Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-08-26

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

All times shown according to UTC.

Time Nick Message
00:00 durham joined #git
00:00 energizer through a proxy
00:01 xaa joined #git
00:06 simulacr why branch name cannot start with '-' char?
00:06 bremner are you sure? is it written in the docs?
00:07 bremner hard to get the command line right, in any case
00:07 bremner huh. seems you're right.
00:08 simulacr interesting result with only one '-', depends on shell maybe
00:10 sangy Hello everyone, how could I try to find a specific line in a file in all commits of all branches? sounds like I could script it, but maybe there's a one-liner that does the trick
00:10 wxg joined #git
00:10 sangy i.e., I'm searching for pillow=* in my requirements.txt in all branches. Right now it is not in master and a dev claims it was merged a while ago
00:11 pijiu joined #git
00:12 kadoban sangy: Did you look at the -S and -G options for 'git log'? Those may be what you're looking for, possibly.
00:12 bremner simulacr: fwiw, "git check-ref-format refs/heads/-foo" succeeds, so in principle it should be allowed
00:12 sangy kadoban: let me take a look. Thanks!
00:13 pijiu joined #git
00:13 qsx '--'?
00:14 energizer qsx http://unix.stackexchange.com/quest​ions/11376/what-does-double-dash-me​an-also-known-as-bare-double-dash
00:14 simulacr qsx: seems "git checkout -b" doesn't support '--'
00:14 kadoban bremner: I doubt it's a good idea at all though
00:14 crayon joined #git
00:15 qsx simulacr: well, it’s part of an option …
00:15 qsx the part after -- would be files
00:15 bremner kadoban: oh sure, it's a lousy idea. But so are lots of things people want to do with git ;)
00:16 energizer So my git clone keeps hanging. Is there another way I can get the repo from a server and still use it like normal?
00:16 kadoban Hehe, yeah, fair enough.
00:16 kadoban energizer: You could possibly use rsync or whatever and then change it from bare to non-bare if necessary I guess.
00:17 simulacr qsx: in my case: git checkout -b -- "${name}" "${sha1}" ->  fatal: '--' is not a valid branch name.
00:17 qsx lol.
00:17 qsx what about refs/heads/--?
00:19 Kaisyu joined #git
00:20 rominronin joined #git
00:21 Hink joined #git
00:21 nettoweb joined #git
00:22 d5sx43 joined #git
00:22 jangeador joined #git
00:22 simulacr it works
00:22 ojacobson man git-check-ref-format
00:22 gitinfo the git-check-ref-format manpage is available at http://jk.gs/git-check-ref-format.html
00:22 ojacobson a branch named dash-dash looks valid to me
00:22 bremner ojacobson: not a complete answer here, since it doesn't mention -
00:22 bremner right
00:23 bremner but the porcelain is confused/fussy
00:23 ojacobson buggy porcelain? that _never_ happens ;)
00:24 ojacobson (I dunno if buggy, or if just awkward)
00:24 bremner well. the error message is kindof a lie
00:24 ojacobson why do you want, or have, a branch named --, anyways?
00:24 HoierM joined #git
00:26 energizer looks like rsync hangs on something in .git/objects/41. Is there anything I can do about that?
00:26 energizer How do I find out what that is?
00:27 simulacr ojacobson: don't want, just some sort of escaping for shell script
00:27 ojacobson 100% it's a large loose object. inspect that dir with`ls -l` on the source repo.
00:27 ojacobson simulacr: neato
00:28 mingrammer joined #git
00:29 ResidentBiscuit joined #git
00:29 dodobrain joined #git
00:29 mikeh joined #git
00:30 Gsham joined #git
00:32 andrewSC_ joined #git
00:32 constant joined #git
00:35 SwiftMatt joined #git
00:36 danslo joined #git
00:38 energizer ojacobson: it's 63M, indeed
00:39 energizer thats big, right?
00:40 ToxicFrog joined #git
00:40 jangeador joined #git
00:40 energizer how can I find out what file it is?
00:40 byteflame joined #git
00:41 tilerendering joined #git
00:45 JamesBaxter joined #git
00:46 red_ joined #git
00:51 alexsystemf__ joined #git
00:51 fakenerd joined #git
00:53 justanotheruser joined #git
00:56 fmcgeough joined #git
01:02 chrisgeorge joined #git
01:02 zmachine joined #git
01:03 anonymuse joined #git
01:04 nettoweb joined #git
01:05 ochorocho__ joined #git
01:07 clemf joined #git
01:11 wxg joined #git
01:13 crayon joined #git
01:13 crayon joined #git
01:13 Manuel15 joined #git
01:13 Manuel15 joined #git
01:13 ytl joined #git
01:15 rehat joined #git
01:16 clemf joined #git
01:17 rehat I'm trying to setup git on my remote web server so I can push updates for my web app.  I am following a tutorial that had me do a init --bare but I guess I can't change branches.  Is it possible to push to a remote clone of my git repository?
01:17 NeXTSUN joined #git
01:17 inflames joined #git
01:18 rehat or do I need to create two separate init --bare folders, one for my live web app and one for developement?
01:19 bronson joined #git
01:21 kadoban rehat: There's two questions intertwined there. Figure out development of the thing, using git for version control and figure out how you want that set up. And figure out deployment, which if you must you can tie into git (though I find is often a terrible idea)
01:22 rominronin joined #git
01:22 nettoweb joined #git
01:22 jangeador joined #git
01:22 kadoban !deploy has some stuff for the second part. Personally I 95% of the time just go with: write an 'rsync' command in a send_it.sh file, and keep it in the repo, run it when I feel like.
01:22 gitinfo Git is not a deployment tool, but you can build one around it (in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/deploy.html
01:25 rehat kadoban: thanks, would the .sh script just do a git clone?  Like what is a modern way to get my local project on to a remote server.  Maybe I am in the wrong room for that haha
01:26 mizu_no_oto joined #git
01:26 rehat I think heroku got me on the wrong path now that I am using DigitalOcean
01:26 kadoban rehat: No, like I said rsync (or whatever else) usually. Whatever deployment tool you like. Git ... isn't really a deployment tool.
01:27 kadoban People use it as one because ... I dunno why.
01:27 malesca joined #git
01:27 rehat ahh just googled rsync ok I will look into that thanks
01:27 phroa So the things you want here are "place to host code" and "deployment script", right?
01:28 malber joined #git
01:28 kadoban For place to host a git repo, there's other stuff to talk about. gitolite is nice, you can just use github or bitbucket if you want, there's a few others.
01:28 KnownSyntax_ joined #git
01:29 rehat nah not looking for git repo hosting, mainly getting my web app onto a webserver.  Totally new to web dev
01:31 constant joined #git
01:32 ertes honestly git is probably a better deployment tool than rsync
01:33 Aces_Charles joined #git
01:36 LeBlaaanc joined #git
01:36 arescorpio joined #git
01:37 danslo joined #git
01:38 AnarchyAo joined #git
01:40 HoierM joined #git
01:40 Mixologic joined #git
01:41 hahuang61 joined #git
01:46 fakenerd joined #git
01:46 durham_ joined #git
01:46 Lyqyd joined #git
01:46 evanwang joined #git
01:47 cqi joined #git
01:52 pks joined #git
01:53 cdown joined #git
01:55 jhass joined #git
01:57 beyertech joined #git
01:57 vassagus joined #git
01:58 Vampire0_ joined #git
01:59 ASOLAr joined #git
01:59 d^sh joined #git
02:00 theverbg_ joined #git
02:02 birdup joined #git
02:07 bronson joined #git
02:08 harish joined #git
02:09 rossome joined #git
02:15 DJ_Spies joined #git
02:17 elementalest joined #git
02:17 elementalest joined #git
02:17 mizu_no_oto joined #git
02:20 rubyonrailed joined #git
02:22 rominronin joined #git
02:24 lagothrix joined #git
02:25 Goplat joined #git
02:25 realz joined #git
02:26 hexagoxel joined #git
02:29 xaviergmail joined #git
02:29 akkad joined #git
02:30 cyphase joined #git
02:30 d1b joined #git
02:32 clemf joined #git
02:33 andrewSC_ joined #git
02:36 kisonecat joined #git
02:37 beyertech joined #git
02:37 annoymouse4210 joined #git
02:37 danslo joined #git
02:38 annoymouse4210 Is it possible to merge 3 commit in test branch to master branch with different commit log?
02:38 beyertech joined #git
02:40 nwkj86 joined #git
02:40 beyertech joined #git
02:40 UserOO7 joined #git
02:40 phroa annoymouse4210: you can !fixup them before merging or you can !rebase them, though I'm not sure what you mean by different
02:40 gitinfo annoymouse4210: So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!
02:40 gitinfo annoymouse4210: 'git rebase' takes away your local commits, updates your branch with new stuff from <upstream> (argument), then re-applies your local commits on top. This makes it look like your commits were created "after" the new stuff, and it can look cleaner than doing a !merge. Beware of !rewriting_public_history, though. Not to be confused with !interactive_rebase.
02:40 annoymouse4210 thanks for quick answer phroa
02:43 chachasmooth joined #git
02:43 qqqqq joined #git
02:46 fstd_ joined #git
02:48 dsdeiz joined #git
02:49 MattMaker joined #git
02:50 rehat joined #git
02:50 pijiu joined #git
02:50 SegFaultAX joined #git
02:51 elementalest joined #git
02:51 elementalest joined #git
02:52 akkad joined #git
02:53 Gsham joined #git
02:53 dsdeiz joined #git
02:53 dsdeiz joined #git
02:54 msm joined #git
02:54 rewt joined #git
02:55 sparr joined #git
02:55 [cust] joined #git
02:56 ixian joined #git
02:56 MattE joined #git
02:56 bluj joined #git
02:57 bittin joined #git
02:57 g5pw joined #git
02:57 yoh joined #git
02:59 daurnimator joined #git
03:00 sandstrom joined #git
03:01 aavrug joined #git
03:02 aavrug joined #git
03:02 tHEMtZ joined #git
03:02 ytl_ joined #git
03:03 Karpuragauram joined #git
03:03 benbergman joined #git
03:03 c0ded joined #git
03:04 eggoez joined #git
03:04 SporkWitch joined #git
03:06 Tsutsukakushi joined #git
03:06 kirb joined #git
03:06 elementalest joined #git
03:06 fullstack joined #git
03:06 JCaesar joined #git
03:06 stompleee joined #git
03:07 red_ joined #git
03:07 ochorocho__ joined #git
03:07 tjone270_ joined #git
03:09 kirb joined #git
03:09 marienz joined #git
03:10 chamunks joined #git
03:11 Darren_ joined #git
03:11 justanotheruser joined #git
03:13 motivic joined #git
03:13 iota_pi joined #git
03:13 Electrometro joined #git
03:16 alexsystemf__ joined #git
03:18 bronson joined #git
03:19 navidr joined #git
03:22 elementalest joined #git
03:22 elementalest joined #git
03:23 dreamyspell joined #git
03:23 rubyonrailed joined #git
03:24 rominronin joined #git
03:26 aidalgol joined #git
03:26 griffindy joined #git
03:29 watabou joined #git
03:31 nwkj86 joined #git
03:32 absk007 joined #git
03:33 mehola joined #git
03:34 MattMaker joined #git
03:36 mehola joined #git
03:38 danslo joined #git
03:41 hahuang61 joined #git
03:42 pioiytr joined #git
03:43 nwkj86 joined #git
03:44 ash_workz joined #git
03:50 Atrumx joined #git
03:56 absk007 joined #git
03:56 jangeador joined #git
03:57 krillr joined #git
03:58 Rodya_ joined #git
03:59 MattMaker joined #git
04:00 vassagus joined #git
04:00 thecomedian joined #git
04:03 romerocesar joined #git
04:05 thiago joined #git
04:09 watabou joined #git
04:10 snowkidind joined #git
04:11 Cabanossi joined #git
04:15 sagerdearia joined #git
04:15 _xor joined #git
04:21 rockstar___ joined #git
04:23 Klumben joined #git
04:25 rominronin joined #git
04:33 xaviergmail joined #git
04:33 andrewSC_ joined #git
04:36 a_thakur joined #git
04:39 danslo joined #git
04:40 LUMIA930 joined #git
04:44 cyan__ joined #git
04:46 [cust] joined #git
04:46 romerocesar joined #git
04:47 nidr0x joined #git
04:47 subhojit777 joined #git
04:49 MattMaker joined #git
04:52 realz joined #git
04:52 danimal2 joined #git
04:52 jab416171 joined #git
04:53 a_thakur joined #git
04:59 MattMaker joined #git
05:00 mingrammer joined #git
05:01 vassagus joined #git
05:03 freimatz joined #git
05:04 UserOO7 joined #git
05:05 MlkeiWjRyM joined #git
05:08 ochorocho__ joined #git
05:10 pijiu joined #git
05:11 wilsoncd35 joined #git
05:12 rubyonrailed joined #git
05:12 a_thakur_ joined #git
05:13 psprint_ joined #git
05:15 _sjs joined #git
05:15 lagothrix joined #git
05:16 realz joined #git
05:17 OpPRzIPZZc joined #git
05:20 bronson joined #git
05:21 cqi joined #git
05:22 a_thakur joined #git
05:24 a_thakur joined #git
05:26 rominronin joined #git
05:28 EvilPeng1 joined #git
05:28 Manuel15 joined #git
05:28 Manuel15 joined #git
05:29 MattMaker joined #git
05:30 tvw joined #git
05:31 ferr joined #git
05:33 deutor joined #git
05:35 tanuki joined #git
05:36 EvilPeng1 joined #git
05:39 tanuki joined #git
05:40 danslo joined #git
05:40 newbiee joined #git
05:41 Xizor joined #git
05:41 newbiee wondering how can i do “git diff” only for certain files?
05:41 Grievre There might be a way but it might be easier to just pipe the output to filterdiff
05:41 newbiee i have lot of changes but i want to see diff only for files which matches “Kconfig”
05:42 newbiee Grievre: piping is not outputting the diff. I can see only file names.
05:42 hahuang61 joined #git
05:44 jceb joined #git
05:44 i7c newbiee, of course you can.
05:44 i7c newbiee, git diff <revisions> -- path/to/file
05:45 Grievre newbiee: ?
05:45 ThomasLocke joined #git
05:45 ThomasLocke joined #git
05:45 newbiee i7c: perhaps I was not clear. I have lot of changes in several files but i want to see only changes which are done in file type Kconfig
05:45 i7c Yeah, that’s what my command does.
05:45 newbiee Grievre: git diff|grep Kconfig
05:45 mangolisk joined #git
05:46 i7c git diff -- Kconfig
05:46 i7c oh
05:46 i7c file *type* you say
05:46 i7c What do you think a file type is?
05:46 phanimahesh joined #git
05:46 i7c Maybe you want git diff -- \*.Kconfig
05:47 newbiee i7c: not working.
05:47 i7c That was assuming that the files end in .Kconfig
05:47 newbiee i7c: No “Kconfig” is the full file name. There are many files like that.
