Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-08-24

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

All times shown according to UTC.

Time Nick Message
00:00 monoprotic joined #git
00:00 monoprotic joined #git
00:02 NeverDie joined #git
00:04 mjuszczak joined #git
00:05 diphtherial left #git
00:06 sebboh joined #git
00:08 deeeepsteep joined #git
00:09 mingrammer joined #git
00:11 madewokherd joined #git
00:12 red_ joined #git
00:12 zacts joined #git
00:13 dendazen joined #git
00:15 alansaul joined #git
00:16 Gsham joined #git
00:22 jimi_ joined #git
00:26 anonymuse joined #git
00:26 Kaisyu joined #git
00:26 anonymuse joined #git
00:29 netj joined #git
00:33 mjuszcza_ joined #git
00:33 rominronin joined #git
00:34 mjuszcza_ joined #git
00:36 anuxivm left #git
00:36 mjuszcza_ joined #git
00:38 thiago joined #git
00:38 atomicb0mb joined #git
00:39 jangeador_ joined #git
00:41 SwiftMatt joined #git
00:41 mjuszczak joined #git
00:42 mjuszczak joined #git
00:43 monoprotic joined #git
00:43 mjuszczak joined #git
00:44 tlaxkit joined #git
00:44 linuxmint joined #git
00:44 JamesBaxter joined #git
00:46 mjuszczak joined #git
00:46 mjuszczak joined #git
00:47 mjuszczak joined #git
00:47 ash_workz joined #git
00:48 peepsalot joined #git
00:49 hwrdprkns joined #git
00:49 _sjs joined #git
00:50 tesuji joined #git
00:54 monoprotic joined #git
00:55 frivol joined #git
00:56 ochorocho__ joined #git
00:58 _August_ joined #git
00:59 _August_ anyone here use phpstorm?
00:59 diprotic joined #git
00:59 _August_ im wondering if there is a way to switch branches and keep the same files in my local repo
01:00 _August_ without them switching everytime i checkout a different brance
01:00 _August_ branch
01:01 rewt why would you want to do that?
01:01 heiligenschein joined #git
01:01 modocache joined #git
01:02 antipsychiatry joined #git
01:03 atomicb0mb joined #git
01:03 Vampire0 joined #git
01:05 raijin joined #git
01:07 heiligenschein i want to install the lwfinger/rtlwifi_new gib with following commands http://pastebin.com/4xcquJvb , everything works fine till the command "make", at first i got this error msg http://pastebin.com/YmYDgwah , than somebody told me to install the kernel-headers with sudo apt-get install linux-headers-$(uname -r) , now after installing those i  get this error msg: http://pastebin.com/MQkvUUbW
01:07 heiligenschein os: debian
01:07 Es0teric joined #git
01:09 sssilver joined #git
01:11 raijin joined #git
01:12 d0nn1e joined #git
01:12 HoierM joined #git
01:16 rkazak joined #git
01:17 ytl joined #git
01:17 sebboh joined #git
01:22 mjuszczak joined #git
01:26 harish joined #git
01:26 ahr3n joined #git
01:27 anonymuse joined #git
01:31 modocache joined #git
01:32 daurnimator joined #git
01:35 rominronin joined #git
01:35 RedneckDonut joined #git
01:36 anonymuse joined #git
01:36 fahadash joined #git
01:38 inflames joined #git
01:38 cagmz joined #git
01:39 gmh joined #git
01:46 _August_ joined #git
01:46 jeffreylevesque_ joined #git
01:46 monoprotic 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 fakenerd joined #git
01:52 MattMaker joined #git
01:56 boshhead joined #git
01:56 pks joined #git
01:57 justanotheruser joined #git
01:57 mingrammer joined #git
01:58 evanwang joined #git
02:01 d^sh joined #git
02:02 kendocoding joined #git
02:03 rchavik joined #git
02:06 Dougie187 joined #git
02:11 netj joined #git
02:12 chachasmooth joined #git
02:12 zacts joined #git
02:16 rubyonrailed joined #git
02:17 MattMaker joined #git
02:18 sangy joined #git
02:25 deeeepsteep joined #git
02:28 hexagoxel joined #git
02:28 durham_ joined #git
02:29 Goplat joined #git
02:30 Gsham joined #git
02:30 griffindy joined #git
02:33 sebboh joined #git
02:35 mjuszczak joined #git
02:36 rominronin joined #git
02:36 deeeepsteep joined #git
02:37 MattMaker joined #git
02:39 cqi joined #git
02:40 safe joined #git
02:50 fstd_ joined #git
02:52 mingrammer joined #git
02:52 MattMaker joined #git
02:54 LeBlaaanc joined #git
02:55 jangeador_ joined #git
02:57 ochorocho__ joined #git
02:58 le_melomane joined #git
03:00 LeBlaaan_ joined #git
03:02 aavrug joined #git
03:05 l4v2 joined #git
03:07 MattMaker joined #git
03:08 mokbu joined #git
03:10 raijin joined #git
03:17 zacts joined #git
03:18 rubyonrailed joined #git
03:20 Dougie187 left #git
03:24 mohabaks joined #git
03:31 ljc joined #git
03:35 mjuszczak joined #git
03:36 ibourgeois joined #git
03:37 rominronin joined #git
03:42 MattMaker joined #git
03:44 sebboh joined #git
03:44 AnarchyAo joined #git
03:45 hwrdprkns joined #git
03:45 mehola joined #git
03:46 Es0teric joined #git
03:47 rubie joined #git
03:47 vassagus_ joined #git
03:49 dsantiago joined #git
03:51 afuentes joined #git
03:57 MattMaker joined #git
04:00 thecomedian joined #git
04:01 SpeakerToMeat joined #git
04:05 Es0teric joined #git
04:11 ShekharReddy joined #git
04:12 MattMaker joined #git
04:13 Cabanossi joined #git
04:14 nyexpress joined #git
04:16 stonercoder joined #git
04:16 Peng__ joined #git
04:18 howdoi joined #git
04:24 anarchos2 joined #git
04:24 anarchos2 hi, i'm trying to output all commits with a date using: git log --oneline --all > gitlog
04:25 anarchos2 i tried --date=iso
04:29 anarchos2 ah, i got it.  --pretty=medium works for my needs
04:30 zacts joined #git
04:34 phanimahesh joined #git
04:34 mahendra_ joined #git
04:36 netj joined #git
04:38 rominronin joined #git
04:38 Vaelatern joined #git
04:38 Xizor joined #git
04:40 watabou joined #git
04:40 mingrammer joined #git
04:41 a3Dman_ joined #git
04:42 phanimahesh joined #git
04:46 Xizor joined #git
04:47 cyphase_eviltwin joined #git
04:47 phanimahesh joined #git
04:47 alansaul joined #git
04:47 MattMaker joined #git
04:47 ahr3n joined #git
04:53 fakenerd joined #git
04:53 mahendra_ joined #git
04:54 fakenerd joined #git
04:54 mahendra_ joined #git
04:57 msonntag joined #git
04:58 ochorocho__ joined #git
04:59 subhojit777 joined #git
05:00 leeN joined #git
05:00 _rht joined #git
05:01 raijin joined #git
05:01 sebboh joined #git
05:01 rubie joined #git
05:01 ljc joined #git
05:02 MattMaker joined #git
05:03 jnewt joined #git
05:04 jnewt i am working with a repo if i try to diff a file, all i get is a warning about LF being replaced by CRLF in that file and that the file will have it's original line endings in your working directory.
05:05 jnewt i have core.autocrlf set to false, but status still shows the file as modified
05:05 jnewt is there actually any modifications?  how can i see them?
05:06 rubyonrailed joined #git
05:06 LeBlaaanc joined #git
05:07 freimatz joined #git
05:07 jnewt i may have presented that wrong.  when i change to core.autocrlf false, status shows differences, but diff does not.  otherwise i get an error.   how do i get status and diff to agree?
05:11 Es0teric joined #git
05:18 phanimahesh joined #git
05:22 MattMaker joined #git
05:24 weox joined #git
05:27 parsnipM_ joined #git
05:27 parsnipM_ left #git
05:32 fakenerd joined #git
05:35 mingrammer joined #git
05:37 navidr joined #git
05:39 navidr joined #git
05:39 rominronin joined #git
05:39 tvw joined #git
05:40 ThomasLocke joined #git
05:40 ThomasLocke joined #git
05:41 navidr joined #git
05:42 Rish joined #git
05:43 omegak joined #git
05:43 LeBlaaan_ joined #git
05:44 linuxmint joined #git
05:45 _ng joined #git
05:45 adac joined #git
05:49 ThomasLocke joined #git
05:50 stuh84 joined #git
05:54 netj joined #git
05:55 phanimahesh joined #git
05:55 zarel joined #git
05:55 anarchos2 joined #git
05:57 diogenese joined #git
06:00 hahuang61 joined #git
06:00 kyan joined #git
06:01 deeeepsteep joined #git
06:02 MattMaker joined #git
06:03 jtprog_ joined #git
06:08 satya4ever joined #git
06:11 ochorocho__ joined #git
06:12 beautifulsoup4 joined #git
06:12 MattMaker joined #git
06:14 deepy If someone a good while ago did a merge by replacing all the files with the copy from upstreams and then made a new commit, how would you go about repairing that?
06:16 phanimahesh joined #git
06:16 sebboh joined #git
06:20 konrados joined #git
06:21 AnarchyAo joined #git
06:21 zacts joined #git
06:21 konrados Hi :)
06:21 Andrew_K joined #git
06:22 Ravana joined #git
06:22 konrados May I simply move an entire project, along with the .git directory into another directory?
06:22 MattMaker joined #git
06:22 ShekharReddy joined #git
06:25 roelmonnens joined #git
06:25 Brain joined #git
06:25 JeroenT joined #git
06:25 tjone270_ left #git
06:26 j416 konrados: yes you may
06:26 JeroenT joined #git
06:26 konrados Thank you j416!
06:27 bocaneri joined #git
06:27 j416 konrados: git only cares about the working directory (normally where .git is in) and the .git directory. Where on disk it is located does not matter.
06:28 konrados OK, I've just done it and now testing the results:)
06:32 MattMaker joined #git
06:33 moei joined #git
06:35 indistylo joined #git
06:36 rominronin joined #git
06:39 lacrymology joined #git
06:41 kuldeep_ joined #git
06:41 ochorocho__ joined #git
06:41 Macaveli joined #git
06:44 ochorocho__ joined #git
06:45 ytl joined #git
06:45 sdothum joined #git
06:46 ljc joined #git
06:47 fakenerd joined #git
06:52 rubie joined #git
06:52 submitnine joined #git
06:54 rubyonrailed joined #git
06:55 vassagus joined #git
06:55 tinajohnson___ joined #git
06:56 Sceorem joined #git
06:57 fees joined #git
06:58 mariuscc joined #git
07:00 bentinata joined #git
07:00 pks joined #git
07:00 M-ou-se joined #git
07:01 bentinata Hello, I have a project based on another project (base project). What's the best approach to integrating some code from derivation to base?
07:03 bentinata I am thinking about cherry-picking commit to base repositories, but, suppose there's better flow for this.
07:03 pks joined #git
07:04 zrts joined #git
07:05 rishi joined #git
07:07 heroux joined #git
07:07 Hounddog joined #git
07:10 bronson joined #git
07:10 vassagus_ joined #git
07:10 Darcidride joined #git
07:10 ldlework se
07:11 ldlework (sorry)
07:11 woodruffw joined #git
07:11 woodruffw joined #git
07:12 pks joined #git
07:12 woodruffw joined #git
07:12 woodruffw joined #git
07:13 woodruffw joined #git
07:14 TomyWork joined #git
07:19 alansaul joined #git
07:20 sebboh joined #git
07:23 mingrammer joined #git
07:24 encod3 joined #git
07:26 Balliad joined #git
07:26 Manuel15 joined #git
07:26 Manuel15 joined #git
07:27 kurkale6ka joined #git
07:27 MattMaker joined #git
07:29 jokajak joined #git
07:29 davisonio joined #git
07:30 Balliad joined #git
07:30 weckl joined #git
07:31 Raging_Hog joined #git
07:35 xaa joined #git
07:36 jceb joined #git
07:36 circ-user-cmOBz joined #git
07:37 MattMaker joined #git
07:41 vuoto joined #git
07:43 davisonio joined #git
07:46 Hko joined #git
07:49 adac joined #git
07:51 _ng joined #git
07:53 boshhead_ joined #git
07:54 vuoto joined #git
07:54 infra-red joined #git
07:55 ahmedelgabri joined #git
07:55 heroux joined #git
07:57 MattMaker joined #git
07:57 vassagus joined #git
07:58 lordjancso joined #git
08:00 bongjovi joined #git
08:01 abyx joined #git
08:02 anarchos2 joined #git
08:02 sdothum joined #git
08:02 hahuang61 joined #git
08:04 abyx joined #git
08:05 vuoto joined #git
08:06 elect_ joined #git
08:06 elect_ if I did some modifications and I want to keep them separated by committing to a new branch, how shall I do?
08:07 elect_ http://stackoverflow.com/questions/1394797/move-existing-uncommited-work-to-a-new-branch-in-git?
08:07 MattMaker joined #git
08:08 tobiasvl elect_: yes. !float
08:08 gitinfo elect_: If you have made a change in your working directory and have NOT YET COMMITTED, you may "float" that change over to another (`git checkout oldbranch`) or new (`git checkout -b newbranch`) branch and commit it there.  If the files you changed differ between branches, the checkout will fail.  In that case, `git stash` then checkout, and `git stash apply` and go through normal conflict resolution.
08:08 elect_ ok thanks
08:08 marcogmonteiro joined #git
08:11 disse joined #git
08:11 chll_ joined #git
08:12 boogibugs joined #git
08:12 watabou joined #git
08:16 disse hey guys, I'm quite new to git. I'm struggling for about 2 hours trying simply to send a patch via E-Mail. The patch is located on an embedded tegra. I copied it to my local maschine. I want to send it using thunderbird via exchange which only has imap folders. First of all I tried sending it directly without thunderbird. I was lost at exchange (I'm not responsible that I have to use it :( ) with this error: 5.7.3
08:16 disse Authentication unsuccessful.
