Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2017-01-23

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

All times shown according to UTC.

Time Nick Message
00:00 _28_ria joined #git
00:00 cfoch joined #git
00:00 cfoch Hi
00:00 Gsham joined #git
00:00 ij bremner, Tooooo bad. :)
00:01 cfoch In my current commit I have a function not being called. How can I know through git what was the last commit this function was called? :)
00:01 mizu_no_oto joined #git
00:02 kadoban joined #git
00:03 cyphase joined #git
00:08 pity joined #git
00:09 bmahe joined #git
00:10 arescorpio joined #git
00:10 nowhere_man joined #git
00:14 dirtyroshi joined #git
00:15 voldz1tic3n joined #git
00:15 xlegoman joined #git
00:18 _28_ria joined #git
00:19 ij left #git
00:19 diogenese joined #git
00:22 thiago cfoch: git log -Sfunctionname -p
00:22 justan0theruser joined #git
00:22 Oatmeal joined #git
00:26 timvisher joined #git
00:30 snowkidind joined #git
00:33 str joined #git
00:34 fuzzybear3965 joined #git
00:34 a_thakur joined #git
00:38 Rodya_ joined #git
00:43 Rodya_ joined #git
00:45 Rodya_ joined #git
00:46 JamesBaxter joined #git
00:47 a_thakur_ joined #git
00:48 Gloomy joined #git
00:49 chachasmooth joined #git
00:50 Rodya_ joined #git
00:52 mehola joined #git
00:53 Cthalupa joined #git
00:53 ome joined #git
01:00 zzz joined #git
01:01 xissburg joined #git
01:06 chachasmooth joined #git
01:07 qpdb joined #git
01:09 XenophonF joined #git
01:11 beuker joined #git
01:11 BSaboia joined #git
01:12 Nilesh_ joined #git
01:16 a_thakur joined #git
01:17 kaldoran joined #git
01:17 Puffball joined #git
01:18 sim590 joined #git
01:19 z1haze joined #git
01:19 z1haze i have a release on github of 1.0.0 but i have pushed changes, how do i update my 1.0.0 release tag on github to use the latest version of my code?
01:19 orzo was it a tag or a branch?
01:20 beuker joined #git
01:20 SuDo__ joined #git
01:21 timvisher joined #git
01:22 SuDo__ joined #git
01:22 orzo https://nathanhoad.net/how-​to-delete-a-remote-git-tag
01:22 sunny256 z1haze: You can delete the tag remotely with something like "git push github :v1.0.0" and then push the tag again. It probably works with "git push github v1.0.0 --force" too.
01:23 rubyonrailed joined #git
01:23 rivarun joined #git
01:24 bremner but !rewrite applies
01:24 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
01:25 UnnamedPlayer joined #git
01:25 z1haze bremner, thanks for the input, this is a brand new package of mine, actually my first and I have a couple mistakes in there
01:26 dsdeiz joined #git
01:26 dsdeiz joined #git
01:27 beuker joined #git
01:31 dsdeiz joined #git
01:31 dsdeiz joined #git
01:33 cybex_ joined #git
01:34 z1haze im pretty new to git myself.. so once i create a release tag for lets say, 1.0.0 and then i make changes and push back up to the same branch, do i need to make another release like 1.0.1 ?
01:34 pity joined #git
01:35 cybex_ hi all, git noob here. currently here: https://git-scm.com/book/en/v2/Git-​on-the-Server-Setting-Up-the-Server. What is the purpose of "origin". Can origin be any name?
01:35 cybex_ context: git remote add origin git@gitserver:/srv/git/project.git
01:37 fahadash joined #git
01:40 Redfoxmoon where does the git build system change man files with :: in the filenames to . for cygwin and win32?
01:44 zacts joined #git
01:51 Ryanar joined #git
01:51 mehola joined #git
01:54 phantombeta joined #git
01:54 phantombeta joined #git
01:55 phantombeta left #git
01:58 ash_workz joined #git
02:01 MissionCritical joined #git
02:06 Gsham joined #git
02:08 pur3eval joined #git
02:09 mda1 joined #git
02:09 radialneon joined #git
02:10 atrigent__ joined #git
02:11 chachasmooth joined #git
02:11 timvisher joined #git
02:12 Derperperd joined #git
02:15 _28_ria joined #git
02:17 LordRyan joined #git
02:20 chachasmooth joined #git
02:22 andrew710 joined #git
02:22 safe joined #git
02:23 cybex_ 1170 people and no one helping...weird
02:24 bremner likely no-one knows, especially since it's a repeat question
02:24 pity joined #git
02:24 bremner oh, your question. well, the short answer is yes, origin can be any name, but it's the name chose by default by git clone
02:25 cybex_ bremner: sorry I was actually refering to all questions. xD thanks
02:26 osse joined #git
02:27 hecliunyx joined #git
02:28 finalbeta joined #git
02:29 dirtyroshi joined #git
02:29 Sircle joined #git
02:29 Sircle Hi
02:29 gitinfo Sircle: 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.
02:30 Sircle Is there a post recieve hook in raw git?
02:30 dodobrain joined #git
02:30 kadoban Yes
02:30 d^sh joined #git
02:30 aypea[3] hmm. is there a way to tell git to merge the differences between two points on branch A into branch B such that it's kept track of?
02:31 pks joined #git
02:32 Sircle Can I get the commit message(s) in a post recieve hook? I plan to auto deploy all the nodejs code at a port mentioned in git commit message. e.g: "Fixed layout. Deploy port :3454". I can use regex to get the port. Possible?
02:32 kadoban cherry-pick or rebase that section of the branch onto B, and then either merge it in with --no-ff, or squash it, or merge it --ff-only, whichevery you like
02:32 Sircle kadoban: ok ^
02:32 kadoban That was for aypea[3], the last bits.
02:33 aypea[3] kadoban: thanks. I'll look into that. cherry-picking was what I was thinking of already but I like the idea that it shows up nicely in git rather than as a comment or something.
02:33 kadoban Yeah, preserving history like that can make a lot of sense sometimes
02:33 aypea[3] indeed :)
02:38 Sircle Can I get the commit message(s) in a post recieve hook? I plan to auto deploy and run all the nodejs code at a port mentioned in git commit message. e.g: "Fixed layout. Deploy port :3454". I can use regex to get the port. Possible?
02:38 Guest95139 joined #git
02:38 kadoban Sounds like bad idea even if possible. It's likely possible though.
02:39 Sircle what software is bitbucket using under the hood? Is it open source and I can download it/ use it? Is it like gitlabs?
02:39 kadoban I don't believe much of their software is open source, especially the interesting server bits.
02:40 belak Bitbucket isn't open source, though we use a number of open source libraries under the hood
02:41 Sircle belak:  I have seen a customer implementation of jira + bitbucket looking like git tool + bamboo and all was locally hosted and custom auto deploy configured. I wonder if these were opensource software and anyone can configure/download them?
02:42 belak Bitbucket Server isn't open source, but you can host it yourself
02:46 belak It's not free though
02:46 Sircle hm..
02:46 Sircle niether is bamboo or jira?
02:46 Sircle https://developer.atlassian.com/bitbuck​et/server/docs/latest/how-tos/building-​bitbucket-server-from-source-code.html?​_ga=1.159315388.741441772.1485138814   code is available to commercial licencees
02:46 belak Bamboo and Jira are the same.
02:46 belak Yes, but that's if you have a license
02:46 belak It's not free
02:46 Sircle k
02:46 Rodya_ joined #git
02:48 ilbot3 joined #git
02:48 Topic for #git is now Welcome to #git, the place for git help and a fresh start on a new branch | 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 | This channel may contain users known to the state of California to cause cancer.
02:50 kadoban Sircle: Looks like post-receive gets as input all of the names of refs that are updated, along with their old and new values.
02:50 kadoban Sircle: So you can use that information. At a basic level, something like 'git log ^old new' might do, though ... you're really going to have to engineer the fuck out of this to make it anything like sane or reliable I suspect.
02:50 chachasmooth joined #git
02:51 kadoban I'd be extremely wary of using this for ... anything. Allowing commit messages themselves to execute code on your server sounds insane just on the face of it.
02:52 WhiskerBiscuit joined #git
02:53 Sircle kadoban:  am. can you elaborate? Do you mean I will not be getting a variable that references to commit messages "For that push" only? (In that hook) and I have to search the logs instead? Secondly, do you propose anything else for the hook to know on which port to run the code pushed in the new push every time?
02:54 kadoban Did you look at the manpage for this? You'll have to become quite familiar with it to even attempt this.
02:54 kadoban man githooks
02:54 gitinfo the githooks manpage is available at http://jk.gs/githooks.html
02:54 kadoban And yes, as far as I know it's not going to give you list of commit messages.
02:54 kadoban You'll have to do a bit of work to generate those.
02:54 Vampire0_ joined #git
02:55 kadoban I propose ... just not doing that. Why would the commit messages have to tell this thing what to do?
03:00 Sircle kadoban: I see there is no other way to run a different instance of the app on each push (on a different port that matches the ticket number of project management tool used by dev for that task)...
03:01 Sircle kadoban:  automatic deployment.. I have seen that with jira +bamboo + bitbucket
03:02 Sircle the only difference was that babmoo deployed it on ticket number as port with no need of commit message having the port number. However, in my case I do not have bamboo
03:02 Sircle kadoban:  any bright ideas?
03:03 kadoban Not particularly
03:04 Sircle kadoban:  so I have this as the only solution and would it be correct if I say that on every push , I have to scan all commit messages and then a) get commits messages in this push b) see if any of it had the port number c) deploy the code on that port
03:04 Sircle ?
03:05 aavrug joined #git
03:05 Sircle scan all old and new messages ^
03:06 aavrug joined #git
03:06 gugah joined #git
03:09 sunri5e joined #git
03:10 xrfang joined #git
03:10 Sircle kadoban:  question: Can I get the files changed in that push?
03:11 hexagoxel joined #git
03:11 cyphase joined #git
03:12 timvisher joined #git
03:12 dreiss joined #git
03:13 madewokherd` joined #git
03:13 Derperperd joined #git
03:14 kadoban If you try, and if you make concrete the definition of "changed", sure.
03:15 Sircle kadoban:  ok, where can I read that what variables and methods are availabile in hooks so I can get commit messages and files changes details?
03:16 arnas joined #git
03:16 Sircle kadoban:  I would go with gitlab
03:16 Sircle kadoban:  but I assume that git commands and hooks won't change in gitlab or whatever tool I use
03:16 kadoban Variables and methods? I think you misunderstand hooks. They're simply executable programs that you write, usually written in a "scripting" language such as bash.
03:16 kadoban You get certain inputs, which the manpage lists.
03:18 Vortex34 joined #git
03:18 Sircle yes but there are certain commands of git used inside them. e.g make a tar of code or past it somewhere
03:18 Sircle kadoban:  ok, which man page? man git?
03:18 gitinfo kadoban: the git manpage is available at http://jk.gs/git.html
03:18 kadoban Then you're asking what commands are available in git? That's a big list.
03:19 Sircle kadoban:  hm
03:19 kadoban The hooks manpage was linked above about 25 minutes ago.
03:20 aavrug joined #git
03:22 Sircle k
03:22 Sircle thanks
03:30 fstd_ joined #git
03:30 Derperperd joined #git
03:30 xrfang hi there, I try to clone a specific commit, and followed instruction here: http://stackoverflow.com/questions/​3489173/how-to-clone-git-repository​-with-specific-revision-changeset
03:31 xrfang while doing the last step, git reported: fatal: ambiguous argument 'FETCH_HEAD': unknown revision or path not in the working tree.
03:31 xrfang why?
