySpoon osse:  ok ive done my reset, im now detached from HEAD
13:57 WhereIsMySpoon what do I do to reattach
13:57 osse WhereIsMySpoon: if you were detached before the reset too then nothing is accomplished
13:57 osse reset does not detach
13:57 WhereIsMySpoon er
13:58 WhereIsMySpoon my bad :)
13:58 WhereIsMySpoon i was detached before, sorry :)
13:59 WhereIsMySpoon ah my history is back
13:59 WhereIsMySpoon thank you so very much osse
13:59 Junior joined #git
13:59 WhereIsMySpoon now i wont do a mixed reset and then reset all my staged changes again :P
13:59 WhereIsMySpoon that was exceptionally stupid of me
14:00 rnsanchez_ joined #git
14:00 borekb kadoban: the behavior has been consistent, eol=lf needed to be enumerated for every possible extension to avoid normalization of binary files. now it seems to behave "correctly" which is exciting but I wanted to confirm that
14:01 Darcidride joined #git
14:03 sdothum joined #git
14:04 Rodya_ joined #git
14:04 nickthetait joined #git
14:04 cbreak joined #git
14:04 eb0t joined #git
14:07 dsdeiz joined #git
14:08 borekb this might be it: https://github.com/git/git/blob/master/Documentation/RelNotes/2.10.0.txt#L248
14:09 borekb `* text=auto eol=crlf` used to be handled as `* text eol=crlf`, i.e., declaring all files as text
14:09 borekb release notes specifically mention CRLF while I have experienced the same with LF but it seems to be fixed
14:11 qws-user-1229 joined #git
14:12 jaguarmagenta joined #git
14:13 derk0pf joined #git
14:16 xo_odin joined #git
14:19 CEnnis91 joined #git
14:20 nick|here joined #git
14:23 Mixologic_ joined #git
14:23 anonymuse joined #git
14:24 venmx joined #git
14:24 Rodya_ joined #git
14:26 Nysar joined #git
14:27 qws-user-1228 joined #git
14:27 cdg joined #git
14:28 Qria joined #git
14:29 Khisanth joined #git
14:30 xjletm joined #git
14:31 cdg joined #git
14:32 Gsham joined #git
14:32 systemovich joined #git
14:34 Snugglebash joined #git
14:36 Es0teric joined #git
14:36 dminuoso joined #git
14:36 Gonzo89 joined #git
14:36 dminuoso I have a small 4 line hunk Im trying to split with git add -p, but it's not splitting any further. What can I do?
14:36 ahmed_elgabri joined #git
14:36 ExoUNX joined #git
14:37 selckin edit
14:37 oaao joined #git
14:37 Es0teric joined #git
14:37 dminuoso selckin: And how can I do the same with editing?
14:37 selckin change it to what you want
14:38 selckin exit
14:38 dminuoso Ohh, I can just remove the partial hunk I odnt want right.
14:38 dminuoso It's just editing the staging area, not WT right?
14:38 selckin from add -p
14:38 selckin pick edit
14:39 kadoban dminuoso: Yes, 'e'dit in that interface is just modifying what will be staged.
14:40 KevinDF joined #git
14:40 KevinDF joined #git
14:41 JeroenT joined #git
14:41 ispn joined #git
14:42 lindenle joined #git
14:43 GavinMagnus left #git
14:44 JamesBaxter joined #git
14:47 saml joined #git
14:49 encod3 joined #git
14:52 dimi1947 joined #git
14:53 boombatower joined #git
14:53 Guest10096 joined #git
14:55 Balliad joined #git
14:55 Helithumper joined #git
14:56 Mixologic_ joined #git
14:57 xall joined #git
14:58 JeroenT joined #git
14:58 Derperperd joined #git
14:59 nickthetait joined #git
14:59 thebope joined #git
14:59 cdg_ joined #git
14:59 HoierM joined #git
14:59 AaronMT joined #git
15:00 jnhy joined #git
15:00 ejb joined #git
15:01 Dougie187 joined #git
15:01 meth joined #git
15:04 Sonderblade joined #git
15:06 sangy joined #git
15:07 HardlySeen joined #git
15:07 HardlySeen joined #git
15:09 JeroenT joined #git
15:11 gleachkr[m] joined #git
15:13 Gallardo joined #git
15:14 e14 joined #git
15:15 ResidentBiscuit joined #git
15:17 jbeeze joined #git
15:18 gleachkr[m] left #git
15:18 shashin joined #git
15:21 breezy_ joined #git
15:21 GodGinrai joined #git
15:24 User_007 joined #git
15:24 e14 joined #git
15:25 User_007 left #git
15:26 cqi joined #git
15:26 dviola joined #git
15:26 raspado joined #git
15:26 Rodya_ joined #git
15:31 venmx joined #git
15:33 LeoB joined #git
15:33 LeoB Hello!
15:34 LeoB I am running out of memory when try to 'git gc'
15:34 LeoB how can I limit the ammount of memory 'git gc' uses? Swap is not working well
15:35 Balliad joined #git
15:36 selckin might be able to configure pack.windowMemory etc read git-repack manual, if its in other steps, buy more memory
15:39 fission6 joined #git
15:40 fission6 i did git checkout -b WHATEVER, not on master, i know have a branch with the topic changes I want but it pulled in some extra files on the branch i checkedout on, whats the best way to take the changes from WHATEVER and rebase it with master as the initial branch
15:41 nicksloan joined #git
15:42 elect_ joined #git
15:42 HardlySe1n joined #git
15:42 tyreld joined #git
15:42 raspado joined #git
15:42 subhojit777 joined #git
15:43 jeffreylevesque_ joined #git
15:43 WhereIsMySpoon fission6: can you not switch back to master, branch off that with a new branch with a similar name and cherry pick the commits you want?