08:16 Hetman joined #git
08:16 disse Then I tried to open the email with thunderbird. but opening the patch, it only shows me an email which "is not able to be send"
08:17 vuoto joined #git
08:17 abyx joined #git
08:17 Hetman left #git
08:17 disse additionally after installing the compose plugin in thunderbird I couldn't find the  bar or where to add it in view because I can't even see there a menu called toolbars
08:18 jknetl joined #git
08:19 JeroenT joined #git
08:21 sebboh joined #git
08:22 ifra201 joined #git
08:23 cyan__ joined #git
08:23 robotroll joined #git
08:27 MattMaker joined #git
08:28 heroux joined #git
08:29 warthog9 joined #git
08:32 liuzhen joined #git
08:33 warthog9 joined #git
08:33 mariuscc joined #git
08:34 aiena joined #git
08:34 aiena I checkout an older commit then I made lots of changes to my code. Now I want to commit again my HEAD is in a detached state what does this mean and how do I correctly commit ?
08:34 canton7 !detached
08:34 gitinfo A detached HEAD(aka "no branch") occurs when your HEAD does not point at a branch. New commits will NOT be added to any branch, and can easily be !lost. This can happen if you a) check out a tag, remote tracking branch, or SHA; or b) if you are in a submodule; or you are in the middle of a c) am or d) rebase that is stuck/conflicted. See !reattach
08:35 aiena !reattach
08:35 gitinfo Letters refer to !detached. (a) and (b): 'git checkout branchname' to continue working on another branch, or 'git checkout -b branchname' to start a new one here; (c) git am --continue; (d) git rebase --continue
08:35 maxorator joined #git
08:35 aiena !rebase
08:35 gitinfo 'git rebase' takes away your local commits, updates your branch with new stuff from <upstream> (argument), then re-applies your local commits on top. This makes it look like your commits were created "after" the new stuff, and it can look cleaner than doing a !merge. Beware of !rewriting_public_history, though. Not to be confused with !interactive_rebase.
08:36 aiena canton7: so If I rebase and then commit I wont loose loca changes currently I have only a local git repo
08:37 aiena I mean local untracked changes
08:37 canton7 git won't let you rebase if you have untracked changes
08:37 canton7 the first thing is to create a new branch and commit, so that you don't lose anything
08:38 canton7 git checkout -b branchname; add, commit
08:38 canton7 right, now you'll have a new branch which just contains your new changes. what do you want to do with them? does it make sense to reapply those changes to some other branch? what are you trying to achieve
08:40 ertesx joined #git
08:40 h- joined #git
08:41 aiena canton7: Now I want to remove this branch and place all changes in master
08:41 canton7 aiena, so you want to move those changes onto the top of master, as a new commit?
08:41 canton7 git checkout master; git cherry-pick branchname;
08:41 aiena yes
08:41 canton7 and if that goes successfully, you can delete branchname
08:43 rubyonrailed joined #git
08:43 deepy is there a way to give git diff a list of the files I'm interested in? I have a file containing ~500 files and these are the only ones I'm interested in
08:43 heroux joined #git
08:44 aiena canton7: there is an error how do I resolve the errors
08:44 canton7 aiena, I can't possibly guess what you mean by "there is an error"
08:44 aiena it only gives 3 hints but no other useful info
08:44 canton7 do you mean there is a merge conflict?
08:44 canton7 what exactly did git tell you?
08:44 aiena one sec posting a paste
08:44 canton7 ... remember that I can't read your mind or what's on your screen ;)
08:45 aiena http://paste.opensuse.org/1d65d29b is what git told me branchname is Refactor_data
08:45 aiena canton7: I know pasted
08:46 canton7 right, so you have conflicts. You checked out an older commit, and made some changes. Git tried to make the same changes to the latest commit, but the lines of code that it's trying to change have also been modified between that old commit and the latest commit
08:46 canton7 'git status' will tell you which files have conflicts in them
08:46 canton7 edit each one, look for the merge markers >>>>>>> and <<<<<<<<, and edit those sections so that they look correct
08:47 MattMaker joined #git
08:47 CheckDavid joined #git
08:47 canton7 !eekaconflict
08:47 gitinfo Merge conflicts are a natural part of collaboration. When facing one, *don't panic*. Read "How to resolve conflicts" in man git-merge and http://git-scm.com/book/ch3-2.html#Basic-Merge-Conflicts then carefully go through the conflicts. Picking one side verbatim is not always the right choice! A nice video explaining merge conflicts: https://www.youtube.com/watch?v=zz7NuSCH6II
08:47 canton7 ^ although that's for merge conflicts, not cherry-pick conflicts
08:47 canton7 when you've edited all of the conflicts, 'git add' each conflicted file, and 'git cherry-pick --continue' to finish the cherry-pick
08:49 aiena canton7: http://paste.opensuse.org/27f8d11e I think the unmerged paths are causing the issue
08:49 ahr3n joined #git
08:49 canton7 they *are* the issue. The unmerged paths are the conflicts
08:49 _ng joined #git
08:49 canton7 so 'both added' means that, since that old commit, both master and you have added a file called 'PrintData.php'
08:49 aiena hmm let me readup on merged paths I dont understand what they are and why they are a problem
08:50 canton7 'both modified' means that both master and you edited the same parts of the file 'submitChoice.php'
08:50 aiena and  'added by us'
08:50 canton7 can't remember :P you'll have to check the docs
08:54 rubie joined #git
08:56 armin joined #git
08:57 aiena hmm ok now I know the conflicts exists but now I need to learn how to reasolve them
08:57 Pursuit joined #git
08:58 courrier joined #git
08:58 shirty joined #git
08:59 fakenerd joined #git
08:59 aiena canton7:  for both added if I want to keep the changes in the new branch but remove teh changes in master what do I do ?
09:00 anticom joined #git
09:01 ki0 joined #git
09:01 canton7 aiena, you can 'git checkout --theirs path/to/the/file'
09:01 canton7 but
09:02 canton7 that will choose the whole file from the new branch, and discard the one from master
09:02 canton7 it won't look at each conflict individually
09:02 MattMaker joined #git
09:04 jeffreylevesque joined #git
09:05 itdjj joined #git
09:05 AlvinLee joined #git
09:06 aiena hmm ok I goofed up
09:06 aiena I accidentally deleted a file
09:06 marcogmonteiro joined #git
09:06 aiena so git status shows its deleted
09:06 aiena if I checkout that file can I get it back
09:07 pro_metedor joined #git
09:08 aiena it is an untracked deletion in the working dir
09:08 itdjj Hi everyone, I'm currently trying to write a script which is automatically moving directories/files on multiple branches and merging that afterwards. (release/1.0 -> release/2.0 -> master) While doing that I get merge conflicts, when applying release/1.0 to release/2.0 due to moved/deleted files. Is there any possiblity to automtically solve those conflicts using "mine"?
09:08 _ng joined #git
09:09 _ng joined #git
09:09 AlvinLee left #git
09:09 pro_metedor lacrymology: I had those problems yesterday. I always do a local copy for backup
09:09 itdjj I tried "git merge -s recursive -X ours release/1.0" but that doesn't work, git still returns the merge conflicts.
09:10 yuhlw_ joined #git
09:10 paws_ joined #git
09:10 mingrammer joined #git
09:15 Ricaz joined #git
09:17 spudowiar joined #git
09:17 MattMaker joined #git
09:18 abyx joined #git
09:22 Ed___ joined #git
09:23 Ed___ left #git
09:24 Raging_Hog joined #git
09:26 Mobutils joined #git
09:27 Seveas aiena: git status shows you how to get the file back :)
09:27 MattMaker joined #git
09:31 ljc joined #git
09:31 sebboh joined #git
09:32 Joost joined #git
09:34 adac hmm what is actually causing this? git describe --tags --long
09:34 adac fatal: No names found, cannot describe anything.
09:35 tobiasvl adac: did you give the command a commit(ish) to describe?
09:35 tobiasvl oh wait --tags
09:36 adac tobiasvl, there is simply no tag then?
09:36 tobiasvl for what commit?
09:36 adac git describe --tags --long always should give the latest tag with latest short commit number
09:37 adac but yeah there is obviously no tag
09:37 adac that will explain it :)
09:39 tobiasvl yeah!
09:41 rahtgaz joined #git
09:41 disse joined #git
09:42 anarchos2 joined #git
09:42 Raging_Hog joined #git
09:45 PaulePanter Hi. Is there a way to force deleting a branch I am on? The expected result would be to be in a disconnected state, like checking out a commit, that no branch points to.
09:47 MattMaker joined #git
09:47 lohfu joined #git
09:47 grawity well, just deleting a branch won't achieve that
09:47 grawity when you're on a branch, HEAD actually points to that branch
09:47 grawity if that branch is gone, HEAD doesn't point to a commit, it points nowhere
09:47 grawity but sure, that can be done manually
09:48 FilipNortic_ joined #git
09:48 grawity e.g. `git checkout HEAD^{}` to detach
09:48 grawity oh, there's --detach as well
09:48 tobiasvl oh nice, I was about to suggest "git checkout `git rev-parse HEAD`" or something :P
09:48 grawity you can also mess around with `git update-ref` and `git symbolic-ref`
09:50 fuchstronaut joined #git
09:55 PaulePanter Would proposing `git branch --detach -D branch-i-am-on` to delete the branch and change HEAD to point to the commit be useful?
09:55 PaulePanter My colleague thinks he needs that feature.
09:55 NwS joined #git
09:55 grawity your colleague can use the aliases feature
09:56 PaulePanter He says, that he then does `git checkout branch-i-am-on`, and git sets up the remote tracking branch, which `git reset --hard origin/branch-i-am-on` does not do.
09:56 PaulePanter grawity: I am on git 2.4.1.
09:56 PaulePanter `git branch` does not have a `--detach` switch there.
09:56 grawity to set up the remote-tracking branch, use `git branch -u origin/…`
09:56 grawity `git reset` is not even close
09:56 Wind0r joined #git
10:02 MattMaker joined #git
10:03 daanavitch joined #git
10:03 hahuang61 joined #git
10:03 JeroenT joined #git
10:03 ramsub07 joined #git
10:06 cyan__ joined #git
10:12 lss8 joined #git
10:14 PioneerAxon joined #git
10:14 konrados Morning :)
10:15 lss8 joined #git
10:15 konrados What this precisely mean (after doing git svn commit): "r16835 = 6da29ef91d35b8de5b96d82fcd3588e5d710635b (refs/remotes/git-svn) No changes between 142c043fc0a7ba439ef795611c434d635fa9b945 and refs/remotes/git-svn Resetting to the latest refs/remotes/git-svn"
10:16 konrados I know r16835 is a revision, but then, the next things?
10:16 canton7 sounds like it tried to dcommit 142c04, but it was identical to the previous svn commit (r16835), so it skipped over it
10:16 canton7 never seen it before, but that seems likely from the message
10:17 konrados But it's not true, I did make changes, and now they're on repo
10:17 konrados canton7, ^
10:17 canton7 "on repo"? does that mean they're in svn, they're in the local repo but dcommit is not committing them..?
10:17 canton7 what does !lol output?
10:17 gitinfo A nifty view of branches, tags, and other refs: git log --oneline --graph --decorate --all
10:19 konrados Last to lines, including comments in Polish: * 6da29ef (HEAD, git-svn, master) rt#brak naprawienie sytuacji z configiem * ac65e5c rt#brak Optymalizacja, z 48 zapytań do 7 w tym jedno dla listowania userów
10:20 canton7 pastebin, please
10:20 canton7 I want to see the relationships between things
10:22 MattMaker joined #git
10:22 adac joined #git
10:23 konrados canton7 : https://gist.github.com/konrados/f70707aea1674d532645813fb449e437
10:23 canton7 konrados, ok, so you don't have any commits which aren't in svn
10:23 canton7 where's the issue?
10:25 govg joined #git
10:27 konrados canton7 - so, everything's fine? The "issue" is, that I wanted to understand the message, what does it try to tell me. Like "Resetting to the latest refs/remotes/git-svn""
10:27 Ediz joined #git
10:27 ediz_ joined #git
10:30 xaviergmail joined #git
10:30 canton7 well, I gave you my best guess. You said that it's wrong, but haven't provided any evidence on why it's wrong, so I'm not sure how to continue :P
10:31 rubyonrailed joined #git
10:31 ibourgeois joined #git
10:32 ichux_ joined #git
10:32 konrados but after this: !log
10:32 gitinfo [!irclog] Public logs of #git are kept at: http://colabti.org/irclogger/irclogger_log/git or look at stats for the last 30 days: http://itvends.com/irc/git.html or since Jan 1, 2012: http://itvends.com/irc/git-all.html
10:32 dex1983 joined #git
10:32 dex1983 hi
10:32 sebboh joined #git
10:33 konrados I gave you a log with the first line stating: "6da29ef (HEAD, git-svn, master) rt#brak naprawienie sytuacji z configiem" - and this is my last commit, the one after I got this message, isn't this a prove, canton7?
10:33 canton7 so it looks like it did skip 142c043fc0
10:33 cdown joined #git
10:33 canton7 pastebin the output of 'git show 142c043fc0'
10:33 dex1983 I have renamed a folder name in the repo and commited the change now the file which was in there in this folder and repo was gone I had to recheck it into the repo now it is here again but without the change history why that?
10:36 konrados canton7, - https://gist.github.com/konrados/88540283c8cc0454aa65e0cd4ed5f3af
10:37 ronskisika joined #git
10:38 canton7 and what are the current contents of app/config/parameters.yml ?
10:39 konrados On both my local PC and repo - the same, i.e. what is on link I gave you (the diff after the "show" command), canton7
10:39 canton7 konrados, ok, so the commit 142c043fc really did do nothing
10:39 canton7 there was already a commit in SVN which did exactly the same thing, and added that parameters.yml
10:40 hiroki When working in a repo and wanting to use files from a different repository which is semi-frequently updated, what would be your go-to approach ? 1. Creating a folder within your repo folder that is used to clone the other repo in + using a .gitignore in your main repo to exclude that folder. Or 2. Cloning that folder outside of the main repo and therefore not needing a .gitignore. (Note. I'm not interested in
10:40 hiroki using submodules nor subtrees)
10:41 konrados OK, but I changed that file, i.e removed a few lines, canton7, I did it just before the commit. May I try again, will you have time for this? I will modify some file?
10:41 _ng joined #git
10:41 phanimahesh joined #git
10:41 canton7 konrados, and that commit was successfully dcommitted to svn, by the looks of it
10:41 _ng joined #git
10:42 MattMaker joined #git
10:42 ichux_ joined #git
10:43 stamina joined #git
10:43 tango_ joined #git
10:45 bongjovi joined #git
10:46 jtprog_ joined #git
10:47 Repox joined #git
10:47 cdown joined #git
10:50 JeroenT joined #git
10:52 konrados canton7, I did a test, I modified the same file, did commit and then git svn dcommit and got the almost exactly same message (with the numbers changed). A cooworker told me he doesn't really know "git svn" but his best guess is that after my commit there are no difference between remote repo and my local files and then.... he sets itself an information than there are no changes. What do you think about it?