03:31 suttonwilliamd joined #git
03:32 ahr3n joined #git
03:33 matsaman xrfang: stackoverflow is kind of a joke, but
03:33 matsaman xrfang: if you're going to use it, look for the top-voted answer, not the one with the check
03:35 xrfang matsaman: I used the checked one with HEAD, it worked, but with a specific commit, it don't. actually what I tried to do is this: http://stackoverflow.com/questions/31278902/how-​to-shallow-clone-a-specific-commit-with-depth-1
03:35 xrfang but according to that post, shallow clone not possible with specific commit...
03:37 chachasmooth joined #git
03:38 matsaman why would I care about what you did that is in direct contradiction to what I just said to you? =P
03:38 pur3eval joined #git
03:39 SuDo__ joined #git
03:39 xrfang matsaman: it seems that you hate stackoverflow... I am sorry. I just want to pull a specific commit, and only that commit from a remote repo.
03:40 matsaman well for what it's worth, I'd rather be stuck with stackoverflow than someone who won't listen =)
03:40 T-J-M Hi there, is there someone familer with git-annex? If so, please help me with this problem. Thank you. https://unix.stackexchange.com/questions/339274/
03:44 dreiss joined #git
03:45 a_thakur joined #git
03:49 a_thakur joined #git
03:54 pity joined #git
03:59 timvisher joined #git
03:59 mizu_no_oto joined #git
04:00 prawnzy joined #git
04:03 dreiss joined #git
04:04 pity joined #git
04:09 hexagoxel joined #git
04:13 pity joined #git
04:16 lvh joined #git
04:17 EliasH joined #git
04:18 xall joined #git
04:18 chachasmooth joined #git
04:25 dweelie joined #git
04:27 mohabaks joined #git
04:29 rubyonrailed joined #git
04:30 msonntag joined #git
04:31 sbulage joined #git
04:32 Limes_ joined #git
04:33 apollo11w joined #git
04:34 robotroll joined #git
04:38 pulsefussion77 joined #git
04:38 miczac joined #git
04:39 pur3eval joined #git
04:40 xall joined #git
04:40 Oatmeal joined #git
04:42 pulsefussion77 left #git
04:44 z3t0 joined #git
04:44 hecliunyx joined #git
04:47 whomp joined #git
04:47 Lunatrius` joined #git
04:48 whomp hi, when i apply a patch from a git diff, when there's a merge conflict, it simply tells me that the patch does not apply. but in this case i want it to behave like a commit, where it still does the merge but shows merge conflicts
04:48 whomp any ideas how to do this?
04:50 s2013 joined #git
04:52 xpk joined #git
04:52 xpk joined #git
04:52 kadoban I strongly suspect that that doesn't make sense. A patch isn't a commit, unless it actually applies. Git really only merges commits.
04:53 kadoban A patch that doesn't apply is ... instructions that can't be understood for how to create the content you want.
04:54 timvisher joined #git
04:55 thiago whomp: git apply cannot do that
04:56 thiago instead of git diff, if you use git format-patch, you can pass the -3 argument to git am
04:58 arnas joined #git
05:00 dermoth joined #git
05:04 rubyonrailed joined #git
05:05 whomp what a round-about way to do this
05:08 ayogi joined #git
05:08 theoceaniscool joined #git
05:08 hecliunyx joined #git
05:08 whomp thiago, thanks! i see that there's actually a -3 flag for git apply as well so that solved my problem
05:08 a3Dman joined #git
05:12 z3t0 joined #git
05:17 apollo11w left #git
05:21 rajat joined #git
05:32 hecliunyx joined #git
05:33 ckruczek joined #git
05:33 atrigent joined #git
05:33 xall joined #git
05:35 Sazhen86 joined #git
05:37 timvisher joined #git
05:38 TomTomTosch joined #git
05:40 pur3eval joined #git
05:42 dodobrain joined #git
05:46 ngm joined #git
05:46 ngm can i move all the commits from master to dev ?
05:47 _ikke_ ngm: what do you want to achieve?
05:49 ngm i have a remote repo in which master has single different commit and all my current commits in my master are in remote dev and i have some additional commits too. So iam thinking of doing this let me know the changes to make it work, 1) moving all the commits to dev and checkout to master and set upstream and pull changes 2) checkout to dev and pul
05:49 ngm l all the changes again with --rebase 3) push the new changes
05:51 ngm 3)*  push the new changes  of the dev branch
05:51 ngm _ikke_: do you think my approach is correct?
05:52 ngm local repo: master -> all the remote dev commits + new commits ;  dev-> not yet created
05:53 ngm remote repo: master -> one commits not yet in local ; dev-> all the commits in local master branch without new commits
05:53 _ikke_ you can just rename master to dev if that's you want?
05:54 cyan__ joined #git
05:54 ngm Like rename the master to dev and then create new master and pull all the changes to be in sync?
05:55 ngm and setup the upstream to their respective remote branch, cool will give it a shot
05:56 mohabaks joined #git
05:56 spacelord_ joined #git
05:57 s2013 joined #git
06:05 dec0n joined #git
06:06 rubyonrailed joined #git
06:08 ngm iam going to try the above operations so is there any way to revert them one by one like pop kind of operations?
06:10 _ikke_ ngm: man git reflog
06:10 gitinfo ngm: the git-reflog manpage is available at http://jk.gs/git-reflog.html
06:10 ngm ty
06:10 ngm This is the strategy to get it working.
06:12 ngm 1) create a dev branch and checkout and delete the master, 2) create new orphan master pull the changes and create a dev from the master and checkout. 3) pull the remote dev changes 4) merge the local dev with orphaned dev branch
06:13 ngm does this work
06:13 ngm _ikke_ simply naming the branches wont work as i have to make the changed to HEAD ref too
06:15 brent__ joined #git
06:15 lindenle joined #git
06:16 xiwang_ joined #git
06:18 adymitruk memfrob: I see. would like to get involved in making senes of the changes so you can have a more dependable way of working with git
06:20 s2013 joined #git
06:27 brent__ joined #git
06:28 a3Dman joined #git
06:29 danimal2 joined #git
06:29 sbulage joined #git
06:30 Raging_Hog joined #git
06:34 govg joined #git
06:34 brent__ joined #git
06:34 lyxus joined #git
06:37 jozwior joined #git
06:37 overlord_tm joined #git
06:38 xall joined #git
06:38 timvisher joined #git
06:46 dersand joined #git
06:49 JanC_ joined #git
06:49 zeroed joined #git
06:49 aw1 joined #git
06:51 Junior joined #git
06:51 govg joined #git
06:51 tirej joined #git
06:51 tirej joined #git
06:51 s2013 joined #git
06:55 ThomasLocke joined #git
06:56 jaafar joined #git
06:56 serialoverflow joined #git
06:57 brent__ joined #git
06:59 brent__ joined #git
06:59 zeroed joined #git
07:00 Raging_Hog joined #git
07:01 elementalest joined #git
07:02 ctcrnitv joined #git
07:02 ctcrnitv I pushed to gh-pages about 10 min ago, all I get is a blank page.
07:02 ctcrnitv tried googling for an answer but most issues relate to jekyll
07:03 ctcrnitv but I pushed it just from a local branch
07:04 freimatz joined #git
07:04 JeroenT joined #git
07:05 ctcrnitv please help this assignment is due in 30 min :(
07:05 dminuoso We don't do homework.
07:05 ctcrnitv that's not my question, I did my work, I just need to deploy it
07:05 dminuoso But if you want I can give you some time management tips.
07:05 ctcrnitv get the fuck out of here asshole
07:06 dminuoso If you demand quick support, perhaps you're better off with a consultant.
07:06 jagob joined #git
07:06 Mo joined #git
07:06 menip joined #git
07:07 rubyonrailed joined #git
07:07 ctcrnitv whatever.
07:08 ctcrnitv anyway I also made sure my master branch and gh-pages branch were synched
07:08 ctcrnitv in case someone who's actually helpful wants to chime in
07:08 Silenced joined #git
07:10 _ikke_ ctcrnitv: responding like thise doesn't make people want to help you
07:12 _ikke_ ctcrnitv: But to be constructive, try pushing again (by ammending a commit and then force push)
07:14 Silenced joined #git
07:19 ihradek joined #git
07:20 bket joined #git
07:20 roelmonnens joined #git
07:23 roelmonnens joined #git
07:23 bl0w3d_0ut joined #git
07:26 pur3eval joined #git
07:26 zeroed joined #git
07:27 timvisher joined #git
07:29 rizonz does git only store passwords when using https ?
07:29 rizonz with pull ?
07:29 theoceaniscool joined #git
07:30 fstd_ joined #git
07:31 JeroenT joined #git
07:33 roelmonn_ joined #git
07:33 cdown joined #git
07:35 pR0Ps joined #git
07:40 cagedwisdom joined #git
07:40 govg joined #git
07:41 miczac joined #git
07:41 Sceorem joined #git
07:41 MarkusDBX joined #git
07:42 TomyLobo joined #git
07:42 MarkusDBX When introducing newbies to git, are they in any way helped by gui tools? I find it seems best to just learn the cli, what is your view?
07:42 mehola joined #git
07:43 dminuoso MarkusDBX: Just learn the CLI.
07:43 dminuoso MarkusDBX: Most visual aids are provided by any of the decent...
07:43 dminuoso !books
07:43 gitinfo [!book] 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
07:43 mehola joined #git
07:44 eduardas_m joined #git
07:44 dminuoso MarkusDBX: Once you understand how git works, it won't matter if you use visual or command line tools.]
07:44 MarkusDBX dminuoso: I know all that. mainly I'm just double checking and gathering arguments
07:44 osse MarkusDBX: I don't have much GUI experience, but a lot of the ones I have seen misrepresent Git in unfortunate ways, e.g. when you right click on a file one of the options is "push"
07:44 Derperperd joined #git
07:44 dminuoso osse: Though in all fairness, for SVN TortoiseSVN is awesome.
07:44 dminuoso Sadly git is not SVN.
07:44 mehola joined #git
07:44 MarkusDBX dminuoso: Yeah I know, IMO, what is hard to learn about git is that YOU yourself, setup the workflow. The workflow is the actually work with git imo
07:44 osse dminuoso: heh, that could be. and yes that was a stab at TortoiseGit
07:45 mehola joined #git
07:46 mehola joined #git
07:46 mehola joined #git
07:47 chele joined #git
07:47 MarkusDBX dminuoso: I sometimes find it a challenge to get "wordpress" point/click plugin "developers" on git, if they got zero cli knowledge, but I guess you must just force them?
07:47 mehola joined #git
07:48 mehola joined #git
07:48 MarkusDBX even in 2017?! there is those ftp ninajs still around unfortunately
07:49 mehola joined #git
07:51 whomp joined #git
07:52 Mikerhinos joined #git
07:54 zefferno joined #git
07:56 fstd_ joined #git
07:58 elect__ joined #git
08:01 jozwior joined #git
08:01 pR0Ps joined #git
08:03 xall_ joined #git
08:05 Mo Hi, what do you think about Keyword expansion. hg people say that is a feature to avoid. But so far I had keyword expansion on almost every VCS from cvs, svn to p4 etc. and I think it is mandatory.
08:08 Hudu joined #git
08:08 jnavila joined #git
08:10 unreal joined #git
08:11 Hudu joined #git
08:11 pR0Ps joined #git
08:12 jast Mo: what kinds of keywords were you thinking of?
08:12 jast you can do trickery for keyword expansion, I think we have a resource for that...
08:12 jast !keywords
08:12 Mo date, desc, author etc.
08:13 jast hmm
08:13 jast git has no built-in support for keyword expansion, though
08:15 jast faq keywords
08:15 gitinfo It&#39;s not a good idea but it&#39;s possible; more details available at https://git.wiki.kernel.org​/index.php/GitFaq#keywords
08:15 pur3eval joined #git
08:15 Mo So keyword expansion isn't part of git either? I know scripting is possible for that, and hg also has an extension, but hg people don't recommend that as keywording would not be the best approach:  https://www.mercurial-scm.org/wiki/FeaturesOfL​astResort#But_I_need_keyword_and_EOL_handling.21
08:15 jast I guess I should HTML-decode the output...