15:43 fission6 i guess
15:43 WhereIsMySpoon seems like the fastest way to me
15:43 WhereIsMySpoon im no expert :) but thats what id do
15:44 mischat_ joined #git
15:45 dfaught joined #git
15:45 GodGinrai wouldn't rebase just base your commits in your branch off of master?
15:46 lindenle joined #git
15:47 GodGinrai fission6: the manual has your exact use case in the rebase docs: https://git-scm.com/docs/git-rebase
15:48 fission6 thanks!
15:48 GodGinrai np
15:48 GodGinrai fission6: it requires a special flag for rebase
15:48 GodGinrai fission6: so look your example up in that link
15:49 aarobc joined #git
15:49 dimi1947 joined #git
15:49 nedbat if i use "git clone --depth=1" to get a shallow clone, can I later populate the full history?
15:49 selckin yes
15:50 sybariten joined #git
15:50 nedbat selckin: i wonder how? :)
15:51 selckin don't remember but you can definitly do it, let me google for you
15:51 selckin git fetch --depth, or git fetch --unshallow
15:52 peepsalot joined #git
15:52 hhee joined #git
15:52 GodGinrai wow, there's actually an unshallow flag
15:52 nedbat selckin: thanks :)
15:53 Darcidride joined #git
15:55 dsdeiz joined #git
15:55 dsdeiz joined #git
15:57 derk0pf joined #git
15:57 xall joined #git
15:58 axl_ joined #git
15:58 treehug88 joined #git
15:58 hahuang65 joined #git
16:01 Electrometro joined #git
16:01 hegge joined #git
16:01 HedgeMage joined #git
16:01 Nicmavr joined #git
16:01 jj- joined #git
16:01 m___ joined #git
16:01 Jupelius joined #git
16:01 Vauff joined #git
16:01 harwood joined #git
16:01 gusnan joined #git
16:01 gusnan joined #git
16:01 ixian joined #git
16:01 matthiaskrgr joined #git
16:01 Ineentho[m] joined #git
16:01 reda joined #git
16:02 Kodax joined #git
16:02 ulkesh_ joined #git
16:02 jimi_ joined #git
16:02 vktec joined #git
16:02 SleePy joined #git
16:02 ManiacTwister joined #git
16:02 trobotham joined #git
16:02 jlebar joined #git
16:02 achadwick joined #git
16:02 MalteTimmermann[ joined #git
16:02 noidea joined #git
16:02 noidea joined #git
16:02 jokester joined #git
16:02 elect_ joined #git
16:02 hariel joined #git
16:02 krillr joined #git
16:02 monomon[m] joined #git
16:02 warrantyvoid joined #git
16:02 _ng joined #git
16:02 roger_raymond joined #git
16:02 duvnell joined #git
16:02 romerocesar joined #git
16:02 Ekho joined #git
16:02 mgoodwin joined #git
16:02 unreal joined #git
16:02 ski7777 joined #git
16:02 lmat joined #git
16:02 Orbitrix joined #git
16:02 Cork joined #git
16:02 MrMojito joined #git
16:02 Sigma joined #git
16:02 armin joined #git
16:02 dviola joined #git
16:02 ToxicFrog joined #git
16:02 Siegfried joined #git
16:02 mrkake joined #git
16:02 qpdb joined #git
16:02 eb0t joined #git
16:02 arooni joined #git
16:02 NeverDie joined #git
16:02 dviola joined #git
16:02 Karunamon joined #git
16:02 floatingman joined #git
16:03 thekingofbandit joined #git
16:03 valkyr2e joined #git
16:03 idr0p joined #git
16:03 cmn joined #git
16:03 martin_g joined #git
16:03 heinrich5991 joined #git
16:04 Qria joined #git
16:04 bcardarella joined #git
16:04 mischat joined #git
16:05 ash_workz joined #git
16:05 johnlage joined #git
16:05 Infamous joined #git
16:06 howdoi joined #git
16:06 huhlig joined #git
16:06 pizzaops joined #git
16:06 ggherdov joined #git
16:06 tanuki joined #git
16:06 anonymuse joined #git
16:06 Knirch joined #git
16:07 Rodya_ joined #git
16:07 overlord_tm joined #git
16:08 navidr joined #git
16:08 altendky joined #git
16:08 prosody joined #git
16:08 JackMc joined #git
16:08 kulelu88 joined #git
16:09 likewhoa joined #git
16:09 angular_mike_ joined #git
16:09 rominronin joined #git
16:10 wizonesolutions joined #git
16:10 dfaught joined #git
16:10 Awesomecase joined #git
16:10 roelmonn_ joined #git
16:11 ojdo joined #git
16:12 dfaught joined #git
16:12 infernix joined #git
16:13 e14 joined #git
16:13 Didac joined #git
16:13 InfoTest joined #git
16:15 ResidentBiscuit joined #git
16:15 blackwind_123 joined #git
16:17 NeverTired joined #git
16:18 kanyeezy joined #git
16:21 Brando753-o_O_o joined #git
16:21 PasiFist joined #git
16:21 oskarkv joined #git
16:23 k0ng joined #git
16:24 dsdeiz joined #git
16:24 dsdeiz joined #git
16:27 bsanford joined #git
16:28 kaldoran joined #git
16:28 sonOfRa joined #git
16:28 Sashir joined #git
16:28 ghostlight joined #git
16:28 meauses joined #git
16:31 NwS joined #git
16:31 rivarun joined #git
16:33 re1_ joined #git
16:35 zeroed joined #git
16:35 zeroed joined #git
16:39 Balliad joined #git
16:43 Salander27 joined #git
16:44 juanitotrrj joined #git
16:45 dreiss joined #git
16:46 jstimm joined #git
16:48 e14 joined #git
16:51 LionsMane joined #git
16:52 Rodya_ joined #git
17:02 ejb joined #git
17:03 emg joined #git
17:07 Snugglebash joined #git
17:08 shashin joined #git