10:53 canton7 so his best guess  is what I told you
10:53 canton7 well, what I told you my best guess was
10:53 canton7 !repro your test
10:53 gitinfo [!transcript] Please paste (using https://gist.github.com/ or https://vomitb.in/ ) a transcript ( https://gist.github.com/2415442 ) of your terminal session so we can see exactly what you see
10:56 JeroenT joined #git
10:57 konrados canton7, - https://gist.github.com/konrados/4a2ea8016facacc454946b526b6540a3
11:04 mingrammer joined #git
11:04 rnsanchez joined #git
11:05 cdown joined #git
11:06 Pulp joined #git
11:07 anttim joined #git
11:07 phanimahesh joined #git
11:08 elsevero joined #git
11:10 matumbo joined #git
11:11 fractalizator joined #git
11:11 bruce_lee joined #git
11:12 matumbo hello
11:12 gitinfo matumbo: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.
11:12 kuldeep_ joined #git
11:13 matumbo Anyone here using gpg to sign commits?
11:14 matumbo When I try to commit my work with the "-S" option to sign my commit, it lists my gpg key and says that it needs a passphrase to unlock the secret key. However, I don't get a prompt to enter my password.
11:15 tvw joined #git
11:16 matumbo If I just try to enter it anyways, it is written in cleartext on the screen and it doesn't sign the commit, it just hangs. If I use my gpg key in an arbitrary operation before commiting it works. I'm guessing it is because of the cache
11:17 MattMaker joined #git
11:17 rorro joined #git
11:18 HoierM joined #git
11:18 ojdo joined #git
11:18 SpeakerToMeat joined #git
11:21 a_thakur joined #git
11:23 Jellyg00se joined #git
11:30 Furai joined #git
11:31 Manuel15 joined #git
11:31 Manuel15 joined #git
11:31 j416 matumbo: I don't know the answer, but take a moment to consider the implications of signing commits, rather than tags
11:32 xaviergmail joined #git
11:32 Macaveli joined #git
11:34 matumbo j416: I haven't used tags before in Git so not sure how it compares
11:34 freimatz joined #git
11:35 encod3 joined #git
11:37 MattMaker joined #git
11:37 sdothum joined #git
11:37 matumbo my goal is to use a git repository to distribute a password database and I want to make sure only trusted signers can update the repo
11:37 sebboh joined #git
11:38 matumbo thought that if every commit is signed, they are indivudually approved
11:38 moritz matumbo: are you aware that dedicated secrets management software exist?
11:38 raijin joined #git
11:38 matumbo yeah
11:39 moritz like Vault or Keywhiz
11:39 matumbo moritz: for my purpose I feel that linux pass together with git is best suited
11:40 moritz matumbo: pass is nice, but it commits for you, so it might not be easy to get it to sign commits
11:40 j416 matumbo: signing a commit means you can never rebase it without re-signing, and it also makes patch-based workflows tricky
11:40 moritz also, shouldn't you handle access control at the ssh / git server level?
11:40 j416 matumbo: if anything, I would sign tagged releases, but avoid signing individual commits. But, ymmv, I don't know the nature of your project.
11:42 matumbo moritz: I will handle it with access control but in the event of someone breaching my computer, I'd like to prevent the submission of commits to the repo.
11:43 matumbo I thought that putting the global git config option commit.gpgsign might solve the problem with pass commiting
11:44 matumbo j416: I'm not gonna have any specific releases so tags seems out of place
11:45 fuchstronaut joined #git
11:46 j416 matumbo: in that case, still figure out why you need signed commits
11:47 j416 if someone has access to your computer you have bigger problems than people secretly committing new passwords to your repo..
11:48 matumbo j416: I know, but I'm trying to cover all angles
11:48 j416 commit signing does not prevent creation of commits.
11:48 j416 it only provides a means to verify that the signer is the creator of the commit
11:49 matumbo Couldn't I block commits that are not signed from being merged?
11:49 moritz matumbo: if somebody breaches your computer, they can most likely obtain a memory dump, including decrypted ssh and gpg keys
11:52 moritz first come up with attack scenarios that you want to defend against
11:52 moritz and then ask yourself if signing commits actually prevents that scenario, in case the attacker is intelligent
11:52 MattMaker joined #git
11:53 Spec-Chum joined #git
11:54 moritz or if, like, automatic mails to a mailing list for new commits would be easier. That way you'll notice if somebody commits in your name
11:54 moritz and if you disallow forced pushes, you can easily revert malicious commits
11:54 rahtgaz joined #git
11:54 matumbo are decrypted ssh and gpg keys stored in memory more that momentarily?
11:54 phanimah1sh joined #git
11:55 moritz if you use the corresponding agents (ssh agent, gpg agent), sure
11:55 moritz if not, you have to re-type your password for each commit
11:55 moritz and then an attacker can easily find it through a keylogger
11:55 moritz (unless you use hardware security tokens that contain your keys)
11:56 matumbo I think gpg-agent only stores the password in memory for a default time of 45 sec
11:56 matumbo and one of my next steps is going to be some hardware security token like yubikey perhaps
11:56 tomog999 joined #git
11:57 MattMaker joined #git
11:59 mingrammer joined #git
11:59 leeN joined #git
12:01 fgro joined #git
12:01 yaiyan_ left #git
12:01 Yaiyan joined #git
12:03 Sceorem joined #git
12:03 fgro hi. I have a git repo where the subfolder "html" contains the data that needs to go into my /www (webroot folder) on my server. How can I check this out? I know there is sparse-checkout, but then I end up with "/www/html" .. I do not have access rights to manipulate "/www" folder. so bascially I want to checkout files in local "html" subfolder of repository into remote machine "/www". Possible?
12:03 a3Dman joined #git
12:04 hahuang61 joined #git
12:05 moritz fgro: how about checking it out somewhere, and create a symlink from thatsomewhere/html to /www ?
12:06 moritz or create a proper build process, which creates a tarball or something from the directory, ships it to the remote server, and unpacks it there?
12:06 fgro moritz: that would be cool, but /www is owned by "root:root" which I don't have access to.
12:06 fgro moritz: you second recommentation seems to be the way to go. thx!
12:06 mingrammer joined #git
12:07 rnsanchez joined #git
12:07 moritz fgro: I'm writing a book on such topics, see https://leanpub.com/deploy :-)
12:08 fgro moritz: awesome! I will check it out.
12:08 jor fgro: needs to be a git checkout? can't use git archive and skip the toplevel dir when unpacking with tar?
12:08 moritz it goes the even bigger route, with building .deb packages
12:09 fgro jor: git archive and unpack with tar? not sure what you mean, can you give an example?
12:10 fgro moritz: deb packages ... something I never took time to get into. cetainly helpful.
12:12 MattMaker joined #git
12:13 fgro jor: thanks for the hint with git archive. Never used that before.
12:13 moritz fgro: I think jor means something along the lines of  git archive | tar x html/
12:13 moritz possibly with --strip-components 1
12:15 JeroenT joined #git
12:15 tvw joined #git
12:15 a_thakur joined #git
12:15 mingrammer joined #git
12:17 Shentino joined #git
12:18 matumbo I suppose signing commits with git doesn't support entering a password through ncurses interface used by gpg-agent
12:20 rnsanchez_ joined #git
12:21 johnmilton joined #git
12:21 moritz don't suppose; try it
12:22 robotroll joined #git
12:22 matumbo moritz: well I never get a prompt to enter my password which is the base of my assumption
12:22 johnmilton joined #git
12:23 matumbo If I have used gpg before so that the key is in the cache, I can sign commits but not when it wants the passphrase
12:24 jeffreylevesque joined #git
12:25 stuh84 joined #git
12:25 daniel_rose joined #git
12:26 ASOLAr joined #git
12:26 mikecmpbll joined #git
12:28 aiena when I am using a function from a classes parent inside the class do I need to use parent:: everywhere ?
12:28 t0by joined #git
12:29 _ikke_ aiena: this is #git :-)
12:30 DevAntoine hi
12:30 rkazak joined #git
12:30 DevAntoine is it possible to see all the commit of a specific branch?
12:30 osse git log branch
12:30 DevAntoine I want to rebase them all but don't know where to start
12:30 osse git log master..branch
12:31 DevAntoine hum, I guess it'll be develop in my case
12:31 DevAntoine trying, thanks
12:31 aiena _ikke_: ouch forgive me
12:31 sanketdg joined #git
12:32 aiena so sorry
12:32 MattMaker joined #git
12:32 sanketdg joined #git
12:33 rgrinberg joined #git
12:34 DevAntoine erf, there's too much merge and rebase in it :/
12:34 DevAntoine can I know the first commit of a branch?
12:36 osse not directly, only by inspection
12:36 osse it is probably the last one listed by log develop..branch
12:36 osse unless branch has been merged into development earlier, in which case do git log --graph --oneline --decorate develop branch and use your eyes
12:38 yottabyte joined #git
12:39 fuchstronaut .gitignore should not contain IDE specific folders like .vim_rc, but rather log files and cache folders, right?
12:40 canton7 fuchstronaut, depends. Some IDE specific folders/files are supposed to be per-user, and shouldn't be committed
12:40 moritz fuchstronaut: I have no problem with putting IDE-specific stuff into .gitignore
12:40 anuxivm joined #git
12:41 fuchstronaut Isn't this cluttering the gitignore file? I thought there is something like a 'local' .gitignore somewhere in /.git
12:41 yottabyte hey guys, if I clone a repo, so I'm in master, and I make a new branch, can I just switch branches whenever I want? i.e. if I want to go back to master? or should I clone a repo for master, another for anothe branch, etc? If I can, what happens to my changes made in the branch? do they get "stashed" or something? not sure what it all is
12:41 canton7 fuchstronaut, again, depends. If everyone on your team uses the same IDE, you should all have the same ignores in your .gitignore which relate to the IDE
12:42 moritz fuchstronaut: there is .git/config/exclude (iirc)
12:42 canton7 fuchstronaut, if you're an oddball on the team using your own special editor, then yeah, .git/info/exclude is a good place
12:42 moritz fuchstronaut: but in practise, cluttering .gitignore doesn't seem to be much of a problem
12:42 canton7 you don't want the one idiot on the team committing per-user IDE settings, and everyone else behaving
12:42 fuchstronaut canton7: we are all using different IDEs so I thought .git/config/exclude might be the right places (thanks moritz)
12:42 osse DevAntoine: https://gist.github.com/Osse/f104ea3b8bb015b72b24aca1db6928fa
12:43 canton7 fuchstronaut, in that case, yeah, maybe .git/info/exclude is better
12:43 canton7 putting random stuff into .gitignore ("cluttering") it doesn't really have a cost though
12:43 osse DevAntoine: hmm never mind. it's not easily adaptible
12:43 osse adaptable?
12:43 moritz changable :-)
12:44 matumbo moritz: Found the solution to my signing problem. Apparently I need to export GPG_TTY every time I start a new TTY. export GPG_TTY=$(tty)
12:44 moritz matumbo: wow, that's rather obscure
12:44 ArchNoob joined #git
12:46 matumbo moritz: yeah, was lycky I stumbled across the same problem on superuser
12:46 matumbo lucky*
12:46 Luca__ joined #git
12:47 anonymuse joined #git
12:47 MattMaker joined #git
12:47 sebboh joined #git
12:48 Luca__ hello everybody, I'm new to git and trying to undestand how to use it
12:49 tobiasvl Luca__: !book
12:49 gitinfo Luca__: There are several good books available about git; 'Pro Git' is probably the best: http://git-scm.com/book but also look at !bottomup !cs !gcs !designers !gitt !vcbe and !parable
12:49 grawity Git specifically, or version control in general?
12:49 dendazen joined #git
12:49 Luca__ I use to program my code by myself, we have a wamp server ad what I want is to change code on my development machine and then deploy it on the server
12:49 tobiasvl !deploy
12:49 gitinfo Git is not a deployment tool, but you can build one around it (in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/deploy.html
12:50 _ikke_ Luca__: One advise that we give to all users here, don't use git push to deploy code
12:50 Luca__ I'm totally new on version control, I've installed git and it seems to work fine
12:50 moritz Luca__: I'm writing a book about deployments, with git as the version control system: https://leanpub.com/deploy
12:50 * moritz is in spammy mode today
12:51 GodGinrai joined #git
12:51 Luca__ great, thank you a lot to you all
12:51 Luca__ I'm going to download and take a look to all those books
12:52 Luca__ At the moment I've performed some try and I have some question
12:52 tlaxkit joined #git
12:52 alansaul joined #git
12:53 Luca__ I have committed changes on my pc and committed them to the server but I cannot see any change in the file committed on the remote server
12:53 Luca__ is it normal?
12:53 tobiasvl commited to the server? you mean pushed?
12:53 Luca__ sorry, committed locally and then pushed to the server
12:54 Luca__ anyway, I just changed some image but I can't see the new images online on the server
12:54 tobiasvl I assume the repo on the server is not !bare ?
12:54 gitinfo A bare repository is used to push/fetch(useful for running a git server), and contains only the contents of .git/ from a "normal" repo. Read more: http://bare-vs-nonbare.gitrecipes.de/
12:54 zivester joined #git
12:54 dsdeiz joined #git
12:54 encod3_ joined #git
12:54 Luca__ git says that it's all right and I cannot understand what I'm doing wrong
12:55 canton7 Luca__, read the !deploy link you were given
12:55 gitinfo Luca__: Git is not a deployment tool, but you can build one around it (in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/deploy.html
12:56 DevAntoine when squashing, it's always the commit on top which gets picked and the ones under which are squashed? It can't work reversly (squashed on top and the one at the bottom getting picked)?
12:56 Luca__ ok, I'm going to do it
12:57 JamesBaxter joined #git
12:57 bremner DevAntoine: reorder the commits then squash?