08:15 rominronin joined #git
08:15 osse Mo: there exists clean/smudge filters. and when you create an archive git-archive can do those kinds of replacements
08:15 Tobbi joined #git
08:15 jast in clean/smudge you don't really have reliable access to commit metadata, though
08:15 Mo jast: Ok, same opinion for git people then.. But I don't get why keywording isn't useful.
08:16 jast well, it's difficult to do correctly, and not difficult to do in your own scripts for release handling etc.
08:17 Mo Difficult? Lot of VCS have done that before and extensions exist. At time of submission, the file revisions and meta data are expanded.
08:17 acetakwas joined #git
08:18 jast if you change all files with keywords at each commit, you will always get merge conflicts on all of those files, though
08:18 jast that alone is enough for me to not be interested in keyword expansion
08:18 Mo It's just very useful/important from user point of view. We have lots of files as copies, and you need to know where it was copied from, which revision etc.
08:19 jast that information isn't available in Git, anyway
08:19 jast you can only make guesses
08:19 osse oh THAT kind of expansion
08:19 jast FWIW it's not available in CVS, either
08:19 Mo Merging is part of such a keyword expansion feature.
08:19 jast and in SVN it really depends on *how* you copy
08:20 jast though I'd argue that using 'svn copy' on individual files is a bad idea in SVN. SVN's concept of not distinguishing between copying and branching/tagging never made sense to me, just makes everything terribly complicated.
08:21 Zoxc joined #git
08:21 Mo jast: I was wrong, not CVS, but RCS was the initiator of these keywords.
08:21 dimi1947 joined #git
08:21 jast CVS does have keyword support, but it doesn't know which files got copied where, etc.
08:21 Mo Then CVS was replacing RCS, so I guess CVS also had keywording.
08:21 jast after all, CVS tracks versioning individually for each file
08:22 jast it can't even follow history across renames
08:22 Mo Just tagging... I see some scripts tagging files via make. But I also have lot of shell scripts, XML, configurations.. I use a keyword-header in all those files and it always worked, at leat with hg KeywordExpansion, p4 and svn.
08:23 lordjancso joined #git
08:23 Mo Anyway, I just wanted to see the git point of view, and appearently it matches the hg point of view about keyword expansion.
08:24 jast apparently, yeah :)
08:24 a_thakur joined #git
08:24 MrMojito joined #git
08:25 a_thakur joined #git
08:25 MafiaInc joined #git
08:30 pR0Ps joined #git
08:30 esc joined #git
08:30 esc joined #git
08:30 nowhere_man joined #git
08:31 texinwien joined #git
08:32 xall joined #git
08:33 giraffe joined #git
08:34 Tobbi joined #git
08:34 Ardethian\work joined #git
08:34 Ardethian\work joined #git
08:35 Derperperd joined #git
08:38 reverse_light joined #git
08:44 jknetl joined #git
08:44 pity joined #git
08:47 durham joined #git
08:47 tirej_ joined #git
08:48 msonntag joined #git
08:50 nostrora joined #git
08:51 Macaveli joined #git
08:53 tirej joined #git
08:53 govg joined #git
08:53 jorj joined #git
08:54 Varazir joined #git
08:55 Balliad joined #git
08:56 pR0Ps joined #git
08:57 Varazir Hmm, I forked a repo on github, cloned it down to my local system. Updated one of the files. When doing 'git push' when I'm in the root folder of the local repo it say everything is up to date. I check the website and it's still the old file.
08:58 osse Varazir: did you commit?
08:58 Varazir yes
08:58 jnavila joined #git
08:58 Varazir osse: but getting Your branch is up-to-date with 'origin/master'.
08:58 Varazir Changes not staged for commit:
08:58 Varazir modified:   irssinotifier.pl
08:59 osse read what git status says
08:59 Varazir so it feels it's something I'm missing there
08:59 durham joined #git
09:00 a_thakur joined #git
09:00 Varazir osse: thanks
09:01 reverse_light joined #git
09:01 durham_ joined #git
09:01 reverse_light joined #git
09:02 pity joined #git
09:02 mikecmpbll joined #git
09:02 rumble joined #git
09:04 msonntag joined #git
09:05 Macaveli joined #git
09:06 codemann joined #git
09:07 Arguggi joined #git
09:09 nowhere_man joined #git
09:09 ssitani joined #git
09:10 marcogmonteiro joined #git
09:12 aard_ joined #git
09:15 m4sk1n joined #git
09:17 djb-irc joined #git
09:17 JeroenT joined #git
09:18 JeroenT_ joined #git
09:21 a_thakur joined #git
09:23 Semtex-h joined #git
09:26 pks Is there any mechanism to have wildcards in the http.<url>.* settings? My use case is to match all subdomains of a specific domain, so something like "http://*.example.com".
09:27 Spherical joined #git
09:27 xall joined #git
09:27 Serpent7776 joined #git
09:29 Varazir Hmm, trying to add a new ssh key to my repo using shell commands
09:29 Varazir https://help.github.com/articles/adding-a-new-​ssh-key-to-your-github-account/#platform-linux only shows from gui
09:29 xall_ joined #git
09:29 moritz Varazir: try #github, maybe?
09:29 grawity you're trying to add it to a github repo, specifically?
09:29 Varazir moritz: ahh sorry
09:30 Varazir grawity: yes
09:30 grawity I don't think they have an 'official' CLI
09:30 moritz https://github.com/ingydotnet/git-hub supports adding SSH keys
09:30 Varazir this is for the program not the site my misstake
09:30 grawity moritz: it does now?
09:30 grawity wait, that's not the same as `hub`
09:31 moritz grawity: never said it was :-)
09:31 moritz just something a friend of mine has been contributing to, which is why I know it exists :-)
09:32 Varazir moritz: thanks I'll take a look at it
09:34 cdown joined #git
09:35 Derperperd joined #git
09:36 rivarun joined #git
09:37 sbulage joined #git
09:37 Arguggi joined #git
09:38 troulouliou_div2 joined #git
09:38 freimatz Hi, how can i list tags on remote?
09:38 grawity git ls-remote [name] --tags
09:39 freimatz thx
09:40 timini joined #git
09:43 Arguggi joined #git
09:44 MineCoins joined #git
09:45 dodobrain joined #git
09:45 netj joined #git
09:47 solf1re joined #git
09:48 Arguggi joined #git
09:49 Vgr_ joined #git
09:49 Vgr_ joined #git
09:49 jozwior joined #git
09:49 freimatz after this i found the right words to google ;-)
09:49 Vgr_ joined #git
09:49 Vgr_ joined #git
09:49 Vgr_ joined #git
09:49 Vgr_ joined #git
09:50 dodobrain joined #git
09:50 mehola joined #git
09:50 donught joined #git
09:52 govg joined #git
09:53 ispn joined #git
09:55 rnsanchez joined #git
09:57 UnnamedPlayer joined #git
09:58 theoceaniscool joined #git
10:00 pR0Ps joined #git
10:03 Arguggi joined #git
10:03 pur3eval joined #git
10:03 xall_ joined #git
10:05 Varazir moritz: got a 404 error when trying to add a key
10:05 ac3takwas joined #git
10:05 moritz Varazir: please open an issue on github
10:05 Varazir moritz: will do
10:06 Tobbi joined #git
10:09 byte512 joined #git
10:09 govg joined #git
10:09 Varazir moritz: done
10:11 manuelschneid3r joined #git
10:13 TSS_ joined #git
10:13 giovanni joined #git
10:14 giovanni left #git
10:14 achadwick joined #git
10:14 chll_ joined #git
10:15 awkwardusername joined #git
10:17 marcogmonteiro joined #git
10:17 nettoweb joined #git
10:19 lindenle joined #git
10:21 ericP left #git
10:21 Gloomy joined #git
10:24 roelmonnens joined #git
10:25 roelmonnens joined #git
10:26 roelmonnens joined #git
10:26 chromity joined #git
10:27 suy Do you have some trick to show the contents of a blob without git? Just some party trick if you are peeking at gitdir's internals, or explaining how git works to people.
10:27 suy The best I've found is this: http://unix.stackexchange.com/a/49066
10:28 jast if it happens to be a loose object (not stored in a packfile), it's fairly straightforward, basically just about what's in that answer
10:28 jast objects stored in packfiles are complex and you really don't want to have to do it without git :)
10:29 jast (you'd have to do delta decompression, which uses git's own format, and potentially follow lengthy chains of delta objects)
10:29 suy Yeah, basically I'd do it before any packing, so I'll try with the alias suggested there.
10:29 King_Hual joined #git
10:31 BSAlb joined #git
10:32 ome joined #git
10:32 tvw joined #git
10:33 jast suy: however, loose Git objects have a small header in front of the compressed data that you need to cut off
10:33 jast I believe for "files" (blobs) it's a simple 4-byte length field
10:35 jast if you have a recent openssl installed, 'openssl zlib -d' is quite a bit simpler than having to mess with prepending the gzip header
10:36 Derperperd joined #git
10:38 Snugglebash joined #git
10:39 suy jast: yes, thank you
10:40 pR0Ps joined #git
10:42 miczac joined #git
10:46 tvl joined #git
10:47 govg joined #git
10:47 govg joined #git
10:49 [0xAA] joined #git
10:50 lullu joined #git
10:50 hhee joined #git
10:51 lullu left #git
10:53 YuGiOhJCJ joined #git
10:55 Xenophon1 joined #git
10:58 mvdan joined #git
10:58 mvdan hi! is there anything like 'git branch --merged' but that also works with rebased commits?
10:58 mvdan as in, the commit is the same except the commit author/date and hash
10:59 pks mvdan: take a look at git-cherry(1)
10:59 xall joined #git
11:01 mvdan pks: looks interesting! seems to work on single branches, though
11:01 mvdan anything that I can run from master to get info on all branches?
11:04 pur3eval joined #git
11:04 pks mvdan: nothing I know of. Problem is that this can only work based on patch-IDs, as there is no other information which relates a cherry-picked commit with others. So the information is rather expensive to compute and also not always correct, unfortunately
11:05 mvdan right
11:05 ShalokShalom_ joined #git
11:05 mvdan I was hoping there would be something that would use a heuristic involving author and message :)
11:05 mvdan I'll try my hand at a quick script
11:05 mvdan thanks pks!
11:06 pity joined #git
11:06 pks You should rather try to use the same approach as git-cherry, I'd say. Using the commit author and message requires the commit messages to be sufficiently unique, which is often not the case. You could also script something by using git-patch-id(1), which computes tha patch ID for you
11:07 pR0Ps joined #git
11:08 mvdan pks: I would say it would be fine as long as you also use the author date
11:09 mvdan patch-id looks promising, thanks!
11:10 pks mvdan: okay, also using the author date would probably work in most cases
11:10 ismithers joined #git
11:11 mvdan I was thinking of avoiding the actual diff as rebasing could change it somewhat
11:11 mvdan perhaps that won't be a problem at all
11:12 pks mvdan: well, it is. As I said, patch IDs will not always work. It _does_ ignore things like whitespace when computing the patch ID, but as you said, the diff may still change when cherry-picking/rebasing
11:12 zeroed joined #git
11:12 mvdan thanks for all your help. usually google is enough, but all I was getting is articles on why rebasing is better than merging :)
11:13 pks Yeah. And this is an obvious use case why merging is sometimes better than rebasing... :/
11:13 mvdan well, right now the only disadvantage I see is a lack of a git branch --merged equivalent :)
11:14 pks Both approaches have their own merits and disadvantages
11:14 ToBeCloud joined #git
11:16 mvdan what merits do you see in merging? I'd only ever use it to merge upstream branches, where keeping the hashes is important
11:16 afuentes joined #git
11:16 mvdan otherwise I see rebasing as superior
11:17 Renter ^
11:18 mvdan (would like to be proven wrong)
11:18 pks You retain the actual history. This can be a huge benefit of merging in certain scenarios
11:19 mvdan like what I described above?