17:10 rominronin joined #git
17:10 WillLondon joined #git
17:13 mischat_ joined #git
17:14 Rodya_ joined #git
17:14 madewokherd joined #git
17:15 mischat_ joined #git
17:17 ResidentBiscuit joined #git
17:21 willlondon joined #git
17:24 thiago joined #git
17:24 willlondon left #git
17:24 bejeb joined #git
17:25 willlondon joined #git
17:26 nidr0x joined #git
17:27 Derperperd joined #git
17:28 kpease joined #git
17:30 durham joined #git
17:31 ShekharReddy joined #git
17:32 willlondon joined #git
17:32 willlondon joined #git
17:34 willlondon joined #git
17:35 thiago joined #git
17:36 tnecniv joined #git
17:36 Spec-Chum joined #git
17:37 willlondon joined #git
17:37 crayon joined #git
17:38 svm_invictvs joined #git
17:41 Ice_Strike2 joined #git
17:41 jimi_ joined #git
17:42 FrostCandy joined #git
17:44 jimi_sanchez joined #git
17:45 SpeakerToMeat Question, is there any way to find change sets that change a specific line number in a file?
17:45 cdown joined #git
17:45 sbeller_ joined #git
17:46 grawity -L
17:46 grawity well, git log -L
17:46 robotroll joined #git
17:46 grawity but that seems brittle since line numbers change if lines are added or removed
17:46 jimi joined #git
17:47 Rodya_ joined #git
17:48 chris2 joined #git
17:48 tnecniv joined #git
17:49 jccn joined #git
17:50 cdown_ joined #git
17:51 ochorocho__ joined #git
17:51 anonymuse joined #git
17:53 mischat joined #git
17:59 Es0teric joined #git
18:00 cdown_ joined #git
18:00 nicksloan joined #git
18:00 Gleb` left #git
18:01 dondy joined #git
18:02 GavinMagnus joined #git
18:02 dondy left #git
18:04 dreiss joined #git
18:07 mweshi joined #git
18:08 seishun joined #git
18:08 catern hey #git
18:08 catern I have an idea, maybe you can tell me if something like this exists?
18:08 vuoto joined #git
18:09 anddam catern: yes, I've heard of ideas already so something like that exists
18:09 catern :)
18:09 WhereIsMySpoon lol
18:10 andlabs joined #git
18:11 catern I have a bunch of git repositories (say, Github). I want to be able to look at them all in a read-only way, without consuming massive amounts of disk space (say, through a networked FUSE filesystem). I want to be able to check branches and look at different commits in the repositories (difficult...) and as a bonus, I would like it also if I could just start editing one of the repositories (say, by the repository getting downloaded onto
18:11 catern my local disk on demand)
18:13 catern so, the last part could be done with overlayfs on top of some FUSE filesystem which serves github files
18:13 catern but the tricky part is the ability to check out and inspect arbitrary commits
18:13 ShekharReddy joined #git
18:13 grawity if you want svn, just use github's svn gateway
18:14 catern is there some way I could achieve that, without having to download the whole repository to my local disk? maybe not...
18:14 grawity or `git clone --shared`
18:15 catern grawity: heh, but svn doesn't run over a networked filesystem. git clone --shared is an interesting idea though, hmm...
18:15 duvnell joined #git
18:15 grawity sure it does
18:15 catern oh, i've never used it :)
18:16 grawity I mean, `svn checkout` is practically exactly what you describe: downloads a specific commit on demand
18:16 grawity sure, it won't work with pure git repos
18:16 grawity but it works with github repos
18:16 r00twrh joined #git
18:16 campee joined #git
18:17 catern well, there's a lot more plumbing needed on top of that to get what I described, though... the biggest issue is that I don't *want* to download a commit on demand; I want to change the view that the network filesystem gives me on demand
18:17 duvnell1 joined #git
18:17 campee is it possible to set a file (an ssh key) to mode 0700 in a git repo so that when it gets checked out it gets to set to 0700?
18:18 ilmari no, git does not track permissions
18:18 ilmari only the executable bit
18:18 campee yeah that's all i could find.. oh well. thanks
18:18 catern after all, the sources alone for all github repositories would take up a bunch of disk space
18:18 dsantiago joined #git
18:18 davidfetter_ge joined #git
18:19 davidfetter_ge hi
18:19 catern all that I want to store is a commit hash saying, "here is what I currently am looking at in this repository - no need to have a local copy because it's only read-only"
18:20 grawity that sounds like `git clone --shared`
18:20 Mixologic joined #git
18:20 catern but if I git clone --shared, the actual source file get stored on my local disk. the objects are remote, sure, but the sources alone will take up too much space
18:21 catern a single checkout of the sources* sorry probably that was the wrong terminology
18:21 shgysk8zer0 joined #git
18:21 catern but do you get what I mean?