12:57 MattMaker joined #git
12:57 kadoban joined #git
12:58 bremner DevAntoine: or just change the commit message, probably simpler
12:58 DevAntoine it was just for my knowledge, I don't see any issue working with the first one getting picked and the other ones getting squashed
12:58 jost_ joined #git
12:59 bremner rebasing applies commits in order, top to bottom of the todo list
12:59 jimi_ joined #git
12:59 mohabaks joined #git
13:00 ExoUNX joined #git
13:01 DevAntoine erf, considering there's a rebase in the middle of my commits, I can't squash everything right :/
13:02 DevAntoine I want to be able to merge my working branch with a single commit and not all the "Work in Progress" in it
13:02 yottabyte hey guys, if I clone a repo, so I'm in master, and I make a new branch, can I just switch branches whenever I want? i.e. if I want to go back to master? or should I clone a repo for master, another for anothe branch, etc? If I can, what happens to my changes made in the branch? do they get "stashed" or something? not sure what it all is
13:04 kadoban yottabyte: You can create new branches and switch branches as you like, yes.
13:04 bremner DevAntoine: maybe you want git merge --squash
13:04 kadoban !float might answer some of your questions with what happens to changes. Changes you commit to a branch stay on that branch.
13:04 gitinfo If you have made a change in your working directory and have NOT YET COMMITTED, you may "float" that change over to another (`git checkout oldbranch`) or new (`git checkout -b newbranch`) branch and commit it there.  If the files you changed differ between branches, the checkout will fail.  In that case, `git stash` then checkout, and `git stash apply` and go through normal conflict resolution.
13:04 yottabyte kadoban: what about changes I've made in one branch though?
13:04 DevAntoine bremner: hum, I don't know this one
13:04 yottabyte ah I see
13:05 hahuang61 joined #git
13:05 f0rblaze joined #git
13:05 kadoban yottabyte: Some practice will make it much clearer of course. Also reading pro git (the book on the website) would be a good idea eventually.
13:05 DevAntoine wow bremner, looks like exactly what I need
13:07 DolphinDream git scripting: how can i tell if a pull / merge command has failed due to conflicts ?
13:07 DevAntoine bremner: not exactly actually, because I'll directly merge my working branch
13:07 DevAntoine I want to do a PR before that
13:07 DevAntoine to make that PR I need a single commit
13:09 Macaveli joined #git
13:10 Moggets joined #git
13:10 Darcidride joined #git
13:11 yottabyte kadoban: I feel like I've tried git checkout oldbranch and files did differ, but the checkout didn't fail
13:11 yottabyte or did that gitinfo only mean if there's conflicts, then it'll fail?
13:11 Mogget joined #git
13:11 kadoban Yeah, reread what it says, it's pretty detailed and exact
13:12 mingrammer joined #git
13:13 sanketdg joined #git
13:13 yottabyte I don't know what "If the files you changed differ between branches, the checkout will fail." means
13:13 yottabyte won't they likely differ between branches
13:13 yottabyte or does differ mean conflict here
13:14 kadoban yottabyte: It means if there were changes in both of them
13:14 kadoban Like you changed something and the file is different between the branches as well.
13:14 osse in layman's terms: git diff branch1 branch -- file  prints somethign
13:17 deepy how terrible is git replace to work with? I have an old commit that's missing one parent
13:17 MattMaker joined #git
13:21 ASOLAr joined #git
13:21 cdg joined #git
13:21 tvw joined #git
13:22 Salamander joined #git
13:22 cdg joined #git
13:22 Salamander Hello, I migrated a GitLab project from gitlab.com to a private remote server, I added myself as a new user and I added an ssh key. How do I now update my cloned repo to pull/push to the new remote server?
13:22 duderonomy joined #git
13:23 deepy Salamander: tell your cloned repo to point the remote to the new remote
13:23 kadoban git remote set-url and etc.
13:23 Salamander thank you
13:24 fees joined #git
13:24 pro_metedor joined #git
13:26 Luca__ Thank you everybody, now I have a clearest vision of my problem... a quite common problem for a newbye, if I can say it ;-)
13:26 matumbo Seems like "git config --global commit.gpgsign true" doesn't have any effect for me
13:26 leehambley joined #git
13:26 phanimah1sh did you also set gpgkey?
13:27 Luca__ so I should use bare repo instead of non-bear ones
13:28 Luca__ and I should adopt the first solution mentioned in http://gitolite.com/deploy.html
13:29 rubie joined #git
13:29 Salamander left #git
13:29 Salamander joined #git
13:29 fahadash joined #git
13:29 Salamander left #git
13:30 Luca__ I would also like to implement the post-push routine mentioned in http://bare-vs-nonbare.gitrecipes.de/ but the link leads to a 404 page. It would be great if someone could give the code :)
13:31 DieguezZ joined #git
13:32 MattMaker joined #git
13:33 al-damiri joined #git
13:33 DieguezZ hi, i just rebased accidentally a branch with dev branch, and now i realised i want to do the pull request to another branch (uat). The problem is that i did push -f to my branch after rebasing, and now, when i try to rebase it to uat, its including dev's commits
13:34 Ozd joined #git
13:34 Ozd startups
13:34 i7c frogs
13:34 Ozd lol
13:35 legumbre joined #git
13:35 Ozd misclick
13:35 MadHatter42 joined #git
13:35 i7c ;)
13:37 tsunamie joined #git
13:38 Gamecubic joined #git
13:38 yottabyte left #git
13:39 Dougie187 joined #git
13:40 rnsanchez joined #git
13:41 Gsham joined #git
13:41 Eryn_1983_FL joined #git
13:42 ichux_ joined #git
13:42 matumbo phanimah1sh: yeah, with signingKey
13:44 fmcgeough joined #git
13:47 Spec-Chum joined #git
13:47 MattMaker joined #git
13:48 sebboh joined #git
13:50 Akimb joined #git
13:55 jaafar joined #git
13:57 MDA1 joined #git
13:58 dvl joined #git
14:00 dvl If I have done a git clone dvl@git.example.com:/var/lib/git/the-tools.git --bare .. then brought it into a new git repo via 'git push TestImport7 --mirror'  ... Have I brought in the hooks associated with the original repo?  I think so, yes, based on there being hooks in first repo.
14:00 JeroenT joined #git
14:01 kadoban dvl: clone doesn't bring in hooks, that'd be *horrible*
14:01 kpease joined #git
14:01 thiago joined #git
14:01 kadoban That would mean any repo you clones could just straight-up run arbitrary code on your machine as soon as you do anything with it.
14:02 MattMaker joined #git
14:03 dvl kadoban: OK, now i need to figure out how to being over the commits.
14:03 dvl ... hooks.
14:04 HappyLoa- joined #git
14:04 ash_workz joined #git
14:04 kadoban dvl: They're just files, you can copy them over if you like. They're in .git/hooks/
14:05 _ikke_ kadoban: even the clone would be able to execute code with the post-checkout hook
14:05 le_melomane joined #git
14:06 dvl kadoban: .git/hooks?  I was looking at the-tools.git/hooks   Are we talking about the same place?
14:06 dvl ^ in my particular case
14:06 _ikke_ dvl: bare vs non-bare repo
14:07 dvl _ikke_: Ahh, so the -bare in my clone excludes the hooks by design?
14:07 _ikke_ no
14:07 _ikke_ Just the differing name
14:07 _ikke_ .git/hooks vs the-tools.git/hooks
14:07 dvl _ikke_: Thank you.
14:07 snowkidind joined #git
14:08 grawity $(git rev-parse --git-dir)/hooks
14:08 PioneerAxon joined #git
14:08 _ikke_ hooks are never cloned, like kadoban said
14:08 AaronMT joined #git
14:08 mohabaks joined #git
14:09 rubie joined #git
14:09 kadoban _ikke_: Ah, good point.
14:10 OnOut2k16 left #git
14:11 EduardoMartins joined #git
14:12 fakenerd joined #git
14:12 duderonomy joined #git
14:14 phaleth joined #git
14:14 ResidentBiscuit joined #git
14:14 dsdeiz joined #git
14:15 forgot is --porcelain like --human-readable or the opposite?
14:15 osse opposite
14:16 xaviergmail joined #git
14:17 Neon joined #git
14:17 forgot im confused
14:17 osse me too
14:17 MattMaker joined #git
14:17 forgot now im glad and confused
14:18 osse the term "porcelain" is used to describe the high-level human-friendly commands, such as log and diff. But --porcelain can also mean that the command will print stable output meant to be parsed by scripts
14:18 Gsham joined #git
14:18 kadoban That's ... confusing.
14:18 Neon Let's say I manage to call git branch (from another process) while a git process is currently doing stuff in the same repo, e.g. git checkout foo. Will git branch wait until the other process has finished, before it gives me output or will the output be undefined?
14:19 forgot shouldn't the low-level command be more stable?
14:19 Neon (all on one machine, not remote)
14:20 _ikke_ --porcelain is output meant to build porcelain :-)
14:20 Neon Or more general: Is git protected against multi-process/-thread fuckups?
14:20 vassagus joined #git
14:20 ichux_ joined #git
14:20 grawity most of it isn't multi-threaded in the first place
14:21 grawity the data structures should be somewhat resilient – objects/ is immutable, index has a .lock file, etc.
14:23 Neon Mh, in my special case I have a file change listener on .git/HEAD in one process and another shell where I would switch the branch, e.g. git checkout foo. The listener would then do some stuff as soon as git updates the HEAD with the new branch ref and I need it to do it after git finished with its work. Maybe it'd suffice if it was safe to say that git updates HEAD at the very last.
14:23 zeroed joined #git
14:23 zeroed joined #git
14:23 spudowiar joined #git
14:24 grawity no, I'm pretty sure .git/HEAD would be updated as the first thing, and worktree update left last
14:25 Neon In my listener I call git branch so if that would wait until the other process finished it would fix itself. Just so you understand my first question.
14:25 Neon Okay, then it's good that I thought about that before getting weird results later on.
14:25 grawity you could, for example, monitor .git/logs/HEAD (the reflog) or turn your listener into a post-checkout hook
14:25 forgot So --porcelain really means ‘I am a porcelain’, i.e. ‘Please make your output suitable for porcelains’ (as opposed to people).
14:25 Neon Any ideas for me? :)
14:25 mohabaks joined #git
14:26 grawity forgot: no, usually means the latter
14:26 grawity or, hmm
14:27 grawity I guess .git/HEAD *is* updated at the end
14:27 Neon grawity: Oh, hooks were what I looked at at first. I oversaw the post-checkout hook, though. There is no example file for that.
14:27 infra-red joined #git
14:27 Neon Just found a site that said "look at the examples in the hooks dir" and then I was disappointed, because there was no checkout hook.
14:27 grawity man githooks
14:27 gitinfo the githooks manpage is available at http://jk.gs/githooks.html
14:27 duderonomy joined #git
14:28 morenoh149 joined #git
14:28 forgot grawity: could you illustrate?
14:28 Neon I will use that hook of course. Is the best solution I think. Thanks for helping me, grawity. (That's a rhyme :D)
14:28 grawity ah, I see now – it opens .git/HEAD.lock, writes the new commit there, updates the reflog, then finally renames HEAD.lock to HEAD
14:29 dvl And to complicate things, we're using a Github Enterprise box.. seems hooks are now webhooks, or did I just go down the wrong rabbit hole?
14:29 Neon Probably even more complex then. The FS will count that as delete, then create, so I'd get spammed with updates.
14:29 grawity dvl: that only applies to the server side
14:30 grawity hooks on the client remain the same
14:30 guampa joined #git
14:30 dvl grawity: The hooks are are using are all server side.  I'm moving the repo from a REAL server into a github enterprise appliance.
14:30 grawity ah
14:30 dvl ... the hooks we are....
14:31 grawity what kind of stuff do those hooks do?
14:31 _sjs joined #git
14:31 grawity forgot: `git status --porcelain`, `git diff --word-diff=porcelain`
14:31 dvl grawity: I'm told, but can't confirm because I have not found the hooks in questions yet, they push changes to our public github repo.
14:32 timothy joined #git
14:32 vassagus_ joined #git
14:32 dvl grawity: I found one which does a 'exec /usr/local/bin/git-notify.pl' to email a mailing list.
14:32 timothy hi, is there any way to do not use SHA-256? since sha-1 is also used in signed tags
14:32 le_melomane joined #git
14:32 MattMaker joined #git
14:33 kadoban timothy: What use of SHA-256 are you trying to avoid?
14:33 timothy git uses SHA-1
14:33 grawity you asked about how to /not/ use SHA-256
14:33 grawity good news, Git does not use SHA-256
14:33 timothy oh sorry
14:33 forgot grawity: git-status outputs human-facing words like (use "git add <file>..." to include in what will be committed), --porcelain does not
14:33 timothy I mean
14:33 grawity all its object IDs are SHA-1 currently
14:33 timothy s/way/reason/
14:34 grawity changing that first requires fixing all of Git code to stop assuming object IDs are 20-byte
14:34 osse it's easy.  make SHA256=no :)
14:34 _ikke_ timothy: The discussion is going on to move to SHA256, but it's a long process
14:34 grawity which is being done
14:34 kadoban timothy: Histerical raisins.  And if we're going to upgrade the crypto, probably might as well use something even more modern.
14:34 _ikke_ timothy: with lots of problems to solve
14:34 grawity and somehow dealing with people trying to merge a new-style sha256 repo into an old-style sha1 one, I guess
14:34 kaldoran joined #git
14:35 _ikke_ yes, and dealing with existing signatures
14:35 grawity it'll be changed eventually, I imagine
14:35 grawity but it's *a lot* more work than just changing the signature algo in TLS or PGP
14:36 timothy like sha3? :P
14:37 * grawity doesn't really understand 'newer = better' in crypto
14:37 daniel_rose joined #git
14:37 kadoban Ya, I'd be curious why they're not skipping to sha3 myself.
14:37 grawity especially not when the people who approved sha3 have specifically recommended otherwise
14:38 kadoban NIST says not to use sha3?
14:38 grawity it says sha2 is just as good
14:38 osse as a checksum sha1 is more than good enouh
14:39 kadoban Ah, yeah I guess they did.
14:39 grawity osse: well yes, but Git has been using it as a bit more than just a checksum
14:39 grawity with all the signed tags, signed commits, etc.
14:40 FilipNortic__ joined #git
14:40 osse does it make signing less useful?
14:41 anarchos2 joined #git
14:41 vassagus joined #git
14:41 sebboh joined #git
14:41 osse (honest question; I am clueless)
14:42 grawity only as "secure" as the weakest link
14:42 forgot less secure for sure
14:42 axl_ joined #git
14:42 MattMaker joined #git
14:42 pasturesofplenty joined #git
14:43 grawity since you only sign the commit hash or tree hash, even if the new fancy sha999 signature itself cannot be broken, a particular sha1-identified blob or tree could still be.
14:44 pasturesofplenty this is probably not the right channel but maybe someone can help... it seems like tortoisegit doesnt support using ed25519 keys for a remote ssh git repo? thoughts?
