Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2017-02-20

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

All times shown according to UTC.

Time Nick Message
00:00 kiltzman joined #git
00:02 chipotle joined #git
00:02 hahuang61 joined #git
00:02 dsa joined #git
00:03 kiltzman joined #git
00:03 kiltzman joined #git
00:04 kiltzman joined #git
00:06 kiltzman joined #git
00:07 ozmage joined #git
00:07 DolpheenDream joined #git
00:07 felixjet joined #git
00:08 kiltzman joined #git
00:08 kiltzman joined #git
00:09 chipotle joined #git
00:10 harv joined #git
00:12 MikeChernev joined #git
00:12 Neobenedict joined #git
00:15 Neobenedict i have a directory full of files: ./stuff, and a remote origin that has those files like this: ./git/stuff
00:15 Neobenedict how do I connect the two
00:16 bennyZ joined #git
00:16 matoro joined #git
00:18 ozmage joined #git
00:21 rivarun joined #git
00:22 Neobenedict nvm
00:23 thiago rename one of the directories
00:23 Neobenedict my bad, it put it in the right directory after all
00:26 mehola joined #git
00:26 mehola joined #git
00:30 anuxivm left #git
00:30 MattMaker joined #git
00:37 chipotle joined #git
00:41 kiltzman joined #git
00:42 DIDAVISION joined #git
00:45 col joined #git
00:46 stevenbenner joined #git
00:46 IRCFrEAK joined #git
00:46 clickjac_ joined #git
00:46 MattMaker joined #git
00:47 IRCFrEAK left #git
00:48 ozmage joined #git
00:48 cdown joined #git
00:51 Electrometro joined #git
00:53 Kaisyu joined #git
00:55 ozmage_ joined #git
00:57 mehola joined #git
01:01 matoro joined #git
01:03 LiftLeft joined #git
01:04 Nilesh_ joined #git
01:04 LiftLeft I added a directory to gitignore and I know I want it back in the git repo
01:04 thiago ok, remove it from gitignore and git add it
01:05 LiftLeft that doesn't work
01:05 LiftLeft the directory isn't on bitbucket
01:06 pks joined #git
01:06 LiftLeft it tell me On branch master nothing to commit, working directory clean
01:07 qpdb joined #git
01:08 thiago it does work
01:08 thiago ls the directory. Does it have content?
01:08 cdown_ joined #git
01:09 LiftLeft yes
01:09 LiftLeft thiago: no it doesn't
01:10 LiftLeft it doesn't work
01:10 thiago what happens when you git add?
01:11 LiftLeft nothing
01:11 LiftLeft I tried rename it from public to public2 and back again but that didn't help
01:12 thiago does the directory have any content that is not another directory? Any files?
01:13 LiftLeft it has other directories in it and those directory have files in them
01:13 dsdeiz joined #git
01:13 dsdeiz joined #git
01:13 thiago git add one of those files
01:14 LiftLeft fatal: Pathspec 'public/common/bison.js' is in submodule 'public'
01:16 cqi joined #git
01:16 thiago there you go
01:16 thiago you've got your answer
01:16 LiftLeft what?
01:16 thiago public is a submodule
01:16 thiago you can't add files in a submodule
01:16 LiftLeft ok
01:16 thiago if you didn't know it was a submodule, it's probably because public/.git exists
01:17 LiftLeft ok
01:17 LiftLeft would I just like add to .gitignore public/.git
01:18 thiago rm -rf public/.git
01:19 LiftLeft Is there a way of having git ignore it?
01:19 LiftLeft cuz it was working before
01:19 LiftLeft for some reason
01:20 thiago the problem is not ignoring
01:20 thiago the problem is that public is a submodule while that .git exists
01:21 ozmage joined #git
01:22 LiftLeft I only need public/common in the repo
01:22 thiago rm -rf the rest
01:22 thiago especially the .git dir
01:25 DolpheenDream joined #git
01:27 CussBot joined #git
01:28 kexmex joined #git
01:32 eblip joined #git
01:34 nowhereman joined #git
01:35 ozmage joined #git
01:38 col joined #git
01:39 cdown_ joined #git
01:39 krillr joined #git
01:39 nettoweb joined #git
01:40 LiftLeft how do I revert git the repo to a previous commit, I don't care about anything after it
01:40 marianina8 joined #git
01:43 Ahmed90 joined #git
01:43 heroux joined #git
01:45 ozmage joined #git
01:48 mischat joined #git
01:49 pur3eval joined #git
01:51 gfixler joined #git
01:51 jameser joined #git
01:52 sz0 joined #git
01:52 ozmage joined #git
01:57 dreiss joined #git
01:57 chipotle joined #git
01:58 Lyqyd joined #git
02:00 nevyn LiftLeft: have you pushed?
02:00 nevyn if not... git reset --hard commitish
02:01 packet0 joined #git
02:02 aidalgol joined #git
02:02 DolpheenDream joined #git
02:03 hahuang61 joined #git
02:05 pks joined #git
02:06 Maroloccio_ joined #git
02:08 freedrull joined #git
02:09 rwb joined #git
02:10 Hello71 !fix
02:10 gitinfo [!fixup] 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!
02:11 ramsub07 joined #git
02:12 finalbeta joined #git
02:13 LiftLeft nevyn: I got it working
02:16 Rodya_ joined #git
02:18 shellbofh joined #git
02:23 gelei007 joined #git
02:24 Vampire0 joined #git
02:27 chipotle joined #git
02:29 ToBeCloud joined #git
02:30 Electrometro joined #git
02:31 Wulf4 joined #git
02:39 hhee joined #git
02:40 MattMaker joined #git
02:41 Oatmeal joined #git
02:42 solenodic joined #git
02:49 ilbot3 joined #git
02:49 Topic for #git is now Welcome to #git, the place for git help. Just git help, apparently. | Public logs at http://jk.gs/glog | First visit? Read: http://jk.gs/git | Current stable version: 2.11.0 | Getting "cannot send to channel"? /msg gitinfo .voice | This is serious business. https://madeitwor.se/git-jokes
02:49 mischat joined #git
02:50 rivarun joined #git
02:53 solenodic joined #git
02:57 gelei007 joined #git
02:59 yaewa joined #git
03:02 nettoweb joined #git
03:02 ResidentBiscuit joined #git
03:03 ojdo joined #git
03:04 pks joined #git
03:12 hexagoxel_ joined #git
03:14 nevodka joined #git
03:14 fstd joined #git
03:14 chipotle joined #git
03:16 fstd_ joined #git
03:17 ozmage joined #git
03:17 Vortex34 joined #git
03:18 pandeiro` joined #git
03:21 Forlorn joined #git
03:22 chachasmooth joined #git
03:24 dreiss joined #git
03:25 akkad joined #git
03:28 pur3eval joined #git
03:35 Forlorn joined #git
03:41 MattMaker joined #git
03:44 lb1c joined #git
03:45 trip joined #git
03:47 ecuanaso joined #git
03:49 ozmage joined #git
03:50 mischat joined #git
03:51 RM982 joined #git
03:52 madgoat joined #git
03:53 RM982 good evening, quick question, let's say I    "git add ."   then I remove a couple individual and I add them to .gitignore
03:54 RM982 If I do "git add ." again, will GIT track those files I gitignored?
03:54 netj joined #git
03:54 kadoban RM982: Not if they aren't already tracked, and if they weren't in the index at the time.
03:55 sunri5e joined #git
03:55 roygbiv joined #git
03:56 d0nn1e joined #git
03:57 RM982 kadoban: sorry i don't follow you, I add the whole dir lets say /www, I git remove /www/a.php and also add the path to .gitignore, what happens the next time I run "git add /www"
03:57 madgoat left #git
03:57 RM982 will /www/a.php be tracked or not?
03:58 b3kN joined #git
03:58 kadoban RM982: No
03:59 RM982 kadoban: perfect, thank you
03:59 kadoban Anytime
03:59 aspiers joined #git
03:59 a_thakur joined #git
04:00 dermoth joined #git
04:01 xaviergmail_ joined #git
04:02 Emperor_Earth joined #git
04:02 sunrunner202 joined #git
04:02 sunrunner202 first time PR userhere
04:02 sunrunner202 i've cloned the main repository
04:03 sunrunner202 but how to I get this (https://github.com/wired/colloquypush/pull/46) PR into my working folder?
04:03 kadoban You're trying to fetch it locally, with command line git for instance you mean?
04:03 hahuang61 joined #git
04:04 sunrunner202 kadoban, yea, from the cli
04:04 sunrunner202 did a git clone <url> already
04:05 kadoban https://help.github.com/articles/c​hecking-out-pull-requests-locally/ is the official answer, but I prefer: https://github.com/wired/colloquypush/pull/46
04:05 kadoban If you just want to do it once, there's a single command for that, but I like being able to just get any of them easily typically, which is what the latter allows.
04:07 kadoban Oh, haha, one of those wasn't the right url of course, uhm sec.
