Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-08-08

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

All times shown according to UTC.

Time Nick Message
00:00 joki joined #git
00:00 Jon28__ joined #git
00:03 tesuji joined #git
00:05 watabou joined #git
00:07 kbs joined #git
00:15 xlegoman joined #git
00:18 interne7y joined #git
00:25 Derperperd joined #git
00:26 semantipc joined #git
00:32 Lunatrius` joined #git
00:32 bluezone joined #git
00:35 mischat joined #git
00:39 xall joined #git
00:40 fuchstronaut joined #git
00:40 Seriusam joined #git
00:41 afuentes joined #git
00:41 zacts joined #git
00:43 SoItBegins joined #git
00:44 kwmiebach joined #git
00:47 Vaelatern joined #git
00:47 KeelOfSteel joined #git
00:48 hahuang65 joined #git
00:49 Rodya_ joined #git
00:49 button joined #git
00:55 dsantiago joined #git
00:56 amdi_ joined #git
00:58 xall joined #git
01:02 tesuji_ joined #git
01:11 PerlJam joined #git
01:16 c3l joined #git
01:20 c3l Workflow question: I have a repo that is used by some people. I have some files with helper scripts and notes for development which I want to include in the repo, but I dont want to clutter master (these files will be of no interest / confusing to most users of the repo). What is a good way of including these helper/dev-files? A dev-branch?
01:22 thiago a wiki
01:23 thiago or just keep in the same repo, regular branch
01:23 HedgeMage c3l: I prefer NOT to do that, because at that point you have a branch structure that isn't doing what it's supposed to: represent different versions of the same content.
01:23 HedgeMage c3l: Either keep it in the main repo, clearly marked as dev utils, or make another repo, or stick it in your dev documentation if it's small enough.
01:24 fahadash joined #git
01:26 c3l You make some very good points. Thanks! I had not even considered the wiki-option, and since I'm on github there is already a wiki connected to the repo. I think I'll go with that.
01:28 pks joined #git
01:29 SoItBegins joined #git
01:32 Mannn joined #git
01:35 Dougie187 joined #git
01:36 mischat joined #git
01:38 rgrinberg joined #git
01:41 fuchstronaut joined #git
01:46 star_prone joined #git
01:48 ilbot3 joined #git
01:48 Topic for #git is now Welcome to #git, the place for git help and botany | Public logs at http://goo.gl/BuUi5o | Current stable version: 2.9.2 | First visit? Read: http://jk.gs/git | Getting "cannot send to channel"? /msg gitinfo .voice | Nurture your git-twigs and they will grow into a full branch
01:48 hahuang65 joined #git
01:50 beyertech joined #git
01:50 d^sh joined #git
01:52 shawnbon206 joined #git
01:52 beyertech joined #git
01:53 mehola joined #git
01:56 shawnbon206 question about submodules, I have a submodule included in my project, and according to the .gitmodule documentation if I don't specify a branch, it defaults to master, but when I clone my repo and cd to that sumodule's directory, it shows as detached at a specific commit.
01:57 arescorpio joined #git
01:57 shawnbon206 currently master is at that commit, but I expect git status to say "On branch master", instead it gives me "HEAD detached at fef1c2f"
01:59 mehola joined #git
02:02 SoItBegins joined #git
02:07 star_prone joined #git
02:09 ThinTim joined #git
02:13 [R] shawnbon206: submodules are associated with a specific commit
02:14 JZA joined #git
02:14 shawnbon206 [R] so it's working as expected?
02:14 JZA hi how can I clone a repo with a different folder
02:15 JZA git clone <new_folder> <url>
02:15 JZA or git clone <url> <new_folder>
02:17 shawnbon206 JZA: the second one
02:17 Goplat joined #git
02:17 [R] shawnbon206: yes, submodules are pretty awful
02:19 Izabia joined #git
02:19 interne7y joined #git
02:20 Rodya_ joined #git
02:21 JZA left #git
02:24 _28_ria joined #git
02:26 Sonderblade joined #git
02:27 Derperperd joined #git
02:27 ISmithers joined #git
02:27 thekrynn_ joined #git
02:28 fairuz joined #git
02:33 Dougie187 left #git
02:37 Limes__ joined #git
02:38 mischat joined #git
02:40 jubalh joined #git
02:42 fuchstronaut joined #git
02:43 rickroberts joined #git
02:45 rchavik joined #git
02:46 hexagoxel joined #git
02:46 githogori joined #git
02:46 lagothri1 joined #git
02:47 skylite_ joined #git
02:49 hahuang65 joined #git
02:53 rkazak joined #git
02:58 chachasmooth joined #git
03:01 tesuji joined #git
03:02 fished joined #git
03:02 aavrug joined #git
03:02 mehola joined #git
03:02 Derperperd joined #git
03:03 aavrug joined #git
03:03 Aljone_1 joined #git
03:09 HanClinto joined #git
03:09 slayerjain joined #git
03:10 Derperperd joined #git
03:11 xall joined #git
03:14 SoItBegins joined #git
03:15 beyertech joined #git
03:22 fstd_ joined #git
03:22 sabgenton joined #git
03:26 tanuki joined #git
03:26 sabgenton after  git checkout --orphan foo; git rm -r --force --cache .    and then committing new stuff
03:26 sabgenton I can not checkout the old branch
03:27 sabgenton Well I get the stuff will be overwritten message but it was clean before I left
03:29 sabgenton basically switching from the new orphan back to master  acts like  the files in working are out of sync with masters tip
03:31 Izabia joined #git
03:33 b1tchcakes joined #git
03:38 _sjs joined #git
03:42 shawnbon206 I've tried everything, the submodules are headless after the clone unless I go in and manually set them at master
03:42 thiago shawnbon206: that's the expected situation
03:42 fuchstronaut joined #git
03:43 thiago sabgenton: !float
03:43 gitinfo sabgenton: 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.
03:43 b1tchcakes joined #git
03:43 sabgenton thiago: but the files don't differ and  git still says the will get overwriten
03:44 sabgenton I think it's cause  theres two commit graphs with out  a common root commit
03:44 thiago sabgenton: don't doubt git
03:44 thiago Derperperd: the files will be overwritten
03:44 dikiaap joined #git
03:44 b1tchcakes joined #git
03:44 Derperperd thiago: what lol
03:45 sabgenton thiago: no Im just saying is that why its complaining?
03:45 sabgenton I'm trying to understand why is happening
03:46 sabgenton the files are definitely Identical it's like a bug in git
03:46 shawnbon206 thiago: really what I am trying to do is have nested repos, i'm trying to get Vundle to install with my vim dotfile directory. I suppose i could do it like "git clone https://url/of/dotfile/dir && .vim/post-clone.sh" ...
03:46 thiago sabgenton: it's complaining because you have a bunch of untracked files that would be overwritten by the checkout
03:47 rossome joined #git
03:49 j416 joined #git
03:50 thiago sabgenton: it's not a bug. Untracked files are not checked for similarity. They are different, period.
03:50 hahuang65 joined #git
03:51 dsantiago joined #git
03:56 dhollinger joined #git
03:56 mischat joined #git
03:57 RedneckDonut joined #git
03:57 dhollinger joined #git
03:58 dhollinger joined #git
04:00 realz joined #git
04:00 [Brain] joined #git
04:00 thecomedian joined #git
04:02 sabgenton thiago: after making some test repos I see the light.  I thought things floated at lest on track files when there where in sync with branch to checkout.  I see now files have to be in sync with HEAD
04:02 sabgenton thiago: thanks
04:03 thekryn__ joined #git
04:05 sangy joined #git
04:06 sabgenton next time I think I will   rm --cache  just the files I want to add then  git rm -rf .  then  add .
04:08 rkazak joined #git
04:08 mehola joined #git
04:08 The_Phoenix joined #git
04:08 star_prone joined #git
04:08 The_Phoenix joined #git
04:08 The_Phoenix joined #git
04:09 lss8 joined #git
04:10 The_Phoenix joined #git
04:12 The_Phoenix joined #git
04:13 The_Phoenix joined #git
04:19 khmarbaise joined #git
04:20 interne7y joined #git
04:24 Motoko joined #git
04:26 Erati joined #git
04:29 star_prone joined #git
04:30 lss8 joined #git
04:33 Cabanossi joined #git
04:34 shgysk8zer0 joined #git
04:36 DJ_Spies joined #git
04:37 a_thakur joined #git
04:39 shawnbon206 thiago: solved  https://github.com/shawnbon​206/.vim/blob/master/vimrc
04:39 a_thakur joined #git
04:40 thiago shawnbon206: sorry, I can't read that file
04:41 shawnbon206 thiago: you mean you need a link to the raw text?
04:41 shawnbon206 https://raw.githubusercontent.co​m/shawnbon206/.vim/master/vimrc
04:42 shawnbon206 anyways, the solution was to stop using submodules and just get something else to do what I wanted
04:43 shawnbon206 thanks for your help
04:43 fuchstronaut joined #git
04:44 a_thakur joined #git
04:45 a_thakur joined #git
04:46 xall joined #git
04:46 sabgenton ok easyer to -rf . then  checkout master:stuff back before adding
04:46 johnny56 joined #git
04:50 lss8 joined #git
04:51 rchavik joined #git
04:51 hahuang65 joined #git
04:53 a_thakur joined #git
04:54 Anupkumar joined #git
04:55 jceb joined #git
04:58 a_thakur joined #git
05:05 govg joined #git
05:07 jceb joined #git
05:08 harish_ joined #git
05:09 Raging_Hog joined #git
05:14 rchavik joined #git
05:14 stuh84 joined #git
05:15 circ-user-aUJuR joined #git
05:17 subhojit777 joined #git
05:25 phanimahesh joined #git
05:34 mehola joined #git
05:39 t-mart_ joined #git
05:39 realz__ joined #git
05:40 zeroed joined #git
05:44 fuchstronaut joined #git
05:44 ffabi joined #git
05:47 dreamyspell_ joined #git
05:50 lss8 joined #git
05:50 jokajak joined #git
05:50 govg joined #git
05:51 hahuang65 joined #git
05:56 sebascar joined #git
05:57 rscata joined #git
06:00 xlegoman joined #git
06:01 satya4ever joined #git
06:03 rkazak joined #git
06:06 HappyLoa- joined #git
06:08 freimatz joined #git
06:09 lss8 joined #git
06:12 Oebele joined #git
06:14 xall joined #git
06:17 seishun joined #git
06:20 blackwind_123 joined #git
06:22 star_prone joined #git
06:22 mischat joined #git
06:23 khmarbaise joined #git
06:28 beyertech joined #git
06:28 ffabi joined #git
06:28 Lvjasha joined #git
06:30 xall joined #git
06:30 jbu joined #git
06:32 jknetl joined #git
06:32 lss8 joined #git
06:35 jbu all: normally my workflow is to merge develop into master, but our customer merged things directly into master so now I need to sync develop with master.  Is the appropriate action to merge master into develop?  It makes sense to me, but I just want a sanity check here...new to git
06:39 ffabi joined #git
06:40 kadoban jbu: Sounds right
06:41 jbu kadoban, ok, thanks.  Being new to git, I find myself second guessing even things that make sense
06:41 rchavik joined #git
06:42 Lvjasha left #git
06:42 kadoban Yeah, double-checking can be a good idea. Also remember that you can just make a copy of the repo and try stuff out if you want, I used to do that a lot before I got confident about how to undo things I'd done that didn't go correctly.