14:44 grawity I think that depends on what version of openssh it ships
14:45 ramsub07 joined #git
14:45 grawity (or does it use libssh?)
14:45 stonerfish joined #git
14:45 pasturesofplenty it uses putty
14:46 pasturesofplenty oh.... wait a minute... i know why... >.<
14:46 grawity ah, well, then give it a more recent putty version
14:46 grawity the releases might not support ed25519 yet, but the "nightly" builds definitely do
14:46 pasturesofplenty yes... thats what i just figured... it has its own copies of putty that are older and dont support ed25519 yet
14:46 grawity (it just needs plink.exe, I guess)
14:47 osse i think you can tell it to use openssh
14:47 osse somewhere in the settings.
14:47 grawity but then you need to obtain a new enough openssh
14:47 pasturesofplenty hmm, im looking in the bin directory for tortoise... it hast tortoisegitplink.exe? is that drastically different from putty's vanilla plink?
14:47 grawity prrrrobably not.
14:48 chachasmooth joined #git
14:48 forgot i think to use new fancy sha999, the whole repo should be rebuilt
14:49 rorro joined #git
14:49 osse the most forceful of force pushes
14:49 forgot like a bitcoin hardfork
14:52 osse Eugene: is it a coincidence that Eugene kay rhymes with Eugene Ray ?
14:54 ochorocho__ joined #git
14:54 DevAntoine I'm trying to understand what the fuck happened: I've merged develop in my working branch. It listed me 5 conflicts which I've solved. After committing the merge I've realized that the merge had literally erased all the work I've done on the working branch by old code from develop
14:54 DevAntoine I can't explain that
14:54 * DevAntoine has already lost 4 hours using PHPStorm's local history to retrieve all the changes
14:54 ASOLAr joined #git
14:54 pasturesofplenty cool, i just copied plink and such from my putty directory and that seemed to basically work
14:55 pasturesofplenty though it looks like maybe it complained at the way plink exited (waitpid)
14:55 kadoban DevAntoine: What commands did you run? Did you resolve the conflicts correctly?
14:55 pasturesofplenty but it did push
14:55 sarbs joined #git
14:55 pasturesofplenty thanks
14:55 DevAntoine kadoban: git merge develop
14:55 DevAntoine kadoban: I did fix the conflicts and then added the conflicted files to the staged changes
14:55 DevAntoine the commit the ùerge
14:56 rpd DevAntoine: you didn't commit your changes before merging?
14:56 dvl I found .git/hooks/pre-rebase but it's not chmod +x ... my reading indicates such hooks are inactive by design.  Am I correct?
14:56 bremner yes
14:57 DevAntoine rpd: everything was pushed
14:57 kadoban DevAntoine: Can you reproduce the behavior, with a !transcript ?
14:57 gitinfo DevAntoine: Please paste (using https://gist.github.com/ or https://vomitb.in/ ) a transcript ( https://gist.github.com/2415442 ) of your terminal session so we can see exactly what you see
14:57 kadoban That doesn't make a lot of sense, and when things don't make a lot of sense, it's rarely the computer that's confused.
14:58 SylvieLorxu joined #git
15:00 ochorocho__ joined #git
15:00 alansaul joined #git
15:00 Rish joined #git
15:01 DevAntoine yeah, I'm pretty sure I did shit somewhere...
15:01 zivester joined #git
15:01 dvl bremner: Merci.
15:02 griffindy joined #git
15:02 LucaTM joined #git
15:03 gynter joined #git
15:04 lacrymology joined #git
15:04 SylvieLorxu Hey everyone, when I do something like "git clone https://github.com/nonexistant/nonexistant", git will ask for my username and password. How can I just make it fail instead of waiting for user input in this case? I'm currently cheaply using it in the background of one of my apps which makes everything lock up because there's no way the user can access it. Yes, I know I should refactor a bit, but one thing at a
15:04 SylvieLorxu time
15:04 SylvieLorxu s/can access it/can enter this info/
15:05 grawity try specifying a dummy username:password in the URL
15:05 daniel_rose joined #git
15:05 grawity afaik, GitHub always reports 401 instead of 404 for "privacy" reasons
15:05 gynter Hello, I'm using git 2.7.4. Currently when using mergetool (vimdiff) I see only 3 windows (LOCAL, MERGED, REMOTE) how can I make git to use vimdiff as 3way merge with 4 windows (LOCAL, BASE, REMOTE, MERGE)?
15:05 hahuang61 joined #git
15:06 gynter I tried changing merge.tool to vimdiff -f -c wincmd J -d $MERGED -M $LOCAL $BASE $REMOTE, but result is thte same
15:06 shinnya joined #git
15:06 Rish joined #git
15:07 osse gynter: are you using Fugitive?
15:07 gynter no
15:07 mingrammer joined #git
15:09 roelmonnens joined #git
15:09 ochorocho__ joined #git
15:09 SylvieLorxu grawity: Unfortunately the URL is user input, so I'd prefer to avoid parsing the URL and then hacking around in it. Isn't there a way I can make git think it's running on a dumb non-interactive terminal or so perhaps? Assuming that could work too. I could probably set an environment variable
15:10 kadoban There's some way to tell git how to get credentials from the user I think, you could just use /bin/false I guess.
15:12 dreamyspell joined #git
15:12 foist joined #git
15:12 jh3 joined #git
15:13 rubyonrailed joined #git
15:13 dimi1947 joined #git
15:13 mjuszczak joined #git
15:14 mohabaks joined #git
15:14 Rish_ joined #git
15:14 DolphinDream git scripting: is there a way to tell if a pull / merge command has failed due to conflicts ?
15:15 charginghawk joined #git
15:16 charginghawk if I accidentally have some large files committed to a branch, and I delete the branch, do those files get removed from the repo?
15:16 dmto joined #git
15:16 osse charginghawk: is the branch merged? No . Is the branch NOT merged? Yes, eventually.
15:17 _ikke_ If that branch was the only one containing the files, then yes
15:17 charginghawk cool
15:17 kadoban charginghawk: Not immediately anyway. There's the reflog if nothing else that keeps them around.
15:17 charginghawk is there a way to check branch sizes?
15:17 MattMaker joined #git
15:18 diphtherial joined #git
15:18 _ikke_ There is no built-in way, but you can calculate it
15:18 diphtherial this is a minor question, but is there a way to check out a repo 'in place' in a directory that's an exact copy of what's already in the repo?
15:18 _ikke_ diphtherial: git checkout <commit>?
15:18 charginghawk how can I calculate it?
15:19 grawity diphtherial: !deploy has some examples
15:19 gitinfo diphtherial: Git is not a deployment tool, but you can build one around it (in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/deploy.html
15:19 diphtherial _ikke_: oh, so you can just check out over existing files?
15:19 SylvieLorxu kadoban: Thanks for the suggestion, I've tried "$ GIT_ASKPASS=/bin/false git clone https://github.com/nonexistant/nonexistant" but unfortunately after git prints "error: unable to read askpass response from '/bin/false'" it goes back to asking interactively so that unfortunately doesn't work :(
15:19 _ikke_ diphtherial: git checkout will update all files to match the commit
15:19 diphtherial grawity: thanks, i'll read that
15:19 _ikke_ diphtherial: Do you mean in an external directory not related to the repository?
15:20 SylvieLorxu Oh hey, using /bin/echo works
15:20 diphtherial _ikke_: well, it's a copy of the files in the repo, but yes, it's not currently a repo
15:20 diphtherial but yes, you answered my question :)
15:20 kadoban SylvieLorxu: That's unfortunate. Maybe there's some protocol it needs to speak, or some other way to override it more strongly.
15:20 kadoban Hah, really? Well ... that's funny.
15:20 SylvieLorxu I wonder what it actually tries to use
15:21 SylvieLorxu Probably defaults to username git, passes that to echo, gits git as password back and thus tries git:git? Hmm
15:21 SylvieLorxu s/gits/gets/
15:21 SylvieLorxu Still, this is a bit ugly lol
15:21 SylvieLorxu And with a bit I mean Oh God this is horrible, haha. But hey, if it works for now...
15:21 kadoban What does /bin/true do?  Or maybe just use cat and have it output some nonsense. Or whatever, I dunno.
15:22 canton7 /bin/true has the best manpage ever
15:22 canton7 "true - do nothing, successfully "
15:23 kadoban Is false  - fail to do nothing?
15:23 canton7 similar to /bin/false: "false - do nothing, unsuccessfully "
15:23 _ikke_ false - do nothing, unsuccessfully
15:23 _ikke_ Too late
15:23 kadoban Hah, nice.
15:23 SylvieLorxu kadoban: /bin/true works too
15:23 kadoban SylvieLorxu: So it must just not like the command failing.
15:23 _ikke_ So /bin/false managed to do something?
15:24 SylvieLorxu kadoban: Seems so, that makes sense, although it makes no sense that git is like "Hey I got no output back I'll use that as password" lol
15:24 ochorocho__ joined #git
15:24 SylvieLorxu Oh well, this hack is good enough for the time being I suppose :P
15:24 SylvieLorxu Thanks
15:24 kadoban Well, it kinda makes sense. A blank password *is* possible.
15:24 kadoban It'd be stupid as hell, but it'd work.
15:25 SylvieLorxu I wonder what it grabs as username though
15:25 canton7 also, https://what.thedailywtf.com/topic/1098/solaris-bin-true, http://git.savannah.gnu.org/cgit/coreutils.git/tree/src/true.c, and http://git.savannah.gnu.org/cgit/coreutils.git/tree/src/false.c
15:25 SylvieLorxu Because I don't have that set globally
15:26 SylvieLorxu Oh God that's a horrible part of Coreutils, heh
15:26 _ng joined #git
15:26 canton7 ... from humble beginnings: http://git.savannah.gnu.org/cgit/coreutils.git/commit/src/true.c?id=5d4eefe8d858f8d832d0038175dd1b461e7004c5
15:27 davisonio joined #git
15:27 _ng joined #git
15:28 boombatower joined #git
15:36 happy-dude joined #git
15:36 miklcct joined #git
15:36 mohabaks joined #git
15:37 MattMaker joined #git
15:39 subhojit777 joined #git
15:39 KevinMGranger compare to an older version of openbsd's true: https://raw.githubusercontent.com/chneukirchen/obase/master/src/usr.bin/true/true.sh
15:39 KevinMGranger And their current version: http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin/true/true.c?rev=1.1&amp;content-type=text/x-cvsweb-markup&amp;only_with_tag=MAIN
15:40 sebboh joined #git
15:40 canton7 but... how will you know what version it is, or how to use it correctly? ><
15:41 hwrdprkns joined #git
15:42 abyx joined #git
15:44 ochorocho__ joined #git
15:45 crayon joined #git
15:45 morenoh149 joined #git
15:47 cdown joined #git
15:47 MattMaker joined #git
15:49 salparadise joined #git
15:49 ochorocho__ joined #git
15:49 EvilPeng1 joined #git
15:50 salparadise hi all, i made a commit on the wrong branch, now I want to go to the correct branch and just grab that one change, im thinking this is easy, but just need the right thing to google
15:50 daniel_rose joined #git
15:50 salparadise is it just a merge of the one commit hash from that branch?
15:50 XiaoMianTiao joined #git
15:50 osse salparadise: cherry-pick
15:50 Kings joined #git
15:50 barajasfab joined #git
15:50 salparadise osse: cool, looking
15:50 salparadise thanks
15:52 HoierM joined #git
15:53 Curly060 joined #git
15:54 fuchstronaut joined #git
15:54 ochorocho__ joined #git
15:54 durham joined #git
15:56 xelxebar joined #git
15:56 dvl Anyone used Github Enterprise appliances?  It seems that hooks are now non-trivial.  They use webhooks. It's not just a copy/paste solution if I'm reading this correctly.
15:56 dvl ^ context: we're moving from a real machine to a GHE appliance.
15:56 salparadise related question, how do I diff just one of the commits in one branch vs. another, I don't want to diff what was done before
15:57 MattMaker joined #git
15:58 _ikke_ diff will diff exactly two commits with eachother
15:59 _ikke_ But all the contents of those commits
16:01 alansaul joined #git
16:02 salparadise _ikke_: yeah nm, actually this cherry-pick looks like worked great, thanks osse, _ikke_
16:02 mingrammer joined #git
16:04 Gsham joined #git
16:05 anuxivm joined #git
16:06 anuxivm joined #git
16:07 MattMaker joined #git
16:10 joki joined #git
16:11 phanimahesh joined #git
16:11 chrisshattuck joined #git
16:12 skylite joined #git
16:12 gynter Hello, I'm using git 2.7.4. Currently when using mergetool (vimdiff) I see only 3 windows (LOCAL, MERGED, REMOTE) how can I make git to use vimdiff as 3way merge with 4 windows (LOCAL, BASE, REMOTE, MERGE)? I tired vimdiff -f -c 'wincmd J' -d '$MERGED' -M '$LOCAL' '$BASE' '$REMOTE' but with no luck.
16:15 blasagna joined #git
16:17 hobodave joined #git
16:17 MattMaker joined #git
16:18 rubie joined #git
16:18 rominronin joined #git
16:19 dreiss joined #git
16:20 navidr joined #git
16:20 osse gynter: what does 'git config merge.tool' say?
16:21 skylite joined #git
16:21 alansaul joined #git
16:22 Guest85 joined #git
16:23 gynter hmm, I had a mistake, i had to set vimdiff cmd to vimdiff ...
16:23 gynter and merge.tool to vimdiff
16:23 gynter but still can't get it working correctly, probably have mistakes in vimdiff arguments
16:23 Endarked joined #git
16:25 diogenese joined #git
16:25 gynter currently I tried cmd = vim -d -c "wincmd l" "$LOCAL" "$MERGED" "$REMOTE" but with no luck, I'm getting error from vim: E471: Argument required: wincmd
16:25 grawity cmd = "vim -d -c \"wincmd l\" \"$LOCAL\" etc."
16:26 osse grawity: just 'vimdiff' is enough
16:26 osse ehh, gynter
16:26 gynter grawity: yea, just figured that out also, thanks :D
16:27 gynter i used single quotes instead of double for -c 'wincmd l'
16:27 McNozes joined #git
16:28 Vgr joined #git
16:28 Vgr joined #git
16:28 Vgr joined #git
16:28 gynter cmd = "vim -f -d -c \"wincmd J\" \"$MERGED\" -M \"$LOCAL\" \"$MERGED\" \"$REMOTE\"" did the trick, thanks!
