Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2017-01-13

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

All times shown according to UTC.

Time Nick Message
00:00 Cerise joined #git
00:00 Cerise joined #git
00:01 sarri joined #git
00:01 sarri joined #git
00:01 rizonz Cerise: like commits show
00:01 rizonz liek on giuthub and such
00:03 osse yes
00:03 phroa while git remotes (that includes github) do know that they are being force-pushed to, I don't believe github makes that information available on your repo's web page, and no git remote transfers that information to local git clones
00:03 zyrb left #git
00:03 phroa if "commit x was force-pushed!" is what you mean
00:03 phroa the actual commits which you're pushing will certainly appear, that's what they're designed to do
00:10 csk157 joined #git
00:11 snowkidind Im weirded out
00:12 snowkidind I have been developing on my local machine, pushing to github and then using git pull to get updates into production
00:12 ixian joined #git
00:12 snowkidind so i go to git pull as usual but somehow it goes like this instead: https://hastebin.com/dewonifixe.sql
00:12 swalladge joined #git
00:13 snowkidind I was pushing to a different server the same website as well so maybe thats what got messed up but I dunno
00:13 snowkidind I just want to update everything with the commit I just pushed to github. what do I do?
00:14 snowkidind (save for trashing the entire dir structure and starting again
00:15 madewokherd joined #git
00:16 snowkidind weird part is that I made no local changes, save for the fact that they probably don't reflect the last commit
00:25 acln joined #git
00:26 imachuchu snowkidind: well git says you do have local changes. Why don't you do a "git difftool" to see what new things you added compared to the commit you have checked out. From the looks of it you might have checked out a different commit as a detached head or copy/pasted code into your repo from a different commit
00:26 snowkidind I figured out what caused it
00:26 snowkidind I changed the group of some of the files
00:26 snowkidind im just going to start over :(
00:27 snowkidind thanks for respondin
00:27 snowkidind g
00:28 lukky513 joined #git
00:32 imachuchu snowkidind: no problem
00:32 gugah joined #git
00:33 _0xbadc0de joined #git
00:35 acln joined #git
00:35 StrongBit_ joined #git
00:38 _0xbadc0de joined #git
00:40 StrongBit_ joined #git
00:44 _0xbadc0de joined #git
00:45 csk157 joined #git
00:46 roygbiv joined #git
00:46 publio joined #git
00:47 voldz1tic3n joined #git
00:48 _0xbadc0de joined #git
00:49 mischat joined #git
00:49 neo_ joined #git
00:50 Nugget_ left #git
00:51 nettoweb joined #git
00:51 Nugget joined #git
00:53 xaviergmail joined #git
00:53 paws_ joined #git
00:53 StrongBit_ joined #git
00:58 rkazak joined #git
00:59 acln joined #git
01:00 Atm0spher1c joined #git
01:02 Kaisyu joined #git
01:04 thiago joined #git
01:06 cdown_ joined #git
01:09 acln joined #git
01:09 _0xbadc0de joined #git
01:09 rossome joined #git
01:10 NeverDie joined #git
01:17 Startrek852 joined #git
01:17 Startrek852 Hello!
01:17 Startrek852 Is it possible to simulate a git push?
01:18 UnnamedPlayer joined #git
01:18 zzz joined #git
01:20 Startrek852 joined #git
01:21 Startrek852 Sorry, had to switch devices.
01:21 voldz1tic3n joined #git
01:21 csk157 joined #git
01:21 sargas joined #git
01:25 boutcher joined #git
01:26 Startrek852 joined #git
01:26 Startrek852 Sorry, lost connection.
01:28 nowhere_man joined #git
01:32 leprechau joined #git
01:32 simonides joined #git
01:37 axisys_ joined #git
01:37 arescorpio joined #git
01:38 axisys_ joined #git
01:39 brent__ joined #git
01:40 Nilesh_ joined #git
01:42 shgysk8zer0 joined #git
01:46 sbeller_ Startrek852: what part of the push?
01:48 Startrek852 sbeller_: What part?
01:49 sbeller_ <Startrek852> Is it possible to simulate a git push?
01:49 sbeller_ what is the simulation trying to archive? (Do you want to know if it would just work or how much bytes are sent over the wire?)
01:49 sbeller_ there is git push --dry-run, how does that differ from what you need for simulating the git push?
01:50 Startrek852 Oh, there is?
01:50 mischat joined #git
01:50 jameser joined #git
01:50 sbeller_ it's documented as "Do everything except actually send the updates."
01:50 sbeller_ which is a bit vague IMO
01:51 arand__ joined #git
01:51 Siegfried left #git
01:53 Startrek852 That sounds like what I want. I'm about to push to a Gerrit for the first time, so I wanted to test some of the options.
01:55 cqi joined #git
01:57 Mikasi joined #git
01:57 xaviergmail joined #git
01:58 csk157 joined #git
02:00 Vampire0 joined #git
02:02 Mobutils joined #git
02:05 Sasazuka joined #git
02:07 gambl0re joined #git
02:09 yohnnyjoe joined #git
02:09 rkazak joined #git
02:11 finalbeta joined #git
02:14 butchster joined #git
02:16 anuxi left #git
02:19 Wulf4 joined #git
02:19 _0xbadc0de joined #git
02:30 davimore_ joined #git
02:34 csk157 joined #git
02:36 Gsham joined #git
02:36 anonymuse joined #git
02:41 impermanence joined #git
02:42 pks joined #git
02:46 d^sh joined #git
02:46 ianmethyst joined #git
02:46 aw1 joined #git
02:46 sargas joined #git
02:46 sdothum joined #git
02:46 sargas joined #git
02:46 Darren_ joined #git
02:46 Gsham joined #git
02:49 ilbot3 joined #git
02:49 Topic for #git is now Welcome to #git, the place for git help and a fresh start on a new branch | Public logs at http://jk.gs/glog | First visit? Read: http://jk.gs/git | Current stable version: 2.11.0 | Getting "cannot send to channel"? /msg gitinfo .voice | This channel may contain users known to the state of California to cause cancer.
02:50 Vampire0_ joined #git
02:51 mischat joined #git
02:51 pks joined #git
02:51 d1z joined #git
02:53 wget_ joined #git
02:55 SpeakerToMeat joined #git
02:57 aavrug joined #git
02:58 Ramen joined #git
02:59 dreiss joined #git
03:10 csk157 joined #git
03:12 Doginal joined #git
03:16 d1z joined #git
03:18 FSadino joined #git
03:18 Derperperd joined #git
03:19 j416 joined #git
03:22 hexagoxel joined #git
03:27 Vortex34 joined #git
03:27 dsantiago joined #git
03:34 sunri5e joined #git
03:36 mizu_no_oto joined #git
03:40 jon-mac joined #git
03:40 albel727 joined #git
03:45 Goplat joined #git
03:50 fstd_ joined #git
03:50 Limes__ joined #git
03:52 mischat joined #git
03:52 subhojit777 joined #git
03:54 ColdKeyboard[a] joined #git
03:55 chachasmooth joined #git
03:55 hasc joined #git
03:56 a_thakur joined #git
03:58 a_thakur_ joined #git
04:00 duderonomy joined #git
04:00 Gsham joined #git
04:01 anonymuse joined #git
04:07 svm_invictvs joined #git
04:07 boutcher joined #git
04:13 safe joined #git
04:14 Lyqyd joined #git
04:17 sargas joined #git
04:19 Derperperd joined #git
04:20 robotroll joined #git
04:27 ReptilianBrain joined #git
04:29 proq joined #git
04:34 ygeLN joined #git
04:35 thiago joined #git
04:39 chachasmooth joined #git
04:45 proq is there a way to run git pull origin master so that it doesn't throw an exception if there is no repo in that folder?  it is breaking my bash script if it does
04:45 proq *when it does
04:45 rewt have the script check if there's a .git dir
04:48 rewt or ask git if there's a repo there:  git status &> /dev/null ; echo $?   results in 0 for yes, and 128 for no
04:50 albel727 joined #git
04:52 mischat joined #git
04:54 rainrawr joined #git
04:55 raspado joined #git
04:56 thiago proq: git rev-parse --show-prefix 2>/dev/null
04:57 UnnamedPlayer joined #git
04:57 jmpp joined #git
04:57 * jmpp greets!
04:58 jmpp what's a quick way to clear *all* changes from my working copy in modern git?
04:58 Gianormagantrous joined #git
04:58 Gianormagantrous joined #git
04:58 jmpp I've been Googling and I see 'git checkout -- .' seems to have been a popular way of doing it
04:58 thiago git reset --hard
04:59 jmpp but on 2.11.0 I get the error: error: pathspec '.' did not match any file(s) known to git.
04:59 jmpp or that, too, yeah
04:59 thiago git checkout -- . does not throw away waht'sbeen git add'ed
04:59 jmpp OK
04:59 thiago you'd need: git checkout HEAD -- .
04:59 jmpp the reset worked, thanks! Wonder why I forgot about that...
05:00 jmpp OKOK
05:03 mehola joined #git
05:04 _0xbadc0de joined #git
05:04 robattila256 joined #git
05:07 lindenle joined #git
05:09 SwiftMatt joined #git
05:12 tyreld joined #git
05:16 robattila256 joined #git
05:18 ni291187 joined #git
05:18 ni291187 left #git
05:19 subhojit777 joined #git
05:19 justan0theruser joined #git
05:19 Derperperd joined #git
05:21 Puffball joined #git
05:23 njbair joined #git
05:23 sbulage joined #git
05:24 ayjay_t joined #git
05:26 Guest27 joined #git
05:26 testuser2 joined #git
05:26 testuser2 left #git
05:28 durham joined #git
05:31 mrt joined #git
05:32 diogenese joined #git
05:34 mrt joined #git
05:35 mrpropa joined #git
05:36 mrpropa joined #git
05:44 Cabanossi joined #git
05:46 vdv joined #git
05:48 NeverTired joined #git
05:49 jmpp joined #git
05:50 skypuke joined #git
05:51 jstein_ joined #git
05:52 acetakwas joined #git
05:53 a_thakur joined #git
05:53 GodGinrai joined #git
05:53 mischat joined #git
05:55 tirej joined #git
05:55 proq ugh.  now that that's working with ruby instead of bash, I'm hitting a repo that is asking for a username and password.  I don't see an arg to abort under that condition  :/
05:57 proq I guess I will just blacklist that repo
05:57 lu_zero joined #git
05:57 lu_zero joined #git
06:00 cyan__ joined #git
06:02 tanuki joined #git
06:04 spacelord_ joined #git
06:09 chachasmooth joined #git
06:15 aidalgol joined #git
06:15 NeverDie joined #git
06:18 rscata joined #git
06:21 Derperperd joined #git
06:22 thiago joined #git
06:23 nowhere_man joined #git
06:26 a_thakur joined #git
06:27 njbair joined #git
06:27 a_thakur joined #git
06:29 ayjay_t joined #git
06:29 overlord_tm joined #git
06:32 bocaneri joined #git
06:34 dreiss joined #git
06:35 qt-x joined #git
06:47 jimi_ joined #git
06:48 jagob joined #git
06:49 tanuki joined #git
06:55 mischat joined #git
06:55 ThomasLocke joined #git
06:55 ThomasLocke joined #git
06:56 pdgwien joined #git
06:58 Raging_Hog joined #git
07:02 zeroed joined #git
07:02 zeroed joined #git
07:03 ki0 joined #git
07:03 eizua joined #git
07:04 done joined #git
07:05 gautamsomani joined #git
07:07 ki0 joined #git
07:08 davimore_ joined #git
07:08 chachasmooth joined #git
07:09 mohabaks joined #git
07:14 chele joined #git
07:16 leeN joined #git
07:16 dreiss joined #git
07:18 zeroed joined #git
07:18 zeroed joined #git
07:21 howdoi joined #git
07:21 Gloomy joined #git
07:22 zeroed joined #git
07:22 zeroed joined #git
07:22 stick-shift joined #git
07:22 metachr0n joined #git
07:22 Derperperd joined #git
07:25 ihradek joined #git
07:26 nidr0x joined #git
07:27 neunon joined #git
07:27 t-mart__ joined #git
07:29 aavrug joined #git
07:30 ReenignE- joined #git
07:30 ReenignE- joined #git
07:30 nitemare joined #git
07:30 qqx joined #git
07:31 Bigcheese joined #git
07:31 qt-x joined #git
07:31 jnavila joined #git
07:31 zifnab joined #git
07:31 codedmart joined #git
07:31 darkmist joined #git
07:31 noahsussman joined #git
07:31 wcpan joined #git
07:31 JeroenT joined #git
07:31 shennyg joined #git
07:31 jgrasser joined #git
07:32 GameGamer43 joined #git
07:32 Fiouz joined #git
07:32 etcinit joined #git
07:32 davisonio joined #git
07:32 _joes__ joined #git
07:32 Forkk joined #git
07:32 MLM_ joined #git
07:32 chachasmooth joined #git
07:32 johtso joined #git
07:32 angelsl joined #git
07:32 flupke joined #git
07:32 addshore joined #git
07:33 sfan786_ joined #git
07:33 kurkale6ka joined #git
07:33 Lowell joined #git
07:34 gyre007 joined #git
07:34 cqi joined #git
07:34 mrnuke joined #git
07:34 ariver joined #git
07:35 martin_g joined #git
07:35 XDS2010 joined #git
07:35 n-st joined #git
07:35 cschneid joined #git
07:35 alphor joined #git
07:35 discopatrick joined #git
07:35 MacGyver joined #git
07:35 troyt joined #git
07:35 barq joined #git
07:35 Samual joined #git
07:35 da5id joined #git
07:36 hexa- joined #git
07:36 Caelum joined #git
07:36 r3c0d3x joined #git
07:36 grawity joined #git
07:37 batrick joined #git
07:37 JeroenT joined #git
07:37 jknetl joined #git
07:38 Nothing4You joined #git
07:38 romerocesar joined #git
07:39 bauruine joined #git
07:39 Guest85605 joined #git
07:41 HalfEatenPie joined #git
07:43 sbulage joined #git
07:43 Guest70788 joined #git
07:46 Sceorem joined #git
07:48 eduardas_m joined #git
07:48 pigeon joined #git
07:51 theoceaniscool joined #git
07:51 NYTimes joined #git
07:53 dminuoso I have a repo with a sub-module "foo". Now I want to swap that "foo" with a local fork on my harddrive, what's the way to do this?