11:19 Oatmeal joined #git
11:19 pks Among others, yeah
11:19 mvdan what others? :)
11:20 pks Merge-upwards work flows, e.g. If they are applicable, merging makes reintegrating changes into newer release-branches and the current line of development a breeze
11:21 pks So yeah, the most important benefit that I personally see is that you can still relate how features came into existence. So you seet he branching point and can easily group commits together by taking a look at the branching structure
11:21 pks Which is lost when you simply rebase and do ff-only merges afterwards
11:21 mvdan right, so either to keep the commit hashes, or to see the tree structure when it's non-trivial
11:21 mvdan makes sense
11:22 mvdan I'm mainly thinking of patches and PRs, where merging seems to be the norm on sites like github
11:22 mvdan so afaik neither applies
11:23 pks For simple bug fixes and one-commit PRs I'd certainly agree
11:23 dec0n joined #git
11:23 jameser joined #git
11:26 chromity joined #git
11:26 MineCoins joined #git
11:26 dami joined #git
11:27 dami hey
11:27 gitinfo dami: 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.
11:27 chromity joined #git
11:27 dami i have branches on a remote, i can only checkout to master or origin/branchname, but i can't check out to branchname
11:27 _ikke_ dami: What happens when you do?
11:27 dami error: pathspec 'cjdns' did not match any file(s) known to git.
11:28 tobiasvl dami: fetch first?
11:28 grawity dami: do you have two remotes?
11:28 dami did a fetch -a, fetch branchname
11:28 dami grawity: yes
11:28 grawity dami: do they have the same branches?
11:28 dami i hope so, checked out branches locally then push --all to the new remote
11:28 chromity joined #git
11:29 dami i just cloned it from fresh and it works now X
11:29 grawity well then that's why
11:29 dami X|, what would the "fix" be for multiple repos though?
11:29 dami afaik git has a local branch as well, can i only switch to remote/branchname when i got more than one?
11:29 _ikke_ dami: explicitly create the branch
11:29 grawity create the local branches by hand, without relying on automagic
11:29 grawity i.e. `git checkout -b cjdns origin/cjdns` or `git checkout -b cjdns foo/cjdns`
11:29 fstd joined #git
11:30 _ikke_ git checkout -b branchname origin/branchname
11:30 chromity joined #git
11:30 grawity if it *had* a local branch, you certainly could just check it out
11:31 chromity joined #git
11:31 dami alright, wasn't aware it tracks them so separately
11:32 chromity joined #git
11:32 dami anyway, thanks
11:33 jameser joined #git
11:33 dami left #git
11:33 chromity joined #git
11:34 fstd joined #git
11:37 SuDo__ joined #git
11:37 Derperperd joined #git
11:38 jameser joined #git
11:41 andrew711 joined #git
11:42 m4sk1n joined #git
11:43 xall_ joined #git
11:43 chipotle joined #git
11:44 JeroenT joined #git
11:45 Macaveli joined #git
11:49 mischat joined #git
11:49 pR0Ps joined #git
11:51 nowhere_man joined #git
11:52 ertesx joined #git
11:53 mischat_ joined #git
12:00 pR0Ps joined #git
12:00 jameser joined #git
12:00 dsdeiz joined #git
12:01 denisMone joined #git
12:02 ac3takwas joined #git
12:03 PtxDK joined #git
12:05 mostlybadfly joined #git
12:07 askb joined #git
12:08 aard__ joined #git
12:09 DrSlony joined #git
12:10 zeroed joined #git
12:10 zeroed joined #git
12:10 jameser joined #git
12:12 a_thakur joined #git
12:13 a_thakur joined #git
12:14 timini Hey im cherry picking a bunch of commits from one branch to another
12:14 DrSlony Hello, I tagged "5.0-gtk2" on branch master and "5.0-gtk3" on branch gtk3. I need to get the branch name in a script. This works if I checkout gtk3 but it fails if I checkout 5.0-gtk3: "git symbolic-ref --short -q HEAD" How do I get the branch name when checking out a tag? Alternatively, how do I get the tag name if I happen to check one out?
12:14 timini but loads of them are empty
12:15 timini how can i find all the empty commits on a branch
12:15 timini i mean commits which will be empty when they are cherry-picked to another branch
12:16 ToxicFrog joined #git
12:16 mentazoom joined #git
12:17 Varazir left #git
12:18 ToxicFrog joined #git
12:21 lindenle joined #git
12:22 Semtex-h joined #git
12:24 str joined #git
12:28 merlac joined #git
12:31 finalbeta joined #git
12:32 intellix joined #git
12:32 jameser_ joined #git
12:33 forrestv joined #git
12:34 timini hey how can I find a commit which will be empty when i cherry-pick it?
12:34 Hudu joined #git
12:34 mvdan pks: thanks again for your help - in an hour I came up with this https://github.com/mvdan/git-picked
12:34 mvdan seems to work well for my needs :)
12:35 Mo joined #git
12:35 Snugglebash joined #git
12:37 anddam left #git
12:38 pR0Ps joined #git
12:38 Derperperd joined #git
12:38 mischat joined #git
12:39 pks mvdan: glad to be of any help :)
12:39 pluszak joined #git
12:40 clmsy joined #git
12:41 Dumblez0r joined #git
12:42 jameser joined #git
12:43 awkwardusername joined #git
12:44 Tobbi joined #git
12:45 awkwardusername joined #git
12:46 atrigent_ joined #git
12:47 pluszak So I have a repo with multiple directories, each directory holds css and js and images and otehr such configuration for a website. Does anyone have a suggestion how to handle it? I don't really think 4000 seperate repositories would be a good idea.
12:48 pluszak But the current system is a killer on it's own
12:48 ProbabilityMoon joined #git
12:48 fstd joined #git
12:49 ntnn joined #git
12:49 moritz how big are they files? how often do they change?
12:50 JeroenT joined #git
12:51 fstd joined #git
12:51 pluszak not that much, math tells me it's some 2 megabytes per directory.
12:51 moritz so 4k directories, each 2MB?
12:52 moritz or how many directories?
12:52 moritz 'cause if it's just O(20MB) total or so, I'd just throw them into a git repo
12:52 osse timini: you can test them, with cherry-pick :p
12:52 moritz if it's more like >= 1GB, I'd find something else
12:53 pur3eval joined #git
12:53 pluszak yes, 8 gigs in total for the repo, 4k directories
12:54 osse pluszak: are there lots of duplicates?
12:54 osse the same CSS For 500 sites etc?
12:54 jast what's the main reason you're putting them in git?
12:55 pluszak osse: yes
12:55 osse pluszak: that will compress quite well
12:55 pluszak jast: the main reason is someone did it this way years ago
12:55 JanC_ joined #git
12:55 jeffreylevesque joined #git
12:55 jast right :)
12:55 pluszak osse: yes it does compress well, however the repo's size is making it hard to pull it locally when you just gotta do one fix
12:56 jast so are these directories updated separately or do they tend to get updates together? (I can guess but I'd rather know for sure)
12:56 zeroed joined #git
12:56 zeroed joined #git
12:56 fstd joined #git
12:57 pluszak Seperately rather. I'd say even that the directories never get updated after the initial work
12:57 pluszak But cause it's so huge somehow we ended with multiple people all working on a single remote repo
12:58 jast right. how many places (filesystem locations/servers) is the repo used in for production/live?
12:58 jast (one/a few/many is all I need to know)
12:58 jast same question about number of people working on the repo
12:58 jast with same accuracy :)
13:00 jast to foreshadow a bit, I think most likely Git isn't the ideal choice of tool, so I'm trying to figure out (a) which tool is a better alternative or (b) whether it's possible to do nasty things with Git that will sort of work in your scenario
13:01 mentazoom joined #git
13:01 muggz joined #git
13:02 pity joined #git
13:03 tobiasBora joined #git
13:06 denete_ joined #git
13:07 timini joined #git
13:08 marcogmonteiro joined #git
13:12 Snugglebash joined #git
13:13 govg joined #git
13:13 MonsterKiller_ joined #git
13:14 synthroid joined #git
13:15 nettoweb1 joined #git
13:17 synthroid joined #git
13:18 ShapeShifter499 joined #git
13:20 fstd joined #git
13:23 MonsterKiller joined #git
13:23 mischat_ joined #git
13:23 ojdo joined #git
13:31 finalbeta joined #git
13:35 cdown_ joined #git
13:36 nowhere_man joined #git
13:36 mohabaks joined #git
13:36 bruce_lee joined #git
13:36 bruce_lee joined #git
13:37 tcorneli joined #git
13:37 blink_ joined #git
13:38 fstd joined #git
13:38 tcorneli joined #git
13:38 Derperperd joined #git
13:42 afuentes joined #git
13:43 fstd joined #git
13:44 timini joined #git
13:45 theoceaniscool joined #git
13:45 ollehar joined #git
13:46 osse sounds like it
13:46 ollehar I have a feature I do not want to carry into another branch because of difficult merge. can I revert it "in" the merge, but keep it in my first branch?
13:46 fstd joined #git
13:47 ollehar new feature in master causes trouble with merge into dev, so I want to revert it "into" dev but keep it in master.
13:47 ollehar possible?
13:47 cdown joined #git
13:47 jeffreylevesque joined #git
13:48 osse ollehar: yes, but then it will be hard to merge it for realsies in the future
13:49 ollehar osse: ok
13:49 ollehar what's the method?
13:49 osse ollehar: git merge --no-commit; tweak the result; git commit
13:49 masuberu joined #git
13:49 solf1re joined #git
13:49 osse or git merge; make changes; git commit --amend
13:49 ollehar the problem is that `git merge` is very heavy right now
13:50 ollehar will check --no-commit option
13:50 DaveTaboola joined #git
13:51 awkwardusername joined #git
13:51 JanC_ joined #git
13:51 osse blink_: you can prepend the rebase todo list with 'exec' lines
13:51 fstd joined #git
13:52 reverse_light joined #git
13:52 mehola joined #git
13:52 osse blink_: what is the use case?
13:53 whomp joined #git
13:55 marcogmonteiro joined #git
13:56 DrSlony osse I tagged a new release (actually two new releases - 5.0-gtk2 in the master branch and 5.0-gtk3 in the gtk3 branch) but of course there was a bug with a script which I now fixed. I'm unclear how to proceed. Normally I would commit the fix to master and merge to gtk3, but does that mean I must make a new tag? Can I move the existing tags to point to the new commit?
13:56 osse you can force-rewrite tags, yes
13:57 DrSlony osse what is the danger with doing that?
13:57 osse DrSlony: mostly the same as force-pushing branches
13:57 osse Plus I think with older gits other people had to git tag -d {tag}; git fetch --tags
13:58 osse IE it wouldn't be automatically updated
13:58 DrSlony What would you recommend I do?
13:58 osse without knowing how many you are, how tags are used, etc etcet c...
13:58 osse i'd say fuckit and force push
14:00 finalbeta joined #git
14:00 DrSlony a friend merged master into his pixelshift branch since it was tagged, if i force-push the tag what can i expect to happen, what problems?
14:02 osse that your friend's tag won't be updated when he fetches
14:02 JamesBaxter joined #git
14:02 DrSlony osse even if he's using a new git client? If so, how does he get his tags to update?