16:28 Vgr joined #git
16:28 gynter s/MERGED/BASE/
16:29 osse :S
16:29 hwrdprkns joined #git
16:29 abyx joined #git
16:29 kpease joined #git
16:29 gynter osse: the issue what I had was that when using default arguments (only vimdiff) then I got 3 windows (LOCAL, MERGE, REMOTE), and no BASE
16:30 madewokherd joined #git
16:30 rorro joined #git
16:30 mjuszczak joined #git
16:30 osse that sounds like a booj in your vim config
16:30 jstein_ joined #git
16:30 Vgr joined #git
16:30 Vgr joined #git
16:31 gynter same ~/.vimrc on two different machines, works on one but not another
16:31 * osse calls Mulder & Scully
16:31 gynter jep
16:31 Vgr joined #git
16:31 hobodave_ joined #git
16:31 thiago joined #git
16:32 blackwind_123 joined #git
16:32 Vgr joined #git
16:32 Vgr joined #git
16:32 Vgr joined #git
16:32 LeBlaaanc joined #git
16:32 toralf joined #git
16:33 Vgr joined #git
16:33 toralf at a fresh cloned git tee the following gives an empty output : git diff --name-only "@{ 2 hour ago }".."@{ 1 hour ago }"
16:33 Vgr joined #git
16:33 toralf at another system, where I cloned teh same tree a while ago, O do see however all changed files
16:33 grawity @{date} searches based on reflog, not the actual commit metadata
16:33 toralf now I do wonder how to convince git to give me teh wsame outout at the freshly cloned git tree
16:34 osse toralf: @{..} uses the reflog. brand new clo.. yes
16:34 toralf grawity: aha
16:34 toralf osse: grawity: what do I need instead
16:34 toralf ?
16:34 grawity try `git log/rev-parse --since= --until=` to figure out which commits you want
16:35 grawity or --min-page= --max-age=
16:35 grawity then diff the first and last results
16:35 freimatz joined #git
16:37 MattMaker joined #git
16:38 sebboh joined #git
16:40 garethdaine joined #git
16:40 zeroed joined #git
16:40 zeroed joined #git
16:43 AnarchyAo joined #git
16:43 SporkWitch does git support markdown in commit comments? does it make sense to use it if the patch is related to a CVE and link to it, or just list the CVE number?
16:43 _ikke_ SporkWitch: Git doesn't matter what's in the comments, but it does by default see # as a comment char
16:44 kadoban AFAIK git wouldn't care at all if you used markdown in your commit comments, does that help?
16:44 canton7 (and convention is to indent code, rather than put it in ```'s)
16:44 canton7 things like github interpret commit messages are markdown, iirc
16:44 SporkWitch canton7: not code, CVE number as a link to the posting
16:44 canton7 the git tooling just sees it as plain text
16:44 _ikke_ It does not interpret it
16:44 canton7 SporkWitch, I got what you meant. I was talking about the one convention which git commit messages *do* have
16:44 SporkWitch fair enough; thanks!
16:44 Noldorin joined #git
16:44 canton7 for that, I wouldn't use a markdown link
16:45 SporkWitch ah!, sorry misunderstood your intent, canton7 lol
16:45 canton7 Makes it slightly harder to read in plain text, and things like github will render links as clickable even without the markdown stuff around them
16:46 SporkWitch understood, just listed the CVE number
16:46 shatnerz joined #git
16:47 shatnerz hey when merging files, is there some easy way to specify if you want to just keep the local or remote version? Right now I'm using mergetool with meld
16:47 rubyonrailed joined #git
16:47 kulelu88 joined #git
16:48 canton7 man git checkout --ours/--theirs
16:48 _ikke_ shatnerz: Do you have a merge conflict or not?
16:48 gitinfo the git-checkout manpage is available at http://jk.gs/git-checkout.html
16:48 shatnerz _ikke_: yes, i want to go through the conflicts
16:48 shatnerz some need manual merging but alot will either keep the remote or local version
16:49 _ikke_ shatnerz: Like canton7 said, you can use git checkout --ours or --theirs to pick one version for the entire file
16:50 p4trix joined #git
16:51 osse Nota bene postscript: It picks the *entire* file. Even the parts where git was able to merge the contents
16:52 MattMaker joined #git
16:52 daynaskully joined #git
16:53 shatnerz osse: so it will contains the changes made by merge? Is there a way to avoid that
16:53 osse shatnerz: no, what I said is just the opposite
16:53 jimi_ joined #git
16:53 jimi_ joined #git
16:54 shatnerz osse: oh ok, perfect
16:54 shatnerz Thanks guys
16:54 GenZai joined #git
16:54 durham joined #git
16:55 fuchstronaut joined #git
16:56 durham joined #git
16:57 d0nn1e joined #git
16:58 durham_ joined #git
17:00 lacrymology joined #git
17:00 C4ta|Work joined #git
17:02 MattMaker joined #git
17:03 AnarchyAo joined #git
17:04 Catachan_ joined #git
17:04 atomicb0mb joined #git
17:04 g38 joined #git
17:05 Sasazuka joined #git
17:08 g38 hello. I am at a loss: in my local repo (connected to a remote repo, but I don't care about the remote, am happy to destroy it) I have a big file which was committed a long time ago and I want to remove it completely from the repo including the index. I have applied filter-branch and gc but the file is still in .git/objects/ (in view of its size). Is it because of its presence in the remote repo? removing the remote repo doesn't seem t
17:08 g38 o solve this problem. help!
17:09 C4ta|Work joined #git
17:09 osse g38: man git-filter-branch
17:09 C4ta|Work joined #git
17:09 gitinfo g38: the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html
17:09 osse see the checklist for shrinking near the bottom
17:10 ams__ joined #git
17:10 lvns joined #git
17:10 osse " People expect the resulting repository to be smaller than the original, but you need a few more steps to actually make it smaller, because Git tries hard not to lose your objects until you tell it to."
17:10 chachasmooth joined #git
17:10 g38 I've seen this, and tried cloning the repo, to no avail. but I will check the first steps given there (before cloning the repo)
17:11 jangeador joined #git
17:11 skylite joined #git
17:11 osse did you remember to use file :/// ?
17:11 mehola joined #git
17:12 g38 no I didn't, I'll look at this, thanks
17:13 g38 ah, that did the trick!  1) thanks so much!  2) if it should depend on this, then something is really broken
17:14 osse g38: by default git clone just hard links stuff to be fast. either --no-local or file:/// disables that
17:14 osse you may have noticed that the clone you just made took longer
17:15 osse man git-clone for deets
17:15 gitinfo the git-clone manpage is available at http://jk.gs/git-clone.html
17:15 eivarv joined #git
17:15 slayerjain joined #git
17:15 g38 ah, I see. hard linking by default seems like a strange idea, I would not have been happy to find the files in my original repo changed after I modified them in the cloned repo.
17:16 osse g38: it only hard links stuff in .git/objects. and stuff in there never change after they're first written
17:16 g38 ah, I see. okay, then. thanks for all your tips.
17:17 MattMaker joined #git
17:17 Akimb joined #git
17:17 osse \bde (nada|rien)\b
17:17 skalpin is there a command that will let me delete all local branches where the upstream has been deleted?
17:18 _ikke_ skalpin: No, there is no way to differentiate between those branches, and branches you created locally and never had a remote
17:18 hobodave joined #git
17:19 _ikke_ Exception is for branches which had upstream information set
17:19 InfoTest joined #git
17:19 osse and for those you can do  git branch -vv | awk '/: gone]/{print $1}' | xargs git branch -d
17:19 skalpin git branch -vv shows that the remote branch is "gone"
17:19 _ikke_ right
17:19 rominronin joined #git
17:19 _ikke_ osse: It would be nice if you could add that info to git for-each-ref ;-)
17:20 osse ;_;
17:20 Darcidride joined #git
17:20 skalpin problem #2, I'm on windows so I don't have awk or xargs
17:20 _ikke_ skalpin: Do you have git bash installed?
17:20 osse _ikke_: there was some talk of unifying for-each-ref's formatting with some other stuff. Maybe I've stepped on some toes, or maybe my patch has more far-reaching effects than I had considered and hence it was just ignored
17:21 rubie joined #git
17:21 _ikke_ osse: my experience is that they would not ignore patches in such a case
17:21 skalpin problem #3 I'm a liar, I do have git bash and I do have awk xargs
17:21 _ikke_ I think it just snowed under, some large patch series appeared
17:22 osse _ikke_: I was thinking it got ignored because it's obviously stupid for the reasons I listed maybe
17:23 osse skalpin: just promise me to never have a commit with ': gone]' somewhere in the first line of the commit message
17:24 davisonio joined #git
17:24 _ikke_ haha
17:24 osse can be made more robust my I'm feeling a sudden onset of laziness
17:24 * skalpin holds hand up
17:24 skalpin osse: I promise
17:25 hwrd_ joined #git
17:26 mingrammer joined #git
17:26 osse _ikke_: i can send it in later in the week
17:26 osse maybe I should check pu/next
17:27 _ikke_ I already checked
17:27 le_melomane joined #git
17:27 MattMaker joined #git
17:27 osse <3
17:27 sebboh joined #git
17:27 _ikke_ osse: typically you'd first see a branch appear, later it might either merged to next or to pu
17:28 osse _ikke_: no I mean check pu/next for stuff that might conflict
17:29 _ikke_ ah
17:30 osse skalpin: more robust. | awk '$4 == "gone]" { print $1 }' |
17:32 rnsanchez_ joined #git
17:33 _ikke_ osse: merge conflict with pu
17:33 hwrdprkns joined #git
17:34 _ikke_ someone changed the code in ref-filter.c
17:34 ArchNoob joined #git
17:35 cyan__ joined #git
17:35 dsdeiz_ joined #git
17:36 _rnsanchez_ joined #git
17:37 osse thanks, I'll take a look
17:39 osse ahh, this is the consolidation stuff I was thinking of
17:40 _ikke_ kn/ref-filter-branch-list
17:40 rnsanchez_ joined #git
17:40 DieguezZ joined #git
17:40 _ikke_ That's the branch it came from
17:41 daniel_rose joined #git
17:41 osse _ikke_: heh
17:41 osse https://github.com/git/git/commit/0c832b0a2d83466f96cd4d0e0e5706a52a6645bf
17:42 atomicb0mb joined #git
17:42 _ikke_ ah, so it's already implemented
17:42 osse back to my "ignored" theory
17:42 MattMaker joined #git
17:42 xaviergmail joined #git
17:42 osse s/theory/hypothesis/
17:42 rorro joined #git
17:44 _ikke_ Or postponed ;-)
17:44 C4ta|Work joined #git
17:44 osse But this is not the same!!
17:45 osse Mine was optional. This adds gone as appropriate, but there's no way to ignore it
17:45 _ikke_ Then you can mention that :P
17:46 LionsMane joined #git
17:46 zivester joined #git
17:46 osse [PATCH v3] "Changes since v2: angrier commit message"
17:47 Nugget heh
17:47 _ikke_ source?
17:47 Nugget related: https://twitter.com/gitlost
17:48 piling joined #git
17:49 Gsham joined #git
17:49 Manuel15 joined #git
17:49 Manuel15 joined #git
17:52 rubyonrailed joined #git
17:53 adac joined #git
17:53 multi_io joined #git
17:56 geekbri joined #git
17:56 tabrez joined #git
17:57 moneylotion joined #git
17:57 MattMaker joined #git
17:59 peepsalot joined #git
17:59 piratepete joined #git
17:59 mohabaks joined #git
18:00 piratepete if I have fork me/otherapp can I set remote them/otherapp? and still have remote me/otherapp?
18:01 piratepete I'm trying to understand re-basing a fork before I actually do it
18:01 JordiGH joined #git
18:01 roelmonnens joined #git
18:01 _ikke_ piratepete: A repository can have multiple remotes
18:01 JordiGH What do bare repos do with HEAD? Can I change where HEAD is with a push?
18:01 piratepete _ikke_ perfect! I
18:01 _ikke_ JordiGH: The branch checked out on clone is based on what HEAD points to
18:02 _ikke_ JordiGH: No, you cannot update it by pushing
18:02 piratepete I'm getting my next question
18:02 kbs Nugget: heh, entertaining
18:02 JordiGH _ikke_: Oh, I see. And how is it changed, then? Someone with shell access has to go and modify it?
18:03 _ikke_ JordiGH: yes, git symbolic-ref
18:03 _ikke_ JordiGH: github exposes it in the webinterface
18:03 piratepete I can alias it like "git remote upstream http://github.com/them/someapp?
18:03 JordiGH No amount of --forcing can change it, eh?
18:03 JordiGH _ikke_: Alright, thanks.
18:03 _ikke_ JordiGH: No, you cannot push symbolic refs
18:03 JordiGH left #git
18:03 _ikke_ piratepete: git remote add <name> <url>
18:04 hahuang61 joined #git
18:04 piratepete And, git fetch upstream (from my local fork gets latest in main repo that I forked)
18:05 Rodya_ joined #git
18:05 piratepete git checkout master and then git rebase upstream/master makes my local master contain the latest of the forked repo?
18:06 gusnan joined #git
18:06 jaafar joined #git
18:06 _ikke_ yes
18:06 piratepete _ikke_ thanks :-)
18:06 osse _ikke_: Hah! Just got an email from Peff himsleff
18:06 _ikke_ osse: nice
18:06 _ikke_ See it was not ignored :P
18:07 osse He even commented on v1
18:07 inflames joined #git
18:08 _ikke_ osse: usually you reply to v1 on a v2 patch, so that they stay together
18:09 osse I want to reply but I am a noob. I used to use gmane but now I can't
18:09 osse should I do reply-to-all in gmail ?
18:09 _ikke_ yes
18:09 _ikke_ ah, reply to v2 too
18:10 osse i will
18:10 _ikke_ I mean, he replied to v2
18:10 zacts joined #git
18:10 _ikke_ https://public-inbox.org/git/20160824180730.mqlvhkthg4imd2vy@sigill.intra.peff.net/T/#t
18:10 tilerendering joined #git
18:11 tilerendering hey
18:11 _ikke_ hello
18:11 tilerendering for some reason, when I update my git repository from my IDE, it seems to have the effect that all files pushed may then only be pulled in a way that the pulled files require root permission. is that possible somehow ?
18:11 Gsham joined #git
18:12 EduardoMartins1 joined #git
18:12 _ikke_ if it pushes as a different linux user, then yes
18:12 _ikke_ (and permissions are set very restrictive)
18:13 tilerendering ok but how do I configure it in a way that all users from group say www-data can push and pull and write etc. ?