07:53 dminuoso (I want to test whether the application works with my new module "foo" as intended before I shove it into the CI chain)
07:53 mircean joined #git
07:53 mircean left #git
07:53 mircean joined #git
07:54 mircean left #git
07:55 davimore__ joined #git
07:55 rivarun joined #git
07:55 mgedmin cd foo && git remote add localfork ~/path/to/local/fork; git fetch localfork; git pull --ff-only localfork/master
07:55 mischat joined #git
07:55 mircean joined #git
07:55 mgedmin then do the test, then use cd .. && git submodule update to switch back to the usual submodule content
07:56 mircean left #git
07:56 mircean joined #git
07:57 mircean joined #git
07:58 mircean left #git
07:58 jnavila_ joined #git
07:59 dminuoso mgedmin: Ahh interesting. I did not expect it to be that trivial.
08:01 barq joined #git
08:01 usil joined #git
08:03 Snugglebash joined #git
08:04 cschneid joined #git
08:04 Samual joined #git
08:04 alphor joined #git
08:04 mrnuke joined #git
08:04 Samual joined #git
08:04 Tobbi joined #git
08:05 NYTimes joined #git
08:05 TomyWork joined #git
08:06 donught joined #git
08:06 qt-x joined #git
08:06 elect joined #git
08:06 _0xbadc0de joined #git
08:06 r3c0d3x joined #git
08:07 usil hello. I have been working on branch on two different computers. I accidentally make a force push to remote from comupter where working branch was behind. Is it possible to get prevoius remote state before I make a force push? Currently I don't have an access to other computer?
08:07 gambl0re joined #git
08:07 dminuoso !password
08:07 SJr joined #git
08:08 dminuoso Can someone trigger the factoid for purging sensitive information from git?
08:08 qt-x joined #git
08:08 osse !sensitive
08:08 gitinfo [!filter_sensitive] You can use filter-branch to remove sensitive data from a repository's history. https://help.github.com/articles/remove-sensitive-data/
08:09 dminuoso Thanks
08:09 thiago joined #git
08:10 lindenle joined #git
08:11 csk157 joined #git
08:12 eizua joined #git
08:12 karstensrage_ joined #git
08:12 MrWoohoo joined #git
08:13 PHPanos joined #git
08:13 StrongBit_ joined #git
08:13 marchelly joined #git
08:14 raedah joined #git
08:14 mgedmin joined #git
08:15 dersand joined #git
08:15 a_thakur joined #git
08:15 raedah how do I take a diff between two branches and apply it to a branch as a commit or save it to apply later?
08:15 moritz go to the target branch
08:16 moritz git diff A B | git apply
08:16 moritz git commit -am "it's da diff from A to B"
08:16 nostrora joined #git
08:17 dsantiago joined #git
08:17 dsdeiz joined #git
08:17 dsdeiz joined #git
08:18 a_thakur joined #git
08:19 navidr joined #git
08:19 JeroenT joined #git
08:19 manuelschneid3r joined #git
08:20 eduardas_m joined #git
08:21 meloc joined #git
08:22 dsdeiz_ joined #git
08:23 Derperperd joined #git
08:23 swalladge joined #git
08:26 raedah moritz: easier than I thought. thks
08:29 Anja joined #git
08:29 strixx joined #git
08:32 moritz raedah: you're welcome
08:37 eduardas_m joined #git
08:37 danslo joined #git
08:39 mischat joined #git
08:39 OptimusMKD joined #git
08:40 strixx joined #git
08:40 ayjay_t joined #git
08:42 Derperperd joined #git
08:43 usil left #git
08:48 flaviodesousa joined #git
08:48 styx_ joined #git
08:49 aard_ joined #git
08:50 sim642 I merged B into A and made a commit on A after that. How do I move the commit into B and have it remerged?
08:51 sim642 rebase seems to be very fishy with merges...
08:51 thebope joined #git
08:51 cmalmber_ joined #git
08:52 satifant joined #git
08:53 hhee2 joined #git
08:53 osse sim642: you can merge A into B
08:54 sim642 I want to get rid of that ugliness in my history
08:54 sim642 the commit was supposed to be on B to begin with
08:54 tvw joined #git
08:55 strk joined #git
08:55 prg3 joined #git
08:55 dminuoso If you had a project, and you were responsible for a sub-module all by yourself, would you still follow feature branches, pushing each all the time, or would you just maintain a development branch with release tags?
08:55 dminuoso Feature branch workflow seems a little overkill for a small component that nobody else works on,
08:56 lordjancso joined #git
08:57 sim642 I guess I can manually cherry pick the commit, undo and redo the merge
08:58 strk left #git
09:00 bongjovi joined #git
09:03 sarbs joined #git
09:05 SwiftMatt joined #git
09:06 rivarun joined #git
09:07 serialoverflow joined #git
09:07 styx__ joined #git
09:10 j416 sim590: rebase eats merges.
09:10 j416 be careful
09:13 gfixler joined #git
09:14 zeroed joined #git
09:14 zeroed joined #git
09:14 mikecmpbll joined #git
09:15 King_Hual joined #git
09:15 King_Hual joined #git
09:16 basiclaser joined #git
09:17 styx_ joined #git
09:20 Seveas rebase --preserve-merges exists
09:20 onmeac joined #git
09:22 ki0 joined #git
09:25 marcogmonteiro joined #git
09:26 theoceaniscool joined #git
09:26 styx__ joined #git
09:27 dweelie joined #git
09:28 cspotcode joined #git
09:29 skelterjohn joined #git
09:30 j416 yes, I'd be careful with that as well. I've never had a real use-case for it.
09:30 j416 ymmv I guess :)
09:30 _ikke_ I used it the other day
09:31 _ikke_ I needed to fix a commit before a merge
09:31 j416 I'd just redo the merge :P
09:31 _ikke_ Right, but this worked too
09:31 j416 feels easier and safer
09:31 j416 yeh
09:32 jast the implementation of -p seem solid to me
09:32 jast +s
09:33 jast I shall blame this *pple wireless keyboard for all my typos henceforth
09:33 manuelschneid3r joined #git
09:34 j416 jast: see bugs section of man git-rebase
09:34 gitinfo jast: the git-rebase manpage is available at http://jk.gs/git-rebase.html
09:34 leeN joined #git
09:35 Morfio joined #git
09:35 j416 so I tend to avoid it altogether instead, never needed it :)
09:36 jast okay, yeah, that makes it a little tricky if you use it in conjunction with -i
09:37 jast I've got a great solution, though: let's change the format of the interactive commands list to XML!
09:37 jast makes it super easy to nest the commit correctly !11!
09:38 jast also, you can easily add more operations with clearly defined schemas by adding more XML namespaces
09:38 jast it's so beautiful I might actually be crying a little
09:39 canton7 you know what? I'd actually take that
09:39 canton7 the ability to rearrange the history of multiple branches? you can write it in HL7 and I'd still love it.
09:39 jast canton7: All that is necessary for the triumph of evil is that good men do nothing
09:40 jast lemme whip up something that uses ANSI X.12
09:40 absinthe joined #git
09:41 mischat joined #git
09:42 canton7 (rearranging the history of multiple branches is going to be a PITA to setup without graphical tooling anyway, IMO, so it doesn't matter what langauge git accepts - I'll write the tooling to spit it out)
09:43 nettoweb joined #git
09:43 mischat joined #git
09:48 muthu joined #git
09:49 muthu joined #git
09:51 ayogi joined #git
09:52 thiago joined #git
09:54 sarbs joined #git
09:58 [0xAA] joined #git
09:59 chll_ joined #git
10:01 aax joined #git
10:02 invisbl joined #git
10:06 SantaClauze joined #git
10:08 mischat joined #git
10:08 sarri joined #git
10:08 sarri joined #git
10:09 cdown joined #git
10:10 cdg joined #git
10:11 davimore joined #git
10:11 cdg joined #git
10:11 pks joined #git
10:14 astrofog joined #git
10:15 tirej joined #git
10:17 chele i am trying to understand why rebase doesn't work and merge does work when you have unstaged changes in your local master and you run git pull (--rebase).
10:17 chele any thoughts?
10:17 manuelschneid3r joined #git
10:18 mrkake joined #git
10:20 NitroLux joined #git
10:21 danslo joined #git
10:21 jast chele: rebase doesn't like doing anything when you have any uncommitted changes. you can use --autostash to make rebase stash your changes and automatically restore them if the rebase completes successfully.
10:23 jast the reason is that the way merge works, it's easy for it to know exactly which files will need to be updated, so it can distinguish situations in which your unstaged changes touch one of the files updated by the merge from situations in which they touch only unrelated files
10:23 SwiftMatt joined #git
10:25 t0by joined #git
10:25 t0by joined #git
10:25 jast rebase, on the other hand, works by resetting your head to a new upstream (up to that point a similar check would be straightforward, but resetting without throwing away unstaged changes actually isn't that easy, so that's another complication), and then look through all the commits that are going to be re-applied, before even doing anything, just to know which files will be affected
10:27 a_thakur joined #git
10:29 cyan__ joined #git
10:29 jast anyway, that'd be rather overkill, so it doesn't actually do that, instead it just refuses to touch anything with uncommitted changes
10:32 marchelly Hi, I'm not able to normally update .gitignore I have webroot dir and bare repo dir. I'm changing .gitignore adding /files/* into it to ignore all the files on this folder. Then I'm doing git rm -r --cached . and git add . and git commit and git push --all to bare repo.
