Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2017-03-14

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

All times shown according to UTC.

Time Nick Message
00:01 guzzlefry joined #git
00:02 cdown joined #git
00:03 guzzlefry Say I'm on the 'development' branch, and I need to move a PNG file from another branch into the 'development' branch. The directory location of the image would be the same. Is there a quick way to do that?
00:07 xaviergmail joined #git
00:07 rewt guzzlefry: git checkout branch -- file
00:07 Es0teric joined #git
00:08 Noodlewitt joined #git
00:09 kenny joined #git
00:10 guzzlefry thanks
00:10 rmbeer left #git
00:11 whaley joined #git
00:12 kenny Hello, I used to 'git add -i' -> 'patch' and it would ask me which files. Now it doesn't. Anyone know if some setting changed that I missed?
00:12 guzzlefry left #git
00:14 mizu_no_oto joined #git
00:15 anuxivm joined #git
00:16 viktor[m] joined #git
00:16 kenny nm, I'm now seeing a thread on the mailing list, so it looks like it's a bug.
00:16 Murr joined #git
00:17 viktor[m] joined #git
00:17 gitgud joined #git
00:17 gitgud hello all
00:19 mikecmpbll joined #git
00:19 rominronin joined #git
00:20 jstimm joined #git
00:20 oich joined #git
00:22 ljc joined #git
00:23 viktor[m] left #git
00:23 xissburg_ joined #git
00:24 Kaisyu joined #git
00:25 kbs joined #git
00:26 nevodka joined #git
00:27 kenny left #git
00:27 cdown_ joined #git
00:28 jdelreal joined #git
00:28 kbs having a brain fart - let's say X and Y are two different commits, pointing to a common parent B. Merging X and Y can result in tree that's different from rebasing (say) Y onto X right?
00:31 masuberu joined #git
00:31 ojacobson kbs: generally not
00:32 ojacobson the exceptions mostly involve conflicts
00:35 kbs ojacobson: right, that's what I was wondering - i.e., if at the end the same tree merge algorithms get used in a different sequence, or there's an entirely different approach to rebasing
00:36 kaldoran joined #git
00:36 kbs (in my quick test, it appeared to me that exactly the same merge conflict arises whether rebasing or merging)
00:37 gitgud left #git
00:38 Cabanossi joined #git
00:39 finder2017 joined #git
00:39 apeman joined #git
00:40 apeman left #git
00:41 kbs (to put it differently :-) when I read rebasing == "replaying a sequence of diffs" and I read merging == "3-way-merge" - trying to think of a concrete example when they would be different
00:43 finder2017 left #git
00:44 ojacobson rebase repeatedly applies 3-way merge
00:44 ojacobson merge applies it in one step
00:45 kbs ah, that's very succinctly clarified - thanks! my exact question
00:45 Whiskey is thee anyway to change pushed commit messages?
00:45 ojacobson non-conflicting merges are generally composable, so the resulting tree should be the same
00:45 Whiskey Whit other words, not the file just the message that was given
00:45 kbs ojacobson: *nod* perfectly explained, thank you
00:46 brent__ joined #git
00:50 cdown joined #git
00:52 kbs Whiskey: you can git-commit --amend, but that will create a *new* commit, and subsequent potential cries of anguish from others if you've already pushed it.
00:52 impliedn joined #git
00:53 menip joined #git
00:53 livingstn joined #git
00:54 dreiss joined #git
00:55 masuberu joined #git
00:56 oaaao joined #git
01:01 Whiskey i was after "git rebase -i HEAD~x"
01:04 BackEndCoder joined #git
01:04 dendazen joined #git
01:05 dviola joined #git
01:08 Cabanossi joined #git
01:08 elricsfate joined #git
01:08 elricsfate Hello all
01:09 jimklo joined #git
01:09 elricsfate I recently forked a dead repository that had some open pull requests. I'd like to integrate said pull requests into my fork. How can I do this while providing proper attribution?
01:09 cbreak elricsfate: git merge
01:10 cbreak elricsfate: do a git ls-remote of the remote, fetch the ref you want, then merge
01:12 kbs joined #git
01:13 elricsfate So I can look at the latest commit and then just fetch? Sweet.
01:13 elricsfate Thank you for the help cbreak
01:13 cbreak elricsfate: at least on github, pull requests get a ref
01:14 elricsfate Is there any reason I should ref the pull instead of just refing the commit on the forked repo of the user who submitted said pull?
01:14 cbreak no reason
01:14 cbreak you can get it from where ever you want
01:14 elricsfate Sweet, thanks
01:14 cbreak but it's more convenient to get them from the source
01:14 cbreak (the one you forked from)
01:14 cbreak since that's your upstream already
01:15 alexandre9099 joined #git
01:15 a_thakur joined #git
01:16 Olipro joined #git
01:18 gugah joined #git
01:18 roelmonnens joined #git
01:18 TbobbyZ_ joined #git
01:19 jimklo joined #git
01:20 cqi joined #git
01:21 jimklo joined #git
01:22 Koterpillar joined #git
01:23 Lyqyd joined #git
01:23 glebihan_ joined #git
01:24 jimklo joined #git
01:24 am55_ joined #git
01:24 Koterpillar We've changed the default branch of a GitHub repository (from 'dev' to 'develop'), and now 'git gc' fails with: fatal: bad object refs/remotes/origin/HEAD error: failed to run repack. How do I fix this?
01:29 jimklo joined #git
01:29 ankit01ojha joined #git
01:30 Levex joined #git
01:30 variable joined #git
01:31 elricsfate Alright, so I'm guessing I need to checkout the remote branch/repo, fetch the commit, and then merge into my master?
01:31 elricsfate cbreak:
01:31 variable Is it possible to construct a branch given a sequence of shas without actually changing working directory?
01:32 variable in other words I want eax/sxnopll/k1/eff495a3ef to point to 4ce88f43c0 -> 5eb3f1239a ...
01:32 variable obviously the shas might change
01:32 cbreak Koterpillar: git remote prune? or maybe just remove the remote and add it again
01:32 cbreak elricsfate: nah
01:32 ojacobson variable: a branch points to a single commit. `git branch` can create branches for existing commits.
01:32 cbreak elricsfate: no need to check out anythning
01:32 cbreak elricsfate: just fetch the ref you want and merge it
01:32 cbreak git fetch remotehere refhere
01:32 ojacobson If the commits you want aren't already _topologically_ related, you'll need to do more work, which will likely require the working directory
01:32 variable ojacobson what I actually want is to "rebase" a branch without checking out the working copy
01:33 elricsfate sweet, thanks cbreak
01:33 ojacobson patch application more or less requires a work tree, sorry
01:33 jrm joined #git
01:33 variable :( okay
01:33 ojacobson git worktree?
01:33 variable can I try applying the patch to the index
01:33 variable worktree won't work, because I'm actually trying to avoid the I/O overhead
01:33 variable not just messing up my current worktree
01:35 kyan joined #git
01:36 variable ojacobson even more specifically I'm trying to figure out how to speed up https://github.com/unixdotfiles/bin/blob/master/git-all-rebase
01:37 Sasazuka joined #git
01:37 variable I've used worktree + parallelism but the I/O cost is high
01:37 elricsfate That doesn't seem to work cbreak but I may be doing it incorrectly.
01:37 elricsfate git fetch git@github.com:craigwillis85/terraform-provider-vultr.git 3894645d69f9b06968acfff45d8fb1ce502ba2c3
01:37 cbreak elricsfate: ref
01:37 cbreak not hash
01:38 Koterpillar cbreak: that didn't change anything
01:38 cbreak Koterpillar: removing the remote?
01:38 CheckDavid joined #git
01:38 Koterpillar yeah
01:39 Koterpillar "git remote -v" lists nothing, but "git gc" still complains
01:39 elricsfate Sorry, you said trhe value I received from ls-remote right?
01:39 cbreak elricsfate: no
01:40 cbreak elricsfate: the ref
01:40 cbreak elricsfate: something like "master" or "refs/pull/5"
01:40 elricsfate Ah just read that, thanks
01:40 a3Dman joined #git
01:41 elricsfate git fetch git@github.com:craigwillis85/terraform-provider-vultr.git master worked. Thanks again
01:41 clickjack joined #git
01:41 luke_nukem joined #git
01:42 elricsfate Ok sorry
01:42 elricsfate That actually didn't work
01:42 anuxivm left #git
01:42 elricsfate I saw it perform a fetch but then checking the changed files the changes aren't actually there
01:43 elricsfate I should see the changes on my local repo after the fetch right?
01:44 svm_invictvs joined #git
01:44 clickjack joined #git
01:45 * Rotaerk mutates constant.
01:45 * constant explodes
01:47 * Rotaerk is so destructive..
01:47 jstimm joined #git
01:48 Sasazuka joined #git
01:48 elricsfate Is there something I'm missing? I definitely pointed to the master branch on the remote repo but it's not changing my local repo on fetch
01:48 cbreak elricsfate: no
01:48 cbreak elricsfate: as I said
01:48 cbreak fetch, then merge
01:48 elricsfate Ah gotcha
01:48 elricsfate is there anyway to test the fetch before I actually merge into my repo?
01:48 cbreak fetch will fetch into FETCH_HEAD
01:48 cbreak you can merge that afterwards
01:49 cbreak elricsfate: you can use FETCH_HEAD like any other ref
01:49 cbreak run git log on it, make a local branch from it, what ever
01:49 elricsfate git fetch git@github.com:craigwillis85/terraform-provider-vultr.git master
01:49 elricsfate git merge
01:49 elricsfate like so?
01:49 elricsfate Sorry for multiline
01:51 cdg joined #git
01:51 rominronin joined #git
01:52 cbreak fetch will fetch into FETCH_HEAD
01:52 cbreak you can merge that afterwards
01:52 cbreak elricsfate: you can use FETCH_HEAD like any other ref
01:53 elricsfate Sorry, I'm really not following. I'm guessing you mean I need to do something like git merge FETCH_HEAD?
01:53 cbreak yes
01:53 elricsfate Gotcha, so fetch first then merge said fetch by referencing. TRying that now
01:54 am55_ joined #git
01:55 elricsfate Ah I see, this merge is local until I push lol
01:55 elricsfate Now I can test the merge and then scrap the changes if they are breaking
01:55 elricsfate Thanks for the help!
01:56 dsdeiz joined #git
01:57 robinsmidsrod joined #git
02:00 a3Dman joined #git
02:00 matoro joined #git
02:02 Lunatrius joined #git
02:08 dsdeiz joined #git
02:08 dsdeiz joined #git
02:09 w9qbj left #git
02:11 shinnya joined #git
02:11 star_prone joined #git
02:13 dsdeiz joined #git
02:14 Noldorin joined #git
02:15 finalbeta joined #git
02:18 Zialus joined #git
02:19 cdg joined #git
02:19 pks joined #git
02:21 cdg_ joined #git
02:22 jameser joined #git
02:23 gugah joined #git
02:23 chars6 joined #git
02:25 chars6 is there any way to fake a git push? i'm only pushing to test my deploy script
02:26 d^sh joined #git
02:27 ojacobson chars6: you can run your hooks directly; the environment setup isn't hard
02:28 Koterpillar chars6: push to another remote
02:28 rkazak_ joined #git
02:29 chars6 thanks ojacobson
02:38 Cabanossi joined #git
02:40 jameser_ joined #git
02:49 ilbot3 joined #git
02:49 Topic for #git is now Welcome to #git, the place for git help. Just git help, apparently. | We are aware of the SHA-1 collision. See: http://jk.gs/t/a4 | Public logs at http://jk.gs/glog | First visit? Read: http://jk.gs/git | Current stable version: 2.12.0 | Getting "cannot send to channel"? /msg gitinfo .voice | This is serious business. https://madeitwor.se/git-jokes
02:49 jameser joined #git
02:50 fuzzybear3965 joined #git
02:51 ayogi joined #git
02:51 sunri5e joined #git
02:52 kbs joined #git
02:55 venmx joined #git
02:56 hahuang61 joined #git
02:57 variable joined #git
02:58 fstd_ joined #git
02:59 CEnnis91 joined #git
03:00 variable joined #git
03:00 Goplat joined #git
03:00 cqi joined #git
03:00 andrew710 joined #git
03:02 variable joined #git
03:04 variable joined #git
03:05 hahuang61 joined #git
03:05 variable joined #git
03:06 moei joined #git
03:08 Cabanossi joined #git
03:09 lucido-cl joined #git
03:13 roelmonnens joined #git
03:16 BlackMaria joined #git
03:16 a_thakur joined #git
03:19 matoro joined #git
03:19 govg joined #git
03:21 Vortex34 joined #git
03:21 variable joined #git
03:21 chipotle joined #git
03:22 lagothri1 joined #git
03:22 rominronin joined #git
03:23 TbobbyZ joined #git
03:24 variable joined #git
03:27 Derperperd joined #git
03:29 salamanderrake joined #git
03:29 bsanford joined #git
03:32 chachasmooth joined #git
03:35 ojdo joined #git
03:36 Vampire0_ joined #git
03:36 Lyqyd joined #git
03:37 xaviergmail joined #git
03:38 Cabanossi joined #git
03:39 Noodlewitt joined #git
03:39 govg joined #git
03:40 rchavik joined #git
03:40 xall_ joined #git
03:41 matoro joined #git
03:42 xall joined #git
03:45 duderonomy joined #git
03:46 clemf joined #git
03:47 om_henners joined #git
03:50 jstimm joined #git
03:52 Saturn812 joined #git
03:54 hexagoxel joined #git
03:55 Emperor_Earth joined #git
03:56 DarkPsydeLord joined #git
03:56 eclecticjohny joined #git
03:59 a_thakur joined #git
04:04 mehola joined #git
04:06 jameser_ joined #git
04:07 Cabanossi joined #git
04:07 NeXTSUN joined #git
04:10 jstimm joined #git
04:13 fletom joined #git
04:15 jameser joined #git
04:19 jameser joined #git
04:21 ahrs_ joined #git
04:28 impliednude joined #git
04:28 jameser joined #git
04:32 jameser joined #git
04:32 govg joined #git
04:35 cluelessperson Hey guys, I have two repos,  I would like to fork them into a single repo, in sub directories, keeping their histories.  How can I do that?