18:13 _ikke_ there are two approaches to this
18:13 roelmonnens joined #git
18:14 tilerendering I use gitolite
18:14 roelmonnens joined #git
18:14 _ikke_ That's the good approach :-)
18:15 dsantiago joined #git
18:15 _ikke_ Let everone push as the git user
18:15 tilerendering see I need all users who have the permission to push and pull to the repo say webdev to have all write and read perms
18:15 _ikke_ No
18:15 tilerendering ok
18:15 tilerendering all as the git user ok ?
18:15 _ikke_ You should always push using git@server
18:15 choki joined #git
18:15 tilerendering alright
18:15 tilerendering where do I configure this locally at the client side ?
18:15 _ikke_ git remote
18:16 boombatower joined #git
18:16 tilerendering can you give me the params ?
18:16 _ikke_ git remote set-url <name> <url>
18:16 tilerendering can I get the current status of name and url ?
18:17 toralf left #git
18:17 _ikke_ git remote -v
18:17 MattMaker joined #git
18:17 roelmonnens joined #git
18:18 tilerendering hm taht s very strange.
18:18 _ikke_ what is?
18:18 tilerendering because it says origin git@myhost.com (fetch)
18:18 sebboh joined #git
18:18 tilerendering then origin git@myhost.com (push)
18:18 _ikke_ ok, that looks good
18:18 tilerendering so it seems it s already configured like that
18:19 tilerendering this is true for my workstation and the workstation of my colleague
18:19 tilerendering same results
18:19 _ikke_ Now to find out how your IDE pushes then
18:19 rominronin joined #git
18:19 moneylotion joined #git
18:19 boombatower joined #git
18:19 leehambley joined #git
18:20 mingrammer joined #git
18:20 charging1awk joined #git
18:21 durham joined #git
18:21 romankl joined #git
18:21 charging1awk left #git
18:21 tilerendering hm it seems to not work with just one user, where I get: error: insufficient permission for adding an object to repository database .git/objects
18:21 _ikke_ Then you need to check his settings
18:22 dviola joined #git
18:22 tilerendering they are the same, that´s what is strange
18:23 romankl joined #git
18:23 _ikke_ can you take a look at .git/objects?
18:23 _ikke_ ls -ld .git/objects
18:23 fished joined #git
18:23 tilerendering as the concerned user ? no
18:23 daumie joined #git
18:23 tilerendering even though he s a member of www-data
18:23 tilerendering it s really strange
18:23 _ikke_ That would not matter
18:23 daumie Hello guyz
18:23 tilerendering ok
18:23 fished joined #git
18:23 _ikke_ because it should always push as git
18:23 _ikke_ so the git user needs permissions
18:24 daumie what's the first step in solving a merge conflict?
18:24 tilerendering yeah I have drwxrwxr-x 127 www-data www-data 4096 Aug 24 20:21 .git/objects
18:24 tilerendering the missing w maybe
18:25 fished joined #git
18:26 Darcidride joined #git
18:26 tilerendering i have no clue anymore
18:26 osse _ikke_: replied
18:26 osse I wonder if it looks good on the list
18:27 _ikke_ No reply yet
18:27 osse me *worried*
18:27 _ikke_ Nope, not yet
18:27 _ikke_ https://public-inbox.org/git/
18:28 MattMaker joined #git
18:28 _ikke_ Now
18:29 osse Yaaay
18:29 HalfEatenPie joined #git
18:29 lolmac joined #git
18:29 osse OOps. I forgot to say "_ikke_ discovered"
18:29 JrCs joined #git
18:29 _ikke_ Don't mind
18:29 _ikke_ Was your idea to even check :P
18:29 hahuang61 joined #git
18:29 _ikke_ I just did the git merge
18:30 blackwell joined #git
18:31 Peng joined #git
18:31 rgrinberg joined #git
18:31 tilerendering ok the user cant even do a local commit
18:31 rnsanchez joined #git
18:33 Torrone joined #git
18:34 lvns joined #git
18:34 Torrone hello, why does this keep happening to me? I ran git reset --hard <commit> and now it won't let me push changes claiming 'Everything up-to-date'. Wat do? Thanks
18:35 _ikke_ Torrone: It probably means that the thing you try to push *is* up-to-date
18:35 grawity sounds like that commit has already been pushed?
18:35 osse _ikke_: "If you're willing and able to do so[review], that would be a big help.  Thanks." Weeelll I'm fucked
18:35 moneylotion joined #git
18:35 Torrone grawity, no, it isn't there, _ikke_ I've just made a commit with multiple insertions (10 files changed, 383 insertions(+), 340 deletions(-))
18:36 _ikke_ Torrone: Then verify what you are pushing
18:36 Torrone git status says HEAD detached from 0f2fd2c
18:36 _ikke_ aha
18:36 Torrone in RED
18:36 _ikke_ So your commits don't live on a branch
18:37 Torrone _ikke_, maybe...
18:37 choki_ joined #git
18:37 Torrone I don't know
18:38 vassagus_ joined #git
18:38 Torrone wat do? I don't want to delete all my files like the last time :P
18:38 atomicb0mb joined #git
18:40 _ikke_ get your commits back on a branch
18:40 milki Torrone: !reattach
18:40 gitinfo Torrone: Letters refer to !detached. (a) and (b): 'git checkout branchname' to continue working on another branch, or 'git checkout -b branchname' to start a new one here; (c) git am --continue; (d) git rebase --continue
18:40 _ikke_ git branch <tmp>
18:40 dtrott joined #git
18:40 Torrone ok
18:41 Torrone still detatched
18:41 _ikke_ yes
18:41 _ikke_ that's step one
18:41 rorro joined #git
18:41 _ikke_ git log --oneline <branch>..tmp
18:42 mjuszczak joined #git
18:42 dtrott I want to merge a release version, $TAG onto the master branch overwritting all contents on master, is there a better way of doing it than this?   :    git checkout master  ;     git merge `git commit-tree -p master -p $TAG -m "$MSG" "${TAG}^{tree}"`
18:42 Torrone got all my commits
18:42 _ikke_ ok
18:42 _ikke_ git rebase <branch> tmp
18:43 Torrone says 'Current branch HEAD is up to date.'
18:44 daku_ joined #git
18:44 robinson joined #git
18:44 Torrone git status is still detached
18:45 _ikke_ Ok, then git checkout master, git merge tmp
18:45 daku_ Anyone use Git on Visual Studio Team Services with Visual Studio Code on Ubuntu?
18:46 skalpin daku_: no, but I'm using git in vsts
18:46 vuoto joined #git
18:46 ahmedelgabri joined #git
18:46 Torrone _ikke_, still nothing to commit :|
18:47 Torrone but the new commit is shown in local git status
18:47 _ikke_ what do you mean, nothing to commit
18:47 _ikke_ I mean, what did the merge do?
18:47 daku_ skalpin, how did you add the remote repo? when I use git remote add <URL> I'm told its not a repo, but thats literally what VSTS is telling me to use
18:47 Torrone git status
18:47 Torrone On branch master
18:47 Torrone nothing to commit, working directory clean
18:47 MattMaker joined #git
18:47 _ikke_ and you did git merge tmp?
18:47 Torrone yes
18:47 _ikke_ git show
18:48 Ediz joined #git
18:48 _ikke_ does it show the latest commit you expect?
18:48 Torrone yes
18:48 _ikke_ ok
18:48 _ikke_ git push now?
18:48 Torrone _ikke_, ok looks good now :)
18:49 skalpin daku_: I cloned from vsts, but you should be able to 'git remote add "https://<yourproject>.visualstudio.com/_git/<projectname>
18:49 vassagus joined #git
18:49 Rodya_ joined #git
18:50 daku_ yeah I get "fatal: Not a git repository (or any parent up to mount point /home/daku)
18:50 daku_ Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).
18:50 daku_ "
18:50 daku_ skalpan Sorry if I'm doing something retarded. I've never used git before.
18:51 osse _ikke_: hmm, I got a new reply at 20:34 but it's not visible on the thingy. Later emails are :S
18:51 cdown_ joined #git
18:51 milki daku_: your not running the command from your git repo
18:51 milki e
18:51 ArchNoob joined #git
18:51 _ikke_ osse: was the mailing-list cc'ed?
18:52 daku_ milki, thanks, maybe I should just run through a git guide of sorts since I'm still a noob at source control in general.
18:52 SteffanW joined #git
18:53 moneylotion joined #git
18:53 cdown__ joined #git
18:53 osse _ikke_: yes. there were to cc's. first the guy with the topic in pu and then the list
18:53 osse maybe that affects it
18:54 _ikke_ should not
18:54 sagerdearia joined #git
18:54 skalpin daku_: can you clone the repo?
18:55 daku_ skalpin yeah,git clone works, which is why im confused on the remote add haha
18:56 skalpin daku_: if the clone worked, then your remote will be setup as origin. You can see it with 'git remote -vv'
18:56 Caelum joined #git
18:56 grawity `git clone` does not automatically cd you into the repo
18:56 grawity you got to do that yourself
18:56 fuchstronaut joined #git
18:57 MattMaker joined #git
18:58 Dougie187 left #git
18:58 daku_ grawity, skalpin *MEGA FACEPALM* Sorry dudes, its one of those days
18:59 vassagus joined #git
18:59 daku_ grawity, skalpin CD to directory worked. Just need to figure out getting VS Code to work with it.
18:59 roelmonnens joined #git
19:00 zeroed joined #git
19:00 skalpin daku_: when you use vscode, file -> open folder... and open the whole folder. Then when you click on the git icon on the left you should see if there are any changes to commit
19:01 skalpin daku_: it won't give you any git info if you open one file
19:01 _ikke_ osse: e-mail finally camei n
19:03 daku_ skalpin, guys, thank you haha. You saved me a grueling time with google.
19:03 skalpin daku_: np
19:06 Repox joined #git
19:06 Infamous joined #git
19:07 sebboh joined #git
19:08 vassagus joined #git
19:09 kenansulayman joined #git
19:10 adac joined #git
19:10 roelmonnens joined #git
19:12 mr_quincy left #git
19:15 StuartMI joined #git
19:16 daku_ left #git
19:17 MattMaker joined #git
19:18 choki joined #git
19:19 lvns joined #git
19:20 rominronin joined #git
19:20 hwrdprkns joined #git
19:21 andlabs joined #git
19:21 morenoh149 joined #git
19:21 andlabs Googling is failing me so: is there a way I can tell git that a given file MUST be cahnged as part of a commit?
19:21 andlabs fail if the file hasn't been changed
19:21 bittin joined #git
19:22 kadoban andlabs: With a hook I guess, sounds weird though.
19:23 lvns joined #git
19:23 vuoto joined #git
19:24 andlabs ok
19:24 hwrdprkns joined #git
19:27 hwrdprkns joined #git
19:27 MattMaker joined #git
19:30 vassagus_ joined #git
19:31 DsAAzvTVQt joined #git
19:35 m0viefreak joined #git
19:37 alansaul joined #git
19:38 Ozd_ joined #git
19:38 Infamous joined #git
19:38 kenansulayman joined #git
19:39 fuchstronaut joined #git
19:40 joki joined #git
19:40 VaUsfOUbRl joined #git
19:40 jtprog_ joined #git
19:40 raijin joined #git
19:41 mkelly joined #git
19:42 MattMaker joined #git
19:43 vuoto joined #git
19:43 kfunk joined #git
19:44 anth0ny joined #git
19:45 vuoto joined #git
19:46 vuoto joined #git
19:47 xcGlQHdZGT joined #git
19:47 daniel_rose joined #git
19:47 daniel_rose joined #git
19:48 vuoto joined #git
19:49 osse andlabs: printf '%s\n' '#!/bin/sh' '! git diff --quiet --staged -- a/given/file' > .git/hooks/pre-commit && chmod +x .git/hooks/pre-commit
19:49 DieguezZ joined #git
19:49 osse andlabs: ez. though you might want to print an error message
19:51 supsup joined #git
19:53 supsup how can i utilize a grep of commit messages to then get a list of all files from those commits?
19:53 _ikke_ git log -S "string" --name-only
19:53 _ikke_ something like that
19:53 rpd that's right
19:53 _ikke_ -S looks for commits that eithers adds or removes string
19:53 supsup is -S short for --grep?
19:53 johnmilton joined #git
19:54 Dougie187 joined #git
19:54 _ikke_ no
19:54 supsup i was thinking I had to pipe th is to something:  git log --all --grep='somevalue'
19:54 _ikke_ --pick-axe
19:54 rpd supsup: you could just call git log <hash-ish> --name-only on every commit you grep
19:54 supsup cool going to try it
19:54 _ikke_ --grep searches for commit messages
19:54 _ikke_ -S searches in the changes
19:54 supsup yeah someone put a string in the commit message
19:54 supsup they want me to find all the commits with that message
19:55 supsup then i need to get a list of files from all those commits
19:55 sebboh joined #git
19:55 Infamous joined #git
19:55 rpd supsup: if your earlier query gets you the commits you want, just add --name-only to only get a list of files
19:55 supsup i did and it kind of works
19:56 supsup but it returns the commit # also and author and date
19:56 supsup do i have to use sed?
19:56 _ikke_ no
19:56 roelmonnens joined #git
19:56 _ikke_ What do you want to output?
19:56 supsup just the file paths
19:57 supsup then i'm going to do a uniq over them
19:57 _ikke_ git log --format=
19:57 _ikke_ Then you can use uniq
19:57 MattMaker joined #git
19:57 _ikke_ (though, you might need to sort first)
19:58 supsup what is the format for just path?
19:59 _ikke_ --name-only
19:59 _ikke_ and then --format= to supress other output
19:59 fuchstronaut joined #git
19:59 le_melomane joined #git
19:59 supsup still does author etc
19:59 supsup commit / author/date
20:00 supsup msg and path
20:00 supsup trying this:  git log --all --grep='evilpsudotag'  --name-only --format=
20:01 andlabs osse: yeah, I got it already
20:01 andlabs now how do I make sure it gets pushed to origin
20:01 TooLmaN joined #git
20:01 _ikke_ "git log --name-only --format=" works for me
20:02 slassh joined #git
20:02 A763736 joined #git
20:02 A763736 left #git
20:03 supsup tried exact thing still has commit #/auth/date/commit msg + path
20:03 supsup does using grep change it?