05:48 i7c They're all called Kconfig?
05:48 newbiee i7c: yes
05:48 gynter left #git
05:48 Grievre newbiee: I didn't say grep, I said filterdiff
05:48 newbiee i7c: there are several directories and every directory might have Kconfig file
05:49 MattMaker joined #git
05:49 newbiee Grievre: git diff|filterdiff Kconfig ?
05:49 shaggycat joined #git
05:49 i7c git diff -- \*Kconfig then
05:49 i7c without the .
05:50 newbiee i7c: working thanks.
05:51 newbiee i7c: mind explaining that command?
05:51 i7c Well it limits the diff to the path after the --, where you actually want *Kconfig (* as in shell globbing. It expands to any path here)
05:52 rchavik joined #git
05:52 i7c Tho shells interpret * so you have to escape it.
05:52 i7c That’s why it’s \*
05:52 i7c So it’s basically limited to all paths in the diff that end in Kconfig
05:54 i7c newbiee, if you want to read more about it check out !man gitglossary there’s a section called "pathspec"
05:54 gitinfo newbiee: The git man pages are available online at http://jk.gs/git.html. Or were you looking for the "man git-foo" syntax(without the !) ?
05:54 gitinfo newbiee: the gitglossary manpage is available at http://jk.gs/gitglossary.html
05:55 asc232 joined #git
05:55 newbiee i7c: thanks for explanation. Works like a charm.
05:55 i7c :)
05:59 dreiss joined #git
05:59 konrados joined #git
06:01 sandstrom joined #git
06:01 vassagus joined #git
06:02 MrWoohoo joined #git
06:06 i7c newbiee, btw, there are many commands that work with those -- limitting/path
06:07 i7c log, show, difftool, checkout to name a few
06:07 newbiee i7c: yes reading that documentation.
06:08 cpg left #git
06:08 newbiee i7c: git diff \*Kconfig
06:08 newbiee i7c: this is also doing the same
06:09 i7c newbiee, yup, if you’re lucky. ;)
06:09 i7c I mean yeah it probably always works with this specific path.
06:09 ochorocho__ joined #git
06:09 i7c newbiee, sometimes there are ambiguities between paths and refs (branchnames, tags).
06:09 i7c Say you have a file called master and you do stuff like git diff master
06:10 i7c In that case the -- really has an effect in that it tells git that master is a file: git diff -- master
06:10 Hink joined #git
06:11 fusionx8_ joined #git
06:11 jknetl joined #git
06:14 MattMaker joined #git
06:14 newbiee i7c: understood. — basically means that the text following it is a file. Without it, it means it is a branchname or tag.
06:14 alaing joined #git
06:14 alaing good morning
06:14 i7c Yup.
06:15 i7c Just that git supports lazyasses and tries using it as a path if it didn’t match any branch/tag.
06:15 roelmonnens joined #git
06:15 i7c Without the --.
06:15 watabou joined #git
06:15 _sjs joined #git
06:16 Sapphomo joined #git
06:16 Repox joined #git
06:17 a_thakur joined #git
06:17 Mogget joined #git
06:19 alaing Hi, I have a repo which I have created three branches, (master, one-page, triage), one-page and triage are just different user journeys through the app. for the time being master will be inline with one-page but eventually I'll need to switch it over to triage. Because they are two completely different codebases (i,e one will have more functionality/web pages etc) i dont think I would be able to merge triage into master. would this b
06:19 alaing e a rebase? or simply delete master branch and create a new master based on triage?
06:19 fakenerd joined #git
06:20 newbiee i7c: wondeirng how can i add one more kind of file along with Kconfig. I want to do git diff for Kconfig and defconfig
06:20 red_ joined #git
06:21 i7c newbiee, exactly how one would expect. ;) -- \*Kconfig \*defconfig
06:22 newbiee i7c: git diff -- \*Kconfig \*defconfig
06:22 i7c mhm yup
06:23 corpsicle joined #git
06:24 i7c alaing, one-page and triage are different branches and master currently is identical to one-page?
06:25 ollien joined #git
06:25 alaing ic7 yes thats right
06:25 alaing triage is a spike to test a theory on user journey and will be developed over time
06:26 moneylotion joined #git
06:26 ProbabilityMoon joined #git
06:26 alaing and will eventually completely replace one-page
06:26 i7c Ok so master is just another name for one-page right now and eventually you want it to be just another name for triage.
06:27 rominronin joined #git
06:27 i7c So, that would just be a reset of master (same as "deleting and recreating master").
06:27 i7c It has some implications tho.
06:28 i7c Anybode who builds stuff based on master will be in hell after the switch.
06:28 Macaveli joined #git
06:28 Hounddog joined #git
06:29 i7c alaing, may I ask what’s the sense of that master branch at the moment?
06:30 alaing i7c: oh I see, you make a good point about people building stuff based on master.
06:31 tanuki joined #git
06:32 i7c I might add that this is not really an issue of git but it kind of is in the nature of the situation.
06:32 i7c If you replace your code base, work based on the old code base is somehow in vain.
06:32 satya4ever joined #git
06:32 nvngpt31 joined #git
06:34 JeroenT joined #git
06:34 andrewSC_ joined #git
06:34 i7c alaing, commits that would make sense on both branches can possibly be cherry-picked. Also note that there are merge options where you can say "if in conflict, always take the version of this branch".
06:35 i7c Which is a bit gentler than a full replacement.
06:38 leeN joined #git
06:38 alaing i7c: I was hoping 'master' would be the most stable branch which would be deployed, devs would branch off one-page and merge their work back into it before it gets merged into master
06:39 alaing i7c I was thinking about cherry picking commits between one-page and triage if it affected both branches
06:40 t0by joined #git
06:40 danslo joined #git
06:41 i7c Makes sense. And you can always try a merge. If it goes wrong, abort it. :D
06:41 alaing i7c is there a way of saying merge triage into master but if triage doesn't have the file (delete it) or theres a conflict then use triage
06:41 z8z joined #git
06:42 alaing i haven't used advanced merge options before
06:43 _sjs joined #git
06:43 j416 alaing: fwiw, you don't need to delete a branch to make it point to some other commit
06:43 j416 alaing: just use git-reset
06:44 MattMaker joined #git
06:44 submitnine joined #git
06:46 rkazak joined #git
06:46 stuh84 joined #git
06:48 mingrammer joined #git
06:48 caffeinatedcode joined #git
06:48 caffeinatedcode joined #git
06:51 mariuscc joined #git
06:55 rominronin joined #git
06:55 blackwind_123 joined #git
06:57 lindenle joined #git
06:58 i7c alaing, yes. There is -Xtheirs and -Xours (see man git-merge)
06:58 gitinfo alaing: the git-merge manpage is available at http://jk.gs/git-merge.html
06:59 i7c It’s described as options to the recursive merge strategy.
06:59 LUMIA930 git checkout <LATEST cartodb-postgresql tag>
06:59 LUMIA930 what has to be filled in there
06:59 j416 joined #git
07:00 LUMIA930 tag : represents last commit ?
07:00 alaing i7c thank you so much for your help. i need to get off the train but i'll read that a bit later
07:00 beyertech joined #git
07:00 LUMIA930 and should i use <> too in command ?
07:01 i7c No!
07:01 LUMIA930 git checkout <LATEST cartodb-postgresql 9ee3125>
07:01 i7c < and > will be interpreted by the shell and do funny things.
07:01 LUMIA930 git checkout LATEST cartodb-postgresql 9ee3125
07:01 i7c You might wanna check git tag  to look for the right tag.
07:01 LUMIA930 so the above is the correct one ?
07:02 LUMIA930 the latest was this : found from Git : 9ee3125
07:02 i7c Then it’s git checkout 9ee3125, if that’s what you want.
07:02 sandstrom joined #git
07:02 vassagus joined #git
07:02 caffeinatedcode joined #git
07:02 caffeinatedcode joined #git
07:03 LUMIA930 i7c: Thanks
07:03 LUMIA930 root@GRMCSVR07:~/cartodb-postgresql# git checkout 9ee3125
07:03 LUMIA930 Note: checking out '9ee3125'.
07:03 LUMIA930 You are in 'detached HEAD' state. You can look around, make experimental
07:03 LUMIA930 changes and commit them, and you can discard any commits you make in this
07:03 LUMIA930 state without impacting any branches by performing another checkout.
07:03 LUMIA930 If you want to create a new branch to retain commits you create, you may
07:03 LUMIA930 do so (now or later) by using -b with the checkout command again. Example:
07:03 LUMIA930 git checkout -b new_branch_name
07:03 LUMIA930 HEAD is now at 9ee3125... Relase 0.17.1
07:03 a_thakur joined #git
07:03 LUMIA930 That worked for me :)
07:04 i7c Don’t paste that stuff here please. Use something like this https://gist.github.com/ and paste only the link if you need to show output.
07:04 Mixologic joined #git
07:05 a_thakur joined #git
07:06 tanuki joined #git
07:06 PHPanos joined #git
07:07 deutor left #git
07:07 Darren_ joined #git
07:08 Sceorem joined #git
07:08 encod3 joined #git
07:08 energizer left #git
07:09 dreiss joined #git
07:10 ochorocho__ joined #git
07:10 Darcidride joined #git
07:10 manishs1 joined #git
07:12 TomyWork joined #git
07:12 ASOLAr joined #git
07:13 elect joined #git
07:13 manishs1 left #git
07:14 MattMaker joined #git
07:14 a_thakur joined #git
07:18 jaafar joined #git
07:21 iskander_work joined #git
07:21 ShapeShifter499 joined #git
07:22 iskander_work hello
07:22 gitinfo iskander_work: 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.
07:23 bronson joined #git
07:23 iskander_work is there a better way to create patches from commits that match a specific re than to make first a list of the commits with 'git log' and then iterate over the list and create a patch for every commit in the list ?
07:24 Sapphomo joined #git
07:24 JeroenT joined #git
07:24 a_thakur_ joined #git
07:24 iskander_work eg: git log --reverse --format=%H --grep='toaster' > commit-list
07:25 iskander_work for c in $(cat commit-list); do git format-patch -1 $c; done
07:26 osse git rev-list --reverse --grep='toaster'
07:27 osse git rev-list --reverse --grep='toaster' | while read c; do git format-patch -1 $c; done
07:27 pro_metedor joined #git
07:27 Balliad joined #git
07:28 vuoto joined #git
07:28 osse Long story short: not really
07:28 osse but you can make the shell parts prettier :p
07:28 i7c Mhmmm, toast.
07:28 constant joined #git
07:29 iskander_work osse: thanks
07:29 osse I could use some toast. Didn't have breakfast :/
07:29 i7c Ditto.
07:30 iskander_work unfortunately, it's yocto toaster :)
07:30 dreamyspell joined #git
07:30 Sapphomo joined #git
07:31 mariuscc joined #git
07:32 iskander_work off-topic: guys, are you familiar with ag, the silver searcher
07:32 iskander_work i'm having a problem with -Q and -g flags
07:32 andrewSC_ joined #git
07:32 ki0 joined #git
07:33 ASOLAr joined #git
07:34 elementalest joined #git
07:34 jast I've, uh, heard of it
07:34 alaing joined #git
07:34 alaing joined #git
07:34 ki0 joined #git
07:34 encod3 joined #git
07:34 i7c I use it because it’s shorter than typing grep.
07:35 elementalest joined #git
07:35 elementalest joined #git
07:35 iskander_work yeah, it's better for programmers
07:35 jast i7c: don't like aliases? :}
07:35 i7c It also makes beautiful colors.
07:35 i7c jast, I was half-kidding. And the colors!
07:35 elementalest joined #git
07:35 elementalest joined #git
07:36 jast grep colourizes, too, if you let it
07:36 elementalest joined #git
07:37 elementalest joined #git
07:37 i7c But but but :'s
07:38 osse do u even git grep
07:38 Hink_ joined #git
07:38 elementalest joined #git
07:38 elementalest joined #git
07:39 NeverDie_ joined #git
07:41 tanuki joined #git
07:41 danslo joined #git
07:42 Joost osse: TIL about git-grep
07:42 Joost thanks
07:42 a_thakur joined #git
07:42 mingrammer joined #git
07:42 JeroenT joined #git
07:43 hahuang61 joined #git
07:43 ahmedelgabri joined #git
07:43 elementalest joined #git
07:43 elementalest joined #git
07:44 MattMaker joined #git
07:49 circ-user-cmOBz joined #git
07:49 Andrew_K joined #git
07:49 Andrew_K joined #git
07:51 psacrifice joined #git
07:52 ImJune joined #git
07:53 vuoto joined #git
07:55 BlackMaria joined #git
07:55 Hello71 joined #git
07:55 BrianBlaze420 joined #git
07:57 yazdmich joined #git
07:58 cqi joined #git
08:00 rubyonrailed joined #git
08:03 vassagus joined #git
08:03 osse Joost: de nada. Add --cached for even more speed
08:03 osse I suppose ag effectively does the git grep thing since it respects .gitignore though
08:04 Joost these things often make me wonder how mnay git tools I'm unaware off
08:04 Joost of*
08:04 KeelOfSteel joined #git
08:04 Joost maybe I should read a manpage
08:04 Hobbyboy joined #git
08:05 Hounddog during initial upload of our project i apparently uploaded some plugin which had a .git file inside. Now its there as a submodule. What would be the cleanest way to remove this and add the actual files?
08:06 marcm joined #git
08:06 marcogmonteiro joined #git
08:09 xaviergmail joined #git
08:10 KeelOfSteel joined #git
08:10 maxorator joined #git
08:11 ASOLAr joined #git
08:12 GavinMagnus joined #git
08:14 Rish joined #git
08:15 JeroenT joined #git
08:16 _ikke_ Hounddog: git rm --cached path/to/dir
08:16 _ikke_ remove the .git file / folder and add again
08:16 Hounddog _ikke_: thats what i did now did a commit with removed and added and commit --ammend
08:17 Quaeram joined #git
08:19 Rish joined #git
08:19 mikecmpbll joined #git
08:20 freimatz joined #git
08:22 constant joined #git
08:23 matumbo Found the reason why commit.gpgsign had no effect for me. Turned out that it was not yet implemented in my git version from the repo. I wrongly assumed that I would get some kind of warning when using an unsupported config option
08:24 osse ugh
08:24 osse matumbo: yeah you can do 'git config --global lol.rofl topkek' and git won't say a peep
08:24 osse but I assume git commit -S would have complained
08:25 matumbo good to know :) git commit -S -m "perfect comment" . signed the commit with no problems
08:29 warthog9 joined #git
08:31 wrouesnel joined #git
08:34 MattMaker joined #git
08:40 DevAntoine joined #git
08:40 DevAntoine hi
08:40 padarom joined #git
08:40 ertesx joined #git
08:41 DevAntoine I've done a git stash pop on an old branch. I've got conflicts, I just want to undo this stash pop to delete this old branch. The popped file are not committed. How can I do that?