18:23 sylario joined #git
18:23 grawity well in that case you need something beyond git's abilities
18:24 grawity I think I've seen a FUSE-based tool that works like this
18:24 grawity i.e. on-demand working tree
18:24 catern I would be very interested in that
18:24 grawity probably https://github.com/PressLabs/gitfs
18:25 jnewt_ joined #git
18:25 catern interesting
18:26 catern that's a little weird though as it abstracts over git
18:26 grawity well, you do need to have a way to ask it for a specific commit...
18:26 grawity should be good enough for read-only access
18:26 catern (well, I mean that it seems to remove git from the equation)
18:27 catern yeah, I will look at it
18:28 catern hmm yeah I don't think this has a way to ask it for a specific commit without remounting
18:28 grawity seems to be cd $commitid
18:28 jnewt__ joined #git
18:29 catern are you talking about https://github.com/g2p/git-fs ?
18:29 grawity I'm talking about the demo at https://www.presslabs.com/gitfs/
18:29 grawity they all are similar though
18:29 catern oh yeah you're right
18:30 robotroll joined #git
18:31 Kodax joined #git
18:31 Sasazuka joined #git
18:32 adac joined #git
18:32 jnewt_ joined #git
18:33 DolphinDream howdy
18:33 duvnell joined #git
18:36 e14 joined #git
18:36 fahadash joined #git
18:37 WayToDoor joined #git
18:37 jnewt__ joined #git
18:39 duvnell joined #git
18:39 catern it's really not as transparent as i'd like though
18:40 catern that kind of magic (the commit hash is in the filesystem) isn't really what I want, since I would like to be able to just start editing the repository in-place and it magically makes a local clone for me
18:41 catern and that becomes pretty weird and tricky to use, I think, if I can "just start editing" at some arbitrary mnt/repo/hash/ path, rather than mnt/repo
18:41 rominronin joined #git
18:42 Ragnor joined #git
18:43 grawity mnt/repo/HEAD
18:44 cagedwisdom joined #git
18:44 spacelord_ joined #git
18:45 catern but how would that work?
18:46 catern in practice let's say I'm inspecting some file in some repo at some hash, so I'm looking at mnt/repo/hash/src/something.c
18:46 tristanp joined #git
18:46 catern I edit mnt/repo/hash/src/something.c and what happens?
18:46 renlo_ joined #git
18:47 renlo_ where are the relationships between commits stored? Ie, that Branch A has a HEAD -> commit 1, and then that commit has a parent commit 2, then that commit has a parent commit 3?
18:47 catern does mnt/repo/hash now contain a git repository or what? I think it's much clearer what should happen when you're just dealing with mnt/repo. with just mnt/repo, if I edited something, it's intuitive to me that mnt/repo now points to a locally-checked out repo
18:48 catern locally cloned git repo*
18:49 osse renlo_: Each commit knows its parent(s)
18:49 grawity renlo_: in the commits themselves
18:49 eivarv joined #git
18:49 renlo_ thank you
18:49 grawity see https://git-scm.com/book/en/v2/Git-Internals-Git-Objects
18:52 meLon joined #git
18:53 SwiftMatt joined #git
18:53 cttttt joined #git
18:55 ResidentBiscuit joined #git
18:56 ResidentBiscuit joined #git
18:56 ejb joined #git
18:56 ResidentBiscuit joined #git
18:56 ResidentBiscuit joined #git
18:57 ResidentBiscuit joined #git
18:57 ResidentBiscuit joined #git
18:57 cttttt Got a possibly dumb question.  I'm in the midst of a rebase, and there are a few files that are just fine in the branch I'm rebasing onto.  Looking at the docs, my best option would be to run `git checkout --ours <file>` in these cases.  I'm further learning that `git checkout --theirs <file>` grabs the version in the branch that is being rebased.  Just
18:57 cttttt wondering what to do if I accidentally run either of these and want to the conflicting file back.  Do I need to abort my rebase and start all the way from the beginning?
18:58 mikecmpbll joined #git
18:58 govg joined #git
19:02 sean` joined #git
19:04 ok91 joined #git
19:04 sean` left #git
19:04 dodobrain joined #git
19:06 ahmed_elgabri joined #git
19:06 moritz cttttt: git checkout --merge iirc
19:07 cttttt @moritz Perfect.  Thanks.  I'll give it a shot next time I accidentally run `git rebase --abort` and flip my desk again :)  Should be a coming up in a few minutes :)
19:07 seanhagen joined #git
19:08 cttttt @moritz Just gave it a shot on this hop of the rebase, and it does exactly that.  My patch markers are back in all their glory.  Thanks!