14:03 HardlySeen joined #git
14:03 osse DrSlony: no, not even. i think it's only a problem with older gits. and I already told you how to deal with it
14:03 osse git tag -d {tag}; git fetch --tags
14:04 DrSlony ok, great, thank you
14:11 navidr joined #git
14:12 jagob joined #git
14:14 Ryanar joined #git
14:14 jimi_ joined #git
14:14 DrSlony osse one more unrelated question. We have a script which we run upon new release, this script generates a file which contains some metadata (branch name, commit hash) which we then include in the tarball so that people can compile the program and have the metadata without having git installed, as is the case with the package maintainers who build for ubuntu. However a part of the script (git symbolic-ref --short
14:15 DrSlony -q HEADdoesn't work when you check out a tag, it only works when you check out a branch. The script used to do "git checkout <tag>" but I read today in the manual that "You can’t really check out a tag in Git" and the proper way to do it is to "git checkout -b master-5.0 5.0-gtk2".
14:15 jbbat_ joined #git
14:15 osse easy fix: remove the branch name
14:16 osse use tag name instead
14:16 fstd joined #git
14:16 Celelibi Is there a hook run on git add?
14:17 _geoff joined #git
14:17 DrSlony osse that is probably the cleanest solution
14:18 osse DrSlony: it is also the most sensible!
14:18 osse that's what tags are for
14:18 kadoban No. Earliest is pre-commit I believe, or that one that does commit messages more likely. There are smudge/clean filters if you need that, DrSlony
14:19 DrSlony kadoban sorry I didn't get a word of that
14:19 Celelibi Maybe this was ment to me?
14:20 kadoban Oh, yes
14:20 ollehar left #git
14:20 bgerber joined #git
14:20 Celelibi What are those clean filters?
14:21 kadoban I've used the approximately once, you're better off getting a description from whatever docs are around. They're part of .gitattributes stuff
14:21 osse the filters to clean up the mess that the smudge filters have made
14:21 kadoban Maybe man git attributes
14:21 gitinfo the git manpage is available at http://jk.gs/git.html
14:22 kadoban Nope ... that didn't work.
14:22 fstd joined #git
14:23 awkwardusername joined #git
14:23 cqi joined #git
14:24 eduardas_m joined #git
14:24 lindenle joined #git
14:27 charlesr joined #git
14:27 sigma__ joined #git
14:28 sigma__ I'm getting the message "your branch is head of origin/master by 10 commits" because I've been using `git fetch origin master` without configuring my remote-tracking branch as described here: https://git-scm.com/docs/git-fetch#_config​ured_remote_tracking_branches_a_id_crtb_a
14:29 sigma__ I know this now, but how do I correct the issue so I can get back on track? Should I just do `git pull origin master`?
14:29 psyb0t_ joined #git
14:29 Gsham joined #git
14:30 sigma__ I should note that I plan on using `git fetch origin` going forward, instead of managing remote-tracking branch configurations on my servers. I'm basically looking for a way to trigger a remote-tracking branch update once so that I can continue from here using `git fetch origin`.
14:31 kpease joined #git
14:32 kpease_ joined #git
14:33 Es0teric joined #git
14:33 todd_dsm joined #git
14:36 ikelso joined #git
14:36 pur3eval joined #git
14:36 clmsy joined #git
14:39 Derperperd joined #git
14:39 anonymuse joined #git
14:39 Es0teric joined #git
14:39 awkwardusername joined #git
14:41 timini joined #git
14:41 nowhere_man joined #git
14:44 GodGinrai joined #git
14:44 Kulrak joined #git
14:44 Kulrak joined #git
14:46 kpease joined #git
14:47 LionsMane joined #git
14:47 kpease_ joined #git
14:48 govg joined #git
14:48 ShekharReddy joined #git
14:48 peepsalot joined #git
14:48 bgerber joined #git
14:49 mazert_ joined #git
14:49 sammyo joined #git
14:49 ThomasLocke_ joined #git
14:49 ThomasLocke_ joined #git
14:50 onmeac joined #git
14:50 overlord_tm joined #git
14:51 blackwind_123 joined #git
14:51 Ryanar joined #git
14:53 kpease joined #git
14:53 nickabbey joined #git
14:55 bannakaffalatta joined #git
14:56 lvh joined #git
14:57 GodGinrai joined #git
14:57 Jesus^ joined #git
14:59 Jackneilll i had a submodule for a repo which i forked and now want that submodule to point to my fork. what can i do? just rm and add again?
14:59 Jackneilll (i didnt change anything.)
14:59 Dougie187 joined #git
15:01 _geoff_ joined #git
15:01 e14 joined #git
15:01 Jesus^ left #git
15:02 rominronin joined #git
15:02 pluszak jast: well, basically just a staging and live server, local edit would be cool. And we have like...half a dozen people working at it any given time? I know someother tool would be better but git is still a step above nfs
15:05 Es0teric joined #git
15:06 tinanoo joined #git
15:06 thebope joined #git
15:06 peeps[lappy] joined #git
15:07 DrSlony osse what would happen if I make tag x in master, a colleague makes tag y in his branch, and then we merge the two?
15:08 GodGinrai DrSlony: merge the branches?
15:10 DrSlony yes
15:11 pluszak DrSlony: Well, the tags will stay on the commits they were added on
15:12 pluszak DrSlony: what did you think is going to happen?
15:12 Es0teric joined #git
15:14 bgerber joined #git
15:14 DrSlony pluszak my bad, i wasn't clear. I meant what would happen when I merge the branches and then type "git describe"?
15:14 Kulrak joined #git
15:14 Kulrak joined #git
15:15 AaronMT joined #git
15:17 cdg joined #git
15:21 ianmethyst joined #git
15:24 xissburg joined #git
15:24 LeBlaaanc joined #git
15:24 DrSlony We would like to know which branch a build comes from without having to examine the hash. "git symbolic-ref --short -q HEAD" works, but not if you "git checkout <tag>". Then again, the manual says you should not checkout a tag. So I'm wondering, should we carry on extracting the branch name and storing that in a metadata file, or should we stop using branch names entirely and use tags? If tags, and if there are
15:24 DrSlony two main branches (master tagged "5.0-gtk2" and gtk3 tagged "5.0-gtk3") with a bunch of other feature branches, should we tag those feature branches too? What would you recommend the tag name to be?\
15:24 isueue joined #git
15:24 bongjovi joined #git
15:25 pluszak well, there are several things in play here
15:26 timvisher joined #git
15:27 sower joined #git
15:27 ankit01ojha joined #git
15:28 pluszak DrSlony: where does it say you shouldn't checkout a tag?
15:28 sower test?
15:28 sower I'm registered, cool cool
15:29 DrSlony pluszak bottom paragraph https://git-scm.com/book/en/v2/Git-Basics-Tagging
15:29 texinwien_ joined #git
15:29 DrSlony "You can’t really check out a tag in Git, since they can’t be moved around. If you want to put a version of your repository in your working directory that looks like a specific tag (...)"
15:30 sower I know that this is more githib than git but - I'm not sure how to go about using issues properly? I thought that I was about to @ people that had commented in the repo previously in order to ping them - and that i would be able to assign an issue to people that had been involved in the project?
15:31 DrSlony sower you need to add contributors to your project if you want to assign issues to them. You can @mention any github user though, even ones completely unrelated to your project.
15:31 pluszak DrSlony: probably that could have been phrased better. You can checkout a tag but it's not a branch so it will be detached
15:31 voldz1tic3n joined #git
15:31 sower DrSlony:  OK cool - ill test contributor adding now
15:32 xall_ joined #git
15:32 ToxicFrog DrSlony: that's badly phrased; it means "you can `git checkout` a tag but it behaves the same as `git checkout <sha>` rather than checking out a branch, i.e. you are in a "detached HEAD" state and any commits you make will be branchless"
15:34 vdl joined #git
15:34 treia joined #git
15:34 DrSlony ok
15:36 douglascorrea joined #git
15:38 cdown_ joined #git
15:39 Kulrak joined #git
15:39 Kulrak joined #git
15:40 Derperperd joined #git
15:40 sower DrSlony:  OK I tested that - but  i don't want them to have full push access with the repo?
15:40 sower I basically just want to be able to make an issue then say "BLah you're doing X"
15:41 sower that kind of thing
15:42 GodGinrai sower: you can give contributors readonly access
15:42 sower GodGinrai:  how so?
15:42 GodGinrai sower: the same place where you add them
15:43 GodGinrai when you add them, you can choose if they have read&write or readonly
15:43 sower https://github.com/blah/bl​ah/settings/collaboration << in that link?
15:43 sower ill try again - i didn't catch that when i just tested it though
15:44 sower NO
15:44 sower Sorry, just 'no'
15:44 sower caps was no - but I can't see an option for that
15:44 sower GodGinrai: are you sure?
15:44 tang^ joined #git
15:45 borkr joined #git
15:45 nettoweb joined #git
15:45 icey joined #git
15:45 icey joined #git
15:45 mischat joined #git
15:45 GodGinrai sower: ah, it seems that granularity is only afforded to organizations
15:46 nneul joined #git
15:46 Gloomy joined #git
15:47 sower I'm not an organisation :/ ok tho
15:47 sower cheers, any suggested work flows for this?
15:47 sower maybe i'm going about it the wrong way
15:47 tinanoo joined #git
15:48 GodGinrai according to github, tho', @mentions should let you mention *any* user on github
15:49 danmackay joined #git
15:49 sower yeah but they don't pop up , so I need to remember the people... but yeah - i might just use that and make custom labels for this
15:52 pur3eval joined #git
15:52 ash_workz joined #git
15:54 whomp joined #git
15:55 diogenese joined #git
15:56 rajat joined #git
15:58 timvisher joined #git
15:59 gajus joined #git
16:00 ikelso joined #git
16:00 apotry joined #git
16:01 b100s joined #git
16:01 b100s hi2all
16:01 ikelso joined #git
16:01 b100s why can be 'git merge branchA' is up to date, but 'git diff branchA' gives me a lot changes ?
16:02 grawity because merges usually keep changes from *both* sides
16:02 rominronin joined #git
16:03 b100s grawity, sorry, cant get what do you mean
16:03 nowhere_man joined #git
16:03 grawity so the previously-happened merge imported changes from branchA *and* kept changes which happened in master
16:03 roelmonn_ joined #git
16:04 grawity if master added 'B', and branchA added 'A', and there was a merge of branchA into master, then master would have commits which add both 'A' and 'B'
16:04 charlesr joined #git
16:04 grawity so there's nothing more to merge because it already has all the changes from A
16:04 grawity but it's still different because it *also* has all the changes from B
16:05 grawity well, it's a guess really
16:07 jknetl joined #git
16:08 onmeac left #git
16:08 onmeac joined #git
16:09 onmeac joined #git
16:09 shinnya joined #git
16:09 cite joined #git
16:10 zeroshft joined #git
16:10 cite joined #git
16:11 finalbeta joined #git
16:12 b100s grawity, yes, sure you are right, but once i have a very interesting thing: we have branchA and branchB started from master and both has few commits; merge brnachB into branchA; revert merge commit; try to merge again - already up to date; but actually there isn changes from branchB; isnt it?
16:12 MapMan Hi! Can someone help me checkout this repo: https://github.com/jacobdufault/fullinspector ? There's a manual at the bottom how to check it out. It says to do a sparse checkout but I follow the guide and my subrepo is empty :( Any ideas?