04:08 kadoban https://gist.github.com/piscisaureus/3342247 that was supposed to be the second url
04:10 kiltzman joined #git
04:11 kiltzman joined #git
04:12 kiltzman joined #git
04:13 kiltzman joined #git
04:13 kiltzman joined #git
04:14 kiltzman joined #git
04:14 sunri5e joined #git
04:14 sbulage joined #git
04:16 Rodya_ joined #git
04:22 hahuang61 joined #git
04:22 ozmage joined #git
04:24 chipotle joined #git
04:30 miczac joined #git
04:32 robotroll joined #git
04:33 urbanmonk joined #git
04:34 chipotle_ joined #git
04:36 urbanmonk joined #git
04:40 marianina8 joined #git
04:45 Doginal joined #git
04:46 matoro joined #git
04:48 govg joined #git
04:49 venmx joined #git
04:50 madprops joined #git
04:50 madprops can't seem ignore a file that is already commited. .. it contains a default password in the repo, and i want to change it locally and then ignore it. i added the file in a line in .gitignore "config.json" (lines are not separated by commas or anything) then I did git reset HEAD and commit again but it keeps updating the file in the repo
04:52 WayToDoor joined #git
04:52 col joined #git
04:52 madprops this is my gitignore http://gsick.com/paste/198/
04:53 cqi joined #git
04:53 madprops i tried prepending / to config.json too
04:53 kasual joined #git
04:56 ayogi joined #git
04:58 Cabanoss- joined #git
05:03 sunrunner202 i'm geetting a "fatal: Refusing to fetch into current branch refs/heads/master of non-bare repository" error trying to follow that guide
05:05 a_thakur joined #git
05:09 a_thakur_ joined #git
05:11 MattMaker joined #git
05:12 cqi joined #git
05:18 jmesmon joined #git
05:18 MineCoins joined #git
05:20 cdown_ joined #git
05:21 sunrunner202 joined #git
05:26 pur3eval joined #git
05:28 Murr joined #git
05:33 Rodya_ joined #git
05:35 cdown joined #git
05:38 a_thakur joined #git
05:38 cdown_ joined #git
05:40 xall joined #git
05:41 Doginal joined #git
05:41 basiclaser joined #git
05:42 a_thakur joined #git
05:46 a_thakur joined #git
05:51 ozmage joined #git
05:51 pks joined #git
05:51 solenodic joined #git
05:58 gelei007 joined #git
05:58 ash_workz joined #git
05:58 SwiftMatt joined #git
05:59 SantaClauze joined #git
06:04 xall how can i match a filetype (or pattern) with a git diff. something like
06:04 xall git diff . ../../../src/ch05-routing/temperature-by-zip --html
06:05 dec0n joined #git
06:05 z8z joined #git
06:06 ResidentBiscuit joined #git
06:07 Doginal joined #git
06:07 blahdodo joined #git
06:07 xaviergmail_ joined #git
06:11 ych joined #git
06:12 mohabaks joined #git
06:12 mohabaks left #git
06:13 peepsalot joined #git
06:14 howdoi joined #git
06:15 rscata joined #git
06:16 MineCoins joined #git
06:17 armyriad joined #git
06:19 lordjancso joined #git
06:21 gelei007 joined #git
06:21 overlord_tm joined #git
06:26 safe joined #git
06:36 jnavila joined #git
06:37 sarri joined #git
06:37 sarri joined #git
06:38 Rodya_ joined #git
06:39 zeroed joined #git
06:39 qt-x joined #git
06:40 dersand joined #git
06:41 jccn joined #git
06:42 gelei007 joined #git
06:45 Doginal joined #git
06:49 aidalgol joined #git
06:50 svm_invictvs joined #git
06:53 ShapeShifter499 joined #git
06:54 Junior joined #git
06:55 dminuoso joined #git
06:55 overlord_tm joined #git
06:57 ThomasLocke joined #git
06:57 ThomasLocke joined #git
07:00 zeroed joined #git
07:01 Raging_Hog joined #git
07:05 gelei007 joined #git
07:06 venmx joined #git
07:06 blahdodo joined #git
07:10 User458764 joined #git
07:10 pelegreno joined #git
07:13 hanthings joined #git
07:14 MattMaker joined #git
07:14 JeroenT joined #git
07:16 text1 joined #git
07:20 EV3RYDAYPR0GRESS joined #git
07:20 toglo joined #git
07:23 EvaLuAtor joined #git
07:24 finalbeta joined #git
07:24 EvaLuAtor Hello. I'd like to ignore everything beneath bin/ but keep bin/Debug/*.exe files. I've tried this http://sprunge.us/ReYa but git doesn't let me add the exe files (without -f). What should be proper rules for this situation?
07:26 cbreak EvaLuAtor: you shouldn't commit derived files anyway
07:27 cbreak EvaLuAtor: it's extremely bad form and can cause all kinds of nasty issues
07:27 EvaLuAtor cbreak: I'm fully aware of it, but I need to do it anyway
07:27 cbreak then add -f is ok, isn't it?
07:27 chele joined #git
07:27 cbreak once a file is tracked, you don't have to use -f again
07:27 chele joined #git
07:29 EvaLuAtor cbreak: thank you, you're right
07:30 EvaLuAtor but I'm still curious how to unignore the files several levels beneath the ignored directory
07:30 MattMaker joined #git
07:30 EvaLuAtor for other situations than binaries
07:30 acetakwas joined #git
07:33 cbreak I think git doesn't even look into ignored directories
07:33 cbreak since they are ignored
07:33 cbreak so you can't unignore anything in them
07:36 ToxicFrog joined #git
07:38 basdb joined #git
07:39 Tobbi joined #git
07:40 jast you need these rules, I think: /bin/* - !/bin/Debug - /bin/Debug/* - !/bin/Debug/*.exe
07:41 EvaLuAtor jast: I've already tried these and it doesn't work
07:42 jast it's different from what you say you've tried
07:42 EvaLuAtor I've tried it too
07:42 EvaLuAtor However, if I keep only these two, it works: bin/Debug/* - !bin/Debug/*.exe
07:42 gitinfo [!binary] Storing binary files in git causes repo balloon, because they do not compress/diff well. In other words, each time you change a file the repo will grow by the size of the file. See !annex for some solutions
07:42 jast it's definitely not going to work with less than four rules, in any case
07:42 DistantS1ar left #git
07:43 Rodya_ joined #git
07:44 jast the details are important. if you use '/bin', it will fail, only '/bin/*' will do the right thing
07:45 jnavila joined #git
07:45 EvaLuAtor as I said, I've tried exactly what you wrote before asking the question, and git still treats bin/Debug/*.exe as ignored files
07:45 jast for the sake of completeness, I just tested it, byte for byte, and it works
07:46 EvaLuAtor well, to be honest, I didn't have there leading '/'s, I'll try with them...
07:46 jast here's my test repo: https://jk.gs/foorepo.tar.gz
07:47 jast and it works without the leading slashes, too
07:48 jast here's my second test repo: https://jk.gs/foorepo2.tar.gz
07:48 EvaLuAtor jast: you're right, it works, I didn't what I did wrong, I had it commented there, exactly what you wrote but something was different
07:48 EvaLuAtor s/didn't/don't know/
07:49 EvaLuAtor sorry and thank you for your time and help
07:49 jast as long as it works :)
07:50 zeroed joined #git
07:50 zeroed joined #git
07:51 dreiss joined #git
07:52 aw1 joined #git
07:53 User458764 joined #git
07:53 Masber joined #git
07:55 jozwior joined #git
07:56 jagob joined #git
07:56 xaviergmail_ joined #git
07:56 xall joined #git
08:02 rossome joined #git
08:03 theoceaniscool joined #git
08:06 zefferno joined #git
08:07 roelmonnens joined #git
08:10 pur3eval joined #git
08:12 King_Hual joined #git
08:14 kiltzman joined #git
08:14 kiltzman joined #git
08:15 Tobbi joined #git
08:16 MineCoins joined #git
08:16 buckowski joined #git
08:18 kiltzman joined #git
08:19 kiltzman joined #git
08:19 elect joined #git
08:20 kiltzman joined #git
08:21 kiltzman joined #git
08:22 kiltzman joined #git
08:24 ozmage joined #git
08:24 dimi1947 joined #git
08:26 djb-irc joined #git
08:30 def_jam joined #git
08:34 armyriad joined #git
08:35 venmx joined #git
08:38 afuentes joined #git
08:38 zeroed joined #git
08:41 myme left #git
08:41 flobin joined #git
08:42 JeroenT_ joined #git
08:42 garethdaine joined #git
08:42 eb0t joined #git
08:43 dimi1947 I tried to squash 2 commits at rebase and take a look a message I got:https://1drv.ms/i/s!AjCBrCpLQye-g1Gxt3C4QegInUje
08:44 Balliad joined #git
08:45 djb-irc joined #git
08:45 Rodya_ joined #git
08:46 a_thakur joined #git
08:46 osse dimi1947: that's a vim error message. do you have a different vim open somewhere, editing a config message?
08:47 osse if not it's probably safe to go for D
08:47 dimi1947 osse:you mean deleting it?
08:47 Ruler- That's one suspicious url btw
08:47 Ruler- asking my password
08:48 osse dimi1947: yes
08:48 sunrunner202 joined #git
08:48 dimi1947 osse:ok..thanks i will do so
08:51 MattMaker joined #git
08:51 j7k6 joined #git
08:51 m4sk1n joined #git
08:52 MattMaker joined #git
08:52 dimi1947 osse:one last question....where I will type exactly the message for the new commit in the screen I have in front of me:https://1drv.ms/i/s!AjCBrCpLQye-g1JEisKEF82Btzmo
08:52 lss8 joined #git
08:52 osse At the top, just like normal
08:52 MattMaker joined #git
08:53 MattMaker joined #git
08:55 nd_ joined #git
08:55 Andrew_K joined #git
08:58 oskarkv joined #git
08:58 DieguezZ joined #git
09:00 govg joined #git
09:00 dimi1947 osse:take a look at the error I got after saving the message for the new commit(the result of the squash):https://1drv.ms/i/s!AjCBrCpLQye-g1OJ-hQc4pA6In4u
09:01 _ng joined #git
09:01 osse git doesn't like your commit message
09:03 Junior joined #git
09:03 dc5ala joined #git
09:04 solenodic joined #git
09:04 hhee2 joined #git
09:06 eb0t joined #git
09:06 MineCoins joined #git
09:07 Dumblez0r joined #git
09:07 dimi1947 osse:why?
09:08 osse I don't know
09:09 rwp It wasn't the commit message.  It was an error invoking the editor 'vi'.
09:09 dimi1947 rwp:how do I resolve it?
09:10 rwp That depends upon why it is having the error.
09:10 osse oh right.
09:10 osse i thought it printed that message anyway
09:10 rwp Did you try invoking "vi" yourself and seeing what error happens when you do it?
09:11 dimi1947 rwp:I tried second time and it worked...
09:11 a_thakur joined #git
09:12 sunrunner20_ joined #git
09:13 dimi1947 is there any chance that the error was due to the commit message being exactly the same with another commit?
09:17 anon861 what is the current state with FF's vimperator support?
09:17 marcogmonteiro joined #git
09:18 mikecmpbll joined #git
09:18 anon861 wrong channel, sorry
09:18 anon861 left #git
09:19 c0c0 joined #git
09:19 zeroed joined #git
09:19 zeroed joined #git
09:20 Snugglebash joined #git
09:21 Snugglebash joined #git
09:21 Snugglebash joined #git
09:22 rwp There is no requirement that commit messages be unique. So no.
09:22 * rwp bids everyone good night
09:22 suttonwilliamd joined #git
09:23 MineCoins joined #git
09:27 mischat joined #git
09:30 mischat joined #git
09:31 eblip joined #git
09:32 acetakwas joined #git
09:34 seni joined #git
09:35 mkoskar joined #git
09:36 seni joined #git
09:37 Atrumx joined #git
09:38 jelle[m] left #git
09:40 pbrewczynski joined #git
09:42 suttonwilliamd_ joined #git
09:42 nettoweb joined #git
09:45 bruce_lee joined #git
09:48 neeri joined #git
09:49 Rodya_ joined #git
09:50 acetakwas joined #git
09:52 notebox joined #git
09:53 s1scha joined #git
09:54 ferr1 joined #git
09:55 Tobbi joined #git
09:55 dmc joined #git
09:57 TomyWork joined #git
09:58 clmsy joined #git
10:01 flaviodesousa joined #git
10:03 DaveTaboola joined #git
10:05 JeroenT joined #git
10:06 JeroenT joined #git
10:07 garethdaine joined #git
10:07 garethdaine joined #git
10:07 spinningarrow[m] left #git
10:07 kikijiki[m] left #git
10:07 unilovelight[m] left #git
10:12 visual joined #git
10:12 visual heya
10:13 visual after accepting a merge request on gitlab, in a way that branch->master
10:13 visual do i just checkout master on the local machine, and it will have all the updated changes?