19:11 jimi joined #git
19:11 UserOO7 joined #git
19:13 overlord_tm joined #git
19:13 paultreny joined #git
19:13 tnecniv joined #git
19:17 cdg joined #git
19:18 ResidentBiscuit joined #git
19:18 jimi_sanchez joined #git
19:18 b1tchcakes joined #git
19:19 Macaveli joined #git
19:19 jstimm joined #git
19:20 jimi joined #git
19:22 shashin joined #git
19:22 jimi_sanchez joined #git
19:25 jimi joined #git
19:25 garethdaine joined #git
19:26 overlord_tm joined #git
19:28 duvnell joined #git
19:29 ok91 joined #git
19:30 jimi joined #git
19:31 Silox| joined #git
19:32 jimi joined #git
19:33 hahuang65 joined #git
19:33 jnewt_ joined #git
19:35 Oebele joined #git
19:35 endiruna joined #git
19:39 quackgyver joined #git
19:40 ahmedelgabri joined #git
19:41 mdw joined #git
19:42 nickabbey joined #git
19:42 ResidentBiscuit joined #git
19:42 troppie joined #git
19:43 seishun joined #git
19:44 multi_io_ joined #git
19:45 nikivi joined #git
19:46 ok91 joined #git
19:46 jnewt__ joined #git
19:47 Anja_ joined #git
19:47 Jackneill joined #git
19:47 nicksloan joined #git
19:49 Masber joined #git
19:49 renlo_ joined #git
19:49 jnewt joined #git
19:50 troppie joined #git
19:51 jnewt_ joined #git
19:52 jnavila joined #git
19:53 spacelord_ joined #git
19:54 ugjka joined #git
19:54 dodobrain joined #git
19:56 Learath2 joined #git
19:57 apotry joined #git
19:58 ljubak joined #git
19:58 troppie joined #git
19:59 roelmonnens joined #git
19:59 ok91 joined #git
20:01 JanC_ joined #git
20:02 thebope joined #git
20:03 satifant joined #git
20:12 mdw joined #git
20:13 rominronin joined #git
20:13 nwkj86 joined #git
20:14 jaguarmagenta joined #git
20:15 cyan__ joined #git
20:17 sssilver joined #git
20:18 ash_workz joined #git
20:18 jnewt__ joined #git
20:18 refried_ joined #git
20:19 dreiss joined #git
20:20 msonntag joined #git
20:20 nikivi joined #git
20:21 aidalgol joined #git
20:21 derpsauce__ joined #git
20:23 overlordtm joined #git
20:24 paul424_ joined #git
20:24 overlord_tm joined #git
20:26 anonymuse joined #git
20:26 Phylock joined #git
20:28 JSharpe joined #git
20:28 spacelord_ joined #git
20:29 overlord_tm joined #git
20:29 mischat joined #git
20:29 overlord_tm joined #git
20:32 aswen joined #git
20:33 overlordtm joined #git
20:33 dsantiago joined #git
20:34 alex1a joined #git
20:35 Peng_ joined #git
20:35 ahmed_elgabri joined #git
20:40 dsop joined #git
20:41 overlord_tm joined #git
20:42 paul424 joined #git
20:42 mattp_ joined #git
20:43 dsdeiz joined #git
20:43 dsdeiz joined #git
20:44 overlord_tm joined #git
20:47 willlondon joined #git
20:47 paul424 Hello, I am banned on that channel , can you read my message ?
20:48 ResidentBiscuit I dont know who you are but I can read your message
20:48 perlpilot paul424: I can see you :)
20:48 _ikke_ paul424: Sometimes it looks like you're banned when you are not registered and for example try to change your nick
20:49 paul424 ok . thanks
20:50 willlondon joined #git
20:50 askb joined #git
20:51 qqx joined #git
20:51 noecc left #git
20:54 overlord_tm joined #git
20:56 ejb joined #git
20:56 dsantiago joined #git
20:57 overlord_tm joined #git
20:57 ahmed_elgabri joined #git
20:58 rorro joined #git
20:58 vuoto joined #git
20:59 axl_ left #git
21:00 railssmith joined #git
21:00 eb0t joined #git
21:01 Guest79017 joined #git
21:03 ahmedelgabri joined #git
21:05 overlord_tm joined #git
21:07 shgysk8zer0 joined #git
21:08 eb0t joined #git
21:08 axl_ joined #git
21:09 jnewt_ joined #git
21:09 Rodya_ joined #git
21:10 dreiss joined #git
21:10 Xenoxsis joined #git
21:11 einnjo joined #git
21:14 overlord_tm joined #git
21:15 kyan joined #git
21:16 johnny56_ joined #git
21:16 dsantiago joined #git
21:17 relipse joined #git
21:17 relipse how do I destroy a commit (with a password)
21:19 _ikke_ !sensitive
21:19 gitinfo [!filter_sensitive] You can use filter-branch to remove sensitive data from a repository's history. https://help.github.com/articles/remove-sensitive-data/
21:19 _ikke_ !bfg
21:19 gitinfo A tool designed to remove large files, or passwords from history: https://rtyley.github.io/bfg-repo-cleaner/ (!rewrite applies)
21:19 m0viefreak joined #git
21:19 Turner92 left #git
21:19 PresidentBiscuit joined #git
21:19 jnewt__ joined #git
21:20 adac joined #git
21:21 jnewt joined #git
21:25 exarch joined #git
21:26 nowhere_man joined #git
21:27 Snugglebash joined #git
21:27 anonymuse joined #git
21:29 axl_ joined #git
21:30 overlord_tm joined #git
21:31 salamanderrake joined #git
21:32 mischat joined #git
21:34 overlord_tm joined #git
21:35 fahadash joined #git
21:35 JSharpeNT joined #git
21:35 kpease_ joined #git
21:36 overlord_tm joined #git
21:37 hariel joined #git
21:41 ahmed_elgabri joined #git
21:42 cdg joined #git
21:42 Guest82103 joined #git
21:42 Guest82103 joined #git
21:43 dsdeiz joined #git
21:43 dsdeiz joined #git
21:43 dunpeal joined #git
21:45 rominronin joined #git
21:45 ToBeCloud joined #git
21:46 nidr0x joined #git
21:49 cagedwisdom joined #git
21:49 overlord_tm joined #git
21:53 Gsham joined #git
21:53 Derperperd joined #git
21:55 mquin joined #git
21:56 dikiaap joined #git
21:59 rwb joined #git
22:03 fmcgeough joined #git
22:04 UserOO7 joined #git
22:05 malo[m] joined #git
22:06 hhee joined #git
22:06 blackwind_123 joined #git
22:09 Rodya_ joined #git
22:09 moei joined #git
22:09 dunpeal Hi. I have quite a few branches that are alive, but not actively being worked on. I'd like to hide them from the `git branch` list, so only active branches are displayed. Is it a reasonable solution to convert them to tags?