06:42 star_prone joined #git
06:44 Sceorem joined #git
06:45 fuchstronaut joined #git
06:46 lss8 joined #git
06:47 zeroed joined #git
06:48 ffabi joined #git
06:48 Lvjasha joined #git
06:52 hahuang65 joined #git
06:55 ffabi joined #git
06:55 submitnine joined #git
06:56 lss8 joined #git
07:01 skasturi joined #git
07:02 Darcidride joined #git
07:06 lss8 joined #git
07:09 tanuki joined #git
07:10 mehola joined #git
07:11 govg joined #git
07:11 DevAntoine joined #git
07:16 tachoknight_ joined #git
07:17 lordjancso joined #git
07:19 djb-irc joined #git
07:22 star_prone joined #git
07:23 xall joined #git
07:25 a_thakur_ joined #git
07:26 ahmedelgabri joined #git
07:26 mariuscc joined #git
07:26 aidalgol joined #git
07:28 Balliad joined #git
07:28 ahmedelgabri joined #git
07:32 LiohAu joined #git
07:35 jbu kadoban, sorry, another newb question, but by "copy of the repo" you just mean make a new branch or...?
07:36 encod3 joined #git
07:36 kadoban jbu: That's an option that's always useful, but if you're really not sure if you're going to screw something up, you can just plain copy the entire directory it's in elsewhere and try there. It takes longer than making a branch, so you stop using it eventually once you get more comfortable.
07:37 jas4711 joined #git
07:38 a_thakur joined #git
07:39 sagerdearia joined #git
07:41 HappyLoa- joined #git
07:45 fuchstronaut joined #git
07:48 encod3 joined #git
07:52 Seveas joined #git
07:53 hahuang65 joined #git
07:57 underpantz_ joined #git
07:57 dreiss joined #git
07:58 TheBrayn joined #git
08:00 ki0 joined #git
08:01 xaa joined #git
08:01 lss8 joined #git
08:04 PCatinean joined #git
08:05 PCatinean How does one remove a folder from a git repository from the tree and history
08:05 PCatinean Like it never existed? :))
08:05 kadoban PCatinean: Look into 'git filter-branch'. It'll let you do that.
08:05 kadoban commit or index filter IIRC
08:06 Seveas PCatinean: or use the bfg repo cleaner if you don't want to dabble in black magic
08:06 selckin !filter_sub
08:06 gitinfo [!filter_subdirectory] You can use filter-branch's subdirectory filter to split a directory from an existing repository into a new repository, keeping all history. https://help.github.com/articles/splittin​g-a-subfolder-out-into-a-new-repository/
08:06 selckin !split_repo
08:06 gitinfo [!split_repository] You can use filter-branch to split an existing repository into smaller new repositories. See http://teach.github.com/art​icles/lesson-filter-branch/ and !filter_subdirectory
08:06 p4trix joined #git
08:07 dreiss joined #git
08:07 PCatinean dabble with black magic haha :))
08:07 sanketdg joined #git
08:07 PCatinean bfg repo cleaner, I assume it's a gui?
08:07 PCatinean http://stackoverflow.com/questions/872565/remove-s​ensitive-files-and-their-commits-from-git-history <-- like this?
08:07 kadoban Deleting a folder in filter-branch isn't very hard ... it's like one of the examples in the manpage.
08:08 Seveas PCatinean: neither bfg nor filter-branch are a gui
08:09 PCatinean I see
08:09 PCatinean The thing is I made a fork from my repo and want to share only 3 folders out of 5
08:09 PCatinean And I thought I would make a new branch, remove the two I don't need and push to the fork and then cherry-pick commits related only to the first 3 instead of all 5
08:09 marcogmonteiro joined #git
08:10 PCatinean That way I can work normally on the main repo and pull just the updates for the first 3 folder in the fork
08:12 Seveas that sounds painful.
08:12 Seveas why not just split the repo in two?
08:12 PCatinean I guess I could do that but I hate having many repos
08:12 bongjovi joined #git
08:12 Seveas 2 != many :P
08:12 PCatinean but what if somebody would want access only to one or two of the modules
08:12 PCatinean then I have to split it into multiple ones
08:12 tomog999_ joined #git
08:13 PCatinean and I end up with 1 repo per module
08:13 Seveas sounds like not a horrible a idea. Putting unrelated stuff in separate repos
08:13 kadoban Indeed
08:13 PCatinean and modules have dependencies between themselves
08:14 PCatinean So what would be the proper approach to this?
08:14 Seveas proper dependency management
08:14 PCatinean I've asked this question before but many suggested submodules, that means 1 repo per module
08:14 cbreak submodules work
08:14 Seveas submodules are bad
08:14 kadoban 'git subtree' works
08:14 cbreak they will maintain correct dependencies
08:14 PCatinean submodules I think are great when you incorporate libraries or other dependencies maintained by someone eslse
08:14 cbreak without the bloated overhead of subtree merging
08:15 Seveas the only good thing about submodules is removing them :)
08:15 cbreak PCatinean: they are also ok if they are maintained by you
08:15 cbreak PCatinean: the main point is that you want separatedness but strong binding
08:15 PCatinean but here I have BASE module (module 1 and module 2) that depend on BASE, then 2 more modules that can work with module 1 and module 2
08:15 cbreak that's what submodules are good for
08:15 Nitesh joined #git
08:16 PCatinean cbreak, so one repo per module then right?
08:16 cbreak sure
08:16 PCatinean and one that gather all of them in one place
08:16 cbreak if you want
08:16 cbreak or you can gather them where you need them
08:16 PCatinean and if I want a fork that has 2 of them
08:16 cbreak sure, that's possible... but that sounds bad
08:16 PCatinean I would make a repo and clone the two submodules
08:16 cbreak two of the same?
08:17 PCatinean then how I would give someone access to a repo with a custom configuration of the entire range of modules?
08:17 GavinMagnus joined #git
08:17 ISmithers joined #git
08:17 cbreak if you use submodules, you have a repository that has the modules as submodules
08:17 cbreak you can have multiple such repos
08:17 cbreak but I'd avoid having multiple copies of the same submodule
08:18 PCatinean that's why I don't like the idea of submodules that much
08:18 cbreak that issue isn't with submodules
08:18 cbreak it's with having modules
08:18 cbreak it's the same no matter what technique you use
08:19 PCatinean A customer is given the option to buy any of the six modules. One chooses two, another three, and one other two and one all of them
08:19 t0by joined #git
08:19 PCatinean I guess then one repo per module and a stable and a development branch perhaps
08:19 cbreak one repo per customer too? :)
08:19 PCatinean no just one repo per module
08:20 cbreak you'd use both
08:20 PCatinean and you give the customer access to the repos in question
08:20 cbreak if you'd use submodules
08:20 clmsy joined #git
08:20 PCatinean both what?
08:20 jguillen joined #git
08:20 cbreak both repo-per-module and repo-per-customer
08:21 cbreak the latter would bind the others together
08:21 cbreak you could bind them together intrinsically
08:21 cbreak i.e. by having moduleB have moduleA as submodule because A is a dependency
08:21 cbreak but that is troublesome if A is a dependency of multiple modules (see what I said above)
08:22 interne7y joined #git
08:23 Nitesh joined #git
08:24 PCatinean hmm
08:26 nub joined #git
08:27 dsdeiz joined #git
08:27 dsdeiz joined #git
08:28 ferr joined #git
08:32 star_prone joined #git
08:33 ki0 joined #git
08:33 cqi joined #git
08:34 ki0 joined #git
08:35 Anticom joined #git
08:37 LarArT joined #git
08:39 linuxmint joined #git
08:41 govg joined #git
08:42 pa joined #git
08:42 pa hello
08:42 gitinfo pa: 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.
08:42 ericsupreme joined #git
08:43 pa thanks :-) my question is: i have a git repo, with a master branch. i create a local feature branch say featureA on top of it. Then featureB based on featureA and featureC based on featureB.  If i am to modify (e.g., commit --amend) featureA, how do i rebase featureB on the modified featureA?
08:44 _ikke_ git rebase featureA featureB
08:44 chll_ joined #git
08:44 pa oh.. that easy? wow :-)
08:45 pa thanks!
08:45 pa _ikke_, does that update also featureC?
08:46 kadoban pa: No, you have to do a cascading series of rebases (in the correct order)
08:46 _ikke_ No
08:46 fuchstronaut joined #git
08:46 pa ah okay, thanks
08:47 harish_ joined #git
08:47 daGrevis joined #git
08:48 daGrevis hello! i have two branches - build and production. i want production to point to build branch. whats the best way to do it? i found out i can do git branch -f production $(commit-from-build-branch), but maybe there's something better?
08:48 garethdaine joined #git
08:49 nilg joined #git
08:49 p4trix joined #git
08:49 pa and to push changes, can i just push feature3 at the end of the rebase, or should i push every feature branch?
08:49 daGrevis as described here https://stackoverflow.com/questions​/5471174/git-move-branch-pointer-to​-different-commit-without-checkout
08:50 _ikke_ pa: every feature branch to update where the point to in the remote
08:50 wrouesnel1 joined #git
08:50 _ikke_ daGrevis: there are several ways to do this, but none of them is better than the other
08:53 daGrevis my coworkers are using an app called gitup. theres an gui option to change the tip of the branch. i wonder how the app does that.
08:54 mehola joined #git
08:54 hahuang65 joined #git
08:54 _ikke_ daGrevis: In the end, a branch is just a pointer to a commit (most basic form is a file with a hash in it)
08:54 kadoban Doesn't sound like something you really need a GUI for, it's ... pretty much trivial on the CLI.
08:55 kadoban The most used way, I think, by the way is 'git reset --hard'. It's the one I always use anyway.
08:55 daGrevis ye I prefer CLI myself too
08:55 daGrevis can I git reset to a commit thats not in the branch I'm currently in?
