Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2017-02-10

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

All times shown according to UTC.

Time Nick Message
00:04 endiruna joined #git
00:05 solenodic joined #git
00:05 pR0Ps joined #git
00:05 spacelord_ joined #git
00:10 e01 joined #git
00:10 e01 joined #git
00:10 rivarun joined #git
00:10 nickabbey joined #git
00:12 dyce[m] joined #git
00:13 tomek joined #git
00:13 tomek joined #git
00:14 madewokherd joined #git
00:14 jstimm joined #git
00:14 zivester joined #git
00:15 dyce[m] is it possible to pull a repo's remotes?
00:15 dyce[m] rather than only getting origin as a remote
00:16 endiruna joined #git
00:16 rewt man git pull
00:16 gitinfo the git-pull manpage is available at http://jk.gs/git-pull.html
00:16 rewt --all
00:17 rewt but at that point, i'd separate pull into fetch and merge
00:17 ojacobson dyce[m]: you can get the commits, but not the urls
00:17 dyce[m] i suppose that's for security reasons
00:18 dyce[m] i guess i can just commit a script that adds them
00:18 ojacobson nope
00:18 ojacobson it's because the urls are stored somewhere fetch doesn't have any way to look
00:19 mephistorium joined #git
00:21 ojacobson if you invent a way to store git-config keys in the object store, then you'd be able to fetch 'em
00:22 ojacobson there isn't one but that doesn't mean there can't be one
00:22 notebox joined #git
00:22 beuker joined #git
00:24 Darcidride joined #git
00:25 Dougie187 left #git
00:31 chriskos joined #git
00:33 thierryp joined #git
00:35 pandbearit joined #git
00:35 pandabearit_ joined #git
00:36 Bob8989 joined #git
00:36 jccn joined #git
00:37 Guest21558 joined #git
00:40 raynold ahh it's a wonderful day
00:40 zoph correct
00:40 Derperperd joined #git
00:42 ojacobson There's a theorem, whose name I won't repeat here, that strongly suggests that it's either always a wonderful day somewhere on the surface of the earth, or not a day at all
00:42 Doginal joined #git
00:42 MrWoohoo joined #git
00:42 ojacobson assuming happiness is a vector quantity, anyways ;)
00:45 Levex joined #git
00:47 leprechau joined #git
00:47 pur3eval joined #git
00:48 ISmithers joined #git
00:48 SwiftMatt joined #git
00:48 matoro joined #git
00:50 bernardio joined #git
00:52 mischat joined #git
00:54 cdown_ joined #git
00:55 watabou joined #git
01:00 hahuang65 joined #git
01:01 ejnahc joined #git
01:02 pandbearit joined #git
01:02 AndrewAlexMac joined #git
01:02 pandabearit_ joined #git
01:04 rivarun joined #git
01:06 rwin joined #git
01:06 cdown_ joined #git
01:06 Guest77315 joined #git
01:07 rwin left #git
01:07 solrize joined #git
01:08 pandabearit_ joined #git
01:09 chachasmooth joined #git
01:10 Sasazuka joined #git
01:13 selckin joined #git
01:13 Hello71 joined #git
01:14 rivarun joined #git
01:17 metachr0n joined #git
01:19 nettoweb joined #git
01:20 artgoeshere joined #git
01:22 pR0Ps joined #git
01:25 reverse_light joined #git
01:26 ozmage__ joined #git
01:27 ozmage__ joined #git
01:28 Sasazuka joined #git
01:28 clemf joined #git
01:28 mda1 joined #git
01:31 jimklo_ joined #git
01:33 cdown joined #git
01:35 thierryp joined #git
01:38 tang^_ joined #git
01:40 pilgo joined #git
01:40 jameser joined #git
01:40 Sasazuka_ joined #git
01:41 pilgo Hi all.
01:41 pilgo I’ve got a project and inside it, I have src/plugins
01:41 pilgo That’s a submodule
01:42 pilgo It always shows up as a changed file and I normally commit it
01:42 arescorpio joined #git
01:43 matoro joined #git
01:43 pilgo Right now it’s showing up as modified:   src/plugins (untracked content)
01:43 Vinnie_win joined #git
01:45 pilgo `git submodule update` gives me fatal: Needed a single revision\ Unable to find current revision in submodule path 'src/plugins'
01:45 pilgo Any ideas on how to sort this out?
01:46 thiago joined #git
01:47 Vampire0_ joined #git
01:48 pur3eval joined #git
01:49 ojacobson cd src/plugins && git status
01:51 gelei007 joined #git
01:51 pilgo ojacobson: http://dpaste.com/1YCCFR6
01:52 thiago joined #git
01:53 pilgo `git br` returns nothing
01:54 kaiten02 joined #git
01:54 mischat joined #git
01:54 pilgo I have a Gitup (visual git tool I use mostly for commit and viewing history) open in the plugins dir and I’m seeing the history fine. It does tell me that I’m detached from HEAD and on some commit, though
01:55 olbat joined #git
01:55 madduck joined #git
01:55 madduck joined #git
01:55 Darren_ joined #git
01:57 rmichnik joined #git
01:57 ToBeCloud joined #git
01:58 watabou joined #git
01:59 pilgo ojacobson: Ideas?
01:59 nickabbey joined #git
02:00 ojacobson The report in the outer repo that src/plugins has untracked content is correct
02:00 ojacobson _why_ it has untracked content is something I would have expected you to know - it's on your disk :)
02:00 rmichnik left #git
02:04 nettoweb joined #git
02:05 rossome joined #git
02:08 joshszep joined #git
02:09 joshszep joined #git
02:10 netj joined #git
02:10 joshszep joined #git
02:11 joshszep joined #git
02:11 Sasazuka joined #git
02:15 peeps[lappy] joined #git
02:17 cyphase joined #git
02:18 psyb0t joined #git
02:18 finalbeta joined #git
02:19 pilgo ojacobson: How do I get my submodule back?!
02:21 svm_invictvs joined #git
02:22 ojacobson It's right where you left it
02:22 ojacobson The uncommitted change is "you made a bunch of changes to the submodule's files" - in this case creating a bunch of new files
02:23 pilgo But when I got to plugins and git br, none of the branches show up
02:24 chriskos joined #git
02:24 ojacobson Try `git branch --all` - a submodule is, in a lot of ways, a normal clone, so it has its own suite of branches and remotes
02:25 SwiftMatt joined #git
02:25 jameser joined #git
02:26 pilgo ojacobson: nope
02:26 shgysk8zer0_ joined #git
02:26 pilgo Someone it’s not the submodule at all
02:26 pilgo as if I deinited it
02:28 gelei007 joined #git
02:30 thiago even if you deinit'ed it, the parent module still thinks it's a submodule and will not go inside that dir
02:30 thiago you need to git rm on the parent too
02:33 skinkitten joined #git
02:34 durham_ joined #git
02:35 shgysk8zer0_ joined #git
02:37 Vampire0 joined #git
02:39 skinkitten_ joined #git
02:39 duderono_ joined #git
02:46 chriskos joined #git
02:46 sammyo joined #git
02:48 ilbot3 joined #git
02:48 Topic for #git is now Welcome to #git, the place for git help and a bag of wheeee | Public logs at http://jk.gs/glog | First visit? Read: http://jk.gs/git | Current stable version: 2.11.0 | Getting "cannot send to channel"? /msg gitinfo .voice | Are you even reading this? Prove it by submitting to https://madeitwor.se/git-jokes
02:52 duderonomy joined #git
02:52 cdg joined #git
02:52 cqi joined #git
02:53 glspi joined #git
02:53 Noldorin joined #git
02:55 pks joined #git
02:55 skinkitten joined #git
02:55 mischat joined #git
02:56 fhackdroid joined #git
02:56 fstd_ joined #git
02:58 d^sh joined #git
02:59 watabou joined #git
02:59 rchavik joined #git
03:02 ojdo joined #git
03:02 rivarun joined #git
03:03 basiclaser joined #git
03:04 Levex joined #git
03:04 gelei007 joined #git
03:05 ccyw9LmF5IC2KnsU joined #git
03:09 watabou joined #git
03:09 skinkitten joined #git
03:13 dreiss joined #git
03:15 aavrug joined #git
03:15 TSS_ joined #git
03:16 aavrug joined #git
03:17 Vortex34 joined #git
03:18 Andrew_K joined #git
03:18 jameser joined #git
03:20 durham joined #git
03:20 durham joined #git
03:21 NullableTruth joined #git
03:26 safe joined #git
03:26 jimsio joined #git
03:27 ahr3n joined #git
03:28 plugwash joined #git
03:28 plugwash can I just hardlink the contents of objects/pack from an existing git repo to a new one?
03:29 kadoban plugwash: If you clone on the same fs, git will do it automatically, if that's what you're trying to do.
03:31 plugwash The thing is I don't want to clone everything, I want the commits but not the branches or tags
03:31 ikelso joined #git
03:32 TSS_ joined #git
03:32 gugah joined #git
03:36 navidr joined #git
03:37 pur3eval joined #git
03:37 thierryp joined #git
03:38 kadoban Sounds a lot like a clone and then minor cleanup.
03:41 sunri5e joined #git
03:45 chriskos joined #git
03:47 plugwash I guess I could do it that way round...
03:47 hexagoxel joined #git
03:48 gelei007 joined #git
03:48 Vinnie_win_i joined #git
03:49 govg joined #git
03:53 d1z joined #git
03:56 chachasmooth joined #git
03:56 mischat joined #git
03:58 ikelso joined #git
04:05 nevodka joined #git
04:05 tilerendering_ joined #git
04:06 mizu_no_oto joined #git
04:12 rivarun joined #git
04:13 howdoi joined #git
04:14 NeverDie joined #git
04:14 nd_ joined #git
04:15 daurnimator hrm. I'm trying to re-merge in a subtree and I'm getting fatal: refusing to merge unrelated histories
04:17 solenodic joined #git
04:17 plugwash If you want to merge unrelated histories you have to use a special option
04:18 Goplat joined #git
04:18 daurnimator plugwash: yes but that options is for  'git merge'. I don't know how to pass it to 'git subtree pull'
04:18 a_thakur joined #git
04:19 a_thakur joined #git
04:22 ccyw9LmF5IC2KnsU --allow-unrelated-histories ?
04:22 ccyw9LmF5IC2KnsU http://stackoverflow.com/questions/37937984​/git-refusing-to-merge-unrelated-histories
04:23 LeBlaaanc joined #git
04:23 ccyw9LmF5IC2KnsU oh nvmd that was the pull one you were talking about
04:25 daurnimator this appears to be the same issue I'm hitting: https://stackoverflow.com/questions/​39281079/git-subtree-error-fatal-ref​using-to-merge-unrelated-histories
04:28 ccyw9LmF5IC2KnsU yeah im seeing a few with the exact same issue https://public-inbox.org/git/1BED74C8-2​E4C-4CA7-B785-F0666B69C1A7@gmail.com/T/
04:28 gfixler joined #git
04:28 mehola joined #git
04:29 mehola joined #git
04:31 daurnimator eek, that mail web interface confuses the hell out of me
04:32 ccyw9LmF5IC2KnsU yeah that was an ugly one
04:32 jon-mac joined #git
04:33 Bizkit joined #git
04:37 ozmage joined #git
04:37 Cabanoss- joined #git
04:38 daurnimator hrm. however that was apprently fixed in 2.10?