10:32 marchelly on next step from another server I'm doing clone and /files/ folder and files still exists
10:34 chele jast: i see. thanks. seems to be a design problem?
10:34 chele s/problem/limitation/
10:34 ispn joined #git
10:34 jast chele: yeah, pretty much
10:36 tobiasvl marchelly: did you run "git status" before committing to see what would happen? "git add ." does not stage removals. you want "git add -u". you might want to do "git add -A" to do both
10:37 jast well 'git rm -r --cached .' would take care of removing :}
10:37 tobiasvl ah sorry
10:37 Lynxium joined #git
10:37 Serpent7776 joined #git
10:38 Lynxium joined #git
10:38 minikN joined #git
10:39 marchelly jast, so what is missed?
10:39 jast marchelly: you might want to review the contents of the commit you made, see if the removals are actually in there: git log --name-status ("D" lines are removals)
10:39 marchelly tobiasvl, yes, I cando git add -A . but as jast toldgit rm doing this
10:39 kyellow joined #git
10:39 jast in the repo where you did the 'git rm' etc.
10:40 absinthe joined #git
10:40 kyellow hi
10:40 gitinfo kyellow: 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.
10:40 kyellow hi git team
10:41 canton7 (note we're almost all users of git, not devs)
10:41 _ikke_ woohoo, we're a team :P
10:41 jast if by "git team" you mean "git developers"... most of them are not here
10:41 jast we're the git family, really :}
10:41 ayogi joined #git
10:42 chandu20 joined #git
10:42 kyellow My git server has version 1.7.12, I want to upgrade it to the latest version.
10:43 marchelly jast, no D there, lookslike the problem is in my .gitignore file. It's like http://pastebin.com/7r5n7qNH
10:43 _ikke_ kyellow: What OS?
10:43 tobiasvl marchelly: gitignore isn't applied to tracked files
10:43 kyellow What steps i should follow to avoid suffering?
10:44 kyellow my Os Debian
10:44 _ikke_ kyellow: git should be backwards compattible iirc
10:44 aw1 joined #git
10:44 _ikke_ i'm not aware of any big changes
10:44 kyellow 8
10:44 kyellow :D
10:44 Sceorem_ joined #git
10:45 marchelly tobiasvl, what is solution forthis situation? Actually I understand, that something wrong in logic but I don't know the right way for this
10:46 rnsanchez joined #git
10:46 kyellow Are we sure that my upgrade will not be harmed?
10:46 NYTimes left #git
10:47 tobiasvl marchelly: hard to say – if you did run the commands you said you did it should have worked I think. do you still have your terminal backlog so you can confirm? reset back to the previous commit and re-do the procedure, then run "git status" before committing to make sure
10:47 _ikke_ git is pretty concervative. All the bigger changes are not related to the repository format
10:47 absinthe joined #git
10:47 marchelly tobiasvl, got you. let me check it one more time
10:48 absinthe joined #git
10:48 borkr joined #git
10:49 ayogi joined #git
10:49 djb-irc joined #git
10:50 multi_io_ left #git
10:52 marchelly tobiasvl, works for me now I think git add -A . does the trick. thank you.
10:53 marchelly actually maybe you can suggest something for other issue.
10:53 marchelly let me explaing my environment
10:53 marchelly I have webroot 800gb size (images, pdf and so on). I'm going to work with php. So I'm creating .gitignore like * !*.php on remote, add . and commiting init commit.
10:54 marchelly then creating bare repo
10:55 marchelly I'm on local end. I have only php files here. I'm editing local .gitignore and adding !*.html here, then git add -f . && git commit && git push.
10:56 marchelly and now I want git give me back all html files from original 800gb repo that where ignored before.
10:56 tobiasvl how are you creating the local .gitignore?
10:57 marchelly could I do this using some commands, or what hooksI should add to git.
10:57 marchelly I've cloned itfrom remote barerepo
10:57 tobiasvl ok so it's the same .gitignore
10:57 tobiasvl I don't really understand the issue
10:58 tobiasvl go to the 800 gb repo, pull, add all the html files, commit?
10:59 marchelly tobiasvl, correct, you understand everyhing right. And now imagine the situation that I (as developer do not have access to that 800G original repo )
10:59 tobiasvl well if you don't have access to the html files, you can't add them to the repo
11:01 YuGiOhJCJ joined #git
11:02 marchelly tobiasvl, yes, I as developer do not have access to that repo and that files, but I do need. So as git administrator (let's give such a role :) ) I should update my git post-receive hook and add some lines there. For example go to that 800G repo, pull, add all (as per new gitignore it's html files) and commit? is this good way?
11:04 mehola joined #git
11:04 tobiasvl this sounds like a very weird setup. do you really need to do this?
11:05 tobiasvl that said, in theory you can do something like this, but I would not recommend you do an automatic pull (let alone in another repo) inside the hook. what if there's a conflict?
11:05 tobiasvl marchelly: also read !deploy
11:05 gitinfo marchelly: 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
11:06 cdown joined #git
11:08 Sceorem__ joined #git
11:10 marchelly tobiasvl, I've already had a discussion about deploy with moritz and agreed that the best way is to use some deployment tool, like goCD
11:10 astrofog joined #git
11:11 tobiasvl yeah
11:12 tobiasvl I guess you _could_ have a post-receive hook that goes to the other repo and does "git fetch; git reset --hard origin/master" and then adds the html files, if it never gets any local changes you care about
11:12 tobiasvl but a deploy tool is obviously better
11:13 tobiasvl just having a hook that operates on another repo gives me the creeps to begin with
11:14 marchelly :)
11:14 UberEber joined #git
11:15 jmpp joined #git
11:15 jmpp joined #git
11:15 marchelly actually yes, I've agreed with myselfthat there are no localchanges never ever, so this hook _could_ work just fine.
11:16 marchelly tobiasvl, hook that operates on anotherrepo islike Russian rulette, you know :)
11:17 JeroenT joined #git
11:18 sbulage joined #git
11:18 JeroenT joined #git
11:18 csk157 joined #git
11:18 osse if you cd to a different repo in a hook remember to unset GIT_DIR and such
11:20 ferr1 joined #git
11:22 marchelly osse, thanks
11:24 UberEber good morning
11:24 raspado joined #git
11:24 UberEber anyone around to help a newbie with an implementation strategy?
11:25 UberEber maybe I should just post the situation?
11:26 jast good idea
11:27 UberEber kk...here goes...ty Jast was not sure my post was visible ...here goes
11:28 UberEber I am not the server admin and he does not know about git and I want to run parallel to prove it with work for them...
11:28 UberEber the files are autocad, the drawings, the post processed files for CNC, and an info file about the CNC file
11:29 UberEber I can alter the CNC and Info file only
11:29 UberEber the CNC file is generated by a post processor after autocad creates the dimension etc
11:30 UberEber there are 3 win directory levels with about 20 directories in each level totalling 3 gig of data
11:31 UberEber I have a complete image directly off the server of all the info
11:32 UberEber if i create a repository of the whole thing ...and in a month from now get a new file/copy of the info..
11:32 jast what is being used now, instead of git?
11:32 UberEber and then copy/replace the entire directory into my repository directories containing git ...will it break git or will the files that have changed show up?
11:33 jast as long as you don't overwrite the .git subdir you should be fine
11:34 UberEber so if a file in a subdir is changed git will say newfile or changed or what?
11:34 UberEber ty Jast
11:34 jast git does special magic to make detection of changes faster, but it's not evil black magic ;)
11:37 Snugglebash joined #git
11:38 Snugglebash joined #git
11:39 UberEber if a tracked file gets replaced by a windows copy replace from a non git user (user did not pull file) what happens?
11:39 khelkun joined #git
11:39 Snugglebash joined #git
11:39 pijiu joined #git
11:40 Snugglebash joined #git
11:40 fahadash joined #git
11:40 Snugglebash joined #git
11:41 thiago joined #git
11:41 Snugglebash joined #git
11:41 osse UberEber: the local git will see the file as changed
11:42 Snugglebash joined #git
11:43 Snugglebash joined #git
11:43 joki joined #git
11:43 UberEber osse: ty ...will local git have the original version for history purposes and thus will the original file be able to be repulled and recommitted in the future if needed?
11:44 Snugglebash joined #git
11:44 Walex joined #git
11:45 osse UberEber: not sure i understand the question.
11:45 thierryp_ joined #git
11:45 osse fi you make changes to a file git will see it's changed. if you replace the file with a different one git will just think you edited it a lot
11:45 _0xbadc0de joined #git
11:46 osse looking at logs etc. works just as well independently
11:47 cdg_ joined #git
11:48 blgria joined #git
11:49 wendelmineiro joined #git
11:49 UberEber osse: so I .git track file A in master...a non .git user (did not take it into a branch) ...changes the file....git sees it as changed ...but can I revert back to the old file?
11:49 osse UberEber: git checkout -- file
11:50 UberEber osse: yes I will checkout files ... but at work others do not know how to do that yet... what happens when the non-.git user overwrites a file ...I can not see how .git can revert to the old one
11:51 osse UberEber: it cannot be done automatically
11:51 afuentes joined #git
11:51 osse someone needs to check
11:51 ertesx joined #git
11:51 tobiasvl do people have write access to the deployed code? what's the scenario here
11:52 UberEber scenario is ....
11:52 UberEber I have copy of server info ...engineering drawings etc
11:53 UberEber others are NOT using git
11:53 UberEber I want to create a git repo of the entire info
11:53 UberEber when a non git user changes a file / replaces it with new
11:54 UberEber i believe git sees it as changed ...and I retrack that file ...but can it undo the non git users change?
11:55 bremner git is really not designed to help you managed a shared work directory
11:56 bremner it is oriented to helping you manage your own copy
11:56 UberEber ty bremner ...maybe there is still a way...
11:56 bremner well, any round peg can be smashed into a square hole eventually.
11:56 UberEber I get the info on a flash drive ...all 3 gig of files monthly ...
11:57 Shentino joined #git
11:57 nikivi how can I revert to a specific commit that I've done
11:57 UberEber most of the files do not change
11:57 nikivi I have tried git commit {unique commit id}
11:57 nikivi but that didn't work
11:57 tobiasvl nikivi: !revert
11:57 gitinfo nikivi: That's a rather ambiguous question... options: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard uncommitted changes in the working tree [git reset --hard]; c) undo committing [git reset --soft HEAD^]; d) restore staged versions of files [git checkout -p]; e) move the current branch to a different point(possibly losing commits)[git reset --hard $COMMIT]?
11:57 UberEber could I .git the monthly update and merge them some how?
11:57 Serg_Penguin joined #git
11:57 nikivi can I git reset --hard to some point
11:58 nikivi like specify to what commit I want to reset?
11:58 bremner UberEber: ok, that should work if you only want git to track changes in your local copy. But if you ignore all changes other people make, why download the 3G
11:58 tobiasvl UberEber: why don't you just change the workflow to a fully git environment?
11:58 tobiasvl nikivi: yes that it what you can do, see e) in the above gitinfo reply ^
11:58 tobiasvl syntax
11:59 csk157_ joined #git
11:59 Serg_Penguin hi ! how can I return several commits to 'modified' status ? I don't want to lose changes, just to re-group them - misclicked files in GUI
11:59 UberEber bremner: I want to track all changes of everyone ...but I get full copies of the info once per month as a file copy
11:59 UberEber tobiasv1: thats what I want to do ...but need to prove to the bosses first
12:00 UberEber run parralel for six months and show them
12:00 UberEber omg ...truly they still use typewriters in my company...NO JOKE
12:00 tobiasvl Serg_Penguin: !unstage
12:00 gitinfo Serg_Penguin: To unstage a file from the index(in other words, to undo an accidental `git add foo/bar.txt`), use: git reset -- foo/bar.txt
12:01 gabi_ joined #git
12:01 tobiasvl Serg_Penguin: oh, commits?