08:57 _ikke_ sure
08:57 govg joined #git
08:58 cbreak daGrevis: you can reset to any commit in the repository
08:58 cbreak daGrevis: and of course, you can git fetch any commit into your repo you want :)
08:58 Ezzy joined #git
08:58 PCatinean cbreak, so I'm making a new project on bitbucket and from what I understood the first step is to make a repository for each module
09:00 fate_ joined #git
09:00 FilipNortic_ joined #git
09:00 hahuang65 joined #git
09:00 playfullyExist joined #git
09:03 p4trick joined #git
09:03 leeN joined #git
09:03 daGrevis thanks for all the fish, you guys
09:07 davisonio joined #git
09:10 gfixler joined #git
09:10 Brain joined #git
09:10 The_Phoenix1 joined #git
09:10 zeroed joined #git
09:11 a_thakur joined #git
09:11 p4trix joined #git
09:13 iamtakingiteasy joined #git
09:17 TomyWork joined #git
09:17 xall joined #git
09:17 encod3 joined #git
09:18 darkbit joined #git
09:19 eijk joined #git
09:19 ahmedelgabri joined #git
09:21 Phlogistique joined #git
09:21 Phlogistique hi
09:21 encod3_ joined #git
09:22 jceb joined #git
09:22 Phlogistique I often use a long-winded incantation like "git fetch && git checkout -f && git reset --hard @{u} && git clean -fdx || git clone ..." to make sure I have a clean git repository
09:22 Phlogistique is there any command I missed that does just that? update a git repository or download it, and make sure it's clean
09:24 notebox joined #git
09:24 p4trick joined #git
09:25 satifant joined #git
09:26 Anticom Phlogistique: dunno if this is what you want but usually `git reset --hard HEAD ; git pull` should do the same
09:26 Anticom or if you want to be extra careful just delete the repository and clone it again
09:27 Anticom However this shouldn't be neccesary unless some other guys do some --foce pushes rewriting existing history
09:27 Phlogistique Anticom: the point is that I don't want to clone it again if I can avoid it
09:27 Anticom Well define what you mean with 'clean' repository?
09:27 Anticom Git takes care that no existing history is ever rewritten on a remote (unless you --force it)
09:28 LeMike joined #git
09:28 Anticom So in theory a simple git pull should do the trick
09:28 Anticom if your working copy is dirty you could shelve or discard your local changes
09:28 Anticom given that you don't want to commit them
09:28 Phlogistique yes, I already do that
09:28 Phlogistique as explained above
09:29 Phlogistique it's just that I use a long-winded incantation to make sure to catch any edge case. I wondered if there is a name for this action
09:29 Anticom Phlogistique: btw you could add a git alias to your ~/.gitconfig that does all those command with a convenient shortcut
09:30 Phlogistique Yes, that's what I do
09:30 Phlogistique though this command is more useful on buildfarms than on my workstation
09:30 Anticom hm on a build server we usually do clean builds
09:31 Anticom meaning rm bulid ; git clone ... build ; ...
09:31 _ikke_ Phlogistique: I would recommend using git archive rather than git clone + git pull
09:31 Phlogistique yes. If this can be done without paying the download penalty, it's better. That's exactly what the incantation above is for
09:32 Phlogistique _ikke_: but then you have to redo the download every time
09:32 _ikke_ Phlogistique: You can also create a bare repository and use git archive on that
09:33 _ikke_ Not bothering with a working tree
09:33 Anderson69s_Deb joined #git
09:33 Phlogistique indeed
09:33 _ikke_ Just fetch on the bare repo and create an archive of the version you need
09:33 carto joined #git
09:33 Phlogistique or check out the tree
09:34 _ikke_ Whatever floats your boat :-)
09:35 _ikke_ I usually do something like git archive <commit> --prefix=version/ | tar ...
09:35 p4trix joined #git
09:35 Tachyon_ joined #git
09:36 djb-irc joined #git
09:36 tvw joined #git
09:37 SoItBegins left #git
09:37 Phlogistique _ikke_: that's indeed less error-prone that a string of reset-checkout-clean
09:37 mikecmpbll joined #git
09:37 encod3 joined #git
09:37 ams__ joined #git
09:37 Phlogistique there might be a performance penalty though? I should try that on Linux and see if it matters at all
09:37 GavinMagnus joined #git
09:38 _ikke_ Unless your repository is huge, I would think the difference is negligable
09:39 Phlogistique yes. I meant "on the Linux git repository"
09:40 p4trick joined #git
09:42 [Brain] joined #git
09:44 nilg joined #git
09:45 dsdeiz joined #git
09:46 sebascar joined #git
09:47 fuchstronaut joined #git
09:48 razieliyo joined #git
09:50 dsdeiz joined #git
09:54 Izabia joined #git
09:56 dsdeiz joined #git
09:59 ljc joined #git
09:59 star_prone joined #git
10:00 ninjaaron joined #git
10:01 ninjaaron I have a branch that I want to keep getting changes from master on, but it also has it's own stuff that I don't want to have merged. I know this is basic git stuff, but I'm sort of a newb, and I forgot this.
10:02 hahuang65 joined #git
10:03 tobiasvl ninjaaron: rebase?
10:05 jhass or merge master into it
10:06 ninjaaron jhass: is that like `git merge master`?
10:06 davisonio joined #git
10:06 jhass yes
10:07 ninjaaron I _think_ that is what I'm trying to do.
10:07 rahtgaz joined #git
10:07 Silenced_v2 joined #git
10:12 Silenced_v2 joined #git
10:12 ninjaaron ok, another question: if I have unstaged changes and I checkout a different branch, what happens?
10:13 jhass if the committed version between the current and the target branch doesn't differ, they're carried over
10:13 jhass else the checkout is aborted
10:14 ninjaaron ok. great.
10:14 carto left #git
10:15 ninjaaron I've used git to do dotfiles for a few years, but I'm only just learning to use it for real.
10:15 fuchstronaut joined #git
10:15 tvw joined #git
10:16 achadwick joined #git
10:18 ninjaaron I need to read Pro Git again.
10:22 Gloomy joined #git
10:23 interne7y joined #git
10:26 jagob joined #git
10:26 freimatz joined #git
10:28 govg joined #git
10:31 nazarewk joined #git
10:33 seishun joined #git
10:35 Darkhunter joined #git
10:39 dlaboratory joined #git
10:42 Finnoloid joined #git
10:43 dikiaap joined #git
10:44 djb-irc joined #git
10:51 stoopkid joined #git
10:51 daGrevis left #git
10:52 tabrez joined #git
10:57 Raging_Hog joined #git
10:59 djb-irc joined #git
11:00 vuoto joined #git
11:00 st0opkid joined #git
11:01 fahadash joined #git
11:02 ahmedelgabri joined #git
11:02 hahuang65 joined #git
11:03 Anticom Why can't --dirty be used with git describe <hash> ?
11:05 Seveas becuase that'd make no sense
11:06 Anticom Seveas: Why? if i checkout revision x and then change some files without staging / commiting them and do a git describe --dirty i'd expect to get <checked-out-commit-hash>-dirty
11:06 Anticom Why does this not make sense?
11:07 Seveas see the manpage entry for --dirty
11:08 Seveas Describe the working tree. It means describe HEAD and appends <mark>
11:08 Seveas --dirty implies HEAD, so --dirty <hash> is nonsense
11:08 Anticom Seveas: but even `git describe HEAD --dirty` doesn't work
11:08 vuoto joined #git
11:09 sebascar joined #git
11:09 Seveas that's a bit naughty, but somewhat understandable
11:09 jceb joined #git
11:09 Seveas (it's easier to reject any commitish than whitelist HEAD)
11:09 Anticom hm sure
11:09 Anticom however i still don't quite understand *why* this doesn't make sense
11:10 Anticom why does the <mark> have to be related to HEAD?
11:13 Seveas because the working tree is related to HEAD
11:14 heftig the working tree is dirty if it differs from the index
11:14 Seveas your working tree will always have changes compared to older commits
11:14 BlessJah I'd say index worktree is related to index
11:14 heftig the index is a tree (not a commit) related to HEAD
11:14 heftig it doesn't make sense to describe an arbitrary commit, but then look for differences between tree and index
11:14 heftig working tree and index
11:14 Seveas BlessJah: they're all related :)
11:15 Seveas one big happy family
11:15 BlessJah is that why git cli syntax is kinda retarded?
11:16 heftig Seveas: git commit ancestries tend to be rather incestuous :p
11:16 Seveas BlessJah: no, there are many reasons for that :)
11:16 fuchstronaut joined #git
11:16 Seveas heftig: heh.
11:16 Anticom Well basically i'd like to a) describe my commit to use that as version information for packaging and b) i'd like to append -dirty to the version if my working copy isn't clean to ensure there are no packages published from any developer who packaged that from a dirty repo
11:17 Anticom IMO it would be more convenient if i could do this in one command
11:17 Anticom i guess i'll have to do two git describes then
11:17 Seveas Anticom: that makes no sense
11:17 Anticom Well I'm not a git pro
11:17 Seveas if you use a commit to create a package from, the use the files from that commit, not the working tree
11:17 Seveas so you can't have a 'dirty' package.
11:18 zarubin joined #git
11:18 BlessJah ++ for not allowing 'dirty' packages
11:18 ojdo joined #git
11:18 Anticom Seveas: I'm using this ( https://github.com/rpavlik/cmake-modules/blo​b/master/GetGitRevisionDescription.cmake#L83 ) cmake module to get the version information
11:18 BlessJah maintenance would be nightmare as you won't be able to reproduce or even identify changes
11:19 Anticom and there ${hash} is passed to git describe without me being in control
11:19 Seveas yuck, cmake
11:19 Anticom i wanted to avoid modifying external packages if possible
11:19 Seveas not even going to try to decipher that, sorry.
11:19 Anticom hence i was asking
11:19 Lunatrius joined #git
11:19 Anticom basically i's just doing `git describe <hash> <extra arguments>`
11:20 a_thakur joined #git
11:20 Anticom BlessJah: this was the idea, yea :)
11:20 geekbri joined #git
11:21 BlessJah package creation is something that should happen on CI where dev won't be able to poison env
11:22 str joined #git
11:23 sebascar joined #git
11:24 rafalcpp how to make git always by default use --show-siganture e.g. in git log?
11:25 robi251 joined #git
11:26 uservblbla joined #git
11:26 Lakshya joined #git
11:26 marcogmonteiro joined #git
11:27 Lakshya left #git
11:28 uservblbla left #git
11:29 Seveas Anticom: instead of trying to do this with git-describe, why not exit earlier in the build process if the output of git statu -s isn't empty?
11:30 Lakshya joined #git
11:30 Anticom Seveas: because that would force all devs to commit all changes before doing a build which is kind of odd if you're building the software to check everything is okay but you're not allowed to do that unless you commit your (potentially faulty) code
11:31 sanketdg joined #git
11:32 playfullyExist joined #git
11:38 ninjaaron I just tried `time echo -n | python` on 2 and 3. 2 was 25ms, 3 was 69 ms (perl did .4 ms!). This seems to be consistant with startup times I've observed on trivial scripts. Any idea what makes the difference?