04:35 variable joined #git
04:35 Koterpillar We've changed the default branch of a GitHub repository (from 'dev' to 'develop'), and now 'git gc' fails with: fatal: bad object refs/remotes/origin/HEAD error: failed to run repack. How do I fix this?
04:35 cluelessperson [ Repo A, Repo B] => New Repo/{A, B}
04:36 kisonecat joined #git
04:36 gfixler joined #git
04:36 Koterpillar cluelessperson: http://stackoverflow.com/questions/6426247/merge-git-repository-in-subdirectory ?
04:37 cluelessperson Koterpillar: thanks
04:37 Nilesh_ joined #git
04:39 jameser joined #git
04:43 ahrs joined #git
04:44 impliedn joined #git
04:45 jimklo joined #git
04:45 ISmithers joined #git
04:46 bocaneri joined #git
04:47 dsantiag_ joined #git
04:50 dreiss joined #git
04:50 blunaxela joined #git
04:50 jameser joined #git
04:52 nevodka joined #git
04:55 rominronin joined #git
04:57 impliednude joined #git
04:57 TbobbyZ joined #git
04:58 mehola joined #git
04:58 jameser_ joined #git
04:58 roelmonnens joined #git
05:01 Gsham joined #git
05:01 jameser joined #git
05:07 Cabanossi joined #git
05:07 ayogi joined #git
05:08 TbobbyZ joined #git
05:11 venmx joined #git
05:13 moei joined #git
05:16 a_thakur joined #git
05:17 matoro joined #git
05:20 howdoi joined #git
05:22 eroux joined #git
05:23 garruth joined #git
05:29 lucido-cl joined #git
05:29 Murr- joined #git
05:31 xaviergmail joined #git
05:31 mehola joined #git
05:38 Blendify joined #git
05:39 Blendify Can I un commit a local commit?
05:39 garruth joined #git
05:39 _ikke_ Blendify: is it the last commit?
05:39 Blendify yes
05:39 _ikke_ git reset HEAD~
05:40 Blendify this will still keep the modified files?
05:40 gtristan joined #git
05:40 _ikke_ yes
05:40 _ikke_ if you want to keep them staged, you can add --soft
05:40 Blendify thanks
05:41 Blendify I need to stop using git add .
05:41 absinthe joined #git
05:43 Blendify ok oh do I unstage a particular file
05:43 Blendify How*
05:43 garruthc joined #git
05:43 grawity git reset
05:44 cagedwisdom joined #git
05:45 kadoban Blendify: 'git status' actually tells you, in case you forget
05:45 garruth joined #git
05:46 Blendify I must have missed that
05:50 d0nn1e joined #git
05:51 TheMattyMonster joined #git
05:52 Cabanossi joined #git
05:53 unenforcer joined #git
05:54 armyriad joined #git
06:00 mehola joined #git
06:03 fedepad joined #git
06:04 dec0n joined #git
06:11 rscata joined #git
06:13 naf joined #git
06:14 overlord_tm joined #git
06:19 jagob joined #git
06:22 kasual joined #git
06:26 g0d355__ joined #git
06:27 rominronin joined #git
06:34 hexagoxel joined #git
06:36 sarri joined #git
06:36 sarri joined #git
06:37 dersand joined #git
06:39 jungsubk joined #git
06:40 a3Dman joined #git
06:41 roelmonnens joined #git
06:44 moei joined #git
06:46 kexmex joined #git
06:48 jstimm joined #git
06:48 ti2 I'd like to store annotated tag commits on a git server where I may only push branch refs to.  Is there a possibility to prevent tag commits from being purged by 'gc' without destroying the possibility to clone the entire repo?
06:48 ti2 (I tried to link a tag commit as a 'pseudo' file, committed the tree and pushed that; but it breaks 'git clone' on that repo.)
06:50 ThomasLocke_ joined #git
06:50 ThomasLocke_ joined #git
06:51 mda1 joined #git
06:52 ThomasLocke joined #git
06:54 a_thakur joined #git
06:55 Raging_Hog joined #git
06:55 kharloss joined #git
06:55 _ng joined #git
06:56 chachasmooth joined #git
06:56 fedepad joined #git
06:59 Silmarilion joined #git
06:59 babis joined #git
07:01 freimatz joined #git
07:01 navidr joined #git
07:04 rchavik joined #git
07:05 grawity there's no such thing as "tag commits", only commit objects and tag objects
07:05 babis joined #git
07:05 grawity you should be able to point a branch ref to the tag though
07:08 ti2 No; if I'm doing something like 'git update-ref refs/heads/my/tag refs/tags/mytag' I get an error message:
07:08 ti2 ...trying to write non-commit object ... to branch 'refs/heads/my/tag'
07:09 TbobbyZ joined #git
07:10 Vortex34 joined #git
07:10 ti2 To correct my mistakenly wording: I created a tree (with 'git update-index --add --cacheinfo ...' and 'git write-tree') with a "file" pointing to a tag object; but that does not work well when attempting to clone the repository.
07:10 babis joined #git
07:10 grawity hmm I suppose it got more sanity checks
07:11 grawity are you specifically limited to branches, or only "not tags"?
07:11 ti2 I may only push branch refs; pushing tags (or other things like 'refs/spam') uploads the objects but reference updating fails.
07:11 senaps joined #git
07:12 ankit01ojha joined #git
07:12 venmx joined #git
07:12 ti2 Thus, actually I just need a way to prevent the uploaded tag objects to not be purged from 'git gc'
07:12 ti2 (s/not //)
07:13 grawity how about refs/foo/spam?
07:13 grawity i.e. with at least 3 path components
07:13 Learath2 joined #git
07:14 imack joined #git
07:14 ti2 No, only refs/heads/...
07:15 grawity you're probably out of luck then
07:15 grawity dump `git cat-file tag foo` into regular files, and when need arises re-add it with `git hash-object -w -t tag`
07:17 ti2 That is at least a straight-forward fallback ... ;-)  Thanks a lot!
07:17 athan joined #git
07:17 athan hey all, I'm in a bit of a pickle
07:17 kharloss joined #git
07:17 athan how do I reset to HEAD, and ignore any untracked changes I've made?
07:17 roelmonnens joined #git
07:18 grawity ignore as in, throw them away, or keep them?
07:18 athan I've done some permissions... uh... destruction that I'd like to just throw away
07:18 athan grawity: just toss em
07:18 grawity git checkout -f
07:18 grawity or, git reset --hard
07:18 athan I've tried `git reset HEAD --hard` and `git checkout --force master`
07:18 ISmithers joined #git
07:18 athan and they won't throw them away
07:18 roelmonnens joined #git
07:18 athan they just keep the untracked changes
07:19 grawity git doesn't track permissions beyond a+x so it doesn't consider them as 'changes'
07:19 athan oh god
07:19 athan whyyyyyy
07:19 athan shit
07:19 athan even then `git status` purs a lot of red
07:19 kharloss joined #git
07:19 athan pours*
07:20 grawity also, "untracked" files aren't ... well, aren't tracked – they're not in HEAD by definition
07:20 grawity they're not part of the repo at all
07:20 athan hmm
07:20 athan so if I add them, then reset to the previous commit...
07:20 grawity does it show "untracked files" or "unstaged changes"?
07:20 athan it might work?
07:20 grawity different things
07:20 athan untracked files
07:20 JeroenT joined #git
07:21 grawity the only thing Git can do with those is delete them
07:21 athan with git clean?
07:21 _ikke_ yes
07:21 grawity yeah, `git clean -ndx` if you want
07:21 grawity -n is the pre-check, -f actually deletes them
07:21 grawity since that's the one place where Git history wouldn't protect against mistakes :D
07:22 grawity hmm maybe without -x is better? habit.
07:23 athan heh, thank you grawity I was pulling my hare out (for magic tricks)
07:24 rokups joined #git
07:24 t0by joined #git
07:24 t0by joined #git
07:25 JeroenT joined #git
07:26 aidalgol joined #git
07:28 svm_invictvs joined #git
07:28 qt-x joined #git
07:30 rominronin joined #git
07:30 kharloss_ joined #git
07:31 masuberu joined #git
07:32 jnavila joined #git
07:33 diogenese joined #git
07:34 hashtagstrashtag joined #git
07:34 kharloss joined #git
07:36 JeroenT joined #git
07:37 imack joined #git
07:40 fedepad joined #git
07:41 chele joined #git
07:44 manuelschneid3r joined #git
07:48 absinthe joined #git
07:50 sz0 joined #git
07:51 srandon1213 joined #git
07:51 miczac joined #git
07:51 srandon1213 hello all, is it possible to use github API to create my own program which does things on the command line  ?
07:52 srandon1213 or i violate some kind of agreement/license  ?
07:52 Cabanossi joined #git
07:52 absinthe joined #git
07:53 grawity well, that's the whole point of an API, isn't it?
07:53 grawity https://hub.github.com/ http://seveas.github.io/git-spindle/github.html
07:53 djb-irc joined #git
07:55 TomyWork joined #git
07:56 moritz but you should read github's Terms of Service before using it
07:56 moritz independently of whether you use it through the web interface or the API
08:01 Brski joined #git
08:01 Raging_Hog joined #git
08:01 star_prone joined #git
08:02 kexmex joined #git
08:04 variable joined #git
08:04 variable how can I get git rebase to act like git read-tree --aggressive ?
08:04 variable in particular I want this behavior
08:05 kharloss joined #git
08:05 variable o   when one side removes a path and the other side leaves the path unmodified. The resolution is to remove that path.
08:06 star_prone joined #git
08:11 heftig variable: that should already be the case
08:11 Learath2 joined #git
08:13 j08nY joined #git
08:15 xaviergmail joined #git
08:17 elect joined #git
08:19 dvaske joined #git
08:20 tkcrib joined #git
08:22 Cabanossi joined #git
08:23 finalbeta joined #git
08:24 SantaClauze joined #git
08:24 moL-Ou joined #git
08:25 PHPanos joined #git
08:25 xall joined #git
08:26 cqi joined #git
08:28 theoceaniscool joined #git
08:28 zefferno joined #git
08:28 acetakwas joined #git
08:29 adac joined #git
08:30 babis joined #git
08:30 moL-Ou joined #git
08:30 manuelschneid3r joined #git
08:31 athan left #git
08:31 marianina8 joined #git
08:32 kent\n joined #git
08:32 overlord_tm joined #git
08:36 RalphJr45 joined #git
08:40 netj joined #git
08:43 _ADN_ joined #git
08:44 valize joined #git
08:45 flaviodesousa joined #git
08:46 Balliad joined #git
08:48 sarri joined #git
08:48 sarri joined #git
08:49 gtristan joined #git
08:51 kurkale6ka joined #git
08:52 penanze joined #git
08:56 TomyWork joined #git
08:58 sbasso joined #git
09:04 darkbit joined #git
09:04 zerow joined #git
09:07 Andrew_K joined #git
09:09 troulouliou_div2 joined #git
09:09 jstimm joined #git
09:11 ph88_ joined #git
09:11 ph88_ when i do git checkout githash  does that change the files to that point in time ?
09:12 selckin yes
09:12 JeroenT_ joined #git
09:12 venmx joined #git
09:14 Dasua joined #git
09:16 ExpiredDev joined #git
09:16 a_thakur joined #git
09:16 CaptainHeavy joined #git
09:16 a_thakur joined #git
09:18 _ADN_ joined #git
09:19 robattila256 joined #git
09:20 netj joined #git
09:25 vamiry_ joined #git
09:26 Lunatrius` joined #git
09:26 zxd joined #git
09:26 Hez joined #git
09:26 zxd hi
09:27 zxd does 'git log' sort the commits by date?
09:27 selckin no
09:27 zxd how does it sort?
09:27 grawity Commit Ordering
09:27 grawity By default, the commits are shown in reverse chronological order.
09:27 grawity see `man git log` for options
09:27 gitinfo the git-log manpage is available at http://jk.gs/git-log.html
09:27 zxd if I have a 'Merge branch' commit
09:28 grawity with merges you'll want --graph though
09:28 jast by default, the ordering by date doesn't stop at merges, so your merged commits will be interspersed with the other commits
09:28 jast --graph is pretty much the only sane alternative if you ask me
09:29 zxd so it does order by date by default
09:29 jast yes
09:29 jast the confusion likely stems from --date-order which does something slightly different
09:30 jast or at least the manpage seems to imply that. I haven't actually looked at what the code does.
09:31 selckin well using the date to pick when to show items from 2 streams that are not ordered by date
09:31 jast yeah, I just verified that. the output of 'git log' and 'git log --date-order' can be different.
09:31 selckin right?