10:13 xaviergmail_ joined #git
10:13 DaveTaboola joined #git
10:15 tobiasvl visual: no, you will need to checkout and then pull (or otherwise fetch + merge) from the gitlab remote too
10:15 acetakwas joined #git
10:16 Snugglebash joined #git
10:16 visual awesome possum, that worked
10:16 visual thanks tobiasvl
10:17 acetakwas joined #git
10:18 cliluw joined #git
10:19 irqq joined #git
10:22 chll_ joined #git
10:24 dcpc007 joined #git
10:24 ceprius_ joined #git
10:26 aleb If I run "git fetch origin refs/notes/a" the command returns error code 128 because the refs/notes/a refspec does not exist in the origin remote repo. Can I use the 128 code to reliably detect when this happens, is it used for other types of errors?
10:27 Tobbi joined #git
10:27 aleb I could also look into parsing the printed error message, but it does not seem reliable:
10:27 aleb fatal: Couldn't find remote ref refs/notes/a
10:27 aleb fatal: The remote end hung up unexpectedly
10:30 manuelschneid3r joined #git
10:38 tvw joined #git
10:40 theelous3 joined #git
10:41 Tobbi joined #git
10:41 theelous3 Hi. How can I pull up the diff between my local master and remote master?
10:41 manuelschneid3r joined #git
10:42 marcogmonteiro joined #git
10:43 Impaloo joined #git
10:43 moei joined #git
10:52 waits_ joined #git
10:52 waits_ Hello
10:52 gitinfo waits_: 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.
10:52 waits_ I have two commits and like to keep the last but delete the second to last. Easiest way to do it?
10:54 Rodya_ joined #git
10:55 garethdaine joined #git
10:55 MattMaker joined #git
10:55 serialoverflow joined #git
10:56 tobiasvl theelous3: git diff local_master..remote_master
10:56 clickjack joined #git
10:56 tobiasvl waits_: delete how? delete it completely from the history, which will involve rewriting history, which is not recommended if it has been pushed?
10:58 lu_na joined #git
10:59 JeroenT joined #git
10:59 aLeSD joined #git
10:59 aLeSD hi all
10:59 aLeSD is it possible to change the commit message of an old pushed commit ?
11:00 aleb aLeSD: you can if you really must, but it could be complicated
11:01 tobiasvl aLeSD: it involves a !rewrite of history
11:01 gitinfo aLeSD: 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
11:01 dcpc007 pushed commit modify is not recommanded and can be very complicated
11:02 waits_ tobiasvl: it's still on my localhost, I'm going to rewrite history and the order of the commits then delete the last commit
11:02 dcpc007 a solution we used here, was to be sure that no dev had local work/branch, and then delete repo and clone again after master dev do the push
11:02 tobiasvl waits_: OK, then it's fine. do an !interactive_rebase
11:02 gitinfo waits_: Interactive rebase sounds similar to rebase but has completely different abilities. It can do this to commits: change the order, squash some of them together, remove some, add random existing commits from other branches, split them, and more... it's very powerful. Documentation is in the section "Interactive Rebase" in 'man git-rebase'.
11:02 jagob joined #git
11:07 Rico left #git
11:08 clmsy joined #git
11:08 rnsanchez joined #git
11:12 mischat_ joined #git
11:12 Celelibi Hm, hey, the script sample for the pre-push hook is almost exactly what I need to prevenit pushing some commits.
11:12 masuberu joined #git
11:13 Celelibi I just have to change "WIP" with something like "nopush"
11:13 Tobbi joined #git
11:15 zefferno joined #git
11:15 metachr0n joined #git
11:16 blackwind_123 joined #git
11:19 nettoweb1 joined #git
11:21 CheckDavid joined #git
11:21 MineCoins joined #git
11:23 rchavik joined #git
11:25 gambl0re joined #git
11:25 gambl0re left #git
11:31 acetakwas joined #git
11:32 Tobbi joined #git
11:33 jameser joined #git
11:35 DaveTaboola joined #git
11:39 theoceaniscool joined #git
11:39 JeroenT joined #git
11:40 tvw joined #git
11:42 ffledgling joined #git
11:44 dminuoso I have a git submodule pointing at a local repository, how do I update this submodule when I commit a change in that repository?
11:44 a_thakur joined #git
11:44 ffledgling Hello, I'm facing an odd issue with a git repo. I clone this repo, and it has a directory in it called themes/blackburn. It appears to be a submodule, but when I run `git submodule status` I get an error that says |fatal: no submodule mapping found in .gitmodules for path 'themes/blackburn'|. I want to get rid of the submodule entirely and commit the as part of the repo-proper
11:45 ffledgling Any idea what I need to do? I can't seem to deinit the submodule and `git rm -f --cached themes/blackburn` does something, but I'm not quite sure what
11:47 dminuoso ffledgling: Is the submodule in .git/config?
11:47 visual left #git
11:48 ffledgling dminuoso: I don't see it there, no
11:48 rorro joined #git
11:49 osse ffledgling: the rm is correcty
11:49 osse s/y//
11:49 rafalcpp git describe returns which tag, if there are several tags on current commit?
11:49 eb0t joined #git
11:49 dminuoso ffledgling: You still need to remove your .git/modules/<name> then
11:49 achadwick joined #git
11:50 rafalcpp how to use git describe so that it will return the latest (by date) tag on current branch. Btw the tag should be signed, an current version should be tagged, I would also like to assert this is the case
11:50 ffledgling osse: okay, but where exactly is git storing information about the submodule in the first place? There's no .gitmodules file, nothing in .git/config and I don't see much git ls-tree
11:50 osse ffledgling: it's in the tree
11:50 ffledgling dminuoso: There's no .git/modules directory
11:51 osse what does git ls-tree HEAD parentdir  say?
11:51 ffledgling osse: is git ls-tree master:themes/blackburn the right way to look at it?
11:51 nettoweb joined #git
11:51 osse no
11:52 nettoweb1 joined #git
11:52 ffledgling osse: https://paste.fedoraproject.org/paste/DE1​n91HqEcxBOKyj9fyLhV5M1UNdIGYhyRLivL9gydE= (. being the root directory)
11:53 ffledgling Not sure if that's what you meant by parentdir
11:53 osse I mean themes
11:53 aleb Why are some strings used for die() calls in remote.c not translated?  https://www.irccloud.com/pastebin/vvrxNd2v/
11:54 ffledgling osse: https://paste.fedoraproject.org/paste/OZv​d8PcJG82MPCj5qQQwB15M1UNdIGYhyRLivL9gydE= ? It just says "commit"
11:54 osse ffledgling: that's what I was after. That means it's a submodule
11:54 ffledgling I see
11:55 osse that's how submodules are stored. You're just missing all that other stuff
11:55 ffledgling In case it were a regular directory, it would've said "tree" then?
11:55 osse yes
11:55 osse This is what happens when you do 'git add path/to/repo' without using git submodule add
11:55 ffledgling and commit it and push it?
11:55 MattMaker joined #git
11:55 a_thakur_ joined #git
11:55 osse hmm?
11:56 ffledgling So this is a fresh clone of a repo
11:56 ffledgling I haven't done any git add or similar yet.
11:56 osse No I mean that's how you ended up in this situation. Someone else added that submodule wrongly
11:56 ffledgling Right
11:57 ffledgling So to rid myself of it, I need to do a git rm -f ? That's it?
11:57 osse yep
11:57 eblip joined #git
11:57 nettoweb joined #git
11:57 suttonwilliamd joined #git
11:57 flobin left #git
11:57 Rodya_ joined #git
11:57 ffledgling Okay, when I actually run that I still see the folder and contents lying around, that's the expected behaviour then?
11:58 osse when you use --cached yes
11:58 ffledgling ah
11:58 ffledgling I still get a warning about Submodule changes to be committed
11:58 osse well, you need to commit also
11:59 ffledgling Indeed, I'm just a little confused about why git think it's a submodule I guess
12:00 ffledgling Hopefully a regular git commit works
12:01 clickjack joined #git
12:02 osse It thinks it's a submodule because it's stored the way submodules are
12:03 osse except for the lack of convenience you get by having it properly in .git/config, it IS a submodule
12:03 ffledgling osse: so essentially whenever I git a dir with a .git inside it, it will always be added as a submodule and not as a regular tree?
12:03 eblip joined #git
12:03 osse yes
12:03 ffledgling *`git add` a dir
12:04 mischat joined #git
12:04 osse imho that should produce a warning if a submodule isn't already added for that path, but alas
12:04 suy I've hit into this problem sometimes: git log some/submodule shows that the last commit on master touching the submodule making it -bbbbbb+aaaaaa. But the current state is bbbbbb. Somebody screwed during a merge conflict resolution. But how can I find it? I can't make git log print which commit made it bbbbb.
12:06 ffledgling osse: I would've thought git add would always do trees and submodules would require a submodule specific command (git submodule add or something), but then I find submodules fairly complex in general
12:07 osse ffledgling: git submodule add does all the .gitmodules and .git/config stuff (and the initial add). But 'git add path' is the correct way to make changes afterwards
12:07 ffledgling ah
12:08 ffledgling on a related note, what's the correct/best-practice way to have a third party repo as a dependency that one would like to pull from, but also maintain one's own patches atop it?
12:09 zefferno joined #git
12:09 ffledgling for example in this case, blackburn is a theme from github, but I'd like to modify some stuff and have my patches applied for personal use (assuming the burden of resolving conflicts and such is on me)
12:10 Tobbi joined #git
12:10 osse have your own fork hosted somewhere and then add that fork as a submodule
12:11 ffledgling and that fork is periodically `git pull --rebase`d ?
12:12 osse that part I don't have strong feelings about. do what feels right.
12:12 osse or least wrong.
12:12 ffledgling I see
12:13 DobrijUser|2 joined #git
12:15 aleb Why are some strings used for die() calls in remote.c not translated?  https://www.irccloud.com/pastebin/vvrxNd2v/  For example: die("Couldn't find remote ref %s", name);
12:20 YuGiOhJCJ joined #git
12:22 a_thakur joined #git
12:23 kiltzman joined #git
12:24 garethdaine joined #git
12:25 ahr3n joined #git
12:25 xaviergmail_ joined #git
12:25 navidr joined #git
12:29 acetakwas joined #git
12:31 bernardio joined #git
12:33 roelmonn_ joined #git
12:36 Tobbi joined #git
12:42 Akuli joined #git
12:42 Akuli can i amend into a commit i have pulled from github?