04:38 daurnimator but I'm on 2.11 with a similar issue
04:39 ccyw9LmF5IC2KnsU yeah im on 2.11.1 and the man page doesn't mention --allow-unrelated-histories
04:41 Emperor_Earth joined #git
04:42 cagmz joined #git
04:44 patarr joined #git
04:46 ozmage joined #git
04:46 chriskos joined #git
04:49 Darren_ joined #git
04:50 sbulage joined #git
04:51 daurnimator hrm. how can I remove a commit sitting behind a merge?
04:52 daurnimator A => B => M <= C.  I want to delete B.
04:53 pur3eval joined #git
04:53 daurnimator if I do 'git rebase -i' and do a squash (my normal approach for such things) then it removes the merge
04:54 ojacobson daurnimator: you'll probably have to redo the merge, but try --preserve-merges first
04:54 ojacobson If M had conflicts, you'll run into problems, but it's worth trying
04:54 Milos joined #git
04:54 Zarthus there's always the safe git-revert, if by "remove" you don't mind it sitting in the history.
04:55 daurnimator Zarthus: no the point is to remove it from history
04:55 daurnimator (B was a temporary commit to get around a git bug... I'm hoping to now get rid of it)
04:55 matoro joined #git
04:57 mischat joined #git
04:57 xall joined #git
05:00 Muzer joined #git
05:02 chriskos joined #git
05:04 eroux joined #git
05:06 Levex joined #git
05:07 Cabanossi joined #git
05:08 gelei007 joined #git
05:08 SwiftMatt joined #git
05:10 ketas- joined #git
05:10 jimsio joined #git
05:10 Xethron joined #git
05:10 cebor joined #git
05:11 demize joined #git
05:11 j416 joined #git
05:12 jameser joined #git
05:14 leonarth joined #git
05:16 duderonomy joined #git
05:18 nicolle joined #git
05:23 daurnimator is it possible to sqash two merges together?
05:24 daurnimator ==> I used a merge strategy (subtree) that only allows 2 parents. but I'd like to merge multiple branches. so i'm wondering if I can just do multiple merges and then squash them together
05:27 jameser joined #git
05:28 chipotle joined #git
05:29 Levex joined #git
05:31 eroux joined #git
05:35 Nilesh_ joined #git
05:38 chriskos joined #git
05:38 amagawdd joined #git
05:39 thierryp joined #git
05:40 gelei007 joined #git
05:41 amagawdd joined #git
05:44 xall_ joined #git
05:52 robotroll joined #git
05:53 eroux joined #git
05:58 mischat joined #git
06:02 dec0n joined #git
06:03 sbulage joined #git
06:03 nicolle joined #git
06:08 solenodic joined #git
06:08 pur3eval joined #git
06:11 rscata joined #git
06:19 ayogi joined #git
06:23 nckpz joined #git
06:23 bocaneri joined #git
06:24 jameser joined #git
06:25 pietv joined #git
06:26 Jason joined #git
06:26 Jason hi all!  so i've got a question: i gpg sign my commits.  but then when I `git format-patch ...`, my gpg signature does not seem to be present.  why is this, and how can my signature be included?
06:27 grawity it's not added because it applies to the _commit object_
06:27 grawity i.e. the internal Git structures
06:27 grawity you cannot directly copy it on top of a patch
06:28 amagawdd joined #git
06:28 Jason grawity: ahh. so this leads to another question.  i don't have direct push access to the repo, and so i'm sending patches to another maintainer and they will commit for me.
06:28 MineCoins joined #git
06:28 grawity if you want to sign emailed patches, it's best to use PGP/MIME for that, which needs to be done by the mail app itself
06:28 Jason is there no way to ensure that my commits are gpg signed?
06:28 grawity no there isn't
06:28 grawity because they're not *your* commits
06:29 grawity they're still committed by that other person
06:29 grawity (they have you as the 'author' but other dev as 'committer')
06:29 Jason you're right -- i don't know why i didn't think about that.  i must be tired.
06:29 Jason thank you much grawity :-P
06:29 grawity Jason: so what I'd suggest is email a pull request instead
06:29 grawity i.e. push to your own repo, and tell the dev to pull from it
06:29 Jason right.  i'm using format-patch for that, getting a little patch file, and shooting it over
06:29 Jason oh?
06:29 Jason that is indeed a good idea
06:29 grawity similar to how linux.git works
06:30 Jason yes, that is a good idea.
06:30 grawity you can even do it automatically using `git request-pull` (that's the original way of doing "pull requests" before github got to it)
06:30 Jason right.  so basically host my own repo, and have the maintainers of the other repo pull my commits
06:30 grawity yes
06:30 Jason that sounds like a great idea.
06:30 Jason thank you much grawity
06:31 grawity so basically a) push to a branch, tell someone to `git pull <url> <branch>`
06:32 grawity or b), create a signed tag (e.g. for-upstream-20170230), push that tag, and again tell someone to `git pull <url> <tag>`
06:33 Jason grawity: tags == branches?
06:33 Jason i'm researching it and it looks like tags are for releases, and branches are for WIPs
06:33 Jason essentially in my use case i'm contributing to a project but cannot commit to the repo itself
06:34 Jason having my own repo and the maintainer doing pull requests sounds like the best idea
06:34 grawity tags are slightly different, yes
06:34 grawity but as far as `git pull` is concerned, all refs can be merged all the same
06:34 grawity and signed tags for PRs are a recent linux.git invention
06:34 Jason nod
06:35 grawity (since devs there rarely sign individual commits, but they _do_ want to sign these merge requests)
06:36 grawity so well, use whichever works for you
06:36 ojacobson A tag is a stable name for a commit, a branch is a movable name for a commit (by default)
06:36 ojacobson _generally_ a branch represents a monotonically-expanding set of commits (the history of that branch) but even that's subject to workflow conventions :)
06:36 grawity yeah; they mainly use tags since Git already knows how to sign and annotate them
06:37 Jason so a tag would basically be for a release
06:37 grawity in the usual non-pull-request use of tags, yes
06:38 chriskos joined #git
06:38 ojacobson yeah; the tag model for pull requests in linux.git is largely so that the original dev can continue working on the branch without disturbing the pull request
06:39 ojacobson (and so that the request can be authenticated)
06:39 Raging_Hog joined #git
06:40 gelei007 joined #git
06:41 chipotle joined #git
06:42 staafl Why does 'git rebase' ignore merges? Can't it just take the diff to the first parent and proceed as normal?
06:43 grawity it doesn't ignore merges, it flattens them
06:43 grawity and for `git rebase` that sounds like a feature tbh
06:45 staafl This guy claims "Merge-preserving rebase is willing to replay (some) merge commits, whereas normal rebase completely ignores merge commits."
06:45 staafl http://stackoverflow.com/questions/15915430/what-e​xactly-does-gits-rebase-preserve-merges-do-and-why
06:45 patarr joined #git
06:45 staafl I guess I need to check what happens myself to understand
06:46 Darren_ joined #git
06:47 ojacobson staafl: the problem is conflicts
06:47 ojacobson re-resolving the merge by reperforming it is easy enough, _but_
06:47 ojacobson re-resolving the merge by re-performing it, then transplanting the conflict resolution, rarely works
06:48 ojacobson changes in the underlying context make "just reuse the first-parent diff" risky, too
06:48 minn joined #git
06:51 ThomasLocke joined #git
06:51 ThomasLocke joined #git
06:52 brent__ joined #git
06:52 zeroed joined #git
06:56 staafl ojacobson, so throwing it away is safer? :-)
06:57 qt-x joined #git
06:57 ojacobson You could make a case for having rebase refuse to operate on commit subgraphs containing merges. (I think you can even make rebase do that, if you like, but don't quote me on it)
06:58 Jason so i have my two remotes now: my (more recent) repository, and the older group repository.  how does the group repository accept my changes from my commits in my newer repository?
06:58 ojacobson I don't have a strong opinion, personally; the behaviour it has is well-specified and no harder to remember than any of Git's numerous other idiosyncrasies
06:58 overlord_tm joined #git
06:59 staafl ojacobson, I don't really understand why a merge commit should be any more risky in terms of applying one more diff in a sequence
06:59 mischat joined #git
06:59 staafl That is, the first parent diff
06:59 staafl From the point of view of the code that's being constructed, it's just another set of changes being applied
07:00 chriskos joined #git
07:00 ojacobson "risk" may have been the wrong metaphor
07:01 ojacobson it's complex to do and it's not clear if any specific strategy will produce more-correct or more-intuitive results than any other
07:01 ojacobson blindly using the first-parent diff is almost certainly the wrong choice, especially if the other arms of the merge are also in the set of commits being rebased
07:01 amagawdd joined #git
07:01 ojacobson (this problem is hard enough with git-filter-branch, which already does more generalized processing of merges)
07:02 ojacobson re-performing the merge is an option (--preserve-merges, which even tries to reconstruct rebased merges from branches included in the rebase set)
07:02 ojacobson Make an alias for rebase that uses it and live with it for a while, see if it works out for you :)
07:03 JeroenT joined #git
07:04 JeroenT joined #git
07:04 staafl ojacobson, I'm actually trying to figure out a simple workflow to explain to people at my company
07:05 staafl I.e. when can you rebase on an upstream
07:05 staafl It seems you must not have performed any merges since the upstream or the rebased history will miss the merged changes
07:06 ojacobson github-flow is often a good starting point; it's easy to say "you can rebase your features but not any shared branches" if only small numbers of people work on any single feature, or if the team decomposes features to single-person tasks
07:06 staafl IMO this is too complicated to watch out for
07:06 ojacobson merging with upstream and rebasing on upstream are… more or less mutually-exclusive workflow choices. Mixing them requires a fairly solid understanding of git's behaviour.
07:07 staafl ojacobson, so what if you need to merge master into your feature branch for synchronizing
07:07 ojacobson merging to integrate downstream (eg. to bring a feature back into its parent branch) and rebasing to integrate upstream (to bring changes from the parent branch into a feature in progress) is pretty common
07:07 ojacobson "just use merge" is also workable, if you don't mind somewhat abstruse commit graphs
07:07 staafl ojacobson, obviously you can't rebase on master
07:07 ojacobson Why not?