09:33 jast I do wonder about the difference, though... truly ordering by timestamp only would be quite inefficient because you have to walk the full graph before outputting anything, and it clearly doesn't actually do that
09:35 Junior joined #git
09:37 jameser joined #git
09:40 m0rdonez joined #git
09:40 ph88_ left #git
09:41 hexagoxel joined #git
09:41 JeroenT joined #git
09:42 whaley joined #git
09:46 vamiry joined #git
09:47 ShapeShifter499 joined #git
09:47 chll joined #git
09:48 utanapischti joined #git
09:49 m0rdonez joined #git
09:49 ferr1 joined #git
09:50 goTAN joined #git
09:50 nivag joined #git
09:51 ExpiredDev joined #git
09:51 jameser joined #git
09:52 marcogmonteiro joined #git
09:53 ExpiredDev Hi everyone! Does anyone know any resource regarding executing update hooks on bitbucket server? In particular, I am trying to find out how to tag merged branch. Any ideas?
09:54 Snugglebash joined #git
09:55 BenFarley joined #git
09:57 vamiry_ joined #git
10:00 borkr joined #git
10:01 pluszak left #git
10:01 doublebit joined #git
10:01 evie_hammond joined #git
10:02 kpease joined #git
10:04 kexmex joined #git
10:05 moritz ExpiredDev: I think bitbucket allows webhooks
10:09 Murii joined #git
10:10 aard_ joined #git
10:11 rchavik joined #git
10:13 cdown joined #git
10:14 TbobbyZ joined #git
10:14 ExpiredDev @moritz, yes, it does, but it seems it's more complicated than just adding script on the git server, isn't it?
10:16 P4nther joined #git
10:17 moritz ExpiredDev: you have to have a HTTP server that accepts the requests that bitbucket sends you
10:17 moritz ExpiredDev: and then does the actual work
10:18 BackEndCoder joined #git
10:19 cdown joined #git
10:22 marcogmonteiro joined #git
10:22 lb1c joined #git
10:23 Lunatrius joined #git
10:23 cdown_ joined #git
10:25 Snugglebash joined #git
10:26 dave0x6d joined #git
10:29 irqq joined #git
10:29 sbasso joined #git
10:30 achadwick joined #git
10:32 rossome joined #git
10:33 marianina8 joined #git
10:35 fedepad joined #git
10:35 cdown_ joined #git
10:42 clmsy joined #git
10:43 goTAN joined #git
10:43 timwis hey folks, i'm rewriting my app in a different framework, and there won't really be any shared code between versions -- to start that branch, should i create a regular branch and delete everything, or create an orphan branch?
10:44 goTAN_ joined #git
10:44 acetakwas joined #git
10:45 llamapixel Is there a way I can see the total size of the data I am adding when commiting or adding?
10:45 llamapixel git add .
10:45 llamapixel git commit -m “commit message here”
10:46 llamapixel I see files changed / insertions and deletions
10:46 llamapixel eg :  23 files changed, 68 insertions(+), 38 deletions(-)
10:46 boshhead joined #git
10:46 llamapixel compare the disk sizes between the work load?
10:47 canton7 llamapixel, "size" in what sense? Is this size of the diff, size of the modified files, size after delta-compression...?
10:47 canton7 size in bytes? characters? lines?
10:47 JeroenT joined #git
10:47 Snugglebash joined #git
10:47 llamapixel size of the commited data / modified files not worrying about removals
10:48 King_Hual joined #git
10:48 madprops joined #git
10:48 canton7 "committed data" is ambiguous. What exactly do you mean by that? :P
10:48 llamapixel Ok, how much traffic is being uploaded.
10:48 askb joined #git
10:48 canton7 oh, that's a completely different question
10:48 canton7 when you commit, no traffic is uploaded anywhere
10:48 tvw joined #git
10:49 llamapixel when I push then
10:49 canton7 when you push, git negotiates with the repo you're pushing to, and figures out the minimum amount of data it can get away with sending, such that the remote repo can still figure out what was pushed
10:49 canton7 so if you push to an empty repo, that's completely different to pushing to a repo which has the previous commit
10:49 llamapixel Is there a measure of how much it pushes in a typical already started repo?
10:50 llamapixel Is it possible to display the pushed amount of traffic data?
10:50 mkoskar joined #git
10:51 canton7 not that I know of. What is sent is compressed as well, so you'll have to guess the output of the compression algorithm
10:51 llamapixel ok thanks mate.
10:51 tobiasvl you can look at the packfile http://stackoverflow.com/questions/5013894/finding-the-size-of-a-push
10:51 llamapixel ah thansk tobiasvl
10:51 tobiasvl it's not perfect
10:51 canton7 when you push, doesn't it give you the size of the push?
10:52 canton7 "Writing objects: 100% (12/12), xxx KiB, done."
10:52 llamapixel It tells me objects and speed
10:52 sbasso joined #git
10:52 llamapixel ah i see
10:52 tobiasvl yeah, but I thought llamapixel wanted to know it before pushing ("when committing")
10:52 llamapixel yes
10:52 canton7 yeah, I was responding to "<llamapixel> Is it possible to display the pushed amount of traffic data?"
10:52 llamapixel I can live with the after but would like to know the before.
10:52 canton7 "pushed" being in the past tense there
10:52 tobiasvl yeah that's true
10:53 llamapixel Thankyou very much guys, I didn’t even have my question straight.
10:53 Levex joined #git
10:54 canton7 tobiasvl's link looks decent, although it is of course an approximation
10:54 llamapixel tobiasvl: I will try that link command next pre -push
10:54 YuGiOhJCJ joined #git
10:55 cdown joined #git
10:55 masuberu joined #git
10:59 llamapixel I can actually try it now with my GIT testing folder and making fake files with . pwgen 16 100 > file.txt
11:00 llamapixel Thanks, that gives me enough content just before the push… time to crash out.
11:01 waveclaw joined #git
11:01 pl Hi. I'm maintaining a "long term fork" of a repository, where I have an upstream branch from which I cherry-pick the commits I need. Sometimes it's difficult to see what's new in the upstream branch since the last time I cherry-picked. Is there a way to list the new commits on a branch since the latest commits the two branches have in common?
11:02 pl *since the last commit they have in common
11:02 acetakwas joined #git
11:02 moritz that won't help you much, since cherry-picking changes the identity of commits
11:03 moritz so such an command will show all the cherry-picked commits as new
11:03 pl then no, it won't help.
11:03 pl any suggestion? maybe on a different workflow?
11:03 m0rdonez joined #git
11:06 bremner with sufficient scripting you can identify patches whose patch-id is not present in your current branch
11:06 osse pl: you can use the git cherry command to weed out commits you've cherry-picked already
11:07 Cabanossi joined #git
11:08 pl Still this cherry-pick based workflow feels clumsy. But I don't think there's an alternative, as upstream and the fork won't merge to a common state at any point
11:08 tcorneli joined #git
11:08 osse you can merge from the upstream into the fork
11:10 Tobbi joined #git
11:11 borkr joined #git
11:11 theoceaniscool joined #git
11:13 eahmedshendy joined #git
11:13 sbasso joined #git
11:13 venmx joined #git
11:13 pl osse, that's probably the best solution
11:14 qt-x joined #git
11:14 adymitruk joined #git
11:14 jungsubk joined #git
11:15 bttf joined #git
11:15 tmg joined #git
11:16 NullableTruth joined #git
11:17 pl osse, maybe what I would like is not a "merge" but an "advance one commit"
11:17 pl in order to be able to have more control on what is merged
11:17 pl is there anything like that?
11:18 cdown_ joined #git
11:19 osse cherry-pick...
11:19 bremner you could rebase -i the upstream onto your branch before merging, if there's <= around 50 new commits
11:19 bremner and then throw away that rebased branch
11:24 acetakwas joined #git
11:25 jameser joined #git
11:27 jeffreylevesque_ joined #git
11:30 xaviergmail joined #git
11:34 marianina8 joined #git
11:34 TbobbyZ joined #git
11:34 miha_S7 joined #git
11:34 mda1 joined #git
11:35 a3Dman joined #git
11:35 dsdeiz joined #git
11:39 nyuszika7h joined #git
11:41 jstimm joined #git
11:41 Shakeel_ joined #git
11:47 c0ded joined #git
11:48 rnsanchez joined #git
11:48 venmx joined #git
11:52 JeroenT joined #git
11:54 finalbeta joined #git
11:55 a_thakur joined #git
12:00 dviola joined #git
12:01 raijin joined #git
12:01 _Cyclone_ joined #git
12:02 doublebit joined #git
12:03 w9qbj joined #git
12:03 ankit01ojha joined #git
12:03 ams__ joined #git
12:03 doublebit joined #git
12:04 roelmonn_ joined #git
12:06 roelmon__ joined #git
12:07 Cabanossi joined #git
12:09 finalbeta joined #git
12:10 Snugglebash joined #git
12:11 Snugglebash joined #git
12:11 publio joined #git
12:12 jameser joined #git
12:12 marcogmonteiro joined #git
12:13 EV3RYDAYPR0GRESS joined #git
12:14 ti2 left #git
12:14 fstd joined #git
12:18 aruns joined #git
12:18 aruns Hi, running Git version 2.10.1 on OS X 10.11.6 (El Capitan).
12:18 aruns Does anyone know if it's possible to get the total file size of a local commit object?
12:19 V`ger joined #git
12:19 miczac joined #git
12:20 aruns Ah, I can do git cat-file -s *commit hash*
12:20 aruns But that just returns a number and doesn't indicate byte unit.
12:20 aruns I have committed about around 48 PDFs.
12:21 aruns I am guessing it would be in MB most likely?
12:22 _ikke_ aruns: I would expect bytws
12:22 _ikke_ bytes
12:22 _ikke_ correct, it's in bytes
12:23 aruns _ikke_: Hmm, well it's 46 PDFs in total
12:23 _ikke_ aruns: You should provide the blob hash
12:23 Tobbi joined #git
12:23 _ikke_ the commit size is the size of the commit object, not all files referenced by the commit
12:23 aruns Ah I see.
12:24 aruns Whereas the blob is the combined file size of all the files in the commit.
12:26 _ikke_ No, each file is its own blob
12:26 ebcz joined #git
12:26 Kasreyn joined #git
12:27 rossome joined #git
12:28 idmistir joined #git
12:28 idmistir joined #git
12:28 aruns Ah OK
12:29 roelmonnens joined #git
12:29 netj joined #git
12:29 acetakwas joined #git
12:29 Kasreyn say i just started a big compile and fetched a branch that doesn't exist locally (it's not possible to do git log branchname), how can i make it available without first checking it out (since there is a compile on the current content) ?
12:31 MarioBranco joined #git
12:31 _ikke_ Kasreyn: git log origin/<branchname>
12:32 Kasreyn _ikke_: ok yeah but that's not the point.. i want to push (while the compile is running) some notes that never got pushed for that branch
12:33 koneko joined #git
12:33 Sound joined #git
12:34 _ikke_ if git log branch is already not working, how can there be something that you haven't pushed yet?
12:35 Kasreyn _ikke_: rebased branch..
12:36 TbobbyZ joined #git
12:36 MarioBranco joined #git
12:37 Rodya_ joined #git
12:38 Kasreyn what i mean is that the branch was pushed (notes not included) and then rebased somewhere else
12:39 porsche944 joined #git
12:39 Kasreyn (unfortunate.. i know!)
12:39 finalbeta joined #git
12:40 _ikke_ How would you normally push the notes?
12:41 Kasreyn _ikke_: by adding "push  = +refs/notes/*:refs/notes/*" to .git/config
12:41 darkbit joined #git
12:41 _ikke_ Kasreyn: Then you can just use git push origin refs/notes/*:refs/notes/*
12:41 _ikke_ No need to create a branch for that
12:42 _ikke_ (you might need to escape the *)
12:42 Kasreyn except .config is not always in sync working across multiple computers so in this case notes never get pushed
12:44 ThomasLocke joined #git
12:44 gugah joined #git
12:44 inflames joined #git
12:44 Kasreyn _ikke_: ok! thanks :)
12:45 albel727 joined #git
12:45 MarioBranco joined #git
12:45 brent__ joined #git
12:46 Kasreyn _ikke_: uhm no.. "error: failed to push some refs to.." "hint: Updates were rejected because the remote contains work that you do not have locally"
12:46 nyuszika7h joined #git
12:47 _ikke_ Kasreyn: forgot the + before
12:47 _ikke_ +refs/notes/\*:refs/notes/\*
12:50 Kasreyn _ikke_: thanks.. my computer is slow.. i'll tell you if it worked ;)
12:50 JeroenT_ joined #git
12:50 vifon joined #git
12:51 bn_ joined #git
12:52 vifon Hi. I've got a silly sounding question. How can I get the list of commits being part of a feature branch that was already merged? "git log master..HEAD" won't work because master already contains my commits after the merge. Sounds simple but every solution I can think of quickly gets pretty complex.
12:52 Stutters joined #git
12:52 dandaman joined #git
12:52 Cabanossi joined #git
12:54 Kasreyn _ikke_: it works :)
12:54 moritz vifon: if the branch wasn't deleted, git log branchname helps
12:55 Dougie187 joined #git
12:55 moritz vifon: if it was deleted, you have to search for the merge commit, and explore from there on. A graphical git client makes that easier
12:55 osse vifon: git log x..HEAD where x is the merge commit~1
12:55 eroussel joined #git
12:55 vifon But it still shows the commits from "below" the fork point.