12:45 jagob joined #git
12:45 DolpheenDream joined #git
12:46 tocix joined #git
12:46 tocix Hello.
12:46 gitinfo tocix: 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.
12:47 King_Hual joined #git
12:47 tocix How can I add a file to git but then stop tracking changes for it so it will always be at the version when I added it?
12:48 Akuli just add it, git remembers what it was like when you added it
12:48 tocix Akuli:  if someone changes the file and commits it, I want git to ignore the changes.
12:48 tocix and keep the file at the version I added it
12:48 Akuli oh i see, then i have no idea :(
12:49 tocix Alternatively, how can I get git to automagically execute a script after I clone a repo?
12:50 Cavallari joined #git
12:50 tocix Not on the server.
12:50 osse Akuli: you can but it may not be a good idea.
12:50 osse Akuli: !rewrite
12:50 gitinfo Akuli: 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
12:51 Akuli osse, ok, i'll tar and netcat the repo to my other computer and keep working on it there
12:51 osse Akuli: is it your own commit? Are you the only developer?
12:51 Akuli yes and yes
12:52 osse Akuli: amend as much as you like
12:52 bernardio joined #git
12:52 Akuli ok
12:52 Akuli but won't it be looking weird on github?
12:52 osse no
12:52 osse it'll be like the original commit was never there
12:53 Akuli if someone goes to read it like "oh there that is" and then "i have a problem with this" and then "oh it got magically fix even though i checked out the same commit 0_o"
12:53 osse it will be a different commit
12:53 osse the checksum will be different.
12:53 Akuli oh
12:53 osse the new one replaces the old
12:53 Akuli and then that guy wondering won't be able to checkout the original commit?
12:54 osse Akuli: he will never know it was there
12:54 osse unless he found your repo before you amended, but cloned it after you did
12:54 osse unlikely
12:55 synthroid joined #git
12:55 Akuli i thought about cloning it before the amend and then after the amend and trying to checkout the state when he cloned it
12:55 Akuli unlikely to happen though, i'll amend
12:55 lektrik joined #git
12:55 Akuli or you know
12:55 * Akuli branches
12:55 osse yes, that will be screwy for that guy
12:55 osse but you said you're the only one so...
12:55 lu_na joined #git
12:56 bvtton joined #git
12:56 synthroi_ joined #git
12:56 MattMaker joined #git
12:58 cmalmber joined #git
12:58 leeN joined #git
12:59 DaveTaboola joined #git
13:01 nicoulaj joined #git
13:02 Rodya_ joined #git
13:03 venmx joined #git
13:06 j3r3mias joined #git
13:09 borkr joined #git
13:10 mikecmpb_ joined #git
13:13 Dumblez0r joined #git
13:14 urbanmonk joined #git
13:16 theelous3 left #git
13:17 Balliad joined #git
13:17 eb0t joined #git
13:18 aw1 joined #git
13:19 mischat_ joined #git
13:22 Gloomy joined #git
13:23 jameser joined #git
13:25 insiderz joined #git
13:26 kexmex joined #git
13:26 j7k6 joined #git
13:27 insiderz joined #git
13:28 dsdeiz joined #git
13:29 synthroid joined #git
13:31 gugah joined #git
13:31 Cavallari joined #git
13:32 finalbeta joined #git
13:35 nettoweb joined #git
13:36 newbie1 joined #git
13:36 Cavallari joined #git
13:37 hhee2 joined #git
13:38 finalbeta joined #git
13:39 troulouliou_div2 joined #git
13:39 denisMone joined #git
13:44 jameser joined #git
13:44 kexmex joined #git
13:48 marcogmonteiro joined #git
13:51 kadoban joined #git
13:53 dcpc007 left #git
13:54 jimsio joined #git
13:55 marlinc joined #git
13:55 jameser joined #git
13:57 MattMaker joined #git
13:59 GodGinrai joined #git
14:00 mohabaks joined #git
14:01 mohabaks left #git
14:01 mischat joined #git
14:02 seni joined #git
14:02 Snugglebash joined #git
14:05 suttonwilliamd joined #git
14:05 Rodya_ joined #git
14:06 Utjet joined #git
14:06 suttonwilliamd joined #git
14:06 Snugglebash joined #git
14:07 Gsham joined #git
14:08 Gloomy joined #git
14:09 courrier joined #git
14:09 kexmex joined #git
14:10 marcogmonteiro joined #git
14:11 treia joined #git
14:12 eang joined #git
14:13 aspiers joined #git
14:15 tlaxkit joined #git
14:15 mikecmpbll joined #git
14:18 c0c0 joined #git
14:20 jameser joined #git
14:21 jimi_ joined #git
14:21 ShekharReddy joined #git
14:21 explody joined #git
14:21 sbulage joined #git
14:23 Snugglebash joined #git
14:24 Hozy joined #git
14:24 theoceaniscool joined #git
14:26 explody I’ve gotten myself into a bit of a pickle, wondering if anyone has an idea (or can tell me I’m wrong).  I have a repo on github forked from someone else, then cloned to numerous repos separate from GH, and heavy changes were made, occasionally reaaching back to the upstream to merge it in, but not (like a dummy) always pushing local it back to the GH fork.  Basically, the GH fork was mostly to track the upstream repo.  Eventually, I got
14:26 explody point where I was going to stick all the changes back out on GH, but I had to clean up some junk including a few files with sensitive data in them.
14:26 notebox joined #git
14:27 chele joined #git
14:27 explody So, I filtered the local branch and stripped out a few things, not realizing that all the SHA stamps, including from the upstream, would get rewritten from the point that I started filtering things out of the index
14:27 joco42 joined #git
14:29 explody end result - all the code is there and fine but I rewrote like 800 SHA stamps going back a year or so.  I went to do a final merge of recent changes from upstream and git thinks it has to merge about a zillion things, presumably because the SHAs are all different now, yielding a *ton* of duplicate commits where the content and timestamp are identical but the SHAs are different.
14:30 roelmonnens joined #git
14:32 joco42 left #git
14:33 explody I don’t care about the SHAs of the repos that I own, I can force push those just fine, but SHAs from commits merged from the upstream also changed, which is going to make the repo messy at best (dup commits) or at worst, make it a HUGE PITA to merge from upstream later on
14:33 Rr4sT joined #git
14:33 bannakaffalatta joined #git
14:34 tobiasvl yeah
14:35 tobiasvl there's no good solution for that, unfortunately. git needs to rewrite all the child commits too (every coomit points back to its parent, so if the parent changes, so does the child)
14:35 shinnya joined #git
14:35 explody so I learned :(  It makes total sense, I just wasn’t thinking
14:36 explody I was contemplating rewinding a repo back to the common ancestor and doing git-am to replay practically everything up to current
14:36 explody manually D:
14:37 acetakwas joined #git
14:38 lucido-cwl joined #git
14:38 lucido-cwl Hi how do I add a certificate to git on linux?
14:38 explody of course, you can’t export merges as patches it would be several hundred format-patch, git-am, merge, patch, am, merge etc
14:38 texinwien_ joined #git
14:38 lucido-cwl I'm trying to use an https remote
14:39 explody lucido-cwl: a client cert or a CA cert for the server?
14:39 Seveas_ explody: rewinding + git am may be the best way out. It's certainly a reasonable approach
14:39 explody I was afraid of that
14:40 * explody polishes his footgun
14:40 lucido-cwl I have a .p12 file
14:43 roelmonn_ joined #git
14:44 glspi joined #git
14:44 _ikke_ lucido-cwl: git uses curl
14:44 _ikke_ lucido-cwl: refer to you distributions documentation how to add a system-wide trusted certificate
14:45 _ikke_ lucido-cwl: You probably have to convert pkcs12 to pem
14:45 Seveas_ lucido-cwl: man got-config, search for http.sslCA
14:45 explody lucido-cwl:  Alternatively, you can tell git to use your own set of CA certs, e.g. http://stackoverflow.com/questions/907237​6/configure-git-to-accept-a-particular-se​lf-signed-server-certificate-for-a-partic
14:45 Derperperd joined #git
14:46 explody Convert your p12 to pem:  https://www.sslshopper.com/article​-most-common-openssl-commands.html
14:47 lucido-cwl thank you
14:47 dendazen joined #git
14:48 Seveas explody: https://thenanfang.com/wp-conten​t/uploads/2014/06/footgun-01.jpg
14:52 bongjovi joined #git
14:55 GodGinrai joined #git
14:55 mischat_ joined #git
14:57 User458764 joined #git
14:58 MattMaker joined #git
14:58 explody Hey, wait… could I take my “polluted” repo, add the pristine upstream as a remote and interactively rebase?
14:59 sagerdearia joined #git
14:59 explody I could squash any dups, and wouldn’t have to patch/apply
14:59 explody lets find out
15:02 rasto joined #git
15:04 venmx joined #git
15:04 Guest9696 joined #git
15:05 lucido-cwl I installed the cert, but still get: fatal: https://myrepository.git/info/refs not valid: is this a git repository?
15:05 lucido-cwl on push origin master
15:06 explody that URL does not look right at all
15:06 explody do you mean https://somewebserver.com/path/to/respoitory.git?
15:06 lucido-cwl explody, hehe, yes
15:07 explody is that on bitbucket?
15:07 lucido-cwl explody, it is
15:07 explody https://confluence.atlassian.com/bbkb/info-refs-n​ot-valid-is-this-a-git-repository-313467635.html
15:07 Rodya_ joined #git
15:08 ResidentBiscuit joined #git
15:08 Cavallari1 joined #git
15:09 ecuanaso joined #git
15:10 marianina8 joined #git
15:11 Ryanar joined #git
15:14 jstimm joined #git
15:16 whaley joined #git
15:19 chipotle joined #git
15:19 eb0t joined #git
15:19 shgysk8zer0_ joined #git
15:20 explody wow, the rebase might actually work
15:20 explody tedious as hell but less tedious that patching/applying
15:22 noteness joined #git
15:23 jdl joined #git
15:24 F0rTh3J3st joined #git
15:26 ecuanaso joined #git
15:29 col joined #git
15:29 Gloomy joined #git
15:29 Tobbi joined #git
15:30 LyndsySimon joined #git
15:30 LyndsySimon Anyone know how to remove a commit from a remote repo?
15:31 LyndsySimon I pushed a branch to GitHub, used `git reset --hard HEAD~1 && git push -f` locally, and now I want to remove the now-orphaned commit from the remote.
15:32 ghukjs joined #git
15:32 ghukjs left #git
15:33 miczac joined #git
15:39 theoceaniscool joined #git
15:39 ayogi joined #git
15:39 LyndsySimon FWIW, this isn't an emergency in this case. I just don't know how to do it and can see a potential future where the task might be critical.