11:38 ninjaaron s/.4/4
11:39 BlessJah Anticom: build != test != package
11:39 BlessJah last step should only involve creating archive, be it zip, deb or rpm
11:39 BlessJah .1
11:39 bartosz_ joined #git
11:39 mda joined #git
11:40 selckin ninjaaron: strace python
11:40 ninjaaron I'm in totally not the channel I thought.
11:40 ninjaaron :D
11:40 rorro joined #git
11:43 Seveas ninjaaron: your timing is flawed
11:43 Seveas 1) 'time' is not a benchmark, 2) time foo | python doesn't time python
11:44 ninjaaron Seveas: 1. true enough 2. on my system, it give output for both processes.
11:44 anuxivm joined #git
11:45 elementalest joined #git
11:46 mlopes joined #git
11:47 kc8wxm joined #git
11:47 Seveas $ (time python2 -c1; time python3 -c1; time perl -e1) 2>&1 | grep real
11:47 Seveas real0m0.013s
11:47 Seveas real0m0.021s
11:47 Seveas real0m0.002s
11:50 vuoto joined #git
11:50 ninjaaron so yeah, I got slightly different times with that, but almost the same time difference.
11:51 Seveas doesn't surprise me too much. Python loads an awful number of things during startup
11:52 satya4ever joined #git
11:52 Seveas seahawk:~/git/git (gitweb-subdirs)$ strace -f -eopen python -c1 2>&1 | grep '^open(.* = [0-9]' | wc -l
11:52 Seveas 68
11:52 Seveas seahawk:~/git/git (gitweb-subdirs)$ strace -f -eopen perl -e1 2>&1 | grep '^open(.* = [0-9]' | wc -l
11:52 Seveas 9
11:53 mohabaks joined #git
11:53 drbean joined #git
11:53 ninjaaron It seems like ther is an inverse correlation between how fast these interpreted languages start up and how fast they are after that.
11:54 davisonio joined #git
11:55 qsx ghc -e 1  0,11s user 0,02s system 95% cpu 0,136 total
11:55 cqi joined #git
11:56 Seveas qsx: /join #darcs
11:56 qsx lol nope
11:57 qsx well that doesn’t really count anyway, haskell isn’t an interpreted language
11:57 qsx … neither is perl, for that matter.
11:57 Sceorem joined #git
11:58 ninjaaron qsx: does perl compile to machine code prior to runtime?
11:58 qsx ninjaaron: i think it’s more of a bytecode-vm, but yes
11:58 Seveas as do python and ruby
11:58 Seveas the line between 'interpreted' and 'compiled' is blurry, even more so for perl
11:58 iota_pi joined #git
11:58 Seveas because some perl code 'runs' at 'compile time'
11:59 qsx http://dpaste.com/3YQ625J
11:59 qsx well
11:59 xlegoman joined #git
11:59 qsx this might explain it a bit.
11:59 star_prone joined #git
11:59 sin3t joined #git
12:00 ChoHag joined #git
12:00 robi251 joined #git
12:00 sin3t joined #git
12:00 ninjaaron I do like that perl compiles regex literals ahead of time.
12:00 ninjaaron very handy.
12:00 ChoHag If I have a branch with a few commits on it, how can I break that branch up into two parallel branches?
12:01 ChoHag Conveniently the code which should be on the other branch is all under one subdirectory.
12:01 Seveas ChoHag: reset to the start of that branch, make new branch, cherry-pick commits that touch that branch, repeat whole thing for second branch
12:02 Seveas (assuming all commits only touch either one or the other side)
12:02 ChoHag The commits are all mingled together.
12:02 ChoHag For each commit I need to split it into one of each branch.
12:03 pa so i tried to do as you guys suggested. If i have local featureBranchA and local featureBranchB, and i change featureBranchA, then i rebase featureBranchB with "git rebase featureBranchA featureBranchB". But every time i do this, there are conflicts on the changes i make. Like git always say that the modified files in both branches are changed, while in fact the files in featureBranchB are just the old version of what was in featureBranchA
12:03 ChoHag It's not a huge problem losing or squashing history from the individual commits.
12:03 pa is there any way to explain this to git, so that it can do a better job at rebasing?
12:03 hahuang65 joined #git
12:04 tvw joined #git
12:05 anuxivm left #git
12:05 mischat joined #git
12:05 psprint joined #git
12:06 psprint Hello. If there is ur.*.insteadOf configured for a remote, will the url be used in FETCH_HEAD after git fetch/pull?
12:06 pa even more because i end up fixing the conflicts, and then i have to always do git rebase --skip, because nothing has changed
12:10 anuxivm joined #git
12:10 sdothum joined #git
12:11 pa kadoban, ^
12:11 kyan joined #git
12:12 Darcidride_ joined #git
12:13 JanC_ joined #git
12:14 cqi joined #git
12:14 davisonio joined #git
12:17 fuchstronaut joined #git
12:17 vuoto joined #git
12:20 LiohAu joined #git
12:22 zarubin joined #git
12:23 Maple__ joined #git
12:23 kedare joined #git
12:23 interne7y joined #git
12:24 synthroid joined #git
12:28 danslo joined #git
12:28 p4trix joined #git
12:29 stasmakarov joined #git
12:30 pasadinhas joined #git
12:32 pioiytr_ joined #git
12:33 wxg joined #git
12:35 pa actually, i was wondering: would there be an alternative to this approach? like if i have commit A, and then commit B, and i realize there is a change i have to do it on A, can i do the change to B and propagate it to A, so that no conflicts will happen?
12:37 _Mica_ joined #git
12:38 Softeisbieger joined #git
12:39 _ikke_ When reapplying commits, git only looks at the changes the commits make themselves compared to the previous commit. You have to make sure those changes do not conflict with each other
12:39 star_prone joined #git
12:39 HoierM joined #git
12:40 pa _ikke_, yes, exactly, so if i change the base (A), then naturally there will be conflicts when rebasing B onto the new A. But if i could change B, and then somehow tell git "look, this change belongs to A, actually", then there wouldnt be conflicts, right?
12:43 veilig2000 joined #git
12:47 byteflame joined #git
12:48 _ikke_ pa: No, not every change will cause conflicts
12:48 synthroi_ joined #git
12:51 _Mica_ left #git
12:51 dsdeiz joined #git
12:53 basiclaser joined #git
12:54 synthroid joined #git
12:55 synthroi_ joined #git
12:55 eizua joined #git
12:55 davisonio joined #git
12:56 _ikke_ pa: when git reports conflicts, it means that a change in the same file on the same lines has occured in both branches
12:57 pa _ikke_, well, in my case i had A -> B, and i didnt change in B the lines i wanted to update in A
12:57 pa but still, rebase stopped and i had conflict blocks to resolve
12:58 eijk joined #git
12:58 heftig pa: it's enough for changed lines to be adjacent to cause a conflict
12:59 pa ah i see.. heftig but so there's no option like "git add file" and then commit the file to this change and the previous N along the history?
13:00 heftig huh?
13:00 heftig could you elaborate?
13:01 pa so if i have A -> B -> C, and i i make a change in C that should rather belong to A, could i rather be on C, commit the change and somehow tell git to propagate it back in the history?
13:01 tobiasvl pa: there is something called filter-branch which can do that
13:01 pa like for the previous 2 changes
13:01 CheckDavid joined #git
13:01 pa ah, interesting
13:01 DolphinDreamm joined #git
13:02 jimi_ joined #git
13:02 ktosiek left #git
13:03 anuxivm left #git
13:04 danslo pa: what I sometimes do in that situation, is just add a commit D with that change, start an interactive rebase, then squash D into A. careful when rewriting history though :)
13:04 hahuang65 joined #git
13:04 HanClinto joined #git
13:04 livingstn joined #git
13:04 pa danslo, ah interesting. so you rebase B into D before squashing?
13:05 pa rather "B on top of D"
13:05 pa no actually that might not work
13:05 pa or
13:05 pa yes it would
13:05 pa if you change only that small thing
13:06 ksolv joined #git
13:07 al-damiri joined #git
13:09 _ikke_ pa: look at git commit --fixup
13:09 _ikke_ and git rebase --autosquash
13:09 pcre joined #git
13:10 govg joined #git
13:11 timvishe` joined #git
13:11 danslo _ikke_: oh. nice
13:12 circ-user-aUJuR joined #git
13:15 beyertech joined #git
13:15 azerus joined #git
13:16 beyertech joined #git
13:18 Raging_Hog joined #git
13:18 fuchstronaut joined #git
13:19 Darcidride__ joined #git
13:20 Raging_Hog joined #git
13:21 ToBeCloud joined #git
13:22 robi251 joined #git
13:23 DolphinDreamm howdy
13:23 DolphinDreamm for scripting how do you check if a repo is empty ?
13:24 Segfault_ joined #git
13:24 Gamecubic joined #git
13:24 star_prone joined #git
13:25 ExoUNX joined #git
13:25 larsxschneider joined #git
13:25 eijk_ joined #git
13:27 rubyonrailed joined #git
13:27 larsxschneider left #git
13:28 nilsi joined #git
13:29 kyan joined #git
13:30 HappyLoa- joined #git
13:31 GavinMagnus left #git
13:32 Raging_Hog joined #git
13:33 thekrynn joined #git
13:33 _ikke_ I think easieast is to check if HEAD points to a valid branch
13:33 Anticom DolphinDreamm: define empty
13:34 _ikke_ although that can also happen when using git checkout --orphan
13:34 Gloomy joined #git
13:34 _ikke_ Why do you need to know this?
13:35 DolphinDreamm Anticom, a new repo was just created and has  no objects in it
13:36 neurodrone_ joined #git
13:36 m0nkmaster joined #git
13:37 Anticom DolphinDreamm: http://stackoverflow.com/a/8874599/966530
13:37 lb DolphinDreamm http://superuser.com/questions/352289/b​ash-scripting-test-for-empty-directory
13:37 Anticom not the prettiest solution but should do the trick
13:37 Anticom also i wouldn't locate objects via find
13:37 GavinMagnus joined #git
13:37 _ikke_ git rev-list --objects --count
13:38 Anticom _ikke_: complains if no commit at all because there's no HEAD yet :)
13:38 jaafar joined #git
13:38 lb http://stackoverflow.com/questions/​5491832/how-can-i-check-whether-a-g​it-repository-has-any-commits-in-it
13:38 Anticom _ikke_: oh nvm, i wen't w/o --objects and that failed
13:39 _ikke_ --objects overrules --count, you'd need to add | wc -l
13:39 DolphinDreamm how about git count-objects ?