16:12 b100s looks likes git just try to search HEAD commit of branchB in branchA
16:12 grawity b100s: but the commits are there
16:12 b100s that all what it do
16:12 Nilesh_ joined #git
16:13 b100s grawity, yes, and knowing that is so sad; git no as smart as i suppose )
16:14 GodGinrai That's not a matter of git being "smart".  You shouldn't just be merging and reverting things without thinking it through
16:17 brent__ joined #git
16:18 CheckDavid joined #git
16:19 a_thakur joined #git
16:19 Snugglebash joined #git
16:19 ash_workz joined #git
16:19 troydm joined #git
16:19 madewokherd joined #git
16:19 GodGinrai b100s: this is a good resource discussing the issue: https://www.kernel.org/pub/software/scm/​git/docs/howto/revert-a-faulty-merge.txt
16:20 geoid_ joined #git
16:20 b100s GodGinrai, thanks
16:21 GodGinrai np
16:21 troydm hey all! is it possible to have submodule folders in git repository to be completly ignored by main repo, because everytime I commit something to submodule I see my main repo now has a change and needs an commit to update latest commit number in submodule?
16:21 denisMone joined #git
16:22 troydm I want to manage them completly separately
16:23 troydm also not I don't have .gitmodules file in my main repo
16:23 troydm *note*
16:24 moritz troydm: just get rid of the submodule, ignore the path in the parent module, and git clone it separately
16:24 govg joined #git
16:25 suttonwilliamd joined #git
16:26 Derperperd joined #git
16:26 douglascorrea joined #git
16:26 ResidentBiscuit joined #git
16:26 pur3eval joined #git
16:27 svm_invictvs joined #git
16:28 s2013 joined #git
16:28 eeyee joined #git
16:30 eeyee left #git
16:30 EliasH joined #git
16:32 xall_ joined #git
16:32 Snugglebash joined #git
16:32 Es0teric joined #git
16:33 hobodave joined #git
16:34 jccn joined #git
16:37 troydm moritz: ahh, right, probably got added earlier, thx
16:37 troydm left #git
16:38 timvisher joined #git
16:39 denete__ joined #git
16:39 nettoweb joined #git
16:42 sigma__ left #git
16:43 cjohnson yep just don't use submodules if you don't want submodules
16:44 Electrometro joined #git
16:47 yehowyada joined #git
16:48 sower joined #git
16:49 Balliad joined #git
16:54 rivarun joined #git
16:55 Es0teric how do i undo the last commit i made?
16:55 whitby Es0teric: do you want to get rid of it completely or do you want undo and then redo with modifications?
16:56 GodGinrai Es0teric: there are many different ways, all which you would use under different circumstances
16:56 Es0teric i want to get rid of it completely, whitby because i commited in the wrong branch
16:56 GodGinrai Es0teric: first... have you already pushed this commit?
16:56 Es0teric GodGinrai no
16:56 GodGinrai then a git reset should work fine
16:56 GodGinrai `git reset HEAD~1`
16:56 GodGinrai use --hard to reset the files, too
16:56 Rodya_ joined #git
16:59 Hedgework joined #git
17:01 mathematic-alpha joined #git
17:01 mathematic-alpha left #git
17:06 cdg joined #git
17:07 jagob joined #git
17:07 skift joined #git
17:09 mischat_ joined #git
17:11 mischat_ joined #git
17:12 timvisher joined #git
17:13 nickabbey joined #git
17:14 Guest36461 joined #git
17:15 strugee joined #git
17:18 nostrora joined #git
17:21 imachuchu joined #git
17:21 Jfs joined #git
17:23 Ranmaru joined #git
17:25 rubyonrailed joined #git
17:29 mischat joined #git
17:29 ikelso joined #git
17:32 redhedded1 joined #git
17:32 polyrob_ joined #git
17:34 rominron_ joined #git
17:35 codemann joined #git
17:36 anonymuse joined #git
17:36 thiago joined #git
17:38 Leonardo_0112 joined #git
17:40 robotroll joined #git
17:40 m0viefreak joined #git
17:43 MapMan Hi! I have a repo with a subrepo. We've decide to get rid of the subrepo and simply make the files a part of our repo. What's the best way to handle this and are there implications for repo users?
17:43 dendazen joined #git
17:44 Leonardo_0112_ joined #git
17:45 chardan joined #git
17:45 BackEndCoder joined #git
17:46 Leonardo_0112 left #git
17:48 hahuang61 joined #git
17:49 doener joined #git
17:49 Eugene MapMan - what type of subrepo? submodules?
17:52 energizer joined #git
17:54 mehola joined #git
17:55 tyreld joined #git
17:55 MapMan Eugene: yes.
17:55 MapMan http://stackoverflow.com/questions/​1260748/how-do-i-remove-a-submodule
17:55 Eugene Use !submodule_rm to get rid of it, then just copy the files in like normal.
17:55 gitinfo [!submodules_rm] To delete a submodule, do this in the superproject: Edit/delete .gitmodules to remove the submodule, and `git add .gitmodules`. Then `rm -rf submodulepath; git rm -f --cached submodulepath; git commit -am "Removed submodules!"`  Inspect .git/config for "submodule" entries to remove.  Inspect .git/modules for caches to remove.  Possible alternatives to submodules: "!gitslave" or "!subtree"
17:56 pur3eval joined #git
17:57 MapMan Second answer seemed pretty good and that's what I did. What are implications for other users? As in, if they reclone the repo they will still have the submodule on say, master branch, right?
17:58 MapMan If they want to get rid of the submodule they have to switch to branch that doesn't have it but that creates a conflict as the subrepo files would be overwritten so they have to remove them first.
17:58 MapMan any way to make this process easier?
17:58 mda1 joined #git
17:59 Eugene Submodules kinda suck, sorry.
18:00 nd_ joined #git
18:00 nickabbey joined #git
18:00 timvisher joined #git
18:01 Cabanossi joined #git
18:01 Hudu joined #git
18:01 MapMan Eugene: got it, thanks :P
18:03 chipotle joined #git
18:04 aard_ joined #git
18:05 afuentes joined #git
18:06 whomp joined #git
18:06 ash_workz do notes for `git notes` get pushed?
18:06 lindenle joined #git
18:07 Starky joined #git
18:07 invisbl_ joined #git
18:07 Eugene Not by default; only refs/heads/ (aka branches) are pushed. You can of course push them manually.
18:08 mikecmpbll joined #git
18:09 jnavila joined #git
18:10 roelmonnens joined #git
18:11 bamb joined #git
18:11 bamb left #git
18:11 neilthereildeil joined #git
18:11 neilthereildeil hi
18:11 ash_workz --edit-description can't get pushed at all?
18:12 neilthereildeil i just ran git reset --hard, but refs/heads/master is not changing
18:12 neilthereildeil how can i change that ref?
18:15 boombatower joined #git
18:16 neilthereildeil how can i change /refs/heads/master?
18:16 aard__ joined #git
18:17 spacelord_ joined #git
18:19 Cabanossi joined #git
18:20 nidr0x joined #git
18:21 osse neilthereildeil: check it out, then reset
18:23 osse neilthereildeil: and you need to provide an argument
18:23 kyan joined #git
18:26 thiago joined #git
18:26 Cabanossi joined #git
18:26 Balliad joined #git
18:26 neilthereildeil osse: refs/heads/master points to 2a5c20db9da6a0ab136b958a972bc1702c3824f1. I want to make it point to 865cfeeaddede23ee9004075efc57fc146a7ec73
18:27 neilthereildeil i just checkout out 865cfeeaddede23ee9004075efc57fc146a7ec73 and git reset --hard 865cfeeaddede23ee9004075efc57fc146a7ec73, but it didnt work
18:27 neilthereildeil master still points to 2a5c
18:27 osse git checkout refs/heads/master; git reset --hard 865cfeeaddede23ee9004075efc57fc146a7ec73
18:27 osse sorry
18:27 osse git checkout master; git reset --hard 865cfeeaddede23ee9004075efc57fc146a7ec73
18:27 neilthereildeil whats the difference between the 2 commands?
18:27 osse I'm not sure. I just know the latter one works
18:28 osse ahh, using refs/heads/master detaches, just like your attempt does
18:28 neilthereildeil ok that worked
18:28 neilthereildeil thanks
18:28 neilthereildeil detaching wont owrk?
18:28 osse no
18:28 osse detached = HEAD points directly to a commit instead of a branch
18:28 osse reset changes what HEAD points to
18:28 hasc joined #git
18:29 HardlySeen joined #git
18:29 neilthereildeil ohh ok
18:29 neilthereildeil so always make sure HEAD and master are synchronized when trynna reset master?
18:29 osse yes
18:29 neilthereildeil k thx
18:32 synthroid joined #git
18:32 kulelu88 joined #git
18:33 canuk joined #git
18:34 MapMan How do I merge Foo into Bar but exclude commit X from Foo?
18:34 neilthereildeil left #git
18:35 alphor joined #git
18:35 GodGinrai MapMan: revert commit X either before or after the merge
18:36 MapMan I guess creating a new branch, rebasing and then merging is the way to go?
18:37 str joined #git
18:39 duderonomy joined #git
18:40 Vampire0 MapMan, do you want to exclude commit X or the changes from commit X?
18:42 timvisher joined #git
18:42 MapMan Vampire0: I want both commit and the changes gone
18:43 MapMan I tried rebasing and dropping the commit I don't want but the rebase fails...
18:43 danmackay joined #git
18:46 Vampire0 MapMan, either cherry-pick all commits from Bar, except of X, or do an interactive rebase removing X before merging, or do an interactive rebase of Bar onto Foo, dropping X. If you did changes after X, that touch the same lines than X did, you will of course get conflicts that you have to resolve
18:47 aard_ joined #git
18:47 emg joined #git
18:47 Balliad joined #git
18:48 cdown joined #git
18:48 aard__ joined #git
18:49 djb-irc joined #git
18:50 peepsalot joined #git
18:51 nickabbey joined #git
18:52 lvh joined #git
18:54 webstrand joined #git
18:54 jimi_ joined #git
18:55 jknetl joined #git
18:55 invisbl joined #git
18:56 aard_ joined #git
18:56 invisbl joined #git
18:57 nettoweb joined #git
18:58 FeyFre joined #git
18:59 FeyFre I accidentally commited(and pushed into remote) big binary file(60m). How I can get rid of it? (remote is fully controllable by me)
18:59 Spherical joined #git
18:59 _ikke_ Is it the last commit?
18:59 osse FeyFre: amend the commit that introduced it then force push
19:00 aard__ joined #git
19:00 FeyFre no. already made few commits after
19:00 _ikke_ then use rebase
19:00 Starky joined #git
19:01 linuxmodder joined #git
19:01 FeyFre it sounds for me as noice.. I am not able to to do it w/o instructions
19:01 lvh left #git
19:02 osse FeyFre: if X is the troublesome commit:   git rebase -i X~1; mark the first commit 'edit'; save and quit. git rm file; git commit --amend; git rebase --continue; git push -f
19:02 FeyFre thanks, I will try
19:04 msafi joined #git
19:04 Tobbi joined #git
19:05 msafi Could someone tell me how to read or what it means "@ -11,6 +11,6 @@"? https://i.imgur.com/KqMBU02.png
19:05 rominron_ joined #git
19:06 osse msafi: the line starting with 'location: Settings..' is the 11th line in the file, both before and after
19:06 msafi osse: it's not the 11th line in the file though?
19:07 osse are you sure?
19:07 msafi Sorry, you're right
19:07 msafi It is
19:07 osse and the number of lines you're looking at (including the context) is 6 lines both
19:07 osse both before and after
19:08 djb-irc joined #git
19:08 mischat joined #git
19:09 msafi Not sure what you mean by the second part...why are there -11 and +11?
19:09 msafi And I don't count 6 lines
19:09 ToxicFrog That just means it starts at line 11 in both the "before" file (-) and the "after" file (+)
19:09 osse I suppose the - and + just indicate that the numbers apply to the before and after
19:09 osse msafi: count the - and + separately
19:09 osse msafi: add the contexrt
19:10 FeyFre fatal: Needed a single revision
19:10 FeyFre invalid upstream 08f63ef600e967d48862bf333dfebba4c57492a3~1
19:10 ToxicFrog And the ,6 means that the diff hunk is 6 lines long; if you split it into two versions, "before" and "after", each one would be 6 lines.