07:07 staafl because you'll probably be rebasing some published history
07:08 ojacobson If your feature diverges from master at some point in the past and is linear subsequently, as is fairly conventional, `git rebase origin/master` will recreate the same linear branch, diverging on a newer commit from master
07:08 staafl If you've ever pushed your feature branch, master..feature/X contains published commits
07:08 ojacobson (I'm presupposing a git-flow-like or github-flow-like ontology of branches, if that doesn't apply or if you don't want to use it, ignore me :)
07:08 staafl ojacobson, that's what we use
07:08 ChiLL_SurF joined #git
07:08 staafl but note my last comment
07:09 ojacobson rebasing and force-pushing features is somewhat widespread for github-flow-like workflows
07:09 staafl ojacobson, ugh
07:09 ojacobson it works fine provided most of the interaction with un-merged features is read-only
07:09 staafl So only one person ever checks out a feature branch
07:09 ojacobson Only one person ever creates new commits on an un-merged feature branch
07:09 ojacobson That requires a level of planning and task division that some teams take to and some teams don't
07:09 staafl ojacobson, well that's not acceptable here
07:09 ojacobson yeah
07:10 staafl so looks like you either don't sync with master or don't rebase
07:10 ojacobson if you have groups working together on long-lived branches, starting from "don't rebase" is a good idea
07:10 DieguezZ joined #git
07:10 ojacobson (or at least "don't force-push")
07:11 ojacobson _someone_ should learn how to repair the consequences of rebasing something previously-published - probably by undoing the rebase, and transplanting any new commits back to the un-rebased state
07:11 ojacobson because a "no-rebase" guideline is an invitation for some bright spark who Knows Better to make a real mess
07:11 staafl ojacobson, unfortunately I'm doing a git training today and the whole situation is quite complicated to explain to newbies
07:11 staafl In a memorable way
07:12 zeroed joined #git
07:12 zeroed joined #git
07:12 ojacobson git's UI is incredibly hostile, surprising nobody
07:12 ojacobson _powerful_, but in no way ergonomic
07:13 ojacobson fortunately? the underlying model isn't actually all that complicated, it can be explained in full in a handful of bullet points
07:13 ojacobson unfortunately, all the sophistication ends up in the brains of the users, instead
07:13 ojacobson (it has to live somewhere, and it's not in git(1))
07:13 chele joined #git
07:14 osse some of it is in gitglossary
07:14 staafl ojacobson, indeed
07:14 staafl Thanks for the discussion, that was helpful
07:15 osse (I prefer git's approach of one manpage per subcommand plus a handful extra)
07:17 aw1 joined #git
07:18 GodGinrai joined #git
07:20 SwiftMatt joined #git
07:20 Guest9 joined #git
07:21 sea-gull joined #git
07:24 elect joined #git
07:26 sekire joined #git
07:27 staafl ojacobson, one question
07:27 ojacobson hm?
07:28 staafl ojacobson, why is force pushing ok if no one else commits to the branch?
07:28 staafl They'll still have a bogus merge when they pull
07:29 ojacobson Generally, don't use pull (or merge, equivalently) to review other peoples' branches, in github-flow-like workflows. Only use it to integrate incoming features into their target branches.
07:29 staafl Actually N bogus merges for different people
07:29 ojacobson Use `checkout` to visit the branch temporarily, and throw away the result
07:29 staafl Great, more tricky rules
07:29 ojacobson That way the lateral movement of the branch tip is irrelevant: you're never merging it unless you plan to keep the branch in that state forever
07:29 ojacobson (by integrating it upstream)
07:30 ojacobson This is, basically, what github's big green merge button does anyways :)
07:30 staafl ojacobson, yeah, but I think I'll recommend avoiding rebase at my company
07:31 nckpz joined #git
07:31 Levex joined #git
07:31 staafl Too easy for a sleepy dev to tangle things up
07:31 ojacobson Sounds good
07:31 blahdodo joined #git
07:31 ojacobson you can always triangulate on a workflow by loosening or tightening the recommended rules, too
07:32 staafl If people have the patience to adapt
07:32 a_thakur joined #git
07:34 thierryp joined #git
07:34 jnavila joined #git
07:35 jnavila joined #git
07:37 a_thakur_ joined #git
07:37 osse what you can't do is detect whether someone has rebased locally \o/
07:38 a_thakur_ joined #git
07:41 staafl True :-)
07:42 ojacobson I mean, you mostly also don't _care_
07:42 ojacobson if someone rebases locally and you never noticed, there's no distinction between that and writing the rebased commits in the first place
07:42 ojacobson (you can detect most rebases by comparing author timestamp and commit timestamp, though)
07:43 staafl A tangential question, is a three-way-merge of two commits equivalent to rebasing one on the other in terms of patch application?
07:43 Wulf joined #git
07:43 staafl I.e. are rebase or merge equivalent, or is one more likely to get into conflicts/erroneous resolutions in some cases?
07:43 osse If there are no conflicts, yes
07:44 staafl osse, as in, mathematically equivalent - you get the same result with both algorithms?
07:45 staafl And what about conflicts, is there an advantage to either
07:45 sargentmki joined #git
07:46 osse staafl: If two changes don't overlap during a merge then Git will apply both. Repeat. This should be the same as you rebased one branch on top of the other
07:47 staafl osse, what about the same change applied in both branches?
07:47 staafl Is there a pathological case where you'll get two applications in rebase?
07:48 osse With rebase you might have to resolve the "same"  conflict several times. And there may be changes that conflict temporarily but are  changed later, in which case rebase would give you a conflict and merge wouldn't
07:49 osse staafl: If both branches have the same change then git's job just becomes easier during a merge. In a rebase git will detect that a the change is already applied and will drop it.
07:49 ojacobson staafl: I don't think there's a formal proof but, for example, darcs (another SCM system) builds a formal patch algebra around the assumption that patch sequences are equivalent to the diff of the endpoints and seems to work wll
07:49 ojacobson well
07:50 ojacobson intuitively, a sequence of patches should produce the same ultimate result
07:50 ojacobson from there, concluding that rebase and merge are equivalent as to the final state of the tree is pretty straightforward (though osse is right about conflicts)
07:54 pR0Ps joined #git
07:55 ihradek joined #git
07:55 zeroed joined #git
07:55 zeroed joined #git
07:56 ivohradek joined #git
07:57 pur3eval joined #git
07:57 staafl Thanks guys :-)
07:59 Alienpruts joined #git
07:59 Balliad joined #git
08:00 mischat joined #git
08:01 gelei007 joined #git
08:07 ysh joined #git
08:07 chipotle joined #git
08:10 tterrag|alt left #git
08:11 mkoskar joined #git
08:11 theoceaniscool joined #git
08:14 pR0Ps joined #git
08:19 TomyWork joined #git
08:22 freimatz joined #git
08:22 pR0Ps joined #git
08:23 SwiftMatt joined #git
08:26 troulouliou_div2 joined #git
08:28 Raed joined #git
08:34 sadme joined #git
08:35 sadme hi, does anyone know how i can tell git that my file is not a symlink?
08:35 osse sadme: what's the use case?
08:36 osse sadme: you cannot have a symlink in a repo and make git pretend it's a file
08:36 sadme no, i have a regular file in a repo that git thinks is a symlink
08:36 osse :O
08:37 osse can you show me  ''ls -l file'' and ''git ls-tree HEAD -- file''
08:39 sadme i should note that i am currently using a windows machine. http://pastebin.com/fA444yMq
08:39 osse sadme: git status ?
08:40 sadme oh, status says " typechange: jquery.js" i haven't seen that before
08:40 osse \o/
08:41 watabou joined #git
08:41 sadme thanks, that seems to have done the trick
08:45 mozzarel1 joined #git
08:46 Dumblez0r joined #git
08:46 patarr joined #git
08:48 Raging_Hog joined #git
08:50 sostenuto joined #git
08:52 aidalgol joined #git
08:54 chipotle joined #git
08:56 elect joined #git
08:57 staafl joined #git
08:57 Derperperd joined #git
08:59 Raging_Hog joined #git
08:59 joshszep joined #git
08:59 aw1 how do i get "git rev-parse" to show the current branch and the SHA1 of the HEAD with a single invocation? "git rev-parse --abbrev-ref HEAD" shows the branch and "git rev-parse --short HEAD" gives me the SHA1 .. can't seem to combine the two in one invocation
08:59 ysh joined #git
09:02 Ovsyanka joined #git
09:03 a-l-e joined #git
09:03 MineCoins joined #git
09:03 solenodic joined #git
09:04 overlord_tm joined #git
09:07 algun_ joined #git
09:07 sea-gull joined #git
09:07 mikecmpbll joined #git
09:07 a-l-e joined #git
09:07 brent__ joined #git
09:08 ac3takwas joined #git
09:08 bongjovi joined #git
09:09 miczac joined #git
09:09 Anticom joined #git
09:10 Shashikant86 joined #git
09:10 osse There doesn't seem to be a way
09:11 osse you can parse "git log -1 --format='%H %D'" but then you might as well run two rev-parse
09:11 osse if you care about performance then maybe symbolic-ref HEAD is better to get the name
09:16 thierryp joined #git
09:19 GodGinrai joined #git
09:20 mda1 joined #git
09:21 TomyWork joined #git
09:22 marcogmonteiro joined #git
09:22 afuentes joined #git
09:26 DBag joined #git
09:29 kasual joined #git
09:30 Levex joined #git
09:32 bjoe2k4 joined #git
09:32 Dandelock joined #git
09:33 Tobbi joined #git
09:33 madduck joined #git
09:33 freezer joined #git
09:35 mda1_ joined #git
09:35 warthog9 joined #git
09:36 dsdeiz joined #git
09:36 dsdeiz joined #git
09:38 madduck joined #git
09:42 nettoweb joined #git
09:44 orzo joined #git
09:44 borkr joined #git
09:44 CustosLimen hi
09:45 mitzip joined #git
09:45 CustosLimen if I clone a git repo with submodules using --recursive on clone - and the submodule commit changes - should git pull get the new submodule commit ?
09:45 miczac joined #git
09:45 CustosLimen or do I have to do something different with git pull ?
09:46 Serpent7776 joined #git
09:46 pur3eval joined #git
09:46 osse CustosLimen: git submodule update
09:47 osse (after pull, checkout, etc)
09:47 CustosLimen ok - can I do this as part of updated with config ?
09:47 CustosLimen as part of pull I mean ?
09:47 chipotle joined #git
09:48 MattMaker joined #git
09:48 Tobbi joined #git
09:49 osse CustosLimen: AFAIK it's common to add 'git submodule update' to the post-checkout and post-merge hooks
09:49 CustosLimen there is fetch.recurseSubmodules
09:49 King_Hual joined #git
09:49 King_Hual joined #git
09:49 Vgr_ joined #git
09:49 Vgr_ joined #git
09:50 wisefool joined #git
09:50 dhollinger joined #git
09:50 Vgr_ joined #git
09:50 Vgr_ joined #git
09:51 quizzi joined #git
09:51 sadme left #git
09:53 ngm joined #git
09:55 watabou joined #git
09:56 hardik joined #git
09:57 Dumblez0r joined #git
09:58 diogenese joined #git
09:59 ocbtec joined #git
10:00 Serpent joined #git
10:02 masuberu joined #git
10:02 solenodic joined #git
10:07 palash joined #git
10:13 jameser joined #git
10:14 blackwind_123 joined #git
10:15 achadwick joined #git
10:18 ferr1 joined #git
10:19 sitaram joined #git
10:20 Raging_Hog joined #git
10:21 theoceaniscool joined #git
10:21 solenodic joined #git
10:23 manuelschneid3r joined #git
10:24 rnsanchez joined #git
10:25 Snugglebash joined #git
10:28 mehola joined #git
10:30 mehola joined #git
10:33 mischat joined #git
10:33 Shashikant86 joined #git
10:34 mischat joined #git
10:36 chele joined #git
10:36 gechr joined #git
10:40 marcogmonteiro joined #git
10:41 thejoecarroll_ joined #git
10:45 jameser joined #git
10:47 patarr joined #git
10:48 Anticom joined #git
10:49 kristofferR joined #git
10:50 marlinc joined #git
10:51 tvw joined #git
10:52 blokeybloke joined #git
10:52 cdown joined #git
10:54 covey joined #git
10:58 Derperperd joined #git
10:59 Spuri0us joined #git
11:01 tvw joined #git
11:02 t0by joined #git
11:02 t0by joined #git
11:06 chele_ joined #git
11:06 chele joined #git
11:07 notebox joined #git
11:07 Shashikant86 joined #git
11:08 brent__ joined #git
11:09 Levex joined #git
11:09 irqq joined #git
11:10 jjthegerman joined #git
11:10 SwiftMatt joined #git
11:11 cdown_ joined #git
11:11 Tobbi joined #git
11:11 nattyg93 joined #git
11:12 bongjovi joined #git
11:14 staafl joined #git
11:14 nattyg93 joined #git
11:15 nattyg93 joined #git
11:15 Bizkit joined #git
11:15 watabou joined #git
11:16 overlord_tm joined #git
11:16 dviola joined #git
11:17 _ADN_ joined #git
11:17 jjthegerman left #git
11:18 chachasmooth joined #git
11:18 JJG joined #git
11:18 nattyg93_ joined #git
11:20 GodGinrai joined #git
11:20 JJG left #git
11:20 mda1 joined #git
11:21 JJG joined #git
11:21 JJG left #git
11:21 rivarun joined #git
11:21 JJtG joined #git
11:21 nattyg93_ Hello
11:21 gitinfo nattyg93_: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.