22:11 dminuoso dunpeal: You could rename them to refs/stuff/branch I guess
22:12 perlpilot dunpeal: What's a "non-active branch"?  Why are you keeping them around?
22:12 dminuoso perlpilot: Plenty reasons. Features you are not working on right now.
22:12 dunpeal dminuoso: how is that better than making them tags?  BTW, if they're not tags or branches, they won't be tracked by origin, and may possibly be garbage-collected, no?
22:13 dminuoso dunpeal: It's not I suppose.
22:13 dminuoso Just trying to come up with ideas.
22:13 dunpeal perlpilot: yup, like dminuoso said, most common is features that are low priority, and will eventually get done, but not right now.
22:13 SirCmpwn joined #git
22:14 Vinnie_win_x joined #git
22:14 dunpeal I have branches with changes that eventually will be completed and merged, but haven't been worked on for months, and probably won't be for some months more.
22:14 perlpilot dunpeal: I'd probably make a git alias that only showed the branches that fit my definition of "active"
22:14 dodobrain joined #git
22:14 dminuoso dunpeal: You could make a new repo.
22:15 dminuoso I mean it's not very pretty, but that would be a fairly quick way to tackle it.
22:15 nedbat dunpeal: you could make make a new branch from "foo" called "inactive/foo" and then not worry about it.
22:16 dodobrain joined #git
22:17 dunpeal nedbat: sure, but I'm not sure I see the downside of converting it to a tag
22:17 nedbat dunpeal: ok
22:17 dunpeal I could create such a prefix, then modify the way I use `git branch` and some other branch-displaying/processing scripts to ignore the inactive ones
22:17 gargola joined #git
22:18 dunpeal But it seems like just converting to tags is very easy, cheap, and simple.
22:18 kadoban dunpeal: Well ... it's not really a tag in semantics at all, so I don't see the benefit. In a while aren't you going to be asking "hey I have a million tags that are stale, what can I do about them?"
22:19 dunpeal kadoban: this might be YMMV, but I feel like tags are much more likely to multiply, and I rarely care about how many there are, or displaying some/all of them.
22:19 cheater joined #git
22:19 cheater hi
22:19 nickabbey joined #git
22:20 dunpeal for example, most GUIs assume you have a small number of branches, and having too many breaks the GUI, while no such assumptions or requirements are made about the number of tags.
22:20 kadoban dunpeal: Seems pretty YMMVy. Also, what happens when they get pushed places, especially when they later need to change but can't?
22:20 dunpeal kadoban: later I delete the tag, and create a branch in its stead.
22:20 PresidentBiscuit joined #git
22:21 kadoban dunpeal: And the tag in everyone else's repo?
22:21 dunpeal Sure. How is that different from any other tag that gets used for a while and deleted?
22:21 cheater i have a clone of a remote repository on my disk, called x. i would like to do git clone x y to quickly get another clone, but have y still have the remotes of x, rather than have x as its upstream - how would i do that?
22:21 dunpeal What costs / downsides do this incur?
22:22 dunpeal cheater: just copy the /x directory
22:22 cheater that's not going to do any linking
22:22 dunpeal cheater: what linking?
22:23 cheater as i understand, when you clone, files in the new one are links to files in the old one
22:23 dunpeal cheater: no, that is incorrect.
22:24 dunpeal clone creates a copy of the repository you cloned, files and all.
22:24 dunpeal you might be confusing with various COW file system features, that exist on a lower abstraction level than the one in which you are operating.
22:25 dunpeal (and therefore shouldn't concern you)
22:25 misha joined #git
22:26 ToxicFrog dunpeal: if you clone a local filesystem path, it will use hardlinks for the contents of .git/objects where possible
22:26 Vinnie_win joined #git
22:26 ToxicFrog cheater is correct here.
22:27 gchristensen joined #git
22:27 gchristensen Hi, I'd like to be signing and verifying `git-notes` commits. The only way I've found to do this is by checking out refs/notes/commits myself and commiting my notes directly with `git-commit`. Is there an easier way to do this?
22:28 mischat joined #git
22:28 mischat joined #git
22:29 dunpeal ToxicFrog: sure, but won't that be also done on any COW filesystem if you just copy as well?
22:31 Hello71 only if you ask it
22:31 Hello71 the filesystem can't automatically know to reflink
22:31 ToxicFrog dunpeal: not generally, no. If you 'cp' a file on ZFS or btrfs, the blocks are written anew.
22:32 Hello71 isn't that what I said
22:32 dunpeal ToxicFrog: thanks, I learned something.
22:32 ToxicFrog copy-on-write applies to how modifications to existing data and snapshots and stuff work. It doesn't imply deduplication.