19:10 FeyFre rebase failed
19:10 Limes_ joined #git
19:11 osse FeyFre: !repro
19:11 gitinfo FeyFre: [!transcript] 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
19:11 Cabanossi joined #git
19:11 Sasazuka joined #git
19:11 Vampire0 FeyFre, was the problematic commit the first commit in the repo?
19:11 msafi I think it makes a little more sense now. Thanks!
19:12 FeyFre actually yes :)
19:12 osse msafi: so if you removed the two lines but only added one, it would say -11,6 +11,5
19:12 msafi osse: I see!
19:13 chachasmooth joined #git
19:13 osse msafi: and finally if you added some more lines later in the file it would say (eg.)  -50,7 +49,7
19:13 osse msafi: because the first difference made the new file one line shorter
19:13 bannakaffalatta joined #git
19:14 Vampire0 FeyFre, then you have to use --root instead of X~1
19:14 msafi osse: that makes sense
19:17 guest3456 joined #git
19:18 guest3456 my git bash console on windows only shows 16 colors
19:18 guest3456 which looks gross when it opens Vim to do a commit message
19:18 guest3456 how do i get better colors ?
19:19 osse guest3456: do you use the windows console or mintty ?
19:19 ikelso joined #git
19:19 guest3456 i'm guessing windows console
19:19 guest3456 i have no idea what mintty is
19:19 krillr joined #git
19:19 osse in that case 16 is all you're gonna get
19:19 guest3456 what is mintty and how/why would i use it
19:19 SuDo__ joined #git
19:19 ToxicFrog guest3456: mintty is a more capable terminal emulator
19:19 tang^ when installing Git for Windows, you are asked whether to use MinTTY or Windows Console for bash
19:20 ToxicFrog It supports features that cmd.exe doesn't, like more than 16 colours
19:20 osse guest3456: it's what you get when you pick "Git Bash" from the start menu instead of "Git CMD"
19:20 guest3456 ah
19:20 guest3456 fk
19:20 guest3456 so i need to reinstall
19:20 FeyFre osse, vampire0, thanks it worked
19:20 ToxicFrog (cmd.exe is, frankly, an awful TE and I wish it would die already)
19:20 guest3456 i dont have 'Git CMD' in my start menu
19:20 osse what do you have then
19:20 ToxicFrog (MS should have put it out to pasture and shipped rxvt instead in XP or something)
19:21 texinwien joined #git
19:21 guest3456 only 'Git Bash'
19:21 timvisher joined #git
19:21 guest3456 and 'Git GUI'
19:21 FeyFre now other question: that particular file still stored inside .git (in remote also) - can I clean it out?
19:21 guest3456 ah git --version 1.9.4
19:21 osse guest3456: open git bash, click on the git icon in the corner and pick options...
19:21 guest3456 i guess i should upgrade
19:21 osse guest3456: oh, that's ooold. uninstall and get the new one
19:22 guest3456 alright let me do that and ill choose mintty
19:22 guest3456 brb
19:22 g5pw joined #git
19:23 FeyFre no questions: git gc did the trick
19:23 guest3456 hrm
19:23 bttf joined #git
19:23 jaafar joined #git
19:23 guest3456 looks like it wants to install into AppData instead of Program Files
19:24 osse ehh
19:24 guest3456 maybe i should uninstall the old version
19:24 osse are you getting that github stuff?
19:24 guest3456 i dont think this will overwrite
19:24 Balliad joined #git
19:24 guest3456 no
19:24 guest3456 although
19:25 guest3456 the installer didn't prompt for admin credentials
19:25 guest3456 so maybe thats why
19:25 guest3456 right click run as admin, now it installs in program files
19:25 ToxicFrog Yeah, if you want to install as not-admin (or write the installed files afterwards) it has to go in not-program-files, and AppData is a sensible place for that
19:25 guest3456 but
19:25 Muzer joined #git
19:25 guest3456 my old install was into program files (x86)
19:25 guest3456 so i guess i should just uninstall
19:25 ToxicFrog Although IME most installers will prompt to sudo rather than defaulting to AppData
19:26 osse TIL
19:26 manuelschneid3r joined #git
19:26 osse Almost all installers I've seen in my life default to Program Files regardless
19:26 ssudo___ joined #git
19:26 ikelso joined #git
19:26 osse Maybe I've been admin all my life, haha
19:27 guest3456 yikes
19:27 Eugene Actually its %programfiles%, but what do I know
19:27 guest3456 control panel reports
19:27 guest3456 Git version 1.9.4preview20140815
19:27 guest3456 thats old
19:27 ssudo___ joined #git
19:27 _ikke_ Atom by default installs in appdata
19:27 superguest joined #git
19:27 ToxicFrog osse: pre-windows...Vista, I want to say? in practice either program files was world writeable or everyone was running as admin
19:27 ToxicFrog At some point they implemented greater separation of powers and added a sudo-alike (UAC)
19:28 codemann- joined #git
19:28 ToxicFrog This is the "program xyz wants to make changes to your system, approve/deny?" prompt you get when running installers on modern windows.
19:28 Eugene Program Files has never been world-writable under a NT kernel; you're thinking of 95/98 & derivatives
19:28 ToxicFrog Eugene: yes, win9x is what I was referring to with "world writeable"
19:28 guest3456 correct
19:28 grawity Eugene: unless you start off with a FAT filesystem
19:28 superguest how do I unset / remove a config variable?  I accidentally did "git config global.username 'Foo Bar'"
19:28 guest3456 but in WinXP, default users were 'power users' which were mini-admins
19:29 Eugene AppData was introduced in the XPish days as the Normal User place; some apps take advantage of this("Install for Everyone or Just You?")
19:29 osse superguest: --unset
19:29 guest3456 win Vista introduced true 'limited' users who didn't have write access
19:29 grawity yeah, in XP you couldn't do stuff without being an admin or runas'ing every damn thing :/
19:29 guest3456 XP you COULD do stuff
19:29 superguest thanks!
19:29 guest3456 because you were a 'power user'
19:29 Eugene I think it was Vista that brought in redirection, wherein non-priv users could "Install to Program Files", but in reality its inside AppData
19:30 kpease_ joined #git
19:30 Eugene Very little has actually changed in the underlying WIn32 APIs between XP and 10.... the difference is that apps are now aware of how to handle failures; not Windows.
19:30 Sudo___ joined #git
19:30 grawity guest3456: I mean, if you tried to have a secure configuration
19:30 guest3456 ok this new git install is defaulting to 'use git from windows cmd prompt'
19:31 guest3456 saying it will add PATH vars
19:31 osse pick the middle thing
19:31 denete_ joined #git
19:31 timini joined #git
19:32 jnavila joined #git
19:32 Sudo___ joined #git
19:33 m4sk1n joined #git
19:34 denete_ joined #git
19:34 guest3456 mintty is default choice now too
19:34 YuGiOhJCJ joined #git
19:35 interrobangd joined #git
19:35 Sudo___ joined #git
19:35 codemann_ joined #git
19:36 denete_ joined #git
19:37 guest3456 hrm ok
19:37 guest3456 looks like a different terminal now
19:37 guest3456 but it must not be respecting my vimrc
19:38 Sudo___ joined #git
19:39 Anja joined #git
19:40 rivarun joined #git
19:40 Sudo___ joined #git
19:41 felher_ joined #git
19:46 Sazhen86 joined #git
19:47 aswen joined #git
19:48 donught joined #git
19:49 jbitdrop joined #git
19:49 ocbtec joined #git
19:50 ESphynx joined #git
19:51 ESphynx hey guys, is there a way to search your stashes to only see those that modify a certain file
19:51 aard_ joined #git
19:51 hagridaaron joined #git
19:52 nowhere_man joined #git
19:52 GodGinrai ESphynx: http://stackoverflow.com/questions/211​30765/stashed-some-code-but-dont-know-​which-stash-it-is-in-how-to-figure-out
19:52 ESphynx thanks GodGinrai
19:52 GodGinrai np
19:52 ESphynx GodGinrai: in other words, no?
19:53 ESphynx I did know about git stash show -p, and on Windows without |grep
19:53 GodGinrai this is *list* not show
19:53 GodGinrai how was that answer a no?
19:53 ikelso joined #git
19:54 ESphynx GodGinrai: well what I would hope for would be a 'git stash show -p -- src/foo/here/
19:55 ESphynx like most other git commands have
19:56 cdown joined #git
19:56 GodGinrai yea, I don't think that git has a command that does exactly what you want.
19:57 GodGinrai But hey, unix philosophy
19:57 ESphynx it would be a welcomed addition :)
19:57 GodGinrai just chain it together with other utilities
19:57 ESphynx what's the Unix philosophy? :P
19:57 bremner use grep
19:57 ESphynx yes except I'm not on Unix and don't have those utilities ;)
19:57 timvisher joined #git
19:57 bremner no, that's the answer to your question, of what is the unix philosophy
19:57 GodGinrai ESphynx: even if you are on windows, "git bash" comes with grep
19:58 durham joined #git
19:58 ESphynx I don't use git bash :P
19:58 ikelso joined #git
19:58 multi_io joined #git
19:58 GodGinrai ESphynx: that doesn't mean grep wasn't installed
19:58 GodGinrai Also, if you are using git in powershell or the like... do yourself a favor and stop
19:59 ESphynx GodEater: I use git in good old vanilla cmd!
19:59 GodGinrai that's probably the worst way you could use it
20:00 b100s joined #git
20:00 ESphynx well my philosophy is that tools should work in the native environmnt
20:00 Balliad joined #git
20:01 GodGinrai they do work
20:01 GodGinrai and if you had installed grep into your path (as was an option afforded to you in the git for windows install) you could use that in cmd as well
20:01 pandeiro joined #git
20:01 sargas joined #git
20:01 ESphynx they do, except for some weird issues sometimes when scrolling and you exit and it looks like you're still inside the scrolling or something
20:02 ESphynx GodGinrai: that option puts sh in the path and is source to a good 90% of the techsupport I have to do as they break MinGW
20:02 GodGinrai grep doesn't have scrolling
20:02 Darcidride joined #git
20:02 ESphynx I meant 'git log' and the lik
20:03 mischat joined #git
20:04 GodGinrai we weren't discussing `git log`, but git stash
20:04 Tuha joined #git
20:04 ESphynx copied grep to my path :P grep loses all the pretty colors :P
20:05 GodGinrai function over form ;)
20:05 GodGinrai there may be a way to keep the colors, but I don't care enough to delve into that
20:05 ESphynx my brain takes 10x more times to parse a diff without colors :P
20:06 ESphynx but thanks...
20:06 ESphynx my main point though was to manifest my wish for path options in git stash show / list :P
20:06 boombatower_ joined #git
20:07 ESphynx I feel sometimes that git stash is like a second class citizen
20:07 ESphynx like not being able to git stash pop when there is a conflict... why couldn't it put conflict markers just like commits
20:07 chipotle joined #git
20:08 charlesr joined #git
20:09 e14 joined #git
20:09 kadoban It does, doesn't it?
20:09 anonymuse joined #git
20:09 ESphynx kadoban: last I checked it refuses to apply/pop :)
20:10 Tobbi joined #git
20:10 Eugene That's because stash is a second-class citizen and you should just commit & cherry-pick instead
20:10 ESphynx which makes it very hard to pop multiple stashes which modify the same files, you need to mak a temp commit
20:10 Eugene stash actually does make commits; you can cherry-pick them by ID
20:11 kadoban AFAIK it only refuses when you have uncommited changes in the working directory, which it does with any kind of mergy thing when there's conflicts.
20:11 ESphynx Eugene: oh you can eh?