11:22 laxman joined #git
11:22 nattyg93_ I've run into an issue which I can't seem to find the answer to. I was hoping someone here could help me.
11:23 JJtG Hi there guys first time here and I would have a question. I have been googling and trying to fix it for the past 2 hours with no results thus far.  Im trying to undo every commit, every change, simply everything that is in any way shape or form changed from what is on the server
11:23 osse JJtG: git fetch; git reset --hard origin/master
11:23 JJtG all the command like git reset --hard and git clean -f -d and whatnot that people in forums suggested did not work
11:24 selckin define did not work
11:24 JJtG osse, if I run git status it says your branch origin/master have diverged and have 3 and 20000 different commits each
11:24 canton7 in that case, you're going to have to explain exactly what makes you think they didn't work - what you expected, how the outcome was different - otherwise we're just going to be repeating what you've already read
11:25 canton7 JJtG, what does git say when you run 'git reset --hard origin/master'? Can you pastebin the output?
11:25 jameser joined #git
11:26 JJtG HEAD is now at "hash" merge branch "master"
11:26 reynir joined #git
11:26 reynir How do I diff a file F1 on branch B1 with a file F2 on branch B2 in git?
11:26 osse now git status again
11:26 osse reynir: git diff B1:F1 B2:F2
11:26 canton7 JJtG, maybe you have a local branch called 'origin/master' - it's easy to create one by mistake. What about 'git reset --hard refs/remotes/origin/master' ?
11:26 reynir osse: Awesome! Thank you
11:27 mischat joined #git
11:27 JJtG osse, same as before "origin/release" have diverged ...
11:27 JJtG canton7, I will try it, one sec
11:27 osse JJtG: don't lie. you said origin/master earlier, now it's origin/release :O
11:27 canton7 JJtG, ('git branch' will list 'origin/master' if you do have a local branch called 'origin/master')
11:27 osse JJtG: git reset --hard origin/release then
11:28 canton7 and yeah - what's going on with master vs release here?
11:28 JJtG oh yeah sorry that was a typo from me, its says origin/release
11:28 canton7 JJtG, wait, step back
11:28 canton7 JJtG, what's the name of the branch you're currently on?
11:29 canton7 what's the name of the branch you want to make it the same as? do you want to make it the same as origin/release, or origin/master?
11:29 JJtG good question, let me google the command for that :P ( i have no clue what I am doing )
11:29 canton7 'git status' tells you what branch you're currently on
11:29 canton7 (and 'git branch' will as well)
11:29 JJtG im on release branch
11:29 nattyg93_ Hi guys, I have a weird issue which I haven't been able to figure out. I'm pretty new to git. I'm running git from the command line on my linux box on the bash shell. When I create two new branches (based off master), say, b1 and b2. I then switch to b1, create a new directory, say, temp and add a new file in temp, say temp.b1 and commit it. Now I switch to b2 and add a directory with the same name (temp) and add a new file, say, temp
11:30 canton7 JJtG, ok, so you're on release. You want to make it look the same as the release branch on origin?
11:30 mehola joined #git
11:30 nattyg93_ Now if I switch back to b1, and type "ls -a" nothing shows up not even ". .." like is normal for an empty directory.
11:30 JJtG I assume so, I dont do any code changes whatsoever but somehow my repo still changed from the servers
11:31 canton7 JJtG, ok, so 'git reset --hard origin/release' to make your 'release' branch look the same as the 'release' branch on 'origin'
11:31 selckin nattyg93_: can you post a transcript (copy paste from terminal) to show the problem
11:31 laxman joined #git
11:32 JJtG canton7, nice now git status says its clean
11:32 canton7 JJtG, cool
11:32 laxman Hi Guys
11:32 gitinfo laxman: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.
11:32 lb1c joined #git
11:32 canton7 JJtG, so you got confused between the 'release' branch and the 'master' branch
11:32 JJtG canton7, I guess thats that. Thanks for the help
11:32 JJtG osse, ty to you too
11:32 nattyg93_ selckin yeah, sure. just give me a moment.
11:33 Taikibansei joined #git
11:33 laxman fatal: Unable to find remote helper for 'https'
11:33 laxman i geeting this when iam pushing to bit bucket
11:33 jost__ joined #git
11:33 JJtG canton7, I guess so. I dont i dont work with git but p4v, but need git in order to compile latest. But it can be confusing sometimes
11:33 laxman please help me guys
11:34 osse JJtG: sorry for calling you a liar
11:34 osse i was half-kidding
11:34 JJtG osse, lol, I dont care :D
11:34 canton7 laxman, what OS?
11:34 laxman CentOS 6.8
11:35 pur3eval joined #git
11:35 canton7 laxman, apparently you need to install curl-devel
11:35 canton7 laxman, did you build git from source, or install using a package?
11:36 yooy joined #git
11:36 laxman installed canton7
11:36 laxman still not working ?
11:36 canton7 laxman, weird - I was reading stackoverflow.com/questions/8329485/unable-to-​find-remote-helper-for-https-during-git-clone, but it's all about building from source
11:37 canton7 maybe there's a separate git-http package of some sort you need?
11:37 canton7 laxman, does git-remote-https exist in /usr/libexec/git-core ?
11:38 canton7 laxman, looks like you might need to uninstall/reinstall git (now that you've installed curl-devel)
11:38 mda1_ joined #git
11:39 jameser joined #git
11:39 laxman Yup .It is there canton7
11:39 cyphaw joined #git
11:39 solenodic joined #git
11:40 canton7 laxman, is it your machine, or a VPS?
11:40 cyphaw Hi, i have a question about git
11:40 nattyg93_ selckin: here is the pastebin: http://pastebin.com/QEqBnbA6
11:40 laxman VPS canton7
11:40 canton7 laxman, cloud linux?
11:40 freimatz joined #git
11:41 laxman digital ocean canton7
11:41 selckin nattyg93_: cd ..; ls
11:42 selckin nattyg93_: you are in the old removed directory, not in the new directory git created
11:42 cyphaw i have some files commited (.project, .classpath) that i put in gitignore. I want to remove them from git tracking, but so that they don t get erased when someone else does a git pull
11:42 nattyg93_ Yeah, that's what I suspected.
11:42 canton7 laxman, and you definitely installed git by typing 'yum install git'?
11:43 nckpz_ joined #git
11:43 nattyg93_ selckin: if you see at the bottom there, when I run "cd ../newdir" then run ls the file shows up.
11:43 laxman Yes canto7
11:43 nckpz__ joined #git
11:43 selckin nattyg93_: yea, so there is no problem
11:43 laxman yes canton7
11:43 nattyg93_ selckin: so am I always just expected to always run something like "cd ." whenever I'm changing branches?
11:44 osse cyphaw: Unfortunately that's not possible. But it's fairly easy for anyone to resurrect them after pulling
11:44 osse And it would only be the one time
11:44 cyphaw osse: so i should git rm --cached them and tell people to save them, git pull and restore them?
11:44 selckin nattyg93_: do it from the root if you're paranoid, but thats just how the filesystem works, if the parent of both has the directly it would work
11:45 canton7 laxman, I'm out of ideas then I'm afraid
11:45 patarr joined #git
11:45 selckin nattyg93_: no in general it doesn't happen often
11:45 hhee joined #git
11:45 osse cyphaw: yes. But they can also get them from git itself without backing them up
11:45 osse cyphaw: Bur i guess they have useful changes in them...
11:46 cyphaw osse: yes, i think. thanks
11:46 laxman i installed curl-devel  after git installed  is ok  canton7 ?
11:46 nattyg93_ selckin: Ok, that makes sense. I assumed that since it's something that I've come up against in my first week of using git that it must be something that happens a lot.
11:46 canton7 laxman, I did suggest uninstalling/reinstalling git, but I don't know that it will make much difference
11:46 CustosLimen so can I put hooks into a git repo that people get when they clone it ?
11:47 CustosLimen or should hooks be setup explicitly by client ?
11:47 nattyg93_ selckin: But yeah, I guess swapping between branches expecting to end up in the new directory when git is removing the directory and adding a new one was a little naive.
11:47 canton7 CustosLimen, no, that would be a security hole. There are tools for allowing people to easier set up hooks from a set stored in the repo, or you can write a script while when run will install the hooks
11:47 nattyg93_ selckin: thank you very much for you help!
11:47 selckin nattyg93_: thats how linux works, on windows it would error not letting it remove the dir while its open
11:47 dendazen joined #git
11:47 laxman how to unstall git ?
11:47 laxman canton7
11:48 canton7 laxman, yum remove
11:48 nattyg93_ selckin: Do you know if it's specifically an ext4 thing? Or would you expect to see the same thing on other filesystems?
11:49 laxman removed .still git command working
11:49 mischat_ joined #git
11:49 laxman conton7
11:49 selckin nattyg93_: its standard, you're allowed to remove open files
11:49 laxman canton7
11:49 selckin nattyg93_: which is usually a good thing
11:49 canton7 "still git command working"?
11:49 canton7 so you removed the package, and didn't reinstall it, and git still works?
11:49 laxman yup canton7
11:49 canton7 laxman, so you have two git installations?
11:50 OMGOMG this sounds like a job for #fedora
11:50 nattyg93_ selckin: Yeah, actually, I do that all the time. I should have realised.
11:50 laxman dont know canton7
11:50 laxman two ?
11:50 laxman strange
11:50 bcc joined #git
11:50 bcc joined #git
11:51 canton7 laxman, well if you uninstalled the git package, but you can still run git commands, then obiously you still have a git installation somewhere
11:51 CustosLimen canton7, can I put hooks inside ~/.git/ ?
11:51 nattyg93_ selckin: Again, thank you for your help!
11:51 canton7 CustosLimen, in your repo? sure
11:51 CustosLimen canton7, no in my home dir
11:51 canton7 oh, sorry. No.
11:51 laxman so how remove that one only canton7
11:51 canton7 laxman, dunno - how did you install it?
11:52 canton7 CustosLimen, actually core.hooksPath exists in man git config
11:52 gitinfo CustosLimen: the git-config manpage is available at http://jk.gs/git-config.html
11:52 canton7 so I was wrong
11:52 selckin nattyg93_: np
11:53 Levex joined #git
11:53 ertesx joined #git
11:54 laxman dno .if git installed from source .then how to uninstall  canton7
11:54 laxman maybe my friend install before from source
11:55 canton7 laxman, never done it, but there's some stuff here: http://serverfault.com/questions/444289/how​-to-uninstall-git-installed-from-the-source
11:56 chipotle joined #git
11:59 blokeybloke joined #git
12:00 cris_ joined #git
12:01 solenodic joined #git
12:02 Levex joined #git
12:04 aw1 joined #git
12:06 CheckDavid joined #git
12:11 jimi_ joined #git
12:14 ivh joined #git
12:16 watabou joined #git
12:17 ivh Hi. I have a repo where master tracks an svn-repo with "git svn", and I work in feature branches that I merge with --squash. All is good, but I wonder how to continue with the branch afterwards. Should I rebase it onto the squashed commit, or merge it? Does it make sense to get the squashed commit back into the branch?