12:01 Serg_Penguin i want to revert them from commits, not staging
12:01 tobiasvl sorry
12:02 tobiasvl Serg_Penguin: git reset --soft COMMIT_YOU_WANT_TO_REVERT_TO
12:02 strixx joined #git
12:02 UberEber thanx all ....I've got a handle on what I need to study next
12:02 UberEber have a great day
12:04 nd_ joined #git
12:04 Serg_Penguin tobiasvl, thanks a lot ! i tried to google but only found how to revert, i.e. destroy chenges, considering them harmful
12:05 rchavik joined #git
12:05 tobiasvl Serg_Penguin: yeah, reset --hard will destroy changes. never use --hard or --force in commands unless you know what you're doing. --soft is safe though
12:05 JeroenT joined #git
12:08 pks joined #git
12:13 zeroed joined #git
12:13 zeroed joined #git
12:14 elsevero joined #git
12:14 ankit01ojha joined #git
12:14 ankit01ojha Hello :)
12:14 gitinfo ankit01ojha: 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.
12:16 ankit01ojha I am having  a problem the problem is like this - I deleted my old local repo and I have pushed some code from it on github which has made a pull request page , now I cloned a new one and but I want to push my code to the previous pull request page.
12:17 ankit01ojha how can I do this?
12:18 tobiasvl ankit01ojha: you can just push to the branch the pull request is from to update it
12:18 Snugglebash joined #git
12:19 raspado joined #git
12:21 Cavallari joined #git
12:22 byte512 joined #git
12:27 Mir4g3 joined #git
12:28 drbean joined #git
12:28 sgfgdf joined #git
12:28 manuelschneid3r joined #git
12:28 ToBeCloud joined #git
12:29 jameser joined #git
12:30 JeroenT joined #git
12:30 sgfgdf hello, guys! i have a feature branch which at some point was merged of development instead of rebased. is it possible to do a clean rebase and show only code related to the particular feature?
12:32 rominronin joined #git
12:33 muthu joined #git
12:34 osse sgfgdf: You can rebase afterwards without problems
12:35 theoceaniscool joined #git
12:36 sgfgdf osse: but will it remove those merges and when make diff on the two branches show only related changes?
12:36 i_dont_know_how_ joined #git
12:36 osse Yes
12:36 z1haze_work i have a weird situation with a git repo here.. Someone has somehow committed and pushed the .git folder to the repository, and i cant figure out how to remove it
12:37 i_dont_know_how_ git rm?
12:37 marchelly tobiasvl, osse I'm fixing post-receive for that autofetch/push. Curently there is a checkout -f now, should I keep it? As it will be treated as "local changes" and just do git pull?
12:38 z1haze_work then i would be removing the only .git folder i have..
12:38 z1haze_work i dont understand how that even ended up in the repository
12:39 z1haze_work ive googled it several ways and nothing comes up.. maybe this is a first
12:39 Shentino joined #git
12:39 osse z1haze_work: Where is it located?
12:39 i_dont_know_how_ Hey guys, i'm trying to checkout a specific folder from a remote branch and merge it into another, according to an article i read, i can use "git checkout <branch> -- <path to file or directory>", which i've attempted, and git tells me "fatal: cannot update paths and switch to origin/<remote branch> at the same time" and "Did you intent to checkout '<path to folder>' which cannot be resolved as commit?"
12:39 marlun joined #git
12:39 i_dont_know_how_ z1haze_work: i could have sworn there's a way to git rm without removing it locall
12:39 i_dont_know_how_ y
12:39 osse i_dont_know_how_: --cached
12:39 watmm1 left #git
12:40 i_dont_know_how_ i've never even heard of someone getting a .git folder in the repo before, lol
12:40 marlun How do I see the commits that has come in on master after I created <branch>?
12:40 i_dont_know_how_ i suppose it's always technically there though, right?
12:40 jimi_ joined #git
12:40 marchelly tobiasvl, osse here is it http://pastebin.com/V6iGu6db
12:40 i_dont_know_how_ i mean git's gotta know about the repo so it has to have it, maybe it's always there but just hidden?
12:40 marlun The ones that I would "get" if I rebased on master.
12:40 i_dont_know_how_ my only other guess would be to gitignore .git but that seems like it'd probably break something lol
12:41 JeroenT joined #git
12:41 djb-irc joined #git
12:42 tobiasvl marlun: git diff branch..master ?
12:42 i_dont_know_how_ z1haze_work: git rm -r --cached .git would be my assumtion on how to remove it, but like i said, i dont know if that's actually supposed to be there and just hidden or not?
12:42 marchelly tobiasvl, whoat do you mean?
12:42 marchelly oh, it's not for me
12:43 marchelly okay
12:43 i_dont_know_how_ What's the deal with nobody else having voice, is it only unregistered users get voice so they can speak because non-authed users are muted?
12:44 sarri joined #git
12:44 sarri joined #git
12:44 marlun tobiasvl: ....I think I made it harder in my head :) thanks!
12:44 z1haze_work yea i have no clue how this happened
12:45 maret joined #git
12:45 z1haze_work but i dont know how to delete it either.. someone suggested with --get-dir im gonna try that
12:45 osse marchelly: i have no idea what i'm looking at :O
12:45 tobiasvl i_dont_know_how_: yeah registered users can speak normally, unregistered once can do /q gitinfo .voice
12:45 osse z1haze_work: Where is it located?
12:45 i_dont_know_how_ z1haze_work: we just told you how to delete it.
12:45 i_dont_know_how_ z1haze_work: git rm -r --cached .git would be my assumtion on how to remove it, but like i said, i dont know if that's actually supposed to be there and just hidden or not?
12:45 marchelly osse, this is post-receive script
12:46 osse if there is a .git that is not the main one at top level then Git thinks it's a submodule
12:46 z1haze_work so --cached deletes the remote instance?
12:46 i_dont_know_how_ yes, and leaves the local.
12:46 buzzLightBeer joined #git
12:46 i_dont_know_how_ so you can purge a file from the repo without deleting it locally.
12:46 z1haze_work i gotcha
12:46 newbie1 joined #git
12:46 i_dont_know_how_ anyone have any suggestions about my issue?
12:47 osse that's a bit misleading. you still have to commit and push
12:47 i_dont_know_how_ what, me?
12:47 i_dont_know_how_ i mean, yeah, of course you have to commit the changes you made and push them for it to take effect in the repo
12:47 i_dont_know_how_ but that;s pretty much common sense and goes without saying, no?
12:47 z1haze_work so i get .git did not match any files
12:48 osse i_dont_know_how_: sure. i just thought it was weird talking about the "remote instance"
12:48 i_dont_know_how_ then .git isn't there, as far as git is concerned
12:48 osse i_dont_know_how_: as for your checkout thing, seems like a typo
12:48 osse can you !repro
12:48 gitinfo [!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
12:48 Tobbi joined #git
12:49 marchelly osse, yep, have to push
12:50 i_dont_know_how_ ot
12:50 i_dont_know_how_ it's definitley not a type on the branch or anything
12:50 jameser joined #git
12:50 i_dont_know_how_ i copied the text for branch and filepath straight from github
12:51 osse is there a -b involved?
12:51 i_dont_know_how_ no
12:51 osse i don't know then
12:51 i_dont_know_how_ my exact command was
12:51 i_dont_know_how_ oh i lied
12:51 i_dont_know_how_ i did it with and without b
12:51 i_dont_know_how_ neither worked
12:51 i_dont_know_how_ "git checkout -b origin/shipping-replacement -- app/code/Revival/ShippingNames"
12:51 osse in any case -b is wrong
12:51 osse do you get the same error without it?
12:52 i_dont_know_how_ error: pathspec 'app/code/Revival/ShippingNames' did not match any file(s) known to git.
12:52 i_dont_know_how_ except they're in that repo
12:52 i_dont_know_how_ in that branch
12:52 osse that's better
12:52 i_dont_know_how_ :v
12:52 osse try :/app/code/Revival/ShippingNames
12:52 i_dont_know_how_ cant's switch to /app/ no such file or directory
12:53 osse I mean git checkout origin/shipping-replacement -- :/app/code/Revival/ShippingNames
12:53 i_dont_know_how_ or did i need to literally put ":/app/"
12:53 i_dont_know_how_ oh
12:53 i_dont_know_how_ you did want the colon
12:53 i_dont_know_how_ okay i thought you had an extra space there
12:53 jast yep
12:53 i_dont_know_how_ hey look it didn't error
12:53 osse are you currently in a subdir =
12:53 i_dont_know_how_ aaaaaaaaaaay \o/
12:53 jast then it probably did its job :)
12:53 i_dont_know_how_ it did indeed
12:54 i_dont_know_how_ thanks guys!
12:54 jast for future reference, in 'git checkout foo -- bar', the bar should normally be relative to the current directory
12:54 i_dont_know_how_ oh, should it?
12:54 i_dont_know_how_ i thought it was just the path on remote
12:54 jast and the :/ basically bypasses that
12:54 i_dont_know_how_ i was inside app/code/Revival
12:54 mehola joined #git
12:54 i_dont_know_how_ because i didn't want it to pull somewhere else on accident
12:55 osse there is no pull, or merge, or commit creation
12:55 osse all you do is get files from a different branch and dump them into your current one
12:56 jameser joined #git
12:57 ojdo joined #git
12:58 Snugglebash joined #git
12:59 synthroid joined #git
13:00 i_dont_know_how_ the general concept of branches just kinda bothers me though lol
13:00 i_dont_know_how_ you have multiple file versions on multiple branches
13:00 i_dont_know_how_ and at any given time git can just swap what you're looking at in the local filesystem
13:01 Tee_Pee joined #git
13:01 i_dont_know_how_ which is pretty strange to me
13:01 osse git only does that if you ask it to
13:01 bremner no more than your editor can change a file. There's no magic.
13:01 osse and yes, git can maintain several branches, and what you have on disk is one of them at a time
13:01 [0xAA] joined #git
13:01 osse {deal with it gif here}
13:02 i_dont_know_how_ well yeah, but still, it's just strange that you basically have multiple what i used to think to be a repo on one repo
13:03 i_dont_know_how_ anyways, my days of lurking in irc are long over, i just popped in to get some help
13:03 i_dont_know_how_ it was nice talking to you guys, thanks for all the help!
13:03 i_dont_know_how_ Have a nice day :)
13:04 i_dont_know_how_ left #git
13:06 bremner Today on News 11, more ways that git is not svn.
13:06 rchavik joined #git
13:07 [0xAA] joined #git
13:08 zyrb joined #git
13:12 marchelly in another words if I already made checkout do I need to make pull? Or if I'm doing pull should i do checkout before it. Next step will be add -A . && commit && push back
13:17 noecc joined #git
13:18 synthroi_ joined #git
13:18 michael_mbp joined #git
13:18 zyrb left #git
13:20 sgfgdf osse: i endup in a lot of conflicts. and resolved all of them by using git checkout --ours ... and continue rebase. most of the time i needed to do git rebase --skip, because there was nothing to apply. is that the right approach if i want to keep the code from my branch?
13:24 shiriru joined #git
13:24 tobiasvl marchelly: well you never fetch, do you?