13:39 _ikke_ "count unpacked number of objects"
13:39 psprint After git clone, I see additional branches only as remotes/origin/<branchname> (git branch -avv). Why clone didn't download the branches? How to download them now?
13:39 _ikke_ packed objects are ignored
13:40 _ikke_ psprint: it did download the branches
13:40 DolphinDreamm ah.. so i can have objects/commits but not be listed via count-objects ?
13:40 _ikke_ !remote_tracking
13:40 gitinfo [!remote_tracking_branch] Remote-tracking branches (branches which start with e.g. 'origin/', listed by 'git branch -r') are read-only mirrors of the corresponding branches in another repository. They're updated by 'git fetch'. You can't edit them directly (trying to check them out results in a !detached HEAD), but you can create a new local branch based on a remote-tracking branch using e.g. 'git checkout -b <branch> <remote
13:40 Raging_Hog joined #git
13:40 psprint _ikke_: ok, I get what you mean, but why no local branches?
13:40 _ikke_ psprint: git by default only creates one local branch
13:40 _ikke_ The others you can just create when you need them
13:40 Anticom psprint: read the last scentence of gitinfo
13:42 psprint Anticom: worked, thanks
13:42 DolphinDreamm _ikke_, the git rev-list --objects --count gives me 0  even though there are 508 objects .
13:43 osse DolphinDreamm: add an --all
13:43 osse or HEAD
13:43 osse or origin/master
13:43 submitnine joined #git
13:44 DolphinDreamm osse, im trying to figure out if the repo has NO commits. i.e. a empty repo was created on remote server.. and cloned locally, thus it has no commits yet.
13:44 Dougie187 joined #git
13:46 PCatinean I guess this is the only approach to my problem: 1 repo for each module for stable builds, 1 repo for each module for development, 1 repo for each module but for a different version
13:46 PCatinean Sounds kinda hard to maintain imo
13:48 danslo PCatinean
13:48 danslo oops, sorry for highlighting
13:49 xlegoman joined #git
13:52 a_thakur joined #git
13:53 osse DolphinDreamm: that's easy enough; check if <git>/objects only contains the empty directories info and pack
13:54 m0nkmaster joined #git
13:54 zeroed joined #git
13:54 DolphinDreamm osse, i used git count-objects but i guess this only reports unpacked objects.. as for checking for info/pack folders could it be that the the repo has commits but are all packed ?
13:55 dob1 joined #git
13:55 osse DolphinDreamm: that's why I said that info and pack should be empty
13:55 _ikke_ yes, just after you cloned, all objects are packed
13:55 PCatinean danslo, had a suggestion?
13:56 ash_workz joined #git
13:56 phaleth joined #git
13:56 cdg joined #git
13:56 AaronMT joined #git
13:56 dob1 hi, i am a beginner about git, or at list i used list most of the times via egit (eclipse plugins), now i am using it from the command line (1.9.x version, windows) and i have doubts about the "git add" command,  the . or *, how to manage new/deleted files, there are 20000 tutorials, can someone give me an advice to what to read?
13:56 dob1 *least
13:57 AaronMT joined #git
13:57 playfullyExist joined #git
13:57 bremner dob1: !book
13:57 gitinfo dob1: There are several good books available about git; 'Pro Git' is probably the best: http://git-scm.com/book but also look at !bottomup !cs !gcs !designers !gitt !vcbe and !parable
13:57 bremner dob1: man gittutorial is also ok, and shorter
13:57 gitinfo dob1: the gittutorial manpage is available at http://jk.gs/gittutorial.html
13:58 bremner look for "MAKING CHANGES"
13:58 _ikke_ dob1: (I strongly recommend updating to the latest git for windows version, which should be around 2.9)
13:59 dob1 thanks for the links/informations
14:00 GodGinrai joined #git
14:01 thiago joined #git
14:01 mkoskar joined #git
14:03 kaldoran joined #git
14:05 hahuang65 joined #git
14:05 ash_workz joined #git
14:05 davisonio joined #git
14:05 sangy joined #git
14:07 beyertech joined #git
14:08 ResidentBiscuit joined #git
14:08 l4v2 joined #git
14:13 star_prone joined #git
14:16 Derperperd joined #git
14:18 shinnya joined #git
14:18 cdown joined #git
14:18 ahmedelgabri joined #git
14:18 fuchstronaut joined #git
14:19 tabrez joined #git
14:20 cdown_ joined #git
14:20 Gloomy joined #git
14:21 midgetspy joined #git
14:21 beyertech joined #git
14:22 zarubin joined #git
14:24 interne7y joined #git
14:25 moei joined #git
14:26 playfullyExist joined #git
14:27 TooLmaN joined #git
14:28 playfullyExist I am really confused as to why github doesn't track contents of one of my folders I commited
14:28 playfullyExist I have this repository I created
14:28 playfullyExist https://github.com/nikitavoloboev/treehouse-ios
14:29 playfullyExist and here is the tracked folder : http://i.imgur.com/eGlHYOM.png
14:29 playfullyExist I did git add .
14:29 playfullyExist git commit . -m 'init'
14:29 playfullyExist and git push
14:29 zeroed joined #git
14:29 zeroed joined #git
14:30 huptoka joined #git
14:30 _ikke_ playfullyExist: what does git status say?
14:31 playfullyExist _ikke_ http://i.imgur.com/a88zd2W.png
14:31 axl_ joined #git
14:31 _ikke_ what does git ls-tree HEAD InteractiveStory return?
14:32 playfullyExist _ikke_ http://i.imgur.com/TrmGJr6.png
14:35 Kicer86 joined #git
14:36 _ikke_ Right
14:36 _ikke_ You have a .git dir inside InteractiveStory
14:36 _ikke_ That means git will ignore the contents of that directory
14:38 midgetspy hey all I'm having trouble with git failing to do a rebase. I'm starting with a clean working dir (no untracked files, no pending changes) and doing an interactive rebase (git rebase -i <sha>) and partway through it dies saying my local changes would be overwritten by merge, please commit or stash. I have tried on two machines including from a fresh clone. git 2.9.2 on Windows... any ideas? I've
14:38 midgetspy searched and only found some mentions of background processes doing timestamp modifications but that doesn't seem to be my problem as far as I can tell
14:39 _ikke_ Files no longer being ignored can cause this
14:39 _ikke_ or ignored files being tracked
14:39 codeman joined #git
14:40 nikitasius joined #git
14:40 Gitzilla joined #git
14:43 Impaloo joined #git
14:45 Impaloo joined #git
14:47 PCatinean joined #git
14:49 Impaloo joined #git
14:50 playfullyExist _ikke_ thank you, this worked :)
14:50 avar Is there no -F like git commit -F for git-merge?
14:50 Cata|Work does anyone know how to fix bad object headers?
14:53 Impaloo joined #git
14:54 avar To answer my own question, seems not and I need git merge --no-edit ... && git commit --amend -F ...
14:55 fmcgeough joined #git
14:57 GetteracWork joined #git
14:58 thefifthsetpin joined #git
14:58 LionsMane joined #git
14:59 nydel07 joined #git
14:59 _ikke_ avar: when you use --no-edit, --amend is not necessary
14:59 _ikke_ (you'd change the previous commit)
15:01 ArSn joined #git
15:02 GetteracWork Hi, so I reverted a commit in a remote repo (in Gitlab), then fetched and merged the branch into my local repo... now the local branch and the remote branch have different files and yet doing a "git status" locally says they're in sync.  What the heck is going on? I can't push my local files because everything is supposedly up to date.
15:02 damjan what different files?
15:03 avar _ikke_: that's what I want
15:03 ahmedelgabri joined #git
15:04 GetteracWork damjan: The goal was to roll back to version X, but i incorrectly did a revert.. now all the changes in that reverted commit are gone on the remote repo.
15:05 Codfection joined #git
15:05 damjan that's how revert works, right?
15:05 hahuang65 joined #git
15:06 GetteracWork damjan: i guess so... i fetched/merged to get my local repo in sync, but the revert didn't show up, now the files are out of sync.
15:06 cdg joined #git
15:06 damjan huh
15:07 damjan what is a gitlab revert then?
15:07 GetteracWork damjan: i have no idea... it created a new remote branch and such. Not sure what it did behind the scenes.
15:08 damjan did you then sync that remote branch?
15:10 GetteracWork damjan: i fetched all of the remote branches, and merged master, not the new branch... wasn't sure what the new branch was for...
15:11 tyreld joined #git
15:11 happy-dude joined #git
15:11 damjan I don't know what you did on gitlab
15:12 GetteracWork damjan: me either :(
15:12 ahmedelgabri joined #git
15:12 wget joined #git
15:13 _Mica_ joined #git
15:13 _Mica_ left #git
15:13 wget Hi guys. Wondering if there is a typo here: https://github.com/progit/progit2/blame/master/b​ook/03-git-branching/sections/nutshell.asc#L174
15:13 wget "in actuality"
15:13 GetteracWork i don't even know how to clean up this mess... i just want to go back before the revert... should i "git reset" to a version prior to the revert?
15:13 wget First time I heard such a statement in English.
15:14 wget I'm not a native, so cannot really say...
15:14 damjan GetteracWork: but you said you don't have the revert in your local repo
15:15 perlpilot wget: looks fine to me, if a little wordy.
15:15 Guest55 joined #git
15:15 Noldorin joined #git
15:16 wget perlpilot: I would have simply used "actually". "In actuality" sounds like the guy was translating litterally from French ;-) (I'm French native)
15:17 perlpilot wget: Aye, me too, but there's nothing wrong with it as written except perhaps "style"  :)
15:18 wget perlpilot: ok, let's leave it as-it then ;-) But such a statement was hurting my eyes. It lets me remember the old ages when I was learning English :D "false friends , etc."
15:18 dikiaap joined #git
15:19 fuchstronaut joined #git
15:20 madewokherd joined #git
15:21 mischat joined #git
15:21 kc8wxm left #git
15:22 danslo joined #git
15:22 dreiss joined #git
15:22 zarubin joined #git
15:23 veilig2000 joined #git
15:23 Raging_Hog joined #git
15:24 gargola joined #git
15:24 midgetspy I have a branch off master which I want to maintain but skip a couple commits that are in master. What's the best way to do this so that when I merge new commits from master it'll just fast forward and won't try to merge the missing commit? If I make a branch then rebase and skip the commits I don't want will that work?
15:25 osse midgetspy: revert them in your branch
15:27 midgetspy osse: yeah that'll work, I thought there might be a better way. thanks.