12:18 Tobbi joined #git
12:18 mizu_no_oto joined #git
12:19 reynir left #git
12:20 Snugglebash joined #git
12:21 mischat joined #git
12:22 mischat joined #git
12:22 mischat joined #git
12:23 gelei007 joined #git
12:23 treia joined #git
12:25 mischat joined #git
12:26 mischat joined #git
12:27 mischat joined #git
12:28 mischat joined #git
12:28 enkrypt joined #git
12:28 inflames joined #git
12:29 mischat joined #git
12:29 jameser joined #git
12:30 laxman Hi canton7. i taken new fresh server .in that pushing is fine .i am able push code bit-bucket .i think problem two gits in old server
12:30 mischat joined #git
12:30 notebox joined #git
12:30 manuelschneid3r joined #git
12:33 texinwien joined #git
12:33 mischat joined #git
12:34 Dumblez0r joined #git
12:34 mischat joined #git
12:35 Carlrobertoh joined #git
12:36 ruxu joined #git
12:37 mischat joined #git
12:38 manuels joined #git
12:40 mizu_no_oto joined #git
12:40 finalbeta joined #git
12:41 Carlrobertoh Hey. I have git fetch and git rebase a project but i don't have the files which are blue https://www.upload.ee/image/6658047/Untitled.jpeg
12:41 nettoweb1 joined #git
12:41 _ikke_ Carlrobertoh: master is currently checked out
12:42 ac3takwas joined #git
12:42 Carlrobertoh thank you
12:42 Carlrobertoh got it working
12:42 lordjancso joined #git
12:43 finalbeta joined #git
12:45 leeN joined #git
12:46 Lunatrius` joined #git
12:46 ruxu joined #git
12:47 re1 joined #git
12:49 Darcidride joined #git
12:49 DolphinDream joined #git
12:50 Naan joined #git
12:50 cdown joined #git
12:53 ac3takwas joined #git
12:53 DonaldJTrump joined #git
12:53 Snugglebash joined #git
12:54 Shashikant86 joined #git
12:56 ysh joined #git
12:58 UrsoBranco joined #git
12:59 Derperperd joined #git
13:00 Raging_Hog joined #git
13:00 mischat joined #git
13:01 Levex joined #git
13:03 DonaldJTrump left #git
13:04 F0rTh3J3st joined #git
13:05 cdown joined #git
13:05 Tobbi joined #git
13:06 Spuri0us joined #git
13:06 wiesel joined #git
13:08 ketas joined #git
13:08 synthroid joined #git
13:10 cdown joined #git
13:10 m4sk1n joined #git
13:13 theoceaniscool joined #git
13:13 synthroi_ joined #git
13:15 jameser joined #git
13:15 nowhereman joined #git
13:17 GavinMagnus joined #git
13:17 GavinMagnus left #git
13:18 watabou joined #git
13:19 jameser joined #git
13:19 cdown joined #git
13:20 Gord joined #git
13:20 selckin probably no, but is there an option to ignore custom diff utils that are configured ? (like diff.zip.textconv = unzip -c -a), since i'm trying to search for something with git log -S and the repo has a corrupt zip file that errors out
13:21 GodGinrai joined #git
13:23 qqx selckin: Have you tried --no-textconv ?
13:23 pur3eval joined #git
13:24 brent__ joined #git
13:24 selckin qqx: sweet, thanks!
13:25 jeffreylevesque joined #git
13:28 Raging_Hog joined #git
13:30 Dumblez0r joined #git
13:30 MineCoins joined #git
13:31 synthroid joined #git
13:41 Snugglebash joined #git
13:41 jagob joined #git
13:41 synthroi_ joined #git
13:42 GodGinrai joined #git
13:42 jorj joined #git
13:45 notebox joined #git
13:46 basiclaser joined #git
13:47 GodGinrai joined #git
13:48 sarri joined #git
13:48 sarri joined #git
13:49 samgoody joined #git
13:49 al-damiri joined #git
13:50 dendazen joined #git
13:50 samgoody Hi. I just received a repo, and a bunch of standalone files that the project owner doesn't know if they were experiments of new ideas, or old code that is obsolete
13:50 samgoody if old, it would be in old commits.
13:51 troulouliou_div2 joined #git
13:51 samgoody Is there any way to search through all committed versions of a file and see if one matches the exsting file
13:51 samgoody Or if one commit is closer to said file than current commit?
13:51 osse samgoody: it is possible via a healthy dose of scripting
13:52 ash_workz joined #git
13:52 osse but git does not have  ''git find-similar-file''  :p
13:52 brent__ joined #git
13:52 Levex joined #git
13:53 selckin just delete it, not commited, not tested, not in ci, not in use, be gone
13:54 samgoody I could, but figured if there was some way to check, thats better than grocking the code to see if it has anything new
13:54 samgoody oh, well. thanks.
13:55 sbulage joined #git
13:58 osse samgoody: git rev-list HEAD -- path/to/file | while read hash; do echo "$hash";  git diff --stat --no-index "$hash":path/to/file  ../codedump/path/to/same/file; done  # See which hash produces the lowest numbers. That hash has the file closest to your file
13:58 scoobertron joined #git
13:58 dendazen joined #git
14:01 Shashikant86 joined #git
14:06 borkr joined #git
14:06 dsdeiz joined #git
14:06 dsdeiz joined #git
14:07 cdown joined #git
14:09 Cienias joined #git
14:09 Cienias hello
14:09 gitinfo Cienias: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.
14:11 ysh joined #git
14:11 Cienias i have a local branch not pushed yet so it's not tracking any remote, i have cherry picked some changes form other branch to this one, but now if i do git status it does not show anything, i guess the chages are already commited
14:11 Cienias but is there a way to see the changes ?
14:12 sbulage joined #git
14:12 Cienias can i make all the commits unstaged again ?
14:12 heroux joined #git
14:13 publio joined #git
14:16 Tobbi joined #git
14:16 watabou joined #git
14:17 Guest77315 left #git
14:17 mgver joined #git
14:17 * mgver Ouvindo no EasyPlayer: Nenhuma mp3 em andamento!
14:18 whitby Cienias: you can see the changes by running `git show <hash>`
14:18 Cienias really ?
14:19 Cienias 'git show #'
14:19 Cienias like that ?
14:19 whitby Yes
14:19 _ikke_ hash is the commit id
14:19 _ikke_ not a pound-sign
14:19 whitby Er, yeah, what _ikke_ said
14:19 Cienias hmmm but then i need to know the commit id
14:20 _ikke_ git log
14:20 whitby What "changes" do you want to see? You can run `git diff` as well
14:20 whitby e.g. `git diff my-local-branch..master`
14:20 Gsham joined #git
14:21 Cienias i think my 2 commits from other banch were automatically commited (source tree) to my current branch an git status does not show anything cause branch is not tracked yet
14:21 nettoweb joined #git
14:22 Cienias actually can someone confirm, if branch is not tracked and you comit some changes and run git status it will not show anything right ?
14:22 thierryp joined #git
14:22 whitby Yes, `git cherry-pick` will auto-commit
14:22 whitby Yes, if you have no changes to your files locally then git status will not show anything
14:22 Cienias whitby: shame it would be nice if the changes were unstanged so you can actually see them
14:23 Cienias whitby: no i mean even with local chages it does not show
14:23 ShekharReddy joined #git
14:23 solenodic joined #git
14:23 whitby man git cherry-pick
14:23 gitinfo the git-cherry-pick manpage is available at http://jk.gs/git-cherry-pick.html
14:24 whitby There is a flag you can pass to cherry-pick to not commit it
14:24 relipse joined #git
14:25 relipse if my git says "Your branch is ahead of origin/foobar by 10 commits" how do i determine what changed locally?
14:25 whitby If you have local changes that you've committed then git status won't show 'em. That is, `git status` shows only what hasn't been committed. It doesn't care about remotes
14:25 BPL joined #git
14:25 whitby relipse: git diff foo..origin/foo
14:25 whitby Or origin/foo..foo
14:25 ShapeShifter499 joined #git
14:25 whitby One of the two, I can never remember which
14:26 whitby Cienias: from the manpage linked above, you can pass `--no-commit` to cherry-pick
14:26 Cassiopaya joined #git
14:27 tmsmith joined #git
14:32 odbh joined #git
14:32 JamesBaxter joined #git
14:32 boombatower joined #git
14:34 thierryp joined #git
14:34 Murii joined #git
14:34 Cienias btw if i have commited changes and i switch branches will those changes affect new branch ?
14:35 mindfart joined #git
14:35 _ikke_ if you switch to another existing branch, then no
14:38 themeek joined #git
14:39 remote joined #git
14:41 Ranhir joined #git
14:41 themeek left #git
14:41 Remramm joined #git
14:41 Remramm joined #git
14:42 Dougie187 joined #git
14:43 DolphinDream joined #git
14:43 DolphinDream howdy
14:43 IZh joined #git
14:44 DolphinDream i'm having this argument with a coworker about whether to add the exclude files like (*.o, *.a, *.so) to each individual repo's .gitignore or add it to the global .gitignore. What is the recommended use?
14:44 IZh Hi. I have a question, what is the best way to implement 2-stage review process workflow?
14:45 * mgedmin adds them to each repo's .gitignore
14:46 freimatz joined #git
14:46 IZh First, let's considering 1-stage. There is a master repo. When somebody needs to fix a feature, it needs to fork it, then create a feature branch (to keep master branch clean), then make a pull request, change local commits during review, and finally merge to remote master.
14:46 IZh But what abot adding one extra level?
14:47 solenodic joined #git
14:47 DolphinDream mgedmin: and your rational is?
14:48 IZh Suppose, there is a master repo A. We've forked a repo B, keeping i in sync with A. Then local developer cloned B as C, makes some changes that had passed internal review, and were merged to B. How to get it back to A?
14:48 IZh And what to do if A's reviewers ask to change some things?
14:49 DolphinDream why adding the same (development platform specific) exclude patterns to every single repos which they in fact are the same for all repos. Each individual repo can have extra exclude patters specific to that repo of course.
14:49 whitby DolphinDream: my approach is this: if everyone who uses the project will generate the files (e.g. build output, *.o, *.so) then it should be explicitly in the project's ignore. If it's Jim or Jane's IDE that nobody else uses that puts shit everywhere then they should add it to their own global ignore (e.g. .idea folders, .vscode, *.swp)
14:49 Sound joined #git
14:49 whitby It's not perfect, but it's my starting point
14:50 Tobbi joined #git
14:50 whitby IZh: what are you trying to do and why are you forcing people to fork things?
14:51 ash_workz joined #git
14:52 thierryp joined #git
14:53 mischat joined #git
14:53 accessun joined #git
14:53 IZh whitby: Consider two companies, like Google and Foo Inc. Google owns a master repo. Foo Inc makes a fork of it. Then employees of Foo Inc implementing some feature. So the code will need to pass internal Foo Inc's review. Then they want to merge changes back to Google's repo. But Google will also review changes, and could ask to change something, so the code must be changes.