12:55 vifon I would like to cut it off.
12:56 bn_ hi, is there an easy way to backport (the most recent version of) a file?
12:56 vifon osse: Any way to easily find merge commit~1?
12:56 osse vifon: you find the merge commit and then you write ~1 behind it
12:56 vifon -_-
12:57 osse ti was only partly a joke
12:57 osse oooooooooh
12:57 osse wait a minute. I think I once wrote a script for exactly this
12:57 vifon Ok then, how to get the child commit of my commit, which would be the merge commit.
12:57 livingstn joined #git
12:57 vifon Possibly without a fruit salad of rev-list arguments.
12:57 JeroenT joined #git
12:58 osse it's not possible. you can only look for commits who have this commit as parent
12:58 osse i've already written what you need. I just need to find it.
12:59 vifon Is that really so non-trivial? That's basically what Github shows under any PR. My gut feeling tells me we're missing something.
12:59 osse what I wrote isn't trivial, but it's not hard either
12:59 amrits joined #git
13:00 j08nY joined #git
13:01 osse vifon: try this: http://sprunge.us/JJcX
13:01 leeN joined #git
13:01 osse shit
13:01 bn_ I'm guessing I'd need to go to HEAD and stage the file in question and do a non-hard reset to the older version to preserve the file during the reset? or stash it?
13:01 osse vifon: I mean this: http://sprunge.us/VGjK
13:02 osse bn_: what do you mean by backport?
13:03 bn_ osse: recommit a newer version of a file w/ no changes to an older version
13:03 CEnnis91 joined #git
13:03 bn_ s/older version/older release/
13:04 osse bn_: git checkout somecommit -- somefile && git commit -m 'Bring back somefile from the good old days of somecommit'
13:04 xaa joined #git
13:04 vifon osse: I get "oops" on my branch though.
13:04 a3Dman joined #git
13:04 osse how do you invoke it?
13:05 vifon ./VGjK HEAD
13:05 osse and master is the branch it's merged in ?
13:05 bn_ osse:  uh.. you mean from the new days? :)
13:05 nettoweb joined #git
13:05 vifon Yes.
13:05 osse bn_: sure, whatever day somecommit comes from
13:06 osse vifon: does 'git branch --contains HEAD' print masteR ?
13:06 adac joined #git
13:06 Snugglebash joined #git
13:06 vifon Yes.
13:06 osse oh I see a typo
13:07 osse vifon: http://sprunge.us/XHWP
13:08 vifon Still oops.
13:09 MarioBranco joined #git
13:09 osse new typo: http://sprunge.us/gXgJ
13:09 bonsairoot joined #git
13:09 osse the script assumed master was checked out...
13:10 * osse flogs self
13:11 bn_ osse: ok, thanks.  just wanted to confirm this would take a newer commit (file) to an older release of the branch (I'm guessing I probably should create a new (topic) branch once I revert to older release and before I do the checkout)
13:12 Cavallari joined #git
13:13 vifon osse: Still oops. :D
13:13 osse igive up
13:13 vifon :D
13:13 osse the point is: you can use git log to find the correct commit
13:13 cdown joined #git
13:13 vifon I'll experiment myself.
13:13 osse for some reason it doesn't work here
13:14 osse is the repo public ?
13:14 vifon No.
13:14 jast I don't actually understand what output you're looking for
13:15 osse vifon: can you show me what happens if you do bash -x ./script HEAD
13:15 jast from your example, I'm assuming you have the feature branch checked out?
13:15 osse vifon: it will only print hashes anyway
13:15 xgen joined #git
13:16 fedepad_ joined #git
13:16 vifon osse: https://vifon.einval.eu/upload/2017-03-14-141611_584x144.png (sorry for using a screenshot)
13:17 bonsairoot Hi. How would I transform such an entry in .hgignore to .gitignore glob syntax: ^folder1/problem[0-9]*/.*.-test/\.obj/
13:18 osse vifon: fuck...
13:18 kharloss joined #git
13:18 osse vifon: the first typo wasn't actually a typo. the second one was, but now the first typo ruins it
13:18 pl left #git
13:19 osse vifon: change $2 to $3 in the awk code
13:19 vifon In which version?
13:19 osse the newest
13:19 vifon Works!
13:19 jast bonsairoot: 'an arbitrary sequence of digits' can't be expressed in gitignore patterns
13:20 vifon osse: Thanks a lot. I'll save it as ~/dl/git-wtf and analyze later. :)
13:20 osse vifon: as you might have guessed, I wrote this so it works regardless of whether it's merged or not.
13:20 vifon Yes, I noticed it already.
13:20 Brski_ joined #git
13:20 osse vifon: should probably modify it to assume HEAD is the feature branch, master the base branch, but so that you can override both.
13:20 osse currently you have to write HEAD
13:20 osse or branchname
13:20 vifon It still blows my mind it cannot be done with some clever git-log flag.
13:20 vifon osse: Yep. That was my plan.
13:21 osse vifon: there is no way a commit can know its children. it's not doable
13:21 osse github probably caches it
13:21 vifon Commit? No. A git tool? Why not...
13:21 hexagoxel joined #git
13:21 vifon Reverse ancestry etc.
13:22 bonsairoot jast, hmm ok ... could I do:  **/*-test/.obj/
13:22 osse vifon: sure, you could write something that maintains a database
13:22 jast bonsairoot: that should work, yes
13:22 bonsairoot jast, thanks
13:22 a_thakur joined #git
13:22 osse OR: you can find the commit you want by looking at the commits' parents, by using some clever git-log flag :P
13:23 jast what information are we hunting for, here?
13:23 vifon Anyway, I don't mean getting a child per se. I mean getting a diff of a branch without the "parent branch" (if I may call it so).
13:24 osse jast: show the commits merged in from a given feature branch
13:24 jast commits up to an earlier merge base?
13:24 osse vifon: you want the diff?? I had to change it from diff to log ...
13:24 vifon Diff, commit list... Both, I guess. I depends.
13:25 osse in any case you have to determine where the branch started
13:25 osse ( in the human sense of the word )
13:25 jast well, that would include all commits that existed in the main branch at the time the feature branch was created, I guess
13:25 vifon I already changed your script to use "$@" instead of "diff".
13:25 vifon (or log, depending on a version)
13:26 vifon jast: That's the main problem. I want to cut them off from the output.
13:26 jast that's basically the intersection of two things, isn't it
13:26 jast yeah, but there is no record in Git of when the branch started
13:26 osse jast: basically what we want is the commits in:  $merge^1..$merge^2
13:26 vifon Yep.
13:26 osse my script finds that
13:27 vifon What the actual fsck... http://stackoverflow.com/a/30998048
13:28 jast so, find the most recent merge that brings in commits that exist on the feature branch?
13:29 zxd how to print files added/removed/modified in each commit   , 'git log' shows only comments
13:29 grawity --stat
13:29 jast stat generates diffstats only, probably --name-status is closer to what you want
13:29 OnceMe joined #git
13:30 jast or just 'git whatchanged' which is a little more verbose but easier to type ;)
13:31 bonsairoot jast, is there a way to *hello.py but there has to be something in front of hello ... or do I have to make both entries *hello.py and !hello.py
13:31 TbobbyZ joined #git
13:31 jast bonsairoot: ?*hello.py
13:31 bonsairoot jast, nice thanks
13:32 jast good old trick from back when you had to use globs for everything :)
13:32 jast also from Solr queries I've had to write...
13:33 sbasso joined #git
13:35 martinhez joined #git
13:35 dandaman joined #git
13:36 ExoUNX joined #git
13:41 OnceMe when I type git status
13:41 OnceMe why do I get modified:   bootstrap/cache/.gitignore ?
13:41 OnceMe I havent modified it, and I initially pushed all those files
13:41 redhedded1 joined #git
13:42 osse OnceMe: what does git diff say ?
13:42 livingstn joined #git
13:42 OnceMe diff --git a/bootstrap/cache/.gitignore b/bootstrap/cache/.gitignore
13:42 OnceMe old mode 100755
13:42 OnceMe new mode 100644
13:42 OnceMe it must be due to my rsync
13:42 OnceMe I have rsync between vagrant and my host machine
13:43 OnceMe perms are different, but that shouldnt bother git
13:44 qqx OnceMe: The executable bit is changed, and git does consider that to be a modification.
13:45 OnceMe qqx: I understand
13:45 OnceMe but how to fix this?
13:45 jast Git does track whether a file is executable, but you're right that the rest of the permissions don't matter to git
13:45 jast well, honestly the new permissions are probably more correct
13:45 OnceMe because I have /var/www/laravel synced from local to vagrant
13:45 jast if you want to keep it the way it is, mark the file as executable
13:46 Snugglebash joined #git
13:46 OnceMe how?
13:46 jast and if both sides have proper filesystems (i.e. not something like Windows's NTFS), I think your syncing is doing something wrong :)
13:47 jast chmod +x .gitignore
13:47 h12o joined #git
13:47 OnceMe its nfs
13:47 OnceMe on vagrant
13:47 OnceMe on host is ext4
13:47 jast NFS is a network protocol, not a physical filesystem
13:47 RichiH "< jast> chmod +x .gitignore"
13:48 * RichiH weeps about not being able to change /topic any more
13:48 jast and it should be quite capable of keeping track of executable bits, so this still shouldn't be happening
13:48 OnceMe I have a bit of .gitignores http://pastebin.com/NsDQ8eFx
13:48 jast RichiH: what, are you trying to preserve this monstrosity? :P
13:48 RichiH yes.
13:48 OnceMe config.vm.synced_folder "/var/www/laravel", "/var/www/laravel", :mount_options => ["dmode=777","fmode=666"]
13:49 OnceMe thats how I sync between host and guest
13:49 enginpost joined #git
13:49 jast I guess we can preserve it in the void at least
13:49 jast .random add < jast> chmod +x .gitignore
13:49 gitinfo jast: Okay.
13:49 manuelschneid3r joined #git
13:49 bannakaffalatta joined #git
13:50 jast OnceMe: well, those options explicitly ask the sync mechanism to screw up the permissions
13:50 jast for some values of "screw up", I guess
13:50 jast it will strip executable bits from all synced files
13:50 jast though that doesn't explain why .gitignore had one in the first place
13:50 Ownix Stupid question, im using Git Extensions, when checking out a remote branch it gives me the option to "Create a local branch with the name: xyz" and "checkout remote branch". What would the later do?
13:51 Ownix How can you checkout a remote branch without having a local copy of it?
13:51 jast Ownix: Git actually keeps local mirrors of the remote branches
13:51 enginpost I want to overwrite my remote origin with my local branch and basically start over
13:51 jast they're meant purely as mirrors, so you can't normally edit them
13:51 enginpost how would I do that?
13:51 Ownix jast: even if I have never checked out?
13:51 IntruderSRB joined #git
13:51 jast and if you check them out, you'll get a !detached HEAD
13:51 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
13:51 jast yes, mirrors get updated whenever you 'fetch' (or 'pull' without arguments)
13:52 jast see 'git branch -r' for a list
13:52 jast enginpost: something involving 'git push --force'
13:52 jast the ultimate tool for destroying remote history
13:52 OnceMe so what to do?
13:53 Ownix So jast, youre saying if I chose the last option ill get a detached head?  https://usercontent.irccloud-cdn.com/file/YZFpFVIN/
13:54 enginpost jast: would you do that from the branch locally?
13:54 jast Ownix: I think so. I've never used Git Extensions but it's the only thing I can see it doing
13:54 jast enginpost: yeah. basically it will force the remote to have the same history in the branch as locally.
13:55 jast if you have an old Git it will even overwrite all the other branches, too, in one fell swoop
13:55 jast (pre 2.0)
13:55 jast or if you have customized push's behaviour in your config
13:55 Ownix So the only reason I would do that is if I wanted to test someones changes for a PR but I didnt want the files on my system because clutter?
13:56 jast Ownix: the difference in how much data it takes up is negligible. Git already has all the data retrieved at that point. the only reason to detach HEAD is if you want to have a quick look without cluttering your list of local branches.
13:56 OnceMe jast: so should I remove mount perms sync?
13:57 enginpost jast: that worked ... thanks
13:58 jast OnceMe: you can try. there's a chance you won't be able to use the files anymore if the user IDs don't map, which is probably why it's configured that way
13:58 Raziel_17743 joined #git
13:58 OnceMe jast: no
13:58 jast other alternatives include telling Git to ignore changes to permissions
13:58 OnceMe I configured it that way because I had issue with perms in storage
13:59 Ownix "the only reason to detach HEAD is if you want to have a quick look without cluttering your list of local branches." my list is horrid :(
14:00 OnceMe jast: ok this worked
14:00 OnceMe now I only get this two files:
14:00 OnceMe Uncommitted changes in working copy:
14:00 OnceMe storage/oauth-private.key
14:00 OnceMe storage/oauth-public.key
14:01 OnceMe ok fixed that too
14:01 OnceMe indeed it was perms issue
14:01 Gsham joined #git
14:02 hasc joined #git
14:08 OnceMe if master is branch, what is remote origin?
14:10 jstimm joined #git
14:11 jast I'm not entirely sure what you mean
14:11 Levex joined #git
14:11 ayogi joined #git
14:12 LeBlaaanc joined #git
14:12 V`ger left #git
14:13 TheMattyMonster joined #git
14:13 bannakaf_ joined #git
14:15 Rotaerk OnceMe, sounds like you need to do some reading about git; branches and remotes are different things
14:16 RichiH jast: good, good
14:16 hasc joined #git
14:16 RichiH future generations will drink from your knowledge
14:17 gtristan joined #git
14:17 Brski__ joined #git
14:18 OnceMe any good link?