22:32 ToxicFrog (although ZFS has a separate deduplication setting you can turn on, if you have lots and lots of RAM)
22:33 cheater COW will not make the files under .git/objects the same inode number
22:34 cheater 5142550 xxxxx-clone2/.git/objects/26/4131970f1bbb70e7166e626bdd64885e2063bb
22:34 cheater 5142550 xxxxx-clone3/.git/objects/26/4131970f1bbb70e7166e626bdd64885e2063bb
22:34 cheater 5142550 xxxxx-clone/.git/objects/26/4131970f1bbb70e7166e626bdd64885e2063bb
22:34 cheater 5142550 xxxxx/.git/objects/26/4131970f1bbb70e7166e626bdd64885e2063bb
22:34 cheater i cloned xxxxx to xxx-clone, cloned that to clone2, and then cloned the original to clone3
22:36 ToxicFrog cheater: anyways, to answer your earlier question, the easiest way is probably to 'git clone' the original and then overwrite its .git/config with the original's .git/config
22:37 cheater ToxicFrog: that's what i've been doing, but i was hoping there was a command line switch
22:37 ToxicFrog Not as far as I know.
22:39 cheater thanks ToxicFrog
22:40 cheater btw, git clone also abandons any uncommitted changes from the original, right?
22:41 cheater so that's another difference to just copying
22:42 derk0pf joined #git
22:43 venmx joined #git
22:43 dunpeal cheater: you can just clone, then `git remote set-url`
22:44 dunpeal and yes, git clone should only clone the actual repo, and not the working directory or any changes thereto that aren't commited
22:46 jstein_ joined #git
22:46 cheater i think set-url will be more involved than just copying the old .git/config over the new one because i'd have to look up the url
22:47 jacob123 joined #git
22:48 whlai I have/had a whole *bunch* of dangling objects and did a: git gc --prune=now
22:48 whlai Why does it take so long?
22:49 whlai Shouldn't it be a simple operation?
22:49 Hello71 has to recompress your packs
22:49 whlai Why?
22:49 Hello71 because you told it to
22:49 whlai Aren't I just deleting those danglinng objects?
22:49 Hello71 they might be in a pack
22:49 Oatmeal joined #git
22:49 Darren_ joined #git
22:50 whlai can you ELI5 it to me what I'm doing with that command?
22:50 ToxicFrog ELI5?
22:50 Hello71 if you want to just delete the unpacked unreachable objects then git prune
22:50 whlai Explain it like I'm 5
22:50 whlai what does it mean to be "packed"
22:51 Hello71 like a seekable tarball
22:51 _28_ria joined #git
22:51 sea-gull_ joined #git
22:51 whlai ok. so it compresses all the changes into various compressed objects
22:52 Hello71 no, it compresses *objects* into *packs*
22:52 ToxicFrog whlai: accessing lots of individual files can be slow. Git periodically combines those files into "packfiles" which can be accessed more efficiently and compressed more effectively. They're in .git/objects/pack/.
22:52 Hello71 accessing the files isn't slow
22:52 ToxicFrog ("those files" == the files representing the various objects in the repo)
22:52 Hello71 it just takes up lots of space because git doesn't store diffs
22:52 ToxicFrog Hello71: I'm making the possibility-unwarranted assumption here that calling readdir() a bunch is more expensive than reading a packidx
22:53 Derperperd joined #git
22:53 compeman joined #git
22:53 jacob123 joined #git
22:53 Hello71 in Linux readdir uses the getdents system call
22:53 whlai what is the most that can be pushed at once and where are the bottlenecks?
22:54 Hello71 so the overhead is "about the same" as calling a similarly complex function to parse your own structure
22:54 Hello71 on Windows I think there are also more efficient interfaces than FindNextFile
22:54 ToxicFrog Aah.
22:54 johnmilton joined #git
22:55 Hello71 whlai: no limit "at once", but you cannot store more than some amount of data in Git (a very large amount)
22:55 Hello71 assuming you are not using FAT
22:55 whlai how large?
22:55 Hello71 large enough
22:56 whlai under 100 G?
22:56 Hello71 if you run into it with present-day hardware you will have other problems
22:56 Hello71 much higher than that.
22:56 whlai gotcha
22:56 Hello71 I think at least exabytes
22:56 whlai my bottlenecks are what? RAM?
22:56 whlai whoa
22:56 enleeten joined #git
22:57 Hello71 AFAIK the only starting-to-be-obsolete limitation in Git is SHA1
22:57 whlai how so?
22:57 arescorpio joined #git
22:57 Siegfried joined #git
22:57 whlai where would SHA1 come into play as a limiter?
22:57 ToxicFrog Once you get above a certain size some git commands start to misbehave, though -- "git gc" requires an unreasonable amount of RAM once your repo is in the hundreds of GB and tens of millions of objects range, for example
22:58 Hello71 it is considered deprecated in terms of cryptographic security
22:58 whlai yeah, but how does that limit git?
22:58 dzho if you can't trust your hashes, what is the point of hashes?
22:58 Hello71 git GPG is based on the SHA-1 hashes
22:58 armin joined #git
22:59 dzho and that
22:59 gchristensen and you don't really want "The data model that Git uses ensures the cryptographic integrity of every bit of your project." to become false
22:59 whlai where is that used? for data transport between local and remote?
22:59 Hello71 for everything
23:00 Hello71 you can prove that your git history is exactly identical to someone else's by sharing git rev-parse HEAD
23:00 whlai So it deprecated, meaning an up-to-date version of git isn't using it, eys?
23:00 gchristensen git uses sha-1
23:00 whlai still?
23:00 dzho forever
23:00 gchristensen sha1 as a cryptographically secure hash is considered deprecated
23:01 whlai i know that
23:01 whlai so why is git still using it
23:01 Derperperd joined #git
23:01 fmeerkoetter joined #git
23:01 Hello71 deprecated is not such an accurate term I used, "considered insufficient for high-security applications"
23:01 dzho because it has always used it
23:01 anonymuse joined #git
23:01 whlai aren't there plans to replace it ?