13:24 tobiasvl fetch, then checkout. don't pull, that does a merge, which can fail if there are conflicts
13:25 tobiasBora joined #git
13:25 finalbeta joined #git
13:26 tobiasvl checkout to the remote-tracking branch, that is
13:26 thiago joined #git
13:26 osse sgfgdf: in that case i have misunderstood the situation
13:27 osse sgfgdf: I thought you wanted to do 'git rebase blah' but did 'git merge blah' by accident
13:28 dsdeiz joined #git
13:28 dsdeiz joined #git
13:28 csk157_ joined #git
13:30 JeroenT joined #git
13:30 mizu_no_oto joined #git
13:33 maret joined #git
13:37 maret joined #git
13:38 SlashLife^m joined #git
13:38 SlashLife^m Morning :)
13:40 sarri joined #git
13:40 sarri joined #git
13:40 osse Has broken
13:40 osse Like the first morning
13:40 SlashLife^m Are there any things which commit objects will always reference? I already read through the internal workings of a Git repo on git-scm.com and now I'm pondering if I could use a server side hook to achieve what I want. (following)
13:41 SlashLife^m osse: SlashLife has spoken like the first slash?
13:41 SlashLife^m Oh, like the first life. :(
13:41 grawity SlashLife^m: what *do* you want?
13:41 sdothum joined #git
13:41 SlashLife^m Basically what I'm thinking is to have short-lived branches for features which then get merged back into the master branch. Basically on the master branch: One commit = one feature.
13:42 SlashLife^m Now I'd like to have "revision numbers" on the master branch, according to the commits made. Hashes are a bit tedious to grasp as a human being.
13:42 SlashLife^m My idea would be to let a server side hook change a counter in a file and add that file to every commit added to the master branch.
13:42 SlashLife^m s/added/merged/
13:43 grawity that might get annoying
13:43 brent__ joined #git
13:43 SlashLife^m In what way?
13:43 grawity if all merges are done *to* master, how about just calculating it with `git rev-list --count HEAD`
13:44 Zamerick joined #git
13:44 grawity or even `git describe`, which counts revisions since the last tag
13:44 grawity hmm *rev-list --count --first-parent
13:45 SlashLife^m Basically I'd like to use that file to build the revision number into the project, so that when you run the software, you can refer to that revision - and that people without access to the repo can still have a grasp of deltas of separate builds.
13:45 grawity `git describe` sounds good enough for that
13:47 synthroid joined #git
13:47 osse gcc -DGIT_VERSION=\"$(git describe)\"" ez
13:47 PHPanos joined #git
13:47 Gsham joined #git
13:48 JeroenT joined #git
13:48 jameser joined #git
13:49 SlashLife^m "git describe" just by itself just gives me an error message ("fatal: No names found, cannot describe anything."); git describe --all just gives me "heads/master". Maybe I'm misunderstanding git describe.
13:50 osse SlashLife^m: ah, because you have no tags
13:50 osse try git describe --always
13:50 Seveas SlashLife^m: you don't have any tags for your releases?
13:50 SlashLife^m Seveas: This is a test repo; there is no real repo yet, we're still planning.
13:50 SlashLife^m git describe --always gives me a shortened hash.
13:50 Seveas SlashLife^m: git describe works best on tags
13:51 Seveas $ git describe
13:51 Seveas v2.11.0-295-gd7dffce1ce
13:51 osse SlashLife^m: it'll continue to do that until you get a tag. after that it starts to make more sense
13:51 SlashLife^m Seveas: I can add a "root" tag, that wouldn't be a problem. But I'd like the revision number to manifest inside the repo as well.
13:51 Seveas this means "295 commits ahead of v.2.11.0, at that sha1"
13:52 Seveas SlashLife^m: you'll have to manually do that. Automatic version numbers make very little sense in a distributed branchy repo
13:52 jeffreylevesque joined #git
13:52 * osse is somewhat disappointed in how this is handled in git.git itself
13:52 SlashLife^m As I said: I'd do that on the central repo as a server side hook for one specific branch only
13:53 SlashLife^m All versions which leave the dev team for testing will come from that branch or an even higher level one.
13:55 osse but is there in your opinion a disadvantage to not doing it centrally on just one branch, but have it as part of the regular build process?
13:55 osse because it sounds like that would be easier
13:55 yinflying2016 joined #git
13:56 GodGinrai joined #git
13:56 mehola joined #git
13:57 theoceaniscool joined #git
13:57 SlashLife^m It would definitely be easier, yes, and it's a nice and elegant solution.
13:58 SlashLife^m We have some ... "specific" requirements when it comes to documenting stuff, though, so if we use this to refer to versions, it makes some bureaucracy stuff much simpler for us, if we have a simple source for the number and can lateron still point to that.
13:59 SlashLife^m With the "git describe" one we'd need to set up a document describing in detail why we can trust that value not to possibly be a different one in the future.
13:59 SlashLife^m (As a rough outline.)
14:00 osse "It's a cryptographic checksum", with links to what those words mean. The End!
14:00 osse :D
14:01 SlashLife^m With the counter we can relatively simply point at the documented process that external builds can only come from this or a higher level branch, and point at the history that each commit has its own, unique, always increasing number.
14:01 lindenle joined #git
14:02 SlashLife^m Anyway; thanks for the input ... guess I'll have to go through gits guts a bit more. :)
14:02 osse Depending on how set in stone the bureaucracy is it might be easier to do something else
14:02 osse But based on what reqs you have git describe will get the job done
14:03 osse "The bureaucracy is expanding to meet the needs of the expanding bureaucracy"
14:03 SlashLife^m Maybe a post-receive hook amending a commit with an incremented revision number could make this a lot easier than attempting to twist the blobs manually.
14:03 osse that'll royally screw everyone over
14:03 eizua joined #git
14:03 osse that's basically history rewriting, so when people pull it will be a mess
14:04 synthroi_ joined #git
14:04 osse if you add a new commit, then it might work
14:04 SlashLife^m Yeah, basically the moment that commit comes, it'd be amended instantly.
14:05 osse and if you twist blobs manually then Git consideres the repo corrupted because checksums no longer match
14:05 tobiasvl then nobody can pull from the repo ever, basically
14:05 osse SlashLife^m: "instantly" is already too late
14:05 SlashLife^m Ugh. Checksums. m( Right.
14:06 osse SlashLife^m: yup, the very same checksums that makes me say you can trust the git describe information forever :)
14:06 netj joined #git
14:06 canton7 SlashLife^m, hg has revision numbers..
14:07 SlashLife^m Great \o/ Not an option. /o\
14:08 canton7 oh well, worth a punt
14:08 SlashLife^m Mhh. Maybe a client side hook to increment the revision number pre-push, as well as a server-side hook to just check it.
14:08 mischat_ joined #git
14:08 SlashLife^m canton7: I'm thankful for all constructive suggestions, even if I have to reject them. :)
14:09 SlashLife^m Anyway; I'll go back to attempting to wrap my head around how to best solve this to get these requirements implemented.
14:09 SlashLife^m Thanks for the help.
14:10 canton7 SlashLife^m, what's a "higher level branch" btw?
14:10 SlashLife^m Roughly: Features -> Dev -> Reviewed -> Release (although that'll likely be tags on the Reviewed branch)
14:11 anonymuse joined #git
14:12 SlashLife^m Basically a lot of --no-ff going on to have a chance for more documentation.
14:12 canton7 from what you were saying earlier, that means that once you've done a release from Reviewed you can never do a release from Dev, ever
14:12 canton7 ... so there's some information which I'm missing
14:13 marchelly joined #git
14:13 regedit joined #git
14:14 ExoUNX joined #git
14:14 SlashLife^m canton7: Features would get merged back into Dev (=master), Dev would repeatedly get merged into Reviewed, occasionally commits in reviewed will get a release tag.
14:15 SlashLife^m Dev -> Reviewed basically could be ff merges, but additional documentation is required.
14:15 Snugglebash joined #git
14:16 canton7 ok, and the incrementing revision number is needed to prove that..?
14:16 dmadaan_ joined #git
14:16 tobiasvl that basically sounds like regular gitflow, doesn't it?
14:16 Snugglebash joined #git
14:16 dmadaan_ left #git
14:17 jimi_ joined #git
14:17 mischat joined #git
14:17 Samy__ joined #git
14:17 Snugglebash joined #git
14:17 SlashLife^m Doesn't GitFlow create a branch per release?
14:17 Samy__ hi, can we make a github page for any repo or only our username?
14:17 canton7 yeha, it does
14:18 Gsham joined #git
14:18 SlashLife^m We're in the lucky position (though we have other caveats) to not have to support multiple software versions in parallel.
14:18 tobiasvl yeah, therefore "basically"
14:18 SlashLife^m Yeah, okay.
14:18 SlashLife^m Yeah, I looked a lot at GitFlow and while we can't use it exactly, it was quite useful.
14:19 a_thakur joined #git
14:19 tobiasvl !flow has some other similar suggestions
14:19 gitinfo [!gitflow] The description of the gitflow branch workflow model is at http://nvie.com/posts/a-successful-git-branching-model/ while a tool to help implement this workflow is at https://github.com/petervanderdoes/gitflow  See http://sethrobertson.github.com/GitBestPractices/#workflow for other workflow suggestions/references
14:20 hasc joined #git
14:20 mischat joined #git
14:21 SlashLife^m Cool, thanks!
14:21 nevodka joined #git
14:22 PHPanos joined #git
14:23 Bizkit joined #git
14:23 aax joined #git
14:24 mischat joined #git
14:24 cdg joined #git
14:25 Kaisyu joined #git
14:26 d1z joined #git
14:26 pockets joined #git
14:27 maret joined #git
14:28 SpeakerToMeat joined #git
14:29 ShekharReddy joined #git
14:30 jnavila_ joined #git
14:33 todd_dsm joined #git
14:33 Dougie187 joined #git
14:33 fahadash joined #git
14:34 boombatower joined #git
14:37 crose joined #git
14:37 ChiLL_SurF joined #git
14:37 jameser joined #git
14:40 snowkidind joined #git
14:42 treia joined #git
14:42 anonymus_ joined #git
14:42 HardlySeen joined #git
14:43 Discovery joined #git
14:44 ash_workz joined #git
14:45 ayjay_t joined #git
14:45 Ryanar joined #git
14:45 theoceaniscool joined #git
14:47 markus3 joined #git
14:47 mischat joined #git
14:48 anonymuse joined #git
14:48 mischat joined #git
14:51 markus3 hi! is there a way to squash fixup commits while keeping the merge structure? I use the --no-ff when merging so every feature is in a branch starting from and merged to master. The original commits are in those feature branches.
14:51 markus3 I tried the -p from here: http://stackoverflow.com/questions/37069340/git-squash-commits-in-history-keeping-graph-structure#37069869, but ended up in a weird state
14:54 a_thakur joined #git
14:56 a_thakur joined #git
14:58 publio joined #git
15:01 StrongBit_ joined #git
15:01 Es0teric joined #git
15:01 anonymuse joined #git
15:01 chardan joined #git
15:02 chardan_ joined #git
15:04 a_thakur joined #git
15:05 AaronMT joined #git
15:06 xmj joined #git
15:06 xmj moin
15:07 xmj is there a way to see if a clone has been fully completed?
15:07 xmj i mean, other than checking path/.git is nonzero
15:07 aavrug joined #git
15:08 jast semi-educated guess: check that path/.git/HEAD exists
15:08 jast though it might actually exist before the clone is completely finished
15:09 Murii joined #git
15:11 Fuco joined #git
15:12 xmj gotcha
15:12 LBV_User joined #git
15:12 xmj that could work
15:12 thiago joined #git
15:12 Fuco left #git
15:12 Fuco joined #git
15:13 Fuco left #git
15:14 Fuco joined #git
15:14 Fuco I have a repository with the .git directory deleted.  Now the origin/master is some 400 commits ahead.  If I clone that and copy the .git over, can I expect it to work?