15:27 vassagus joined #git
15:28 m0nkmaster joined #git
15:28 claudiop joined #git
15:30 cdg joined #git
15:31 synthroid joined #git
15:31 str joined #git
15:32 tristanp joined #git
15:32 axl__ joined #git
15:33 InfoTest joined #git
15:34 smithbone joined #git
15:35 rwb joined #git
15:35 angular_mike any git svn alternatives for cygwin?
15:37 selckin !msysgit
15:37 gitinfo Git for Windows, formerly referred to as msysgit, is available at http://git-scm.com/downloads
15:37 burtons joined #git
15:37 GavinMagnus joined #git
15:39 damjan angular_mike: or that new shiny WSL (windows subsystem for linux) or as they called it bash on windows
15:39 davisonio joined #git
15:39 nilsi_ joined #git
15:40 subhojit777 joined #git
15:41 nilg joined #git
15:42 GetteracWork damjan: yeah i don't know how the history of the two repos got out of sync... ended up trashing the local repo and starting over from the remote. :X
15:43 zacts joined #git
15:46 tcurdt joined #git
15:47 tcurdt 'git log' shows tags that are missing from 'git tag' ... how can that be?
15:47 tcurdt I am the only one working on that repo btw
15:48 angular_mike damjan: what SVN features it got?
15:48 damjan angular_mike: it's got everything that you have on linux
15:48 osse tcurdt: maybe they aren't actually tags but branches
15:49 tcurdt osse nope
15:49 angular_mike damjan: but I don't need it
15:49 angular_mike damjan: I need what I don't have
15:49 osse tcurdt: can you show
15:50 rorro joined #git
15:51 tcurdt osse https://gist.github.com/tcurdt/2​dfbb51529dc815d87f72933969b2656
15:51 lss8 joined #git
15:51 tcurdt the log is just a partial output of course
15:51 tcurdt so it says "(tag: alpha-1.0.0-154)"
15:51 tcurdt but there is so such tag
15:52 tcurdt (although it actually should be there)
15:52 osse tcurdt: take a closer look at line 12
15:52 tcurdt darn!
15:52 osse \o/
15:52 Cata|Work has anyone ever had a .pack file be corrupted?
15:52 seishun joined #git
15:52 tcurdt tricked by sorting!
15:53 swift1 joined #git
15:53 _ikke_ Cata|Work: I haven't experienced it myself
15:53 GetteracWork damjan: thanks for your help and suggestions!
15:53 GetteracWork left #git
15:55 Seveas Cata|Work: sure, given enough time, all files go corrupt. Haven't seen git corrupt one yet though.
15:55 Cata|Work I have a file that everytime I commit it I get a corrupt .pack file so my pushes fail.
15:56 Cata|Work I've tried reseting the branch and deleting the .pack file before retrying the whole thing, but it just generated a new corrupt file with the same name (must be that the names are shas or CRCs
15:56 Electrometro joined #git
15:56 crayon joined #git
15:56 Seveas err, deleting the .pack? That's bad, now your repo is missing data..
15:57 The_Phoenix joined #git
15:57 vassagus joined #git
15:58 The_Phoenix joined #git
15:58 Seveas what was the original error you got? Git also speaks the packfile protocol during push, so if push was complaining, it may have complained about the in-flight pack
15:59 Darren_ joined #git
15:59 Remram[m] joined #git
15:59 The_Phoenix joined #git
16:00 ahmedelgabri joined #git
16:01 LeBlaaanc joined #git
16:01 Cata|Work the original error was bad object header.
16:04 LeBlaaan_ joined #git
16:04 Cata|Work it is also a fairly large file
16:04 Cata|Work but, I haven't had any other issues with large files in this repo
16:05 Seveas try running git fsck. I'm expecting it to throw tons of unrecoverable errors now that you deleted the .pack file
16:05 davisonio joined #git
16:06 Cata|Work I removed the .pack file after reseting the repo
16:06 Cata|Work I'm running fsck right now
16:06 Seveas there's no such thing as 'reseting the repo'. What do you mean woith that?
16:06 hahuang65 joined #git
16:07 Cata|Work I did a hard reset on the branch to a commit that didn't have the issue.
16:07 Seveas that only moves a branch pointer, doesn't do anything with your packfiles.
16:08 Cata|Work of course it doesn't... well the only error so far is a bad object header identical to the original error.
16:08 Seveas pastebin the output
16:08 Cata|Work I guess if I have to I'll blow away the repo and clone it again.
16:08 Cata|Work but I'll pastebin the output first once it finishes.
16:09 Seveas if there's nothing in there you want to keep, that's definitely the easiest option.
16:10 vuoto joined #git
16:11 Cata|Work http://pastebin.com/iVQkKEnx
16:15 vuoto joined #git
16:17 Electrometro joined #git
16:20 fuchstronaut joined #git
16:22 TomyLobo joined #git
16:22 zarubin joined #git
16:23 tcurdt what I am wondering: how is it possible to amend a commit message - wouldn't that also change the sha?
16:25 vuoto joined #git
16:26 phanimahesh joined #git
16:27 dreiss joined #git
16:32 davisonio joined #git
16:33 netj joined #git
16:33 durham joined #git
16:33 b1tchcakes joined #git
16:39 Seveas it does
16:39 Seveas git commit --amend creates a new commit
16:40 Seveas search for 'changing the latest commit' on https://git.seveas.net/undoing-all-kinds-of-​mistakes.html#undoing-all-kinds-of-mistakes and click the 'git commit --amend' link
16:41 intellix joined #git
16:42 zacts joined #git
16:44 dhollinger joined #git
16:44 EvilPenguin joined #git
16:46 Seveas Cata|Work: hmm, is this now a different corrupt packfile or did you not remove the first broken packfile?
16:47 Electrometro joined #git
16:48 Cata|Work it appears to have either been regenerated or it was added back when I did a pull
16:49 orbitphreak joined #git
16:49 Cata|Work the filename matches the original corrupt file
16:51 eivarv joined #git
16:52 daynaskully joined #git
16:53 csd_ joined #git
16:54 Mannn joined #git
16:55 csd_ My repo has a tests/ folder with another small git repo inside of it. I've added a config file to it, e.g. tests/subRepo/conf/conf.yml. Git doesn't want to let me git add tests/* because it wants it to be a submodule. What's the best thing for me to do? I don't think I can make it a submodule without also having to create a script to copy the config file from a parent directory, right?
16:56 csd_ the sub repo is small so i'm contemplating just deleting the .git directory but i feel like a better way to do it exists
16:58 SteffanW2 joined #git
16:58 tcurdt Seveas awesome. great link. thanks!
16:59 kus joined #git
16:59 alfy joined #git
16:59 tjone270 joined #git
17:00 tcurdt shouldn't these all be the same? `git hash-object FILE` == `git ls-files -s FILE` == `git rev-parse :FILE` == `(echo -ne "blob `wc -c < FILE`\0"; cat FILE) | openssl dgst -sha1`
17:00 daveomcd joined #git
17:00 durham_ joined #git
17:00 durham_ joined #git
17:01 tcurdt or in other words: I would love to understand how to match the `openssl dgst -sha1` to what's in the index
17:01 daveomcd if I have my master branch, make a new branch called "branch1" then make checkout/make changes to "branch1" -- proceed to checkout master, make changes on master... how do I get those changes over to "branch1" w/o affecting master?
17:01 Cata|Work merge master into Branch1
17:02 Cata|Work it won't add a commit to the master branch
17:02 daveomcd Cata|Work, whats difference between doing - git merge origin/master && git rebase master ?
17:02 daveomcd i saw both of those options and got confused
17:02 Cata|Work I'm not the one to really ask about that
17:02 daveomcd ahh ok
17:02 daveomcd thanks!
17:03 Darren_ joined #git
17:03 LeBlaaanc joined #git
17:03 lolmac joined #git
17:03 slayerjain joined #git
17:03 Cata|Work you are welcome
17:06 jbitdrop joined #git
17:06 synthmeat joined #git
17:07 hahuang65 joined #git
17:10 nitric joined #git
17:12 jbitdrop joined #git
17:12 bsanford joined #git
17:14 crose joined #git
17:14 fahadash joined #git
17:17 VladGh joined #git
17:17 kbs joined #git
17:21 EvilPenguin joined #git
17:22 zarubin joined #git
17:23 danslo_ joined #git
17:23 phanimahesh joined #git
17:23 durham joined #git
17:30 tvw joined #git
17:31 nikitha joined #git
17:31 zacts joined #git
17:32 Limes__ joined #git
17:33 robotroll joined #git
17:34 VladGh joined #git
17:35 VladGh joined #git
17:36 jottr joined #git
17:36 ferr joined #git
17:38 VladGh_ joined #git
17:38 phanimahesh joined #git
17:41 sweatsuit joined #git
17:42 LionsMane joined #git
17:44 Seveas tcurdt: there's some zlib'ing involved
17:45 Seveas though I'm not sure if hashing is done before or after
17:46 _ikke_ before
17:46 _ikke_ https://stackoverflow.com/questions/1869​885/why-is-the-python-calculated-hashlib​-sha1-different-from-git-hash-object-for
17:46 Seveas I keep forgetting :)
17:46 jost joined #git
17:46 _ikke_ me too :P
17:46 _ikke_ I asked that question myself
17:46 _ikke_ had to look it up
17:47 Seveas heh
17:47 thiago joined #git
17:47 dmto joined #git
17:48 parasite joined #git
17:48 tango_ joined #git
17:49 eivarv joined #git
17:50 LeBlaaanc joined #git
17:50 dreiss joined #git
17:50 darkbit joined #git
17:52 showaz joined #git
17:57 shinnya joined #git
17:59 mikecmpbll joined #git
17:59 hamster joined #git
17:59 _ikke_ Makes sense, that way your hash is not depending on the compression method used, and you can change it later
18:00 _ikke_ ie, the compression is a implementation detail
18:01 sybariten can "origin" and "master" be explained better in newbie layman terms? I have .ini files on my PC that i have changed slightly, i have the same .ini files on my laptop that i may have changed slightly, and i have a repo on bitbucket. Now it's time to chose what should take precedence
18:01 sybariten ANd i know there are diff commands, but i'm a bit confused
18:02 osse sybariten: origin is a common name for a remote repository
18:02 osse sybariten: master is a branch
18:02 osse it's apples and oranges
18:02 osse or apples and lobsters, as my math teacher once put it
18:03 sybariten aha aha
18:04 sybariten is there a way to see how many branches i have? In order to understand what a branch is.
18:04 osse 'git branch'
18:04 faux___ joined #git
18:04 osse -r or -a optional
18:04 sybariten says master....