08:42 nwkj86 joined #git
08:43 spudowiar joined #git
08:44 MattMaker joined #git
08:44 ASOLAr joined #git
08:46 dreiss joined #git
08:46 osse DevAntoine: do undo the stash you can simpyl nuke the changes
08:46 osse git checkout HEAD -- . or reser --hard
08:46 osse reset
08:47 DevAntoine osse: what about the popped files not comitted?
08:48 padarom joined #git
08:48 osse nothing
08:48 padarom Hey there, I have a conceptual question about my Git workflow. I'm developing an app that we have one basic version of and make modifications and publish it in a version for customers. The main development is in the master branch, each customer's customizations is within a "customer-foo", "customer-bar", ... branch. There are some files in the master branch, that I occasionally edit in there, but I don't want
08:48 padarom these changes to override what I did in the customer branches when merging. What would be the best way to do that?
08:48 DevAntoine osse: they won't be erased?
08:48 osse DevAntoine: i don't think so. not sure
08:49 _ikke_ padarom: imo, branches are not a good way to distinguish between features
08:52 padarom _ikke_: These modifications mainly are configuration differences, like different images or a different app name. When creating a customer version of the app, I so far just create a new branch, modify three files and that's it. I wouldn't usually need to touch these anymore, just merge more features developed in master back into these customer branches
08:53 achlys joined #git
08:54 _ikke_ Right, but it's the merge part that's the problem
08:54 _ikke_ git merge merges complete trees
08:54 _ikke_ it has no ignore option
08:54 MattMaker joined #git
08:55 padarom _ikke_: Okay, that's what I wanted to find out :) Do you have an idea of how I could do it differently?
08:55 cbreak merging into those "fork" branches should be no problem
08:55 xaa joined #git
08:55 cbreak merging them into something else would be a problem
08:55 iskander_work _ikke_: what would be a better approach to manage features ? i've got the same problem with different features for different customers
08:55 cbreak you can also consider cherry-picking, or not messing around in such an amateurish way in the first place
08:56 cbreak create the customer specific parts in a different repository and use yours as submodule
08:56 cbreak or completely separate
08:56 _ikke_ cbreak: agree
08:56 _ikke_ These are ideally not part of the project source
08:56 cbreak they could be untracked I guess
08:56 _ikke_ Just a 'customer package'
08:56 romerocesar joined #git
08:57 ASOLAr joined #git
08:57 padarom The problem is that some of these files are needed for the app to function properly. I can't develop in the master branch without having them, as I won't be able to successfully compile and test my changes
08:58 padarom Submoduling could be an option, I haven't considered that before
08:58 cbreak padarom: or have a default version of the file
08:58 cbreak and if some other version is present, the default isn't used
08:59 cbreak if you use submodules, I'd have your code be the submodule of the user code, as if it was a library
08:59 rchavik joined #git
08:59 cbreak that way, user code always gets the version it was developed for, not the newest version, with which it might or might not be compatible
09:00 achlys joined #git
09:00 Torrone joined #git
09:01 courrier joined #git
09:01 harish__ joined #git
09:01 padarom I considered something like default versions before, but I don't know if the iOS app structure allows for that. I'll do some research on it and also see if submodules could work in this case
09:01 padarom Thanks so far
09:01 diogenese joined #git
09:03 ImJune joined #git
09:04 achlys joined #git
09:04 vassagus joined #git
09:04 MattMaker joined #git
09:05 pkpafh joined #git
09:07 kirb joined #git
09:07 kirb joined #git
09:09 jost joined #git
09:11 ochorocho__ joined #git
09:12 JeroenT joined #git
09:13 matumbo If there exists one, where can I find the git.spec file ?
09:16 chandu20 joined #git
09:17 Seveas matumbo: in the .src.rpm of your distribution
09:18 Seveas there used to be a git.spec in git.git, but it was unmaintained for years. It was removed earlier this year.
09:18 osse Seveas: jeff King said test_config had problems, so you got pwned
09:19 Seveas osse: not as much as you, your whole patch got dropped :P
09:20 osse :(
09:20 mingrammer joined #git
09:21 matumbo Seveas: Okey, I downloaded the latest rpm from rpmfind.net for git and tried unpacking it but I only find a usr and etc folder inside with to spec file
09:21 cbreak what's a git.spec file for?
09:21 Seveas matumbo: that's the binary rpm, not the source rpm. And don't use rpmfind.net, use your distro's archive.
09:21 Seveas cbreak: for building rpm's
09:21 matumbo oh okey
09:21 matumbo thanks
09:22 fiddlerwoaroof joined #git
09:22 * osse used rpmfind at his previous job and brutally installed all sorts of crap
09:22 osse good times
09:23 * Seveas slaps osse
09:23 netj joined #git
09:24 osse Seveas: they asked for it when they made developers use a desktop subscription thingy instead of a workstation subscription thingy and had some sort of local copy of the workstation repo with outdated packages in it
09:24 osse there was something fishy going on there
09:24 MattMaker joined #git
09:24 Seveas sounds like a dodgy RHEL setup
09:24 osse And /usr/local was the mount point for an NFS share...
09:25 osse Not sure why we used RHEL to begin with.
09:26 chandu20 joined #git
09:26 ichux_ joined #git
09:26 Seveas we don't force devs to use a standard os
09:27 osse A colleague added some centos repos and stuff. One day he got prompted about installing a huge amount of updates. That was the day he accidentally stopped using RHEL 5 and started using CentOS 6 :P
09:27 Seveas awesome
09:28 matumbo Where would I find the source rpms for centos7 ?
09:29 steven joined #git
09:29 osse Seveas: the basic idea was sound. We were doing embedded stuff and this was a VM. So it was an easy way of giving all developers a standard environment with everything they needed. But then why use a NFS share? And why RHEL? And why auth with some central server?
09:29 steven if I run git ls-remote tag I see refs/tags/1.4.3^{}
09:30 achlys joined #git
09:30 steven whats the meaning of ^{}
09:30 steven ?
09:30 osse Seveas: it never made sense to me. But then again I'm not a sysadmin/devops guy.
09:30 osse Oh, I know. Fucking ClearCase.
09:31 osse Now I remember.
09:31 notebox joined #git
09:31 Seveas steven: peeled tag, so the sha of the object the tag points to
09:32 subhojit777 joined #git
09:32 Seveas for lightweight tags it's the same, for annotated tags it's the object the tag points to instead of the tag object
09:32 steven but the {} is empty?
09:32 grawity sure; it doesn't have to have anything inside
09:32 grawity if you specify a type, it peels until it finds that type
09:32 steven I see.
09:33 grawity if you don't, it peels until it finds a non-tag
09:33 wxg joined #git
09:33 Seveas matumbo: yum install yum-utils, yumdownloader --source git
09:34 matumbo Seveas: thanks, found that vault.centos.org seems to provide source rpms
09:34 Seveas vault is for obsolete versions
09:35 matumbo It has the latest version that is available in the standard repo
09:35 star_prone joined #git
09:37 tvw joined #git
09:38 matumbo anyhow, seems like it might be to much work to be worth it to convert the latest git source code to a rpm
09:38 Seveas matumbo: if you want that, you can simply grab the centos 7 packages from https://build.opensuse.org/pro​ject/show/home:seveas:git-next
09:39 Seveas that's always updated with the latest git 'next' branch
09:39 michele joined #git
09:39 michele hey!
09:39 matumbo anything similar for the latest stable version ?
09:40 achlys joined #git
09:40 matumbo or which version is in the next branch?
09:40 michele how can I see which commit produced the diff in a file I am viewing the diff between two branches? e.g.: git diff branchB -- myFile < which commit produced this difference?
09:40 matumbo nvm, assuming it would be the next :P
09:42 danslo_ joined #git
09:44 hahuang61 joined #git
09:46 mikeh_ joined #git
09:47 robotroll joined #git
09:48 cbreak michele: git log branchA...branchB
09:49 rubyonrailed joined #git
09:49 cbreak michele: you can further limit this with -- filename
09:49 cbreak and get per-commit patches with -p
09:49 michele thanks cbreak
09:52 catsup joined #git
09:53 mikeh joined #git
09:54 MattMaker joined #git
09:57 achlys joined #git
09:59 mikeh_ joined #git
10:03 eoli3n joined #git
10:03 eoli3n hi
10:03 eoli3n i want to do my first pull request, but i dont understand something
10:03 vuoto joined #git
10:03 eoli3n i forked a repo
10:03 eoli3n the git cloned it on my computer
10:03 eoli3n created a new branch
10:04 eoli3n the git add, commit my modifications
10:04 eoli3n and then i wanted on the original repo, to do a pull request on github website
10:04 eoli3n but i dont see my branch
10:04 grawity you need to actually push that new branch (with your commits) to your fork on github
10:04 achlys joined #git
10:04 eoli3n i did
10:04 eoli3n i think
10:04 eoli3n let me paste you, a sec
10:04 MattMaker joined #git
10:05 vassagus joined #git
10:05 eoli3n git push --set-upstream origin push-ansible-pull ?
10:05 eoli3n when i git pushed, it tells me to do this think
10:05 eoli3n thing
10:05 grawity that should work, I think
10:05 eoli3n push-ansible-pull is the name of my branch
10:05 mariuscc joined #git
10:05 grawity (I'd have shortened it to `git push -u ...` though)
10:05 eoli3n so on github website, on the webpage of original repo, when i go to pull request, i should see my branch ?
10:06 grawity hmm
10:06 grawity you should see it on your fork, at least
10:06 eoli3n yep, but how to submit it to the original repo ?
10:06 canton7 you should see that branch in your fork. it normally pops up on the main page of your fork, with a button to create a PR
10:06 grawity well, that's what the pull request is for.
10:06 eoli3n oh ok, i need to do this on my repo
10:06 grawity you have a branch on your fork, you PR it, and they merge it into the original repo's master
10:07 canton7 you can also go to the original repo, Pull Requests, New pull request, click 'compare across forks', select your fork, and your branch
10:07 eoli3n https://github.com/ansible​/ansible-examples/pull/166
10:07 eoli3n worked :D
10:07 grawity that's the long way around
10:08 eoli3n my first PULL REQUEST huhu
10:08 eoli3n thx
10:08 canton7 ... or go to your fork, branches, find the branch, 'new pull request'
10:08 canton7 so, 3 ways
10:08 grawity the shortest way is, if you visit the original repo soon after pushing, usually Github shows the yellow line with "recently pushed branches" and offers to PR them
10:09 canton7 yeah, that's way 1 which I described ebfreo
10:09 canton7 *before
10:09 MattMaker joined #git
10:09 grawity ah I thought it only shows up when visiting the original
10:09 canton7 ah, sorry, you said original
10:09 canton7 hadn't noticed that one
10:09 canton7 ok, 4 ways!
10:11 courrier joined #git
10:13 anticom joined #git
10:14 MattMaker joined #git
10:15 sans_s3r1f joined #git
10:15 Anticom joined #git
10:15 michele left #git
10:19 PioneerAxon joined #git
10:21 hyperair joined #git
10:22 inflames joined #git
10:23 Guest23766 joined #git
10:24 mikeh joined #git
10:24 jmd joined #git
10:25 jmd Is git rebase an atomic operation ?
10:25 bremner no
10:25 bremner well, not afaik.
10:25 _ikke_ From what standpoint
10:25 bronson joined #git
10:25 _ikke_ Moving the rebased ref can be atomic I guess?
10:25 _ikke_ It either points to the old history, or to the new history
10:25 grawity but I don't think `git rebase` actually does that
10:26 grawity it rolls back the branch, then cherry-picks commits by one
10:26 jmd from the standpoint of any processes not under control of git
10:26 i7c HEAD changes many times during a rebase.
10:26 i7c I doubt that even checkout is atomic from that point.
10:26 grawity hmm, I suppose if it detached HEAD first, then at least the branch update could become atomic
10:26 _ikke_ jmd: If you purely look at refs, it can be
10:26 grawity or maybe it does do that
10:26 _ikke_ But the rest of the repo can be in many states
10:26 grawity but atomic HEAD update would require a separate worktree
10:27 bremner _ikke_: even rebase -i?
10:27 jmd I mean looking at the working directory
10:27 grawity no, afaik that's not atomic
10:27 _ikke_ bremner: The ref is updated after the last commit has been rebased
10:27 bremner ook. I guess I was thinking of HEAD
10:27 grawity jmd: you could probably use `git worktree` to implement atomic rebases
10:27 padarom exit
10:27 _ikke_ yeah, HEAD moves all over
10:28 i7c I guess you can even catch the working directory in a half updated state when you perform checkout?
10:28 _ikke_ So that's why I was asking from what point of view
10:28 _ikke_ i7c: correct
10:28 alansaul joined #git
10:28 i7c So the answer for jmd is a definitive "no".
10:28 jmd Ok thanks.
10:28 i7c Because rebase will also update the working dir at some point.
10:28 jmd (that would explain why half our tests failed last night)
10:28 _ikke_ I think quite regularly even
10:29 _ikke_ afaik it applies commits in the working tree, not just in memory
10:29 MattMaker joined #git
10:29 grawity automatic tests really ought to run out of their own clone...
10:29 _ikke_ yes
10:29 osse the branch update is atomic, but I think wasn't in earlier version of git
10:29 watabou joined #git
10:29 _ikke_ They don't even require a repository imo
10:29 osse if you rebase foo then foo's reflog would be a mess. but it's not now
10:30 i7c But HEAD’s reflog is a mess. :)
10:30 _ikke_ each applied commit will result in a reflog entry
10:30 osse it is? :O
10:30 _ikke_ iirc
10:30 i7c I think you can see the whole walk in HEAD’s reflog.
10:30 _ikke_ you cannot do git reset --hard @{1} iirc to undo the rebase
10:31 osse leis
10:32 _ikke_ lais
10:32 i7c Okay not the whole walk, indeed.
10:33 i7c I mean you don’t see where it finds the common base, because I think that’s just a graph operation and does nothing with the refs.
10:33 i7c But the application of the single commits is visible in HEAD’s reflog.
10:34 i7c Whereas the branch’s reflog just contains one extra entry.
10:34 joegreen88 joined #git
10:34 _ikke_ makes sense, because the branch is only updated once
10:34 _ikke_ but HEAD many times
10:34 mariuscc joined #git
10:37 osse hmm indeed
10:37 osse I thought it disabled reflog stuff for HEAD or something
10:38 mikeh_ joined #git
10:40 _ikke_ osse: It does allow you to retrieve back rebased commits
10:40 _ikke_ ie, when you abort a rebase, the reflog of HEAD still contains the intermediate commits
10:40 tHEMtZ joined #git
10:41 nikivi joined #git
10:42 mikeh joined #git
10:43 ASOLAr joined #git
10:44 JeroenT joined #git
10:45 kurkale6ka joined #git
10:47 mikeh_ joined #git
10:47 mingrammer joined #git
10:49 nanga joined #git
10:49 ljc joined #git
10:49 wrouesnel1 joined #git
10:49 mikecmpb_ joined #git
10:50 mikeh joined #git
10:52 Rodya_ joined #git
10:52 anttim joined #git
10:54 pro_metedor joined #git
10:54 noom joined #git
10:54 noom hey #git
10:54 gitinfo noom: 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.