15:15 Fuco I know the commit that is checked out
15:18 dviola joined #git
15:19 xmj Fuco: if you didnt change anything locally (e.g. commit stuff yourself) that should work
15:20 nettoweb joined #git
15:22 shinnya joined #git
15:22 exarch joined #git
15:24 gautamsomani joined #git
15:24 siddart joined #git
15:24 Raed joined #git
15:25 al-damiri joined #git
15:25 cyan__ joined #git
15:26 zivester joined #git
15:26 noecc joined #git
15:27 GodGinrai joined #git
15:28 multi_io joined #git
15:28 LeBlaaanc joined #git
15:30 trasz_ joined #git
15:30 daynaskully joined #git
15:30 trasz_ hi, i have a 730MB repo that's empty (as in, 'ls -al' shows only the .git directory); there's a single commit, and a single branch.  what gives?
15:31 chardan_ joined #git
15:32 jast trasz_: either it's a big commit :) or there are unreferenced objects (git keeps things around for quite a while for safety reasons)
15:32 cdown joined #git
15:32 jast for details see the section about shrinking a repository in man git-filter-branch
15:32 gitinfo the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html
15:33 roelmonnens joined #git
15:34 oskarkv joined #git
15:37 minikN joined #git
15:39 robotroll joined #git
15:39 mizu_no_oto joined #git
15:40 voldz1tic3n joined #git
15:40 timvisher joined #git
15:42 LionsMane joined #git
15:43 danslo joined #git
15:44 ResidentBiscuit joined #git
15:48 Raging_Hog joined #git
15:48 phanimahesh joined #git
15:48 mischat_ joined #git
15:49 Darren_ joined #git
15:49 grilled-cheese joined #git
15:53 parasite_ joined #git
15:55 timvisher joined #git
15:55 Domhack joined #git
15:57 karstensrage_ left #git
15:58 mehola joined #git
16:01 nickabbey joined #git
16:03 aarobc joined #git
16:05 Es0teric joined #git
16:09 erebel55 joined #git
16:09 erebel55 lo
16:10 osse l
16:11 erebel55 does anyone know how to get a more detailed breakdown of what is taking up space in a git repo? specifically when it comes to commit history? My repo is 8.3GB but when I pull it is only 4.7 GB. So I would like to figure out what I can do to reduce the size of it some
16:11 osse erebel55: hmm, what do you mean? what are those numbers?
16:11 osse everything is 8.3 and .git alone is 4.7 ?
16:12 erebel55 *when I clone it is only 4.7 GB
16:12 erebel55 yeah gitlab is telling me that my repo size is 8.3GB
16:12 osse no clue how gitlab calculates that
16:12 crayon joined #git
16:13 osse but if a fresh clone is 4.7GB then that's close to how compact it's going to get without removing stuff from history
16:15 erebel55 okay, well when I look at .git it is 4.76GB. I also have a file in git lfs but it is only 500MB. I guess I will have to ask gitlab how they are getting 8.3 then because it doesn't make sense to me
16:15 osse erebel55: maybe it simply hasn't been cleaned in a while (git gc and so on)
16:16 osse erebel55: out of quriosity, how large is the checkout?
16:16 osse ie the whole shebang minus .git
16:16 erebel55 yeah I just ran that because it ballooned to 14GB, because of a big binary. But I removed the binary and used bfg to get rid of that in the commit history. But gitlab is now reading 8.3
16:16 brent__ joined #git
16:17 erebel55 the checkout is 2.25 GB
16:17 gugah joined #git
16:18 jottr joined #git
16:18 csk157_ joined #git
16:18 Samy__ hi, can we make a github page for any repo ?
16:19 ash_workz drat `09:14 MSTMajor service outage.` from github
16:19 refried_ joined #git
16:19 seni joined #git
16:19 seni could I stash changes, hard reset a couple commits back, then apply the changes again?
16:19 osse yes
16:19 seni nice
16:20 Oebele joined #git
16:22 Silox| joined #git
16:22 erebel55 osse, do you think the repo size is .git + checkout + git lfs files?
16:22 fission6 joined #git
16:22 fission6 whats the deal?
16:22 osse erebel55: that sounds reasonable. but i have no idea really
16:23 erebel55 okay, I'll ask gitlab
16:23 thekingofbandit joined #git
16:23 erebel55 thank you
16:23 Samy__ is github down? https://github.com/ is not showing page
16:23 Xe Samy__: https://status.github.com/
16:24 Samy__ Xe: that shows an error  16:14 UTC Major service outage.  ?
16:24 Xe Samy__: http://i.imgur.com/8mXFt5N.png
16:24 retroj joined #git
16:24 Xe github is down right now
16:24 Xe go take a long lunch break
16:24 Xe it'll be back when you're done
16:25 Samy__ :D
16:25 Samy__ what's the reason ? just curious
16:25 retroj joined #git
16:25 retroj hello.  was rsync removed from the Git for Windows distribution, and if so, is there another source for that binary?
16:26 stevenbenner joined #git
16:26 Xe the reason is that github is down, if you check status.github.com you'll find out information just the same anyone else does
16:26 mogio joined #git
16:26 TomyLobo joined #git
16:28 mischat joined #git
16:29 d1z joined #git
16:30 pa joined #git
16:30 pa hi
16:30 pa sorry for coming here with help requests.. I keep getting my git repo in the state that pushing to the remote complains with "no changes made". It's the second time in two days, and i don't know a) why and b) how to get it unstuck
16:31 pa i even tried to checkout the base, then create a whole new patch, with some differences, and push as new, but nothing
16:31 pa how can i diagnose how git got stuck?
16:31 osse pa: have you made a commit?
16:32 pa yes
16:32 pa and i'm trying to push it
16:32 canton7 output of 'git status'?
16:32 pa nothing to commit, working directory clean
16:32 perlpilot pa: you're not using some tool that pushes for you in the background (several editor plugins do that)
16:32 canton7 pa, no the whole thing, please
16:33 momo joined #git
16:33 osse and output of   git log @{u}..@
16:33 pa perlpilot, no, just git push
16:33 Raimondi joined #git
16:33 tang^ joined #git
16:33 canton7 pa, my suspicion is that you don't have a branch checked out, and the output of 'git status' will indicate that
16:33 Raimondi left #git
16:33 jccn joined #git
16:34 pa osse, that line tells me "fatal: no upstream configured for branch 'rotatingMapGestureArea'
16:34 pa (which is my local branch i created)
16:34 canton7 pa, git version?
16:34 osse pa: in that case you might need to do  git push -u origin rotatingMapGestureArea
16:34 strixx joined #git
16:34 pa canton7, 2.7.4
16:35 canton7 'git config push.default' ?
16:35 retroj looks like i was mistaken about rsync being in Git for Windows - that binary came from git-annex for windows.  problem solved :-)
16:35 pa canton7, that is empty
16:35 canton7 (I thought 'simple' complained if the current branch didn't exist on the remote?)
16:36 jast no, it complains if the names are different
16:36 pa i can try to make a patch and apply it on top of the base i want
16:36 canton7 pa, try osse's command
16:36 theoceaniscool joined #git
16:36 osse Yes! \o/
16:36 canton7 pa, that will create a new remote branch on the remote if one doesn't already exist
16:36 canton7 (and set up stuff so that 'git pull' (with no arguments) fetches from that remote branch)
16:36 nell57 joined #git
16:36 jast 'simple' doesn't auto-guess the upstream
16:37 canton7 ah ok, so if the branch doesn't exist on the remote it just silently exits? Seems a bit counter-intuitive
16:38 jast it doesn't do an existence check at all
16:38 jast 'simple' is exactly equivalent to 'upstream' except it refuses to do anything if the remote branch name differs from the local branch name
16:38 osse I get "fatal: The current branch local has no upstream branch"
16:38 pa question: how do i apply the result of git show > somefile?
16:38 Malgorath joined #git
16:39 grill joined #git
16:39 jast oh, right... well, it may be a git 1.x client
16:39 howdoi joined #git
16:39 grill hi. what's the quickest way to get a list of files that have changed in my branch throughout its existence?
16:39 jast meaning it'd still default to 'matching'
16:39 canton7 jast no, it's 2.7.4 - I checked that
16:39 csk157_ joined #git
16:39 jast grill: git whatchanged?
16:39 _ikke_ git diff --name-status <upstream>..<branch>
16:39 _ikke_ grill: ^
16:39 osse pa: the usual patch program can do that. the gitty way would be using 'git apply'
16:40 grill cool. thanks
16:40 pa ah thanks i try git apply
16:40 pa lets see
16:40 _ikke_ grill: use three
16:40 _ikke_ three dots instead of two
16:40 Malgorath Is there a webserver that can look at git branches individually? like something I can run locally when I do a branch and want to have someone test that branch in my (using php on this for now) malgorath-101 branch. Now if I don't change branches Apache runs it fine, but if I need to work on another branch while waiting on testing for 101 I can't just work on another branch without it switching in apache to the new branch
16:40 cdown joined #git
16:40 osse pa: this is unrelated to your original question right...?
16:41 overlord_tm joined #git
16:42 grawity Malgorath: so, one that would take files directly from the repo without needing a workingtree?
16:42 _ikke_ Malgorath: Am in understanding that you have a git repository in your apache webroot?
16:42 grawity _ikke_: tbf, it's one of the ¡deploy methods
16:43 _ikke_ grawity: Just trying to understand his question
16:43 Cavallari joined #git
16:43 grawity Malgorath: that's possible but only works well with entirely static sites
16:43 Malgorath _ikke_, no its not in my root, i have several paths in my server right now
16:43 _ikke_ Malgorath: Not sure I follow, but you might want to look into man git work-tree
16:43 gitinfo Malgorath: the git manpage is available at http://jk.gs/git.html
16:43 _ikke_ git worktree
16:43 _ikke_ man git worktree
16:43 gitinfo the git-worktree manpage is available at http://jk.gs/git-worktree.html
16:44 jast yeah, I don't understand the question, either. (why? git doesn't interact with web servers on its own...)
16:45 grawity I've served small static sites out of git, with some custom scripting, but...
16:45 grill ok, so what's the easiest way to see a file's changes before and after a given commit?
16:45 grawity grill: git log -p file.c
16:46 grawity Malgorath: actually maybe you'd find gitfs (FUSE) useful
16:46 jast git show <commit> -- <file>
16:46 zeroed joined #git
16:47 subhojit777 joined #git
16:47 Malgorath jast, here is the scenario: I have a development web server that I work on.  I have multiple projects under a localdomain (10.0.2.150/projecst/*)  now if I am working on project for person 1 it might be pathed 10.0.2.150/projects/person-1/
16:47 grill grawity not sure that command is entirely correct
16:48 raspado joined #git
16:48 Malgorath If I am in /home/malgorath/projects/person-1/ and I git branch here, it will only display the files in that branch not files I have continued to change in another branch, I'd like something that does http://10.0.2.150/malgorath/projects/person-1/malgorath-101/
16:48 Malgorath where malgorath-101 is the branch name
16:49 grawity yeah, use gitfs
16:49 inflames joined #git
16:50 freimatz joined #git
16:50 Malgorath grawity, yeah this looks actually ncie for the fact that sometimes I work with turnips that don't know what git even is.
16:50 HardlySe1n joined #git
16:51 grill ok.so what's the best way to show a git before and after a given commit?
16:51 maret joined #git
16:51 retroj left #git
16:51 osse grill: before:   git show abc123~1:file;  after: git show abc123:file
16:52 osse where abc123 is the commit
16:52 grill well, i really want to same sort of view that i'd get when i did a git diff before a commit
16:52 grill can i see changes to every file that was changed?