15:40 kbs joined #git
15:40 xaviergmail_ joined #git
15:40 troulouliou_div2 joined #git
15:41 solenodic joined #git
15:43 canton7 LyndsySimon, as in, remove the now-orphaned commit from the remote repo?
15:43 LyndsySimon Exactly.
15:43 canton7 so you've already removed all references to it on the remote?
15:43 govg joined #git
15:44 canton7 you wait until the remote repo gcs it :P github doesn't allow you to do that. "normal" git remotes will let you do the necessary reflog expire and prune commands, but they won't let you fetch an unreferenced object *anyway* so there's little point
15:44 cdown joined #git
15:44 LyndsySimon Yes, as far as I know. It was in a branch, I pushed it, then I remove it and force pushed the branch.
15:44 LyndsySimon canton7: GitHub *does* allow you to reference an orphaned commit
15:45 canton7 (github used to let you view unreferenced commits in the web viewer thingy, but wouldn't let you fetch them into a git repo. I don't know if that changed)
15:45 LyndsySimon I an envision a situation where, for instance, you accidentally commit a huge file and don't want to download it each time your do an initial fetch.
15:45 canton7 the initial fetch will not include unreferenced objects
15:45 mischat joined #git
15:47 LyndsySimon OK, then imagine you've been asked to remove a password that was accidentally pushed. Ignore the fact that doing so it pointless and that the password should simply be changed :)
15:47 m4sk1n joined #git
15:48 patarr joined #git
15:48 jast it's a case of what the hosting provider happens to allow you to do
15:48 rwb joined #git
15:49 linuxpjohnston joined #git
15:49 LyndsySimon So there's no git command to force a gc on the remote?
15:49 jast if you remove all references to the commit and they still let you view it, and you don't have shell access to the server, you're out of luck
15:50 synthroid joined #git
15:50 jast no... keep in mind that gc can consume a lot of CPU, RAM and disk I/O, so allowing random people to trigger it remotely is basically asking for DoS
15:50 jast additionally 'git gc' doesn't immediately remove unreferenced objects
15:51 Gustavo6046 joined #git
15:52 dreiss joined #git
15:53 LyndsySimon Right - and a commit not on a branch is still not technically "orphaned" as long as it is referenced in the reflog.
15:54 jast yes, though reflogs are not enabled in bare repositories by default, I beileve
15:54 LyndsySimon True. I was just thinking that you can't even remove it from your local repo unless you remove it from the reflog and manually run git gc.
15:54 LyndsySimon OK, I guess the answer is "it's not possible".
15:55 LyndsySimon Out of curiosity, I did set up a GitHub repo with this state. I'm curious to see if it will even garbage collect, especially since I've referenced the commit in an Issue.
15:55 LyndsySimon Here's the link if you're curious as well: https://github.com/lyndsysi​mon/git-remove-commit-test
15:57 jimklo joined #git
15:58 mischat_ joined #git
15:59 MattMaker joined #git
16:00 LyndsySimon Thanks canton7 and jast. Like I said, this wasn't a pressing issue for me right now but it's nice to know how it works ahead of time.
16:00 roelmonnens joined #git
16:01 canton7 LyndsySimon, do bear in mind what I said though - remote repos won't allow you to *fetch* an unreferenced commit at all, because of precisely this issue
16:01 LyndsySimon Interesting.
16:01 canton7 ... it's an odditiy that github's web ui lets you see unreferenced commits (or used to, at least)
16:02 jdl You can see them if you can name them....
16:02 jdl git show <sha1> will do it.
16:02 canton7 jdl, this is a remote repo - you can't 'git show' in it
16:02 jdl I understand.
16:02 LyndsySimon canton7: that doesn't appear to be true.
16:03 canton7 LyndsySimon, pretty sure it is ;) !repro
16:03 gitinfo LyndsySimon: [!transcript] Please paste (using https://gist.github.com/ or similar) a transcript ( https://git.io/viMGr ) of your terminal session so we can see exactly what you see
16:03 LyndsySimon canton7: `git fetch origin b415f25e9c04d21ecc8b5eb4964c1615cbe10398` assigns the commit to FETCH_HEAD, which I can then check out
16:03 solarfly joined #git
16:05 LyndsySimon I'll give you an image, it's not like you'll need to copy/paste it.
16:05 LyndsySimon https://usercontent.irccloud-cdn.co​m/file/PL7letPM/git%20checkout%20of​%20a%20remote%20orphaned%20commit
16:06 canton7 LyndsySimon, output of 'git fetch && git branch -a --contains b415f25e9c' ?
16:07 LyndsySimon Nothing, when I check out the default branch first
16:08 miczac joined #git
16:08 Derperperd joined #git
16:08 roelmonn_ joined #git
16:09 LyndsySimon Hmm. Even more weirdness. It works for a repo that contains that commit locally, but not for a fresh repo.
16:10 Glooomy joined #git
16:10 LyndsySimon https://usercontent.irccloud-cdn.com/file/​MnycWvfn/Testing%20with%20a%20fresh%20repo
16:10 crose joined #git
16:10 Rodya_ joined #git
16:10 LyndsySimon Note that my PS1 is set up so that the block after the command is red for a non-zero exit code, so the last command failed.
16:11 LyndsySimon The happy faces are random for a zero exit status. It amuses me, but is slightly embarrassing now that I see it here, lol.
16:11 canton7 LyndsySimon, this is what I see: https://gist.github.com/canton7/​11bf3c7373753ab50a3fdf4f13da884e
16:12 canton7 hmm, although, the full sha worked. I was 99% sure that was blocked by git. Wonder what's going on...
16:13 LyndsySimon I bet it won't work from a newly-created repo.
16:13 Glooomy joined #git
16:13 LyndsySimon My theory is that it's cached, because you've previously fetched the remote where it was included.
16:13 * LyndsySimon shrugs.
16:16 kst123 joined #git
16:16 Anticom joined #git
16:17 elect joined #git
16:20 jstimm joined #git
16:20 math-alpha joined #git
16:21 solenodic joined #git
16:25 Akuli joined #git
16:25 Akuli does git push care about which branch i'm on?
16:25 elect joined #git
16:27 scoobertron joined #git
16:30 jason237 joined #git
16:40 unreal joined #git
16:41 canton7 Akuli, 'git push' (without other arguments) may do, depending on the value of the 'push.default' config
16:41 Akuli git config push.default says simple
16:41 Akuli what does that mean?
16:41 canton7 man git config for the description
16:41 gitinfo the git-config manpage is available at http://jk.gs/git-config.html
16:42 Akuli the git-push man page didn't mention that 0_o
16:42 thebird joined #git
16:42 jccn joined #git
16:43 Akuli i was on branch cleanup and i did git push origin master
16:43 Akuli what did it do?
16:43 canton7 Akuli, from the top of man git push: "When the command line does not specify what to push with <refspec>... arguments or --all, --mirror, --tags options, the command finds the default <refspec> by consulting remote.*.push configuration, and if it is not found, honors push.default configuration to decide what to push (See git-config(1) for the meaning of push.default)."
16:43 gitinfo Akuli: the git-push manpage is available at http://jk.gs/git-push.html
16:44 canton7 Akuli, if you git 'git push origin master', then the push.default stuff didn't apply: that *only* applies if you type 'git push' (without other arguments), or 'git push remotename'
16:44 canton7 Akuli, 'git push origin master' will always push the local master branch to a remote branch called 'master' on the remote 'origin'
16:45 Akuli oh so local master to remote master?
16:45 Akuli it doesn't care about my branch?
16:45 canton7 actually I'm not entirely correct
16:45 Akuli thanks that was all i wanted to know :)
16:45 Akuli hmm
16:45 Akuli ok
16:45 canton7 " If git push [<repository>] without any <refspec> argument is set to update some ref at the destination with <src> with remote.<repository>.push configuration variable, :<dst> part can be omitted—such a push will update a ref that <src> normally updates without any <refspec> on the command line. Otherwise, missing :<dst> means to update the same ref as the <src>."
16:46 canton7 so if you have a bit of config 'remote.origin.push' (which is set automatically when you create a new local branch from a remote branch, or when you use git branch --set-upstream-to or git push -u), then it will honour that
16:46 canton7 otherwise, it will do as I described
16:46 canton7 but yeah, it doesn't care which branch you have checked out
16:47 Akuli ok
16:47 ecuanaso joined #git
16:47 canton7 ah, no, I misread even that - that's for 'git push origin', not 'git push origin master'
16:47 canton7 so yes: master to master
16:48 Akuli and git push on branch cleanup is cleanup to cleanup?
16:49 peacememories joined #git
16:49 canton7 no: that's when the push.default stuff we just discussed kicks in
16:50 nettoweb1 joined #git
16:51 AustinMatherne joined #git
16:51 col joined #git
16:52 sz0 joined #git
16:53 jagob joined #git
16:53 klin joined #git
16:54 Guest5348 joined #git
16:55 Domhack joined #git
16:56 jstimm joined #git
16:56 robotroll joined #git
16:57 solenodic joined #git
16:58 publio joined #git
16:59 shinnya joined #git
17:00 nowhereman joined #git
17:00 Hozy joined #git
17:00 MattMaker joined #git
17:00 UTAN_dev_ joined #git
17:09 rwb1 joined #git
17:09 venmx joined #git
17:09 re1 joined #git
17:09 nephyrin joined #git
17:09 boshhead_ joined #git
17:09 seveneve2 joined #git
17:09 vmiklos joined #git
17:09 re1 joined #git
17:09 fuchstro1aut joined #git
17:09 S0bait joined #git
17:09 S0bait Hello, how can I share a user across newly installed oses on raspberry Pis? For example, we need to be able to git clone from our private users, at this moment, we will get denied of permission as the user is not added to the github.
17:09 tabakhase__ joined #git
17:09 [sarri] joined #git
17:09 Domhack_ joined #git
17:09 GodEater joined #git
17:09 GodEater joined #git
17:09 tmz joined #git
17:09 Akuli left #git
17:09 oskarkv joined #git
17:09 Learath2 joined #git
17:09 t-mart joined #git
17:09 Cthalupa joined #git
17:09 xaviergmail_ joined #git
17:09 umbSublime joined #git
17:09 umbSublime joined #git
17:09 denisMone joined #git
17:09 sfix joined #git
17:09 todd_dsm joined #git
17:10 ToxicFrog joined #git
17:10 alcohol joined #git
17:10 musicmatze joined #git
17:10 HelloShitty joined #git
17:10 andschwa joined #git
17:10 LyndsySimon S0bait: I don't know if I fully understand your problem, but I *think* what you're looking for is a way to let users of a cloned system pull access to a private GitHub repo. If so, you might consider using a deploy key.