10:55 osse _ikke_: sure, I agree it's useful. I just didn't know Git did it
10:56 noom I'm having a very simple problem but can't find out how to solve it. I've issued a 'git add .' and I found out that it included files that I don't want to be tracked; how do I untrack them?
10:56 nanga joined #git
10:56 nanga joined #git
10:56 _ikke_ noom: git status would tell you
10:56 _ikke_ git reset <file>
10:56 noom it doesn't
10:56 noom thanks !
10:57 noom ah yes it does
10:57 _ikke_ :-)
10:57 noom my bad I didn't see that :')
10:57 mikeh_ joined #git
10:57 _ikke_ noom: You see it soo often that you start to ignore it
10:57 _ikke_ That's why we point to it, so that next time, you know where to look
10:57 osse _ikke_: is there a way to discover what mailing lists are available on public-inbox, or do I just have to know?
10:58 romerocesar joined #git
10:58 noom yep thanks
10:58 _ikke_ osse: wouldn't know
10:58 _ikke_ osse: there are some mailing list topics about it
10:58 noom but anyway it doesn't work ... I'd like to untrack a whole folder
10:58 noom and doing git reset -- folder doesn't do anything
10:59 tobiasvl noom: folders aren't tracked, just the files inside it
10:59 tobiasvl git reset -- folder/*
10:59 mikeh_ joined #git
10:59 MattMaker joined #git
11:00 _ikke_ ah, untracking is something different
11:00 noom "no matches found: home/woybi/.config/nvim/*"
11:00 _ikke_ tobiasvl: git would recursively reset files in a folder
11:00 _ikke_ tobiasvl: so that's not necessary
11:00 noom I didn't commited it, I just want to untrack them
11:01 tobiasvl _ikke_: ah, alright
11:01 noom how can I do then ? :/
11:02 _ikke_ git rm --cached <dir>
11:02 noom ah, 'git reset' without parameters did the trick
11:02 _ikke_ ok
11:02 tobiasvl noom: jfyi, if you haven't commited those files, they're not tracked. you want to unstage them
11:02 _ikke_ tobiasvl: If they're in the index, they're tracked
11:03 noom ah you're right tobiasvl sorry
11:03 noom hm :')
11:03 osse _ikke_: hmm, you shoukd know this. To me you're "the public-inbox guy"
11:03 _ikke_ osse: I'm not :P
11:03 _ikke_ osse: Eric Wong is :P
11:05 vassagus joined #git
11:05 mikeh joined #git
11:05 PioneerAxon joined #git
11:05 bremner so is the git list really going to use public-inbox?
11:07 _ikke_ Yeah, I think they are going to
11:07 anonymuse joined #git
11:07 Torrone joined #git
11:08 ahmed_elgabri joined #git
11:09 garyserj joined #git
11:09 noom left #git
11:10 mikeh_ joined #git
11:10 osse is it something that they "use"? isn't that just a mirror thingy?
11:11 _ikke_ osse: The always 'used' gmane
11:11 _ikke_ as a default reference to e-mails
11:11 mohabaks joined #git
11:12 garyserj I have a project where I accidentally added a file to the repository that I now realise I don't want there. Is there any way that I can remove it from all the commits?
11:12 osse Ahh, that kind of use.
11:12 ochorocho__ joined #git
11:12 _ikke_ public-inbox is replacing that since gmane doesn't offer an http interface anymore
11:12 encod3 joined #git
11:12 LucaTM joined #git
11:12 JeroenT joined #git
11:12 ASOLAr joined #git
11:12 _ikke_ garyserj: !BFG
11:12 gitinfo garyserj: [!bfg] A tool designed to remove large files, or passwords from history: https://rtyley.github.io/bfg-repo-cleaner/ (!rewrite applies)
11:12 osse garyserj: Pretend it's sensitive and follow this guide: !sensitive
11:12 gitinfo garyserj: [!filter_sensitive] You can use filter-branch to remove sensitive data from a repository's history. https://help.github.com/arti​cles/remove-sensitive-data/
11:12 osse _ikke_: the gmane guy got burned out :(
11:12 _ikke_ yeah
11:13 garyserj thanks
11:13 lss8 joined #git
11:13 grawity at least gmane still offers nntp... nice to see that public-inbox also does
11:13 grawity I guess git@ chose it because it's git-based and all
11:13 bremner also, really ugle
11:13 bremner err. ugly. So that's a big plus.
11:14 grawity I wouldn't call it ugly, more like overloaded with info and links and stuff
11:14 ecraven joined #git
11:14 grawity mail-archive.com is ugly
11:14 MattMaker joined #git
11:14 bremner there are many kinds of ugly ;)
11:14 ecraven left #git
11:15 rnsanchez joined #git
11:15 bremner I just mean it's completely css/div/span/whatever free.
11:15 HoierM joined #git
11:17 mikeh joined #git
11:19 osse I'm no web designer, but I'm pretty sure css is preferred over using stuff like <b>...</b>
11:19 osse <b> and <font> etc. are equivalent to committing generated files
11:19 sdothum joined #git
11:20 ahmed_elgabri joined #git
11:20 i7c Not nice but everybody does it?
11:21 jstein_ joined #git
11:21 jceb joined #git
11:22 Torrone joined #git
11:22 mikeh_ joined #git
11:24 vuoto joined #git
11:24 Spec-Chum joined #git
11:24 achlys joined #git
11:25 njbair joined #git
11:27 bronson joined #git
11:27 mikeh joined #git
11:27 AlexanderS joined #git
11:30 Raging_Hog joined #git
11:30 ASOLAr joined #git
11:31 watabou joined #git
11:34 PHPanos joined #git
11:34 osse they do?
11:37 rubyonrailed joined #git
11:39 MattMaker joined #git
11:40 lss8 joined #git
11:41 i7c I’ve seen it.
11:42 mingrammer joined #git
11:42 encod3 joined #git
11:44 encod3_ joined #git
11:44 achlys joined #git
11:44 hahuang61 joined #git
11:45 danslo_ joined #git
11:46 Manuel15 joined #git
11:46 Manuel15 joined #git
11:46 encod3 joined #git
11:46 Jellyg00se joined #git
11:47 vuoto joined #git
11:47 encod3 joined #git
11:48 rahtgaz joined #git
11:52 Dro joined #git
11:53 preyalone joined #git
11:54 ahmed_elgabri joined #git
11:54 cbreak i7c: these days, almost everything's css, with random java script thrown in for no reason
11:55 i7c I really try to stay away from webdev.
11:55 i7c I think this migh be quite accurate: https://i.imgur.com/Q3cUg29.gif
11:55 osse i7c: <em> is fine. <b> is sin.
11:55 mohabaks joined #git
11:56 achlys joined #git
11:57 sanketdg joined #git
11:57 tomog999 joined #git
11:58 d0nn1e joined #git
11:58 sanketdg joined #git
11:59 jimi_ joined #git
11:59 ljc joined #git
11:59 sanketdg joined #git
12:00 sanketdg joined #git
12:01 mikeh__ joined #git
12:01 ASOLAr joined #git
12:01 andrewSC_ joined #git
12:01 synthroid joined #git
12:02 jast webdev isn't really worse than anything else
12:02 jast because......... everything sucks! \o/
12:02 achlys joined #git
12:02 i7c I agree.
12:02 sanketdg joined #git
12:02 ASOLAr joined #git
12:03 jast I do web dev, though I tend to focus more on the backend side of things
12:03 i7c But if you look at C you think "ok, it was invented a long time ago, they didn’t know better then."
12:03 i7c But when you look at js
12:03 * i7c shakes head.
12:03 jast JS isn't so bad. for the most part, anyway.
12:03 i7c It’s cancer.
12:03 sevenfourk joined #git
12:04 jast the only really terrible thing is the lengths it goes to with implicit type conversion
12:04 jast and the web APIs, but that's not javascript's fault
12:04 phanimahesh joined #git
12:04 MattMaker joined #git
12:05 PioneerAxon joined #git
12:06 vassagus joined #git
12:06 Sceorem joined #git
12:07 garyserj joined #git
12:07 i7c Well, I disklike it just like everything else, just more.
12:07 i7c But it’s good that you don’t because somebody gotta do the job. :p
12:08 jast I leave most of it to the people who think adding 50 GB worth of frameworks to each and every website is a good idea
12:08 i7c Sounds ... awesome. :)
12:08 jast I like my internets extra chunky
12:09 jor asm.js :)
12:10 Torrone joined #git
12:10 hobodave joined #git
12:13 daniel-wtd joined #git
12:14 MattMaker joined #git
12:14 rgrinberg joined #git
12:15 anonymuse joined #git
12:15 solf1re joined #git
12:16 Semo_amin joined #git
12:21 mattcen joined #git
12:23 radsy joined #git
12:24 andlabs joined #git
12:24 daniel_rose joined #git
12:26 jhe_ joined #git
12:26 star_prone joined #git
12:27 johnmilton joined #git
12:28 bronson joined #git
12:29 hellyeah joined #git
12:29 hellyeah git commit -m -> m is parameter for messaging what you did last
12:29 johnmilton joined #git
12:29 hellyeah is there any other parameters commonly used beside -m?
12:30 rory hellyeah: check "git help commit" there are a few more options
12:31 rory hellyeah: I say don't use -m, better to edit a full commit message, but different devs/teams work differently
12:31 watabou joined #git
12:33 hellyeah so doing git commit and editting files is more preferrable
12:35 EY hellyeah, git commit -m restricts how much you can type easily, which makes typing good commit messages difficult.
12:36 EY hellyeah, so it's better that you use the editor that `git commit` brings up.
12:36 _ikke_ EY: I usually do git commit -m "foo" -e
12:36 _ikke_ To open the editor
12:36 hellyeah ic
12:36 jeffreylevesque joined #git
12:36 jmd EY: How does it restrict you?
12:37 hellyeah EY:  restrict how many characters you type?
12:37 EY jmd, unless you know the control characters to do new lines and indentation in the command line, it's difficult.
12:37 jmd ??
12:37 _ikke_ EY: git commit -m "blablabla
12:37 _ikke_
12:37 _ikke_ bla bla bla
12:37 _ikke_ "
12:38 jmd You do a new line by hitting the <Enter> key.
12:38 jmd At least on all machines I've ever played with.
12:38 EY _ikke_, yep, but not as easy as revising what you have typed in a previous line.
12:39 MattMaker joined #git
12:39 gelei007 joined #git
12:39 jmd Similarly indents are done with the <Tab> key.
12:39 _ikke_ ctrl+x ctrl-e to open editor
12:39 EY _ikke_, didn't know about the -e option though.
12:39 hellyeah with git commit, you just type what you change it is like kind of release note right
12:39 hellyeah _
12:39 hellyeah ?
12:40 inflames joined #git
12:40 _ikke_ hellyeah: It's mostly *why* you change things
12:40 hellyeah ic
12:40 _ikke_ what you change you can read from the diff
12:40 hellyeah do we edit readme.md?
12:41 Gamecubic joined #git
12:41 jceb joined #git
12:41 hellyeah one more question will you do seperate branch for every single change then combine it with master branch?
12:41 radsy joined #git
12:41 _ikke_ There are different workflows possible
12:41 _ikke_ each with their own advantages and disadvantages
12:42 star_prone joined #git
12:42 anonymuse joined #git
12:42 hellyeah well i am new you know better than me definitely
12:42 ahmed_elgabri joined #git
12:43 _ikke_ !workflow
12:43 gitinfo Finding the right workflow for you is critical for the success of any SCM project.  Git is very flexible with respect to workflow.  See http://sethrobertson.github.c​om/GitBestPractices/#workflow for a list of references about choosing branching and distributed workflows.
12:43 hellyeah what are those workflows you meantioned? Can you lead me to any link?
12:43 _ikke_ ^^
12:43 hellyeah ic
12:43 hellyeah thanks
12:44 hellyeah i will read that link
12:45 hellyeah why you need git pull oprigin master
12:45 GodGinrai joined #git
12:45 hellyeah if you commit every chance in local repo first?
12:45 Darcidride_ joined #git
12:46 hellyeah i am sorry for n newbie question
12:47 jmd hellyeah: Typically, one commits to the local repo, then either merges or rebases against the remove, then pushes.
12:48 ljc joined #git
12:48 jmd (I never understood why git pull does a merge my default.  In 99% of cases it is the wrong thing to do IMO)
12:48 anth0ny_ joined #git
12:48 star_prone joined #git
12:49 hellyeah merge is just merging brach
12:49 specing left #git
12:49 hellyeah i didnt hear rebase
12:49 hellyeah in theory we do some changes in local and puısh to remote
12:49 MattMaker joined #git
12:49 jmd Yes.
12:49 hellyeah so we have last update in local so do we really need to pull smt from remote?
12:49 achlys joined #git
12:49 jmd In case anyone else has pushed something since you last pulled.
12:50 hellyeah ah
12:50 hellyeah someone did a change in your project right
12:50 hellyeah do some push request
12:50 maxorator joined #git
12:50 jmd What I normally do is "git pull --rebase && git push"
12:50 ahmed_elgabri joined #git
12:50 bruce_lee joined #git
12:50 sanketdg joined #git
12:51 hellyeah so rebase is like merging local with remote is that correct?
12:52 jmd Rebasing means that all your local commits are rolled back and then re-applied on top of the new version of your local repo.
12:53 jimi_ joined #git
12:54 star_prone joined #git
12:54 EduardoMartins joined #git
12:54 hellyeah ic
12:56 MegaManSec joined #git
12:56 MegaManSec if you, for example, have a single file in a git repo, is it possible to tell how much data has been saved by not saving it X-commit times, over and over?
12:58 jost_ joined #git
12:59 grawity the simple option would be to count all commits with that file (minus one) and multiply by file size
12:59 grawity but that doesn't take into account the space saved by packfiles, which also avoid saving /similar/ files
13:01 ToBeCloud joined #git
13:01 Darcidride__ joined #git
13:02 cdg joined #git
13:02 synchronous joined #git
13:03 sangy joined #git
13:04 LUMIA930 joined #git
13:05 achlys joined #git
13:07 vassagus joined #git
13:07 achlys joined #git
13:09 JamesBaxter joined #git
13:09 Lunatrius joined #git
13:09 atomicb0mb joined #git
13:11 ExoUNX joined #git
13:12 EisNerd_ joined #git
13:12 ljc joined #git
13:13 ochorocho__ joined #git
13:13 mingrammer joined #git
13:14 Yuri6037 joined #git
13:14 Yuri6037 Hey all I have a problem
13:14 Yuri6037 The git client wants to upload VS runtimes !!!!!!