16:52 jast git show <commit>
16:53 grill there we go
16:53 grill wonderful
16:53 mischat_ joined #git
16:54 mischat__ joined #git
16:56 nidr0x joined #git
16:57 Snugglebash joined #git
16:57 Snugglebash joined #git
16:58 Snugglebash joined #git
17:00 csk157_ joined #git
17:00 mischat joined #git
17:01 pa osse, no, it was
17:02 pa i resorted to extracting the patch, and applying it to another cloned tree
17:02 pa that worked
17:02 cdleonard joined #git
17:02 pa not sure what got fucked up in my normal tree
17:02 osse the problem is that it doesn't scale
17:02 osse git didn't know what you wanted to push
17:02 osse or, rather where I guess
17:03 osse The solution to that is  git push -u origin rotatingMapGestureArea  and after that  git push  will work as normal
17:03 svm_invictvs joined #git
17:03 thiago joined #git
17:04 rkazak joined #git
17:05 Gsham joined #git
17:06 synthroid joined #git
17:06 nidr0x joined #git
17:08 Electrometro joined #git
17:08 Snugglebash joined #git
17:11 HardlySeen joined #git
17:17 krillr joined #git
17:18 emg joined #git
17:21 csk157_ joined #git
17:23 kulelu88 joined #git
17:23 stick-shift joined #git
17:24 jozwior joined #git
17:24 gautamsomani joined #git
17:26 blackjid joined #git
17:27 chipotle_ joined #git
17:27 nidr0x joined #git
17:33 trasz_ jast: hi, still here?  sorry, i had an unexpected meeting.
17:33 jast trasz_: busy now, sorry
17:33 trasz_ jast: so, wouldn't 'git fsck' find the unreferenced objects?
17:34 Guest54781 joined #git
17:34 jast there are different levels of unreferenced, keyword 'reflogs'
17:34 zeroed joined #git
17:34 zeroed joined #git
17:34 trasz_ jast: also, there was just one commit, and it apparently didn't touch any files at all.
17:36 diogenese joined #git
17:37 svm_invictvs joined #git
17:39 JanC_ joined #git
17:41 z1haze_work left #git
17:43 DolphinDream howdy gitsters
17:43 durham joined #git
17:43 DolphinDream how do i apply commits on one repo on a different clone of the same repo ?
17:44 DolphinDream I have repo1->A->B->C .. and repo2->A-D-E  .. I want to take the B->C commits from repo1 and apply them on top of A in repo2
17:45 _ikke_ DolphinDream: Can you fetch from repo A in repo B
17:50 mischat_ joined #git
17:51 trasz_ jast: damn, figured this out, i just forgot to pass -a to 'git branch'.
17:53 NeverDie joined #git
18:01 jnavila_ joined #git
18:02 gautamsomani left #git
18:02 gautamsomani joined #git
18:03 danslo joined #git
18:05 zeroed joined #git
18:07 sargas joined #git
18:08 HardlySeen joined #git
18:11 jnavila joined #git
18:12 DolphinDream _ikke_: can you do it by creating some patches ? what if the two repos are not on the same machine/network
18:12 Gsham joined #git
18:13 romerocesar joined #git
18:16 Spacew00t joined #git
18:16 duderonomy joined #git
18:20 Gloomy joined #git
18:21 mischat joined #git
18:21 Ahd joined #git
18:21 Gloomy joined #git
18:21 chachasmooth joined #git
18:22 Gloomy joined #git
18:23 judson_ joined #git
18:23 Gloomy joined #git
18:23 sargas joined #git
18:23 Gloomy joined #git
18:24 grill joined #git
18:24 phanimahesh joined #git
18:26 vdv joined #git
18:26 jstimm joined #git
18:28 Mobutils joined #git
18:28 jmpp joined #git
18:29 rorro joined #git
18:29 Gsham joined #git
18:33 chipotle joined #git
18:33 csk157_ joined #git
18:33 Hhks joined #git
18:34 Es0teric joined #git
18:34 thebope joined #git
18:35 theoceaniscool joined #git
18:36 jnavila joined #git
18:43 hahuang65 joined #git
18:43 theoceaniscool_ joined #git
18:43 mehola joined #git
18:44 aw1 joined #git
18:45 madewokherd joined #git
18:45 inflames joined #git
18:45 syg_ joined #git
18:46 m0viefreak joined #git
18:47 zzz joined #git
18:47 jost_ joined #git
18:49 ajf- joined #git
18:51 dsantiago joined #git
18:53 ki0 joined #git
18:53 wget joined #git
18:53 wget joined #git
18:54 Es0teric joined #git
18:56 rkazak joined #git
18:56 marc_v92 joined #git
19:02 druonysus joined #git
19:03 Mikerhinos joined #git
19:06 LionsMane joined #git
19:07 trasz_ left #git
19:08 Gsham joined #git
19:09 romerocesar joined #git
19:10 govg joined #git
19:11 glebihan joined #git
19:11 chipotle joined #git
19:12 Snugglebash joined #git
19:14 ayjay_t joined #git
19:15 Ryanar joined #git
19:15 maret joined #git
19:15 voldz1tic3n joined #git
19:15 PioneerAxon joined #git
19:17 jstein_ joined #git
19:17 Sasazuka joined #git
19:18 Gsham joined #git
19:23 hashpuppy joined #git
19:23 davimore_ joined #git
19:24 ayjay_t joined #git
19:27 acetakwas joined #git
19:29 gautamsomani joined #git
19:32 theoceaniscool joined #git
19:33 finalbeta joined #git
19:34 erebel55 joined #git
19:36 quesker joined #git
19:36 quesker can you have more than one git remote somehow?
19:37 robattila256 joined #git
19:39 nzc joined #git
19:40 jozwior joined #git
19:41 cyphase joined #git
19:42 jozwior joined #git
19:42 nwkj86 joined #git
19:45 ToxicFrog Oh, definitely.
19:45 ToxicFrog I have five in one of my repos.
19:46 ToxicFrog You can use "git remote" to view and manipulate them.
19:46 ToxicFrog (see 'git help remote'/'man git-remote' for details)
19:46 gitinfo the git-remote manpage is available at http://jk.gs/git-remote.html
19:46 sofly joined #git
19:46 sofly joined #git
19:47 mizu_no_oto joined #git
19:48 gugah joined #git
19:50 cwg joined #git
19:51 elsevero joined #git
19:51 voldz1tic3n joined #git
19:53 multi_io_ joined #git
19:53 mischat joined #git
19:55 dreiss joined #git
19:57 cwg Hi, is there a something out there that helps to work on a project (under git) from multiple machines?  Let me explain what I mean:
19:58 cwg The usual solution to the sync problem seems to be to pull/push manually, but this does not sync the stash, nor remotes and other config, nor untracked files.
19:59 warthog9 joined #git
19:59 cwg Alternatively, one can use something like unison to sync the .git database together with the working directory, but while this works when one is disciplined, it's obviously dangerous.
19:59 BlessJah cwg: what you need is more branches
20:00 cwg BlessJah: you mean per machine branches?
20:00 BlessJah no, branches in general
20:00 cwg In my main project I have 24, is this enough? ;-)
20:01 elsevero can anyone help me with a hint? I had one that website bookmarked but can’t find it right now :( its referring to git transferring commits to another user
20:01 BlessJah create a remote you will use to sync your machines so you wouldn't annoy coworkers
20:01 BlessJah cwg: I doubt anyway that you have enough machines for that to be a problem
20:02 cwg I have 2 at home and 2 at work...
20:02 BlessJah and you really use 4 of them to develop that one project?
20:02 BlessJah I mean develop, not test or sth
20:03 cwg No, but it would be cool to know that if my laptop breaks down, I can just sit down at the desktop and continue.
20:03 jnavila joined #git
20:03 ToxicFrog cwg: my approach here is to use syncthing
20:03 cwg I've had an idea in the back of my head for years, perhaps someone here will find it interesting.
20:04 ToxicFrog + syncthing-inotify
20:04 BlessJah cwg: what is your recovery point objective?
20:04 ToxicFrog It's continuous sync, unlike unison, so you don't need to remember to run a thing; all you need to remember is not to edit simultaneously on multiple machines.
20:05 * BlessJah senses !xy
20:05 BlessJah !xy
20:05 gitinfo Woah, slow down for a bit. Are you sure that you need to jump through that particular hoop to achieve your goal? We suspect you don't, so why don't you back up a bit and tell us about the overall objective...
20:05 cwg ToxicFrog: I know about it, and I guess it works if you let it enough time to sync before continuing to work.  It's good to know that it works in practice with git (unison does)
20:06 ToxicFrog cwg: git is just files on disk, any sync program should properly sync your git state
20:06 danslo joined #git
20:06 cwg ToxicFrog: the problem is that it's several files that are inter-related.
20:07 ToxicFrog BlessJah: at least for me, the use case is "not have to care, at all, whether I'm working on my laptop or sshed into my server when I'm working on stuff in ~/devel, while at the same time being able to use the laptop without network connectivity"
20:07 ToxicFrog #1 rules out things that require explicit sync (git push/pull, unison), and #2 rules out just using sshfs.
20:07 druonysus joined #git
20:07 cwg If you change a bit here, and a bit over there, tools like syncthing may merge and break the repo.
20:07 ToxicFrog cwg: I did point out that you have to avoid simultaneous edits in two places
20:07 grawity in my experience, dropbox works as long as you wait for the other end to sync
20:08 BlessJah ToxicFrog: I hate master-master file-level synchronisation, the git is meant to solve "sometimes I work on laptop and sometimes on server" problem
20:08 grawity index messes up quickly, otherwise
20:08 hahuang65 joined #git
20:08 ToxicFrog That said, ST will never attempt to merge file contents; the most recently updated one wins and the other gets written to a .sync-conflict file
20:08 finalbeta joined #git
20:08 ToxicFrog I can also say from personal experience that Dropbox and BTSync also work for this, with the same caveats about simultaneous edits.
20:09 ToxicFrog BlessJah: it does not in fact solve it very well, for the reasons cwg has pointed out
20:09 cwg OK, yes, so sync tools work, but they are not optimal and one has to be careful.  And since the history is already git, it should be possible to build a tool that does the syncing using git.
20:09 ianmethyst joined #git
20:09 anonymus_ joined #git
20:09 cwg I was thinking of some "metagit" that tracks the metadata (remotes, creation and deletion of branches) of several git repos.
20:09 sargas joined #git
20:10 BlessJah this is why I asked about RPO
20:10 ToxicFrog cwg: AIUI you want to sync not just the history but also the stashes, working directory state, etc too, though?
20:10 cube`` joined #git
20:10 BlessJah how much data are you ok to loose
20:10 cube`` joined #git
20:10 cube`` joined #git
20:10 cube`` hey guys, i have github desktop on a mac and i have a forked repo and i made a branch there and did some changes and loaded the foreign changes that happened in the original repo and now when i want to sync my own fork i get "Authentication Failed: You may not have permission to access elixir-school. Check Preferences to make sure you’re still logged in."
20:10 ToxicFrog Or have I misunderstood?
20:10 cwg Perpahs you are aware of git-annex (the annex stuff is not important here).  It has a command "git annex sync" that syncs git branches in a peer-to-peer way.
20:11 cwg It's described here: http://git-annex.branchable.com/sync/
20:11 cwg Again, the point is not the annex, but the peer-to-peer way of using git.
20:12 cwg This scheme seems to work well, but it does not sync the metadata.  For example, there's no way to propagate the deletion of a branch.
20:13 cwg The idea is basically to do a tool that automates the manual pulling/pushing when working with multiple machines.
20:13 Raaa joined #git
20:14 ToxicFrog cwg: from your original statement, though, even that won't be enough.
20:14 cwg why not?