18:04 osse then you only have one
18:05 sybariten on my PC... lets see what the lappie says
18:05 sybariten says master there too
18:06 osse important point:  branches are only local. the lappie master is different from whatever other masters there are
18:07 mischat joined #git
18:08 hahuang65 joined #git
18:09 tang^ git origin -v will show you your origins. git branch --all will show you branches local and remote
18:09 LeBlaaanc joined #git
18:10 _ikke_ git origin                                                                                                                    ~/projects/git.git
18:10 _ikke_ git: 'origin' is not a git command. See 'git --help'.
18:10 _ikke_ You probably meant remote
18:11 brandonw joined #git
18:12 ahmedelgabri joined #git
18:13 * tang^ blinks
18:13 tang^ oops. sorry... yes I did mean git remote
18:14 davisonio joined #git
18:15 GenZai joined #git
18:18 LeBlaaanc joined #git
18:18 Sasazuka joined #git
18:20 Derperperd joined #git
18:21 fuchstronaut joined #git
18:21 fuchstronaut joined #git
18:21 sybariten osse: ok, good to know!  :-)
18:22 zarubin joined #git
18:23 j416 joined #git
18:23 daniel_rose joined #git
18:24 LeBlaaanc joined #git
18:24 jubalh joined #git
18:24 jubalh joined #git
18:25 Balliad joined #git
18:26 interne7y joined #git
18:27 docnvk joined #git
18:27 str joined #git
18:27 LeBlaaan_ joined #git
18:29 burtons joined #git
18:29 LeBlaaanc joined #git
18:32 paul424 joined #git
18:33 Aljone_1 left #git
18:34 jimi_ joined #git
18:36 kulelu88 joined #git
18:37 jimi_ joined #git
18:39 AaronMT joined #git
18:39 burtons joined #git
18:40 anth0ny joined #git
18:41 zivester joined #git
18:41 kfunk joined #git
18:41 Cassiopaya joined #git
18:42 dikiaap joined #git
18:42 AaronMT_ joined #git
18:44 IlNominePatre joined #git
18:45 cdg joined #git
18:48 GenZai joined #git
18:49 danslo_ joined #git
18:49 star_prone joined #git
18:49 abec0_ joined #git
18:49 harv_ joined #git
18:49 rasto_ joined #git
18:49 Ariadeno_ joined #git
18:49 victorbjelkholm_ joined #git
18:49 Patrick- joined #git
18:50 spb_ joined #git
18:50 larrybolt` joined #git
18:50 HedgeMage_ joined #git
18:50 BeerSerc_ joined #git
18:50 adante_ joined #git
18:50 heinrich5991_ joined #git
18:50 legumbre_ joined #git
18:50 BrianBla- joined #git
18:50 ChoHag_ joined #git
18:50 boshhead_ joined #git
18:50 thejoecarroll_ joined #git
18:50 milki_ joined #git
18:50 yossarian joined #git
18:50 zapb_ joined #git
18:50 suy joined #git
18:50 Karunamon joined #git
18:50 RiPer joined #git
18:50 nd_ joined #git
18:50 tinanoo joined #git
18:50 warthog9 joined #git
18:50 ngladitz joined #git
18:50 ngladitz joined #git
18:50 led_belly joined #git
18:50 cebor joined #git
18:50 MacGyver joined #git
18:50 fakenullie joined #git
18:50 rafalcpp joined #git
18:50 woodruffw joined #git
18:50 jleclanche joined #git
18:50 maqr joined #git
18:50 Xandaros_ joined #git
18:50 lebster joined #git
18:50 snowinferno joined #git
18:51 jumpman joined #git
18:51 malesca joined #git
18:52 MLM joined #git
18:52 ELLIOTTCABLE joined #git
18:52 ki0 joined #git
18:52 Tratos joined #git
18:53 zyxep joined #git
18:53 nikitha joined #git
18:53 jaeckel joined #git
18:54 saidelike_ joined #git
18:54 kapu joined #git
18:55 gusnan joined #git
18:55 nikitha joined #git
18:56 burtons joined #git
18:56 nikitha joined #git
18:57 srcerer joined #git
18:57 Electrometro joined #git
18:58 nikitha joined #git
18:59 ShekharReddy joined #git
18:59 Peuc joined #git
19:01 nikitha joined #git
19:02 LeBlaaanc joined #git
19:03 nikitha joined #git
19:03 Mission-Critical joined #git
19:05 nikitha joined #git
19:05 LeBlaaan_ joined #git
19:06 davisonio joined #git
19:07 nikitha joined #git
19:08 mischat joined #git
19:08 LeBlaaanc joined #git
19:09 nikitha joined #git
19:09 noteness joined #git
19:09 sulky joined #git
19:11 tcurdt Seveas _ikke_ but then (echo -ne "blob `wc -c < FILE`\0"; cat FILE) | openssl dgst -sha1` should give the git sha - but seems like it doesn't. Or what am I missing?
19:11 nikitha joined #git
19:13 vassagus joined #git
19:13 nikitha joined #git
19:14 veilig2000 joined #git
19:14 LeBlaaan_ joined #git
19:14 jwest joined #git
19:15 Dyonisos joined #git
19:15 xall joined #git
19:15 burtons joined #git
19:16 nikitha joined #git
19:17 mustmodify joined #git
19:17 mustmodify Can branches include the # symbol? I'm having trouble checking one out...
19:17 _ikke_ mustmodify: yes
19:17 _ikke_ mustmodify: Try escaping it, most likely your shell is seeing it as a comment character
19:17 xall joined #git
19:18 mustmodify d'oh! I had two consoles open and I did `git fetch` on the wrong repo!
19:18 _ikke_ ah
19:19 Darcidride joined #git
19:19 mustmodify That was really confusing.
19:19 nikitha joined #git
19:19 LiohAu joined #git
19:19 Anupkumar joined #git
19:20 tgunr joined #git
19:23 Balliad joined #git
19:23 star_prone joined #git
19:24 mustmodify left #git
19:24 Mannn joined #git
19:31 nwkj86 joined #git
19:33 davidosomething joined #git
19:34 BrianBlaze420 joined #git
19:34 netj joined #git
19:35 skasturi joined #git
19:35 Spec-Chum joined #git
19:35 Pulp joined #git
19:37 madebymarkca joined #git
19:37 medeirosthiiago joined #git
19:38 osse tcurdt: you're not missing anything. it gives the right answer
19:40 osse tcurdt: try this instead, in case the echo in your shell is acting up
19:40 osse { echo -ne "blob `wc -c < git.c`\0"; cat git.c; } | openssl dgst -sha1
19:40 osse oopd
19:41 osse { printf 'blob %s\0' $(wc -c < FILE); cat FILE; } | openssl dgst -sha1
19:41 EvilPeng1 joined #git
19:42 Rodya_ joined #git
19:44 tcurdt osse ha! interesting! https://gist.github.com/tcurdt/d​8edbaa7d10ac5e850f7358e19d4d5f2
19:44 tcurdt the printf one worked OK
19:45 osse echo must die
19:45 vassagus joined #git
19:45 tcurdt :)
19:46 nilg joined #git
19:46 tcurdt I am still curious what's wrong there ... I guess echo is not printing the \0 correctly
19:46 cdg joined #git
19:47 osse tcurdt: what does this print?   echo -ne "\0" | hexdump
19:47 jubalh joined #git
19:47 lebster joined #git
19:47 ELLIOTTCABLE joined #git
19:47 MLM joined #git
19:47 ShekharReddy joined #git
19:47 Anupkumar joined #git
19:47 _ikke_ 00000000: 30                                       0
19:47 _ikke_ a real 0
19:48 tcurdt 0000000 00
19:48 tcurdt odd
19:48 _ikke_ quoted a null byte
19:48 osse tcurdt: that's a null byte
19:49 tcurdt osse yes, which is what it should be (I guess) ... which is why it's odd
19:49 paul424 joined #git
19:50 tcurdt hexdump is great to debug this
19:51 tcurdt ha! found it
19:51 osse nbsp ?
19:52 tcurdt "wc -c < Gemfile" has white space before the actual filesize
19:53 osse ugh
19:54 osse and even then it only works because I didn't quote the $()
19:55 tcurdt I will never understand this whitespace business in some cmd line tools
19:55 osse my wc doesn't do this
19:56 * osse shakes fist at BSD
19:56 _ikke_ Mine neither
19:57 osse but your echo -ne "\0" prints an ascii zero :p
19:57 osse the whole world's gone to hell!
19:57 tcurdt :)
19:57 jkt joined #git
19:57 sybariten i try to do a "git pull" but git warns me that a txt file will be overwritten by the remote file. But i want that file to be overwritten in this case. What do i do?
19:57 sybariten it says i should commit or stash...
19:57 _ikke_ osse: No, that was echo -ne \0
19:57 osse tcurdt:  printf '<%s>\n' "$(wc -c < Gemfile)" plz
19:57 _ikke_ echo -ne "\0" prints the right thing
19:58 osse ah
19:58 tcurdt osse <      56>
19:58 tcurdt that's the problem
19:58 osse uuugh
19:58 _ikke_ wc --version
19:59 tcurdt (echo -ne "blob `wc -c < Gemfile | tr -d '[[:space:]]'`\0"; cat Gemfile) | openssl dgst -sha1 ... does work
19:59 osse tcurdt: intersting that its size is identical to the number I typed at random when testing
19:59 patrick-sartent joined #git
19:59 nitric sybariten: do what git says  :-)  "git stash; git pull" is the quickest and easiest, and you can just forget about the stash if you don't care about the changes you had
20:00 tcurdt _ikke_ my wc doesn't have --version option :)
20:00 osse some echoes will print  "-ne blob 56\0"
20:00 _ikke_ tcurdt: what distro?
20:00 tcurdt _ikke_ macOS 10.11.6 :)
20:01 sybariten nitric: uh well just "git pull" didnt work, appearently. It didnt want to get the remote file.
20:01 _ikke_ tcurdt: that explains :P
20:01 _ikke_ sorry
20:01 tcurdt indeed
20:01 osse IOW: BSD version of various tools
20:01 nitric sybariten: is the text file not under git's control on your working copy? if that's the case, just delete it
20:02 nitric sybariten: git got the remote file just fine, it didn't want to stomp on a file you already had locally
20:02 ELFrederich joined #git
20:02 ELFrederich anyone here write their own LFS plugin to hook up to a different kind of system?