13:15 grawity use .gitignore
13:15 dunk I wonder if someone could tell me how it's possible that an interactive rebase could fail to reorder my last two commits when they do not edit the same lines? (they do edit adjacent lines though)
13:15 Yuri6037 It says it can't upload the repo because a file called UMod.VS.db is too large which is VS runtime !!!!
13:15 Repox joined #git
13:15 Yuri6037 How can I use that .gitignore I don't understand anything of that format
13:15 grawity dunk: when applying patches, git also looks at the surrounding context lines
13:16 dunk grawity: can I disable that somehow? or lessen it?
13:16 grawity dunk: not sure
13:16 dunk Obvs it still needs to find the position in the file
13:16 grawity it's partly a safety thing, to avoid applying wrong patches
13:16 dunk yeah
13:16 grawity especially if the patched location has moved and git has to look for it
13:17 achlys joined #git
13:17 star_prone joined #git
13:18 tigos2 joined #git
13:18 raijin joined #git
13:19 MattMaker joined #git
13:19 a_thakur joined #git
13:20 Yuri6037 Anyone please ? How should I make that .gitignore
13:20 DevAntoine joined #git
13:21 spinningarrow Yuri6037: The simplest way is just to list the files you want to ignore, one per line
13:22 Yuri6037 Where should I do it ? In .git in UMod root ?
13:23 spinningarrow If you put the name of a folder followed by a / (e.g. folder-name/) that folder will be ignored
13:23 spinningarrow No it is in the root folder of your git repository
13:23 JeroenT joined #git
13:23 Yuri6037 Ok so UMod root then thanks
13:23 spinningarrow Create a new file called ".gitignore" and put the stuff there
13:24 spinningarrow What is UMod?
13:24 MattMaker joined #git
13:24 Yuri6037 It's the name of my project
13:24 spinningarrow Ah, got it
13:25 Yuri6037 The only problem is that windows does not allow creation of .gitignore as file already exists but is invisible
13:25 Yuri6037 Here for my project : https://github.com/StoneLineDevTeam/UMod
13:25 simulacr is there a some wrapper/way to move forward/backward all tree files between commits? for example 'latestcommit -> oldcommit -> newerthanoldcommit'
13:25 spinningarrow If it already exists, you can just edit the existing one
13:25 Yuri6037 That's what I'd like to update
13:25 rubyonrailed joined #git
13:25 spinningarrow Yuri6037: You can even edit it directly on github
13:25 Yuri6037 The problem is that windows explorer doesn't recognize this file it's not here
13:26 spinningarrow You can open it using a text editor like Sublime Text. Open the project in the editor and you'll see it there
13:26 Yuri6037 and I can't do directly from server as my git client has again a push problem it absolutely needs -f otherwise it asks pull and then asks me to overwrite the whole local repo and destroy again my docs
13:27 spinningarrow You can also show hidden files in Windows
13:28 spinningarrow One thing to note - if a file is already being tracked in git, simply putting it in .gitignore won't do the trick. You'll also need to run `git rm --cached <filename>` and then commit that change
13:28 Yuri6037 .gitignore stil does not show on explorer I can not access it
13:28 star_prone joined #git
13:28 fahadash joined #git
13:28 HanClinto joined #git
13:29 shinnya joined #git
13:30 hellyeah does it sound logical?
13:31 hellyeah if someone do "push request" and maek some changes on my code. I keep this remote changes in a brach and my local files in another brach and compare both files about  what difference they made etc. And then merge it with master brach
13:31 mingrammer joined #git
13:32 mingrammer joined #git
13:32 watabou joined #git
13:33 grawity that's how regular pull requests work, don't they?
13:34 MattMaker joined #git
13:35 hellyeah grawity:  i am new in that business
13:37 Yuri6037 I arrived to access the .gitignore and added the UMod.VC.db
13:38 Yuri6037 however the git rm --cached UMod.VC.db only prints rm 'UMod.VC.db' and when I git push I have still the same number of files than before I typed the command
13:38 achlys joined #git
13:39 liali joined #git
13:40 Yuri6037 The command does not remove from the commit
13:41 grawity it does, you need to actually `git commit` the removal
13:41 grawity probably with --amend, in this case
13:41 Yuri6037 The problem is that I have new files modified
13:41 stonerfish joined #git
13:41 Yuri6037 If I commit it will add my new modified files which are not done and should not be uploaded
13:43 spinningarrow Don't stage them; only commit the removed file
13:43 Yuri6037 How git does not recognise git commit --ammend
13:44 spinningarrow The commit will only consist of staged files (those that have been added to the index using 'git add')
13:44 Yuri6037 Oh no no
13:44 spinningarrow perhaps a type? single m instead of double in 'amend'
13:44 spinningarrow typo*
13:44 Yuri6037 It's back wanting to delete all my local files
13:45 Yuri6037 "Merge branch 'UMod_Alpha' of https://github.com/StoneLineDevTeam/UMod into UMod_Alpha"
13:45 Torrone_1 joined #git
13:45 hahuang61 joined #git
13:45 Yuri6037 I fucking don't want any merge I only want it to push without that shit .VC.db that's all why is it so complicated
13:45 Yuri6037 ?
13:46 encod3_ joined #git
13:46 caffeinatedcode joined #git
13:46 caffeinatedcode joined #git
13:47 Yuri6037 I think I've no choice I will need to create a new repo or a new branch or give up with git
13:47 star_prone joined #git
13:49 ljc joined #git
13:49 Yuri6037 I tried with Git GUI but I have a message I don't understand now !!!
13:49 spinningarrow What is the message?
13:50 Yuri6037 "This repository has approximately 512 loose objects."
13:50 Yuri6037 What does that means, that it it's going to destroy my work again ?
13:51 Yuri6037 I had so many issues with git that now when I see "loose", "merge", "remove" I get angry
13:51 a_thakur joined #git
13:51 spinningarrow It's unlikely that your work will be destroyed
13:51 Yuri6037 Git already removed my repo twice (fortunatly HDD backup) corrupted .ccp and .h files
13:51 spinningarrow But I'd suggest recloning and then applying your changes on top of the fresh clone
13:52 spinningarrow What steps got you into this state, do you remember?
13:53 a_thakur joined #git
13:53 ljc joined #git
13:54 Yuri6037 Yes I remember I got HDD failure so I backed up on remote GitHub.com before HDD died, then I got new HDD restored from my NAS backup (I also made one before HDD failure), then I reinstalled git client on system HDD then spent almost 2 months thinking/redesigning large parts of the project...
13:55 Yuri6037 Then came yesterday evening after finishing VMTConverter.cpp and .h the time to upload so I oppened git bash (right-click Git Bash here)
13:55 Yuri6037 then I done like beore git add --all which worked no issues
13:56 Yuri6037 then I done git commit -am "I don't remember exactly what I inputed here"
13:56 Yuri6037 Then I done git push which resulted in error that my local repo wasn't up to date (normal I changed the Readme.md
13:57 Yuri6037 directly on internet not in local
13:57 Yuri6037 But when I tried git pull it said it wanted to delete EVERY SINGLE files that I created in those 2 monthes I opened task manager and Killprocess
13:57 star_prone joined #git
13:59 Yuri6037 So I searched on internet how to force push and added like said -f after push and it worked but this morning I got a problem that UMod.VC.db was not ignored
14:00 Yuri6037 Anyways should I copress or not the database ? (that's last info in the message box of the GUI)
14:00 ShekharReddy joined #git
14:00 al-damiri joined #git
14:01 Yuri6037 Apparently from Internet it's not going to delete my files
14:01 p4trix joined #git
14:01 JanC_ joined #git
14:02 AaronMT joined #git
14:03 ash_workz joined #git
14:03 AaronMT joined #git
14:04 MattMaker joined #git
14:04 Yuri6037 Ok if your shit app realy wants to merge and delete my files ok I'll let do it but with a backup....
14:04 Gsham joined #git
14:10 ResidentBiscuit joined #git
14:12 vassagus joined #git
14:13 raijin joined #git
14:13 Yuri6037 From what my net graph says me it's sending 64 451 bytes per seconds but Git GUI does not display any progress
14:15 kaldoran joined #git
14:15 star_prone joined #git
14:16 Endarked joined #git
14:19 daniel-wtd joined #git
14:20 star_prone joined #git
14:24 govg joined #git
14:24 harish__ joined #git
14:24 star_prone joined #git
14:25 l4v2 joined #git
14:26 Prisoner-2460_1 joined #git
14:26 ASOLAr joined #git
14:27 Prisoner-2460_1 yesterday tried to squash. foolishly just ran git rebase -i, got nervous and did :q. then made my commit and it _seemed_ to have done what i wanted. is it possilbe i screwed with anythign that I might be unable to see right now
14:30 kanzure p/win 6
14:30 james_checkout joined #git
14:31 danemacmillan Has anyone here used GitHub's LFS plugin? How well does it handle very large directories of, say, 40GB. The files themselves are mostly average size JPEGs.
14:31 danemacmillan Would it be a good use case
14:31 james_checkout hit
14:31 james_checkout hello all
14:32 james_checkout i think i just made the biggest mistake ever
14:32 james_checkout i have been working on a project for a day or two
14:32 james_checkout without committing much (first mistake)
14:32 ASOLAr joined #git
14:32 james_checkout i was getting ready to commit my work
14:32 james_checkout but needed to checkout a file that i didn't want to commit
14:33 james_checkout keep in mind all of my changes were not staged
14:33 danemacmillan lol
14:33 danemacmillan I see where this is going
14:33 james_checkout so i did the worse possible thing ever...
14:33 james_checkout git checkout -- .
14:33 kpease joined #git
14:33 james_checkout is there any saving this?
14:33 james_checkout or is my life over?
14:33 danemacmillan Unfortunately, no
14:34 danemacmillan Over
14:34 MattMaker joined #git
14:34 danemacmillan Unless you do some disk recovery stuff--but that's outside of git
14:35 danemacmillan I don't envy you. You've just made my day more tolerable.
14:35 theverbg_ joined #git
14:36 fmcgeough joined #git
14:36 Prisoner-2460_1 is there a way to just view the changes in a state of head through reflog?
14:36 morenoh149 joined #git
14:38 shaggycat joined #git
14:38 kpease joined #git
14:38 mizu_no_oto joined #git
14:38 theverbg_ joined #git
14:38 ochorocho__ joined #git
14:39 [cust] joined #git
14:40 phaleth joined #git
14:41 GavinMagnus left #git
14:41 docnvk joined #git
14:42 james_checkout WHYYY
14:42 james_checkout this is the saddest day of my life
14:43 _ikke_ james_checkout: Did you ever git add those changes?
14:43 _ikke_ james_checkout: Does your editor keep history? (yesterday someone managed to recover through PHPStorm history)
14:43 CheckDavid joined #git
14:44 nathdwek joined #git
14:44 MattMaker joined #git
14:46 nathdwek Hi there, does anyone know if there is a possibility to emulate the behaviour of git add -p with git diff, and be prompted for each patch whether to keep it or not?
14:46 _ikke_ nathdwek: keep in what sense?
14:46 nathdwek do a partial checkout
14:46 _ikke_ git checkout -p
14:47 nathdwek As in, I'm not ready to commit yet, but I'd like to continue working while removing blocks of modifications
14:47 navidr joined #git
14:47 nathdwek Oh this exists?
14:47 _ikke_ yes'
14:47 nathdwek AWESOME
14:47 _ikke_ and git reset -p too
14:47 xaviergmail joined #git
14:47 nathdwek thanks. Bye
14:47 _ikke_ (patches you say yes to will be reverted)
14:47 nathdwek Yeah the prompt is explicit, I'll find my way
14:48 _ikke_ okasy
14:48 _ikke_ success
14:48 rubyonrailed joined #git
14:48 thiago joined #git
14:50 gargola joined #git
14:50 james_checkout @_ikki_
14:50 james_checkout im using sublime
14:51 Eryn_1983_FL joined #git
14:51 star_prone joined #git
14:52 Remram[m] IntelliJ IDEs keep local history but probably no one else
14:52 Remram[m] (PyCharm, PHPStorm, ...)
14:53 _ikke_ If you had the files open, sometimes you can use undo to get it back
14:54 preyalone joined #git
14:54 MattMaker joined #git
14:55 james_checkout well i guess the lesson here is to not use that damn '.'
14:56 AaronMT_ joined #git
14:56 _ikke_ or use git stash + git stash drop
14:56 vassagus joined #git
14:56 jor james_checkout: also throwing away work, and redoing it from memory gives a much better end result then the original... (just trying to cheer you up ;)
14:56 james_checkout thanks jor
14:56 james_checkout yeah prob won't take me as long
14:56 james_checkout i just have to let go at this point
14:57 bsanford joined #git
14:57 vassagus joined #git
14:58 spinningarrow james_checkout: if you ever did a git diff in your terminal, you could potentially recover your data from there… but I admit it's a long shot
14:58 AaronMT joined #git
14:59 hobodave joined #git
15:00 james_checkout yeah git diff would of been nice had i done it. i really love the suggestions guys. :)
15:01 eycsigfy joined #git
15:02 cydrobolt joined #git
15:02 cydrobolt joined #git
15:03 daniel-wtd joined #git
15:03 umbSublime joined #git
15:03 nevyn joined #git
15:03 madewokherd joined #git
15:03 tomzx joined #git
15:04 umbSublime_ joined #git
15:04 MattMaker joined #git
15:06 umbSublime_ joined #git
15:06 rkazak joined #git
15:07 cdown joined #git
15:08 dustinm joined #git
15:08 Dro joined #git
15:09 MattMaker joined #git
15:10 clemf joined #git
15:11 Anticom Hi all. I've got a strange issue: I've got https://github.com/splunk/pion as a submodule in my project. Now when i do 'git checkout 5.0.1' and then run 'git describe --exact-match $(git rev-parse HEAD)' i get "fatal: no Tag matches exactly '086cca38007bedcef532e45433f4ba1573d51c01'
15:11 Anticom Any ideas why that is?
15:12 le_melomane joined #git
15:13 kaldoran_ joined #git
15:15 fakenerd joined #git
15:15 griffindy joined #git
15:16 _ikke_ Do you run it in the submodule, or in the parent repository?
15:16 inflames joined #git
15:17 InfoTest joined #git
15:20 kaldoran_ joined #git
15:20 Anticom _ikke_: it's in my PS1, so in pwd which is the submodule itself
15:22 Anticom FYI i haven't had this issue yet in other submodules etc.
15:22 osse Anticom: what about git describe --exact-match HEAD
15:22 grawity refs/tags/5.0.1 is a thin tag, `git describe` ignores those by defaults
15:22 Anticom osse same
15:22 osse Anticom: add --tags
15:22 kaldoran_ joined #git
15:22 Anticom grawity: ah, i was thinking that those tags might be some imported ones by git svn... afaik they're not imported "propery" by any means
15:23 _ikke_ There are two types of tags in git
15:23 Anticom osse: that fixed it :)
15:23 _ikke_ one is just a ref, pointing to a commit
15:23 grawity they *might* be imported properly, it's just there isn't enough to import
15:23 shgysk8zer0 joined #git
15:23 _ikke_ the other is a ref, pointing to a tag object, pointing to a commit
15:24 marc_v92 joined #git
15:24 kaldoran_ joined #git
15:24 Anticom _ikke_: where the latter has a commit message aswell while the former has not, right?