20:15 ToxicFrog You said you wanted to sync not just commits but also the stash, remotes and other config (presumably including hooks and submodules), and untracked files
20:15 ToxicFrog None of which are in scope for push/pull no matter how you automate it
20:16 sq-one joined #git
20:16 ToxicFrog I mean, all of this can be worked around (use temporary branches instead of stashes, commit untracked files to one of these branches before syncing, copy .git/config and .git/hooks into the worktree and do the same there, reverse this procedure when receiving a sync), but at some point you have to ask "wouldn't it be easier to just use a networked filesystem or something"
20:17 cwg ToxicFrog: that's why my idea is to do a meta-repository that keeps this "metadata".  The meta repo is synced, and applied to the real working repos.
20:17 ToxicFrog Aah.
20:17 dreiss joined #git
20:18 cwg ToxicFrog: imagine a repo into which the config files from several repos have been checked in and that is used to sync these.
20:19 cwg ... of course this may be quite complicated in reality.  One may have to normalize the config files in some way before syncing them, for example, and perhaps also remove "local" parts that are specific to a single machine.
20:21 snits joined #git
20:21 hasc_ joined #git
20:21 cwg I find it a bit surprising that no one has implemented something like this, even a simple version.  But they I may miss some problems.
20:21 hasc_ joined #git
20:22 nowhere_man joined #git
20:24 adino joined #git
20:26 Es0teric joined #git
20:27 overlord_tm joined #git
20:27 cyan__ joined #git
20:29 marchelly joined #git
20:29 hasc_ joined #git
20:31 elsevero what are the other options for transferring commits (author) to another user?
20:33 hasc joined #git
20:34 Merh5 joined #git
20:34 imachuchu elsevero: ? Like you want to rewrite a bunch of commits to change their author from one person to another?
20:35 marchelly Hi,
20:35 marchelly When I'm writing git hook post-receive, isthat a way to know that this commit is related to updating .gitignore?
20:35 elsevero imachuchu: yes, author (name & email).
20:35 _ikke_ DolphinDream: Sure. lookt at man git format-patch and man git am
20:35 gitinfo DolphinDream: the git-format-patch manpage is available at http://jk.gs/git-format-patch.html
20:35 gitinfo DolphinDream: the git-am manpage is available at http://jk.gs/git-am.html
20:36 elsevero imachuchu: tried http://stackoverflow.com/a/750182/850491 but still I’m able to see the old author
20:36 marchelly for example for master branch I'm doing this if [[ $ref =~ .*/master$ ]];
20:37 marchelly I want the same but to know, that .gitignore file is updated
20:40 raedah what is a good way to maintain a patch that I want to have handy to apply over another branch?
20:40 _ikke_ marchelly: You nead to read from stdin what refs are updated
20:40 _ikke_ marchelly: Then for each ref, check the diff between the old and new value
20:40 raedah Right now I just have the patch as its own branch, but i am having to do a lot of rebasing and such just to maintain it.
20:41 _ikke_ raedah: man git format-patch
20:41 gitinfo raedah: the git-format-patch manpage is available at http://jk.gs/git-format-patch.html
20:42 elsevero imachuchu: do you have any other idea? know any cli tool
20:42 marchelly _ikke_, that's the simpliest solution? Is in not an overhead?
20:42 _ikke_ marchelly: That's the only solution
20:43 _ikke_ marchelly: If you want to check if there is *any* commit that changes it, you have to iterate each commit too
20:44 marchelly _ikke_, got you. thanks.
20:44 _ikke_ marchelly: note that the hook receives on stdin what refs are being updated
20:44 jimi_ joined #git
20:45 cwg left #git
20:46 marchelly _ikke_, got it. Maybe I should do another way if it's easier and wait for some <actionword> in commit? like UPDATEIGNORE ?
20:46 Atm0spher1c joined #git
20:46 hhee2 joined #git
20:46 rkazak joined #git
20:46 marchelly comment I mean
20:47 judson joined #git
20:47 _ikke_ marchelly: That's about the same 'overhead'
20:47 imachuchu elsevero: did you try the linked github script (https://help.github.com/articles/changing-author-info/)?
20:47 marchelly :)
20:47 ggherdov Hello, what's the way to re-sync all tags with a given remote? say I want to discard all local tags, and get the afresh from remote
20:48 _ikke_ marchelly: if you use git diff --name-only, git can just do a tree diff, which should be fairly cheap
20:48 anonymuse joined #git
20:49 marchelly I'm not sure I can use this. let me think, thank you.
20:50 Guest6615 left #git
20:51 imachuchu elsevero: also note that the old commits will still be around, just that now your branches and tags will be pointing at new ones. The old commits will still be in the repo but will be cleaned up eventually. Also everybody who has cloned the repo will need to pull down the changes and git will whine because history has changed
20:51 _ikke_ ggherdov: git for-each-ref --format='delete %(refname)' refs/tags | git update-ref --stdin
20:51 ariver joined #git
20:51 _ikke_ that will delete all local tags (untested)
20:52 Electrometro joined #git
20:52 elsevero imachuchu: I’m fine with that. someone recommded me a tool, can’t find it right now
20:53 cdg joined #git
20:55 strixx joined #git
20:55 _ikke_ marchelly: something like https://gist.github.com/Ikke/7bb9f213b9bfda89b2bbc43aefd6bfd0
20:56 _ikke_ marchelly: might be edge cases I've not dealth with, but this is the basis
20:56 Jackneill joined #git
20:57 marchelly yes, this make sense, I'll try to use it.
20:58 _ikke_ marchelly: I've updated it
20:59 Jackneill joined #git
21:00 marchelly _ikke_, yep, that could make a trick. I'll test this a little bit later and let you know.
21:01 Kktq9 joined #git
21:02 shgysk8zer0 joined #git
21:04 kevwil joined #git
21:04 nilg joined #git
21:06 mikecmpbll joined #git
21:07 enckse joined #git
21:08 storrgie joined #git
21:10 oaao joined #git
21:12 mentazoom joined #git
21:12 blackwind_123 joined #git
21:13 Guest54781 joined #git
21:14 enckse joined #git
21:14 hasc joined #git
21:15 storrgie joined #git
21:18 pa left #git
21:18 cdown joined #git
21:21 marchelly _ikke_, I'm not sure about this line if git diff -q ${commit}~1 $commit .gitignore
21:22 _ikke_ marchelly: diff the commit against it's parent
21:23 re1 joined #git
21:23 voldz1tic3n joined #git
21:23 _ikke_ -q == --quiet, turning off output and enabling --exit-code
21:24 _ikke_ so if there is a difference, it will return a non-zero exit code
21:26 Sasazuka joined #git
21:30 marchelly _ikke_, I don't know what I'm doing wrong but I got usage: git diff [<options>] [<commit> [<commit>]] [--] [<path>...] on every run of git diff
21:30 theoceaniscool joined #git
21:30 _ikke_ hmm
21:30 marchelly I'm doing git diff -q b605803ad17b8c8d177694a293f10ded2130b062~1 b605803ad17b8c8d177694a293f10ded2130b062 .gitignore
21:31 marchelly adding -- before /gitignore made no sense
21:31 chachasmooth joined #git
21:31 imachuchu marchelly: did you try adding -- before .gitignore?
21:31 _ikke_ marchelly: sorry, -q does not exist. Use --quiet
21:32 marchelly yes
21:32 marchelly this works fine already read man
21:32 marchelly that's works git diff --quiet b605803ad17b8c8d177694a293f10ded2130b062~1 b605803ad17b8c8d177694a293f10ded2130b062 .gitignore
21:32 marchelly even without --
21:32 _ikke_ yeah
21:32 _ikke_ -- is not required in this case
21:33 _ikke_ because there is no ambiguity
21:35 ggherdov _ikke_: thanks for the command to delete tags. To get them from remote I do `git fetch -t <remote>`, right?
21:35 _ikke_ ggherdov: just git fetch should be enough already
21:35 _ikke_ git fetch <remote>
21:35 ggherdov _ikke_: ok
21:37 chosafine joined #git
21:42 SwiftMatt joined #git
21:43 marchelly _ikke_, works. the only thing is that:  if git diff --quiet $oldval $commit .gitignore then echo "ignore not changed"
21:43 glebihan_ joined #git
21:44 marchelly but that's very good trick, not overhead
21:44 _ikke_ marchelly: Ah, right, if ! git diff ..
21:44 _ikke_ marchelly: np
21:45 marchelly yep
21:48 flavi0 joined #git
21:49 d1z joined #git
21:51 minikN joined #git
21:55 thiago joined #git
21:55 hasc_ joined #git
21:55 jeffreylevesque joined #git
21:56 nettoweb joined #git
21:56 rgrinberg joined #git
21:56 rgrinberg I have this generated file that I'm trying to version with git that doesn't have a newline at the end of the file. But everytime I commit it and regenerate, it seems to be added. Anyway to stop this?
21:57 beneter joined #git
21:58 phanimahesh joined #git
21:59 mischat joined #git
21:59 _ikke_ git does not add newlines
21:59 _ikke_ So you need to find out what does
21:59 Dougie187 left #git
21:59 raspado joined #git
22:01 rgrinberg I see. So it's not a setting. I'm committing these files from magit, maybe doing it from the command line will not add it
22:01 rgrinberg I see. So it's not a setting. I'm committing these files from magit, maybe doing it from the command line will not add it
22:02 _ikke_ right
22:03 pbrewczynski joined #git
22:03 Sasazuka_ joined #git
22:07 fatalhalt joined #git
22:08 fatalhalt joined #git
22:08 rgrinberg Damn, something is still adding them
22:11 tang^ generated file? does it really need to be committed?
22:11 Sasazuka joined #git
22:12 rgrinberg tang^: Absolutely.
22:13 tang^ just asking because I don't use git for deployment and generated files here only get deployed
22:13 rgrinberg Argh, i'm an idiot. it wasn't a newline thing just 2 uuid's on the last line. They looked the same to me on first glance
22:15 gusnan joined #git
22:16 gambl0re joined #git
22:17 Anja joined #git
22:17 finalbeta joined #git
22:17 ayjay_t joined #git
22:18 Derperperd joined #git
22:19 Kaisyu joined #git
22:26 mischat joined #git
22:28 ki0 joined #git
22:31 gugah joined #git
22:31 nowhere_man joined #git
22:36 chosafine joined #git
22:36 chosafine joined #git
22:36 cyan__ joined #git
22:38 KV joined #git
22:39 Tanfrieze joined #git
22:39 cxreg2 left #git
22:39 cxreg joined #git
22:39 BSaboia joined #git
22:42 GavinMagnus joined #git
22:45 karstensrage joined #git
22:47 kevwil joined #git
22:47 Sasazuka joined #git
22:51 LaptopBrain joined #git
22:57 kevwil joined #git
22:59 brent__ joined #git
23:01 fmeerkoetter joined #git
23:04 rivarun joined #git
23:08 kevwil joined #git
23:14 anonymuse joined #git
23:18 kevwil joined #git
23:18 ljubak joined #git
23:18 hahuang61 joined #git
23:22 Sasazuka_ joined #git
23:23 snowkidind joined #git
23:25 Derperperd joined #git
23:26 YuGiOhJCJ joined #git
23:28 kevwil joined #git
23:29 cagmz joined #git
23:30 SwiftMatt joined #git
23:32 jimi_ joined #git
23:32 voldz1tic3n joined #git
23:33 durham_ joined #git
23:40 UserOO7 joined #git
23:40 cdown In Tox, is there some reasonable way to "append" another dependency to the default testenv dependencies, or another command to the default testenv's commands?
23:40 ghormoon joined #git
23:42 BlessJah cdown: I believe you wanted to ask in #python, not #git.
23:42 cdown BlessJah: What on earth
23:42 cdown I... don't know how this happened.
23:43 cdown Sorry about that.
23:48 drizbit joined #git
23:49 yqt joined #git
23:54 ljubak joined #git
23:56 thiago joined #git
23:59 bernardio joined #git
23:59 vdv joined #git

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