20:03 MattMaker joined #git
20:03 cyan__ joined #git
20:03 SwiftMatt joined #git
20:03 ahmedelgabri joined #git
20:05 scalability-junk left #git
20:05 vuoto joined #git
20:06 vuoto joined #git
20:07 zarubin joined #git
20:08 kbs joined #git
20:08 mischat joined #git
20:09 hahuang65 joined #git
20:13 mikecmpbll joined #git
20:16 phoenix301 joined #git
20:17 star_prone joined #git
20:19 Dougie187 joined #git
20:19 star_prone joined #git
20:20 MattMaker joined #git
20:22 aguimaraes1986 joined #git
20:23 fuchstronaut joined #git
20:23 m0viefreak joined #git
20:24 dsantiago joined #git
20:26 burtons joined #git
20:30 Darcidride joined #git
20:31 Joel joined #git
20:32 p4trix joined #git
20:33 mweshi joined #git
20:33 nwkj86 joined #git
20:33 mikecmpbll joined #git
20:34 fuchstronaut joined #git
20:35 MattMaker joined #git
20:35 kpease joined #git
20:37 kpease joined #git
20:38 MattMaker joined #git
20:40 l4v2 joined #git
20:44 mikecmpbll joined #git
20:46 vuoto joined #git
20:47 aidalgol joined #git
20:50 SwiftMatt joined #git
20:51 ash_workz can i discard changes that are unstaged?
20:52 preaction git checkout can do that for an entire file
20:52 nitric ash_workz: "git checkout -p" will let you discard unstaged hunks, too
20:52 ash_workz thanks!
20:53 de-vri-es joined #git
20:53 mohabaks joined #git
20:53 zonedigital joined #git
20:54 SonikBoom joined #git
20:58 Dougie187 joined #git
21:02 durham_ joined #git
21:04 Marbug_ joined #git
21:06 dsdeiz joined #git
21:06 dsdeiz joined #git
21:10 johnmilton joined #git
21:10 d4rklit3 joined #git
21:25 mikecmpbll joined #git
21:25 occupant so I'm trying to config git-multimail to NOT show the diffs in the emails it sends. the docs seem to suggest that this is the default state and that 'refchangeShowLog' defaults to false, but it has the same behavior whether it's set true or false.
21:25 hahuang65 joined #git
21:25 vassagus joined #git
21:26 durham joined #git
21:27 durham_ joined #git
21:27 case39 joined #git
21:27 zonedigital joined #git
21:31 bjpenn joined #git
21:32 durham joined #git
21:32 bjpenn if i removed a file from the repo, and find out maybe a week later, after a million changes were made to the repo
21:32 bjpenn how do i get that file back?
21:33 HedgeMage bjpenn: depends on what you want the history to look like...you may just want to check out the last commit in which the file existed, copy the file to a safe place, check out HEAD, and commit the file with an appropriate commit message.
21:33 HedgeMage that's the easiest thing to do
21:33 d4rklit3 joined #git
21:33 radsy joined #git
21:41 CheckDavid joined #git
21:42 bjpenn HedgeMage: thanks, was wondering if thers a way to just cherry pick just one file or something
21:42 bjpenn but i guess i can live with the solution you mentioned
21:42 HedgeMage bjpenn: No.  You can cherry pick a single commit, but not only one file from that commit.
21:42 bjpenn HedgeMage: any easy way to locate the last commit in which the file existed?
21:43 bjpenn like if you have no experience with teh code base, lets say its the 2nd day of your new job
21:43 bjpenn and you had to restore this file
21:43 bjpenn out of a huge project with a million commits
21:43 HedgeMage hrm...I know how to do it with reposurgeon, but I'd like to think there's a command in git for that
21:43 wxg joined #git
21:43 bjpenn same here
21:46 ki0 joined #git
21:47 nitric bjpenn: c'mon man, google: http://stackoverflow.com/a/1113140/2926055
21:47 ahr3n joined #git
21:48 bjpenn nitric: thanks!!!
21:49 dreamyspell joined #git
21:49 tcurdt so everyone says "don't store binary files in git" but I am wondering about some of the details
21:51 tcurdt let's say I have a huge file - which will never change and that everyone (that has a copy of the repo) will need
21:51 tcurdt what are the drawbacks?
21:52 danslo_ joined #git
21:52 kadoban tcurdt: None, in that specific case.
21:52 tcurdt will a "git status" always have to calculate its sha1 for example?
21:52 kadoban I like to say ... don't store binary files that are large and will change often (or if you need to merge them). Other combinations can be okay.
21:52 kadoban tcurdt: No, it's smarter than that.
21:53 tcurdt I was hoping that :)
21:54 Dougie187 left #git
21:54 kadoban So the main problems are that git can't really compress binary file diffs much at all, so each time you change one it's pretty much like storing another complete copy in history. Also, perhaps more importantly, git has no idea at all how to merge two binary files, so you have to do that yourself ... which usually is kind of impossible to do manually, though sometimes not.
21:55 Impaloo joined #git
21:55 darkbit joined #git
21:56 tcurdt kadoban yes, that's clear
21:57 Faylite joined #git
21:59 kpease joined #git
22:03 tcurdt still wondering if using "git annex" would be a better choice anyway ... never used annex before though
22:04 sweatsuit joined #git
22:06 blackwind_123 joined #git
22:06 boombatower joined #git
22:10 Mannn joined #git
22:10 hahuang61 joined #git
22:10 ryez joined #git
22:11 beyertech joined #git
22:13 dsantiago joined #git
22:13 axl_ joined #git
22:15 sagerdearia joined #git
22:15 unreal joined #git
22:17 Faylite joined #git
22:18 sweatsuit joined #git
22:19 xlegoman joined #git
22:20 sweatsuit joined #git
22:21 Faylite joined #git
22:22 sweatsuit joined #git
22:26 ribasushi joined #git
22:26 mrpackethead joined #git
22:26 mrpackethead on github, is is possible to invite an external person to a private repo, but restrict what they can do.
22:27 HedgeMage mrpackethead: Nope...they have commit access or they don't.
22:27 mrpackethead I want to control pull request approvals
22:27 beyertech joined #git
22:27 HedgeMage mrpackethead: That's why I have my primary hosting on my own server, and use github as a mirror only.
22:27 nitric_ joined #git
22:28 interne7y joined #git
22:28 mozzarella joined #git
22:29 msafi joined #git
22:30 msafi Hey folks, I have the following graph in my git repo, I want to remove all but the last commit
22:30 spudowiar joined #git
22:30 msafi https://i.imgur.com/QiPzeBe.png
22:30 msafi Anyone know how to do that?
22:33 osse msafi: git rebase --root. Mark the two latter commits with 's'
22:33 osse msafi: or rm -rf .git; git init; git add .; git commit -m 'first commit'
22:33 osse msafi: or git reset HEAD~2; git commit --amend
22:37 amikiri joined #git
22:37 skapunker joined #git
22:37 nitric joined #git
22:38 msafi osse: I kinda tried all solutions, and this is what I currently have https://i.imgur.com/SmdvW6N.png
22:38 msafi I guess it's close enough :D
22:38 osse my answer to this problem is exactly the same as earlier
22:38 osse (except with HEAD~1 this time)
22:40 msafi osse: https://i.imgur.com/YaXeJQ9.png
22:40 msafi Oh sorry
22:41 msafi osse: the problem is the owner of the first commit is not me
22:41 osse msafi: Git doesn't care
22:41 msafi How can I get it to show me as the owner then?
22:42 osse msafi: with my last solution, add --reset-author to git commit
22:42 fmcgeough joined #git
22:42 msafi osse: I've already pushed. Now I have one commit only
22:43 osse Use the --force
22:43 osse !rewrite
22:43 gitinfo 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
22:44 msafi osse: git commit --amend --reset-author?
22:44 Darren_ joined #git
22:44 msafi osse: git commit --amend --reset-author --force?
22:44 b1tchcakes joined #git
22:45 osse yes, and no respectively
22:45 osse I meant git push --force
22:46 jjasinski joined #git
22:47 msafi osse: that did it. Thank you!
22:47 NeXTSUN joined #git
22:48 Faylite_ joined #git
22:48 HappyLoa- joined #git
22:50 crayon joined #git
22:53 ikmaak joined #git
22:53 kbs joined #git
22:55 medeirosthiiago joined #git
23:01 durham_ joined #git
23:01 shakeel joined #git
23:02 shakeel hi
23:02 gitinfo shakeel: 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.
23:02 amikiri joined #git
23:06 _28_ria joined #git
23:07 rubyonrailed joined #git
23:09 Guest9 joined #git
23:09 RedneckDonut joined #git
23:11 hahuang61 joined #git
23:14 fahadash joined #git
23:15 hashpuppy joined #git
23:15 dsantiago joined #git
23:21 mablae__ joined #git
23:22 vuoto joined #git
23:23 flaguy48 left #git
23:23 mweshi joined #git
23:24 BSAlb joined #git
23:24 BSAlb joined #git
23:30 mweshi joined #git
23:30 e89c8c95a050 joined #git
23:31 mweshi joined #git
23:32 flaguy48 joined #git
23:32 Peuc joined #git
23:36 dsdeiz joined #git
23:36 dsdeiz joined #git
23:36 ansimation joined #git
23:37 johnnyfive joined #git
23:37 HanClinto joined #git
23:37 kadoban joined #git
23:38 xlegoman joined #git
23:39 adambrenecki joined #git
23:40 abyx joined #git
23:41 abyx aghh, is there a way to use multiple git accounts on the same machine without much trouble?
23:42 solrize git accounts??
23:42 solrize you mean repos?
23:42 adambrenecki abyx: what do you mean by accoints
23:42 adambrenecki Gah, accounts
23:43 abyx i have work and personal github user accounts, i want to push using work credentials during the day, and personal credentials when not at work
23:43 abyx (and commit under different names as well)
23:43 solrize oh you mean github accounts
23:43 adambrenecki For accounts on eg github/gitlab/etc: just put the remote URL as "HTTPS://username@github.com/..."
23:43 damjan abyx: [user]
23:43 damjan email/name can be setup globaly or in a repo
23:44 adambrenecki If you're doing it on the same repo, you'd need to set two remotes
23:44 damjan ssh keys are used automatically
23:44 rubyonrailed joined #git
23:44 adambrenecki Not over HTTPS, that uses basic auth
23:45 ShekharReddy I want to make a branch into a separate repo on github, anyone suggest me how to get to this point
23:46 adambrenecki abyx: What's the aim here - to separate which commits were made at home vs work?
23:47 abyx adambrenecki, yep
23:47 ShekharReddy the current repo is on github i forked it and changed as per my requirements on a separate branch and now i want to publish the changes as a fresh repo on github. Is this possible(not manually)?
23:50 mweshi joined #git
23:50 adambrenecki Yep, so it probably matters more what name/email is attached to the commits rather than which account you use to push them
23:51 adambrenecki Best solution is make a clone that you use at home and set the user.name and user.email settings locally like damjan suggested
23:52 adambrenecki Because that's actually stored on the commit, which account you use to push isn't
23:54 EvilPeng1 joined #git
23:55 zacts joined #git
23:57 cqi joined #git
23:57 cqi_ joined #git
23:58 mweshi joined #git
23:59 sebascar joined #git

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