15:25 Anticom wasn't that one of the differences?
15:25 _ikke_ Yes, the tag object has a message
15:25 EvilPeng1 joined #git
15:25 kaldoran_ joined #git
15:25 ASOLAr joined #git
15:25 Anticom git skills 1up
15:25 _ikke_ and author+date information
15:25 _ikke_ and can be signed
15:25 Anticom _ikke_: signing is one thing i really have to look into
15:26 _ikke_ Never bothered to do
15:26 _ikke_ I don't have projects where it matters
15:26 grawity I don't either, but Github's "[verified]" tags look pretty
15:26 grawity it's like the blue ✔ in Twitter
15:26 Anticom well we used to do a lot of reviews prior to commiting to svn. Maybe they want to keep this. idk
15:26 grawity makes you feel leet
15:26 _ikke_ gamification
15:27 Anticom but shouldn't be neccesary since you can't push to develop directly. We're enforcing PRs so the review happens there anyway
15:27 kaldoran left #git
15:28 _ikke_ yeah
15:28 _ikke_ Review is more important than signature
15:29 Anthaas joined #git
15:30 Anticom Oh I've got another question about git-flow branching model
15:30 elect joined #git
15:30 Anticom We want to use tags for versioning
15:30 bronson joined #git
15:30 Anticom However the tags are only added to commits on master iirc
15:31 daniel-wtd joined #git
15:31 Anticom How would i get a reference to a prior tag in development branch or feature/etc. branch?
15:31 grawity tags aren't limited to a specific branch
15:31 dmto joined #git
15:32 grawity if a commit ID gets tagged, it's tagged in /all/ branches it is in
15:32 Anticom but i need a forward reference to my commit right?
15:32 Anticom which i wouldn't have if i never merged master back to develop
15:32 jguillen joined #git
15:32 mikecmpbll joined #git
15:32 grawity well then, it makes sense that you won't find the prior tag
15:33 Anticom grawity: that's my point
15:33 _ikke_ Anticom: nothing prevents your from merging master back to develop
15:33 Anticom but from how i got it i'm not supposed to merge master to develop
15:33 Anticom so IMO it's a flaw in that model
15:33 _ikke_ Anticom: Just merge master back into develop, problem solved
15:35 mikeh__ joined #git
15:35 yaun joined #git
15:35 watabou joined #git
15:35 Gitzilla joined #git
15:35 napping joined #git
15:35 Anticom And what would be the correct order then if i had a hotfix? (reference image: http://nvie.com/img/git-model@2x.png) Merge hotfix to develop + master, then merge master back to develop i suppose?
15:36 _ikke_ Anticom: If you merge master to develop, it's not necessary to seperately merge hotfix into develop
15:36 DonRichie joined #git
15:36 Anticom but otherwise it wouldn't comply with the model
15:37 Zamyatin joined #git
15:37 Anticom :)
15:37 _ikke_ Doesn't matter
15:37 _ikke_ it's a model, not law
15:38 sssilver joined #git
15:38 _ikke_ Adopt it to your needs
15:38 Anticom fair enough
15:38 _ikke_ https://github.com/nvie/gitflow/wiki/FAQ
15:39 _ikke_ It talks about this probelm
15:39 _ikke_ proble
15:39 MattMaker joined #git
15:39 sssilver hey guys, having this problem with my rebase... can't make sense of it -- can anyone help? seems trivial https://dpaste.de/wJjc
15:39 mingrammer joined #git
15:39 sssilver and also, what do I do with file 'a'?
15:39 romerocesar joined #git
15:40 Anticom thank's for the link _ikke_
15:41 cagmz joined #git
15:41 napping Is there any way to have only a subdirectory checked out into the working directory?
15:41 _ikke_ napping: sparse checkouts
15:42 napping yes, but if sparse-checkout is just /some/deep/path/, can those files be checked out to the root of the working directory?
15:43 grawity no
15:43 Dalvini joined #git
15:43 Dalvini left #git
15:44 napping hmm, making a symlink to that directory works better than expected
15:45 le_melomane joined #git
15:46 marc_v92 joined #git
15:46 LionsMane joined #git
15:47 leeN joined #git
15:47 raijin joined #git
15:48 ReenignE joined #git
15:48 ReenignE joined #git
15:49 MattMaker joined #git
15:50 durham joined #git
15:52 anonymuse joined #git
15:53 hahuang61 joined #git
15:57 chrisshattuck joined #git
15:59 MattMaker joined #git
16:00 Manuel15 joined #git
16:00 Manuel15 joined #git
16:00 ASOLAr joined #git
16:02 watabou joined #git
16:03 TinFury joined #git
16:04 TinFury I think I'm forked how do I check?
16:04 t0by joined #git
16:04 grawity see if you have any stab wounds
16:05 osse it's a dead giveaway if you have three or four of them, tiny, evenly spaced along a straight line
16:05 Gsham joined #git
16:06 kba joined #git
16:06 napping See whether "git status" says you are "On branch master"
16:06 TinFury napping, thanks
16:07 TinFury It says i'm on master but I'm ahead by 14 commits
16:07 TinFury How could that be... my production server never gets changed.
16:07 TinFury Well.... shouldn't
16:07 TinFury it's only used by myself.
16:07 axl0909_ joined #git
16:08 napping Are you recently learning git?
16:08 osse ahead compared to what? origin/master?
16:08 TinFury napping, not really but I use it in a very simple way.
16:08 ojacobson !lol is good for visualizing the commit graph
16:08 TinFury I have a dev server and i use git to backup my dev
16:08 gitinfo A nifty view of branches, tags, and other refs: git log --oneline --graph --decorate --all
16:08 TinFury Then on production I just "git pull origin master"
16:08 axl0909_ Hello, I have a question about git diff, and reading the man pages wasn't of much help
16:08 TinFury And that's it
16:08 TinFury I don't fork or have multiple users...
16:08 napping Very simple, if you haven't seen that before
16:09 liefer joined #git
16:09 TinFury Just commit push and pull
16:09 yaun left #git
16:09 napping "ahead" means the directory you are in has more commits than the repository you cloned it from
16:09 axl0909_ To be precise, I don't understand what git diff prints as output if 3 commits are given as arguments
16:09 thiago joined #git
16:09 TinFury napping, but I don't commit on the production server
16:09 axl0909_ Could anyone perhaps help?
16:09 napping which server are you seeing that message on?
16:10 TinFury my production website. On that server I ONLY do git pull origin master
16:10 liefer Im trying to push to github, but it keeps using the wrong account! I've tried setting the remote to https://username@github/... with no luck. How do i change the username? git config user.name shows my real name, not my username
16:10 napping well, you somehow have more commits that "origin" has
16:11 napping axl0909_: I don't think git diff can take more than two commits. How are you running it?
16:11 TinFury napping, Wouldn't I have to do a git commit on that server to have commits?
16:12 axl0909_ @napping exactly my problem :P
16:12 axl0909_ I will explain the situation better
16:12 mycyber joined #git
16:12 osse it's not documented and seems to behave weirdly
16:13 napping TinFury: origin/master is what your local repository remembers as where origin's master branch was the last time you fetched or pulled
16:13 axl0909_ @napping I have two branches, a and b, and their merge-base. What I want is to calculate which files "git status" will print after attempting a "git merge --no-commit b" while on branch a.
16:13 napping So if your dev server moved master back a few commits and then production checked again you might end up like this
16:13 Fernando-Basso joined #git
16:14 TinFury napping, Ahh... that did happen. but only by one commit
16:14 MattMaker joined #git
16:14 TinFury I get this "Your branch is ahead of 'origin/master' by 14 commits.'
16:14 TinFury That's too much.
16:14 phlix joined #git
16:15 axl0909_ my guess was that it prints whatever a "git diff a...b" would print
16:15 phlix I am having a hard time trying to update my fork with master. Issue is that master deleted a bunch of unused files. So if I try to merge master or anything into my new fork I think it assumes since those files still exist on my fork that they are to be merged in... Is the only way to fix this is to delete my fork and re create it?
16:16 axl0909_ however this is not the case, because if a and b have both modified a file foo (since merge-base) but the end result is the same at both branches, "git diff a...b" will include foo in the output, while "git status" after attempting merge won't.
16:17 Lachezar joined #git
16:17 napping What exactly is "attempting merge"? The merge reported conflicts and didn't finish?
16:18 osse the only thing I've managed to find is this:  ''git diff x y z'' prints the same as ''git diff -R x y'' but only if z is more recent than x and y, otherwise it prints nothing
16:18 axl0909_ It is "git merge --no-commit b"
16:18 Lachezar Hey all. I got myself a predicament: I've got about 1300 repositories that I attempted to convert from working to bare (mv <repo>/.git <repo>.git; rm -r <repo>; git config ... bare)'
16:18 napping well, that set's things up so anything that merged successfully appears up to date, so it makes sense that file wouldn't show in git status
16:18 axl0909_ It is not a necessary to have conflicts
16:18 dreiss joined #git
16:19 axl0909_ Yeah, I agree to that
16:19 Lachezar Afterwards I found out that there is --mirror, and not just --bare, so I tried to replace the refs on all repositories, but now it seems none of the repos has any tags/branches/etc. However I can see objects (packs) inside.
16:19 MattMaker joined #git
16:20 Lachezar Is it possible to somehow rebuild a repository without the need to download everything (its 170+ GB of data).
16:20 axl0909_ However is there a way to get the same result using git diff? I mean, find which files have been changes in branch b since merge-base, and exclude the files that are up-to-date with branch a?
16:20 thiago Lachezar: define "rebuild"
16:20 axl0909_ I tried achieving that with a 3 argument diff
16:20 axl0909_ and in my surprise I got the results I needed
16:21 thiago axl0909_: git diff a...b
16:21 thiago axl0909_: not sure
16:21 axl0909_ But since I do not understand how it works, I do not know if I will always get the expected results
16:21 napping TinFury: I'd check where you are at on both servers. If you are ahead like that, git pull will leave you alone rather than rolling back to match the server
16:21 Lachezar thiago: Well… I'd like to have the repository to look as if it was 'git clone --mirror'
16:22 thiago Lachezar: then you have to download everything
16:22 ytl joined #git
16:22 Lachezar thiago: Data too?
16:22 axl0909_ @thiago, I explained above that in case a file has been changed in a and b, but is of the end result is the same at both branches, your solution will print that file too
16:22 thiago Lachezar: in order to have everything, you have to download everything
16:22 thiago Lachezar: well, yeah
16:22 napping axl0909_: I doubt any simple 3-argument diff will exactly match what you see after merge, because what you see after merge  includes anything the merge strategy successfully handled
16:22 axl0909_ @thiago although that file is in fact up too date with branch a
16:22 thiago Lachezar: you've just defined that you want an exact copy. That means everything.
16:22 AnarchyAo joined #git
16:23 thiago axl0909_: so you want it to show the diff between B and C, except where A and B are the same?
16:23 TinFury napping, oh... thanks.. trying to figure it out
16:23 Lachezar thiago: No, thjat
16:23 axl0909_ @napping if there are conflicts, after merge if you run git status you can also see conflicting files
16:23 napping axl0909_: if you specifically want to see what merge leaves as the conflict when it doesn't succeed, there's probably a tool that does that
16:23 Lachezar thiago: No, that's not what I was saying/writing.
16:23 axl0909_ @thiago exactly!
16:23 thiago Lachezar: so what do you wat?
16:23 napping axl0909_: the tricky case is when there  are multiple ancestors
16:23 thiago axl0909_: write a script
16:24 thiago axl0909_: PS: don't put @ before people's nicknames
16:24 phanimahesh joined #git
16:24 phlix anyone have a tutorial or something that can show me how to update my fork with master.. Problem is master has deleted files and folders. When I try to merge my fork its thinking those files that need to be nuked should be added.. not sure how to do this.
16:24 axl0909_ thiago oh ok sorry for @
16:24 thiago Lachezar: what do you want then?
16:25 axl0909_ thiago yeah the script is the way I will probably go for. I was just wondering if that would be possible with a git command
16:25 Lachezar thiago: I want to *not* need to readownload 170+ GB of data that I already have.
16:26 axl0909_ thiago and the fact that in the man pages there is nothing about 3 arguments, but it works, really left me wondering what it does :P
16:26 thiago Lachezar: if you already have it, you don't need to download it
16:27 thiago Lachezar: but I don't get what you want to do
16:27 Lachezar thiago: Yeah, you don't.
16:27 thiago Lachezar: do you have a repository that you cloned from somewhere, and now you want to transform it to be as if you had cloned --mirror?
16:27 axl0909_ napping: right now, if I want to find which files will be in conflict and git will not be able to autoresolve after a merge is the following
16:28 snowkidind joined #git
16:28 Lachezar thiago: Yes!
16:28 harish joined #git
16:28 thiago Lachezar: rename all the refs in refs/remotes/origin/* to be refs/heads/*
16:28 Lachezar thiago: well… ~1300 repositories
16:28 axl0909_ napping: Since there is not git merge --dry-run, I do "git merge --no-commit", then "git ls-files -u" and finally "git merge --abort"
16:29 thiago Lachezar: then update the fetch spec (remote.origin.fetch) to be +refs/*:refs/*
16:29 thiago Lachezar: then fetch again
16:29 phlix is deleting my fork the only way to accomplish this?
16:29 thiago phlix: define "update"
16:29 axl0909_ However my original question has not been answered.. Does anyone know  what a git diff with 3 arguments actually prints?
16:30 marc_v92 joined #git
16:30 Lachezar thiago: Rename refs… Can't just delete all refs?
16:30 thiago phlix: do you have changes in your repository that the upstream doesn't have?
16:30 thiago Lachezar: if you do that, you delete the objects they refer to
16:30 phlix thiago: I don't think so
16:30 thiago Lachezar: you need to have those refs when you do the fetch
16:30 Lachezar thiago: Ahhh…
16:30 thiago phlix: then git fetch from the upstream, then push to your repo
16:30 phlix it actually says that my version of master is 27 commits ahead which I am not sure why it says this
16:30 axl0909_ osse said something  of interest earlier.. I will check that
16:30 phlix I have tried that
16:31 thiago Lachezar: it might be enough to update the fetchspec and then fetch
16:31 Lachezar thiago: Well… If I do not do a 'git gc' those should be around?
16:31 osse axl0909_: not very interesting. just random experiments for three minutes.
16:31 Lachezar thiago: I've already updated the fetch specs (got help in #git yesterday.
16:31 phanimah1sh joined #git
16:31 thiago phlix: pull from the upstream, then push to your repo
16:31 thiago Lachezar: then just git fetch
16:32 phlix but when i pull from my upstream and look at the files that should be removed they still exist.