14:54 texinwien_ joined #git
14:54 DBag joined #git
14:54 IZh So it's a two-stage review between 2 companies: internal and external.
14:54 IZh And what is the best way to setup workflow in this case?
14:55 whitby There isn't a
14:55 whitby "best" way
14:55 whitby It will depend heavily on the internal practices of Foo Inc.
14:56 whitby Google could say, we only accept changes via PR and Foo Inc can come about a PR how ever they like
14:56 chipotle joined #git
14:56 IZh Internal practices are: mirroring Google's repo, then creating a feature branch for each bug, then merging to local master.
14:56 IZh whitby: Suppose, we can build the workflow of two companies from scratch.
14:57 IZh whitby: All is set in stone, that the Google's internal workflow is to create a feature branch for each bug, then merging it to master.
14:57 IZh whitby: Foo Inc tries to mimic this workflow.
14:57 IZh whitby: With additional level of indirection.
14:58 IZh whitby: Ideally, Foo Inc should look as a single developer from Google's point of view.
14:59 IZh whitby: So internal Foo Inc's review process isn't visible to Google.
14:59 Levex joined #git
15:00 nickabbey joined #git
15:00 IZh whitby: So Foo Inc forks the master repo, creates a feature branch then makes a pull request. Google reviews it, Foo In revrites the commits until Google is satisfied, then the commits are merged to Google's master.
15:00 IZh whitby: But at the same time Foo Inc needs similar internal review process.
15:00 IZh (To avoid sending total rubish to Google)
15:02 AaronMT joined #git
15:02 whitby You don't need to prefix every msg with my nick :(
15:02 Derperperd joined #git
15:02 whitby But yeah, I think you have a fine understanding of how that would work. What's the question?
15:04 IZh whitby: The question is that at the time of merge request to Google, the feature branch are merged to local master (from which the request is made). But what to to if Google asks to change something? Is it possible to change commits after branch merge?
15:04 gitstudent joined #git
15:04 ocbtec joined #git
15:04 whitby > Is it possible to change commits after branch merge?
15:04 whitby Yes
15:05 IZh Probably need to test it with 3 servers. :-)
15:05 Lunatrius` joined #git
15:06 IZh How to update pull request on remote server?
15:06 jameser joined #git
15:07 gugah joined #git
15:07 IZh I mean, when you fork a project within e.g. github, then makes a pull request, then makes a forced push to your branch, GH updates pull request automatically.
15:07 IZh But there would be 2 independent servers: Google's and Foo Inc's.
15:08 ysh joined #git
15:08 whitby You can do a PR to the branch that you had submitted a PR from
15:09 jozwior joined #git
15:09 IZh Thanks. Need to test it.
15:09 Noldorin joined #git
15:09 shinnya joined #git
15:11 LeBlaaanc joined #git
15:12 l4v2 joined #git
15:12 pur3eval joined #git
15:13 Rodya_ joined #git
15:15 mohabaks joined #git
15:15 holodoc joined #git
15:16 Snugglebash joined #git
15:17 watabou joined #git
15:17 nickabbey joined #git
15:19 metachr0n joined #git
15:21 Derperperd joined #git
15:22 Snugglebash joined #git
15:25 texinwien joined #git
15:26 miczac joined #git
15:26 unbalancedparen joined #git
15:27 cdown joined #git
15:28 miczac_ joined #git
15:32 dfas joined #git
15:33 Levex joined #git
15:34 nettoweb joined #git
15:35 ivh left #git
15:37 shgysk8zer0_ joined #git
15:37 MrLawrence joined #git
15:40 jimi_ joined #git
15:40 patarr joined #git
15:40 mischat joined #git
15:40 mohabaks joined #git
15:45 staafl joined #git
15:46 patarr joined #git
15:48 aca20031 joined #git
15:51 shgysk8zer0_ joined #git
15:52 dabura667 joined #git
15:53 blackwind_123 joined #git
15:57 mischat joined #git
15:58 dersand joined #git
15:59 durham joined #git
15:59 p4trix joined #git
15:59 mnaser joined #git
16:00 nowhereman joined #git
16:03 DBag left #git
16:05 skalpin Is there a way for me to create and use variables in the gitconfig?
16:05 robattila256 joined #git
16:05 brokensyntax joined #git
16:06 Shashikant86 joined #git
16:06 patarr joined #git
16:07 IZh left #git
16:07 clemf joined #git
16:08 Jackneill joined #git
16:09 jagob joined #git
16:09 Snugglebash joined #git
16:10 Xunie joined #git
16:11 JanC_ joined #git
16:13 pur3eval joined #git
16:13 Rodya_ joined #git
16:14 Vampire0_ joined #git
16:15 crose joined #git
16:17 dreiss joined #git
16:18 cdg joined #git
16:18 watabou joined #git
16:20 Vampire0 joined #git
16:21 mischat joined #git
16:21 jimklo joined #git
16:22 thiago joined #git
16:22 mischat joined #git
16:22 Rodya_ joined #git
16:23 ayogi joined #git
16:23 Levex joined #git
16:23 Darren_ joined #git
16:26 cdown joined #git
16:26 Vampire0_ joined #git
16:27 Gsham joined #git
16:30 hobodave joined #git
16:31 unbalancedparen joined #git
16:31 Xunie` joined #git
16:32 _ikke_ skalpin: No, the format does not allow variables
16:33 tang^ joined #git
16:33 synthroid joined #git
16:35 jozwior joined #git
16:35 skalpin The best I've come up with is under [ailas] to add variable = "!echo my-string" and then later use it with thing = "!git rebase origin/$(git my-string)"
16:35 jccn joined #git
16:36 _ikke_ of course you can add your own config items there
16:36 _ikke_ git config my.string value
16:38 amagawdd joined #git
16:38 brent__ joined #git
16:38 madewokherd joined #git
16:40 mgedmin DolphinDream: my rationale is same as whitbys: if every user of this git repo will have to ignore those files, they shouldn't be forced to manually update their global .gitignore
16:41 solenodic joined #git
16:42 Levex joined #git
16:43 sagerdearia joined #git
16:44 skalpin _ikke_: I would I use it in an alias then?
16:45 F0rTh3J3st_ joined #git
16:46 jozwior joined #git
16:47 Rodya_ joined #git
16:47 scarcry joined #git
16:48 F0rTh3J3st_ joined #git
16:48 Xunie joined #git
16:48 ankit01ojha joined #git
16:50 Levex joined #git
16:50 TomyWork i have a clone of a repo within the work tree of another repo
16:51 TomyWork is there any easy way to make that a submodule?
16:52 _ikke_ skalpin: similiar, $(git config my.variable)
16:52 F0rTh3J3st joined #git
16:53 skalpin _ikke_: I actually like that a little bit better, thanks for the idea!
16:53 Rodya_ joined #git
16:55 svm_invictvs joined #git
16:55 YuGiOhJCJ joined #git
16:56 brent__ joined #git
16:56 skinkitten joined #git
17:00 gugah joined #git
17:03 wildleaf_ joined #git
17:04 SteenJobs_ joined #git
17:04 wildleaf_ I need some help. Whenever I push my code up to my repo it changes all of the formatting. Anyone know why?
17:04 TomyWork are you on windows?
17:04 SteenJobs_ anyone with iOS experience - when i git diff my project.pbxproj, the diff output is supressed - any idea how to see the diff?
17:05 TomyWork SteenJobs_ you use git on iOS?
17:05 SteenJobs_ TomyWork: what do you mean?
17:05 TomyWork i didnt know there was a port for it
17:06 TomyWork anyway, if you dont see the output, it could be that there isnt any or that it's not a text file in any encodinggit diff can process
17:06 TomyWork utf-16, for instance
17:06 wildleaf_ TomyWork: no I'm on os x
17:06 SteenJobs_ like not ON iOS, but it’s for an Xcode/iOS project
17:07 SteenJobs_ oh interesting
17:07 SteenJobs_ it says:
17:07 SteenJobs_ well it treats them as “binary files"
17:07 SteenJobs_ maybe that’s why?
17:07 TomyWork wildleaf_ then please define what you mean by "push", "code", "my repo", "changes" and "all of the formatting"
17:07 SteenJobs_ this is the only filetype that gives me an issue
17:07 IndigoTiger joined #git
17:07 DBag joined #git
17:07 TomyWork SteenJobs_ can you instruct xcode to save it as utf-8?
17:07 SteenJobs_ hmm good question, i’ll look into it
17:08 TomyWork but first of all, confirm that it's utf-16
17:09 cdown joined #git
17:09 wildleaf_ TomyWork: when i 'git push' my 'js' code to the 'remote repo' the way the code is formatted i.e. the spaces and tabs completely changes.
17:09 TomyWork wildleaf_ if "git push" changed your code, it would also have to change your commit IDs
17:10 TomyWork I assume it doesn't and call bullshit :)
17:10 wildleaf_ TomyWork: maybe i'm in the wrong place actually.. because when i 'git diff' the code also looks different. it must be my editor?
17:10 _ikke_ TomyWork: Not really welcoming for users to call their observations bullshit
17:10 TomyWork _ikke_ just their conclusions
17:11 TomyWork _ikke_ apparently i was right, as well
17:11 wildleaf_ welp, thanks, i guess? :)
17:12 TomyWork wildleaf_ yeah, some editors rearrange all the code when saving. that's probably it
17:12 SteenJobs TomyWork: the file is utf-8
17:13 TomyWork SteenJobs huh? odd, does it contain any binary parts, then?
17:13 tolland left #git
17:13 TomyWork wildleaf_ try using a dumber editor, like mcedit or nano or something
17:13 TomyWork i dunno what you have available
17:14 SteenJobs TomyWork: don’t think so - file -I output is “text/plain; charset=utf-8”, so not sure what’s causing git to think it’s a binary file
17:14 _ikke_ TomyWork: You can be right, but it's still not kind
17:14 _ikke_ We try to be a welcoming community here
17:14 wildleaf_ will do
17:14 TomyWork wildleaf_ see if you can reproduce the problem with such an editor. if not, it's the editor
17:14 Murr joined #git
17:14 wildleaf_ +1
17:15 rath joined #git
17:15 Levex joined #git
17:18 spacelord_ joined #git
17:18 Murr joined #git
17:19 watabou joined #git
17:20 mischat joined #git
17:21 SteenJobs TomyWork: figured it out
17:22 Oatmeal joined #git
17:22 SteenJobs TomyWork: https://git-scm.com/book/en/v2/​Customizing-Git-Git-Attributes - under ‘Identifying Binary Files’
17:22 SteenJobs checked the .gitattributes file and lo and behold it had *.pbxproj binary
17:24 cdown joined #git
17:24 skinkitten joined #git
17:25 jnavila joined #git
17:28 jccn joined #git
17:28 TomyWork oww
17:29 TomyWork ok i kinda assumed it was your project and you knew what you checked in :)
17:30 _ikke_ SteenJobs already left
17:32 gugah joined #git
17:33 troulouliou_div2 joined #git
17:34 Karazhan joined #git
17:34 DBag joined #git
17:39 mikecmpbll joined #git
17:40 mischat joined #git
17:41 solenodic joined #git
17:44 canuk joined #git
17:48 Rodya_ joined #git
17:48 oncall-pokemon joined #git
17:50 BackEndCoder joined #git
17:50 joki joined #git
17:51 jimsio joined #git
17:52 chardan joined #git
17:53 jozwior joined #git
17:53 bgerber joined #git
17:54 pandeiro joined #git
17:58 wrkrcoop joined #git
17:59 wrkrcoop i hear ppl at work say ‘we have to merge down’ what the hell does that mean?