14:19 cdg joined #git
14:20 JeroenT joined #git
14:20 MarioBranco joined #git
14:21 bitmod joined #git
14:21 kharloss joined #git
14:21 MarioBranco joined #git
14:24 DarkPsydeLord joined #git
14:25 manithree joined #git
14:25 jstimm joined #git
14:26 joelhed joined #git
14:27 joelhed is it possible to get the remotes from a remote?
14:28 _ikke_ You mean the remotes that a remote repository has?
14:28 joelhed exactly! I can
14:30 Squarism joined #git
14:30 Squarism i have a tendency not pushing branches i work with
14:30 Squarism anyone know some software / trick to avoid such
14:30 Squarism misstakes
14:31 Rusky_ joined #git
14:32 joelhed I can't push changes directly to the local repo, since it has unstaged changes, so I thought I could push directly to the "original" origin
14:32 canton7 Squarism, there's a built-in thingy to put the current status of your repo in your command prompt. so it will show "(master u=)" if you're on the master branch and up to date, "(master u+1)" if you've got an unpushed commit, etc
14:32 mikecmpbll joined #git
14:32 _ikke_ joelhed: there is no way in git to get this information
14:32 jast joelhed: not really. you can see which remote branches the remote repository has mirrored from elsewhere, though, if any: git ls-remote origin (they'll show up as refs/remotes/foo/bar)
14:32 cdg joined #git
14:32 jast you can't get at the original remote URLs or anything
14:32 canton7 Squarism, I find that to be a helpful reminder that I've got uncommitted/unpushed stuff
14:32 _ikke_ Right, but that does not tell you where that remote actually lives
14:32 jagob joined #git
14:33 Squarism canton7, sounds like a good idea!
14:33 fstd joined #git
14:33 ij canton7, Where is that "builtin thingy"?
14:33 canton7 it's the __git_ps1 bash function. can't remember where it's defined
14:33 joelhed Alright, thanks!
14:33 Rusky_ joined #git
14:33 jast git:: HEAD:contrib/completion/
14:33 gitinfo Git web link: https://github.com/git/git/tree/HEAD/contrib/completion/
14:33 canton7 it's in git-prompt.sh, apparently
14:34 canton7 aha
14:34 canton7 Squarism, !ps1
14:34 gitinfo Squarism: [!prompt] The bash completion script developed with git (http://en.newinstance.it/2010/05/23/git-autocompletion-and-enhanced-bash-prompt/) and the one for zsh shipped with zsh (http://gitready.com/advanced/2009/01/28/zsh-git-status.html) both support showing the status of your repo in the prompt string, at a glance
14:34 jast finally this feature has a syntax I can actually remember
14:34 fstd joined #git
14:34 Squarism canton7, thanks again
14:35 jast hum, those links look slightly outdated...
14:35 Perspicaciosity joined #git
14:36 nowhereman joined #git
14:36 jast ah, but the first one uses the completion/prompt files shipped with git, so that's okay
14:36 chachasmooth joined #git
14:37 fstd joined #git
14:37 AaronMT joined #git
14:37 tvw joined #git
14:38 thebird joined #git
14:39 Levex joined #git
14:39 TbobbyZ joined #git
14:41 fstd joined #git
14:42 kousu joined #git
14:46 kousu I misused `git merge -s ours topic-branch` trying to avoid dealing unrelated merge conflicts from a different topic branch. But -s ours just chose everything on my master branch, ignoring the changes on topic-branch. Oops! So now I need to undo that, but `git merge topic-branch` says "Already up-to-date."
14:46 kousu Is there a way I can force the merge? Should I try to revert it?
14:48 sbasso joined #git
14:48 Naan joined #git
14:48 ispn joined #git
14:48 fstd joined #git
14:48 JeroenT joined #git
14:49 al-damiri joined #git
14:49 h12o joined #git
14:50 marianina8 joined #git
14:51 a_thakur joined #git
14:52 Dumblez0r joined #git
14:54 jstimm joined #git
14:54 andrew710 joined #git
15:00 JeroenT_ joined #git
15:01 qqx kousu: Have you pushed the merge?
15:01 kousu qqx: yes :(
15:01 kousu and there's work on top of it now
15:02 kousu If I `git diff topic-branch` I see lots of changes, but `git merge topic-branch` says no because it is only looking at the commit IDs I guess.
15:02 qqx Unfortunately there isn't really a good way to deal wit that. Even if you revert the merge, the work from that branch will still be part of your history and so not considered for future merges.
15:02 dreiss joined #git
15:03 qqx Perhaps cherry picking the changes from the topic would work.
15:03 eclecticjohny joined #git
15:04 andrew710 joined #git
15:05 kousu I thought about trying that, but I was worried about screwing it up more
15:05 kousu I'll give it a shot though :)
15:06 qqx At least until you push changes, it generally possible to recover from most screw ups.
15:07 _rgn joined #git
15:11 xaviergmail joined #git
15:12 Shatnerz0 joined #git
15:15 mizu_no_oto joined #git
15:16 GodGinrai joined #git
15:17 zxd d
15:18 xpk joined #git
15:19 enginpost I am doing something wrong.  If I am locally on a branch called old_branch and I want to push that up to origin and overwrite it, how do I do that?  if I write git push --force, that just pushes by local old_branch up to the remote repo. How do I replace origin master with old_branch?
15:19 xpk joined #git
15:20 jast enginpost: git push --force origin old_branch:master
15:20 enginpost jast: how do you read that such that it makes sense to me
15:20 jast push to origin, source is old_branch, destination is master
15:20 qqx I'd recommend using --force-with-lease rather than just --force .
15:21 enginpost aaaaah
15:21 enginpost jast: awesome
15:21 enginpost makes so much more sense
15:21 LBV_User joined #git
15:21 finalbeta joined #git
15:21 jast I don't recommend using this too much because you'll get your branches mixed up at some point :)
15:21 theoceaniscool joined #git
15:22 thebird joined #git
15:22 enginpost jast: does it replace the commit history with the old_branch history then?
15:23 fahadash joined #git
15:25 svm_invictvs joined #git
15:27 rvgate_ joined #git
15:29 impliednude joined #git
15:30 Gsham joined #git
15:31 Gustavo6046 joined #git
15:32 boombatower joined #git
15:33 brent__ joined #git
15:33 ResidentBiscuit joined #git
15:34 brent__ joined #git
15:34 jast enginpost: yes
15:34 brent__ joined #git
15:34 lmatteis joined #git
15:36 tang^ joined #git
15:38 absinthe joined #git
15:38 TbobbyZ joined #git
15:38 Snugglebash joined #git
15:39 tachoknight__ joined #git
15:40 zivester joined #git
15:41 raspado joined #git
15:44 re1 joined #git
15:45 absinthe joined #git
15:46 jstimm joined #git
15:47 Tobbi joined #git
15:47 _rgn joined #git
15:47 _rgn joined #git
15:47 miczac_ joined #git
15:47 jnewt joined #git
15:48 matoro joined #git
15:49 navidr joined #git
15:50 dviola joined #git
15:51 finalbeta joined #git
15:51 xissburg joined #git
15:52 WeiJunLi joined #git
15:56 OnceMe how can I check last commits?
15:56 mizu_no_oto joined #git
15:56 osse define check
15:57 OnceMe osse: I want to see all files which have been changed in the last commit
15:57 osse OnceMe: git show
15:57 osse git show --name-only
15:57 osse git show --stat
15:58 osse for machine consuption: git diff -z --name-only HEAD~ HEAD
15:58 OnceMe and in the previous commit?
15:58 osse git show ... HEAD~
15:58 cdg joined #git
15:58 OnceMe git show ... HEAD~
15:58 OnceMe returns nothin
15:58 s2013 joined #git
15:58 osse bullshit
15:58 osse doesn't it even print an error ?
15:59 OnceMe no
15:59 troulouliou_div2 joined #git
15:59 OnceMe I typed literally "git show ... HEAD~"
15:59 dreiss joined #git
15:59 osse yes, I didn't bother typing the two variants one more time
15:59 miczac joined #git
15:59 osse but ok:
15:59 osse git show --name-only HEAd~
15:59 osse git show --stat HEAd~
16:00 osse correct the 'd'
16:00 OnceMe why HEAd?
16:00 duderonomy joined #git
16:00 osse to make you understand what I mean instead of copying it blindly
16:00 kbs joined #git
16:00 dviola joined #git
16:01 OnceMe ok shows previous commit
16:01 osse Success!
16:01 koneko joined #git
16:01 OnceMe how to go back to the last commit
16:01 OnceMe not last commit 2 times
16:02 OnceMe git show . HEAD~ ?
16:02 osse ?
16:02 h12o joined #git
16:02 roelmonn_ joined #git
16:03 khfeng_ joined #git
16:03 osse you don't have to "go back". this command doesn't actually change anything. it just prints information
16:04 OnceMe osse: ok
16:04 OnceMe but how can I go to see 3rd commit from the back
16:04 invisbl joined #git
16:04 wrouesnel1 joined #git
16:04 OnceMe if git show --name-only is last commit
16:05 OnceMe and before the last commit is show --name-only HEAD~
16:05 OnceMe how do I get the third one?
16:05 OnceMe HEAD~~?
16:05 jast HEAD~2 or HEAD~~
16:05 perlpilot HEAD~# where # is some number
16:05 OnceMe thanks
16:05 Elzington joined #git
16:06 jast man gitrevisions for all the details on these types of abbreviations
16:06 gitinfo the gitrevisions manpage is available at http://jk.gs/gitrevisions.html
16:06 Vauff joined #git
16:06 clemf joined #git
16:06 OnceMe how can I master git?
16:07 oich joined #git
16:07 Rodya_ joined #git
16:07 perlpilot OnceMe: practice.  lots and lots and LOTS of practice  :)
16:07 OnceMe when I read git-show it gives me bunch of params and options
16:07 OnceMe I'm just lost
16:10 Noldorin joined #git
16:11 tekniq joined #git
16:14 dviola joined #git
16:15 AlHafoudh joined #git
16:16 svm_invictvs joined #git
16:16 kbs (imo at least) git is best learnt at the lowest level - it's easy to understand what it's doing. General git commands are ... interesting, and sub-optimal as a way to understand git.
16:17 muhannad_______ joined #git
16:19 jast 'git show' simply has a lot of options. what exactly were you looking for?
16:19 marianina8 joined #git
16:20 durham joined #git
16:22 madewokherd joined #git
16:22 Cabanossi joined #git
16:23 zxd what's the difference between 'merge --squash' and  rebase [<upstream>] [<branch>]
16:23 grawity merge with --squash compresses everything into a single commit and applies it on top
16:23 grawity rebase preserves the commit separation -- if you had 10 before, you'll have 10 afterwards
16:24 cdg_ joined #git
16:24 thebird joined #git
16:25 guampa joined #git
16:25 re1 joined #git
16:25 zxd grawity: merge with --squash  creates a commit?
16:26 Vampire0 zxd yes
16:26 ayogi joined #git
16:26 Vampire0 zxd, except you use --no-commit
16:26 re1 joined #git
16:26 zxd with --no-commit it's just like copying files over to the working tree and staging them with git add ?
16:27 dglambert joined #git
16:27 dglambert yo
16:27 Vampire0 zxd, should be, yes
16:27 rominronin joined #git
16:27 gitinfo dglambert: [!welcome] Welcome to #git, a place full of helpful gits. If you have a question,  feel free to just go ahead and ask—somebody should answer shortly.  For more info on this channel, see http://jk.gs/git/  Take backups (type !backup to learn how) before taking advice.
16:27 rominronin joined #git
16:27 zxd except it has all the candies of checking conflicts and such
16:28 re1 joined #git
16:28 Vampire0 zxd, should have, yes. But rebase also has if that was your concern
16:28 dglambert I just did a git push --set-upstream origin MyBranch
16:29 jstimm joined #git
16:29 dglambert is the origin maintained between branches or globally? So if I go back to MyOtherBranch, and push origin, will it try to push to my newly set origin, or does each branch have an origin
16:30 Levex joined #git
16:30 re1 joined #git
16:31 Vampire0 dglambert, "origin" is just a name for a remote repository of which you can have 0 to any number, named however you like. "origin" just happens to be the conventional default name. The upstream relation is just for one branch. You say "this local branch tracks that remote branch in that remote repository" which then is used for push, pull, merge and rebase by default if you don't specify anything
16:31 Levex joined #git
16:31 tmsmith joined #git
16:31 _ikke_ (also for git status)
16:32 Vampire0 right :-)
16:32 durham_ joined #git
16:32 h12o joined #git
16:32 Vampire0 dglambert, so if you have "MyBranch" checked out and do "git push", you push to the upstream branch configured for that branch. If you do "git push origin MyBranch:foo" you push to the remote branch "foo", no matter what upstream is defined
16:32 Darcidride joined #git
16:33 Vampire0 dglambert, if you have any other branch checked out and try to "git push" it will tell you that it doesn't know where to push to, as no upstream branch is defined
16:33 dglambert got it
16:33 dglambert so the upstream is maintained at the branch level
16:33 dglambert but I can push to other branches using MyBranch:foo
16:34 _ikke_ correct
16:35 Vampire0 dglambert, exactly
16:35 mns joined #git
16:36 Vampire0 dglambert, you can always use explicit refspecs. You can also push any arbitrary SHA to a remote branch, does not have to be a branch name actually
16:36 svm_invictvs joined #git
16:36 OnceMe is remotes/origin/master origin/master the same?