16:32 axl0909_ osse: well, it's a start.. I've been trying to decode the output for some time now :P
16:32 bronson joined #git
16:32 thiago phlix: your expectation is wrong
16:32 thiago phlix: git pull did the right thing
16:33 ash_workz joined #git
16:33 MattMaker joined #git
16:34 xaa joined #git
16:34 mingrammer joined #git
16:34 thiago phlix: if you're not seeing what you expected to see, you must have fetched from the wrong repository
16:34 chrisshattuck joined #git
16:34 phlix hmm ok let me try a few things
16:35 Lachezar thiago: I did, but it looks different… There are links for branches, upstream… and there are remotes/origin/* branches
16:35 thiago Lachezar: what looks different?
16:36 nitric joined #git
16:36 Lachezar thiago: The repository meta data: config, refs. I am comparing a 'converted' repository for one of the projects with a fresh 'clone --mirror'
16:37 zacts joined #git
16:37 napping axl0909_: looks to me like git-diff A B C prints a unified diff with two columns of +/ /- instead of three, with the first showing what A changed from B, and the second what B changed from C
16:37 osse that would make sense
16:37 osse I never tested with commits in a "V" so to speak
16:37 napping or wait, maybe first column considers "B" the newer
16:38 thiago Lachezar: what's different in the config?
16:39 Lachezar thiago: missing 'remote.origin.mirror=true', and there is a [branch "master"] section that is not there in the --mirror
16:39 MattMaker joined #git
16:39 thiago Lachezar: then set that config
16:40 Lachezar thiago: That I think I can do, how to remove the [branch "master"] section?
16:40 thiago Lachezar: ignore that
16:40 thiago Lachezar: I recommend you do a fetch first without the mirror=true option, then add it, then fetch again
16:40 blackwind_123 joined #git
16:41 napping Lachezar: the claim that deleting refs will delete objects is wrong (as long as you don't run git gc)
16:41 cyphase joined #git
16:41 thiago true
16:41 Prisoner-2460_1 joined #git
16:41 thiago but if the objects aren't referenced, then git fetch will download them again
16:41 kaldoran_ joined #git
16:41 thiago as if they had been deleted
16:42 napping is that how it goes?
16:42 Lachezar thiago: So basically the objects are there, but are inaccessible?
16:42 thiago unreachable
16:42 Darren_ joined #git
16:42 chimay joined #git
16:42 thiago if you know them by name, you can still access them
16:42 thiago but Git doesn't try to find objects like that when fetching
16:43 dviola joined #git
16:43 Darcidride joined #git
16:43 thiago the fetch algorithm basically sends to the other side "here's what I have, send me what I don't"
16:43 thiago if you delete the refs before fetching, then your client won't send anythng to the server, so the server needs to send everyhing again
16:46 napping do you have any packed-refs yet?
16:46 armyriad joined #git
16:46 DolphinDream how do I find out if a given commit was cherry picked into another branch ?
16:46 napping If not, I'd suggest just making a new subdirectory under refs, and moving everything existing under there
16:46 osse DolphinDream: git cherry
16:47 diogenese joined #git
16:47 napping thiago: git-fetch-pack documents say what it sends as available bases - all local refs
16:47 SteffanW joined #git
16:49 napping Lachezar: did you get something working?
16:49 MattMaker joined #git
16:49 t0by joined #git
16:50 vuoto joined #git
16:50 napping Is it an acceptable download to run git clone --mirror --depth=1 ?
16:51 mostlybadfly joined #git
16:52 osse axl0909_: you can do git diff A B C D E F G H
16:52 osse axl0909_: from what I can see in the source it does N - 2 three way diffs
16:52 vuoto joined #git
16:52 osse i mean N - 1
16:52 daynaskully joined #git
16:52 napping which three-way diffs?
16:53 osse no clue :P
16:53 osse This greek reads like C code to me
16:54 osse seems like it does A..B and A...C and A...D etc.
16:55 LUMIA930 joined #git
16:57 dviola can git-revert take a file, so I can only revert a file?
16:57 dviola or should I use checkout instead?
16:57 dviola I want to do: git revert commit -- path/to/file
16:57 dviola or something like that
16:58 napping git checkout commit -- path/to/file works
16:58 dviola thanks
16:58 osse kinda sorta
16:58 napping that is, it overwrites working directory with the old version
16:58 HoierM joined #git
16:58 RedNamalas joined #git
16:59 MattMaker joined #git
17:00 DolphinDream osse, cherry huh? how exactly  ? git cherry master SomeSHA ?
17:00 dviola I should be able to use HEAD^ in this example if I want to revert to the previous commit from HEAD right?
17:01 napping you write the commit any way you want (actually any "tree-ish" for a path). Just be careful you don't have unsaved work in that file
17:01 dviola ok
17:02 dviola HEAD^ refers to the previous commit from head though, right?
17:02 dviola oh yeah, git show shows that
17:02 napping man gitrevisions shows you all the options
17:02 gitinfo the gitrevisions manpage is available at http://jk.gs/gitrevisions.html
17:03 a_thakur joined #git
17:03 kadoban joined #git
17:03 dviola thanks
17:03 Sasazuka joined #git
17:05 Mixologic joined #git
17:11 a_thakur joined #git
17:13 elementalest joined #git
17:14 elementalest joined #git
17:15 elementalest joined #git
17:15 elementalest joined #git
17:15 elementalest joined #git
17:15 elementalest joined #git
17:16 aax joined #git
17:16 elementalest joined #git
17:16 elementalest joined #git
17:17 mohabaks joined #git
17:23 vuoto joined #git
17:24 MattMaker joined #git
17:24 astrofog joined #git
17:24 marc_v92 joined #git
17:25 codeamend joined #git
17:25 codeamend whats up folks
17:26 Spec-Chum joined #git
17:26 Rodya_ joined #git
17:27 kbs This is more a crypto than a git thing, but out of curiosity - has someone formally analyzed the git merkle DAG for what properties it may offer? (e.g. assuming collision-resistance, cannot create two trees with the same root hash, etc.)
17:27 AaronMT_ joined #git
17:27 codeamend I commited 3 times on the wrong branch and need to move these commits to the right branch. I have never pushed anything. Is this called a rebase?
17:28 mingrammer joined #git
17:28 axl0909_ osse: napping: it seems that git diff A B C D prints a combined diff
17:28 daey joined #git
17:28 thiago codeamend: it's easier to visualise a cherry-pick sequence
17:28 axl0909_ osse napping:  where each column is how each parent sees modifications
17:29 codeamend Cool. Ill do some reading. Thanks
17:29 thiago codeamend: check out the correct branch, cherry-pick the commits, then check out the wrong one and reset it back to what it should be
17:29 MattMaker joined #git
17:29 axl0909_ osse napping for example first column of -/+ show how B views modifications made by A
17:30 axl0909_ osse napping for example second column of -/+ show how C views modifications made by A
17:30 morenoh149 joined #git
17:30 axl0909_ osse napping etc.. It is clear to me now. Thanks!
17:30 Peuc joined #git
17:32 andrewSC left #git
17:35 vuoto joined #git
17:35 vuoto joined #git
17:35 codeamend thanks thiago
17:36 RedNamalas left #git
17:38 morenoh149 joined #git
17:38 realz joined #git
17:38 ki0 joined #git
17:39 MattMaker joined #git
17:42 Mobutils joined #git
17:43 daey if i add/delete/rename files and then run "git add -u" does it take care of all the file changes in the git db?
17:44 piling joined #git
17:44 thiago type git status and it will tell you
17:45 vuoto joined #git
17:45 thiago but, no, it won't catch everything
17:45 thiago new files won't be added
17:45 Mixologic joined #git
17:45 daey ty
17:46 raijin joined #git
17:47 ASOLAr joined #git
17:48 bronson joined #git
17:50 mohabaks joined #git
17:52 subhojit777 joined #git
17:53 terminal_echo joined #git
17:53 vuoto joined #git
17:54 multi_io joined #git
17:54 sr___ joined #git
17:54 Torrone_1 joined #git
17:55 _ikke_ git add -A would
17:55 trolling42 joined #git
17:56 Mixologic joined #git
17:59 trolling42 guys anyone know the difference between committing with -m flag and writing commit message in an editor ?
18:00 _ikke_ trolling42: There is no difference, except that writing good commit messages is easier in the editor
18:00 trolling42 _ikke_, Does it make a difference re index or what info is stored ?
18:00 _ikke_ no
18:01 trolling42 _ikke_, thanks
18:01 garethdaine joined #git
18:02 Mobutils joined #git
18:06 subhojit_ joined #git
18:07 xender left #git
18:08 Torrone joined #git
18:09 MattMaker joined #git
18:10 Torrone joined #git
18:11 dmto joined #git
18:11 Prisoner-2460_1 joined #git
18:11 DrAwesome joined #git
18:12 LeBlaaanc joined #git
18:13 shaggycat joined #git
18:14 LeBlaaanc joined #git
18:15 fusionx86 joined #git
18:19 MattMaker joined #git
18:20 nidr0x joined #git
18:21 tvw joined #git
18:22 EvilPenguin joined #git
18:22 jimi_ joined #git
18:22 nidr0x joined #git
18:23 nitin joined #git
18:24 star_prone joined #git
18:25 star_prone joined #git
18:25 nidr0x joined #git
18:28 Rodya_ joined #git
18:28 marc_v92 joined #git
18:30 GavinMagnus joined #git
18:31 cdg joined #git
18:32 mohabaks joined #git
18:33 ahmedelgabri joined #git
18:33 ki0 joined #git
18:33 MattMaker joined #git
18:34 Stace joined #git
18:34 Stace If I have cherry picked a commit on a branch to my master, and then I rebase the branch, what happens to the commit?
18:34 Stace wouldn't it technically be empty?
18:35 _ikke_ Stace: Yeah, git would skip those automatically
18:35 Stace ok, so it would just disappear?
18:35 osse si senor
18:35 Stace ok, thanks
18:35 stomplee joined #git
18:36 Stace oh, I see, I just did it
18:36 Stace it just didn't apply the commit in question
18:38 rath joined #git
18:39 moneylotion joined #git
18:39 ahmed_elgabri joined #git
18:40 Orion3k joined #git
18:40 subhojit777 joined #git
18:41 texasmynsted joined #git
18:41 subhojit777 joined #git
18:42 subhojit777 joined #git
18:43 texasmynsted what do you recommend for using git locally for a project that is managed in svn?  I see there are at least two "solutions" but I am looking for this simplest way that I can use git locally, then when I am ready to "merge my changes" I would do something to merge them into the svn repo.
18:43 grawity git-svn seems to be the most popular; it can commit back
18:43 _ikke_ git-svn
18:43 _ikke_ It uses rebase to keep history flat
18:43 subhojit777 joined #git
18:44 mizu_no_oto joined #git
18:44 texasmynsted So I could safely use that and not bring any git artifacts into the svn repo?
18:45 texasmynsted I guess I need to read up on how that works
18:45 subhojit777 joined #git
18:46 LionsMane joined #git
18:49 MattMaker joined #git
18:50 zacts joined #git
18:52 editshare-ericf joined #git
18:55 cyan__ joined #git
18:56 Torrone joined #git
18:57 xaviergmail joined #git
18:59 Mixologic joined #git
18:59 MattMaker joined #git
19:00 mingrammer joined #git
19:00 a3Dman_ joined #git
19:01 romerocesar joined #git
19:02 Stace` joined #git
19:04 byteflame joined #git
19:05 GenZai joined #git
19:05 satifant joined #git
19:07 kfunk joined #git
19:09 MattMaker joined #git
19:11 moneylotion joined #git
19:12 robotroll joined #git
19:12 boombatower joined #git
19:13 johnmilton joined #git
19:13 an_ joined #git
19:14 an_ joined #git
19:14 marc_v92 joined #git
19:15 bronson joined #git
19:18 JamesBaxter joined #git
19:18 pasky joined #git
19:19 MattMaker joined #git
19:20 raijin joined #git
19:21 elementalest joined #git
19:23 rbern joined #git
19:25 krs93_ joined #git
19:27 sandstrom joined #git
19:29 Xizor joined #git
19:30 Electrometro joined #git
19:31 Rodya_ joined #git
19:32 Rodya__ joined #git
19:34 Torrone joined #git
19:34 johnmilton joined #git
19:35 nitin joined #git
19:37 Duder9999 joined #git
19:38 cbreak git svn works
19:38 cbreak the only git artefacts you'll bring in are special meta data lines in the commits
19:39 normalra joined #git
19:40 bronson joined #git
19:41 normalra hello, how may i force git clone to return a non-zero exit code when an "empty" repo has been cloned (which does invoke a warning)?
19:42 the193rd joined #git
19:42 ArchNoob joined #git
19:42 durham_ joined #git
19:42 Seveas normalra: as that is not an error, a non-zero exitcode would be wrong
19:43 normalra Seveas: i understand. how may i check whether the cloned repo is an "empty" repo?
19:43 cbreak normalra: in an empty repo, HEAD is pointing to an unborn branch
19:43 Seveas git rev-parse HEAD
19:43 cbreak it's a symbolic ref that doesn't exist
19:43 Seveas if that returns nonzero, the repo is empty (or severely broken)
19:43 normalra cbreak: Seveas thank you :)
19:44 cbreak it's the same state that you get in if you git checkout --orphan blah
19:44 MattMaker joined #git
19:44 cbreak so it doesn't mean that the repo is empty
19:44 Seveas cbreak: hmm, true, or that. Not quite severely broken
19:44 the193rd Maybe a noobish question: I have a LaTeX repo with 2 branches of very different section order. What would be the best workflow to integrate new content into both branches? (knowing that I work on a file which has in fact the same order). I don't want to end up with the same structure in both branches when merging
19:44 cbreak but if you have this state after a clone, then it was likely empty
19:44 Seveas cbreak: could it be anything but empty?
19:44 normalra cbreak: yeah, the test will always be invoked right after a clone
19:45 cbreak Seveas: HEAD on the remote could point to something stupid
19:45 Seveas ah, yes
19:45 osse ! { git clone ... 2>&1 | grep -q 'You appear to have cloned an empty repository'; }
19:45 osse ez
19:45 Seveas so, git-for-each ref giving no output, combined with rev-parse HEAD failing would be a tad more robust
19:45 cbreak yes
19:46 * Seveas slaps osse
19:46 dsantiag_ joined #git
19:46 Seveas I think that's the second time today I had to do that...
19:46 osse :(
19:46 normalra osse: The Shell Way :I
19:47 Seveas actually, the thing that triggers the warning doesn't even look at HEAD, just at the amount of refs
19:48 osse normalra: the { } aren't needed, by the way
19:48 Seveas if (refs) { ... } else { warning(_("You appear to have cloned an empty repository.")); }
19:48 pasky joined #git
19:49 Repox joined #git
19:49 MattMaker joined #git
19:49 osse Seveas: here's something to dig into. I was never satified with the conclusion we reached earlier.  What the hell does ''git diff A B C D E...'' do?