17:59 wrkrcoop google no say
17:59 _ikke_ no clue either
18:00 wrkrcoop cool
18:02 pur3eval joined #git
18:02 cdown joined #git
18:07 TomyWork wrkrcoop the only place i know where that makes sense is when merging image layers in Gimp
18:07 TomyWork but Git? nope
18:08 wrkrcoop haha cool.
18:08 wrkrcoop im going to ask them next time
18:10 mohabaks joined #git
18:12 raynold joined #git
18:13 rudi_s wrkrcoop: I guess "merge down" means that they have to merge patches to other branches, e.g. you have a master and a release branch and if you fix a security issue on the master branch, you have to merge those changes "down" to the release branch so that you can release the fixes. However you can't really do that in Git and have to merge "up", so you can fix it on the release branch and then merge that= ...
18:13 rudi_s ... into the master branch.
18:13 Xunie joined #git
18:13 Bizkit joined #git
18:13 thierryp joined #git
18:13 wrkrcoop rudi_s: thank you that makes sense
18:14 muldover joined #git
18:14 e14 joined #git
18:16 chachasmooth joined #git
18:17 rudi_s wrkrcoop: But why not ask them, maybe they're using the term differently.
18:17 wrkrcoop rudi_s: i will, just didn’t want to look like a nub
18:18 nickabbey joined #git
18:18 duderonomy joined #git
18:18 spacelord_ joined #git
18:18 thiago joined #git
18:19 Guest30 joined #git
18:20 tterrag joined #git
18:21 watabou joined #git
18:22 dreiss joined #git
18:25 gopar joined #git
18:26 e14 joined #git
18:29 e01 joined #git
18:30 Xandaros joined #git
18:31 johnnyfive joined #git
18:32 matoro joined #git
18:36 Guest85 joined #git
18:36 fatalhalt joined #git
18:36 inflames joined #git
18:36 borkr joined #git
18:38 mohabaks joined #git
18:40 Guest30 joined #git
18:41 Tobbi joined #git
18:42 solenodic joined #git
18:44 johnnyfive joined #git
18:44 johnnyfive joined #git
18:44 SteenJobs joined #git
18:47 Naan joined #git
18:48 aartist joined #git
18:51 kutenai I'm not sure how to do a diff that shows the changes from a potential merge. Assume I have a branch 'feature/x', and develop has moved forward since feature/x started. If I do a 'git diff feature/x..develop', I see many changes, which include all of the changes from develop  back to the merge base. If, instead I checkout develop, then do 'git merge --no-ff
18:51 kutenai --no-commit feature/x', I get just the changes that would be applied if feature/x is merged onto develop. QUESTION is there some way to do a 'git diff <???>' that would show me the changes that would be applied to develop if I merge feature/x into it?
18:52 kutenai That is about as detailed of a question as I come up with ;-)
18:52 joshszep joined #git
18:53 nickabbey joined #git
18:55 Levex joined #git
18:57 skinkitten joined #git
18:57 kutenai Hmm. it seems the simple answer is to use three dots... I tried this before and it didn't work as I expected.. but I must have done something wrong.. I tried it again, and i did just what I expected... oh well.
18:57 dil3mm4 joined #git
19:00 HardlySeen joined #git
19:01 matoro joined #git
19:02 Levex joined #git
19:02 pur3eval joined #git
19:03 Guest9 joined #git
19:05 ych joined #git
19:07 systemovich joined #git
19:08 cdown joined #git
19:08 SwiftMatt joined #git
19:09 hobodave_ joined #git
19:10 mischat joined #git
19:10 joki joined #git
19:10 gugah joined #git
19:11 frew joined #git
19:12 cdown_ joined #git
19:12 frew left #git
19:14 TomyLobo joined #git
19:16 JonelethIrenicus joined #git
19:17 JonelethIrenicus how do i use git lfs?
19:17 bezhermoso joined #git
19:19 ozmage joined #git
19:19 peeps[lappy] joined #git
19:21 Zarthus did you read the instructions that came with installing it
19:22 apotry joined #git
19:22 miczac joined #git
19:23 watabou joined #git
19:23 JonelethIrenicus Zarthus: yeah the instructions suck
19:24 vuoto joined #git
19:24 Zarthus what part are you stuck at
19:24 Zarthus https://git-lfs.github.com/ looks pretty clear
19:25 setuid joined #git
19:25 JonelethIrenicus well i installed the deb via their script you get with curl and git lfs isn't available
19:25 theoceaniscool joined #git
19:26 JonelethIrenicus just says The repository is setup! You can now install packages.
19:26 Zarthus I haven't installed it as a .deb before, https://github.com/git-lfs/git​-lfs/blob/master/INSTALLING.md is the relevant documentation though
19:26 cdown joined #git
19:27 Zarthus if you installed it from a .deb my only thought is that it's not available in your $PATH
19:27 spacelord_ joined #git
19:29 JonelethIrenicus Zarthus: alright i get it
19:29 JonelethIrenicus the "deb" just adds a an apt source
19:30 Zarthus lol
19:30 Zarthus that's a bit weird
19:30 bezhermoso I dunno if this belongs here but I'll give it a shot: on MacOS `taskgated` complains with `MacOS error: -67062.` whenever I perform `git status` or `git ls-files` etc. Anyone else know what might be causing this?
19:30 Vampire0 joined #git
19:30 DDBag joined #git
19:30 JonelethIrenicus it doesn't actually install the package
19:30 JonelethIrenicus haha
19:31 WaReZ joined #git
19:31 flavius joined #git
19:31 matsaman joined #git
19:33 irqq joined #git
19:33 qqx That seems to be fairly standard. Adding an apt source allows providing updates for it. And separating that from the actual software avoids repeatedly messing with the sources as that happens.
19:36 Macaveli joined #git
19:37 Macaveli joined #git
19:38 Macaveli joined #git
19:39 _ADN_ joined #git
19:42 MineCoins joined #git
19:42 miczac joined #git
19:43 Macaveli joined #git
19:43 Gsham joined #git
19:43 Zarthus qqx: Sure, but it's unexpected when the instructions read as "You can install Git LFS ..." and then the instructions don't install git-lfs.
19:43 Zarthus qqx: https://github.com/git-lfs/git-lfs#getting-started for reference.
19:47 cyberspectre joined #git
19:48 Levex joined #git
19:48 skinkitten_ joined #git
19:48 Gsham joined #git
19:48 cyberspectre left #git
19:48 spacelord_ joined #git
19:48 Vampire0_ joined #git
19:52 tyreld joined #git
19:53 chipotle joined #git
19:56 Mikerhinos joined #git
19:57 gareppa joined #git
19:58 gugah joined #git
20:02 urda joined #git
20:05 aw1 joined #git
20:06 Synthead joined #git
20:09 Sasazuka joined #git
20:09 dsdeiz joined #git
20:09 dsdeiz joined #git
20:11 SteenJobs left #git
20:15 joshszep joined #git
20:15 blokeybloke joined #git
20:15 Spuri0us joined #git
20:19 WakiMiko joined #git
20:19 user__ joined #git
20:20 DBag joined #git
20:21 Vampire0 joined #git
20:22 DDDBag joined #git
20:23 Derperperd joined #git
20:24 WakiMiko joined #git
20:25 watabou joined #git
20:27 spacelord_ joined #git
20:27 netj joined #git
20:27 e01 joined #git
20:27 e01 joined #git
20:29 chachasmooth joined #git
20:29 peeps[lappy] joined #git
20:29 Doow joined #git
20:36 oich joined #git
20:39 oich in a submodule where HEAD is detached HEAD means the current commit?
20:40 ALVAN joined #git
20:40 oich ok I read that that is what head means
20:40 ALVAN hi ... is possible to make a file private inside a project ?
20:40 ALVAN lets say i have a passwd file inside a project
20:41 zoph add it to the .gitignore
20:41 _ikke_ ALVAN: You can not prevent a tracked file from being pushed
20:41 markus-k joined #git
20:41 _ikke_ ALVAN: so like zoph suggested, don't track it
20:41 halp joined #git
20:41 peepsalot joined #git
20:42 ALVAN ok thanks
20:43 mrVandermare joined #git
20:43 matoro joined #git
20:43 halp hi. is there a way to update a commit that happened a few commits back?
20:43 _ikke_ halp: sure. Question is if it's sensible (mainly depends on whether you pushed it or not)
20:44 chachasmooth joined #git
20:44 halp _ikke_: i figured it's possible, but yeah i don't know how much trouble it's worth. it is pushed
20:44 _ikke_ halp: Might others have fetched that commit?
20:44 halp indeed they have
20:45 _ikke_ Are you prepared to explain them how to make use of the new history you introduce?
20:46 oich So, I have a file in working tree that is modified. If I use git checkout -- path. It still says it is modified. (this in a submodule with detached head) If I use get reset HEAD path, it is no longer reported as modified. how does that work?
20:46 _ikke_ oich: Do you have a !transcript?
20:46 gitinfo oich: Please paste (using https://gist.github.com/ or similar) a transcript ( https://git.io/viMGr ) of your terminal session so we can see exactly what you see
20:46 halp _ikke_: yes, it's one person who's probably okay with it, assuming it's not a huge change. is there a name for this type of action? i'm not sure exactly how it would work, and how the new history would work
20:47 _ikke_ halp: It's called rewriting history
20:47 _ikke_ halp: you can do it with several command, but in this case, you want to use interactive rebase
20:48 halp _ikke_: i assume it would also be possible to insert a commit back in history too
20:48 _ikke_ halp: correct
20:48 NullableTruth joined #git
20:48 Rodya_ joined #git
20:49 Vicentius joined #git
20:49 halp _ikke_: i am looking at this SO discussion and it looks simple enough: https://stackoverflow.com/questions/32​315156/how-to-inject-a-commit-between-​some-two-arbitrary-commits-in-the-past -- how would the consequences be for my coworker and how can he update his repo?
20:50 _ikke_ halp: in simple cases, it's enough for them to do git pull --rebase while having the updated branch checked out
20:51 b1tchcakes joined #git
20:51 pur3eval joined #git
20:51 Macaveli joined #git
20:51 halp _ikke_: ok, this is a simple case so i will try that. thank you very much for your help!
20:51 _ikke_ no problem
20:51 drodger joined #git
20:52 _ikke_ halp: One tip, learn to use git push --force-with-lease <remote> <branch> instead of --force / -f
20:53 yqt joined #git
20:54 overlord_tm joined #git
20:54 halp _ikke_: will that be necessary for thsi rebase?
20:54 _ikke_ yes
20:55 _ikke_ if you rewrite history that you have pushed, you are required to force push
20:55 halp i see. thanks again!
20:55 _ikke_ even if you do git commit --amend on a commit that has been pushed, it is necessary
20:56 Flaghacker joined #git
20:56 oich _ikke_ I don't have a transcript unfortunately. What I have typed from memory is different because git was suggesting using git reset HEAD as a fix, whereas if I try to reproduce it suggests using git add or git checkout. https://gist.github.com/anonymous​/a1b4a7b0f5cf7fd3a4e6272f5e8fc1a0
20:57 crose joined #git
20:57 _ikke_ Right, I was mostly interested in what things like git status said
20:58 Flaghacker Hey guys! I want to have a file in my repository that doesn't show up in any normal branch but has it's own history in a separate branch, but I want the most recent version of that file to still show up on my local filesystem when I'm in a normal branch. Is there a way to do this?