20:11 ESphynx well it would be convenient if you could just git stash pop or apply a bunch of stashes withou having to commit
20:12 potato__ joined #git
20:12 kpease joined #git
20:12 kadoban It'd be very lossy if anything at all went wrong
20:13 douglascorrea joined #git
20:13 Eugene ESphynx - yup; git show stash@{0} will tell you all about the commit it creaetes
20:14 ESphynx Eugene: well that does save a step
20:15 Eugene I never use stash; I jsut create a temp branch(see also !float) and commit whatever I have, then sort it out later.
20:15 gitinfo If you have made a change in your working directory and have NOT YET COMMITTED, you may "float" that change over to another (`git checkout oldbranch`) or new (`git checkout -b newbranch`) branch and commit it there.  If the files you changed differ between branches, the checkout will fail.  In that case, `git stash` then checkout, and `git stash apply` and go through normal conflict resolution.
20:15 Zmr joined #git
20:15 Singmyr joined #git
20:16 Eugene Its basically the same thing as a stash, but with a branch handle - way more useful
20:16 kpease joined #git
20:17 ESphynx Eugene: probably a habbit I should take...
20:17 ESphynx though with git 'git stash' is usually a 'aaaarrggghhhh...'  thing
20:17 valkyr2e joined #git
20:17 masuberu joined #git
20:18 astrofog joined #git
20:18 GodGinrai I always just `git stash; git stash branch <newbranch>`
20:19 ESphynx GodGinrai: didn't know you could do that... still learning things after 7 years of using Git daily :P
20:20 durham joined #git
20:20 potato__ good evening
20:20 GodGinrai oh yea, I love the branching command
20:20 GodGinrai so useful
20:20 mfuoy joined #git
20:21 durham joined #git
20:22 Tobbi joined #git
20:23 heroux joined #git
20:27 kpease_ joined #git
20:29 MrWoohoo joined #git
20:30 dreaddymck joined #git
20:30 onmeac joined #git
20:31 MTecknology joined #git
20:31 whomp joined #git
20:34 bremner joined #git
20:34 bremner joined #git
20:36 timvisher joined #git
20:37 rominron_ joined #git
20:39 Tobbi joined #git
20:40 ToBeCloud joined #git
20:43 ikelso joined #git
20:44 chardan joined #git
20:47 Murii joined #git
20:48 _rht joined #git
20:49 anonymuse joined #git
20:51 aidalgol joined #git
20:52 leah2 joined #git
20:53 chachasmooth joined #git
20:53 douglascorrea joined #git
20:54 lvh joined #git
20:54 gugah joined #git
20:54 t0by joined #git
20:54 t0by joined #git
20:56 Tobbi joined #git
20:56 satifant joined #git
21:01 jccn joined #git
21:01 matsaman joined #git
21:01 chachasmooth joined #git
21:02 PettanShoutaKun joined #git
21:03 Hudu joined #git
21:05 PettanShouta-Kun joined #git
21:05 msonntag joined #git
21:05 j0hnny_P joined #git
21:05 dsdeiz joined #git
21:05 dsdeiz joined #git
21:06 njbair joined #git
21:07 miczac joined #git
21:08 druonysus_ joined #git
21:08 EliasH joined #git
21:10 pandeiro joined #git
21:10 SwiftMatt joined #git
21:10 fieldmapper joined #git
21:11 xissburg is mercurial a copycat of git?
21:12 e14 joined #git
21:12 bremner probably not, why not ask them?
21:12 osse no, development started around the same time
21:13 kadoban IIRC both git and mercurial started because of the same event (bitkeeper being douches)
21:13 xissburg it's too similar (?)
21:13 kadoban So they have shared history, and I doubt either was created in a vacuum.
21:14 xissburg who needs mercurial when you have git? ^^
21:14 kadoban I'm sure they ask a similar question.
21:14 bremner xissburg: now your just trolling, and trolling the wrong people.
21:15 bremner *you're
21:15 xissburg haha no
21:16 msonntag joined #git
21:17 cdg joined #git
21:17 matsaman the basic idea of git seems great to me
21:18 matsaman some implementation details bother me
21:18 Gloomy joined #git
21:21 timvisher joined #git
21:25 tinanoo joined #git
21:26 Bizkit joined #git
21:30 bgerber joined #git
21:31 jstein_ joined #git
21:31 frobnic joined #git
21:34 tterrag joined #git
21:36 chachasmooth joined #git
21:40 WifiBrain joined #git
21:40 chachasmooth_ joined #git
21:40 tang^ they were created within days of each other
21:43 pbrewczynski joined #git
21:46 Starky joined #git
21:47 douglascorrea joined #git
21:47 _geoff joined #git
21:47 douglascorrea joined #git
21:50 d^sh joined #git
21:54 donutttt joined #git
21:55 roko joined #git
21:55 roko so I did git checkout HEAD~2. How do I get back to HEAD?
21:55 mehola joined #git
21:56 chachasmooth joined #git
21:56 kadoban roko: You likely want to checkout whatever branch you were on. 'git checkout somebranch'
21:57 kadoban if you're not sure what branch, there's ways to find out, but that's likely not too sane so I'll skip that
21:57 roko kadoban: Thanks! worked
21:57 kadoban Cool
21:58 roko Also how can I check the diff between 2 commits for a single file?
21:58 EliasH left #git
21:59 fatalhalt joined #git
22:00 matsaman joined #git
22:00 donutttt If I've got a forked repo, cloned locally. There have been commits from others since I cloned the master repo, I wonder, what's the best way to get the updates from the remote repo?
22:01 canuk joined #git
22:01 ash_workz joined #git
22:01 kadoban donutttt: Add upstream as a remote, fetch (probably with --all), ... do whatever you want to modify actual branches, probably a merge or rebase.
22:02 kadoban roko: Should be 'git diff someref someother -- somepath/to/somefile'
22:02 ispn joined #git
22:02 timvisher joined #git
22:02 donutttt kadoban: awesome, thanks for that, so the upstream to add, that should be the original repo from which mine was forked?
22:03 kadoban Yes, sounds right
22:04 kadoban Wherever you want to get "updates" from really though.
22:04 mentazoom joined #git
22:04 kadoban git doesn't care, as long as there is shared history it makes sense. Hell even if there isn't git won't care, but it may make less sense in practice.
22:05 donutttt ah awesome, thanks, really appreciate it
22:05 d^sh joined #git
22:06 blackwind_123 joined #git
22:07 chardan joined #git
22:07 Zamyatin joined #git
22:07 cdg joined #git
22:08 robattila256 joined #git
22:09 rominronin joined #git
22:11 Starky joined #git
22:15 le_melomane joined #git
22:15 Limes_ joined #git
22:16 Starky joined #git
22:17 ShalokShalom joined #git
22:17 Guest36461 joined #git
22:17 tvw joined #git
22:17 Samual joined #git
22:18 cr34ton joined #git
22:18 Atm0spher1c joined #git
22:21 sballert joined #git
22:23 GodGinrai joined #git
22:24 ShalokShalom joined #git
22:25 Limes__ joined #git
22:26 nsh joined #git
22:29 sballert joined #git
22:30 sballert joined #git
22:30 timvisher joined #git
22:32 MTecknology left #git
22:32 sballert joined #git
22:32 askb joined #git
22:32 Gsham joined #git
22:33 ResidentBiscuit joined #git
22:33 netj joined #git
22:35 pity joined #git
22:40 AlgorithmWizard left #git
22:42 Puffball joined #git
22:45 oven joined #git
22:45 Starky joined #git
22:48 mischat joined #git
22:48 raspado joined #git
22:49 jeffreylevesque joined #git
22:50 Naan joined #git
22:52 ESphynx left #git
22:55 megamnrld joined #git
22:55 jwmann joined #git
22:56 Samual joined #git
22:56 Rodya_ joined #git
23:00 duderonomy joined #git
23:01 fmeerkoetter joined #git
23:02 Bizkit joined #git
23:02 timvisher joined #git
23:02 Dougie187 left #git
23:03 ikelso joined #git
23:03 ikelso joined #git
23:06 Gsham joined #git
23:08 linuxmodder is there a easily to read guide on git config --global || --local  options
23:09 linuxmodder man git config seems to be lean
23:09 gitinfo the git-config manpage is available at http://jk.gs/git-config.html
23:09 jantje joined #git
23:09 linuxmodder seems I was using git config not git-config in the manpage invocation thanks
23:14 linuxmodder anyone able to explain the need / difference of gitsshcommand if remote | origin is git+ssh:// or over ssh using user 'git'
23:15 kadoban What?
23:18 netj joined #git
23:22 lvh left #git
23:23 jimi_ joined #git
23:24 PettanShoutaKun joined #git
23:25 fieldmapper joined #git
23:26 Gsham joined #git
23:28 publio joined #git
23:28 nettoweb joined #git
23:30 jantje joined #git
23:31 Gsham joined #git
23:33 kyan joined #git
23:33 dsdeiz joined #git
23:35 mikecmpbll joined #git
23:36 fieldmapper joined #git
23:36 NeverTired joined #git
23:37 cbreak haven't seen git+ssh in years
23:38 dsdeiz joined #git
23:38 dsdeiz joined #git
23:40 dzho linuxmodder: what's "gitsshcommand"?
23:40 suttonwilliamd joined #git
23:41 cyan__ joined #git
23:41 rominronin joined #git
23:41 linuxmodder dzho,  pretty much my question $GIT_SSH seems sufficient not sure why you'd need $GIT_SSH_COMMAND
23:42 nowhere_man joined #git
23:43 guest3456 left #git
23:43 dzho $GIT_SSH_COMMAND takes precedence over $GIT_SSH, and is interpreted by the shell, which allows additional
23:43 dzho arguments to be included.
23:44 dzho so, if you wanted to pass a port number or private key without having to wrap it, seems to me
23:44 dzho linuxmodder: just for example
23:45 Derperperd joined #git
23:47 linuxmodder wouldn't one jsut set up a gpgtunnel | sshtunnel first tho adn use that for all said git?
23:47 shinnya joined #git
23:48 * dzho shrugs
23:48 dzho I just set up my ssh remotes as remotes with "git remote add" and have an agent handle the keys
23:49 dzho I don't know why I'd bother setting up a tunnel separately, but I suppose if I were scripting this maybe I'd want to be able to pass in ssh -p 6666 -i /path/to/my/key via the longer option?
23:50 timvisher joined #git
23:51 Eugene xissburg - the basic core of git was knocked out over a weekend by linus. Mercurial was invented independently a few days later, unaware. Both use an architecture(object-model) inspired by BitKeeper, and you can pretty freely convert a history between them.
23:52 rivarun joined #git
23:52 justanotheruser joined #git
23:53 xissburg so git is a copycat of bitkeeper
23:53 Eugene Only in the sense that Ford is a copycat of Mercedes-Benz for having 4 wheels and an engine, yes
23:54 xissburg heh
23:54 dendazen joined #git
23:54 Eugene Linus wrote a tool that fit his use case, and we're all gits for following-along
23:54 martin290 joined #git
23:55 dzho wrote the tool, and then got someone else to maintain, even
23:55 dzho you gotta admit, the man has style
23:55 xissburg Linus is a genius
23:55 xissburg and a dick
23:55 martin290 hey everyone, i'm a  total noob to git. i removed a folder by accident, and now i'm trying to rollback (revert) my changes, and ignore the file i want to delete (my end goal is to NOT push a folder to git)
23:55 xissburg Christ, people!
23:55 roelmonnens joined #git
23:56 martin290 but i keep getting errors when i try to revert my changes -- it says it will overwrite files, which is fine with me
23:56 denete_ joined #git
23:57 redhedded1 joined #git
23:58 mizu_no_oto joined #git
23:58 Guest55 joined #git

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