17:11 mattcen joined #git
17:11 LyndsySimon From the private GitHub repo, choose Settings, then Deploy Keys. Add a new key that's read-only.
17:11 MikeChernev joined #git
17:11 kini joined #git
17:11 JanC_ joined #git
17:11 peeps[lappy] joined #git
17:12 LyndsySimon You'll need to take the private key portion of that and include it in ~/.ssh of your RPi's OS image, have it included by default in ssh-agent, and finally configure it to use that key by default when attempting to connect to GitHub.
17:12 OfficialLeibniz joined #git
17:12 S0bait LyndsySimon: So I am trying to have all of our code on github, then when someone is installing a new Raspberry PI OS, they run ascript which installs git and clones all of the repositories. Problem is that the user will need to add their key to the repo. If i go the deploy key route, wont i still have to copy over the private and public key?
17:12 PioneerAxon joined #git
17:12 BlackMaria joined #git
17:14 LyndsySimon Yes. I assumed you were using an OS image on the RPis.
17:14 bambams joined #git
17:14 noodle joined #git
17:15 ojacobson S0bait: You have the option of either making the repo public (so that clients need no special access to clone it) or baking a credential into the OS image (so the repo can be private)
17:15 Rodya_ joined #git
17:15 tabakhase__ joined #git
17:15 ojacobson the former means you can clone it without your special raspi, which _I_ think is a feature but you might not
17:15 ojacobson the latter rules out stock raspian distros
17:15 a_thakur joined #git
17:18 Rodya_ joined #git
17:18 zzz joined #git
17:18 irqq joined #git
17:19 govg joined #git
17:21 svm_invictvs joined #git
17:24 rwb1 joined #git
17:26 raynold joined #git
17:26 mischat joined #git
17:27 madewokherd joined #git
17:30 Atemu joined #git
17:30 raspado joined #git
17:31 anry joined #git
17:31 a_thakur joined #git
17:35 thebird joined #git
17:35 brent__ joined #git
17:37 Motoko joined #git
17:38 irqq joined #git
17:39 User458764 joined #git
17:39 LeBlaaanc joined #git
17:39 nettoweb joined #git
17:40 gugah joined #git
17:41 hhee2 joined #git
17:41 j7k6 joined #git
17:41 bill99 joined #git
17:42 irqq_ joined #git
17:43 zeroed joined #git
17:43 ffilozov joined #git
17:44 puff Afternoon all.
17:44 aw1 joined #git
17:45 aw1 joined #git
17:45 roelmonnens joined #git
17:46 aw1 joined #git
17:47 fmeerkoetter joined #git
17:47 clemf joined #git
17:47 aw1 joined #git
17:48 aw1 joined #git
17:48 S0bait joined #git
17:48 d0nn1e joined #git
17:48 phy1729 joined #git
17:49 aw1 joined #git
17:51 User458764 joined #git
17:51 SwiftMatt joined #git
17:52 Nerbrun joined #git
17:53 ertesx joined #git
17:54 retroj joined #git
17:56 cdown joined #git
17:57 ecuanaso joined #git
17:58 lhx left #git
17:59 retroj hello.  i am getting an error that i don't understand with 'git submodule add': { git submodule add -b 4.7.2 --depth 1 git@github.com:WordPress/WordPress.git 4.7..2 } reports: fatal: Cannot update paths and switch to branch '4.7.2' at the same time.  Did you intend to checkout 'origin/4.7.2' which cannot be resolved as a commit?  Unable to checkout submodule 'wordpress/4.7.2'.  What does this mean?  Is it possible to make a submodule
17:59 retroj that tracks a tag?
18:00 ramsub07 joined #git
18:00 clickjack joined #git
18:01 MattMaker joined #git
18:01 jnavila joined #git
18:03 mloy joined #git
18:03 jstimm joined #git
18:03 roelmonnens joined #git
18:04 mikecmpbll joined #git
18:05 grawity since tags usually point to the same commit and never ever change
18:05 heftig joined #git
18:06 grawity it's already kind of how submodules work by default
18:06 grawity i.e. they're already pinned to a specific commit *unless* you specify a branch
18:06 grawity so... just add it without -b foo
18:07 HurricaneHarry joined #git
18:08 retroj grawity: i was hoping to reduce the size of the .git directory by using a tag name with --depth 1, but i see what you mean.
18:08 User458764 joined #git
18:09 grawity that's pointless
18:09 thebird joined #git
18:09 retroj howso?
18:09 kexmex joined #git
18:09 grawity cloning (including shallow) already works in terms of commit IDs
18:10 grawity i.e. the client doesn't *ask* for a tag, it asks for a commit and the server just verifies that the commit is reachable from some branch/tag
18:10 suttonwilliamd joined #git
18:10 m4sk1n joined #git
18:10 lb1c38 joined #git
18:10 grawity so there should be no difference at all between shallow cloning a tag by name vs by commit-hash
18:11 King_Hual joined #git
18:11 retroj grawity: how would i create the submodule as a shallow clone at a given tag?
18:11 roelmonnens joined #git
18:11 suttonwilliamd joined #git
18:11 heftig joined #git
18:12 eblip joined #git
18:12 suttonwilliamd joined #git
18:13 wilbert joined #git
18:14 TomyLobo joined #git
18:15 rwb joined #git
18:16 a_thakur joined #git
18:19 serialoverflow joined #git
18:19 solenodic joined #git
18:24 bvtton joined #git
18:25 Electrometro joined #git
18:28 Glooomy joined #git
18:28 tyreld joined #git
18:28 a_thakur joined #git
18:31 sparr joined #git
18:38 Doginal joined #git
18:40 mohabaks joined #git
18:40 cbreak remote: you can git clone submodules manually
18:40 aspiers joined #git
18:40 cbreak sorry, retroj
18:44 PioneerAxon joined #git
18:46 jottr joined #git
18:47 esha_ joined #git
18:49 BlessJah joined #git
18:49 sveinse joined #git
18:50 esha_ joined #git
18:52 finalbeta joined #git
18:52 retroj cbreak: i was wondering about that.  how?
18:53 S0bait When I do git status i just see "./" how do I get it to actually tell me what changed and what got added.. this sucks.. http://prntscr.com/eb8ct2
18:53 cbreak retroj: git init a new repo
18:53 cbreak git remote add the remote
18:53 cbreak set the fetch spec to fetch the ref you want
18:53 cbreak set the configuration for shallow fetches
18:53 cbreak then fetch
18:53 Todd joined #git
18:54 YuGiOhJCJ joined #git
18:54 retroj cbreak: how do i configure that as a submodule though?
18:54 Todd hello
18:54 gitinfo Todd: 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.
18:55 cbreak retroj: git submodule add
18:55 esha_ joined #git
18:55 cbreak retroj: everyone else that clones your repo and then does a submodule init / update gets a normal repository of course
18:57 garethdaine joined #git
19:00 retroj cbreak: 'git submodule add <repo>' performs a git-clone.  i didn't see a way around that.
19:01 j7k6 joined #git
19:01 Oebele joined #git
19:02 nixjdm joined #git
19:02 MattMaker joined #git
19:03 diogenese joined #git
19:04 theoceaniscool joined #git
19:05 venmx joined #git
19:05 Derperperd joined #git
19:06 esha_ joined #git
19:06 apotry joined #git
19:07 cbreak retroj: it doesn't
19:07 cbreak retroj: you can specify an existing repo
19:07 matoro joined #git
19:07 cbreak the man page describes this in detail, read it :)
19:09 KevinDF joined #git
19:10 KevinDF joined #git
19:10 hhee2 joined #git
19:12 xaviergmail_ joined #git
19:15 Learath2 joined #git
19:18 reynierpm joined #git
19:18 reynierpm hi there
19:19 metalraiden34 joined #git
19:19 reynierpm I have one branch1 with some changes commited but not push, I am working in a branch2 I haven't commit anything here yet, I want to merge changes from branch2 into branch1 and dismiss every change in branch1, how can I do this?
19:20 cbreak reynierpm: if you haven't committed anything in branch2, then there's nothing to merge
19:20 glspi_ joined #git
19:20 cbreak reynierpm: why do you need branch 1?
19:20 cbreak just git checkout master
19:20 cbreak and commit there
19:20 cbreak or something like git checkout -b branchname master
19:21 cbreak if you want a new branch named branchname
19:22 reynierpm cbreak, because that branch already exists on remote so I don't want to create a new one just for a few changes
19:22 reynierpm cbreak, I will commit changes on branch2
19:22 cbreak you'll have to force push anyway if you want to remove old history
19:23 max42 joined #git
19:23 reynierpm cbreak, ummm ya, you're right so what's the command to push a branch to remote and set the up-stream all in one?
19:24 cbreak git push -fu origin branchname
19:24 glspi_ left #git
19:24 ecuanaso joined #git
19:24 cbreak where origin is the remote name
19:24 diogenese joined #git
19:24 glspi joined #git
19:24 cbreak -f means force, -u means set-upstream
19:24 reynierpm gotcha
19:24 jnavila joined #git
19:25 reynierpm I believe origin should be origin/<repo_name> right?
19:25 lhx joined #git
19:28 irqq joined #git
19:28 lhx left #git
19:28 miczac joined #git
19:31 j7k6 joined #git
19:31 solenodic joined #git
19:32 max42 I have a file commited as 644 to the repo. When I do chmod u+x file, git will show the mode as changed to 755 while ls -la shows rwxr--r--. Even after commiting ls -la shows rwxr--r-- while git says it's 755. Is this a bug or feature?
19:32 robattila256 joined #git
19:32 Eugene max42 - git does not actually store the full octal permission; only the +x bit.
19:32 Random832 max42, files in repositories aren't allowed to have a mode other than 644 and 755.
19:33 Eugene Your umask setting is applied, which strips the excess permissions(to match that of the directory normally)
19:33 max42 to the point, thanks
19:33 Random832 Eugene, oddly, it actually does store several digits of redundant modes in ASCII octal encoding.
19:33 Random832 like physically, in the format of the tree object
19:33 sveinse We have a git:// server that has been taken down. Can I redirect it (the best) or proxy it to the new server?
19:33 ramsub07 joined #git
19:33 Eugene Random832 - thats how it is encoded in the commit object yes; if you dig through the code you'll see that its ignored
19:33 Random832 and despite the hash being in binary
19:34 Random832 yes it's just kind of a weird gotcha
19:34 Eugene There's also a bitfield for symlinks, which are stored as a plaintext file with the "name" of the destination as the contents
19:34 Random832 something that has caused me no end of grief on Windows, though it might actually be mingw's fault.