16:36 OnceMe git branch -r gives without remotes/
16:36 OnceMe where git branch -a gives with remotes/
16:37 _ikke_ OnceMe: refs/remotes/origin/master
16:37 OnceMe refs/remotes/?
16:37 _ikke_ That's the complete name
16:37 dglambert wait on, git push origin MyBranch:foo
16:37 Vampire0 OnceMe, if "origin/master" is not ambiguous, it is the same and actually fully refs/remotes/origin/master as _ikke_ said
16:37 _ikke_ refs/heads/master == master
16:38 sixtel joined #git
16:38 robotroll joined #git
16:38 dglambert git push [myRemoteRepositoryAlias] [myLocalBranchAlias]:[myRemoteBranchAlias]
16:38 MacWinner joined #git
16:38 dglambert git push repo branch:branch
16:39 OnceMe ok
16:39 OnceMe my branch locally and on remote are called master
16:39 OnceMe is that fine?
16:39 _ikke_ yup, that's not a problem
16:39 OnceMe cool thanks
16:39 OnceMe I did arc land master
16:39 Vampire0 dglambert, yes, is there a question hidden?
16:39 OnceMe and was worried it went pushed to something else, but when I checked with git show its all good
16:39 gitinfo OnceMe: This channel tracks karma based on who has gotten lots of thanks for being helpful. If you want to help someone reach karmic nirvana, please mention their name when thanking them with "thank you", "thankyou", "thanks", "thx" or "cheers". Try ".karma <nick>" or ".topkarma" to show karma status of a person.
16:39 dglambert Yes, is my understanding/example correct?
16:40 Vampire0 yes
16:40 tmsmith joined #git
16:40 _ikke_ git push <remote> <refspec>; where refspec is <srcspec:remotespec>
16:40 dglambert so I can cross push from one branch to another?
16:40 YuGiOhJCJ joined #git
16:40 _ikke_ dglambert: You totally can
16:41 dglambert so does it merge under the hood?
16:41 _ikke_ No
16:41 Vampire0 dglambert, as I said, you can even do "git push origin 234abc:foo" to push the commit with SHA 234abc to the remote branch foo.
16:41 _ikke_ It either rejects, or you have to force push
16:41 dglambert so does it just overwrite it?
16:41 _ikke_ git push cannot merge
16:41 Vampire0 dglambert, if have no branch called 234abc that is
16:41 _ikke_ if you force push, then yes
16:42 durham joined #git
16:42 TikityTik joined #git
16:42 dglambert so it will create branch 234abc in the remote repo
16:42 dglambert or overwrite it if you force push
16:43 Rodya_ joined #git
16:43 Vampire0 dglambert, nope, it will create or update branch "foo"
16:43 TomyLobo joined #git
16:44 Vampire0 dglambert, left is local right is remote
16:44 dglambert oh sorry, I didn't see your example
16:44 Vampire0 dglambert, 234abc was an example for any arbitrary SHA or actually any commit-ish
16:44 dglambert oh
16:44 dglambert ok got it
16:45 dglambert I think I got the basics though, git push [repo] [branch]:[branch]
16:45 dglambert or git push [repo] [sha]:[branch]
16:45 dglambert where sha can reference a commit
16:45 jast or tag or whatever you can come up with that resolves to a commit
16:46 _ikke_ dglambert: git calls all those "comittish"
16:46 eclecticjohny joined #git
16:46 dglambert got it
16:46 _ikke_ anything to can be resolved to a commit
16:46 dglambert makes sense
16:46 _ikke_ s/to/that/
16:46 dglambert thanks _ikke_
16:46 dglambert thanks Vampire0
16:47 Vampire0 dglambert, yw
16:47 WizJin joined #git
16:47 dglambert .topkarma
16:47 gitinfo top karmic beings of past 30 days ('all' arg to see totals): _‍ikke_: 5,  o‍sse: 3,  j‍ast: 2,  v‍ampire0: 1,  q‍qx: 1
16:47 * _ikke_ whistles
16:47 jast neat, I haven't managed the top 5 in a while
16:47 jast .topkarma all
16:47 gitinfo top karmic beings of all time: o‍sse: 268,  _‍ikke_: 234,  c‍break: 144,  s‍eveas: 115,  c‍anton7: 111
16:47 jast still not in there, of course
16:47 Vampire0 _ikke_, _ikke_, _ikke_ yaaay :-D
16:47 mizu_no_oto joined #git
16:48 jast .karma
16:48 gitinfo the Genuine Real Life Karma™ REST API results are back! j‍ast: 96 (2 in past 30 days) (2 given out)
16:48 Vampire0 .karma
16:48 gitinfo the Genuine Real Life Karma™ REST API results are back! v‍ampire0: 29 (1 in past 30 days) (3 given out)
16:48 bremner karma++
16:48 arand__ joined #git
16:48 Vampire0 Ouuuh, long way to go :-D
16:48 lucido-cl joined #git
16:48 jast karma is a lot of work :}
16:48 dglambert I thought I remembered a karma system in irc way back when I first started, I think it was this channel but I drifted away after I got the hang of things
16:49 _ikke_ it sure is
16:49 dglambert then I switched to tfs for the past 3 years
16:49 jast and cheaters will be hanged
16:49 jast the karma system here doesn't get that much attention, really
16:50 _ikke_ jast: That's what you think :P
16:50 jast I made the numbers increase fairly slowly to reduce the instant gratification
16:50 Levex joined #git
16:50 _ikke_ yeah, that's smart
16:50 TbobbyZ joined #git
16:50 jast that's me :P
16:50 _ikke_ lol
16:52 rwp .karma
16:52 gitinfo the Genuine Real Life Karma™ REST API results are back! r‍wp: 1 (0 in past 30 days) (0 given out)
16:52 h12o joined #git
16:52 dev42 joined #git
16:53 orbyt_ joined #git
16:53 TbobbyZ joined #git
16:53 Silenced joined #git
16:54 Brski_ joined #git
16:54 flopsec joined #git
16:55 dev42 left #git
16:55 lrochfort joined #git
16:55 archaic joined #git
16:57 lrochfort We have a feature branch that multiple devs commit to. Periodically, we need to get it up-to-date with master. Given that we can't rebase pushed commits, is a merge of origin/master into the feature branch the best option?
16:57 eroussel joined #git
16:57 Eugene Yup.
16:58 grawity it's not unusual to rebase feature branches, although it's much more annoying than just merging
16:58 lrochfort grawity: Would rebasing require a force push?
16:58 ij Of the feature branch, yes.
16:58 grawity among other things, yes
16:58 Darren_ joined #git
16:58 lrochfort Which then causes pain for the other developers.
16:59 lrochfort Could you expand on the other things?
16:59 ij If they pull with rebase, not too much pain.
16:59 grawity mainly just fallout from the force-push; !rewriting
16:59 gitinfo [!rewrite] Rewriting public history is not recommended. Everyone who has pulled the old history will have to do work (and you'll have to tell them to), so it's infinitely better to just move on. If you must, you can use `git push --force-with-lease <remote> <branch>` to force (and the remote may reject that, anyway). See http://goo.gl/waqum
17:00 lrochfort ij: What do you mean by not too much pain?
17:00 ij Probably no pain at all.
17:01 jstimm joined #git
17:02 lrochfort ij: Sorry, could you elaborate? So, somebody rebased onto master, then force pushed. What would be required of the other devs?
17:02 ij If you're into this kind of thing, you can try it out by making two local copies of some repo and make diverging branches on each and then pull from each other.
17:02 warthog9 joined #git
17:03 lrochfort I'll try, thanks.
17:04 Kobaz joined #git
17:06 hahuang61 joined #git
17:07 koneko joined #git
17:08 gitinfo lrochfort: This channel tracks karma based on who has gotten lots of thanks for being helpful. If you want to help someone reach karmic nirvana, please mention their name when thanking them with "thank you", "thankyou", "thanks", "thx" or "cheers". Try ".karma <nick>" or ".topkarma" to show karma status of a person.
17:08 gugah_ joined #git
17:09 raijin joined #git
17:10 leeN joined #git
17:10 Urbany joined #git
17:10 ij lrochfort, try this out in an empty directyory: http://sprunge.us/QVVA
17:10 ij and then do a git log in a
17:12 h12o joined #git
17:13 dviola joined #git
17:14 multi_io joined #git
17:15 ij A better version: http://sprunge.us/ZJiB
17:16 nohop is there a git command that returns a non-zero value when there is no tag for the current head ?
17:16 kousu thanks qqx, cherry-picking seems to have solved my bad merge well enough
17:18 davimore__ joined #git
17:18 chachasmooth joined #git
17:18 eh3- joined #git
17:20 eroux joined #git
17:20 dreiss joined #git
17:20 jungsubk joined #git
17:22 mizu_no_oto joined #git
17:23 yoh joined #git
17:23 raijin joined #git
17:23 Miika--_ joined #git
17:23 Miika--_ left #git
17:23 Silenced joined #git
17:24 thiago joined #git
17:26 yoh use-case: upon 'submodule update', I want to introduce a logic to make submodule's HEAD to stay within the branch even if remote end "progressed".  so I check if updated hexsha is within the history for the branch, and then  update-ref refs/heads/BRANCH CURRENT_HEXSHA and then  to get into the branch   update-ref HEAD refs/heads/BRANCH .... and the update of HEAD worked once but now just doesn't do anything... if
17:26 yoh I just echo "refs/heads/BRANCH" > .git/HEAD it seems to be what I need ,but not koshre
17:26 yoh how could I figure out why  update-ref HEAD doesnt do anything?
17:27 yoh here is a bit more of info etc:  http://paste.debian.net/919893/
17:28 bitmod guys if i create a new file in a project and don't stage it, should git diff show the contents of that file?
17:28 cdg joined #git
17:28 bitmod at the moments it's just giving me a blank
17:28 psprint joined #git
17:29 yoh bitmod: git diff operates only on files known (tracked) by git
17:29 yoh until you 'git add' that new file, git does not know about it
17:29 _rgn joined #git
17:29 livingstn joined #git
17:29 bitmod yoh: ah ok. Although even after adding it i'm still getting nothing when i run git diff
17:30 yoh bitmod: ;) that is a different case -- when you 'git add' it gets staged to be committed
17:30 yoh git diff by default shows difference from what you have in the tree to the state what to be committed
17:30 yoh so unless you add more changes to that file -- git diff would be empty
17:30 tang^ a new file has nothing to diff against
17:30 yoh git diff --cached, would show difference from what was staged to be committed, so try that one ;)
17:31 bitmod thanks!
17:31 qqx You could also use `git add -N` to tell git that you want to track the file, but not stage the contents.
17:32 flavius joined #git
17:33 sword joined #git
17:33 h12o joined #git
17:34 Tobbi joined #git
17:34 cliluw joined #git
17:34 yoh so anyone has a clue when 'update-ref' doesn't actually do anything asked from it? ;)
17:35 grawity note that HEAD is usually a symbolic-ref, not a regular one
17:36 gugah__ joined #git
17:37 yoh grawity: ah, doh, bleh... ;) thanks ... lets try corresponding update
17:37 yoh grawity: worked! THANKS!
17:38 marianina8 joined #git
17:38 lrochfort left #git
17:40 bitmod yoh: also, when i create a new file and run git diff, it lists the file (which is untracked). If i then add the file, and create a new file, git diff lists no files. Should this be happening?
17:41 fedepad joined #git
17:41 a_thakur joined #git
17:41 cahoots joined #git
17:41 moritz bitmod: git diff --staged shows things that you already git add'ed
17:42 corentin or --cached
17:42 cahoots if no branches point at a commit anymore (a commit with large files that i want to get rid of), no one will git pull it anymore, correct?
17:42 moritz bitmod: also, 'git status' to get an overview of what happened in the repo (both unstaged changes and staged chagnes)
17:42 cahoots also, is there a way to make sure that no branches point to it anymore?
17:42 moritz corentin: that was never intuitive to me. People call the index "staging area", but it's not really a cache, IMHO
17:43 bitmod moritz: yeah that part i get, and git diff --staged shows the file that i "git add"ed, but shouldn't git diff be showing the second file i created that hasn't been staged yet?
17:43 corentin moritz: agreed
17:43 robotroll joined #git
17:43 moritz bitmod: no
17:44 moritz bitmod: until you start tracking it with 'git add', git doesn't know and care about the file
17:44 bitmod moritz: ah right, has to be tracked
17:44 Sasazuka joined #git
17:44 moritz bitmod: hence, use 'git status' if you also care about untracked files
17:45 matoro joined #git
17:47 bitmod moritz: i think i get it now, so git diff --staged is the difference between the last commit and the current tracked files, and git diff is the difference between the state of the files when they were tracked and their modified state?
17:48 porsche944 joined #git
17:48 qqx bitmod: That's correct. You can also use `git diff HEAD` to show the difference between the last commit and your working files (but still only ones that are tracked).
17:49 Snugglebash joined #git
17:49 clemf joined #git
17:49 Yewbacca joined #git
17:49 systemovich joined #git
17:50 Yewbacca Hey what's the strongest git gc command to pack everything even new changes?
17:50 Yewbacca git gc --aggressive, or more?