19:50 Seveas normalra: so the most correct thing to do is if [ -z "$(git ls-remote origin)" ]; then echo 'Empty repo!'; done
19:50 osse seems to do diff A...B; diff A...C; diff A...D, etc. etc. but combine it all into one huge diff
19:51 Seveas osse: hmm, that shold be equivalent to git diff A B -- C D E, no?
19:51 garethdaine joined #git
19:51 osse that's what I thought
19:51 normalra Seveas: thank you, i assume `git rev-parse HEAD` is a no-go?
19:51 normalra osse: thanks, i'll see whether i end up using the shell :)
19:51 Seveas osse: is the ... part of your actual command?
19:52 Seveas normalra: git rev-parse HEAD isn't a no-go, but it's not what git actually does to check whether the repo is empty
19:53 Seveas osse: git diff HEAD HEAD^ HEAD^^ shows the HEAD HEAD^ diff with doubled +/- prefixes
19:54 Seveas HEAD HEAD^ HEAD^^ HEAD^^^ causes triple prefixes
19:54 Seveas this smells buggy
19:54 osse Seveas: https://github.com/git/git/blob/​master/builtin/diff.c#L449-L468
19:54 osse i think it's *kind of* on purpose
19:55 mingrammer joined #git
19:55 Seveas it triggers the --cc behaviour
19:55 Seveas so it is attempting to construct merges
19:56 Seveas https://dpaste.de/EDaj/raw
19:57 normalra Seveas: well, it does seem like `git ls-remote origin` is simple too.
19:57 Seveas osse, yeah, it's doing a merge of B C D E into A
19:58 Seveas so this does seem intentional
19:58 osse hmmm
19:58 osse Interesting
19:58 Seveas but the synopsis in the manpage doesn't list this option
19:59 osse this doesn't either :D https://github.com/git/git/blob/​master/builtin/diff.c#L269-L290
19:59 Khisanth joined #git
19:59 MattMaker joined #git
19:59 osse N = 0, 1 or 2. Other cases are errors
20:00 docnvk joined #git
20:03 theverbg_ joined #git
20:05 cdg joined #git
20:05 docnvk joined #git
20:07 energizer joined #git
20:08 energizer I erased some large files in my repo history, now when I try to push to github it says your branch and origin/dev have diverged. How can I deal?
20:09 thiago energizer: are you aware that you are !rewriting history?
20:09 gitinfo energizer: [!rewrite] Rewriting public history is not recommended. Everyone who has pulled the old history will have to do work (and you'll have to tell them to), so it's infinitely better to just move on. If you must, you can use `git push --force-with-lease <remote> <branch>` to force (and the remote may reject that, anyway). See http://goo.gl/waqum
20:09 energizer thiago: ya the purpose was because the files were really big, so rsync wouldnt permit cloning the repo
20:10 thiago energizer: do you understand that everyone who has cloned it will need to delete and re-clone?
20:10 thiago if you understand that, use the force
20:10 thiago git push -f
20:10 energizer thiago: i don't think i actually realized that, but i don't think anybody's cloned this one
20:11 energizer so if i'm sure i want to do this, do git push -f?
20:11 ki0 joined #git
20:11 thiago add -f to the push command you were using
20:11 ash_workz joined #git
20:12 energizer ok, is there a way to make sure it's the case that nobody's cloned it?
20:12 thiago no
20:12 energizer from github
20:12 thiago you know your team
20:12 energizer ok i will ask them
20:14 raijin joined #git
20:14 MattMaker joined #git
20:14 Impaloo thiago: ppl wont have to re-clone the entire repo
20:14 thiago Impaloo: it's easier to explain this way
20:15 energizer Impaloo: what is true instead
20:15 thiago Impaloo: too many conditions on what to do
20:19 choki joined #git
20:21 xnrand joined #git
20:23 Torrone_1 joined #git
20:23 mooru joined #git
20:23 anonymuse joined #git
20:24 GavinMagnus left #git
20:26 EY joined #git
20:26 daniel-wtd joined #git
20:26 fusionx86 joined #git
20:27 LionsMane joined #git
20:28 pasky joined #git
20:29 hackel joined #git
20:29 byteflame joined #git
20:30 energizer I have a 144MB .pack file. Is that normal?
20:32 Seveas if the entire history of the repo, in delta-compressed and gzipped format, could be that size, then it's normal
20:32 hackel I've got a project with 17 other repositories cloned in a subdirectory.  Is there a way to convert these to submodules automatically, without having to look up the repository URL for each one and add it manually?
20:33 janx joined #git
20:33 maxorator joined #git
20:33 _ikke_ hackel: There is nothing built-in to do that
20:33 Anupkumar joined #git
20:33 kwmiebach joined #git
20:33 Mixologic joined #git
20:34 _ikke_ hackel: basically it's matter of add those directories, and generating a .gitmodules file
20:34 lebster joined #git
20:34 stealthii joined #git
20:34 hackel _ikke_: Ah, damn, okay.  So probably best to either suck it up and do it manually or write a bash script to read all the .git/config files?
20:35 huhlig joined #git
20:36 _ikke_ yes
20:36 nwkj86 joined #git
20:36 bbhoss joined #git
20:36 lambofgud joined #git
20:37 pizzaops joined #git
20:37 devhost joined #git
20:38 freimatz joined #git
20:39 hackel Cool, thanks.
20:39 MattMaker joined #git
20:41 yoyoyo___ joined #git
20:41 ahmed_elgabri joined #git
20:42 l4v2 joined #git
20:42 zarel joined #git
20:42 rfv joined #git
20:43 vassagus joined #git
20:44 yoh I wondered -- what would be the easiest way to deploy most recent git release within travis precise environment?
20:44 cstrahan joined #git
20:46 Seveas yoh: build from source. Or find a way to not need it, which is likely easier :)
20:46 danslo_ joined #git
20:46 Spec-Chum joined #git
20:46 Lowell joined #git
20:47 pfg joined #git
20:47 fahadash joined #git
20:49 cloudcell_ joined #git
20:49 moneylotion joined #git
20:52 NeXTSUN joined #git
20:52 yoh Saveas: the purpose is to react to changes in git behavior in upcoming release before release hits the wild.  so can't really "not needing it" by design ;)  ok -- I will consider build from source. cheers
20:55 anttim joined #git
20:56 mizu_no_oto joined #git
20:56 monoprotic joined #git
20:56 monoprotic joined #git
20:57 paperziggurat joined #git
20:57 vassagus joined #git
20:57 ahmed_elgabri joined #git
20:58 Gsham joined #git
20:59 martin_g joined #git
21:00 barq joined #git
21:01 anti_TORTURE joined #git
21:01 durham joined #git
21:02 aendrew joined #git
21:03 romerocesar joined #git
21:03 sssilver_ joined #git
21:04 ulkesh joined #git
21:06 lambofgud left #git
21:06 dreiss joined #git
21:06 ramsub07 joined #git
21:07 AnarchyAo joined #git
21:07 durham joined #git
21:07 ramsub07 hello. I have rebased my branch with master a few days back. How do i get the summary of all the commits from master applied to my branch post rebase ?
21:08 star_prone joined #git
21:08 Lewix joined #git
21:08 osse ramsub07: try git log branch@{1}..master
21:08 yoh left #git
21:09 MattMaker joined #git
21:10 sebhoss joined #git
21:11 ramsub07 osse: doesn't show anything
21:12 osse ramsub07: git reflog branch, find the entry before the rebase, git log branch@{4}..master or something
21:12 unholycrab joined #git
21:13 sssilver joined #git
21:13 spudowiar joined #git
21:13 cory_ joined #git
21:13 mischat joined #git
21:13 joseph_arnstein joined #git
21:14 dsantiago joined #git
21:14 yoan joined #git
21:16 spudowiar joined #git
21:24 GenZai joined #git
21:24 vassagus_ joined #git
21:29 durham joined #git
21:30 rubyonrailed joined #git
21:30 mikeric joined #git
21:31 a10c joined #git
21:34 TinFury joined #git
21:38 sagerdearia joined #git
21:39 alansaul joined #git
21:39 Gurkenglas joined #git
21:39 jstein__ joined #git
21:39 MattMaker joined #git
21:39 Gurkenglas Will some command allow me to play around with the file system as it would be if I merged two branches together? As git checkout does for commits.
21:40 nivag joined #git
21:42 osse Gurkenglas: git checkout -b tmp branch1; git merge branch2; play play play; git checkout -; git branch -D tmp
21:42 osse Gurkenglas: short answer: no
21:42 blackwell joined #git
21:42 mingrammer joined #git
21:43 andlabs joined #git
21:44 cyan__ joined #git
21:46 dsdeiz joined #git
21:48 UTAN_dev joined #git
21:49 Darren_ joined #git
21:53 EisNerd_ joined #git
21:55 UTAN_dev_ joined #git
21:55 raijin joined #git
21:56 xaviergmail joined #git
21:57 delmicio joined #git
21:58 jimmt joined #git
21:58 ahmed_elgabri joined #git
21:58 delmicio hola
21:58 xnrand joined #git
21:58 delmicio does anobody knows why isnt working this command: git archive HEAD $(git diff --name-only --diff-filter=ACMRT HEAD~..HEAD) | tar -czf parche.tgz --mode=0644
21:59 jimmt git is acting like a folder is in gitignore when I don't see any relevant definition in any of the .gitignore files under the root
21:59 osse jimmt: git check-ignore -v path/to/dir
22:00 Puffball joined #git
22:00 anonymuse joined #git
22:01 jimmt seems like it's only some files including the new ones that are there and ignored
22:01 osse check-ignore -v path/to/specific/file then
22:01 jimmt my global gitignore was doing this
22:01 jimmt weird
22:02 jimmt something put a global gitignore in my Documents directory that ignored all DLL files, wow
22:02 UTAN_dev joined #git
22:03 kadoban Doesn't sound like something you'd usually want to track.
22:04 jimmt they're how dependencies are stored in a c# project
22:04 jimmt for one example
22:04 realz__ joined #git
22:05 UTAN_dev_ joined #git
22:05 Derperperd joined #git
22:07 realz joined #git
22:08 osse why not use that github thing that keeps 5 of githubs servers running at full load 24/7 ?
22:09 MattMaker joined #git
22:09 mikeric joined #git
22:10 spudowiar1 joined #git
22:10 osse never mind. I thought it was for C# but it was for something else
22:10 str joined #git
22:10 leeN joined #git
22:10 delmicio I belive the tar option is the problem
22:10 delmicio git archive HEAD $(git diff --name-only --diff-filter=ACMRT HEAD~..HEAD) | tar -czf parche.tgz --mode=644
22:11 xelxebar joined #git
22:11 osse delmicio: git archive already produces a tar archive. pipe it directly to gzip instead of tar -z
22:12 delmicio but that way I can't use --mode=644
22:12 osse find another way then
22:12 osse in this case you don't even need git archive
22:12 delmicio that's why i'm trying to do ¬¬
22:12 UTAN_dev joined #git
22:13 acqyl joined #git
22:13 delmicio I come from git archive HEAD | gzip > parche.tgz
22:13 osse git diff --name-only --diff-filter=ACMRT HEAD~..HEAD | xargs tar -czf parche.tgz --mode=644
22:13 chrisshattuck joined #git
22:14 delmicio oh I see
22:14 rudi_s delmicio: umask?
22:14 rudi_s Or why exactly do you need 644?
22:14 vassagus joined #git
22:16 delmicio because then I upload the files to our server and extract
22:16 delmicio using my local files permissions
22:17 delmicio that's a problem for me
22:17 rudi_s delmicio: And why aren't the permissions already correct? Do you have a strict umask?
22:18 delmicio osse thanks you, I'm going your way
22:18 osse Ah. Fleedwood Mac!
22:19 btree joined #git
22:19 delmicio i'm on windows and not an expert with console handling
22:21 _sjs joined #git
22:24 WayToDoor joined #git
22:25 whomp_ joined #git
22:25 whomp_ hi, i'm looking at someone's pr and they moved large methods between files, with a few changes in there
22:26 whomp_ is there an easy way to do a diff only certain line ranges, e.g., git diff fileA[25-50] fileB[60-85]
22:28 thiago joined #git
22:29 l4v2 joined #git
22:29 xunie-laptop joined #git
22:35 liquidjamm joined #git
22:37 mingrammer joined #git
22:38 ASOLAr joined #git
22:38 btree left #git
22:38 red_ joined #git
22:39 liquidjamm joined #git
22:39 MattMaker joined #git
22:40 Noldorin joined #git
22:48 pijiu joined #git
22:48 anonymuse joined #git
22:54 MattMaker joined #git
22:59 MattMaker joined #git
22:59 sssilver_ joined #git
23:00 ghoti joined #git
23:00 ahmed_elgabri joined #git
23:01 rubyonrailed joined #git
23:03 romerocesar joined #git
23:07 xaviergmail joined #git
23:08 bronson joined #git
23:08 zacts joined #git
23:15 Lewix joined #git
23:15 huhlig joined #git
23:15 devhost joined #git
23:15 durham joined #git
23:15 cloudcell_ joined #git
23:18 sebhoss joined #git
23:18 kwmiebach joined #git
23:18 MattMaker joined #git
23:20 lebster joined #git
23:20 Mixologic_ joined #git
23:20 tvw joined #git
23:22 moneylotion joined #git
23:24 MattMaker joined #git
23:25 zacts joined #git
23:26 ulkesh joined #git
23:27 madewokherd joined #git
23:28 pfg joined #git
23:29 thiago joined #git
23:29 Ahmed90 joined #git
23:29 ffabi joined #git
23:29 rfv joined #git
23:29 krs93 joined #git
23:30 janx joined #git
23:31 Electrom_ joined #git
23:31 Anupkumar joined #git
23:31 beyertech joined #git
23:33 martin_g joined #git
23:34 angelsl_ joined #git
23:34 MattMaker joined #git
23:34 [diecast]_ joined #git
23:35 mikeh__ joined #git
23:35 b1tchcakes joined #git
23:37 moneylotion joined #git
23:37 Lowell joined #git
23:37 vcoinminer joined #git
23:38 timwis joined #git
23:41 aendrew joined #git
23:41 pizzaops joined #git
23:41 stealthii joined #git
23:42 bbhoss joined #git
23:42 imdsm_ joined #git
23:43 hazelux joined #git
23:44 MattMaker joined #git
23:45 joseph_arnstein joined #git
23:46 Jellyg00se joined #git
23:47 sargas joined #git
23:51 MaddHatter joined #git
23:52 skylite joined #git
23:52 MaddHatter Anybody aware of a filter-branch implementation capable of removing parts of a tree but that doesn't require a checkout/working dir?
23:52 Darren_ joined #git
23:53 MaddHatter file system operations are slow, and most of the i/o filter-branch does isn't really necessary for what I'm trying to do
23:53 bremner you know about index-filter, right?
23:53 MaddHatter bremner: I don't think so... I'll look
23:53 anttim joined #git
23:54 MaddHatter ah, perfect
23:58 Manuel15 joined #git
23:58 Manuel15 joined #git

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