20:03 roelmonnens joined #git
20:03 supsup git log --name-only --format='' --grep='matchme'
20:03 vuoto joined #git
20:03 Torrone joined #git
20:05 vuoto joined #git
20:05 ASOLAr joined #git
20:06 supsup guess i can use sed just seems odd
20:06 renlo joined #git
20:06 _ikke_ adding --grep does not add output
20:06 _ikke_ what version of git do you have?
20:07 supsup 1.9.5.msygit.1
20:07 _ikke_ A bit old
20:07 lvns joined #git
20:07 mjuszczak joined #git
20:07 _ikke_ we're currently at 2.9.3
20:07 MattMaker joined #git
20:08 supsup alright i'll upgrade
20:08 gynter left #git
20:08 mingrammer joined #git
20:08 hwrdprkns joined #git
20:08 cdg joined #git
20:09 lvns joined #git
20:09 andlabs so how do I push my hooks into origin so they're avialable everywhere?
20:09 andlabs to everyone
20:10 _ikke_ andlabs: git does not allow you to clone hooks as is
20:10 _ikke_ andlabs: You can always add them to the working tree
20:10 _ikke_ and commit them
20:10 _ikke_ but people have to put them in place themselfs
20:11 supsup yep
20:11 supsup version thanks
20:11 _ikke_ alright
20:12 db42 joined #git
20:12 supsup so yeah guy said he tagged all his stuff....but he used a comment to tag it. is that real life?
20:12 db42 how do i copy my current branch into another name in a remote ?
20:13 _ikke_ git push <remote> <branch>:<othername>
20:14 db42 amazing !
20:16 db42 no googling found me this answer so easy, maybe there should be a shell script git-irc-help <question> that throws back the first reply :)
20:16 renlo is anyone else getting asked to log in for every single fricken git-lfs file?
20:16 supsup so is there a way to feed into git checkout -- a list of paths from this file I just exported?
20:17 supsup maybe with xargs
20:17 _ikke_ yeah
20:17 _ikke_ What do you want to achieve?
20:17 fuchstronaut joined #git
20:17 jh3 joined #git
20:17 supsup i want to checkout from stage branch into master all these files they committed into stage
20:17 supsup like in master :  git checkout stage -- (feed in all these files)
20:18 supsup normally i just do it file by file but there is a ton of files
20:19 supsup like i have a matches.txt file that has a list of all the file paths.
20:19 rgrinberg joined #git
20:19 anonymuse joined #git
20:20 evie_hammond joined #git
20:21 tgy left #git
20:21 rominronin joined #git
20:21 LeMike joined #git
20:22 mattcen joined #git
20:22 MattMaker joined #git
20:24 zacts joined #git
20:26 alansaul joined #git
20:27 abyx joined #git
20:28 rominronin joined #git
20:28 fuchstronaut joined #git
20:28 mohabaks joined #git
20:29 db42 left #git
20:29 daniel_rose joined #git
20:29 clandestino joined #git
20:30 p4trix joined #git
20:31 xaviergmail joined #git
20:31 lvns_ joined #git
20:33 aidalgol joined #git
20:33 jangeador joined #git
20:34 zarel joined #git
20:35 Endarked_ joined #git
20:36 LeMike_ joined #git
20:37 MattMaker joined #git
20:38 raijin joined #git
20:38 g38 joined #git
20:38 LeMike joined #git
20:39 Jon28 joined #git
20:39 duderonomy joined #git
20:39 thiago joined #git
20:42 sebboh joined #git
20:42 stamina joined #git
20:43 githogori joined #git
20:45 tvw joined #git
20:47 robotroll joined #git
20:50 alansaul joined #git
20:50 cdown_ joined #git
20:51 Fazer2 joined #git
20:52 d4rklit3 joined #git
20:52 Fazer2 hi, is it possible to change git gui background and font colors? the white background is too blinding
20:52 jstein_ joined #git
20:52 MattMaker joined #git
20:53 cdown__ joined #git
20:55 _ikke_ Fazer2: It folows GTK settings
20:55 jtprog_ joined #git
20:56 renlo joined #git
20:58 j416 joined #git
20:58 LeMike joined #git
20:59 LeMike joined #git
21:02 hiroki A GIT (+1.8.2) folder -> "/tmp/repo.git/{folder1,folder2,folder3}" coupled with a '/tmp/repo.git/.gitignore' containing 'folder2/**' will ignore the complete folder2 folder + everything inside of it recursively / until infinity. Is that correct ?
21:03 mingrammer joined #git
21:03 afuentes joined #git
21:03 supsup left #git
21:04 ojacobson I would spell that '/folder2/' instead of 'folder2/**'
21:04 ojacobson the leading slash anchors it to the root of the repo, and the trailing slash is sufficient for "this directory and all of its contents"
21:05 hiroki Ah great ! Did not know that ojacobson. The behaviour of not needing **, is that pre 1.8.2 as well ?
21:05 ojacobson I believe so
21:06 hiroki Thanks a bunch
21:06 dreiss joined #git
21:07 Fazer2 _ikke_: interesting, I'll check it on linux later, I guess you don't know how to change it on windows?
21:07 alansaul joined #git
21:09 raijin joined #git
21:09 j416 joined #git
21:09 ochorocho__ joined #git
21:11 hwrdprkns joined #git
21:12 MattMaker joined #git
21:12 Gsham joined #git
21:12 ASOLAr joined #git
21:13 srcerer_ joined #git
21:15 mohabaks joined #git
21:15 bittin joined #git
21:15 Puffball joined #git
21:15 ajf- joined #git
21:20 davisonio joined #git
21:21 renlo joined #git
21:22 hahuang61 joined #git
21:22 abyx joined #git
21:22 MattMaker joined #git
21:23 str joined #git
21:23 rbern joined #git
21:25 zacts joined #git
21:26 fusionx86 joined #git
21:26 macivy joined #git
21:28 rgrinberg joined #git
21:29 anarchos2 joined #git
21:30 sebboh joined #git
21:34 hwrdprkns joined #git
21:36 Bennux joined #git
21:37 MattMaker joined #git
21:37 Torrone joined #git
21:38 realz joined #git
21:39 chachasmooth joined #git
21:40 spudowiar joined #git
21:41 MrWoohoo joined #git
21:41 elementalest joined #git
21:41 j416_ joined #git
21:42 duderonomy joined #git
21:43 button joined #git
21:45 Samual joined #git
21:45 astrofog joined #git
21:46 patarr joined #git
21:47 j416 joined #git
21:48 jimi_ joined #git
21:50 Spec-Chum joined #git
21:51 anonymuse joined #git
21:54 dsantiag_ joined #git
21:55 dsantiag_ joined #git
21:57 MattMaker joined #git
21:59 inflames joined #git
22:04 fuchstronaut joined #git
22:04 Manuel15 joined #git
22:04 Manuel15 joined #git
22:05 nitric joined #git
22:05 ahr3n joined #git
22:06 beyertech joined #git
22:09 _August_ joined #git
22:09 j416 joined #git
22:09 whomp joined #git
22:10 _August_ need confirmation on merging a branch with another branch. I am trying to merge branch 1.2 to 1.3. Do i checkout 1.3 first and then merge or the other way around?
22:11 stomplee joined #git
22:12 whomp i just ran "merge master", fixed some merge conflicts, and pushed it, and i get: https://gist.github.com/michaeleiselsc/7c9a8886b730d7e125a9c220ede3fe3c
22:12 whomp any ideas where this big file would come from?
22:12 MattMaker joined #git
22:13 habbasi joined #git
22:13 beyertech_ joined #git
22:14 habbasi Hello. Kind of a weird question, how do I get a date for a specific commit?
22:14 morenoh149 joined #git
22:14 shinnya joined #git
22:14 vassagus joined #git
22:14 chachasmooth joined #git
22:14 habbasi I tried git show -s --format=%ci commit but it says "unknown revision or path not in the working tree"
22:14 axl__ joined #git
22:15 kuldeep joined #git
22:15 johnny56_ joined #git
22:16 nitric joined #git
22:16 beyertech joined #git
22:17 sebboh joined #git
22:18 kus joined #git
22:18 kbs whomp: not that I know a solution, but out of curiosity have you already taken a look at that blob to see what sort of a file it is?
22:19 rpd habbasi: that command is fine except that you don't specify the object you want to show
22:19 rpd just how to show it
22:19 rpd unless I'm misunderstanding
22:19 beyertech_ joined #git
22:19 rpd habbasi: does `git show <commit-ish>` show the commit
22:21 beyertech joined #git
22:23 habbasi rpd: Same output.
22:23 habbasi rpd: I'm using a 7-char hex string...
22:23 habbasi rpd: Is that weird?
22:24 rbern joined #git
22:25 andlabs joined #git
22:26 madewokherd joined #git
22:26 Dreamer3 joined #git
22:26 habbasi git log | grep <7-char thing> shows nothing either. :/
22:27 Salander27 joined #git
22:27 Torrone joined #git
22:27 MattMaker joined #git
22:27 mohabaks joined #git
22:27 chachasmooth joined #git
22:28 rpd uh
22:28 rpd where did you get that then?
22:29 rpd Sounds like the commit you're talking about doesn't exist in that branch if I had to guess
22:29 daynaskully joined #git
22:31 renlo joined #git
22:31 Gsham joined #git
22:32 inflames joined #git
22:34 renlo joined #git
22:35 mjuszczak joined #git
22:36 cdg joined #git
22:38 rkazak joined #git
22:38 cdown_ joined #git
22:39 FuzzySockets joined #git
22:42 MattMaker joined #git
22:42 moritz joined #git
22:43 tshirtman joined #git
22:44 j416 joined #git
22:45 tshirtman so, when i do git pull, i get displayed the range of commits that were pulled on my current branch (e.g  "bcce38b..998fe00" ), i was thinking maybe it would be nice to have a hook to auto display git log of this range, if somebody knows if there is a nice way to do that?
22:45 skylite joined #git
22:46 kadoban Step 1 would be: for all that is good and holy, don't use "git pull"
22:46 Timvde kadoban: why not?
22:46 ojacobson fortunately, that output comes from git fetch, which is reasonable even for beginners
22:47 tshirtman i know what git pull does, and i know how to abort if i don't want a merge, and i can use pull --rebase if i want a rebase
22:47 ojacobson Timvde: correctly reasoning about git pull involves correctly reasoning abou two and a half other tools. The people who have learned how those other tools work can and will ignore advice to not use git pull, though I'd agree it could be phrased better
22:47 ibourgeois joined #git
22:47 kadoban Timvde: It's extremely overused, and I find it more confusing rather than helpful.
22:47 Timvde ojacobson: So the advice is actually: "Don't teach git pull to beginners"?
22:48 Timvde kadoban: It's probably what most people want, though. Well, personally, I prefer git pull --rebase, but still
22:48 whomp so i did a git cp with a merge commit, and it tells me i need to use the -m option, and i see online that i can do -m 1 or 2. but how do i know which branch 1 and 2 refer to?
22:48 ojacobson Timvde: that's how I usually pitch it, yeah
22:48 ojacobson "probably what people want" is great until it's not; then users need to be able to reason about the theory
22:49 ojacobson (or remain reliant on external help, like this channel, indefinitely)
22:49 tshirtman ok, i guess that was the wrong question to ask
22:49 ojacobson tshirtman: it's okay, pull is a live wire
22:49 ojacobson the short version is "huh, maybe." Try suggesting it on the !list - maybe demo it with an alias or a wrapper script
22:49 gitinfo [!mailing_list] The mailing list can be reached via git@vger.kernel.org. You don't need to subscribe to the list, you will always be put in cc on reply. Read archives at http://j.mp/gitlist
22:49 ojacobson parsing porcelain output isn't great but it'll work for a demo
22:49 ojacobson (I suggest the ML because nobody who works on Git is a regular here)
22:50 tshirtman ok, thanks for the suggestion :)
22:50 tshirtman sure, i can hack a demo with grep and cut i guess :)
22:50 Timvde ojacobson: True, I've had to help enough people out of trouble for a variety of reasons :P
22:51 j416 joined #git
22:51 whomp MLs are so bad
22:51 tshirtman i least i had not much trouble fixing a wrong pull/merge with git, and i never managed (ever again) an svn merge after trying once
22:51 tshirtman or tried*
22:52 MattMaker joined #git
22:52 Gitzilla joined #git
22:56 tshirtman ojacobson: not sure who is to ping about that, but the link to the archive of the ML is dead, i guess because gmane…
22:56 duderonomy joined #git
22:56 k6ekb joined #git
22:58 j416 joined #git
22:59 Torrone joined #git
23:02 MattMaker joined #git
23:05 renlo joined #git
23:06 ojacobson oh, crud. Me, I guess…
23:06 ojacobson Now, if only I could remember how to edit factoids
23:08 energizer joined #git
23:09 chachasmooth joined #git
23:09 energizer When I'm writing a commit message, my text editor window's background is all green. Why is this?
23:10 sebboh joined #git
23:10 HedgeMage energizer: that's probably a setting in your editor.
23:10 ochorocho__ joined #git
23:10 HedgeMage energizer: git, unless you have told it otherwise, will inherit your system default text editor.
23:10 whomp joined #git
23:11 energizer i dont see anything that looks plausible in .emacs
23:12 axl_ joined #git
23:12 MattMaker joined #git
23:13 HedgeMage hrm... I know that's not the default if you use magit for emacs, but I'm not sure about git-mode...next time you find yourself editing a commit message, check what major and minor modes are active, that may provide a clue
23:14 duderonomy joined #git
23:15 energizer HedgeMage: good idea, ill do that
23:15 ASOLAr joined #git
23:15 HedgeMage :)
23:15 HedgeMage at that point, #emacs may be more help than we are
23:16 sargas joined #git
23:16 Torrone joined #git
23:17 dendazen joined #git
23:20 dsdeiz joined #git
23:21 fahadash joined #git
23:21 atomicb0mb joined #git
23:22 ToxicFrog joined #git
23:22 MattMaker joined #git
23:25 hahuang61 joined #git
23:26 mkelly left #git
23:30 batrick joined #git
23:30 Torrone_1 joined #git
23:31 cagedwisdom joined #git
23:31 cagedwisdom left #git
23:34 Torrone joined #git
23:38 vassagus_ joined #git
23:39 fusionx8_ joined #git
23:40 Torrone joined #git
23:41 jstimm joined #git
23:42 MattMaker joined #git
23:44 Torrone_1 joined #git
23:45 skylite joined #git
23:49 Darren_ joined #git
23:51 duderonomy joined #git
23:53 Torrone joined #git
23:54 nydel07 joined #git
23:55 SpecChum joined #git
23:57 MattMaker joined #git
23:59 whomp joined #git

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