19:34 Eugene And some other things that I can't remember off the top of my head - been a while since I read this manpage
19:34 joeco joined #git
19:35 Eugene Yes; its a mingw/Win32 API problem. There is no direct conversion between UNIX symlinks and NTFS junction points / shortcuts
19:35 Eugene Solution: dont use symlinks
19:35 kadoban reynierpm: git doesn't really have that concept, of remotes with multiple repos on them. All it knows about is remotes, each one specifies an exact remote repo. You can specify URLs directly, in a few formats, or usually you have aliases for them, like 'origin'. You can query/modify them with 'git remote' subcommands.
19:35 Eugene kadoban - sure it does, you can do a multi-fetch or multi-push just fine. It doesn't make sense from a UI perspective, but it works
19:36 Eugene Just specify multiple url entries
19:36 Eugene (don't do this)
19:36 Random832 submodules, which are the only one that aren't stored in a way "equivalent" to a UNIX octal value (directories get 040000 and symlinks 120000)
19:36 Eugene You can also do insane things with url != pushurl (again, don't if you like sanity)
19:36 kadoban Eugene: I mean hierarchal remotes, where you specify the origin and then on that remote you specify what repo you're talking about.
19:36 Eugene submodules are their own breed of insanity
19:37 Eugene kadoban - you can do that with a ssh_config entry. Example: https://vomitb.in/E9Nt492V1t
19:37 Forlorn joined #git
19:37 iLembus joined #git
19:37 Eugene `git remote add github github:User/repo.git`
19:38 Eugene You can't have collisions of the repo name, because that would be dumb(I believe the last-defined-one would take precedence?)
19:39 kadoban *shrug* we're talking about different things, read what I was responding to if you want.
19:39 retroj cbreak: hrm.. i can't find that feature.  i have git version 2.11.0 and the man page for 'git submodule' shows that 'add' requires an origin repository as an argument
19:39 Eugene Different way to approach it
19:39 Eugene Or I'm not drunk enough to understand. EIther way: more booze
19:40 aw1 joined #git
19:40 kadoban More booze sounds like a brilliant idea right now
19:40 aw1 joined #git
19:41 garethdaine joined #git
19:41 * phy1729 submits "<Eugene> Or I'm not drunk enough to understand. EIther way: more booze" to git-jokes
19:41 j7k6_ joined #git
19:42 Eugene Alcoholism is no joke
19:42 Eugene You need to have years of practice and a good support network if you want to be a successful alcoholic
19:43 venmx joined #git
19:45 retroj cbreak: aha!  found it.  sorry for the confusion - the relevant sentence in the manual for my situation is buried where i didn't expect it
19:45 sparr what's the most straightforward way to make the working directory represent the merged state of multiple branches, without making any changes to those branches?
19:49 ffilozov sparr, Wouldn't checking out the merge commit do what you want?
19:49 ffilozov Or you mean, before everything is merged?
19:50 maurer joined #git
19:52 max42 left #git
19:52 maurer I have repository A, and under A, I want several checkouts of B at different refs, e.g. A/foo/B@ref1, A/foo/B@ref2, A/bar/B@ref3,...
19:52 ffilozov You might want to make a temporary branch, merge everything, and checkout the temporary branch.
19:53 maurer I intend to use git submodule for this, but I'm wondering if there's a way to know git won't fetch a full copy of the repository for each submodule
19:54 BytesAndCoffee left #git
19:55 sveinse What is the transport on a git:// URL?
19:55 eroux joined #git
19:55 grawity just raw TCP
19:56 sveinse Can it be proxied?
19:56 iLembus_ joined #git
19:56 grawity probably, if you have a custom-built proxy
19:56 grawity or if you have SOCKS/HTTP-CONNECT/something that does arbitrary TCP
19:58 MarioBranco joined #git
19:58 jjasinski joined #git
19:58 duderonomy joined #git
19:58 Derperperd joined #git
19:58 weep joined #git
19:58 sveinse hmm. We have a git server that has been taken down in favour of a new server with better security and backup. Thing is that the URL changes as well. I hoped I could setup a http redirect, only to find that it uses git:// which is not http
19:59 esha_ joined #git
19:59 grawity git config url.git://newserver/.insteadOf git://oldserver/
19:59 retroj maurer: i am trying to do exactly the same thing.  it's possible to configure the submodules to be shallow locally, but there doesn't seem to be a way to have shallow submodules of references other than HEAD when cloning
19:59 sparr ffilozov: before everything is merged. I've got a dozen feature branches, and I want to compile them all together.
20:00 sparr I could make a new branch and merge all the others into it, but that's a lot of background work I'd prefer to avoid
20:00 grawity or just set up a generic TCP proxy, like rinetd?
20:00 maurer retroj: Hm, that is not great for what I'm trying to do. Maybe I'll just make a separate update.sh script that syncs everything up
20:01 sveinse grawity: I can't change the instruction/documentation already put into existing releases and documents, so git clone must work out of box.
20:01 maurer (the repository I've got referenced multiply is 602M, though an individual checkout is not so big, but I intend to reference it maybe hundreds of times)
20:01 ffilozov sparr, That's a lot less work than manually selecting files from each branch and putting them in your working directory. If you don't have any merge conflicts, merging all your features in one branch is easy.
20:01 sveinse BTW how does git respond to http redirect? failure or warning and follows the redirect?
20:01 maurer (so cloning it each time when someone goes to set up the repo is no-go)
20:02 nidr0x joined #git
20:02 esha_ joined #git
20:02 sparr ffilozov: ok, assuming I do that... how do I blow that branch away later, so I'm not keeping a history of all those merge commits?
20:03 MattMaker joined #git
20:03 safe joined #git
20:03 suttonwilliamd joined #git
20:03 ffilozov Just a regular branch delete, git branch -D <branch_name>
20:04 suttonwilliamd joined #git
20:04 sparr can I skip the commit message prompt for the merges?
20:05 mehola joined #git
20:05 mehola joined #git
20:05 ffilozov I don't think so.
20:05 maurer retroj: I figured it out!
20:06 maurer retroj: git submodule add --force --name common-repo-name path/to/checkout
20:06 maurer retroj: git submodule add --force --name common-repo-name path/to/othercheckout
20:06 MarioBranco joined #git
20:06 ffilozov sparr, There might be a way with some flags like --no-edit, but I'm not certain.
20:06 retroj maurer: interesting!  can you use -b?
20:06 maurer Oh hm, this might not actually work
20:07 maurer I thought it worked based on the error message
20:07 maurer let me see
20:08 roelmonnens joined #git
20:09 maurer retroj: Hm, this may not do quite what I want, checking out a different revision of one submodule seems to change the other
20:09 retroj maurer: yes, when they share a name, that means they share a .git directory
20:09 metalraiden34 joined #git
20:09 maurer Yes, and I want them to do
20:09 maurer *to do that
20:09 maurer I don't want them to share a worktree though
20:10 retroj right..
20:10 Balliad joined #git
20:10 [Brain] joined #git
20:11 aw1 joined #git
20:12 retroj maurer: i wondered if there was a way to use git submodule --reference to do that
20:13 retroj maurer: however, i think that would only optimize files locally, not reduce the time and disk space of a clone
20:14 pur3eval joined #git
20:15 jimklo joined #git
20:16 eroux joined #git
20:17 ixpict joined #git
20:18 Tobbi joined #git
20:18 alexandre9099 joined #git
20:19 SirOliver joined #git
20:20 clemf joined #git
20:21 cbreak retroj: it doesn't matter where the submodule HEAD points to
20:21 cbreak retroj: because after update it will point at the correct hash
20:22 User458764 joined #git
20:23 Sample joined #git
20:24 aidalgol joined #git
20:24 Sample joined #git
20:25 shgysk8zer0 joined #git
20:25 RasmusWL joined #git
20:26 suttonwilliamd joined #git
20:26 RasmusWL left #git
20:26 BlessJah joined #git
20:26 retroj maurer: i would love it if there were submodule.<name>.share options that allowed submodules to share an object store, but each have an independent checkout ref
20:26 matoro joined #git
20:28 reynierpm left #git
20:28 dviola joined #git
20:28 retroj cbreak: it's good enough for me to create a new submodule by performing a 'git clone -b' followed by 'git submodule add -b' on that same directory.
20:29 finalbeta joined #git
20:29 retroj cbreak: it seems that shallow submodules are not possible in this scenario (such that git-clone --shallow-submodules) could be used, but i can live with that.
20:31 maurer retroj: https://github.com/maurer/b​ugs/blob/master/setup.bash
20:31 maurer I'm testing this to see if it works
20:31 maurer it seems like it does
20:32 jimi_ joined #git
20:32 maurer They'll each have .git directories, which isn't optimal
20:32 maurer but they'll share an object store
20:32 maurer which is 90% o fthe space
20:32 retroj that could be the solution
20:33 HurricaneHarry joined #git
20:33 maurer The second one needs a --init too
20:34 eroux joined #git
20:36 MrWoohoo joined #git
20:36 retroj maurer: trying it... you may be my hero :-)
20:38 physikoi joined #git
20:39 retroj maurer: my hero :-D
20:39 physikoi hi #git. Anyone familiar with `git grep`? Currently I've got this -- `git grep -Ew -f migrate.txt -h old/_user-settings.scss`. What I'm wondering is if I could augment this with a commandline option specifying that I want each match to begin with a, in this case, "$". Thoughts?
20:43 MattMaker joined #git
20:43 venmx joined #git
20:43 justanotheruser joined #git
20:44 User458764 joined #git
20:44 Ryanar joined #git
20:45 osse physikoi: you could use shell magic to edit the file
20:45 P4nther joined #git
20:45 fragMental joined #git
20:45 osse git grep -Ew -f <(sed 's/^/$/' migrate.txt) -h old/_user-settings.scss
20:46 physikoi osse: ohhhhh. i don't know why, but i wasn't thinking of the file as stdin
20:46 physikoi osse: thanks, gonna try that now!
20:47 osse oh, hmm.  sed 's/^/$/' migrate.txt | git grep -Ew -f - -h old/_user-settings.scss  might work too
20:47 jim joined #git
20:48 rewt physikoi, that was not stdin.  <( ) creats a "virtual file" whose contents is the output of the enclosed command
20:49 jim hi, haven't run into this before, but I have a merge conflict, here's the output of an attempted git pull: http://termbin.com/ro5v
20:49 jim how can I fix this?