23:01 Muzer joined #git
23:01 gchristensen dzho: I could imagine a way to move forward away from sha1
23:01 dzho to not use it might mean changing every repo ever
23:01 dzho object by object
23:01 Hello71 because nobody has put in the (tremendous) work needed to change it in Git, and it is not a huge issue
23:01 osse There are plans to make git more hash agnostic
23:02 Hello71 yes, but those are a long ways off
23:02 dzho well, that leaves us wondering where the balance is between "not a huge issue" and "a long ways off"
23:02 osse At least they're working on not hardcoding "20"  eberyw :P
23:02 kadoban The places where it matters if it's cryptographically secure, in git, are fairly limited thankfully. Though not nonexistant.
23:02 Hello71 the largest user of Git, the Linux kernel, I believe does not use GPG signatures
23:02 osse Everywhere
23:03 osse For the most part it's only used as a checksum
23:03 Hello71 eh? the design of (present-day) Git requires that hashes are unique
23:03 Hello71 everything is addressed by its hash
23:03 qqx Even GPG signatures don't really help, since when used with git they're generally just signing the ID of a git object which will still be a SHA1.
23:03 whlai are you saying that they are also using sha-1 for the transportation layer from remote to local and vice versa?
23:03 Hello71 but my point is that they rely on other methods to ensure security
23:04 Hello71 whlai: that doesn't really make sense
23:04 Hello71 hashes are not encryption
23:04 whlai but you can hash data for transposrt
23:05 railssmi_ joined #git
23:05 kadoban You can, for some purposes. Here I suspect it's not particularly relevant to the dicussion though.
23:06 dzho whlai: so, here's the thing--someone could modify a commit object in git while keeping the hashes for it the same, if they threw enough computing power at the problem, making the change very hard to find via normal use of git.
23:06 Hello71 generally Git transfers are secured by TLS or SSH
23:06 Snugglebash joined #git
23:06 dzho the change *could* be introduced during transfer, but it needn't be, and yeah that threat can be reasonably covered by transport encryption.
23:06 whlai ok. nm. I got confused for a sec. Transposrt will be determined by the remote repo, right? i.e. it'll use the algos that they have chosen for their TLS implementation.
23:07 mellernoia joined #git
23:07 dzho I think you're barking up the wrong tree by focussing on transport.
23:07 kadoban TLS or SSH, more often probably SSH I think
23:07 osse Isn't the checksum used in tcp packets very primitive? Wake up sheeple!
23:07 dzho goodness knows we haven't had *any* problem with any of the internet protocols lately.
23:08 whlai I'm trying to understand the attack vectors on private repos
23:08 PresidentBiscuit joined #git
23:08 dzho why just private repos?
23:08 whlai Because I'm assuming security on the local host and remote repo
23:09 dzho why would you do that?
23:09 whlai *good security
23:09 osse If you have write access in the first place and is able to create a commit impersonating someone else that contains harmful code that matches a sha1 in the repo already, then...
23:09 whlai Well, I know my own box. The other side is Atlassian and I'm giving them the benefit of the doubt for my thought experiment.
23:10 dzho "good" security isn't just preventing a thing from happening, it's also detecting that a thing has occurred so it can be mitigated.
23:10 whlai Right, but I don't have other users with write.
23:10 whlai So it's "good" for my use.
23:10 whlai and I log everything as well
23:10 BlessJah whlai: sha1 hash is recursively calculated using sha1 hashes of tree-like structure of directories and files
23:11 dzho it's turtles, all the way down
23:11 qqx whlai: If your threat model is that no untrusted users can write to any repo you interact with, SHA1 should be fine for a long time to come.
23:11 BlessJah assuming attacker were to replace some of the files (just some, so you would not notice), he would need to find collisions all the way down
23:11 ochorocho__ joined #git
23:12 ToxicFrog BlessJah: "all the way down" is probably depth 1, though
23:12 railssmith joined #git
23:12 ToxicFrog Like, if I want to clandestinely modify some files, I only need to find collisions for those files
23:12 andlabs joined #git
23:13 ToxicFrog They have the same hash as the original, so the trees that reference them don't change
23:13 Hello71 that's assuming they haven't been subsequently modified
23:13 BlessJah yep, that's about right
23:13 Hello71 if you want to persist those modifications you'll have to find collisions for all subsequent versions of that file
23:13 dwmw2_gone joined #git
23:13 Hello71 or more precisely that blob
23:13 BlessJah you're storing binary data there, don't you?
23:13 BlessJah :>
23:13 whlai :D
23:14 BlessJah with text data finding meaningful collision is harder than with binary where you can play with paddings and many container will simply ignore trailing garbage
23:16 Noldorin joined #git
23:16 rominronin joined #git
23:17 whlai thanks yall.
23:17 whlai im vamoose
23:20 Dougie187 left #git
23:22 cheater left #git
23:24 WaReZ joined #git
23:27 Oatmeal joined #git
23:27 derk0pf joined #git
23:28 nowhere_man joined #git
23:31 fatalhalt joined #git
23:33 nowhere_man joined #git
23:33 railssmi_ joined #git
23:37 a3Dman joined #git
23:42 einnjo joined #git
23:44 Emperor_Earth joined #git
23:47 anonymuse joined #git
23:47 tribly joined #git
23:52 justanotheruser joined #git
23:55 renlo joined #git
23:55 Darren_ joined #git
23:56 TomyLobo joined #git
23:57 ShapeShifter499 joined #git

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