17:50 bitmod qqx: fantastic, thanks
17:51 gopar joined #git
17:51 Kasreyn left #git
17:52 chrisgeorge joined #git
17:53 jophish joined #git
17:54 Yewbacca git gc --aggressive --prune=all ?
17:55 duderonomy joined #git
17:55 Cabanossi joined #git
17:56 a_thakur joined #git
17:57 mda1 joined #git
17:58 zzz joined #git
18:03 _ikke_ I would not use --prune=all
18:03 h12o joined #git
18:05 steve joined #git
18:06 chardan joined #git
18:07 NeXTSUN joined #git
18:08 filcab joined #git
18:08 nd_ joined #git
18:11 chachasmooth joined #git
18:12 thiago joined #git
18:13 chardan joined #git
18:14 h12o joined #git
18:16 jason237 joined #git
18:17 Es0teric joined #git
18:18 julia_c joined #git
18:18 paultreny joined #git
18:19 paultreny joined #git
18:22 Tatou joined #git
18:22 bitmod _ikke_, how would you suggest a beginner uses git? should i track every file after i change it and commit sporadically, or do it in a more organize manner?
18:22 Aeterne joined #git
18:22 bitmod organized*
18:23 matoro joined #git
18:24 h12o joined #git
18:24 madewokherd try to commit at a time when things at least build, and if possible limit commits to a single logical change
18:25 madewokherd but that can be difficult with a new project, sometimes I end up with several changes in one commit
18:25 Levex joined #git
18:26 _ikke_ madewokherd: Yeah, I struggle with that too, consise commits in a starting project
18:27 xall joined #git
18:31 raijin joined #git
18:32 leeN joined #git
18:33 raijin joined #git
18:33 kbs joined #git
18:36 kexmex joined #git
18:38 Levex joined #git
18:40 stoopkid joined #git
18:42 ash_workz joined #git
18:44 h12o joined #git
18:44 ankit01ojha joined #git
18:50 durham joined #git
18:51 luke_nukem joined #git
18:52 mknod joined #git
18:53 bitmod _ikke_, how would you suggest a beginner uses git? should i track every file after i change it and commit sporadically, or do it in a more organize manner?
18:54 hasc joined #git
18:55 _ADN_ joined #git
18:56 GodGinrai joined #git
18:56 GodGinrai bitmod: Use commits to track meaningful changes
18:57 GodGinrai bitmod: if you make enough changes to one file that you think it is important to have its own commit, do so.  If you think it is part of a more sweeping set of changes, then commit once you have made all of the changes
18:57 h12o joined #git
18:58 kasual joined #git
18:58 TbobbyZ joined #git
18:58 Levex joined #git
18:58 kadoban bitmod: err on the side of commiting too often, especially at first.
18:59 chrisgeorge_ joined #git
18:59 kadoban You often see beginners with hours-long work sessions trying to then commit a bunch of stuff, and it's ... not particularly pretty. It makes later things you might want to do more annoying.
19:00 Vampire0_ joined #git
19:01 nivag_ joined #git
19:01 rob joined #git
19:02 nivag joined #git
19:02 bitmod thanks for the advice
19:03 chachasmooth joined #git
19:06 aidalgol joined #git
19:06 fedepad joined #git
19:06 satifant joined #git
19:07 Endarked joined #git
19:08 xaviergmail joined #git
19:12 clemf joined #git
19:13 clemf joined #git
19:14 miha_S7_ joined #git
19:14 chachasmooth joined #git
19:15 diogenese joined #git
19:17 raijin joined #git
19:17 troulouliou_dev joined #git
19:18 h12o joined #git
19:18 shakalaka joined #git
19:19 WeiJunLi_ joined #git
19:19 Balliad joined #git
19:24 Atm0spher1c joined #git
19:24 miczac joined #git
19:25 finalbeta joined #git
19:25 Nugget joined #git
19:26 hotpancakes joined #git
19:26 WeiJunLi_ left #git
19:27 chrisgeorge joined #git
19:28 glowdemon1 joined #git
19:29 porsche944 joined #git
19:29 chachasmooth joined #git
19:31 hotpancakes joined #git
19:32 hotpancakes Hey everyone. Quick question. I'm working on a project and I have work that's not ready to be committed, but I'd like to try something experimental out. Is there any way I can create a branch from the current state of my project, not from my last commit?
19:33 hotpancakes I just want to spend an hour trying something out, with the knowledge that I can easily get rid of it and get back to my current state.
19:33 relipse joined #git
19:33 Vampire0 hotpancakes, yes, just give it as additional option
19:33 qqx hotpancakes: The way I generally handle that is to create a branch, commit, then `git checkout -` to get back to the prev branch.
19:33 relipse is there a reason why my branch isn't listed under git branch and it won't let me do a git merge either
19:33 relipse i tried git fetch
19:34 clemf joined #git
19:34 Vampire0 hotpancakes, `git branch foo <commit-ish-to-start-foo-from>` or `git checkout -b foo <commit-ish-to-start-foo-from>`
19:34 Vampire0 relipse, you mean a remote branch?
19:34 Vampire0 relipse, !where
19:34 gitinfo relipse: [!wherearemybranches] Git clones all branches, but initially makes only the main branch (usually 'master') available as a normal local branch. To work on another branch, simply 'git checkout <branch>' (or, in versions older than 1.6.6, git checkout -b <branch> origin/<branch>). This works even if the branch doesn't show up in the normal list of branches. For technical details, see !rtb
19:35 clemf joined #git
19:35 qqx Then if I later decide I want to go back to what I had before I do `git reset --hard <the new branch> && git reset HEAD~`.
19:35 relipse ok so git merge origin/mybranch
19:35 durham joined #git
19:35 gugah joined #git
19:37 eroussel joined #git
19:38 cr34ton joined #git
19:38 hotpancakes qqx: thanks, what does `git checkout -` do?
19:38 h12o joined #git
19:38 qqx That checks out the previous branch.
19:40 hotpancakes qqx: but wouldn't that prevent me from getting my existing staging area back?
19:40 navidr joined #git
19:40 miczac joined #git
19:41 irqq joined #git
19:41 chachasmooth joined #git
19:42 hotpancakes Vampire0: after I checkout a branch from my last commit on master, how would I return to master while keeping my current staging area?
19:43 qqx Is it the existing staged content that you want to base your trial work on or do you want5 to get back to that later.
19:43 robattila256 joined #git
19:43 moritz hotpancakes: !float
19:43 gitinfo hotpancakes: 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.
19:43 PHPanos joined #git
19:43 hotpancakes qqx: both
19:44 Vampire0 hotpancakes, you are probably after man git stash
19:44 gitinfo hotpancakes: the git-stash manpage is available at http://jk.gs/git-stash.html
19:45 durham joined #git
19:46 xissburg joined #git
19:46 prg3 joined #git
19:48 Vampire0 hotpancakes, alternatively you can of course also make some commits on your current branch and later revert them again
19:48 Vampire0 hotpancakes, just like you prefer
19:48 timetrex joined #git
19:49 timetrex left #git
19:49 qqx Yeah, stash can work well if the content is likely to be short-lived. If I expect to want to hang onto the content longer I prefer to make commits with decent commit messages to help me remember the status.
19:50 chachasmooth joined #git
19:51 rkazak_ joined #git
19:54 OnceMe while git pull works and sync my current branch with the master one
19:54 OnceMe why would I use git pull origin master?
19:54 OnceMe it will do the same job
19:55 cdown joined #git
19:56 hotpancakes Ah okay. So I just stashed all my uncommitted work, checked out a branch (which preserves staging area), and made some changes. To revert back to original, I checked out master, discarded all local changes, and then re-applied my saved stash. Seems like a kludge, but it works.
19:57 hahuang61 joined #git
19:57 hotpancakes Sorry if this seems basic, still learning. :)
19:57 Snugglebash joined #git
19:58 venmx joined #git
19:58 foul_owl Assuming I have two repos. Repo a and repo b. repo a is a legacy repo on a legacy git hosting service. repo b is an empty repo created on gitlab. How do I copy repo a to repo b? I tried adding a new remote and doing a push to repo b, but that only gets master...it doesn't move all branches. Thanks!
19:58 h12o joined #git
19:59 foul_owl repo a has about 75 branches
19:59 moritz foul_owl: you start with a git clone --miror a
19:59 moritz foul_owl: and then git remote add b url-to-b
20:00 moritz git push --mirror
20:00 moritz git push --mirror b
20:00 foul_owl Ahhhhhhh ok so what I need is the --mirror option!
20:00 foul_owl Awesome
20:00 foul_owl Thank you!! :)
20:00 qqx Not necessarily
20:00 moritz you might even be able to do use git push --miror b without a bare repo, dunno
20:00 qqx You could also do `git push b refs/remotes/a/*:refs/heads/*`.
20:01 qqx That should work from the existing clone.
20:01 foul_owl Thanks!! :)
20:01 t0by joined #git
20:01 t0by joined #git
20:02 sunri5e joined #git
20:03 adac joined #git
20:07 Vampire0 hotpancakes, git stash should also stash the staged changes, except you told it to leave the stashed changes in place. With --keep-index if I remember correctly
20:10 venmx joined #git
20:12 jstein_ joined #git
20:12 cdg joined #git
20:12 clemf_ joined #git
20:12 tyreld joined #git
20:14 circ-user-Jhrga joined #git
20:15 Naan joined #git
20:17 a_thakur joined #git
20:17 venmx joined #git
20:18 aielima joined #git
20:19 h12o joined #git
20:20 dreiss joined #git
20:20 chachasmooth joined #git
20:20 enginpost_ joined #git
20:21 sherm4n4tor_ joined #git
20:21 enginpost_ sherm4n4tor: hey
20:22 mar77i joined #git
20:22 enginpost_ sherm4n4tor: are you there
20:22 pbrewczynski joined #git
20:22 enginpost_ sherm4n4tor: over here man
20:23 sherm4n4tor enginpost: leave me alone
20:23 sherm4n4tor enginpost: jk
20:24 matoro joined #git
20:25 enginpost_ left #git
20:25 enginpost_ joined #git
20:25 sherm4n4tor left #git
20:25 sherm4n4tor joined #git
20:25 enginpost_ joined #git
20:26 _ikke_ ..
20:26 enginpost_ joined #git
20:26 cdown_ joined #git
20:26 g0d355__ joined #git
20:26 hotpancakes Vampire0: Yeah, thanks! Are there any other ways you'd recommend doing this without using stash?
20:27 dminuoso You mean like with github?
20:27 dminuoso :|
20:28 sherm4n4tor I'm having trouble with git to ignore the bin and obj folders from a Visual Studio project.  Can someone help me?
20:32 glowdemon1 joined #git
20:33 zzz joined #git
20:34 minn joined #git
20:35 cdown_ joined #git
20:36 overlord_tm joined #git
20:36 strugee joined #git
20:36 zeroed joined #git
20:36 zeroed joined #git
20:37 re1 joined #git
20:39 btree joined #git
20:40 jstimm joined #git
20:41 btree left #git
20:42 _Vi joined #git
20:42 apotry joined #git
20:42 _Vi Shall Git merge also merge submodules if the such merge inside submodule would be fast-forward?
20:43 GodGinrai joined #git
20:44 _ikke_ No, git merge does not look inside submodules
20:46 _Vi _ikke_, Is it a good idea to have a mode that looks inside submodules during merge and handles fast-forward cases automatically, without even marking it as conflict?
20:46 apotry joined #git
20:46 _ikke_ no
20:46 _Vi What are issues?
20:47 _ikke_ In general, git does not care what happens in the submodule. It only tracks what commit it needs to checkout
20:47 _Vi As far as I been told, Sourcetree does that. Why command-line Git client shouldn't?
20:47 Levex joined #git
20:48 _ikke_ Never heard of that
20:48 _Vi For example, it may be `git merge ... --recurse-submodules`.
20:49 zzz joined #git
20:49 askb joined #git
20:51 irqq joined #git
20:51 _Vi Lesser idea: should `git merge-tool` go into submodules and show how diverging they are, offering not a choice between "aaaa" and "bbbb", but between "aaaa (3 commits since ccccc)" and "bbbb (103 commits since ccccc)". Does it sound nice?
20:53 _Vi The big idea may be transparent submodules: users pull, edit, commit and push code without even realizing that there are submodules. `git push --recurse-submodules` is a step in that direction.
20:54 sherm4n4tor joined #git
20:57 griffindy joined #git
20:58 wrouesnel joined #git
20:59 tmg joined #git
20:59 lamer14895246260 joined #git
20:59 lamer14895246260 joined #git
21:00 tyreld joined #git
21:00 chipotle joined #git
21:01 mizu_no_oto joined #git
21:01 durham_ joined #git
21:02 _ikke_ Right, but it's a bit orthogornal to the original idea behind them
21:03 _Vi _ikke_, The idea is sharing code between projects in organized way, without duplication (like with subtree).
21:04 relipse joined #git
21:04 _Vi That way submodule-unaware developer may edit code for submodule-aware maintainer of side project.
21:05 _Vi s/(like with subtree)/(unlike with subtree)./
21:05 relipse i have a qa branch but i'm concerned if i merge in another branch and then go and update the other branch and try to merge it in qa again it will mess up. Is this true?
21:05 relipse i want to merge and then if there are code changes merge again
21:07 dviola joined #git
21:07 buzzLightBeer joined #git
21:07 _Vi relipse, The second merge should handle the new changes since the first merge.