20:49 LiftLeft joined #git
20:49 rewt git rebase
20:49 osse jim: !eek
20:49 gitinfo jim: [!eekaconflict] Merge conflicts are a natural part of collaboration. When facing one, *don't panic*. Read "How to resolve conflicts" in man git-merge and http://git-scm.com/book/ch3-​2.html#Basic-Merge-Conflicts then carefully go through the conflicts. Picking one side verbatim is not always the right choice! A nice video explaining merge conflicts: https://www.youtube.com/watch?v=zz7NuSCH6II
20:50 jim thanks
20:50 jstimm joined #git
20:51 jim the basic deal is I accidentally committed to a repo in a working dir which is an older copy... I want to get the changes made on both repos if I can
20:52 osse you will
20:52 osse run git diff and see what's going on
20:52 physikoi rewt: oh, thanks for the disabusement
20:53 aspiers joined #git
20:54 aw1 joined #git
20:54 rewt aye, it's easy to misread that... it comes in really handy when you want to, for example, diff 2 files that have different timestamps at the start of each line:  vimdiff <( cut -f1 file1.txt) <(cut -f1 file2.txt)
20:55 retroj maurer: what will your procedure be when you want to add a new submodule later?  both to update the master nixpkgs and add the new one?
20:56 diogenese joined #git
20:57 eang left #git
20:59 glebihan joined #git
21:00 maurer retroj: At the moment, yeah, what I'll have to do is to bump nixpkgs master, then add the new submodule
21:00 maurer I'm kind of wondering if I should have the reference nixpkgs not be a submodule
21:00 maurer and instead just a clone created by setup.bash
21:00 maurer and have an update.bash which git pulls inside nixpkgs
21:00 maurer then git submodule updates the rest
21:00 retroj for me, it's convenient for the reference repo to be a submodule itself
21:01 retroj (i'm working with a tree of wordpress versions)
21:02 chitopunk joined #git
21:03 glebihan_ joined #git
21:05 mischat joined #git
21:05 User458764 joined #git
21:06 Isla_de_Muerte joined #git
21:08 miczac joined #git
21:09 matoro joined #git
21:10 ecuanaso joined #git
21:13 retroj maurer: after adding new submodules and updating the reference module, what is the 'git pull' procedure to update clones?
21:15 retroj maurer: would it be just this? { git pull; git submodule update --init --reference ./nixpkgs }
21:15 dsdeiz joined #git
21:15 dsdeiz joined #git
21:15 maurer I bleieve so
21:15 maurer *believe
21:15 maurer Worst case would be
21:16 maurer git pull; git submodule update --init nixpkgs; git submodule update --init --reference ./nixpkgs
21:18 retroj maurer: what happens if you never do { update --init nixpkgs } but only do { update --init --reference ./nixpkgs }.  is that an error or will git do what you mean?
21:18 retroj (i'll try that here)
21:18 d1z joined #git
21:18 d1z can somebody explain this format to me? 2017-01-22T16:20:31-05:00
21:18 maurer Uh, if you've never initialized ./nixpkgs, I'll expect it'll error
21:18 maurer since nixpkgs can't use itself as a reference
21:18 d1z the date is clear, I'm talking about everything after T
21:19 maurer If you've initialized it, but not updated it, I expect it will just end up with the new submodules having some objects not properly centralized
21:19 d1z 16:20:31 I guess equals 4:20 PM 31 seconds
21:20 _ikke_ correct
21:20 _ikke_ and the -05:00 means the timezone offset
21:20 _ikke_ GMT -5
21:20 _ikke_ rather, UTC -5
21:20 d1z alright
21:20 GGMethos joined #git
21:21 retroj maurer: oh, i see.  thank you!
21:22 clemf joined #git
21:22 explody joined #git
21:22 metalraiden34 joined #git
21:24 glebihan joined #git
21:29 aspiers joined #git
21:29 m0viefreak joined #git
21:30 Atm0spher1c joined #git
21:32 jim thanks, I think I'm ok now
21:37 suttonwilliamd_ joined #git
21:39 ljc joined #git
21:41 pur3eval joined #git
21:41 texinwien_ joined #git
21:42 j7k6_ joined #git
21:43 nowhereman joined #git
21:44 rorro joined #git
21:48 F0rTh3J3st_ joined #git
21:51 lb1c38 joined #git
21:53 enckse joined #git
21:54 ecuanaso joined #git
21:54 storrgie joined #git
21:55 elijah__ joined #git
21:55 maroloccio joined #git
21:56 maroloccio hi. explained simply, difference between (1) and (2)? (1) git submodule update --init (2) git submodule update // typical use case for (1) and (2)?
21:58 elijah joined #git
21:59 xaviergmail_ joined #git
22:00 _ikke_ maroloccio: You need to run git submodule init before git submodule update for new submodules
22:00 _ikke_ git submodule update --init does that automatically
22:00 hasc joined #git
22:01 wgrant joined #git
22:01 maroloccio _ikke_: thank you
22:01 hasc joined #git
22:02 mischat joined #git
22:04 raspado joined #git
22:05 blackwind_123 joined #git
22:07 jim .lskeys
22:11 suttonwilliamd_ joined #git
22:13 miczac joined #git
22:14 User458764 joined #git
22:14 jim .lskeys
22:15 _ikke_ ..
22:15 jim argh, sorry
22:15 _ikke_ ;-)
22:18 SheldonCooper joined #git
22:19 SheldonCooper quick question, I'm writing some unit tests for some custom git operations scripts
22:20 SheldonCooper doe git diff and git format-patch use the network or does it use the local fetch of the remotes
22:20 SheldonCooper because the tests will run in a sandbox without net access
22:20 m0viefreak no net access needed for those
22:20 SheldonCooper good
22:21 SheldonCooper thanks :)
22:22 glspi_ joined #git
22:22 _ikke_ only commands like git fetch, git push, in some cases git remote, git ls-remote require network access
22:22 _ikke_ (and git pull, because it does a git fetcH)
22:24 wgrant joined #git
22:25 F0rTh3J3st joined #git
22:26 User458764 joined #git
22:26 z8z joined #git
22:29 Rodya_ joined #git
22:32 amikiri joined #git
22:32 miczac joined #git
22:33 war_pigs joined #git
22:33 metalraiden34 joined #git
22:33 tvw joined #git
22:34 tvw joined #git
22:36 jimi_ joined #git
22:40 ozmage joined #git
22:40 nettoweb1 joined #git
22:42 nowhereman joined #git
22:43 MattMaker joined #git
22:43 ozmage_ joined #git
22:43 crst_ joined #git
22:44 Doginal left #git
22:45 crst_ Hi, when I git clone a project from github and then I make a small change, how can I keep that change and still being able to sync the project from github to get the newest update?
22:46 kadoban crst_: You should do your changes in a branch. You shouldn't commit to the master branch at all, or whatever the default branch is called.
22:46 kadoban crst_: Then in your branch, you can either rebase it on top of master occasionally to get updates from upstream, or you can merge master into your branch, whichever you prefer. Usually the latter if you're sharing your branch with others, the former otherwise.
22:48 kent\n kadoban: you know you can still do that on master right?, you just rebase/merge against 'origin/master' instead of 'master'
22:48 gugah joined #git
22:48 anuxivm joined #git
22:49 _ikke_ kent\n: sure, but it's easier / more clear if you keep it separate
22:50 kadoban kent\n: Yes, but it's confusing and people screw it up a lot.
22:50 SwiftMatt joined #git
22:50 _ikke_ kent\n: one git pull, and you merged in new history with that change
22:51 kent\n yeah. :)
22:51 kent\n I personally get more paranoid about pushing to the wrong place
22:51 kent\n anything that happens locally can be fixed ;)
22:52 ffilozov crst_, You might find this useful https://help.github.com/articles/syncing-a-fork/
22:54 bannakaffalatta joined #git
22:55 ozmage joined #git
22:56 Rodya_ joined #git
22:56 ozmage joined #git
22:56 mentazoom joined #git
23:01 fmeerkoetter joined #git
23:02 theoceaniscool joined #git
23:04 ecuanaso joined #git
23:05 ozmage joined #git
23:05 justan0theruser joined #git
23:06 ozmage joined #git
23:06 crst_ thanks for your help guys! I tried kadoban's method, made a branch, changed something, commit, "git branch --set-upstream-to=origin/master test", git rebase, git pull... etc... can I see git diff master my branch to check what happened?
23:07 Taikibansei joined #git
23:09 crst_ damn, files got deleted... on my branch
23:09 metalraiden34 joined #git
23:10 crst_ mhmm, no they'r still there, only in my test branch.. wow so comfusing
23:11 ToBeCloud joined #git
23:11 crst_ I see they coexist, nice
23:13 Murr joined #git
23:14 shgysk8zer0_ joined #git
23:14 rvgate_ joined #git
23:15 PtxDK joined #git
23:16 RichiH joined #git
23:16 reactormonk joined #git
23:16 ecuanaso joined #git
23:18 crst_ git is sooooo complicated, you guys must be geniuses being able to comprehend it. I've been using it for a couple of years for my personal projects and still no clue.
23:18 steven joined #git
23:19 _W_ joined #git
23:21 kadoban Most people seem to attain a kind of basic level of competence in it and never touch the "fun" parts. If you hang out here much, you'll definitely learn a lot. I know I have.
23:24 Electrometro joined #git
23:26 crst_ I remember trying to remove some big files from history once... another time, merging commits.. wow, it was hell and I probably messed everything up. I hope I don't have to rely on getting things back and find out what I need is gone :)   I'm sure when you spend a lot of time on it you can get good though.
23:29 menip joined #git
23:30 crst_ And not to forget some projects being totally messed up after setting up team collaboration. Turned out to be unsyncable after a couple of commits and had to be "reset" manually.
23:30 crst_ :D
23:31 ISmithers joined #git
23:33 askb joined #git
23:35 venmx joined #git
23:37 d1z joined #git
23:38 heftig joined #git
23:38 kadoban Usually that happens when someone changes published history.
23:39 SCHAAP137 joined #git
23:43 overlord_tm joined #git
23:43 j7k6 joined #git
23:43 crst_ kadoban: yeah, I probably did that.
23:45 bannakaffalatta joined #git
23:45 maroloccio joined #git
23:47 matoro joined #git
23:47 dsdeiz joined #git
23:48 dreiss joined #git
23:48 kexmex joined #git
23:50 solenodic joined #git
23:53 Gloomy joined #git
23:53 Derperperd joined #git
23:53 mischat joined #git
23:53 Gloomy joined #git
23:54 Gloomy joined #git
23:55 acidjnk22 joined #git
23:55 pks joined #git
23:55 acidjnk22 hello
23:55 gitinfo acidjnk22: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.
23:55 Gloomy joined #git
23:56 Gloomy joined #git
23:56 acidjnk I just did a merge from hell. You know. One branch developed apart from the starting revision for 4 months, and did so in a single, massive commit. Bringing that together with master after months.
23:57 shinnya joined #git
23:57 anuxivm left #git
23:57 Derperperd joined #git
23:57 acidjnk I did the conflict resolution now. Should I do a commit, with hundreds of compile errors, and then fix compile errors in a different commit? Or fix it first and then commit?
23:57 Gloomy joined #git
23:57 Kaisyu joined #git
23:57 mikecmpbll joined #git
23:58 menip joined #git
23:59 venmx joined #git

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