20:58 m0viefreak joined #git
20:58 _ikke_ Flaghacker: Not in an easy way
20:59 _ikke_ Flaghacker: You can always checkout the file from that branch
20:59 _ikke_ but updating the file on that branch requires you to checkout that branch
20:59 _ikke_ (perhaps man git worktree can work for that)
20:59 gitinfo the git-worktree manpage is available at http://jk.gs/git-worktree.html
20:59 Flaghacker And whenever I check out a different branch I'll have to update that file too?
21:00 _ikke_ Flaghacker: Make sure that file is not tracked, and you won't have to
21:00 Flaghacker Oh okay that doesn't sound too bad, I wont update that file very often anyways.
21:01 _ikke_ But it means you always need to make sure you get this file from that branch whenever you make a new clone or try to deploy
21:02 _ikke_ oich: could it be that that file was staged?
21:02 chipotle joined #git
21:02 _ikke_ oich: That would explain why git checkout <file> would not do anything
21:04 oich _kke_ my memory is that it was modified. It had been set to assume-unchanged. to update the submodule it complains the file wlll be overwritten. so I --no-assume-unchanged then checkout then the scenario I described where it still say it is modified after checkout
21:04 oich _ikke_ .. I mean
21:05 cdg joined #git
21:06 dave0x6d joined #git
21:06 MattMake_ joined #git
21:07 cdg_ joined #git
21:07 BlackMaria_netsp joined #git
21:08 tobiasvl oich: !assume_unchanged
21:08 gitinfo oich: git update-index --assume-unchanged is sometimes suggested for keeping yourself from committing changes to a file, but it's actually a promise to Git that the file is identical to the committed version, to reduce unnecessary disk reads. Not only will Git feel free to discard the flag, it will also happily overwrite your changes whenever a checkout/merge/rebase/... updates it.
21:09 Tobbi joined #git
21:11 cdg__ joined #git
21:11 oich tobiasvl I might be missing the point. I unset assume-unchanged before checkout so that any differences between it and head would be undone, or so I thought
21:13 BlackMaria joined #git
21:13 Levex joined #git
21:14 oich "Git will fail (gracefully) in case it needs to modify this file in the index e.g. when merging in a commit; thus, in case the assumed-untracked file is changed upstream,you will need to handle the situation manually." so I "handled the situation manually" by unsetting the bit and using checkout
21:16 cdg joined #git
21:16 jimi_ joined #git
21:19 miczac_ joined #git
21:20 gopar joined #git
21:21 LiftLeft joined #git
21:22 theoceaniscool joined #git
21:23 fuzzybear3965 joined #git
21:27 watabou joined #git
21:28 oich _ikke_ it must have been that it was staged. yeah. okay that explains what I saw
21:29 kristofferR joined #git
21:29 joshszep joined #git
21:33 oich thanks I have to log off
21:33 oich left #git
21:37 Ryanar joined #git
21:37 BSaboia joined #git
21:37 spacelord_ joined #git
21:38 renee80 left #git
21:38 tvw joined #git
21:40 hobodave joined #git
21:42 diogenese joined #git
21:43 askb joined #git
21:43 ShapeShifter499 joined #git
21:44 joshszep joined #git
21:45 e01 joined #git
21:45 e01 joined #git
21:46 crayon joined #git
21:47 robotroll joined #git
21:48 Vampire0_ joined #git
21:49 gugah joined #git
21:52 pur3eval joined #git
21:52 aruns_ joined #git
21:53 aruns_ Hi, quick question.
21:53 aruns_ Working locally on my portfolio site.
21:53 aruns_ Using SASS to compile CSS styles.
21:53 aruns_ I put down style.css in my .gitignore
21:53 aruns_ And committed / pushed to repo.
21:53 aruns_ But Git's still tracking it?
21:54 aruns_ This is my .gitignore file
21:54 aruns_ https://bpaste.net/show/b16d156ea215
21:54 aruns_ Unless you're not supposed to have blank lines?
21:57 ferr1 joined #git
21:59 _ikke_ aruns_: !ignore_tracked
21:59 gitinfo aruns_: Git only applies ignore patterns to untracked files. You can't use ignore patterns to ignore changes to files that are already tracked by git. To remove files only from git, but keeping them on disk, use git rm --cached <file>. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.
21:59 aruns_ Ah figured out now.
21:59 aruns_ _ikke_: Thanks it's working now.
22:00 Vicentius joined #git
22:02 LaptopBrain joined #git
22:04 mellernoia joined #git
22:05 ramonn334 joined #git
22:06 ramonn334 left #git
22:06 ramonn334 joined #git
22:06 Derperperd joined #git
22:07 cdg_ joined #git
22:12 cdg joined #git
22:13 pingwindyktator joined #git
22:13 dreiss joined #git
22:14 salamanderrake joined #git
22:16 jagob joined #git
22:20 BSAlb joined #git
22:21 endiruna joined #git
22:23 cdg_ joined #git
22:24 ALVAN left #git
22:25 JonelethIrenicus left #git
22:27 netj joined #git
22:27 skinkitten joined #git
22:28 watabou joined #git
22:30 kadoban joined #git
22:30 skinkitten_ joined #git
22:32 salamanderrake joined #git
22:33 Rodya_ joined #git
22:33 hhee joined #git
22:34 jstimm joined #git
22:34 Taikibansei joined #git
22:35 jeffreylevesque joined #git
22:36 eldritch joined #git
22:37 pfloyd joined #git
22:37 gugah joined #git
22:37 BSAlb joined #git
22:38 cdg joined #git
22:38 pfloyd I have a situation where someone merged branch a onto branch b later than they wanted (they meant to merge a tag of branch a from a specific release).  they've since made other commits/etc to branch b.  what are my options to kind of undoing that merge without subsequently corrupting branch a when branch b merges back into a ?
22:39 pfloyd oh and they've already merged b back to a...
22:40 _ikke_ !situation
22:40 gitinfo Please pastebin the relevant output of `git log -n 20 --all --graph --format="%h %p %d"` to give us an idea about what your situation is (increase the -n argument if your problem is more complex than 20 commits can explain)
22:40 pfloyd I can't show that output here as the repository is confidential.
22:41 _ikke_ pfloyd: Look at the output of that command
22:41 _ikke_ Unless branch names are also confidential, it should not expose any information
22:41 _ikke_ if branchnames are a problem, remove the %d from the formt
22:41 gugah joined #git
22:41 pfloyd well there are confidential tags.  I can filter through sed I guess hang on
22:41 _ikke_ ok
22:41 amagawdd joined #git
22:45 pfloyd _ikke_: have to go back 432 sheesh.  that's the nature of using merge commits (long story)
22:45 _ikke_ Feel free to condense it a bit
22:46 Jason left #git
22:46 gopar joined #git
22:46 _ikke_ I don't need straight lines of lots of commits
22:46 BSaboia joined #git
22:47 pfloyd it's a pretty simple situation, let me show you a quick pic
22:47 matoro joined #git
22:47 _ikke_ ok
22:48 sargentmki joined #git
22:49 pfloyd _ikke_: http://imgur.com/a/uz0pc
22:50 _ikke_ I think it's doable, but not easy
22:50 pfloyd so basically, they had pulled changes from the A branch before they ran git merge A.  they meant to merge where the green branch point was, not the red brnach point.  now they've made commits since, and merged those back into A.  The problem is the changes between the green and red branch points are incompatible with the changes they made in their B branch
22:51 hobodave joined #git
22:52 _ikke_ pfloyd: is history going from right to left?
22:52 pfloyd left to right
22:53 pur3eval joined #git
22:53 bezhermoso joined #git
22:54 spacelord_ joined #git
22:54 _ikke_ One question, has this history been shared yet?
22:54 pfloyd yes
22:55 _ikke_ That does make it tougher to rewrite history
22:55 pfloyd yeah :/
22:55 pfloyd my suggestion was to deal with the fallout and intercept the next point they can from A
22:56 _ikke_ rewriting history with merges is going to be a pain, even more so when it has been shared
22:57 pfloyd yeah, I guess I'll just have to tell them to live with it. was afraid of that, but wanted to confirm :)
22:57 _ikke_ I would rewrite A to remove the merge from B into A
22:58 _ikke_ s/rewrite/rebase
22:59 cagomez joined #git
22:59 _ikke_ then rebase B just before the merge, then merge in the correct commit, and continue the rebase
22:59 _ikke_ and then redo the merge from B into A
23:00 pfloyd hmm so rebase A back to prior to the merge of B and merge the commits that have happened in A, rebase B back to before the merge from A and re-merge the commits in B, then merge B back to A again?
23:00 _ikke_ the rebase on A would just remove the merge, so it would still contain all commits, except the merge
23:01 fmeerkoetter joined #git
23:01 pfloyd ah ok what about the other merges that have happened?
23:02 GodGinra1 joined #git
23:03 pfloyd A has continued on since B merged back to A, by quite a bit actually
23:03 _ikke_ You can try to redo that merge later again with a rebase
23:03 llamapixel This seems to help with more complex layouts
23:03 llamapixel http://asciiflow.com/
23:04 _ikke_ so rebase -i until the commit that it was merged into, and then do the merge, and continue
23:04 llamapixel Not saying this is one just throwing it out there.
23:04 _ikke_ I know that tool
23:06 pfloyd thanks guys, appreciate it. at this point the ROI of doing all that along with the risk to the integrity of branch A and having to pause any pushes to A is not worth it, but they have to make that call I guess (proj mgmt)
23:08 Darcidride joined #git
23:10 morenoh149 joined #git
23:12 morenoh149 how can I create a commit based off the diff of two branches ?
23:12 nickabbey joined #git
23:12 morenoh149 I accidentally ammended a commit a shouldn't have
23:16 mda1 joined #git
23:20 dreiss joined #git
23:22 Sasazuka joined #git
23:24 e01 joined #git
23:25 e01 joined #git
23:25 e01 joined #git
23:26 inflames joined #git
23:28 cagomez joined #git
23:29 Vampire0_ joined #git
23:30 watabou joined #git
23:35 mehola joined #git
23:35 MrWoohoo joined #git
23:36 green_ joined #git
23:36 dendazen joined #git
23:36 NeverDie joined #git
23:39 mephistorium joined #git
23:42 green_ What's a good way to "clone the regdb from git;  Is there a good tutorial or instructions for cloning, running make etc
23:43 bremner regdb?
23:43 green_ I am trying to edit regdb.c
23:43 green_ regd.c
23:44 green_ https://git.kernel.org/cgit/linux/ker​nel/git/linville/wireless-regdb.git/
23:44 matoro joined #git
23:45 green_ I am running mint 18.1
23:46 Vinnie_win joined #git
23:46 bremner sadly just because a project uses git doesn't mean #git knows about it. It sounds more like a question for linux kernel people.
23:48 ketas joined #git
23:49 ysh joined #git
23:51 green_ ok. is there a channel that would be appropriate
23:51 green_ I have instructions to clone the regdb, than make, copy the signing key; I don't know how to do it though
23:53 hahuang65 joined #git
23:53 pur3eval joined #git
23:53 chipotle joined #git
23:55 bremner there's ##workingset for help with make and similar tools. If you can't actually clone the repo, then someone here can probably help you.
23:56 hexagoxel joined #git
23:59 Derperperd joined #git

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