21:07 _ikke_ _Vi: https://public-inbox.org/git/20170306205919.9713-1-sbeller@google.com/T/#t a thread that is talking about integrating submodules more into other git commands
21:08 blackwind_123 joined #git
21:08 _ikke_ https://public-inbox.org/git/CAGZ79kZmWaobW9e4iPY05y0N6PLcFphGnZmDHtrGKeV0Up70vg@mail.gmail.com/
21:10 Tobbi joined #git
21:11 fedepad joined #git
21:11 durham joined #git
21:12 jost_ joined #git
21:13 relipse thanks _Vi
21:15 LBV_User joined #git
21:15 dviola joined #git
21:16 _Vi _ikke_, Next idea: Tech "git submodule update --merge" to resolve conflicts in current repository. Now it just says "Skipping unmerged submodule subm".
21:17 _Vi _ikke_, Do you like the idea of using some "git submodule" subcommand (e.g. "update") to automatically handle conflicts in merges involving submodules?
21:17 d0nn1e joined #git
21:17 gabbaell joined #git
21:18 dreiss joined #git
21:19 _Vi It may even be called "git submodule merge".
21:20 _ikke_ You can always propose it on the mailing list
21:20 dsdeiz joined #git
21:21 karoal joined #git
21:21 peterbecich joined #git
21:21 karoal quick question: is it bad practice to have changes in the commit which are not related to the commit message?
21:22 _Vi Yes, but consider asking here first a good idea. Otherwise I would be proposing the "git merge" change, for example.
21:22 karoal even if they are very minor and don't deserve a commit by themselves?
21:22 _Vi karoal, Yes, bad idea.
21:22 _Vi karoal, Whitespace changes?
21:22 karoal _Vi: thank you. Yes, for example
21:22 _Vi karoal, If you commit gets reverted ("git revert"), that unrelated minor change gets reverted too (or prohibits clean revert).
21:23 _ikke_ Yeah, and it clobbers diffs
21:23 _Vi karoal, If your commit cherry-picked somewhere where those little unrelated change is unapplicable (no such file) => conflict.
21:23 karoal _Vi: thanks a lot. Explains a lot to me. So I should just roll back the unnecessary changes?
21:23 _ikke_ It's easier to review when those are separated (though I must say I'm guilty myself of combining them)
21:23 _Vi karoal, Have you already published it (git push)?
21:24 karoal _Vi: no, I try to be careful when I push
21:24 karoal _Vi: that's why I asked :) thanks
21:24 _Vi karoal, If not published then you can roll back and re-organize your commits. If something goes wrong, you can access your old commit using "git reflog".
21:24 Topcat joined #git
21:25 Cabanossi joined #git
21:25 _Vi karoal, Note that if you have multiple little changes (e.g. style fixes here and there), they may be contained within one commit, without separate "Style fixes on file goo.c", "Style fixes for file bar.c".
21:26 Tobbi joined #git
21:26 karoal _Vi: cheers for the advice.
21:26 Topcat evening. I've an existing repo. I want to now work from the dev branch and replace all of master with a subtree tracking a folder. How do I shift off master? Any merge will keep some files if there's no conflict right?
21:27 _Vi Topcat, Should dev branch contain unmodified (unshifted) view?
21:28 Topcat yeh
21:28 _Vi And each merge from dev to master is a subtree merge?
21:28 vedranm joined #git
21:28 powermac joined #git
21:29 Topcat yes
21:29 _Vi Maybe it should be split in two repositores connected by submodules?
21:29 Topcat to restate clearly- dev now is the source I want from now on. master, 1 commit behind, i no longer care about and want to move it to this dev state but only the public folder.
21:30 _Vi public/private? Is there any sensitive information in current or previous Git files?
21:30 Topcat no, i've moved a handmade site to a static generated one and want to take advantage of github pages master deploy
21:31 powermac hi all, my git project lives in my home directory, but it's internal structure is a little annoying, when I clone my master, to /home/ I get /home/foo/foo_project/<stuff>
21:31 hahuang61 joined #git
21:31 powermac how can I get rid of 'foo_project'?
21:31 Topcat I considered just deleting the branch master then doing the subtree push
21:31 _Vi powermac, You can use "sparse checkout" feature.
21:31 _ikke_ I don't think that does what he wants
21:31 _Vi powermac, Or you actually want the "foo_project" to disappear?
21:32 _Vi powermac, (not only from your checkout, but from some central Git repository as well)
21:32 powermac _Vi: From some central git repo as well
21:32 chardan joined #git
21:33 _Vi powermac, If this "foo_project" is not big (weighting the repository down) or contain sensitive information that is to be removed, then probably just "git rm".
21:33 _Vi (assuming nothing important remains there)
21:33 adymitruk joined #git
21:33 Anja joined #git
21:34 _Vi Topcat, Maybe you can use "git commit-tree" command to manually create a new commit specifying which tree (found by "git ls-tree") it should be based on.
21:34 Topcat never heard of that, off to read
21:34 _Vi Topcat, But I'm not sure if the whole idea would work well.
21:34 powermac _vi /foo/foo_project/ contains all of my files, I want all of my files ot live in /foo
21:35 _Vi Topcat, Don't forget to specify parent with "-p" and to update master with it (git update-ref).
21:35 _Vi powermac, Then maybe go into "/foo/foo_project" and do like "git mv * ../"?
21:36 powermac ohhh, fancy
21:36 hotpancakes joined #git
21:36 _Vi powermac, Note that hidden (filename begin with a ".") files will remain inside "foo_project".
21:37 bitmod in the command "git reset HEAD myfile", how can i interpret it?
21:37 bitmod HEAD is the last commit right?
21:37 _Vi bitmod, Remove "myfile" from the pending commit.
21:37 _Vi bitmod, Yes.
21:38 _Vi bitmod, Actualy file "myfile" in filesystem is not affected, but no longer marked "to be commited".
21:38 bitmod _Vi: i guess i'm trying to picture what happens in my head. In the context of "git for ages 4 and up", how would that be visualized?
21:38 Topcat HEAD is where git is now, not necessarily your state
21:38 bitmod _Vi: in other words remove myfile from tracked right?
21:38 livingstn joined #git
21:39 bitmod Topcat: as in the currently checked out branch?
21:39 Topcat bitmod: look through the output of git reflog, you'll see it's where you've moved through git
21:39 bitmod Topcat: how can i do that?
21:40 Topcat just type git reflog
21:40 _Vi bitmod, Between the commit history and your files there is special intermediate zone called "index". Files marked for committing are stored there. But it can even be third version of the file (not your and not what in previous commit). "git add" copies the file from your zone to index zone, "git commit" copies all files from index zone to the new commit. "git reset" (without --hard) copies from last commit to index zone.
21:40 bitmod ah ok
21:40 m0viefreak joined #git
21:40 bitmod _Vi: "index" = "staging" right?
21:40 _Vi bitmod, "in other words remove myfile from tracked right?" -> No.
21:40 _Vi bitmod, '"index" = "staging"' -> Yes.
21:41 bitmod _Vi: doesn't using "git reset HEAD myfile" untrack that file though?
21:41 _Vi bitmod, No.
21:42 _Vi bitmod, The file will still be tracked, but changes in it marked for the next commit with "git add" will be unmarked.
21:42 _Vi bitmod, If "git status" tells nothing about this file then "git reset HEAD thefile" should do nothing.
21:42 bitmod that's confusing, git displays a message "Untracked files", with myfile listed underneath
21:43 _Vi bitmod, Then "git reset HEAD myfile" may fail outright, as far as I understand.
21:43 _Vi Because of the file is probably not in index and not in the last commit.
21:44 _Vi bitmod, Experimented: "git reset HEAD unrelated_file" does nothing if the file is untracked or just not exist.
21:45 bitmod _Vi: but if it's tracked, it untracks it doesn't it?
21:45 bitmod i'm confused
21:45 robattila256 joined #git
21:45 _Vi bitmod, No. For untracking a tracked file there is "git rm --cached". (or just "git rm" if you want to remove the file in addition to untracking it).
21:46 _Vi bitmod, "git reset HEAD myfile" is only about changes to a file, not about the presence of a file in Git.
21:46 afuentes joined #git
21:46 bitmod i'm going to have to go and contemplate this deeply
21:47 _Vi bitmod, Unfortunately the terminology and command-line UI is Git is not the best part. Commands may mean drastically different things (from user perspective) based on options involved.
21:47 _Vi s/UI is Git/UI in Git/
21:48 bitmod yeah, it's my second day using git and i'm still trying to understand the big feature, not intuitive yet
21:48 _Vi bitmod, Have you tried reading a book like "Pro Git"?
21:49 bitmod _Vi: yeah, reading that at the moment
21:49 LeBlaaanc joined #git
21:50 _Vi I expect it to give in-depth knowledge about Git. Don't try to read manpages straight away, before tutorial and grasping terminology.
21:52 roelmonnens joined #git
21:52 miha_S7_ joined #git
21:53 h12o joined #git
21:54 venmx joined #git
21:56 JanC_ joined #git
21:56 foreal joined #git
21:57 DGMurdockIII joined #git
21:57 DGMurdockIII git don't do proper binary diffs, which is annoying unless im missing somthing?
21:57 Topcat _Vi: commit-tree might have worked but since I've not pushed the repo anywhere I decided to back it up, delete master and subtree master from dev with the folder prefix. looks nice. master 1 commit dev 44 commits
21:58 Topcat DGMurdockIII: I dont belive git touches binaries at all. it commits the whole thing each change
21:58 inflames joined #git
21:58 DGMurdockIII ok
21:59 DGMurdockIII is a version control system that dose?
22:00 durham joined #git
22:01 venmx joined #git
22:02 patrick99e99 joined #git
22:02 alphor joined #git
22:02 cdg joined #git
22:03 _ikke_ DGMurdockIII: Git allows you to specify a custom diff / merge driver
22:03 patrick99e99 hey everyone.. a coworker just did a git mv file.js file.es6, and noticed that the file's history is lost...  Stackoverflow says this is an unfortunate side effect of git mv...  Is there any way around this?
22:03 _ikke_ patrick99e99: history is not lost
22:04 acetakwas joined #git
22:04 _ikke_ Many commands have parameters that tells git to follow files across renames
22:05 DGMurdockIII ok
22:05 DGMurdockIII _ikke_, even on binary files?
22:06 _ikke_ DGMurdockIII: Sure
22:06 cdown joined #git
22:10 TomyLobo2 joined #git
22:10 durham joined #git
22:11 ISmithers joined #git
22:11 miczac joined #git
22:12 jimi_ joined #git
22:13 Anja joined #git
22:16 bannakaffalatta joined #git
22:16 chachasmooth joined #git
22:17 minn joined #git
22:20 Dougie187 left #git
22:21 cyrus_mc joined #git
22:22 souLJin joined #git
22:22 cr34ton joined #git
22:24 Sound joined #git
22:27 eroussel joined #git
22:31 impliednude joined #git
22:32 DolpheenDream joined #git
22:33 souLJin left #git
22:39 eh3 joined #git
22:40 NeXTSUN joined #git
22:42 hotpanca_ joined #git
22:49 Noodlewitt joined #git
22:50 anuxivm joined #git
22:51 livingstn joined #git
22:54 BackEndCoder joined #git
22:54 mizu_no_oto joined #git
22:54 menip joined #git
22:54 livingstn joined #git
22:54 Rodya_ joined #git
22:55 arescorpio joined #git
22:55 hasc joined #git
22:59 cyrus_mc left #git
23:02 cyphase joined #git
23:06 miha_S7_ joined #git
23:06 thiago joined #git
23:08 brent__ joined #git
23:09 hahuang61 joined #git
23:10 dviola joined #git
23:11 bocaneri joined #git
23:13 ReptilianBrain joined #git
23:14 a0x10c joined #git
23:14 jimi_ joined #git
23:22 madewokherd joined #git
23:24 fragMental joined #git
23:24 Auea joined #git
23:24 livingstn joined #git
23:25 clemf joined #git
23:26 rschneberger joined #git
23:28 Kaisyu joined #git
23:28 cdown joined #git
23:29 gopar joined #git
23:33 msonntag joined #git
23:36 SwiftMatt joined #git
23:38 raijin joined #git
23:38 shgysk8zer0_ joined #git
23:39 xaviergmail_ joined #git
23:41 peterbecich joined #git
23:43 Gsham joined #git
23:45 theoceaniscool joined #git
23:47 Rodya_ joined #git
23:48 hotpancakes joined #git
23:48 Noodlewitt joined #git
23:50 mysticjohn joined #git
23:51 graSP_ joined #git
23:51 Noodlewitt joined #git
23:51 minn Suppose I have a topology like "A -> B -> C -> D" and file F was modified in B and D. My understanding is that the parent reported by git log (with %P) will be C and A. Is there any way to get the hash of the ancestor on the branch (e.g., report B as the ancestor for D and nothing for B, assuming B was created in B)?
23:52 minn s/assuming B/assuming F/
23:52 Fallen0223 joined #git
23:53 ertes joined #git
23:53 thiago don't assume, test
23:55 x5a joined #git
23:56 DGMurdockIII left #git
23:57 VladGh joined #git
23:58 dave0x6d joined #git
23:58 Fallen0223 joined #git
23:59 minn Well, I just tested it and it confirmed my understanding. So I guess the question remains, is there any way to report the ancestor at which the file was last modified (so that the commit hash and ancestor hash chain together)?

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